Commit 03c751a5e10caafbb6d1afcaf1ea67f2153c3193
Exists in
ti-lsk-linux-4.1.y
and in
10 other branches
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs
Pull btrfs fixes from Chris Mason: "None of these are huge, but my commit does fix a regression from 3.18 that could cause lost files during log replay. This also adds Dave Sterba to the list of Btrfs maintainers. It doesn't mean we're doing things differently, but Dave has really been helping with the maintainer workload for years" * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs: Btrfs: don't delay inode ref updates during log replay Btrfs: correctly get tree level in tree_backref_for_extent Btrfs: call inode_dec_link_count() on mkdir error path Btrfs: abort transaction if we don't find the block group Btrfs, scrub: uninitialized variable in scrub_extent_for_parity() Btrfs: add more maintainers
Showing 6 changed files Inline Diff
MAINTAINERS
| 1 | 1 | ||
| 2 | 2 | ||
| 3 | List of maintainers and how to submit kernel changes | 3 | List of maintainers and how to submit kernel changes |
| 4 | 4 | ||
| 5 | Please try to follow the guidelines below. This will make things | 5 | Please try to follow the guidelines below. This will make things |
| 6 | easier on the maintainers. Not all of these guidelines matter for every | 6 | easier on the maintainers. Not all of these guidelines matter for every |
| 7 | trivial patch so apply some common sense. | 7 | trivial patch so apply some common sense. |
| 8 | 8 | ||
| 9 | 1. Always _test_ your changes, however small, on at least 4 or | 9 | 1. Always _test_ your changes, however small, on at least 4 or |
| 10 | 5 people, preferably many more. | 10 | 5 people, preferably many more. |
| 11 | 11 | ||
| 12 | 2. Try to release a few ALPHA test versions to the net. Announce | 12 | 2. Try to release a few ALPHA test versions to the net. Announce |
| 13 | them onto the kernel channel and await results. This is especially | 13 | them onto the kernel channel and await results. This is especially |
| 14 | important for device drivers, because often that's the only way | 14 | important for device drivers, because often that's the only way |
| 15 | you will find things like the fact version 3 firmware needs | 15 | you will find things like the fact version 3 firmware needs |
| 16 | a magic fix you didn't know about, or some clown changed the | 16 | a magic fix you didn't know about, or some clown changed the |
| 17 | chips on a board and not its name. (Don't laugh! Look at the | 17 | chips on a board and not its name. (Don't laugh! Look at the |
| 18 | SMC etherpower for that.) | 18 | SMC etherpower for that.) |
| 19 | 19 | ||
| 20 | 3. Make sure your changes compile correctly in multiple | 20 | 3. Make sure your changes compile correctly in multiple |
| 21 | configurations. In particular check that changes work both as a | 21 | configurations. In particular check that changes work both as a |
| 22 | module and built into the kernel. | 22 | module and built into the kernel. |
| 23 | 23 | ||
| 24 | 4. When you are happy with a change make it generally available for | 24 | 4. When you are happy with a change make it generally available for |
| 25 | testing and await feedback. | 25 | testing and await feedback. |
| 26 | 26 | ||
| 27 | 5. Make a patch available to the relevant maintainer in the list. Use | 27 | 5. Make a patch available to the relevant maintainer in the list. Use |
| 28 | 'diff -u' to make the patch easy to merge. Be prepared to get your | 28 | 'diff -u' to make the patch easy to merge. Be prepared to get your |
| 29 | changes sent back with seemingly silly requests about formatting | 29 | changes sent back with seemingly silly requests about formatting |
| 30 | and variable names. These aren't as silly as they seem. One | 30 | and variable names. These aren't as silly as they seem. One |
| 31 | job the maintainers (and especially Linus) do is to keep things | 31 | job the maintainers (and especially Linus) do is to keep things |
| 32 | looking the same. Sometimes this means that the clever hack in | 32 | looking the same. Sometimes this means that the clever hack in |
| 33 | your driver to get around a problem actually needs to become a | 33 | your driver to get around a problem actually needs to become a |
| 34 | generalized kernel feature ready for next time. | 34 | generalized kernel feature ready for next time. |
| 35 | 35 | ||
| 36 | PLEASE check your patch with the automated style checker | 36 | PLEASE check your patch with the automated style checker |
| 37 | (scripts/checkpatch.pl) to catch trival style violations. | 37 | (scripts/checkpatch.pl) to catch trival style violations. |
| 38 | See Documentation/CodingStyle for guidance here. | 38 | See Documentation/CodingStyle for guidance here. |
| 39 | 39 | ||
| 40 | PLEASE CC: the maintainers and mailing lists that are generated | 40 | PLEASE CC: the maintainers and mailing lists that are generated |
| 41 | by scripts/get_maintainer.pl. The results returned by the | 41 | by scripts/get_maintainer.pl. The results returned by the |
| 42 | script will be best if you have git installed and are making | 42 | script will be best if you have git installed and are making |
| 43 | your changes in a branch derived from Linus' latest git tree. | 43 | your changes in a branch derived from Linus' latest git tree. |
| 44 | See Documentation/SubmittingPatches for details. | 44 | See Documentation/SubmittingPatches for details. |
| 45 | 45 | ||
| 46 | PLEASE try to include any credit lines you want added with the | 46 | PLEASE try to include any credit lines you want added with the |
| 47 | patch. It avoids people being missed off by mistake and makes | 47 | patch. It avoids people being missed off by mistake and makes |
| 48 | it easier to know who wants adding and who doesn't. | 48 | it easier to know who wants adding and who doesn't. |
| 49 | 49 | ||
| 50 | PLEASE document known bugs. If it doesn't work for everything | 50 | PLEASE document known bugs. If it doesn't work for everything |
| 51 | or does something very odd once a month document it. | 51 | or does something very odd once a month document it. |
| 52 | 52 | ||
| 53 | PLEASE remember that submissions must be made under the terms | 53 | PLEASE remember that submissions must be made under the terms |
| 54 | of the OSDL certificate of contribution and should include a | 54 | of the OSDL certificate of contribution and should include a |
| 55 | Signed-off-by: line. The current version of this "Developer's | 55 | Signed-off-by: line. The current version of this "Developer's |
| 56 | Certificate of Origin" (DCO) is listed in the file | 56 | Certificate of Origin" (DCO) is listed in the file |
| 57 | Documentation/SubmittingPatches. | 57 | Documentation/SubmittingPatches. |
| 58 | 58 | ||
| 59 | 6. Make sure you have the right to send any changes you make. If you | 59 | 6. Make sure you have the right to send any changes you make. If you |
| 60 | do changes at work you may find your employer owns the patch | 60 | do changes at work you may find your employer owns the patch |
| 61 | not you. | 61 | not you. |
| 62 | 62 | ||
| 63 | 7. When sending security related changes or reports to a maintainer | 63 | 7. When sending security related changes or reports to a maintainer |
| 64 | please Cc: security@kernel.org, especially if the maintainer | 64 | please Cc: security@kernel.org, especially if the maintainer |
| 65 | does not respond. | 65 | does not respond. |
| 66 | 66 | ||
| 67 | 8. Happy hacking. | 67 | 8. Happy hacking. |
| 68 | 68 | ||
| 69 | Descriptions of section entries: | 69 | Descriptions of section entries: |
| 70 | 70 | ||
| 71 | P: Person (obsolete) | 71 | P: Person (obsolete) |
| 72 | M: Mail patches to: FullName <address@domain> | 72 | M: Mail patches to: FullName <address@domain> |
| 73 | R: Designated reviewer: FullName <address@domain> | 73 | R: Designated reviewer: FullName <address@domain> |
| 74 | These reviewers should be CCed on patches. | 74 | These reviewers should be CCed on patches. |
| 75 | L: Mailing list that is relevant to this area | 75 | L: Mailing list that is relevant to this area |
| 76 | W: Web-page with status/info | 76 | W: Web-page with status/info |
| 77 | Q: Patchwork web based patch tracking system site | 77 | Q: Patchwork web based patch tracking system site |
| 78 | T: SCM tree type and location. | 78 | T: SCM tree type and location. |
| 79 | Type is one of: git, hg, quilt, stgit, topgit | 79 | Type is one of: git, hg, quilt, stgit, topgit |
| 80 | S: Status, one of the following: | 80 | S: Status, one of the following: |
| 81 | Supported: Someone is actually paid to look after this. | 81 | Supported: Someone is actually paid to look after this. |
| 82 | Maintained: Someone actually looks after it. | 82 | Maintained: Someone actually looks after it. |
| 83 | Odd Fixes: It has a maintainer but they don't have time to do | 83 | Odd Fixes: It has a maintainer but they don't have time to do |
| 84 | much other than throw the odd patch in. See below.. | 84 | much other than throw the odd patch in. See below.. |
| 85 | Orphan: No current maintainer [but maybe you could take the | 85 | Orphan: No current maintainer [but maybe you could take the |
| 86 | role as you write your new code]. | 86 | role as you write your new code]. |
| 87 | Obsolete: Old code. Something tagged obsolete generally means | 87 | Obsolete: Old code. Something tagged obsolete generally means |
| 88 | it has been replaced by a better system and you | 88 | it has been replaced by a better system and you |
| 89 | should be using that. | 89 | should be using that. |
| 90 | F: Files and directories with wildcard patterns. | 90 | F: Files and directories with wildcard patterns. |
| 91 | A trailing slash includes all files and subdirectory files. | 91 | A trailing slash includes all files and subdirectory files. |
| 92 | F: drivers/net/ all files in and below drivers/net | 92 | F: drivers/net/ all files in and below drivers/net |
| 93 | F: drivers/net/* all files in drivers/net, but not below | 93 | F: drivers/net/* all files in drivers/net, but not below |
| 94 | F: */net/* all files in "any top level directory"/net | 94 | F: */net/* all files in "any top level directory"/net |
| 95 | One pattern per line. Multiple F: lines acceptable. | 95 | One pattern per line. Multiple F: lines acceptable. |
| 96 | N: Files and directories with regex patterns. | 96 | N: Files and directories with regex patterns. |
| 97 | N: [^a-z]tegra all files whose path contains the word tegra | 97 | N: [^a-z]tegra all files whose path contains the word tegra |
| 98 | One pattern per line. Multiple N: lines acceptable. | 98 | One pattern per line. Multiple N: lines acceptable. |
| 99 | scripts/get_maintainer.pl has different behavior for files that | 99 | scripts/get_maintainer.pl has different behavior for files that |
| 100 | match F: pattern and matches of N: patterns. By default, | 100 | match F: pattern and matches of N: patterns. By default, |
| 101 | get_maintainer will not look at git log history when an F: pattern | 101 | get_maintainer will not look at git log history when an F: pattern |
| 102 | match occurs. When an N: match occurs, git log history is used | 102 | match occurs. When an N: match occurs, git log history is used |
| 103 | to also notify the people that have git commit signatures. | 103 | to also notify the people that have git commit signatures. |
| 104 | X: Files and directories that are NOT maintained, same rules as F: | 104 | X: Files and directories that are NOT maintained, same rules as F: |
| 105 | Files exclusions are tested before file matches. | 105 | Files exclusions are tested before file matches. |
| 106 | Can be useful for excluding a specific subdirectory, for instance: | 106 | Can be useful for excluding a specific subdirectory, for instance: |
| 107 | F: net/ | 107 | F: net/ |
| 108 | X: net/ipv6/ | 108 | X: net/ipv6/ |
| 109 | matches all files in and below net excluding net/ipv6/ | 109 | matches all files in and below net excluding net/ipv6/ |
| 110 | K: Keyword perl extended regex pattern to match content in a | 110 | K: Keyword perl extended regex pattern to match content in a |
| 111 | patch or file. For instance: | 111 | patch or file. For instance: |
| 112 | K: of_get_profile | 112 | K: of_get_profile |
| 113 | matches patches or files that contain "of_get_profile" | 113 | matches patches or files that contain "of_get_profile" |
| 114 | K: \b(printk|pr_(info|err))\b | 114 | K: \b(printk|pr_(info|err))\b |
| 115 | matches patches or files that contain one or more of the words | 115 | matches patches or files that contain one or more of the words |
| 116 | printk, pr_info or pr_err | 116 | printk, pr_info or pr_err |
| 117 | One regex pattern per line. Multiple K: lines acceptable. | 117 | One regex pattern per line. Multiple K: lines acceptable. |
| 118 | 118 | ||
| 119 | Note: For the hard of thinking, this list is meant to remain in alphabetical | 119 | Note: For the hard of thinking, this list is meant to remain in alphabetical |
| 120 | order. If you could add yourselves to it in alphabetical order that would be | 120 | order. If you could add yourselves to it in alphabetical order that would be |
| 121 | so much easier [Ed] | 121 | so much easier [Ed] |
| 122 | 122 | ||
| 123 | Maintainers List (try to look for most precise areas first) | 123 | Maintainers List (try to look for most precise areas first) |
| 124 | 124 | ||
| 125 | ----------------------------------- | 125 | ----------------------------------- |
| 126 | 126 | ||
| 127 | 3C59X NETWORK DRIVER | 127 | 3C59X NETWORK DRIVER |
| 128 | M: Steffen Klassert <klassert@mathematik.tu-chemnitz.de> | 128 | M: Steffen Klassert <klassert@mathematik.tu-chemnitz.de> |
| 129 | L: netdev@vger.kernel.org | 129 | L: netdev@vger.kernel.org |
| 130 | S: Maintained | 130 | S: Maintained |
| 131 | F: Documentation/networking/vortex.txt | 131 | F: Documentation/networking/vortex.txt |
| 132 | F: drivers/net/ethernet/3com/3c59x.c | 132 | F: drivers/net/ethernet/3com/3c59x.c |
| 133 | 133 | ||
| 134 | 3CR990 NETWORK DRIVER | 134 | 3CR990 NETWORK DRIVER |
| 135 | M: David Dillow <dave@thedillows.org> | 135 | M: David Dillow <dave@thedillows.org> |
| 136 | L: netdev@vger.kernel.org | 136 | L: netdev@vger.kernel.org |
| 137 | S: Maintained | 137 | S: Maintained |
| 138 | F: drivers/net/ethernet/3com/typhoon* | 138 | F: drivers/net/ethernet/3com/typhoon* |
| 139 | 139 | ||
| 140 | 3WARE SAS/SATA-RAID SCSI DRIVERS (3W-XXXX, 3W-9XXX, 3W-SAS) | 140 | 3WARE SAS/SATA-RAID SCSI DRIVERS (3W-XXXX, 3W-9XXX, 3W-SAS) |
| 141 | M: Adam Radford <linuxraid@lsi.com> | 141 | M: Adam Radford <linuxraid@lsi.com> |
| 142 | L: linux-scsi@vger.kernel.org | 142 | L: linux-scsi@vger.kernel.org |
| 143 | W: http://www.lsi.com | 143 | W: http://www.lsi.com |
| 144 | S: Supported | 144 | S: Supported |
| 145 | F: drivers/scsi/3w-* | 145 | F: drivers/scsi/3w-* |
| 146 | 146 | ||
| 147 | 53C700 AND 53C700-66 SCSI DRIVER | 147 | 53C700 AND 53C700-66 SCSI DRIVER |
| 148 | M: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> | 148 | M: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> |
| 149 | L: linux-scsi@vger.kernel.org | 149 | L: linux-scsi@vger.kernel.org |
| 150 | S: Maintained | 150 | S: Maintained |
| 151 | F: drivers/scsi/53c700* | 151 | F: drivers/scsi/53c700* |
| 152 | 152 | ||
| 153 | 6LOWPAN GENERIC (BTLE/IEEE 802.15.4) | 153 | 6LOWPAN GENERIC (BTLE/IEEE 802.15.4) |
| 154 | M: Alexander Aring <alex.aring@gmail.com> | 154 | M: Alexander Aring <alex.aring@gmail.com> |
| 155 | M: Jukka Rissanen <jukka.rissanen@linux.intel.com> | 155 | M: Jukka Rissanen <jukka.rissanen@linux.intel.com> |
| 156 | L: linux-bluetooth@vger.kernel.org | 156 | L: linux-bluetooth@vger.kernel.org |
| 157 | L: linux-wpan@vger.kernel.org | 157 | L: linux-wpan@vger.kernel.org |
| 158 | S: Maintained | 158 | S: Maintained |
| 159 | F: net/6lowpan/ | 159 | F: net/6lowpan/ |
| 160 | F: include/net/6lowpan.h | 160 | F: include/net/6lowpan.h |
| 161 | 161 | ||
| 162 | 6PACK NETWORK DRIVER FOR AX.25 | 162 | 6PACK NETWORK DRIVER FOR AX.25 |
| 163 | M: Andreas Koensgen <ajk@comnets.uni-bremen.de> | 163 | M: Andreas Koensgen <ajk@comnets.uni-bremen.de> |
| 164 | L: linux-hams@vger.kernel.org | 164 | L: linux-hams@vger.kernel.org |
| 165 | S: Maintained | 165 | S: Maintained |
| 166 | F: drivers/net/hamradio/6pack.c | 166 | F: drivers/net/hamradio/6pack.c |
| 167 | 167 | ||
| 168 | 8169 10/100/1000 GIGABIT ETHERNET DRIVER | 168 | 8169 10/100/1000 GIGABIT ETHERNET DRIVER |
| 169 | M: Realtek linux nic maintainers <nic_swsd@realtek.com> | 169 | M: Realtek linux nic maintainers <nic_swsd@realtek.com> |
| 170 | L: netdev@vger.kernel.org | 170 | L: netdev@vger.kernel.org |
| 171 | S: Maintained | 171 | S: Maintained |
| 172 | F: drivers/net/ethernet/realtek/r8169.c | 172 | F: drivers/net/ethernet/realtek/r8169.c |
| 173 | 173 | ||
| 174 | 8250/16?50 (AND CLONE UARTS) SERIAL DRIVER | 174 | 8250/16?50 (AND CLONE UARTS) SERIAL DRIVER |
| 175 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 175 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 176 | L: linux-serial@vger.kernel.org | 176 | L: linux-serial@vger.kernel.org |
| 177 | W: http://serial.sourceforge.net | 177 | W: http://serial.sourceforge.net |
| 178 | S: Maintained | 178 | S: Maintained |
| 179 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git | 179 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git |
| 180 | F: drivers/tty/serial/8250* | 180 | F: drivers/tty/serial/8250* |
| 181 | F: include/linux/serial_8250.h | 181 | F: include/linux/serial_8250.h |
| 182 | 182 | ||
| 183 | 8390 NETWORK DRIVERS [WD80x3/SMC-ELITE, SMC-ULTRA, NE2000, 3C503, etc.] | 183 | 8390 NETWORK DRIVERS [WD80x3/SMC-ELITE, SMC-ULTRA, NE2000, 3C503, etc.] |
| 184 | L: netdev@vger.kernel.org | 184 | L: netdev@vger.kernel.org |
| 185 | S: Orphan / Obsolete | 185 | S: Orphan / Obsolete |
| 186 | F: drivers/net/ethernet/8390/ | 186 | F: drivers/net/ethernet/8390/ |
| 187 | 187 | ||
| 188 | 9P FILE SYSTEM | 188 | 9P FILE SYSTEM |
| 189 | M: Eric Van Hensbergen <ericvh@gmail.com> | 189 | M: Eric Van Hensbergen <ericvh@gmail.com> |
| 190 | M: Ron Minnich <rminnich@sandia.gov> | 190 | M: Ron Minnich <rminnich@sandia.gov> |
| 191 | M: Latchesar Ionkov <lucho@ionkov.net> | 191 | M: Latchesar Ionkov <lucho@ionkov.net> |
| 192 | L: v9fs-developer@lists.sourceforge.net | 192 | L: v9fs-developer@lists.sourceforge.net |
| 193 | W: http://swik.net/v9fs | 193 | W: http://swik.net/v9fs |
| 194 | Q: http://patchwork.kernel.org/project/v9fs-devel/list/ | 194 | Q: http://patchwork.kernel.org/project/v9fs-devel/list/ |
| 195 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs.git | 195 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs.git |
| 196 | S: Maintained | 196 | S: Maintained |
| 197 | F: Documentation/filesystems/9p.txt | 197 | F: Documentation/filesystems/9p.txt |
| 198 | F: fs/9p/ | 198 | F: fs/9p/ |
| 199 | F: net/9p/ | 199 | F: net/9p/ |
| 200 | F: include/net/9p/ | 200 | F: include/net/9p/ |
| 201 | F: include/uapi/linux/virtio_9p.h | 201 | F: include/uapi/linux/virtio_9p.h |
| 202 | F: include/trace/events/9p.h | 202 | F: include/trace/events/9p.h |
| 203 | 203 | ||
| 204 | 204 | ||
| 205 | A8293 MEDIA DRIVER | 205 | A8293 MEDIA DRIVER |
| 206 | M: Antti Palosaari <crope@iki.fi> | 206 | M: Antti Palosaari <crope@iki.fi> |
| 207 | L: linux-media@vger.kernel.org | 207 | L: linux-media@vger.kernel.org |
| 208 | W: http://linuxtv.org/ | 208 | W: http://linuxtv.org/ |
| 209 | W: http://palosaari.fi/linux/ | 209 | W: http://palosaari.fi/linux/ |
| 210 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 210 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 211 | T: git git://linuxtv.org/anttip/media_tree.git | 211 | T: git git://linuxtv.org/anttip/media_tree.git |
| 212 | S: Maintained | 212 | S: Maintained |
| 213 | F: drivers/media/dvb-frontends/a8293* | 213 | F: drivers/media/dvb-frontends/a8293* |
| 214 | 214 | ||
| 215 | AACRAID SCSI RAID DRIVER | 215 | AACRAID SCSI RAID DRIVER |
| 216 | M: Adaptec OEM Raid Solutions <aacraid@adaptec.com> | 216 | M: Adaptec OEM Raid Solutions <aacraid@adaptec.com> |
| 217 | L: linux-scsi@vger.kernel.org | 217 | L: linux-scsi@vger.kernel.org |
| 218 | W: http://www.adaptec.com/ | 218 | W: http://www.adaptec.com/ |
| 219 | S: Supported | 219 | S: Supported |
| 220 | F: Documentation/scsi/aacraid.txt | 220 | F: Documentation/scsi/aacraid.txt |
| 221 | F: drivers/scsi/aacraid/ | 221 | F: drivers/scsi/aacraid/ |
| 222 | 222 | ||
| 223 | ABI/API | 223 | ABI/API |
| 224 | L: linux-api@vger.kernel.org | 224 | L: linux-api@vger.kernel.org |
| 225 | F: Documentation/ABI/ | 225 | F: Documentation/ABI/ |
| 226 | F: include/linux/syscalls.h | 226 | F: include/linux/syscalls.h |
| 227 | F: include/uapi/ | 227 | F: include/uapi/ |
| 228 | F: kernel/sys_ni.c | 228 | F: kernel/sys_ni.c |
| 229 | 229 | ||
| 230 | ABIT UGURU 1,2 HARDWARE MONITOR DRIVER | 230 | ABIT UGURU 1,2 HARDWARE MONITOR DRIVER |
| 231 | M: Hans de Goede <hdegoede@redhat.com> | 231 | M: Hans de Goede <hdegoede@redhat.com> |
| 232 | L: lm-sensors@lm-sensors.org | 232 | L: lm-sensors@lm-sensors.org |
| 233 | S: Maintained | 233 | S: Maintained |
| 234 | F: drivers/hwmon/abituguru.c | 234 | F: drivers/hwmon/abituguru.c |
| 235 | 235 | ||
| 236 | ABIT UGURU 3 HARDWARE MONITOR DRIVER | 236 | ABIT UGURU 3 HARDWARE MONITOR DRIVER |
| 237 | M: Alistair John Strachan <alistair@devzero.co.uk> | 237 | M: Alistair John Strachan <alistair@devzero.co.uk> |
| 238 | L: lm-sensors@lm-sensors.org | 238 | L: lm-sensors@lm-sensors.org |
| 239 | S: Maintained | 239 | S: Maintained |
| 240 | F: drivers/hwmon/abituguru3.c | 240 | F: drivers/hwmon/abituguru3.c |
| 241 | 241 | ||
| 242 | ACENIC DRIVER | 242 | ACENIC DRIVER |
| 243 | M: Jes Sorensen <jes@trained-monkey.org> | 243 | M: Jes Sorensen <jes@trained-monkey.org> |
| 244 | L: linux-acenic@sunsite.dk | 244 | L: linux-acenic@sunsite.dk |
| 245 | S: Maintained | 245 | S: Maintained |
| 246 | F: drivers/net/ethernet/alteon/acenic* | 246 | F: drivers/net/ethernet/alteon/acenic* |
| 247 | 247 | ||
| 248 | ACER ASPIRE ONE TEMPERATURE AND FAN DRIVER | 248 | ACER ASPIRE ONE TEMPERATURE AND FAN DRIVER |
| 249 | M: Peter Feuerer <peter@piie.net> | 249 | M: Peter Feuerer <peter@piie.net> |
| 250 | L: platform-driver-x86@vger.kernel.org | 250 | L: platform-driver-x86@vger.kernel.org |
| 251 | W: http://piie.net/?section=acerhdf | 251 | W: http://piie.net/?section=acerhdf |
| 252 | S: Maintained | 252 | S: Maintained |
| 253 | F: drivers/platform/x86/acerhdf.c | 253 | F: drivers/platform/x86/acerhdf.c |
| 254 | 254 | ||
| 255 | ACER WMI LAPTOP EXTRAS | 255 | ACER WMI LAPTOP EXTRAS |
| 256 | M: "Lee, Chun-Yi" <jlee@suse.com> | 256 | M: "Lee, Chun-Yi" <jlee@suse.com> |
| 257 | L: platform-driver-x86@vger.kernel.org | 257 | L: platform-driver-x86@vger.kernel.org |
| 258 | S: Maintained | 258 | S: Maintained |
| 259 | F: drivers/platform/x86/acer-wmi.c | 259 | F: drivers/platform/x86/acer-wmi.c |
| 260 | 260 | ||
| 261 | ACPI | 261 | ACPI |
| 262 | M: Rafael J. Wysocki <rjw@rjwysocki.net> | 262 | M: Rafael J. Wysocki <rjw@rjwysocki.net> |
| 263 | M: Len Brown <lenb@kernel.org> | 263 | M: Len Brown <lenb@kernel.org> |
| 264 | L: linux-acpi@vger.kernel.org | 264 | L: linux-acpi@vger.kernel.org |
| 265 | W: https://01.org/linux-acpi | 265 | W: https://01.org/linux-acpi |
| 266 | Q: https://patchwork.kernel.org/project/linux-acpi/list/ | 266 | Q: https://patchwork.kernel.org/project/linux-acpi/list/ |
| 267 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm | 267 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm |
| 268 | S: Supported | 268 | S: Supported |
| 269 | F: drivers/acpi/ | 269 | F: drivers/acpi/ |
| 270 | F: drivers/pnp/pnpacpi/ | 270 | F: drivers/pnp/pnpacpi/ |
| 271 | F: include/linux/acpi.h | 271 | F: include/linux/acpi.h |
| 272 | F: include/acpi/ | 272 | F: include/acpi/ |
| 273 | F: Documentation/acpi | 273 | F: Documentation/acpi |
| 274 | F: Documentation/ABI/testing/sysfs-bus-acpi | 274 | F: Documentation/ABI/testing/sysfs-bus-acpi |
| 275 | F: drivers/pci/*acpi* | 275 | F: drivers/pci/*acpi* |
| 276 | F: drivers/pci/*/*acpi* | 276 | F: drivers/pci/*/*acpi* |
| 277 | F: drivers/pci/*/*/*acpi* | 277 | F: drivers/pci/*/*/*acpi* |
| 278 | F: tools/power/acpi | 278 | F: tools/power/acpi |
| 279 | 279 | ||
| 280 | ACPI COMPONENT ARCHITECTURE (ACPICA) | 280 | ACPI COMPONENT ARCHITECTURE (ACPICA) |
| 281 | M: Robert Moore <robert.moore@intel.com> | 281 | M: Robert Moore <robert.moore@intel.com> |
| 282 | M: Lv Zheng <lv.zheng@intel.com> | 282 | M: Lv Zheng <lv.zheng@intel.com> |
| 283 | M: Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 283 | M: Rafael J. Wysocki <rafael.j.wysocki@intel.com> |
| 284 | L: linux-acpi@vger.kernel.org | 284 | L: linux-acpi@vger.kernel.org |
| 285 | L: devel@acpica.org | 285 | L: devel@acpica.org |
| 286 | W: https://acpica.org/ | 286 | W: https://acpica.org/ |
| 287 | W: https://github.com/acpica/acpica/ | 287 | W: https://github.com/acpica/acpica/ |
| 288 | Q: https://patchwork.kernel.org/project/linux-acpi/list/ | 288 | Q: https://patchwork.kernel.org/project/linux-acpi/list/ |
| 289 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm | 289 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm |
| 290 | S: Supported | 290 | S: Supported |
| 291 | F: drivers/acpi/acpica/ | 291 | F: drivers/acpi/acpica/ |
| 292 | F: include/acpi/ | 292 | F: include/acpi/ |
| 293 | F: tools/power/acpi/ | 293 | F: tools/power/acpi/ |
| 294 | 294 | ||
| 295 | ACPI FAN DRIVER | 295 | ACPI FAN DRIVER |
| 296 | M: Zhang Rui <rui.zhang@intel.com> | 296 | M: Zhang Rui <rui.zhang@intel.com> |
| 297 | L: linux-acpi@vger.kernel.org | 297 | L: linux-acpi@vger.kernel.org |
| 298 | W: https://01.org/linux-acpi | 298 | W: https://01.org/linux-acpi |
| 299 | S: Supported | 299 | S: Supported |
| 300 | F: drivers/acpi/fan.c | 300 | F: drivers/acpi/fan.c |
| 301 | 301 | ||
| 302 | ACPI THERMAL DRIVER | 302 | ACPI THERMAL DRIVER |
| 303 | M: Zhang Rui <rui.zhang@intel.com> | 303 | M: Zhang Rui <rui.zhang@intel.com> |
| 304 | L: linux-acpi@vger.kernel.org | 304 | L: linux-acpi@vger.kernel.org |
| 305 | W: https://01.org/linux-acpi | 305 | W: https://01.org/linux-acpi |
| 306 | S: Supported | 306 | S: Supported |
| 307 | F: drivers/acpi/*thermal* | 307 | F: drivers/acpi/*thermal* |
| 308 | 308 | ||
| 309 | ACPI VIDEO DRIVER | 309 | ACPI VIDEO DRIVER |
| 310 | M: Zhang Rui <rui.zhang@intel.com> | 310 | M: Zhang Rui <rui.zhang@intel.com> |
| 311 | L: linux-acpi@vger.kernel.org | 311 | L: linux-acpi@vger.kernel.org |
| 312 | W: https://01.org/linux-acpi | 312 | W: https://01.org/linux-acpi |
| 313 | S: Supported | 313 | S: Supported |
| 314 | F: drivers/acpi/video.c | 314 | F: drivers/acpi/video.c |
| 315 | 315 | ||
| 316 | ACPI WMI DRIVER | 316 | ACPI WMI DRIVER |
| 317 | L: platform-driver-x86@vger.kernel.org | 317 | L: platform-driver-x86@vger.kernel.org |
| 318 | S: Orphan | 318 | S: Orphan |
| 319 | F: drivers/platform/x86/wmi.c | 319 | F: drivers/platform/x86/wmi.c |
| 320 | 320 | ||
| 321 | AD1889 ALSA SOUND DRIVER | 321 | AD1889 ALSA SOUND DRIVER |
| 322 | M: Thibaut Varene <T-Bone@parisc-linux.org> | 322 | M: Thibaut Varene <T-Bone@parisc-linux.org> |
| 323 | W: http://wiki.parisc-linux.org/AD1889 | 323 | W: http://wiki.parisc-linux.org/AD1889 |
| 324 | L: linux-parisc@vger.kernel.org | 324 | L: linux-parisc@vger.kernel.org |
| 325 | S: Maintained | 325 | S: Maintained |
| 326 | F: sound/pci/ad1889.* | 326 | F: sound/pci/ad1889.* |
| 327 | 327 | ||
| 328 | AD525X ANALOG DEVICES DIGITAL POTENTIOMETERS DRIVER | 328 | AD525X ANALOG DEVICES DIGITAL POTENTIOMETERS DRIVER |
| 329 | M: Michael Hennerich <michael.hennerich@analog.com> | 329 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 330 | W: http://wiki.analog.com/AD5254 | 330 | W: http://wiki.analog.com/AD5254 |
| 331 | W: http://ez.analog.com/community/linux-device-drivers | 331 | W: http://ez.analog.com/community/linux-device-drivers |
| 332 | S: Supported | 332 | S: Supported |
| 333 | F: drivers/misc/ad525x_dpot.c | 333 | F: drivers/misc/ad525x_dpot.c |
| 334 | 334 | ||
| 335 | AD5398 CURRENT REGULATOR DRIVER (AD5398/AD5821) | 335 | AD5398 CURRENT REGULATOR DRIVER (AD5398/AD5821) |
| 336 | M: Michael Hennerich <michael.hennerich@analog.com> | 336 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 337 | W: http://wiki.analog.com/AD5398 | 337 | W: http://wiki.analog.com/AD5398 |
| 338 | W: http://ez.analog.com/community/linux-device-drivers | 338 | W: http://ez.analog.com/community/linux-device-drivers |
| 339 | S: Supported | 339 | S: Supported |
| 340 | F: drivers/regulator/ad5398.c | 340 | F: drivers/regulator/ad5398.c |
| 341 | 341 | ||
| 342 | AD714X CAPACITANCE TOUCH SENSOR DRIVER (AD7142/3/7/8/7A) | 342 | AD714X CAPACITANCE TOUCH SENSOR DRIVER (AD7142/3/7/8/7A) |
| 343 | M: Michael Hennerich <michael.hennerich@analog.com> | 343 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 344 | W: http://wiki.analog.com/AD7142 | 344 | W: http://wiki.analog.com/AD7142 |
| 345 | W: http://ez.analog.com/community/linux-device-drivers | 345 | W: http://ez.analog.com/community/linux-device-drivers |
| 346 | S: Supported | 346 | S: Supported |
| 347 | F: drivers/input/misc/ad714x.c | 347 | F: drivers/input/misc/ad714x.c |
| 348 | 348 | ||
| 349 | AD7877 TOUCHSCREEN DRIVER | 349 | AD7877 TOUCHSCREEN DRIVER |
| 350 | M: Michael Hennerich <michael.hennerich@analog.com> | 350 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 351 | W: http://wiki.analog.com/AD7877 | 351 | W: http://wiki.analog.com/AD7877 |
| 352 | W: http://ez.analog.com/community/linux-device-drivers | 352 | W: http://ez.analog.com/community/linux-device-drivers |
| 353 | S: Supported | 353 | S: Supported |
| 354 | F: drivers/input/touchscreen/ad7877.c | 354 | F: drivers/input/touchscreen/ad7877.c |
| 355 | 355 | ||
| 356 | AD7879 TOUCHSCREEN DRIVER (AD7879/AD7889) | 356 | AD7879 TOUCHSCREEN DRIVER (AD7879/AD7889) |
| 357 | M: Michael Hennerich <michael.hennerich@analog.com> | 357 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 358 | W: http://wiki.analog.com/AD7879 | 358 | W: http://wiki.analog.com/AD7879 |
| 359 | W: http://ez.analog.com/community/linux-device-drivers | 359 | W: http://ez.analog.com/community/linux-device-drivers |
| 360 | S: Supported | 360 | S: Supported |
| 361 | F: drivers/input/touchscreen/ad7879.c | 361 | F: drivers/input/touchscreen/ad7879.c |
| 362 | 362 | ||
| 363 | ADDRESS SPACE LAYOUT RANDOMIZATION (ASLR) | 363 | ADDRESS SPACE LAYOUT RANDOMIZATION (ASLR) |
| 364 | M: Jiri Kosina <jkosina@suse.cz> | 364 | M: Jiri Kosina <jkosina@suse.cz> |
| 365 | S: Maintained | 365 | S: Maintained |
| 366 | 366 | ||
| 367 | ADM1025 HARDWARE MONITOR DRIVER | 367 | ADM1025 HARDWARE MONITOR DRIVER |
| 368 | M: Jean Delvare <jdelvare@suse.de> | 368 | M: Jean Delvare <jdelvare@suse.de> |
| 369 | L: lm-sensors@lm-sensors.org | 369 | L: lm-sensors@lm-sensors.org |
| 370 | S: Maintained | 370 | S: Maintained |
| 371 | F: Documentation/hwmon/adm1025 | 371 | F: Documentation/hwmon/adm1025 |
| 372 | F: drivers/hwmon/adm1025.c | 372 | F: drivers/hwmon/adm1025.c |
| 373 | 373 | ||
| 374 | ADM1029 HARDWARE MONITOR DRIVER | 374 | ADM1029 HARDWARE MONITOR DRIVER |
| 375 | M: Corentin Labbe <clabbe.montjoie@gmail.com> | 375 | M: Corentin Labbe <clabbe.montjoie@gmail.com> |
| 376 | L: lm-sensors@lm-sensors.org | 376 | L: lm-sensors@lm-sensors.org |
| 377 | S: Maintained | 377 | S: Maintained |
| 378 | F: drivers/hwmon/adm1029.c | 378 | F: drivers/hwmon/adm1029.c |
| 379 | 379 | ||
| 380 | ADM8211 WIRELESS DRIVER | 380 | ADM8211 WIRELESS DRIVER |
| 381 | L: linux-wireless@vger.kernel.org | 381 | L: linux-wireless@vger.kernel.org |
| 382 | W: http://wireless.kernel.org/ | 382 | W: http://wireless.kernel.org/ |
| 383 | S: Orphan | 383 | S: Orphan |
| 384 | F: drivers/net/wireless/adm8211.* | 384 | F: drivers/net/wireless/adm8211.* |
| 385 | 385 | ||
| 386 | ADP1653 FLASH CONTROLLER DRIVER | 386 | ADP1653 FLASH CONTROLLER DRIVER |
| 387 | M: Sakari Ailus <sakari.ailus@iki.fi> | 387 | M: Sakari Ailus <sakari.ailus@iki.fi> |
| 388 | L: linux-media@vger.kernel.org | 388 | L: linux-media@vger.kernel.org |
| 389 | S: Maintained | 389 | S: Maintained |
| 390 | F: drivers/media/i2c/adp1653.c | 390 | F: drivers/media/i2c/adp1653.c |
| 391 | F: include/media/adp1653.h | 391 | F: include/media/adp1653.h |
| 392 | 392 | ||
| 393 | ADP5520 BACKLIGHT DRIVER WITH IO EXPANDER (ADP5520/ADP5501) | 393 | ADP5520 BACKLIGHT DRIVER WITH IO EXPANDER (ADP5520/ADP5501) |
| 394 | M: Michael Hennerich <michael.hennerich@analog.com> | 394 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 395 | W: http://wiki.analog.com/ADP5520 | 395 | W: http://wiki.analog.com/ADP5520 |
| 396 | W: http://ez.analog.com/community/linux-device-drivers | 396 | W: http://ez.analog.com/community/linux-device-drivers |
| 397 | S: Supported | 397 | S: Supported |
| 398 | F: drivers/mfd/adp5520.c | 398 | F: drivers/mfd/adp5520.c |
| 399 | F: drivers/video/backlight/adp5520_bl.c | 399 | F: drivers/video/backlight/adp5520_bl.c |
| 400 | F: drivers/leds/leds-adp5520.c | 400 | F: drivers/leds/leds-adp5520.c |
| 401 | F: drivers/gpio/gpio-adp5520.c | 401 | F: drivers/gpio/gpio-adp5520.c |
| 402 | F: drivers/input/keyboard/adp5520-keys.c | 402 | F: drivers/input/keyboard/adp5520-keys.c |
| 403 | 403 | ||
| 404 | ADP5588 QWERTY KEYPAD AND IO EXPANDER DRIVER (ADP5588/ADP5587) | 404 | ADP5588 QWERTY KEYPAD AND IO EXPANDER DRIVER (ADP5588/ADP5587) |
| 405 | M: Michael Hennerich <michael.hennerich@analog.com> | 405 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 406 | W: http://wiki.analog.com/ADP5588 | 406 | W: http://wiki.analog.com/ADP5588 |
| 407 | W: http://ez.analog.com/community/linux-device-drivers | 407 | W: http://ez.analog.com/community/linux-device-drivers |
| 408 | S: Supported | 408 | S: Supported |
| 409 | F: drivers/input/keyboard/adp5588-keys.c | 409 | F: drivers/input/keyboard/adp5588-keys.c |
| 410 | F: drivers/gpio/gpio-adp5588.c | 410 | F: drivers/gpio/gpio-adp5588.c |
| 411 | 411 | ||
| 412 | ADP8860 BACKLIGHT DRIVER (ADP8860/ADP8861/ADP8863) | 412 | ADP8860 BACKLIGHT DRIVER (ADP8860/ADP8861/ADP8863) |
| 413 | M: Michael Hennerich <michael.hennerich@analog.com> | 413 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 414 | W: http://wiki.analog.com/ADP8860 | 414 | W: http://wiki.analog.com/ADP8860 |
| 415 | W: http://ez.analog.com/community/linux-device-drivers | 415 | W: http://ez.analog.com/community/linux-device-drivers |
| 416 | S: Supported | 416 | S: Supported |
| 417 | F: drivers/video/backlight/adp8860_bl.c | 417 | F: drivers/video/backlight/adp8860_bl.c |
| 418 | 418 | ||
| 419 | ADS1015 HARDWARE MONITOR DRIVER | 419 | ADS1015 HARDWARE MONITOR DRIVER |
| 420 | M: Dirk Eibach <eibach@gdsys.de> | 420 | M: Dirk Eibach <eibach@gdsys.de> |
| 421 | L: lm-sensors@lm-sensors.org | 421 | L: lm-sensors@lm-sensors.org |
| 422 | S: Maintained | 422 | S: Maintained |
| 423 | F: Documentation/hwmon/ads1015 | 423 | F: Documentation/hwmon/ads1015 |
| 424 | F: drivers/hwmon/ads1015.c | 424 | F: drivers/hwmon/ads1015.c |
| 425 | F: include/linux/i2c/ads1015.h | 425 | F: include/linux/i2c/ads1015.h |
| 426 | 426 | ||
| 427 | ADT746X FAN DRIVER | 427 | ADT746X FAN DRIVER |
| 428 | M: Colin Leroy <colin@colino.net> | 428 | M: Colin Leroy <colin@colino.net> |
| 429 | S: Maintained | 429 | S: Maintained |
| 430 | F: drivers/macintosh/therm_adt746x.c | 430 | F: drivers/macintosh/therm_adt746x.c |
| 431 | 431 | ||
| 432 | ADT7475 HARDWARE MONITOR DRIVER | 432 | ADT7475 HARDWARE MONITOR DRIVER |
| 433 | M: Jean Delvare <jdelvare@suse.de> | 433 | M: Jean Delvare <jdelvare@suse.de> |
| 434 | L: lm-sensors@lm-sensors.org | 434 | L: lm-sensors@lm-sensors.org |
| 435 | S: Maintained | 435 | S: Maintained |
| 436 | F: Documentation/hwmon/adt7475 | 436 | F: Documentation/hwmon/adt7475 |
| 437 | F: drivers/hwmon/adt7475.c | 437 | F: drivers/hwmon/adt7475.c |
| 438 | 438 | ||
| 439 | ADXL34X THREE-AXIS DIGITAL ACCELEROMETER DRIVER (ADXL345/ADXL346) | 439 | ADXL34X THREE-AXIS DIGITAL ACCELEROMETER DRIVER (ADXL345/ADXL346) |
| 440 | M: Michael Hennerich <michael.hennerich@analog.com> | 440 | M: Michael Hennerich <michael.hennerich@analog.com> |
| 441 | W: http://wiki.analog.com/ADXL345 | 441 | W: http://wiki.analog.com/ADXL345 |
| 442 | W: http://ez.analog.com/community/linux-device-drivers | 442 | W: http://ez.analog.com/community/linux-device-drivers |
| 443 | S: Supported | 443 | S: Supported |
| 444 | F: drivers/input/misc/adxl34x.c | 444 | F: drivers/input/misc/adxl34x.c |
| 445 | 445 | ||
| 446 | ADVANSYS SCSI DRIVER | 446 | ADVANSYS SCSI DRIVER |
| 447 | M: Matthew Wilcox <matthew@wil.cx> | 447 | M: Matthew Wilcox <matthew@wil.cx> |
| 448 | L: linux-scsi@vger.kernel.org | 448 | L: linux-scsi@vger.kernel.org |
| 449 | S: Maintained | 449 | S: Maintained |
| 450 | F: Documentation/scsi/advansys.txt | 450 | F: Documentation/scsi/advansys.txt |
| 451 | F: drivers/scsi/advansys.c | 451 | F: drivers/scsi/advansys.c |
| 452 | 452 | ||
| 453 | AEDSP16 DRIVER | 453 | AEDSP16 DRIVER |
| 454 | M: Riccardo Facchetti <fizban@tin.it> | 454 | M: Riccardo Facchetti <fizban@tin.it> |
| 455 | S: Maintained | 455 | S: Maintained |
| 456 | F: sound/oss/aedsp16.c | 456 | F: sound/oss/aedsp16.c |
| 457 | 457 | ||
| 458 | AF9013 MEDIA DRIVER | 458 | AF9013 MEDIA DRIVER |
| 459 | M: Antti Palosaari <crope@iki.fi> | 459 | M: Antti Palosaari <crope@iki.fi> |
| 460 | L: linux-media@vger.kernel.org | 460 | L: linux-media@vger.kernel.org |
| 461 | W: http://linuxtv.org/ | 461 | W: http://linuxtv.org/ |
| 462 | W: http://palosaari.fi/linux/ | 462 | W: http://palosaari.fi/linux/ |
| 463 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 463 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 464 | T: git git://linuxtv.org/anttip/media_tree.git | 464 | T: git git://linuxtv.org/anttip/media_tree.git |
| 465 | S: Maintained | 465 | S: Maintained |
| 466 | F: drivers/media/dvb-frontends/af9013* | 466 | F: drivers/media/dvb-frontends/af9013* |
| 467 | 467 | ||
| 468 | AF9033 MEDIA DRIVER | 468 | AF9033 MEDIA DRIVER |
| 469 | M: Antti Palosaari <crope@iki.fi> | 469 | M: Antti Palosaari <crope@iki.fi> |
| 470 | L: linux-media@vger.kernel.org | 470 | L: linux-media@vger.kernel.org |
| 471 | W: http://linuxtv.org/ | 471 | W: http://linuxtv.org/ |
| 472 | W: http://palosaari.fi/linux/ | 472 | W: http://palosaari.fi/linux/ |
| 473 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 473 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 474 | T: git git://linuxtv.org/anttip/media_tree.git | 474 | T: git git://linuxtv.org/anttip/media_tree.git |
| 475 | S: Maintained | 475 | S: Maintained |
| 476 | F: drivers/media/dvb-frontends/af9033* | 476 | F: drivers/media/dvb-frontends/af9033* |
| 477 | 477 | ||
| 478 | AFFS FILE SYSTEM | 478 | AFFS FILE SYSTEM |
| 479 | L: linux-fsdevel@vger.kernel.org | 479 | L: linux-fsdevel@vger.kernel.org |
| 480 | S: Orphan | 480 | S: Orphan |
| 481 | F: Documentation/filesystems/affs.txt | 481 | F: Documentation/filesystems/affs.txt |
| 482 | F: fs/affs/ | 482 | F: fs/affs/ |
| 483 | 483 | ||
| 484 | AFS FILESYSTEM & AF_RXRPC SOCKET DOMAIN | 484 | AFS FILESYSTEM & AF_RXRPC SOCKET DOMAIN |
| 485 | M: David Howells <dhowells@redhat.com> | 485 | M: David Howells <dhowells@redhat.com> |
| 486 | L: linux-afs@lists.infradead.org | 486 | L: linux-afs@lists.infradead.org |
| 487 | S: Supported | 487 | S: Supported |
| 488 | F: fs/afs/ | 488 | F: fs/afs/ |
| 489 | F: include/net/af_rxrpc.h | 489 | F: include/net/af_rxrpc.h |
| 490 | F: net/rxrpc/af_rxrpc.c | 490 | F: net/rxrpc/af_rxrpc.c |
| 491 | 491 | ||
| 492 | AGPGART DRIVER | 492 | AGPGART DRIVER |
| 493 | M: David Airlie <airlied@linux.ie> | 493 | M: David Airlie <airlied@linux.ie> |
| 494 | T: git git://people.freedesktop.org/~airlied/linux (part of drm maint) | 494 | T: git git://people.freedesktop.org/~airlied/linux (part of drm maint) |
| 495 | S: Maintained | 495 | S: Maintained |
| 496 | F: drivers/char/agp/ | 496 | F: drivers/char/agp/ |
| 497 | F: include/linux/agp* | 497 | F: include/linux/agp* |
| 498 | F: include/uapi/linux/agp* | 498 | F: include/uapi/linux/agp* |
| 499 | 499 | ||
| 500 | AHA152X SCSI DRIVER | 500 | AHA152X SCSI DRIVER |
| 501 | M: "Juergen E. Fischer" <fischer@norbit.de> | 501 | M: "Juergen E. Fischer" <fischer@norbit.de> |
| 502 | L: linux-scsi@vger.kernel.org | 502 | L: linux-scsi@vger.kernel.org |
| 503 | S: Maintained | 503 | S: Maintained |
| 504 | F: drivers/scsi/aha152x* | 504 | F: drivers/scsi/aha152x* |
| 505 | F: drivers/scsi/pcmcia/aha152x* | 505 | F: drivers/scsi/pcmcia/aha152x* |
| 506 | 506 | ||
| 507 | AIC7XXX / AIC79XX SCSI DRIVER | 507 | AIC7XXX / AIC79XX SCSI DRIVER |
| 508 | M: Hannes Reinecke <hare@suse.de> | 508 | M: Hannes Reinecke <hare@suse.de> |
| 509 | L: linux-scsi@vger.kernel.org | 509 | L: linux-scsi@vger.kernel.org |
| 510 | S: Maintained | 510 | S: Maintained |
| 511 | F: drivers/scsi/aic7xxx/ | 511 | F: drivers/scsi/aic7xxx/ |
| 512 | 512 | ||
| 513 | AIMSLAB FM RADIO RECEIVER DRIVER | 513 | AIMSLAB FM RADIO RECEIVER DRIVER |
| 514 | M: Hans Verkuil <hverkuil@xs4all.nl> | 514 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 515 | L: linux-media@vger.kernel.org | 515 | L: linux-media@vger.kernel.org |
| 516 | T: git git://linuxtv.org/media_tree.git | 516 | T: git git://linuxtv.org/media_tree.git |
| 517 | W: http://linuxtv.org | 517 | W: http://linuxtv.org |
| 518 | S: Maintained | 518 | S: Maintained |
| 519 | F: drivers/media/radio/radio-aimslab* | 519 | F: drivers/media/radio/radio-aimslab* |
| 520 | 520 | ||
| 521 | AIO | 521 | AIO |
| 522 | M: Benjamin LaHaise <bcrl@kvack.org> | 522 | M: Benjamin LaHaise <bcrl@kvack.org> |
| 523 | L: linux-aio@kvack.org | 523 | L: linux-aio@kvack.org |
| 524 | S: Supported | 524 | S: Supported |
| 525 | F: fs/aio.c | 525 | F: fs/aio.c |
| 526 | F: include/linux/*aio*.h | 526 | F: include/linux/*aio*.h |
| 527 | 527 | ||
| 528 | AIRSPY MEDIA DRIVER | 528 | AIRSPY MEDIA DRIVER |
| 529 | M: Antti Palosaari <crope@iki.fi> | 529 | M: Antti Palosaari <crope@iki.fi> |
| 530 | L: linux-media@vger.kernel.org | 530 | L: linux-media@vger.kernel.org |
| 531 | W: http://linuxtv.org/ | 531 | W: http://linuxtv.org/ |
| 532 | W: http://palosaari.fi/linux/ | 532 | W: http://palosaari.fi/linux/ |
| 533 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 533 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 534 | T: git git://linuxtv.org/anttip/media_tree.git | 534 | T: git git://linuxtv.org/anttip/media_tree.git |
| 535 | S: Maintained | 535 | S: Maintained |
| 536 | F: drivers/media/usb/airspy/ | 536 | F: drivers/media/usb/airspy/ |
| 537 | 537 | ||
| 538 | ALCATEL SPEEDTOUCH USB DRIVER | 538 | ALCATEL SPEEDTOUCH USB DRIVER |
| 539 | M: Duncan Sands <duncan.sands@free.fr> | 539 | M: Duncan Sands <duncan.sands@free.fr> |
| 540 | L: linux-usb@vger.kernel.org | 540 | L: linux-usb@vger.kernel.org |
| 541 | W: http://www.linux-usb.org/SpeedTouch/ | 541 | W: http://www.linux-usb.org/SpeedTouch/ |
| 542 | S: Maintained | 542 | S: Maintained |
| 543 | F: drivers/usb/atm/speedtch.c | 543 | F: drivers/usb/atm/speedtch.c |
| 544 | F: drivers/usb/atm/usbatm.c | 544 | F: drivers/usb/atm/usbatm.c |
| 545 | 545 | ||
| 546 | ALCHEMY AU1XX0 MMC DRIVER | 546 | ALCHEMY AU1XX0 MMC DRIVER |
| 547 | M: Manuel Lauss <manuel.lauss@gmail.com> | 547 | M: Manuel Lauss <manuel.lauss@gmail.com> |
| 548 | S: Maintained | 548 | S: Maintained |
| 549 | F: drivers/mmc/host/au1xmmc.c | 549 | F: drivers/mmc/host/au1xmmc.c |
| 550 | 550 | ||
| 551 | ALI1563 I2C DRIVER | 551 | ALI1563 I2C DRIVER |
| 552 | M: Rudolf Marek <r.marek@assembler.cz> | 552 | M: Rudolf Marek <r.marek@assembler.cz> |
| 553 | L: linux-i2c@vger.kernel.org | 553 | L: linux-i2c@vger.kernel.org |
| 554 | S: Maintained | 554 | S: Maintained |
| 555 | F: Documentation/i2c/busses/i2c-ali1563 | 555 | F: Documentation/i2c/busses/i2c-ali1563 |
| 556 | F: drivers/i2c/busses/i2c-ali1563.c | 556 | F: drivers/i2c/busses/i2c-ali1563.c |
| 557 | 557 | ||
| 558 | ALPHA PORT | 558 | ALPHA PORT |
| 559 | M: Richard Henderson <rth@twiddle.net> | 559 | M: Richard Henderson <rth@twiddle.net> |
| 560 | M: Ivan Kokshaysky <ink@jurassic.park.msu.ru> | 560 | M: Ivan Kokshaysky <ink@jurassic.park.msu.ru> |
| 561 | M: Matt Turner <mattst88@gmail.com> | 561 | M: Matt Turner <mattst88@gmail.com> |
| 562 | S: Odd Fixes | 562 | S: Odd Fixes |
| 563 | L: linux-alpha@vger.kernel.org | 563 | L: linux-alpha@vger.kernel.org |
| 564 | F: arch/alpha/ | 564 | F: arch/alpha/ |
| 565 | 565 | ||
| 566 | ALTERA TRIPLE SPEED ETHERNET DRIVER | 566 | ALTERA TRIPLE SPEED ETHERNET DRIVER |
| 567 | M: Vince Bridgers <vbridger@opensource.altera.com> | 567 | M: Vince Bridgers <vbridger@opensource.altera.com> |
| 568 | L: netdev@vger.kernel.org | 568 | L: netdev@vger.kernel.org |
| 569 | L: nios2-dev@lists.rocketboards.org (moderated for non-subscribers) | 569 | L: nios2-dev@lists.rocketboards.org (moderated for non-subscribers) |
| 570 | S: Maintained | 570 | S: Maintained |
| 571 | F: drivers/net/ethernet/altera/ | 571 | F: drivers/net/ethernet/altera/ |
| 572 | 572 | ||
| 573 | ALTERA UART/JTAG UART SERIAL DRIVERS | 573 | ALTERA UART/JTAG UART SERIAL DRIVERS |
| 574 | M: Tobias Klauser <tklauser@distanz.ch> | 574 | M: Tobias Klauser <tklauser@distanz.ch> |
| 575 | L: linux-serial@vger.kernel.org | 575 | L: linux-serial@vger.kernel.org |
| 576 | L: nios2-dev@lists.rocketboards.org (moderated for non-subscribers) | 576 | L: nios2-dev@lists.rocketboards.org (moderated for non-subscribers) |
| 577 | S: Maintained | 577 | S: Maintained |
| 578 | F: drivers/tty/serial/altera_uart.c | 578 | F: drivers/tty/serial/altera_uart.c |
| 579 | F: drivers/tty/serial/altera_jtaguart.c | 579 | F: drivers/tty/serial/altera_jtaguart.c |
| 580 | F: include/linux/altera_uart.h | 580 | F: include/linux/altera_uart.h |
| 581 | F: include/linux/altera_jtaguart.h | 581 | F: include/linux/altera_jtaguart.h |
| 582 | 582 | ||
| 583 | AMD CRYPTOGRAPHIC COPROCESSOR (CCP) DRIVER | 583 | AMD CRYPTOGRAPHIC COPROCESSOR (CCP) DRIVER |
| 584 | M: Tom Lendacky <thomas.lendacky@amd.com> | 584 | M: Tom Lendacky <thomas.lendacky@amd.com> |
| 585 | L: linux-crypto@vger.kernel.org | 585 | L: linux-crypto@vger.kernel.org |
| 586 | S: Supported | 586 | S: Supported |
| 587 | F: drivers/crypto/ccp/ | 587 | F: drivers/crypto/ccp/ |
| 588 | F: include/linux/ccp.h | 588 | F: include/linux/ccp.h |
| 589 | 589 | ||
| 590 | AMD FAM15H PROCESSOR POWER MONITORING DRIVER | 590 | AMD FAM15H PROCESSOR POWER MONITORING DRIVER |
| 591 | M: Andreas Herrmann <herrmann.der.user@googlemail.com> | 591 | M: Andreas Herrmann <herrmann.der.user@googlemail.com> |
| 592 | L: lm-sensors@lm-sensors.org | 592 | L: lm-sensors@lm-sensors.org |
| 593 | S: Maintained | 593 | S: Maintained |
| 594 | F: Documentation/hwmon/fam15h_power | 594 | F: Documentation/hwmon/fam15h_power |
| 595 | F: drivers/hwmon/fam15h_power.c | 595 | F: drivers/hwmon/fam15h_power.c |
| 596 | 596 | ||
| 597 | AMD GEODE CS5536 USB DEVICE CONTROLLER DRIVER | 597 | AMD GEODE CS5536 USB DEVICE CONTROLLER DRIVER |
| 598 | M: Thomas Dahlmann <dahlmann.thomas@arcor.de> | 598 | M: Thomas Dahlmann <dahlmann.thomas@arcor.de> |
| 599 | L: linux-geode@lists.infradead.org (moderated for non-subscribers) | 599 | L: linux-geode@lists.infradead.org (moderated for non-subscribers) |
| 600 | S: Supported | 600 | S: Supported |
| 601 | F: drivers/usb/gadget/udc/amd5536udc.* | 601 | F: drivers/usb/gadget/udc/amd5536udc.* |
| 602 | 602 | ||
| 603 | AMD GEODE PROCESSOR/CHIPSET SUPPORT | 603 | AMD GEODE PROCESSOR/CHIPSET SUPPORT |
| 604 | P: Andres Salomon <dilinger@queued.net> | 604 | P: Andres Salomon <dilinger@queued.net> |
| 605 | L: linux-geode@lists.infradead.org (moderated for non-subscribers) | 605 | L: linux-geode@lists.infradead.org (moderated for non-subscribers) |
| 606 | W: http://www.amd.com/us-en/ConnectivitySolutions/TechnicalResources/0,,50_2334_2452_11363,00.html | 606 | W: http://www.amd.com/us-en/ConnectivitySolutions/TechnicalResources/0,,50_2334_2452_11363,00.html |
| 607 | S: Supported | 607 | S: Supported |
| 608 | F: drivers/char/hw_random/geode-rng.c | 608 | F: drivers/char/hw_random/geode-rng.c |
| 609 | F: drivers/crypto/geode* | 609 | F: drivers/crypto/geode* |
| 610 | F: drivers/video/fbdev/geode/ | 610 | F: drivers/video/fbdev/geode/ |
| 611 | F: arch/x86/include/asm/geode.h | 611 | F: arch/x86/include/asm/geode.h |
| 612 | 612 | ||
| 613 | AMD IOMMU (AMD-VI) | 613 | AMD IOMMU (AMD-VI) |
| 614 | M: Joerg Roedel <joro@8bytes.org> | 614 | M: Joerg Roedel <joro@8bytes.org> |
| 615 | L: iommu@lists.linux-foundation.org | 615 | L: iommu@lists.linux-foundation.org |
| 616 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git | 616 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git |
| 617 | S: Maintained | 617 | S: Maintained |
| 618 | F: drivers/iommu/amd_iommu*.[ch] | 618 | F: drivers/iommu/amd_iommu*.[ch] |
| 619 | F: include/linux/amd-iommu.h | 619 | F: include/linux/amd-iommu.h |
| 620 | 620 | ||
| 621 | AMD KFD | 621 | AMD KFD |
| 622 | M: Oded Gabbay <oded.gabbay@amd.com> | 622 | M: Oded Gabbay <oded.gabbay@amd.com> |
| 623 | L: dri-devel@lists.freedesktop.org | 623 | L: dri-devel@lists.freedesktop.org |
| 624 | T: git git://people.freedesktop.org/~gabbayo/linux.git | 624 | T: git git://people.freedesktop.org/~gabbayo/linux.git |
| 625 | S: Supported | 625 | S: Supported |
| 626 | F: drivers/gpu/drm/amd/amdkfd/ | 626 | F: drivers/gpu/drm/amd/amdkfd/ |
| 627 | F: drivers/gpu/drm/radeon/radeon_kfd.c | 627 | F: drivers/gpu/drm/radeon/radeon_kfd.c |
| 628 | F: drivers/gpu/drm/radeon/radeon_kfd.h | 628 | F: drivers/gpu/drm/radeon/radeon_kfd.h |
| 629 | F: include/uapi/linux/kfd_ioctl.h | 629 | F: include/uapi/linux/kfd_ioctl.h |
| 630 | 630 | ||
| 631 | AMD MICROCODE UPDATE SUPPORT | 631 | AMD MICROCODE UPDATE SUPPORT |
| 632 | M: Andreas Herrmann <herrmann.der.user@googlemail.com> | 632 | M: Andreas Herrmann <herrmann.der.user@googlemail.com> |
| 633 | L: amd64-microcode@amd64.org | 633 | L: amd64-microcode@amd64.org |
| 634 | S: Maintained | 634 | S: Maintained |
| 635 | F: arch/x86/kernel/cpu/microcode/amd* | 635 | F: arch/x86/kernel/cpu/microcode/amd* |
| 636 | 636 | ||
| 637 | AMD XGBE DRIVER | 637 | AMD XGBE DRIVER |
| 638 | M: Tom Lendacky <thomas.lendacky@amd.com> | 638 | M: Tom Lendacky <thomas.lendacky@amd.com> |
| 639 | L: netdev@vger.kernel.org | 639 | L: netdev@vger.kernel.org |
| 640 | S: Supported | 640 | S: Supported |
| 641 | F: drivers/net/ethernet/amd/xgbe/ | 641 | F: drivers/net/ethernet/amd/xgbe/ |
| 642 | F: drivers/net/phy/amd-xgbe-phy.c | 642 | F: drivers/net/phy/amd-xgbe-phy.c |
| 643 | 643 | ||
| 644 | AMS (Apple Motion Sensor) DRIVER | 644 | AMS (Apple Motion Sensor) DRIVER |
| 645 | M: Michael Hanselmann <linux-kernel@hansmi.ch> | 645 | M: Michael Hanselmann <linux-kernel@hansmi.ch> |
| 646 | S: Supported | 646 | S: Supported |
| 647 | F: drivers/macintosh/ams/ | 647 | F: drivers/macintosh/ams/ |
| 648 | 648 | ||
| 649 | AMSO1100 RNIC DRIVER | 649 | AMSO1100 RNIC DRIVER |
| 650 | M: Tom Tucker <tom@opengridcomputing.com> | 650 | M: Tom Tucker <tom@opengridcomputing.com> |
| 651 | M: Steve Wise <swise@opengridcomputing.com> | 651 | M: Steve Wise <swise@opengridcomputing.com> |
| 652 | L: linux-rdma@vger.kernel.org | 652 | L: linux-rdma@vger.kernel.org |
| 653 | S: Maintained | 653 | S: Maintained |
| 654 | F: drivers/infiniband/hw/amso1100/ | 654 | F: drivers/infiniband/hw/amso1100/ |
| 655 | 655 | ||
| 656 | ANALOG DEVICES INC AD9389B DRIVER | 656 | ANALOG DEVICES INC AD9389B DRIVER |
| 657 | M: Hans Verkuil <hans.verkuil@cisco.com> | 657 | M: Hans Verkuil <hans.verkuil@cisco.com> |
| 658 | L: linux-media@vger.kernel.org | 658 | L: linux-media@vger.kernel.org |
| 659 | S: Maintained | 659 | S: Maintained |
| 660 | F: drivers/media/i2c/ad9389b* | 660 | F: drivers/media/i2c/ad9389b* |
| 661 | 661 | ||
| 662 | ANALOG DEVICES INC ADV7511 DRIVER | 662 | ANALOG DEVICES INC ADV7511 DRIVER |
| 663 | M: Hans Verkuil <hans.verkuil@cisco.com> | 663 | M: Hans Verkuil <hans.verkuil@cisco.com> |
| 664 | L: linux-media@vger.kernel.org | 664 | L: linux-media@vger.kernel.org |
| 665 | S: Maintained | 665 | S: Maintained |
| 666 | F: drivers/media/i2c/adv7511* | 666 | F: drivers/media/i2c/adv7511* |
| 667 | 667 | ||
| 668 | ANALOG DEVICES INC ADV7604 DRIVER | 668 | ANALOG DEVICES INC ADV7604 DRIVER |
| 669 | M: Hans Verkuil <hans.verkuil@cisco.com> | 669 | M: Hans Verkuil <hans.verkuil@cisco.com> |
| 670 | L: linux-media@vger.kernel.org | 670 | L: linux-media@vger.kernel.org |
| 671 | S: Maintained | 671 | S: Maintained |
| 672 | F: drivers/media/i2c/adv7604* | 672 | F: drivers/media/i2c/adv7604* |
| 673 | 673 | ||
| 674 | ANALOG DEVICES INC ADV7842 DRIVER | 674 | ANALOG DEVICES INC ADV7842 DRIVER |
| 675 | M: Hans Verkuil <hans.verkuil@cisco.com> | 675 | M: Hans Verkuil <hans.verkuil@cisco.com> |
| 676 | L: linux-media@vger.kernel.org | 676 | L: linux-media@vger.kernel.org |
| 677 | S: Maintained | 677 | S: Maintained |
| 678 | F: drivers/media/i2c/adv7842* | 678 | F: drivers/media/i2c/adv7842* |
| 679 | 679 | ||
| 680 | ANALOG DEVICES INC ASOC CODEC DRIVERS | 680 | ANALOG DEVICES INC ASOC CODEC DRIVERS |
| 681 | M: Lars-Peter Clausen <lars@metafoo.de> | 681 | M: Lars-Peter Clausen <lars@metafoo.de> |
| 682 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 682 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 683 | W: http://wiki.analog.com/ | 683 | W: http://wiki.analog.com/ |
| 684 | W: http://ez.analog.com/community/linux-device-drivers | 684 | W: http://ez.analog.com/community/linux-device-drivers |
| 685 | S: Supported | 685 | S: Supported |
| 686 | F: sound/soc/codecs/adau* | 686 | F: sound/soc/codecs/adau* |
| 687 | F: sound/soc/codecs/adav* | 687 | F: sound/soc/codecs/adav* |
| 688 | F: sound/soc/codecs/ad1* | 688 | F: sound/soc/codecs/ad1* |
| 689 | F: sound/soc/codecs/ad7* | 689 | F: sound/soc/codecs/ad7* |
| 690 | F: sound/soc/codecs/ssm* | 690 | F: sound/soc/codecs/ssm* |
| 691 | F: sound/soc/codecs/sigmadsp.* | 691 | F: sound/soc/codecs/sigmadsp.* |
| 692 | 692 | ||
| 693 | ANALOG DEVICES INC ASOC DRIVERS | 693 | ANALOG DEVICES INC ASOC DRIVERS |
| 694 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 694 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 695 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 695 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 696 | W: http://blackfin.uclinux.org/ | 696 | W: http://blackfin.uclinux.org/ |
| 697 | S: Supported | 697 | S: Supported |
| 698 | F: sound/soc/blackfin/* | 698 | F: sound/soc/blackfin/* |
| 699 | 699 | ||
| 700 | ANALOG DEVICES INC IIO DRIVERS | 700 | ANALOG DEVICES INC IIO DRIVERS |
| 701 | M: Lars-Peter Clausen <lars@metafoo.de> | 701 | M: Lars-Peter Clausen <lars@metafoo.de> |
| 702 | M: Michael Hennerich <Michael.Hennerich@analog.com> | 702 | M: Michael Hennerich <Michael.Hennerich@analog.com> |
| 703 | W: http://wiki.analog.com/ | 703 | W: http://wiki.analog.com/ |
| 704 | W: http://ez.analog.com/community/linux-device-drivers | 704 | W: http://ez.analog.com/community/linux-device-drivers |
| 705 | S: Supported | 705 | S: Supported |
| 706 | F: drivers/iio/*/ad* | 706 | F: drivers/iio/*/ad* |
| 707 | X: drivers/iio/*/adjd* | 707 | X: drivers/iio/*/adjd* |
| 708 | F: drivers/staging/iio/*/ad* | 708 | F: drivers/staging/iio/*/ad* |
| 709 | F: staging/iio/trigger/iio-trig-bfin-timer.c | 709 | F: staging/iio/trigger/iio-trig-bfin-timer.c |
| 710 | 710 | ||
| 711 | AOA (Apple Onboard Audio) ALSA DRIVER | 711 | AOA (Apple Onboard Audio) ALSA DRIVER |
| 712 | M: Johannes Berg <johannes@sipsolutions.net> | 712 | M: Johannes Berg <johannes@sipsolutions.net> |
| 713 | L: linuxppc-dev@lists.ozlabs.org | 713 | L: linuxppc-dev@lists.ozlabs.org |
| 714 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 714 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 715 | S: Maintained | 715 | S: Maintained |
| 716 | F: sound/aoa/ | 716 | F: sound/aoa/ |
| 717 | 717 | ||
| 718 | APM DRIVER | 718 | APM DRIVER |
| 719 | M: Jiri Kosina <jkosina@suse.cz> | 719 | M: Jiri Kosina <jkosina@suse.cz> |
| 720 | S: Odd fixes | 720 | S: Odd fixes |
| 721 | F: arch/x86/kernel/apm_32.c | 721 | F: arch/x86/kernel/apm_32.c |
| 722 | F: include/linux/apm_bios.h | 722 | F: include/linux/apm_bios.h |
| 723 | F: include/uapi/linux/apm_bios.h | 723 | F: include/uapi/linux/apm_bios.h |
| 724 | F: drivers/char/apm-emulation.c | 724 | F: drivers/char/apm-emulation.c |
| 725 | 725 | ||
| 726 | APPLE BCM5974 MULTITOUCH DRIVER | 726 | APPLE BCM5974 MULTITOUCH DRIVER |
| 727 | M: Henrik Rydberg <rydberg@bitmath.org> | 727 | M: Henrik Rydberg <rydberg@bitmath.org> |
| 728 | L: linux-input@vger.kernel.org | 728 | L: linux-input@vger.kernel.org |
| 729 | S: Odd fixes | 729 | S: Odd fixes |
| 730 | F: drivers/input/mouse/bcm5974.c | 730 | F: drivers/input/mouse/bcm5974.c |
| 731 | 731 | ||
| 732 | APPLE SMC DRIVER | 732 | APPLE SMC DRIVER |
| 733 | M: Henrik Rydberg <rydberg@bitmath.org> | 733 | M: Henrik Rydberg <rydberg@bitmath.org> |
| 734 | L: lm-sensors@lm-sensors.org | 734 | L: lm-sensors@lm-sensors.org |
| 735 | S: Odd fixes | 735 | S: Odd fixes |
| 736 | F: drivers/hwmon/applesmc.c | 736 | F: drivers/hwmon/applesmc.c |
| 737 | 737 | ||
| 738 | APPLETALK NETWORK LAYER | 738 | APPLETALK NETWORK LAYER |
| 739 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> | 739 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> |
| 740 | S: Maintained | 740 | S: Maintained |
| 741 | F: drivers/net/appletalk/ | 741 | F: drivers/net/appletalk/ |
| 742 | F: net/appletalk/ | 742 | F: net/appletalk/ |
| 743 | 743 | ||
| 744 | APPLIED MICRO (APM) X-GENE SOC ETHERNET DRIVER | 744 | APPLIED MICRO (APM) X-GENE SOC ETHERNET DRIVER |
| 745 | M: Iyappan Subramanian <isubramanian@apm.com> | 745 | M: Iyappan Subramanian <isubramanian@apm.com> |
| 746 | M: Keyur Chudgar <kchudgar@apm.com> | 746 | M: Keyur Chudgar <kchudgar@apm.com> |
| 747 | S: Supported | 747 | S: Supported |
| 748 | F: drivers/net/ethernet/apm/xgene/ | 748 | F: drivers/net/ethernet/apm/xgene/ |
| 749 | F: Documentation/devicetree/bindings/net/apm-xgene-enet.txt | 749 | F: Documentation/devicetree/bindings/net/apm-xgene-enet.txt |
| 750 | 750 | ||
| 751 | APTINA CAMERA SENSOR PLL | 751 | APTINA CAMERA SENSOR PLL |
| 752 | M: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> | 752 | M: Laurent Pinchart <Laurent.pinchart@ideasonboard.com> |
| 753 | L: linux-media@vger.kernel.org | 753 | L: linux-media@vger.kernel.org |
| 754 | S: Maintained | 754 | S: Maintained |
| 755 | F: drivers/media/i2c/aptina-pll.* | 755 | F: drivers/media/i2c/aptina-pll.* |
| 756 | 756 | ||
| 757 | ARASAN COMPACT FLASH PATA CONTROLLER | 757 | ARASAN COMPACT FLASH PATA CONTROLLER |
| 758 | M: Viresh Kumar <viresh.linux@gmail.com> | 758 | M: Viresh Kumar <viresh.linux@gmail.com> |
| 759 | L: linux-ide@vger.kernel.org | 759 | L: linux-ide@vger.kernel.org |
| 760 | S: Maintained | 760 | S: Maintained |
| 761 | F: include/linux/pata_arasan_cf_data.h | 761 | F: include/linux/pata_arasan_cf_data.h |
| 762 | F: drivers/ata/pata_arasan_cf.c | 762 | F: drivers/ata/pata_arasan_cf.c |
| 763 | 763 | ||
| 764 | ARC FRAMEBUFFER DRIVER | 764 | ARC FRAMEBUFFER DRIVER |
| 765 | M: Jaya Kumar <jayalk@intworks.biz> | 765 | M: Jaya Kumar <jayalk@intworks.biz> |
| 766 | S: Maintained | 766 | S: Maintained |
| 767 | F: drivers/video/fbdev/arcfb.c | 767 | F: drivers/video/fbdev/arcfb.c |
| 768 | F: drivers/video/fbdev/core/fb_defio.c | 768 | F: drivers/video/fbdev/core/fb_defio.c |
| 769 | 769 | ||
| 770 | ARM MFM AND FLOPPY DRIVERS | 770 | ARM MFM AND FLOPPY DRIVERS |
| 771 | M: Ian Molton <spyro@f2s.com> | 771 | M: Ian Molton <spyro@f2s.com> |
| 772 | S: Maintained | 772 | S: Maintained |
| 773 | F: arch/arm/lib/floppydma.S | 773 | F: arch/arm/lib/floppydma.S |
| 774 | F: arch/arm/include/asm/floppy.h | 774 | F: arch/arm/include/asm/floppy.h |
| 775 | 775 | ||
| 776 | ARM PMU PROFILING AND DEBUGGING | 776 | ARM PMU PROFILING AND DEBUGGING |
| 777 | M: Will Deacon <will.deacon@arm.com> | 777 | M: Will Deacon <will.deacon@arm.com> |
| 778 | S: Maintained | 778 | S: Maintained |
| 779 | F: arch/arm/kernel/perf_event* | 779 | F: arch/arm/kernel/perf_event* |
| 780 | F: arch/arm/oprofile/common.c | 780 | F: arch/arm/oprofile/common.c |
| 781 | F: arch/arm/include/asm/pmu.h | 781 | F: arch/arm/include/asm/pmu.h |
| 782 | F: arch/arm/kernel/hw_breakpoint.c | 782 | F: arch/arm/kernel/hw_breakpoint.c |
| 783 | F: arch/arm/include/asm/hw_breakpoint.h | 783 | F: arch/arm/include/asm/hw_breakpoint.h |
| 784 | 784 | ||
| 785 | ARM PORT | 785 | ARM PORT |
| 786 | M: Russell King <linux@arm.linux.org.uk> | 786 | M: Russell King <linux@arm.linux.org.uk> |
| 787 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 787 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 788 | W: http://www.arm.linux.org.uk/ | 788 | W: http://www.arm.linux.org.uk/ |
| 789 | S: Maintained | 789 | S: Maintained |
| 790 | F: arch/arm/ | 790 | F: arch/arm/ |
| 791 | 791 | ||
| 792 | ARM SUB-ARCHITECTURES | 792 | ARM SUB-ARCHITECTURES |
| 793 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 793 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 794 | S: Maintained | 794 | S: Maintained |
| 795 | F: arch/arm/mach-*/ | 795 | F: arch/arm/mach-*/ |
| 796 | F: arch/arm/plat-*/ | 796 | F: arch/arm/plat-*/ |
| 797 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc.git | 797 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc.git |
| 798 | 798 | ||
| 799 | ARM PRIMECELL AACI PL041 DRIVER | 799 | ARM PRIMECELL AACI PL041 DRIVER |
| 800 | M: Russell King <linux@arm.linux.org.uk> | 800 | M: Russell King <linux@arm.linux.org.uk> |
| 801 | S: Maintained | 801 | S: Maintained |
| 802 | F: sound/arm/aaci.* | 802 | F: sound/arm/aaci.* |
| 803 | 803 | ||
| 804 | ARM PRIMECELL CLCD PL110 DRIVER | 804 | ARM PRIMECELL CLCD PL110 DRIVER |
| 805 | M: Russell King <linux@arm.linux.org.uk> | 805 | M: Russell King <linux@arm.linux.org.uk> |
| 806 | S: Maintained | 806 | S: Maintained |
| 807 | F: drivers/video/fbdev/amba-clcd.* | 807 | F: drivers/video/fbdev/amba-clcd.* |
| 808 | 808 | ||
| 809 | ARM PRIMECELL KMI PL050 DRIVER | 809 | ARM PRIMECELL KMI PL050 DRIVER |
| 810 | M: Russell King <linux@arm.linux.org.uk> | 810 | M: Russell King <linux@arm.linux.org.uk> |
| 811 | S: Maintained | 811 | S: Maintained |
| 812 | F: drivers/input/serio/ambakmi.* | 812 | F: drivers/input/serio/ambakmi.* |
| 813 | F: include/linux/amba/kmi.h | 813 | F: include/linux/amba/kmi.h |
| 814 | 814 | ||
| 815 | ARM PRIMECELL MMCI PL180/1 DRIVER | 815 | ARM PRIMECELL MMCI PL180/1 DRIVER |
| 816 | M: Russell King <linux@arm.linux.org.uk> | 816 | M: Russell King <linux@arm.linux.org.uk> |
| 817 | S: Maintained | 817 | S: Maintained |
| 818 | F: drivers/mmc/host/mmci.* | 818 | F: drivers/mmc/host/mmci.* |
| 819 | F: include/linux/amba/mmci.h | 819 | F: include/linux/amba/mmci.h |
| 820 | 820 | ||
| 821 | ARM PRIMECELL UART PL010 AND PL011 DRIVERS | 821 | ARM PRIMECELL UART PL010 AND PL011 DRIVERS |
| 822 | M: Russell King <linux@arm.linux.org.uk> | 822 | M: Russell King <linux@arm.linux.org.uk> |
| 823 | S: Maintained | 823 | S: Maintained |
| 824 | F: drivers/tty/serial/amba-pl01*.c | 824 | F: drivers/tty/serial/amba-pl01*.c |
| 825 | F: include/linux/amba/serial.h | 825 | F: include/linux/amba/serial.h |
| 826 | 826 | ||
| 827 | ARM PRIMECELL BUS SUPPORT | 827 | ARM PRIMECELL BUS SUPPORT |
| 828 | M: Russell King <linux@arm.linux.org.uk> | 828 | M: Russell King <linux@arm.linux.org.uk> |
| 829 | S: Maintained | 829 | S: Maintained |
| 830 | F: drivers/amba/ | 830 | F: drivers/amba/ |
| 831 | F: include/linux/amba/bus.h | 831 | F: include/linux/amba/bus.h |
| 832 | 832 | ||
| 833 | ARM/ADS SPHERE MACHINE SUPPORT | 833 | ARM/ADS SPHERE MACHINE SUPPORT |
| 834 | M: Lennert Buytenhek <kernel@wantstofly.org> | 834 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 835 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 835 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 836 | S: Maintained | 836 | S: Maintained |
| 837 | 837 | ||
| 838 | ARM/AFEB9260 MACHINE SUPPORT | 838 | ARM/AFEB9260 MACHINE SUPPORT |
| 839 | M: Sergey Lapin <slapin@ossfans.org> | 839 | M: Sergey Lapin <slapin@ossfans.org> |
| 840 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 840 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 841 | S: Maintained | 841 | S: Maintained |
| 842 | 842 | ||
| 843 | ARM/AJECO 1ARM MACHINE SUPPORT | 843 | ARM/AJECO 1ARM MACHINE SUPPORT |
| 844 | M: Lennert Buytenhek <kernel@wantstofly.org> | 844 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 845 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 845 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 846 | S: Maintained | 846 | S: Maintained |
| 847 | 847 | ||
| 848 | ARM/Allwinner A1X SoC support | 848 | ARM/Allwinner A1X SoC support |
| 849 | M: Maxime Ripard <maxime.ripard@free-electrons.com> | 849 | M: Maxime Ripard <maxime.ripard@free-electrons.com> |
| 850 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 850 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 851 | S: Maintained | 851 | S: Maintained |
| 852 | N: sun[x4567]i | 852 | N: sun[x4567]i |
| 853 | 853 | ||
| 854 | ARM/Allwinner SoC Clock Support | 854 | ARM/Allwinner SoC Clock Support |
| 855 | M: Emilio Lรณpez <emilio@elopez.com.ar> | 855 | M: Emilio Lรณpez <emilio@elopez.com.ar> |
| 856 | S: Maintained | 856 | S: Maintained |
| 857 | F: drivers/clk/sunxi/ | 857 | F: drivers/clk/sunxi/ |
| 858 | 858 | ||
| 859 | ARM/Amlogic MesonX SoC support | 859 | ARM/Amlogic MesonX SoC support |
| 860 | M: Carlo Caione <carlo@caione.org> | 860 | M: Carlo Caione <carlo@caione.org> |
| 861 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 861 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 862 | S: Maintained | 862 | S: Maintained |
| 863 | F: drivers/media/rc/meson-ir.c | 863 | F: drivers/media/rc/meson-ir.c |
| 864 | N: meson[x68] | 864 | N: meson[x68] |
| 865 | 865 | ||
| 866 | ARM/ATMEL AT91RM9200 AND AT91SAM ARM ARCHITECTURES | 866 | ARM/ATMEL AT91RM9200 AND AT91SAM ARM ARCHITECTURES |
| 867 | M: Andrew Victor <linux@maxim.org.za> | 867 | M: Andrew Victor <linux@maxim.org.za> |
| 868 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 868 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 869 | M: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com> | 869 | M: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com> |
| 870 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 870 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 871 | W: http://maxim.org.za/at91_26.html | 871 | W: http://maxim.org.za/at91_26.html |
| 872 | W: http://www.linux4sam.org | 872 | W: http://www.linux4sam.org |
| 873 | S: Supported | 873 | S: Supported |
| 874 | F: arch/arm/mach-at91/ | 874 | F: arch/arm/mach-at91/ |
| 875 | F: include/soc/at91/ | 875 | F: include/soc/at91/ |
| 876 | F: arch/arm/boot/dts/at91*.dts | 876 | F: arch/arm/boot/dts/at91*.dts |
| 877 | F: arch/arm/boot/dts/at91*.dtsi | 877 | F: arch/arm/boot/dts/at91*.dtsi |
| 878 | F: arch/arm/boot/dts/sama*.dts | 878 | F: arch/arm/boot/dts/sama*.dts |
| 879 | F: arch/arm/boot/dts/sama*.dtsi | 879 | F: arch/arm/boot/dts/sama*.dtsi |
| 880 | 880 | ||
| 881 | ARM/ATMEL AT91 Clock Support | 881 | ARM/ATMEL AT91 Clock Support |
| 882 | M: Boris Brezillon <boris.brezillon@free-electrons.com> | 882 | M: Boris Brezillon <boris.brezillon@free-electrons.com> |
| 883 | S: Maintained | 883 | S: Maintained |
| 884 | F: drivers/clk/at91 | 884 | F: drivers/clk/at91 |
| 885 | 885 | ||
| 886 | ARM/CALXEDA HIGHBANK ARCHITECTURE | 886 | ARM/CALXEDA HIGHBANK ARCHITECTURE |
| 887 | M: Rob Herring <robh@kernel.org> | 887 | M: Rob Herring <robh@kernel.org> |
| 888 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 888 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 889 | S: Maintained | 889 | S: Maintained |
| 890 | F: arch/arm/mach-highbank/ | 890 | F: arch/arm/mach-highbank/ |
| 891 | 891 | ||
| 892 | ARM/CAVIUM NETWORKS CNS3XXX MACHINE SUPPORT | 892 | ARM/CAVIUM NETWORKS CNS3XXX MACHINE SUPPORT |
| 893 | M: Krzysztof Halasa <khalasa@piap.pl> | 893 | M: Krzysztof Halasa <khalasa@piap.pl> |
| 894 | S: Maintained | 894 | S: Maintained |
| 895 | F: arch/arm/mach-cns3xxx/ | 895 | F: arch/arm/mach-cns3xxx/ |
| 896 | 896 | ||
| 897 | ARM/CIRRUS LOGIC CLPS711X ARM ARCHITECTURE | 897 | ARM/CIRRUS LOGIC CLPS711X ARM ARCHITECTURE |
| 898 | M: Alexander Shiyan <shc_work@mail.ru> | 898 | M: Alexander Shiyan <shc_work@mail.ru> |
| 899 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 899 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 900 | S: Odd Fixes | 900 | S: Odd Fixes |
| 901 | N: clps711x | 901 | N: clps711x |
| 902 | 902 | ||
| 903 | ARM/CIRRUS LOGIC EP93XX ARM ARCHITECTURE | 903 | ARM/CIRRUS LOGIC EP93XX ARM ARCHITECTURE |
| 904 | M: Hartley Sweeten <hsweeten@visionengravers.com> | 904 | M: Hartley Sweeten <hsweeten@visionengravers.com> |
| 905 | M: Ryan Mallon <rmallon@gmail.com> | 905 | M: Ryan Mallon <rmallon@gmail.com> |
| 906 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 906 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 907 | S: Maintained | 907 | S: Maintained |
| 908 | F: arch/arm/mach-ep93xx/ | 908 | F: arch/arm/mach-ep93xx/ |
| 909 | F: arch/arm/mach-ep93xx/include/mach/ | 909 | F: arch/arm/mach-ep93xx/include/mach/ |
| 910 | 910 | ||
| 911 | ARM/CIRRUS LOGIC EDB9315A MACHINE SUPPORT | 911 | ARM/CIRRUS LOGIC EDB9315A MACHINE SUPPORT |
| 912 | M: Lennert Buytenhek <kernel@wantstofly.org> | 912 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 913 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 913 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 914 | S: Maintained | 914 | S: Maintained |
| 915 | 915 | ||
| 916 | ARM/CLKDEV SUPPORT | 916 | ARM/CLKDEV SUPPORT |
| 917 | M: Russell King <linux@arm.linux.org.uk> | 917 | M: Russell King <linux@arm.linux.org.uk> |
| 918 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 918 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 919 | S: Maintained | 919 | S: Maintained |
| 920 | F: arch/arm/include/asm/clkdev.h | 920 | F: arch/arm/include/asm/clkdev.h |
| 921 | F: drivers/clk/clkdev.c | 921 | F: drivers/clk/clkdev.c |
| 922 | 922 | ||
| 923 | ARM/COMPULAB CM-X270/EM-X270 and CM-X300 MACHINE SUPPORT | 923 | ARM/COMPULAB CM-X270/EM-X270 and CM-X300 MACHINE SUPPORT |
| 924 | M: Mike Rapoport <mike@compulab.co.il> | 924 | M: Mike Rapoport <mike@compulab.co.il> |
| 925 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 925 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 926 | S: Maintained | 926 | S: Maintained |
| 927 | 927 | ||
| 928 | ARM/CONTEC MICRO9 MACHINE SUPPORT | 928 | ARM/CONTEC MICRO9 MACHINE SUPPORT |
| 929 | M: Hubert Feurstein <hubert.feurstein@contec.at> | 929 | M: Hubert Feurstein <hubert.feurstein@contec.at> |
| 930 | S: Maintained | 930 | S: Maintained |
| 931 | F: arch/arm/mach-ep93xx/micro9.c | 931 | F: arch/arm/mach-ep93xx/micro9.c |
| 932 | 932 | ||
| 933 | ARM/CORESIGHT FRAMEWORK AND DRIVERS | 933 | ARM/CORESIGHT FRAMEWORK AND DRIVERS |
| 934 | M: Mathieu Poirier <mathieu.poirier@linaro.org> | 934 | M: Mathieu Poirier <mathieu.poirier@linaro.org> |
| 935 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 935 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 936 | S: Maintained | 936 | S: Maintained |
| 937 | F: drivers/coresight/* | 937 | F: drivers/coresight/* |
| 938 | F: Documentation/trace/coresight.txt | 938 | F: Documentation/trace/coresight.txt |
| 939 | F: Documentation/devicetree/bindings/arm/coresight.txt | 939 | F: Documentation/devicetree/bindings/arm/coresight.txt |
| 940 | F: Documentation/ABI/testing/sysfs-bus-coresight-devices-* | 940 | F: Documentation/ABI/testing/sysfs-bus-coresight-devices-* |
| 941 | 941 | ||
| 942 | ARM/CORGI MACHINE SUPPORT | 942 | ARM/CORGI MACHINE SUPPORT |
| 943 | M: Richard Purdie <rpurdie@rpsys.net> | 943 | M: Richard Purdie <rpurdie@rpsys.net> |
| 944 | S: Maintained | 944 | S: Maintained |
| 945 | 945 | ||
| 946 | ARM/CORTINA SYSTEMS GEMINI ARM ARCHITECTURE | 946 | ARM/CORTINA SYSTEMS GEMINI ARM ARCHITECTURE |
| 947 | M: Hans Ulli Kroll <ulli.kroll@googlemail.com> | 947 | M: Hans Ulli Kroll <ulli.kroll@googlemail.com> |
| 948 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 948 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 949 | T: git git://git.berlios.de/gemini-board | 949 | T: git git://git.berlios.de/gemini-board |
| 950 | S: Maintained | 950 | S: Maintained |
| 951 | F: arch/arm/mach-gemini/ | 951 | F: arch/arm/mach-gemini/ |
| 952 | 952 | ||
| 953 | ARM/CSR SIRFPRIMA2 MACHINE SUPPORT | 953 | ARM/CSR SIRFPRIMA2 MACHINE SUPPORT |
| 954 | M: Barry Song <baohua@kernel.org> | 954 | M: Barry Song <baohua@kernel.org> |
| 955 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 955 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 956 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/baohua/linux.git | 956 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/baohua/linux.git |
| 957 | S: Maintained | 957 | S: Maintained |
| 958 | F: arch/arm/mach-prima2/ | 958 | F: arch/arm/mach-prima2/ |
| 959 | F: drivers/clk/sirf/ | 959 | F: drivers/clk/sirf/ |
| 960 | F: drivers/clocksource/timer-prima2.c | 960 | F: drivers/clocksource/timer-prima2.c |
| 961 | F: drivers/clocksource/timer-marco.c | 961 | F: drivers/clocksource/timer-marco.c |
| 962 | N: [^a-z]sirf | 962 | N: [^a-z]sirf |
| 963 | 963 | ||
| 964 | ARM/EBSA110 MACHINE SUPPORT | 964 | ARM/EBSA110 MACHINE SUPPORT |
| 965 | M: Russell King <linux@arm.linux.org.uk> | 965 | M: Russell King <linux@arm.linux.org.uk> |
| 966 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 966 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 967 | W: http://www.arm.linux.org.uk/ | 967 | W: http://www.arm.linux.org.uk/ |
| 968 | S: Maintained | 968 | S: Maintained |
| 969 | F: arch/arm/mach-ebsa110/ | 969 | F: arch/arm/mach-ebsa110/ |
| 970 | F: drivers/net/ethernet/amd/am79c961a.* | 970 | F: drivers/net/ethernet/amd/am79c961a.* |
| 971 | 971 | ||
| 972 | ARM/ENERGY MICRO (SILICON LABS) EFM32 SUPPORT | 972 | ARM/ENERGY MICRO (SILICON LABS) EFM32 SUPPORT |
| 973 | M: Uwe Kleine-Kรถnig <kernel@pengutronix.de> | 973 | M: Uwe Kleine-Kรถnig <kernel@pengutronix.de> |
| 974 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 974 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 975 | S: Maintained | 975 | S: Maintained |
| 976 | N: efm32 | 976 | N: efm32 |
| 977 | 977 | ||
| 978 | ARM/EZX SMARTPHONES (A780, A910, A1200, E680, ROKR E2 and ROKR E6) | 978 | ARM/EZX SMARTPHONES (A780, A910, A1200, E680, ROKR E2 and ROKR E6) |
| 979 | M: Daniel Ribeiro <drwyrm@gmail.com> | 979 | M: Daniel Ribeiro <drwyrm@gmail.com> |
| 980 | M: Stefan Schmidt <stefan@openezx.org> | 980 | M: Stefan Schmidt <stefan@openezx.org> |
| 981 | M: Harald Welte <laforge@openezx.org> | 981 | M: Harald Welte <laforge@openezx.org> |
| 982 | L: openezx-devel@lists.openezx.org (moderated for non-subscribers) | 982 | L: openezx-devel@lists.openezx.org (moderated for non-subscribers) |
| 983 | W: http://www.openezx.org/ | 983 | W: http://www.openezx.org/ |
| 984 | S: Maintained | 984 | S: Maintained |
| 985 | T: topgit git://git.openezx.org/openezx.git | 985 | T: topgit git://git.openezx.org/openezx.git |
| 986 | F: arch/arm/mach-pxa/ezx.c | 986 | F: arch/arm/mach-pxa/ezx.c |
| 987 | 987 | ||
| 988 | ARM/FARADAY FA526 PORT | 988 | ARM/FARADAY FA526 PORT |
| 989 | M: Hans Ulli Kroll <ulli.kroll@googlemail.com> | 989 | M: Hans Ulli Kroll <ulli.kroll@googlemail.com> |
| 990 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 990 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 991 | S: Maintained | 991 | S: Maintained |
| 992 | T: git git://git.berlios.de/gemini-board | 992 | T: git git://git.berlios.de/gemini-board |
| 993 | F: arch/arm/mm/*-fa* | 993 | F: arch/arm/mm/*-fa* |
| 994 | 994 | ||
| 995 | ARM/FOOTBRIDGE ARCHITECTURE | 995 | ARM/FOOTBRIDGE ARCHITECTURE |
| 996 | M: Russell King <linux@arm.linux.org.uk> | 996 | M: Russell King <linux@arm.linux.org.uk> |
| 997 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 997 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 998 | W: http://www.arm.linux.org.uk/ | 998 | W: http://www.arm.linux.org.uk/ |
| 999 | S: Maintained | 999 | S: Maintained |
| 1000 | F: arch/arm/include/asm/hardware/dec21285.h | 1000 | F: arch/arm/include/asm/hardware/dec21285.h |
| 1001 | F: arch/arm/mach-footbridge/ | 1001 | F: arch/arm/mach-footbridge/ |
| 1002 | 1002 | ||
| 1003 | ARM/FREESCALE IMX / MXC ARM ARCHITECTURE | 1003 | ARM/FREESCALE IMX / MXC ARM ARCHITECTURE |
| 1004 | M: Shawn Guo <shawn.guo@linaro.org> | 1004 | M: Shawn Guo <shawn.guo@linaro.org> |
| 1005 | M: Sascha Hauer <kernel@pengutronix.de> | 1005 | M: Sascha Hauer <kernel@pengutronix.de> |
| 1006 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1006 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1007 | S: Maintained | 1007 | S: Maintained |
| 1008 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux.git | 1008 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux.git |
| 1009 | F: arch/arm/mach-imx/ | 1009 | F: arch/arm/mach-imx/ |
| 1010 | F: arch/arm/mach-mxs/ | 1010 | F: arch/arm/mach-mxs/ |
| 1011 | F: arch/arm/boot/dts/imx* | 1011 | F: arch/arm/boot/dts/imx* |
| 1012 | F: arch/arm/configs/imx*_defconfig | 1012 | F: arch/arm/configs/imx*_defconfig |
| 1013 | 1013 | ||
| 1014 | ARM/GLOMATION GESBC9312SX MACHINE SUPPORT | 1014 | ARM/GLOMATION GESBC9312SX MACHINE SUPPORT |
| 1015 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1015 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1016 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1016 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1017 | S: Maintained | 1017 | S: Maintained |
| 1018 | 1018 | ||
| 1019 | ARM/GUMSTIX MACHINE SUPPORT | 1019 | ARM/GUMSTIX MACHINE SUPPORT |
| 1020 | M: Steve Sakoman <sakoman@gmail.com> | 1020 | M: Steve Sakoman <sakoman@gmail.com> |
| 1021 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1021 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1022 | S: Maintained | 1022 | S: Maintained |
| 1023 | 1023 | ||
| 1024 | ARM/H4700 (HP IPAQ HX4700) MACHINE SUPPORT | 1024 | ARM/H4700 (HP IPAQ HX4700) MACHINE SUPPORT |
| 1025 | M: Philipp Zabel <philipp.zabel@gmail.com> | 1025 | M: Philipp Zabel <philipp.zabel@gmail.com> |
| 1026 | M: Paul Parsons <lost.distance@yahoo.com> | 1026 | M: Paul Parsons <lost.distance@yahoo.com> |
| 1027 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1027 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1028 | S: Maintained | 1028 | S: Maintained |
| 1029 | F: arch/arm/mach-pxa/hx4700.c | 1029 | F: arch/arm/mach-pxa/hx4700.c |
| 1030 | F: arch/arm/mach-pxa/include/mach/hx4700.h | 1030 | F: arch/arm/mach-pxa/include/mach/hx4700.h |
| 1031 | F: sound/soc/pxa/hx4700.c | 1031 | F: sound/soc/pxa/hx4700.c |
| 1032 | 1032 | ||
| 1033 | ARM/HISILICON SOC SUPPORT | 1033 | ARM/HISILICON SOC SUPPORT |
| 1034 | M: Wei Xu <xuwei5@hisilicon.com> | 1034 | M: Wei Xu <xuwei5@hisilicon.com> |
| 1035 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1035 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1036 | W: http://www.hisilicon.com | 1036 | W: http://www.hisilicon.com |
| 1037 | S: Supported | 1037 | S: Supported |
| 1038 | T: git git://github.com/hisilicon/linux-hisi.git | 1038 | T: git git://github.com/hisilicon/linux-hisi.git |
| 1039 | F: arch/arm/mach-hisi/ | 1039 | F: arch/arm/mach-hisi/ |
| 1040 | 1040 | ||
| 1041 | ARM/HP JORNADA 7XX MACHINE SUPPORT | 1041 | ARM/HP JORNADA 7XX MACHINE SUPPORT |
| 1042 | M: Kristoffer Ericson <kristoffer.ericson@gmail.com> | 1042 | M: Kristoffer Ericson <kristoffer.ericson@gmail.com> |
| 1043 | W: www.jlime.com | 1043 | W: www.jlime.com |
| 1044 | S: Maintained | 1044 | S: Maintained |
| 1045 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kristoffer/linux-hpc.git | 1045 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kristoffer/linux-hpc.git |
| 1046 | F: arch/arm/mach-sa1100/jornada720.c | 1046 | F: arch/arm/mach-sa1100/jornada720.c |
| 1047 | F: arch/arm/mach-sa1100/include/mach/jornada720.h | 1047 | F: arch/arm/mach-sa1100/include/mach/jornada720.h |
| 1048 | 1048 | ||
| 1049 | ARM/IGEP MACHINE SUPPORT | 1049 | ARM/IGEP MACHINE SUPPORT |
| 1050 | M: Enric Balletbo i Serra <eballetbo@gmail.com> | 1050 | M: Enric Balletbo i Serra <eballetbo@gmail.com> |
| 1051 | M: Javier Martinez Canillas <javier@dowhile0.org> | 1051 | M: Javier Martinez Canillas <javier@dowhile0.org> |
| 1052 | L: linux-omap@vger.kernel.org | 1052 | L: linux-omap@vger.kernel.org |
| 1053 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1053 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1054 | S: Maintained | 1054 | S: Maintained |
| 1055 | F: arch/arm/boot/dts/omap3-igep* | 1055 | F: arch/arm/boot/dts/omap3-igep* |
| 1056 | 1056 | ||
| 1057 | ARM/INCOME PXA270 SUPPORT | 1057 | ARM/INCOME PXA270 SUPPORT |
| 1058 | M: Marek Vasut <marek.vasut@gmail.com> | 1058 | M: Marek Vasut <marek.vasut@gmail.com> |
| 1059 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1059 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1060 | S: Maintained | 1060 | S: Maintained |
| 1061 | F: arch/arm/mach-pxa/colibri-pxa270-income.c | 1061 | F: arch/arm/mach-pxa/colibri-pxa270-income.c |
| 1062 | 1062 | ||
| 1063 | ARM/INTEL IOP32X ARM ARCHITECTURE | 1063 | ARM/INTEL IOP32X ARM ARCHITECTURE |
| 1064 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1064 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1065 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1065 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1066 | S: Maintained | 1066 | S: Maintained |
| 1067 | 1067 | ||
| 1068 | ARM/INTEL IOP33X ARM ARCHITECTURE | 1068 | ARM/INTEL IOP33X ARM ARCHITECTURE |
| 1069 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1069 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1070 | S: Orphan | 1070 | S: Orphan |
| 1071 | 1071 | ||
| 1072 | ARM/INTEL IOP13XX ARM ARCHITECTURE | 1072 | ARM/INTEL IOP13XX ARM ARCHITECTURE |
| 1073 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1073 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1074 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1074 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1075 | S: Maintained | 1075 | S: Maintained |
| 1076 | 1076 | ||
| 1077 | ARM/INTEL IQ81342EX MACHINE SUPPORT | 1077 | ARM/INTEL IQ81342EX MACHINE SUPPORT |
| 1078 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1078 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1079 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1079 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1080 | S: Maintained | 1080 | S: Maintained |
| 1081 | 1081 | ||
| 1082 | ARM/INTEL IXDP2850 MACHINE SUPPORT | 1082 | ARM/INTEL IXDP2850 MACHINE SUPPORT |
| 1083 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1083 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1084 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1084 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1085 | S: Maintained | 1085 | S: Maintained |
| 1086 | 1086 | ||
| 1087 | ARM/INTEL IXP4XX ARM ARCHITECTURE | 1087 | ARM/INTEL IXP4XX ARM ARCHITECTURE |
| 1088 | M: Imre Kaloz <kaloz@openwrt.org> | 1088 | M: Imre Kaloz <kaloz@openwrt.org> |
| 1089 | M: Krzysztof Halasa <khalasa@piap.pl> | 1089 | M: Krzysztof Halasa <khalasa@piap.pl> |
| 1090 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1090 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1091 | S: Maintained | 1091 | S: Maintained |
| 1092 | F: arch/arm/mach-ixp4xx/ | 1092 | F: arch/arm/mach-ixp4xx/ |
| 1093 | 1093 | ||
| 1094 | ARM/INTEL RESEARCH IMOTE/STARGATE 2 MACHINE SUPPORT | 1094 | ARM/INTEL RESEARCH IMOTE/STARGATE 2 MACHINE SUPPORT |
| 1095 | M: Jonathan Cameron <jic23@cam.ac.uk> | 1095 | M: Jonathan Cameron <jic23@cam.ac.uk> |
| 1096 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1096 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1097 | S: Maintained | 1097 | S: Maintained |
| 1098 | F: arch/arm/mach-pxa/stargate2.c | 1098 | F: arch/arm/mach-pxa/stargate2.c |
| 1099 | F: drivers/pcmcia/pxa2xx_stargate2.c | 1099 | F: drivers/pcmcia/pxa2xx_stargate2.c |
| 1100 | 1100 | ||
| 1101 | ARM/INTEL XSC3 (MANZANO) ARM CORE | 1101 | ARM/INTEL XSC3 (MANZANO) ARM CORE |
| 1102 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1102 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1103 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1103 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1104 | S: Maintained | 1104 | S: Maintained |
| 1105 | 1105 | ||
| 1106 | ARM/IP FABRICS DOUBLE ESPRESSO MACHINE SUPPORT | 1106 | ARM/IP FABRICS DOUBLE ESPRESSO MACHINE SUPPORT |
| 1107 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1107 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1108 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1108 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1109 | S: Maintained | 1109 | S: Maintained |
| 1110 | 1110 | ||
| 1111 | ARM/TEXAS INSTRUMENT KEYSTONE ARCHITECTURE | 1111 | ARM/TEXAS INSTRUMENT KEYSTONE ARCHITECTURE |
| 1112 | M: Santosh Shilimkar <ssantosh@kernel.org> | 1112 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 1113 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1113 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1114 | S: Maintained | 1114 | S: Maintained |
| 1115 | F: arch/arm/mach-keystone/ | 1115 | F: arch/arm/mach-keystone/ |
| 1116 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git | 1116 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git |
| 1117 | 1117 | ||
| 1118 | ARM/TEXAS INSTRUMENT KEYSTONE CLOCK FRAMEWORK | 1118 | ARM/TEXAS INSTRUMENT KEYSTONE CLOCK FRAMEWORK |
| 1119 | M: Santosh Shilimkar <ssantosh@kernel.org> | 1119 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 1120 | L: linux-kernel@vger.kernel.org | 1120 | L: linux-kernel@vger.kernel.org |
| 1121 | S: Maintained | 1121 | S: Maintained |
| 1122 | F: drivers/clk/keystone/ | 1122 | F: drivers/clk/keystone/ |
| 1123 | 1123 | ||
| 1124 | ARM/TEXAS INSTRUMENT KEYSTONE ClOCKSOURCE | 1124 | ARM/TEXAS INSTRUMENT KEYSTONE ClOCKSOURCE |
| 1125 | M: Santosh Shilimkar <ssantosh@kernel.org> | 1125 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 1126 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1126 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1127 | L: linux-kernel@vger.kernel.org | 1127 | L: linux-kernel@vger.kernel.org |
| 1128 | S: Maintained | 1128 | S: Maintained |
| 1129 | F: drivers/clocksource/timer-keystone.c | 1129 | F: drivers/clocksource/timer-keystone.c |
| 1130 | 1130 | ||
| 1131 | ARM/TEXAS INSTRUMENT KEYSTONE RESET DRIVER | 1131 | ARM/TEXAS INSTRUMENT KEYSTONE RESET DRIVER |
| 1132 | M: Santosh Shilimkar <ssantosh@kernel.org> | 1132 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 1133 | L: linux-kernel@vger.kernel.org | 1133 | L: linux-kernel@vger.kernel.org |
| 1134 | S: Maintained | 1134 | S: Maintained |
| 1135 | F: drivers/power/reset/keystone-reset.c | 1135 | F: drivers/power/reset/keystone-reset.c |
| 1136 | 1136 | ||
| 1137 | ARM/TEXAS INSTRUMENT AEMIF/EMIF DRIVERS | 1137 | ARM/TEXAS INSTRUMENT AEMIF/EMIF DRIVERS |
| 1138 | M: Santosh Shilimkar <ssantosh@kernel.org> | 1138 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 1139 | L: linux-kernel@vger.kernel.org | 1139 | L: linux-kernel@vger.kernel.org |
| 1140 | S: Maintained | 1140 | S: Maintained |
| 1141 | F: drivers/memory/*emif* | 1141 | F: drivers/memory/*emif* |
| 1142 | 1142 | ||
| 1143 | ARM/LOGICPD PXA270 MACHINE SUPPORT | 1143 | ARM/LOGICPD PXA270 MACHINE SUPPORT |
| 1144 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1144 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1145 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1145 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1146 | S: Maintained | 1146 | S: Maintained |
| 1147 | 1147 | ||
| 1148 | ARM/MAGICIAN MACHINE SUPPORT | 1148 | ARM/MAGICIAN MACHINE SUPPORT |
| 1149 | M: Philipp Zabel <philipp.zabel@gmail.com> | 1149 | M: Philipp Zabel <philipp.zabel@gmail.com> |
| 1150 | S: Maintained | 1150 | S: Maintained |
| 1151 | 1151 | ||
| 1152 | ARM/Marvell Armada 370 and Armada XP SOC support | 1152 | ARM/Marvell Armada 370 and Armada XP SOC support |
| 1153 | M: Jason Cooper <jason@lakedaemon.net> | 1153 | M: Jason Cooper <jason@lakedaemon.net> |
| 1154 | M: Andrew Lunn <andrew@lunn.ch> | 1154 | M: Andrew Lunn <andrew@lunn.ch> |
| 1155 | M: Gregory Clement <gregory.clement@free-electrons.com> | 1155 | M: Gregory Clement <gregory.clement@free-electrons.com> |
| 1156 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> | 1156 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> |
| 1157 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1157 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1158 | S: Maintained | 1158 | S: Maintained |
| 1159 | F: arch/arm/mach-mvebu/ | 1159 | F: arch/arm/mach-mvebu/ |
| 1160 | 1160 | ||
| 1161 | ARM/Marvell Berlin SoC support | 1161 | ARM/Marvell Berlin SoC support |
| 1162 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> | 1162 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> |
| 1163 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1163 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1164 | S: Maintained | 1164 | S: Maintained |
| 1165 | F: arch/arm/mach-berlin/ | 1165 | F: arch/arm/mach-berlin/ |
| 1166 | 1166 | ||
| 1167 | ARM/Marvell Dove/MV78xx0/Orion SOC support | 1167 | ARM/Marvell Dove/MV78xx0/Orion SOC support |
| 1168 | M: Jason Cooper <jason@lakedaemon.net> | 1168 | M: Jason Cooper <jason@lakedaemon.net> |
| 1169 | M: Andrew Lunn <andrew@lunn.ch> | 1169 | M: Andrew Lunn <andrew@lunn.ch> |
| 1170 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> | 1170 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> |
| 1171 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1171 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1172 | S: Maintained | 1172 | S: Maintained |
| 1173 | F: arch/arm/mach-dove/ | 1173 | F: arch/arm/mach-dove/ |
| 1174 | F: arch/arm/mach-mv78xx0/ | 1174 | F: arch/arm/mach-mv78xx0/ |
| 1175 | F: arch/arm/mach-orion5x/ | 1175 | F: arch/arm/mach-orion5x/ |
| 1176 | F: arch/arm/plat-orion/ | 1176 | F: arch/arm/plat-orion/ |
| 1177 | 1177 | ||
| 1178 | ARM/Orion SoC/Technologic Systems TS-78xx platform support | 1178 | ARM/Orion SoC/Technologic Systems TS-78xx platform support |
| 1179 | M: Alexander Clouter <alex@digriz.org.uk> | 1179 | M: Alexander Clouter <alex@digriz.org.uk> |
| 1180 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1180 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1181 | W: http://www.digriz.org.uk/ts78xx/kernel | 1181 | W: http://www.digriz.org.uk/ts78xx/kernel |
| 1182 | S: Maintained | 1182 | S: Maintained |
| 1183 | F: arch/arm/mach-orion5x/ts78xx-* | 1183 | F: arch/arm/mach-orion5x/ts78xx-* |
| 1184 | 1184 | ||
| 1185 | ARM/Mediatek SoC support | 1185 | ARM/Mediatek SoC support |
| 1186 | M: Matthias Brugger <matthias.bgg@gmail.com> | 1186 | M: Matthias Brugger <matthias.bgg@gmail.com> |
| 1187 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1187 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1188 | S: Maintained | 1188 | S: Maintained |
| 1189 | F: arch/arm/boot/dts/mt6* | 1189 | F: arch/arm/boot/dts/mt6* |
| 1190 | F: arch/arm/boot/dts/mt8* | 1190 | F: arch/arm/boot/dts/mt8* |
| 1191 | F: arch/arm/mach-mediatek/ | 1191 | F: arch/arm/mach-mediatek/ |
| 1192 | N: mtk | 1192 | N: mtk |
| 1193 | K: mediatek | 1193 | K: mediatek |
| 1194 | 1194 | ||
| 1195 | ARM/MICREL KS8695 ARCHITECTURE | 1195 | ARM/MICREL KS8695 ARCHITECTURE |
| 1196 | M: Greg Ungerer <gerg@uclinux.org> | 1196 | M: Greg Ungerer <gerg@uclinux.org> |
| 1197 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1197 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1198 | F: arch/arm/mach-ks8695/ | 1198 | F: arch/arm/mach-ks8695/ |
| 1199 | S: Odd Fixes | 1199 | S: Odd Fixes |
| 1200 | 1200 | ||
| 1201 | ARM/MIOA701 MACHINE SUPPORT | 1201 | ARM/MIOA701 MACHINE SUPPORT |
| 1202 | M: Robert Jarzmik <robert.jarzmik@free.fr> | 1202 | M: Robert Jarzmik <robert.jarzmik@free.fr> |
| 1203 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1203 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1204 | F: arch/arm/mach-pxa/mioa701.c | 1204 | F: arch/arm/mach-pxa/mioa701.c |
| 1205 | S: Maintained | 1205 | S: Maintained |
| 1206 | 1206 | ||
| 1207 | ARM/NEC MOBILEPRO 900/c MACHINE SUPPORT | 1207 | ARM/NEC MOBILEPRO 900/c MACHINE SUPPORT |
| 1208 | M: Michael Petchkovsky <mkpetch@internode.on.net> | 1208 | M: Michael Petchkovsky <mkpetch@internode.on.net> |
| 1209 | S: Maintained | 1209 | S: Maintained |
| 1210 | 1210 | ||
| 1211 | ARM/NOMADIK ARCHITECTURE | 1211 | ARM/NOMADIK ARCHITECTURE |
| 1212 | M: Alessandro Rubini <rubini@unipv.it> | 1212 | M: Alessandro Rubini <rubini@unipv.it> |
| 1213 | M: Linus Walleij <linus.walleij@linaro.org> | 1213 | M: Linus Walleij <linus.walleij@linaro.org> |
| 1214 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1214 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1215 | S: Maintained | 1215 | S: Maintained |
| 1216 | F: arch/arm/mach-nomadik/ | 1216 | F: arch/arm/mach-nomadik/ |
| 1217 | F: drivers/pinctrl/nomadik/ | 1217 | F: drivers/pinctrl/nomadik/ |
| 1218 | F: drivers/i2c/busses/i2c-nomadik.c | 1218 | F: drivers/i2c/busses/i2c-nomadik.c |
| 1219 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-nomadik.git | 1219 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-nomadik.git |
| 1220 | 1220 | ||
| 1221 | ARM/OPENMOKO NEO FREERUNNER (GTA02) MACHINE SUPPORT | 1221 | ARM/OPENMOKO NEO FREERUNNER (GTA02) MACHINE SUPPORT |
| 1222 | M: Nelson Castillo <arhuaco@freaks-unidos.net> | 1222 | M: Nelson Castillo <arhuaco@freaks-unidos.net> |
| 1223 | L: openmoko-kernel@lists.openmoko.org (subscribers-only) | 1223 | L: openmoko-kernel@lists.openmoko.org (subscribers-only) |
| 1224 | W: http://wiki.openmoko.org/wiki/Neo_FreeRunner | 1224 | W: http://wiki.openmoko.org/wiki/Neo_FreeRunner |
| 1225 | S: Supported | 1225 | S: Supported |
| 1226 | 1226 | ||
| 1227 | ARM/QUALCOMM MSM MACHINE SUPPORT | 1227 | ARM/QUALCOMM MSM MACHINE SUPPORT |
| 1228 | M: David Brown <davidb@codeaurora.org> | 1228 | M: David Brown <davidb@codeaurora.org> |
| 1229 | M: Daniel Walker <dwalker@fifo99.com> | 1229 | M: Daniel Walker <dwalker@fifo99.com> |
| 1230 | M: Bryan Huntsman <bryanh@codeaurora.org> | 1230 | M: Bryan Huntsman <bryanh@codeaurora.org> |
| 1231 | L: linux-arm-msm@vger.kernel.org | 1231 | L: linux-arm-msm@vger.kernel.org |
| 1232 | F: arch/arm/mach-msm/ | 1232 | F: arch/arm/mach-msm/ |
| 1233 | F: drivers/video/fbdev/msm/ | 1233 | F: drivers/video/fbdev/msm/ |
| 1234 | F: drivers/mmc/host/msm_sdcc.c | 1234 | F: drivers/mmc/host/msm_sdcc.c |
| 1235 | F: drivers/mmc/host/msm_sdcc.h | 1235 | F: drivers/mmc/host/msm_sdcc.h |
| 1236 | F: drivers/tty/serial/msm_serial.h | 1236 | F: drivers/tty/serial/msm_serial.h |
| 1237 | F: drivers/tty/serial/msm_serial.c | 1237 | F: drivers/tty/serial/msm_serial.c |
| 1238 | F: drivers/*/pm8???-* | 1238 | F: drivers/*/pm8???-* |
| 1239 | F: drivers/mfd/ssbi.c | 1239 | F: drivers/mfd/ssbi.c |
| 1240 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davidb/linux-msm.git | 1240 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davidb/linux-msm.git |
| 1241 | S: Maintained | 1241 | S: Maintained |
| 1242 | 1242 | ||
| 1243 | ARM/TOSA MACHINE SUPPORT | 1243 | ARM/TOSA MACHINE SUPPORT |
| 1244 | M: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> | 1244 | M: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> |
| 1245 | M: Dirk Opfer <dirk@opfer-online.de> | 1245 | M: Dirk Opfer <dirk@opfer-online.de> |
| 1246 | S: Maintained | 1246 | S: Maintained |
| 1247 | 1247 | ||
| 1248 | ARM/PALMTX,PALMT5,PALMLD,PALMTE2,PALMTC SUPPORT | 1248 | ARM/PALMTX,PALMT5,PALMLD,PALMTE2,PALMTC SUPPORT |
| 1249 | M: Marek Vasut <marek.vasut@gmail.com> | 1249 | M: Marek Vasut <marek.vasut@gmail.com> |
| 1250 | L: linux-arm-kernel@lists.infradead.org | 1250 | L: linux-arm-kernel@lists.infradead.org |
| 1251 | W: http://hackndev.com | 1251 | W: http://hackndev.com |
| 1252 | S: Maintained | 1252 | S: Maintained |
| 1253 | F: arch/arm/mach-pxa/include/mach/palmtx.h | 1253 | F: arch/arm/mach-pxa/include/mach/palmtx.h |
| 1254 | F: arch/arm/mach-pxa/palmtx.c | 1254 | F: arch/arm/mach-pxa/palmtx.c |
| 1255 | F: arch/arm/mach-pxa/include/mach/palmt5.h | 1255 | F: arch/arm/mach-pxa/include/mach/palmt5.h |
| 1256 | F: arch/arm/mach-pxa/palmt5.c | 1256 | F: arch/arm/mach-pxa/palmt5.c |
| 1257 | F: arch/arm/mach-pxa/include/mach/palmld.h | 1257 | F: arch/arm/mach-pxa/include/mach/palmld.h |
| 1258 | F: arch/arm/mach-pxa/palmld.c | 1258 | F: arch/arm/mach-pxa/palmld.c |
| 1259 | F: arch/arm/mach-pxa/include/mach/palmte2.h | 1259 | F: arch/arm/mach-pxa/include/mach/palmte2.h |
| 1260 | F: arch/arm/mach-pxa/palmte2.c | 1260 | F: arch/arm/mach-pxa/palmte2.c |
| 1261 | F: arch/arm/mach-pxa/include/mach/palmtc.h | 1261 | F: arch/arm/mach-pxa/include/mach/palmtc.h |
| 1262 | F: arch/arm/mach-pxa/palmtc.c | 1262 | F: arch/arm/mach-pxa/palmtc.c |
| 1263 | 1263 | ||
| 1264 | ARM/PALM TREO SUPPORT | 1264 | ARM/PALM TREO SUPPORT |
| 1265 | M: Tomas Cech <sleep_walker@suse.cz> | 1265 | M: Tomas Cech <sleep_walker@suse.cz> |
| 1266 | L: linux-arm-kernel@lists.infradead.org | 1266 | L: linux-arm-kernel@lists.infradead.org |
| 1267 | W: http://hackndev.com | 1267 | W: http://hackndev.com |
| 1268 | S: Maintained | 1268 | S: Maintained |
| 1269 | F: arch/arm/mach-pxa/include/mach/palmtreo.h | 1269 | F: arch/arm/mach-pxa/include/mach/palmtreo.h |
| 1270 | F: arch/arm/mach-pxa/palmtreo.c | 1270 | F: arch/arm/mach-pxa/palmtreo.c |
| 1271 | 1271 | ||
| 1272 | ARM/PALMZ72 SUPPORT | 1272 | ARM/PALMZ72 SUPPORT |
| 1273 | M: Sergey Lapin <slapin@ossfans.org> | 1273 | M: Sergey Lapin <slapin@ossfans.org> |
| 1274 | L: linux-arm-kernel@lists.infradead.org | 1274 | L: linux-arm-kernel@lists.infradead.org |
| 1275 | W: http://hackndev.com | 1275 | W: http://hackndev.com |
| 1276 | S: Maintained | 1276 | S: Maintained |
| 1277 | F: arch/arm/mach-pxa/include/mach/palmz72.h | 1277 | F: arch/arm/mach-pxa/include/mach/palmz72.h |
| 1278 | F: arch/arm/mach-pxa/palmz72.c | 1278 | F: arch/arm/mach-pxa/palmz72.c |
| 1279 | 1279 | ||
| 1280 | ARM/PLEB SUPPORT | 1280 | ARM/PLEB SUPPORT |
| 1281 | M: Peter Chubb <pleb@gelato.unsw.edu.au> | 1281 | M: Peter Chubb <pleb@gelato.unsw.edu.au> |
| 1282 | W: http://www.disy.cse.unsw.edu.au/Hardware/PLEB | 1282 | W: http://www.disy.cse.unsw.edu.au/Hardware/PLEB |
| 1283 | S: Maintained | 1283 | S: Maintained |
| 1284 | 1284 | ||
| 1285 | ARM/PT DIGITAL BOARD PORT | 1285 | ARM/PT DIGITAL BOARD PORT |
| 1286 | M: Stefan Eletzhofer <stefan.eletzhofer@eletztrick.de> | 1286 | M: Stefan Eletzhofer <stefan.eletzhofer@eletztrick.de> |
| 1287 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1287 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1288 | W: http://www.arm.linux.org.uk/ | 1288 | W: http://www.arm.linux.org.uk/ |
| 1289 | S: Maintained | 1289 | S: Maintained |
| 1290 | 1290 | ||
| 1291 | ARM/QUALCOMM SUPPORT | 1291 | ARM/QUALCOMM SUPPORT |
| 1292 | M: Kumar Gala <galak@codeaurora.org> | 1292 | M: Kumar Gala <galak@codeaurora.org> |
| 1293 | M: David Brown <davidb@codeaurora.org> | 1293 | M: David Brown <davidb@codeaurora.org> |
| 1294 | L: linux-arm-msm@vger.kernel.org | 1294 | L: linux-arm-msm@vger.kernel.org |
| 1295 | S: Maintained | 1295 | S: Maintained |
| 1296 | F: arch/arm/mach-qcom/ | 1296 | F: arch/arm/mach-qcom/ |
| 1297 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/galak/linux-qcom.git | 1297 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/galak/linux-qcom.git |
| 1298 | 1298 | ||
| 1299 | ARM/RADISYS ENP2611 MACHINE SUPPORT | 1299 | ARM/RADISYS ENP2611 MACHINE SUPPORT |
| 1300 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1300 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1301 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1301 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1302 | S: Maintained | 1302 | S: Maintained |
| 1303 | 1303 | ||
| 1304 | ARM/RISCPC ARCHITECTURE | 1304 | ARM/RISCPC ARCHITECTURE |
| 1305 | M: Russell King <linux@arm.linux.org.uk> | 1305 | M: Russell King <linux@arm.linux.org.uk> |
| 1306 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1306 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1307 | W: http://www.arm.linux.org.uk/ | 1307 | W: http://www.arm.linux.org.uk/ |
| 1308 | S: Maintained | 1308 | S: Maintained |
| 1309 | F: arch/arm/include/asm/hardware/entry-macro-iomd.S | 1309 | F: arch/arm/include/asm/hardware/entry-macro-iomd.S |
| 1310 | F: arch/arm/include/asm/hardware/ioc.h | 1310 | F: arch/arm/include/asm/hardware/ioc.h |
| 1311 | F: arch/arm/include/asm/hardware/iomd.h | 1311 | F: arch/arm/include/asm/hardware/iomd.h |
| 1312 | F: arch/arm/include/asm/hardware/memc.h | 1312 | F: arch/arm/include/asm/hardware/memc.h |
| 1313 | F: arch/arm/mach-rpc/ | 1313 | F: arch/arm/mach-rpc/ |
| 1314 | F: drivers/net/ethernet/8390/etherh.c | 1314 | F: drivers/net/ethernet/8390/etherh.c |
| 1315 | F: drivers/net/ethernet/i825xx/ether1* | 1315 | F: drivers/net/ethernet/i825xx/ether1* |
| 1316 | F: drivers/net/ethernet/seeq/ether3* | 1316 | F: drivers/net/ethernet/seeq/ether3* |
| 1317 | F: drivers/scsi/arm/ | 1317 | F: drivers/scsi/arm/ |
| 1318 | 1318 | ||
| 1319 | ARM/Rockchip SoC support | 1319 | ARM/Rockchip SoC support |
| 1320 | M: Heiko Stuebner <heiko@sntech.de> | 1320 | M: Heiko Stuebner <heiko@sntech.de> |
| 1321 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1321 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1322 | L: linux-rockchip@lists.infradead.org | 1322 | L: linux-rockchip@lists.infradead.org |
| 1323 | S: Maintained | 1323 | S: Maintained |
| 1324 | F: arch/arm/boot/dts/rk3* | 1324 | F: arch/arm/boot/dts/rk3* |
| 1325 | F: arch/arm/mach-rockchip/ | 1325 | F: arch/arm/mach-rockchip/ |
| 1326 | F: drivers/clk/rockchip/ | 1326 | F: drivers/clk/rockchip/ |
| 1327 | F: drivers/i2c/busses/i2c-rk3x.c | 1327 | F: drivers/i2c/busses/i2c-rk3x.c |
| 1328 | F: drivers/*/*rockchip* | 1328 | F: drivers/*/*rockchip* |
| 1329 | F: drivers/*/*/*rockchip* | 1329 | F: drivers/*/*/*rockchip* |
| 1330 | F: sound/soc/rockchip/ | 1330 | F: sound/soc/rockchip/ |
| 1331 | 1331 | ||
| 1332 | ARM/SAMSUNG EXYNOS ARM ARCHITECTURES | 1332 | ARM/SAMSUNG EXYNOS ARM ARCHITECTURES |
| 1333 | M: Kukjin Kim <kgene@kernel.org> | 1333 | M: Kukjin Kim <kgene@kernel.org> |
| 1334 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1334 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1335 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) | 1335 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) |
| 1336 | S: Maintained | 1336 | S: Maintained |
| 1337 | F: arch/arm/boot/dts/s3c* | 1337 | F: arch/arm/boot/dts/s3c* |
| 1338 | F: arch/arm/boot/dts/exynos* | 1338 | F: arch/arm/boot/dts/exynos* |
| 1339 | F: arch/arm/plat-samsung/ | 1339 | F: arch/arm/plat-samsung/ |
| 1340 | F: arch/arm/mach-s3c24*/ | 1340 | F: arch/arm/mach-s3c24*/ |
| 1341 | F: arch/arm/mach-s3c64xx/ | 1341 | F: arch/arm/mach-s3c64xx/ |
| 1342 | F: arch/arm/mach-s5p*/ | 1342 | F: arch/arm/mach-s5p*/ |
| 1343 | F: arch/arm/mach-exynos*/ | 1343 | F: arch/arm/mach-exynos*/ |
| 1344 | F: drivers/*/*s3c2410* | 1344 | F: drivers/*/*s3c2410* |
| 1345 | F: drivers/*/*/*s3c2410* | 1345 | F: drivers/*/*/*s3c2410* |
| 1346 | F: drivers/spi/spi-s3c* | 1346 | F: drivers/spi/spi-s3c* |
| 1347 | F: sound/soc/samsung/* | 1347 | F: sound/soc/samsung/* |
| 1348 | N: exynos | 1348 | N: exynos |
| 1349 | 1349 | ||
| 1350 | ARM/SAMSUNG MOBILE MACHINE SUPPORT | 1350 | ARM/SAMSUNG MOBILE MACHINE SUPPORT |
| 1351 | M: Kyungmin Park <kyungmin.park@samsung.com> | 1351 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 1352 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1352 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1353 | S: Maintained | 1353 | S: Maintained |
| 1354 | F: arch/arm/mach-s5pv210/ | 1354 | F: arch/arm/mach-s5pv210/ |
| 1355 | 1355 | ||
| 1356 | ARM/SAMSUNG S5P SERIES 2D GRAPHICS ACCELERATION (G2D) SUPPORT | 1356 | ARM/SAMSUNG S5P SERIES 2D GRAPHICS ACCELERATION (G2D) SUPPORT |
| 1357 | M: Kyungmin Park <kyungmin.park@samsung.com> | 1357 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 1358 | M: Kamil Debski <k.debski@samsung.com> | 1358 | M: Kamil Debski <k.debski@samsung.com> |
| 1359 | L: linux-arm-kernel@lists.infradead.org | 1359 | L: linux-arm-kernel@lists.infradead.org |
| 1360 | L: linux-media@vger.kernel.org | 1360 | L: linux-media@vger.kernel.org |
| 1361 | S: Maintained | 1361 | S: Maintained |
| 1362 | F: drivers/media/platform/s5p-g2d/ | 1362 | F: drivers/media/platform/s5p-g2d/ |
| 1363 | 1363 | ||
| 1364 | ARM/SAMSUNG S5P SERIES Multi Format Codec (MFC) SUPPORT | 1364 | ARM/SAMSUNG S5P SERIES Multi Format Codec (MFC) SUPPORT |
| 1365 | M: Kyungmin Park <kyungmin.park@samsung.com> | 1365 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 1366 | M: Kamil Debski <k.debski@samsung.com> | 1366 | M: Kamil Debski <k.debski@samsung.com> |
| 1367 | M: Jeongtae Park <jtp.park@samsung.com> | 1367 | M: Jeongtae Park <jtp.park@samsung.com> |
| 1368 | L: linux-arm-kernel@lists.infradead.org | 1368 | L: linux-arm-kernel@lists.infradead.org |
| 1369 | L: linux-media@vger.kernel.org | 1369 | L: linux-media@vger.kernel.org |
| 1370 | S: Maintained | 1370 | S: Maintained |
| 1371 | F: arch/arm/plat-samsung/s5p-dev-mfc.c | 1371 | F: arch/arm/plat-samsung/s5p-dev-mfc.c |
| 1372 | F: drivers/media/platform/s5p-mfc/ | 1372 | F: drivers/media/platform/s5p-mfc/ |
| 1373 | 1373 | ||
| 1374 | ARM/SAMSUNG S5P SERIES TV SUBSYSTEM SUPPORT | 1374 | ARM/SAMSUNG S5P SERIES TV SUBSYSTEM SUPPORT |
| 1375 | M: Kyungmin Park <kyungmin.park@samsung.com> | 1375 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 1376 | M: Tomasz Stanislawski <t.stanislaws@samsung.com> | 1376 | M: Tomasz Stanislawski <t.stanislaws@samsung.com> |
| 1377 | L: linux-arm-kernel@lists.infradead.org | 1377 | L: linux-arm-kernel@lists.infradead.org |
| 1378 | L: linux-media@vger.kernel.org | 1378 | L: linux-media@vger.kernel.org |
| 1379 | S: Maintained | 1379 | S: Maintained |
| 1380 | F: drivers/media/platform/s5p-tv/ | 1380 | F: drivers/media/platform/s5p-tv/ |
| 1381 | 1381 | ||
| 1382 | ARM/SHMOBILE ARM ARCHITECTURE | 1382 | ARM/SHMOBILE ARM ARCHITECTURE |
| 1383 | M: Simon Horman <horms@verge.net.au> | 1383 | M: Simon Horman <horms@verge.net.au> |
| 1384 | M: Magnus Damm <magnus.damm@gmail.com> | 1384 | M: Magnus Damm <magnus.damm@gmail.com> |
| 1385 | L: linux-sh@vger.kernel.org | 1385 | L: linux-sh@vger.kernel.org |
| 1386 | W: http://oss.renesas.com | 1386 | W: http://oss.renesas.com |
| 1387 | Q: http://patchwork.kernel.org/project/linux-sh/list/ | 1387 | Q: http://patchwork.kernel.org/project/linux-sh/list/ |
| 1388 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas.git next | 1388 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas.git next |
| 1389 | S: Supported | 1389 | S: Supported |
| 1390 | F: arch/arm/boot/dts/emev2* | 1390 | F: arch/arm/boot/dts/emev2* |
| 1391 | F: arch/arm/boot/dts/r7s* | 1391 | F: arch/arm/boot/dts/r7s* |
| 1392 | F: arch/arm/boot/dts/r8a* | 1392 | F: arch/arm/boot/dts/r8a* |
| 1393 | F: arch/arm/boot/dts/sh* | 1393 | F: arch/arm/boot/dts/sh* |
| 1394 | F: arch/arm/configs/ape6evm_defconfig | 1394 | F: arch/arm/configs/ape6evm_defconfig |
| 1395 | F: arch/arm/configs/armadillo800eva_defconfig | 1395 | F: arch/arm/configs/armadillo800eva_defconfig |
| 1396 | F: arch/arm/configs/bockw_defconfig | 1396 | F: arch/arm/configs/bockw_defconfig |
| 1397 | F: arch/arm/configs/kzm9g_defconfig | 1397 | F: arch/arm/configs/kzm9g_defconfig |
| 1398 | F: arch/arm/configs/lager_defconfig | 1398 | F: arch/arm/configs/lager_defconfig |
| 1399 | F: arch/arm/configs/mackerel_defconfig | 1399 | F: arch/arm/configs/mackerel_defconfig |
| 1400 | F: arch/arm/configs/marzen_defconfig | 1400 | F: arch/arm/configs/marzen_defconfig |
| 1401 | F: arch/arm/configs/shmobile_defconfig | 1401 | F: arch/arm/configs/shmobile_defconfig |
| 1402 | F: arch/arm/include/debug/renesas-scif.S | 1402 | F: arch/arm/include/debug/renesas-scif.S |
| 1403 | F: arch/arm/mach-shmobile/ | 1403 | F: arch/arm/mach-shmobile/ |
| 1404 | F: drivers/sh/ | 1404 | F: drivers/sh/ |
| 1405 | 1405 | ||
| 1406 | ARM/SOCFPGA ARCHITECTURE | 1406 | ARM/SOCFPGA ARCHITECTURE |
| 1407 | M: Dinh Nguyen <dinguyen@opensource.altera.com> | 1407 | M: Dinh Nguyen <dinguyen@opensource.altera.com> |
| 1408 | S: Maintained | 1408 | S: Maintained |
| 1409 | F: arch/arm/mach-socfpga/ | 1409 | F: arch/arm/mach-socfpga/ |
| 1410 | W: http://www.rocketboards.org | 1410 | W: http://www.rocketboards.org |
| 1411 | T: git://git.rocketboards.org/linux-socfpga.git | 1411 | T: git://git.rocketboards.org/linux-socfpga.git |
| 1412 | T: git://git.rocketboards.org/linux-socfpga-next.git | 1412 | T: git://git.rocketboards.org/linux-socfpga-next.git |
| 1413 | 1413 | ||
| 1414 | ARM/SOCFPGA CLOCK FRAMEWORK SUPPORT | 1414 | ARM/SOCFPGA CLOCK FRAMEWORK SUPPORT |
| 1415 | M: Dinh Nguyen <dinguyen@opensource.altera.com> | 1415 | M: Dinh Nguyen <dinguyen@opensource.altera.com> |
| 1416 | S: Maintained | 1416 | S: Maintained |
| 1417 | F: drivers/clk/socfpga/ | 1417 | F: drivers/clk/socfpga/ |
| 1418 | 1418 | ||
| 1419 | ARM/SOCFPGA EDAC SUPPORT | 1419 | ARM/SOCFPGA EDAC SUPPORT |
| 1420 | M: Thor Thayer <tthayer@opensource.altera.com> | 1420 | M: Thor Thayer <tthayer@opensource.altera.com> |
| 1421 | S: Maintained | 1421 | S: Maintained |
| 1422 | F: drivers/edac/altera_edac. | 1422 | F: drivers/edac/altera_edac. |
| 1423 | 1423 | ||
| 1424 | ARM/STI ARCHITECTURE | 1424 | ARM/STI ARCHITECTURE |
| 1425 | M: Srinivas Kandagatla <srinivas.kandagatla@gmail.com> | 1425 | M: Srinivas Kandagatla <srinivas.kandagatla@gmail.com> |
| 1426 | M: Maxime Coquelin <maxime.coquelin@st.com> | 1426 | M: Maxime Coquelin <maxime.coquelin@st.com> |
| 1427 | M: Patrice Chotard <patrice.chotard@st.com> | 1427 | M: Patrice Chotard <patrice.chotard@st.com> |
| 1428 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1428 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1429 | L: kernel@stlinux.com | 1429 | L: kernel@stlinux.com |
| 1430 | W: http://www.stlinux.com | 1430 | W: http://www.stlinux.com |
| 1431 | S: Maintained | 1431 | S: Maintained |
| 1432 | F: arch/arm/mach-sti/ | 1432 | F: arch/arm/mach-sti/ |
| 1433 | F: arch/arm/boot/dts/sti* | 1433 | F: arch/arm/boot/dts/sti* |
| 1434 | F: drivers/clocksource/arm_global_timer.c | 1434 | F: drivers/clocksource/arm_global_timer.c |
| 1435 | F: drivers/i2c/busses/i2c-st.c | 1435 | F: drivers/i2c/busses/i2c-st.c |
| 1436 | F: drivers/media/rc/st_rc.c | 1436 | F: drivers/media/rc/st_rc.c |
| 1437 | F: drivers/mmc/host/sdhci-st.c | 1437 | F: drivers/mmc/host/sdhci-st.c |
| 1438 | F: drivers/phy/phy-stih407-usb.c | 1438 | F: drivers/phy/phy-stih407-usb.c |
| 1439 | F: drivers/phy/phy-stih41x-usb.c | 1439 | F: drivers/phy/phy-stih41x-usb.c |
| 1440 | F: drivers/pinctrl/pinctrl-st.c | 1440 | F: drivers/pinctrl/pinctrl-st.c |
| 1441 | F: drivers/reset/sti/ | 1441 | F: drivers/reset/sti/ |
| 1442 | F: drivers/tty/serial/st-asc.c | 1442 | F: drivers/tty/serial/st-asc.c |
| 1443 | F: drivers/usb/dwc3/dwc3-st.c | 1443 | F: drivers/usb/dwc3/dwc3-st.c |
| 1444 | F: drivers/usb/host/ehci-st.c | 1444 | F: drivers/usb/host/ehci-st.c |
| 1445 | F: drivers/usb/host/ohci-st.c | 1445 | F: drivers/usb/host/ohci-st.c |
| 1446 | F: drivers/ata/ahci_st.c | 1446 | F: drivers/ata/ahci_st.c |
| 1447 | 1447 | ||
| 1448 | ARM/TECHNOLOGIC SYSTEMS TS7250 MACHINE SUPPORT | 1448 | ARM/TECHNOLOGIC SYSTEMS TS7250 MACHINE SUPPORT |
| 1449 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1449 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1450 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1450 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1451 | S: Maintained | 1451 | S: Maintained |
| 1452 | 1452 | ||
| 1453 | ARM/TETON BGA MACHINE SUPPORT | 1453 | ARM/TETON BGA MACHINE SUPPORT |
| 1454 | M: "Mark F. Brown" <mark.brown314@gmail.com> | 1454 | M: "Mark F. Brown" <mark.brown314@gmail.com> |
| 1455 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1455 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1456 | S: Maintained | 1456 | S: Maintained |
| 1457 | 1457 | ||
| 1458 | ARM/THECUS N2100 MACHINE SUPPORT | 1458 | ARM/THECUS N2100 MACHINE SUPPORT |
| 1459 | M: Lennert Buytenhek <kernel@wantstofly.org> | 1459 | M: Lennert Buytenhek <kernel@wantstofly.org> |
| 1460 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1460 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1461 | S: Maintained | 1461 | S: Maintained |
| 1462 | 1462 | ||
| 1463 | ARM/NUVOTON W90X900 ARM ARCHITECTURE | 1463 | ARM/NUVOTON W90X900 ARM ARCHITECTURE |
| 1464 | M: Wan ZongShun <mcuos.com@gmail.com> | 1464 | M: Wan ZongShun <mcuos.com@gmail.com> |
| 1465 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1465 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1466 | W: http://www.mcuos.com | 1466 | W: http://www.mcuos.com |
| 1467 | S: Maintained | 1467 | S: Maintained |
| 1468 | F: arch/arm/mach-w90x900/ | 1468 | F: arch/arm/mach-w90x900/ |
| 1469 | F: drivers/input/keyboard/w90p910_keypad.c | 1469 | F: drivers/input/keyboard/w90p910_keypad.c |
| 1470 | F: drivers/input/touchscreen/w90p910_ts.c | 1470 | F: drivers/input/touchscreen/w90p910_ts.c |
| 1471 | F: drivers/watchdog/nuc900_wdt.c | 1471 | F: drivers/watchdog/nuc900_wdt.c |
| 1472 | F: drivers/net/ethernet/nuvoton/w90p910_ether.c | 1472 | F: drivers/net/ethernet/nuvoton/w90p910_ether.c |
| 1473 | F: drivers/mtd/nand/nuc900_nand.c | 1473 | F: drivers/mtd/nand/nuc900_nand.c |
| 1474 | F: drivers/rtc/rtc-nuc900.c | 1474 | F: drivers/rtc/rtc-nuc900.c |
| 1475 | F: drivers/spi/spi-nuc900.c | 1475 | F: drivers/spi/spi-nuc900.c |
| 1476 | F: drivers/usb/host/ehci-w90x900.c | 1476 | F: drivers/usb/host/ehci-w90x900.c |
| 1477 | F: drivers/video/fbdev/nuc900fb.c | 1477 | F: drivers/video/fbdev/nuc900fb.c |
| 1478 | 1478 | ||
| 1479 | ARM/U300 MACHINE SUPPORT | 1479 | ARM/U300 MACHINE SUPPORT |
| 1480 | M: Linus Walleij <linus.walleij@linaro.org> | 1480 | M: Linus Walleij <linus.walleij@linaro.org> |
| 1481 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1481 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1482 | S: Supported | 1482 | S: Supported |
| 1483 | F: arch/arm/mach-u300/ | 1483 | F: arch/arm/mach-u300/ |
| 1484 | F: drivers/clocksource/timer-u300.c | 1484 | F: drivers/clocksource/timer-u300.c |
| 1485 | F: drivers/i2c/busses/i2c-stu300.c | 1485 | F: drivers/i2c/busses/i2c-stu300.c |
| 1486 | F: drivers/rtc/rtc-coh901331.c | 1486 | F: drivers/rtc/rtc-coh901331.c |
| 1487 | F: drivers/watchdog/coh901327_wdt.c | 1487 | F: drivers/watchdog/coh901327_wdt.c |
| 1488 | F: drivers/dma/coh901318* | 1488 | F: drivers/dma/coh901318* |
| 1489 | F: drivers/mfd/ab3100* | 1489 | F: drivers/mfd/ab3100* |
| 1490 | F: drivers/rtc/rtc-ab3100.c | 1490 | F: drivers/rtc/rtc-ab3100.c |
| 1491 | F: drivers/rtc/rtc-coh901331.c | 1491 | F: drivers/rtc/rtc-coh901331.c |
| 1492 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-stericsson.git | 1492 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-stericsson.git |
| 1493 | 1493 | ||
| 1494 | ARM/Ux500 ARM ARCHITECTURE | 1494 | ARM/Ux500 ARM ARCHITECTURE |
| 1495 | M: Linus Walleij <linus.walleij@linaro.org> | 1495 | M: Linus Walleij <linus.walleij@linaro.org> |
| 1496 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1496 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1497 | S: Maintained | 1497 | S: Maintained |
| 1498 | F: arch/arm/mach-ux500/ | 1498 | F: arch/arm/mach-ux500/ |
| 1499 | F: drivers/clocksource/clksrc-dbx500-prcmu.c | 1499 | F: drivers/clocksource/clksrc-dbx500-prcmu.c |
| 1500 | F: drivers/dma/ste_dma40* | 1500 | F: drivers/dma/ste_dma40* |
| 1501 | F: drivers/hwspinlock/u8500_hsem.c | 1501 | F: drivers/hwspinlock/u8500_hsem.c |
| 1502 | F: drivers/mfd/abx500* | 1502 | F: drivers/mfd/abx500* |
| 1503 | F: drivers/mfd/ab8500* | 1503 | F: drivers/mfd/ab8500* |
| 1504 | F: drivers/mfd/dbx500* | 1504 | F: drivers/mfd/dbx500* |
| 1505 | F: drivers/mfd/db8500* | 1505 | F: drivers/mfd/db8500* |
| 1506 | F: drivers/pinctrl/nomadik/pinctrl-ab* | 1506 | F: drivers/pinctrl/nomadik/pinctrl-ab* |
| 1507 | F: drivers/pinctrl/nomadik/pinctrl-nomadik* | 1507 | F: drivers/pinctrl/nomadik/pinctrl-nomadik* |
| 1508 | F: drivers/rtc/rtc-ab8500.c | 1508 | F: drivers/rtc/rtc-ab8500.c |
| 1509 | F: drivers/rtc/rtc-pl031.c | 1509 | F: drivers/rtc/rtc-pl031.c |
| 1510 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-stericsson.git | 1510 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-stericsson.git |
| 1511 | 1511 | ||
| 1512 | ARM/Ux500 CLOCK FRAMEWORK SUPPORT | 1512 | ARM/Ux500 CLOCK FRAMEWORK SUPPORT |
| 1513 | M: Ulf Hansson <ulf.hansson@linaro.org> | 1513 | M: Ulf Hansson <ulf.hansson@linaro.org> |
| 1514 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1514 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1515 | T: git git://git.linaro.org/people/ulfh/clk.git | 1515 | T: git git://git.linaro.org/people/ulfh/clk.git |
| 1516 | S: Maintained | 1516 | S: Maintained |
| 1517 | F: drivers/clk/ux500/ | 1517 | F: drivers/clk/ux500/ |
| 1518 | F: include/linux/platform_data/clk-ux500.h | 1518 | F: include/linux/platform_data/clk-ux500.h |
| 1519 | 1519 | ||
| 1520 | ARM/VERSATILE EXPRESS PLATFORM | 1520 | ARM/VERSATILE EXPRESS PLATFORM |
| 1521 | M: Liviu Dudau <liviu.dudau@arm.com> | 1521 | M: Liviu Dudau <liviu.dudau@arm.com> |
| 1522 | M: Sudeep Holla <sudeep.holla@arm.com> | 1522 | M: Sudeep Holla <sudeep.holla@arm.com> |
| 1523 | M: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> | 1523 | M: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> |
| 1524 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1524 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1525 | S: Maintained | 1525 | S: Maintained |
| 1526 | F: arch/arm/boot/dts/vexpress* | 1526 | F: arch/arm/boot/dts/vexpress* |
| 1527 | F: arch/arm/mach-vexpress/ | 1527 | F: arch/arm/mach-vexpress/ |
| 1528 | F: */*/vexpress* | 1528 | F: */*/vexpress* |
| 1529 | F: */*/*/vexpress* | 1529 | F: */*/*/vexpress* |
| 1530 | F: drivers/clk/versatile/clk-vexpress-osc.c | 1530 | F: drivers/clk/versatile/clk-vexpress-osc.c |
| 1531 | F: drivers/clocksource/versatile.c | 1531 | F: drivers/clocksource/versatile.c |
| 1532 | 1532 | ||
| 1533 | ARM/VFP SUPPORT | 1533 | ARM/VFP SUPPORT |
| 1534 | M: Russell King <linux@arm.linux.org.uk> | 1534 | M: Russell King <linux@arm.linux.org.uk> |
| 1535 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1535 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1536 | W: http://www.arm.linux.org.uk/ | 1536 | W: http://www.arm.linux.org.uk/ |
| 1537 | S: Maintained | 1537 | S: Maintained |
| 1538 | F: arch/arm/vfp/ | 1538 | F: arch/arm/vfp/ |
| 1539 | 1539 | ||
| 1540 | ARM/VOIPAC PXA270 SUPPORT | 1540 | ARM/VOIPAC PXA270 SUPPORT |
| 1541 | M: Marek Vasut <marek.vasut@gmail.com> | 1541 | M: Marek Vasut <marek.vasut@gmail.com> |
| 1542 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1542 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1543 | S: Maintained | 1543 | S: Maintained |
| 1544 | F: arch/arm/mach-pxa/vpac270.c | 1544 | F: arch/arm/mach-pxa/vpac270.c |
| 1545 | F: arch/arm/mach-pxa/include/mach/vpac270.h | 1545 | F: arch/arm/mach-pxa/include/mach/vpac270.h |
| 1546 | 1546 | ||
| 1547 | ARM/VT8500 ARM ARCHITECTURE | 1547 | ARM/VT8500 ARM ARCHITECTURE |
| 1548 | M: Tony Prisk <linux@prisktech.co.nz> | 1548 | M: Tony Prisk <linux@prisktech.co.nz> |
| 1549 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1549 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1550 | S: Maintained | 1550 | S: Maintained |
| 1551 | F: arch/arm/mach-vt8500/ | 1551 | F: arch/arm/mach-vt8500/ |
| 1552 | F: drivers/clocksource/vt8500_timer.c | 1552 | F: drivers/clocksource/vt8500_timer.c |
| 1553 | F: drivers/i2c/busses/i2c-wmt.c | 1553 | F: drivers/i2c/busses/i2c-wmt.c |
| 1554 | F: drivers/mmc/host/wmt-sdmmc.c | 1554 | F: drivers/mmc/host/wmt-sdmmc.c |
| 1555 | F: drivers/pwm/pwm-vt8500.c | 1555 | F: drivers/pwm/pwm-vt8500.c |
| 1556 | F: drivers/rtc/rtc-vt8500.c | 1556 | F: drivers/rtc/rtc-vt8500.c |
| 1557 | F: drivers/tty/serial/vt8500_serial.c | 1557 | F: drivers/tty/serial/vt8500_serial.c |
| 1558 | F: drivers/usb/host/ehci-platform.c | 1558 | F: drivers/usb/host/ehci-platform.c |
| 1559 | F: drivers/usb/host/uhci-platform.c | 1559 | F: drivers/usb/host/uhci-platform.c |
| 1560 | F: drivers/video/fbdev/vt8500lcdfb.* | 1560 | F: drivers/video/fbdev/vt8500lcdfb.* |
| 1561 | F: drivers/video/fbdev/wm8505fb* | 1561 | F: drivers/video/fbdev/wm8505fb* |
| 1562 | F: drivers/video/fbdev/wmt_ge_rops.* | 1562 | F: drivers/video/fbdev/wmt_ge_rops.* |
| 1563 | 1563 | ||
| 1564 | ARM/ZIPIT Z2 SUPPORT | 1564 | ARM/ZIPIT Z2 SUPPORT |
| 1565 | M: Marek Vasut <marek.vasut@gmail.com> | 1565 | M: Marek Vasut <marek.vasut@gmail.com> |
| 1566 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1566 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1567 | S: Maintained | 1567 | S: Maintained |
| 1568 | F: arch/arm/mach-pxa/z2.c | 1568 | F: arch/arm/mach-pxa/z2.c |
| 1569 | F: arch/arm/mach-pxa/include/mach/z2.h | 1569 | F: arch/arm/mach-pxa/include/mach/z2.h |
| 1570 | 1570 | ||
| 1571 | ARM/ZYNQ ARCHITECTURE | 1571 | ARM/ZYNQ ARCHITECTURE |
| 1572 | M: Michal Simek <michal.simek@xilinx.com> | 1572 | M: Michal Simek <michal.simek@xilinx.com> |
| 1573 | R: Sรถren Brinkmann <soren.brinkmann@xilinx.com> | 1573 | R: Sรถren Brinkmann <soren.brinkmann@xilinx.com> |
| 1574 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1574 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1575 | W: http://wiki.xilinx.com | 1575 | W: http://wiki.xilinx.com |
| 1576 | T: git git://git.xilinx.com/linux-xlnx.git | 1576 | T: git git://git.xilinx.com/linux-xlnx.git |
| 1577 | S: Supported | 1577 | S: Supported |
| 1578 | F: arch/arm/mach-zynq/ | 1578 | F: arch/arm/mach-zynq/ |
| 1579 | F: drivers/cpuidle/cpuidle-zynq.c | 1579 | F: drivers/cpuidle/cpuidle-zynq.c |
| 1580 | F: drivers/block/xsysace.c | 1580 | F: drivers/block/xsysace.c |
| 1581 | N: zynq | 1581 | N: zynq |
| 1582 | N: xilinx | 1582 | N: xilinx |
| 1583 | F: drivers/clocksource/cadence_ttc_timer.c | 1583 | F: drivers/clocksource/cadence_ttc_timer.c |
| 1584 | F: drivers/i2c/busses/i2c-cadence.c | 1584 | F: drivers/i2c/busses/i2c-cadence.c |
| 1585 | F: drivers/mmc/host/sdhci-of-arasan.c | 1585 | F: drivers/mmc/host/sdhci-of-arasan.c |
| 1586 | 1586 | ||
| 1587 | ARM SMMU DRIVER | 1587 | ARM SMMU DRIVER |
| 1588 | M: Will Deacon <will.deacon@arm.com> | 1588 | M: Will Deacon <will.deacon@arm.com> |
| 1589 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1589 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1590 | S: Maintained | 1590 | S: Maintained |
| 1591 | F: drivers/iommu/arm-smmu.c | 1591 | F: drivers/iommu/arm-smmu.c |
| 1592 | 1592 | ||
| 1593 | ARM64 PORT (AARCH64 ARCHITECTURE) | 1593 | ARM64 PORT (AARCH64 ARCHITECTURE) |
| 1594 | M: Catalin Marinas <catalin.marinas@arm.com> | 1594 | M: Catalin Marinas <catalin.marinas@arm.com> |
| 1595 | M: Will Deacon <will.deacon@arm.com> | 1595 | M: Will Deacon <will.deacon@arm.com> |
| 1596 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1596 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1597 | S: Maintained | 1597 | S: Maintained |
| 1598 | F: arch/arm64/ | 1598 | F: arch/arm64/ |
| 1599 | F: Documentation/arm64/ | 1599 | F: Documentation/arm64/ |
| 1600 | 1600 | ||
| 1601 | AS3645A LED FLASH CONTROLLER DRIVER | 1601 | AS3645A LED FLASH CONTROLLER DRIVER |
| 1602 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 1602 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 1603 | L: linux-media@vger.kernel.org | 1603 | L: linux-media@vger.kernel.org |
| 1604 | T: git git://linuxtv.org/media_tree.git | 1604 | T: git git://linuxtv.org/media_tree.git |
| 1605 | S: Maintained | 1605 | S: Maintained |
| 1606 | F: drivers/media/i2c/as3645a.c | 1606 | F: drivers/media/i2c/as3645a.c |
| 1607 | F: include/media/as3645a.h | 1607 | F: include/media/as3645a.h |
| 1608 | 1608 | ||
| 1609 | ASC7621 HARDWARE MONITOR DRIVER | 1609 | ASC7621 HARDWARE MONITOR DRIVER |
| 1610 | M: George Joseph <george.joseph@fairview5.com> | 1610 | M: George Joseph <george.joseph@fairview5.com> |
| 1611 | L: lm-sensors@lm-sensors.org | 1611 | L: lm-sensors@lm-sensors.org |
| 1612 | S: Maintained | 1612 | S: Maintained |
| 1613 | F: Documentation/hwmon/asc7621 | 1613 | F: Documentation/hwmon/asc7621 |
| 1614 | F: drivers/hwmon/asc7621.c | 1614 | F: drivers/hwmon/asc7621.c |
| 1615 | 1615 | ||
| 1616 | ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS | 1616 | ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS |
| 1617 | M: Corentin Chary <corentin.chary@gmail.com> | 1617 | M: Corentin Chary <corentin.chary@gmail.com> |
| 1618 | L: acpi4asus-user@lists.sourceforge.net | 1618 | L: acpi4asus-user@lists.sourceforge.net |
| 1619 | L: platform-driver-x86@vger.kernel.org | 1619 | L: platform-driver-x86@vger.kernel.org |
| 1620 | W: http://acpi4asus.sf.net | 1620 | W: http://acpi4asus.sf.net |
| 1621 | S: Maintained | 1621 | S: Maintained |
| 1622 | F: drivers/platform/x86/asus*.c | 1622 | F: drivers/platform/x86/asus*.c |
| 1623 | F: drivers/platform/x86/eeepc*.c | 1623 | F: drivers/platform/x86/eeepc*.c |
| 1624 | 1624 | ||
| 1625 | ASYNCHRONOUS TRANSFERS/TRANSFORMS (IOAT) API | 1625 | ASYNCHRONOUS TRANSFERS/TRANSFORMS (IOAT) API |
| 1626 | R: Dan Williams <dan.j.williams@intel.com> | 1626 | R: Dan Williams <dan.j.williams@intel.com> |
| 1627 | W: http://sourceforge.net/projects/xscaleiop | 1627 | W: http://sourceforge.net/projects/xscaleiop |
| 1628 | S: Odd fixes | 1628 | S: Odd fixes |
| 1629 | F: Documentation/crypto/async-tx-api.txt | 1629 | F: Documentation/crypto/async-tx-api.txt |
| 1630 | F: crypto/async_tx/ | 1630 | F: crypto/async_tx/ |
| 1631 | F: drivers/dma/ | 1631 | F: drivers/dma/ |
| 1632 | F: include/linux/dmaengine.h | 1632 | F: include/linux/dmaengine.h |
| 1633 | F: include/linux/async_tx.h | 1633 | F: include/linux/async_tx.h |
| 1634 | 1634 | ||
| 1635 | AT24 EEPROM DRIVER | 1635 | AT24 EEPROM DRIVER |
| 1636 | M: Wolfram Sang <wsa@the-dreams.de> | 1636 | M: Wolfram Sang <wsa@the-dreams.de> |
| 1637 | L: linux-i2c@vger.kernel.org | 1637 | L: linux-i2c@vger.kernel.org |
| 1638 | S: Maintained | 1638 | S: Maintained |
| 1639 | F: drivers/misc/eeprom/at24.c | 1639 | F: drivers/misc/eeprom/at24.c |
| 1640 | F: include/linux/platform_data/at24.h | 1640 | F: include/linux/platform_data/at24.h |
| 1641 | 1641 | ||
| 1642 | ATA OVER ETHERNET (AOE) DRIVER | 1642 | ATA OVER ETHERNET (AOE) DRIVER |
| 1643 | M: "Ed L. Cashin" <ecashin@coraid.com> | 1643 | M: "Ed L. Cashin" <ecashin@coraid.com> |
| 1644 | W: http://support.coraid.com/support/linux | 1644 | W: http://support.coraid.com/support/linux |
| 1645 | S: Supported | 1645 | S: Supported |
| 1646 | F: Documentation/aoe/ | 1646 | F: Documentation/aoe/ |
| 1647 | F: drivers/block/aoe/ | 1647 | F: drivers/block/aoe/ |
| 1648 | 1648 | ||
| 1649 | ATHEROS ATH GENERIC UTILITIES | 1649 | ATHEROS ATH GENERIC UTILITIES |
| 1650 | M: "Luis R. Rodriguez" <mcgrof@do-not-panic.com> | 1650 | M: "Luis R. Rodriguez" <mcgrof@do-not-panic.com> |
| 1651 | L: linux-wireless@vger.kernel.org | 1651 | L: linux-wireless@vger.kernel.org |
| 1652 | S: Supported | 1652 | S: Supported |
| 1653 | F: drivers/net/wireless/ath/* | 1653 | F: drivers/net/wireless/ath/* |
| 1654 | 1654 | ||
| 1655 | ATHEROS ATH5K WIRELESS DRIVER | 1655 | ATHEROS ATH5K WIRELESS DRIVER |
| 1656 | M: Jiri Slaby <jirislaby@gmail.com> | 1656 | M: Jiri Slaby <jirislaby@gmail.com> |
| 1657 | M: Nick Kossifidis <mickflemm@gmail.com> | 1657 | M: Nick Kossifidis <mickflemm@gmail.com> |
| 1658 | M: "Luis R. Rodriguez" <mcgrof@do-not-panic.com> | 1658 | M: "Luis R. Rodriguez" <mcgrof@do-not-panic.com> |
| 1659 | L: linux-wireless@vger.kernel.org | 1659 | L: linux-wireless@vger.kernel.org |
| 1660 | L: ath5k-devel@lists.ath5k.org | 1660 | L: ath5k-devel@lists.ath5k.org |
| 1661 | W: http://wireless.kernel.org/en/users/Drivers/ath5k | 1661 | W: http://wireless.kernel.org/en/users/Drivers/ath5k |
| 1662 | S: Maintained | 1662 | S: Maintained |
| 1663 | F: drivers/net/wireless/ath/ath5k/ | 1663 | F: drivers/net/wireless/ath/ath5k/ |
| 1664 | 1664 | ||
| 1665 | ATHEROS ATH6KL WIRELESS DRIVER | 1665 | ATHEROS ATH6KL WIRELESS DRIVER |
| 1666 | M: Kalle Valo <kvalo@qca.qualcomm.com> | 1666 | M: Kalle Valo <kvalo@qca.qualcomm.com> |
| 1667 | L: linux-wireless@vger.kernel.org | 1667 | L: linux-wireless@vger.kernel.org |
| 1668 | W: http://wireless.kernel.org/en/users/Drivers/ath6kl | 1668 | W: http://wireless.kernel.org/en/users/Drivers/ath6kl |
| 1669 | T: git git://github.com/kvalo/ath.git | 1669 | T: git git://github.com/kvalo/ath.git |
| 1670 | S: Supported | 1670 | S: Supported |
| 1671 | F: drivers/net/wireless/ath/ath6kl/ | 1671 | F: drivers/net/wireless/ath/ath6kl/ |
| 1672 | 1672 | ||
| 1673 | WILOCITY WIL6210 WIRELESS DRIVER | 1673 | WILOCITY WIL6210 WIRELESS DRIVER |
| 1674 | M: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com> | 1674 | M: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com> |
| 1675 | L: linux-wireless@vger.kernel.org | 1675 | L: linux-wireless@vger.kernel.org |
| 1676 | L: wil6210@qca.qualcomm.com | 1676 | L: wil6210@qca.qualcomm.com |
| 1677 | S: Supported | 1677 | S: Supported |
| 1678 | W: http://wireless.kernel.org/en/users/Drivers/wil6210 | 1678 | W: http://wireless.kernel.org/en/users/Drivers/wil6210 |
| 1679 | F: drivers/net/wireless/ath/wil6210/ | 1679 | F: drivers/net/wireless/ath/wil6210/ |
| 1680 | F: include/uapi/linux/wil6210_uapi.h | 1680 | F: include/uapi/linux/wil6210_uapi.h |
| 1681 | 1681 | ||
| 1682 | CARL9170 LINUX COMMUNITY WIRELESS DRIVER | 1682 | CARL9170 LINUX COMMUNITY WIRELESS DRIVER |
| 1683 | M: Christian Lamparter <chunkeey@googlemail.com> | 1683 | M: Christian Lamparter <chunkeey@googlemail.com> |
| 1684 | L: linux-wireless@vger.kernel.org | 1684 | L: linux-wireless@vger.kernel.org |
| 1685 | W: http://wireless.kernel.org/en/users/Drivers/carl9170 | 1685 | W: http://wireless.kernel.org/en/users/Drivers/carl9170 |
| 1686 | S: Maintained | 1686 | S: Maintained |
| 1687 | F: drivers/net/wireless/ath/carl9170/ | 1687 | F: drivers/net/wireless/ath/carl9170/ |
| 1688 | 1688 | ||
| 1689 | ATK0110 HWMON DRIVER | 1689 | ATK0110 HWMON DRIVER |
| 1690 | M: Luca Tettamanti <kronos.it@gmail.com> | 1690 | M: Luca Tettamanti <kronos.it@gmail.com> |
| 1691 | L: lm-sensors@lm-sensors.org | 1691 | L: lm-sensors@lm-sensors.org |
| 1692 | S: Maintained | 1692 | S: Maintained |
| 1693 | F: drivers/hwmon/asus_atk0110.c | 1693 | F: drivers/hwmon/asus_atk0110.c |
| 1694 | 1694 | ||
| 1695 | ATI_REMOTE2 DRIVER | 1695 | ATI_REMOTE2 DRIVER |
| 1696 | M: Ville Syrjala <syrjala@sci.fi> | 1696 | M: Ville Syrjala <syrjala@sci.fi> |
| 1697 | S: Maintained | 1697 | S: Maintained |
| 1698 | F: drivers/input/misc/ati_remote2.c | 1698 | F: drivers/input/misc/ati_remote2.c |
| 1699 | 1699 | ||
| 1700 | ATLX ETHERNET DRIVERS | 1700 | ATLX ETHERNET DRIVERS |
| 1701 | M: Jay Cliburn <jcliburn@gmail.com> | 1701 | M: Jay Cliburn <jcliburn@gmail.com> |
| 1702 | M: Chris Snook <chris.snook@gmail.com> | 1702 | M: Chris Snook <chris.snook@gmail.com> |
| 1703 | L: netdev@vger.kernel.org | 1703 | L: netdev@vger.kernel.org |
| 1704 | W: http://sourceforge.net/projects/atl1 | 1704 | W: http://sourceforge.net/projects/atl1 |
| 1705 | W: http://atl1.sourceforge.net | 1705 | W: http://atl1.sourceforge.net |
| 1706 | S: Maintained | 1706 | S: Maintained |
| 1707 | F: drivers/net/ethernet/atheros/ | 1707 | F: drivers/net/ethernet/atheros/ |
| 1708 | 1708 | ||
| 1709 | ATM | 1709 | ATM |
| 1710 | M: Chas Williams <chas@cmf.nrl.navy.mil> | 1710 | M: Chas Williams <chas@cmf.nrl.navy.mil> |
| 1711 | L: linux-atm-general@lists.sourceforge.net (moderated for non-subscribers) | 1711 | L: linux-atm-general@lists.sourceforge.net (moderated for non-subscribers) |
| 1712 | L: netdev@vger.kernel.org | 1712 | L: netdev@vger.kernel.org |
| 1713 | W: http://linux-atm.sourceforge.net | 1713 | W: http://linux-atm.sourceforge.net |
| 1714 | S: Maintained | 1714 | S: Maintained |
| 1715 | F: drivers/atm/ | 1715 | F: drivers/atm/ |
| 1716 | F: include/linux/atm* | 1716 | F: include/linux/atm* |
| 1717 | F: include/uapi/linux/atm* | 1717 | F: include/uapi/linux/atm* |
| 1718 | 1718 | ||
| 1719 | ATMEL AT91 / AT32 MCI DRIVER | 1719 | ATMEL AT91 / AT32 MCI DRIVER |
| 1720 | M: Ludovic Desroches <ludovic.desroches@atmel.com> | 1720 | M: Ludovic Desroches <ludovic.desroches@atmel.com> |
| 1721 | S: Maintained | 1721 | S: Maintained |
| 1722 | F: drivers/mmc/host/atmel-mci.c | 1722 | F: drivers/mmc/host/atmel-mci.c |
| 1723 | F: drivers/mmc/host/atmel-mci-regs.h | 1723 | F: drivers/mmc/host/atmel-mci-regs.h |
| 1724 | 1724 | ||
| 1725 | ATMEL AT91 / AT32 SERIAL DRIVER | 1725 | ATMEL AT91 / AT32 SERIAL DRIVER |
| 1726 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1726 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1727 | S: Supported | 1727 | S: Supported |
| 1728 | F: drivers/tty/serial/atmel_serial.c | 1728 | F: drivers/tty/serial/atmel_serial.c |
| 1729 | 1729 | ||
| 1730 | ATMEL Audio ALSA driver | 1730 | ATMEL Audio ALSA driver |
| 1731 | M: Bo Shen <voice.shen@atmel.com> | 1731 | M: Bo Shen <voice.shen@atmel.com> |
| 1732 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 1732 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 1733 | S: Supported | 1733 | S: Supported |
| 1734 | F: sound/soc/atmel | 1734 | F: sound/soc/atmel |
| 1735 | 1735 | ||
| 1736 | ATMEL DMA DRIVER | 1736 | ATMEL DMA DRIVER |
| 1737 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1737 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1738 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1738 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1739 | S: Supported | 1739 | S: Supported |
| 1740 | F: drivers/dma/at_hdmac.c | 1740 | F: drivers/dma/at_hdmac.c |
| 1741 | F: drivers/dma/at_hdmac_regs.h | 1741 | F: drivers/dma/at_hdmac_regs.h |
| 1742 | F: include/linux/platform_data/dma-atmel.h | 1742 | F: include/linux/platform_data/dma-atmel.h |
| 1743 | 1743 | ||
| 1744 | ATMEL XDMA DRIVER | 1744 | ATMEL XDMA DRIVER |
| 1745 | M: Ludovic Desroches <ludovic.desroches@atmel.com> | 1745 | M: Ludovic Desroches <ludovic.desroches@atmel.com> |
| 1746 | L: linux-arm-kernel@lists.infradead.org | 1746 | L: linux-arm-kernel@lists.infradead.org |
| 1747 | L: dmaengine@vger.kernel.org | 1747 | L: dmaengine@vger.kernel.org |
| 1748 | S: Supported | 1748 | S: Supported |
| 1749 | F: drivers/dma/at_xdmac.c | 1749 | F: drivers/dma/at_xdmac.c |
| 1750 | 1750 | ||
| 1751 | ATMEL I2C DRIVER | 1751 | ATMEL I2C DRIVER |
| 1752 | M: Ludovic Desroches <ludovic.desroches@atmel.com> | 1752 | M: Ludovic Desroches <ludovic.desroches@atmel.com> |
| 1753 | L: linux-i2c@vger.kernel.org | 1753 | L: linux-i2c@vger.kernel.org |
| 1754 | S: Supported | 1754 | S: Supported |
| 1755 | F: drivers/i2c/busses/i2c-at91.c | 1755 | F: drivers/i2c/busses/i2c-at91.c |
| 1756 | 1756 | ||
| 1757 | ATMEL ISI DRIVER | 1757 | ATMEL ISI DRIVER |
| 1758 | M: Josh Wu <josh.wu@atmel.com> | 1758 | M: Josh Wu <josh.wu@atmel.com> |
| 1759 | L: linux-media@vger.kernel.org | 1759 | L: linux-media@vger.kernel.org |
| 1760 | S: Supported | 1760 | S: Supported |
| 1761 | F: drivers/media/platform/soc_camera/atmel-isi.c | 1761 | F: drivers/media/platform/soc_camera/atmel-isi.c |
| 1762 | F: include/media/atmel-isi.h | 1762 | F: include/media/atmel-isi.h |
| 1763 | 1763 | ||
| 1764 | ATMEL LCDFB DRIVER | 1764 | ATMEL LCDFB DRIVER |
| 1765 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1765 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1766 | L: linux-fbdev@vger.kernel.org | 1766 | L: linux-fbdev@vger.kernel.org |
| 1767 | S: Maintained | 1767 | S: Maintained |
| 1768 | F: drivers/video/fbdev/atmel_lcdfb.c | 1768 | F: drivers/video/fbdev/atmel_lcdfb.c |
| 1769 | F: include/video/atmel_lcdc.h | 1769 | F: include/video/atmel_lcdc.h |
| 1770 | 1770 | ||
| 1771 | ATMEL MACB ETHERNET DRIVER | 1771 | ATMEL MACB ETHERNET DRIVER |
| 1772 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1772 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1773 | S: Supported | 1773 | S: Supported |
| 1774 | F: drivers/net/ethernet/cadence/ | 1774 | F: drivers/net/ethernet/cadence/ |
| 1775 | 1775 | ||
| 1776 | ATMEL NAND DRIVER | 1776 | ATMEL NAND DRIVER |
| 1777 | M: Josh Wu <josh.wu@atmel.com> | 1777 | M: Josh Wu <josh.wu@atmel.com> |
| 1778 | L: linux-mtd@lists.infradead.org | 1778 | L: linux-mtd@lists.infradead.org |
| 1779 | S: Supported | 1779 | S: Supported |
| 1780 | F: drivers/mtd/nand/atmel_nand* | 1780 | F: drivers/mtd/nand/atmel_nand* |
| 1781 | 1781 | ||
| 1782 | ATMEL SPI DRIVER | 1782 | ATMEL SPI DRIVER |
| 1783 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1783 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1784 | S: Supported | 1784 | S: Supported |
| 1785 | F: drivers/spi/spi-atmel.* | 1785 | F: drivers/spi/spi-atmel.* |
| 1786 | 1786 | ||
| 1787 | ATMEL SSC DRIVER | 1787 | ATMEL SSC DRIVER |
| 1788 | M: Bo Shen <voice.shen@atmel.com> | 1788 | M: Bo Shen <voice.shen@atmel.com> |
| 1789 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1789 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1790 | S: Supported | 1790 | S: Supported |
| 1791 | F: drivers/misc/atmel-ssc.c | 1791 | F: drivers/misc/atmel-ssc.c |
| 1792 | F: include/linux/atmel-ssc.h | 1792 | F: include/linux/atmel-ssc.h |
| 1793 | 1793 | ||
| 1794 | ATMEL Timer Counter (TC) AND CLOCKSOURCE DRIVERS | 1794 | ATMEL Timer Counter (TC) AND CLOCKSOURCE DRIVERS |
| 1795 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1795 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1796 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1796 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1797 | S: Supported | 1797 | S: Supported |
| 1798 | F: drivers/misc/atmel_tclib.c | 1798 | F: drivers/misc/atmel_tclib.c |
| 1799 | F: drivers/clocksource/tcb_clksrc.c | 1799 | F: drivers/clocksource/tcb_clksrc.c |
| 1800 | 1800 | ||
| 1801 | ATMEL USBA UDC DRIVER | 1801 | ATMEL USBA UDC DRIVER |
| 1802 | M: Nicolas Ferre <nicolas.ferre@atmel.com> | 1802 | M: Nicolas Ferre <nicolas.ferre@atmel.com> |
| 1803 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 1803 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 1804 | S: Supported | 1804 | S: Supported |
| 1805 | F: drivers/usb/gadget/udc/atmel_usba_udc.* | 1805 | F: drivers/usb/gadget/udc/atmel_usba_udc.* |
| 1806 | 1806 | ||
| 1807 | ATMEL WIRELESS DRIVER | 1807 | ATMEL WIRELESS DRIVER |
| 1808 | M: Simon Kelley <simon@thekelleys.org.uk> | 1808 | M: Simon Kelley <simon@thekelleys.org.uk> |
| 1809 | L: linux-wireless@vger.kernel.org | 1809 | L: linux-wireless@vger.kernel.org |
| 1810 | W: http://www.thekelleys.org.uk/atmel | 1810 | W: http://www.thekelleys.org.uk/atmel |
| 1811 | W: http://atmelwlandriver.sourceforge.net/ | 1811 | W: http://atmelwlandriver.sourceforge.net/ |
| 1812 | S: Maintained | 1812 | S: Maintained |
| 1813 | F: drivers/net/wireless/atmel* | 1813 | F: drivers/net/wireless/atmel* |
| 1814 | 1814 | ||
| 1815 | ATTO EXPRESSSAS SAS/SATA RAID SCSI DRIVER | 1815 | ATTO EXPRESSSAS SAS/SATA RAID SCSI DRIVER |
| 1816 | M: Bradley Grove <linuxdrivers@attotech.com> | 1816 | M: Bradley Grove <linuxdrivers@attotech.com> |
| 1817 | L: linux-scsi@vger.kernel.org | 1817 | L: linux-scsi@vger.kernel.org |
| 1818 | W: http://www.attotech.com | 1818 | W: http://www.attotech.com |
| 1819 | S: Supported | 1819 | S: Supported |
| 1820 | F: drivers/scsi/esas2r | 1820 | F: drivers/scsi/esas2r |
| 1821 | 1821 | ||
| 1822 | AUDIT SUBSYSTEM | 1822 | AUDIT SUBSYSTEM |
| 1823 | M: Paul Moore <paul@paul-moore.com> | 1823 | M: Paul Moore <paul@paul-moore.com> |
| 1824 | M: Eric Paris <eparis@redhat.com> | 1824 | M: Eric Paris <eparis@redhat.com> |
| 1825 | L: linux-audit@redhat.com (moderated for non-subscribers) | 1825 | L: linux-audit@redhat.com (moderated for non-subscribers) |
| 1826 | W: http://people.redhat.com/sgrubb/audit/ | 1826 | W: http://people.redhat.com/sgrubb/audit/ |
| 1827 | T: git git://git.infradead.org/users/pcmoore/audit | 1827 | T: git git://git.infradead.org/users/pcmoore/audit |
| 1828 | S: Maintained | 1828 | S: Maintained |
| 1829 | F: include/linux/audit.h | 1829 | F: include/linux/audit.h |
| 1830 | F: include/uapi/linux/audit.h | 1830 | F: include/uapi/linux/audit.h |
| 1831 | F: kernel/audit* | 1831 | F: kernel/audit* |
| 1832 | 1832 | ||
| 1833 | AUXILIARY DISPLAY DRIVERS | 1833 | AUXILIARY DISPLAY DRIVERS |
| 1834 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> | 1834 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> |
| 1835 | W: http://miguelojeda.es/auxdisplay.htm | 1835 | W: http://miguelojeda.es/auxdisplay.htm |
| 1836 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm | 1836 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm |
| 1837 | S: Maintained | 1837 | S: Maintained |
| 1838 | F: drivers/auxdisplay/ | 1838 | F: drivers/auxdisplay/ |
| 1839 | F: include/linux/cfag12864b.h | 1839 | F: include/linux/cfag12864b.h |
| 1840 | 1840 | ||
| 1841 | AVR32 ARCHITECTURE | 1841 | AVR32 ARCHITECTURE |
| 1842 | M: Haavard Skinnemoen <hskinnemoen@gmail.com> | 1842 | M: Haavard Skinnemoen <hskinnemoen@gmail.com> |
| 1843 | M: Hans-Christian Egtvedt <egtvedt@samfundet.no> | 1843 | M: Hans-Christian Egtvedt <egtvedt@samfundet.no> |
| 1844 | W: http://www.atmel.com/products/AVR32/ | 1844 | W: http://www.atmel.com/products/AVR32/ |
| 1845 | W: http://mirror.egtvedt.no/avr32linux.org/ | 1845 | W: http://mirror.egtvedt.no/avr32linux.org/ |
| 1846 | W: http://avrfreaks.net/ | 1846 | W: http://avrfreaks.net/ |
| 1847 | S: Maintained | 1847 | S: Maintained |
| 1848 | F: arch/avr32/ | 1848 | F: arch/avr32/ |
| 1849 | 1849 | ||
| 1850 | AVR32/AT32AP MACHINE SUPPORT | 1850 | AVR32/AT32AP MACHINE SUPPORT |
| 1851 | M: Haavard Skinnemoen <hskinnemoen@gmail.com> | 1851 | M: Haavard Skinnemoen <hskinnemoen@gmail.com> |
| 1852 | M: Hans-Christian Egtvedt <egtvedt@samfundet.no> | 1852 | M: Hans-Christian Egtvedt <egtvedt@samfundet.no> |
| 1853 | S: Maintained | 1853 | S: Maintained |
| 1854 | F: arch/avr32/mach-at32ap/ | 1854 | F: arch/avr32/mach-at32ap/ |
| 1855 | 1855 | ||
| 1856 | AX.25 NETWORK LAYER | 1856 | AX.25 NETWORK LAYER |
| 1857 | M: Ralf Baechle <ralf@linux-mips.org> | 1857 | M: Ralf Baechle <ralf@linux-mips.org> |
| 1858 | L: linux-hams@vger.kernel.org | 1858 | L: linux-hams@vger.kernel.org |
| 1859 | W: http://www.linux-ax25.org/ | 1859 | W: http://www.linux-ax25.org/ |
| 1860 | S: Maintained | 1860 | S: Maintained |
| 1861 | F: include/uapi/linux/ax25.h | 1861 | F: include/uapi/linux/ax25.h |
| 1862 | F: include/net/ax25.h | 1862 | F: include/net/ax25.h |
| 1863 | F: net/ax25/ | 1863 | F: net/ax25/ |
| 1864 | 1864 | ||
| 1865 | AZ6007 DVB DRIVER | 1865 | AZ6007 DVB DRIVER |
| 1866 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 1866 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 1867 | L: linux-media@vger.kernel.org | 1867 | L: linux-media@vger.kernel.org |
| 1868 | W: http://linuxtv.org | 1868 | W: http://linuxtv.org |
| 1869 | T: git git://linuxtv.org/media_tree.git | 1869 | T: git git://linuxtv.org/media_tree.git |
| 1870 | S: Maintained | 1870 | S: Maintained |
| 1871 | F: drivers/media/usb/dvb-usb-v2/az6007.c | 1871 | F: drivers/media/usb/dvb-usb-v2/az6007.c |
| 1872 | 1872 | ||
| 1873 | AZTECH FM RADIO RECEIVER DRIVER | 1873 | AZTECH FM RADIO RECEIVER DRIVER |
| 1874 | M: Hans Verkuil <hverkuil@xs4all.nl> | 1874 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 1875 | L: linux-media@vger.kernel.org | 1875 | L: linux-media@vger.kernel.org |
| 1876 | T: git git://linuxtv.org/media_tree.git | 1876 | T: git git://linuxtv.org/media_tree.git |
| 1877 | W: http://linuxtv.org | 1877 | W: http://linuxtv.org |
| 1878 | S: Maintained | 1878 | S: Maintained |
| 1879 | F: drivers/media/radio/radio-aztech* | 1879 | F: drivers/media/radio/radio-aztech* |
| 1880 | 1880 | ||
| 1881 | B43 WIRELESS DRIVER | 1881 | B43 WIRELESS DRIVER |
| 1882 | M: Stefano Brivio <stefano.brivio@polimi.it> | 1882 | M: Stefano Brivio <stefano.brivio@polimi.it> |
| 1883 | L: linux-wireless@vger.kernel.org | 1883 | L: linux-wireless@vger.kernel.org |
| 1884 | L: b43-dev@lists.infradead.org | 1884 | L: b43-dev@lists.infradead.org |
| 1885 | W: http://wireless.kernel.org/en/users/Drivers/b43 | 1885 | W: http://wireless.kernel.org/en/users/Drivers/b43 |
| 1886 | S: Maintained | 1886 | S: Maintained |
| 1887 | F: drivers/net/wireless/b43/ | 1887 | F: drivers/net/wireless/b43/ |
| 1888 | 1888 | ||
| 1889 | B43LEGACY WIRELESS DRIVER | 1889 | B43LEGACY WIRELESS DRIVER |
| 1890 | M: Larry Finger <Larry.Finger@lwfinger.net> | 1890 | M: Larry Finger <Larry.Finger@lwfinger.net> |
| 1891 | M: Stefano Brivio <stefano.brivio@polimi.it> | 1891 | M: Stefano Brivio <stefano.brivio@polimi.it> |
| 1892 | L: linux-wireless@vger.kernel.org | 1892 | L: linux-wireless@vger.kernel.org |
| 1893 | L: b43-dev@lists.infradead.org | 1893 | L: b43-dev@lists.infradead.org |
| 1894 | W: http://wireless.kernel.org/en/users/Drivers/b43 | 1894 | W: http://wireless.kernel.org/en/users/Drivers/b43 |
| 1895 | S: Maintained | 1895 | S: Maintained |
| 1896 | F: drivers/net/wireless/b43legacy/ | 1896 | F: drivers/net/wireless/b43legacy/ |
| 1897 | 1897 | ||
| 1898 | BACKLIGHT CLASS/SUBSYSTEM | 1898 | BACKLIGHT CLASS/SUBSYSTEM |
| 1899 | M: Jingoo Han <jg1.han@samsung.com> | 1899 | M: Jingoo Han <jg1.han@samsung.com> |
| 1900 | M: Lee Jones <lee.jones@linaro.org> | 1900 | M: Lee Jones <lee.jones@linaro.org> |
| 1901 | S: Maintained | 1901 | S: Maintained |
| 1902 | F: drivers/video/backlight/ | 1902 | F: drivers/video/backlight/ |
| 1903 | F: include/linux/backlight.h | 1903 | F: include/linux/backlight.h |
| 1904 | 1904 | ||
| 1905 | BATMAN ADVANCED | 1905 | BATMAN ADVANCED |
| 1906 | M: Marek Lindner <mareklindner@neomailbox.ch> | 1906 | M: Marek Lindner <mareklindner@neomailbox.ch> |
| 1907 | M: Simon Wunderlich <sw@simonwunderlich.de> | 1907 | M: Simon Wunderlich <sw@simonwunderlich.de> |
| 1908 | M: Antonio Quartulli <antonio@meshcoding.com> | 1908 | M: Antonio Quartulli <antonio@meshcoding.com> |
| 1909 | L: b.a.t.m.a.n@lists.open-mesh.org | 1909 | L: b.a.t.m.a.n@lists.open-mesh.org |
| 1910 | W: http://www.open-mesh.org/ | 1910 | W: http://www.open-mesh.org/ |
| 1911 | S: Maintained | 1911 | S: Maintained |
| 1912 | F: net/batman-adv/ | 1912 | F: net/batman-adv/ |
| 1913 | 1913 | ||
| 1914 | BAYCOM/HDLCDRV DRIVERS FOR AX.25 | 1914 | BAYCOM/HDLCDRV DRIVERS FOR AX.25 |
| 1915 | M: Thomas Sailer <t.sailer@alumni.ethz.ch> | 1915 | M: Thomas Sailer <t.sailer@alumni.ethz.ch> |
| 1916 | L: linux-hams@vger.kernel.org | 1916 | L: linux-hams@vger.kernel.org |
| 1917 | W: http://www.baycom.org/~tom/ham/ham.html | 1917 | W: http://www.baycom.org/~tom/ham/ham.html |
| 1918 | S: Maintained | 1918 | S: Maintained |
| 1919 | F: drivers/net/hamradio/baycom* | 1919 | F: drivers/net/hamradio/baycom* |
| 1920 | 1920 | ||
| 1921 | BCACHE (BLOCK LAYER CACHE) | 1921 | BCACHE (BLOCK LAYER CACHE) |
| 1922 | M: Kent Overstreet <kmo@daterainc.com> | 1922 | M: Kent Overstreet <kmo@daterainc.com> |
| 1923 | L: linux-bcache@vger.kernel.org | 1923 | L: linux-bcache@vger.kernel.org |
| 1924 | W: http://bcache.evilpiepirate.org | 1924 | W: http://bcache.evilpiepirate.org |
| 1925 | S: Maintained: | 1925 | S: Maintained: |
| 1926 | F: drivers/md/bcache/ | 1926 | F: drivers/md/bcache/ |
| 1927 | 1927 | ||
| 1928 | BEFS FILE SYSTEM | 1928 | BEFS FILE SYSTEM |
| 1929 | S: Orphan | 1929 | S: Orphan |
| 1930 | F: Documentation/filesystems/befs.txt | 1930 | F: Documentation/filesystems/befs.txt |
| 1931 | F: fs/befs/ | 1931 | F: fs/befs/ |
| 1932 | 1932 | ||
| 1933 | BECKHOFF CX5020 ETHERCAT MASTER DRIVER | 1933 | BECKHOFF CX5020 ETHERCAT MASTER DRIVER |
| 1934 | M: Dariusz Marcinkiewicz <reksio@newterm.pl> | 1934 | M: Dariusz Marcinkiewicz <reksio@newterm.pl> |
| 1935 | L: netdev@vger.kernel.org | 1935 | L: netdev@vger.kernel.org |
| 1936 | S: Maintained | 1936 | S: Maintained |
| 1937 | F: drivers/net/ethernet/ec_bhf.c | 1937 | F: drivers/net/ethernet/ec_bhf.c |
| 1938 | 1938 | ||
| 1939 | BFS FILE SYSTEM | 1939 | BFS FILE SYSTEM |
| 1940 | M: "Tigran A. Aivazian" <tigran@aivazian.fsnet.co.uk> | 1940 | M: "Tigran A. Aivazian" <tigran@aivazian.fsnet.co.uk> |
| 1941 | S: Maintained | 1941 | S: Maintained |
| 1942 | F: Documentation/filesystems/bfs.txt | 1942 | F: Documentation/filesystems/bfs.txt |
| 1943 | F: fs/bfs/ | 1943 | F: fs/bfs/ |
| 1944 | F: include/uapi/linux/bfs_fs.h | 1944 | F: include/uapi/linux/bfs_fs.h |
| 1945 | 1945 | ||
| 1946 | BLACKFIN ARCHITECTURE | 1946 | BLACKFIN ARCHITECTURE |
| 1947 | M: Steven Miao <realmz6@gmail.com> | 1947 | M: Steven Miao <realmz6@gmail.com> |
| 1948 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1948 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1949 | T: git git://git.code.sf.net/p/adi-linux/code | 1949 | T: git git://git.code.sf.net/p/adi-linux/code |
| 1950 | W: http://blackfin.uclinux.org | 1950 | W: http://blackfin.uclinux.org |
| 1951 | S: Supported | 1951 | S: Supported |
| 1952 | F: arch/blackfin/ | 1952 | F: arch/blackfin/ |
| 1953 | 1953 | ||
| 1954 | BLACKFIN EMAC DRIVER | 1954 | BLACKFIN EMAC DRIVER |
| 1955 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1955 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1956 | W: http://blackfin.uclinux.org | 1956 | W: http://blackfin.uclinux.org |
| 1957 | S: Supported | 1957 | S: Supported |
| 1958 | F: drivers/net/ethernet/adi/ | 1958 | F: drivers/net/ethernet/adi/ |
| 1959 | 1959 | ||
| 1960 | BLACKFIN RTC DRIVER | 1960 | BLACKFIN RTC DRIVER |
| 1961 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1961 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1962 | W: http://blackfin.uclinux.org | 1962 | W: http://blackfin.uclinux.org |
| 1963 | S: Supported | 1963 | S: Supported |
| 1964 | F: drivers/rtc/rtc-bfin.c | 1964 | F: drivers/rtc/rtc-bfin.c |
| 1965 | 1965 | ||
| 1966 | BLACKFIN SDH DRIVER | 1966 | BLACKFIN SDH DRIVER |
| 1967 | M: Sonic Zhang <sonic.zhang@analog.com> | 1967 | M: Sonic Zhang <sonic.zhang@analog.com> |
| 1968 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1968 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1969 | W: http://blackfin.uclinux.org | 1969 | W: http://blackfin.uclinux.org |
| 1970 | S: Supported | 1970 | S: Supported |
| 1971 | F: drivers/mmc/host/bfin_sdh.c | 1971 | F: drivers/mmc/host/bfin_sdh.c |
| 1972 | 1972 | ||
| 1973 | BLACKFIN SERIAL DRIVER | 1973 | BLACKFIN SERIAL DRIVER |
| 1974 | M: Sonic Zhang <sonic.zhang@analog.com> | 1974 | M: Sonic Zhang <sonic.zhang@analog.com> |
| 1975 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1975 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1976 | W: http://blackfin.uclinux.org | 1976 | W: http://blackfin.uclinux.org |
| 1977 | S: Supported | 1977 | S: Supported |
| 1978 | F: drivers/tty/serial/bfin_uart.c | 1978 | F: drivers/tty/serial/bfin_uart.c |
| 1979 | 1979 | ||
| 1980 | BLACKFIN WATCHDOG DRIVER | 1980 | BLACKFIN WATCHDOG DRIVER |
| 1981 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1981 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1982 | W: http://blackfin.uclinux.org | 1982 | W: http://blackfin.uclinux.org |
| 1983 | S: Supported | 1983 | S: Supported |
| 1984 | F: drivers/watchdog/bfin_wdt.c | 1984 | F: drivers/watchdog/bfin_wdt.c |
| 1985 | 1985 | ||
| 1986 | BLACKFIN I2C TWI DRIVER | 1986 | BLACKFIN I2C TWI DRIVER |
| 1987 | M: Sonic Zhang <sonic.zhang@analog.com> | 1987 | M: Sonic Zhang <sonic.zhang@analog.com> |
| 1988 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1988 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1989 | W: http://blackfin.uclinux.org/ | 1989 | W: http://blackfin.uclinux.org/ |
| 1990 | S: Supported | 1990 | S: Supported |
| 1991 | F: drivers/i2c/busses/i2c-bfin-twi.c | 1991 | F: drivers/i2c/busses/i2c-bfin-twi.c |
| 1992 | 1992 | ||
| 1993 | BLACKFIN MEDIA DRIVER | 1993 | BLACKFIN MEDIA DRIVER |
| 1994 | M: Scott Jiang <scott.jiang.linux@gmail.com> | 1994 | M: Scott Jiang <scott.jiang.linux@gmail.com> |
| 1995 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) | 1995 | L: adi-buildroot-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 1996 | W: http://blackfin.uclinux.org/ | 1996 | W: http://blackfin.uclinux.org/ |
| 1997 | S: Supported | 1997 | S: Supported |
| 1998 | F: drivers/media/platform/blackfin/ | 1998 | F: drivers/media/platform/blackfin/ |
| 1999 | F: drivers/media/i2c/adv7183* | 1999 | F: drivers/media/i2c/adv7183* |
| 2000 | F: drivers/media/i2c/vs6624* | 2000 | F: drivers/media/i2c/vs6624* |
| 2001 | 2001 | ||
| 2002 | BLINKM RGB LED DRIVER | 2002 | BLINKM RGB LED DRIVER |
| 2003 | M: Jan-Simon Moeller <jansimon.moeller@gmx.de> | 2003 | M: Jan-Simon Moeller <jansimon.moeller@gmx.de> |
| 2004 | S: Maintained | 2004 | S: Maintained |
| 2005 | F: drivers/leds/leds-blinkm.c | 2005 | F: drivers/leds/leds-blinkm.c |
| 2006 | 2006 | ||
| 2007 | BLOCK LAYER | 2007 | BLOCK LAYER |
| 2008 | M: Jens Axboe <axboe@kernel.dk> | 2008 | M: Jens Axboe <axboe@kernel.dk> |
| 2009 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git | 2009 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git |
| 2010 | S: Maintained | 2010 | S: Maintained |
| 2011 | F: block/ | 2011 | F: block/ |
| 2012 | 2012 | ||
| 2013 | BLOCK2MTD DRIVER | 2013 | BLOCK2MTD DRIVER |
| 2014 | M: Joern Engel <joern@lazybastard.org> | 2014 | M: Joern Engel <joern@lazybastard.org> |
| 2015 | L: linux-mtd@lists.infradead.org | 2015 | L: linux-mtd@lists.infradead.org |
| 2016 | S: Maintained | 2016 | S: Maintained |
| 2017 | F: drivers/mtd/devices/block2mtd.c | 2017 | F: drivers/mtd/devices/block2mtd.c |
| 2018 | 2018 | ||
| 2019 | BLUETOOTH DRIVERS | 2019 | BLUETOOTH DRIVERS |
| 2020 | M: Marcel Holtmann <marcel@holtmann.org> | 2020 | M: Marcel Holtmann <marcel@holtmann.org> |
| 2021 | M: Gustavo Padovan <gustavo@padovan.org> | 2021 | M: Gustavo Padovan <gustavo@padovan.org> |
| 2022 | M: Johan Hedberg <johan.hedberg@gmail.com> | 2022 | M: Johan Hedberg <johan.hedberg@gmail.com> |
| 2023 | L: linux-bluetooth@vger.kernel.org | 2023 | L: linux-bluetooth@vger.kernel.org |
| 2024 | W: http://www.bluez.org/ | 2024 | W: http://www.bluez.org/ |
| 2025 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth.git | 2025 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth.git |
| 2026 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git | 2026 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git |
| 2027 | S: Maintained | 2027 | S: Maintained |
| 2028 | F: drivers/bluetooth/ | 2028 | F: drivers/bluetooth/ |
| 2029 | 2029 | ||
| 2030 | BLUETOOTH SUBSYSTEM | 2030 | BLUETOOTH SUBSYSTEM |
| 2031 | M: Marcel Holtmann <marcel@holtmann.org> | 2031 | M: Marcel Holtmann <marcel@holtmann.org> |
| 2032 | M: Gustavo Padovan <gustavo@padovan.org> | 2032 | M: Gustavo Padovan <gustavo@padovan.org> |
| 2033 | M: Johan Hedberg <johan.hedberg@gmail.com> | 2033 | M: Johan Hedberg <johan.hedberg@gmail.com> |
| 2034 | L: linux-bluetooth@vger.kernel.org | 2034 | L: linux-bluetooth@vger.kernel.org |
| 2035 | W: http://www.bluez.org/ | 2035 | W: http://www.bluez.org/ |
| 2036 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth.git | 2036 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth.git |
| 2037 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git | 2037 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git |
| 2038 | S: Maintained | 2038 | S: Maintained |
| 2039 | F: net/bluetooth/ | 2039 | F: net/bluetooth/ |
| 2040 | F: include/net/bluetooth/ | 2040 | F: include/net/bluetooth/ |
| 2041 | 2041 | ||
| 2042 | BONDING DRIVER | 2042 | BONDING DRIVER |
| 2043 | M: Jay Vosburgh <j.vosburgh@gmail.com> | 2043 | M: Jay Vosburgh <j.vosburgh@gmail.com> |
| 2044 | M: Veaceslav Falico <vfalico@gmail.com> | 2044 | M: Veaceslav Falico <vfalico@gmail.com> |
| 2045 | M: Andy Gospodarek <andy@greyhouse.net> | 2045 | M: Andy Gospodarek <andy@greyhouse.net> |
| 2046 | L: netdev@vger.kernel.org | 2046 | L: netdev@vger.kernel.org |
| 2047 | W: http://sourceforge.net/projects/bonding/ | 2047 | W: http://sourceforge.net/projects/bonding/ |
| 2048 | S: Supported | 2048 | S: Supported |
| 2049 | F: drivers/net/bonding/ | 2049 | F: drivers/net/bonding/ |
| 2050 | F: include/uapi/linux/if_bonding.h | 2050 | F: include/uapi/linux/if_bonding.h |
| 2051 | 2051 | ||
| 2052 | BPF (Safe dynamic programs and tools) | 2052 | BPF (Safe dynamic programs and tools) |
| 2053 | M: Alexei Starovoitov <ast@kernel.org> | 2053 | M: Alexei Starovoitov <ast@kernel.org> |
| 2054 | L: netdev@vger.kernel.org | 2054 | L: netdev@vger.kernel.org |
| 2055 | L: linux-kernel@vger.kernel.org | 2055 | L: linux-kernel@vger.kernel.org |
| 2056 | S: Supported | 2056 | S: Supported |
| 2057 | F: kernel/bpf/ | 2057 | F: kernel/bpf/ |
| 2058 | 2058 | ||
| 2059 | BROADCOM B44 10/100 ETHERNET DRIVER | 2059 | BROADCOM B44 10/100 ETHERNET DRIVER |
| 2060 | M: Gary Zambrano <zambrano@broadcom.com> | 2060 | M: Gary Zambrano <zambrano@broadcom.com> |
| 2061 | L: netdev@vger.kernel.org | 2061 | L: netdev@vger.kernel.org |
| 2062 | S: Supported | 2062 | S: Supported |
| 2063 | F: drivers/net/ethernet/broadcom/b44.* | 2063 | F: drivers/net/ethernet/broadcom/b44.* |
| 2064 | 2064 | ||
| 2065 | BROADCOM GENET ETHERNET DRIVER | 2065 | BROADCOM GENET ETHERNET DRIVER |
| 2066 | M: Florian Fainelli <f.fainelli@gmail.com> | 2066 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 2067 | L: netdev@vger.kernel.org | 2067 | L: netdev@vger.kernel.org |
| 2068 | S: Supported | 2068 | S: Supported |
| 2069 | F: drivers/net/ethernet/broadcom/genet/ | 2069 | F: drivers/net/ethernet/broadcom/genet/ |
| 2070 | 2070 | ||
| 2071 | BROADCOM BNX2 GIGABIT ETHERNET DRIVER | 2071 | BROADCOM BNX2 GIGABIT ETHERNET DRIVER |
| 2072 | M: Sony Chacko <sony.chacko@qlogic.com> | 2072 | M: Sony Chacko <sony.chacko@qlogic.com> |
| 2073 | M: Dept-HSGLinuxNICDev@qlogic.com | 2073 | M: Dept-HSGLinuxNICDev@qlogic.com |
| 2074 | L: netdev@vger.kernel.org | 2074 | L: netdev@vger.kernel.org |
| 2075 | S: Supported | 2075 | S: Supported |
| 2076 | F: drivers/net/ethernet/broadcom/bnx2.* | 2076 | F: drivers/net/ethernet/broadcom/bnx2.* |
| 2077 | F: drivers/net/ethernet/broadcom/bnx2_* | 2077 | F: drivers/net/ethernet/broadcom/bnx2_* |
| 2078 | 2078 | ||
| 2079 | BROADCOM BNX2X 10 GIGABIT ETHERNET DRIVER | 2079 | BROADCOM BNX2X 10 GIGABIT ETHERNET DRIVER |
| 2080 | M: Ariel Elior <ariel.elior@qlogic.com> | 2080 | M: Ariel Elior <ariel.elior@qlogic.com> |
| 2081 | L: netdev@vger.kernel.org | 2081 | L: netdev@vger.kernel.org |
| 2082 | S: Supported | 2082 | S: Supported |
| 2083 | F: drivers/net/ethernet/broadcom/bnx2x/ | 2083 | F: drivers/net/ethernet/broadcom/bnx2x/ |
| 2084 | 2084 | ||
| 2085 | BROADCOM BCM281XX/BCM11XXX/BCM216XX ARM ARCHITECTURE | 2085 | BROADCOM BCM281XX/BCM11XXX/BCM216XX ARM ARCHITECTURE |
| 2086 | M: Christian Daudt <bcm@fixthebug.org> | 2086 | M: Christian Daudt <bcm@fixthebug.org> |
| 2087 | M: Matt Porter <mporter@linaro.org> | 2087 | M: Matt Porter <mporter@linaro.org> |
| 2088 | M: Florian Fainelli <f.fainelli@gmail.com> | 2088 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 2089 | L: bcm-kernel-feedback-list@broadcom.com | 2089 | L: bcm-kernel-feedback-list@broadcom.com |
| 2090 | T: git git://github.com/broadcom/mach-bcm | 2090 | T: git git://github.com/broadcom/mach-bcm |
| 2091 | S: Maintained | 2091 | S: Maintained |
| 2092 | F: arch/arm/mach-bcm/ | 2092 | F: arch/arm/mach-bcm/ |
| 2093 | F: arch/arm/boot/dts/bcm113* | 2093 | F: arch/arm/boot/dts/bcm113* |
| 2094 | F: arch/arm/boot/dts/bcm216* | 2094 | F: arch/arm/boot/dts/bcm216* |
| 2095 | F: arch/arm/boot/dts/bcm281* | 2095 | F: arch/arm/boot/dts/bcm281* |
| 2096 | F: arch/arm/configs/bcm_defconfig | 2096 | F: arch/arm/configs/bcm_defconfig |
| 2097 | F: drivers/mmc/host/sdhci-bcm-kona.c | 2097 | F: drivers/mmc/host/sdhci-bcm-kona.c |
| 2098 | F: drivers/clocksource/bcm_kona_timer.c | 2098 | F: drivers/clocksource/bcm_kona_timer.c |
| 2099 | 2099 | ||
| 2100 | BROADCOM BCM2835 ARM ARCHITECTURE | 2100 | BROADCOM BCM2835 ARM ARCHITECTURE |
| 2101 | M: Stephen Warren <swarren@wwwdotorg.org> | 2101 | M: Stephen Warren <swarren@wwwdotorg.org> |
| 2102 | M: Lee Jones <lee@kernel.org> | 2102 | M: Lee Jones <lee@kernel.org> |
| 2103 | L: linux-rpi-kernel@lists.infradead.org (moderated for non-subscribers) | 2103 | L: linux-rpi-kernel@lists.infradead.org (moderated for non-subscribers) |
| 2104 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rpi/linux-rpi.git | 2104 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rpi/linux-rpi.git |
| 2105 | S: Maintained | 2105 | S: Maintained |
| 2106 | N: bcm2835 | 2106 | N: bcm2835 |
| 2107 | 2107 | ||
| 2108 | BROADCOM BCM33XX MIPS ARCHITECTURE | 2108 | BROADCOM BCM33XX MIPS ARCHITECTURE |
| 2109 | M: Kevin Cernekee <cernekee@gmail.com> | 2109 | M: Kevin Cernekee <cernekee@gmail.com> |
| 2110 | L: linux-mips@linux-mips.org | 2110 | L: linux-mips@linux-mips.org |
| 2111 | S: Maintained | 2111 | S: Maintained |
| 2112 | F: arch/mips/bcm3384/* | 2112 | F: arch/mips/bcm3384/* |
| 2113 | F: arch/mips/include/asm/mach-bcm3384/* | 2113 | F: arch/mips/include/asm/mach-bcm3384/* |
| 2114 | F: arch/mips/kernel/*bmips* | 2114 | F: arch/mips/kernel/*bmips* |
| 2115 | 2115 | ||
| 2116 | BROADCOM BCM5301X ARM ARCHITECTURE | 2116 | BROADCOM BCM5301X ARM ARCHITECTURE |
| 2117 | M: Hauke Mehrtens <hauke@hauke-m.de> | 2117 | M: Hauke Mehrtens <hauke@hauke-m.de> |
| 2118 | L: linux-arm-kernel@lists.infradead.org | 2118 | L: linux-arm-kernel@lists.infradead.org |
| 2119 | S: Maintained | 2119 | S: Maintained |
| 2120 | F: arch/arm/mach-bcm/bcm_5301x.c | 2120 | F: arch/arm/mach-bcm/bcm_5301x.c |
| 2121 | F: arch/arm/boot/dts/bcm5301x.dtsi | 2121 | F: arch/arm/boot/dts/bcm5301x.dtsi |
| 2122 | F: arch/arm/boot/dts/bcm470* | 2122 | F: arch/arm/boot/dts/bcm470* |
| 2123 | 2123 | ||
| 2124 | BROADCOM BCM63XX ARM ARCHITECTURE | 2124 | BROADCOM BCM63XX ARM ARCHITECTURE |
| 2125 | M: Florian Fainelli <f.fainelli@gmail.com> | 2125 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 2126 | L: linux-arm-kernel@lists.infradead.org | 2126 | L: linux-arm-kernel@lists.infradead.org |
| 2127 | T: git git://git.github.com/brcm/linux.git | 2127 | T: git git://git.github.com/brcm/linux.git |
| 2128 | S: Maintained | 2128 | S: Maintained |
| 2129 | F: arch/arm/mach-bcm/bcm63xx.c | 2129 | F: arch/arm/mach-bcm/bcm63xx.c |
| 2130 | F: arch/arm/include/debug/bcm63xx.S | 2130 | F: arch/arm/include/debug/bcm63xx.S |
| 2131 | 2131 | ||
| 2132 | BROADCOM BCM63XX/BCM33XX UDC DRIVER | 2132 | BROADCOM BCM63XX/BCM33XX UDC DRIVER |
| 2133 | M: Kevin Cernekee <cernekee@gmail.com> | 2133 | M: Kevin Cernekee <cernekee@gmail.com> |
| 2134 | L: linux-usb@vger.kernel.org | 2134 | L: linux-usb@vger.kernel.org |
| 2135 | S: Maintained | 2135 | S: Maintained |
| 2136 | F: drivers/usb/gadget/udc/bcm63xx_udc.* | 2136 | F: drivers/usb/gadget/udc/bcm63xx_udc.* |
| 2137 | 2137 | ||
| 2138 | BROADCOM BCM7XXX ARM ARCHITECTURE | 2138 | BROADCOM BCM7XXX ARM ARCHITECTURE |
| 2139 | M: Marc Carino <marc.ceeeee@gmail.com> | 2139 | M: Marc Carino <marc.ceeeee@gmail.com> |
| 2140 | M: Brian Norris <computersforpeace@gmail.com> | 2140 | M: Brian Norris <computersforpeace@gmail.com> |
| 2141 | M: Gregory Fong <gregory.0xf0@gmail.com> | 2141 | M: Gregory Fong <gregory.0xf0@gmail.com> |
| 2142 | M: Florian Fainelli <f.fainelli@gmail.com> | 2142 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 2143 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 2143 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 2144 | S: Maintained | 2144 | S: Maintained |
| 2145 | F: arch/arm/mach-bcm/*brcmstb* | 2145 | F: arch/arm/mach-bcm/*brcmstb* |
| 2146 | F: arch/arm/boot/dts/bcm7*.dts* | 2146 | F: arch/arm/boot/dts/bcm7*.dts* |
| 2147 | F: drivers/bus/brcmstb_gisb.c | 2147 | F: drivers/bus/brcmstb_gisb.c |
| 2148 | 2148 | ||
| 2149 | BROADCOM BMIPS MIPS ARCHITECTURE | 2149 | BROADCOM BMIPS MIPS ARCHITECTURE |
| 2150 | M: Kevin Cernekee <cernekee@gmail.com> | 2150 | M: Kevin Cernekee <cernekee@gmail.com> |
| 2151 | M: Florian Fainelli <f.fainelli@gmail.com> | 2151 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 2152 | L: linux-mips@linux-mips.org | 2152 | L: linux-mips@linux-mips.org |
| 2153 | S: Maintained | 2153 | S: Maintained |
| 2154 | F: arch/mips/bmips/* | 2154 | F: arch/mips/bmips/* |
| 2155 | F: arch/mips/include/asm/mach-bmips/* | 2155 | F: arch/mips/include/asm/mach-bmips/* |
| 2156 | F: arch/mips/kernel/*bmips* | 2156 | F: arch/mips/kernel/*bmips* |
| 2157 | F: arch/mips/boot/dts/bcm*.dts* | 2157 | F: arch/mips/boot/dts/bcm*.dts* |
| 2158 | F: drivers/irqchip/irq-bcm7* | 2158 | F: drivers/irqchip/irq-bcm7* |
| 2159 | F: drivers/irqchip/irq-brcmstb* | 2159 | F: drivers/irqchip/irq-brcmstb* |
| 2160 | 2160 | ||
| 2161 | BROADCOM TG3 GIGABIT ETHERNET DRIVER | 2161 | BROADCOM TG3 GIGABIT ETHERNET DRIVER |
| 2162 | M: Prashant Sreedharan <prashant@broadcom.com> | 2162 | M: Prashant Sreedharan <prashant@broadcom.com> |
| 2163 | M: Michael Chan <mchan@broadcom.com> | 2163 | M: Michael Chan <mchan@broadcom.com> |
| 2164 | L: netdev@vger.kernel.org | 2164 | L: netdev@vger.kernel.org |
| 2165 | S: Supported | 2165 | S: Supported |
| 2166 | F: drivers/net/ethernet/broadcom/tg3.* | 2166 | F: drivers/net/ethernet/broadcom/tg3.* |
| 2167 | 2167 | ||
| 2168 | BROADCOM BRCM80211 IEEE802.11n WIRELESS DRIVER | 2168 | BROADCOM BRCM80211 IEEE802.11n WIRELESS DRIVER |
| 2169 | M: Brett Rudley <brudley@broadcom.com> | 2169 | M: Brett Rudley <brudley@broadcom.com> |
| 2170 | M: Arend van Spriel <arend@broadcom.com> | 2170 | M: Arend van Spriel <arend@broadcom.com> |
| 2171 | M: Franky (Zhenhui) Lin <frankyl@broadcom.com> | 2171 | M: Franky (Zhenhui) Lin <frankyl@broadcom.com> |
| 2172 | M: Hante Meuleman <meuleman@broadcom.com> | 2172 | M: Hante Meuleman <meuleman@broadcom.com> |
| 2173 | L: linux-wireless@vger.kernel.org | 2173 | L: linux-wireless@vger.kernel.org |
| 2174 | L: brcm80211-dev-list@broadcom.com | 2174 | L: brcm80211-dev-list@broadcom.com |
| 2175 | S: Supported | 2175 | S: Supported |
| 2176 | F: drivers/net/wireless/brcm80211/ | 2176 | F: drivers/net/wireless/brcm80211/ |
| 2177 | 2177 | ||
| 2178 | BROADCOM BNX2FC 10 GIGABIT FCOE DRIVER | 2178 | BROADCOM BNX2FC 10 GIGABIT FCOE DRIVER |
| 2179 | M: QLogic-Storage-Upstream@qlogic.com | 2179 | M: QLogic-Storage-Upstream@qlogic.com |
| 2180 | L: linux-scsi@vger.kernel.org | 2180 | L: linux-scsi@vger.kernel.org |
| 2181 | S: Supported | 2181 | S: Supported |
| 2182 | F: drivers/scsi/bnx2fc/ | 2182 | F: drivers/scsi/bnx2fc/ |
| 2183 | 2183 | ||
| 2184 | BROADCOM BNX2I 1/10 GIGABIT iSCSI DRIVER | 2184 | BROADCOM BNX2I 1/10 GIGABIT iSCSI DRIVER |
| 2185 | M: QLogic-Storage-Upstream@qlogic.com | 2185 | M: QLogic-Storage-Upstream@qlogic.com |
| 2186 | L: linux-scsi@vger.kernel.org | 2186 | L: linux-scsi@vger.kernel.org |
| 2187 | S: Supported | 2187 | S: Supported |
| 2188 | F: drivers/scsi/bnx2i/ | 2188 | F: drivers/scsi/bnx2i/ |
| 2189 | 2189 | ||
| 2190 | BROADCOM CYGNUS/IPROC ARM ARCHITECTURE | 2190 | BROADCOM CYGNUS/IPROC ARM ARCHITECTURE |
| 2191 | M: Ray Jui <rjui@broadcom.com> | 2191 | M: Ray Jui <rjui@broadcom.com> |
| 2192 | M: Scott Branden <sbranden@broadcom.com> | 2192 | M: Scott Branden <sbranden@broadcom.com> |
| 2193 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 2193 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 2194 | L: bcm-kernel-feedback-list@broadcom.com | 2194 | L: bcm-kernel-feedback-list@broadcom.com |
| 2195 | T: git git://git.github.com/brcm/linux.git | 2195 | T: git git://git.github.com/brcm/linux.git |
| 2196 | S: Maintained | 2196 | S: Maintained |
| 2197 | N: iproc | 2197 | N: iproc |
| 2198 | N: cygnus | 2198 | N: cygnus |
| 2199 | N: bcm9113* | 2199 | N: bcm9113* |
| 2200 | N: bcm9583* | 2200 | N: bcm9583* |
| 2201 | N: bcm583* | 2201 | N: bcm583* |
| 2202 | N: bcm113* | 2202 | N: bcm113* |
| 2203 | 2203 | ||
| 2204 | BROADCOM KONA GPIO DRIVER | 2204 | BROADCOM KONA GPIO DRIVER |
| 2205 | M: Ray Jui <rjui@broadcom.com> | 2205 | M: Ray Jui <rjui@broadcom.com> |
| 2206 | L: bcm-kernel-feedback-list@broadcom.com | 2206 | L: bcm-kernel-feedback-list@broadcom.com |
| 2207 | S: Supported | 2207 | S: Supported |
| 2208 | F: drivers/gpio/gpio-bcm-kona.c | 2208 | F: drivers/gpio/gpio-bcm-kona.c |
| 2209 | F: Documentation/devicetree/bindings/gpio/gpio-bcm-kona.txt | 2209 | F: Documentation/devicetree/bindings/gpio/gpio-bcm-kona.txt |
| 2210 | 2210 | ||
| 2211 | BROADCOM SPECIFIC AMBA DRIVER (BCMA) | 2211 | BROADCOM SPECIFIC AMBA DRIVER (BCMA) |
| 2212 | M: Rafaล Miลecki <zajec5@gmail.com> | 2212 | M: Rafaล Miลecki <zajec5@gmail.com> |
| 2213 | L: linux-wireless@vger.kernel.org | 2213 | L: linux-wireless@vger.kernel.org |
| 2214 | S: Maintained | 2214 | S: Maintained |
| 2215 | F: drivers/bcma/ | 2215 | F: drivers/bcma/ |
| 2216 | F: include/linux/bcma/ | 2216 | F: include/linux/bcma/ |
| 2217 | 2217 | ||
| 2218 | BROADCOM SYSTEMPORT ETHERNET DRIVER | 2218 | BROADCOM SYSTEMPORT ETHERNET DRIVER |
| 2219 | M: Florian Fainelli <f.fainelli@gmail.com> | 2219 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 2220 | L: netdev@vger.kernel.org | 2220 | L: netdev@vger.kernel.org |
| 2221 | S: Supported | 2221 | S: Supported |
| 2222 | F: drivers/net/ethernet/broadcom/bcmsysport.* | 2222 | F: drivers/net/ethernet/broadcom/bcmsysport.* |
| 2223 | 2223 | ||
| 2224 | BROCADE BFA FC SCSI DRIVER | 2224 | BROCADE BFA FC SCSI DRIVER |
| 2225 | M: Anil Gurumurthy <anil.gurumurthy@qlogic.com> | 2225 | M: Anil Gurumurthy <anil.gurumurthy@qlogic.com> |
| 2226 | M: Sudarsana Kalluru <sudarsana.kalluru@qlogic.com> | 2226 | M: Sudarsana Kalluru <sudarsana.kalluru@qlogic.com> |
| 2227 | L: linux-scsi@vger.kernel.org | 2227 | L: linux-scsi@vger.kernel.org |
| 2228 | S: Supported | 2228 | S: Supported |
| 2229 | F: drivers/scsi/bfa/ | 2229 | F: drivers/scsi/bfa/ |
| 2230 | 2230 | ||
| 2231 | BROCADE BNA 10 GIGABIT ETHERNET DRIVER | 2231 | BROCADE BNA 10 GIGABIT ETHERNET DRIVER |
| 2232 | M: Rasesh Mody <rasesh.mody@qlogic.com> | 2232 | M: Rasesh Mody <rasesh.mody@qlogic.com> |
| 2233 | L: netdev@vger.kernel.org | 2233 | L: netdev@vger.kernel.org |
| 2234 | S: Supported | 2234 | S: Supported |
| 2235 | F: drivers/net/ethernet/brocade/bna/ | 2235 | F: drivers/net/ethernet/brocade/bna/ |
| 2236 | 2236 | ||
| 2237 | BSG (block layer generic sg v4 driver) | 2237 | BSG (block layer generic sg v4 driver) |
| 2238 | M: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> | 2238 | M: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> |
| 2239 | L: linux-scsi@vger.kernel.org | 2239 | L: linux-scsi@vger.kernel.org |
| 2240 | S: Supported | 2240 | S: Supported |
| 2241 | F: block/bsg.c | 2241 | F: block/bsg.c |
| 2242 | F: include/linux/bsg.h | 2242 | F: include/linux/bsg.h |
| 2243 | F: include/uapi/linux/bsg.h | 2243 | F: include/uapi/linux/bsg.h |
| 2244 | 2244 | ||
| 2245 | BT87X AUDIO DRIVER | 2245 | BT87X AUDIO DRIVER |
| 2246 | M: Clemens Ladisch <clemens@ladisch.de> | 2246 | M: Clemens Ladisch <clemens@ladisch.de> |
| 2247 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 2247 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 2248 | T: git git://git.alsa-project.org/alsa-kernel.git | 2248 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 2249 | S: Maintained | 2249 | S: Maintained |
| 2250 | F: Documentation/sound/alsa/Bt87x.txt | 2250 | F: Documentation/sound/alsa/Bt87x.txt |
| 2251 | F: sound/pci/bt87x.c | 2251 | F: sound/pci/bt87x.c |
| 2252 | 2252 | ||
| 2253 | BT8XXGPIO DRIVER | 2253 | BT8XXGPIO DRIVER |
| 2254 | M: Michael Buesch <m@bues.ch> | 2254 | M: Michael Buesch <m@bues.ch> |
| 2255 | W: http://bu3sch.de/btgpio.php | 2255 | W: http://bu3sch.de/btgpio.php |
| 2256 | S: Maintained | 2256 | S: Maintained |
| 2257 | F: drivers/gpio/gpio-bt8xx.c | 2257 | F: drivers/gpio/gpio-bt8xx.c |
| 2258 | 2258 | ||
| 2259 | BTRFS FILE SYSTEM | 2259 | BTRFS FILE SYSTEM |
| 2260 | M: Chris Mason <clm@fb.com> | 2260 | M: Chris Mason <clm@fb.com> |
| 2261 | M: Josef Bacik <jbacik@fb.com> | 2261 | M: Josef Bacik <jbacik@fb.com> |
| 2262 | M: David Sterba <dsterba@suse.cz> | ||
| 2262 | L: linux-btrfs@vger.kernel.org | 2263 | L: linux-btrfs@vger.kernel.org |
| 2263 | W: http://btrfs.wiki.kernel.org/ | 2264 | W: http://btrfs.wiki.kernel.org/ |
| 2264 | Q: http://patchwork.kernel.org/project/linux-btrfs/list/ | 2265 | Q: http://patchwork.kernel.org/project/linux-btrfs/list/ |
| 2265 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs.git | 2266 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs.git |
| 2266 | S: Maintained | 2267 | S: Maintained |
| 2267 | F: Documentation/filesystems/btrfs.txt | 2268 | F: Documentation/filesystems/btrfs.txt |
| 2268 | F: fs/btrfs/ | 2269 | F: fs/btrfs/ |
| 2269 | 2270 | ||
| 2270 | BTTV VIDEO4LINUX DRIVER | 2271 | BTTV VIDEO4LINUX DRIVER |
| 2271 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 2272 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 2272 | L: linux-media@vger.kernel.org | 2273 | L: linux-media@vger.kernel.org |
| 2273 | W: http://linuxtv.org | 2274 | W: http://linuxtv.org |
| 2274 | T: git git://linuxtv.org/media_tree.git | 2275 | T: git git://linuxtv.org/media_tree.git |
| 2275 | S: Odd fixes | 2276 | S: Odd fixes |
| 2276 | F: Documentation/video4linux/bttv/ | 2277 | F: Documentation/video4linux/bttv/ |
| 2277 | F: drivers/media/pci/bt8xx/bttv* | 2278 | F: drivers/media/pci/bt8xx/bttv* |
| 2278 | 2279 | ||
| 2279 | BUSLOGIC SCSI DRIVER | 2280 | BUSLOGIC SCSI DRIVER |
| 2280 | M: Khalid Aziz <khalid@gonehiking.org> | 2281 | M: Khalid Aziz <khalid@gonehiking.org> |
| 2281 | L: linux-scsi@vger.kernel.org | 2282 | L: linux-scsi@vger.kernel.org |
| 2282 | S: Maintained | 2283 | S: Maintained |
| 2283 | F: drivers/scsi/BusLogic.* | 2284 | F: drivers/scsi/BusLogic.* |
| 2284 | F: drivers/scsi/FlashPoint.* | 2285 | F: drivers/scsi/FlashPoint.* |
| 2285 | 2286 | ||
| 2286 | C-MEDIA CMI8788 DRIVER | 2287 | C-MEDIA CMI8788 DRIVER |
| 2287 | M: Clemens Ladisch <clemens@ladisch.de> | 2288 | M: Clemens Ladisch <clemens@ladisch.de> |
| 2288 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 2289 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 2289 | T: git git://git.alsa-project.org/alsa-kernel.git | 2290 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 2290 | S: Maintained | 2291 | S: Maintained |
| 2291 | F: sound/pci/oxygen/ | 2292 | F: sound/pci/oxygen/ |
| 2292 | 2293 | ||
| 2293 | C6X ARCHITECTURE | 2294 | C6X ARCHITECTURE |
| 2294 | M: Mark Salter <msalter@redhat.com> | 2295 | M: Mark Salter <msalter@redhat.com> |
| 2295 | M: Aurelien Jacquiot <a-jacquiot@ti.com> | 2296 | M: Aurelien Jacquiot <a-jacquiot@ti.com> |
| 2296 | L: linux-c6x-dev@linux-c6x.org | 2297 | L: linux-c6x-dev@linux-c6x.org |
| 2297 | W: http://www.linux-c6x.org/wiki/index.php/Main_Page | 2298 | W: http://www.linux-c6x.org/wiki/index.php/Main_Page |
| 2298 | S: Maintained | 2299 | S: Maintained |
| 2299 | F: arch/c6x/ | 2300 | F: arch/c6x/ |
| 2300 | 2301 | ||
| 2301 | CACHEFILES: FS-CACHE BACKEND FOR CACHING ON MOUNTED FILESYSTEMS | 2302 | CACHEFILES: FS-CACHE BACKEND FOR CACHING ON MOUNTED FILESYSTEMS |
| 2302 | M: David Howells <dhowells@redhat.com> | 2303 | M: David Howells <dhowells@redhat.com> |
| 2303 | L: linux-cachefs@redhat.com | 2304 | L: linux-cachefs@redhat.com |
| 2304 | S: Supported | 2305 | S: Supported |
| 2305 | F: Documentation/filesystems/caching/cachefiles.txt | 2306 | F: Documentation/filesystems/caching/cachefiles.txt |
| 2306 | F: fs/cachefiles/ | 2307 | F: fs/cachefiles/ |
| 2307 | 2308 | ||
| 2308 | CADET FM/AM RADIO RECEIVER DRIVER | 2309 | CADET FM/AM RADIO RECEIVER DRIVER |
| 2309 | M: Hans Verkuil <hverkuil@xs4all.nl> | 2310 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 2310 | L: linux-media@vger.kernel.org | 2311 | L: linux-media@vger.kernel.org |
| 2311 | T: git git://linuxtv.org/media_tree.git | 2312 | T: git git://linuxtv.org/media_tree.git |
| 2312 | W: http://linuxtv.org | 2313 | W: http://linuxtv.org |
| 2313 | S: Maintained | 2314 | S: Maintained |
| 2314 | F: drivers/media/radio/radio-cadet* | 2315 | F: drivers/media/radio/radio-cadet* |
| 2315 | 2316 | ||
| 2316 | CAFE CMOS INTEGRATED CAMERA CONTROLLER DRIVER | 2317 | CAFE CMOS INTEGRATED CAMERA CONTROLLER DRIVER |
| 2317 | M: Jonathan Corbet <corbet@lwn.net> | 2318 | M: Jonathan Corbet <corbet@lwn.net> |
| 2318 | L: linux-media@vger.kernel.org | 2319 | L: linux-media@vger.kernel.org |
| 2319 | T: git git://linuxtv.org/media_tree.git | 2320 | T: git git://linuxtv.org/media_tree.git |
| 2320 | S: Maintained | 2321 | S: Maintained |
| 2321 | F: Documentation/video4linux/cafe_ccic | 2322 | F: Documentation/video4linux/cafe_ccic |
| 2322 | F: drivers/media/platform/marvell-ccic/ | 2323 | F: drivers/media/platform/marvell-ccic/ |
| 2323 | 2324 | ||
| 2324 | CAIF NETWORK LAYER | 2325 | CAIF NETWORK LAYER |
| 2325 | M: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no> | 2326 | M: Dmitry Tarnyagin <dmitry.tarnyagin@lockless.no> |
| 2326 | L: netdev@vger.kernel.org | 2327 | L: netdev@vger.kernel.org |
| 2327 | S: Supported | 2328 | S: Supported |
| 2328 | F: Documentation/networking/caif/ | 2329 | F: Documentation/networking/caif/ |
| 2329 | F: drivers/net/caif/ | 2330 | F: drivers/net/caif/ |
| 2330 | F: include/uapi/linux/caif/ | 2331 | F: include/uapi/linux/caif/ |
| 2331 | F: include/net/caif/ | 2332 | F: include/net/caif/ |
| 2332 | F: net/caif/ | 2333 | F: net/caif/ |
| 2333 | 2334 | ||
| 2334 | CALGARY x86-64 IOMMU | 2335 | CALGARY x86-64 IOMMU |
| 2335 | M: Muli Ben-Yehuda <muli@il.ibm.com> | 2336 | M: Muli Ben-Yehuda <muli@il.ibm.com> |
| 2336 | M: "Jon D. Mason" <jdmason@kudzu.us> | 2337 | M: "Jon D. Mason" <jdmason@kudzu.us> |
| 2337 | L: discuss@x86-64.org | 2338 | L: discuss@x86-64.org |
| 2338 | S: Maintained | 2339 | S: Maintained |
| 2339 | F: arch/x86/kernel/pci-calgary_64.c | 2340 | F: arch/x86/kernel/pci-calgary_64.c |
| 2340 | F: arch/x86/kernel/tce_64.c | 2341 | F: arch/x86/kernel/tce_64.c |
| 2341 | F: arch/x86/include/asm/calgary.h | 2342 | F: arch/x86/include/asm/calgary.h |
| 2342 | F: arch/x86/include/asm/tce.h | 2343 | F: arch/x86/include/asm/tce.h |
| 2343 | 2344 | ||
| 2344 | CAN NETWORK LAYER | 2345 | CAN NETWORK LAYER |
| 2345 | M: Oliver Hartkopp <socketcan@hartkopp.net> | 2346 | M: Oliver Hartkopp <socketcan@hartkopp.net> |
| 2346 | L: linux-can@vger.kernel.org | 2347 | L: linux-can@vger.kernel.org |
| 2347 | W: http://gitorious.org/linux-can | 2348 | W: http://gitorious.org/linux-can |
| 2348 | T: git git://gitorious.org/linux-can/linux-can-next.git | 2349 | T: git git://gitorious.org/linux-can/linux-can-next.git |
| 2349 | S: Maintained | 2350 | S: Maintained |
| 2350 | F: Documentation/networking/can.txt | 2351 | F: Documentation/networking/can.txt |
| 2351 | F: net/can/ | 2352 | F: net/can/ |
| 2352 | F: include/linux/can/core.h | 2353 | F: include/linux/can/core.h |
| 2353 | F: include/uapi/linux/can.h | 2354 | F: include/uapi/linux/can.h |
| 2354 | F: include/uapi/linux/can/bcm.h | 2355 | F: include/uapi/linux/can/bcm.h |
| 2355 | F: include/uapi/linux/can/raw.h | 2356 | F: include/uapi/linux/can/raw.h |
| 2356 | F: include/uapi/linux/can/gw.h | 2357 | F: include/uapi/linux/can/gw.h |
| 2357 | 2358 | ||
| 2358 | CAN NETWORK DRIVERS | 2359 | CAN NETWORK DRIVERS |
| 2359 | M: Wolfgang Grandegger <wg@grandegger.com> | 2360 | M: Wolfgang Grandegger <wg@grandegger.com> |
| 2360 | M: Marc Kleine-Budde <mkl@pengutronix.de> | 2361 | M: Marc Kleine-Budde <mkl@pengutronix.de> |
| 2361 | L: linux-can@vger.kernel.org | 2362 | L: linux-can@vger.kernel.org |
| 2362 | W: http://gitorious.org/linux-can | 2363 | W: http://gitorious.org/linux-can |
| 2363 | T: git git://gitorious.org/linux-can/linux-can-next.git | 2364 | T: git git://gitorious.org/linux-can/linux-can-next.git |
| 2364 | S: Maintained | 2365 | S: Maintained |
| 2365 | F: drivers/net/can/ | 2366 | F: drivers/net/can/ |
| 2366 | F: include/linux/can/dev.h | 2367 | F: include/linux/can/dev.h |
| 2367 | F: include/linux/can/platform/ | 2368 | F: include/linux/can/platform/ |
| 2368 | F: include/uapi/linux/can/error.h | 2369 | F: include/uapi/linux/can/error.h |
| 2369 | F: include/uapi/linux/can/netlink.h | 2370 | F: include/uapi/linux/can/netlink.h |
| 2370 | 2371 | ||
| 2371 | CAPABILITIES | 2372 | CAPABILITIES |
| 2372 | M: Serge Hallyn <serge.hallyn@canonical.com> | 2373 | M: Serge Hallyn <serge.hallyn@canonical.com> |
| 2373 | L: linux-security-module@vger.kernel.org | 2374 | L: linux-security-module@vger.kernel.org |
| 2374 | S: Supported | 2375 | S: Supported |
| 2375 | F: include/linux/capability.h | 2376 | F: include/linux/capability.h |
| 2376 | F: include/uapi/linux/capability.h | 2377 | F: include/uapi/linux/capability.h |
| 2377 | F: security/capability.c | 2378 | F: security/capability.c |
| 2378 | F: security/commoncap.c | 2379 | F: security/commoncap.c |
| 2379 | F: kernel/capability.c | 2380 | F: kernel/capability.c |
| 2380 | 2381 | ||
| 2381 | CC2520 IEEE-802.15.4 RADIO DRIVER | 2382 | CC2520 IEEE-802.15.4 RADIO DRIVER |
| 2382 | M: Varka Bhadram <varkabhadram@gmail.com> | 2383 | M: Varka Bhadram <varkabhadram@gmail.com> |
| 2383 | L: linux-wpan@vger.kernel.org | 2384 | L: linux-wpan@vger.kernel.org |
| 2384 | S: Maintained | 2385 | S: Maintained |
| 2385 | F: drivers/net/ieee802154/cc2520.c | 2386 | F: drivers/net/ieee802154/cc2520.c |
| 2386 | F: include/linux/spi/cc2520.h | 2387 | F: include/linux/spi/cc2520.h |
| 2387 | F: Documentation/devicetree/bindings/net/ieee802154/cc2520.txt | 2388 | F: Documentation/devicetree/bindings/net/ieee802154/cc2520.txt |
| 2388 | 2389 | ||
| 2389 | CELL BROADBAND ENGINE ARCHITECTURE | 2390 | CELL BROADBAND ENGINE ARCHITECTURE |
| 2390 | M: Arnd Bergmann <arnd@arndb.de> | 2391 | M: Arnd Bergmann <arnd@arndb.de> |
| 2391 | L: linuxppc-dev@lists.ozlabs.org | 2392 | L: linuxppc-dev@lists.ozlabs.org |
| 2392 | L: cbe-oss-dev@lists.ozlabs.org | 2393 | L: cbe-oss-dev@lists.ozlabs.org |
| 2393 | W: http://www.ibm.com/developerworks/power/cell/ | 2394 | W: http://www.ibm.com/developerworks/power/cell/ |
| 2394 | S: Supported | 2395 | S: Supported |
| 2395 | F: arch/powerpc/include/asm/cell*.h | 2396 | F: arch/powerpc/include/asm/cell*.h |
| 2396 | F: arch/powerpc/include/asm/spu*.h | 2397 | F: arch/powerpc/include/asm/spu*.h |
| 2397 | F: arch/powerpc/include/uapi/asm/spu*.h | 2398 | F: arch/powerpc/include/uapi/asm/spu*.h |
| 2398 | F: arch/powerpc/oprofile/*cell* | 2399 | F: arch/powerpc/oprofile/*cell* |
| 2399 | F: arch/powerpc/platforms/cell/ | 2400 | F: arch/powerpc/platforms/cell/ |
| 2400 | 2401 | ||
| 2401 | CEPH DISTRIBUTED FILE SYSTEM CLIENT | 2402 | CEPH DISTRIBUTED FILE SYSTEM CLIENT |
| 2402 | M: Sage Weil <sage@inktank.com> | 2403 | M: Sage Weil <sage@inktank.com> |
| 2403 | L: ceph-devel@vger.kernel.org | 2404 | L: ceph-devel@vger.kernel.org |
| 2404 | W: http://ceph.com/ | 2405 | W: http://ceph.com/ |
| 2405 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git | 2406 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git |
| 2406 | S: Supported | 2407 | S: Supported |
| 2407 | F: Documentation/filesystems/ceph.txt | 2408 | F: Documentation/filesystems/ceph.txt |
| 2408 | F: fs/ceph/ | 2409 | F: fs/ceph/ |
| 2409 | F: net/ceph/ | 2410 | F: net/ceph/ |
| 2410 | F: include/linux/ceph/ | 2411 | F: include/linux/ceph/ |
| 2411 | F: include/linux/crush/ | 2412 | F: include/linux/crush/ |
| 2412 | 2413 | ||
| 2413 | CERTIFIED WIRELESS USB (WUSB) SUBSYSTEM: | 2414 | CERTIFIED WIRELESS USB (WUSB) SUBSYSTEM: |
| 2414 | L: linux-usb@vger.kernel.org | 2415 | L: linux-usb@vger.kernel.org |
| 2415 | S: Orphan | 2416 | S: Orphan |
| 2416 | F: Documentation/usb/WUSB-Design-overview.txt | 2417 | F: Documentation/usb/WUSB-Design-overview.txt |
| 2417 | F: Documentation/usb/wusb-cbaf | 2418 | F: Documentation/usb/wusb-cbaf |
| 2418 | F: drivers/usb/host/hwa-hc.c | 2419 | F: drivers/usb/host/hwa-hc.c |
| 2419 | F: drivers/usb/host/whci/ | 2420 | F: drivers/usb/host/whci/ |
| 2420 | F: drivers/usb/wusbcore/ | 2421 | F: drivers/usb/wusbcore/ |
| 2421 | F: include/linux/usb/wusb* | 2422 | F: include/linux/usb/wusb* |
| 2422 | 2423 | ||
| 2423 | CFAG12864B LCD DRIVER | 2424 | CFAG12864B LCD DRIVER |
| 2424 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> | 2425 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> |
| 2425 | W: http://miguelojeda.es/auxdisplay.htm | 2426 | W: http://miguelojeda.es/auxdisplay.htm |
| 2426 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm | 2427 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm |
| 2427 | S: Maintained | 2428 | S: Maintained |
| 2428 | F: drivers/auxdisplay/cfag12864b.c | 2429 | F: drivers/auxdisplay/cfag12864b.c |
| 2429 | F: include/linux/cfag12864b.h | 2430 | F: include/linux/cfag12864b.h |
| 2430 | 2431 | ||
| 2431 | CFAG12864BFB LCD FRAMEBUFFER DRIVER | 2432 | CFAG12864BFB LCD FRAMEBUFFER DRIVER |
| 2432 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> | 2433 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> |
| 2433 | W: http://miguelojeda.es/auxdisplay.htm | 2434 | W: http://miguelojeda.es/auxdisplay.htm |
| 2434 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm | 2435 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm |
| 2435 | S: Maintained | 2436 | S: Maintained |
| 2436 | F: drivers/auxdisplay/cfag12864bfb.c | 2437 | F: drivers/auxdisplay/cfag12864bfb.c |
| 2437 | F: include/linux/cfag12864b.h | 2438 | F: include/linux/cfag12864b.h |
| 2438 | 2439 | ||
| 2439 | CFG80211 and NL80211 | 2440 | CFG80211 and NL80211 |
| 2440 | M: Johannes Berg <johannes@sipsolutions.net> | 2441 | M: Johannes Berg <johannes@sipsolutions.net> |
| 2441 | L: linux-wireless@vger.kernel.org | 2442 | L: linux-wireless@vger.kernel.org |
| 2442 | W: http://wireless.kernel.org/ | 2443 | W: http://wireless.kernel.org/ |
| 2443 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211.git | 2444 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211.git |
| 2444 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next.git | 2445 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next.git |
| 2445 | S: Maintained | 2446 | S: Maintained |
| 2446 | F: include/uapi/linux/nl80211.h | 2447 | F: include/uapi/linux/nl80211.h |
| 2447 | F: include/net/cfg80211.h | 2448 | F: include/net/cfg80211.h |
| 2448 | F: net/wireless/* | 2449 | F: net/wireless/* |
| 2449 | X: net/wireless/wext* | 2450 | X: net/wireless/wext* |
| 2450 | 2451 | ||
| 2451 | CHAR and MISC DRIVERS | 2452 | CHAR and MISC DRIVERS |
| 2452 | M: Arnd Bergmann <arnd@arndb.de> | 2453 | M: Arnd Bergmann <arnd@arndb.de> |
| 2453 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2454 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 2454 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc.git | 2455 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc.git |
| 2455 | S: Supported | 2456 | S: Supported |
| 2456 | F: drivers/char/* | 2457 | F: drivers/char/* |
| 2457 | F: drivers/misc/* | 2458 | F: drivers/misc/* |
| 2458 | F: include/linux/miscdevice.h | 2459 | F: include/linux/miscdevice.h |
| 2459 | 2460 | ||
| 2460 | CHECKPATCH | 2461 | CHECKPATCH |
| 2461 | M: Andy Whitcroft <apw@canonical.com> | 2462 | M: Andy Whitcroft <apw@canonical.com> |
| 2462 | M: Joe Perches <joe@perches.com> | 2463 | M: Joe Perches <joe@perches.com> |
| 2463 | S: Maintained | 2464 | S: Maintained |
| 2464 | F: scripts/checkpatch.pl | 2465 | F: scripts/checkpatch.pl |
| 2465 | 2466 | ||
| 2466 | CHINESE DOCUMENTATION | 2467 | CHINESE DOCUMENTATION |
| 2467 | M: Harry Wei <harryxiyou@gmail.com> | 2468 | M: Harry Wei <harryxiyou@gmail.com> |
| 2468 | L: xiyoulinuxkernelgroup@googlegroups.com (subscribers-only) | 2469 | L: xiyoulinuxkernelgroup@googlegroups.com (subscribers-only) |
| 2469 | L: linux-kernel@zh-kernel.org (moderated for non-subscribers) | 2470 | L: linux-kernel@zh-kernel.org (moderated for non-subscribers) |
| 2470 | S: Maintained | 2471 | S: Maintained |
| 2471 | F: Documentation/zh_CN/ | 2472 | F: Documentation/zh_CN/ |
| 2472 | 2473 | ||
| 2473 | CHIPIDEA USB HIGH SPEED DUAL ROLE CONTROLLER | 2474 | CHIPIDEA USB HIGH SPEED DUAL ROLE CONTROLLER |
| 2474 | M: Peter Chen <Peter.Chen@freescale.com> | 2475 | M: Peter Chen <Peter.Chen@freescale.com> |
| 2475 | T: git git://github.com/hzpeterchen/linux-usb.git | 2476 | T: git git://github.com/hzpeterchen/linux-usb.git |
| 2476 | L: linux-usb@vger.kernel.org | 2477 | L: linux-usb@vger.kernel.org |
| 2477 | S: Maintained | 2478 | S: Maintained |
| 2478 | F: drivers/usb/chipidea/ | 2479 | F: drivers/usb/chipidea/ |
| 2479 | 2480 | ||
| 2480 | CHROME HARDWARE PLATFORM SUPPORT | 2481 | CHROME HARDWARE PLATFORM SUPPORT |
| 2481 | M: Olof Johansson <olof@lixom.net> | 2482 | M: Olof Johansson <olof@lixom.net> |
| 2482 | S: Maintained | 2483 | S: Maintained |
| 2483 | F: drivers/platform/chrome/ | 2484 | F: drivers/platform/chrome/ |
| 2484 | 2485 | ||
| 2485 | CISCO VIC ETHERNET NIC DRIVER | 2486 | CISCO VIC ETHERNET NIC DRIVER |
| 2486 | M: Christian Benvenuti <benve@cisco.com> | 2487 | M: Christian Benvenuti <benve@cisco.com> |
| 2487 | M: Sujith Sankar <ssujith@cisco.com> | 2488 | M: Sujith Sankar <ssujith@cisco.com> |
| 2488 | M: Govindarajulu Varadarajan <_govind@gmx.com> | 2489 | M: Govindarajulu Varadarajan <_govind@gmx.com> |
| 2489 | M: Neel Patel <neepatel@cisco.com> | 2490 | M: Neel Patel <neepatel@cisco.com> |
| 2490 | S: Supported | 2491 | S: Supported |
| 2491 | F: drivers/net/ethernet/cisco/enic/ | 2492 | F: drivers/net/ethernet/cisco/enic/ |
| 2492 | 2493 | ||
| 2493 | CISCO VIC LOW LATENCY NIC DRIVER | 2494 | CISCO VIC LOW LATENCY NIC DRIVER |
| 2494 | M: Upinder Malhi <umalhi@cisco.com> | 2495 | M: Upinder Malhi <umalhi@cisco.com> |
| 2495 | S: Supported | 2496 | S: Supported |
| 2496 | F: drivers/infiniband/hw/usnic | 2497 | F: drivers/infiniband/hw/usnic |
| 2497 | 2498 | ||
| 2498 | CIRRUS LOGIC EP93XX ETHERNET DRIVER | 2499 | CIRRUS LOGIC EP93XX ETHERNET DRIVER |
| 2499 | M: Hartley Sweeten <hsweeten@visionengravers.com> | 2500 | M: Hartley Sweeten <hsweeten@visionengravers.com> |
| 2500 | L: netdev@vger.kernel.org | 2501 | L: netdev@vger.kernel.org |
| 2501 | S: Maintained | 2502 | S: Maintained |
| 2502 | F: drivers/net/ethernet/cirrus/ep93xx_eth.c | 2503 | F: drivers/net/ethernet/cirrus/ep93xx_eth.c |
| 2503 | 2504 | ||
| 2504 | CIRRUS LOGIC AUDIO CODEC DRIVERS | 2505 | CIRRUS LOGIC AUDIO CODEC DRIVERS |
| 2505 | M: Brian Austin <brian.austin@cirrus.com> | 2506 | M: Brian Austin <brian.austin@cirrus.com> |
| 2506 | M: Paul Handrigan <Paul.Handrigan@cirrus.com> | 2507 | M: Paul Handrigan <Paul.Handrigan@cirrus.com> |
| 2507 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 2508 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 2508 | S: Maintained | 2509 | S: Maintained |
| 2509 | F: sound/soc/codecs/cs* | 2510 | F: sound/soc/codecs/cs* |
| 2510 | 2511 | ||
| 2511 | CLEANCACHE API | 2512 | CLEANCACHE API |
| 2512 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 2513 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 2513 | L: linux-kernel@vger.kernel.org | 2514 | L: linux-kernel@vger.kernel.org |
| 2514 | S: Maintained | 2515 | S: Maintained |
| 2515 | F: mm/cleancache.c | 2516 | F: mm/cleancache.c |
| 2516 | F: include/linux/cleancache.h | 2517 | F: include/linux/cleancache.h |
| 2517 | 2518 | ||
| 2518 | CLK API | 2519 | CLK API |
| 2519 | M: Russell King <linux@arm.linux.org.uk> | 2520 | M: Russell King <linux@arm.linux.org.uk> |
| 2520 | S: Maintained | 2521 | S: Maintained |
| 2521 | F: include/linux/clk.h | 2522 | F: include/linux/clk.h |
| 2522 | 2523 | ||
| 2523 | CLOCKSOURCE, CLOCKEVENT DRIVERS | 2524 | CLOCKSOURCE, CLOCKEVENT DRIVERS |
| 2524 | M: Daniel Lezcano <daniel.lezcano@linaro.org> | 2525 | M: Daniel Lezcano <daniel.lezcano@linaro.org> |
| 2525 | M: Thomas Gleixner <tglx@linutronix.de> | 2526 | M: Thomas Gleixner <tglx@linutronix.de> |
| 2526 | L: linux-kernel@vger.kernel.org | 2527 | L: linux-kernel@vger.kernel.org |
| 2527 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core | 2528 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core |
| 2528 | S: Supported | 2529 | S: Supported |
| 2529 | F: drivers/clocksource | 2530 | F: drivers/clocksource |
| 2530 | 2531 | ||
| 2531 | CISCO FCOE HBA DRIVER | 2532 | CISCO FCOE HBA DRIVER |
| 2532 | M: Hiral Patel <hiralpat@cisco.com> | 2533 | M: Hiral Patel <hiralpat@cisco.com> |
| 2533 | M: Suma Ramars <sramars@cisco.com> | 2534 | M: Suma Ramars <sramars@cisco.com> |
| 2534 | M: Brian Uchino <buchino@cisco.com> | 2535 | M: Brian Uchino <buchino@cisco.com> |
| 2535 | L: linux-scsi@vger.kernel.org | 2536 | L: linux-scsi@vger.kernel.org |
| 2536 | S: Supported | 2537 | S: Supported |
| 2537 | F: drivers/scsi/fnic/ | 2538 | F: drivers/scsi/fnic/ |
| 2538 | 2539 | ||
| 2539 | CMPC ACPI DRIVER | 2540 | CMPC ACPI DRIVER |
| 2540 | M: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com> | 2541 | M: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com> |
| 2541 | M: Daniel Oliveira Nascimento <don@syst.com.br> | 2542 | M: Daniel Oliveira Nascimento <don@syst.com.br> |
| 2542 | L: platform-driver-x86@vger.kernel.org | 2543 | L: platform-driver-x86@vger.kernel.org |
| 2543 | S: Supported | 2544 | S: Supported |
| 2544 | F: drivers/platform/x86/classmate-laptop.c | 2545 | F: drivers/platform/x86/classmate-laptop.c |
| 2545 | 2546 | ||
| 2546 | COCCINELLE/Semantic Patches (SmPL) | 2547 | COCCINELLE/Semantic Patches (SmPL) |
| 2547 | M: Julia Lawall <Julia.Lawall@lip6.fr> | 2548 | M: Julia Lawall <Julia.Lawall@lip6.fr> |
| 2548 | M: Gilles Muller <Gilles.Muller@lip6.fr> | 2549 | M: Gilles Muller <Gilles.Muller@lip6.fr> |
| 2549 | M: Nicolas Palix <nicolas.palix@imag.fr> | 2550 | M: Nicolas Palix <nicolas.palix@imag.fr> |
| 2550 | M: Michal Marek <mmarek@suse.cz> | 2551 | M: Michal Marek <mmarek@suse.cz> |
| 2551 | L: cocci@systeme.lip6.fr (moderated for non-subscribers) | 2552 | L: cocci@systeme.lip6.fr (moderated for non-subscribers) |
| 2552 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git misc | 2553 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git misc |
| 2553 | W: http://coccinelle.lip6.fr/ | 2554 | W: http://coccinelle.lip6.fr/ |
| 2554 | S: Supported | 2555 | S: Supported |
| 2555 | F: Documentation/coccinelle.txt | 2556 | F: Documentation/coccinelle.txt |
| 2556 | F: scripts/coccinelle/ | 2557 | F: scripts/coccinelle/ |
| 2557 | F: scripts/coccicheck | 2558 | F: scripts/coccicheck |
| 2558 | 2559 | ||
| 2559 | CODA FILE SYSTEM | 2560 | CODA FILE SYSTEM |
| 2560 | M: Jan Harkes <jaharkes@cs.cmu.edu> | 2561 | M: Jan Harkes <jaharkes@cs.cmu.edu> |
| 2561 | M: coda@cs.cmu.edu | 2562 | M: coda@cs.cmu.edu |
| 2562 | L: codalist@coda.cs.cmu.edu | 2563 | L: codalist@coda.cs.cmu.edu |
| 2563 | W: http://www.coda.cs.cmu.edu/ | 2564 | W: http://www.coda.cs.cmu.edu/ |
| 2564 | S: Maintained | 2565 | S: Maintained |
| 2565 | F: Documentation/filesystems/coda.txt | 2566 | F: Documentation/filesystems/coda.txt |
| 2566 | F: fs/coda/ | 2567 | F: fs/coda/ |
| 2567 | F: include/linux/coda*.h | 2568 | F: include/linux/coda*.h |
| 2568 | F: include/uapi/linux/coda*.h | 2569 | F: include/uapi/linux/coda*.h |
| 2569 | 2570 | ||
| 2570 | CODA V4L2 MEM2MEM DRIVER | 2571 | CODA V4L2 MEM2MEM DRIVER |
| 2571 | M: Philipp Zabel <p.zabel@pengutronix.de> | 2572 | M: Philipp Zabel <p.zabel@pengutronix.de> |
| 2572 | L: linux-media@vger.kernel.org | 2573 | L: linux-media@vger.kernel.org |
| 2573 | S: Maintained | 2574 | S: Maintained |
| 2574 | F: Documentation/devicetree/bindings/media/coda.txt | 2575 | F: Documentation/devicetree/bindings/media/coda.txt |
| 2575 | F: drivers/media/platform/coda/ | 2576 | F: drivers/media/platform/coda/ |
| 2576 | 2577 | ||
| 2577 | COMMON CLK FRAMEWORK | 2578 | COMMON CLK FRAMEWORK |
| 2578 | M: Mike Turquette <mturquette@linaro.org> | 2579 | M: Mike Turquette <mturquette@linaro.org> |
| 2579 | M: Stephen Boyd <sboyd@codeaurora.org> | 2580 | M: Stephen Boyd <sboyd@codeaurora.org> |
| 2580 | L: linux-kernel@vger.kernel.org | 2581 | L: linux-kernel@vger.kernel.org |
| 2581 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux.git | 2582 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux.git |
| 2582 | S: Maintained | 2583 | S: Maintained |
| 2583 | F: drivers/clk/ | 2584 | F: drivers/clk/ |
| 2584 | X: drivers/clk/clkdev.c | 2585 | X: drivers/clk/clkdev.c |
| 2585 | F: include/linux/clk-pr* | 2586 | F: include/linux/clk-pr* |
| 2586 | F: include/linux/clk/ | 2587 | F: include/linux/clk/ |
| 2587 | 2588 | ||
| 2588 | COMMON INTERNET FILE SYSTEM (CIFS) | 2589 | COMMON INTERNET FILE SYSTEM (CIFS) |
| 2589 | M: Steve French <sfrench@samba.org> | 2590 | M: Steve French <sfrench@samba.org> |
| 2590 | L: linux-cifs@vger.kernel.org | 2591 | L: linux-cifs@vger.kernel.org |
| 2591 | L: samba-technical@lists.samba.org (moderated for non-subscribers) | 2592 | L: samba-technical@lists.samba.org (moderated for non-subscribers) |
| 2592 | W: http://linux-cifs.samba.org/ | 2593 | W: http://linux-cifs.samba.org/ |
| 2593 | T: git git://git.samba.org/sfrench/cifs-2.6.git | 2594 | T: git git://git.samba.org/sfrench/cifs-2.6.git |
| 2594 | S: Supported | 2595 | S: Supported |
| 2595 | F: Documentation/filesystems/cifs/ | 2596 | F: Documentation/filesystems/cifs/ |
| 2596 | F: fs/cifs/ | 2597 | F: fs/cifs/ |
| 2597 | 2598 | ||
| 2598 | COMPACTPCI HOTPLUG CORE | 2599 | COMPACTPCI HOTPLUG CORE |
| 2599 | M: Scott Murray <scott@spiteful.org> | 2600 | M: Scott Murray <scott@spiteful.org> |
| 2600 | L: linux-pci@vger.kernel.org | 2601 | L: linux-pci@vger.kernel.org |
| 2601 | S: Maintained | 2602 | S: Maintained |
| 2602 | F: drivers/pci/hotplug/cpci_hotplug* | 2603 | F: drivers/pci/hotplug/cpci_hotplug* |
| 2603 | 2604 | ||
| 2604 | COMPACTPCI HOTPLUG ZIATECH ZT5550 DRIVER | 2605 | COMPACTPCI HOTPLUG ZIATECH ZT5550 DRIVER |
| 2605 | M: Scott Murray <scott@spiteful.org> | 2606 | M: Scott Murray <scott@spiteful.org> |
| 2606 | L: linux-pci@vger.kernel.org | 2607 | L: linux-pci@vger.kernel.org |
| 2607 | S: Maintained | 2608 | S: Maintained |
| 2608 | F: drivers/pci/hotplug/cpcihp_zt5550.* | 2609 | F: drivers/pci/hotplug/cpcihp_zt5550.* |
| 2609 | 2610 | ||
| 2610 | COMPACTPCI HOTPLUG GENERIC DRIVER | 2611 | COMPACTPCI HOTPLUG GENERIC DRIVER |
| 2611 | M: Scott Murray <scott@spiteful.org> | 2612 | M: Scott Murray <scott@spiteful.org> |
| 2612 | L: linux-pci@vger.kernel.org | 2613 | L: linux-pci@vger.kernel.org |
| 2613 | S: Maintained | 2614 | S: Maintained |
| 2614 | F: drivers/pci/hotplug/cpcihp_generic.c | 2615 | F: drivers/pci/hotplug/cpcihp_generic.c |
| 2615 | 2616 | ||
| 2616 | COMPAL LAPTOP SUPPORT | 2617 | COMPAL LAPTOP SUPPORT |
| 2617 | M: Cezary Jackiewicz <cezary.jackiewicz@gmail.com> | 2618 | M: Cezary Jackiewicz <cezary.jackiewicz@gmail.com> |
| 2618 | L: platform-driver-x86@vger.kernel.org | 2619 | L: platform-driver-x86@vger.kernel.org |
| 2619 | S: Maintained | 2620 | S: Maintained |
| 2620 | F: drivers/platform/x86/compal-laptop.c | 2621 | F: drivers/platform/x86/compal-laptop.c |
| 2621 | 2622 | ||
| 2622 | CONEXANT ACCESSRUNNER USB DRIVER | 2623 | CONEXANT ACCESSRUNNER USB DRIVER |
| 2623 | M: Simon Arlott <cxacru@fire.lp0.eu> | 2624 | M: Simon Arlott <cxacru@fire.lp0.eu> |
| 2624 | L: accessrunner-general@lists.sourceforge.net | 2625 | L: accessrunner-general@lists.sourceforge.net |
| 2625 | W: http://accessrunner.sourceforge.net/ | 2626 | W: http://accessrunner.sourceforge.net/ |
| 2626 | S: Maintained | 2627 | S: Maintained |
| 2627 | F: drivers/usb/atm/cxacru.c | 2628 | F: drivers/usb/atm/cxacru.c |
| 2628 | 2629 | ||
| 2629 | CONFIGFS | 2630 | CONFIGFS |
| 2630 | M: Joel Becker <jlbec@evilplan.org> | 2631 | M: Joel Becker <jlbec@evilplan.org> |
| 2631 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/configfs.git | 2632 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/configfs.git |
| 2632 | S: Supported | 2633 | S: Supported |
| 2633 | F: fs/configfs/ | 2634 | F: fs/configfs/ |
| 2634 | F: include/linux/configfs.h | 2635 | F: include/linux/configfs.h |
| 2635 | 2636 | ||
| 2636 | CONNECTOR | 2637 | CONNECTOR |
| 2637 | M: Evgeniy Polyakov <zbr@ioremap.net> | 2638 | M: Evgeniy Polyakov <zbr@ioremap.net> |
| 2638 | L: netdev@vger.kernel.org | 2639 | L: netdev@vger.kernel.org |
| 2639 | S: Maintained | 2640 | S: Maintained |
| 2640 | F: drivers/connector/ | 2641 | F: drivers/connector/ |
| 2641 | 2642 | ||
| 2642 | CONTROL GROUP (CGROUP) | 2643 | CONTROL GROUP (CGROUP) |
| 2643 | M: Tejun Heo <tj@kernel.org> | 2644 | M: Tejun Heo <tj@kernel.org> |
| 2644 | M: Li Zefan <lizefan@huawei.com> | 2645 | M: Li Zefan <lizefan@huawei.com> |
| 2645 | L: cgroups@vger.kernel.org | 2646 | L: cgroups@vger.kernel.org |
| 2646 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git | 2647 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git |
| 2647 | S: Maintained | 2648 | S: Maintained |
| 2648 | F: Documentation/cgroups/ | 2649 | F: Documentation/cgroups/ |
| 2649 | F: include/linux/cgroup* | 2650 | F: include/linux/cgroup* |
| 2650 | F: kernel/cgroup* | 2651 | F: kernel/cgroup* |
| 2651 | 2652 | ||
| 2652 | CONTROL GROUP - CPUSET | 2653 | CONTROL GROUP - CPUSET |
| 2653 | M: Li Zefan <lizefan@huawei.com> | 2654 | M: Li Zefan <lizefan@huawei.com> |
| 2654 | L: cgroups@vger.kernel.org | 2655 | L: cgroups@vger.kernel.org |
| 2655 | W: http://www.bullopensource.org/cpuset/ | 2656 | W: http://www.bullopensource.org/cpuset/ |
| 2656 | W: http://oss.sgi.com/projects/cpusets/ | 2657 | W: http://oss.sgi.com/projects/cpusets/ |
| 2657 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git | 2658 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git |
| 2658 | S: Maintained | 2659 | S: Maintained |
| 2659 | F: Documentation/cgroups/cpusets.txt | 2660 | F: Documentation/cgroups/cpusets.txt |
| 2660 | F: include/linux/cpuset.h | 2661 | F: include/linux/cpuset.h |
| 2661 | F: kernel/cpuset.c | 2662 | F: kernel/cpuset.c |
| 2662 | 2663 | ||
| 2663 | CONTROL GROUP - MEMORY RESOURCE CONTROLLER (MEMCG) | 2664 | CONTROL GROUP - MEMORY RESOURCE CONTROLLER (MEMCG) |
| 2664 | M: Johannes Weiner <hannes@cmpxchg.org> | 2665 | M: Johannes Weiner <hannes@cmpxchg.org> |
| 2665 | M: Michal Hocko <mhocko@suse.cz> | 2666 | M: Michal Hocko <mhocko@suse.cz> |
| 2666 | L: cgroups@vger.kernel.org | 2667 | L: cgroups@vger.kernel.org |
| 2667 | L: linux-mm@kvack.org | 2668 | L: linux-mm@kvack.org |
| 2668 | S: Maintained | 2669 | S: Maintained |
| 2669 | F: mm/memcontrol.c | 2670 | F: mm/memcontrol.c |
| 2670 | F: mm/swap_cgroup.c | 2671 | F: mm/swap_cgroup.c |
| 2671 | 2672 | ||
| 2672 | CORETEMP HARDWARE MONITORING DRIVER | 2673 | CORETEMP HARDWARE MONITORING DRIVER |
| 2673 | M: Fenghua Yu <fenghua.yu@intel.com> | 2674 | M: Fenghua Yu <fenghua.yu@intel.com> |
| 2674 | L: lm-sensors@lm-sensors.org | 2675 | L: lm-sensors@lm-sensors.org |
| 2675 | S: Maintained | 2676 | S: Maintained |
| 2676 | F: Documentation/hwmon/coretemp | 2677 | F: Documentation/hwmon/coretemp |
| 2677 | F: drivers/hwmon/coretemp.c | 2678 | F: drivers/hwmon/coretemp.c |
| 2678 | 2679 | ||
| 2679 | COSA/SRP SYNC SERIAL DRIVER | 2680 | COSA/SRP SYNC SERIAL DRIVER |
| 2680 | M: Jan "Yenya" Kasprzak <kas@fi.muni.cz> | 2681 | M: Jan "Yenya" Kasprzak <kas@fi.muni.cz> |
| 2681 | W: http://www.fi.muni.cz/~kas/cosa/ | 2682 | W: http://www.fi.muni.cz/~kas/cosa/ |
| 2682 | S: Maintained | 2683 | S: Maintained |
| 2683 | F: drivers/net/wan/cosa* | 2684 | F: drivers/net/wan/cosa* |
| 2684 | 2685 | ||
| 2685 | CPMAC ETHERNET DRIVER | 2686 | CPMAC ETHERNET DRIVER |
| 2686 | M: Florian Fainelli <florian@openwrt.org> | 2687 | M: Florian Fainelli <florian@openwrt.org> |
| 2687 | L: netdev@vger.kernel.org | 2688 | L: netdev@vger.kernel.org |
| 2688 | S: Maintained | 2689 | S: Maintained |
| 2689 | F: drivers/net/ethernet/ti/cpmac.c | 2690 | F: drivers/net/ethernet/ti/cpmac.c |
| 2690 | 2691 | ||
| 2691 | CPU FREQUENCY DRIVERS | 2692 | CPU FREQUENCY DRIVERS |
| 2692 | M: Rafael J. Wysocki <rjw@rjwysocki.net> | 2693 | M: Rafael J. Wysocki <rjw@rjwysocki.net> |
| 2693 | M: Viresh Kumar <viresh.kumar@linaro.org> | 2694 | M: Viresh Kumar <viresh.kumar@linaro.org> |
| 2694 | L: linux-pm@vger.kernel.org | 2695 | L: linux-pm@vger.kernel.org |
| 2695 | S: Maintained | 2696 | S: Maintained |
| 2696 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git | 2697 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git |
| 2697 | T: git git://git.linaro.org/people/vireshk/linux.git (For ARM Updates) | 2698 | T: git git://git.linaro.org/people/vireshk/linux.git (For ARM Updates) |
| 2698 | F: drivers/cpufreq/ | 2699 | F: drivers/cpufreq/ |
| 2699 | F: include/linux/cpufreq.h | 2700 | F: include/linux/cpufreq.h |
| 2700 | 2701 | ||
| 2701 | CPU FREQUENCY DRIVERS - ARM BIG LITTLE | 2702 | CPU FREQUENCY DRIVERS - ARM BIG LITTLE |
| 2702 | M: Viresh Kumar <viresh.kumar@linaro.org> | 2703 | M: Viresh Kumar <viresh.kumar@linaro.org> |
| 2703 | M: Sudeep Holla <sudeep.holla@arm.com> | 2704 | M: Sudeep Holla <sudeep.holla@arm.com> |
| 2704 | L: linux-pm@vger.kernel.org | 2705 | L: linux-pm@vger.kernel.org |
| 2705 | W: http://www.arm.com/products/processors/technologies/biglittleprocessing.php | 2706 | W: http://www.arm.com/products/processors/technologies/biglittleprocessing.php |
| 2706 | S: Maintained | 2707 | S: Maintained |
| 2707 | F: drivers/cpufreq/arm_big_little.h | 2708 | F: drivers/cpufreq/arm_big_little.h |
| 2708 | F: drivers/cpufreq/arm_big_little.c | 2709 | F: drivers/cpufreq/arm_big_little.c |
| 2709 | F: drivers/cpufreq/arm_big_little_dt.c | 2710 | F: drivers/cpufreq/arm_big_little_dt.c |
| 2710 | 2711 | ||
| 2711 | CPUIDLE DRIVER - ARM BIG LITTLE | 2712 | CPUIDLE DRIVER - ARM BIG LITTLE |
| 2712 | M: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> | 2713 | M: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> |
| 2713 | M: Daniel Lezcano <daniel.lezcano@linaro.org> | 2714 | M: Daniel Lezcano <daniel.lezcano@linaro.org> |
| 2714 | L: linux-pm@vger.kernel.org | 2715 | L: linux-pm@vger.kernel.org |
| 2715 | L: linux-arm-kernel@lists.infradead.org | 2716 | L: linux-arm-kernel@lists.infradead.org |
| 2716 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git | 2717 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git |
| 2717 | S: Maintained | 2718 | S: Maintained |
| 2718 | F: drivers/cpuidle/cpuidle-big_little.c | 2719 | F: drivers/cpuidle/cpuidle-big_little.c |
| 2719 | 2720 | ||
| 2720 | CPUIDLE DRIVER - ARM EXYNOS | 2721 | CPUIDLE DRIVER - ARM EXYNOS |
| 2721 | M: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> | 2722 | M: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> |
| 2722 | M: Daniel Lezcano <daniel.lezcano@linaro.org> | 2723 | M: Daniel Lezcano <daniel.lezcano@linaro.org> |
| 2723 | M: Kukjin Kim <kgene@kernel.org> | 2724 | M: Kukjin Kim <kgene@kernel.org> |
| 2724 | L: linux-pm@vger.kernel.org | 2725 | L: linux-pm@vger.kernel.org |
| 2725 | L: linux-samsung-soc@vger.kernel.org | 2726 | L: linux-samsung-soc@vger.kernel.org |
| 2726 | S: Supported | 2727 | S: Supported |
| 2727 | F: drivers/cpuidle/cpuidle-exynos.c | 2728 | F: drivers/cpuidle/cpuidle-exynos.c |
| 2728 | F: arch/arm/mach-exynos/pm.c | 2729 | F: arch/arm/mach-exynos/pm.c |
| 2729 | 2730 | ||
| 2730 | CPUIDLE DRIVERS | 2731 | CPUIDLE DRIVERS |
| 2731 | M: Rafael J. Wysocki <rjw@rjwysocki.net> | 2732 | M: Rafael J. Wysocki <rjw@rjwysocki.net> |
| 2732 | M: Daniel Lezcano <daniel.lezcano@linaro.org> | 2733 | M: Daniel Lezcano <daniel.lezcano@linaro.org> |
| 2733 | L: linux-pm@vger.kernel.org | 2734 | L: linux-pm@vger.kernel.org |
| 2734 | S: Maintained | 2735 | S: Maintained |
| 2735 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git | 2736 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git |
| 2736 | F: drivers/cpuidle/* | 2737 | F: drivers/cpuidle/* |
| 2737 | F: include/linux/cpuidle.h | 2738 | F: include/linux/cpuidle.h |
| 2738 | 2739 | ||
| 2739 | CPUID/MSR DRIVER | 2740 | CPUID/MSR DRIVER |
| 2740 | M: "H. Peter Anvin" <hpa@zytor.com> | 2741 | M: "H. Peter Anvin" <hpa@zytor.com> |
| 2741 | S: Maintained | 2742 | S: Maintained |
| 2742 | F: arch/x86/kernel/cpuid.c | 2743 | F: arch/x86/kernel/cpuid.c |
| 2743 | F: arch/x86/kernel/msr.c | 2744 | F: arch/x86/kernel/msr.c |
| 2744 | 2745 | ||
| 2745 | CPU POWER MONITORING SUBSYSTEM | 2746 | CPU POWER MONITORING SUBSYSTEM |
| 2746 | M: Thomas Renninger <trenn@suse.de> | 2747 | M: Thomas Renninger <trenn@suse.de> |
| 2747 | L: linux-pm@vger.kernel.org | 2748 | L: linux-pm@vger.kernel.org |
| 2748 | S: Maintained | 2749 | S: Maintained |
| 2749 | F: tools/power/cpupower/ | 2750 | F: tools/power/cpupower/ |
| 2750 | 2751 | ||
| 2751 | CRAMFS FILESYSTEM | 2752 | CRAMFS FILESYSTEM |
| 2752 | W: http://sourceforge.net/projects/cramfs/ | 2753 | W: http://sourceforge.net/projects/cramfs/ |
| 2753 | S: Orphan / Obsolete | 2754 | S: Orphan / Obsolete |
| 2754 | F: Documentation/filesystems/cramfs.txt | 2755 | F: Documentation/filesystems/cramfs.txt |
| 2755 | F: fs/cramfs/ | 2756 | F: fs/cramfs/ |
| 2756 | 2757 | ||
| 2757 | CRIS PORT | 2758 | CRIS PORT |
| 2758 | M: Mikael Starvik <starvik@axis.com> | 2759 | M: Mikael Starvik <starvik@axis.com> |
| 2759 | M: Jesper Nilsson <jesper.nilsson@axis.com> | 2760 | M: Jesper Nilsson <jesper.nilsson@axis.com> |
| 2760 | L: linux-cris-kernel@axis.com | 2761 | L: linux-cris-kernel@axis.com |
| 2761 | W: http://developer.axis.com | 2762 | W: http://developer.axis.com |
| 2762 | S: Maintained | 2763 | S: Maintained |
| 2763 | F: arch/cris/ | 2764 | F: arch/cris/ |
| 2764 | F: drivers/tty/serial/crisv10.* | 2765 | F: drivers/tty/serial/crisv10.* |
| 2765 | 2766 | ||
| 2766 | CRYPTO API | 2767 | CRYPTO API |
| 2767 | M: Herbert Xu <herbert@gondor.apana.org.au> | 2768 | M: Herbert Xu <herbert@gondor.apana.org.au> |
| 2768 | M: "David S. Miller" <davem@davemloft.net> | 2769 | M: "David S. Miller" <davem@davemloft.net> |
| 2769 | L: linux-crypto@vger.kernel.org | 2770 | L: linux-crypto@vger.kernel.org |
| 2770 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6.git | 2771 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6.git |
| 2771 | S: Maintained | 2772 | S: Maintained |
| 2772 | F: Documentation/crypto/ | 2773 | F: Documentation/crypto/ |
| 2773 | F: arch/*/crypto/ | 2774 | F: arch/*/crypto/ |
| 2774 | F: crypto/ | 2775 | F: crypto/ |
| 2775 | F: drivers/crypto/ | 2776 | F: drivers/crypto/ |
| 2776 | F: include/crypto/ | 2777 | F: include/crypto/ |
| 2777 | 2778 | ||
| 2778 | CRYPTOGRAPHIC RANDOM NUMBER GENERATOR | 2779 | CRYPTOGRAPHIC RANDOM NUMBER GENERATOR |
| 2779 | M: Neil Horman <nhorman@tuxdriver.com> | 2780 | M: Neil Horman <nhorman@tuxdriver.com> |
| 2780 | L: linux-crypto@vger.kernel.org | 2781 | L: linux-crypto@vger.kernel.org |
| 2781 | S: Maintained | 2782 | S: Maintained |
| 2782 | F: crypto/ansi_cprng.c | 2783 | F: crypto/ansi_cprng.c |
| 2783 | F: crypto/rng.c | 2784 | F: crypto/rng.c |
| 2784 | 2785 | ||
| 2785 | CS5535 Audio ALSA driver | 2786 | CS5535 Audio ALSA driver |
| 2786 | M: Jaya Kumar <jayakumar.alsa@gmail.com> | 2787 | M: Jaya Kumar <jayakumar.alsa@gmail.com> |
| 2787 | S: Maintained | 2788 | S: Maintained |
| 2788 | F: sound/pci/cs5535audio/ | 2789 | F: sound/pci/cs5535audio/ |
| 2789 | 2790 | ||
| 2790 | CW1200 WLAN driver | 2791 | CW1200 WLAN driver |
| 2791 | M: Solomon Peachy <pizza@shaftnet.org> | 2792 | M: Solomon Peachy <pizza@shaftnet.org> |
| 2792 | S: Maintained | 2793 | S: Maintained |
| 2793 | F: drivers/net/wireless/cw1200/ | 2794 | F: drivers/net/wireless/cw1200/ |
| 2794 | 2795 | ||
| 2795 | CX18 VIDEO4LINUX DRIVER | 2796 | CX18 VIDEO4LINUX DRIVER |
| 2796 | M: Andy Walls <awalls@md.metrocast.net> | 2797 | M: Andy Walls <awalls@md.metrocast.net> |
| 2797 | L: ivtv-devel@ivtvdriver.org (subscribers-only) | 2798 | L: ivtv-devel@ivtvdriver.org (subscribers-only) |
| 2798 | L: linux-media@vger.kernel.org | 2799 | L: linux-media@vger.kernel.org |
| 2799 | T: git git://linuxtv.org/media_tree.git | 2800 | T: git git://linuxtv.org/media_tree.git |
| 2800 | W: http://linuxtv.org | 2801 | W: http://linuxtv.org |
| 2801 | W: http://www.ivtvdriver.org/index.php/Cx18 | 2802 | W: http://www.ivtvdriver.org/index.php/Cx18 |
| 2802 | S: Maintained | 2803 | S: Maintained |
| 2803 | F: Documentation/video4linux/cx18.txt | 2804 | F: Documentation/video4linux/cx18.txt |
| 2804 | F: drivers/media/pci/cx18/ | 2805 | F: drivers/media/pci/cx18/ |
| 2805 | F: include/uapi/linux/ivtv* | 2806 | F: include/uapi/linux/ivtv* |
| 2806 | 2807 | ||
| 2807 | CX2341X MPEG ENCODER HELPER MODULE | 2808 | CX2341X MPEG ENCODER HELPER MODULE |
| 2808 | M: Hans Verkuil <hverkuil@xs4all.nl> | 2809 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 2809 | L: linux-media@vger.kernel.org | 2810 | L: linux-media@vger.kernel.org |
| 2810 | T: git git://linuxtv.org/media_tree.git | 2811 | T: git git://linuxtv.org/media_tree.git |
| 2811 | W: http://linuxtv.org | 2812 | W: http://linuxtv.org |
| 2812 | S: Maintained | 2813 | S: Maintained |
| 2813 | F: drivers/media/common/cx2341x* | 2814 | F: drivers/media/common/cx2341x* |
| 2814 | F: include/media/cx2341x* | 2815 | F: include/media/cx2341x* |
| 2815 | 2816 | ||
| 2816 | CX88 VIDEO4LINUX DRIVER | 2817 | CX88 VIDEO4LINUX DRIVER |
| 2817 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 2818 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 2818 | L: linux-media@vger.kernel.org | 2819 | L: linux-media@vger.kernel.org |
| 2819 | W: http://linuxtv.org | 2820 | W: http://linuxtv.org |
| 2820 | T: git git://linuxtv.org/media_tree.git | 2821 | T: git git://linuxtv.org/media_tree.git |
| 2821 | S: Odd fixes | 2822 | S: Odd fixes |
| 2822 | F: Documentation/video4linux/cx88/ | 2823 | F: Documentation/video4linux/cx88/ |
| 2823 | F: drivers/media/pci/cx88/ | 2824 | F: drivers/media/pci/cx88/ |
| 2824 | 2825 | ||
| 2825 | CXD2820R MEDIA DRIVER | 2826 | CXD2820R MEDIA DRIVER |
| 2826 | M: Antti Palosaari <crope@iki.fi> | 2827 | M: Antti Palosaari <crope@iki.fi> |
| 2827 | L: linux-media@vger.kernel.org | 2828 | L: linux-media@vger.kernel.org |
| 2828 | W: http://linuxtv.org/ | 2829 | W: http://linuxtv.org/ |
| 2829 | W: http://palosaari.fi/linux/ | 2830 | W: http://palosaari.fi/linux/ |
| 2830 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 2831 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 2831 | T: git git://linuxtv.org/anttip/media_tree.git | 2832 | T: git git://linuxtv.org/anttip/media_tree.git |
| 2832 | S: Maintained | 2833 | S: Maintained |
| 2833 | F: drivers/media/dvb-frontends/cxd2820r* | 2834 | F: drivers/media/dvb-frontends/cxd2820r* |
| 2834 | 2835 | ||
| 2835 | CXGB3 ETHERNET DRIVER (CXGB3) | 2836 | CXGB3 ETHERNET DRIVER (CXGB3) |
| 2836 | M: Santosh Raspatur <santosh@chelsio.com> | 2837 | M: Santosh Raspatur <santosh@chelsio.com> |
| 2837 | L: netdev@vger.kernel.org | 2838 | L: netdev@vger.kernel.org |
| 2838 | W: http://www.chelsio.com | 2839 | W: http://www.chelsio.com |
| 2839 | S: Supported | 2840 | S: Supported |
| 2840 | F: drivers/net/ethernet/chelsio/cxgb3/ | 2841 | F: drivers/net/ethernet/chelsio/cxgb3/ |
| 2841 | 2842 | ||
| 2842 | CXGB3 ISCSI DRIVER (CXGB3I) | 2843 | CXGB3 ISCSI DRIVER (CXGB3I) |
| 2843 | M: Karen Xie <kxie@chelsio.com> | 2844 | M: Karen Xie <kxie@chelsio.com> |
| 2844 | L: linux-scsi@vger.kernel.org | 2845 | L: linux-scsi@vger.kernel.org |
| 2845 | W: http://www.chelsio.com | 2846 | W: http://www.chelsio.com |
| 2846 | S: Supported | 2847 | S: Supported |
| 2847 | F: drivers/scsi/cxgbi/cxgb3i | 2848 | F: drivers/scsi/cxgbi/cxgb3i |
| 2848 | 2849 | ||
| 2849 | CXGB3 IWARP RNIC DRIVER (IW_CXGB3) | 2850 | CXGB3 IWARP RNIC DRIVER (IW_CXGB3) |
| 2850 | M: Steve Wise <swise@chelsio.com> | 2851 | M: Steve Wise <swise@chelsio.com> |
| 2851 | L: linux-rdma@vger.kernel.org | 2852 | L: linux-rdma@vger.kernel.org |
| 2852 | W: http://www.openfabrics.org | 2853 | W: http://www.openfabrics.org |
| 2853 | S: Supported | 2854 | S: Supported |
| 2854 | F: drivers/infiniband/hw/cxgb3/ | 2855 | F: drivers/infiniband/hw/cxgb3/ |
| 2855 | 2856 | ||
| 2856 | CXGB4 ETHERNET DRIVER (CXGB4) | 2857 | CXGB4 ETHERNET DRIVER (CXGB4) |
| 2857 | M: Hariprasad S <hariprasad@chelsio.com> | 2858 | M: Hariprasad S <hariprasad@chelsio.com> |
| 2858 | L: netdev@vger.kernel.org | 2859 | L: netdev@vger.kernel.org |
| 2859 | W: http://www.chelsio.com | 2860 | W: http://www.chelsio.com |
| 2860 | S: Supported | 2861 | S: Supported |
| 2861 | F: drivers/net/ethernet/chelsio/cxgb4/ | 2862 | F: drivers/net/ethernet/chelsio/cxgb4/ |
| 2862 | 2863 | ||
| 2863 | CXGB4 ISCSI DRIVER (CXGB4I) | 2864 | CXGB4 ISCSI DRIVER (CXGB4I) |
| 2864 | M: Karen Xie <kxie@chelsio.com> | 2865 | M: Karen Xie <kxie@chelsio.com> |
| 2865 | L: linux-scsi@vger.kernel.org | 2866 | L: linux-scsi@vger.kernel.org |
| 2866 | W: http://www.chelsio.com | 2867 | W: http://www.chelsio.com |
| 2867 | S: Supported | 2868 | S: Supported |
| 2868 | F: drivers/scsi/cxgbi/cxgb4i | 2869 | F: drivers/scsi/cxgbi/cxgb4i |
| 2869 | 2870 | ||
| 2870 | CXGB4 IWARP RNIC DRIVER (IW_CXGB4) | 2871 | CXGB4 IWARP RNIC DRIVER (IW_CXGB4) |
| 2871 | M: Steve Wise <swise@chelsio.com> | 2872 | M: Steve Wise <swise@chelsio.com> |
| 2872 | L: linux-rdma@vger.kernel.org | 2873 | L: linux-rdma@vger.kernel.org |
| 2873 | W: http://www.openfabrics.org | 2874 | W: http://www.openfabrics.org |
| 2874 | S: Supported | 2875 | S: Supported |
| 2875 | F: drivers/infiniband/hw/cxgb4/ | 2876 | F: drivers/infiniband/hw/cxgb4/ |
| 2876 | 2877 | ||
| 2877 | CXGB4VF ETHERNET DRIVER (CXGB4VF) | 2878 | CXGB4VF ETHERNET DRIVER (CXGB4VF) |
| 2878 | M: Casey Leedom <leedom@chelsio.com> | 2879 | M: Casey Leedom <leedom@chelsio.com> |
| 2879 | L: netdev@vger.kernel.org | 2880 | L: netdev@vger.kernel.org |
| 2880 | W: http://www.chelsio.com | 2881 | W: http://www.chelsio.com |
| 2881 | S: Supported | 2882 | S: Supported |
| 2882 | F: drivers/net/ethernet/chelsio/cxgb4vf/ | 2883 | F: drivers/net/ethernet/chelsio/cxgb4vf/ |
| 2883 | 2884 | ||
| 2884 | CXL (IBM Coherent Accelerator Processor Interface CAPI) DRIVER | 2885 | CXL (IBM Coherent Accelerator Processor Interface CAPI) DRIVER |
| 2885 | M: Ian Munsie <imunsie@au1.ibm.com> | 2886 | M: Ian Munsie <imunsie@au1.ibm.com> |
| 2886 | M: Michael Neuling <mikey@neuling.org> | 2887 | M: Michael Neuling <mikey@neuling.org> |
| 2887 | L: linuxppc-dev@lists.ozlabs.org | 2888 | L: linuxppc-dev@lists.ozlabs.org |
| 2888 | S: Supported | 2889 | S: Supported |
| 2889 | F: drivers/misc/cxl/ | 2890 | F: drivers/misc/cxl/ |
| 2890 | F: include/misc/cxl.h | 2891 | F: include/misc/cxl.h |
| 2891 | F: include/uapi/misc/cxl.h | 2892 | F: include/uapi/misc/cxl.h |
| 2892 | F: Documentation/powerpc/cxl.txt | 2893 | F: Documentation/powerpc/cxl.txt |
| 2893 | F: Documentation/powerpc/cxl.txt | 2894 | F: Documentation/powerpc/cxl.txt |
| 2894 | F: Documentation/ABI/testing/sysfs-class-cxl | 2895 | F: Documentation/ABI/testing/sysfs-class-cxl |
| 2895 | 2896 | ||
| 2896 | STMMAC ETHERNET DRIVER | 2897 | STMMAC ETHERNET DRIVER |
| 2897 | M: Giuseppe Cavallaro <peppe.cavallaro@st.com> | 2898 | M: Giuseppe Cavallaro <peppe.cavallaro@st.com> |
| 2898 | L: netdev@vger.kernel.org | 2899 | L: netdev@vger.kernel.org |
| 2899 | W: http://www.stlinux.com | 2900 | W: http://www.stlinux.com |
| 2900 | S: Supported | 2901 | S: Supported |
| 2901 | F: drivers/net/ethernet/stmicro/stmmac/ | 2902 | F: drivers/net/ethernet/stmicro/stmmac/ |
| 2902 | 2903 | ||
| 2903 | CYBERPRO FB DRIVER | 2904 | CYBERPRO FB DRIVER |
| 2904 | M: Russell King <linux@arm.linux.org.uk> | 2905 | M: Russell King <linux@arm.linux.org.uk> |
| 2905 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 2906 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 2906 | W: http://www.arm.linux.org.uk/ | 2907 | W: http://www.arm.linux.org.uk/ |
| 2907 | S: Maintained | 2908 | S: Maintained |
| 2908 | F: drivers/video/fbdev/cyber2000fb.* | 2909 | F: drivers/video/fbdev/cyber2000fb.* |
| 2909 | 2910 | ||
| 2910 | CYCLADES ASYNC MUX DRIVER | 2911 | CYCLADES ASYNC MUX DRIVER |
| 2911 | W: http://www.cyclades.com/ | 2912 | W: http://www.cyclades.com/ |
| 2912 | S: Orphan | 2913 | S: Orphan |
| 2913 | F: drivers/tty/cyclades.c | 2914 | F: drivers/tty/cyclades.c |
| 2914 | F: include/linux/cyclades.h | 2915 | F: include/linux/cyclades.h |
| 2915 | F: include/uapi/linux/cyclades.h | 2916 | F: include/uapi/linux/cyclades.h |
| 2916 | 2917 | ||
| 2917 | CYCLADES PC300 DRIVER | 2918 | CYCLADES PC300 DRIVER |
| 2918 | W: http://www.cyclades.com/ | 2919 | W: http://www.cyclades.com/ |
| 2919 | S: Orphan | 2920 | S: Orphan |
| 2920 | F: drivers/net/wan/pc300* | 2921 | F: drivers/net/wan/pc300* |
| 2921 | 2922 | ||
| 2922 | CYPRESS_FIRMWARE MEDIA DRIVER | 2923 | CYPRESS_FIRMWARE MEDIA DRIVER |
| 2923 | M: Antti Palosaari <crope@iki.fi> | 2924 | M: Antti Palosaari <crope@iki.fi> |
| 2924 | L: linux-media@vger.kernel.org | 2925 | L: linux-media@vger.kernel.org |
| 2925 | W: http://linuxtv.org/ | 2926 | W: http://linuxtv.org/ |
| 2926 | W: http://palosaari.fi/linux/ | 2927 | W: http://palosaari.fi/linux/ |
| 2927 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 2928 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 2928 | T: git git://linuxtv.org/anttip/media_tree.git | 2929 | T: git git://linuxtv.org/anttip/media_tree.git |
| 2929 | S: Maintained | 2930 | S: Maintained |
| 2930 | F: drivers/media/common/cypress_firmware* | 2931 | F: drivers/media/common/cypress_firmware* |
| 2931 | 2932 | ||
| 2932 | CYTTSP TOUCHSCREEN DRIVER | 2933 | CYTTSP TOUCHSCREEN DRIVER |
| 2933 | M: Ferruh Yigit <fery@cypress.com> | 2934 | M: Ferruh Yigit <fery@cypress.com> |
| 2934 | L: linux-input@vger.kernel.org | 2935 | L: linux-input@vger.kernel.org |
| 2935 | S: Supported | 2936 | S: Supported |
| 2936 | F: drivers/input/touchscreen/cyttsp* | 2937 | F: drivers/input/touchscreen/cyttsp* |
| 2937 | F: include/linux/input/cyttsp.h | 2938 | F: include/linux/input/cyttsp.h |
| 2938 | 2939 | ||
| 2939 | DAMA SLAVE for AX.25 | 2940 | DAMA SLAVE for AX.25 |
| 2940 | M: Joerg Reuter <jreuter@yaina.de> | 2941 | M: Joerg Reuter <jreuter@yaina.de> |
| 2941 | W: http://yaina.de/jreuter/ | 2942 | W: http://yaina.de/jreuter/ |
| 2942 | W: http://www.qsl.net/dl1bke/ | 2943 | W: http://www.qsl.net/dl1bke/ |
| 2943 | L: linux-hams@vger.kernel.org | 2944 | L: linux-hams@vger.kernel.org |
| 2944 | S: Maintained | 2945 | S: Maintained |
| 2945 | F: net/ax25/af_ax25.c | 2946 | F: net/ax25/af_ax25.c |
| 2946 | F: net/ax25/ax25_dev.c | 2947 | F: net/ax25/ax25_dev.c |
| 2947 | F: net/ax25/ax25_ds_* | 2948 | F: net/ax25/ax25_ds_* |
| 2948 | F: net/ax25/ax25_in.c | 2949 | F: net/ax25/ax25_in.c |
| 2949 | F: net/ax25/ax25_out.c | 2950 | F: net/ax25/ax25_out.c |
| 2950 | F: net/ax25/ax25_timer.c | 2951 | F: net/ax25/ax25_timer.c |
| 2951 | F: net/ax25/sysctl_net_ax25.c | 2952 | F: net/ax25/sysctl_net_ax25.c |
| 2952 | 2953 | ||
| 2953 | DAVICOM FAST ETHERNET (DMFE) NETWORK DRIVER | 2954 | DAVICOM FAST ETHERNET (DMFE) NETWORK DRIVER |
| 2954 | L: netdev@vger.kernel.org | 2955 | L: netdev@vger.kernel.org |
| 2955 | S: Orphan | 2956 | S: Orphan |
| 2956 | F: Documentation/networking/dmfe.txt | 2957 | F: Documentation/networking/dmfe.txt |
| 2957 | F: drivers/net/ethernet/dec/tulip/dmfe.c | 2958 | F: drivers/net/ethernet/dec/tulip/dmfe.c |
| 2958 | 2959 | ||
| 2959 | DC390/AM53C974 SCSI driver | 2960 | DC390/AM53C974 SCSI driver |
| 2960 | M: Hannes Reinecke <hare@suse.de> | 2961 | M: Hannes Reinecke <hare@suse.de> |
| 2961 | L: linux-scsi@vger.kernel.org | 2962 | L: linux-scsi@vger.kernel.org |
| 2962 | S: Maintained | 2963 | S: Maintained |
| 2963 | F: drivers/scsi/am53c974.c | 2964 | F: drivers/scsi/am53c974.c |
| 2964 | 2965 | ||
| 2965 | DC395x SCSI driver | 2966 | DC395x SCSI driver |
| 2966 | M: Oliver Neukum <oliver@neukum.org> | 2967 | M: Oliver Neukum <oliver@neukum.org> |
| 2967 | M: Ali Akcaagac <aliakc@web.de> | 2968 | M: Ali Akcaagac <aliakc@web.de> |
| 2968 | M: Jamie Lenehan <lenehan@twibble.org> | 2969 | M: Jamie Lenehan <lenehan@twibble.org> |
| 2969 | L: dc395x@twibble.org | 2970 | L: dc395x@twibble.org |
| 2970 | W: http://twibble.org/dist/dc395x/ | 2971 | W: http://twibble.org/dist/dc395x/ |
| 2971 | W: http://lists.twibble.org/mailman/listinfo/dc395x/ | 2972 | W: http://lists.twibble.org/mailman/listinfo/dc395x/ |
| 2972 | S: Maintained | 2973 | S: Maintained |
| 2973 | F: Documentation/scsi/dc395x.txt | 2974 | F: Documentation/scsi/dc395x.txt |
| 2974 | F: drivers/scsi/dc395x.* | 2975 | F: drivers/scsi/dc395x.* |
| 2975 | 2976 | ||
| 2976 | DCCP PROTOCOL | 2977 | DCCP PROTOCOL |
| 2977 | M: Gerrit Renker <gerrit@erg.abdn.ac.uk> | 2978 | M: Gerrit Renker <gerrit@erg.abdn.ac.uk> |
| 2978 | L: dccp@vger.kernel.org | 2979 | L: dccp@vger.kernel.org |
| 2979 | W: http://www.linuxfoundation.org/collaborate/workgroups/networking/dccp | 2980 | W: http://www.linuxfoundation.org/collaborate/workgroups/networking/dccp |
| 2980 | S: Maintained | 2981 | S: Maintained |
| 2981 | F: include/linux/dccp.h | 2982 | F: include/linux/dccp.h |
| 2982 | F: include/uapi/linux/dccp.h | 2983 | F: include/uapi/linux/dccp.h |
| 2983 | F: include/linux/tfrc.h | 2984 | F: include/linux/tfrc.h |
| 2984 | F: net/dccp/ | 2985 | F: net/dccp/ |
| 2985 | 2986 | ||
| 2986 | DECnet NETWORK LAYER | 2987 | DECnet NETWORK LAYER |
| 2987 | W: http://linux-decnet.sourceforge.net | 2988 | W: http://linux-decnet.sourceforge.net |
| 2988 | L: linux-decnet-user@lists.sourceforge.net | 2989 | L: linux-decnet-user@lists.sourceforge.net |
| 2989 | S: Orphan | 2990 | S: Orphan |
| 2990 | F: Documentation/networking/decnet.txt | 2991 | F: Documentation/networking/decnet.txt |
| 2991 | F: net/decnet/ | 2992 | F: net/decnet/ |
| 2992 | 2993 | ||
| 2993 | DECSTATION PLATFORM SUPPORT | 2994 | DECSTATION PLATFORM SUPPORT |
| 2994 | M: "Maciej W. Rozycki" <macro@linux-mips.org> | 2995 | M: "Maciej W. Rozycki" <macro@linux-mips.org> |
| 2995 | L: linux-mips@linux-mips.org | 2996 | L: linux-mips@linux-mips.org |
| 2996 | W: http://www.linux-mips.org/wiki/DECstation | 2997 | W: http://www.linux-mips.org/wiki/DECstation |
| 2997 | S: Maintained | 2998 | S: Maintained |
| 2998 | F: arch/mips/dec/ | 2999 | F: arch/mips/dec/ |
| 2999 | F: arch/mips/include/asm/dec/ | 3000 | F: arch/mips/include/asm/dec/ |
| 3000 | F: arch/mips/include/asm/mach-dec/ | 3001 | F: arch/mips/include/asm/mach-dec/ |
| 3001 | 3002 | ||
| 3002 | DEFXX FDDI NETWORK DRIVER | 3003 | DEFXX FDDI NETWORK DRIVER |
| 3003 | M: "Maciej W. Rozycki" <macro@linux-mips.org> | 3004 | M: "Maciej W. Rozycki" <macro@linux-mips.org> |
| 3004 | S: Maintained | 3005 | S: Maintained |
| 3005 | F: drivers/net/fddi/defxx.* | 3006 | F: drivers/net/fddi/defxx.* |
| 3006 | 3007 | ||
| 3007 | DELL LAPTOP DRIVER | 3008 | DELL LAPTOP DRIVER |
| 3008 | M: Matthew Garrett <mjg59@srcf.ucam.org> | 3009 | M: Matthew Garrett <mjg59@srcf.ucam.org> |
| 3009 | L: platform-driver-x86@vger.kernel.org | 3010 | L: platform-driver-x86@vger.kernel.org |
| 3010 | S: Maintained | 3011 | S: Maintained |
| 3011 | F: drivers/platform/x86/dell-laptop.c | 3012 | F: drivers/platform/x86/dell-laptop.c |
| 3012 | 3013 | ||
| 3013 | DELL LAPTOP SMM DRIVER | 3014 | DELL LAPTOP SMM DRIVER |
| 3014 | M: Guenter Roeck <linux@roeck-us.net> | 3015 | M: Guenter Roeck <linux@roeck-us.net> |
| 3015 | F: drivers/char/i8k.c | 3016 | F: drivers/char/i8k.c |
| 3016 | F: include/uapi/linux/i8k.h | 3017 | F: include/uapi/linux/i8k.h |
| 3017 | 3018 | ||
| 3018 | DELL SYSTEMS MANAGEMENT BASE DRIVER (dcdbas) | 3019 | DELL SYSTEMS MANAGEMENT BASE DRIVER (dcdbas) |
| 3019 | M: Doug Warzecha <Douglas_Warzecha@dell.com> | 3020 | M: Doug Warzecha <Douglas_Warzecha@dell.com> |
| 3020 | S: Maintained | 3021 | S: Maintained |
| 3021 | F: Documentation/dcdbas.txt | 3022 | F: Documentation/dcdbas.txt |
| 3022 | F: drivers/firmware/dcdbas.* | 3023 | F: drivers/firmware/dcdbas.* |
| 3023 | 3024 | ||
| 3024 | DELL WMI EXTRAS DRIVER | 3025 | DELL WMI EXTRAS DRIVER |
| 3025 | M: Matthew Garrett <mjg59@srcf.ucam.org> | 3026 | M: Matthew Garrett <mjg59@srcf.ucam.org> |
| 3026 | S: Maintained | 3027 | S: Maintained |
| 3027 | F: drivers/platform/x86/dell-wmi.c | 3028 | F: drivers/platform/x86/dell-wmi.c |
| 3028 | 3029 | ||
| 3029 | DESIGNWARE USB2 DRD IP DRIVER | 3030 | DESIGNWARE USB2 DRD IP DRIVER |
| 3030 | M: Paul Zimmerman <paulz@synopsys.com> | 3031 | M: Paul Zimmerman <paulz@synopsys.com> |
| 3031 | L: linux-usb@vger.kernel.org | 3032 | L: linux-usb@vger.kernel.org |
| 3032 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git | 3033 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git |
| 3033 | S: Maintained | 3034 | S: Maintained |
| 3034 | F: drivers/usb/dwc2/ | 3035 | F: drivers/usb/dwc2/ |
| 3035 | 3036 | ||
| 3036 | DESIGNWARE USB3 DRD IP DRIVER | 3037 | DESIGNWARE USB3 DRD IP DRIVER |
| 3037 | M: Felipe Balbi <balbi@ti.com> | 3038 | M: Felipe Balbi <balbi@ti.com> |
| 3038 | L: linux-usb@vger.kernel.org | 3039 | L: linux-usb@vger.kernel.org |
| 3039 | L: linux-omap@vger.kernel.org | 3040 | L: linux-omap@vger.kernel.org |
| 3040 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git | 3041 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git |
| 3041 | S: Maintained | 3042 | S: Maintained |
| 3042 | F: drivers/usb/dwc3/ | 3043 | F: drivers/usb/dwc3/ |
| 3043 | 3044 | ||
| 3044 | DEVICE COREDUMP (DEV_COREDUMP) | 3045 | DEVICE COREDUMP (DEV_COREDUMP) |
| 3045 | M: Johannes Berg <johannes@sipsolutions.net> | 3046 | M: Johannes Berg <johannes@sipsolutions.net> |
| 3046 | L: linux-kernel@vger.kernel.org | 3047 | L: linux-kernel@vger.kernel.org |
| 3047 | S: Maintained | 3048 | S: Maintained |
| 3048 | F: drivers/base/devcoredump.c | 3049 | F: drivers/base/devcoredump.c |
| 3049 | F: include/linux/devcoredump.h | 3050 | F: include/linux/devcoredump.h |
| 3050 | 3051 | ||
| 3051 | DEVICE FREQUENCY (DEVFREQ) | 3052 | DEVICE FREQUENCY (DEVFREQ) |
| 3052 | M: MyungJoo Ham <myungjoo.ham@samsung.com> | 3053 | M: MyungJoo Ham <myungjoo.ham@samsung.com> |
| 3053 | M: Kyungmin Park <kyungmin.park@samsung.com> | 3054 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 3054 | L: linux-pm@vger.kernel.org | 3055 | L: linux-pm@vger.kernel.org |
| 3055 | S: Maintained | 3056 | S: Maintained |
| 3056 | F: drivers/devfreq/ | 3057 | F: drivers/devfreq/ |
| 3057 | 3058 | ||
| 3058 | DEVICE NUMBER REGISTRY | 3059 | DEVICE NUMBER REGISTRY |
| 3059 | M: Torben Mathiasen <device@lanana.org> | 3060 | M: Torben Mathiasen <device@lanana.org> |
| 3060 | W: http://lanana.org/docs/device-list/index.html | 3061 | W: http://lanana.org/docs/device-list/index.html |
| 3061 | S: Maintained | 3062 | S: Maintained |
| 3062 | 3063 | ||
| 3063 | DEVICE-MAPPER (LVM) | 3064 | DEVICE-MAPPER (LVM) |
| 3064 | M: Alasdair Kergon <agk@redhat.com> | 3065 | M: Alasdair Kergon <agk@redhat.com> |
| 3065 | M: Mike Snitzer <snitzer@redhat.com> | 3066 | M: Mike Snitzer <snitzer@redhat.com> |
| 3066 | M: dm-devel@redhat.com | 3067 | M: dm-devel@redhat.com |
| 3067 | L: dm-devel@redhat.com | 3068 | L: dm-devel@redhat.com |
| 3068 | W: http://sources.redhat.com/dm | 3069 | W: http://sources.redhat.com/dm |
| 3069 | Q: http://patchwork.kernel.org/project/dm-devel/list/ | 3070 | Q: http://patchwork.kernel.org/project/dm-devel/list/ |
| 3070 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git | 3071 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git |
| 3071 | T: quilt http://people.redhat.com/agk/patches/linux/editing/ | 3072 | T: quilt http://people.redhat.com/agk/patches/linux/editing/ |
| 3072 | S: Maintained | 3073 | S: Maintained |
| 3073 | F: Documentation/device-mapper/ | 3074 | F: Documentation/device-mapper/ |
| 3074 | F: drivers/md/dm* | 3075 | F: drivers/md/dm* |
| 3075 | F: drivers/md/persistent-data/ | 3076 | F: drivers/md/persistent-data/ |
| 3076 | F: include/linux/device-mapper.h | 3077 | F: include/linux/device-mapper.h |
| 3077 | F: include/linux/dm-*.h | 3078 | F: include/linux/dm-*.h |
| 3078 | F: include/uapi/linux/dm-*.h | 3079 | F: include/uapi/linux/dm-*.h |
| 3079 | 3080 | ||
| 3080 | DIALOG SEMICONDUCTOR DRIVERS | 3081 | DIALOG SEMICONDUCTOR DRIVERS |
| 3081 | M: Support Opensource <support.opensource@diasemi.com> | 3082 | M: Support Opensource <support.opensource@diasemi.com> |
| 3082 | W: http://www.dialog-semiconductor.com/products | 3083 | W: http://www.dialog-semiconductor.com/products |
| 3083 | S: Supported | 3084 | S: Supported |
| 3084 | F: Documentation/hwmon/da90?? | 3085 | F: Documentation/hwmon/da90?? |
| 3085 | F: drivers/gpio/gpio-da90??.c | 3086 | F: drivers/gpio/gpio-da90??.c |
| 3086 | F: drivers/hwmon/da90??-hwmon.c | 3087 | F: drivers/hwmon/da90??-hwmon.c |
| 3087 | F: drivers/input/misc/da90??_onkey.c | 3088 | F: drivers/input/misc/da90??_onkey.c |
| 3088 | F: drivers/input/touchscreen/da9052_tsi.c | 3089 | F: drivers/input/touchscreen/da9052_tsi.c |
| 3089 | F: drivers/leds/leds-da90??.c | 3090 | F: drivers/leds/leds-da90??.c |
| 3090 | F: drivers/mfd/da903x.c | 3091 | F: drivers/mfd/da903x.c |
| 3091 | F: drivers/mfd/da90??-*.c | 3092 | F: drivers/mfd/da90??-*.c |
| 3092 | F: drivers/power/da9052-battery.c | 3093 | F: drivers/power/da9052-battery.c |
| 3093 | F: drivers/regulator/da903x.c | 3094 | F: drivers/regulator/da903x.c |
| 3094 | F: drivers/regulator/da9???-regulator.[ch] | 3095 | F: drivers/regulator/da9???-regulator.[ch] |
| 3095 | F: drivers/rtc/rtc-da90??.c | 3096 | F: drivers/rtc/rtc-da90??.c |
| 3096 | F: drivers/video/backlight/da90??_bl.c | 3097 | F: drivers/video/backlight/da90??_bl.c |
| 3097 | F: drivers/watchdog/da90??_wdt.c | 3098 | F: drivers/watchdog/da90??_wdt.c |
| 3098 | F: include/linux/mfd/da903x.h | 3099 | F: include/linux/mfd/da903x.h |
| 3099 | F: include/linux/mfd/da9052/ | 3100 | F: include/linux/mfd/da9052/ |
| 3100 | F: include/linux/mfd/da9055/ | 3101 | F: include/linux/mfd/da9055/ |
| 3101 | F: include/linux/mfd/da9063/ | 3102 | F: include/linux/mfd/da9063/ |
| 3102 | F: include/sound/da[79]*.h | 3103 | F: include/sound/da[79]*.h |
| 3103 | F: sound/soc/codecs/da[79]*.[ch] | 3104 | F: sound/soc/codecs/da[79]*.[ch] |
| 3104 | 3105 | ||
| 3105 | DIGI NEO AND CLASSIC PCI PRODUCTS | 3106 | DIGI NEO AND CLASSIC PCI PRODUCTS |
| 3106 | M: Lidza Louina <lidza.louina@gmail.com> | 3107 | M: Lidza Louina <lidza.louina@gmail.com> |
| 3107 | M: Mark Hounschell <markh@compro.net> | 3108 | M: Mark Hounschell <markh@compro.net> |
| 3108 | L: driverdev-devel@linuxdriverproject.org | 3109 | L: driverdev-devel@linuxdriverproject.org |
| 3109 | S: Maintained | 3110 | S: Maintained |
| 3110 | F: drivers/staging/dgnc/ | 3111 | F: drivers/staging/dgnc/ |
| 3111 | 3112 | ||
| 3112 | DIGI EPCA PCI PRODUCTS | 3113 | DIGI EPCA PCI PRODUCTS |
| 3113 | M: Lidza Louina <lidza.louina@gmail.com> | 3114 | M: Lidza Louina <lidza.louina@gmail.com> |
| 3114 | M: Mark Hounschell <markh@compro.net> | 3115 | M: Mark Hounschell <markh@compro.net> |
| 3115 | M: Daeseok Youn <daeseok.youn@gmail.com> | 3116 | M: Daeseok Youn <daeseok.youn@gmail.com> |
| 3116 | L: driverdev-devel@linuxdriverproject.org | 3117 | L: driverdev-devel@linuxdriverproject.org |
| 3117 | S: Maintained | 3118 | S: Maintained |
| 3118 | F: drivers/staging/dgap/ | 3119 | F: drivers/staging/dgap/ |
| 3119 | 3120 | ||
| 3120 | DIOLAN U2C-12 I2C DRIVER | 3121 | DIOLAN U2C-12 I2C DRIVER |
| 3121 | M: Guenter Roeck <linux@roeck-us.net> | 3122 | M: Guenter Roeck <linux@roeck-us.net> |
| 3122 | L: linux-i2c@vger.kernel.org | 3123 | L: linux-i2c@vger.kernel.org |
| 3123 | S: Maintained | 3124 | S: Maintained |
| 3124 | F: drivers/i2c/busses/i2c-diolan-u2c.c | 3125 | F: drivers/i2c/busses/i2c-diolan-u2c.c |
| 3125 | 3126 | ||
| 3126 | DIRECTORY NOTIFICATION (DNOTIFY) | 3127 | DIRECTORY NOTIFICATION (DNOTIFY) |
| 3127 | M: Eric Paris <eparis@parisplace.org> | 3128 | M: Eric Paris <eparis@parisplace.org> |
| 3128 | S: Maintained | 3129 | S: Maintained |
| 3129 | F: Documentation/filesystems/dnotify.txt | 3130 | F: Documentation/filesystems/dnotify.txt |
| 3130 | F: fs/notify/dnotify/ | 3131 | F: fs/notify/dnotify/ |
| 3131 | F: include/linux/dnotify.h | 3132 | F: include/linux/dnotify.h |
| 3132 | 3133 | ||
| 3133 | DISK GEOMETRY AND PARTITION HANDLING | 3134 | DISK GEOMETRY AND PARTITION HANDLING |
| 3134 | M: Andries Brouwer <aeb@cwi.nl> | 3135 | M: Andries Brouwer <aeb@cwi.nl> |
| 3135 | W: http://www.win.tue.nl/~aeb/linux/Large-Disk.html | 3136 | W: http://www.win.tue.nl/~aeb/linux/Large-Disk.html |
| 3136 | W: http://www.win.tue.nl/~aeb/linux/zip/zip-1.html | 3137 | W: http://www.win.tue.nl/~aeb/linux/zip/zip-1.html |
| 3137 | W: http://www.win.tue.nl/~aeb/partitions/partition_types-1.html | 3138 | W: http://www.win.tue.nl/~aeb/partitions/partition_types-1.html |
| 3138 | S: Maintained | 3139 | S: Maintained |
| 3139 | 3140 | ||
| 3140 | DISKQUOTA | 3141 | DISKQUOTA |
| 3141 | M: Jan Kara <jack@suse.cz> | 3142 | M: Jan Kara <jack@suse.cz> |
| 3142 | S: Maintained | 3143 | S: Maintained |
| 3143 | F: Documentation/filesystems/quota.txt | 3144 | F: Documentation/filesystems/quota.txt |
| 3144 | F: fs/quota/ | 3145 | F: fs/quota/ |
| 3145 | F: include/linux/quota*.h | 3146 | F: include/linux/quota*.h |
| 3146 | F: include/uapi/linux/quota*.h | 3147 | F: include/uapi/linux/quota*.h |
| 3147 | 3148 | ||
| 3148 | DISPLAYLINK USB 2.0 FRAMEBUFFER DRIVER (UDLFB) | 3149 | DISPLAYLINK USB 2.0 FRAMEBUFFER DRIVER (UDLFB) |
| 3149 | M: Bernie Thompson <bernie@plugable.com> | 3150 | M: Bernie Thompson <bernie@plugable.com> |
| 3150 | L: linux-fbdev@vger.kernel.org | 3151 | L: linux-fbdev@vger.kernel.org |
| 3151 | S: Maintained | 3152 | S: Maintained |
| 3152 | W: http://plugable.com/category/projects/udlfb/ | 3153 | W: http://plugable.com/category/projects/udlfb/ |
| 3153 | F: drivers/video/fbdev/udlfb.c | 3154 | F: drivers/video/fbdev/udlfb.c |
| 3154 | F: include/video/udlfb.h | 3155 | F: include/video/udlfb.h |
| 3155 | F: Documentation/fb/udlfb.txt | 3156 | F: Documentation/fb/udlfb.txt |
| 3156 | 3157 | ||
| 3157 | DISTRIBUTED LOCK MANAGER (DLM) | 3158 | DISTRIBUTED LOCK MANAGER (DLM) |
| 3158 | M: Christine Caulfield <ccaulfie@redhat.com> | 3159 | M: Christine Caulfield <ccaulfie@redhat.com> |
| 3159 | M: David Teigland <teigland@redhat.com> | 3160 | M: David Teigland <teigland@redhat.com> |
| 3160 | L: cluster-devel@redhat.com | 3161 | L: cluster-devel@redhat.com |
| 3161 | W: http://sources.redhat.com/cluster/ | 3162 | W: http://sources.redhat.com/cluster/ |
| 3162 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm.git | 3163 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm.git |
| 3163 | S: Supported | 3164 | S: Supported |
| 3164 | F: fs/dlm/ | 3165 | F: fs/dlm/ |
| 3165 | 3166 | ||
| 3166 | DMA BUFFER SHARING FRAMEWORK | 3167 | DMA BUFFER SHARING FRAMEWORK |
| 3167 | M: Sumit Semwal <sumit.semwal@linaro.org> | 3168 | M: Sumit Semwal <sumit.semwal@linaro.org> |
| 3168 | S: Maintained | 3169 | S: Maintained |
| 3169 | L: linux-media@vger.kernel.org | 3170 | L: linux-media@vger.kernel.org |
| 3170 | L: dri-devel@lists.freedesktop.org | 3171 | L: dri-devel@lists.freedesktop.org |
| 3171 | L: linaro-mm-sig@lists.linaro.org (moderated for non-subscribers) | 3172 | L: linaro-mm-sig@lists.linaro.org (moderated for non-subscribers) |
| 3172 | F: drivers/dma-buf/ | 3173 | F: drivers/dma-buf/ |
| 3173 | F: include/linux/dma-buf* | 3174 | F: include/linux/dma-buf* |
| 3174 | F: include/linux/reservation.h | 3175 | F: include/linux/reservation.h |
| 3175 | F: include/linux/*fence.h | 3176 | F: include/linux/*fence.h |
| 3176 | F: Documentation/dma-buf-sharing.txt | 3177 | F: Documentation/dma-buf-sharing.txt |
| 3177 | T: git git://git.linaro.org/people/sumitsemwal/linux-dma-buf.git | 3178 | T: git git://git.linaro.org/people/sumitsemwal/linux-dma-buf.git |
| 3178 | 3179 | ||
| 3179 | DMA GENERIC OFFLOAD ENGINE SUBSYSTEM | 3180 | DMA GENERIC OFFLOAD ENGINE SUBSYSTEM |
| 3180 | M: Vinod Koul <vinod.koul@intel.com> | 3181 | M: Vinod Koul <vinod.koul@intel.com> |
| 3181 | L: dmaengine@vger.kernel.org | 3182 | L: dmaengine@vger.kernel.org |
| 3182 | Q: https://patchwork.kernel.org/project/linux-dmaengine/list/ | 3183 | Q: https://patchwork.kernel.org/project/linux-dmaengine/list/ |
| 3183 | S: Maintained | 3184 | S: Maintained |
| 3184 | F: drivers/dma/ | 3185 | F: drivers/dma/ |
| 3185 | F: include/linux/dma* | 3186 | F: include/linux/dma* |
| 3186 | F: Documentation/dmaengine/ | 3187 | F: Documentation/dmaengine/ |
| 3187 | T: git git://git.infradead.org/users/vkoul/slave-dma.git | 3188 | T: git git://git.infradead.org/users/vkoul/slave-dma.git |
| 3188 | 3189 | ||
| 3189 | DME1737 HARDWARE MONITOR DRIVER | 3190 | DME1737 HARDWARE MONITOR DRIVER |
| 3190 | M: Juerg Haefliger <juergh@gmail.com> | 3191 | M: Juerg Haefliger <juergh@gmail.com> |
| 3191 | L: lm-sensors@lm-sensors.org | 3192 | L: lm-sensors@lm-sensors.org |
| 3192 | S: Maintained | 3193 | S: Maintained |
| 3193 | F: Documentation/hwmon/dme1737 | 3194 | F: Documentation/hwmon/dme1737 |
| 3194 | F: drivers/hwmon/dme1737.c | 3195 | F: drivers/hwmon/dme1737.c |
| 3195 | 3196 | ||
| 3196 | DOCKING STATION DRIVER | 3197 | DOCKING STATION DRIVER |
| 3197 | M: Shaohua Li <shaohua.li@intel.com> | 3198 | M: Shaohua Li <shaohua.li@intel.com> |
| 3198 | L: linux-acpi@vger.kernel.org | 3199 | L: linux-acpi@vger.kernel.org |
| 3199 | S: Supported | 3200 | S: Supported |
| 3200 | F: drivers/acpi/dock.c | 3201 | F: drivers/acpi/dock.c |
| 3201 | 3202 | ||
| 3202 | DOCUMENTATION | 3203 | DOCUMENTATION |
| 3203 | M: Jonathan Corbet <corbet@lwn.net> | 3204 | M: Jonathan Corbet <corbet@lwn.net> |
| 3204 | L: linux-doc@vger.kernel.org | 3205 | L: linux-doc@vger.kernel.org |
| 3205 | S: Maintained | 3206 | S: Maintained |
| 3206 | F: Documentation/ | 3207 | F: Documentation/ |
| 3207 | X: Documentation/ABI/ | 3208 | X: Documentation/ABI/ |
| 3208 | X: Documentation/devicetree/ | 3209 | X: Documentation/devicetree/ |
| 3209 | X: Documentation/[a-z][a-z]_[A-Z][A-Z]/ | 3210 | X: Documentation/[a-z][a-z]_[A-Z][A-Z]/ |
| 3210 | 3211 | ||
| 3211 | DOUBLETALK DRIVER | 3212 | DOUBLETALK DRIVER |
| 3212 | M: "James R. Van Zandt" <jrv@vanzandt.mv.com> | 3213 | M: "James R. Van Zandt" <jrv@vanzandt.mv.com> |
| 3213 | L: blinux-list@redhat.com | 3214 | L: blinux-list@redhat.com |
| 3214 | S: Maintained | 3215 | S: Maintained |
| 3215 | F: drivers/char/dtlk.c | 3216 | F: drivers/char/dtlk.c |
| 3216 | F: include/linux/dtlk.h | 3217 | F: include/linux/dtlk.h |
| 3217 | 3218 | ||
| 3218 | DPT_I2O SCSI RAID DRIVER | 3219 | DPT_I2O SCSI RAID DRIVER |
| 3219 | M: Adaptec OEM Raid Solutions <aacraid@adaptec.com> | 3220 | M: Adaptec OEM Raid Solutions <aacraid@adaptec.com> |
| 3220 | L: linux-scsi@vger.kernel.org | 3221 | L: linux-scsi@vger.kernel.org |
| 3221 | W: http://www.adaptec.com/ | 3222 | W: http://www.adaptec.com/ |
| 3222 | S: Maintained | 3223 | S: Maintained |
| 3223 | F: drivers/scsi/dpt* | 3224 | F: drivers/scsi/dpt* |
| 3224 | F: drivers/scsi/dpt/ | 3225 | F: drivers/scsi/dpt/ |
| 3225 | 3226 | ||
| 3226 | DRBD DRIVER | 3227 | DRBD DRIVER |
| 3227 | P: Philipp Reisner | 3228 | P: Philipp Reisner |
| 3228 | P: Lars Ellenberg | 3229 | P: Lars Ellenberg |
| 3229 | M: drbd-dev@lists.linbit.com | 3230 | M: drbd-dev@lists.linbit.com |
| 3230 | L: drbd-user@lists.linbit.com | 3231 | L: drbd-user@lists.linbit.com |
| 3231 | W: http://www.drbd.org | 3232 | W: http://www.drbd.org |
| 3232 | T: git git://git.drbd.org/linux-2.6-drbd.git drbd | 3233 | T: git git://git.drbd.org/linux-2.6-drbd.git drbd |
| 3233 | T: git git://git.drbd.org/drbd-8.3.git | 3234 | T: git git://git.drbd.org/drbd-8.3.git |
| 3234 | S: Supported | 3235 | S: Supported |
| 3235 | F: drivers/block/drbd/ | 3236 | F: drivers/block/drbd/ |
| 3236 | F: lib/lru_cache.c | 3237 | F: lib/lru_cache.c |
| 3237 | F: Documentation/blockdev/drbd/ | 3238 | F: Documentation/blockdev/drbd/ |
| 3238 | 3239 | ||
| 3239 | DRIVER CORE, KOBJECTS, DEBUGFS AND SYSFS | 3240 | DRIVER CORE, KOBJECTS, DEBUGFS AND SYSFS |
| 3240 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 3241 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 3241 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git | 3242 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git |
| 3242 | S: Supported | 3243 | S: Supported |
| 3243 | F: Documentation/kobject.txt | 3244 | F: Documentation/kobject.txt |
| 3244 | F: drivers/base/ | 3245 | F: drivers/base/ |
| 3245 | F: fs/sysfs/ | 3246 | F: fs/sysfs/ |
| 3246 | F: fs/debugfs/ | 3247 | F: fs/debugfs/ |
| 3247 | F: include/linux/kobj* | 3248 | F: include/linux/kobj* |
| 3248 | F: include/linux/debugfs.h | 3249 | F: include/linux/debugfs.h |
| 3249 | F: lib/kobj* | 3250 | F: lib/kobj* |
| 3250 | 3251 | ||
| 3251 | DRM DRIVERS | 3252 | DRM DRIVERS |
| 3252 | M: David Airlie <airlied@linux.ie> | 3253 | M: David Airlie <airlied@linux.ie> |
| 3253 | L: dri-devel@lists.freedesktop.org | 3254 | L: dri-devel@lists.freedesktop.org |
| 3254 | T: git git://people.freedesktop.org/~airlied/linux | 3255 | T: git git://people.freedesktop.org/~airlied/linux |
| 3255 | S: Maintained | 3256 | S: Maintained |
| 3256 | F: drivers/gpu/drm/ | 3257 | F: drivers/gpu/drm/ |
| 3257 | F: drivers/gpu/vga/ | 3258 | F: drivers/gpu/vga/ |
| 3258 | F: include/drm/ | 3259 | F: include/drm/ |
| 3259 | F: include/uapi/drm/ | 3260 | F: include/uapi/drm/ |
| 3260 | 3261 | ||
| 3261 | RADEON DRM DRIVERS | 3262 | RADEON DRM DRIVERS |
| 3262 | M: Alex Deucher <alexander.deucher@amd.com> | 3263 | M: Alex Deucher <alexander.deucher@amd.com> |
| 3263 | M: Christian Kรถnig <christian.koenig@amd.com> | 3264 | M: Christian Kรถnig <christian.koenig@amd.com> |
| 3264 | L: dri-devel@lists.freedesktop.org | 3265 | L: dri-devel@lists.freedesktop.org |
| 3265 | T: git git://people.freedesktop.org/~agd5f/linux | 3266 | T: git git://people.freedesktop.org/~agd5f/linux |
| 3266 | S: Supported | 3267 | S: Supported |
| 3267 | F: drivers/gpu/drm/radeon/ | 3268 | F: drivers/gpu/drm/radeon/ |
| 3268 | F: include/uapi/drm/radeon* | 3269 | F: include/uapi/drm/radeon* |
| 3269 | 3270 | ||
| 3270 | DRM PANEL DRIVERS | 3271 | DRM PANEL DRIVERS |
| 3271 | M: Thierry Reding <thierry.reding@gmail.com> | 3272 | M: Thierry Reding <thierry.reding@gmail.com> |
| 3272 | L: dri-devel@lists.freedesktop.org | 3273 | L: dri-devel@lists.freedesktop.org |
| 3273 | T: git git://anongit.freedesktop.org/tegra/linux.git | 3274 | T: git git://anongit.freedesktop.org/tegra/linux.git |
| 3274 | S: Maintained | 3275 | S: Maintained |
| 3275 | F: drivers/gpu/drm/drm_panel.c | 3276 | F: drivers/gpu/drm/drm_panel.c |
| 3276 | F: drivers/gpu/drm/panel/ | 3277 | F: drivers/gpu/drm/panel/ |
| 3277 | F: include/drm/drm_panel.h | 3278 | F: include/drm/drm_panel.h |
| 3278 | F: Documentation/devicetree/bindings/panel/ | 3279 | F: Documentation/devicetree/bindings/panel/ |
| 3279 | 3280 | ||
| 3280 | INTEL DRM DRIVERS (excluding Poulsbo, Moorestown and derivative chipsets) | 3281 | INTEL DRM DRIVERS (excluding Poulsbo, Moorestown and derivative chipsets) |
| 3281 | M: Daniel Vetter <daniel.vetter@intel.com> | 3282 | M: Daniel Vetter <daniel.vetter@intel.com> |
| 3282 | M: Jani Nikula <jani.nikula@linux.intel.com> | 3283 | M: Jani Nikula <jani.nikula@linux.intel.com> |
| 3283 | L: intel-gfx@lists.freedesktop.org | 3284 | L: intel-gfx@lists.freedesktop.org |
| 3284 | L: dri-devel@lists.freedesktop.org | 3285 | L: dri-devel@lists.freedesktop.org |
| 3285 | Q: http://patchwork.freedesktop.org/project/intel-gfx/ | 3286 | Q: http://patchwork.freedesktop.org/project/intel-gfx/ |
| 3286 | T: git git://anongit.freedesktop.org/drm-intel | 3287 | T: git git://anongit.freedesktop.org/drm-intel |
| 3287 | S: Supported | 3288 | S: Supported |
| 3288 | F: drivers/gpu/drm/i915/ | 3289 | F: drivers/gpu/drm/i915/ |
| 3289 | F: include/drm/i915* | 3290 | F: include/drm/i915* |
| 3290 | F: include/uapi/drm/i915* | 3291 | F: include/uapi/drm/i915* |
| 3291 | 3292 | ||
| 3292 | DRM DRIVERS FOR EXYNOS | 3293 | DRM DRIVERS FOR EXYNOS |
| 3293 | M: Inki Dae <inki.dae@samsung.com> | 3294 | M: Inki Dae <inki.dae@samsung.com> |
| 3294 | M: Joonyoung Shim <jy0922.shim@samsung.com> | 3295 | M: Joonyoung Shim <jy0922.shim@samsung.com> |
| 3295 | M: Seung-Woo Kim <sw0312.kim@samsung.com> | 3296 | M: Seung-Woo Kim <sw0312.kim@samsung.com> |
| 3296 | M: Kyungmin Park <kyungmin.park@samsung.com> | 3297 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 3297 | L: dri-devel@lists.freedesktop.org | 3298 | L: dri-devel@lists.freedesktop.org |
| 3298 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos.git | 3299 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos.git |
| 3299 | S: Supported | 3300 | S: Supported |
| 3300 | F: drivers/gpu/drm/exynos/ | 3301 | F: drivers/gpu/drm/exynos/ |
| 3301 | F: include/drm/exynos* | 3302 | F: include/drm/exynos* |
| 3302 | F: include/uapi/drm/exynos* | 3303 | F: include/uapi/drm/exynos* |
| 3303 | 3304 | ||
| 3304 | DRM DRIVERS FOR FREESCALE IMX | 3305 | DRM DRIVERS FOR FREESCALE IMX |
| 3305 | M: Philipp Zabel <p.zabel@pengutronix.de> | 3306 | M: Philipp Zabel <p.zabel@pengutronix.de> |
| 3306 | L: dri-devel@lists.freedesktop.org | 3307 | L: dri-devel@lists.freedesktop.org |
| 3307 | S: Maintained | 3308 | S: Maintained |
| 3308 | F: drivers/gpu/drm/imx/ | 3309 | F: drivers/gpu/drm/imx/ |
| 3309 | F: Documentation/devicetree/bindings/drm/imx/ | 3310 | F: Documentation/devicetree/bindings/drm/imx/ |
| 3310 | 3311 | ||
| 3311 | DRM DRIVERS FOR NVIDIA TEGRA | 3312 | DRM DRIVERS FOR NVIDIA TEGRA |
| 3312 | M: Thierry Reding <thierry.reding@gmail.com> | 3313 | M: Thierry Reding <thierry.reding@gmail.com> |
| 3313 | M: Terje Bergstrรถm <tbergstrom@nvidia.com> | 3314 | M: Terje Bergstrรถm <tbergstrom@nvidia.com> |
| 3314 | L: dri-devel@lists.freedesktop.org | 3315 | L: dri-devel@lists.freedesktop.org |
| 3315 | L: linux-tegra@vger.kernel.org | 3316 | L: linux-tegra@vger.kernel.org |
| 3316 | T: git git://anongit.freedesktop.org/tegra/linux.git | 3317 | T: git git://anongit.freedesktop.org/tegra/linux.git |
| 3317 | S: Supported | 3318 | S: Supported |
| 3318 | F: drivers/gpu/drm/tegra/ | 3319 | F: drivers/gpu/drm/tegra/ |
| 3319 | F: drivers/gpu/host1x/ | 3320 | F: drivers/gpu/host1x/ |
| 3320 | F: include/linux/host1x.h | 3321 | F: include/linux/host1x.h |
| 3321 | F: include/uapi/drm/tegra_drm.h | 3322 | F: include/uapi/drm/tegra_drm.h |
| 3322 | F: Documentation/devicetree/bindings/gpu/nvidia,tegra20-host1x.txt | 3323 | F: Documentation/devicetree/bindings/gpu/nvidia,tegra20-host1x.txt |
| 3323 | 3324 | ||
| 3324 | DRM DRIVERS FOR RENESAS | 3325 | DRM DRIVERS FOR RENESAS |
| 3325 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 3326 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 3326 | L: dri-devel@lists.freedesktop.org | 3327 | L: dri-devel@lists.freedesktop.org |
| 3327 | L: linux-sh@vger.kernel.org | 3328 | L: linux-sh@vger.kernel.org |
| 3328 | T: git git://people.freedesktop.org/~airlied/linux | 3329 | T: git git://people.freedesktop.org/~airlied/linux |
| 3329 | S: Supported | 3330 | S: Supported |
| 3330 | F: drivers/gpu/drm/rcar-du/ | 3331 | F: drivers/gpu/drm/rcar-du/ |
| 3331 | F: drivers/gpu/drm/shmobile/ | 3332 | F: drivers/gpu/drm/shmobile/ |
| 3332 | F: include/linux/platform_data/rcar-du.h | 3333 | F: include/linux/platform_data/rcar-du.h |
| 3333 | F: include/linux/platform_data/shmob_drm.h | 3334 | F: include/linux/platform_data/shmob_drm.h |
| 3334 | 3335 | ||
| 3335 | DSBR100 USB FM RADIO DRIVER | 3336 | DSBR100 USB FM RADIO DRIVER |
| 3336 | M: Alexey Klimov <klimov.linux@gmail.com> | 3337 | M: Alexey Klimov <klimov.linux@gmail.com> |
| 3337 | L: linux-media@vger.kernel.org | 3338 | L: linux-media@vger.kernel.org |
| 3338 | T: git git://linuxtv.org/media_tree.git | 3339 | T: git git://linuxtv.org/media_tree.git |
| 3339 | S: Maintained | 3340 | S: Maintained |
| 3340 | F: drivers/media/radio/dsbr100.c | 3341 | F: drivers/media/radio/dsbr100.c |
| 3341 | 3342 | ||
| 3342 | DSCC4 DRIVER | 3343 | DSCC4 DRIVER |
| 3343 | M: Francois Romieu <romieu@fr.zoreil.com> | 3344 | M: Francois Romieu <romieu@fr.zoreil.com> |
| 3344 | L: netdev@vger.kernel.org | 3345 | L: netdev@vger.kernel.org |
| 3345 | S: Maintained | 3346 | S: Maintained |
| 3346 | F: drivers/net/wan/dscc4.c | 3347 | F: drivers/net/wan/dscc4.c |
| 3347 | 3348 | ||
| 3348 | DVB_USB_AF9015 MEDIA DRIVER | 3349 | DVB_USB_AF9015 MEDIA DRIVER |
| 3349 | M: Antti Palosaari <crope@iki.fi> | 3350 | M: Antti Palosaari <crope@iki.fi> |
| 3350 | L: linux-media@vger.kernel.org | 3351 | L: linux-media@vger.kernel.org |
| 3351 | W: http://linuxtv.org/ | 3352 | W: http://linuxtv.org/ |
| 3352 | W: http://palosaari.fi/linux/ | 3353 | W: http://palosaari.fi/linux/ |
| 3353 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3354 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3354 | T: git git://linuxtv.org/anttip/media_tree.git | 3355 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3355 | S: Maintained | 3356 | S: Maintained |
| 3356 | F: drivers/media/usb/dvb-usb-v2/af9015* | 3357 | F: drivers/media/usb/dvb-usb-v2/af9015* |
| 3357 | 3358 | ||
| 3358 | DVB_USB_AF9035 MEDIA DRIVER | 3359 | DVB_USB_AF9035 MEDIA DRIVER |
| 3359 | M: Antti Palosaari <crope@iki.fi> | 3360 | M: Antti Palosaari <crope@iki.fi> |
| 3360 | L: linux-media@vger.kernel.org | 3361 | L: linux-media@vger.kernel.org |
| 3361 | W: http://linuxtv.org/ | 3362 | W: http://linuxtv.org/ |
| 3362 | W: http://palosaari.fi/linux/ | 3363 | W: http://palosaari.fi/linux/ |
| 3363 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3364 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3364 | T: git git://linuxtv.org/anttip/media_tree.git | 3365 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3365 | S: Maintained | 3366 | S: Maintained |
| 3366 | F: drivers/media/usb/dvb-usb-v2/af9035* | 3367 | F: drivers/media/usb/dvb-usb-v2/af9035* |
| 3367 | 3368 | ||
| 3368 | DVB_USB_ANYSEE MEDIA DRIVER | 3369 | DVB_USB_ANYSEE MEDIA DRIVER |
| 3369 | M: Antti Palosaari <crope@iki.fi> | 3370 | M: Antti Palosaari <crope@iki.fi> |
| 3370 | L: linux-media@vger.kernel.org | 3371 | L: linux-media@vger.kernel.org |
| 3371 | W: http://linuxtv.org/ | 3372 | W: http://linuxtv.org/ |
| 3372 | W: http://palosaari.fi/linux/ | 3373 | W: http://palosaari.fi/linux/ |
| 3373 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3374 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3374 | T: git git://linuxtv.org/anttip/media_tree.git | 3375 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3375 | S: Maintained | 3376 | S: Maintained |
| 3376 | F: drivers/media/usb/dvb-usb-v2/anysee* | 3377 | F: drivers/media/usb/dvb-usb-v2/anysee* |
| 3377 | 3378 | ||
| 3378 | DVB_USB_AU6610 MEDIA DRIVER | 3379 | DVB_USB_AU6610 MEDIA DRIVER |
| 3379 | M: Antti Palosaari <crope@iki.fi> | 3380 | M: Antti Palosaari <crope@iki.fi> |
| 3380 | L: linux-media@vger.kernel.org | 3381 | L: linux-media@vger.kernel.org |
| 3381 | W: http://linuxtv.org/ | 3382 | W: http://linuxtv.org/ |
| 3382 | W: http://palosaari.fi/linux/ | 3383 | W: http://palosaari.fi/linux/ |
| 3383 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3384 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3384 | T: git git://linuxtv.org/anttip/media_tree.git | 3385 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3385 | S: Maintained | 3386 | S: Maintained |
| 3386 | F: drivers/media/usb/dvb-usb-v2/au6610* | 3387 | F: drivers/media/usb/dvb-usb-v2/au6610* |
| 3387 | 3388 | ||
| 3388 | DVB_USB_CE6230 MEDIA DRIVER | 3389 | DVB_USB_CE6230 MEDIA DRIVER |
| 3389 | M: Antti Palosaari <crope@iki.fi> | 3390 | M: Antti Palosaari <crope@iki.fi> |
| 3390 | L: linux-media@vger.kernel.org | 3391 | L: linux-media@vger.kernel.org |
| 3391 | W: http://linuxtv.org/ | 3392 | W: http://linuxtv.org/ |
| 3392 | W: http://palosaari.fi/linux/ | 3393 | W: http://palosaari.fi/linux/ |
| 3393 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3394 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3394 | T: git git://linuxtv.org/anttip/media_tree.git | 3395 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3395 | S: Maintained | 3396 | S: Maintained |
| 3396 | F: drivers/media/usb/dvb-usb-v2/ce6230* | 3397 | F: drivers/media/usb/dvb-usb-v2/ce6230* |
| 3397 | 3398 | ||
| 3398 | DVB_USB_CXUSB MEDIA DRIVER | 3399 | DVB_USB_CXUSB MEDIA DRIVER |
| 3399 | M: Michael Krufky <mkrufky@linuxtv.org> | 3400 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 3400 | L: linux-media@vger.kernel.org | 3401 | L: linux-media@vger.kernel.org |
| 3401 | W: http://linuxtv.org/ | 3402 | W: http://linuxtv.org/ |
| 3402 | W: http://github.com/mkrufky | 3403 | W: http://github.com/mkrufky |
| 3403 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3404 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3404 | T: git git://linuxtv.org/media_tree.git | 3405 | T: git git://linuxtv.org/media_tree.git |
| 3405 | S: Maintained | 3406 | S: Maintained |
| 3406 | F: drivers/media/usb/dvb-usb/cxusb* | 3407 | F: drivers/media/usb/dvb-usb/cxusb* |
| 3407 | 3408 | ||
| 3408 | DVB_USB_EC168 MEDIA DRIVER | 3409 | DVB_USB_EC168 MEDIA DRIVER |
| 3409 | M: Antti Palosaari <crope@iki.fi> | 3410 | M: Antti Palosaari <crope@iki.fi> |
| 3410 | L: linux-media@vger.kernel.org | 3411 | L: linux-media@vger.kernel.org |
| 3411 | W: http://linuxtv.org/ | 3412 | W: http://linuxtv.org/ |
| 3412 | W: http://palosaari.fi/linux/ | 3413 | W: http://palosaari.fi/linux/ |
| 3413 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3414 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3414 | T: git git://linuxtv.org/anttip/media_tree.git | 3415 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3415 | S: Maintained | 3416 | S: Maintained |
| 3416 | F: drivers/media/usb/dvb-usb-v2/ec168* | 3417 | F: drivers/media/usb/dvb-usb-v2/ec168* |
| 3417 | 3418 | ||
| 3418 | DVB_USB_GL861 MEDIA DRIVER | 3419 | DVB_USB_GL861 MEDIA DRIVER |
| 3419 | M: Antti Palosaari <crope@iki.fi> | 3420 | M: Antti Palosaari <crope@iki.fi> |
| 3420 | L: linux-media@vger.kernel.org | 3421 | L: linux-media@vger.kernel.org |
| 3421 | W: http://linuxtv.org/ | 3422 | W: http://linuxtv.org/ |
| 3422 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3423 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3423 | T: git git://linuxtv.org/anttip/media_tree.git | 3424 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3424 | S: Maintained | 3425 | S: Maintained |
| 3425 | F: drivers/media/usb/dvb-usb-v2/gl861* | 3426 | F: drivers/media/usb/dvb-usb-v2/gl861* |
| 3426 | 3427 | ||
| 3427 | DVB_USB_MXL111SF MEDIA DRIVER | 3428 | DVB_USB_MXL111SF MEDIA DRIVER |
| 3428 | M: Michael Krufky <mkrufky@linuxtv.org> | 3429 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 3429 | L: linux-media@vger.kernel.org | 3430 | L: linux-media@vger.kernel.org |
| 3430 | W: http://linuxtv.org/ | 3431 | W: http://linuxtv.org/ |
| 3431 | W: http://github.com/mkrufky | 3432 | W: http://github.com/mkrufky |
| 3432 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3433 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3433 | T: git git://linuxtv.org/mkrufky/mxl111sf.git | 3434 | T: git git://linuxtv.org/mkrufky/mxl111sf.git |
| 3434 | S: Maintained | 3435 | S: Maintained |
| 3435 | F: drivers/media/usb/dvb-usb-v2/mxl111sf* | 3436 | F: drivers/media/usb/dvb-usb-v2/mxl111sf* |
| 3436 | 3437 | ||
| 3437 | DVB_USB_RTL28XXU MEDIA DRIVER | 3438 | DVB_USB_RTL28XXU MEDIA DRIVER |
| 3438 | M: Antti Palosaari <crope@iki.fi> | 3439 | M: Antti Palosaari <crope@iki.fi> |
| 3439 | L: linux-media@vger.kernel.org | 3440 | L: linux-media@vger.kernel.org |
| 3440 | W: http://linuxtv.org/ | 3441 | W: http://linuxtv.org/ |
| 3441 | W: http://palosaari.fi/linux/ | 3442 | W: http://palosaari.fi/linux/ |
| 3442 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3443 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3443 | T: git git://linuxtv.org/anttip/media_tree.git | 3444 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3444 | S: Maintained | 3445 | S: Maintained |
| 3445 | F: drivers/media/usb/dvb-usb-v2/rtl28xxu* | 3446 | F: drivers/media/usb/dvb-usb-v2/rtl28xxu* |
| 3446 | 3447 | ||
| 3447 | DVB_USB_V2 MEDIA DRIVER | 3448 | DVB_USB_V2 MEDIA DRIVER |
| 3448 | M: Antti Palosaari <crope@iki.fi> | 3449 | M: Antti Palosaari <crope@iki.fi> |
| 3449 | L: linux-media@vger.kernel.org | 3450 | L: linux-media@vger.kernel.org |
| 3450 | W: http://linuxtv.org/ | 3451 | W: http://linuxtv.org/ |
| 3451 | W: http://palosaari.fi/linux/ | 3452 | W: http://palosaari.fi/linux/ |
| 3452 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3453 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3453 | T: git git://linuxtv.org/anttip/media_tree.git | 3454 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3454 | S: Maintained | 3455 | S: Maintained |
| 3455 | F: drivers/media/usb/dvb-usb-v2/dvb_usb* | 3456 | F: drivers/media/usb/dvb-usb-v2/dvb_usb* |
| 3456 | F: drivers/media/usb/dvb-usb-v2/usb_urb.c | 3457 | F: drivers/media/usb/dvb-usb-v2/usb_urb.c |
| 3457 | 3458 | ||
| 3458 | DYNAMIC DEBUG | 3459 | DYNAMIC DEBUG |
| 3459 | M: Jason Baron <jbaron@akamai.com> | 3460 | M: Jason Baron <jbaron@akamai.com> |
| 3460 | S: Maintained | 3461 | S: Maintained |
| 3461 | F: lib/dynamic_debug.c | 3462 | F: lib/dynamic_debug.c |
| 3462 | F: include/linux/dynamic_debug.h | 3463 | F: include/linux/dynamic_debug.h |
| 3463 | 3464 | ||
| 3464 | DZ DECSTATION DZ11 SERIAL DRIVER | 3465 | DZ DECSTATION DZ11 SERIAL DRIVER |
| 3465 | M: "Maciej W. Rozycki" <macro@linux-mips.org> | 3466 | M: "Maciej W. Rozycki" <macro@linux-mips.org> |
| 3466 | S: Maintained | 3467 | S: Maintained |
| 3467 | F: drivers/tty/serial/dz.* | 3468 | F: drivers/tty/serial/dz.* |
| 3468 | 3469 | ||
| 3469 | E4000 MEDIA DRIVER | 3470 | E4000 MEDIA DRIVER |
| 3470 | M: Antti Palosaari <crope@iki.fi> | 3471 | M: Antti Palosaari <crope@iki.fi> |
| 3471 | L: linux-media@vger.kernel.org | 3472 | L: linux-media@vger.kernel.org |
| 3472 | W: http://linuxtv.org/ | 3473 | W: http://linuxtv.org/ |
| 3473 | W: http://palosaari.fi/linux/ | 3474 | W: http://palosaari.fi/linux/ |
| 3474 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3475 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3475 | T: git git://linuxtv.org/anttip/media_tree.git | 3476 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3476 | S: Maintained | 3477 | S: Maintained |
| 3477 | F: drivers/media/tuners/e4000* | 3478 | F: drivers/media/tuners/e4000* |
| 3478 | 3479 | ||
| 3479 | EATA ISA/EISA/PCI SCSI DRIVER | 3480 | EATA ISA/EISA/PCI SCSI DRIVER |
| 3480 | M: Dario Ballabio <ballabio_dario@emc.com> | 3481 | M: Dario Ballabio <ballabio_dario@emc.com> |
| 3481 | L: linux-scsi@vger.kernel.org | 3482 | L: linux-scsi@vger.kernel.org |
| 3482 | S: Maintained | 3483 | S: Maintained |
| 3483 | F: drivers/scsi/eata.c | 3484 | F: drivers/scsi/eata.c |
| 3484 | 3485 | ||
| 3485 | EC100 MEDIA DRIVER | 3486 | EC100 MEDIA DRIVER |
| 3486 | M: Antti Palosaari <crope@iki.fi> | 3487 | M: Antti Palosaari <crope@iki.fi> |
| 3487 | L: linux-media@vger.kernel.org | 3488 | L: linux-media@vger.kernel.org |
| 3488 | W: http://linuxtv.org/ | 3489 | W: http://linuxtv.org/ |
| 3489 | W: http://palosaari.fi/linux/ | 3490 | W: http://palosaari.fi/linux/ |
| 3490 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3491 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3491 | T: git git://linuxtv.org/anttip/media_tree.git | 3492 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3492 | S: Maintained | 3493 | S: Maintained |
| 3493 | F: drivers/media/dvb-frontends/ec100* | 3494 | F: drivers/media/dvb-frontends/ec100* |
| 3494 | 3495 | ||
| 3495 | ECRYPT FILE SYSTEM | 3496 | ECRYPT FILE SYSTEM |
| 3496 | M: Tyler Hicks <tyhicks@canonical.com> | 3497 | M: Tyler Hicks <tyhicks@canonical.com> |
| 3497 | L: ecryptfs@vger.kernel.org | 3498 | L: ecryptfs@vger.kernel.org |
| 3498 | W: http://ecryptfs.org | 3499 | W: http://ecryptfs.org |
| 3499 | W: https://launchpad.net/ecryptfs | 3500 | W: https://launchpad.net/ecryptfs |
| 3500 | S: Supported | 3501 | S: Supported |
| 3501 | F: Documentation/filesystems/ecryptfs.txt | 3502 | F: Documentation/filesystems/ecryptfs.txt |
| 3502 | F: fs/ecryptfs/ | 3503 | F: fs/ecryptfs/ |
| 3503 | 3504 | ||
| 3504 | EDAC-CORE | 3505 | EDAC-CORE |
| 3505 | M: Doug Thompson <dougthompson@xmission.com> | 3506 | M: Doug Thompson <dougthompson@xmission.com> |
| 3506 | M: Borislav Petkov <bp@alien8.de> | 3507 | M: Borislav Petkov <bp@alien8.de> |
| 3507 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3508 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3508 | L: linux-edac@vger.kernel.org | 3509 | L: linux-edac@vger.kernel.org |
| 3509 | W: bluesmoke.sourceforge.net | 3510 | W: bluesmoke.sourceforge.net |
| 3510 | S: Supported | 3511 | S: Supported |
| 3511 | F: Documentation/edac.txt | 3512 | F: Documentation/edac.txt |
| 3512 | F: drivers/edac/ | 3513 | F: drivers/edac/ |
| 3513 | F: include/linux/edac.h | 3514 | F: include/linux/edac.h |
| 3514 | 3515 | ||
| 3515 | EDAC-AMD64 | 3516 | EDAC-AMD64 |
| 3516 | M: Doug Thompson <dougthompson@xmission.com> | 3517 | M: Doug Thompson <dougthompson@xmission.com> |
| 3517 | M: Borislav Petkov <bp@alien8.de> | 3518 | M: Borislav Petkov <bp@alien8.de> |
| 3518 | L: linux-edac@vger.kernel.org | 3519 | L: linux-edac@vger.kernel.org |
| 3519 | W: bluesmoke.sourceforge.net | 3520 | W: bluesmoke.sourceforge.net |
| 3520 | S: Maintained | 3521 | S: Maintained |
| 3521 | F: drivers/edac/amd64_edac* | 3522 | F: drivers/edac/amd64_edac* |
| 3522 | 3523 | ||
| 3523 | EDAC-CALXEDA | 3524 | EDAC-CALXEDA |
| 3524 | M: Doug Thompson <dougthompson@xmission.com> | 3525 | M: Doug Thompson <dougthompson@xmission.com> |
| 3525 | M: Robert Richter <rric@kernel.org> | 3526 | M: Robert Richter <rric@kernel.org> |
| 3526 | L: linux-edac@vger.kernel.org | 3527 | L: linux-edac@vger.kernel.org |
| 3527 | W: bluesmoke.sourceforge.net | 3528 | W: bluesmoke.sourceforge.net |
| 3528 | S: Maintained | 3529 | S: Maintained |
| 3529 | F: drivers/edac/highbank* | 3530 | F: drivers/edac/highbank* |
| 3530 | 3531 | ||
| 3531 | EDAC-CAVIUM | 3532 | EDAC-CAVIUM |
| 3532 | M: Ralf Baechle <ralf@linux-mips.org> | 3533 | M: Ralf Baechle <ralf@linux-mips.org> |
| 3533 | M: David Daney <david.daney@cavium.com> | 3534 | M: David Daney <david.daney@cavium.com> |
| 3534 | L: linux-edac@vger.kernel.org | 3535 | L: linux-edac@vger.kernel.org |
| 3535 | L: linux-mips@linux-mips.org | 3536 | L: linux-mips@linux-mips.org |
| 3536 | W: bluesmoke.sourceforge.net | 3537 | W: bluesmoke.sourceforge.net |
| 3537 | S: Supported | 3538 | S: Supported |
| 3538 | F: drivers/edac/octeon_edac* | 3539 | F: drivers/edac/octeon_edac* |
| 3539 | 3540 | ||
| 3540 | EDAC-E752X | 3541 | EDAC-E752X |
| 3541 | M: Mark Gross <mark.gross@intel.com> | 3542 | M: Mark Gross <mark.gross@intel.com> |
| 3542 | M: Doug Thompson <dougthompson@xmission.com> | 3543 | M: Doug Thompson <dougthompson@xmission.com> |
| 3543 | L: linux-edac@vger.kernel.org | 3544 | L: linux-edac@vger.kernel.org |
| 3544 | W: bluesmoke.sourceforge.net | 3545 | W: bluesmoke.sourceforge.net |
| 3545 | S: Maintained | 3546 | S: Maintained |
| 3546 | F: drivers/edac/e752x_edac.c | 3547 | F: drivers/edac/e752x_edac.c |
| 3547 | 3548 | ||
| 3548 | EDAC-E7XXX | 3549 | EDAC-E7XXX |
| 3549 | M: Doug Thompson <dougthompson@xmission.com> | 3550 | M: Doug Thompson <dougthompson@xmission.com> |
| 3550 | L: linux-edac@vger.kernel.org | 3551 | L: linux-edac@vger.kernel.org |
| 3551 | W: bluesmoke.sourceforge.net | 3552 | W: bluesmoke.sourceforge.net |
| 3552 | S: Maintained | 3553 | S: Maintained |
| 3553 | F: drivers/edac/e7xxx_edac.c | 3554 | F: drivers/edac/e7xxx_edac.c |
| 3554 | 3555 | ||
| 3555 | EDAC-GHES | 3556 | EDAC-GHES |
| 3556 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3557 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3557 | L: linux-edac@vger.kernel.org | 3558 | L: linux-edac@vger.kernel.org |
| 3558 | W: bluesmoke.sourceforge.net | 3559 | W: bluesmoke.sourceforge.net |
| 3559 | S: Maintained | 3560 | S: Maintained |
| 3560 | F: drivers/edac/ghes_edac.c | 3561 | F: drivers/edac/ghes_edac.c |
| 3561 | 3562 | ||
| 3562 | EDAC-I82443BXGX | 3563 | EDAC-I82443BXGX |
| 3563 | M: Tim Small <tim@buttersideup.com> | 3564 | M: Tim Small <tim@buttersideup.com> |
| 3564 | L: linux-edac@vger.kernel.org | 3565 | L: linux-edac@vger.kernel.org |
| 3565 | W: bluesmoke.sourceforge.net | 3566 | W: bluesmoke.sourceforge.net |
| 3566 | S: Maintained | 3567 | S: Maintained |
| 3567 | F: drivers/edac/i82443bxgx_edac.c | 3568 | F: drivers/edac/i82443bxgx_edac.c |
| 3568 | 3569 | ||
| 3569 | EDAC-I3000 | 3570 | EDAC-I3000 |
| 3570 | M: Jason Uhlenkott <juhlenko@akamai.com> | 3571 | M: Jason Uhlenkott <juhlenko@akamai.com> |
| 3571 | L: linux-edac@vger.kernel.org | 3572 | L: linux-edac@vger.kernel.org |
| 3572 | W: bluesmoke.sourceforge.net | 3573 | W: bluesmoke.sourceforge.net |
| 3573 | S: Maintained | 3574 | S: Maintained |
| 3574 | F: drivers/edac/i3000_edac.c | 3575 | F: drivers/edac/i3000_edac.c |
| 3575 | 3576 | ||
| 3576 | EDAC-I5000 | 3577 | EDAC-I5000 |
| 3577 | M: Doug Thompson <dougthompson@xmission.com> | 3578 | M: Doug Thompson <dougthompson@xmission.com> |
| 3578 | L: linux-edac@vger.kernel.org | 3579 | L: linux-edac@vger.kernel.org |
| 3579 | W: bluesmoke.sourceforge.net | 3580 | W: bluesmoke.sourceforge.net |
| 3580 | S: Maintained | 3581 | S: Maintained |
| 3581 | F: drivers/edac/i5000_edac.c | 3582 | F: drivers/edac/i5000_edac.c |
| 3582 | 3583 | ||
| 3583 | EDAC-I5400 | 3584 | EDAC-I5400 |
| 3584 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3585 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3585 | L: linux-edac@vger.kernel.org | 3586 | L: linux-edac@vger.kernel.org |
| 3586 | W: bluesmoke.sourceforge.net | 3587 | W: bluesmoke.sourceforge.net |
| 3587 | S: Maintained | 3588 | S: Maintained |
| 3588 | F: drivers/edac/i5400_edac.c | 3589 | F: drivers/edac/i5400_edac.c |
| 3589 | 3590 | ||
| 3590 | EDAC-I7300 | 3591 | EDAC-I7300 |
| 3591 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3592 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3592 | L: linux-edac@vger.kernel.org | 3593 | L: linux-edac@vger.kernel.org |
| 3593 | W: bluesmoke.sourceforge.net | 3594 | W: bluesmoke.sourceforge.net |
| 3594 | S: Maintained | 3595 | S: Maintained |
| 3595 | F: drivers/edac/i7300_edac.c | 3596 | F: drivers/edac/i7300_edac.c |
| 3596 | 3597 | ||
| 3597 | EDAC-I7CORE | 3598 | EDAC-I7CORE |
| 3598 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3599 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3599 | L: linux-edac@vger.kernel.org | 3600 | L: linux-edac@vger.kernel.org |
| 3600 | W: bluesmoke.sourceforge.net | 3601 | W: bluesmoke.sourceforge.net |
| 3601 | S: Maintained | 3602 | S: Maintained |
| 3602 | F: drivers/edac/i7core_edac.c | 3603 | F: drivers/edac/i7core_edac.c |
| 3603 | 3604 | ||
| 3604 | EDAC-I82975X | 3605 | EDAC-I82975X |
| 3605 | M: Ranganathan Desikan <ravi@jetztechnologies.com> | 3606 | M: Ranganathan Desikan <ravi@jetztechnologies.com> |
| 3606 | M: "Arvind R." <arvino55@gmail.com> | 3607 | M: "Arvind R." <arvino55@gmail.com> |
| 3607 | L: linux-edac@vger.kernel.org | 3608 | L: linux-edac@vger.kernel.org |
| 3608 | W: bluesmoke.sourceforge.net | 3609 | W: bluesmoke.sourceforge.net |
| 3609 | S: Maintained | 3610 | S: Maintained |
| 3610 | F: drivers/edac/i82975x_edac.c | 3611 | F: drivers/edac/i82975x_edac.c |
| 3611 | 3612 | ||
| 3612 | EDAC-IE31200 | 3613 | EDAC-IE31200 |
| 3613 | M: Jason Baron <jbaron@akamai.com> | 3614 | M: Jason Baron <jbaron@akamai.com> |
| 3614 | L: linux-edac@vger.kernel.org | 3615 | L: linux-edac@vger.kernel.org |
| 3615 | W: bluesmoke.sourceforge.net | 3616 | W: bluesmoke.sourceforge.net |
| 3616 | S: Maintained | 3617 | S: Maintained |
| 3617 | F: drivers/edac/ie31200_edac.c | 3618 | F: drivers/edac/ie31200_edac.c |
| 3618 | 3619 | ||
| 3619 | EDAC-MPC85XX | 3620 | EDAC-MPC85XX |
| 3620 | M: Johannes Thumshirn <johannes.thumshirn@men.de> | 3621 | M: Johannes Thumshirn <johannes.thumshirn@men.de> |
| 3621 | L: linux-edac@vger.kernel.org | 3622 | L: linux-edac@vger.kernel.org |
| 3622 | W: bluesmoke.sourceforge.net | 3623 | W: bluesmoke.sourceforge.net |
| 3623 | S: Maintained | 3624 | S: Maintained |
| 3624 | F: drivers/edac/mpc85xx_edac.[ch] | 3625 | F: drivers/edac/mpc85xx_edac.[ch] |
| 3625 | 3626 | ||
| 3626 | EDAC-PASEMI | 3627 | EDAC-PASEMI |
| 3627 | M: Egor Martovetsky <egor@pasemi.com> | 3628 | M: Egor Martovetsky <egor@pasemi.com> |
| 3628 | L: linux-edac@vger.kernel.org | 3629 | L: linux-edac@vger.kernel.org |
| 3629 | W: bluesmoke.sourceforge.net | 3630 | W: bluesmoke.sourceforge.net |
| 3630 | S: Maintained | 3631 | S: Maintained |
| 3631 | F: drivers/edac/pasemi_edac.c | 3632 | F: drivers/edac/pasemi_edac.c |
| 3632 | 3633 | ||
| 3633 | EDAC-R82600 | 3634 | EDAC-R82600 |
| 3634 | M: Tim Small <tim@buttersideup.com> | 3635 | M: Tim Small <tim@buttersideup.com> |
| 3635 | L: linux-edac@vger.kernel.org | 3636 | L: linux-edac@vger.kernel.org |
| 3636 | W: bluesmoke.sourceforge.net | 3637 | W: bluesmoke.sourceforge.net |
| 3637 | S: Maintained | 3638 | S: Maintained |
| 3638 | F: drivers/edac/r82600_edac.c | 3639 | F: drivers/edac/r82600_edac.c |
| 3639 | 3640 | ||
| 3640 | EDAC-SBRIDGE | 3641 | EDAC-SBRIDGE |
| 3641 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3642 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3642 | L: linux-edac@vger.kernel.org | 3643 | L: linux-edac@vger.kernel.org |
| 3643 | W: bluesmoke.sourceforge.net | 3644 | W: bluesmoke.sourceforge.net |
| 3644 | S: Maintained | 3645 | S: Maintained |
| 3645 | F: drivers/edac/sb_edac.c | 3646 | F: drivers/edac/sb_edac.c |
| 3646 | 3647 | ||
| 3647 | EDIROL UA-101/UA-1000 DRIVER | 3648 | EDIROL UA-101/UA-1000 DRIVER |
| 3648 | M: Clemens Ladisch <clemens@ladisch.de> | 3649 | M: Clemens Ladisch <clemens@ladisch.de> |
| 3649 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 3650 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 3650 | T: git git://git.alsa-project.org/alsa-kernel.git | 3651 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 3651 | S: Maintained | 3652 | S: Maintained |
| 3652 | F: sound/usb/misc/ua101.c | 3653 | F: sound/usb/misc/ua101.c |
| 3653 | 3654 | ||
| 3654 | EXTENSIBLE FIRMWARE INTERFACE (EFI) | 3655 | EXTENSIBLE FIRMWARE INTERFACE (EFI) |
| 3655 | M: Matt Fleming <matt.fleming@intel.com> | 3656 | M: Matt Fleming <matt.fleming@intel.com> |
| 3656 | L: linux-efi@vger.kernel.org | 3657 | L: linux-efi@vger.kernel.org |
| 3657 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi.git | 3658 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi.git |
| 3658 | S: Maintained | 3659 | S: Maintained |
| 3659 | F: Documentation/efi-stub.txt | 3660 | F: Documentation/efi-stub.txt |
| 3660 | F: arch/ia64/kernel/efi.c | 3661 | F: arch/ia64/kernel/efi.c |
| 3661 | F: arch/x86/boot/compressed/eboot.[ch] | 3662 | F: arch/x86/boot/compressed/eboot.[ch] |
| 3662 | F: arch/x86/include/asm/efi.h | 3663 | F: arch/x86/include/asm/efi.h |
| 3663 | F: arch/x86/platform/efi/* | 3664 | F: arch/x86/platform/efi/* |
| 3664 | F: drivers/firmware/efi/* | 3665 | F: drivers/firmware/efi/* |
| 3665 | F: include/linux/efi*.h | 3666 | F: include/linux/efi*.h |
| 3666 | 3667 | ||
| 3667 | EFI VARIABLE FILESYSTEM | 3668 | EFI VARIABLE FILESYSTEM |
| 3668 | M: Matthew Garrett <matthew.garrett@nebula.com> | 3669 | M: Matthew Garrett <matthew.garrett@nebula.com> |
| 3669 | M: Jeremy Kerr <jk@ozlabs.org> | 3670 | M: Jeremy Kerr <jk@ozlabs.org> |
| 3670 | M: Matt Fleming <matt.fleming@intel.com> | 3671 | M: Matt Fleming <matt.fleming@intel.com> |
| 3671 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi.git | 3672 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mfleming/efi.git |
| 3672 | L: linux-efi@vger.kernel.org | 3673 | L: linux-efi@vger.kernel.org |
| 3673 | S: Maintained | 3674 | S: Maintained |
| 3674 | F: fs/efivarfs/ | 3675 | F: fs/efivarfs/ |
| 3675 | 3676 | ||
| 3676 | EFIFB FRAMEBUFFER DRIVER | 3677 | EFIFB FRAMEBUFFER DRIVER |
| 3677 | L: linux-fbdev@vger.kernel.org | 3678 | L: linux-fbdev@vger.kernel.org |
| 3678 | M: Peter Jones <pjones@redhat.com> | 3679 | M: Peter Jones <pjones@redhat.com> |
| 3679 | S: Maintained | 3680 | S: Maintained |
| 3680 | F: drivers/video/fbdev/efifb.c | 3681 | F: drivers/video/fbdev/efifb.c |
| 3681 | 3682 | ||
| 3682 | EFS FILESYSTEM | 3683 | EFS FILESYSTEM |
| 3683 | W: http://aeschi.ch.eu.org/efs/ | 3684 | W: http://aeschi.ch.eu.org/efs/ |
| 3684 | S: Orphan | 3685 | S: Orphan |
| 3685 | F: fs/efs/ | 3686 | F: fs/efs/ |
| 3686 | 3687 | ||
| 3687 | EHCA (IBM GX bus InfiniBand adapter) DRIVER | 3688 | EHCA (IBM GX bus InfiniBand adapter) DRIVER |
| 3688 | M: Hoang-Nam Nguyen <hnguyen@de.ibm.com> | 3689 | M: Hoang-Nam Nguyen <hnguyen@de.ibm.com> |
| 3689 | M: Christoph Raisch <raisch@de.ibm.com> | 3690 | M: Christoph Raisch <raisch@de.ibm.com> |
| 3690 | L: linux-rdma@vger.kernel.org | 3691 | L: linux-rdma@vger.kernel.org |
| 3691 | S: Supported | 3692 | S: Supported |
| 3692 | F: drivers/infiniband/hw/ehca/ | 3693 | F: drivers/infiniband/hw/ehca/ |
| 3693 | 3694 | ||
| 3694 | EHEA (IBM pSeries eHEA 10Gb ethernet adapter) DRIVER | 3695 | EHEA (IBM pSeries eHEA 10Gb ethernet adapter) DRIVER |
| 3695 | M: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com> | 3696 | M: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com> |
| 3696 | L: netdev@vger.kernel.org | 3697 | L: netdev@vger.kernel.org |
| 3697 | S: Maintained | 3698 | S: Maintained |
| 3698 | F: drivers/net/ethernet/ibm/ehea/ | 3699 | F: drivers/net/ethernet/ibm/ehea/ |
| 3699 | 3700 | ||
| 3700 | EM28XX VIDEO4LINUX DRIVER | 3701 | EM28XX VIDEO4LINUX DRIVER |
| 3701 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 3702 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 3702 | L: linux-media@vger.kernel.org | 3703 | L: linux-media@vger.kernel.org |
| 3703 | W: http://linuxtv.org | 3704 | W: http://linuxtv.org |
| 3704 | T: git git://linuxtv.org/media_tree.git | 3705 | T: git git://linuxtv.org/media_tree.git |
| 3705 | S: Maintained | 3706 | S: Maintained |
| 3706 | F: drivers/media/usb/em28xx/ | 3707 | F: drivers/media/usb/em28xx/ |
| 3707 | 3708 | ||
| 3708 | EMBEDDED LINUX | 3709 | EMBEDDED LINUX |
| 3709 | M: Paul Gortmaker <paul.gortmaker@windriver.com> | 3710 | M: Paul Gortmaker <paul.gortmaker@windriver.com> |
| 3710 | M: Matt Mackall <mpm@selenic.com> | 3711 | M: Matt Mackall <mpm@selenic.com> |
| 3711 | M: David Woodhouse <dwmw2@infradead.org> | 3712 | M: David Woodhouse <dwmw2@infradead.org> |
| 3712 | L: linux-embedded@vger.kernel.org | 3713 | L: linux-embedded@vger.kernel.org |
| 3713 | S: Maintained | 3714 | S: Maintained |
| 3714 | 3715 | ||
| 3715 | EMULEX LPFC FC SCSI DRIVER | 3716 | EMULEX LPFC FC SCSI DRIVER |
| 3716 | M: James Smart <james.smart@emulex.com> | 3717 | M: James Smart <james.smart@emulex.com> |
| 3717 | L: linux-scsi@vger.kernel.org | 3718 | L: linux-scsi@vger.kernel.org |
| 3718 | W: http://sourceforge.net/projects/lpfcxxxx | 3719 | W: http://sourceforge.net/projects/lpfcxxxx |
| 3719 | S: Supported | 3720 | S: Supported |
| 3720 | F: drivers/scsi/lpfc/ | 3721 | F: drivers/scsi/lpfc/ |
| 3721 | 3722 | ||
| 3722 | ENE CB710 FLASH CARD READER DRIVER | 3723 | ENE CB710 FLASH CARD READER DRIVER |
| 3723 | M: Michaล Mirosลaw <mirq-linux@rere.qmqm.pl> | 3724 | M: Michaล Mirosลaw <mirq-linux@rere.qmqm.pl> |
| 3724 | S: Maintained | 3725 | S: Maintained |
| 3725 | F: drivers/misc/cb710/ | 3726 | F: drivers/misc/cb710/ |
| 3726 | F: drivers/mmc/host/cb710-mmc.* | 3727 | F: drivers/mmc/host/cb710-mmc.* |
| 3727 | F: include/linux/cb710.h | 3728 | F: include/linux/cb710.h |
| 3728 | 3729 | ||
| 3729 | ENE KB2426 (ENE0100/ENE020XX) INFRARED RECEIVER | 3730 | ENE KB2426 (ENE0100/ENE020XX) INFRARED RECEIVER |
| 3730 | M: Maxim Levitsky <maximlevitsky@gmail.com> | 3731 | M: Maxim Levitsky <maximlevitsky@gmail.com> |
| 3731 | S: Maintained | 3732 | S: Maintained |
| 3732 | F: drivers/media/rc/ene_ir.* | 3733 | F: drivers/media/rc/ene_ir.* |
| 3733 | 3734 | ||
| 3734 | ENHANCED ERROR HANDLING (EEH) | 3735 | ENHANCED ERROR HANDLING (EEH) |
| 3735 | M: Gavin Shan <shangw@linux.vnet.ibm.com> | 3736 | M: Gavin Shan <shangw@linux.vnet.ibm.com> |
| 3736 | L: linuxppc-dev@lists.ozlabs.org | 3737 | L: linuxppc-dev@lists.ozlabs.org |
| 3737 | S: Supported | 3738 | S: Supported |
| 3738 | F: Documentation/powerpc/eeh-pci-error-recovery.txt | 3739 | F: Documentation/powerpc/eeh-pci-error-recovery.txt |
| 3739 | F: arch/powerpc/kernel/eeh*.c | 3740 | F: arch/powerpc/kernel/eeh*.c |
| 3740 | 3741 | ||
| 3741 | EPSON S1D13XXX FRAMEBUFFER DRIVER | 3742 | EPSON S1D13XXX FRAMEBUFFER DRIVER |
| 3742 | M: Kristoffer Ericson <kristoffer.ericson@gmail.com> | 3743 | M: Kristoffer Ericson <kristoffer.ericson@gmail.com> |
| 3743 | S: Maintained | 3744 | S: Maintained |
| 3744 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kristoffer/linux-hpc.git | 3745 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kristoffer/linux-hpc.git |
| 3745 | F: drivers/video/fbdev/s1d13xxxfb.c | 3746 | F: drivers/video/fbdev/s1d13xxxfb.c |
| 3746 | F: include/video/s1d13xxxfb.h | 3747 | F: include/video/s1d13xxxfb.h |
| 3747 | 3748 | ||
| 3748 | ET131X NETWORK DRIVER | 3749 | ET131X NETWORK DRIVER |
| 3749 | M: Mark Einon <mark.einon@gmail.com> | 3750 | M: Mark Einon <mark.einon@gmail.com> |
| 3750 | S: Odd Fixes | 3751 | S: Odd Fixes |
| 3751 | F: drivers/net/ethernet/agere/ | 3752 | F: drivers/net/ethernet/agere/ |
| 3752 | 3753 | ||
| 3753 | ETHERNET BRIDGE | 3754 | ETHERNET BRIDGE |
| 3754 | M: Stephen Hemminger <stephen@networkplumber.org> | 3755 | M: Stephen Hemminger <stephen@networkplumber.org> |
| 3755 | L: bridge@lists.linux-foundation.org | 3756 | L: bridge@lists.linux-foundation.org |
| 3756 | L: netdev@vger.kernel.org | 3757 | L: netdev@vger.kernel.org |
| 3757 | W: http://www.linuxfoundation.org/en/Net:Bridge | 3758 | W: http://www.linuxfoundation.org/en/Net:Bridge |
| 3758 | S: Maintained | 3759 | S: Maintained |
| 3759 | F: include/linux/netfilter_bridge/ | 3760 | F: include/linux/netfilter_bridge/ |
| 3760 | F: net/bridge/ | 3761 | F: net/bridge/ |
| 3761 | 3762 | ||
| 3762 | ETHERNET PHY LIBRARY | 3763 | ETHERNET PHY LIBRARY |
| 3763 | M: Florian Fainelli <f.fainelli@gmail.com> | 3764 | M: Florian Fainelli <f.fainelli@gmail.com> |
| 3764 | L: netdev@vger.kernel.org | 3765 | L: netdev@vger.kernel.org |
| 3765 | S: Maintained | 3766 | S: Maintained |
| 3766 | F: include/linux/phy.h | 3767 | F: include/linux/phy.h |
| 3767 | F: include/linux/phy_fixed.h | 3768 | F: include/linux/phy_fixed.h |
| 3768 | F: drivers/net/phy/ | 3769 | F: drivers/net/phy/ |
| 3769 | F: Documentation/networking/phy.txt | 3770 | F: Documentation/networking/phy.txt |
| 3770 | F: drivers/of/of_mdio.c | 3771 | F: drivers/of/of_mdio.c |
| 3771 | F: drivers/of/of_net.c | 3772 | F: drivers/of/of_net.c |
| 3772 | 3773 | ||
| 3773 | EXT2 FILE SYSTEM | 3774 | EXT2 FILE SYSTEM |
| 3774 | M: Jan Kara <jack@suse.cz> | 3775 | M: Jan Kara <jack@suse.cz> |
| 3775 | L: linux-ext4@vger.kernel.org | 3776 | L: linux-ext4@vger.kernel.org |
| 3776 | S: Maintained | 3777 | S: Maintained |
| 3777 | F: Documentation/filesystems/ext2.txt | 3778 | F: Documentation/filesystems/ext2.txt |
| 3778 | F: fs/ext2/ | 3779 | F: fs/ext2/ |
| 3779 | F: include/linux/ext2* | 3780 | F: include/linux/ext2* |
| 3780 | 3781 | ||
| 3781 | EXT3 FILE SYSTEM | 3782 | EXT3 FILE SYSTEM |
| 3782 | M: Jan Kara <jack@suse.cz> | 3783 | M: Jan Kara <jack@suse.cz> |
| 3783 | M: Andrew Morton <akpm@linux-foundation.org> | 3784 | M: Andrew Morton <akpm@linux-foundation.org> |
| 3784 | M: Andreas Dilger <adilger.kernel@dilger.ca> | 3785 | M: Andreas Dilger <adilger.kernel@dilger.ca> |
| 3785 | L: linux-ext4@vger.kernel.org | 3786 | L: linux-ext4@vger.kernel.org |
| 3786 | S: Maintained | 3787 | S: Maintained |
| 3787 | F: Documentation/filesystems/ext3.txt | 3788 | F: Documentation/filesystems/ext3.txt |
| 3788 | F: fs/ext3/ | 3789 | F: fs/ext3/ |
| 3789 | 3790 | ||
| 3790 | EXT4 FILE SYSTEM | 3791 | EXT4 FILE SYSTEM |
| 3791 | M: "Theodore Ts'o" <tytso@mit.edu> | 3792 | M: "Theodore Ts'o" <tytso@mit.edu> |
| 3792 | M: Andreas Dilger <adilger.kernel@dilger.ca> | 3793 | M: Andreas Dilger <adilger.kernel@dilger.ca> |
| 3793 | L: linux-ext4@vger.kernel.org | 3794 | L: linux-ext4@vger.kernel.org |
| 3794 | W: http://ext4.wiki.kernel.org | 3795 | W: http://ext4.wiki.kernel.org |
| 3795 | Q: http://patchwork.ozlabs.org/project/linux-ext4/list/ | 3796 | Q: http://patchwork.ozlabs.org/project/linux-ext4/list/ |
| 3796 | S: Maintained | 3797 | S: Maintained |
| 3797 | F: Documentation/filesystems/ext4.txt | 3798 | F: Documentation/filesystems/ext4.txt |
| 3798 | F: fs/ext4/ | 3799 | F: fs/ext4/ |
| 3799 | 3800 | ||
| 3800 | Extended Verification Module (EVM) | 3801 | Extended Verification Module (EVM) |
| 3801 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> | 3802 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> |
| 3802 | L: linux-ima-devel@lists.sourceforge.net | 3803 | L: linux-ima-devel@lists.sourceforge.net |
| 3803 | L: linux-security-module@vger.kernel.org | 3804 | L: linux-security-module@vger.kernel.org |
| 3804 | S: Supported | 3805 | S: Supported |
| 3805 | F: security/integrity/evm/ | 3806 | F: security/integrity/evm/ |
| 3806 | 3807 | ||
| 3807 | EXTERNAL CONNECTOR SUBSYSTEM (EXTCON) | 3808 | EXTERNAL CONNECTOR SUBSYSTEM (EXTCON) |
| 3808 | M: MyungJoo Ham <myungjoo.ham@samsung.com> | 3809 | M: MyungJoo Ham <myungjoo.ham@samsung.com> |
| 3809 | M: Chanwoo Choi <cw00.choi@samsung.com> | 3810 | M: Chanwoo Choi <cw00.choi@samsung.com> |
| 3810 | L: linux-kernel@vger.kernel.org | 3811 | L: linux-kernel@vger.kernel.org |
| 3811 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/chanwoo/extcon.git | 3812 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/chanwoo/extcon.git |
| 3812 | S: Maintained | 3813 | S: Maintained |
| 3813 | F: drivers/extcon/ | 3814 | F: drivers/extcon/ |
| 3814 | F: Documentation/extcon/ | 3815 | F: Documentation/extcon/ |
| 3815 | 3816 | ||
| 3816 | EXYNOS DP DRIVER | 3817 | EXYNOS DP DRIVER |
| 3817 | M: Jingoo Han <jg1.han@samsung.com> | 3818 | M: Jingoo Han <jg1.han@samsung.com> |
| 3818 | L: dri-devel@lists.freedesktop.org | 3819 | L: dri-devel@lists.freedesktop.org |
| 3819 | S: Maintained | 3820 | S: Maintained |
| 3820 | F: drivers/gpu/drm/exynos/exynos_dp* | 3821 | F: drivers/gpu/drm/exynos/exynos_dp* |
| 3821 | 3822 | ||
| 3822 | EXYNOS MIPI DISPLAY DRIVERS | 3823 | EXYNOS MIPI DISPLAY DRIVERS |
| 3823 | M: Inki Dae <inki.dae@samsung.com> | 3824 | M: Inki Dae <inki.dae@samsung.com> |
| 3824 | M: Donghwa Lee <dh09.lee@samsung.com> | 3825 | M: Donghwa Lee <dh09.lee@samsung.com> |
| 3825 | M: Kyungmin Park <kyungmin.park@samsung.com> | 3826 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 3826 | L: linux-fbdev@vger.kernel.org | 3827 | L: linux-fbdev@vger.kernel.org |
| 3827 | S: Maintained | 3828 | S: Maintained |
| 3828 | F: drivers/video/fbdev/exynos/exynos_mipi* | 3829 | F: drivers/video/fbdev/exynos/exynos_mipi* |
| 3829 | F: include/video/exynos_mipi* | 3830 | F: include/video/exynos_mipi* |
| 3830 | 3831 | ||
| 3831 | F71805F HARDWARE MONITORING DRIVER | 3832 | F71805F HARDWARE MONITORING DRIVER |
| 3832 | M: Jean Delvare <jdelvare@suse.de> | 3833 | M: Jean Delvare <jdelvare@suse.de> |
| 3833 | L: lm-sensors@lm-sensors.org | 3834 | L: lm-sensors@lm-sensors.org |
| 3834 | S: Maintained | 3835 | S: Maintained |
| 3835 | F: Documentation/hwmon/f71805f | 3836 | F: Documentation/hwmon/f71805f |
| 3836 | F: drivers/hwmon/f71805f.c | 3837 | F: drivers/hwmon/f71805f.c |
| 3837 | 3838 | ||
| 3838 | FC0011 TUNER DRIVER | 3839 | FC0011 TUNER DRIVER |
| 3839 | M: Michael Buesch <m@bues.ch> | 3840 | M: Michael Buesch <m@bues.ch> |
| 3840 | L: linux-media@vger.kernel.org | 3841 | L: linux-media@vger.kernel.org |
| 3841 | S: Maintained | 3842 | S: Maintained |
| 3842 | F: drivers/media/tuners/fc0011.h | 3843 | F: drivers/media/tuners/fc0011.h |
| 3843 | F: drivers/media/tuners/fc0011.c | 3844 | F: drivers/media/tuners/fc0011.c |
| 3844 | 3845 | ||
| 3845 | FC2580 MEDIA DRIVER | 3846 | FC2580 MEDIA DRIVER |
| 3846 | M: Antti Palosaari <crope@iki.fi> | 3847 | M: Antti Palosaari <crope@iki.fi> |
| 3847 | L: linux-media@vger.kernel.org | 3848 | L: linux-media@vger.kernel.org |
| 3848 | W: http://linuxtv.org/ | 3849 | W: http://linuxtv.org/ |
| 3849 | W: http://palosaari.fi/linux/ | 3850 | W: http://palosaari.fi/linux/ |
| 3850 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 3851 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 3851 | T: git git://linuxtv.org/anttip/media_tree.git | 3852 | T: git git://linuxtv.org/anttip/media_tree.git |
| 3852 | S: Maintained | 3853 | S: Maintained |
| 3853 | F: drivers/media/tuners/fc2580* | 3854 | F: drivers/media/tuners/fc2580* |
| 3854 | 3855 | ||
| 3855 | FANOTIFY | 3856 | FANOTIFY |
| 3856 | M: Eric Paris <eparis@redhat.com> | 3857 | M: Eric Paris <eparis@redhat.com> |
| 3857 | S: Maintained | 3858 | S: Maintained |
| 3858 | F: fs/notify/fanotify/ | 3859 | F: fs/notify/fanotify/ |
| 3859 | F: include/linux/fanotify.h | 3860 | F: include/linux/fanotify.h |
| 3860 | F: include/uapi/linux/fanotify.h | 3861 | F: include/uapi/linux/fanotify.h |
| 3861 | 3862 | ||
| 3862 | FARSYNC SYNCHRONOUS DRIVER | 3863 | FARSYNC SYNCHRONOUS DRIVER |
| 3863 | M: Kevin Curtis <kevin.curtis@farsite.co.uk> | 3864 | M: Kevin Curtis <kevin.curtis@farsite.co.uk> |
| 3864 | W: http://www.farsite.co.uk/ | 3865 | W: http://www.farsite.co.uk/ |
| 3865 | S: Supported | 3866 | S: Supported |
| 3866 | F: drivers/net/wan/farsync.* | 3867 | F: drivers/net/wan/farsync.* |
| 3867 | 3868 | ||
| 3868 | FAULT INJECTION SUPPORT | 3869 | FAULT INJECTION SUPPORT |
| 3869 | M: Akinobu Mita <akinobu.mita@gmail.com> | 3870 | M: Akinobu Mita <akinobu.mita@gmail.com> |
| 3870 | S: Supported | 3871 | S: Supported |
| 3871 | F: Documentation/fault-injection/ | 3872 | F: Documentation/fault-injection/ |
| 3872 | F: lib/fault-inject.c | 3873 | F: lib/fault-inject.c |
| 3873 | 3874 | ||
| 3874 | FCOE SUBSYSTEM (libfc, libfcoe, fcoe) | 3875 | FCOE SUBSYSTEM (libfc, libfcoe, fcoe) |
| 3875 | M: Robert Love <robert.w.love@intel.com> | 3876 | M: Robert Love <robert.w.love@intel.com> |
| 3876 | L: fcoe-devel@open-fcoe.org | 3877 | L: fcoe-devel@open-fcoe.org |
| 3877 | W: www.Open-FCoE.org | 3878 | W: www.Open-FCoE.org |
| 3878 | S: Supported | 3879 | S: Supported |
| 3879 | F: drivers/scsi/libfc/ | 3880 | F: drivers/scsi/libfc/ |
| 3880 | F: drivers/scsi/fcoe/ | 3881 | F: drivers/scsi/fcoe/ |
| 3881 | F: include/scsi/fc/ | 3882 | F: include/scsi/fc/ |
| 3882 | F: include/scsi/libfc.h | 3883 | F: include/scsi/libfc.h |
| 3883 | F: include/scsi/libfcoe.h | 3884 | F: include/scsi/libfcoe.h |
| 3884 | F: include/uapi/scsi/fc/ | 3885 | F: include/uapi/scsi/fc/ |
| 3885 | 3886 | ||
| 3886 | FILE LOCKING (flock() and fcntl()/lockf()) | 3887 | FILE LOCKING (flock() and fcntl()/lockf()) |
| 3887 | M: Jeff Layton <jlayton@poochiereds.net> | 3888 | M: Jeff Layton <jlayton@poochiereds.net> |
| 3888 | M: J. Bruce Fields <bfields@fieldses.org> | 3889 | M: J. Bruce Fields <bfields@fieldses.org> |
| 3889 | L: linux-fsdevel@vger.kernel.org | 3890 | L: linux-fsdevel@vger.kernel.org |
| 3890 | S: Maintained | 3891 | S: Maintained |
| 3891 | F: include/linux/fcntl.h | 3892 | F: include/linux/fcntl.h |
| 3892 | F: include/linux/fs.h | 3893 | F: include/linux/fs.h |
| 3893 | F: include/uapi/linux/fcntl.h | 3894 | F: include/uapi/linux/fcntl.h |
| 3894 | F: include/uapi/linux/fs.h | 3895 | F: include/uapi/linux/fs.h |
| 3895 | F: fs/fcntl.c | 3896 | F: fs/fcntl.c |
| 3896 | F: fs/locks.c | 3897 | F: fs/locks.c |
| 3897 | 3898 | ||
| 3898 | FILESYSTEMS (VFS and infrastructure) | 3899 | FILESYSTEMS (VFS and infrastructure) |
| 3899 | M: Alexander Viro <viro@zeniv.linux.org.uk> | 3900 | M: Alexander Viro <viro@zeniv.linux.org.uk> |
| 3900 | L: linux-fsdevel@vger.kernel.org | 3901 | L: linux-fsdevel@vger.kernel.org |
| 3901 | S: Maintained | 3902 | S: Maintained |
| 3902 | F: fs/* | 3903 | F: fs/* |
| 3903 | 3904 | ||
| 3904 | FINTEK F75375S HARDWARE MONITOR AND FAN CONTROLLER DRIVER | 3905 | FINTEK F75375S HARDWARE MONITOR AND FAN CONTROLLER DRIVER |
| 3905 | M: Riku Voipio <riku.voipio@iki.fi> | 3906 | M: Riku Voipio <riku.voipio@iki.fi> |
| 3906 | L: lm-sensors@lm-sensors.org | 3907 | L: lm-sensors@lm-sensors.org |
| 3907 | S: Maintained | 3908 | S: Maintained |
| 3908 | F: drivers/hwmon/f75375s.c | 3909 | F: drivers/hwmon/f75375s.c |
| 3909 | F: include/linux/f75375s.h | 3910 | F: include/linux/f75375s.h |
| 3910 | 3911 | ||
| 3911 | FIREWIRE AUDIO DRIVERS | 3912 | FIREWIRE AUDIO DRIVERS |
| 3912 | M: Clemens Ladisch <clemens@ladisch.de> | 3913 | M: Clemens Ladisch <clemens@ladisch.de> |
| 3913 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 3914 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 3914 | T: git git://git.alsa-project.org/alsa-kernel.git | 3915 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 3915 | S: Maintained | 3916 | S: Maintained |
| 3916 | F: sound/firewire/ | 3917 | F: sound/firewire/ |
| 3917 | 3918 | ||
| 3918 | FIREWIRE MEDIA DRIVERS (firedtv) | 3919 | FIREWIRE MEDIA DRIVERS (firedtv) |
| 3919 | M: Stefan Richter <stefanr@s5r6.in-berlin.de> | 3920 | M: Stefan Richter <stefanr@s5r6.in-berlin.de> |
| 3920 | L: linux-media@vger.kernel.org | 3921 | L: linux-media@vger.kernel.org |
| 3921 | L: linux1394-devel@lists.sourceforge.net | 3922 | L: linux1394-devel@lists.sourceforge.net |
| 3922 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media.git | 3923 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media.git |
| 3923 | S: Maintained | 3924 | S: Maintained |
| 3924 | F: drivers/media/firewire/ | 3925 | F: drivers/media/firewire/ |
| 3925 | 3926 | ||
| 3926 | FIREWIRE SBP-2 TARGET | 3927 | FIREWIRE SBP-2 TARGET |
| 3927 | M: Chris Boot <bootc@bootc.net> | 3928 | M: Chris Boot <bootc@bootc.net> |
| 3928 | L: linux-scsi@vger.kernel.org | 3929 | L: linux-scsi@vger.kernel.org |
| 3929 | L: target-devel@vger.kernel.org | 3930 | L: target-devel@vger.kernel.org |
| 3930 | L: linux1394-devel@lists.sourceforge.net | 3931 | L: linux1394-devel@lists.sourceforge.net |
| 3931 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/nab/lio-core-2.6.git master | 3932 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/nab/lio-core-2.6.git master |
| 3932 | S: Maintained | 3933 | S: Maintained |
| 3933 | F: drivers/target/sbp/ | 3934 | F: drivers/target/sbp/ |
| 3934 | 3935 | ||
| 3935 | FIREWIRE SUBSYSTEM | 3936 | FIREWIRE SUBSYSTEM |
| 3936 | M: Stefan Richter <stefanr@s5r6.in-berlin.de> | 3937 | M: Stefan Richter <stefanr@s5r6.in-berlin.de> |
| 3937 | L: linux1394-devel@lists.sourceforge.net | 3938 | L: linux1394-devel@lists.sourceforge.net |
| 3938 | W: http://ieee1394.wiki.kernel.org/ | 3939 | W: http://ieee1394.wiki.kernel.org/ |
| 3939 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394.git | 3940 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394.git |
| 3940 | S: Maintained | 3941 | S: Maintained |
| 3941 | F: drivers/firewire/ | 3942 | F: drivers/firewire/ |
| 3942 | F: include/linux/firewire.h | 3943 | F: include/linux/firewire.h |
| 3943 | F: include/uapi/linux/firewire*.h | 3944 | F: include/uapi/linux/firewire*.h |
| 3944 | F: tools/firewire/ | 3945 | F: tools/firewire/ |
| 3945 | 3946 | ||
| 3946 | FIRMWARE LOADER (request_firmware) | 3947 | FIRMWARE LOADER (request_firmware) |
| 3947 | M: Ming Lei <ming.lei@canonical.com> | 3948 | M: Ming Lei <ming.lei@canonical.com> |
| 3948 | L: linux-kernel@vger.kernel.org | 3949 | L: linux-kernel@vger.kernel.org |
| 3949 | S: Maintained | 3950 | S: Maintained |
| 3950 | F: Documentation/firmware_class/ | 3951 | F: Documentation/firmware_class/ |
| 3951 | F: drivers/base/firmware*.c | 3952 | F: drivers/base/firmware*.c |
| 3952 | F: include/linux/firmware.h | 3953 | F: include/linux/firmware.h |
| 3953 | 3954 | ||
| 3954 | FLASH ADAPTER DRIVER (IBM Flash Adapter 900GB Full Height PCI Flash Card) | 3955 | FLASH ADAPTER DRIVER (IBM Flash Adapter 900GB Full Height PCI Flash Card) |
| 3955 | M: Joshua Morris <josh.h.morris@us.ibm.com> | 3956 | M: Joshua Morris <josh.h.morris@us.ibm.com> |
| 3956 | M: Philip Kelleher <pjk1939@linux.vnet.ibm.com> | 3957 | M: Philip Kelleher <pjk1939@linux.vnet.ibm.com> |
| 3957 | S: Maintained | 3958 | S: Maintained |
| 3958 | F: drivers/block/rsxx/ | 3959 | F: drivers/block/rsxx/ |
| 3959 | 3960 | ||
| 3960 | FLOPPY DRIVER | 3961 | FLOPPY DRIVER |
| 3961 | M: Jiri Kosina <jkosina@suse.cz> | 3962 | M: Jiri Kosina <jkosina@suse.cz> |
| 3962 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/floppy.git | 3963 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/floppy.git |
| 3963 | S: Odd fixes | 3964 | S: Odd fixes |
| 3964 | F: drivers/block/floppy.c | 3965 | F: drivers/block/floppy.c |
| 3965 | 3966 | ||
| 3966 | FMC SUBSYSTEM | 3967 | FMC SUBSYSTEM |
| 3967 | M: Alessandro Rubini <rubini@gnudd.com> | 3968 | M: Alessandro Rubini <rubini@gnudd.com> |
| 3968 | W: http://www.ohwr.org/projects/fmc-bus | 3969 | W: http://www.ohwr.org/projects/fmc-bus |
| 3969 | S: Supported | 3970 | S: Supported |
| 3970 | F: drivers/fmc/ | 3971 | F: drivers/fmc/ |
| 3971 | F: include/linux/fmc*.h | 3972 | F: include/linux/fmc*.h |
| 3972 | F: include/linux/ipmi-fru.h | 3973 | F: include/linux/ipmi-fru.h |
| 3973 | K: fmc_d.*register | 3974 | K: fmc_d.*register |
| 3974 | 3975 | ||
| 3975 | FPU EMULATOR | 3976 | FPU EMULATOR |
| 3976 | M: Bill Metzenthen <billm@melbpc.org.au> | 3977 | M: Bill Metzenthen <billm@melbpc.org.au> |
| 3977 | W: http://floatingpoint.sourceforge.net/emulator/index.html | 3978 | W: http://floatingpoint.sourceforge.net/emulator/index.html |
| 3978 | S: Maintained | 3979 | S: Maintained |
| 3979 | F: arch/x86/math-emu/ | 3980 | F: arch/x86/math-emu/ |
| 3980 | 3981 | ||
| 3981 | FRAME RELAY DLCI/FRAD (Sangoma drivers too) | 3982 | FRAME RELAY DLCI/FRAD (Sangoma drivers too) |
| 3982 | L: netdev@vger.kernel.org | 3983 | L: netdev@vger.kernel.org |
| 3983 | S: Orphan | 3984 | S: Orphan |
| 3984 | F: drivers/net/wan/dlci.c | 3985 | F: drivers/net/wan/dlci.c |
| 3985 | F: drivers/net/wan/sdla.c | 3986 | F: drivers/net/wan/sdla.c |
| 3986 | 3987 | ||
| 3987 | FRAMEBUFFER LAYER | 3988 | FRAMEBUFFER LAYER |
| 3988 | M: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com> | 3989 | M: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com> |
| 3989 | M: Tomi Valkeinen <tomi.valkeinen@ti.com> | 3990 | M: Tomi Valkeinen <tomi.valkeinen@ti.com> |
| 3990 | L: linux-fbdev@vger.kernel.org | 3991 | L: linux-fbdev@vger.kernel.org |
| 3991 | W: http://linux-fbdev.sourceforge.net/ | 3992 | W: http://linux-fbdev.sourceforge.net/ |
| 3992 | Q: http://patchwork.kernel.org/project/linux-fbdev/list/ | 3993 | Q: http://patchwork.kernel.org/project/linux-fbdev/list/ |
| 3993 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/plagnioj/linux-fbdev.git | 3994 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/plagnioj/linux-fbdev.git |
| 3994 | S: Maintained | 3995 | S: Maintained |
| 3995 | F: Documentation/fb/ | 3996 | F: Documentation/fb/ |
| 3996 | F: Documentation/devicetree/bindings/fb/ | 3997 | F: Documentation/devicetree/bindings/fb/ |
| 3997 | F: drivers/video/ | 3998 | F: drivers/video/ |
| 3998 | F: include/video/ | 3999 | F: include/video/ |
| 3999 | F: include/linux/fb.h | 4000 | F: include/linux/fb.h |
| 4000 | F: include/uapi/video/ | 4001 | F: include/uapi/video/ |
| 4001 | F: include/uapi/linux/fb.h | 4002 | F: include/uapi/linux/fb.h |
| 4002 | 4003 | ||
| 4003 | FREESCALE DIU FRAMEBUFFER DRIVER | 4004 | FREESCALE DIU FRAMEBUFFER DRIVER |
| 4004 | M: Timur Tabi <timur@tabi.org> | 4005 | M: Timur Tabi <timur@tabi.org> |
| 4005 | L: linux-fbdev@vger.kernel.org | 4006 | L: linux-fbdev@vger.kernel.org |
| 4006 | S: Maintained | 4007 | S: Maintained |
| 4007 | F: drivers/video/fbdev/fsl-diu-fb.* | 4008 | F: drivers/video/fbdev/fsl-diu-fb.* |
| 4008 | 4009 | ||
| 4009 | FREESCALE DMA DRIVER | 4010 | FREESCALE DMA DRIVER |
| 4010 | M: Li Yang <leoli@freescale.com> | 4011 | M: Li Yang <leoli@freescale.com> |
| 4011 | M: Zhang Wei <zw@zh-kernel.org> | 4012 | M: Zhang Wei <zw@zh-kernel.org> |
| 4012 | L: linuxppc-dev@lists.ozlabs.org | 4013 | L: linuxppc-dev@lists.ozlabs.org |
| 4013 | S: Maintained | 4014 | S: Maintained |
| 4014 | F: drivers/dma/fsldma.* | 4015 | F: drivers/dma/fsldma.* |
| 4015 | 4016 | ||
| 4016 | FREESCALE I2C CPM DRIVER | 4017 | FREESCALE I2C CPM DRIVER |
| 4017 | M: Jochen Friedrich <jochen@scram.de> | 4018 | M: Jochen Friedrich <jochen@scram.de> |
| 4018 | L: linuxppc-dev@lists.ozlabs.org | 4019 | L: linuxppc-dev@lists.ozlabs.org |
| 4019 | L: linux-i2c@vger.kernel.org | 4020 | L: linux-i2c@vger.kernel.org |
| 4020 | S: Maintained | 4021 | S: Maintained |
| 4021 | F: drivers/i2c/busses/i2c-cpm.c | 4022 | F: drivers/i2c/busses/i2c-cpm.c |
| 4022 | 4023 | ||
| 4023 | FREESCALE IMX / MXC FRAMEBUFFER DRIVER | 4024 | FREESCALE IMX / MXC FRAMEBUFFER DRIVER |
| 4024 | M: Sascha Hauer <kernel@pengutronix.de> | 4025 | M: Sascha Hauer <kernel@pengutronix.de> |
| 4025 | L: linux-fbdev@vger.kernel.org | 4026 | L: linux-fbdev@vger.kernel.org |
| 4026 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 4027 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 4027 | S: Maintained | 4028 | S: Maintained |
| 4028 | F: include/linux/platform_data/video-imxfb.h | 4029 | F: include/linux/platform_data/video-imxfb.h |
| 4029 | F: drivers/video/fbdev/imxfb.c | 4030 | F: drivers/video/fbdev/imxfb.c |
| 4030 | 4031 | ||
| 4031 | FREESCALE SOC FS_ENET DRIVER | 4032 | FREESCALE SOC FS_ENET DRIVER |
| 4032 | M: Pantelis Antoniou <pantelis.antoniou@gmail.com> | 4033 | M: Pantelis Antoniou <pantelis.antoniou@gmail.com> |
| 4033 | M: Vitaly Bordug <vbordug@ru.mvista.com> | 4034 | M: Vitaly Bordug <vbordug@ru.mvista.com> |
| 4034 | L: linuxppc-dev@lists.ozlabs.org | 4035 | L: linuxppc-dev@lists.ozlabs.org |
| 4035 | L: netdev@vger.kernel.org | 4036 | L: netdev@vger.kernel.org |
| 4036 | S: Maintained | 4037 | S: Maintained |
| 4037 | F: drivers/net/ethernet/freescale/fs_enet/ | 4038 | F: drivers/net/ethernet/freescale/fs_enet/ |
| 4038 | F: include/linux/fs_enet_pd.h | 4039 | F: include/linux/fs_enet_pd.h |
| 4039 | 4040 | ||
| 4040 | FREESCALE QUICC ENGINE LIBRARY | 4041 | FREESCALE QUICC ENGINE LIBRARY |
| 4041 | L: linuxppc-dev@lists.ozlabs.org | 4042 | L: linuxppc-dev@lists.ozlabs.org |
| 4042 | S: Orphan | 4043 | S: Orphan |
| 4043 | F: arch/powerpc/sysdev/qe_lib/ | 4044 | F: arch/powerpc/sysdev/qe_lib/ |
| 4044 | F: arch/powerpc/include/asm/*qe.h | 4045 | F: arch/powerpc/include/asm/*qe.h |
| 4045 | 4046 | ||
| 4046 | FREESCALE USB PERIPHERAL DRIVERS | 4047 | FREESCALE USB PERIPHERAL DRIVERS |
| 4047 | M: Li Yang <leoli@freescale.com> | 4048 | M: Li Yang <leoli@freescale.com> |
| 4048 | L: linux-usb@vger.kernel.org | 4049 | L: linux-usb@vger.kernel.org |
| 4049 | L: linuxppc-dev@lists.ozlabs.org | 4050 | L: linuxppc-dev@lists.ozlabs.org |
| 4050 | S: Maintained | 4051 | S: Maintained |
| 4051 | F: drivers/usb/gadget/udc/fsl* | 4052 | F: drivers/usb/gadget/udc/fsl* |
| 4052 | 4053 | ||
| 4053 | FREESCALE QUICC ENGINE UCC ETHERNET DRIVER | 4054 | FREESCALE QUICC ENGINE UCC ETHERNET DRIVER |
| 4054 | M: Li Yang <leoli@freescale.com> | 4055 | M: Li Yang <leoli@freescale.com> |
| 4055 | L: netdev@vger.kernel.org | 4056 | L: netdev@vger.kernel.org |
| 4056 | L: linuxppc-dev@lists.ozlabs.org | 4057 | L: linuxppc-dev@lists.ozlabs.org |
| 4057 | S: Maintained | 4058 | S: Maintained |
| 4058 | F: drivers/net/ethernet/freescale/ucc_geth* | 4059 | F: drivers/net/ethernet/freescale/ucc_geth* |
| 4059 | 4060 | ||
| 4060 | FREESCALE QUICC ENGINE UCC UART DRIVER | 4061 | FREESCALE QUICC ENGINE UCC UART DRIVER |
| 4061 | M: Timur Tabi <timur@tabi.org> | 4062 | M: Timur Tabi <timur@tabi.org> |
| 4062 | L: linuxppc-dev@lists.ozlabs.org | 4063 | L: linuxppc-dev@lists.ozlabs.org |
| 4063 | S: Maintained | 4064 | S: Maintained |
| 4064 | F: drivers/tty/serial/ucc_uart.c | 4065 | F: drivers/tty/serial/ucc_uart.c |
| 4065 | 4066 | ||
| 4066 | FREESCALE SOC SOUND DRIVERS | 4067 | FREESCALE SOC SOUND DRIVERS |
| 4067 | M: Timur Tabi <timur@tabi.org> | 4068 | M: Timur Tabi <timur@tabi.org> |
| 4068 | M: Nicolin Chen <nicoleotsuka@gmail.com> | 4069 | M: Nicolin Chen <nicoleotsuka@gmail.com> |
| 4069 | M: Xiubo Li <Xiubo.Lee@gmail.com> | 4070 | M: Xiubo Li <Xiubo.Lee@gmail.com> |
| 4070 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 4071 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 4071 | L: linuxppc-dev@lists.ozlabs.org | 4072 | L: linuxppc-dev@lists.ozlabs.org |
| 4072 | S: Maintained | 4073 | S: Maintained |
| 4073 | F: sound/soc/fsl/fsl* | 4074 | F: sound/soc/fsl/fsl* |
| 4074 | F: sound/soc/fsl/imx* | 4075 | F: sound/soc/fsl/imx* |
| 4075 | F: sound/soc/fsl/mpc8610_hpcd.c | 4076 | F: sound/soc/fsl/mpc8610_hpcd.c |
| 4076 | 4077 | ||
| 4077 | FREEVXFS FILESYSTEM | 4078 | FREEVXFS FILESYSTEM |
| 4078 | M: Christoph Hellwig <hch@infradead.org> | 4079 | M: Christoph Hellwig <hch@infradead.org> |
| 4079 | W: ftp://ftp.openlinux.org/pub/people/hch/vxfs | 4080 | W: ftp://ftp.openlinux.org/pub/people/hch/vxfs |
| 4080 | S: Maintained | 4081 | S: Maintained |
| 4081 | F: fs/freevxfs/ | 4082 | F: fs/freevxfs/ |
| 4082 | 4083 | ||
| 4083 | FREEZER | 4084 | FREEZER |
| 4084 | M: "Rafael J. Wysocki" <rjw@rjwysocki.net> | 4085 | M: "Rafael J. Wysocki" <rjw@rjwysocki.net> |
| 4085 | M: Pavel Machek <pavel@ucw.cz> | 4086 | M: Pavel Machek <pavel@ucw.cz> |
| 4086 | L: linux-pm@vger.kernel.org | 4087 | L: linux-pm@vger.kernel.org |
| 4087 | S: Supported | 4088 | S: Supported |
| 4088 | F: Documentation/power/freezing-of-tasks.txt | 4089 | F: Documentation/power/freezing-of-tasks.txt |
| 4089 | F: include/linux/freezer.h | 4090 | F: include/linux/freezer.h |
| 4090 | F: kernel/freezer.c | 4091 | F: kernel/freezer.c |
| 4091 | 4092 | ||
| 4092 | FRONTSWAP API | 4093 | FRONTSWAP API |
| 4093 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 4094 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 4094 | L: linux-kernel@vger.kernel.org | 4095 | L: linux-kernel@vger.kernel.org |
| 4095 | S: Maintained | 4096 | S: Maintained |
| 4096 | F: mm/frontswap.c | 4097 | F: mm/frontswap.c |
| 4097 | F: include/linux/frontswap.h | 4098 | F: include/linux/frontswap.h |
| 4098 | 4099 | ||
| 4099 | FS-CACHE: LOCAL CACHING FOR NETWORK FILESYSTEMS | 4100 | FS-CACHE: LOCAL CACHING FOR NETWORK FILESYSTEMS |
| 4100 | M: David Howells <dhowells@redhat.com> | 4101 | M: David Howells <dhowells@redhat.com> |
| 4101 | L: linux-cachefs@redhat.com | 4102 | L: linux-cachefs@redhat.com |
| 4102 | S: Supported | 4103 | S: Supported |
| 4103 | F: Documentation/filesystems/caching/ | 4104 | F: Documentation/filesystems/caching/ |
| 4104 | F: fs/fscache/ | 4105 | F: fs/fscache/ |
| 4105 | F: include/linux/fscache*.h | 4106 | F: include/linux/fscache*.h |
| 4106 | 4107 | ||
| 4107 | F2FS FILE SYSTEM | 4108 | F2FS FILE SYSTEM |
| 4108 | M: Jaegeuk Kim <jaegeuk@kernel.org> | 4109 | M: Jaegeuk Kim <jaegeuk@kernel.org> |
| 4109 | M: Changman Lee <cm224.lee@samsung.com> | 4110 | M: Changman Lee <cm224.lee@samsung.com> |
| 4110 | L: linux-f2fs-devel@lists.sourceforge.net | 4111 | L: linux-f2fs-devel@lists.sourceforge.net |
| 4111 | W: http://en.wikipedia.org/wiki/F2FS | 4112 | W: http://en.wikipedia.org/wiki/F2FS |
| 4112 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git | 4113 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs.git |
| 4113 | S: Maintained | 4114 | S: Maintained |
| 4114 | F: Documentation/filesystems/f2fs.txt | 4115 | F: Documentation/filesystems/f2fs.txt |
| 4115 | F: Documentation/ABI/testing/sysfs-fs-f2fs | 4116 | F: Documentation/ABI/testing/sysfs-fs-f2fs |
| 4116 | F: fs/f2fs/ | 4117 | F: fs/f2fs/ |
| 4117 | F: include/linux/f2fs_fs.h | 4118 | F: include/linux/f2fs_fs.h |
| 4118 | 4119 | ||
| 4119 | FUJITSU FR-V (FRV) PORT | 4120 | FUJITSU FR-V (FRV) PORT |
| 4120 | M: David Howells <dhowells@redhat.com> | 4121 | M: David Howells <dhowells@redhat.com> |
| 4121 | S: Maintained | 4122 | S: Maintained |
| 4122 | F: arch/frv/ | 4123 | F: arch/frv/ |
| 4123 | 4124 | ||
| 4124 | FUJITSU LAPTOP EXTRAS | 4125 | FUJITSU LAPTOP EXTRAS |
| 4125 | M: Jonathan Woithe <jwoithe@just42.net> | 4126 | M: Jonathan Woithe <jwoithe@just42.net> |
| 4126 | L: platform-driver-x86@vger.kernel.org | 4127 | L: platform-driver-x86@vger.kernel.org |
| 4127 | S: Maintained | 4128 | S: Maintained |
| 4128 | F: drivers/platform/x86/fujitsu-laptop.c | 4129 | F: drivers/platform/x86/fujitsu-laptop.c |
| 4129 | 4130 | ||
| 4130 | FUJITSU M-5MO LS CAMERA ISP DRIVER | 4131 | FUJITSU M-5MO LS CAMERA ISP DRIVER |
| 4131 | M: Kyungmin Park <kyungmin.park@samsung.com> | 4132 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 4132 | M: Heungjun Kim <riverful.kim@samsung.com> | 4133 | M: Heungjun Kim <riverful.kim@samsung.com> |
| 4133 | L: linux-media@vger.kernel.org | 4134 | L: linux-media@vger.kernel.org |
| 4134 | S: Maintained | 4135 | S: Maintained |
| 4135 | F: drivers/media/i2c/m5mols/ | 4136 | F: drivers/media/i2c/m5mols/ |
| 4136 | F: include/media/m5mols.h | 4137 | F: include/media/m5mols.h |
| 4137 | 4138 | ||
| 4138 | FUJITSU TABLET EXTRAS | 4139 | FUJITSU TABLET EXTRAS |
| 4139 | M: Robert Gerlach <khnz@gmx.de> | 4140 | M: Robert Gerlach <khnz@gmx.de> |
| 4140 | L: platform-driver-x86@vger.kernel.org | 4141 | L: platform-driver-x86@vger.kernel.org |
| 4141 | S: Maintained | 4142 | S: Maintained |
| 4142 | F: drivers/platform/x86/fujitsu-tablet.c | 4143 | F: drivers/platform/x86/fujitsu-tablet.c |
| 4143 | 4144 | ||
| 4144 | FUSE: FILESYSTEM IN USERSPACE | 4145 | FUSE: FILESYSTEM IN USERSPACE |
| 4145 | M: Miklos Szeredi <miklos@szeredi.hu> | 4146 | M: Miklos Szeredi <miklos@szeredi.hu> |
| 4146 | L: fuse-devel@lists.sourceforge.net | 4147 | L: fuse-devel@lists.sourceforge.net |
| 4147 | W: http://fuse.sourceforge.net/ | 4148 | W: http://fuse.sourceforge.net/ |
| 4148 | S: Maintained | 4149 | S: Maintained |
| 4149 | F: fs/fuse/ | 4150 | F: fs/fuse/ |
| 4150 | F: include/uapi/linux/fuse.h | 4151 | F: include/uapi/linux/fuse.h |
| 4151 | 4152 | ||
| 4152 | FUTURE DOMAIN TMC-16x0 SCSI DRIVER (16-bit) | 4153 | FUTURE DOMAIN TMC-16x0 SCSI DRIVER (16-bit) |
| 4153 | M: Rik Faith <faith@cs.unc.edu> | 4154 | M: Rik Faith <faith@cs.unc.edu> |
| 4154 | L: linux-scsi@vger.kernel.org | 4155 | L: linux-scsi@vger.kernel.org |
| 4155 | S: Odd Fixes (e.g., new signatures) | 4156 | S: Odd Fixes (e.g., new signatures) |
| 4156 | F: drivers/scsi/fdomain.* | 4157 | F: drivers/scsi/fdomain.* |
| 4157 | 4158 | ||
| 4158 | GCOV BASED KERNEL PROFILING | 4159 | GCOV BASED KERNEL PROFILING |
| 4159 | M: Peter Oberparleiter <oberpar@linux.vnet.ibm.com> | 4160 | M: Peter Oberparleiter <oberpar@linux.vnet.ibm.com> |
| 4160 | S: Maintained | 4161 | S: Maintained |
| 4161 | F: kernel/gcov/ | 4162 | F: kernel/gcov/ |
| 4162 | F: Documentation/gcov.txt | 4163 | F: Documentation/gcov.txt |
| 4163 | 4164 | ||
| 4164 | GDT SCSI DISK ARRAY CONTROLLER DRIVER | 4165 | GDT SCSI DISK ARRAY CONTROLLER DRIVER |
| 4165 | M: Achim Leubner <achim_leubner@adaptec.com> | 4166 | M: Achim Leubner <achim_leubner@adaptec.com> |
| 4166 | L: linux-scsi@vger.kernel.org | 4167 | L: linux-scsi@vger.kernel.org |
| 4167 | W: http://www.icp-vortex.com/ | 4168 | W: http://www.icp-vortex.com/ |
| 4168 | S: Supported | 4169 | S: Supported |
| 4169 | F: drivers/scsi/gdt* | 4170 | F: drivers/scsi/gdt* |
| 4170 | 4171 | ||
| 4171 | GEMTEK FM RADIO RECEIVER DRIVER | 4172 | GEMTEK FM RADIO RECEIVER DRIVER |
| 4172 | M: Hans Verkuil <hverkuil@xs4all.nl> | 4173 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 4173 | L: linux-media@vger.kernel.org | 4174 | L: linux-media@vger.kernel.org |
| 4174 | T: git git://linuxtv.org/media_tree.git | 4175 | T: git git://linuxtv.org/media_tree.git |
| 4175 | W: http://linuxtv.org | 4176 | W: http://linuxtv.org |
| 4176 | S: Maintained | 4177 | S: Maintained |
| 4177 | F: drivers/media/radio/radio-gemtek* | 4178 | F: drivers/media/radio/radio-gemtek* |
| 4178 | 4179 | ||
| 4179 | GENERIC GPIO I2C DRIVER | 4180 | GENERIC GPIO I2C DRIVER |
| 4180 | M: Haavard Skinnemoen <hskinnemoen@gmail.com> | 4181 | M: Haavard Skinnemoen <hskinnemoen@gmail.com> |
| 4181 | S: Supported | 4182 | S: Supported |
| 4182 | F: drivers/i2c/busses/i2c-gpio.c | 4183 | F: drivers/i2c/busses/i2c-gpio.c |
| 4183 | F: include/linux/i2c-gpio.h | 4184 | F: include/linux/i2c-gpio.h |
| 4184 | 4185 | ||
| 4185 | GENERIC GPIO I2C MULTIPLEXER DRIVER | 4186 | GENERIC GPIO I2C MULTIPLEXER DRIVER |
| 4186 | M: Peter Korsgaard <peter.korsgaard@barco.com> | 4187 | M: Peter Korsgaard <peter.korsgaard@barco.com> |
| 4187 | L: linux-i2c@vger.kernel.org | 4188 | L: linux-i2c@vger.kernel.org |
| 4188 | S: Supported | 4189 | S: Supported |
| 4189 | F: drivers/i2c/muxes/i2c-mux-gpio.c | 4190 | F: drivers/i2c/muxes/i2c-mux-gpio.c |
| 4190 | F: include/linux/i2c-mux-gpio.h | 4191 | F: include/linux/i2c-mux-gpio.h |
| 4191 | F: Documentation/i2c/muxes/i2c-mux-gpio | 4192 | F: Documentation/i2c/muxes/i2c-mux-gpio |
| 4192 | 4193 | ||
| 4193 | GENERIC HDLC (WAN) DRIVERS | 4194 | GENERIC HDLC (WAN) DRIVERS |
| 4194 | M: Krzysztof Halasa <khc@pm.waw.pl> | 4195 | M: Krzysztof Halasa <khc@pm.waw.pl> |
| 4195 | W: http://www.kernel.org/pub/linux/utils/net/hdlc/ | 4196 | W: http://www.kernel.org/pub/linux/utils/net/hdlc/ |
| 4196 | S: Maintained | 4197 | S: Maintained |
| 4197 | F: drivers/net/wan/c101.c | 4198 | F: drivers/net/wan/c101.c |
| 4198 | F: drivers/net/wan/hd6457* | 4199 | F: drivers/net/wan/hd6457* |
| 4199 | F: drivers/net/wan/hdlc* | 4200 | F: drivers/net/wan/hdlc* |
| 4200 | F: drivers/net/wan/n2.c | 4201 | F: drivers/net/wan/n2.c |
| 4201 | F: drivers/net/wan/pc300too.c | 4202 | F: drivers/net/wan/pc300too.c |
| 4202 | F: drivers/net/wan/pci200syn.c | 4203 | F: drivers/net/wan/pci200syn.c |
| 4203 | F: drivers/net/wan/wanxl* | 4204 | F: drivers/net/wan/wanxl* |
| 4204 | 4205 | ||
| 4205 | GENERIC INCLUDE/ASM HEADER FILES | 4206 | GENERIC INCLUDE/ASM HEADER FILES |
| 4206 | M: Arnd Bergmann <arnd@arndb.de> | 4207 | M: Arnd Bergmann <arnd@arndb.de> |
| 4207 | L: linux-arch@vger.kernel.org | 4208 | L: linux-arch@vger.kernel.org |
| 4208 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic.git | 4209 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic.git |
| 4209 | S: Maintained | 4210 | S: Maintained |
| 4210 | F: include/asm-generic/ | 4211 | F: include/asm-generic/ |
| 4211 | F: include/uapi/asm-generic/ | 4212 | F: include/uapi/asm-generic/ |
| 4212 | 4213 | ||
| 4213 | GENERIC PHY FRAMEWORK | 4214 | GENERIC PHY FRAMEWORK |
| 4214 | M: Kishon Vijay Abraham I <kishon@ti.com> | 4215 | M: Kishon Vijay Abraham I <kishon@ti.com> |
| 4215 | L: linux-kernel@vger.kernel.org | 4216 | L: linux-kernel@vger.kernel.org |
| 4216 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kishon/linux-phy.git | 4217 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kishon/linux-phy.git |
| 4217 | S: Supported | 4218 | S: Supported |
| 4218 | F: drivers/phy/ | 4219 | F: drivers/phy/ |
| 4219 | F: include/linux/phy/ | 4220 | F: include/linux/phy/ |
| 4220 | 4221 | ||
| 4221 | GENERIC UIO DRIVER FOR PCI DEVICES | 4222 | GENERIC UIO DRIVER FOR PCI DEVICES |
| 4222 | M: "Michael S. Tsirkin" <mst@redhat.com> | 4223 | M: "Michael S. Tsirkin" <mst@redhat.com> |
| 4223 | L: kvm@vger.kernel.org | 4224 | L: kvm@vger.kernel.org |
| 4224 | S: Supported | 4225 | S: Supported |
| 4225 | F: drivers/uio/uio_pci_generic.c | 4226 | F: drivers/uio/uio_pci_generic.c |
| 4226 | 4227 | ||
| 4227 | GET_MAINTAINER SCRIPT | 4228 | GET_MAINTAINER SCRIPT |
| 4228 | M: Joe Perches <joe@perches.com> | 4229 | M: Joe Perches <joe@perches.com> |
| 4229 | S: Maintained | 4230 | S: Maintained |
| 4230 | F: scripts/get_maintainer.pl | 4231 | F: scripts/get_maintainer.pl |
| 4231 | 4232 | ||
| 4232 | GFS2 FILE SYSTEM | 4233 | GFS2 FILE SYSTEM |
| 4233 | M: Steven Whitehouse <swhiteho@redhat.com> | 4234 | M: Steven Whitehouse <swhiteho@redhat.com> |
| 4234 | L: cluster-devel@redhat.com | 4235 | L: cluster-devel@redhat.com |
| 4235 | W: http://sources.redhat.com/cluster/ | 4236 | W: http://sources.redhat.com/cluster/ |
| 4236 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes.git | 4237 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes.git |
| 4237 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw.git | 4238 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-nmw.git |
| 4238 | S: Supported | 4239 | S: Supported |
| 4239 | F: Documentation/filesystems/gfs2*.txt | 4240 | F: Documentation/filesystems/gfs2*.txt |
| 4240 | F: fs/gfs2/ | 4241 | F: fs/gfs2/ |
| 4241 | F: include/uapi/linux/gfs2_ondisk.h | 4242 | F: include/uapi/linux/gfs2_ondisk.h |
| 4242 | 4243 | ||
| 4243 | GIGASET ISDN DRIVERS | 4244 | GIGASET ISDN DRIVERS |
| 4244 | M: Hansjoerg Lipp <hjlipp@web.de> | 4245 | M: Hansjoerg Lipp <hjlipp@web.de> |
| 4245 | M: Tilman Schmidt <tilman@imap.cc> | 4246 | M: Tilman Schmidt <tilman@imap.cc> |
| 4246 | L: gigaset307x-common@lists.sourceforge.net | 4247 | L: gigaset307x-common@lists.sourceforge.net |
| 4247 | W: http://gigaset307x.sourceforge.net/ | 4248 | W: http://gigaset307x.sourceforge.net/ |
| 4248 | S: Maintained | 4249 | S: Maintained |
| 4249 | F: Documentation/isdn/README.gigaset | 4250 | F: Documentation/isdn/README.gigaset |
| 4250 | F: drivers/isdn/gigaset/ | 4251 | F: drivers/isdn/gigaset/ |
| 4251 | F: include/uapi/linux/gigaset_dev.h | 4252 | F: include/uapi/linux/gigaset_dev.h |
| 4252 | 4253 | ||
| 4253 | GO7007 MPEG CODEC | 4254 | GO7007 MPEG CODEC |
| 4254 | M: Hans Verkuil <hans.verkuil@cisco.com> | 4255 | M: Hans Verkuil <hans.verkuil@cisco.com> |
| 4255 | L: linux-media@vger.kernel.org | 4256 | L: linux-media@vger.kernel.org |
| 4256 | S: Maintained | 4257 | S: Maintained |
| 4257 | F: drivers/media/usb/go7007/ | 4258 | F: drivers/media/usb/go7007/ |
| 4258 | 4259 | ||
| 4259 | GOODIX TOUCHSCREEN | 4260 | GOODIX TOUCHSCREEN |
| 4260 | M: Bastien Nocera <hadess@hadess.net> | 4261 | M: Bastien Nocera <hadess@hadess.net> |
| 4261 | L: linux-input@vger.kernel.org | 4262 | L: linux-input@vger.kernel.org |
| 4262 | S: Maintained | 4263 | S: Maintained |
| 4263 | F: drivers/input/touchscreen/goodix.c | 4264 | F: drivers/input/touchscreen/goodix.c |
| 4264 | 4265 | ||
| 4265 | GPIO SUBSYSTEM | 4266 | GPIO SUBSYSTEM |
| 4266 | M: Linus Walleij <linus.walleij@linaro.org> | 4267 | M: Linus Walleij <linus.walleij@linaro.org> |
| 4267 | M: Alexandre Courbot <gnurou@gmail.com> | 4268 | M: Alexandre Courbot <gnurou@gmail.com> |
| 4268 | L: linux-gpio@vger.kernel.org | 4269 | L: linux-gpio@vger.kernel.org |
| 4269 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio.git | 4270 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio.git |
| 4270 | S: Maintained | 4271 | S: Maintained |
| 4271 | F: Documentation/gpio/ | 4272 | F: Documentation/gpio/ |
| 4272 | F: drivers/gpio/ | 4273 | F: drivers/gpio/ |
| 4273 | F: include/linux/gpio/ | 4274 | F: include/linux/gpio/ |
| 4274 | F: include/linux/gpio.h | 4275 | F: include/linux/gpio.h |
| 4275 | F: include/asm-generic/gpio.h | 4276 | F: include/asm-generic/gpio.h |
| 4276 | 4277 | ||
| 4277 | GRE DEMULTIPLEXER DRIVER | 4278 | GRE DEMULTIPLEXER DRIVER |
| 4278 | M: Dmitry Kozlov <xeb@mail.ru> | 4279 | M: Dmitry Kozlov <xeb@mail.ru> |
| 4279 | L: netdev@vger.kernel.org | 4280 | L: netdev@vger.kernel.org |
| 4280 | S: Maintained | 4281 | S: Maintained |
| 4281 | F: net/ipv4/gre_demux.c | 4282 | F: net/ipv4/gre_demux.c |
| 4282 | F: net/ipv4/gre_offload.c | 4283 | F: net/ipv4/gre_offload.c |
| 4283 | F: include/net/gre.h | 4284 | F: include/net/gre.h |
| 4284 | 4285 | ||
| 4285 | GRETH 10/100/1G Ethernet MAC device driver | 4286 | GRETH 10/100/1G Ethernet MAC device driver |
| 4286 | M: Kristoffer Glembo <kristoffer@gaisler.com> | 4287 | M: Kristoffer Glembo <kristoffer@gaisler.com> |
| 4287 | L: netdev@vger.kernel.org | 4288 | L: netdev@vger.kernel.org |
| 4288 | S: Maintained | 4289 | S: Maintained |
| 4289 | F: drivers/net/ethernet/aeroflex/ | 4290 | F: drivers/net/ethernet/aeroflex/ |
| 4290 | 4291 | ||
| 4291 | GSPCA FINEPIX SUBDRIVER | 4292 | GSPCA FINEPIX SUBDRIVER |
| 4292 | M: Frank Zago <frank@zago.net> | 4293 | M: Frank Zago <frank@zago.net> |
| 4293 | L: linux-media@vger.kernel.org | 4294 | L: linux-media@vger.kernel.org |
| 4294 | T: git git://linuxtv.org/media_tree.git | 4295 | T: git git://linuxtv.org/media_tree.git |
| 4295 | S: Maintained | 4296 | S: Maintained |
| 4296 | F: drivers/media/usb/gspca/finepix.c | 4297 | F: drivers/media/usb/gspca/finepix.c |
| 4297 | 4298 | ||
| 4298 | GSPCA GL860 SUBDRIVER | 4299 | GSPCA GL860 SUBDRIVER |
| 4299 | M: Olivier Lorin <o.lorin@laposte.net> | 4300 | M: Olivier Lorin <o.lorin@laposte.net> |
| 4300 | L: linux-media@vger.kernel.org | 4301 | L: linux-media@vger.kernel.org |
| 4301 | T: git git://linuxtv.org/media_tree.git | 4302 | T: git git://linuxtv.org/media_tree.git |
| 4302 | S: Maintained | 4303 | S: Maintained |
| 4303 | F: drivers/media/usb/gspca/gl860/ | 4304 | F: drivers/media/usb/gspca/gl860/ |
| 4304 | 4305 | ||
| 4305 | GSPCA M5602 SUBDRIVER | 4306 | GSPCA M5602 SUBDRIVER |
| 4306 | M: Erik Andren <erik.andren@gmail.com> | 4307 | M: Erik Andren <erik.andren@gmail.com> |
| 4307 | L: linux-media@vger.kernel.org | 4308 | L: linux-media@vger.kernel.org |
| 4308 | T: git git://linuxtv.org/media_tree.git | 4309 | T: git git://linuxtv.org/media_tree.git |
| 4309 | S: Maintained | 4310 | S: Maintained |
| 4310 | F: drivers/media/usb/gspca/m5602/ | 4311 | F: drivers/media/usb/gspca/m5602/ |
| 4311 | 4312 | ||
| 4312 | GSPCA PAC207 SONIXB SUBDRIVER | 4313 | GSPCA PAC207 SONIXB SUBDRIVER |
| 4313 | M: Hans de Goede <hdegoede@redhat.com> | 4314 | M: Hans de Goede <hdegoede@redhat.com> |
| 4314 | L: linux-media@vger.kernel.org | 4315 | L: linux-media@vger.kernel.org |
| 4315 | T: git git://linuxtv.org/media_tree.git | 4316 | T: git git://linuxtv.org/media_tree.git |
| 4316 | S: Maintained | 4317 | S: Maintained |
| 4317 | F: drivers/media/usb/gspca/pac207.c | 4318 | F: drivers/media/usb/gspca/pac207.c |
| 4318 | 4319 | ||
| 4319 | GSPCA SN9C20X SUBDRIVER | 4320 | GSPCA SN9C20X SUBDRIVER |
| 4320 | M: Brian Johnson <brijohn@gmail.com> | 4321 | M: Brian Johnson <brijohn@gmail.com> |
| 4321 | L: linux-media@vger.kernel.org | 4322 | L: linux-media@vger.kernel.org |
| 4322 | T: git git://linuxtv.org/media_tree.git | 4323 | T: git git://linuxtv.org/media_tree.git |
| 4323 | S: Maintained | 4324 | S: Maintained |
| 4324 | F: drivers/media/usb/gspca/sn9c20x.c | 4325 | F: drivers/media/usb/gspca/sn9c20x.c |
| 4325 | 4326 | ||
| 4326 | GSPCA T613 SUBDRIVER | 4327 | GSPCA T613 SUBDRIVER |
| 4327 | M: Leandro Costantino <lcostantino@gmail.com> | 4328 | M: Leandro Costantino <lcostantino@gmail.com> |
| 4328 | L: linux-media@vger.kernel.org | 4329 | L: linux-media@vger.kernel.org |
| 4329 | T: git git://linuxtv.org/media_tree.git | 4330 | T: git git://linuxtv.org/media_tree.git |
| 4330 | S: Maintained | 4331 | S: Maintained |
| 4331 | F: drivers/media/usb/gspca/t613.c | 4332 | F: drivers/media/usb/gspca/t613.c |
| 4332 | 4333 | ||
| 4333 | GSPCA USB WEBCAM DRIVER | 4334 | GSPCA USB WEBCAM DRIVER |
| 4334 | M: Hans de Goede <hdegoede@redhat.com> | 4335 | M: Hans de Goede <hdegoede@redhat.com> |
| 4335 | L: linux-media@vger.kernel.org | 4336 | L: linux-media@vger.kernel.org |
| 4336 | T: git git://linuxtv.org/media_tree.git | 4337 | T: git git://linuxtv.org/media_tree.git |
| 4337 | S: Maintained | 4338 | S: Maintained |
| 4338 | F: drivers/media/usb/gspca/ | 4339 | F: drivers/media/usb/gspca/ |
| 4339 | 4340 | ||
| 4340 | GUID PARTITION TABLE (GPT) | 4341 | GUID PARTITION TABLE (GPT) |
| 4341 | M: Davidlohr Bueso <davidlohr@hp.com> | 4342 | M: Davidlohr Bueso <davidlohr@hp.com> |
| 4342 | L: linux-efi@vger.kernel.org | 4343 | L: linux-efi@vger.kernel.org |
| 4343 | S: Maintained | 4344 | S: Maintained |
| 4344 | F: block/partitions/efi.* | 4345 | F: block/partitions/efi.* |
| 4345 | 4346 | ||
| 4346 | STK1160 USB VIDEO CAPTURE DRIVER | 4347 | STK1160 USB VIDEO CAPTURE DRIVER |
| 4347 | M: Ezequiel Garcia <elezegarcia@gmail.com> | 4348 | M: Ezequiel Garcia <elezegarcia@gmail.com> |
| 4348 | L: linux-media@vger.kernel.org | 4349 | L: linux-media@vger.kernel.org |
| 4349 | T: git git://linuxtv.org/media_tree.git | 4350 | T: git git://linuxtv.org/media_tree.git |
| 4350 | S: Maintained | 4351 | S: Maintained |
| 4351 | F: drivers/media/usb/stk1160/ | 4352 | F: drivers/media/usb/stk1160/ |
| 4352 | 4353 | ||
| 4353 | HARD DRIVE ACTIVE PROTECTION SYSTEM (HDAPS) DRIVER | 4354 | HARD DRIVE ACTIVE PROTECTION SYSTEM (HDAPS) DRIVER |
| 4354 | M: Frank Seidel <frank@f-seidel.de> | 4355 | M: Frank Seidel <frank@f-seidel.de> |
| 4355 | L: platform-driver-x86@vger.kernel.org | 4356 | L: platform-driver-x86@vger.kernel.org |
| 4356 | W: http://www.kernel.org/pub/linux/kernel/people/fseidel/hdaps/ | 4357 | W: http://www.kernel.org/pub/linux/kernel/people/fseidel/hdaps/ |
| 4357 | S: Maintained | 4358 | S: Maintained |
| 4358 | F: drivers/platform/x86/hdaps.c | 4359 | F: drivers/platform/x86/hdaps.c |
| 4359 | 4360 | ||
| 4360 | HDPVR USB VIDEO ENCODER DRIVER | 4361 | HDPVR USB VIDEO ENCODER DRIVER |
| 4361 | M: Hans Verkuil <hverkuil@xs4all.nl> | 4362 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 4362 | L: linux-media@vger.kernel.org | 4363 | L: linux-media@vger.kernel.org |
| 4363 | T: git git://linuxtv.org/media_tree.git | 4364 | T: git git://linuxtv.org/media_tree.git |
| 4364 | W: http://linuxtv.org | 4365 | W: http://linuxtv.org |
| 4365 | S: Odd Fixes | 4366 | S: Odd Fixes |
| 4366 | F: drivers/media/usb/hdpvr/ | 4367 | F: drivers/media/usb/hdpvr/ |
| 4367 | 4368 | ||
| 4368 | HWPOISON MEMORY FAILURE HANDLING | 4369 | HWPOISON MEMORY FAILURE HANDLING |
| 4369 | M: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> | 4370 | M: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> |
| 4370 | L: linux-mm@kvack.org | 4371 | L: linux-mm@kvack.org |
| 4371 | S: Maintained | 4372 | S: Maintained |
| 4372 | F: mm/memory-failure.c | 4373 | F: mm/memory-failure.c |
| 4373 | F: mm/hwpoison-inject.c | 4374 | F: mm/hwpoison-inject.c |
| 4374 | 4375 | ||
| 4375 | HYPERVISOR VIRTUAL CONSOLE DRIVER | 4376 | HYPERVISOR VIRTUAL CONSOLE DRIVER |
| 4376 | L: linuxppc-dev@lists.ozlabs.org | 4377 | L: linuxppc-dev@lists.ozlabs.org |
| 4377 | S: Odd Fixes | 4378 | S: Odd Fixes |
| 4378 | F: drivers/tty/hvc/ | 4379 | F: drivers/tty/hvc/ |
| 4379 | 4380 | ||
| 4380 | HACKRF MEDIA DRIVER | 4381 | HACKRF MEDIA DRIVER |
| 4381 | M: Antti Palosaari <crope@iki.fi> | 4382 | M: Antti Palosaari <crope@iki.fi> |
| 4382 | L: linux-media@vger.kernel.org | 4383 | L: linux-media@vger.kernel.org |
| 4383 | W: http://linuxtv.org/ | 4384 | W: http://linuxtv.org/ |
| 4384 | W: http://palosaari.fi/linux/ | 4385 | W: http://palosaari.fi/linux/ |
| 4385 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 4386 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 4386 | T: git git://linuxtv.org/anttip/media_tree.git | 4387 | T: git git://linuxtv.org/anttip/media_tree.git |
| 4387 | S: Maintained | 4388 | S: Maintained |
| 4388 | F: drivers/media/usb/hackrf/ | 4389 | F: drivers/media/usb/hackrf/ |
| 4389 | 4390 | ||
| 4390 | HARDWARE MONITORING | 4391 | HARDWARE MONITORING |
| 4391 | M: Jean Delvare <jdelvare@suse.de> | 4392 | M: Jean Delvare <jdelvare@suse.de> |
| 4392 | M: Guenter Roeck <linux@roeck-us.net> | 4393 | M: Guenter Roeck <linux@roeck-us.net> |
| 4393 | L: lm-sensors@lm-sensors.org | 4394 | L: lm-sensors@lm-sensors.org |
| 4394 | W: http://www.lm-sensors.org/ | 4395 | W: http://www.lm-sensors.org/ |
| 4395 | T: quilt kernel.org/pub/linux/kernel/people/jdelvare/linux-2.6/jdelvare-hwmon/ | 4396 | T: quilt kernel.org/pub/linux/kernel/people/jdelvare/linux-2.6/jdelvare-hwmon/ |
| 4396 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging.git | 4397 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging.git |
| 4397 | S: Maintained | 4398 | S: Maintained |
| 4398 | F: Documentation/hwmon/ | 4399 | F: Documentation/hwmon/ |
| 4399 | F: drivers/hwmon/ | 4400 | F: drivers/hwmon/ |
| 4400 | F: include/linux/hwmon*.h | 4401 | F: include/linux/hwmon*.h |
| 4401 | 4402 | ||
| 4402 | HARDWARE RANDOM NUMBER GENERATOR CORE | 4403 | HARDWARE RANDOM NUMBER GENERATOR CORE |
| 4403 | M: Matt Mackall <mpm@selenic.com> | 4404 | M: Matt Mackall <mpm@selenic.com> |
| 4404 | M: Herbert Xu <herbert@gondor.apana.org.au> | 4405 | M: Herbert Xu <herbert@gondor.apana.org.au> |
| 4405 | S: Odd fixes | 4406 | S: Odd fixes |
| 4406 | F: Documentation/hw_random.txt | 4407 | F: Documentation/hw_random.txt |
| 4407 | F: drivers/char/hw_random/ | 4408 | F: drivers/char/hw_random/ |
| 4408 | F: include/linux/hw_random.h | 4409 | F: include/linux/hw_random.h |
| 4409 | 4410 | ||
| 4410 | HARDWARE SPINLOCK CORE | 4411 | HARDWARE SPINLOCK CORE |
| 4411 | M: Ohad Ben-Cohen <ohad@wizery.com> | 4412 | M: Ohad Ben-Cohen <ohad@wizery.com> |
| 4412 | S: Maintained | 4413 | S: Maintained |
| 4413 | F: Documentation/hwspinlock.txt | 4414 | F: Documentation/hwspinlock.txt |
| 4414 | F: drivers/hwspinlock/hwspinlock_* | 4415 | F: drivers/hwspinlock/hwspinlock_* |
| 4415 | F: include/linux/hwspinlock.h | 4416 | F: include/linux/hwspinlock.h |
| 4416 | 4417 | ||
| 4417 | HARMONY SOUND DRIVER | 4418 | HARMONY SOUND DRIVER |
| 4418 | L: linux-parisc@vger.kernel.org | 4419 | L: linux-parisc@vger.kernel.org |
| 4419 | S: Maintained | 4420 | S: Maintained |
| 4420 | F: sound/parisc/harmony.* | 4421 | F: sound/parisc/harmony.* |
| 4421 | 4422 | ||
| 4422 | HD29L2 MEDIA DRIVER | 4423 | HD29L2 MEDIA DRIVER |
| 4423 | M: Antti Palosaari <crope@iki.fi> | 4424 | M: Antti Palosaari <crope@iki.fi> |
| 4424 | L: linux-media@vger.kernel.org | 4425 | L: linux-media@vger.kernel.org |
| 4425 | W: http://linuxtv.org/ | 4426 | W: http://linuxtv.org/ |
| 4426 | W: http://palosaari.fi/linux/ | 4427 | W: http://palosaari.fi/linux/ |
| 4427 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 4428 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 4428 | T: git git://linuxtv.org/anttip/media_tree.git | 4429 | T: git git://linuxtv.org/anttip/media_tree.git |
| 4429 | S: Maintained | 4430 | S: Maintained |
| 4430 | F: drivers/media/dvb-frontends/hd29l2* | 4431 | F: drivers/media/dvb-frontends/hd29l2* |
| 4431 | 4432 | ||
| 4432 | HEWLETT-PACKARD SMART2 RAID DRIVER | 4433 | HEWLETT-PACKARD SMART2 RAID DRIVER |
| 4433 | L: iss_storagedev@hp.com | 4434 | L: iss_storagedev@hp.com |
| 4434 | S: Orphan | 4435 | S: Orphan |
| 4435 | F: Documentation/blockdev/cpqarray.txt | 4436 | F: Documentation/blockdev/cpqarray.txt |
| 4436 | F: drivers/block/cpqarray.* | 4437 | F: drivers/block/cpqarray.* |
| 4437 | 4438 | ||
| 4438 | HEWLETT-PACKARD SMART ARRAY RAID DRIVER (hpsa) | 4439 | HEWLETT-PACKARD SMART ARRAY RAID DRIVER (hpsa) |
| 4439 | M: Don Brace <don.brace@pmcs.com> | 4440 | M: Don Brace <don.brace@pmcs.com> |
| 4440 | L: iss_storagedev@hp.com | 4441 | L: iss_storagedev@hp.com |
| 4441 | L: storagedev@pmcs.com | 4442 | L: storagedev@pmcs.com |
| 4442 | L: linux-scsi@vger.kernel.org | 4443 | L: linux-scsi@vger.kernel.org |
| 4443 | S: Supported | 4444 | S: Supported |
| 4444 | F: Documentation/scsi/hpsa.txt | 4445 | F: Documentation/scsi/hpsa.txt |
| 4445 | F: drivers/scsi/hpsa*.[ch] | 4446 | F: drivers/scsi/hpsa*.[ch] |
| 4446 | F: include/linux/cciss*.h | 4447 | F: include/linux/cciss*.h |
| 4447 | F: include/uapi/linux/cciss*.h | 4448 | F: include/uapi/linux/cciss*.h |
| 4448 | 4449 | ||
| 4449 | HEWLETT-PACKARD SMART CISS RAID DRIVER (cciss) | 4450 | HEWLETT-PACKARD SMART CISS RAID DRIVER (cciss) |
| 4450 | M: Don Brace <don.brace@pmcs.com> | 4451 | M: Don Brace <don.brace@pmcs.com> |
| 4451 | L: iss_storagedev@hp.com | 4452 | L: iss_storagedev@hp.com |
| 4452 | L: storagedev@pmcs.com | 4453 | L: storagedev@pmcs.com |
| 4453 | L: linux-scsi@vger.kernel.org | 4454 | L: linux-scsi@vger.kernel.org |
| 4454 | S: Supported | 4455 | S: Supported |
| 4455 | F: Documentation/blockdev/cciss.txt | 4456 | F: Documentation/blockdev/cciss.txt |
| 4456 | F: drivers/block/cciss* | 4457 | F: drivers/block/cciss* |
| 4457 | F: include/linux/cciss_ioctl.h | 4458 | F: include/linux/cciss_ioctl.h |
| 4458 | F: include/uapi/linux/cciss_ioctl.h | 4459 | F: include/uapi/linux/cciss_ioctl.h |
| 4459 | 4460 | ||
| 4460 | HFS FILESYSTEM | 4461 | HFS FILESYSTEM |
| 4461 | L: linux-fsdevel@vger.kernel.org | 4462 | L: linux-fsdevel@vger.kernel.org |
| 4462 | S: Orphan | 4463 | S: Orphan |
| 4463 | F: Documentation/filesystems/hfs.txt | 4464 | F: Documentation/filesystems/hfs.txt |
| 4464 | F: fs/hfs/ | 4465 | F: fs/hfs/ |
| 4465 | 4466 | ||
| 4466 | HFSPLUS FILESYSTEM | 4467 | HFSPLUS FILESYSTEM |
| 4467 | L: linux-fsdevel@vger.kernel.org | 4468 | L: linux-fsdevel@vger.kernel.org |
| 4468 | S: Orphan | 4469 | S: Orphan |
| 4469 | F: Documentation/filesystems/hfsplus.txt | 4470 | F: Documentation/filesystems/hfsplus.txt |
| 4470 | F: fs/hfsplus/ | 4471 | F: fs/hfsplus/ |
| 4471 | 4472 | ||
| 4472 | HGA FRAMEBUFFER DRIVER | 4473 | HGA FRAMEBUFFER DRIVER |
| 4473 | M: Ferenc Bakonyi <fero@drama.obuda.kando.hu> | 4474 | M: Ferenc Bakonyi <fero@drama.obuda.kando.hu> |
| 4474 | L: linux-nvidia@lists.surfsouth.com | 4475 | L: linux-nvidia@lists.surfsouth.com |
| 4475 | W: http://drama.obuda.kando.hu/~fero/cgi-bin/hgafb.shtml | 4476 | W: http://drama.obuda.kando.hu/~fero/cgi-bin/hgafb.shtml |
| 4476 | S: Maintained | 4477 | S: Maintained |
| 4477 | F: drivers/video/fbdev/hgafb.c | 4478 | F: drivers/video/fbdev/hgafb.c |
| 4478 | 4479 | ||
| 4479 | HIBERNATION (aka Software Suspend, aka swsusp) | 4480 | HIBERNATION (aka Software Suspend, aka swsusp) |
| 4480 | M: "Rafael J. Wysocki" <rjw@rjwysocki.net> | 4481 | M: "Rafael J. Wysocki" <rjw@rjwysocki.net> |
| 4481 | M: Pavel Machek <pavel@ucw.cz> | 4482 | M: Pavel Machek <pavel@ucw.cz> |
| 4482 | L: linux-pm@vger.kernel.org | 4483 | L: linux-pm@vger.kernel.org |
| 4483 | S: Supported | 4484 | S: Supported |
| 4484 | F: arch/x86/power/ | 4485 | F: arch/x86/power/ |
| 4485 | F: drivers/base/power/ | 4486 | F: drivers/base/power/ |
| 4486 | F: kernel/power/ | 4487 | F: kernel/power/ |
| 4487 | F: include/linux/suspend.h | 4488 | F: include/linux/suspend.h |
| 4488 | F: include/linux/freezer.h | 4489 | F: include/linux/freezer.h |
| 4489 | F: include/linux/pm.h | 4490 | F: include/linux/pm.h |
| 4490 | F: arch/*/include/asm/suspend*.h | 4491 | F: arch/*/include/asm/suspend*.h |
| 4491 | 4492 | ||
| 4492 | HID CORE LAYER | 4493 | HID CORE LAYER |
| 4493 | M: Jiri Kosina <jkosina@suse.cz> | 4494 | M: Jiri Kosina <jkosina@suse.cz> |
| 4494 | L: linux-input@vger.kernel.org | 4495 | L: linux-input@vger.kernel.org |
| 4495 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git | 4496 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git |
| 4496 | S: Maintained | 4497 | S: Maintained |
| 4497 | F: drivers/hid/ | 4498 | F: drivers/hid/ |
| 4498 | F: include/linux/hid* | 4499 | F: include/linux/hid* |
| 4499 | F: include/uapi/linux/hid* | 4500 | F: include/uapi/linux/hid* |
| 4500 | 4501 | ||
| 4501 | HIGH-RESOLUTION TIMERS, CLOCKEVENTS, DYNTICKS | 4502 | HIGH-RESOLUTION TIMERS, CLOCKEVENTS, DYNTICKS |
| 4502 | M: Thomas Gleixner <tglx@linutronix.de> | 4503 | M: Thomas Gleixner <tglx@linutronix.de> |
| 4503 | L: linux-kernel@vger.kernel.org | 4504 | L: linux-kernel@vger.kernel.org |
| 4504 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core | 4505 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core |
| 4505 | S: Maintained | 4506 | S: Maintained |
| 4506 | F: Documentation/timers/ | 4507 | F: Documentation/timers/ |
| 4507 | F: kernel/time/hrtimer.c | 4508 | F: kernel/time/hrtimer.c |
| 4508 | F: kernel/time/clockevents.c | 4509 | F: kernel/time/clockevents.c |
| 4509 | F: kernel/time/tick*.* | 4510 | F: kernel/time/tick*.* |
| 4510 | F: kernel/time/timer_*.c | 4511 | F: kernel/time/timer_*.c |
| 4511 | F: include/linux/clockchips.h | 4512 | F: include/linux/clockchips.h |
| 4512 | F: include/linux/hrtimer.h | 4513 | F: include/linux/hrtimer.h |
| 4513 | 4514 | ||
| 4514 | HIGH-SPEED SCC DRIVER FOR AX.25 | 4515 | HIGH-SPEED SCC DRIVER FOR AX.25 |
| 4515 | L: linux-hams@vger.kernel.org | 4516 | L: linux-hams@vger.kernel.org |
| 4516 | S: Orphan | 4517 | S: Orphan |
| 4517 | F: drivers/net/hamradio/dmascc.c | 4518 | F: drivers/net/hamradio/dmascc.c |
| 4518 | F: drivers/net/hamradio/scc.c | 4519 | F: drivers/net/hamradio/scc.c |
| 4519 | 4520 | ||
| 4520 | HIGHPOINT ROCKETRAID 3xxx RAID DRIVER | 4521 | HIGHPOINT ROCKETRAID 3xxx RAID DRIVER |
| 4521 | M: HighPoint Linux Team <linux@highpoint-tech.com> | 4522 | M: HighPoint Linux Team <linux@highpoint-tech.com> |
| 4522 | W: http://www.highpoint-tech.com | 4523 | W: http://www.highpoint-tech.com |
| 4523 | S: Supported | 4524 | S: Supported |
| 4524 | F: Documentation/scsi/hptiop.txt | 4525 | F: Documentation/scsi/hptiop.txt |
| 4525 | F: drivers/scsi/hptiop.c | 4526 | F: drivers/scsi/hptiop.c |
| 4526 | 4527 | ||
| 4527 | HIPPI | 4528 | HIPPI |
| 4528 | M: Jes Sorensen <jes@trained-monkey.org> | 4529 | M: Jes Sorensen <jes@trained-monkey.org> |
| 4529 | L: linux-hippi@sunsite.dk | 4530 | L: linux-hippi@sunsite.dk |
| 4530 | S: Maintained | 4531 | S: Maintained |
| 4531 | F: include/linux/hippidevice.h | 4532 | F: include/linux/hippidevice.h |
| 4532 | F: include/uapi/linux/if_hippi.h | 4533 | F: include/uapi/linux/if_hippi.h |
| 4533 | F: net/802/hippi.c | 4534 | F: net/802/hippi.c |
| 4534 | F: drivers/net/hippi/ | 4535 | F: drivers/net/hippi/ |
| 4535 | 4536 | ||
| 4536 | HOST AP DRIVER | 4537 | HOST AP DRIVER |
| 4537 | M: Jouni Malinen <j@w1.fi> | 4538 | M: Jouni Malinen <j@w1.fi> |
| 4538 | L: hostap@shmoo.com (subscribers-only) | 4539 | L: hostap@shmoo.com (subscribers-only) |
| 4539 | L: linux-wireless@vger.kernel.org | 4540 | L: linux-wireless@vger.kernel.org |
| 4540 | W: http://hostap.epitest.fi/ | 4541 | W: http://hostap.epitest.fi/ |
| 4541 | S: Maintained | 4542 | S: Maintained |
| 4542 | F: drivers/net/wireless/hostap/ | 4543 | F: drivers/net/wireless/hostap/ |
| 4543 | 4544 | ||
| 4544 | HP COMPAQ TC1100 TABLET WMI EXTRAS DRIVER | 4545 | HP COMPAQ TC1100 TABLET WMI EXTRAS DRIVER |
| 4545 | L: platform-driver-x86@vger.kernel.org | 4546 | L: platform-driver-x86@vger.kernel.org |
| 4546 | S: Orphan | 4547 | S: Orphan |
| 4547 | F: drivers/platform/x86/tc1100-wmi.c | 4548 | F: drivers/platform/x86/tc1100-wmi.c |
| 4548 | 4549 | ||
| 4549 | HP100: Driver for HP 10/100 Mbit/s Voice Grade Network Adapter Series | 4550 | HP100: Driver for HP 10/100 Mbit/s Voice Grade Network Adapter Series |
| 4550 | M: Jaroslav Kysela <perex@perex.cz> | 4551 | M: Jaroslav Kysela <perex@perex.cz> |
| 4551 | S: Maintained | 4552 | S: Maintained |
| 4552 | F: drivers/net/ethernet/hp/hp100.* | 4553 | F: drivers/net/ethernet/hp/hp100.* |
| 4553 | 4554 | ||
| 4554 | HPET: High Precision Event Timers driver | 4555 | HPET: High Precision Event Timers driver |
| 4555 | M: Clemens Ladisch <clemens@ladisch.de> | 4556 | M: Clemens Ladisch <clemens@ladisch.de> |
| 4556 | S: Maintained | 4557 | S: Maintained |
| 4557 | F: Documentation/timers/hpet.txt | 4558 | F: Documentation/timers/hpet.txt |
| 4558 | F: drivers/char/hpet.c | 4559 | F: drivers/char/hpet.c |
| 4559 | F: include/linux/hpet.h | 4560 | F: include/linux/hpet.h |
| 4560 | F: include/uapi/linux/hpet.h | 4561 | F: include/uapi/linux/hpet.h |
| 4561 | 4562 | ||
| 4562 | HPET: x86 | 4563 | HPET: x86 |
| 4563 | S: Orphan | 4564 | S: Orphan |
| 4564 | F: arch/x86/kernel/hpet.c | 4565 | F: arch/x86/kernel/hpet.c |
| 4565 | F: arch/x86/include/asm/hpet.h | 4566 | F: arch/x86/include/asm/hpet.h |
| 4566 | 4567 | ||
| 4567 | HPFS FILESYSTEM | 4568 | HPFS FILESYSTEM |
| 4568 | M: Mikulas Patocka <mikulas@artax.karlin.mff.cuni.cz> | 4569 | M: Mikulas Patocka <mikulas@artax.karlin.mff.cuni.cz> |
| 4569 | W: http://artax.karlin.mff.cuni.cz/~mikulas/vyplody/hpfs/index-e.cgi | 4570 | W: http://artax.karlin.mff.cuni.cz/~mikulas/vyplody/hpfs/index-e.cgi |
| 4570 | S: Maintained | 4571 | S: Maintained |
| 4571 | F: fs/hpfs/ | 4572 | F: fs/hpfs/ |
| 4572 | 4573 | ||
| 4573 | HSI SUBSYSTEM | 4574 | HSI SUBSYSTEM |
| 4574 | M: Sebastian Reichel <sre@kernel.org> | 4575 | M: Sebastian Reichel <sre@kernel.org> |
| 4575 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-hsi.git | 4576 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sre/linux-hsi.git |
| 4576 | S: Maintained | 4577 | S: Maintained |
| 4577 | F: Documentation/ABI/testing/sysfs-bus-hsi | 4578 | F: Documentation/ABI/testing/sysfs-bus-hsi |
| 4578 | F: Documentation/hsi.txt | 4579 | F: Documentation/hsi.txt |
| 4579 | F: drivers/hsi/ | 4580 | F: drivers/hsi/ |
| 4580 | F: include/linux/hsi/ | 4581 | F: include/linux/hsi/ |
| 4581 | F: include/uapi/linux/hsi/ | 4582 | F: include/uapi/linux/hsi/ |
| 4582 | 4583 | ||
| 4583 | HSO 3G MODEM DRIVER | 4584 | HSO 3G MODEM DRIVER |
| 4584 | M: Jan Dumon <j.dumon@option.com> | 4585 | M: Jan Dumon <j.dumon@option.com> |
| 4585 | W: http://www.pharscape.org | 4586 | W: http://www.pharscape.org |
| 4586 | S: Maintained | 4587 | S: Maintained |
| 4587 | F: drivers/net/usb/hso.c | 4588 | F: drivers/net/usb/hso.c |
| 4588 | 4589 | ||
| 4589 | HSR NETWORK PROTOCOL | 4590 | HSR NETWORK PROTOCOL |
| 4590 | M: Arvid Brodin <arvid.brodin@alten.se> | 4591 | M: Arvid Brodin <arvid.brodin@alten.se> |
| 4591 | L: netdev@vger.kernel.org | 4592 | L: netdev@vger.kernel.org |
| 4592 | S: Maintained | 4593 | S: Maintained |
| 4593 | F: net/hsr/ | 4594 | F: net/hsr/ |
| 4594 | 4595 | ||
| 4595 | HTCPEN TOUCHSCREEN DRIVER | 4596 | HTCPEN TOUCHSCREEN DRIVER |
| 4596 | M: Pau Oliva Fora <pof@eslack.org> | 4597 | M: Pau Oliva Fora <pof@eslack.org> |
| 4597 | L: linux-input@vger.kernel.org | 4598 | L: linux-input@vger.kernel.org |
| 4598 | S: Maintained | 4599 | S: Maintained |
| 4599 | F: drivers/input/touchscreen/htcpen.c | 4600 | F: drivers/input/touchscreen/htcpen.c |
| 4600 | 4601 | ||
| 4601 | HUGETLB FILESYSTEM | 4602 | HUGETLB FILESYSTEM |
| 4602 | M: Nadia Yvette Chambers <nyc@holomorphy.com> | 4603 | M: Nadia Yvette Chambers <nyc@holomorphy.com> |
| 4603 | S: Maintained | 4604 | S: Maintained |
| 4604 | F: fs/hugetlbfs/ | 4605 | F: fs/hugetlbfs/ |
| 4605 | 4606 | ||
| 4606 | Hyper-V CORE AND DRIVERS | 4607 | Hyper-V CORE AND DRIVERS |
| 4607 | M: K. Y. Srinivasan <kys@microsoft.com> | 4608 | M: K. Y. Srinivasan <kys@microsoft.com> |
| 4608 | M: Haiyang Zhang <haiyangz@microsoft.com> | 4609 | M: Haiyang Zhang <haiyangz@microsoft.com> |
| 4609 | L: devel@linuxdriverproject.org | 4610 | L: devel@linuxdriverproject.org |
| 4610 | S: Maintained | 4611 | S: Maintained |
| 4611 | F: arch/x86/include/asm/mshyperv.h | 4612 | F: arch/x86/include/asm/mshyperv.h |
| 4612 | F: arch/x86/include/uapi/asm/hyperv.h | 4613 | F: arch/x86/include/uapi/asm/hyperv.h |
| 4613 | F: arch/x86/kernel/cpu/mshyperv.c | 4614 | F: arch/x86/kernel/cpu/mshyperv.c |
| 4614 | F: drivers/hid/hid-hyperv.c | 4615 | F: drivers/hid/hid-hyperv.c |
| 4615 | F: drivers/hv/ | 4616 | F: drivers/hv/ |
| 4616 | F: drivers/input/serio/hyperv-keyboard.c | 4617 | F: drivers/input/serio/hyperv-keyboard.c |
| 4617 | F: drivers/net/hyperv/ | 4618 | F: drivers/net/hyperv/ |
| 4618 | F: drivers/scsi/storvsc_drv.c | 4619 | F: drivers/scsi/storvsc_drv.c |
| 4619 | F: drivers/video/fbdev/hyperv_fb.c | 4620 | F: drivers/video/fbdev/hyperv_fb.c |
| 4620 | F: include/linux/hyperv.h | 4621 | F: include/linux/hyperv.h |
| 4621 | F: tools/hv/ | 4622 | F: tools/hv/ |
| 4622 | 4623 | ||
| 4623 | I2C OVER PARALLEL PORT | 4624 | I2C OVER PARALLEL PORT |
| 4624 | M: Jean Delvare <jdelvare@suse.de> | 4625 | M: Jean Delvare <jdelvare@suse.de> |
| 4625 | L: linux-i2c@vger.kernel.org | 4626 | L: linux-i2c@vger.kernel.org |
| 4626 | S: Maintained | 4627 | S: Maintained |
| 4627 | F: Documentation/i2c/busses/i2c-parport | 4628 | F: Documentation/i2c/busses/i2c-parport |
| 4628 | F: Documentation/i2c/busses/i2c-parport-light | 4629 | F: Documentation/i2c/busses/i2c-parport-light |
| 4629 | F: drivers/i2c/busses/i2c-parport.c | 4630 | F: drivers/i2c/busses/i2c-parport.c |
| 4630 | F: drivers/i2c/busses/i2c-parport-light.c | 4631 | F: drivers/i2c/busses/i2c-parport-light.c |
| 4631 | 4632 | ||
| 4632 | I2C/SMBUS CONTROLLER DRIVERS FOR PC | 4633 | I2C/SMBUS CONTROLLER DRIVERS FOR PC |
| 4633 | M: Jean Delvare <jdelvare@suse.de> | 4634 | M: Jean Delvare <jdelvare@suse.de> |
| 4634 | L: linux-i2c@vger.kernel.org | 4635 | L: linux-i2c@vger.kernel.org |
| 4635 | S: Maintained | 4636 | S: Maintained |
| 4636 | F: Documentation/i2c/busses/i2c-ali1535 | 4637 | F: Documentation/i2c/busses/i2c-ali1535 |
| 4637 | F: Documentation/i2c/busses/i2c-ali1563 | 4638 | F: Documentation/i2c/busses/i2c-ali1563 |
| 4638 | F: Documentation/i2c/busses/i2c-ali15x3 | 4639 | F: Documentation/i2c/busses/i2c-ali15x3 |
| 4639 | F: Documentation/i2c/busses/i2c-amd756 | 4640 | F: Documentation/i2c/busses/i2c-amd756 |
| 4640 | F: Documentation/i2c/busses/i2c-amd8111 | 4641 | F: Documentation/i2c/busses/i2c-amd8111 |
| 4641 | F: Documentation/i2c/busses/i2c-i801 | 4642 | F: Documentation/i2c/busses/i2c-i801 |
| 4642 | F: Documentation/i2c/busses/i2c-nforce2 | 4643 | F: Documentation/i2c/busses/i2c-nforce2 |
| 4643 | F: Documentation/i2c/busses/i2c-piix4 | 4644 | F: Documentation/i2c/busses/i2c-piix4 |
| 4644 | F: Documentation/i2c/busses/i2c-sis5595 | 4645 | F: Documentation/i2c/busses/i2c-sis5595 |
| 4645 | F: Documentation/i2c/busses/i2c-sis630 | 4646 | F: Documentation/i2c/busses/i2c-sis630 |
| 4646 | F: Documentation/i2c/busses/i2c-sis96x | 4647 | F: Documentation/i2c/busses/i2c-sis96x |
| 4647 | F: Documentation/i2c/busses/i2c-via | 4648 | F: Documentation/i2c/busses/i2c-via |
| 4648 | F: Documentation/i2c/busses/i2c-viapro | 4649 | F: Documentation/i2c/busses/i2c-viapro |
| 4649 | F: drivers/i2c/busses/i2c-ali1535.c | 4650 | F: drivers/i2c/busses/i2c-ali1535.c |
| 4650 | F: drivers/i2c/busses/i2c-ali1563.c | 4651 | F: drivers/i2c/busses/i2c-ali1563.c |
| 4651 | F: drivers/i2c/busses/i2c-ali15x3.c | 4652 | F: drivers/i2c/busses/i2c-ali15x3.c |
| 4652 | F: drivers/i2c/busses/i2c-amd756.c | 4653 | F: drivers/i2c/busses/i2c-amd756.c |
| 4653 | F: drivers/i2c/busses/i2c-amd756-s4882.c | 4654 | F: drivers/i2c/busses/i2c-amd756-s4882.c |
| 4654 | F: drivers/i2c/busses/i2c-amd8111.c | 4655 | F: drivers/i2c/busses/i2c-amd8111.c |
| 4655 | F: drivers/i2c/busses/i2c-i801.c | 4656 | F: drivers/i2c/busses/i2c-i801.c |
| 4656 | F: drivers/i2c/busses/i2c-isch.c | 4657 | F: drivers/i2c/busses/i2c-isch.c |
| 4657 | F: drivers/i2c/busses/i2c-nforce2.c | 4658 | F: drivers/i2c/busses/i2c-nforce2.c |
| 4658 | F: drivers/i2c/busses/i2c-nforce2-s4985.c | 4659 | F: drivers/i2c/busses/i2c-nforce2-s4985.c |
| 4659 | F: drivers/i2c/busses/i2c-piix4.c | 4660 | F: drivers/i2c/busses/i2c-piix4.c |
| 4660 | F: drivers/i2c/busses/i2c-sis5595.c | 4661 | F: drivers/i2c/busses/i2c-sis5595.c |
| 4661 | F: drivers/i2c/busses/i2c-sis630.c | 4662 | F: drivers/i2c/busses/i2c-sis630.c |
| 4662 | F: drivers/i2c/busses/i2c-sis96x.c | 4663 | F: drivers/i2c/busses/i2c-sis96x.c |
| 4663 | F: drivers/i2c/busses/i2c-via.c | 4664 | F: drivers/i2c/busses/i2c-via.c |
| 4664 | F: drivers/i2c/busses/i2c-viapro.c | 4665 | F: drivers/i2c/busses/i2c-viapro.c |
| 4665 | 4666 | ||
| 4666 | I2C/SMBUS ISMT DRIVER | 4667 | I2C/SMBUS ISMT DRIVER |
| 4667 | M: Seth Heasley <seth.heasley@intel.com> | 4668 | M: Seth Heasley <seth.heasley@intel.com> |
| 4668 | M: Neil Horman <nhorman@tuxdriver.com> | 4669 | M: Neil Horman <nhorman@tuxdriver.com> |
| 4669 | L: linux-i2c@vger.kernel.org | 4670 | L: linux-i2c@vger.kernel.org |
| 4670 | F: drivers/i2c/busses/i2c-ismt.c | 4671 | F: drivers/i2c/busses/i2c-ismt.c |
| 4671 | F: Documentation/i2c/busses/i2c-ismt | 4672 | F: Documentation/i2c/busses/i2c-ismt |
| 4672 | 4673 | ||
| 4673 | I2C/SMBUS STUB DRIVER | 4674 | I2C/SMBUS STUB DRIVER |
| 4674 | M: Jean Delvare <jdelvare@suse.de> | 4675 | M: Jean Delvare <jdelvare@suse.de> |
| 4675 | L: linux-i2c@vger.kernel.org | 4676 | L: linux-i2c@vger.kernel.org |
| 4676 | S: Maintained | 4677 | S: Maintained |
| 4677 | F: drivers/i2c/i2c-stub.c | 4678 | F: drivers/i2c/i2c-stub.c |
| 4678 | 4679 | ||
| 4679 | I2C SUBSYSTEM | 4680 | I2C SUBSYSTEM |
| 4680 | M: Wolfram Sang <wsa@the-dreams.de> | 4681 | M: Wolfram Sang <wsa@the-dreams.de> |
| 4681 | L: linux-i2c@vger.kernel.org | 4682 | L: linux-i2c@vger.kernel.org |
| 4682 | W: https://i2c.wiki.kernel.org/ | 4683 | W: https://i2c.wiki.kernel.org/ |
| 4683 | Q: https://patchwork.ozlabs.org/project/linux-i2c/list/ | 4684 | Q: https://patchwork.ozlabs.org/project/linux-i2c/list/ |
| 4684 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git | 4685 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git |
| 4685 | S: Maintained | 4686 | S: Maintained |
| 4686 | F: Documentation/devicetree/bindings/i2c/ | 4687 | F: Documentation/devicetree/bindings/i2c/ |
| 4687 | F: Documentation/i2c/ | 4688 | F: Documentation/i2c/ |
| 4688 | F: drivers/i2c/ | 4689 | F: drivers/i2c/ |
| 4689 | F: include/linux/i2c.h | 4690 | F: include/linux/i2c.h |
| 4690 | F: include/linux/i2c-*.h | 4691 | F: include/linux/i2c-*.h |
| 4691 | F: include/uapi/linux/i2c.h | 4692 | F: include/uapi/linux/i2c.h |
| 4692 | F: include/uapi/linux/i2c-*.h | 4693 | F: include/uapi/linux/i2c-*.h |
| 4693 | 4694 | ||
| 4694 | I2C ACPI SUPPORT | 4695 | I2C ACPI SUPPORT |
| 4695 | M: Mika Westerberg <mika.westerberg@linux.intel.com> | 4696 | M: Mika Westerberg <mika.westerberg@linux.intel.com> |
| 4696 | L: linux-i2c@vger.kernel.org | 4697 | L: linux-i2c@vger.kernel.org |
| 4697 | L: linux-acpi@vger.kernel.org | 4698 | L: linux-acpi@vger.kernel.org |
| 4698 | S: Maintained | 4699 | S: Maintained |
| 4699 | 4700 | ||
| 4700 | I2C-TAOS-EVM DRIVER | 4701 | I2C-TAOS-EVM DRIVER |
| 4701 | M: Jean Delvare <jdelvare@suse.de> | 4702 | M: Jean Delvare <jdelvare@suse.de> |
| 4702 | L: linux-i2c@vger.kernel.org | 4703 | L: linux-i2c@vger.kernel.org |
| 4703 | S: Maintained | 4704 | S: Maintained |
| 4704 | F: Documentation/i2c/busses/i2c-taos-evm | 4705 | F: Documentation/i2c/busses/i2c-taos-evm |
| 4705 | F: drivers/i2c/busses/i2c-taos-evm.c | 4706 | F: drivers/i2c/busses/i2c-taos-evm.c |
| 4706 | 4707 | ||
| 4707 | I2C-TINY-USB DRIVER | 4708 | I2C-TINY-USB DRIVER |
| 4708 | M: Till Harbaum <till@harbaum.org> | 4709 | M: Till Harbaum <till@harbaum.org> |
| 4709 | L: linux-i2c@vger.kernel.org | 4710 | L: linux-i2c@vger.kernel.org |
| 4710 | W: http://www.harbaum.org/till/i2c_tiny_usb | 4711 | W: http://www.harbaum.org/till/i2c_tiny_usb |
| 4711 | S: Maintained | 4712 | S: Maintained |
| 4712 | F: drivers/i2c/busses/i2c-tiny-usb.c | 4713 | F: drivers/i2c/busses/i2c-tiny-usb.c |
| 4713 | 4714 | ||
| 4714 | i386 BOOT CODE | 4715 | i386 BOOT CODE |
| 4715 | M: "H. Peter Anvin" <hpa@zytor.com> | 4716 | M: "H. Peter Anvin" <hpa@zytor.com> |
| 4716 | S: Maintained | 4717 | S: Maintained |
| 4717 | F: arch/x86/boot/ | 4718 | F: arch/x86/boot/ |
| 4718 | 4719 | ||
| 4719 | i386 SETUP CODE / CPU ERRATA WORKAROUNDS | 4720 | i386 SETUP CODE / CPU ERRATA WORKAROUNDS |
| 4720 | M: "H. Peter Anvin" <hpa@zytor.com> | 4721 | M: "H. Peter Anvin" <hpa@zytor.com> |
| 4721 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup.git | 4722 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup.git |
| 4722 | S: Maintained | 4723 | S: Maintained |
| 4723 | 4724 | ||
| 4724 | IA64 (Itanium) PLATFORM | 4725 | IA64 (Itanium) PLATFORM |
| 4725 | M: Tony Luck <tony.luck@intel.com> | 4726 | M: Tony Luck <tony.luck@intel.com> |
| 4726 | M: Fenghua Yu <fenghua.yu@intel.com> | 4727 | M: Fenghua Yu <fenghua.yu@intel.com> |
| 4727 | L: linux-ia64@vger.kernel.org | 4728 | L: linux-ia64@vger.kernel.org |
| 4728 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux.git | 4729 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux.git |
| 4729 | S: Maintained | 4730 | S: Maintained |
| 4730 | F: arch/ia64/ | 4731 | F: arch/ia64/ |
| 4731 | 4732 | ||
| 4732 | IBM Power in-Nest Crypto Acceleration | 4733 | IBM Power in-Nest Crypto Acceleration |
| 4733 | M: Marcelo Henrique Cerri <mhcerri@linux.vnet.ibm.com> | 4734 | M: Marcelo Henrique Cerri <mhcerri@linux.vnet.ibm.com> |
| 4734 | M: Fionnuala Gunter <fin@linux.vnet.ibm.com> | 4735 | M: Fionnuala Gunter <fin@linux.vnet.ibm.com> |
| 4735 | L: linux-crypto@vger.kernel.org | 4736 | L: linux-crypto@vger.kernel.org |
| 4736 | S: Supported | 4737 | S: Supported |
| 4737 | F: drivers/crypto/nx/ | 4738 | F: drivers/crypto/nx/ |
| 4738 | 4739 | ||
| 4739 | IBM Power 842 compression accelerator | 4740 | IBM Power 842 compression accelerator |
| 4740 | M: Dan Streetman <ddstreet@us.ibm.com> | 4741 | M: Dan Streetman <ddstreet@us.ibm.com> |
| 4741 | S: Supported | 4742 | S: Supported |
| 4742 | F: drivers/crypto/nx/nx-842.c | 4743 | F: drivers/crypto/nx/nx-842.c |
| 4743 | F: include/linux/nx842.h | 4744 | F: include/linux/nx842.h |
| 4744 | 4745 | ||
| 4745 | IBM Power Linux RAID adapter | 4746 | IBM Power Linux RAID adapter |
| 4746 | M: Brian King <brking@us.ibm.com> | 4747 | M: Brian King <brking@us.ibm.com> |
| 4747 | S: Supported | 4748 | S: Supported |
| 4748 | F: drivers/scsi/ipr.* | 4749 | F: drivers/scsi/ipr.* |
| 4749 | 4750 | ||
| 4750 | IBM Power Virtual Ethernet Device Driver | 4751 | IBM Power Virtual Ethernet Device Driver |
| 4751 | M: Santiago Leon <santil@linux.vnet.ibm.com> | 4752 | M: Santiago Leon <santil@linux.vnet.ibm.com> |
| 4752 | L: netdev@vger.kernel.org | 4753 | L: netdev@vger.kernel.org |
| 4753 | S: Supported | 4754 | S: Supported |
| 4754 | F: drivers/net/ethernet/ibm/ibmveth.* | 4755 | F: drivers/net/ethernet/ibm/ibmveth.* |
| 4755 | 4756 | ||
| 4756 | IBM Power Virtual SCSI Device Drivers | 4757 | IBM Power Virtual SCSI Device Drivers |
| 4757 | M: Nathan Fontenot <nfont@linux.vnet.ibm.com> | 4758 | M: Nathan Fontenot <nfont@linux.vnet.ibm.com> |
| 4758 | L: linux-scsi@vger.kernel.org | 4759 | L: linux-scsi@vger.kernel.org |
| 4759 | S: Supported | 4760 | S: Supported |
| 4760 | F: drivers/scsi/ibmvscsi/ibmvscsi* | 4761 | F: drivers/scsi/ibmvscsi/ibmvscsi* |
| 4761 | F: drivers/scsi/ibmvscsi/viosrp.h | 4762 | F: drivers/scsi/ibmvscsi/viosrp.h |
| 4762 | 4763 | ||
| 4763 | IBM Power Virtual FC Device Drivers | 4764 | IBM Power Virtual FC Device Drivers |
| 4764 | M: Brian King <brking@linux.vnet.ibm.com> | 4765 | M: Brian King <brking@linux.vnet.ibm.com> |
| 4765 | L: linux-scsi@vger.kernel.org | 4766 | L: linux-scsi@vger.kernel.org |
| 4766 | S: Supported | 4767 | S: Supported |
| 4767 | F: drivers/scsi/ibmvscsi/ibmvfc* | 4768 | F: drivers/scsi/ibmvscsi/ibmvfc* |
| 4768 | 4769 | ||
| 4769 | IBM ServeRAID RAID DRIVER | 4770 | IBM ServeRAID RAID DRIVER |
| 4770 | S: Orphan | 4771 | S: Orphan |
| 4771 | F: drivers/scsi/ips.* | 4772 | F: drivers/scsi/ips.* |
| 4772 | 4773 | ||
| 4773 | ICH LPC AND GPIO DRIVER | 4774 | ICH LPC AND GPIO DRIVER |
| 4774 | M: Peter Tyser <ptyser@xes-inc.com> | 4775 | M: Peter Tyser <ptyser@xes-inc.com> |
| 4775 | S: Maintained | 4776 | S: Maintained |
| 4776 | F: drivers/mfd/lpc_ich.c | 4777 | F: drivers/mfd/lpc_ich.c |
| 4777 | F: drivers/gpio/gpio-ich.c | 4778 | F: drivers/gpio/gpio-ich.c |
| 4778 | 4779 | ||
| 4779 | IDE SUBSYSTEM | 4780 | IDE SUBSYSTEM |
| 4780 | M: "David S. Miller" <davem@davemloft.net> | 4781 | M: "David S. Miller" <davem@davemloft.net> |
| 4781 | L: linux-ide@vger.kernel.org | 4782 | L: linux-ide@vger.kernel.org |
| 4782 | Q: http://patchwork.ozlabs.org/project/linux-ide/list/ | 4783 | Q: http://patchwork.ozlabs.org/project/linux-ide/list/ |
| 4783 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide.git | 4784 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide.git |
| 4784 | S: Maintained | 4785 | S: Maintained |
| 4785 | F: Documentation/ide/ | 4786 | F: Documentation/ide/ |
| 4786 | F: drivers/ide/ | 4787 | F: drivers/ide/ |
| 4787 | F: include/linux/ide.h | 4788 | F: include/linux/ide.h |
| 4788 | 4789 | ||
| 4789 | IDEAPAD LAPTOP EXTRAS DRIVER | 4790 | IDEAPAD LAPTOP EXTRAS DRIVER |
| 4790 | M: Ike Panhc <ike.pan@canonical.com> | 4791 | M: Ike Panhc <ike.pan@canonical.com> |
| 4791 | L: platform-driver-x86@vger.kernel.org | 4792 | L: platform-driver-x86@vger.kernel.org |
| 4792 | W: http://launchpad.net/ideapad-laptop | 4793 | W: http://launchpad.net/ideapad-laptop |
| 4793 | S: Maintained | 4794 | S: Maintained |
| 4794 | F: drivers/platform/x86/ideapad-laptop.c | 4795 | F: drivers/platform/x86/ideapad-laptop.c |
| 4795 | 4796 | ||
| 4796 | IDEAPAD LAPTOP SLIDEBAR DRIVER | 4797 | IDEAPAD LAPTOP SLIDEBAR DRIVER |
| 4797 | M: Andrey Moiseev <o2g.org.ru@gmail.com> | 4798 | M: Andrey Moiseev <o2g.org.ru@gmail.com> |
| 4798 | L: linux-input@vger.kernel.org | 4799 | L: linux-input@vger.kernel.org |
| 4799 | W: https://github.com/o2genum/ideapad-slidebar | 4800 | W: https://github.com/o2genum/ideapad-slidebar |
| 4800 | S: Maintained | 4801 | S: Maintained |
| 4801 | F: drivers/input/misc/ideapad_slidebar.c | 4802 | F: drivers/input/misc/ideapad_slidebar.c |
| 4802 | 4803 | ||
| 4803 | IDE/ATAPI DRIVERS | 4804 | IDE/ATAPI DRIVERS |
| 4804 | M: Borislav Petkov <bp@alien8.de> | 4805 | M: Borislav Petkov <bp@alien8.de> |
| 4805 | L: linux-ide@vger.kernel.org | 4806 | L: linux-ide@vger.kernel.org |
| 4806 | S: Maintained | 4807 | S: Maintained |
| 4807 | F: Documentation/cdrom/ide-cd | 4808 | F: Documentation/cdrom/ide-cd |
| 4808 | F: drivers/ide/ide-cd* | 4809 | F: drivers/ide/ide-cd* |
| 4809 | 4810 | ||
| 4810 | IDLE-I7300 | 4811 | IDLE-I7300 |
| 4811 | M: Andy Henroid <andrew.d.henroid@intel.com> | 4812 | M: Andy Henroid <andrew.d.henroid@intel.com> |
| 4812 | L: linux-pm@vger.kernel.org | 4813 | L: linux-pm@vger.kernel.org |
| 4813 | S: Supported | 4814 | S: Supported |
| 4814 | F: drivers/idle/i7300_idle.c | 4815 | F: drivers/idle/i7300_idle.c |
| 4815 | 4816 | ||
| 4816 | IEEE 802.15.4 SUBSYSTEM | 4817 | IEEE 802.15.4 SUBSYSTEM |
| 4817 | M: Alexander Aring <alex.aring@gmail.com> | 4818 | M: Alexander Aring <alex.aring@gmail.com> |
| 4818 | L: linux-wpan@vger.kernel.org | 4819 | L: linux-wpan@vger.kernel.org |
| 4819 | W: https://github.com/linux-wpan | 4820 | W: https://github.com/linux-wpan |
| 4820 | T: git git://github.com/linux-wpan/linux-wpan-next.git | 4821 | T: git git://github.com/linux-wpan/linux-wpan-next.git |
| 4821 | S: Maintained | 4822 | S: Maintained |
| 4822 | F: net/ieee802154/ | 4823 | F: net/ieee802154/ |
| 4823 | F: net/mac802154/ | 4824 | F: net/mac802154/ |
| 4824 | F: drivers/net/ieee802154/ | 4825 | F: drivers/net/ieee802154/ |
| 4825 | F: include/linux/nl802154.h | 4826 | F: include/linux/nl802154.h |
| 4826 | F: include/linux/ieee802154.h | 4827 | F: include/linux/ieee802154.h |
| 4827 | F: include/net/nl802154.h | 4828 | F: include/net/nl802154.h |
| 4828 | F: include/net/mac802154.h | 4829 | F: include/net/mac802154.h |
| 4829 | F: include/net/af_ieee802154.h | 4830 | F: include/net/af_ieee802154.h |
| 4830 | F: include/net/cfg802154.h | 4831 | F: include/net/cfg802154.h |
| 4831 | F: include/net/ieee802154_netdev.h | 4832 | F: include/net/ieee802154_netdev.h |
| 4832 | F: Documentation/networking/ieee802154.txt | 4833 | F: Documentation/networking/ieee802154.txt |
| 4833 | 4834 | ||
| 4834 | IGORPLUG-USB IR RECEIVER | 4835 | IGORPLUG-USB IR RECEIVER |
| 4835 | M: Sean Young <sean@mess.org> | 4836 | M: Sean Young <sean@mess.org> |
| 4836 | L: linux-media@vger.kernel.org | 4837 | L: linux-media@vger.kernel.org |
| 4837 | S: Maintained | 4838 | S: Maintained |
| 4838 | F: drivers/media/rc/igorplugusb.c | 4839 | F: drivers/media/rc/igorplugusb.c |
| 4839 | 4840 | ||
| 4840 | IGUANAWORKS USB IR TRANSCEIVER | 4841 | IGUANAWORKS USB IR TRANSCEIVER |
| 4841 | M: Sean Young <sean@mess.org> | 4842 | M: Sean Young <sean@mess.org> |
| 4842 | L: linux-media@vger.kernel.org | 4843 | L: linux-media@vger.kernel.org |
| 4843 | S: Maintained | 4844 | S: Maintained |
| 4844 | F: drivers/media/rc/iguanair.c | 4845 | F: drivers/media/rc/iguanair.c |
| 4845 | 4846 | ||
| 4846 | IIO SUBSYSTEM AND DRIVERS | 4847 | IIO SUBSYSTEM AND DRIVERS |
| 4847 | M: Jonathan Cameron <jic23@kernel.org> | 4848 | M: Jonathan Cameron <jic23@kernel.org> |
| 4848 | R: Hartmut Knaack <knaack.h@gmx.de> | 4849 | R: Hartmut Knaack <knaack.h@gmx.de> |
| 4849 | R: Lars-Peter Clausen <lars@metafoo.de> | 4850 | R: Lars-Peter Clausen <lars@metafoo.de> |
| 4850 | R: Peter Meerwald <pmeerw@pmeerw.net> | 4851 | R: Peter Meerwald <pmeerw@pmeerw.net> |
| 4851 | L: linux-iio@vger.kernel.org | 4852 | L: linux-iio@vger.kernel.org |
| 4852 | S: Maintained | 4853 | S: Maintained |
| 4853 | F: drivers/iio/ | 4854 | F: drivers/iio/ |
| 4854 | F: drivers/staging/iio/ | 4855 | F: drivers/staging/iio/ |
| 4855 | F: include/linux/iio/ | 4856 | F: include/linux/iio/ |
| 4856 | 4857 | ||
| 4857 | IKANOS/ADI EAGLE ADSL USB DRIVER | 4858 | IKANOS/ADI EAGLE ADSL USB DRIVER |
| 4858 | M: Matthieu Castet <castet.matthieu@free.fr> | 4859 | M: Matthieu Castet <castet.matthieu@free.fr> |
| 4859 | M: Stanislaw Gruszka <stf_xl@wp.pl> | 4860 | M: Stanislaw Gruszka <stf_xl@wp.pl> |
| 4860 | S: Maintained | 4861 | S: Maintained |
| 4861 | F: drivers/usb/atm/ueagle-atm.c | 4862 | F: drivers/usb/atm/ueagle-atm.c |
| 4862 | 4863 | ||
| 4863 | INA209 HARDWARE MONITOR DRIVER | 4864 | INA209 HARDWARE MONITOR DRIVER |
| 4864 | M: Guenter Roeck <linux@roeck-us.net> | 4865 | M: Guenter Roeck <linux@roeck-us.net> |
| 4865 | L: lm-sensors@lm-sensors.org | 4866 | L: lm-sensors@lm-sensors.org |
| 4866 | S: Maintained | 4867 | S: Maintained |
| 4867 | F: Documentation/hwmon/ina209 | 4868 | F: Documentation/hwmon/ina209 |
| 4868 | F: Documentation/devicetree/bindings/i2c/ina209.txt | 4869 | F: Documentation/devicetree/bindings/i2c/ina209.txt |
| 4869 | F: drivers/hwmon/ina209.c | 4870 | F: drivers/hwmon/ina209.c |
| 4870 | 4871 | ||
| 4871 | INA2XX HARDWARE MONITOR DRIVER | 4872 | INA2XX HARDWARE MONITOR DRIVER |
| 4872 | M: Guenter Roeck <linux@roeck-us.net> | 4873 | M: Guenter Roeck <linux@roeck-us.net> |
| 4873 | L: lm-sensors@lm-sensors.org | 4874 | L: lm-sensors@lm-sensors.org |
| 4874 | S: Maintained | 4875 | S: Maintained |
| 4875 | F: Documentation/hwmon/ina2xx | 4876 | F: Documentation/hwmon/ina2xx |
| 4876 | F: drivers/hwmon/ina2xx.c | 4877 | F: drivers/hwmon/ina2xx.c |
| 4877 | F: include/linux/platform_data/ina2xx.h | 4878 | F: include/linux/platform_data/ina2xx.h |
| 4878 | 4879 | ||
| 4879 | INDUSTRY PACK SUBSYSTEM (IPACK) | 4880 | INDUSTRY PACK SUBSYSTEM (IPACK) |
| 4880 | M: Samuel Iglesias Gonsalvez <siglesias@igalia.com> | 4881 | M: Samuel Iglesias Gonsalvez <siglesias@igalia.com> |
| 4881 | M: Jens Taprogge <jens.taprogge@taprogge.org> | 4882 | M: Jens Taprogge <jens.taprogge@taprogge.org> |
| 4882 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 4883 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 4883 | L: industrypack-devel@lists.sourceforge.net | 4884 | L: industrypack-devel@lists.sourceforge.net |
| 4884 | W: http://industrypack.sourceforge.net | 4885 | W: http://industrypack.sourceforge.net |
| 4885 | S: Maintained | 4886 | S: Maintained |
| 4886 | F: drivers/ipack/ | 4887 | F: drivers/ipack/ |
| 4887 | 4888 | ||
| 4888 | INTEGRITY MEASUREMENT ARCHITECTURE (IMA) | 4889 | INTEGRITY MEASUREMENT ARCHITECTURE (IMA) |
| 4889 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> | 4890 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> |
| 4890 | M: Dmitry Kasatkin <d.kasatkin@samsung.com> | 4891 | M: Dmitry Kasatkin <d.kasatkin@samsung.com> |
| 4891 | L: linux-ima-devel@lists.sourceforge.net | 4892 | L: linux-ima-devel@lists.sourceforge.net |
| 4892 | L: linux-ima-user@lists.sourceforge.net | 4893 | L: linux-ima-user@lists.sourceforge.net |
| 4893 | L: linux-security-module@vger.kernel.org | 4894 | L: linux-security-module@vger.kernel.org |
| 4894 | S: Supported | 4895 | S: Supported |
| 4895 | F: security/integrity/ima/ | 4896 | F: security/integrity/ima/ |
| 4896 | 4897 | ||
| 4897 | IMGTEC IR DECODER DRIVER | 4898 | IMGTEC IR DECODER DRIVER |
| 4898 | M: James Hogan <james.hogan@imgtec.com> | 4899 | M: James Hogan <james.hogan@imgtec.com> |
| 4899 | S: Maintained | 4900 | S: Maintained |
| 4900 | F: drivers/media/rc/img-ir/ | 4901 | F: drivers/media/rc/img-ir/ |
| 4901 | 4902 | ||
| 4902 | IMS TWINTURBO FRAMEBUFFER DRIVER | 4903 | IMS TWINTURBO FRAMEBUFFER DRIVER |
| 4903 | L: linux-fbdev@vger.kernel.org | 4904 | L: linux-fbdev@vger.kernel.org |
| 4904 | S: Orphan | 4905 | S: Orphan |
| 4905 | F: drivers/video/fbdev/imsttfb.c | 4906 | F: drivers/video/fbdev/imsttfb.c |
| 4906 | 4907 | ||
| 4907 | INFINIBAND SUBSYSTEM | 4908 | INFINIBAND SUBSYSTEM |
| 4908 | M: Roland Dreier <roland@kernel.org> | 4909 | M: Roland Dreier <roland@kernel.org> |
| 4909 | M: Sean Hefty <sean.hefty@intel.com> | 4910 | M: Sean Hefty <sean.hefty@intel.com> |
| 4910 | M: Hal Rosenstock <hal.rosenstock@gmail.com> | 4911 | M: Hal Rosenstock <hal.rosenstock@gmail.com> |
| 4911 | L: linux-rdma@vger.kernel.org | 4912 | L: linux-rdma@vger.kernel.org |
| 4912 | W: http://www.openfabrics.org/ | 4913 | W: http://www.openfabrics.org/ |
| 4913 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ | 4914 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ |
| 4914 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git | 4915 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband.git |
| 4915 | S: Supported | 4916 | S: Supported |
| 4916 | F: Documentation/infiniband/ | 4917 | F: Documentation/infiniband/ |
| 4917 | F: drivers/infiniband/ | 4918 | F: drivers/infiniband/ |
| 4918 | F: include/uapi/linux/if_infiniband.h | 4919 | F: include/uapi/linux/if_infiniband.h |
| 4919 | 4920 | ||
| 4920 | INOTIFY | 4921 | INOTIFY |
| 4921 | M: John McCutchan <john@johnmccutchan.com> | 4922 | M: John McCutchan <john@johnmccutchan.com> |
| 4922 | M: Robert Love <rlove@rlove.org> | 4923 | M: Robert Love <rlove@rlove.org> |
| 4923 | M: Eric Paris <eparis@parisplace.org> | 4924 | M: Eric Paris <eparis@parisplace.org> |
| 4924 | S: Maintained | 4925 | S: Maintained |
| 4925 | F: Documentation/filesystems/inotify.txt | 4926 | F: Documentation/filesystems/inotify.txt |
| 4926 | F: fs/notify/inotify/ | 4927 | F: fs/notify/inotify/ |
| 4927 | F: include/linux/inotify.h | 4928 | F: include/linux/inotify.h |
| 4928 | F: include/uapi/linux/inotify.h | 4929 | F: include/uapi/linux/inotify.h |
| 4929 | 4930 | ||
| 4930 | INPUT (KEYBOARD, MOUSE, JOYSTICK, TOUCHSCREEN) DRIVERS | 4931 | INPUT (KEYBOARD, MOUSE, JOYSTICK, TOUCHSCREEN) DRIVERS |
| 4931 | M: Dmitry Torokhov <dmitry.torokhov@gmail.com> | 4932 | M: Dmitry Torokhov <dmitry.torokhov@gmail.com> |
| 4932 | M: Dmitry Torokhov <dtor@mail.ru> | 4933 | M: Dmitry Torokhov <dtor@mail.ru> |
| 4933 | L: linux-input@vger.kernel.org | 4934 | L: linux-input@vger.kernel.org |
| 4934 | Q: http://patchwork.kernel.org/project/linux-input/list/ | 4935 | Q: http://patchwork.kernel.org/project/linux-input/list/ |
| 4935 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input.git | 4936 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input.git |
| 4936 | S: Maintained | 4937 | S: Maintained |
| 4937 | F: drivers/input/ | 4938 | F: drivers/input/ |
| 4938 | F: include/linux/input.h | 4939 | F: include/linux/input.h |
| 4939 | F: include/uapi/linux/input.h | 4940 | F: include/uapi/linux/input.h |
| 4940 | F: include/linux/input/ | 4941 | F: include/linux/input/ |
| 4941 | 4942 | ||
| 4942 | INPUT MULTITOUCH (MT) PROTOCOL | 4943 | INPUT MULTITOUCH (MT) PROTOCOL |
| 4943 | M: Henrik Rydberg <rydberg@bitmath.org> | 4944 | M: Henrik Rydberg <rydberg@bitmath.org> |
| 4944 | L: linux-input@vger.kernel.org | 4945 | L: linux-input@vger.kernel.org |
| 4945 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rydberg/input-mt.git | 4946 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rydberg/input-mt.git |
| 4946 | S: Odd fixes | 4947 | S: Odd fixes |
| 4947 | F: Documentation/input/multi-touch-protocol.txt | 4948 | F: Documentation/input/multi-touch-protocol.txt |
| 4948 | F: drivers/input/input-mt.c | 4949 | F: drivers/input/input-mt.c |
| 4949 | K: \b(ABS|SYN)_MT_ | 4950 | K: \b(ABS|SYN)_MT_ |
| 4950 | 4951 | ||
| 4951 | INTEL C600 SERIES SAS CONTROLLER DRIVER | 4952 | INTEL C600 SERIES SAS CONTROLLER DRIVER |
| 4952 | M: Intel SCU Linux support <intel-linux-scu@intel.com> | 4953 | M: Intel SCU Linux support <intel-linux-scu@intel.com> |
| 4953 | M: Artur Paszkiewicz <artur.paszkiewicz@intel.com> | 4954 | M: Artur Paszkiewicz <artur.paszkiewicz@intel.com> |
| 4954 | M: Dave Jiang <dave.jiang@intel.com> | 4955 | M: Dave Jiang <dave.jiang@intel.com> |
| 4955 | L: linux-scsi@vger.kernel.org | 4956 | L: linux-scsi@vger.kernel.org |
| 4956 | T: git git://git.code.sf.net/p/intel-sas/isci | 4957 | T: git git://git.code.sf.net/p/intel-sas/isci |
| 4957 | S: Supported | 4958 | S: Supported |
| 4958 | F: drivers/scsi/isci/ | 4959 | F: drivers/scsi/isci/ |
| 4959 | 4960 | ||
| 4960 | INTEL IDLE DRIVER | 4961 | INTEL IDLE DRIVER |
| 4961 | M: Len Brown <lenb@kernel.org> | 4962 | M: Len Brown <lenb@kernel.org> |
| 4962 | L: linux-pm@vger.kernel.org | 4963 | L: linux-pm@vger.kernel.org |
| 4963 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux.git | 4964 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux.git |
| 4964 | S: Supported | 4965 | S: Supported |
| 4965 | F: drivers/idle/intel_idle.c | 4966 | F: drivers/idle/intel_idle.c |
| 4966 | 4967 | ||
| 4967 | INTEL PSTATE DRIVER | 4968 | INTEL PSTATE DRIVER |
| 4968 | M: Kristen Carlson Accardi <kristen@linux.intel.com> | 4969 | M: Kristen Carlson Accardi <kristen@linux.intel.com> |
| 4969 | L: linux-pm@vger.kernel.org | 4970 | L: linux-pm@vger.kernel.org |
| 4970 | S: Supported | 4971 | S: Supported |
| 4971 | F: drivers/cpufreq/intel_pstate.c | 4972 | F: drivers/cpufreq/intel_pstate.c |
| 4972 | 4973 | ||
| 4973 | INTEL FRAMEBUFFER DRIVER (excluding 810 and 815) | 4974 | INTEL FRAMEBUFFER DRIVER (excluding 810 and 815) |
| 4974 | M: Maik Broemme <mbroemme@plusserver.de> | 4975 | M: Maik Broemme <mbroemme@plusserver.de> |
| 4975 | L: linux-fbdev@vger.kernel.org | 4976 | L: linux-fbdev@vger.kernel.org |
| 4976 | S: Maintained | 4977 | S: Maintained |
| 4977 | F: Documentation/fb/intelfb.txt | 4978 | F: Documentation/fb/intelfb.txt |
| 4978 | F: drivers/video/fbdev/intelfb/ | 4979 | F: drivers/video/fbdev/intelfb/ |
| 4979 | 4980 | ||
| 4980 | INTEL 810/815 FRAMEBUFFER DRIVER | 4981 | INTEL 810/815 FRAMEBUFFER DRIVER |
| 4981 | M: Antonino Daplas <adaplas@gmail.com> | 4982 | M: Antonino Daplas <adaplas@gmail.com> |
| 4982 | L: linux-fbdev@vger.kernel.org | 4983 | L: linux-fbdev@vger.kernel.org |
| 4983 | S: Maintained | 4984 | S: Maintained |
| 4984 | F: drivers/video/fbdev/i810/ | 4985 | F: drivers/video/fbdev/i810/ |
| 4985 | 4986 | ||
| 4986 | INTEL MENLOW THERMAL DRIVER | 4987 | INTEL MENLOW THERMAL DRIVER |
| 4987 | M: Sujith Thomas <sujith.thomas@intel.com> | 4988 | M: Sujith Thomas <sujith.thomas@intel.com> |
| 4988 | L: platform-driver-x86@vger.kernel.org | 4989 | L: platform-driver-x86@vger.kernel.org |
| 4989 | W: https://01.org/linux-acpi | 4990 | W: https://01.org/linux-acpi |
| 4990 | S: Supported | 4991 | S: Supported |
| 4991 | F: drivers/platform/x86/intel_menlow.c | 4992 | F: drivers/platform/x86/intel_menlow.c |
| 4992 | 4993 | ||
| 4993 | INTEL IA32 MICROCODE UPDATE SUPPORT | 4994 | INTEL IA32 MICROCODE UPDATE SUPPORT |
| 4994 | M: Tigran Aivazian <tigran@aivazian.fsnet.co.uk> | 4995 | M: Tigran Aivazian <tigran@aivazian.fsnet.co.uk> |
| 4995 | S: Maintained | 4996 | S: Maintained |
| 4996 | F: arch/x86/kernel/cpu/microcode/core* | 4997 | F: arch/x86/kernel/cpu/microcode/core* |
| 4997 | F: arch/x86/kernel/cpu/microcode/intel* | 4998 | F: arch/x86/kernel/cpu/microcode/intel* |
| 4998 | 4999 | ||
| 4999 | INTEL I/OAT DMA DRIVER | 5000 | INTEL I/OAT DMA DRIVER |
| 5000 | M: Dave Jiang <dave.jiang@intel.com> | 5001 | M: Dave Jiang <dave.jiang@intel.com> |
| 5001 | R: Dan Williams <dan.j.williams@intel.com> | 5002 | R: Dan Williams <dan.j.williams@intel.com> |
| 5002 | L: dmaengine@vger.kernel.org | 5003 | L: dmaengine@vger.kernel.org |
| 5003 | Q: https://patchwork.kernel.org/project/linux-dmaengine/list/ | 5004 | Q: https://patchwork.kernel.org/project/linux-dmaengine/list/ |
| 5004 | S: Supported | 5005 | S: Supported |
| 5005 | F: drivers/dma/ioat* | 5006 | F: drivers/dma/ioat* |
| 5006 | 5007 | ||
| 5007 | INTEL IOMMU (VT-d) | 5008 | INTEL IOMMU (VT-d) |
| 5008 | M: David Woodhouse <dwmw2@infradead.org> | 5009 | M: David Woodhouse <dwmw2@infradead.org> |
| 5009 | L: iommu@lists.linux-foundation.org | 5010 | L: iommu@lists.linux-foundation.org |
| 5010 | T: git git://git.infradead.org/iommu-2.6.git | 5011 | T: git git://git.infradead.org/iommu-2.6.git |
| 5011 | S: Supported | 5012 | S: Supported |
| 5012 | F: drivers/iommu/intel-iommu.c | 5013 | F: drivers/iommu/intel-iommu.c |
| 5013 | F: include/linux/intel-iommu.h | 5014 | F: include/linux/intel-iommu.h |
| 5014 | 5015 | ||
| 5015 | INTEL IOP-ADMA DMA DRIVER | 5016 | INTEL IOP-ADMA DMA DRIVER |
| 5016 | R: Dan Williams <dan.j.williams@intel.com> | 5017 | R: Dan Williams <dan.j.williams@intel.com> |
| 5017 | S: Odd fixes | 5018 | S: Odd fixes |
| 5018 | F: drivers/dma/iop-adma.c | 5019 | F: drivers/dma/iop-adma.c |
| 5019 | 5020 | ||
| 5020 | INTEL IXP4XX QMGR, NPE, ETHERNET and HSS SUPPORT | 5021 | INTEL IXP4XX QMGR, NPE, ETHERNET and HSS SUPPORT |
| 5021 | M: Krzysztof Halasa <khalasa@piap.pl> | 5022 | M: Krzysztof Halasa <khalasa@piap.pl> |
| 5022 | S: Maintained | 5023 | S: Maintained |
| 5023 | F: arch/arm/mach-ixp4xx/include/mach/qmgr.h | 5024 | F: arch/arm/mach-ixp4xx/include/mach/qmgr.h |
| 5024 | F: arch/arm/mach-ixp4xx/include/mach/npe.h | 5025 | F: arch/arm/mach-ixp4xx/include/mach/npe.h |
| 5025 | F: arch/arm/mach-ixp4xx/ixp4xx_qmgr.c | 5026 | F: arch/arm/mach-ixp4xx/ixp4xx_qmgr.c |
| 5026 | F: arch/arm/mach-ixp4xx/ixp4xx_npe.c | 5027 | F: arch/arm/mach-ixp4xx/ixp4xx_npe.c |
| 5027 | F: drivers/net/ethernet/xscale/ixp4xx_eth.c | 5028 | F: drivers/net/ethernet/xscale/ixp4xx_eth.c |
| 5028 | F: drivers/net/wan/ixp4xx_hss.c | 5029 | F: drivers/net/wan/ixp4xx_hss.c |
| 5029 | 5030 | ||
| 5030 | INTEL IXP4XX RANDOM NUMBER GENERATOR SUPPORT | 5031 | INTEL IXP4XX RANDOM NUMBER GENERATOR SUPPORT |
| 5031 | M: Deepak Saxena <dsaxena@plexity.net> | 5032 | M: Deepak Saxena <dsaxena@plexity.net> |
| 5032 | S: Maintained | 5033 | S: Maintained |
| 5033 | F: drivers/char/hw_random/ixp4xx-rng.c | 5034 | F: drivers/char/hw_random/ixp4xx-rng.c |
| 5034 | 5035 | ||
| 5035 | INTEL ETHERNET DRIVERS (e100/e1000/e1000e/fm10k/igb/igbvf/ixgb/ixgbe/ixgbevf/i40e/i40evf) | 5036 | INTEL ETHERNET DRIVERS (e100/e1000/e1000e/fm10k/igb/igbvf/ixgb/ixgbe/ixgbevf/i40e/i40evf) |
| 5036 | M: Jeff Kirsher <jeffrey.t.kirsher@intel.com> | 5037 | M: Jeff Kirsher <jeffrey.t.kirsher@intel.com> |
| 5037 | M: Jesse Brandeburg <jesse.brandeburg@intel.com> | 5038 | M: Jesse Brandeburg <jesse.brandeburg@intel.com> |
| 5038 | M: Bruce Allan <bruce.w.allan@intel.com> | 5039 | M: Bruce Allan <bruce.w.allan@intel.com> |
| 5039 | M: Carolyn Wyborny <carolyn.wyborny@intel.com> | 5040 | M: Carolyn Wyborny <carolyn.wyborny@intel.com> |
| 5040 | M: Don Skidmore <donald.c.skidmore@intel.com> | 5041 | M: Don Skidmore <donald.c.skidmore@intel.com> |
| 5041 | M: Greg Rose <gregory.v.rose@intel.com> | 5042 | M: Greg Rose <gregory.v.rose@intel.com> |
| 5042 | M: Matthew Vick <matthew.vick@intel.com> | 5043 | M: Matthew Vick <matthew.vick@intel.com> |
| 5043 | M: John Ronciak <john.ronciak@intel.com> | 5044 | M: John Ronciak <john.ronciak@intel.com> |
| 5044 | M: Mitch Williams <mitch.a.williams@intel.com> | 5045 | M: Mitch Williams <mitch.a.williams@intel.com> |
| 5045 | M: Linux NICS <linux.nics@intel.com> | 5046 | M: Linux NICS <linux.nics@intel.com> |
| 5046 | L: e1000-devel@lists.sourceforge.net | 5047 | L: e1000-devel@lists.sourceforge.net |
| 5047 | W: http://www.intel.com/support/feedback.htm | 5048 | W: http://www.intel.com/support/feedback.htm |
| 5048 | W: http://e1000.sourceforge.net/ | 5049 | W: http://e1000.sourceforge.net/ |
| 5049 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net.git | 5050 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net.git |
| 5050 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next.git | 5051 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next.git |
| 5051 | S: Supported | 5052 | S: Supported |
| 5052 | F: Documentation/networking/e100.txt | 5053 | F: Documentation/networking/e100.txt |
| 5053 | F: Documentation/networking/e1000.txt | 5054 | F: Documentation/networking/e1000.txt |
| 5054 | F: Documentation/networking/e1000e.txt | 5055 | F: Documentation/networking/e1000e.txt |
| 5055 | F: Documentation/networking/igb.txt | 5056 | F: Documentation/networking/igb.txt |
| 5056 | F: Documentation/networking/igbvf.txt | 5057 | F: Documentation/networking/igbvf.txt |
| 5057 | F: Documentation/networking/ixgb.txt | 5058 | F: Documentation/networking/ixgb.txt |
| 5058 | F: Documentation/networking/ixgbe.txt | 5059 | F: Documentation/networking/ixgbe.txt |
| 5059 | F: Documentation/networking/ixgbevf.txt | 5060 | F: Documentation/networking/ixgbevf.txt |
| 5060 | F: Documentation/networking/i40e.txt | 5061 | F: Documentation/networking/i40e.txt |
| 5061 | F: Documentation/networking/i40evf.txt | 5062 | F: Documentation/networking/i40evf.txt |
| 5062 | F: drivers/net/ethernet/intel/ | 5063 | F: drivers/net/ethernet/intel/ |
| 5063 | F: drivers/net/ethernet/intel/*/ | 5064 | F: drivers/net/ethernet/intel/*/ |
| 5064 | 5065 | ||
| 5065 | INTEL-MID GPIO DRIVER | 5066 | INTEL-MID GPIO DRIVER |
| 5066 | M: David Cohen <david.a.cohen@linux.intel.com> | 5067 | M: David Cohen <david.a.cohen@linux.intel.com> |
| 5067 | L: linux-gpio@vger.kernel.org | 5068 | L: linux-gpio@vger.kernel.org |
| 5068 | S: Maintained | 5069 | S: Maintained |
| 5069 | F: drivers/gpio/gpio-intel-mid.c | 5070 | F: drivers/gpio/gpio-intel-mid.c |
| 5070 | 5071 | ||
| 5071 | INTEL PRO/WIRELESS 2100, 2200BG, 2915ABG NETWORK CONNECTION SUPPORT | 5072 | INTEL PRO/WIRELESS 2100, 2200BG, 2915ABG NETWORK CONNECTION SUPPORT |
| 5072 | M: Stanislav Yakovlev <stas.yakovlev@gmail.com> | 5073 | M: Stanislav Yakovlev <stas.yakovlev@gmail.com> |
| 5073 | L: linux-wireless@vger.kernel.org | 5074 | L: linux-wireless@vger.kernel.org |
| 5074 | S: Maintained | 5075 | S: Maintained |
| 5075 | F: Documentation/networking/README.ipw2100 | 5076 | F: Documentation/networking/README.ipw2100 |
| 5076 | F: Documentation/networking/README.ipw2200 | 5077 | F: Documentation/networking/README.ipw2200 |
| 5077 | F: drivers/net/wireless/ipw2x00/ | 5078 | F: drivers/net/wireless/ipw2x00/ |
| 5078 | 5079 | ||
| 5079 | INTEL(R) TRUSTED EXECUTION TECHNOLOGY (TXT) | 5080 | INTEL(R) TRUSTED EXECUTION TECHNOLOGY (TXT) |
| 5080 | M: Richard L Maliszewski <richard.l.maliszewski@intel.com> | 5081 | M: Richard L Maliszewski <richard.l.maliszewski@intel.com> |
| 5081 | M: Gang Wei <gang.wei@intel.com> | 5082 | M: Gang Wei <gang.wei@intel.com> |
| 5082 | M: Shane Wang <shane.wang@intel.com> | 5083 | M: Shane Wang <shane.wang@intel.com> |
| 5083 | L: tboot-devel@lists.sourceforge.net | 5084 | L: tboot-devel@lists.sourceforge.net |
| 5084 | W: http://tboot.sourceforge.net | 5085 | W: http://tboot.sourceforge.net |
| 5085 | T: hg http://tboot.hg.sourceforge.net:8000/hgroot/tboot/tboot | 5086 | T: hg http://tboot.hg.sourceforge.net:8000/hgroot/tboot/tboot |
| 5086 | S: Supported | 5087 | S: Supported |
| 5087 | F: Documentation/intel_txt.txt | 5088 | F: Documentation/intel_txt.txt |
| 5088 | F: include/linux/tboot.h | 5089 | F: include/linux/tboot.h |
| 5089 | F: arch/x86/kernel/tboot.c | 5090 | F: arch/x86/kernel/tboot.c |
| 5090 | 5091 | ||
| 5091 | INTEL WIRELESS WIMAX CONNECTION 2400 | 5092 | INTEL WIRELESS WIMAX CONNECTION 2400 |
| 5092 | M: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com> | 5093 | M: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com> |
| 5093 | M: linux-wimax@intel.com | 5094 | M: linux-wimax@intel.com |
| 5094 | L: wimax@linuxwimax.org (subscribers-only) | 5095 | L: wimax@linuxwimax.org (subscribers-only) |
| 5095 | S: Supported | 5096 | S: Supported |
| 5096 | W: http://linuxwimax.org | 5097 | W: http://linuxwimax.org |
| 5097 | F: Documentation/wimax/README.i2400m | 5098 | F: Documentation/wimax/README.i2400m |
| 5098 | F: drivers/net/wimax/i2400m/ | 5099 | F: drivers/net/wimax/i2400m/ |
| 5099 | F: include/uapi/linux/wimax/i2400m.h | 5100 | F: include/uapi/linux/wimax/i2400m.h |
| 5100 | 5101 | ||
| 5101 | INTEL WIRELESS 3945ABG/BG, 4965AGN (iwlegacy) | 5102 | INTEL WIRELESS 3945ABG/BG, 4965AGN (iwlegacy) |
| 5102 | M: Stanislaw Gruszka <sgruszka@redhat.com> | 5103 | M: Stanislaw Gruszka <sgruszka@redhat.com> |
| 5103 | L: linux-wireless@vger.kernel.org | 5104 | L: linux-wireless@vger.kernel.org |
| 5104 | S: Supported | 5105 | S: Supported |
| 5105 | F: drivers/net/wireless/iwlegacy/ | 5106 | F: drivers/net/wireless/iwlegacy/ |
| 5106 | 5107 | ||
| 5107 | INTEL WIRELESS WIFI LINK (iwlwifi) | 5108 | INTEL WIRELESS WIFI LINK (iwlwifi) |
| 5108 | M: Johannes Berg <johannes.berg@intel.com> | 5109 | M: Johannes Berg <johannes.berg@intel.com> |
| 5109 | M: Emmanuel Grumbach <emmanuel.grumbach@intel.com> | 5110 | M: Emmanuel Grumbach <emmanuel.grumbach@intel.com> |
| 5110 | M: Intel Linux Wireless <ilw@linux.intel.com> | 5111 | M: Intel Linux Wireless <ilw@linux.intel.com> |
| 5111 | L: linux-wireless@vger.kernel.org | 5112 | L: linux-wireless@vger.kernel.org |
| 5112 | W: http://intellinuxwireless.org | 5113 | W: http://intellinuxwireless.org |
| 5113 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi.git | 5114 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi.git |
| 5114 | S: Supported | 5115 | S: Supported |
| 5115 | F: drivers/net/wireless/iwlwifi/ | 5116 | F: drivers/net/wireless/iwlwifi/ |
| 5116 | 5117 | ||
| 5117 | INTEL MANAGEMENT ENGINE (mei) | 5118 | INTEL MANAGEMENT ENGINE (mei) |
| 5118 | M: Tomas Winkler <tomas.winkler@intel.com> | 5119 | M: Tomas Winkler <tomas.winkler@intel.com> |
| 5119 | L: linux-kernel@vger.kernel.org | 5120 | L: linux-kernel@vger.kernel.org |
| 5120 | S: Supported | 5121 | S: Supported |
| 5121 | F: include/uapi/linux/mei.h | 5122 | F: include/uapi/linux/mei.h |
| 5122 | F: drivers/misc/mei/* | 5123 | F: drivers/misc/mei/* |
| 5123 | F: Documentation/misc-devices/mei/* | 5124 | F: Documentation/misc-devices/mei/* |
| 5124 | 5125 | ||
| 5125 | IOC3 ETHERNET DRIVER | 5126 | IOC3 ETHERNET DRIVER |
| 5126 | M: Ralf Baechle <ralf@linux-mips.org> | 5127 | M: Ralf Baechle <ralf@linux-mips.org> |
| 5127 | L: linux-mips@linux-mips.org | 5128 | L: linux-mips@linux-mips.org |
| 5128 | S: Maintained | 5129 | S: Maintained |
| 5129 | F: drivers/net/ethernet/sgi/ioc3-eth.c | 5130 | F: drivers/net/ethernet/sgi/ioc3-eth.c |
| 5130 | 5131 | ||
| 5131 | IOC3 SERIAL DRIVER | 5132 | IOC3 SERIAL DRIVER |
| 5132 | M: Pat Gefre <pfg@sgi.com> | 5133 | M: Pat Gefre <pfg@sgi.com> |
| 5133 | L: linux-serial@vger.kernel.org | 5134 | L: linux-serial@vger.kernel.org |
| 5134 | S: Maintained | 5135 | S: Maintained |
| 5135 | F: drivers/tty/serial/ioc3_serial.c | 5136 | F: drivers/tty/serial/ioc3_serial.c |
| 5136 | 5137 | ||
| 5137 | IOMMU DRIVERS | 5138 | IOMMU DRIVERS |
| 5138 | M: Joerg Roedel <joro@8bytes.org> | 5139 | M: Joerg Roedel <joro@8bytes.org> |
| 5139 | L: iommu@lists.linux-foundation.org | 5140 | L: iommu@lists.linux-foundation.org |
| 5140 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git | 5141 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu.git |
| 5141 | S: Maintained | 5142 | S: Maintained |
| 5142 | F: drivers/iommu/ | 5143 | F: drivers/iommu/ |
| 5143 | 5144 | ||
| 5144 | IP MASQUERADING | 5145 | IP MASQUERADING |
| 5145 | M: Juanjo Ciarlante <jjciarla@raiz.uncu.edu.ar> | 5146 | M: Juanjo Ciarlante <jjciarla@raiz.uncu.edu.ar> |
| 5146 | S: Maintained | 5147 | S: Maintained |
| 5147 | F: net/ipv4/netfilter/ipt_MASQUERADE.c | 5148 | F: net/ipv4/netfilter/ipt_MASQUERADE.c |
| 5148 | 5149 | ||
| 5149 | IP1000A 10/100/1000 GIGABIT ETHERNET DRIVER | 5150 | IP1000A 10/100/1000 GIGABIT ETHERNET DRIVER |
| 5150 | M: Francois Romieu <romieu@fr.zoreil.com> | 5151 | M: Francois Romieu <romieu@fr.zoreil.com> |
| 5151 | M: Sorbica Shieh <sorbica@icplus.com.tw> | 5152 | M: Sorbica Shieh <sorbica@icplus.com.tw> |
| 5152 | L: netdev@vger.kernel.org | 5153 | L: netdev@vger.kernel.org |
| 5153 | S: Maintained | 5154 | S: Maintained |
| 5154 | F: drivers/net/ethernet/icplus/ipg.* | 5155 | F: drivers/net/ethernet/icplus/ipg.* |
| 5155 | 5156 | ||
| 5156 | IPATH DRIVER | 5157 | IPATH DRIVER |
| 5157 | M: Mike Marciniszyn <infinipath@intel.com> | 5158 | M: Mike Marciniszyn <infinipath@intel.com> |
| 5158 | L: linux-rdma@vger.kernel.org | 5159 | L: linux-rdma@vger.kernel.org |
| 5159 | S: Maintained | 5160 | S: Maintained |
| 5160 | F: drivers/infiniband/hw/ipath/ | 5161 | F: drivers/infiniband/hw/ipath/ |
| 5161 | 5162 | ||
| 5162 | IPMI SUBSYSTEM | 5163 | IPMI SUBSYSTEM |
| 5163 | M: Corey Minyard <minyard@acm.org> | 5164 | M: Corey Minyard <minyard@acm.org> |
| 5164 | L: openipmi-developer@lists.sourceforge.net (moderated for non-subscribers) | 5165 | L: openipmi-developer@lists.sourceforge.net (moderated for non-subscribers) |
| 5165 | W: http://openipmi.sourceforge.net/ | 5166 | W: http://openipmi.sourceforge.net/ |
| 5166 | S: Supported | 5167 | S: Supported |
| 5167 | F: Documentation/IPMI.txt | 5168 | F: Documentation/IPMI.txt |
| 5168 | F: drivers/char/ipmi/ | 5169 | F: drivers/char/ipmi/ |
| 5169 | F: include/linux/ipmi* | 5170 | F: include/linux/ipmi* |
| 5170 | F: include/uapi/linux/ipmi* | 5171 | F: include/uapi/linux/ipmi* |
| 5171 | 5172 | ||
| 5172 | IPS SCSI RAID DRIVER | 5173 | IPS SCSI RAID DRIVER |
| 5173 | M: Adaptec OEM Raid Solutions <aacraid@adaptec.com> | 5174 | M: Adaptec OEM Raid Solutions <aacraid@adaptec.com> |
| 5174 | L: linux-scsi@vger.kernel.org | 5175 | L: linux-scsi@vger.kernel.org |
| 5175 | W: http://www.adaptec.com/ | 5176 | W: http://www.adaptec.com/ |
| 5176 | S: Maintained | 5177 | S: Maintained |
| 5177 | F: drivers/scsi/ips* | 5178 | F: drivers/scsi/ips* |
| 5178 | 5179 | ||
| 5179 | IPVS | 5180 | IPVS |
| 5180 | M: Wensong Zhang <wensong@linux-vs.org> | 5181 | M: Wensong Zhang <wensong@linux-vs.org> |
| 5181 | M: Simon Horman <horms@verge.net.au> | 5182 | M: Simon Horman <horms@verge.net.au> |
| 5182 | M: Julian Anastasov <ja@ssi.bg> | 5183 | M: Julian Anastasov <ja@ssi.bg> |
| 5183 | L: netdev@vger.kernel.org | 5184 | L: netdev@vger.kernel.org |
| 5184 | L: lvs-devel@vger.kernel.org | 5185 | L: lvs-devel@vger.kernel.org |
| 5185 | S: Maintained | 5186 | S: Maintained |
| 5186 | F: Documentation/networking/ipvs-sysctl.txt | 5187 | F: Documentation/networking/ipvs-sysctl.txt |
| 5187 | F: include/net/ip_vs.h | 5188 | F: include/net/ip_vs.h |
| 5188 | F: include/uapi/linux/ip_vs.h | 5189 | F: include/uapi/linux/ip_vs.h |
| 5189 | F: net/netfilter/ipvs/ | 5190 | F: net/netfilter/ipvs/ |
| 5190 | 5191 | ||
| 5191 | IPWIRELESS DRIVER | 5192 | IPWIRELESS DRIVER |
| 5192 | M: Jiri Kosina <jkosina@suse.cz> | 5193 | M: Jiri Kosina <jkosina@suse.cz> |
| 5193 | M: David Sterba <dsterba@suse.cz> | 5194 | M: David Sterba <dsterba@suse.cz> |
| 5194 | S: Odd Fixes | 5195 | S: Odd Fixes |
| 5195 | F: drivers/tty/ipwireless/ | 5196 | F: drivers/tty/ipwireless/ |
| 5196 | 5197 | ||
| 5197 | IPX NETWORK LAYER | 5198 | IPX NETWORK LAYER |
| 5198 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> | 5199 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> |
| 5199 | L: netdev@vger.kernel.org | 5200 | L: netdev@vger.kernel.org |
| 5200 | S: Maintained | 5201 | S: Maintained |
| 5201 | F: include/net/ipx.h | 5202 | F: include/net/ipx.h |
| 5202 | F: include/uapi/linux/ipx.h | 5203 | F: include/uapi/linux/ipx.h |
| 5203 | F: net/ipx/ | 5204 | F: net/ipx/ |
| 5204 | 5205 | ||
| 5205 | IRDA SUBSYSTEM | 5206 | IRDA SUBSYSTEM |
| 5206 | M: Samuel Ortiz <samuel@sortiz.org> | 5207 | M: Samuel Ortiz <samuel@sortiz.org> |
| 5207 | L: irda-users@lists.sourceforge.net (subscribers-only) | 5208 | L: irda-users@lists.sourceforge.net (subscribers-only) |
| 5208 | L: netdev@vger.kernel.org | 5209 | L: netdev@vger.kernel.org |
| 5209 | W: http://irda.sourceforge.net/ | 5210 | W: http://irda.sourceforge.net/ |
| 5210 | S: Maintained | 5211 | S: Maintained |
| 5211 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sameo/irda-2.6.git | 5212 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sameo/irda-2.6.git |
| 5212 | F: Documentation/networking/irda.txt | 5213 | F: Documentation/networking/irda.txt |
| 5213 | F: drivers/net/irda/ | 5214 | F: drivers/net/irda/ |
| 5214 | F: include/net/irda/ | 5215 | F: include/net/irda/ |
| 5215 | F: net/irda/ | 5216 | F: net/irda/ |
| 5216 | 5217 | ||
| 5217 | IRQ SUBSYSTEM | 5218 | IRQ SUBSYSTEM |
| 5218 | M: Thomas Gleixner <tglx@linutronix.de> | 5219 | M: Thomas Gleixner <tglx@linutronix.de> |
| 5219 | L: linux-kernel@vger.kernel.org | 5220 | L: linux-kernel@vger.kernel.org |
| 5220 | S: Maintained | 5221 | S: Maintained |
| 5221 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git irq/core | 5222 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git irq/core |
| 5222 | F: kernel/irq/ | 5223 | F: kernel/irq/ |
| 5223 | 5224 | ||
| 5224 | IRQCHIP DRIVERS | 5225 | IRQCHIP DRIVERS |
| 5225 | M: Thomas Gleixner <tglx@linutronix.de> | 5226 | M: Thomas Gleixner <tglx@linutronix.de> |
| 5226 | M: Jason Cooper <jason@lakedaemon.net> | 5227 | M: Jason Cooper <jason@lakedaemon.net> |
| 5227 | L: linux-kernel@vger.kernel.org | 5228 | L: linux-kernel@vger.kernel.org |
| 5228 | S: Maintained | 5229 | S: Maintained |
| 5229 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git irq/core | 5230 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git irq/core |
| 5230 | T: git git://git.infradead.org/users/jcooper/linux.git irqchip/core | 5231 | T: git git://git.infradead.org/users/jcooper/linux.git irqchip/core |
| 5231 | F: Documentation/devicetree/bindings/interrupt-controller/ | 5232 | F: Documentation/devicetree/bindings/interrupt-controller/ |
| 5232 | F: drivers/irqchip/ | 5233 | F: drivers/irqchip/ |
| 5233 | 5234 | ||
| 5234 | IRQ DOMAINS (IRQ NUMBER MAPPING LIBRARY) | 5235 | IRQ DOMAINS (IRQ NUMBER MAPPING LIBRARY) |
| 5235 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> | 5236 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> |
| 5236 | S: Maintained | 5237 | S: Maintained |
| 5237 | F: Documentation/IRQ-domain.txt | 5238 | F: Documentation/IRQ-domain.txt |
| 5238 | F: include/linux/irqdomain.h | 5239 | F: include/linux/irqdomain.h |
| 5239 | F: kernel/irq/irqdomain.c | 5240 | F: kernel/irq/irqdomain.c |
| 5240 | 5241 | ||
| 5241 | ISAPNP | 5242 | ISAPNP |
| 5242 | M: Jaroslav Kysela <perex@perex.cz> | 5243 | M: Jaroslav Kysela <perex@perex.cz> |
| 5243 | S: Maintained | 5244 | S: Maintained |
| 5244 | F: Documentation/isapnp.txt | 5245 | F: Documentation/isapnp.txt |
| 5245 | F: drivers/pnp/isapnp/ | 5246 | F: drivers/pnp/isapnp/ |
| 5246 | F: include/linux/isapnp.h | 5247 | F: include/linux/isapnp.h |
| 5247 | 5248 | ||
| 5248 | ISA RADIO MODULE | 5249 | ISA RADIO MODULE |
| 5249 | M: Hans Verkuil <hverkuil@xs4all.nl> | 5250 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 5250 | L: linux-media@vger.kernel.org | 5251 | L: linux-media@vger.kernel.org |
| 5251 | T: git git://linuxtv.org/media_tree.git | 5252 | T: git git://linuxtv.org/media_tree.git |
| 5252 | W: http://linuxtv.org | 5253 | W: http://linuxtv.org |
| 5253 | S: Maintained | 5254 | S: Maintained |
| 5254 | F: drivers/media/radio/radio-isa* | 5255 | F: drivers/media/radio/radio-isa* |
| 5255 | 5256 | ||
| 5256 | iSCSI BOOT FIRMWARE TABLE (iBFT) DRIVER | 5257 | iSCSI BOOT FIRMWARE TABLE (iBFT) DRIVER |
| 5257 | M: Peter Jones <pjones@redhat.com> | 5258 | M: Peter Jones <pjones@redhat.com> |
| 5258 | M: Konrad Rzeszutek Wilk <konrad@kernel.org> | 5259 | M: Konrad Rzeszutek Wilk <konrad@kernel.org> |
| 5259 | S: Maintained | 5260 | S: Maintained |
| 5260 | F: drivers/firmware/iscsi_ibft* | 5261 | F: drivers/firmware/iscsi_ibft* |
| 5261 | 5262 | ||
| 5262 | ISCSI | 5263 | ISCSI |
| 5263 | M: Mike Christie <michaelc@cs.wisc.edu> | 5264 | M: Mike Christie <michaelc@cs.wisc.edu> |
| 5264 | L: open-iscsi@googlegroups.com | 5265 | L: open-iscsi@googlegroups.com |
| 5265 | W: www.open-iscsi.org | 5266 | W: www.open-iscsi.org |
| 5266 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mnc/linux-2.6-iscsi.git | 5267 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mnc/linux-2.6-iscsi.git |
| 5267 | S: Maintained | 5268 | S: Maintained |
| 5268 | F: drivers/scsi/*iscsi* | 5269 | F: drivers/scsi/*iscsi* |
| 5269 | F: include/scsi/*iscsi* | 5270 | F: include/scsi/*iscsi* |
| 5270 | 5271 | ||
| 5271 | ISCSI EXTENSIONS FOR RDMA (ISER) INITIATOR | 5272 | ISCSI EXTENSIONS FOR RDMA (ISER) INITIATOR |
| 5272 | M: Or Gerlitz <ogerlitz@mellanox.com> | 5273 | M: Or Gerlitz <ogerlitz@mellanox.com> |
| 5273 | M: Sagi Grimberg <sagig@mellanox.com> | 5274 | M: Sagi Grimberg <sagig@mellanox.com> |
| 5274 | M: Roi Dayan <roid@mellanox.com> | 5275 | M: Roi Dayan <roid@mellanox.com> |
| 5275 | L: linux-rdma@vger.kernel.org | 5276 | L: linux-rdma@vger.kernel.org |
| 5276 | S: Supported | 5277 | S: Supported |
| 5277 | W: http://www.openfabrics.org | 5278 | W: http://www.openfabrics.org |
| 5278 | W: www.open-iscsi.org | 5279 | W: www.open-iscsi.org |
| 5279 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ | 5280 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ |
| 5280 | F: drivers/infiniband/ulp/iser/ | 5281 | F: drivers/infiniband/ulp/iser/ |
| 5281 | 5282 | ||
| 5282 | ISDN SUBSYSTEM | 5283 | ISDN SUBSYSTEM |
| 5283 | M: Karsten Keil <isdn@linux-pingi.de> | 5284 | M: Karsten Keil <isdn@linux-pingi.de> |
| 5284 | L: isdn4linux@listserv.isdn4linux.de (subscribers-only) | 5285 | L: isdn4linux@listserv.isdn4linux.de (subscribers-only) |
| 5285 | L: netdev@vger.kernel.org | 5286 | L: netdev@vger.kernel.org |
| 5286 | W: http://www.isdn4linux.de | 5287 | W: http://www.isdn4linux.de |
| 5287 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kkeil/isdn-2.6.git | 5288 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kkeil/isdn-2.6.git |
| 5288 | S: Maintained | 5289 | S: Maintained |
| 5289 | F: Documentation/isdn/ | 5290 | F: Documentation/isdn/ |
| 5290 | F: drivers/isdn/ | 5291 | F: drivers/isdn/ |
| 5291 | F: include/linux/isdn.h | 5292 | F: include/linux/isdn.h |
| 5292 | F: include/linux/isdn/ | 5293 | F: include/linux/isdn/ |
| 5293 | F: include/uapi/linux/isdn.h | 5294 | F: include/uapi/linux/isdn.h |
| 5294 | F: include/uapi/linux/isdn/ | 5295 | F: include/uapi/linux/isdn/ |
| 5295 | 5296 | ||
| 5296 | ISDN SUBSYSTEM (Eicon active card driver) | 5297 | ISDN SUBSYSTEM (Eicon active card driver) |
| 5297 | M: Armin Schindler <mac@melware.de> | 5298 | M: Armin Schindler <mac@melware.de> |
| 5298 | L: isdn4linux@listserv.isdn4linux.de (subscribers-only) | 5299 | L: isdn4linux@listserv.isdn4linux.de (subscribers-only) |
| 5299 | W: http://www.melware.de | 5300 | W: http://www.melware.de |
| 5300 | S: Maintained | 5301 | S: Maintained |
| 5301 | F: drivers/isdn/hardware/eicon/ | 5302 | F: drivers/isdn/hardware/eicon/ |
| 5302 | 5303 | ||
| 5303 | IT87 HARDWARE MONITORING DRIVER | 5304 | IT87 HARDWARE MONITORING DRIVER |
| 5304 | M: Jean Delvare <jdelvare@suse.de> | 5305 | M: Jean Delvare <jdelvare@suse.de> |
| 5305 | L: lm-sensors@lm-sensors.org | 5306 | L: lm-sensors@lm-sensors.org |
| 5306 | S: Maintained | 5307 | S: Maintained |
| 5307 | F: Documentation/hwmon/it87 | 5308 | F: Documentation/hwmon/it87 |
| 5308 | F: drivers/hwmon/it87.c | 5309 | F: drivers/hwmon/it87.c |
| 5309 | 5310 | ||
| 5310 | IT913X MEDIA DRIVER | 5311 | IT913X MEDIA DRIVER |
| 5311 | M: Antti Palosaari <crope@iki.fi> | 5312 | M: Antti Palosaari <crope@iki.fi> |
| 5312 | L: linux-media@vger.kernel.org | 5313 | L: linux-media@vger.kernel.org |
| 5313 | W: http://linuxtv.org/ | 5314 | W: http://linuxtv.org/ |
| 5314 | W: http://palosaari.fi/linux/ | 5315 | W: http://palosaari.fi/linux/ |
| 5315 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5316 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5316 | T: git git://linuxtv.org/anttip/media_tree.git | 5317 | T: git git://linuxtv.org/anttip/media_tree.git |
| 5317 | S: Maintained | 5318 | S: Maintained |
| 5318 | F: drivers/media/tuners/it913x* | 5319 | F: drivers/media/tuners/it913x* |
| 5319 | 5320 | ||
| 5320 | IVTV VIDEO4LINUX DRIVER | 5321 | IVTV VIDEO4LINUX DRIVER |
| 5321 | M: Andy Walls <awalls@md.metrocast.net> | 5322 | M: Andy Walls <awalls@md.metrocast.net> |
| 5322 | L: ivtv-devel@ivtvdriver.org (subscribers-only) | 5323 | L: ivtv-devel@ivtvdriver.org (subscribers-only) |
| 5323 | L: linux-media@vger.kernel.org | 5324 | L: linux-media@vger.kernel.org |
| 5324 | T: git git://linuxtv.org/media_tree.git | 5325 | T: git git://linuxtv.org/media_tree.git |
| 5325 | W: http://www.ivtvdriver.org | 5326 | W: http://www.ivtvdriver.org |
| 5326 | S: Maintained | 5327 | S: Maintained |
| 5327 | F: Documentation/video4linux/*.ivtv | 5328 | F: Documentation/video4linux/*.ivtv |
| 5328 | F: drivers/media/pci/ivtv/ | 5329 | F: drivers/media/pci/ivtv/ |
| 5329 | F: include/uapi/linux/ivtv* | 5330 | F: include/uapi/linux/ivtv* |
| 5330 | 5331 | ||
| 5331 | IX2505V MEDIA DRIVER | 5332 | IX2505V MEDIA DRIVER |
| 5332 | M: Malcolm Priestley <tvboxspy@gmail.com> | 5333 | M: Malcolm Priestley <tvboxspy@gmail.com> |
| 5333 | L: linux-media@vger.kernel.org | 5334 | L: linux-media@vger.kernel.org |
| 5334 | W: http://linuxtv.org/ | 5335 | W: http://linuxtv.org/ |
| 5335 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5336 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5336 | S: Maintained | 5337 | S: Maintained |
| 5337 | F: drivers/media/dvb-frontends/ix2505v* | 5338 | F: drivers/media/dvb-frontends/ix2505v* |
| 5338 | 5339 | ||
| 5339 | JC42.4 TEMPERATURE SENSOR DRIVER | 5340 | JC42.4 TEMPERATURE SENSOR DRIVER |
| 5340 | M: Guenter Roeck <linux@roeck-us.net> | 5341 | M: Guenter Roeck <linux@roeck-us.net> |
| 5341 | L: lm-sensors@lm-sensors.org | 5342 | L: lm-sensors@lm-sensors.org |
| 5342 | S: Maintained | 5343 | S: Maintained |
| 5343 | F: drivers/hwmon/jc42.c | 5344 | F: drivers/hwmon/jc42.c |
| 5344 | F: Documentation/hwmon/jc42 | 5345 | F: Documentation/hwmon/jc42 |
| 5345 | 5346 | ||
| 5346 | JFS FILESYSTEM | 5347 | JFS FILESYSTEM |
| 5347 | M: Dave Kleikamp <shaggy@kernel.org> | 5348 | M: Dave Kleikamp <shaggy@kernel.org> |
| 5348 | L: jfs-discussion@lists.sourceforge.net | 5349 | L: jfs-discussion@lists.sourceforge.net |
| 5349 | W: http://jfs.sourceforge.net/ | 5350 | W: http://jfs.sourceforge.net/ |
| 5350 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/shaggy/jfs-2.6.git | 5351 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/shaggy/jfs-2.6.git |
| 5351 | S: Maintained | 5352 | S: Maintained |
| 5352 | F: Documentation/filesystems/jfs.txt | 5353 | F: Documentation/filesystems/jfs.txt |
| 5353 | F: fs/jfs/ | 5354 | F: fs/jfs/ |
| 5354 | 5355 | ||
| 5355 | JME NETWORK DRIVER | 5356 | JME NETWORK DRIVER |
| 5356 | M: Guo-Fu Tseng <cooldavid@cooldavid.org> | 5357 | M: Guo-Fu Tseng <cooldavid@cooldavid.org> |
| 5357 | L: netdev@vger.kernel.org | 5358 | L: netdev@vger.kernel.org |
| 5358 | S: Maintained | 5359 | S: Maintained |
| 5359 | F: drivers/net/ethernet/jme.* | 5360 | F: drivers/net/ethernet/jme.* |
| 5360 | 5361 | ||
| 5361 | JOURNALLING FLASH FILE SYSTEM V2 (JFFS2) | 5362 | JOURNALLING FLASH FILE SYSTEM V2 (JFFS2) |
| 5362 | M: David Woodhouse <dwmw2@infradead.org> | 5363 | M: David Woodhouse <dwmw2@infradead.org> |
| 5363 | L: linux-mtd@lists.infradead.org | 5364 | L: linux-mtd@lists.infradead.org |
| 5364 | W: http://www.linux-mtd.infradead.org/doc/jffs2.html | 5365 | W: http://www.linux-mtd.infradead.org/doc/jffs2.html |
| 5365 | S: Maintained | 5366 | S: Maintained |
| 5366 | F: fs/jffs2/ | 5367 | F: fs/jffs2/ |
| 5367 | F: include/uapi/linux/jffs2.h | 5368 | F: include/uapi/linux/jffs2.h |
| 5368 | 5369 | ||
| 5369 | JOURNALLING LAYER FOR BLOCK DEVICES (JBD) | 5370 | JOURNALLING LAYER FOR BLOCK DEVICES (JBD) |
| 5370 | M: Andrew Morton <akpm@linux-foundation.org> | 5371 | M: Andrew Morton <akpm@linux-foundation.org> |
| 5371 | M: Jan Kara <jack@suse.cz> | 5372 | M: Jan Kara <jack@suse.cz> |
| 5372 | L: linux-ext4@vger.kernel.org | 5373 | L: linux-ext4@vger.kernel.org |
| 5373 | S: Maintained | 5374 | S: Maintained |
| 5374 | F: fs/jbd/ | 5375 | F: fs/jbd/ |
| 5375 | F: include/linux/jbd.h | 5376 | F: include/linux/jbd.h |
| 5376 | 5377 | ||
| 5377 | JOURNALLING LAYER FOR BLOCK DEVICES (JBD2) | 5378 | JOURNALLING LAYER FOR BLOCK DEVICES (JBD2) |
| 5378 | M: "Theodore Ts'o" <tytso@mit.edu> | 5379 | M: "Theodore Ts'o" <tytso@mit.edu> |
| 5379 | L: linux-ext4@vger.kernel.org | 5380 | L: linux-ext4@vger.kernel.org |
| 5380 | S: Maintained | 5381 | S: Maintained |
| 5381 | F: fs/jbd2/ | 5382 | F: fs/jbd2/ |
| 5382 | F: include/linux/jbd2.h | 5383 | F: include/linux/jbd2.h |
| 5383 | 5384 | ||
| 5384 | JSM Neo PCI based serial card | 5385 | JSM Neo PCI based serial card |
| 5385 | M: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com> | 5386 | M: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com> |
| 5386 | L: linux-serial@vger.kernel.org | 5387 | L: linux-serial@vger.kernel.org |
| 5387 | S: Maintained | 5388 | S: Maintained |
| 5388 | F: drivers/tty/serial/jsm/ | 5389 | F: drivers/tty/serial/jsm/ |
| 5389 | 5390 | ||
| 5390 | K10TEMP HARDWARE MONITORING DRIVER | 5391 | K10TEMP HARDWARE MONITORING DRIVER |
| 5391 | M: Clemens Ladisch <clemens@ladisch.de> | 5392 | M: Clemens Ladisch <clemens@ladisch.de> |
| 5392 | L: lm-sensors@lm-sensors.org | 5393 | L: lm-sensors@lm-sensors.org |
| 5393 | S: Maintained | 5394 | S: Maintained |
| 5394 | F: Documentation/hwmon/k10temp | 5395 | F: Documentation/hwmon/k10temp |
| 5395 | F: drivers/hwmon/k10temp.c | 5396 | F: drivers/hwmon/k10temp.c |
| 5396 | 5397 | ||
| 5397 | K8TEMP HARDWARE MONITORING DRIVER | 5398 | K8TEMP HARDWARE MONITORING DRIVER |
| 5398 | M: Rudolf Marek <r.marek@assembler.cz> | 5399 | M: Rudolf Marek <r.marek@assembler.cz> |
| 5399 | L: lm-sensors@lm-sensors.org | 5400 | L: lm-sensors@lm-sensors.org |
| 5400 | S: Maintained | 5401 | S: Maintained |
| 5401 | F: Documentation/hwmon/k8temp | 5402 | F: Documentation/hwmon/k8temp |
| 5402 | F: drivers/hwmon/k8temp.c | 5403 | F: drivers/hwmon/k8temp.c |
| 5403 | 5404 | ||
| 5404 | KCONFIG | 5405 | KCONFIG |
| 5405 | M: "Yann E. MORIN" <yann.morin.1998@free.fr> | 5406 | M: "Yann E. MORIN" <yann.morin.1998@free.fr> |
| 5406 | L: linux-kbuild@vger.kernel.org | 5407 | L: linux-kbuild@vger.kernel.org |
| 5407 | T: git git://gitorious.org/linux-kconfig/linux-kconfig | 5408 | T: git git://gitorious.org/linux-kconfig/linux-kconfig |
| 5408 | S: Maintained | 5409 | S: Maintained |
| 5409 | F: Documentation/kbuild/kconfig-language.txt | 5410 | F: Documentation/kbuild/kconfig-language.txt |
| 5410 | F: scripts/kconfig/ | 5411 | F: scripts/kconfig/ |
| 5411 | 5412 | ||
| 5412 | KDUMP | 5413 | KDUMP |
| 5413 | M: Vivek Goyal <vgoyal@redhat.com> | 5414 | M: Vivek Goyal <vgoyal@redhat.com> |
| 5414 | M: Haren Myneni <hbabu@us.ibm.com> | 5415 | M: Haren Myneni <hbabu@us.ibm.com> |
| 5415 | L: kexec@lists.infradead.org | 5416 | L: kexec@lists.infradead.org |
| 5416 | W: http://lse.sourceforge.net/kdump/ | 5417 | W: http://lse.sourceforge.net/kdump/ |
| 5417 | S: Maintained | 5418 | S: Maintained |
| 5418 | F: Documentation/kdump/ | 5419 | F: Documentation/kdump/ |
| 5419 | 5420 | ||
| 5420 | KEENE FM RADIO TRANSMITTER DRIVER | 5421 | KEENE FM RADIO TRANSMITTER DRIVER |
| 5421 | M: Hans Verkuil <hverkuil@xs4all.nl> | 5422 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 5422 | L: linux-media@vger.kernel.org | 5423 | L: linux-media@vger.kernel.org |
| 5423 | T: git git://linuxtv.org/media_tree.git | 5424 | T: git git://linuxtv.org/media_tree.git |
| 5424 | W: http://linuxtv.org | 5425 | W: http://linuxtv.org |
| 5425 | S: Maintained | 5426 | S: Maintained |
| 5426 | F: drivers/media/radio/radio-keene* | 5427 | F: drivers/media/radio/radio-keene* |
| 5427 | 5428 | ||
| 5428 | KERNEL AUTOMOUNTER v4 (AUTOFS4) | 5429 | KERNEL AUTOMOUNTER v4 (AUTOFS4) |
| 5429 | M: Ian Kent <raven@themaw.net> | 5430 | M: Ian Kent <raven@themaw.net> |
| 5430 | L: autofs@vger.kernel.org | 5431 | L: autofs@vger.kernel.org |
| 5431 | S: Maintained | 5432 | S: Maintained |
| 5432 | F: fs/autofs4/ | 5433 | F: fs/autofs4/ |
| 5433 | 5434 | ||
| 5434 | KERNEL BUILD + files below scripts/ (unless maintained elsewhere) | 5435 | KERNEL BUILD + files below scripts/ (unless maintained elsewhere) |
| 5435 | M: Michal Marek <mmarek@suse.cz> | 5436 | M: Michal Marek <mmarek@suse.cz> |
| 5436 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git for-next | 5437 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git for-next |
| 5437 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git rc-fixes | 5438 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild.git rc-fixes |
| 5438 | L: linux-kbuild@vger.kernel.org | 5439 | L: linux-kbuild@vger.kernel.org |
| 5439 | S: Maintained | 5440 | S: Maintained |
| 5440 | F: Documentation/kbuild/ | 5441 | F: Documentation/kbuild/ |
| 5441 | F: Makefile | 5442 | F: Makefile |
| 5442 | F: scripts/Makefile.* | 5443 | F: scripts/Makefile.* |
| 5443 | F: scripts/basic/ | 5444 | F: scripts/basic/ |
| 5444 | F: scripts/mk* | 5445 | F: scripts/mk* |
| 5445 | F: scripts/package/ | 5446 | F: scripts/package/ |
| 5446 | 5447 | ||
| 5447 | KERNEL JANITORS | 5448 | KERNEL JANITORS |
| 5448 | L: kernel-janitors@vger.kernel.org | 5449 | L: kernel-janitors@vger.kernel.org |
| 5449 | W: http://kernelnewbies.org/KernelJanitors | 5450 | W: http://kernelnewbies.org/KernelJanitors |
| 5450 | S: Odd Fixes | 5451 | S: Odd Fixes |
| 5451 | 5452 | ||
| 5452 | KERNEL NFSD, SUNRPC, AND LOCKD SERVERS | 5453 | KERNEL NFSD, SUNRPC, AND LOCKD SERVERS |
| 5453 | M: "J. Bruce Fields" <bfields@fieldses.org> | 5454 | M: "J. Bruce Fields" <bfields@fieldses.org> |
| 5454 | L: linux-nfs@vger.kernel.org | 5455 | L: linux-nfs@vger.kernel.org |
| 5455 | W: http://nfs.sourceforge.net/ | 5456 | W: http://nfs.sourceforge.net/ |
| 5456 | S: Supported | 5457 | S: Supported |
| 5457 | F: fs/nfsd/ | 5458 | F: fs/nfsd/ |
| 5458 | F: include/uapi/linux/nfsd/ | 5459 | F: include/uapi/linux/nfsd/ |
| 5459 | F: fs/lockd/ | 5460 | F: fs/lockd/ |
| 5460 | F: fs/nfs_common/ | 5461 | F: fs/nfs_common/ |
| 5461 | F: net/sunrpc/ | 5462 | F: net/sunrpc/ |
| 5462 | F: include/linux/lockd/ | 5463 | F: include/linux/lockd/ |
| 5463 | F: include/linux/sunrpc/ | 5464 | F: include/linux/sunrpc/ |
| 5464 | F: include/uapi/linux/sunrpc/ | 5465 | F: include/uapi/linux/sunrpc/ |
| 5465 | 5466 | ||
| 5466 | KERNEL SELFTEST FRAMEWORK | 5467 | KERNEL SELFTEST FRAMEWORK |
| 5467 | M: Shuah Khan <shuahkh@osg.samsung.com> | 5468 | M: Shuah Khan <shuahkh@osg.samsung.com> |
| 5468 | L: linux-api@vger.kernel.org | 5469 | L: linux-api@vger.kernel.org |
| 5469 | T: git git://git.kernel.org/pub/scm/shuah/linux-kselftest | 5470 | T: git git://git.kernel.org/pub/scm/shuah/linux-kselftest |
| 5470 | S: Maintained | 5471 | S: Maintained |
| 5471 | F: tools/testing/selftests | 5472 | F: tools/testing/selftests |
| 5472 | 5473 | ||
| 5473 | KERNEL VIRTUAL MACHINE (KVM) | 5474 | KERNEL VIRTUAL MACHINE (KVM) |
| 5474 | M: Gleb Natapov <gleb@kernel.org> | 5475 | M: Gleb Natapov <gleb@kernel.org> |
| 5475 | M: Paolo Bonzini <pbonzini@redhat.com> | 5476 | M: Paolo Bonzini <pbonzini@redhat.com> |
| 5476 | L: kvm@vger.kernel.org | 5477 | L: kvm@vger.kernel.org |
| 5477 | W: http://www.linux-kvm.org | 5478 | W: http://www.linux-kvm.org |
| 5478 | T: git git://git.kernel.org/pub/scm/virt/kvm/kvm.git | 5479 | T: git git://git.kernel.org/pub/scm/virt/kvm/kvm.git |
| 5479 | S: Supported | 5480 | S: Supported |
| 5480 | F: Documentation/*/kvm*.txt | 5481 | F: Documentation/*/kvm*.txt |
| 5481 | F: Documentation/virtual/kvm/ | 5482 | F: Documentation/virtual/kvm/ |
| 5482 | F: arch/*/kvm/ | 5483 | F: arch/*/kvm/ |
| 5483 | F: arch/*/include/asm/kvm* | 5484 | F: arch/*/include/asm/kvm* |
| 5484 | F: include/linux/kvm* | 5485 | F: include/linux/kvm* |
| 5485 | F: include/uapi/linux/kvm* | 5486 | F: include/uapi/linux/kvm* |
| 5486 | F: virt/kvm/ | 5487 | F: virt/kvm/ |
| 5487 | 5488 | ||
| 5488 | KERNEL VIRTUAL MACHINE (KVM) FOR AMD-V | 5489 | KERNEL VIRTUAL MACHINE (KVM) FOR AMD-V |
| 5489 | M: Joerg Roedel <joro@8bytes.org> | 5490 | M: Joerg Roedel <joro@8bytes.org> |
| 5490 | L: kvm@vger.kernel.org | 5491 | L: kvm@vger.kernel.org |
| 5491 | W: http://kvm.qumranet.com | 5492 | W: http://kvm.qumranet.com |
| 5492 | S: Maintained | 5493 | S: Maintained |
| 5493 | F: arch/x86/include/asm/svm.h | 5494 | F: arch/x86/include/asm/svm.h |
| 5494 | F: arch/x86/kvm/svm.c | 5495 | F: arch/x86/kvm/svm.c |
| 5495 | 5496 | ||
| 5496 | KERNEL VIRTUAL MACHINE (KVM) FOR POWERPC | 5497 | KERNEL VIRTUAL MACHINE (KVM) FOR POWERPC |
| 5497 | M: Alexander Graf <agraf@suse.de> | 5498 | M: Alexander Graf <agraf@suse.de> |
| 5498 | L: kvm-ppc@vger.kernel.org | 5499 | L: kvm-ppc@vger.kernel.org |
| 5499 | W: http://kvm.qumranet.com | 5500 | W: http://kvm.qumranet.com |
| 5500 | T: git git://github.com/agraf/linux-2.6.git | 5501 | T: git git://github.com/agraf/linux-2.6.git |
| 5501 | S: Supported | 5502 | S: Supported |
| 5502 | F: arch/powerpc/include/asm/kvm* | 5503 | F: arch/powerpc/include/asm/kvm* |
| 5503 | F: arch/powerpc/kvm/ | 5504 | F: arch/powerpc/kvm/ |
| 5504 | 5505 | ||
| 5505 | KERNEL VIRTUAL MACHINE for s390 (KVM/s390) | 5506 | KERNEL VIRTUAL MACHINE for s390 (KVM/s390) |
| 5506 | M: Christian Borntraeger <borntraeger@de.ibm.com> | 5507 | M: Christian Borntraeger <borntraeger@de.ibm.com> |
| 5507 | M: Cornelia Huck <cornelia.huck@de.ibm.com> | 5508 | M: Cornelia Huck <cornelia.huck@de.ibm.com> |
| 5508 | M: linux390@de.ibm.com | 5509 | M: linux390@de.ibm.com |
| 5509 | L: linux-s390@vger.kernel.org | 5510 | L: linux-s390@vger.kernel.org |
| 5510 | W: http://www.ibm.com/developerworks/linux/linux390/ | 5511 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 5511 | S: Supported | 5512 | S: Supported |
| 5512 | F: Documentation/s390/kvm.txt | 5513 | F: Documentation/s390/kvm.txt |
| 5513 | F: arch/s390/include/asm/kvm* | 5514 | F: arch/s390/include/asm/kvm* |
| 5514 | F: arch/s390/kvm/ | 5515 | F: arch/s390/kvm/ |
| 5515 | F: drivers/s390/kvm/ | 5516 | F: drivers/s390/kvm/ |
| 5516 | 5517 | ||
| 5517 | KERNEL VIRTUAL MACHINE (KVM) FOR ARM | 5518 | KERNEL VIRTUAL MACHINE (KVM) FOR ARM |
| 5518 | M: Christoffer Dall <christoffer.dall@linaro.org> | 5519 | M: Christoffer Dall <christoffer.dall@linaro.org> |
| 5519 | M: Marc Zyngier <marc.zyngier@arm.com> | 5520 | M: Marc Zyngier <marc.zyngier@arm.com> |
| 5520 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 5521 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 5521 | L: kvmarm@lists.cs.columbia.edu | 5522 | L: kvmarm@lists.cs.columbia.edu |
| 5522 | W: http://systems.cs.columbia.edu/projects/kvm-arm | 5523 | W: http://systems.cs.columbia.edu/projects/kvm-arm |
| 5523 | S: Supported | 5524 | S: Supported |
| 5524 | F: arch/arm/include/uapi/asm/kvm* | 5525 | F: arch/arm/include/uapi/asm/kvm* |
| 5525 | F: arch/arm/include/asm/kvm* | 5526 | F: arch/arm/include/asm/kvm* |
| 5526 | F: arch/arm/kvm/ | 5527 | F: arch/arm/kvm/ |
| 5527 | F: virt/kvm/arm/ | 5528 | F: virt/kvm/arm/ |
| 5528 | F: include/kvm/arm_* | 5529 | F: include/kvm/arm_* |
| 5529 | 5530 | ||
| 5530 | KERNEL VIRTUAL MACHINE FOR ARM64 (KVM/arm64) | 5531 | KERNEL VIRTUAL MACHINE FOR ARM64 (KVM/arm64) |
| 5531 | M: Christoffer Dall <christoffer.dall@linaro.org> | 5532 | M: Christoffer Dall <christoffer.dall@linaro.org> |
| 5532 | M: Marc Zyngier <marc.zyngier@arm.com> | 5533 | M: Marc Zyngier <marc.zyngier@arm.com> |
| 5533 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 5534 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 5534 | L: kvmarm@lists.cs.columbia.edu | 5535 | L: kvmarm@lists.cs.columbia.edu |
| 5535 | S: Maintained | 5536 | S: Maintained |
| 5536 | F: arch/arm64/include/uapi/asm/kvm* | 5537 | F: arch/arm64/include/uapi/asm/kvm* |
| 5537 | F: arch/arm64/include/asm/kvm* | 5538 | F: arch/arm64/include/asm/kvm* |
| 5538 | F: arch/arm64/kvm/ | 5539 | F: arch/arm64/kvm/ |
| 5539 | 5540 | ||
| 5540 | KEXEC | 5541 | KEXEC |
| 5541 | M: Eric Biederman <ebiederm@xmission.com> | 5542 | M: Eric Biederman <ebiederm@xmission.com> |
| 5542 | W: http://kernel.org/pub/linux/utils/kernel/kexec/ | 5543 | W: http://kernel.org/pub/linux/utils/kernel/kexec/ |
| 5543 | L: kexec@lists.infradead.org | 5544 | L: kexec@lists.infradead.org |
| 5544 | S: Maintained | 5545 | S: Maintained |
| 5545 | F: include/linux/kexec.h | 5546 | F: include/linux/kexec.h |
| 5546 | F: include/uapi/linux/kexec.h | 5547 | F: include/uapi/linux/kexec.h |
| 5547 | F: kernel/kexec.c | 5548 | F: kernel/kexec.c |
| 5548 | 5549 | ||
| 5549 | KEYS/KEYRINGS: | 5550 | KEYS/KEYRINGS: |
| 5550 | M: David Howells <dhowells@redhat.com> | 5551 | M: David Howells <dhowells@redhat.com> |
| 5551 | L: keyrings@linux-nfs.org | 5552 | L: keyrings@linux-nfs.org |
| 5552 | S: Maintained | 5553 | S: Maintained |
| 5553 | F: Documentation/security/keys.txt | 5554 | F: Documentation/security/keys.txt |
| 5554 | F: include/linux/key.h | 5555 | F: include/linux/key.h |
| 5555 | F: include/linux/key-type.h | 5556 | F: include/linux/key-type.h |
| 5556 | F: include/keys/ | 5557 | F: include/keys/ |
| 5557 | F: security/keys/ | 5558 | F: security/keys/ |
| 5558 | 5559 | ||
| 5559 | KEYS-TRUSTED | 5560 | KEYS-TRUSTED |
| 5560 | M: David Safford <safford@us.ibm.com> | 5561 | M: David Safford <safford@us.ibm.com> |
| 5561 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> | 5562 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> |
| 5562 | L: linux-security-module@vger.kernel.org | 5563 | L: linux-security-module@vger.kernel.org |
| 5563 | L: keyrings@linux-nfs.org | 5564 | L: keyrings@linux-nfs.org |
| 5564 | S: Supported | 5565 | S: Supported |
| 5565 | F: Documentation/security/keys-trusted-encrypted.txt | 5566 | F: Documentation/security/keys-trusted-encrypted.txt |
| 5566 | F: include/keys/trusted-type.h | 5567 | F: include/keys/trusted-type.h |
| 5567 | F: security/keys/trusted.c | 5568 | F: security/keys/trusted.c |
| 5568 | F: security/keys/trusted.h | 5569 | F: security/keys/trusted.h |
| 5569 | 5570 | ||
| 5570 | KEYS-ENCRYPTED | 5571 | KEYS-ENCRYPTED |
| 5571 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> | 5572 | M: Mimi Zohar <zohar@linux.vnet.ibm.com> |
| 5572 | M: David Safford <safford@us.ibm.com> | 5573 | M: David Safford <safford@us.ibm.com> |
| 5573 | L: linux-security-module@vger.kernel.org | 5574 | L: linux-security-module@vger.kernel.org |
| 5574 | L: keyrings@linux-nfs.org | 5575 | L: keyrings@linux-nfs.org |
| 5575 | S: Supported | 5576 | S: Supported |
| 5576 | F: Documentation/security/keys-trusted-encrypted.txt | 5577 | F: Documentation/security/keys-trusted-encrypted.txt |
| 5577 | F: include/keys/encrypted-type.h | 5578 | F: include/keys/encrypted-type.h |
| 5578 | F: security/keys/encrypted-keys/ | 5579 | F: security/keys/encrypted-keys/ |
| 5579 | 5580 | ||
| 5580 | KGDB / KDB /debug_core | 5581 | KGDB / KDB /debug_core |
| 5581 | M: Jason Wessel <jason.wessel@windriver.com> | 5582 | M: Jason Wessel <jason.wessel@windriver.com> |
| 5582 | W: http://kgdb.wiki.kernel.org/ | 5583 | W: http://kgdb.wiki.kernel.org/ |
| 5583 | L: kgdb-bugreport@lists.sourceforge.net | 5584 | L: kgdb-bugreport@lists.sourceforge.net |
| 5584 | S: Maintained | 5585 | S: Maintained |
| 5585 | F: Documentation/DocBook/kgdb.tmpl | 5586 | F: Documentation/DocBook/kgdb.tmpl |
| 5586 | F: drivers/misc/kgdbts.c | 5587 | F: drivers/misc/kgdbts.c |
| 5587 | F: drivers/tty/serial/kgdboc.c | 5588 | F: drivers/tty/serial/kgdboc.c |
| 5588 | F: include/linux/kdb.h | 5589 | F: include/linux/kdb.h |
| 5589 | F: include/linux/kgdb.h | 5590 | F: include/linux/kgdb.h |
| 5590 | F: kernel/debug/ | 5591 | F: kernel/debug/ |
| 5591 | 5592 | ||
| 5592 | KMEMCHECK | 5593 | KMEMCHECK |
| 5593 | M: Vegard Nossum <vegardno@ifi.uio.no> | 5594 | M: Vegard Nossum <vegardno@ifi.uio.no> |
| 5594 | M: Pekka Enberg <penberg@kernel.org> | 5595 | M: Pekka Enberg <penberg@kernel.org> |
| 5595 | S: Maintained | 5596 | S: Maintained |
| 5596 | F: Documentation/kmemcheck.txt | 5597 | F: Documentation/kmemcheck.txt |
| 5597 | F: arch/x86/include/asm/kmemcheck.h | 5598 | F: arch/x86/include/asm/kmemcheck.h |
| 5598 | F: arch/x86/mm/kmemcheck/ | 5599 | F: arch/x86/mm/kmemcheck/ |
| 5599 | F: include/linux/kmemcheck.h | 5600 | F: include/linux/kmemcheck.h |
| 5600 | F: mm/kmemcheck.c | 5601 | F: mm/kmemcheck.c |
| 5601 | 5602 | ||
| 5602 | KMEMLEAK | 5603 | KMEMLEAK |
| 5603 | M: Catalin Marinas <catalin.marinas@arm.com> | 5604 | M: Catalin Marinas <catalin.marinas@arm.com> |
| 5604 | S: Maintained | 5605 | S: Maintained |
| 5605 | F: Documentation/kmemleak.txt | 5606 | F: Documentation/kmemleak.txt |
| 5606 | F: include/linux/kmemleak.h | 5607 | F: include/linux/kmemleak.h |
| 5607 | F: mm/kmemleak.c | 5608 | F: mm/kmemleak.c |
| 5608 | F: mm/kmemleak-test.c | 5609 | F: mm/kmemleak-test.c |
| 5609 | 5610 | ||
| 5610 | KPROBES | 5611 | KPROBES |
| 5611 | M: Ananth N Mavinakayanahalli <ananth@in.ibm.com> | 5612 | M: Ananth N Mavinakayanahalli <ananth@in.ibm.com> |
| 5612 | M: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> | 5613 | M: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com> |
| 5613 | M: "David S. Miller" <davem@davemloft.net> | 5614 | M: "David S. Miller" <davem@davemloft.net> |
| 5614 | M: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> | 5615 | M: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> |
| 5615 | S: Maintained | 5616 | S: Maintained |
| 5616 | F: Documentation/kprobes.txt | 5617 | F: Documentation/kprobes.txt |
| 5617 | F: include/linux/kprobes.h | 5618 | F: include/linux/kprobes.h |
| 5618 | F: kernel/kprobes.c | 5619 | F: kernel/kprobes.c |
| 5619 | 5620 | ||
| 5620 | KS0108 LCD CONTROLLER DRIVER | 5621 | KS0108 LCD CONTROLLER DRIVER |
| 5621 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> | 5622 | M: Miguel Ojeda Sandonis <miguel.ojeda.sandonis@gmail.com> |
| 5622 | W: http://miguelojeda.es/auxdisplay.htm | 5623 | W: http://miguelojeda.es/auxdisplay.htm |
| 5623 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm | 5624 | W: http://jair.lab.fi.uva.es/~migojed/auxdisplay.htm |
| 5624 | S: Maintained | 5625 | S: Maintained |
| 5625 | F: Documentation/auxdisplay/ks0108 | 5626 | F: Documentation/auxdisplay/ks0108 |
| 5626 | F: drivers/auxdisplay/ks0108.c | 5627 | F: drivers/auxdisplay/ks0108.c |
| 5627 | F: include/linux/ks0108.h | 5628 | F: include/linux/ks0108.h |
| 5628 | 5629 | ||
| 5629 | LAPB module | 5630 | LAPB module |
| 5630 | L: linux-x25@vger.kernel.org | 5631 | L: linux-x25@vger.kernel.org |
| 5631 | S: Orphan | 5632 | S: Orphan |
| 5632 | F: Documentation/networking/lapb-module.txt | 5633 | F: Documentation/networking/lapb-module.txt |
| 5633 | F: include/*/lapb.h | 5634 | F: include/*/lapb.h |
| 5634 | F: net/lapb/ | 5635 | F: net/lapb/ |
| 5635 | 5636 | ||
| 5636 | LASI 53c700 driver for PARISC | 5637 | LASI 53c700 driver for PARISC |
| 5637 | M: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> | 5638 | M: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> |
| 5638 | L: linux-scsi@vger.kernel.org | 5639 | L: linux-scsi@vger.kernel.org |
| 5639 | S: Maintained | 5640 | S: Maintained |
| 5640 | F: Documentation/scsi/53c700.txt | 5641 | F: Documentation/scsi/53c700.txt |
| 5641 | F: drivers/scsi/53c700* | 5642 | F: drivers/scsi/53c700* |
| 5642 | 5643 | ||
| 5643 | LED SUBSYSTEM | 5644 | LED SUBSYSTEM |
| 5644 | M: Bryan Wu <cooloney@gmail.com> | 5645 | M: Bryan Wu <cooloney@gmail.com> |
| 5645 | M: Richard Purdie <rpurdie@rpsys.net> | 5646 | M: Richard Purdie <rpurdie@rpsys.net> |
| 5646 | L: linux-leds@vger.kernel.org | 5647 | L: linux-leds@vger.kernel.org |
| 5647 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/linux-leds.git | 5648 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/linux-leds.git |
| 5648 | S: Maintained | 5649 | S: Maintained |
| 5649 | F: drivers/leds/ | 5650 | F: drivers/leds/ |
| 5650 | F: include/linux/leds.h | 5651 | F: include/linux/leds.h |
| 5651 | 5652 | ||
| 5652 | LEGACY EEPROM DRIVER | 5653 | LEGACY EEPROM DRIVER |
| 5653 | M: Jean Delvare <jdelvare@suse.de> | 5654 | M: Jean Delvare <jdelvare@suse.de> |
| 5654 | S: Maintained | 5655 | S: Maintained |
| 5655 | F: Documentation/misc-devices/eeprom | 5656 | F: Documentation/misc-devices/eeprom |
| 5656 | F: drivers/misc/eeprom/eeprom.c | 5657 | F: drivers/misc/eeprom/eeprom.c |
| 5657 | 5658 | ||
| 5658 | LEGO USB Tower driver | 5659 | LEGO USB Tower driver |
| 5659 | M: Juergen Stuber <starblue@users.sourceforge.net> | 5660 | M: Juergen Stuber <starblue@users.sourceforge.net> |
| 5660 | L: legousb-devel@lists.sourceforge.net | 5661 | L: legousb-devel@lists.sourceforge.net |
| 5661 | W: http://legousb.sourceforge.net/ | 5662 | W: http://legousb.sourceforge.net/ |
| 5662 | S: Maintained | 5663 | S: Maintained |
| 5663 | F: drivers/usb/misc/legousbtower.c | 5664 | F: drivers/usb/misc/legousbtower.c |
| 5664 | 5665 | ||
| 5665 | LG2160 MEDIA DRIVER | 5666 | LG2160 MEDIA DRIVER |
| 5666 | M: Michael Krufky <mkrufky@linuxtv.org> | 5667 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 5667 | L: linux-media@vger.kernel.org | 5668 | L: linux-media@vger.kernel.org |
| 5668 | W: http://linuxtv.org/ | 5669 | W: http://linuxtv.org/ |
| 5669 | W: http://github.com/mkrufky | 5670 | W: http://github.com/mkrufky |
| 5670 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5671 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5671 | T: git git://linuxtv.org/mkrufky/tuners.git | 5672 | T: git git://linuxtv.org/mkrufky/tuners.git |
| 5672 | S: Maintained | 5673 | S: Maintained |
| 5673 | F: drivers/media/dvb-frontends/lg2160.* | 5674 | F: drivers/media/dvb-frontends/lg2160.* |
| 5674 | 5675 | ||
| 5675 | LGDT3305 MEDIA DRIVER | 5676 | LGDT3305 MEDIA DRIVER |
| 5676 | M: Michael Krufky <mkrufky@linuxtv.org> | 5677 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 5677 | L: linux-media@vger.kernel.org | 5678 | L: linux-media@vger.kernel.org |
| 5678 | W: http://linuxtv.org/ | 5679 | W: http://linuxtv.org/ |
| 5679 | W: http://github.com/mkrufky | 5680 | W: http://github.com/mkrufky |
| 5680 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5681 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5681 | T: git git://linuxtv.org/mkrufky/tuners.git | 5682 | T: git git://linuxtv.org/mkrufky/tuners.git |
| 5682 | S: Maintained | 5683 | S: Maintained |
| 5683 | F: drivers/media/dvb-frontends/lgdt3305.* | 5684 | F: drivers/media/dvb-frontends/lgdt3305.* |
| 5684 | 5685 | ||
| 5685 | LGUEST | 5686 | LGUEST |
| 5686 | M: Rusty Russell <rusty@rustcorp.com.au> | 5687 | M: Rusty Russell <rusty@rustcorp.com.au> |
| 5687 | L: lguest@lists.ozlabs.org | 5688 | L: lguest@lists.ozlabs.org |
| 5688 | W: http://lguest.ozlabs.org/ | 5689 | W: http://lguest.ozlabs.org/ |
| 5689 | S: Odd Fixes | 5690 | S: Odd Fixes |
| 5690 | F: arch/x86/include/asm/lguest*.h | 5691 | F: arch/x86/include/asm/lguest*.h |
| 5691 | F: arch/x86/lguest/ | 5692 | F: arch/x86/lguest/ |
| 5692 | F: drivers/lguest/ | 5693 | F: drivers/lguest/ |
| 5693 | F: include/linux/lguest*.h | 5694 | F: include/linux/lguest*.h |
| 5694 | F: tools/lguest/ | 5695 | F: tools/lguest/ |
| 5695 | 5696 | ||
| 5696 | LIBLOCKDEP | 5697 | LIBLOCKDEP |
| 5697 | M: Sasha Levin <sasha.levin@oracle.com> | 5698 | M: Sasha Levin <sasha.levin@oracle.com> |
| 5698 | S: Maintained | 5699 | S: Maintained |
| 5699 | F: tools/lib/lockdep/ | 5700 | F: tools/lib/lockdep/ |
| 5700 | 5701 | ||
| 5701 | LINUX FOR IBM pSERIES (RS/6000) | 5702 | LINUX FOR IBM pSERIES (RS/6000) |
| 5702 | M: Paul Mackerras <paulus@au.ibm.com> | 5703 | M: Paul Mackerras <paulus@au.ibm.com> |
| 5703 | W: http://www.ibm.com/linux/ltc/projects/ppc | 5704 | W: http://www.ibm.com/linux/ltc/projects/ppc |
| 5704 | S: Supported | 5705 | S: Supported |
| 5705 | F: arch/powerpc/boot/rs6000.h | 5706 | F: arch/powerpc/boot/rs6000.h |
| 5706 | 5707 | ||
| 5707 | LINUX FOR POWERPC (32-BIT AND 64-BIT) | 5708 | LINUX FOR POWERPC (32-BIT AND 64-BIT) |
| 5708 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> | 5709 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> |
| 5709 | M: Paul Mackerras <paulus@samba.org> | 5710 | M: Paul Mackerras <paulus@samba.org> |
| 5710 | M: Michael Ellerman <mpe@ellerman.id.au> | 5711 | M: Michael Ellerman <mpe@ellerman.id.au> |
| 5711 | W: http://www.penguinppc.org/ | 5712 | W: http://www.penguinppc.org/ |
| 5712 | L: linuxppc-dev@lists.ozlabs.org | 5713 | L: linuxppc-dev@lists.ozlabs.org |
| 5713 | Q: http://patchwork.ozlabs.org/project/linuxppc-dev/list/ | 5714 | Q: http://patchwork.ozlabs.org/project/linuxppc-dev/list/ |
| 5714 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc.git | 5715 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc.git |
| 5715 | S: Supported | 5716 | S: Supported |
| 5716 | F: Documentation/powerpc/ | 5717 | F: Documentation/powerpc/ |
| 5717 | F: arch/powerpc/ | 5718 | F: arch/powerpc/ |
| 5718 | 5719 | ||
| 5719 | LINUX FOR POWER MACINTOSH | 5720 | LINUX FOR POWER MACINTOSH |
| 5720 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> | 5721 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> |
| 5721 | W: http://www.penguinppc.org/ | 5722 | W: http://www.penguinppc.org/ |
| 5722 | L: linuxppc-dev@lists.ozlabs.org | 5723 | L: linuxppc-dev@lists.ozlabs.org |
| 5723 | S: Maintained | 5724 | S: Maintained |
| 5724 | F: arch/powerpc/platforms/powermac/ | 5725 | F: arch/powerpc/platforms/powermac/ |
| 5725 | F: drivers/macintosh/ | 5726 | F: drivers/macintosh/ |
| 5726 | 5727 | ||
| 5727 | LINUX FOR POWERPC EMBEDDED MPC5XXX | 5728 | LINUX FOR POWERPC EMBEDDED MPC5XXX |
| 5728 | M: Anatolij Gustschin <agust@denx.de> | 5729 | M: Anatolij Gustschin <agust@denx.de> |
| 5729 | L: linuxppc-dev@lists.ozlabs.org | 5730 | L: linuxppc-dev@lists.ozlabs.org |
| 5730 | T: git git://git.denx.de/linux-denx-agust.git | 5731 | T: git git://git.denx.de/linux-denx-agust.git |
| 5731 | S: Maintained | 5732 | S: Maintained |
| 5732 | F: arch/powerpc/platforms/512x/ | 5733 | F: arch/powerpc/platforms/512x/ |
| 5733 | F: arch/powerpc/platforms/52xx/ | 5734 | F: arch/powerpc/platforms/52xx/ |
| 5734 | 5735 | ||
| 5735 | LINUX FOR POWERPC EMBEDDED PPC4XX | 5736 | LINUX FOR POWERPC EMBEDDED PPC4XX |
| 5736 | M: Alistair Popple <alistair@popple.id.au> | 5737 | M: Alistair Popple <alistair@popple.id.au> |
| 5737 | M: Matt Porter <mporter@kernel.crashing.org> | 5738 | M: Matt Porter <mporter@kernel.crashing.org> |
| 5738 | W: http://www.penguinppc.org/ | 5739 | W: http://www.penguinppc.org/ |
| 5739 | L: linuxppc-dev@lists.ozlabs.org | 5740 | L: linuxppc-dev@lists.ozlabs.org |
| 5740 | S: Maintained | 5741 | S: Maintained |
| 5741 | F: arch/powerpc/platforms/40x/ | 5742 | F: arch/powerpc/platforms/40x/ |
| 5742 | F: arch/powerpc/platforms/44x/ | 5743 | F: arch/powerpc/platforms/44x/ |
| 5743 | 5744 | ||
| 5744 | LINUX FOR POWERPC EMBEDDED XILINX VIRTEX | 5745 | LINUX FOR POWERPC EMBEDDED XILINX VIRTEX |
| 5745 | L: linuxppc-dev@lists.ozlabs.org | 5746 | L: linuxppc-dev@lists.ozlabs.org |
| 5746 | S: Orphan | 5747 | S: Orphan |
| 5747 | F: arch/powerpc/*/*virtex* | 5748 | F: arch/powerpc/*/*virtex* |
| 5748 | F: arch/powerpc/*/*/*virtex* | 5749 | F: arch/powerpc/*/*/*virtex* |
| 5749 | 5750 | ||
| 5750 | LINUX FOR POWERPC EMBEDDED PPC8XX | 5751 | LINUX FOR POWERPC EMBEDDED PPC8XX |
| 5751 | M: Vitaly Bordug <vitb@kernel.crashing.org> | 5752 | M: Vitaly Bordug <vitb@kernel.crashing.org> |
| 5752 | W: http://www.penguinppc.org/ | 5753 | W: http://www.penguinppc.org/ |
| 5753 | L: linuxppc-dev@lists.ozlabs.org | 5754 | L: linuxppc-dev@lists.ozlabs.org |
| 5754 | S: Maintained | 5755 | S: Maintained |
| 5755 | F: arch/powerpc/platforms/8xx/ | 5756 | F: arch/powerpc/platforms/8xx/ |
| 5756 | 5757 | ||
| 5757 | LINUX FOR POWERPC EMBEDDED PPC83XX AND PPC85XX | 5758 | LINUX FOR POWERPC EMBEDDED PPC83XX AND PPC85XX |
| 5758 | M: Scott Wood <scottwood@freescale.com> | 5759 | M: Scott Wood <scottwood@freescale.com> |
| 5759 | M: Kumar Gala <galak@kernel.crashing.org> | 5760 | M: Kumar Gala <galak@kernel.crashing.org> |
| 5760 | W: http://www.penguinppc.org/ | 5761 | W: http://www.penguinppc.org/ |
| 5761 | L: linuxppc-dev@lists.ozlabs.org | 5762 | L: linuxppc-dev@lists.ozlabs.org |
| 5762 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/scottwood/linux.git | 5763 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/scottwood/linux.git |
| 5763 | S: Maintained | 5764 | S: Maintained |
| 5764 | F: arch/powerpc/platforms/83xx/ | 5765 | F: arch/powerpc/platforms/83xx/ |
| 5765 | F: arch/powerpc/platforms/85xx/ | 5766 | F: arch/powerpc/platforms/85xx/ |
| 5766 | 5767 | ||
| 5767 | LINUX FOR POWERPC PA SEMI PWRFICIENT | 5768 | LINUX FOR POWERPC PA SEMI PWRFICIENT |
| 5768 | M: Olof Johansson <olof@lixom.net> | 5769 | M: Olof Johansson <olof@lixom.net> |
| 5769 | L: linuxppc-dev@lists.ozlabs.org | 5770 | L: linuxppc-dev@lists.ozlabs.org |
| 5770 | S: Maintained | 5771 | S: Maintained |
| 5771 | F: arch/powerpc/platforms/pasemi/ | 5772 | F: arch/powerpc/platforms/pasemi/ |
| 5772 | F: drivers/*/*pasemi* | 5773 | F: drivers/*/*pasemi* |
| 5773 | F: drivers/*/*/*pasemi* | 5774 | F: drivers/*/*/*pasemi* |
| 5774 | 5775 | ||
| 5775 | LINUX SECURITY MODULE (LSM) FRAMEWORK | 5776 | LINUX SECURITY MODULE (LSM) FRAMEWORK |
| 5776 | M: Chris Wright <chrisw@sous-sol.org> | 5777 | M: Chris Wright <chrisw@sous-sol.org> |
| 5777 | L: linux-security-module@vger.kernel.org | 5778 | L: linux-security-module@vger.kernel.org |
| 5778 | S: Supported | 5779 | S: Supported |
| 5779 | 5780 | ||
| 5780 | LIS3LV02D ACCELEROMETER DRIVER | 5781 | LIS3LV02D ACCELEROMETER DRIVER |
| 5781 | M: Eric Piel <eric.piel@tremplin-utc.net> | 5782 | M: Eric Piel <eric.piel@tremplin-utc.net> |
| 5782 | S: Maintained | 5783 | S: Maintained |
| 5783 | F: Documentation/misc-devices/lis3lv02d | 5784 | F: Documentation/misc-devices/lis3lv02d |
| 5784 | F: drivers/misc/lis3lv02d/ | 5785 | F: drivers/misc/lis3lv02d/ |
| 5785 | F: drivers/platform/x86/hp_accel.c | 5786 | F: drivers/platform/x86/hp_accel.c |
| 5786 | 5787 | ||
| 5787 | LLC (802.2) | 5788 | LLC (802.2) |
| 5788 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> | 5789 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> |
| 5789 | S: Maintained | 5790 | S: Maintained |
| 5790 | F: include/linux/llc.h | 5791 | F: include/linux/llc.h |
| 5791 | F: include/uapi/linux/llc.h | 5792 | F: include/uapi/linux/llc.h |
| 5792 | F: include/net/llc* | 5793 | F: include/net/llc* |
| 5793 | F: net/llc/ | 5794 | F: net/llc/ |
| 5794 | 5795 | ||
| 5795 | LM73 HARDWARE MONITOR DRIVER | 5796 | LM73 HARDWARE MONITOR DRIVER |
| 5796 | M: Guillaume Ligneul <guillaume.ligneul@gmail.com> | 5797 | M: Guillaume Ligneul <guillaume.ligneul@gmail.com> |
| 5797 | L: lm-sensors@lm-sensors.org | 5798 | L: lm-sensors@lm-sensors.org |
| 5798 | S: Maintained | 5799 | S: Maintained |
| 5799 | F: drivers/hwmon/lm73.c | 5800 | F: drivers/hwmon/lm73.c |
| 5800 | 5801 | ||
| 5801 | LM78 HARDWARE MONITOR DRIVER | 5802 | LM78 HARDWARE MONITOR DRIVER |
| 5802 | M: Jean Delvare <jdelvare@suse.de> | 5803 | M: Jean Delvare <jdelvare@suse.de> |
| 5803 | L: lm-sensors@lm-sensors.org | 5804 | L: lm-sensors@lm-sensors.org |
| 5804 | S: Maintained | 5805 | S: Maintained |
| 5805 | F: Documentation/hwmon/lm78 | 5806 | F: Documentation/hwmon/lm78 |
| 5806 | F: drivers/hwmon/lm78.c | 5807 | F: drivers/hwmon/lm78.c |
| 5807 | 5808 | ||
| 5808 | LM83 HARDWARE MONITOR DRIVER | 5809 | LM83 HARDWARE MONITOR DRIVER |
| 5809 | M: Jean Delvare <jdelvare@suse.de> | 5810 | M: Jean Delvare <jdelvare@suse.de> |
| 5810 | L: lm-sensors@lm-sensors.org | 5811 | L: lm-sensors@lm-sensors.org |
| 5811 | S: Maintained | 5812 | S: Maintained |
| 5812 | F: Documentation/hwmon/lm83 | 5813 | F: Documentation/hwmon/lm83 |
| 5813 | F: drivers/hwmon/lm83.c | 5814 | F: drivers/hwmon/lm83.c |
| 5814 | 5815 | ||
| 5815 | LM90 HARDWARE MONITOR DRIVER | 5816 | LM90 HARDWARE MONITOR DRIVER |
| 5816 | M: Jean Delvare <jdelvare@suse.de> | 5817 | M: Jean Delvare <jdelvare@suse.de> |
| 5817 | L: lm-sensors@lm-sensors.org | 5818 | L: lm-sensors@lm-sensors.org |
| 5818 | S: Maintained | 5819 | S: Maintained |
| 5819 | F: Documentation/hwmon/lm90 | 5820 | F: Documentation/hwmon/lm90 |
| 5820 | F: Documentation/devicetree/bindings/hwmon/lm90.txt | 5821 | F: Documentation/devicetree/bindings/hwmon/lm90.txt |
| 5821 | F: drivers/hwmon/lm90.c | 5822 | F: drivers/hwmon/lm90.c |
| 5822 | 5823 | ||
| 5823 | LM95234 HARDWARE MONITOR DRIVER | 5824 | LM95234 HARDWARE MONITOR DRIVER |
| 5824 | M: Guenter Roeck <linux@roeck-us.net> | 5825 | M: Guenter Roeck <linux@roeck-us.net> |
| 5825 | L: lm-sensors@lm-sensors.org | 5826 | L: lm-sensors@lm-sensors.org |
| 5826 | S: Maintained | 5827 | S: Maintained |
| 5827 | F: Documentation/hwmon/lm95234 | 5828 | F: Documentation/hwmon/lm95234 |
| 5828 | F: drivers/hwmon/lm95234.c | 5829 | F: drivers/hwmon/lm95234.c |
| 5829 | 5830 | ||
| 5830 | LME2510 MEDIA DRIVER | 5831 | LME2510 MEDIA DRIVER |
| 5831 | M: Malcolm Priestley <tvboxspy@gmail.com> | 5832 | M: Malcolm Priestley <tvboxspy@gmail.com> |
| 5832 | L: linux-media@vger.kernel.org | 5833 | L: linux-media@vger.kernel.org |
| 5833 | W: http://linuxtv.org/ | 5834 | W: http://linuxtv.org/ |
| 5834 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5835 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5835 | S: Maintained | 5836 | S: Maintained |
| 5836 | F: drivers/media/usb/dvb-usb-v2/lmedm04* | 5837 | F: drivers/media/usb/dvb-usb-v2/lmedm04* |
| 5837 | 5838 | ||
| 5838 | LOCKDEP AND LOCKSTAT | 5839 | LOCKDEP AND LOCKSTAT |
| 5839 | M: Peter Zijlstra <peterz@infradead.org> | 5840 | M: Peter Zijlstra <peterz@infradead.org> |
| 5840 | M: Ingo Molnar <mingo@redhat.com> | 5841 | M: Ingo Molnar <mingo@redhat.com> |
| 5841 | L: linux-kernel@vger.kernel.org | 5842 | L: linux-kernel@vger.kernel.org |
| 5842 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git core/locking | 5843 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git core/locking |
| 5843 | S: Maintained | 5844 | S: Maintained |
| 5844 | F: Documentation/locking/lockdep*.txt | 5845 | F: Documentation/locking/lockdep*.txt |
| 5845 | F: Documentation/locking/lockstat.txt | 5846 | F: Documentation/locking/lockstat.txt |
| 5846 | F: include/linux/lockdep.h | 5847 | F: include/linux/lockdep.h |
| 5847 | F: kernel/locking/ | 5848 | F: kernel/locking/ |
| 5848 | 5849 | ||
| 5849 | LOGICAL DISK MANAGER SUPPORT (LDM, Windows 2000/XP/Vista Dynamic Disks) | 5850 | LOGICAL DISK MANAGER SUPPORT (LDM, Windows 2000/XP/Vista Dynamic Disks) |
| 5850 | M: "Richard Russon (FlatCap)" <ldm@flatcap.org> | 5851 | M: "Richard Russon (FlatCap)" <ldm@flatcap.org> |
| 5851 | L: linux-ntfs-dev@lists.sourceforge.net | 5852 | L: linux-ntfs-dev@lists.sourceforge.net |
| 5852 | W: http://www.linux-ntfs.org/content/view/19/37/ | 5853 | W: http://www.linux-ntfs.org/content/view/19/37/ |
| 5853 | S: Maintained | 5854 | S: Maintained |
| 5854 | F: Documentation/ldm.txt | 5855 | F: Documentation/ldm.txt |
| 5855 | F: block/partitions/ldm.* | 5856 | F: block/partitions/ldm.* |
| 5856 | 5857 | ||
| 5857 | LogFS | 5858 | LogFS |
| 5858 | M: Joern Engel <joern@logfs.org> | 5859 | M: Joern Engel <joern@logfs.org> |
| 5859 | M: Prasad Joshi <prasadjoshi.linux@gmail.com> | 5860 | M: Prasad Joshi <prasadjoshi.linux@gmail.com> |
| 5860 | L: logfs@logfs.org | 5861 | L: logfs@logfs.org |
| 5861 | W: logfs.org | 5862 | W: logfs.org |
| 5862 | S: Maintained | 5863 | S: Maintained |
| 5863 | F: fs/logfs/ | 5864 | F: fs/logfs/ |
| 5864 | 5865 | ||
| 5865 | LPC32XX MACHINE SUPPORT | 5866 | LPC32XX MACHINE SUPPORT |
| 5866 | M: Roland Stigge <stigge@antcom.de> | 5867 | M: Roland Stigge <stigge@antcom.de> |
| 5867 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 5868 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 5868 | S: Maintained | 5869 | S: Maintained |
| 5869 | F: arch/arm/mach-lpc32xx/ | 5870 | F: arch/arm/mach-lpc32xx/ |
| 5870 | 5871 | ||
| 5871 | LSILOGIC MPT FUSION DRIVERS (FC/SAS/SPI) | 5872 | LSILOGIC MPT FUSION DRIVERS (FC/SAS/SPI) |
| 5872 | M: Nagalakshmi Nandigama <nagalakshmi.nandigama@avagotech.com> | 5873 | M: Nagalakshmi Nandigama <nagalakshmi.nandigama@avagotech.com> |
| 5873 | M: Praveen Krishnamoorthy <praveen.krishnamoorthy@avagotech.com> | 5874 | M: Praveen Krishnamoorthy <praveen.krishnamoorthy@avagotech.com> |
| 5874 | M: Sreekanth Reddy <sreekanth.reddy@avagotech.com> | 5875 | M: Sreekanth Reddy <sreekanth.reddy@avagotech.com> |
| 5875 | M: Abhijit Mahajan <abhijit.mahajan@avagotech.com> | 5876 | M: Abhijit Mahajan <abhijit.mahajan@avagotech.com> |
| 5876 | L: MPT-FusionLinux.pdl@avagotech.com | 5877 | L: MPT-FusionLinux.pdl@avagotech.com |
| 5877 | L: linux-scsi@vger.kernel.org | 5878 | L: linux-scsi@vger.kernel.org |
| 5878 | W: http://www.lsilogic.com/support | 5879 | W: http://www.lsilogic.com/support |
| 5879 | S: Supported | 5880 | S: Supported |
| 5880 | F: drivers/message/fusion/ | 5881 | F: drivers/message/fusion/ |
| 5881 | F: drivers/scsi/mpt2sas/ | 5882 | F: drivers/scsi/mpt2sas/ |
| 5882 | F: drivers/scsi/mpt3sas/ | 5883 | F: drivers/scsi/mpt3sas/ |
| 5883 | 5884 | ||
| 5884 | LSILOGIC/SYMBIOS/NCR 53C8XX and 53C1010 PCI-SCSI drivers | 5885 | LSILOGIC/SYMBIOS/NCR 53C8XX and 53C1010 PCI-SCSI drivers |
| 5885 | M: Matthew Wilcox <matthew@wil.cx> | 5886 | M: Matthew Wilcox <matthew@wil.cx> |
| 5886 | L: linux-scsi@vger.kernel.org | 5887 | L: linux-scsi@vger.kernel.org |
| 5887 | S: Maintained | 5888 | S: Maintained |
| 5888 | F: drivers/scsi/sym53c8xx_2/ | 5889 | F: drivers/scsi/sym53c8xx_2/ |
| 5889 | 5890 | ||
| 5890 | LTC4261 HARDWARE MONITOR DRIVER | 5891 | LTC4261 HARDWARE MONITOR DRIVER |
| 5891 | M: Guenter Roeck <linux@roeck-us.net> | 5892 | M: Guenter Roeck <linux@roeck-us.net> |
| 5892 | L: lm-sensors@lm-sensors.org | 5893 | L: lm-sensors@lm-sensors.org |
| 5893 | S: Maintained | 5894 | S: Maintained |
| 5894 | F: Documentation/hwmon/ltc4261 | 5895 | F: Documentation/hwmon/ltc4261 |
| 5895 | F: drivers/hwmon/ltc4261.c | 5896 | F: drivers/hwmon/ltc4261.c |
| 5896 | 5897 | ||
| 5897 | LTP (Linux Test Project) | 5898 | LTP (Linux Test Project) |
| 5898 | M: Mike Frysinger <vapier@gentoo.org> | 5899 | M: Mike Frysinger <vapier@gentoo.org> |
| 5899 | M: Cyril Hrubis <chrubis@suse.cz> | 5900 | M: Cyril Hrubis <chrubis@suse.cz> |
| 5900 | M: Wanlong Gao <gaowanlong@cn.fujitsu.com> | 5901 | M: Wanlong Gao <gaowanlong@cn.fujitsu.com> |
| 5901 | M: Jan Stancek <jstancek@redhat.com> | 5902 | M: Jan Stancek <jstancek@redhat.com> |
| 5902 | M: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com> | 5903 | M: Stanislav Kholmanskikh <stanislav.kholmanskikh@oracle.com> |
| 5903 | M: Alexey Kodanev <alexey.kodanev@oracle.com> | 5904 | M: Alexey Kodanev <alexey.kodanev@oracle.com> |
| 5904 | L: ltp-list@lists.sourceforge.net (subscribers-only) | 5905 | L: ltp-list@lists.sourceforge.net (subscribers-only) |
| 5905 | W: http://linux-test-project.github.io/ | 5906 | W: http://linux-test-project.github.io/ |
| 5906 | T: git git://github.com/linux-test-project/ltp.git | 5907 | T: git git://github.com/linux-test-project/ltp.git |
| 5907 | S: Maintained | 5908 | S: Maintained |
| 5908 | 5909 | ||
| 5909 | M32R ARCHITECTURE | 5910 | M32R ARCHITECTURE |
| 5910 | W: http://www.linux-m32r.org/ | 5911 | W: http://www.linux-m32r.org/ |
| 5911 | S: Orphan | 5912 | S: Orphan |
| 5912 | F: arch/m32r/ | 5913 | F: arch/m32r/ |
| 5913 | 5914 | ||
| 5914 | M68K ARCHITECTURE | 5915 | M68K ARCHITECTURE |
| 5915 | M: Geert Uytterhoeven <geert@linux-m68k.org> | 5916 | M: Geert Uytterhoeven <geert@linux-m68k.org> |
| 5916 | L: linux-m68k@lists.linux-m68k.org | 5917 | L: linux-m68k@lists.linux-m68k.org |
| 5917 | W: http://www.linux-m68k.org/ | 5918 | W: http://www.linux-m68k.org/ |
| 5918 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git | 5919 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/geert/linux-m68k.git |
| 5919 | S: Maintained | 5920 | S: Maintained |
| 5920 | F: arch/m68k/ | 5921 | F: arch/m68k/ |
| 5921 | F: drivers/zorro/ | 5922 | F: drivers/zorro/ |
| 5922 | 5923 | ||
| 5923 | M68K ON APPLE MACINTOSH | 5924 | M68K ON APPLE MACINTOSH |
| 5924 | M: Joshua Thompson <funaho@jurai.org> | 5925 | M: Joshua Thompson <funaho@jurai.org> |
| 5925 | W: http://www.mac.linux-m68k.org/ | 5926 | W: http://www.mac.linux-m68k.org/ |
| 5926 | L: linux-m68k@lists.linux-m68k.org | 5927 | L: linux-m68k@lists.linux-m68k.org |
| 5927 | S: Maintained | 5928 | S: Maintained |
| 5928 | F: arch/m68k/mac/ | 5929 | F: arch/m68k/mac/ |
| 5929 | 5930 | ||
| 5930 | M68K ON HP9000/300 | 5931 | M68K ON HP9000/300 |
| 5931 | M: Philip Blundell <philb@gnu.org> | 5932 | M: Philip Blundell <philb@gnu.org> |
| 5932 | W: http://www.tazenda.demon.co.uk/phil/linux-hp | 5933 | W: http://www.tazenda.demon.co.uk/phil/linux-hp |
| 5933 | S: Maintained | 5934 | S: Maintained |
| 5934 | F: arch/m68k/hp300/ | 5935 | F: arch/m68k/hp300/ |
| 5935 | 5936 | ||
| 5936 | M88DS3103 MEDIA DRIVER | 5937 | M88DS3103 MEDIA DRIVER |
| 5937 | M: Antti Palosaari <crope@iki.fi> | 5938 | M: Antti Palosaari <crope@iki.fi> |
| 5938 | L: linux-media@vger.kernel.org | 5939 | L: linux-media@vger.kernel.org |
| 5939 | W: http://linuxtv.org/ | 5940 | W: http://linuxtv.org/ |
| 5940 | W: http://palosaari.fi/linux/ | 5941 | W: http://palosaari.fi/linux/ |
| 5941 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5942 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5942 | T: git git://linuxtv.org/anttip/media_tree.git | 5943 | T: git git://linuxtv.org/anttip/media_tree.git |
| 5943 | S: Maintained | 5944 | S: Maintained |
| 5944 | F: drivers/media/dvb-frontends/m88ds3103* | 5945 | F: drivers/media/dvb-frontends/m88ds3103* |
| 5945 | 5946 | ||
| 5946 | M88RS2000 MEDIA DRIVER | 5947 | M88RS2000 MEDIA DRIVER |
| 5947 | M: Malcolm Priestley <tvboxspy@gmail.com> | 5948 | M: Malcolm Priestley <tvboxspy@gmail.com> |
| 5948 | L: linux-media@vger.kernel.org | 5949 | L: linux-media@vger.kernel.org |
| 5949 | W: http://linuxtv.org/ | 5950 | W: http://linuxtv.org/ |
| 5950 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5951 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5951 | S: Maintained | 5952 | S: Maintained |
| 5952 | F: drivers/media/dvb-frontends/m88rs2000* | 5953 | F: drivers/media/dvb-frontends/m88rs2000* |
| 5953 | 5954 | ||
| 5954 | M88TS2022 MEDIA DRIVER | 5955 | M88TS2022 MEDIA DRIVER |
| 5955 | M: Antti Palosaari <crope@iki.fi> | 5956 | M: Antti Palosaari <crope@iki.fi> |
| 5956 | L: linux-media@vger.kernel.org | 5957 | L: linux-media@vger.kernel.org |
| 5957 | W: http://linuxtv.org/ | 5958 | W: http://linuxtv.org/ |
| 5958 | W: http://palosaari.fi/linux/ | 5959 | W: http://palosaari.fi/linux/ |
| 5959 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 5960 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 5960 | T: git git://linuxtv.org/anttip/media_tree.git | 5961 | T: git git://linuxtv.org/anttip/media_tree.git |
| 5961 | S: Maintained | 5962 | S: Maintained |
| 5962 | F: drivers/media/tuners/m88ts2022* | 5963 | F: drivers/media/tuners/m88ts2022* |
| 5963 | 5964 | ||
| 5964 | MA901 MASTERKIT USB FM RADIO DRIVER | 5965 | MA901 MASTERKIT USB FM RADIO DRIVER |
| 5965 | M: Alexey Klimov <klimov.linux@gmail.com> | 5966 | M: Alexey Klimov <klimov.linux@gmail.com> |
| 5966 | L: linux-media@vger.kernel.org | 5967 | L: linux-media@vger.kernel.org |
| 5967 | T: git git://linuxtv.org/media_tree.git | 5968 | T: git git://linuxtv.org/media_tree.git |
| 5968 | S: Maintained | 5969 | S: Maintained |
| 5969 | F: drivers/media/radio/radio-ma901.c | 5970 | F: drivers/media/radio/radio-ma901.c |
| 5970 | 5971 | ||
| 5971 | MAC80211 | 5972 | MAC80211 |
| 5972 | M: Johannes Berg <johannes@sipsolutions.net> | 5973 | M: Johannes Berg <johannes@sipsolutions.net> |
| 5973 | L: linux-wireless@vger.kernel.org | 5974 | L: linux-wireless@vger.kernel.org |
| 5974 | W: http://wireless.kernel.org/ | 5975 | W: http://wireless.kernel.org/ |
| 5975 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211.git | 5976 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211.git |
| 5976 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next.git | 5977 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next.git |
| 5977 | S: Maintained | 5978 | S: Maintained |
| 5978 | F: Documentation/networking/mac80211-injection.txt | 5979 | F: Documentation/networking/mac80211-injection.txt |
| 5979 | F: include/net/mac80211.h | 5980 | F: include/net/mac80211.h |
| 5980 | F: net/mac80211/ | 5981 | F: net/mac80211/ |
| 5981 | 5982 | ||
| 5982 | MACVLAN DRIVER | 5983 | MACVLAN DRIVER |
| 5983 | M: Patrick McHardy <kaber@trash.net> | 5984 | M: Patrick McHardy <kaber@trash.net> |
| 5984 | L: netdev@vger.kernel.org | 5985 | L: netdev@vger.kernel.org |
| 5985 | S: Maintained | 5986 | S: Maintained |
| 5986 | F: drivers/net/macvlan.c | 5987 | F: drivers/net/macvlan.c |
| 5987 | F: include/linux/if_macvlan.h | 5988 | F: include/linux/if_macvlan.h |
| 5988 | 5989 | ||
| 5989 | MAILBOX API | 5990 | MAILBOX API |
| 5990 | M: Jassi Brar <jassisinghbrar@gmail.com> | 5991 | M: Jassi Brar <jassisinghbrar@gmail.com> |
| 5991 | L: linux-kernel@vger.kernel.org | 5992 | L: linux-kernel@vger.kernel.org |
| 5992 | S: Maintained | 5993 | S: Maintained |
| 5993 | F: drivers/mailbox/ | 5994 | F: drivers/mailbox/ |
| 5994 | F: include/linux/mailbox_client.h | 5995 | F: include/linux/mailbox_client.h |
| 5995 | F: include/linux/mailbox_controller.h | 5996 | F: include/linux/mailbox_controller.h |
| 5996 | 5997 | ||
| 5997 | MAN-PAGES: MANUAL PAGES FOR LINUX -- Sections 2, 3, 4, 5, and 7 | 5998 | MAN-PAGES: MANUAL PAGES FOR LINUX -- Sections 2, 3, 4, 5, and 7 |
| 5998 | M: Michael Kerrisk <mtk.manpages@gmail.com> | 5999 | M: Michael Kerrisk <mtk.manpages@gmail.com> |
| 5999 | W: http://www.kernel.org/doc/man-pages | 6000 | W: http://www.kernel.org/doc/man-pages |
| 6000 | L: linux-man@vger.kernel.org | 6001 | L: linux-man@vger.kernel.org |
| 6001 | S: Maintained | 6002 | S: Maintained |
| 6002 | 6003 | ||
| 6003 | MARVELL ARMADA DRM SUPPORT | 6004 | MARVELL ARMADA DRM SUPPORT |
| 6004 | M: Russell King <rmk+kernel@arm.linux.org.uk> | 6005 | M: Russell King <rmk+kernel@arm.linux.org.uk> |
| 6005 | S: Maintained | 6006 | S: Maintained |
| 6006 | F: drivers/gpu/drm/armada/ | 6007 | F: drivers/gpu/drm/armada/ |
| 6007 | 6008 | ||
| 6008 | MARVELL 88E6352 DSA support | 6009 | MARVELL 88E6352 DSA support |
| 6009 | M: Guenter Roeck <linux@roeck-us.net> | 6010 | M: Guenter Roeck <linux@roeck-us.net> |
| 6010 | S: Maintained | 6011 | S: Maintained |
| 6011 | F: drivers/net/dsa/mv88e6352.c | 6012 | F: drivers/net/dsa/mv88e6352.c |
| 6012 | 6013 | ||
| 6013 | MARVELL GIGABIT ETHERNET DRIVERS (skge/sky2) | 6014 | MARVELL GIGABIT ETHERNET DRIVERS (skge/sky2) |
| 6014 | M: Mirko Lindner <mlindner@marvell.com> | 6015 | M: Mirko Lindner <mlindner@marvell.com> |
| 6015 | M: Stephen Hemminger <stephen@networkplumber.org> | 6016 | M: Stephen Hemminger <stephen@networkplumber.org> |
| 6016 | L: netdev@vger.kernel.org | 6017 | L: netdev@vger.kernel.org |
| 6017 | S: Maintained | 6018 | S: Maintained |
| 6018 | F: drivers/net/ethernet/marvell/sk* | 6019 | F: drivers/net/ethernet/marvell/sk* |
| 6019 | 6020 | ||
| 6020 | MARVELL LIBERTAS WIRELESS DRIVER | 6021 | MARVELL LIBERTAS WIRELESS DRIVER |
| 6021 | L: libertas-dev@lists.infradead.org | 6022 | L: libertas-dev@lists.infradead.org |
| 6022 | S: Orphan | 6023 | S: Orphan |
| 6023 | F: drivers/net/wireless/libertas/ | 6024 | F: drivers/net/wireless/libertas/ |
| 6024 | 6025 | ||
| 6025 | MARVELL MV643XX ETHERNET DRIVER | 6026 | MARVELL MV643XX ETHERNET DRIVER |
| 6026 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> | 6027 | M: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> |
| 6027 | L: netdev@vger.kernel.org | 6028 | L: netdev@vger.kernel.org |
| 6028 | S: Maintained | 6029 | S: Maintained |
| 6029 | F: drivers/net/ethernet/marvell/mv643xx_eth.* | 6030 | F: drivers/net/ethernet/marvell/mv643xx_eth.* |
| 6030 | F: include/linux/mv643xx.h | 6031 | F: include/linux/mv643xx.h |
| 6031 | 6032 | ||
| 6032 | MARVELL MVNETA ETHERNET DRIVER | 6033 | MARVELL MVNETA ETHERNET DRIVER |
| 6033 | M: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> | 6034 | M: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> |
| 6034 | L: netdev@vger.kernel.org | 6035 | L: netdev@vger.kernel.org |
| 6035 | S: Maintained | 6036 | S: Maintained |
| 6036 | F: drivers/net/ethernet/marvell/mvneta.* | 6037 | F: drivers/net/ethernet/marvell/mvneta.* |
| 6037 | 6038 | ||
| 6038 | MARVELL MWIFIEX WIRELESS DRIVER | 6039 | MARVELL MWIFIEX WIRELESS DRIVER |
| 6039 | M: Amitkumar Karwar <akarwar@marvell.com> | 6040 | M: Amitkumar Karwar <akarwar@marvell.com> |
| 6040 | M: Avinash Patil <patila@marvell.com> | 6041 | M: Avinash Patil <patila@marvell.com> |
| 6041 | L: linux-wireless@vger.kernel.org | 6042 | L: linux-wireless@vger.kernel.org |
| 6042 | S: Maintained | 6043 | S: Maintained |
| 6043 | F: drivers/net/wireless/mwifiex/ | 6044 | F: drivers/net/wireless/mwifiex/ |
| 6044 | 6045 | ||
| 6045 | MARVELL MWL8K WIRELESS DRIVER | 6046 | MARVELL MWL8K WIRELESS DRIVER |
| 6046 | M: Lennert Buytenhek <buytenh@wantstofly.org> | 6047 | M: Lennert Buytenhek <buytenh@wantstofly.org> |
| 6047 | L: linux-wireless@vger.kernel.org | 6048 | L: linux-wireless@vger.kernel.org |
| 6048 | S: Odd Fixes | 6049 | S: Odd Fixes |
| 6049 | F: drivers/net/wireless/mwl8k.c | 6050 | F: drivers/net/wireless/mwl8k.c |
| 6050 | 6051 | ||
| 6051 | MARVELL SOC MMC/SD/SDIO CONTROLLER DRIVER | 6052 | MARVELL SOC MMC/SD/SDIO CONTROLLER DRIVER |
| 6052 | M: Nicolas Pitre <nico@fluxnic.net> | 6053 | M: Nicolas Pitre <nico@fluxnic.net> |
| 6053 | S: Odd Fixes | 6054 | S: Odd Fixes |
| 6054 | F: drivers/mmc/host/mvsdio.* | 6055 | F: drivers/mmc/host/mvsdio.* |
| 6055 | 6056 | ||
| 6056 | MATROX FRAMEBUFFER DRIVER | 6057 | MATROX FRAMEBUFFER DRIVER |
| 6057 | L: linux-fbdev@vger.kernel.org | 6058 | L: linux-fbdev@vger.kernel.org |
| 6058 | S: Orphan | 6059 | S: Orphan |
| 6059 | F: drivers/video/fbdev/matrox/matroxfb_* | 6060 | F: drivers/video/fbdev/matrox/matroxfb_* |
| 6060 | F: include/uapi/linux/matroxfb.h | 6061 | F: include/uapi/linux/matroxfb.h |
| 6061 | 6062 | ||
| 6062 | MAX16065 HARDWARE MONITOR DRIVER | 6063 | MAX16065 HARDWARE MONITOR DRIVER |
| 6063 | M: Guenter Roeck <linux@roeck-us.net> | 6064 | M: Guenter Roeck <linux@roeck-us.net> |
| 6064 | L: lm-sensors@lm-sensors.org | 6065 | L: lm-sensors@lm-sensors.org |
| 6065 | S: Maintained | 6066 | S: Maintained |
| 6066 | F: Documentation/hwmon/max16065 | 6067 | F: Documentation/hwmon/max16065 |
| 6067 | F: drivers/hwmon/max16065.c | 6068 | F: drivers/hwmon/max16065.c |
| 6068 | 6069 | ||
| 6069 | MAX6650 HARDWARE MONITOR AND FAN CONTROLLER DRIVER | 6070 | MAX6650 HARDWARE MONITOR AND FAN CONTROLLER DRIVER |
| 6070 | M: "Hans J. Koch" <hjk@hansjkoch.de> | 6071 | M: "Hans J. Koch" <hjk@hansjkoch.de> |
| 6071 | L: lm-sensors@lm-sensors.org | 6072 | L: lm-sensors@lm-sensors.org |
| 6072 | S: Maintained | 6073 | S: Maintained |
| 6073 | F: Documentation/hwmon/max6650 | 6074 | F: Documentation/hwmon/max6650 |
| 6074 | F: drivers/hwmon/max6650.c | 6075 | F: drivers/hwmon/max6650.c |
| 6075 | 6076 | ||
| 6076 | MAX6697 HARDWARE MONITOR DRIVER | 6077 | MAX6697 HARDWARE MONITOR DRIVER |
| 6077 | M: Guenter Roeck <linux@roeck-us.net> | 6078 | M: Guenter Roeck <linux@roeck-us.net> |
| 6078 | L: lm-sensors@lm-sensors.org | 6079 | L: lm-sensors@lm-sensors.org |
| 6079 | S: Maintained | 6080 | S: Maintained |
| 6080 | F: Documentation/hwmon/max6697 | 6081 | F: Documentation/hwmon/max6697 |
| 6081 | F: Documentation/devicetree/bindings/i2c/max6697.txt | 6082 | F: Documentation/devicetree/bindings/i2c/max6697.txt |
| 6082 | F: drivers/hwmon/max6697.c | 6083 | F: drivers/hwmon/max6697.c |
| 6083 | F: include/linux/platform_data/max6697.h | 6084 | F: include/linux/platform_data/max6697.h |
| 6084 | 6085 | ||
| 6085 | MAXIRADIO FM RADIO RECEIVER DRIVER | 6086 | MAXIRADIO FM RADIO RECEIVER DRIVER |
| 6086 | M: Hans Verkuil <hverkuil@xs4all.nl> | 6087 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 6087 | L: linux-media@vger.kernel.org | 6088 | L: linux-media@vger.kernel.org |
| 6088 | T: git git://linuxtv.org/media_tree.git | 6089 | T: git git://linuxtv.org/media_tree.git |
| 6089 | W: http://linuxtv.org | 6090 | W: http://linuxtv.org |
| 6090 | S: Maintained | 6091 | S: Maintained |
| 6091 | F: drivers/media/radio/radio-maxiradio* | 6092 | F: drivers/media/radio/radio-maxiradio* |
| 6092 | 6093 | ||
| 6093 | MEDIA INPUT INFRASTRUCTURE (V4L/DVB) | 6094 | MEDIA INPUT INFRASTRUCTURE (V4L/DVB) |
| 6094 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 6095 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 6095 | P: LinuxTV.org Project | 6096 | P: LinuxTV.org Project |
| 6096 | L: linux-media@vger.kernel.org | 6097 | L: linux-media@vger.kernel.org |
| 6097 | W: http://linuxtv.org | 6098 | W: http://linuxtv.org |
| 6098 | Q: http://patchwork.kernel.org/project/linux-media/list/ | 6099 | Q: http://patchwork.kernel.org/project/linux-media/list/ |
| 6099 | T: git git://linuxtv.org/media_tree.git | 6100 | T: git git://linuxtv.org/media_tree.git |
| 6100 | S: Maintained | 6101 | S: Maintained |
| 6101 | F: Documentation/dvb/ | 6102 | F: Documentation/dvb/ |
| 6102 | F: Documentation/video4linux/ | 6103 | F: Documentation/video4linux/ |
| 6103 | F: Documentation/DocBook/media/ | 6104 | F: Documentation/DocBook/media/ |
| 6104 | F: drivers/media/ | 6105 | F: drivers/media/ |
| 6105 | F: drivers/staging/media/ | 6106 | F: drivers/staging/media/ |
| 6106 | F: include/media/ | 6107 | F: include/media/ |
| 6107 | F: include/uapi/linux/dvb/ | 6108 | F: include/uapi/linux/dvb/ |
| 6108 | F: include/uapi/linux/videodev2.h | 6109 | F: include/uapi/linux/videodev2.h |
| 6109 | F: include/uapi/linux/media.h | 6110 | F: include/uapi/linux/media.h |
| 6110 | F: include/uapi/linux/v4l2-* | 6111 | F: include/uapi/linux/v4l2-* |
| 6111 | F: include/uapi/linux/meye.h | 6112 | F: include/uapi/linux/meye.h |
| 6112 | F: include/uapi/linux/ivtv* | 6113 | F: include/uapi/linux/ivtv* |
| 6113 | F: include/uapi/linux/uvcvideo.h | 6114 | F: include/uapi/linux/uvcvideo.h |
| 6114 | 6115 | ||
| 6115 | MEDIAVISION PRO MOVIE STUDIO DRIVER | 6116 | MEDIAVISION PRO MOVIE STUDIO DRIVER |
| 6116 | M: Hans Verkuil <hverkuil@xs4all.nl> | 6117 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 6117 | L: linux-media@vger.kernel.org | 6118 | L: linux-media@vger.kernel.org |
| 6118 | T: git git://linuxtv.org/media_tree.git | 6119 | T: git git://linuxtv.org/media_tree.git |
| 6119 | W: http://linuxtv.org | 6120 | W: http://linuxtv.org |
| 6120 | S: Odd Fixes | 6121 | S: Odd Fixes |
| 6121 | F: drivers/media/parport/pms* | 6122 | F: drivers/media/parport/pms* |
| 6122 | 6123 | ||
| 6123 | MEGARAID SCSI/SAS DRIVERS | 6124 | MEGARAID SCSI/SAS DRIVERS |
| 6124 | M: Kashyap Desai <kashyap.desai@avagotech.com> | 6125 | M: Kashyap Desai <kashyap.desai@avagotech.com> |
| 6125 | M: Sumit Saxena <sumit.saxena@avagotech.com> | 6126 | M: Sumit Saxena <sumit.saxena@avagotech.com> |
| 6126 | M: Uday Lingala <uday.lingala@avagotech.com> | 6127 | M: Uday Lingala <uday.lingala@avagotech.com> |
| 6127 | L: megaraidlinux.pdl@avagotech.com | 6128 | L: megaraidlinux.pdl@avagotech.com |
| 6128 | L: linux-scsi@vger.kernel.org | 6129 | L: linux-scsi@vger.kernel.org |
| 6129 | W: http://www.lsi.com | 6130 | W: http://www.lsi.com |
| 6130 | S: Maintained | 6131 | S: Maintained |
| 6131 | F: Documentation/scsi/megaraid.txt | 6132 | F: Documentation/scsi/megaraid.txt |
| 6132 | F: drivers/scsi/megaraid.* | 6133 | F: drivers/scsi/megaraid.* |
| 6133 | F: drivers/scsi/megaraid/ | 6134 | F: drivers/scsi/megaraid/ |
| 6134 | 6135 | ||
| 6135 | MELLANOX ETHERNET DRIVER (mlx4_en) | 6136 | MELLANOX ETHERNET DRIVER (mlx4_en) |
| 6136 | M: Amir Vadai <amirv@mellanox.com> | 6137 | M: Amir Vadai <amirv@mellanox.com> |
| 6137 | L: netdev@vger.kernel.org | 6138 | L: netdev@vger.kernel.org |
| 6138 | S: Supported | 6139 | S: Supported |
| 6139 | W: http://www.mellanox.com | 6140 | W: http://www.mellanox.com |
| 6140 | Q: http://patchwork.ozlabs.org/project/netdev/list/ | 6141 | Q: http://patchwork.ozlabs.org/project/netdev/list/ |
| 6141 | F: drivers/net/ethernet/mellanox/mlx4/en_* | 6142 | F: drivers/net/ethernet/mellanox/mlx4/en_* |
| 6142 | 6143 | ||
| 6143 | MEMORY MANAGEMENT | 6144 | MEMORY MANAGEMENT |
| 6144 | L: linux-mm@kvack.org | 6145 | L: linux-mm@kvack.org |
| 6145 | W: http://www.linux-mm.org | 6146 | W: http://www.linux-mm.org |
| 6146 | S: Maintained | 6147 | S: Maintained |
| 6147 | F: include/linux/mm.h | 6148 | F: include/linux/mm.h |
| 6148 | F: include/linux/gfp.h | 6149 | F: include/linux/gfp.h |
| 6149 | F: include/linux/mmzone.h | 6150 | F: include/linux/mmzone.h |
| 6150 | F: include/linux/memory_hotplug.h | 6151 | F: include/linux/memory_hotplug.h |
| 6151 | F: include/linux/vmalloc.h | 6152 | F: include/linux/vmalloc.h |
| 6152 | F: mm/ | 6153 | F: mm/ |
| 6153 | 6154 | ||
| 6154 | MEMORY TECHNOLOGY DEVICES (MTD) | 6155 | MEMORY TECHNOLOGY DEVICES (MTD) |
| 6155 | M: David Woodhouse <dwmw2@infradead.org> | 6156 | M: David Woodhouse <dwmw2@infradead.org> |
| 6156 | M: Brian Norris <computersforpeace@gmail.com> | 6157 | M: Brian Norris <computersforpeace@gmail.com> |
| 6157 | L: linux-mtd@lists.infradead.org | 6158 | L: linux-mtd@lists.infradead.org |
| 6158 | W: http://www.linux-mtd.infradead.org/ | 6159 | W: http://www.linux-mtd.infradead.org/ |
| 6159 | Q: http://patchwork.ozlabs.org/project/linux-mtd/list/ | 6160 | Q: http://patchwork.ozlabs.org/project/linux-mtd/list/ |
| 6160 | T: git git://git.infradead.org/linux-mtd.git | 6161 | T: git git://git.infradead.org/linux-mtd.git |
| 6161 | T: git git://git.infradead.org/l2-mtd.git | 6162 | T: git git://git.infradead.org/l2-mtd.git |
| 6162 | S: Maintained | 6163 | S: Maintained |
| 6163 | F: drivers/mtd/ | 6164 | F: drivers/mtd/ |
| 6164 | F: include/linux/mtd/ | 6165 | F: include/linux/mtd/ |
| 6165 | F: include/uapi/mtd/ | 6166 | F: include/uapi/mtd/ |
| 6166 | 6167 | ||
| 6167 | MEN A21 WATCHDOG DRIVER | 6168 | MEN A21 WATCHDOG DRIVER |
| 6168 | M: Johannes Thumshirn <johannes.thumshirn@men.de> | 6169 | M: Johannes Thumshirn <johannes.thumshirn@men.de> |
| 6169 | L: linux-watchdog@vger.kernel.org | 6170 | L: linux-watchdog@vger.kernel.org |
| 6170 | S: Supported | 6171 | S: Supported |
| 6171 | F: drivers/watchdog/mena21_wdt.c | 6172 | F: drivers/watchdog/mena21_wdt.c |
| 6172 | 6173 | ||
| 6173 | MEN CHAMELEON BUS (mcb) | 6174 | MEN CHAMELEON BUS (mcb) |
| 6174 | M: Johannes Thumshirn <johannes.thumshirn@men.de> | 6175 | M: Johannes Thumshirn <johannes.thumshirn@men.de> |
| 6175 | S: Supported | 6176 | S: Supported |
| 6176 | F: drivers/mcb/ | 6177 | F: drivers/mcb/ |
| 6177 | F: include/linux/mcb.h | 6178 | F: include/linux/mcb.h |
| 6178 | 6179 | ||
| 6179 | MEN F21BMC (Board Management Controller) | 6180 | MEN F21BMC (Board Management Controller) |
| 6180 | M: Andreas Werner <andreas.werner@men.de> | 6181 | M: Andreas Werner <andreas.werner@men.de> |
| 6181 | S: Supported | 6182 | S: Supported |
| 6182 | F: drivers/mfd/menf21bmc.c | 6183 | F: drivers/mfd/menf21bmc.c |
| 6183 | F: drivers/watchdog/menf21bmc_wdt.c | 6184 | F: drivers/watchdog/menf21bmc_wdt.c |
| 6184 | F: drivers/leds/leds-menf21bmc.c | 6185 | F: drivers/leds/leds-menf21bmc.c |
| 6185 | F: drivers/hwmon/menf21bmc_hwmon.c | 6186 | F: drivers/hwmon/menf21bmc_hwmon.c |
| 6186 | F: Documentation/hwmon/menf21bmc | 6187 | F: Documentation/hwmon/menf21bmc |
| 6187 | 6188 | ||
| 6188 | METAG ARCHITECTURE | 6189 | METAG ARCHITECTURE |
| 6189 | M: James Hogan <james.hogan@imgtec.com> | 6190 | M: James Hogan <james.hogan@imgtec.com> |
| 6190 | L: linux-metag@vger.kernel.org | 6191 | L: linux-metag@vger.kernel.org |
| 6191 | S: Supported | 6192 | S: Supported |
| 6192 | F: arch/metag/ | 6193 | F: arch/metag/ |
| 6193 | F: Documentation/metag/ | 6194 | F: Documentation/metag/ |
| 6194 | F: Documentation/devicetree/bindings/metag/ | 6195 | F: Documentation/devicetree/bindings/metag/ |
| 6195 | F: drivers/clocksource/metag_generic.c | 6196 | F: drivers/clocksource/metag_generic.c |
| 6196 | F: drivers/irqchip/irq-metag.c | 6197 | F: drivers/irqchip/irq-metag.c |
| 6197 | F: drivers/irqchip/irq-metag-ext.c | 6198 | F: drivers/irqchip/irq-metag-ext.c |
| 6198 | F: drivers/tty/metag_da.c | 6199 | F: drivers/tty/metag_da.c |
| 6199 | 6200 | ||
| 6200 | MICROBLAZE ARCHITECTURE | 6201 | MICROBLAZE ARCHITECTURE |
| 6201 | M: Michal Simek <monstr@monstr.eu> | 6202 | M: Michal Simek <monstr@monstr.eu> |
| 6202 | W: http://www.monstr.eu/fdt/ | 6203 | W: http://www.monstr.eu/fdt/ |
| 6203 | T: git git://git.monstr.eu/linux-2.6-microblaze.git | 6204 | T: git git://git.monstr.eu/linux-2.6-microblaze.git |
| 6204 | S: Supported | 6205 | S: Supported |
| 6205 | F: arch/microblaze/ | 6206 | F: arch/microblaze/ |
| 6206 | 6207 | ||
| 6207 | MICROTEK X6 SCANNER | 6208 | MICROTEK X6 SCANNER |
| 6208 | M: Oliver Neukum <oliver@neukum.org> | 6209 | M: Oliver Neukum <oliver@neukum.org> |
| 6209 | S: Maintained | 6210 | S: Maintained |
| 6210 | F: drivers/usb/image/microtek.* | 6211 | F: drivers/usb/image/microtek.* |
| 6211 | 6212 | ||
| 6212 | MIPS | 6213 | MIPS |
| 6213 | M: Ralf Baechle <ralf@linux-mips.org> | 6214 | M: Ralf Baechle <ralf@linux-mips.org> |
| 6214 | L: linux-mips@linux-mips.org | 6215 | L: linux-mips@linux-mips.org |
| 6215 | W: http://www.linux-mips.org/ | 6216 | W: http://www.linux-mips.org/ |
| 6216 | T: git git://git.linux-mips.org/pub/scm/ralf/linux.git | 6217 | T: git git://git.linux-mips.org/pub/scm/ralf/linux.git |
| 6217 | Q: http://patchwork.linux-mips.org/project/linux-mips/list/ | 6218 | Q: http://patchwork.linux-mips.org/project/linux-mips/list/ |
| 6218 | S: Supported | 6219 | S: Supported |
| 6219 | F: Documentation/mips/ | 6220 | F: Documentation/mips/ |
| 6220 | F: arch/mips/ | 6221 | F: arch/mips/ |
| 6221 | 6222 | ||
| 6222 | MIROSOUND PCM20 FM RADIO RECEIVER DRIVER | 6223 | MIROSOUND PCM20 FM RADIO RECEIVER DRIVER |
| 6223 | M: Hans Verkuil <hverkuil@xs4all.nl> | 6224 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 6224 | L: linux-media@vger.kernel.org | 6225 | L: linux-media@vger.kernel.org |
| 6225 | T: git git://linuxtv.org/media_tree.git | 6226 | T: git git://linuxtv.org/media_tree.git |
| 6226 | W: http://linuxtv.org | 6227 | W: http://linuxtv.org |
| 6227 | S: Odd Fixes | 6228 | S: Odd Fixes |
| 6228 | F: drivers/media/radio/radio-miropcm20* | 6229 | F: drivers/media/radio/radio-miropcm20* |
| 6229 | 6230 | ||
| 6230 | Mellanox MLX5 core VPI driver | 6231 | Mellanox MLX5 core VPI driver |
| 6231 | M: Eli Cohen <eli@mellanox.com> | 6232 | M: Eli Cohen <eli@mellanox.com> |
| 6232 | L: netdev@vger.kernel.org | 6233 | L: netdev@vger.kernel.org |
| 6233 | L: linux-rdma@vger.kernel.org | 6234 | L: linux-rdma@vger.kernel.org |
| 6234 | W: http://www.mellanox.com | 6235 | W: http://www.mellanox.com |
| 6235 | Q: http://patchwork.ozlabs.org/project/netdev/list/ | 6236 | Q: http://patchwork.ozlabs.org/project/netdev/list/ |
| 6236 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ | 6237 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ |
| 6237 | T: git git://openfabrics.org/~eli/connect-ib.git | 6238 | T: git git://openfabrics.org/~eli/connect-ib.git |
| 6238 | S: Supported | 6239 | S: Supported |
| 6239 | F: drivers/net/ethernet/mellanox/mlx5/core/ | 6240 | F: drivers/net/ethernet/mellanox/mlx5/core/ |
| 6240 | F: include/linux/mlx5/ | 6241 | F: include/linux/mlx5/ |
| 6241 | 6242 | ||
| 6242 | Mellanox MLX5 IB driver | 6243 | Mellanox MLX5 IB driver |
| 6243 | M: Eli Cohen <eli@mellanox.com> | 6244 | M: Eli Cohen <eli@mellanox.com> |
| 6244 | L: linux-rdma@vger.kernel.org | 6245 | L: linux-rdma@vger.kernel.org |
| 6245 | W: http://www.mellanox.com | 6246 | W: http://www.mellanox.com |
| 6246 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ | 6247 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ |
| 6247 | T: git git://openfabrics.org/~eli/connect-ib.git | 6248 | T: git git://openfabrics.org/~eli/connect-ib.git |
| 6248 | S: Supported | 6249 | S: Supported |
| 6249 | F: include/linux/mlx5/ | 6250 | F: include/linux/mlx5/ |
| 6250 | F: drivers/infiniband/hw/mlx5/ | 6251 | F: drivers/infiniband/hw/mlx5/ |
| 6251 | 6252 | ||
| 6252 | MN88472 MEDIA DRIVER | 6253 | MN88472 MEDIA DRIVER |
| 6253 | M: Antti Palosaari <crope@iki.fi> | 6254 | M: Antti Palosaari <crope@iki.fi> |
| 6254 | L: linux-media@vger.kernel.org | 6255 | L: linux-media@vger.kernel.org |
| 6255 | W: http://linuxtv.org/ | 6256 | W: http://linuxtv.org/ |
| 6256 | W: http://palosaari.fi/linux/ | 6257 | W: http://palosaari.fi/linux/ |
| 6257 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 6258 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 6258 | T: git git://linuxtv.org/anttip/media_tree.git | 6259 | T: git git://linuxtv.org/anttip/media_tree.git |
| 6259 | S: Maintained | 6260 | S: Maintained |
| 6260 | F: drivers/staging/media/mn88472/ | 6261 | F: drivers/staging/media/mn88472/ |
| 6261 | F: drivers/media/dvb-frontends/mn88472.h | 6262 | F: drivers/media/dvb-frontends/mn88472.h |
| 6262 | 6263 | ||
| 6263 | MN88473 MEDIA DRIVER | 6264 | MN88473 MEDIA DRIVER |
| 6264 | M: Antti Palosaari <crope@iki.fi> | 6265 | M: Antti Palosaari <crope@iki.fi> |
| 6265 | L: linux-media@vger.kernel.org | 6266 | L: linux-media@vger.kernel.org |
| 6266 | W: http://linuxtv.org/ | 6267 | W: http://linuxtv.org/ |
| 6267 | W: http://palosaari.fi/linux/ | 6268 | W: http://palosaari.fi/linux/ |
| 6268 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 6269 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 6269 | T: git git://linuxtv.org/anttip/media_tree.git | 6270 | T: git git://linuxtv.org/anttip/media_tree.git |
| 6270 | S: Maintained | 6271 | S: Maintained |
| 6271 | F: drivers/staging/media/mn88473/ | 6272 | F: drivers/staging/media/mn88473/ |
| 6272 | F: drivers/media/dvb-frontends/mn88473.h | 6273 | F: drivers/media/dvb-frontends/mn88473.h |
| 6273 | 6274 | ||
| 6274 | MODULE SUPPORT | 6275 | MODULE SUPPORT |
| 6275 | M: Rusty Russell <rusty@rustcorp.com.au> | 6276 | M: Rusty Russell <rusty@rustcorp.com.au> |
| 6276 | S: Maintained | 6277 | S: Maintained |
| 6277 | F: include/linux/module.h | 6278 | F: include/linux/module.h |
| 6278 | F: kernel/module.c | 6279 | F: kernel/module.c |
| 6279 | 6280 | ||
| 6280 | MOTION EYE VAIO PICTUREBOOK CAMERA DRIVER | 6281 | MOTION EYE VAIO PICTUREBOOK CAMERA DRIVER |
| 6281 | W: http://popies.net/meye/ | 6282 | W: http://popies.net/meye/ |
| 6282 | S: Orphan | 6283 | S: Orphan |
| 6283 | F: Documentation/video4linux/meye.txt | 6284 | F: Documentation/video4linux/meye.txt |
| 6284 | F: drivers/media/pci/meye/ | 6285 | F: drivers/media/pci/meye/ |
| 6285 | F: include/uapi/linux/meye.h | 6286 | F: include/uapi/linux/meye.h |
| 6286 | 6287 | ||
| 6287 | MOXA SMARTIO/INDUSTIO/INTELLIO SERIAL CARD | 6288 | MOXA SMARTIO/INDUSTIO/INTELLIO SERIAL CARD |
| 6288 | M: Jiri Slaby <jirislaby@gmail.com> | 6289 | M: Jiri Slaby <jirislaby@gmail.com> |
| 6289 | S: Maintained | 6290 | S: Maintained |
| 6290 | F: Documentation/serial/moxa-smartio | 6291 | F: Documentation/serial/moxa-smartio |
| 6291 | F: drivers/tty/mxser.* | 6292 | F: drivers/tty/mxser.* |
| 6292 | 6293 | ||
| 6293 | MR800 AVERMEDIA USB FM RADIO DRIVER | 6294 | MR800 AVERMEDIA USB FM RADIO DRIVER |
| 6294 | M: Alexey Klimov <klimov.linux@gmail.com> | 6295 | M: Alexey Klimov <klimov.linux@gmail.com> |
| 6295 | L: linux-media@vger.kernel.org | 6296 | L: linux-media@vger.kernel.org |
| 6296 | T: git git://linuxtv.org/media_tree.git | 6297 | T: git git://linuxtv.org/media_tree.git |
| 6297 | S: Maintained | 6298 | S: Maintained |
| 6298 | F: drivers/media/radio/radio-mr800.c | 6299 | F: drivers/media/radio/radio-mr800.c |
| 6299 | 6300 | ||
| 6300 | MRF24J40 IEEE 802.15.4 RADIO DRIVER | 6301 | MRF24J40 IEEE 802.15.4 RADIO DRIVER |
| 6301 | M: Alan Ott <alan@signal11.us> | 6302 | M: Alan Ott <alan@signal11.us> |
| 6302 | L: linux-wpan@vger.kernel.org | 6303 | L: linux-wpan@vger.kernel.org |
| 6303 | S: Maintained | 6304 | S: Maintained |
| 6304 | F: drivers/net/ieee802154/mrf24j40.c | 6305 | F: drivers/net/ieee802154/mrf24j40.c |
| 6305 | 6306 | ||
| 6306 | MSI LAPTOP SUPPORT | 6307 | MSI LAPTOP SUPPORT |
| 6307 | M: "Lee, Chun-Yi" <jlee@suse.com> | 6308 | M: "Lee, Chun-Yi" <jlee@suse.com> |
| 6308 | L: platform-driver-x86@vger.kernel.org | 6309 | L: platform-driver-x86@vger.kernel.org |
| 6309 | S: Maintained | 6310 | S: Maintained |
| 6310 | F: drivers/platform/x86/msi-laptop.c | 6311 | F: drivers/platform/x86/msi-laptop.c |
| 6311 | 6312 | ||
| 6312 | MSI WMI SUPPORT | 6313 | MSI WMI SUPPORT |
| 6313 | M: Anisse Astier <anisse@astier.eu> | 6314 | M: Anisse Astier <anisse@astier.eu> |
| 6314 | L: platform-driver-x86@vger.kernel.org | 6315 | L: platform-driver-x86@vger.kernel.org |
| 6315 | S: Supported | 6316 | S: Supported |
| 6316 | F: drivers/platform/x86/msi-wmi.c | 6317 | F: drivers/platform/x86/msi-wmi.c |
| 6317 | 6318 | ||
| 6318 | MSI001 MEDIA DRIVER | 6319 | MSI001 MEDIA DRIVER |
| 6319 | M: Antti Palosaari <crope@iki.fi> | 6320 | M: Antti Palosaari <crope@iki.fi> |
| 6320 | L: linux-media@vger.kernel.org | 6321 | L: linux-media@vger.kernel.org |
| 6321 | W: http://linuxtv.org/ | 6322 | W: http://linuxtv.org/ |
| 6322 | W: http://palosaari.fi/linux/ | 6323 | W: http://palosaari.fi/linux/ |
| 6323 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 6324 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 6324 | T: git git://linuxtv.org/anttip/media_tree.git | 6325 | T: git git://linuxtv.org/anttip/media_tree.git |
| 6325 | S: Maintained | 6326 | S: Maintained |
| 6326 | F: drivers/media/tuners/msi001* | 6327 | F: drivers/media/tuners/msi001* |
| 6327 | 6328 | ||
| 6328 | MSI2500 MEDIA DRIVER | 6329 | MSI2500 MEDIA DRIVER |
| 6329 | M: Antti Palosaari <crope@iki.fi> | 6330 | M: Antti Palosaari <crope@iki.fi> |
| 6330 | L: linux-media@vger.kernel.org | 6331 | L: linux-media@vger.kernel.org |
| 6331 | W: http://linuxtv.org/ | 6332 | W: http://linuxtv.org/ |
| 6332 | W: http://palosaari.fi/linux/ | 6333 | W: http://palosaari.fi/linux/ |
| 6333 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 6334 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 6334 | T: git git://linuxtv.org/anttip/media_tree.git | 6335 | T: git git://linuxtv.org/anttip/media_tree.git |
| 6335 | S: Maintained | 6336 | S: Maintained |
| 6336 | F: drivers/media/usb/msi2500/ | 6337 | F: drivers/media/usb/msi2500/ |
| 6337 | 6338 | ||
| 6338 | MT9M032 APTINA SENSOR DRIVER | 6339 | MT9M032 APTINA SENSOR DRIVER |
| 6339 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 6340 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 6340 | L: linux-media@vger.kernel.org | 6341 | L: linux-media@vger.kernel.org |
| 6341 | T: git git://linuxtv.org/media_tree.git | 6342 | T: git git://linuxtv.org/media_tree.git |
| 6342 | S: Maintained | 6343 | S: Maintained |
| 6343 | F: drivers/media/i2c/mt9m032.c | 6344 | F: drivers/media/i2c/mt9m032.c |
| 6344 | F: include/media/mt9m032.h | 6345 | F: include/media/mt9m032.h |
| 6345 | 6346 | ||
| 6346 | MT9P031 APTINA CAMERA SENSOR | 6347 | MT9P031 APTINA CAMERA SENSOR |
| 6347 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 6348 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 6348 | L: linux-media@vger.kernel.org | 6349 | L: linux-media@vger.kernel.org |
| 6349 | T: git git://linuxtv.org/media_tree.git | 6350 | T: git git://linuxtv.org/media_tree.git |
| 6350 | S: Maintained | 6351 | S: Maintained |
| 6351 | F: drivers/media/i2c/mt9p031.c | 6352 | F: drivers/media/i2c/mt9p031.c |
| 6352 | F: include/media/mt9p031.h | 6353 | F: include/media/mt9p031.h |
| 6353 | 6354 | ||
| 6354 | MT9T001 APTINA CAMERA SENSOR | 6355 | MT9T001 APTINA CAMERA SENSOR |
| 6355 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 6356 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 6356 | L: linux-media@vger.kernel.org | 6357 | L: linux-media@vger.kernel.org |
| 6357 | T: git git://linuxtv.org/media_tree.git | 6358 | T: git git://linuxtv.org/media_tree.git |
| 6358 | S: Maintained | 6359 | S: Maintained |
| 6359 | F: drivers/media/i2c/mt9t001.c | 6360 | F: drivers/media/i2c/mt9t001.c |
| 6360 | F: include/media/mt9t001.h | 6361 | F: include/media/mt9t001.h |
| 6361 | 6362 | ||
| 6362 | MT9V032 APTINA CAMERA SENSOR | 6363 | MT9V032 APTINA CAMERA SENSOR |
| 6363 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 6364 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 6364 | L: linux-media@vger.kernel.org | 6365 | L: linux-media@vger.kernel.org |
| 6365 | T: git git://linuxtv.org/media_tree.git | 6366 | T: git git://linuxtv.org/media_tree.git |
| 6366 | S: Maintained | 6367 | S: Maintained |
| 6367 | F: drivers/media/i2c/mt9v032.c | 6368 | F: drivers/media/i2c/mt9v032.c |
| 6368 | F: include/media/mt9v032.h | 6369 | F: include/media/mt9v032.h |
| 6369 | 6370 | ||
| 6370 | MULTIFUNCTION DEVICES (MFD) | 6371 | MULTIFUNCTION DEVICES (MFD) |
| 6371 | M: Samuel Ortiz <sameo@linux.intel.com> | 6372 | M: Samuel Ortiz <sameo@linux.intel.com> |
| 6372 | M: Lee Jones <lee.jones@linaro.org> | 6373 | M: Lee Jones <lee.jones@linaro.org> |
| 6373 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git | 6374 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd.git |
| 6374 | S: Supported | 6375 | S: Supported |
| 6375 | F: drivers/mfd/ | 6376 | F: drivers/mfd/ |
| 6376 | F: include/linux/mfd/ | 6377 | F: include/linux/mfd/ |
| 6377 | 6378 | ||
| 6378 | MULTIMEDIA CARD (MMC), SECURE DIGITAL (SD) AND SDIO SUBSYSTEM | 6379 | MULTIMEDIA CARD (MMC), SECURE DIGITAL (SD) AND SDIO SUBSYSTEM |
| 6379 | M: Chris Ball <chris@printf.net> | 6380 | M: Chris Ball <chris@printf.net> |
| 6380 | M: Ulf Hansson <ulf.hansson@linaro.org> | 6381 | M: Ulf Hansson <ulf.hansson@linaro.org> |
| 6381 | L: linux-mmc@vger.kernel.org | 6382 | L: linux-mmc@vger.kernel.org |
| 6382 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc.git | 6383 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc.git |
| 6383 | T: git git://git.linaro.org/people/ulf.hansson/mmc.git | 6384 | T: git git://git.linaro.org/people/ulf.hansson/mmc.git |
| 6384 | S: Maintained | 6385 | S: Maintained |
| 6385 | F: drivers/mmc/ | 6386 | F: drivers/mmc/ |
| 6386 | F: include/linux/mmc/ | 6387 | F: include/linux/mmc/ |
| 6387 | F: include/uapi/linux/mmc/ | 6388 | F: include/uapi/linux/mmc/ |
| 6388 | 6389 | ||
| 6389 | MULTIMEDIA CARD (MMC) ETC. OVER SPI | 6390 | MULTIMEDIA CARD (MMC) ETC. OVER SPI |
| 6390 | S: Orphan | 6391 | S: Orphan |
| 6391 | F: drivers/mmc/host/mmc_spi.c | 6392 | F: drivers/mmc/host/mmc_spi.c |
| 6392 | F: include/linux/spi/mmc_spi.h | 6393 | F: include/linux/spi/mmc_spi.h |
| 6393 | 6394 | ||
| 6394 | MULTISOUND SOUND DRIVER | 6395 | MULTISOUND SOUND DRIVER |
| 6395 | M: Andrew Veliath <andrewtv@usa.net> | 6396 | M: Andrew Veliath <andrewtv@usa.net> |
| 6396 | S: Maintained | 6397 | S: Maintained |
| 6397 | F: Documentation/sound/oss/MultiSound | 6398 | F: Documentation/sound/oss/MultiSound |
| 6398 | F: sound/oss/msnd* | 6399 | F: sound/oss/msnd* |
| 6399 | 6400 | ||
| 6400 | MULTITECH MULTIPORT CARD (ISICOM) | 6401 | MULTITECH MULTIPORT CARD (ISICOM) |
| 6401 | S: Orphan | 6402 | S: Orphan |
| 6402 | F: drivers/tty/isicom.c | 6403 | F: drivers/tty/isicom.c |
| 6403 | F: include/linux/isicom.h | 6404 | F: include/linux/isicom.h |
| 6404 | 6405 | ||
| 6405 | MUSB MULTIPOINT HIGH SPEED DUAL-ROLE CONTROLLER | 6406 | MUSB MULTIPOINT HIGH SPEED DUAL-ROLE CONTROLLER |
| 6406 | M: Felipe Balbi <balbi@ti.com> | 6407 | M: Felipe Balbi <balbi@ti.com> |
| 6407 | L: linux-usb@vger.kernel.org | 6408 | L: linux-usb@vger.kernel.org |
| 6408 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git | 6409 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git |
| 6409 | S: Maintained | 6410 | S: Maintained |
| 6410 | F: drivers/usb/musb/ | 6411 | F: drivers/usb/musb/ |
| 6411 | 6412 | ||
| 6412 | MXL5007T MEDIA DRIVER | 6413 | MXL5007T MEDIA DRIVER |
| 6413 | M: Michael Krufky <mkrufky@linuxtv.org> | 6414 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 6414 | L: linux-media@vger.kernel.org | 6415 | L: linux-media@vger.kernel.org |
| 6415 | W: http://linuxtv.org/ | 6416 | W: http://linuxtv.org/ |
| 6416 | W: http://github.com/mkrufky | 6417 | W: http://github.com/mkrufky |
| 6417 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 6418 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 6418 | T: git git://linuxtv.org/mkrufky/tuners.git | 6419 | T: git git://linuxtv.org/mkrufky/tuners.git |
| 6419 | S: Maintained | 6420 | S: Maintained |
| 6420 | F: drivers/media/tuners/mxl5007t.* | 6421 | F: drivers/media/tuners/mxl5007t.* |
| 6421 | 6422 | ||
| 6422 | MYRICOM MYRI-10G 10GbE DRIVER (MYRI10GE) | 6423 | MYRICOM MYRI-10G 10GbE DRIVER (MYRI10GE) |
| 6423 | M: Hyong-Youb Kim <hykim@myri.com> | 6424 | M: Hyong-Youb Kim <hykim@myri.com> |
| 6424 | L: netdev@vger.kernel.org | 6425 | L: netdev@vger.kernel.org |
| 6425 | W: https://www.myricom.com/support/downloads/myri10ge.html | 6426 | W: https://www.myricom.com/support/downloads/myri10ge.html |
| 6426 | S: Supported | 6427 | S: Supported |
| 6427 | F: drivers/net/ethernet/myricom/myri10ge/ | 6428 | F: drivers/net/ethernet/myricom/myri10ge/ |
| 6428 | 6429 | ||
| 6429 | NATSEMI ETHERNET DRIVER (DP8381x) | 6430 | NATSEMI ETHERNET DRIVER (DP8381x) |
| 6430 | S: Orphan | 6431 | S: Orphan |
| 6431 | F: drivers/net/ethernet/natsemi/natsemi.c | 6432 | F: drivers/net/ethernet/natsemi/natsemi.c |
| 6432 | 6433 | ||
| 6433 | NATIVE INSTRUMENTS USB SOUND INTERFACE DRIVER | 6434 | NATIVE INSTRUMENTS USB SOUND INTERFACE DRIVER |
| 6434 | M: Daniel Mack <zonque@gmail.com> | 6435 | M: Daniel Mack <zonque@gmail.com> |
| 6435 | S: Maintained | 6436 | S: Maintained |
| 6436 | L: alsa-devel@alsa-project.org | 6437 | L: alsa-devel@alsa-project.org |
| 6437 | W: http://www.native-instruments.com | 6438 | W: http://www.native-instruments.com |
| 6438 | F: sound/usb/caiaq/ | 6439 | F: sound/usb/caiaq/ |
| 6439 | 6440 | ||
| 6440 | NCP FILESYSTEM | 6441 | NCP FILESYSTEM |
| 6441 | M: Petr Vandrovec <petr@vandrovec.name> | 6442 | M: Petr Vandrovec <petr@vandrovec.name> |
| 6442 | S: Odd Fixes | 6443 | S: Odd Fixes |
| 6443 | F: fs/ncpfs/ | 6444 | F: fs/ncpfs/ |
| 6444 | 6445 | ||
| 6445 | NCR 5380 SCSI DRIVERS | 6446 | NCR 5380 SCSI DRIVERS |
| 6446 | M: Finn Thain <fthain@telegraphics.com.au> | 6447 | M: Finn Thain <fthain@telegraphics.com.au> |
| 6447 | M: Michael Schmitz <schmitzmic@gmail.com> | 6448 | M: Michael Schmitz <schmitzmic@gmail.com> |
| 6448 | L: linux-scsi@vger.kernel.org | 6449 | L: linux-scsi@vger.kernel.org |
| 6449 | S: Maintained | 6450 | S: Maintained |
| 6450 | F: Documentation/scsi/g_NCR5380.txt | 6451 | F: Documentation/scsi/g_NCR5380.txt |
| 6451 | F: drivers/scsi/NCR5380.* | 6452 | F: drivers/scsi/NCR5380.* |
| 6452 | F: drivers/scsi/arm/cumana_1.c | 6453 | F: drivers/scsi/arm/cumana_1.c |
| 6453 | F: drivers/scsi/arm/oak.c | 6454 | F: drivers/scsi/arm/oak.c |
| 6454 | F: drivers/scsi/atari_NCR5380.c | 6455 | F: drivers/scsi/atari_NCR5380.c |
| 6455 | F: drivers/scsi/atari_scsi.* | 6456 | F: drivers/scsi/atari_scsi.* |
| 6456 | F: drivers/scsi/dmx3191d.c | 6457 | F: drivers/scsi/dmx3191d.c |
| 6457 | F: drivers/scsi/dtc.* | 6458 | F: drivers/scsi/dtc.* |
| 6458 | F: drivers/scsi/g_NCR5380.* | 6459 | F: drivers/scsi/g_NCR5380.* |
| 6459 | F: drivers/scsi/g_NCR5380_mmio.c | 6460 | F: drivers/scsi/g_NCR5380_mmio.c |
| 6460 | F: drivers/scsi/mac_scsi.* | 6461 | F: drivers/scsi/mac_scsi.* |
| 6461 | F: drivers/scsi/pas16.* | 6462 | F: drivers/scsi/pas16.* |
| 6462 | F: drivers/scsi/sun3_scsi.* | 6463 | F: drivers/scsi/sun3_scsi.* |
| 6463 | F: drivers/scsi/sun3_scsi_vme.c | 6464 | F: drivers/scsi/sun3_scsi_vme.c |
| 6464 | F: drivers/scsi/t128.* | 6465 | F: drivers/scsi/t128.* |
| 6465 | 6466 | ||
| 6466 | NCR DUAL 700 SCSI DRIVER (MICROCHANNEL) | 6467 | NCR DUAL 700 SCSI DRIVER (MICROCHANNEL) |
| 6467 | M: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> | 6468 | M: "James E.J. Bottomley" <James.Bottomley@HansenPartnership.com> |
| 6468 | L: linux-scsi@vger.kernel.org | 6469 | L: linux-scsi@vger.kernel.org |
| 6469 | S: Maintained | 6470 | S: Maintained |
| 6470 | F: drivers/scsi/NCR_D700.* | 6471 | F: drivers/scsi/NCR_D700.* |
| 6471 | 6472 | ||
| 6472 | NCT6775 HARDWARE MONITOR DRIVER | 6473 | NCT6775 HARDWARE MONITOR DRIVER |
| 6473 | M: Guenter Roeck <linux@roeck-us.net> | 6474 | M: Guenter Roeck <linux@roeck-us.net> |
| 6474 | L: lm-sensors@lm-sensors.org | 6475 | L: lm-sensors@lm-sensors.org |
| 6475 | S: Maintained | 6476 | S: Maintained |
| 6476 | F: Documentation/hwmon/nct6775 | 6477 | F: Documentation/hwmon/nct6775 |
| 6477 | F: drivers/hwmon/nct6775.c | 6478 | F: drivers/hwmon/nct6775.c |
| 6478 | 6479 | ||
| 6479 | NETEFFECT IWARP RNIC DRIVER (IW_NES) | 6480 | NETEFFECT IWARP RNIC DRIVER (IW_NES) |
| 6480 | M: Faisal Latif <faisal.latif@intel.com> | 6481 | M: Faisal Latif <faisal.latif@intel.com> |
| 6481 | L: linux-rdma@vger.kernel.org | 6482 | L: linux-rdma@vger.kernel.org |
| 6482 | W: http://www.intel.com/Products/Server/Adapters/Server-Cluster/Server-Cluster-overview.htm | 6483 | W: http://www.intel.com/Products/Server/Adapters/Server-Cluster/Server-Cluster-overview.htm |
| 6483 | S: Supported | 6484 | S: Supported |
| 6484 | F: drivers/infiniband/hw/nes/ | 6485 | F: drivers/infiniband/hw/nes/ |
| 6485 | 6486 | ||
| 6486 | NETEM NETWORK EMULATOR | 6487 | NETEM NETWORK EMULATOR |
| 6487 | M: Stephen Hemminger <stephen@networkplumber.org> | 6488 | M: Stephen Hemminger <stephen@networkplumber.org> |
| 6488 | L: netem@lists.linux-foundation.org | 6489 | L: netem@lists.linux-foundation.org |
| 6489 | S: Maintained | 6490 | S: Maintained |
| 6490 | F: net/sched/sch_netem.c | 6491 | F: net/sched/sch_netem.c |
| 6491 | 6492 | ||
| 6492 | NETERION 10GbE DRIVERS (s2io/vxge) | 6493 | NETERION 10GbE DRIVERS (s2io/vxge) |
| 6493 | M: Jon Mason <jdmason@kudzu.us> | 6494 | M: Jon Mason <jdmason@kudzu.us> |
| 6494 | L: netdev@vger.kernel.org | 6495 | L: netdev@vger.kernel.org |
| 6495 | S: Supported | 6496 | S: Supported |
| 6496 | F: Documentation/networking/s2io.txt | 6497 | F: Documentation/networking/s2io.txt |
| 6497 | F: Documentation/networking/vxge.txt | 6498 | F: Documentation/networking/vxge.txt |
| 6498 | F: drivers/net/ethernet/neterion/ | 6499 | F: drivers/net/ethernet/neterion/ |
| 6499 | 6500 | ||
| 6500 | NETFILTER ({IP,IP6,ARP,EB,NF}TABLES) | 6501 | NETFILTER ({IP,IP6,ARP,EB,NF}TABLES) |
| 6501 | M: Pablo Neira Ayuso <pablo@netfilter.org> | 6502 | M: Pablo Neira Ayuso <pablo@netfilter.org> |
| 6502 | M: Patrick McHardy <kaber@trash.net> | 6503 | M: Patrick McHardy <kaber@trash.net> |
| 6503 | M: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> | 6504 | M: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> |
| 6504 | L: netfilter-devel@vger.kernel.org | 6505 | L: netfilter-devel@vger.kernel.org |
| 6505 | L: coreteam@netfilter.org | 6506 | L: coreteam@netfilter.org |
| 6506 | W: http://www.netfilter.org/ | 6507 | W: http://www.netfilter.org/ |
| 6507 | W: http://www.iptables.org/ | 6508 | W: http://www.iptables.org/ |
| 6508 | Q: http://patchwork.ozlabs.org/project/netfilter-devel/list/ | 6509 | Q: http://patchwork.ozlabs.org/project/netfilter-devel/list/ |
| 6509 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git | 6510 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git |
| 6510 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next.git | 6511 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next.git |
| 6511 | S: Supported | 6512 | S: Supported |
| 6512 | F: include/linux/netfilter* | 6513 | F: include/linux/netfilter* |
| 6513 | F: include/linux/netfilter/ | 6514 | F: include/linux/netfilter/ |
| 6514 | F: include/net/netfilter/ | 6515 | F: include/net/netfilter/ |
| 6515 | F: include/uapi/linux/netfilter* | 6516 | F: include/uapi/linux/netfilter* |
| 6516 | F: include/uapi/linux/netfilter/ | 6517 | F: include/uapi/linux/netfilter/ |
| 6517 | F: net/*/netfilter.c | 6518 | F: net/*/netfilter.c |
| 6518 | F: net/*/netfilter/ | 6519 | F: net/*/netfilter/ |
| 6519 | F: net/netfilter/ | 6520 | F: net/netfilter/ |
| 6520 | 6521 | ||
| 6521 | NETLABEL | 6522 | NETLABEL |
| 6522 | M: Paul Moore <paul@paul-moore.com> | 6523 | M: Paul Moore <paul@paul-moore.com> |
| 6523 | W: http://netlabel.sf.net | 6524 | W: http://netlabel.sf.net |
| 6524 | L: netdev@vger.kernel.org | 6525 | L: netdev@vger.kernel.org |
| 6525 | S: Maintained | 6526 | S: Maintained |
| 6526 | F: Documentation/netlabel/ | 6527 | F: Documentation/netlabel/ |
| 6527 | F: include/net/netlabel.h | 6528 | F: include/net/netlabel.h |
| 6528 | F: net/netlabel/ | 6529 | F: net/netlabel/ |
| 6529 | 6530 | ||
| 6530 | NETROM NETWORK LAYER | 6531 | NETROM NETWORK LAYER |
| 6531 | M: Ralf Baechle <ralf@linux-mips.org> | 6532 | M: Ralf Baechle <ralf@linux-mips.org> |
| 6532 | L: linux-hams@vger.kernel.org | 6533 | L: linux-hams@vger.kernel.org |
| 6533 | W: http://www.linux-ax25.org/ | 6534 | W: http://www.linux-ax25.org/ |
| 6534 | S: Maintained | 6535 | S: Maintained |
| 6535 | F: include/net/netrom.h | 6536 | F: include/net/netrom.h |
| 6536 | F: include/uapi/linux/netrom.h | 6537 | F: include/uapi/linux/netrom.h |
| 6537 | F: net/netrom/ | 6538 | F: net/netrom/ |
| 6538 | 6539 | ||
| 6539 | NETWORK BLOCK DEVICE (NBD) | 6540 | NETWORK BLOCK DEVICE (NBD) |
| 6540 | M: Paul Clements <Paul.Clements@steeleye.com> | 6541 | M: Paul Clements <Paul.Clements@steeleye.com> |
| 6541 | S: Maintained | 6542 | S: Maintained |
| 6542 | L: nbd-general@lists.sourceforge.net | 6543 | L: nbd-general@lists.sourceforge.net |
| 6543 | F: Documentation/blockdev/nbd.txt | 6544 | F: Documentation/blockdev/nbd.txt |
| 6544 | F: drivers/block/nbd.c | 6545 | F: drivers/block/nbd.c |
| 6545 | F: include/linux/nbd.h | 6546 | F: include/linux/nbd.h |
| 6546 | F: include/uapi/linux/nbd.h | 6547 | F: include/uapi/linux/nbd.h |
| 6547 | 6548 | ||
| 6548 | NETWORK DROP MONITOR | 6549 | NETWORK DROP MONITOR |
| 6549 | M: Neil Horman <nhorman@tuxdriver.com> | 6550 | M: Neil Horman <nhorman@tuxdriver.com> |
| 6550 | L: netdev@vger.kernel.org | 6551 | L: netdev@vger.kernel.org |
| 6551 | S: Maintained | 6552 | S: Maintained |
| 6552 | W: https://fedorahosted.org/dropwatch/ | 6553 | W: https://fedorahosted.org/dropwatch/ |
| 6553 | F: net/core/drop_monitor.c | 6554 | F: net/core/drop_monitor.c |
| 6554 | 6555 | ||
| 6555 | NETWORKING [GENERAL] | 6556 | NETWORKING [GENERAL] |
| 6556 | M: "David S. Miller" <davem@davemloft.net> | 6557 | M: "David S. Miller" <davem@davemloft.net> |
| 6557 | L: netdev@vger.kernel.org | 6558 | L: netdev@vger.kernel.org |
| 6558 | W: http://www.linuxfoundation.org/en/Net | 6559 | W: http://www.linuxfoundation.org/en/Net |
| 6559 | Q: http://patchwork.ozlabs.org/project/netdev/list/ | 6560 | Q: http://patchwork.ozlabs.org/project/netdev/list/ |
| 6560 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git | 6561 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git |
| 6561 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git | 6562 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git |
| 6562 | S: Maintained | 6563 | S: Maintained |
| 6563 | F: net/ | 6564 | F: net/ |
| 6564 | F: include/net/ | 6565 | F: include/net/ |
| 6565 | F: include/linux/in.h | 6566 | F: include/linux/in.h |
| 6566 | F: include/linux/net.h | 6567 | F: include/linux/net.h |
| 6567 | F: include/linux/netdevice.h | 6568 | F: include/linux/netdevice.h |
| 6568 | F: include/uapi/linux/in.h | 6569 | F: include/uapi/linux/in.h |
| 6569 | F: include/uapi/linux/net.h | 6570 | F: include/uapi/linux/net.h |
| 6570 | F: include/uapi/linux/netdevice.h | 6571 | F: include/uapi/linux/netdevice.h |
| 6571 | F: tools/net/ | 6572 | F: tools/net/ |
| 6572 | F: tools/testing/selftests/net/ | 6573 | F: tools/testing/selftests/net/ |
| 6573 | F: lib/random32.c | 6574 | F: lib/random32.c |
| 6574 | F: lib/test_bpf.c | 6575 | F: lib/test_bpf.c |
| 6575 | 6576 | ||
| 6576 | NETWORKING [IPv4/IPv6] | 6577 | NETWORKING [IPv4/IPv6] |
| 6577 | M: "David S. Miller" <davem@davemloft.net> | 6578 | M: "David S. Miller" <davem@davemloft.net> |
| 6578 | M: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru> | 6579 | M: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru> |
| 6579 | M: James Morris <jmorris@namei.org> | 6580 | M: James Morris <jmorris@namei.org> |
| 6580 | M: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org> | 6581 | M: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org> |
| 6581 | M: Patrick McHardy <kaber@trash.net> | 6582 | M: Patrick McHardy <kaber@trash.net> |
| 6582 | L: netdev@vger.kernel.org | 6583 | L: netdev@vger.kernel.org |
| 6583 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git | 6584 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git |
| 6584 | S: Maintained | 6585 | S: Maintained |
| 6585 | F: net/ipv4/ | 6586 | F: net/ipv4/ |
| 6586 | F: net/ipv6/ | 6587 | F: net/ipv6/ |
| 6587 | F: include/net/ip* | 6588 | F: include/net/ip* |
| 6588 | F: arch/x86/net/* | 6589 | F: arch/x86/net/* |
| 6589 | 6590 | ||
| 6590 | NETWORKING [IPSEC] | 6591 | NETWORKING [IPSEC] |
| 6591 | M: Steffen Klassert <steffen.klassert@secunet.com> | 6592 | M: Steffen Klassert <steffen.klassert@secunet.com> |
| 6592 | M: Herbert Xu <herbert@gondor.apana.org.au> | 6593 | M: Herbert Xu <herbert@gondor.apana.org.au> |
| 6593 | M: "David S. Miller" <davem@davemloft.net> | 6594 | M: "David S. Miller" <davem@davemloft.net> |
| 6594 | L: netdev@vger.kernel.org | 6595 | L: netdev@vger.kernel.org |
| 6595 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec.git | 6596 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec.git |
| 6596 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next.git | 6597 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next.git |
| 6597 | S: Maintained | 6598 | S: Maintained |
| 6598 | F: net/core/flow.c | 6599 | F: net/core/flow.c |
| 6599 | F: net/xfrm/ | 6600 | F: net/xfrm/ |
| 6600 | F: net/key/ | 6601 | F: net/key/ |
| 6601 | F: net/ipv4/xfrm* | 6602 | F: net/ipv4/xfrm* |
| 6602 | F: net/ipv4/esp4.c | 6603 | F: net/ipv4/esp4.c |
| 6603 | F: net/ipv4/ah4.c | 6604 | F: net/ipv4/ah4.c |
| 6604 | F: net/ipv4/ipcomp.c | 6605 | F: net/ipv4/ipcomp.c |
| 6605 | F: net/ipv4/ip_vti.c | 6606 | F: net/ipv4/ip_vti.c |
| 6606 | F: net/ipv6/xfrm* | 6607 | F: net/ipv6/xfrm* |
| 6607 | F: net/ipv6/esp6.c | 6608 | F: net/ipv6/esp6.c |
| 6608 | F: net/ipv6/ah6.c | 6609 | F: net/ipv6/ah6.c |
| 6609 | F: net/ipv6/ipcomp6.c | 6610 | F: net/ipv6/ipcomp6.c |
| 6610 | F: net/ipv6/ip6_vti.c | 6611 | F: net/ipv6/ip6_vti.c |
| 6611 | F: include/uapi/linux/xfrm.h | 6612 | F: include/uapi/linux/xfrm.h |
| 6612 | F: include/net/xfrm.h | 6613 | F: include/net/xfrm.h |
| 6613 | 6614 | ||
| 6614 | NETWORKING [LABELED] (NetLabel, CIPSO, Labeled IPsec, SECMARK) | 6615 | NETWORKING [LABELED] (NetLabel, CIPSO, Labeled IPsec, SECMARK) |
| 6615 | M: Paul Moore <paul@paul-moore.com> | 6616 | M: Paul Moore <paul@paul-moore.com> |
| 6616 | L: netdev@vger.kernel.org | 6617 | L: netdev@vger.kernel.org |
| 6617 | S: Maintained | 6618 | S: Maintained |
| 6618 | 6619 | ||
| 6619 | NETWORKING [WIRELESS] | 6620 | NETWORKING [WIRELESS] |
| 6620 | L: linux-wireless@vger.kernel.org | 6621 | L: linux-wireless@vger.kernel.org |
| 6621 | Q: http://patchwork.kernel.org/project/linux-wireless/list/ | 6622 | Q: http://patchwork.kernel.org/project/linux-wireless/list/ |
| 6622 | 6623 | ||
| 6623 | NETWORKING DRIVERS | 6624 | NETWORKING DRIVERS |
| 6624 | L: netdev@vger.kernel.org | 6625 | L: netdev@vger.kernel.org |
| 6625 | W: http://www.linuxfoundation.org/en/Net | 6626 | W: http://www.linuxfoundation.org/en/Net |
| 6626 | Q: http://patchwork.ozlabs.org/project/netdev/list/ | 6627 | Q: http://patchwork.ozlabs.org/project/netdev/list/ |
| 6627 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git | 6628 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git |
| 6628 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git | 6629 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git |
| 6629 | S: Odd Fixes | 6630 | S: Odd Fixes |
| 6630 | F: drivers/net/ | 6631 | F: drivers/net/ |
| 6631 | F: include/linux/if_* | 6632 | F: include/linux/if_* |
| 6632 | F: include/linux/netdevice.h | 6633 | F: include/linux/netdevice.h |
| 6633 | F: include/linux/arcdevice.h | 6634 | F: include/linux/arcdevice.h |
| 6634 | F: include/linux/etherdevice.h | 6635 | F: include/linux/etherdevice.h |
| 6635 | F: include/linux/fcdevice.h | 6636 | F: include/linux/fcdevice.h |
| 6636 | F: include/linux/fddidevice.h | 6637 | F: include/linux/fddidevice.h |
| 6637 | F: include/linux/hippidevice.h | 6638 | F: include/linux/hippidevice.h |
| 6638 | F: include/linux/inetdevice.h | 6639 | F: include/linux/inetdevice.h |
| 6639 | F: include/uapi/linux/if_* | 6640 | F: include/uapi/linux/if_* |
| 6640 | F: include/uapi/linux/netdevice.h | 6641 | F: include/uapi/linux/netdevice.h |
| 6641 | 6642 | ||
| 6642 | NETWORKING DRIVERS (WIRELESS) | 6643 | NETWORKING DRIVERS (WIRELESS) |
| 6643 | M: Kalle Valo <kvalo@codeaurora.org> | 6644 | M: Kalle Valo <kvalo@codeaurora.org> |
| 6644 | L: linux-wireless@vger.kernel.org | 6645 | L: linux-wireless@vger.kernel.org |
| 6645 | Q: http://patchwork.kernel.org/project/linux-wireless/list/ | 6646 | Q: http://patchwork.kernel.org/project/linux-wireless/list/ |
| 6646 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers.git/ | 6647 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers.git/ |
| 6647 | S: Maintained | 6648 | S: Maintained |
| 6648 | F: drivers/net/wireless/ | 6649 | F: drivers/net/wireless/ |
| 6649 | 6650 | ||
| 6650 | NETXEN (1/10) GbE SUPPORT | 6651 | NETXEN (1/10) GbE SUPPORT |
| 6651 | M: Manish Chopra <manish.chopra@qlogic.com> | 6652 | M: Manish Chopra <manish.chopra@qlogic.com> |
| 6652 | M: Sony Chacko <sony.chacko@qlogic.com> | 6653 | M: Sony Chacko <sony.chacko@qlogic.com> |
| 6653 | M: Rajesh Borundia <rajesh.borundia@qlogic.com> | 6654 | M: Rajesh Borundia <rajesh.borundia@qlogic.com> |
| 6654 | L: netdev@vger.kernel.org | 6655 | L: netdev@vger.kernel.org |
| 6655 | W: http://www.qlogic.com | 6656 | W: http://www.qlogic.com |
| 6656 | S: Supported | 6657 | S: Supported |
| 6657 | F: drivers/net/ethernet/qlogic/netxen/ | 6658 | F: drivers/net/ethernet/qlogic/netxen/ |
| 6658 | 6659 | ||
| 6659 | NFC SUBSYSTEM | 6660 | NFC SUBSYSTEM |
| 6660 | M: Lauro Ramos Venancio <lauro.venancio@openbossa.org> | 6661 | M: Lauro Ramos Venancio <lauro.venancio@openbossa.org> |
| 6661 | M: Aloisio Almeida Jr <aloisio.almeida@openbossa.org> | 6662 | M: Aloisio Almeida Jr <aloisio.almeida@openbossa.org> |
| 6662 | M: Samuel Ortiz <sameo@linux.intel.com> | 6663 | M: Samuel Ortiz <sameo@linux.intel.com> |
| 6663 | L: linux-wireless@vger.kernel.org | 6664 | L: linux-wireless@vger.kernel.org |
| 6664 | L: linux-nfc@lists.01.org (subscribers-only) | 6665 | L: linux-nfc@lists.01.org (subscribers-only) |
| 6665 | S: Supported | 6666 | S: Supported |
| 6666 | F: net/nfc/ | 6667 | F: net/nfc/ |
| 6667 | F: include/net/nfc/ | 6668 | F: include/net/nfc/ |
| 6668 | F: include/uapi/linux/nfc.h | 6669 | F: include/uapi/linux/nfc.h |
| 6669 | F: drivers/nfc/ | 6670 | F: drivers/nfc/ |
| 6670 | F: include/linux/platform_data/pn544.h | 6671 | F: include/linux/platform_data/pn544.h |
| 6671 | F: Documentation/devicetree/bindings/net/nfc/ | 6672 | F: Documentation/devicetree/bindings/net/nfc/ |
| 6672 | 6673 | ||
| 6673 | NFS, SUNRPC, AND LOCKD CLIENTS | 6674 | NFS, SUNRPC, AND LOCKD CLIENTS |
| 6674 | M: Trond Myklebust <trond.myklebust@primarydata.com> | 6675 | M: Trond Myklebust <trond.myklebust@primarydata.com> |
| 6675 | L: linux-nfs@vger.kernel.org | 6676 | L: linux-nfs@vger.kernel.org |
| 6676 | W: http://client.linux-nfs.org | 6677 | W: http://client.linux-nfs.org |
| 6677 | T: git git://git.linux-nfs.org/projects/trondmy/linux-nfs.git | 6678 | T: git git://git.linux-nfs.org/projects/trondmy/linux-nfs.git |
| 6678 | S: Maintained | 6679 | S: Maintained |
| 6679 | F: fs/lockd/ | 6680 | F: fs/lockd/ |
| 6680 | F: fs/nfs/ | 6681 | F: fs/nfs/ |
| 6681 | F: fs/nfs_common/ | 6682 | F: fs/nfs_common/ |
| 6682 | F: net/sunrpc/ | 6683 | F: net/sunrpc/ |
| 6683 | F: include/linux/lockd/ | 6684 | F: include/linux/lockd/ |
| 6684 | F: include/linux/nfs* | 6685 | F: include/linux/nfs* |
| 6685 | F: include/linux/sunrpc/ | 6686 | F: include/linux/sunrpc/ |
| 6686 | F: include/uapi/linux/nfs* | 6687 | F: include/uapi/linux/nfs* |
| 6687 | F: include/uapi/linux/sunrpc/ | 6688 | F: include/uapi/linux/sunrpc/ |
| 6688 | 6689 | ||
| 6689 | NILFS2 FILESYSTEM | 6690 | NILFS2 FILESYSTEM |
| 6690 | M: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> | 6691 | M: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> |
| 6691 | L: linux-nilfs@vger.kernel.org | 6692 | L: linux-nilfs@vger.kernel.org |
| 6692 | W: http://nilfs.sourceforge.net/ | 6693 | W: http://nilfs.sourceforge.net/ |
| 6693 | T: git git://github.com/konis/nilfs2.git | 6694 | T: git git://github.com/konis/nilfs2.git |
| 6694 | S: Supported | 6695 | S: Supported |
| 6695 | F: Documentation/filesystems/nilfs2.txt | 6696 | F: Documentation/filesystems/nilfs2.txt |
| 6696 | F: fs/nilfs2/ | 6697 | F: fs/nilfs2/ |
| 6697 | F: include/linux/nilfs2_fs.h | 6698 | F: include/linux/nilfs2_fs.h |
| 6698 | 6699 | ||
| 6699 | NINJA SCSI-3 / NINJA SCSI-32Bi (16bit/CardBus) PCMCIA SCSI HOST ADAPTER DRIVER | 6700 | NINJA SCSI-3 / NINJA SCSI-32Bi (16bit/CardBus) PCMCIA SCSI HOST ADAPTER DRIVER |
| 6700 | M: YOKOTA Hiroshi <yokota@netlab.is.tsukuba.ac.jp> | 6701 | M: YOKOTA Hiroshi <yokota@netlab.is.tsukuba.ac.jp> |
| 6701 | W: http://www.netlab.is.tsukuba.ac.jp/~yokota/izumi/ninja/ | 6702 | W: http://www.netlab.is.tsukuba.ac.jp/~yokota/izumi/ninja/ |
| 6702 | S: Maintained | 6703 | S: Maintained |
| 6703 | F: Documentation/scsi/NinjaSCSI.txt | 6704 | F: Documentation/scsi/NinjaSCSI.txt |
| 6704 | F: drivers/scsi/pcmcia/nsp_* | 6705 | F: drivers/scsi/pcmcia/nsp_* |
| 6705 | 6706 | ||
| 6706 | NINJA SCSI-32Bi/UDE PCI/CARDBUS SCSI HOST ADAPTER DRIVER | 6707 | NINJA SCSI-32Bi/UDE PCI/CARDBUS SCSI HOST ADAPTER DRIVER |
| 6707 | M: GOTO Masanori <gotom@debian.or.jp> | 6708 | M: GOTO Masanori <gotom@debian.or.jp> |
| 6708 | M: YOKOTA Hiroshi <yokota@netlab.is.tsukuba.ac.jp> | 6709 | M: YOKOTA Hiroshi <yokota@netlab.is.tsukuba.ac.jp> |
| 6709 | W: http://www.netlab.is.tsukuba.ac.jp/~yokota/izumi/ninja/ | 6710 | W: http://www.netlab.is.tsukuba.ac.jp/~yokota/izumi/ninja/ |
| 6710 | S: Maintained | 6711 | S: Maintained |
| 6711 | F: Documentation/scsi/NinjaSCSI.txt | 6712 | F: Documentation/scsi/NinjaSCSI.txt |
| 6712 | F: drivers/scsi/nsp32* | 6713 | F: drivers/scsi/nsp32* |
| 6713 | 6714 | ||
| 6714 | NIOS2 ARCHITECTURE | 6715 | NIOS2 ARCHITECTURE |
| 6715 | M: Ley Foon Tan <lftan@altera.com> | 6716 | M: Ley Foon Tan <lftan@altera.com> |
| 6716 | L: nios2-dev@lists.rocketboards.org (moderated for non-subscribers) | 6717 | L: nios2-dev@lists.rocketboards.org (moderated for non-subscribers) |
| 6717 | T: git git://git.rocketboards.org/linux-socfpga.git | 6718 | T: git git://git.rocketboards.org/linux-socfpga.git |
| 6718 | S: Maintained | 6719 | S: Maintained |
| 6719 | F: arch/nios2/ | 6720 | F: arch/nios2/ |
| 6720 | 6721 | ||
| 6721 | NTB DRIVER | 6722 | NTB DRIVER |
| 6722 | M: Jon Mason <jdmason@kudzu.us> | 6723 | M: Jon Mason <jdmason@kudzu.us> |
| 6723 | M: Dave Jiang <dave.jiang@intel.com> | 6724 | M: Dave Jiang <dave.jiang@intel.com> |
| 6724 | S: Supported | 6725 | S: Supported |
| 6725 | W: https://github.com/jonmason/ntb/wiki | 6726 | W: https://github.com/jonmason/ntb/wiki |
| 6726 | T: git git://github.com/jonmason/ntb.git | 6727 | T: git git://github.com/jonmason/ntb.git |
| 6727 | F: drivers/ntb/ | 6728 | F: drivers/ntb/ |
| 6728 | F: drivers/net/ntb_netdev.c | 6729 | F: drivers/net/ntb_netdev.c |
| 6729 | F: include/linux/ntb.h | 6730 | F: include/linux/ntb.h |
| 6730 | 6731 | ||
| 6731 | NTFS FILESYSTEM | 6732 | NTFS FILESYSTEM |
| 6732 | M: Anton Altaparmakov <anton@tuxera.com> | 6733 | M: Anton Altaparmakov <anton@tuxera.com> |
| 6733 | L: linux-ntfs-dev@lists.sourceforge.net | 6734 | L: linux-ntfs-dev@lists.sourceforge.net |
| 6734 | W: http://www.tuxera.com/ | 6735 | W: http://www.tuxera.com/ |
| 6735 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/aia21/ntfs.git | 6736 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/aia21/ntfs.git |
| 6736 | S: Supported | 6737 | S: Supported |
| 6737 | F: Documentation/filesystems/ntfs.txt | 6738 | F: Documentation/filesystems/ntfs.txt |
| 6738 | F: fs/ntfs/ | 6739 | F: fs/ntfs/ |
| 6739 | 6740 | ||
| 6740 | NVIDIA (rivafb and nvidiafb) FRAMEBUFFER DRIVER | 6741 | NVIDIA (rivafb and nvidiafb) FRAMEBUFFER DRIVER |
| 6741 | M: Antonino Daplas <adaplas@gmail.com> | 6742 | M: Antonino Daplas <adaplas@gmail.com> |
| 6742 | L: linux-fbdev@vger.kernel.org | 6743 | L: linux-fbdev@vger.kernel.org |
| 6743 | S: Maintained | 6744 | S: Maintained |
| 6744 | F: drivers/video/fbdev/riva/ | 6745 | F: drivers/video/fbdev/riva/ |
| 6745 | F: drivers/video/fbdev/nvidia/ | 6746 | F: drivers/video/fbdev/nvidia/ |
| 6746 | 6747 | ||
| 6747 | NVM EXPRESS DRIVER | 6748 | NVM EXPRESS DRIVER |
| 6748 | M: Matthew Wilcox <willy@linux.intel.com> | 6749 | M: Matthew Wilcox <willy@linux.intel.com> |
| 6749 | L: linux-nvme@lists.infradead.org | 6750 | L: linux-nvme@lists.infradead.org |
| 6750 | T: git git://git.infradead.org/users/willy/linux-nvme.git | 6751 | T: git git://git.infradead.org/users/willy/linux-nvme.git |
| 6751 | S: Supported | 6752 | S: Supported |
| 6752 | F: drivers/block/nvme* | 6753 | F: drivers/block/nvme* |
| 6753 | F: include/linux/nvme.h | 6754 | F: include/linux/nvme.h |
| 6754 | 6755 | ||
| 6755 | NXP TDA998X DRM DRIVER | 6756 | NXP TDA998X DRM DRIVER |
| 6756 | M: Russell King <rmk+kernel@arm.linux.org.uk> | 6757 | M: Russell King <rmk+kernel@arm.linux.org.uk> |
| 6757 | S: Supported | 6758 | S: Supported |
| 6758 | F: drivers/gpu/drm/i2c/tda998x_drv.c | 6759 | F: drivers/gpu/drm/i2c/tda998x_drv.c |
| 6759 | F: include/drm/i2c/tda998x.h | 6760 | F: include/drm/i2c/tda998x.h |
| 6760 | 6761 | ||
| 6761 | NXP TFA9879 DRIVER | 6762 | NXP TFA9879 DRIVER |
| 6762 | M: Peter Rosin <peda@axentia.se> | 6763 | M: Peter Rosin <peda@axentia.se> |
| 6763 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 6764 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 6764 | S: Maintained | 6765 | S: Maintained |
| 6765 | F: sound/soc/codecs/tfa9879* | 6766 | F: sound/soc/codecs/tfa9879* |
| 6766 | 6767 | ||
| 6767 | OMAP SUPPORT | 6768 | OMAP SUPPORT |
| 6768 | M: Tony Lindgren <tony@atomide.com> | 6769 | M: Tony Lindgren <tony@atomide.com> |
| 6769 | L: linux-omap@vger.kernel.org | 6770 | L: linux-omap@vger.kernel.org |
| 6770 | W: http://www.muru.com/linux/omap/ | 6771 | W: http://www.muru.com/linux/omap/ |
| 6771 | W: http://linux.omap.com/ | 6772 | W: http://linux.omap.com/ |
| 6772 | Q: http://patchwork.kernel.org/project/linux-omap/list/ | 6773 | Q: http://patchwork.kernel.org/project/linux-omap/list/ |
| 6773 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap.git | 6774 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap.git |
| 6774 | S: Maintained | 6775 | S: Maintained |
| 6775 | F: arch/arm/*omap*/ | 6776 | F: arch/arm/*omap*/ |
| 6776 | F: drivers/i2c/busses/i2c-omap.c | 6777 | F: drivers/i2c/busses/i2c-omap.c |
| 6777 | F: drivers/irqchip/irq-omap-intc.c | 6778 | F: drivers/irqchip/irq-omap-intc.c |
| 6778 | F: drivers/mfd/*omap*.c | 6779 | F: drivers/mfd/*omap*.c |
| 6779 | F: drivers/mfd/menelaus.c | 6780 | F: drivers/mfd/menelaus.c |
| 6780 | F: drivers/mfd/palmas.c | 6781 | F: drivers/mfd/palmas.c |
| 6781 | F: drivers/mfd/tps65217.c | 6782 | F: drivers/mfd/tps65217.c |
| 6782 | F: drivers/mfd/tps65218.c | 6783 | F: drivers/mfd/tps65218.c |
| 6783 | F: drivers/mfd/tps65910.c | 6784 | F: drivers/mfd/tps65910.c |
| 6784 | F: drivers/mfd/twl-core.[ch] | 6785 | F: drivers/mfd/twl-core.[ch] |
| 6785 | F: drivers/mfd/twl4030*.c | 6786 | F: drivers/mfd/twl4030*.c |
| 6786 | F: drivers/mfd/twl6030*.c | 6787 | F: drivers/mfd/twl6030*.c |
| 6787 | F: drivers/mfd/twl6040*.c | 6788 | F: drivers/mfd/twl6040*.c |
| 6788 | F: drivers/regulator/palmas-regulator*.c | 6789 | F: drivers/regulator/palmas-regulator*.c |
| 6789 | F: drivers/regulator/pbias-regulator.c | 6790 | F: drivers/regulator/pbias-regulator.c |
| 6790 | F: drivers/regulator/tps65217-regulator.c | 6791 | F: drivers/regulator/tps65217-regulator.c |
| 6791 | F: drivers/regulator/tps65218-regulator.c | 6792 | F: drivers/regulator/tps65218-regulator.c |
| 6792 | F: drivers/regulator/tps65910-regulator.c | 6793 | F: drivers/regulator/tps65910-regulator.c |
| 6793 | F: drivers/regulator/twl-regulator.c | 6794 | F: drivers/regulator/twl-regulator.c |
| 6794 | F: include/linux/i2c-omap.h | 6795 | F: include/linux/i2c-omap.h |
| 6795 | 6796 | ||
| 6796 | OMAP DEVICE TREE SUPPORT | 6797 | OMAP DEVICE TREE SUPPORT |
| 6797 | M: Benoรฎt Cousson <bcousson@baylibre.com> | 6798 | M: Benoรฎt Cousson <bcousson@baylibre.com> |
| 6798 | M: Tony Lindgren <tony@atomide.com> | 6799 | M: Tony Lindgren <tony@atomide.com> |
| 6799 | L: linux-omap@vger.kernel.org | 6800 | L: linux-omap@vger.kernel.org |
| 6800 | L: devicetree@vger.kernel.org | 6801 | L: devicetree@vger.kernel.org |
| 6801 | S: Maintained | 6802 | S: Maintained |
| 6802 | F: arch/arm/boot/dts/*omap* | 6803 | F: arch/arm/boot/dts/*omap* |
| 6803 | F: arch/arm/boot/dts/*am3* | 6804 | F: arch/arm/boot/dts/*am3* |
| 6804 | F: arch/arm/boot/dts/*am4* | 6805 | F: arch/arm/boot/dts/*am4* |
| 6805 | F: arch/arm/boot/dts/*am5* | 6806 | F: arch/arm/boot/dts/*am5* |
| 6806 | F: arch/arm/boot/dts/*dra7* | 6807 | F: arch/arm/boot/dts/*dra7* |
| 6807 | 6808 | ||
| 6808 | OMAP CLOCK FRAMEWORK SUPPORT | 6809 | OMAP CLOCK FRAMEWORK SUPPORT |
| 6809 | M: Paul Walmsley <paul@pwsan.com> | 6810 | M: Paul Walmsley <paul@pwsan.com> |
| 6810 | L: linux-omap@vger.kernel.org | 6811 | L: linux-omap@vger.kernel.org |
| 6811 | S: Maintained | 6812 | S: Maintained |
| 6812 | F: arch/arm/*omap*/*clock* | 6813 | F: arch/arm/*omap*/*clock* |
| 6813 | 6814 | ||
| 6814 | OMAP POWER MANAGEMENT SUPPORT | 6815 | OMAP POWER MANAGEMENT SUPPORT |
| 6815 | M: Kevin Hilman <khilman@deeprootsystems.com> | 6816 | M: Kevin Hilman <khilman@deeprootsystems.com> |
| 6816 | L: linux-omap@vger.kernel.org | 6817 | L: linux-omap@vger.kernel.org |
| 6817 | S: Maintained | 6818 | S: Maintained |
| 6818 | F: arch/arm/*omap*/*pm* | 6819 | F: arch/arm/*omap*/*pm* |
| 6819 | F: drivers/cpufreq/omap-cpufreq.c | 6820 | F: drivers/cpufreq/omap-cpufreq.c |
| 6820 | 6821 | ||
| 6821 | OMAP POWERDOMAIN SOC ADAPTATION LAYER SUPPORT | 6822 | OMAP POWERDOMAIN SOC ADAPTATION LAYER SUPPORT |
| 6822 | M: Rajendra Nayak <rnayak@ti.com> | 6823 | M: Rajendra Nayak <rnayak@ti.com> |
| 6823 | M: Paul Walmsley <paul@pwsan.com> | 6824 | M: Paul Walmsley <paul@pwsan.com> |
| 6824 | L: linux-omap@vger.kernel.org | 6825 | L: linux-omap@vger.kernel.org |
| 6825 | S: Maintained | 6826 | S: Maintained |
| 6826 | F: arch/arm/mach-omap2/prm* | 6827 | F: arch/arm/mach-omap2/prm* |
| 6827 | 6828 | ||
| 6828 | OMAP AUDIO SUPPORT | 6829 | OMAP AUDIO SUPPORT |
| 6829 | M: Peter Ujfalusi <peter.ujfalusi@ti.com> | 6830 | M: Peter Ujfalusi <peter.ujfalusi@ti.com> |
| 6830 | M: Jarkko Nikula <jarkko.nikula@bitmer.com> | 6831 | M: Jarkko Nikula <jarkko.nikula@bitmer.com> |
| 6831 | L: alsa-devel@alsa-project.org (subscribers-only) | 6832 | L: alsa-devel@alsa-project.org (subscribers-only) |
| 6832 | L: linux-omap@vger.kernel.org | 6833 | L: linux-omap@vger.kernel.org |
| 6833 | S: Maintained | 6834 | S: Maintained |
| 6834 | F: sound/soc/omap/ | 6835 | F: sound/soc/omap/ |
| 6835 | 6836 | ||
| 6836 | OMAP GENERAL PURPOSE MEMORY CONTROLLER SUPPORT | 6837 | OMAP GENERAL PURPOSE MEMORY CONTROLLER SUPPORT |
| 6837 | M: Roger Quadros <rogerq@ti.com> | 6838 | M: Roger Quadros <rogerq@ti.com> |
| 6838 | M: Tony Lindgren <tony@atomide.com> | 6839 | M: Tony Lindgren <tony@atomide.com> |
| 6839 | L: linux-omap@vger.kernel.org | 6840 | L: linux-omap@vger.kernel.org |
| 6840 | S: Maintained | 6841 | S: Maintained |
| 6841 | F: drivers/memory/omap-gpmc.c | 6842 | F: drivers/memory/omap-gpmc.c |
| 6842 | F: arch/arm/mach-omap2/*gpmc* | 6843 | F: arch/arm/mach-omap2/*gpmc* |
| 6843 | 6844 | ||
| 6844 | OMAP FRAMEBUFFER SUPPORT | 6845 | OMAP FRAMEBUFFER SUPPORT |
| 6845 | M: Tomi Valkeinen <tomi.valkeinen@ti.com> | 6846 | M: Tomi Valkeinen <tomi.valkeinen@ti.com> |
| 6846 | L: linux-fbdev@vger.kernel.org | 6847 | L: linux-fbdev@vger.kernel.org |
| 6847 | L: linux-omap@vger.kernel.org | 6848 | L: linux-omap@vger.kernel.org |
| 6848 | S: Maintained | 6849 | S: Maintained |
| 6849 | F: drivers/video/fbdev/omap/ | 6850 | F: drivers/video/fbdev/omap/ |
| 6850 | 6851 | ||
| 6851 | OMAP DISPLAY SUBSYSTEM and FRAMEBUFFER SUPPORT (DSS2) | 6852 | OMAP DISPLAY SUBSYSTEM and FRAMEBUFFER SUPPORT (DSS2) |
| 6852 | M: Tomi Valkeinen <tomi.valkeinen@ti.com> | 6853 | M: Tomi Valkeinen <tomi.valkeinen@ti.com> |
| 6853 | L: linux-omap@vger.kernel.org | 6854 | L: linux-omap@vger.kernel.org |
| 6854 | L: linux-fbdev@vger.kernel.org | 6855 | L: linux-fbdev@vger.kernel.org |
| 6855 | S: Maintained | 6856 | S: Maintained |
| 6856 | F: drivers/video/fbdev/omap2/ | 6857 | F: drivers/video/fbdev/omap2/ |
| 6857 | F: Documentation/arm/OMAP/DSS | 6858 | F: Documentation/arm/OMAP/DSS |
| 6858 | 6859 | ||
| 6859 | OMAP HARDWARE SPINLOCK SUPPORT | 6860 | OMAP HARDWARE SPINLOCK SUPPORT |
| 6860 | M: Ohad Ben-Cohen <ohad@wizery.com> | 6861 | M: Ohad Ben-Cohen <ohad@wizery.com> |
| 6861 | L: linux-omap@vger.kernel.org | 6862 | L: linux-omap@vger.kernel.org |
| 6862 | S: Maintained | 6863 | S: Maintained |
| 6863 | F: drivers/hwspinlock/omap_hwspinlock.c | 6864 | F: drivers/hwspinlock/omap_hwspinlock.c |
| 6864 | F: arch/arm/mach-omap2/hwspinlock.c | 6865 | F: arch/arm/mach-omap2/hwspinlock.c |
| 6865 | 6866 | ||
| 6866 | OMAP MMC SUPPORT | 6867 | OMAP MMC SUPPORT |
| 6867 | M: Jarkko Lavinen <jarkko.lavinen@nokia.com> | 6868 | M: Jarkko Lavinen <jarkko.lavinen@nokia.com> |
| 6868 | L: linux-omap@vger.kernel.org | 6869 | L: linux-omap@vger.kernel.org |
| 6869 | S: Maintained | 6870 | S: Maintained |
| 6870 | F: drivers/mmc/host/omap.c | 6871 | F: drivers/mmc/host/omap.c |
| 6871 | 6872 | ||
| 6872 | OMAP HS MMC SUPPORT | 6873 | OMAP HS MMC SUPPORT |
| 6873 | L: linux-mmc@vger.kernel.org | 6874 | L: linux-mmc@vger.kernel.org |
| 6874 | L: linux-omap@vger.kernel.org | 6875 | L: linux-omap@vger.kernel.org |
| 6875 | S: Orphan | 6876 | S: Orphan |
| 6876 | F: drivers/mmc/host/omap_hsmmc.c | 6877 | F: drivers/mmc/host/omap_hsmmc.c |
| 6877 | 6878 | ||
| 6878 | OMAP RANDOM NUMBER GENERATOR SUPPORT | 6879 | OMAP RANDOM NUMBER GENERATOR SUPPORT |
| 6879 | M: Deepak Saxena <dsaxena@plexity.net> | 6880 | M: Deepak Saxena <dsaxena@plexity.net> |
| 6880 | S: Maintained | 6881 | S: Maintained |
| 6881 | F: drivers/char/hw_random/omap-rng.c | 6882 | F: drivers/char/hw_random/omap-rng.c |
| 6882 | 6883 | ||
| 6883 | OMAP HWMOD SUPPORT | 6884 | OMAP HWMOD SUPPORT |
| 6884 | M: Benoรฎt Cousson <bcousson@baylibre.com> | 6885 | M: Benoรฎt Cousson <bcousson@baylibre.com> |
| 6885 | M: Paul Walmsley <paul@pwsan.com> | 6886 | M: Paul Walmsley <paul@pwsan.com> |
| 6886 | L: linux-omap@vger.kernel.org | 6887 | L: linux-omap@vger.kernel.org |
| 6887 | S: Maintained | 6888 | S: Maintained |
| 6888 | F: arch/arm/mach-omap2/omap_hwmod.* | 6889 | F: arch/arm/mach-omap2/omap_hwmod.* |
| 6889 | 6890 | ||
| 6890 | OMAP HWMOD DATA FOR OMAP4-BASED DEVICES | 6891 | OMAP HWMOD DATA FOR OMAP4-BASED DEVICES |
| 6891 | M: Benoรฎt Cousson <bcousson@baylibre.com> | 6892 | M: Benoรฎt Cousson <bcousson@baylibre.com> |
| 6892 | L: linux-omap@vger.kernel.org | 6893 | L: linux-omap@vger.kernel.org |
| 6893 | S: Maintained | 6894 | S: Maintained |
| 6894 | F: arch/arm/mach-omap2/omap_hwmod_44xx_data.c | 6895 | F: arch/arm/mach-omap2/omap_hwmod_44xx_data.c |
| 6895 | 6896 | ||
| 6896 | OMAP IMAGING SUBSYSTEM (OMAP3 ISP and OMAP4 ISS) | 6897 | OMAP IMAGING SUBSYSTEM (OMAP3 ISP and OMAP4 ISS) |
| 6897 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 6898 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 6898 | L: linux-media@vger.kernel.org | 6899 | L: linux-media@vger.kernel.org |
| 6899 | S: Maintained | 6900 | S: Maintained |
| 6900 | F: drivers/media/platform/omap3isp/ | 6901 | F: drivers/media/platform/omap3isp/ |
| 6901 | F: drivers/staging/media/omap4iss/ | 6902 | F: drivers/staging/media/omap4iss/ |
| 6902 | 6903 | ||
| 6903 | OMAP USB SUPPORT | 6904 | OMAP USB SUPPORT |
| 6904 | M: Felipe Balbi <balbi@ti.com> | 6905 | M: Felipe Balbi <balbi@ti.com> |
| 6905 | L: linux-usb@vger.kernel.org | 6906 | L: linux-usb@vger.kernel.org |
| 6906 | L: linux-omap@vger.kernel.org | 6907 | L: linux-omap@vger.kernel.org |
| 6907 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git | 6908 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git |
| 6908 | S: Maintained | 6909 | S: Maintained |
| 6909 | F: drivers/usb/*/*omap* | 6910 | F: drivers/usb/*/*omap* |
| 6910 | F: arch/arm/*omap*/usb* | 6911 | F: arch/arm/*omap*/usb* |
| 6911 | 6912 | ||
| 6912 | OMAP GPIO DRIVER | 6913 | OMAP GPIO DRIVER |
| 6913 | M: Javier Martinez Canillas <javier@dowhile0.org> | 6914 | M: Javier Martinez Canillas <javier@dowhile0.org> |
| 6914 | M: Santosh Shilimkar <ssantosh@kernel.org> | 6915 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 6915 | M: Kevin Hilman <khilman@deeprootsystems.com> | 6916 | M: Kevin Hilman <khilman@deeprootsystems.com> |
| 6916 | L: linux-omap@vger.kernel.org | 6917 | L: linux-omap@vger.kernel.org |
| 6917 | S: Maintained | 6918 | S: Maintained |
| 6918 | F: drivers/gpio/gpio-omap.c | 6919 | F: drivers/gpio/gpio-omap.c |
| 6919 | 6920 | ||
| 6920 | OMAP/NEWFLOW NANOBONE MACHINE SUPPORT | 6921 | OMAP/NEWFLOW NANOBONE MACHINE SUPPORT |
| 6921 | M: Mark Jackson <mpfj@newflow.co.uk> | 6922 | M: Mark Jackson <mpfj@newflow.co.uk> |
| 6922 | L: linux-omap@vger.kernel.org | 6923 | L: linux-omap@vger.kernel.org |
| 6923 | S: Maintained | 6924 | S: Maintained |
| 6924 | F: arch/arm/boot/dts/am335x-nano.dts | 6925 | F: arch/arm/boot/dts/am335x-nano.dts |
| 6925 | 6926 | ||
| 6926 | OMFS FILESYSTEM | 6927 | OMFS FILESYSTEM |
| 6927 | M: Bob Copeland <me@bobcopeland.com> | 6928 | M: Bob Copeland <me@bobcopeland.com> |
| 6928 | L: linux-karma-devel@lists.sourceforge.net | 6929 | L: linux-karma-devel@lists.sourceforge.net |
| 6929 | S: Maintained | 6930 | S: Maintained |
| 6930 | F: Documentation/filesystems/omfs.txt | 6931 | F: Documentation/filesystems/omfs.txt |
| 6931 | F: fs/omfs/ | 6932 | F: fs/omfs/ |
| 6932 | 6933 | ||
| 6933 | OMNIKEY CARDMAN 4000 DRIVER | 6934 | OMNIKEY CARDMAN 4000 DRIVER |
| 6934 | M: Harald Welte <laforge@gnumonks.org> | 6935 | M: Harald Welte <laforge@gnumonks.org> |
| 6935 | S: Maintained | 6936 | S: Maintained |
| 6936 | F: drivers/char/pcmcia/cm4000_cs.c | 6937 | F: drivers/char/pcmcia/cm4000_cs.c |
| 6937 | F: include/linux/cm4000_cs.h | 6938 | F: include/linux/cm4000_cs.h |
| 6938 | F: include/uapi/linux/cm4000_cs.h | 6939 | F: include/uapi/linux/cm4000_cs.h |
| 6939 | 6940 | ||
| 6940 | OMNIKEY CARDMAN 4040 DRIVER | 6941 | OMNIKEY CARDMAN 4040 DRIVER |
| 6941 | M: Harald Welte <laforge@gnumonks.org> | 6942 | M: Harald Welte <laforge@gnumonks.org> |
| 6942 | S: Maintained | 6943 | S: Maintained |
| 6943 | F: drivers/char/pcmcia/cm4040_cs.* | 6944 | F: drivers/char/pcmcia/cm4040_cs.* |
| 6944 | 6945 | ||
| 6945 | OMNIVISION OV7670 SENSOR DRIVER | 6946 | OMNIVISION OV7670 SENSOR DRIVER |
| 6946 | M: Jonathan Corbet <corbet@lwn.net> | 6947 | M: Jonathan Corbet <corbet@lwn.net> |
| 6947 | L: linux-media@vger.kernel.org | 6948 | L: linux-media@vger.kernel.org |
| 6948 | T: git git://linuxtv.org/media_tree.git | 6949 | T: git git://linuxtv.org/media_tree.git |
| 6949 | S: Maintained | 6950 | S: Maintained |
| 6950 | F: drivers/media/i2c/ov7670.c | 6951 | F: drivers/media/i2c/ov7670.c |
| 6951 | 6952 | ||
| 6952 | ONENAND FLASH DRIVER | 6953 | ONENAND FLASH DRIVER |
| 6953 | M: Kyungmin Park <kyungmin.park@samsung.com> | 6954 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 6954 | L: linux-mtd@lists.infradead.org | 6955 | L: linux-mtd@lists.infradead.org |
| 6955 | S: Maintained | 6956 | S: Maintained |
| 6956 | F: drivers/mtd/onenand/ | 6957 | F: drivers/mtd/onenand/ |
| 6957 | F: include/linux/mtd/onenand*.h | 6958 | F: include/linux/mtd/onenand*.h |
| 6958 | 6959 | ||
| 6959 | ONSTREAM SCSI TAPE DRIVER | 6960 | ONSTREAM SCSI TAPE DRIVER |
| 6960 | M: Willem Riede <osst@riede.org> | 6961 | M: Willem Riede <osst@riede.org> |
| 6961 | L: osst-users@lists.sourceforge.net | 6962 | L: osst-users@lists.sourceforge.net |
| 6962 | L: linux-scsi@vger.kernel.org | 6963 | L: linux-scsi@vger.kernel.org |
| 6963 | S: Maintained | 6964 | S: Maintained |
| 6964 | F: Documentation/scsi/osst.txt | 6965 | F: Documentation/scsi/osst.txt |
| 6965 | F: drivers/scsi/osst.* | 6966 | F: drivers/scsi/osst.* |
| 6966 | F: drivers/scsi/osst_*.h | 6967 | F: drivers/scsi/osst_*.h |
| 6967 | F: drivers/scsi/st.h | 6968 | F: drivers/scsi/st.h |
| 6968 | 6969 | ||
| 6969 | OPENCORES I2C BUS DRIVER | 6970 | OPENCORES I2C BUS DRIVER |
| 6970 | M: Peter Korsgaard <jacmet@sunsite.dk> | 6971 | M: Peter Korsgaard <jacmet@sunsite.dk> |
| 6971 | L: linux-i2c@vger.kernel.org | 6972 | L: linux-i2c@vger.kernel.org |
| 6972 | S: Maintained | 6973 | S: Maintained |
| 6973 | F: Documentation/i2c/busses/i2c-ocores | 6974 | F: Documentation/i2c/busses/i2c-ocores |
| 6974 | F: drivers/i2c/busses/i2c-ocores.c | 6975 | F: drivers/i2c/busses/i2c-ocores.c |
| 6975 | 6976 | ||
| 6976 | OPEN FIRMWARE AND FLATTENED DEVICE TREE | 6977 | OPEN FIRMWARE AND FLATTENED DEVICE TREE |
| 6977 | M: Grant Likely <grant.likely@linaro.org> | 6978 | M: Grant Likely <grant.likely@linaro.org> |
| 6978 | M: Rob Herring <robh+dt@kernel.org> | 6979 | M: Rob Herring <robh+dt@kernel.org> |
| 6979 | L: devicetree@vger.kernel.org | 6980 | L: devicetree@vger.kernel.org |
| 6980 | W: http://fdt.secretlab.ca | 6981 | W: http://fdt.secretlab.ca |
| 6981 | T: git git://git.secretlab.ca/git/linux-2.6.git | 6982 | T: git git://git.secretlab.ca/git/linux-2.6.git |
| 6982 | S: Maintained | 6983 | S: Maintained |
| 6983 | F: drivers/of/ | 6984 | F: drivers/of/ |
| 6984 | F: include/linux/of*.h | 6985 | F: include/linux/of*.h |
| 6985 | F: scripts/dtc/ | 6986 | F: scripts/dtc/ |
| 6986 | K: of_get_property | 6987 | K: of_get_property |
| 6987 | K: of_match_table | 6988 | K: of_match_table |
| 6988 | 6989 | ||
| 6989 | OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS | 6990 | OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS |
| 6990 | M: Rob Herring <robh+dt@kernel.org> | 6991 | M: Rob Herring <robh+dt@kernel.org> |
| 6991 | M: Pawel Moll <pawel.moll@arm.com> | 6992 | M: Pawel Moll <pawel.moll@arm.com> |
| 6992 | M: Mark Rutland <mark.rutland@arm.com> | 6993 | M: Mark Rutland <mark.rutland@arm.com> |
| 6993 | M: Ian Campbell <ijc+devicetree@hellion.org.uk> | 6994 | M: Ian Campbell <ijc+devicetree@hellion.org.uk> |
| 6994 | M: Kumar Gala <galak@codeaurora.org> | 6995 | M: Kumar Gala <galak@codeaurora.org> |
| 6995 | L: devicetree@vger.kernel.org | 6996 | L: devicetree@vger.kernel.org |
| 6996 | S: Maintained | 6997 | S: Maintained |
| 6997 | F: Documentation/devicetree/ | 6998 | F: Documentation/devicetree/ |
| 6998 | F: arch/*/boot/dts/ | 6999 | F: arch/*/boot/dts/ |
| 6999 | F: include/dt-bindings/ | 7000 | F: include/dt-bindings/ |
| 7000 | 7001 | ||
| 7001 | OPENRISC ARCHITECTURE | 7002 | OPENRISC ARCHITECTURE |
| 7002 | M: Jonas Bonn <jonas@southpole.se> | 7003 | M: Jonas Bonn <jonas@southpole.se> |
| 7003 | W: http://openrisc.net | 7004 | W: http://openrisc.net |
| 7004 | L: linux@lists.openrisc.net (moderated for non-subscribers) | 7005 | L: linux@lists.openrisc.net (moderated for non-subscribers) |
| 7005 | S: Maintained | 7006 | S: Maintained |
| 7006 | T: git git://openrisc.net/~jonas/linux | 7007 | T: git git://openrisc.net/~jonas/linux |
| 7007 | F: arch/openrisc/ | 7008 | F: arch/openrisc/ |
| 7008 | 7009 | ||
| 7009 | OPENVSWITCH | 7010 | OPENVSWITCH |
| 7010 | M: Pravin Shelar <pshelar@nicira.com> | 7011 | M: Pravin Shelar <pshelar@nicira.com> |
| 7011 | L: dev@openvswitch.org | 7012 | L: dev@openvswitch.org |
| 7012 | W: http://openvswitch.org | 7013 | W: http://openvswitch.org |
| 7013 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/pshelar/openvswitch.git | 7014 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/pshelar/openvswitch.git |
| 7014 | S: Maintained | 7015 | S: Maintained |
| 7015 | F: net/openvswitch/ | 7016 | F: net/openvswitch/ |
| 7016 | 7017 | ||
| 7017 | OPL4 DRIVER | 7018 | OPL4 DRIVER |
| 7018 | M: Clemens Ladisch <clemens@ladisch.de> | 7019 | M: Clemens Ladisch <clemens@ladisch.de> |
| 7019 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 7020 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 7020 | T: git git://git.alsa-project.org/alsa-kernel.git | 7021 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 7021 | S: Maintained | 7022 | S: Maintained |
| 7022 | F: sound/drivers/opl4/ | 7023 | F: sound/drivers/opl4/ |
| 7023 | 7024 | ||
| 7024 | OPROFILE | 7025 | OPROFILE |
| 7025 | M: Robert Richter <rric@kernel.org> | 7026 | M: Robert Richter <rric@kernel.org> |
| 7026 | L: oprofile-list@lists.sf.net | 7027 | L: oprofile-list@lists.sf.net |
| 7027 | S: Maintained | 7028 | S: Maintained |
| 7028 | F: arch/*/include/asm/oprofile*.h | 7029 | F: arch/*/include/asm/oprofile*.h |
| 7029 | F: arch/*/oprofile/ | 7030 | F: arch/*/oprofile/ |
| 7030 | F: drivers/oprofile/ | 7031 | F: drivers/oprofile/ |
| 7031 | F: include/linux/oprofile.h | 7032 | F: include/linux/oprofile.h |
| 7032 | 7033 | ||
| 7033 | ORACLE CLUSTER FILESYSTEM 2 (OCFS2) | 7034 | ORACLE CLUSTER FILESYSTEM 2 (OCFS2) |
| 7034 | M: Mark Fasheh <mfasheh@suse.com> | 7035 | M: Mark Fasheh <mfasheh@suse.com> |
| 7035 | M: Joel Becker <jlbec@evilplan.org> | 7036 | M: Joel Becker <jlbec@evilplan.org> |
| 7036 | L: ocfs2-devel@oss.oracle.com (moderated for non-subscribers) | 7037 | L: ocfs2-devel@oss.oracle.com (moderated for non-subscribers) |
| 7037 | W: http://oss.oracle.com/projects/ocfs2/ | 7038 | W: http://oss.oracle.com/projects/ocfs2/ |
| 7038 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2.git | 7039 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2.git |
| 7039 | S: Supported | 7040 | S: Supported |
| 7040 | F: Documentation/filesystems/ocfs2.txt | 7041 | F: Documentation/filesystems/ocfs2.txt |
| 7041 | F: Documentation/filesystems/dlmfs.txt | 7042 | F: Documentation/filesystems/dlmfs.txt |
| 7042 | F: fs/ocfs2/ | 7043 | F: fs/ocfs2/ |
| 7043 | 7044 | ||
| 7044 | ORINOCO DRIVER | 7045 | ORINOCO DRIVER |
| 7045 | L: linux-wireless@vger.kernel.org | 7046 | L: linux-wireless@vger.kernel.org |
| 7046 | W: http://wireless.kernel.org/en/users/Drivers/orinoco | 7047 | W: http://wireless.kernel.org/en/users/Drivers/orinoco |
| 7047 | W: http://www.nongnu.org/orinoco/ | 7048 | W: http://www.nongnu.org/orinoco/ |
| 7048 | S: Orphan | 7049 | S: Orphan |
| 7049 | F: drivers/net/wireless/orinoco/ | 7050 | F: drivers/net/wireless/orinoco/ |
| 7050 | 7051 | ||
| 7051 | OSD LIBRARY and FILESYSTEM | 7052 | OSD LIBRARY and FILESYSTEM |
| 7052 | M: Boaz Harrosh <ooo@electrozaur.com> | 7053 | M: Boaz Harrosh <ooo@electrozaur.com> |
| 7053 | M: Benny Halevy <bhalevy@primarydata.com> | 7054 | M: Benny Halevy <bhalevy@primarydata.com> |
| 7054 | L: osd-dev@open-osd.org | 7055 | L: osd-dev@open-osd.org |
| 7055 | W: http://open-osd.org | 7056 | W: http://open-osd.org |
| 7056 | T: git git://git.open-osd.org/open-osd.git | 7057 | T: git git://git.open-osd.org/open-osd.git |
| 7057 | S: Maintained | 7058 | S: Maintained |
| 7058 | F: drivers/scsi/osd/ | 7059 | F: drivers/scsi/osd/ |
| 7059 | F: include/scsi/osd_* | 7060 | F: include/scsi/osd_* |
| 7060 | F: fs/exofs/ | 7061 | F: fs/exofs/ |
| 7061 | 7062 | ||
| 7062 | OVERLAY FILESYSTEM | 7063 | OVERLAY FILESYSTEM |
| 7063 | M: Miklos Szeredi <miklos@szeredi.hu> | 7064 | M: Miklos Szeredi <miklos@szeredi.hu> |
| 7064 | L: linux-unionfs@vger.kernel.org | 7065 | L: linux-unionfs@vger.kernel.org |
| 7065 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs.git | 7066 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs.git |
| 7066 | S: Supported | 7067 | S: Supported |
| 7067 | F: fs/overlayfs/ | 7068 | F: fs/overlayfs/ |
| 7068 | F: Documentation/filesystems/overlayfs.txt | 7069 | F: Documentation/filesystems/overlayfs.txt |
| 7069 | 7070 | ||
| 7070 | P54 WIRELESS DRIVER | 7071 | P54 WIRELESS DRIVER |
| 7071 | M: Christian Lamparter <chunkeey@googlemail.com> | 7072 | M: Christian Lamparter <chunkeey@googlemail.com> |
| 7072 | L: linux-wireless@vger.kernel.org | 7073 | L: linux-wireless@vger.kernel.org |
| 7073 | W: http://wireless.kernel.org/en/users/Drivers/p54 | 7074 | W: http://wireless.kernel.org/en/users/Drivers/p54 |
| 7074 | S: Maintained | 7075 | S: Maintained |
| 7075 | F: drivers/net/wireless/p54/ | 7076 | F: drivers/net/wireless/p54/ |
| 7076 | 7077 | ||
| 7077 | PA SEMI ETHERNET DRIVER | 7078 | PA SEMI ETHERNET DRIVER |
| 7078 | M: Olof Johansson <olof@lixom.net> | 7079 | M: Olof Johansson <olof@lixom.net> |
| 7079 | L: netdev@vger.kernel.org | 7080 | L: netdev@vger.kernel.org |
| 7080 | S: Maintained | 7081 | S: Maintained |
| 7081 | F: drivers/net/ethernet/pasemi/* | 7082 | F: drivers/net/ethernet/pasemi/* |
| 7082 | 7083 | ||
| 7083 | PA SEMI SMBUS DRIVER | 7084 | PA SEMI SMBUS DRIVER |
| 7084 | M: Olof Johansson <olof@lixom.net> | 7085 | M: Olof Johansson <olof@lixom.net> |
| 7085 | L: linux-i2c@vger.kernel.org | 7086 | L: linux-i2c@vger.kernel.org |
| 7086 | S: Maintained | 7087 | S: Maintained |
| 7087 | F: drivers/i2c/busses/i2c-pasemi.c | 7088 | F: drivers/i2c/busses/i2c-pasemi.c |
| 7088 | 7089 | ||
| 7089 | PADATA PARALLEL EXECUTION MECHANISM | 7090 | PADATA PARALLEL EXECUTION MECHANISM |
| 7090 | M: Steffen Klassert <steffen.klassert@secunet.com> | 7091 | M: Steffen Klassert <steffen.klassert@secunet.com> |
| 7091 | L: linux-crypto@vger.kernel.org | 7092 | L: linux-crypto@vger.kernel.org |
| 7092 | S: Maintained | 7093 | S: Maintained |
| 7093 | F: kernel/padata.c | 7094 | F: kernel/padata.c |
| 7094 | F: include/linux/padata.h | 7095 | F: include/linux/padata.h |
| 7095 | F: Documentation/padata.txt | 7096 | F: Documentation/padata.txt |
| 7096 | 7097 | ||
| 7097 | PANASONIC LAPTOP ACPI EXTRAS DRIVER | 7098 | PANASONIC LAPTOP ACPI EXTRAS DRIVER |
| 7098 | M: Harald Welte <laforge@gnumonks.org> | 7099 | M: Harald Welte <laforge@gnumonks.org> |
| 7099 | L: platform-driver-x86@vger.kernel.org | 7100 | L: platform-driver-x86@vger.kernel.org |
| 7100 | S: Maintained | 7101 | S: Maintained |
| 7101 | F: drivers/platform/x86/panasonic-laptop.c | 7102 | F: drivers/platform/x86/panasonic-laptop.c |
| 7102 | 7103 | ||
| 7103 | PANASONIC MN10300/AM33/AM34 PORT | 7104 | PANASONIC MN10300/AM33/AM34 PORT |
| 7104 | M: David Howells <dhowells@redhat.com> | 7105 | M: David Howells <dhowells@redhat.com> |
| 7105 | M: Koichi Yasutake <yasutake.koichi@jp.panasonic.com> | 7106 | M: Koichi Yasutake <yasutake.koichi@jp.panasonic.com> |
| 7106 | L: linux-am33-list@redhat.com (moderated for non-subscribers) | 7107 | L: linux-am33-list@redhat.com (moderated for non-subscribers) |
| 7107 | W: ftp://ftp.redhat.com/pub/redhat/gnupro/AM33/ | 7108 | W: ftp://ftp.redhat.com/pub/redhat/gnupro/AM33/ |
| 7108 | S: Maintained | 7109 | S: Maintained |
| 7109 | F: Documentation/mn10300/ | 7110 | F: Documentation/mn10300/ |
| 7110 | F: arch/mn10300/ | 7111 | F: arch/mn10300/ |
| 7111 | 7112 | ||
| 7112 | PARALLEL PORT SUPPORT | 7113 | PARALLEL PORT SUPPORT |
| 7113 | L: linux-parport@lists.infradead.org (subscribers-only) | 7114 | L: linux-parport@lists.infradead.org (subscribers-only) |
| 7114 | S: Orphan | 7115 | S: Orphan |
| 7115 | F: drivers/parport/ | 7116 | F: drivers/parport/ |
| 7116 | F: include/linux/parport*.h | 7117 | F: include/linux/parport*.h |
| 7117 | F: drivers/char/ppdev.c | 7118 | F: drivers/char/ppdev.c |
| 7118 | F: include/uapi/linux/ppdev.h | 7119 | F: include/uapi/linux/ppdev.h |
| 7119 | 7120 | ||
| 7120 | PARAVIRT_OPS INTERFACE | 7121 | PARAVIRT_OPS INTERFACE |
| 7121 | M: Jeremy Fitzhardinge <jeremy@goop.org> | 7122 | M: Jeremy Fitzhardinge <jeremy@goop.org> |
| 7122 | M: Chris Wright <chrisw@sous-sol.org> | 7123 | M: Chris Wright <chrisw@sous-sol.org> |
| 7123 | M: Alok Kataria <akataria@vmware.com> | 7124 | M: Alok Kataria <akataria@vmware.com> |
| 7124 | M: Rusty Russell <rusty@rustcorp.com.au> | 7125 | M: Rusty Russell <rusty@rustcorp.com.au> |
| 7125 | L: virtualization@lists.linux-foundation.org | 7126 | L: virtualization@lists.linux-foundation.org |
| 7126 | S: Supported | 7127 | S: Supported |
| 7127 | F: Documentation/ia64/paravirt_ops.txt | 7128 | F: Documentation/ia64/paravirt_ops.txt |
| 7128 | F: arch/*/kernel/paravirt* | 7129 | F: arch/*/kernel/paravirt* |
| 7129 | F: arch/*/include/asm/paravirt.h | 7130 | F: arch/*/include/asm/paravirt.h |
| 7130 | 7131 | ||
| 7131 | PARIDE DRIVERS FOR PARALLEL PORT IDE DEVICES | 7132 | PARIDE DRIVERS FOR PARALLEL PORT IDE DEVICES |
| 7132 | M: Tim Waugh <tim@cyberelk.net> | 7133 | M: Tim Waugh <tim@cyberelk.net> |
| 7133 | L: linux-parport@lists.infradead.org (subscribers-only) | 7134 | L: linux-parport@lists.infradead.org (subscribers-only) |
| 7134 | W: http://www.torque.net/linux-pp.html | 7135 | W: http://www.torque.net/linux-pp.html |
| 7135 | S: Maintained | 7136 | S: Maintained |
| 7136 | F: Documentation/blockdev/paride.txt | 7137 | F: Documentation/blockdev/paride.txt |
| 7137 | F: drivers/block/paride/ | 7138 | F: drivers/block/paride/ |
| 7138 | 7139 | ||
| 7139 | PARISC ARCHITECTURE | 7140 | PARISC ARCHITECTURE |
| 7140 | M: "James E.J. Bottomley" <jejb@parisc-linux.org> | 7141 | M: "James E.J. Bottomley" <jejb@parisc-linux.org> |
| 7141 | M: Helge Deller <deller@gmx.de> | 7142 | M: Helge Deller <deller@gmx.de> |
| 7142 | L: linux-parisc@vger.kernel.org | 7143 | L: linux-parisc@vger.kernel.org |
| 7143 | W: http://www.parisc-linux.org/ | 7144 | W: http://www.parisc-linux.org/ |
| 7144 | Q: http://patchwork.kernel.org/project/linux-parisc/list/ | 7145 | Q: http://patchwork.kernel.org/project/linux-parisc/list/ |
| 7145 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/parisc-2.6.git | 7146 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/parisc-2.6.git |
| 7146 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux.git | 7147 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux.git |
| 7147 | S: Maintained | 7148 | S: Maintained |
| 7148 | F: arch/parisc/ | 7149 | F: arch/parisc/ |
| 7149 | F: Documentation/parisc/ | 7150 | F: Documentation/parisc/ |
| 7150 | F: drivers/parisc/ | 7151 | F: drivers/parisc/ |
| 7151 | F: drivers/char/agp/parisc-agp.c | 7152 | F: drivers/char/agp/parisc-agp.c |
| 7152 | F: drivers/input/serio/gscps2.c | 7153 | F: drivers/input/serio/gscps2.c |
| 7153 | F: drivers/parport/parport_gsc.* | 7154 | F: drivers/parport/parport_gsc.* |
| 7154 | F: drivers/tty/serial/8250/8250_gsc.c | 7155 | F: drivers/tty/serial/8250/8250_gsc.c |
| 7155 | F: drivers/video/fbdev/sti* | 7156 | F: drivers/video/fbdev/sti* |
| 7156 | F: drivers/video/console/sti* | 7157 | F: drivers/video/console/sti* |
| 7157 | F: drivers/video/logo/logo_parisc* | 7158 | F: drivers/video/logo/logo_parisc* |
| 7158 | 7159 | ||
| 7159 | PC87360 HARDWARE MONITORING DRIVER | 7160 | PC87360 HARDWARE MONITORING DRIVER |
| 7160 | M: Jim Cromie <jim.cromie@gmail.com> | 7161 | M: Jim Cromie <jim.cromie@gmail.com> |
| 7161 | L: lm-sensors@lm-sensors.org | 7162 | L: lm-sensors@lm-sensors.org |
| 7162 | S: Maintained | 7163 | S: Maintained |
| 7163 | F: Documentation/hwmon/pc87360 | 7164 | F: Documentation/hwmon/pc87360 |
| 7164 | F: drivers/hwmon/pc87360.c | 7165 | F: drivers/hwmon/pc87360.c |
| 7165 | 7166 | ||
| 7166 | PC8736x GPIO DRIVER | 7167 | PC8736x GPIO DRIVER |
| 7167 | M: Jim Cromie <jim.cromie@gmail.com> | 7168 | M: Jim Cromie <jim.cromie@gmail.com> |
| 7168 | S: Maintained | 7169 | S: Maintained |
| 7169 | F: drivers/char/pc8736x_gpio.c | 7170 | F: drivers/char/pc8736x_gpio.c |
| 7170 | 7171 | ||
| 7171 | PC87427 HARDWARE MONITORING DRIVER | 7172 | PC87427 HARDWARE MONITORING DRIVER |
| 7172 | M: Jean Delvare <jdelvare@suse.de> | 7173 | M: Jean Delvare <jdelvare@suse.de> |
| 7173 | L: lm-sensors@lm-sensors.org | 7174 | L: lm-sensors@lm-sensors.org |
| 7174 | S: Maintained | 7175 | S: Maintained |
| 7175 | F: Documentation/hwmon/pc87427 | 7176 | F: Documentation/hwmon/pc87427 |
| 7176 | F: drivers/hwmon/pc87427.c | 7177 | F: drivers/hwmon/pc87427.c |
| 7177 | 7178 | ||
| 7178 | PCA9532 LED DRIVER | 7179 | PCA9532 LED DRIVER |
| 7179 | M: Riku Voipio <riku.voipio@iki.fi> | 7180 | M: Riku Voipio <riku.voipio@iki.fi> |
| 7180 | S: Maintained | 7181 | S: Maintained |
| 7181 | F: drivers/leds/leds-pca9532.c | 7182 | F: drivers/leds/leds-pca9532.c |
| 7182 | F: include/linux/leds-pca9532.h | 7183 | F: include/linux/leds-pca9532.h |
| 7183 | 7184 | ||
| 7184 | PCA9541 I2C BUS MASTER SELECTOR DRIVER | 7185 | PCA9541 I2C BUS MASTER SELECTOR DRIVER |
| 7185 | M: Guenter Roeck <linux@roeck-us.net> | 7186 | M: Guenter Roeck <linux@roeck-us.net> |
| 7186 | L: linux-i2c@vger.kernel.org | 7187 | L: linux-i2c@vger.kernel.org |
| 7187 | S: Maintained | 7188 | S: Maintained |
| 7188 | F: drivers/i2c/muxes/i2c-mux-pca9541.c | 7189 | F: drivers/i2c/muxes/i2c-mux-pca9541.c |
| 7189 | 7190 | ||
| 7190 | PCDP - PRIMARY CONSOLE AND DEBUG PORT | 7191 | PCDP - PRIMARY CONSOLE AND DEBUG PORT |
| 7191 | M: Khalid Aziz <khalid@gonehiking.org> | 7192 | M: Khalid Aziz <khalid@gonehiking.org> |
| 7192 | S: Maintained | 7193 | S: Maintained |
| 7193 | F: drivers/firmware/pcdp.* | 7194 | F: drivers/firmware/pcdp.* |
| 7194 | 7195 | ||
| 7195 | PCI ERROR RECOVERY | 7196 | PCI ERROR RECOVERY |
| 7196 | M: Linas Vepstas <linasvepstas@gmail.com> | 7197 | M: Linas Vepstas <linasvepstas@gmail.com> |
| 7197 | L: linux-pci@vger.kernel.org | 7198 | L: linux-pci@vger.kernel.org |
| 7198 | S: Supported | 7199 | S: Supported |
| 7199 | F: Documentation/PCI/pci-error-recovery.txt | 7200 | F: Documentation/PCI/pci-error-recovery.txt |
| 7200 | 7201 | ||
| 7201 | PCI SUBSYSTEM | 7202 | PCI SUBSYSTEM |
| 7202 | M: Bjorn Helgaas <bhelgaas@google.com> | 7203 | M: Bjorn Helgaas <bhelgaas@google.com> |
| 7203 | L: linux-pci@vger.kernel.org | 7204 | L: linux-pci@vger.kernel.org |
| 7204 | Q: http://patchwork.ozlabs.org/project/linux-pci/list/ | 7205 | Q: http://patchwork.ozlabs.org/project/linux-pci/list/ |
| 7205 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci.git | 7206 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci.git |
| 7206 | S: Supported | 7207 | S: Supported |
| 7207 | F: Documentation/PCI/ | 7208 | F: Documentation/PCI/ |
| 7208 | F: drivers/pci/ | 7209 | F: drivers/pci/ |
| 7209 | F: include/linux/pci* | 7210 | F: include/linux/pci* |
| 7210 | F: arch/x86/pci/ | 7211 | F: arch/x86/pci/ |
| 7211 | F: arch/x86/kernel/quirks.c | 7212 | F: arch/x86/kernel/quirks.c |
| 7212 | 7213 | ||
| 7213 | PCI DRIVER FOR APPLIEDMICRO XGENE | 7214 | PCI DRIVER FOR APPLIEDMICRO XGENE |
| 7214 | M: Tanmay Inamdar <tinamdar@apm.com> | 7215 | M: Tanmay Inamdar <tinamdar@apm.com> |
| 7215 | L: linux-pci@vger.kernel.org | 7216 | L: linux-pci@vger.kernel.org |
| 7216 | L: linux-arm-kernel@lists.infradead.org | 7217 | L: linux-arm-kernel@lists.infradead.org |
| 7217 | S: Maintained | 7218 | S: Maintained |
| 7218 | F: Documentation/devicetree/bindings/pci/xgene-pci.txt | 7219 | F: Documentation/devicetree/bindings/pci/xgene-pci.txt |
| 7219 | F: drivers/pci/host/pci-xgene.c | 7220 | F: drivers/pci/host/pci-xgene.c |
| 7220 | 7221 | ||
| 7221 | PCI DRIVER FOR FREESCALE LAYERSCAPE | 7222 | PCI DRIVER FOR FREESCALE LAYERSCAPE |
| 7222 | M: Minghuan Lian <minghuan.Lian@freescale.com> | 7223 | M: Minghuan Lian <minghuan.Lian@freescale.com> |
| 7223 | M: Mingkai Hu <mingkai.hu@freescale.com> | 7224 | M: Mingkai Hu <mingkai.hu@freescale.com> |
| 7224 | M: Roy Zang <tie-fei.zang@freescale.com> | 7225 | M: Roy Zang <tie-fei.zang@freescale.com> |
| 7225 | L: linuxppc-dev@lists.ozlabs.org | 7226 | L: linuxppc-dev@lists.ozlabs.org |
| 7226 | L: linux-pci@vger.kernel.org | 7227 | L: linux-pci@vger.kernel.org |
| 7227 | L: linux-arm-kernel@lists.infradead.org | 7228 | L: linux-arm-kernel@lists.infradead.org |
| 7228 | S: Maintained | 7229 | S: Maintained |
| 7229 | F: drivers/pci/host/*layerscape* | 7230 | F: drivers/pci/host/*layerscape* |
| 7230 | 7231 | ||
| 7231 | PCI DRIVER FOR IMX6 | 7232 | PCI DRIVER FOR IMX6 |
| 7232 | M: Richard Zhu <r65037@freescale.com> | 7233 | M: Richard Zhu <r65037@freescale.com> |
| 7233 | M: Lucas Stach <l.stach@pengutronix.de> | 7234 | M: Lucas Stach <l.stach@pengutronix.de> |
| 7234 | L: linux-pci@vger.kernel.org | 7235 | L: linux-pci@vger.kernel.org |
| 7235 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7236 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7236 | S: Maintained | 7237 | S: Maintained |
| 7237 | F: drivers/pci/host/*imx6* | 7238 | F: drivers/pci/host/*imx6* |
| 7238 | 7239 | ||
| 7239 | PCI DRIVER FOR TI KEYSTONE | 7240 | PCI DRIVER FOR TI KEYSTONE |
| 7240 | M: Murali Karicheri <m-karicheri2@ti.com> | 7241 | M: Murali Karicheri <m-karicheri2@ti.com> |
| 7241 | L: linux-pci@vger.kernel.org | 7242 | L: linux-pci@vger.kernel.org |
| 7242 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7243 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7243 | S: Maintained | 7244 | S: Maintained |
| 7244 | F: drivers/pci/host/*keystone* | 7245 | F: drivers/pci/host/*keystone* |
| 7245 | 7246 | ||
| 7246 | PCI DRIVER FOR MVEBU (Marvell Armada 370 and Armada XP SOC support) | 7247 | PCI DRIVER FOR MVEBU (Marvell Armada 370 and Armada XP SOC support) |
| 7247 | M: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> | 7248 | M: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> |
| 7248 | M: Jason Cooper <jason@lakedaemon.net> | 7249 | M: Jason Cooper <jason@lakedaemon.net> |
| 7249 | L: linux-pci@vger.kernel.org | 7250 | L: linux-pci@vger.kernel.org |
| 7250 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7251 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7251 | S: Maintained | 7252 | S: Maintained |
| 7252 | F: drivers/pci/host/*mvebu* | 7253 | F: drivers/pci/host/*mvebu* |
| 7253 | 7254 | ||
| 7254 | PCI DRIVER FOR NVIDIA TEGRA | 7255 | PCI DRIVER FOR NVIDIA TEGRA |
| 7255 | M: Thierry Reding <thierry.reding@gmail.com> | 7256 | M: Thierry Reding <thierry.reding@gmail.com> |
| 7256 | L: linux-tegra@vger.kernel.org | 7257 | L: linux-tegra@vger.kernel.org |
| 7257 | L: linux-pci@vger.kernel.org | 7258 | L: linux-pci@vger.kernel.org |
| 7258 | S: Supported | 7259 | S: Supported |
| 7259 | F: Documentation/devicetree/bindings/pci/nvidia,tegra20-pcie.txt | 7260 | F: Documentation/devicetree/bindings/pci/nvidia,tegra20-pcie.txt |
| 7260 | F: drivers/pci/host/pci-tegra.c | 7261 | F: drivers/pci/host/pci-tegra.c |
| 7261 | 7262 | ||
| 7262 | PCI DRIVER FOR TI DRA7XX | 7263 | PCI DRIVER FOR TI DRA7XX |
| 7263 | M: Kishon Vijay Abraham I <kishon@ti.com> | 7264 | M: Kishon Vijay Abraham I <kishon@ti.com> |
| 7264 | L: linux-omap@vger.kernel.org | 7265 | L: linux-omap@vger.kernel.org |
| 7265 | L: linux-pci@vger.kernel.org | 7266 | L: linux-pci@vger.kernel.org |
| 7266 | S: Supported | 7267 | S: Supported |
| 7267 | F: Documentation/devicetree/bindings/pci/ti-pci.txt | 7268 | F: Documentation/devicetree/bindings/pci/ti-pci.txt |
| 7268 | F: drivers/pci/host/pci-dra7xx.c | 7269 | F: drivers/pci/host/pci-dra7xx.c |
| 7269 | 7270 | ||
| 7270 | PCI DRIVER FOR RENESAS R-CAR | 7271 | PCI DRIVER FOR RENESAS R-CAR |
| 7271 | M: Simon Horman <horms@verge.net.au> | 7272 | M: Simon Horman <horms@verge.net.au> |
| 7272 | L: linux-pci@vger.kernel.org | 7273 | L: linux-pci@vger.kernel.org |
| 7273 | L: linux-sh@vger.kernel.org | 7274 | L: linux-sh@vger.kernel.org |
| 7274 | S: Maintained | 7275 | S: Maintained |
| 7275 | F: drivers/pci/host/*rcar* | 7276 | F: drivers/pci/host/*rcar* |
| 7276 | 7277 | ||
| 7277 | PCI DRIVER FOR SAMSUNG EXYNOS | 7278 | PCI DRIVER FOR SAMSUNG EXYNOS |
| 7278 | M: Jingoo Han <jg1.han@samsung.com> | 7279 | M: Jingoo Han <jg1.han@samsung.com> |
| 7279 | L: linux-pci@vger.kernel.org | 7280 | L: linux-pci@vger.kernel.org |
| 7280 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7281 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7281 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) | 7282 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) |
| 7282 | S: Maintained | 7283 | S: Maintained |
| 7283 | F: drivers/pci/host/pci-exynos.c | 7284 | F: drivers/pci/host/pci-exynos.c |
| 7284 | 7285 | ||
| 7285 | PCI DRIVER FOR SYNOPSIS DESIGNWARE | 7286 | PCI DRIVER FOR SYNOPSIS DESIGNWARE |
| 7286 | M: Mohit Kumar <mohit.kumar@st.com> | 7287 | M: Mohit Kumar <mohit.kumar@st.com> |
| 7287 | M: Jingoo Han <jg1.han@samsung.com> | 7288 | M: Jingoo Han <jg1.han@samsung.com> |
| 7288 | L: linux-pci@vger.kernel.org | 7289 | L: linux-pci@vger.kernel.org |
| 7289 | S: Maintained | 7290 | S: Maintained |
| 7290 | F: drivers/pci/host/*designware* | 7291 | F: drivers/pci/host/*designware* |
| 7291 | 7292 | ||
| 7292 | PCI DRIVER FOR GENERIC OF HOSTS | 7293 | PCI DRIVER FOR GENERIC OF HOSTS |
| 7293 | M: Will Deacon <will.deacon@arm.com> | 7294 | M: Will Deacon <will.deacon@arm.com> |
| 7294 | L: linux-pci@vger.kernel.org | 7295 | L: linux-pci@vger.kernel.org |
| 7295 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7296 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7296 | S: Maintained | 7297 | S: Maintained |
| 7297 | F: Documentation/devicetree/bindings/pci/host-generic-pci.txt | 7298 | F: Documentation/devicetree/bindings/pci/host-generic-pci.txt |
| 7298 | F: drivers/pci/host/pci-host-generic.c | 7299 | F: drivers/pci/host/pci-host-generic.c |
| 7299 | 7300 | ||
| 7300 | PCIE DRIVER FOR ST SPEAR13XX | 7301 | PCIE DRIVER FOR ST SPEAR13XX |
| 7301 | M: Mohit Kumar <mohit.kumar@st.com> | 7302 | M: Mohit Kumar <mohit.kumar@st.com> |
| 7302 | L: linux-pci@vger.kernel.org | 7303 | L: linux-pci@vger.kernel.org |
| 7303 | S: Maintained | 7304 | S: Maintained |
| 7304 | F: drivers/pci/host/*spear* | 7305 | F: drivers/pci/host/*spear* |
| 7305 | 7306 | ||
| 7306 | PCMCIA SUBSYSTEM | 7307 | PCMCIA SUBSYSTEM |
| 7307 | P: Linux PCMCIA Team | 7308 | P: Linux PCMCIA Team |
| 7308 | L: linux-pcmcia@lists.infradead.org | 7309 | L: linux-pcmcia@lists.infradead.org |
| 7309 | W: http://lists.infradead.org/mailman/listinfo/linux-pcmcia | 7310 | W: http://lists.infradead.org/mailman/listinfo/linux-pcmcia |
| 7310 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6.git | 7311 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6.git |
| 7311 | S: Maintained | 7312 | S: Maintained |
| 7312 | F: Documentation/pcmcia/ | 7313 | F: Documentation/pcmcia/ |
| 7313 | F: drivers/pcmcia/ | 7314 | F: drivers/pcmcia/ |
| 7314 | F: include/pcmcia/ | 7315 | F: include/pcmcia/ |
| 7315 | 7316 | ||
| 7316 | PCNET32 NETWORK DRIVER | 7317 | PCNET32 NETWORK DRIVER |
| 7317 | M: Don Fry <pcnet32@frontier.com> | 7318 | M: Don Fry <pcnet32@frontier.com> |
| 7318 | L: netdev@vger.kernel.org | 7319 | L: netdev@vger.kernel.org |
| 7319 | S: Maintained | 7320 | S: Maintained |
| 7320 | F: drivers/net/ethernet/amd/pcnet32.c | 7321 | F: drivers/net/ethernet/amd/pcnet32.c |
| 7321 | 7322 | ||
| 7322 | PCRYPT PARALLEL CRYPTO ENGINE | 7323 | PCRYPT PARALLEL CRYPTO ENGINE |
| 7323 | M: Steffen Klassert <steffen.klassert@secunet.com> | 7324 | M: Steffen Klassert <steffen.klassert@secunet.com> |
| 7324 | L: linux-crypto@vger.kernel.org | 7325 | L: linux-crypto@vger.kernel.org |
| 7325 | S: Maintained | 7326 | S: Maintained |
| 7326 | F: crypto/pcrypt.c | 7327 | F: crypto/pcrypt.c |
| 7327 | F: include/crypto/pcrypt.h | 7328 | F: include/crypto/pcrypt.h |
| 7328 | 7329 | ||
| 7329 | PER-CPU MEMORY ALLOCATOR | 7330 | PER-CPU MEMORY ALLOCATOR |
| 7330 | M: Tejun Heo <tj@kernel.org> | 7331 | M: Tejun Heo <tj@kernel.org> |
| 7331 | M: Christoph Lameter <cl@linux-foundation.org> | 7332 | M: Christoph Lameter <cl@linux-foundation.org> |
| 7332 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git | 7333 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu.git |
| 7333 | S: Maintained | 7334 | S: Maintained |
| 7334 | F: include/linux/percpu*.h | 7335 | F: include/linux/percpu*.h |
| 7335 | F: mm/percpu*.c | 7336 | F: mm/percpu*.c |
| 7336 | F: arch/*/include/asm/percpu.h | 7337 | F: arch/*/include/asm/percpu.h |
| 7337 | 7338 | ||
| 7338 | PER-TASK DELAY ACCOUNTING | 7339 | PER-TASK DELAY ACCOUNTING |
| 7339 | M: Balbir Singh <bsingharora@gmail.com> | 7340 | M: Balbir Singh <bsingharora@gmail.com> |
| 7340 | S: Maintained | 7341 | S: Maintained |
| 7341 | F: include/linux/delayacct.h | 7342 | F: include/linux/delayacct.h |
| 7342 | F: kernel/delayacct.c | 7343 | F: kernel/delayacct.c |
| 7343 | 7344 | ||
| 7344 | PERFORMANCE EVENTS SUBSYSTEM | 7345 | PERFORMANCE EVENTS SUBSYSTEM |
| 7345 | M: Peter Zijlstra <a.p.zijlstra@chello.nl> | 7346 | M: Peter Zijlstra <a.p.zijlstra@chello.nl> |
| 7346 | M: Paul Mackerras <paulus@samba.org> | 7347 | M: Paul Mackerras <paulus@samba.org> |
| 7347 | M: Ingo Molnar <mingo@redhat.com> | 7348 | M: Ingo Molnar <mingo@redhat.com> |
| 7348 | M: Arnaldo Carvalho de Melo <acme@kernel.org> | 7349 | M: Arnaldo Carvalho de Melo <acme@kernel.org> |
| 7349 | L: linux-kernel@vger.kernel.org | 7350 | L: linux-kernel@vger.kernel.org |
| 7350 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core | 7351 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core |
| 7351 | S: Supported | 7352 | S: Supported |
| 7352 | F: kernel/events/* | 7353 | F: kernel/events/* |
| 7353 | F: include/linux/perf_event.h | 7354 | F: include/linux/perf_event.h |
| 7354 | F: include/uapi/linux/perf_event.h | 7355 | F: include/uapi/linux/perf_event.h |
| 7355 | F: arch/*/kernel/perf_event*.c | 7356 | F: arch/*/kernel/perf_event*.c |
| 7356 | F: arch/*/kernel/*/perf_event*.c | 7357 | F: arch/*/kernel/*/perf_event*.c |
| 7357 | F: arch/*/kernel/*/*/perf_event*.c | 7358 | F: arch/*/kernel/*/*/perf_event*.c |
| 7358 | F: arch/*/include/asm/perf_event.h | 7359 | F: arch/*/include/asm/perf_event.h |
| 7359 | F: arch/*/kernel/perf_callchain.c | 7360 | F: arch/*/kernel/perf_callchain.c |
| 7360 | F: tools/perf/ | 7361 | F: tools/perf/ |
| 7361 | 7362 | ||
| 7362 | PERSONALITY HANDLING | 7363 | PERSONALITY HANDLING |
| 7363 | M: Christoph Hellwig <hch@infradead.org> | 7364 | M: Christoph Hellwig <hch@infradead.org> |
| 7364 | L: linux-abi-devel@lists.sourceforge.net | 7365 | L: linux-abi-devel@lists.sourceforge.net |
| 7365 | S: Maintained | 7366 | S: Maintained |
| 7366 | F: include/linux/personality.h | 7367 | F: include/linux/personality.h |
| 7367 | F: include/uapi/linux/personality.h | 7368 | F: include/uapi/linux/personality.h |
| 7368 | 7369 | ||
| 7369 | PHONET PROTOCOL | 7370 | PHONET PROTOCOL |
| 7370 | M: Remi Denis-Courmont <courmisch@gmail.com> | 7371 | M: Remi Denis-Courmont <courmisch@gmail.com> |
| 7371 | S: Supported | 7372 | S: Supported |
| 7372 | F: Documentation/networking/phonet.txt | 7373 | F: Documentation/networking/phonet.txt |
| 7373 | F: include/linux/phonet.h | 7374 | F: include/linux/phonet.h |
| 7374 | F: include/net/phonet/ | 7375 | F: include/net/phonet/ |
| 7375 | F: include/uapi/linux/phonet.h | 7376 | F: include/uapi/linux/phonet.h |
| 7376 | F: net/phonet/ | 7377 | F: net/phonet/ |
| 7377 | 7378 | ||
| 7378 | PHRAM MTD DRIVER | 7379 | PHRAM MTD DRIVER |
| 7379 | M: Joern Engel <joern@lazybastard.org> | 7380 | M: Joern Engel <joern@lazybastard.org> |
| 7380 | L: linux-mtd@lists.infradead.org | 7381 | L: linux-mtd@lists.infradead.org |
| 7381 | S: Maintained | 7382 | S: Maintained |
| 7382 | F: drivers/mtd/devices/phram.c | 7383 | F: drivers/mtd/devices/phram.c |
| 7383 | 7384 | ||
| 7384 | PICOLCD HID DRIVER | 7385 | PICOLCD HID DRIVER |
| 7385 | M: Bruno Prรฉmont <bonbons@linux-vserver.org> | 7386 | M: Bruno Prรฉmont <bonbons@linux-vserver.org> |
| 7386 | L: linux-input@vger.kernel.org | 7387 | L: linux-input@vger.kernel.org |
| 7387 | S: Maintained | 7388 | S: Maintained |
| 7388 | F: drivers/hid/hid-picolcd* | 7389 | F: drivers/hid/hid-picolcd* |
| 7389 | 7390 | ||
| 7390 | PICOXCELL SUPPORT | 7391 | PICOXCELL SUPPORT |
| 7391 | M: Jamie Iles <jamie@jamieiles.com> | 7392 | M: Jamie Iles <jamie@jamieiles.com> |
| 7392 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7393 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7393 | T: git git://github.com/jamieiles/linux-2.6-ji.git | 7394 | T: git git://github.com/jamieiles/linux-2.6-ji.git |
| 7394 | S: Supported | 7395 | S: Supported |
| 7395 | F: arch/arm/boot/dts/picoxcell* | 7396 | F: arch/arm/boot/dts/picoxcell* |
| 7396 | F: arch/arm/mach-picoxcell/ | 7397 | F: arch/arm/mach-picoxcell/ |
| 7397 | F: drivers/crypto/picoxcell* | 7398 | F: drivers/crypto/picoxcell* |
| 7398 | 7399 | ||
| 7399 | PIN CONTROL SUBSYSTEM | 7400 | PIN CONTROL SUBSYSTEM |
| 7400 | M: Linus Walleij <linus.walleij@linaro.org> | 7401 | M: Linus Walleij <linus.walleij@linaro.org> |
| 7401 | L: linux-gpio@vger.kernel.org | 7402 | L: linux-gpio@vger.kernel.org |
| 7402 | S: Maintained | 7403 | S: Maintained |
| 7403 | F: drivers/pinctrl/ | 7404 | F: drivers/pinctrl/ |
| 7404 | F: include/linux/pinctrl/ | 7405 | F: include/linux/pinctrl/ |
| 7405 | 7406 | ||
| 7406 | PIN CONTROLLER - ATMEL AT91 | 7407 | PIN CONTROLLER - ATMEL AT91 |
| 7407 | M: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com> | 7408 | M: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com> |
| 7408 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7409 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7409 | S: Maintained | 7410 | S: Maintained |
| 7410 | F: drivers/pinctrl/pinctrl-at91.* | 7411 | F: drivers/pinctrl/pinctrl-at91.* |
| 7411 | 7412 | ||
| 7412 | PIN CONTROLLER - INTEL | 7413 | PIN CONTROLLER - INTEL |
| 7413 | M: Mika Westerberg <mika.westerberg@linux.intel.com> | 7414 | M: Mika Westerberg <mika.westerberg@linux.intel.com> |
| 7414 | M: Heikki Krogerus <heikki.krogerus@linux.intel.com> | 7415 | M: Heikki Krogerus <heikki.krogerus@linux.intel.com> |
| 7415 | S: Maintained | 7416 | S: Maintained |
| 7416 | F: drivers/pinctrl/intel/ | 7417 | F: drivers/pinctrl/intel/ |
| 7417 | 7418 | ||
| 7418 | PIN CONTROLLER - RENESAS | 7419 | PIN CONTROLLER - RENESAS |
| 7419 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 7420 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 7420 | L: linux-sh@vger.kernel.org | 7421 | L: linux-sh@vger.kernel.org |
| 7421 | S: Maintained | 7422 | S: Maintained |
| 7422 | F: drivers/pinctrl/sh-pfc/ | 7423 | F: drivers/pinctrl/sh-pfc/ |
| 7423 | 7424 | ||
| 7424 | PIN CONTROLLER - SAMSUNG | 7425 | PIN CONTROLLER - SAMSUNG |
| 7425 | M: Tomasz Figa <tomasz.figa@gmail.com> | 7426 | M: Tomasz Figa <tomasz.figa@gmail.com> |
| 7426 | M: Thomas Abraham <thomas.abraham@linaro.org> | 7427 | M: Thomas Abraham <thomas.abraham@linaro.org> |
| 7427 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7428 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7428 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) | 7429 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) |
| 7429 | S: Maintained | 7430 | S: Maintained |
| 7430 | F: drivers/pinctrl/samsung/ | 7431 | F: drivers/pinctrl/samsung/ |
| 7431 | 7432 | ||
| 7432 | PIN CONTROLLER - ST SPEAR | 7433 | PIN CONTROLLER - ST SPEAR |
| 7433 | M: Viresh Kumar <viresh.linux@gmail.com> | 7434 | M: Viresh Kumar <viresh.linux@gmail.com> |
| 7434 | L: spear-devel@list.st.com | 7435 | L: spear-devel@list.st.com |
| 7435 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7436 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7436 | W: http://www.st.com/spear | 7437 | W: http://www.st.com/spear |
| 7437 | S: Maintained | 7438 | S: Maintained |
| 7438 | F: drivers/pinctrl/spear/ | 7439 | F: drivers/pinctrl/spear/ |
| 7439 | 7440 | ||
| 7440 | PKTCDVD DRIVER | 7441 | PKTCDVD DRIVER |
| 7441 | M: Jiri Kosina <jkosina@suse.cz> | 7442 | M: Jiri Kosina <jkosina@suse.cz> |
| 7442 | S: Maintained | 7443 | S: Maintained |
| 7443 | F: drivers/block/pktcdvd.c | 7444 | F: drivers/block/pktcdvd.c |
| 7444 | F: include/linux/pktcdvd.h | 7445 | F: include/linux/pktcdvd.h |
| 7445 | F: include/uapi/linux/pktcdvd.h | 7446 | F: include/uapi/linux/pktcdvd.h |
| 7446 | 7447 | ||
| 7447 | PKUNITY SOC DRIVERS | 7448 | PKUNITY SOC DRIVERS |
| 7448 | M: Guan Xuetao <gxt@mprc.pku.edu.cn> | 7449 | M: Guan Xuetao <gxt@mprc.pku.edu.cn> |
| 7449 | W: http://mprc.pku.edu.cn/~guanxuetao/linux | 7450 | W: http://mprc.pku.edu.cn/~guanxuetao/linux |
| 7450 | S: Maintained | 7451 | S: Maintained |
| 7451 | T: git git://github.com/gxt/linux.git | 7452 | T: git git://github.com/gxt/linux.git |
| 7452 | F: drivers/input/serio/i8042-unicore32io.h | 7453 | F: drivers/input/serio/i8042-unicore32io.h |
| 7453 | F: drivers/i2c/busses/i2c-puv3.c | 7454 | F: drivers/i2c/busses/i2c-puv3.c |
| 7454 | F: drivers/video/fbdev/fb-puv3.c | 7455 | F: drivers/video/fbdev/fb-puv3.c |
| 7455 | F: drivers/rtc/rtc-puv3.c | 7456 | F: drivers/rtc/rtc-puv3.c |
| 7456 | 7457 | ||
| 7457 | PMBUS HARDWARE MONITORING DRIVERS | 7458 | PMBUS HARDWARE MONITORING DRIVERS |
| 7458 | M: Guenter Roeck <linux@roeck-us.net> | 7459 | M: Guenter Roeck <linux@roeck-us.net> |
| 7459 | L: lm-sensors@lm-sensors.org | 7460 | L: lm-sensors@lm-sensors.org |
| 7460 | W: http://www.lm-sensors.org/ | 7461 | W: http://www.lm-sensors.org/ |
| 7461 | W: http://www.roeck-us.net/linux/drivers/ | 7462 | W: http://www.roeck-us.net/linux/drivers/ |
| 7462 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging.git | 7463 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging.git |
| 7463 | S: Maintained | 7464 | S: Maintained |
| 7464 | F: Documentation/hwmon/pmbus | 7465 | F: Documentation/hwmon/pmbus |
| 7465 | F: drivers/hwmon/pmbus/ | 7466 | F: drivers/hwmon/pmbus/ |
| 7466 | F: include/linux/i2c/pmbus.h | 7467 | F: include/linux/i2c/pmbus.h |
| 7467 | 7468 | ||
| 7468 | PMC SIERRA MaxRAID DRIVER | 7469 | PMC SIERRA MaxRAID DRIVER |
| 7469 | M: Anil Ravindranath <anil_ravindranath@pmc-sierra.com> | 7470 | M: Anil Ravindranath <anil_ravindranath@pmc-sierra.com> |
| 7470 | L: linux-scsi@vger.kernel.org | 7471 | L: linux-scsi@vger.kernel.org |
| 7471 | W: http://www.pmc-sierra.com/ | 7472 | W: http://www.pmc-sierra.com/ |
| 7472 | S: Supported | 7473 | S: Supported |
| 7473 | F: drivers/scsi/pmcraid.* | 7474 | F: drivers/scsi/pmcraid.* |
| 7474 | 7475 | ||
| 7475 | PMC SIERRA PM8001 DRIVER | 7476 | PMC SIERRA PM8001 DRIVER |
| 7476 | M: xjtuwjp@gmail.com | 7477 | M: xjtuwjp@gmail.com |
| 7477 | M: lindar_liu@usish.com | 7478 | M: lindar_liu@usish.com |
| 7478 | L: pmchba@pmcs.com | 7479 | L: pmchba@pmcs.com |
| 7479 | L: linux-scsi@vger.kernel.org | 7480 | L: linux-scsi@vger.kernel.org |
| 7480 | S: Supported | 7481 | S: Supported |
| 7481 | F: drivers/scsi/pm8001/ | 7482 | F: drivers/scsi/pm8001/ |
| 7482 | 7483 | ||
| 7483 | POSIX CLOCKS and TIMERS | 7484 | POSIX CLOCKS and TIMERS |
| 7484 | M: Thomas Gleixner <tglx@linutronix.de> | 7485 | M: Thomas Gleixner <tglx@linutronix.de> |
| 7485 | L: linux-kernel@vger.kernel.org | 7486 | L: linux-kernel@vger.kernel.org |
| 7486 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core | 7487 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core |
| 7487 | S: Maintained | 7488 | S: Maintained |
| 7488 | F: fs/timerfd.c | 7489 | F: fs/timerfd.c |
| 7489 | F: include/linux/timer* | 7490 | F: include/linux/timer* |
| 7490 | F: kernel/time/*timer* | 7491 | F: kernel/time/*timer* |
| 7491 | 7492 | ||
| 7492 | POWER SUPPLY CLASS/SUBSYSTEM and DRIVERS | 7493 | POWER SUPPLY CLASS/SUBSYSTEM and DRIVERS |
| 7493 | M: Sebastian Reichel <sre@kernel.org> | 7494 | M: Sebastian Reichel <sre@kernel.org> |
| 7494 | M: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> | 7495 | M: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> |
| 7495 | M: David Woodhouse <dwmw2@infradead.org> | 7496 | M: David Woodhouse <dwmw2@infradead.org> |
| 7496 | L: linux-pm@vger.kernel.org | 7497 | L: linux-pm@vger.kernel.org |
| 7497 | T: git git://git.infradead.org/battery-2.6.git | 7498 | T: git git://git.infradead.org/battery-2.6.git |
| 7498 | S: Maintained | 7499 | S: Maintained |
| 7499 | F: include/linux/power_supply.h | 7500 | F: include/linux/power_supply.h |
| 7500 | F: drivers/power/ | 7501 | F: drivers/power/ |
| 7501 | 7502 | ||
| 7502 | PNP SUPPORT | 7503 | PNP SUPPORT |
| 7503 | M: Rafael J. Wysocki <rafael.j.wysocki@intel.com> | 7504 | M: Rafael J. Wysocki <rafael.j.wysocki@intel.com> |
| 7504 | S: Maintained | 7505 | S: Maintained |
| 7505 | F: drivers/pnp/ | 7506 | F: drivers/pnp/ |
| 7506 | 7507 | ||
| 7507 | PNXxxxx I2C DRIVER | 7508 | PNXxxxx I2C DRIVER |
| 7508 | M: Vitaly Wool <vitalywool@gmail.com> | 7509 | M: Vitaly Wool <vitalywool@gmail.com> |
| 7509 | L: linux-i2c@vger.kernel.org | 7510 | L: linux-i2c@vger.kernel.org |
| 7510 | S: Maintained | 7511 | S: Maintained |
| 7511 | F: drivers/i2c/busses/i2c-pnx.c | 7512 | F: drivers/i2c/busses/i2c-pnx.c |
| 7512 | 7513 | ||
| 7513 | PPP PROTOCOL DRIVERS AND COMPRESSORS | 7514 | PPP PROTOCOL DRIVERS AND COMPRESSORS |
| 7514 | M: Paul Mackerras <paulus@samba.org> | 7515 | M: Paul Mackerras <paulus@samba.org> |
| 7515 | L: linux-ppp@vger.kernel.org | 7516 | L: linux-ppp@vger.kernel.org |
| 7516 | S: Maintained | 7517 | S: Maintained |
| 7517 | F: drivers/net/ppp/ppp_* | 7518 | F: drivers/net/ppp/ppp_* |
| 7518 | 7519 | ||
| 7519 | PPP OVER ATM (RFC 2364) | 7520 | PPP OVER ATM (RFC 2364) |
| 7520 | M: Mitchell Blank Jr <mitch@sfgoth.com> | 7521 | M: Mitchell Blank Jr <mitch@sfgoth.com> |
| 7521 | S: Maintained | 7522 | S: Maintained |
| 7522 | F: net/atm/pppoatm.c | 7523 | F: net/atm/pppoatm.c |
| 7523 | F: include/uapi/linux/atmppp.h | 7524 | F: include/uapi/linux/atmppp.h |
| 7524 | 7525 | ||
| 7525 | PPP OVER ETHERNET | 7526 | PPP OVER ETHERNET |
| 7526 | M: Michal Ostrowski <mostrows@earthlink.net> | 7527 | M: Michal Ostrowski <mostrows@earthlink.net> |
| 7527 | S: Maintained | 7528 | S: Maintained |
| 7528 | F: drivers/net/ppp/pppoe.c | 7529 | F: drivers/net/ppp/pppoe.c |
| 7529 | F: drivers/net/ppp/pppox.c | 7530 | F: drivers/net/ppp/pppox.c |
| 7530 | 7531 | ||
| 7531 | PPP OVER L2TP | 7532 | PPP OVER L2TP |
| 7532 | M: James Chapman <jchapman@katalix.com> | 7533 | M: James Chapman <jchapman@katalix.com> |
| 7533 | S: Maintained | 7534 | S: Maintained |
| 7534 | F: net/l2tp/l2tp_ppp.c | 7535 | F: net/l2tp/l2tp_ppp.c |
| 7535 | F: include/linux/if_pppol2tp.h | 7536 | F: include/linux/if_pppol2tp.h |
| 7536 | F: include/uapi/linux/if_pppol2tp.h | 7537 | F: include/uapi/linux/if_pppol2tp.h |
| 7537 | 7538 | ||
| 7538 | PPS SUPPORT | 7539 | PPS SUPPORT |
| 7539 | M: Rodolfo Giometti <giometti@enneenne.com> | 7540 | M: Rodolfo Giometti <giometti@enneenne.com> |
| 7540 | W: http://wiki.enneenne.com/index.php/LinuxPPS_support | 7541 | W: http://wiki.enneenne.com/index.php/LinuxPPS_support |
| 7541 | L: linuxpps@ml.enneenne.com (subscribers-only) | 7542 | L: linuxpps@ml.enneenne.com (subscribers-only) |
| 7542 | S: Maintained | 7543 | S: Maintained |
| 7543 | F: Documentation/pps/ | 7544 | F: Documentation/pps/ |
| 7544 | F: drivers/pps/ | 7545 | F: drivers/pps/ |
| 7545 | F: include/linux/pps*.h | 7546 | F: include/linux/pps*.h |
| 7546 | 7547 | ||
| 7547 | PPTP DRIVER | 7548 | PPTP DRIVER |
| 7548 | M: Dmitry Kozlov <xeb@mail.ru> | 7549 | M: Dmitry Kozlov <xeb@mail.ru> |
| 7549 | L: netdev@vger.kernel.org | 7550 | L: netdev@vger.kernel.org |
| 7550 | S: Maintained | 7551 | S: Maintained |
| 7551 | F: drivers/net/ppp/pptp.c | 7552 | F: drivers/net/ppp/pptp.c |
| 7552 | W: http://sourceforge.net/projects/accel-pptp | 7553 | W: http://sourceforge.net/projects/accel-pptp |
| 7553 | 7554 | ||
| 7554 | PREEMPTIBLE KERNEL | 7555 | PREEMPTIBLE KERNEL |
| 7555 | M: Robert Love <rml@tech9.net> | 7556 | M: Robert Love <rml@tech9.net> |
| 7556 | L: kpreempt-tech@lists.sourceforge.net | 7557 | L: kpreempt-tech@lists.sourceforge.net |
| 7557 | W: ftp://ftp.kernel.org/pub/linux/kernel/people/rml/preempt-kernel | 7558 | W: ftp://ftp.kernel.org/pub/linux/kernel/people/rml/preempt-kernel |
| 7558 | S: Supported | 7559 | S: Supported |
| 7559 | F: Documentation/preempt-locking.txt | 7560 | F: Documentation/preempt-locking.txt |
| 7560 | F: include/linux/preempt.h | 7561 | F: include/linux/preempt.h |
| 7561 | 7562 | ||
| 7562 | PRISM54 WIRELESS DRIVER | 7563 | PRISM54 WIRELESS DRIVER |
| 7563 | M: "Luis R. Rodriguez" <mcgrof@gmail.com> | 7564 | M: "Luis R. Rodriguez" <mcgrof@gmail.com> |
| 7564 | L: linux-wireless@vger.kernel.org | 7565 | L: linux-wireless@vger.kernel.org |
| 7565 | W: http://wireless.kernel.org/en/users/Drivers/p54 | 7566 | W: http://wireless.kernel.org/en/users/Drivers/p54 |
| 7566 | S: Obsolete | 7567 | S: Obsolete |
| 7567 | F: drivers/net/wireless/prism54/ | 7568 | F: drivers/net/wireless/prism54/ |
| 7568 | 7569 | ||
| 7569 | PROMISE SATA TX2/TX4 CONTROLLER LIBATA DRIVER | 7570 | PROMISE SATA TX2/TX4 CONTROLLER LIBATA DRIVER |
| 7570 | M: Mikael Pettersson <mikpelinux@gmail.com> | 7571 | M: Mikael Pettersson <mikpelinux@gmail.com> |
| 7571 | L: linux-ide@vger.kernel.org | 7572 | L: linux-ide@vger.kernel.org |
| 7572 | S: Maintained | 7573 | S: Maintained |
| 7573 | F: drivers/ata/sata_promise.* | 7574 | F: drivers/ata/sata_promise.* |
| 7574 | 7575 | ||
| 7575 | PS3 NETWORK SUPPORT | 7576 | PS3 NETWORK SUPPORT |
| 7576 | M: Geoff Levand <geoff@infradead.org> | 7577 | M: Geoff Levand <geoff@infradead.org> |
| 7577 | L: netdev@vger.kernel.org | 7578 | L: netdev@vger.kernel.org |
| 7578 | L: cbe-oss-dev@lists.ozlabs.org | 7579 | L: cbe-oss-dev@lists.ozlabs.org |
| 7579 | S: Maintained | 7580 | S: Maintained |
| 7580 | F: drivers/net/ethernet/toshiba/ps3_gelic_net.* | 7581 | F: drivers/net/ethernet/toshiba/ps3_gelic_net.* |
| 7581 | 7582 | ||
| 7582 | PS3 PLATFORM SUPPORT | 7583 | PS3 PLATFORM SUPPORT |
| 7583 | M: Geoff Levand <geoff@infradead.org> | 7584 | M: Geoff Levand <geoff@infradead.org> |
| 7584 | L: linuxppc-dev@lists.ozlabs.org | 7585 | L: linuxppc-dev@lists.ozlabs.org |
| 7585 | L: cbe-oss-dev@lists.ozlabs.org | 7586 | L: cbe-oss-dev@lists.ozlabs.org |
| 7586 | S: Maintained | 7587 | S: Maintained |
| 7587 | F: arch/powerpc/boot/ps3* | 7588 | F: arch/powerpc/boot/ps3* |
| 7588 | F: arch/powerpc/include/asm/lv1call.h | 7589 | F: arch/powerpc/include/asm/lv1call.h |
| 7589 | F: arch/powerpc/include/asm/ps3*.h | 7590 | F: arch/powerpc/include/asm/ps3*.h |
| 7590 | F: arch/powerpc/platforms/ps3/ | 7591 | F: arch/powerpc/platforms/ps3/ |
| 7591 | F: drivers/*/ps3* | 7592 | F: drivers/*/ps3* |
| 7592 | F: drivers/ps3/ | 7593 | F: drivers/ps3/ |
| 7593 | F: drivers/rtc/rtc-ps3.c | 7594 | F: drivers/rtc/rtc-ps3.c |
| 7594 | F: drivers/usb/host/*ps3.c | 7595 | F: drivers/usb/host/*ps3.c |
| 7595 | F: sound/ppc/snd_ps3* | 7596 | F: sound/ppc/snd_ps3* |
| 7596 | 7597 | ||
| 7597 | PS3VRAM DRIVER | 7598 | PS3VRAM DRIVER |
| 7598 | M: Jim Paris <jim@jtan.com> | 7599 | M: Jim Paris <jim@jtan.com> |
| 7599 | L: cbe-oss-dev@lists.ozlabs.org | 7600 | L: cbe-oss-dev@lists.ozlabs.org |
| 7600 | S: Maintained | 7601 | S: Maintained |
| 7601 | F: drivers/block/ps3vram.c | 7602 | F: drivers/block/ps3vram.c |
| 7602 | 7603 | ||
| 7603 | PSTORE FILESYSTEM | 7604 | PSTORE FILESYSTEM |
| 7604 | M: Anton Vorontsov <anton@enomsg.org> | 7605 | M: Anton Vorontsov <anton@enomsg.org> |
| 7605 | M: Colin Cross <ccross@android.com> | 7606 | M: Colin Cross <ccross@android.com> |
| 7606 | M: Kees Cook <keescook@chromium.org> | 7607 | M: Kees Cook <keescook@chromium.org> |
| 7607 | M: Tony Luck <tony.luck@intel.com> | 7608 | M: Tony Luck <tony.luck@intel.com> |
| 7608 | S: Maintained | 7609 | S: Maintained |
| 7609 | T: git git://git.infradead.org/users/cbou/linux-pstore.git | 7610 | T: git git://git.infradead.org/users/cbou/linux-pstore.git |
| 7610 | F: fs/pstore/ | 7611 | F: fs/pstore/ |
| 7611 | F: include/linux/pstore* | 7612 | F: include/linux/pstore* |
| 7612 | F: drivers/firmware/efi/efi-pstore.c | 7613 | F: drivers/firmware/efi/efi-pstore.c |
| 7613 | F: drivers/acpi/apei/erst.c | 7614 | F: drivers/acpi/apei/erst.c |
| 7614 | 7615 | ||
| 7615 | PTP HARDWARE CLOCK SUPPORT | 7616 | PTP HARDWARE CLOCK SUPPORT |
| 7616 | M: Richard Cochran <richardcochran@gmail.com> | 7617 | M: Richard Cochran <richardcochran@gmail.com> |
| 7617 | L: netdev@vger.kernel.org | 7618 | L: netdev@vger.kernel.org |
| 7618 | S: Maintained | 7619 | S: Maintained |
| 7619 | W: http://linuxptp.sourceforge.net/ | 7620 | W: http://linuxptp.sourceforge.net/ |
| 7620 | F: Documentation/ABI/testing/sysfs-ptp | 7621 | F: Documentation/ABI/testing/sysfs-ptp |
| 7621 | F: Documentation/ptp/* | 7622 | F: Documentation/ptp/* |
| 7622 | F: drivers/net/ethernet/freescale/gianfar_ptp.c | 7623 | F: drivers/net/ethernet/freescale/gianfar_ptp.c |
| 7623 | F: drivers/net/phy/dp83640* | 7624 | F: drivers/net/phy/dp83640* |
| 7624 | F: drivers/ptp/* | 7625 | F: drivers/ptp/* |
| 7625 | F: include/linux/ptp_cl* | 7626 | F: include/linux/ptp_cl* |
| 7626 | 7627 | ||
| 7627 | PTRACE SUPPORT | 7628 | PTRACE SUPPORT |
| 7628 | M: Roland McGrath <roland@hack.frob.com> | 7629 | M: Roland McGrath <roland@hack.frob.com> |
| 7629 | M: Oleg Nesterov <oleg@redhat.com> | 7630 | M: Oleg Nesterov <oleg@redhat.com> |
| 7630 | S: Maintained | 7631 | S: Maintained |
| 7631 | F: include/asm-generic/syscall.h | 7632 | F: include/asm-generic/syscall.h |
| 7632 | F: include/linux/ptrace.h | 7633 | F: include/linux/ptrace.h |
| 7633 | F: include/linux/regset.h | 7634 | F: include/linux/regset.h |
| 7634 | F: include/linux/tracehook.h | 7635 | F: include/linux/tracehook.h |
| 7635 | F: include/uapi/linux/ptrace.h | 7636 | F: include/uapi/linux/ptrace.h |
| 7636 | F: kernel/ptrace.c | 7637 | F: kernel/ptrace.c |
| 7637 | 7638 | ||
| 7638 | PVRUSB2 VIDEO4LINUX DRIVER | 7639 | PVRUSB2 VIDEO4LINUX DRIVER |
| 7639 | M: Mike Isely <isely@pobox.com> | 7640 | M: Mike Isely <isely@pobox.com> |
| 7640 | L: pvrusb2@isely.net (subscribers-only) | 7641 | L: pvrusb2@isely.net (subscribers-only) |
| 7641 | L: linux-media@vger.kernel.org | 7642 | L: linux-media@vger.kernel.org |
| 7642 | W: http://www.isely.net/pvrusb2/ | 7643 | W: http://www.isely.net/pvrusb2/ |
| 7643 | T: git git://linuxtv.org/media_tree.git | 7644 | T: git git://linuxtv.org/media_tree.git |
| 7644 | S: Maintained | 7645 | S: Maintained |
| 7645 | F: Documentation/video4linux/README.pvrusb2 | 7646 | F: Documentation/video4linux/README.pvrusb2 |
| 7646 | F: drivers/media/usb/pvrusb2/ | 7647 | F: drivers/media/usb/pvrusb2/ |
| 7647 | 7648 | ||
| 7648 | PWC WEBCAM DRIVER | 7649 | PWC WEBCAM DRIVER |
| 7649 | M: Hans de Goede <hdegoede@redhat.com> | 7650 | M: Hans de Goede <hdegoede@redhat.com> |
| 7650 | L: linux-media@vger.kernel.org | 7651 | L: linux-media@vger.kernel.org |
| 7651 | T: git git://linuxtv.org/media_tree.git | 7652 | T: git git://linuxtv.org/media_tree.git |
| 7652 | S: Maintained | 7653 | S: Maintained |
| 7653 | F: drivers/media/usb/pwc/* | 7654 | F: drivers/media/usb/pwc/* |
| 7654 | 7655 | ||
| 7655 | PWM FAN DRIVER | 7656 | PWM FAN DRIVER |
| 7656 | M: Kamil Debski <k.debski@samsung.com> | 7657 | M: Kamil Debski <k.debski@samsung.com> |
| 7657 | L: lm-sensors@lm-sensors.org | 7658 | L: lm-sensors@lm-sensors.org |
| 7658 | S: Supported | 7659 | S: Supported |
| 7659 | F: Documentation/devicetree/bindings/hwmon/pwm-fan.txt | 7660 | F: Documentation/devicetree/bindings/hwmon/pwm-fan.txt |
| 7660 | F: Documentation/hwmon/pwm-fan | 7661 | F: Documentation/hwmon/pwm-fan |
| 7661 | F: drivers/hwmon/pwm-fan.c | 7662 | F: drivers/hwmon/pwm-fan.c |
| 7662 | 7663 | ||
| 7663 | PWM SUBSYSTEM | 7664 | PWM SUBSYSTEM |
| 7664 | M: Thierry Reding <thierry.reding@gmail.com> | 7665 | M: Thierry Reding <thierry.reding@gmail.com> |
| 7665 | L: linux-pwm@vger.kernel.org | 7666 | L: linux-pwm@vger.kernel.org |
| 7666 | S: Maintained | 7667 | S: Maintained |
| 7667 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm.git | 7668 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm.git |
| 7668 | F: Documentation/pwm.txt | 7669 | F: Documentation/pwm.txt |
| 7669 | F: Documentation/devicetree/bindings/pwm/ | 7670 | F: Documentation/devicetree/bindings/pwm/ |
| 7670 | F: include/linux/pwm.h | 7671 | F: include/linux/pwm.h |
| 7671 | F: drivers/pwm/ | 7672 | F: drivers/pwm/ |
| 7672 | F: drivers/video/backlight/pwm_bl.c | 7673 | F: drivers/video/backlight/pwm_bl.c |
| 7673 | F: include/linux/pwm_backlight.h | 7674 | F: include/linux/pwm_backlight.h |
| 7674 | 7675 | ||
| 7675 | PXA2xx/PXA3xx SUPPORT | 7676 | PXA2xx/PXA3xx SUPPORT |
| 7676 | M: Daniel Mack <daniel@zonque.org> | 7677 | M: Daniel Mack <daniel@zonque.org> |
| 7677 | M: Haojian Zhuang <haojian.zhuang@gmail.com> | 7678 | M: Haojian Zhuang <haojian.zhuang@gmail.com> |
| 7678 | M: Robert Jarzmik <robert.jarzmik@free.fr> | 7679 | M: Robert Jarzmik <robert.jarzmik@free.fr> |
| 7679 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7680 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7680 | T: git git://github.com/hzhuang1/linux.git | 7681 | T: git git://github.com/hzhuang1/linux.git |
| 7681 | T: git git://github.com/rjarzmik/linux.git | 7682 | T: git git://github.com/rjarzmik/linux.git |
| 7682 | S: Maintained | 7683 | S: Maintained |
| 7683 | F: arch/arm/mach-pxa/ | 7684 | F: arch/arm/mach-pxa/ |
| 7684 | F: drivers/pcmcia/pxa2xx* | 7685 | F: drivers/pcmcia/pxa2xx* |
| 7685 | F: drivers/spi/spi-pxa2xx* | 7686 | F: drivers/spi/spi-pxa2xx* |
| 7686 | F: drivers/usb/gadget/udc/pxa2* | 7687 | F: drivers/usb/gadget/udc/pxa2* |
| 7687 | F: include/sound/pxa2xx-lib.h | 7688 | F: include/sound/pxa2xx-lib.h |
| 7688 | F: sound/arm/pxa* | 7689 | F: sound/arm/pxa* |
| 7689 | F: sound/soc/pxa/ | 7690 | F: sound/soc/pxa/ |
| 7690 | 7691 | ||
| 7691 | PXA3xx NAND FLASH DRIVER | 7692 | PXA3xx NAND FLASH DRIVER |
| 7692 | M: Ezequiel Garcia <ezequiel.garcia@free-electrons.com> | 7693 | M: Ezequiel Garcia <ezequiel.garcia@free-electrons.com> |
| 7693 | L: linux-mtd@lists.infradead.org | 7694 | L: linux-mtd@lists.infradead.org |
| 7694 | S: Maintained | 7695 | S: Maintained |
| 7695 | F: drivers/mtd/nand/pxa3xx_nand.c | 7696 | F: drivers/mtd/nand/pxa3xx_nand.c |
| 7696 | 7697 | ||
| 7697 | MMP SUPPORT | 7698 | MMP SUPPORT |
| 7698 | M: Eric Miao <eric.y.miao@gmail.com> | 7699 | M: Eric Miao <eric.y.miao@gmail.com> |
| 7699 | M: Haojian Zhuang <haojian.zhuang@gmail.com> | 7700 | M: Haojian Zhuang <haojian.zhuang@gmail.com> |
| 7700 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 7701 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 7701 | T: git git://github.com/hzhuang1/linux.git | 7702 | T: git git://github.com/hzhuang1/linux.git |
| 7702 | T: git git://git.linaro.org/people/ycmiao/pxa-linux.git | 7703 | T: git git://git.linaro.org/people/ycmiao/pxa-linux.git |
| 7703 | S: Maintained | 7704 | S: Maintained |
| 7704 | F: arch/arm/mach-mmp/ | 7705 | F: arch/arm/mach-mmp/ |
| 7705 | 7706 | ||
| 7706 | PXA MMCI DRIVER | 7707 | PXA MMCI DRIVER |
| 7707 | S: Orphan | 7708 | S: Orphan |
| 7708 | 7709 | ||
| 7709 | PXA RTC DRIVER | 7710 | PXA RTC DRIVER |
| 7710 | M: Robert Jarzmik <robert.jarzmik@free.fr> | 7711 | M: Robert Jarzmik <robert.jarzmik@free.fr> |
| 7711 | L: rtc-linux@googlegroups.com | 7712 | L: rtc-linux@googlegroups.com |
| 7712 | S: Maintained | 7713 | S: Maintained |
| 7713 | 7714 | ||
| 7714 | QAT DRIVER | 7715 | QAT DRIVER |
| 7715 | M: Tadeusz Struk <tadeusz.struk@intel.com> | 7716 | M: Tadeusz Struk <tadeusz.struk@intel.com> |
| 7716 | L: qat-linux@intel.com | 7717 | L: qat-linux@intel.com |
| 7717 | S: Supported | 7718 | S: Supported |
| 7718 | F: drivers/crypto/qat/ | 7719 | F: drivers/crypto/qat/ |
| 7719 | 7720 | ||
| 7720 | QIB DRIVER | 7721 | QIB DRIVER |
| 7721 | M: Mike Marciniszyn <infinipath@intel.com> | 7722 | M: Mike Marciniszyn <infinipath@intel.com> |
| 7722 | L: linux-rdma@vger.kernel.org | 7723 | L: linux-rdma@vger.kernel.org |
| 7723 | S: Supported | 7724 | S: Supported |
| 7724 | F: drivers/infiniband/hw/qib/ | 7725 | F: drivers/infiniband/hw/qib/ |
| 7725 | 7726 | ||
| 7726 | QLOGIC QLA1280 SCSI DRIVER | 7727 | QLOGIC QLA1280 SCSI DRIVER |
| 7727 | M: Michael Reed <mdr@sgi.com> | 7728 | M: Michael Reed <mdr@sgi.com> |
| 7728 | L: linux-scsi@vger.kernel.org | 7729 | L: linux-scsi@vger.kernel.org |
| 7729 | S: Maintained | 7730 | S: Maintained |
| 7730 | F: drivers/scsi/qla1280.[ch] | 7731 | F: drivers/scsi/qla1280.[ch] |
| 7731 | 7732 | ||
| 7732 | QLOGIC QLA2XXX FC-SCSI DRIVER | 7733 | QLOGIC QLA2XXX FC-SCSI DRIVER |
| 7733 | M: qla2xxx-upstream@qlogic.com | 7734 | M: qla2xxx-upstream@qlogic.com |
| 7734 | L: linux-scsi@vger.kernel.org | 7735 | L: linux-scsi@vger.kernel.org |
| 7735 | S: Supported | 7736 | S: Supported |
| 7736 | F: Documentation/scsi/LICENSE.qla2xxx | 7737 | F: Documentation/scsi/LICENSE.qla2xxx |
| 7737 | F: drivers/scsi/qla2xxx/ | 7738 | F: drivers/scsi/qla2xxx/ |
| 7738 | 7739 | ||
| 7739 | QLOGIC QLA4XXX iSCSI DRIVER | 7740 | QLOGIC QLA4XXX iSCSI DRIVER |
| 7740 | M: Vikas Chaudhary <vikas.chaudhary@qlogic.com> | 7741 | M: Vikas Chaudhary <vikas.chaudhary@qlogic.com> |
| 7741 | M: iscsi-driver@qlogic.com | 7742 | M: iscsi-driver@qlogic.com |
| 7742 | L: linux-scsi@vger.kernel.org | 7743 | L: linux-scsi@vger.kernel.org |
| 7743 | S: Supported | 7744 | S: Supported |
| 7744 | F: Documentation/scsi/LICENSE.qla4xxx | 7745 | F: Documentation/scsi/LICENSE.qla4xxx |
| 7745 | F: drivers/scsi/qla4xxx/ | 7746 | F: drivers/scsi/qla4xxx/ |
| 7746 | 7747 | ||
| 7747 | QLOGIC QLA3XXX NETWORK DRIVER | 7748 | QLOGIC QLA3XXX NETWORK DRIVER |
| 7748 | M: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com> | 7749 | M: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com> |
| 7749 | M: Ron Mercer <ron.mercer@qlogic.com> | 7750 | M: Ron Mercer <ron.mercer@qlogic.com> |
| 7750 | M: linux-driver@qlogic.com | 7751 | M: linux-driver@qlogic.com |
| 7751 | L: netdev@vger.kernel.org | 7752 | L: netdev@vger.kernel.org |
| 7752 | S: Supported | 7753 | S: Supported |
| 7753 | F: Documentation/networking/LICENSE.qla3xxx | 7754 | F: Documentation/networking/LICENSE.qla3xxx |
| 7754 | F: drivers/net/ethernet/qlogic/qla3xxx.* | 7755 | F: drivers/net/ethernet/qlogic/qla3xxx.* |
| 7755 | 7756 | ||
| 7756 | QLOGIC QLCNIC (1/10)Gb ETHERNET DRIVER | 7757 | QLOGIC QLCNIC (1/10)Gb ETHERNET DRIVER |
| 7757 | M: Shahed Shaikh <shahed.shaikh@qlogic.com> | 7758 | M: Shahed Shaikh <shahed.shaikh@qlogic.com> |
| 7758 | M: Dept-GELinuxNICDev@qlogic.com | 7759 | M: Dept-GELinuxNICDev@qlogic.com |
| 7759 | L: netdev@vger.kernel.org | 7760 | L: netdev@vger.kernel.org |
| 7760 | S: Supported | 7761 | S: Supported |
| 7761 | F: drivers/net/ethernet/qlogic/qlcnic/ | 7762 | F: drivers/net/ethernet/qlogic/qlcnic/ |
| 7762 | 7763 | ||
| 7763 | QLOGIC QLGE 10Gb ETHERNET DRIVER | 7764 | QLOGIC QLGE 10Gb ETHERNET DRIVER |
| 7764 | M: Harish Patil <harish.patil@qlogic.com> | 7765 | M: Harish Patil <harish.patil@qlogic.com> |
| 7765 | M: Sudarsana Kalluru <sudarsana.kalluru@qlogic.com> | 7766 | M: Sudarsana Kalluru <sudarsana.kalluru@qlogic.com> |
| 7766 | M: Dept-GELinuxNICDev@qlogic.com | 7767 | M: Dept-GELinuxNICDev@qlogic.com |
| 7767 | M: linux-driver@qlogic.com | 7768 | M: linux-driver@qlogic.com |
| 7768 | L: netdev@vger.kernel.org | 7769 | L: netdev@vger.kernel.org |
| 7769 | S: Supported | 7770 | S: Supported |
| 7770 | F: drivers/net/ethernet/qlogic/qlge/ | 7771 | F: drivers/net/ethernet/qlogic/qlge/ |
| 7771 | 7772 | ||
| 7772 | QNX4 FILESYSTEM | 7773 | QNX4 FILESYSTEM |
| 7773 | M: Anders Larsen <al@alarsen.net> | 7774 | M: Anders Larsen <al@alarsen.net> |
| 7774 | W: http://www.alarsen.net/linux/qnx4fs/ | 7775 | W: http://www.alarsen.net/linux/qnx4fs/ |
| 7775 | S: Maintained | 7776 | S: Maintained |
| 7776 | F: fs/qnx4/ | 7777 | F: fs/qnx4/ |
| 7777 | F: include/uapi/linux/qnx4_fs.h | 7778 | F: include/uapi/linux/qnx4_fs.h |
| 7778 | F: include/uapi/linux/qnxtypes.h | 7779 | F: include/uapi/linux/qnxtypes.h |
| 7779 | 7780 | ||
| 7780 | QT1010 MEDIA DRIVER | 7781 | QT1010 MEDIA DRIVER |
| 7781 | M: Antti Palosaari <crope@iki.fi> | 7782 | M: Antti Palosaari <crope@iki.fi> |
| 7782 | L: linux-media@vger.kernel.org | 7783 | L: linux-media@vger.kernel.org |
| 7783 | W: http://linuxtv.org/ | 7784 | W: http://linuxtv.org/ |
| 7784 | W: http://palosaari.fi/linux/ | 7785 | W: http://palosaari.fi/linux/ |
| 7785 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 7786 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 7786 | T: git git://linuxtv.org/anttip/media_tree.git | 7787 | T: git git://linuxtv.org/anttip/media_tree.git |
| 7787 | S: Maintained | 7788 | S: Maintained |
| 7788 | F: drivers/media/tuners/qt1010* | 7789 | F: drivers/media/tuners/qt1010* |
| 7789 | 7790 | ||
| 7790 | QUALCOMM ATHEROS ATH9K WIRELESS DRIVER | 7791 | QUALCOMM ATHEROS ATH9K WIRELESS DRIVER |
| 7791 | M: QCA ath9k Development <ath9k-devel@qca.qualcomm.com> | 7792 | M: QCA ath9k Development <ath9k-devel@qca.qualcomm.com> |
| 7792 | L: linux-wireless@vger.kernel.org | 7793 | L: linux-wireless@vger.kernel.org |
| 7793 | L: ath9k-devel@lists.ath9k.org | 7794 | L: ath9k-devel@lists.ath9k.org |
| 7794 | W: http://wireless.kernel.org/en/users/Drivers/ath9k | 7795 | W: http://wireless.kernel.org/en/users/Drivers/ath9k |
| 7795 | S: Supported | 7796 | S: Supported |
| 7796 | F: drivers/net/wireless/ath/ath9k/ | 7797 | F: drivers/net/wireless/ath/ath9k/ |
| 7797 | 7798 | ||
| 7798 | QUALCOMM ATHEROS ATH10K WIRELESS DRIVER | 7799 | QUALCOMM ATHEROS ATH10K WIRELESS DRIVER |
| 7799 | M: Kalle Valo <kvalo@qca.qualcomm.com> | 7800 | M: Kalle Valo <kvalo@qca.qualcomm.com> |
| 7800 | L: ath10k@lists.infradead.org | 7801 | L: ath10k@lists.infradead.org |
| 7801 | W: http://wireless.kernel.org/en/users/Drivers/ath10k | 7802 | W: http://wireless.kernel.org/en/users/Drivers/ath10k |
| 7802 | T: git git://github.com/kvalo/ath.git | 7803 | T: git git://github.com/kvalo/ath.git |
| 7803 | S: Supported | 7804 | S: Supported |
| 7804 | F: drivers/net/wireless/ath/ath10k/ | 7805 | F: drivers/net/wireless/ath/ath10k/ |
| 7805 | 7806 | ||
| 7806 | QUALCOMM HEXAGON ARCHITECTURE | 7807 | QUALCOMM HEXAGON ARCHITECTURE |
| 7807 | M: Richard Kuo <rkuo@codeaurora.org> | 7808 | M: Richard Kuo <rkuo@codeaurora.org> |
| 7808 | L: linux-hexagon@vger.kernel.org | 7809 | L: linux-hexagon@vger.kernel.org |
| 7809 | S: Supported | 7810 | S: Supported |
| 7810 | F: arch/hexagon/ | 7811 | F: arch/hexagon/ |
| 7811 | 7812 | ||
| 7812 | QUALCOMM WCN36XX WIRELESS DRIVER | 7813 | QUALCOMM WCN36XX WIRELESS DRIVER |
| 7813 | M: Eugene Krasnikov <k.eugene.e@gmail.com> | 7814 | M: Eugene Krasnikov <k.eugene.e@gmail.com> |
| 7814 | L: wcn36xx@lists.infradead.org | 7815 | L: wcn36xx@lists.infradead.org |
| 7815 | W: http://wireless.kernel.org/en/users/Drivers/wcn36xx | 7816 | W: http://wireless.kernel.org/en/users/Drivers/wcn36xx |
| 7816 | T: git git://github.com/KrasnikovEugene/wcn36xx.git | 7817 | T: git git://github.com/KrasnikovEugene/wcn36xx.git |
| 7817 | S: Supported | 7818 | S: Supported |
| 7818 | F: drivers/net/wireless/ath/wcn36xx/ | 7819 | F: drivers/net/wireless/ath/wcn36xx/ |
| 7819 | 7820 | ||
| 7820 | QUICKCAM PARALLEL PORT WEBCAMS | 7821 | QUICKCAM PARALLEL PORT WEBCAMS |
| 7821 | M: Hans Verkuil <hverkuil@xs4all.nl> | 7822 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 7822 | L: linux-media@vger.kernel.org | 7823 | L: linux-media@vger.kernel.org |
| 7823 | T: git git://linuxtv.org/media_tree.git | 7824 | T: git git://linuxtv.org/media_tree.git |
| 7824 | W: http://linuxtv.org | 7825 | W: http://linuxtv.org |
| 7825 | S: Odd Fixes | 7826 | S: Odd Fixes |
| 7826 | F: drivers/media/parport/*-qcam* | 7827 | F: drivers/media/parport/*-qcam* |
| 7827 | 7828 | ||
| 7828 | RADOS BLOCK DEVICE (RBD) | 7829 | RADOS BLOCK DEVICE (RBD) |
| 7829 | M: Yehuda Sadeh <yehuda@inktank.com> | 7830 | M: Yehuda Sadeh <yehuda@inktank.com> |
| 7830 | M: Sage Weil <sage@inktank.com> | 7831 | M: Sage Weil <sage@inktank.com> |
| 7831 | M: Alex Elder <elder@kernel.org> | 7832 | M: Alex Elder <elder@kernel.org> |
| 7832 | M: ceph-devel@vger.kernel.org | 7833 | M: ceph-devel@vger.kernel.org |
| 7833 | W: http://ceph.com/ | 7834 | W: http://ceph.com/ |
| 7834 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git | 7835 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client.git |
| 7835 | S: Supported | 7836 | S: Supported |
| 7836 | F: drivers/block/rbd.c | 7837 | F: drivers/block/rbd.c |
| 7837 | F: drivers/block/rbd_types.h | 7838 | F: drivers/block/rbd_types.h |
| 7838 | 7839 | ||
| 7839 | RADEON FRAMEBUFFER DISPLAY DRIVER | 7840 | RADEON FRAMEBUFFER DISPLAY DRIVER |
| 7840 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> | 7841 | M: Benjamin Herrenschmidt <benh@kernel.crashing.org> |
| 7841 | L: linux-fbdev@vger.kernel.org | 7842 | L: linux-fbdev@vger.kernel.org |
| 7842 | S: Maintained | 7843 | S: Maintained |
| 7843 | F: drivers/video/fbdev/aty/radeon* | 7844 | F: drivers/video/fbdev/aty/radeon* |
| 7844 | F: include/uapi/linux/radeonfb.h | 7845 | F: include/uapi/linux/radeonfb.h |
| 7845 | 7846 | ||
| 7846 | RADIOSHARK RADIO DRIVER | 7847 | RADIOSHARK RADIO DRIVER |
| 7847 | M: Hans de Goede <hdegoede@redhat.com> | 7848 | M: Hans de Goede <hdegoede@redhat.com> |
| 7848 | L: linux-media@vger.kernel.org | 7849 | L: linux-media@vger.kernel.org |
| 7849 | T: git git://linuxtv.org/media_tree.git | 7850 | T: git git://linuxtv.org/media_tree.git |
| 7850 | S: Maintained | 7851 | S: Maintained |
| 7851 | F: drivers/media/radio/radio-shark.c | 7852 | F: drivers/media/radio/radio-shark.c |
| 7852 | 7853 | ||
| 7853 | RADIOSHARK2 RADIO DRIVER | 7854 | RADIOSHARK2 RADIO DRIVER |
| 7854 | M: Hans de Goede <hdegoede@redhat.com> | 7855 | M: Hans de Goede <hdegoede@redhat.com> |
| 7855 | L: linux-media@vger.kernel.org | 7856 | L: linux-media@vger.kernel.org |
| 7856 | T: git git://linuxtv.org/media_tree.git | 7857 | T: git git://linuxtv.org/media_tree.git |
| 7857 | S: Maintained | 7858 | S: Maintained |
| 7858 | F: drivers/media/radio/radio-shark2.c | 7859 | F: drivers/media/radio/radio-shark2.c |
| 7859 | F: drivers/media/radio/radio-tea5777.c | 7860 | F: drivers/media/radio/radio-tea5777.c |
| 7860 | 7861 | ||
| 7861 | RAGE128 FRAMEBUFFER DISPLAY DRIVER | 7862 | RAGE128 FRAMEBUFFER DISPLAY DRIVER |
| 7862 | M: Paul Mackerras <paulus@samba.org> | 7863 | M: Paul Mackerras <paulus@samba.org> |
| 7863 | L: linux-fbdev@vger.kernel.org | 7864 | L: linux-fbdev@vger.kernel.org |
| 7864 | S: Maintained | 7865 | S: Maintained |
| 7865 | F: drivers/video/fbdev/aty/aty128fb.c | 7866 | F: drivers/video/fbdev/aty/aty128fb.c |
| 7866 | 7867 | ||
| 7867 | RALINK RT2X00 WIRELESS LAN DRIVER | 7868 | RALINK RT2X00 WIRELESS LAN DRIVER |
| 7868 | P: rt2x00 project | 7869 | P: rt2x00 project |
| 7869 | M: Stanislaw Gruszka <sgruszka@redhat.com> | 7870 | M: Stanislaw Gruszka <sgruszka@redhat.com> |
| 7870 | M: Helmut Schaa <helmut.schaa@googlemail.com> | 7871 | M: Helmut Schaa <helmut.schaa@googlemail.com> |
| 7871 | L: linux-wireless@vger.kernel.org | 7872 | L: linux-wireless@vger.kernel.org |
| 7872 | L: users@rt2x00.serialmonkey.com (moderated for non-subscribers) | 7873 | L: users@rt2x00.serialmonkey.com (moderated for non-subscribers) |
| 7873 | W: http://rt2x00.serialmonkey.com/ | 7874 | W: http://rt2x00.serialmonkey.com/ |
| 7874 | S: Maintained | 7875 | S: Maintained |
| 7875 | F: drivers/net/wireless/rt2x00/ | 7876 | F: drivers/net/wireless/rt2x00/ |
| 7876 | 7877 | ||
| 7877 | RAMDISK RAM BLOCK DEVICE DRIVER | 7878 | RAMDISK RAM BLOCK DEVICE DRIVER |
| 7878 | M: Nick Piggin <npiggin@kernel.dk> | 7879 | M: Nick Piggin <npiggin@kernel.dk> |
| 7879 | S: Maintained | 7880 | S: Maintained |
| 7880 | F: Documentation/blockdev/ramdisk.txt | 7881 | F: Documentation/blockdev/ramdisk.txt |
| 7881 | F: drivers/block/brd.c | 7882 | F: drivers/block/brd.c |
| 7882 | 7883 | ||
| 7883 | RANDOM NUMBER DRIVER | 7884 | RANDOM NUMBER DRIVER |
| 7884 | M: "Theodore Ts'o" <tytso@mit.edu> | 7885 | M: "Theodore Ts'o" <tytso@mit.edu> |
| 7885 | S: Maintained | 7886 | S: Maintained |
| 7886 | F: drivers/char/random.c | 7887 | F: drivers/char/random.c |
| 7887 | 7888 | ||
| 7888 | RAPIDIO SUBSYSTEM | 7889 | RAPIDIO SUBSYSTEM |
| 7889 | M: Matt Porter <mporter@kernel.crashing.org> | 7890 | M: Matt Porter <mporter@kernel.crashing.org> |
| 7890 | M: Alexandre Bounine <alexandre.bounine@idt.com> | 7891 | M: Alexandre Bounine <alexandre.bounine@idt.com> |
| 7891 | S: Maintained | 7892 | S: Maintained |
| 7892 | F: drivers/rapidio/ | 7893 | F: drivers/rapidio/ |
| 7893 | 7894 | ||
| 7894 | RAYLINK/WEBGEAR 802.11 WIRELESS LAN DRIVER | 7895 | RAYLINK/WEBGEAR 802.11 WIRELESS LAN DRIVER |
| 7895 | L: linux-wireless@vger.kernel.org | 7896 | L: linux-wireless@vger.kernel.org |
| 7896 | S: Orphan | 7897 | S: Orphan |
| 7897 | F: drivers/net/wireless/ray* | 7898 | F: drivers/net/wireless/ray* |
| 7898 | 7899 | ||
| 7899 | RCUTORTURE MODULE | 7900 | RCUTORTURE MODULE |
| 7900 | M: Josh Triplett <josh@joshtriplett.org> | 7901 | M: Josh Triplett <josh@joshtriplett.org> |
| 7901 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> | 7902 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
| 7902 | L: linux-kernel@vger.kernel.org | 7903 | L: linux-kernel@vger.kernel.org |
| 7903 | S: Supported | 7904 | S: Supported |
| 7904 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git | 7905 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git |
| 7905 | F: Documentation/RCU/torture.txt | 7906 | F: Documentation/RCU/torture.txt |
| 7906 | F: kernel/rcu/rcutorture.c | 7907 | F: kernel/rcu/rcutorture.c |
| 7907 | 7908 | ||
| 7908 | RCUTORTURE TEST FRAMEWORK | 7909 | RCUTORTURE TEST FRAMEWORK |
| 7909 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> | 7910 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
| 7910 | M: Josh Triplett <josh@joshtriplett.org> | 7911 | M: Josh Triplett <josh@joshtriplett.org> |
| 7911 | R: Steven Rostedt <rostedt@goodmis.org> | 7912 | R: Steven Rostedt <rostedt@goodmis.org> |
| 7912 | R: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> | 7913 | R: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> |
| 7913 | R: Lai Jiangshan <laijs@cn.fujitsu.com> | 7914 | R: Lai Jiangshan <laijs@cn.fujitsu.com> |
| 7914 | L: linux-kernel@vger.kernel.org | 7915 | L: linux-kernel@vger.kernel.org |
| 7915 | S: Supported | 7916 | S: Supported |
| 7916 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git | 7917 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git |
| 7917 | F: tools/testing/selftests/rcutorture | 7918 | F: tools/testing/selftests/rcutorture |
| 7918 | 7919 | ||
| 7919 | RDC R-321X SoC | 7920 | RDC R-321X SoC |
| 7920 | M: Florian Fainelli <florian@openwrt.org> | 7921 | M: Florian Fainelli <florian@openwrt.org> |
| 7921 | S: Maintained | 7922 | S: Maintained |
| 7922 | 7923 | ||
| 7923 | RDC R6040 FAST ETHERNET DRIVER | 7924 | RDC R6040 FAST ETHERNET DRIVER |
| 7924 | M: Florian Fainelli <florian@openwrt.org> | 7925 | M: Florian Fainelli <florian@openwrt.org> |
| 7925 | L: netdev@vger.kernel.org | 7926 | L: netdev@vger.kernel.org |
| 7926 | S: Maintained | 7927 | S: Maintained |
| 7927 | F: drivers/net/ethernet/rdc/r6040.c | 7928 | F: drivers/net/ethernet/rdc/r6040.c |
| 7928 | 7929 | ||
| 7929 | RDS - RELIABLE DATAGRAM SOCKETS | 7930 | RDS - RELIABLE DATAGRAM SOCKETS |
| 7930 | M: Chien Yen <chien.yen@oracle.com> | 7931 | M: Chien Yen <chien.yen@oracle.com> |
| 7931 | L: rds-devel@oss.oracle.com (moderated for non-subscribers) | 7932 | L: rds-devel@oss.oracle.com (moderated for non-subscribers) |
| 7932 | S: Supported | 7933 | S: Supported |
| 7933 | F: net/rds/ | 7934 | F: net/rds/ |
| 7934 | 7935 | ||
| 7935 | READ-COPY UPDATE (RCU) | 7936 | READ-COPY UPDATE (RCU) |
| 7936 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> | 7937 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
| 7937 | M: Josh Triplett <josh@joshtriplett.org> | 7938 | M: Josh Triplett <josh@joshtriplett.org> |
| 7938 | R: Steven Rostedt <rostedt@goodmis.org> | 7939 | R: Steven Rostedt <rostedt@goodmis.org> |
| 7939 | R: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> | 7940 | R: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> |
| 7940 | R: Lai Jiangshan <laijs@cn.fujitsu.com> | 7941 | R: Lai Jiangshan <laijs@cn.fujitsu.com> |
| 7941 | L: linux-kernel@vger.kernel.org | 7942 | L: linux-kernel@vger.kernel.org |
| 7942 | W: http://www.rdrop.com/users/paulmck/RCU/ | 7943 | W: http://www.rdrop.com/users/paulmck/RCU/ |
| 7943 | S: Supported | 7944 | S: Supported |
| 7944 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git | 7945 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git |
| 7945 | F: Documentation/RCU/ | 7946 | F: Documentation/RCU/ |
| 7946 | X: Documentation/RCU/torture.txt | 7947 | X: Documentation/RCU/torture.txt |
| 7947 | F: include/linux/rcu* | 7948 | F: include/linux/rcu* |
| 7948 | X: include/linux/srcu.h | 7949 | X: include/linux/srcu.h |
| 7949 | F: kernel/rcu/ | 7950 | F: kernel/rcu/ |
| 7950 | X: kernel/torture.c | 7951 | X: kernel/torture.c |
| 7951 | 7952 | ||
| 7952 | REAL TIME CLOCK (RTC) SUBSYSTEM | 7953 | REAL TIME CLOCK (RTC) SUBSYSTEM |
| 7953 | M: Alessandro Zummo <a.zummo@towertech.it> | 7954 | M: Alessandro Zummo <a.zummo@towertech.it> |
| 7954 | L: rtc-linux@googlegroups.com | 7955 | L: rtc-linux@googlegroups.com |
| 7955 | Q: http://patchwork.ozlabs.org/project/rtc-linux/list/ | 7956 | Q: http://patchwork.ozlabs.org/project/rtc-linux/list/ |
| 7956 | S: Maintained | 7957 | S: Maintained |
| 7957 | F: Documentation/rtc.txt | 7958 | F: Documentation/rtc.txt |
| 7958 | F: drivers/rtc/ | 7959 | F: drivers/rtc/ |
| 7959 | F: include/linux/rtc.h | 7960 | F: include/linux/rtc.h |
| 7960 | F: include/uapi/linux/rtc.h | 7961 | F: include/uapi/linux/rtc.h |
| 7961 | 7962 | ||
| 7962 | REALTEK AUDIO CODECS | 7963 | REALTEK AUDIO CODECS |
| 7963 | M: Bard Liao <bardliao@realtek.com> | 7964 | M: Bard Liao <bardliao@realtek.com> |
| 7964 | M: Oder Chiou <oder_chiou@realtek.com> | 7965 | M: Oder Chiou <oder_chiou@realtek.com> |
| 7965 | S: Maintained | 7966 | S: Maintained |
| 7966 | F: sound/soc/codecs/rt* | 7967 | F: sound/soc/codecs/rt* |
| 7967 | F: include/sound/rt*.h | 7968 | F: include/sound/rt*.h |
| 7968 | 7969 | ||
| 7969 | REISERFS FILE SYSTEM | 7970 | REISERFS FILE SYSTEM |
| 7970 | L: reiserfs-devel@vger.kernel.org | 7971 | L: reiserfs-devel@vger.kernel.org |
| 7971 | S: Supported | 7972 | S: Supported |
| 7972 | F: fs/reiserfs/ | 7973 | F: fs/reiserfs/ |
| 7973 | 7974 | ||
| 7974 | REGISTER MAP ABSTRACTION | 7975 | REGISTER MAP ABSTRACTION |
| 7975 | M: Mark Brown <broonie@kernel.org> | 7976 | M: Mark Brown <broonie@kernel.org> |
| 7976 | L: linux-kernel@vger.kernel.org | 7977 | L: linux-kernel@vger.kernel.org |
| 7977 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap.git | 7978 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap.git |
| 7978 | S: Supported | 7979 | S: Supported |
| 7979 | F: drivers/base/regmap/ | 7980 | F: drivers/base/regmap/ |
| 7980 | F: include/linux/regmap.h | 7981 | F: include/linux/regmap.h |
| 7981 | 7982 | ||
| 7982 | REMOTE PROCESSOR (REMOTEPROC) SUBSYSTEM | 7983 | REMOTE PROCESSOR (REMOTEPROC) SUBSYSTEM |
| 7983 | M: Ohad Ben-Cohen <ohad@wizery.com> | 7984 | M: Ohad Ben-Cohen <ohad@wizery.com> |
| 7984 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ohad/remoteproc.git | 7985 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ohad/remoteproc.git |
| 7985 | S: Maintained | 7986 | S: Maintained |
| 7986 | F: drivers/remoteproc/ | 7987 | F: drivers/remoteproc/ |
| 7987 | F: Documentation/remoteproc.txt | 7988 | F: Documentation/remoteproc.txt |
| 7988 | F: include/linux/remoteproc.h | 7989 | F: include/linux/remoteproc.h |
| 7989 | 7990 | ||
| 7990 | REMOTE PROCESSOR MESSAGING (RPMSG) SUBSYSTEM | 7991 | REMOTE PROCESSOR MESSAGING (RPMSG) SUBSYSTEM |
| 7991 | M: Ohad Ben-Cohen <ohad@wizery.com> | 7992 | M: Ohad Ben-Cohen <ohad@wizery.com> |
| 7992 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ohad/rpmsg.git | 7993 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ohad/rpmsg.git |
| 7993 | S: Maintained | 7994 | S: Maintained |
| 7994 | F: drivers/rpmsg/ | 7995 | F: drivers/rpmsg/ |
| 7995 | F: Documentation/rpmsg.txt | 7996 | F: Documentation/rpmsg.txt |
| 7996 | F: include/linux/rpmsg.h | 7997 | F: include/linux/rpmsg.h |
| 7997 | 7998 | ||
| 7998 | RESET CONTROLLER FRAMEWORK | 7999 | RESET CONTROLLER FRAMEWORK |
| 7999 | M: Philipp Zabel <p.zabel@pengutronix.de> | 8000 | M: Philipp Zabel <p.zabel@pengutronix.de> |
| 8000 | S: Maintained | 8001 | S: Maintained |
| 8001 | F: drivers/reset/ | 8002 | F: drivers/reset/ |
| 8002 | F: Documentation/devicetree/bindings/reset/ | 8003 | F: Documentation/devicetree/bindings/reset/ |
| 8003 | F: include/linux/reset.h | 8004 | F: include/linux/reset.h |
| 8004 | F: include/linux/reset-controller.h | 8005 | F: include/linux/reset-controller.h |
| 8005 | 8006 | ||
| 8006 | RFKILL | 8007 | RFKILL |
| 8007 | M: Johannes Berg <johannes@sipsolutions.net> | 8008 | M: Johannes Berg <johannes@sipsolutions.net> |
| 8008 | L: linux-wireless@vger.kernel.org | 8009 | L: linux-wireless@vger.kernel.org |
| 8009 | W: http://wireless.kernel.org/ | 8010 | W: http://wireless.kernel.org/ |
| 8010 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211.git | 8011 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211.git |
| 8011 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next.git | 8012 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next.git |
| 8012 | S: Maintained | 8013 | S: Maintained |
| 8013 | F: Documentation/rfkill.txt | 8014 | F: Documentation/rfkill.txt |
| 8014 | F: net/rfkill/ | 8015 | F: net/rfkill/ |
| 8015 | 8016 | ||
| 8016 | RICOH SMARTMEDIA/XD DRIVER | 8017 | RICOH SMARTMEDIA/XD DRIVER |
| 8017 | M: Maxim Levitsky <maximlevitsky@gmail.com> | 8018 | M: Maxim Levitsky <maximlevitsky@gmail.com> |
| 8018 | S: Maintained | 8019 | S: Maintained |
| 8019 | F: drivers/mtd/nand/r852.c | 8020 | F: drivers/mtd/nand/r852.c |
| 8020 | F: drivers/mtd/nand/r852.h | 8021 | F: drivers/mtd/nand/r852.h |
| 8021 | 8022 | ||
| 8022 | RICOH R5C592 MEMORYSTICK DRIVER | 8023 | RICOH R5C592 MEMORYSTICK DRIVER |
| 8023 | M: Maxim Levitsky <maximlevitsky@gmail.com> | 8024 | M: Maxim Levitsky <maximlevitsky@gmail.com> |
| 8024 | S: Maintained | 8025 | S: Maintained |
| 8025 | F: drivers/memstick/host/r592.* | 8026 | F: drivers/memstick/host/r592.* |
| 8026 | 8027 | ||
| 8027 | ROCCAT DRIVERS | 8028 | ROCCAT DRIVERS |
| 8028 | M: Stefan Achatz <erazor_de@users.sourceforge.net> | 8029 | M: Stefan Achatz <erazor_de@users.sourceforge.net> |
| 8029 | W: http://sourceforge.net/projects/roccat/ | 8030 | W: http://sourceforge.net/projects/roccat/ |
| 8030 | S: Maintained | 8031 | S: Maintained |
| 8031 | F: drivers/hid/hid-roccat* | 8032 | F: drivers/hid/hid-roccat* |
| 8032 | F: include/linux/hid-roccat* | 8033 | F: include/linux/hid-roccat* |
| 8033 | F: Documentation/ABI/*/sysfs-driver-hid-roccat* | 8034 | F: Documentation/ABI/*/sysfs-driver-hid-roccat* |
| 8034 | 8035 | ||
| 8035 | ROCKER DRIVER | 8036 | ROCKER DRIVER |
| 8036 | M: Jiri Pirko <jiri@resnulli.us> | 8037 | M: Jiri Pirko <jiri@resnulli.us> |
| 8037 | M: Scott Feldman <sfeldma@gmail.com> | 8038 | M: Scott Feldman <sfeldma@gmail.com> |
| 8038 | L: netdev@vger.kernel.org | 8039 | L: netdev@vger.kernel.org |
| 8039 | S: Supported | 8040 | S: Supported |
| 8040 | F: drivers/net/ethernet/rocker/ | 8041 | F: drivers/net/ethernet/rocker/ |
| 8041 | 8042 | ||
| 8042 | ROCKETPORT DRIVER | 8043 | ROCKETPORT DRIVER |
| 8043 | P: Comtrol Corp. | 8044 | P: Comtrol Corp. |
| 8044 | W: http://www.comtrol.com | 8045 | W: http://www.comtrol.com |
| 8045 | S: Maintained | 8046 | S: Maintained |
| 8046 | F: Documentation/serial/rocket.txt | 8047 | F: Documentation/serial/rocket.txt |
| 8047 | F: drivers/tty/rocket* | 8048 | F: drivers/tty/rocket* |
| 8048 | 8049 | ||
| 8049 | ROCKETPORT EXPRESS/INFINITY DRIVER | 8050 | ROCKETPORT EXPRESS/INFINITY DRIVER |
| 8050 | M: Kevin Cernekee <cernekee@gmail.com> | 8051 | M: Kevin Cernekee <cernekee@gmail.com> |
| 8051 | L: linux-serial@vger.kernel.org | 8052 | L: linux-serial@vger.kernel.org |
| 8052 | S: Odd Fixes | 8053 | S: Odd Fixes |
| 8053 | F: drivers/tty/serial/rp2.* | 8054 | F: drivers/tty/serial/rp2.* |
| 8054 | 8055 | ||
| 8055 | ROSE NETWORK LAYER | 8056 | ROSE NETWORK LAYER |
| 8056 | M: Ralf Baechle <ralf@linux-mips.org> | 8057 | M: Ralf Baechle <ralf@linux-mips.org> |
| 8057 | L: linux-hams@vger.kernel.org | 8058 | L: linux-hams@vger.kernel.org |
| 8058 | W: http://www.linux-ax25.org/ | 8059 | W: http://www.linux-ax25.org/ |
| 8059 | S: Maintained | 8060 | S: Maintained |
| 8060 | F: include/net/rose.h | 8061 | F: include/net/rose.h |
| 8061 | F: include/uapi/linux/rose.h | 8062 | F: include/uapi/linux/rose.h |
| 8062 | F: net/rose/ | 8063 | F: net/rose/ |
| 8063 | 8064 | ||
| 8064 | RTL2830 MEDIA DRIVER | 8065 | RTL2830 MEDIA DRIVER |
| 8065 | M: Antti Palosaari <crope@iki.fi> | 8066 | M: Antti Palosaari <crope@iki.fi> |
| 8066 | L: linux-media@vger.kernel.org | 8067 | L: linux-media@vger.kernel.org |
| 8067 | W: http://linuxtv.org/ | 8068 | W: http://linuxtv.org/ |
| 8068 | W: http://palosaari.fi/linux/ | 8069 | W: http://palosaari.fi/linux/ |
| 8069 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8070 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8070 | T: git git://linuxtv.org/anttip/media_tree.git | 8071 | T: git git://linuxtv.org/anttip/media_tree.git |
| 8071 | S: Maintained | 8072 | S: Maintained |
| 8072 | F: drivers/media/dvb-frontends/rtl2830* | 8073 | F: drivers/media/dvb-frontends/rtl2830* |
| 8073 | 8074 | ||
| 8074 | RTL2832 MEDIA DRIVER | 8075 | RTL2832 MEDIA DRIVER |
| 8075 | M: Antti Palosaari <crope@iki.fi> | 8076 | M: Antti Palosaari <crope@iki.fi> |
| 8076 | L: linux-media@vger.kernel.org | 8077 | L: linux-media@vger.kernel.org |
| 8077 | W: http://linuxtv.org/ | 8078 | W: http://linuxtv.org/ |
| 8078 | W: http://palosaari.fi/linux/ | 8079 | W: http://palosaari.fi/linux/ |
| 8079 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8080 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8080 | T: git git://linuxtv.org/anttip/media_tree.git | 8081 | T: git git://linuxtv.org/anttip/media_tree.git |
| 8081 | S: Maintained | 8082 | S: Maintained |
| 8082 | F: drivers/media/dvb-frontends/rtl2832* | 8083 | F: drivers/media/dvb-frontends/rtl2832* |
| 8083 | 8084 | ||
| 8084 | RTL2832_SDR MEDIA DRIVER | 8085 | RTL2832_SDR MEDIA DRIVER |
| 8085 | M: Antti Palosaari <crope@iki.fi> | 8086 | M: Antti Palosaari <crope@iki.fi> |
| 8086 | L: linux-media@vger.kernel.org | 8087 | L: linux-media@vger.kernel.org |
| 8087 | W: http://linuxtv.org/ | 8088 | W: http://linuxtv.org/ |
| 8088 | W: http://palosaari.fi/linux/ | 8089 | W: http://palosaari.fi/linux/ |
| 8089 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8090 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8090 | T: git git://linuxtv.org/anttip/media_tree.git | 8091 | T: git git://linuxtv.org/anttip/media_tree.git |
| 8091 | S: Maintained | 8092 | S: Maintained |
| 8092 | F: drivers/media/dvb-frontends/rtl2832_sdr* | 8093 | F: drivers/media/dvb-frontends/rtl2832_sdr* |
| 8093 | 8094 | ||
| 8094 | RTL8180 WIRELESS DRIVER | 8095 | RTL8180 WIRELESS DRIVER |
| 8095 | L: linux-wireless@vger.kernel.org | 8096 | L: linux-wireless@vger.kernel.org |
| 8096 | W: http://wireless.kernel.org/ | 8097 | W: http://wireless.kernel.org/ |
| 8097 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git | 8098 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git |
| 8098 | S: Orphan | 8099 | S: Orphan |
| 8099 | F: drivers/net/wireless/rtl818x/rtl8180/ | 8100 | F: drivers/net/wireless/rtl818x/rtl8180/ |
| 8100 | 8101 | ||
| 8101 | RTL8187 WIRELESS DRIVER | 8102 | RTL8187 WIRELESS DRIVER |
| 8102 | M: Herton Ronaldo Krzesinski <herton@canonical.com> | 8103 | M: Herton Ronaldo Krzesinski <herton@canonical.com> |
| 8103 | M: Hin-Tak Leung <htl10@users.sourceforge.net> | 8104 | M: Hin-Tak Leung <htl10@users.sourceforge.net> |
| 8104 | M: Larry Finger <Larry.Finger@lwfinger.net> | 8105 | M: Larry Finger <Larry.Finger@lwfinger.net> |
| 8105 | L: linux-wireless@vger.kernel.org | 8106 | L: linux-wireless@vger.kernel.org |
| 8106 | W: http://wireless.kernel.org/ | 8107 | W: http://wireless.kernel.org/ |
| 8107 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git | 8108 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git |
| 8108 | S: Maintained | 8109 | S: Maintained |
| 8109 | F: drivers/net/wireless/rtl818x/rtl8187/ | 8110 | F: drivers/net/wireless/rtl818x/rtl8187/ |
| 8110 | 8111 | ||
| 8111 | RTL8192CE WIRELESS DRIVER | 8112 | RTL8192CE WIRELESS DRIVER |
| 8112 | M: Larry Finger <Larry.Finger@lwfinger.net> | 8113 | M: Larry Finger <Larry.Finger@lwfinger.net> |
| 8113 | M: Chaoming Li <chaoming_li@realsil.com.cn> | 8114 | M: Chaoming Li <chaoming_li@realsil.com.cn> |
| 8114 | L: linux-wireless@vger.kernel.org | 8115 | L: linux-wireless@vger.kernel.org |
| 8115 | W: http://wireless.kernel.org/ | 8116 | W: http://wireless.kernel.org/ |
| 8116 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git | 8117 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git |
| 8117 | S: Maintained | 8118 | S: Maintained |
| 8118 | F: drivers/net/wireless/rtlwifi/ | 8119 | F: drivers/net/wireless/rtlwifi/ |
| 8119 | F: drivers/net/wireless/rtlwifi/rtl8192ce/ | 8120 | F: drivers/net/wireless/rtlwifi/rtl8192ce/ |
| 8120 | 8121 | ||
| 8121 | S3 SAVAGE FRAMEBUFFER DRIVER | 8122 | S3 SAVAGE FRAMEBUFFER DRIVER |
| 8122 | M: Antonino Daplas <adaplas@gmail.com> | 8123 | M: Antonino Daplas <adaplas@gmail.com> |
| 8123 | L: linux-fbdev@vger.kernel.org | 8124 | L: linux-fbdev@vger.kernel.org |
| 8124 | S: Maintained | 8125 | S: Maintained |
| 8125 | F: drivers/video/fbdev/savage/ | 8126 | F: drivers/video/fbdev/savage/ |
| 8126 | 8127 | ||
| 8127 | S390 | 8128 | S390 |
| 8128 | M: Martin Schwidefsky <schwidefsky@de.ibm.com> | 8129 | M: Martin Schwidefsky <schwidefsky@de.ibm.com> |
| 8129 | M: Heiko Carstens <heiko.carstens@de.ibm.com> | 8130 | M: Heiko Carstens <heiko.carstens@de.ibm.com> |
| 8130 | M: linux390@de.ibm.com | 8131 | M: linux390@de.ibm.com |
| 8131 | L: linux-s390@vger.kernel.org | 8132 | L: linux-s390@vger.kernel.org |
| 8132 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8133 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8133 | S: Supported | 8134 | S: Supported |
| 8134 | F: arch/s390/ | 8135 | F: arch/s390/ |
| 8135 | F: drivers/s390/ | 8136 | F: drivers/s390/ |
| 8136 | F: Documentation/s390/ | 8137 | F: Documentation/s390/ |
| 8137 | F: Documentation/DocBook/s390* | 8138 | F: Documentation/DocBook/s390* |
| 8138 | 8139 | ||
| 8139 | S390 COMMON I/O LAYER | 8140 | S390 COMMON I/O LAYER |
| 8140 | M: Sebastian Ott <sebott@linux.vnet.ibm.com> | 8141 | M: Sebastian Ott <sebott@linux.vnet.ibm.com> |
| 8141 | M: Peter Oberparleiter <oberpar@linux.vnet.ibm.com> | 8142 | M: Peter Oberparleiter <oberpar@linux.vnet.ibm.com> |
| 8142 | L: linux-s390@vger.kernel.org | 8143 | L: linux-s390@vger.kernel.org |
| 8143 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8144 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8144 | S: Supported | 8145 | S: Supported |
| 8145 | F: drivers/s390/cio/ | 8146 | F: drivers/s390/cio/ |
| 8146 | 8147 | ||
| 8147 | S390 DASD DRIVER | 8148 | S390 DASD DRIVER |
| 8148 | M: Stefan Weinhuber <wein@de.ibm.com> | 8149 | M: Stefan Weinhuber <wein@de.ibm.com> |
| 8149 | M: Stefan Haberland <stefan.haberland@de.ibm.com> | 8150 | M: Stefan Haberland <stefan.haberland@de.ibm.com> |
| 8150 | L: linux-s390@vger.kernel.org | 8151 | L: linux-s390@vger.kernel.org |
| 8151 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8152 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8152 | S: Supported | 8153 | S: Supported |
| 8153 | F: drivers/s390/block/dasd* | 8154 | F: drivers/s390/block/dasd* |
| 8154 | F: block/partitions/ibm.c | 8155 | F: block/partitions/ibm.c |
| 8155 | 8156 | ||
| 8156 | S390 NETWORK DRIVERS | 8157 | S390 NETWORK DRIVERS |
| 8157 | M: Ursula Braun <ursula.braun@de.ibm.com> | 8158 | M: Ursula Braun <ursula.braun@de.ibm.com> |
| 8158 | M: Frank Blaschka <blaschka@linux.vnet.ibm.com> | 8159 | M: Frank Blaschka <blaschka@linux.vnet.ibm.com> |
| 8159 | M: linux390@de.ibm.com | 8160 | M: linux390@de.ibm.com |
| 8160 | L: linux-s390@vger.kernel.org | 8161 | L: linux-s390@vger.kernel.org |
| 8161 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8162 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8162 | S: Supported | 8163 | S: Supported |
| 8163 | F: drivers/s390/net/ | 8164 | F: drivers/s390/net/ |
| 8164 | 8165 | ||
| 8165 | S390 PCI SUBSYSTEM | 8166 | S390 PCI SUBSYSTEM |
| 8166 | M: Sebastian Ott <sebott@linux.vnet.ibm.com> | 8167 | M: Sebastian Ott <sebott@linux.vnet.ibm.com> |
| 8167 | M: Gerald Schaefer <gerald.schaefer@de.ibm.com> | 8168 | M: Gerald Schaefer <gerald.schaefer@de.ibm.com> |
| 8168 | L: linux-s390@vger.kernel.org | 8169 | L: linux-s390@vger.kernel.org |
| 8169 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8170 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8170 | S: Supported | 8171 | S: Supported |
| 8171 | F: arch/s390/pci/ | 8172 | F: arch/s390/pci/ |
| 8172 | F: drivers/pci/hotplug/s390_pci_hpc.c | 8173 | F: drivers/pci/hotplug/s390_pci_hpc.c |
| 8173 | 8174 | ||
| 8174 | S390 ZCRYPT DRIVER | 8175 | S390 ZCRYPT DRIVER |
| 8175 | M: Ingo Tuchscherer <ingo.tuchscherer@de.ibm.com> | 8176 | M: Ingo Tuchscherer <ingo.tuchscherer@de.ibm.com> |
| 8176 | M: linux390@de.ibm.com | 8177 | M: linux390@de.ibm.com |
| 8177 | L: linux-s390@vger.kernel.org | 8178 | L: linux-s390@vger.kernel.org |
| 8178 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8179 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8179 | S: Supported | 8180 | S: Supported |
| 8180 | F: drivers/s390/crypto/ | 8181 | F: drivers/s390/crypto/ |
| 8181 | 8182 | ||
| 8182 | S390 ZFCP DRIVER | 8183 | S390 ZFCP DRIVER |
| 8183 | M: Steffen Maier <maier@linux.vnet.ibm.com> | 8184 | M: Steffen Maier <maier@linux.vnet.ibm.com> |
| 8184 | M: linux390@de.ibm.com | 8185 | M: linux390@de.ibm.com |
| 8185 | L: linux-s390@vger.kernel.org | 8186 | L: linux-s390@vger.kernel.org |
| 8186 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8187 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8187 | S: Supported | 8188 | S: Supported |
| 8188 | F: drivers/s390/scsi/zfcp_* | 8189 | F: drivers/s390/scsi/zfcp_* |
| 8189 | 8190 | ||
| 8190 | S390 IUCV NETWORK LAYER | 8191 | S390 IUCV NETWORK LAYER |
| 8191 | M: Ursula Braun <ursula.braun@de.ibm.com> | 8192 | M: Ursula Braun <ursula.braun@de.ibm.com> |
| 8192 | M: linux390@de.ibm.com | 8193 | M: linux390@de.ibm.com |
| 8193 | L: linux-s390@vger.kernel.org | 8194 | L: linux-s390@vger.kernel.org |
| 8194 | W: http://www.ibm.com/developerworks/linux/linux390/ | 8195 | W: http://www.ibm.com/developerworks/linux/linux390/ |
| 8195 | S: Supported | 8196 | S: Supported |
| 8196 | F: drivers/s390/net/*iucv* | 8197 | F: drivers/s390/net/*iucv* |
| 8197 | F: include/net/iucv/ | 8198 | F: include/net/iucv/ |
| 8198 | F: net/iucv/ | 8199 | F: net/iucv/ |
| 8199 | 8200 | ||
| 8200 | S3C24XX SD/MMC Driver | 8201 | S3C24XX SD/MMC Driver |
| 8201 | M: Ben Dooks <ben-linux@fluff.org> | 8202 | M: Ben Dooks <ben-linux@fluff.org> |
| 8202 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 8203 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 8203 | S: Supported | 8204 | S: Supported |
| 8204 | F: drivers/mmc/host/s3cmci.* | 8205 | F: drivers/mmc/host/s3cmci.* |
| 8205 | 8206 | ||
| 8206 | SAA6588 RDS RECEIVER DRIVER | 8207 | SAA6588 RDS RECEIVER DRIVER |
| 8207 | M: Hans Verkuil <hverkuil@xs4all.nl> | 8208 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 8208 | L: linux-media@vger.kernel.org | 8209 | L: linux-media@vger.kernel.org |
| 8209 | T: git git://linuxtv.org/media_tree.git | 8210 | T: git git://linuxtv.org/media_tree.git |
| 8210 | W: http://linuxtv.org | 8211 | W: http://linuxtv.org |
| 8211 | S: Odd Fixes | 8212 | S: Odd Fixes |
| 8212 | F: drivers/media/i2c/saa6588* | 8213 | F: drivers/media/i2c/saa6588* |
| 8213 | 8214 | ||
| 8214 | SAA7134 VIDEO4LINUX DRIVER | 8215 | SAA7134 VIDEO4LINUX DRIVER |
| 8215 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 8216 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 8216 | L: linux-media@vger.kernel.org | 8217 | L: linux-media@vger.kernel.org |
| 8217 | W: http://linuxtv.org | 8218 | W: http://linuxtv.org |
| 8218 | T: git git://linuxtv.org/media_tree.git | 8219 | T: git git://linuxtv.org/media_tree.git |
| 8219 | S: Odd fixes | 8220 | S: Odd fixes |
| 8220 | F: Documentation/video4linux/*.saa7134 | 8221 | F: Documentation/video4linux/*.saa7134 |
| 8221 | F: drivers/media/pci/saa7134/ | 8222 | F: drivers/media/pci/saa7134/ |
| 8222 | 8223 | ||
| 8223 | SAA7146 VIDEO4LINUX-2 DRIVER | 8224 | SAA7146 VIDEO4LINUX-2 DRIVER |
| 8224 | M: Hans Verkuil <hverkuil@xs4all.nl> | 8225 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 8225 | L: linux-media@vger.kernel.org | 8226 | L: linux-media@vger.kernel.org |
| 8226 | T: git git://linuxtv.org/media_tree.git | 8227 | T: git git://linuxtv.org/media_tree.git |
| 8227 | S: Maintained | 8228 | S: Maintained |
| 8228 | F: drivers/media/common/saa7146/ | 8229 | F: drivers/media/common/saa7146/ |
| 8229 | F: drivers/media/pci/saa7146/ | 8230 | F: drivers/media/pci/saa7146/ |
| 8230 | F: include/media/saa7146* | 8231 | F: include/media/saa7146* |
| 8231 | 8232 | ||
| 8232 | SAMSUNG LAPTOP DRIVER | 8233 | SAMSUNG LAPTOP DRIVER |
| 8233 | M: Corentin Chary <corentin.chary@gmail.com> | 8234 | M: Corentin Chary <corentin.chary@gmail.com> |
| 8234 | L: platform-driver-x86@vger.kernel.org | 8235 | L: platform-driver-x86@vger.kernel.org |
| 8235 | S: Maintained | 8236 | S: Maintained |
| 8236 | F: drivers/platform/x86/samsung-laptop.c | 8237 | F: drivers/platform/x86/samsung-laptop.c |
| 8237 | 8238 | ||
| 8238 | SAMSUNG AUDIO (ASoC) DRIVERS | 8239 | SAMSUNG AUDIO (ASoC) DRIVERS |
| 8239 | M: Sangbeom Kim <sbkim73@samsung.com> | 8240 | M: Sangbeom Kim <sbkim73@samsung.com> |
| 8240 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 8241 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 8241 | S: Supported | 8242 | S: Supported |
| 8242 | F: sound/soc/samsung/ | 8243 | F: sound/soc/samsung/ |
| 8243 | 8244 | ||
| 8244 | SAMSUNG FRAMEBUFFER DRIVER | 8245 | SAMSUNG FRAMEBUFFER DRIVER |
| 8245 | M: Jingoo Han <jg1.han@samsung.com> | 8246 | M: Jingoo Han <jg1.han@samsung.com> |
| 8246 | L: linux-fbdev@vger.kernel.org | 8247 | L: linux-fbdev@vger.kernel.org |
| 8247 | S: Maintained | 8248 | S: Maintained |
| 8248 | F: drivers/video/fbdev/s3c-fb.c | 8249 | F: drivers/video/fbdev/s3c-fb.c |
| 8249 | 8250 | ||
| 8250 | SAMSUNG MULTIFUNCTION DEVICE DRIVERS | 8251 | SAMSUNG MULTIFUNCTION DEVICE DRIVERS |
| 8251 | M: Sangbeom Kim <sbkim73@samsung.com> | 8252 | M: Sangbeom Kim <sbkim73@samsung.com> |
| 8252 | L: linux-kernel@vger.kernel.org | 8253 | L: linux-kernel@vger.kernel.org |
| 8253 | S: Supported | 8254 | S: Supported |
| 8254 | F: drivers/mfd/sec*.c | 8255 | F: drivers/mfd/sec*.c |
| 8255 | F: drivers/regulator/s2m*.c | 8256 | F: drivers/regulator/s2m*.c |
| 8256 | F: drivers/regulator/s5m*.c | 8257 | F: drivers/regulator/s5m*.c |
| 8257 | F: include/linux/mfd/samsung/ | 8258 | F: include/linux/mfd/samsung/ |
| 8258 | 8259 | ||
| 8259 | SAMSUNG S5P/EXYNOS4 SOC SERIES CAMERA SUBSYSTEM DRIVERS | 8260 | SAMSUNG S5P/EXYNOS4 SOC SERIES CAMERA SUBSYSTEM DRIVERS |
| 8260 | M: Kyungmin Park <kyungmin.park@samsung.com> | 8261 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 8261 | M: Sylwester Nawrocki <s.nawrocki@samsung.com> | 8262 | M: Sylwester Nawrocki <s.nawrocki@samsung.com> |
| 8262 | L: linux-media@vger.kernel.org | 8263 | L: linux-media@vger.kernel.org |
| 8263 | Q: https://patchwork.linuxtv.org/project/linux-media/list/ | 8264 | Q: https://patchwork.linuxtv.org/project/linux-media/list/ |
| 8264 | S: Supported | 8265 | S: Supported |
| 8265 | F: drivers/media/platform/exynos4-is/ | 8266 | F: drivers/media/platform/exynos4-is/ |
| 8266 | 8267 | ||
| 8267 | SAMSUNG S3C24XX/S3C64XX SOC SERIES CAMIF DRIVER | 8268 | SAMSUNG S3C24XX/S3C64XX SOC SERIES CAMIF DRIVER |
| 8268 | M: Sylwester Nawrocki <sylvester.nawrocki@gmail.com> | 8269 | M: Sylwester Nawrocki <sylvester.nawrocki@gmail.com> |
| 8269 | L: linux-media@vger.kernel.org | 8270 | L: linux-media@vger.kernel.org |
| 8270 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) | 8271 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) |
| 8271 | S: Maintained | 8272 | S: Maintained |
| 8272 | F: drivers/media/platform/s3c-camif/ | 8273 | F: drivers/media/platform/s3c-camif/ |
| 8273 | F: include/media/s3c_camif.h | 8274 | F: include/media/s3c_camif.h |
| 8274 | 8275 | ||
| 8275 | SAMSUNG S5C73M3 CAMERA DRIVER | 8276 | SAMSUNG S5C73M3 CAMERA DRIVER |
| 8276 | M: Kyungmin Park <kyungmin.park@samsung.com> | 8277 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 8277 | M: Andrzej Hajda <a.hajda@samsung.com> | 8278 | M: Andrzej Hajda <a.hajda@samsung.com> |
| 8278 | L: linux-media@vger.kernel.org | 8279 | L: linux-media@vger.kernel.org |
| 8279 | S: Supported | 8280 | S: Supported |
| 8280 | F: drivers/media/i2c/s5c73m3/* | 8281 | F: drivers/media/i2c/s5c73m3/* |
| 8281 | 8282 | ||
| 8282 | SAMSUNG S5K5BAF CAMERA DRIVER | 8283 | SAMSUNG S5K5BAF CAMERA DRIVER |
| 8283 | M: Kyungmin Park <kyungmin.park@samsung.com> | 8284 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 8284 | M: Andrzej Hajda <a.hajda@samsung.com> | 8285 | M: Andrzej Hajda <a.hajda@samsung.com> |
| 8285 | L: linux-media@vger.kernel.org | 8286 | L: linux-media@vger.kernel.org |
| 8286 | S: Supported | 8287 | S: Supported |
| 8287 | F: drivers/media/i2c/s5k5baf.c | 8288 | F: drivers/media/i2c/s5k5baf.c |
| 8288 | 8289 | ||
| 8289 | SAMSUNG SOC CLOCK DRIVERS | 8290 | SAMSUNG SOC CLOCK DRIVERS |
| 8290 | M: Sylwester Nawrocki <s.nawrocki@samsung.com> | 8291 | M: Sylwester Nawrocki <s.nawrocki@samsung.com> |
| 8291 | M: Tomasz Figa <tomasz.figa@gmail.com> | 8292 | M: Tomasz Figa <tomasz.figa@gmail.com> |
| 8292 | S: Supported | 8293 | S: Supported |
| 8293 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) | 8294 | L: linux-samsung-soc@vger.kernel.org (moderated for non-subscribers) |
| 8294 | F: drivers/clk/samsung/ | 8295 | F: drivers/clk/samsung/ |
| 8295 | 8296 | ||
| 8296 | SAMSUNG SXGBE DRIVERS | 8297 | SAMSUNG SXGBE DRIVERS |
| 8297 | M: Byungho An <bh74.an@samsung.com> | 8298 | M: Byungho An <bh74.an@samsung.com> |
| 8298 | M: Girish K S <ks.giri@samsung.com> | 8299 | M: Girish K S <ks.giri@samsung.com> |
| 8299 | M: Vipul Pandya <vipul.pandya@samsung.com> | 8300 | M: Vipul Pandya <vipul.pandya@samsung.com> |
| 8300 | S: Supported | 8301 | S: Supported |
| 8301 | L: netdev@vger.kernel.org | 8302 | L: netdev@vger.kernel.org |
| 8302 | F: drivers/net/ethernet/samsung/sxgbe/ | 8303 | F: drivers/net/ethernet/samsung/sxgbe/ |
| 8303 | 8304 | ||
| 8304 | SAMSUNG USB2 PHY DRIVER | 8305 | SAMSUNG USB2 PHY DRIVER |
| 8305 | M: Kamil Debski <k.debski@samsung.com> | 8306 | M: Kamil Debski <k.debski@samsung.com> |
| 8306 | L: linux-kernel@vger.kernel.org | 8307 | L: linux-kernel@vger.kernel.org |
| 8307 | S: Supported | 8308 | S: Supported |
| 8308 | F: Documentation/devicetree/bindings/phy/samsung-phy.txt | 8309 | F: Documentation/devicetree/bindings/phy/samsung-phy.txt |
| 8309 | F: Documentation/phy/samsung-usb2.txt | 8310 | F: Documentation/phy/samsung-usb2.txt |
| 8310 | F: drivers/phy/phy-exynos4210-usb2.c | 8311 | F: drivers/phy/phy-exynos4210-usb2.c |
| 8311 | F: drivers/phy/phy-exynos4x12-usb2.c | 8312 | F: drivers/phy/phy-exynos4x12-usb2.c |
| 8312 | F: drivers/phy/phy-exynos5250-usb2.c | 8313 | F: drivers/phy/phy-exynos5250-usb2.c |
| 8313 | F: drivers/phy/phy-s5pv210-usb2.c | 8314 | F: drivers/phy/phy-s5pv210-usb2.c |
| 8314 | F: drivers/phy/phy-samsung-usb2.c | 8315 | F: drivers/phy/phy-samsung-usb2.c |
| 8315 | F: drivers/phy/phy-samsung-usb2.h | 8316 | F: drivers/phy/phy-samsung-usb2.h |
| 8316 | 8317 | ||
| 8317 | SERIAL DRIVERS | 8318 | SERIAL DRIVERS |
| 8318 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 8319 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 8319 | L: linux-serial@vger.kernel.org | 8320 | L: linux-serial@vger.kernel.org |
| 8320 | S: Maintained | 8321 | S: Maintained |
| 8321 | F: drivers/tty/serial/ | 8322 | F: drivers/tty/serial/ |
| 8322 | 8323 | ||
| 8323 | SYNOPSYS DESIGNWARE DMAC DRIVER | 8324 | SYNOPSYS DESIGNWARE DMAC DRIVER |
| 8324 | M: Viresh Kumar <viresh.linux@gmail.com> | 8325 | M: Viresh Kumar <viresh.linux@gmail.com> |
| 8325 | M: Andy Shevchenko <andriy.shevchenko@linux.intel.com> | 8326 | M: Andy Shevchenko <andriy.shevchenko@linux.intel.com> |
| 8326 | S: Maintained | 8327 | S: Maintained |
| 8327 | F: include/linux/platform_data/dma-dw.h | 8328 | F: include/linux/platform_data/dma-dw.h |
| 8328 | F: drivers/dma/dw/ | 8329 | F: drivers/dma/dw/ |
| 8329 | 8330 | ||
| 8330 | SYNOPSYS DESIGNWARE MMC/SD/SDIO DRIVER | 8331 | SYNOPSYS DESIGNWARE MMC/SD/SDIO DRIVER |
| 8331 | M: Seungwon Jeon <tgih.jun@samsung.com> | 8332 | M: Seungwon Jeon <tgih.jun@samsung.com> |
| 8332 | M: Jaehoon Chung <jh80.chung@samsung.com> | 8333 | M: Jaehoon Chung <jh80.chung@samsung.com> |
| 8333 | L: linux-mmc@vger.kernel.org | 8334 | L: linux-mmc@vger.kernel.org |
| 8334 | S: Maintained | 8335 | S: Maintained |
| 8335 | F: include/linux/mmc/dw_mmc.h | 8336 | F: include/linux/mmc/dw_mmc.h |
| 8336 | F: drivers/mmc/host/dw_mmc* | 8337 | F: drivers/mmc/host/dw_mmc* |
| 8337 | 8338 | ||
| 8338 | THUNDERBOLT DRIVER | 8339 | THUNDERBOLT DRIVER |
| 8339 | M: Andreas Noever <andreas.noever@gmail.com> | 8340 | M: Andreas Noever <andreas.noever@gmail.com> |
| 8340 | S: Maintained | 8341 | S: Maintained |
| 8341 | F: drivers/thunderbolt/ | 8342 | F: drivers/thunderbolt/ |
| 8342 | 8343 | ||
| 8343 | TIMEKEEPING, CLOCKSOURCE CORE, NTP | 8344 | TIMEKEEPING, CLOCKSOURCE CORE, NTP |
| 8344 | M: John Stultz <john.stultz@linaro.org> | 8345 | M: John Stultz <john.stultz@linaro.org> |
| 8345 | M: Thomas Gleixner <tglx@linutronix.de> | 8346 | M: Thomas Gleixner <tglx@linutronix.de> |
| 8346 | L: linux-kernel@vger.kernel.org | 8347 | L: linux-kernel@vger.kernel.org |
| 8347 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core | 8348 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git timers/core |
| 8348 | S: Supported | 8349 | S: Supported |
| 8349 | F: include/linux/clocksource.h | 8350 | F: include/linux/clocksource.h |
| 8350 | F: include/linux/time.h | 8351 | F: include/linux/time.h |
| 8351 | F: include/linux/timex.h | 8352 | F: include/linux/timex.h |
| 8352 | F: include/uapi/linux/time.h | 8353 | F: include/uapi/linux/time.h |
| 8353 | F: include/uapi/linux/timex.h | 8354 | F: include/uapi/linux/timex.h |
| 8354 | F: kernel/time/clocksource.c | 8355 | F: kernel/time/clocksource.c |
| 8355 | F: kernel/time/time*.c | 8356 | F: kernel/time/time*.c |
| 8356 | F: kernel/time/ntp.c | 8357 | F: kernel/time/ntp.c |
| 8357 | 8358 | ||
| 8358 | TLG2300 VIDEO4LINUX-2 DRIVER | 8359 | TLG2300 VIDEO4LINUX-2 DRIVER |
| 8359 | M: Huang Shijie <shijie8@gmail.com> | 8360 | M: Huang Shijie <shijie8@gmail.com> |
| 8360 | M: Hans Verkuil <hverkuil@xs4all.nl> | 8361 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 8361 | S: Odd Fixes | 8362 | S: Odd Fixes |
| 8362 | F: drivers/media/usb/tlg2300/ | 8363 | F: drivers/media/usb/tlg2300/ |
| 8363 | 8364 | ||
| 8364 | SC1200 WDT DRIVER | 8365 | SC1200 WDT DRIVER |
| 8365 | M: Zwane Mwaikambo <zwanem@gmail.com> | 8366 | M: Zwane Mwaikambo <zwanem@gmail.com> |
| 8366 | S: Maintained | 8367 | S: Maintained |
| 8367 | F: drivers/watchdog/sc1200wdt.c | 8368 | F: drivers/watchdog/sc1200wdt.c |
| 8368 | 8369 | ||
| 8369 | SCHEDULER | 8370 | SCHEDULER |
| 8370 | M: Ingo Molnar <mingo@redhat.com> | 8371 | M: Ingo Molnar <mingo@redhat.com> |
| 8371 | M: Peter Zijlstra <peterz@infradead.org> | 8372 | M: Peter Zijlstra <peterz@infradead.org> |
| 8372 | L: linux-kernel@vger.kernel.org | 8373 | L: linux-kernel@vger.kernel.org |
| 8373 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git sched/core | 8374 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git sched/core |
| 8374 | S: Maintained | 8375 | S: Maintained |
| 8375 | F: kernel/sched/ | 8376 | F: kernel/sched/ |
| 8376 | F: include/linux/sched.h | 8377 | F: include/linux/sched.h |
| 8377 | F: include/uapi/linux/sched.h | 8378 | F: include/uapi/linux/sched.h |
| 8378 | F: include/linux/wait.h | 8379 | F: include/linux/wait.h |
| 8379 | 8380 | ||
| 8380 | SCORE ARCHITECTURE | 8381 | SCORE ARCHITECTURE |
| 8381 | M: Chen Liqin <liqin.linux@gmail.com> | 8382 | M: Chen Liqin <liqin.linux@gmail.com> |
| 8382 | M: Lennox Wu <lennox.wu@gmail.com> | 8383 | M: Lennox Wu <lennox.wu@gmail.com> |
| 8383 | W: http://www.sunplus.com | 8384 | W: http://www.sunplus.com |
| 8384 | S: Supported | 8385 | S: Supported |
| 8385 | F: arch/score/ | 8386 | F: arch/score/ |
| 8386 | 8387 | ||
| 8387 | SCSI CDROM DRIVER | 8388 | SCSI CDROM DRIVER |
| 8388 | M: Jens Axboe <axboe@kernel.dk> | 8389 | M: Jens Axboe <axboe@kernel.dk> |
| 8389 | L: linux-scsi@vger.kernel.org | 8390 | L: linux-scsi@vger.kernel.org |
| 8390 | W: http://www.kernel.dk | 8391 | W: http://www.kernel.dk |
| 8391 | S: Maintained | 8392 | S: Maintained |
| 8392 | F: drivers/scsi/sr* | 8393 | F: drivers/scsi/sr* |
| 8393 | 8394 | ||
| 8394 | SCSI RDMA PROTOCOL (SRP) INITIATOR | 8395 | SCSI RDMA PROTOCOL (SRP) INITIATOR |
| 8395 | M: Bart Van Assche <bvanassche@acm.org> | 8396 | M: Bart Van Assche <bvanassche@acm.org> |
| 8396 | L: linux-rdma@vger.kernel.org | 8397 | L: linux-rdma@vger.kernel.org |
| 8397 | S: Supported | 8398 | S: Supported |
| 8398 | W: http://www.openfabrics.org | 8399 | W: http://www.openfabrics.org |
| 8399 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ | 8400 | Q: http://patchwork.kernel.org/project/linux-rdma/list/ |
| 8400 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/dad/srp-initiator.git | 8401 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/dad/srp-initiator.git |
| 8401 | F: drivers/infiniband/ulp/srp/ | 8402 | F: drivers/infiniband/ulp/srp/ |
| 8402 | F: include/scsi/srp.h | 8403 | F: include/scsi/srp.h |
| 8403 | 8404 | ||
| 8404 | SCSI SG DRIVER | 8405 | SCSI SG DRIVER |
| 8405 | M: Doug Gilbert <dgilbert@interlog.com> | 8406 | M: Doug Gilbert <dgilbert@interlog.com> |
| 8406 | L: linux-scsi@vger.kernel.org | 8407 | L: linux-scsi@vger.kernel.org |
| 8407 | W: http://sg.danny.cz/sg | 8408 | W: http://sg.danny.cz/sg |
| 8408 | S: Maintained | 8409 | S: Maintained |
| 8409 | F: Documentation/scsi/scsi-generic.txt | 8410 | F: Documentation/scsi/scsi-generic.txt |
| 8410 | F: drivers/scsi/sg.c | 8411 | F: drivers/scsi/sg.c |
| 8411 | F: include/scsi/sg.h | 8412 | F: include/scsi/sg.h |
| 8412 | 8413 | ||
| 8413 | SCSI SUBSYSTEM | 8414 | SCSI SUBSYSTEM |
| 8414 | M: "James E.J. Bottomley" <JBottomley@parallels.com> | 8415 | M: "James E.J. Bottomley" <JBottomley@parallels.com> |
| 8415 | L: linux-scsi@vger.kernel.org | 8416 | L: linux-scsi@vger.kernel.org |
| 8416 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6.git | 8417 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6.git |
| 8417 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6.git | 8418 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6.git |
| 8418 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-pending-2.6.git | 8419 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-pending-2.6.git |
| 8419 | S: Maintained | 8420 | S: Maintained |
| 8420 | F: drivers/scsi/ | 8421 | F: drivers/scsi/ |
| 8421 | F: include/scsi/ | 8422 | F: include/scsi/ |
| 8422 | 8423 | ||
| 8423 | SCSI TAPE DRIVER | 8424 | SCSI TAPE DRIVER |
| 8424 | M: Kai Mรคkisara <Kai.Makisara@kolumbus.fi> | 8425 | M: Kai Mรคkisara <Kai.Makisara@kolumbus.fi> |
| 8425 | L: linux-scsi@vger.kernel.org | 8426 | L: linux-scsi@vger.kernel.org |
| 8426 | S: Maintained | 8427 | S: Maintained |
| 8427 | F: Documentation/scsi/st.txt | 8428 | F: Documentation/scsi/st.txt |
| 8428 | F: drivers/scsi/st.* | 8429 | F: drivers/scsi/st.* |
| 8429 | F: drivers/scsi/st_*.h | 8430 | F: drivers/scsi/st_*.h |
| 8430 | 8431 | ||
| 8431 | SCTP PROTOCOL | 8432 | SCTP PROTOCOL |
| 8432 | M: Vlad Yasevich <vyasevich@gmail.com> | 8433 | M: Vlad Yasevich <vyasevich@gmail.com> |
| 8433 | M: Neil Horman <nhorman@tuxdriver.com> | 8434 | M: Neil Horman <nhorman@tuxdriver.com> |
| 8434 | L: linux-sctp@vger.kernel.org | 8435 | L: linux-sctp@vger.kernel.org |
| 8435 | W: http://lksctp.sourceforge.net | 8436 | W: http://lksctp.sourceforge.net |
| 8436 | S: Maintained | 8437 | S: Maintained |
| 8437 | F: Documentation/networking/sctp.txt | 8438 | F: Documentation/networking/sctp.txt |
| 8438 | F: include/linux/sctp.h | 8439 | F: include/linux/sctp.h |
| 8439 | F: include/uapi/linux/sctp.h | 8440 | F: include/uapi/linux/sctp.h |
| 8440 | F: include/net/sctp/ | 8441 | F: include/net/sctp/ |
| 8441 | F: net/sctp/ | 8442 | F: net/sctp/ |
| 8442 | 8443 | ||
| 8443 | SCx200 CPU SUPPORT | 8444 | SCx200 CPU SUPPORT |
| 8444 | M: Jim Cromie <jim.cromie@gmail.com> | 8445 | M: Jim Cromie <jim.cromie@gmail.com> |
| 8445 | S: Odd Fixes | 8446 | S: Odd Fixes |
| 8446 | F: Documentation/i2c/busses/scx200_acb | 8447 | F: Documentation/i2c/busses/scx200_acb |
| 8447 | F: arch/x86/platform/scx200/ | 8448 | F: arch/x86/platform/scx200/ |
| 8448 | F: drivers/watchdog/scx200_wdt.c | 8449 | F: drivers/watchdog/scx200_wdt.c |
| 8449 | F: drivers/i2c/busses/scx200* | 8450 | F: drivers/i2c/busses/scx200* |
| 8450 | F: drivers/mtd/maps/scx200_docflash.c | 8451 | F: drivers/mtd/maps/scx200_docflash.c |
| 8451 | F: include/linux/scx200.h | 8452 | F: include/linux/scx200.h |
| 8452 | 8453 | ||
| 8453 | SCx200 GPIO DRIVER | 8454 | SCx200 GPIO DRIVER |
| 8454 | M: Jim Cromie <jim.cromie@gmail.com> | 8455 | M: Jim Cromie <jim.cromie@gmail.com> |
| 8455 | S: Maintained | 8456 | S: Maintained |
| 8456 | F: drivers/char/scx200_gpio.c | 8457 | F: drivers/char/scx200_gpio.c |
| 8457 | F: include/linux/scx200_gpio.h | 8458 | F: include/linux/scx200_gpio.h |
| 8458 | 8459 | ||
| 8459 | SCx200 HRT CLOCKSOURCE DRIVER | 8460 | SCx200 HRT CLOCKSOURCE DRIVER |
| 8460 | M: Jim Cromie <jim.cromie@gmail.com> | 8461 | M: Jim Cromie <jim.cromie@gmail.com> |
| 8461 | S: Maintained | 8462 | S: Maintained |
| 8462 | F: drivers/clocksource/scx200_hrt.c | 8463 | F: drivers/clocksource/scx200_hrt.c |
| 8463 | 8464 | ||
| 8464 | SDRICOH_CS MMC/SD HOST CONTROLLER INTERFACE DRIVER | 8465 | SDRICOH_CS MMC/SD HOST CONTROLLER INTERFACE DRIVER |
| 8465 | M: Sascha Sommer <saschasommer@freenet.de> | 8466 | M: Sascha Sommer <saschasommer@freenet.de> |
| 8466 | L: sdricohcs-devel@lists.sourceforge.net (subscribers-only) | 8467 | L: sdricohcs-devel@lists.sourceforge.net (subscribers-only) |
| 8467 | S: Maintained | 8468 | S: Maintained |
| 8468 | F: drivers/mmc/host/sdricoh_cs.c | 8469 | F: drivers/mmc/host/sdricoh_cs.c |
| 8469 | 8470 | ||
| 8470 | SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) DRIVER | 8471 | SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) DRIVER |
| 8471 | M: Chris Ball <chris@printf.net> | 8472 | M: Chris Ball <chris@printf.net> |
| 8472 | L: linux-mmc@vger.kernel.org | 8473 | L: linux-mmc@vger.kernel.org |
| 8473 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc.git | 8474 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc.git |
| 8474 | S: Maintained | 8475 | S: Maintained |
| 8475 | F: drivers/mmc/host/sdhci.* | 8476 | F: drivers/mmc/host/sdhci.* |
| 8476 | F: drivers/mmc/host/sdhci-pltfm.[ch] | 8477 | F: drivers/mmc/host/sdhci-pltfm.[ch] |
| 8477 | 8478 | ||
| 8478 | SECURE COMPUTING | 8479 | SECURE COMPUTING |
| 8479 | M: Kees Cook <keescook@chromium.org> | 8480 | M: Kees Cook <keescook@chromium.org> |
| 8480 | R: Andy Lutomirski <luto@amacapital.net> | 8481 | R: Andy Lutomirski <luto@amacapital.net> |
| 8481 | R: Will Drewry <wad@chromium.org> | 8482 | R: Will Drewry <wad@chromium.org> |
| 8482 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git seccomp | 8483 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git seccomp |
| 8483 | S: Supported | 8484 | S: Supported |
| 8484 | F: kernel/seccomp.c | 8485 | F: kernel/seccomp.c |
| 8485 | F: include/uapi/linux/seccomp.h | 8486 | F: include/uapi/linux/seccomp.h |
| 8486 | F: include/linux/seccomp.h | 8487 | F: include/linux/seccomp.h |
| 8487 | K: \bsecure_computing | 8488 | K: \bsecure_computing |
| 8488 | K: \bTIF_SECCOMP\b | 8489 | K: \bTIF_SECCOMP\b |
| 8489 | 8490 | ||
| 8490 | SECURE DIGITAL HOST CONTROLLER INTERFACE, OPEN FIRMWARE BINDINGS (SDHCI-OF) | 8491 | SECURE DIGITAL HOST CONTROLLER INTERFACE, OPEN FIRMWARE BINDINGS (SDHCI-OF) |
| 8491 | M: Anton Vorontsov <anton@enomsg.org> | 8492 | M: Anton Vorontsov <anton@enomsg.org> |
| 8492 | L: linuxppc-dev@lists.ozlabs.org | 8493 | L: linuxppc-dev@lists.ozlabs.org |
| 8493 | L: linux-mmc@vger.kernel.org | 8494 | L: linux-mmc@vger.kernel.org |
| 8494 | S: Maintained | 8495 | S: Maintained |
| 8495 | F: drivers/mmc/host/sdhci-pltfm.[ch] | 8496 | F: drivers/mmc/host/sdhci-pltfm.[ch] |
| 8496 | 8497 | ||
| 8497 | SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) SAMSUNG DRIVER | 8498 | SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) SAMSUNG DRIVER |
| 8498 | M: Ben Dooks <ben-linux@fluff.org> | 8499 | M: Ben Dooks <ben-linux@fluff.org> |
| 8499 | L: linux-mmc@vger.kernel.org | 8500 | L: linux-mmc@vger.kernel.org |
| 8500 | S: Maintained | 8501 | S: Maintained |
| 8501 | F: drivers/mmc/host/sdhci-s3c.c | 8502 | F: drivers/mmc/host/sdhci-s3c.c |
| 8502 | 8503 | ||
| 8503 | SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) ST SPEAR DRIVER | 8504 | SECURE DIGITAL HOST CONTROLLER INTERFACE (SDHCI) ST SPEAR DRIVER |
| 8504 | M: Viresh Kumar <viresh.linux@gmail.com> | 8505 | M: Viresh Kumar <viresh.linux@gmail.com> |
| 8505 | L: spear-devel@list.st.com | 8506 | L: spear-devel@list.st.com |
| 8506 | L: linux-mmc@vger.kernel.org | 8507 | L: linux-mmc@vger.kernel.org |
| 8507 | S: Maintained | 8508 | S: Maintained |
| 8508 | F: drivers/mmc/host/sdhci-spear.c | 8509 | F: drivers/mmc/host/sdhci-spear.c |
| 8509 | 8510 | ||
| 8510 | SECURITY SUBSYSTEM | 8511 | SECURITY SUBSYSTEM |
| 8511 | M: James Morris <james.l.morris@oracle.com> | 8512 | M: James Morris <james.l.morris@oracle.com> |
| 8512 | M: Serge E. Hallyn <serge@hallyn.com> | 8513 | M: Serge E. Hallyn <serge@hallyn.com> |
| 8513 | L: linux-security-module@vger.kernel.org (suggested Cc:) | 8514 | L: linux-security-module@vger.kernel.org (suggested Cc:) |
| 8514 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security.git | 8515 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security.git |
| 8515 | W: http://kernsec.org/ | 8516 | W: http://kernsec.org/ |
| 8516 | S: Supported | 8517 | S: Supported |
| 8517 | F: security/ | 8518 | F: security/ |
| 8518 | 8519 | ||
| 8519 | SECURITY CONTACT | 8520 | SECURITY CONTACT |
| 8520 | M: Security Officers <security@kernel.org> | 8521 | M: Security Officers <security@kernel.org> |
| 8521 | S: Supported | 8522 | S: Supported |
| 8522 | 8523 | ||
| 8523 | SELINUX SECURITY MODULE | 8524 | SELINUX SECURITY MODULE |
| 8524 | M: Paul Moore <paul@paul-moore.com> | 8525 | M: Paul Moore <paul@paul-moore.com> |
| 8525 | M: Stephen Smalley <sds@tycho.nsa.gov> | 8526 | M: Stephen Smalley <sds@tycho.nsa.gov> |
| 8526 | M: Eric Paris <eparis@parisplace.org> | 8527 | M: Eric Paris <eparis@parisplace.org> |
| 8527 | L: selinux@tycho.nsa.gov (moderated for non-subscribers) | 8528 | L: selinux@tycho.nsa.gov (moderated for non-subscribers) |
| 8528 | W: http://selinuxproject.org | 8529 | W: http://selinuxproject.org |
| 8529 | T: git git://git.infradead.org/users/pcmoore/selinux | 8530 | T: git git://git.infradead.org/users/pcmoore/selinux |
| 8530 | S: Supported | 8531 | S: Supported |
| 8531 | F: include/linux/selinux* | 8532 | F: include/linux/selinux* |
| 8532 | F: security/selinux/ | 8533 | F: security/selinux/ |
| 8533 | F: scripts/selinux/ | 8534 | F: scripts/selinux/ |
| 8534 | 8535 | ||
| 8535 | APPARMOR SECURITY MODULE | 8536 | APPARMOR SECURITY MODULE |
| 8536 | M: John Johansen <john.johansen@canonical.com> | 8537 | M: John Johansen <john.johansen@canonical.com> |
| 8537 | L: apparmor@lists.ubuntu.com (subscribers-only, general discussion) | 8538 | L: apparmor@lists.ubuntu.com (subscribers-only, general discussion) |
| 8538 | W: apparmor.wiki.kernel.org | 8539 | W: apparmor.wiki.kernel.org |
| 8539 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jj/apparmor-dev.git | 8540 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jj/apparmor-dev.git |
| 8540 | S: Supported | 8541 | S: Supported |
| 8541 | F: security/apparmor/ | 8542 | F: security/apparmor/ |
| 8542 | 8543 | ||
| 8543 | SENSABLE PHANTOM | 8544 | SENSABLE PHANTOM |
| 8544 | M: Jiri Slaby <jirislaby@gmail.com> | 8545 | M: Jiri Slaby <jirislaby@gmail.com> |
| 8545 | S: Maintained | 8546 | S: Maintained |
| 8546 | F: drivers/misc/phantom.c | 8547 | F: drivers/misc/phantom.c |
| 8547 | F: include/uapi/linux/phantom.h | 8548 | F: include/uapi/linux/phantom.h |
| 8548 | 8549 | ||
| 8549 | SERIAL ATA (SATA) SUBSYSTEM | 8550 | SERIAL ATA (SATA) SUBSYSTEM |
| 8550 | M: Tejun Heo <tj@kernel.org> | 8551 | M: Tejun Heo <tj@kernel.org> |
| 8551 | L: linux-ide@vger.kernel.org | 8552 | L: linux-ide@vger.kernel.org |
| 8552 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata.git | 8553 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata.git |
| 8553 | S: Supported | 8554 | S: Supported |
| 8554 | F: drivers/ata/ | 8555 | F: drivers/ata/ |
| 8555 | F: include/linux/ata.h | 8556 | F: include/linux/ata.h |
| 8556 | F: include/linux/libata.h | 8557 | F: include/linux/libata.h |
| 8557 | 8558 | ||
| 8558 | SERIAL ATA AHCI PLATFORM devices support | 8559 | SERIAL ATA AHCI PLATFORM devices support |
| 8559 | M: Hans de Goede <hdegoede@redhat.com> | 8560 | M: Hans de Goede <hdegoede@redhat.com> |
| 8560 | M: Tejun Heo <tj@kernel.org> | 8561 | M: Tejun Heo <tj@kernel.org> |
| 8561 | L: linux-ide@vger.kernel.org | 8562 | L: linux-ide@vger.kernel.org |
| 8562 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata.git | 8563 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata.git |
| 8563 | S: Supported | 8564 | S: Supported |
| 8564 | F: drivers/ata/ahci_platform.c | 8565 | F: drivers/ata/ahci_platform.c |
| 8565 | F: drivers/ata/libahci_platform.c | 8566 | F: drivers/ata/libahci_platform.c |
| 8566 | F: include/linux/ahci_platform.h | 8567 | F: include/linux/ahci_platform.h |
| 8567 | 8568 | ||
| 8568 | SERVER ENGINES 10Gbps iSCSI - BladeEngine 2 DRIVER | 8569 | SERVER ENGINES 10Gbps iSCSI - BladeEngine 2 DRIVER |
| 8569 | M: Jayamohan Kallickal <jayamohan.kallickal@emulex.com> | 8570 | M: Jayamohan Kallickal <jayamohan.kallickal@emulex.com> |
| 8570 | L: linux-scsi@vger.kernel.org | 8571 | L: linux-scsi@vger.kernel.org |
| 8571 | W: http://www.emulex.com | 8572 | W: http://www.emulex.com |
| 8572 | S: Supported | 8573 | S: Supported |
| 8573 | F: drivers/scsi/be2iscsi/ | 8574 | F: drivers/scsi/be2iscsi/ |
| 8574 | 8575 | ||
| 8575 | SERVER ENGINES 10Gbps NIC - BladeEngine 2 DRIVER | 8576 | SERVER ENGINES 10Gbps NIC - BladeEngine 2 DRIVER |
| 8576 | M: Sathya Perla <sathya.perla@emulex.com> | 8577 | M: Sathya Perla <sathya.perla@emulex.com> |
| 8577 | M: Subbu Seetharaman <subbu.seetharaman@emulex.com> | 8578 | M: Subbu Seetharaman <subbu.seetharaman@emulex.com> |
| 8578 | M: Ajit Khaparde <ajit.khaparde@emulex.com> | 8579 | M: Ajit Khaparde <ajit.khaparde@emulex.com> |
| 8579 | L: netdev@vger.kernel.org | 8580 | L: netdev@vger.kernel.org |
| 8580 | W: http://www.emulex.com | 8581 | W: http://www.emulex.com |
| 8581 | S: Supported | 8582 | S: Supported |
| 8582 | F: drivers/net/ethernet/emulex/benet/ | 8583 | F: drivers/net/ethernet/emulex/benet/ |
| 8583 | 8584 | ||
| 8584 | SFC NETWORK DRIVER | 8585 | SFC NETWORK DRIVER |
| 8585 | M: Solarflare linux maintainers <linux-net-drivers@solarflare.com> | 8586 | M: Solarflare linux maintainers <linux-net-drivers@solarflare.com> |
| 8586 | M: Shradha Shah <sshah@solarflare.com> | 8587 | M: Shradha Shah <sshah@solarflare.com> |
| 8587 | L: netdev@vger.kernel.org | 8588 | L: netdev@vger.kernel.org |
| 8588 | S: Supported | 8589 | S: Supported |
| 8589 | F: drivers/net/ethernet/sfc/ | 8590 | F: drivers/net/ethernet/sfc/ |
| 8590 | 8591 | ||
| 8591 | SGI GRU DRIVER | 8592 | SGI GRU DRIVER |
| 8592 | M: Dimitri Sivanich <sivanich@sgi.com> | 8593 | M: Dimitri Sivanich <sivanich@sgi.com> |
| 8593 | S: Maintained | 8594 | S: Maintained |
| 8594 | F: drivers/misc/sgi-gru/ | 8595 | F: drivers/misc/sgi-gru/ |
| 8595 | 8596 | ||
| 8596 | SGI SN-IA64 (Altix) SERIAL CONSOLE DRIVER | 8597 | SGI SN-IA64 (Altix) SERIAL CONSOLE DRIVER |
| 8597 | M: Pat Gefre <pfg@sgi.com> | 8598 | M: Pat Gefre <pfg@sgi.com> |
| 8598 | L: linux-ia64@vger.kernel.org | 8599 | L: linux-ia64@vger.kernel.org |
| 8599 | S: Supported | 8600 | S: Supported |
| 8600 | F: Documentation/ia64/serial.txt | 8601 | F: Documentation/ia64/serial.txt |
| 8601 | F: drivers/tty/serial/ioc?_serial.c | 8602 | F: drivers/tty/serial/ioc?_serial.c |
| 8602 | F: include/linux/ioc?.h | 8603 | F: include/linux/ioc?.h |
| 8603 | 8604 | ||
| 8604 | SGI XP/XPC/XPNET DRIVER | 8605 | SGI XP/XPC/XPNET DRIVER |
| 8605 | M: Cliff Whickman <cpw@sgi.com> | 8606 | M: Cliff Whickman <cpw@sgi.com> |
| 8606 | M: Robin Holt <robinmholt@gmail.com> | 8607 | M: Robin Holt <robinmholt@gmail.com> |
| 8607 | S: Maintained | 8608 | S: Maintained |
| 8608 | F: drivers/misc/sgi-xp/ | 8609 | F: drivers/misc/sgi-xp/ |
| 8609 | 8610 | ||
| 8610 | SI2157 MEDIA DRIVER | 8611 | SI2157 MEDIA DRIVER |
| 8611 | M: Antti Palosaari <crope@iki.fi> | 8612 | M: Antti Palosaari <crope@iki.fi> |
| 8612 | L: linux-media@vger.kernel.org | 8613 | L: linux-media@vger.kernel.org |
| 8613 | W: http://linuxtv.org/ | 8614 | W: http://linuxtv.org/ |
| 8614 | W: http://palosaari.fi/linux/ | 8615 | W: http://palosaari.fi/linux/ |
| 8615 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8616 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8616 | T: git git://linuxtv.org/anttip/media_tree.git | 8617 | T: git git://linuxtv.org/anttip/media_tree.git |
| 8617 | S: Maintained | 8618 | S: Maintained |
| 8618 | F: drivers/media/tuners/si2157* | 8619 | F: drivers/media/tuners/si2157* |
| 8619 | 8620 | ||
| 8620 | SI2168 MEDIA DRIVER | 8621 | SI2168 MEDIA DRIVER |
| 8621 | M: Antti Palosaari <crope@iki.fi> | 8622 | M: Antti Palosaari <crope@iki.fi> |
| 8622 | L: linux-media@vger.kernel.org | 8623 | L: linux-media@vger.kernel.org |
| 8623 | W: http://linuxtv.org/ | 8624 | W: http://linuxtv.org/ |
| 8624 | W: http://palosaari.fi/linux/ | 8625 | W: http://palosaari.fi/linux/ |
| 8625 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8626 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8626 | T: git git://linuxtv.org/anttip/media_tree.git | 8627 | T: git git://linuxtv.org/anttip/media_tree.git |
| 8627 | S: Maintained | 8628 | S: Maintained |
| 8628 | F: drivers/media/dvb-frontends/si2168* | 8629 | F: drivers/media/dvb-frontends/si2168* |
| 8629 | 8630 | ||
| 8630 | SI470X FM RADIO RECEIVER I2C DRIVER | 8631 | SI470X FM RADIO RECEIVER I2C DRIVER |
| 8631 | M: Hans Verkuil <hverkuil@xs4all.nl> | 8632 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 8632 | L: linux-media@vger.kernel.org | 8633 | L: linux-media@vger.kernel.org |
| 8633 | T: git git://linuxtv.org/media_tree.git | 8634 | T: git git://linuxtv.org/media_tree.git |
| 8634 | W: http://linuxtv.org | 8635 | W: http://linuxtv.org |
| 8635 | S: Odd Fixes | 8636 | S: Odd Fixes |
| 8636 | F: drivers/media/radio/si470x/radio-si470x-i2c.c | 8637 | F: drivers/media/radio/si470x/radio-si470x-i2c.c |
| 8637 | 8638 | ||
| 8638 | SI470X FM RADIO RECEIVER USB DRIVER | 8639 | SI470X FM RADIO RECEIVER USB DRIVER |
| 8639 | M: Hans Verkuil <hverkuil@xs4all.nl> | 8640 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 8640 | L: linux-media@vger.kernel.org | 8641 | L: linux-media@vger.kernel.org |
| 8641 | T: git git://linuxtv.org/media_tree.git | 8642 | T: git git://linuxtv.org/media_tree.git |
| 8642 | W: http://linuxtv.org | 8643 | W: http://linuxtv.org |
| 8643 | S: Maintained | 8644 | S: Maintained |
| 8644 | F: drivers/media/radio/si470x/radio-si470x-common.c | 8645 | F: drivers/media/radio/si470x/radio-si470x-common.c |
| 8645 | F: drivers/media/radio/si470x/radio-si470x.h | 8646 | F: drivers/media/radio/si470x/radio-si470x.h |
| 8646 | F: drivers/media/radio/si470x/radio-si470x-usb.c | 8647 | F: drivers/media/radio/si470x/radio-si470x-usb.c |
| 8647 | 8648 | ||
| 8648 | SI4713 FM RADIO TRANSMITTER I2C DRIVER | 8649 | SI4713 FM RADIO TRANSMITTER I2C DRIVER |
| 8649 | M: Eduardo Valentin <edubezval@gmail.com> | 8650 | M: Eduardo Valentin <edubezval@gmail.com> |
| 8650 | L: linux-media@vger.kernel.org | 8651 | L: linux-media@vger.kernel.org |
| 8651 | T: git git://linuxtv.org/media_tree.git | 8652 | T: git git://linuxtv.org/media_tree.git |
| 8652 | W: http://linuxtv.org | 8653 | W: http://linuxtv.org |
| 8653 | S: Odd Fixes | 8654 | S: Odd Fixes |
| 8654 | F: drivers/media/radio/si4713/si4713.? | 8655 | F: drivers/media/radio/si4713/si4713.? |
| 8655 | 8656 | ||
| 8656 | SI4713 FM RADIO TRANSMITTER PLATFORM DRIVER | 8657 | SI4713 FM RADIO TRANSMITTER PLATFORM DRIVER |
| 8657 | M: Eduardo Valentin <edubezval@gmail.com> | 8658 | M: Eduardo Valentin <edubezval@gmail.com> |
| 8658 | L: linux-media@vger.kernel.org | 8659 | L: linux-media@vger.kernel.org |
| 8659 | T: git git://linuxtv.org/media_tree.git | 8660 | T: git git://linuxtv.org/media_tree.git |
| 8660 | W: http://linuxtv.org | 8661 | W: http://linuxtv.org |
| 8661 | S: Odd Fixes | 8662 | S: Odd Fixes |
| 8662 | F: drivers/media/radio/si4713/radio-platform-si4713.c | 8663 | F: drivers/media/radio/si4713/radio-platform-si4713.c |
| 8663 | 8664 | ||
| 8664 | SI4713 FM RADIO TRANSMITTER USB DRIVER | 8665 | SI4713 FM RADIO TRANSMITTER USB DRIVER |
| 8665 | M: Hans Verkuil <hverkuil@xs4all.nl> | 8666 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 8666 | L: linux-media@vger.kernel.org | 8667 | L: linux-media@vger.kernel.org |
| 8667 | T: git git://linuxtv.org/media_tree.git | 8668 | T: git git://linuxtv.org/media_tree.git |
| 8668 | W: http://linuxtv.org | 8669 | W: http://linuxtv.org |
| 8669 | S: Maintained | 8670 | S: Maintained |
| 8670 | F: drivers/media/radio/si4713/radio-usb-si4713.c | 8671 | F: drivers/media/radio/si4713/radio-usb-si4713.c |
| 8671 | 8672 | ||
| 8672 | SIANO DVB DRIVER | 8673 | SIANO DVB DRIVER |
| 8673 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 8674 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 8674 | L: linux-media@vger.kernel.org | 8675 | L: linux-media@vger.kernel.org |
| 8675 | W: http://linuxtv.org | 8676 | W: http://linuxtv.org |
| 8676 | T: git git://linuxtv.org/media_tree.git | 8677 | T: git git://linuxtv.org/media_tree.git |
| 8677 | S: Odd fixes | 8678 | S: Odd fixes |
| 8678 | F: drivers/media/common/siano/ | 8679 | F: drivers/media/common/siano/ |
| 8679 | F: drivers/media/usb/siano/ | 8680 | F: drivers/media/usb/siano/ |
| 8680 | F: drivers/media/usb/siano/ | 8681 | F: drivers/media/usb/siano/ |
| 8681 | F: drivers/media/mmc/siano/ | 8682 | F: drivers/media/mmc/siano/ |
| 8682 | 8683 | ||
| 8683 | SIMPLEFB FB DRIVER | 8684 | SIMPLEFB FB DRIVER |
| 8684 | M: Hans de Goede <hdegoede@redhat.com> | 8685 | M: Hans de Goede <hdegoede@redhat.com> |
| 8685 | L: linux-fbdev@vger.kernel.org | 8686 | L: linux-fbdev@vger.kernel.org |
| 8686 | S: Maintained | 8687 | S: Maintained |
| 8687 | F: Documentation/devicetree/bindings/video/simple-framebuffer.txt | 8688 | F: Documentation/devicetree/bindings/video/simple-framebuffer.txt |
| 8688 | F: drivers/video/fbdev/simplefb.c | 8689 | F: drivers/video/fbdev/simplefb.c |
| 8689 | F: include/linux/platform_data/simplefb.h | 8690 | F: include/linux/platform_data/simplefb.h |
| 8690 | 8691 | ||
| 8691 | SH_VEU V4L2 MEM2MEM DRIVER | 8692 | SH_VEU V4L2 MEM2MEM DRIVER |
| 8692 | L: linux-media@vger.kernel.org | 8693 | L: linux-media@vger.kernel.org |
| 8693 | S: Orphan | 8694 | S: Orphan |
| 8694 | F: drivers/media/platform/sh_veu.c | 8695 | F: drivers/media/platform/sh_veu.c |
| 8695 | 8696 | ||
| 8696 | SH_VOU V4L2 OUTPUT DRIVER | 8697 | SH_VOU V4L2 OUTPUT DRIVER |
| 8697 | L: linux-media@vger.kernel.org | 8698 | L: linux-media@vger.kernel.org |
| 8698 | S: Orphan | 8699 | S: Orphan |
| 8699 | F: drivers/media/platform/sh_vou.c | 8700 | F: drivers/media/platform/sh_vou.c |
| 8700 | F: include/media/sh_vou.h | 8701 | F: include/media/sh_vou.h |
| 8701 | 8702 | ||
| 8702 | SIMPLE FIRMWARE INTERFACE (SFI) | 8703 | SIMPLE FIRMWARE INTERFACE (SFI) |
| 8703 | M: Len Brown <lenb@kernel.org> | 8704 | M: Len Brown <lenb@kernel.org> |
| 8704 | L: sfi-devel@simplefirmware.org | 8705 | L: sfi-devel@simplefirmware.org |
| 8705 | W: http://simplefirmware.org/ | 8706 | W: http://simplefirmware.org/ |
| 8706 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-sfi-2.6.git | 8707 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-sfi-2.6.git |
| 8707 | S: Supported | 8708 | S: Supported |
| 8708 | F: arch/x86/platform/sfi/ | 8709 | F: arch/x86/platform/sfi/ |
| 8709 | F: drivers/sfi/ | 8710 | F: drivers/sfi/ |
| 8710 | F: include/linux/sfi*.h | 8711 | F: include/linux/sfi*.h |
| 8711 | 8712 | ||
| 8712 | SIMTEC EB110ATX (Chalice CATS) | 8713 | SIMTEC EB110ATX (Chalice CATS) |
| 8713 | P: Ben Dooks | 8714 | P: Ben Dooks |
| 8714 | P: Vincent Sanders <vince@simtec.co.uk> | 8715 | P: Vincent Sanders <vince@simtec.co.uk> |
| 8715 | M: Simtec Linux Team <linux@simtec.co.uk> | 8716 | M: Simtec Linux Team <linux@simtec.co.uk> |
| 8716 | W: http://www.simtec.co.uk/products/EB110ATX/ | 8717 | W: http://www.simtec.co.uk/products/EB110ATX/ |
| 8717 | S: Supported | 8718 | S: Supported |
| 8718 | 8719 | ||
| 8719 | SIMTEC EB2410ITX (BAST) | 8720 | SIMTEC EB2410ITX (BAST) |
| 8720 | P: Ben Dooks | 8721 | P: Ben Dooks |
| 8721 | P: Vincent Sanders <vince@simtec.co.uk> | 8722 | P: Vincent Sanders <vince@simtec.co.uk> |
| 8722 | M: Simtec Linux Team <linux@simtec.co.uk> | 8723 | M: Simtec Linux Team <linux@simtec.co.uk> |
| 8723 | W: http://www.simtec.co.uk/products/EB2410ITX/ | 8724 | W: http://www.simtec.co.uk/products/EB2410ITX/ |
| 8724 | S: Supported | 8725 | S: Supported |
| 8725 | F: arch/arm/mach-s3c24xx/mach-bast.c | 8726 | F: arch/arm/mach-s3c24xx/mach-bast.c |
| 8726 | F: arch/arm/mach-s3c24xx/bast-ide.c | 8727 | F: arch/arm/mach-s3c24xx/bast-ide.c |
| 8727 | F: arch/arm/mach-s3c24xx/bast-irq.c | 8728 | F: arch/arm/mach-s3c24xx/bast-irq.c |
| 8728 | 8729 | ||
| 8729 | TI DAVINCI MACHINE SUPPORT | 8730 | TI DAVINCI MACHINE SUPPORT |
| 8730 | M: Sekhar Nori <nsekhar@ti.com> | 8731 | M: Sekhar Nori <nsekhar@ti.com> |
| 8731 | M: Kevin Hilman <khilman@deeprootsystems.com> | 8732 | M: Kevin Hilman <khilman@deeprootsystems.com> |
| 8732 | T: git git://gitorious.org/linux-davinci/linux-davinci.git | 8733 | T: git git://gitorious.org/linux-davinci/linux-davinci.git |
| 8733 | Q: http://patchwork.kernel.org/project/linux-davinci/list/ | 8734 | Q: http://patchwork.kernel.org/project/linux-davinci/list/ |
| 8734 | S: Supported | 8735 | S: Supported |
| 8735 | F: arch/arm/mach-davinci/ | 8736 | F: arch/arm/mach-davinci/ |
| 8736 | F: drivers/i2c/busses/i2c-davinci.c | 8737 | F: drivers/i2c/busses/i2c-davinci.c |
| 8737 | 8738 | ||
| 8738 | TI DAVINCI SERIES MEDIA DRIVER | 8739 | TI DAVINCI SERIES MEDIA DRIVER |
| 8739 | M: Lad, Prabhakar <prabhakar.csengg@gmail.com> | 8740 | M: Lad, Prabhakar <prabhakar.csengg@gmail.com> |
| 8740 | L: linux-media@vger.kernel.org | 8741 | L: linux-media@vger.kernel.org |
| 8741 | W: http://linuxtv.org/ | 8742 | W: http://linuxtv.org/ |
| 8742 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8743 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8743 | T: git git://linuxtv.org/mhadli/v4l-dvb-davinci_devices.git | 8744 | T: git git://linuxtv.org/mhadli/v4l-dvb-davinci_devices.git |
| 8744 | S: Maintained | 8745 | S: Maintained |
| 8745 | F: drivers/media/platform/davinci/ | 8746 | F: drivers/media/platform/davinci/ |
| 8746 | F: include/media/davinci/ | 8747 | F: include/media/davinci/ |
| 8747 | 8748 | ||
| 8748 | SIS 190 ETHERNET DRIVER | 8749 | SIS 190 ETHERNET DRIVER |
| 8749 | M: Francois Romieu <romieu@fr.zoreil.com> | 8750 | M: Francois Romieu <romieu@fr.zoreil.com> |
| 8750 | L: netdev@vger.kernel.org | 8751 | L: netdev@vger.kernel.org |
| 8751 | S: Maintained | 8752 | S: Maintained |
| 8752 | F: drivers/net/ethernet/sis/sis190.c | 8753 | F: drivers/net/ethernet/sis/sis190.c |
| 8753 | 8754 | ||
| 8754 | SIS 900/7016 FAST ETHERNET DRIVER | 8755 | SIS 900/7016 FAST ETHERNET DRIVER |
| 8755 | M: Daniele Venzano <venza@brownhat.org> | 8756 | M: Daniele Venzano <venza@brownhat.org> |
| 8756 | W: http://www.brownhat.org/sis900.html | 8757 | W: http://www.brownhat.org/sis900.html |
| 8757 | L: netdev@vger.kernel.org | 8758 | L: netdev@vger.kernel.org |
| 8758 | S: Maintained | 8759 | S: Maintained |
| 8759 | F: drivers/net/ethernet/sis/sis900.* | 8760 | F: drivers/net/ethernet/sis/sis900.* |
| 8760 | 8761 | ||
| 8761 | SIS FRAMEBUFFER DRIVER | 8762 | SIS FRAMEBUFFER DRIVER |
| 8762 | M: Thomas Winischhofer <thomas@winischhofer.net> | 8763 | M: Thomas Winischhofer <thomas@winischhofer.net> |
| 8763 | W: http://www.winischhofer.net/linuxsisvga.shtml | 8764 | W: http://www.winischhofer.net/linuxsisvga.shtml |
| 8764 | S: Maintained | 8765 | S: Maintained |
| 8765 | F: Documentation/fb/sisfb.txt | 8766 | F: Documentation/fb/sisfb.txt |
| 8766 | F: drivers/video/fbdev/sis/ | 8767 | F: drivers/video/fbdev/sis/ |
| 8767 | F: include/video/sisfb.h | 8768 | F: include/video/sisfb.h |
| 8768 | 8769 | ||
| 8769 | SIS USB2VGA DRIVER | 8770 | SIS USB2VGA DRIVER |
| 8770 | M: Thomas Winischhofer <thomas@winischhofer.net> | 8771 | M: Thomas Winischhofer <thomas@winischhofer.net> |
| 8771 | W: http://www.winischhofer.at/linuxsisusbvga.shtml | 8772 | W: http://www.winischhofer.at/linuxsisusbvga.shtml |
| 8772 | S: Maintained | 8773 | S: Maintained |
| 8773 | F: drivers/usb/misc/sisusbvga/ | 8774 | F: drivers/usb/misc/sisusbvga/ |
| 8774 | 8775 | ||
| 8775 | SLAB ALLOCATOR | 8776 | SLAB ALLOCATOR |
| 8776 | M: Christoph Lameter <cl@linux.com> | 8777 | M: Christoph Lameter <cl@linux.com> |
| 8777 | M: Pekka Enberg <penberg@kernel.org> | 8778 | M: Pekka Enberg <penberg@kernel.org> |
| 8778 | M: David Rientjes <rientjes@google.com> | 8779 | M: David Rientjes <rientjes@google.com> |
| 8779 | M: Joonsoo Kim <iamjoonsoo.kim@lge.com> | 8780 | M: Joonsoo Kim <iamjoonsoo.kim@lge.com> |
| 8780 | M: Andrew Morton <akpm@linux-foundation.org> | 8781 | M: Andrew Morton <akpm@linux-foundation.org> |
| 8781 | L: linux-mm@kvack.org | 8782 | L: linux-mm@kvack.org |
| 8782 | S: Maintained | 8783 | S: Maintained |
| 8783 | F: include/linux/sl?b*.h | 8784 | F: include/linux/sl?b*.h |
| 8784 | F: mm/sl?b* | 8785 | F: mm/sl?b* |
| 8785 | 8786 | ||
| 8786 | SLEEPABLE READ-COPY UPDATE (SRCU) | 8787 | SLEEPABLE READ-COPY UPDATE (SRCU) |
| 8787 | M: Lai Jiangshan <laijs@cn.fujitsu.com> | 8788 | M: Lai Jiangshan <laijs@cn.fujitsu.com> |
| 8788 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> | 8789 | M: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
| 8789 | M: Josh Triplett <josh@joshtriplett.org> | 8790 | M: Josh Triplett <josh@joshtriplett.org> |
| 8790 | R: Steven Rostedt <rostedt@goodmis.org> | 8791 | R: Steven Rostedt <rostedt@goodmis.org> |
| 8791 | R: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> | 8792 | R: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> |
| 8792 | L: linux-kernel@vger.kernel.org | 8793 | L: linux-kernel@vger.kernel.org |
| 8793 | W: http://www.rdrop.com/users/paulmck/RCU/ | 8794 | W: http://www.rdrop.com/users/paulmck/RCU/ |
| 8794 | S: Supported | 8795 | S: Supported |
| 8795 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git | 8796 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git |
| 8796 | F: include/linux/srcu.h | 8797 | F: include/linux/srcu.h |
| 8797 | F: kernel/rcu/srcu.c | 8798 | F: kernel/rcu/srcu.c |
| 8798 | 8799 | ||
| 8799 | SMACK SECURITY MODULE | 8800 | SMACK SECURITY MODULE |
| 8800 | M: Casey Schaufler <casey@schaufler-ca.com> | 8801 | M: Casey Schaufler <casey@schaufler-ca.com> |
| 8801 | L: linux-security-module@vger.kernel.org | 8802 | L: linux-security-module@vger.kernel.org |
| 8802 | W: http://schaufler-ca.com | 8803 | W: http://schaufler-ca.com |
| 8803 | T: git git://git.gitorious.org/smack-next/kernel.git | 8804 | T: git git://git.gitorious.org/smack-next/kernel.git |
| 8804 | S: Maintained | 8805 | S: Maintained |
| 8805 | F: Documentation/security/Smack.txt | 8806 | F: Documentation/security/Smack.txt |
| 8806 | F: security/smack/ | 8807 | F: security/smack/ |
| 8807 | 8808 | ||
| 8808 | DRIVERS FOR ADAPTIVE VOLTAGE SCALING (AVS) | 8809 | DRIVERS FOR ADAPTIVE VOLTAGE SCALING (AVS) |
| 8809 | M: Kevin Hilman <khilman@kernel.org> | 8810 | M: Kevin Hilman <khilman@kernel.org> |
| 8810 | M: Nishanth Menon <nm@ti.com> | 8811 | M: Nishanth Menon <nm@ti.com> |
| 8811 | S: Maintained | 8812 | S: Maintained |
| 8812 | F: drivers/power/avs/ | 8813 | F: drivers/power/avs/ |
| 8813 | F: include/linux/power/smartreflex.h | 8814 | F: include/linux/power/smartreflex.h |
| 8814 | L: linux-pm@vger.kernel.org | 8815 | L: linux-pm@vger.kernel.org |
| 8815 | 8816 | ||
| 8816 | SMC91x ETHERNET DRIVER | 8817 | SMC91x ETHERNET DRIVER |
| 8817 | M: Nicolas Pitre <nico@fluxnic.net> | 8818 | M: Nicolas Pitre <nico@fluxnic.net> |
| 8818 | S: Odd Fixes | 8819 | S: Odd Fixes |
| 8819 | F: drivers/net/ethernet/smsc/smc91x.* | 8820 | F: drivers/net/ethernet/smsc/smc91x.* |
| 8820 | 8821 | ||
| 8821 | SMIA AND SMIA++ IMAGE SENSOR DRIVER | 8822 | SMIA AND SMIA++ IMAGE SENSOR DRIVER |
| 8822 | M: Sakari Ailus <sakari.ailus@iki.fi> | 8823 | M: Sakari Ailus <sakari.ailus@iki.fi> |
| 8823 | L: linux-media@vger.kernel.org | 8824 | L: linux-media@vger.kernel.org |
| 8824 | S: Maintained | 8825 | S: Maintained |
| 8825 | F: drivers/media/i2c/smiapp/ | 8826 | F: drivers/media/i2c/smiapp/ |
| 8826 | F: include/media/smiapp.h | 8827 | F: include/media/smiapp.h |
| 8827 | F: drivers/media/i2c/smiapp-pll.c | 8828 | F: drivers/media/i2c/smiapp-pll.c |
| 8828 | F: drivers/media/i2c/smiapp-pll.h | 8829 | F: drivers/media/i2c/smiapp-pll.h |
| 8829 | 8830 | ||
| 8830 | SMM665 HARDWARE MONITOR DRIVER | 8831 | SMM665 HARDWARE MONITOR DRIVER |
| 8831 | M: Guenter Roeck <linux@roeck-us.net> | 8832 | M: Guenter Roeck <linux@roeck-us.net> |
| 8832 | L: lm-sensors@lm-sensors.org | 8833 | L: lm-sensors@lm-sensors.org |
| 8833 | S: Maintained | 8834 | S: Maintained |
| 8834 | F: Documentation/hwmon/smm665 | 8835 | F: Documentation/hwmon/smm665 |
| 8835 | F: drivers/hwmon/smm665.c | 8836 | F: drivers/hwmon/smm665.c |
| 8836 | 8837 | ||
| 8837 | SMSC EMC2103 HARDWARE MONITOR DRIVER | 8838 | SMSC EMC2103 HARDWARE MONITOR DRIVER |
| 8838 | M: Steve Glendinning <steve.glendinning@shawell.net> | 8839 | M: Steve Glendinning <steve.glendinning@shawell.net> |
| 8839 | L: lm-sensors@lm-sensors.org | 8840 | L: lm-sensors@lm-sensors.org |
| 8840 | S: Maintained | 8841 | S: Maintained |
| 8841 | F: Documentation/hwmon/emc2103 | 8842 | F: Documentation/hwmon/emc2103 |
| 8842 | F: drivers/hwmon/emc2103.c | 8843 | F: drivers/hwmon/emc2103.c |
| 8843 | 8844 | ||
| 8844 | SMSC SCH5627 HARDWARE MONITOR DRIVER | 8845 | SMSC SCH5627 HARDWARE MONITOR DRIVER |
| 8845 | M: Hans de Goede <hdegoede@redhat.com> | 8846 | M: Hans de Goede <hdegoede@redhat.com> |
| 8846 | L: lm-sensors@lm-sensors.org | 8847 | L: lm-sensors@lm-sensors.org |
| 8847 | S: Supported | 8848 | S: Supported |
| 8848 | F: Documentation/hwmon/sch5627 | 8849 | F: Documentation/hwmon/sch5627 |
| 8849 | F: drivers/hwmon/sch5627.c | 8850 | F: drivers/hwmon/sch5627.c |
| 8850 | 8851 | ||
| 8851 | SMSC47B397 HARDWARE MONITOR DRIVER | 8852 | SMSC47B397 HARDWARE MONITOR DRIVER |
| 8852 | M: Jean Delvare <jdelvare@suse.de> | 8853 | M: Jean Delvare <jdelvare@suse.de> |
| 8853 | L: lm-sensors@lm-sensors.org | 8854 | L: lm-sensors@lm-sensors.org |
| 8854 | S: Maintained | 8855 | S: Maintained |
| 8855 | F: Documentation/hwmon/smsc47b397 | 8856 | F: Documentation/hwmon/smsc47b397 |
| 8856 | F: drivers/hwmon/smsc47b397.c | 8857 | F: drivers/hwmon/smsc47b397.c |
| 8857 | 8858 | ||
| 8858 | SMSC911x ETHERNET DRIVER | 8859 | SMSC911x ETHERNET DRIVER |
| 8859 | M: Steve Glendinning <steve.glendinning@shawell.net> | 8860 | M: Steve Glendinning <steve.glendinning@shawell.net> |
| 8860 | L: netdev@vger.kernel.org | 8861 | L: netdev@vger.kernel.org |
| 8861 | S: Maintained | 8862 | S: Maintained |
| 8862 | F: include/linux/smsc911x.h | 8863 | F: include/linux/smsc911x.h |
| 8863 | F: drivers/net/ethernet/smsc/smsc911x.* | 8864 | F: drivers/net/ethernet/smsc/smsc911x.* |
| 8864 | 8865 | ||
| 8865 | SMSC9420 PCI ETHERNET DRIVER | 8866 | SMSC9420 PCI ETHERNET DRIVER |
| 8866 | M: Steve Glendinning <steve.glendinning@shawell.net> | 8867 | M: Steve Glendinning <steve.glendinning@shawell.net> |
| 8867 | L: netdev@vger.kernel.org | 8868 | L: netdev@vger.kernel.org |
| 8868 | S: Maintained | 8869 | S: Maintained |
| 8869 | F: drivers/net/ethernet/smsc/smsc9420.* | 8870 | F: drivers/net/ethernet/smsc/smsc9420.* |
| 8870 | 8871 | ||
| 8871 | SMSC UFX6000 and UFX7000 USB to VGA DRIVER | 8872 | SMSC UFX6000 and UFX7000 USB to VGA DRIVER |
| 8872 | M: Steve Glendinning <steve.glendinning@shawell.net> | 8873 | M: Steve Glendinning <steve.glendinning@shawell.net> |
| 8873 | L: linux-fbdev@vger.kernel.org | 8874 | L: linux-fbdev@vger.kernel.org |
| 8874 | S: Maintained | 8875 | S: Maintained |
| 8875 | F: drivers/video/fbdev/smscufx.c | 8876 | F: drivers/video/fbdev/smscufx.c |
| 8876 | 8877 | ||
| 8877 | SOC-CAMERA V4L2 SUBSYSTEM | 8878 | SOC-CAMERA V4L2 SUBSYSTEM |
| 8878 | M: Guennadi Liakhovetski <g.liakhovetski@gmx.de> | 8879 | M: Guennadi Liakhovetski <g.liakhovetski@gmx.de> |
| 8879 | L: linux-media@vger.kernel.org | 8880 | L: linux-media@vger.kernel.org |
| 8880 | T: git git://linuxtv.org/media_tree.git | 8881 | T: git git://linuxtv.org/media_tree.git |
| 8881 | S: Maintained | 8882 | S: Maintained |
| 8882 | F: include/media/soc* | 8883 | F: include/media/soc* |
| 8883 | F: drivers/media/i2c/soc_camera/ | 8884 | F: drivers/media/i2c/soc_camera/ |
| 8884 | F: drivers/media/platform/soc_camera/ | 8885 | F: drivers/media/platform/soc_camera/ |
| 8885 | 8886 | ||
| 8886 | SOEKRIS NET48XX LED SUPPORT | 8887 | SOEKRIS NET48XX LED SUPPORT |
| 8887 | M: Chris Boot <bootc@bootc.net> | 8888 | M: Chris Boot <bootc@bootc.net> |
| 8888 | S: Maintained | 8889 | S: Maintained |
| 8889 | F: drivers/leds/leds-net48xx.c | 8890 | F: drivers/leds/leds-net48xx.c |
| 8890 | 8891 | ||
| 8891 | SOFTLOGIC 6x10 MPEG CODEC | 8892 | SOFTLOGIC 6x10 MPEG CODEC |
| 8892 | M: Bluecherry Maintainers <maintainers@bluecherrydvr.com> | 8893 | M: Bluecherry Maintainers <maintainers@bluecherrydvr.com> |
| 8893 | M: Andrey Utkin <andrey.utkin@corp.bluecherry.net> | 8894 | M: Andrey Utkin <andrey.utkin@corp.bluecherry.net> |
| 8894 | M: Andrey Utkin <andrey.krieger.utkin@gmail.com> | 8895 | M: Andrey Utkin <andrey.krieger.utkin@gmail.com> |
| 8895 | L: linux-media@vger.kernel.org | 8896 | L: linux-media@vger.kernel.org |
| 8896 | S: Supported | 8897 | S: Supported |
| 8897 | F: drivers/media/pci/solo6x10/ | 8898 | F: drivers/media/pci/solo6x10/ |
| 8898 | 8899 | ||
| 8899 | SOFTWARE RAID (Multiple Disks) SUPPORT | 8900 | SOFTWARE RAID (Multiple Disks) SUPPORT |
| 8900 | M: Neil Brown <neilb@suse.de> | 8901 | M: Neil Brown <neilb@suse.de> |
| 8901 | L: linux-raid@vger.kernel.org | 8902 | L: linux-raid@vger.kernel.org |
| 8902 | S: Supported | 8903 | S: Supported |
| 8903 | F: drivers/md/ | 8904 | F: drivers/md/ |
| 8904 | F: include/linux/raid/ | 8905 | F: include/linux/raid/ |
| 8905 | F: include/uapi/linux/raid/ | 8906 | F: include/uapi/linux/raid/ |
| 8906 | 8907 | ||
| 8907 | SONIC NETWORK DRIVER | 8908 | SONIC NETWORK DRIVER |
| 8908 | M: Thomas Bogendoerfer <tsbogend@alpha.franken.de> | 8909 | M: Thomas Bogendoerfer <tsbogend@alpha.franken.de> |
| 8909 | L: netdev@vger.kernel.org | 8910 | L: netdev@vger.kernel.org |
| 8910 | S: Maintained | 8911 | S: Maintained |
| 8911 | F: drivers/net/ethernet/natsemi/sonic.* | 8912 | F: drivers/net/ethernet/natsemi/sonic.* |
| 8912 | 8913 | ||
| 8913 | SONICS SILICON BACKPLANE DRIVER (SSB) | 8914 | SONICS SILICON BACKPLANE DRIVER (SSB) |
| 8914 | M: Michael Buesch <m@bues.ch> | 8915 | M: Michael Buesch <m@bues.ch> |
| 8915 | L: netdev@vger.kernel.org | 8916 | L: netdev@vger.kernel.org |
| 8916 | S: Maintained | 8917 | S: Maintained |
| 8917 | F: drivers/ssb/ | 8918 | F: drivers/ssb/ |
| 8918 | F: include/linux/ssb/ | 8919 | F: include/linux/ssb/ |
| 8919 | 8920 | ||
| 8920 | SONY VAIO CONTROL DEVICE DRIVER | 8921 | SONY VAIO CONTROL DEVICE DRIVER |
| 8921 | M: Mattia Dongili <malattia@linux.it> | 8922 | M: Mattia Dongili <malattia@linux.it> |
| 8922 | L: platform-driver-x86@vger.kernel.org | 8923 | L: platform-driver-x86@vger.kernel.org |
| 8923 | W: http://www.linux.it/~malattia/wiki/index.php/Sony_drivers | 8924 | W: http://www.linux.it/~malattia/wiki/index.php/Sony_drivers |
| 8924 | S: Maintained | 8925 | S: Maintained |
| 8925 | F: Documentation/laptops/sony-laptop.txt | 8926 | F: Documentation/laptops/sony-laptop.txt |
| 8926 | F: drivers/char/sonypi.c | 8927 | F: drivers/char/sonypi.c |
| 8927 | F: drivers/platform/x86/sony-laptop.c | 8928 | F: drivers/platform/x86/sony-laptop.c |
| 8928 | F: include/linux/sony-laptop.h | 8929 | F: include/linux/sony-laptop.h |
| 8929 | 8930 | ||
| 8930 | SONY MEMORYSTICK CARD SUPPORT | 8931 | SONY MEMORYSTICK CARD SUPPORT |
| 8931 | M: Alex Dubov <oakad@yahoo.com> | 8932 | M: Alex Dubov <oakad@yahoo.com> |
| 8932 | W: http://tifmxx.berlios.de/ | 8933 | W: http://tifmxx.berlios.de/ |
| 8933 | S: Maintained | 8934 | S: Maintained |
| 8934 | F: drivers/memstick/host/tifm_ms.c | 8935 | F: drivers/memstick/host/tifm_ms.c |
| 8935 | 8936 | ||
| 8936 | SONY MEMORYSTICK STANDARD SUPPORT | 8937 | SONY MEMORYSTICK STANDARD SUPPORT |
| 8937 | M: Maxim Levitsky <maximlevitsky@gmail.com> | 8938 | M: Maxim Levitsky <maximlevitsky@gmail.com> |
| 8938 | S: Maintained | 8939 | S: Maintained |
| 8939 | F: drivers/memstick/core/ms_block.* | 8940 | F: drivers/memstick/core/ms_block.* |
| 8940 | 8941 | ||
| 8941 | SOUND | 8942 | SOUND |
| 8942 | M: Jaroslav Kysela <perex@perex.cz> | 8943 | M: Jaroslav Kysela <perex@perex.cz> |
| 8943 | M: Takashi Iwai <tiwai@suse.de> | 8944 | M: Takashi Iwai <tiwai@suse.de> |
| 8944 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 8945 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 8945 | W: http://www.alsa-project.org/ | 8946 | W: http://www.alsa-project.org/ |
| 8946 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git | 8947 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git |
| 8947 | T: git git://git.alsa-project.org/alsa-kernel.git | 8948 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 8948 | Q: http://patchwork.kernel.org/project/alsa-devel/list/ | 8949 | Q: http://patchwork.kernel.org/project/alsa-devel/list/ |
| 8949 | S: Maintained | 8950 | S: Maintained |
| 8950 | F: Documentation/sound/ | 8951 | F: Documentation/sound/ |
| 8951 | F: include/sound/ | 8952 | F: include/sound/ |
| 8952 | F: include/uapi/sound/ | 8953 | F: include/uapi/sound/ |
| 8953 | F: sound/ | 8954 | F: sound/ |
| 8954 | 8955 | ||
| 8955 | SOUND - COMPRESSED AUDIO | 8956 | SOUND - COMPRESSED AUDIO |
| 8956 | M: Vinod Koul <vinod.koul@intel.com> | 8957 | M: Vinod Koul <vinod.koul@intel.com> |
| 8957 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 8958 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 8958 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git | 8959 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git |
| 8959 | S: Supported | 8960 | S: Supported |
| 8960 | F: Documentation/sound/alsa/compress_offload.txt | 8961 | F: Documentation/sound/alsa/compress_offload.txt |
| 8961 | F: include/sound/compress_driver.h | 8962 | F: include/sound/compress_driver.h |
| 8962 | F: include/uapi/sound/compress_* | 8963 | F: include/uapi/sound/compress_* |
| 8963 | F: sound/core/compress_offload.c | 8964 | F: sound/core/compress_offload.c |
| 8964 | F: sound/soc/soc-compress.c | 8965 | F: sound/soc/soc-compress.c |
| 8965 | 8966 | ||
| 8966 | SOUND - SOC LAYER / DYNAMIC AUDIO POWER MANAGEMENT (ASoC) | 8967 | SOUND - SOC LAYER / DYNAMIC AUDIO POWER MANAGEMENT (ASoC) |
| 8967 | M: Liam Girdwood <lgirdwood@gmail.com> | 8968 | M: Liam Girdwood <lgirdwood@gmail.com> |
| 8968 | M: Mark Brown <broonie@kernel.org> | 8969 | M: Mark Brown <broonie@kernel.org> |
| 8969 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git | 8970 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git |
| 8970 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 8971 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 8971 | W: http://alsa-project.org/main/index.php/ASoC | 8972 | W: http://alsa-project.org/main/index.php/ASoC |
| 8972 | S: Supported | 8973 | S: Supported |
| 8973 | F: Documentation/sound/alsa/soc/ | 8974 | F: Documentation/sound/alsa/soc/ |
| 8974 | F: sound/soc/ | 8975 | F: sound/soc/ |
| 8975 | F: include/sound/soc* | 8976 | F: include/sound/soc* |
| 8976 | 8977 | ||
| 8977 | SOUND - DMAENGINE HELPERS | 8978 | SOUND - DMAENGINE HELPERS |
| 8978 | M: Lars-Peter Clausen <lars@metafoo.de> | 8979 | M: Lars-Peter Clausen <lars@metafoo.de> |
| 8979 | S: Supported | 8980 | S: Supported |
| 8980 | F: include/sound/dmaengine_pcm.h | 8981 | F: include/sound/dmaengine_pcm.h |
| 8981 | F: sound/core/pcm_dmaengine.c | 8982 | F: sound/core/pcm_dmaengine.c |
| 8982 | F: sound/soc/soc-generic-dmaengine-pcm.c | 8983 | F: sound/soc/soc-generic-dmaengine-pcm.c |
| 8983 | 8984 | ||
| 8984 | SP2 MEDIA DRIVER | 8985 | SP2 MEDIA DRIVER |
| 8985 | M: Olli Salonen <olli.salonen@iki.fi> | 8986 | M: Olli Salonen <olli.salonen@iki.fi> |
| 8986 | L: linux-media@vger.kernel.org | 8987 | L: linux-media@vger.kernel.org |
| 8987 | W: http://linuxtv.org/ | 8988 | W: http://linuxtv.org/ |
| 8988 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 8989 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 8989 | S: Maintained | 8990 | S: Maintained |
| 8990 | F: drivers/media/dvb-frontends/sp2* | 8991 | F: drivers/media/dvb-frontends/sp2* |
| 8991 | 8992 | ||
| 8992 | SPARC + UltraSPARC (sparc/sparc64) | 8993 | SPARC + UltraSPARC (sparc/sparc64) |
| 8993 | M: "David S. Miller" <davem@davemloft.net> | 8994 | M: "David S. Miller" <davem@davemloft.net> |
| 8994 | L: sparclinux@vger.kernel.org | 8995 | L: sparclinux@vger.kernel.org |
| 8995 | Q: http://patchwork.ozlabs.org/project/sparclinux/list/ | 8996 | Q: http://patchwork.ozlabs.org/project/sparclinux/list/ |
| 8996 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc.git | 8997 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc.git |
| 8997 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-next.git | 8998 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-next.git |
| 8998 | S: Maintained | 8999 | S: Maintained |
| 8999 | F: arch/sparc/ | 9000 | F: arch/sparc/ |
| 9000 | F: drivers/sbus/ | 9001 | F: drivers/sbus/ |
| 9001 | 9002 | ||
| 9002 | SPARC SERIAL DRIVERS | 9003 | SPARC SERIAL DRIVERS |
| 9003 | M: "David S. Miller" <davem@davemloft.net> | 9004 | M: "David S. Miller" <davem@davemloft.net> |
| 9004 | L: sparclinux@vger.kernel.org | 9005 | L: sparclinux@vger.kernel.org |
| 9005 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc.git | 9006 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc.git |
| 9006 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-next.git | 9007 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-next.git |
| 9007 | S: Maintained | 9008 | S: Maintained |
| 9008 | F: include/linux/sunserialcore.h | 9009 | F: include/linux/sunserialcore.h |
| 9009 | F: drivers/tty/serial/suncore.c | 9010 | F: drivers/tty/serial/suncore.c |
| 9010 | F: drivers/tty/serial/sunhv.c | 9011 | F: drivers/tty/serial/sunhv.c |
| 9011 | F: drivers/tty/serial/sunsab.c | 9012 | F: drivers/tty/serial/sunsab.c |
| 9012 | F: drivers/tty/serial/sunsab.h | 9013 | F: drivers/tty/serial/sunsab.h |
| 9013 | F: drivers/tty/serial/sunsu.c | 9014 | F: drivers/tty/serial/sunsu.c |
| 9014 | F: drivers/tty/serial/sunzilog.c | 9015 | F: drivers/tty/serial/sunzilog.c |
| 9015 | F: drivers/tty/serial/sunzilog.h | 9016 | F: drivers/tty/serial/sunzilog.h |
| 9016 | 9017 | ||
| 9017 | SPARSE CHECKER | 9018 | SPARSE CHECKER |
| 9018 | M: "Christopher Li" <sparse@chrisli.org> | 9019 | M: "Christopher Li" <sparse@chrisli.org> |
| 9019 | L: linux-sparse@vger.kernel.org | 9020 | L: linux-sparse@vger.kernel.org |
| 9020 | W: https://sparse.wiki.kernel.org/ | 9021 | W: https://sparse.wiki.kernel.org/ |
| 9021 | T: git git://git.kernel.org/pub/scm/devel/sparse/sparse.git | 9022 | T: git git://git.kernel.org/pub/scm/devel/sparse/sparse.git |
| 9022 | T: git git://git.kernel.org/pub/scm/devel/sparse/chrisl/sparse.git | 9023 | T: git git://git.kernel.org/pub/scm/devel/sparse/chrisl/sparse.git |
| 9023 | S: Maintained | 9024 | S: Maintained |
| 9024 | F: include/linux/compiler.h | 9025 | F: include/linux/compiler.h |
| 9025 | 9026 | ||
| 9026 | SPEAR PLATFORM SUPPORT | 9027 | SPEAR PLATFORM SUPPORT |
| 9027 | M: Viresh Kumar <viresh.linux@gmail.com> | 9028 | M: Viresh Kumar <viresh.linux@gmail.com> |
| 9028 | M: Shiraz Hashim <shiraz.linux.kernel@gmail.com> | 9029 | M: Shiraz Hashim <shiraz.linux.kernel@gmail.com> |
| 9029 | L: spear-devel@list.st.com | 9030 | L: spear-devel@list.st.com |
| 9030 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 9031 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 9031 | W: http://www.st.com/spear | 9032 | W: http://www.st.com/spear |
| 9032 | S: Maintained | 9033 | S: Maintained |
| 9033 | F: arch/arm/mach-spear/ | 9034 | F: arch/arm/mach-spear/ |
| 9034 | 9035 | ||
| 9035 | SPEAR CLOCK FRAMEWORK SUPPORT | 9036 | SPEAR CLOCK FRAMEWORK SUPPORT |
| 9036 | M: Viresh Kumar <viresh.linux@gmail.com> | 9037 | M: Viresh Kumar <viresh.linux@gmail.com> |
| 9037 | L: spear-devel@list.st.com | 9038 | L: spear-devel@list.st.com |
| 9038 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 9039 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 9039 | W: http://www.st.com/spear | 9040 | W: http://www.st.com/spear |
| 9040 | S: Maintained | 9041 | S: Maintained |
| 9041 | F: drivers/clk/spear/ | 9042 | F: drivers/clk/spear/ |
| 9042 | 9043 | ||
| 9043 | SPI SUBSYSTEM | 9044 | SPI SUBSYSTEM |
| 9044 | M: Mark Brown <broonie@kernel.org> | 9045 | M: Mark Brown <broonie@kernel.org> |
| 9045 | L: linux-spi@vger.kernel.org | 9046 | L: linux-spi@vger.kernel.org |
| 9046 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git | 9047 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git |
| 9047 | Q: http://patchwork.kernel.org/project/spi-devel-general/list/ | 9048 | Q: http://patchwork.kernel.org/project/spi-devel-general/list/ |
| 9048 | S: Maintained | 9049 | S: Maintained |
| 9049 | F: Documentation/spi/ | 9050 | F: Documentation/spi/ |
| 9050 | F: drivers/spi/ | 9051 | F: drivers/spi/ |
| 9051 | F: include/linux/spi/ | 9052 | F: include/linux/spi/ |
| 9052 | F: include/uapi/linux/spi/ | 9053 | F: include/uapi/linux/spi/ |
| 9053 | 9054 | ||
| 9054 | SPIDERNET NETWORK DRIVER for CELL | 9055 | SPIDERNET NETWORK DRIVER for CELL |
| 9055 | M: Ishizaki Kou <kou.ishizaki@toshiba.co.jp> | 9056 | M: Ishizaki Kou <kou.ishizaki@toshiba.co.jp> |
| 9056 | M: Jens Osterkamp <jens@de.ibm.com> | 9057 | M: Jens Osterkamp <jens@de.ibm.com> |
| 9057 | L: netdev@vger.kernel.org | 9058 | L: netdev@vger.kernel.org |
| 9058 | S: Supported | 9059 | S: Supported |
| 9059 | F: Documentation/networking/spider_net.txt | 9060 | F: Documentation/networking/spider_net.txt |
| 9060 | F: drivers/net/ethernet/toshiba/spider_net* | 9061 | F: drivers/net/ethernet/toshiba/spider_net* |
| 9061 | 9062 | ||
| 9062 | SPU FILE SYSTEM | 9063 | SPU FILE SYSTEM |
| 9063 | M: Jeremy Kerr <jk@ozlabs.org> | 9064 | M: Jeremy Kerr <jk@ozlabs.org> |
| 9064 | L: linuxppc-dev@lists.ozlabs.org | 9065 | L: linuxppc-dev@lists.ozlabs.org |
| 9065 | L: cbe-oss-dev@lists.ozlabs.org | 9066 | L: cbe-oss-dev@lists.ozlabs.org |
| 9066 | W: http://www.ibm.com/developerworks/power/cell/ | 9067 | W: http://www.ibm.com/developerworks/power/cell/ |
| 9067 | S: Supported | 9068 | S: Supported |
| 9068 | F: Documentation/filesystems/spufs.txt | 9069 | F: Documentation/filesystems/spufs.txt |
| 9069 | F: arch/powerpc/platforms/cell/spufs/ | 9070 | F: arch/powerpc/platforms/cell/spufs/ |
| 9070 | 9071 | ||
| 9071 | SQUASHFS FILE SYSTEM | 9072 | SQUASHFS FILE SYSTEM |
| 9072 | M: Phillip Lougher <phillip@squashfs.org.uk> | 9073 | M: Phillip Lougher <phillip@squashfs.org.uk> |
| 9073 | L: squashfs-devel@lists.sourceforge.net (subscribers-only) | 9074 | L: squashfs-devel@lists.sourceforge.net (subscribers-only) |
| 9074 | W: http://squashfs.org.uk | 9075 | W: http://squashfs.org.uk |
| 9075 | S: Maintained | 9076 | S: Maintained |
| 9076 | F: Documentation/filesystems/squashfs.txt | 9077 | F: Documentation/filesystems/squashfs.txt |
| 9077 | F: fs/squashfs/ | 9078 | F: fs/squashfs/ |
| 9078 | 9079 | ||
| 9079 | SRM (Alpha) environment access | 9080 | SRM (Alpha) environment access |
| 9080 | M: Jan-Benedict Glaw <jbglaw@lug-owl.de> | 9081 | M: Jan-Benedict Glaw <jbglaw@lug-owl.de> |
| 9081 | S: Maintained | 9082 | S: Maintained |
| 9082 | F: arch/alpha/kernel/srm_env.c | 9083 | F: arch/alpha/kernel/srm_env.c |
| 9083 | 9084 | ||
| 9084 | STABLE BRANCH | 9085 | STABLE BRANCH |
| 9085 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 9086 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 9086 | L: stable@vger.kernel.org | 9087 | L: stable@vger.kernel.org |
| 9087 | S: Supported | 9088 | S: Supported |
| 9088 | F: Documentation/stable_kernel_rules.txt | 9089 | F: Documentation/stable_kernel_rules.txt |
| 9089 | 9090 | ||
| 9090 | STAGING SUBSYSTEM | 9091 | STAGING SUBSYSTEM |
| 9091 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 9092 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 9092 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git | 9093 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git |
| 9093 | L: devel@driverdev.osuosl.org | 9094 | L: devel@driverdev.osuosl.org |
| 9094 | S: Supported | 9095 | S: Supported |
| 9095 | F: drivers/staging/ | 9096 | F: drivers/staging/ |
| 9096 | 9097 | ||
| 9097 | STAGING - COMEDI | 9098 | STAGING - COMEDI |
| 9098 | M: Ian Abbott <abbotti@mev.co.uk> | 9099 | M: Ian Abbott <abbotti@mev.co.uk> |
| 9099 | M: H Hartley Sweeten <hsweeten@visionengravers.com> | 9100 | M: H Hartley Sweeten <hsweeten@visionengravers.com> |
| 9100 | S: Odd Fixes | 9101 | S: Odd Fixes |
| 9101 | F: drivers/staging/comedi/ | 9102 | F: drivers/staging/comedi/ |
| 9102 | 9103 | ||
| 9103 | STAGING - FLARION FT1000 DRIVERS | 9104 | STAGING - FLARION FT1000 DRIVERS |
| 9104 | M: Marek Belisko <marek.belisko@gmail.com> | 9105 | M: Marek Belisko <marek.belisko@gmail.com> |
| 9105 | S: Odd Fixes | 9106 | S: Odd Fixes |
| 9106 | F: drivers/staging/ft1000/ | 9107 | F: drivers/staging/ft1000/ |
| 9107 | 9108 | ||
| 9108 | STAGING - INDUSTRIAL IO | 9109 | STAGING - INDUSTRIAL IO |
| 9109 | M: Jonathan Cameron <jic23@kernel.org> | 9110 | M: Jonathan Cameron <jic23@kernel.org> |
| 9110 | L: linux-iio@vger.kernel.org | 9111 | L: linux-iio@vger.kernel.org |
| 9111 | S: Odd Fixes | 9112 | S: Odd Fixes |
| 9112 | F: drivers/staging/iio/ | 9113 | F: drivers/staging/iio/ |
| 9113 | 9114 | ||
| 9114 | STAGING - LIRC (LINUX INFRARED REMOTE CONTROL) DRIVERS | 9115 | STAGING - LIRC (LINUX INFRARED REMOTE CONTROL) DRIVERS |
| 9115 | M: Jarod Wilson <jarod@wilsonet.com> | 9116 | M: Jarod Wilson <jarod@wilsonet.com> |
| 9116 | W: http://www.lirc.org/ | 9117 | W: http://www.lirc.org/ |
| 9117 | S: Odd Fixes | 9118 | S: Odd Fixes |
| 9118 | F: drivers/staging/media/lirc/ | 9119 | F: drivers/staging/media/lirc/ |
| 9119 | 9120 | ||
| 9120 | STAGING - LUSTRE PARALLEL FILESYSTEM | 9121 | STAGING - LUSTRE PARALLEL FILESYSTEM |
| 9121 | M: Oleg Drokin <oleg.drokin@intel.com> | 9122 | M: Oleg Drokin <oleg.drokin@intel.com> |
| 9122 | M: Andreas Dilger <andreas.dilger@intel.com> | 9123 | M: Andreas Dilger <andreas.dilger@intel.com> |
| 9123 | L: HPDD-discuss@lists.01.org (moderated for non-subscribers) | 9124 | L: HPDD-discuss@lists.01.org (moderated for non-subscribers) |
| 9124 | W: http://lustre.opensfs.org/ | 9125 | W: http://lustre.opensfs.org/ |
| 9125 | S: Maintained | 9126 | S: Maintained |
| 9126 | F: drivers/staging/lustre | 9127 | F: drivers/staging/lustre |
| 9127 | 9128 | ||
| 9128 | STAGING - NVIDIA COMPLIANT EMBEDDED CONTROLLER INTERFACE (nvec) | 9129 | STAGING - NVIDIA COMPLIANT EMBEDDED CONTROLLER INTERFACE (nvec) |
| 9129 | M: Julian Andres Klode <jak@jak-linux.org> | 9130 | M: Julian Andres Klode <jak@jak-linux.org> |
| 9130 | M: Marc Dietrich <marvin24@gmx.de> | 9131 | M: Marc Dietrich <marvin24@gmx.de> |
| 9131 | L: ac100@lists.launchpad.net (moderated for non-subscribers) | 9132 | L: ac100@lists.launchpad.net (moderated for non-subscribers) |
| 9132 | L: linux-tegra@vger.kernel.org | 9133 | L: linux-tegra@vger.kernel.org |
| 9133 | S: Maintained | 9134 | S: Maintained |
| 9134 | F: drivers/staging/nvec/ | 9135 | F: drivers/staging/nvec/ |
| 9135 | 9136 | ||
| 9136 | STAGING - OLPC SECONDARY DISPLAY CONTROLLER (DCON) | 9137 | STAGING - OLPC SECONDARY DISPLAY CONTROLLER (DCON) |
| 9137 | M: Jens Frederich <jfrederich@gmail.com> | 9138 | M: Jens Frederich <jfrederich@gmail.com> |
| 9138 | M: Daniel Drake <dsd@laptop.org> | 9139 | M: Daniel Drake <dsd@laptop.org> |
| 9139 | M: Jon Nettleton <jon.nettleton@gmail.com> | 9140 | M: Jon Nettleton <jon.nettleton@gmail.com> |
| 9140 | W: http://wiki.laptop.org/go/DCON | 9141 | W: http://wiki.laptop.org/go/DCON |
| 9141 | S: Maintained | 9142 | S: Maintained |
| 9142 | F: drivers/staging/olpc_dcon/ | 9143 | F: drivers/staging/olpc_dcon/ |
| 9143 | 9144 | ||
| 9144 | STAGING - OZMO DEVICES USB OVER WIFI DRIVER | 9145 | STAGING - OZMO DEVICES USB OVER WIFI DRIVER |
| 9145 | M: Shigekatsu Tateno <shigekatsu.tateno@atmel.com> | 9146 | M: Shigekatsu Tateno <shigekatsu.tateno@atmel.com> |
| 9146 | S: Maintained | 9147 | S: Maintained |
| 9147 | F: drivers/staging/ozwpan/ | 9148 | F: drivers/staging/ozwpan/ |
| 9148 | 9149 | ||
| 9149 | STAGING - PARALLEL LCD/KEYPAD PANEL DRIVER | 9150 | STAGING - PARALLEL LCD/KEYPAD PANEL DRIVER |
| 9150 | M: Willy Tarreau <willy@meta-x.org> | 9151 | M: Willy Tarreau <willy@meta-x.org> |
| 9151 | S: Odd Fixes | 9152 | S: Odd Fixes |
| 9152 | F: drivers/staging/panel/ | 9153 | F: drivers/staging/panel/ |
| 9153 | 9154 | ||
| 9154 | STAGING - REALTEK RTL8712U DRIVERS | 9155 | STAGING - REALTEK RTL8712U DRIVERS |
| 9155 | M: Larry Finger <Larry.Finger@lwfinger.net> | 9156 | M: Larry Finger <Larry.Finger@lwfinger.net> |
| 9156 | M: Florian Schilhabel <florian.c.schilhabel@googlemail.com>. | 9157 | M: Florian Schilhabel <florian.c.schilhabel@googlemail.com>. |
| 9157 | S: Odd Fixes | 9158 | S: Odd Fixes |
| 9158 | F: drivers/staging/rtl8712/ | 9159 | F: drivers/staging/rtl8712/ |
| 9159 | 9160 | ||
| 9160 | STAGING - REALTEK RTL8723U WIRELESS DRIVER | 9161 | STAGING - REALTEK RTL8723U WIRELESS DRIVER |
| 9161 | M: Larry Finger <Larry.Finger@lwfinger.net> | 9162 | M: Larry Finger <Larry.Finger@lwfinger.net> |
| 9162 | M: Jes Sorensen <Jes.Sorensen@redhat.com> | 9163 | M: Jes Sorensen <Jes.Sorensen@redhat.com> |
| 9163 | L: linux-wireless@vger.kernel.org | 9164 | L: linux-wireless@vger.kernel.org |
| 9164 | S: Maintained | 9165 | S: Maintained |
| 9165 | F: drivers/staging/rtl8723au/ | 9166 | F: drivers/staging/rtl8723au/ |
| 9166 | 9167 | ||
| 9167 | STAGING - SLICOSS | 9168 | STAGING - SLICOSS |
| 9168 | M: Lior Dotan <liodot@gmail.com> | 9169 | M: Lior Dotan <liodot@gmail.com> |
| 9169 | M: Christopher Harrer <charrer@alacritech.com> | 9170 | M: Christopher Harrer <charrer@alacritech.com> |
| 9170 | S: Odd Fixes | 9171 | S: Odd Fixes |
| 9171 | F: drivers/staging/slicoss/ | 9172 | F: drivers/staging/slicoss/ |
| 9172 | 9173 | ||
| 9173 | STAGING - SPEAKUP CONSOLE SPEECH DRIVER | 9174 | STAGING - SPEAKUP CONSOLE SPEECH DRIVER |
| 9174 | M: William Hubbs <w.d.hubbs@gmail.com> | 9175 | M: William Hubbs <w.d.hubbs@gmail.com> |
| 9175 | M: Chris Brannon <chris@the-brannons.com> | 9176 | M: Chris Brannon <chris@the-brannons.com> |
| 9176 | M: Kirk Reiser <kirk@reisers.ca> | 9177 | M: Kirk Reiser <kirk@reisers.ca> |
| 9177 | M: Samuel Thibault <samuel.thibault@ens-lyon.org> | 9178 | M: Samuel Thibault <samuel.thibault@ens-lyon.org> |
| 9178 | L: speakup@linux-speakup.org | 9179 | L: speakup@linux-speakup.org |
| 9179 | W: http://www.linux-speakup.org/ | 9180 | W: http://www.linux-speakup.org/ |
| 9180 | S: Odd Fixes | 9181 | S: Odd Fixes |
| 9181 | F: drivers/staging/speakup/ | 9182 | F: drivers/staging/speakup/ |
| 9182 | 9183 | ||
| 9183 | STAGING - VIA VT665X DRIVERS | 9184 | STAGING - VIA VT665X DRIVERS |
| 9184 | M: Forest Bond <forest@alittletooquiet.net> | 9185 | M: Forest Bond <forest@alittletooquiet.net> |
| 9185 | S: Odd Fixes | 9186 | S: Odd Fixes |
| 9186 | F: drivers/staging/vt665?/ | 9187 | F: drivers/staging/vt665?/ |
| 9187 | 9188 | ||
| 9188 | STAGING - XGI Z7,Z9,Z11 PCI DISPLAY DRIVER | 9189 | STAGING - XGI Z7,Z9,Z11 PCI DISPLAY DRIVER |
| 9189 | M: Arnaud Patard <arnaud.patard@rtp-net.org> | 9190 | M: Arnaud Patard <arnaud.patard@rtp-net.org> |
| 9190 | S: Odd Fixes | 9191 | S: Odd Fixes |
| 9191 | F: drivers/staging/xgifb/ | 9192 | F: drivers/staging/xgifb/ |
| 9192 | 9193 | ||
| 9193 | STARFIRE/DURALAN NETWORK DRIVER | 9194 | STARFIRE/DURALAN NETWORK DRIVER |
| 9194 | M: Ion Badulescu <ionut@badula.org> | 9195 | M: Ion Badulescu <ionut@badula.org> |
| 9195 | S: Odd Fixes | 9196 | S: Odd Fixes |
| 9196 | F: drivers/net/ethernet/adaptec/starfire* | 9197 | F: drivers/net/ethernet/adaptec/starfire* |
| 9197 | 9198 | ||
| 9198 | SUN3/3X | 9199 | SUN3/3X |
| 9199 | M: Sam Creasey <sammy@sammy.net> | 9200 | M: Sam Creasey <sammy@sammy.net> |
| 9200 | W: http://sammy.net/sun3/ | 9201 | W: http://sammy.net/sun3/ |
| 9201 | S: Maintained | 9202 | S: Maintained |
| 9202 | F: arch/m68k/kernel/*sun3* | 9203 | F: arch/m68k/kernel/*sun3* |
| 9203 | F: arch/m68k/sun3*/ | 9204 | F: arch/m68k/sun3*/ |
| 9204 | F: arch/m68k/include/asm/sun3* | 9205 | F: arch/m68k/include/asm/sun3* |
| 9205 | F: drivers/net/ethernet/i825xx/sun3* | 9206 | F: drivers/net/ethernet/i825xx/sun3* |
| 9206 | 9207 | ||
| 9207 | SUNDANCE NETWORK DRIVER | 9208 | SUNDANCE NETWORK DRIVER |
| 9208 | M: Denis Kirjanov <kda@linux-powerpc.org> | 9209 | M: Denis Kirjanov <kda@linux-powerpc.org> |
| 9209 | L: netdev@vger.kernel.org | 9210 | L: netdev@vger.kernel.org |
| 9210 | S: Maintained | 9211 | S: Maintained |
| 9211 | F: drivers/net/ethernet/dlink/sundance.c | 9212 | F: drivers/net/ethernet/dlink/sundance.c |
| 9212 | 9213 | ||
| 9213 | SUPERH | 9214 | SUPERH |
| 9214 | L: linux-sh@vger.kernel.org | 9215 | L: linux-sh@vger.kernel.org |
| 9215 | W: http://www.linux-sh.org | 9216 | W: http://www.linux-sh.org |
| 9216 | Q: http://patchwork.kernel.org/project/linux-sh/list/ | 9217 | Q: http://patchwork.kernel.org/project/linux-sh/list/ |
| 9217 | S: Orphan | 9218 | S: Orphan |
| 9218 | F: Documentation/sh/ | 9219 | F: Documentation/sh/ |
| 9219 | F: arch/sh/ | 9220 | F: arch/sh/ |
| 9220 | F: drivers/sh/ | 9221 | F: drivers/sh/ |
| 9221 | 9222 | ||
| 9222 | SUSPEND TO RAM | 9223 | SUSPEND TO RAM |
| 9223 | M: "Rafael J. Wysocki" <rjw@rjwysocki.net> | 9224 | M: "Rafael J. Wysocki" <rjw@rjwysocki.net> |
| 9224 | M: Len Brown <len.brown@intel.com> | 9225 | M: Len Brown <len.brown@intel.com> |
| 9225 | M: Pavel Machek <pavel@ucw.cz> | 9226 | M: Pavel Machek <pavel@ucw.cz> |
| 9226 | L: linux-pm@vger.kernel.org | 9227 | L: linux-pm@vger.kernel.org |
| 9227 | S: Supported | 9228 | S: Supported |
| 9228 | F: Documentation/power/ | 9229 | F: Documentation/power/ |
| 9229 | F: arch/x86/kernel/acpi/ | 9230 | F: arch/x86/kernel/acpi/ |
| 9230 | F: drivers/base/power/ | 9231 | F: drivers/base/power/ |
| 9231 | F: kernel/power/ | 9232 | F: kernel/power/ |
| 9232 | F: include/linux/suspend.h | 9233 | F: include/linux/suspend.h |
| 9233 | F: include/linux/freezer.h | 9234 | F: include/linux/freezer.h |
| 9234 | F: include/linux/pm.h | 9235 | F: include/linux/pm.h |
| 9235 | 9236 | ||
| 9236 | SVGA HANDLING | 9237 | SVGA HANDLING |
| 9237 | M: Martin Mares <mj@ucw.cz> | 9238 | M: Martin Mares <mj@ucw.cz> |
| 9238 | L: linux-video@atrey.karlin.mff.cuni.cz | 9239 | L: linux-video@atrey.karlin.mff.cuni.cz |
| 9239 | S: Maintained | 9240 | S: Maintained |
| 9240 | F: Documentation/svga.txt | 9241 | F: Documentation/svga.txt |
| 9241 | F: arch/x86/boot/video* | 9242 | F: arch/x86/boot/video* |
| 9242 | 9243 | ||
| 9243 | SWIOTLB SUBSYSTEM | 9244 | SWIOTLB SUBSYSTEM |
| 9244 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 9245 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 9245 | L: linux-kernel@vger.kernel.org | 9246 | L: linux-kernel@vger.kernel.org |
| 9246 | S: Supported | 9247 | S: Supported |
| 9247 | F: lib/swiotlb.c | 9248 | F: lib/swiotlb.c |
| 9248 | F: arch/*/kernel/pci-swiotlb.c | 9249 | F: arch/*/kernel/pci-swiotlb.c |
| 9249 | F: include/linux/swiotlb.h | 9250 | F: include/linux/swiotlb.h |
| 9250 | 9251 | ||
| 9251 | SWITCHDEV | 9252 | SWITCHDEV |
| 9252 | M: Jiri Pirko <jiri@resnulli.us> | 9253 | M: Jiri Pirko <jiri@resnulli.us> |
| 9253 | L: netdev@vger.kernel.org | 9254 | L: netdev@vger.kernel.org |
| 9254 | S: Supported | 9255 | S: Supported |
| 9255 | F: net/switchdev/ | 9256 | F: net/switchdev/ |
| 9256 | F: include/net/switchdev.h | 9257 | F: include/net/switchdev.h |
| 9257 | 9258 | ||
| 9258 | SYNOPSYS ARC ARCHITECTURE | 9259 | SYNOPSYS ARC ARCHITECTURE |
| 9259 | M: Vineet Gupta <vgupta@synopsys.com> | 9260 | M: Vineet Gupta <vgupta@synopsys.com> |
| 9260 | S: Supported | 9261 | S: Supported |
| 9261 | F: arch/arc/ | 9262 | F: arch/arc/ |
| 9262 | F: Documentation/devicetree/bindings/arc/ | 9263 | F: Documentation/devicetree/bindings/arc/ |
| 9263 | F: drivers/tty/serial/arc_uart.c | 9264 | F: drivers/tty/serial/arc_uart.c |
| 9264 | 9265 | ||
| 9265 | SYSV FILESYSTEM | 9266 | SYSV FILESYSTEM |
| 9266 | M: Christoph Hellwig <hch@infradead.org> | 9267 | M: Christoph Hellwig <hch@infradead.org> |
| 9267 | S: Maintained | 9268 | S: Maintained |
| 9268 | F: Documentation/filesystems/sysv-fs.txt | 9269 | F: Documentation/filesystems/sysv-fs.txt |
| 9269 | F: fs/sysv/ | 9270 | F: fs/sysv/ |
| 9270 | F: include/linux/sysv_fs.h | 9271 | F: include/linux/sysv_fs.h |
| 9271 | 9272 | ||
| 9272 | TARGET SUBSYSTEM | 9273 | TARGET SUBSYSTEM |
| 9273 | M: Nicholas A. Bellinger <nab@linux-iscsi.org> | 9274 | M: Nicholas A. Bellinger <nab@linux-iscsi.org> |
| 9274 | L: linux-scsi@vger.kernel.org | 9275 | L: linux-scsi@vger.kernel.org |
| 9275 | L: target-devel@vger.kernel.org | 9276 | L: target-devel@vger.kernel.org |
| 9276 | W: http://www.linux-iscsi.org | 9277 | W: http://www.linux-iscsi.org |
| 9277 | W: http://groups.google.com/group/linux-iscsi-target-dev | 9278 | W: http://groups.google.com/group/linux-iscsi-target-dev |
| 9278 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending.git master | 9279 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending.git master |
| 9279 | S: Supported | 9280 | S: Supported |
| 9280 | F: drivers/target/ | 9281 | F: drivers/target/ |
| 9281 | F: include/target/ | 9282 | F: include/target/ |
| 9282 | F: Documentation/target/ | 9283 | F: Documentation/target/ |
| 9283 | 9284 | ||
| 9284 | TASKSTATS STATISTICS INTERFACE | 9285 | TASKSTATS STATISTICS INTERFACE |
| 9285 | M: Balbir Singh <bsingharora@gmail.com> | 9286 | M: Balbir Singh <bsingharora@gmail.com> |
| 9286 | S: Maintained | 9287 | S: Maintained |
| 9287 | F: Documentation/accounting/taskstats* | 9288 | F: Documentation/accounting/taskstats* |
| 9288 | F: include/linux/taskstats* | 9289 | F: include/linux/taskstats* |
| 9289 | F: kernel/taskstats.c | 9290 | F: kernel/taskstats.c |
| 9290 | 9291 | ||
| 9291 | TC CLASSIFIER | 9292 | TC CLASSIFIER |
| 9292 | M: Jamal Hadi Salim <jhs@mojatatu.com> | 9293 | M: Jamal Hadi Salim <jhs@mojatatu.com> |
| 9293 | L: netdev@vger.kernel.org | 9294 | L: netdev@vger.kernel.org |
| 9294 | S: Maintained | 9295 | S: Maintained |
| 9295 | F: include/net/pkt_cls.h | 9296 | F: include/net/pkt_cls.h |
| 9296 | F: include/uapi/linux/pkt_cls.h | 9297 | F: include/uapi/linux/pkt_cls.h |
| 9297 | F: net/sched/ | 9298 | F: net/sched/ |
| 9298 | 9299 | ||
| 9299 | TCP LOW PRIORITY MODULE | 9300 | TCP LOW PRIORITY MODULE |
| 9300 | M: "Wong Hoi Sing, Edison" <hswong3i@gmail.com> | 9301 | M: "Wong Hoi Sing, Edison" <hswong3i@gmail.com> |
| 9301 | M: "Hung Hing Lun, Mike" <hlhung3i@gmail.com> | 9302 | M: "Hung Hing Lun, Mike" <hlhung3i@gmail.com> |
| 9302 | W: http://tcp-lp-mod.sourceforge.net/ | 9303 | W: http://tcp-lp-mod.sourceforge.net/ |
| 9303 | S: Maintained | 9304 | S: Maintained |
| 9304 | F: net/ipv4/tcp_lp.c | 9305 | F: net/ipv4/tcp_lp.c |
| 9305 | 9306 | ||
| 9306 | TDA10071 MEDIA DRIVER | 9307 | TDA10071 MEDIA DRIVER |
| 9307 | M: Antti Palosaari <crope@iki.fi> | 9308 | M: Antti Palosaari <crope@iki.fi> |
| 9308 | L: linux-media@vger.kernel.org | 9309 | L: linux-media@vger.kernel.org |
| 9309 | W: http://linuxtv.org/ | 9310 | W: http://linuxtv.org/ |
| 9310 | W: http://palosaari.fi/linux/ | 9311 | W: http://palosaari.fi/linux/ |
| 9311 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9312 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9312 | T: git git://linuxtv.org/anttip/media_tree.git | 9313 | T: git git://linuxtv.org/anttip/media_tree.git |
| 9313 | S: Maintained | 9314 | S: Maintained |
| 9314 | F: drivers/media/dvb-frontends/tda10071* | 9315 | F: drivers/media/dvb-frontends/tda10071* |
| 9315 | 9316 | ||
| 9316 | TDA18212 MEDIA DRIVER | 9317 | TDA18212 MEDIA DRIVER |
| 9317 | M: Antti Palosaari <crope@iki.fi> | 9318 | M: Antti Palosaari <crope@iki.fi> |
| 9318 | L: linux-media@vger.kernel.org | 9319 | L: linux-media@vger.kernel.org |
| 9319 | W: http://linuxtv.org/ | 9320 | W: http://linuxtv.org/ |
| 9320 | W: http://palosaari.fi/linux/ | 9321 | W: http://palosaari.fi/linux/ |
| 9321 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9322 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9322 | T: git git://linuxtv.org/anttip/media_tree.git | 9323 | T: git git://linuxtv.org/anttip/media_tree.git |
| 9323 | S: Maintained | 9324 | S: Maintained |
| 9324 | F: drivers/media/tuners/tda18212* | 9325 | F: drivers/media/tuners/tda18212* |
| 9325 | 9326 | ||
| 9326 | TDA18218 MEDIA DRIVER | 9327 | TDA18218 MEDIA DRIVER |
| 9327 | M: Antti Palosaari <crope@iki.fi> | 9328 | M: Antti Palosaari <crope@iki.fi> |
| 9328 | L: linux-media@vger.kernel.org | 9329 | L: linux-media@vger.kernel.org |
| 9329 | W: http://linuxtv.org/ | 9330 | W: http://linuxtv.org/ |
| 9330 | W: http://palosaari.fi/linux/ | 9331 | W: http://palosaari.fi/linux/ |
| 9331 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9332 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9332 | T: git git://linuxtv.org/anttip/media_tree.git | 9333 | T: git git://linuxtv.org/anttip/media_tree.git |
| 9333 | S: Maintained | 9334 | S: Maintained |
| 9334 | F: drivers/media/tuners/tda18218* | 9335 | F: drivers/media/tuners/tda18218* |
| 9335 | 9336 | ||
| 9336 | TDA18271 MEDIA DRIVER | 9337 | TDA18271 MEDIA DRIVER |
| 9337 | M: Michael Krufky <mkrufky@linuxtv.org> | 9338 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 9338 | L: linux-media@vger.kernel.org | 9339 | L: linux-media@vger.kernel.org |
| 9339 | W: http://linuxtv.org/ | 9340 | W: http://linuxtv.org/ |
| 9340 | W: http://github.com/mkrufky | 9341 | W: http://github.com/mkrufky |
| 9341 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9342 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9342 | T: git git://linuxtv.org/mkrufky/tuners.git | 9343 | T: git git://linuxtv.org/mkrufky/tuners.git |
| 9343 | S: Maintained | 9344 | S: Maintained |
| 9344 | F: drivers/media/tuners/tda18271* | 9345 | F: drivers/media/tuners/tda18271* |
| 9345 | 9346 | ||
| 9346 | TDA827x MEDIA DRIVER | 9347 | TDA827x MEDIA DRIVER |
| 9347 | M: Michael Krufky <mkrufky@linuxtv.org> | 9348 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 9348 | L: linux-media@vger.kernel.org | 9349 | L: linux-media@vger.kernel.org |
| 9349 | W: http://linuxtv.org/ | 9350 | W: http://linuxtv.org/ |
| 9350 | W: http://github.com/mkrufky | 9351 | W: http://github.com/mkrufky |
| 9351 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9352 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9352 | T: git git://linuxtv.org/mkrufky/tuners.git | 9353 | T: git git://linuxtv.org/mkrufky/tuners.git |
| 9353 | S: Maintained | 9354 | S: Maintained |
| 9354 | F: drivers/media/tuners/tda8290.* | 9355 | F: drivers/media/tuners/tda8290.* |
| 9355 | 9356 | ||
| 9356 | TDA8290 MEDIA DRIVER | 9357 | TDA8290 MEDIA DRIVER |
| 9357 | M: Michael Krufky <mkrufky@linuxtv.org> | 9358 | M: Michael Krufky <mkrufky@linuxtv.org> |
| 9358 | L: linux-media@vger.kernel.org | 9359 | L: linux-media@vger.kernel.org |
| 9359 | W: http://linuxtv.org/ | 9360 | W: http://linuxtv.org/ |
| 9360 | W: http://github.com/mkrufky | 9361 | W: http://github.com/mkrufky |
| 9361 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9362 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9362 | T: git git://linuxtv.org/mkrufky/tuners.git | 9363 | T: git git://linuxtv.org/mkrufky/tuners.git |
| 9363 | S: Maintained | 9364 | S: Maintained |
| 9364 | F: drivers/media/tuners/tda8290.* | 9365 | F: drivers/media/tuners/tda8290.* |
| 9365 | 9366 | ||
| 9366 | TDA9840 MEDIA DRIVER | 9367 | TDA9840 MEDIA DRIVER |
| 9367 | M: Hans Verkuil <hverkuil@xs4all.nl> | 9368 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 9368 | L: linux-media@vger.kernel.org | 9369 | L: linux-media@vger.kernel.org |
| 9369 | T: git git://linuxtv.org/media_tree.git | 9370 | T: git git://linuxtv.org/media_tree.git |
| 9370 | W: http://linuxtv.org | 9371 | W: http://linuxtv.org |
| 9371 | S: Maintained | 9372 | S: Maintained |
| 9372 | F: drivers/media/i2c/tda9840* | 9373 | F: drivers/media/i2c/tda9840* |
| 9373 | 9374 | ||
| 9374 | TEA5761 TUNER DRIVER | 9375 | TEA5761 TUNER DRIVER |
| 9375 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 9376 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 9376 | L: linux-media@vger.kernel.org | 9377 | L: linux-media@vger.kernel.org |
| 9377 | W: http://linuxtv.org | 9378 | W: http://linuxtv.org |
| 9378 | T: git git://linuxtv.org/media_tree.git | 9379 | T: git git://linuxtv.org/media_tree.git |
| 9379 | S: Odd fixes | 9380 | S: Odd fixes |
| 9380 | F: drivers/media/tuners/tea5761.* | 9381 | F: drivers/media/tuners/tea5761.* |
| 9381 | 9382 | ||
| 9382 | TEA5767 TUNER DRIVER | 9383 | TEA5767 TUNER DRIVER |
| 9383 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 9384 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 9384 | L: linux-media@vger.kernel.org | 9385 | L: linux-media@vger.kernel.org |
| 9385 | W: http://linuxtv.org | 9386 | W: http://linuxtv.org |
| 9386 | T: git git://linuxtv.org/media_tree.git | 9387 | T: git git://linuxtv.org/media_tree.git |
| 9387 | S: Maintained | 9388 | S: Maintained |
| 9388 | F: drivers/media/tuners/tea5767.* | 9389 | F: drivers/media/tuners/tea5767.* |
| 9389 | 9390 | ||
| 9390 | TEA6415C MEDIA DRIVER | 9391 | TEA6415C MEDIA DRIVER |
| 9391 | M: Hans Verkuil <hverkuil@xs4all.nl> | 9392 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 9392 | L: linux-media@vger.kernel.org | 9393 | L: linux-media@vger.kernel.org |
| 9393 | T: git git://linuxtv.org/media_tree.git | 9394 | T: git git://linuxtv.org/media_tree.git |
| 9394 | W: http://linuxtv.org | 9395 | W: http://linuxtv.org |
| 9395 | S: Maintained | 9396 | S: Maintained |
| 9396 | F: drivers/media/i2c/tea6415c* | 9397 | F: drivers/media/i2c/tea6415c* |
| 9397 | 9398 | ||
| 9398 | TEA6420 MEDIA DRIVER | 9399 | TEA6420 MEDIA DRIVER |
| 9399 | M: Hans Verkuil <hverkuil@xs4all.nl> | 9400 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 9400 | L: linux-media@vger.kernel.org | 9401 | L: linux-media@vger.kernel.org |
| 9401 | T: git git://linuxtv.org/media_tree.git | 9402 | T: git git://linuxtv.org/media_tree.git |
| 9402 | W: http://linuxtv.org | 9403 | W: http://linuxtv.org |
| 9403 | S: Maintained | 9404 | S: Maintained |
| 9404 | F: drivers/media/i2c/tea6420* | 9405 | F: drivers/media/i2c/tea6420* |
| 9405 | 9406 | ||
| 9406 | TEAM DRIVER | 9407 | TEAM DRIVER |
| 9407 | M: Jiri Pirko <jiri@resnulli.us> | 9408 | M: Jiri Pirko <jiri@resnulli.us> |
| 9408 | L: netdev@vger.kernel.org | 9409 | L: netdev@vger.kernel.org |
| 9409 | S: Supported | 9410 | S: Supported |
| 9410 | F: drivers/net/team/ | 9411 | F: drivers/net/team/ |
| 9411 | F: include/linux/if_team.h | 9412 | F: include/linux/if_team.h |
| 9412 | F: include/uapi/linux/if_team.h | 9413 | F: include/uapi/linux/if_team.h |
| 9413 | 9414 | ||
| 9414 | TECHNOLOGIC SYSTEMS TS-5500 PLATFORM SUPPORT | 9415 | TECHNOLOGIC SYSTEMS TS-5500 PLATFORM SUPPORT |
| 9415 | M: Savoir-faire Linux Inc. <kernel@savoirfairelinux.com> | 9416 | M: Savoir-faire Linux Inc. <kernel@savoirfairelinux.com> |
| 9416 | S: Maintained | 9417 | S: Maintained |
| 9417 | F: arch/x86/platform/ts5500/ | 9418 | F: arch/x86/platform/ts5500/ |
| 9418 | 9419 | ||
| 9419 | TECHNOTREND USB IR RECEIVER | 9420 | TECHNOTREND USB IR RECEIVER |
| 9420 | M: Sean Young <sean@mess.org> | 9421 | M: Sean Young <sean@mess.org> |
| 9421 | L: linux-media@vger.kernel.org | 9422 | L: linux-media@vger.kernel.org |
| 9422 | S: Maintained | 9423 | S: Maintained |
| 9423 | F: drivers/media/rc/ttusbir.c | 9424 | F: drivers/media/rc/ttusbir.c |
| 9424 | 9425 | ||
| 9425 | TEGRA ARCHITECTURE SUPPORT | 9426 | TEGRA ARCHITECTURE SUPPORT |
| 9426 | M: Stephen Warren <swarren@wwwdotorg.org> | 9427 | M: Stephen Warren <swarren@wwwdotorg.org> |
| 9427 | M: Thierry Reding <thierry.reding@gmail.com> | 9428 | M: Thierry Reding <thierry.reding@gmail.com> |
| 9428 | M: Alexandre Courbot <gnurou@gmail.com> | 9429 | M: Alexandre Courbot <gnurou@gmail.com> |
| 9429 | L: linux-tegra@vger.kernel.org | 9430 | L: linux-tegra@vger.kernel.org |
| 9430 | Q: http://patchwork.ozlabs.org/project/linux-tegra/list/ | 9431 | Q: http://patchwork.ozlabs.org/project/linux-tegra/list/ |
| 9431 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git | 9432 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux.git |
| 9432 | S: Supported | 9433 | S: Supported |
| 9433 | N: [^a-z]tegra | 9434 | N: [^a-z]tegra |
| 9434 | 9435 | ||
| 9435 | TEGRA CLOCK DRIVER | 9436 | TEGRA CLOCK DRIVER |
| 9436 | M: Peter De Schrijver <pdeschrijver@nvidia.com> | 9437 | M: Peter De Schrijver <pdeschrijver@nvidia.com> |
| 9437 | M: Prashant Gaikwad <pgaikwad@nvidia.com> | 9438 | M: Prashant Gaikwad <pgaikwad@nvidia.com> |
| 9438 | S: Supported | 9439 | S: Supported |
| 9439 | F: drivers/clk/tegra/ | 9440 | F: drivers/clk/tegra/ |
| 9440 | 9441 | ||
| 9441 | TEGRA DMA DRIVER | 9442 | TEGRA DMA DRIVER |
| 9442 | M: Laxman Dewangan <ldewangan@nvidia.com> | 9443 | M: Laxman Dewangan <ldewangan@nvidia.com> |
| 9443 | S: Supported | 9444 | S: Supported |
| 9444 | F: drivers/dma/tegra20-apb-dma.c | 9445 | F: drivers/dma/tegra20-apb-dma.c |
| 9445 | 9446 | ||
| 9446 | TEGRA I2C DRIVER | 9447 | TEGRA I2C DRIVER |
| 9447 | M: Laxman Dewangan <ldewangan@nvidia.com> | 9448 | M: Laxman Dewangan <ldewangan@nvidia.com> |
| 9448 | S: Supported | 9449 | S: Supported |
| 9449 | F: drivers/i2c/busses/i2c-tegra.c | 9450 | F: drivers/i2c/busses/i2c-tegra.c |
| 9450 | 9451 | ||
| 9451 | TEGRA IOMMU DRIVERS | 9452 | TEGRA IOMMU DRIVERS |
| 9452 | M: Hiroshi Doyu <hdoyu@nvidia.com> | 9453 | M: Hiroshi Doyu <hdoyu@nvidia.com> |
| 9453 | S: Supported | 9454 | S: Supported |
| 9454 | F: drivers/iommu/tegra* | 9455 | F: drivers/iommu/tegra* |
| 9455 | 9456 | ||
| 9456 | TEGRA KBC DRIVER | 9457 | TEGRA KBC DRIVER |
| 9457 | M: Rakesh Iyer <riyer@nvidia.com> | 9458 | M: Rakesh Iyer <riyer@nvidia.com> |
| 9458 | M: Laxman Dewangan <ldewangan@nvidia.com> | 9459 | M: Laxman Dewangan <ldewangan@nvidia.com> |
| 9459 | S: Supported | 9460 | S: Supported |
| 9460 | F: drivers/input/keyboard/tegra-kbc.c | 9461 | F: drivers/input/keyboard/tegra-kbc.c |
| 9461 | 9462 | ||
| 9462 | TEGRA PWM DRIVER | 9463 | TEGRA PWM DRIVER |
| 9463 | M: Thierry Reding <thierry.reding@gmail.com> | 9464 | M: Thierry Reding <thierry.reding@gmail.com> |
| 9464 | S: Supported | 9465 | S: Supported |
| 9465 | F: drivers/pwm/pwm-tegra.c | 9466 | F: drivers/pwm/pwm-tegra.c |
| 9466 | 9467 | ||
| 9467 | TEGRA SERIAL DRIVER | 9468 | TEGRA SERIAL DRIVER |
| 9468 | M: Laxman Dewangan <ldewangan@nvidia.com> | 9469 | M: Laxman Dewangan <ldewangan@nvidia.com> |
| 9469 | S: Supported | 9470 | S: Supported |
| 9470 | F: drivers/tty/serial/serial-tegra.c | 9471 | F: drivers/tty/serial/serial-tegra.c |
| 9471 | 9472 | ||
| 9472 | TEGRA SPI DRIVER | 9473 | TEGRA SPI DRIVER |
| 9473 | M: Laxman Dewangan <ldewangan@nvidia.com> | 9474 | M: Laxman Dewangan <ldewangan@nvidia.com> |
| 9474 | S: Supported | 9475 | S: Supported |
| 9475 | F: drivers/spi/spi-tegra* | 9476 | F: drivers/spi/spi-tegra* |
| 9476 | 9477 | ||
| 9477 | TEHUTI ETHERNET DRIVER | 9478 | TEHUTI ETHERNET DRIVER |
| 9478 | M: Andy Gospodarek <andy@greyhouse.net> | 9479 | M: Andy Gospodarek <andy@greyhouse.net> |
| 9479 | L: netdev@vger.kernel.org | 9480 | L: netdev@vger.kernel.org |
| 9480 | S: Supported | 9481 | S: Supported |
| 9481 | F: drivers/net/ethernet/tehuti/* | 9482 | F: drivers/net/ethernet/tehuti/* |
| 9482 | 9483 | ||
| 9483 | Telecom Clock Driver for MCPL0010 | 9484 | Telecom Clock Driver for MCPL0010 |
| 9484 | M: Mark Gross <mark.gross@intel.com> | 9485 | M: Mark Gross <mark.gross@intel.com> |
| 9485 | S: Supported | 9486 | S: Supported |
| 9486 | F: drivers/char/tlclk.c | 9487 | F: drivers/char/tlclk.c |
| 9487 | 9488 | ||
| 9488 | TENSILICA XTENSA PORT (xtensa) | 9489 | TENSILICA XTENSA PORT (xtensa) |
| 9489 | M: Chris Zankel <chris@zankel.net> | 9490 | M: Chris Zankel <chris@zankel.net> |
| 9490 | M: Max Filippov <jcmvbkbc@gmail.com> | 9491 | M: Max Filippov <jcmvbkbc@gmail.com> |
| 9491 | L: linux-xtensa@linux-xtensa.org | 9492 | L: linux-xtensa@linux-xtensa.org |
| 9492 | S: Maintained | 9493 | S: Maintained |
| 9493 | F: arch/xtensa/ | 9494 | F: arch/xtensa/ |
| 9494 | F: drivers/irqchip/irq-xtensa-* | 9495 | F: drivers/irqchip/irq-xtensa-* |
| 9495 | 9496 | ||
| 9496 | THANKO'S RAREMONO AM/FM/SW RADIO RECEIVER USB DRIVER | 9497 | THANKO'S RAREMONO AM/FM/SW RADIO RECEIVER USB DRIVER |
| 9497 | M: Hans Verkuil <hverkuil@xs4all.nl> | 9498 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 9498 | L: linux-media@vger.kernel.org | 9499 | L: linux-media@vger.kernel.org |
| 9499 | T: git git://linuxtv.org/media_tree.git | 9500 | T: git git://linuxtv.org/media_tree.git |
| 9500 | W: http://linuxtv.org | 9501 | W: http://linuxtv.org |
| 9501 | S: Maintained | 9502 | S: Maintained |
| 9502 | F: drivers/media/radio/radio-raremono.c | 9503 | F: drivers/media/radio/radio-raremono.c |
| 9503 | 9504 | ||
| 9504 | THERMAL | 9505 | THERMAL |
| 9505 | M: Zhang Rui <rui.zhang@intel.com> | 9506 | M: Zhang Rui <rui.zhang@intel.com> |
| 9506 | M: Eduardo Valentin <edubezval@gmail.com> | 9507 | M: Eduardo Valentin <edubezval@gmail.com> |
| 9507 | L: linux-pm@vger.kernel.org | 9508 | L: linux-pm@vger.kernel.org |
| 9508 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux.git | 9509 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux.git |
| 9509 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/evalenti/linux-soc-thermal.git | 9510 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/evalenti/linux-soc-thermal.git |
| 9510 | Q: https://patchwork.kernel.org/project/linux-pm/list/ | 9511 | Q: https://patchwork.kernel.org/project/linux-pm/list/ |
| 9511 | S: Supported | 9512 | S: Supported |
| 9512 | F: drivers/thermal/ | 9513 | F: drivers/thermal/ |
| 9513 | F: include/linux/thermal.h | 9514 | F: include/linux/thermal.h |
| 9514 | F: include/uapi/linux/thermal.h | 9515 | F: include/uapi/linux/thermal.h |
| 9515 | F: include/linux/cpu_cooling.h | 9516 | F: include/linux/cpu_cooling.h |
| 9516 | F: Documentation/devicetree/bindings/thermal/ | 9517 | F: Documentation/devicetree/bindings/thermal/ |
| 9517 | 9518 | ||
| 9518 | THINGM BLINK(1) USB RGB LED DRIVER | 9519 | THINGM BLINK(1) USB RGB LED DRIVER |
| 9519 | M: Vivien Didelot <vivien.didelot@savoirfairelinux.com> | 9520 | M: Vivien Didelot <vivien.didelot@savoirfairelinux.com> |
| 9520 | S: Maintained | 9521 | S: Maintained |
| 9521 | F: drivers/hid/hid-thingm.c | 9522 | F: drivers/hid/hid-thingm.c |
| 9522 | 9523 | ||
| 9523 | THINKPAD ACPI EXTRAS DRIVER | 9524 | THINKPAD ACPI EXTRAS DRIVER |
| 9524 | M: Henrique de Moraes Holschuh <ibm-acpi@hmh.eng.br> | 9525 | M: Henrique de Moraes Holschuh <ibm-acpi@hmh.eng.br> |
| 9525 | L: ibm-acpi-devel@lists.sourceforge.net | 9526 | L: ibm-acpi-devel@lists.sourceforge.net |
| 9526 | L: platform-driver-x86@vger.kernel.org | 9527 | L: platform-driver-x86@vger.kernel.org |
| 9527 | W: http://ibm-acpi.sourceforge.net | 9528 | W: http://ibm-acpi.sourceforge.net |
| 9528 | W: http://thinkwiki.org/wiki/Ibm-acpi | 9529 | W: http://thinkwiki.org/wiki/Ibm-acpi |
| 9529 | T: git git://repo.or.cz/linux-2.6/linux-acpi-2.6/ibm-acpi-2.6.git | 9530 | T: git git://repo.or.cz/linux-2.6/linux-acpi-2.6/ibm-acpi-2.6.git |
| 9530 | S: Maintained | 9531 | S: Maintained |
| 9531 | F: drivers/platform/x86/thinkpad_acpi.c | 9532 | F: drivers/platform/x86/thinkpad_acpi.c |
| 9532 | 9533 | ||
| 9533 | TI BANDGAP AND THERMAL DRIVER | 9534 | TI BANDGAP AND THERMAL DRIVER |
| 9534 | M: Eduardo Valentin <edubezval@gmail.com> | 9535 | M: Eduardo Valentin <edubezval@gmail.com> |
| 9535 | L: linux-pm@vger.kernel.org | 9536 | L: linux-pm@vger.kernel.org |
| 9536 | S: Supported | 9537 | S: Supported |
| 9537 | F: drivers/thermal/ti-soc-thermal/ | 9538 | F: drivers/thermal/ti-soc-thermal/ |
| 9538 | 9539 | ||
| 9539 | TI CLOCK DRIVER | 9540 | TI CLOCK DRIVER |
| 9540 | M: Tero Kristo <t-kristo@ti.com> | 9541 | M: Tero Kristo <t-kristo@ti.com> |
| 9541 | L: linux-omap@vger.kernel.org | 9542 | L: linux-omap@vger.kernel.org |
| 9542 | S: Maintained | 9543 | S: Maintained |
| 9543 | F: drivers/clk/ti/ | 9544 | F: drivers/clk/ti/ |
| 9544 | F: include/linux/clk/ti.h | 9545 | F: include/linux/clk/ti.h |
| 9545 | 9546 | ||
| 9546 | TI FLASH MEDIA INTERFACE DRIVER | 9547 | TI FLASH MEDIA INTERFACE DRIVER |
| 9547 | M: Alex Dubov <oakad@yahoo.com> | 9548 | M: Alex Dubov <oakad@yahoo.com> |
| 9548 | S: Maintained | 9549 | S: Maintained |
| 9549 | F: drivers/misc/tifm* | 9550 | F: drivers/misc/tifm* |
| 9550 | F: drivers/mmc/host/tifm_sd.c | 9551 | F: drivers/mmc/host/tifm_sd.c |
| 9551 | F: include/linux/tifm.h | 9552 | F: include/linux/tifm.h |
| 9552 | 9553 | ||
| 9553 | TI KEYSTONE MULTICORE NAVIGATOR DRIVERS | 9554 | TI KEYSTONE MULTICORE NAVIGATOR DRIVERS |
| 9554 | M: Santosh Shilimkar <ssantosh@kernel.org> | 9555 | M: Santosh Shilimkar <ssantosh@kernel.org> |
| 9555 | L: linux-kernel@vger.kernel.org | 9556 | L: linux-kernel@vger.kernel.org |
| 9556 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) | 9557 | L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) |
| 9557 | S: Maintained | 9558 | S: Maintained |
| 9558 | F: drivers/soc/ti/* | 9559 | F: drivers/soc/ti/* |
| 9559 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git | 9560 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone.git |
| 9560 | 9561 | ||
| 9561 | 9562 | ||
| 9562 | TI LM49xxx FAMILY ASoC CODEC DRIVERS | 9563 | TI LM49xxx FAMILY ASoC CODEC DRIVERS |
| 9563 | M: M R Swami Reddy <mr.swami.reddy@ti.com> | 9564 | M: M R Swami Reddy <mr.swami.reddy@ti.com> |
| 9564 | M: Vishwas A Deshpande <vishwas.a.deshpande@ti.com> | 9565 | M: Vishwas A Deshpande <vishwas.a.deshpande@ti.com> |
| 9565 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 9566 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 9566 | S: Maintained | 9567 | S: Maintained |
| 9567 | F: sound/soc/codecs/lm49453* | 9568 | F: sound/soc/codecs/lm49453* |
| 9568 | F: sound/soc/codecs/isabelle* | 9569 | F: sound/soc/codecs/isabelle* |
| 9569 | 9570 | ||
| 9570 | TI LP855x BACKLIGHT DRIVER | 9571 | TI LP855x BACKLIGHT DRIVER |
| 9571 | M: Milo Kim <milo.kim@ti.com> | 9572 | M: Milo Kim <milo.kim@ti.com> |
| 9572 | S: Maintained | 9573 | S: Maintained |
| 9573 | F: Documentation/backlight/lp855x-driver.txt | 9574 | F: Documentation/backlight/lp855x-driver.txt |
| 9574 | F: drivers/video/backlight/lp855x_bl.c | 9575 | F: drivers/video/backlight/lp855x_bl.c |
| 9575 | F: include/linux/platform_data/lp855x.h | 9576 | F: include/linux/platform_data/lp855x.h |
| 9576 | 9577 | ||
| 9577 | TI LP8727 CHARGER DRIVER | 9578 | TI LP8727 CHARGER DRIVER |
| 9578 | M: Milo Kim <milo.kim@ti.com> | 9579 | M: Milo Kim <milo.kim@ti.com> |
| 9579 | S: Maintained | 9580 | S: Maintained |
| 9580 | F: drivers/power/lp8727_charger.c | 9581 | F: drivers/power/lp8727_charger.c |
| 9581 | F: include/linux/platform_data/lp8727.h | 9582 | F: include/linux/platform_data/lp8727.h |
| 9582 | 9583 | ||
| 9583 | TI LP8788 MFD DRIVER | 9584 | TI LP8788 MFD DRIVER |
| 9584 | M: Milo Kim <milo.kim@ti.com> | 9585 | M: Milo Kim <milo.kim@ti.com> |
| 9585 | S: Maintained | 9586 | S: Maintained |
| 9586 | F: drivers/iio/adc/lp8788_adc.c | 9587 | F: drivers/iio/adc/lp8788_adc.c |
| 9587 | F: drivers/leds/leds-lp8788.c | 9588 | F: drivers/leds/leds-lp8788.c |
| 9588 | F: drivers/mfd/lp8788*.c | 9589 | F: drivers/mfd/lp8788*.c |
| 9589 | F: drivers/power/lp8788-charger.c | 9590 | F: drivers/power/lp8788-charger.c |
| 9590 | F: drivers/regulator/lp8788-*.c | 9591 | F: drivers/regulator/lp8788-*.c |
| 9591 | F: include/linux/mfd/lp8788*.h | 9592 | F: include/linux/mfd/lp8788*.h |
| 9592 | 9593 | ||
| 9593 | TI TWL4030 SERIES SOC CODEC DRIVER | 9594 | TI TWL4030 SERIES SOC CODEC DRIVER |
| 9594 | M: Peter Ujfalusi <peter.ujfalusi@ti.com> | 9595 | M: Peter Ujfalusi <peter.ujfalusi@ti.com> |
| 9595 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 9596 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 9596 | S: Maintained | 9597 | S: Maintained |
| 9597 | F: sound/soc/codecs/twl4030* | 9598 | F: sound/soc/codecs/twl4030* |
| 9598 | 9599 | ||
| 9599 | TI WILINK WIRELESS DRIVERS | 9600 | TI WILINK WIRELESS DRIVERS |
| 9600 | L: linux-wireless@vger.kernel.org | 9601 | L: linux-wireless@vger.kernel.org |
| 9601 | W: http://wireless.kernel.org/en/users/Drivers/wl12xx | 9602 | W: http://wireless.kernel.org/en/users/Drivers/wl12xx |
| 9602 | W: http://wireless.kernel.org/en/users/Drivers/wl1251 | 9603 | W: http://wireless.kernel.org/en/users/Drivers/wl1251 |
| 9603 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/luca/wl12xx.git | 9604 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/luca/wl12xx.git |
| 9604 | S: Orphan | 9605 | S: Orphan |
| 9605 | F: drivers/net/wireless/ti/ | 9606 | F: drivers/net/wireless/ti/ |
| 9606 | F: include/linux/wl12xx.h | 9607 | F: include/linux/wl12xx.h |
| 9607 | 9608 | ||
| 9608 | TIPC NETWORK LAYER | 9609 | TIPC NETWORK LAYER |
| 9609 | M: Jon Maloy <jon.maloy@ericsson.com> | 9610 | M: Jon Maloy <jon.maloy@ericsson.com> |
| 9610 | M: Allan Stephens <allan.stephens@windriver.com> | 9611 | M: Allan Stephens <allan.stephens@windriver.com> |
| 9611 | L: netdev@vger.kernel.org (core kernel code) | 9612 | L: netdev@vger.kernel.org (core kernel code) |
| 9612 | L: tipc-discussion@lists.sourceforge.net (user apps, general discussion) | 9613 | L: tipc-discussion@lists.sourceforge.net (user apps, general discussion) |
| 9613 | W: http://tipc.sourceforge.net/ | 9614 | W: http://tipc.sourceforge.net/ |
| 9614 | S: Maintained | 9615 | S: Maintained |
| 9615 | F: include/uapi/linux/tipc*.h | 9616 | F: include/uapi/linux/tipc*.h |
| 9616 | F: net/tipc/ | 9617 | F: net/tipc/ |
| 9617 | 9618 | ||
| 9618 | TILE ARCHITECTURE | 9619 | TILE ARCHITECTURE |
| 9619 | M: Chris Metcalf <cmetcalf@ezchip.com> | 9620 | M: Chris Metcalf <cmetcalf@ezchip.com> |
| 9620 | W: http://www.tilera.com/scm/ | 9621 | W: http://www.tilera.com/scm/ |
| 9621 | S: Supported | 9622 | S: Supported |
| 9622 | F: arch/tile/ | 9623 | F: arch/tile/ |
| 9623 | F: drivers/char/tile-srom.c | 9624 | F: drivers/char/tile-srom.c |
| 9624 | F: drivers/edac/tile_edac.c | 9625 | F: drivers/edac/tile_edac.c |
| 9625 | F: drivers/net/ethernet/tile/ | 9626 | F: drivers/net/ethernet/tile/ |
| 9626 | F: drivers/rtc/rtc-tile.c | 9627 | F: drivers/rtc/rtc-tile.c |
| 9627 | F: drivers/tty/hvc/hvc_tile.c | 9628 | F: drivers/tty/hvc/hvc_tile.c |
| 9628 | F: drivers/tty/serial/tilegx.c | 9629 | F: drivers/tty/serial/tilegx.c |
| 9629 | F: drivers/usb/host/*-tilegx.c | 9630 | F: drivers/usb/host/*-tilegx.c |
| 9630 | F: include/linux/usb/tilegx.h | 9631 | F: include/linux/usb/tilegx.h |
| 9631 | 9632 | ||
| 9632 | TLAN NETWORK DRIVER | 9633 | TLAN NETWORK DRIVER |
| 9633 | M: Samuel Chessman <chessman@tux.org> | 9634 | M: Samuel Chessman <chessman@tux.org> |
| 9634 | L: tlan-devel@lists.sourceforge.net (subscribers-only) | 9635 | L: tlan-devel@lists.sourceforge.net (subscribers-only) |
| 9635 | W: http://sourceforge.net/projects/tlan/ | 9636 | W: http://sourceforge.net/projects/tlan/ |
| 9636 | S: Maintained | 9637 | S: Maintained |
| 9637 | F: Documentation/networking/tlan.txt | 9638 | F: Documentation/networking/tlan.txt |
| 9638 | F: drivers/net/ethernet/ti/tlan.* | 9639 | F: drivers/net/ethernet/ti/tlan.* |
| 9639 | 9640 | ||
| 9640 | TOMOYO SECURITY MODULE | 9641 | TOMOYO SECURITY MODULE |
| 9641 | M: Kentaro Takeda <takedakn@nttdata.co.jp> | 9642 | M: Kentaro Takeda <takedakn@nttdata.co.jp> |
| 9642 | M: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> | 9643 | M: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> |
| 9643 | L: tomoyo-dev-en@lists.sourceforge.jp (subscribers-only, for developers in English) | 9644 | L: tomoyo-dev-en@lists.sourceforge.jp (subscribers-only, for developers in English) |
| 9644 | L: tomoyo-users-en@lists.sourceforge.jp (subscribers-only, for users in English) | 9645 | L: tomoyo-users-en@lists.sourceforge.jp (subscribers-only, for users in English) |
| 9645 | L: tomoyo-dev@lists.sourceforge.jp (subscribers-only, for developers in Japanese) | 9646 | L: tomoyo-dev@lists.sourceforge.jp (subscribers-only, for developers in Japanese) |
| 9646 | L: tomoyo-users@lists.sourceforge.jp (subscribers-only, for users in Japanese) | 9647 | L: tomoyo-users@lists.sourceforge.jp (subscribers-only, for users in Japanese) |
| 9647 | W: http://tomoyo.sourceforge.jp/ | 9648 | W: http://tomoyo.sourceforge.jp/ |
| 9648 | T: quilt http://svn.sourceforge.jp/svnroot/tomoyo/trunk/2.5.x/tomoyo-lsm/patches/ | 9649 | T: quilt http://svn.sourceforge.jp/svnroot/tomoyo/trunk/2.5.x/tomoyo-lsm/patches/ |
| 9649 | S: Maintained | 9650 | S: Maintained |
| 9650 | F: security/tomoyo/ | 9651 | F: security/tomoyo/ |
| 9651 | 9652 | ||
| 9652 | TOPSTAR LAPTOP EXTRAS DRIVER | 9653 | TOPSTAR LAPTOP EXTRAS DRIVER |
| 9653 | M: Herton Ronaldo Krzesinski <herton@canonical.com> | 9654 | M: Herton Ronaldo Krzesinski <herton@canonical.com> |
| 9654 | L: platform-driver-x86@vger.kernel.org | 9655 | L: platform-driver-x86@vger.kernel.org |
| 9655 | S: Maintained | 9656 | S: Maintained |
| 9656 | F: drivers/platform/x86/topstar-laptop.c | 9657 | F: drivers/platform/x86/topstar-laptop.c |
| 9657 | 9658 | ||
| 9658 | TOSHIBA ACPI EXTRAS DRIVER | 9659 | TOSHIBA ACPI EXTRAS DRIVER |
| 9659 | L: platform-driver-x86@vger.kernel.org | 9660 | L: platform-driver-x86@vger.kernel.org |
| 9660 | S: Orphan | 9661 | S: Orphan |
| 9661 | F: drivers/platform/x86/toshiba_acpi.c | 9662 | F: drivers/platform/x86/toshiba_acpi.c |
| 9662 | 9663 | ||
| 9663 | TOSHIBA SMM DRIVER | 9664 | TOSHIBA SMM DRIVER |
| 9664 | M: Jonathan Buzzard <jonathan@buzzard.org.uk> | 9665 | M: Jonathan Buzzard <jonathan@buzzard.org.uk> |
| 9665 | L: tlinux-users@tce.toshiba-dme.co.jp | 9666 | L: tlinux-users@tce.toshiba-dme.co.jp |
| 9666 | W: http://www.buzzard.org.uk/toshiba/ | 9667 | W: http://www.buzzard.org.uk/toshiba/ |
| 9667 | S: Maintained | 9668 | S: Maintained |
| 9668 | F: drivers/char/toshiba.c | 9669 | F: drivers/char/toshiba.c |
| 9669 | F: include/linux/toshiba.h | 9670 | F: include/linux/toshiba.h |
| 9670 | F: include/uapi/linux/toshiba.h | 9671 | F: include/uapi/linux/toshiba.h |
| 9671 | 9672 | ||
| 9672 | TMIO MMC DRIVER | 9673 | TMIO MMC DRIVER |
| 9673 | M: Ian Molton <ian.molton@codethink.co.uk> | 9674 | M: Ian Molton <ian.molton@codethink.co.uk> |
| 9674 | L: linux-mmc@vger.kernel.org | 9675 | L: linux-mmc@vger.kernel.org |
| 9675 | S: Maintained | 9676 | S: Maintained |
| 9676 | F: drivers/mmc/host/tmio_mmc* | 9677 | F: drivers/mmc/host/tmio_mmc* |
| 9677 | F: drivers/mmc/host/sh_mobile_sdhi.c | 9678 | F: drivers/mmc/host/sh_mobile_sdhi.c |
| 9678 | F: include/linux/mmc/tmio.h | 9679 | F: include/linux/mmc/tmio.h |
| 9679 | F: include/linux/mmc/sh_mobile_sdhi.h | 9680 | F: include/linux/mmc/sh_mobile_sdhi.h |
| 9680 | 9681 | ||
| 9681 | TMP401 HARDWARE MONITOR DRIVER | 9682 | TMP401 HARDWARE MONITOR DRIVER |
| 9682 | M: Guenter Roeck <linux@roeck-us.net> | 9683 | M: Guenter Roeck <linux@roeck-us.net> |
| 9683 | L: lm-sensors@lm-sensors.org | 9684 | L: lm-sensors@lm-sensors.org |
| 9684 | S: Maintained | 9685 | S: Maintained |
| 9685 | F: Documentation/hwmon/tmp401 | 9686 | F: Documentation/hwmon/tmp401 |
| 9686 | F: drivers/hwmon/tmp401.c | 9687 | F: drivers/hwmon/tmp401.c |
| 9687 | 9688 | ||
| 9688 | TMPFS (SHMEM FILESYSTEM) | 9689 | TMPFS (SHMEM FILESYSTEM) |
| 9689 | M: Hugh Dickins <hughd@google.com> | 9690 | M: Hugh Dickins <hughd@google.com> |
| 9690 | L: linux-mm@kvack.org | 9691 | L: linux-mm@kvack.org |
| 9691 | S: Maintained | 9692 | S: Maintained |
| 9692 | F: include/linux/shmem_fs.h | 9693 | F: include/linux/shmem_fs.h |
| 9693 | F: mm/shmem.c | 9694 | F: mm/shmem.c |
| 9694 | 9695 | ||
| 9695 | TM6000 VIDEO4LINUX DRIVER | 9696 | TM6000 VIDEO4LINUX DRIVER |
| 9696 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 9697 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 9697 | L: linux-media@vger.kernel.org | 9698 | L: linux-media@vger.kernel.org |
| 9698 | W: http://linuxtv.org | 9699 | W: http://linuxtv.org |
| 9699 | T: git git://linuxtv.org/media_tree.git | 9700 | T: git git://linuxtv.org/media_tree.git |
| 9700 | S: Odd fixes | 9701 | S: Odd fixes |
| 9701 | F: drivers/media/usb/tm6000/ | 9702 | F: drivers/media/usb/tm6000/ |
| 9702 | 9703 | ||
| 9703 | TW68 VIDEO4LINUX DRIVER | 9704 | TW68 VIDEO4LINUX DRIVER |
| 9704 | M: Hans Verkuil <hverkuil@xs4all.nl> | 9705 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 9705 | L: linux-media@vger.kernel.org | 9706 | L: linux-media@vger.kernel.org |
| 9706 | T: git git://linuxtv.org/media_tree.git | 9707 | T: git git://linuxtv.org/media_tree.git |
| 9707 | W: http://linuxtv.org | 9708 | W: http://linuxtv.org |
| 9708 | S: Odd Fixes | 9709 | S: Odd Fixes |
| 9709 | F: drivers/media/pci/tw68/ | 9710 | F: drivers/media/pci/tw68/ |
| 9710 | 9711 | ||
| 9711 | TPM DEVICE DRIVER | 9712 | TPM DEVICE DRIVER |
| 9712 | M: Peter Huewe <peterhuewe@gmx.de> | 9713 | M: Peter Huewe <peterhuewe@gmx.de> |
| 9713 | M: Ashley Lai <ashley@ashleylai.com> | 9714 | M: Ashley Lai <ashley@ashleylai.com> |
| 9714 | M: Marcel Selhorst <tpmdd@selhorst.net> | 9715 | M: Marcel Selhorst <tpmdd@selhorst.net> |
| 9715 | W: http://tpmdd.sourceforge.net | 9716 | W: http://tpmdd.sourceforge.net |
| 9716 | L: tpmdd-devel@lists.sourceforge.net (moderated for non-subscribers) | 9717 | L: tpmdd-devel@lists.sourceforge.net (moderated for non-subscribers) |
| 9717 | S: Maintained | 9718 | S: Maintained |
| 9718 | F: drivers/char/tpm/ | 9719 | F: drivers/char/tpm/ |
| 9719 | 9720 | ||
| 9720 | TRACING | 9721 | TRACING |
| 9721 | M: Steven Rostedt <rostedt@goodmis.org> | 9722 | M: Steven Rostedt <rostedt@goodmis.org> |
| 9722 | M: Ingo Molnar <mingo@redhat.com> | 9723 | M: Ingo Molnar <mingo@redhat.com> |
| 9723 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core | 9724 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git perf/core |
| 9724 | S: Maintained | 9725 | S: Maintained |
| 9725 | F: Documentation/trace/ftrace.txt | 9726 | F: Documentation/trace/ftrace.txt |
| 9726 | F: arch/*/*/*/ftrace.h | 9727 | F: arch/*/*/*/ftrace.h |
| 9727 | F: arch/*/kernel/ftrace.c | 9728 | F: arch/*/kernel/ftrace.c |
| 9728 | F: include/*/ftrace.h | 9729 | F: include/*/ftrace.h |
| 9729 | F: include/linux/trace*.h | 9730 | F: include/linux/trace*.h |
| 9730 | F: include/trace/ | 9731 | F: include/trace/ |
| 9731 | F: kernel/trace/ | 9732 | F: kernel/trace/ |
| 9732 | F: tools/testing/selftests/ftrace/ | 9733 | F: tools/testing/selftests/ftrace/ |
| 9733 | 9734 | ||
| 9734 | TRIVIAL PATCHES | 9735 | TRIVIAL PATCHES |
| 9735 | M: Jiri Kosina <trivial@kernel.org> | 9736 | M: Jiri Kosina <trivial@kernel.org> |
| 9736 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial.git | 9737 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial.git |
| 9737 | S: Maintained | 9738 | S: Maintained |
| 9738 | K: ^Subject:.*(?i)trivial | 9739 | K: ^Subject:.*(?i)trivial |
| 9739 | 9740 | ||
| 9740 | TTY LAYER | 9741 | TTY LAYER |
| 9741 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 9742 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 9742 | M: Jiri Slaby <jslaby@suse.cz> | 9743 | M: Jiri Slaby <jslaby@suse.cz> |
| 9743 | S: Supported | 9744 | S: Supported |
| 9744 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git | 9745 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git |
| 9745 | F: drivers/tty/ | 9746 | F: drivers/tty/ |
| 9746 | F: drivers/tty/serial/serial_core.c | 9747 | F: drivers/tty/serial/serial_core.c |
| 9747 | F: include/linux/serial_core.h | 9748 | F: include/linux/serial_core.h |
| 9748 | F: include/linux/serial.h | 9749 | F: include/linux/serial.h |
| 9749 | F: include/linux/tty.h | 9750 | F: include/linux/tty.h |
| 9750 | F: include/uapi/linux/serial_core.h | 9751 | F: include/uapi/linux/serial_core.h |
| 9751 | F: include/uapi/linux/serial.h | 9752 | F: include/uapi/linux/serial.h |
| 9752 | F: include/uapi/linux/tty.h | 9753 | F: include/uapi/linux/tty.h |
| 9753 | 9754 | ||
| 9754 | TUA9001 MEDIA DRIVER | 9755 | TUA9001 MEDIA DRIVER |
| 9755 | M: Antti Palosaari <crope@iki.fi> | 9756 | M: Antti Palosaari <crope@iki.fi> |
| 9756 | L: linux-media@vger.kernel.org | 9757 | L: linux-media@vger.kernel.org |
| 9757 | W: http://linuxtv.org/ | 9758 | W: http://linuxtv.org/ |
| 9758 | W: http://palosaari.fi/linux/ | 9759 | W: http://palosaari.fi/linux/ |
| 9759 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ | 9760 | Q: http://patchwork.linuxtv.org/project/linux-media/list/ |
| 9760 | T: git git://linuxtv.org/anttip/media_tree.git | 9761 | T: git git://linuxtv.org/anttip/media_tree.git |
| 9761 | S: Maintained | 9762 | S: Maintained |
| 9762 | F: drivers/media/tuners/tua9001* | 9763 | F: drivers/media/tuners/tua9001* |
| 9763 | 9764 | ||
| 9764 | TULIP NETWORK DRIVERS | 9765 | TULIP NETWORK DRIVERS |
| 9765 | M: Grant Grundler <grundler@parisc-linux.org> | 9766 | M: Grant Grundler <grundler@parisc-linux.org> |
| 9766 | L: netdev@vger.kernel.org | 9767 | L: netdev@vger.kernel.org |
| 9767 | S: Maintained | 9768 | S: Maintained |
| 9768 | F: drivers/net/ethernet/dec/tulip/ | 9769 | F: drivers/net/ethernet/dec/tulip/ |
| 9769 | 9770 | ||
| 9770 | TUN/TAP driver | 9771 | TUN/TAP driver |
| 9771 | M: Maxim Krasnyansky <maxk@qti.qualcomm.com> | 9772 | M: Maxim Krasnyansky <maxk@qti.qualcomm.com> |
| 9772 | W: http://vtun.sourceforge.net/tun | 9773 | W: http://vtun.sourceforge.net/tun |
| 9773 | S: Maintained | 9774 | S: Maintained |
| 9774 | F: Documentation/networking/tuntap.txt | 9775 | F: Documentation/networking/tuntap.txt |
| 9775 | F: arch/um/os-Linux/drivers/ | 9776 | F: arch/um/os-Linux/drivers/ |
| 9776 | 9777 | ||
| 9777 | TURBOCHANNEL SUBSYSTEM | 9778 | TURBOCHANNEL SUBSYSTEM |
| 9778 | M: "Maciej W. Rozycki" <macro@linux-mips.org> | 9779 | M: "Maciej W. Rozycki" <macro@linux-mips.org> |
| 9779 | M: Ralf Baechle <ralf@linux-mips.org> | 9780 | M: Ralf Baechle <ralf@linux-mips.org> |
| 9780 | L: linux-mips@linux-mips.org | 9781 | L: linux-mips@linux-mips.org |
| 9781 | Q: http://patchwork.linux-mips.org/project/linux-mips/list/ | 9782 | Q: http://patchwork.linux-mips.org/project/linux-mips/list/ |
| 9782 | S: Maintained | 9783 | S: Maintained |
| 9783 | F: drivers/tc/ | 9784 | F: drivers/tc/ |
| 9784 | F: include/linux/tc.h | 9785 | F: include/linux/tc.h |
| 9785 | 9786 | ||
| 9786 | U14-34F SCSI DRIVER | 9787 | U14-34F SCSI DRIVER |
| 9787 | M: Dario Ballabio <ballabio_dario@emc.com> | 9788 | M: Dario Ballabio <ballabio_dario@emc.com> |
| 9788 | L: linux-scsi@vger.kernel.org | 9789 | L: linux-scsi@vger.kernel.org |
| 9789 | S: Maintained | 9790 | S: Maintained |
| 9790 | F: drivers/scsi/u14-34f.c | 9791 | F: drivers/scsi/u14-34f.c |
| 9791 | 9792 | ||
| 9792 | UBI FILE SYSTEM (UBIFS) | 9793 | UBI FILE SYSTEM (UBIFS) |
| 9793 | M: Artem Bityutskiy <dedekind1@gmail.com> | 9794 | M: Artem Bityutskiy <dedekind1@gmail.com> |
| 9794 | M: Adrian Hunter <adrian.hunter@intel.com> | 9795 | M: Adrian Hunter <adrian.hunter@intel.com> |
| 9795 | L: linux-mtd@lists.infradead.org | 9796 | L: linux-mtd@lists.infradead.org |
| 9796 | T: git git://git.infradead.org/ubifs-2.6.git | 9797 | T: git git://git.infradead.org/ubifs-2.6.git |
| 9797 | W: http://www.linux-mtd.infradead.org/doc/ubifs.html | 9798 | W: http://www.linux-mtd.infradead.org/doc/ubifs.html |
| 9798 | S: Maintained | 9799 | S: Maintained |
| 9799 | F: Documentation/filesystems/ubifs.txt | 9800 | F: Documentation/filesystems/ubifs.txt |
| 9800 | F: fs/ubifs/ | 9801 | F: fs/ubifs/ |
| 9801 | 9802 | ||
| 9802 | UCLINUX (AND M68KNOMMU) | 9803 | UCLINUX (AND M68KNOMMU) |
| 9803 | M: Greg Ungerer <gerg@uclinux.org> | 9804 | M: Greg Ungerer <gerg@uclinux.org> |
| 9804 | W: http://www.uclinux.org/ | 9805 | W: http://www.uclinux.org/ |
| 9805 | L: uclinux-dev@uclinux.org (subscribers-only) | 9806 | L: uclinux-dev@uclinux.org (subscribers-only) |
| 9806 | S: Maintained | 9807 | S: Maintained |
| 9807 | F: arch/m68k/*/*_no.* | 9808 | F: arch/m68k/*/*_no.* |
| 9808 | F: arch/m68k/include/asm/*_no.* | 9809 | F: arch/m68k/include/asm/*_no.* |
| 9809 | 9810 | ||
| 9810 | UDF FILESYSTEM | 9811 | UDF FILESYSTEM |
| 9811 | M: Jan Kara <jack@suse.cz> | 9812 | M: Jan Kara <jack@suse.cz> |
| 9812 | S: Maintained | 9813 | S: Maintained |
| 9813 | F: Documentation/filesystems/udf.txt | 9814 | F: Documentation/filesystems/udf.txt |
| 9814 | F: fs/udf/ | 9815 | F: fs/udf/ |
| 9815 | 9816 | ||
| 9816 | UFS FILESYSTEM | 9817 | UFS FILESYSTEM |
| 9817 | M: Evgeniy Dushistov <dushistov@mail.ru> | 9818 | M: Evgeniy Dushistov <dushistov@mail.ru> |
| 9818 | S: Maintained | 9819 | S: Maintained |
| 9819 | F: Documentation/filesystems/ufs.txt | 9820 | F: Documentation/filesystems/ufs.txt |
| 9820 | F: fs/ufs/ | 9821 | F: fs/ufs/ |
| 9821 | 9822 | ||
| 9822 | UHID USERSPACE HID IO DRIVER: | 9823 | UHID USERSPACE HID IO DRIVER: |
| 9823 | M: David Herrmann <dh.herrmann@googlemail.com> | 9824 | M: David Herrmann <dh.herrmann@googlemail.com> |
| 9824 | L: linux-input@vger.kernel.org | 9825 | L: linux-input@vger.kernel.org |
| 9825 | S: Maintained | 9826 | S: Maintained |
| 9826 | F: drivers/hid/uhid.c | 9827 | F: drivers/hid/uhid.c |
| 9827 | F: include/uapi/linux/uhid.h | 9828 | F: include/uapi/linux/uhid.h |
| 9828 | 9829 | ||
| 9829 | ULTRA-WIDEBAND (UWB) SUBSYSTEM: | 9830 | ULTRA-WIDEBAND (UWB) SUBSYSTEM: |
| 9830 | L: linux-usb@vger.kernel.org | 9831 | L: linux-usb@vger.kernel.org |
| 9831 | S: Orphan | 9832 | S: Orphan |
| 9832 | F: drivers/uwb/ | 9833 | F: drivers/uwb/ |
| 9833 | F: include/linux/uwb.h | 9834 | F: include/linux/uwb.h |
| 9834 | F: include/linux/uwb/ | 9835 | F: include/linux/uwb/ |
| 9835 | 9836 | ||
| 9836 | UNICORE32 ARCHITECTURE: | 9837 | UNICORE32 ARCHITECTURE: |
| 9837 | M: Guan Xuetao <gxt@mprc.pku.edu.cn> | 9838 | M: Guan Xuetao <gxt@mprc.pku.edu.cn> |
| 9838 | W: http://mprc.pku.edu.cn/~guanxuetao/linux | 9839 | W: http://mprc.pku.edu.cn/~guanxuetao/linux |
| 9839 | S: Maintained | 9840 | S: Maintained |
| 9840 | T: git git://github.com/gxt/linux.git | 9841 | T: git git://github.com/gxt/linux.git |
| 9841 | F: arch/unicore32/ | 9842 | F: arch/unicore32/ |
| 9842 | 9843 | ||
| 9843 | UNIFDEF | 9844 | UNIFDEF |
| 9844 | M: Tony Finch <dot@dotat.at> | 9845 | M: Tony Finch <dot@dotat.at> |
| 9845 | W: http://dotat.at/prog/unifdef | 9846 | W: http://dotat.at/prog/unifdef |
| 9846 | S: Maintained | 9847 | S: Maintained |
| 9847 | F: scripts/unifdef.c | 9848 | F: scripts/unifdef.c |
| 9848 | 9849 | ||
| 9849 | UNIFORM CDROM DRIVER | 9850 | UNIFORM CDROM DRIVER |
| 9850 | M: Jens Axboe <axboe@kernel.dk> | 9851 | M: Jens Axboe <axboe@kernel.dk> |
| 9851 | W: http://www.kernel.dk | 9852 | W: http://www.kernel.dk |
| 9852 | S: Maintained | 9853 | S: Maintained |
| 9853 | F: Documentation/cdrom/ | 9854 | F: Documentation/cdrom/ |
| 9854 | F: drivers/cdrom/cdrom.c | 9855 | F: drivers/cdrom/cdrom.c |
| 9855 | F: include/linux/cdrom.h | 9856 | F: include/linux/cdrom.h |
| 9856 | F: include/uapi/linux/cdrom.h | 9857 | F: include/uapi/linux/cdrom.h |
| 9857 | 9858 | ||
| 9858 | UNISYS S-PAR DRIVERS | 9859 | UNISYS S-PAR DRIVERS |
| 9859 | M: Benjamin Romer <benjamin.romer@unisys.com> | 9860 | M: Benjamin Romer <benjamin.romer@unisys.com> |
| 9860 | M: David Kershner <david.kershner@unisys.com> | 9861 | M: David Kershner <david.kershner@unisys.com> |
| 9861 | L: sparmaintainer@unisys.com (Unisys internal) | 9862 | L: sparmaintainer@unisys.com (Unisys internal) |
| 9862 | S: Supported | 9863 | S: Supported |
| 9863 | F: drivers/staging/unisys/ | 9864 | F: drivers/staging/unisys/ |
| 9864 | 9865 | ||
| 9865 | UNIVERSAL FLASH STORAGE HOST CONTROLLER DRIVER | 9866 | UNIVERSAL FLASH STORAGE HOST CONTROLLER DRIVER |
| 9866 | M: Vinayak Holikatti <vinholikatti@gmail.com> | 9867 | M: Vinayak Holikatti <vinholikatti@gmail.com> |
| 9867 | L: linux-scsi@vger.kernel.org | 9868 | L: linux-scsi@vger.kernel.org |
| 9868 | S: Supported | 9869 | S: Supported |
| 9869 | F: Documentation/scsi/ufs.txt | 9870 | F: Documentation/scsi/ufs.txt |
| 9870 | F: drivers/scsi/ufs/ | 9871 | F: drivers/scsi/ufs/ |
| 9871 | 9872 | ||
| 9872 | UNSORTED BLOCK IMAGES (UBI) | 9873 | UNSORTED BLOCK IMAGES (UBI) |
| 9873 | M: Artem Bityutskiy <dedekind1@gmail.com> | 9874 | M: Artem Bityutskiy <dedekind1@gmail.com> |
| 9874 | W: http://www.linux-mtd.infradead.org/ | 9875 | W: http://www.linux-mtd.infradead.org/ |
| 9875 | L: linux-mtd@lists.infradead.org | 9876 | L: linux-mtd@lists.infradead.org |
| 9876 | T: git git://git.infradead.org/ubifs-2.6.git | 9877 | T: git git://git.infradead.org/ubifs-2.6.git |
| 9877 | S: Maintained | 9878 | S: Maintained |
| 9878 | F: drivers/mtd/ubi/ | 9879 | F: drivers/mtd/ubi/ |
| 9879 | F: include/linux/mtd/ubi.h | 9880 | F: include/linux/mtd/ubi.h |
| 9880 | F: include/uapi/mtd/ubi-user.h | 9881 | F: include/uapi/mtd/ubi-user.h |
| 9881 | 9882 | ||
| 9882 | UNSORTED BLOCK IMAGES (UBI) Fastmap | 9883 | UNSORTED BLOCK IMAGES (UBI) Fastmap |
| 9883 | M: Richard Weinberger <richard@nod.at> | 9884 | M: Richard Weinberger <richard@nod.at> |
| 9884 | L: linux-mtd@lists.infradead.org | 9885 | L: linux-mtd@lists.infradead.org |
| 9885 | S: Maintained | 9886 | S: Maintained |
| 9886 | F: drivers/mtd/ubi/fastmap.c | 9887 | F: drivers/mtd/ubi/fastmap.c |
| 9887 | 9888 | ||
| 9888 | USB ACM DRIVER | 9889 | USB ACM DRIVER |
| 9889 | M: Oliver Neukum <oliver@neukum.org> | 9890 | M: Oliver Neukum <oliver@neukum.org> |
| 9890 | L: linux-usb@vger.kernel.org | 9891 | L: linux-usb@vger.kernel.org |
| 9891 | S: Maintained | 9892 | S: Maintained |
| 9892 | F: Documentation/usb/acm.txt | 9893 | F: Documentation/usb/acm.txt |
| 9893 | F: drivers/usb/class/cdc-acm.* | 9894 | F: drivers/usb/class/cdc-acm.* |
| 9894 | 9895 | ||
| 9895 | USB AR5523 WIRELESS DRIVER | 9896 | USB AR5523 WIRELESS DRIVER |
| 9896 | M: Pontus Fuchs <pontus.fuchs@gmail.com> | 9897 | M: Pontus Fuchs <pontus.fuchs@gmail.com> |
| 9897 | L: linux-wireless@vger.kernel.org | 9898 | L: linux-wireless@vger.kernel.org |
| 9898 | S: Maintained | 9899 | S: Maintained |
| 9899 | F: drivers/net/wireless/ath/ar5523/ | 9900 | F: drivers/net/wireless/ath/ar5523/ |
| 9900 | 9901 | ||
| 9901 | USB ATTACHED SCSI | 9902 | USB ATTACHED SCSI |
| 9902 | M: Hans de Goede <hdegoede@redhat.com> | 9903 | M: Hans de Goede <hdegoede@redhat.com> |
| 9903 | M: Gerd Hoffmann <kraxel@redhat.com> | 9904 | M: Gerd Hoffmann <kraxel@redhat.com> |
| 9904 | L: linux-usb@vger.kernel.org | 9905 | L: linux-usb@vger.kernel.org |
| 9905 | L: linux-scsi@vger.kernel.org | 9906 | L: linux-scsi@vger.kernel.org |
| 9906 | S: Maintained | 9907 | S: Maintained |
| 9907 | F: drivers/usb/storage/uas.c | 9908 | F: drivers/usb/storage/uas.c |
| 9908 | 9909 | ||
| 9909 | USB CDC ETHERNET DRIVER | 9910 | USB CDC ETHERNET DRIVER |
| 9910 | M: Oliver Neukum <oliver@neukum.org> | 9911 | M: Oliver Neukum <oliver@neukum.org> |
| 9911 | L: linux-usb@vger.kernel.org | 9912 | L: linux-usb@vger.kernel.org |
| 9912 | S: Maintained | 9913 | S: Maintained |
| 9913 | F: drivers/net/usb/cdc_*.c | 9914 | F: drivers/net/usb/cdc_*.c |
| 9914 | F: include/uapi/linux/usb/cdc.h | 9915 | F: include/uapi/linux/usb/cdc.h |
| 9915 | 9916 | ||
| 9916 | USB CYPRESS C67X00 DRIVER | 9917 | USB CYPRESS C67X00 DRIVER |
| 9917 | M: Peter Korsgaard <jacmet@sunsite.dk> | 9918 | M: Peter Korsgaard <jacmet@sunsite.dk> |
| 9918 | L: linux-usb@vger.kernel.org | 9919 | L: linux-usb@vger.kernel.org |
| 9919 | S: Maintained | 9920 | S: Maintained |
| 9920 | F: drivers/usb/c67x00/ | 9921 | F: drivers/usb/c67x00/ |
| 9921 | 9922 | ||
| 9922 | USB DAVICOM DM9601 DRIVER | 9923 | USB DAVICOM DM9601 DRIVER |
| 9923 | M: Peter Korsgaard <jacmet@sunsite.dk> | 9924 | M: Peter Korsgaard <jacmet@sunsite.dk> |
| 9924 | L: netdev@vger.kernel.org | 9925 | L: netdev@vger.kernel.org |
| 9925 | W: http://www.linux-usb.org/usbnet | 9926 | W: http://www.linux-usb.org/usbnet |
| 9926 | S: Maintained | 9927 | S: Maintained |
| 9927 | F: drivers/net/usb/dm9601.c | 9928 | F: drivers/net/usb/dm9601.c |
| 9928 | 9929 | ||
| 9929 | USB DIAMOND RIO500 DRIVER | 9930 | USB DIAMOND RIO500 DRIVER |
| 9930 | M: Cesar Miquel <miquel@df.uba.ar> | 9931 | M: Cesar Miquel <miquel@df.uba.ar> |
| 9931 | L: rio500-users@lists.sourceforge.net | 9932 | L: rio500-users@lists.sourceforge.net |
| 9932 | W: http://rio500.sourceforge.net | 9933 | W: http://rio500.sourceforge.net |
| 9933 | S: Maintained | 9934 | S: Maintained |
| 9934 | F: drivers/usb/misc/rio500* | 9935 | F: drivers/usb/misc/rio500* |
| 9935 | 9936 | ||
| 9936 | USB EHCI DRIVER | 9937 | USB EHCI DRIVER |
| 9937 | M: Alan Stern <stern@rowland.harvard.edu> | 9938 | M: Alan Stern <stern@rowland.harvard.edu> |
| 9938 | L: linux-usb@vger.kernel.org | 9939 | L: linux-usb@vger.kernel.org |
| 9939 | S: Maintained | 9940 | S: Maintained |
| 9940 | F: Documentation/usb/ehci.txt | 9941 | F: Documentation/usb/ehci.txt |
| 9941 | F: drivers/usb/host/ehci* | 9942 | F: drivers/usb/host/ehci* |
| 9942 | 9943 | ||
| 9943 | USB GADGET/PERIPHERAL SUBSYSTEM | 9944 | USB GADGET/PERIPHERAL SUBSYSTEM |
| 9944 | M: Felipe Balbi <balbi@ti.com> | 9945 | M: Felipe Balbi <balbi@ti.com> |
| 9945 | L: linux-usb@vger.kernel.org | 9946 | L: linux-usb@vger.kernel.org |
| 9946 | W: http://www.linux-usb.org/gadget | 9947 | W: http://www.linux-usb.org/gadget |
| 9947 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git | 9948 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git |
| 9948 | S: Maintained | 9949 | S: Maintained |
| 9949 | F: drivers/usb/gadget/ | 9950 | F: drivers/usb/gadget/ |
| 9950 | F: include/linux/usb/gadget* | 9951 | F: include/linux/usb/gadget* |
| 9951 | 9952 | ||
| 9952 | USB HID/HIDBP DRIVERS (USB KEYBOARDS, MICE, REMOTE CONTROLS, ...) | 9953 | USB HID/HIDBP DRIVERS (USB KEYBOARDS, MICE, REMOTE CONTROLS, ...) |
| 9953 | M: Jiri Kosina <jkosina@suse.cz> | 9954 | M: Jiri Kosina <jkosina@suse.cz> |
| 9954 | L: linux-usb@vger.kernel.org | 9955 | L: linux-usb@vger.kernel.org |
| 9955 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git | 9956 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid.git |
| 9956 | S: Maintained | 9957 | S: Maintained |
| 9957 | F: Documentation/hid/hiddev.txt | 9958 | F: Documentation/hid/hiddev.txt |
| 9958 | F: drivers/hid/usbhid/ | 9959 | F: drivers/hid/usbhid/ |
| 9959 | 9960 | ||
| 9960 | USB ISP116X DRIVER | 9961 | USB ISP116X DRIVER |
| 9961 | M: Olav Kongas <ok@artecdesign.ee> | 9962 | M: Olav Kongas <ok@artecdesign.ee> |
| 9962 | L: linux-usb@vger.kernel.org | 9963 | L: linux-usb@vger.kernel.org |
| 9963 | S: Maintained | 9964 | S: Maintained |
| 9964 | F: drivers/usb/host/isp116x* | 9965 | F: drivers/usb/host/isp116x* |
| 9965 | F: include/linux/usb/isp116x.h | 9966 | F: include/linux/usb/isp116x.h |
| 9966 | 9967 | ||
| 9967 | USB MASS STORAGE DRIVER | 9968 | USB MASS STORAGE DRIVER |
| 9968 | M: Matthew Dharm <mdharm-usb@one-eyed-alien.net> | 9969 | M: Matthew Dharm <mdharm-usb@one-eyed-alien.net> |
| 9969 | L: linux-usb@vger.kernel.org | 9970 | L: linux-usb@vger.kernel.org |
| 9970 | L: usb-storage@lists.one-eyed-alien.net | 9971 | L: usb-storage@lists.one-eyed-alien.net |
| 9971 | S: Maintained | 9972 | S: Maintained |
| 9972 | W: http://www.one-eyed-alien.net/~mdharm/linux-usb/ | 9973 | W: http://www.one-eyed-alien.net/~mdharm/linux-usb/ |
| 9973 | F: drivers/usb/storage/ | 9974 | F: drivers/usb/storage/ |
| 9974 | 9975 | ||
| 9975 | USB MIDI DRIVER | 9976 | USB MIDI DRIVER |
| 9976 | M: Clemens Ladisch <clemens@ladisch.de> | 9977 | M: Clemens Ladisch <clemens@ladisch.de> |
| 9977 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) | 9978 | L: alsa-devel@alsa-project.org (moderated for non-subscribers) |
| 9978 | T: git git://git.alsa-project.org/alsa-kernel.git | 9979 | T: git git://git.alsa-project.org/alsa-kernel.git |
| 9979 | S: Maintained | 9980 | S: Maintained |
| 9980 | F: sound/usb/midi.* | 9981 | F: sound/usb/midi.* |
| 9981 | 9982 | ||
| 9982 | USB NETWORKING DRIVERS | 9983 | USB NETWORKING DRIVERS |
| 9983 | L: linux-usb@vger.kernel.org | 9984 | L: linux-usb@vger.kernel.org |
| 9984 | S: Odd Fixes | 9985 | S: Odd Fixes |
| 9985 | F: drivers/net/usb/ | 9986 | F: drivers/net/usb/ |
| 9986 | 9987 | ||
| 9987 | USB OHCI DRIVER | 9988 | USB OHCI DRIVER |
| 9988 | M: Alan Stern <stern@rowland.harvard.edu> | 9989 | M: Alan Stern <stern@rowland.harvard.edu> |
| 9989 | L: linux-usb@vger.kernel.org | 9990 | L: linux-usb@vger.kernel.org |
| 9990 | S: Maintained | 9991 | S: Maintained |
| 9991 | F: Documentation/usb/ohci.txt | 9992 | F: Documentation/usb/ohci.txt |
| 9992 | F: drivers/usb/host/ohci* | 9993 | F: drivers/usb/host/ohci* |
| 9993 | 9994 | ||
| 9994 | USB OVER IP DRIVER | 9995 | USB OVER IP DRIVER |
| 9995 | M: Valentina Manea <valentina.manea.m@gmail.com> | 9996 | M: Valentina Manea <valentina.manea.m@gmail.com> |
| 9996 | M: Shuah Khan <shuah.kh@samsung.com> | 9997 | M: Shuah Khan <shuah.kh@samsung.com> |
| 9997 | L: linux-usb@vger.kernel.org | 9998 | L: linux-usb@vger.kernel.org |
| 9998 | S: Maintained | 9999 | S: Maintained |
| 9999 | F: drivers/usb/usbip/ | 10000 | F: drivers/usb/usbip/ |
| 10000 | F: tools/usb/usbip/ | 10001 | F: tools/usb/usbip/ |
| 10001 | 10002 | ||
| 10002 | USB PEGASUS DRIVER | 10003 | USB PEGASUS DRIVER |
| 10003 | M: Petko Manolov <petkan@nucleusys.com> | 10004 | M: Petko Manolov <petkan@nucleusys.com> |
| 10004 | L: linux-usb@vger.kernel.org | 10005 | L: linux-usb@vger.kernel.org |
| 10005 | L: netdev@vger.kernel.org | 10006 | L: netdev@vger.kernel.org |
| 10006 | T: git git://github.com/petkan/pegasus.git | 10007 | T: git git://github.com/petkan/pegasus.git |
| 10007 | W: https://github.com/petkan/pegasus | 10008 | W: https://github.com/petkan/pegasus |
| 10008 | S: Maintained | 10009 | S: Maintained |
| 10009 | F: drivers/net/usb/pegasus.* | 10010 | F: drivers/net/usb/pegasus.* |
| 10010 | 10011 | ||
| 10011 | USB PHY LAYER | 10012 | USB PHY LAYER |
| 10012 | M: Felipe Balbi <balbi@ti.com> | 10013 | M: Felipe Balbi <balbi@ti.com> |
| 10013 | L: linux-usb@vger.kernel.org | 10014 | L: linux-usb@vger.kernel.org |
| 10014 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git | 10015 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git |
| 10015 | S: Maintained | 10016 | S: Maintained |
| 10016 | F: drivers/usb/phy/ | 10017 | F: drivers/usb/phy/ |
| 10017 | 10018 | ||
| 10018 | USB PRINTER DRIVER (usblp) | 10019 | USB PRINTER DRIVER (usblp) |
| 10019 | M: Pete Zaitcev <zaitcev@redhat.com> | 10020 | M: Pete Zaitcev <zaitcev@redhat.com> |
| 10020 | L: linux-usb@vger.kernel.org | 10021 | L: linux-usb@vger.kernel.org |
| 10021 | S: Supported | 10022 | S: Supported |
| 10022 | F: drivers/usb/class/usblp.c | 10023 | F: drivers/usb/class/usblp.c |
| 10023 | 10024 | ||
| 10024 | USB RTL8150 DRIVER | 10025 | USB RTL8150 DRIVER |
| 10025 | M: Petko Manolov <petkan@nucleusys.com> | 10026 | M: Petko Manolov <petkan@nucleusys.com> |
| 10026 | L: linux-usb@vger.kernel.org | 10027 | L: linux-usb@vger.kernel.org |
| 10027 | L: netdev@vger.kernel.org | 10028 | L: netdev@vger.kernel.org |
| 10028 | T: git git://github.com/petkan/rtl8150.git | 10029 | T: git git://github.com/petkan/rtl8150.git |
| 10029 | W: https://github.com/petkan/rtl8150 | 10030 | W: https://github.com/petkan/rtl8150 |
| 10030 | S: Maintained | 10031 | S: Maintained |
| 10031 | F: drivers/net/usb/rtl8150.c | 10032 | F: drivers/net/usb/rtl8150.c |
| 10032 | 10033 | ||
| 10033 | USB SERIAL SUBSYSTEM | 10034 | USB SERIAL SUBSYSTEM |
| 10034 | M: Johan Hovold <johan@kernel.org> | 10035 | M: Johan Hovold <johan@kernel.org> |
| 10035 | L: linux-usb@vger.kernel.org | 10036 | L: linux-usb@vger.kernel.org |
| 10036 | S: Maintained | 10037 | S: Maintained |
| 10037 | F: Documentation/usb/usb-serial.txt | 10038 | F: Documentation/usb/usb-serial.txt |
| 10038 | F: drivers/usb/serial/ | 10039 | F: drivers/usb/serial/ |
| 10039 | F: include/linux/usb/serial.h | 10040 | F: include/linux/usb/serial.h |
| 10040 | 10041 | ||
| 10041 | USB SMSC75XX ETHERNET DRIVER | 10042 | USB SMSC75XX ETHERNET DRIVER |
| 10042 | M: Steve Glendinning <steve.glendinning@shawell.net> | 10043 | M: Steve Glendinning <steve.glendinning@shawell.net> |
| 10043 | L: netdev@vger.kernel.org | 10044 | L: netdev@vger.kernel.org |
| 10044 | S: Maintained | 10045 | S: Maintained |
| 10045 | F: drivers/net/usb/smsc75xx.* | 10046 | F: drivers/net/usb/smsc75xx.* |
| 10046 | 10047 | ||
| 10047 | USB SMSC95XX ETHERNET DRIVER | 10048 | USB SMSC95XX ETHERNET DRIVER |
| 10048 | M: Steve Glendinning <steve.glendinning@shawell.net> | 10049 | M: Steve Glendinning <steve.glendinning@shawell.net> |
| 10049 | L: netdev@vger.kernel.org | 10050 | L: netdev@vger.kernel.org |
| 10050 | S: Maintained | 10051 | S: Maintained |
| 10051 | F: drivers/net/usb/smsc95xx.* | 10052 | F: drivers/net/usb/smsc95xx.* |
| 10052 | 10053 | ||
| 10053 | USB SUBSYSTEM | 10054 | USB SUBSYSTEM |
| 10054 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 10055 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 10055 | L: linux-usb@vger.kernel.org | 10056 | L: linux-usb@vger.kernel.org |
| 10056 | W: http://www.linux-usb.org | 10057 | W: http://www.linux-usb.org |
| 10057 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb.git | 10058 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb.git |
| 10058 | S: Supported | 10059 | S: Supported |
| 10059 | F: Documentation/usb/ | 10060 | F: Documentation/usb/ |
| 10060 | F: drivers/usb/ | 10061 | F: drivers/usb/ |
| 10061 | F: include/linux/usb.h | 10062 | F: include/linux/usb.h |
| 10062 | F: include/linux/usb/ | 10063 | F: include/linux/usb/ |
| 10063 | 10064 | ||
| 10064 | USB UHCI DRIVER | 10065 | USB UHCI DRIVER |
| 10065 | M: Alan Stern <stern@rowland.harvard.edu> | 10066 | M: Alan Stern <stern@rowland.harvard.edu> |
| 10066 | L: linux-usb@vger.kernel.org | 10067 | L: linux-usb@vger.kernel.org |
| 10067 | S: Maintained | 10068 | S: Maintained |
| 10068 | F: drivers/usb/host/uhci* | 10069 | F: drivers/usb/host/uhci* |
| 10069 | 10070 | ||
| 10070 | USB "USBNET" DRIVER FRAMEWORK | 10071 | USB "USBNET" DRIVER FRAMEWORK |
| 10071 | M: Oliver Neukum <oneukum@suse.de> | 10072 | M: Oliver Neukum <oneukum@suse.de> |
| 10072 | L: netdev@vger.kernel.org | 10073 | L: netdev@vger.kernel.org |
| 10073 | W: http://www.linux-usb.org/usbnet | 10074 | W: http://www.linux-usb.org/usbnet |
| 10074 | S: Maintained | 10075 | S: Maintained |
| 10075 | F: drivers/net/usb/usbnet.c | 10076 | F: drivers/net/usb/usbnet.c |
| 10076 | F: include/linux/usb/usbnet.h | 10077 | F: include/linux/usb/usbnet.h |
| 10077 | 10078 | ||
| 10078 | USB VIDEO CLASS | 10079 | USB VIDEO CLASS |
| 10079 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 10080 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 10080 | L: linux-uvc-devel@lists.sourceforge.net (subscribers-only) | 10081 | L: linux-uvc-devel@lists.sourceforge.net (subscribers-only) |
| 10081 | L: linux-media@vger.kernel.org | 10082 | L: linux-media@vger.kernel.org |
| 10082 | T: git git://linuxtv.org/media_tree.git | 10083 | T: git git://linuxtv.org/media_tree.git |
| 10083 | W: http://www.ideasonboard.org/uvc/ | 10084 | W: http://www.ideasonboard.org/uvc/ |
| 10084 | S: Maintained | 10085 | S: Maintained |
| 10085 | F: drivers/media/usb/uvc/ | 10086 | F: drivers/media/usb/uvc/ |
| 10086 | F: include/uapi/linux/uvcvideo.h | 10087 | F: include/uapi/linux/uvcvideo.h |
| 10087 | 10088 | ||
| 10088 | USB VISION DRIVER | 10089 | USB VISION DRIVER |
| 10089 | M: Hans Verkuil <hverkuil@xs4all.nl> | 10090 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 10090 | L: linux-media@vger.kernel.org | 10091 | L: linux-media@vger.kernel.org |
| 10091 | T: git git://linuxtv.org/media_tree.git | 10092 | T: git git://linuxtv.org/media_tree.git |
| 10092 | W: http://linuxtv.org | 10093 | W: http://linuxtv.org |
| 10093 | S: Odd Fixes | 10094 | S: Odd Fixes |
| 10094 | F: drivers/media/usb/usbvision/ | 10095 | F: drivers/media/usb/usbvision/ |
| 10095 | 10096 | ||
| 10096 | USB WEBCAM GADGET | 10097 | USB WEBCAM GADGET |
| 10097 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> | 10098 | M: Laurent Pinchart <laurent.pinchart@ideasonboard.com> |
| 10098 | L: linux-usb@vger.kernel.org | 10099 | L: linux-usb@vger.kernel.org |
| 10099 | S: Maintained | 10100 | S: Maintained |
| 10100 | F: drivers/usb/gadget/function/*uvc* | 10101 | F: drivers/usb/gadget/function/*uvc* |
| 10101 | F: drivers/usb/gadget/legacy/webcam.c | 10102 | F: drivers/usb/gadget/legacy/webcam.c |
| 10102 | 10103 | ||
| 10103 | USB WIRELESS RNDIS DRIVER (rndis_wlan) | 10104 | USB WIRELESS RNDIS DRIVER (rndis_wlan) |
| 10104 | M: Jussi Kivilinna <jussi.kivilinna@iki.fi> | 10105 | M: Jussi Kivilinna <jussi.kivilinna@iki.fi> |
| 10105 | L: linux-wireless@vger.kernel.org | 10106 | L: linux-wireless@vger.kernel.org |
| 10106 | S: Maintained | 10107 | S: Maintained |
| 10107 | F: drivers/net/wireless/rndis_wlan.c | 10108 | F: drivers/net/wireless/rndis_wlan.c |
| 10108 | 10109 | ||
| 10109 | USB XHCI DRIVER | 10110 | USB XHCI DRIVER |
| 10110 | M: Mathias Nyman <mathias.nyman@intel.com> | 10111 | M: Mathias Nyman <mathias.nyman@intel.com> |
| 10111 | L: linux-usb@vger.kernel.org | 10112 | L: linux-usb@vger.kernel.org |
| 10112 | S: Supported | 10113 | S: Supported |
| 10113 | F: drivers/usb/host/xhci* | 10114 | F: drivers/usb/host/xhci* |
| 10114 | F: drivers/usb/host/pci-quirks* | 10115 | F: drivers/usb/host/pci-quirks* |
| 10115 | 10116 | ||
| 10116 | USB ZD1201 DRIVER | 10117 | USB ZD1201 DRIVER |
| 10117 | L: linux-wireless@vger.kernel.org | 10118 | L: linux-wireless@vger.kernel.org |
| 10118 | W: http://linux-lc100020.sourceforge.net | 10119 | W: http://linux-lc100020.sourceforge.net |
| 10119 | S: Orphan | 10120 | S: Orphan |
| 10120 | F: drivers/net/wireless/zd1201.* | 10121 | F: drivers/net/wireless/zd1201.* |
| 10121 | 10122 | ||
| 10122 | USB ZR364XX DRIVER | 10123 | USB ZR364XX DRIVER |
| 10123 | M: Antoine Jacquet <royale@zerezo.com> | 10124 | M: Antoine Jacquet <royale@zerezo.com> |
| 10124 | L: linux-usb@vger.kernel.org | 10125 | L: linux-usb@vger.kernel.org |
| 10125 | L: linux-media@vger.kernel.org | 10126 | L: linux-media@vger.kernel.org |
| 10126 | T: git git://linuxtv.org/media_tree.git | 10127 | T: git git://linuxtv.org/media_tree.git |
| 10127 | W: http://royale.zerezo.com/zr364xx/ | 10128 | W: http://royale.zerezo.com/zr364xx/ |
| 10128 | S: Maintained | 10129 | S: Maintained |
| 10129 | F: Documentation/video4linux/zr364xx.txt | 10130 | F: Documentation/video4linux/zr364xx.txt |
| 10130 | F: drivers/media/usb/zr364xx/ | 10131 | F: drivers/media/usb/zr364xx/ |
| 10131 | 10132 | ||
| 10132 | USER-MODE LINUX (UML) | 10133 | USER-MODE LINUX (UML) |
| 10133 | M: Jeff Dike <jdike@addtoit.com> | 10134 | M: Jeff Dike <jdike@addtoit.com> |
| 10134 | M: Richard Weinberger <richard@nod.at> | 10135 | M: Richard Weinberger <richard@nod.at> |
| 10135 | L: user-mode-linux-devel@lists.sourceforge.net | 10136 | L: user-mode-linux-devel@lists.sourceforge.net |
| 10136 | L: user-mode-linux-user@lists.sourceforge.net | 10137 | L: user-mode-linux-user@lists.sourceforge.net |
| 10137 | W: http://user-mode-linux.sourceforge.net | 10138 | W: http://user-mode-linux.sourceforge.net |
| 10138 | S: Maintained | 10139 | S: Maintained |
| 10139 | F: Documentation/virtual/uml/ | 10140 | F: Documentation/virtual/uml/ |
| 10140 | F: arch/um/ | 10141 | F: arch/um/ |
| 10141 | F: arch/x86/um/ | 10142 | F: arch/x86/um/ |
| 10142 | F: fs/hostfs/ | 10143 | F: fs/hostfs/ |
| 10143 | F: fs/hppfs/ | 10144 | F: fs/hppfs/ |
| 10144 | 10145 | ||
| 10145 | USERSPACE I/O (UIO) | 10146 | USERSPACE I/O (UIO) |
| 10146 | M: "Hans J. Koch" <hjk@hansjkoch.de> | 10147 | M: "Hans J. Koch" <hjk@hansjkoch.de> |
| 10147 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 10148 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 10148 | S: Maintained | 10149 | S: Maintained |
| 10149 | F: Documentation/DocBook/uio-howto.tmpl | 10150 | F: Documentation/DocBook/uio-howto.tmpl |
| 10150 | F: drivers/uio/ | 10151 | F: drivers/uio/ |
| 10151 | F: include/linux/uio*.h | 10152 | F: include/linux/uio*.h |
| 10152 | 10153 | ||
| 10153 | UTIL-LINUX PACKAGE | 10154 | UTIL-LINUX PACKAGE |
| 10154 | M: Karel Zak <kzak@redhat.com> | 10155 | M: Karel Zak <kzak@redhat.com> |
| 10155 | L: util-linux@vger.kernel.org | 10156 | L: util-linux@vger.kernel.org |
| 10156 | W: http://en.wikipedia.org/wiki/Util-linux | 10157 | W: http://en.wikipedia.org/wiki/Util-linux |
| 10157 | T: git git://git.kernel.org/pub/scm/utils/util-linux/util-linux.git | 10158 | T: git git://git.kernel.org/pub/scm/utils/util-linux/util-linux.git |
| 10158 | S: Maintained | 10159 | S: Maintained |
| 10159 | 10160 | ||
| 10160 | UVESAFB DRIVER | 10161 | UVESAFB DRIVER |
| 10161 | M: Michal Januszewski <spock@gentoo.org> | 10162 | M: Michal Januszewski <spock@gentoo.org> |
| 10162 | L: linux-fbdev@vger.kernel.org | 10163 | L: linux-fbdev@vger.kernel.org |
| 10163 | W: http://dev.gentoo.org/~spock/projects/uvesafb/ | 10164 | W: http://dev.gentoo.org/~spock/projects/uvesafb/ |
| 10164 | S: Maintained | 10165 | S: Maintained |
| 10165 | F: Documentation/fb/uvesafb.txt | 10166 | F: Documentation/fb/uvesafb.txt |
| 10166 | F: drivers/video/fbdev/uvesafb.* | 10167 | F: drivers/video/fbdev/uvesafb.* |
| 10167 | 10168 | ||
| 10168 | VFAT/FAT/MSDOS FILESYSTEM | 10169 | VFAT/FAT/MSDOS FILESYSTEM |
| 10169 | M: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> | 10170 | M: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp> |
| 10170 | S: Maintained | 10171 | S: Maintained |
| 10171 | F: Documentation/filesystems/vfat.txt | 10172 | F: Documentation/filesystems/vfat.txt |
| 10172 | F: fs/fat/ | 10173 | F: fs/fat/ |
| 10173 | 10174 | ||
| 10174 | VFIO DRIVER | 10175 | VFIO DRIVER |
| 10175 | M: Alex Williamson <alex.williamson@redhat.com> | 10176 | M: Alex Williamson <alex.williamson@redhat.com> |
| 10176 | L: kvm@vger.kernel.org | 10177 | L: kvm@vger.kernel.org |
| 10177 | S: Maintained | 10178 | S: Maintained |
| 10178 | F: Documentation/vfio.txt | 10179 | F: Documentation/vfio.txt |
| 10179 | F: drivers/vfio/ | 10180 | F: drivers/vfio/ |
| 10180 | F: include/linux/vfio.h | 10181 | F: include/linux/vfio.h |
| 10181 | F: include/uapi/linux/vfio.h | 10182 | F: include/uapi/linux/vfio.h |
| 10182 | 10183 | ||
| 10183 | VIDEOBUF2 FRAMEWORK | 10184 | VIDEOBUF2 FRAMEWORK |
| 10184 | M: Pawel Osciak <pawel@osciak.com> | 10185 | M: Pawel Osciak <pawel@osciak.com> |
| 10185 | M: Marek Szyprowski <m.szyprowski@samsung.com> | 10186 | M: Marek Szyprowski <m.szyprowski@samsung.com> |
| 10186 | M: Kyungmin Park <kyungmin.park@samsung.com> | 10187 | M: Kyungmin Park <kyungmin.park@samsung.com> |
| 10187 | L: linux-media@vger.kernel.org | 10188 | L: linux-media@vger.kernel.org |
| 10188 | S: Maintained | 10189 | S: Maintained |
| 10189 | F: drivers/media/v4l2-core/videobuf2-* | 10190 | F: drivers/media/v4l2-core/videobuf2-* |
| 10190 | F: include/media/videobuf2-* | 10191 | F: include/media/videobuf2-* |
| 10191 | 10192 | ||
| 10192 | VIRTIO CONSOLE DRIVER | 10193 | VIRTIO CONSOLE DRIVER |
| 10193 | M: Amit Shah <amit.shah@redhat.com> | 10194 | M: Amit Shah <amit.shah@redhat.com> |
| 10194 | L: virtualization@lists.linux-foundation.org | 10195 | L: virtualization@lists.linux-foundation.org |
| 10195 | S: Maintained | 10196 | S: Maintained |
| 10196 | F: drivers/char/virtio_console.c | 10197 | F: drivers/char/virtio_console.c |
| 10197 | F: include/linux/virtio_console.h | 10198 | F: include/linux/virtio_console.h |
| 10198 | F: include/uapi/linux/virtio_console.h | 10199 | F: include/uapi/linux/virtio_console.h |
| 10199 | 10200 | ||
| 10200 | VIRTIO CORE, NET AND BLOCK DRIVERS | 10201 | VIRTIO CORE, NET AND BLOCK DRIVERS |
| 10201 | M: Rusty Russell <rusty@rustcorp.com.au> | 10202 | M: Rusty Russell <rusty@rustcorp.com.au> |
| 10202 | M: "Michael S. Tsirkin" <mst@redhat.com> | 10203 | M: "Michael S. Tsirkin" <mst@redhat.com> |
| 10203 | L: virtualization@lists.linux-foundation.org | 10204 | L: virtualization@lists.linux-foundation.org |
| 10204 | S: Maintained | 10205 | S: Maintained |
| 10205 | F: drivers/virtio/ | 10206 | F: drivers/virtio/ |
| 10206 | F: tools/virtio/ | 10207 | F: tools/virtio/ |
| 10207 | F: drivers/net/virtio_net.c | 10208 | F: drivers/net/virtio_net.c |
| 10208 | F: drivers/block/virtio_blk.c | 10209 | F: drivers/block/virtio_blk.c |
| 10209 | F: include/linux/virtio_*.h | 10210 | F: include/linux/virtio_*.h |
| 10210 | F: include/uapi/linux/virtio_*.h | 10211 | F: include/uapi/linux/virtio_*.h |
| 10211 | 10212 | ||
| 10212 | VIRTIO HOST (VHOST) | 10213 | VIRTIO HOST (VHOST) |
| 10213 | M: "Michael S. Tsirkin" <mst@redhat.com> | 10214 | M: "Michael S. Tsirkin" <mst@redhat.com> |
| 10214 | L: kvm@vger.kernel.org | 10215 | L: kvm@vger.kernel.org |
| 10215 | L: virtualization@lists.linux-foundation.org | 10216 | L: virtualization@lists.linux-foundation.org |
| 10216 | L: netdev@vger.kernel.org | 10217 | L: netdev@vger.kernel.org |
| 10217 | S: Maintained | 10218 | S: Maintained |
| 10218 | F: drivers/vhost/ | 10219 | F: drivers/vhost/ |
| 10219 | F: include/uapi/linux/vhost.h | 10220 | F: include/uapi/linux/vhost.h |
| 10220 | 10221 | ||
| 10221 | VIA RHINE NETWORK DRIVER | 10222 | VIA RHINE NETWORK DRIVER |
| 10222 | M: Roger Luethi <rl@hellgate.ch> | 10223 | M: Roger Luethi <rl@hellgate.ch> |
| 10223 | S: Maintained | 10224 | S: Maintained |
| 10224 | F: drivers/net/ethernet/via/via-rhine.c | 10225 | F: drivers/net/ethernet/via/via-rhine.c |
| 10225 | 10226 | ||
| 10226 | VIA SD/MMC CARD CONTROLLER DRIVER | 10227 | VIA SD/MMC CARD CONTROLLER DRIVER |
| 10227 | M: Bruce Chang <brucechang@via.com.tw> | 10228 | M: Bruce Chang <brucechang@via.com.tw> |
| 10228 | M: Harald Welte <HaraldWelte@viatech.com> | 10229 | M: Harald Welte <HaraldWelte@viatech.com> |
| 10229 | S: Maintained | 10230 | S: Maintained |
| 10230 | F: drivers/mmc/host/via-sdmmc.c | 10231 | F: drivers/mmc/host/via-sdmmc.c |
| 10231 | 10232 | ||
| 10232 | VIA UNICHROME(PRO)/CHROME9 FRAMEBUFFER DRIVER | 10233 | VIA UNICHROME(PRO)/CHROME9 FRAMEBUFFER DRIVER |
| 10233 | M: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> | 10234 | M: Florian Tobias Schandinat <FlorianSchandinat@gmx.de> |
| 10234 | L: linux-fbdev@vger.kernel.org | 10235 | L: linux-fbdev@vger.kernel.org |
| 10235 | S: Maintained | 10236 | S: Maintained |
| 10236 | F: include/linux/via-core.h | 10237 | F: include/linux/via-core.h |
| 10237 | F: include/linux/via-gpio.h | 10238 | F: include/linux/via-gpio.h |
| 10238 | F: include/linux/via_i2c.h | 10239 | F: include/linux/via_i2c.h |
| 10239 | F: drivers/video/fbdev/via/ | 10240 | F: drivers/video/fbdev/via/ |
| 10240 | 10241 | ||
| 10241 | VIA VELOCITY NETWORK DRIVER | 10242 | VIA VELOCITY NETWORK DRIVER |
| 10242 | M: Francois Romieu <romieu@fr.zoreil.com> | 10243 | M: Francois Romieu <romieu@fr.zoreil.com> |
| 10243 | L: netdev@vger.kernel.org | 10244 | L: netdev@vger.kernel.org |
| 10244 | S: Maintained | 10245 | S: Maintained |
| 10245 | F: drivers/net/ethernet/via/via-velocity.* | 10246 | F: drivers/net/ethernet/via/via-velocity.* |
| 10246 | 10247 | ||
| 10247 | VIVID VIRTUAL VIDEO DRIVER | 10248 | VIVID VIRTUAL VIDEO DRIVER |
| 10248 | M: Hans Verkuil <hverkuil@xs4all.nl> | 10249 | M: Hans Verkuil <hverkuil@xs4all.nl> |
| 10249 | L: linux-media@vger.kernel.org | 10250 | L: linux-media@vger.kernel.org |
| 10250 | T: git git://linuxtv.org/media_tree.git | 10251 | T: git git://linuxtv.org/media_tree.git |
| 10251 | W: http://linuxtv.org | 10252 | W: http://linuxtv.org |
| 10252 | S: Maintained | 10253 | S: Maintained |
| 10253 | F: drivers/media/platform/vivid/* | 10254 | F: drivers/media/platform/vivid/* |
| 10254 | 10255 | ||
| 10255 | VLAN (802.1Q) | 10256 | VLAN (802.1Q) |
| 10256 | M: Patrick McHardy <kaber@trash.net> | 10257 | M: Patrick McHardy <kaber@trash.net> |
| 10257 | L: netdev@vger.kernel.org | 10258 | L: netdev@vger.kernel.org |
| 10258 | S: Maintained | 10259 | S: Maintained |
| 10259 | F: drivers/net/macvlan.c | 10260 | F: drivers/net/macvlan.c |
| 10260 | F: include/linux/if_*vlan.h | 10261 | F: include/linux/if_*vlan.h |
| 10261 | F: net/8021q/ | 10262 | F: net/8021q/ |
| 10262 | 10263 | ||
| 10263 | VLYNQ BUS | 10264 | VLYNQ BUS |
| 10264 | M: Florian Fainelli <florian@openwrt.org> | 10265 | M: Florian Fainelli <florian@openwrt.org> |
| 10265 | L: openwrt-devel@lists.openwrt.org (subscribers-only) | 10266 | L: openwrt-devel@lists.openwrt.org (subscribers-only) |
| 10266 | S: Maintained | 10267 | S: Maintained |
| 10267 | F: drivers/vlynq/vlynq.c | 10268 | F: drivers/vlynq/vlynq.c |
| 10268 | F: include/linux/vlynq.h | 10269 | F: include/linux/vlynq.h |
| 10269 | 10270 | ||
| 10270 | VME SUBSYSTEM | 10271 | VME SUBSYSTEM |
| 10271 | M: Martyn Welch <martyn.welch@ge.com> | 10272 | M: Martyn Welch <martyn.welch@ge.com> |
| 10272 | M: Manohar Vanga <manohar.vanga@gmail.com> | 10273 | M: Manohar Vanga <manohar.vanga@gmail.com> |
| 10273 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 10274 | M: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
| 10274 | L: devel@driverdev.osuosl.org | 10275 | L: devel@driverdev.osuosl.org |
| 10275 | S: Maintained | 10276 | S: Maintained |
| 10276 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git | 10277 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git |
| 10277 | F: Documentation/vme_api.txt | 10278 | F: Documentation/vme_api.txt |
| 10278 | F: drivers/staging/vme/ | 10279 | F: drivers/staging/vme/ |
| 10279 | F: drivers/vme/ | 10280 | F: drivers/vme/ |
| 10280 | F: include/linux/vme* | 10281 | F: include/linux/vme* |
| 10281 | 10282 | ||
| 10282 | VMWARE HYPERVISOR INTERFACE | 10283 | VMWARE HYPERVISOR INTERFACE |
| 10283 | M: Alok Kataria <akataria@vmware.com> | 10284 | M: Alok Kataria <akataria@vmware.com> |
| 10284 | L: virtualization@lists.linux-foundation.org | 10285 | L: virtualization@lists.linux-foundation.org |
| 10285 | S: Supported | 10286 | S: Supported |
| 10286 | F: arch/x86/kernel/cpu/vmware.c | 10287 | F: arch/x86/kernel/cpu/vmware.c |
| 10287 | 10288 | ||
| 10288 | VMWARE BALLOON DRIVER | 10289 | VMWARE BALLOON DRIVER |
| 10289 | M: Xavier Deguillard <xdeguillard@vmware.com> | 10290 | M: Xavier Deguillard <xdeguillard@vmware.com> |
| 10290 | M: Philip Moltmann <moltmann@vmware.com> | 10291 | M: Philip Moltmann <moltmann@vmware.com> |
| 10291 | M: "VMware, Inc." <pv-drivers@vmware.com> | 10292 | M: "VMware, Inc." <pv-drivers@vmware.com> |
| 10292 | L: linux-kernel@vger.kernel.org | 10293 | L: linux-kernel@vger.kernel.org |
| 10293 | S: Maintained | 10294 | S: Maintained |
| 10294 | F: drivers/misc/vmw_balloon.c | 10295 | F: drivers/misc/vmw_balloon.c |
| 10295 | 10296 | ||
| 10296 | VMWARE VMXNET3 ETHERNET DRIVER | 10297 | VMWARE VMXNET3 ETHERNET DRIVER |
| 10297 | M: Shreyas Bhatewara <sbhatewara@vmware.com> | 10298 | M: Shreyas Bhatewara <sbhatewara@vmware.com> |
| 10298 | M: "VMware, Inc." <pv-drivers@vmware.com> | 10299 | M: "VMware, Inc." <pv-drivers@vmware.com> |
| 10299 | L: netdev@vger.kernel.org | 10300 | L: netdev@vger.kernel.org |
| 10300 | S: Maintained | 10301 | S: Maintained |
| 10301 | F: drivers/net/vmxnet3/ | 10302 | F: drivers/net/vmxnet3/ |
| 10302 | 10303 | ||
| 10303 | VMware PVSCSI driver | 10304 | VMware PVSCSI driver |
| 10304 | M: Arvind Kumar <arvindkumar@vmware.com> | 10305 | M: Arvind Kumar <arvindkumar@vmware.com> |
| 10305 | M: VMware PV-Drivers <pv-drivers@vmware.com> | 10306 | M: VMware PV-Drivers <pv-drivers@vmware.com> |
| 10306 | L: linux-scsi@vger.kernel.org | 10307 | L: linux-scsi@vger.kernel.org |
| 10307 | S: Maintained | 10308 | S: Maintained |
| 10308 | F: drivers/scsi/vmw_pvscsi.c | 10309 | F: drivers/scsi/vmw_pvscsi.c |
| 10309 | F: drivers/scsi/vmw_pvscsi.h | 10310 | F: drivers/scsi/vmw_pvscsi.h |
| 10310 | 10311 | ||
| 10311 | VOLTAGE AND CURRENT REGULATOR FRAMEWORK | 10312 | VOLTAGE AND CURRENT REGULATOR FRAMEWORK |
| 10312 | M: Liam Girdwood <lgirdwood@gmail.com> | 10313 | M: Liam Girdwood <lgirdwood@gmail.com> |
| 10313 | M: Mark Brown <broonie@kernel.org> | 10314 | M: Mark Brown <broonie@kernel.org> |
| 10314 | L: linux-kernel@vger.kernel.org | 10315 | L: linux-kernel@vger.kernel.org |
| 10315 | W: http://opensource.wolfsonmicro.com/node/15 | 10316 | W: http://opensource.wolfsonmicro.com/node/15 |
| 10316 | W: http://www.slimlogic.co.uk/?p=48 | 10317 | W: http://www.slimlogic.co.uk/?p=48 |
| 10317 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator.git | 10318 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator.git |
| 10318 | S: Supported | 10319 | S: Supported |
| 10319 | F: drivers/regulator/ | 10320 | F: drivers/regulator/ |
| 10320 | F: include/linux/regulator/ | 10321 | F: include/linux/regulator/ |
| 10321 | 10322 | ||
| 10322 | VT1211 HARDWARE MONITOR DRIVER | 10323 | VT1211 HARDWARE MONITOR DRIVER |
| 10323 | M: Juerg Haefliger <juergh@gmail.com> | 10324 | M: Juerg Haefliger <juergh@gmail.com> |
| 10324 | L: lm-sensors@lm-sensors.org | 10325 | L: lm-sensors@lm-sensors.org |
| 10325 | S: Maintained | 10326 | S: Maintained |
| 10326 | F: Documentation/hwmon/vt1211 | 10327 | F: Documentation/hwmon/vt1211 |
| 10327 | F: drivers/hwmon/vt1211.c | 10328 | F: drivers/hwmon/vt1211.c |
| 10328 | 10329 | ||
| 10329 | VT8231 HARDWARE MONITOR DRIVER | 10330 | VT8231 HARDWARE MONITOR DRIVER |
| 10330 | M: Roger Lucas <vt8231@hiddenengine.co.uk> | 10331 | M: Roger Lucas <vt8231@hiddenengine.co.uk> |
| 10331 | L: lm-sensors@lm-sensors.org | 10332 | L: lm-sensors@lm-sensors.org |
| 10332 | S: Maintained | 10333 | S: Maintained |
| 10333 | F: drivers/hwmon/vt8231.c | 10334 | F: drivers/hwmon/vt8231.c |
| 10334 | 10335 | ||
| 10335 | VUB300 USB to SDIO/SD/MMC bridge chip | 10336 | VUB300 USB to SDIO/SD/MMC bridge chip |
| 10336 | M: Tony Olech <tony.olech@elandigitalsystems.com> | 10337 | M: Tony Olech <tony.olech@elandigitalsystems.com> |
| 10337 | L: linux-mmc@vger.kernel.org | 10338 | L: linux-mmc@vger.kernel.org |
| 10338 | L: linux-usb@vger.kernel.org | 10339 | L: linux-usb@vger.kernel.org |
| 10339 | S: Supported | 10340 | S: Supported |
| 10340 | F: drivers/mmc/host/vub300.c | 10341 | F: drivers/mmc/host/vub300.c |
| 10341 | 10342 | ||
| 10342 | W1 DALLAS'S 1-WIRE BUS | 10343 | W1 DALLAS'S 1-WIRE BUS |
| 10343 | M: Evgeniy Polyakov <zbr@ioremap.net> | 10344 | M: Evgeniy Polyakov <zbr@ioremap.net> |
| 10344 | S: Maintained | 10345 | S: Maintained |
| 10345 | F: Documentation/w1/ | 10346 | F: Documentation/w1/ |
| 10346 | F: drivers/w1/ | 10347 | F: drivers/w1/ |
| 10347 | 10348 | ||
| 10348 | W83791D HARDWARE MONITORING DRIVER | 10349 | W83791D HARDWARE MONITORING DRIVER |
| 10349 | M: Marc Hulsman <m.hulsman@tudelft.nl> | 10350 | M: Marc Hulsman <m.hulsman@tudelft.nl> |
| 10350 | L: lm-sensors@lm-sensors.org | 10351 | L: lm-sensors@lm-sensors.org |
| 10351 | S: Maintained | 10352 | S: Maintained |
| 10352 | F: Documentation/hwmon/w83791d | 10353 | F: Documentation/hwmon/w83791d |
| 10353 | F: drivers/hwmon/w83791d.c | 10354 | F: drivers/hwmon/w83791d.c |
| 10354 | 10355 | ||
| 10355 | W83793 HARDWARE MONITORING DRIVER | 10356 | W83793 HARDWARE MONITORING DRIVER |
| 10356 | M: Rudolf Marek <r.marek@assembler.cz> | 10357 | M: Rudolf Marek <r.marek@assembler.cz> |
| 10357 | L: lm-sensors@lm-sensors.org | 10358 | L: lm-sensors@lm-sensors.org |
| 10358 | S: Maintained | 10359 | S: Maintained |
| 10359 | F: Documentation/hwmon/w83793 | 10360 | F: Documentation/hwmon/w83793 |
| 10360 | F: drivers/hwmon/w83793.c | 10361 | F: drivers/hwmon/w83793.c |
| 10361 | 10362 | ||
| 10362 | W83795 HARDWARE MONITORING DRIVER | 10363 | W83795 HARDWARE MONITORING DRIVER |
| 10363 | M: Jean Delvare <jdelvare@suse.de> | 10364 | M: Jean Delvare <jdelvare@suse.de> |
| 10364 | L: lm-sensors@lm-sensors.org | 10365 | L: lm-sensors@lm-sensors.org |
| 10365 | S: Maintained | 10366 | S: Maintained |
| 10366 | F: drivers/hwmon/w83795.c | 10367 | F: drivers/hwmon/w83795.c |
| 10367 | 10368 | ||
| 10368 | W83L51xD SD/MMC CARD INTERFACE DRIVER | 10369 | W83L51xD SD/MMC CARD INTERFACE DRIVER |
| 10369 | M: Pierre Ossman <pierre@ossman.eu> | 10370 | M: Pierre Ossman <pierre@ossman.eu> |
| 10370 | S: Maintained | 10371 | S: Maintained |
| 10371 | F: drivers/mmc/host/wbsd.* | 10372 | F: drivers/mmc/host/wbsd.* |
| 10372 | 10373 | ||
| 10373 | WACOM PROTOCOL 4 SERIAL TABLETS | 10374 | WACOM PROTOCOL 4 SERIAL TABLETS |
| 10374 | M: Julian Squires <julian@cipht.net> | 10375 | M: Julian Squires <julian@cipht.net> |
| 10375 | M: Hans de Goede <hdegoede@redhat.com> | 10376 | M: Hans de Goede <hdegoede@redhat.com> |
| 10376 | L: linux-input@vger.kernel.org | 10377 | L: linux-input@vger.kernel.org |
| 10377 | S: Maintained | 10378 | S: Maintained |
| 10378 | F: drivers/input/tablet/wacom_serial4.c | 10379 | F: drivers/input/tablet/wacom_serial4.c |
| 10379 | 10380 | ||
| 10380 | WATCHDOG DEVICE DRIVERS | 10381 | WATCHDOG DEVICE DRIVERS |
| 10381 | M: Wim Van Sebroeck <wim@iguana.be> | 10382 | M: Wim Van Sebroeck <wim@iguana.be> |
| 10382 | L: linux-watchdog@vger.kernel.org | 10383 | L: linux-watchdog@vger.kernel.org |
| 10383 | W: http://www.linux-watchdog.org/ | 10384 | W: http://www.linux-watchdog.org/ |
| 10384 | T: git git://www.linux-watchdog.org/linux-watchdog.git | 10385 | T: git git://www.linux-watchdog.org/linux-watchdog.git |
| 10385 | S: Maintained | 10386 | S: Maintained |
| 10386 | F: Documentation/watchdog/ | 10387 | F: Documentation/watchdog/ |
| 10387 | F: drivers/watchdog/ | 10388 | F: drivers/watchdog/ |
| 10388 | F: include/linux/watchdog.h | 10389 | F: include/linux/watchdog.h |
| 10389 | F: include/uapi/linux/watchdog.h | 10390 | F: include/uapi/linux/watchdog.h |
| 10390 | 10391 | ||
| 10391 | WD7000 SCSI DRIVER | 10392 | WD7000 SCSI DRIVER |
| 10392 | M: Miroslav Zagorac <zaga@fly.cc.fer.hr> | 10393 | M: Miroslav Zagorac <zaga@fly.cc.fer.hr> |
| 10393 | L: linux-scsi@vger.kernel.org | 10394 | L: linux-scsi@vger.kernel.org |
| 10394 | S: Maintained | 10395 | S: Maintained |
| 10395 | F: drivers/scsi/wd7000.c | 10396 | F: drivers/scsi/wd7000.c |
| 10396 | 10397 | ||
| 10397 | WIIMOTE HID DRIVER | 10398 | WIIMOTE HID DRIVER |
| 10398 | M: David Herrmann <dh.herrmann@googlemail.com> | 10399 | M: David Herrmann <dh.herrmann@googlemail.com> |
| 10399 | L: linux-input@vger.kernel.org | 10400 | L: linux-input@vger.kernel.org |
| 10400 | S: Maintained | 10401 | S: Maintained |
| 10401 | F: drivers/hid/hid-wiimote* | 10402 | F: drivers/hid/hid-wiimote* |
| 10402 | 10403 | ||
| 10403 | WINBOND CIR DRIVER | 10404 | WINBOND CIR DRIVER |
| 10404 | M: David Hรคrdeman <david@hardeman.nu> | 10405 | M: David Hรคrdeman <david@hardeman.nu> |
| 10405 | S: Maintained | 10406 | S: Maintained |
| 10406 | F: drivers/media/rc/winbond-cir.c | 10407 | F: drivers/media/rc/winbond-cir.c |
| 10407 | 10408 | ||
| 10408 | WIMAX STACK | 10409 | WIMAX STACK |
| 10409 | M: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com> | 10410 | M: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com> |
| 10410 | M: linux-wimax@intel.com | 10411 | M: linux-wimax@intel.com |
| 10411 | L: wimax@linuxwimax.org (subscribers-only) | 10412 | L: wimax@linuxwimax.org (subscribers-only) |
| 10412 | S: Supported | 10413 | S: Supported |
| 10413 | W: http://linuxwimax.org | 10414 | W: http://linuxwimax.org |
| 10414 | F: Documentation/wimax/README.wimax | 10415 | F: Documentation/wimax/README.wimax |
| 10415 | F: include/linux/wimax/debug.h | 10416 | F: include/linux/wimax/debug.h |
| 10416 | F: include/net/wimax.h | 10417 | F: include/net/wimax.h |
| 10417 | F: include/uapi/linux/wimax.h | 10418 | F: include/uapi/linux/wimax.h |
| 10418 | F: net/wimax/ | 10419 | F: net/wimax/ |
| 10419 | 10420 | ||
| 10420 | WISTRON LAPTOP BUTTON DRIVER | 10421 | WISTRON LAPTOP BUTTON DRIVER |
| 10421 | M: Miloslav Trmac <mitr@volny.cz> | 10422 | M: Miloslav Trmac <mitr@volny.cz> |
| 10422 | S: Maintained | 10423 | S: Maintained |
| 10423 | F: drivers/input/misc/wistron_btns.c | 10424 | F: drivers/input/misc/wistron_btns.c |
| 10424 | 10425 | ||
| 10425 | WL3501 WIRELESS PCMCIA CARD DRIVER | 10426 | WL3501 WIRELESS PCMCIA CARD DRIVER |
| 10426 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> | 10427 | M: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> |
| 10427 | L: linux-wireless@vger.kernel.org | 10428 | L: linux-wireless@vger.kernel.org |
| 10428 | W: http://oops.ghostprotocols.net:81/blog | 10429 | W: http://oops.ghostprotocols.net:81/blog |
| 10429 | S: Maintained | 10430 | S: Maintained |
| 10430 | F: drivers/net/wireless/wl3501* | 10431 | F: drivers/net/wireless/wl3501* |
| 10431 | 10432 | ||
| 10432 | WM97XX TOUCHSCREEN DRIVERS | 10433 | WM97XX TOUCHSCREEN DRIVERS |
| 10433 | M: Mark Brown <broonie@kernel.org> | 10434 | M: Mark Brown <broonie@kernel.org> |
| 10434 | M: Liam Girdwood <lrg@slimlogic.co.uk> | 10435 | M: Liam Girdwood <lrg@slimlogic.co.uk> |
| 10435 | L: linux-input@vger.kernel.org | 10436 | L: linux-input@vger.kernel.org |
| 10436 | T: git git://opensource.wolfsonmicro.com/linux-2.6-touch | 10437 | T: git git://opensource.wolfsonmicro.com/linux-2.6-touch |
| 10437 | W: http://opensource.wolfsonmicro.com/node/7 | 10438 | W: http://opensource.wolfsonmicro.com/node/7 |
| 10438 | S: Supported | 10439 | S: Supported |
| 10439 | F: drivers/input/touchscreen/*wm97* | 10440 | F: drivers/input/touchscreen/*wm97* |
| 10440 | F: include/linux/wm97xx.h | 10441 | F: include/linux/wm97xx.h |
| 10441 | 10442 | ||
| 10442 | WOLFSON MICROELECTRONICS DRIVERS | 10443 | WOLFSON MICROELECTRONICS DRIVERS |
| 10443 | L: patches@opensource.wolfsonmicro.com | 10444 | L: patches@opensource.wolfsonmicro.com |
| 10444 | T: git git://opensource.wolfsonmicro.com/linux-2.6-asoc | 10445 | T: git git://opensource.wolfsonmicro.com/linux-2.6-asoc |
| 10445 | T: git git://opensource.wolfsonmicro.com/linux-2.6-audioplus | 10446 | T: git git://opensource.wolfsonmicro.com/linux-2.6-audioplus |
| 10446 | W: http://opensource.wolfsonmicro.com/content/linux-drivers-wolfson-devices | 10447 | W: http://opensource.wolfsonmicro.com/content/linux-drivers-wolfson-devices |
| 10447 | S: Supported | 10448 | S: Supported |
| 10448 | F: Documentation/hwmon/wm83?? | 10449 | F: Documentation/hwmon/wm83?? |
| 10449 | F: arch/arm/mach-s3c64xx/mach-crag6410* | 10450 | F: arch/arm/mach-s3c64xx/mach-crag6410* |
| 10450 | F: drivers/clk/clk-wm83*.c | 10451 | F: drivers/clk/clk-wm83*.c |
| 10451 | F: drivers/extcon/extcon-arizona.c | 10452 | F: drivers/extcon/extcon-arizona.c |
| 10452 | F: drivers/leds/leds-wm83*.c | 10453 | F: drivers/leds/leds-wm83*.c |
| 10453 | F: drivers/gpio/gpio-*wm*.c | 10454 | F: drivers/gpio/gpio-*wm*.c |
| 10454 | F: drivers/gpio/gpio-arizona.c | 10455 | F: drivers/gpio/gpio-arizona.c |
| 10455 | F: drivers/hwmon/wm83??-hwmon.c | 10456 | F: drivers/hwmon/wm83??-hwmon.c |
| 10456 | F: drivers/input/misc/wm831x-on.c | 10457 | F: drivers/input/misc/wm831x-on.c |
| 10457 | F: drivers/input/touchscreen/wm831x-ts.c | 10458 | F: drivers/input/touchscreen/wm831x-ts.c |
| 10458 | F: drivers/input/touchscreen/wm97*.c | 10459 | F: drivers/input/touchscreen/wm97*.c |
| 10459 | F: drivers/mfd/arizona* | 10460 | F: drivers/mfd/arizona* |
| 10460 | F: drivers/mfd/wm*.c | 10461 | F: drivers/mfd/wm*.c |
| 10461 | F: drivers/power/wm83*.c | 10462 | F: drivers/power/wm83*.c |
| 10462 | F: drivers/rtc/rtc-wm83*.c | 10463 | F: drivers/rtc/rtc-wm83*.c |
| 10463 | F: drivers/regulator/wm8*.c | 10464 | F: drivers/regulator/wm8*.c |
| 10464 | F: drivers/video/backlight/wm83*_bl.c | 10465 | F: drivers/video/backlight/wm83*_bl.c |
| 10465 | F: drivers/watchdog/wm83*_wdt.c | 10466 | F: drivers/watchdog/wm83*_wdt.c |
| 10466 | F: include/linux/mfd/arizona/ | 10467 | F: include/linux/mfd/arizona/ |
| 10467 | F: include/linux/mfd/wm831x/ | 10468 | F: include/linux/mfd/wm831x/ |
| 10468 | F: include/linux/mfd/wm8350/ | 10469 | F: include/linux/mfd/wm8350/ |
| 10469 | F: include/linux/mfd/wm8400* | 10470 | F: include/linux/mfd/wm8400* |
| 10470 | F: include/linux/wm97xx.h | 10471 | F: include/linux/wm97xx.h |
| 10471 | F: include/sound/wm????.h | 10472 | F: include/sound/wm????.h |
| 10472 | F: sound/soc/codecs/arizona.? | 10473 | F: sound/soc/codecs/arizona.? |
| 10473 | F: sound/soc/codecs/wm* | 10474 | F: sound/soc/codecs/wm* |
| 10474 | 10475 | ||
| 10475 | WORKQUEUE | 10476 | WORKQUEUE |
| 10476 | M: Tejun Heo <tj@kernel.org> | 10477 | M: Tejun Heo <tj@kernel.org> |
| 10477 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git | 10478 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git |
| 10478 | S: Maintained | 10479 | S: Maintained |
| 10479 | F: include/linux/workqueue.h | 10480 | F: include/linux/workqueue.h |
| 10480 | F: kernel/workqueue.c | 10481 | F: kernel/workqueue.c |
| 10481 | F: Documentation/workqueue.txt | 10482 | F: Documentation/workqueue.txt |
| 10482 | 10483 | ||
| 10483 | X.25 NETWORK LAYER | 10484 | X.25 NETWORK LAYER |
| 10484 | M: Andrew Hendry <andrew.hendry@gmail.com> | 10485 | M: Andrew Hendry <andrew.hendry@gmail.com> |
| 10485 | L: linux-x25@vger.kernel.org | 10486 | L: linux-x25@vger.kernel.org |
| 10486 | S: Odd Fixes | 10487 | S: Odd Fixes |
| 10487 | F: Documentation/networking/x25* | 10488 | F: Documentation/networking/x25* |
| 10488 | F: include/net/x25* | 10489 | F: include/net/x25* |
| 10489 | F: net/x25/ | 10490 | F: net/x25/ |
| 10490 | 10491 | ||
| 10491 | X86 ARCHITECTURE (32-BIT AND 64-BIT) | 10492 | X86 ARCHITECTURE (32-BIT AND 64-BIT) |
| 10492 | M: Thomas Gleixner <tglx@linutronix.de> | 10493 | M: Thomas Gleixner <tglx@linutronix.de> |
| 10493 | M: Ingo Molnar <mingo@redhat.com> | 10494 | M: Ingo Molnar <mingo@redhat.com> |
| 10494 | M: "H. Peter Anvin" <hpa@zytor.com> | 10495 | M: "H. Peter Anvin" <hpa@zytor.com> |
| 10495 | M: x86@kernel.org | 10496 | M: x86@kernel.org |
| 10496 | L: linux-kernel@vger.kernel.org | 10497 | L: linux-kernel@vger.kernel.org |
| 10497 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/core | 10498 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/core |
| 10498 | S: Maintained | 10499 | S: Maintained |
| 10499 | F: Documentation/x86/ | 10500 | F: Documentation/x86/ |
| 10500 | F: arch/x86/ | 10501 | F: arch/x86/ |
| 10501 | 10502 | ||
| 10502 | X86 PLATFORM DRIVERS | 10503 | X86 PLATFORM DRIVERS |
| 10503 | M: Darren Hart <dvhart@infradead.org> | 10504 | M: Darren Hart <dvhart@infradead.org> |
| 10504 | L: platform-driver-x86@vger.kernel.org | 10505 | L: platform-driver-x86@vger.kernel.org |
| 10505 | T: git git://git.infradead.org/users/dvhart/linux-platform-drivers-x86.git | 10506 | T: git git://git.infradead.org/users/dvhart/linux-platform-drivers-x86.git |
| 10506 | S: Maintained | 10507 | S: Maintained |
| 10507 | F: drivers/platform/x86/ | 10508 | F: drivers/platform/x86/ |
| 10508 | 10509 | ||
| 10509 | X86 MCE INFRASTRUCTURE | 10510 | X86 MCE INFRASTRUCTURE |
| 10510 | M: Tony Luck <tony.luck@intel.com> | 10511 | M: Tony Luck <tony.luck@intel.com> |
| 10511 | M: Borislav Petkov <bp@alien8.de> | 10512 | M: Borislav Petkov <bp@alien8.de> |
| 10512 | L: linux-edac@vger.kernel.org | 10513 | L: linux-edac@vger.kernel.org |
| 10513 | S: Maintained | 10514 | S: Maintained |
| 10514 | F: arch/x86/kernel/cpu/mcheck/* | 10515 | F: arch/x86/kernel/cpu/mcheck/* |
| 10515 | 10516 | ||
| 10516 | X86 VDSO | 10517 | X86 VDSO |
| 10517 | M: Andy Lutomirski <luto@amacapital.net> | 10518 | M: Andy Lutomirski <luto@amacapital.net> |
| 10518 | L: linux-kernel@vger.kernel.org | 10519 | L: linux-kernel@vger.kernel.org |
| 10519 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/vdso | 10520 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86/vdso |
| 10520 | S: Maintained | 10521 | S: Maintained |
| 10521 | F: arch/x86/vdso/ | 10522 | F: arch/x86/vdso/ |
| 10522 | 10523 | ||
| 10523 | XC2028/3028 TUNER DRIVER | 10524 | XC2028/3028 TUNER DRIVER |
| 10524 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> | 10525 | M: Mauro Carvalho Chehab <mchehab@osg.samsung.com> |
| 10525 | L: linux-media@vger.kernel.org | 10526 | L: linux-media@vger.kernel.org |
| 10526 | W: http://linuxtv.org | 10527 | W: http://linuxtv.org |
| 10527 | T: git git://linuxtv.org/media_tree.git | 10528 | T: git git://linuxtv.org/media_tree.git |
| 10528 | S: Maintained | 10529 | S: Maintained |
| 10529 | F: drivers/media/tuners/tuner-xc2028.* | 10530 | F: drivers/media/tuners/tuner-xc2028.* |
| 10530 | 10531 | ||
| 10531 | XEN HYPERVISOR INTERFACE | 10532 | XEN HYPERVISOR INTERFACE |
| 10532 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 10533 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 10533 | M: Boris Ostrovsky <boris.ostrovsky@oracle.com> | 10534 | M: Boris Ostrovsky <boris.ostrovsky@oracle.com> |
| 10534 | M: David Vrabel <david.vrabel@citrix.com> | 10535 | M: David Vrabel <david.vrabel@citrix.com> |
| 10535 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10536 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10536 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip.git | 10537 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip.git |
| 10537 | S: Supported | 10538 | S: Supported |
| 10538 | F: arch/x86/xen/ | 10539 | F: arch/x86/xen/ |
| 10539 | F: drivers/*/xen-*front.c | 10540 | F: drivers/*/xen-*front.c |
| 10540 | F: drivers/xen/ | 10541 | F: drivers/xen/ |
| 10541 | F: arch/x86/include/asm/xen/ | 10542 | F: arch/x86/include/asm/xen/ |
| 10542 | F: include/xen/ | 10543 | F: include/xen/ |
| 10543 | F: include/uapi/xen/ | 10544 | F: include/uapi/xen/ |
| 10544 | 10545 | ||
| 10545 | XEN HYPERVISOR ARM | 10546 | XEN HYPERVISOR ARM |
| 10546 | M: Stefano Stabellini <stefano.stabellini@eu.citrix.com> | 10547 | M: Stefano Stabellini <stefano.stabellini@eu.citrix.com> |
| 10547 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10548 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10548 | S: Supported | 10549 | S: Supported |
| 10549 | F: arch/arm/xen/ | 10550 | F: arch/arm/xen/ |
| 10550 | F: arch/arm/include/asm/xen/ | 10551 | F: arch/arm/include/asm/xen/ |
| 10551 | 10552 | ||
| 10552 | XEN HYPERVISOR ARM64 | 10553 | XEN HYPERVISOR ARM64 |
| 10553 | M: Stefano Stabellini <stefano.stabellini@eu.citrix.com> | 10554 | M: Stefano Stabellini <stefano.stabellini@eu.citrix.com> |
| 10554 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10555 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10555 | S: Supported | 10556 | S: Supported |
| 10556 | F: arch/arm64/xen/ | 10557 | F: arch/arm64/xen/ |
| 10557 | F: arch/arm64/include/asm/xen/ | 10558 | F: arch/arm64/include/asm/xen/ |
| 10558 | 10559 | ||
| 10559 | XEN NETWORK BACKEND DRIVER | 10560 | XEN NETWORK BACKEND DRIVER |
| 10560 | M: Ian Campbell <ian.campbell@citrix.com> | 10561 | M: Ian Campbell <ian.campbell@citrix.com> |
| 10561 | M: Wei Liu <wei.liu2@citrix.com> | 10562 | M: Wei Liu <wei.liu2@citrix.com> |
| 10562 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10563 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10563 | L: netdev@vger.kernel.org | 10564 | L: netdev@vger.kernel.org |
| 10564 | S: Supported | 10565 | S: Supported |
| 10565 | F: drivers/net/xen-netback/* | 10566 | F: drivers/net/xen-netback/* |
| 10566 | 10567 | ||
| 10567 | XEN PCI SUBSYSTEM | 10568 | XEN PCI SUBSYSTEM |
| 10568 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 10569 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 10569 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10570 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10570 | S: Supported | 10571 | S: Supported |
| 10571 | F: arch/x86/pci/*xen* | 10572 | F: arch/x86/pci/*xen* |
| 10572 | F: drivers/pci/*xen* | 10573 | F: drivers/pci/*xen* |
| 10573 | 10574 | ||
| 10574 | XEN BLOCK SUBSYSTEM | 10575 | XEN BLOCK SUBSYSTEM |
| 10575 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 10576 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 10576 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10577 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10577 | S: Supported | 10578 | S: Supported |
| 10578 | F: drivers/block/xen-blkback/* | 10579 | F: drivers/block/xen-blkback/* |
| 10579 | F: drivers/block/xen* | 10580 | F: drivers/block/xen* |
| 10580 | 10581 | ||
| 10581 | XEN PVSCSI DRIVERS | 10582 | XEN PVSCSI DRIVERS |
| 10582 | M: Juergen Gross <jgross@suse.com> | 10583 | M: Juergen Gross <jgross@suse.com> |
| 10583 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10584 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10584 | L: linux-scsi@vger.kernel.org | 10585 | L: linux-scsi@vger.kernel.org |
| 10585 | S: Supported | 10586 | S: Supported |
| 10586 | F: drivers/scsi/xen-scsifront.c | 10587 | F: drivers/scsi/xen-scsifront.c |
| 10587 | F: drivers/xen/xen-scsiback.c | 10588 | F: drivers/xen/xen-scsiback.c |
| 10588 | F: include/xen/interface/io/vscsiif.h | 10589 | F: include/xen/interface/io/vscsiif.h |
| 10589 | 10590 | ||
| 10590 | XEN SWIOTLB SUBSYSTEM | 10591 | XEN SWIOTLB SUBSYSTEM |
| 10591 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> | 10592 | M: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> |
| 10592 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) | 10593 | L: xen-devel@lists.xenproject.org (moderated for non-subscribers) |
| 10593 | S: Supported | 10594 | S: Supported |
| 10594 | F: arch/x86/xen/*swiotlb* | 10595 | F: arch/x86/xen/*swiotlb* |
| 10595 | F: drivers/xen/*swiotlb* | 10596 | F: drivers/xen/*swiotlb* |
| 10596 | 10597 | ||
| 10597 | XFS FILESYSTEM | 10598 | XFS FILESYSTEM |
| 10598 | P: Silicon Graphics Inc | 10599 | P: Silicon Graphics Inc |
| 10599 | M: Dave Chinner <david@fromorbit.com> | 10600 | M: Dave Chinner <david@fromorbit.com> |
| 10600 | M: xfs@oss.sgi.com | 10601 | M: xfs@oss.sgi.com |
| 10601 | L: xfs@oss.sgi.com | 10602 | L: xfs@oss.sgi.com |
| 10602 | W: http://oss.sgi.com/projects/xfs | 10603 | W: http://oss.sgi.com/projects/xfs |
| 10603 | T: git git://oss.sgi.com/xfs/xfs.git | 10604 | T: git git://oss.sgi.com/xfs/xfs.git |
| 10604 | S: Supported | 10605 | S: Supported |
| 10605 | F: Documentation/filesystems/xfs.txt | 10606 | F: Documentation/filesystems/xfs.txt |
| 10606 | F: fs/xfs/ | 10607 | F: fs/xfs/ |
| 10607 | 10608 | ||
| 10608 | XILINX AXI ETHERNET DRIVER | 10609 | XILINX AXI ETHERNET DRIVER |
| 10609 | M: Anirudha Sarangi <anirudh@xilinx.com> | 10610 | M: Anirudha Sarangi <anirudh@xilinx.com> |
| 10610 | M: John Linn <John.Linn@xilinx.com> | 10611 | M: John Linn <John.Linn@xilinx.com> |
| 10611 | S: Maintained | 10612 | S: Maintained |
| 10612 | F: drivers/net/ethernet/xilinx/xilinx_axienet* | 10613 | F: drivers/net/ethernet/xilinx/xilinx_axienet* |
| 10613 | 10614 | ||
| 10614 | XILINX UARTLITE SERIAL DRIVER | 10615 | XILINX UARTLITE SERIAL DRIVER |
| 10615 | M: Peter Korsgaard <jacmet@sunsite.dk> | 10616 | M: Peter Korsgaard <jacmet@sunsite.dk> |
| 10616 | L: linux-serial@vger.kernel.org | 10617 | L: linux-serial@vger.kernel.org |
| 10617 | S: Maintained | 10618 | S: Maintained |
| 10618 | F: drivers/tty/serial/uartlite.c | 10619 | F: drivers/tty/serial/uartlite.c |
| 10619 | 10620 | ||
| 10620 | XILLYBUS DRIVER | 10621 | XILLYBUS DRIVER |
| 10621 | M: Eli Billauer <eli.billauer@gmail.com> | 10622 | M: Eli Billauer <eli.billauer@gmail.com> |
| 10622 | L: linux-kernel@vger.kernel.org | 10623 | L: linux-kernel@vger.kernel.org |
| 10623 | S: Supported | 10624 | S: Supported |
| 10624 | F: drivers/char/xillybus/ | 10625 | F: drivers/char/xillybus/ |
| 10625 | 10626 | ||
| 10626 | XTENSA XTFPGA PLATFORM SUPPORT | 10627 | XTENSA XTFPGA PLATFORM SUPPORT |
| 10627 | M: Max Filippov <jcmvbkbc@gmail.com> | 10628 | M: Max Filippov <jcmvbkbc@gmail.com> |
| 10628 | L: linux-xtensa@linux-xtensa.org | 10629 | L: linux-xtensa@linux-xtensa.org |
| 10629 | S: Maintained | 10630 | S: Maintained |
| 10630 | F: drivers/spi/spi-xtensa-xtfpga.c | 10631 | F: drivers/spi/spi-xtensa-xtfpga.c |
| 10631 | 10632 | ||
| 10632 | YAM DRIVER FOR AX.25 | 10633 | YAM DRIVER FOR AX.25 |
| 10633 | M: Jean-Paul Roubelat <jpr@f6fbb.org> | 10634 | M: Jean-Paul Roubelat <jpr@f6fbb.org> |
| 10634 | L: linux-hams@vger.kernel.org | 10635 | L: linux-hams@vger.kernel.org |
| 10635 | S: Maintained | 10636 | S: Maintained |
| 10636 | F: drivers/net/hamradio/yam* | 10637 | F: drivers/net/hamradio/yam* |
| 10637 | F: include/linux/yam.h | 10638 | F: include/linux/yam.h |
| 10638 | 10639 | ||
| 10639 | YEALINK PHONE DRIVER | 10640 | YEALINK PHONE DRIVER |
| 10640 | M: Henk Vergonet <Henk.Vergonet@gmail.com> | 10641 | M: Henk Vergonet <Henk.Vergonet@gmail.com> |
| 10641 | L: usbb2k-api-dev@nongnu.org | 10642 | L: usbb2k-api-dev@nongnu.org |
| 10642 | S: Maintained | 10643 | S: Maintained |
| 10643 | F: Documentation/input/yealink.txt | 10644 | F: Documentation/input/yealink.txt |
| 10644 | F: drivers/input/misc/yealink.* | 10645 | F: drivers/input/misc/yealink.* |
| 10645 | 10646 | ||
| 10646 | Z8530 DRIVER FOR AX.25 | 10647 | Z8530 DRIVER FOR AX.25 |
| 10647 | M: Joerg Reuter <jreuter@yaina.de> | 10648 | M: Joerg Reuter <jreuter@yaina.de> |
| 10648 | W: http://yaina.de/jreuter/ | 10649 | W: http://yaina.de/jreuter/ |
| 10649 | W: http://www.qsl.net/dl1bke/ | 10650 | W: http://www.qsl.net/dl1bke/ |
| 10650 | L: linux-hams@vger.kernel.org | 10651 | L: linux-hams@vger.kernel.org |
| 10651 | S: Maintained | 10652 | S: Maintained |
| 10652 | F: Documentation/networking/z8530drv.txt | 10653 | F: Documentation/networking/z8530drv.txt |
| 10653 | F: drivers/net/hamradio/*scc.c | 10654 | F: drivers/net/hamradio/*scc.c |
| 10654 | F: drivers/net/hamradio/z8530.h | 10655 | F: drivers/net/hamradio/z8530.h |
| 10655 | 10656 | ||
| 10656 | ZBUD COMPRESSED PAGE ALLOCATOR | 10657 | ZBUD COMPRESSED PAGE ALLOCATOR |
| 10657 | M: Seth Jennings <sjennings@variantweb.net> | 10658 | M: Seth Jennings <sjennings@variantweb.net> |
| 10658 | L: linux-mm@kvack.org | 10659 | L: linux-mm@kvack.org |
| 10659 | S: Maintained | 10660 | S: Maintained |
| 10660 | F: mm/zbud.c | 10661 | F: mm/zbud.c |
| 10661 | F: include/linux/zbud.h | 10662 | F: include/linux/zbud.h |
| 10662 | 10663 | ||
| 10663 | ZD1211RW WIRELESS DRIVER | 10664 | ZD1211RW WIRELESS DRIVER |
| 10664 | M: Daniel Drake <dsd@gentoo.org> | 10665 | M: Daniel Drake <dsd@gentoo.org> |
| 10665 | M: Ulrich Kunitz <kune@deine-taler.de> | 10666 | M: Ulrich Kunitz <kune@deine-taler.de> |
| 10666 | W: http://zd1211.ath.cx/wiki/DriverRewrite | 10667 | W: http://zd1211.ath.cx/wiki/DriverRewrite |
| 10667 | L: linux-wireless@vger.kernel.org | 10668 | L: linux-wireless@vger.kernel.org |
| 10668 | L: zd1211-devs@lists.sourceforge.net (subscribers-only) | 10669 | L: zd1211-devs@lists.sourceforge.net (subscribers-only) |
| 10669 | S: Maintained | 10670 | S: Maintained |
| 10670 | F: drivers/net/wireless/zd1211rw/ | 10671 | F: drivers/net/wireless/zd1211rw/ |
| 10671 | 10672 | ||
| 10672 | ZR36067 VIDEO FOR LINUX DRIVER | 10673 | ZR36067 VIDEO FOR LINUX DRIVER |
| 10673 | L: mjpeg-users@lists.sourceforge.net | 10674 | L: mjpeg-users@lists.sourceforge.net |
| 10674 | L: linux-media@vger.kernel.org | 10675 | L: linux-media@vger.kernel.org |
| 10675 | W: http://mjpeg.sourceforge.net/driver-zoran/ | 10676 | W: http://mjpeg.sourceforge.net/driver-zoran/ |
| 10676 | T: hg http://linuxtv.org/hg/v4l-dvb | 10677 | T: hg http://linuxtv.org/hg/v4l-dvb |
| 10677 | S: Odd Fixes | 10678 | S: Odd Fixes |
| 10678 | F: drivers/media/pci/zoran/ | 10679 | F: drivers/media/pci/zoran/ |
| 10679 | 10680 | ||
| 10680 | ZRAM COMPRESSED RAM BLOCK DEVICE DRVIER | 10681 | ZRAM COMPRESSED RAM BLOCK DEVICE DRVIER |
| 10681 | M: Minchan Kim <minchan@kernel.org> | 10682 | M: Minchan Kim <minchan@kernel.org> |
| 10682 | M: Nitin Gupta <ngupta@vflare.org> | 10683 | M: Nitin Gupta <ngupta@vflare.org> |
| 10683 | L: linux-kernel@vger.kernel.org | 10684 | L: linux-kernel@vger.kernel.org |
| 10684 | S: Maintained | 10685 | S: Maintained |
| 10685 | F: drivers/block/zram/ | 10686 | F: drivers/block/zram/ |
| 10686 | F: Documentation/blockdev/zram.txt | 10687 | F: Documentation/blockdev/zram.txt |
| 10687 | 10688 | ||
| 10688 | ZS DECSTATION Z85C30 SERIAL DRIVER | 10689 | ZS DECSTATION Z85C30 SERIAL DRIVER |
| 10689 | M: "Maciej W. Rozycki" <macro@linux-mips.org> | 10690 | M: "Maciej W. Rozycki" <macro@linux-mips.org> |
| 10690 | S: Maintained | 10691 | S: Maintained |
| 10691 | F: drivers/tty/serial/zs.* | 10692 | F: drivers/tty/serial/zs.* |
| 10692 | 10693 | ||
| 10693 | ZSMALLOC COMPRESSED SLAB MEMORY ALLOCATOR | 10694 | ZSMALLOC COMPRESSED SLAB MEMORY ALLOCATOR |
| 10694 | M: Minchan Kim <minchan@kernel.org> | 10695 | M: Minchan Kim <minchan@kernel.org> |
| 10695 | M: Nitin Gupta <ngupta@vflare.org> | 10696 | M: Nitin Gupta <ngupta@vflare.org> |
| 10696 | L: linux-mm@kvack.org | 10697 | L: linux-mm@kvack.org |
| 10697 | S: Maintained | 10698 | S: Maintained |
| 10698 | F: mm/zsmalloc.c | 10699 | F: mm/zsmalloc.c |
| 10699 | F: include/linux/zsmalloc.h | 10700 | F: include/linux/zsmalloc.h |
| 10700 | 10701 | ||
| 10701 | ZSWAP COMPRESSED SWAP CACHING | 10702 | ZSWAP COMPRESSED SWAP CACHING |
| 10702 | M: Seth Jennings <sjennings@variantweb.net> | 10703 | M: Seth Jennings <sjennings@variantweb.net> |
| 10703 | L: linux-mm@kvack.org | 10704 | L: linux-mm@kvack.org |
| 10704 | S: Maintained | 10705 | S: Maintained |
| 10705 | F: mm/zswap.c | 10706 | F: mm/zswap.c |
| 10706 | 10707 | ||
| 10707 | THE REST | 10708 | THE REST |
| 10708 | M: Linus Torvalds <torvalds@linux-foundation.org> | 10709 | M: Linus Torvalds <torvalds@linux-foundation.org> |
| 10709 | L: linux-kernel@vger.kernel.org | 10710 | L: linux-kernel@vger.kernel.org |
| 10710 | Q: http://patchwork.kernel.org/project/LKML/list/ | 10711 | Q: http://patchwork.kernel.org/project/LKML/list/ |
| 10711 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git | 10712 | T: git git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git |
| 10712 | S: Buried alive in reporters | 10713 | S: Buried alive in reporters |
| 10713 | F: * | 10714 | F: * |
| 10714 | F: */ | 10715 | F: */ |
| 10715 | 10716 |
fs/btrfs/backref.c
| 1 | /* | 1 | /* |
| 2 | * Copyright (C) 2011 STRATO. All rights reserved. | 2 | * Copyright (C) 2011 STRATO. All rights reserved. |
| 3 | * | 3 | * |
| 4 | * This program is free software; you can redistribute it and/or | 4 | * This program is free software; you can redistribute it and/or |
| 5 | * modify it under the terms of the GNU General Public | 5 | * modify it under the terms of the GNU General Public |
| 6 | * License v2 as published by the Free Software Foundation. | 6 | * License v2 as published by the Free Software Foundation. |
| 7 | * | 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, | 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 11 | * General Public License for more details. | 11 | * General Public License for more details. |
| 12 | * | 12 | * |
| 13 | * You should have received a copy of the GNU General Public | 13 | * You should have received a copy of the GNU General Public |
| 14 | * License along with this program; if not, write to the | 14 | * License along with this program; if not, write to the |
| 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 16 | * Boston, MA 021110-1307, USA. | 16 | * Boston, MA 021110-1307, USA. |
| 17 | */ | 17 | */ |
| 18 | 18 | ||
| 19 | #include <linux/vmalloc.h> | 19 | #include <linux/vmalloc.h> |
| 20 | #include "ctree.h" | 20 | #include "ctree.h" |
| 21 | #include "disk-io.h" | 21 | #include "disk-io.h" |
| 22 | #include "backref.h" | 22 | #include "backref.h" |
| 23 | #include "ulist.h" | 23 | #include "ulist.h" |
| 24 | #include "transaction.h" | 24 | #include "transaction.h" |
| 25 | #include "delayed-ref.h" | 25 | #include "delayed-ref.h" |
| 26 | #include "locking.h" | 26 | #include "locking.h" |
| 27 | 27 | ||
| 28 | /* Just an arbitrary number so we can be sure this happened */ | 28 | /* Just an arbitrary number so we can be sure this happened */ |
| 29 | #define BACKREF_FOUND_SHARED 6 | 29 | #define BACKREF_FOUND_SHARED 6 |
| 30 | 30 | ||
| 31 | struct extent_inode_elem { | 31 | struct extent_inode_elem { |
| 32 | u64 inum; | 32 | u64 inum; |
| 33 | u64 offset; | 33 | u64 offset; |
| 34 | struct extent_inode_elem *next; | 34 | struct extent_inode_elem *next; |
| 35 | }; | 35 | }; |
| 36 | 36 | ||
| 37 | static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb, | 37 | static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb, |
| 38 | struct btrfs_file_extent_item *fi, | 38 | struct btrfs_file_extent_item *fi, |
| 39 | u64 extent_item_pos, | 39 | u64 extent_item_pos, |
| 40 | struct extent_inode_elem **eie) | 40 | struct extent_inode_elem **eie) |
| 41 | { | 41 | { |
| 42 | u64 offset = 0; | 42 | u64 offset = 0; |
| 43 | struct extent_inode_elem *e; | 43 | struct extent_inode_elem *e; |
| 44 | 44 | ||
| 45 | if (!btrfs_file_extent_compression(eb, fi) && | 45 | if (!btrfs_file_extent_compression(eb, fi) && |
| 46 | !btrfs_file_extent_encryption(eb, fi) && | 46 | !btrfs_file_extent_encryption(eb, fi) && |
| 47 | !btrfs_file_extent_other_encoding(eb, fi)) { | 47 | !btrfs_file_extent_other_encoding(eb, fi)) { |
| 48 | u64 data_offset; | 48 | u64 data_offset; |
| 49 | u64 data_len; | 49 | u64 data_len; |
| 50 | 50 | ||
| 51 | data_offset = btrfs_file_extent_offset(eb, fi); | 51 | data_offset = btrfs_file_extent_offset(eb, fi); |
| 52 | data_len = btrfs_file_extent_num_bytes(eb, fi); | 52 | data_len = btrfs_file_extent_num_bytes(eb, fi); |
| 53 | 53 | ||
| 54 | if (extent_item_pos < data_offset || | 54 | if (extent_item_pos < data_offset || |
| 55 | extent_item_pos >= data_offset + data_len) | 55 | extent_item_pos >= data_offset + data_len) |
| 56 | return 1; | 56 | return 1; |
| 57 | offset = extent_item_pos - data_offset; | 57 | offset = extent_item_pos - data_offset; |
| 58 | } | 58 | } |
| 59 | 59 | ||
| 60 | e = kmalloc(sizeof(*e), GFP_NOFS); | 60 | e = kmalloc(sizeof(*e), GFP_NOFS); |
| 61 | if (!e) | 61 | if (!e) |
| 62 | return -ENOMEM; | 62 | return -ENOMEM; |
| 63 | 63 | ||
| 64 | e->next = *eie; | 64 | e->next = *eie; |
| 65 | e->inum = key->objectid; | 65 | e->inum = key->objectid; |
| 66 | e->offset = key->offset + offset; | 66 | e->offset = key->offset + offset; |
| 67 | *eie = e; | 67 | *eie = e; |
| 68 | 68 | ||
| 69 | return 0; | 69 | return 0; |
| 70 | } | 70 | } |
| 71 | 71 | ||
| 72 | static void free_inode_elem_list(struct extent_inode_elem *eie) | 72 | static void free_inode_elem_list(struct extent_inode_elem *eie) |
| 73 | { | 73 | { |
| 74 | struct extent_inode_elem *eie_next; | 74 | struct extent_inode_elem *eie_next; |
| 75 | 75 | ||
| 76 | for (; eie; eie = eie_next) { | 76 | for (; eie; eie = eie_next) { |
| 77 | eie_next = eie->next; | 77 | eie_next = eie->next; |
| 78 | kfree(eie); | 78 | kfree(eie); |
| 79 | } | 79 | } |
| 80 | } | 80 | } |
| 81 | 81 | ||
| 82 | static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte, | 82 | static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte, |
| 83 | u64 extent_item_pos, | 83 | u64 extent_item_pos, |
| 84 | struct extent_inode_elem **eie) | 84 | struct extent_inode_elem **eie) |
| 85 | { | 85 | { |
| 86 | u64 disk_byte; | 86 | u64 disk_byte; |
| 87 | struct btrfs_key key; | 87 | struct btrfs_key key; |
| 88 | struct btrfs_file_extent_item *fi; | 88 | struct btrfs_file_extent_item *fi; |
| 89 | int slot; | 89 | int slot; |
| 90 | int nritems; | 90 | int nritems; |
| 91 | int extent_type; | 91 | int extent_type; |
| 92 | int ret; | 92 | int ret; |
| 93 | 93 | ||
| 94 | /* | 94 | /* |
| 95 | * from the shared data ref, we only have the leaf but we need | 95 | * from the shared data ref, we only have the leaf but we need |
| 96 | * the key. thus, we must look into all items and see that we | 96 | * the key. thus, we must look into all items and see that we |
| 97 | * find one (some) with a reference to our extent item. | 97 | * find one (some) with a reference to our extent item. |
| 98 | */ | 98 | */ |
| 99 | nritems = btrfs_header_nritems(eb); | 99 | nritems = btrfs_header_nritems(eb); |
| 100 | for (slot = 0; slot < nritems; ++slot) { | 100 | for (slot = 0; slot < nritems; ++slot) { |
| 101 | btrfs_item_key_to_cpu(eb, &key, slot); | 101 | btrfs_item_key_to_cpu(eb, &key, slot); |
| 102 | if (key.type != BTRFS_EXTENT_DATA_KEY) | 102 | if (key.type != BTRFS_EXTENT_DATA_KEY) |
| 103 | continue; | 103 | continue; |
| 104 | fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); | 104 | fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); |
| 105 | extent_type = btrfs_file_extent_type(eb, fi); | 105 | extent_type = btrfs_file_extent_type(eb, fi); |
| 106 | if (extent_type == BTRFS_FILE_EXTENT_INLINE) | 106 | if (extent_type == BTRFS_FILE_EXTENT_INLINE) |
| 107 | continue; | 107 | continue; |
| 108 | /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */ | 108 | /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */ |
| 109 | disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); | 109 | disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); |
| 110 | if (disk_byte != wanted_disk_byte) | 110 | if (disk_byte != wanted_disk_byte) |
| 111 | continue; | 111 | continue; |
| 112 | 112 | ||
| 113 | ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie); | 113 | ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie); |
| 114 | if (ret < 0) | 114 | if (ret < 0) |
| 115 | return ret; | 115 | return ret; |
| 116 | } | 116 | } |
| 117 | 117 | ||
| 118 | return 0; | 118 | return 0; |
| 119 | } | 119 | } |
| 120 | 120 | ||
| 121 | /* | 121 | /* |
| 122 | * this structure records all encountered refs on the way up to the root | 122 | * this structure records all encountered refs on the way up to the root |
| 123 | */ | 123 | */ |
| 124 | struct __prelim_ref { | 124 | struct __prelim_ref { |
| 125 | struct list_head list; | 125 | struct list_head list; |
| 126 | u64 root_id; | 126 | u64 root_id; |
| 127 | struct btrfs_key key_for_search; | 127 | struct btrfs_key key_for_search; |
| 128 | int level; | 128 | int level; |
| 129 | int count; | 129 | int count; |
| 130 | struct extent_inode_elem *inode_list; | 130 | struct extent_inode_elem *inode_list; |
| 131 | u64 parent; | 131 | u64 parent; |
| 132 | u64 wanted_disk_byte; | 132 | u64 wanted_disk_byte; |
| 133 | }; | 133 | }; |
| 134 | 134 | ||
| 135 | static struct kmem_cache *btrfs_prelim_ref_cache; | 135 | static struct kmem_cache *btrfs_prelim_ref_cache; |
| 136 | 136 | ||
| 137 | int __init btrfs_prelim_ref_init(void) | 137 | int __init btrfs_prelim_ref_init(void) |
| 138 | { | 138 | { |
| 139 | btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref", | 139 | btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref", |
| 140 | sizeof(struct __prelim_ref), | 140 | sizeof(struct __prelim_ref), |
| 141 | 0, | 141 | 0, |
| 142 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, | 142 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, |
| 143 | NULL); | 143 | NULL); |
| 144 | if (!btrfs_prelim_ref_cache) | 144 | if (!btrfs_prelim_ref_cache) |
| 145 | return -ENOMEM; | 145 | return -ENOMEM; |
| 146 | return 0; | 146 | return 0; |
| 147 | } | 147 | } |
| 148 | 148 | ||
| 149 | void btrfs_prelim_ref_exit(void) | 149 | void btrfs_prelim_ref_exit(void) |
| 150 | { | 150 | { |
| 151 | if (btrfs_prelim_ref_cache) | 151 | if (btrfs_prelim_ref_cache) |
| 152 | kmem_cache_destroy(btrfs_prelim_ref_cache); | 152 | kmem_cache_destroy(btrfs_prelim_ref_cache); |
| 153 | } | 153 | } |
| 154 | 154 | ||
| 155 | /* | 155 | /* |
| 156 | * the rules for all callers of this function are: | 156 | * the rules for all callers of this function are: |
| 157 | * - obtaining the parent is the goal | 157 | * - obtaining the parent is the goal |
| 158 | * - if you add a key, you must know that it is a correct key | 158 | * - if you add a key, you must know that it is a correct key |
| 159 | * - if you cannot add the parent or a correct key, then we will look into the | 159 | * - if you cannot add the parent or a correct key, then we will look into the |
| 160 | * block later to set a correct key | 160 | * block later to set a correct key |
| 161 | * | 161 | * |
| 162 | * delayed refs | 162 | * delayed refs |
| 163 | * ============ | 163 | * ============ |
| 164 | * backref type | shared | indirect | shared | indirect | 164 | * backref type | shared | indirect | shared | indirect |
| 165 | * information | tree | tree | data | data | 165 | * information | tree | tree | data | data |
| 166 | * --------------------+--------+----------+--------+---------- | 166 | * --------------------+--------+----------+--------+---------- |
| 167 | * parent logical | y | - | - | - | 167 | * parent logical | y | - | - | - |
| 168 | * key to resolve | - | y | y | y | 168 | * key to resolve | - | y | y | y |
| 169 | * tree block logical | - | - | - | - | 169 | * tree block logical | - | - | - | - |
| 170 | * root for resolving | y | y | y | y | 170 | * root for resolving | y | y | y | y |
| 171 | * | 171 | * |
| 172 | * - column 1: we've the parent -> done | 172 | * - column 1: we've the parent -> done |
| 173 | * - column 2, 3, 4: we use the key to find the parent | 173 | * - column 2, 3, 4: we use the key to find the parent |
| 174 | * | 174 | * |
| 175 | * on disk refs (inline or keyed) | 175 | * on disk refs (inline or keyed) |
| 176 | * ============================== | 176 | * ============================== |
| 177 | * backref type | shared | indirect | shared | indirect | 177 | * backref type | shared | indirect | shared | indirect |
| 178 | * information | tree | tree | data | data | 178 | * information | tree | tree | data | data |
| 179 | * --------------------+--------+----------+--------+---------- | 179 | * --------------------+--------+----------+--------+---------- |
| 180 | * parent logical | y | - | y | - | 180 | * parent logical | y | - | y | - |
| 181 | * key to resolve | - | - | - | y | 181 | * key to resolve | - | - | - | y |
| 182 | * tree block logical | y | y | y | y | 182 | * tree block logical | y | y | y | y |
| 183 | * root for resolving | - | y | y | y | 183 | * root for resolving | - | y | y | y |
| 184 | * | 184 | * |
| 185 | * - column 1, 3: we've the parent -> done | 185 | * - column 1, 3: we've the parent -> done |
| 186 | * - column 2: we take the first key from the block to find the parent | 186 | * - column 2: we take the first key from the block to find the parent |
| 187 | * (see __add_missing_keys) | 187 | * (see __add_missing_keys) |
| 188 | * - column 4: we use the key to find the parent | 188 | * - column 4: we use the key to find the parent |
| 189 | * | 189 | * |
| 190 | * additional information that's available but not required to find the parent | 190 | * additional information that's available but not required to find the parent |
| 191 | * block might help in merging entries to gain some speed. | 191 | * block might help in merging entries to gain some speed. |
| 192 | */ | 192 | */ |
| 193 | 193 | ||
| 194 | static int __add_prelim_ref(struct list_head *head, u64 root_id, | 194 | static int __add_prelim_ref(struct list_head *head, u64 root_id, |
| 195 | struct btrfs_key *key, int level, | 195 | struct btrfs_key *key, int level, |
| 196 | u64 parent, u64 wanted_disk_byte, int count, | 196 | u64 parent, u64 wanted_disk_byte, int count, |
| 197 | gfp_t gfp_mask) | 197 | gfp_t gfp_mask) |
| 198 | { | 198 | { |
| 199 | struct __prelim_ref *ref; | 199 | struct __prelim_ref *ref; |
| 200 | 200 | ||
| 201 | if (root_id == BTRFS_DATA_RELOC_TREE_OBJECTID) | 201 | if (root_id == BTRFS_DATA_RELOC_TREE_OBJECTID) |
| 202 | return 0; | 202 | return 0; |
| 203 | 203 | ||
| 204 | ref = kmem_cache_alloc(btrfs_prelim_ref_cache, gfp_mask); | 204 | ref = kmem_cache_alloc(btrfs_prelim_ref_cache, gfp_mask); |
| 205 | if (!ref) | 205 | if (!ref) |
| 206 | return -ENOMEM; | 206 | return -ENOMEM; |
| 207 | 207 | ||
| 208 | ref->root_id = root_id; | 208 | ref->root_id = root_id; |
| 209 | if (key) | 209 | if (key) |
| 210 | ref->key_for_search = *key; | 210 | ref->key_for_search = *key; |
| 211 | else | 211 | else |
| 212 | memset(&ref->key_for_search, 0, sizeof(ref->key_for_search)); | 212 | memset(&ref->key_for_search, 0, sizeof(ref->key_for_search)); |
| 213 | 213 | ||
| 214 | ref->inode_list = NULL; | 214 | ref->inode_list = NULL; |
| 215 | ref->level = level; | 215 | ref->level = level; |
| 216 | ref->count = count; | 216 | ref->count = count; |
| 217 | ref->parent = parent; | 217 | ref->parent = parent; |
| 218 | ref->wanted_disk_byte = wanted_disk_byte; | 218 | ref->wanted_disk_byte = wanted_disk_byte; |
| 219 | list_add_tail(&ref->list, head); | 219 | list_add_tail(&ref->list, head); |
| 220 | 220 | ||
| 221 | return 0; | 221 | return 0; |
| 222 | } | 222 | } |
| 223 | 223 | ||
| 224 | static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path, | 224 | static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path, |
| 225 | struct ulist *parents, struct __prelim_ref *ref, | 225 | struct ulist *parents, struct __prelim_ref *ref, |
| 226 | int level, u64 time_seq, const u64 *extent_item_pos, | 226 | int level, u64 time_seq, const u64 *extent_item_pos, |
| 227 | u64 total_refs) | 227 | u64 total_refs) |
| 228 | { | 228 | { |
| 229 | int ret = 0; | 229 | int ret = 0; |
| 230 | int slot; | 230 | int slot; |
| 231 | struct extent_buffer *eb; | 231 | struct extent_buffer *eb; |
| 232 | struct btrfs_key key; | 232 | struct btrfs_key key; |
| 233 | struct btrfs_key *key_for_search = &ref->key_for_search; | 233 | struct btrfs_key *key_for_search = &ref->key_for_search; |
| 234 | struct btrfs_file_extent_item *fi; | 234 | struct btrfs_file_extent_item *fi; |
| 235 | struct extent_inode_elem *eie = NULL, *old = NULL; | 235 | struct extent_inode_elem *eie = NULL, *old = NULL; |
| 236 | u64 disk_byte; | 236 | u64 disk_byte; |
| 237 | u64 wanted_disk_byte = ref->wanted_disk_byte; | 237 | u64 wanted_disk_byte = ref->wanted_disk_byte; |
| 238 | u64 count = 0; | 238 | u64 count = 0; |
| 239 | 239 | ||
| 240 | if (level != 0) { | 240 | if (level != 0) { |
| 241 | eb = path->nodes[level]; | 241 | eb = path->nodes[level]; |
| 242 | ret = ulist_add(parents, eb->start, 0, GFP_NOFS); | 242 | ret = ulist_add(parents, eb->start, 0, GFP_NOFS); |
| 243 | if (ret < 0) | 243 | if (ret < 0) |
| 244 | return ret; | 244 | return ret; |
| 245 | return 0; | 245 | return 0; |
| 246 | } | 246 | } |
| 247 | 247 | ||
| 248 | /* | 248 | /* |
| 249 | * We normally enter this function with the path already pointing to | 249 | * We normally enter this function with the path already pointing to |
| 250 | * the first item to check. But sometimes, we may enter it with | 250 | * the first item to check. But sometimes, we may enter it with |
| 251 | * slot==nritems. In that case, go to the next leaf before we continue. | 251 | * slot==nritems. In that case, go to the next leaf before we continue. |
| 252 | */ | 252 | */ |
| 253 | if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) | 253 | if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) |
| 254 | ret = btrfs_next_old_leaf(root, path, time_seq); | 254 | ret = btrfs_next_old_leaf(root, path, time_seq); |
| 255 | 255 | ||
| 256 | while (!ret && count < total_refs) { | 256 | while (!ret && count < total_refs) { |
| 257 | eb = path->nodes[0]; | 257 | eb = path->nodes[0]; |
| 258 | slot = path->slots[0]; | 258 | slot = path->slots[0]; |
| 259 | 259 | ||
| 260 | btrfs_item_key_to_cpu(eb, &key, slot); | 260 | btrfs_item_key_to_cpu(eb, &key, slot); |
| 261 | 261 | ||
| 262 | if (key.objectid != key_for_search->objectid || | 262 | if (key.objectid != key_for_search->objectid || |
| 263 | key.type != BTRFS_EXTENT_DATA_KEY) | 263 | key.type != BTRFS_EXTENT_DATA_KEY) |
| 264 | break; | 264 | break; |
| 265 | 265 | ||
| 266 | fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); | 266 | fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item); |
| 267 | disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); | 267 | disk_byte = btrfs_file_extent_disk_bytenr(eb, fi); |
| 268 | 268 | ||
| 269 | if (disk_byte == wanted_disk_byte) { | 269 | if (disk_byte == wanted_disk_byte) { |
| 270 | eie = NULL; | 270 | eie = NULL; |
| 271 | old = NULL; | 271 | old = NULL; |
| 272 | count++; | 272 | count++; |
| 273 | if (extent_item_pos) { | 273 | if (extent_item_pos) { |
| 274 | ret = check_extent_in_eb(&key, eb, fi, | 274 | ret = check_extent_in_eb(&key, eb, fi, |
| 275 | *extent_item_pos, | 275 | *extent_item_pos, |
| 276 | &eie); | 276 | &eie); |
| 277 | if (ret < 0) | 277 | if (ret < 0) |
| 278 | break; | 278 | break; |
| 279 | } | 279 | } |
| 280 | if (ret > 0) | 280 | if (ret > 0) |
| 281 | goto next; | 281 | goto next; |
| 282 | ret = ulist_add_merge_ptr(parents, eb->start, | 282 | ret = ulist_add_merge_ptr(parents, eb->start, |
| 283 | eie, (void **)&old, GFP_NOFS); | 283 | eie, (void **)&old, GFP_NOFS); |
| 284 | if (ret < 0) | 284 | if (ret < 0) |
| 285 | break; | 285 | break; |
| 286 | if (!ret && extent_item_pos) { | 286 | if (!ret && extent_item_pos) { |
| 287 | while (old->next) | 287 | while (old->next) |
| 288 | old = old->next; | 288 | old = old->next; |
| 289 | old->next = eie; | 289 | old->next = eie; |
| 290 | } | 290 | } |
| 291 | eie = NULL; | 291 | eie = NULL; |
| 292 | } | 292 | } |
| 293 | next: | 293 | next: |
| 294 | ret = btrfs_next_old_item(root, path, time_seq); | 294 | ret = btrfs_next_old_item(root, path, time_seq); |
| 295 | } | 295 | } |
| 296 | 296 | ||
| 297 | if (ret > 0) | 297 | if (ret > 0) |
| 298 | ret = 0; | 298 | ret = 0; |
| 299 | else if (ret < 0) | 299 | else if (ret < 0) |
| 300 | free_inode_elem_list(eie); | 300 | free_inode_elem_list(eie); |
| 301 | return ret; | 301 | return ret; |
| 302 | } | 302 | } |
| 303 | 303 | ||
| 304 | /* | 304 | /* |
| 305 | * resolve an indirect backref in the form (root_id, key, level) | 305 | * resolve an indirect backref in the form (root_id, key, level) |
| 306 | * to a logical address | 306 | * to a logical address |
| 307 | */ | 307 | */ |
| 308 | static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info, | 308 | static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info, |
| 309 | struct btrfs_path *path, u64 time_seq, | 309 | struct btrfs_path *path, u64 time_seq, |
| 310 | struct __prelim_ref *ref, | 310 | struct __prelim_ref *ref, |
| 311 | struct ulist *parents, | 311 | struct ulist *parents, |
| 312 | const u64 *extent_item_pos, u64 total_refs) | 312 | const u64 *extent_item_pos, u64 total_refs) |
| 313 | { | 313 | { |
| 314 | struct btrfs_root *root; | 314 | struct btrfs_root *root; |
| 315 | struct btrfs_key root_key; | 315 | struct btrfs_key root_key; |
| 316 | struct extent_buffer *eb; | 316 | struct extent_buffer *eb; |
| 317 | int ret = 0; | 317 | int ret = 0; |
| 318 | int root_level; | 318 | int root_level; |
| 319 | int level = ref->level; | 319 | int level = ref->level; |
| 320 | int index; | 320 | int index; |
| 321 | 321 | ||
| 322 | root_key.objectid = ref->root_id; | 322 | root_key.objectid = ref->root_id; |
| 323 | root_key.type = BTRFS_ROOT_ITEM_KEY; | 323 | root_key.type = BTRFS_ROOT_ITEM_KEY; |
| 324 | root_key.offset = (u64)-1; | 324 | root_key.offset = (u64)-1; |
| 325 | 325 | ||
| 326 | index = srcu_read_lock(&fs_info->subvol_srcu); | 326 | index = srcu_read_lock(&fs_info->subvol_srcu); |
| 327 | 327 | ||
| 328 | root = btrfs_read_fs_root_no_name(fs_info, &root_key); | 328 | root = btrfs_read_fs_root_no_name(fs_info, &root_key); |
| 329 | if (IS_ERR(root)) { | 329 | if (IS_ERR(root)) { |
| 330 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 330 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 331 | ret = PTR_ERR(root); | 331 | ret = PTR_ERR(root); |
| 332 | goto out; | 332 | goto out; |
| 333 | } | 333 | } |
| 334 | 334 | ||
| 335 | if (path->search_commit_root) | 335 | if (path->search_commit_root) |
| 336 | root_level = btrfs_header_level(root->commit_root); | 336 | root_level = btrfs_header_level(root->commit_root); |
| 337 | else | 337 | else |
| 338 | root_level = btrfs_old_root_level(root, time_seq); | 338 | root_level = btrfs_old_root_level(root, time_seq); |
| 339 | 339 | ||
| 340 | if (root_level + 1 == level) { | 340 | if (root_level + 1 == level) { |
| 341 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 341 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 342 | goto out; | 342 | goto out; |
| 343 | } | 343 | } |
| 344 | 344 | ||
| 345 | path->lowest_level = level; | 345 | path->lowest_level = level; |
| 346 | ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq); | 346 | ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq); |
| 347 | 347 | ||
| 348 | /* root node has been locked, we can release @subvol_srcu safely here */ | 348 | /* root node has been locked, we can release @subvol_srcu safely here */ |
| 349 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 349 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 350 | 350 | ||
| 351 | pr_debug("search slot in root %llu (level %d, ref count %d) returned " | 351 | pr_debug("search slot in root %llu (level %d, ref count %d) returned " |
| 352 | "%d for key (%llu %u %llu)\n", | 352 | "%d for key (%llu %u %llu)\n", |
| 353 | ref->root_id, level, ref->count, ret, | 353 | ref->root_id, level, ref->count, ret, |
| 354 | ref->key_for_search.objectid, ref->key_for_search.type, | 354 | ref->key_for_search.objectid, ref->key_for_search.type, |
| 355 | ref->key_for_search.offset); | 355 | ref->key_for_search.offset); |
| 356 | if (ret < 0) | 356 | if (ret < 0) |
| 357 | goto out; | 357 | goto out; |
| 358 | 358 | ||
| 359 | eb = path->nodes[level]; | 359 | eb = path->nodes[level]; |
| 360 | while (!eb) { | 360 | while (!eb) { |
| 361 | if (WARN_ON(!level)) { | 361 | if (WARN_ON(!level)) { |
| 362 | ret = 1; | 362 | ret = 1; |
| 363 | goto out; | 363 | goto out; |
| 364 | } | 364 | } |
| 365 | level--; | 365 | level--; |
| 366 | eb = path->nodes[level]; | 366 | eb = path->nodes[level]; |
| 367 | } | 367 | } |
| 368 | 368 | ||
| 369 | ret = add_all_parents(root, path, parents, ref, level, time_seq, | 369 | ret = add_all_parents(root, path, parents, ref, level, time_seq, |
| 370 | extent_item_pos, total_refs); | 370 | extent_item_pos, total_refs); |
| 371 | out: | 371 | out: |
| 372 | path->lowest_level = 0; | 372 | path->lowest_level = 0; |
| 373 | btrfs_release_path(path); | 373 | btrfs_release_path(path); |
| 374 | return ret; | 374 | return ret; |
| 375 | } | 375 | } |
| 376 | 376 | ||
| 377 | /* | 377 | /* |
| 378 | * resolve all indirect backrefs from the list | 378 | * resolve all indirect backrefs from the list |
| 379 | */ | 379 | */ |
| 380 | static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info, | 380 | static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info, |
| 381 | struct btrfs_path *path, u64 time_seq, | 381 | struct btrfs_path *path, u64 time_seq, |
| 382 | struct list_head *head, | 382 | struct list_head *head, |
| 383 | const u64 *extent_item_pos, u64 total_refs, | 383 | const u64 *extent_item_pos, u64 total_refs, |
| 384 | u64 root_objectid) | 384 | u64 root_objectid) |
| 385 | { | 385 | { |
| 386 | int err; | 386 | int err; |
| 387 | int ret = 0; | 387 | int ret = 0; |
| 388 | struct __prelim_ref *ref; | 388 | struct __prelim_ref *ref; |
| 389 | struct __prelim_ref *ref_safe; | 389 | struct __prelim_ref *ref_safe; |
| 390 | struct __prelim_ref *new_ref; | 390 | struct __prelim_ref *new_ref; |
| 391 | struct ulist *parents; | 391 | struct ulist *parents; |
| 392 | struct ulist_node *node; | 392 | struct ulist_node *node; |
| 393 | struct ulist_iterator uiter; | 393 | struct ulist_iterator uiter; |
| 394 | 394 | ||
| 395 | parents = ulist_alloc(GFP_NOFS); | 395 | parents = ulist_alloc(GFP_NOFS); |
| 396 | if (!parents) | 396 | if (!parents) |
| 397 | return -ENOMEM; | 397 | return -ENOMEM; |
| 398 | 398 | ||
| 399 | /* | 399 | /* |
| 400 | * _safe allows us to insert directly after the current item without | 400 | * _safe allows us to insert directly after the current item without |
| 401 | * iterating over the newly inserted items. | 401 | * iterating over the newly inserted items. |
| 402 | * we're also allowed to re-assign ref during iteration. | 402 | * we're also allowed to re-assign ref during iteration. |
| 403 | */ | 403 | */ |
| 404 | list_for_each_entry_safe(ref, ref_safe, head, list) { | 404 | list_for_each_entry_safe(ref, ref_safe, head, list) { |
| 405 | if (ref->parent) /* already direct */ | 405 | if (ref->parent) /* already direct */ |
| 406 | continue; | 406 | continue; |
| 407 | if (ref->count == 0) | 407 | if (ref->count == 0) |
| 408 | continue; | 408 | continue; |
| 409 | if (root_objectid && ref->root_id != root_objectid) { | 409 | if (root_objectid && ref->root_id != root_objectid) { |
| 410 | ret = BACKREF_FOUND_SHARED; | 410 | ret = BACKREF_FOUND_SHARED; |
| 411 | goto out; | 411 | goto out; |
| 412 | } | 412 | } |
| 413 | err = __resolve_indirect_ref(fs_info, path, time_seq, ref, | 413 | err = __resolve_indirect_ref(fs_info, path, time_seq, ref, |
| 414 | parents, extent_item_pos, | 414 | parents, extent_item_pos, |
| 415 | total_refs); | 415 | total_refs); |
| 416 | /* | 416 | /* |
| 417 | * we can only tolerate ENOENT,otherwise,we should catch error | 417 | * we can only tolerate ENOENT,otherwise,we should catch error |
| 418 | * and return directly. | 418 | * and return directly. |
| 419 | */ | 419 | */ |
| 420 | if (err == -ENOENT) { | 420 | if (err == -ENOENT) { |
| 421 | continue; | 421 | continue; |
| 422 | } else if (err) { | 422 | } else if (err) { |
| 423 | ret = err; | 423 | ret = err; |
| 424 | goto out; | 424 | goto out; |
| 425 | } | 425 | } |
| 426 | 426 | ||
| 427 | /* we put the first parent into the ref at hand */ | 427 | /* we put the first parent into the ref at hand */ |
| 428 | ULIST_ITER_INIT(&uiter); | 428 | ULIST_ITER_INIT(&uiter); |
| 429 | node = ulist_next(parents, &uiter); | 429 | node = ulist_next(parents, &uiter); |
| 430 | ref->parent = node ? node->val : 0; | 430 | ref->parent = node ? node->val : 0; |
| 431 | ref->inode_list = node ? | 431 | ref->inode_list = node ? |
| 432 | (struct extent_inode_elem *)(uintptr_t)node->aux : NULL; | 432 | (struct extent_inode_elem *)(uintptr_t)node->aux : NULL; |
| 433 | 433 | ||
| 434 | /* additional parents require new refs being added here */ | 434 | /* additional parents require new refs being added here */ |
| 435 | while ((node = ulist_next(parents, &uiter))) { | 435 | while ((node = ulist_next(parents, &uiter))) { |
| 436 | new_ref = kmem_cache_alloc(btrfs_prelim_ref_cache, | 436 | new_ref = kmem_cache_alloc(btrfs_prelim_ref_cache, |
| 437 | GFP_NOFS); | 437 | GFP_NOFS); |
| 438 | if (!new_ref) { | 438 | if (!new_ref) { |
| 439 | ret = -ENOMEM; | 439 | ret = -ENOMEM; |
| 440 | goto out; | 440 | goto out; |
| 441 | } | 441 | } |
| 442 | memcpy(new_ref, ref, sizeof(*ref)); | 442 | memcpy(new_ref, ref, sizeof(*ref)); |
| 443 | new_ref->parent = node->val; | 443 | new_ref->parent = node->val; |
| 444 | new_ref->inode_list = (struct extent_inode_elem *) | 444 | new_ref->inode_list = (struct extent_inode_elem *) |
| 445 | (uintptr_t)node->aux; | 445 | (uintptr_t)node->aux; |
| 446 | list_add(&new_ref->list, &ref->list); | 446 | list_add(&new_ref->list, &ref->list); |
| 447 | } | 447 | } |
| 448 | ulist_reinit(parents); | 448 | ulist_reinit(parents); |
| 449 | } | 449 | } |
| 450 | out: | 450 | out: |
| 451 | ulist_free(parents); | 451 | ulist_free(parents); |
| 452 | return ret; | 452 | return ret; |
| 453 | } | 453 | } |
| 454 | 454 | ||
| 455 | static inline int ref_for_same_block(struct __prelim_ref *ref1, | 455 | static inline int ref_for_same_block(struct __prelim_ref *ref1, |
| 456 | struct __prelim_ref *ref2) | 456 | struct __prelim_ref *ref2) |
| 457 | { | 457 | { |
| 458 | if (ref1->level != ref2->level) | 458 | if (ref1->level != ref2->level) |
| 459 | return 0; | 459 | return 0; |
| 460 | if (ref1->root_id != ref2->root_id) | 460 | if (ref1->root_id != ref2->root_id) |
| 461 | return 0; | 461 | return 0; |
| 462 | if (ref1->key_for_search.type != ref2->key_for_search.type) | 462 | if (ref1->key_for_search.type != ref2->key_for_search.type) |
| 463 | return 0; | 463 | return 0; |
| 464 | if (ref1->key_for_search.objectid != ref2->key_for_search.objectid) | 464 | if (ref1->key_for_search.objectid != ref2->key_for_search.objectid) |
| 465 | return 0; | 465 | return 0; |
| 466 | if (ref1->key_for_search.offset != ref2->key_for_search.offset) | 466 | if (ref1->key_for_search.offset != ref2->key_for_search.offset) |
| 467 | return 0; | 467 | return 0; |
| 468 | if (ref1->parent != ref2->parent) | 468 | if (ref1->parent != ref2->parent) |
| 469 | return 0; | 469 | return 0; |
| 470 | 470 | ||
| 471 | return 1; | 471 | return 1; |
| 472 | } | 472 | } |
| 473 | 473 | ||
| 474 | /* | 474 | /* |
| 475 | * read tree blocks and add keys where required. | 475 | * read tree blocks and add keys where required. |
| 476 | */ | 476 | */ |
| 477 | static int __add_missing_keys(struct btrfs_fs_info *fs_info, | 477 | static int __add_missing_keys(struct btrfs_fs_info *fs_info, |
| 478 | struct list_head *head) | 478 | struct list_head *head) |
| 479 | { | 479 | { |
| 480 | struct list_head *pos; | 480 | struct list_head *pos; |
| 481 | struct extent_buffer *eb; | 481 | struct extent_buffer *eb; |
| 482 | 482 | ||
| 483 | list_for_each(pos, head) { | 483 | list_for_each(pos, head) { |
| 484 | struct __prelim_ref *ref; | 484 | struct __prelim_ref *ref; |
| 485 | ref = list_entry(pos, struct __prelim_ref, list); | 485 | ref = list_entry(pos, struct __prelim_ref, list); |
| 486 | 486 | ||
| 487 | if (ref->parent) | 487 | if (ref->parent) |
| 488 | continue; | 488 | continue; |
| 489 | if (ref->key_for_search.type) | 489 | if (ref->key_for_search.type) |
| 490 | continue; | 490 | continue; |
| 491 | BUG_ON(!ref->wanted_disk_byte); | 491 | BUG_ON(!ref->wanted_disk_byte); |
| 492 | eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte, | 492 | eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte, |
| 493 | 0); | 493 | 0); |
| 494 | if (!eb || !extent_buffer_uptodate(eb)) { | 494 | if (!eb || !extent_buffer_uptodate(eb)) { |
| 495 | free_extent_buffer(eb); | 495 | free_extent_buffer(eb); |
| 496 | return -EIO; | 496 | return -EIO; |
| 497 | } | 497 | } |
| 498 | btrfs_tree_read_lock(eb); | 498 | btrfs_tree_read_lock(eb); |
| 499 | if (btrfs_header_level(eb) == 0) | 499 | if (btrfs_header_level(eb) == 0) |
| 500 | btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0); | 500 | btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0); |
| 501 | else | 501 | else |
| 502 | btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0); | 502 | btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0); |
| 503 | btrfs_tree_read_unlock(eb); | 503 | btrfs_tree_read_unlock(eb); |
| 504 | free_extent_buffer(eb); | 504 | free_extent_buffer(eb); |
| 505 | } | 505 | } |
| 506 | return 0; | 506 | return 0; |
| 507 | } | 507 | } |
| 508 | 508 | ||
| 509 | /* | 509 | /* |
| 510 | * merge two lists of backrefs and adjust counts accordingly | 510 | * merge two lists of backrefs and adjust counts accordingly |
| 511 | * | 511 | * |
| 512 | * mode = 1: merge identical keys, if key is set | 512 | * mode = 1: merge identical keys, if key is set |
| 513 | * FIXME: if we add more keys in __add_prelim_ref, we can merge more here. | 513 | * FIXME: if we add more keys in __add_prelim_ref, we can merge more here. |
| 514 | * additionally, we could even add a key range for the blocks we | 514 | * additionally, we could even add a key range for the blocks we |
| 515 | * looked into to merge even more (-> replace unresolved refs by those | 515 | * looked into to merge even more (-> replace unresolved refs by those |
| 516 | * having a parent). | 516 | * having a parent). |
| 517 | * mode = 2: merge identical parents | 517 | * mode = 2: merge identical parents |
| 518 | */ | 518 | */ |
| 519 | static void __merge_refs(struct list_head *head, int mode) | 519 | static void __merge_refs(struct list_head *head, int mode) |
| 520 | { | 520 | { |
| 521 | struct list_head *pos1; | 521 | struct list_head *pos1; |
| 522 | 522 | ||
| 523 | list_for_each(pos1, head) { | 523 | list_for_each(pos1, head) { |
| 524 | struct list_head *n2; | 524 | struct list_head *n2; |
| 525 | struct list_head *pos2; | 525 | struct list_head *pos2; |
| 526 | struct __prelim_ref *ref1; | 526 | struct __prelim_ref *ref1; |
| 527 | 527 | ||
| 528 | ref1 = list_entry(pos1, struct __prelim_ref, list); | 528 | ref1 = list_entry(pos1, struct __prelim_ref, list); |
| 529 | 529 | ||
| 530 | for (pos2 = pos1->next, n2 = pos2->next; pos2 != head; | 530 | for (pos2 = pos1->next, n2 = pos2->next; pos2 != head; |
| 531 | pos2 = n2, n2 = pos2->next) { | 531 | pos2 = n2, n2 = pos2->next) { |
| 532 | struct __prelim_ref *ref2; | 532 | struct __prelim_ref *ref2; |
| 533 | struct __prelim_ref *xchg; | 533 | struct __prelim_ref *xchg; |
| 534 | struct extent_inode_elem *eie; | 534 | struct extent_inode_elem *eie; |
| 535 | 535 | ||
| 536 | ref2 = list_entry(pos2, struct __prelim_ref, list); | 536 | ref2 = list_entry(pos2, struct __prelim_ref, list); |
| 537 | 537 | ||
| 538 | if (mode == 1) { | 538 | if (mode == 1) { |
| 539 | if (!ref_for_same_block(ref1, ref2)) | 539 | if (!ref_for_same_block(ref1, ref2)) |
| 540 | continue; | 540 | continue; |
| 541 | if (!ref1->parent && ref2->parent) { | 541 | if (!ref1->parent && ref2->parent) { |
| 542 | xchg = ref1; | 542 | xchg = ref1; |
| 543 | ref1 = ref2; | 543 | ref1 = ref2; |
| 544 | ref2 = xchg; | 544 | ref2 = xchg; |
| 545 | } | 545 | } |
| 546 | } else { | 546 | } else { |
| 547 | if (ref1->parent != ref2->parent) | 547 | if (ref1->parent != ref2->parent) |
| 548 | continue; | 548 | continue; |
| 549 | } | 549 | } |
| 550 | 550 | ||
| 551 | eie = ref1->inode_list; | 551 | eie = ref1->inode_list; |
| 552 | while (eie && eie->next) | 552 | while (eie && eie->next) |
| 553 | eie = eie->next; | 553 | eie = eie->next; |
| 554 | if (eie) | 554 | if (eie) |
| 555 | eie->next = ref2->inode_list; | 555 | eie->next = ref2->inode_list; |
| 556 | else | 556 | else |
| 557 | ref1->inode_list = ref2->inode_list; | 557 | ref1->inode_list = ref2->inode_list; |
| 558 | ref1->count += ref2->count; | 558 | ref1->count += ref2->count; |
| 559 | 559 | ||
| 560 | list_del(&ref2->list); | 560 | list_del(&ref2->list); |
| 561 | kmem_cache_free(btrfs_prelim_ref_cache, ref2); | 561 | kmem_cache_free(btrfs_prelim_ref_cache, ref2); |
| 562 | } | 562 | } |
| 563 | 563 | ||
| 564 | } | 564 | } |
| 565 | } | 565 | } |
| 566 | 566 | ||
| 567 | /* | 567 | /* |
| 568 | * add all currently queued delayed refs from this head whose seq nr is | 568 | * add all currently queued delayed refs from this head whose seq nr is |
| 569 | * smaller or equal that seq to the list | 569 | * smaller or equal that seq to the list |
| 570 | */ | 570 | */ |
| 571 | static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq, | 571 | static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq, |
| 572 | struct list_head *prefs, u64 *total_refs, | 572 | struct list_head *prefs, u64 *total_refs, |
| 573 | u64 inum) | 573 | u64 inum) |
| 574 | { | 574 | { |
| 575 | struct btrfs_delayed_extent_op *extent_op = head->extent_op; | 575 | struct btrfs_delayed_extent_op *extent_op = head->extent_op; |
| 576 | struct rb_node *n = &head->node.rb_node; | 576 | struct rb_node *n = &head->node.rb_node; |
| 577 | struct btrfs_key key; | 577 | struct btrfs_key key; |
| 578 | struct btrfs_key op_key = {0}; | 578 | struct btrfs_key op_key = {0}; |
| 579 | int sgn; | 579 | int sgn; |
| 580 | int ret = 0; | 580 | int ret = 0; |
| 581 | 581 | ||
| 582 | if (extent_op && extent_op->update_key) | 582 | if (extent_op && extent_op->update_key) |
| 583 | btrfs_disk_key_to_cpu(&op_key, &extent_op->key); | 583 | btrfs_disk_key_to_cpu(&op_key, &extent_op->key); |
| 584 | 584 | ||
| 585 | spin_lock(&head->lock); | 585 | spin_lock(&head->lock); |
| 586 | n = rb_first(&head->ref_root); | 586 | n = rb_first(&head->ref_root); |
| 587 | while (n) { | 587 | while (n) { |
| 588 | struct btrfs_delayed_ref_node *node; | 588 | struct btrfs_delayed_ref_node *node; |
| 589 | node = rb_entry(n, struct btrfs_delayed_ref_node, | 589 | node = rb_entry(n, struct btrfs_delayed_ref_node, |
| 590 | rb_node); | 590 | rb_node); |
| 591 | n = rb_next(n); | 591 | n = rb_next(n); |
| 592 | if (node->seq > seq) | 592 | if (node->seq > seq) |
| 593 | continue; | 593 | continue; |
| 594 | 594 | ||
| 595 | switch (node->action) { | 595 | switch (node->action) { |
| 596 | case BTRFS_ADD_DELAYED_EXTENT: | 596 | case BTRFS_ADD_DELAYED_EXTENT: |
| 597 | case BTRFS_UPDATE_DELAYED_HEAD: | 597 | case BTRFS_UPDATE_DELAYED_HEAD: |
| 598 | WARN_ON(1); | 598 | WARN_ON(1); |
| 599 | continue; | 599 | continue; |
| 600 | case BTRFS_ADD_DELAYED_REF: | 600 | case BTRFS_ADD_DELAYED_REF: |
| 601 | sgn = 1; | 601 | sgn = 1; |
| 602 | break; | 602 | break; |
| 603 | case BTRFS_DROP_DELAYED_REF: | 603 | case BTRFS_DROP_DELAYED_REF: |
| 604 | sgn = -1; | 604 | sgn = -1; |
| 605 | break; | 605 | break; |
| 606 | default: | 606 | default: |
| 607 | BUG_ON(1); | 607 | BUG_ON(1); |
| 608 | } | 608 | } |
| 609 | *total_refs += (node->ref_mod * sgn); | 609 | *total_refs += (node->ref_mod * sgn); |
| 610 | switch (node->type) { | 610 | switch (node->type) { |
| 611 | case BTRFS_TREE_BLOCK_REF_KEY: { | 611 | case BTRFS_TREE_BLOCK_REF_KEY: { |
| 612 | struct btrfs_delayed_tree_ref *ref; | 612 | struct btrfs_delayed_tree_ref *ref; |
| 613 | 613 | ||
| 614 | ref = btrfs_delayed_node_to_tree_ref(node); | 614 | ref = btrfs_delayed_node_to_tree_ref(node); |
| 615 | ret = __add_prelim_ref(prefs, ref->root, &op_key, | 615 | ret = __add_prelim_ref(prefs, ref->root, &op_key, |
| 616 | ref->level + 1, 0, node->bytenr, | 616 | ref->level + 1, 0, node->bytenr, |
| 617 | node->ref_mod * sgn, GFP_ATOMIC); | 617 | node->ref_mod * sgn, GFP_ATOMIC); |
| 618 | break; | 618 | break; |
| 619 | } | 619 | } |
| 620 | case BTRFS_SHARED_BLOCK_REF_KEY: { | 620 | case BTRFS_SHARED_BLOCK_REF_KEY: { |
| 621 | struct btrfs_delayed_tree_ref *ref; | 621 | struct btrfs_delayed_tree_ref *ref; |
| 622 | 622 | ||
| 623 | ref = btrfs_delayed_node_to_tree_ref(node); | 623 | ref = btrfs_delayed_node_to_tree_ref(node); |
| 624 | ret = __add_prelim_ref(prefs, ref->root, NULL, | 624 | ret = __add_prelim_ref(prefs, ref->root, NULL, |
| 625 | ref->level + 1, ref->parent, | 625 | ref->level + 1, ref->parent, |
| 626 | node->bytenr, | 626 | node->bytenr, |
| 627 | node->ref_mod * sgn, GFP_ATOMIC); | 627 | node->ref_mod * sgn, GFP_ATOMIC); |
| 628 | break; | 628 | break; |
| 629 | } | 629 | } |
| 630 | case BTRFS_EXTENT_DATA_REF_KEY: { | 630 | case BTRFS_EXTENT_DATA_REF_KEY: { |
| 631 | struct btrfs_delayed_data_ref *ref; | 631 | struct btrfs_delayed_data_ref *ref; |
| 632 | ref = btrfs_delayed_node_to_data_ref(node); | 632 | ref = btrfs_delayed_node_to_data_ref(node); |
| 633 | 633 | ||
| 634 | key.objectid = ref->objectid; | 634 | key.objectid = ref->objectid; |
| 635 | key.type = BTRFS_EXTENT_DATA_KEY; | 635 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 636 | key.offset = ref->offset; | 636 | key.offset = ref->offset; |
| 637 | 637 | ||
| 638 | /* | 638 | /* |
| 639 | * Found a inum that doesn't match our known inum, we | 639 | * Found a inum that doesn't match our known inum, we |
| 640 | * know it's shared. | 640 | * know it's shared. |
| 641 | */ | 641 | */ |
| 642 | if (inum && ref->objectid != inum) { | 642 | if (inum && ref->objectid != inum) { |
| 643 | ret = BACKREF_FOUND_SHARED; | 643 | ret = BACKREF_FOUND_SHARED; |
| 644 | break; | 644 | break; |
| 645 | } | 645 | } |
| 646 | 646 | ||
| 647 | ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0, | 647 | ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0, |
| 648 | node->bytenr, | 648 | node->bytenr, |
| 649 | node->ref_mod * sgn, GFP_ATOMIC); | 649 | node->ref_mod * sgn, GFP_ATOMIC); |
| 650 | break; | 650 | break; |
| 651 | } | 651 | } |
| 652 | case BTRFS_SHARED_DATA_REF_KEY: { | 652 | case BTRFS_SHARED_DATA_REF_KEY: { |
| 653 | struct btrfs_delayed_data_ref *ref; | 653 | struct btrfs_delayed_data_ref *ref; |
| 654 | 654 | ||
| 655 | ref = btrfs_delayed_node_to_data_ref(node); | 655 | ref = btrfs_delayed_node_to_data_ref(node); |
| 656 | 656 | ||
| 657 | key.objectid = ref->objectid; | 657 | key.objectid = ref->objectid; |
| 658 | key.type = BTRFS_EXTENT_DATA_KEY; | 658 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 659 | key.offset = ref->offset; | 659 | key.offset = ref->offset; |
| 660 | ret = __add_prelim_ref(prefs, ref->root, &key, 0, | 660 | ret = __add_prelim_ref(prefs, ref->root, &key, 0, |
| 661 | ref->parent, node->bytenr, | 661 | ref->parent, node->bytenr, |
| 662 | node->ref_mod * sgn, GFP_ATOMIC); | 662 | node->ref_mod * sgn, GFP_ATOMIC); |
| 663 | break; | 663 | break; |
| 664 | } | 664 | } |
| 665 | default: | 665 | default: |
| 666 | WARN_ON(1); | 666 | WARN_ON(1); |
| 667 | } | 667 | } |
| 668 | if (ret) | 668 | if (ret) |
| 669 | break; | 669 | break; |
| 670 | } | 670 | } |
| 671 | spin_unlock(&head->lock); | 671 | spin_unlock(&head->lock); |
| 672 | return ret; | 672 | return ret; |
| 673 | } | 673 | } |
| 674 | 674 | ||
| 675 | /* | 675 | /* |
| 676 | * add all inline backrefs for bytenr to the list | 676 | * add all inline backrefs for bytenr to the list |
| 677 | */ | 677 | */ |
| 678 | static int __add_inline_refs(struct btrfs_fs_info *fs_info, | 678 | static int __add_inline_refs(struct btrfs_fs_info *fs_info, |
| 679 | struct btrfs_path *path, u64 bytenr, | 679 | struct btrfs_path *path, u64 bytenr, |
| 680 | int *info_level, struct list_head *prefs, | 680 | int *info_level, struct list_head *prefs, |
| 681 | u64 *total_refs, u64 inum) | 681 | u64 *total_refs, u64 inum) |
| 682 | { | 682 | { |
| 683 | int ret = 0; | 683 | int ret = 0; |
| 684 | int slot; | 684 | int slot; |
| 685 | struct extent_buffer *leaf; | 685 | struct extent_buffer *leaf; |
| 686 | struct btrfs_key key; | 686 | struct btrfs_key key; |
| 687 | struct btrfs_key found_key; | 687 | struct btrfs_key found_key; |
| 688 | unsigned long ptr; | 688 | unsigned long ptr; |
| 689 | unsigned long end; | 689 | unsigned long end; |
| 690 | struct btrfs_extent_item *ei; | 690 | struct btrfs_extent_item *ei; |
| 691 | u64 flags; | 691 | u64 flags; |
| 692 | u64 item_size; | 692 | u64 item_size; |
| 693 | 693 | ||
| 694 | /* | 694 | /* |
| 695 | * enumerate all inline refs | 695 | * enumerate all inline refs |
| 696 | */ | 696 | */ |
| 697 | leaf = path->nodes[0]; | 697 | leaf = path->nodes[0]; |
| 698 | slot = path->slots[0]; | 698 | slot = path->slots[0]; |
| 699 | 699 | ||
| 700 | item_size = btrfs_item_size_nr(leaf, slot); | 700 | item_size = btrfs_item_size_nr(leaf, slot); |
| 701 | BUG_ON(item_size < sizeof(*ei)); | 701 | BUG_ON(item_size < sizeof(*ei)); |
| 702 | 702 | ||
| 703 | ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item); | 703 | ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item); |
| 704 | flags = btrfs_extent_flags(leaf, ei); | 704 | flags = btrfs_extent_flags(leaf, ei); |
| 705 | *total_refs += btrfs_extent_refs(leaf, ei); | 705 | *total_refs += btrfs_extent_refs(leaf, ei); |
| 706 | btrfs_item_key_to_cpu(leaf, &found_key, slot); | 706 | btrfs_item_key_to_cpu(leaf, &found_key, slot); |
| 707 | 707 | ||
| 708 | ptr = (unsigned long)(ei + 1); | 708 | ptr = (unsigned long)(ei + 1); |
| 709 | end = (unsigned long)ei + item_size; | 709 | end = (unsigned long)ei + item_size; |
| 710 | 710 | ||
| 711 | if (found_key.type == BTRFS_EXTENT_ITEM_KEY && | 711 | if (found_key.type == BTRFS_EXTENT_ITEM_KEY && |
| 712 | flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { | 712 | flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { |
| 713 | struct btrfs_tree_block_info *info; | 713 | struct btrfs_tree_block_info *info; |
| 714 | 714 | ||
| 715 | info = (struct btrfs_tree_block_info *)ptr; | 715 | info = (struct btrfs_tree_block_info *)ptr; |
| 716 | *info_level = btrfs_tree_block_level(leaf, info); | 716 | *info_level = btrfs_tree_block_level(leaf, info); |
| 717 | ptr += sizeof(struct btrfs_tree_block_info); | 717 | ptr += sizeof(struct btrfs_tree_block_info); |
| 718 | BUG_ON(ptr > end); | 718 | BUG_ON(ptr > end); |
| 719 | } else if (found_key.type == BTRFS_METADATA_ITEM_KEY) { | 719 | } else if (found_key.type == BTRFS_METADATA_ITEM_KEY) { |
| 720 | *info_level = found_key.offset; | 720 | *info_level = found_key.offset; |
| 721 | } else { | 721 | } else { |
| 722 | BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA)); | 722 | BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA)); |
| 723 | } | 723 | } |
| 724 | 724 | ||
| 725 | while (ptr < end) { | 725 | while (ptr < end) { |
| 726 | struct btrfs_extent_inline_ref *iref; | 726 | struct btrfs_extent_inline_ref *iref; |
| 727 | u64 offset; | 727 | u64 offset; |
| 728 | int type; | 728 | int type; |
| 729 | 729 | ||
| 730 | iref = (struct btrfs_extent_inline_ref *)ptr; | 730 | iref = (struct btrfs_extent_inline_ref *)ptr; |
| 731 | type = btrfs_extent_inline_ref_type(leaf, iref); | 731 | type = btrfs_extent_inline_ref_type(leaf, iref); |
| 732 | offset = btrfs_extent_inline_ref_offset(leaf, iref); | 732 | offset = btrfs_extent_inline_ref_offset(leaf, iref); |
| 733 | 733 | ||
| 734 | switch (type) { | 734 | switch (type) { |
| 735 | case BTRFS_SHARED_BLOCK_REF_KEY: | 735 | case BTRFS_SHARED_BLOCK_REF_KEY: |
| 736 | ret = __add_prelim_ref(prefs, 0, NULL, | 736 | ret = __add_prelim_ref(prefs, 0, NULL, |
| 737 | *info_level + 1, offset, | 737 | *info_level + 1, offset, |
| 738 | bytenr, 1, GFP_NOFS); | 738 | bytenr, 1, GFP_NOFS); |
| 739 | break; | 739 | break; |
| 740 | case BTRFS_SHARED_DATA_REF_KEY: { | 740 | case BTRFS_SHARED_DATA_REF_KEY: { |
| 741 | struct btrfs_shared_data_ref *sdref; | 741 | struct btrfs_shared_data_ref *sdref; |
| 742 | int count; | 742 | int count; |
| 743 | 743 | ||
| 744 | sdref = (struct btrfs_shared_data_ref *)(iref + 1); | 744 | sdref = (struct btrfs_shared_data_ref *)(iref + 1); |
| 745 | count = btrfs_shared_data_ref_count(leaf, sdref); | 745 | count = btrfs_shared_data_ref_count(leaf, sdref); |
| 746 | ret = __add_prelim_ref(prefs, 0, NULL, 0, offset, | 746 | ret = __add_prelim_ref(prefs, 0, NULL, 0, offset, |
| 747 | bytenr, count, GFP_NOFS); | 747 | bytenr, count, GFP_NOFS); |
| 748 | break; | 748 | break; |
| 749 | } | 749 | } |
| 750 | case BTRFS_TREE_BLOCK_REF_KEY: | 750 | case BTRFS_TREE_BLOCK_REF_KEY: |
| 751 | ret = __add_prelim_ref(prefs, offset, NULL, | 751 | ret = __add_prelim_ref(prefs, offset, NULL, |
| 752 | *info_level + 1, 0, | 752 | *info_level + 1, 0, |
| 753 | bytenr, 1, GFP_NOFS); | 753 | bytenr, 1, GFP_NOFS); |
| 754 | break; | 754 | break; |
| 755 | case BTRFS_EXTENT_DATA_REF_KEY: { | 755 | case BTRFS_EXTENT_DATA_REF_KEY: { |
| 756 | struct btrfs_extent_data_ref *dref; | 756 | struct btrfs_extent_data_ref *dref; |
| 757 | int count; | 757 | int count; |
| 758 | u64 root; | 758 | u64 root; |
| 759 | 759 | ||
| 760 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); | 760 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 761 | count = btrfs_extent_data_ref_count(leaf, dref); | 761 | count = btrfs_extent_data_ref_count(leaf, dref); |
| 762 | key.objectid = btrfs_extent_data_ref_objectid(leaf, | 762 | key.objectid = btrfs_extent_data_ref_objectid(leaf, |
| 763 | dref); | 763 | dref); |
| 764 | key.type = BTRFS_EXTENT_DATA_KEY; | 764 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 765 | key.offset = btrfs_extent_data_ref_offset(leaf, dref); | 765 | key.offset = btrfs_extent_data_ref_offset(leaf, dref); |
| 766 | 766 | ||
| 767 | if (inum && key.objectid != inum) { | 767 | if (inum && key.objectid != inum) { |
| 768 | ret = BACKREF_FOUND_SHARED; | 768 | ret = BACKREF_FOUND_SHARED; |
| 769 | break; | 769 | break; |
| 770 | } | 770 | } |
| 771 | 771 | ||
| 772 | root = btrfs_extent_data_ref_root(leaf, dref); | 772 | root = btrfs_extent_data_ref_root(leaf, dref); |
| 773 | ret = __add_prelim_ref(prefs, root, &key, 0, 0, | 773 | ret = __add_prelim_ref(prefs, root, &key, 0, 0, |
| 774 | bytenr, count, GFP_NOFS); | 774 | bytenr, count, GFP_NOFS); |
| 775 | break; | 775 | break; |
| 776 | } | 776 | } |
| 777 | default: | 777 | default: |
| 778 | WARN_ON(1); | 778 | WARN_ON(1); |
| 779 | } | 779 | } |
| 780 | if (ret) | 780 | if (ret) |
| 781 | return ret; | 781 | return ret; |
| 782 | ptr += btrfs_extent_inline_ref_size(type); | 782 | ptr += btrfs_extent_inline_ref_size(type); |
| 783 | } | 783 | } |
| 784 | 784 | ||
| 785 | return 0; | 785 | return 0; |
| 786 | } | 786 | } |
| 787 | 787 | ||
| 788 | /* | 788 | /* |
| 789 | * add all non-inline backrefs for bytenr to the list | 789 | * add all non-inline backrefs for bytenr to the list |
| 790 | */ | 790 | */ |
| 791 | static int __add_keyed_refs(struct btrfs_fs_info *fs_info, | 791 | static int __add_keyed_refs(struct btrfs_fs_info *fs_info, |
| 792 | struct btrfs_path *path, u64 bytenr, | 792 | struct btrfs_path *path, u64 bytenr, |
| 793 | int info_level, struct list_head *prefs, u64 inum) | 793 | int info_level, struct list_head *prefs, u64 inum) |
| 794 | { | 794 | { |
| 795 | struct btrfs_root *extent_root = fs_info->extent_root; | 795 | struct btrfs_root *extent_root = fs_info->extent_root; |
| 796 | int ret; | 796 | int ret; |
| 797 | int slot; | 797 | int slot; |
| 798 | struct extent_buffer *leaf; | 798 | struct extent_buffer *leaf; |
| 799 | struct btrfs_key key; | 799 | struct btrfs_key key; |
| 800 | 800 | ||
| 801 | while (1) { | 801 | while (1) { |
| 802 | ret = btrfs_next_item(extent_root, path); | 802 | ret = btrfs_next_item(extent_root, path); |
| 803 | if (ret < 0) | 803 | if (ret < 0) |
| 804 | break; | 804 | break; |
| 805 | if (ret) { | 805 | if (ret) { |
| 806 | ret = 0; | 806 | ret = 0; |
| 807 | break; | 807 | break; |
| 808 | } | 808 | } |
| 809 | 809 | ||
| 810 | slot = path->slots[0]; | 810 | slot = path->slots[0]; |
| 811 | leaf = path->nodes[0]; | 811 | leaf = path->nodes[0]; |
| 812 | btrfs_item_key_to_cpu(leaf, &key, slot); | 812 | btrfs_item_key_to_cpu(leaf, &key, slot); |
| 813 | 813 | ||
| 814 | if (key.objectid != bytenr) | 814 | if (key.objectid != bytenr) |
| 815 | break; | 815 | break; |
| 816 | if (key.type < BTRFS_TREE_BLOCK_REF_KEY) | 816 | if (key.type < BTRFS_TREE_BLOCK_REF_KEY) |
| 817 | continue; | 817 | continue; |
| 818 | if (key.type > BTRFS_SHARED_DATA_REF_KEY) | 818 | if (key.type > BTRFS_SHARED_DATA_REF_KEY) |
| 819 | break; | 819 | break; |
| 820 | 820 | ||
| 821 | switch (key.type) { | 821 | switch (key.type) { |
| 822 | case BTRFS_SHARED_BLOCK_REF_KEY: | 822 | case BTRFS_SHARED_BLOCK_REF_KEY: |
| 823 | ret = __add_prelim_ref(prefs, 0, NULL, | 823 | ret = __add_prelim_ref(prefs, 0, NULL, |
| 824 | info_level + 1, key.offset, | 824 | info_level + 1, key.offset, |
| 825 | bytenr, 1, GFP_NOFS); | 825 | bytenr, 1, GFP_NOFS); |
| 826 | break; | 826 | break; |
| 827 | case BTRFS_SHARED_DATA_REF_KEY: { | 827 | case BTRFS_SHARED_DATA_REF_KEY: { |
| 828 | struct btrfs_shared_data_ref *sdref; | 828 | struct btrfs_shared_data_ref *sdref; |
| 829 | int count; | 829 | int count; |
| 830 | 830 | ||
| 831 | sdref = btrfs_item_ptr(leaf, slot, | 831 | sdref = btrfs_item_ptr(leaf, slot, |
| 832 | struct btrfs_shared_data_ref); | 832 | struct btrfs_shared_data_ref); |
| 833 | count = btrfs_shared_data_ref_count(leaf, sdref); | 833 | count = btrfs_shared_data_ref_count(leaf, sdref); |
| 834 | ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset, | 834 | ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset, |
| 835 | bytenr, count, GFP_NOFS); | 835 | bytenr, count, GFP_NOFS); |
| 836 | break; | 836 | break; |
| 837 | } | 837 | } |
| 838 | case BTRFS_TREE_BLOCK_REF_KEY: | 838 | case BTRFS_TREE_BLOCK_REF_KEY: |
| 839 | ret = __add_prelim_ref(prefs, key.offset, NULL, | 839 | ret = __add_prelim_ref(prefs, key.offset, NULL, |
| 840 | info_level + 1, 0, | 840 | info_level + 1, 0, |
| 841 | bytenr, 1, GFP_NOFS); | 841 | bytenr, 1, GFP_NOFS); |
| 842 | break; | 842 | break; |
| 843 | case BTRFS_EXTENT_DATA_REF_KEY: { | 843 | case BTRFS_EXTENT_DATA_REF_KEY: { |
| 844 | struct btrfs_extent_data_ref *dref; | 844 | struct btrfs_extent_data_ref *dref; |
| 845 | int count; | 845 | int count; |
| 846 | u64 root; | 846 | u64 root; |
| 847 | 847 | ||
| 848 | dref = btrfs_item_ptr(leaf, slot, | 848 | dref = btrfs_item_ptr(leaf, slot, |
| 849 | struct btrfs_extent_data_ref); | 849 | struct btrfs_extent_data_ref); |
| 850 | count = btrfs_extent_data_ref_count(leaf, dref); | 850 | count = btrfs_extent_data_ref_count(leaf, dref); |
| 851 | key.objectid = btrfs_extent_data_ref_objectid(leaf, | 851 | key.objectid = btrfs_extent_data_ref_objectid(leaf, |
| 852 | dref); | 852 | dref); |
| 853 | key.type = BTRFS_EXTENT_DATA_KEY; | 853 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 854 | key.offset = btrfs_extent_data_ref_offset(leaf, dref); | 854 | key.offset = btrfs_extent_data_ref_offset(leaf, dref); |
| 855 | 855 | ||
| 856 | if (inum && key.objectid != inum) { | 856 | if (inum && key.objectid != inum) { |
| 857 | ret = BACKREF_FOUND_SHARED; | 857 | ret = BACKREF_FOUND_SHARED; |
| 858 | break; | 858 | break; |
| 859 | } | 859 | } |
| 860 | 860 | ||
| 861 | root = btrfs_extent_data_ref_root(leaf, dref); | 861 | root = btrfs_extent_data_ref_root(leaf, dref); |
| 862 | ret = __add_prelim_ref(prefs, root, &key, 0, 0, | 862 | ret = __add_prelim_ref(prefs, root, &key, 0, 0, |
| 863 | bytenr, count, GFP_NOFS); | 863 | bytenr, count, GFP_NOFS); |
| 864 | break; | 864 | break; |
| 865 | } | 865 | } |
| 866 | default: | 866 | default: |
| 867 | WARN_ON(1); | 867 | WARN_ON(1); |
| 868 | } | 868 | } |
| 869 | if (ret) | 869 | if (ret) |
| 870 | return ret; | 870 | return ret; |
| 871 | 871 | ||
| 872 | } | 872 | } |
| 873 | 873 | ||
| 874 | return ret; | 874 | return ret; |
| 875 | } | 875 | } |
| 876 | 876 | ||
| 877 | /* | 877 | /* |
| 878 | * this adds all existing backrefs (inline backrefs, backrefs and delayed | 878 | * this adds all existing backrefs (inline backrefs, backrefs and delayed |
| 879 | * refs) for the given bytenr to the refs list, merges duplicates and resolves | 879 | * refs) for the given bytenr to the refs list, merges duplicates and resolves |
| 880 | * indirect refs to their parent bytenr. | 880 | * indirect refs to their parent bytenr. |
| 881 | * When roots are found, they're added to the roots list | 881 | * When roots are found, they're added to the roots list |
| 882 | * | 882 | * |
| 883 | * FIXME some caching might speed things up | 883 | * FIXME some caching might speed things up |
| 884 | */ | 884 | */ |
| 885 | static int find_parent_nodes(struct btrfs_trans_handle *trans, | 885 | static int find_parent_nodes(struct btrfs_trans_handle *trans, |
| 886 | struct btrfs_fs_info *fs_info, u64 bytenr, | 886 | struct btrfs_fs_info *fs_info, u64 bytenr, |
| 887 | u64 time_seq, struct ulist *refs, | 887 | u64 time_seq, struct ulist *refs, |
| 888 | struct ulist *roots, const u64 *extent_item_pos, | 888 | struct ulist *roots, const u64 *extent_item_pos, |
| 889 | u64 root_objectid, u64 inum) | 889 | u64 root_objectid, u64 inum) |
| 890 | { | 890 | { |
| 891 | struct btrfs_key key; | 891 | struct btrfs_key key; |
| 892 | struct btrfs_path *path; | 892 | struct btrfs_path *path; |
| 893 | struct btrfs_delayed_ref_root *delayed_refs = NULL; | 893 | struct btrfs_delayed_ref_root *delayed_refs = NULL; |
| 894 | struct btrfs_delayed_ref_head *head; | 894 | struct btrfs_delayed_ref_head *head; |
| 895 | int info_level = 0; | 895 | int info_level = 0; |
| 896 | int ret; | 896 | int ret; |
| 897 | struct list_head prefs_delayed; | 897 | struct list_head prefs_delayed; |
| 898 | struct list_head prefs; | 898 | struct list_head prefs; |
| 899 | struct __prelim_ref *ref; | 899 | struct __prelim_ref *ref; |
| 900 | struct extent_inode_elem *eie = NULL; | 900 | struct extent_inode_elem *eie = NULL; |
| 901 | u64 total_refs = 0; | 901 | u64 total_refs = 0; |
| 902 | 902 | ||
| 903 | INIT_LIST_HEAD(&prefs); | 903 | INIT_LIST_HEAD(&prefs); |
| 904 | INIT_LIST_HEAD(&prefs_delayed); | 904 | INIT_LIST_HEAD(&prefs_delayed); |
| 905 | 905 | ||
| 906 | key.objectid = bytenr; | 906 | key.objectid = bytenr; |
| 907 | key.offset = (u64)-1; | 907 | key.offset = (u64)-1; |
| 908 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) | 908 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) |
| 909 | key.type = BTRFS_METADATA_ITEM_KEY; | 909 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 910 | else | 910 | else |
| 911 | key.type = BTRFS_EXTENT_ITEM_KEY; | 911 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 912 | 912 | ||
| 913 | path = btrfs_alloc_path(); | 913 | path = btrfs_alloc_path(); |
| 914 | if (!path) | 914 | if (!path) |
| 915 | return -ENOMEM; | 915 | return -ENOMEM; |
| 916 | if (!trans) { | 916 | if (!trans) { |
| 917 | path->search_commit_root = 1; | 917 | path->search_commit_root = 1; |
| 918 | path->skip_locking = 1; | 918 | path->skip_locking = 1; |
| 919 | } | 919 | } |
| 920 | 920 | ||
| 921 | /* | 921 | /* |
| 922 | * grab both a lock on the path and a lock on the delayed ref head. | 922 | * grab both a lock on the path and a lock on the delayed ref head. |
| 923 | * We need both to get a consistent picture of how the refs look | 923 | * We need both to get a consistent picture of how the refs look |
| 924 | * at a specified point in time | 924 | * at a specified point in time |
| 925 | */ | 925 | */ |
| 926 | again: | 926 | again: |
| 927 | head = NULL; | 927 | head = NULL; |
| 928 | 928 | ||
| 929 | ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0); | 929 | ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0); |
| 930 | if (ret < 0) | 930 | if (ret < 0) |
| 931 | goto out; | 931 | goto out; |
| 932 | BUG_ON(ret == 0); | 932 | BUG_ON(ret == 0); |
| 933 | 933 | ||
| 934 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS | 934 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS |
| 935 | if (trans && likely(trans->type != __TRANS_DUMMY)) { | 935 | if (trans && likely(trans->type != __TRANS_DUMMY)) { |
| 936 | #else | 936 | #else |
| 937 | if (trans) { | 937 | if (trans) { |
| 938 | #endif | 938 | #endif |
| 939 | /* | 939 | /* |
| 940 | * look if there are updates for this ref queued and lock the | 940 | * look if there are updates for this ref queued and lock the |
| 941 | * head | 941 | * head |
| 942 | */ | 942 | */ |
| 943 | delayed_refs = &trans->transaction->delayed_refs; | 943 | delayed_refs = &trans->transaction->delayed_refs; |
| 944 | spin_lock(&delayed_refs->lock); | 944 | spin_lock(&delayed_refs->lock); |
| 945 | head = btrfs_find_delayed_ref_head(trans, bytenr); | 945 | head = btrfs_find_delayed_ref_head(trans, bytenr); |
| 946 | if (head) { | 946 | if (head) { |
| 947 | if (!mutex_trylock(&head->mutex)) { | 947 | if (!mutex_trylock(&head->mutex)) { |
| 948 | atomic_inc(&head->node.refs); | 948 | atomic_inc(&head->node.refs); |
| 949 | spin_unlock(&delayed_refs->lock); | 949 | spin_unlock(&delayed_refs->lock); |
| 950 | 950 | ||
| 951 | btrfs_release_path(path); | 951 | btrfs_release_path(path); |
| 952 | 952 | ||
| 953 | /* | 953 | /* |
| 954 | * Mutex was contended, block until it's | 954 | * Mutex was contended, block until it's |
| 955 | * released and try again | 955 | * released and try again |
| 956 | */ | 956 | */ |
| 957 | mutex_lock(&head->mutex); | 957 | mutex_lock(&head->mutex); |
| 958 | mutex_unlock(&head->mutex); | 958 | mutex_unlock(&head->mutex); |
| 959 | btrfs_put_delayed_ref(&head->node); | 959 | btrfs_put_delayed_ref(&head->node); |
| 960 | goto again; | 960 | goto again; |
| 961 | } | 961 | } |
| 962 | spin_unlock(&delayed_refs->lock); | 962 | spin_unlock(&delayed_refs->lock); |
| 963 | ret = __add_delayed_refs(head, time_seq, | 963 | ret = __add_delayed_refs(head, time_seq, |
| 964 | &prefs_delayed, &total_refs, | 964 | &prefs_delayed, &total_refs, |
| 965 | inum); | 965 | inum); |
| 966 | mutex_unlock(&head->mutex); | 966 | mutex_unlock(&head->mutex); |
| 967 | if (ret) | 967 | if (ret) |
| 968 | goto out; | 968 | goto out; |
| 969 | } else { | 969 | } else { |
| 970 | spin_unlock(&delayed_refs->lock); | 970 | spin_unlock(&delayed_refs->lock); |
| 971 | } | 971 | } |
| 972 | } | 972 | } |
| 973 | 973 | ||
| 974 | if (path->slots[0]) { | 974 | if (path->slots[0]) { |
| 975 | struct extent_buffer *leaf; | 975 | struct extent_buffer *leaf; |
| 976 | int slot; | 976 | int slot; |
| 977 | 977 | ||
| 978 | path->slots[0]--; | 978 | path->slots[0]--; |
| 979 | leaf = path->nodes[0]; | 979 | leaf = path->nodes[0]; |
| 980 | slot = path->slots[0]; | 980 | slot = path->slots[0]; |
| 981 | btrfs_item_key_to_cpu(leaf, &key, slot); | 981 | btrfs_item_key_to_cpu(leaf, &key, slot); |
| 982 | if (key.objectid == bytenr && | 982 | if (key.objectid == bytenr && |
| 983 | (key.type == BTRFS_EXTENT_ITEM_KEY || | 983 | (key.type == BTRFS_EXTENT_ITEM_KEY || |
| 984 | key.type == BTRFS_METADATA_ITEM_KEY)) { | 984 | key.type == BTRFS_METADATA_ITEM_KEY)) { |
| 985 | ret = __add_inline_refs(fs_info, path, bytenr, | 985 | ret = __add_inline_refs(fs_info, path, bytenr, |
| 986 | &info_level, &prefs, | 986 | &info_level, &prefs, |
| 987 | &total_refs, inum); | 987 | &total_refs, inum); |
| 988 | if (ret) | 988 | if (ret) |
| 989 | goto out; | 989 | goto out; |
| 990 | ret = __add_keyed_refs(fs_info, path, bytenr, | 990 | ret = __add_keyed_refs(fs_info, path, bytenr, |
| 991 | info_level, &prefs, inum); | 991 | info_level, &prefs, inum); |
| 992 | if (ret) | 992 | if (ret) |
| 993 | goto out; | 993 | goto out; |
| 994 | } | 994 | } |
| 995 | } | 995 | } |
| 996 | btrfs_release_path(path); | 996 | btrfs_release_path(path); |
| 997 | 997 | ||
| 998 | list_splice_init(&prefs_delayed, &prefs); | 998 | list_splice_init(&prefs_delayed, &prefs); |
| 999 | 999 | ||
| 1000 | ret = __add_missing_keys(fs_info, &prefs); | 1000 | ret = __add_missing_keys(fs_info, &prefs); |
| 1001 | if (ret) | 1001 | if (ret) |
| 1002 | goto out; | 1002 | goto out; |
| 1003 | 1003 | ||
| 1004 | __merge_refs(&prefs, 1); | 1004 | __merge_refs(&prefs, 1); |
| 1005 | 1005 | ||
| 1006 | ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs, | 1006 | ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs, |
| 1007 | extent_item_pos, total_refs, | 1007 | extent_item_pos, total_refs, |
| 1008 | root_objectid); | 1008 | root_objectid); |
| 1009 | if (ret) | 1009 | if (ret) |
| 1010 | goto out; | 1010 | goto out; |
| 1011 | 1011 | ||
| 1012 | __merge_refs(&prefs, 2); | 1012 | __merge_refs(&prefs, 2); |
| 1013 | 1013 | ||
| 1014 | while (!list_empty(&prefs)) { | 1014 | while (!list_empty(&prefs)) { |
| 1015 | ref = list_first_entry(&prefs, struct __prelim_ref, list); | 1015 | ref = list_first_entry(&prefs, struct __prelim_ref, list); |
| 1016 | WARN_ON(ref->count < 0); | 1016 | WARN_ON(ref->count < 0); |
| 1017 | if (roots && ref->count && ref->root_id && ref->parent == 0) { | 1017 | if (roots && ref->count && ref->root_id && ref->parent == 0) { |
| 1018 | if (root_objectid && ref->root_id != root_objectid) { | 1018 | if (root_objectid && ref->root_id != root_objectid) { |
| 1019 | ret = BACKREF_FOUND_SHARED; | 1019 | ret = BACKREF_FOUND_SHARED; |
| 1020 | goto out; | 1020 | goto out; |
| 1021 | } | 1021 | } |
| 1022 | 1022 | ||
| 1023 | /* no parent == root of tree */ | 1023 | /* no parent == root of tree */ |
| 1024 | ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS); | 1024 | ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS); |
| 1025 | if (ret < 0) | 1025 | if (ret < 0) |
| 1026 | goto out; | 1026 | goto out; |
| 1027 | } | 1027 | } |
| 1028 | if (ref->count && ref->parent) { | 1028 | if (ref->count && ref->parent) { |
| 1029 | if (extent_item_pos && !ref->inode_list && | 1029 | if (extent_item_pos && !ref->inode_list && |
| 1030 | ref->level == 0) { | 1030 | ref->level == 0) { |
| 1031 | struct extent_buffer *eb; | 1031 | struct extent_buffer *eb; |
| 1032 | 1032 | ||
| 1033 | eb = read_tree_block(fs_info->extent_root, | 1033 | eb = read_tree_block(fs_info->extent_root, |
| 1034 | ref->parent, 0); | 1034 | ref->parent, 0); |
| 1035 | if (!eb || !extent_buffer_uptodate(eb)) { | 1035 | if (!eb || !extent_buffer_uptodate(eb)) { |
| 1036 | free_extent_buffer(eb); | 1036 | free_extent_buffer(eb); |
| 1037 | ret = -EIO; | 1037 | ret = -EIO; |
| 1038 | goto out; | 1038 | goto out; |
| 1039 | } | 1039 | } |
| 1040 | btrfs_tree_read_lock(eb); | 1040 | btrfs_tree_read_lock(eb); |
| 1041 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); | 1041 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); |
| 1042 | ret = find_extent_in_eb(eb, bytenr, | 1042 | ret = find_extent_in_eb(eb, bytenr, |
| 1043 | *extent_item_pos, &eie); | 1043 | *extent_item_pos, &eie); |
| 1044 | btrfs_tree_read_unlock_blocking(eb); | 1044 | btrfs_tree_read_unlock_blocking(eb); |
| 1045 | free_extent_buffer(eb); | 1045 | free_extent_buffer(eb); |
| 1046 | if (ret < 0) | 1046 | if (ret < 0) |
| 1047 | goto out; | 1047 | goto out; |
| 1048 | ref->inode_list = eie; | 1048 | ref->inode_list = eie; |
| 1049 | } | 1049 | } |
| 1050 | ret = ulist_add_merge_ptr(refs, ref->parent, | 1050 | ret = ulist_add_merge_ptr(refs, ref->parent, |
| 1051 | ref->inode_list, | 1051 | ref->inode_list, |
| 1052 | (void **)&eie, GFP_NOFS); | 1052 | (void **)&eie, GFP_NOFS); |
| 1053 | if (ret < 0) | 1053 | if (ret < 0) |
| 1054 | goto out; | 1054 | goto out; |
| 1055 | if (!ret && extent_item_pos) { | 1055 | if (!ret && extent_item_pos) { |
| 1056 | /* | 1056 | /* |
| 1057 | * we've recorded that parent, so we must extend | 1057 | * we've recorded that parent, so we must extend |
| 1058 | * its inode list here | 1058 | * its inode list here |
| 1059 | */ | 1059 | */ |
| 1060 | BUG_ON(!eie); | 1060 | BUG_ON(!eie); |
| 1061 | while (eie->next) | 1061 | while (eie->next) |
| 1062 | eie = eie->next; | 1062 | eie = eie->next; |
| 1063 | eie->next = ref->inode_list; | 1063 | eie->next = ref->inode_list; |
| 1064 | } | 1064 | } |
| 1065 | eie = NULL; | 1065 | eie = NULL; |
| 1066 | } | 1066 | } |
| 1067 | list_del(&ref->list); | 1067 | list_del(&ref->list); |
| 1068 | kmem_cache_free(btrfs_prelim_ref_cache, ref); | 1068 | kmem_cache_free(btrfs_prelim_ref_cache, ref); |
| 1069 | } | 1069 | } |
| 1070 | 1070 | ||
| 1071 | out: | 1071 | out: |
| 1072 | btrfs_free_path(path); | 1072 | btrfs_free_path(path); |
| 1073 | while (!list_empty(&prefs)) { | 1073 | while (!list_empty(&prefs)) { |
| 1074 | ref = list_first_entry(&prefs, struct __prelim_ref, list); | 1074 | ref = list_first_entry(&prefs, struct __prelim_ref, list); |
| 1075 | list_del(&ref->list); | 1075 | list_del(&ref->list); |
| 1076 | kmem_cache_free(btrfs_prelim_ref_cache, ref); | 1076 | kmem_cache_free(btrfs_prelim_ref_cache, ref); |
| 1077 | } | 1077 | } |
| 1078 | while (!list_empty(&prefs_delayed)) { | 1078 | while (!list_empty(&prefs_delayed)) { |
| 1079 | ref = list_first_entry(&prefs_delayed, struct __prelim_ref, | 1079 | ref = list_first_entry(&prefs_delayed, struct __prelim_ref, |
| 1080 | list); | 1080 | list); |
| 1081 | list_del(&ref->list); | 1081 | list_del(&ref->list); |
| 1082 | kmem_cache_free(btrfs_prelim_ref_cache, ref); | 1082 | kmem_cache_free(btrfs_prelim_ref_cache, ref); |
| 1083 | } | 1083 | } |
| 1084 | if (ret < 0) | 1084 | if (ret < 0) |
| 1085 | free_inode_elem_list(eie); | 1085 | free_inode_elem_list(eie); |
| 1086 | return ret; | 1086 | return ret; |
| 1087 | } | 1087 | } |
| 1088 | 1088 | ||
| 1089 | static void free_leaf_list(struct ulist *blocks) | 1089 | static void free_leaf_list(struct ulist *blocks) |
| 1090 | { | 1090 | { |
| 1091 | struct ulist_node *node = NULL; | 1091 | struct ulist_node *node = NULL; |
| 1092 | struct extent_inode_elem *eie; | 1092 | struct extent_inode_elem *eie; |
| 1093 | struct ulist_iterator uiter; | 1093 | struct ulist_iterator uiter; |
| 1094 | 1094 | ||
| 1095 | ULIST_ITER_INIT(&uiter); | 1095 | ULIST_ITER_INIT(&uiter); |
| 1096 | while ((node = ulist_next(blocks, &uiter))) { | 1096 | while ((node = ulist_next(blocks, &uiter))) { |
| 1097 | if (!node->aux) | 1097 | if (!node->aux) |
| 1098 | continue; | 1098 | continue; |
| 1099 | eie = (struct extent_inode_elem *)(uintptr_t)node->aux; | 1099 | eie = (struct extent_inode_elem *)(uintptr_t)node->aux; |
| 1100 | free_inode_elem_list(eie); | 1100 | free_inode_elem_list(eie); |
| 1101 | node->aux = 0; | 1101 | node->aux = 0; |
| 1102 | } | 1102 | } |
| 1103 | 1103 | ||
| 1104 | ulist_free(blocks); | 1104 | ulist_free(blocks); |
| 1105 | } | 1105 | } |
| 1106 | 1106 | ||
| 1107 | /* | 1107 | /* |
| 1108 | * Finds all leafs with a reference to the specified combination of bytenr and | 1108 | * Finds all leafs with a reference to the specified combination of bytenr and |
| 1109 | * offset. key_list_head will point to a list of corresponding keys (caller must | 1109 | * offset. key_list_head will point to a list of corresponding keys (caller must |
| 1110 | * free each list element). The leafs will be stored in the leafs ulist, which | 1110 | * free each list element). The leafs will be stored in the leafs ulist, which |
| 1111 | * must be freed with ulist_free. | 1111 | * must be freed with ulist_free. |
| 1112 | * | 1112 | * |
| 1113 | * returns 0 on success, <0 on error | 1113 | * returns 0 on success, <0 on error |
| 1114 | */ | 1114 | */ |
| 1115 | static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans, | 1115 | static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans, |
| 1116 | struct btrfs_fs_info *fs_info, u64 bytenr, | 1116 | struct btrfs_fs_info *fs_info, u64 bytenr, |
| 1117 | u64 time_seq, struct ulist **leafs, | 1117 | u64 time_seq, struct ulist **leafs, |
| 1118 | const u64 *extent_item_pos) | 1118 | const u64 *extent_item_pos) |
| 1119 | { | 1119 | { |
| 1120 | int ret; | 1120 | int ret; |
| 1121 | 1121 | ||
| 1122 | *leafs = ulist_alloc(GFP_NOFS); | 1122 | *leafs = ulist_alloc(GFP_NOFS); |
| 1123 | if (!*leafs) | 1123 | if (!*leafs) |
| 1124 | return -ENOMEM; | 1124 | return -ENOMEM; |
| 1125 | 1125 | ||
| 1126 | ret = find_parent_nodes(trans, fs_info, bytenr, | 1126 | ret = find_parent_nodes(trans, fs_info, bytenr, |
| 1127 | time_seq, *leafs, NULL, extent_item_pos, 0, 0); | 1127 | time_seq, *leafs, NULL, extent_item_pos, 0, 0); |
| 1128 | if (ret < 0 && ret != -ENOENT) { | 1128 | if (ret < 0 && ret != -ENOENT) { |
| 1129 | free_leaf_list(*leafs); | 1129 | free_leaf_list(*leafs); |
| 1130 | return ret; | 1130 | return ret; |
| 1131 | } | 1131 | } |
| 1132 | 1132 | ||
| 1133 | return 0; | 1133 | return 0; |
| 1134 | } | 1134 | } |
| 1135 | 1135 | ||
| 1136 | /* | 1136 | /* |
| 1137 | * walk all backrefs for a given extent to find all roots that reference this | 1137 | * walk all backrefs for a given extent to find all roots that reference this |
| 1138 | * extent. Walking a backref means finding all extents that reference this | 1138 | * extent. Walking a backref means finding all extents that reference this |
| 1139 | * extent and in turn walk the backrefs of those, too. Naturally this is a | 1139 | * extent and in turn walk the backrefs of those, too. Naturally this is a |
| 1140 | * recursive process, but here it is implemented in an iterative fashion: We | 1140 | * recursive process, but here it is implemented in an iterative fashion: We |
| 1141 | * find all referencing extents for the extent in question and put them on a | 1141 | * find all referencing extents for the extent in question and put them on a |
| 1142 | * list. In turn, we find all referencing extents for those, further appending | 1142 | * list. In turn, we find all referencing extents for those, further appending |
| 1143 | * to the list. The way we iterate the list allows adding more elements after | 1143 | * to the list. The way we iterate the list allows adding more elements after |
| 1144 | * the current while iterating. The process stops when we reach the end of the | 1144 | * the current while iterating. The process stops when we reach the end of the |
| 1145 | * list. Found roots are added to the roots list. | 1145 | * list. Found roots are added to the roots list. |
| 1146 | * | 1146 | * |
| 1147 | * returns 0 on success, < 0 on error. | 1147 | * returns 0 on success, < 0 on error. |
| 1148 | */ | 1148 | */ |
| 1149 | static int __btrfs_find_all_roots(struct btrfs_trans_handle *trans, | 1149 | static int __btrfs_find_all_roots(struct btrfs_trans_handle *trans, |
| 1150 | struct btrfs_fs_info *fs_info, u64 bytenr, | 1150 | struct btrfs_fs_info *fs_info, u64 bytenr, |
| 1151 | u64 time_seq, struct ulist **roots) | 1151 | u64 time_seq, struct ulist **roots) |
| 1152 | { | 1152 | { |
| 1153 | struct ulist *tmp; | 1153 | struct ulist *tmp; |
| 1154 | struct ulist_node *node = NULL; | 1154 | struct ulist_node *node = NULL; |
| 1155 | struct ulist_iterator uiter; | 1155 | struct ulist_iterator uiter; |
| 1156 | int ret; | 1156 | int ret; |
| 1157 | 1157 | ||
| 1158 | tmp = ulist_alloc(GFP_NOFS); | 1158 | tmp = ulist_alloc(GFP_NOFS); |
| 1159 | if (!tmp) | 1159 | if (!tmp) |
| 1160 | return -ENOMEM; | 1160 | return -ENOMEM; |
| 1161 | *roots = ulist_alloc(GFP_NOFS); | 1161 | *roots = ulist_alloc(GFP_NOFS); |
| 1162 | if (!*roots) { | 1162 | if (!*roots) { |
| 1163 | ulist_free(tmp); | 1163 | ulist_free(tmp); |
| 1164 | return -ENOMEM; | 1164 | return -ENOMEM; |
| 1165 | } | 1165 | } |
| 1166 | 1166 | ||
| 1167 | ULIST_ITER_INIT(&uiter); | 1167 | ULIST_ITER_INIT(&uiter); |
| 1168 | while (1) { | 1168 | while (1) { |
| 1169 | ret = find_parent_nodes(trans, fs_info, bytenr, | 1169 | ret = find_parent_nodes(trans, fs_info, bytenr, |
| 1170 | time_seq, tmp, *roots, NULL, 0, 0); | 1170 | time_seq, tmp, *roots, NULL, 0, 0); |
| 1171 | if (ret < 0 && ret != -ENOENT) { | 1171 | if (ret < 0 && ret != -ENOENT) { |
| 1172 | ulist_free(tmp); | 1172 | ulist_free(tmp); |
| 1173 | ulist_free(*roots); | 1173 | ulist_free(*roots); |
| 1174 | return ret; | 1174 | return ret; |
| 1175 | } | 1175 | } |
| 1176 | node = ulist_next(tmp, &uiter); | 1176 | node = ulist_next(tmp, &uiter); |
| 1177 | if (!node) | 1177 | if (!node) |
| 1178 | break; | 1178 | break; |
| 1179 | bytenr = node->val; | 1179 | bytenr = node->val; |
| 1180 | cond_resched(); | 1180 | cond_resched(); |
| 1181 | } | 1181 | } |
| 1182 | 1182 | ||
| 1183 | ulist_free(tmp); | 1183 | ulist_free(tmp); |
| 1184 | return 0; | 1184 | return 0; |
| 1185 | } | 1185 | } |
| 1186 | 1186 | ||
| 1187 | int btrfs_find_all_roots(struct btrfs_trans_handle *trans, | 1187 | int btrfs_find_all_roots(struct btrfs_trans_handle *trans, |
| 1188 | struct btrfs_fs_info *fs_info, u64 bytenr, | 1188 | struct btrfs_fs_info *fs_info, u64 bytenr, |
| 1189 | u64 time_seq, struct ulist **roots) | 1189 | u64 time_seq, struct ulist **roots) |
| 1190 | { | 1190 | { |
| 1191 | int ret; | 1191 | int ret; |
| 1192 | 1192 | ||
| 1193 | if (!trans) | 1193 | if (!trans) |
| 1194 | down_read(&fs_info->commit_root_sem); | 1194 | down_read(&fs_info->commit_root_sem); |
| 1195 | ret = __btrfs_find_all_roots(trans, fs_info, bytenr, time_seq, roots); | 1195 | ret = __btrfs_find_all_roots(trans, fs_info, bytenr, time_seq, roots); |
| 1196 | if (!trans) | 1196 | if (!trans) |
| 1197 | up_read(&fs_info->commit_root_sem); | 1197 | up_read(&fs_info->commit_root_sem); |
| 1198 | return ret; | 1198 | return ret; |
| 1199 | } | 1199 | } |
| 1200 | 1200 | ||
| 1201 | int btrfs_check_shared(struct btrfs_trans_handle *trans, | 1201 | int btrfs_check_shared(struct btrfs_trans_handle *trans, |
| 1202 | struct btrfs_fs_info *fs_info, u64 root_objectid, | 1202 | struct btrfs_fs_info *fs_info, u64 root_objectid, |
| 1203 | u64 inum, u64 bytenr) | 1203 | u64 inum, u64 bytenr) |
| 1204 | { | 1204 | { |
| 1205 | struct ulist *tmp = NULL; | 1205 | struct ulist *tmp = NULL; |
| 1206 | struct ulist *roots = NULL; | 1206 | struct ulist *roots = NULL; |
| 1207 | struct ulist_iterator uiter; | 1207 | struct ulist_iterator uiter; |
| 1208 | struct ulist_node *node; | 1208 | struct ulist_node *node; |
| 1209 | struct seq_list elem = {}; | 1209 | struct seq_list elem = {}; |
| 1210 | int ret = 0; | 1210 | int ret = 0; |
| 1211 | 1211 | ||
| 1212 | tmp = ulist_alloc(GFP_NOFS); | 1212 | tmp = ulist_alloc(GFP_NOFS); |
| 1213 | roots = ulist_alloc(GFP_NOFS); | 1213 | roots = ulist_alloc(GFP_NOFS); |
| 1214 | if (!tmp || !roots) { | 1214 | if (!tmp || !roots) { |
| 1215 | ulist_free(tmp); | 1215 | ulist_free(tmp); |
| 1216 | ulist_free(roots); | 1216 | ulist_free(roots); |
| 1217 | return -ENOMEM; | 1217 | return -ENOMEM; |
| 1218 | } | 1218 | } |
| 1219 | 1219 | ||
| 1220 | if (trans) | 1220 | if (trans) |
| 1221 | btrfs_get_tree_mod_seq(fs_info, &elem); | 1221 | btrfs_get_tree_mod_seq(fs_info, &elem); |
| 1222 | else | 1222 | else |
| 1223 | down_read(&fs_info->commit_root_sem); | 1223 | down_read(&fs_info->commit_root_sem); |
| 1224 | ULIST_ITER_INIT(&uiter); | 1224 | ULIST_ITER_INIT(&uiter); |
| 1225 | while (1) { | 1225 | while (1) { |
| 1226 | ret = find_parent_nodes(trans, fs_info, bytenr, elem.seq, tmp, | 1226 | ret = find_parent_nodes(trans, fs_info, bytenr, elem.seq, tmp, |
| 1227 | roots, NULL, root_objectid, inum); | 1227 | roots, NULL, root_objectid, inum); |
| 1228 | if (ret == BACKREF_FOUND_SHARED) { | 1228 | if (ret == BACKREF_FOUND_SHARED) { |
| 1229 | ret = 1; | 1229 | ret = 1; |
| 1230 | break; | 1230 | break; |
| 1231 | } | 1231 | } |
| 1232 | if (ret < 0 && ret != -ENOENT) | 1232 | if (ret < 0 && ret != -ENOENT) |
| 1233 | break; | 1233 | break; |
| 1234 | node = ulist_next(tmp, &uiter); | 1234 | node = ulist_next(tmp, &uiter); |
| 1235 | if (!node) | 1235 | if (!node) |
| 1236 | break; | 1236 | break; |
| 1237 | bytenr = node->val; | 1237 | bytenr = node->val; |
| 1238 | cond_resched(); | 1238 | cond_resched(); |
| 1239 | } | 1239 | } |
| 1240 | if (trans) | 1240 | if (trans) |
| 1241 | btrfs_put_tree_mod_seq(fs_info, &elem); | 1241 | btrfs_put_tree_mod_seq(fs_info, &elem); |
| 1242 | else | 1242 | else |
| 1243 | up_read(&fs_info->commit_root_sem); | 1243 | up_read(&fs_info->commit_root_sem); |
| 1244 | ulist_free(tmp); | 1244 | ulist_free(tmp); |
| 1245 | ulist_free(roots); | 1245 | ulist_free(roots); |
| 1246 | return ret; | 1246 | return ret; |
| 1247 | } | 1247 | } |
| 1248 | 1248 | ||
| 1249 | /* | 1249 | /* |
| 1250 | * this makes the path point to (inum INODE_ITEM ioff) | 1250 | * this makes the path point to (inum INODE_ITEM ioff) |
| 1251 | */ | 1251 | */ |
| 1252 | int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, | 1252 | int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, |
| 1253 | struct btrfs_path *path) | 1253 | struct btrfs_path *path) |
| 1254 | { | 1254 | { |
| 1255 | struct btrfs_key key; | 1255 | struct btrfs_key key; |
| 1256 | return btrfs_find_item(fs_root, path, inum, ioff, | 1256 | return btrfs_find_item(fs_root, path, inum, ioff, |
| 1257 | BTRFS_INODE_ITEM_KEY, &key); | 1257 | BTRFS_INODE_ITEM_KEY, &key); |
| 1258 | } | 1258 | } |
| 1259 | 1259 | ||
| 1260 | static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, | 1260 | static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root, |
| 1261 | struct btrfs_path *path, | 1261 | struct btrfs_path *path, |
| 1262 | struct btrfs_key *found_key) | 1262 | struct btrfs_key *found_key) |
| 1263 | { | 1263 | { |
| 1264 | return btrfs_find_item(fs_root, path, inum, ioff, | 1264 | return btrfs_find_item(fs_root, path, inum, ioff, |
| 1265 | BTRFS_INODE_REF_KEY, found_key); | 1265 | BTRFS_INODE_REF_KEY, found_key); |
| 1266 | } | 1266 | } |
| 1267 | 1267 | ||
| 1268 | int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid, | 1268 | int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid, |
| 1269 | u64 start_off, struct btrfs_path *path, | 1269 | u64 start_off, struct btrfs_path *path, |
| 1270 | struct btrfs_inode_extref **ret_extref, | 1270 | struct btrfs_inode_extref **ret_extref, |
| 1271 | u64 *found_off) | 1271 | u64 *found_off) |
| 1272 | { | 1272 | { |
| 1273 | int ret, slot; | 1273 | int ret, slot; |
| 1274 | struct btrfs_key key; | 1274 | struct btrfs_key key; |
| 1275 | struct btrfs_key found_key; | 1275 | struct btrfs_key found_key; |
| 1276 | struct btrfs_inode_extref *extref; | 1276 | struct btrfs_inode_extref *extref; |
| 1277 | struct extent_buffer *leaf; | 1277 | struct extent_buffer *leaf; |
| 1278 | unsigned long ptr; | 1278 | unsigned long ptr; |
| 1279 | 1279 | ||
| 1280 | key.objectid = inode_objectid; | 1280 | key.objectid = inode_objectid; |
| 1281 | key.type = BTRFS_INODE_EXTREF_KEY; | 1281 | key.type = BTRFS_INODE_EXTREF_KEY; |
| 1282 | key.offset = start_off; | 1282 | key.offset = start_off; |
| 1283 | 1283 | ||
| 1284 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 1284 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 1285 | if (ret < 0) | 1285 | if (ret < 0) |
| 1286 | return ret; | 1286 | return ret; |
| 1287 | 1287 | ||
| 1288 | while (1) { | 1288 | while (1) { |
| 1289 | leaf = path->nodes[0]; | 1289 | leaf = path->nodes[0]; |
| 1290 | slot = path->slots[0]; | 1290 | slot = path->slots[0]; |
| 1291 | if (slot >= btrfs_header_nritems(leaf)) { | 1291 | if (slot >= btrfs_header_nritems(leaf)) { |
| 1292 | /* | 1292 | /* |
| 1293 | * If the item at offset is not found, | 1293 | * If the item at offset is not found, |
| 1294 | * btrfs_search_slot will point us to the slot | 1294 | * btrfs_search_slot will point us to the slot |
| 1295 | * where it should be inserted. In our case | 1295 | * where it should be inserted. In our case |
| 1296 | * that will be the slot directly before the | 1296 | * that will be the slot directly before the |
| 1297 | * next INODE_REF_KEY_V2 item. In the case | 1297 | * next INODE_REF_KEY_V2 item. In the case |
| 1298 | * that we're pointing to the last slot in a | 1298 | * that we're pointing to the last slot in a |
| 1299 | * leaf, we must move one leaf over. | 1299 | * leaf, we must move one leaf over. |
| 1300 | */ | 1300 | */ |
| 1301 | ret = btrfs_next_leaf(root, path); | 1301 | ret = btrfs_next_leaf(root, path); |
| 1302 | if (ret) { | 1302 | if (ret) { |
| 1303 | if (ret >= 1) | 1303 | if (ret >= 1) |
| 1304 | ret = -ENOENT; | 1304 | ret = -ENOENT; |
| 1305 | break; | 1305 | break; |
| 1306 | } | 1306 | } |
| 1307 | continue; | 1307 | continue; |
| 1308 | } | 1308 | } |
| 1309 | 1309 | ||
| 1310 | btrfs_item_key_to_cpu(leaf, &found_key, slot); | 1310 | btrfs_item_key_to_cpu(leaf, &found_key, slot); |
| 1311 | 1311 | ||
| 1312 | /* | 1312 | /* |
| 1313 | * Check that we're still looking at an extended ref key for | 1313 | * Check that we're still looking at an extended ref key for |
| 1314 | * this particular objectid. If we have different | 1314 | * this particular objectid. If we have different |
| 1315 | * objectid or type then there are no more to be found | 1315 | * objectid or type then there are no more to be found |
| 1316 | * in the tree and we can exit. | 1316 | * in the tree and we can exit. |
| 1317 | */ | 1317 | */ |
| 1318 | ret = -ENOENT; | 1318 | ret = -ENOENT; |
| 1319 | if (found_key.objectid != inode_objectid) | 1319 | if (found_key.objectid != inode_objectid) |
| 1320 | break; | 1320 | break; |
| 1321 | if (found_key.type != BTRFS_INODE_EXTREF_KEY) | 1321 | if (found_key.type != BTRFS_INODE_EXTREF_KEY) |
| 1322 | break; | 1322 | break; |
| 1323 | 1323 | ||
| 1324 | ret = 0; | 1324 | ret = 0; |
| 1325 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); | 1325 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); |
| 1326 | extref = (struct btrfs_inode_extref *)ptr; | 1326 | extref = (struct btrfs_inode_extref *)ptr; |
| 1327 | *ret_extref = extref; | 1327 | *ret_extref = extref; |
| 1328 | if (found_off) | 1328 | if (found_off) |
| 1329 | *found_off = found_key.offset; | 1329 | *found_off = found_key.offset; |
| 1330 | break; | 1330 | break; |
| 1331 | } | 1331 | } |
| 1332 | 1332 | ||
| 1333 | return ret; | 1333 | return ret; |
| 1334 | } | 1334 | } |
| 1335 | 1335 | ||
| 1336 | /* | 1336 | /* |
| 1337 | * this iterates to turn a name (from iref/extref) into a full filesystem path. | 1337 | * this iterates to turn a name (from iref/extref) into a full filesystem path. |
| 1338 | * Elements of the path are separated by '/' and the path is guaranteed to be | 1338 | * Elements of the path are separated by '/' and the path is guaranteed to be |
| 1339 | * 0-terminated. the path is only given within the current file system. | 1339 | * 0-terminated. the path is only given within the current file system. |
| 1340 | * Therefore, it never starts with a '/'. the caller is responsible to provide | 1340 | * Therefore, it never starts with a '/'. the caller is responsible to provide |
| 1341 | * "size" bytes in "dest". the dest buffer will be filled backwards. finally, | 1341 | * "size" bytes in "dest". the dest buffer will be filled backwards. finally, |
| 1342 | * the start point of the resulting string is returned. this pointer is within | 1342 | * the start point of the resulting string is returned. this pointer is within |
| 1343 | * dest, normally. | 1343 | * dest, normally. |
| 1344 | * in case the path buffer would overflow, the pointer is decremented further | 1344 | * in case the path buffer would overflow, the pointer is decremented further |
| 1345 | * as if output was written to the buffer, though no more output is actually | 1345 | * as if output was written to the buffer, though no more output is actually |
| 1346 | * generated. that way, the caller can determine how much space would be | 1346 | * generated. that way, the caller can determine how much space would be |
| 1347 | * required for the path to fit into the buffer. in that case, the returned | 1347 | * required for the path to fit into the buffer. in that case, the returned |
| 1348 | * value will be smaller than dest. callers must check this! | 1348 | * value will be smaller than dest. callers must check this! |
| 1349 | */ | 1349 | */ |
| 1350 | char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path, | 1350 | char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path, |
| 1351 | u32 name_len, unsigned long name_off, | 1351 | u32 name_len, unsigned long name_off, |
| 1352 | struct extent_buffer *eb_in, u64 parent, | 1352 | struct extent_buffer *eb_in, u64 parent, |
| 1353 | char *dest, u32 size) | 1353 | char *dest, u32 size) |
| 1354 | { | 1354 | { |
| 1355 | int slot; | 1355 | int slot; |
| 1356 | u64 next_inum; | 1356 | u64 next_inum; |
| 1357 | int ret; | 1357 | int ret; |
| 1358 | s64 bytes_left = ((s64)size) - 1; | 1358 | s64 bytes_left = ((s64)size) - 1; |
| 1359 | struct extent_buffer *eb = eb_in; | 1359 | struct extent_buffer *eb = eb_in; |
| 1360 | struct btrfs_key found_key; | 1360 | struct btrfs_key found_key; |
| 1361 | int leave_spinning = path->leave_spinning; | 1361 | int leave_spinning = path->leave_spinning; |
| 1362 | struct btrfs_inode_ref *iref; | 1362 | struct btrfs_inode_ref *iref; |
| 1363 | 1363 | ||
| 1364 | if (bytes_left >= 0) | 1364 | if (bytes_left >= 0) |
| 1365 | dest[bytes_left] = '\0'; | 1365 | dest[bytes_left] = '\0'; |
| 1366 | 1366 | ||
| 1367 | path->leave_spinning = 1; | 1367 | path->leave_spinning = 1; |
| 1368 | while (1) { | 1368 | while (1) { |
| 1369 | bytes_left -= name_len; | 1369 | bytes_left -= name_len; |
| 1370 | if (bytes_left >= 0) | 1370 | if (bytes_left >= 0) |
| 1371 | read_extent_buffer(eb, dest + bytes_left, | 1371 | read_extent_buffer(eb, dest + bytes_left, |
| 1372 | name_off, name_len); | 1372 | name_off, name_len); |
| 1373 | if (eb != eb_in) { | 1373 | if (eb != eb_in) { |
| 1374 | btrfs_tree_read_unlock_blocking(eb); | 1374 | btrfs_tree_read_unlock_blocking(eb); |
| 1375 | free_extent_buffer(eb); | 1375 | free_extent_buffer(eb); |
| 1376 | } | 1376 | } |
| 1377 | ret = inode_ref_info(parent, 0, fs_root, path, &found_key); | 1377 | ret = inode_ref_info(parent, 0, fs_root, path, &found_key); |
| 1378 | if (ret > 0) | 1378 | if (ret > 0) |
| 1379 | ret = -ENOENT; | 1379 | ret = -ENOENT; |
| 1380 | if (ret) | 1380 | if (ret) |
| 1381 | break; | 1381 | break; |
| 1382 | 1382 | ||
| 1383 | next_inum = found_key.offset; | 1383 | next_inum = found_key.offset; |
| 1384 | 1384 | ||
| 1385 | /* regular exit ahead */ | 1385 | /* regular exit ahead */ |
| 1386 | if (parent == next_inum) | 1386 | if (parent == next_inum) |
| 1387 | break; | 1387 | break; |
| 1388 | 1388 | ||
| 1389 | slot = path->slots[0]; | 1389 | slot = path->slots[0]; |
| 1390 | eb = path->nodes[0]; | 1390 | eb = path->nodes[0]; |
| 1391 | /* make sure we can use eb after releasing the path */ | 1391 | /* make sure we can use eb after releasing the path */ |
| 1392 | if (eb != eb_in) { | 1392 | if (eb != eb_in) { |
| 1393 | atomic_inc(&eb->refs); | 1393 | atomic_inc(&eb->refs); |
| 1394 | btrfs_tree_read_lock(eb); | 1394 | btrfs_tree_read_lock(eb); |
| 1395 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); | 1395 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); |
| 1396 | } | 1396 | } |
| 1397 | btrfs_release_path(path); | 1397 | btrfs_release_path(path); |
| 1398 | iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); | 1398 | iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); |
| 1399 | 1399 | ||
| 1400 | name_len = btrfs_inode_ref_name_len(eb, iref); | 1400 | name_len = btrfs_inode_ref_name_len(eb, iref); |
| 1401 | name_off = (unsigned long)(iref + 1); | 1401 | name_off = (unsigned long)(iref + 1); |
| 1402 | 1402 | ||
| 1403 | parent = next_inum; | 1403 | parent = next_inum; |
| 1404 | --bytes_left; | 1404 | --bytes_left; |
| 1405 | if (bytes_left >= 0) | 1405 | if (bytes_left >= 0) |
| 1406 | dest[bytes_left] = '/'; | 1406 | dest[bytes_left] = '/'; |
| 1407 | } | 1407 | } |
| 1408 | 1408 | ||
| 1409 | btrfs_release_path(path); | 1409 | btrfs_release_path(path); |
| 1410 | path->leave_spinning = leave_spinning; | 1410 | path->leave_spinning = leave_spinning; |
| 1411 | 1411 | ||
| 1412 | if (ret) | 1412 | if (ret) |
| 1413 | return ERR_PTR(ret); | 1413 | return ERR_PTR(ret); |
| 1414 | 1414 | ||
| 1415 | return dest + bytes_left; | 1415 | return dest + bytes_left; |
| 1416 | } | 1416 | } |
| 1417 | 1417 | ||
| 1418 | /* | 1418 | /* |
| 1419 | * this makes the path point to (logical EXTENT_ITEM *) | 1419 | * this makes the path point to (logical EXTENT_ITEM *) |
| 1420 | * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for | 1420 | * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for |
| 1421 | * tree blocks and <0 on error. | 1421 | * tree blocks and <0 on error. |
| 1422 | */ | 1422 | */ |
| 1423 | int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical, | 1423 | int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical, |
| 1424 | struct btrfs_path *path, struct btrfs_key *found_key, | 1424 | struct btrfs_path *path, struct btrfs_key *found_key, |
| 1425 | u64 *flags_ret) | 1425 | u64 *flags_ret) |
| 1426 | { | 1426 | { |
| 1427 | int ret; | 1427 | int ret; |
| 1428 | u64 flags; | 1428 | u64 flags; |
| 1429 | u64 size = 0; | 1429 | u64 size = 0; |
| 1430 | u32 item_size; | 1430 | u32 item_size; |
| 1431 | struct extent_buffer *eb; | 1431 | struct extent_buffer *eb; |
| 1432 | struct btrfs_extent_item *ei; | 1432 | struct btrfs_extent_item *ei; |
| 1433 | struct btrfs_key key; | 1433 | struct btrfs_key key; |
| 1434 | 1434 | ||
| 1435 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) | 1435 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) |
| 1436 | key.type = BTRFS_METADATA_ITEM_KEY; | 1436 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 1437 | else | 1437 | else |
| 1438 | key.type = BTRFS_EXTENT_ITEM_KEY; | 1438 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 1439 | key.objectid = logical; | 1439 | key.objectid = logical; |
| 1440 | key.offset = (u64)-1; | 1440 | key.offset = (u64)-1; |
| 1441 | 1441 | ||
| 1442 | ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0); | 1442 | ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0); |
| 1443 | if (ret < 0) | 1443 | if (ret < 0) |
| 1444 | return ret; | 1444 | return ret; |
| 1445 | 1445 | ||
| 1446 | ret = btrfs_previous_extent_item(fs_info->extent_root, path, 0); | 1446 | ret = btrfs_previous_extent_item(fs_info->extent_root, path, 0); |
| 1447 | if (ret) { | 1447 | if (ret) { |
| 1448 | if (ret > 0) | 1448 | if (ret > 0) |
| 1449 | ret = -ENOENT; | 1449 | ret = -ENOENT; |
| 1450 | return ret; | 1450 | return ret; |
| 1451 | } | 1451 | } |
| 1452 | btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]); | 1452 | btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]); |
| 1453 | if (found_key->type == BTRFS_METADATA_ITEM_KEY) | 1453 | if (found_key->type == BTRFS_METADATA_ITEM_KEY) |
| 1454 | size = fs_info->extent_root->nodesize; | 1454 | size = fs_info->extent_root->nodesize; |
| 1455 | else if (found_key->type == BTRFS_EXTENT_ITEM_KEY) | 1455 | else if (found_key->type == BTRFS_EXTENT_ITEM_KEY) |
| 1456 | size = found_key->offset; | 1456 | size = found_key->offset; |
| 1457 | 1457 | ||
| 1458 | if (found_key->objectid > logical || | 1458 | if (found_key->objectid > logical || |
| 1459 | found_key->objectid + size <= logical) { | 1459 | found_key->objectid + size <= logical) { |
| 1460 | pr_debug("logical %llu is not within any extent\n", logical); | 1460 | pr_debug("logical %llu is not within any extent\n", logical); |
| 1461 | return -ENOENT; | 1461 | return -ENOENT; |
| 1462 | } | 1462 | } |
| 1463 | 1463 | ||
| 1464 | eb = path->nodes[0]; | 1464 | eb = path->nodes[0]; |
| 1465 | item_size = btrfs_item_size_nr(eb, path->slots[0]); | 1465 | item_size = btrfs_item_size_nr(eb, path->slots[0]); |
| 1466 | BUG_ON(item_size < sizeof(*ei)); | 1466 | BUG_ON(item_size < sizeof(*ei)); |
| 1467 | 1467 | ||
| 1468 | ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); | 1468 | ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); |
| 1469 | flags = btrfs_extent_flags(eb, ei); | 1469 | flags = btrfs_extent_flags(eb, ei); |
| 1470 | 1470 | ||
| 1471 | pr_debug("logical %llu is at position %llu within the extent (%llu " | 1471 | pr_debug("logical %llu is at position %llu within the extent (%llu " |
| 1472 | "EXTENT_ITEM %llu) flags %#llx size %u\n", | 1472 | "EXTENT_ITEM %llu) flags %#llx size %u\n", |
| 1473 | logical, logical - found_key->objectid, found_key->objectid, | 1473 | logical, logical - found_key->objectid, found_key->objectid, |
| 1474 | found_key->offset, flags, item_size); | 1474 | found_key->offset, flags, item_size); |
| 1475 | 1475 | ||
| 1476 | WARN_ON(!flags_ret); | 1476 | WARN_ON(!flags_ret); |
| 1477 | if (flags_ret) { | 1477 | if (flags_ret) { |
| 1478 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) | 1478 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) |
| 1479 | *flags_ret = BTRFS_EXTENT_FLAG_TREE_BLOCK; | 1479 | *flags_ret = BTRFS_EXTENT_FLAG_TREE_BLOCK; |
| 1480 | else if (flags & BTRFS_EXTENT_FLAG_DATA) | 1480 | else if (flags & BTRFS_EXTENT_FLAG_DATA) |
| 1481 | *flags_ret = BTRFS_EXTENT_FLAG_DATA; | 1481 | *flags_ret = BTRFS_EXTENT_FLAG_DATA; |
| 1482 | else | 1482 | else |
| 1483 | BUG_ON(1); | 1483 | BUG_ON(1); |
| 1484 | return 0; | 1484 | return 0; |
| 1485 | } | 1485 | } |
| 1486 | 1486 | ||
| 1487 | return -EIO; | 1487 | return -EIO; |
| 1488 | } | 1488 | } |
| 1489 | 1489 | ||
| 1490 | /* | 1490 | /* |
| 1491 | * helper function to iterate extent inline refs. ptr must point to a 0 value | 1491 | * helper function to iterate extent inline refs. ptr must point to a 0 value |
| 1492 | * for the first call and may be modified. it is used to track state. | 1492 | * for the first call and may be modified. it is used to track state. |
| 1493 | * if more refs exist, 0 is returned and the next call to | 1493 | * if more refs exist, 0 is returned and the next call to |
| 1494 | * __get_extent_inline_ref must pass the modified ptr parameter to get the | 1494 | * __get_extent_inline_ref must pass the modified ptr parameter to get the |
| 1495 | * next ref. after the last ref was processed, 1 is returned. | 1495 | * next ref. after the last ref was processed, 1 is returned. |
| 1496 | * returns <0 on error | 1496 | * returns <0 on error |
| 1497 | */ | 1497 | */ |
| 1498 | static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb, | 1498 | static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb, |
| 1499 | struct btrfs_key *key, | 1499 | struct btrfs_key *key, |
| 1500 | struct btrfs_extent_item *ei, u32 item_size, | 1500 | struct btrfs_extent_item *ei, u32 item_size, |
| 1501 | struct btrfs_extent_inline_ref **out_eiref, | 1501 | struct btrfs_extent_inline_ref **out_eiref, |
| 1502 | int *out_type) | 1502 | int *out_type) |
| 1503 | { | 1503 | { |
| 1504 | unsigned long end; | 1504 | unsigned long end; |
| 1505 | u64 flags; | 1505 | u64 flags; |
| 1506 | struct btrfs_tree_block_info *info; | 1506 | struct btrfs_tree_block_info *info; |
| 1507 | 1507 | ||
| 1508 | if (!*ptr) { | 1508 | if (!*ptr) { |
| 1509 | /* first call */ | 1509 | /* first call */ |
| 1510 | flags = btrfs_extent_flags(eb, ei); | 1510 | flags = btrfs_extent_flags(eb, ei); |
| 1511 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { | 1511 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { |
| 1512 | if (key->type == BTRFS_METADATA_ITEM_KEY) { | 1512 | if (key->type == BTRFS_METADATA_ITEM_KEY) { |
| 1513 | /* a skinny metadata extent */ | 1513 | /* a skinny metadata extent */ |
| 1514 | *out_eiref = | 1514 | *out_eiref = |
| 1515 | (struct btrfs_extent_inline_ref *)(ei + 1); | 1515 | (struct btrfs_extent_inline_ref *)(ei + 1); |
| 1516 | } else { | 1516 | } else { |
| 1517 | WARN_ON(key->type != BTRFS_EXTENT_ITEM_KEY); | 1517 | WARN_ON(key->type != BTRFS_EXTENT_ITEM_KEY); |
| 1518 | info = (struct btrfs_tree_block_info *)(ei + 1); | 1518 | info = (struct btrfs_tree_block_info *)(ei + 1); |
| 1519 | *out_eiref = | 1519 | *out_eiref = |
| 1520 | (struct btrfs_extent_inline_ref *)(info + 1); | 1520 | (struct btrfs_extent_inline_ref *)(info + 1); |
| 1521 | } | 1521 | } |
| 1522 | } else { | 1522 | } else { |
| 1523 | *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1); | 1523 | *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1); |
| 1524 | } | 1524 | } |
| 1525 | *ptr = (unsigned long)*out_eiref; | 1525 | *ptr = (unsigned long)*out_eiref; |
| 1526 | if ((unsigned long)(*ptr) >= (unsigned long)ei + item_size) | 1526 | if ((unsigned long)(*ptr) >= (unsigned long)ei + item_size) |
| 1527 | return -ENOENT; | 1527 | return -ENOENT; |
| 1528 | } | 1528 | } |
| 1529 | 1529 | ||
| 1530 | end = (unsigned long)ei + item_size; | 1530 | end = (unsigned long)ei + item_size; |
| 1531 | *out_eiref = (struct btrfs_extent_inline_ref *)(*ptr); | 1531 | *out_eiref = (struct btrfs_extent_inline_ref *)(*ptr); |
| 1532 | *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref); | 1532 | *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref); |
| 1533 | 1533 | ||
| 1534 | *ptr += btrfs_extent_inline_ref_size(*out_type); | 1534 | *ptr += btrfs_extent_inline_ref_size(*out_type); |
| 1535 | WARN_ON(*ptr > end); | 1535 | WARN_ON(*ptr > end); |
| 1536 | if (*ptr == end) | 1536 | if (*ptr == end) |
| 1537 | return 1; /* last */ | 1537 | return 1; /* last */ |
| 1538 | 1538 | ||
| 1539 | return 0; | 1539 | return 0; |
| 1540 | } | 1540 | } |
| 1541 | 1541 | ||
| 1542 | /* | 1542 | /* |
| 1543 | * reads the tree block backref for an extent. tree level and root are returned | 1543 | * reads the tree block backref for an extent. tree level and root are returned |
| 1544 | * through out_level and out_root. ptr must point to a 0 value for the first | 1544 | * through out_level and out_root. ptr must point to a 0 value for the first |
| 1545 | * call and may be modified (see __get_extent_inline_ref comment). | 1545 | * call and may be modified (see __get_extent_inline_ref comment). |
| 1546 | * returns 0 if data was provided, 1 if there was no more data to provide or | 1546 | * returns 0 if data was provided, 1 if there was no more data to provide or |
| 1547 | * <0 on error. | 1547 | * <0 on error. |
| 1548 | */ | 1548 | */ |
| 1549 | int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb, | 1549 | int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb, |
| 1550 | struct btrfs_key *key, struct btrfs_extent_item *ei, | 1550 | struct btrfs_key *key, struct btrfs_extent_item *ei, |
| 1551 | u32 item_size, u64 *out_root, u8 *out_level) | 1551 | u32 item_size, u64 *out_root, u8 *out_level) |
| 1552 | { | 1552 | { |
| 1553 | int ret; | 1553 | int ret; |
| 1554 | int type; | 1554 | int type; |
| 1555 | struct btrfs_tree_block_info *info; | ||
| 1556 | struct btrfs_extent_inline_ref *eiref; | 1555 | struct btrfs_extent_inline_ref *eiref; |
| 1557 | 1556 | ||
| 1558 | if (*ptr == (unsigned long)-1) | 1557 | if (*ptr == (unsigned long)-1) |
| 1559 | return 1; | 1558 | return 1; |
| 1560 | 1559 | ||
| 1561 | while (1) { | 1560 | while (1) { |
| 1562 | ret = __get_extent_inline_ref(ptr, eb, key, ei, item_size, | 1561 | ret = __get_extent_inline_ref(ptr, eb, key, ei, item_size, |
| 1563 | &eiref, &type); | 1562 | &eiref, &type); |
| 1564 | if (ret < 0) | 1563 | if (ret < 0) |
| 1565 | return ret; | 1564 | return ret; |
| 1566 | 1565 | ||
| 1567 | if (type == BTRFS_TREE_BLOCK_REF_KEY || | 1566 | if (type == BTRFS_TREE_BLOCK_REF_KEY || |
| 1568 | type == BTRFS_SHARED_BLOCK_REF_KEY) | 1567 | type == BTRFS_SHARED_BLOCK_REF_KEY) |
| 1569 | break; | 1568 | break; |
| 1570 | 1569 | ||
| 1571 | if (ret == 1) | 1570 | if (ret == 1) |
| 1572 | return 1; | 1571 | return 1; |
| 1573 | } | 1572 | } |
| 1574 | 1573 | ||
| 1575 | /* we can treat both ref types equally here */ | 1574 | /* we can treat both ref types equally here */ |
| 1576 | info = (struct btrfs_tree_block_info *)(ei + 1); | ||
| 1577 | *out_root = btrfs_extent_inline_ref_offset(eb, eiref); | 1575 | *out_root = btrfs_extent_inline_ref_offset(eb, eiref); |
| 1578 | *out_level = btrfs_tree_block_level(eb, info); | 1576 | |
| 1577 | if (key->type == BTRFS_EXTENT_ITEM_KEY) { | ||
| 1578 | struct btrfs_tree_block_info *info; | ||
| 1579 | |||
| 1580 | info = (struct btrfs_tree_block_info *)(ei + 1); | ||
| 1581 | *out_level = btrfs_tree_block_level(eb, info); | ||
| 1582 | } else { | ||
| 1583 | ASSERT(key->type == BTRFS_METADATA_ITEM_KEY); | ||
| 1584 | *out_level = (u8)key->offset; | ||
| 1585 | } | ||
| 1579 | 1586 | ||
| 1580 | if (ret == 1) | 1587 | if (ret == 1) |
| 1581 | *ptr = (unsigned long)-1; | 1588 | *ptr = (unsigned long)-1; |
| 1582 | 1589 | ||
| 1583 | return 0; | 1590 | return 0; |
| 1584 | } | 1591 | } |
| 1585 | 1592 | ||
| 1586 | static int iterate_leaf_refs(struct extent_inode_elem *inode_list, | 1593 | static int iterate_leaf_refs(struct extent_inode_elem *inode_list, |
| 1587 | u64 root, u64 extent_item_objectid, | 1594 | u64 root, u64 extent_item_objectid, |
| 1588 | iterate_extent_inodes_t *iterate, void *ctx) | 1595 | iterate_extent_inodes_t *iterate, void *ctx) |
| 1589 | { | 1596 | { |
| 1590 | struct extent_inode_elem *eie; | 1597 | struct extent_inode_elem *eie; |
| 1591 | int ret = 0; | 1598 | int ret = 0; |
| 1592 | 1599 | ||
| 1593 | for (eie = inode_list; eie; eie = eie->next) { | 1600 | for (eie = inode_list; eie; eie = eie->next) { |
| 1594 | pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), " | 1601 | pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), " |
| 1595 | "root %llu\n", extent_item_objectid, | 1602 | "root %llu\n", extent_item_objectid, |
| 1596 | eie->inum, eie->offset, root); | 1603 | eie->inum, eie->offset, root); |
| 1597 | ret = iterate(eie->inum, eie->offset, root, ctx); | 1604 | ret = iterate(eie->inum, eie->offset, root, ctx); |
| 1598 | if (ret) { | 1605 | if (ret) { |
| 1599 | pr_debug("stopping iteration for %llu due to ret=%d\n", | 1606 | pr_debug("stopping iteration for %llu due to ret=%d\n", |
| 1600 | extent_item_objectid, ret); | 1607 | extent_item_objectid, ret); |
| 1601 | break; | 1608 | break; |
| 1602 | } | 1609 | } |
| 1603 | } | 1610 | } |
| 1604 | 1611 | ||
| 1605 | return ret; | 1612 | return ret; |
| 1606 | } | 1613 | } |
| 1607 | 1614 | ||
| 1608 | /* | 1615 | /* |
| 1609 | * calls iterate() for every inode that references the extent identified by | 1616 | * calls iterate() for every inode that references the extent identified by |
| 1610 | * the given parameters. | 1617 | * the given parameters. |
| 1611 | * when the iterator function returns a non-zero value, iteration stops. | 1618 | * when the iterator function returns a non-zero value, iteration stops. |
| 1612 | */ | 1619 | */ |
| 1613 | int iterate_extent_inodes(struct btrfs_fs_info *fs_info, | 1620 | int iterate_extent_inodes(struct btrfs_fs_info *fs_info, |
| 1614 | u64 extent_item_objectid, u64 extent_item_pos, | 1621 | u64 extent_item_objectid, u64 extent_item_pos, |
| 1615 | int search_commit_root, | 1622 | int search_commit_root, |
| 1616 | iterate_extent_inodes_t *iterate, void *ctx) | 1623 | iterate_extent_inodes_t *iterate, void *ctx) |
| 1617 | { | 1624 | { |
| 1618 | int ret; | 1625 | int ret; |
| 1619 | struct btrfs_trans_handle *trans = NULL; | 1626 | struct btrfs_trans_handle *trans = NULL; |
| 1620 | struct ulist *refs = NULL; | 1627 | struct ulist *refs = NULL; |
| 1621 | struct ulist *roots = NULL; | 1628 | struct ulist *roots = NULL; |
| 1622 | struct ulist_node *ref_node = NULL; | 1629 | struct ulist_node *ref_node = NULL; |
| 1623 | struct ulist_node *root_node = NULL; | 1630 | struct ulist_node *root_node = NULL; |
| 1624 | struct seq_list tree_mod_seq_elem = {}; | 1631 | struct seq_list tree_mod_seq_elem = {}; |
| 1625 | struct ulist_iterator ref_uiter; | 1632 | struct ulist_iterator ref_uiter; |
| 1626 | struct ulist_iterator root_uiter; | 1633 | struct ulist_iterator root_uiter; |
| 1627 | 1634 | ||
| 1628 | pr_debug("resolving all inodes for extent %llu\n", | 1635 | pr_debug("resolving all inodes for extent %llu\n", |
| 1629 | extent_item_objectid); | 1636 | extent_item_objectid); |
| 1630 | 1637 | ||
| 1631 | if (!search_commit_root) { | 1638 | if (!search_commit_root) { |
| 1632 | trans = btrfs_join_transaction(fs_info->extent_root); | 1639 | trans = btrfs_join_transaction(fs_info->extent_root); |
| 1633 | if (IS_ERR(trans)) | 1640 | if (IS_ERR(trans)) |
| 1634 | return PTR_ERR(trans); | 1641 | return PTR_ERR(trans); |
| 1635 | btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem); | 1642 | btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem); |
| 1636 | } else { | 1643 | } else { |
| 1637 | down_read(&fs_info->commit_root_sem); | 1644 | down_read(&fs_info->commit_root_sem); |
| 1638 | } | 1645 | } |
| 1639 | 1646 | ||
| 1640 | ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid, | 1647 | ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid, |
| 1641 | tree_mod_seq_elem.seq, &refs, | 1648 | tree_mod_seq_elem.seq, &refs, |
| 1642 | &extent_item_pos); | 1649 | &extent_item_pos); |
| 1643 | if (ret) | 1650 | if (ret) |
| 1644 | goto out; | 1651 | goto out; |
| 1645 | 1652 | ||
| 1646 | ULIST_ITER_INIT(&ref_uiter); | 1653 | ULIST_ITER_INIT(&ref_uiter); |
| 1647 | while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) { | 1654 | while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) { |
| 1648 | ret = __btrfs_find_all_roots(trans, fs_info, ref_node->val, | 1655 | ret = __btrfs_find_all_roots(trans, fs_info, ref_node->val, |
| 1649 | tree_mod_seq_elem.seq, &roots); | 1656 | tree_mod_seq_elem.seq, &roots); |
| 1650 | if (ret) | 1657 | if (ret) |
| 1651 | break; | 1658 | break; |
| 1652 | ULIST_ITER_INIT(&root_uiter); | 1659 | ULIST_ITER_INIT(&root_uiter); |
| 1653 | while (!ret && (root_node = ulist_next(roots, &root_uiter))) { | 1660 | while (!ret && (root_node = ulist_next(roots, &root_uiter))) { |
| 1654 | pr_debug("root %llu references leaf %llu, data list " | 1661 | pr_debug("root %llu references leaf %llu, data list " |
| 1655 | "%#llx\n", root_node->val, ref_node->val, | 1662 | "%#llx\n", root_node->val, ref_node->val, |
| 1656 | ref_node->aux); | 1663 | ref_node->aux); |
| 1657 | ret = iterate_leaf_refs((struct extent_inode_elem *) | 1664 | ret = iterate_leaf_refs((struct extent_inode_elem *) |
| 1658 | (uintptr_t)ref_node->aux, | 1665 | (uintptr_t)ref_node->aux, |
| 1659 | root_node->val, | 1666 | root_node->val, |
| 1660 | extent_item_objectid, | 1667 | extent_item_objectid, |
| 1661 | iterate, ctx); | 1668 | iterate, ctx); |
| 1662 | } | 1669 | } |
| 1663 | ulist_free(roots); | 1670 | ulist_free(roots); |
| 1664 | } | 1671 | } |
| 1665 | 1672 | ||
| 1666 | free_leaf_list(refs); | 1673 | free_leaf_list(refs); |
| 1667 | out: | 1674 | out: |
| 1668 | if (!search_commit_root) { | 1675 | if (!search_commit_root) { |
| 1669 | btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem); | 1676 | btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem); |
| 1670 | btrfs_end_transaction(trans, fs_info->extent_root); | 1677 | btrfs_end_transaction(trans, fs_info->extent_root); |
| 1671 | } else { | 1678 | } else { |
| 1672 | up_read(&fs_info->commit_root_sem); | 1679 | up_read(&fs_info->commit_root_sem); |
| 1673 | } | 1680 | } |
| 1674 | 1681 | ||
| 1675 | return ret; | 1682 | return ret; |
| 1676 | } | 1683 | } |
| 1677 | 1684 | ||
| 1678 | int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info, | 1685 | int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info, |
| 1679 | struct btrfs_path *path, | 1686 | struct btrfs_path *path, |
| 1680 | iterate_extent_inodes_t *iterate, void *ctx) | 1687 | iterate_extent_inodes_t *iterate, void *ctx) |
| 1681 | { | 1688 | { |
| 1682 | int ret; | 1689 | int ret; |
| 1683 | u64 extent_item_pos; | 1690 | u64 extent_item_pos; |
| 1684 | u64 flags = 0; | 1691 | u64 flags = 0; |
| 1685 | struct btrfs_key found_key; | 1692 | struct btrfs_key found_key; |
| 1686 | int search_commit_root = path->search_commit_root; | 1693 | int search_commit_root = path->search_commit_root; |
| 1687 | 1694 | ||
| 1688 | ret = extent_from_logical(fs_info, logical, path, &found_key, &flags); | 1695 | ret = extent_from_logical(fs_info, logical, path, &found_key, &flags); |
| 1689 | btrfs_release_path(path); | 1696 | btrfs_release_path(path); |
| 1690 | if (ret < 0) | 1697 | if (ret < 0) |
| 1691 | return ret; | 1698 | return ret; |
| 1692 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) | 1699 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) |
| 1693 | return -EINVAL; | 1700 | return -EINVAL; |
| 1694 | 1701 | ||
| 1695 | extent_item_pos = logical - found_key.objectid; | 1702 | extent_item_pos = logical - found_key.objectid; |
| 1696 | ret = iterate_extent_inodes(fs_info, found_key.objectid, | 1703 | ret = iterate_extent_inodes(fs_info, found_key.objectid, |
| 1697 | extent_item_pos, search_commit_root, | 1704 | extent_item_pos, search_commit_root, |
| 1698 | iterate, ctx); | 1705 | iterate, ctx); |
| 1699 | 1706 | ||
| 1700 | return ret; | 1707 | return ret; |
| 1701 | } | 1708 | } |
| 1702 | 1709 | ||
| 1703 | typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off, | 1710 | typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off, |
| 1704 | struct extent_buffer *eb, void *ctx); | 1711 | struct extent_buffer *eb, void *ctx); |
| 1705 | 1712 | ||
| 1706 | static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root, | 1713 | static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root, |
| 1707 | struct btrfs_path *path, | 1714 | struct btrfs_path *path, |
| 1708 | iterate_irefs_t *iterate, void *ctx) | 1715 | iterate_irefs_t *iterate, void *ctx) |
| 1709 | { | 1716 | { |
| 1710 | int ret = 0; | 1717 | int ret = 0; |
| 1711 | int slot; | 1718 | int slot; |
| 1712 | u32 cur; | 1719 | u32 cur; |
| 1713 | u32 len; | 1720 | u32 len; |
| 1714 | u32 name_len; | 1721 | u32 name_len; |
| 1715 | u64 parent = 0; | 1722 | u64 parent = 0; |
| 1716 | int found = 0; | 1723 | int found = 0; |
| 1717 | struct extent_buffer *eb; | 1724 | struct extent_buffer *eb; |
| 1718 | struct btrfs_item *item; | 1725 | struct btrfs_item *item; |
| 1719 | struct btrfs_inode_ref *iref; | 1726 | struct btrfs_inode_ref *iref; |
| 1720 | struct btrfs_key found_key; | 1727 | struct btrfs_key found_key; |
| 1721 | 1728 | ||
| 1722 | while (!ret) { | 1729 | while (!ret) { |
| 1723 | ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path, | 1730 | ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path, |
| 1724 | &found_key); | 1731 | &found_key); |
| 1725 | if (ret < 0) | 1732 | if (ret < 0) |
| 1726 | break; | 1733 | break; |
| 1727 | if (ret) { | 1734 | if (ret) { |
| 1728 | ret = found ? 0 : -ENOENT; | 1735 | ret = found ? 0 : -ENOENT; |
| 1729 | break; | 1736 | break; |
| 1730 | } | 1737 | } |
| 1731 | ++found; | 1738 | ++found; |
| 1732 | 1739 | ||
| 1733 | parent = found_key.offset; | 1740 | parent = found_key.offset; |
| 1734 | slot = path->slots[0]; | 1741 | slot = path->slots[0]; |
| 1735 | eb = btrfs_clone_extent_buffer(path->nodes[0]); | 1742 | eb = btrfs_clone_extent_buffer(path->nodes[0]); |
| 1736 | if (!eb) { | 1743 | if (!eb) { |
| 1737 | ret = -ENOMEM; | 1744 | ret = -ENOMEM; |
| 1738 | break; | 1745 | break; |
| 1739 | } | 1746 | } |
| 1740 | extent_buffer_get(eb); | 1747 | extent_buffer_get(eb); |
| 1741 | btrfs_tree_read_lock(eb); | 1748 | btrfs_tree_read_lock(eb); |
| 1742 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); | 1749 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); |
| 1743 | btrfs_release_path(path); | 1750 | btrfs_release_path(path); |
| 1744 | 1751 | ||
| 1745 | item = btrfs_item_nr(slot); | 1752 | item = btrfs_item_nr(slot); |
| 1746 | iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); | 1753 | iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref); |
| 1747 | 1754 | ||
| 1748 | for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) { | 1755 | for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) { |
| 1749 | name_len = btrfs_inode_ref_name_len(eb, iref); | 1756 | name_len = btrfs_inode_ref_name_len(eb, iref); |
| 1750 | /* path must be released before calling iterate()! */ | 1757 | /* path must be released before calling iterate()! */ |
| 1751 | pr_debug("following ref at offset %u for inode %llu in " | 1758 | pr_debug("following ref at offset %u for inode %llu in " |
| 1752 | "tree %llu\n", cur, found_key.objectid, | 1759 | "tree %llu\n", cur, found_key.objectid, |
| 1753 | fs_root->objectid); | 1760 | fs_root->objectid); |
| 1754 | ret = iterate(parent, name_len, | 1761 | ret = iterate(parent, name_len, |
| 1755 | (unsigned long)(iref + 1), eb, ctx); | 1762 | (unsigned long)(iref + 1), eb, ctx); |
| 1756 | if (ret) | 1763 | if (ret) |
| 1757 | break; | 1764 | break; |
| 1758 | len = sizeof(*iref) + name_len; | 1765 | len = sizeof(*iref) + name_len; |
| 1759 | iref = (struct btrfs_inode_ref *)((char *)iref + len); | 1766 | iref = (struct btrfs_inode_ref *)((char *)iref + len); |
| 1760 | } | 1767 | } |
| 1761 | btrfs_tree_read_unlock_blocking(eb); | 1768 | btrfs_tree_read_unlock_blocking(eb); |
| 1762 | free_extent_buffer(eb); | 1769 | free_extent_buffer(eb); |
| 1763 | } | 1770 | } |
| 1764 | 1771 | ||
| 1765 | btrfs_release_path(path); | 1772 | btrfs_release_path(path); |
| 1766 | 1773 | ||
| 1767 | return ret; | 1774 | return ret; |
| 1768 | } | 1775 | } |
| 1769 | 1776 | ||
| 1770 | static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root, | 1777 | static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root, |
| 1771 | struct btrfs_path *path, | 1778 | struct btrfs_path *path, |
| 1772 | iterate_irefs_t *iterate, void *ctx) | 1779 | iterate_irefs_t *iterate, void *ctx) |
| 1773 | { | 1780 | { |
| 1774 | int ret; | 1781 | int ret; |
| 1775 | int slot; | 1782 | int slot; |
| 1776 | u64 offset = 0; | 1783 | u64 offset = 0; |
| 1777 | u64 parent; | 1784 | u64 parent; |
| 1778 | int found = 0; | 1785 | int found = 0; |
| 1779 | struct extent_buffer *eb; | 1786 | struct extent_buffer *eb; |
| 1780 | struct btrfs_inode_extref *extref; | 1787 | struct btrfs_inode_extref *extref; |
| 1781 | struct extent_buffer *leaf; | 1788 | struct extent_buffer *leaf; |
| 1782 | u32 item_size; | 1789 | u32 item_size; |
| 1783 | u32 cur_offset; | 1790 | u32 cur_offset; |
| 1784 | unsigned long ptr; | 1791 | unsigned long ptr; |
| 1785 | 1792 | ||
| 1786 | while (1) { | 1793 | while (1) { |
| 1787 | ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref, | 1794 | ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref, |
| 1788 | &offset); | 1795 | &offset); |
| 1789 | if (ret < 0) | 1796 | if (ret < 0) |
| 1790 | break; | 1797 | break; |
| 1791 | if (ret) { | 1798 | if (ret) { |
| 1792 | ret = found ? 0 : -ENOENT; | 1799 | ret = found ? 0 : -ENOENT; |
| 1793 | break; | 1800 | break; |
| 1794 | } | 1801 | } |
| 1795 | ++found; | 1802 | ++found; |
| 1796 | 1803 | ||
| 1797 | slot = path->slots[0]; | 1804 | slot = path->slots[0]; |
| 1798 | eb = btrfs_clone_extent_buffer(path->nodes[0]); | 1805 | eb = btrfs_clone_extent_buffer(path->nodes[0]); |
| 1799 | if (!eb) { | 1806 | if (!eb) { |
| 1800 | ret = -ENOMEM; | 1807 | ret = -ENOMEM; |
| 1801 | break; | 1808 | break; |
| 1802 | } | 1809 | } |
| 1803 | extent_buffer_get(eb); | 1810 | extent_buffer_get(eb); |
| 1804 | 1811 | ||
| 1805 | btrfs_tree_read_lock(eb); | 1812 | btrfs_tree_read_lock(eb); |
| 1806 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); | 1813 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); |
| 1807 | btrfs_release_path(path); | 1814 | btrfs_release_path(path); |
| 1808 | 1815 | ||
| 1809 | leaf = path->nodes[0]; | 1816 | leaf = path->nodes[0]; |
| 1810 | item_size = btrfs_item_size_nr(leaf, slot); | 1817 | item_size = btrfs_item_size_nr(leaf, slot); |
| 1811 | ptr = btrfs_item_ptr_offset(leaf, slot); | 1818 | ptr = btrfs_item_ptr_offset(leaf, slot); |
| 1812 | cur_offset = 0; | 1819 | cur_offset = 0; |
| 1813 | 1820 | ||
| 1814 | while (cur_offset < item_size) { | 1821 | while (cur_offset < item_size) { |
| 1815 | u32 name_len; | 1822 | u32 name_len; |
| 1816 | 1823 | ||
| 1817 | extref = (struct btrfs_inode_extref *)(ptr + cur_offset); | 1824 | extref = (struct btrfs_inode_extref *)(ptr + cur_offset); |
| 1818 | parent = btrfs_inode_extref_parent(eb, extref); | 1825 | parent = btrfs_inode_extref_parent(eb, extref); |
| 1819 | name_len = btrfs_inode_extref_name_len(eb, extref); | 1826 | name_len = btrfs_inode_extref_name_len(eb, extref); |
| 1820 | ret = iterate(parent, name_len, | 1827 | ret = iterate(parent, name_len, |
| 1821 | (unsigned long)&extref->name, eb, ctx); | 1828 | (unsigned long)&extref->name, eb, ctx); |
| 1822 | if (ret) | 1829 | if (ret) |
| 1823 | break; | 1830 | break; |
| 1824 | 1831 | ||
| 1825 | cur_offset += btrfs_inode_extref_name_len(leaf, extref); | 1832 | cur_offset += btrfs_inode_extref_name_len(leaf, extref); |
| 1826 | cur_offset += sizeof(*extref); | 1833 | cur_offset += sizeof(*extref); |
| 1827 | } | 1834 | } |
| 1828 | btrfs_tree_read_unlock_blocking(eb); | 1835 | btrfs_tree_read_unlock_blocking(eb); |
| 1829 | free_extent_buffer(eb); | 1836 | free_extent_buffer(eb); |
| 1830 | 1837 | ||
| 1831 | offset++; | 1838 | offset++; |
| 1832 | } | 1839 | } |
| 1833 | 1840 | ||
| 1834 | btrfs_release_path(path); | 1841 | btrfs_release_path(path); |
| 1835 | 1842 | ||
| 1836 | return ret; | 1843 | return ret; |
| 1837 | } | 1844 | } |
| 1838 | 1845 | ||
| 1839 | static int iterate_irefs(u64 inum, struct btrfs_root *fs_root, | 1846 | static int iterate_irefs(u64 inum, struct btrfs_root *fs_root, |
| 1840 | struct btrfs_path *path, iterate_irefs_t *iterate, | 1847 | struct btrfs_path *path, iterate_irefs_t *iterate, |
| 1841 | void *ctx) | 1848 | void *ctx) |
| 1842 | { | 1849 | { |
| 1843 | int ret; | 1850 | int ret; |
| 1844 | int found_refs = 0; | 1851 | int found_refs = 0; |
| 1845 | 1852 | ||
| 1846 | ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx); | 1853 | ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx); |
| 1847 | if (!ret) | 1854 | if (!ret) |
| 1848 | ++found_refs; | 1855 | ++found_refs; |
| 1849 | else if (ret != -ENOENT) | 1856 | else if (ret != -ENOENT) |
| 1850 | return ret; | 1857 | return ret; |
| 1851 | 1858 | ||
| 1852 | ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx); | 1859 | ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx); |
| 1853 | if (ret == -ENOENT && found_refs) | 1860 | if (ret == -ENOENT && found_refs) |
| 1854 | return 0; | 1861 | return 0; |
| 1855 | 1862 | ||
| 1856 | return ret; | 1863 | return ret; |
| 1857 | } | 1864 | } |
| 1858 | 1865 | ||
| 1859 | /* | 1866 | /* |
| 1860 | * returns 0 if the path could be dumped (probably truncated) | 1867 | * returns 0 if the path could be dumped (probably truncated) |
| 1861 | * returns <0 in case of an error | 1868 | * returns <0 in case of an error |
| 1862 | */ | 1869 | */ |
| 1863 | static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off, | 1870 | static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off, |
| 1864 | struct extent_buffer *eb, void *ctx) | 1871 | struct extent_buffer *eb, void *ctx) |
| 1865 | { | 1872 | { |
| 1866 | struct inode_fs_paths *ipath = ctx; | 1873 | struct inode_fs_paths *ipath = ctx; |
| 1867 | char *fspath; | 1874 | char *fspath; |
| 1868 | char *fspath_min; | 1875 | char *fspath_min; |
| 1869 | int i = ipath->fspath->elem_cnt; | 1876 | int i = ipath->fspath->elem_cnt; |
| 1870 | const int s_ptr = sizeof(char *); | 1877 | const int s_ptr = sizeof(char *); |
| 1871 | u32 bytes_left; | 1878 | u32 bytes_left; |
| 1872 | 1879 | ||
| 1873 | bytes_left = ipath->fspath->bytes_left > s_ptr ? | 1880 | bytes_left = ipath->fspath->bytes_left > s_ptr ? |
| 1874 | ipath->fspath->bytes_left - s_ptr : 0; | 1881 | ipath->fspath->bytes_left - s_ptr : 0; |
| 1875 | 1882 | ||
| 1876 | fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr; | 1883 | fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr; |
| 1877 | fspath = btrfs_ref_to_path(ipath->fs_root, ipath->btrfs_path, name_len, | 1884 | fspath = btrfs_ref_to_path(ipath->fs_root, ipath->btrfs_path, name_len, |
| 1878 | name_off, eb, inum, fspath_min, bytes_left); | 1885 | name_off, eb, inum, fspath_min, bytes_left); |
| 1879 | if (IS_ERR(fspath)) | 1886 | if (IS_ERR(fspath)) |
| 1880 | return PTR_ERR(fspath); | 1887 | return PTR_ERR(fspath); |
| 1881 | 1888 | ||
| 1882 | if (fspath > fspath_min) { | 1889 | if (fspath > fspath_min) { |
| 1883 | ipath->fspath->val[i] = (u64)(unsigned long)fspath; | 1890 | ipath->fspath->val[i] = (u64)(unsigned long)fspath; |
| 1884 | ++ipath->fspath->elem_cnt; | 1891 | ++ipath->fspath->elem_cnt; |
| 1885 | ipath->fspath->bytes_left = fspath - fspath_min; | 1892 | ipath->fspath->bytes_left = fspath - fspath_min; |
| 1886 | } else { | 1893 | } else { |
| 1887 | ++ipath->fspath->elem_missed; | 1894 | ++ipath->fspath->elem_missed; |
| 1888 | ipath->fspath->bytes_missing += fspath_min - fspath; | 1895 | ipath->fspath->bytes_missing += fspath_min - fspath; |
| 1889 | ipath->fspath->bytes_left = 0; | 1896 | ipath->fspath->bytes_left = 0; |
| 1890 | } | 1897 | } |
| 1891 | 1898 | ||
| 1892 | return 0; | 1899 | return 0; |
| 1893 | } | 1900 | } |
| 1894 | 1901 | ||
| 1895 | /* | 1902 | /* |
| 1896 | * this dumps all file system paths to the inode into the ipath struct, provided | 1903 | * this dumps all file system paths to the inode into the ipath struct, provided |
| 1897 | * is has been created large enough. each path is zero-terminated and accessed | 1904 | * is has been created large enough. each path is zero-terminated and accessed |
| 1898 | * from ipath->fspath->val[i]. | 1905 | * from ipath->fspath->val[i]. |
| 1899 | * when it returns, there are ipath->fspath->elem_cnt number of paths available | 1906 | * when it returns, there are ipath->fspath->elem_cnt number of paths available |
| 1900 | * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the | 1907 | * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the |
| 1901 | * number of missed paths in recored in ipath->fspath->elem_missed, otherwise, | 1908 | * number of missed paths in recored in ipath->fspath->elem_missed, otherwise, |
| 1902 | * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would | 1909 | * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would |
| 1903 | * have been needed to return all paths. | 1910 | * have been needed to return all paths. |
| 1904 | */ | 1911 | */ |
| 1905 | int paths_from_inode(u64 inum, struct inode_fs_paths *ipath) | 1912 | int paths_from_inode(u64 inum, struct inode_fs_paths *ipath) |
| 1906 | { | 1913 | { |
| 1907 | return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path, | 1914 | return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path, |
| 1908 | inode_to_path, ipath); | 1915 | inode_to_path, ipath); |
| 1909 | } | 1916 | } |
| 1910 | 1917 | ||
| 1911 | struct btrfs_data_container *init_data_container(u32 total_bytes) | 1918 | struct btrfs_data_container *init_data_container(u32 total_bytes) |
| 1912 | { | 1919 | { |
| 1913 | struct btrfs_data_container *data; | 1920 | struct btrfs_data_container *data; |
| 1914 | size_t alloc_bytes; | 1921 | size_t alloc_bytes; |
| 1915 | 1922 | ||
| 1916 | alloc_bytes = max_t(size_t, total_bytes, sizeof(*data)); | 1923 | alloc_bytes = max_t(size_t, total_bytes, sizeof(*data)); |
| 1917 | data = vmalloc(alloc_bytes); | 1924 | data = vmalloc(alloc_bytes); |
| 1918 | if (!data) | 1925 | if (!data) |
| 1919 | return ERR_PTR(-ENOMEM); | 1926 | return ERR_PTR(-ENOMEM); |
| 1920 | 1927 | ||
| 1921 | if (total_bytes >= sizeof(*data)) { | 1928 | if (total_bytes >= sizeof(*data)) { |
| 1922 | data->bytes_left = total_bytes - sizeof(*data); | 1929 | data->bytes_left = total_bytes - sizeof(*data); |
| 1923 | data->bytes_missing = 0; | 1930 | data->bytes_missing = 0; |
| 1924 | } else { | 1931 | } else { |
| 1925 | data->bytes_missing = sizeof(*data) - total_bytes; | 1932 | data->bytes_missing = sizeof(*data) - total_bytes; |
| 1926 | data->bytes_left = 0; | 1933 | data->bytes_left = 0; |
| 1927 | } | 1934 | } |
| 1928 | 1935 | ||
| 1929 | data->elem_cnt = 0; | 1936 | data->elem_cnt = 0; |
| 1930 | data->elem_missed = 0; | 1937 | data->elem_missed = 0; |
| 1931 | 1938 | ||
| 1932 | return data; | 1939 | return data; |
| 1933 | } | 1940 | } |
| 1934 | 1941 | ||
| 1935 | /* | 1942 | /* |
| 1936 | * allocates space to return multiple file system paths for an inode. | 1943 | * allocates space to return multiple file system paths for an inode. |
| 1937 | * total_bytes to allocate are passed, note that space usable for actual path | 1944 | * total_bytes to allocate are passed, note that space usable for actual path |
| 1938 | * information will be total_bytes - sizeof(struct inode_fs_paths). | 1945 | * information will be total_bytes - sizeof(struct inode_fs_paths). |
| 1939 | * the returned pointer must be freed with free_ipath() in the end. | 1946 | * the returned pointer must be freed with free_ipath() in the end. |
| 1940 | */ | 1947 | */ |
| 1941 | struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root, | 1948 | struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root, |
| 1942 | struct btrfs_path *path) | 1949 | struct btrfs_path *path) |
| 1943 | { | 1950 | { |
| 1944 | struct inode_fs_paths *ifp; | 1951 | struct inode_fs_paths *ifp; |
| 1945 | struct btrfs_data_container *fspath; | 1952 | struct btrfs_data_container *fspath; |
| 1946 | 1953 | ||
| 1947 | fspath = init_data_container(total_bytes); | 1954 | fspath = init_data_container(total_bytes); |
| 1948 | if (IS_ERR(fspath)) | 1955 | if (IS_ERR(fspath)) |
| 1949 | return (void *)fspath; | 1956 | return (void *)fspath; |
| 1950 | 1957 | ||
| 1951 | ifp = kmalloc(sizeof(*ifp), GFP_NOFS); | 1958 | ifp = kmalloc(sizeof(*ifp), GFP_NOFS); |
| 1952 | if (!ifp) { | 1959 | if (!ifp) { |
| 1953 | kfree(fspath); | 1960 | kfree(fspath); |
| 1954 | return ERR_PTR(-ENOMEM); | 1961 | return ERR_PTR(-ENOMEM); |
| 1955 | } | 1962 | } |
| 1956 | 1963 | ||
| 1957 | ifp->btrfs_path = path; | 1964 | ifp->btrfs_path = path; |
| 1958 | ifp->fspath = fspath; | 1965 | ifp->fspath = fspath; |
| 1959 | ifp->fs_root = fs_root; | 1966 | ifp->fs_root = fs_root; |
| 1960 | 1967 | ||
| 1961 | return ifp; | 1968 | return ifp; |
| 1962 | } | 1969 | } |
| 1963 | 1970 | ||
| 1964 | void free_ipath(struct inode_fs_paths *ipath) | 1971 | void free_ipath(struct inode_fs_paths *ipath) |
| 1965 | { | 1972 | { |
| 1966 | if (!ipath) | 1973 | if (!ipath) |
| 1967 | return; | 1974 | return; |
| 1968 | vfree(ipath->fspath); | 1975 | vfree(ipath->fspath); |
| 1969 | kfree(ipath); | 1976 | kfree(ipath); |
fs/btrfs/delayed-inode.c
| 1 | /* | 1 | /* |
| 2 | * Copyright (C) 2011 Fujitsu. All rights reserved. | 2 | * Copyright (C) 2011 Fujitsu. All rights reserved. |
| 3 | * Written by Miao Xie <miaox@cn.fujitsu.com> | 3 | * Written by Miao Xie <miaox@cn.fujitsu.com> |
| 4 | * | 4 | * |
| 5 | * This program is free software; you can redistribute it and/or | 5 | * This program is free software; you can redistribute it and/or |
| 6 | * modify it under the terms of the GNU General Public | 6 | * modify it under the terms of the GNU General Public |
| 7 | * License v2 as published by the Free Software Foundation. | 7 | * License v2 as published by the Free Software Foundation. |
| 8 | * | 8 | * |
| 9 | * This program is distributed in the hope that it will be useful, | 9 | * This program is distributed in the hope that it will be useful, |
| 10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | 10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | 11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 12 | * General Public License for more details. | 12 | * General Public License for more details. |
| 13 | * | 13 | * |
| 14 | * You should have received a copy of the GNU General Public | 14 | * You should have received a copy of the GNU General Public |
| 15 | * License along with this program; if not, write to the | 15 | * License along with this program; if not, write to the |
| 16 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | 16 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 17 | * Boston, MA 021110-1307, USA. | 17 | * Boston, MA 021110-1307, USA. |
| 18 | */ | 18 | */ |
| 19 | 19 | ||
| 20 | #include <linux/slab.h> | 20 | #include <linux/slab.h> |
| 21 | #include "delayed-inode.h" | 21 | #include "delayed-inode.h" |
| 22 | #include "disk-io.h" | 22 | #include "disk-io.h" |
| 23 | #include "transaction.h" | 23 | #include "transaction.h" |
| 24 | #include "ctree.h" | 24 | #include "ctree.h" |
| 25 | 25 | ||
| 26 | #define BTRFS_DELAYED_WRITEBACK 512 | 26 | #define BTRFS_DELAYED_WRITEBACK 512 |
| 27 | #define BTRFS_DELAYED_BACKGROUND 128 | 27 | #define BTRFS_DELAYED_BACKGROUND 128 |
| 28 | #define BTRFS_DELAYED_BATCH 16 | 28 | #define BTRFS_DELAYED_BATCH 16 |
| 29 | 29 | ||
| 30 | static struct kmem_cache *delayed_node_cache; | 30 | static struct kmem_cache *delayed_node_cache; |
| 31 | 31 | ||
| 32 | int __init btrfs_delayed_inode_init(void) | 32 | int __init btrfs_delayed_inode_init(void) |
| 33 | { | 33 | { |
| 34 | delayed_node_cache = kmem_cache_create("btrfs_delayed_node", | 34 | delayed_node_cache = kmem_cache_create("btrfs_delayed_node", |
| 35 | sizeof(struct btrfs_delayed_node), | 35 | sizeof(struct btrfs_delayed_node), |
| 36 | 0, | 36 | 0, |
| 37 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, | 37 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, |
| 38 | NULL); | 38 | NULL); |
| 39 | if (!delayed_node_cache) | 39 | if (!delayed_node_cache) |
| 40 | return -ENOMEM; | 40 | return -ENOMEM; |
| 41 | return 0; | 41 | return 0; |
| 42 | } | 42 | } |
| 43 | 43 | ||
| 44 | void btrfs_delayed_inode_exit(void) | 44 | void btrfs_delayed_inode_exit(void) |
| 45 | { | 45 | { |
| 46 | if (delayed_node_cache) | 46 | if (delayed_node_cache) |
| 47 | kmem_cache_destroy(delayed_node_cache); | 47 | kmem_cache_destroy(delayed_node_cache); |
| 48 | } | 48 | } |
| 49 | 49 | ||
| 50 | static inline void btrfs_init_delayed_node( | 50 | static inline void btrfs_init_delayed_node( |
| 51 | struct btrfs_delayed_node *delayed_node, | 51 | struct btrfs_delayed_node *delayed_node, |
| 52 | struct btrfs_root *root, u64 inode_id) | 52 | struct btrfs_root *root, u64 inode_id) |
| 53 | { | 53 | { |
| 54 | delayed_node->root = root; | 54 | delayed_node->root = root; |
| 55 | delayed_node->inode_id = inode_id; | 55 | delayed_node->inode_id = inode_id; |
| 56 | atomic_set(&delayed_node->refs, 0); | 56 | atomic_set(&delayed_node->refs, 0); |
| 57 | delayed_node->count = 0; | 57 | delayed_node->count = 0; |
| 58 | delayed_node->flags = 0; | 58 | delayed_node->flags = 0; |
| 59 | delayed_node->ins_root = RB_ROOT; | 59 | delayed_node->ins_root = RB_ROOT; |
| 60 | delayed_node->del_root = RB_ROOT; | 60 | delayed_node->del_root = RB_ROOT; |
| 61 | mutex_init(&delayed_node->mutex); | 61 | mutex_init(&delayed_node->mutex); |
| 62 | delayed_node->index_cnt = 0; | 62 | delayed_node->index_cnt = 0; |
| 63 | INIT_LIST_HEAD(&delayed_node->n_list); | 63 | INIT_LIST_HEAD(&delayed_node->n_list); |
| 64 | INIT_LIST_HEAD(&delayed_node->p_list); | 64 | INIT_LIST_HEAD(&delayed_node->p_list); |
| 65 | delayed_node->bytes_reserved = 0; | 65 | delayed_node->bytes_reserved = 0; |
| 66 | memset(&delayed_node->inode_item, 0, sizeof(delayed_node->inode_item)); | 66 | memset(&delayed_node->inode_item, 0, sizeof(delayed_node->inode_item)); |
| 67 | } | 67 | } |
| 68 | 68 | ||
| 69 | static inline int btrfs_is_continuous_delayed_item( | 69 | static inline int btrfs_is_continuous_delayed_item( |
| 70 | struct btrfs_delayed_item *item1, | 70 | struct btrfs_delayed_item *item1, |
| 71 | struct btrfs_delayed_item *item2) | 71 | struct btrfs_delayed_item *item2) |
| 72 | { | 72 | { |
| 73 | if (item1->key.type == BTRFS_DIR_INDEX_KEY && | 73 | if (item1->key.type == BTRFS_DIR_INDEX_KEY && |
| 74 | item1->key.objectid == item2->key.objectid && | 74 | item1->key.objectid == item2->key.objectid && |
| 75 | item1->key.type == item2->key.type && | 75 | item1->key.type == item2->key.type && |
| 76 | item1->key.offset + 1 == item2->key.offset) | 76 | item1->key.offset + 1 == item2->key.offset) |
| 77 | return 1; | 77 | return 1; |
| 78 | return 0; | 78 | return 0; |
| 79 | } | 79 | } |
| 80 | 80 | ||
| 81 | static inline struct btrfs_delayed_root *btrfs_get_delayed_root( | 81 | static inline struct btrfs_delayed_root *btrfs_get_delayed_root( |
| 82 | struct btrfs_root *root) | 82 | struct btrfs_root *root) |
| 83 | { | 83 | { |
| 84 | return root->fs_info->delayed_root; | 84 | return root->fs_info->delayed_root; |
| 85 | } | 85 | } |
| 86 | 86 | ||
| 87 | static struct btrfs_delayed_node *btrfs_get_delayed_node(struct inode *inode) | 87 | static struct btrfs_delayed_node *btrfs_get_delayed_node(struct inode *inode) |
| 88 | { | 88 | { |
| 89 | struct btrfs_inode *btrfs_inode = BTRFS_I(inode); | 89 | struct btrfs_inode *btrfs_inode = BTRFS_I(inode); |
| 90 | struct btrfs_root *root = btrfs_inode->root; | 90 | struct btrfs_root *root = btrfs_inode->root; |
| 91 | u64 ino = btrfs_ino(inode); | 91 | u64 ino = btrfs_ino(inode); |
| 92 | struct btrfs_delayed_node *node; | 92 | struct btrfs_delayed_node *node; |
| 93 | 93 | ||
| 94 | node = ACCESS_ONCE(btrfs_inode->delayed_node); | 94 | node = ACCESS_ONCE(btrfs_inode->delayed_node); |
| 95 | if (node) { | 95 | if (node) { |
| 96 | atomic_inc(&node->refs); | 96 | atomic_inc(&node->refs); |
| 97 | return node; | 97 | return node; |
| 98 | } | 98 | } |
| 99 | 99 | ||
| 100 | spin_lock(&root->inode_lock); | 100 | spin_lock(&root->inode_lock); |
| 101 | node = radix_tree_lookup(&root->delayed_nodes_tree, ino); | 101 | node = radix_tree_lookup(&root->delayed_nodes_tree, ino); |
| 102 | if (node) { | 102 | if (node) { |
| 103 | if (btrfs_inode->delayed_node) { | 103 | if (btrfs_inode->delayed_node) { |
| 104 | atomic_inc(&node->refs); /* can be accessed */ | 104 | atomic_inc(&node->refs); /* can be accessed */ |
| 105 | BUG_ON(btrfs_inode->delayed_node != node); | 105 | BUG_ON(btrfs_inode->delayed_node != node); |
| 106 | spin_unlock(&root->inode_lock); | 106 | spin_unlock(&root->inode_lock); |
| 107 | return node; | 107 | return node; |
| 108 | } | 108 | } |
| 109 | btrfs_inode->delayed_node = node; | 109 | btrfs_inode->delayed_node = node; |
| 110 | /* can be accessed and cached in the inode */ | 110 | /* can be accessed and cached in the inode */ |
| 111 | atomic_add(2, &node->refs); | 111 | atomic_add(2, &node->refs); |
| 112 | spin_unlock(&root->inode_lock); | 112 | spin_unlock(&root->inode_lock); |
| 113 | return node; | 113 | return node; |
| 114 | } | 114 | } |
| 115 | spin_unlock(&root->inode_lock); | 115 | spin_unlock(&root->inode_lock); |
| 116 | 116 | ||
| 117 | return NULL; | 117 | return NULL; |
| 118 | } | 118 | } |
| 119 | 119 | ||
| 120 | /* Will return either the node or PTR_ERR(-ENOMEM) */ | 120 | /* Will return either the node or PTR_ERR(-ENOMEM) */ |
| 121 | static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node( | 121 | static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node( |
| 122 | struct inode *inode) | 122 | struct inode *inode) |
| 123 | { | 123 | { |
| 124 | struct btrfs_delayed_node *node; | 124 | struct btrfs_delayed_node *node; |
| 125 | struct btrfs_inode *btrfs_inode = BTRFS_I(inode); | 125 | struct btrfs_inode *btrfs_inode = BTRFS_I(inode); |
| 126 | struct btrfs_root *root = btrfs_inode->root; | 126 | struct btrfs_root *root = btrfs_inode->root; |
| 127 | u64 ino = btrfs_ino(inode); | 127 | u64 ino = btrfs_ino(inode); |
| 128 | int ret; | 128 | int ret; |
| 129 | 129 | ||
| 130 | again: | 130 | again: |
| 131 | node = btrfs_get_delayed_node(inode); | 131 | node = btrfs_get_delayed_node(inode); |
| 132 | if (node) | 132 | if (node) |
| 133 | return node; | 133 | return node; |
| 134 | 134 | ||
| 135 | node = kmem_cache_alloc(delayed_node_cache, GFP_NOFS); | 135 | node = kmem_cache_alloc(delayed_node_cache, GFP_NOFS); |
| 136 | if (!node) | 136 | if (!node) |
| 137 | return ERR_PTR(-ENOMEM); | 137 | return ERR_PTR(-ENOMEM); |
| 138 | btrfs_init_delayed_node(node, root, ino); | 138 | btrfs_init_delayed_node(node, root, ino); |
| 139 | 139 | ||
| 140 | /* cached in the btrfs inode and can be accessed */ | 140 | /* cached in the btrfs inode and can be accessed */ |
| 141 | atomic_add(2, &node->refs); | 141 | atomic_add(2, &node->refs); |
| 142 | 142 | ||
| 143 | ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); | 143 | ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); |
| 144 | if (ret) { | 144 | if (ret) { |
| 145 | kmem_cache_free(delayed_node_cache, node); | 145 | kmem_cache_free(delayed_node_cache, node); |
| 146 | return ERR_PTR(ret); | 146 | return ERR_PTR(ret); |
| 147 | } | 147 | } |
| 148 | 148 | ||
| 149 | spin_lock(&root->inode_lock); | 149 | spin_lock(&root->inode_lock); |
| 150 | ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node); | 150 | ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node); |
| 151 | if (ret == -EEXIST) { | 151 | if (ret == -EEXIST) { |
| 152 | spin_unlock(&root->inode_lock); | 152 | spin_unlock(&root->inode_lock); |
| 153 | kmem_cache_free(delayed_node_cache, node); | 153 | kmem_cache_free(delayed_node_cache, node); |
| 154 | radix_tree_preload_end(); | 154 | radix_tree_preload_end(); |
| 155 | goto again; | 155 | goto again; |
| 156 | } | 156 | } |
| 157 | btrfs_inode->delayed_node = node; | 157 | btrfs_inode->delayed_node = node; |
| 158 | spin_unlock(&root->inode_lock); | 158 | spin_unlock(&root->inode_lock); |
| 159 | radix_tree_preload_end(); | 159 | radix_tree_preload_end(); |
| 160 | 160 | ||
| 161 | return node; | 161 | return node; |
| 162 | } | 162 | } |
| 163 | 163 | ||
| 164 | /* | 164 | /* |
| 165 | * Call it when holding delayed_node->mutex | 165 | * Call it when holding delayed_node->mutex |
| 166 | * | 166 | * |
| 167 | * If mod = 1, add this node into the prepared list. | 167 | * If mod = 1, add this node into the prepared list. |
| 168 | */ | 168 | */ |
| 169 | static void btrfs_queue_delayed_node(struct btrfs_delayed_root *root, | 169 | static void btrfs_queue_delayed_node(struct btrfs_delayed_root *root, |
| 170 | struct btrfs_delayed_node *node, | 170 | struct btrfs_delayed_node *node, |
| 171 | int mod) | 171 | int mod) |
| 172 | { | 172 | { |
| 173 | spin_lock(&root->lock); | 173 | spin_lock(&root->lock); |
| 174 | if (test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) { | 174 | if (test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) { |
| 175 | if (!list_empty(&node->p_list)) | 175 | if (!list_empty(&node->p_list)) |
| 176 | list_move_tail(&node->p_list, &root->prepare_list); | 176 | list_move_tail(&node->p_list, &root->prepare_list); |
| 177 | else if (mod) | 177 | else if (mod) |
| 178 | list_add_tail(&node->p_list, &root->prepare_list); | 178 | list_add_tail(&node->p_list, &root->prepare_list); |
| 179 | } else { | 179 | } else { |
| 180 | list_add_tail(&node->n_list, &root->node_list); | 180 | list_add_tail(&node->n_list, &root->node_list); |
| 181 | list_add_tail(&node->p_list, &root->prepare_list); | 181 | list_add_tail(&node->p_list, &root->prepare_list); |
| 182 | atomic_inc(&node->refs); /* inserted into list */ | 182 | atomic_inc(&node->refs); /* inserted into list */ |
| 183 | root->nodes++; | 183 | root->nodes++; |
| 184 | set_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags); | 184 | set_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags); |
| 185 | } | 185 | } |
| 186 | spin_unlock(&root->lock); | 186 | spin_unlock(&root->lock); |
| 187 | } | 187 | } |
| 188 | 188 | ||
| 189 | /* Call it when holding delayed_node->mutex */ | 189 | /* Call it when holding delayed_node->mutex */ |
| 190 | static void btrfs_dequeue_delayed_node(struct btrfs_delayed_root *root, | 190 | static void btrfs_dequeue_delayed_node(struct btrfs_delayed_root *root, |
| 191 | struct btrfs_delayed_node *node) | 191 | struct btrfs_delayed_node *node) |
| 192 | { | 192 | { |
| 193 | spin_lock(&root->lock); | 193 | spin_lock(&root->lock); |
| 194 | if (test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) { | 194 | if (test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) { |
| 195 | root->nodes--; | 195 | root->nodes--; |
| 196 | atomic_dec(&node->refs); /* not in the list */ | 196 | atomic_dec(&node->refs); /* not in the list */ |
| 197 | list_del_init(&node->n_list); | 197 | list_del_init(&node->n_list); |
| 198 | if (!list_empty(&node->p_list)) | 198 | if (!list_empty(&node->p_list)) |
| 199 | list_del_init(&node->p_list); | 199 | list_del_init(&node->p_list); |
| 200 | clear_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags); | 200 | clear_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags); |
| 201 | } | 201 | } |
| 202 | spin_unlock(&root->lock); | 202 | spin_unlock(&root->lock); |
| 203 | } | 203 | } |
| 204 | 204 | ||
| 205 | static struct btrfs_delayed_node *btrfs_first_delayed_node( | 205 | static struct btrfs_delayed_node *btrfs_first_delayed_node( |
| 206 | struct btrfs_delayed_root *delayed_root) | 206 | struct btrfs_delayed_root *delayed_root) |
| 207 | { | 207 | { |
| 208 | struct list_head *p; | 208 | struct list_head *p; |
| 209 | struct btrfs_delayed_node *node = NULL; | 209 | struct btrfs_delayed_node *node = NULL; |
| 210 | 210 | ||
| 211 | spin_lock(&delayed_root->lock); | 211 | spin_lock(&delayed_root->lock); |
| 212 | if (list_empty(&delayed_root->node_list)) | 212 | if (list_empty(&delayed_root->node_list)) |
| 213 | goto out; | 213 | goto out; |
| 214 | 214 | ||
| 215 | p = delayed_root->node_list.next; | 215 | p = delayed_root->node_list.next; |
| 216 | node = list_entry(p, struct btrfs_delayed_node, n_list); | 216 | node = list_entry(p, struct btrfs_delayed_node, n_list); |
| 217 | atomic_inc(&node->refs); | 217 | atomic_inc(&node->refs); |
| 218 | out: | 218 | out: |
| 219 | spin_unlock(&delayed_root->lock); | 219 | spin_unlock(&delayed_root->lock); |
| 220 | 220 | ||
| 221 | return node; | 221 | return node; |
| 222 | } | 222 | } |
| 223 | 223 | ||
| 224 | static struct btrfs_delayed_node *btrfs_next_delayed_node( | 224 | static struct btrfs_delayed_node *btrfs_next_delayed_node( |
| 225 | struct btrfs_delayed_node *node) | 225 | struct btrfs_delayed_node *node) |
| 226 | { | 226 | { |
| 227 | struct btrfs_delayed_root *delayed_root; | 227 | struct btrfs_delayed_root *delayed_root; |
| 228 | struct list_head *p; | 228 | struct list_head *p; |
| 229 | struct btrfs_delayed_node *next = NULL; | 229 | struct btrfs_delayed_node *next = NULL; |
| 230 | 230 | ||
| 231 | delayed_root = node->root->fs_info->delayed_root; | 231 | delayed_root = node->root->fs_info->delayed_root; |
| 232 | spin_lock(&delayed_root->lock); | 232 | spin_lock(&delayed_root->lock); |
| 233 | if (!test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) { | 233 | if (!test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) { |
| 234 | /* not in the list */ | 234 | /* not in the list */ |
| 235 | if (list_empty(&delayed_root->node_list)) | 235 | if (list_empty(&delayed_root->node_list)) |
| 236 | goto out; | 236 | goto out; |
| 237 | p = delayed_root->node_list.next; | 237 | p = delayed_root->node_list.next; |
| 238 | } else if (list_is_last(&node->n_list, &delayed_root->node_list)) | 238 | } else if (list_is_last(&node->n_list, &delayed_root->node_list)) |
| 239 | goto out; | 239 | goto out; |
| 240 | else | 240 | else |
| 241 | p = node->n_list.next; | 241 | p = node->n_list.next; |
| 242 | 242 | ||
| 243 | next = list_entry(p, struct btrfs_delayed_node, n_list); | 243 | next = list_entry(p, struct btrfs_delayed_node, n_list); |
| 244 | atomic_inc(&next->refs); | 244 | atomic_inc(&next->refs); |
| 245 | out: | 245 | out: |
| 246 | spin_unlock(&delayed_root->lock); | 246 | spin_unlock(&delayed_root->lock); |
| 247 | 247 | ||
| 248 | return next; | 248 | return next; |
| 249 | } | 249 | } |
| 250 | 250 | ||
| 251 | static void __btrfs_release_delayed_node( | 251 | static void __btrfs_release_delayed_node( |
| 252 | struct btrfs_delayed_node *delayed_node, | 252 | struct btrfs_delayed_node *delayed_node, |
| 253 | int mod) | 253 | int mod) |
| 254 | { | 254 | { |
| 255 | struct btrfs_delayed_root *delayed_root; | 255 | struct btrfs_delayed_root *delayed_root; |
| 256 | 256 | ||
| 257 | if (!delayed_node) | 257 | if (!delayed_node) |
| 258 | return; | 258 | return; |
| 259 | 259 | ||
| 260 | delayed_root = delayed_node->root->fs_info->delayed_root; | 260 | delayed_root = delayed_node->root->fs_info->delayed_root; |
| 261 | 261 | ||
| 262 | mutex_lock(&delayed_node->mutex); | 262 | mutex_lock(&delayed_node->mutex); |
| 263 | if (delayed_node->count) | 263 | if (delayed_node->count) |
| 264 | btrfs_queue_delayed_node(delayed_root, delayed_node, mod); | 264 | btrfs_queue_delayed_node(delayed_root, delayed_node, mod); |
| 265 | else | 265 | else |
| 266 | btrfs_dequeue_delayed_node(delayed_root, delayed_node); | 266 | btrfs_dequeue_delayed_node(delayed_root, delayed_node); |
| 267 | mutex_unlock(&delayed_node->mutex); | 267 | mutex_unlock(&delayed_node->mutex); |
| 268 | 268 | ||
| 269 | if (atomic_dec_and_test(&delayed_node->refs)) { | 269 | if (atomic_dec_and_test(&delayed_node->refs)) { |
| 270 | bool free = false; | 270 | bool free = false; |
| 271 | struct btrfs_root *root = delayed_node->root; | 271 | struct btrfs_root *root = delayed_node->root; |
| 272 | spin_lock(&root->inode_lock); | 272 | spin_lock(&root->inode_lock); |
| 273 | if (atomic_read(&delayed_node->refs) == 0) { | 273 | if (atomic_read(&delayed_node->refs) == 0) { |
| 274 | radix_tree_delete(&root->delayed_nodes_tree, | 274 | radix_tree_delete(&root->delayed_nodes_tree, |
| 275 | delayed_node->inode_id); | 275 | delayed_node->inode_id); |
| 276 | free = true; | 276 | free = true; |
| 277 | } | 277 | } |
| 278 | spin_unlock(&root->inode_lock); | 278 | spin_unlock(&root->inode_lock); |
| 279 | if (free) | 279 | if (free) |
| 280 | kmem_cache_free(delayed_node_cache, delayed_node); | 280 | kmem_cache_free(delayed_node_cache, delayed_node); |
| 281 | } | 281 | } |
| 282 | } | 282 | } |
| 283 | 283 | ||
| 284 | static inline void btrfs_release_delayed_node(struct btrfs_delayed_node *node) | 284 | static inline void btrfs_release_delayed_node(struct btrfs_delayed_node *node) |
| 285 | { | 285 | { |
| 286 | __btrfs_release_delayed_node(node, 0); | 286 | __btrfs_release_delayed_node(node, 0); |
| 287 | } | 287 | } |
| 288 | 288 | ||
| 289 | static struct btrfs_delayed_node *btrfs_first_prepared_delayed_node( | 289 | static struct btrfs_delayed_node *btrfs_first_prepared_delayed_node( |
| 290 | struct btrfs_delayed_root *delayed_root) | 290 | struct btrfs_delayed_root *delayed_root) |
| 291 | { | 291 | { |
| 292 | struct list_head *p; | 292 | struct list_head *p; |
| 293 | struct btrfs_delayed_node *node = NULL; | 293 | struct btrfs_delayed_node *node = NULL; |
| 294 | 294 | ||
| 295 | spin_lock(&delayed_root->lock); | 295 | spin_lock(&delayed_root->lock); |
| 296 | if (list_empty(&delayed_root->prepare_list)) | 296 | if (list_empty(&delayed_root->prepare_list)) |
| 297 | goto out; | 297 | goto out; |
| 298 | 298 | ||
| 299 | p = delayed_root->prepare_list.next; | 299 | p = delayed_root->prepare_list.next; |
| 300 | list_del_init(p); | 300 | list_del_init(p); |
| 301 | node = list_entry(p, struct btrfs_delayed_node, p_list); | 301 | node = list_entry(p, struct btrfs_delayed_node, p_list); |
| 302 | atomic_inc(&node->refs); | 302 | atomic_inc(&node->refs); |
| 303 | out: | 303 | out: |
| 304 | spin_unlock(&delayed_root->lock); | 304 | spin_unlock(&delayed_root->lock); |
| 305 | 305 | ||
| 306 | return node; | 306 | return node; |
| 307 | } | 307 | } |
| 308 | 308 | ||
| 309 | static inline void btrfs_release_prepared_delayed_node( | 309 | static inline void btrfs_release_prepared_delayed_node( |
| 310 | struct btrfs_delayed_node *node) | 310 | struct btrfs_delayed_node *node) |
| 311 | { | 311 | { |
| 312 | __btrfs_release_delayed_node(node, 1); | 312 | __btrfs_release_delayed_node(node, 1); |
| 313 | } | 313 | } |
| 314 | 314 | ||
| 315 | static struct btrfs_delayed_item *btrfs_alloc_delayed_item(u32 data_len) | 315 | static struct btrfs_delayed_item *btrfs_alloc_delayed_item(u32 data_len) |
| 316 | { | 316 | { |
| 317 | struct btrfs_delayed_item *item; | 317 | struct btrfs_delayed_item *item; |
| 318 | item = kmalloc(sizeof(*item) + data_len, GFP_NOFS); | 318 | item = kmalloc(sizeof(*item) + data_len, GFP_NOFS); |
| 319 | if (item) { | 319 | if (item) { |
| 320 | item->data_len = data_len; | 320 | item->data_len = data_len; |
| 321 | item->ins_or_del = 0; | 321 | item->ins_or_del = 0; |
| 322 | item->bytes_reserved = 0; | 322 | item->bytes_reserved = 0; |
| 323 | item->delayed_node = NULL; | 323 | item->delayed_node = NULL; |
| 324 | atomic_set(&item->refs, 1); | 324 | atomic_set(&item->refs, 1); |
| 325 | } | 325 | } |
| 326 | return item; | 326 | return item; |
| 327 | } | 327 | } |
| 328 | 328 | ||
| 329 | /* | 329 | /* |
| 330 | * __btrfs_lookup_delayed_item - look up the delayed item by key | 330 | * __btrfs_lookup_delayed_item - look up the delayed item by key |
| 331 | * @delayed_node: pointer to the delayed node | 331 | * @delayed_node: pointer to the delayed node |
| 332 | * @key: the key to look up | 332 | * @key: the key to look up |
| 333 | * @prev: used to store the prev item if the right item isn't found | 333 | * @prev: used to store the prev item if the right item isn't found |
| 334 | * @next: used to store the next item if the right item isn't found | 334 | * @next: used to store the next item if the right item isn't found |
| 335 | * | 335 | * |
| 336 | * Note: if we don't find the right item, we will return the prev item and | 336 | * Note: if we don't find the right item, we will return the prev item and |
| 337 | * the next item. | 337 | * the next item. |
| 338 | */ | 338 | */ |
| 339 | static struct btrfs_delayed_item *__btrfs_lookup_delayed_item( | 339 | static struct btrfs_delayed_item *__btrfs_lookup_delayed_item( |
| 340 | struct rb_root *root, | 340 | struct rb_root *root, |
| 341 | struct btrfs_key *key, | 341 | struct btrfs_key *key, |
| 342 | struct btrfs_delayed_item **prev, | 342 | struct btrfs_delayed_item **prev, |
| 343 | struct btrfs_delayed_item **next) | 343 | struct btrfs_delayed_item **next) |
| 344 | { | 344 | { |
| 345 | struct rb_node *node, *prev_node = NULL; | 345 | struct rb_node *node, *prev_node = NULL; |
| 346 | struct btrfs_delayed_item *delayed_item = NULL; | 346 | struct btrfs_delayed_item *delayed_item = NULL; |
| 347 | int ret = 0; | 347 | int ret = 0; |
| 348 | 348 | ||
| 349 | node = root->rb_node; | 349 | node = root->rb_node; |
| 350 | 350 | ||
| 351 | while (node) { | 351 | while (node) { |
| 352 | delayed_item = rb_entry(node, struct btrfs_delayed_item, | 352 | delayed_item = rb_entry(node, struct btrfs_delayed_item, |
| 353 | rb_node); | 353 | rb_node); |
| 354 | prev_node = node; | 354 | prev_node = node; |
| 355 | ret = btrfs_comp_cpu_keys(&delayed_item->key, key); | 355 | ret = btrfs_comp_cpu_keys(&delayed_item->key, key); |
| 356 | if (ret < 0) | 356 | if (ret < 0) |
| 357 | node = node->rb_right; | 357 | node = node->rb_right; |
| 358 | else if (ret > 0) | 358 | else if (ret > 0) |
| 359 | node = node->rb_left; | 359 | node = node->rb_left; |
| 360 | else | 360 | else |
| 361 | return delayed_item; | 361 | return delayed_item; |
| 362 | } | 362 | } |
| 363 | 363 | ||
| 364 | if (prev) { | 364 | if (prev) { |
| 365 | if (!prev_node) | 365 | if (!prev_node) |
| 366 | *prev = NULL; | 366 | *prev = NULL; |
| 367 | else if (ret < 0) | 367 | else if (ret < 0) |
| 368 | *prev = delayed_item; | 368 | *prev = delayed_item; |
| 369 | else if ((node = rb_prev(prev_node)) != NULL) { | 369 | else if ((node = rb_prev(prev_node)) != NULL) { |
| 370 | *prev = rb_entry(node, struct btrfs_delayed_item, | 370 | *prev = rb_entry(node, struct btrfs_delayed_item, |
| 371 | rb_node); | 371 | rb_node); |
| 372 | } else | 372 | } else |
| 373 | *prev = NULL; | 373 | *prev = NULL; |
| 374 | } | 374 | } |
| 375 | 375 | ||
| 376 | if (next) { | 376 | if (next) { |
| 377 | if (!prev_node) | 377 | if (!prev_node) |
| 378 | *next = NULL; | 378 | *next = NULL; |
| 379 | else if (ret > 0) | 379 | else if (ret > 0) |
| 380 | *next = delayed_item; | 380 | *next = delayed_item; |
| 381 | else if ((node = rb_next(prev_node)) != NULL) { | 381 | else if ((node = rb_next(prev_node)) != NULL) { |
| 382 | *next = rb_entry(node, struct btrfs_delayed_item, | 382 | *next = rb_entry(node, struct btrfs_delayed_item, |
| 383 | rb_node); | 383 | rb_node); |
| 384 | } else | 384 | } else |
| 385 | *next = NULL; | 385 | *next = NULL; |
| 386 | } | 386 | } |
| 387 | return NULL; | 387 | return NULL; |
| 388 | } | 388 | } |
| 389 | 389 | ||
| 390 | static struct btrfs_delayed_item *__btrfs_lookup_delayed_insertion_item( | 390 | static struct btrfs_delayed_item *__btrfs_lookup_delayed_insertion_item( |
| 391 | struct btrfs_delayed_node *delayed_node, | 391 | struct btrfs_delayed_node *delayed_node, |
| 392 | struct btrfs_key *key) | 392 | struct btrfs_key *key) |
| 393 | { | 393 | { |
| 394 | struct btrfs_delayed_item *item; | 394 | struct btrfs_delayed_item *item; |
| 395 | 395 | ||
| 396 | item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key, | 396 | item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key, |
| 397 | NULL, NULL); | 397 | NULL, NULL); |
| 398 | return item; | 398 | return item; |
| 399 | } | 399 | } |
| 400 | 400 | ||
| 401 | static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node, | 401 | static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node, |
| 402 | struct btrfs_delayed_item *ins, | 402 | struct btrfs_delayed_item *ins, |
| 403 | int action) | 403 | int action) |
| 404 | { | 404 | { |
| 405 | struct rb_node **p, *node; | 405 | struct rb_node **p, *node; |
| 406 | struct rb_node *parent_node = NULL; | 406 | struct rb_node *parent_node = NULL; |
| 407 | struct rb_root *root; | 407 | struct rb_root *root; |
| 408 | struct btrfs_delayed_item *item; | 408 | struct btrfs_delayed_item *item; |
| 409 | int cmp; | 409 | int cmp; |
| 410 | 410 | ||
| 411 | if (action == BTRFS_DELAYED_INSERTION_ITEM) | 411 | if (action == BTRFS_DELAYED_INSERTION_ITEM) |
| 412 | root = &delayed_node->ins_root; | 412 | root = &delayed_node->ins_root; |
| 413 | else if (action == BTRFS_DELAYED_DELETION_ITEM) | 413 | else if (action == BTRFS_DELAYED_DELETION_ITEM) |
| 414 | root = &delayed_node->del_root; | 414 | root = &delayed_node->del_root; |
| 415 | else | 415 | else |
| 416 | BUG(); | 416 | BUG(); |
| 417 | p = &root->rb_node; | 417 | p = &root->rb_node; |
| 418 | node = &ins->rb_node; | 418 | node = &ins->rb_node; |
| 419 | 419 | ||
| 420 | while (*p) { | 420 | while (*p) { |
| 421 | parent_node = *p; | 421 | parent_node = *p; |
| 422 | item = rb_entry(parent_node, struct btrfs_delayed_item, | 422 | item = rb_entry(parent_node, struct btrfs_delayed_item, |
| 423 | rb_node); | 423 | rb_node); |
| 424 | 424 | ||
| 425 | cmp = btrfs_comp_cpu_keys(&item->key, &ins->key); | 425 | cmp = btrfs_comp_cpu_keys(&item->key, &ins->key); |
| 426 | if (cmp < 0) | 426 | if (cmp < 0) |
| 427 | p = &(*p)->rb_right; | 427 | p = &(*p)->rb_right; |
| 428 | else if (cmp > 0) | 428 | else if (cmp > 0) |
| 429 | p = &(*p)->rb_left; | 429 | p = &(*p)->rb_left; |
| 430 | else | 430 | else |
| 431 | return -EEXIST; | 431 | return -EEXIST; |
| 432 | } | 432 | } |
| 433 | 433 | ||
| 434 | rb_link_node(node, parent_node, p); | 434 | rb_link_node(node, parent_node, p); |
| 435 | rb_insert_color(node, root); | 435 | rb_insert_color(node, root); |
| 436 | ins->delayed_node = delayed_node; | 436 | ins->delayed_node = delayed_node; |
| 437 | ins->ins_or_del = action; | 437 | ins->ins_or_del = action; |
| 438 | 438 | ||
| 439 | if (ins->key.type == BTRFS_DIR_INDEX_KEY && | 439 | if (ins->key.type == BTRFS_DIR_INDEX_KEY && |
| 440 | action == BTRFS_DELAYED_INSERTION_ITEM && | 440 | action == BTRFS_DELAYED_INSERTION_ITEM && |
| 441 | ins->key.offset >= delayed_node->index_cnt) | 441 | ins->key.offset >= delayed_node->index_cnt) |
| 442 | delayed_node->index_cnt = ins->key.offset + 1; | 442 | delayed_node->index_cnt = ins->key.offset + 1; |
| 443 | 443 | ||
| 444 | delayed_node->count++; | 444 | delayed_node->count++; |
| 445 | atomic_inc(&delayed_node->root->fs_info->delayed_root->items); | 445 | atomic_inc(&delayed_node->root->fs_info->delayed_root->items); |
| 446 | return 0; | 446 | return 0; |
| 447 | } | 447 | } |
| 448 | 448 | ||
| 449 | static int __btrfs_add_delayed_insertion_item(struct btrfs_delayed_node *node, | 449 | static int __btrfs_add_delayed_insertion_item(struct btrfs_delayed_node *node, |
| 450 | struct btrfs_delayed_item *item) | 450 | struct btrfs_delayed_item *item) |
| 451 | { | 451 | { |
| 452 | return __btrfs_add_delayed_item(node, item, | 452 | return __btrfs_add_delayed_item(node, item, |
| 453 | BTRFS_DELAYED_INSERTION_ITEM); | 453 | BTRFS_DELAYED_INSERTION_ITEM); |
| 454 | } | 454 | } |
| 455 | 455 | ||
| 456 | static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node, | 456 | static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node, |
| 457 | struct btrfs_delayed_item *item) | 457 | struct btrfs_delayed_item *item) |
| 458 | { | 458 | { |
| 459 | return __btrfs_add_delayed_item(node, item, | 459 | return __btrfs_add_delayed_item(node, item, |
| 460 | BTRFS_DELAYED_DELETION_ITEM); | 460 | BTRFS_DELAYED_DELETION_ITEM); |
| 461 | } | 461 | } |
| 462 | 462 | ||
| 463 | static void finish_one_item(struct btrfs_delayed_root *delayed_root) | 463 | static void finish_one_item(struct btrfs_delayed_root *delayed_root) |
| 464 | { | 464 | { |
| 465 | int seq = atomic_inc_return(&delayed_root->items_seq); | 465 | int seq = atomic_inc_return(&delayed_root->items_seq); |
| 466 | if ((atomic_dec_return(&delayed_root->items) < | 466 | if ((atomic_dec_return(&delayed_root->items) < |
| 467 | BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0) && | 467 | BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0) && |
| 468 | waitqueue_active(&delayed_root->wait)) | 468 | waitqueue_active(&delayed_root->wait)) |
| 469 | wake_up(&delayed_root->wait); | 469 | wake_up(&delayed_root->wait); |
| 470 | } | 470 | } |
| 471 | 471 | ||
| 472 | static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) | 472 | static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) |
| 473 | { | 473 | { |
| 474 | struct rb_root *root; | 474 | struct rb_root *root; |
| 475 | struct btrfs_delayed_root *delayed_root; | 475 | struct btrfs_delayed_root *delayed_root; |
| 476 | 476 | ||
| 477 | delayed_root = delayed_item->delayed_node->root->fs_info->delayed_root; | 477 | delayed_root = delayed_item->delayed_node->root->fs_info->delayed_root; |
| 478 | 478 | ||
| 479 | BUG_ON(!delayed_root); | 479 | BUG_ON(!delayed_root); |
| 480 | BUG_ON(delayed_item->ins_or_del != BTRFS_DELAYED_DELETION_ITEM && | 480 | BUG_ON(delayed_item->ins_or_del != BTRFS_DELAYED_DELETION_ITEM && |
| 481 | delayed_item->ins_or_del != BTRFS_DELAYED_INSERTION_ITEM); | 481 | delayed_item->ins_or_del != BTRFS_DELAYED_INSERTION_ITEM); |
| 482 | 482 | ||
| 483 | if (delayed_item->ins_or_del == BTRFS_DELAYED_INSERTION_ITEM) | 483 | if (delayed_item->ins_or_del == BTRFS_DELAYED_INSERTION_ITEM) |
| 484 | root = &delayed_item->delayed_node->ins_root; | 484 | root = &delayed_item->delayed_node->ins_root; |
| 485 | else | 485 | else |
| 486 | root = &delayed_item->delayed_node->del_root; | 486 | root = &delayed_item->delayed_node->del_root; |
| 487 | 487 | ||
| 488 | rb_erase(&delayed_item->rb_node, root); | 488 | rb_erase(&delayed_item->rb_node, root); |
| 489 | delayed_item->delayed_node->count--; | 489 | delayed_item->delayed_node->count--; |
| 490 | 490 | ||
| 491 | finish_one_item(delayed_root); | 491 | finish_one_item(delayed_root); |
| 492 | } | 492 | } |
| 493 | 493 | ||
| 494 | static void btrfs_release_delayed_item(struct btrfs_delayed_item *item) | 494 | static void btrfs_release_delayed_item(struct btrfs_delayed_item *item) |
| 495 | { | 495 | { |
| 496 | if (item) { | 496 | if (item) { |
| 497 | __btrfs_remove_delayed_item(item); | 497 | __btrfs_remove_delayed_item(item); |
| 498 | if (atomic_dec_and_test(&item->refs)) | 498 | if (atomic_dec_and_test(&item->refs)) |
| 499 | kfree(item); | 499 | kfree(item); |
| 500 | } | 500 | } |
| 501 | } | 501 | } |
| 502 | 502 | ||
| 503 | static struct btrfs_delayed_item *__btrfs_first_delayed_insertion_item( | 503 | static struct btrfs_delayed_item *__btrfs_first_delayed_insertion_item( |
| 504 | struct btrfs_delayed_node *delayed_node) | 504 | struct btrfs_delayed_node *delayed_node) |
| 505 | { | 505 | { |
| 506 | struct rb_node *p; | 506 | struct rb_node *p; |
| 507 | struct btrfs_delayed_item *item = NULL; | 507 | struct btrfs_delayed_item *item = NULL; |
| 508 | 508 | ||
| 509 | p = rb_first(&delayed_node->ins_root); | 509 | p = rb_first(&delayed_node->ins_root); |
| 510 | if (p) | 510 | if (p) |
| 511 | item = rb_entry(p, struct btrfs_delayed_item, rb_node); | 511 | item = rb_entry(p, struct btrfs_delayed_item, rb_node); |
| 512 | 512 | ||
| 513 | return item; | 513 | return item; |
| 514 | } | 514 | } |
| 515 | 515 | ||
| 516 | static struct btrfs_delayed_item *__btrfs_first_delayed_deletion_item( | 516 | static struct btrfs_delayed_item *__btrfs_first_delayed_deletion_item( |
| 517 | struct btrfs_delayed_node *delayed_node) | 517 | struct btrfs_delayed_node *delayed_node) |
| 518 | { | 518 | { |
| 519 | struct rb_node *p; | 519 | struct rb_node *p; |
| 520 | struct btrfs_delayed_item *item = NULL; | 520 | struct btrfs_delayed_item *item = NULL; |
| 521 | 521 | ||
| 522 | p = rb_first(&delayed_node->del_root); | 522 | p = rb_first(&delayed_node->del_root); |
| 523 | if (p) | 523 | if (p) |
| 524 | item = rb_entry(p, struct btrfs_delayed_item, rb_node); | 524 | item = rb_entry(p, struct btrfs_delayed_item, rb_node); |
| 525 | 525 | ||
| 526 | return item; | 526 | return item; |
| 527 | } | 527 | } |
| 528 | 528 | ||
| 529 | static struct btrfs_delayed_item *__btrfs_next_delayed_item( | 529 | static struct btrfs_delayed_item *__btrfs_next_delayed_item( |
| 530 | struct btrfs_delayed_item *item) | 530 | struct btrfs_delayed_item *item) |
| 531 | { | 531 | { |
| 532 | struct rb_node *p; | 532 | struct rb_node *p; |
| 533 | struct btrfs_delayed_item *next = NULL; | 533 | struct btrfs_delayed_item *next = NULL; |
| 534 | 534 | ||
| 535 | p = rb_next(&item->rb_node); | 535 | p = rb_next(&item->rb_node); |
| 536 | if (p) | 536 | if (p) |
| 537 | next = rb_entry(p, struct btrfs_delayed_item, rb_node); | 537 | next = rb_entry(p, struct btrfs_delayed_item, rb_node); |
| 538 | 538 | ||
| 539 | return next; | 539 | return next; |
| 540 | } | 540 | } |
| 541 | 541 | ||
| 542 | static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, | 542 | static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, |
| 543 | struct btrfs_root *root, | 543 | struct btrfs_root *root, |
| 544 | struct btrfs_delayed_item *item) | 544 | struct btrfs_delayed_item *item) |
| 545 | { | 545 | { |
| 546 | struct btrfs_block_rsv *src_rsv; | 546 | struct btrfs_block_rsv *src_rsv; |
| 547 | struct btrfs_block_rsv *dst_rsv; | 547 | struct btrfs_block_rsv *dst_rsv; |
| 548 | u64 num_bytes; | 548 | u64 num_bytes; |
| 549 | int ret; | 549 | int ret; |
| 550 | 550 | ||
| 551 | if (!trans->bytes_reserved) | 551 | if (!trans->bytes_reserved) |
| 552 | return 0; | 552 | return 0; |
| 553 | 553 | ||
| 554 | src_rsv = trans->block_rsv; | 554 | src_rsv = trans->block_rsv; |
| 555 | dst_rsv = &root->fs_info->delayed_block_rsv; | 555 | dst_rsv = &root->fs_info->delayed_block_rsv; |
| 556 | 556 | ||
| 557 | num_bytes = btrfs_calc_trans_metadata_size(root, 1); | 557 | num_bytes = btrfs_calc_trans_metadata_size(root, 1); |
| 558 | ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); | 558 | ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); |
| 559 | if (!ret) { | 559 | if (!ret) { |
| 560 | trace_btrfs_space_reservation(root->fs_info, "delayed_item", | 560 | trace_btrfs_space_reservation(root->fs_info, "delayed_item", |
| 561 | item->key.objectid, | 561 | item->key.objectid, |
| 562 | num_bytes, 1); | 562 | num_bytes, 1); |
| 563 | item->bytes_reserved = num_bytes; | 563 | item->bytes_reserved = num_bytes; |
| 564 | } | 564 | } |
| 565 | 565 | ||
| 566 | return ret; | 566 | return ret; |
| 567 | } | 567 | } |
| 568 | 568 | ||
| 569 | static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, | 569 | static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, |
| 570 | struct btrfs_delayed_item *item) | 570 | struct btrfs_delayed_item *item) |
| 571 | { | 571 | { |
| 572 | struct btrfs_block_rsv *rsv; | 572 | struct btrfs_block_rsv *rsv; |
| 573 | 573 | ||
| 574 | if (!item->bytes_reserved) | 574 | if (!item->bytes_reserved) |
| 575 | return; | 575 | return; |
| 576 | 576 | ||
| 577 | rsv = &root->fs_info->delayed_block_rsv; | 577 | rsv = &root->fs_info->delayed_block_rsv; |
| 578 | trace_btrfs_space_reservation(root->fs_info, "delayed_item", | 578 | trace_btrfs_space_reservation(root->fs_info, "delayed_item", |
| 579 | item->key.objectid, item->bytes_reserved, | 579 | item->key.objectid, item->bytes_reserved, |
| 580 | 0); | 580 | 0); |
| 581 | btrfs_block_rsv_release(root, rsv, | 581 | btrfs_block_rsv_release(root, rsv, |
| 582 | item->bytes_reserved); | 582 | item->bytes_reserved); |
| 583 | } | 583 | } |
| 584 | 584 | ||
| 585 | static int btrfs_delayed_inode_reserve_metadata( | 585 | static int btrfs_delayed_inode_reserve_metadata( |
| 586 | struct btrfs_trans_handle *trans, | 586 | struct btrfs_trans_handle *trans, |
| 587 | struct btrfs_root *root, | 587 | struct btrfs_root *root, |
| 588 | struct inode *inode, | 588 | struct inode *inode, |
| 589 | struct btrfs_delayed_node *node) | 589 | struct btrfs_delayed_node *node) |
| 590 | { | 590 | { |
| 591 | struct btrfs_block_rsv *src_rsv; | 591 | struct btrfs_block_rsv *src_rsv; |
| 592 | struct btrfs_block_rsv *dst_rsv; | 592 | struct btrfs_block_rsv *dst_rsv; |
| 593 | u64 num_bytes; | 593 | u64 num_bytes; |
| 594 | int ret; | 594 | int ret; |
| 595 | bool release = false; | 595 | bool release = false; |
| 596 | 596 | ||
| 597 | src_rsv = trans->block_rsv; | 597 | src_rsv = trans->block_rsv; |
| 598 | dst_rsv = &root->fs_info->delayed_block_rsv; | 598 | dst_rsv = &root->fs_info->delayed_block_rsv; |
| 599 | 599 | ||
| 600 | num_bytes = btrfs_calc_trans_metadata_size(root, 1); | 600 | num_bytes = btrfs_calc_trans_metadata_size(root, 1); |
| 601 | 601 | ||
| 602 | /* | 602 | /* |
| 603 | * btrfs_dirty_inode will update the inode under btrfs_join_transaction | 603 | * btrfs_dirty_inode will update the inode under btrfs_join_transaction |
| 604 | * which doesn't reserve space for speed. This is a problem since we | 604 | * which doesn't reserve space for speed. This is a problem since we |
| 605 | * still need to reserve space for this update, so try to reserve the | 605 | * still need to reserve space for this update, so try to reserve the |
| 606 | * space. | 606 | * space. |
| 607 | * | 607 | * |
| 608 | * Now if src_rsv == delalloc_block_rsv we'll let it just steal since | 608 | * Now if src_rsv == delalloc_block_rsv we'll let it just steal since |
| 609 | * we're accounted for. | 609 | * we're accounted for. |
| 610 | */ | 610 | */ |
| 611 | if (!src_rsv || (!trans->bytes_reserved && | 611 | if (!src_rsv || (!trans->bytes_reserved && |
| 612 | src_rsv->type != BTRFS_BLOCK_RSV_DELALLOC)) { | 612 | src_rsv->type != BTRFS_BLOCK_RSV_DELALLOC)) { |
| 613 | ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes, | 613 | ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes, |
| 614 | BTRFS_RESERVE_NO_FLUSH); | 614 | BTRFS_RESERVE_NO_FLUSH); |
| 615 | /* | 615 | /* |
| 616 | * Since we're under a transaction reserve_metadata_bytes could | 616 | * Since we're under a transaction reserve_metadata_bytes could |
| 617 | * try to commit the transaction which will make it return | 617 | * try to commit the transaction which will make it return |
| 618 | * EAGAIN to make us stop the transaction we have, so return | 618 | * EAGAIN to make us stop the transaction we have, so return |
| 619 | * ENOSPC instead so that btrfs_dirty_inode knows what to do. | 619 | * ENOSPC instead so that btrfs_dirty_inode knows what to do. |
| 620 | */ | 620 | */ |
| 621 | if (ret == -EAGAIN) | 621 | if (ret == -EAGAIN) |
| 622 | ret = -ENOSPC; | 622 | ret = -ENOSPC; |
| 623 | if (!ret) { | 623 | if (!ret) { |
| 624 | node->bytes_reserved = num_bytes; | 624 | node->bytes_reserved = num_bytes; |
| 625 | trace_btrfs_space_reservation(root->fs_info, | 625 | trace_btrfs_space_reservation(root->fs_info, |
| 626 | "delayed_inode", | 626 | "delayed_inode", |
| 627 | btrfs_ino(inode), | 627 | btrfs_ino(inode), |
| 628 | num_bytes, 1); | 628 | num_bytes, 1); |
| 629 | } | 629 | } |
| 630 | return ret; | 630 | return ret; |
| 631 | } else if (src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) { | 631 | } else if (src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) { |
| 632 | spin_lock(&BTRFS_I(inode)->lock); | 632 | spin_lock(&BTRFS_I(inode)->lock); |
| 633 | if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, | 633 | if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, |
| 634 | &BTRFS_I(inode)->runtime_flags)) { | 634 | &BTRFS_I(inode)->runtime_flags)) { |
| 635 | spin_unlock(&BTRFS_I(inode)->lock); | 635 | spin_unlock(&BTRFS_I(inode)->lock); |
| 636 | release = true; | 636 | release = true; |
| 637 | goto migrate; | 637 | goto migrate; |
| 638 | } | 638 | } |
| 639 | spin_unlock(&BTRFS_I(inode)->lock); | 639 | spin_unlock(&BTRFS_I(inode)->lock); |
| 640 | 640 | ||
| 641 | /* Ok we didn't have space pre-reserved. This shouldn't happen | 641 | /* Ok we didn't have space pre-reserved. This shouldn't happen |
| 642 | * too often but it can happen if we do delalloc to an existing | 642 | * too often but it can happen if we do delalloc to an existing |
| 643 | * inode which gets dirtied because of the time update, and then | 643 | * inode which gets dirtied because of the time update, and then |
| 644 | * isn't touched again until after the transaction commits and | 644 | * isn't touched again until after the transaction commits and |
| 645 | * then we try to write out the data. First try to be nice and | 645 | * then we try to write out the data. First try to be nice and |
| 646 | * reserve something strictly for us. If not be a pain and try | 646 | * reserve something strictly for us. If not be a pain and try |
| 647 | * to steal from the delalloc block rsv. | 647 | * to steal from the delalloc block rsv. |
| 648 | */ | 648 | */ |
| 649 | ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes, | 649 | ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes, |
| 650 | BTRFS_RESERVE_NO_FLUSH); | 650 | BTRFS_RESERVE_NO_FLUSH); |
| 651 | if (!ret) | 651 | if (!ret) |
| 652 | goto out; | 652 | goto out; |
| 653 | 653 | ||
| 654 | ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); | 654 | ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); |
| 655 | if (!WARN_ON(ret)) | 655 | if (!WARN_ON(ret)) |
| 656 | goto out; | 656 | goto out; |
| 657 | 657 | ||
| 658 | /* | 658 | /* |
| 659 | * Ok this is a problem, let's just steal from the global rsv | 659 | * Ok this is a problem, let's just steal from the global rsv |
| 660 | * since this really shouldn't happen that often. | 660 | * since this really shouldn't happen that often. |
| 661 | */ | 661 | */ |
| 662 | ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, | 662 | ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv, |
| 663 | dst_rsv, num_bytes); | 663 | dst_rsv, num_bytes); |
| 664 | goto out; | 664 | goto out; |
| 665 | } | 665 | } |
| 666 | 666 | ||
| 667 | migrate: | 667 | migrate: |
| 668 | ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); | 668 | ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); |
| 669 | 669 | ||
| 670 | out: | 670 | out: |
| 671 | /* | 671 | /* |
| 672 | * Migrate only takes a reservation, it doesn't touch the size of the | 672 | * Migrate only takes a reservation, it doesn't touch the size of the |
| 673 | * block_rsv. This is to simplify people who don't normally have things | 673 | * block_rsv. This is to simplify people who don't normally have things |
| 674 | * migrated from their block rsv. If they go to release their | 674 | * migrated from their block rsv. If they go to release their |
| 675 | * reservation, that will decrease the size as well, so if migrate | 675 | * reservation, that will decrease the size as well, so if migrate |
| 676 | * reduced size we'd end up with a negative size. But for the | 676 | * reduced size we'd end up with a negative size. But for the |
| 677 | * delalloc_meta_reserved stuff we will only know to drop 1 reservation, | 677 | * delalloc_meta_reserved stuff we will only know to drop 1 reservation, |
| 678 | * but we could in fact do this reserve/migrate dance several times | 678 | * but we could in fact do this reserve/migrate dance several times |
| 679 | * between the time we did the original reservation and we'd clean it | 679 | * between the time we did the original reservation and we'd clean it |
| 680 | * up. So to take care of this, release the space for the meta | 680 | * up. So to take care of this, release the space for the meta |
| 681 | * reservation here. I think it may be time for a documentation page on | 681 | * reservation here. I think it may be time for a documentation page on |
| 682 | * how block rsvs. work. | 682 | * how block rsvs. work. |
| 683 | */ | 683 | */ |
| 684 | if (!ret) { | 684 | if (!ret) { |
| 685 | trace_btrfs_space_reservation(root->fs_info, "delayed_inode", | 685 | trace_btrfs_space_reservation(root->fs_info, "delayed_inode", |
| 686 | btrfs_ino(inode), num_bytes, 1); | 686 | btrfs_ino(inode), num_bytes, 1); |
| 687 | node->bytes_reserved = num_bytes; | 687 | node->bytes_reserved = num_bytes; |
| 688 | } | 688 | } |
| 689 | 689 | ||
| 690 | if (release) { | 690 | if (release) { |
| 691 | trace_btrfs_space_reservation(root->fs_info, "delalloc", | 691 | trace_btrfs_space_reservation(root->fs_info, "delalloc", |
| 692 | btrfs_ino(inode), num_bytes, 0); | 692 | btrfs_ino(inode), num_bytes, 0); |
| 693 | btrfs_block_rsv_release(root, src_rsv, num_bytes); | 693 | btrfs_block_rsv_release(root, src_rsv, num_bytes); |
| 694 | } | 694 | } |
| 695 | 695 | ||
| 696 | return ret; | 696 | return ret; |
| 697 | } | 697 | } |
| 698 | 698 | ||
| 699 | static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root, | 699 | static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root, |
| 700 | struct btrfs_delayed_node *node) | 700 | struct btrfs_delayed_node *node) |
| 701 | { | 701 | { |
| 702 | struct btrfs_block_rsv *rsv; | 702 | struct btrfs_block_rsv *rsv; |
| 703 | 703 | ||
| 704 | if (!node->bytes_reserved) | 704 | if (!node->bytes_reserved) |
| 705 | return; | 705 | return; |
| 706 | 706 | ||
| 707 | rsv = &root->fs_info->delayed_block_rsv; | 707 | rsv = &root->fs_info->delayed_block_rsv; |
| 708 | trace_btrfs_space_reservation(root->fs_info, "delayed_inode", | 708 | trace_btrfs_space_reservation(root->fs_info, "delayed_inode", |
| 709 | node->inode_id, node->bytes_reserved, 0); | 709 | node->inode_id, node->bytes_reserved, 0); |
| 710 | btrfs_block_rsv_release(root, rsv, | 710 | btrfs_block_rsv_release(root, rsv, |
| 711 | node->bytes_reserved); | 711 | node->bytes_reserved); |
| 712 | node->bytes_reserved = 0; | 712 | node->bytes_reserved = 0; |
| 713 | } | 713 | } |
| 714 | 714 | ||
| 715 | /* | 715 | /* |
| 716 | * This helper will insert some continuous items into the same leaf according | 716 | * This helper will insert some continuous items into the same leaf according |
| 717 | * to the free space of the leaf. | 717 | * to the free space of the leaf. |
| 718 | */ | 718 | */ |
| 719 | static int btrfs_batch_insert_items(struct btrfs_root *root, | 719 | static int btrfs_batch_insert_items(struct btrfs_root *root, |
| 720 | struct btrfs_path *path, | 720 | struct btrfs_path *path, |
| 721 | struct btrfs_delayed_item *item) | 721 | struct btrfs_delayed_item *item) |
| 722 | { | 722 | { |
| 723 | struct btrfs_delayed_item *curr, *next; | 723 | struct btrfs_delayed_item *curr, *next; |
| 724 | int free_space; | 724 | int free_space; |
| 725 | int total_data_size = 0, total_size = 0; | 725 | int total_data_size = 0, total_size = 0; |
| 726 | struct extent_buffer *leaf; | 726 | struct extent_buffer *leaf; |
| 727 | char *data_ptr; | 727 | char *data_ptr; |
| 728 | struct btrfs_key *keys; | 728 | struct btrfs_key *keys; |
| 729 | u32 *data_size; | 729 | u32 *data_size; |
| 730 | struct list_head head; | 730 | struct list_head head; |
| 731 | int slot; | 731 | int slot; |
| 732 | int nitems; | 732 | int nitems; |
| 733 | int i; | 733 | int i; |
| 734 | int ret = 0; | 734 | int ret = 0; |
| 735 | 735 | ||
| 736 | BUG_ON(!path->nodes[0]); | 736 | BUG_ON(!path->nodes[0]); |
| 737 | 737 | ||
| 738 | leaf = path->nodes[0]; | 738 | leaf = path->nodes[0]; |
| 739 | free_space = btrfs_leaf_free_space(root, leaf); | 739 | free_space = btrfs_leaf_free_space(root, leaf); |
| 740 | INIT_LIST_HEAD(&head); | 740 | INIT_LIST_HEAD(&head); |
| 741 | 741 | ||
| 742 | next = item; | 742 | next = item; |
| 743 | nitems = 0; | 743 | nitems = 0; |
| 744 | 744 | ||
| 745 | /* | 745 | /* |
| 746 | * count the number of the continuous items that we can insert in batch | 746 | * count the number of the continuous items that we can insert in batch |
| 747 | */ | 747 | */ |
| 748 | while (total_size + next->data_len + sizeof(struct btrfs_item) <= | 748 | while (total_size + next->data_len + sizeof(struct btrfs_item) <= |
| 749 | free_space) { | 749 | free_space) { |
| 750 | total_data_size += next->data_len; | 750 | total_data_size += next->data_len; |
| 751 | total_size += next->data_len + sizeof(struct btrfs_item); | 751 | total_size += next->data_len + sizeof(struct btrfs_item); |
| 752 | list_add_tail(&next->tree_list, &head); | 752 | list_add_tail(&next->tree_list, &head); |
| 753 | nitems++; | 753 | nitems++; |
| 754 | 754 | ||
| 755 | curr = next; | 755 | curr = next; |
| 756 | next = __btrfs_next_delayed_item(curr); | 756 | next = __btrfs_next_delayed_item(curr); |
| 757 | if (!next) | 757 | if (!next) |
| 758 | break; | 758 | break; |
| 759 | 759 | ||
| 760 | if (!btrfs_is_continuous_delayed_item(curr, next)) | 760 | if (!btrfs_is_continuous_delayed_item(curr, next)) |
| 761 | break; | 761 | break; |
| 762 | } | 762 | } |
| 763 | 763 | ||
| 764 | if (!nitems) { | 764 | if (!nitems) { |
| 765 | ret = 0; | 765 | ret = 0; |
| 766 | goto out; | 766 | goto out; |
| 767 | } | 767 | } |
| 768 | 768 | ||
| 769 | /* | 769 | /* |
| 770 | * we need allocate some memory space, but it might cause the task | 770 | * we need allocate some memory space, but it might cause the task |
| 771 | * to sleep, so we set all locked nodes in the path to blocking locks | 771 | * to sleep, so we set all locked nodes in the path to blocking locks |
| 772 | * first. | 772 | * first. |
| 773 | */ | 773 | */ |
| 774 | btrfs_set_path_blocking(path); | 774 | btrfs_set_path_blocking(path); |
| 775 | 775 | ||
| 776 | keys = kmalloc_array(nitems, sizeof(struct btrfs_key), GFP_NOFS); | 776 | keys = kmalloc_array(nitems, sizeof(struct btrfs_key), GFP_NOFS); |
| 777 | if (!keys) { | 777 | if (!keys) { |
| 778 | ret = -ENOMEM; | 778 | ret = -ENOMEM; |
| 779 | goto out; | 779 | goto out; |
| 780 | } | 780 | } |
| 781 | 781 | ||
| 782 | data_size = kmalloc_array(nitems, sizeof(u32), GFP_NOFS); | 782 | data_size = kmalloc_array(nitems, sizeof(u32), GFP_NOFS); |
| 783 | if (!data_size) { | 783 | if (!data_size) { |
| 784 | ret = -ENOMEM; | 784 | ret = -ENOMEM; |
| 785 | goto error; | 785 | goto error; |
| 786 | } | 786 | } |
| 787 | 787 | ||
| 788 | /* get keys of all the delayed items */ | 788 | /* get keys of all the delayed items */ |
| 789 | i = 0; | 789 | i = 0; |
| 790 | list_for_each_entry(next, &head, tree_list) { | 790 | list_for_each_entry(next, &head, tree_list) { |
| 791 | keys[i] = next->key; | 791 | keys[i] = next->key; |
| 792 | data_size[i] = next->data_len; | 792 | data_size[i] = next->data_len; |
| 793 | i++; | 793 | i++; |
| 794 | } | 794 | } |
| 795 | 795 | ||
| 796 | /* reset all the locked nodes in the patch to spinning locks. */ | 796 | /* reset all the locked nodes in the patch to spinning locks. */ |
| 797 | btrfs_clear_path_blocking(path, NULL, 0); | 797 | btrfs_clear_path_blocking(path, NULL, 0); |
| 798 | 798 | ||
| 799 | /* insert the keys of the items */ | 799 | /* insert the keys of the items */ |
| 800 | setup_items_for_insert(root, path, keys, data_size, | 800 | setup_items_for_insert(root, path, keys, data_size, |
| 801 | total_data_size, total_size, nitems); | 801 | total_data_size, total_size, nitems); |
| 802 | 802 | ||
| 803 | /* insert the dir index items */ | 803 | /* insert the dir index items */ |
| 804 | slot = path->slots[0]; | 804 | slot = path->slots[0]; |
| 805 | list_for_each_entry_safe(curr, next, &head, tree_list) { | 805 | list_for_each_entry_safe(curr, next, &head, tree_list) { |
| 806 | data_ptr = btrfs_item_ptr(leaf, slot, char); | 806 | data_ptr = btrfs_item_ptr(leaf, slot, char); |
| 807 | write_extent_buffer(leaf, &curr->data, | 807 | write_extent_buffer(leaf, &curr->data, |
| 808 | (unsigned long)data_ptr, | 808 | (unsigned long)data_ptr, |
| 809 | curr->data_len); | 809 | curr->data_len); |
| 810 | slot++; | 810 | slot++; |
| 811 | 811 | ||
| 812 | btrfs_delayed_item_release_metadata(root, curr); | 812 | btrfs_delayed_item_release_metadata(root, curr); |
| 813 | 813 | ||
| 814 | list_del(&curr->tree_list); | 814 | list_del(&curr->tree_list); |
| 815 | btrfs_release_delayed_item(curr); | 815 | btrfs_release_delayed_item(curr); |
| 816 | } | 816 | } |
| 817 | 817 | ||
| 818 | error: | 818 | error: |
| 819 | kfree(data_size); | 819 | kfree(data_size); |
| 820 | kfree(keys); | 820 | kfree(keys); |
| 821 | out: | 821 | out: |
| 822 | return ret; | 822 | return ret; |
| 823 | } | 823 | } |
| 824 | 824 | ||
| 825 | /* | 825 | /* |
| 826 | * This helper can just do simple insertion that needn't extend item for new | 826 | * This helper can just do simple insertion that needn't extend item for new |
| 827 | * data, such as directory name index insertion, inode insertion. | 827 | * data, such as directory name index insertion, inode insertion. |
| 828 | */ | 828 | */ |
| 829 | static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans, | 829 | static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans, |
| 830 | struct btrfs_root *root, | 830 | struct btrfs_root *root, |
| 831 | struct btrfs_path *path, | 831 | struct btrfs_path *path, |
| 832 | struct btrfs_delayed_item *delayed_item) | 832 | struct btrfs_delayed_item *delayed_item) |
| 833 | { | 833 | { |
| 834 | struct extent_buffer *leaf; | 834 | struct extent_buffer *leaf; |
| 835 | char *ptr; | 835 | char *ptr; |
| 836 | int ret; | 836 | int ret; |
| 837 | 837 | ||
| 838 | ret = btrfs_insert_empty_item(trans, root, path, &delayed_item->key, | 838 | ret = btrfs_insert_empty_item(trans, root, path, &delayed_item->key, |
| 839 | delayed_item->data_len); | 839 | delayed_item->data_len); |
| 840 | if (ret < 0 && ret != -EEXIST) | 840 | if (ret < 0 && ret != -EEXIST) |
| 841 | return ret; | 841 | return ret; |
| 842 | 842 | ||
| 843 | leaf = path->nodes[0]; | 843 | leaf = path->nodes[0]; |
| 844 | 844 | ||
| 845 | ptr = btrfs_item_ptr(leaf, path->slots[0], char); | 845 | ptr = btrfs_item_ptr(leaf, path->slots[0], char); |
| 846 | 846 | ||
| 847 | write_extent_buffer(leaf, delayed_item->data, (unsigned long)ptr, | 847 | write_extent_buffer(leaf, delayed_item->data, (unsigned long)ptr, |
| 848 | delayed_item->data_len); | 848 | delayed_item->data_len); |
| 849 | btrfs_mark_buffer_dirty(leaf); | 849 | btrfs_mark_buffer_dirty(leaf); |
| 850 | 850 | ||
| 851 | btrfs_delayed_item_release_metadata(root, delayed_item); | 851 | btrfs_delayed_item_release_metadata(root, delayed_item); |
| 852 | return 0; | 852 | return 0; |
| 853 | } | 853 | } |
| 854 | 854 | ||
| 855 | /* | 855 | /* |
| 856 | * we insert an item first, then if there are some continuous items, we try | 856 | * we insert an item first, then if there are some continuous items, we try |
| 857 | * to insert those items into the same leaf. | 857 | * to insert those items into the same leaf. |
| 858 | */ | 858 | */ |
| 859 | static int btrfs_insert_delayed_items(struct btrfs_trans_handle *trans, | 859 | static int btrfs_insert_delayed_items(struct btrfs_trans_handle *trans, |
| 860 | struct btrfs_path *path, | 860 | struct btrfs_path *path, |
| 861 | struct btrfs_root *root, | 861 | struct btrfs_root *root, |
| 862 | struct btrfs_delayed_node *node) | 862 | struct btrfs_delayed_node *node) |
| 863 | { | 863 | { |
| 864 | struct btrfs_delayed_item *curr, *prev; | 864 | struct btrfs_delayed_item *curr, *prev; |
| 865 | int ret = 0; | 865 | int ret = 0; |
| 866 | 866 | ||
| 867 | do_again: | 867 | do_again: |
| 868 | mutex_lock(&node->mutex); | 868 | mutex_lock(&node->mutex); |
| 869 | curr = __btrfs_first_delayed_insertion_item(node); | 869 | curr = __btrfs_first_delayed_insertion_item(node); |
| 870 | if (!curr) | 870 | if (!curr) |
| 871 | goto insert_end; | 871 | goto insert_end; |
| 872 | 872 | ||
| 873 | ret = btrfs_insert_delayed_item(trans, root, path, curr); | 873 | ret = btrfs_insert_delayed_item(trans, root, path, curr); |
| 874 | if (ret < 0) { | 874 | if (ret < 0) { |
| 875 | btrfs_release_path(path); | 875 | btrfs_release_path(path); |
| 876 | goto insert_end; | 876 | goto insert_end; |
| 877 | } | 877 | } |
| 878 | 878 | ||
| 879 | prev = curr; | 879 | prev = curr; |
| 880 | curr = __btrfs_next_delayed_item(prev); | 880 | curr = __btrfs_next_delayed_item(prev); |
| 881 | if (curr && btrfs_is_continuous_delayed_item(prev, curr)) { | 881 | if (curr && btrfs_is_continuous_delayed_item(prev, curr)) { |
| 882 | /* insert the continuous items into the same leaf */ | 882 | /* insert the continuous items into the same leaf */ |
| 883 | path->slots[0]++; | 883 | path->slots[0]++; |
| 884 | btrfs_batch_insert_items(root, path, curr); | 884 | btrfs_batch_insert_items(root, path, curr); |
| 885 | } | 885 | } |
| 886 | btrfs_release_delayed_item(prev); | 886 | btrfs_release_delayed_item(prev); |
| 887 | btrfs_mark_buffer_dirty(path->nodes[0]); | 887 | btrfs_mark_buffer_dirty(path->nodes[0]); |
| 888 | 888 | ||
| 889 | btrfs_release_path(path); | 889 | btrfs_release_path(path); |
| 890 | mutex_unlock(&node->mutex); | 890 | mutex_unlock(&node->mutex); |
| 891 | goto do_again; | 891 | goto do_again; |
| 892 | 892 | ||
| 893 | insert_end: | 893 | insert_end: |
| 894 | mutex_unlock(&node->mutex); | 894 | mutex_unlock(&node->mutex); |
| 895 | return ret; | 895 | return ret; |
| 896 | } | 896 | } |
| 897 | 897 | ||
| 898 | static int btrfs_batch_delete_items(struct btrfs_trans_handle *trans, | 898 | static int btrfs_batch_delete_items(struct btrfs_trans_handle *trans, |
| 899 | struct btrfs_root *root, | 899 | struct btrfs_root *root, |
| 900 | struct btrfs_path *path, | 900 | struct btrfs_path *path, |
| 901 | struct btrfs_delayed_item *item) | 901 | struct btrfs_delayed_item *item) |
| 902 | { | 902 | { |
| 903 | struct btrfs_delayed_item *curr, *next; | 903 | struct btrfs_delayed_item *curr, *next; |
| 904 | struct extent_buffer *leaf; | 904 | struct extent_buffer *leaf; |
| 905 | struct btrfs_key key; | 905 | struct btrfs_key key; |
| 906 | struct list_head head; | 906 | struct list_head head; |
| 907 | int nitems, i, last_item; | 907 | int nitems, i, last_item; |
| 908 | int ret = 0; | 908 | int ret = 0; |
| 909 | 909 | ||
| 910 | BUG_ON(!path->nodes[0]); | 910 | BUG_ON(!path->nodes[0]); |
| 911 | 911 | ||
| 912 | leaf = path->nodes[0]; | 912 | leaf = path->nodes[0]; |
| 913 | 913 | ||
| 914 | i = path->slots[0]; | 914 | i = path->slots[0]; |
| 915 | last_item = btrfs_header_nritems(leaf) - 1; | 915 | last_item = btrfs_header_nritems(leaf) - 1; |
| 916 | if (i > last_item) | 916 | if (i > last_item) |
| 917 | return -ENOENT; /* FIXME: Is errno suitable? */ | 917 | return -ENOENT; /* FIXME: Is errno suitable? */ |
| 918 | 918 | ||
| 919 | next = item; | 919 | next = item; |
| 920 | INIT_LIST_HEAD(&head); | 920 | INIT_LIST_HEAD(&head); |
| 921 | btrfs_item_key_to_cpu(leaf, &key, i); | 921 | btrfs_item_key_to_cpu(leaf, &key, i); |
| 922 | nitems = 0; | 922 | nitems = 0; |
| 923 | /* | 923 | /* |
| 924 | * count the number of the dir index items that we can delete in batch | 924 | * count the number of the dir index items that we can delete in batch |
| 925 | */ | 925 | */ |
| 926 | while (btrfs_comp_cpu_keys(&next->key, &key) == 0) { | 926 | while (btrfs_comp_cpu_keys(&next->key, &key) == 0) { |
| 927 | list_add_tail(&next->tree_list, &head); | 927 | list_add_tail(&next->tree_list, &head); |
| 928 | nitems++; | 928 | nitems++; |
| 929 | 929 | ||
| 930 | curr = next; | 930 | curr = next; |
| 931 | next = __btrfs_next_delayed_item(curr); | 931 | next = __btrfs_next_delayed_item(curr); |
| 932 | if (!next) | 932 | if (!next) |
| 933 | break; | 933 | break; |
| 934 | 934 | ||
| 935 | if (!btrfs_is_continuous_delayed_item(curr, next)) | 935 | if (!btrfs_is_continuous_delayed_item(curr, next)) |
| 936 | break; | 936 | break; |
| 937 | 937 | ||
| 938 | i++; | 938 | i++; |
| 939 | if (i > last_item) | 939 | if (i > last_item) |
| 940 | break; | 940 | break; |
| 941 | btrfs_item_key_to_cpu(leaf, &key, i); | 941 | btrfs_item_key_to_cpu(leaf, &key, i); |
| 942 | } | 942 | } |
| 943 | 943 | ||
| 944 | if (!nitems) | 944 | if (!nitems) |
| 945 | return 0; | 945 | return 0; |
| 946 | 946 | ||
| 947 | ret = btrfs_del_items(trans, root, path, path->slots[0], nitems); | 947 | ret = btrfs_del_items(trans, root, path, path->slots[0], nitems); |
| 948 | if (ret) | 948 | if (ret) |
| 949 | goto out; | 949 | goto out; |
| 950 | 950 | ||
| 951 | list_for_each_entry_safe(curr, next, &head, tree_list) { | 951 | list_for_each_entry_safe(curr, next, &head, tree_list) { |
| 952 | btrfs_delayed_item_release_metadata(root, curr); | 952 | btrfs_delayed_item_release_metadata(root, curr); |
| 953 | list_del(&curr->tree_list); | 953 | list_del(&curr->tree_list); |
| 954 | btrfs_release_delayed_item(curr); | 954 | btrfs_release_delayed_item(curr); |
| 955 | } | 955 | } |
| 956 | 956 | ||
| 957 | out: | 957 | out: |
| 958 | return ret; | 958 | return ret; |
| 959 | } | 959 | } |
| 960 | 960 | ||
| 961 | static int btrfs_delete_delayed_items(struct btrfs_trans_handle *trans, | 961 | static int btrfs_delete_delayed_items(struct btrfs_trans_handle *trans, |
| 962 | struct btrfs_path *path, | 962 | struct btrfs_path *path, |
| 963 | struct btrfs_root *root, | 963 | struct btrfs_root *root, |
| 964 | struct btrfs_delayed_node *node) | 964 | struct btrfs_delayed_node *node) |
| 965 | { | 965 | { |
| 966 | struct btrfs_delayed_item *curr, *prev; | 966 | struct btrfs_delayed_item *curr, *prev; |
| 967 | int ret = 0; | 967 | int ret = 0; |
| 968 | 968 | ||
| 969 | do_again: | 969 | do_again: |
| 970 | mutex_lock(&node->mutex); | 970 | mutex_lock(&node->mutex); |
| 971 | curr = __btrfs_first_delayed_deletion_item(node); | 971 | curr = __btrfs_first_delayed_deletion_item(node); |
| 972 | if (!curr) | 972 | if (!curr) |
| 973 | goto delete_fail; | 973 | goto delete_fail; |
| 974 | 974 | ||
| 975 | ret = btrfs_search_slot(trans, root, &curr->key, path, -1, 1); | 975 | ret = btrfs_search_slot(trans, root, &curr->key, path, -1, 1); |
| 976 | if (ret < 0) | 976 | if (ret < 0) |
| 977 | goto delete_fail; | 977 | goto delete_fail; |
| 978 | else if (ret > 0) { | 978 | else if (ret > 0) { |
| 979 | /* | 979 | /* |
| 980 | * can't find the item which the node points to, so this node | 980 | * can't find the item which the node points to, so this node |
| 981 | * is invalid, just drop it. | 981 | * is invalid, just drop it. |
| 982 | */ | 982 | */ |
| 983 | prev = curr; | 983 | prev = curr; |
| 984 | curr = __btrfs_next_delayed_item(prev); | 984 | curr = __btrfs_next_delayed_item(prev); |
| 985 | btrfs_release_delayed_item(prev); | 985 | btrfs_release_delayed_item(prev); |
| 986 | ret = 0; | 986 | ret = 0; |
| 987 | btrfs_release_path(path); | 987 | btrfs_release_path(path); |
| 988 | if (curr) { | 988 | if (curr) { |
| 989 | mutex_unlock(&node->mutex); | 989 | mutex_unlock(&node->mutex); |
| 990 | goto do_again; | 990 | goto do_again; |
| 991 | } else | 991 | } else |
| 992 | goto delete_fail; | 992 | goto delete_fail; |
| 993 | } | 993 | } |
| 994 | 994 | ||
| 995 | btrfs_batch_delete_items(trans, root, path, curr); | 995 | btrfs_batch_delete_items(trans, root, path, curr); |
| 996 | btrfs_release_path(path); | 996 | btrfs_release_path(path); |
| 997 | mutex_unlock(&node->mutex); | 997 | mutex_unlock(&node->mutex); |
| 998 | goto do_again; | 998 | goto do_again; |
| 999 | 999 | ||
| 1000 | delete_fail: | 1000 | delete_fail: |
| 1001 | btrfs_release_path(path); | 1001 | btrfs_release_path(path); |
| 1002 | mutex_unlock(&node->mutex); | 1002 | mutex_unlock(&node->mutex); |
| 1003 | return ret; | 1003 | return ret; |
| 1004 | } | 1004 | } |
| 1005 | 1005 | ||
| 1006 | static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) | 1006 | static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) |
| 1007 | { | 1007 | { |
| 1008 | struct btrfs_delayed_root *delayed_root; | 1008 | struct btrfs_delayed_root *delayed_root; |
| 1009 | 1009 | ||
| 1010 | if (delayed_node && | 1010 | if (delayed_node && |
| 1011 | test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { | 1011 | test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { |
| 1012 | BUG_ON(!delayed_node->root); | 1012 | BUG_ON(!delayed_node->root); |
| 1013 | clear_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags); | 1013 | clear_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags); |
| 1014 | delayed_node->count--; | 1014 | delayed_node->count--; |
| 1015 | 1015 | ||
| 1016 | delayed_root = delayed_node->root->fs_info->delayed_root; | 1016 | delayed_root = delayed_node->root->fs_info->delayed_root; |
| 1017 | finish_one_item(delayed_root); | 1017 | finish_one_item(delayed_root); |
| 1018 | } | 1018 | } |
| 1019 | } | 1019 | } |
| 1020 | 1020 | ||
| 1021 | static void btrfs_release_delayed_iref(struct btrfs_delayed_node *delayed_node) | 1021 | static void btrfs_release_delayed_iref(struct btrfs_delayed_node *delayed_node) |
| 1022 | { | 1022 | { |
| 1023 | struct btrfs_delayed_root *delayed_root; | 1023 | struct btrfs_delayed_root *delayed_root; |
| 1024 | 1024 | ||
| 1025 | ASSERT(delayed_node->root); | 1025 | ASSERT(delayed_node->root); |
| 1026 | clear_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags); | 1026 | clear_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags); |
| 1027 | delayed_node->count--; | 1027 | delayed_node->count--; |
| 1028 | 1028 | ||
| 1029 | delayed_root = delayed_node->root->fs_info->delayed_root; | 1029 | delayed_root = delayed_node->root->fs_info->delayed_root; |
| 1030 | finish_one_item(delayed_root); | 1030 | finish_one_item(delayed_root); |
| 1031 | } | 1031 | } |
| 1032 | 1032 | ||
| 1033 | static int __btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, | 1033 | static int __btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, |
| 1034 | struct btrfs_root *root, | 1034 | struct btrfs_root *root, |
| 1035 | struct btrfs_path *path, | 1035 | struct btrfs_path *path, |
| 1036 | struct btrfs_delayed_node *node) | 1036 | struct btrfs_delayed_node *node) |
| 1037 | { | 1037 | { |
| 1038 | struct btrfs_key key; | 1038 | struct btrfs_key key; |
| 1039 | struct btrfs_inode_item *inode_item; | 1039 | struct btrfs_inode_item *inode_item; |
| 1040 | struct extent_buffer *leaf; | 1040 | struct extent_buffer *leaf; |
| 1041 | int mod; | 1041 | int mod; |
| 1042 | int ret; | 1042 | int ret; |
| 1043 | 1043 | ||
| 1044 | key.objectid = node->inode_id; | 1044 | key.objectid = node->inode_id; |
| 1045 | key.type = BTRFS_INODE_ITEM_KEY; | 1045 | key.type = BTRFS_INODE_ITEM_KEY; |
| 1046 | key.offset = 0; | 1046 | key.offset = 0; |
| 1047 | 1047 | ||
| 1048 | if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags)) | 1048 | if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags)) |
| 1049 | mod = -1; | 1049 | mod = -1; |
| 1050 | else | 1050 | else |
| 1051 | mod = 1; | 1051 | mod = 1; |
| 1052 | 1052 | ||
| 1053 | ret = btrfs_lookup_inode(trans, root, path, &key, mod); | 1053 | ret = btrfs_lookup_inode(trans, root, path, &key, mod); |
| 1054 | if (ret > 0) { | 1054 | if (ret > 0) { |
| 1055 | btrfs_release_path(path); | 1055 | btrfs_release_path(path); |
| 1056 | return -ENOENT; | 1056 | return -ENOENT; |
| 1057 | } else if (ret < 0) { | 1057 | } else if (ret < 0) { |
| 1058 | return ret; | 1058 | return ret; |
| 1059 | } | 1059 | } |
| 1060 | 1060 | ||
| 1061 | leaf = path->nodes[0]; | 1061 | leaf = path->nodes[0]; |
| 1062 | inode_item = btrfs_item_ptr(leaf, path->slots[0], | 1062 | inode_item = btrfs_item_ptr(leaf, path->slots[0], |
| 1063 | struct btrfs_inode_item); | 1063 | struct btrfs_inode_item); |
| 1064 | write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item, | 1064 | write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item, |
| 1065 | sizeof(struct btrfs_inode_item)); | 1065 | sizeof(struct btrfs_inode_item)); |
| 1066 | btrfs_mark_buffer_dirty(leaf); | 1066 | btrfs_mark_buffer_dirty(leaf); |
| 1067 | 1067 | ||
| 1068 | if (!test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags)) | 1068 | if (!test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags)) |
| 1069 | goto no_iref; | 1069 | goto no_iref; |
| 1070 | 1070 | ||
| 1071 | path->slots[0]++; | 1071 | path->slots[0]++; |
| 1072 | if (path->slots[0] >= btrfs_header_nritems(leaf)) | 1072 | if (path->slots[0] >= btrfs_header_nritems(leaf)) |
| 1073 | goto search; | 1073 | goto search; |
| 1074 | again: | 1074 | again: |
| 1075 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 1075 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 1076 | if (key.objectid != node->inode_id) | 1076 | if (key.objectid != node->inode_id) |
| 1077 | goto out; | 1077 | goto out; |
| 1078 | 1078 | ||
| 1079 | if (key.type != BTRFS_INODE_REF_KEY && | 1079 | if (key.type != BTRFS_INODE_REF_KEY && |
| 1080 | key.type != BTRFS_INODE_EXTREF_KEY) | 1080 | key.type != BTRFS_INODE_EXTREF_KEY) |
| 1081 | goto out; | 1081 | goto out; |
| 1082 | 1082 | ||
| 1083 | /* | 1083 | /* |
| 1084 | * Delayed iref deletion is for the inode who has only one link, | 1084 | * Delayed iref deletion is for the inode who has only one link, |
| 1085 | * so there is only one iref. The case that several irefs are | 1085 | * so there is only one iref. The case that several irefs are |
| 1086 | * in the same item doesn't exist. | 1086 | * in the same item doesn't exist. |
| 1087 | */ | 1087 | */ |
| 1088 | btrfs_del_item(trans, root, path); | 1088 | btrfs_del_item(trans, root, path); |
| 1089 | out: | 1089 | out: |
| 1090 | btrfs_release_delayed_iref(node); | 1090 | btrfs_release_delayed_iref(node); |
| 1091 | no_iref: | 1091 | no_iref: |
| 1092 | btrfs_release_path(path); | 1092 | btrfs_release_path(path); |
| 1093 | err_out: | 1093 | err_out: |
| 1094 | btrfs_delayed_inode_release_metadata(root, node); | 1094 | btrfs_delayed_inode_release_metadata(root, node); |
| 1095 | btrfs_release_delayed_inode(node); | 1095 | btrfs_release_delayed_inode(node); |
| 1096 | 1096 | ||
| 1097 | return ret; | 1097 | return ret; |
| 1098 | 1098 | ||
| 1099 | search: | 1099 | search: |
| 1100 | btrfs_release_path(path); | 1100 | btrfs_release_path(path); |
| 1101 | 1101 | ||
| 1102 | key.type = BTRFS_INODE_EXTREF_KEY; | 1102 | key.type = BTRFS_INODE_EXTREF_KEY; |
| 1103 | key.offset = -1; | 1103 | key.offset = -1; |
| 1104 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 1104 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 1105 | if (ret < 0) | 1105 | if (ret < 0) |
| 1106 | goto err_out; | 1106 | goto err_out; |
| 1107 | ASSERT(ret); | 1107 | ASSERT(ret); |
| 1108 | 1108 | ||
| 1109 | ret = 0; | 1109 | ret = 0; |
| 1110 | leaf = path->nodes[0]; | 1110 | leaf = path->nodes[0]; |
| 1111 | path->slots[0]--; | 1111 | path->slots[0]--; |
| 1112 | goto again; | 1112 | goto again; |
| 1113 | } | 1113 | } |
| 1114 | 1114 | ||
| 1115 | static inline int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, | 1115 | static inline int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, |
| 1116 | struct btrfs_root *root, | 1116 | struct btrfs_root *root, |
| 1117 | struct btrfs_path *path, | 1117 | struct btrfs_path *path, |
| 1118 | struct btrfs_delayed_node *node) | 1118 | struct btrfs_delayed_node *node) |
| 1119 | { | 1119 | { |
| 1120 | int ret; | 1120 | int ret; |
| 1121 | 1121 | ||
| 1122 | mutex_lock(&node->mutex); | 1122 | mutex_lock(&node->mutex); |
| 1123 | if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &node->flags)) { | 1123 | if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &node->flags)) { |
| 1124 | mutex_unlock(&node->mutex); | 1124 | mutex_unlock(&node->mutex); |
| 1125 | return 0; | 1125 | return 0; |
| 1126 | } | 1126 | } |
| 1127 | 1127 | ||
| 1128 | ret = __btrfs_update_delayed_inode(trans, root, path, node); | 1128 | ret = __btrfs_update_delayed_inode(trans, root, path, node); |
| 1129 | mutex_unlock(&node->mutex); | 1129 | mutex_unlock(&node->mutex); |
| 1130 | return ret; | 1130 | return ret; |
| 1131 | } | 1131 | } |
| 1132 | 1132 | ||
| 1133 | static inline int | 1133 | static inline int |
| 1134 | __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, | 1134 | __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, |
| 1135 | struct btrfs_path *path, | 1135 | struct btrfs_path *path, |
| 1136 | struct btrfs_delayed_node *node) | 1136 | struct btrfs_delayed_node *node) |
| 1137 | { | 1137 | { |
| 1138 | int ret; | 1138 | int ret; |
| 1139 | 1139 | ||
| 1140 | ret = btrfs_insert_delayed_items(trans, path, node->root, node); | 1140 | ret = btrfs_insert_delayed_items(trans, path, node->root, node); |
| 1141 | if (ret) | 1141 | if (ret) |
| 1142 | return ret; | 1142 | return ret; |
| 1143 | 1143 | ||
| 1144 | ret = btrfs_delete_delayed_items(trans, path, node->root, node); | 1144 | ret = btrfs_delete_delayed_items(trans, path, node->root, node); |
| 1145 | if (ret) | 1145 | if (ret) |
| 1146 | return ret; | 1146 | return ret; |
| 1147 | 1147 | ||
| 1148 | ret = btrfs_update_delayed_inode(trans, node->root, path, node); | 1148 | ret = btrfs_update_delayed_inode(trans, node->root, path, node); |
| 1149 | return ret; | 1149 | return ret; |
| 1150 | } | 1150 | } |
| 1151 | 1151 | ||
| 1152 | /* | 1152 | /* |
| 1153 | * Called when committing the transaction. | 1153 | * Called when committing the transaction. |
| 1154 | * Returns 0 on success. | 1154 | * Returns 0 on success. |
| 1155 | * Returns < 0 on error and returns with an aborted transaction with any | 1155 | * Returns < 0 on error and returns with an aborted transaction with any |
| 1156 | * outstanding delayed items cleaned up. | 1156 | * outstanding delayed items cleaned up. |
| 1157 | */ | 1157 | */ |
| 1158 | static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans, | 1158 | static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans, |
| 1159 | struct btrfs_root *root, int nr) | 1159 | struct btrfs_root *root, int nr) |
| 1160 | { | 1160 | { |
| 1161 | struct btrfs_delayed_root *delayed_root; | 1161 | struct btrfs_delayed_root *delayed_root; |
| 1162 | struct btrfs_delayed_node *curr_node, *prev_node; | 1162 | struct btrfs_delayed_node *curr_node, *prev_node; |
| 1163 | struct btrfs_path *path; | 1163 | struct btrfs_path *path; |
| 1164 | struct btrfs_block_rsv *block_rsv; | 1164 | struct btrfs_block_rsv *block_rsv; |
| 1165 | int ret = 0; | 1165 | int ret = 0; |
| 1166 | bool count = (nr > 0); | 1166 | bool count = (nr > 0); |
| 1167 | 1167 | ||
| 1168 | if (trans->aborted) | 1168 | if (trans->aborted) |
| 1169 | return -EIO; | 1169 | return -EIO; |
| 1170 | 1170 | ||
| 1171 | path = btrfs_alloc_path(); | 1171 | path = btrfs_alloc_path(); |
| 1172 | if (!path) | 1172 | if (!path) |
| 1173 | return -ENOMEM; | 1173 | return -ENOMEM; |
| 1174 | path->leave_spinning = 1; | 1174 | path->leave_spinning = 1; |
| 1175 | 1175 | ||
| 1176 | block_rsv = trans->block_rsv; | 1176 | block_rsv = trans->block_rsv; |
| 1177 | trans->block_rsv = &root->fs_info->delayed_block_rsv; | 1177 | trans->block_rsv = &root->fs_info->delayed_block_rsv; |
| 1178 | 1178 | ||
| 1179 | delayed_root = btrfs_get_delayed_root(root); | 1179 | delayed_root = btrfs_get_delayed_root(root); |
| 1180 | 1180 | ||
| 1181 | curr_node = btrfs_first_delayed_node(delayed_root); | 1181 | curr_node = btrfs_first_delayed_node(delayed_root); |
| 1182 | while (curr_node && (!count || (count && nr--))) { | 1182 | while (curr_node && (!count || (count && nr--))) { |
| 1183 | ret = __btrfs_commit_inode_delayed_items(trans, path, | 1183 | ret = __btrfs_commit_inode_delayed_items(trans, path, |
| 1184 | curr_node); | 1184 | curr_node); |
| 1185 | if (ret) { | 1185 | if (ret) { |
| 1186 | btrfs_release_delayed_node(curr_node); | 1186 | btrfs_release_delayed_node(curr_node); |
| 1187 | curr_node = NULL; | 1187 | curr_node = NULL; |
| 1188 | btrfs_abort_transaction(trans, root, ret); | 1188 | btrfs_abort_transaction(trans, root, ret); |
| 1189 | break; | 1189 | break; |
| 1190 | } | 1190 | } |
| 1191 | 1191 | ||
| 1192 | prev_node = curr_node; | 1192 | prev_node = curr_node; |
| 1193 | curr_node = btrfs_next_delayed_node(curr_node); | 1193 | curr_node = btrfs_next_delayed_node(curr_node); |
| 1194 | btrfs_release_delayed_node(prev_node); | 1194 | btrfs_release_delayed_node(prev_node); |
| 1195 | } | 1195 | } |
| 1196 | 1196 | ||
| 1197 | if (curr_node) | 1197 | if (curr_node) |
| 1198 | btrfs_release_delayed_node(curr_node); | 1198 | btrfs_release_delayed_node(curr_node); |
| 1199 | btrfs_free_path(path); | 1199 | btrfs_free_path(path); |
| 1200 | trans->block_rsv = block_rsv; | 1200 | trans->block_rsv = block_rsv; |
| 1201 | 1201 | ||
| 1202 | return ret; | 1202 | return ret; |
| 1203 | } | 1203 | } |
| 1204 | 1204 | ||
| 1205 | int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, | 1205 | int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, |
| 1206 | struct btrfs_root *root) | 1206 | struct btrfs_root *root) |
| 1207 | { | 1207 | { |
| 1208 | return __btrfs_run_delayed_items(trans, root, -1); | 1208 | return __btrfs_run_delayed_items(trans, root, -1); |
| 1209 | } | 1209 | } |
| 1210 | 1210 | ||
| 1211 | int btrfs_run_delayed_items_nr(struct btrfs_trans_handle *trans, | 1211 | int btrfs_run_delayed_items_nr(struct btrfs_trans_handle *trans, |
| 1212 | struct btrfs_root *root, int nr) | 1212 | struct btrfs_root *root, int nr) |
| 1213 | { | 1213 | { |
| 1214 | return __btrfs_run_delayed_items(trans, root, nr); | 1214 | return __btrfs_run_delayed_items(trans, root, nr); |
| 1215 | } | 1215 | } |
| 1216 | 1216 | ||
| 1217 | int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, | 1217 | int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, |
| 1218 | struct inode *inode) | 1218 | struct inode *inode) |
| 1219 | { | 1219 | { |
| 1220 | struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); | 1220 | struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); |
| 1221 | struct btrfs_path *path; | 1221 | struct btrfs_path *path; |
| 1222 | struct btrfs_block_rsv *block_rsv; | 1222 | struct btrfs_block_rsv *block_rsv; |
| 1223 | int ret; | 1223 | int ret; |
| 1224 | 1224 | ||
| 1225 | if (!delayed_node) | 1225 | if (!delayed_node) |
| 1226 | return 0; | 1226 | return 0; |
| 1227 | 1227 | ||
| 1228 | mutex_lock(&delayed_node->mutex); | 1228 | mutex_lock(&delayed_node->mutex); |
| 1229 | if (!delayed_node->count) { | 1229 | if (!delayed_node->count) { |
| 1230 | mutex_unlock(&delayed_node->mutex); | 1230 | mutex_unlock(&delayed_node->mutex); |
| 1231 | btrfs_release_delayed_node(delayed_node); | 1231 | btrfs_release_delayed_node(delayed_node); |
| 1232 | return 0; | 1232 | return 0; |
| 1233 | } | 1233 | } |
| 1234 | mutex_unlock(&delayed_node->mutex); | 1234 | mutex_unlock(&delayed_node->mutex); |
| 1235 | 1235 | ||
| 1236 | path = btrfs_alloc_path(); | 1236 | path = btrfs_alloc_path(); |
| 1237 | if (!path) { | 1237 | if (!path) { |
| 1238 | btrfs_release_delayed_node(delayed_node); | 1238 | btrfs_release_delayed_node(delayed_node); |
| 1239 | return -ENOMEM; | 1239 | return -ENOMEM; |
| 1240 | } | 1240 | } |
| 1241 | path->leave_spinning = 1; | 1241 | path->leave_spinning = 1; |
| 1242 | 1242 | ||
| 1243 | block_rsv = trans->block_rsv; | 1243 | block_rsv = trans->block_rsv; |
| 1244 | trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv; | 1244 | trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv; |
| 1245 | 1245 | ||
| 1246 | ret = __btrfs_commit_inode_delayed_items(trans, path, delayed_node); | 1246 | ret = __btrfs_commit_inode_delayed_items(trans, path, delayed_node); |
| 1247 | 1247 | ||
| 1248 | btrfs_release_delayed_node(delayed_node); | 1248 | btrfs_release_delayed_node(delayed_node); |
| 1249 | btrfs_free_path(path); | 1249 | btrfs_free_path(path); |
| 1250 | trans->block_rsv = block_rsv; | 1250 | trans->block_rsv = block_rsv; |
| 1251 | 1251 | ||
| 1252 | return ret; | 1252 | return ret; |
| 1253 | } | 1253 | } |
| 1254 | 1254 | ||
| 1255 | int btrfs_commit_inode_delayed_inode(struct inode *inode) | 1255 | int btrfs_commit_inode_delayed_inode(struct inode *inode) |
| 1256 | { | 1256 | { |
| 1257 | struct btrfs_trans_handle *trans; | 1257 | struct btrfs_trans_handle *trans; |
| 1258 | struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); | 1258 | struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); |
| 1259 | struct btrfs_path *path; | 1259 | struct btrfs_path *path; |
| 1260 | struct btrfs_block_rsv *block_rsv; | 1260 | struct btrfs_block_rsv *block_rsv; |
| 1261 | int ret; | 1261 | int ret; |
| 1262 | 1262 | ||
| 1263 | if (!delayed_node) | 1263 | if (!delayed_node) |
| 1264 | return 0; | 1264 | return 0; |
| 1265 | 1265 | ||
| 1266 | mutex_lock(&delayed_node->mutex); | 1266 | mutex_lock(&delayed_node->mutex); |
| 1267 | if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { | 1267 | if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { |
| 1268 | mutex_unlock(&delayed_node->mutex); | 1268 | mutex_unlock(&delayed_node->mutex); |
| 1269 | btrfs_release_delayed_node(delayed_node); | 1269 | btrfs_release_delayed_node(delayed_node); |
| 1270 | return 0; | 1270 | return 0; |
| 1271 | } | 1271 | } |
| 1272 | mutex_unlock(&delayed_node->mutex); | 1272 | mutex_unlock(&delayed_node->mutex); |
| 1273 | 1273 | ||
| 1274 | trans = btrfs_join_transaction(delayed_node->root); | 1274 | trans = btrfs_join_transaction(delayed_node->root); |
| 1275 | if (IS_ERR(trans)) { | 1275 | if (IS_ERR(trans)) { |
| 1276 | ret = PTR_ERR(trans); | 1276 | ret = PTR_ERR(trans); |
| 1277 | goto out; | 1277 | goto out; |
| 1278 | } | 1278 | } |
| 1279 | 1279 | ||
| 1280 | path = btrfs_alloc_path(); | 1280 | path = btrfs_alloc_path(); |
| 1281 | if (!path) { | 1281 | if (!path) { |
| 1282 | ret = -ENOMEM; | 1282 | ret = -ENOMEM; |
| 1283 | goto trans_out; | 1283 | goto trans_out; |
| 1284 | } | 1284 | } |
| 1285 | path->leave_spinning = 1; | 1285 | path->leave_spinning = 1; |
| 1286 | 1286 | ||
| 1287 | block_rsv = trans->block_rsv; | 1287 | block_rsv = trans->block_rsv; |
| 1288 | trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv; | 1288 | trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv; |
| 1289 | 1289 | ||
| 1290 | mutex_lock(&delayed_node->mutex); | 1290 | mutex_lock(&delayed_node->mutex); |
| 1291 | if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) | 1291 | if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) |
| 1292 | ret = __btrfs_update_delayed_inode(trans, delayed_node->root, | 1292 | ret = __btrfs_update_delayed_inode(trans, delayed_node->root, |
| 1293 | path, delayed_node); | 1293 | path, delayed_node); |
| 1294 | else | 1294 | else |
| 1295 | ret = 0; | 1295 | ret = 0; |
| 1296 | mutex_unlock(&delayed_node->mutex); | 1296 | mutex_unlock(&delayed_node->mutex); |
| 1297 | 1297 | ||
| 1298 | btrfs_free_path(path); | 1298 | btrfs_free_path(path); |
| 1299 | trans->block_rsv = block_rsv; | 1299 | trans->block_rsv = block_rsv; |
| 1300 | trans_out: | 1300 | trans_out: |
| 1301 | btrfs_end_transaction(trans, delayed_node->root); | 1301 | btrfs_end_transaction(trans, delayed_node->root); |
| 1302 | btrfs_btree_balance_dirty(delayed_node->root); | 1302 | btrfs_btree_balance_dirty(delayed_node->root); |
| 1303 | out: | 1303 | out: |
| 1304 | btrfs_release_delayed_node(delayed_node); | 1304 | btrfs_release_delayed_node(delayed_node); |
| 1305 | 1305 | ||
| 1306 | return ret; | 1306 | return ret; |
| 1307 | } | 1307 | } |
| 1308 | 1308 | ||
| 1309 | void btrfs_remove_delayed_node(struct inode *inode) | 1309 | void btrfs_remove_delayed_node(struct inode *inode) |
| 1310 | { | 1310 | { |
| 1311 | struct btrfs_delayed_node *delayed_node; | 1311 | struct btrfs_delayed_node *delayed_node; |
| 1312 | 1312 | ||
| 1313 | delayed_node = ACCESS_ONCE(BTRFS_I(inode)->delayed_node); | 1313 | delayed_node = ACCESS_ONCE(BTRFS_I(inode)->delayed_node); |
| 1314 | if (!delayed_node) | 1314 | if (!delayed_node) |
| 1315 | return; | 1315 | return; |
| 1316 | 1316 | ||
| 1317 | BTRFS_I(inode)->delayed_node = NULL; | 1317 | BTRFS_I(inode)->delayed_node = NULL; |
| 1318 | btrfs_release_delayed_node(delayed_node); | 1318 | btrfs_release_delayed_node(delayed_node); |
| 1319 | } | 1319 | } |
| 1320 | 1320 | ||
| 1321 | struct btrfs_async_delayed_work { | 1321 | struct btrfs_async_delayed_work { |
| 1322 | struct btrfs_delayed_root *delayed_root; | 1322 | struct btrfs_delayed_root *delayed_root; |
| 1323 | int nr; | 1323 | int nr; |
| 1324 | struct btrfs_work work; | 1324 | struct btrfs_work work; |
| 1325 | }; | 1325 | }; |
| 1326 | 1326 | ||
| 1327 | static void btrfs_async_run_delayed_root(struct btrfs_work *work) | 1327 | static void btrfs_async_run_delayed_root(struct btrfs_work *work) |
| 1328 | { | 1328 | { |
| 1329 | struct btrfs_async_delayed_work *async_work; | 1329 | struct btrfs_async_delayed_work *async_work; |
| 1330 | struct btrfs_delayed_root *delayed_root; | 1330 | struct btrfs_delayed_root *delayed_root; |
| 1331 | struct btrfs_trans_handle *trans; | 1331 | struct btrfs_trans_handle *trans; |
| 1332 | struct btrfs_path *path; | 1332 | struct btrfs_path *path; |
| 1333 | struct btrfs_delayed_node *delayed_node = NULL; | 1333 | struct btrfs_delayed_node *delayed_node = NULL; |
| 1334 | struct btrfs_root *root; | 1334 | struct btrfs_root *root; |
| 1335 | struct btrfs_block_rsv *block_rsv; | 1335 | struct btrfs_block_rsv *block_rsv; |
| 1336 | int total_done = 0; | 1336 | int total_done = 0; |
| 1337 | 1337 | ||
| 1338 | async_work = container_of(work, struct btrfs_async_delayed_work, work); | 1338 | async_work = container_of(work, struct btrfs_async_delayed_work, work); |
| 1339 | delayed_root = async_work->delayed_root; | 1339 | delayed_root = async_work->delayed_root; |
| 1340 | 1340 | ||
| 1341 | path = btrfs_alloc_path(); | 1341 | path = btrfs_alloc_path(); |
| 1342 | if (!path) | 1342 | if (!path) |
| 1343 | goto out; | 1343 | goto out; |
| 1344 | 1344 | ||
| 1345 | again: | 1345 | again: |
| 1346 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND / 2) | 1346 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND / 2) |
| 1347 | goto free_path; | 1347 | goto free_path; |
| 1348 | 1348 | ||
| 1349 | delayed_node = btrfs_first_prepared_delayed_node(delayed_root); | 1349 | delayed_node = btrfs_first_prepared_delayed_node(delayed_root); |
| 1350 | if (!delayed_node) | 1350 | if (!delayed_node) |
| 1351 | goto free_path; | 1351 | goto free_path; |
| 1352 | 1352 | ||
| 1353 | path->leave_spinning = 1; | 1353 | path->leave_spinning = 1; |
| 1354 | root = delayed_node->root; | 1354 | root = delayed_node->root; |
| 1355 | 1355 | ||
| 1356 | trans = btrfs_join_transaction(root); | 1356 | trans = btrfs_join_transaction(root); |
| 1357 | if (IS_ERR(trans)) | 1357 | if (IS_ERR(trans)) |
| 1358 | goto release_path; | 1358 | goto release_path; |
| 1359 | 1359 | ||
| 1360 | block_rsv = trans->block_rsv; | 1360 | block_rsv = trans->block_rsv; |
| 1361 | trans->block_rsv = &root->fs_info->delayed_block_rsv; | 1361 | trans->block_rsv = &root->fs_info->delayed_block_rsv; |
| 1362 | 1362 | ||
| 1363 | __btrfs_commit_inode_delayed_items(trans, path, delayed_node); | 1363 | __btrfs_commit_inode_delayed_items(trans, path, delayed_node); |
| 1364 | 1364 | ||
| 1365 | trans->block_rsv = block_rsv; | 1365 | trans->block_rsv = block_rsv; |
| 1366 | btrfs_end_transaction(trans, root); | 1366 | btrfs_end_transaction(trans, root); |
| 1367 | btrfs_btree_balance_dirty_nodelay(root); | 1367 | btrfs_btree_balance_dirty_nodelay(root); |
| 1368 | 1368 | ||
| 1369 | release_path: | 1369 | release_path: |
| 1370 | btrfs_release_path(path); | 1370 | btrfs_release_path(path); |
| 1371 | total_done++; | 1371 | total_done++; |
| 1372 | 1372 | ||
| 1373 | btrfs_release_prepared_delayed_node(delayed_node); | 1373 | btrfs_release_prepared_delayed_node(delayed_node); |
| 1374 | if (async_work->nr == 0 || total_done < async_work->nr) | 1374 | if (async_work->nr == 0 || total_done < async_work->nr) |
| 1375 | goto again; | 1375 | goto again; |
| 1376 | 1376 | ||
| 1377 | free_path: | 1377 | free_path: |
| 1378 | btrfs_free_path(path); | 1378 | btrfs_free_path(path); |
| 1379 | out: | 1379 | out: |
| 1380 | wake_up(&delayed_root->wait); | 1380 | wake_up(&delayed_root->wait); |
| 1381 | kfree(async_work); | 1381 | kfree(async_work); |
| 1382 | } | 1382 | } |
| 1383 | 1383 | ||
| 1384 | 1384 | ||
| 1385 | static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root, | 1385 | static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root, |
| 1386 | struct btrfs_root *root, int nr) | 1386 | struct btrfs_root *root, int nr) |
| 1387 | { | 1387 | { |
| 1388 | struct btrfs_async_delayed_work *async_work; | 1388 | struct btrfs_async_delayed_work *async_work; |
| 1389 | 1389 | ||
| 1390 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) | 1390 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) |
| 1391 | return 0; | 1391 | return 0; |
| 1392 | 1392 | ||
| 1393 | async_work = kmalloc(sizeof(*async_work), GFP_NOFS); | 1393 | async_work = kmalloc(sizeof(*async_work), GFP_NOFS); |
| 1394 | if (!async_work) | 1394 | if (!async_work) |
| 1395 | return -ENOMEM; | 1395 | return -ENOMEM; |
| 1396 | 1396 | ||
| 1397 | async_work->delayed_root = delayed_root; | 1397 | async_work->delayed_root = delayed_root; |
| 1398 | btrfs_init_work(&async_work->work, btrfs_delayed_meta_helper, | 1398 | btrfs_init_work(&async_work->work, btrfs_delayed_meta_helper, |
| 1399 | btrfs_async_run_delayed_root, NULL, NULL); | 1399 | btrfs_async_run_delayed_root, NULL, NULL); |
| 1400 | async_work->nr = nr; | 1400 | async_work->nr = nr; |
| 1401 | 1401 | ||
| 1402 | btrfs_queue_work(root->fs_info->delayed_workers, &async_work->work); | 1402 | btrfs_queue_work(root->fs_info->delayed_workers, &async_work->work); |
| 1403 | return 0; | 1403 | return 0; |
| 1404 | } | 1404 | } |
| 1405 | 1405 | ||
| 1406 | void btrfs_assert_delayed_root_empty(struct btrfs_root *root) | 1406 | void btrfs_assert_delayed_root_empty(struct btrfs_root *root) |
| 1407 | { | 1407 | { |
| 1408 | struct btrfs_delayed_root *delayed_root; | 1408 | struct btrfs_delayed_root *delayed_root; |
| 1409 | delayed_root = btrfs_get_delayed_root(root); | 1409 | delayed_root = btrfs_get_delayed_root(root); |
| 1410 | WARN_ON(btrfs_first_delayed_node(delayed_root)); | 1410 | WARN_ON(btrfs_first_delayed_node(delayed_root)); |
| 1411 | } | 1411 | } |
| 1412 | 1412 | ||
| 1413 | static int could_end_wait(struct btrfs_delayed_root *delayed_root, int seq) | 1413 | static int could_end_wait(struct btrfs_delayed_root *delayed_root, int seq) |
| 1414 | { | 1414 | { |
| 1415 | int val = atomic_read(&delayed_root->items_seq); | 1415 | int val = atomic_read(&delayed_root->items_seq); |
| 1416 | 1416 | ||
| 1417 | if (val < seq || val >= seq + BTRFS_DELAYED_BATCH) | 1417 | if (val < seq || val >= seq + BTRFS_DELAYED_BATCH) |
| 1418 | return 1; | 1418 | return 1; |
| 1419 | 1419 | ||
| 1420 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) | 1420 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) |
| 1421 | return 1; | 1421 | return 1; |
| 1422 | 1422 | ||
| 1423 | return 0; | 1423 | return 0; |
| 1424 | } | 1424 | } |
| 1425 | 1425 | ||
| 1426 | void btrfs_balance_delayed_items(struct btrfs_root *root) | 1426 | void btrfs_balance_delayed_items(struct btrfs_root *root) |
| 1427 | { | 1427 | { |
| 1428 | struct btrfs_delayed_root *delayed_root; | 1428 | struct btrfs_delayed_root *delayed_root; |
| 1429 | 1429 | ||
| 1430 | delayed_root = btrfs_get_delayed_root(root); | 1430 | delayed_root = btrfs_get_delayed_root(root); |
| 1431 | 1431 | ||
| 1432 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) | 1432 | if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) |
| 1433 | return; | 1433 | return; |
| 1434 | 1434 | ||
| 1435 | if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) { | 1435 | if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) { |
| 1436 | int seq; | 1436 | int seq; |
| 1437 | int ret; | 1437 | int ret; |
| 1438 | 1438 | ||
| 1439 | seq = atomic_read(&delayed_root->items_seq); | 1439 | seq = atomic_read(&delayed_root->items_seq); |
| 1440 | 1440 | ||
| 1441 | ret = btrfs_wq_run_delayed_node(delayed_root, root, 0); | 1441 | ret = btrfs_wq_run_delayed_node(delayed_root, root, 0); |
| 1442 | if (ret) | 1442 | if (ret) |
| 1443 | return; | 1443 | return; |
| 1444 | 1444 | ||
| 1445 | wait_event_interruptible(delayed_root->wait, | 1445 | wait_event_interruptible(delayed_root->wait, |
| 1446 | could_end_wait(delayed_root, seq)); | 1446 | could_end_wait(delayed_root, seq)); |
| 1447 | return; | 1447 | return; |
| 1448 | } | 1448 | } |
| 1449 | 1449 | ||
| 1450 | btrfs_wq_run_delayed_node(delayed_root, root, BTRFS_DELAYED_BATCH); | 1450 | btrfs_wq_run_delayed_node(delayed_root, root, BTRFS_DELAYED_BATCH); |
| 1451 | } | 1451 | } |
| 1452 | 1452 | ||
| 1453 | /* Will return 0 or -ENOMEM */ | 1453 | /* Will return 0 or -ENOMEM */ |
| 1454 | int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, | 1454 | int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, |
| 1455 | struct btrfs_root *root, const char *name, | 1455 | struct btrfs_root *root, const char *name, |
| 1456 | int name_len, struct inode *dir, | 1456 | int name_len, struct inode *dir, |
| 1457 | struct btrfs_disk_key *disk_key, u8 type, | 1457 | struct btrfs_disk_key *disk_key, u8 type, |
| 1458 | u64 index) | 1458 | u64 index) |
| 1459 | { | 1459 | { |
| 1460 | struct btrfs_delayed_node *delayed_node; | 1460 | struct btrfs_delayed_node *delayed_node; |
| 1461 | struct btrfs_delayed_item *delayed_item; | 1461 | struct btrfs_delayed_item *delayed_item; |
| 1462 | struct btrfs_dir_item *dir_item; | 1462 | struct btrfs_dir_item *dir_item; |
| 1463 | int ret; | 1463 | int ret; |
| 1464 | 1464 | ||
| 1465 | delayed_node = btrfs_get_or_create_delayed_node(dir); | 1465 | delayed_node = btrfs_get_or_create_delayed_node(dir); |
| 1466 | if (IS_ERR(delayed_node)) | 1466 | if (IS_ERR(delayed_node)) |
| 1467 | return PTR_ERR(delayed_node); | 1467 | return PTR_ERR(delayed_node); |
| 1468 | 1468 | ||
| 1469 | delayed_item = btrfs_alloc_delayed_item(sizeof(*dir_item) + name_len); | 1469 | delayed_item = btrfs_alloc_delayed_item(sizeof(*dir_item) + name_len); |
| 1470 | if (!delayed_item) { | 1470 | if (!delayed_item) { |
| 1471 | ret = -ENOMEM; | 1471 | ret = -ENOMEM; |
| 1472 | goto release_node; | 1472 | goto release_node; |
| 1473 | } | 1473 | } |
| 1474 | 1474 | ||
| 1475 | delayed_item->key.objectid = btrfs_ino(dir); | 1475 | delayed_item->key.objectid = btrfs_ino(dir); |
| 1476 | delayed_item->key.type = BTRFS_DIR_INDEX_KEY; | 1476 | delayed_item->key.type = BTRFS_DIR_INDEX_KEY; |
| 1477 | delayed_item->key.offset = index; | 1477 | delayed_item->key.offset = index; |
| 1478 | 1478 | ||
| 1479 | dir_item = (struct btrfs_dir_item *)delayed_item->data; | 1479 | dir_item = (struct btrfs_dir_item *)delayed_item->data; |
| 1480 | dir_item->location = *disk_key; | 1480 | dir_item->location = *disk_key; |
| 1481 | btrfs_set_stack_dir_transid(dir_item, trans->transid); | 1481 | btrfs_set_stack_dir_transid(dir_item, trans->transid); |
| 1482 | btrfs_set_stack_dir_data_len(dir_item, 0); | 1482 | btrfs_set_stack_dir_data_len(dir_item, 0); |
| 1483 | btrfs_set_stack_dir_name_len(dir_item, name_len); | 1483 | btrfs_set_stack_dir_name_len(dir_item, name_len); |
| 1484 | btrfs_set_stack_dir_type(dir_item, type); | 1484 | btrfs_set_stack_dir_type(dir_item, type); |
| 1485 | memcpy((char *)(dir_item + 1), name, name_len); | 1485 | memcpy((char *)(dir_item + 1), name, name_len); |
| 1486 | 1486 | ||
| 1487 | ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item); | 1487 | ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item); |
| 1488 | /* | 1488 | /* |
| 1489 | * we have reserved enough space when we start a new transaction, | 1489 | * we have reserved enough space when we start a new transaction, |
| 1490 | * so reserving metadata failure is impossible | 1490 | * so reserving metadata failure is impossible |
| 1491 | */ | 1491 | */ |
| 1492 | BUG_ON(ret); | 1492 | BUG_ON(ret); |
| 1493 | 1493 | ||
| 1494 | 1494 | ||
| 1495 | mutex_lock(&delayed_node->mutex); | 1495 | mutex_lock(&delayed_node->mutex); |
| 1496 | ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item); | 1496 | ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item); |
| 1497 | if (unlikely(ret)) { | 1497 | if (unlikely(ret)) { |
| 1498 | btrfs_err(root->fs_info, "err add delayed dir index item(name: %.*s) " | 1498 | btrfs_err(root->fs_info, "err add delayed dir index item(name: %.*s) " |
| 1499 | "into the insertion tree of the delayed node" | 1499 | "into the insertion tree of the delayed node" |
| 1500 | "(root id: %llu, inode id: %llu, errno: %d)", | 1500 | "(root id: %llu, inode id: %llu, errno: %d)", |
| 1501 | name_len, name, delayed_node->root->objectid, | 1501 | name_len, name, delayed_node->root->objectid, |
| 1502 | delayed_node->inode_id, ret); | 1502 | delayed_node->inode_id, ret); |
| 1503 | BUG(); | 1503 | BUG(); |
| 1504 | } | 1504 | } |
| 1505 | mutex_unlock(&delayed_node->mutex); | 1505 | mutex_unlock(&delayed_node->mutex); |
| 1506 | 1506 | ||
| 1507 | release_node: | 1507 | release_node: |
| 1508 | btrfs_release_delayed_node(delayed_node); | 1508 | btrfs_release_delayed_node(delayed_node); |
| 1509 | return ret; | 1509 | return ret; |
| 1510 | } | 1510 | } |
| 1511 | 1511 | ||
| 1512 | static int btrfs_delete_delayed_insertion_item(struct btrfs_root *root, | 1512 | static int btrfs_delete_delayed_insertion_item(struct btrfs_root *root, |
| 1513 | struct btrfs_delayed_node *node, | 1513 | struct btrfs_delayed_node *node, |
| 1514 | struct btrfs_key *key) | 1514 | struct btrfs_key *key) |
| 1515 | { | 1515 | { |
| 1516 | struct btrfs_delayed_item *item; | 1516 | struct btrfs_delayed_item *item; |
| 1517 | 1517 | ||
| 1518 | mutex_lock(&node->mutex); | 1518 | mutex_lock(&node->mutex); |
| 1519 | item = __btrfs_lookup_delayed_insertion_item(node, key); | 1519 | item = __btrfs_lookup_delayed_insertion_item(node, key); |
| 1520 | if (!item) { | 1520 | if (!item) { |
| 1521 | mutex_unlock(&node->mutex); | 1521 | mutex_unlock(&node->mutex); |
| 1522 | return 1; | 1522 | return 1; |
| 1523 | } | 1523 | } |
| 1524 | 1524 | ||
| 1525 | btrfs_delayed_item_release_metadata(root, item); | 1525 | btrfs_delayed_item_release_metadata(root, item); |
| 1526 | btrfs_release_delayed_item(item); | 1526 | btrfs_release_delayed_item(item); |
| 1527 | mutex_unlock(&node->mutex); | 1527 | mutex_unlock(&node->mutex); |
| 1528 | return 0; | 1528 | return 0; |
| 1529 | } | 1529 | } |
| 1530 | 1530 | ||
| 1531 | int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans, | 1531 | int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans, |
| 1532 | struct btrfs_root *root, struct inode *dir, | 1532 | struct btrfs_root *root, struct inode *dir, |
| 1533 | u64 index) | 1533 | u64 index) |
| 1534 | { | 1534 | { |
| 1535 | struct btrfs_delayed_node *node; | 1535 | struct btrfs_delayed_node *node; |
| 1536 | struct btrfs_delayed_item *item; | 1536 | struct btrfs_delayed_item *item; |
| 1537 | struct btrfs_key item_key; | 1537 | struct btrfs_key item_key; |
| 1538 | int ret; | 1538 | int ret; |
| 1539 | 1539 | ||
| 1540 | node = btrfs_get_or_create_delayed_node(dir); | 1540 | node = btrfs_get_or_create_delayed_node(dir); |
| 1541 | if (IS_ERR(node)) | 1541 | if (IS_ERR(node)) |
| 1542 | return PTR_ERR(node); | 1542 | return PTR_ERR(node); |
| 1543 | 1543 | ||
| 1544 | item_key.objectid = btrfs_ino(dir); | 1544 | item_key.objectid = btrfs_ino(dir); |
| 1545 | item_key.type = BTRFS_DIR_INDEX_KEY; | 1545 | item_key.type = BTRFS_DIR_INDEX_KEY; |
| 1546 | item_key.offset = index; | 1546 | item_key.offset = index; |
| 1547 | 1547 | ||
| 1548 | ret = btrfs_delete_delayed_insertion_item(root, node, &item_key); | 1548 | ret = btrfs_delete_delayed_insertion_item(root, node, &item_key); |
| 1549 | if (!ret) | 1549 | if (!ret) |
| 1550 | goto end; | 1550 | goto end; |
| 1551 | 1551 | ||
| 1552 | item = btrfs_alloc_delayed_item(0); | 1552 | item = btrfs_alloc_delayed_item(0); |
| 1553 | if (!item) { | 1553 | if (!item) { |
| 1554 | ret = -ENOMEM; | 1554 | ret = -ENOMEM; |
| 1555 | goto end; | 1555 | goto end; |
| 1556 | } | 1556 | } |
| 1557 | 1557 | ||
| 1558 | item->key = item_key; | 1558 | item->key = item_key; |
| 1559 | 1559 | ||
| 1560 | ret = btrfs_delayed_item_reserve_metadata(trans, root, item); | 1560 | ret = btrfs_delayed_item_reserve_metadata(trans, root, item); |
| 1561 | /* | 1561 | /* |
| 1562 | * we have reserved enough space when we start a new transaction, | 1562 | * we have reserved enough space when we start a new transaction, |
| 1563 | * so reserving metadata failure is impossible. | 1563 | * so reserving metadata failure is impossible. |
| 1564 | */ | 1564 | */ |
| 1565 | BUG_ON(ret); | 1565 | BUG_ON(ret); |
| 1566 | 1566 | ||
| 1567 | mutex_lock(&node->mutex); | 1567 | mutex_lock(&node->mutex); |
| 1568 | ret = __btrfs_add_delayed_deletion_item(node, item); | 1568 | ret = __btrfs_add_delayed_deletion_item(node, item); |
| 1569 | if (unlikely(ret)) { | 1569 | if (unlikely(ret)) { |
| 1570 | btrfs_err(root->fs_info, "err add delayed dir index item(index: %llu) " | 1570 | btrfs_err(root->fs_info, "err add delayed dir index item(index: %llu) " |
| 1571 | "into the deletion tree of the delayed node" | 1571 | "into the deletion tree of the delayed node" |
| 1572 | "(root id: %llu, inode id: %llu, errno: %d)", | 1572 | "(root id: %llu, inode id: %llu, errno: %d)", |
| 1573 | index, node->root->objectid, node->inode_id, | 1573 | index, node->root->objectid, node->inode_id, |
| 1574 | ret); | 1574 | ret); |
| 1575 | BUG(); | 1575 | BUG(); |
| 1576 | } | 1576 | } |
| 1577 | mutex_unlock(&node->mutex); | 1577 | mutex_unlock(&node->mutex); |
| 1578 | end: | 1578 | end: |
| 1579 | btrfs_release_delayed_node(node); | 1579 | btrfs_release_delayed_node(node); |
| 1580 | return ret; | 1580 | return ret; |
| 1581 | } | 1581 | } |
| 1582 | 1582 | ||
| 1583 | int btrfs_inode_delayed_dir_index_count(struct inode *inode) | 1583 | int btrfs_inode_delayed_dir_index_count(struct inode *inode) |
| 1584 | { | 1584 | { |
| 1585 | struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); | 1585 | struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); |
| 1586 | 1586 | ||
| 1587 | if (!delayed_node) | 1587 | if (!delayed_node) |
| 1588 | return -ENOENT; | 1588 | return -ENOENT; |
| 1589 | 1589 | ||
| 1590 | /* | 1590 | /* |
| 1591 | * Since we have held i_mutex of this directory, it is impossible that | 1591 | * Since we have held i_mutex of this directory, it is impossible that |
| 1592 | * a new directory index is added into the delayed node and index_cnt | 1592 | * a new directory index is added into the delayed node and index_cnt |
| 1593 | * is updated now. So we needn't lock the delayed node. | 1593 | * is updated now. So we needn't lock the delayed node. |
| 1594 | */ | 1594 | */ |
| 1595 | if (!delayed_node->index_cnt) { | 1595 | if (!delayed_node->index_cnt) { |
| 1596 | btrfs_release_delayed_node(delayed_node); | 1596 | btrfs_release_delayed_node(delayed_node); |
| 1597 | return -EINVAL; | 1597 | return -EINVAL; |
| 1598 | } | 1598 | } |
| 1599 | 1599 | ||
| 1600 | BTRFS_I(inode)->index_cnt = delayed_node->index_cnt; | 1600 | BTRFS_I(inode)->index_cnt = delayed_node->index_cnt; |
| 1601 | btrfs_release_delayed_node(delayed_node); | 1601 | btrfs_release_delayed_node(delayed_node); |
| 1602 | return 0; | 1602 | return 0; |
| 1603 | } | 1603 | } |
| 1604 | 1604 | ||
| 1605 | void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, | 1605 | void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, |
| 1606 | struct list_head *del_list) | 1606 | struct list_head *del_list) |
| 1607 | { | 1607 | { |
| 1608 | struct btrfs_delayed_node *delayed_node; | 1608 | struct btrfs_delayed_node *delayed_node; |
| 1609 | struct btrfs_delayed_item *item; | 1609 | struct btrfs_delayed_item *item; |
| 1610 | 1610 | ||
| 1611 | delayed_node = btrfs_get_delayed_node(inode); | 1611 | delayed_node = btrfs_get_delayed_node(inode); |
| 1612 | if (!delayed_node) | 1612 | if (!delayed_node) |
| 1613 | return; | 1613 | return; |
| 1614 | 1614 | ||
| 1615 | mutex_lock(&delayed_node->mutex); | 1615 | mutex_lock(&delayed_node->mutex); |
| 1616 | item = __btrfs_first_delayed_insertion_item(delayed_node); | 1616 | item = __btrfs_first_delayed_insertion_item(delayed_node); |
| 1617 | while (item) { | 1617 | while (item) { |
| 1618 | atomic_inc(&item->refs); | 1618 | atomic_inc(&item->refs); |
| 1619 | list_add_tail(&item->readdir_list, ins_list); | 1619 | list_add_tail(&item->readdir_list, ins_list); |
| 1620 | item = __btrfs_next_delayed_item(item); | 1620 | item = __btrfs_next_delayed_item(item); |
| 1621 | } | 1621 | } |
| 1622 | 1622 | ||
| 1623 | item = __btrfs_first_delayed_deletion_item(delayed_node); | 1623 | item = __btrfs_first_delayed_deletion_item(delayed_node); |
| 1624 | while (item) { | 1624 | while (item) { |
| 1625 | atomic_inc(&item->refs); | 1625 | atomic_inc(&item->refs); |
| 1626 | list_add_tail(&item->readdir_list, del_list); | 1626 | list_add_tail(&item->readdir_list, del_list); |
| 1627 | item = __btrfs_next_delayed_item(item); | 1627 | item = __btrfs_next_delayed_item(item); |
| 1628 | } | 1628 | } |
| 1629 | mutex_unlock(&delayed_node->mutex); | 1629 | mutex_unlock(&delayed_node->mutex); |
| 1630 | /* | 1630 | /* |
| 1631 | * This delayed node is still cached in the btrfs inode, so refs | 1631 | * This delayed node is still cached in the btrfs inode, so refs |
| 1632 | * must be > 1 now, and we needn't check it is going to be freed | 1632 | * must be > 1 now, and we needn't check it is going to be freed |
| 1633 | * or not. | 1633 | * or not. |
| 1634 | * | 1634 | * |
| 1635 | * Besides that, this function is used to read dir, we do not | 1635 | * Besides that, this function is used to read dir, we do not |
| 1636 | * insert/delete delayed items in this period. So we also needn't | 1636 | * insert/delete delayed items in this period. So we also needn't |
| 1637 | * requeue or dequeue this delayed node. | 1637 | * requeue or dequeue this delayed node. |
| 1638 | */ | 1638 | */ |
| 1639 | atomic_dec(&delayed_node->refs); | 1639 | atomic_dec(&delayed_node->refs); |
| 1640 | } | 1640 | } |
| 1641 | 1641 | ||
| 1642 | void btrfs_put_delayed_items(struct list_head *ins_list, | 1642 | void btrfs_put_delayed_items(struct list_head *ins_list, |
| 1643 | struct list_head *del_list) | 1643 | struct list_head *del_list) |
| 1644 | { | 1644 | { |
| 1645 | struct btrfs_delayed_item *curr, *next; | 1645 | struct btrfs_delayed_item *curr, *next; |
| 1646 | 1646 | ||
| 1647 | list_for_each_entry_safe(curr, next, ins_list, readdir_list) { | 1647 | list_for_each_entry_safe(curr, next, ins_list, readdir_list) { |
| 1648 | list_del(&curr->readdir_list); | 1648 | list_del(&curr->readdir_list); |
| 1649 | if (atomic_dec_and_test(&curr->refs)) | 1649 | if (atomic_dec_and_test(&curr->refs)) |
| 1650 | kfree(curr); | 1650 | kfree(curr); |
| 1651 | } | 1651 | } |
| 1652 | 1652 | ||
| 1653 | list_for_each_entry_safe(curr, next, del_list, readdir_list) { | 1653 | list_for_each_entry_safe(curr, next, del_list, readdir_list) { |
| 1654 | list_del(&curr->readdir_list); | 1654 | list_del(&curr->readdir_list); |
| 1655 | if (atomic_dec_and_test(&curr->refs)) | 1655 | if (atomic_dec_and_test(&curr->refs)) |
| 1656 | kfree(curr); | 1656 | kfree(curr); |
| 1657 | } | 1657 | } |
| 1658 | } | 1658 | } |
| 1659 | 1659 | ||
| 1660 | int btrfs_should_delete_dir_index(struct list_head *del_list, | 1660 | int btrfs_should_delete_dir_index(struct list_head *del_list, |
| 1661 | u64 index) | 1661 | u64 index) |
| 1662 | { | 1662 | { |
| 1663 | struct btrfs_delayed_item *curr, *next; | 1663 | struct btrfs_delayed_item *curr, *next; |
| 1664 | int ret; | 1664 | int ret; |
| 1665 | 1665 | ||
| 1666 | if (list_empty(del_list)) | 1666 | if (list_empty(del_list)) |
| 1667 | return 0; | 1667 | return 0; |
| 1668 | 1668 | ||
| 1669 | list_for_each_entry_safe(curr, next, del_list, readdir_list) { | 1669 | list_for_each_entry_safe(curr, next, del_list, readdir_list) { |
| 1670 | if (curr->key.offset > index) | 1670 | if (curr->key.offset > index) |
| 1671 | break; | 1671 | break; |
| 1672 | 1672 | ||
| 1673 | list_del(&curr->readdir_list); | 1673 | list_del(&curr->readdir_list); |
| 1674 | ret = (curr->key.offset == index); | 1674 | ret = (curr->key.offset == index); |
| 1675 | 1675 | ||
| 1676 | if (atomic_dec_and_test(&curr->refs)) | 1676 | if (atomic_dec_and_test(&curr->refs)) |
| 1677 | kfree(curr); | 1677 | kfree(curr); |
| 1678 | 1678 | ||
| 1679 | if (ret) | 1679 | if (ret) |
| 1680 | return 1; | 1680 | return 1; |
| 1681 | else | 1681 | else |
| 1682 | continue; | 1682 | continue; |
| 1683 | } | 1683 | } |
| 1684 | return 0; | 1684 | return 0; |
| 1685 | } | 1685 | } |
| 1686 | 1686 | ||
| 1687 | /* | 1687 | /* |
| 1688 | * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree | 1688 | * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree |
| 1689 | * | 1689 | * |
| 1690 | */ | 1690 | */ |
| 1691 | int btrfs_readdir_delayed_dir_index(struct dir_context *ctx, | 1691 | int btrfs_readdir_delayed_dir_index(struct dir_context *ctx, |
| 1692 | struct list_head *ins_list) | 1692 | struct list_head *ins_list) |
| 1693 | { | 1693 | { |
| 1694 | struct btrfs_dir_item *di; | 1694 | struct btrfs_dir_item *di; |
| 1695 | struct btrfs_delayed_item *curr, *next; | 1695 | struct btrfs_delayed_item *curr, *next; |
| 1696 | struct btrfs_key location; | 1696 | struct btrfs_key location; |
| 1697 | char *name; | 1697 | char *name; |
| 1698 | int name_len; | 1698 | int name_len; |
| 1699 | int over = 0; | 1699 | int over = 0; |
| 1700 | unsigned char d_type; | 1700 | unsigned char d_type; |
| 1701 | 1701 | ||
| 1702 | if (list_empty(ins_list)) | 1702 | if (list_empty(ins_list)) |
| 1703 | return 0; | 1703 | return 0; |
| 1704 | 1704 | ||
| 1705 | /* | 1705 | /* |
| 1706 | * Changing the data of the delayed item is impossible. So | 1706 | * Changing the data of the delayed item is impossible. So |
| 1707 | * we needn't lock them. And we have held i_mutex of the | 1707 | * we needn't lock them. And we have held i_mutex of the |
| 1708 | * directory, nobody can delete any directory indexes now. | 1708 | * directory, nobody can delete any directory indexes now. |
| 1709 | */ | 1709 | */ |
| 1710 | list_for_each_entry_safe(curr, next, ins_list, readdir_list) { | 1710 | list_for_each_entry_safe(curr, next, ins_list, readdir_list) { |
| 1711 | list_del(&curr->readdir_list); | 1711 | list_del(&curr->readdir_list); |
| 1712 | 1712 | ||
| 1713 | if (curr->key.offset < ctx->pos) { | 1713 | if (curr->key.offset < ctx->pos) { |
| 1714 | if (atomic_dec_and_test(&curr->refs)) | 1714 | if (atomic_dec_and_test(&curr->refs)) |
| 1715 | kfree(curr); | 1715 | kfree(curr); |
| 1716 | continue; | 1716 | continue; |
| 1717 | } | 1717 | } |
| 1718 | 1718 | ||
| 1719 | ctx->pos = curr->key.offset; | 1719 | ctx->pos = curr->key.offset; |
| 1720 | 1720 | ||
| 1721 | di = (struct btrfs_dir_item *)curr->data; | 1721 | di = (struct btrfs_dir_item *)curr->data; |
| 1722 | name = (char *)(di + 1); | 1722 | name = (char *)(di + 1); |
| 1723 | name_len = btrfs_stack_dir_name_len(di); | 1723 | name_len = btrfs_stack_dir_name_len(di); |
| 1724 | 1724 | ||
| 1725 | d_type = btrfs_filetype_table[di->type]; | 1725 | d_type = btrfs_filetype_table[di->type]; |
| 1726 | btrfs_disk_key_to_cpu(&location, &di->location); | 1726 | btrfs_disk_key_to_cpu(&location, &di->location); |
| 1727 | 1727 | ||
| 1728 | over = !dir_emit(ctx, name, name_len, | 1728 | over = !dir_emit(ctx, name, name_len, |
| 1729 | location.objectid, d_type); | 1729 | location.objectid, d_type); |
| 1730 | 1730 | ||
| 1731 | if (atomic_dec_and_test(&curr->refs)) | 1731 | if (atomic_dec_and_test(&curr->refs)) |
| 1732 | kfree(curr); | 1732 | kfree(curr); |
| 1733 | 1733 | ||
| 1734 | if (over) | 1734 | if (over) |
| 1735 | return 1; | 1735 | return 1; |
| 1736 | } | 1736 | } |
| 1737 | return 0; | 1737 | return 0; |
| 1738 | } | 1738 | } |
| 1739 | 1739 | ||
| 1740 | static void fill_stack_inode_item(struct btrfs_trans_handle *trans, | 1740 | static void fill_stack_inode_item(struct btrfs_trans_handle *trans, |
| 1741 | struct btrfs_inode_item *inode_item, | 1741 | struct btrfs_inode_item *inode_item, |
| 1742 | struct inode *inode) | 1742 | struct inode *inode) |
| 1743 | { | 1743 | { |
| 1744 | btrfs_set_stack_inode_uid(inode_item, i_uid_read(inode)); | 1744 | btrfs_set_stack_inode_uid(inode_item, i_uid_read(inode)); |
| 1745 | btrfs_set_stack_inode_gid(inode_item, i_gid_read(inode)); | 1745 | btrfs_set_stack_inode_gid(inode_item, i_gid_read(inode)); |
| 1746 | btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size); | 1746 | btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size); |
| 1747 | btrfs_set_stack_inode_mode(inode_item, inode->i_mode); | 1747 | btrfs_set_stack_inode_mode(inode_item, inode->i_mode); |
| 1748 | btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink); | 1748 | btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink); |
| 1749 | btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode)); | 1749 | btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode)); |
| 1750 | btrfs_set_stack_inode_generation(inode_item, | 1750 | btrfs_set_stack_inode_generation(inode_item, |
| 1751 | BTRFS_I(inode)->generation); | 1751 | BTRFS_I(inode)->generation); |
| 1752 | btrfs_set_stack_inode_sequence(inode_item, inode->i_version); | 1752 | btrfs_set_stack_inode_sequence(inode_item, inode->i_version); |
| 1753 | btrfs_set_stack_inode_transid(inode_item, trans->transid); | 1753 | btrfs_set_stack_inode_transid(inode_item, trans->transid); |
| 1754 | btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev); | 1754 | btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev); |
| 1755 | btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags); | 1755 | btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags); |
| 1756 | btrfs_set_stack_inode_block_group(inode_item, 0); | 1756 | btrfs_set_stack_inode_block_group(inode_item, 0); |
| 1757 | 1757 | ||
| 1758 | btrfs_set_stack_timespec_sec(btrfs_inode_atime(inode_item), | 1758 | btrfs_set_stack_timespec_sec(btrfs_inode_atime(inode_item), |
| 1759 | inode->i_atime.tv_sec); | 1759 | inode->i_atime.tv_sec); |
| 1760 | btrfs_set_stack_timespec_nsec(btrfs_inode_atime(inode_item), | 1760 | btrfs_set_stack_timespec_nsec(btrfs_inode_atime(inode_item), |
| 1761 | inode->i_atime.tv_nsec); | 1761 | inode->i_atime.tv_nsec); |
| 1762 | 1762 | ||
| 1763 | btrfs_set_stack_timespec_sec(btrfs_inode_mtime(inode_item), | 1763 | btrfs_set_stack_timespec_sec(btrfs_inode_mtime(inode_item), |
| 1764 | inode->i_mtime.tv_sec); | 1764 | inode->i_mtime.tv_sec); |
| 1765 | btrfs_set_stack_timespec_nsec(btrfs_inode_mtime(inode_item), | 1765 | btrfs_set_stack_timespec_nsec(btrfs_inode_mtime(inode_item), |
| 1766 | inode->i_mtime.tv_nsec); | 1766 | inode->i_mtime.tv_nsec); |
| 1767 | 1767 | ||
| 1768 | btrfs_set_stack_timespec_sec(btrfs_inode_ctime(inode_item), | 1768 | btrfs_set_stack_timespec_sec(btrfs_inode_ctime(inode_item), |
| 1769 | inode->i_ctime.tv_sec); | 1769 | inode->i_ctime.tv_sec); |
| 1770 | btrfs_set_stack_timespec_nsec(btrfs_inode_ctime(inode_item), | 1770 | btrfs_set_stack_timespec_nsec(btrfs_inode_ctime(inode_item), |
| 1771 | inode->i_ctime.tv_nsec); | 1771 | inode->i_ctime.tv_nsec); |
| 1772 | } | 1772 | } |
| 1773 | 1773 | ||
| 1774 | int btrfs_fill_inode(struct inode *inode, u32 *rdev) | 1774 | int btrfs_fill_inode(struct inode *inode, u32 *rdev) |
| 1775 | { | 1775 | { |
| 1776 | struct btrfs_delayed_node *delayed_node; | 1776 | struct btrfs_delayed_node *delayed_node; |
| 1777 | struct btrfs_inode_item *inode_item; | 1777 | struct btrfs_inode_item *inode_item; |
| 1778 | struct btrfs_timespec *tspec; | 1778 | struct btrfs_timespec *tspec; |
| 1779 | 1779 | ||
| 1780 | delayed_node = btrfs_get_delayed_node(inode); | 1780 | delayed_node = btrfs_get_delayed_node(inode); |
| 1781 | if (!delayed_node) | 1781 | if (!delayed_node) |
| 1782 | return -ENOENT; | 1782 | return -ENOENT; |
| 1783 | 1783 | ||
| 1784 | mutex_lock(&delayed_node->mutex); | 1784 | mutex_lock(&delayed_node->mutex); |
| 1785 | if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { | 1785 | if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { |
| 1786 | mutex_unlock(&delayed_node->mutex); | 1786 | mutex_unlock(&delayed_node->mutex); |
| 1787 | btrfs_release_delayed_node(delayed_node); | 1787 | btrfs_release_delayed_node(delayed_node); |
| 1788 | return -ENOENT; | 1788 | return -ENOENT; |
| 1789 | } | 1789 | } |
| 1790 | 1790 | ||
| 1791 | inode_item = &delayed_node->inode_item; | 1791 | inode_item = &delayed_node->inode_item; |
| 1792 | 1792 | ||
| 1793 | i_uid_write(inode, btrfs_stack_inode_uid(inode_item)); | 1793 | i_uid_write(inode, btrfs_stack_inode_uid(inode_item)); |
| 1794 | i_gid_write(inode, btrfs_stack_inode_gid(inode_item)); | 1794 | i_gid_write(inode, btrfs_stack_inode_gid(inode_item)); |
| 1795 | btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item)); | 1795 | btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item)); |
| 1796 | inode->i_mode = btrfs_stack_inode_mode(inode_item); | 1796 | inode->i_mode = btrfs_stack_inode_mode(inode_item); |
| 1797 | set_nlink(inode, btrfs_stack_inode_nlink(inode_item)); | 1797 | set_nlink(inode, btrfs_stack_inode_nlink(inode_item)); |
| 1798 | inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item)); | 1798 | inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item)); |
| 1799 | BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item); | 1799 | BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item); |
| 1800 | inode->i_version = btrfs_stack_inode_sequence(inode_item); | 1800 | inode->i_version = btrfs_stack_inode_sequence(inode_item); |
| 1801 | inode->i_rdev = 0; | 1801 | inode->i_rdev = 0; |
| 1802 | *rdev = btrfs_stack_inode_rdev(inode_item); | 1802 | *rdev = btrfs_stack_inode_rdev(inode_item); |
| 1803 | BTRFS_I(inode)->flags = btrfs_stack_inode_flags(inode_item); | 1803 | BTRFS_I(inode)->flags = btrfs_stack_inode_flags(inode_item); |
| 1804 | 1804 | ||
| 1805 | tspec = btrfs_inode_atime(inode_item); | 1805 | tspec = btrfs_inode_atime(inode_item); |
| 1806 | inode->i_atime.tv_sec = btrfs_stack_timespec_sec(tspec); | 1806 | inode->i_atime.tv_sec = btrfs_stack_timespec_sec(tspec); |
| 1807 | inode->i_atime.tv_nsec = btrfs_stack_timespec_nsec(tspec); | 1807 | inode->i_atime.tv_nsec = btrfs_stack_timespec_nsec(tspec); |
| 1808 | 1808 | ||
| 1809 | tspec = btrfs_inode_mtime(inode_item); | 1809 | tspec = btrfs_inode_mtime(inode_item); |
| 1810 | inode->i_mtime.tv_sec = btrfs_stack_timespec_sec(tspec); | 1810 | inode->i_mtime.tv_sec = btrfs_stack_timespec_sec(tspec); |
| 1811 | inode->i_mtime.tv_nsec = btrfs_stack_timespec_nsec(tspec); | 1811 | inode->i_mtime.tv_nsec = btrfs_stack_timespec_nsec(tspec); |
| 1812 | 1812 | ||
| 1813 | tspec = btrfs_inode_ctime(inode_item); | 1813 | tspec = btrfs_inode_ctime(inode_item); |
| 1814 | inode->i_ctime.tv_sec = btrfs_stack_timespec_sec(tspec); | 1814 | inode->i_ctime.tv_sec = btrfs_stack_timespec_sec(tspec); |
| 1815 | inode->i_ctime.tv_nsec = btrfs_stack_timespec_nsec(tspec); | 1815 | inode->i_ctime.tv_nsec = btrfs_stack_timespec_nsec(tspec); |
| 1816 | 1816 | ||
| 1817 | inode->i_generation = BTRFS_I(inode)->generation; | 1817 | inode->i_generation = BTRFS_I(inode)->generation; |
| 1818 | BTRFS_I(inode)->index_cnt = (u64)-1; | 1818 | BTRFS_I(inode)->index_cnt = (u64)-1; |
| 1819 | 1819 | ||
| 1820 | mutex_unlock(&delayed_node->mutex); | 1820 | mutex_unlock(&delayed_node->mutex); |
| 1821 | btrfs_release_delayed_node(delayed_node); | 1821 | btrfs_release_delayed_node(delayed_node); |
| 1822 | return 0; | 1822 | return 0; |
| 1823 | } | 1823 | } |
| 1824 | 1824 | ||
| 1825 | int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, | 1825 | int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, |
| 1826 | struct btrfs_root *root, struct inode *inode) | 1826 | struct btrfs_root *root, struct inode *inode) |
| 1827 | { | 1827 | { |
| 1828 | struct btrfs_delayed_node *delayed_node; | 1828 | struct btrfs_delayed_node *delayed_node; |
| 1829 | int ret = 0; | 1829 | int ret = 0; |
| 1830 | 1830 | ||
| 1831 | delayed_node = btrfs_get_or_create_delayed_node(inode); | 1831 | delayed_node = btrfs_get_or_create_delayed_node(inode); |
| 1832 | if (IS_ERR(delayed_node)) | 1832 | if (IS_ERR(delayed_node)) |
| 1833 | return PTR_ERR(delayed_node); | 1833 | return PTR_ERR(delayed_node); |
| 1834 | 1834 | ||
| 1835 | mutex_lock(&delayed_node->mutex); | 1835 | mutex_lock(&delayed_node->mutex); |
| 1836 | if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { | 1836 | if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { |
| 1837 | fill_stack_inode_item(trans, &delayed_node->inode_item, inode); | 1837 | fill_stack_inode_item(trans, &delayed_node->inode_item, inode); |
| 1838 | goto release_node; | 1838 | goto release_node; |
| 1839 | } | 1839 | } |
| 1840 | 1840 | ||
| 1841 | ret = btrfs_delayed_inode_reserve_metadata(trans, root, inode, | 1841 | ret = btrfs_delayed_inode_reserve_metadata(trans, root, inode, |
| 1842 | delayed_node); | 1842 | delayed_node); |
| 1843 | if (ret) | 1843 | if (ret) |
| 1844 | goto release_node; | 1844 | goto release_node; |
| 1845 | 1845 | ||
| 1846 | fill_stack_inode_item(trans, &delayed_node->inode_item, inode); | 1846 | fill_stack_inode_item(trans, &delayed_node->inode_item, inode); |
| 1847 | set_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags); | 1847 | set_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags); |
| 1848 | delayed_node->count++; | 1848 | delayed_node->count++; |
| 1849 | atomic_inc(&root->fs_info->delayed_root->items); | 1849 | atomic_inc(&root->fs_info->delayed_root->items); |
| 1850 | release_node: | 1850 | release_node: |
| 1851 | mutex_unlock(&delayed_node->mutex); | 1851 | mutex_unlock(&delayed_node->mutex); |
| 1852 | btrfs_release_delayed_node(delayed_node); | 1852 | btrfs_release_delayed_node(delayed_node); |
| 1853 | return ret; | 1853 | return ret; |
| 1854 | } | 1854 | } |
| 1855 | 1855 | ||
| 1856 | int btrfs_delayed_delete_inode_ref(struct inode *inode) | 1856 | int btrfs_delayed_delete_inode_ref(struct inode *inode) |
| 1857 | { | 1857 | { |
| 1858 | struct btrfs_delayed_node *delayed_node; | 1858 | struct btrfs_delayed_node *delayed_node; |
| 1859 | 1859 | ||
| 1860 | /* | ||
| 1861 | * we don't do delayed inode updates during log recovery because it | ||
| 1862 | * leads to enospc problems. This means we also can't do | ||
| 1863 | * delayed inode refs | ||
| 1864 | */ | ||
| 1865 | if (BTRFS_I(inode)->root->fs_info->log_root_recovering) | ||
| 1866 | return -EAGAIN; | ||
| 1867 | |||
| 1860 | delayed_node = btrfs_get_or_create_delayed_node(inode); | 1868 | delayed_node = btrfs_get_or_create_delayed_node(inode); |
| 1861 | if (IS_ERR(delayed_node)) | 1869 | if (IS_ERR(delayed_node)) |
| 1862 | return PTR_ERR(delayed_node); | 1870 | return PTR_ERR(delayed_node); |
| 1863 | 1871 | ||
| 1864 | /* | 1872 | /* |
| 1865 | * We don't reserve space for inode ref deletion is because: | 1873 | * We don't reserve space for inode ref deletion is because: |
| 1866 | * - We ONLY do async inode ref deletion for the inode who has only | 1874 | * - We ONLY do async inode ref deletion for the inode who has only |
| 1867 | * one link(i_nlink == 1), it means there is only one inode ref. | 1875 | * one link(i_nlink == 1), it means there is only one inode ref. |
| 1868 | * And in most case, the inode ref and the inode item are in the | 1876 | * And in most case, the inode ref and the inode item are in the |
| 1869 | * same leaf, and we will deal with them at the same time. | 1877 | * same leaf, and we will deal with them at the same time. |
| 1870 | * Since we are sure we will reserve the space for the inode item, | 1878 | * Since we are sure we will reserve the space for the inode item, |
| 1871 | * it is unnecessary to reserve space for inode ref deletion. | 1879 | * it is unnecessary to reserve space for inode ref deletion. |
| 1872 | * - If the inode ref and the inode item are not in the same leaf, | 1880 | * - If the inode ref and the inode item are not in the same leaf, |
| 1873 | * We also needn't worry about enospc problem, because we reserve | 1881 | * We also needn't worry about enospc problem, because we reserve |
| 1874 | * much more space for the inode update than it needs. | 1882 | * much more space for the inode update than it needs. |
| 1875 | * - At the worst, we can steal some space from the global reservation. | 1883 | * - At the worst, we can steal some space from the global reservation. |
| 1876 | * It is very rare. | 1884 | * It is very rare. |
| 1877 | */ | 1885 | */ |
| 1878 | mutex_lock(&delayed_node->mutex); | 1886 | mutex_lock(&delayed_node->mutex); |
| 1879 | if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags)) | 1887 | if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags)) |
| 1880 | goto release_node; | 1888 | goto release_node; |
| 1881 | 1889 | ||
| 1882 | set_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags); | 1890 | set_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags); |
| 1883 | delayed_node->count++; | 1891 | delayed_node->count++; |
| 1884 | atomic_inc(&BTRFS_I(inode)->root->fs_info->delayed_root->items); | 1892 | atomic_inc(&BTRFS_I(inode)->root->fs_info->delayed_root->items); |
| 1885 | release_node: | 1893 | release_node: |
| 1886 | mutex_unlock(&delayed_node->mutex); | 1894 | mutex_unlock(&delayed_node->mutex); |
| 1887 | btrfs_release_delayed_node(delayed_node); | 1895 | btrfs_release_delayed_node(delayed_node); |
| 1888 | return 0; | 1896 | return 0; |
| 1889 | } | 1897 | } |
| 1890 | 1898 | ||
| 1891 | static void __btrfs_kill_delayed_node(struct btrfs_delayed_node *delayed_node) | 1899 | static void __btrfs_kill_delayed_node(struct btrfs_delayed_node *delayed_node) |
| 1892 | { | 1900 | { |
| 1893 | struct btrfs_root *root = delayed_node->root; | 1901 | struct btrfs_root *root = delayed_node->root; |
| 1894 | struct btrfs_delayed_item *curr_item, *prev_item; | 1902 | struct btrfs_delayed_item *curr_item, *prev_item; |
| 1895 | 1903 | ||
| 1896 | mutex_lock(&delayed_node->mutex); | 1904 | mutex_lock(&delayed_node->mutex); |
| 1897 | curr_item = __btrfs_first_delayed_insertion_item(delayed_node); | 1905 | curr_item = __btrfs_first_delayed_insertion_item(delayed_node); |
| 1898 | while (curr_item) { | 1906 | while (curr_item) { |
| 1899 | btrfs_delayed_item_release_metadata(root, curr_item); | 1907 | btrfs_delayed_item_release_metadata(root, curr_item); |
| 1900 | prev_item = curr_item; | 1908 | prev_item = curr_item; |
| 1901 | curr_item = __btrfs_next_delayed_item(prev_item); | 1909 | curr_item = __btrfs_next_delayed_item(prev_item); |
| 1902 | btrfs_release_delayed_item(prev_item); | 1910 | btrfs_release_delayed_item(prev_item); |
| 1903 | } | 1911 | } |
| 1904 | 1912 | ||
| 1905 | curr_item = __btrfs_first_delayed_deletion_item(delayed_node); | 1913 | curr_item = __btrfs_first_delayed_deletion_item(delayed_node); |
| 1906 | while (curr_item) { | 1914 | while (curr_item) { |
| 1907 | btrfs_delayed_item_release_metadata(root, curr_item); | 1915 | btrfs_delayed_item_release_metadata(root, curr_item); |
| 1908 | prev_item = curr_item; | 1916 | prev_item = curr_item; |
| 1909 | curr_item = __btrfs_next_delayed_item(prev_item); | 1917 | curr_item = __btrfs_next_delayed_item(prev_item); |
| 1910 | btrfs_release_delayed_item(prev_item); | 1918 | btrfs_release_delayed_item(prev_item); |
| 1911 | } | 1919 | } |
| 1912 | 1920 | ||
| 1913 | if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags)) | 1921 | if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags)) |
| 1914 | btrfs_release_delayed_iref(delayed_node); | 1922 | btrfs_release_delayed_iref(delayed_node); |
| 1915 | 1923 | ||
| 1916 | if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { | 1924 | if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) { |
| 1917 | btrfs_delayed_inode_release_metadata(root, delayed_node); | 1925 | btrfs_delayed_inode_release_metadata(root, delayed_node); |
| 1918 | btrfs_release_delayed_inode(delayed_node); | 1926 | btrfs_release_delayed_inode(delayed_node); |
| 1919 | } | 1927 | } |
| 1920 | mutex_unlock(&delayed_node->mutex); | 1928 | mutex_unlock(&delayed_node->mutex); |
| 1921 | } | 1929 | } |
| 1922 | 1930 | ||
| 1923 | void btrfs_kill_delayed_inode_items(struct inode *inode) | 1931 | void btrfs_kill_delayed_inode_items(struct inode *inode) |
| 1924 | { | 1932 | { |
| 1925 | struct btrfs_delayed_node *delayed_node; | 1933 | struct btrfs_delayed_node *delayed_node; |
| 1926 | 1934 | ||
| 1927 | delayed_node = btrfs_get_delayed_node(inode); | 1935 | delayed_node = btrfs_get_delayed_node(inode); |
| 1928 | if (!delayed_node) | 1936 | if (!delayed_node) |
| 1929 | return; | 1937 | return; |
| 1930 | 1938 | ||
| 1931 | __btrfs_kill_delayed_node(delayed_node); | 1939 | __btrfs_kill_delayed_node(delayed_node); |
| 1932 | btrfs_release_delayed_node(delayed_node); | 1940 | btrfs_release_delayed_node(delayed_node); |
| 1933 | } | 1941 | } |
| 1934 | 1942 | ||
| 1935 | void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) | 1943 | void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) |
| 1936 | { | 1944 | { |
| 1937 | u64 inode_id = 0; | 1945 | u64 inode_id = 0; |
| 1938 | struct btrfs_delayed_node *delayed_nodes[8]; | 1946 | struct btrfs_delayed_node *delayed_nodes[8]; |
| 1939 | int i, n; | 1947 | int i, n; |
| 1940 | 1948 | ||
| 1941 | while (1) { | 1949 | while (1) { |
| 1942 | spin_lock(&root->inode_lock); | 1950 | spin_lock(&root->inode_lock); |
| 1943 | n = radix_tree_gang_lookup(&root->delayed_nodes_tree, | 1951 | n = radix_tree_gang_lookup(&root->delayed_nodes_tree, |
| 1944 | (void **)delayed_nodes, inode_id, | 1952 | (void **)delayed_nodes, inode_id, |
| 1945 | ARRAY_SIZE(delayed_nodes)); | 1953 | ARRAY_SIZE(delayed_nodes)); |
| 1946 | if (!n) { | 1954 | if (!n) { |
| 1947 | spin_unlock(&root->inode_lock); | 1955 | spin_unlock(&root->inode_lock); |
| 1948 | break; | 1956 | break; |
| 1949 | } | 1957 | } |
| 1950 | 1958 | ||
| 1951 | inode_id = delayed_nodes[n - 1]->inode_id + 1; | 1959 | inode_id = delayed_nodes[n - 1]->inode_id + 1; |
| 1952 | 1960 | ||
| 1953 | for (i = 0; i < n; i++) | 1961 | for (i = 0; i < n; i++) |
| 1954 | atomic_inc(&delayed_nodes[i]->refs); | 1962 | atomic_inc(&delayed_nodes[i]->refs); |
| 1955 | spin_unlock(&root->inode_lock); | 1963 | spin_unlock(&root->inode_lock); |
| 1956 | 1964 | ||
| 1957 | for (i = 0; i < n; i++) { | 1965 | for (i = 0; i < n; i++) { |
| 1958 | __btrfs_kill_delayed_node(delayed_nodes[i]); | 1966 | __btrfs_kill_delayed_node(delayed_nodes[i]); |
| 1959 | btrfs_release_delayed_node(delayed_nodes[i]); | 1967 | btrfs_release_delayed_node(delayed_nodes[i]); |
| 1960 | } | 1968 | } |
| 1961 | } | 1969 | } |
| 1962 | } | 1970 | } |
| 1963 | 1971 | ||
| 1964 | void btrfs_destroy_delayed_inodes(struct btrfs_root *root) | 1972 | void btrfs_destroy_delayed_inodes(struct btrfs_root *root) |
| 1965 | { | 1973 | { |
| 1966 | struct btrfs_delayed_root *delayed_root; | 1974 | struct btrfs_delayed_root *delayed_root; |
| 1967 | struct btrfs_delayed_node *curr_node, *prev_node; | 1975 | struct btrfs_delayed_node *curr_node, *prev_node; |
| 1968 | 1976 | ||
| 1969 | delayed_root = btrfs_get_delayed_root(root); | 1977 | delayed_root = btrfs_get_delayed_root(root); |
| 1970 | 1978 | ||
| 1971 | curr_node = btrfs_first_delayed_node(delayed_root); | 1979 | curr_node = btrfs_first_delayed_node(delayed_root); |
| 1972 | while (curr_node) { | 1980 | while (curr_node) { |
| 1973 | __btrfs_kill_delayed_node(curr_node); | 1981 | __btrfs_kill_delayed_node(curr_node); |
| 1974 | 1982 | ||
| 1975 | prev_node = curr_node; | 1983 | prev_node = curr_node; |
| 1976 | curr_node = btrfs_next_delayed_node(curr_node); | 1984 | curr_node = btrfs_next_delayed_node(curr_node); |
| 1977 | btrfs_release_delayed_node(prev_node); | 1985 | btrfs_release_delayed_node(prev_node); |
| 1978 | } | 1986 | } |
| 1979 | } | 1987 | } |
| 1980 | 1988 | ||
| 1981 | 1989 |
fs/btrfs/extent-tree.c
| 1 | /* | 1 | /* |
| 2 | * Copyright (C) 2007 Oracle. All rights reserved. | 2 | * Copyright (C) 2007 Oracle. All rights reserved. |
| 3 | * | 3 | * |
| 4 | * This program is free software; you can redistribute it and/or | 4 | * This program is free software; you can redistribute it and/or |
| 5 | * modify it under the terms of the GNU General Public | 5 | * modify it under the terms of the GNU General Public |
| 6 | * License v2 as published by the Free Software Foundation. | 6 | * License v2 as published by the Free Software Foundation. |
| 7 | * | 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, | 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 11 | * General Public License for more details. | 11 | * General Public License for more details. |
| 12 | * | 12 | * |
| 13 | * You should have received a copy of the GNU General Public | 13 | * You should have received a copy of the GNU General Public |
| 14 | * License along with this program; if not, write to the | 14 | * License along with this program; if not, write to the |
| 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 16 | * Boston, MA 021110-1307, USA. | 16 | * Boston, MA 021110-1307, USA. |
| 17 | */ | 17 | */ |
| 18 | #include <linux/sched.h> | 18 | #include <linux/sched.h> |
| 19 | #include <linux/pagemap.h> | 19 | #include <linux/pagemap.h> |
| 20 | #include <linux/writeback.h> | 20 | #include <linux/writeback.h> |
| 21 | #include <linux/blkdev.h> | 21 | #include <linux/blkdev.h> |
| 22 | #include <linux/sort.h> | 22 | #include <linux/sort.h> |
| 23 | #include <linux/rcupdate.h> | 23 | #include <linux/rcupdate.h> |
| 24 | #include <linux/kthread.h> | 24 | #include <linux/kthread.h> |
| 25 | #include <linux/slab.h> | 25 | #include <linux/slab.h> |
| 26 | #include <linux/ratelimit.h> | 26 | #include <linux/ratelimit.h> |
| 27 | #include <linux/percpu_counter.h> | 27 | #include <linux/percpu_counter.h> |
| 28 | #include "hash.h" | 28 | #include "hash.h" |
| 29 | #include "tree-log.h" | 29 | #include "tree-log.h" |
| 30 | #include "disk-io.h" | 30 | #include "disk-io.h" |
| 31 | #include "print-tree.h" | 31 | #include "print-tree.h" |
| 32 | #include "volumes.h" | 32 | #include "volumes.h" |
| 33 | #include "raid56.h" | 33 | #include "raid56.h" |
| 34 | #include "locking.h" | 34 | #include "locking.h" |
| 35 | #include "free-space-cache.h" | 35 | #include "free-space-cache.h" |
| 36 | #include "math.h" | 36 | #include "math.h" |
| 37 | #include "sysfs.h" | 37 | #include "sysfs.h" |
| 38 | #include "qgroup.h" | 38 | #include "qgroup.h" |
| 39 | 39 | ||
| 40 | #undef SCRAMBLE_DELAYED_REFS | 40 | #undef SCRAMBLE_DELAYED_REFS |
| 41 | 41 | ||
| 42 | /* | 42 | /* |
| 43 | * control flags for do_chunk_alloc's force field | 43 | * control flags for do_chunk_alloc's force field |
| 44 | * CHUNK_ALLOC_NO_FORCE means to only allocate a chunk | 44 | * CHUNK_ALLOC_NO_FORCE means to only allocate a chunk |
| 45 | * if we really need one. | 45 | * if we really need one. |
| 46 | * | 46 | * |
| 47 | * CHUNK_ALLOC_LIMITED means to only try and allocate one | 47 | * CHUNK_ALLOC_LIMITED means to only try and allocate one |
| 48 | * if we have very few chunks already allocated. This is | 48 | * if we have very few chunks already allocated. This is |
| 49 | * used as part of the clustering code to help make sure | 49 | * used as part of the clustering code to help make sure |
| 50 | * we have a good pool of storage to cluster in, without | 50 | * we have a good pool of storage to cluster in, without |
| 51 | * filling the FS with empty chunks | 51 | * filling the FS with empty chunks |
| 52 | * | 52 | * |
| 53 | * CHUNK_ALLOC_FORCE means it must try to allocate one | 53 | * CHUNK_ALLOC_FORCE means it must try to allocate one |
| 54 | * | 54 | * |
| 55 | */ | 55 | */ |
| 56 | enum { | 56 | enum { |
| 57 | CHUNK_ALLOC_NO_FORCE = 0, | 57 | CHUNK_ALLOC_NO_FORCE = 0, |
| 58 | CHUNK_ALLOC_LIMITED = 1, | 58 | CHUNK_ALLOC_LIMITED = 1, |
| 59 | CHUNK_ALLOC_FORCE = 2, | 59 | CHUNK_ALLOC_FORCE = 2, |
| 60 | }; | 60 | }; |
| 61 | 61 | ||
| 62 | /* | 62 | /* |
| 63 | * Control how reservations are dealt with. | 63 | * Control how reservations are dealt with. |
| 64 | * | 64 | * |
| 65 | * RESERVE_FREE - freeing a reservation. | 65 | * RESERVE_FREE - freeing a reservation. |
| 66 | * RESERVE_ALLOC - allocating space and we need to update bytes_may_use for | 66 | * RESERVE_ALLOC - allocating space and we need to update bytes_may_use for |
| 67 | * ENOSPC accounting | 67 | * ENOSPC accounting |
| 68 | * RESERVE_ALLOC_NO_ACCOUNT - allocating space and we should not update | 68 | * RESERVE_ALLOC_NO_ACCOUNT - allocating space and we should not update |
| 69 | * bytes_may_use as the ENOSPC accounting is done elsewhere | 69 | * bytes_may_use as the ENOSPC accounting is done elsewhere |
| 70 | */ | 70 | */ |
| 71 | enum { | 71 | enum { |
| 72 | RESERVE_FREE = 0, | 72 | RESERVE_FREE = 0, |
| 73 | RESERVE_ALLOC = 1, | 73 | RESERVE_ALLOC = 1, |
| 74 | RESERVE_ALLOC_NO_ACCOUNT = 2, | 74 | RESERVE_ALLOC_NO_ACCOUNT = 2, |
| 75 | }; | 75 | }; |
| 76 | 76 | ||
| 77 | static int update_block_group(struct btrfs_root *root, | 77 | static int update_block_group(struct btrfs_root *root, |
| 78 | u64 bytenr, u64 num_bytes, int alloc); | 78 | u64 bytenr, u64 num_bytes, int alloc); |
| 79 | static int __btrfs_free_extent(struct btrfs_trans_handle *trans, | 79 | static int __btrfs_free_extent(struct btrfs_trans_handle *trans, |
| 80 | struct btrfs_root *root, | 80 | struct btrfs_root *root, |
| 81 | u64 bytenr, u64 num_bytes, u64 parent, | 81 | u64 bytenr, u64 num_bytes, u64 parent, |
| 82 | u64 root_objectid, u64 owner_objectid, | 82 | u64 root_objectid, u64 owner_objectid, |
| 83 | u64 owner_offset, int refs_to_drop, | 83 | u64 owner_offset, int refs_to_drop, |
| 84 | struct btrfs_delayed_extent_op *extra_op, | 84 | struct btrfs_delayed_extent_op *extra_op, |
| 85 | int no_quota); | 85 | int no_quota); |
| 86 | static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op, | 86 | static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op, |
| 87 | struct extent_buffer *leaf, | 87 | struct extent_buffer *leaf, |
| 88 | struct btrfs_extent_item *ei); | 88 | struct btrfs_extent_item *ei); |
| 89 | static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, | 89 | static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, |
| 90 | struct btrfs_root *root, | 90 | struct btrfs_root *root, |
| 91 | u64 parent, u64 root_objectid, | 91 | u64 parent, u64 root_objectid, |
| 92 | u64 flags, u64 owner, u64 offset, | 92 | u64 flags, u64 owner, u64 offset, |
| 93 | struct btrfs_key *ins, int ref_mod); | 93 | struct btrfs_key *ins, int ref_mod); |
| 94 | static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, | 94 | static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, |
| 95 | struct btrfs_root *root, | 95 | struct btrfs_root *root, |
| 96 | u64 parent, u64 root_objectid, | 96 | u64 parent, u64 root_objectid, |
| 97 | u64 flags, struct btrfs_disk_key *key, | 97 | u64 flags, struct btrfs_disk_key *key, |
| 98 | int level, struct btrfs_key *ins, | 98 | int level, struct btrfs_key *ins, |
| 99 | int no_quota); | 99 | int no_quota); |
| 100 | static int do_chunk_alloc(struct btrfs_trans_handle *trans, | 100 | static int do_chunk_alloc(struct btrfs_trans_handle *trans, |
| 101 | struct btrfs_root *extent_root, u64 flags, | 101 | struct btrfs_root *extent_root, u64 flags, |
| 102 | int force); | 102 | int force); |
| 103 | static int find_next_key(struct btrfs_path *path, int level, | 103 | static int find_next_key(struct btrfs_path *path, int level, |
| 104 | struct btrfs_key *key); | 104 | struct btrfs_key *key); |
| 105 | static void dump_space_info(struct btrfs_space_info *info, u64 bytes, | 105 | static void dump_space_info(struct btrfs_space_info *info, u64 bytes, |
| 106 | int dump_block_groups); | 106 | int dump_block_groups); |
| 107 | static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, | 107 | static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, |
| 108 | u64 num_bytes, int reserve, | 108 | u64 num_bytes, int reserve, |
| 109 | int delalloc); | 109 | int delalloc); |
| 110 | static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, | 110 | static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, |
| 111 | u64 num_bytes); | 111 | u64 num_bytes); |
| 112 | int btrfs_pin_extent(struct btrfs_root *root, | 112 | int btrfs_pin_extent(struct btrfs_root *root, |
| 113 | u64 bytenr, u64 num_bytes, int reserved); | 113 | u64 bytenr, u64 num_bytes, int reserved); |
| 114 | 114 | ||
| 115 | static noinline int | 115 | static noinline int |
| 116 | block_group_cache_done(struct btrfs_block_group_cache *cache) | 116 | block_group_cache_done(struct btrfs_block_group_cache *cache) |
| 117 | { | 117 | { |
| 118 | smp_mb(); | 118 | smp_mb(); |
| 119 | return cache->cached == BTRFS_CACHE_FINISHED || | 119 | return cache->cached == BTRFS_CACHE_FINISHED || |
| 120 | cache->cached == BTRFS_CACHE_ERROR; | 120 | cache->cached == BTRFS_CACHE_ERROR; |
| 121 | } | 121 | } |
| 122 | 122 | ||
| 123 | static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits) | 123 | static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits) |
| 124 | { | 124 | { |
| 125 | return (cache->flags & bits) == bits; | 125 | return (cache->flags & bits) == bits; |
| 126 | } | 126 | } |
| 127 | 127 | ||
| 128 | static void btrfs_get_block_group(struct btrfs_block_group_cache *cache) | 128 | static void btrfs_get_block_group(struct btrfs_block_group_cache *cache) |
| 129 | { | 129 | { |
| 130 | atomic_inc(&cache->count); | 130 | atomic_inc(&cache->count); |
| 131 | } | 131 | } |
| 132 | 132 | ||
| 133 | void btrfs_put_block_group(struct btrfs_block_group_cache *cache) | 133 | void btrfs_put_block_group(struct btrfs_block_group_cache *cache) |
| 134 | { | 134 | { |
| 135 | if (atomic_dec_and_test(&cache->count)) { | 135 | if (atomic_dec_and_test(&cache->count)) { |
| 136 | WARN_ON(cache->pinned > 0); | 136 | WARN_ON(cache->pinned > 0); |
| 137 | WARN_ON(cache->reserved > 0); | 137 | WARN_ON(cache->reserved > 0); |
| 138 | kfree(cache->free_space_ctl); | 138 | kfree(cache->free_space_ctl); |
| 139 | kfree(cache); | 139 | kfree(cache); |
| 140 | } | 140 | } |
| 141 | } | 141 | } |
| 142 | 142 | ||
| 143 | /* | 143 | /* |
| 144 | * this adds the block group to the fs_info rb tree for the block group | 144 | * this adds the block group to the fs_info rb tree for the block group |
| 145 | * cache | 145 | * cache |
| 146 | */ | 146 | */ |
| 147 | static int btrfs_add_block_group_cache(struct btrfs_fs_info *info, | 147 | static int btrfs_add_block_group_cache(struct btrfs_fs_info *info, |
| 148 | struct btrfs_block_group_cache *block_group) | 148 | struct btrfs_block_group_cache *block_group) |
| 149 | { | 149 | { |
| 150 | struct rb_node **p; | 150 | struct rb_node **p; |
| 151 | struct rb_node *parent = NULL; | 151 | struct rb_node *parent = NULL; |
| 152 | struct btrfs_block_group_cache *cache; | 152 | struct btrfs_block_group_cache *cache; |
| 153 | 153 | ||
| 154 | spin_lock(&info->block_group_cache_lock); | 154 | spin_lock(&info->block_group_cache_lock); |
| 155 | p = &info->block_group_cache_tree.rb_node; | 155 | p = &info->block_group_cache_tree.rb_node; |
| 156 | 156 | ||
| 157 | while (*p) { | 157 | while (*p) { |
| 158 | parent = *p; | 158 | parent = *p; |
| 159 | cache = rb_entry(parent, struct btrfs_block_group_cache, | 159 | cache = rb_entry(parent, struct btrfs_block_group_cache, |
| 160 | cache_node); | 160 | cache_node); |
| 161 | if (block_group->key.objectid < cache->key.objectid) { | 161 | if (block_group->key.objectid < cache->key.objectid) { |
| 162 | p = &(*p)->rb_left; | 162 | p = &(*p)->rb_left; |
| 163 | } else if (block_group->key.objectid > cache->key.objectid) { | 163 | } else if (block_group->key.objectid > cache->key.objectid) { |
| 164 | p = &(*p)->rb_right; | 164 | p = &(*p)->rb_right; |
| 165 | } else { | 165 | } else { |
| 166 | spin_unlock(&info->block_group_cache_lock); | 166 | spin_unlock(&info->block_group_cache_lock); |
| 167 | return -EEXIST; | 167 | return -EEXIST; |
| 168 | } | 168 | } |
| 169 | } | 169 | } |
| 170 | 170 | ||
| 171 | rb_link_node(&block_group->cache_node, parent, p); | 171 | rb_link_node(&block_group->cache_node, parent, p); |
| 172 | rb_insert_color(&block_group->cache_node, | 172 | rb_insert_color(&block_group->cache_node, |
| 173 | &info->block_group_cache_tree); | 173 | &info->block_group_cache_tree); |
| 174 | 174 | ||
| 175 | if (info->first_logical_byte > block_group->key.objectid) | 175 | if (info->first_logical_byte > block_group->key.objectid) |
| 176 | info->first_logical_byte = block_group->key.objectid; | 176 | info->first_logical_byte = block_group->key.objectid; |
| 177 | 177 | ||
| 178 | spin_unlock(&info->block_group_cache_lock); | 178 | spin_unlock(&info->block_group_cache_lock); |
| 179 | 179 | ||
| 180 | return 0; | 180 | return 0; |
| 181 | } | 181 | } |
| 182 | 182 | ||
| 183 | /* | 183 | /* |
| 184 | * This will return the block group at or after bytenr if contains is 0, else | 184 | * This will return the block group at or after bytenr if contains is 0, else |
| 185 | * it will return the block group that contains the bytenr | 185 | * it will return the block group that contains the bytenr |
| 186 | */ | 186 | */ |
| 187 | static struct btrfs_block_group_cache * | 187 | static struct btrfs_block_group_cache * |
| 188 | block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr, | 188 | block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr, |
| 189 | int contains) | 189 | int contains) |
| 190 | { | 190 | { |
| 191 | struct btrfs_block_group_cache *cache, *ret = NULL; | 191 | struct btrfs_block_group_cache *cache, *ret = NULL; |
| 192 | struct rb_node *n; | 192 | struct rb_node *n; |
| 193 | u64 end, start; | 193 | u64 end, start; |
| 194 | 194 | ||
| 195 | spin_lock(&info->block_group_cache_lock); | 195 | spin_lock(&info->block_group_cache_lock); |
| 196 | n = info->block_group_cache_tree.rb_node; | 196 | n = info->block_group_cache_tree.rb_node; |
| 197 | 197 | ||
| 198 | while (n) { | 198 | while (n) { |
| 199 | cache = rb_entry(n, struct btrfs_block_group_cache, | 199 | cache = rb_entry(n, struct btrfs_block_group_cache, |
| 200 | cache_node); | 200 | cache_node); |
| 201 | end = cache->key.objectid + cache->key.offset - 1; | 201 | end = cache->key.objectid + cache->key.offset - 1; |
| 202 | start = cache->key.objectid; | 202 | start = cache->key.objectid; |
| 203 | 203 | ||
| 204 | if (bytenr < start) { | 204 | if (bytenr < start) { |
| 205 | if (!contains && (!ret || start < ret->key.objectid)) | 205 | if (!contains && (!ret || start < ret->key.objectid)) |
| 206 | ret = cache; | 206 | ret = cache; |
| 207 | n = n->rb_left; | 207 | n = n->rb_left; |
| 208 | } else if (bytenr > start) { | 208 | } else if (bytenr > start) { |
| 209 | if (contains && bytenr <= end) { | 209 | if (contains && bytenr <= end) { |
| 210 | ret = cache; | 210 | ret = cache; |
| 211 | break; | 211 | break; |
| 212 | } | 212 | } |
| 213 | n = n->rb_right; | 213 | n = n->rb_right; |
| 214 | } else { | 214 | } else { |
| 215 | ret = cache; | 215 | ret = cache; |
| 216 | break; | 216 | break; |
| 217 | } | 217 | } |
| 218 | } | 218 | } |
| 219 | if (ret) { | 219 | if (ret) { |
| 220 | btrfs_get_block_group(ret); | 220 | btrfs_get_block_group(ret); |
| 221 | if (bytenr == 0 && info->first_logical_byte > ret->key.objectid) | 221 | if (bytenr == 0 && info->first_logical_byte > ret->key.objectid) |
| 222 | info->first_logical_byte = ret->key.objectid; | 222 | info->first_logical_byte = ret->key.objectid; |
| 223 | } | 223 | } |
| 224 | spin_unlock(&info->block_group_cache_lock); | 224 | spin_unlock(&info->block_group_cache_lock); |
| 225 | 225 | ||
| 226 | return ret; | 226 | return ret; |
| 227 | } | 227 | } |
| 228 | 228 | ||
| 229 | static int add_excluded_extent(struct btrfs_root *root, | 229 | static int add_excluded_extent(struct btrfs_root *root, |
| 230 | u64 start, u64 num_bytes) | 230 | u64 start, u64 num_bytes) |
| 231 | { | 231 | { |
| 232 | u64 end = start + num_bytes - 1; | 232 | u64 end = start + num_bytes - 1; |
| 233 | set_extent_bits(&root->fs_info->freed_extents[0], | 233 | set_extent_bits(&root->fs_info->freed_extents[0], |
| 234 | start, end, EXTENT_UPTODATE, GFP_NOFS); | 234 | start, end, EXTENT_UPTODATE, GFP_NOFS); |
| 235 | set_extent_bits(&root->fs_info->freed_extents[1], | 235 | set_extent_bits(&root->fs_info->freed_extents[1], |
| 236 | start, end, EXTENT_UPTODATE, GFP_NOFS); | 236 | start, end, EXTENT_UPTODATE, GFP_NOFS); |
| 237 | return 0; | 237 | return 0; |
| 238 | } | 238 | } |
| 239 | 239 | ||
| 240 | static void free_excluded_extents(struct btrfs_root *root, | 240 | static void free_excluded_extents(struct btrfs_root *root, |
| 241 | struct btrfs_block_group_cache *cache) | 241 | struct btrfs_block_group_cache *cache) |
| 242 | { | 242 | { |
| 243 | u64 start, end; | 243 | u64 start, end; |
| 244 | 244 | ||
| 245 | start = cache->key.objectid; | 245 | start = cache->key.objectid; |
| 246 | end = start + cache->key.offset - 1; | 246 | end = start + cache->key.offset - 1; |
| 247 | 247 | ||
| 248 | clear_extent_bits(&root->fs_info->freed_extents[0], | 248 | clear_extent_bits(&root->fs_info->freed_extents[0], |
| 249 | start, end, EXTENT_UPTODATE, GFP_NOFS); | 249 | start, end, EXTENT_UPTODATE, GFP_NOFS); |
| 250 | clear_extent_bits(&root->fs_info->freed_extents[1], | 250 | clear_extent_bits(&root->fs_info->freed_extents[1], |
| 251 | start, end, EXTENT_UPTODATE, GFP_NOFS); | 251 | start, end, EXTENT_UPTODATE, GFP_NOFS); |
| 252 | } | 252 | } |
| 253 | 253 | ||
| 254 | static int exclude_super_stripes(struct btrfs_root *root, | 254 | static int exclude_super_stripes(struct btrfs_root *root, |
| 255 | struct btrfs_block_group_cache *cache) | 255 | struct btrfs_block_group_cache *cache) |
| 256 | { | 256 | { |
| 257 | u64 bytenr; | 257 | u64 bytenr; |
| 258 | u64 *logical; | 258 | u64 *logical; |
| 259 | int stripe_len; | 259 | int stripe_len; |
| 260 | int i, nr, ret; | 260 | int i, nr, ret; |
| 261 | 261 | ||
| 262 | if (cache->key.objectid < BTRFS_SUPER_INFO_OFFSET) { | 262 | if (cache->key.objectid < BTRFS_SUPER_INFO_OFFSET) { |
| 263 | stripe_len = BTRFS_SUPER_INFO_OFFSET - cache->key.objectid; | 263 | stripe_len = BTRFS_SUPER_INFO_OFFSET - cache->key.objectid; |
| 264 | cache->bytes_super += stripe_len; | 264 | cache->bytes_super += stripe_len; |
| 265 | ret = add_excluded_extent(root, cache->key.objectid, | 265 | ret = add_excluded_extent(root, cache->key.objectid, |
| 266 | stripe_len); | 266 | stripe_len); |
| 267 | if (ret) | 267 | if (ret) |
| 268 | return ret; | 268 | return ret; |
| 269 | } | 269 | } |
| 270 | 270 | ||
| 271 | for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { | 271 | for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { |
| 272 | bytenr = btrfs_sb_offset(i); | 272 | bytenr = btrfs_sb_offset(i); |
| 273 | ret = btrfs_rmap_block(&root->fs_info->mapping_tree, | 273 | ret = btrfs_rmap_block(&root->fs_info->mapping_tree, |
| 274 | cache->key.objectid, bytenr, | 274 | cache->key.objectid, bytenr, |
| 275 | 0, &logical, &nr, &stripe_len); | 275 | 0, &logical, &nr, &stripe_len); |
| 276 | if (ret) | 276 | if (ret) |
| 277 | return ret; | 277 | return ret; |
| 278 | 278 | ||
| 279 | while (nr--) { | 279 | while (nr--) { |
| 280 | u64 start, len; | 280 | u64 start, len; |
| 281 | 281 | ||
| 282 | if (logical[nr] > cache->key.objectid + | 282 | if (logical[nr] > cache->key.objectid + |
| 283 | cache->key.offset) | 283 | cache->key.offset) |
| 284 | continue; | 284 | continue; |
| 285 | 285 | ||
| 286 | if (logical[nr] + stripe_len <= cache->key.objectid) | 286 | if (logical[nr] + stripe_len <= cache->key.objectid) |
| 287 | continue; | 287 | continue; |
| 288 | 288 | ||
| 289 | start = logical[nr]; | 289 | start = logical[nr]; |
| 290 | if (start < cache->key.objectid) { | 290 | if (start < cache->key.objectid) { |
| 291 | start = cache->key.objectid; | 291 | start = cache->key.objectid; |
| 292 | len = (logical[nr] + stripe_len) - start; | 292 | len = (logical[nr] + stripe_len) - start; |
| 293 | } else { | 293 | } else { |
| 294 | len = min_t(u64, stripe_len, | 294 | len = min_t(u64, stripe_len, |
| 295 | cache->key.objectid + | 295 | cache->key.objectid + |
| 296 | cache->key.offset - start); | 296 | cache->key.offset - start); |
| 297 | } | 297 | } |
| 298 | 298 | ||
| 299 | cache->bytes_super += len; | 299 | cache->bytes_super += len; |
| 300 | ret = add_excluded_extent(root, start, len); | 300 | ret = add_excluded_extent(root, start, len); |
| 301 | if (ret) { | 301 | if (ret) { |
| 302 | kfree(logical); | 302 | kfree(logical); |
| 303 | return ret; | 303 | return ret; |
| 304 | } | 304 | } |
| 305 | } | 305 | } |
| 306 | 306 | ||
| 307 | kfree(logical); | 307 | kfree(logical); |
| 308 | } | 308 | } |
| 309 | return 0; | 309 | return 0; |
| 310 | } | 310 | } |
| 311 | 311 | ||
| 312 | static struct btrfs_caching_control * | 312 | static struct btrfs_caching_control * |
| 313 | get_caching_control(struct btrfs_block_group_cache *cache) | 313 | get_caching_control(struct btrfs_block_group_cache *cache) |
| 314 | { | 314 | { |
| 315 | struct btrfs_caching_control *ctl; | 315 | struct btrfs_caching_control *ctl; |
| 316 | 316 | ||
| 317 | spin_lock(&cache->lock); | 317 | spin_lock(&cache->lock); |
| 318 | if (!cache->caching_ctl) { | 318 | if (!cache->caching_ctl) { |
| 319 | spin_unlock(&cache->lock); | 319 | spin_unlock(&cache->lock); |
| 320 | return NULL; | 320 | return NULL; |
| 321 | } | 321 | } |
| 322 | 322 | ||
| 323 | ctl = cache->caching_ctl; | 323 | ctl = cache->caching_ctl; |
| 324 | atomic_inc(&ctl->count); | 324 | atomic_inc(&ctl->count); |
| 325 | spin_unlock(&cache->lock); | 325 | spin_unlock(&cache->lock); |
| 326 | return ctl; | 326 | return ctl; |
| 327 | } | 327 | } |
| 328 | 328 | ||
| 329 | static void put_caching_control(struct btrfs_caching_control *ctl) | 329 | static void put_caching_control(struct btrfs_caching_control *ctl) |
| 330 | { | 330 | { |
| 331 | if (atomic_dec_and_test(&ctl->count)) | 331 | if (atomic_dec_and_test(&ctl->count)) |
| 332 | kfree(ctl); | 332 | kfree(ctl); |
| 333 | } | 333 | } |
| 334 | 334 | ||
| 335 | /* | 335 | /* |
| 336 | * this is only called by cache_block_group, since we could have freed extents | 336 | * this is only called by cache_block_group, since we could have freed extents |
| 337 | * we need to check the pinned_extents for any extents that can't be used yet | 337 | * we need to check the pinned_extents for any extents that can't be used yet |
| 338 | * since their free space will be released as soon as the transaction commits. | 338 | * since their free space will be released as soon as the transaction commits. |
| 339 | */ | 339 | */ |
| 340 | static u64 add_new_free_space(struct btrfs_block_group_cache *block_group, | 340 | static u64 add_new_free_space(struct btrfs_block_group_cache *block_group, |
| 341 | struct btrfs_fs_info *info, u64 start, u64 end) | 341 | struct btrfs_fs_info *info, u64 start, u64 end) |
| 342 | { | 342 | { |
| 343 | u64 extent_start, extent_end, size, total_added = 0; | 343 | u64 extent_start, extent_end, size, total_added = 0; |
| 344 | int ret; | 344 | int ret; |
| 345 | 345 | ||
| 346 | while (start < end) { | 346 | while (start < end) { |
| 347 | ret = find_first_extent_bit(info->pinned_extents, start, | 347 | ret = find_first_extent_bit(info->pinned_extents, start, |
| 348 | &extent_start, &extent_end, | 348 | &extent_start, &extent_end, |
| 349 | EXTENT_DIRTY | EXTENT_UPTODATE, | 349 | EXTENT_DIRTY | EXTENT_UPTODATE, |
| 350 | NULL); | 350 | NULL); |
| 351 | if (ret) | 351 | if (ret) |
| 352 | break; | 352 | break; |
| 353 | 353 | ||
| 354 | if (extent_start <= start) { | 354 | if (extent_start <= start) { |
| 355 | start = extent_end + 1; | 355 | start = extent_end + 1; |
| 356 | } else if (extent_start > start && extent_start < end) { | 356 | } else if (extent_start > start && extent_start < end) { |
| 357 | size = extent_start - start; | 357 | size = extent_start - start; |
| 358 | total_added += size; | 358 | total_added += size; |
| 359 | ret = btrfs_add_free_space(block_group, start, | 359 | ret = btrfs_add_free_space(block_group, start, |
| 360 | size); | 360 | size); |
| 361 | BUG_ON(ret); /* -ENOMEM or logic error */ | 361 | BUG_ON(ret); /* -ENOMEM or logic error */ |
| 362 | start = extent_end + 1; | 362 | start = extent_end + 1; |
| 363 | } else { | 363 | } else { |
| 364 | break; | 364 | break; |
| 365 | } | 365 | } |
| 366 | } | 366 | } |
| 367 | 367 | ||
| 368 | if (start < end) { | 368 | if (start < end) { |
| 369 | size = end - start; | 369 | size = end - start; |
| 370 | total_added += size; | 370 | total_added += size; |
| 371 | ret = btrfs_add_free_space(block_group, start, size); | 371 | ret = btrfs_add_free_space(block_group, start, size); |
| 372 | BUG_ON(ret); /* -ENOMEM or logic error */ | 372 | BUG_ON(ret); /* -ENOMEM or logic error */ |
| 373 | } | 373 | } |
| 374 | 374 | ||
| 375 | return total_added; | 375 | return total_added; |
| 376 | } | 376 | } |
| 377 | 377 | ||
| 378 | static noinline void caching_thread(struct btrfs_work *work) | 378 | static noinline void caching_thread(struct btrfs_work *work) |
| 379 | { | 379 | { |
| 380 | struct btrfs_block_group_cache *block_group; | 380 | struct btrfs_block_group_cache *block_group; |
| 381 | struct btrfs_fs_info *fs_info; | 381 | struct btrfs_fs_info *fs_info; |
| 382 | struct btrfs_caching_control *caching_ctl; | 382 | struct btrfs_caching_control *caching_ctl; |
| 383 | struct btrfs_root *extent_root; | 383 | struct btrfs_root *extent_root; |
| 384 | struct btrfs_path *path; | 384 | struct btrfs_path *path; |
| 385 | struct extent_buffer *leaf; | 385 | struct extent_buffer *leaf; |
| 386 | struct btrfs_key key; | 386 | struct btrfs_key key; |
| 387 | u64 total_found = 0; | 387 | u64 total_found = 0; |
| 388 | u64 last = 0; | 388 | u64 last = 0; |
| 389 | u32 nritems; | 389 | u32 nritems; |
| 390 | int ret = -ENOMEM; | 390 | int ret = -ENOMEM; |
| 391 | 391 | ||
| 392 | caching_ctl = container_of(work, struct btrfs_caching_control, work); | 392 | caching_ctl = container_of(work, struct btrfs_caching_control, work); |
| 393 | block_group = caching_ctl->block_group; | 393 | block_group = caching_ctl->block_group; |
| 394 | fs_info = block_group->fs_info; | 394 | fs_info = block_group->fs_info; |
| 395 | extent_root = fs_info->extent_root; | 395 | extent_root = fs_info->extent_root; |
| 396 | 396 | ||
| 397 | path = btrfs_alloc_path(); | 397 | path = btrfs_alloc_path(); |
| 398 | if (!path) | 398 | if (!path) |
| 399 | goto out; | 399 | goto out; |
| 400 | 400 | ||
| 401 | last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET); | 401 | last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET); |
| 402 | 402 | ||
| 403 | /* | 403 | /* |
| 404 | * We don't want to deadlock with somebody trying to allocate a new | 404 | * We don't want to deadlock with somebody trying to allocate a new |
| 405 | * extent for the extent root while also trying to search the extent | 405 | * extent for the extent root while also trying to search the extent |
| 406 | * root to add free space. So we skip locking and search the commit | 406 | * root to add free space. So we skip locking and search the commit |
| 407 | * root, since its read-only | 407 | * root, since its read-only |
| 408 | */ | 408 | */ |
| 409 | path->skip_locking = 1; | 409 | path->skip_locking = 1; |
| 410 | path->search_commit_root = 1; | 410 | path->search_commit_root = 1; |
| 411 | path->reada = 1; | 411 | path->reada = 1; |
| 412 | 412 | ||
| 413 | key.objectid = last; | 413 | key.objectid = last; |
| 414 | key.offset = 0; | 414 | key.offset = 0; |
| 415 | key.type = BTRFS_EXTENT_ITEM_KEY; | 415 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 416 | again: | 416 | again: |
| 417 | mutex_lock(&caching_ctl->mutex); | 417 | mutex_lock(&caching_ctl->mutex); |
| 418 | /* need to make sure the commit_root doesn't disappear */ | 418 | /* need to make sure the commit_root doesn't disappear */ |
| 419 | down_read(&fs_info->commit_root_sem); | 419 | down_read(&fs_info->commit_root_sem); |
| 420 | 420 | ||
| 421 | next: | 421 | next: |
| 422 | ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0); | 422 | ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0); |
| 423 | if (ret < 0) | 423 | if (ret < 0) |
| 424 | goto err; | 424 | goto err; |
| 425 | 425 | ||
| 426 | leaf = path->nodes[0]; | 426 | leaf = path->nodes[0]; |
| 427 | nritems = btrfs_header_nritems(leaf); | 427 | nritems = btrfs_header_nritems(leaf); |
| 428 | 428 | ||
| 429 | while (1) { | 429 | while (1) { |
| 430 | if (btrfs_fs_closing(fs_info) > 1) { | 430 | if (btrfs_fs_closing(fs_info) > 1) { |
| 431 | last = (u64)-1; | 431 | last = (u64)-1; |
| 432 | break; | 432 | break; |
| 433 | } | 433 | } |
| 434 | 434 | ||
| 435 | if (path->slots[0] < nritems) { | 435 | if (path->slots[0] < nritems) { |
| 436 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 436 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 437 | } else { | 437 | } else { |
| 438 | ret = find_next_key(path, 0, &key); | 438 | ret = find_next_key(path, 0, &key); |
| 439 | if (ret) | 439 | if (ret) |
| 440 | break; | 440 | break; |
| 441 | 441 | ||
| 442 | if (need_resched() || | 442 | if (need_resched() || |
| 443 | rwsem_is_contended(&fs_info->commit_root_sem)) { | 443 | rwsem_is_contended(&fs_info->commit_root_sem)) { |
| 444 | caching_ctl->progress = last; | 444 | caching_ctl->progress = last; |
| 445 | btrfs_release_path(path); | 445 | btrfs_release_path(path); |
| 446 | up_read(&fs_info->commit_root_sem); | 446 | up_read(&fs_info->commit_root_sem); |
| 447 | mutex_unlock(&caching_ctl->mutex); | 447 | mutex_unlock(&caching_ctl->mutex); |
| 448 | cond_resched(); | 448 | cond_resched(); |
| 449 | goto again; | 449 | goto again; |
| 450 | } | 450 | } |
| 451 | 451 | ||
| 452 | ret = btrfs_next_leaf(extent_root, path); | 452 | ret = btrfs_next_leaf(extent_root, path); |
| 453 | if (ret < 0) | 453 | if (ret < 0) |
| 454 | goto err; | 454 | goto err; |
| 455 | if (ret) | 455 | if (ret) |
| 456 | break; | 456 | break; |
| 457 | leaf = path->nodes[0]; | 457 | leaf = path->nodes[0]; |
| 458 | nritems = btrfs_header_nritems(leaf); | 458 | nritems = btrfs_header_nritems(leaf); |
| 459 | continue; | 459 | continue; |
| 460 | } | 460 | } |
| 461 | 461 | ||
| 462 | if (key.objectid < last) { | 462 | if (key.objectid < last) { |
| 463 | key.objectid = last; | 463 | key.objectid = last; |
| 464 | key.offset = 0; | 464 | key.offset = 0; |
| 465 | key.type = BTRFS_EXTENT_ITEM_KEY; | 465 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 466 | 466 | ||
| 467 | caching_ctl->progress = last; | 467 | caching_ctl->progress = last; |
| 468 | btrfs_release_path(path); | 468 | btrfs_release_path(path); |
| 469 | goto next; | 469 | goto next; |
| 470 | } | 470 | } |
| 471 | 471 | ||
| 472 | if (key.objectid < block_group->key.objectid) { | 472 | if (key.objectid < block_group->key.objectid) { |
| 473 | path->slots[0]++; | 473 | path->slots[0]++; |
| 474 | continue; | 474 | continue; |
| 475 | } | 475 | } |
| 476 | 476 | ||
| 477 | if (key.objectid >= block_group->key.objectid + | 477 | if (key.objectid >= block_group->key.objectid + |
| 478 | block_group->key.offset) | 478 | block_group->key.offset) |
| 479 | break; | 479 | break; |
| 480 | 480 | ||
| 481 | if (key.type == BTRFS_EXTENT_ITEM_KEY || | 481 | if (key.type == BTRFS_EXTENT_ITEM_KEY || |
| 482 | key.type == BTRFS_METADATA_ITEM_KEY) { | 482 | key.type == BTRFS_METADATA_ITEM_KEY) { |
| 483 | total_found += add_new_free_space(block_group, | 483 | total_found += add_new_free_space(block_group, |
| 484 | fs_info, last, | 484 | fs_info, last, |
| 485 | key.objectid); | 485 | key.objectid); |
| 486 | if (key.type == BTRFS_METADATA_ITEM_KEY) | 486 | if (key.type == BTRFS_METADATA_ITEM_KEY) |
| 487 | last = key.objectid + | 487 | last = key.objectid + |
| 488 | fs_info->tree_root->nodesize; | 488 | fs_info->tree_root->nodesize; |
| 489 | else | 489 | else |
| 490 | last = key.objectid + key.offset; | 490 | last = key.objectid + key.offset; |
| 491 | 491 | ||
| 492 | if (total_found > (1024 * 1024 * 2)) { | 492 | if (total_found > (1024 * 1024 * 2)) { |
| 493 | total_found = 0; | 493 | total_found = 0; |
| 494 | wake_up(&caching_ctl->wait); | 494 | wake_up(&caching_ctl->wait); |
| 495 | } | 495 | } |
| 496 | } | 496 | } |
| 497 | path->slots[0]++; | 497 | path->slots[0]++; |
| 498 | } | 498 | } |
| 499 | ret = 0; | 499 | ret = 0; |
| 500 | 500 | ||
| 501 | total_found += add_new_free_space(block_group, fs_info, last, | 501 | total_found += add_new_free_space(block_group, fs_info, last, |
| 502 | block_group->key.objectid + | 502 | block_group->key.objectid + |
| 503 | block_group->key.offset); | 503 | block_group->key.offset); |
| 504 | caching_ctl->progress = (u64)-1; | 504 | caching_ctl->progress = (u64)-1; |
| 505 | 505 | ||
| 506 | spin_lock(&block_group->lock); | 506 | spin_lock(&block_group->lock); |
| 507 | block_group->caching_ctl = NULL; | 507 | block_group->caching_ctl = NULL; |
| 508 | block_group->cached = BTRFS_CACHE_FINISHED; | 508 | block_group->cached = BTRFS_CACHE_FINISHED; |
| 509 | spin_unlock(&block_group->lock); | 509 | spin_unlock(&block_group->lock); |
| 510 | 510 | ||
| 511 | err: | 511 | err: |
| 512 | btrfs_free_path(path); | 512 | btrfs_free_path(path); |
| 513 | up_read(&fs_info->commit_root_sem); | 513 | up_read(&fs_info->commit_root_sem); |
| 514 | 514 | ||
| 515 | free_excluded_extents(extent_root, block_group); | 515 | free_excluded_extents(extent_root, block_group); |
| 516 | 516 | ||
| 517 | mutex_unlock(&caching_ctl->mutex); | 517 | mutex_unlock(&caching_ctl->mutex); |
| 518 | out: | 518 | out: |
| 519 | if (ret) { | 519 | if (ret) { |
| 520 | spin_lock(&block_group->lock); | 520 | spin_lock(&block_group->lock); |
| 521 | block_group->caching_ctl = NULL; | 521 | block_group->caching_ctl = NULL; |
| 522 | block_group->cached = BTRFS_CACHE_ERROR; | 522 | block_group->cached = BTRFS_CACHE_ERROR; |
| 523 | spin_unlock(&block_group->lock); | 523 | spin_unlock(&block_group->lock); |
| 524 | } | 524 | } |
| 525 | wake_up(&caching_ctl->wait); | 525 | wake_up(&caching_ctl->wait); |
| 526 | 526 | ||
| 527 | put_caching_control(caching_ctl); | 527 | put_caching_control(caching_ctl); |
| 528 | btrfs_put_block_group(block_group); | 528 | btrfs_put_block_group(block_group); |
| 529 | } | 529 | } |
| 530 | 530 | ||
| 531 | static int cache_block_group(struct btrfs_block_group_cache *cache, | 531 | static int cache_block_group(struct btrfs_block_group_cache *cache, |
| 532 | int load_cache_only) | 532 | int load_cache_only) |
| 533 | { | 533 | { |
| 534 | DEFINE_WAIT(wait); | 534 | DEFINE_WAIT(wait); |
| 535 | struct btrfs_fs_info *fs_info = cache->fs_info; | 535 | struct btrfs_fs_info *fs_info = cache->fs_info; |
| 536 | struct btrfs_caching_control *caching_ctl; | 536 | struct btrfs_caching_control *caching_ctl; |
| 537 | int ret = 0; | 537 | int ret = 0; |
| 538 | 538 | ||
| 539 | caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); | 539 | caching_ctl = kzalloc(sizeof(*caching_ctl), GFP_NOFS); |
| 540 | if (!caching_ctl) | 540 | if (!caching_ctl) |
| 541 | return -ENOMEM; | 541 | return -ENOMEM; |
| 542 | 542 | ||
| 543 | INIT_LIST_HEAD(&caching_ctl->list); | 543 | INIT_LIST_HEAD(&caching_ctl->list); |
| 544 | mutex_init(&caching_ctl->mutex); | 544 | mutex_init(&caching_ctl->mutex); |
| 545 | init_waitqueue_head(&caching_ctl->wait); | 545 | init_waitqueue_head(&caching_ctl->wait); |
| 546 | caching_ctl->block_group = cache; | 546 | caching_ctl->block_group = cache; |
| 547 | caching_ctl->progress = cache->key.objectid; | 547 | caching_ctl->progress = cache->key.objectid; |
| 548 | atomic_set(&caching_ctl->count, 1); | 548 | atomic_set(&caching_ctl->count, 1); |
| 549 | btrfs_init_work(&caching_ctl->work, btrfs_cache_helper, | 549 | btrfs_init_work(&caching_ctl->work, btrfs_cache_helper, |
| 550 | caching_thread, NULL, NULL); | 550 | caching_thread, NULL, NULL); |
| 551 | 551 | ||
| 552 | spin_lock(&cache->lock); | 552 | spin_lock(&cache->lock); |
| 553 | /* | 553 | /* |
| 554 | * This should be a rare occasion, but this could happen I think in the | 554 | * This should be a rare occasion, but this could happen I think in the |
| 555 | * case where one thread starts to load the space cache info, and then | 555 | * case where one thread starts to load the space cache info, and then |
| 556 | * some other thread starts a transaction commit which tries to do an | 556 | * some other thread starts a transaction commit which tries to do an |
| 557 | * allocation while the other thread is still loading the space cache | 557 | * allocation while the other thread is still loading the space cache |
| 558 | * info. The previous loop should have kept us from choosing this block | 558 | * info. The previous loop should have kept us from choosing this block |
| 559 | * group, but if we've moved to the state where we will wait on caching | 559 | * group, but if we've moved to the state where we will wait on caching |
| 560 | * block groups we need to first check if we're doing a fast load here, | 560 | * block groups we need to first check if we're doing a fast load here, |
| 561 | * so we can wait for it to finish, otherwise we could end up allocating | 561 | * so we can wait for it to finish, otherwise we could end up allocating |
| 562 | * from a block group who's cache gets evicted for one reason or | 562 | * from a block group who's cache gets evicted for one reason or |
| 563 | * another. | 563 | * another. |
| 564 | */ | 564 | */ |
| 565 | while (cache->cached == BTRFS_CACHE_FAST) { | 565 | while (cache->cached == BTRFS_CACHE_FAST) { |
| 566 | struct btrfs_caching_control *ctl; | 566 | struct btrfs_caching_control *ctl; |
| 567 | 567 | ||
| 568 | ctl = cache->caching_ctl; | 568 | ctl = cache->caching_ctl; |
| 569 | atomic_inc(&ctl->count); | 569 | atomic_inc(&ctl->count); |
| 570 | prepare_to_wait(&ctl->wait, &wait, TASK_UNINTERRUPTIBLE); | 570 | prepare_to_wait(&ctl->wait, &wait, TASK_UNINTERRUPTIBLE); |
| 571 | spin_unlock(&cache->lock); | 571 | spin_unlock(&cache->lock); |
| 572 | 572 | ||
| 573 | schedule(); | 573 | schedule(); |
| 574 | 574 | ||
| 575 | finish_wait(&ctl->wait, &wait); | 575 | finish_wait(&ctl->wait, &wait); |
| 576 | put_caching_control(ctl); | 576 | put_caching_control(ctl); |
| 577 | spin_lock(&cache->lock); | 577 | spin_lock(&cache->lock); |
| 578 | } | 578 | } |
| 579 | 579 | ||
| 580 | if (cache->cached != BTRFS_CACHE_NO) { | 580 | if (cache->cached != BTRFS_CACHE_NO) { |
| 581 | spin_unlock(&cache->lock); | 581 | spin_unlock(&cache->lock); |
| 582 | kfree(caching_ctl); | 582 | kfree(caching_ctl); |
| 583 | return 0; | 583 | return 0; |
| 584 | } | 584 | } |
| 585 | WARN_ON(cache->caching_ctl); | 585 | WARN_ON(cache->caching_ctl); |
| 586 | cache->caching_ctl = caching_ctl; | 586 | cache->caching_ctl = caching_ctl; |
| 587 | cache->cached = BTRFS_CACHE_FAST; | 587 | cache->cached = BTRFS_CACHE_FAST; |
| 588 | spin_unlock(&cache->lock); | 588 | spin_unlock(&cache->lock); |
| 589 | 589 | ||
| 590 | if (fs_info->mount_opt & BTRFS_MOUNT_SPACE_CACHE) { | 590 | if (fs_info->mount_opt & BTRFS_MOUNT_SPACE_CACHE) { |
| 591 | mutex_lock(&caching_ctl->mutex); | 591 | mutex_lock(&caching_ctl->mutex); |
| 592 | ret = load_free_space_cache(fs_info, cache); | 592 | ret = load_free_space_cache(fs_info, cache); |
| 593 | 593 | ||
| 594 | spin_lock(&cache->lock); | 594 | spin_lock(&cache->lock); |
| 595 | if (ret == 1) { | 595 | if (ret == 1) { |
| 596 | cache->caching_ctl = NULL; | 596 | cache->caching_ctl = NULL; |
| 597 | cache->cached = BTRFS_CACHE_FINISHED; | 597 | cache->cached = BTRFS_CACHE_FINISHED; |
| 598 | cache->last_byte_to_unpin = (u64)-1; | 598 | cache->last_byte_to_unpin = (u64)-1; |
| 599 | caching_ctl->progress = (u64)-1; | 599 | caching_ctl->progress = (u64)-1; |
| 600 | } else { | 600 | } else { |
| 601 | if (load_cache_only) { | 601 | if (load_cache_only) { |
| 602 | cache->caching_ctl = NULL; | 602 | cache->caching_ctl = NULL; |
| 603 | cache->cached = BTRFS_CACHE_NO; | 603 | cache->cached = BTRFS_CACHE_NO; |
| 604 | } else { | 604 | } else { |
| 605 | cache->cached = BTRFS_CACHE_STARTED; | 605 | cache->cached = BTRFS_CACHE_STARTED; |
| 606 | cache->has_caching_ctl = 1; | 606 | cache->has_caching_ctl = 1; |
| 607 | } | 607 | } |
| 608 | } | 608 | } |
| 609 | spin_unlock(&cache->lock); | 609 | spin_unlock(&cache->lock); |
| 610 | mutex_unlock(&caching_ctl->mutex); | 610 | mutex_unlock(&caching_ctl->mutex); |
| 611 | 611 | ||
| 612 | wake_up(&caching_ctl->wait); | 612 | wake_up(&caching_ctl->wait); |
| 613 | if (ret == 1) { | 613 | if (ret == 1) { |
| 614 | put_caching_control(caching_ctl); | 614 | put_caching_control(caching_ctl); |
| 615 | free_excluded_extents(fs_info->extent_root, cache); | 615 | free_excluded_extents(fs_info->extent_root, cache); |
| 616 | return 0; | 616 | return 0; |
| 617 | } | 617 | } |
| 618 | } else { | 618 | } else { |
| 619 | /* | 619 | /* |
| 620 | * We are not going to do the fast caching, set cached to the | 620 | * We are not going to do the fast caching, set cached to the |
| 621 | * appropriate value and wakeup any waiters. | 621 | * appropriate value and wakeup any waiters. |
| 622 | */ | 622 | */ |
| 623 | spin_lock(&cache->lock); | 623 | spin_lock(&cache->lock); |
| 624 | if (load_cache_only) { | 624 | if (load_cache_only) { |
| 625 | cache->caching_ctl = NULL; | 625 | cache->caching_ctl = NULL; |
| 626 | cache->cached = BTRFS_CACHE_NO; | 626 | cache->cached = BTRFS_CACHE_NO; |
| 627 | } else { | 627 | } else { |
| 628 | cache->cached = BTRFS_CACHE_STARTED; | 628 | cache->cached = BTRFS_CACHE_STARTED; |
| 629 | cache->has_caching_ctl = 1; | 629 | cache->has_caching_ctl = 1; |
| 630 | } | 630 | } |
| 631 | spin_unlock(&cache->lock); | 631 | spin_unlock(&cache->lock); |
| 632 | wake_up(&caching_ctl->wait); | 632 | wake_up(&caching_ctl->wait); |
| 633 | } | 633 | } |
| 634 | 634 | ||
| 635 | if (load_cache_only) { | 635 | if (load_cache_only) { |
| 636 | put_caching_control(caching_ctl); | 636 | put_caching_control(caching_ctl); |
| 637 | return 0; | 637 | return 0; |
| 638 | } | 638 | } |
| 639 | 639 | ||
| 640 | down_write(&fs_info->commit_root_sem); | 640 | down_write(&fs_info->commit_root_sem); |
| 641 | atomic_inc(&caching_ctl->count); | 641 | atomic_inc(&caching_ctl->count); |
| 642 | list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups); | 642 | list_add_tail(&caching_ctl->list, &fs_info->caching_block_groups); |
| 643 | up_write(&fs_info->commit_root_sem); | 643 | up_write(&fs_info->commit_root_sem); |
| 644 | 644 | ||
| 645 | btrfs_get_block_group(cache); | 645 | btrfs_get_block_group(cache); |
| 646 | 646 | ||
| 647 | btrfs_queue_work(fs_info->caching_workers, &caching_ctl->work); | 647 | btrfs_queue_work(fs_info->caching_workers, &caching_ctl->work); |
| 648 | 648 | ||
| 649 | return ret; | 649 | return ret; |
| 650 | } | 650 | } |
| 651 | 651 | ||
| 652 | /* | 652 | /* |
| 653 | * return the block group that starts at or after bytenr | 653 | * return the block group that starts at or after bytenr |
| 654 | */ | 654 | */ |
| 655 | static struct btrfs_block_group_cache * | 655 | static struct btrfs_block_group_cache * |
| 656 | btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr) | 656 | btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr) |
| 657 | { | 657 | { |
| 658 | struct btrfs_block_group_cache *cache; | 658 | struct btrfs_block_group_cache *cache; |
| 659 | 659 | ||
| 660 | cache = block_group_cache_tree_search(info, bytenr, 0); | 660 | cache = block_group_cache_tree_search(info, bytenr, 0); |
| 661 | 661 | ||
| 662 | return cache; | 662 | return cache; |
| 663 | } | 663 | } |
| 664 | 664 | ||
| 665 | /* | 665 | /* |
| 666 | * return the block group that contains the given bytenr | 666 | * return the block group that contains the given bytenr |
| 667 | */ | 667 | */ |
| 668 | struct btrfs_block_group_cache *btrfs_lookup_block_group( | 668 | struct btrfs_block_group_cache *btrfs_lookup_block_group( |
| 669 | struct btrfs_fs_info *info, | 669 | struct btrfs_fs_info *info, |
| 670 | u64 bytenr) | 670 | u64 bytenr) |
| 671 | { | 671 | { |
| 672 | struct btrfs_block_group_cache *cache; | 672 | struct btrfs_block_group_cache *cache; |
| 673 | 673 | ||
| 674 | cache = block_group_cache_tree_search(info, bytenr, 1); | 674 | cache = block_group_cache_tree_search(info, bytenr, 1); |
| 675 | 675 | ||
| 676 | return cache; | 676 | return cache; |
| 677 | } | 677 | } |
| 678 | 678 | ||
| 679 | static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info, | 679 | static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info, |
| 680 | u64 flags) | 680 | u64 flags) |
| 681 | { | 681 | { |
| 682 | struct list_head *head = &info->space_info; | 682 | struct list_head *head = &info->space_info; |
| 683 | struct btrfs_space_info *found; | 683 | struct btrfs_space_info *found; |
| 684 | 684 | ||
| 685 | flags &= BTRFS_BLOCK_GROUP_TYPE_MASK; | 685 | flags &= BTRFS_BLOCK_GROUP_TYPE_MASK; |
| 686 | 686 | ||
| 687 | rcu_read_lock(); | 687 | rcu_read_lock(); |
| 688 | list_for_each_entry_rcu(found, head, list) { | 688 | list_for_each_entry_rcu(found, head, list) { |
| 689 | if (found->flags & flags) { | 689 | if (found->flags & flags) { |
| 690 | rcu_read_unlock(); | 690 | rcu_read_unlock(); |
| 691 | return found; | 691 | return found; |
| 692 | } | 692 | } |
| 693 | } | 693 | } |
| 694 | rcu_read_unlock(); | 694 | rcu_read_unlock(); |
| 695 | return NULL; | 695 | return NULL; |
| 696 | } | 696 | } |
| 697 | 697 | ||
| 698 | /* | 698 | /* |
| 699 | * after adding space to the filesystem, we need to clear the full flags | 699 | * after adding space to the filesystem, we need to clear the full flags |
| 700 | * on all the space infos. | 700 | * on all the space infos. |
| 701 | */ | 701 | */ |
| 702 | void btrfs_clear_space_info_full(struct btrfs_fs_info *info) | 702 | void btrfs_clear_space_info_full(struct btrfs_fs_info *info) |
| 703 | { | 703 | { |
| 704 | struct list_head *head = &info->space_info; | 704 | struct list_head *head = &info->space_info; |
| 705 | struct btrfs_space_info *found; | 705 | struct btrfs_space_info *found; |
| 706 | 706 | ||
| 707 | rcu_read_lock(); | 707 | rcu_read_lock(); |
| 708 | list_for_each_entry_rcu(found, head, list) | 708 | list_for_each_entry_rcu(found, head, list) |
| 709 | found->full = 0; | 709 | found->full = 0; |
| 710 | rcu_read_unlock(); | 710 | rcu_read_unlock(); |
| 711 | } | 711 | } |
| 712 | 712 | ||
| 713 | /* simple helper to search for an existing data extent at a given offset */ | 713 | /* simple helper to search for an existing data extent at a given offset */ |
| 714 | int btrfs_lookup_data_extent(struct btrfs_root *root, u64 start, u64 len) | 714 | int btrfs_lookup_data_extent(struct btrfs_root *root, u64 start, u64 len) |
| 715 | { | 715 | { |
| 716 | int ret; | 716 | int ret; |
| 717 | struct btrfs_key key; | 717 | struct btrfs_key key; |
| 718 | struct btrfs_path *path; | 718 | struct btrfs_path *path; |
| 719 | 719 | ||
| 720 | path = btrfs_alloc_path(); | 720 | path = btrfs_alloc_path(); |
| 721 | if (!path) | 721 | if (!path) |
| 722 | return -ENOMEM; | 722 | return -ENOMEM; |
| 723 | 723 | ||
| 724 | key.objectid = start; | 724 | key.objectid = start; |
| 725 | key.offset = len; | 725 | key.offset = len; |
| 726 | key.type = BTRFS_EXTENT_ITEM_KEY; | 726 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 727 | ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path, | 727 | ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path, |
| 728 | 0, 0); | 728 | 0, 0); |
| 729 | btrfs_free_path(path); | 729 | btrfs_free_path(path); |
| 730 | return ret; | 730 | return ret; |
| 731 | } | 731 | } |
| 732 | 732 | ||
| 733 | /* | 733 | /* |
| 734 | * helper function to lookup reference count and flags of a tree block. | 734 | * helper function to lookup reference count and flags of a tree block. |
| 735 | * | 735 | * |
| 736 | * the head node for delayed ref is used to store the sum of all the | 736 | * the head node for delayed ref is used to store the sum of all the |
| 737 | * reference count modifications queued up in the rbtree. the head | 737 | * reference count modifications queued up in the rbtree. the head |
| 738 | * node may also store the extent flags to set. This way you can check | 738 | * node may also store the extent flags to set. This way you can check |
| 739 | * to see what the reference count and extent flags would be if all of | 739 | * to see what the reference count and extent flags would be if all of |
| 740 | * the delayed refs are not processed. | 740 | * the delayed refs are not processed. |
| 741 | */ | 741 | */ |
| 742 | int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, | 742 | int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, |
| 743 | struct btrfs_root *root, u64 bytenr, | 743 | struct btrfs_root *root, u64 bytenr, |
| 744 | u64 offset, int metadata, u64 *refs, u64 *flags) | 744 | u64 offset, int metadata, u64 *refs, u64 *flags) |
| 745 | { | 745 | { |
| 746 | struct btrfs_delayed_ref_head *head; | 746 | struct btrfs_delayed_ref_head *head; |
| 747 | struct btrfs_delayed_ref_root *delayed_refs; | 747 | struct btrfs_delayed_ref_root *delayed_refs; |
| 748 | struct btrfs_path *path; | 748 | struct btrfs_path *path; |
| 749 | struct btrfs_extent_item *ei; | 749 | struct btrfs_extent_item *ei; |
| 750 | struct extent_buffer *leaf; | 750 | struct extent_buffer *leaf; |
| 751 | struct btrfs_key key; | 751 | struct btrfs_key key; |
| 752 | u32 item_size; | 752 | u32 item_size; |
| 753 | u64 num_refs; | 753 | u64 num_refs; |
| 754 | u64 extent_flags; | 754 | u64 extent_flags; |
| 755 | int ret; | 755 | int ret; |
| 756 | 756 | ||
| 757 | /* | 757 | /* |
| 758 | * If we don't have skinny metadata, don't bother doing anything | 758 | * If we don't have skinny metadata, don't bother doing anything |
| 759 | * different | 759 | * different |
| 760 | */ | 760 | */ |
| 761 | if (metadata && !btrfs_fs_incompat(root->fs_info, SKINNY_METADATA)) { | 761 | if (metadata && !btrfs_fs_incompat(root->fs_info, SKINNY_METADATA)) { |
| 762 | offset = root->nodesize; | 762 | offset = root->nodesize; |
| 763 | metadata = 0; | 763 | metadata = 0; |
| 764 | } | 764 | } |
| 765 | 765 | ||
| 766 | path = btrfs_alloc_path(); | 766 | path = btrfs_alloc_path(); |
| 767 | if (!path) | 767 | if (!path) |
| 768 | return -ENOMEM; | 768 | return -ENOMEM; |
| 769 | 769 | ||
| 770 | if (!trans) { | 770 | if (!trans) { |
| 771 | path->skip_locking = 1; | 771 | path->skip_locking = 1; |
| 772 | path->search_commit_root = 1; | 772 | path->search_commit_root = 1; |
| 773 | } | 773 | } |
| 774 | 774 | ||
| 775 | search_again: | 775 | search_again: |
| 776 | key.objectid = bytenr; | 776 | key.objectid = bytenr; |
| 777 | key.offset = offset; | 777 | key.offset = offset; |
| 778 | if (metadata) | 778 | if (metadata) |
| 779 | key.type = BTRFS_METADATA_ITEM_KEY; | 779 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 780 | else | 780 | else |
| 781 | key.type = BTRFS_EXTENT_ITEM_KEY; | 781 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 782 | 782 | ||
| 783 | ret = btrfs_search_slot(trans, root->fs_info->extent_root, | 783 | ret = btrfs_search_slot(trans, root->fs_info->extent_root, |
| 784 | &key, path, 0, 0); | 784 | &key, path, 0, 0); |
| 785 | if (ret < 0) | 785 | if (ret < 0) |
| 786 | goto out_free; | 786 | goto out_free; |
| 787 | 787 | ||
| 788 | if (ret > 0 && metadata && key.type == BTRFS_METADATA_ITEM_KEY) { | 788 | if (ret > 0 && metadata && key.type == BTRFS_METADATA_ITEM_KEY) { |
| 789 | if (path->slots[0]) { | 789 | if (path->slots[0]) { |
| 790 | path->slots[0]--; | 790 | path->slots[0]--; |
| 791 | btrfs_item_key_to_cpu(path->nodes[0], &key, | 791 | btrfs_item_key_to_cpu(path->nodes[0], &key, |
| 792 | path->slots[0]); | 792 | path->slots[0]); |
| 793 | if (key.objectid == bytenr && | 793 | if (key.objectid == bytenr && |
| 794 | key.type == BTRFS_EXTENT_ITEM_KEY && | 794 | key.type == BTRFS_EXTENT_ITEM_KEY && |
| 795 | key.offset == root->nodesize) | 795 | key.offset == root->nodesize) |
| 796 | ret = 0; | 796 | ret = 0; |
| 797 | } | 797 | } |
| 798 | } | 798 | } |
| 799 | 799 | ||
| 800 | if (ret == 0) { | 800 | if (ret == 0) { |
| 801 | leaf = path->nodes[0]; | 801 | leaf = path->nodes[0]; |
| 802 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 802 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 803 | if (item_size >= sizeof(*ei)) { | 803 | if (item_size >= sizeof(*ei)) { |
| 804 | ei = btrfs_item_ptr(leaf, path->slots[0], | 804 | ei = btrfs_item_ptr(leaf, path->slots[0], |
| 805 | struct btrfs_extent_item); | 805 | struct btrfs_extent_item); |
| 806 | num_refs = btrfs_extent_refs(leaf, ei); | 806 | num_refs = btrfs_extent_refs(leaf, ei); |
| 807 | extent_flags = btrfs_extent_flags(leaf, ei); | 807 | extent_flags = btrfs_extent_flags(leaf, ei); |
| 808 | } else { | 808 | } else { |
| 809 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 809 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 810 | struct btrfs_extent_item_v0 *ei0; | 810 | struct btrfs_extent_item_v0 *ei0; |
| 811 | BUG_ON(item_size != sizeof(*ei0)); | 811 | BUG_ON(item_size != sizeof(*ei0)); |
| 812 | ei0 = btrfs_item_ptr(leaf, path->slots[0], | 812 | ei0 = btrfs_item_ptr(leaf, path->slots[0], |
| 813 | struct btrfs_extent_item_v0); | 813 | struct btrfs_extent_item_v0); |
| 814 | num_refs = btrfs_extent_refs_v0(leaf, ei0); | 814 | num_refs = btrfs_extent_refs_v0(leaf, ei0); |
| 815 | /* FIXME: this isn't correct for data */ | 815 | /* FIXME: this isn't correct for data */ |
| 816 | extent_flags = BTRFS_BLOCK_FLAG_FULL_BACKREF; | 816 | extent_flags = BTRFS_BLOCK_FLAG_FULL_BACKREF; |
| 817 | #else | 817 | #else |
| 818 | BUG(); | 818 | BUG(); |
| 819 | #endif | 819 | #endif |
| 820 | } | 820 | } |
| 821 | BUG_ON(num_refs == 0); | 821 | BUG_ON(num_refs == 0); |
| 822 | } else { | 822 | } else { |
| 823 | num_refs = 0; | 823 | num_refs = 0; |
| 824 | extent_flags = 0; | 824 | extent_flags = 0; |
| 825 | ret = 0; | 825 | ret = 0; |
| 826 | } | 826 | } |
| 827 | 827 | ||
| 828 | if (!trans) | 828 | if (!trans) |
| 829 | goto out; | 829 | goto out; |
| 830 | 830 | ||
| 831 | delayed_refs = &trans->transaction->delayed_refs; | 831 | delayed_refs = &trans->transaction->delayed_refs; |
| 832 | spin_lock(&delayed_refs->lock); | 832 | spin_lock(&delayed_refs->lock); |
| 833 | head = btrfs_find_delayed_ref_head(trans, bytenr); | 833 | head = btrfs_find_delayed_ref_head(trans, bytenr); |
| 834 | if (head) { | 834 | if (head) { |
| 835 | if (!mutex_trylock(&head->mutex)) { | 835 | if (!mutex_trylock(&head->mutex)) { |
| 836 | atomic_inc(&head->node.refs); | 836 | atomic_inc(&head->node.refs); |
| 837 | spin_unlock(&delayed_refs->lock); | 837 | spin_unlock(&delayed_refs->lock); |
| 838 | 838 | ||
| 839 | btrfs_release_path(path); | 839 | btrfs_release_path(path); |
| 840 | 840 | ||
| 841 | /* | 841 | /* |
| 842 | * Mutex was contended, block until it's released and try | 842 | * Mutex was contended, block until it's released and try |
| 843 | * again | 843 | * again |
| 844 | */ | 844 | */ |
| 845 | mutex_lock(&head->mutex); | 845 | mutex_lock(&head->mutex); |
| 846 | mutex_unlock(&head->mutex); | 846 | mutex_unlock(&head->mutex); |
| 847 | btrfs_put_delayed_ref(&head->node); | 847 | btrfs_put_delayed_ref(&head->node); |
| 848 | goto search_again; | 848 | goto search_again; |
| 849 | } | 849 | } |
| 850 | spin_lock(&head->lock); | 850 | spin_lock(&head->lock); |
| 851 | if (head->extent_op && head->extent_op->update_flags) | 851 | if (head->extent_op && head->extent_op->update_flags) |
| 852 | extent_flags |= head->extent_op->flags_to_set; | 852 | extent_flags |= head->extent_op->flags_to_set; |
| 853 | else | 853 | else |
| 854 | BUG_ON(num_refs == 0); | 854 | BUG_ON(num_refs == 0); |
| 855 | 855 | ||
| 856 | num_refs += head->node.ref_mod; | 856 | num_refs += head->node.ref_mod; |
| 857 | spin_unlock(&head->lock); | 857 | spin_unlock(&head->lock); |
| 858 | mutex_unlock(&head->mutex); | 858 | mutex_unlock(&head->mutex); |
| 859 | } | 859 | } |
| 860 | spin_unlock(&delayed_refs->lock); | 860 | spin_unlock(&delayed_refs->lock); |
| 861 | out: | 861 | out: |
| 862 | WARN_ON(num_refs == 0); | 862 | WARN_ON(num_refs == 0); |
| 863 | if (refs) | 863 | if (refs) |
| 864 | *refs = num_refs; | 864 | *refs = num_refs; |
| 865 | if (flags) | 865 | if (flags) |
| 866 | *flags = extent_flags; | 866 | *flags = extent_flags; |
| 867 | out_free: | 867 | out_free: |
| 868 | btrfs_free_path(path); | 868 | btrfs_free_path(path); |
| 869 | return ret; | 869 | return ret; |
| 870 | } | 870 | } |
| 871 | 871 | ||
| 872 | /* | 872 | /* |
| 873 | * Back reference rules. Back refs have three main goals: | 873 | * Back reference rules. Back refs have three main goals: |
| 874 | * | 874 | * |
| 875 | * 1) differentiate between all holders of references to an extent so that | 875 | * 1) differentiate between all holders of references to an extent so that |
| 876 | * when a reference is dropped we can make sure it was a valid reference | 876 | * when a reference is dropped we can make sure it was a valid reference |
| 877 | * before freeing the extent. | 877 | * before freeing the extent. |
| 878 | * | 878 | * |
| 879 | * 2) Provide enough information to quickly find the holders of an extent | 879 | * 2) Provide enough information to quickly find the holders of an extent |
| 880 | * if we notice a given block is corrupted or bad. | 880 | * if we notice a given block is corrupted or bad. |
| 881 | * | 881 | * |
| 882 | * 3) Make it easy to migrate blocks for FS shrinking or storage pool | 882 | * 3) Make it easy to migrate blocks for FS shrinking or storage pool |
| 883 | * maintenance. This is actually the same as #2, but with a slightly | 883 | * maintenance. This is actually the same as #2, but with a slightly |
| 884 | * different use case. | 884 | * different use case. |
| 885 | * | 885 | * |
| 886 | * There are two kinds of back refs. The implicit back refs is optimized | 886 | * There are two kinds of back refs. The implicit back refs is optimized |
| 887 | * for pointers in non-shared tree blocks. For a given pointer in a block, | 887 | * for pointers in non-shared tree blocks. For a given pointer in a block, |
| 888 | * back refs of this kind provide information about the block's owner tree | 888 | * back refs of this kind provide information about the block's owner tree |
| 889 | * and the pointer's key. These information allow us to find the block by | 889 | * and the pointer's key. These information allow us to find the block by |
| 890 | * b-tree searching. The full back refs is for pointers in tree blocks not | 890 | * b-tree searching. The full back refs is for pointers in tree blocks not |
| 891 | * referenced by their owner trees. The location of tree block is recorded | 891 | * referenced by their owner trees. The location of tree block is recorded |
| 892 | * in the back refs. Actually the full back refs is generic, and can be | 892 | * in the back refs. Actually the full back refs is generic, and can be |
| 893 | * used in all cases the implicit back refs is used. The major shortcoming | 893 | * used in all cases the implicit back refs is used. The major shortcoming |
| 894 | * of the full back refs is its overhead. Every time a tree block gets | 894 | * of the full back refs is its overhead. Every time a tree block gets |
| 895 | * COWed, we have to update back refs entry for all pointers in it. | 895 | * COWed, we have to update back refs entry for all pointers in it. |
| 896 | * | 896 | * |
| 897 | * For a newly allocated tree block, we use implicit back refs for | 897 | * For a newly allocated tree block, we use implicit back refs for |
| 898 | * pointers in it. This means most tree related operations only involve | 898 | * pointers in it. This means most tree related operations only involve |
| 899 | * implicit back refs. For a tree block created in old transaction, the | 899 | * implicit back refs. For a tree block created in old transaction, the |
| 900 | * only way to drop a reference to it is COW it. So we can detect the | 900 | * only way to drop a reference to it is COW it. So we can detect the |
| 901 | * event that tree block loses its owner tree's reference and do the | 901 | * event that tree block loses its owner tree's reference and do the |
| 902 | * back refs conversion. | 902 | * back refs conversion. |
| 903 | * | 903 | * |
| 904 | * When a tree block is COW'd through a tree, there are four cases: | 904 | * When a tree block is COW'd through a tree, there are four cases: |
| 905 | * | 905 | * |
| 906 | * The reference count of the block is one and the tree is the block's | 906 | * The reference count of the block is one and the tree is the block's |
| 907 | * owner tree. Nothing to do in this case. | 907 | * owner tree. Nothing to do in this case. |
| 908 | * | 908 | * |
| 909 | * The reference count of the block is one and the tree is not the | 909 | * The reference count of the block is one and the tree is not the |
| 910 | * block's owner tree. In this case, full back refs is used for pointers | 910 | * block's owner tree. In this case, full back refs is used for pointers |
| 911 | * in the block. Remove these full back refs, add implicit back refs for | 911 | * in the block. Remove these full back refs, add implicit back refs for |
| 912 | * every pointers in the new block. | 912 | * every pointers in the new block. |
| 913 | * | 913 | * |
| 914 | * The reference count of the block is greater than one and the tree is | 914 | * The reference count of the block is greater than one and the tree is |
| 915 | * the block's owner tree. In this case, implicit back refs is used for | 915 | * the block's owner tree. In this case, implicit back refs is used for |
| 916 | * pointers in the block. Add full back refs for every pointers in the | 916 | * pointers in the block. Add full back refs for every pointers in the |
| 917 | * block, increase lower level extents' reference counts. The original | 917 | * block, increase lower level extents' reference counts. The original |
| 918 | * implicit back refs are entailed to the new block. | 918 | * implicit back refs are entailed to the new block. |
| 919 | * | 919 | * |
| 920 | * The reference count of the block is greater than one and the tree is | 920 | * The reference count of the block is greater than one and the tree is |
| 921 | * not the block's owner tree. Add implicit back refs for every pointer in | 921 | * not the block's owner tree. Add implicit back refs for every pointer in |
| 922 | * the new block, increase lower level extents' reference count. | 922 | * the new block, increase lower level extents' reference count. |
| 923 | * | 923 | * |
| 924 | * Back Reference Key composing: | 924 | * Back Reference Key composing: |
| 925 | * | 925 | * |
| 926 | * The key objectid corresponds to the first byte in the extent, | 926 | * The key objectid corresponds to the first byte in the extent, |
| 927 | * The key type is used to differentiate between types of back refs. | 927 | * The key type is used to differentiate between types of back refs. |
| 928 | * There are different meanings of the key offset for different types | 928 | * There are different meanings of the key offset for different types |
| 929 | * of back refs. | 929 | * of back refs. |
| 930 | * | 930 | * |
| 931 | * File extents can be referenced by: | 931 | * File extents can be referenced by: |
| 932 | * | 932 | * |
| 933 | * - multiple snapshots, subvolumes, or different generations in one subvol | 933 | * - multiple snapshots, subvolumes, or different generations in one subvol |
| 934 | * - different files inside a single subvolume | 934 | * - different files inside a single subvolume |
| 935 | * - different offsets inside a file (bookend extents in file.c) | 935 | * - different offsets inside a file (bookend extents in file.c) |
| 936 | * | 936 | * |
| 937 | * The extent ref structure for the implicit back refs has fields for: | 937 | * The extent ref structure for the implicit back refs has fields for: |
| 938 | * | 938 | * |
| 939 | * - Objectid of the subvolume root | 939 | * - Objectid of the subvolume root |
| 940 | * - objectid of the file holding the reference | 940 | * - objectid of the file holding the reference |
| 941 | * - original offset in the file | 941 | * - original offset in the file |
| 942 | * - how many bookend extents | 942 | * - how many bookend extents |
| 943 | * | 943 | * |
| 944 | * The key offset for the implicit back refs is hash of the first | 944 | * The key offset for the implicit back refs is hash of the first |
| 945 | * three fields. | 945 | * three fields. |
| 946 | * | 946 | * |
| 947 | * The extent ref structure for the full back refs has field for: | 947 | * The extent ref structure for the full back refs has field for: |
| 948 | * | 948 | * |
| 949 | * - number of pointers in the tree leaf | 949 | * - number of pointers in the tree leaf |
| 950 | * | 950 | * |
| 951 | * The key offset for the implicit back refs is the first byte of | 951 | * The key offset for the implicit back refs is the first byte of |
| 952 | * the tree leaf | 952 | * the tree leaf |
| 953 | * | 953 | * |
| 954 | * When a file extent is allocated, The implicit back refs is used. | 954 | * When a file extent is allocated, The implicit back refs is used. |
| 955 | * the fields are filled in: | 955 | * the fields are filled in: |
| 956 | * | 956 | * |
| 957 | * (root_key.objectid, inode objectid, offset in file, 1) | 957 | * (root_key.objectid, inode objectid, offset in file, 1) |
| 958 | * | 958 | * |
| 959 | * When a file extent is removed file truncation, we find the | 959 | * When a file extent is removed file truncation, we find the |
| 960 | * corresponding implicit back refs and check the following fields: | 960 | * corresponding implicit back refs and check the following fields: |
| 961 | * | 961 | * |
| 962 | * (btrfs_header_owner(leaf), inode objectid, offset in file) | 962 | * (btrfs_header_owner(leaf), inode objectid, offset in file) |
| 963 | * | 963 | * |
| 964 | * Btree extents can be referenced by: | 964 | * Btree extents can be referenced by: |
| 965 | * | 965 | * |
| 966 | * - Different subvolumes | 966 | * - Different subvolumes |
| 967 | * | 967 | * |
| 968 | * Both the implicit back refs and the full back refs for tree blocks | 968 | * Both the implicit back refs and the full back refs for tree blocks |
| 969 | * only consist of key. The key offset for the implicit back refs is | 969 | * only consist of key. The key offset for the implicit back refs is |
| 970 | * objectid of block's owner tree. The key offset for the full back refs | 970 | * objectid of block's owner tree. The key offset for the full back refs |
| 971 | * is the first byte of parent block. | 971 | * is the first byte of parent block. |
| 972 | * | 972 | * |
| 973 | * When implicit back refs is used, information about the lowest key and | 973 | * When implicit back refs is used, information about the lowest key and |
| 974 | * level of the tree block are required. These information are stored in | 974 | * level of the tree block are required. These information are stored in |
| 975 | * tree block info structure. | 975 | * tree block info structure. |
| 976 | */ | 976 | */ |
| 977 | 977 | ||
| 978 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 978 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 979 | static int convert_extent_item_v0(struct btrfs_trans_handle *trans, | 979 | static int convert_extent_item_v0(struct btrfs_trans_handle *trans, |
| 980 | struct btrfs_root *root, | 980 | struct btrfs_root *root, |
| 981 | struct btrfs_path *path, | 981 | struct btrfs_path *path, |
| 982 | u64 owner, u32 extra_size) | 982 | u64 owner, u32 extra_size) |
| 983 | { | 983 | { |
| 984 | struct btrfs_extent_item *item; | 984 | struct btrfs_extent_item *item; |
| 985 | struct btrfs_extent_item_v0 *ei0; | 985 | struct btrfs_extent_item_v0 *ei0; |
| 986 | struct btrfs_extent_ref_v0 *ref0; | 986 | struct btrfs_extent_ref_v0 *ref0; |
| 987 | struct btrfs_tree_block_info *bi; | 987 | struct btrfs_tree_block_info *bi; |
| 988 | struct extent_buffer *leaf; | 988 | struct extent_buffer *leaf; |
| 989 | struct btrfs_key key; | 989 | struct btrfs_key key; |
| 990 | struct btrfs_key found_key; | 990 | struct btrfs_key found_key; |
| 991 | u32 new_size = sizeof(*item); | 991 | u32 new_size = sizeof(*item); |
| 992 | u64 refs; | 992 | u64 refs; |
| 993 | int ret; | 993 | int ret; |
| 994 | 994 | ||
| 995 | leaf = path->nodes[0]; | 995 | leaf = path->nodes[0]; |
| 996 | BUG_ON(btrfs_item_size_nr(leaf, path->slots[0]) != sizeof(*ei0)); | 996 | BUG_ON(btrfs_item_size_nr(leaf, path->slots[0]) != sizeof(*ei0)); |
| 997 | 997 | ||
| 998 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 998 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 999 | ei0 = btrfs_item_ptr(leaf, path->slots[0], | 999 | ei0 = btrfs_item_ptr(leaf, path->slots[0], |
| 1000 | struct btrfs_extent_item_v0); | 1000 | struct btrfs_extent_item_v0); |
| 1001 | refs = btrfs_extent_refs_v0(leaf, ei0); | 1001 | refs = btrfs_extent_refs_v0(leaf, ei0); |
| 1002 | 1002 | ||
| 1003 | if (owner == (u64)-1) { | 1003 | if (owner == (u64)-1) { |
| 1004 | while (1) { | 1004 | while (1) { |
| 1005 | if (path->slots[0] >= btrfs_header_nritems(leaf)) { | 1005 | if (path->slots[0] >= btrfs_header_nritems(leaf)) { |
| 1006 | ret = btrfs_next_leaf(root, path); | 1006 | ret = btrfs_next_leaf(root, path); |
| 1007 | if (ret < 0) | 1007 | if (ret < 0) |
| 1008 | return ret; | 1008 | return ret; |
| 1009 | BUG_ON(ret > 0); /* Corruption */ | 1009 | BUG_ON(ret > 0); /* Corruption */ |
| 1010 | leaf = path->nodes[0]; | 1010 | leaf = path->nodes[0]; |
| 1011 | } | 1011 | } |
| 1012 | btrfs_item_key_to_cpu(leaf, &found_key, | 1012 | btrfs_item_key_to_cpu(leaf, &found_key, |
| 1013 | path->slots[0]); | 1013 | path->slots[0]); |
| 1014 | BUG_ON(key.objectid != found_key.objectid); | 1014 | BUG_ON(key.objectid != found_key.objectid); |
| 1015 | if (found_key.type != BTRFS_EXTENT_REF_V0_KEY) { | 1015 | if (found_key.type != BTRFS_EXTENT_REF_V0_KEY) { |
| 1016 | path->slots[0]++; | 1016 | path->slots[0]++; |
| 1017 | continue; | 1017 | continue; |
| 1018 | } | 1018 | } |
| 1019 | ref0 = btrfs_item_ptr(leaf, path->slots[0], | 1019 | ref0 = btrfs_item_ptr(leaf, path->slots[0], |
| 1020 | struct btrfs_extent_ref_v0); | 1020 | struct btrfs_extent_ref_v0); |
| 1021 | owner = btrfs_ref_objectid_v0(leaf, ref0); | 1021 | owner = btrfs_ref_objectid_v0(leaf, ref0); |
| 1022 | break; | 1022 | break; |
| 1023 | } | 1023 | } |
| 1024 | } | 1024 | } |
| 1025 | btrfs_release_path(path); | 1025 | btrfs_release_path(path); |
| 1026 | 1026 | ||
| 1027 | if (owner < BTRFS_FIRST_FREE_OBJECTID) | 1027 | if (owner < BTRFS_FIRST_FREE_OBJECTID) |
| 1028 | new_size += sizeof(*bi); | 1028 | new_size += sizeof(*bi); |
| 1029 | 1029 | ||
| 1030 | new_size -= sizeof(*ei0); | 1030 | new_size -= sizeof(*ei0); |
| 1031 | ret = btrfs_search_slot(trans, root, &key, path, | 1031 | ret = btrfs_search_slot(trans, root, &key, path, |
| 1032 | new_size + extra_size, 1); | 1032 | new_size + extra_size, 1); |
| 1033 | if (ret < 0) | 1033 | if (ret < 0) |
| 1034 | return ret; | 1034 | return ret; |
| 1035 | BUG_ON(ret); /* Corruption */ | 1035 | BUG_ON(ret); /* Corruption */ |
| 1036 | 1036 | ||
| 1037 | btrfs_extend_item(root, path, new_size); | 1037 | btrfs_extend_item(root, path, new_size); |
| 1038 | 1038 | ||
| 1039 | leaf = path->nodes[0]; | 1039 | leaf = path->nodes[0]; |
| 1040 | item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 1040 | item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 1041 | btrfs_set_extent_refs(leaf, item, refs); | 1041 | btrfs_set_extent_refs(leaf, item, refs); |
| 1042 | /* FIXME: get real generation */ | 1042 | /* FIXME: get real generation */ |
| 1043 | btrfs_set_extent_generation(leaf, item, 0); | 1043 | btrfs_set_extent_generation(leaf, item, 0); |
| 1044 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 1044 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 1045 | btrfs_set_extent_flags(leaf, item, | 1045 | btrfs_set_extent_flags(leaf, item, |
| 1046 | BTRFS_EXTENT_FLAG_TREE_BLOCK | | 1046 | BTRFS_EXTENT_FLAG_TREE_BLOCK | |
| 1047 | BTRFS_BLOCK_FLAG_FULL_BACKREF); | 1047 | BTRFS_BLOCK_FLAG_FULL_BACKREF); |
| 1048 | bi = (struct btrfs_tree_block_info *)(item + 1); | 1048 | bi = (struct btrfs_tree_block_info *)(item + 1); |
| 1049 | /* FIXME: get first key of the block */ | 1049 | /* FIXME: get first key of the block */ |
| 1050 | memset_extent_buffer(leaf, 0, (unsigned long)bi, sizeof(*bi)); | 1050 | memset_extent_buffer(leaf, 0, (unsigned long)bi, sizeof(*bi)); |
| 1051 | btrfs_set_tree_block_level(leaf, bi, (int)owner); | 1051 | btrfs_set_tree_block_level(leaf, bi, (int)owner); |
| 1052 | } else { | 1052 | } else { |
| 1053 | btrfs_set_extent_flags(leaf, item, BTRFS_EXTENT_FLAG_DATA); | 1053 | btrfs_set_extent_flags(leaf, item, BTRFS_EXTENT_FLAG_DATA); |
| 1054 | } | 1054 | } |
| 1055 | btrfs_mark_buffer_dirty(leaf); | 1055 | btrfs_mark_buffer_dirty(leaf); |
| 1056 | return 0; | 1056 | return 0; |
| 1057 | } | 1057 | } |
| 1058 | #endif | 1058 | #endif |
| 1059 | 1059 | ||
| 1060 | static u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset) | 1060 | static u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset) |
| 1061 | { | 1061 | { |
| 1062 | u32 high_crc = ~(u32)0; | 1062 | u32 high_crc = ~(u32)0; |
| 1063 | u32 low_crc = ~(u32)0; | 1063 | u32 low_crc = ~(u32)0; |
| 1064 | __le64 lenum; | 1064 | __le64 lenum; |
| 1065 | 1065 | ||
| 1066 | lenum = cpu_to_le64(root_objectid); | 1066 | lenum = cpu_to_le64(root_objectid); |
| 1067 | high_crc = btrfs_crc32c(high_crc, &lenum, sizeof(lenum)); | 1067 | high_crc = btrfs_crc32c(high_crc, &lenum, sizeof(lenum)); |
| 1068 | lenum = cpu_to_le64(owner); | 1068 | lenum = cpu_to_le64(owner); |
| 1069 | low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum)); | 1069 | low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum)); |
| 1070 | lenum = cpu_to_le64(offset); | 1070 | lenum = cpu_to_le64(offset); |
| 1071 | low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum)); | 1071 | low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum)); |
| 1072 | 1072 | ||
| 1073 | return ((u64)high_crc << 31) ^ (u64)low_crc; | 1073 | return ((u64)high_crc << 31) ^ (u64)low_crc; |
| 1074 | } | 1074 | } |
| 1075 | 1075 | ||
| 1076 | static u64 hash_extent_data_ref_item(struct extent_buffer *leaf, | 1076 | static u64 hash_extent_data_ref_item(struct extent_buffer *leaf, |
| 1077 | struct btrfs_extent_data_ref *ref) | 1077 | struct btrfs_extent_data_ref *ref) |
| 1078 | { | 1078 | { |
| 1079 | return hash_extent_data_ref(btrfs_extent_data_ref_root(leaf, ref), | 1079 | return hash_extent_data_ref(btrfs_extent_data_ref_root(leaf, ref), |
| 1080 | btrfs_extent_data_ref_objectid(leaf, ref), | 1080 | btrfs_extent_data_ref_objectid(leaf, ref), |
| 1081 | btrfs_extent_data_ref_offset(leaf, ref)); | 1081 | btrfs_extent_data_ref_offset(leaf, ref)); |
| 1082 | } | 1082 | } |
| 1083 | 1083 | ||
| 1084 | static int match_extent_data_ref(struct extent_buffer *leaf, | 1084 | static int match_extent_data_ref(struct extent_buffer *leaf, |
| 1085 | struct btrfs_extent_data_ref *ref, | 1085 | struct btrfs_extent_data_ref *ref, |
| 1086 | u64 root_objectid, u64 owner, u64 offset) | 1086 | u64 root_objectid, u64 owner, u64 offset) |
| 1087 | { | 1087 | { |
| 1088 | if (btrfs_extent_data_ref_root(leaf, ref) != root_objectid || | 1088 | if (btrfs_extent_data_ref_root(leaf, ref) != root_objectid || |
| 1089 | btrfs_extent_data_ref_objectid(leaf, ref) != owner || | 1089 | btrfs_extent_data_ref_objectid(leaf, ref) != owner || |
| 1090 | btrfs_extent_data_ref_offset(leaf, ref) != offset) | 1090 | btrfs_extent_data_ref_offset(leaf, ref) != offset) |
| 1091 | return 0; | 1091 | return 0; |
| 1092 | return 1; | 1092 | return 1; |
| 1093 | } | 1093 | } |
| 1094 | 1094 | ||
| 1095 | static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans, | 1095 | static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans, |
| 1096 | struct btrfs_root *root, | 1096 | struct btrfs_root *root, |
| 1097 | struct btrfs_path *path, | 1097 | struct btrfs_path *path, |
| 1098 | u64 bytenr, u64 parent, | 1098 | u64 bytenr, u64 parent, |
| 1099 | u64 root_objectid, | 1099 | u64 root_objectid, |
| 1100 | u64 owner, u64 offset) | 1100 | u64 owner, u64 offset) |
| 1101 | { | 1101 | { |
| 1102 | struct btrfs_key key; | 1102 | struct btrfs_key key; |
| 1103 | struct btrfs_extent_data_ref *ref; | 1103 | struct btrfs_extent_data_ref *ref; |
| 1104 | struct extent_buffer *leaf; | 1104 | struct extent_buffer *leaf; |
| 1105 | u32 nritems; | 1105 | u32 nritems; |
| 1106 | int ret; | 1106 | int ret; |
| 1107 | int recow; | 1107 | int recow; |
| 1108 | int err = -ENOENT; | 1108 | int err = -ENOENT; |
| 1109 | 1109 | ||
| 1110 | key.objectid = bytenr; | 1110 | key.objectid = bytenr; |
| 1111 | if (parent) { | 1111 | if (parent) { |
| 1112 | key.type = BTRFS_SHARED_DATA_REF_KEY; | 1112 | key.type = BTRFS_SHARED_DATA_REF_KEY; |
| 1113 | key.offset = parent; | 1113 | key.offset = parent; |
| 1114 | } else { | 1114 | } else { |
| 1115 | key.type = BTRFS_EXTENT_DATA_REF_KEY; | 1115 | key.type = BTRFS_EXTENT_DATA_REF_KEY; |
| 1116 | key.offset = hash_extent_data_ref(root_objectid, | 1116 | key.offset = hash_extent_data_ref(root_objectid, |
| 1117 | owner, offset); | 1117 | owner, offset); |
| 1118 | } | 1118 | } |
| 1119 | again: | 1119 | again: |
| 1120 | recow = 0; | 1120 | recow = 0; |
| 1121 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 1121 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 1122 | if (ret < 0) { | 1122 | if (ret < 0) { |
| 1123 | err = ret; | 1123 | err = ret; |
| 1124 | goto fail; | 1124 | goto fail; |
| 1125 | } | 1125 | } |
| 1126 | 1126 | ||
| 1127 | if (parent) { | 1127 | if (parent) { |
| 1128 | if (!ret) | 1128 | if (!ret) |
| 1129 | return 0; | 1129 | return 0; |
| 1130 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 1130 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 1131 | key.type = BTRFS_EXTENT_REF_V0_KEY; | 1131 | key.type = BTRFS_EXTENT_REF_V0_KEY; |
| 1132 | btrfs_release_path(path); | 1132 | btrfs_release_path(path); |
| 1133 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 1133 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 1134 | if (ret < 0) { | 1134 | if (ret < 0) { |
| 1135 | err = ret; | 1135 | err = ret; |
| 1136 | goto fail; | 1136 | goto fail; |
| 1137 | } | 1137 | } |
| 1138 | if (!ret) | 1138 | if (!ret) |
| 1139 | return 0; | 1139 | return 0; |
| 1140 | #endif | 1140 | #endif |
| 1141 | goto fail; | 1141 | goto fail; |
| 1142 | } | 1142 | } |
| 1143 | 1143 | ||
| 1144 | leaf = path->nodes[0]; | 1144 | leaf = path->nodes[0]; |
| 1145 | nritems = btrfs_header_nritems(leaf); | 1145 | nritems = btrfs_header_nritems(leaf); |
| 1146 | while (1) { | 1146 | while (1) { |
| 1147 | if (path->slots[0] >= nritems) { | 1147 | if (path->slots[0] >= nritems) { |
| 1148 | ret = btrfs_next_leaf(root, path); | 1148 | ret = btrfs_next_leaf(root, path); |
| 1149 | if (ret < 0) | 1149 | if (ret < 0) |
| 1150 | err = ret; | 1150 | err = ret; |
| 1151 | if (ret) | 1151 | if (ret) |
| 1152 | goto fail; | 1152 | goto fail; |
| 1153 | 1153 | ||
| 1154 | leaf = path->nodes[0]; | 1154 | leaf = path->nodes[0]; |
| 1155 | nritems = btrfs_header_nritems(leaf); | 1155 | nritems = btrfs_header_nritems(leaf); |
| 1156 | recow = 1; | 1156 | recow = 1; |
| 1157 | } | 1157 | } |
| 1158 | 1158 | ||
| 1159 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 1159 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 1160 | if (key.objectid != bytenr || | 1160 | if (key.objectid != bytenr || |
| 1161 | key.type != BTRFS_EXTENT_DATA_REF_KEY) | 1161 | key.type != BTRFS_EXTENT_DATA_REF_KEY) |
| 1162 | goto fail; | 1162 | goto fail; |
| 1163 | 1163 | ||
| 1164 | ref = btrfs_item_ptr(leaf, path->slots[0], | 1164 | ref = btrfs_item_ptr(leaf, path->slots[0], |
| 1165 | struct btrfs_extent_data_ref); | 1165 | struct btrfs_extent_data_ref); |
| 1166 | 1166 | ||
| 1167 | if (match_extent_data_ref(leaf, ref, root_objectid, | 1167 | if (match_extent_data_ref(leaf, ref, root_objectid, |
| 1168 | owner, offset)) { | 1168 | owner, offset)) { |
| 1169 | if (recow) { | 1169 | if (recow) { |
| 1170 | btrfs_release_path(path); | 1170 | btrfs_release_path(path); |
| 1171 | goto again; | 1171 | goto again; |
| 1172 | } | 1172 | } |
| 1173 | err = 0; | 1173 | err = 0; |
| 1174 | break; | 1174 | break; |
| 1175 | } | 1175 | } |
| 1176 | path->slots[0]++; | 1176 | path->slots[0]++; |
| 1177 | } | 1177 | } |
| 1178 | fail: | 1178 | fail: |
| 1179 | return err; | 1179 | return err; |
| 1180 | } | 1180 | } |
| 1181 | 1181 | ||
| 1182 | static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, | 1182 | static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, |
| 1183 | struct btrfs_root *root, | 1183 | struct btrfs_root *root, |
| 1184 | struct btrfs_path *path, | 1184 | struct btrfs_path *path, |
| 1185 | u64 bytenr, u64 parent, | 1185 | u64 bytenr, u64 parent, |
| 1186 | u64 root_objectid, u64 owner, | 1186 | u64 root_objectid, u64 owner, |
| 1187 | u64 offset, int refs_to_add) | 1187 | u64 offset, int refs_to_add) |
| 1188 | { | 1188 | { |
| 1189 | struct btrfs_key key; | 1189 | struct btrfs_key key; |
| 1190 | struct extent_buffer *leaf; | 1190 | struct extent_buffer *leaf; |
| 1191 | u32 size; | 1191 | u32 size; |
| 1192 | u32 num_refs; | 1192 | u32 num_refs; |
| 1193 | int ret; | 1193 | int ret; |
| 1194 | 1194 | ||
| 1195 | key.objectid = bytenr; | 1195 | key.objectid = bytenr; |
| 1196 | if (parent) { | 1196 | if (parent) { |
| 1197 | key.type = BTRFS_SHARED_DATA_REF_KEY; | 1197 | key.type = BTRFS_SHARED_DATA_REF_KEY; |
| 1198 | key.offset = parent; | 1198 | key.offset = parent; |
| 1199 | size = sizeof(struct btrfs_shared_data_ref); | 1199 | size = sizeof(struct btrfs_shared_data_ref); |
| 1200 | } else { | 1200 | } else { |
| 1201 | key.type = BTRFS_EXTENT_DATA_REF_KEY; | 1201 | key.type = BTRFS_EXTENT_DATA_REF_KEY; |
| 1202 | key.offset = hash_extent_data_ref(root_objectid, | 1202 | key.offset = hash_extent_data_ref(root_objectid, |
| 1203 | owner, offset); | 1203 | owner, offset); |
| 1204 | size = sizeof(struct btrfs_extent_data_ref); | 1204 | size = sizeof(struct btrfs_extent_data_ref); |
| 1205 | } | 1205 | } |
| 1206 | 1206 | ||
| 1207 | ret = btrfs_insert_empty_item(trans, root, path, &key, size); | 1207 | ret = btrfs_insert_empty_item(trans, root, path, &key, size); |
| 1208 | if (ret && ret != -EEXIST) | 1208 | if (ret && ret != -EEXIST) |
| 1209 | goto fail; | 1209 | goto fail; |
| 1210 | 1210 | ||
| 1211 | leaf = path->nodes[0]; | 1211 | leaf = path->nodes[0]; |
| 1212 | if (parent) { | 1212 | if (parent) { |
| 1213 | struct btrfs_shared_data_ref *ref; | 1213 | struct btrfs_shared_data_ref *ref; |
| 1214 | ref = btrfs_item_ptr(leaf, path->slots[0], | 1214 | ref = btrfs_item_ptr(leaf, path->slots[0], |
| 1215 | struct btrfs_shared_data_ref); | 1215 | struct btrfs_shared_data_ref); |
| 1216 | if (ret == 0) { | 1216 | if (ret == 0) { |
| 1217 | btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add); | 1217 | btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add); |
| 1218 | } else { | 1218 | } else { |
| 1219 | num_refs = btrfs_shared_data_ref_count(leaf, ref); | 1219 | num_refs = btrfs_shared_data_ref_count(leaf, ref); |
| 1220 | num_refs += refs_to_add; | 1220 | num_refs += refs_to_add; |
| 1221 | btrfs_set_shared_data_ref_count(leaf, ref, num_refs); | 1221 | btrfs_set_shared_data_ref_count(leaf, ref, num_refs); |
| 1222 | } | 1222 | } |
| 1223 | } else { | 1223 | } else { |
| 1224 | struct btrfs_extent_data_ref *ref; | 1224 | struct btrfs_extent_data_ref *ref; |
| 1225 | while (ret == -EEXIST) { | 1225 | while (ret == -EEXIST) { |
| 1226 | ref = btrfs_item_ptr(leaf, path->slots[0], | 1226 | ref = btrfs_item_ptr(leaf, path->slots[0], |
| 1227 | struct btrfs_extent_data_ref); | 1227 | struct btrfs_extent_data_ref); |
| 1228 | if (match_extent_data_ref(leaf, ref, root_objectid, | 1228 | if (match_extent_data_ref(leaf, ref, root_objectid, |
| 1229 | owner, offset)) | 1229 | owner, offset)) |
| 1230 | break; | 1230 | break; |
| 1231 | btrfs_release_path(path); | 1231 | btrfs_release_path(path); |
| 1232 | key.offset++; | 1232 | key.offset++; |
| 1233 | ret = btrfs_insert_empty_item(trans, root, path, &key, | 1233 | ret = btrfs_insert_empty_item(trans, root, path, &key, |
| 1234 | size); | 1234 | size); |
| 1235 | if (ret && ret != -EEXIST) | 1235 | if (ret && ret != -EEXIST) |
| 1236 | goto fail; | 1236 | goto fail; |
| 1237 | 1237 | ||
| 1238 | leaf = path->nodes[0]; | 1238 | leaf = path->nodes[0]; |
| 1239 | } | 1239 | } |
| 1240 | ref = btrfs_item_ptr(leaf, path->slots[0], | 1240 | ref = btrfs_item_ptr(leaf, path->slots[0], |
| 1241 | struct btrfs_extent_data_ref); | 1241 | struct btrfs_extent_data_ref); |
| 1242 | if (ret == 0) { | 1242 | if (ret == 0) { |
| 1243 | btrfs_set_extent_data_ref_root(leaf, ref, | 1243 | btrfs_set_extent_data_ref_root(leaf, ref, |
| 1244 | root_objectid); | 1244 | root_objectid); |
| 1245 | btrfs_set_extent_data_ref_objectid(leaf, ref, owner); | 1245 | btrfs_set_extent_data_ref_objectid(leaf, ref, owner); |
| 1246 | btrfs_set_extent_data_ref_offset(leaf, ref, offset); | 1246 | btrfs_set_extent_data_ref_offset(leaf, ref, offset); |
| 1247 | btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add); | 1247 | btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add); |
| 1248 | } else { | 1248 | } else { |
| 1249 | num_refs = btrfs_extent_data_ref_count(leaf, ref); | 1249 | num_refs = btrfs_extent_data_ref_count(leaf, ref); |
| 1250 | num_refs += refs_to_add; | 1250 | num_refs += refs_to_add; |
| 1251 | btrfs_set_extent_data_ref_count(leaf, ref, num_refs); | 1251 | btrfs_set_extent_data_ref_count(leaf, ref, num_refs); |
| 1252 | } | 1252 | } |
| 1253 | } | 1253 | } |
| 1254 | btrfs_mark_buffer_dirty(leaf); | 1254 | btrfs_mark_buffer_dirty(leaf); |
| 1255 | ret = 0; | 1255 | ret = 0; |
| 1256 | fail: | 1256 | fail: |
| 1257 | btrfs_release_path(path); | 1257 | btrfs_release_path(path); |
| 1258 | return ret; | 1258 | return ret; |
| 1259 | } | 1259 | } |
| 1260 | 1260 | ||
| 1261 | static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans, | 1261 | static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans, |
| 1262 | struct btrfs_root *root, | 1262 | struct btrfs_root *root, |
| 1263 | struct btrfs_path *path, | 1263 | struct btrfs_path *path, |
| 1264 | int refs_to_drop, int *last_ref) | 1264 | int refs_to_drop, int *last_ref) |
| 1265 | { | 1265 | { |
| 1266 | struct btrfs_key key; | 1266 | struct btrfs_key key; |
| 1267 | struct btrfs_extent_data_ref *ref1 = NULL; | 1267 | struct btrfs_extent_data_ref *ref1 = NULL; |
| 1268 | struct btrfs_shared_data_ref *ref2 = NULL; | 1268 | struct btrfs_shared_data_ref *ref2 = NULL; |
| 1269 | struct extent_buffer *leaf; | 1269 | struct extent_buffer *leaf; |
| 1270 | u32 num_refs = 0; | 1270 | u32 num_refs = 0; |
| 1271 | int ret = 0; | 1271 | int ret = 0; |
| 1272 | 1272 | ||
| 1273 | leaf = path->nodes[0]; | 1273 | leaf = path->nodes[0]; |
| 1274 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 1274 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 1275 | 1275 | ||
| 1276 | if (key.type == BTRFS_EXTENT_DATA_REF_KEY) { | 1276 | if (key.type == BTRFS_EXTENT_DATA_REF_KEY) { |
| 1277 | ref1 = btrfs_item_ptr(leaf, path->slots[0], | 1277 | ref1 = btrfs_item_ptr(leaf, path->slots[0], |
| 1278 | struct btrfs_extent_data_ref); | 1278 | struct btrfs_extent_data_ref); |
| 1279 | num_refs = btrfs_extent_data_ref_count(leaf, ref1); | 1279 | num_refs = btrfs_extent_data_ref_count(leaf, ref1); |
| 1280 | } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) { | 1280 | } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) { |
| 1281 | ref2 = btrfs_item_ptr(leaf, path->slots[0], | 1281 | ref2 = btrfs_item_ptr(leaf, path->slots[0], |
| 1282 | struct btrfs_shared_data_ref); | 1282 | struct btrfs_shared_data_ref); |
| 1283 | num_refs = btrfs_shared_data_ref_count(leaf, ref2); | 1283 | num_refs = btrfs_shared_data_ref_count(leaf, ref2); |
| 1284 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 1284 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 1285 | } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) { | 1285 | } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) { |
| 1286 | struct btrfs_extent_ref_v0 *ref0; | 1286 | struct btrfs_extent_ref_v0 *ref0; |
| 1287 | ref0 = btrfs_item_ptr(leaf, path->slots[0], | 1287 | ref0 = btrfs_item_ptr(leaf, path->slots[0], |
| 1288 | struct btrfs_extent_ref_v0); | 1288 | struct btrfs_extent_ref_v0); |
| 1289 | num_refs = btrfs_ref_count_v0(leaf, ref0); | 1289 | num_refs = btrfs_ref_count_v0(leaf, ref0); |
| 1290 | #endif | 1290 | #endif |
| 1291 | } else { | 1291 | } else { |
| 1292 | BUG(); | 1292 | BUG(); |
| 1293 | } | 1293 | } |
| 1294 | 1294 | ||
| 1295 | BUG_ON(num_refs < refs_to_drop); | 1295 | BUG_ON(num_refs < refs_to_drop); |
| 1296 | num_refs -= refs_to_drop; | 1296 | num_refs -= refs_to_drop; |
| 1297 | 1297 | ||
| 1298 | if (num_refs == 0) { | 1298 | if (num_refs == 0) { |
| 1299 | ret = btrfs_del_item(trans, root, path); | 1299 | ret = btrfs_del_item(trans, root, path); |
| 1300 | *last_ref = 1; | 1300 | *last_ref = 1; |
| 1301 | } else { | 1301 | } else { |
| 1302 | if (key.type == BTRFS_EXTENT_DATA_REF_KEY) | 1302 | if (key.type == BTRFS_EXTENT_DATA_REF_KEY) |
| 1303 | btrfs_set_extent_data_ref_count(leaf, ref1, num_refs); | 1303 | btrfs_set_extent_data_ref_count(leaf, ref1, num_refs); |
| 1304 | else if (key.type == BTRFS_SHARED_DATA_REF_KEY) | 1304 | else if (key.type == BTRFS_SHARED_DATA_REF_KEY) |
| 1305 | btrfs_set_shared_data_ref_count(leaf, ref2, num_refs); | 1305 | btrfs_set_shared_data_ref_count(leaf, ref2, num_refs); |
| 1306 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 1306 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 1307 | else { | 1307 | else { |
| 1308 | struct btrfs_extent_ref_v0 *ref0; | 1308 | struct btrfs_extent_ref_v0 *ref0; |
| 1309 | ref0 = btrfs_item_ptr(leaf, path->slots[0], | 1309 | ref0 = btrfs_item_ptr(leaf, path->slots[0], |
| 1310 | struct btrfs_extent_ref_v0); | 1310 | struct btrfs_extent_ref_v0); |
| 1311 | btrfs_set_ref_count_v0(leaf, ref0, num_refs); | 1311 | btrfs_set_ref_count_v0(leaf, ref0, num_refs); |
| 1312 | } | 1312 | } |
| 1313 | #endif | 1313 | #endif |
| 1314 | btrfs_mark_buffer_dirty(leaf); | 1314 | btrfs_mark_buffer_dirty(leaf); |
| 1315 | } | 1315 | } |
| 1316 | return ret; | 1316 | return ret; |
| 1317 | } | 1317 | } |
| 1318 | 1318 | ||
| 1319 | static noinline u32 extent_data_ref_count(struct btrfs_root *root, | 1319 | static noinline u32 extent_data_ref_count(struct btrfs_root *root, |
| 1320 | struct btrfs_path *path, | 1320 | struct btrfs_path *path, |
| 1321 | struct btrfs_extent_inline_ref *iref) | 1321 | struct btrfs_extent_inline_ref *iref) |
| 1322 | { | 1322 | { |
| 1323 | struct btrfs_key key; | 1323 | struct btrfs_key key; |
| 1324 | struct extent_buffer *leaf; | 1324 | struct extent_buffer *leaf; |
| 1325 | struct btrfs_extent_data_ref *ref1; | 1325 | struct btrfs_extent_data_ref *ref1; |
| 1326 | struct btrfs_shared_data_ref *ref2; | 1326 | struct btrfs_shared_data_ref *ref2; |
| 1327 | u32 num_refs = 0; | 1327 | u32 num_refs = 0; |
| 1328 | 1328 | ||
| 1329 | leaf = path->nodes[0]; | 1329 | leaf = path->nodes[0]; |
| 1330 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 1330 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 1331 | if (iref) { | 1331 | if (iref) { |
| 1332 | if (btrfs_extent_inline_ref_type(leaf, iref) == | 1332 | if (btrfs_extent_inline_ref_type(leaf, iref) == |
| 1333 | BTRFS_EXTENT_DATA_REF_KEY) { | 1333 | BTRFS_EXTENT_DATA_REF_KEY) { |
| 1334 | ref1 = (struct btrfs_extent_data_ref *)(&iref->offset); | 1334 | ref1 = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 1335 | num_refs = btrfs_extent_data_ref_count(leaf, ref1); | 1335 | num_refs = btrfs_extent_data_ref_count(leaf, ref1); |
| 1336 | } else { | 1336 | } else { |
| 1337 | ref2 = (struct btrfs_shared_data_ref *)(iref + 1); | 1337 | ref2 = (struct btrfs_shared_data_ref *)(iref + 1); |
| 1338 | num_refs = btrfs_shared_data_ref_count(leaf, ref2); | 1338 | num_refs = btrfs_shared_data_ref_count(leaf, ref2); |
| 1339 | } | 1339 | } |
| 1340 | } else if (key.type == BTRFS_EXTENT_DATA_REF_KEY) { | 1340 | } else if (key.type == BTRFS_EXTENT_DATA_REF_KEY) { |
| 1341 | ref1 = btrfs_item_ptr(leaf, path->slots[0], | 1341 | ref1 = btrfs_item_ptr(leaf, path->slots[0], |
| 1342 | struct btrfs_extent_data_ref); | 1342 | struct btrfs_extent_data_ref); |
| 1343 | num_refs = btrfs_extent_data_ref_count(leaf, ref1); | 1343 | num_refs = btrfs_extent_data_ref_count(leaf, ref1); |
| 1344 | } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) { | 1344 | } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) { |
| 1345 | ref2 = btrfs_item_ptr(leaf, path->slots[0], | 1345 | ref2 = btrfs_item_ptr(leaf, path->slots[0], |
| 1346 | struct btrfs_shared_data_ref); | 1346 | struct btrfs_shared_data_ref); |
| 1347 | num_refs = btrfs_shared_data_ref_count(leaf, ref2); | 1347 | num_refs = btrfs_shared_data_ref_count(leaf, ref2); |
| 1348 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 1348 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 1349 | } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) { | 1349 | } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) { |
| 1350 | struct btrfs_extent_ref_v0 *ref0; | 1350 | struct btrfs_extent_ref_v0 *ref0; |
| 1351 | ref0 = btrfs_item_ptr(leaf, path->slots[0], | 1351 | ref0 = btrfs_item_ptr(leaf, path->slots[0], |
| 1352 | struct btrfs_extent_ref_v0); | 1352 | struct btrfs_extent_ref_v0); |
| 1353 | num_refs = btrfs_ref_count_v0(leaf, ref0); | 1353 | num_refs = btrfs_ref_count_v0(leaf, ref0); |
| 1354 | #endif | 1354 | #endif |
| 1355 | } else { | 1355 | } else { |
| 1356 | WARN_ON(1); | 1356 | WARN_ON(1); |
| 1357 | } | 1357 | } |
| 1358 | return num_refs; | 1358 | return num_refs; |
| 1359 | } | 1359 | } |
| 1360 | 1360 | ||
| 1361 | static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans, | 1361 | static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans, |
| 1362 | struct btrfs_root *root, | 1362 | struct btrfs_root *root, |
| 1363 | struct btrfs_path *path, | 1363 | struct btrfs_path *path, |
| 1364 | u64 bytenr, u64 parent, | 1364 | u64 bytenr, u64 parent, |
| 1365 | u64 root_objectid) | 1365 | u64 root_objectid) |
| 1366 | { | 1366 | { |
| 1367 | struct btrfs_key key; | 1367 | struct btrfs_key key; |
| 1368 | int ret; | 1368 | int ret; |
| 1369 | 1369 | ||
| 1370 | key.objectid = bytenr; | 1370 | key.objectid = bytenr; |
| 1371 | if (parent) { | 1371 | if (parent) { |
| 1372 | key.type = BTRFS_SHARED_BLOCK_REF_KEY; | 1372 | key.type = BTRFS_SHARED_BLOCK_REF_KEY; |
| 1373 | key.offset = parent; | 1373 | key.offset = parent; |
| 1374 | } else { | 1374 | } else { |
| 1375 | key.type = BTRFS_TREE_BLOCK_REF_KEY; | 1375 | key.type = BTRFS_TREE_BLOCK_REF_KEY; |
| 1376 | key.offset = root_objectid; | 1376 | key.offset = root_objectid; |
| 1377 | } | 1377 | } |
| 1378 | 1378 | ||
| 1379 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 1379 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 1380 | if (ret > 0) | 1380 | if (ret > 0) |
| 1381 | ret = -ENOENT; | 1381 | ret = -ENOENT; |
| 1382 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 1382 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 1383 | if (ret == -ENOENT && parent) { | 1383 | if (ret == -ENOENT && parent) { |
| 1384 | btrfs_release_path(path); | 1384 | btrfs_release_path(path); |
| 1385 | key.type = BTRFS_EXTENT_REF_V0_KEY; | 1385 | key.type = BTRFS_EXTENT_REF_V0_KEY; |
| 1386 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 1386 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 1387 | if (ret > 0) | 1387 | if (ret > 0) |
| 1388 | ret = -ENOENT; | 1388 | ret = -ENOENT; |
| 1389 | } | 1389 | } |
| 1390 | #endif | 1390 | #endif |
| 1391 | return ret; | 1391 | return ret; |
| 1392 | } | 1392 | } |
| 1393 | 1393 | ||
| 1394 | static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans, | 1394 | static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans, |
| 1395 | struct btrfs_root *root, | 1395 | struct btrfs_root *root, |
| 1396 | struct btrfs_path *path, | 1396 | struct btrfs_path *path, |
| 1397 | u64 bytenr, u64 parent, | 1397 | u64 bytenr, u64 parent, |
| 1398 | u64 root_objectid) | 1398 | u64 root_objectid) |
| 1399 | { | 1399 | { |
| 1400 | struct btrfs_key key; | 1400 | struct btrfs_key key; |
| 1401 | int ret; | 1401 | int ret; |
| 1402 | 1402 | ||
| 1403 | key.objectid = bytenr; | 1403 | key.objectid = bytenr; |
| 1404 | if (parent) { | 1404 | if (parent) { |
| 1405 | key.type = BTRFS_SHARED_BLOCK_REF_KEY; | 1405 | key.type = BTRFS_SHARED_BLOCK_REF_KEY; |
| 1406 | key.offset = parent; | 1406 | key.offset = parent; |
| 1407 | } else { | 1407 | } else { |
| 1408 | key.type = BTRFS_TREE_BLOCK_REF_KEY; | 1408 | key.type = BTRFS_TREE_BLOCK_REF_KEY; |
| 1409 | key.offset = root_objectid; | 1409 | key.offset = root_objectid; |
| 1410 | } | 1410 | } |
| 1411 | 1411 | ||
| 1412 | ret = btrfs_insert_empty_item(trans, root, path, &key, 0); | 1412 | ret = btrfs_insert_empty_item(trans, root, path, &key, 0); |
| 1413 | btrfs_release_path(path); | 1413 | btrfs_release_path(path); |
| 1414 | return ret; | 1414 | return ret; |
| 1415 | } | 1415 | } |
| 1416 | 1416 | ||
| 1417 | static inline int extent_ref_type(u64 parent, u64 owner) | 1417 | static inline int extent_ref_type(u64 parent, u64 owner) |
| 1418 | { | 1418 | { |
| 1419 | int type; | 1419 | int type; |
| 1420 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 1420 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 1421 | if (parent > 0) | 1421 | if (parent > 0) |
| 1422 | type = BTRFS_SHARED_BLOCK_REF_KEY; | 1422 | type = BTRFS_SHARED_BLOCK_REF_KEY; |
| 1423 | else | 1423 | else |
| 1424 | type = BTRFS_TREE_BLOCK_REF_KEY; | 1424 | type = BTRFS_TREE_BLOCK_REF_KEY; |
| 1425 | } else { | 1425 | } else { |
| 1426 | if (parent > 0) | 1426 | if (parent > 0) |
| 1427 | type = BTRFS_SHARED_DATA_REF_KEY; | 1427 | type = BTRFS_SHARED_DATA_REF_KEY; |
| 1428 | else | 1428 | else |
| 1429 | type = BTRFS_EXTENT_DATA_REF_KEY; | 1429 | type = BTRFS_EXTENT_DATA_REF_KEY; |
| 1430 | } | 1430 | } |
| 1431 | return type; | 1431 | return type; |
| 1432 | } | 1432 | } |
| 1433 | 1433 | ||
| 1434 | static int find_next_key(struct btrfs_path *path, int level, | 1434 | static int find_next_key(struct btrfs_path *path, int level, |
| 1435 | struct btrfs_key *key) | 1435 | struct btrfs_key *key) |
| 1436 | 1436 | ||
| 1437 | { | 1437 | { |
| 1438 | for (; level < BTRFS_MAX_LEVEL; level++) { | 1438 | for (; level < BTRFS_MAX_LEVEL; level++) { |
| 1439 | if (!path->nodes[level]) | 1439 | if (!path->nodes[level]) |
| 1440 | break; | 1440 | break; |
| 1441 | if (path->slots[level] + 1 >= | 1441 | if (path->slots[level] + 1 >= |
| 1442 | btrfs_header_nritems(path->nodes[level])) | 1442 | btrfs_header_nritems(path->nodes[level])) |
| 1443 | continue; | 1443 | continue; |
| 1444 | if (level == 0) | 1444 | if (level == 0) |
| 1445 | btrfs_item_key_to_cpu(path->nodes[level], key, | 1445 | btrfs_item_key_to_cpu(path->nodes[level], key, |
| 1446 | path->slots[level] + 1); | 1446 | path->slots[level] + 1); |
| 1447 | else | 1447 | else |
| 1448 | btrfs_node_key_to_cpu(path->nodes[level], key, | 1448 | btrfs_node_key_to_cpu(path->nodes[level], key, |
| 1449 | path->slots[level] + 1); | 1449 | path->slots[level] + 1); |
| 1450 | return 0; | 1450 | return 0; |
| 1451 | } | 1451 | } |
| 1452 | return 1; | 1452 | return 1; |
| 1453 | } | 1453 | } |
| 1454 | 1454 | ||
| 1455 | /* | 1455 | /* |
| 1456 | * look for inline back ref. if back ref is found, *ref_ret is set | 1456 | * look for inline back ref. if back ref is found, *ref_ret is set |
| 1457 | * to the address of inline back ref, and 0 is returned. | 1457 | * to the address of inline back ref, and 0 is returned. |
| 1458 | * | 1458 | * |
| 1459 | * if back ref isn't found, *ref_ret is set to the address where it | 1459 | * if back ref isn't found, *ref_ret is set to the address where it |
| 1460 | * should be inserted, and -ENOENT is returned. | 1460 | * should be inserted, and -ENOENT is returned. |
| 1461 | * | 1461 | * |
| 1462 | * if insert is true and there are too many inline back refs, the path | 1462 | * if insert is true and there are too many inline back refs, the path |
| 1463 | * points to the extent item, and -EAGAIN is returned. | 1463 | * points to the extent item, and -EAGAIN is returned. |
| 1464 | * | 1464 | * |
| 1465 | * NOTE: inline back refs are ordered in the same way that back ref | 1465 | * NOTE: inline back refs are ordered in the same way that back ref |
| 1466 | * items in the tree are ordered. | 1466 | * items in the tree are ordered. |
| 1467 | */ | 1467 | */ |
| 1468 | static noinline_for_stack | 1468 | static noinline_for_stack |
| 1469 | int lookup_inline_extent_backref(struct btrfs_trans_handle *trans, | 1469 | int lookup_inline_extent_backref(struct btrfs_trans_handle *trans, |
| 1470 | struct btrfs_root *root, | 1470 | struct btrfs_root *root, |
| 1471 | struct btrfs_path *path, | 1471 | struct btrfs_path *path, |
| 1472 | struct btrfs_extent_inline_ref **ref_ret, | 1472 | struct btrfs_extent_inline_ref **ref_ret, |
| 1473 | u64 bytenr, u64 num_bytes, | 1473 | u64 bytenr, u64 num_bytes, |
| 1474 | u64 parent, u64 root_objectid, | 1474 | u64 parent, u64 root_objectid, |
| 1475 | u64 owner, u64 offset, int insert) | 1475 | u64 owner, u64 offset, int insert) |
| 1476 | { | 1476 | { |
| 1477 | struct btrfs_key key; | 1477 | struct btrfs_key key; |
| 1478 | struct extent_buffer *leaf; | 1478 | struct extent_buffer *leaf; |
| 1479 | struct btrfs_extent_item *ei; | 1479 | struct btrfs_extent_item *ei; |
| 1480 | struct btrfs_extent_inline_ref *iref; | 1480 | struct btrfs_extent_inline_ref *iref; |
| 1481 | u64 flags; | 1481 | u64 flags; |
| 1482 | u64 item_size; | 1482 | u64 item_size; |
| 1483 | unsigned long ptr; | 1483 | unsigned long ptr; |
| 1484 | unsigned long end; | 1484 | unsigned long end; |
| 1485 | int extra_size; | 1485 | int extra_size; |
| 1486 | int type; | 1486 | int type; |
| 1487 | int want; | 1487 | int want; |
| 1488 | int ret; | 1488 | int ret; |
| 1489 | int err = 0; | 1489 | int err = 0; |
| 1490 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, | 1490 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, |
| 1491 | SKINNY_METADATA); | 1491 | SKINNY_METADATA); |
| 1492 | 1492 | ||
| 1493 | key.objectid = bytenr; | 1493 | key.objectid = bytenr; |
| 1494 | key.type = BTRFS_EXTENT_ITEM_KEY; | 1494 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 1495 | key.offset = num_bytes; | 1495 | key.offset = num_bytes; |
| 1496 | 1496 | ||
| 1497 | want = extent_ref_type(parent, owner); | 1497 | want = extent_ref_type(parent, owner); |
| 1498 | if (insert) { | 1498 | if (insert) { |
| 1499 | extra_size = btrfs_extent_inline_ref_size(want); | 1499 | extra_size = btrfs_extent_inline_ref_size(want); |
| 1500 | path->keep_locks = 1; | 1500 | path->keep_locks = 1; |
| 1501 | } else | 1501 | } else |
| 1502 | extra_size = -1; | 1502 | extra_size = -1; |
| 1503 | 1503 | ||
| 1504 | /* | 1504 | /* |
| 1505 | * Owner is our parent level, so we can just add one to get the level | 1505 | * Owner is our parent level, so we can just add one to get the level |
| 1506 | * for the block we are interested in. | 1506 | * for the block we are interested in. |
| 1507 | */ | 1507 | */ |
| 1508 | if (skinny_metadata && owner < BTRFS_FIRST_FREE_OBJECTID) { | 1508 | if (skinny_metadata && owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 1509 | key.type = BTRFS_METADATA_ITEM_KEY; | 1509 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 1510 | key.offset = owner; | 1510 | key.offset = owner; |
| 1511 | } | 1511 | } |
| 1512 | 1512 | ||
| 1513 | again: | 1513 | again: |
| 1514 | ret = btrfs_search_slot(trans, root, &key, path, extra_size, 1); | 1514 | ret = btrfs_search_slot(trans, root, &key, path, extra_size, 1); |
| 1515 | if (ret < 0) { | 1515 | if (ret < 0) { |
| 1516 | err = ret; | 1516 | err = ret; |
| 1517 | goto out; | 1517 | goto out; |
| 1518 | } | 1518 | } |
| 1519 | 1519 | ||
| 1520 | /* | 1520 | /* |
| 1521 | * We may be a newly converted file system which still has the old fat | 1521 | * We may be a newly converted file system which still has the old fat |
| 1522 | * extent entries for metadata, so try and see if we have one of those. | 1522 | * extent entries for metadata, so try and see if we have one of those. |
| 1523 | */ | 1523 | */ |
| 1524 | if (ret > 0 && skinny_metadata) { | 1524 | if (ret > 0 && skinny_metadata) { |
| 1525 | skinny_metadata = false; | 1525 | skinny_metadata = false; |
| 1526 | if (path->slots[0]) { | 1526 | if (path->slots[0]) { |
| 1527 | path->slots[0]--; | 1527 | path->slots[0]--; |
| 1528 | btrfs_item_key_to_cpu(path->nodes[0], &key, | 1528 | btrfs_item_key_to_cpu(path->nodes[0], &key, |
| 1529 | path->slots[0]); | 1529 | path->slots[0]); |
| 1530 | if (key.objectid == bytenr && | 1530 | if (key.objectid == bytenr && |
| 1531 | key.type == BTRFS_EXTENT_ITEM_KEY && | 1531 | key.type == BTRFS_EXTENT_ITEM_KEY && |
| 1532 | key.offset == num_bytes) | 1532 | key.offset == num_bytes) |
| 1533 | ret = 0; | 1533 | ret = 0; |
| 1534 | } | 1534 | } |
| 1535 | if (ret) { | 1535 | if (ret) { |
| 1536 | key.objectid = bytenr; | 1536 | key.objectid = bytenr; |
| 1537 | key.type = BTRFS_EXTENT_ITEM_KEY; | 1537 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 1538 | key.offset = num_bytes; | 1538 | key.offset = num_bytes; |
| 1539 | btrfs_release_path(path); | 1539 | btrfs_release_path(path); |
| 1540 | goto again; | 1540 | goto again; |
| 1541 | } | 1541 | } |
| 1542 | } | 1542 | } |
| 1543 | 1543 | ||
| 1544 | if (ret && !insert) { | 1544 | if (ret && !insert) { |
| 1545 | err = -ENOENT; | 1545 | err = -ENOENT; |
| 1546 | goto out; | 1546 | goto out; |
| 1547 | } else if (WARN_ON(ret)) { | 1547 | } else if (WARN_ON(ret)) { |
| 1548 | err = -EIO; | 1548 | err = -EIO; |
| 1549 | goto out; | 1549 | goto out; |
| 1550 | } | 1550 | } |
| 1551 | 1551 | ||
| 1552 | leaf = path->nodes[0]; | 1552 | leaf = path->nodes[0]; |
| 1553 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 1553 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 1554 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 1554 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 1555 | if (item_size < sizeof(*ei)) { | 1555 | if (item_size < sizeof(*ei)) { |
| 1556 | if (!insert) { | 1556 | if (!insert) { |
| 1557 | err = -ENOENT; | 1557 | err = -ENOENT; |
| 1558 | goto out; | 1558 | goto out; |
| 1559 | } | 1559 | } |
| 1560 | ret = convert_extent_item_v0(trans, root, path, owner, | 1560 | ret = convert_extent_item_v0(trans, root, path, owner, |
| 1561 | extra_size); | 1561 | extra_size); |
| 1562 | if (ret < 0) { | 1562 | if (ret < 0) { |
| 1563 | err = ret; | 1563 | err = ret; |
| 1564 | goto out; | 1564 | goto out; |
| 1565 | } | 1565 | } |
| 1566 | leaf = path->nodes[0]; | 1566 | leaf = path->nodes[0]; |
| 1567 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 1567 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 1568 | } | 1568 | } |
| 1569 | #endif | 1569 | #endif |
| 1570 | BUG_ON(item_size < sizeof(*ei)); | 1570 | BUG_ON(item_size < sizeof(*ei)); |
| 1571 | 1571 | ||
| 1572 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 1572 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 1573 | flags = btrfs_extent_flags(leaf, ei); | 1573 | flags = btrfs_extent_flags(leaf, ei); |
| 1574 | 1574 | ||
| 1575 | ptr = (unsigned long)(ei + 1); | 1575 | ptr = (unsigned long)(ei + 1); |
| 1576 | end = (unsigned long)ei + item_size; | 1576 | end = (unsigned long)ei + item_size; |
| 1577 | 1577 | ||
| 1578 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK && !skinny_metadata) { | 1578 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK && !skinny_metadata) { |
| 1579 | ptr += sizeof(struct btrfs_tree_block_info); | 1579 | ptr += sizeof(struct btrfs_tree_block_info); |
| 1580 | BUG_ON(ptr > end); | 1580 | BUG_ON(ptr > end); |
| 1581 | } | 1581 | } |
| 1582 | 1582 | ||
| 1583 | err = -ENOENT; | 1583 | err = -ENOENT; |
| 1584 | while (1) { | 1584 | while (1) { |
| 1585 | if (ptr >= end) { | 1585 | if (ptr >= end) { |
| 1586 | WARN_ON(ptr > end); | 1586 | WARN_ON(ptr > end); |
| 1587 | break; | 1587 | break; |
| 1588 | } | 1588 | } |
| 1589 | iref = (struct btrfs_extent_inline_ref *)ptr; | 1589 | iref = (struct btrfs_extent_inline_ref *)ptr; |
| 1590 | type = btrfs_extent_inline_ref_type(leaf, iref); | 1590 | type = btrfs_extent_inline_ref_type(leaf, iref); |
| 1591 | if (want < type) | 1591 | if (want < type) |
| 1592 | break; | 1592 | break; |
| 1593 | if (want > type) { | 1593 | if (want > type) { |
| 1594 | ptr += btrfs_extent_inline_ref_size(type); | 1594 | ptr += btrfs_extent_inline_ref_size(type); |
| 1595 | continue; | 1595 | continue; |
| 1596 | } | 1596 | } |
| 1597 | 1597 | ||
| 1598 | if (type == BTRFS_EXTENT_DATA_REF_KEY) { | 1598 | if (type == BTRFS_EXTENT_DATA_REF_KEY) { |
| 1599 | struct btrfs_extent_data_ref *dref; | 1599 | struct btrfs_extent_data_ref *dref; |
| 1600 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); | 1600 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 1601 | if (match_extent_data_ref(leaf, dref, root_objectid, | 1601 | if (match_extent_data_ref(leaf, dref, root_objectid, |
| 1602 | owner, offset)) { | 1602 | owner, offset)) { |
| 1603 | err = 0; | 1603 | err = 0; |
| 1604 | break; | 1604 | break; |
| 1605 | } | 1605 | } |
| 1606 | if (hash_extent_data_ref_item(leaf, dref) < | 1606 | if (hash_extent_data_ref_item(leaf, dref) < |
| 1607 | hash_extent_data_ref(root_objectid, owner, offset)) | 1607 | hash_extent_data_ref(root_objectid, owner, offset)) |
| 1608 | break; | 1608 | break; |
| 1609 | } else { | 1609 | } else { |
| 1610 | u64 ref_offset; | 1610 | u64 ref_offset; |
| 1611 | ref_offset = btrfs_extent_inline_ref_offset(leaf, iref); | 1611 | ref_offset = btrfs_extent_inline_ref_offset(leaf, iref); |
| 1612 | if (parent > 0) { | 1612 | if (parent > 0) { |
| 1613 | if (parent == ref_offset) { | 1613 | if (parent == ref_offset) { |
| 1614 | err = 0; | 1614 | err = 0; |
| 1615 | break; | 1615 | break; |
| 1616 | } | 1616 | } |
| 1617 | if (ref_offset < parent) | 1617 | if (ref_offset < parent) |
| 1618 | break; | 1618 | break; |
| 1619 | } else { | 1619 | } else { |
| 1620 | if (root_objectid == ref_offset) { | 1620 | if (root_objectid == ref_offset) { |
| 1621 | err = 0; | 1621 | err = 0; |
| 1622 | break; | 1622 | break; |
| 1623 | } | 1623 | } |
| 1624 | if (ref_offset < root_objectid) | 1624 | if (ref_offset < root_objectid) |
| 1625 | break; | 1625 | break; |
| 1626 | } | 1626 | } |
| 1627 | } | 1627 | } |
| 1628 | ptr += btrfs_extent_inline_ref_size(type); | 1628 | ptr += btrfs_extent_inline_ref_size(type); |
| 1629 | } | 1629 | } |
| 1630 | if (err == -ENOENT && insert) { | 1630 | if (err == -ENOENT && insert) { |
| 1631 | if (item_size + extra_size >= | 1631 | if (item_size + extra_size >= |
| 1632 | BTRFS_MAX_EXTENT_ITEM_SIZE(root)) { | 1632 | BTRFS_MAX_EXTENT_ITEM_SIZE(root)) { |
| 1633 | err = -EAGAIN; | 1633 | err = -EAGAIN; |
| 1634 | goto out; | 1634 | goto out; |
| 1635 | } | 1635 | } |
| 1636 | /* | 1636 | /* |
| 1637 | * To add new inline back ref, we have to make sure | 1637 | * To add new inline back ref, we have to make sure |
| 1638 | * there is no corresponding back ref item. | 1638 | * there is no corresponding back ref item. |
| 1639 | * For simplicity, we just do not add new inline back | 1639 | * For simplicity, we just do not add new inline back |
| 1640 | * ref if there is any kind of item for this block | 1640 | * ref if there is any kind of item for this block |
| 1641 | */ | 1641 | */ |
| 1642 | if (find_next_key(path, 0, &key) == 0 && | 1642 | if (find_next_key(path, 0, &key) == 0 && |
| 1643 | key.objectid == bytenr && | 1643 | key.objectid == bytenr && |
| 1644 | key.type < BTRFS_BLOCK_GROUP_ITEM_KEY) { | 1644 | key.type < BTRFS_BLOCK_GROUP_ITEM_KEY) { |
| 1645 | err = -EAGAIN; | 1645 | err = -EAGAIN; |
| 1646 | goto out; | 1646 | goto out; |
| 1647 | } | 1647 | } |
| 1648 | } | 1648 | } |
| 1649 | *ref_ret = (struct btrfs_extent_inline_ref *)ptr; | 1649 | *ref_ret = (struct btrfs_extent_inline_ref *)ptr; |
| 1650 | out: | 1650 | out: |
| 1651 | if (insert) { | 1651 | if (insert) { |
| 1652 | path->keep_locks = 0; | 1652 | path->keep_locks = 0; |
| 1653 | btrfs_unlock_up_safe(path, 1); | 1653 | btrfs_unlock_up_safe(path, 1); |
| 1654 | } | 1654 | } |
| 1655 | return err; | 1655 | return err; |
| 1656 | } | 1656 | } |
| 1657 | 1657 | ||
| 1658 | /* | 1658 | /* |
| 1659 | * helper to add new inline back ref | 1659 | * helper to add new inline back ref |
| 1660 | */ | 1660 | */ |
| 1661 | static noinline_for_stack | 1661 | static noinline_for_stack |
| 1662 | void setup_inline_extent_backref(struct btrfs_root *root, | 1662 | void setup_inline_extent_backref(struct btrfs_root *root, |
| 1663 | struct btrfs_path *path, | 1663 | struct btrfs_path *path, |
| 1664 | struct btrfs_extent_inline_ref *iref, | 1664 | struct btrfs_extent_inline_ref *iref, |
| 1665 | u64 parent, u64 root_objectid, | 1665 | u64 parent, u64 root_objectid, |
| 1666 | u64 owner, u64 offset, int refs_to_add, | 1666 | u64 owner, u64 offset, int refs_to_add, |
| 1667 | struct btrfs_delayed_extent_op *extent_op) | 1667 | struct btrfs_delayed_extent_op *extent_op) |
| 1668 | { | 1668 | { |
| 1669 | struct extent_buffer *leaf; | 1669 | struct extent_buffer *leaf; |
| 1670 | struct btrfs_extent_item *ei; | 1670 | struct btrfs_extent_item *ei; |
| 1671 | unsigned long ptr; | 1671 | unsigned long ptr; |
| 1672 | unsigned long end; | 1672 | unsigned long end; |
| 1673 | unsigned long item_offset; | 1673 | unsigned long item_offset; |
| 1674 | u64 refs; | 1674 | u64 refs; |
| 1675 | int size; | 1675 | int size; |
| 1676 | int type; | 1676 | int type; |
| 1677 | 1677 | ||
| 1678 | leaf = path->nodes[0]; | 1678 | leaf = path->nodes[0]; |
| 1679 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 1679 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 1680 | item_offset = (unsigned long)iref - (unsigned long)ei; | 1680 | item_offset = (unsigned long)iref - (unsigned long)ei; |
| 1681 | 1681 | ||
| 1682 | type = extent_ref_type(parent, owner); | 1682 | type = extent_ref_type(parent, owner); |
| 1683 | size = btrfs_extent_inline_ref_size(type); | 1683 | size = btrfs_extent_inline_ref_size(type); |
| 1684 | 1684 | ||
| 1685 | btrfs_extend_item(root, path, size); | 1685 | btrfs_extend_item(root, path, size); |
| 1686 | 1686 | ||
| 1687 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 1687 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 1688 | refs = btrfs_extent_refs(leaf, ei); | 1688 | refs = btrfs_extent_refs(leaf, ei); |
| 1689 | refs += refs_to_add; | 1689 | refs += refs_to_add; |
| 1690 | btrfs_set_extent_refs(leaf, ei, refs); | 1690 | btrfs_set_extent_refs(leaf, ei, refs); |
| 1691 | if (extent_op) | 1691 | if (extent_op) |
| 1692 | __run_delayed_extent_op(extent_op, leaf, ei); | 1692 | __run_delayed_extent_op(extent_op, leaf, ei); |
| 1693 | 1693 | ||
| 1694 | ptr = (unsigned long)ei + item_offset; | 1694 | ptr = (unsigned long)ei + item_offset; |
| 1695 | end = (unsigned long)ei + btrfs_item_size_nr(leaf, path->slots[0]); | 1695 | end = (unsigned long)ei + btrfs_item_size_nr(leaf, path->slots[0]); |
| 1696 | if (ptr < end - size) | 1696 | if (ptr < end - size) |
| 1697 | memmove_extent_buffer(leaf, ptr + size, ptr, | 1697 | memmove_extent_buffer(leaf, ptr + size, ptr, |
| 1698 | end - size - ptr); | 1698 | end - size - ptr); |
| 1699 | 1699 | ||
| 1700 | iref = (struct btrfs_extent_inline_ref *)ptr; | 1700 | iref = (struct btrfs_extent_inline_ref *)ptr; |
| 1701 | btrfs_set_extent_inline_ref_type(leaf, iref, type); | 1701 | btrfs_set_extent_inline_ref_type(leaf, iref, type); |
| 1702 | if (type == BTRFS_EXTENT_DATA_REF_KEY) { | 1702 | if (type == BTRFS_EXTENT_DATA_REF_KEY) { |
| 1703 | struct btrfs_extent_data_ref *dref; | 1703 | struct btrfs_extent_data_ref *dref; |
| 1704 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); | 1704 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 1705 | btrfs_set_extent_data_ref_root(leaf, dref, root_objectid); | 1705 | btrfs_set_extent_data_ref_root(leaf, dref, root_objectid); |
| 1706 | btrfs_set_extent_data_ref_objectid(leaf, dref, owner); | 1706 | btrfs_set_extent_data_ref_objectid(leaf, dref, owner); |
| 1707 | btrfs_set_extent_data_ref_offset(leaf, dref, offset); | 1707 | btrfs_set_extent_data_ref_offset(leaf, dref, offset); |
| 1708 | btrfs_set_extent_data_ref_count(leaf, dref, refs_to_add); | 1708 | btrfs_set_extent_data_ref_count(leaf, dref, refs_to_add); |
| 1709 | } else if (type == BTRFS_SHARED_DATA_REF_KEY) { | 1709 | } else if (type == BTRFS_SHARED_DATA_REF_KEY) { |
| 1710 | struct btrfs_shared_data_ref *sref; | 1710 | struct btrfs_shared_data_ref *sref; |
| 1711 | sref = (struct btrfs_shared_data_ref *)(iref + 1); | 1711 | sref = (struct btrfs_shared_data_ref *)(iref + 1); |
| 1712 | btrfs_set_shared_data_ref_count(leaf, sref, refs_to_add); | 1712 | btrfs_set_shared_data_ref_count(leaf, sref, refs_to_add); |
| 1713 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); | 1713 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); |
| 1714 | } else if (type == BTRFS_SHARED_BLOCK_REF_KEY) { | 1714 | } else if (type == BTRFS_SHARED_BLOCK_REF_KEY) { |
| 1715 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); | 1715 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); |
| 1716 | } else { | 1716 | } else { |
| 1717 | btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid); | 1717 | btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid); |
| 1718 | } | 1718 | } |
| 1719 | btrfs_mark_buffer_dirty(leaf); | 1719 | btrfs_mark_buffer_dirty(leaf); |
| 1720 | } | 1720 | } |
| 1721 | 1721 | ||
| 1722 | static int lookup_extent_backref(struct btrfs_trans_handle *trans, | 1722 | static int lookup_extent_backref(struct btrfs_trans_handle *trans, |
| 1723 | struct btrfs_root *root, | 1723 | struct btrfs_root *root, |
| 1724 | struct btrfs_path *path, | 1724 | struct btrfs_path *path, |
| 1725 | struct btrfs_extent_inline_ref **ref_ret, | 1725 | struct btrfs_extent_inline_ref **ref_ret, |
| 1726 | u64 bytenr, u64 num_bytes, u64 parent, | 1726 | u64 bytenr, u64 num_bytes, u64 parent, |
| 1727 | u64 root_objectid, u64 owner, u64 offset) | 1727 | u64 root_objectid, u64 owner, u64 offset) |
| 1728 | { | 1728 | { |
| 1729 | int ret; | 1729 | int ret; |
| 1730 | 1730 | ||
| 1731 | ret = lookup_inline_extent_backref(trans, root, path, ref_ret, | 1731 | ret = lookup_inline_extent_backref(trans, root, path, ref_ret, |
| 1732 | bytenr, num_bytes, parent, | 1732 | bytenr, num_bytes, parent, |
| 1733 | root_objectid, owner, offset, 0); | 1733 | root_objectid, owner, offset, 0); |
| 1734 | if (ret != -ENOENT) | 1734 | if (ret != -ENOENT) |
| 1735 | return ret; | 1735 | return ret; |
| 1736 | 1736 | ||
| 1737 | btrfs_release_path(path); | 1737 | btrfs_release_path(path); |
| 1738 | *ref_ret = NULL; | 1738 | *ref_ret = NULL; |
| 1739 | 1739 | ||
| 1740 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 1740 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 1741 | ret = lookup_tree_block_ref(trans, root, path, bytenr, parent, | 1741 | ret = lookup_tree_block_ref(trans, root, path, bytenr, parent, |
| 1742 | root_objectid); | 1742 | root_objectid); |
| 1743 | } else { | 1743 | } else { |
| 1744 | ret = lookup_extent_data_ref(trans, root, path, bytenr, parent, | 1744 | ret = lookup_extent_data_ref(trans, root, path, bytenr, parent, |
| 1745 | root_objectid, owner, offset); | 1745 | root_objectid, owner, offset); |
| 1746 | } | 1746 | } |
| 1747 | return ret; | 1747 | return ret; |
| 1748 | } | 1748 | } |
| 1749 | 1749 | ||
| 1750 | /* | 1750 | /* |
| 1751 | * helper to update/remove inline back ref | 1751 | * helper to update/remove inline back ref |
| 1752 | */ | 1752 | */ |
| 1753 | static noinline_for_stack | 1753 | static noinline_for_stack |
| 1754 | void update_inline_extent_backref(struct btrfs_root *root, | 1754 | void update_inline_extent_backref(struct btrfs_root *root, |
| 1755 | struct btrfs_path *path, | 1755 | struct btrfs_path *path, |
| 1756 | struct btrfs_extent_inline_ref *iref, | 1756 | struct btrfs_extent_inline_ref *iref, |
| 1757 | int refs_to_mod, | 1757 | int refs_to_mod, |
| 1758 | struct btrfs_delayed_extent_op *extent_op, | 1758 | struct btrfs_delayed_extent_op *extent_op, |
| 1759 | int *last_ref) | 1759 | int *last_ref) |
| 1760 | { | 1760 | { |
| 1761 | struct extent_buffer *leaf; | 1761 | struct extent_buffer *leaf; |
| 1762 | struct btrfs_extent_item *ei; | 1762 | struct btrfs_extent_item *ei; |
| 1763 | struct btrfs_extent_data_ref *dref = NULL; | 1763 | struct btrfs_extent_data_ref *dref = NULL; |
| 1764 | struct btrfs_shared_data_ref *sref = NULL; | 1764 | struct btrfs_shared_data_ref *sref = NULL; |
| 1765 | unsigned long ptr; | 1765 | unsigned long ptr; |
| 1766 | unsigned long end; | 1766 | unsigned long end; |
| 1767 | u32 item_size; | 1767 | u32 item_size; |
| 1768 | int size; | 1768 | int size; |
| 1769 | int type; | 1769 | int type; |
| 1770 | u64 refs; | 1770 | u64 refs; |
| 1771 | 1771 | ||
| 1772 | leaf = path->nodes[0]; | 1772 | leaf = path->nodes[0]; |
| 1773 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 1773 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 1774 | refs = btrfs_extent_refs(leaf, ei); | 1774 | refs = btrfs_extent_refs(leaf, ei); |
| 1775 | WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0); | 1775 | WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0); |
| 1776 | refs += refs_to_mod; | 1776 | refs += refs_to_mod; |
| 1777 | btrfs_set_extent_refs(leaf, ei, refs); | 1777 | btrfs_set_extent_refs(leaf, ei, refs); |
| 1778 | if (extent_op) | 1778 | if (extent_op) |
| 1779 | __run_delayed_extent_op(extent_op, leaf, ei); | 1779 | __run_delayed_extent_op(extent_op, leaf, ei); |
| 1780 | 1780 | ||
| 1781 | type = btrfs_extent_inline_ref_type(leaf, iref); | 1781 | type = btrfs_extent_inline_ref_type(leaf, iref); |
| 1782 | 1782 | ||
| 1783 | if (type == BTRFS_EXTENT_DATA_REF_KEY) { | 1783 | if (type == BTRFS_EXTENT_DATA_REF_KEY) { |
| 1784 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); | 1784 | dref = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 1785 | refs = btrfs_extent_data_ref_count(leaf, dref); | 1785 | refs = btrfs_extent_data_ref_count(leaf, dref); |
| 1786 | } else if (type == BTRFS_SHARED_DATA_REF_KEY) { | 1786 | } else if (type == BTRFS_SHARED_DATA_REF_KEY) { |
| 1787 | sref = (struct btrfs_shared_data_ref *)(iref + 1); | 1787 | sref = (struct btrfs_shared_data_ref *)(iref + 1); |
| 1788 | refs = btrfs_shared_data_ref_count(leaf, sref); | 1788 | refs = btrfs_shared_data_ref_count(leaf, sref); |
| 1789 | } else { | 1789 | } else { |
| 1790 | refs = 1; | 1790 | refs = 1; |
| 1791 | BUG_ON(refs_to_mod != -1); | 1791 | BUG_ON(refs_to_mod != -1); |
| 1792 | } | 1792 | } |
| 1793 | 1793 | ||
| 1794 | BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod); | 1794 | BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod); |
| 1795 | refs += refs_to_mod; | 1795 | refs += refs_to_mod; |
| 1796 | 1796 | ||
| 1797 | if (refs > 0) { | 1797 | if (refs > 0) { |
| 1798 | if (type == BTRFS_EXTENT_DATA_REF_KEY) | 1798 | if (type == BTRFS_EXTENT_DATA_REF_KEY) |
| 1799 | btrfs_set_extent_data_ref_count(leaf, dref, refs); | 1799 | btrfs_set_extent_data_ref_count(leaf, dref, refs); |
| 1800 | else | 1800 | else |
| 1801 | btrfs_set_shared_data_ref_count(leaf, sref, refs); | 1801 | btrfs_set_shared_data_ref_count(leaf, sref, refs); |
| 1802 | } else { | 1802 | } else { |
| 1803 | *last_ref = 1; | 1803 | *last_ref = 1; |
| 1804 | size = btrfs_extent_inline_ref_size(type); | 1804 | size = btrfs_extent_inline_ref_size(type); |
| 1805 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 1805 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 1806 | ptr = (unsigned long)iref; | 1806 | ptr = (unsigned long)iref; |
| 1807 | end = (unsigned long)ei + item_size; | 1807 | end = (unsigned long)ei + item_size; |
| 1808 | if (ptr + size < end) | 1808 | if (ptr + size < end) |
| 1809 | memmove_extent_buffer(leaf, ptr, ptr + size, | 1809 | memmove_extent_buffer(leaf, ptr, ptr + size, |
| 1810 | end - ptr - size); | 1810 | end - ptr - size); |
| 1811 | item_size -= size; | 1811 | item_size -= size; |
| 1812 | btrfs_truncate_item(root, path, item_size, 1); | 1812 | btrfs_truncate_item(root, path, item_size, 1); |
| 1813 | } | 1813 | } |
| 1814 | btrfs_mark_buffer_dirty(leaf); | 1814 | btrfs_mark_buffer_dirty(leaf); |
| 1815 | } | 1815 | } |
| 1816 | 1816 | ||
| 1817 | static noinline_for_stack | 1817 | static noinline_for_stack |
| 1818 | int insert_inline_extent_backref(struct btrfs_trans_handle *trans, | 1818 | int insert_inline_extent_backref(struct btrfs_trans_handle *trans, |
| 1819 | struct btrfs_root *root, | 1819 | struct btrfs_root *root, |
| 1820 | struct btrfs_path *path, | 1820 | struct btrfs_path *path, |
| 1821 | u64 bytenr, u64 num_bytes, u64 parent, | 1821 | u64 bytenr, u64 num_bytes, u64 parent, |
| 1822 | u64 root_objectid, u64 owner, | 1822 | u64 root_objectid, u64 owner, |
| 1823 | u64 offset, int refs_to_add, | 1823 | u64 offset, int refs_to_add, |
| 1824 | struct btrfs_delayed_extent_op *extent_op) | 1824 | struct btrfs_delayed_extent_op *extent_op) |
| 1825 | { | 1825 | { |
| 1826 | struct btrfs_extent_inline_ref *iref; | 1826 | struct btrfs_extent_inline_ref *iref; |
| 1827 | int ret; | 1827 | int ret; |
| 1828 | 1828 | ||
| 1829 | ret = lookup_inline_extent_backref(trans, root, path, &iref, | 1829 | ret = lookup_inline_extent_backref(trans, root, path, &iref, |
| 1830 | bytenr, num_bytes, parent, | 1830 | bytenr, num_bytes, parent, |
| 1831 | root_objectid, owner, offset, 1); | 1831 | root_objectid, owner, offset, 1); |
| 1832 | if (ret == 0) { | 1832 | if (ret == 0) { |
| 1833 | BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID); | 1833 | BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID); |
| 1834 | update_inline_extent_backref(root, path, iref, | 1834 | update_inline_extent_backref(root, path, iref, |
| 1835 | refs_to_add, extent_op, NULL); | 1835 | refs_to_add, extent_op, NULL); |
| 1836 | } else if (ret == -ENOENT) { | 1836 | } else if (ret == -ENOENT) { |
| 1837 | setup_inline_extent_backref(root, path, iref, parent, | 1837 | setup_inline_extent_backref(root, path, iref, parent, |
| 1838 | root_objectid, owner, offset, | 1838 | root_objectid, owner, offset, |
| 1839 | refs_to_add, extent_op); | 1839 | refs_to_add, extent_op); |
| 1840 | ret = 0; | 1840 | ret = 0; |
| 1841 | } | 1841 | } |
| 1842 | return ret; | 1842 | return ret; |
| 1843 | } | 1843 | } |
| 1844 | 1844 | ||
| 1845 | static int insert_extent_backref(struct btrfs_trans_handle *trans, | 1845 | static int insert_extent_backref(struct btrfs_trans_handle *trans, |
| 1846 | struct btrfs_root *root, | 1846 | struct btrfs_root *root, |
| 1847 | struct btrfs_path *path, | 1847 | struct btrfs_path *path, |
| 1848 | u64 bytenr, u64 parent, u64 root_objectid, | 1848 | u64 bytenr, u64 parent, u64 root_objectid, |
| 1849 | u64 owner, u64 offset, int refs_to_add) | 1849 | u64 owner, u64 offset, int refs_to_add) |
| 1850 | { | 1850 | { |
| 1851 | int ret; | 1851 | int ret; |
| 1852 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 1852 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 1853 | BUG_ON(refs_to_add != 1); | 1853 | BUG_ON(refs_to_add != 1); |
| 1854 | ret = insert_tree_block_ref(trans, root, path, bytenr, | 1854 | ret = insert_tree_block_ref(trans, root, path, bytenr, |
| 1855 | parent, root_objectid); | 1855 | parent, root_objectid); |
| 1856 | } else { | 1856 | } else { |
| 1857 | ret = insert_extent_data_ref(trans, root, path, bytenr, | 1857 | ret = insert_extent_data_ref(trans, root, path, bytenr, |
| 1858 | parent, root_objectid, | 1858 | parent, root_objectid, |
| 1859 | owner, offset, refs_to_add); | 1859 | owner, offset, refs_to_add); |
| 1860 | } | 1860 | } |
| 1861 | return ret; | 1861 | return ret; |
| 1862 | } | 1862 | } |
| 1863 | 1863 | ||
| 1864 | static int remove_extent_backref(struct btrfs_trans_handle *trans, | 1864 | static int remove_extent_backref(struct btrfs_trans_handle *trans, |
| 1865 | struct btrfs_root *root, | 1865 | struct btrfs_root *root, |
| 1866 | struct btrfs_path *path, | 1866 | struct btrfs_path *path, |
| 1867 | struct btrfs_extent_inline_ref *iref, | 1867 | struct btrfs_extent_inline_ref *iref, |
| 1868 | int refs_to_drop, int is_data, int *last_ref) | 1868 | int refs_to_drop, int is_data, int *last_ref) |
| 1869 | { | 1869 | { |
| 1870 | int ret = 0; | 1870 | int ret = 0; |
| 1871 | 1871 | ||
| 1872 | BUG_ON(!is_data && refs_to_drop != 1); | 1872 | BUG_ON(!is_data && refs_to_drop != 1); |
| 1873 | if (iref) { | 1873 | if (iref) { |
| 1874 | update_inline_extent_backref(root, path, iref, | 1874 | update_inline_extent_backref(root, path, iref, |
| 1875 | -refs_to_drop, NULL, last_ref); | 1875 | -refs_to_drop, NULL, last_ref); |
| 1876 | } else if (is_data) { | 1876 | } else if (is_data) { |
| 1877 | ret = remove_extent_data_ref(trans, root, path, refs_to_drop, | 1877 | ret = remove_extent_data_ref(trans, root, path, refs_to_drop, |
| 1878 | last_ref); | 1878 | last_ref); |
| 1879 | } else { | 1879 | } else { |
| 1880 | *last_ref = 1; | 1880 | *last_ref = 1; |
| 1881 | ret = btrfs_del_item(trans, root, path); | 1881 | ret = btrfs_del_item(trans, root, path); |
| 1882 | } | 1882 | } |
| 1883 | return ret; | 1883 | return ret; |
| 1884 | } | 1884 | } |
| 1885 | 1885 | ||
| 1886 | static int btrfs_issue_discard(struct block_device *bdev, | 1886 | static int btrfs_issue_discard(struct block_device *bdev, |
| 1887 | u64 start, u64 len) | 1887 | u64 start, u64 len) |
| 1888 | { | 1888 | { |
| 1889 | return blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0); | 1889 | return blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_NOFS, 0); |
| 1890 | } | 1890 | } |
| 1891 | 1891 | ||
| 1892 | int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, | 1892 | int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, |
| 1893 | u64 num_bytes, u64 *actual_bytes) | 1893 | u64 num_bytes, u64 *actual_bytes) |
| 1894 | { | 1894 | { |
| 1895 | int ret; | 1895 | int ret; |
| 1896 | u64 discarded_bytes = 0; | 1896 | u64 discarded_bytes = 0; |
| 1897 | struct btrfs_bio *bbio = NULL; | 1897 | struct btrfs_bio *bbio = NULL; |
| 1898 | 1898 | ||
| 1899 | 1899 | ||
| 1900 | /* Tell the block device(s) that the sectors can be discarded */ | 1900 | /* Tell the block device(s) that the sectors can be discarded */ |
| 1901 | ret = btrfs_map_block(root->fs_info, REQ_DISCARD, | 1901 | ret = btrfs_map_block(root->fs_info, REQ_DISCARD, |
| 1902 | bytenr, &num_bytes, &bbio, 0); | 1902 | bytenr, &num_bytes, &bbio, 0); |
| 1903 | /* Error condition is -ENOMEM */ | 1903 | /* Error condition is -ENOMEM */ |
| 1904 | if (!ret) { | 1904 | if (!ret) { |
| 1905 | struct btrfs_bio_stripe *stripe = bbio->stripes; | 1905 | struct btrfs_bio_stripe *stripe = bbio->stripes; |
| 1906 | int i; | 1906 | int i; |
| 1907 | 1907 | ||
| 1908 | 1908 | ||
| 1909 | for (i = 0; i < bbio->num_stripes; i++, stripe++) { | 1909 | for (i = 0; i < bbio->num_stripes; i++, stripe++) { |
| 1910 | if (!stripe->dev->can_discard) | 1910 | if (!stripe->dev->can_discard) |
| 1911 | continue; | 1911 | continue; |
| 1912 | 1912 | ||
| 1913 | ret = btrfs_issue_discard(stripe->dev->bdev, | 1913 | ret = btrfs_issue_discard(stripe->dev->bdev, |
| 1914 | stripe->physical, | 1914 | stripe->physical, |
| 1915 | stripe->length); | 1915 | stripe->length); |
| 1916 | if (!ret) | 1916 | if (!ret) |
| 1917 | discarded_bytes += stripe->length; | 1917 | discarded_bytes += stripe->length; |
| 1918 | else if (ret != -EOPNOTSUPP) | 1918 | else if (ret != -EOPNOTSUPP) |
| 1919 | break; /* Logic errors or -ENOMEM, or -EIO but I don't know how that could happen JDM */ | 1919 | break; /* Logic errors or -ENOMEM, or -EIO but I don't know how that could happen JDM */ |
| 1920 | 1920 | ||
| 1921 | /* | 1921 | /* |
| 1922 | * Just in case we get back EOPNOTSUPP for some reason, | 1922 | * Just in case we get back EOPNOTSUPP for some reason, |
| 1923 | * just ignore the return value so we don't screw up | 1923 | * just ignore the return value so we don't screw up |
| 1924 | * people calling discard_extent. | 1924 | * people calling discard_extent. |
| 1925 | */ | 1925 | */ |
| 1926 | ret = 0; | 1926 | ret = 0; |
| 1927 | } | 1927 | } |
| 1928 | kfree(bbio); | 1928 | kfree(bbio); |
| 1929 | } | 1929 | } |
| 1930 | 1930 | ||
| 1931 | if (actual_bytes) | 1931 | if (actual_bytes) |
| 1932 | *actual_bytes = discarded_bytes; | 1932 | *actual_bytes = discarded_bytes; |
| 1933 | 1933 | ||
| 1934 | 1934 | ||
| 1935 | if (ret == -EOPNOTSUPP) | 1935 | if (ret == -EOPNOTSUPP) |
| 1936 | ret = 0; | 1936 | ret = 0; |
| 1937 | return ret; | 1937 | return ret; |
| 1938 | } | 1938 | } |
| 1939 | 1939 | ||
| 1940 | /* Can return -ENOMEM */ | 1940 | /* Can return -ENOMEM */ |
| 1941 | int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, | 1941 | int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, |
| 1942 | struct btrfs_root *root, | 1942 | struct btrfs_root *root, |
| 1943 | u64 bytenr, u64 num_bytes, u64 parent, | 1943 | u64 bytenr, u64 num_bytes, u64 parent, |
| 1944 | u64 root_objectid, u64 owner, u64 offset, | 1944 | u64 root_objectid, u64 owner, u64 offset, |
| 1945 | int no_quota) | 1945 | int no_quota) |
| 1946 | { | 1946 | { |
| 1947 | int ret; | 1947 | int ret; |
| 1948 | struct btrfs_fs_info *fs_info = root->fs_info; | 1948 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 1949 | 1949 | ||
| 1950 | BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID && | 1950 | BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID && |
| 1951 | root_objectid == BTRFS_TREE_LOG_OBJECTID); | 1951 | root_objectid == BTRFS_TREE_LOG_OBJECTID); |
| 1952 | 1952 | ||
| 1953 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 1953 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 1954 | ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, | 1954 | ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, |
| 1955 | num_bytes, | 1955 | num_bytes, |
| 1956 | parent, root_objectid, (int)owner, | 1956 | parent, root_objectid, (int)owner, |
| 1957 | BTRFS_ADD_DELAYED_REF, NULL, no_quota); | 1957 | BTRFS_ADD_DELAYED_REF, NULL, no_quota); |
| 1958 | } else { | 1958 | } else { |
| 1959 | ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, | 1959 | ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, |
| 1960 | num_bytes, | 1960 | num_bytes, |
| 1961 | parent, root_objectid, owner, offset, | 1961 | parent, root_objectid, owner, offset, |
| 1962 | BTRFS_ADD_DELAYED_REF, NULL, no_quota); | 1962 | BTRFS_ADD_DELAYED_REF, NULL, no_quota); |
| 1963 | } | 1963 | } |
| 1964 | return ret; | 1964 | return ret; |
| 1965 | } | 1965 | } |
| 1966 | 1966 | ||
| 1967 | static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, | 1967 | static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, |
| 1968 | struct btrfs_root *root, | 1968 | struct btrfs_root *root, |
| 1969 | u64 bytenr, u64 num_bytes, | 1969 | u64 bytenr, u64 num_bytes, |
| 1970 | u64 parent, u64 root_objectid, | 1970 | u64 parent, u64 root_objectid, |
| 1971 | u64 owner, u64 offset, int refs_to_add, | 1971 | u64 owner, u64 offset, int refs_to_add, |
| 1972 | int no_quota, | 1972 | int no_quota, |
| 1973 | struct btrfs_delayed_extent_op *extent_op) | 1973 | struct btrfs_delayed_extent_op *extent_op) |
| 1974 | { | 1974 | { |
| 1975 | struct btrfs_fs_info *fs_info = root->fs_info; | 1975 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 1976 | struct btrfs_path *path; | 1976 | struct btrfs_path *path; |
| 1977 | struct extent_buffer *leaf; | 1977 | struct extent_buffer *leaf; |
| 1978 | struct btrfs_extent_item *item; | 1978 | struct btrfs_extent_item *item; |
| 1979 | struct btrfs_key key; | 1979 | struct btrfs_key key; |
| 1980 | u64 refs; | 1980 | u64 refs; |
| 1981 | int ret; | 1981 | int ret; |
| 1982 | enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_ADD_EXCL; | 1982 | enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_ADD_EXCL; |
| 1983 | 1983 | ||
| 1984 | path = btrfs_alloc_path(); | 1984 | path = btrfs_alloc_path(); |
| 1985 | if (!path) | 1985 | if (!path) |
| 1986 | return -ENOMEM; | 1986 | return -ENOMEM; |
| 1987 | 1987 | ||
| 1988 | if (!is_fstree(root_objectid) || !root->fs_info->quota_enabled) | 1988 | if (!is_fstree(root_objectid) || !root->fs_info->quota_enabled) |
| 1989 | no_quota = 1; | 1989 | no_quota = 1; |
| 1990 | 1990 | ||
| 1991 | path->reada = 1; | 1991 | path->reada = 1; |
| 1992 | path->leave_spinning = 1; | 1992 | path->leave_spinning = 1; |
| 1993 | /* this will setup the path even if it fails to insert the back ref */ | 1993 | /* this will setup the path even if it fails to insert the back ref */ |
| 1994 | ret = insert_inline_extent_backref(trans, fs_info->extent_root, path, | 1994 | ret = insert_inline_extent_backref(trans, fs_info->extent_root, path, |
| 1995 | bytenr, num_bytes, parent, | 1995 | bytenr, num_bytes, parent, |
| 1996 | root_objectid, owner, offset, | 1996 | root_objectid, owner, offset, |
| 1997 | refs_to_add, extent_op); | 1997 | refs_to_add, extent_op); |
| 1998 | if ((ret < 0 && ret != -EAGAIN) || (!ret && no_quota)) | 1998 | if ((ret < 0 && ret != -EAGAIN) || (!ret && no_quota)) |
| 1999 | goto out; | 1999 | goto out; |
| 2000 | /* | 2000 | /* |
| 2001 | * Ok we were able to insert an inline extent and it appears to be a new | 2001 | * Ok we were able to insert an inline extent and it appears to be a new |
| 2002 | * reference, deal with the qgroup accounting. | 2002 | * reference, deal with the qgroup accounting. |
| 2003 | */ | 2003 | */ |
| 2004 | if (!ret && !no_quota) { | 2004 | if (!ret && !no_quota) { |
| 2005 | ASSERT(root->fs_info->quota_enabled); | 2005 | ASSERT(root->fs_info->quota_enabled); |
| 2006 | leaf = path->nodes[0]; | 2006 | leaf = path->nodes[0]; |
| 2007 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 2007 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 2008 | item = btrfs_item_ptr(leaf, path->slots[0], | 2008 | item = btrfs_item_ptr(leaf, path->slots[0], |
| 2009 | struct btrfs_extent_item); | 2009 | struct btrfs_extent_item); |
| 2010 | if (btrfs_extent_refs(leaf, item) > (u64)refs_to_add) | 2010 | if (btrfs_extent_refs(leaf, item) > (u64)refs_to_add) |
| 2011 | type = BTRFS_QGROUP_OPER_ADD_SHARED; | 2011 | type = BTRFS_QGROUP_OPER_ADD_SHARED; |
| 2012 | btrfs_release_path(path); | 2012 | btrfs_release_path(path); |
| 2013 | 2013 | ||
| 2014 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, | 2014 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, |
| 2015 | bytenr, num_bytes, type, 0); | 2015 | bytenr, num_bytes, type, 0); |
| 2016 | goto out; | 2016 | goto out; |
| 2017 | } | 2017 | } |
| 2018 | 2018 | ||
| 2019 | /* | 2019 | /* |
| 2020 | * Ok we had -EAGAIN which means we didn't have space to insert and | 2020 | * Ok we had -EAGAIN which means we didn't have space to insert and |
| 2021 | * inline extent ref, so just update the reference count and add a | 2021 | * inline extent ref, so just update the reference count and add a |
| 2022 | * normal backref. | 2022 | * normal backref. |
| 2023 | */ | 2023 | */ |
| 2024 | leaf = path->nodes[0]; | 2024 | leaf = path->nodes[0]; |
| 2025 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 2025 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 2026 | item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 2026 | item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 2027 | refs = btrfs_extent_refs(leaf, item); | 2027 | refs = btrfs_extent_refs(leaf, item); |
| 2028 | if (refs) | 2028 | if (refs) |
| 2029 | type = BTRFS_QGROUP_OPER_ADD_SHARED; | 2029 | type = BTRFS_QGROUP_OPER_ADD_SHARED; |
| 2030 | btrfs_set_extent_refs(leaf, item, refs + refs_to_add); | 2030 | btrfs_set_extent_refs(leaf, item, refs + refs_to_add); |
| 2031 | if (extent_op) | 2031 | if (extent_op) |
| 2032 | __run_delayed_extent_op(extent_op, leaf, item); | 2032 | __run_delayed_extent_op(extent_op, leaf, item); |
| 2033 | 2033 | ||
| 2034 | btrfs_mark_buffer_dirty(leaf); | 2034 | btrfs_mark_buffer_dirty(leaf); |
| 2035 | btrfs_release_path(path); | 2035 | btrfs_release_path(path); |
| 2036 | 2036 | ||
| 2037 | if (!no_quota) { | 2037 | if (!no_quota) { |
| 2038 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, | 2038 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, |
| 2039 | bytenr, num_bytes, type, 0); | 2039 | bytenr, num_bytes, type, 0); |
| 2040 | if (ret) | 2040 | if (ret) |
| 2041 | goto out; | 2041 | goto out; |
| 2042 | } | 2042 | } |
| 2043 | 2043 | ||
| 2044 | path->reada = 1; | 2044 | path->reada = 1; |
| 2045 | path->leave_spinning = 1; | 2045 | path->leave_spinning = 1; |
| 2046 | /* now insert the actual backref */ | 2046 | /* now insert the actual backref */ |
| 2047 | ret = insert_extent_backref(trans, root->fs_info->extent_root, | 2047 | ret = insert_extent_backref(trans, root->fs_info->extent_root, |
| 2048 | path, bytenr, parent, root_objectid, | 2048 | path, bytenr, parent, root_objectid, |
| 2049 | owner, offset, refs_to_add); | 2049 | owner, offset, refs_to_add); |
| 2050 | if (ret) | 2050 | if (ret) |
| 2051 | btrfs_abort_transaction(trans, root, ret); | 2051 | btrfs_abort_transaction(trans, root, ret); |
| 2052 | out: | 2052 | out: |
| 2053 | btrfs_free_path(path); | 2053 | btrfs_free_path(path); |
| 2054 | return ret; | 2054 | return ret; |
| 2055 | } | 2055 | } |
| 2056 | 2056 | ||
| 2057 | static int run_delayed_data_ref(struct btrfs_trans_handle *trans, | 2057 | static int run_delayed_data_ref(struct btrfs_trans_handle *trans, |
| 2058 | struct btrfs_root *root, | 2058 | struct btrfs_root *root, |
| 2059 | struct btrfs_delayed_ref_node *node, | 2059 | struct btrfs_delayed_ref_node *node, |
| 2060 | struct btrfs_delayed_extent_op *extent_op, | 2060 | struct btrfs_delayed_extent_op *extent_op, |
| 2061 | int insert_reserved) | 2061 | int insert_reserved) |
| 2062 | { | 2062 | { |
| 2063 | int ret = 0; | 2063 | int ret = 0; |
| 2064 | struct btrfs_delayed_data_ref *ref; | 2064 | struct btrfs_delayed_data_ref *ref; |
| 2065 | struct btrfs_key ins; | 2065 | struct btrfs_key ins; |
| 2066 | u64 parent = 0; | 2066 | u64 parent = 0; |
| 2067 | u64 ref_root = 0; | 2067 | u64 ref_root = 0; |
| 2068 | u64 flags = 0; | 2068 | u64 flags = 0; |
| 2069 | 2069 | ||
| 2070 | ins.objectid = node->bytenr; | 2070 | ins.objectid = node->bytenr; |
| 2071 | ins.offset = node->num_bytes; | 2071 | ins.offset = node->num_bytes; |
| 2072 | ins.type = BTRFS_EXTENT_ITEM_KEY; | 2072 | ins.type = BTRFS_EXTENT_ITEM_KEY; |
| 2073 | 2073 | ||
| 2074 | ref = btrfs_delayed_node_to_data_ref(node); | 2074 | ref = btrfs_delayed_node_to_data_ref(node); |
| 2075 | trace_run_delayed_data_ref(node, ref, node->action); | 2075 | trace_run_delayed_data_ref(node, ref, node->action); |
| 2076 | 2076 | ||
| 2077 | if (node->type == BTRFS_SHARED_DATA_REF_KEY) | 2077 | if (node->type == BTRFS_SHARED_DATA_REF_KEY) |
| 2078 | parent = ref->parent; | 2078 | parent = ref->parent; |
| 2079 | ref_root = ref->root; | 2079 | ref_root = ref->root; |
| 2080 | 2080 | ||
| 2081 | if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) { | 2081 | if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) { |
| 2082 | if (extent_op) | 2082 | if (extent_op) |
| 2083 | flags |= extent_op->flags_to_set; | 2083 | flags |= extent_op->flags_to_set; |
| 2084 | ret = alloc_reserved_file_extent(trans, root, | 2084 | ret = alloc_reserved_file_extent(trans, root, |
| 2085 | parent, ref_root, flags, | 2085 | parent, ref_root, flags, |
| 2086 | ref->objectid, ref->offset, | 2086 | ref->objectid, ref->offset, |
| 2087 | &ins, node->ref_mod); | 2087 | &ins, node->ref_mod); |
| 2088 | } else if (node->action == BTRFS_ADD_DELAYED_REF) { | 2088 | } else if (node->action == BTRFS_ADD_DELAYED_REF) { |
| 2089 | ret = __btrfs_inc_extent_ref(trans, root, node->bytenr, | 2089 | ret = __btrfs_inc_extent_ref(trans, root, node->bytenr, |
| 2090 | node->num_bytes, parent, | 2090 | node->num_bytes, parent, |
| 2091 | ref_root, ref->objectid, | 2091 | ref_root, ref->objectid, |
| 2092 | ref->offset, node->ref_mod, | 2092 | ref->offset, node->ref_mod, |
| 2093 | node->no_quota, extent_op); | 2093 | node->no_quota, extent_op); |
| 2094 | } else if (node->action == BTRFS_DROP_DELAYED_REF) { | 2094 | } else if (node->action == BTRFS_DROP_DELAYED_REF) { |
| 2095 | ret = __btrfs_free_extent(trans, root, node->bytenr, | 2095 | ret = __btrfs_free_extent(trans, root, node->bytenr, |
| 2096 | node->num_bytes, parent, | 2096 | node->num_bytes, parent, |
| 2097 | ref_root, ref->objectid, | 2097 | ref_root, ref->objectid, |
| 2098 | ref->offset, node->ref_mod, | 2098 | ref->offset, node->ref_mod, |
| 2099 | extent_op, node->no_quota); | 2099 | extent_op, node->no_quota); |
| 2100 | } else { | 2100 | } else { |
| 2101 | BUG(); | 2101 | BUG(); |
| 2102 | } | 2102 | } |
| 2103 | return ret; | 2103 | return ret; |
| 2104 | } | 2104 | } |
| 2105 | 2105 | ||
| 2106 | static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op, | 2106 | static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op, |
| 2107 | struct extent_buffer *leaf, | 2107 | struct extent_buffer *leaf, |
| 2108 | struct btrfs_extent_item *ei) | 2108 | struct btrfs_extent_item *ei) |
| 2109 | { | 2109 | { |
| 2110 | u64 flags = btrfs_extent_flags(leaf, ei); | 2110 | u64 flags = btrfs_extent_flags(leaf, ei); |
| 2111 | if (extent_op->update_flags) { | 2111 | if (extent_op->update_flags) { |
| 2112 | flags |= extent_op->flags_to_set; | 2112 | flags |= extent_op->flags_to_set; |
| 2113 | btrfs_set_extent_flags(leaf, ei, flags); | 2113 | btrfs_set_extent_flags(leaf, ei, flags); |
| 2114 | } | 2114 | } |
| 2115 | 2115 | ||
| 2116 | if (extent_op->update_key) { | 2116 | if (extent_op->update_key) { |
| 2117 | struct btrfs_tree_block_info *bi; | 2117 | struct btrfs_tree_block_info *bi; |
| 2118 | BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)); | 2118 | BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)); |
| 2119 | bi = (struct btrfs_tree_block_info *)(ei + 1); | 2119 | bi = (struct btrfs_tree_block_info *)(ei + 1); |
| 2120 | btrfs_set_tree_block_key(leaf, bi, &extent_op->key); | 2120 | btrfs_set_tree_block_key(leaf, bi, &extent_op->key); |
| 2121 | } | 2121 | } |
| 2122 | } | 2122 | } |
| 2123 | 2123 | ||
| 2124 | static int run_delayed_extent_op(struct btrfs_trans_handle *trans, | 2124 | static int run_delayed_extent_op(struct btrfs_trans_handle *trans, |
| 2125 | struct btrfs_root *root, | 2125 | struct btrfs_root *root, |
| 2126 | struct btrfs_delayed_ref_node *node, | 2126 | struct btrfs_delayed_ref_node *node, |
| 2127 | struct btrfs_delayed_extent_op *extent_op) | 2127 | struct btrfs_delayed_extent_op *extent_op) |
| 2128 | { | 2128 | { |
| 2129 | struct btrfs_key key; | 2129 | struct btrfs_key key; |
| 2130 | struct btrfs_path *path; | 2130 | struct btrfs_path *path; |
| 2131 | struct btrfs_extent_item *ei; | 2131 | struct btrfs_extent_item *ei; |
| 2132 | struct extent_buffer *leaf; | 2132 | struct extent_buffer *leaf; |
| 2133 | u32 item_size; | 2133 | u32 item_size; |
| 2134 | int ret; | 2134 | int ret; |
| 2135 | int err = 0; | 2135 | int err = 0; |
| 2136 | int metadata = !extent_op->is_data; | 2136 | int metadata = !extent_op->is_data; |
| 2137 | 2137 | ||
| 2138 | if (trans->aborted) | 2138 | if (trans->aborted) |
| 2139 | return 0; | 2139 | return 0; |
| 2140 | 2140 | ||
| 2141 | if (metadata && !btrfs_fs_incompat(root->fs_info, SKINNY_METADATA)) | 2141 | if (metadata && !btrfs_fs_incompat(root->fs_info, SKINNY_METADATA)) |
| 2142 | metadata = 0; | 2142 | metadata = 0; |
| 2143 | 2143 | ||
| 2144 | path = btrfs_alloc_path(); | 2144 | path = btrfs_alloc_path(); |
| 2145 | if (!path) | 2145 | if (!path) |
| 2146 | return -ENOMEM; | 2146 | return -ENOMEM; |
| 2147 | 2147 | ||
| 2148 | key.objectid = node->bytenr; | 2148 | key.objectid = node->bytenr; |
| 2149 | 2149 | ||
| 2150 | if (metadata) { | 2150 | if (metadata) { |
| 2151 | key.type = BTRFS_METADATA_ITEM_KEY; | 2151 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 2152 | key.offset = extent_op->level; | 2152 | key.offset = extent_op->level; |
| 2153 | } else { | 2153 | } else { |
| 2154 | key.type = BTRFS_EXTENT_ITEM_KEY; | 2154 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 2155 | key.offset = node->num_bytes; | 2155 | key.offset = node->num_bytes; |
| 2156 | } | 2156 | } |
| 2157 | 2157 | ||
| 2158 | again: | 2158 | again: |
| 2159 | path->reada = 1; | 2159 | path->reada = 1; |
| 2160 | path->leave_spinning = 1; | 2160 | path->leave_spinning = 1; |
| 2161 | ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, | 2161 | ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key, |
| 2162 | path, 0, 1); | 2162 | path, 0, 1); |
| 2163 | if (ret < 0) { | 2163 | if (ret < 0) { |
| 2164 | err = ret; | 2164 | err = ret; |
| 2165 | goto out; | 2165 | goto out; |
| 2166 | } | 2166 | } |
| 2167 | if (ret > 0) { | 2167 | if (ret > 0) { |
| 2168 | if (metadata) { | 2168 | if (metadata) { |
| 2169 | if (path->slots[0] > 0) { | 2169 | if (path->slots[0] > 0) { |
| 2170 | path->slots[0]--; | 2170 | path->slots[0]--; |
| 2171 | btrfs_item_key_to_cpu(path->nodes[0], &key, | 2171 | btrfs_item_key_to_cpu(path->nodes[0], &key, |
| 2172 | path->slots[0]); | 2172 | path->slots[0]); |
| 2173 | if (key.objectid == node->bytenr && | 2173 | if (key.objectid == node->bytenr && |
| 2174 | key.type == BTRFS_EXTENT_ITEM_KEY && | 2174 | key.type == BTRFS_EXTENT_ITEM_KEY && |
| 2175 | key.offset == node->num_bytes) | 2175 | key.offset == node->num_bytes) |
| 2176 | ret = 0; | 2176 | ret = 0; |
| 2177 | } | 2177 | } |
| 2178 | if (ret > 0) { | 2178 | if (ret > 0) { |
| 2179 | btrfs_release_path(path); | 2179 | btrfs_release_path(path); |
| 2180 | metadata = 0; | 2180 | metadata = 0; |
| 2181 | 2181 | ||
| 2182 | key.objectid = node->bytenr; | 2182 | key.objectid = node->bytenr; |
| 2183 | key.offset = node->num_bytes; | 2183 | key.offset = node->num_bytes; |
| 2184 | key.type = BTRFS_EXTENT_ITEM_KEY; | 2184 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 2185 | goto again; | 2185 | goto again; |
| 2186 | } | 2186 | } |
| 2187 | } else { | 2187 | } else { |
| 2188 | err = -EIO; | 2188 | err = -EIO; |
| 2189 | goto out; | 2189 | goto out; |
| 2190 | } | 2190 | } |
| 2191 | } | 2191 | } |
| 2192 | 2192 | ||
| 2193 | leaf = path->nodes[0]; | 2193 | leaf = path->nodes[0]; |
| 2194 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 2194 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 2195 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 2195 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 2196 | if (item_size < sizeof(*ei)) { | 2196 | if (item_size < sizeof(*ei)) { |
| 2197 | ret = convert_extent_item_v0(trans, root->fs_info->extent_root, | 2197 | ret = convert_extent_item_v0(trans, root->fs_info->extent_root, |
| 2198 | path, (u64)-1, 0); | 2198 | path, (u64)-1, 0); |
| 2199 | if (ret < 0) { | 2199 | if (ret < 0) { |
| 2200 | err = ret; | 2200 | err = ret; |
| 2201 | goto out; | 2201 | goto out; |
| 2202 | } | 2202 | } |
| 2203 | leaf = path->nodes[0]; | 2203 | leaf = path->nodes[0]; |
| 2204 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 2204 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 2205 | } | 2205 | } |
| 2206 | #endif | 2206 | #endif |
| 2207 | BUG_ON(item_size < sizeof(*ei)); | 2207 | BUG_ON(item_size < sizeof(*ei)); |
| 2208 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 2208 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 2209 | __run_delayed_extent_op(extent_op, leaf, ei); | 2209 | __run_delayed_extent_op(extent_op, leaf, ei); |
| 2210 | 2210 | ||
| 2211 | btrfs_mark_buffer_dirty(leaf); | 2211 | btrfs_mark_buffer_dirty(leaf); |
| 2212 | out: | 2212 | out: |
| 2213 | btrfs_free_path(path); | 2213 | btrfs_free_path(path); |
| 2214 | return err; | 2214 | return err; |
| 2215 | } | 2215 | } |
| 2216 | 2216 | ||
| 2217 | static int run_delayed_tree_ref(struct btrfs_trans_handle *trans, | 2217 | static int run_delayed_tree_ref(struct btrfs_trans_handle *trans, |
| 2218 | struct btrfs_root *root, | 2218 | struct btrfs_root *root, |
| 2219 | struct btrfs_delayed_ref_node *node, | 2219 | struct btrfs_delayed_ref_node *node, |
| 2220 | struct btrfs_delayed_extent_op *extent_op, | 2220 | struct btrfs_delayed_extent_op *extent_op, |
| 2221 | int insert_reserved) | 2221 | int insert_reserved) |
| 2222 | { | 2222 | { |
| 2223 | int ret = 0; | 2223 | int ret = 0; |
| 2224 | struct btrfs_delayed_tree_ref *ref; | 2224 | struct btrfs_delayed_tree_ref *ref; |
| 2225 | struct btrfs_key ins; | 2225 | struct btrfs_key ins; |
| 2226 | u64 parent = 0; | 2226 | u64 parent = 0; |
| 2227 | u64 ref_root = 0; | 2227 | u64 ref_root = 0; |
| 2228 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, | 2228 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, |
| 2229 | SKINNY_METADATA); | 2229 | SKINNY_METADATA); |
| 2230 | 2230 | ||
| 2231 | ref = btrfs_delayed_node_to_tree_ref(node); | 2231 | ref = btrfs_delayed_node_to_tree_ref(node); |
| 2232 | trace_run_delayed_tree_ref(node, ref, node->action); | 2232 | trace_run_delayed_tree_ref(node, ref, node->action); |
| 2233 | 2233 | ||
| 2234 | if (node->type == BTRFS_SHARED_BLOCK_REF_KEY) | 2234 | if (node->type == BTRFS_SHARED_BLOCK_REF_KEY) |
| 2235 | parent = ref->parent; | 2235 | parent = ref->parent; |
| 2236 | ref_root = ref->root; | 2236 | ref_root = ref->root; |
| 2237 | 2237 | ||
| 2238 | ins.objectid = node->bytenr; | 2238 | ins.objectid = node->bytenr; |
| 2239 | if (skinny_metadata) { | 2239 | if (skinny_metadata) { |
| 2240 | ins.offset = ref->level; | 2240 | ins.offset = ref->level; |
| 2241 | ins.type = BTRFS_METADATA_ITEM_KEY; | 2241 | ins.type = BTRFS_METADATA_ITEM_KEY; |
| 2242 | } else { | 2242 | } else { |
| 2243 | ins.offset = node->num_bytes; | 2243 | ins.offset = node->num_bytes; |
| 2244 | ins.type = BTRFS_EXTENT_ITEM_KEY; | 2244 | ins.type = BTRFS_EXTENT_ITEM_KEY; |
| 2245 | } | 2245 | } |
| 2246 | 2246 | ||
| 2247 | BUG_ON(node->ref_mod != 1); | 2247 | BUG_ON(node->ref_mod != 1); |
| 2248 | if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) { | 2248 | if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) { |
| 2249 | BUG_ON(!extent_op || !extent_op->update_flags); | 2249 | BUG_ON(!extent_op || !extent_op->update_flags); |
| 2250 | ret = alloc_reserved_tree_block(trans, root, | 2250 | ret = alloc_reserved_tree_block(trans, root, |
| 2251 | parent, ref_root, | 2251 | parent, ref_root, |
| 2252 | extent_op->flags_to_set, | 2252 | extent_op->flags_to_set, |
| 2253 | &extent_op->key, | 2253 | &extent_op->key, |
| 2254 | ref->level, &ins, | 2254 | ref->level, &ins, |
| 2255 | node->no_quota); | 2255 | node->no_quota); |
| 2256 | } else if (node->action == BTRFS_ADD_DELAYED_REF) { | 2256 | } else if (node->action == BTRFS_ADD_DELAYED_REF) { |
| 2257 | ret = __btrfs_inc_extent_ref(trans, root, node->bytenr, | 2257 | ret = __btrfs_inc_extent_ref(trans, root, node->bytenr, |
| 2258 | node->num_bytes, parent, ref_root, | 2258 | node->num_bytes, parent, ref_root, |
| 2259 | ref->level, 0, 1, node->no_quota, | 2259 | ref->level, 0, 1, node->no_quota, |
| 2260 | extent_op); | 2260 | extent_op); |
| 2261 | } else if (node->action == BTRFS_DROP_DELAYED_REF) { | 2261 | } else if (node->action == BTRFS_DROP_DELAYED_REF) { |
| 2262 | ret = __btrfs_free_extent(trans, root, node->bytenr, | 2262 | ret = __btrfs_free_extent(trans, root, node->bytenr, |
| 2263 | node->num_bytes, parent, ref_root, | 2263 | node->num_bytes, parent, ref_root, |
| 2264 | ref->level, 0, 1, extent_op, | 2264 | ref->level, 0, 1, extent_op, |
| 2265 | node->no_quota); | 2265 | node->no_quota); |
| 2266 | } else { | 2266 | } else { |
| 2267 | BUG(); | 2267 | BUG(); |
| 2268 | } | 2268 | } |
| 2269 | return ret; | 2269 | return ret; |
| 2270 | } | 2270 | } |
| 2271 | 2271 | ||
| 2272 | /* helper function to actually process a single delayed ref entry */ | 2272 | /* helper function to actually process a single delayed ref entry */ |
| 2273 | static int run_one_delayed_ref(struct btrfs_trans_handle *trans, | 2273 | static int run_one_delayed_ref(struct btrfs_trans_handle *trans, |
| 2274 | struct btrfs_root *root, | 2274 | struct btrfs_root *root, |
| 2275 | struct btrfs_delayed_ref_node *node, | 2275 | struct btrfs_delayed_ref_node *node, |
| 2276 | struct btrfs_delayed_extent_op *extent_op, | 2276 | struct btrfs_delayed_extent_op *extent_op, |
| 2277 | int insert_reserved) | 2277 | int insert_reserved) |
| 2278 | { | 2278 | { |
| 2279 | int ret = 0; | 2279 | int ret = 0; |
| 2280 | 2280 | ||
| 2281 | if (trans->aborted) { | 2281 | if (trans->aborted) { |
| 2282 | if (insert_reserved) | 2282 | if (insert_reserved) |
| 2283 | btrfs_pin_extent(root, node->bytenr, | 2283 | btrfs_pin_extent(root, node->bytenr, |
| 2284 | node->num_bytes, 1); | 2284 | node->num_bytes, 1); |
| 2285 | return 0; | 2285 | return 0; |
| 2286 | } | 2286 | } |
| 2287 | 2287 | ||
| 2288 | if (btrfs_delayed_ref_is_head(node)) { | 2288 | if (btrfs_delayed_ref_is_head(node)) { |
| 2289 | struct btrfs_delayed_ref_head *head; | 2289 | struct btrfs_delayed_ref_head *head; |
| 2290 | /* | 2290 | /* |
| 2291 | * we've hit the end of the chain and we were supposed | 2291 | * we've hit the end of the chain and we were supposed |
| 2292 | * to insert this extent into the tree. But, it got | 2292 | * to insert this extent into the tree. But, it got |
| 2293 | * deleted before we ever needed to insert it, so all | 2293 | * deleted before we ever needed to insert it, so all |
| 2294 | * we have to do is clean up the accounting | 2294 | * we have to do is clean up the accounting |
| 2295 | */ | 2295 | */ |
| 2296 | BUG_ON(extent_op); | 2296 | BUG_ON(extent_op); |
| 2297 | head = btrfs_delayed_node_to_head(node); | 2297 | head = btrfs_delayed_node_to_head(node); |
| 2298 | trace_run_delayed_ref_head(node, head, node->action); | 2298 | trace_run_delayed_ref_head(node, head, node->action); |
| 2299 | 2299 | ||
| 2300 | if (insert_reserved) { | 2300 | if (insert_reserved) { |
| 2301 | btrfs_pin_extent(root, node->bytenr, | 2301 | btrfs_pin_extent(root, node->bytenr, |
| 2302 | node->num_bytes, 1); | 2302 | node->num_bytes, 1); |
| 2303 | if (head->is_data) { | 2303 | if (head->is_data) { |
| 2304 | ret = btrfs_del_csums(trans, root, | 2304 | ret = btrfs_del_csums(trans, root, |
| 2305 | node->bytenr, | 2305 | node->bytenr, |
| 2306 | node->num_bytes); | 2306 | node->num_bytes); |
| 2307 | } | 2307 | } |
| 2308 | } | 2308 | } |
| 2309 | return ret; | 2309 | return ret; |
| 2310 | } | 2310 | } |
| 2311 | 2311 | ||
| 2312 | if (node->type == BTRFS_TREE_BLOCK_REF_KEY || | 2312 | if (node->type == BTRFS_TREE_BLOCK_REF_KEY || |
| 2313 | node->type == BTRFS_SHARED_BLOCK_REF_KEY) | 2313 | node->type == BTRFS_SHARED_BLOCK_REF_KEY) |
| 2314 | ret = run_delayed_tree_ref(trans, root, node, extent_op, | 2314 | ret = run_delayed_tree_ref(trans, root, node, extent_op, |
| 2315 | insert_reserved); | 2315 | insert_reserved); |
| 2316 | else if (node->type == BTRFS_EXTENT_DATA_REF_KEY || | 2316 | else if (node->type == BTRFS_EXTENT_DATA_REF_KEY || |
| 2317 | node->type == BTRFS_SHARED_DATA_REF_KEY) | 2317 | node->type == BTRFS_SHARED_DATA_REF_KEY) |
| 2318 | ret = run_delayed_data_ref(trans, root, node, extent_op, | 2318 | ret = run_delayed_data_ref(trans, root, node, extent_op, |
| 2319 | insert_reserved); | 2319 | insert_reserved); |
| 2320 | else | 2320 | else |
| 2321 | BUG(); | 2321 | BUG(); |
| 2322 | return ret; | 2322 | return ret; |
| 2323 | } | 2323 | } |
| 2324 | 2324 | ||
| 2325 | static noinline struct btrfs_delayed_ref_node * | 2325 | static noinline struct btrfs_delayed_ref_node * |
| 2326 | select_delayed_ref(struct btrfs_delayed_ref_head *head) | 2326 | select_delayed_ref(struct btrfs_delayed_ref_head *head) |
| 2327 | { | 2327 | { |
| 2328 | struct rb_node *node; | 2328 | struct rb_node *node; |
| 2329 | struct btrfs_delayed_ref_node *ref, *last = NULL;; | 2329 | struct btrfs_delayed_ref_node *ref, *last = NULL;; |
| 2330 | 2330 | ||
| 2331 | /* | 2331 | /* |
| 2332 | * select delayed ref of type BTRFS_ADD_DELAYED_REF first. | 2332 | * select delayed ref of type BTRFS_ADD_DELAYED_REF first. |
| 2333 | * this prevents ref count from going down to zero when | 2333 | * this prevents ref count from going down to zero when |
| 2334 | * there still are pending delayed ref. | 2334 | * there still are pending delayed ref. |
| 2335 | */ | 2335 | */ |
| 2336 | node = rb_first(&head->ref_root); | 2336 | node = rb_first(&head->ref_root); |
| 2337 | while (node) { | 2337 | while (node) { |
| 2338 | ref = rb_entry(node, struct btrfs_delayed_ref_node, | 2338 | ref = rb_entry(node, struct btrfs_delayed_ref_node, |
| 2339 | rb_node); | 2339 | rb_node); |
| 2340 | if (ref->action == BTRFS_ADD_DELAYED_REF) | 2340 | if (ref->action == BTRFS_ADD_DELAYED_REF) |
| 2341 | return ref; | 2341 | return ref; |
| 2342 | else if (last == NULL) | 2342 | else if (last == NULL) |
| 2343 | last = ref; | 2343 | last = ref; |
| 2344 | node = rb_next(node); | 2344 | node = rb_next(node); |
| 2345 | } | 2345 | } |
| 2346 | return last; | 2346 | return last; |
| 2347 | } | 2347 | } |
| 2348 | 2348 | ||
| 2349 | /* | 2349 | /* |
| 2350 | * Returns 0 on success or if called with an already aborted transaction. | 2350 | * Returns 0 on success or if called with an already aborted transaction. |
| 2351 | * Returns -ENOMEM or -EIO on failure and will abort the transaction. | 2351 | * Returns -ENOMEM or -EIO on failure and will abort the transaction. |
| 2352 | */ | 2352 | */ |
| 2353 | static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, | 2353 | static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, |
| 2354 | struct btrfs_root *root, | 2354 | struct btrfs_root *root, |
| 2355 | unsigned long nr) | 2355 | unsigned long nr) |
| 2356 | { | 2356 | { |
| 2357 | struct btrfs_delayed_ref_root *delayed_refs; | 2357 | struct btrfs_delayed_ref_root *delayed_refs; |
| 2358 | struct btrfs_delayed_ref_node *ref; | 2358 | struct btrfs_delayed_ref_node *ref; |
| 2359 | struct btrfs_delayed_ref_head *locked_ref = NULL; | 2359 | struct btrfs_delayed_ref_head *locked_ref = NULL; |
| 2360 | struct btrfs_delayed_extent_op *extent_op; | 2360 | struct btrfs_delayed_extent_op *extent_op; |
| 2361 | struct btrfs_fs_info *fs_info = root->fs_info; | 2361 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 2362 | ktime_t start = ktime_get(); | 2362 | ktime_t start = ktime_get(); |
| 2363 | int ret; | 2363 | int ret; |
| 2364 | unsigned long count = 0; | 2364 | unsigned long count = 0; |
| 2365 | unsigned long actual_count = 0; | 2365 | unsigned long actual_count = 0; |
| 2366 | int must_insert_reserved = 0; | 2366 | int must_insert_reserved = 0; |
| 2367 | 2367 | ||
| 2368 | delayed_refs = &trans->transaction->delayed_refs; | 2368 | delayed_refs = &trans->transaction->delayed_refs; |
| 2369 | while (1) { | 2369 | while (1) { |
| 2370 | if (!locked_ref) { | 2370 | if (!locked_ref) { |
| 2371 | if (count >= nr) | 2371 | if (count >= nr) |
| 2372 | break; | 2372 | break; |
| 2373 | 2373 | ||
| 2374 | spin_lock(&delayed_refs->lock); | 2374 | spin_lock(&delayed_refs->lock); |
| 2375 | locked_ref = btrfs_select_ref_head(trans); | 2375 | locked_ref = btrfs_select_ref_head(trans); |
| 2376 | if (!locked_ref) { | 2376 | if (!locked_ref) { |
| 2377 | spin_unlock(&delayed_refs->lock); | 2377 | spin_unlock(&delayed_refs->lock); |
| 2378 | break; | 2378 | break; |
| 2379 | } | 2379 | } |
| 2380 | 2380 | ||
| 2381 | /* grab the lock that says we are going to process | 2381 | /* grab the lock that says we are going to process |
| 2382 | * all the refs for this head */ | 2382 | * all the refs for this head */ |
| 2383 | ret = btrfs_delayed_ref_lock(trans, locked_ref); | 2383 | ret = btrfs_delayed_ref_lock(trans, locked_ref); |
| 2384 | spin_unlock(&delayed_refs->lock); | 2384 | spin_unlock(&delayed_refs->lock); |
| 2385 | /* | 2385 | /* |
| 2386 | * we may have dropped the spin lock to get the head | 2386 | * we may have dropped the spin lock to get the head |
| 2387 | * mutex lock, and that might have given someone else | 2387 | * mutex lock, and that might have given someone else |
| 2388 | * time to free the head. If that's true, it has been | 2388 | * time to free the head. If that's true, it has been |
| 2389 | * removed from our list and we can move on. | 2389 | * removed from our list and we can move on. |
| 2390 | */ | 2390 | */ |
| 2391 | if (ret == -EAGAIN) { | 2391 | if (ret == -EAGAIN) { |
| 2392 | locked_ref = NULL; | 2392 | locked_ref = NULL; |
| 2393 | count++; | 2393 | count++; |
| 2394 | continue; | 2394 | continue; |
| 2395 | } | 2395 | } |
| 2396 | } | 2396 | } |
| 2397 | 2397 | ||
| 2398 | /* | 2398 | /* |
| 2399 | * We need to try and merge add/drops of the same ref since we | 2399 | * We need to try and merge add/drops of the same ref since we |
| 2400 | * can run into issues with relocate dropping the implicit ref | 2400 | * can run into issues with relocate dropping the implicit ref |
| 2401 | * and then it being added back again before the drop can | 2401 | * and then it being added back again before the drop can |
| 2402 | * finish. If we merged anything we need to re-loop so we can | 2402 | * finish. If we merged anything we need to re-loop so we can |
| 2403 | * get a good ref. | 2403 | * get a good ref. |
| 2404 | */ | 2404 | */ |
| 2405 | spin_lock(&locked_ref->lock); | 2405 | spin_lock(&locked_ref->lock); |
| 2406 | btrfs_merge_delayed_refs(trans, fs_info, delayed_refs, | 2406 | btrfs_merge_delayed_refs(trans, fs_info, delayed_refs, |
| 2407 | locked_ref); | 2407 | locked_ref); |
| 2408 | 2408 | ||
| 2409 | /* | 2409 | /* |
| 2410 | * locked_ref is the head node, so we have to go one | 2410 | * locked_ref is the head node, so we have to go one |
| 2411 | * node back for any delayed ref updates | 2411 | * node back for any delayed ref updates |
| 2412 | */ | 2412 | */ |
| 2413 | ref = select_delayed_ref(locked_ref); | 2413 | ref = select_delayed_ref(locked_ref); |
| 2414 | 2414 | ||
| 2415 | if (ref && ref->seq && | 2415 | if (ref && ref->seq && |
| 2416 | btrfs_check_delayed_seq(fs_info, delayed_refs, ref->seq)) { | 2416 | btrfs_check_delayed_seq(fs_info, delayed_refs, ref->seq)) { |
| 2417 | spin_unlock(&locked_ref->lock); | 2417 | spin_unlock(&locked_ref->lock); |
| 2418 | btrfs_delayed_ref_unlock(locked_ref); | 2418 | btrfs_delayed_ref_unlock(locked_ref); |
| 2419 | spin_lock(&delayed_refs->lock); | 2419 | spin_lock(&delayed_refs->lock); |
| 2420 | locked_ref->processing = 0; | 2420 | locked_ref->processing = 0; |
| 2421 | delayed_refs->num_heads_ready++; | 2421 | delayed_refs->num_heads_ready++; |
| 2422 | spin_unlock(&delayed_refs->lock); | 2422 | spin_unlock(&delayed_refs->lock); |
| 2423 | locked_ref = NULL; | 2423 | locked_ref = NULL; |
| 2424 | cond_resched(); | 2424 | cond_resched(); |
| 2425 | count++; | 2425 | count++; |
| 2426 | continue; | 2426 | continue; |
| 2427 | } | 2427 | } |
| 2428 | 2428 | ||
| 2429 | /* | 2429 | /* |
| 2430 | * record the must insert reserved flag before we | 2430 | * record the must insert reserved flag before we |
| 2431 | * drop the spin lock. | 2431 | * drop the spin lock. |
| 2432 | */ | 2432 | */ |
| 2433 | must_insert_reserved = locked_ref->must_insert_reserved; | 2433 | must_insert_reserved = locked_ref->must_insert_reserved; |
| 2434 | locked_ref->must_insert_reserved = 0; | 2434 | locked_ref->must_insert_reserved = 0; |
| 2435 | 2435 | ||
| 2436 | extent_op = locked_ref->extent_op; | 2436 | extent_op = locked_ref->extent_op; |
| 2437 | locked_ref->extent_op = NULL; | 2437 | locked_ref->extent_op = NULL; |
| 2438 | 2438 | ||
| 2439 | if (!ref) { | 2439 | if (!ref) { |
| 2440 | 2440 | ||
| 2441 | 2441 | ||
| 2442 | /* All delayed refs have been processed, Go ahead | 2442 | /* All delayed refs have been processed, Go ahead |
| 2443 | * and send the head node to run_one_delayed_ref, | 2443 | * and send the head node to run_one_delayed_ref, |
| 2444 | * so that any accounting fixes can happen | 2444 | * so that any accounting fixes can happen |
| 2445 | */ | 2445 | */ |
| 2446 | ref = &locked_ref->node; | 2446 | ref = &locked_ref->node; |
| 2447 | 2447 | ||
| 2448 | if (extent_op && must_insert_reserved) { | 2448 | if (extent_op && must_insert_reserved) { |
| 2449 | btrfs_free_delayed_extent_op(extent_op); | 2449 | btrfs_free_delayed_extent_op(extent_op); |
| 2450 | extent_op = NULL; | 2450 | extent_op = NULL; |
| 2451 | } | 2451 | } |
| 2452 | 2452 | ||
| 2453 | if (extent_op) { | 2453 | if (extent_op) { |
| 2454 | spin_unlock(&locked_ref->lock); | 2454 | spin_unlock(&locked_ref->lock); |
| 2455 | ret = run_delayed_extent_op(trans, root, | 2455 | ret = run_delayed_extent_op(trans, root, |
| 2456 | ref, extent_op); | 2456 | ref, extent_op); |
| 2457 | btrfs_free_delayed_extent_op(extent_op); | 2457 | btrfs_free_delayed_extent_op(extent_op); |
| 2458 | 2458 | ||
| 2459 | if (ret) { | 2459 | if (ret) { |
| 2460 | /* | 2460 | /* |
| 2461 | * Need to reset must_insert_reserved if | 2461 | * Need to reset must_insert_reserved if |
| 2462 | * there was an error so the abort stuff | 2462 | * there was an error so the abort stuff |
| 2463 | * can cleanup the reserved space | 2463 | * can cleanup the reserved space |
| 2464 | * properly. | 2464 | * properly. |
| 2465 | */ | 2465 | */ |
| 2466 | if (must_insert_reserved) | 2466 | if (must_insert_reserved) |
| 2467 | locked_ref->must_insert_reserved = 1; | 2467 | locked_ref->must_insert_reserved = 1; |
| 2468 | locked_ref->processing = 0; | 2468 | locked_ref->processing = 0; |
| 2469 | btrfs_debug(fs_info, "run_delayed_extent_op returned %d", ret); | 2469 | btrfs_debug(fs_info, "run_delayed_extent_op returned %d", ret); |
| 2470 | btrfs_delayed_ref_unlock(locked_ref); | 2470 | btrfs_delayed_ref_unlock(locked_ref); |
| 2471 | return ret; | 2471 | return ret; |
| 2472 | } | 2472 | } |
| 2473 | continue; | 2473 | continue; |
| 2474 | } | 2474 | } |
| 2475 | 2475 | ||
| 2476 | /* | 2476 | /* |
| 2477 | * Need to drop our head ref lock and re-aqcuire the | 2477 | * Need to drop our head ref lock and re-aqcuire the |
| 2478 | * delayed ref lock and then re-check to make sure | 2478 | * delayed ref lock and then re-check to make sure |
| 2479 | * nobody got added. | 2479 | * nobody got added. |
| 2480 | */ | 2480 | */ |
| 2481 | spin_unlock(&locked_ref->lock); | 2481 | spin_unlock(&locked_ref->lock); |
| 2482 | spin_lock(&delayed_refs->lock); | 2482 | spin_lock(&delayed_refs->lock); |
| 2483 | spin_lock(&locked_ref->lock); | 2483 | spin_lock(&locked_ref->lock); |
| 2484 | if (rb_first(&locked_ref->ref_root) || | 2484 | if (rb_first(&locked_ref->ref_root) || |
| 2485 | locked_ref->extent_op) { | 2485 | locked_ref->extent_op) { |
| 2486 | spin_unlock(&locked_ref->lock); | 2486 | spin_unlock(&locked_ref->lock); |
| 2487 | spin_unlock(&delayed_refs->lock); | 2487 | spin_unlock(&delayed_refs->lock); |
| 2488 | continue; | 2488 | continue; |
| 2489 | } | 2489 | } |
| 2490 | ref->in_tree = 0; | 2490 | ref->in_tree = 0; |
| 2491 | delayed_refs->num_heads--; | 2491 | delayed_refs->num_heads--; |
| 2492 | rb_erase(&locked_ref->href_node, | 2492 | rb_erase(&locked_ref->href_node, |
| 2493 | &delayed_refs->href_root); | 2493 | &delayed_refs->href_root); |
| 2494 | spin_unlock(&delayed_refs->lock); | 2494 | spin_unlock(&delayed_refs->lock); |
| 2495 | } else { | 2495 | } else { |
| 2496 | actual_count++; | 2496 | actual_count++; |
| 2497 | ref->in_tree = 0; | 2497 | ref->in_tree = 0; |
| 2498 | rb_erase(&ref->rb_node, &locked_ref->ref_root); | 2498 | rb_erase(&ref->rb_node, &locked_ref->ref_root); |
| 2499 | } | 2499 | } |
| 2500 | atomic_dec(&delayed_refs->num_entries); | 2500 | atomic_dec(&delayed_refs->num_entries); |
| 2501 | 2501 | ||
| 2502 | if (!btrfs_delayed_ref_is_head(ref)) { | 2502 | if (!btrfs_delayed_ref_is_head(ref)) { |
| 2503 | /* | 2503 | /* |
| 2504 | * when we play the delayed ref, also correct the | 2504 | * when we play the delayed ref, also correct the |
| 2505 | * ref_mod on head | 2505 | * ref_mod on head |
| 2506 | */ | 2506 | */ |
| 2507 | switch (ref->action) { | 2507 | switch (ref->action) { |
| 2508 | case BTRFS_ADD_DELAYED_REF: | 2508 | case BTRFS_ADD_DELAYED_REF: |
| 2509 | case BTRFS_ADD_DELAYED_EXTENT: | 2509 | case BTRFS_ADD_DELAYED_EXTENT: |
| 2510 | locked_ref->node.ref_mod -= ref->ref_mod; | 2510 | locked_ref->node.ref_mod -= ref->ref_mod; |
| 2511 | break; | 2511 | break; |
| 2512 | case BTRFS_DROP_DELAYED_REF: | 2512 | case BTRFS_DROP_DELAYED_REF: |
| 2513 | locked_ref->node.ref_mod += ref->ref_mod; | 2513 | locked_ref->node.ref_mod += ref->ref_mod; |
| 2514 | break; | 2514 | break; |
| 2515 | default: | 2515 | default: |
| 2516 | WARN_ON(1); | 2516 | WARN_ON(1); |
| 2517 | } | 2517 | } |
| 2518 | } | 2518 | } |
| 2519 | spin_unlock(&locked_ref->lock); | 2519 | spin_unlock(&locked_ref->lock); |
| 2520 | 2520 | ||
| 2521 | ret = run_one_delayed_ref(trans, root, ref, extent_op, | 2521 | ret = run_one_delayed_ref(trans, root, ref, extent_op, |
| 2522 | must_insert_reserved); | 2522 | must_insert_reserved); |
| 2523 | 2523 | ||
| 2524 | btrfs_free_delayed_extent_op(extent_op); | 2524 | btrfs_free_delayed_extent_op(extent_op); |
| 2525 | if (ret) { | 2525 | if (ret) { |
| 2526 | locked_ref->processing = 0; | 2526 | locked_ref->processing = 0; |
| 2527 | btrfs_delayed_ref_unlock(locked_ref); | 2527 | btrfs_delayed_ref_unlock(locked_ref); |
| 2528 | btrfs_put_delayed_ref(ref); | 2528 | btrfs_put_delayed_ref(ref); |
| 2529 | btrfs_debug(fs_info, "run_one_delayed_ref returned %d", ret); | 2529 | btrfs_debug(fs_info, "run_one_delayed_ref returned %d", ret); |
| 2530 | return ret; | 2530 | return ret; |
| 2531 | } | 2531 | } |
| 2532 | 2532 | ||
| 2533 | /* | 2533 | /* |
| 2534 | * If this node is a head, that means all the refs in this head | 2534 | * If this node is a head, that means all the refs in this head |
| 2535 | * have been dealt with, and we will pick the next head to deal | 2535 | * have been dealt with, and we will pick the next head to deal |
| 2536 | * with, so we must unlock the head and drop it from the cluster | 2536 | * with, so we must unlock the head and drop it from the cluster |
| 2537 | * list before we release it. | 2537 | * list before we release it. |
| 2538 | */ | 2538 | */ |
| 2539 | if (btrfs_delayed_ref_is_head(ref)) { | 2539 | if (btrfs_delayed_ref_is_head(ref)) { |
| 2540 | btrfs_delayed_ref_unlock(locked_ref); | 2540 | btrfs_delayed_ref_unlock(locked_ref); |
| 2541 | locked_ref = NULL; | 2541 | locked_ref = NULL; |
| 2542 | } | 2542 | } |
| 2543 | btrfs_put_delayed_ref(ref); | 2543 | btrfs_put_delayed_ref(ref); |
| 2544 | count++; | 2544 | count++; |
| 2545 | cond_resched(); | 2545 | cond_resched(); |
| 2546 | } | 2546 | } |
| 2547 | 2547 | ||
| 2548 | /* | 2548 | /* |
| 2549 | * We don't want to include ref heads since we can have empty ref heads | 2549 | * We don't want to include ref heads since we can have empty ref heads |
| 2550 | * and those will drastically skew our runtime down since we just do | 2550 | * and those will drastically skew our runtime down since we just do |
| 2551 | * accounting, no actual extent tree updates. | 2551 | * accounting, no actual extent tree updates. |
| 2552 | */ | 2552 | */ |
| 2553 | if (actual_count > 0) { | 2553 | if (actual_count > 0) { |
| 2554 | u64 runtime = ktime_to_ns(ktime_sub(ktime_get(), start)); | 2554 | u64 runtime = ktime_to_ns(ktime_sub(ktime_get(), start)); |
| 2555 | u64 avg; | 2555 | u64 avg; |
| 2556 | 2556 | ||
| 2557 | /* | 2557 | /* |
| 2558 | * We weigh the current average higher than our current runtime | 2558 | * We weigh the current average higher than our current runtime |
| 2559 | * to avoid large swings in the average. | 2559 | * to avoid large swings in the average. |
| 2560 | */ | 2560 | */ |
| 2561 | spin_lock(&delayed_refs->lock); | 2561 | spin_lock(&delayed_refs->lock); |
| 2562 | avg = fs_info->avg_delayed_ref_runtime * 3 + runtime; | 2562 | avg = fs_info->avg_delayed_ref_runtime * 3 + runtime; |
| 2563 | avg = div64_u64(avg, 4); | 2563 | avg = div64_u64(avg, 4); |
| 2564 | fs_info->avg_delayed_ref_runtime = avg; | 2564 | fs_info->avg_delayed_ref_runtime = avg; |
| 2565 | spin_unlock(&delayed_refs->lock); | 2565 | spin_unlock(&delayed_refs->lock); |
| 2566 | } | 2566 | } |
| 2567 | return 0; | 2567 | return 0; |
| 2568 | } | 2568 | } |
| 2569 | 2569 | ||
| 2570 | #ifdef SCRAMBLE_DELAYED_REFS | 2570 | #ifdef SCRAMBLE_DELAYED_REFS |
| 2571 | /* | 2571 | /* |
| 2572 | * Normally delayed refs get processed in ascending bytenr order. This | 2572 | * Normally delayed refs get processed in ascending bytenr order. This |
| 2573 | * correlates in most cases to the order added. To expose dependencies on this | 2573 | * correlates in most cases to the order added. To expose dependencies on this |
| 2574 | * order, we start to process the tree in the middle instead of the beginning | 2574 | * order, we start to process the tree in the middle instead of the beginning |
| 2575 | */ | 2575 | */ |
| 2576 | static u64 find_middle(struct rb_root *root) | 2576 | static u64 find_middle(struct rb_root *root) |
| 2577 | { | 2577 | { |
| 2578 | struct rb_node *n = root->rb_node; | 2578 | struct rb_node *n = root->rb_node; |
| 2579 | struct btrfs_delayed_ref_node *entry; | 2579 | struct btrfs_delayed_ref_node *entry; |
| 2580 | int alt = 1; | 2580 | int alt = 1; |
| 2581 | u64 middle; | 2581 | u64 middle; |
| 2582 | u64 first = 0, last = 0; | 2582 | u64 first = 0, last = 0; |
| 2583 | 2583 | ||
| 2584 | n = rb_first(root); | 2584 | n = rb_first(root); |
| 2585 | if (n) { | 2585 | if (n) { |
| 2586 | entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); | 2586 | entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); |
| 2587 | first = entry->bytenr; | 2587 | first = entry->bytenr; |
| 2588 | } | 2588 | } |
| 2589 | n = rb_last(root); | 2589 | n = rb_last(root); |
| 2590 | if (n) { | 2590 | if (n) { |
| 2591 | entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); | 2591 | entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); |
| 2592 | last = entry->bytenr; | 2592 | last = entry->bytenr; |
| 2593 | } | 2593 | } |
| 2594 | n = root->rb_node; | 2594 | n = root->rb_node; |
| 2595 | 2595 | ||
| 2596 | while (n) { | 2596 | while (n) { |
| 2597 | entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); | 2597 | entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node); |
| 2598 | WARN_ON(!entry->in_tree); | 2598 | WARN_ON(!entry->in_tree); |
| 2599 | 2599 | ||
| 2600 | middle = entry->bytenr; | 2600 | middle = entry->bytenr; |
| 2601 | 2601 | ||
| 2602 | if (alt) | 2602 | if (alt) |
| 2603 | n = n->rb_left; | 2603 | n = n->rb_left; |
| 2604 | else | 2604 | else |
| 2605 | n = n->rb_right; | 2605 | n = n->rb_right; |
| 2606 | 2606 | ||
| 2607 | alt = 1 - alt; | 2607 | alt = 1 - alt; |
| 2608 | } | 2608 | } |
| 2609 | return middle; | 2609 | return middle; |
| 2610 | } | 2610 | } |
| 2611 | #endif | 2611 | #endif |
| 2612 | 2612 | ||
| 2613 | static inline u64 heads_to_leaves(struct btrfs_root *root, u64 heads) | 2613 | static inline u64 heads_to_leaves(struct btrfs_root *root, u64 heads) |
| 2614 | { | 2614 | { |
| 2615 | u64 num_bytes; | 2615 | u64 num_bytes; |
| 2616 | 2616 | ||
| 2617 | num_bytes = heads * (sizeof(struct btrfs_extent_item) + | 2617 | num_bytes = heads * (sizeof(struct btrfs_extent_item) + |
| 2618 | sizeof(struct btrfs_extent_inline_ref)); | 2618 | sizeof(struct btrfs_extent_inline_ref)); |
| 2619 | if (!btrfs_fs_incompat(root->fs_info, SKINNY_METADATA)) | 2619 | if (!btrfs_fs_incompat(root->fs_info, SKINNY_METADATA)) |
| 2620 | num_bytes += heads * sizeof(struct btrfs_tree_block_info); | 2620 | num_bytes += heads * sizeof(struct btrfs_tree_block_info); |
| 2621 | 2621 | ||
| 2622 | /* | 2622 | /* |
| 2623 | * We don't ever fill up leaves all the way so multiply by 2 just to be | 2623 | * We don't ever fill up leaves all the way so multiply by 2 just to be |
| 2624 | * closer to what we're really going to want to ouse. | 2624 | * closer to what we're really going to want to ouse. |
| 2625 | */ | 2625 | */ |
| 2626 | return div64_u64(num_bytes, BTRFS_LEAF_DATA_SIZE(root)); | 2626 | return div64_u64(num_bytes, BTRFS_LEAF_DATA_SIZE(root)); |
| 2627 | } | 2627 | } |
| 2628 | 2628 | ||
| 2629 | int btrfs_check_space_for_delayed_refs(struct btrfs_trans_handle *trans, | 2629 | int btrfs_check_space_for_delayed_refs(struct btrfs_trans_handle *trans, |
| 2630 | struct btrfs_root *root) | 2630 | struct btrfs_root *root) |
| 2631 | { | 2631 | { |
| 2632 | struct btrfs_block_rsv *global_rsv; | 2632 | struct btrfs_block_rsv *global_rsv; |
| 2633 | u64 num_heads = trans->transaction->delayed_refs.num_heads_ready; | 2633 | u64 num_heads = trans->transaction->delayed_refs.num_heads_ready; |
| 2634 | u64 num_bytes; | 2634 | u64 num_bytes; |
| 2635 | int ret = 0; | 2635 | int ret = 0; |
| 2636 | 2636 | ||
| 2637 | num_bytes = btrfs_calc_trans_metadata_size(root, 1); | 2637 | num_bytes = btrfs_calc_trans_metadata_size(root, 1); |
| 2638 | num_heads = heads_to_leaves(root, num_heads); | 2638 | num_heads = heads_to_leaves(root, num_heads); |
| 2639 | if (num_heads > 1) | 2639 | if (num_heads > 1) |
| 2640 | num_bytes += (num_heads - 1) * root->nodesize; | 2640 | num_bytes += (num_heads - 1) * root->nodesize; |
| 2641 | num_bytes <<= 1; | 2641 | num_bytes <<= 1; |
| 2642 | global_rsv = &root->fs_info->global_block_rsv; | 2642 | global_rsv = &root->fs_info->global_block_rsv; |
| 2643 | 2643 | ||
| 2644 | /* | 2644 | /* |
| 2645 | * If we can't allocate any more chunks lets make sure we have _lots_ of | 2645 | * If we can't allocate any more chunks lets make sure we have _lots_ of |
| 2646 | * wiggle room since running delayed refs can create more delayed refs. | 2646 | * wiggle room since running delayed refs can create more delayed refs. |
| 2647 | */ | 2647 | */ |
| 2648 | if (global_rsv->space_info->full) | 2648 | if (global_rsv->space_info->full) |
| 2649 | num_bytes <<= 1; | 2649 | num_bytes <<= 1; |
| 2650 | 2650 | ||
| 2651 | spin_lock(&global_rsv->lock); | 2651 | spin_lock(&global_rsv->lock); |
| 2652 | if (global_rsv->reserved <= num_bytes) | 2652 | if (global_rsv->reserved <= num_bytes) |
| 2653 | ret = 1; | 2653 | ret = 1; |
| 2654 | spin_unlock(&global_rsv->lock); | 2654 | spin_unlock(&global_rsv->lock); |
| 2655 | return ret; | 2655 | return ret; |
| 2656 | } | 2656 | } |
| 2657 | 2657 | ||
| 2658 | int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans, | 2658 | int btrfs_should_throttle_delayed_refs(struct btrfs_trans_handle *trans, |
| 2659 | struct btrfs_root *root) | 2659 | struct btrfs_root *root) |
| 2660 | { | 2660 | { |
| 2661 | struct btrfs_fs_info *fs_info = root->fs_info; | 2661 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 2662 | u64 num_entries = | 2662 | u64 num_entries = |
| 2663 | atomic_read(&trans->transaction->delayed_refs.num_entries); | 2663 | atomic_read(&trans->transaction->delayed_refs.num_entries); |
| 2664 | u64 avg_runtime; | 2664 | u64 avg_runtime; |
| 2665 | u64 val; | 2665 | u64 val; |
| 2666 | 2666 | ||
| 2667 | smp_mb(); | 2667 | smp_mb(); |
| 2668 | avg_runtime = fs_info->avg_delayed_ref_runtime; | 2668 | avg_runtime = fs_info->avg_delayed_ref_runtime; |
| 2669 | val = num_entries * avg_runtime; | 2669 | val = num_entries * avg_runtime; |
| 2670 | if (num_entries * avg_runtime >= NSEC_PER_SEC) | 2670 | if (num_entries * avg_runtime >= NSEC_PER_SEC) |
| 2671 | return 1; | 2671 | return 1; |
| 2672 | if (val >= NSEC_PER_SEC / 2) | 2672 | if (val >= NSEC_PER_SEC / 2) |
| 2673 | return 2; | 2673 | return 2; |
| 2674 | 2674 | ||
| 2675 | return btrfs_check_space_for_delayed_refs(trans, root); | 2675 | return btrfs_check_space_for_delayed_refs(trans, root); |
| 2676 | } | 2676 | } |
| 2677 | 2677 | ||
| 2678 | struct async_delayed_refs { | 2678 | struct async_delayed_refs { |
| 2679 | struct btrfs_root *root; | 2679 | struct btrfs_root *root; |
| 2680 | int count; | 2680 | int count; |
| 2681 | int error; | 2681 | int error; |
| 2682 | int sync; | 2682 | int sync; |
| 2683 | struct completion wait; | 2683 | struct completion wait; |
| 2684 | struct btrfs_work work; | 2684 | struct btrfs_work work; |
| 2685 | }; | 2685 | }; |
| 2686 | 2686 | ||
| 2687 | static void delayed_ref_async_start(struct btrfs_work *work) | 2687 | static void delayed_ref_async_start(struct btrfs_work *work) |
| 2688 | { | 2688 | { |
| 2689 | struct async_delayed_refs *async; | 2689 | struct async_delayed_refs *async; |
| 2690 | struct btrfs_trans_handle *trans; | 2690 | struct btrfs_trans_handle *trans; |
| 2691 | int ret; | 2691 | int ret; |
| 2692 | 2692 | ||
| 2693 | async = container_of(work, struct async_delayed_refs, work); | 2693 | async = container_of(work, struct async_delayed_refs, work); |
| 2694 | 2694 | ||
| 2695 | trans = btrfs_join_transaction(async->root); | 2695 | trans = btrfs_join_transaction(async->root); |
| 2696 | if (IS_ERR(trans)) { | 2696 | if (IS_ERR(trans)) { |
| 2697 | async->error = PTR_ERR(trans); | 2697 | async->error = PTR_ERR(trans); |
| 2698 | goto done; | 2698 | goto done; |
| 2699 | } | 2699 | } |
| 2700 | 2700 | ||
| 2701 | /* | 2701 | /* |
| 2702 | * trans->sync means that when we call end_transaciton, we won't | 2702 | * trans->sync means that when we call end_transaciton, we won't |
| 2703 | * wait on delayed refs | 2703 | * wait on delayed refs |
| 2704 | */ | 2704 | */ |
| 2705 | trans->sync = true; | 2705 | trans->sync = true; |
| 2706 | ret = btrfs_run_delayed_refs(trans, async->root, async->count); | 2706 | ret = btrfs_run_delayed_refs(trans, async->root, async->count); |
| 2707 | if (ret) | 2707 | if (ret) |
| 2708 | async->error = ret; | 2708 | async->error = ret; |
| 2709 | 2709 | ||
| 2710 | ret = btrfs_end_transaction(trans, async->root); | 2710 | ret = btrfs_end_transaction(trans, async->root); |
| 2711 | if (ret && !async->error) | 2711 | if (ret && !async->error) |
| 2712 | async->error = ret; | 2712 | async->error = ret; |
| 2713 | done: | 2713 | done: |
| 2714 | if (async->sync) | 2714 | if (async->sync) |
| 2715 | complete(&async->wait); | 2715 | complete(&async->wait); |
| 2716 | else | 2716 | else |
| 2717 | kfree(async); | 2717 | kfree(async); |
| 2718 | } | 2718 | } |
| 2719 | 2719 | ||
| 2720 | int btrfs_async_run_delayed_refs(struct btrfs_root *root, | 2720 | int btrfs_async_run_delayed_refs(struct btrfs_root *root, |
| 2721 | unsigned long count, int wait) | 2721 | unsigned long count, int wait) |
| 2722 | { | 2722 | { |
| 2723 | struct async_delayed_refs *async; | 2723 | struct async_delayed_refs *async; |
| 2724 | int ret; | 2724 | int ret; |
| 2725 | 2725 | ||
| 2726 | async = kmalloc(sizeof(*async), GFP_NOFS); | 2726 | async = kmalloc(sizeof(*async), GFP_NOFS); |
| 2727 | if (!async) | 2727 | if (!async) |
| 2728 | return -ENOMEM; | 2728 | return -ENOMEM; |
| 2729 | 2729 | ||
| 2730 | async->root = root->fs_info->tree_root; | 2730 | async->root = root->fs_info->tree_root; |
| 2731 | async->count = count; | 2731 | async->count = count; |
| 2732 | async->error = 0; | 2732 | async->error = 0; |
| 2733 | if (wait) | 2733 | if (wait) |
| 2734 | async->sync = 1; | 2734 | async->sync = 1; |
| 2735 | else | 2735 | else |
| 2736 | async->sync = 0; | 2736 | async->sync = 0; |
| 2737 | init_completion(&async->wait); | 2737 | init_completion(&async->wait); |
| 2738 | 2738 | ||
| 2739 | btrfs_init_work(&async->work, btrfs_extent_refs_helper, | 2739 | btrfs_init_work(&async->work, btrfs_extent_refs_helper, |
| 2740 | delayed_ref_async_start, NULL, NULL); | 2740 | delayed_ref_async_start, NULL, NULL); |
| 2741 | 2741 | ||
| 2742 | btrfs_queue_work(root->fs_info->extent_workers, &async->work); | 2742 | btrfs_queue_work(root->fs_info->extent_workers, &async->work); |
| 2743 | 2743 | ||
| 2744 | if (wait) { | 2744 | if (wait) { |
| 2745 | wait_for_completion(&async->wait); | 2745 | wait_for_completion(&async->wait); |
| 2746 | ret = async->error; | 2746 | ret = async->error; |
| 2747 | kfree(async); | 2747 | kfree(async); |
| 2748 | return ret; | 2748 | return ret; |
| 2749 | } | 2749 | } |
| 2750 | return 0; | 2750 | return 0; |
| 2751 | } | 2751 | } |
| 2752 | 2752 | ||
| 2753 | /* | 2753 | /* |
| 2754 | * this starts processing the delayed reference count updates and | 2754 | * this starts processing the delayed reference count updates and |
| 2755 | * extent insertions we have queued up so far. count can be | 2755 | * extent insertions we have queued up so far. count can be |
| 2756 | * 0, which means to process everything in the tree at the start | 2756 | * 0, which means to process everything in the tree at the start |
| 2757 | * of the run (but not newly added entries), or it can be some target | 2757 | * of the run (but not newly added entries), or it can be some target |
| 2758 | * number you'd like to process. | 2758 | * number you'd like to process. |
| 2759 | * | 2759 | * |
| 2760 | * Returns 0 on success or if called with an aborted transaction | 2760 | * Returns 0 on success or if called with an aborted transaction |
| 2761 | * Returns <0 on error and aborts the transaction | 2761 | * Returns <0 on error and aborts the transaction |
| 2762 | */ | 2762 | */ |
| 2763 | int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, | 2763 | int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, |
| 2764 | struct btrfs_root *root, unsigned long count) | 2764 | struct btrfs_root *root, unsigned long count) |
| 2765 | { | 2765 | { |
| 2766 | struct rb_node *node; | 2766 | struct rb_node *node; |
| 2767 | struct btrfs_delayed_ref_root *delayed_refs; | 2767 | struct btrfs_delayed_ref_root *delayed_refs; |
| 2768 | struct btrfs_delayed_ref_head *head; | 2768 | struct btrfs_delayed_ref_head *head; |
| 2769 | int ret; | 2769 | int ret; |
| 2770 | int run_all = count == (unsigned long)-1; | 2770 | int run_all = count == (unsigned long)-1; |
| 2771 | int run_most = 0; | 2771 | int run_most = 0; |
| 2772 | 2772 | ||
| 2773 | /* We'll clean this up in btrfs_cleanup_transaction */ | 2773 | /* We'll clean this up in btrfs_cleanup_transaction */ |
| 2774 | if (trans->aborted) | 2774 | if (trans->aborted) |
| 2775 | return 0; | 2775 | return 0; |
| 2776 | 2776 | ||
| 2777 | if (root == root->fs_info->extent_root) | 2777 | if (root == root->fs_info->extent_root) |
| 2778 | root = root->fs_info->tree_root; | 2778 | root = root->fs_info->tree_root; |
| 2779 | 2779 | ||
| 2780 | delayed_refs = &trans->transaction->delayed_refs; | 2780 | delayed_refs = &trans->transaction->delayed_refs; |
| 2781 | if (count == 0) { | 2781 | if (count == 0) { |
| 2782 | count = atomic_read(&delayed_refs->num_entries) * 2; | 2782 | count = atomic_read(&delayed_refs->num_entries) * 2; |
| 2783 | run_most = 1; | 2783 | run_most = 1; |
| 2784 | } | 2784 | } |
| 2785 | 2785 | ||
| 2786 | again: | 2786 | again: |
| 2787 | #ifdef SCRAMBLE_DELAYED_REFS | 2787 | #ifdef SCRAMBLE_DELAYED_REFS |
| 2788 | delayed_refs->run_delayed_start = find_middle(&delayed_refs->root); | 2788 | delayed_refs->run_delayed_start = find_middle(&delayed_refs->root); |
| 2789 | #endif | 2789 | #endif |
| 2790 | ret = __btrfs_run_delayed_refs(trans, root, count); | 2790 | ret = __btrfs_run_delayed_refs(trans, root, count); |
| 2791 | if (ret < 0) { | 2791 | if (ret < 0) { |
| 2792 | btrfs_abort_transaction(trans, root, ret); | 2792 | btrfs_abort_transaction(trans, root, ret); |
| 2793 | return ret; | 2793 | return ret; |
| 2794 | } | 2794 | } |
| 2795 | 2795 | ||
| 2796 | if (run_all) { | 2796 | if (run_all) { |
| 2797 | if (!list_empty(&trans->new_bgs)) | 2797 | if (!list_empty(&trans->new_bgs)) |
| 2798 | btrfs_create_pending_block_groups(trans, root); | 2798 | btrfs_create_pending_block_groups(trans, root); |
| 2799 | 2799 | ||
| 2800 | spin_lock(&delayed_refs->lock); | 2800 | spin_lock(&delayed_refs->lock); |
| 2801 | node = rb_first(&delayed_refs->href_root); | 2801 | node = rb_first(&delayed_refs->href_root); |
| 2802 | if (!node) { | 2802 | if (!node) { |
| 2803 | spin_unlock(&delayed_refs->lock); | 2803 | spin_unlock(&delayed_refs->lock); |
| 2804 | goto out; | 2804 | goto out; |
| 2805 | } | 2805 | } |
| 2806 | count = (unsigned long)-1; | 2806 | count = (unsigned long)-1; |
| 2807 | 2807 | ||
| 2808 | while (node) { | 2808 | while (node) { |
| 2809 | head = rb_entry(node, struct btrfs_delayed_ref_head, | 2809 | head = rb_entry(node, struct btrfs_delayed_ref_head, |
| 2810 | href_node); | 2810 | href_node); |
| 2811 | if (btrfs_delayed_ref_is_head(&head->node)) { | 2811 | if (btrfs_delayed_ref_is_head(&head->node)) { |
| 2812 | struct btrfs_delayed_ref_node *ref; | 2812 | struct btrfs_delayed_ref_node *ref; |
| 2813 | 2813 | ||
| 2814 | ref = &head->node; | 2814 | ref = &head->node; |
| 2815 | atomic_inc(&ref->refs); | 2815 | atomic_inc(&ref->refs); |
| 2816 | 2816 | ||
| 2817 | spin_unlock(&delayed_refs->lock); | 2817 | spin_unlock(&delayed_refs->lock); |
| 2818 | /* | 2818 | /* |
| 2819 | * Mutex was contended, block until it's | 2819 | * Mutex was contended, block until it's |
| 2820 | * released and try again | 2820 | * released and try again |
| 2821 | */ | 2821 | */ |
| 2822 | mutex_lock(&head->mutex); | 2822 | mutex_lock(&head->mutex); |
| 2823 | mutex_unlock(&head->mutex); | 2823 | mutex_unlock(&head->mutex); |
| 2824 | 2824 | ||
| 2825 | btrfs_put_delayed_ref(ref); | 2825 | btrfs_put_delayed_ref(ref); |
| 2826 | cond_resched(); | 2826 | cond_resched(); |
| 2827 | goto again; | 2827 | goto again; |
| 2828 | } else { | 2828 | } else { |
| 2829 | WARN_ON(1); | 2829 | WARN_ON(1); |
| 2830 | } | 2830 | } |
| 2831 | node = rb_next(node); | 2831 | node = rb_next(node); |
| 2832 | } | 2832 | } |
| 2833 | spin_unlock(&delayed_refs->lock); | 2833 | spin_unlock(&delayed_refs->lock); |
| 2834 | cond_resched(); | 2834 | cond_resched(); |
| 2835 | goto again; | 2835 | goto again; |
| 2836 | } | 2836 | } |
| 2837 | out: | 2837 | out: |
| 2838 | ret = btrfs_delayed_qgroup_accounting(trans, root->fs_info); | 2838 | ret = btrfs_delayed_qgroup_accounting(trans, root->fs_info); |
| 2839 | if (ret) | 2839 | if (ret) |
| 2840 | return ret; | 2840 | return ret; |
| 2841 | assert_qgroups_uptodate(trans); | 2841 | assert_qgroups_uptodate(trans); |
| 2842 | return 0; | 2842 | return 0; |
| 2843 | } | 2843 | } |
| 2844 | 2844 | ||
| 2845 | int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans, | 2845 | int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans, |
| 2846 | struct btrfs_root *root, | 2846 | struct btrfs_root *root, |
| 2847 | u64 bytenr, u64 num_bytes, u64 flags, | 2847 | u64 bytenr, u64 num_bytes, u64 flags, |
| 2848 | int level, int is_data) | 2848 | int level, int is_data) |
| 2849 | { | 2849 | { |
| 2850 | struct btrfs_delayed_extent_op *extent_op; | 2850 | struct btrfs_delayed_extent_op *extent_op; |
| 2851 | int ret; | 2851 | int ret; |
| 2852 | 2852 | ||
| 2853 | extent_op = btrfs_alloc_delayed_extent_op(); | 2853 | extent_op = btrfs_alloc_delayed_extent_op(); |
| 2854 | if (!extent_op) | 2854 | if (!extent_op) |
| 2855 | return -ENOMEM; | 2855 | return -ENOMEM; |
| 2856 | 2856 | ||
| 2857 | extent_op->flags_to_set = flags; | 2857 | extent_op->flags_to_set = flags; |
| 2858 | extent_op->update_flags = 1; | 2858 | extent_op->update_flags = 1; |
| 2859 | extent_op->update_key = 0; | 2859 | extent_op->update_key = 0; |
| 2860 | extent_op->is_data = is_data ? 1 : 0; | 2860 | extent_op->is_data = is_data ? 1 : 0; |
| 2861 | extent_op->level = level; | 2861 | extent_op->level = level; |
| 2862 | 2862 | ||
| 2863 | ret = btrfs_add_delayed_extent_op(root->fs_info, trans, bytenr, | 2863 | ret = btrfs_add_delayed_extent_op(root->fs_info, trans, bytenr, |
| 2864 | num_bytes, extent_op); | 2864 | num_bytes, extent_op); |
| 2865 | if (ret) | 2865 | if (ret) |
| 2866 | btrfs_free_delayed_extent_op(extent_op); | 2866 | btrfs_free_delayed_extent_op(extent_op); |
| 2867 | return ret; | 2867 | return ret; |
| 2868 | } | 2868 | } |
| 2869 | 2869 | ||
| 2870 | static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, | 2870 | static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, |
| 2871 | struct btrfs_root *root, | 2871 | struct btrfs_root *root, |
| 2872 | struct btrfs_path *path, | 2872 | struct btrfs_path *path, |
| 2873 | u64 objectid, u64 offset, u64 bytenr) | 2873 | u64 objectid, u64 offset, u64 bytenr) |
| 2874 | { | 2874 | { |
| 2875 | struct btrfs_delayed_ref_head *head; | 2875 | struct btrfs_delayed_ref_head *head; |
| 2876 | struct btrfs_delayed_ref_node *ref; | 2876 | struct btrfs_delayed_ref_node *ref; |
| 2877 | struct btrfs_delayed_data_ref *data_ref; | 2877 | struct btrfs_delayed_data_ref *data_ref; |
| 2878 | struct btrfs_delayed_ref_root *delayed_refs; | 2878 | struct btrfs_delayed_ref_root *delayed_refs; |
| 2879 | struct rb_node *node; | 2879 | struct rb_node *node; |
| 2880 | int ret = 0; | 2880 | int ret = 0; |
| 2881 | 2881 | ||
| 2882 | delayed_refs = &trans->transaction->delayed_refs; | 2882 | delayed_refs = &trans->transaction->delayed_refs; |
| 2883 | spin_lock(&delayed_refs->lock); | 2883 | spin_lock(&delayed_refs->lock); |
| 2884 | head = btrfs_find_delayed_ref_head(trans, bytenr); | 2884 | head = btrfs_find_delayed_ref_head(trans, bytenr); |
| 2885 | if (!head) { | 2885 | if (!head) { |
| 2886 | spin_unlock(&delayed_refs->lock); | 2886 | spin_unlock(&delayed_refs->lock); |
| 2887 | return 0; | 2887 | return 0; |
| 2888 | } | 2888 | } |
| 2889 | 2889 | ||
| 2890 | if (!mutex_trylock(&head->mutex)) { | 2890 | if (!mutex_trylock(&head->mutex)) { |
| 2891 | atomic_inc(&head->node.refs); | 2891 | atomic_inc(&head->node.refs); |
| 2892 | spin_unlock(&delayed_refs->lock); | 2892 | spin_unlock(&delayed_refs->lock); |
| 2893 | 2893 | ||
| 2894 | btrfs_release_path(path); | 2894 | btrfs_release_path(path); |
| 2895 | 2895 | ||
| 2896 | /* | 2896 | /* |
| 2897 | * Mutex was contended, block until it's released and let | 2897 | * Mutex was contended, block until it's released and let |
| 2898 | * caller try again | 2898 | * caller try again |
| 2899 | */ | 2899 | */ |
| 2900 | mutex_lock(&head->mutex); | 2900 | mutex_lock(&head->mutex); |
| 2901 | mutex_unlock(&head->mutex); | 2901 | mutex_unlock(&head->mutex); |
| 2902 | btrfs_put_delayed_ref(&head->node); | 2902 | btrfs_put_delayed_ref(&head->node); |
| 2903 | return -EAGAIN; | 2903 | return -EAGAIN; |
| 2904 | } | 2904 | } |
| 2905 | spin_unlock(&delayed_refs->lock); | 2905 | spin_unlock(&delayed_refs->lock); |
| 2906 | 2906 | ||
| 2907 | spin_lock(&head->lock); | 2907 | spin_lock(&head->lock); |
| 2908 | node = rb_first(&head->ref_root); | 2908 | node = rb_first(&head->ref_root); |
| 2909 | while (node) { | 2909 | while (node) { |
| 2910 | ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node); | 2910 | ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node); |
| 2911 | node = rb_next(node); | 2911 | node = rb_next(node); |
| 2912 | 2912 | ||
| 2913 | /* If it's a shared ref we know a cross reference exists */ | 2913 | /* If it's a shared ref we know a cross reference exists */ |
| 2914 | if (ref->type != BTRFS_EXTENT_DATA_REF_KEY) { | 2914 | if (ref->type != BTRFS_EXTENT_DATA_REF_KEY) { |
| 2915 | ret = 1; | 2915 | ret = 1; |
| 2916 | break; | 2916 | break; |
| 2917 | } | 2917 | } |
| 2918 | 2918 | ||
| 2919 | data_ref = btrfs_delayed_node_to_data_ref(ref); | 2919 | data_ref = btrfs_delayed_node_to_data_ref(ref); |
| 2920 | 2920 | ||
| 2921 | /* | 2921 | /* |
| 2922 | * If our ref doesn't match the one we're currently looking at | 2922 | * If our ref doesn't match the one we're currently looking at |
| 2923 | * then we have a cross reference. | 2923 | * then we have a cross reference. |
| 2924 | */ | 2924 | */ |
| 2925 | if (data_ref->root != root->root_key.objectid || | 2925 | if (data_ref->root != root->root_key.objectid || |
| 2926 | data_ref->objectid != objectid || | 2926 | data_ref->objectid != objectid || |
| 2927 | data_ref->offset != offset) { | 2927 | data_ref->offset != offset) { |
| 2928 | ret = 1; | 2928 | ret = 1; |
| 2929 | break; | 2929 | break; |
| 2930 | } | 2930 | } |
| 2931 | } | 2931 | } |
| 2932 | spin_unlock(&head->lock); | 2932 | spin_unlock(&head->lock); |
| 2933 | mutex_unlock(&head->mutex); | 2933 | mutex_unlock(&head->mutex); |
| 2934 | return ret; | 2934 | return ret; |
| 2935 | } | 2935 | } |
| 2936 | 2936 | ||
| 2937 | static noinline int check_committed_ref(struct btrfs_trans_handle *trans, | 2937 | static noinline int check_committed_ref(struct btrfs_trans_handle *trans, |
| 2938 | struct btrfs_root *root, | 2938 | struct btrfs_root *root, |
| 2939 | struct btrfs_path *path, | 2939 | struct btrfs_path *path, |
| 2940 | u64 objectid, u64 offset, u64 bytenr) | 2940 | u64 objectid, u64 offset, u64 bytenr) |
| 2941 | { | 2941 | { |
| 2942 | struct btrfs_root *extent_root = root->fs_info->extent_root; | 2942 | struct btrfs_root *extent_root = root->fs_info->extent_root; |
| 2943 | struct extent_buffer *leaf; | 2943 | struct extent_buffer *leaf; |
| 2944 | struct btrfs_extent_data_ref *ref; | 2944 | struct btrfs_extent_data_ref *ref; |
| 2945 | struct btrfs_extent_inline_ref *iref; | 2945 | struct btrfs_extent_inline_ref *iref; |
| 2946 | struct btrfs_extent_item *ei; | 2946 | struct btrfs_extent_item *ei; |
| 2947 | struct btrfs_key key; | 2947 | struct btrfs_key key; |
| 2948 | u32 item_size; | 2948 | u32 item_size; |
| 2949 | int ret; | 2949 | int ret; |
| 2950 | 2950 | ||
| 2951 | key.objectid = bytenr; | 2951 | key.objectid = bytenr; |
| 2952 | key.offset = (u64)-1; | 2952 | key.offset = (u64)-1; |
| 2953 | key.type = BTRFS_EXTENT_ITEM_KEY; | 2953 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 2954 | 2954 | ||
| 2955 | ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0); | 2955 | ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0); |
| 2956 | if (ret < 0) | 2956 | if (ret < 0) |
| 2957 | goto out; | 2957 | goto out; |
| 2958 | BUG_ON(ret == 0); /* Corruption */ | 2958 | BUG_ON(ret == 0); /* Corruption */ |
| 2959 | 2959 | ||
| 2960 | ret = -ENOENT; | 2960 | ret = -ENOENT; |
| 2961 | if (path->slots[0] == 0) | 2961 | if (path->slots[0] == 0) |
| 2962 | goto out; | 2962 | goto out; |
| 2963 | 2963 | ||
| 2964 | path->slots[0]--; | 2964 | path->slots[0]--; |
| 2965 | leaf = path->nodes[0]; | 2965 | leaf = path->nodes[0]; |
| 2966 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 2966 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 2967 | 2967 | ||
| 2968 | if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY) | 2968 | if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY) |
| 2969 | goto out; | 2969 | goto out; |
| 2970 | 2970 | ||
| 2971 | ret = 1; | 2971 | ret = 1; |
| 2972 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | 2972 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); |
| 2973 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 2973 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 2974 | if (item_size < sizeof(*ei)) { | 2974 | if (item_size < sizeof(*ei)) { |
| 2975 | WARN_ON(item_size != sizeof(struct btrfs_extent_item_v0)); | 2975 | WARN_ON(item_size != sizeof(struct btrfs_extent_item_v0)); |
| 2976 | goto out; | 2976 | goto out; |
| 2977 | } | 2977 | } |
| 2978 | #endif | 2978 | #endif |
| 2979 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); | 2979 | ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); |
| 2980 | 2980 | ||
| 2981 | if (item_size != sizeof(*ei) + | 2981 | if (item_size != sizeof(*ei) + |
| 2982 | btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY)) | 2982 | btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY)) |
| 2983 | goto out; | 2983 | goto out; |
| 2984 | 2984 | ||
| 2985 | if (btrfs_extent_generation(leaf, ei) <= | 2985 | if (btrfs_extent_generation(leaf, ei) <= |
| 2986 | btrfs_root_last_snapshot(&root->root_item)) | 2986 | btrfs_root_last_snapshot(&root->root_item)) |
| 2987 | goto out; | 2987 | goto out; |
| 2988 | 2988 | ||
| 2989 | iref = (struct btrfs_extent_inline_ref *)(ei + 1); | 2989 | iref = (struct btrfs_extent_inline_ref *)(ei + 1); |
| 2990 | if (btrfs_extent_inline_ref_type(leaf, iref) != | 2990 | if (btrfs_extent_inline_ref_type(leaf, iref) != |
| 2991 | BTRFS_EXTENT_DATA_REF_KEY) | 2991 | BTRFS_EXTENT_DATA_REF_KEY) |
| 2992 | goto out; | 2992 | goto out; |
| 2993 | 2993 | ||
| 2994 | ref = (struct btrfs_extent_data_ref *)(&iref->offset); | 2994 | ref = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 2995 | if (btrfs_extent_refs(leaf, ei) != | 2995 | if (btrfs_extent_refs(leaf, ei) != |
| 2996 | btrfs_extent_data_ref_count(leaf, ref) || | 2996 | btrfs_extent_data_ref_count(leaf, ref) || |
| 2997 | btrfs_extent_data_ref_root(leaf, ref) != | 2997 | btrfs_extent_data_ref_root(leaf, ref) != |
| 2998 | root->root_key.objectid || | 2998 | root->root_key.objectid || |
| 2999 | btrfs_extent_data_ref_objectid(leaf, ref) != objectid || | 2999 | btrfs_extent_data_ref_objectid(leaf, ref) != objectid || |
| 3000 | btrfs_extent_data_ref_offset(leaf, ref) != offset) | 3000 | btrfs_extent_data_ref_offset(leaf, ref) != offset) |
| 3001 | goto out; | 3001 | goto out; |
| 3002 | 3002 | ||
| 3003 | ret = 0; | 3003 | ret = 0; |
| 3004 | out: | 3004 | out: |
| 3005 | return ret; | 3005 | return ret; |
| 3006 | } | 3006 | } |
| 3007 | 3007 | ||
| 3008 | int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, | 3008 | int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, |
| 3009 | struct btrfs_root *root, | 3009 | struct btrfs_root *root, |
| 3010 | u64 objectid, u64 offset, u64 bytenr) | 3010 | u64 objectid, u64 offset, u64 bytenr) |
| 3011 | { | 3011 | { |
| 3012 | struct btrfs_path *path; | 3012 | struct btrfs_path *path; |
| 3013 | int ret; | 3013 | int ret; |
| 3014 | int ret2; | 3014 | int ret2; |
| 3015 | 3015 | ||
| 3016 | path = btrfs_alloc_path(); | 3016 | path = btrfs_alloc_path(); |
| 3017 | if (!path) | 3017 | if (!path) |
| 3018 | return -ENOENT; | 3018 | return -ENOENT; |
| 3019 | 3019 | ||
| 3020 | do { | 3020 | do { |
| 3021 | ret = check_committed_ref(trans, root, path, objectid, | 3021 | ret = check_committed_ref(trans, root, path, objectid, |
| 3022 | offset, bytenr); | 3022 | offset, bytenr); |
| 3023 | if (ret && ret != -ENOENT) | 3023 | if (ret && ret != -ENOENT) |
| 3024 | goto out; | 3024 | goto out; |
| 3025 | 3025 | ||
| 3026 | ret2 = check_delayed_ref(trans, root, path, objectid, | 3026 | ret2 = check_delayed_ref(trans, root, path, objectid, |
| 3027 | offset, bytenr); | 3027 | offset, bytenr); |
| 3028 | } while (ret2 == -EAGAIN); | 3028 | } while (ret2 == -EAGAIN); |
| 3029 | 3029 | ||
| 3030 | if (ret2 && ret2 != -ENOENT) { | 3030 | if (ret2 && ret2 != -ENOENT) { |
| 3031 | ret = ret2; | 3031 | ret = ret2; |
| 3032 | goto out; | 3032 | goto out; |
| 3033 | } | 3033 | } |
| 3034 | 3034 | ||
| 3035 | if (ret != -ENOENT || ret2 != -ENOENT) | 3035 | if (ret != -ENOENT || ret2 != -ENOENT) |
| 3036 | ret = 0; | 3036 | ret = 0; |
| 3037 | out: | 3037 | out: |
| 3038 | btrfs_free_path(path); | 3038 | btrfs_free_path(path); |
| 3039 | if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) | 3039 | if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) |
| 3040 | WARN_ON(ret > 0); | 3040 | WARN_ON(ret > 0); |
| 3041 | return ret; | 3041 | return ret; |
| 3042 | } | 3042 | } |
| 3043 | 3043 | ||
| 3044 | static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, | 3044 | static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, |
| 3045 | struct btrfs_root *root, | 3045 | struct btrfs_root *root, |
| 3046 | struct extent_buffer *buf, | 3046 | struct extent_buffer *buf, |
| 3047 | int full_backref, int inc) | 3047 | int full_backref, int inc) |
| 3048 | { | 3048 | { |
| 3049 | u64 bytenr; | 3049 | u64 bytenr; |
| 3050 | u64 num_bytes; | 3050 | u64 num_bytes; |
| 3051 | u64 parent; | 3051 | u64 parent; |
| 3052 | u64 ref_root; | 3052 | u64 ref_root; |
| 3053 | u32 nritems; | 3053 | u32 nritems; |
| 3054 | struct btrfs_key key; | 3054 | struct btrfs_key key; |
| 3055 | struct btrfs_file_extent_item *fi; | 3055 | struct btrfs_file_extent_item *fi; |
| 3056 | int i; | 3056 | int i; |
| 3057 | int level; | 3057 | int level; |
| 3058 | int ret = 0; | 3058 | int ret = 0; |
| 3059 | int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *, | 3059 | int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *, |
| 3060 | u64, u64, u64, u64, u64, u64, int); | 3060 | u64, u64, u64, u64, u64, u64, int); |
| 3061 | 3061 | ||
| 3062 | 3062 | ||
| 3063 | if (btrfs_test_is_dummy_root(root)) | 3063 | if (btrfs_test_is_dummy_root(root)) |
| 3064 | return 0; | 3064 | return 0; |
| 3065 | 3065 | ||
| 3066 | ref_root = btrfs_header_owner(buf); | 3066 | ref_root = btrfs_header_owner(buf); |
| 3067 | nritems = btrfs_header_nritems(buf); | 3067 | nritems = btrfs_header_nritems(buf); |
| 3068 | level = btrfs_header_level(buf); | 3068 | level = btrfs_header_level(buf); |
| 3069 | 3069 | ||
| 3070 | if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state) && level == 0) | 3070 | if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state) && level == 0) |
| 3071 | return 0; | 3071 | return 0; |
| 3072 | 3072 | ||
| 3073 | if (inc) | 3073 | if (inc) |
| 3074 | process_func = btrfs_inc_extent_ref; | 3074 | process_func = btrfs_inc_extent_ref; |
| 3075 | else | 3075 | else |
| 3076 | process_func = btrfs_free_extent; | 3076 | process_func = btrfs_free_extent; |
| 3077 | 3077 | ||
| 3078 | if (full_backref) | 3078 | if (full_backref) |
| 3079 | parent = buf->start; | 3079 | parent = buf->start; |
| 3080 | else | 3080 | else |
| 3081 | parent = 0; | 3081 | parent = 0; |
| 3082 | 3082 | ||
| 3083 | for (i = 0; i < nritems; i++) { | 3083 | for (i = 0; i < nritems; i++) { |
| 3084 | if (level == 0) { | 3084 | if (level == 0) { |
| 3085 | btrfs_item_key_to_cpu(buf, &key, i); | 3085 | btrfs_item_key_to_cpu(buf, &key, i); |
| 3086 | if (key.type != BTRFS_EXTENT_DATA_KEY) | 3086 | if (key.type != BTRFS_EXTENT_DATA_KEY) |
| 3087 | continue; | 3087 | continue; |
| 3088 | fi = btrfs_item_ptr(buf, i, | 3088 | fi = btrfs_item_ptr(buf, i, |
| 3089 | struct btrfs_file_extent_item); | 3089 | struct btrfs_file_extent_item); |
| 3090 | if (btrfs_file_extent_type(buf, fi) == | 3090 | if (btrfs_file_extent_type(buf, fi) == |
| 3091 | BTRFS_FILE_EXTENT_INLINE) | 3091 | BTRFS_FILE_EXTENT_INLINE) |
| 3092 | continue; | 3092 | continue; |
| 3093 | bytenr = btrfs_file_extent_disk_bytenr(buf, fi); | 3093 | bytenr = btrfs_file_extent_disk_bytenr(buf, fi); |
| 3094 | if (bytenr == 0) | 3094 | if (bytenr == 0) |
| 3095 | continue; | 3095 | continue; |
| 3096 | 3096 | ||
| 3097 | num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi); | 3097 | num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi); |
| 3098 | key.offset -= btrfs_file_extent_offset(buf, fi); | 3098 | key.offset -= btrfs_file_extent_offset(buf, fi); |
| 3099 | ret = process_func(trans, root, bytenr, num_bytes, | 3099 | ret = process_func(trans, root, bytenr, num_bytes, |
| 3100 | parent, ref_root, key.objectid, | 3100 | parent, ref_root, key.objectid, |
| 3101 | key.offset, 1); | 3101 | key.offset, 1); |
| 3102 | if (ret) | 3102 | if (ret) |
| 3103 | goto fail; | 3103 | goto fail; |
| 3104 | } else { | 3104 | } else { |
| 3105 | bytenr = btrfs_node_blockptr(buf, i); | 3105 | bytenr = btrfs_node_blockptr(buf, i); |
| 3106 | num_bytes = root->nodesize; | 3106 | num_bytes = root->nodesize; |
| 3107 | ret = process_func(trans, root, bytenr, num_bytes, | 3107 | ret = process_func(trans, root, bytenr, num_bytes, |
| 3108 | parent, ref_root, level - 1, 0, | 3108 | parent, ref_root, level - 1, 0, |
| 3109 | 1); | 3109 | 1); |
| 3110 | if (ret) | 3110 | if (ret) |
| 3111 | goto fail; | 3111 | goto fail; |
| 3112 | } | 3112 | } |
| 3113 | } | 3113 | } |
| 3114 | return 0; | 3114 | return 0; |
| 3115 | fail: | 3115 | fail: |
| 3116 | return ret; | 3116 | return ret; |
| 3117 | } | 3117 | } |
| 3118 | 3118 | ||
| 3119 | int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, | 3119 | int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
| 3120 | struct extent_buffer *buf, int full_backref) | 3120 | struct extent_buffer *buf, int full_backref) |
| 3121 | { | 3121 | { |
| 3122 | return __btrfs_mod_ref(trans, root, buf, full_backref, 1); | 3122 | return __btrfs_mod_ref(trans, root, buf, full_backref, 1); |
| 3123 | } | 3123 | } |
| 3124 | 3124 | ||
| 3125 | int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, | 3125 | int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
| 3126 | struct extent_buffer *buf, int full_backref) | 3126 | struct extent_buffer *buf, int full_backref) |
| 3127 | { | 3127 | { |
| 3128 | return __btrfs_mod_ref(trans, root, buf, full_backref, 0); | 3128 | return __btrfs_mod_ref(trans, root, buf, full_backref, 0); |
| 3129 | } | 3129 | } |
| 3130 | 3130 | ||
| 3131 | static int write_one_cache_group(struct btrfs_trans_handle *trans, | 3131 | static int write_one_cache_group(struct btrfs_trans_handle *trans, |
| 3132 | struct btrfs_root *root, | 3132 | struct btrfs_root *root, |
| 3133 | struct btrfs_path *path, | 3133 | struct btrfs_path *path, |
| 3134 | struct btrfs_block_group_cache *cache) | 3134 | struct btrfs_block_group_cache *cache) |
| 3135 | { | 3135 | { |
| 3136 | int ret; | 3136 | int ret; |
| 3137 | struct btrfs_root *extent_root = root->fs_info->extent_root; | 3137 | struct btrfs_root *extent_root = root->fs_info->extent_root; |
| 3138 | unsigned long bi; | 3138 | unsigned long bi; |
| 3139 | struct extent_buffer *leaf; | 3139 | struct extent_buffer *leaf; |
| 3140 | 3140 | ||
| 3141 | ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1); | 3141 | ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1); |
| 3142 | if (ret < 0) | 3142 | if (ret) { |
| 3143 | if (ret > 0) | ||
| 3144 | ret = -ENOENT; | ||
| 3143 | goto fail; | 3145 | goto fail; |
| 3144 | BUG_ON(ret); /* Corruption */ | 3146 | } |
| 3145 | 3147 | ||
| 3146 | leaf = path->nodes[0]; | 3148 | leaf = path->nodes[0]; |
| 3147 | bi = btrfs_item_ptr_offset(leaf, path->slots[0]); | 3149 | bi = btrfs_item_ptr_offset(leaf, path->slots[0]); |
| 3148 | write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item)); | 3150 | write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item)); |
| 3149 | btrfs_mark_buffer_dirty(leaf); | 3151 | btrfs_mark_buffer_dirty(leaf); |
| 3150 | btrfs_release_path(path); | 3152 | btrfs_release_path(path); |
| 3151 | fail: | 3153 | fail: |
| 3152 | if (ret) { | 3154 | if (ret) |
| 3153 | btrfs_abort_transaction(trans, root, ret); | 3155 | btrfs_abort_transaction(trans, root, ret); |
| 3154 | return ret; | 3156 | return ret; |
| 3155 | } | ||
| 3156 | return 0; | ||
| 3157 | 3157 | ||
| 3158 | } | 3158 | } |
| 3159 | 3159 | ||
| 3160 | static struct btrfs_block_group_cache * | 3160 | static struct btrfs_block_group_cache * |
| 3161 | next_block_group(struct btrfs_root *root, | 3161 | next_block_group(struct btrfs_root *root, |
| 3162 | struct btrfs_block_group_cache *cache) | 3162 | struct btrfs_block_group_cache *cache) |
| 3163 | { | 3163 | { |
| 3164 | struct rb_node *node; | 3164 | struct rb_node *node; |
| 3165 | 3165 | ||
| 3166 | spin_lock(&root->fs_info->block_group_cache_lock); | 3166 | spin_lock(&root->fs_info->block_group_cache_lock); |
| 3167 | 3167 | ||
| 3168 | /* If our block group was removed, we need a full search. */ | 3168 | /* If our block group was removed, we need a full search. */ |
| 3169 | if (RB_EMPTY_NODE(&cache->cache_node)) { | 3169 | if (RB_EMPTY_NODE(&cache->cache_node)) { |
| 3170 | const u64 next_bytenr = cache->key.objectid + cache->key.offset; | 3170 | const u64 next_bytenr = cache->key.objectid + cache->key.offset; |
| 3171 | 3171 | ||
| 3172 | spin_unlock(&root->fs_info->block_group_cache_lock); | 3172 | spin_unlock(&root->fs_info->block_group_cache_lock); |
| 3173 | btrfs_put_block_group(cache); | 3173 | btrfs_put_block_group(cache); |
| 3174 | cache = btrfs_lookup_first_block_group(root->fs_info, | 3174 | cache = btrfs_lookup_first_block_group(root->fs_info, |
| 3175 | next_bytenr); | 3175 | next_bytenr); |
| 3176 | return cache; | 3176 | return cache; |
| 3177 | } | 3177 | } |
| 3178 | node = rb_next(&cache->cache_node); | 3178 | node = rb_next(&cache->cache_node); |
| 3179 | btrfs_put_block_group(cache); | 3179 | btrfs_put_block_group(cache); |
| 3180 | if (node) { | 3180 | if (node) { |
| 3181 | cache = rb_entry(node, struct btrfs_block_group_cache, | 3181 | cache = rb_entry(node, struct btrfs_block_group_cache, |
| 3182 | cache_node); | 3182 | cache_node); |
| 3183 | btrfs_get_block_group(cache); | 3183 | btrfs_get_block_group(cache); |
| 3184 | } else | 3184 | } else |
| 3185 | cache = NULL; | 3185 | cache = NULL; |
| 3186 | spin_unlock(&root->fs_info->block_group_cache_lock); | 3186 | spin_unlock(&root->fs_info->block_group_cache_lock); |
| 3187 | return cache; | 3187 | return cache; |
| 3188 | } | 3188 | } |
| 3189 | 3189 | ||
| 3190 | static int cache_save_setup(struct btrfs_block_group_cache *block_group, | 3190 | static int cache_save_setup(struct btrfs_block_group_cache *block_group, |
| 3191 | struct btrfs_trans_handle *trans, | 3191 | struct btrfs_trans_handle *trans, |
| 3192 | struct btrfs_path *path) | 3192 | struct btrfs_path *path) |
| 3193 | { | 3193 | { |
| 3194 | struct btrfs_root *root = block_group->fs_info->tree_root; | 3194 | struct btrfs_root *root = block_group->fs_info->tree_root; |
| 3195 | struct inode *inode = NULL; | 3195 | struct inode *inode = NULL; |
| 3196 | u64 alloc_hint = 0; | 3196 | u64 alloc_hint = 0; |
| 3197 | int dcs = BTRFS_DC_ERROR; | 3197 | int dcs = BTRFS_DC_ERROR; |
| 3198 | int num_pages = 0; | 3198 | int num_pages = 0; |
| 3199 | int retries = 0; | 3199 | int retries = 0; |
| 3200 | int ret = 0; | 3200 | int ret = 0; |
| 3201 | 3201 | ||
| 3202 | /* | 3202 | /* |
| 3203 | * If this block group is smaller than 100 megs don't bother caching the | 3203 | * If this block group is smaller than 100 megs don't bother caching the |
| 3204 | * block group. | 3204 | * block group. |
| 3205 | */ | 3205 | */ |
| 3206 | if (block_group->key.offset < (100 * 1024 * 1024)) { | 3206 | if (block_group->key.offset < (100 * 1024 * 1024)) { |
| 3207 | spin_lock(&block_group->lock); | 3207 | spin_lock(&block_group->lock); |
| 3208 | block_group->disk_cache_state = BTRFS_DC_WRITTEN; | 3208 | block_group->disk_cache_state = BTRFS_DC_WRITTEN; |
| 3209 | spin_unlock(&block_group->lock); | 3209 | spin_unlock(&block_group->lock); |
| 3210 | return 0; | 3210 | return 0; |
| 3211 | } | 3211 | } |
| 3212 | 3212 | ||
| 3213 | again: | 3213 | again: |
| 3214 | inode = lookup_free_space_inode(root, block_group, path); | 3214 | inode = lookup_free_space_inode(root, block_group, path); |
| 3215 | if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { | 3215 | if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { |
| 3216 | ret = PTR_ERR(inode); | 3216 | ret = PTR_ERR(inode); |
| 3217 | btrfs_release_path(path); | 3217 | btrfs_release_path(path); |
| 3218 | goto out; | 3218 | goto out; |
| 3219 | } | 3219 | } |
| 3220 | 3220 | ||
| 3221 | if (IS_ERR(inode)) { | 3221 | if (IS_ERR(inode)) { |
| 3222 | BUG_ON(retries); | 3222 | BUG_ON(retries); |
| 3223 | retries++; | 3223 | retries++; |
| 3224 | 3224 | ||
| 3225 | if (block_group->ro) | 3225 | if (block_group->ro) |
| 3226 | goto out_free; | 3226 | goto out_free; |
| 3227 | 3227 | ||
| 3228 | ret = create_free_space_inode(root, trans, block_group, path); | 3228 | ret = create_free_space_inode(root, trans, block_group, path); |
| 3229 | if (ret) | 3229 | if (ret) |
| 3230 | goto out_free; | 3230 | goto out_free; |
| 3231 | goto again; | 3231 | goto again; |
| 3232 | } | 3232 | } |
| 3233 | 3233 | ||
| 3234 | /* We've already setup this transaction, go ahead and exit */ | 3234 | /* We've already setup this transaction, go ahead and exit */ |
| 3235 | if (block_group->cache_generation == trans->transid && | 3235 | if (block_group->cache_generation == trans->transid && |
| 3236 | i_size_read(inode)) { | 3236 | i_size_read(inode)) { |
| 3237 | dcs = BTRFS_DC_SETUP; | 3237 | dcs = BTRFS_DC_SETUP; |
| 3238 | goto out_put; | 3238 | goto out_put; |
| 3239 | } | 3239 | } |
| 3240 | 3240 | ||
| 3241 | /* | 3241 | /* |
| 3242 | * We want to set the generation to 0, that way if anything goes wrong | 3242 | * We want to set the generation to 0, that way if anything goes wrong |
| 3243 | * from here on out we know not to trust this cache when we load up next | 3243 | * from here on out we know not to trust this cache when we load up next |
| 3244 | * time. | 3244 | * time. |
| 3245 | */ | 3245 | */ |
| 3246 | BTRFS_I(inode)->generation = 0; | 3246 | BTRFS_I(inode)->generation = 0; |
| 3247 | ret = btrfs_update_inode(trans, root, inode); | 3247 | ret = btrfs_update_inode(trans, root, inode); |
| 3248 | WARN_ON(ret); | 3248 | WARN_ON(ret); |
| 3249 | 3249 | ||
| 3250 | if (i_size_read(inode) > 0) { | 3250 | if (i_size_read(inode) > 0) { |
| 3251 | ret = btrfs_check_trunc_cache_free_space(root, | 3251 | ret = btrfs_check_trunc_cache_free_space(root, |
| 3252 | &root->fs_info->global_block_rsv); | 3252 | &root->fs_info->global_block_rsv); |
| 3253 | if (ret) | 3253 | if (ret) |
| 3254 | goto out_put; | 3254 | goto out_put; |
| 3255 | 3255 | ||
| 3256 | ret = btrfs_truncate_free_space_cache(root, trans, inode); | 3256 | ret = btrfs_truncate_free_space_cache(root, trans, inode); |
| 3257 | if (ret) | 3257 | if (ret) |
| 3258 | goto out_put; | 3258 | goto out_put; |
| 3259 | } | 3259 | } |
| 3260 | 3260 | ||
| 3261 | spin_lock(&block_group->lock); | 3261 | spin_lock(&block_group->lock); |
| 3262 | if (block_group->cached != BTRFS_CACHE_FINISHED || | 3262 | if (block_group->cached != BTRFS_CACHE_FINISHED || |
| 3263 | !btrfs_test_opt(root, SPACE_CACHE) || | 3263 | !btrfs_test_opt(root, SPACE_CACHE) || |
| 3264 | block_group->delalloc_bytes) { | 3264 | block_group->delalloc_bytes) { |
| 3265 | /* | 3265 | /* |
| 3266 | * don't bother trying to write stuff out _if_ | 3266 | * don't bother trying to write stuff out _if_ |
| 3267 | * a) we're not cached, | 3267 | * a) we're not cached, |
| 3268 | * b) we're with nospace_cache mount option. | 3268 | * b) we're with nospace_cache mount option. |
| 3269 | */ | 3269 | */ |
| 3270 | dcs = BTRFS_DC_WRITTEN; | 3270 | dcs = BTRFS_DC_WRITTEN; |
| 3271 | spin_unlock(&block_group->lock); | 3271 | spin_unlock(&block_group->lock); |
| 3272 | goto out_put; | 3272 | goto out_put; |
| 3273 | } | 3273 | } |
| 3274 | spin_unlock(&block_group->lock); | 3274 | spin_unlock(&block_group->lock); |
| 3275 | 3275 | ||
| 3276 | /* | 3276 | /* |
| 3277 | * Try to preallocate enough space based on how big the block group is. | 3277 | * Try to preallocate enough space based on how big the block group is. |
| 3278 | * Keep in mind this has to include any pinned space which could end up | 3278 | * Keep in mind this has to include any pinned space which could end up |
| 3279 | * taking up quite a bit since it's not folded into the other space | 3279 | * taking up quite a bit since it's not folded into the other space |
| 3280 | * cache. | 3280 | * cache. |
| 3281 | */ | 3281 | */ |
| 3282 | num_pages = (int)div64_u64(block_group->key.offset, 256 * 1024 * 1024); | 3282 | num_pages = (int)div64_u64(block_group->key.offset, 256 * 1024 * 1024); |
| 3283 | if (!num_pages) | 3283 | if (!num_pages) |
| 3284 | num_pages = 1; | 3284 | num_pages = 1; |
| 3285 | 3285 | ||
| 3286 | num_pages *= 16; | 3286 | num_pages *= 16; |
| 3287 | num_pages *= PAGE_CACHE_SIZE; | 3287 | num_pages *= PAGE_CACHE_SIZE; |
| 3288 | 3288 | ||
| 3289 | ret = btrfs_check_data_free_space(inode, num_pages); | 3289 | ret = btrfs_check_data_free_space(inode, num_pages); |
| 3290 | if (ret) | 3290 | if (ret) |
| 3291 | goto out_put; | 3291 | goto out_put; |
| 3292 | 3292 | ||
| 3293 | ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, num_pages, | 3293 | ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, num_pages, |
| 3294 | num_pages, num_pages, | 3294 | num_pages, num_pages, |
| 3295 | &alloc_hint); | 3295 | &alloc_hint); |
| 3296 | if (!ret) | 3296 | if (!ret) |
| 3297 | dcs = BTRFS_DC_SETUP; | 3297 | dcs = BTRFS_DC_SETUP; |
| 3298 | btrfs_free_reserved_data_space(inode, num_pages); | 3298 | btrfs_free_reserved_data_space(inode, num_pages); |
| 3299 | 3299 | ||
| 3300 | out_put: | 3300 | out_put: |
| 3301 | iput(inode); | 3301 | iput(inode); |
| 3302 | out_free: | 3302 | out_free: |
| 3303 | btrfs_release_path(path); | 3303 | btrfs_release_path(path); |
| 3304 | out: | 3304 | out: |
| 3305 | spin_lock(&block_group->lock); | 3305 | spin_lock(&block_group->lock); |
| 3306 | if (!ret && dcs == BTRFS_DC_SETUP) | 3306 | if (!ret && dcs == BTRFS_DC_SETUP) |
| 3307 | block_group->cache_generation = trans->transid; | 3307 | block_group->cache_generation = trans->transid; |
| 3308 | block_group->disk_cache_state = dcs; | 3308 | block_group->disk_cache_state = dcs; |
| 3309 | spin_unlock(&block_group->lock); | 3309 | spin_unlock(&block_group->lock); |
| 3310 | 3310 | ||
| 3311 | return ret; | 3311 | return ret; |
| 3312 | } | 3312 | } |
| 3313 | 3313 | ||
| 3314 | int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans, | 3314 | int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans, |
| 3315 | struct btrfs_root *root) | 3315 | struct btrfs_root *root) |
| 3316 | { | 3316 | { |
| 3317 | struct btrfs_block_group_cache *cache; | 3317 | struct btrfs_block_group_cache *cache; |
| 3318 | int err = 0; | 3318 | int err = 0; |
| 3319 | struct btrfs_path *path; | 3319 | struct btrfs_path *path; |
| 3320 | u64 last = 0; | 3320 | u64 last = 0; |
| 3321 | 3321 | ||
| 3322 | path = btrfs_alloc_path(); | 3322 | path = btrfs_alloc_path(); |
| 3323 | if (!path) | 3323 | if (!path) |
| 3324 | return -ENOMEM; | 3324 | return -ENOMEM; |
| 3325 | 3325 | ||
| 3326 | again: | 3326 | again: |
| 3327 | while (1) { | 3327 | while (1) { |
| 3328 | cache = btrfs_lookup_first_block_group(root->fs_info, last); | 3328 | cache = btrfs_lookup_first_block_group(root->fs_info, last); |
| 3329 | while (cache) { | 3329 | while (cache) { |
| 3330 | if (cache->disk_cache_state == BTRFS_DC_CLEAR) | 3330 | if (cache->disk_cache_state == BTRFS_DC_CLEAR) |
| 3331 | break; | 3331 | break; |
| 3332 | cache = next_block_group(root, cache); | 3332 | cache = next_block_group(root, cache); |
| 3333 | } | 3333 | } |
| 3334 | if (!cache) { | 3334 | if (!cache) { |
| 3335 | if (last == 0) | 3335 | if (last == 0) |
| 3336 | break; | 3336 | break; |
| 3337 | last = 0; | 3337 | last = 0; |
| 3338 | continue; | 3338 | continue; |
| 3339 | } | 3339 | } |
| 3340 | err = cache_save_setup(cache, trans, path); | 3340 | err = cache_save_setup(cache, trans, path); |
| 3341 | last = cache->key.objectid + cache->key.offset; | 3341 | last = cache->key.objectid + cache->key.offset; |
| 3342 | btrfs_put_block_group(cache); | 3342 | btrfs_put_block_group(cache); |
| 3343 | } | 3343 | } |
| 3344 | 3344 | ||
| 3345 | while (1) { | 3345 | while (1) { |
| 3346 | if (last == 0) { | 3346 | if (last == 0) { |
| 3347 | err = btrfs_run_delayed_refs(trans, root, | 3347 | err = btrfs_run_delayed_refs(trans, root, |
| 3348 | (unsigned long)-1); | 3348 | (unsigned long)-1); |
| 3349 | if (err) /* File system offline */ | 3349 | if (err) /* File system offline */ |
| 3350 | goto out; | 3350 | goto out; |
| 3351 | } | 3351 | } |
| 3352 | 3352 | ||
| 3353 | cache = btrfs_lookup_first_block_group(root->fs_info, last); | 3353 | cache = btrfs_lookup_first_block_group(root->fs_info, last); |
| 3354 | while (cache) { | 3354 | while (cache) { |
| 3355 | if (cache->disk_cache_state == BTRFS_DC_CLEAR) { | 3355 | if (cache->disk_cache_state == BTRFS_DC_CLEAR) { |
| 3356 | btrfs_put_block_group(cache); | 3356 | btrfs_put_block_group(cache); |
| 3357 | goto again; | 3357 | goto again; |
| 3358 | } | 3358 | } |
| 3359 | 3359 | ||
| 3360 | if (cache->dirty) | 3360 | if (cache->dirty) |
| 3361 | break; | 3361 | break; |
| 3362 | cache = next_block_group(root, cache); | 3362 | cache = next_block_group(root, cache); |
| 3363 | } | 3363 | } |
| 3364 | if (!cache) { | 3364 | if (!cache) { |
| 3365 | if (last == 0) | 3365 | if (last == 0) |
| 3366 | break; | 3366 | break; |
| 3367 | last = 0; | 3367 | last = 0; |
| 3368 | continue; | 3368 | continue; |
| 3369 | } | 3369 | } |
| 3370 | 3370 | ||
| 3371 | if (cache->disk_cache_state == BTRFS_DC_SETUP) | 3371 | if (cache->disk_cache_state == BTRFS_DC_SETUP) |
| 3372 | cache->disk_cache_state = BTRFS_DC_NEED_WRITE; | 3372 | cache->disk_cache_state = BTRFS_DC_NEED_WRITE; |
| 3373 | cache->dirty = 0; | 3373 | cache->dirty = 0; |
| 3374 | last = cache->key.objectid + cache->key.offset; | 3374 | last = cache->key.objectid + cache->key.offset; |
| 3375 | 3375 | ||
| 3376 | err = write_one_cache_group(trans, root, path, cache); | 3376 | err = write_one_cache_group(trans, root, path, cache); |
| 3377 | btrfs_put_block_group(cache); | 3377 | btrfs_put_block_group(cache); |
| 3378 | if (err) /* File system offline */ | 3378 | if (err) /* File system offline */ |
| 3379 | goto out; | 3379 | goto out; |
| 3380 | } | 3380 | } |
| 3381 | 3381 | ||
| 3382 | while (1) { | 3382 | while (1) { |
| 3383 | /* | 3383 | /* |
| 3384 | * I don't think this is needed since we're just marking our | 3384 | * I don't think this is needed since we're just marking our |
| 3385 | * preallocated extent as written, but just in case it can't | 3385 | * preallocated extent as written, but just in case it can't |
| 3386 | * hurt. | 3386 | * hurt. |
| 3387 | */ | 3387 | */ |
| 3388 | if (last == 0) { | 3388 | if (last == 0) { |
| 3389 | err = btrfs_run_delayed_refs(trans, root, | 3389 | err = btrfs_run_delayed_refs(trans, root, |
| 3390 | (unsigned long)-1); | 3390 | (unsigned long)-1); |
| 3391 | if (err) /* File system offline */ | 3391 | if (err) /* File system offline */ |
| 3392 | goto out; | 3392 | goto out; |
| 3393 | } | 3393 | } |
| 3394 | 3394 | ||
| 3395 | cache = btrfs_lookup_first_block_group(root->fs_info, last); | 3395 | cache = btrfs_lookup_first_block_group(root->fs_info, last); |
| 3396 | while (cache) { | 3396 | while (cache) { |
| 3397 | /* | 3397 | /* |
| 3398 | * Really this shouldn't happen, but it could if we | 3398 | * Really this shouldn't happen, but it could if we |
| 3399 | * couldn't write the entire preallocated extent and | 3399 | * couldn't write the entire preallocated extent and |
| 3400 | * splitting the extent resulted in a new block. | 3400 | * splitting the extent resulted in a new block. |
| 3401 | */ | 3401 | */ |
| 3402 | if (cache->dirty) { | 3402 | if (cache->dirty) { |
| 3403 | btrfs_put_block_group(cache); | 3403 | btrfs_put_block_group(cache); |
| 3404 | goto again; | 3404 | goto again; |
| 3405 | } | 3405 | } |
| 3406 | if (cache->disk_cache_state == BTRFS_DC_NEED_WRITE) | 3406 | if (cache->disk_cache_state == BTRFS_DC_NEED_WRITE) |
| 3407 | break; | 3407 | break; |
| 3408 | cache = next_block_group(root, cache); | 3408 | cache = next_block_group(root, cache); |
| 3409 | } | 3409 | } |
| 3410 | if (!cache) { | 3410 | if (!cache) { |
| 3411 | if (last == 0) | 3411 | if (last == 0) |
| 3412 | break; | 3412 | break; |
| 3413 | last = 0; | 3413 | last = 0; |
| 3414 | continue; | 3414 | continue; |
| 3415 | } | 3415 | } |
| 3416 | 3416 | ||
| 3417 | err = btrfs_write_out_cache(root, trans, cache, path); | 3417 | err = btrfs_write_out_cache(root, trans, cache, path); |
| 3418 | 3418 | ||
| 3419 | /* | 3419 | /* |
| 3420 | * If we didn't have an error then the cache state is still | 3420 | * If we didn't have an error then the cache state is still |
| 3421 | * NEED_WRITE, so we can set it to WRITTEN. | 3421 | * NEED_WRITE, so we can set it to WRITTEN. |
| 3422 | */ | 3422 | */ |
| 3423 | if (!err && cache->disk_cache_state == BTRFS_DC_NEED_WRITE) | 3423 | if (!err && cache->disk_cache_state == BTRFS_DC_NEED_WRITE) |
| 3424 | cache->disk_cache_state = BTRFS_DC_WRITTEN; | 3424 | cache->disk_cache_state = BTRFS_DC_WRITTEN; |
| 3425 | last = cache->key.objectid + cache->key.offset; | 3425 | last = cache->key.objectid + cache->key.offset; |
| 3426 | btrfs_put_block_group(cache); | 3426 | btrfs_put_block_group(cache); |
| 3427 | } | 3427 | } |
| 3428 | out: | 3428 | out: |
| 3429 | 3429 | ||
| 3430 | btrfs_free_path(path); | 3430 | btrfs_free_path(path); |
| 3431 | return err; | 3431 | return err; |
| 3432 | } | 3432 | } |
| 3433 | 3433 | ||
| 3434 | int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr) | 3434 | int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr) |
| 3435 | { | 3435 | { |
| 3436 | struct btrfs_block_group_cache *block_group; | 3436 | struct btrfs_block_group_cache *block_group; |
| 3437 | int readonly = 0; | 3437 | int readonly = 0; |
| 3438 | 3438 | ||
| 3439 | block_group = btrfs_lookup_block_group(root->fs_info, bytenr); | 3439 | block_group = btrfs_lookup_block_group(root->fs_info, bytenr); |
| 3440 | if (!block_group || block_group->ro) | 3440 | if (!block_group || block_group->ro) |
| 3441 | readonly = 1; | 3441 | readonly = 1; |
| 3442 | if (block_group) | 3442 | if (block_group) |
| 3443 | btrfs_put_block_group(block_group); | 3443 | btrfs_put_block_group(block_group); |
| 3444 | return readonly; | 3444 | return readonly; |
| 3445 | } | 3445 | } |
| 3446 | 3446 | ||
| 3447 | static const char *alloc_name(u64 flags) | 3447 | static const char *alloc_name(u64 flags) |
| 3448 | { | 3448 | { |
| 3449 | switch (flags) { | 3449 | switch (flags) { |
| 3450 | case BTRFS_BLOCK_GROUP_METADATA|BTRFS_BLOCK_GROUP_DATA: | 3450 | case BTRFS_BLOCK_GROUP_METADATA|BTRFS_BLOCK_GROUP_DATA: |
| 3451 | return "mixed"; | 3451 | return "mixed"; |
| 3452 | case BTRFS_BLOCK_GROUP_METADATA: | 3452 | case BTRFS_BLOCK_GROUP_METADATA: |
| 3453 | return "metadata"; | 3453 | return "metadata"; |
| 3454 | case BTRFS_BLOCK_GROUP_DATA: | 3454 | case BTRFS_BLOCK_GROUP_DATA: |
| 3455 | return "data"; | 3455 | return "data"; |
| 3456 | case BTRFS_BLOCK_GROUP_SYSTEM: | 3456 | case BTRFS_BLOCK_GROUP_SYSTEM: |
| 3457 | return "system"; | 3457 | return "system"; |
| 3458 | default: | 3458 | default: |
| 3459 | WARN_ON(1); | 3459 | WARN_ON(1); |
| 3460 | return "invalid-combination"; | 3460 | return "invalid-combination"; |
| 3461 | }; | 3461 | }; |
| 3462 | } | 3462 | } |
| 3463 | 3463 | ||
| 3464 | static int update_space_info(struct btrfs_fs_info *info, u64 flags, | 3464 | static int update_space_info(struct btrfs_fs_info *info, u64 flags, |
| 3465 | u64 total_bytes, u64 bytes_used, | 3465 | u64 total_bytes, u64 bytes_used, |
| 3466 | struct btrfs_space_info **space_info) | 3466 | struct btrfs_space_info **space_info) |
| 3467 | { | 3467 | { |
| 3468 | struct btrfs_space_info *found; | 3468 | struct btrfs_space_info *found; |
| 3469 | int i; | 3469 | int i; |
| 3470 | int factor; | 3470 | int factor; |
| 3471 | int ret; | 3471 | int ret; |
| 3472 | 3472 | ||
| 3473 | if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | | 3473 | if (flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | |
| 3474 | BTRFS_BLOCK_GROUP_RAID10)) | 3474 | BTRFS_BLOCK_GROUP_RAID10)) |
| 3475 | factor = 2; | 3475 | factor = 2; |
| 3476 | else | 3476 | else |
| 3477 | factor = 1; | 3477 | factor = 1; |
| 3478 | 3478 | ||
| 3479 | found = __find_space_info(info, flags); | 3479 | found = __find_space_info(info, flags); |
| 3480 | if (found) { | 3480 | if (found) { |
| 3481 | spin_lock(&found->lock); | 3481 | spin_lock(&found->lock); |
| 3482 | found->total_bytes += total_bytes; | 3482 | found->total_bytes += total_bytes; |
| 3483 | found->disk_total += total_bytes * factor; | 3483 | found->disk_total += total_bytes * factor; |
| 3484 | found->bytes_used += bytes_used; | 3484 | found->bytes_used += bytes_used; |
| 3485 | found->disk_used += bytes_used * factor; | 3485 | found->disk_used += bytes_used * factor; |
| 3486 | found->full = 0; | 3486 | found->full = 0; |
| 3487 | spin_unlock(&found->lock); | 3487 | spin_unlock(&found->lock); |
| 3488 | *space_info = found; | 3488 | *space_info = found; |
| 3489 | return 0; | 3489 | return 0; |
| 3490 | } | 3490 | } |
| 3491 | found = kzalloc(sizeof(*found), GFP_NOFS); | 3491 | found = kzalloc(sizeof(*found), GFP_NOFS); |
| 3492 | if (!found) | 3492 | if (!found) |
| 3493 | return -ENOMEM; | 3493 | return -ENOMEM; |
| 3494 | 3494 | ||
| 3495 | ret = percpu_counter_init(&found->total_bytes_pinned, 0, GFP_KERNEL); | 3495 | ret = percpu_counter_init(&found->total_bytes_pinned, 0, GFP_KERNEL); |
| 3496 | if (ret) { | 3496 | if (ret) { |
| 3497 | kfree(found); | 3497 | kfree(found); |
| 3498 | return ret; | 3498 | return ret; |
| 3499 | } | 3499 | } |
| 3500 | 3500 | ||
| 3501 | for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) | 3501 | for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) |
| 3502 | INIT_LIST_HEAD(&found->block_groups[i]); | 3502 | INIT_LIST_HEAD(&found->block_groups[i]); |
| 3503 | init_rwsem(&found->groups_sem); | 3503 | init_rwsem(&found->groups_sem); |
| 3504 | spin_lock_init(&found->lock); | 3504 | spin_lock_init(&found->lock); |
| 3505 | found->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK; | 3505 | found->flags = flags & BTRFS_BLOCK_GROUP_TYPE_MASK; |
| 3506 | found->total_bytes = total_bytes; | 3506 | found->total_bytes = total_bytes; |
| 3507 | found->disk_total = total_bytes * factor; | 3507 | found->disk_total = total_bytes * factor; |
| 3508 | found->bytes_used = bytes_used; | 3508 | found->bytes_used = bytes_used; |
| 3509 | found->disk_used = bytes_used * factor; | 3509 | found->disk_used = bytes_used * factor; |
| 3510 | found->bytes_pinned = 0; | 3510 | found->bytes_pinned = 0; |
| 3511 | found->bytes_reserved = 0; | 3511 | found->bytes_reserved = 0; |
| 3512 | found->bytes_readonly = 0; | 3512 | found->bytes_readonly = 0; |
| 3513 | found->bytes_may_use = 0; | 3513 | found->bytes_may_use = 0; |
| 3514 | found->full = 0; | 3514 | found->full = 0; |
| 3515 | found->force_alloc = CHUNK_ALLOC_NO_FORCE; | 3515 | found->force_alloc = CHUNK_ALLOC_NO_FORCE; |
| 3516 | found->chunk_alloc = 0; | 3516 | found->chunk_alloc = 0; |
| 3517 | found->flush = 0; | 3517 | found->flush = 0; |
| 3518 | init_waitqueue_head(&found->wait); | 3518 | init_waitqueue_head(&found->wait); |
| 3519 | INIT_LIST_HEAD(&found->ro_bgs); | 3519 | INIT_LIST_HEAD(&found->ro_bgs); |
| 3520 | 3520 | ||
| 3521 | ret = kobject_init_and_add(&found->kobj, &space_info_ktype, | 3521 | ret = kobject_init_and_add(&found->kobj, &space_info_ktype, |
| 3522 | info->space_info_kobj, "%s", | 3522 | info->space_info_kobj, "%s", |
| 3523 | alloc_name(found->flags)); | 3523 | alloc_name(found->flags)); |
| 3524 | if (ret) { | 3524 | if (ret) { |
| 3525 | kfree(found); | 3525 | kfree(found); |
| 3526 | return ret; | 3526 | return ret; |
| 3527 | } | 3527 | } |
| 3528 | 3528 | ||
| 3529 | *space_info = found; | 3529 | *space_info = found; |
| 3530 | list_add_rcu(&found->list, &info->space_info); | 3530 | list_add_rcu(&found->list, &info->space_info); |
| 3531 | if (flags & BTRFS_BLOCK_GROUP_DATA) | 3531 | if (flags & BTRFS_BLOCK_GROUP_DATA) |
| 3532 | info->data_sinfo = found; | 3532 | info->data_sinfo = found; |
| 3533 | 3533 | ||
| 3534 | return ret; | 3534 | return ret; |
| 3535 | } | 3535 | } |
| 3536 | 3536 | ||
| 3537 | static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) | 3537 | static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) |
| 3538 | { | 3538 | { |
| 3539 | u64 extra_flags = chunk_to_extended(flags) & | 3539 | u64 extra_flags = chunk_to_extended(flags) & |
| 3540 | BTRFS_EXTENDED_PROFILE_MASK; | 3540 | BTRFS_EXTENDED_PROFILE_MASK; |
| 3541 | 3541 | ||
| 3542 | write_seqlock(&fs_info->profiles_lock); | 3542 | write_seqlock(&fs_info->profiles_lock); |
| 3543 | if (flags & BTRFS_BLOCK_GROUP_DATA) | 3543 | if (flags & BTRFS_BLOCK_GROUP_DATA) |
| 3544 | fs_info->avail_data_alloc_bits |= extra_flags; | 3544 | fs_info->avail_data_alloc_bits |= extra_flags; |
| 3545 | if (flags & BTRFS_BLOCK_GROUP_METADATA) | 3545 | if (flags & BTRFS_BLOCK_GROUP_METADATA) |
| 3546 | fs_info->avail_metadata_alloc_bits |= extra_flags; | 3546 | fs_info->avail_metadata_alloc_bits |= extra_flags; |
| 3547 | if (flags & BTRFS_BLOCK_GROUP_SYSTEM) | 3547 | if (flags & BTRFS_BLOCK_GROUP_SYSTEM) |
| 3548 | fs_info->avail_system_alloc_bits |= extra_flags; | 3548 | fs_info->avail_system_alloc_bits |= extra_flags; |
| 3549 | write_sequnlock(&fs_info->profiles_lock); | 3549 | write_sequnlock(&fs_info->profiles_lock); |
| 3550 | } | 3550 | } |
| 3551 | 3551 | ||
| 3552 | /* | 3552 | /* |
| 3553 | * returns target flags in extended format or 0 if restripe for this | 3553 | * returns target flags in extended format or 0 if restripe for this |
| 3554 | * chunk_type is not in progress | 3554 | * chunk_type is not in progress |
| 3555 | * | 3555 | * |
| 3556 | * should be called with either volume_mutex or balance_lock held | 3556 | * should be called with either volume_mutex or balance_lock held |
| 3557 | */ | 3557 | */ |
| 3558 | static u64 get_restripe_target(struct btrfs_fs_info *fs_info, u64 flags) | 3558 | static u64 get_restripe_target(struct btrfs_fs_info *fs_info, u64 flags) |
| 3559 | { | 3559 | { |
| 3560 | struct btrfs_balance_control *bctl = fs_info->balance_ctl; | 3560 | struct btrfs_balance_control *bctl = fs_info->balance_ctl; |
| 3561 | u64 target = 0; | 3561 | u64 target = 0; |
| 3562 | 3562 | ||
| 3563 | if (!bctl) | 3563 | if (!bctl) |
| 3564 | return 0; | 3564 | return 0; |
| 3565 | 3565 | ||
| 3566 | if (flags & BTRFS_BLOCK_GROUP_DATA && | 3566 | if (flags & BTRFS_BLOCK_GROUP_DATA && |
| 3567 | bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) { | 3567 | bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) { |
| 3568 | target = BTRFS_BLOCK_GROUP_DATA | bctl->data.target; | 3568 | target = BTRFS_BLOCK_GROUP_DATA | bctl->data.target; |
| 3569 | } else if (flags & BTRFS_BLOCK_GROUP_SYSTEM && | 3569 | } else if (flags & BTRFS_BLOCK_GROUP_SYSTEM && |
| 3570 | bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) { | 3570 | bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) { |
| 3571 | target = BTRFS_BLOCK_GROUP_SYSTEM | bctl->sys.target; | 3571 | target = BTRFS_BLOCK_GROUP_SYSTEM | bctl->sys.target; |
| 3572 | } else if (flags & BTRFS_BLOCK_GROUP_METADATA && | 3572 | } else if (flags & BTRFS_BLOCK_GROUP_METADATA && |
| 3573 | bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) { | 3573 | bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) { |
| 3574 | target = BTRFS_BLOCK_GROUP_METADATA | bctl->meta.target; | 3574 | target = BTRFS_BLOCK_GROUP_METADATA | bctl->meta.target; |
| 3575 | } | 3575 | } |
| 3576 | 3576 | ||
| 3577 | return target; | 3577 | return target; |
| 3578 | } | 3578 | } |
| 3579 | 3579 | ||
| 3580 | /* | 3580 | /* |
| 3581 | * @flags: available profiles in extended format (see ctree.h) | 3581 | * @flags: available profiles in extended format (see ctree.h) |
| 3582 | * | 3582 | * |
| 3583 | * Returns reduced profile in chunk format. If profile changing is in | 3583 | * Returns reduced profile in chunk format. If profile changing is in |
| 3584 | * progress (either running or paused) picks the target profile (if it's | 3584 | * progress (either running or paused) picks the target profile (if it's |
| 3585 | * already available), otherwise falls back to plain reducing. | 3585 | * already available), otherwise falls back to plain reducing. |
| 3586 | */ | 3586 | */ |
| 3587 | static u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) | 3587 | static u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) |
| 3588 | { | 3588 | { |
| 3589 | u64 num_devices = root->fs_info->fs_devices->rw_devices; | 3589 | u64 num_devices = root->fs_info->fs_devices->rw_devices; |
| 3590 | u64 target; | 3590 | u64 target; |
| 3591 | u64 tmp; | 3591 | u64 tmp; |
| 3592 | 3592 | ||
| 3593 | /* | 3593 | /* |
| 3594 | * see if restripe for this chunk_type is in progress, if so | 3594 | * see if restripe for this chunk_type is in progress, if so |
| 3595 | * try to reduce to the target profile | 3595 | * try to reduce to the target profile |
| 3596 | */ | 3596 | */ |
| 3597 | spin_lock(&root->fs_info->balance_lock); | 3597 | spin_lock(&root->fs_info->balance_lock); |
| 3598 | target = get_restripe_target(root->fs_info, flags); | 3598 | target = get_restripe_target(root->fs_info, flags); |
| 3599 | if (target) { | 3599 | if (target) { |
| 3600 | /* pick target profile only if it's already available */ | 3600 | /* pick target profile only if it's already available */ |
| 3601 | if ((flags & target) & BTRFS_EXTENDED_PROFILE_MASK) { | 3601 | if ((flags & target) & BTRFS_EXTENDED_PROFILE_MASK) { |
| 3602 | spin_unlock(&root->fs_info->balance_lock); | 3602 | spin_unlock(&root->fs_info->balance_lock); |
| 3603 | return extended_to_chunk(target); | 3603 | return extended_to_chunk(target); |
| 3604 | } | 3604 | } |
| 3605 | } | 3605 | } |
| 3606 | spin_unlock(&root->fs_info->balance_lock); | 3606 | spin_unlock(&root->fs_info->balance_lock); |
| 3607 | 3607 | ||
| 3608 | /* First, mask out the RAID levels which aren't possible */ | 3608 | /* First, mask out the RAID levels which aren't possible */ |
| 3609 | if (num_devices == 1) | 3609 | if (num_devices == 1) |
| 3610 | flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0 | | 3610 | flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0 | |
| 3611 | BTRFS_BLOCK_GROUP_RAID5); | 3611 | BTRFS_BLOCK_GROUP_RAID5); |
| 3612 | if (num_devices < 3) | 3612 | if (num_devices < 3) |
| 3613 | flags &= ~BTRFS_BLOCK_GROUP_RAID6; | 3613 | flags &= ~BTRFS_BLOCK_GROUP_RAID6; |
| 3614 | if (num_devices < 4) | 3614 | if (num_devices < 4) |
| 3615 | flags &= ~BTRFS_BLOCK_GROUP_RAID10; | 3615 | flags &= ~BTRFS_BLOCK_GROUP_RAID10; |
| 3616 | 3616 | ||
| 3617 | tmp = flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID0 | | 3617 | tmp = flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID0 | |
| 3618 | BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID5 | | 3618 | BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID5 | |
| 3619 | BTRFS_BLOCK_GROUP_RAID6 | BTRFS_BLOCK_GROUP_RAID10); | 3619 | BTRFS_BLOCK_GROUP_RAID6 | BTRFS_BLOCK_GROUP_RAID10); |
| 3620 | flags &= ~tmp; | 3620 | flags &= ~tmp; |
| 3621 | 3621 | ||
| 3622 | if (tmp & BTRFS_BLOCK_GROUP_RAID6) | 3622 | if (tmp & BTRFS_BLOCK_GROUP_RAID6) |
| 3623 | tmp = BTRFS_BLOCK_GROUP_RAID6; | 3623 | tmp = BTRFS_BLOCK_GROUP_RAID6; |
| 3624 | else if (tmp & BTRFS_BLOCK_GROUP_RAID5) | 3624 | else if (tmp & BTRFS_BLOCK_GROUP_RAID5) |
| 3625 | tmp = BTRFS_BLOCK_GROUP_RAID5; | 3625 | tmp = BTRFS_BLOCK_GROUP_RAID5; |
| 3626 | else if (tmp & BTRFS_BLOCK_GROUP_RAID10) | 3626 | else if (tmp & BTRFS_BLOCK_GROUP_RAID10) |
| 3627 | tmp = BTRFS_BLOCK_GROUP_RAID10; | 3627 | tmp = BTRFS_BLOCK_GROUP_RAID10; |
| 3628 | else if (tmp & BTRFS_BLOCK_GROUP_RAID1) | 3628 | else if (tmp & BTRFS_BLOCK_GROUP_RAID1) |
| 3629 | tmp = BTRFS_BLOCK_GROUP_RAID1; | 3629 | tmp = BTRFS_BLOCK_GROUP_RAID1; |
| 3630 | else if (tmp & BTRFS_BLOCK_GROUP_RAID0) | 3630 | else if (tmp & BTRFS_BLOCK_GROUP_RAID0) |
| 3631 | tmp = BTRFS_BLOCK_GROUP_RAID0; | 3631 | tmp = BTRFS_BLOCK_GROUP_RAID0; |
| 3632 | 3632 | ||
| 3633 | return extended_to_chunk(flags | tmp); | 3633 | return extended_to_chunk(flags | tmp); |
| 3634 | } | 3634 | } |
| 3635 | 3635 | ||
| 3636 | static u64 get_alloc_profile(struct btrfs_root *root, u64 orig_flags) | 3636 | static u64 get_alloc_profile(struct btrfs_root *root, u64 orig_flags) |
| 3637 | { | 3637 | { |
| 3638 | unsigned seq; | 3638 | unsigned seq; |
| 3639 | u64 flags; | 3639 | u64 flags; |
| 3640 | 3640 | ||
| 3641 | do { | 3641 | do { |
| 3642 | flags = orig_flags; | 3642 | flags = orig_flags; |
| 3643 | seq = read_seqbegin(&root->fs_info->profiles_lock); | 3643 | seq = read_seqbegin(&root->fs_info->profiles_lock); |
| 3644 | 3644 | ||
| 3645 | if (flags & BTRFS_BLOCK_GROUP_DATA) | 3645 | if (flags & BTRFS_BLOCK_GROUP_DATA) |
| 3646 | flags |= root->fs_info->avail_data_alloc_bits; | 3646 | flags |= root->fs_info->avail_data_alloc_bits; |
| 3647 | else if (flags & BTRFS_BLOCK_GROUP_SYSTEM) | 3647 | else if (flags & BTRFS_BLOCK_GROUP_SYSTEM) |
| 3648 | flags |= root->fs_info->avail_system_alloc_bits; | 3648 | flags |= root->fs_info->avail_system_alloc_bits; |
| 3649 | else if (flags & BTRFS_BLOCK_GROUP_METADATA) | 3649 | else if (flags & BTRFS_BLOCK_GROUP_METADATA) |
| 3650 | flags |= root->fs_info->avail_metadata_alloc_bits; | 3650 | flags |= root->fs_info->avail_metadata_alloc_bits; |
| 3651 | } while (read_seqretry(&root->fs_info->profiles_lock, seq)); | 3651 | } while (read_seqretry(&root->fs_info->profiles_lock, seq)); |
| 3652 | 3652 | ||
| 3653 | return btrfs_reduce_alloc_profile(root, flags); | 3653 | return btrfs_reduce_alloc_profile(root, flags); |
| 3654 | } | 3654 | } |
| 3655 | 3655 | ||
| 3656 | u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) | 3656 | u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) |
| 3657 | { | 3657 | { |
| 3658 | u64 flags; | 3658 | u64 flags; |
| 3659 | u64 ret; | 3659 | u64 ret; |
| 3660 | 3660 | ||
| 3661 | if (data) | 3661 | if (data) |
| 3662 | flags = BTRFS_BLOCK_GROUP_DATA; | 3662 | flags = BTRFS_BLOCK_GROUP_DATA; |
| 3663 | else if (root == root->fs_info->chunk_root) | 3663 | else if (root == root->fs_info->chunk_root) |
| 3664 | flags = BTRFS_BLOCK_GROUP_SYSTEM; | 3664 | flags = BTRFS_BLOCK_GROUP_SYSTEM; |
| 3665 | else | 3665 | else |
| 3666 | flags = BTRFS_BLOCK_GROUP_METADATA; | 3666 | flags = BTRFS_BLOCK_GROUP_METADATA; |
| 3667 | 3667 | ||
| 3668 | ret = get_alloc_profile(root, flags); | 3668 | ret = get_alloc_profile(root, flags); |
| 3669 | return ret; | 3669 | return ret; |
| 3670 | } | 3670 | } |
| 3671 | 3671 | ||
| 3672 | /* | 3672 | /* |
| 3673 | * This will check the space that the inode allocates from to make sure we have | 3673 | * This will check the space that the inode allocates from to make sure we have |
| 3674 | * enough space for bytes. | 3674 | * enough space for bytes. |
| 3675 | */ | 3675 | */ |
| 3676 | int btrfs_check_data_free_space(struct inode *inode, u64 bytes) | 3676 | int btrfs_check_data_free_space(struct inode *inode, u64 bytes) |
| 3677 | { | 3677 | { |
| 3678 | struct btrfs_space_info *data_sinfo; | 3678 | struct btrfs_space_info *data_sinfo; |
| 3679 | struct btrfs_root *root = BTRFS_I(inode)->root; | 3679 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 3680 | struct btrfs_fs_info *fs_info = root->fs_info; | 3680 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 3681 | u64 used; | 3681 | u64 used; |
| 3682 | int ret = 0, committed = 0, alloc_chunk = 1; | 3682 | int ret = 0, committed = 0, alloc_chunk = 1; |
| 3683 | 3683 | ||
| 3684 | /* make sure bytes are sectorsize aligned */ | 3684 | /* make sure bytes are sectorsize aligned */ |
| 3685 | bytes = ALIGN(bytes, root->sectorsize); | 3685 | bytes = ALIGN(bytes, root->sectorsize); |
| 3686 | 3686 | ||
| 3687 | if (btrfs_is_free_space_inode(inode)) { | 3687 | if (btrfs_is_free_space_inode(inode)) { |
| 3688 | committed = 1; | 3688 | committed = 1; |
| 3689 | ASSERT(current->journal_info); | 3689 | ASSERT(current->journal_info); |
| 3690 | } | 3690 | } |
| 3691 | 3691 | ||
| 3692 | data_sinfo = fs_info->data_sinfo; | 3692 | data_sinfo = fs_info->data_sinfo; |
| 3693 | if (!data_sinfo) | 3693 | if (!data_sinfo) |
| 3694 | goto alloc; | 3694 | goto alloc; |
| 3695 | 3695 | ||
| 3696 | again: | 3696 | again: |
| 3697 | /* make sure we have enough space to handle the data first */ | 3697 | /* make sure we have enough space to handle the data first */ |
| 3698 | spin_lock(&data_sinfo->lock); | 3698 | spin_lock(&data_sinfo->lock); |
| 3699 | used = data_sinfo->bytes_used + data_sinfo->bytes_reserved + | 3699 | used = data_sinfo->bytes_used + data_sinfo->bytes_reserved + |
| 3700 | data_sinfo->bytes_pinned + data_sinfo->bytes_readonly + | 3700 | data_sinfo->bytes_pinned + data_sinfo->bytes_readonly + |
| 3701 | data_sinfo->bytes_may_use; | 3701 | data_sinfo->bytes_may_use; |
| 3702 | 3702 | ||
| 3703 | if (used + bytes > data_sinfo->total_bytes) { | 3703 | if (used + bytes > data_sinfo->total_bytes) { |
| 3704 | struct btrfs_trans_handle *trans; | 3704 | struct btrfs_trans_handle *trans; |
| 3705 | 3705 | ||
| 3706 | /* | 3706 | /* |
| 3707 | * if we don't have enough free bytes in this space then we need | 3707 | * if we don't have enough free bytes in this space then we need |
| 3708 | * to alloc a new chunk. | 3708 | * to alloc a new chunk. |
| 3709 | */ | 3709 | */ |
| 3710 | if (!data_sinfo->full && alloc_chunk) { | 3710 | if (!data_sinfo->full && alloc_chunk) { |
| 3711 | u64 alloc_target; | 3711 | u64 alloc_target; |
| 3712 | 3712 | ||
| 3713 | data_sinfo->force_alloc = CHUNK_ALLOC_FORCE; | 3713 | data_sinfo->force_alloc = CHUNK_ALLOC_FORCE; |
| 3714 | spin_unlock(&data_sinfo->lock); | 3714 | spin_unlock(&data_sinfo->lock); |
| 3715 | alloc: | 3715 | alloc: |
| 3716 | alloc_target = btrfs_get_alloc_profile(root, 1); | 3716 | alloc_target = btrfs_get_alloc_profile(root, 1); |
| 3717 | /* | 3717 | /* |
| 3718 | * It is ugly that we don't call nolock join | 3718 | * It is ugly that we don't call nolock join |
| 3719 | * transaction for the free space inode case here. | 3719 | * transaction for the free space inode case here. |
| 3720 | * But it is safe because we only do the data space | 3720 | * But it is safe because we only do the data space |
| 3721 | * reservation for the free space cache in the | 3721 | * reservation for the free space cache in the |
| 3722 | * transaction context, the common join transaction | 3722 | * transaction context, the common join transaction |
| 3723 | * just increase the counter of the current transaction | 3723 | * just increase the counter of the current transaction |
| 3724 | * handler, doesn't try to acquire the trans_lock of | 3724 | * handler, doesn't try to acquire the trans_lock of |
| 3725 | * the fs. | 3725 | * the fs. |
| 3726 | */ | 3726 | */ |
| 3727 | trans = btrfs_join_transaction(root); | 3727 | trans = btrfs_join_transaction(root); |
| 3728 | if (IS_ERR(trans)) | 3728 | if (IS_ERR(trans)) |
| 3729 | return PTR_ERR(trans); | 3729 | return PTR_ERR(trans); |
| 3730 | 3730 | ||
| 3731 | ret = do_chunk_alloc(trans, root->fs_info->extent_root, | 3731 | ret = do_chunk_alloc(trans, root->fs_info->extent_root, |
| 3732 | alloc_target, | 3732 | alloc_target, |
| 3733 | CHUNK_ALLOC_NO_FORCE); | 3733 | CHUNK_ALLOC_NO_FORCE); |
| 3734 | btrfs_end_transaction(trans, root); | 3734 | btrfs_end_transaction(trans, root); |
| 3735 | if (ret < 0) { | 3735 | if (ret < 0) { |
| 3736 | if (ret != -ENOSPC) | 3736 | if (ret != -ENOSPC) |
| 3737 | return ret; | 3737 | return ret; |
| 3738 | else | 3738 | else |
| 3739 | goto commit_trans; | 3739 | goto commit_trans; |
| 3740 | } | 3740 | } |
| 3741 | 3741 | ||
| 3742 | if (!data_sinfo) | 3742 | if (!data_sinfo) |
| 3743 | data_sinfo = fs_info->data_sinfo; | 3743 | data_sinfo = fs_info->data_sinfo; |
| 3744 | 3744 | ||
| 3745 | goto again; | 3745 | goto again; |
| 3746 | } | 3746 | } |
| 3747 | 3747 | ||
| 3748 | /* | 3748 | /* |
| 3749 | * If we don't have enough pinned space to deal with this | 3749 | * If we don't have enough pinned space to deal with this |
| 3750 | * allocation don't bother committing the transaction. | 3750 | * allocation don't bother committing the transaction. |
| 3751 | */ | 3751 | */ |
| 3752 | if (percpu_counter_compare(&data_sinfo->total_bytes_pinned, | 3752 | if (percpu_counter_compare(&data_sinfo->total_bytes_pinned, |
| 3753 | bytes) < 0) | 3753 | bytes) < 0) |
| 3754 | committed = 1; | 3754 | committed = 1; |
| 3755 | spin_unlock(&data_sinfo->lock); | 3755 | spin_unlock(&data_sinfo->lock); |
| 3756 | 3756 | ||
| 3757 | /* commit the current transaction and try again */ | 3757 | /* commit the current transaction and try again */ |
| 3758 | commit_trans: | 3758 | commit_trans: |
| 3759 | if (!committed && | 3759 | if (!committed && |
| 3760 | !atomic_read(&root->fs_info->open_ioctl_trans)) { | 3760 | !atomic_read(&root->fs_info->open_ioctl_trans)) { |
| 3761 | committed = 1; | 3761 | committed = 1; |
| 3762 | 3762 | ||
| 3763 | trans = btrfs_join_transaction(root); | 3763 | trans = btrfs_join_transaction(root); |
| 3764 | if (IS_ERR(trans)) | 3764 | if (IS_ERR(trans)) |
| 3765 | return PTR_ERR(trans); | 3765 | return PTR_ERR(trans); |
| 3766 | ret = btrfs_commit_transaction(trans, root); | 3766 | ret = btrfs_commit_transaction(trans, root); |
| 3767 | if (ret) | 3767 | if (ret) |
| 3768 | return ret; | 3768 | return ret; |
| 3769 | goto again; | 3769 | goto again; |
| 3770 | } | 3770 | } |
| 3771 | 3771 | ||
| 3772 | trace_btrfs_space_reservation(root->fs_info, | 3772 | trace_btrfs_space_reservation(root->fs_info, |
| 3773 | "space_info:enospc", | 3773 | "space_info:enospc", |
| 3774 | data_sinfo->flags, bytes, 1); | 3774 | data_sinfo->flags, bytes, 1); |
| 3775 | return -ENOSPC; | 3775 | return -ENOSPC; |
| 3776 | } | 3776 | } |
| 3777 | data_sinfo->bytes_may_use += bytes; | 3777 | data_sinfo->bytes_may_use += bytes; |
| 3778 | trace_btrfs_space_reservation(root->fs_info, "space_info", | 3778 | trace_btrfs_space_reservation(root->fs_info, "space_info", |
| 3779 | data_sinfo->flags, bytes, 1); | 3779 | data_sinfo->flags, bytes, 1); |
| 3780 | spin_unlock(&data_sinfo->lock); | 3780 | spin_unlock(&data_sinfo->lock); |
| 3781 | 3781 | ||
| 3782 | return 0; | 3782 | return 0; |
| 3783 | } | 3783 | } |
| 3784 | 3784 | ||
| 3785 | /* | 3785 | /* |
| 3786 | * Called if we need to clear a data reservation for this inode. | 3786 | * Called if we need to clear a data reservation for this inode. |
| 3787 | */ | 3787 | */ |
| 3788 | void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes) | 3788 | void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes) |
| 3789 | { | 3789 | { |
| 3790 | struct btrfs_root *root = BTRFS_I(inode)->root; | 3790 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 3791 | struct btrfs_space_info *data_sinfo; | 3791 | struct btrfs_space_info *data_sinfo; |
| 3792 | 3792 | ||
| 3793 | /* make sure bytes are sectorsize aligned */ | 3793 | /* make sure bytes are sectorsize aligned */ |
| 3794 | bytes = ALIGN(bytes, root->sectorsize); | 3794 | bytes = ALIGN(bytes, root->sectorsize); |
| 3795 | 3795 | ||
| 3796 | data_sinfo = root->fs_info->data_sinfo; | 3796 | data_sinfo = root->fs_info->data_sinfo; |
| 3797 | spin_lock(&data_sinfo->lock); | 3797 | spin_lock(&data_sinfo->lock); |
| 3798 | WARN_ON(data_sinfo->bytes_may_use < bytes); | 3798 | WARN_ON(data_sinfo->bytes_may_use < bytes); |
| 3799 | data_sinfo->bytes_may_use -= bytes; | 3799 | data_sinfo->bytes_may_use -= bytes; |
| 3800 | trace_btrfs_space_reservation(root->fs_info, "space_info", | 3800 | trace_btrfs_space_reservation(root->fs_info, "space_info", |
| 3801 | data_sinfo->flags, bytes, 0); | 3801 | data_sinfo->flags, bytes, 0); |
| 3802 | spin_unlock(&data_sinfo->lock); | 3802 | spin_unlock(&data_sinfo->lock); |
| 3803 | } | 3803 | } |
| 3804 | 3804 | ||
| 3805 | static void force_metadata_allocation(struct btrfs_fs_info *info) | 3805 | static void force_metadata_allocation(struct btrfs_fs_info *info) |
| 3806 | { | 3806 | { |
| 3807 | struct list_head *head = &info->space_info; | 3807 | struct list_head *head = &info->space_info; |
| 3808 | struct btrfs_space_info *found; | 3808 | struct btrfs_space_info *found; |
| 3809 | 3809 | ||
| 3810 | rcu_read_lock(); | 3810 | rcu_read_lock(); |
| 3811 | list_for_each_entry_rcu(found, head, list) { | 3811 | list_for_each_entry_rcu(found, head, list) { |
| 3812 | if (found->flags & BTRFS_BLOCK_GROUP_METADATA) | 3812 | if (found->flags & BTRFS_BLOCK_GROUP_METADATA) |
| 3813 | found->force_alloc = CHUNK_ALLOC_FORCE; | 3813 | found->force_alloc = CHUNK_ALLOC_FORCE; |
| 3814 | } | 3814 | } |
| 3815 | rcu_read_unlock(); | 3815 | rcu_read_unlock(); |
| 3816 | } | 3816 | } |
| 3817 | 3817 | ||
| 3818 | static inline u64 calc_global_rsv_need_space(struct btrfs_block_rsv *global) | 3818 | static inline u64 calc_global_rsv_need_space(struct btrfs_block_rsv *global) |
| 3819 | { | 3819 | { |
| 3820 | return (global->size << 1); | 3820 | return (global->size << 1); |
| 3821 | } | 3821 | } |
| 3822 | 3822 | ||
| 3823 | static int should_alloc_chunk(struct btrfs_root *root, | 3823 | static int should_alloc_chunk(struct btrfs_root *root, |
| 3824 | struct btrfs_space_info *sinfo, int force) | 3824 | struct btrfs_space_info *sinfo, int force) |
| 3825 | { | 3825 | { |
| 3826 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; | 3826 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; |
| 3827 | u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly; | 3827 | u64 num_bytes = sinfo->total_bytes - sinfo->bytes_readonly; |
| 3828 | u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved; | 3828 | u64 num_allocated = sinfo->bytes_used + sinfo->bytes_reserved; |
| 3829 | u64 thresh; | 3829 | u64 thresh; |
| 3830 | 3830 | ||
| 3831 | if (force == CHUNK_ALLOC_FORCE) | 3831 | if (force == CHUNK_ALLOC_FORCE) |
| 3832 | return 1; | 3832 | return 1; |
| 3833 | 3833 | ||
| 3834 | /* | 3834 | /* |
| 3835 | * We need to take into account the global rsv because for all intents | 3835 | * We need to take into account the global rsv because for all intents |
| 3836 | * and purposes it's used space. Don't worry about locking the | 3836 | * and purposes it's used space. Don't worry about locking the |
| 3837 | * global_rsv, it doesn't change except when the transaction commits. | 3837 | * global_rsv, it doesn't change except when the transaction commits. |
| 3838 | */ | 3838 | */ |
| 3839 | if (sinfo->flags & BTRFS_BLOCK_GROUP_METADATA) | 3839 | if (sinfo->flags & BTRFS_BLOCK_GROUP_METADATA) |
| 3840 | num_allocated += calc_global_rsv_need_space(global_rsv); | 3840 | num_allocated += calc_global_rsv_need_space(global_rsv); |
| 3841 | 3841 | ||
| 3842 | /* | 3842 | /* |
| 3843 | * in limited mode, we want to have some free space up to | 3843 | * in limited mode, we want to have some free space up to |
| 3844 | * about 1% of the FS size. | 3844 | * about 1% of the FS size. |
| 3845 | */ | 3845 | */ |
| 3846 | if (force == CHUNK_ALLOC_LIMITED) { | 3846 | if (force == CHUNK_ALLOC_LIMITED) { |
| 3847 | thresh = btrfs_super_total_bytes(root->fs_info->super_copy); | 3847 | thresh = btrfs_super_total_bytes(root->fs_info->super_copy); |
| 3848 | thresh = max_t(u64, 64 * 1024 * 1024, | 3848 | thresh = max_t(u64, 64 * 1024 * 1024, |
| 3849 | div_factor_fine(thresh, 1)); | 3849 | div_factor_fine(thresh, 1)); |
| 3850 | 3850 | ||
| 3851 | if (num_bytes - num_allocated < thresh) | 3851 | if (num_bytes - num_allocated < thresh) |
| 3852 | return 1; | 3852 | return 1; |
| 3853 | } | 3853 | } |
| 3854 | 3854 | ||
| 3855 | if (num_allocated + 2 * 1024 * 1024 < div_factor(num_bytes, 8)) | 3855 | if (num_allocated + 2 * 1024 * 1024 < div_factor(num_bytes, 8)) |
| 3856 | return 0; | 3856 | return 0; |
| 3857 | return 1; | 3857 | return 1; |
| 3858 | } | 3858 | } |
| 3859 | 3859 | ||
| 3860 | static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type) | 3860 | static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type) |
| 3861 | { | 3861 | { |
| 3862 | u64 num_dev; | 3862 | u64 num_dev; |
| 3863 | 3863 | ||
| 3864 | if (type & (BTRFS_BLOCK_GROUP_RAID10 | | 3864 | if (type & (BTRFS_BLOCK_GROUP_RAID10 | |
| 3865 | BTRFS_BLOCK_GROUP_RAID0 | | 3865 | BTRFS_BLOCK_GROUP_RAID0 | |
| 3866 | BTRFS_BLOCK_GROUP_RAID5 | | 3866 | BTRFS_BLOCK_GROUP_RAID5 | |
| 3867 | BTRFS_BLOCK_GROUP_RAID6)) | 3867 | BTRFS_BLOCK_GROUP_RAID6)) |
| 3868 | num_dev = root->fs_info->fs_devices->rw_devices; | 3868 | num_dev = root->fs_info->fs_devices->rw_devices; |
| 3869 | else if (type & BTRFS_BLOCK_GROUP_RAID1) | 3869 | else if (type & BTRFS_BLOCK_GROUP_RAID1) |
| 3870 | num_dev = 2; | 3870 | num_dev = 2; |
| 3871 | else | 3871 | else |
| 3872 | num_dev = 1; /* DUP or single */ | 3872 | num_dev = 1; /* DUP or single */ |
| 3873 | 3873 | ||
| 3874 | /* metadata for updaing devices and chunk tree */ | 3874 | /* metadata for updaing devices and chunk tree */ |
| 3875 | return btrfs_calc_trans_metadata_size(root, num_dev + 1); | 3875 | return btrfs_calc_trans_metadata_size(root, num_dev + 1); |
| 3876 | } | 3876 | } |
| 3877 | 3877 | ||
| 3878 | static void check_system_chunk(struct btrfs_trans_handle *trans, | 3878 | static void check_system_chunk(struct btrfs_trans_handle *trans, |
| 3879 | struct btrfs_root *root, u64 type) | 3879 | struct btrfs_root *root, u64 type) |
| 3880 | { | 3880 | { |
| 3881 | struct btrfs_space_info *info; | 3881 | struct btrfs_space_info *info; |
| 3882 | u64 left; | 3882 | u64 left; |
| 3883 | u64 thresh; | 3883 | u64 thresh; |
| 3884 | 3884 | ||
| 3885 | info = __find_space_info(root->fs_info, BTRFS_BLOCK_GROUP_SYSTEM); | 3885 | info = __find_space_info(root->fs_info, BTRFS_BLOCK_GROUP_SYSTEM); |
| 3886 | spin_lock(&info->lock); | 3886 | spin_lock(&info->lock); |
| 3887 | left = info->total_bytes - info->bytes_used - info->bytes_pinned - | 3887 | left = info->total_bytes - info->bytes_used - info->bytes_pinned - |
| 3888 | info->bytes_reserved - info->bytes_readonly; | 3888 | info->bytes_reserved - info->bytes_readonly; |
| 3889 | spin_unlock(&info->lock); | 3889 | spin_unlock(&info->lock); |
| 3890 | 3890 | ||
| 3891 | thresh = get_system_chunk_thresh(root, type); | 3891 | thresh = get_system_chunk_thresh(root, type); |
| 3892 | if (left < thresh && btrfs_test_opt(root, ENOSPC_DEBUG)) { | 3892 | if (left < thresh && btrfs_test_opt(root, ENOSPC_DEBUG)) { |
| 3893 | btrfs_info(root->fs_info, "left=%llu, need=%llu, flags=%llu", | 3893 | btrfs_info(root->fs_info, "left=%llu, need=%llu, flags=%llu", |
| 3894 | left, thresh, type); | 3894 | left, thresh, type); |
| 3895 | dump_space_info(info, 0, 0); | 3895 | dump_space_info(info, 0, 0); |
| 3896 | } | 3896 | } |
| 3897 | 3897 | ||
| 3898 | if (left < thresh) { | 3898 | if (left < thresh) { |
| 3899 | u64 flags; | 3899 | u64 flags; |
| 3900 | 3900 | ||
| 3901 | flags = btrfs_get_alloc_profile(root->fs_info->chunk_root, 0); | 3901 | flags = btrfs_get_alloc_profile(root->fs_info->chunk_root, 0); |
| 3902 | btrfs_alloc_chunk(trans, root, flags); | 3902 | btrfs_alloc_chunk(trans, root, flags); |
| 3903 | } | 3903 | } |
| 3904 | } | 3904 | } |
| 3905 | 3905 | ||
| 3906 | static int do_chunk_alloc(struct btrfs_trans_handle *trans, | 3906 | static int do_chunk_alloc(struct btrfs_trans_handle *trans, |
| 3907 | struct btrfs_root *extent_root, u64 flags, int force) | 3907 | struct btrfs_root *extent_root, u64 flags, int force) |
| 3908 | { | 3908 | { |
| 3909 | struct btrfs_space_info *space_info; | 3909 | struct btrfs_space_info *space_info; |
| 3910 | struct btrfs_fs_info *fs_info = extent_root->fs_info; | 3910 | struct btrfs_fs_info *fs_info = extent_root->fs_info; |
| 3911 | int wait_for_alloc = 0; | 3911 | int wait_for_alloc = 0; |
| 3912 | int ret = 0; | 3912 | int ret = 0; |
| 3913 | 3913 | ||
| 3914 | /* Don't re-enter if we're already allocating a chunk */ | 3914 | /* Don't re-enter if we're already allocating a chunk */ |
| 3915 | if (trans->allocating_chunk) | 3915 | if (trans->allocating_chunk) |
| 3916 | return -ENOSPC; | 3916 | return -ENOSPC; |
| 3917 | 3917 | ||
| 3918 | space_info = __find_space_info(extent_root->fs_info, flags); | 3918 | space_info = __find_space_info(extent_root->fs_info, flags); |
| 3919 | if (!space_info) { | 3919 | if (!space_info) { |
| 3920 | ret = update_space_info(extent_root->fs_info, flags, | 3920 | ret = update_space_info(extent_root->fs_info, flags, |
| 3921 | 0, 0, &space_info); | 3921 | 0, 0, &space_info); |
| 3922 | BUG_ON(ret); /* -ENOMEM */ | 3922 | BUG_ON(ret); /* -ENOMEM */ |
| 3923 | } | 3923 | } |
| 3924 | BUG_ON(!space_info); /* Logic error */ | 3924 | BUG_ON(!space_info); /* Logic error */ |
| 3925 | 3925 | ||
| 3926 | again: | 3926 | again: |
| 3927 | spin_lock(&space_info->lock); | 3927 | spin_lock(&space_info->lock); |
| 3928 | if (force < space_info->force_alloc) | 3928 | if (force < space_info->force_alloc) |
| 3929 | force = space_info->force_alloc; | 3929 | force = space_info->force_alloc; |
| 3930 | if (space_info->full) { | 3930 | if (space_info->full) { |
| 3931 | if (should_alloc_chunk(extent_root, space_info, force)) | 3931 | if (should_alloc_chunk(extent_root, space_info, force)) |
| 3932 | ret = -ENOSPC; | 3932 | ret = -ENOSPC; |
| 3933 | else | 3933 | else |
| 3934 | ret = 0; | 3934 | ret = 0; |
| 3935 | spin_unlock(&space_info->lock); | 3935 | spin_unlock(&space_info->lock); |
| 3936 | return ret; | 3936 | return ret; |
| 3937 | } | 3937 | } |
| 3938 | 3938 | ||
| 3939 | if (!should_alloc_chunk(extent_root, space_info, force)) { | 3939 | if (!should_alloc_chunk(extent_root, space_info, force)) { |
| 3940 | spin_unlock(&space_info->lock); | 3940 | spin_unlock(&space_info->lock); |
| 3941 | return 0; | 3941 | return 0; |
| 3942 | } else if (space_info->chunk_alloc) { | 3942 | } else if (space_info->chunk_alloc) { |
| 3943 | wait_for_alloc = 1; | 3943 | wait_for_alloc = 1; |
| 3944 | } else { | 3944 | } else { |
| 3945 | space_info->chunk_alloc = 1; | 3945 | space_info->chunk_alloc = 1; |
| 3946 | } | 3946 | } |
| 3947 | 3947 | ||
| 3948 | spin_unlock(&space_info->lock); | 3948 | spin_unlock(&space_info->lock); |
| 3949 | 3949 | ||
| 3950 | mutex_lock(&fs_info->chunk_mutex); | 3950 | mutex_lock(&fs_info->chunk_mutex); |
| 3951 | 3951 | ||
| 3952 | /* | 3952 | /* |
| 3953 | * The chunk_mutex is held throughout the entirety of a chunk | 3953 | * The chunk_mutex is held throughout the entirety of a chunk |
| 3954 | * allocation, so once we've acquired the chunk_mutex we know that the | 3954 | * allocation, so once we've acquired the chunk_mutex we know that the |
| 3955 | * other guy is done and we need to recheck and see if we should | 3955 | * other guy is done and we need to recheck and see if we should |
| 3956 | * allocate. | 3956 | * allocate. |
| 3957 | */ | 3957 | */ |
| 3958 | if (wait_for_alloc) { | 3958 | if (wait_for_alloc) { |
| 3959 | mutex_unlock(&fs_info->chunk_mutex); | 3959 | mutex_unlock(&fs_info->chunk_mutex); |
| 3960 | wait_for_alloc = 0; | 3960 | wait_for_alloc = 0; |
| 3961 | goto again; | 3961 | goto again; |
| 3962 | } | 3962 | } |
| 3963 | 3963 | ||
| 3964 | trans->allocating_chunk = true; | 3964 | trans->allocating_chunk = true; |
| 3965 | 3965 | ||
| 3966 | /* | 3966 | /* |
| 3967 | * If we have mixed data/metadata chunks we want to make sure we keep | 3967 | * If we have mixed data/metadata chunks we want to make sure we keep |
| 3968 | * allocating mixed chunks instead of individual chunks. | 3968 | * allocating mixed chunks instead of individual chunks. |
| 3969 | */ | 3969 | */ |
| 3970 | if (btrfs_mixed_space_info(space_info)) | 3970 | if (btrfs_mixed_space_info(space_info)) |
| 3971 | flags |= (BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA); | 3971 | flags |= (BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA); |
| 3972 | 3972 | ||
| 3973 | /* | 3973 | /* |
| 3974 | * if we're doing a data chunk, go ahead and make sure that | 3974 | * if we're doing a data chunk, go ahead and make sure that |
| 3975 | * we keep a reasonable number of metadata chunks allocated in the | 3975 | * we keep a reasonable number of metadata chunks allocated in the |
| 3976 | * FS as well. | 3976 | * FS as well. |
| 3977 | */ | 3977 | */ |
| 3978 | if (flags & BTRFS_BLOCK_GROUP_DATA && fs_info->metadata_ratio) { | 3978 | if (flags & BTRFS_BLOCK_GROUP_DATA && fs_info->metadata_ratio) { |
| 3979 | fs_info->data_chunk_allocations++; | 3979 | fs_info->data_chunk_allocations++; |
| 3980 | if (!(fs_info->data_chunk_allocations % | 3980 | if (!(fs_info->data_chunk_allocations % |
| 3981 | fs_info->metadata_ratio)) | 3981 | fs_info->metadata_ratio)) |
| 3982 | force_metadata_allocation(fs_info); | 3982 | force_metadata_allocation(fs_info); |
| 3983 | } | 3983 | } |
| 3984 | 3984 | ||
| 3985 | /* | 3985 | /* |
| 3986 | * Check if we have enough space in SYSTEM chunk because we may need | 3986 | * Check if we have enough space in SYSTEM chunk because we may need |
| 3987 | * to update devices. | 3987 | * to update devices. |
| 3988 | */ | 3988 | */ |
| 3989 | check_system_chunk(trans, extent_root, flags); | 3989 | check_system_chunk(trans, extent_root, flags); |
| 3990 | 3990 | ||
| 3991 | ret = btrfs_alloc_chunk(trans, extent_root, flags); | 3991 | ret = btrfs_alloc_chunk(trans, extent_root, flags); |
| 3992 | trans->allocating_chunk = false; | 3992 | trans->allocating_chunk = false; |
| 3993 | 3993 | ||
| 3994 | spin_lock(&space_info->lock); | 3994 | spin_lock(&space_info->lock); |
| 3995 | if (ret < 0 && ret != -ENOSPC) | 3995 | if (ret < 0 && ret != -ENOSPC) |
| 3996 | goto out; | 3996 | goto out; |
| 3997 | if (ret) | 3997 | if (ret) |
| 3998 | space_info->full = 1; | 3998 | space_info->full = 1; |
| 3999 | else | 3999 | else |
| 4000 | ret = 1; | 4000 | ret = 1; |
| 4001 | 4001 | ||
| 4002 | space_info->force_alloc = CHUNK_ALLOC_NO_FORCE; | 4002 | space_info->force_alloc = CHUNK_ALLOC_NO_FORCE; |
| 4003 | out: | 4003 | out: |
| 4004 | space_info->chunk_alloc = 0; | 4004 | space_info->chunk_alloc = 0; |
| 4005 | spin_unlock(&space_info->lock); | 4005 | spin_unlock(&space_info->lock); |
| 4006 | mutex_unlock(&fs_info->chunk_mutex); | 4006 | mutex_unlock(&fs_info->chunk_mutex); |
| 4007 | return ret; | 4007 | return ret; |
| 4008 | } | 4008 | } |
| 4009 | 4009 | ||
| 4010 | static int can_overcommit(struct btrfs_root *root, | 4010 | static int can_overcommit(struct btrfs_root *root, |
| 4011 | struct btrfs_space_info *space_info, u64 bytes, | 4011 | struct btrfs_space_info *space_info, u64 bytes, |
| 4012 | enum btrfs_reserve_flush_enum flush) | 4012 | enum btrfs_reserve_flush_enum flush) |
| 4013 | { | 4013 | { |
| 4014 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; | 4014 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; |
| 4015 | u64 profile = btrfs_get_alloc_profile(root, 0); | 4015 | u64 profile = btrfs_get_alloc_profile(root, 0); |
| 4016 | u64 space_size; | 4016 | u64 space_size; |
| 4017 | u64 avail; | 4017 | u64 avail; |
| 4018 | u64 used; | 4018 | u64 used; |
| 4019 | 4019 | ||
| 4020 | used = space_info->bytes_used + space_info->bytes_reserved + | 4020 | used = space_info->bytes_used + space_info->bytes_reserved + |
| 4021 | space_info->bytes_pinned + space_info->bytes_readonly; | 4021 | space_info->bytes_pinned + space_info->bytes_readonly; |
| 4022 | 4022 | ||
| 4023 | /* | 4023 | /* |
| 4024 | * We only want to allow over committing if we have lots of actual space | 4024 | * We only want to allow over committing if we have lots of actual space |
| 4025 | * free, but if we don't have enough space to handle the global reserve | 4025 | * free, but if we don't have enough space to handle the global reserve |
| 4026 | * space then we could end up having a real enospc problem when trying | 4026 | * space then we could end up having a real enospc problem when trying |
| 4027 | * to allocate a chunk or some other such important allocation. | 4027 | * to allocate a chunk or some other such important allocation. |
| 4028 | */ | 4028 | */ |
| 4029 | spin_lock(&global_rsv->lock); | 4029 | spin_lock(&global_rsv->lock); |
| 4030 | space_size = calc_global_rsv_need_space(global_rsv); | 4030 | space_size = calc_global_rsv_need_space(global_rsv); |
| 4031 | spin_unlock(&global_rsv->lock); | 4031 | spin_unlock(&global_rsv->lock); |
| 4032 | if (used + space_size >= space_info->total_bytes) | 4032 | if (used + space_size >= space_info->total_bytes) |
| 4033 | return 0; | 4033 | return 0; |
| 4034 | 4034 | ||
| 4035 | used += space_info->bytes_may_use; | 4035 | used += space_info->bytes_may_use; |
| 4036 | 4036 | ||
| 4037 | spin_lock(&root->fs_info->free_chunk_lock); | 4037 | spin_lock(&root->fs_info->free_chunk_lock); |
| 4038 | avail = root->fs_info->free_chunk_space; | 4038 | avail = root->fs_info->free_chunk_space; |
| 4039 | spin_unlock(&root->fs_info->free_chunk_lock); | 4039 | spin_unlock(&root->fs_info->free_chunk_lock); |
| 4040 | 4040 | ||
| 4041 | /* | 4041 | /* |
| 4042 | * If we have dup, raid1 or raid10 then only half of the free | 4042 | * If we have dup, raid1 or raid10 then only half of the free |
| 4043 | * space is actually useable. For raid56, the space info used | 4043 | * space is actually useable. For raid56, the space info used |
| 4044 | * doesn't include the parity drive, so we don't have to | 4044 | * doesn't include the parity drive, so we don't have to |
| 4045 | * change the math | 4045 | * change the math |
| 4046 | */ | 4046 | */ |
| 4047 | if (profile & (BTRFS_BLOCK_GROUP_DUP | | 4047 | if (profile & (BTRFS_BLOCK_GROUP_DUP | |
| 4048 | BTRFS_BLOCK_GROUP_RAID1 | | 4048 | BTRFS_BLOCK_GROUP_RAID1 | |
| 4049 | BTRFS_BLOCK_GROUP_RAID10)) | 4049 | BTRFS_BLOCK_GROUP_RAID10)) |
| 4050 | avail >>= 1; | 4050 | avail >>= 1; |
| 4051 | 4051 | ||
| 4052 | /* | 4052 | /* |
| 4053 | * If we aren't flushing all things, let us overcommit up to | 4053 | * If we aren't flushing all things, let us overcommit up to |
| 4054 | * 1/2th of the space. If we can flush, don't let us overcommit | 4054 | * 1/2th of the space. If we can flush, don't let us overcommit |
| 4055 | * too much, let it overcommit up to 1/8 of the space. | 4055 | * too much, let it overcommit up to 1/8 of the space. |
| 4056 | */ | 4056 | */ |
| 4057 | if (flush == BTRFS_RESERVE_FLUSH_ALL) | 4057 | if (flush == BTRFS_RESERVE_FLUSH_ALL) |
| 4058 | avail >>= 3; | 4058 | avail >>= 3; |
| 4059 | else | 4059 | else |
| 4060 | avail >>= 1; | 4060 | avail >>= 1; |
| 4061 | 4061 | ||
| 4062 | if (used + bytes < space_info->total_bytes + avail) | 4062 | if (used + bytes < space_info->total_bytes + avail) |
| 4063 | return 1; | 4063 | return 1; |
| 4064 | return 0; | 4064 | return 0; |
| 4065 | } | 4065 | } |
| 4066 | 4066 | ||
| 4067 | static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root, | 4067 | static void btrfs_writeback_inodes_sb_nr(struct btrfs_root *root, |
| 4068 | unsigned long nr_pages, int nr_items) | 4068 | unsigned long nr_pages, int nr_items) |
| 4069 | { | 4069 | { |
| 4070 | struct super_block *sb = root->fs_info->sb; | 4070 | struct super_block *sb = root->fs_info->sb; |
| 4071 | 4071 | ||
| 4072 | if (down_read_trylock(&sb->s_umount)) { | 4072 | if (down_read_trylock(&sb->s_umount)) { |
| 4073 | writeback_inodes_sb_nr(sb, nr_pages, WB_REASON_FS_FREE_SPACE); | 4073 | writeback_inodes_sb_nr(sb, nr_pages, WB_REASON_FS_FREE_SPACE); |
| 4074 | up_read(&sb->s_umount); | 4074 | up_read(&sb->s_umount); |
| 4075 | } else { | 4075 | } else { |
| 4076 | /* | 4076 | /* |
| 4077 | * We needn't worry the filesystem going from r/w to r/o though | 4077 | * We needn't worry the filesystem going from r/w to r/o though |
| 4078 | * we don't acquire ->s_umount mutex, because the filesystem | 4078 | * we don't acquire ->s_umount mutex, because the filesystem |
| 4079 | * should guarantee the delalloc inodes list be empty after | 4079 | * should guarantee the delalloc inodes list be empty after |
| 4080 | * the filesystem is readonly(all dirty pages are written to | 4080 | * the filesystem is readonly(all dirty pages are written to |
| 4081 | * the disk). | 4081 | * the disk). |
| 4082 | */ | 4082 | */ |
| 4083 | btrfs_start_delalloc_roots(root->fs_info, 0, nr_items); | 4083 | btrfs_start_delalloc_roots(root->fs_info, 0, nr_items); |
| 4084 | if (!current->journal_info) | 4084 | if (!current->journal_info) |
| 4085 | btrfs_wait_ordered_roots(root->fs_info, nr_items); | 4085 | btrfs_wait_ordered_roots(root->fs_info, nr_items); |
| 4086 | } | 4086 | } |
| 4087 | } | 4087 | } |
| 4088 | 4088 | ||
| 4089 | static inline int calc_reclaim_items_nr(struct btrfs_root *root, u64 to_reclaim) | 4089 | static inline int calc_reclaim_items_nr(struct btrfs_root *root, u64 to_reclaim) |
| 4090 | { | 4090 | { |
| 4091 | u64 bytes; | 4091 | u64 bytes; |
| 4092 | int nr; | 4092 | int nr; |
| 4093 | 4093 | ||
| 4094 | bytes = btrfs_calc_trans_metadata_size(root, 1); | 4094 | bytes = btrfs_calc_trans_metadata_size(root, 1); |
| 4095 | nr = (int)div64_u64(to_reclaim, bytes); | 4095 | nr = (int)div64_u64(to_reclaim, bytes); |
| 4096 | if (!nr) | 4096 | if (!nr) |
| 4097 | nr = 1; | 4097 | nr = 1; |
| 4098 | return nr; | 4098 | return nr; |
| 4099 | } | 4099 | } |
| 4100 | 4100 | ||
| 4101 | #define EXTENT_SIZE_PER_ITEM (256 * 1024) | 4101 | #define EXTENT_SIZE_PER_ITEM (256 * 1024) |
| 4102 | 4102 | ||
| 4103 | /* | 4103 | /* |
| 4104 | * shrink metadata reservation for delalloc | 4104 | * shrink metadata reservation for delalloc |
| 4105 | */ | 4105 | */ |
| 4106 | static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, | 4106 | static void shrink_delalloc(struct btrfs_root *root, u64 to_reclaim, u64 orig, |
| 4107 | bool wait_ordered) | 4107 | bool wait_ordered) |
| 4108 | { | 4108 | { |
| 4109 | struct btrfs_block_rsv *block_rsv; | 4109 | struct btrfs_block_rsv *block_rsv; |
| 4110 | struct btrfs_space_info *space_info; | 4110 | struct btrfs_space_info *space_info; |
| 4111 | struct btrfs_trans_handle *trans; | 4111 | struct btrfs_trans_handle *trans; |
| 4112 | u64 delalloc_bytes; | 4112 | u64 delalloc_bytes; |
| 4113 | u64 max_reclaim; | 4113 | u64 max_reclaim; |
| 4114 | long time_left; | 4114 | long time_left; |
| 4115 | unsigned long nr_pages; | 4115 | unsigned long nr_pages; |
| 4116 | int loops; | 4116 | int loops; |
| 4117 | int items; | 4117 | int items; |
| 4118 | enum btrfs_reserve_flush_enum flush; | 4118 | enum btrfs_reserve_flush_enum flush; |
| 4119 | 4119 | ||
| 4120 | /* Calc the number of the pages we need flush for space reservation */ | 4120 | /* Calc the number of the pages we need flush for space reservation */ |
| 4121 | items = calc_reclaim_items_nr(root, to_reclaim); | 4121 | items = calc_reclaim_items_nr(root, to_reclaim); |
| 4122 | to_reclaim = items * EXTENT_SIZE_PER_ITEM; | 4122 | to_reclaim = items * EXTENT_SIZE_PER_ITEM; |
| 4123 | 4123 | ||
| 4124 | trans = (struct btrfs_trans_handle *)current->journal_info; | 4124 | trans = (struct btrfs_trans_handle *)current->journal_info; |
| 4125 | block_rsv = &root->fs_info->delalloc_block_rsv; | 4125 | block_rsv = &root->fs_info->delalloc_block_rsv; |
| 4126 | space_info = block_rsv->space_info; | 4126 | space_info = block_rsv->space_info; |
| 4127 | 4127 | ||
| 4128 | delalloc_bytes = percpu_counter_sum_positive( | 4128 | delalloc_bytes = percpu_counter_sum_positive( |
| 4129 | &root->fs_info->delalloc_bytes); | 4129 | &root->fs_info->delalloc_bytes); |
| 4130 | if (delalloc_bytes == 0) { | 4130 | if (delalloc_bytes == 0) { |
| 4131 | if (trans) | 4131 | if (trans) |
| 4132 | return; | 4132 | return; |
| 4133 | if (wait_ordered) | 4133 | if (wait_ordered) |
| 4134 | btrfs_wait_ordered_roots(root->fs_info, items); | 4134 | btrfs_wait_ordered_roots(root->fs_info, items); |
| 4135 | return; | 4135 | return; |
| 4136 | } | 4136 | } |
| 4137 | 4137 | ||
| 4138 | loops = 0; | 4138 | loops = 0; |
| 4139 | while (delalloc_bytes && loops < 3) { | 4139 | while (delalloc_bytes && loops < 3) { |
| 4140 | max_reclaim = min(delalloc_bytes, to_reclaim); | 4140 | max_reclaim = min(delalloc_bytes, to_reclaim); |
| 4141 | nr_pages = max_reclaim >> PAGE_CACHE_SHIFT; | 4141 | nr_pages = max_reclaim >> PAGE_CACHE_SHIFT; |
| 4142 | btrfs_writeback_inodes_sb_nr(root, nr_pages, items); | 4142 | btrfs_writeback_inodes_sb_nr(root, nr_pages, items); |
| 4143 | /* | 4143 | /* |
| 4144 | * We need to wait for the async pages to actually start before | 4144 | * We need to wait for the async pages to actually start before |
| 4145 | * we do anything. | 4145 | * we do anything. |
| 4146 | */ | 4146 | */ |
| 4147 | max_reclaim = atomic_read(&root->fs_info->async_delalloc_pages); | 4147 | max_reclaim = atomic_read(&root->fs_info->async_delalloc_pages); |
| 4148 | if (!max_reclaim) | 4148 | if (!max_reclaim) |
| 4149 | goto skip_async; | 4149 | goto skip_async; |
| 4150 | 4150 | ||
| 4151 | if (max_reclaim <= nr_pages) | 4151 | if (max_reclaim <= nr_pages) |
| 4152 | max_reclaim = 0; | 4152 | max_reclaim = 0; |
| 4153 | else | 4153 | else |
| 4154 | max_reclaim -= nr_pages; | 4154 | max_reclaim -= nr_pages; |
| 4155 | 4155 | ||
| 4156 | wait_event(root->fs_info->async_submit_wait, | 4156 | wait_event(root->fs_info->async_submit_wait, |
| 4157 | atomic_read(&root->fs_info->async_delalloc_pages) <= | 4157 | atomic_read(&root->fs_info->async_delalloc_pages) <= |
| 4158 | (int)max_reclaim); | 4158 | (int)max_reclaim); |
| 4159 | skip_async: | 4159 | skip_async: |
| 4160 | if (!trans) | 4160 | if (!trans) |
| 4161 | flush = BTRFS_RESERVE_FLUSH_ALL; | 4161 | flush = BTRFS_RESERVE_FLUSH_ALL; |
| 4162 | else | 4162 | else |
| 4163 | flush = BTRFS_RESERVE_NO_FLUSH; | 4163 | flush = BTRFS_RESERVE_NO_FLUSH; |
| 4164 | spin_lock(&space_info->lock); | 4164 | spin_lock(&space_info->lock); |
| 4165 | if (can_overcommit(root, space_info, orig, flush)) { | 4165 | if (can_overcommit(root, space_info, orig, flush)) { |
| 4166 | spin_unlock(&space_info->lock); | 4166 | spin_unlock(&space_info->lock); |
| 4167 | break; | 4167 | break; |
| 4168 | } | 4168 | } |
| 4169 | spin_unlock(&space_info->lock); | 4169 | spin_unlock(&space_info->lock); |
| 4170 | 4170 | ||
| 4171 | loops++; | 4171 | loops++; |
| 4172 | if (wait_ordered && !trans) { | 4172 | if (wait_ordered && !trans) { |
| 4173 | btrfs_wait_ordered_roots(root->fs_info, items); | 4173 | btrfs_wait_ordered_roots(root->fs_info, items); |
| 4174 | } else { | 4174 | } else { |
| 4175 | time_left = schedule_timeout_killable(1); | 4175 | time_left = schedule_timeout_killable(1); |
| 4176 | if (time_left) | 4176 | if (time_left) |
| 4177 | break; | 4177 | break; |
| 4178 | } | 4178 | } |
| 4179 | delalloc_bytes = percpu_counter_sum_positive( | 4179 | delalloc_bytes = percpu_counter_sum_positive( |
| 4180 | &root->fs_info->delalloc_bytes); | 4180 | &root->fs_info->delalloc_bytes); |
| 4181 | } | 4181 | } |
| 4182 | } | 4182 | } |
| 4183 | 4183 | ||
| 4184 | /** | 4184 | /** |
| 4185 | * maybe_commit_transaction - possibly commit the transaction if its ok to | 4185 | * maybe_commit_transaction - possibly commit the transaction if its ok to |
| 4186 | * @root - the root we're allocating for | 4186 | * @root - the root we're allocating for |
| 4187 | * @bytes - the number of bytes we want to reserve | 4187 | * @bytes - the number of bytes we want to reserve |
| 4188 | * @force - force the commit | 4188 | * @force - force the commit |
| 4189 | * | 4189 | * |
| 4190 | * This will check to make sure that committing the transaction will actually | 4190 | * This will check to make sure that committing the transaction will actually |
| 4191 | * get us somewhere and then commit the transaction if it does. Otherwise it | 4191 | * get us somewhere and then commit the transaction if it does. Otherwise it |
| 4192 | * will return -ENOSPC. | 4192 | * will return -ENOSPC. |
| 4193 | */ | 4193 | */ |
| 4194 | static int may_commit_transaction(struct btrfs_root *root, | 4194 | static int may_commit_transaction(struct btrfs_root *root, |
| 4195 | struct btrfs_space_info *space_info, | 4195 | struct btrfs_space_info *space_info, |
| 4196 | u64 bytes, int force) | 4196 | u64 bytes, int force) |
| 4197 | { | 4197 | { |
| 4198 | struct btrfs_block_rsv *delayed_rsv = &root->fs_info->delayed_block_rsv; | 4198 | struct btrfs_block_rsv *delayed_rsv = &root->fs_info->delayed_block_rsv; |
| 4199 | struct btrfs_trans_handle *trans; | 4199 | struct btrfs_trans_handle *trans; |
| 4200 | 4200 | ||
| 4201 | trans = (struct btrfs_trans_handle *)current->journal_info; | 4201 | trans = (struct btrfs_trans_handle *)current->journal_info; |
| 4202 | if (trans) | 4202 | if (trans) |
| 4203 | return -EAGAIN; | 4203 | return -EAGAIN; |
| 4204 | 4204 | ||
| 4205 | if (force) | 4205 | if (force) |
| 4206 | goto commit; | 4206 | goto commit; |
| 4207 | 4207 | ||
| 4208 | /* See if there is enough pinned space to make this reservation */ | 4208 | /* See if there is enough pinned space to make this reservation */ |
| 4209 | if (percpu_counter_compare(&space_info->total_bytes_pinned, | 4209 | if (percpu_counter_compare(&space_info->total_bytes_pinned, |
| 4210 | bytes) >= 0) | 4210 | bytes) >= 0) |
| 4211 | goto commit; | 4211 | goto commit; |
| 4212 | 4212 | ||
| 4213 | /* | 4213 | /* |
| 4214 | * See if there is some space in the delayed insertion reservation for | 4214 | * See if there is some space in the delayed insertion reservation for |
| 4215 | * this reservation. | 4215 | * this reservation. |
| 4216 | */ | 4216 | */ |
| 4217 | if (space_info != delayed_rsv->space_info) | 4217 | if (space_info != delayed_rsv->space_info) |
| 4218 | return -ENOSPC; | 4218 | return -ENOSPC; |
| 4219 | 4219 | ||
| 4220 | spin_lock(&delayed_rsv->lock); | 4220 | spin_lock(&delayed_rsv->lock); |
| 4221 | if (percpu_counter_compare(&space_info->total_bytes_pinned, | 4221 | if (percpu_counter_compare(&space_info->total_bytes_pinned, |
| 4222 | bytes - delayed_rsv->size) >= 0) { | 4222 | bytes - delayed_rsv->size) >= 0) { |
| 4223 | spin_unlock(&delayed_rsv->lock); | 4223 | spin_unlock(&delayed_rsv->lock); |
| 4224 | return -ENOSPC; | 4224 | return -ENOSPC; |
| 4225 | } | 4225 | } |
| 4226 | spin_unlock(&delayed_rsv->lock); | 4226 | spin_unlock(&delayed_rsv->lock); |
| 4227 | 4227 | ||
| 4228 | commit: | 4228 | commit: |
| 4229 | trans = btrfs_join_transaction(root); | 4229 | trans = btrfs_join_transaction(root); |
| 4230 | if (IS_ERR(trans)) | 4230 | if (IS_ERR(trans)) |
| 4231 | return -ENOSPC; | 4231 | return -ENOSPC; |
| 4232 | 4232 | ||
| 4233 | return btrfs_commit_transaction(trans, root); | 4233 | return btrfs_commit_transaction(trans, root); |
| 4234 | } | 4234 | } |
| 4235 | 4235 | ||
| 4236 | enum flush_state { | 4236 | enum flush_state { |
| 4237 | FLUSH_DELAYED_ITEMS_NR = 1, | 4237 | FLUSH_DELAYED_ITEMS_NR = 1, |
| 4238 | FLUSH_DELAYED_ITEMS = 2, | 4238 | FLUSH_DELAYED_ITEMS = 2, |
| 4239 | FLUSH_DELALLOC = 3, | 4239 | FLUSH_DELALLOC = 3, |
| 4240 | FLUSH_DELALLOC_WAIT = 4, | 4240 | FLUSH_DELALLOC_WAIT = 4, |
| 4241 | ALLOC_CHUNK = 5, | 4241 | ALLOC_CHUNK = 5, |
| 4242 | COMMIT_TRANS = 6, | 4242 | COMMIT_TRANS = 6, |
| 4243 | }; | 4243 | }; |
| 4244 | 4244 | ||
| 4245 | static int flush_space(struct btrfs_root *root, | 4245 | static int flush_space(struct btrfs_root *root, |
| 4246 | struct btrfs_space_info *space_info, u64 num_bytes, | 4246 | struct btrfs_space_info *space_info, u64 num_bytes, |
| 4247 | u64 orig_bytes, int state) | 4247 | u64 orig_bytes, int state) |
| 4248 | { | 4248 | { |
| 4249 | struct btrfs_trans_handle *trans; | 4249 | struct btrfs_trans_handle *trans; |
| 4250 | int nr; | 4250 | int nr; |
| 4251 | int ret = 0; | 4251 | int ret = 0; |
| 4252 | 4252 | ||
| 4253 | switch (state) { | 4253 | switch (state) { |
| 4254 | case FLUSH_DELAYED_ITEMS_NR: | 4254 | case FLUSH_DELAYED_ITEMS_NR: |
| 4255 | case FLUSH_DELAYED_ITEMS: | 4255 | case FLUSH_DELAYED_ITEMS: |
| 4256 | if (state == FLUSH_DELAYED_ITEMS_NR) | 4256 | if (state == FLUSH_DELAYED_ITEMS_NR) |
| 4257 | nr = calc_reclaim_items_nr(root, num_bytes) * 2; | 4257 | nr = calc_reclaim_items_nr(root, num_bytes) * 2; |
| 4258 | else | 4258 | else |
| 4259 | nr = -1; | 4259 | nr = -1; |
| 4260 | 4260 | ||
| 4261 | trans = btrfs_join_transaction(root); | 4261 | trans = btrfs_join_transaction(root); |
| 4262 | if (IS_ERR(trans)) { | 4262 | if (IS_ERR(trans)) { |
| 4263 | ret = PTR_ERR(trans); | 4263 | ret = PTR_ERR(trans); |
| 4264 | break; | 4264 | break; |
| 4265 | } | 4265 | } |
| 4266 | ret = btrfs_run_delayed_items_nr(trans, root, nr); | 4266 | ret = btrfs_run_delayed_items_nr(trans, root, nr); |
| 4267 | btrfs_end_transaction(trans, root); | 4267 | btrfs_end_transaction(trans, root); |
| 4268 | break; | 4268 | break; |
| 4269 | case FLUSH_DELALLOC: | 4269 | case FLUSH_DELALLOC: |
| 4270 | case FLUSH_DELALLOC_WAIT: | 4270 | case FLUSH_DELALLOC_WAIT: |
| 4271 | shrink_delalloc(root, num_bytes * 2, orig_bytes, | 4271 | shrink_delalloc(root, num_bytes * 2, orig_bytes, |
| 4272 | state == FLUSH_DELALLOC_WAIT); | 4272 | state == FLUSH_DELALLOC_WAIT); |
| 4273 | break; | 4273 | break; |
| 4274 | case ALLOC_CHUNK: | 4274 | case ALLOC_CHUNK: |
| 4275 | trans = btrfs_join_transaction(root); | 4275 | trans = btrfs_join_transaction(root); |
| 4276 | if (IS_ERR(trans)) { | 4276 | if (IS_ERR(trans)) { |
| 4277 | ret = PTR_ERR(trans); | 4277 | ret = PTR_ERR(trans); |
| 4278 | break; | 4278 | break; |
| 4279 | } | 4279 | } |
| 4280 | ret = do_chunk_alloc(trans, root->fs_info->extent_root, | 4280 | ret = do_chunk_alloc(trans, root->fs_info->extent_root, |
| 4281 | btrfs_get_alloc_profile(root, 0), | 4281 | btrfs_get_alloc_profile(root, 0), |
| 4282 | CHUNK_ALLOC_NO_FORCE); | 4282 | CHUNK_ALLOC_NO_FORCE); |
| 4283 | btrfs_end_transaction(trans, root); | 4283 | btrfs_end_transaction(trans, root); |
| 4284 | if (ret == -ENOSPC) | 4284 | if (ret == -ENOSPC) |
| 4285 | ret = 0; | 4285 | ret = 0; |
| 4286 | break; | 4286 | break; |
| 4287 | case COMMIT_TRANS: | 4287 | case COMMIT_TRANS: |
| 4288 | ret = may_commit_transaction(root, space_info, orig_bytes, 0); | 4288 | ret = may_commit_transaction(root, space_info, orig_bytes, 0); |
| 4289 | break; | 4289 | break; |
| 4290 | default: | 4290 | default: |
| 4291 | ret = -ENOSPC; | 4291 | ret = -ENOSPC; |
| 4292 | break; | 4292 | break; |
| 4293 | } | 4293 | } |
| 4294 | 4294 | ||
| 4295 | return ret; | 4295 | return ret; |
| 4296 | } | 4296 | } |
| 4297 | 4297 | ||
| 4298 | static inline u64 | 4298 | static inline u64 |
| 4299 | btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, | 4299 | btrfs_calc_reclaim_metadata_size(struct btrfs_root *root, |
| 4300 | struct btrfs_space_info *space_info) | 4300 | struct btrfs_space_info *space_info) |
| 4301 | { | 4301 | { |
| 4302 | u64 used; | 4302 | u64 used; |
| 4303 | u64 expected; | 4303 | u64 expected; |
| 4304 | u64 to_reclaim; | 4304 | u64 to_reclaim; |
| 4305 | 4305 | ||
| 4306 | to_reclaim = min_t(u64, num_online_cpus() * 1024 * 1024, | 4306 | to_reclaim = min_t(u64, num_online_cpus() * 1024 * 1024, |
| 4307 | 16 * 1024 * 1024); | 4307 | 16 * 1024 * 1024); |
| 4308 | spin_lock(&space_info->lock); | 4308 | spin_lock(&space_info->lock); |
| 4309 | if (can_overcommit(root, space_info, to_reclaim, | 4309 | if (can_overcommit(root, space_info, to_reclaim, |
| 4310 | BTRFS_RESERVE_FLUSH_ALL)) { | 4310 | BTRFS_RESERVE_FLUSH_ALL)) { |
| 4311 | to_reclaim = 0; | 4311 | to_reclaim = 0; |
| 4312 | goto out; | 4312 | goto out; |
| 4313 | } | 4313 | } |
| 4314 | 4314 | ||
| 4315 | used = space_info->bytes_used + space_info->bytes_reserved + | 4315 | used = space_info->bytes_used + space_info->bytes_reserved + |
| 4316 | space_info->bytes_pinned + space_info->bytes_readonly + | 4316 | space_info->bytes_pinned + space_info->bytes_readonly + |
| 4317 | space_info->bytes_may_use; | 4317 | space_info->bytes_may_use; |
| 4318 | if (can_overcommit(root, space_info, 1024 * 1024, | 4318 | if (can_overcommit(root, space_info, 1024 * 1024, |
| 4319 | BTRFS_RESERVE_FLUSH_ALL)) | 4319 | BTRFS_RESERVE_FLUSH_ALL)) |
| 4320 | expected = div_factor_fine(space_info->total_bytes, 95); | 4320 | expected = div_factor_fine(space_info->total_bytes, 95); |
| 4321 | else | 4321 | else |
| 4322 | expected = div_factor_fine(space_info->total_bytes, 90); | 4322 | expected = div_factor_fine(space_info->total_bytes, 90); |
| 4323 | 4323 | ||
| 4324 | if (used > expected) | 4324 | if (used > expected) |
| 4325 | to_reclaim = used - expected; | 4325 | to_reclaim = used - expected; |
| 4326 | else | 4326 | else |
| 4327 | to_reclaim = 0; | 4327 | to_reclaim = 0; |
| 4328 | to_reclaim = min(to_reclaim, space_info->bytes_may_use + | 4328 | to_reclaim = min(to_reclaim, space_info->bytes_may_use + |
| 4329 | space_info->bytes_reserved); | 4329 | space_info->bytes_reserved); |
| 4330 | out: | 4330 | out: |
| 4331 | spin_unlock(&space_info->lock); | 4331 | spin_unlock(&space_info->lock); |
| 4332 | 4332 | ||
| 4333 | return to_reclaim; | 4333 | return to_reclaim; |
| 4334 | } | 4334 | } |
| 4335 | 4335 | ||
| 4336 | static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, | 4336 | static inline int need_do_async_reclaim(struct btrfs_space_info *space_info, |
| 4337 | struct btrfs_fs_info *fs_info, u64 used) | 4337 | struct btrfs_fs_info *fs_info, u64 used) |
| 4338 | { | 4338 | { |
| 4339 | return (used >= div_factor_fine(space_info->total_bytes, 98) && | 4339 | return (used >= div_factor_fine(space_info->total_bytes, 98) && |
| 4340 | !btrfs_fs_closing(fs_info) && | 4340 | !btrfs_fs_closing(fs_info) && |
| 4341 | !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); | 4341 | !test_bit(BTRFS_FS_STATE_REMOUNTING, &fs_info->fs_state)); |
| 4342 | } | 4342 | } |
| 4343 | 4343 | ||
| 4344 | static int btrfs_need_do_async_reclaim(struct btrfs_space_info *space_info, | 4344 | static int btrfs_need_do_async_reclaim(struct btrfs_space_info *space_info, |
| 4345 | struct btrfs_fs_info *fs_info, | 4345 | struct btrfs_fs_info *fs_info, |
| 4346 | int flush_state) | 4346 | int flush_state) |
| 4347 | { | 4347 | { |
| 4348 | u64 used; | 4348 | u64 used; |
| 4349 | 4349 | ||
| 4350 | spin_lock(&space_info->lock); | 4350 | spin_lock(&space_info->lock); |
| 4351 | /* | 4351 | /* |
| 4352 | * We run out of space and have not got any free space via flush_space, | 4352 | * We run out of space and have not got any free space via flush_space, |
| 4353 | * so don't bother doing async reclaim. | 4353 | * so don't bother doing async reclaim. |
| 4354 | */ | 4354 | */ |
| 4355 | if (flush_state > COMMIT_TRANS && space_info->full) { | 4355 | if (flush_state > COMMIT_TRANS && space_info->full) { |
| 4356 | spin_unlock(&space_info->lock); | 4356 | spin_unlock(&space_info->lock); |
| 4357 | return 0; | 4357 | return 0; |
| 4358 | } | 4358 | } |
| 4359 | 4359 | ||
| 4360 | used = space_info->bytes_used + space_info->bytes_reserved + | 4360 | used = space_info->bytes_used + space_info->bytes_reserved + |
| 4361 | space_info->bytes_pinned + space_info->bytes_readonly + | 4361 | space_info->bytes_pinned + space_info->bytes_readonly + |
| 4362 | space_info->bytes_may_use; | 4362 | space_info->bytes_may_use; |
| 4363 | if (need_do_async_reclaim(space_info, fs_info, used)) { | 4363 | if (need_do_async_reclaim(space_info, fs_info, used)) { |
| 4364 | spin_unlock(&space_info->lock); | 4364 | spin_unlock(&space_info->lock); |
| 4365 | return 1; | 4365 | return 1; |
| 4366 | } | 4366 | } |
| 4367 | spin_unlock(&space_info->lock); | 4367 | spin_unlock(&space_info->lock); |
| 4368 | 4368 | ||
| 4369 | return 0; | 4369 | return 0; |
| 4370 | } | 4370 | } |
| 4371 | 4371 | ||
| 4372 | static void btrfs_async_reclaim_metadata_space(struct work_struct *work) | 4372 | static void btrfs_async_reclaim_metadata_space(struct work_struct *work) |
| 4373 | { | 4373 | { |
| 4374 | struct btrfs_fs_info *fs_info; | 4374 | struct btrfs_fs_info *fs_info; |
| 4375 | struct btrfs_space_info *space_info; | 4375 | struct btrfs_space_info *space_info; |
| 4376 | u64 to_reclaim; | 4376 | u64 to_reclaim; |
| 4377 | int flush_state; | 4377 | int flush_state; |
| 4378 | 4378 | ||
| 4379 | fs_info = container_of(work, struct btrfs_fs_info, async_reclaim_work); | 4379 | fs_info = container_of(work, struct btrfs_fs_info, async_reclaim_work); |
| 4380 | space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); | 4380 | space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); |
| 4381 | 4381 | ||
| 4382 | to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, | 4382 | to_reclaim = btrfs_calc_reclaim_metadata_size(fs_info->fs_root, |
| 4383 | space_info); | 4383 | space_info); |
| 4384 | if (!to_reclaim) | 4384 | if (!to_reclaim) |
| 4385 | return; | 4385 | return; |
| 4386 | 4386 | ||
| 4387 | flush_state = FLUSH_DELAYED_ITEMS_NR; | 4387 | flush_state = FLUSH_DELAYED_ITEMS_NR; |
| 4388 | do { | 4388 | do { |
| 4389 | flush_space(fs_info->fs_root, space_info, to_reclaim, | 4389 | flush_space(fs_info->fs_root, space_info, to_reclaim, |
| 4390 | to_reclaim, flush_state); | 4390 | to_reclaim, flush_state); |
| 4391 | flush_state++; | 4391 | flush_state++; |
| 4392 | if (!btrfs_need_do_async_reclaim(space_info, fs_info, | 4392 | if (!btrfs_need_do_async_reclaim(space_info, fs_info, |
| 4393 | flush_state)) | 4393 | flush_state)) |
| 4394 | return; | 4394 | return; |
| 4395 | } while (flush_state <= COMMIT_TRANS); | 4395 | } while (flush_state <= COMMIT_TRANS); |
| 4396 | 4396 | ||
| 4397 | if (btrfs_need_do_async_reclaim(space_info, fs_info, flush_state)) | 4397 | if (btrfs_need_do_async_reclaim(space_info, fs_info, flush_state)) |
| 4398 | queue_work(system_unbound_wq, work); | 4398 | queue_work(system_unbound_wq, work); |
| 4399 | } | 4399 | } |
| 4400 | 4400 | ||
| 4401 | void btrfs_init_async_reclaim_work(struct work_struct *work) | 4401 | void btrfs_init_async_reclaim_work(struct work_struct *work) |
| 4402 | { | 4402 | { |
| 4403 | INIT_WORK(work, btrfs_async_reclaim_metadata_space); | 4403 | INIT_WORK(work, btrfs_async_reclaim_metadata_space); |
| 4404 | } | 4404 | } |
| 4405 | 4405 | ||
| 4406 | /** | 4406 | /** |
| 4407 | * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space | 4407 | * reserve_metadata_bytes - try to reserve bytes from the block_rsv's space |
| 4408 | * @root - the root we're allocating for | 4408 | * @root - the root we're allocating for |
| 4409 | * @block_rsv - the block_rsv we're allocating for | 4409 | * @block_rsv - the block_rsv we're allocating for |
| 4410 | * @orig_bytes - the number of bytes we want | 4410 | * @orig_bytes - the number of bytes we want |
| 4411 | * @flush - whether or not we can flush to make our reservation | 4411 | * @flush - whether or not we can flush to make our reservation |
| 4412 | * | 4412 | * |
| 4413 | * This will reserve orgi_bytes number of bytes from the space info associated | 4413 | * This will reserve orgi_bytes number of bytes from the space info associated |
| 4414 | * with the block_rsv. If there is not enough space it will make an attempt to | 4414 | * with the block_rsv. If there is not enough space it will make an attempt to |
| 4415 | * flush out space to make room. It will do this by flushing delalloc if | 4415 | * flush out space to make room. It will do this by flushing delalloc if |
| 4416 | * possible or committing the transaction. If flush is 0 then no attempts to | 4416 | * possible or committing the transaction. If flush is 0 then no attempts to |
| 4417 | * regain reservations will be made and this will fail if there is not enough | 4417 | * regain reservations will be made and this will fail if there is not enough |
| 4418 | * space already. | 4418 | * space already. |
| 4419 | */ | 4419 | */ |
| 4420 | static int reserve_metadata_bytes(struct btrfs_root *root, | 4420 | static int reserve_metadata_bytes(struct btrfs_root *root, |
| 4421 | struct btrfs_block_rsv *block_rsv, | 4421 | struct btrfs_block_rsv *block_rsv, |
| 4422 | u64 orig_bytes, | 4422 | u64 orig_bytes, |
| 4423 | enum btrfs_reserve_flush_enum flush) | 4423 | enum btrfs_reserve_flush_enum flush) |
| 4424 | { | 4424 | { |
| 4425 | struct btrfs_space_info *space_info = block_rsv->space_info; | 4425 | struct btrfs_space_info *space_info = block_rsv->space_info; |
| 4426 | u64 used; | 4426 | u64 used; |
| 4427 | u64 num_bytes = orig_bytes; | 4427 | u64 num_bytes = orig_bytes; |
| 4428 | int flush_state = FLUSH_DELAYED_ITEMS_NR; | 4428 | int flush_state = FLUSH_DELAYED_ITEMS_NR; |
| 4429 | int ret = 0; | 4429 | int ret = 0; |
| 4430 | bool flushing = false; | 4430 | bool flushing = false; |
| 4431 | 4431 | ||
| 4432 | again: | 4432 | again: |
| 4433 | ret = 0; | 4433 | ret = 0; |
| 4434 | spin_lock(&space_info->lock); | 4434 | spin_lock(&space_info->lock); |
| 4435 | /* | 4435 | /* |
| 4436 | * We only want to wait if somebody other than us is flushing and we | 4436 | * We only want to wait if somebody other than us is flushing and we |
| 4437 | * are actually allowed to flush all things. | 4437 | * are actually allowed to flush all things. |
| 4438 | */ | 4438 | */ |
| 4439 | while (flush == BTRFS_RESERVE_FLUSH_ALL && !flushing && | 4439 | while (flush == BTRFS_RESERVE_FLUSH_ALL && !flushing && |
| 4440 | space_info->flush) { | 4440 | space_info->flush) { |
| 4441 | spin_unlock(&space_info->lock); | 4441 | spin_unlock(&space_info->lock); |
| 4442 | /* | 4442 | /* |
| 4443 | * If we have a trans handle we can't wait because the flusher | 4443 | * If we have a trans handle we can't wait because the flusher |
| 4444 | * may have to commit the transaction, which would mean we would | 4444 | * may have to commit the transaction, which would mean we would |
| 4445 | * deadlock since we are waiting for the flusher to finish, but | 4445 | * deadlock since we are waiting for the flusher to finish, but |
| 4446 | * hold the current transaction open. | 4446 | * hold the current transaction open. |
| 4447 | */ | 4447 | */ |
| 4448 | if (current->journal_info) | 4448 | if (current->journal_info) |
| 4449 | return -EAGAIN; | 4449 | return -EAGAIN; |
| 4450 | ret = wait_event_killable(space_info->wait, !space_info->flush); | 4450 | ret = wait_event_killable(space_info->wait, !space_info->flush); |
| 4451 | /* Must have been killed, return */ | 4451 | /* Must have been killed, return */ |
| 4452 | if (ret) | 4452 | if (ret) |
| 4453 | return -EINTR; | 4453 | return -EINTR; |
| 4454 | 4454 | ||
| 4455 | spin_lock(&space_info->lock); | 4455 | spin_lock(&space_info->lock); |
| 4456 | } | 4456 | } |
| 4457 | 4457 | ||
| 4458 | ret = -ENOSPC; | 4458 | ret = -ENOSPC; |
| 4459 | used = space_info->bytes_used + space_info->bytes_reserved + | 4459 | used = space_info->bytes_used + space_info->bytes_reserved + |
| 4460 | space_info->bytes_pinned + space_info->bytes_readonly + | 4460 | space_info->bytes_pinned + space_info->bytes_readonly + |
| 4461 | space_info->bytes_may_use; | 4461 | space_info->bytes_may_use; |
| 4462 | 4462 | ||
| 4463 | /* | 4463 | /* |
| 4464 | * The idea here is that we've not already over-reserved the block group | 4464 | * The idea here is that we've not already over-reserved the block group |
| 4465 | * then we can go ahead and save our reservation first and then start | 4465 | * then we can go ahead and save our reservation first and then start |
| 4466 | * flushing if we need to. Otherwise if we've already overcommitted | 4466 | * flushing if we need to. Otherwise if we've already overcommitted |
| 4467 | * lets start flushing stuff first and then come back and try to make | 4467 | * lets start flushing stuff first and then come back and try to make |
| 4468 | * our reservation. | 4468 | * our reservation. |
| 4469 | */ | 4469 | */ |
| 4470 | if (used <= space_info->total_bytes) { | 4470 | if (used <= space_info->total_bytes) { |
| 4471 | if (used + orig_bytes <= space_info->total_bytes) { | 4471 | if (used + orig_bytes <= space_info->total_bytes) { |
| 4472 | space_info->bytes_may_use += orig_bytes; | 4472 | space_info->bytes_may_use += orig_bytes; |
| 4473 | trace_btrfs_space_reservation(root->fs_info, | 4473 | trace_btrfs_space_reservation(root->fs_info, |
| 4474 | "space_info", space_info->flags, orig_bytes, 1); | 4474 | "space_info", space_info->flags, orig_bytes, 1); |
| 4475 | ret = 0; | 4475 | ret = 0; |
| 4476 | } else { | 4476 | } else { |
| 4477 | /* | 4477 | /* |
| 4478 | * Ok set num_bytes to orig_bytes since we aren't | 4478 | * Ok set num_bytes to orig_bytes since we aren't |
| 4479 | * overocmmitted, this way we only try and reclaim what | 4479 | * overocmmitted, this way we only try and reclaim what |
| 4480 | * we need. | 4480 | * we need. |
| 4481 | */ | 4481 | */ |
| 4482 | num_bytes = orig_bytes; | 4482 | num_bytes = orig_bytes; |
| 4483 | } | 4483 | } |
| 4484 | } else { | 4484 | } else { |
| 4485 | /* | 4485 | /* |
| 4486 | * Ok we're over committed, set num_bytes to the overcommitted | 4486 | * Ok we're over committed, set num_bytes to the overcommitted |
| 4487 | * amount plus the amount of bytes that we need for this | 4487 | * amount plus the amount of bytes that we need for this |
| 4488 | * reservation. | 4488 | * reservation. |
| 4489 | */ | 4489 | */ |
| 4490 | num_bytes = used - space_info->total_bytes + | 4490 | num_bytes = used - space_info->total_bytes + |
| 4491 | (orig_bytes * 2); | 4491 | (orig_bytes * 2); |
| 4492 | } | 4492 | } |
| 4493 | 4493 | ||
| 4494 | if (ret && can_overcommit(root, space_info, orig_bytes, flush)) { | 4494 | if (ret && can_overcommit(root, space_info, orig_bytes, flush)) { |
| 4495 | space_info->bytes_may_use += orig_bytes; | 4495 | space_info->bytes_may_use += orig_bytes; |
| 4496 | trace_btrfs_space_reservation(root->fs_info, "space_info", | 4496 | trace_btrfs_space_reservation(root->fs_info, "space_info", |
| 4497 | space_info->flags, orig_bytes, | 4497 | space_info->flags, orig_bytes, |
| 4498 | 1); | 4498 | 1); |
| 4499 | ret = 0; | 4499 | ret = 0; |
| 4500 | } | 4500 | } |
| 4501 | 4501 | ||
| 4502 | /* | 4502 | /* |
| 4503 | * Couldn't make our reservation, save our place so while we're trying | 4503 | * Couldn't make our reservation, save our place so while we're trying |
| 4504 | * to reclaim space we can actually use it instead of somebody else | 4504 | * to reclaim space we can actually use it instead of somebody else |
| 4505 | * stealing it from us. | 4505 | * stealing it from us. |
| 4506 | * | 4506 | * |
| 4507 | * We make the other tasks wait for the flush only when we can flush | 4507 | * We make the other tasks wait for the flush only when we can flush |
| 4508 | * all things. | 4508 | * all things. |
| 4509 | */ | 4509 | */ |
| 4510 | if (ret && flush != BTRFS_RESERVE_NO_FLUSH) { | 4510 | if (ret && flush != BTRFS_RESERVE_NO_FLUSH) { |
| 4511 | flushing = true; | 4511 | flushing = true; |
| 4512 | space_info->flush = 1; | 4512 | space_info->flush = 1; |
| 4513 | } else if (!ret && space_info->flags & BTRFS_BLOCK_GROUP_METADATA) { | 4513 | } else if (!ret && space_info->flags & BTRFS_BLOCK_GROUP_METADATA) { |
| 4514 | used += orig_bytes; | 4514 | used += orig_bytes; |
| 4515 | /* | 4515 | /* |
| 4516 | * We will do the space reservation dance during log replay, | 4516 | * We will do the space reservation dance during log replay, |
| 4517 | * which means we won't have fs_info->fs_root set, so don't do | 4517 | * which means we won't have fs_info->fs_root set, so don't do |
| 4518 | * the async reclaim as we will panic. | 4518 | * the async reclaim as we will panic. |
| 4519 | */ | 4519 | */ |
| 4520 | if (!root->fs_info->log_root_recovering && | 4520 | if (!root->fs_info->log_root_recovering && |
| 4521 | need_do_async_reclaim(space_info, root->fs_info, used) && | 4521 | need_do_async_reclaim(space_info, root->fs_info, used) && |
| 4522 | !work_busy(&root->fs_info->async_reclaim_work)) | 4522 | !work_busy(&root->fs_info->async_reclaim_work)) |
| 4523 | queue_work(system_unbound_wq, | 4523 | queue_work(system_unbound_wq, |
| 4524 | &root->fs_info->async_reclaim_work); | 4524 | &root->fs_info->async_reclaim_work); |
| 4525 | } | 4525 | } |
| 4526 | spin_unlock(&space_info->lock); | 4526 | spin_unlock(&space_info->lock); |
| 4527 | 4527 | ||
| 4528 | if (!ret || flush == BTRFS_RESERVE_NO_FLUSH) | 4528 | if (!ret || flush == BTRFS_RESERVE_NO_FLUSH) |
| 4529 | goto out; | 4529 | goto out; |
| 4530 | 4530 | ||
| 4531 | ret = flush_space(root, space_info, num_bytes, orig_bytes, | 4531 | ret = flush_space(root, space_info, num_bytes, orig_bytes, |
| 4532 | flush_state); | 4532 | flush_state); |
| 4533 | flush_state++; | 4533 | flush_state++; |
| 4534 | 4534 | ||
| 4535 | /* | 4535 | /* |
| 4536 | * If we are FLUSH_LIMIT, we can not flush delalloc, or the deadlock | 4536 | * If we are FLUSH_LIMIT, we can not flush delalloc, or the deadlock |
| 4537 | * would happen. So skip delalloc flush. | 4537 | * would happen. So skip delalloc flush. |
| 4538 | */ | 4538 | */ |
| 4539 | if (flush == BTRFS_RESERVE_FLUSH_LIMIT && | 4539 | if (flush == BTRFS_RESERVE_FLUSH_LIMIT && |
| 4540 | (flush_state == FLUSH_DELALLOC || | 4540 | (flush_state == FLUSH_DELALLOC || |
| 4541 | flush_state == FLUSH_DELALLOC_WAIT)) | 4541 | flush_state == FLUSH_DELALLOC_WAIT)) |
| 4542 | flush_state = ALLOC_CHUNK; | 4542 | flush_state = ALLOC_CHUNK; |
| 4543 | 4543 | ||
| 4544 | if (!ret) | 4544 | if (!ret) |
| 4545 | goto again; | 4545 | goto again; |
| 4546 | else if (flush == BTRFS_RESERVE_FLUSH_LIMIT && | 4546 | else if (flush == BTRFS_RESERVE_FLUSH_LIMIT && |
| 4547 | flush_state < COMMIT_TRANS) | 4547 | flush_state < COMMIT_TRANS) |
| 4548 | goto again; | 4548 | goto again; |
| 4549 | else if (flush == BTRFS_RESERVE_FLUSH_ALL && | 4549 | else if (flush == BTRFS_RESERVE_FLUSH_ALL && |
| 4550 | flush_state <= COMMIT_TRANS) | 4550 | flush_state <= COMMIT_TRANS) |
| 4551 | goto again; | 4551 | goto again; |
| 4552 | 4552 | ||
| 4553 | out: | 4553 | out: |
| 4554 | if (ret == -ENOSPC && | 4554 | if (ret == -ENOSPC && |
| 4555 | unlikely(root->orphan_cleanup_state == ORPHAN_CLEANUP_STARTED)) { | 4555 | unlikely(root->orphan_cleanup_state == ORPHAN_CLEANUP_STARTED)) { |
| 4556 | struct btrfs_block_rsv *global_rsv = | 4556 | struct btrfs_block_rsv *global_rsv = |
| 4557 | &root->fs_info->global_block_rsv; | 4557 | &root->fs_info->global_block_rsv; |
| 4558 | 4558 | ||
| 4559 | if (block_rsv != global_rsv && | 4559 | if (block_rsv != global_rsv && |
| 4560 | !block_rsv_use_bytes(global_rsv, orig_bytes)) | 4560 | !block_rsv_use_bytes(global_rsv, orig_bytes)) |
| 4561 | ret = 0; | 4561 | ret = 0; |
| 4562 | } | 4562 | } |
| 4563 | if (ret == -ENOSPC) | 4563 | if (ret == -ENOSPC) |
| 4564 | trace_btrfs_space_reservation(root->fs_info, | 4564 | trace_btrfs_space_reservation(root->fs_info, |
| 4565 | "space_info:enospc", | 4565 | "space_info:enospc", |
| 4566 | space_info->flags, orig_bytes, 1); | 4566 | space_info->flags, orig_bytes, 1); |
| 4567 | if (flushing) { | 4567 | if (flushing) { |
| 4568 | spin_lock(&space_info->lock); | 4568 | spin_lock(&space_info->lock); |
| 4569 | space_info->flush = 0; | 4569 | space_info->flush = 0; |
| 4570 | wake_up_all(&space_info->wait); | 4570 | wake_up_all(&space_info->wait); |
| 4571 | spin_unlock(&space_info->lock); | 4571 | spin_unlock(&space_info->lock); |
| 4572 | } | 4572 | } |
| 4573 | return ret; | 4573 | return ret; |
| 4574 | } | 4574 | } |
| 4575 | 4575 | ||
| 4576 | static struct btrfs_block_rsv *get_block_rsv( | 4576 | static struct btrfs_block_rsv *get_block_rsv( |
| 4577 | const struct btrfs_trans_handle *trans, | 4577 | const struct btrfs_trans_handle *trans, |
| 4578 | const struct btrfs_root *root) | 4578 | const struct btrfs_root *root) |
| 4579 | { | 4579 | { |
| 4580 | struct btrfs_block_rsv *block_rsv = NULL; | 4580 | struct btrfs_block_rsv *block_rsv = NULL; |
| 4581 | 4581 | ||
| 4582 | if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) | 4582 | if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) |
| 4583 | block_rsv = trans->block_rsv; | 4583 | block_rsv = trans->block_rsv; |
| 4584 | 4584 | ||
| 4585 | if (root == root->fs_info->csum_root && trans->adding_csums) | 4585 | if (root == root->fs_info->csum_root && trans->adding_csums) |
| 4586 | block_rsv = trans->block_rsv; | 4586 | block_rsv = trans->block_rsv; |
| 4587 | 4587 | ||
| 4588 | if (root == root->fs_info->uuid_root) | 4588 | if (root == root->fs_info->uuid_root) |
| 4589 | block_rsv = trans->block_rsv; | 4589 | block_rsv = trans->block_rsv; |
| 4590 | 4590 | ||
| 4591 | if (!block_rsv) | 4591 | if (!block_rsv) |
| 4592 | block_rsv = root->block_rsv; | 4592 | block_rsv = root->block_rsv; |
| 4593 | 4593 | ||
| 4594 | if (!block_rsv) | 4594 | if (!block_rsv) |
| 4595 | block_rsv = &root->fs_info->empty_block_rsv; | 4595 | block_rsv = &root->fs_info->empty_block_rsv; |
| 4596 | 4596 | ||
| 4597 | return block_rsv; | 4597 | return block_rsv; |
| 4598 | } | 4598 | } |
| 4599 | 4599 | ||
| 4600 | static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, | 4600 | static int block_rsv_use_bytes(struct btrfs_block_rsv *block_rsv, |
| 4601 | u64 num_bytes) | 4601 | u64 num_bytes) |
| 4602 | { | 4602 | { |
| 4603 | int ret = -ENOSPC; | 4603 | int ret = -ENOSPC; |
| 4604 | spin_lock(&block_rsv->lock); | 4604 | spin_lock(&block_rsv->lock); |
| 4605 | if (block_rsv->reserved >= num_bytes) { | 4605 | if (block_rsv->reserved >= num_bytes) { |
| 4606 | block_rsv->reserved -= num_bytes; | 4606 | block_rsv->reserved -= num_bytes; |
| 4607 | if (block_rsv->reserved < block_rsv->size) | 4607 | if (block_rsv->reserved < block_rsv->size) |
| 4608 | block_rsv->full = 0; | 4608 | block_rsv->full = 0; |
| 4609 | ret = 0; | 4609 | ret = 0; |
| 4610 | } | 4610 | } |
| 4611 | spin_unlock(&block_rsv->lock); | 4611 | spin_unlock(&block_rsv->lock); |
| 4612 | return ret; | 4612 | return ret; |
| 4613 | } | 4613 | } |
| 4614 | 4614 | ||
| 4615 | static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv, | 4615 | static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv, |
| 4616 | u64 num_bytes, int update_size) | 4616 | u64 num_bytes, int update_size) |
| 4617 | { | 4617 | { |
| 4618 | spin_lock(&block_rsv->lock); | 4618 | spin_lock(&block_rsv->lock); |
| 4619 | block_rsv->reserved += num_bytes; | 4619 | block_rsv->reserved += num_bytes; |
| 4620 | if (update_size) | 4620 | if (update_size) |
| 4621 | block_rsv->size += num_bytes; | 4621 | block_rsv->size += num_bytes; |
| 4622 | else if (block_rsv->reserved >= block_rsv->size) | 4622 | else if (block_rsv->reserved >= block_rsv->size) |
| 4623 | block_rsv->full = 1; | 4623 | block_rsv->full = 1; |
| 4624 | spin_unlock(&block_rsv->lock); | 4624 | spin_unlock(&block_rsv->lock); |
| 4625 | } | 4625 | } |
| 4626 | 4626 | ||
| 4627 | int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, | 4627 | int btrfs_cond_migrate_bytes(struct btrfs_fs_info *fs_info, |
| 4628 | struct btrfs_block_rsv *dest, u64 num_bytes, | 4628 | struct btrfs_block_rsv *dest, u64 num_bytes, |
| 4629 | int min_factor) | 4629 | int min_factor) |
| 4630 | { | 4630 | { |
| 4631 | struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; | 4631 | struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; |
| 4632 | u64 min_bytes; | 4632 | u64 min_bytes; |
| 4633 | 4633 | ||
| 4634 | if (global_rsv->space_info != dest->space_info) | 4634 | if (global_rsv->space_info != dest->space_info) |
| 4635 | return -ENOSPC; | 4635 | return -ENOSPC; |
| 4636 | 4636 | ||
| 4637 | spin_lock(&global_rsv->lock); | 4637 | spin_lock(&global_rsv->lock); |
| 4638 | min_bytes = div_factor(global_rsv->size, min_factor); | 4638 | min_bytes = div_factor(global_rsv->size, min_factor); |
| 4639 | if (global_rsv->reserved < min_bytes + num_bytes) { | 4639 | if (global_rsv->reserved < min_bytes + num_bytes) { |
| 4640 | spin_unlock(&global_rsv->lock); | 4640 | spin_unlock(&global_rsv->lock); |
| 4641 | return -ENOSPC; | 4641 | return -ENOSPC; |
| 4642 | } | 4642 | } |
| 4643 | global_rsv->reserved -= num_bytes; | 4643 | global_rsv->reserved -= num_bytes; |
| 4644 | if (global_rsv->reserved < global_rsv->size) | 4644 | if (global_rsv->reserved < global_rsv->size) |
| 4645 | global_rsv->full = 0; | 4645 | global_rsv->full = 0; |
| 4646 | spin_unlock(&global_rsv->lock); | 4646 | spin_unlock(&global_rsv->lock); |
| 4647 | 4647 | ||
| 4648 | block_rsv_add_bytes(dest, num_bytes, 1); | 4648 | block_rsv_add_bytes(dest, num_bytes, 1); |
| 4649 | return 0; | 4649 | return 0; |
| 4650 | } | 4650 | } |
| 4651 | 4651 | ||
| 4652 | static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, | 4652 | static void block_rsv_release_bytes(struct btrfs_fs_info *fs_info, |
| 4653 | struct btrfs_block_rsv *block_rsv, | 4653 | struct btrfs_block_rsv *block_rsv, |
| 4654 | struct btrfs_block_rsv *dest, u64 num_bytes) | 4654 | struct btrfs_block_rsv *dest, u64 num_bytes) |
| 4655 | { | 4655 | { |
| 4656 | struct btrfs_space_info *space_info = block_rsv->space_info; | 4656 | struct btrfs_space_info *space_info = block_rsv->space_info; |
| 4657 | 4657 | ||
| 4658 | spin_lock(&block_rsv->lock); | 4658 | spin_lock(&block_rsv->lock); |
| 4659 | if (num_bytes == (u64)-1) | 4659 | if (num_bytes == (u64)-1) |
| 4660 | num_bytes = block_rsv->size; | 4660 | num_bytes = block_rsv->size; |
| 4661 | block_rsv->size -= num_bytes; | 4661 | block_rsv->size -= num_bytes; |
| 4662 | if (block_rsv->reserved >= block_rsv->size) { | 4662 | if (block_rsv->reserved >= block_rsv->size) { |
| 4663 | num_bytes = block_rsv->reserved - block_rsv->size; | 4663 | num_bytes = block_rsv->reserved - block_rsv->size; |
| 4664 | block_rsv->reserved = block_rsv->size; | 4664 | block_rsv->reserved = block_rsv->size; |
| 4665 | block_rsv->full = 1; | 4665 | block_rsv->full = 1; |
| 4666 | } else { | 4666 | } else { |
| 4667 | num_bytes = 0; | 4667 | num_bytes = 0; |
| 4668 | } | 4668 | } |
| 4669 | spin_unlock(&block_rsv->lock); | 4669 | spin_unlock(&block_rsv->lock); |
| 4670 | 4670 | ||
| 4671 | if (num_bytes > 0) { | 4671 | if (num_bytes > 0) { |
| 4672 | if (dest) { | 4672 | if (dest) { |
| 4673 | spin_lock(&dest->lock); | 4673 | spin_lock(&dest->lock); |
| 4674 | if (!dest->full) { | 4674 | if (!dest->full) { |
| 4675 | u64 bytes_to_add; | 4675 | u64 bytes_to_add; |
| 4676 | 4676 | ||
| 4677 | bytes_to_add = dest->size - dest->reserved; | 4677 | bytes_to_add = dest->size - dest->reserved; |
| 4678 | bytes_to_add = min(num_bytes, bytes_to_add); | 4678 | bytes_to_add = min(num_bytes, bytes_to_add); |
| 4679 | dest->reserved += bytes_to_add; | 4679 | dest->reserved += bytes_to_add; |
| 4680 | if (dest->reserved >= dest->size) | 4680 | if (dest->reserved >= dest->size) |
| 4681 | dest->full = 1; | 4681 | dest->full = 1; |
| 4682 | num_bytes -= bytes_to_add; | 4682 | num_bytes -= bytes_to_add; |
| 4683 | } | 4683 | } |
| 4684 | spin_unlock(&dest->lock); | 4684 | spin_unlock(&dest->lock); |
| 4685 | } | 4685 | } |
| 4686 | if (num_bytes) { | 4686 | if (num_bytes) { |
| 4687 | spin_lock(&space_info->lock); | 4687 | spin_lock(&space_info->lock); |
| 4688 | space_info->bytes_may_use -= num_bytes; | 4688 | space_info->bytes_may_use -= num_bytes; |
| 4689 | trace_btrfs_space_reservation(fs_info, "space_info", | 4689 | trace_btrfs_space_reservation(fs_info, "space_info", |
| 4690 | space_info->flags, num_bytes, 0); | 4690 | space_info->flags, num_bytes, 0); |
| 4691 | spin_unlock(&space_info->lock); | 4691 | spin_unlock(&space_info->lock); |
| 4692 | } | 4692 | } |
| 4693 | } | 4693 | } |
| 4694 | } | 4694 | } |
| 4695 | 4695 | ||
| 4696 | static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src, | 4696 | static int block_rsv_migrate_bytes(struct btrfs_block_rsv *src, |
| 4697 | struct btrfs_block_rsv *dst, u64 num_bytes) | 4697 | struct btrfs_block_rsv *dst, u64 num_bytes) |
| 4698 | { | 4698 | { |
| 4699 | int ret; | 4699 | int ret; |
| 4700 | 4700 | ||
| 4701 | ret = block_rsv_use_bytes(src, num_bytes); | 4701 | ret = block_rsv_use_bytes(src, num_bytes); |
| 4702 | if (ret) | 4702 | if (ret) |
| 4703 | return ret; | 4703 | return ret; |
| 4704 | 4704 | ||
| 4705 | block_rsv_add_bytes(dst, num_bytes, 1); | 4705 | block_rsv_add_bytes(dst, num_bytes, 1); |
| 4706 | return 0; | 4706 | return 0; |
| 4707 | } | 4707 | } |
| 4708 | 4708 | ||
| 4709 | void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv, unsigned short type) | 4709 | void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv, unsigned short type) |
| 4710 | { | 4710 | { |
| 4711 | memset(rsv, 0, sizeof(*rsv)); | 4711 | memset(rsv, 0, sizeof(*rsv)); |
| 4712 | spin_lock_init(&rsv->lock); | 4712 | spin_lock_init(&rsv->lock); |
| 4713 | rsv->type = type; | 4713 | rsv->type = type; |
| 4714 | } | 4714 | } |
| 4715 | 4715 | ||
| 4716 | struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root, | 4716 | struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root, |
| 4717 | unsigned short type) | 4717 | unsigned short type) |
| 4718 | { | 4718 | { |
| 4719 | struct btrfs_block_rsv *block_rsv; | 4719 | struct btrfs_block_rsv *block_rsv; |
| 4720 | struct btrfs_fs_info *fs_info = root->fs_info; | 4720 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 4721 | 4721 | ||
| 4722 | block_rsv = kmalloc(sizeof(*block_rsv), GFP_NOFS); | 4722 | block_rsv = kmalloc(sizeof(*block_rsv), GFP_NOFS); |
| 4723 | if (!block_rsv) | 4723 | if (!block_rsv) |
| 4724 | return NULL; | 4724 | return NULL; |
| 4725 | 4725 | ||
| 4726 | btrfs_init_block_rsv(block_rsv, type); | 4726 | btrfs_init_block_rsv(block_rsv, type); |
| 4727 | block_rsv->space_info = __find_space_info(fs_info, | 4727 | block_rsv->space_info = __find_space_info(fs_info, |
| 4728 | BTRFS_BLOCK_GROUP_METADATA); | 4728 | BTRFS_BLOCK_GROUP_METADATA); |
| 4729 | return block_rsv; | 4729 | return block_rsv; |
| 4730 | } | 4730 | } |
| 4731 | 4731 | ||
| 4732 | void btrfs_free_block_rsv(struct btrfs_root *root, | 4732 | void btrfs_free_block_rsv(struct btrfs_root *root, |
| 4733 | struct btrfs_block_rsv *rsv) | 4733 | struct btrfs_block_rsv *rsv) |
| 4734 | { | 4734 | { |
| 4735 | if (!rsv) | 4735 | if (!rsv) |
| 4736 | return; | 4736 | return; |
| 4737 | btrfs_block_rsv_release(root, rsv, (u64)-1); | 4737 | btrfs_block_rsv_release(root, rsv, (u64)-1); |
| 4738 | kfree(rsv); | 4738 | kfree(rsv); |
| 4739 | } | 4739 | } |
| 4740 | 4740 | ||
| 4741 | int btrfs_block_rsv_add(struct btrfs_root *root, | 4741 | int btrfs_block_rsv_add(struct btrfs_root *root, |
| 4742 | struct btrfs_block_rsv *block_rsv, u64 num_bytes, | 4742 | struct btrfs_block_rsv *block_rsv, u64 num_bytes, |
| 4743 | enum btrfs_reserve_flush_enum flush) | 4743 | enum btrfs_reserve_flush_enum flush) |
| 4744 | { | 4744 | { |
| 4745 | int ret; | 4745 | int ret; |
| 4746 | 4746 | ||
| 4747 | if (num_bytes == 0) | 4747 | if (num_bytes == 0) |
| 4748 | return 0; | 4748 | return 0; |
| 4749 | 4749 | ||
| 4750 | ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); | 4750 | ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); |
| 4751 | if (!ret) { | 4751 | if (!ret) { |
| 4752 | block_rsv_add_bytes(block_rsv, num_bytes, 1); | 4752 | block_rsv_add_bytes(block_rsv, num_bytes, 1); |
| 4753 | return 0; | 4753 | return 0; |
| 4754 | } | 4754 | } |
| 4755 | 4755 | ||
| 4756 | return ret; | 4756 | return ret; |
| 4757 | } | 4757 | } |
| 4758 | 4758 | ||
| 4759 | int btrfs_block_rsv_check(struct btrfs_root *root, | 4759 | int btrfs_block_rsv_check(struct btrfs_root *root, |
| 4760 | struct btrfs_block_rsv *block_rsv, int min_factor) | 4760 | struct btrfs_block_rsv *block_rsv, int min_factor) |
| 4761 | { | 4761 | { |
| 4762 | u64 num_bytes = 0; | 4762 | u64 num_bytes = 0; |
| 4763 | int ret = -ENOSPC; | 4763 | int ret = -ENOSPC; |
| 4764 | 4764 | ||
| 4765 | if (!block_rsv) | 4765 | if (!block_rsv) |
| 4766 | return 0; | 4766 | return 0; |
| 4767 | 4767 | ||
| 4768 | spin_lock(&block_rsv->lock); | 4768 | spin_lock(&block_rsv->lock); |
| 4769 | num_bytes = div_factor(block_rsv->size, min_factor); | 4769 | num_bytes = div_factor(block_rsv->size, min_factor); |
| 4770 | if (block_rsv->reserved >= num_bytes) | 4770 | if (block_rsv->reserved >= num_bytes) |
| 4771 | ret = 0; | 4771 | ret = 0; |
| 4772 | spin_unlock(&block_rsv->lock); | 4772 | spin_unlock(&block_rsv->lock); |
| 4773 | 4773 | ||
| 4774 | return ret; | 4774 | return ret; |
| 4775 | } | 4775 | } |
| 4776 | 4776 | ||
| 4777 | int btrfs_block_rsv_refill(struct btrfs_root *root, | 4777 | int btrfs_block_rsv_refill(struct btrfs_root *root, |
| 4778 | struct btrfs_block_rsv *block_rsv, u64 min_reserved, | 4778 | struct btrfs_block_rsv *block_rsv, u64 min_reserved, |
| 4779 | enum btrfs_reserve_flush_enum flush) | 4779 | enum btrfs_reserve_flush_enum flush) |
| 4780 | { | 4780 | { |
| 4781 | u64 num_bytes = 0; | 4781 | u64 num_bytes = 0; |
| 4782 | int ret = -ENOSPC; | 4782 | int ret = -ENOSPC; |
| 4783 | 4783 | ||
| 4784 | if (!block_rsv) | 4784 | if (!block_rsv) |
| 4785 | return 0; | 4785 | return 0; |
| 4786 | 4786 | ||
| 4787 | spin_lock(&block_rsv->lock); | 4787 | spin_lock(&block_rsv->lock); |
| 4788 | num_bytes = min_reserved; | 4788 | num_bytes = min_reserved; |
| 4789 | if (block_rsv->reserved >= num_bytes) | 4789 | if (block_rsv->reserved >= num_bytes) |
| 4790 | ret = 0; | 4790 | ret = 0; |
| 4791 | else | 4791 | else |
| 4792 | num_bytes -= block_rsv->reserved; | 4792 | num_bytes -= block_rsv->reserved; |
| 4793 | spin_unlock(&block_rsv->lock); | 4793 | spin_unlock(&block_rsv->lock); |
| 4794 | 4794 | ||
| 4795 | if (!ret) | 4795 | if (!ret) |
| 4796 | return 0; | 4796 | return 0; |
| 4797 | 4797 | ||
| 4798 | ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); | 4798 | ret = reserve_metadata_bytes(root, block_rsv, num_bytes, flush); |
| 4799 | if (!ret) { | 4799 | if (!ret) { |
| 4800 | block_rsv_add_bytes(block_rsv, num_bytes, 0); | 4800 | block_rsv_add_bytes(block_rsv, num_bytes, 0); |
| 4801 | return 0; | 4801 | return 0; |
| 4802 | } | 4802 | } |
| 4803 | 4803 | ||
| 4804 | return ret; | 4804 | return ret; |
| 4805 | } | 4805 | } |
| 4806 | 4806 | ||
| 4807 | int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, | 4807 | int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, |
| 4808 | struct btrfs_block_rsv *dst_rsv, | 4808 | struct btrfs_block_rsv *dst_rsv, |
| 4809 | u64 num_bytes) | 4809 | u64 num_bytes) |
| 4810 | { | 4810 | { |
| 4811 | return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); | 4811 | return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); |
| 4812 | } | 4812 | } |
| 4813 | 4813 | ||
| 4814 | void btrfs_block_rsv_release(struct btrfs_root *root, | 4814 | void btrfs_block_rsv_release(struct btrfs_root *root, |
| 4815 | struct btrfs_block_rsv *block_rsv, | 4815 | struct btrfs_block_rsv *block_rsv, |
| 4816 | u64 num_bytes) | 4816 | u64 num_bytes) |
| 4817 | { | 4817 | { |
| 4818 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; | 4818 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; |
| 4819 | if (global_rsv == block_rsv || | 4819 | if (global_rsv == block_rsv || |
| 4820 | block_rsv->space_info != global_rsv->space_info) | 4820 | block_rsv->space_info != global_rsv->space_info) |
| 4821 | global_rsv = NULL; | 4821 | global_rsv = NULL; |
| 4822 | block_rsv_release_bytes(root->fs_info, block_rsv, global_rsv, | 4822 | block_rsv_release_bytes(root->fs_info, block_rsv, global_rsv, |
| 4823 | num_bytes); | 4823 | num_bytes); |
| 4824 | } | 4824 | } |
| 4825 | 4825 | ||
| 4826 | /* | 4826 | /* |
| 4827 | * helper to calculate size of global block reservation. | 4827 | * helper to calculate size of global block reservation. |
| 4828 | * the desired value is sum of space used by extent tree, | 4828 | * the desired value is sum of space used by extent tree, |
| 4829 | * checksum tree and root tree | 4829 | * checksum tree and root tree |
| 4830 | */ | 4830 | */ |
| 4831 | static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) | 4831 | static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) |
| 4832 | { | 4832 | { |
| 4833 | struct btrfs_space_info *sinfo; | 4833 | struct btrfs_space_info *sinfo; |
| 4834 | u64 num_bytes; | 4834 | u64 num_bytes; |
| 4835 | u64 meta_used; | 4835 | u64 meta_used; |
| 4836 | u64 data_used; | 4836 | u64 data_used; |
| 4837 | int csum_size = btrfs_super_csum_size(fs_info->super_copy); | 4837 | int csum_size = btrfs_super_csum_size(fs_info->super_copy); |
| 4838 | 4838 | ||
| 4839 | sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); | 4839 | sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); |
| 4840 | spin_lock(&sinfo->lock); | 4840 | spin_lock(&sinfo->lock); |
| 4841 | data_used = sinfo->bytes_used; | 4841 | data_used = sinfo->bytes_used; |
| 4842 | spin_unlock(&sinfo->lock); | 4842 | spin_unlock(&sinfo->lock); |
| 4843 | 4843 | ||
| 4844 | sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); | 4844 | sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); |
| 4845 | spin_lock(&sinfo->lock); | 4845 | spin_lock(&sinfo->lock); |
| 4846 | if (sinfo->flags & BTRFS_BLOCK_GROUP_DATA) | 4846 | if (sinfo->flags & BTRFS_BLOCK_GROUP_DATA) |
| 4847 | data_used = 0; | 4847 | data_used = 0; |
| 4848 | meta_used = sinfo->bytes_used; | 4848 | meta_used = sinfo->bytes_used; |
| 4849 | spin_unlock(&sinfo->lock); | 4849 | spin_unlock(&sinfo->lock); |
| 4850 | 4850 | ||
| 4851 | num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) * | 4851 | num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) * |
| 4852 | csum_size * 2; | 4852 | csum_size * 2; |
| 4853 | num_bytes += div64_u64(data_used + meta_used, 50); | 4853 | num_bytes += div64_u64(data_used + meta_used, 50); |
| 4854 | 4854 | ||
| 4855 | if (num_bytes * 3 > meta_used) | 4855 | if (num_bytes * 3 > meta_used) |
| 4856 | num_bytes = div64_u64(meta_used, 3); | 4856 | num_bytes = div64_u64(meta_used, 3); |
| 4857 | 4857 | ||
| 4858 | return ALIGN(num_bytes, fs_info->extent_root->nodesize << 10); | 4858 | return ALIGN(num_bytes, fs_info->extent_root->nodesize << 10); |
| 4859 | } | 4859 | } |
| 4860 | 4860 | ||
| 4861 | static void update_global_block_rsv(struct btrfs_fs_info *fs_info) | 4861 | static void update_global_block_rsv(struct btrfs_fs_info *fs_info) |
| 4862 | { | 4862 | { |
| 4863 | struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv; | 4863 | struct btrfs_block_rsv *block_rsv = &fs_info->global_block_rsv; |
| 4864 | struct btrfs_space_info *sinfo = block_rsv->space_info; | 4864 | struct btrfs_space_info *sinfo = block_rsv->space_info; |
| 4865 | u64 num_bytes; | 4865 | u64 num_bytes; |
| 4866 | 4866 | ||
| 4867 | num_bytes = calc_global_metadata_size(fs_info); | 4867 | num_bytes = calc_global_metadata_size(fs_info); |
| 4868 | 4868 | ||
| 4869 | spin_lock(&sinfo->lock); | 4869 | spin_lock(&sinfo->lock); |
| 4870 | spin_lock(&block_rsv->lock); | 4870 | spin_lock(&block_rsv->lock); |
| 4871 | 4871 | ||
| 4872 | block_rsv->size = min_t(u64, num_bytes, 512 * 1024 * 1024); | 4872 | block_rsv->size = min_t(u64, num_bytes, 512 * 1024 * 1024); |
| 4873 | 4873 | ||
| 4874 | num_bytes = sinfo->bytes_used + sinfo->bytes_pinned + | 4874 | num_bytes = sinfo->bytes_used + sinfo->bytes_pinned + |
| 4875 | sinfo->bytes_reserved + sinfo->bytes_readonly + | 4875 | sinfo->bytes_reserved + sinfo->bytes_readonly + |
| 4876 | sinfo->bytes_may_use; | 4876 | sinfo->bytes_may_use; |
| 4877 | 4877 | ||
| 4878 | if (sinfo->total_bytes > num_bytes) { | 4878 | if (sinfo->total_bytes > num_bytes) { |
| 4879 | num_bytes = sinfo->total_bytes - num_bytes; | 4879 | num_bytes = sinfo->total_bytes - num_bytes; |
| 4880 | block_rsv->reserved += num_bytes; | 4880 | block_rsv->reserved += num_bytes; |
| 4881 | sinfo->bytes_may_use += num_bytes; | 4881 | sinfo->bytes_may_use += num_bytes; |
| 4882 | trace_btrfs_space_reservation(fs_info, "space_info", | 4882 | trace_btrfs_space_reservation(fs_info, "space_info", |
| 4883 | sinfo->flags, num_bytes, 1); | 4883 | sinfo->flags, num_bytes, 1); |
| 4884 | } | 4884 | } |
| 4885 | 4885 | ||
| 4886 | if (block_rsv->reserved >= block_rsv->size) { | 4886 | if (block_rsv->reserved >= block_rsv->size) { |
| 4887 | num_bytes = block_rsv->reserved - block_rsv->size; | 4887 | num_bytes = block_rsv->reserved - block_rsv->size; |
| 4888 | sinfo->bytes_may_use -= num_bytes; | 4888 | sinfo->bytes_may_use -= num_bytes; |
| 4889 | trace_btrfs_space_reservation(fs_info, "space_info", | 4889 | trace_btrfs_space_reservation(fs_info, "space_info", |
| 4890 | sinfo->flags, num_bytes, 0); | 4890 | sinfo->flags, num_bytes, 0); |
| 4891 | block_rsv->reserved = block_rsv->size; | 4891 | block_rsv->reserved = block_rsv->size; |
| 4892 | block_rsv->full = 1; | 4892 | block_rsv->full = 1; |
| 4893 | } | 4893 | } |
| 4894 | 4894 | ||
| 4895 | spin_unlock(&block_rsv->lock); | 4895 | spin_unlock(&block_rsv->lock); |
| 4896 | spin_unlock(&sinfo->lock); | 4896 | spin_unlock(&sinfo->lock); |
| 4897 | } | 4897 | } |
| 4898 | 4898 | ||
| 4899 | static void init_global_block_rsv(struct btrfs_fs_info *fs_info) | 4899 | static void init_global_block_rsv(struct btrfs_fs_info *fs_info) |
| 4900 | { | 4900 | { |
| 4901 | struct btrfs_space_info *space_info; | 4901 | struct btrfs_space_info *space_info; |
| 4902 | 4902 | ||
| 4903 | space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); | 4903 | space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_SYSTEM); |
| 4904 | fs_info->chunk_block_rsv.space_info = space_info; | 4904 | fs_info->chunk_block_rsv.space_info = space_info; |
| 4905 | 4905 | ||
| 4906 | space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); | 4906 | space_info = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_METADATA); |
| 4907 | fs_info->global_block_rsv.space_info = space_info; | 4907 | fs_info->global_block_rsv.space_info = space_info; |
| 4908 | fs_info->delalloc_block_rsv.space_info = space_info; | 4908 | fs_info->delalloc_block_rsv.space_info = space_info; |
| 4909 | fs_info->trans_block_rsv.space_info = space_info; | 4909 | fs_info->trans_block_rsv.space_info = space_info; |
| 4910 | fs_info->empty_block_rsv.space_info = space_info; | 4910 | fs_info->empty_block_rsv.space_info = space_info; |
| 4911 | fs_info->delayed_block_rsv.space_info = space_info; | 4911 | fs_info->delayed_block_rsv.space_info = space_info; |
| 4912 | 4912 | ||
| 4913 | fs_info->extent_root->block_rsv = &fs_info->global_block_rsv; | 4913 | fs_info->extent_root->block_rsv = &fs_info->global_block_rsv; |
| 4914 | fs_info->csum_root->block_rsv = &fs_info->global_block_rsv; | 4914 | fs_info->csum_root->block_rsv = &fs_info->global_block_rsv; |
| 4915 | fs_info->dev_root->block_rsv = &fs_info->global_block_rsv; | 4915 | fs_info->dev_root->block_rsv = &fs_info->global_block_rsv; |
| 4916 | fs_info->tree_root->block_rsv = &fs_info->global_block_rsv; | 4916 | fs_info->tree_root->block_rsv = &fs_info->global_block_rsv; |
| 4917 | if (fs_info->quota_root) | 4917 | if (fs_info->quota_root) |
| 4918 | fs_info->quota_root->block_rsv = &fs_info->global_block_rsv; | 4918 | fs_info->quota_root->block_rsv = &fs_info->global_block_rsv; |
| 4919 | fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv; | 4919 | fs_info->chunk_root->block_rsv = &fs_info->chunk_block_rsv; |
| 4920 | 4920 | ||
| 4921 | update_global_block_rsv(fs_info); | 4921 | update_global_block_rsv(fs_info); |
| 4922 | } | 4922 | } |
| 4923 | 4923 | ||
| 4924 | static void release_global_block_rsv(struct btrfs_fs_info *fs_info) | 4924 | static void release_global_block_rsv(struct btrfs_fs_info *fs_info) |
| 4925 | { | 4925 | { |
| 4926 | block_rsv_release_bytes(fs_info, &fs_info->global_block_rsv, NULL, | 4926 | block_rsv_release_bytes(fs_info, &fs_info->global_block_rsv, NULL, |
| 4927 | (u64)-1); | 4927 | (u64)-1); |
| 4928 | WARN_ON(fs_info->delalloc_block_rsv.size > 0); | 4928 | WARN_ON(fs_info->delalloc_block_rsv.size > 0); |
| 4929 | WARN_ON(fs_info->delalloc_block_rsv.reserved > 0); | 4929 | WARN_ON(fs_info->delalloc_block_rsv.reserved > 0); |
| 4930 | WARN_ON(fs_info->trans_block_rsv.size > 0); | 4930 | WARN_ON(fs_info->trans_block_rsv.size > 0); |
| 4931 | WARN_ON(fs_info->trans_block_rsv.reserved > 0); | 4931 | WARN_ON(fs_info->trans_block_rsv.reserved > 0); |
| 4932 | WARN_ON(fs_info->chunk_block_rsv.size > 0); | 4932 | WARN_ON(fs_info->chunk_block_rsv.size > 0); |
| 4933 | WARN_ON(fs_info->chunk_block_rsv.reserved > 0); | 4933 | WARN_ON(fs_info->chunk_block_rsv.reserved > 0); |
| 4934 | WARN_ON(fs_info->delayed_block_rsv.size > 0); | 4934 | WARN_ON(fs_info->delayed_block_rsv.size > 0); |
| 4935 | WARN_ON(fs_info->delayed_block_rsv.reserved > 0); | 4935 | WARN_ON(fs_info->delayed_block_rsv.reserved > 0); |
| 4936 | } | 4936 | } |
| 4937 | 4937 | ||
| 4938 | void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, | 4938 | void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans, |
| 4939 | struct btrfs_root *root) | 4939 | struct btrfs_root *root) |
| 4940 | { | 4940 | { |
| 4941 | if (!trans->block_rsv) | 4941 | if (!trans->block_rsv) |
| 4942 | return; | 4942 | return; |
| 4943 | 4943 | ||
| 4944 | if (!trans->bytes_reserved) | 4944 | if (!trans->bytes_reserved) |
| 4945 | return; | 4945 | return; |
| 4946 | 4946 | ||
| 4947 | trace_btrfs_space_reservation(root->fs_info, "transaction", | 4947 | trace_btrfs_space_reservation(root->fs_info, "transaction", |
| 4948 | trans->transid, trans->bytes_reserved, 0); | 4948 | trans->transid, trans->bytes_reserved, 0); |
| 4949 | btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved); | 4949 | btrfs_block_rsv_release(root, trans->block_rsv, trans->bytes_reserved); |
| 4950 | trans->bytes_reserved = 0; | 4950 | trans->bytes_reserved = 0; |
| 4951 | } | 4951 | } |
| 4952 | 4952 | ||
| 4953 | /* Can only return 0 or -ENOSPC */ | 4953 | /* Can only return 0 or -ENOSPC */ |
| 4954 | int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, | 4954 | int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, |
| 4955 | struct inode *inode) | 4955 | struct inode *inode) |
| 4956 | { | 4956 | { |
| 4957 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4957 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4958 | struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root); | 4958 | struct btrfs_block_rsv *src_rsv = get_block_rsv(trans, root); |
| 4959 | struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv; | 4959 | struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv; |
| 4960 | 4960 | ||
| 4961 | /* | 4961 | /* |
| 4962 | * We need to hold space in order to delete our orphan item once we've | 4962 | * We need to hold space in order to delete our orphan item once we've |
| 4963 | * added it, so this takes the reservation so we can release it later | 4963 | * added it, so this takes the reservation so we can release it later |
| 4964 | * when we are truly done with the orphan item. | 4964 | * when we are truly done with the orphan item. |
| 4965 | */ | 4965 | */ |
| 4966 | u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); | 4966 | u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); |
| 4967 | trace_btrfs_space_reservation(root->fs_info, "orphan", | 4967 | trace_btrfs_space_reservation(root->fs_info, "orphan", |
| 4968 | btrfs_ino(inode), num_bytes, 1); | 4968 | btrfs_ino(inode), num_bytes, 1); |
| 4969 | return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); | 4969 | return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); |
| 4970 | } | 4970 | } |
| 4971 | 4971 | ||
| 4972 | void btrfs_orphan_release_metadata(struct inode *inode) | 4972 | void btrfs_orphan_release_metadata(struct inode *inode) |
| 4973 | { | 4973 | { |
| 4974 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4974 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4975 | u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); | 4975 | u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); |
| 4976 | trace_btrfs_space_reservation(root->fs_info, "orphan", | 4976 | trace_btrfs_space_reservation(root->fs_info, "orphan", |
| 4977 | btrfs_ino(inode), num_bytes, 0); | 4977 | btrfs_ino(inode), num_bytes, 0); |
| 4978 | btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes); | 4978 | btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes); |
| 4979 | } | 4979 | } |
| 4980 | 4980 | ||
| 4981 | /* | 4981 | /* |
| 4982 | * btrfs_subvolume_reserve_metadata() - reserve space for subvolume operation | 4982 | * btrfs_subvolume_reserve_metadata() - reserve space for subvolume operation |
| 4983 | * root: the root of the parent directory | 4983 | * root: the root of the parent directory |
| 4984 | * rsv: block reservation | 4984 | * rsv: block reservation |
| 4985 | * items: the number of items that we need do reservation | 4985 | * items: the number of items that we need do reservation |
| 4986 | * qgroup_reserved: used to return the reserved size in qgroup | 4986 | * qgroup_reserved: used to return the reserved size in qgroup |
| 4987 | * | 4987 | * |
| 4988 | * This function is used to reserve the space for snapshot/subvolume | 4988 | * This function is used to reserve the space for snapshot/subvolume |
| 4989 | * creation and deletion. Those operations are different with the | 4989 | * creation and deletion. Those operations are different with the |
| 4990 | * common file/directory operations, they change two fs/file trees | 4990 | * common file/directory operations, they change two fs/file trees |
| 4991 | * and root tree, the number of items that the qgroup reserves is | 4991 | * and root tree, the number of items that the qgroup reserves is |
| 4992 | * different with the free space reservation. So we can not use | 4992 | * different with the free space reservation. So we can not use |
| 4993 | * the space reseravtion mechanism in start_transaction(). | 4993 | * the space reseravtion mechanism in start_transaction(). |
| 4994 | */ | 4994 | */ |
| 4995 | int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, | 4995 | int btrfs_subvolume_reserve_metadata(struct btrfs_root *root, |
| 4996 | struct btrfs_block_rsv *rsv, | 4996 | struct btrfs_block_rsv *rsv, |
| 4997 | int items, | 4997 | int items, |
| 4998 | u64 *qgroup_reserved, | 4998 | u64 *qgroup_reserved, |
| 4999 | bool use_global_rsv) | 4999 | bool use_global_rsv) |
| 5000 | { | 5000 | { |
| 5001 | u64 num_bytes; | 5001 | u64 num_bytes; |
| 5002 | int ret; | 5002 | int ret; |
| 5003 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; | 5003 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; |
| 5004 | 5004 | ||
| 5005 | if (root->fs_info->quota_enabled) { | 5005 | if (root->fs_info->quota_enabled) { |
| 5006 | /* One for parent inode, two for dir entries */ | 5006 | /* One for parent inode, two for dir entries */ |
| 5007 | num_bytes = 3 * root->nodesize; | 5007 | num_bytes = 3 * root->nodesize; |
| 5008 | ret = btrfs_qgroup_reserve(root, num_bytes); | 5008 | ret = btrfs_qgroup_reserve(root, num_bytes); |
| 5009 | if (ret) | 5009 | if (ret) |
| 5010 | return ret; | 5010 | return ret; |
| 5011 | } else { | 5011 | } else { |
| 5012 | num_bytes = 0; | 5012 | num_bytes = 0; |
| 5013 | } | 5013 | } |
| 5014 | 5014 | ||
| 5015 | *qgroup_reserved = num_bytes; | 5015 | *qgroup_reserved = num_bytes; |
| 5016 | 5016 | ||
| 5017 | num_bytes = btrfs_calc_trans_metadata_size(root, items); | 5017 | num_bytes = btrfs_calc_trans_metadata_size(root, items); |
| 5018 | rsv->space_info = __find_space_info(root->fs_info, | 5018 | rsv->space_info = __find_space_info(root->fs_info, |
| 5019 | BTRFS_BLOCK_GROUP_METADATA); | 5019 | BTRFS_BLOCK_GROUP_METADATA); |
| 5020 | ret = btrfs_block_rsv_add(root, rsv, num_bytes, | 5020 | ret = btrfs_block_rsv_add(root, rsv, num_bytes, |
| 5021 | BTRFS_RESERVE_FLUSH_ALL); | 5021 | BTRFS_RESERVE_FLUSH_ALL); |
| 5022 | 5022 | ||
| 5023 | if (ret == -ENOSPC && use_global_rsv) | 5023 | if (ret == -ENOSPC && use_global_rsv) |
| 5024 | ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes); | 5024 | ret = btrfs_block_rsv_migrate(global_rsv, rsv, num_bytes); |
| 5025 | 5025 | ||
| 5026 | if (ret) { | 5026 | if (ret) { |
| 5027 | if (*qgroup_reserved) | 5027 | if (*qgroup_reserved) |
| 5028 | btrfs_qgroup_free(root, *qgroup_reserved); | 5028 | btrfs_qgroup_free(root, *qgroup_reserved); |
| 5029 | } | 5029 | } |
| 5030 | 5030 | ||
| 5031 | return ret; | 5031 | return ret; |
| 5032 | } | 5032 | } |
| 5033 | 5033 | ||
| 5034 | void btrfs_subvolume_release_metadata(struct btrfs_root *root, | 5034 | void btrfs_subvolume_release_metadata(struct btrfs_root *root, |
| 5035 | struct btrfs_block_rsv *rsv, | 5035 | struct btrfs_block_rsv *rsv, |
| 5036 | u64 qgroup_reserved) | 5036 | u64 qgroup_reserved) |
| 5037 | { | 5037 | { |
| 5038 | btrfs_block_rsv_release(root, rsv, (u64)-1); | 5038 | btrfs_block_rsv_release(root, rsv, (u64)-1); |
| 5039 | if (qgroup_reserved) | 5039 | if (qgroup_reserved) |
| 5040 | btrfs_qgroup_free(root, qgroup_reserved); | 5040 | btrfs_qgroup_free(root, qgroup_reserved); |
| 5041 | } | 5041 | } |
| 5042 | 5042 | ||
| 5043 | /** | 5043 | /** |
| 5044 | * drop_outstanding_extent - drop an outstanding extent | 5044 | * drop_outstanding_extent - drop an outstanding extent |
| 5045 | * @inode: the inode we're dropping the extent for | 5045 | * @inode: the inode we're dropping the extent for |
| 5046 | * | 5046 | * |
| 5047 | * This is called when we are freeing up an outstanding extent, either called | 5047 | * This is called when we are freeing up an outstanding extent, either called |
| 5048 | * after an error or after an extent is written. This will return the number of | 5048 | * after an error or after an extent is written. This will return the number of |
| 5049 | * reserved extents that need to be freed. This must be called with | 5049 | * reserved extents that need to be freed. This must be called with |
| 5050 | * BTRFS_I(inode)->lock held. | 5050 | * BTRFS_I(inode)->lock held. |
| 5051 | */ | 5051 | */ |
| 5052 | static unsigned drop_outstanding_extent(struct inode *inode) | 5052 | static unsigned drop_outstanding_extent(struct inode *inode) |
| 5053 | { | 5053 | { |
| 5054 | unsigned drop_inode_space = 0; | 5054 | unsigned drop_inode_space = 0; |
| 5055 | unsigned dropped_extents = 0; | 5055 | unsigned dropped_extents = 0; |
| 5056 | 5056 | ||
| 5057 | BUG_ON(!BTRFS_I(inode)->outstanding_extents); | 5057 | BUG_ON(!BTRFS_I(inode)->outstanding_extents); |
| 5058 | BTRFS_I(inode)->outstanding_extents--; | 5058 | BTRFS_I(inode)->outstanding_extents--; |
| 5059 | 5059 | ||
| 5060 | if (BTRFS_I(inode)->outstanding_extents == 0 && | 5060 | if (BTRFS_I(inode)->outstanding_extents == 0 && |
| 5061 | test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, | 5061 | test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED, |
| 5062 | &BTRFS_I(inode)->runtime_flags)) | 5062 | &BTRFS_I(inode)->runtime_flags)) |
| 5063 | drop_inode_space = 1; | 5063 | drop_inode_space = 1; |
| 5064 | 5064 | ||
| 5065 | /* | 5065 | /* |
| 5066 | * If we have more or the same amount of outsanding extents than we have | 5066 | * If we have more or the same amount of outsanding extents than we have |
| 5067 | * reserved then we need to leave the reserved extents count alone. | 5067 | * reserved then we need to leave the reserved extents count alone. |
| 5068 | */ | 5068 | */ |
| 5069 | if (BTRFS_I(inode)->outstanding_extents >= | 5069 | if (BTRFS_I(inode)->outstanding_extents >= |
| 5070 | BTRFS_I(inode)->reserved_extents) | 5070 | BTRFS_I(inode)->reserved_extents) |
| 5071 | return drop_inode_space; | 5071 | return drop_inode_space; |
| 5072 | 5072 | ||
| 5073 | dropped_extents = BTRFS_I(inode)->reserved_extents - | 5073 | dropped_extents = BTRFS_I(inode)->reserved_extents - |
| 5074 | BTRFS_I(inode)->outstanding_extents; | 5074 | BTRFS_I(inode)->outstanding_extents; |
| 5075 | BTRFS_I(inode)->reserved_extents -= dropped_extents; | 5075 | BTRFS_I(inode)->reserved_extents -= dropped_extents; |
| 5076 | return dropped_extents + drop_inode_space; | 5076 | return dropped_extents + drop_inode_space; |
| 5077 | } | 5077 | } |
| 5078 | 5078 | ||
| 5079 | /** | 5079 | /** |
| 5080 | * calc_csum_metadata_size - return the amount of metada space that must be | 5080 | * calc_csum_metadata_size - return the amount of metada space that must be |
| 5081 | * reserved/free'd for the given bytes. | 5081 | * reserved/free'd for the given bytes. |
| 5082 | * @inode: the inode we're manipulating | 5082 | * @inode: the inode we're manipulating |
| 5083 | * @num_bytes: the number of bytes in question | 5083 | * @num_bytes: the number of bytes in question |
| 5084 | * @reserve: 1 if we are reserving space, 0 if we are freeing space | 5084 | * @reserve: 1 if we are reserving space, 0 if we are freeing space |
| 5085 | * | 5085 | * |
| 5086 | * This adjusts the number of csum_bytes in the inode and then returns the | 5086 | * This adjusts the number of csum_bytes in the inode and then returns the |
| 5087 | * correct amount of metadata that must either be reserved or freed. We | 5087 | * correct amount of metadata that must either be reserved or freed. We |
| 5088 | * calculate how many checksums we can fit into one leaf and then divide the | 5088 | * calculate how many checksums we can fit into one leaf and then divide the |
| 5089 | * number of bytes that will need to be checksumed by this value to figure out | 5089 | * number of bytes that will need to be checksumed by this value to figure out |
| 5090 | * how many checksums will be required. If we are adding bytes then the number | 5090 | * how many checksums will be required. If we are adding bytes then the number |
| 5091 | * may go up and we will return the number of additional bytes that must be | 5091 | * may go up and we will return the number of additional bytes that must be |
| 5092 | * reserved. If it is going down we will return the number of bytes that must | 5092 | * reserved. If it is going down we will return the number of bytes that must |
| 5093 | * be freed. | 5093 | * be freed. |
| 5094 | * | 5094 | * |
| 5095 | * This must be called with BTRFS_I(inode)->lock held. | 5095 | * This must be called with BTRFS_I(inode)->lock held. |
| 5096 | */ | 5096 | */ |
| 5097 | static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes, | 5097 | static u64 calc_csum_metadata_size(struct inode *inode, u64 num_bytes, |
| 5098 | int reserve) | 5098 | int reserve) |
| 5099 | { | 5099 | { |
| 5100 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5100 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5101 | u64 csum_size; | 5101 | u64 csum_size; |
| 5102 | int num_csums_per_leaf; | 5102 | int num_csums_per_leaf; |
| 5103 | int num_csums; | 5103 | int num_csums; |
| 5104 | int old_csums; | 5104 | int old_csums; |
| 5105 | 5105 | ||
| 5106 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM && | 5106 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM && |
| 5107 | BTRFS_I(inode)->csum_bytes == 0) | 5107 | BTRFS_I(inode)->csum_bytes == 0) |
| 5108 | return 0; | 5108 | return 0; |
| 5109 | 5109 | ||
| 5110 | old_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); | 5110 | old_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); |
| 5111 | if (reserve) | 5111 | if (reserve) |
| 5112 | BTRFS_I(inode)->csum_bytes += num_bytes; | 5112 | BTRFS_I(inode)->csum_bytes += num_bytes; |
| 5113 | else | 5113 | else |
| 5114 | BTRFS_I(inode)->csum_bytes -= num_bytes; | 5114 | BTRFS_I(inode)->csum_bytes -= num_bytes; |
| 5115 | csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item); | 5115 | csum_size = BTRFS_LEAF_DATA_SIZE(root) - sizeof(struct btrfs_item); |
| 5116 | num_csums_per_leaf = (int)div64_u64(csum_size, | 5116 | num_csums_per_leaf = (int)div64_u64(csum_size, |
| 5117 | sizeof(struct btrfs_csum_item) + | 5117 | sizeof(struct btrfs_csum_item) + |
| 5118 | sizeof(struct btrfs_disk_key)); | 5118 | sizeof(struct btrfs_disk_key)); |
| 5119 | num_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); | 5119 | num_csums = (int)div64_u64(BTRFS_I(inode)->csum_bytes, root->sectorsize); |
| 5120 | num_csums = num_csums + num_csums_per_leaf - 1; | 5120 | num_csums = num_csums + num_csums_per_leaf - 1; |
| 5121 | num_csums = num_csums / num_csums_per_leaf; | 5121 | num_csums = num_csums / num_csums_per_leaf; |
| 5122 | 5122 | ||
| 5123 | old_csums = old_csums + num_csums_per_leaf - 1; | 5123 | old_csums = old_csums + num_csums_per_leaf - 1; |
| 5124 | old_csums = old_csums / num_csums_per_leaf; | 5124 | old_csums = old_csums / num_csums_per_leaf; |
| 5125 | 5125 | ||
| 5126 | /* No change, no need to reserve more */ | 5126 | /* No change, no need to reserve more */ |
| 5127 | if (old_csums == num_csums) | 5127 | if (old_csums == num_csums) |
| 5128 | return 0; | 5128 | return 0; |
| 5129 | 5129 | ||
| 5130 | if (reserve) | 5130 | if (reserve) |
| 5131 | return btrfs_calc_trans_metadata_size(root, | 5131 | return btrfs_calc_trans_metadata_size(root, |
| 5132 | num_csums - old_csums); | 5132 | num_csums - old_csums); |
| 5133 | 5133 | ||
| 5134 | return btrfs_calc_trans_metadata_size(root, old_csums - num_csums); | 5134 | return btrfs_calc_trans_metadata_size(root, old_csums - num_csums); |
| 5135 | } | 5135 | } |
| 5136 | 5136 | ||
| 5137 | int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) | 5137 | int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) |
| 5138 | { | 5138 | { |
| 5139 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5139 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5140 | struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; | 5140 | struct btrfs_block_rsv *block_rsv = &root->fs_info->delalloc_block_rsv; |
| 5141 | u64 to_reserve = 0; | 5141 | u64 to_reserve = 0; |
| 5142 | u64 csum_bytes; | 5142 | u64 csum_bytes; |
| 5143 | unsigned nr_extents = 0; | 5143 | unsigned nr_extents = 0; |
| 5144 | int extra_reserve = 0; | 5144 | int extra_reserve = 0; |
| 5145 | enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_FLUSH_ALL; | 5145 | enum btrfs_reserve_flush_enum flush = BTRFS_RESERVE_FLUSH_ALL; |
| 5146 | int ret = 0; | 5146 | int ret = 0; |
| 5147 | bool delalloc_lock = true; | 5147 | bool delalloc_lock = true; |
| 5148 | u64 to_free = 0; | 5148 | u64 to_free = 0; |
| 5149 | unsigned dropped; | 5149 | unsigned dropped; |
| 5150 | 5150 | ||
| 5151 | /* If we are a free space inode we need to not flush since we will be in | 5151 | /* If we are a free space inode we need to not flush since we will be in |
| 5152 | * the middle of a transaction commit. We also don't need the delalloc | 5152 | * the middle of a transaction commit. We also don't need the delalloc |
| 5153 | * mutex since we won't race with anybody. We need this mostly to make | 5153 | * mutex since we won't race with anybody. We need this mostly to make |
| 5154 | * lockdep shut its filthy mouth. | 5154 | * lockdep shut its filthy mouth. |
| 5155 | */ | 5155 | */ |
| 5156 | if (btrfs_is_free_space_inode(inode)) { | 5156 | if (btrfs_is_free_space_inode(inode)) { |
| 5157 | flush = BTRFS_RESERVE_NO_FLUSH; | 5157 | flush = BTRFS_RESERVE_NO_FLUSH; |
| 5158 | delalloc_lock = false; | 5158 | delalloc_lock = false; |
| 5159 | } | 5159 | } |
| 5160 | 5160 | ||
| 5161 | if (flush != BTRFS_RESERVE_NO_FLUSH && | 5161 | if (flush != BTRFS_RESERVE_NO_FLUSH && |
| 5162 | btrfs_transaction_in_commit(root->fs_info)) | 5162 | btrfs_transaction_in_commit(root->fs_info)) |
| 5163 | schedule_timeout(1); | 5163 | schedule_timeout(1); |
| 5164 | 5164 | ||
| 5165 | if (delalloc_lock) | 5165 | if (delalloc_lock) |
| 5166 | mutex_lock(&BTRFS_I(inode)->delalloc_mutex); | 5166 | mutex_lock(&BTRFS_I(inode)->delalloc_mutex); |
| 5167 | 5167 | ||
| 5168 | num_bytes = ALIGN(num_bytes, root->sectorsize); | 5168 | num_bytes = ALIGN(num_bytes, root->sectorsize); |
| 5169 | 5169 | ||
| 5170 | spin_lock(&BTRFS_I(inode)->lock); | 5170 | spin_lock(&BTRFS_I(inode)->lock); |
| 5171 | BTRFS_I(inode)->outstanding_extents++; | 5171 | BTRFS_I(inode)->outstanding_extents++; |
| 5172 | 5172 | ||
| 5173 | if (BTRFS_I(inode)->outstanding_extents > | 5173 | if (BTRFS_I(inode)->outstanding_extents > |
| 5174 | BTRFS_I(inode)->reserved_extents) | 5174 | BTRFS_I(inode)->reserved_extents) |
| 5175 | nr_extents = BTRFS_I(inode)->outstanding_extents - | 5175 | nr_extents = BTRFS_I(inode)->outstanding_extents - |
| 5176 | BTRFS_I(inode)->reserved_extents; | 5176 | BTRFS_I(inode)->reserved_extents; |
| 5177 | 5177 | ||
| 5178 | /* | 5178 | /* |
| 5179 | * Add an item to reserve for updating the inode when we complete the | 5179 | * Add an item to reserve for updating the inode when we complete the |
| 5180 | * delalloc io. | 5180 | * delalloc io. |
| 5181 | */ | 5181 | */ |
| 5182 | if (!test_bit(BTRFS_INODE_DELALLOC_META_RESERVED, | 5182 | if (!test_bit(BTRFS_INODE_DELALLOC_META_RESERVED, |
| 5183 | &BTRFS_I(inode)->runtime_flags)) { | 5183 | &BTRFS_I(inode)->runtime_flags)) { |
| 5184 | nr_extents++; | 5184 | nr_extents++; |
| 5185 | extra_reserve = 1; | 5185 | extra_reserve = 1; |
| 5186 | } | 5186 | } |
| 5187 | 5187 | ||
| 5188 | to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); | 5188 | to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); |
| 5189 | to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); | 5189 | to_reserve += calc_csum_metadata_size(inode, num_bytes, 1); |
| 5190 | csum_bytes = BTRFS_I(inode)->csum_bytes; | 5190 | csum_bytes = BTRFS_I(inode)->csum_bytes; |
| 5191 | spin_unlock(&BTRFS_I(inode)->lock); | 5191 | spin_unlock(&BTRFS_I(inode)->lock); |
| 5192 | 5192 | ||
| 5193 | if (root->fs_info->quota_enabled) { | 5193 | if (root->fs_info->quota_enabled) { |
| 5194 | ret = btrfs_qgroup_reserve(root, num_bytes + | 5194 | ret = btrfs_qgroup_reserve(root, num_bytes + |
| 5195 | nr_extents * root->nodesize); | 5195 | nr_extents * root->nodesize); |
| 5196 | if (ret) | 5196 | if (ret) |
| 5197 | goto out_fail; | 5197 | goto out_fail; |
| 5198 | } | 5198 | } |
| 5199 | 5199 | ||
| 5200 | ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); | 5200 | ret = reserve_metadata_bytes(root, block_rsv, to_reserve, flush); |
| 5201 | if (unlikely(ret)) { | 5201 | if (unlikely(ret)) { |
| 5202 | if (root->fs_info->quota_enabled) | 5202 | if (root->fs_info->quota_enabled) |
| 5203 | btrfs_qgroup_free(root, num_bytes + | 5203 | btrfs_qgroup_free(root, num_bytes + |
| 5204 | nr_extents * root->nodesize); | 5204 | nr_extents * root->nodesize); |
| 5205 | goto out_fail; | 5205 | goto out_fail; |
| 5206 | } | 5206 | } |
| 5207 | 5207 | ||
| 5208 | spin_lock(&BTRFS_I(inode)->lock); | 5208 | spin_lock(&BTRFS_I(inode)->lock); |
| 5209 | if (extra_reserve) { | 5209 | if (extra_reserve) { |
| 5210 | set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, | 5210 | set_bit(BTRFS_INODE_DELALLOC_META_RESERVED, |
| 5211 | &BTRFS_I(inode)->runtime_flags); | 5211 | &BTRFS_I(inode)->runtime_flags); |
| 5212 | nr_extents--; | 5212 | nr_extents--; |
| 5213 | } | 5213 | } |
| 5214 | BTRFS_I(inode)->reserved_extents += nr_extents; | 5214 | BTRFS_I(inode)->reserved_extents += nr_extents; |
| 5215 | spin_unlock(&BTRFS_I(inode)->lock); | 5215 | spin_unlock(&BTRFS_I(inode)->lock); |
| 5216 | 5216 | ||
| 5217 | if (delalloc_lock) | 5217 | if (delalloc_lock) |
| 5218 | mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); | 5218 | mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); |
| 5219 | 5219 | ||
| 5220 | if (to_reserve) | 5220 | if (to_reserve) |
| 5221 | trace_btrfs_space_reservation(root->fs_info, "delalloc", | 5221 | trace_btrfs_space_reservation(root->fs_info, "delalloc", |
| 5222 | btrfs_ino(inode), to_reserve, 1); | 5222 | btrfs_ino(inode), to_reserve, 1); |
| 5223 | block_rsv_add_bytes(block_rsv, to_reserve, 1); | 5223 | block_rsv_add_bytes(block_rsv, to_reserve, 1); |
| 5224 | 5224 | ||
| 5225 | return 0; | 5225 | return 0; |
| 5226 | 5226 | ||
| 5227 | out_fail: | 5227 | out_fail: |
| 5228 | spin_lock(&BTRFS_I(inode)->lock); | 5228 | spin_lock(&BTRFS_I(inode)->lock); |
| 5229 | dropped = drop_outstanding_extent(inode); | 5229 | dropped = drop_outstanding_extent(inode); |
| 5230 | /* | 5230 | /* |
| 5231 | * If the inodes csum_bytes is the same as the original | 5231 | * If the inodes csum_bytes is the same as the original |
| 5232 | * csum_bytes then we know we haven't raced with any free()ers | 5232 | * csum_bytes then we know we haven't raced with any free()ers |
| 5233 | * so we can just reduce our inodes csum bytes and carry on. | 5233 | * so we can just reduce our inodes csum bytes and carry on. |
| 5234 | */ | 5234 | */ |
| 5235 | if (BTRFS_I(inode)->csum_bytes == csum_bytes) { | 5235 | if (BTRFS_I(inode)->csum_bytes == csum_bytes) { |
| 5236 | calc_csum_metadata_size(inode, num_bytes, 0); | 5236 | calc_csum_metadata_size(inode, num_bytes, 0); |
| 5237 | } else { | 5237 | } else { |
| 5238 | u64 orig_csum_bytes = BTRFS_I(inode)->csum_bytes; | 5238 | u64 orig_csum_bytes = BTRFS_I(inode)->csum_bytes; |
| 5239 | u64 bytes; | 5239 | u64 bytes; |
| 5240 | 5240 | ||
| 5241 | /* | 5241 | /* |
| 5242 | * This is tricky, but first we need to figure out how much we | 5242 | * This is tricky, but first we need to figure out how much we |
| 5243 | * free'd from any free-ers that occured during this | 5243 | * free'd from any free-ers that occured during this |
| 5244 | * reservation, so we reset ->csum_bytes to the csum_bytes | 5244 | * reservation, so we reset ->csum_bytes to the csum_bytes |
| 5245 | * before we dropped our lock, and then call the free for the | 5245 | * before we dropped our lock, and then call the free for the |
| 5246 | * number of bytes that were freed while we were trying our | 5246 | * number of bytes that were freed while we were trying our |
| 5247 | * reservation. | 5247 | * reservation. |
| 5248 | */ | 5248 | */ |
| 5249 | bytes = csum_bytes - BTRFS_I(inode)->csum_bytes; | 5249 | bytes = csum_bytes - BTRFS_I(inode)->csum_bytes; |
| 5250 | BTRFS_I(inode)->csum_bytes = csum_bytes; | 5250 | BTRFS_I(inode)->csum_bytes = csum_bytes; |
| 5251 | to_free = calc_csum_metadata_size(inode, bytes, 0); | 5251 | to_free = calc_csum_metadata_size(inode, bytes, 0); |
| 5252 | 5252 | ||
| 5253 | 5253 | ||
| 5254 | /* | 5254 | /* |
| 5255 | * Now we need to see how much we would have freed had we not | 5255 | * Now we need to see how much we would have freed had we not |
| 5256 | * been making this reservation and our ->csum_bytes were not | 5256 | * been making this reservation and our ->csum_bytes were not |
| 5257 | * artificially inflated. | 5257 | * artificially inflated. |
| 5258 | */ | 5258 | */ |
| 5259 | BTRFS_I(inode)->csum_bytes = csum_bytes - num_bytes; | 5259 | BTRFS_I(inode)->csum_bytes = csum_bytes - num_bytes; |
| 5260 | bytes = csum_bytes - orig_csum_bytes; | 5260 | bytes = csum_bytes - orig_csum_bytes; |
| 5261 | bytes = calc_csum_metadata_size(inode, bytes, 0); | 5261 | bytes = calc_csum_metadata_size(inode, bytes, 0); |
| 5262 | 5262 | ||
| 5263 | /* | 5263 | /* |
| 5264 | * Now reset ->csum_bytes to what it should be. If bytes is | 5264 | * Now reset ->csum_bytes to what it should be. If bytes is |
| 5265 | * more than to_free then we would have free'd more space had we | 5265 | * more than to_free then we would have free'd more space had we |
| 5266 | * not had an artificially high ->csum_bytes, so we need to free | 5266 | * not had an artificially high ->csum_bytes, so we need to free |
| 5267 | * the remainder. If bytes is the same or less then we don't | 5267 | * the remainder. If bytes is the same or less then we don't |
| 5268 | * need to do anything, the other free-ers did the correct | 5268 | * need to do anything, the other free-ers did the correct |
| 5269 | * thing. | 5269 | * thing. |
| 5270 | */ | 5270 | */ |
| 5271 | BTRFS_I(inode)->csum_bytes = orig_csum_bytes - num_bytes; | 5271 | BTRFS_I(inode)->csum_bytes = orig_csum_bytes - num_bytes; |
| 5272 | if (bytes > to_free) | 5272 | if (bytes > to_free) |
| 5273 | to_free = bytes - to_free; | 5273 | to_free = bytes - to_free; |
| 5274 | else | 5274 | else |
| 5275 | to_free = 0; | 5275 | to_free = 0; |
| 5276 | } | 5276 | } |
| 5277 | spin_unlock(&BTRFS_I(inode)->lock); | 5277 | spin_unlock(&BTRFS_I(inode)->lock); |
| 5278 | if (dropped) | 5278 | if (dropped) |
| 5279 | to_free += btrfs_calc_trans_metadata_size(root, dropped); | 5279 | to_free += btrfs_calc_trans_metadata_size(root, dropped); |
| 5280 | 5280 | ||
| 5281 | if (to_free) { | 5281 | if (to_free) { |
| 5282 | btrfs_block_rsv_release(root, block_rsv, to_free); | 5282 | btrfs_block_rsv_release(root, block_rsv, to_free); |
| 5283 | trace_btrfs_space_reservation(root->fs_info, "delalloc", | 5283 | trace_btrfs_space_reservation(root->fs_info, "delalloc", |
| 5284 | btrfs_ino(inode), to_free, 0); | 5284 | btrfs_ino(inode), to_free, 0); |
| 5285 | } | 5285 | } |
| 5286 | if (delalloc_lock) | 5286 | if (delalloc_lock) |
| 5287 | mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); | 5287 | mutex_unlock(&BTRFS_I(inode)->delalloc_mutex); |
| 5288 | return ret; | 5288 | return ret; |
| 5289 | } | 5289 | } |
| 5290 | 5290 | ||
| 5291 | /** | 5291 | /** |
| 5292 | * btrfs_delalloc_release_metadata - release a metadata reservation for an inode | 5292 | * btrfs_delalloc_release_metadata - release a metadata reservation for an inode |
| 5293 | * @inode: the inode to release the reservation for | 5293 | * @inode: the inode to release the reservation for |
| 5294 | * @num_bytes: the number of bytes we're releasing | 5294 | * @num_bytes: the number of bytes we're releasing |
| 5295 | * | 5295 | * |
| 5296 | * This will release the metadata reservation for an inode. This can be called | 5296 | * This will release the metadata reservation for an inode. This can be called |
| 5297 | * once we complete IO for a given set of bytes to release their metadata | 5297 | * once we complete IO for a given set of bytes to release their metadata |
| 5298 | * reservations. | 5298 | * reservations. |
| 5299 | */ | 5299 | */ |
| 5300 | void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) | 5300 | void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) |
| 5301 | { | 5301 | { |
| 5302 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5302 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5303 | u64 to_free = 0; | 5303 | u64 to_free = 0; |
| 5304 | unsigned dropped; | 5304 | unsigned dropped; |
| 5305 | 5305 | ||
| 5306 | num_bytes = ALIGN(num_bytes, root->sectorsize); | 5306 | num_bytes = ALIGN(num_bytes, root->sectorsize); |
| 5307 | spin_lock(&BTRFS_I(inode)->lock); | 5307 | spin_lock(&BTRFS_I(inode)->lock); |
| 5308 | dropped = drop_outstanding_extent(inode); | 5308 | dropped = drop_outstanding_extent(inode); |
| 5309 | 5309 | ||
| 5310 | if (num_bytes) | 5310 | if (num_bytes) |
| 5311 | to_free = calc_csum_metadata_size(inode, num_bytes, 0); | 5311 | to_free = calc_csum_metadata_size(inode, num_bytes, 0); |
| 5312 | spin_unlock(&BTRFS_I(inode)->lock); | 5312 | spin_unlock(&BTRFS_I(inode)->lock); |
| 5313 | if (dropped > 0) | 5313 | if (dropped > 0) |
| 5314 | to_free += btrfs_calc_trans_metadata_size(root, dropped); | 5314 | to_free += btrfs_calc_trans_metadata_size(root, dropped); |
| 5315 | 5315 | ||
| 5316 | trace_btrfs_space_reservation(root->fs_info, "delalloc", | 5316 | trace_btrfs_space_reservation(root->fs_info, "delalloc", |
| 5317 | btrfs_ino(inode), to_free, 0); | 5317 | btrfs_ino(inode), to_free, 0); |
| 5318 | if (root->fs_info->quota_enabled) { | 5318 | if (root->fs_info->quota_enabled) { |
| 5319 | btrfs_qgroup_free(root, num_bytes + | 5319 | btrfs_qgroup_free(root, num_bytes + |
| 5320 | dropped * root->nodesize); | 5320 | dropped * root->nodesize); |
| 5321 | } | 5321 | } |
| 5322 | 5322 | ||
| 5323 | btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv, | 5323 | btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv, |
| 5324 | to_free); | 5324 | to_free); |
| 5325 | } | 5325 | } |
| 5326 | 5326 | ||
| 5327 | /** | 5327 | /** |
| 5328 | * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc | 5328 | * btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc |
| 5329 | * @inode: inode we're writing to | 5329 | * @inode: inode we're writing to |
| 5330 | * @num_bytes: the number of bytes we want to allocate | 5330 | * @num_bytes: the number of bytes we want to allocate |
| 5331 | * | 5331 | * |
| 5332 | * This will do the following things | 5332 | * This will do the following things |
| 5333 | * | 5333 | * |
| 5334 | * o reserve space in the data space info for num_bytes | 5334 | * o reserve space in the data space info for num_bytes |
| 5335 | * o reserve space in the metadata space info based on number of outstanding | 5335 | * o reserve space in the metadata space info based on number of outstanding |
| 5336 | * extents and how much csums will be needed | 5336 | * extents and how much csums will be needed |
| 5337 | * o add to the inodes ->delalloc_bytes | 5337 | * o add to the inodes ->delalloc_bytes |
| 5338 | * o add it to the fs_info's delalloc inodes list. | 5338 | * o add it to the fs_info's delalloc inodes list. |
| 5339 | * | 5339 | * |
| 5340 | * This will return 0 for success and -ENOSPC if there is no space left. | 5340 | * This will return 0 for success and -ENOSPC if there is no space left. |
| 5341 | */ | 5341 | */ |
| 5342 | int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes) | 5342 | int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes) |
| 5343 | { | 5343 | { |
| 5344 | int ret; | 5344 | int ret; |
| 5345 | 5345 | ||
| 5346 | ret = btrfs_check_data_free_space(inode, num_bytes); | 5346 | ret = btrfs_check_data_free_space(inode, num_bytes); |
| 5347 | if (ret) | 5347 | if (ret) |
| 5348 | return ret; | 5348 | return ret; |
| 5349 | 5349 | ||
| 5350 | ret = btrfs_delalloc_reserve_metadata(inode, num_bytes); | 5350 | ret = btrfs_delalloc_reserve_metadata(inode, num_bytes); |
| 5351 | if (ret) { | 5351 | if (ret) { |
| 5352 | btrfs_free_reserved_data_space(inode, num_bytes); | 5352 | btrfs_free_reserved_data_space(inode, num_bytes); |
| 5353 | return ret; | 5353 | return ret; |
| 5354 | } | 5354 | } |
| 5355 | 5355 | ||
| 5356 | return 0; | 5356 | return 0; |
| 5357 | } | 5357 | } |
| 5358 | 5358 | ||
| 5359 | /** | 5359 | /** |
| 5360 | * btrfs_delalloc_release_space - release data and metadata space for delalloc | 5360 | * btrfs_delalloc_release_space - release data and metadata space for delalloc |
| 5361 | * @inode: inode we're releasing space for | 5361 | * @inode: inode we're releasing space for |
| 5362 | * @num_bytes: the number of bytes we want to free up | 5362 | * @num_bytes: the number of bytes we want to free up |
| 5363 | * | 5363 | * |
| 5364 | * This must be matched with a call to btrfs_delalloc_reserve_space. This is | 5364 | * This must be matched with a call to btrfs_delalloc_reserve_space. This is |
| 5365 | * called in the case that we don't need the metadata AND data reservations | 5365 | * called in the case that we don't need the metadata AND data reservations |
| 5366 | * anymore. So if there is an error or we insert an inline extent. | 5366 | * anymore. So if there is an error or we insert an inline extent. |
| 5367 | * | 5367 | * |
| 5368 | * This function will release the metadata space that was not used and will | 5368 | * This function will release the metadata space that was not used and will |
| 5369 | * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes | 5369 | * decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes |
| 5370 | * list if there are no delalloc bytes left. | 5370 | * list if there are no delalloc bytes left. |
| 5371 | */ | 5371 | */ |
| 5372 | void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes) | 5372 | void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes) |
| 5373 | { | 5373 | { |
| 5374 | btrfs_delalloc_release_metadata(inode, num_bytes); | 5374 | btrfs_delalloc_release_metadata(inode, num_bytes); |
| 5375 | btrfs_free_reserved_data_space(inode, num_bytes); | 5375 | btrfs_free_reserved_data_space(inode, num_bytes); |
| 5376 | } | 5376 | } |
| 5377 | 5377 | ||
| 5378 | static int update_block_group(struct btrfs_root *root, | 5378 | static int update_block_group(struct btrfs_root *root, |
| 5379 | u64 bytenr, u64 num_bytes, int alloc) | 5379 | u64 bytenr, u64 num_bytes, int alloc) |
| 5380 | { | 5380 | { |
| 5381 | struct btrfs_block_group_cache *cache = NULL; | 5381 | struct btrfs_block_group_cache *cache = NULL; |
| 5382 | struct btrfs_fs_info *info = root->fs_info; | 5382 | struct btrfs_fs_info *info = root->fs_info; |
| 5383 | u64 total = num_bytes; | 5383 | u64 total = num_bytes; |
| 5384 | u64 old_val; | 5384 | u64 old_val; |
| 5385 | u64 byte_in_group; | 5385 | u64 byte_in_group; |
| 5386 | int factor; | 5386 | int factor; |
| 5387 | 5387 | ||
| 5388 | /* block accounting for super block */ | 5388 | /* block accounting for super block */ |
| 5389 | spin_lock(&info->delalloc_root_lock); | 5389 | spin_lock(&info->delalloc_root_lock); |
| 5390 | old_val = btrfs_super_bytes_used(info->super_copy); | 5390 | old_val = btrfs_super_bytes_used(info->super_copy); |
| 5391 | if (alloc) | 5391 | if (alloc) |
| 5392 | old_val += num_bytes; | 5392 | old_val += num_bytes; |
| 5393 | else | 5393 | else |
| 5394 | old_val -= num_bytes; | 5394 | old_val -= num_bytes; |
| 5395 | btrfs_set_super_bytes_used(info->super_copy, old_val); | 5395 | btrfs_set_super_bytes_used(info->super_copy, old_val); |
| 5396 | spin_unlock(&info->delalloc_root_lock); | 5396 | spin_unlock(&info->delalloc_root_lock); |
| 5397 | 5397 | ||
| 5398 | while (total) { | 5398 | while (total) { |
| 5399 | cache = btrfs_lookup_block_group(info, bytenr); | 5399 | cache = btrfs_lookup_block_group(info, bytenr); |
| 5400 | if (!cache) | 5400 | if (!cache) |
| 5401 | return -ENOENT; | 5401 | return -ENOENT; |
| 5402 | if (cache->flags & (BTRFS_BLOCK_GROUP_DUP | | 5402 | if (cache->flags & (BTRFS_BLOCK_GROUP_DUP | |
| 5403 | BTRFS_BLOCK_GROUP_RAID1 | | 5403 | BTRFS_BLOCK_GROUP_RAID1 | |
| 5404 | BTRFS_BLOCK_GROUP_RAID10)) | 5404 | BTRFS_BLOCK_GROUP_RAID10)) |
| 5405 | factor = 2; | 5405 | factor = 2; |
| 5406 | else | 5406 | else |
| 5407 | factor = 1; | 5407 | factor = 1; |
| 5408 | /* | 5408 | /* |
| 5409 | * If this block group has free space cache written out, we | 5409 | * If this block group has free space cache written out, we |
| 5410 | * need to make sure to load it if we are removing space. This | 5410 | * need to make sure to load it if we are removing space. This |
| 5411 | * is because we need the unpinning stage to actually add the | 5411 | * is because we need the unpinning stage to actually add the |
| 5412 | * space back to the block group, otherwise we will leak space. | 5412 | * space back to the block group, otherwise we will leak space. |
| 5413 | */ | 5413 | */ |
| 5414 | if (!alloc && cache->cached == BTRFS_CACHE_NO) | 5414 | if (!alloc && cache->cached == BTRFS_CACHE_NO) |
| 5415 | cache_block_group(cache, 1); | 5415 | cache_block_group(cache, 1); |
| 5416 | 5416 | ||
| 5417 | byte_in_group = bytenr - cache->key.objectid; | 5417 | byte_in_group = bytenr - cache->key.objectid; |
| 5418 | WARN_ON(byte_in_group > cache->key.offset); | 5418 | WARN_ON(byte_in_group > cache->key.offset); |
| 5419 | 5419 | ||
| 5420 | spin_lock(&cache->space_info->lock); | 5420 | spin_lock(&cache->space_info->lock); |
| 5421 | spin_lock(&cache->lock); | 5421 | spin_lock(&cache->lock); |
| 5422 | 5422 | ||
| 5423 | if (btrfs_test_opt(root, SPACE_CACHE) && | 5423 | if (btrfs_test_opt(root, SPACE_CACHE) && |
| 5424 | cache->disk_cache_state < BTRFS_DC_CLEAR) | 5424 | cache->disk_cache_state < BTRFS_DC_CLEAR) |
| 5425 | cache->disk_cache_state = BTRFS_DC_CLEAR; | 5425 | cache->disk_cache_state = BTRFS_DC_CLEAR; |
| 5426 | 5426 | ||
| 5427 | cache->dirty = 1; | 5427 | cache->dirty = 1; |
| 5428 | old_val = btrfs_block_group_used(&cache->item); | 5428 | old_val = btrfs_block_group_used(&cache->item); |
| 5429 | num_bytes = min(total, cache->key.offset - byte_in_group); | 5429 | num_bytes = min(total, cache->key.offset - byte_in_group); |
| 5430 | if (alloc) { | 5430 | if (alloc) { |
| 5431 | old_val += num_bytes; | 5431 | old_val += num_bytes; |
| 5432 | btrfs_set_block_group_used(&cache->item, old_val); | 5432 | btrfs_set_block_group_used(&cache->item, old_val); |
| 5433 | cache->reserved -= num_bytes; | 5433 | cache->reserved -= num_bytes; |
| 5434 | cache->space_info->bytes_reserved -= num_bytes; | 5434 | cache->space_info->bytes_reserved -= num_bytes; |
| 5435 | cache->space_info->bytes_used += num_bytes; | 5435 | cache->space_info->bytes_used += num_bytes; |
| 5436 | cache->space_info->disk_used += num_bytes * factor; | 5436 | cache->space_info->disk_used += num_bytes * factor; |
| 5437 | spin_unlock(&cache->lock); | 5437 | spin_unlock(&cache->lock); |
| 5438 | spin_unlock(&cache->space_info->lock); | 5438 | spin_unlock(&cache->space_info->lock); |
| 5439 | } else { | 5439 | } else { |
| 5440 | old_val -= num_bytes; | 5440 | old_val -= num_bytes; |
| 5441 | btrfs_set_block_group_used(&cache->item, old_val); | 5441 | btrfs_set_block_group_used(&cache->item, old_val); |
| 5442 | cache->pinned += num_bytes; | 5442 | cache->pinned += num_bytes; |
| 5443 | cache->space_info->bytes_pinned += num_bytes; | 5443 | cache->space_info->bytes_pinned += num_bytes; |
| 5444 | cache->space_info->bytes_used -= num_bytes; | 5444 | cache->space_info->bytes_used -= num_bytes; |
| 5445 | cache->space_info->disk_used -= num_bytes * factor; | 5445 | cache->space_info->disk_used -= num_bytes * factor; |
| 5446 | spin_unlock(&cache->lock); | 5446 | spin_unlock(&cache->lock); |
| 5447 | spin_unlock(&cache->space_info->lock); | 5447 | spin_unlock(&cache->space_info->lock); |
| 5448 | 5448 | ||
| 5449 | set_extent_dirty(info->pinned_extents, | 5449 | set_extent_dirty(info->pinned_extents, |
| 5450 | bytenr, bytenr + num_bytes - 1, | 5450 | bytenr, bytenr + num_bytes - 1, |
| 5451 | GFP_NOFS | __GFP_NOFAIL); | 5451 | GFP_NOFS | __GFP_NOFAIL); |
| 5452 | /* | 5452 | /* |
| 5453 | * No longer have used bytes in this block group, queue | 5453 | * No longer have used bytes in this block group, queue |
| 5454 | * it for deletion. | 5454 | * it for deletion. |
| 5455 | */ | 5455 | */ |
| 5456 | if (old_val == 0) { | 5456 | if (old_val == 0) { |
| 5457 | spin_lock(&info->unused_bgs_lock); | 5457 | spin_lock(&info->unused_bgs_lock); |
| 5458 | if (list_empty(&cache->bg_list)) { | 5458 | if (list_empty(&cache->bg_list)) { |
| 5459 | btrfs_get_block_group(cache); | 5459 | btrfs_get_block_group(cache); |
| 5460 | list_add_tail(&cache->bg_list, | 5460 | list_add_tail(&cache->bg_list, |
| 5461 | &info->unused_bgs); | 5461 | &info->unused_bgs); |
| 5462 | } | 5462 | } |
| 5463 | spin_unlock(&info->unused_bgs_lock); | 5463 | spin_unlock(&info->unused_bgs_lock); |
| 5464 | } | 5464 | } |
| 5465 | } | 5465 | } |
| 5466 | btrfs_put_block_group(cache); | 5466 | btrfs_put_block_group(cache); |
| 5467 | total -= num_bytes; | 5467 | total -= num_bytes; |
| 5468 | bytenr += num_bytes; | 5468 | bytenr += num_bytes; |
| 5469 | } | 5469 | } |
| 5470 | return 0; | 5470 | return 0; |
| 5471 | } | 5471 | } |
| 5472 | 5472 | ||
| 5473 | static u64 first_logical_byte(struct btrfs_root *root, u64 search_start) | 5473 | static u64 first_logical_byte(struct btrfs_root *root, u64 search_start) |
| 5474 | { | 5474 | { |
| 5475 | struct btrfs_block_group_cache *cache; | 5475 | struct btrfs_block_group_cache *cache; |
| 5476 | u64 bytenr; | 5476 | u64 bytenr; |
| 5477 | 5477 | ||
| 5478 | spin_lock(&root->fs_info->block_group_cache_lock); | 5478 | spin_lock(&root->fs_info->block_group_cache_lock); |
| 5479 | bytenr = root->fs_info->first_logical_byte; | 5479 | bytenr = root->fs_info->first_logical_byte; |
| 5480 | spin_unlock(&root->fs_info->block_group_cache_lock); | 5480 | spin_unlock(&root->fs_info->block_group_cache_lock); |
| 5481 | 5481 | ||
| 5482 | if (bytenr < (u64)-1) | 5482 | if (bytenr < (u64)-1) |
| 5483 | return bytenr; | 5483 | return bytenr; |
| 5484 | 5484 | ||
| 5485 | cache = btrfs_lookup_first_block_group(root->fs_info, search_start); | 5485 | cache = btrfs_lookup_first_block_group(root->fs_info, search_start); |
| 5486 | if (!cache) | 5486 | if (!cache) |
| 5487 | return 0; | 5487 | return 0; |
| 5488 | 5488 | ||
| 5489 | bytenr = cache->key.objectid; | 5489 | bytenr = cache->key.objectid; |
| 5490 | btrfs_put_block_group(cache); | 5490 | btrfs_put_block_group(cache); |
| 5491 | 5491 | ||
| 5492 | return bytenr; | 5492 | return bytenr; |
| 5493 | } | 5493 | } |
| 5494 | 5494 | ||
| 5495 | static int pin_down_extent(struct btrfs_root *root, | 5495 | static int pin_down_extent(struct btrfs_root *root, |
| 5496 | struct btrfs_block_group_cache *cache, | 5496 | struct btrfs_block_group_cache *cache, |
| 5497 | u64 bytenr, u64 num_bytes, int reserved) | 5497 | u64 bytenr, u64 num_bytes, int reserved) |
| 5498 | { | 5498 | { |
| 5499 | spin_lock(&cache->space_info->lock); | 5499 | spin_lock(&cache->space_info->lock); |
| 5500 | spin_lock(&cache->lock); | 5500 | spin_lock(&cache->lock); |
| 5501 | cache->pinned += num_bytes; | 5501 | cache->pinned += num_bytes; |
| 5502 | cache->space_info->bytes_pinned += num_bytes; | 5502 | cache->space_info->bytes_pinned += num_bytes; |
| 5503 | if (reserved) { | 5503 | if (reserved) { |
| 5504 | cache->reserved -= num_bytes; | 5504 | cache->reserved -= num_bytes; |
| 5505 | cache->space_info->bytes_reserved -= num_bytes; | 5505 | cache->space_info->bytes_reserved -= num_bytes; |
| 5506 | } | 5506 | } |
| 5507 | spin_unlock(&cache->lock); | 5507 | spin_unlock(&cache->lock); |
| 5508 | spin_unlock(&cache->space_info->lock); | 5508 | spin_unlock(&cache->space_info->lock); |
| 5509 | 5509 | ||
| 5510 | set_extent_dirty(root->fs_info->pinned_extents, bytenr, | 5510 | set_extent_dirty(root->fs_info->pinned_extents, bytenr, |
| 5511 | bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); | 5511 | bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL); |
| 5512 | if (reserved) | 5512 | if (reserved) |
| 5513 | trace_btrfs_reserved_extent_free(root, bytenr, num_bytes); | 5513 | trace_btrfs_reserved_extent_free(root, bytenr, num_bytes); |
| 5514 | return 0; | 5514 | return 0; |
| 5515 | } | 5515 | } |
| 5516 | 5516 | ||
| 5517 | /* | 5517 | /* |
| 5518 | * this function must be called within transaction | 5518 | * this function must be called within transaction |
| 5519 | */ | 5519 | */ |
| 5520 | int btrfs_pin_extent(struct btrfs_root *root, | 5520 | int btrfs_pin_extent(struct btrfs_root *root, |
| 5521 | u64 bytenr, u64 num_bytes, int reserved) | 5521 | u64 bytenr, u64 num_bytes, int reserved) |
| 5522 | { | 5522 | { |
| 5523 | struct btrfs_block_group_cache *cache; | 5523 | struct btrfs_block_group_cache *cache; |
| 5524 | 5524 | ||
| 5525 | cache = btrfs_lookup_block_group(root->fs_info, bytenr); | 5525 | cache = btrfs_lookup_block_group(root->fs_info, bytenr); |
| 5526 | BUG_ON(!cache); /* Logic error */ | 5526 | BUG_ON(!cache); /* Logic error */ |
| 5527 | 5527 | ||
| 5528 | pin_down_extent(root, cache, bytenr, num_bytes, reserved); | 5528 | pin_down_extent(root, cache, bytenr, num_bytes, reserved); |
| 5529 | 5529 | ||
| 5530 | btrfs_put_block_group(cache); | 5530 | btrfs_put_block_group(cache); |
| 5531 | return 0; | 5531 | return 0; |
| 5532 | } | 5532 | } |
| 5533 | 5533 | ||
| 5534 | /* | 5534 | /* |
| 5535 | * this function must be called within transaction | 5535 | * this function must be called within transaction |
| 5536 | */ | 5536 | */ |
| 5537 | int btrfs_pin_extent_for_log_replay(struct btrfs_root *root, | 5537 | int btrfs_pin_extent_for_log_replay(struct btrfs_root *root, |
| 5538 | u64 bytenr, u64 num_bytes) | 5538 | u64 bytenr, u64 num_bytes) |
| 5539 | { | 5539 | { |
| 5540 | struct btrfs_block_group_cache *cache; | 5540 | struct btrfs_block_group_cache *cache; |
| 5541 | int ret; | 5541 | int ret; |
| 5542 | 5542 | ||
| 5543 | cache = btrfs_lookup_block_group(root->fs_info, bytenr); | 5543 | cache = btrfs_lookup_block_group(root->fs_info, bytenr); |
| 5544 | if (!cache) | 5544 | if (!cache) |
| 5545 | return -EINVAL; | 5545 | return -EINVAL; |
| 5546 | 5546 | ||
| 5547 | /* | 5547 | /* |
| 5548 | * pull in the free space cache (if any) so that our pin | 5548 | * pull in the free space cache (if any) so that our pin |
| 5549 | * removes the free space from the cache. We have load_only set | 5549 | * removes the free space from the cache. We have load_only set |
| 5550 | * to one because the slow code to read in the free extents does check | 5550 | * to one because the slow code to read in the free extents does check |
| 5551 | * the pinned extents. | 5551 | * the pinned extents. |
| 5552 | */ | 5552 | */ |
| 5553 | cache_block_group(cache, 1); | 5553 | cache_block_group(cache, 1); |
| 5554 | 5554 | ||
| 5555 | pin_down_extent(root, cache, bytenr, num_bytes, 0); | 5555 | pin_down_extent(root, cache, bytenr, num_bytes, 0); |
| 5556 | 5556 | ||
| 5557 | /* remove us from the free space cache (if we're there at all) */ | 5557 | /* remove us from the free space cache (if we're there at all) */ |
| 5558 | ret = btrfs_remove_free_space(cache, bytenr, num_bytes); | 5558 | ret = btrfs_remove_free_space(cache, bytenr, num_bytes); |
| 5559 | btrfs_put_block_group(cache); | 5559 | btrfs_put_block_group(cache); |
| 5560 | return ret; | 5560 | return ret; |
| 5561 | } | 5561 | } |
| 5562 | 5562 | ||
| 5563 | static int __exclude_logged_extent(struct btrfs_root *root, u64 start, u64 num_bytes) | 5563 | static int __exclude_logged_extent(struct btrfs_root *root, u64 start, u64 num_bytes) |
| 5564 | { | 5564 | { |
| 5565 | int ret; | 5565 | int ret; |
| 5566 | struct btrfs_block_group_cache *block_group; | 5566 | struct btrfs_block_group_cache *block_group; |
| 5567 | struct btrfs_caching_control *caching_ctl; | 5567 | struct btrfs_caching_control *caching_ctl; |
| 5568 | 5568 | ||
| 5569 | block_group = btrfs_lookup_block_group(root->fs_info, start); | 5569 | block_group = btrfs_lookup_block_group(root->fs_info, start); |
| 5570 | if (!block_group) | 5570 | if (!block_group) |
| 5571 | return -EINVAL; | 5571 | return -EINVAL; |
| 5572 | 5572 | ||
| 5573 | cache_block_group(block_group, 0); | 5573 | cache_block_group(block_group, 0); |
| 5574 | caching_ctl = get_caching_control(block_group); | 5574 | caching_ctl = get_caching_control(block_group); |
| 5575 | 5575 | ||
| 5576 | if (!caching_ctl) { | 5576 | if (!caching_ctl) { |
| 5577 | /* Logic error */ | 5577 | /* Logic error */ |
| 5578 | BUG_ON(!block_group_cache_done(block_group)); | 5578 | BUG_ON(!block_group_cache_done(block_group)); |
| 5579 | ret = btrfs_remove_free_space(block_group, start, num_bytes); | 5579 | ret = btrfs_remove_free_space(block_group, start, num_bytes); |
| 5580 | } else { | 5580 | } else { |
| 5581 | mutex_lock(&caching_ctl->mutex); | 5581 | mutex_lock(&caching_ctl->mutex); |
| 5582 | 5582 | ||
| 5583 | if (start >= caching_ctl->progress) { | 5583 | if (start >= caching_ctl->progress) { |
| 5584 | ret = add_excluded_extent(root, start, num_bytes); | 5584 | ret = add_excluded_extent(root, start, num_bytes); |
| 5585 | } else if (start + num_bytes <= caching_ctl->progress) { | 5585 | } else if (start + num_bytes <= caching_ctl->progress) { |
| 5586 | ret = btrfs_remove_free_space(block_group, | 5586 | ret = btrfs_remove_free_space(block_group, |
| 5587 | start, num_bytes); | 5587 | start, num_bytes); |
| 5588 | } else { | 5588 | } else { |
| 5589 | num_bytes = caching_ctl->progress - start; | 5589 | num_bytes = caching_ctl->progress - start; |
| 5590 | ret = btrfs_remove_free_space(block_group, | 5590 | ret = btrfs_remove_free_space(block_group, |
| 5591 | start, num_bytes); | 5591 | start, num_bytes); |
| 5592 | if (ret) | 5592 | if (ret) |
| 5593 | goto out_lock; | 5593 | goto out_lock; |
| 5594 | 5594 | ||
| 5595 | num_bytes = (start + num_bytes) - | 5595 | num_bytes = (start + num_bytes) - |
| 5596 | caching_ctl->progress; | 5596 | caching_ctl->progress; |
| 5597 | start = caching_ctl->progress; | 5597 | start = caching_ctl->progress; |
| 5598 | ret = add_excluded_extent(root, start, num_bytes); | 5598 | ret = add_excluded_extent(root, start, num_bytes); |
| 5599 | } | 5599 | } |
| 5600 | out_lock: | 5600 | out_lock: |
| 5601 | mutex_unlock(&caching_ctl->mutex); | 5601 | mutex_unlock(&caching_ctl->mutex); |
| 5602 | put_caching_control(caching_ctl); | 5602 | put_caching_control(caching_ctl); |
| 5603 | } | 5603 | } |
| 5604 | btrfs_put_block_group(block_group); | 5604 | btrfs_put_block_group(block_group); |
| 5605 | return ret; | 5605 | return ret; |
| 5606 | } | 5606 | } |
| 5607 | 5607 | ||
| 5608 | int btrfs_exclude_logged_extents(struct btrfs_root *log, | 5608 | int btrfs_exclude_logged_extents(struct btrfs_root *log, |
| 5609 | struct extent_buffer *eb) | 5609 | struct extent_buffer *eb) |
| 5610 | { | 5610 | { |
| 5611 | struct btrfs_file_extent_item *item; | 5611 | struct btrfs_file_extent_item *item; |
| 5612 | struct btrfs_key key; | 5612 | struct btrfs_key key; |
| 5613 | int found_type; | 5613 | int found_type; |
| 5614 | int i; | 5614 | int i; |
| 5615 | 5615 | ||
| 5616 | if (!btrfs_fs_incompat(log->fs_info, MIXED_GROUPS)) | 5616 | if (!btrfs_fs_incompat(log->fs_info, MIXED_GROUPS)) |
| 5617 | return 0; | 5617 | return 0; |
| 5618 | 5618 | ||
| 5619 | for (i = 0; i < btrfs_header_nritems(eb); i++) { | 5619 | for (i = 0; i < btrfs_header_nritems(eb); i++) { |
| 5620 | btrfs_item_key_to_cpu(eb, &key, i); | 5620 | btrfs_item_key_to_cpu(eb, &key, i); |
| 5621 | if (key.type != BTRFS_EXTENT_DATA_KEY) | 5621 | if (key.type != BTRFS_EXTENT_DATA_KEY) |
| 5622 | continue; | 5622 | continue; |
| 5623 | item = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item); | 5623 | item = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item); |
| 5624 | found_type = btrfs_file_extent_type(eb, item); | 5624 | found_type = btrfs_file_extent_type(eb, item); |
| 5625 | if (found_type == BTRFS_FILE_EXTENT_INLINE) | 5625 | if (found_type == BTRFS_FILE_EXTENT_INLINE) |
| 5626 | continue; | 5626 | continue; |
| 5627 | if (btrfs_file_extent_disk_bytenr(eb, item) == 0) | 5627 | if (btrfs_file_extent_disk_bytenr(eb, item) == 0) |
| 5628 | continue; | 5628 | continue; |
| 5629 | key.objectid = btrfs_file_extent_disk_bytenr(eb, item); | 5629 | key.objectid = btrfs_file_extent_disk_bytenr(eb, item); |
| 5630 | key.offset = btrfs_file_extent_disk_num_bytes(eb, item); | 5630 | key.offset = btrfs_file_extent_disk_num_bytes(eb, item); |
| 5631 | __exclude_logged_extent(log, key.objectid, key.offset); | 5631 | __exclude_logged_extent(log, key.objectid, key.offset); |
| 5632 | } | 5632 | } |
| 5633 | 5633 | ||
| 5634 | return 0; | 5634 | return 0; |
| 5635 | } | 5635 | } |
| 5636 | 5636 | ||
| 5637 | /** | 5637 | /** |
| 5638 | * btrfs_update_reserved_bytes - update the block_group and space info counters | 5638 | * btrfs_update_reserved_bytes - update the block_group and space info counters |
| 5639 | * @cache: The cache we are manipulating | 5639 | * @cache: The cache we are manipulating |
| 5640 | * @num_bytes: The number of bytes in question | 5640 | * @num_bytes: The number of bytes in question |
| 5641 | * @reserve: One of the reservation enums | 5641 | * @reserve: One of the reservation enums |
| 5642 | * @delalloc: The blocks are allocated for the delalloc write | 5642 | * @delalloc: The blocks are allocated for the delalloc write |
| 5643 | * | 5643 | * |
| 5644 | * This is called by the allocator when it reserves space, or by somebody who is | 5644 | * This is called by the allocator when it reserves space, or by somebody who is |
| 5645 | * freeing space that was never actually used on disk. For example if you | 5645 | * freeing space that was never actually used on disk. For example if you |
| 5646 | * reserve some space for a new leaf in transaction A and before transaction A | 5646 | * reserve some space for a new leaf in transaction A and before transaction A |
| 5647 | * commits you free that leaf, you call this with reserve set to 0 in order to | 5647 | * commits you free that leaf, you call this with reserve set to 0 in order to |
| 5648 | * clear the reservation. | 5648 | * clear the reservation. |
| 5649 | * | 5649 | * |
| 5650 | * Metadata reservations should be called with RESERVE_ALLOC so we do the proper | 5650 | * Metadata reservations should be called with RESERVE_ALLOC so we do the proper |
| 5651 | * ENOSPC accounting. For data we handle the reservation through clearing the | 5651 | * ENOSPC accounting. For data we handle the reservation through clearing the |
| 5652 | * delalloc bits in the io_tree. We have to do this since we could end up | 5652 | * delalloc bits in the io_tree. We have to do this since we could end up |
| 5653 | * allocating less disk space for the amount of data we have reserved in the | 5653 | * allocating less disk space for the amount of data we have reserved in the |
| 5654 | * case of compression. | 5654 | * case of compression. |
| 5655 | * | 5655 | * |
| 5656 | * If this is a reservation and the block group has become read only we cannot | 5656 | * If this is a reservation and the block group has become read only we cannot |
| 5657 | * make the reservation and return -EAGAIN, otherwise this function always | 5657 | * make the reservation and return -EAGAIN, otherwise this function always |
| 5658 | * succeeds. | 5658 | * succeeds. |
| 5659 | */ | 5659 | */ |
| 5660 | static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, | 5660 | static int btrfs_update_reserved_bytes(struct btrfs_block_group_cache *cache, |
| 5661 | u64 num_bytes, int reserve, int delalloc) | 5661 | u64 num_bytes, int reserve, int delalloc) |
| 5662 | { | 5662 | { |
| 5663 | struct btrfs_space_info *space_info = cache->space_info; | 5663 | struct btrfs_space_info *space_info = cache->space_info; |
| 5664 | int ret = 0; | 5664 | int ret = 0; |
| 5665 | 5665 | ||
| 5666 | spin_lock(&space_info->lock); | 5666 | spin_lock(&space_info->lock); |
| 5667 | spin_lock(&cache->lock); | 5667 | spin_lock(&cache->lock); |
| 5668 | if (reserve != RESERVE_FREE) { | 5668 | if (reserve != RESERVE_FREE) { |
| 5669 | if (cache->ro) { | 5669 | if (cache->ro) { |
| 5670 | ret = -EAGAIN; | 5670 | ret = -EAGAIN; |
| 5671 | } else { | 5671 | } else { |
| 5672 | cache->reserved += num_bytes; | 5672 | cache->reserved += num_bytes; |
| 5673 | space_info->bytes_reserved += num_bytes; | 5673 | space_info->bytes_reserved += num_bytes; |
| 5674 | if (reserve == RESERVE_ALLOC) { | 5674 | if (reserve == RESERVE_ALLOC) { |
| 5675 | trace_btrfs_space_reservation(cache->fs_info, | 5675 | trace_btrfs_space_reservation(cache->fs_info, |
| 5676 | "space_info", space_info->flags, | 5676 | "space_info", space_info->flags, |
| 5677 | num_bytes, 0); | 5677 | num_bytes, 0); |
| 5678 | space_info->bytes_may_use -= num_bytes; | 5678 | space_info->bytes_may_use -= num_bytes; |
| 5679 | } | 5679 | } |
| 5680 | 5680 | ||
| 5681 | if (delalloc) | 5681 | if (delalloc) |
| 5682 | cache->delalloc_bytes += num_bytes; | 5682 | cache->delalloc_bytes += num_bytes; |
| 5683 | } | 5683 | } |
| 5684 | } else { | 5684 | } else { |
| 5685 | if (cache->ro) | 5685 | if (cache->ro) |
| 5686 | space_info->bytes_readonly += num_bytes; | 5686 | space_info->bytes_readonly += num_bytes; |
| 5687 | cache->reserved -= num_bytes; | 5687 | cache->reserved -= num_bytes; |
| 5688 | space_info->bytes_reserved -= num_bytes; | 5688 | space_info->bytes_reserved -= num_bytes; |
| 5689 | 5689 | ||
| 5690 | if (delalloc) | 5690 | if (delalloc) |
| 5691 | cache->delalloc_bytes -= num_bytes; | 5691 | cache->delalloc_bytes -= num_bytes; |
| 5692 | } | 5692 | } |
| 5693 | spin_unlock(&cache->lock); | 5693 | spin_unlock(&cache->lock); |
| 5694 | spin_unlock(&space_info->lock); | 5694 | spin_unlock(&space_info->lock); |
| 5695 | return ret; | 5695 | return ret; |
| 5696 | } | 5696 | } |
| 5697 | 5697 | ||
| 5698 | void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, | 5698 | void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans, |
| 5699 | struct btrfs_root *root) | 5699 | struct btrfs_root *root) |
| 5700 | { | 5700 | { |
| 5701 | struct btrfs_fs_info *fs_info = root->fs_info; | 5701 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 5702 | struct btrfs_caching_control *next; | 5702 | struct btrfs_caching_control *next; |
| 5703 | struct btrfs_caching_control *caching_ctl; | 5703 | struct btrfs_caching_control *caching_ctl; |
| 5704 | struct btrfs_block_group_cache *cache; | 5704 | struct btrfs_block_group_cache *cache; |
| 5705 | 5705 | ||
| 5706 | down_write(&fs_info->commit_root_sem); | 5706 | down_write(&fs_info->commit_root_sem); |
| 5707 | 5707 | ||
| 5708 | list_for_each_entry_safe(caching_ctl, next, | 5708 | list_for_each_entry_safe(caching_ctl, next, |
| 5709 | &fs_info->caching_block_groups, list) { | 5709 | &fs_info->caching_block_groups, list) { |
| 5710 | cache = caching_ctl->block_group; | 5710 | cache = caching_ctl->block_group; |
| 5711 | if (block_group_cache_done(cache)) { | 5711 | if (block_group_cache_done(cache)) { |
| 5712 | cache->last_byte_to_unpin = (u64)-1; | 5712 | cache->last_byte_to_unpin = (u64)-1; |
| 5713 | list_del_init(&caching_ctl->list); | 5713 | list_del_init(&caching_ctl->list); |
| 5714 | put_caching_control(caching_ctl); | 5714 | put_caching_control(caching_ctl); |
| 5715 | } else { | 5715 | } else { |
| 5716 | cache->last_byte_to_unpin = caching_ctl->progress; | 5716 | cache->last_byte_to_unpin = caching_ctl->progress; |
| 5717 | } | 5717 | } |
| 5718 | } | 5718 | } |
| 5719 | 5719 | ||
| 5720 | if (fs_info->pinned_extents == &fs_info->freed_extents[0]) | 5720 | if (fs_info->pinned_extents == &fs_info->freed_extents[0]) |
| 5721 | fs_info->pinned_extents = &fs_info->freed_extents[1]; | 5721 | fs_info->pinned_extents = &fs_info->freed_extents[1]; |
| 5722 | else | 5722 | else |
| 5723 | fs_info->pinned_extents = &fs_info->freed_extents[0]; | 5723 | fs_info->pinned_extents = &fs_info->freed_extents[0]; |
| 5724 | 5724 | ||
| 5725 | up_write(&fs_info->commit_root_sem); | 5725 | up_write(&fs_info->commit_root_sem); |
| 5726 | 5726 | ||
| 5727 | update_global_block_rsv(fs_info); | 5727 | update_global_block_rsv(fs_info); |
| 5728 | } | 5728 | } |
| 5729 | 5729 | ||
| 5730 | static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, | 5730 | static int unpin_extent_range(struct btrfs_root *root, u64 start, u64 end, |
| 5731 | const bool return_free_space) | 5731 | const bool return_free_space) |
| 5732 | { | 5732 | { |
| 5733 | struct btrfs_fs_info *fs_info = root->fs_info; | 5733 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 5734 | struct btrfs_block_group_cache *cache = NULL; | 5734 | struct btrfs_block_group_cache *cache = NULL; |
| 5735 | struct btrfs_space_info *space_info; | 5735 | struct btrfs_space_info *space_info; |
| 5736 | struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; | 5736 | struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv; |
| 5737 | u64 len; | 5737 | u64 len; |
| 5738 | bool readonly; | 5738 | bool readonly; |
| 5739 | 5739 | ||
| 5740 | while (start <= end) { | 5740 | while (start <= end) { |
| 5741 | readonly = false; | 5741 | readonly = false; |
| 5742 | if (!cache || | 5742 | if (!cache || |
| 5743 | start >= cache->key.objectid + cache->key.offset) { | 5743 | start >= cache->key.objectid + cache->key.offset) { |
| 5744 | if (cache) | 5744 | if (cache) |
| 5745 | btrfs_put_block_group(cache); | 5745 | btrfs_put_block_group(cache); |
| 5746 | cache = btrfs_lookup_block_group(fs_info, start); | 5746 | cache = btrfs_lookup_block_group(fs_info, start); |
| 5747 | BUG_ON(!cache); /* Logic error */ | 5747 | BUG_ON(!cache); /* Logic error */ |
| 5748 | } | 5748 | } |
| 5749 | 5749 | ||
| 5750 | len = cache->key.objectid + cache->key.offset - start; | 5750 | len = cache->key.objectid + cache->key.offset - start; |
| 5751 | len = min(len, end + 1 - start); | 5751 | len = min(len, end + 1 - start); |
| 5752 | 5752 | ||
| 5753 | if (start < cache->last_byte_to_unpin) { | 5753 | if (start < cache->last_byte_to_unpin) { |
| 5754 | len = min(len, cache->last_byte_to_unpin - start); | 5754 | len = min(len, cache->last_byte_to_unpin - start); |
| 5755 | if (return_free_space) | 5755 | if (return_free_space) |
| 5756 | btrfs_add_free_space(cache, start, len); | 5756 | btrfs_add_free_space(cache, start, len); |
| 5757 | } | 5757 | } |
| 5758 | 5758 | ||
| 5759 | start += len; | 5759 | start += len; |
| 5760 | space_info = cache->space_info; | 5760 | space_info = cache->space_info; |
| 5761 | 5761 | ||
| 5762 | spin_lock(&space_info->lock); | 5762 | spin_lock(&space_info->lock); |
| 5763 | spin_lock(&cache->lock); | 5763 | spin_lock(&cache->lock); |
| 5764 | cache->pinned -= len; | 5764 | cache->pinned -= len; |
| 5765 | space_info->bytes_pinned -= len; | 5765 | space_info->bytes_pinned -= len; |
| 5766 | percpu_counter_add(&space_info->total_bytes_pinned, -len); | 5766 | percpu_counter_add(&space_info->total_bytes_pinned, -len); |
| 5767 | if (cache->ro) { | 5767 | if (cache->ro) { |
| 5768 | space_info->bytes_readonly += len; | 5768 | space_info->bytes_readonly += len; |
| 5769 | readonly = true; | 5769 | readonly = true; |
| 5770 | } | 5770 | } |
| 5771 | spin_unlock(&cache->lock); | 5771 | spin_unlock(&cache->lock); |
| 5772 | if (!readonly && global_rsv->space_info == space_info) { | 5772 | if (!readonly && global_rsv->space_info == space_info) { |
| 5773 | spin_lock(&global_rsv->lock); | 5773 | spin_lock(&global_rsv->lock); |
| 5774 | if (!global_rsv->full) { | 5774 | if (!global_rsv->full) { |
| 5775 | len = min(len, global_rsv->size - | 5775 | len = min(len, global_rsv->size - |
| 5776 | global_rsv->reserved); | 5776 | global_rsv->reserved); |
| 5777 | global_rsv->reserved += len; | 5777 | global_rsv->reserved += len; |
| 5778 | space_info->bytes_may_use += len; | 5778 | space_info->bytes_may_use += len; |
| 5779 | if (global_rsv->reserved >= global_rsv->size) | 5779 | if (global_rsv->reserved >= global_rsv->size) |
| 5780 | global_rsv->full = 1; | 5780 | global_rsv->full = 1; |
| 5781 | } | 5781 | } |
| 5782 | spin_unlock(&global_rsv->lock); | 5782 | spin_unlock(&global_rsv->lock); |
| 5783 | } | 5783 | } |
| 5784 | spin_unlock(&space_info->lock); | 5784 | spin_unlock(&space_info->lock); |
| 5785 | } | 5785 | } |
| 5786 | 5786 | ||
| 5787 | if (cache) | 5787 | if (cache) |
| 5788 | btrfs_put_block_group(cache); | 5788 | btrfs_put_block_group(cache); |
| 5789 | return 0; | 5789 | return 0; |
| 5790 | } | 5790 | } |
| 5791 | 5791 | ||
| 5792 | int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, | 5792 | int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans, |
| 5793 | struct btrfs_root *root) | 5793 | struct btrfs_root *root) |
| 5794 | { | 5794 | { |
| 5795 | struct btrfs_fs_info *fs_info = root->fs_info; | 5795 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 5796 | struct extent_io_tree *unpin; | 5796 | struct extent_io_tree *unpin; |
| 5797 | u64 start; | 5797 | u64 start; |
| 5798 | u64 end; | 5798 | u64 end; |
| 5799 | int ret; | 5799 | int ret; |
| 5800 | 5800 | ||
| 5801 | if (trans->aborted) | 5801 | if (trans->aborted) |
| 5802 | return 0; | 5802 | return 0; |
| 5803 | 5803 | ||
| 5804 | if (fs_info->pinned_extents == &fs_info->freed_extents[0]) | 5804 | if (fs_info->pinned_extents == &fs_info->freed_extents[0]) |
| 5805 | unpin = &fs_info->freed_extents[1]; | 5805 | unpin = &fs_info->freed_extents[1]; |
| 5806 | else | 5806 | else |
| 5807 | unpin = &fs_info->freed_extents[0]; | 5807 | unpin = &fs_info->freed_extents[0]; |
| 5808 | 5808 | ||
| 5809 | while (1) { | 5809 | while (1) { |
| 5810 | ret = find_first_extent_bit(unpin, 0, &start, &end, | 5810 | ret = find_first_extent_bit(unpin, 0, &start, &end, |
| 5811 | EXTENT_DIRTY, NULL); | 5811 | EXTENT_DIRTY, NULL); |
| 5812 | if (ret) | 5812 | if (ret) |
| 5813 | break; | 5813 | break; |
| 5814 | 5814 | ||
| 5815 | if (btrfs_test_opt(root, DISCARD)) | 5815 | if (btrfs_test_opt(root, DISCARD)) |
| 5816 | ret = btrfs_discard_extent(root, start, | 5816 | ret = btrfs_discard_extent(root, start, |
| 5817 | end + 1 - start, NULL); | 5817 | end + 1 - start, NULL); |
| 5818 | 5818 | ||
| 5819 | clear_extent_dirty(unpin, start, end, GFP_NOFS); | 5819 | clear_extent_dirty(unpin, start, end, GFP_NOFS); |
| 5820 | unpin_extent_range(root, start, end, true); | 5820 | unpin_extent_range(root, start, end, true); |
| 5821 | cond_resched(); | 5821 | cond_resched(); |
| 5822 | } | 5822 | } |
| 5823 | 5823 | ||
| 5824 | return 0; | 5824 | return 0; |
| 5825 | } | 5825 | } |
| 5826 | 5826 | ||
| 5827 | static void add_pinned_bytes(struct btrfs_fs_info *fs_info, u64 num_bytes, | 5827 | static void add_pinned_bytes(struct btrfs_fs_info *fs_info, u64 num_bytes, |
| 5828 | u64 owner, u64 root_objectid) | 5828 | u64 owner, u64 root_objectid) |
| 5829 | { | 5829 | { |
| 5830 | struct btrfs_space_info *space_info; | 5830 | struct btrfs_space_info *space_info; |
| 5831 | u64 flags; | 5831 | u64 flags; |
| 5832 | 5832 | ||
| 5833 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 5833 | if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 5834 | if (root_objectid == BTRFS_CHUNK_TREE_OBJECTID) | 5834 | if (root_objectid == BTRFS_CHUNK_TREE_OBJECTID) |
| 5835 | flags = BTRFS_BLOCK_GROUP_SYSTEM; | 5835 | flags = BTRFS_BLOCK_GROUP_SYSTEM; |
| 5836 | else | 5836 | else |
| 5837 | flags = BTRFS_BLOCK_GROUP_METADATA; | 5837 | flags = BTRFS_BLOCK_GROUP_METADATA; |
| 5838 | } else { | 5838 | } else { |
| 5839 | flags = BTRFS_BLOCK_GROUP_DATA; | 5839 | flags = BTRFS_BLOCK_GROUP_DATA; |
| 5840 | } | 5840 | } |
| 5841 | 5841 | ||
| 5842 | space_info = __find_space_info(fs_info, flags); | 5842 | space_info = __find_space_info(fs_info, flags); |
| 5843 | BUG_ON(!space_info); /* Logic bug */ | 5843 | BUG_ON(!space_info); /* Logic bug */ |
| 5844 | percpu_counter_add(&space_info->total_bytes_pinned, num_bytes); | 5844 | percpu_counter_add(&space_info->total_bytes_pinned, num_bytes); |
| 5845 | } | 5845 | } |
| 5846 | 5846 | ||
| 5847 | 5847 | ||
| 5848 | static int __btrfs_free_extent(struct btrfs_trans_handle *trans, | 5848 | static int __btrfs_free_extent(struct btrfs_trans_handle *trans, |
| 5849 | struct btrfs_root *root, | 5849 | struct btrfs_root *root, |
| 5850 | u64 bytenr, u64 num_bytes, u64 parent, | 5850 | u64 bytenr, u64 num_bytes, u64 parent, |
| 5851 | u64 root_objectid, u64 owner_objectid, | 5851 | u64 root_objectid, u64 owner_objectid, |
| 5852 | u64 owner_offset, int refs_to_drop, | 5852 | u64 owner_offset, int refs_to_drop, |
| 5853 | struct btrfs_delayed_extent_op *extent_op, | 5853 | struct btrfs_delayed_extent_op *extent_op, |
| 5854 | int no_quota) | 5854 | int no_quota) |
| 5855 | { | 5855 | { |
| 5856 | struct btrfs_key key; | 5856 | struct btrfs_key key; |
| 5857 | struct btrfs_path *path; | 5857 | struct btrfs_path *path; |
| 5858 | struct btrfs_fs_info *info = root->fs_info; | 5858 | struct btrfs_fs_info *info = root->fs_info; |
| 5859 | struct btrfs_root *extent_root = info->extent_root; | 5859 | struct btrfs_root *extent_root = info->extent_root; |
| 5860 | struct extent_buffer *leaf; | 5860 | struct extent_buffer *leaf; |
| 5861 | struct btrfs_extent_item *ei; | 5861 | struct btrfs_extent_item *ei; |
| 5862 | struct btrfs_extent_inline_ref *iref; | 5862 | struct btrfs_extent_inline_ref *iref; |
| 5863 | int ret; | 5863 | int ret; |
| 5864 | int is_data; | 5864 | int is_data; |
| 5865 | int extent_slot = 0; | 5865 | int extent_slot = 0; |
| 5866 | int found_extent = 0; | 5866 | int found_extent = 0; |
| 5867 | int num_to_del = 1; | 5867 | int num_to_del = 1; |
| 5868 | u32 item_size; | 5868 | u32 item_size; |
| 5869 | u64 refs; | 5869 | u64 refs; |
| 5870 | int last_ref = 0; | 5870 | int last_ref = 0; |
| 5871 | enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_SUB_EXCL; | 5871 | enum btrfs_qgroup_operation_type type = BTRFS_QGROUP_OPER_SUB_EXCL; |
| 5872 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, | 5872 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, |
| 5873 | SKINNY_METADATA); | 5873 | SKINNY_METADATA); |
| 5874 | 5874 | ||
| 5875 | if (!info->quota_enabled || !is_fstree(root_objectid)) | 5875 | if (!info->quota_enabled || !is_fstree(root_objectid)) |
| 5876 | no_quota = 1; | 5876 | no_quota = 1; |
| 5877 | 5877 | ||
| 5878 | path = btrfs_alloc_path(); | 5878 | path = btrfs_alloc_path(); |
| 5879 | if (!path) | 5879 | if (!path) |
| 5880 | return -ENOMEM; | 5880 | return -ENOMEM; |
| 5881 | 5881 | ||
| 5882 | path->reada = 1; | 5882 | path->reada = 1; |
| 5883 | path->leave_spinning = 1; | 5883 | path->leave_spinning = 1; |
| 5884 | 5884 | ||
| 5885 | is_data = owner_objectid >= BTRFS_FIRST_FREE_OBJECTID; | 5885 | is_data = owner_objectid >= BTRFS_FIRST_FREE_OBJECTID; |
| 5886 | BUG_ON(!is_data && refs_to_drop != 1); | 5886 | BUG_ON(!is_data && refs_to_drop != 1); |
| 5887 | 5887 | ||
| 5888 | if (is_data) | 5888 | if (is_data) |
| 5889 | skinny_metadata = 0; | 5889 | skinny_metadata = 0; |
| 5890 | 5890 | ||
| 5891 | ret = lookup_extent_backref(trans, extent_root, path, &iref, | 5891 | ret = lookup_extent_backref(trans, extent_root, path, &iref, |
| 5892 | bytenr, num_bytes, parent, | 5892 | bytenr, num_bytes, parent, |
| 5893 | root_objectid, owner_objectid, | 5893 | root_objectid, owner_objectid, |
| 5894 | owner_offset); | 5894 | owner_offset); |
| 5895 | if (ret == 0) { | 5895 | if (ret == 0) { |
| 5896 | extent_slot = path->slots[0]; | 5896 | extent_slot = path->slots[0]; |
| 5897 | while (extent_slot >= 0) { | 5897 | while (extent_slot >= 0) { |
| 5898 | btrfs_item_key_to_cpu(path->nodes[0], &key, | 5898 | btrfs_item_key_to_cpu(path->nodes[0], &key, |
| 5899 | extent_slot); | 5899 | extent_slot); |
| 5900 | if (key.objectid != bytenr) | 5900 | if (key.objectid != bytenr) |
| 5901 | break; | 5901 | break; |
| 5902 | if (key.type == BTRFS_EXTENT_ITEM_KEY && | 5902 | if (key.type == BTRFS_EXTENT_ITEM_KEY && |
| 5903 | key.offset == num_bytes) { | 5903 | key.offset == num_bytes) { |
| 5904 | found_extent = 1; | 5904 | found_extent = 1; |
| 5905 | break; | 5905 | break; |
| 5906 | } | 5906 | } |
| 5907 | if (key.type == BTRFS_METADATA_ITEM_KEY && | 5907 | if (key.type == BTRFS_METADATA_ITEM_KEY && |
| 5908 | key.offset == owner_objectid) { | 5908 | key.offset == owner_objectid) { |
| 5909 | found_extent = 1; | 5909 | found_extent = 1; |
| 5910 | break; | 5910 | break; |
| 5911 | } | 5911 | } |
| 5912 | if (path->slots[0] - extent_slot > 5) | 5912 | if (path->slots[0] - extent_slot > 5) |
| 5913 | break; | 5913 | break; |
| 5914 | extent_slot--; | 5914 | extent_slot--; |
| 5915 | } | 5915 | } |
| 5916 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 5916 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 5917 | item_size = btrfs_item_size_nr(path->nodes[0], extent_slot); | 5917 | item_size = btrfs_item_size_nr(path->nodes[0], extent_slot); |
| 5918 | if (found_extent && item_size < sizeof(*ei)) | 5918 | if (found_extent && item_size < sizeof(*ei)) |
| 5919 | found_extent = 0; | 5919 | found_extent = 0; |
| 5920 | #endif | 5920 | #endif |
| 5921 | if (!found_extent) { | 5921 | if (!found_extent) { |
| 5922 | BUG_ON(iref); | 5922 | BUG_ON(iref); |
| 5923 | ret = remove_extent_backref(trans, extent_root, path, | 5923 | ret = remove_extent_backref(trans, extent_root, path, |
| 5924 | NULL, refs_to_drop, | 5924 | NULL, refs_to_drop, |
| 5925 | is_data, &last_ref); | 5925 | is_data, &last_ref); |
| 5926 | if (ret) { | 5926 | if (ret) { |
| 5927 | btrfs_abort_transaction(trans, extent_root, ret); | 5927 | btrfs_abort_transaction(trans, extent_root, ret); |
| 5928 | goto out; | 5928 | goto out; |
| 5929 | } | 5929 | } |
| 5930 | btrfs_release_path(path); | 5930 | btrfs_release_path(path); |
| 5931 | path->leave_spinning = 1; | 5931 | path->leave_spinning = 1; |
| 5932 | 5932 | ||
| 5933 | key.objectid = bytenr; | 5933 | key.objectid = bytenr; |
| 5934 | key.type = BTRFS_EXTENT_ITEM_KEY; | 5934 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 5935 | key.offset = num_bytes; | 5935 | key.offset = num_bytes; |
| 5936 | 5936 | ||
| 5937 | if (!is_data && skinny_metadata) { | 5937 | if (!is_data && skinny_metadata) { |
| 5938 | key.type = BTRFS_METADATA_ITEM_KEY; | 5938 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 5939 | key.offset = owner_objectid; | 5939 | key.offset = owner_objectid; |
| 5940 | } | 5940 | } |
| 5941 | 5941 | ||
| 5942 | ret = btrfs_search_slot(trans, extent_root, | 5942 | ret = btrfs_search_slot(trans, extent_root, |
| 5943 | &key, path, -1, 1); | 5943 | &key, path, -1, 1); |
| 5944 | if (ret > 0 && skinny_metadata && path->slots[0]) { | 5944 | if (ret > 0 && skinny_metadata && path->slots[0]) { |
| 5945 | /* | 5945 | /* |
| 5946 | * Couldn't find our skinny metadata item, | 5946 | * Couldn't find our skinny metadata item, |
| 5947 | * see if we have ye olde extent item. | 5947 | * see if we have ye olde extent item. |
| 5948 | */ | 5948 | */ |
| 5949 | path->slots[0]--; | 5949 | path->slots[0]--; |
| 5950 | btrfs_item_key_to_cpu(path->nodes[0], &key, | 5950 | btrfs_item_key_to_cpu(path->nodes[0], &key, |
| 5951 | path->slots[0]); | 5951 | path->slots[0]); |
| 5952 | if (key.objectid == bytenr && | 5952 | if (key.objectid == bytenr && |
| 5953 | key.type == BTRFS_EXTENT_ITEM_KEY && | 5953 | key.type == BTRFS_EXTENT_ITEM_KEY && |
| 5954 | key.offset == num_bytes) | 5954 | key.offset == num_bytes) |
| 5955 | ret = 0; | 5955 | ret = 0; |
| 5956 | } | 5956 | } |
| 5957 | 5957 | ||
| 5958 | if (ret > 0 && skinny_metadata) { | 5958 | if (ret > 0 && skinny_metadata) { |
| 5959 | skinny_metadata = false; | 5959 | skinny_metadata = false; |
| 5960 | key.objectid = bytenr; | 5960 | key.objectid = bytenr; |
| 5961 | key.type = BTRFS_EXTENT_ITEM_KEY; | 5961 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 5962 | key.offset = num_bytes; | 5962 | key.offset = num_bytes; |
| 5963 | btrfs_release_path(path); | 5963 | btrfs_release_path(path); |
| 5964 | ret = btrfs_search_slot(trans, extent_root, | 5964 | ret = btrfs_search_slot(trans, extent_root, |
| 5965 | &key, path, -1, 1); | 5965 | &key, path, -1, 1); |
| 5966 | } | 5966 | } |
| 5967 | 5967 | ||
| 5968 | if (ret) { | 5968 | if (ret) { |
| 5969 | btrfs_err(info, "umm, got %d back from search, was looking for %llu", | 5969 | btrfs_err(info, "umm, got %d back from search, was looking for %llu", |
| 5970 | ret, bytenr); | 5970 | ret, bytenr); |
| 5971 | if (ret > 0) | 5971 | if (ret > 0) |
| 5972 | btrfs_print_leaf(extent_root, | 5972 | btrfs_print_leaf(extent_root, |
| 5973 | path->nodes[0]); | 5973 | path->nodes[0]); |
| 5974 | } | 5974 | } |
| 5975 | if (ret < 0) { | 5975 | if (ret < 0) { |
| 5976 | btrfs_abort_transaction(trans, extent_root, ret); | 5976 | btrfs_abort_transaction(trans, extent_root, ret); |
| 5977 | goto out; | 5977 | goto out; |
| 5978 | } | 5978 | } |
| 5979 | extent_slot = path->slots[0]; | 5979 | extent_slot = path->slots[0]; |
| 5980 | } | 5980 | } |
| 5981 | } else if (WARN_ON(ret == -ENOENT)) { | 5981 | } else if (WARN_ON(ret == -ENOENT)) { |
| 5982 | btrfs_print_leaf(extent_root, path->nodes[0]); | 5982 | btrfs_print_leaf(extent_root, path->nodes[0]); |
| 5983 | btrfs_err(info, | 5983 | btrfs_err(info, |
| 5984 | "unable to find ref byte nr %llu parent %llu root %llu owner %llu offset %llu", | 5984 | "unable to find ref byte nr %llu parent %llu root %llu owner %llu offset %llu", |
| 5985 | bytenr, parent, root_objectid, owner_objectid, | 5985 | bytenr, parent, root_objectid, owner_objectid, |
| 5986 | owner_offset); | 5986 | owner_offset); |
| 5987 | btrfs_abort_transaction(trans, extent_root, ret); | 5987 | btrfs_abort_transaction(trans, extent_root, ret); |
| 5988 | goto out; | 5988 | goto out; |
| 5989 | } else { | 5989 | } else { |
| 5990 | btrfs_abort_transaction(trans, extent_root, ret); | 5990 | btrfs_abort_transaction(trans, extent_root, ret); |
| 5991 | goto out; | 5991 | goto out; |
| 5992 | } | 5992 | } |
| 5993 | 5993 | ||
| 5994 | leaf = path->nodes[0]; | 5994 | leaf = path->nodes[0]; |
| 5995 | item_size = btrfs_item_size_nr(leaf, extent_slot); | 5995 | item_size = btrfs_item_size_nr(leaf, extent_slot); |
| 5996 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | 5996 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 |
| 5997 | if (item_size < sizeof(*ei)) { | 5997 | if (item_size < sizeof(*ei)) { |
| 5998 | BUG_ON(found_extent || extent_slot != path->slots[0]); | 5998 | BUG_ON(found_extent || extent_slot != path->slots[0]); |
| 5999 | ret = convert_extent_item_v0(trans, extent_root, path, | 5999 | ret = convert_extent_item_v0(trans, extent_root, path, |
| 6000 | owner_objectid, 0); | 6000 | owner_objectid, 0); |
| 6001 | if (ret < 0) { | 6001 | if (ret < 0) { |
| 6002 | btrfs_abort_transaction(trans, extent_root, ret); | 6002 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6003 | goto out; | 6003 | goto out; |
| 6004 | } | 6004 | } |
| 6005 | 6005 | ||
| 6006 | btrfs_release_path(path); | 6006 | btrfs_release_path(path); |
| 6007 | path->leave_spinning = 1; | 6007 | path->leave_spinning = 1; |
| 6008 | 6008 | ||
| 6009 | key.objectid = bytenr; | 6009 | key.objectid = bytenr; |
| 6010 | key.type = BTRFS_EXTENT_ITEM_KEY; | 6010 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 6011 | key.offset = num_bytes; | 6011 | key.offset = num_bytes; |
| 6012 | 6012 | ||
| 6013 | ret = btrfs_search_slot(trans, extent_root, &key, path, | 6013 | ret = btrfs_search_slot(trans, extent_root, &key, path, |
| 6014 | -1, 1); | 6014 | -1, 1); |
| 6015 | if (ret) { | 6015 | if (ret) { |
| 6016 | btrfs_err(info, "umm, got %d back from search, was looking for %llu", | 6016 | btrfs_err(info, "umm, got %d back from search, was looking for %llu", |
| 6017 | ret, bytenr); | 6017 | ret, bytenr); |
| 6018 | btrfs_print_leaf(extent_root, path->nodes[0]); | 6018 | btrfs_print_leaf(extent_root, path->nodes[0]); |
| 6019 | } | 6019 | } |
| 6020 | if (ret < 0) { | 6020 | if (ret < 0) { |
| 6021 | btrfs_abort_transaction(trans, extent_root, ret); | 6021 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6022 | goto out; | 6022 | goto out; |
| 6023 | } | 6023 | } |
| 6024 | 6024 | ||
| 6025 | extent_slot = path->slots[0]; | 6025 | extent_slot = path->slots[0]; |
| 6026 | leaf = path->nodes[0]; | 6026 | leaf = path->nodes[0]; |
| 6027 | item_size = btrfs_item_size_nr(leaf, extent_slot); | 6027 | item_size = btrfs_item_size_nr(leaf, extent_slot); |
| 6028 | } | 6028 | } |
| 6029 | #endif | 6029 | #endif |
| 6030 | BUG_ON(item_size < sizeof(*ei)); | 6030 | BUG_ON(item_size < sizeof(*ei)); |
| 6031 | ei = btrfs_item_ptr(leaf, extent_slot, | 6031 | ei = btrfs_item_ptr(leaf, extent_slot, |
| 6032 | struct btrfs_extent_item); | 6032 | struct btrfs_extent_item); |
| 6033 | if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID && | 6033 | if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID && |
| 6034 | key.type == BTRFS_EXTENT_ITEM_KEY) { | 6034 | key.type == BTRFS_EXTENT_ITEM_KEY) { |
| 6035 | struct btrfs_tree_block_info *bi; | 6035 | struct btrfs_tree_block_info *bi; |
| 6036 | BUG_ON(item_size < sizeof(*ei) + sizeof(*bi)); | 6036 | BUG_ON(item_size < sizeof(*ei) + sizeof(*bi)); |
| 6037 | bi = (struct btrfs_tree_block_info *)(ei + 1); | 6037 | bi = (struct btrfs_tree_block_info *)(ei + 1); |
| 6038 | WARN_ON(owner_objectid != btrfs_tree_block_level(leaf, bi)); | 6038 | WARN_ON(owner_objectid != btrfs_tree_block_level(leaf, bi)); |
| 6039 | } | 6039 | } |
| 6040 | 6040 | ||
| 6041 | refs = btrfs_extent_refs(leaf, ei); | 6041 | refs = btrfs_extent_refs(leaf, ei); |
| 6042 | if (refs < refs_to_drop) { | 6042 | if (refs < refs_to_drop) { |
| 6043 | btrfs_err(info, "trying to drop %d refs but we only have %Lu " | 6043 | btrfs_err(info, "trying to drop %d refs but we only have %Lu " |
| 6044 | "for bytenr %Lu", refs_to_drop, refs, bytenr); | 6044 | "for bytenr %Lu", refs_to_drop, refs, bytenr); |
| 6045 | ret = -EINVAL; | 6045 | ret = -EINVAL; |
| 6046 | btrfs_abort_transaction(trans, extent_root, ret); | 6046 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6047 | goto out; | 6047 | goto out; |
| 6048 | } | 6048 | } |
| 6049 | refs -= refs_to_drop; | 6049 | refs -= refs_to_drop; |
| 6050 | 6050 | ||
| 6051 | if (refs > 0) { | 6051 | if (refs > 0) { |
| 6052 | type = BTRFS_QGROUP_OPER_SUB_SHARED; | 6052 | type = BTRFS_QGROUP_OPER_SUB_SHARED; |
| 6053 | if (extent_op) | 6053 | if (extent_op) |
| 6054 | __run_delayed_extent_op(extent_op, leaf, ei); | 6054 | __run_delayed_extent_op(extent_op, leaf, ei); |
| 6055 | /* | 6055 | /* |
| 6056 | * In the case of inline back ref, reference count will | 6056 | * In the case of inline back ref, reference count will |
| 6057 | * be updated by remove_extent_backref | 6057 | * be updated by remove_extent_backref |
| 6058 | */ | 6058 | */ |
| 6059 | if (iref) { | 6059 | if (iref) { |
| 6060 | BUG_ON(!found_extent); | 6060 | BUG_ON(!found_extent); |
| 6061 | } else { | 6061 | } else { |
| 6062 | btrfs_set_extent_refs(leaf, ei, refs); | 6062 | btrfs_set_extent_refs(leaf, ei, refs); |
| 6063 | btrfs_mark_buffer_dirty(leaf); | 6063 | btrfs_mark_buffer_dirty(leaf); |
| 6064 | } | 6064 | } |
| 6065 | if (found_extent) { | 6065 | if (found_extent) { |
| 6066 | ret = remove_extent_backref(trans, extent_root, path, | 6066 | ret = remove_extent_backref(trans, extent_root, path, |
| 6067 | iref, refs_to_drop, | 6067 | iref, refs_to_drop, |
| 6068 | is_data, &last_ref); | 6068 | is_data, &last_ref); |
| 6069 | if (ret) { | 6069 | if (ret) { |
| 6070 | btrfs_abort_transaction(trans, extent_root, ret); | 6070 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6071 | goto out; | 6071 | goto out; |
| 6072 | } | 6072 | } |
| 6073 | } | 6073 | } |
| 6074 | add_pinned_bytes(root->fs_info, -num_bytes, owner_objectid, | 6074 | add_pinned_bytes(root->fs_info, -num_bytes, owner_objectid, |
| 6075 | root_objectid); | 6075 | root_objectid); |
| 6076 | } else { | 6076 | } else { |
| 6077 | if (found_extent) { | 6077 | if (found_extent) { |
| 6078 | BUG_ON(is_data && refs_to_drop != | 6078 | BUG_ON(is_data && refs_to_drop != |
| 6079 | extent_data_ref_count(root, path, iref)); | 6079 | extent_data_ref_count(root, path, iref)); |
| 6080 | if (iref) { | 6080 | if (iref) { |
| 6081 | BUG_ON(path->slots[0] != extent_slot); | 6081 | BUG_ON(path->slots[0] != extent_slot); |
| 6082 | } else { | 6082 | } else { |
| 6083 | BUG_ON(path->slots[0] != extent_slot + 1); | 6083 | BUG_ON(path->slots[0] != extent_slot + 1); |
| 6084 | path->slots[0] = extent_slot; | 6084 | path->slots[0] = extent_slot; |
| 6085 | num_to_del = 2; | 6085 | num_to_del = 2; |
| 6086 | } | 6086 | } |
| 6087 | } | 6087 | } |
| 6088 | 6088 | ||
| 6089 | last_ref = 1; | 6089 | last_ref = 1; |
| 6090 | ret = btrfs_del_items(trans, extent_root, path, path->slots[0], | 6090 | ret = btrfs_del_items(trans, extent_root, path, path->slots[0], |
| 6091 | num_to_del); | 6091 | num_to_del); |
| 6092 | if (ret) { | 6092 | if (ret) { |
| 6093 | btrfs_abort_transaction(trans, extent_root, ret); | 6093 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6094 | goto out; | 6094 | goto out; |
| 6095 | } | 6095 | } |
| 6096 | btrfs_release_path(path); | 6096 | btrfs_release_path(path); |
| 6097 | 6097 | ||
| 6098 | if (is_data) { | 6098 | if (is_data) { |
| 6099 | ret = btrfs_del_csums(trans, root, bytenr, num_bytes); | 6099 | ret = btrfs_del_csums(trans, root, bytenr, num_bytes); |
| 6100 | if (ret) { | 6100 | if (ret) { |
| 6101 | btrfs_abort_transaction(trans, extent_root, ret); | 6101 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6102 | goto out; | 6102 | goto out; |
| 6103 | } | 6103 | } |
| 6104 | } | 6104 | } |
| 6105 | 6105 | ||
| 6106 | ret = update_block_group(root, bytenr, num_bytes, 0); | 6106 | ret = update_block_group(root, bytenr, num_bytes, 0); |
| 6107 | if (ret) { | 6107 | if (ret) { |
| 6108 | btrfs_abort_transaction(trans, extent_root, ret); | 6108 | btrfs_abort_transaction(trans, extent_root, ret); |
| 6109 | goto out; | 6109 | goto out; |
| 6110 | } | 6110 | } |
| 6111 | } | 6111 | } |
| 6112 | btrfs_release_path(path); | 6112 | btrfs_release_path(path); |
| 6113 | 6113 | ||
| 6114 | /* Deal with the quota accounting */ | 6114 | /* Deal with the quota accounting */ |
| 6115 | if (!ret && last_ref && !no_quota) { | 6115 | if (!ret && last_ref && !no_quota) { |
| 6116 | int mod_seq = 0; | 6116 | int mod_seq = 0; |
| 6117 | 6117 | ||
| 6118 | if (owner_objectid >= BTRFS_FIRST_FREE_OBJECTID && | 6118 | if (owner_objectid >= BTRFS_FIRST_FREE_OBJECTID && |
| 6119 | type == BTRFS_QGROUP_OPER_SUB_SHARED) | 6119 | type == BTRFS_QGROUP_OPER_SUB_SHARED) |
| 6120 | mod_seq = 1; | 6120 | mod_seq = 1; |
| 6121 | 6121 | ||
| 6122 | ret = btrfs_qgroup_record_ref(trans, info, root_objectid, | 6122 | ret = btrfs_qgroup_record_ref(trans, info, root_objectid, |
| 6123 | bytenr, num_bytes, type, | 6123 | bytenr, num_bytes, type, |
| 6124 | mod_seq); | 6124 | mod_seq); |
| 6125 | } | 6125 | } |
| 6126 | out: | 6126 | out: |
| 6127 | btrfs_free_path(path); | 6127 | btrfs_free_path(path); |
| 6128 | return ret; | 6128 | return ret; |
| 6129 | } | 6129 | } |
| 6130 | 6130 | ||
| 6131 | /* | 6131 | /* |
| 6132 | * when we free an block, it is possible (and likely) that we free the last | 6132 | * when we free an block, it is possible (and likely) that we free the last |
| 6133 | * delayed ref for that extent as well. This searches the delayed ref tree for | 6133 | * delayed ref for that extent as well. This searches the delayed ref tree for |
| 6134 | * a given extent, and if there are no other delayed refs to be processed, it | 6134 | * a given extent, and if there are no other delayed refs to be processed, it |
| 6135 | * removes it from the tree. | 6135 | * removes it from the tree. |
| 6136 | */ | 6136 | */ |
| 6137 | static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans, | 6137 | static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans, |
| 6138 | struct btrfs_root *root, u64 bytenr) | 6138 | struct btrfs_root *root, u64 bytenr) |
| 6139 | { | 6139 | { |
| 6140 | struct btrfs_delayed_ref_head *head; | 6140 | struct btrfs_delayed_ref_head *head; |
| 6141 | struct btrfs_delayed_ref_root *delayed_refs; | 6141 | struct btrfs_delayed_ref_root *delayed_refs; |
| 6142 | int ret = 0; | 6142 | int ret = 0; |
| 6143 | 6143 | ||
| 6144 | delayed_refs = &trans->transaction->delayed_refs; | 6144 | delayed_refs = &trans->transaction->delayed_refs; |
| 6145 | spin_lock(&delayed_refs->lock); | 6145 | spin_lock(&delayed_refs->lock); |
| 6146 | head = btrfs_find_delayed_ref_head(trans, bytenr); | 6146 | head = btrfs_find_delayed_ref_head(trans, bytenr); |
| 6147 | if (!head) | 6147 | if (!head) |
| 6148 | goto out_delayed_unlock; | 6148 | goto out_delayed_unlock; |
| 6149 | 6149 | ||
| 6150 | spin_lock(&head->lock); | 6150 | spin_lock(&head->lock); |
| 6151 | if (rb_first(&head->ref_root)) | 6151 | if (rb_first(&head->ref_root)) |
| 6152 | goto out; | 6152 | goto out; |
| 6153 | 6153 | ||
| 6154 | if (head->extent_op) { | 6154 | if (head->extent_op) { |
| 6155 | if (!head->must_insert_reserved) | 6155 | if (!head->must_insert_reserved) |
| 6156 | goto out; | 6156 | goto out; |
| 6157 | btrfs_free_delayed_extent_op(head->extent_op); | 6157 | btrfs_free_delayed_extent_op(head->extent_op); |
| 6158 | head->extent_op = NULL; | 6158 | head->extent_op = NULL; |
| 6159 | } | 6159 | } |
| 6160 | 6160 | ||
| 6161 | /* | 6161 | /* |
| 6162 | * waiting for the lock here would deadlock. If someone else has it | 6162 | * waiting for the lock here would deadlock. If someone else has it |
| 6163 | * locked they are already in the process of dropping it anyway | 6163 | * locked they are already in the process of dropping it anyway |
| 6164 | */ | 6164 | */ |
| 6165 | if (!mutex_trylock(&head->mutex)) | 6165 | if (!mutex_trylock(&head->mutex)) |
| 6166 | goto out; | 6166 | goto out; |
| 6167 | 6167 | ||
| 6168 | /* | 6168 | /* |
| 6169 | * at this point we have a head with no other entries. Go | 6169 | * at this point we have a head with no other entries. Go |
| 6170 | * ahead and process it. | 6170 | * ahead and process it. |
| 6171 | */ | 6171 | */ |
| 6172 | head->node.in_tree = 0; | 6172 | head->node.in_tree = 0; |
| 6173 | rb_erase(&head->href_node, &delayed_refs->href_root); | 6173 | rb_erase(&head->href_node, &delayed_refs->href_root); |
| 6174 | 6174 | ||
| 6175 | atomic_dec(&delayed_refs->num_entries); | 6175 | atomic_dec(&delayed_refs->num_entries); |
| 6176 | 6176 | ||
| 6177 | /* | 6177 | /* |
| 6178 | * we don't take a ref on the node because we're removing it from the | 6178 | * we don't take a ref on the node because we're removing it from the |
| 6179 | * tree, so we just steal the ref the tree was holding. | 6179 | * tree, so we just steal the ref the tree was holding. |
| 6180 | */ | 6180 | */ |
| 6181 | delayed_refs->num_heads--; | 6181 | delayed_refs->num_heads--; |
| 6182 | if (head->processing == 0) | 6182 | if (head->processing == 0) |
| 6183 | delayed_refs->num_heads_ready--; | 6183 | delayed_refs->num_heads_ready--; |
| 6184 | head->processing = 0; | 6184 | head->processing = 0; |
| 6185 | spin_unlock(&head->lock); | 6185 | spin_unlock(&head->lock); |
| 6186 | spin_unlock(&delayed_refs->lock); | 6186 | spin_unlock(&delayed_refs->lock); |
| 6187 | 6187 | ||
| 6188 | BUG_ON(head->extent_op); | 6188 | BUG_ON(head->extent_op); |
| 6189 | if (head->must_insert_reserved) | 6189 | if (head->must_insert_reserved) |
| 6190 | ret = 1; | 6190 | ret = 1; |
| 6191 | 6191 | ||
| 6192 | mutex_unlock(&head->mutex); | 6192 | mutex_unlock(&head->mutex); |
| 6193 | btrfs_put_delayed_ref(&head->node); | 6193 | btrfs_put_delayed_ref(&head->node); |
| 6194 | return ret; | 6194 | return ret; |
| 6195 | out: | 6195 | out: |
| 6196 | spin_unlock(&head->lock); | 6196 | spin_unlock(&head->lock); |
| 6197 | 6197 | ||
| 6198 | out_delayed_unlock: | 6198 | out_delayed_unlock: |
| 6199 | spin_unlock(&delayed_refs->lock); | 6199 | spin_unlock(&delayed_refs->lock); |
| 6200 | return 0; | 6200 | return 0; |
| 6201 | } | 6201 | } |
| 6202 | 6202 | ||
| 6203 | void btrfs_free_tree_block(struct btrfs_trans_handle *trans, | 6203 | void btrfs_free_tree_block(struct btrfs_trans_handle *trans, |
| 6204 | struct btrfs_root *root, | 6204 | struct btrfs_root *root, |
| 6205 | struct extent_buffer *buf, | 6205 | struct extent_buffer *buf, |
| 6206 | u64 parent, int last_ref) | 6206 | u64 parent, int last_ref) |
| 6207 | { | 6207 | { |
| 6208 | struct btrfs_block_group_cache *cache = NULL; | 6208 | struct btrfs_block_group_cache *cache = NULL; |
| 6209 | int pin = 1; | 6209 | int pin = 1; |
| 6210 | int ret; | 6210 | int ret; |
| 6211 | 6211 | ||
| 6212 | if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { | 6212 | if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { |
| 6213 | ret = btrfs_add_delayed_tree_ref(root->fs_info, trans, | 6213 | ret = btrfs_add_delayed_tree_ref(root->fs_info, trans, |
| 6214 | buf->start, buf->len, | 6214 | buf->start, buf->len, |
| 6215 | parent, root->root_key.objectid, | 6215 | parent, root->root_key.objectid, |
| 6216 | btrfs_header_level(buf), | 6216 | btrfs_header_level(buf), |
| 6217 | BTRFS_DROP_DELAYED_REF, NULL, 0); | 6217 | BTRFS_DROP_DELAYED_REF, NULL, 0); |
| 6218 | BUG_ON(ret); /* -ENOMEM */ | 6218 | BUG_ON(ret); /* -ENOMEM */ |
| 6219 | } | 6219 | } |
| 6220 | 6220 | ||
| 6221 | if (!last_ref) | 6221 | if (!last_ref) |
| 6222 | return; | 6222 | return; |
| 6223 | 6223 | ||
| 6224 | cache = btrfs_lookup_block_group(root->fs_info, buf->start); | 6224 | cache = btrfs_lookup_block_group(root->fs_info, buf->start); |
| 6225 | 6225 | ||
| 6226 | if (btrfs_header_generation(buf) == trans->transid) { | 6226 | if (btrfs_header_generation(buf) == trans->transid) { |
| 6227 | if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { | 6227 | if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) { |
| 6228 | ret = check_ref_cleanup(trans, root, buf->start); | 6228 | ret = check_ref_cleanup(trans, root, buf->start); |
| 6229 | if (!ret) | 6229 | if (!ret) |
| 6230 | goto out; | 6230 | goto out; |
| 6231 | } | 6231 | } |
| 6232 | 6232 | ||
| 6233 | if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) { | 6233 | if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) { |
| 6234 | pin_down_extent(root, cache, buf->start, buf->len, 1); | 6234 | pin_down_extent(root, cache, buf->start, buf->len, 1); |
| 6235 | goto out; | 6235 | goto out; |
| 6236 | } | 6236 | } |
| 6237 | 6237 | ||
| 6238 | WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)); | 6238 | WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)); |
| 6239 | 6239 | ||
| 6240 | btrfs_add_free_space(cache, buf->start, buf->len); | 6240 | btrfs_add_free_space(cache, buf->start, buf->len); |
| 6241 | btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE, 0); | 6241 | btrfs_update_reserved_bytes(cache, buf->len, RESERVE_FREE, 0); |
| 6242 | trace_btrfs_reserved_extent_free(root, buf->start, buf->len); | 6242 | trace_btrfs_reserved_extent_free(root, buf->start, buf->len); |
| 6243 | pin = 0; | 6243 | pin = 0; |
| 6244 | } | 6244 | } |
| 6245 | out: | 6245 | out: |
| 6246 | if (pin) | 6246 | if (pin) |
| 6247 | add_pinned_bytes(root->fs_info, buf->len, | 6247 | add_pinned_bytes(root->fs_info, buf->len, |
| 6248 | btrfs_header_level(buf), | 6248 | btrfs_header_level(buf), |
| 6249 | root->root_key.objectid); | 6249 | root->root_key.objectid); |
| 6250 | 6250 | ||
| 6251 | /* | 6251 | /* |
| 6252 | * Deleting the buffer, clear the corrupt flag since it doesn't matter | 6252 | * Deleting the buffer, clear the corrupt flag since it doesn't matter |
| 6253 | * anymore. | 6253 | * anymore. |
| 6254 | */ | 6254 | */ |
| 6255 | clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags); | 6255 | clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags); |
| 6256 | btrfs_put_block_group(cache); | 6256 | btrfs_put_block_group(cache); |
| 6257 | } | 6257 | } |
| 6258 | 6258 | ||
| 6259 | /* Can return -ENOMEM */ | 6259 | /* Can return -ENOMEM */ |
| 6260 | int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, | 6260 | int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
| 6261 | u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, | 6261 | u64 bytenr, u64 num_bytes, u64 parent, u64 root_objectid, |
| 6262 | u64 owner, u64 offset, int no_quota) | 6262 | u64 owner, u64 offset, int no_quota) |
| 6263 | { | 6263 | { |
| 6264 | int ret; | 6264 | int ret; |
| 6265 | struct btrfs_fs_info *fs_info = root->fs_info; | 6265 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 6266 | 6266 | ||
| 6267 | if (btrfs_test_is_dummy_root(root)) | 6267 | if (btrfs_test_is_dummy_root(root)) |
| 6268 | return 0; | 6268 | return 0; |
| 6269 | 6269 | ||
| 6270 | add_pinned_bytes(root->fs_info, num_bytes, owner, root_objectid); | 6270 | add_pinned_bytes(root->fs_info, num_bytes, owner, root_objectid); |
| 6271 | 6271 | ||
| 6272 | /* | 6272 | /* |
| 6273 | * tree log blocks never actually go into the extent allocation | 6273 | * tree log blocks never actually go into the extent allocation |
| 6274 | * tree, just update pinning info and exit early. | 6274 | * tree, just update pinning info and exit early. |
| 6275 | */ | 6275 | */ |
| 6276 | if (root_objectid == BTRFS_TREE_LOG_OBJECTID) { | 6276 | if (root_objectid == BTRFS_TREE_LOG_OBJECTID) { |
| 6277 | WARN_ON(owner >= BTRFS_FIRST_FREE_OBJECTID); | 6277 | WARN_ON(owner >= BTRFS_FIRST_FREE_OBJECTID); |
| 6278 | /* unlocks the pinned mutex */ | 6278 | /* unlocks the pinned mutex */ |
| 6279 | btrfs_pin_extent(root, bytenr, num_bytes, 1); | 6279 | btrfs_pin_extent(root, bytenr, num_bytes, 1); |
| 6280 | ret = 0; | 6280 | ret = 0; |
| 6281 | } else if (owner < BTRFS_FIRST_FREE_OBJECTID) { | 6281 | } else if (owner < BTRFS_FIRST_FREE_OBJECTID) { |
| 6282 | ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, | 6282 | ret = btrfs_add_delayed_tree_ref(fs_info, trans, bytenr, |
| 6283 | num_bytes, | 6283 | num_bytes, |
| 6284 | parent, root_objectid, (int)owner, | 6284 | parent, root_objectid, (int)owner, |
| 6285 | BTRFS_DROP_DELAYED_REF, NULL, no_quota); | 6285 | BTRFS_DROP_DELAYED_REF, NULL, no_quota); |
| 6286 | } else { | 6286 | } else { |
| 6287 | ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, | 6287 | ret = btrfs_add_delayed_data_ref(fs_info, trans, bytenr, |
| 6288 | num_bytes, | 6288 | num_bytes, |
| 6289 | parent, root_objectid, owner, | 6289 | parent, root_objectid, owner, |
| 6290 | offset, BTRFS_DROP_DELAYED_REF, | 6290 | offset, BTRFS_DROP_DELAYED_REF, |
| 6291 | NULL, no_quota); | 6291 | NULL, no_quota); |
| 6292 | } | 6292 | } |
| 6293 | return ret; | 6293 | return ret; |
| 6294 | } | 6294 | } |
| 6295 | 6295 | ||
| 6296 | /* | 6296 | /* |
| 6297 | * when we wait for progress in the block group caching, its because | 6297 | * when we wait for progress in the block group caching, its because |
| 6298 | * our allocation attempt failed at least once. So, we must sleep | 6298 | * our allocation attempt failed at least once. So, we must sleep |
| 6299 | * and let some progress happen before we try again. | 6299 | * and let some progress happen before we try again. |
| 6300 | * | 6300 | * |
| 6301 | * This function will sleep at least once waiting for new free space to | 6301 | * This function will sleep at least once waiting for new free space to |
| 6302 | * show up, and then it will check the block group free space numbers | 6302 | * show up, and then it will check the block group free space numbers |
| 6303 | * for our min num_bytes. Another option is to have it go ahead | 6303 | * for our min num_bytes. Another option is to have it go ahead |
| 6304 | * and look in the rbtree for a free extent of a given size, but this | 6304 | * and look in the rbtree for a free extent of a given size, but this |
| 6305 | * is a good start. | 6305 | * is a good start. |
| 6306 | * | 6306 | * |
| 6307 | * Callers of this must check if cache->cached == BTRFS_CACHE_ERROR before using | 6307 | * Callers of this must check if cache->cached == BTRFS_CACHE_ERROR before using |
| 6308 | * any of the information in this block group. | 6308 | * any of the information in this block group. |
| 6309 | */ | 6309 | */ |
| 6310 | static noinline void | 6310 | static noinline void |
| 6311 | wait_block_group_cache_progress(struct btrfs_block_group_cache *cache, | 6311 | wait_block_group_cache_progress(struct btrfs_block_group_cache *cache, |
| 6312 | u64 num_bytes) | 6312 | u64 num_bytes) |
| 6313 | { | 6313 | { |
| 6314 | struct btrfs_caching_control *caching_ctl; | 6314 | struct btrfs_caching_control *caching_ctl; |
| 6315 | 6315 | ||
| 6316 | caching_ctl = get_caching_control(cache); | 6316 | caching_ctl = get_caching_control(cache); |
| 6317 | if (!caching_ctl) | 6317 | if (!caching_ctl) |
| 6318 | return; | 6318 | return; |
| 6319 | 6319 | ||
| 6320 | wait_event(caching_ctl->wait, block_group_cache_done(cache) || | 6320 | wait_event(caching_ctl->wait, block_group_cache_done(cache) || |
| 6321 | (cache->free_space_ctl->free_space >= num_bytes)); | 6321 | (cache->free_space_ctl->free_space >= num_bytes)); |
| 6322 | 6322 | ||
| 6323 | put_caching_control(caching_ctl); | 6323 | put_caching_control(caching_ctl); |
| 6324 | } | 6324 | } |
| 6325 | 6325 | ||
| 6326 | static noinline int | 6326 | static noinline int |
| 6327 | wait_block_group_cache_done(struct btrfs_block_group_cache *cache) | 6327 | wait_block_group_cache_done(struct btrfs_block_group_cache *cache) |
| 6328 | { | 6328 | { |
| 6329 | struct btrfs_caching_control *caching_ctl; | 6329 | struct btrfs_caching_control *caching_ctl; |
| 6330 | int ret = 0; | 6330 | int ret = 0; |
| 6331 | 6331 | ||
| 6332 | caching_ctl = get_caching_control(cache); | 6332 | caching_ctl = get_caching_control(cache); |
| 6333 | if (!caching_ctl) | 6333 | if (!caching_ctl) |
| 6334 | return (cache->cached == BTRFS_CACHE_ERROR) ? -EIO : 0; | 6334 | return (cache->cached == BTRFS_CACHE_ERROR) ? -EIO : 0; |
| 6335 | 6335 | ||
| 6336 | wait_event(caching_ctl->wait, block_group_cache_done(cache)); | 6336 | wait_event(caching_ctl->wait, block_group_cache_done(cache)); |
| 6337 | if (cache->cached == BTRFS_CACHE_ERROR) | 6337 | if (cache->cached == BTRFS_CACHE_ERROR) |
| 6338 | ret = -EIO; | 6338 | ret = -EIO; |
| 6339 | put_caching_control(caching_ctl); | 6339 | put_caching_control(caching_ctl); |
| 6340 | return ret; | 6340 | return ret; |
| 6341 | } | 6341 | } |
| 6342 | 6342 | ||
| 6343 | int __get_raid_index(u64 flags) | 6343 | int __get_raid_index(u64 flags) |
| 6344 | { | 6344 | { |
| 6345 | if (flags & BTRFS_BLOCK_GROUP_RAID10) | 6345 | if (flags & BTRFS_BLOCK_GROUP_RAID10) |
| 6346 | return BTRFS_RAID_RAID10; | 6346 | return BTRFS_RAID_RAID10; |
| 6347 | else if (flags & BTRFS_BLOCK_GROUP_RAID1) | 6347 | else if (flags & BTRFS_BLOCK_GROUP_RAID1) |
| 6348 | return BTRFS_RAID_RAID1; | 6348 | return BTRFS_RAID_RAID1; |
| 6349 | else if (flags & BTRFS_BLOCK_GROUP_DUP) | 6349 | else if (flags & BTRFS_BLOCK_GROUP_DUP) |
| 6350 | return BTRFS_RAID_DUP; | 6350 | return BTRFS_RAID_DUP; |
| 6351 | else if (flags & BTRFS_BLOCK_GROUP_RAID0) | 6351 | else if (flags & BTRFS_BLOCK_GROUP_RAID0) |
| 6352 | return BTRFS_RAID_RAID0; | 6352 | return BTRFS_RAID_RAID0; |
| 6353 | else if (flags & BTRFS_BLOCK_GROUP_RAID5) | 6353 | else if (flags & BTRFS_BLOCK_GROUP_RAID5) |
| 6354 | return BTRFS_RAID_RAID5; | 6354 | return BTRFS_RAID_RAID5; |
| 6355 | else if (flags & BTRFS_BLOCK_GROUP_RAID6) | 6355 | else if (flags & BTRFS_BLOCK_GROUP_RAID6) |
| 6356 | return BTRFS_RAID_RAID6; | 6356 | return BTRFS_RAID_RAID6; |
| 6357 | 6357 | ||
| 6358 | return BTRFS_RAID_SINGLE; /* BTRFS_BLOCK_GROUP_SINGLE */ | 6358 | return BTRFS_RAID_SINGLE; /* BTRFS_BLOCK_GROUP_SINGLE */ |
| 6359 | } | 6359 | } |
| 6360 | 6360 | ||
| 6361 | int get_block_group_index(struct btrfs_block_group_cache *cache) | 6361 | int get_block_group_index(struct btrfs_block_group_cache *cache) |
| 6362 | { | 6362 | { |
| 6363 | return __get_raid_index(cache->flags); | 6363 | return __get_raid_index(cache->flags); |
| 6364 | } | 6364 | } |
| 6365 | 6365 | ||
| 6366 | static const char *btrfs_raid_type_names[BTRFS_NR_RAID_TYPES] = { | 6366 | static const char *btrfs_raid_type_names[BTRFS_NR_RAID_TYPES] = { |
| 6367 | [BTRFS_RAID_RAID10] = "raid10", | 6367 | [BTRFS_RAID_RAID10] = "raid10", |
| 6368 | [BTRFS_RAID_RAID1] = "raid1", | 6368 | [BTRFS_RAID_RAID1] = "raid1", |
| 6369 | [BTRFS_RAID_DUP] = "dup", | 6369 | [BTRFS_RAID_DUP] = "dup", |
| 6370 | [BTRFS_RAID_RAID0] = "raid0", | 6370 | [BTRFS_RAID_RAID0] = "raid0", |
| 6371 | [BTRFS_RAID_SINGLE] = "single", | 6371 | [BTRFS_RAID_SINGLE] = "single", |
| 6372 | [BTRFS_RAID_RAID5] = "raid5", | 6372 | [BTRFS_RAID_RAID5] = "raid5", |
| 6373 | [BTRFS_RAID_RAID6] = "raid6", | 6373 | [BTRFS_RAID_RAID6] = "raid6", |
| 6374 | }; | 6374 | }; |
| 6375 | 6375 | ||
| 6376 | static const char *get_raid_name(enum btrfs_raid_types type) | 6376 | static const char *get_raid_name(enum btrfs_raid_types type) |
| 6377 | { | 6377 | { |
| 6378 | if (type >= BTRFS_NR_RAID_TYPES) | 6378 | if (type >= BTRFS_NR_RAID_TYPES) |
| 6379 | return NULL; | 6379 | return NULL; |
| 6380 | 6380 | ||
| 6381 | return btrfs_raid_type_names[type]; | 6381 | return btrfs_raid_type_names[type]; |
| 6382 | } | 6382 | } |
| 6383 | 6383 | ||
| 6384 | enum btrfs_loop_type { | 6384 | enum btrfs_loop_type { |
| 6385 | LOOP_CACHING_NOWAIT = 0, | 6385 | LOOP_CACHING_NOWAIT = 0, |
| 6386 | LOOP_CACHING_WAIT = 1, | 6386 | LOOP_CACHING_WAIT = 1, |
| 6387 | LOOP_ALLOC_CHUNK = 2, | 6387 | LOOP_ALLOC_CHUNK = 2, |
| 6388 | LOOP_NO_EMPTY_SIZE = 3, | 6388 | LOOP_NO_EMPTY_SIZE = 3, |
| 6389 | }; | 6389 | }; |
| 6390 | 6390 | ||
| 6391 | static inline void | 6391 | static inline void |
| 6392 | btrfs_lock_block_group(struct btrfs_block_group_cache *cache, | 6392 | btrfs_lock_block_group(struct btrfs_block_group_cache *cache, |
| 6393 | int delalloc) | 6393 | int delalloc) |
| 6394 | { | 6394 | { |
| 6395 | if (delalloc) | 6395 | if (delalloc) |
| 6396 | down_read(&cache->data_rwsem); | 6396 | down_read(&cache->data_rwsem); |
| 6397 | } | 6397 | } |
| 6398 | 6398 | ||
| 6399 | static inline void | 6399 | static inline void |
| 6400 | btrfs_grab_block_group(struct btrfs_block_group_cache *cache, | 6400 | btrfs_grab_block_group(struct btrfs_block_group_cache *cache, |
| 6401 | int delalloc) | 6401 | int delalloc) |
| 6402 | { | 6402 | { |
| 6403 | btrfs_get_block_group(cache); | 6403 | btrfs_get_block_group(cache); |
| 6404 | if (delalloc) | 6404 | if (delalloc) |
| 6405 | down_read(&cache->data_rwsem); | 6405 | down_read(&cache->data_rwsem); |
| 6406 | } | 6406 | } |
| 6407 | 6407 | ||
| 6408 | static struct btrfs_block_group_cache * | 6408 | static struct btrfs_block_group_cache * |
| 6409 | btrfs_lock_cluster(struct btrfs_block_group_cache *block_group, | 6409 | btrfs_lock_cluster(struct btrfs_block_group_cache *block_group, |
| 6410 | struct btrfs_free_cluster *cluster, | 6410 | struct btrfs_free_cluster *cluster, |
| 6411 | int delalloc) | 6411 | int delalloc) |
| 6412 | { | 6412 | { |
| 6413 | struct btrfs_block_group_cache *used_bg; | 6413 | struct btrfs_block_group_cache *used_bg; |
| 6414 | bool locked = false; | 6414 | bool locked = false; |
| 6415 | again: | 6415 | again: |
| 6416 | spin_lock(&cluster->refill_lock); | 6416 | spin_lock(&cluster->refill_lock); |
| 6417 | if (locked) { | 6417 | if (locked) { |
| 6418 | if (used_bg == cluster->block_group) | 6418 | if (used_bg == cluster->block_group) |
| 6419 | return used_bg; | 6419 | return used_bg; |
| 6420 | 6420 | ||
| 6421 | up_read(&used_bg->data_rwsem); | 6421 | up_read(&used_bg->data_rwsem); |
| 6422 | btrfs_put_block_group(used_bg); | 6422 | btrfs_put_block_group(used_bg); |
| 6423 | } | 6423 | } |
| 6424 | 6424 | ||
| 6425 | used_bg = cluster->block_group; | 6425 | used_bg = cluster->block_group; |
| 6426 | if (!used_bg) | 6426 | if (!used_bg) |
| 6427 | return NULL; | 6427 | return NULL; |
| 6428 | 6428 | ||
| 6429 | if (used_bg == block_group) | 6429 | if (used_bg == block_group) |
| 6430 | return used_bg; | 6430 | return used_bg; |
| 6431 | 6431 | ||
| 6432 | btrfs_get_block_group(used_bg); | 6432 | btrfs_get_block_group(used_bg); |
| 6433 | 6433 | ||
| 6434 | if (!delalloc) | 6434 | if (!delalloc) |
| 6435 | return used_bg; | 6435 | return used_bg; |
| 6436 | 6436 | ||
| 6437 | if (down_read_trylock(&used_bg->data_rwsem)) | 6437 | if (down_read_trylock(&used_bg->data_rwsem)) |
| 6438 | return used_bg; | 6438 | return used_bg; |
| 6439 | 6439 | ||
| 6440 | spin_unlock(&cluster->refill_lock); | 6440 | spin_unlock(&cluster->refill_lock); |
| 6441 | down_read(&used_bg->data_rwsem); | 6441 | down_read(&used_bg->data_rwsem); |
| 6442 | locked = true; | 6442 | locked = true; |
| 6443 | goto again; | 6443 | goto again; |
| 6444 | } | 6444 | } |
| 6445 | 6445 | ||
| 6446 | static inline void | 6446 | static inline void |
| 6447 | btrfs_release_block_group(struct btrfs_block_group_cache *cache, | 6447 | btrfs_release_block_group(struct btrfs_block_group_cache *cache, |
| 6448 | int delalloc) | 6448 | int delalloc) |
| 6449 | { | 6449 | { |
| 6450 | if (delalloc) | 6450 | if (delalloc) |
| 6451 | up_read(&cache->data_rwsem); | 6451 | up_read(&cache->data_rwsem); |
| 6452 | btrfs_put_block_group(cache); | 6452 | btrfs_put_block_group(cache); |
| 6453 | } | 6453 | } |
| 6454 | 6454 | ||
| 6455 | /* | 6455 | /* |
| 6456 | * walks the btree of allocated extents and find a hole of a given size. | 6456 | * walks the btree of allocated extents and find a hole of a given size. |
| 6457 | * The key ins is changed to record the hole: | 6457 | * The key ins is changed to record the hole: |
| 6458 | * ins->objectid == start position | 6458 | * ins->objectid == start position |
| 6459 | * ins->flags = BTRFS_EXTENT_ITEM_KEY | 6459 | * ins->flags = BTRFS_EXTENT_ITEM_KEY |
| 6460 | * ins->offset == the size of the hole. | 6460 | * ins->offset == the size of the hole. |
| 6461 | * Any available blocks before search_start are skipped. | 6461 | * Any available blocks before search_start are skipped. |
| 6462 | * | 6462 | * |
| 6463 | * If there is no suitable free space, we will record the max size of | 6463 | * If there is no suitable free space, we will record the max size of |
| 6464 | * the free space extent currently. | 6464 | * the free space extent currently. |
| 6465 | */ | 6465 | */ |
| 6466 | static noinline int find_free_extent(struct btrfs_root *orig_root, | 6466 | static noinline int find_free_extent(struct btrfs_root *orig_root, |
| 6467 | u64 num_bytes, u64 empty_size, | 6467 | u64 num_bytes, u64 empty_size, |
| 6468 | u64 hint_byte, struct btrfs_key *ins, | 6468 | u64 hint_byte, struct btrfs_key *ins, |
| 6469 | u64 flags, int delalloc) | 6469 | u64 flags, int delalloc) |
| 6470 | { | 6470 | { |
| 6471 | int ret = 0; | 6471 | int ret = 0; |
| 6472 | struct btrfs_root *root = orig_root->fs_info->extent_root; | 6472 | struct btrfs_root *root = orig_root->fs_info->extent_root; |
| 6473 | struct btrfs_free_cluster *last_ptr = NULL; | 6473 | struct btrfs_free_cluster *last_ptr = NULL; |
| 6474 | struct btrfs_block_group_cache *block_group = NULL; | 6474 | struct btrfs_block_group_cache *block_group = NULL; |
| 6475 | u64 search_start = 0; | 6475 | u64 search_start = 0; |
| 6476 | u64 max_extent_size = 0; | 6476 | u64 max_extent_size = 0; |
| 6477 | int empty_cluster = 2 * 1024 * 1024; | 6477 | int empty_cluster = 2 * 1024 * 1024; |
| 6478 | struct btrfs_space_info *space_info; | 6478 | struct btrfs_space_info *space_info; |
| 6479 | int loop = 0; | 6479 | int loop = 0; |
| 6480 | int index = __get_raid_index(flags); | 6480 | int index = __get_raid_index(flags); |
| 6481 | int alloc_type = (flags & BTRFS_BLOCK_GROUP_DATA) ? | 6481 | int alloc_type = (flags & BTRFS_BLOCK_GROUP_DATA) ? |
| 6482 | RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC; | 6482 | RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC; |
| 6483 | bool failed_cluster_refill = false; | 6483 | bool failed_cluster_refill = false; |
| 6484 | bool failed_alloc = false; | 6484 | bool failed_alloc = false; |
| 6485 | bool use_cluster = true; | 6485 | bool use_cluster = true; |
| 6486 | bool have_caching_bg = false; | 6486 | bool have_caching_bg = false; |
| 6487 | 6487 | ||
| 6488 | WARN_ON(num_bytes < root->sectorsize); | 6488 | WARN_ON(num_bytes < root->sectorsize); |
| 6489 | ins->type = BTRFS_EXTENT_ITEM_KEY; | 6489 | ins->type = BTRFS_EXTENT_ITEM_KEY; |
| 6490 | ins->objectid = 0; | 6490 | ins->objectid = 0; |
| 6491 | ins->offset = 0; | 6491 | ins->offset = 0; |
| 6492 | 6492 | ||
| 6493 | trace_find_free_extent(orig_root, num_bytes, empty_size, flags); | 6493 | trace_find_free_extent(orig_root, num_bytes, empty_size, flags); |
| 6494 | 6494 | ||
| 6495 | space_info = __find_space_info(root->fs_info, flags); | 6495 | space_info = __find_space_info(root->fs_info, flags); |
| 6496 | if (!space_info) { | 6496 | if (!space_info) { |
| 6497 | btrfs_err(root->fs_info, "No space info for %llu", flags); | 6497 | btrfs_err(root->fs_info, "No space info for %llu", flags); |
| 6498 | return -ENOSPC; | 6498 | return -ENOSPC; |
| 6499 | } | 6499 | } |
| 6500 | 6500 | ||
| 6501 | /* | 6501 | /* |
| 6502 | * If the space info is for both data and metadata it means we have a | 6502 | * If the space info is for both data and metadata it means we have a |
| 6503 | * small filesystem and we can't use the clustering stuff. | 6503 | * small filesystem and we can't use the clustering stuff. |
| 6504 | */ | 6504 | */ |
| 6505 | if (btrfs_mixed_space_info(space_info)) | 6505 | if (btrfs_mixed_space_info(space_info)) |
| 6506 | use_cluster = false; | 6506 | use_cluster = false; |
| 6507 | 6507 | ||
| 6508 | if (flags & BTRFS_BLOCK_GROUP_METADATA && use_cluster) { | 6508 | if (flags & BTRFS_BLOCK_GROUP_METADATA && use_cluster) { |
| 6509 | last_ptr = &root->fs_info->meta_alloc_cluster; | 6509 | last_ptr = &root->fs_info->meta_alloc_cluster; |
| 6510 | if (!btrfs_test_opt(root, SSD)) | 6510 | if (!btrfs_test_opt(root, SSD)) |
| 6511 | empty_cluster = 64 * 1024; | 6511 | empty_cluster = 64 * 1024; |
| 6512 | } | 6512 | } |
| 6513 | 6513 | ||
| 6514 | if ((flags & BTRFS_BLOCK_GROUP_DATA) && use_cluster && | 6514 | if ((flags & BTRFS_BLOCK_GROUP_DATA) && use_cluster && |
| 6515 | btrfs_test_opt(root, SSD)) { | 6515 | btrfs_test_opt(root, SSD)) { |
| 6516 | last_ptr = &root->fs_info->data_alloc_cluster; | 6516 | last_ptr = &root->fs_info->data_alloc_cluster; |
| 6517 | } | 6517 | } |
| 6518 | 6518 | ||
| 6519 | if (last_ptr) { | 6519 | if (last_ptr) { |
| 6520 | spin_lock(&last_ptr->lock); | 6520 | spin_lock(&last_ptr->lock); |
| 6521 | if (last_ptr->block_group) | 6521 | if (last_ptr->block_group) |
| 6522 | hint_byte = last_ptr->window_start; | 6522 | hint_byte = last_ptr->window_start; |
| 6523 | spin_unlock(&last_ptr->lock); | 6523 | spin_unlock(&last_ptr->lock); |
| 6524 | } | 6524 | } |
| 6525 | 6525 | ||
| 6526 | search_start = max(search_start, first_logical_byte(root, 0)); | 6526 | search_start = max(search_start, first_logical_byte(root, 0)); |
| 6527 | search_start = max(search_start, hint_byte); | 6527 | search_start = max(search_start, hint_byte); |
| 6528 | 6528 | ||
| 6529 | if (!last_ptr) | 6529 | if (!last_ptr) |
| 6530 | empty_cluster = 0; | 6530 | empty_cluster = 0; |
| 6531 | 6531 | ||
| 6532 | if (search_start == hint_byte) { | 6532 | if (search_start == hint_byte) { |
| 6533 | block_group = btrfs_lookup_block_group(root->fs_info, | 6533 | block_group = btrfs_lookup_block_group(root->fs_info, |
| 6534 | search_start); | 6534 | search_start); |
| 6535 | /* | 6535 | /* |
| 6536 | * we don't want to use the block group if it doesn't match our | 6536 | * we don't want to use the block group if it doesn't match our |
| 6537 | * allocation bits, or if its not cached. | 6537 | * allocation bits, or if its not cached. |
| 6538 | * | 6538 | * |
| 6539 | * However if we are re-searching with an ideal block group | 6539 | * However if we are re-searching with an ideal block group |
| 6540 | * picked out then we don't care that the block group is cached. | 6540 | * picked out then we don't care that the block group is cached. |
| 6541 | */ | 6541 | */ |
| 6542 | if (block_group && block_group_bits(block_group, flags) && | 6542 | if (block_group && block_group_bits(block_group, flags) && |
| 6543 | block_group->cached != BTRFS_CACHE_NO) { | 6543 | block_group->cached != BTRFS_CACHE_NO) { |
| 6544 | down_read(&space_info->groups_sem); | 6544 | down_read(&space_info->groups_sem); |
| 6545 | if (list_empty(&block_group->list) || | 6545 | if (list_empty(&block_group->list) || |
| 6546 | block_group->ro) { | 6546 | block_group->ro) { |
| 6547 | /* | 6547 | /* |
| 6548 | * someone is removing this block group, | 6548 | * someone is removing this block group, |
| 6549 | * we can't jump into the have_block_group | 6549 | * we can't jump into the have_block_group |
| 6550 | * target because our list pointers are not | 6550 | * target because our list pointers are not |
| 6551 | * valid | 6551 | * valid |
| 6552 | */ | 6552 | */ |
| 6553 | btrfs_put_block_group(block_group); | 6553 | btrfs_put_block_group(block_group); |
| 6554 | up_read(&space_info->groups_sem); | 6554 | up_read(&space_info->groups_sem); |
| 6555 | } else { | 6555 | } else { |
| 6556 | index = get_block_group_index(block_group); | 6556 | index = get_block_group_index(block_group); |
| 6557 | btrfs_lock_block_group(block_group, delalloc); | 6557 | btrfs_lock_block_group(block_group, delalloc); |
| 6558 | goto have_block_group; | 6558 | goto have_block_group; |
| 6559 | } | 6559 | } |
| 6560 | } else if (block_group) { | 6560 | } else if (block_group) { |
| 6561 | btrfs_put_block_group(block_group); | 6561 | btrfs_put_block_group(block_group); |
| 6562 | } | 6562 | } |
| 6563 | } | 6563 | } |
| 6564 | search: | 6564 | search: |
| 6565 | have_caching_bg = false; | 6565 | have_caching_bg = false; |
| 6566 | down_read(&space_info->groups_sem); | 6566 | down_read(&space_info->groups_sem); |
| 6567 | list_for_each_entry(block_group, &space_info->block_groups[index], | 6567 | list_for_each_entry(block_group, &space_info->block_groups[index], |
| 6568 | list) { | 6568 | list) { |
| 6569 | u64 offset; | 6569 | u64 offset; |
| 6570 | int cached; | 6570 | int cached; |
| 6571 | 6571 | ||
| 6572 | btrfs_grab_block_group(block_group, delalloc); | 6572 | btrfs_grab_block_group(block_group, delalloc); |
| 6573 | search_start = block_group->key.objectid; | 6573 | search_start = block_group->key.objectid; |
| 6574 | 6574 | ||
| 6575 | /* | 6575 | /* |
| 6576 | * this can happen if we end up cycling through all the | 6576 | * this can happen if we end up cycling through all the |
| 6577 | * raid types, but we want to make sure we only allocate | 6577 | * raid types, but we want to make sure we only allocate |
| 6578 | * for the proper type. | 6578 | * for the proper type. |
| 6579 | */ | 6579 | */ |
| 6580 | if (!block_group_bits(block_group, flags)) { | 6580 | if (!block_group_bits(block_group, flags)) { |
| 6581 | u64 extra = BTRFS_BLOCK_GROUP_DUP | | 6581 | u64 extra = BTRFS_BLOCK_GROUP_DUP | |
| 6582 | BTRFS_BLOCK_GROUP_RAID1 | | 6582 | BTRFS_BLOCK_GROUP_RAID1 | |
| 6583 | BTRFS_BLOCK_GROUP_RAID5 | | 6583 | BTRFS_BLOCK_GROUP_RAID5 | |
| 6584 | BTRFS_BLOCK_GROUP_RAID6 | | 6584 | BTRFS_BLOCK_GROUP_RAID6 | |
| 6585 | BTRFS_BLOCK_GROUP_RAID10; | 6585 | BTRFS_BLOCK_GROUP_RAID10; |
| 6586 | 6586 | ||
| 6587 | /* | 6587 | /* |
| 6588 | * if they asked for extra copies and this block group | 6588 | * if they asked for extra copies and this block group |
| 6589 | * doesn't provide them, bail. This does allow us to | 6589 | * doesn't provide them, bail. This does allow us to |
| 6590 | * fill raid0 from raid1. | 6590 | * fill raid0 from raid1. |
| 6591 | */ | 6591 | */ |
| 6592 | if ((flags & extra) && !(block_group->flags & extra)) | 6592 | if ((flags & extra) && !(block_group->flags & extra)) |
| 6593 | goto loop; | 6593 | goto loop; |
| 6594 | } | 6594 | } |
| 6595 | 6595 | ||
| 6596 | have_block_group: | 6596 | have_block_group: |
| 6597 | cached = block_group_cache_done(block_group); | 6597 | cached = block_group_cache_done(block_group); |
| 6598 | if (unlikely(!cached)) { | 6598 | if (unlikely(!cached)) { |
| 6599 | ret = cache_block_group(block_group, 0); | 6599 | ret = cache_block_group(block_group, 0); |
| 6600 | BUG_ON(ret < 0); | 6600 | BUG_ON(ret < 0); |
| 6601 | ret = 0; | 6601 | ret = 0; |
| 6602 | } | 6602 | } |
| 6603 | 6603 | ||
| 6604 | if (unlikely(block_group->cached == BTRFS_CACHE_ERROR)) | 6604 | if (unlikely(block_group->cached == BTRFS_CACHE_ERROR)) |
| 6605 | goto loop; | 6605 | goto loop; |
| 6606 | if (unlikely(block_group->ro)) | 6606 | if (unlikely(block_group->ro)) |
| 6607 | goto loop; | 6607 | goto loop; |
| 6608 | 6608 | ||
| 6609 | /* | 6609 | /* |
| 6610 | * Ok we want to try and use the cluster allocator, so | 6610 | * Ok we want to try and use the cluster allocator, so |
| 6611 | * lets look there | 6611 | * lets look there |
| 6612 | */ | 6612 | */ |
| 6613 | if (last_ptr) { | 6613 | if (last_ptr) { |
| 6614 | struct btrfs_block_group_cache *used_block_group; | 6614 | struct btrfs_block_group_cache *used_block_group; |
| 6615 | unsigned long aligned_cluster; | 6615 | unsigned long aligned_cluster; |
| 6616 | /* | 6616 | /* |
| 6617 | * the refill lock keeps out other | 6617 | * the refill lock keeps out other |
| 6618 | * people trying to start a new cluster | 6618 | * people trying to start a new cluster |
| 6619 | */ | 6619 | */ |
| 6620 | used_block_group = btrfs_lock_cluster(block_group, | 6620 | used_block_group = btrfs_lock_cluster(block_group, |
| 6621 | last_ptr, | 6621 | last_ptr, |
| 6622 | delalloc); | 6622 | delalloc); |
| 6623 | if (!used_block_group) | 6623 | if (!used_block_group) |
| 6624 | goto refill_cluster; | 6624 | goto refill_cluster; |
| 6625 | 6625 | ||
| 6626 | if (used_block_group != block_group && | 6626 | if (used_block_group != block_group && |
| 6627 | (used_block_group->ro || | 6627 | (used_block_group->ro || |
| 6628 | !block_group_bits(used_block_group, flags))) | 6628 | !block_group_bits(used_block_group, flags))) |
| 6629 | goto release_cluster; | 6629 | goto release_cluster; |
| 6630 | 6630 | ||
| 6631 | offset = btrfs_alloc_from_cluster(used_block_group, | 6631 | offset = btrfs_alloc_from_cluster(used_block_group, |
| 6632 | last_ptr, | 6632 | last_ptr, |
| 6633 | num_bytes, | 6633 | num_bytes, |
| 6634 | used_block_group->key.objectid, | 6634 | used_block_group->key.objectid, |
| 6635 | &max_extent_size); | 6635 | &max_extent_size); |
| 6636 | if (offset) { | 6636 | if (offset) { |
| 6637 | /* we have a block, we're done */ | 6637 | /* we have a block, we're done */ |
| 6638 | spin_unlock(&last_ptr->refill_lock); | 6638 | spin_unlock(&last_ptr->refill_lock); |
| 6639 | trace_btrfs_reserve_extent_cluster(root, | 6639 | trace_btrfs_reserve_extent_cluster(root, |
| 6640 | used_block_group, | 6640 | used_block_group, |
| 6641 | search_start, num_bytes); | 6641 | search_start, num_bytes); |
| 6642 | if (used_block_group != block_group) { | 6642 | if (used_block_group != block_group) { |
| 6643 | btrfs_release_block_group(block_group, | 6643 | btrfs_release_block_group(block_group, |
| 6644 | delalloc); | 6644 | delalloc); |
| 6645 | block_group = used_block_group; | 6645 | block_group = used_block_group; |
| 6646 | } | 6646 | } |
| 6647 | goto checks; | 6647 | goto checks; |
| 6648 | } | 6648 | } |
| 6649 | 6649 | ||
| 6650 | WARN_ON(last_ptr->block_group != used_block_group); | 6650 | WARN_ON(last_ptr->block_group != used_block_group); |
| 6651 | release_cluster: | 6651 | release_cluster: |
| 6652 | /* If we are on LOOP_NO_EMPTY_SIZE, we can't | 6652 | /* If we are on LOOP_NO_EMPTY_SIZE, we can't |
| 6653 | * set up a new clusters, so lets just skip it | 6653 | * set up a new clusters, so lets just skip it |
| 6654 | * and let the allocator find whatever block | 6654 | * and let the allocator find whatever block |
| 6655 | * it can find. If we reach this point, we | 6655 | * it can find. If we reach this point, we |
| 6656 | * will have tried the cluster allocator | 6656 | * will have tried the cluster allocator |
| 6657 | * plenty of times and not have found | 6657 | * plenty of times and not have found |
| 6658 | * anything, so we are likely way too | 6658 | * anything, so we are likely way too |
| 6659 | * fragmented for the clustering stuff to find | 6659 | * fragmented for the clustering stuff to find |
| 6660 | * anything. | 6660 | * anything. |
| 6661 | * | 6661 | * |
| 6662 | * However, if the cluster is taken from the | 6662 | * However, if the cluster is taken from the |
| 6663 | * current block group, release the cluster | 6663 | * current block group, release the cluster |
| 6664 | * first, so that we stand a better chance of | 6664 | * first, so that we stand a better chance of |
| 6665 | * succeeding in the unclustered | 6665 | * succeeding in the unclustered |
| 6666 | * allocation. */ | 6666 | * allocation. */ |
| 6667 | if (loop >= LOOP_NO_EMPTY_SIZE && | 6667 | if (loop >= LOOP_NO_EMPTY_SIZE && |
| 6668 | used_block_group != block_group) { | 6668 | used_block_group != block_group) { |
| 6669 | spin_unlock(&last_ptr->refill_lock); | 6669 | spin_unlock(&last_ptr->refill_lock); |
| 6670 | btrfs_release_block_group(used_block_group, | 6670 | btrfs_release_block_group(used_block_group, |
| 6671 | delalloc); | 6671 | delalloc); |
| 6672 | goto unclustered_alloc; | 6672 | goto unclustered_alloc; |
| 6673 | } | 6673 | } |
| 6674 | 6674 | ||
| 6675 | /* | 6675 | /* |
| 6676 | * this cluster didn't work out, free it and | 6676 | * this cluster didn't work out, free it and |
| 6677 | * start over | 6677 | * start over |
| 6678 | */ | 6678 | */ |
| 6679 | btrfs_return_cluster_to_free_space(NULL, last_ptr); | 6679 | btrfs_return_cluster_to_free_space(NULL, last_ptr); |
| 6680 | 6680 | ||
| 6681 | if (used_block_group != block_group) | 6681 | if (used_block_group != block_group) |
| 6682 | btrfs_release_block_group(used_block_group, | 6682 | btrfs_release_block_group(used_block_group, |
| 6683 | delalloc); | 6683 | delalloc); |
| 6684 | refill_cluster: | 6684 | refill_cluster: |
| 6685 | if (loop >= LOOP_NO_EMPTY_SIZE) { | 6685 | if (loop >= LOOP_NO_EMPTY_SIZE) { |
| 6686 | spin_unlock(&last_ptr->refill_lock); | 6686 | spin_unlock(&last_ptr->refill_lock); |
| 6687 | goto unclustered_alloc; | 6687 | goto unclustered_alloc; |
| 6688 | } | 6688 | } |
| 6689 | 6689 | ||
| 6690 | aligned_cluster = max_t(unsigned long, | 6690 | aligned_cluster = max_t(unsigned long, |
| 6691 | empty_cluster + empty_size, | 6691 | empty_cluster + empty_size, |
| 6692 | block_group->full_stripe_len); | 6692 | block_group->full_stripe_len); |
| 6693 | 6693 | ||
| 6694 | /* allocate a cluster in this block group */ | 6694 | /* allocate a cluster in this block group */ |
| 6695 | ret = btrfs_find_space_cluster(root, block_group, | 6695 | ret = btrfs_find_space_cluster(root, block_group, |
| 6696 | last_ptr, search_start, | 6696 | last_ptr, search_start, |
| 6697 | num_bytes, | 6697 | num_bytes, |
| 6698 | aligned_cluster); | 6698 | aligned_cluster); |
| 6699 | if (ret == 0) { | 6699 | if (ret == 0) { |
| 6700 | /* | 6700 | /* |
| 6701 | * now pull our allocation out of this | 6701 | * now pull our allocation out of this |
| 6702 | * cluster | 6702 | * cluster |
| 6703 | */ | 6703 | */ |
| 6704 | offset = btrfs_alloc_from_cluster(block_group, | 6704 | offset = btrfs_alloc_from_cluster(block_group, |
| 6705 | last_ptr, | 6705 | last_ptr, |
| 6706 | num_bytes, | 6706 | num_bytes, |
| 6707 | search_start, | 6707 | search_start, |
| 6708 | &max_extent_size); | 6708 | &max_extent_size); |
| 6709 | if (offset) { | 6709 | if (offset) { |
| 6710 | /* we found one, proceed */ | 6710 | /* we found one, proceed */ |
| 6711 | spin_unlock(&last_ptr->refill_lock); | 6711 | spin_unlock(&last_ptr->refill_lock); |
| 6712 | trace_btrfs_reserve_extent_cluster(root, | 6712 | trace_btrfs_reserve_extent_cluster(root, |
| 6713 | block_group, search_start, | 6713 | block_group, search_start, |
| 6714 | num_bytes); | 6714 | num_bytes); |
| 6715 | goto checks; | 6715 | goto checks; |
| 6716 | } | 6716 | } |
| 6717 | } else if (!cached && loop > LOOP_CACHING_NOWAIT | 6717 | } else if (!cached && loop > LOOP_CACHING_NOWAIT |
| 6718 | && !failed_cluster_refill) { | 6718 | && !failed_cluster_refill) { |
| 6719 | spin_unlock(&last_ptr->refill_lock); | 6719 | spin_unlock(&last_ptr->refill_lock); |
| 6720 | 6720 | ||
| 6721 | failed_cluster_refill = true; | 6721 | failed_cluster_refill = true; |
| 6722 | wait_block_group_cache_progress(block_group, | 6722 | wait_block_group_cache_progress(block_group, |
| 6723 | num_bytes + empty_cluster + empty_size); | 6723 | num_bytes + empty_cluster + empty_size); |
| 6724 | goto have_block_group; | 6724 | goto have_block_group; |
| 6725 | } | 6725 | } |
| 6726 | 6726 | ||
| 6727 | /* | 6727 | /* |
| 6728 | * at this point we either didn't find a cluster | 6728 | * at this point we either didn't find a cluster |
| 6729 | * or we weren't able to allocate a block from our | 6729 | * or we weren't able to allocate a block from our |
| 6730 | * cluster. Free the cluster we've been trying | 6730 | * cluster. Free the cluster we've been trying |
| 6731 | * to use, and go to the next block group | 6731 | * to use, and go to the next block group |
| 6732 | */ | 6732 | */ |
| 6733 | btrfs_return_cluster_to_free_space(NULL, last_ptr); | 6733 | btrfs_return_cluster_to_free_space(NULL, last_ptr); |
| 6734 | spin_unlock(&last_ptr->refill_lock); | 6734 | spin_unlock(&last_ptr->refill_lock); |
| 6735 | goto loop; | 6735 | goto loop; |
| 6736 | } | 6736 | } |
| 6737 | 6737 | ||
| 6738 | unclustered_alloc: | 6738 | unclustered_alloc: |
| 6739 | spin_lock(&block_group->free_space_ctl->tree_lock); | 6739 | spin_lock(&block_group->free_space_ctl->tree_lock); |
| 6740 | if (cached && | 6740 | if (cached && |
| 6741 | block_group->free_space_ctl->free_space < | 6741 | block_group->free_space_ctl->free_space < |
| 6742 | num_bytes + empty_cluster + empty_size) { | 6742 | num_bytes + empty_cluster + empty_size) { |
| 6743 | if (block_group->free_space_ctl->free_space > | 6743 | if (block_group->free_space_ctl->free_space > |
| 6744 | max_extent_size) | 6744 | max_extent_size) |
| 6745 | max_extent_size = | 6745 | max_extent_size = |
| 6746 | block_group->free_space_ctl->free_space; | 6746 | block_group->free_space_ctl->free_space; |
| 6747 | spin_unlock(&block_group->free_space_ctl->tree_lock); | 6747 | spin_unlock(&block_group->free_space_ctl->tree_lock); |
| 6748 | goto loop; | 6748 | goto loop; |
| 6749 | } | 6749 | } |
| 6750 | spin_unlock(&block_group->free_space_ctl->tree_lock); | 6750 | spin_unlock(&block_group->free_space_ctl->tree_lock); |
| 6751 | 6751 | ||
| 6752 | offset = btrfs_find_space_for_alloc(block_group, search_start, | 6752 | offset = btrfs_find_space_for_alloc(block_group, search_start, |
| 6753 | num_bytes, empty_size, | 6753 | num_bytes, empty_size, |
| 6754 | &max_extent_size); | 6754 | &max_extent_size); |
| 6755 | /* | 6755 | /* |
| 6756 | * If we didn't find a chunk, and we haven't failed on this | 6756 | * If we didn't find a chunk, and we haven't failed on this |
| 6757 | * block group before, and this block group is in the middle of | 6757 | * block group before, and this block group is in the middle of |
| 6758 | * caching and we are ok with waiting, then go ahead and wait | 6758 | * caching and we are ok with waiting, then go ahead and wait |
| 6759 | * for progress to be made, and set failed_alloc to true. | 6759 | * for progress to be made, and set failed_alloc to true. |
| 6760 | * | 6760 | * |
| 6761 | * If failed_alloc is true then we've already waited on this | 6761 | * If failed_alloc is true then we've already waited on this |
| 6762 | * block group once and should move on to the next block group. | 6762 | * block group once and should move on to the next block group. |
| 6763 | */ | 6763 | */ |
| 6764 | if (!offset && !failed_alloc && !cached && | 6764 | if (!offset && !failed_alloc && !cached && |
| 6765 | loop > LOOP_CACHING_NOWAIT) { | 6765 | loop > LOOP_CACHING_NOWAIT) { |
| 6766 | wait_block_group_cache_progress(block_group, | 6766 | wait_block_group_cache_progress(block_group, |
| 6767 | num_bytes + empty_size); | 6767 | num_bytes + empty_size); |
| 6768 | failed_alloc = true; | 6768 | failed_alloc = true; |
| 6769 | goto have_block_group; | 6769 | goto have_block_group; |
| 6770 | } else if (!offset) { | 6770 | } else if (!offset) { |
| 6771 | if (!cached) | 6771 | if (!cached) |
| 6772 | have_caching_bg = true; | 6772 | have_caching_bg = true; |
| 6773 | goto loop; | 6773 | goto loop; |
| 6774 | } | 6774 | } |
| 6775 | checks: | 6775 | checks: |
| 6776 | search_start = ALIGN(offset, root->stripesize); | 6776 | search_start = ALIGN(offset, root->stripesize); |
| 6777 | 6777 | ||
| 6778 | /* move on to the next group */ | 6778 | /* move on to the next group */ |
| 6779 | if (search_start + num_bytes > | 6779 | if (search_start + num_bytes > |
| 6780 | block_group->key.objectid + block_group->key.offset) { | 6780 | block_group->key.objectid + block_group->key.offset) { |
| 6781 | btrfs_add_free_space(block_group, offset, num_bytes); | 6781 | btrfs_add_free_space(block_group, offset, num_bytes); |
| 6782 | goto loop; | 6782 | goto loop; |
| 6783 | } | 6783 | } |
| 6784 | 6784 | ||
| 6785 | if (offset < search_start) | 6785 | if (offset < search_start) |
| 6786 | btrfs_add_free_space(block_group, offset, | 6786 | btrfs_add_free_space(block_group, offset, |
| 6787 | search_start - offset); | 6787 | search_start - offset); |
| 6788 | BUG_ON(offset > search_start); | 6788 | BUG_ON(offset > search_start); |
| 6789 | 6789 | ||
| 6790 | ret = btrfs_update_reserved_bytes(block_group, num_bytes, | 6790 | ret = btrfs_update_reserved_bytes(block_group, num_bytes, |
| 6791 | alloc_type, delalloc); | 6791 | alloc_type, delalloc); |
| 6792 | if (ret == -EAGAIN) { | 6792 | if (ret == -EAGAIN) { |
| 6793 | btrfs_add_free_space(block_group, offset, num_bytes); | 6793 | btrfs_add_free_space(block_group, offset, num_bytes); |
| 6794 | goto loop; | 6794 | goto loop; |
| 6795 | } | 6795 | } |
| 6796 | 6796 | ||
| 6797 | /* we are all good, lets return */ | 6797 | /* we are all good, lets return */ |
| 6798 | ins->objectid = search_start; | 6798 | ins->objectid = search_start; |
| 6799 | ins->offset = num_bytes; | 6799 | ins->offset = num_bytes; |
| 6800 | 6800 | ||
| 6801 | trace_btrfs_reserve_extent(orig_root, block_group, | 6801 | trace_btrfs_reserve_extent(orig_root, block_group, |
| 6802 | search_start, num_bytes); | 6802 | search_start, num_bytes); |
| 6803 | btrfs_release_block_group(block_group, delalloc); | 6803 | btrfs_release_block_group(block_group, delalloc); |
| 6804 | break; | 6804 | break; |
| 6805 | loop: | 6805 | loop: |
| 6806 | failed_cluster_refill = false; | 6806 | failed_cluster_refill = false; |
| 6807 | failed_alloc = false; | 6807 | failed_alloc = false; |
| 6808 | BUG_ON(index != get_block_group_index(block_group)); | 6808 | BUG_ON(index != get_block_group_index(block_group)); |
| 6809 | btrfs_release_block_group(block_group, delalloc); | 6809 | btrfs_release_block_group(block_group, delalloc); |
| 6810 | } | 6810 | } |
| 6811 | up_read(&space_info->groups_sem); | 6811 | up_read(&space_info->groups_sem); |
| 6812 | 6812 | ||
| 6813 | if (!ins->objectid && loop >= LOOP_CACHING_WAIT && have_caching_bg) | 6813 | if (!ins->objectid && loop >= LOOP_CACHING_WAIT && have_caching_bg) |
| 6814 | goto search; | 6814 | goto search; |
| 6815 | 6815 | ||
| 6816 | if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES) | 6816 | if (!ins->objectid && ++index < BTRFS_NR_RAID_TYPES) |
| 6817 | goto search; | 6817 | goto search; |
| 6818 | 6818 | ||
| 6819 | /* | 6819 | /* |
| 6820 | * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking | 6820 | * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking |
| 6821 | * caching kthreads as we move along | 6821 | * caching kthreads as we move along |
| 6822 | * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching | 6822 | * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching |
| 6823 | * LOOP_ALLOC_CHUNK, force a chunk allocation and try again | 6823 | * LOOP_ALLOC_CHUNK, force a chunk allocation and try again |
| 6824 | * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try | 6824 | * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try |
| 6825 | * again | 6825 | * again |
| 6826 | */ | 6826 | */ |
| 6827 | if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) { | 6827 | if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) { |
| 6828 | index = 0; | 6828 | index = 0; |
| 6829 | loop++; | 6829 | loop++; |
| 6830 | if (loop == LOOP_ALLOC_CHUNK) { | 6830 | if (loop == LOOP_ALLOC_CHUNK) { |
| 6831 | struct btrfs_trans_handle *trans; | 6831 | struct btrfs_trans_handle *trans; |
| 6832 | int exist = 0; | 6832 | int exist = 0; |
| 6833 | 6833 | ||
| 6834 | trans = current->journal_info; | 6834 | trans = current->journal_info; |
| 6835 | if (trans) | 6835 | if (trans) |
| 6836 | exist = 1; | 6836 | exist = 1; |
| 6837 | else | 6837 | else |
| 6838 | trans = btrfs_join_transaction(root); | 6838 | trans = btrfs_join_transaction(root); |
| 6839 | 6839 | ||
| 6840 | if (IS_ERR(trans)) { | 6840 | if (IS_ERR(trans)) { |
| 6841 | ret = PTR_ERR(trans); | 6841 | ret = PTR_ERR(trans); |
| 6842 | goto out; | 6842 | goto out; |
| 6843 | } | 6843 | } |
| 6844 | 6844 | ||
| 6845 | ret = do_chunk_alloc(trans, root, flags, | 6845 | ret = do_chunk_alloc(trans, root, flags, |
| 6846 | CHUNK_ALLOC_FORCE); | 6846 | CHUNK_ALLOC_FORCE); |
| 6847 | /* | 6847 | /* |
| 6848 | * Do not bail out on ENOSPC since we | 6848 | * Do not bail out on ENOSPC since we |
| 6849 | * can do more things. | 6849 | * can do more things. |
| 6850 | */ | 6850 | */ |
| 6851 | if (ret < 0 && ret != -ENOSPC) | 6851 | if (ret < 0 && ret != -ENOSPC) |
| 6852 | btrfs_abort_transaction(trans, | 6852 | btrfs_abort_transaction(trans, |
| 6853 | root, ret); | 6853 | root, ret); |
| 6854 | else | 6854 | else |
| 6855 | ret = 0; | 6855 | ret = 0; |
| 6856 | if (!exist) | 6856 | if (!exist) |
| 6857 | btrfs_end_transaction(trans, root); | 6857 | btrfs_end_transaction(trans, root); |
| 6858 | if (ret) | 6858 | if (ret) |
| 6859 | goto out; | 6859 | goto out; |
| 6860 | } | 6860 | } |
| 6861 | 6861 | ||
| 6862 | if (loop == LOOP_NO_EMPTY_SIZE) { | 6862 | if (loop == LOOP_NO_EMPTY_SIZE) { |
| 6863 | empty_size = 0; | 6863 | empty_size = 0; |
| 6864 | empty_cluster = 0; | 6864 | empty_cluster = 0; |
| 6865 | } | 6865 | } |
| 6866 | 6866 | ||
| 6867 | goto search; | 6867 | goto search; |
| 6868 | } else if (!ins->objectid) { | 6868 | } else if (!ins->objectid) { |
| 6869 | ret = -ENOSPC; | 6869 | ret = -ENOSPC; |
| 6870 | } else if (ins->objectid) { | 6870 | } else if (ins->objectid) { |
| 6871 | ret = 0; | 6871 | ret = 0; |
| 6872 | } | 6872 | } |
| 6873 | out: | 6873 | out: |
| 6874 | if (ret == -ENOSPC) | 6874 | if (ret == -ENOSPC) |
| 6875 | ins->offset = max_extent_size; | 6875 | ins->offset = max_extent_size; |
| 6876 | return ret; | 6876 | return ret; |
| 6877 | } | 6877 | } |
| 6878 | 6878 | ||
| 6879 | static void dump_space_info(struct btrfs_space_info *info, u64 bytes, | 6879 | static void dump_space_info(struct btrfs_space_info *info, u64 bytes, |
| 6880 | int dump_block_groups) | 6880 | int dump_block_groups) |
| 6881 | { | 6881 | { |
| 6882 | struct btrfs_block_group_cache *cache; | 6882 | struct btrfs_block_group_cache *cache; |
| 6883 | int index = 0; | 6883 | int index = 0; |
| 6884 | 6884 | ||
| 6885 | spin_lock(&info->lock); | 6885 | spin_lock(&info->lock); |
| 6886 | printk(KERN_INFO "BTRFS: space_info %llu has %llu free, is %sfull\n", | 6886 | printk(KERN_INFO "BTRFS: space_info %llu has %llu free, is %sfull\n", |
| 6887 | info->flags, | 6887 | info->flags, |
| 6888 | info->total_bytes - info->bytes_used - info->bytes_pinned - | 6888 | info->total_bytes - info->bytes_used - info->bytes_pinned - |
| 6889 | info->bytes_reserved - info->bytes_readonly, | 6889 | info->bytes_reserved - info->bytes_readonly, |
| 6890 | (info->full) ? "" : "not "); | 6890 | (info->full) ? "" : "not "); |
| 6891 | printk(KERN_INFO "BTRFS: space_info total=%llu, used=%llu, pinned=%llu, " | 6891 | printk(KERN_INFO "BTRFS: space_info total=%llu, used=%llu, pinned=%llu, " |
| 6892 | "reserved=%llu, may_use=%llu, readonly=%llu\n", | 6892 | "reserved=%llu, may_use=%llu, readonly=%llu\n", |
| 6893 | info->total_bytes, info->bytes_used, info->bytes_pinned, | 6893 | info->total_bytes, info->bytes_used, info->bytes_pinned, |
| 6894 | info->bytes_reserved, info->bytes_may_use, | 6894 | info->bytes_reserved, info->bytes_may_use, |
| 6895 | info->bytes_readonly); | 6895 | info->bytes_readonly); |
| 6896 | spin_unlock(&info->lock); | 6896 | spin_unlock(&info->lock); |
| 6897 | 6897 | ||
| 6898 | if (!dump_block_groups) | 6898 | if (!dump_block_groups) |
| 6899 | return; | 6899 | return; |
| 6900 | 6900 | ||
| 6901 | down_read(&info->groups_sem); | 6901 | down_read(&info->groups_sem); |
| 6902 | again: | 6902 | again: |
| 6903 | list_for_each_entry(cache, &info->block_groups[index], list) { | 6903 | list_for_each_entry(cache, &info->block_groups[index], list) { |
| 6904 | spin_lock(&cache->lock); | 6904 | spin_lock(&cache->lock); |
| 6905 | printk(KERN_INFO "BTRFS: " | 6905 | printk(KERN_INFO "BTRFS: " |
| 6906 | "block group %llu has %llu bytes, " | 6906 | "block group %llu has %llu bytes, " |
| 6907 | "%llu used %llu pinned %llu reserved %s\n", | 6907 | "%llu used %llu pinned %llu reserved %s\n", |
| 6908 | cache->key.objectid, cache->key.offset, | 6908 | cache->key.objectid, cache->key.offset, |
| 6909 | btrfs_block_group_used(&cache->item), cache->pinned, | 6909 | btrfs_block_group_used(&cache->item), cache->pinned, |
| 6910 | cache->reserved, cache->ro ? "[readonly]" : ""); | 6910 | cache->reserved, cache->ro ? "[readonly]" : ""); |
| 6911 | btrfs_dump_free_space(cache, bytes); | 6911 | btrfs_dump_free_space(cache, bytes); |
| 6912 | spin_unlock(&cache->lock); | 6912 | spin_unlock(&cache->lock); |
| 6913 | } | 6913 | } |
| 6914 | if (++index < BTRFS_NR_RAID_TYPES) | 6914 | if (++index < BTRFS_NR_RAID_TYPES) |
| 6915 | goto again; | 6915 | goto again; |
| 6916 | up_read(&info->groups_sem); | 6916 | up_read(&info->groups_sem); |
| 6917 | } | 6917 | } |
| 6918 | 6918 | ||
| 6919 | int btrfs_reserve_extent(struct btrfs_root *root, | 6919 | int btrfs_reserve_extent(struct btrfs_root *root, |
| 6920 | u64 num_bytes, u64 min_alloc_size, | 6920 | u64 num_bytes, u64 min_alloc_size, |
| 6921 | u64 empty_size, u64 hint_byte, | 6921 | u64 empty_size, u64 hint_byte, |
| 6922 | struct btrfs_key *ins, int is_data, int delalloc) | 6922 | struct btrfs_key *ins, int is_data, int delalloc) |
| 6923 | { | 6923 | { |
| 6924 | bool final_tried = false; | 6924 | bool final_tried = false; |
| 6925 | u64 flags; | 6925 | u64 flags; |
| 6926 | int ret; | 6926 | int ret; |
| 6927 | 6927 | ||
| 6928 | flags = btrfs_get_alloc_profile(root, is_data); | 6928 | flags = btrfs_get_alloc_profile(root, is_data); |
| 6929 | again: | 6929 | again: |
| 6930 | WARN_ON(num_bytes < root->sectorsize); | 6930 | WARN_ON(num_bytes < root->sectorsize); |
| 6931 | ret = find_free_extent(root, num_bytes, empty_size, hint_byte, ins, | 6931 | ret = find_free_extent(root, num_bytes, empty_size, hint_byte, ins, |
| 6932 | flags, delalloc); | 6932 | flags, delalloc); |
| 6933 | 6933 | ||
| 6934 | if (ret == -ENOSPC) { | 6934 | if (ret == -ENOSPC) { |
| 6935 | if (!final_tried && ins->offset) { | 6935 | if (!final_tried && ins->offset) { |
| 6936 | num_bytes = min(num_bytes >> 1, ins->offset); | 6936 | num_bytes = min(num_bytes >> 1, ins->offset); |
| 6937 | num_bytes = round_down(num_bytes, root->sectorsize); | 6937 | num_bytes = round_down(num_bytes, root->sectorsize); |
| 6938 | num_bytes = max(num_bytes, min_alloc_size); | 6938 | num_bytes = max(num_bytes, min_alloc_size); |
| 6939 | if (num_bytes == min_alloc_size) | 6939 | if (num_bytes == min_alloc_size) |
| 6940 | final_tried = true; | 6940 | final_tried = true; |
| 6941 | goto again; | 6941 | goto again; |
| 6942 | } else if (btrfs_test_opt(root, ENOSPC_DEBUG)) { | 6942 | } else if (btrfs_test_opt(root, ENOSPC_DEBUG)) { |
| 6943 | struct btrfs_space_info *sinfo; | 6943 | struct btrfs_space_info *sinfo; |
| 6944 | 6944 | ||
| 6945 | sinfo = __find_space_info(root->fs_info, flags); | 6945 | sinfo = __find_space_info(root->fs_info, flags); |
| 6946 | btrfs_err(root->fs_info, "allocation failed flags %llu, wanted %llu", | 6946 | btrfs_err(root->fs_info, "allocation failed flags %llu, wanted %llu", |
| 6947 | flags, num_bytes); | 6947 | flags, num_bytes); |
| 6948 | if (sinfo) | 6948 | if (sinfo) |
| 6949 | dump_space_info(sinfo, num_bytes, 1); | 6949 | dump_space_info(sinfo, num_bytes, 1); |
| 6950 | } | 6950 | } |
| 6951 | } | 6951 | } |
| 6952 | 6952 | ||
| 6953 | return ret; | 6953 | return ret; |
| 6954 | } | 6954 | } |
| 6955 | 6955 | ||
| 6956 | static int __btrfs_free_reserved_extent(struct btrfs_root *root, | 6956 | static int __btrfs_free_reserved_extent(struct btrfs_root *root, |
| 6957 | u64 start, u64 len, | 6957 | u64 start, u64 len, |
| 6958 | int pin, int delalloc) | 6958 | int pin, int delalloc) |
| 6959 | { | 6959 | { |
| 6960 | struct btrfs_block_group_cache *cache; | 6960 | struct btrfs_block_group_cache *cache; |
| 6961 | int ret = 0; | 6961 | int ret = 0; |
| 6962 | 6962 | ||
| 6963 | cache = btrfs_lookup_block_group(root->fs_info, start); | 6963 | cache = btrfs_lookup_block_group(root->fs_info, start); |
| 6964 | if (!cache) { | 6964 | if (!cache) { |
| 6965 | btrfs_err(root->fs_info, "Unable to find block group for %llu", | 6965 | btrfs_err(root->fs_info, "Unable to find block group for %llu", |
| 6966 | start); | 6966 | start); |
| 6967 | return -ENOSPC; | 6967 | return -ENOSPC; |
| 6968 | } | 6968 | } |
| 6969 | 6969 | ||
| 6970 | if (btrfs_test_opt(root, DISCARD)) | 6970 | if (btrfs_test_opt(root, DISCARD)) |
| 6971 | ret = btrfs_discard_extent(root, start, len, NULL); | 6971 | ret = btrfs_discard_extent(root, start, len, NULL); |
| 6972 | 6972 | ||
| 6973 | if (pin) | 6973 | if (pin) |
| 6974 | pin_down_extent(root, cache, start, len, 1); | 6974 | pin_down_extent(root, cache, start, len, 1); |
| 6975 | else { | 6975 | else { |
| 6976 | btrfs_add_free_space(cache, start, len); | 6976 | btrfs_add_free_space(cache, start, len); |
| 6977 | btrfs_update_reserved_bytes(cache, len, RESERVE_FREE, delalloc); | 6977 | btrfs_update_reserved_bytes(cache, len, RESERVE_FREE, delalloc); |
| 6978 | } | 6978 | } |
| 6979 | btrfs_put_block_group(cache); | 6979 | btrfs_put_block_group(cache); |
| 6980 | 6980 | ||
| 6981 | trace_btrfs_reserved_extent_free(root, start, len); | 6981 | trace_btrfs_reserved_extent_free(root, start, len); |
| 6982 | 6982 | ||
| 6983 | return ret; | 6983 | return ret; |
| 6984 | } | 6984 | } |
| 6985 | 6985 | ||
| 6986 | int btrfs_free_reserved_extent(struct btrfs_root *root, | 6986 | int btrfs_free_reserved_extent(struct btrfs_root *root, |
| 6987 | u64 start, u64 len, int delalloc) | 6987 | u64 start, u64 len, int delalloc) |
| 6988 | { | 6988 | { |
| 6989 | return __btrfs_free_reserved_extent(root, start, len, 0, delalloc); | 6989 | return __btrfs_free_reserved_extent(root, start, len, 0, delalloc); |
| 6990 | } | 6990 | } |
| 6991 | 6991 | ||
| 6992 | int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root, | 6992 | int btrfs_free_and_pin_reserved_extent(struct btrfs_root *root, |
| 6993 | u64 start, u64 len) | 6993 | u64 start, u64 len) |
| 6994 | { | 6994 | { |
| 6995 | return __btrfs_free_reserved_extent(root, start, len, 1, 0); | 6995 | return __btrfs_free_reserved_extent(root, start, len, 1, 0); |
| 6996 | } | 6996 | } |
| 6997 | 6997 | ||
| 6998 | static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, | 6998 | static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans, |
| 6999 | struct btrfs_root *root, | 6999 | struct btrfs_root *root, |
| 7000 | u64 parent, u64 root_objectid, | 7000 | u64 parent, u64 root_objectid, |
| 7001 | u64 flags, u64 owner, u64 offset, | 7001 | u64 flags, u64 owner, u64 offset, |
| 7002 | struct btrfs_key *ins, int ref_mod) | 7002 | struct btrfs_key *ins, int ref_mod) |
| 7003 | { | 7003 | { |
| 7004 | int ret; | 7004 | int ret; |
| 7005 | struct btrfs_fs_info *fs_info = root->fs_info; | 7005 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 7006 | struct btrfs_extent_item *extent_item; | 7006 | struct btrfs_extent_item *extent_item; |
| 7007 | struct btrfs_extent_inline_ref *iref; | 7007 | struct btrfs_extent_inline_ref *iref; |
| 7008 | struct btrfs_path *path; | 7008 | struct btrfs_path *path; |
| 7009 | struct extent_buffer *leaf; | 7009 | struct extent_buffer *leaf; |
| 7010 | int type; | 7010 | int type; |
| 7011 | u32 size; | 7011 | u32 size; |
| 7012 | 7012 | ||
| 7013 | if (parent > 0) | 7013 | if (parent > 0) |
| 7014 | type = BTRFS_SHARED_DATA_REF_KEY; | 7014 | type = BTRFS_SHARED_DATA_REF_KEY; |
| 7015 | else | 7015 | else |
| 7016 | type = BTRFS_EXTENT_DATA_REF_KEY; | 7016 | type = BTRFS_EXTENT_DATA_REF_KEY; |
| 7017 | 7017 | ||
| 7018 | size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type); | 7018 | size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type); |
| 7019 | 7019 | ||
| 7020 | path = btrfs_alloc_path(); | 7020 | path = btrfs_alloc_path(); |
| 7021 | if (!path) | 7021 | if (!path) |
| 7022 | return -ENOMEM; | 7022 | return -ENOMEM; |
| 7023 | 7023 | ||
| 7024 | path->leave_spinning = 1; | 7024 | path->leave_spinning = 1; |
| 7025 | ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, | 7025 | ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, |
| 7026 | ins, size); | 7026 | ins, size); |
| 7027 | if (ret) { | 7027 | if (ret) { |
| 7028 | btrfs_free_path(path); | 7028 | btrfs_free_path(path); |
| 7029 | return ret; | 7029 | return ret; |
| 7030 | } | 7030 | } |
| 7031 | 7031 | ||
| 7032 | leaf = path->nodes[0]; | 7032 | leaf = path->nodes[0]; |
| 7033 | extent_item = btrfs_item_ptr(leaf, path->slots[0], | 7033 | extent_item = btrfs_item_ptr(leaf, path->slots[0], |
| 7034 | struct btrfs_extent_item); | 7034 | struct btrfs_extent_item); |
| 7035 | btrfs_set_extent_refs(leaf, extent_item, ref_mod); | 7035 | btrfs_set_extent_refs(leaf, extent_item, ref_mod); |
| 7036 | btrfs_set_extent_generation(leaf, extent_item, trans->transid); | 7036 | btrfs_set_extent_generation(leaf, extent_item, trans->transid); |
| 7037 | btrfs_set_extent_flags(leaf, extent_item, | 7037 | btrfs_set_extent_flags(leaf, extent_item, |
| 7038 | flags | BTRFS_EXTENT_FLAG_DATA); | 7038 | flags | BTRFS_EXTENT_FLAG_DATA); |
| 7039 | 7039 | ||
| 7040 | iref = (struct btrfs_extent_inline_ref *)(extent_item + 1); | 7040 | iref = (struct btrfs_extent_inline_ref *)(extent_item + 1); |
| 7041 | btrfs_set_extent_inline_ref_type(leaf, iref, type); | 7041 | btrfs_set_extent_inline_ref_type(leaf, iref, type); |
| 7042 | if (parent > 0) { | 7042 | if (parent > 0) { |
| 7043 | struct btrfs_shared_data_ref *ref; | 7043 | struct btrfs_shared_data_ref *ref; |
| 7044 | ref = (struct btrfs_shared_data_ref *)(iref + 1); | 7044 | ref = (struct btrfs_shared_data_ref *)(iref + 1); |
| 7045 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); | 7045 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); |
| 7046 | btrfs_set_shared_data_ref_count(leaf, ref, ref_mod); | 7046 | btrfs_set_shared_data_ref_count(leaf, ref, ref_mod); |
| 7047 | } else { | 7047 | } else { |
| 7048 | struct btrfs_extent_data_ref *ref; | 7048 | struct btrfs_extent_data_ref *ref; |
| 7049 | ref = (struct btrfs_extent_data_ref *)(&iref->offset); | 7049 | ref = (struct btrfs_extent_data_ref *)(&iref->offset); |
| 7050 | btrfs_set_extent_data_ref_root(leaf, ref, root_objectid); | 7050 | btrfs_set_extent_data_ref_root(leaf, ref, root_objectid); |
| 7051 | btrfs_set_extent_data_ref_objectid(leaf, ref, owner); | 7051 | btrfs_set_extent_data_ref_objectid(leaf, ref, owner); |
| 7052 | btrfs_set_extent_data_ref_offset(leaf, ref, offset); | 7052 | btrfs_set_extent_data_ref_offset(leaf, ref, offset); |
| 7053 | btrfs_set_extent_data_ref_count(leaf, ref, ref_mod); | 7053 | btrfs_set_extent_data_ref_count(leaf, ref, ref_mod); |
| 7054 | } | 7054 | } |
| 7055 | 7055 | ||
| 7056 | btrfs_mark_buffer_dirty(path->nodes[0]); | 7056 | btrfs_mark_buffer_dirty(path->nodes[0]); |
| 7057 | btrfs_free_path(path); | 7057 | btrfs_free_path(path); |
| 7058 | 7058 | ||
| 7059 | /* Always set parent to 0 here since its exclusive anyway. */ | 7059 | /* Always set parent to 0 here since its exclusive anyway. */ |
| 7060 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, | 7060 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, |
| 7061 | ins->objectid, ins->offset, | 7061 | ins->objectid, ins->offset, |
| 7062 | BTRFS_QGROUP_OPER_ADD_EXCL, 0); | 7062 | BTRFS_QGROUP_OPER_ADD_EXCL, 0); |
| 7063 | if (ret) | 7063 | if (ret) |
| 7064 | return ret; | 7064 | return ret; |
| 7065 | 7065 | ||
| 7066 | ret = update_block_group(root, ins->objectid, ins->offset, 1); | 7066 | ret = update_block_group(root, ins->objectid, ins->offset, 1); |
| 7067 | if (ret) { /* -ENOENT, logic error */ | 7067 | if (ret) { /* -ENOENT, logic error */ |
| 7068 | btrfs_err(fs_info, "update block group failed for %llu %llu", | 7068 | btrfs_err(fs_info, "update block group failed for %llu %llu", |
| 7069 | ins->objectid, ins->offset); | 7069 | ins->objectid, ins->offset); |
| 7070 | BUG(); | 7070 | BUG(); |
| 7071 | } | 7071 | } |
| 7072 | trace_btrfs_reserved_extent_alloc(root, ins->objectid, ins->offset); | 7072 | trace_btrfs_reserved_extent_alloc(root, ins->objectid, ins->offset); |
| 7073 | return ret; | 7073 | return ret; |
| 7074 | } | 7074 | } |
| 7075 | 7075 | ||
| 7076 | static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, | 7076 | static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans, |
| 7077 | struct btrfs_root *root, | 7077 | struct btrfs_root *root, |
| 7078 | u64 parent, u64 root_objectid, | 7078 | u64 parent, u64 root_objectid, |
| 7079 | u64 flags, struct btrfs_disk_key *key, | 7079 | u64 flags, struct btrfs_disk_key *key, |
| 7080 | int level, struct btrfs_key *ins, | 7080 | int level, struct btrfs_key *ins, |
| 7081 | int no_quota) | 7081 | int no_quota) |
| 7082 | { | 7082 | { |
| 7083 | int ret; | 7083 | int ret; |
| 7084 | struct btrfs_fs_info *fs_info = root->fs_info; | 7084 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 7085 | struct btrfs_extent_item *extent_item; | 7085 | struct btrfs_extent_item *extent_item; |
| 7086 | struct btrfs_tree_block_info *block_info; | 7086 | struct btrfs_tree_block_info *block_info; |
| 7087 | struct btrfs_extent_inline_ref *iref; | 7087 | struct btrfs_extent_inline_ref *iref; |
| 7088 | struct btrfs_path *path; | 7088 | struct btrfs_path *path; |
| 7089 | struct extent_buffer *leaf; | 7089 | struct extent_buffer *leaf; |
| 7090 | u32 size = sizeof(*extent_item) + sizeof(*iref); | 7090 | u32 size = sizeof(*extent_item) + sizeof(*iref); |
| 7091 | u64 num_bytes = ins->offset; | 7091 | u64 num_bytes = ins->offset; |
| 7092 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, | 7092 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, |
| 7093 | SKINNY_METADATA); | 7093 | SKINNY_METADATA); |
| 7094 | 7094 | ||
| 7095 | if (!skinny_metadata) | 7095 | if (!skinny_metadata) |
| 7096 | size += sizeof(*block_info); | 7096 | size += sizeof(*block_info); |
| 7097 | 7097 | ||
| 7098 | path = btrfs_alloc_path(); | 7098 | path = btrfs_alloc_path(); |
| 7099 | if (!path) { | 7099 | if (!path) { |
| 7100 | btrfs_free_and_pin_reserved_extent(root, ins->objectid, | 7100 | btrfs_free_and_pin_reserved_extent(root, ins->objectid, |
| 7101 | root->nodesize); | 7101 | root->nodesize); |
| 7102 | return -ENOMEM; | 7102 | return -ENOMEM; |
| 7103 | } | 7103 | } |
| 7104 | 7104 | ||
| 7105 | path->leave_spinning = 1; | 7105 | path->leave_spinning = 1; |
| 7106 | ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, | 7106 | ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path, |
| 7107 | ins, size); | 7107 | ins, size); |
| 7108 | if (ret) { | 7108 | if (ret) { |
| 7109 | btrfs_free_and_pin_reserved_extent(root, ins->objectid, | 7109 | btrfs_free_and_pin_reserved_extent(root, ins->objectid, |
| 7110 | root->nodesize); | 7110 | root->nodesize); |
| 7111 | btrfs_free_path(path); | 7111 | btrfs_free_path(path); |
| 7112 | return ret; | 7112 | return ret; |
| 7113 | } | 7113 | } |
| 7114 | 7114 | ||
| 7115 | leaf = path->nodes[0]; | 7115 | leaf = path->nodes[0]; |
| 7116 | extent_item = btrfs_item_ptr(leaf, path->slots[0], | 7116 | extent_item = btrfs_item_ptr(leaf, path->slots[0], |
| 7117 | struct btrfs_extent_item); | 7117 | struct btrfs_extent_item); |
| 7118 | btrfs_set_extent_refs(leaf, extent_item, 1); | 7118 | btrfs_set_extent_refs(leaf, extent_item, 1); |
| 7119 | btrfs_set_extent_generation(leaf, extent_item, trans->transid); | 7119 | btrfs_set_extent_generation(leaf, extent_item, trans->transid); |
| 7120 | btrfs_set_extent_flags(leaf, extent_item, | 7120 | btrfs_set_extent_flags(leaf, extent_item, |
| 7121 | flags | BTRFS_EXTENT_FLAG_TREE_BLOCK); | 7121 | flags | BTRFS_EXTENT_FLAG_TREE_BLOCK); |
| 7122 | 7122 | ||
| 7123 | if (skinny_metadata) { | 7123 | if (skinny_metadata) { |
| 7124 | iref = (struct btrfs_extent_inline_ref *)(extent_item + 1); | 7124 | iref = (struct btrfs_extent_inline_ref *)(extent_item + 1); |
| 7125 | num_bytes = root->nodesize; | 7125 | num_bytes = root->nodesize; |
| 7126 | } else { | 7126 | } else { |
| 7127 | block_info = (struct btrfs_tree_block_info *)(extent_item + 1); | 7127 | block_info = (struct btrfs_tree_block_info *)(extent_item + 1); |
| 7128 | btrfs_set_tree_block_key(leaf, block_info, key); | 7128 | btrfs_set_tree_block_key(leaf, block_info, key); |
| 7129 | btrfs_set_tree_block_level(leaf, block_info, level); | 7129 | btrfs_set_tree_block_level(leaf, block_info, level); |
| 7130 | iref = (struct btrfs_extent_inline_ref *)(block_info + 1); | 7130 | iref = (struct btrfs_extent_inline_ref *)(block_info + 1); |
| 7131 | } | 7131 | } |
| 7132 | 7132 | ||
| 7133 | if (parent > 0) { | 7133 | if (parent > 0) { |
| 7134 | BUG_ON(!(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)); | 7134 | BUG_ON(!(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)); |
| 7135 | btrfs_set_extent_inline_ref_type(leaf, iref, | 7135 | btrfs_set_extent_inline_ref_type(leaf, iref, |
| 7136 | BTRFS_SHARED_BLOCK_REF_KEY); | 7136 | BTRFS_SHARED_BLOCK_REF_KEY); |
| 7137 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); | 7137 | btrfs_set_extent_inline_ref_offset(leaf, iref, parent); |
| 7138 | } else { | 7138 | } else { |
| 7139 | btrfs_set_extent_inline_ref_type(leaf, iref, | 7139 | btrfs_set_extent_inline_ref_type(leaf, iref, |
| 7140 | BTRFS_TREE_BLOCK_REF_KEY); | 7140 | BTRFS_TREE_BLOCK_REF_KEY); |
| 7141 | btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid); | 7141 | btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid); |
| 7142 | } | 7142 | } |
| 7143 | 7143 | ||
| 7144 | btrfs_mark_buffer_dirty(leaf); | 7144 | btrfs_mark_buffer_dirty(leaf); |
| 7145 | btrfs_free_path(path); | 7145 | btrfs_free_path(path); |
| 7146 | 7146 | ||
| 7147 | if (!no_quota) { | 7147 | if (!no_quota) { |
| 7148 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, | 7148 | ret = btrfs_qgroup_record_ref(trans, fs_info, root_objectid, |
| 7149 | ins->objectid, num_bytes, | 7149 | ins->objectid, num_bytes, |
| 7150 | BTRFS_QGROUP_OPER_ADD_EXCL, 0); | 7150 | BTRFS_QGROUP_OPER_ADD_EXCL, 0); |
| 7151 | if (ret) | 7151 | if (ret) |
| 7152 | return ret; | 7152 | return ret; |
| 7153 | } | 7153 | } |
| 7154 | 7154 | ||
| 7155 | ret = update_block_group(root, ins->objectid, root->nodesize, 1); | 7155 | ret = update_block_group(root, ins->objectid, root->nodesize, 1); |
| 7156 | if (ret) { /* -ENOENT, logic error */ | 7156 | if (ret) { /* -ENOENT, logic error */ |
| 7157 | btrfs_err(fs_info, "update block group failed for %llu %llu", | 7157 | btrfs_err(fs_info, "update block group failed for %llu %llu", |
| 7158 | ins->objectid, ins->offset); | 7158 | ins->objectid, ins->offset); |
| 7159 | BUG(); | 7159 | BUG(); |
| 7160 | } | 7160 | } |
| 7161 | 7161 | ||
| 7162 | trace_btrfs_reserved_extent_alloc(root, ins->objectid, root->nodesize); | 7162 | trace_btrfs_reserved_extent_alloc(root, ins->objectid, root->nodesize); |
| 7163 | return ret; | 7163 | return ret; |
| 7164 | } | 7164 | } |
| 7165 | 7165 | ||
| 7166 | int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans, | 7166 | int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans, |
| 7167 | struct btrfs_root *root, | 7167 | struct btrfs_root *root, |
| 7168 | u64 root_objectid, u64 owner, | 7168 | u64 root_objectid, u64 owner, |
| 7169 | u64 offset, struct btrfs_key *ins) | 7169 | u64 offset, struct btrfs_key *ins) |
| 7170 | { | 7170 | { |
| 7171 | int ret; | 7171 | int ret; |
| 7172 | 7172 | ||
| 7173 | BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID); | 7173 | BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID); |
| 7174 | 7174 | ||
| 7175 | ret = btrfs_add_delayed_data_ref(root->fs_info, trans, ins->objectid, | 7175 | ret = btrfs_add_delayed_data_ref(root->fs_info, trans, ins->objectid, |
| 7176 | ins->offset, 0, | 7176 | ins->offset, 0, |
| 7177 | root_objectid, owner, offset, | 7177 | root_objectid, owner, offset, |
| 7178 | BTRFS_ADD_DELAYED_EXTENT, NULL, 0); | 7178 | BTRFS_ADD_DELAYED_EXTENT, NULL, 0); |
| 7179 | return ret; | 7179 | return ret; |
| 7180 | } | 7180 | } |
| 7181 | 7181 | ||
| 7182 | /* | 7182 | /* |
| 7183 | * this is used by the tree logging recovery code. It records that | 7183 | * this is used by the tree logging recovery code. It records that |
| 7184 | * an extent has been allocated and makes sure to clear the free | 7184 | * an extent has been allocated and makes sure to clear the free |
| 7185 | * space cache bits as well | 7185 | * space cache bits as well |
| 7186 | */ | 7186 | */ |
| 7187 | int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, | 7187 | int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans, |
| 7188 | struct btrfs_root *root, | 7188 | struct btrfs_root *root, |
| 7189 | u64 root_objectid, u64 owner, u64 offset, | 7189 | u64 root_objectid, u64 owner, u64 offset, |
| 7190 | struct btrfs_key *ins) | 7190 | struct btrfs_key *ins) |
| 7191 | { | 7191 | { |
| 7192 | int ret; | 7192 | int ret; |
| 7193 | struct btrfs_block_group_cache *block_group; | 7193 | struct btrfs_block_group_cache *block_group; |
| 7194 | 7194 | ||
| 7195 | /* | 7195 | /* |
| 7196 | * Mixed block groups will exclude before processing the log so we only | 7196 | * Mixed block groups will exclude before processing the log so we only |
| 7197 | * need to do the exlude dance if this fs isn't mixed. | 7197 | * need to do the exlude dance if this fs isn't mixed. |
| 7198 | */ | 7198 | */ |
| 7199 | if (!btrfs_fs_incompat(root->fs_info, MIXED_GROUPS)) { | 7199 | if (!btrfs_fs_incompat(root->fs_info, MIXED_GROUPS)) { |
| 7200 | ret = __exclude_logged_extent(root, ins->objectid, ins->offset); | 7200 | ret = __exclude_logged_extent(root, ins->objectid, ins->offset); |
| 7201 | if (ret) | 7201 | if (ret) |
| 7202 | return ret; | 7202 | return ret; |
| 7203 | } | 7203 | } |
| 7204 | 7204 | ||
| 7205 | block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid); | 7205 | block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid); |
| 7206 | if (!block_group) | 7206 | if (!block_group) |
| 7207 | return -EINVAL; | 7207 | return -EINVAL; |
| 7208 | 7208 | ||
| 7209 | ret = btrfs_update_reserved_bytes(block_group, ins->offset, | 7209 | ret = btrfs_update_reserved_bytes(block_group, ins->offset, |
| 7210 | RESERVE_ALLOC_NO_ACCOUNT, 0); | 7210 | RESERVE_ALLOC_NO_ACCOUNT, 0); |
| 7211 | BUG_ON(ret); /* logic error */ | 7211 | BUG_ON(ret); /* logic error */ |
| 7212 | ret = alloc_reserved_file_extent(trans, root, 0, root_objectid, | 7212 | ret = alloc_reserved_file_extent(trans, root, 0, root_objectid, |
| 7213 | 0, owner, offset, ins, 1); | 7213 | 0, owner, offset, ins, 1); |
| 7214 | btrfs_put_block_group(block_group); | 7214 | btrfs_put_block_group(block_group); |
| 7215 | return ret; | 7215 | return ret; |
| 7216 | } | 7216 | } |
| 7217 | 7217 | ||
| 7218 | static struct extent_buffer * | 7218 | static struct extent_buffer * |
| 7219 | btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root, | 7219 | btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
| 7220 | u64 bytenr, u32 blocksize, int level) | 7220 | u64 bytenr, u32 blocksize, int level) |
| 7221 | { | 7221 | { |
| 7222 | struct extent_buffer *buf; | 7222 | struct extent_buffer *buf; |
| 7223 | 7223 | ||
| 7224 | buf = btrfs_find_create_tree_block(root, bytenr, blocksize); | 7224 | buf = btrfs_find_create_tree_block(root, bytenr, blocksize); |
| 7225 | if (!buf) | 7225 | if (!buf) |
| 7226 | return ERR_PTR(-ENOMEM); | 7226 | return ERR_PTR(-ENOMEM); |
| 7227 | btrfs_set_header_generation(buf, trans->transid); | 7227 | btrfs_set_header_generation(buf, trans->transid); |
| 7228 | btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level); | 7228 | btrfs_set_buffer_lockdep_class(root->root_key.objectid, buf, level); |
| 7229 | btrfs_tree_lock(buf); | 7229 | btrfs_tree_lock(buf); |
| 7230 | clean_tree_block(trans, root, buf); | 7230 | clean_tree_block(trans, root, buf); |
| 7231 | clear_bit(EXTENT_BUFFER_STALE, &buf->bflags); | 7231 | clear_bit(EXTENT_BUFFER_STALE, &buf->bflags); |
| 7232 | 7232 | ||
| 7233 | btrfs_set_lock_blocking(buf); | 7233 | btrfs_set_lock_blocking(buf); |
| 7234 | btrfs_set_buffer_uptodate(buf); | 7234 | btrfs_set_buffer_uptodate(buf); |
| 7235 | 7235 | ||
| 7236 | if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) { | 7236 | if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) { |
| 7237 | buf->log_index = root->log_transid % 2; | 7237 | buf->log_index = root->log_transid % 2; |
| 7238 | /* | 7238 | /* |
| 7239 | * we allow two log transactions at a time, use different | 7239 | * we allow two log transactions at a time, use different |
| 7240 | * EXENT bit to differentiate dirty pages. | 7240 | * EXENT bit to differentiate dirty pages. |
| 7241 | */ | 7241 | */ |
| 7242 | if (buf->log_index == 0) | 7242 | if (buf->log_index == 0) |
| 7243 | set_extent_dirty(&root->dirty_log_pages, buf->start, | 7243 | set_extent_dirty(&root->dirty_log_pages, buf->start, |
| 7244 | buf->start + buf->len - 1, GFP_NOFS); | 7244 | buf->start + buf->len - 1, GFP_NOFS); |
| 7245 | else | 7245 | else |
| 7246 | set_extent_new(&root->dirty_log_pages, buf->start, | 7246 | set_extent_new(&root->dirty_log_pages, buf->start, |
| 7247 | buf->start + buf->len - 1, GFP_NOFS); | 7247 | buf->start + buf->len - 1, GFP_NOFS); |
| 7248 | } else { | 7248 | } else { |
| 7249 | buf->log_index = -1; | 7249 | buf->log_index = -1; |
| 7250 | set_extent_dirty(&trans->transaction->dirty_pages, buf->start, | 7250 | set_extent_dirty(&trans->transaction->dirty_pages, buf->start, |
| 7251 | buf->start + buf->len - 1, GFP_NOFS); | 7251 | buf->start + buf->len - 1, GFP_NOFS); |
| 7252 | } | 7252 | } |
| 7253 | trans->blocks_used++; | 7253 | trans->blocks_used++; |
| 7254 | /* this returns a buffer locked for blocking */ | 7254 | /* this returns a buffer locked for blocking */ |
| 7255 | return buf; | 7255 | return buf; |
| 7256 | } | 7256 | } |
| 7257 | 7257 | ||
| 7258 | static struct btrfs_block_rsv * | 7258 | static struct btrfs_block_rsv * |
| 7259 | use_block_rsv(struct btrfs_trans_handle *trans, | 7259 | use_block_rsv(struct btrfs_trans_handle *trans, |
| 7260 | struct btrfs_root *root, u32 blocksize) | 7260 | struct btrfs_root *root, u32 blocksize) |
| 7261 | { | 7261 | { |
| 7262 | struct btrfs_block_rsv *block_rsv; | 7262 | struct btrfs_block_rsv *block_rsv; |
| 7263 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; | 7263 | struct btrfs_block_rsv *global_rsv = &root->fs_info->global_block_rsv; |
| 7264 | int ret; | 7264 | int ret; |
| 7265 | bool global_updated = false; | 7265 | bool global_updated = false; |
| 7266 | 7266 | ||
| 7267 | block_rsv = get_block_rsv(trans, root); | 7267 | block_rsv = get_block_rsv(trans, root); |
| 7268 | 7268 | ||
| 7269 | if (unlikely(block_rsv->size == 0)) | 7269 | if (unlikely(block_rsv->size == 0)) |
| 7270 | goto try_reserve; | 7270 | goto try_reserve; |
| 7271 | again: | 7271 | again: |
| 7272 | ret = block_rsv_use_bytes(block_rsv, blocksize); | 7272 | ret = block_rsv_use_bytes(block_rsv, blocksize); |
| 7273 | if (!ret) | 7273 | if (!ret) |
| 7274 | return block_rsv; | 7274 | return block_rsv; |
| 7275 | 7275 | ||
| 7276 | if (block_rsv->failfast) | 7276 | if (block_rsv->failfast) |
| 7277 | return ERR_PTR(ret); | 7277 | return ERR_PTR(ret); |
| 7278 | 7278 | ||
| 7279 | if (block_rsv->type == BTRFS_BLOCK_RSV_GLOBAL && !global_updated) { | 7279 | if (block_rsv->type == BTRFS_BLOCK_RSV_GLOBAL && !global_updated) { |
| 7280 | global_updated = true; | 7280 | global_updated = true; |
| 7281 | update_global_block_rsv(root->fs_info); | 7281 | update_global_block_rsv(root->fs_info); |
| 7282 | goto again; | 7282 | goto again; |
| 7283 | } | 7283 | } |
| 7284 | 7284 | ||
| 7285 | if (btrfs_test_opt(root, ENOSPC_DEBUG)) { | 7285 | if (btrfs_test_opt(root, ENOSPC_DEBUG)) { |
| 7286 | static DEFINE_RATELIMIT_STATE(_rs, | 7286 | static DEFINE_RATELIMIT_STATE(_rs, |
| 7287 | DEFAULT_RATELIMIT_INTERVAL * 10, | 7287 | DEFAULT_RATELIMIT_INTERVAL * 10, |
| 7288 | /*DEFAULT_RATELIMIT_BURST*/ 1); | 7288 | /*DEFAULT_RATELIMIT_BURST*/ 1); |
| 7289 | if (__ratelimit(&_rs)) | 7289 | if (__ratelimit(&_rs)) |
| 7290 | WARN(1, KERN_DEBUG | 7290 | WARN(1, KERN_DEBUG |
| 7291 | "BTRFS: block rsv returned %d\n", ret); | 7291 | "BTRFS: block rsv returned %d\n", ret); |
| 7292 | } | 7292 | } |
| 7293 | try_reserve: | 7293 | try_reserve: |
| 7294 | ret = reserve_metadata_bytes(root, block_rsv, blocksize, | 7294 | ret = reserve_metadata_bytes(root, block_rsv, blocksize, |
| 7295 | BTRFS_RESERVE_NO_FLUSH); | 7295 | BTRFS_RESERVE_NO_FLUSH); |
| 7296 | if (!ret) | 7296 | if (!ret) |
| 7297 | return block_rsv; | 7297 | return block_rsv; |
| 7298 | /* | 7298 | /* |
| 7299 | * If we couldn't reserve metadata bytes try and use some from | 7299 | * If we couldn't reserve metadata bytes try and use some from |
| 7300 | * the global reserve if its space type is the same as the global | 7300 | * the global reserve if its space type is the same as the global |
| 7301 | * reservation. | 7301 | * reservation. |
| 7302 | */ | 7302 | */ |
| 7303 | if (block_rsv->type != BTRFS_BLOCK_RSV_GLOBAL && | 7303 | if (block_rsv->type != BTRFS_BLOCK_RSV_GLOBAL && |
| 7304 | block_rsv->space_info == global_rsv->space_info) { | 7304 | block_rsv->space_info == global_rsv->space_info) { |
| 7305 | ret = block_rsv_use_bytes(global_rsv, blocksize); | 7305 | ret = block_rsv_use_bytes(global_rsv, blocksize); |
| 7306 | if (!ret) | 7306 | if (!ret) |
| 7307 | return global_rsv; | 7307 | return global_rsv; |
| 7308 | } | 7308 | } |
| 7309 | return ERR_PTR(ret); | 7309 | return ERR_PTR(ret); |
| 7310 | } | 7310 | } |
| 7311 | 7311 | ||
| 7312 | static void unuse_block_rsv(struct btrfs_fs_info *fs_info, | 7312 | static void unuse_block_rsv(struct btrfs_fs_info *fs_info, |
| 7313 | struct btrfs_block_rsv *block_rsv, u32 blocksize) | 7313 | struct btrfs_block_rsv *block_rsv, u32 blocksize) |
| 7314 | { | 7314 | { |
| 7315 | block_rsv_add_bytes(block_rsv, blocksize, 0); | 7315 | block_rsv_add_bytes(block_rsv, blocksize, 0); |
| 7316 | block_rsv_release_bytes(fs_info, block_rsv, NULL, 0); | 7316 | block_rsv_release_bytes(fs_info, block_rsv, NULL, 0); |
| 7317 | } | 7317 | } |
| 7318 | 7318 | ||
| 7319 | /* | 7319 | /* |
| 7320 | * finds a free extent and does all the dirty work required for allocation | 7320 | * finds a free extent and does all the dirty work required for allocation |
| 7321 | * returns the key for the extent through ins, and a tree buffer for | 7321 | * returns the key for the extent through ins, and a tree buffer for |
| 7322 | * the first block of the extent through buf. | 7322 | * the first block of the extent through buf. |
| 7323 | * | 7323 | * |
| 7324 | * returns the tree buffer or NULL. | 7324 | * returns the tree buffer or NULL. |
| 7325 | */ | 7325 | */ |
| 7326 | struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, | 7326 | struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans, |
| 7327 | struct btrfs_root *root, | 7327 | struct btrfs_root *root, |
| 7328 | u64 parent, u64 root_objectid, | 7328 | u64 parent, u64 root_objectid, |
| 7329 | struct btrfs_disk_key *key, int level, | 7329 | struct btrfs_disk_key *key, int level, |
| 7330 | u64 hint, u64 empty_size) | 7330 | u64 hint, u64 empty_size) |
| 7331 | { | 7331 | { |
| 7332 | struct btrfs_key ins; | 7332 | struct btrfs_key ins; |
| 7333 | struct btrfs_block_rsv *block_rsv; | 7333 | struct btrfs_block_rsv *block_rsv; |
| 7334 | struct extent_buffer *buf; | 7334 | struct extent_buffer *buf; |
| 7335 | u64 flags = 0; | 7335 | u64 flags = 0; |
| 7336 | int ret; | 7336 | int ret; |
| 7337 | u32 blocksize = root->nodesize; | 7337 | u32 blocksize = root->nodesize; |
| 7338 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, | 7338 | bool skinny_metadata = btrfs_fs_incompat(root->fs_info, |
| 7339 | SKINNY_METADATA); | 7339 | SKINNY_METADATA); |
| 7340 | 7340 | ||
| 7341 | if (btrfs_test_is_dummy_root(root)) { | 7341 | if (btrfs_test_is_dummy_root(root)) { |
| 7342 | buf = btrfs_init_new_buffer(trans, root, root->alloc_bytenr, | 7342 | buf = btrfs_init_new_buffer(trans, root, root->alloc_bytenr, |
| 7343 | blocksize, level); | 7343 | blocksize, level); |
| 7344 | if (!IS_ERR(buf)) | 7344 | if (!IS_ERR(buf)) |
| 7345 | root->alloc_bytenr += blocksize; | 7345 | root->alloc_bytenr += blocksize; |
| 7346 | return buf; | 7346 | return buf; |
| 7347 | } | 7347 | } |
| 7348 | 7348 | ||
| 7349 | block_rsv = use_block_rsv(trans, root, blocksize); | 7349 | block_rsv = use_block_rsv(trans, root, blocksize); |
| 7350 | if (IS_ERR(block_rsv)) | 7350 | if (IS_ERR(block_rsv)) |
| 7351 | return ERR_CAST(block_rsv); | 7351 | return ERR_CAST(block_rsv); |
| 7352 | 7352 | ||
| 7353 | ret = btrfs_reserve_extent(root, blocksize, blocksize, | 7353 | ret = btrfs_reserve_extent(root, blocksize, blocksize, |
| 7354 | empty_size, hint, &ins, 0, 0); | 7354 | empty_size, hint, &ins, 0, 0); |
| 7355 | if (ret) { | 7355 | if (ret) { |
| 7356 | unuse_block_rsv(root->fs_info, block_rsv, blocksize); | 7356 | unuse_block_rsv(root->fs_info, block_rsv, blocksize); |
| 7357 | return ERR_PTR(ret); | 7357 | return ERR_PTR(ret); |
| 7358 | } | 7358 | } |
| 7359 | 7359 | ||
| 7360 | buf = btrfs_init_new_buffer(trans, root, ins.objectid, | 7360 | buf = btrfs_init_new_buffer(trans, root, ins.objectid, |
| 7361 | blocksize, level); | 7361 | blocksize, level); |
| 7362 | BUG_ON(IS_ERR(buf)); /* -ENOMEM */ | 7362 | BUG_ON(IS_ERR(buf)); /* -ENOMEM */ |
| 7363 | 7363 | ||
| 7364 | if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) { | 7364 | if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) { |
| 7365 | if (parent == 0) | 7365 | if (parent == 0) |
| 7366 | parent = ins.objectid; | 7366 | parent = ins.objectid; |
| 7367 | flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF; | 7367 | flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF; |
| 7368 | } else | 7368 | } else |
| 7369 | BUG_ON(parent > 0); | 7369 | BUG_ON(parent > 0); |
| 7370 | 7370 | ||
| 7371 | if (root_objectid != BTRFS_TREE_LOG_OBJECTID) { | 7371 | if (root_objectid != BTRFS_TREE_LOG_OBJECTID) { |
| 7372 | struct btrfs_delayed_extent_op *extent_op; | 7372 | struct btrfs_delayed_extent_op *extent_op; |
| 7373 | extent_op = btrfs_alloc_delayed_extent_op(); | 7373 | extent_op = btrfs_alloc_delayed_extent_op(); |
| 7374 | BUG_ON(!extent_op); /* -ENOMEM */ | 7374 | BUG_ON(!extent_op); /* -ENOMEM */ |
| 7375 | if (key) | 7375 | if (key) |
| 7376 | memcpy(&extent_op->key, key, sizeof(extent_op->key)); | 7376 | memcpy(&extent_op->key, key, sizeof(extent_op->key)); |
| 7377 | else | 7377 | else |
| 7378 | memset(&extent_op->key, 0, sizeof(extent_op->key)); | 7378 | memset(&extent_op->key, 0, sizeof(extent_op->key)); |
| 7379 | extent_op->flags_to_set = flags; | 7379 | extent_op->flags_to_set = flags; |
| 7380 | if (skinny_metadata) | 7380 | if (skinny_metadata) |
| 7381 | extent_op->update_key = 0; | 7381 | extent_op->update_key = 0; |
| 7382 | else | 7382 | else |
| 7383 | extent_op->update_key = 1; | 7383 | extent_op->update_key = 1; |
| 7384 | extent_op->update_flags = 1; | 7384 | extent_op->update_flags = 1; |
| 7385 | extent_op->is_data = 0; | 7385 | extent_op->is_data = 0; |
| 7386 | extent_op->level = level; | 7386 | extent_op->level = level; |
| 7387 | 7387 | ||
| 7388 | ret = btrfs_add_delayed_tree_ref(root->fs_info, trans, | 7388 | ret = btrfs_add_delayed_tree_ref(root->fs_info, trans, |
| 7389 | ins.objectid, | 7389 | ins.objectid, |
| 7390 | ins.offset, parent, root_objectid, | 7390 | ins.offset, parent, root_objectid, |
| 7391 | level, BTRFS_ADD_DELAYED_EXTENT, | 7391 | level, BTRFS_ADD_DELAYED_EXTENT, |
| 7392 | extent_op, 0); | 7392 | extent_op, 0); |
| 7393 | BUG_ON(ret); /* -ENOMEM */ | 7393 | BUG_ON(ret); /* -ENOMEM */ |
| 7394 | } | 7394 | } |
| 7395 | return buf; | 7395 | return buf; |
| 7396 | } | 7396 | } |
| 7397 | 7397 | ||
| 7398 | struct walk_control { | 7398 | struct walk_control { |
| 7399 | u64 refs[BTRFS_MAX_LEVEL]; | 7399 | u64 refs[BTRFS_MAX_LEVEL]; |
| 7400 | u64 flags[BTRFS_MAX_LEVEL]; | 7400 | u64 flags[BTRFS_MAX_LEVEL]; |
| 7401 | struct btrfs_key update_progress; | 7401 | struct btrfs_key update_progress; |
| 7402 | int stage; | 7402 | int stage; |
| 7403 | int level; | 7403 | int level; |
| 7404 | int shared_level; | 7404 | int shared_level; |
| 7405 | int update_ref; | 7405 | int update_ref; |
| 7406 | int keep_locks; | 7406 | int keep_locks; |
| 7407 | int reada_slot; | 7407 | int reada_slot; |
| 7408 | int reada_count; | 7408 | int reada_count; |
| 7409 | int for_reloc; | 7409 | int for_reloc; |
| 7410 | }; | 7410 | }; |
| 7411 | 7411 | ||
| 7412 | #define DROP_REFERENCE 1 | 7412 | #define DROP_REFERENCE 1 |
| 7413 | #define UPDATE_BACKREF 2 | 7413 | #define UPDATE_BACKREF 2 |
| 7414 | 7414 | ||
| 7415 | static noinline void reada_walk_down(struct btrfs_trans_handle *trans, | 7415 | static noinline void reada_walk_down(struct btrfs_trans_handle *trans, |
| 7416 | struct btrfs_root *root, | 7416 | struct btrfs_root *root, |
| 7417 | struct walk_control *wc, | 7417 | struct walk_control *wc, |
| 7418 | struct btrfs_path *path) | 7418 | struct btrfs_path *path) |
| 7419 | { | 7419 | { |
| 7420 | u64 bytenr; | 7420 | u64 bytenr; |
| 7421 | u64 generation; | 7421 | u64 generation; |
| 7422 | u64 refs; | 7422 | u64 refs; |
| 7423 | u64 flags; | 7423 | u64 flags; |
| 7424 | u32 nritems; | 7424 | u32 nritems; |
| 7425 | u32 blocksize; | 7425 | u32 blocksize; |
| 7426 | struct btrfs_key key; | 7426 | struct btrfs_key key; |
| 7427 | struct extent_buffer *eb; | 7427 | struct extent_buffer *eb; |
| 7428 | int ret; | 7428 | int ret; |
| 7429 | int slot; | 7429 | int slot; |
| 7430 | int nread = 0; | 7430 | int nread = 0; |
| 7431 | 7431 | ||
| 7432 | if (path->slots[wc->level] < wc->reada_slot) { | 7432 | if (path->slots[wc->level] < wc->reada_slot) { |
| 7433 | wc->reada_count = wc->reada_count * 2 / 3; | 7433 | wc->reada_count = wc->reada_count * 2 / 3; |
| 7434 | wc->reada_count = max(wc->reada_count, 2); | 7434 | wc->reada_count = max(wc->reada_count, 2); |
| 7435 | } else { | 7435 | } else { |
| 7436 | wc->reada_count = wc->reada_count * 3 / 2; | 7436 | wc->reada_count = wc->reada_count * 3 / 2; |
| 7437 | wc->reada_count = min_t(int, wc->reada_count, | 7437 | wc->reada_count = min_t(int, wc->reada_count, |
| 7438 | BTRFS_NODEPTRS_PER_BLOCK(root)); | 7438 | BTRFS_NODEPTRS_PER_BLOCK(root)); |
| 7439 | } | 7439 | } |
| 7440 | 7440 | ||
| 7441 | eb = path->nodes[wc->level]; | 7441 | eb = path->nodes[wc->level]; |
| 7442 | nritems = btrfs_header_nritems(eb); | 7442 | nritems = btrfs_header_nritems(eb); |
| 7443 | blocksize = root->nodesize; | 7443 | blocksize = root->nodesize; |
| 7444 | 7444 | ||
| 7445 | for (slot = path->slots[wc->level]; slot < nritems; slot++) { | 7445 | for (slot = path->slots[wc->level]; slot < nritems; slot++) { |
| 7446 | if (nread >= wc->reada_count) | 7446 | if (nread >= wc->reada_count) |
| 7447 | break; | 7447 | break; |
| 7448 | 7448 | ||
| 7449 | cond_resched(); | 7449 | cond_resched(); |
| 7450 | bytenr = btrfs_node_blockptr(eb, slot); | 7450 | bytenr = btrfs_node_blockptr(eb, slot); |
| 7451 | generation = btrfs_node_ptr_generation(eb, slot); | 7451 | generation = btrfs_node_ptr_generation(eb, slot); |
| 7452 | 7452 | ||
| 7453 | if (slot == path->slots[wc->level]) | 7453 | if (slot == path->slots[wc->level]) |
| 7454 | goto reada; | 7454 | goto reada; |
| 7455 | 7455 | ||
| 7456 | if (wc->stage == UPDATE_BACKREF && | 7456 | if (wc->stage == UPDATE_BACKREF && |
| 7457 | generation <= root->root_key.offset) | 7457 | generation <= root->root_key.offset) |
| 7458 | continue; | 7458 | continue; |
| 7459 | 7459 | ||
| 7460 | /* We don't lock the tree block, it's OK to be racy here */ | 7460 | /* We don't lock the tree block, it's OK to be racy here */ |
| 7461 | ret = btrfs_lookup_extent_info(trans, root, bytenr, | 7461 | ret = btrfs_lookup_extent_info(trans, root, bytenr, |
| 7462 | wc->level - 1, 1, &refs, | 7462 | wc->level - 1, 1, &refs, |
| 7463 | &flags); | 7463 | &flags); |
| 7464 | /* We don't care about errors in readahead. */ | 7464 | /* We don't care about errors in readahead. */ |
| 7465 | if (ret < 0) | 7465 | if (ret < 0) |
| 7466 | continue; | 7466 | continue; |
| 7467 | BUG_ON(refs == 0); | 7467 | BUG_ON(refs == 0); |
| 7468 | 7468 | ||
| 7469 | if (wc->stage == DROP_REFERENCE) { | 7469 | if (wc->stage == DROP_REFERENCE) { |
| 7470 | if (refs == 1) | 7470 | if (refs == 1) |
| 7471 | goto reada; | 7471 | goto reada; |
| 7472 | 7472 | ||
| 7473 | if (wc->level == 1 && | 7473 | if (wc->level == 1 && |
| 7474 | (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) | 7474 | (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) |
| 7475 | continue; | 7475 | continue; |
| 7476 | if (!wc->update_ref || | 7476 | if (!wc->update_ref || |
| 7477 | generation <= root->root_key.offset) | 7477 | generation <= root->root_key.offset) |
| 7478 | continue; | 7478 | continue; |
| 7479 | btrfs_node_key_to_cpu(eb, &key, slot); | 7479 | btrfs_node_key_to_cpu(eb, &key, slot); |
| 7480 | ret = btrfs_comp_cpu_keys(&key, | 7480 | ret = btrfs_comp_cpu_keys(&key, |
| 7481 | &wc->update_progress); | 7481 | &wc->update_progress); |
| 7482 | if (ret < 0) | 7482 | if (ret < 0) |
| 7483 | continue; | 7483 | continue; |
| 7484 | } else { | 7484 | } else { |
| 7485 | if (wc->level == 1 && | 7485 | if (wc->level == 1 && |
| 7486 | (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) | 7486 | (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) |
| 7487 | continue; | 7487 | continue; |
| 7488 | } | 7488 | } |
| 7489 | reada: | 7489 | reada: |
| 7490 | readahead_tree_block(root, bytenr, blocksize); | 7490 | readahead_tree_block(root, bytenr, blocksize); |
| 7491 | nread++; | 7491 | nread++; |
| 7492 | } | 7492 | } |
| 7493 | wc->reada_slot = slot; | 7493 | wc->reada_slot = slot; |
| 7494 | } | 7494 | } |
| 7495 | 7495 | ||
| 7496 | static int account_leaf_items(struct btrfs_trans_handle *trans, | 7496 | static int account_leaf_items(struct btrfs_trans_handle *trans, |
| 7497 | struct btrfs_root *root, | 7497 | struct btrfs_root *root, |
| 7498 | struct extent_buffer *eb) | 7498 | struct extent_buffer *eb) |
| 7499 | { | 7499 | { |
| 7500 | int nr = btrfs_header_nritems(eb); | 7500 | int nr = btrfs_header_nritems(eb); |
| 7501 | int i, extent_type, ret; | 7501 | int i, extent_type, ret; |
| 7502 | struct btrfs_key key; | 7502 | struct btrfs_key key; |
| 7503 | struct btrfs_file_extent_item *fi; | 7503 | struct btrfs_file_extent_item *fi; |
| 7504 | u64 bytenr, num_bytes; | 7504 | u64 bytenr, num_bytes; |
| 7505 | 7505 | ||
| 7506 | for (i = 0; i < nr; i++) { | 7506 | for (i = 0; i < nr; i++) { |
| 7507 | btrfs_item_key_to_cpu(eb, &key, i); | 7507 | btrfs_item_key_to_cpu(eb, &key, i); |
| 7508 | 7508 | ||
| 7509 | if (key.type != BTRFS_EXTENT_DATA_KEY) | 7509 | if (key.type != BTRFS_EXTENT_DATA_KEY) |
| 7510 | continue; | 7510 | continue; |
| 7511 | 7511 | ||
| 7512 | fi = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item); | 7512 | fi = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item); |
| 7513 | /* filter out non qgroup-accountable extents */ | 7513 | /* filter out non qgroup-accountable extents */ |
| 7514 | extent_type = btrfs_file_extent_type(eb, fi); | 7514 | extent_type = btrfs_file_extent_type(eb, fi); |
| 7515 | 7515 | ||
| 7516 | if (extent_type == BTRFS_FILE_EXTENT_INLINE) | 7516 | if (extent_type == BTRFS_FILE_EXTENT_INLINE) |
| 7517 | continue; | 7517 | continue; |
| 7518 | 7518 | ||
| 7519 | bytenr = btrfs_file_extent_disk_bytenr(eb, fi); | 7519 | bytenr = btrfs_file_extent_disk_bytenr(eb, fi); |
| 7520 | if (!bytenr) | 7520 | if (!bytenr) |
| 7521 | continue; | 7521 | continue; |
| 7522 | 7522 | ||
| 7523 | num_bytes = btrfs_file_extent_disk_num_bytes(eb, fi); | 7523 | num_bytes = btrfs_file_extent_disk_num_bytes(eb, fi); |
| 7524 | 7524 | ||
| 7525 | ret = btrfs_qgroup_record_ref(trans, root->fs_info, | 7525 | ret = btrfs_qgroup_record_ref(trans, root->fs_info, |
| 7526 | root->objectid, | 7526 | root->objectid, |
| 7527 | bytenr, num_bytes, | 7527 | bytenr, num_bytes, |
| 7528 | BTRFS_QGROUP_OPER_SUB_SUBTREE, 0); | 7528 | BTRFS_QGROUP_OPER_SUB_SUBTREE, 0); |
| 7529 | if (ret) | 7529 | if (ret) |
| 7530 | return ret; | 7530 | return ret; |
| 7531 | } | 7531 | } |
| 7532 | return 0; | 7532 | return 0; |
| 7533 | } | 7533 | } |
| 7534 | 7534 | ||
| 7535 | /* | 7535 | /* |
| 7536 | * Walk up the tree from the bottom, freeing leaves and any interior | 7536 | * Walk up the tree from the bottom, freeing leaves and any interior |
| 7537 | * nodes which have had all slots visited. If a node (leaf or | 7537 | * nodes which have had all slots visited. If a node (leaf or |
| 7538 | * interior) is freed, the node above it will have it's slot | 7538 | * interior) is freed, the node above it will have it's slot |
| 7539 | * incremented. The root node will never be freed. | 7539 | * incremented. The root node will never be freed. |
| 7540 | * | 7540 | * |
| 7541 | * At the end of this function, we should have a path which has all | 7541 | * At the end of this function, we should have a path which has all |
| 7542 | * slots incremented to the next position for a search. If we need to | 7542 | * slots incremented to the next position for a search. If we need to |
| 7543 | * read a new node it will be NULL and the node above it will have the | 7543 | * read a new node it will be NULL and the node above it will have the |
| 7544 | * correct slot selected for a later read. | 7544 | * correct slot selected for a later read. |
| 7545 | * | 7545 | * |
| 7546 | * If we increment the root nodes slot counter past the number of | 7546 | * If we increment the root nodes slot counter past the number of |
| 7547 | * elements, 1 is returned to signal completion of the search. | 7547 | * elements, 1 is returned to signal completion of the search. |
| 7548 | */ | 7548 | */ |
| 7549 | static int adjust_slots_upwards(struct btrfs_root *root, | 7549 | static int adjust_slots_upwards(struct btrfs_root *root, |
| 7550 | struct btrfs_path *path, int root_level) | 7550 | struct btrfs_path *path, int root_level) |
| 7551 | { | 7551 | { |
| 7552 | int level = 0; | 7552 | int level = 0; |
| 7553 | int nr, slot; | 7553 | int nr, slot; |
| 7554 | struct extent_buffer *eb; | 7554 | struct extent_buffer *eb; |
| 7555 | 7555 | ||
| 7556 | if (root_level == 0) | 7556 | if (root_level == 0) |
| 7557 | return 1; | 7557 | return 1; |
| 7558 | 7558 | ||
| 7559 | while (level <= root_level) { | 7559 | while (level <= root_level) { |
| 7560 | eb = path->nodes[level]; | 7560 | eb = path->nodes[level]; |
| 7561 | nr = btrfs_header_nritems(eb); | 7561 | nr = btrfs_header_nritems(eb); |
| 7562 | path->slots[level]++; | 7562 | path->slots[level]++; |
| 7563 | slot = path->slots[level]; | 7563 | slot = path->slots[level]; |
| 7564 | if (slot >= nr || level == 0) { | 7564 | if (slot >= nr || level == 0) { |
| 7565 | /* | 7565 | /* |
| 7566 | * Don't free the root - we will detect this | 7566 | * Don't free the root - we will detect this |
| 7567 | * condition after our loop and return a | 7567 | * condition after our loop and return a |
| 7568 | * positive value for caller to stop walking the tree. | 7568 | * positive value for caller to stop walking the tree. |
| 7569 | */ | 7569 | */ |
| 7570 | if (level != root_level) { | 7570 | if (level != root_level) { |
| 7571 | btrfs_tree_unlock_rw(eb, path->locks[level]); | 7571 | btrfs_tree_unlock_rw(eb, path->locks[level]); |
| 7572 | path->locks[level] = 0; | 7572 | path->locks[level] = 0; |
| 7573 | 7573 | ||
| 7574 | free_extent_buffer(eb); | 7574 | free_extent_buffer(eb); |
| 7575 | path->nodes[level] = NULL; | 7575 | path->nodes[level] = NULL; |
| 7576 | path->slots[level] = 0; | 7576 | path->slots[level] = 0; |
| 7577 | } | 7577 | } |
| 7578 | } else { | 7578 | } else { |
| 7579 | /* | 7579 | /* |
| 7580 | * We have a valid slot to walk back down | 7580 | * We have a valid slot to walk back down |
| 7581 | * from. Stop here so caller can process these | 7581 | * from. Stop here so caller can process these |
| 7582 | * new nodes. | 7582 | * new nodes. |
| 7583 | */ | 7583 | */ |
| 7584 | break; | 7584 | break; |
| 7585 | } | 7585 | } |
| 7586 | 7586 | ||
| 7587 | level++; | 7587 | level++; |
| 7588 | } | 7588 | } |
| 7589 | 7589 | ||
| 7590 | eb = path->nodes[root_level]; | 7590 | eb = path->nodes[root_level]; |
| 7591 | if (path->slots[root_level] >= btrfs_header_nritems(eb)) | 7591 | if (path->slots[root_level] >= btrfs_header_nritems(eb)) |
| 7592 | return 1; | 7592 | return 1; |
| 7593 | 7593 | ||
| 7594 | return 0; | 7594 | return 0; |
| 7595 | } | 7595 | } |
| 7596 | 7596 | ||
| 7597 | /* | 7597 | /* |
| 7598 | * root_eb is the subtree root and is locked before this function is called. | 7598 | * root_eb is the subtree root and is locked before this function is called. |
| 7599 | */ | 7599 | */ |
| 7600 | static int account_shared_subtree(struct btrfs_trans_handle *trans, | 7600 | static int account_shared_subtree(struct btrfs_trans_handle *trans, |
| 7601 | struct btrfs_root *root, | 7601 | struct btrfs_root *root, |
| 7602 | struct extent_buffer *root_eb, | 7602 | struct extent_buffer *root_eb, |
| 7603 | u64 root_gen, | 7603 | u64 root_gen, |
| 7604 | int root_level) | 7604 | int root_level) |
| 7605 | { | 7605 | { |
| 7606 | int ret = 0; | 7606 | int ret = 0; |
| 7607 | int level; | 7607 | int level; |
| 7608 | struct extent_buffer *eb = root_eb; | 7608 | struct extent_buffer *eb = root_eb; |
| 7609 | struct btrfs_path *path = NULL; | 7609 | struct btrfs_path *path = NULL; |
| 7610 | 7610 | ||
| 7611 | BUG_ON(root_level < 0 || root_level > BTRFS_MAX_LEVEL); | 7611 | BUG_ON(root_level < 0 || root_level > BTRFS_MAX_LEVEL); |
| 7612 | BUG_ON(root_eb == NULL); | 7612 | BUG_ON(root_eb == NULL); |
| 7613 | 7613 | ||
| 7614 | if (!root->fs_info->quota_enabled) | 7614 | if (!root->fs_info->quota_enabled) |
| 7615 | return 0; | 7615 | return 0; |
| 7616 | 7616 | ||
| 7617 | if (!extent_buffer_uptodate(root_eb)) { | 7617 | if (!extent_buffer_uptodate(root_eb)) { |
| 7618 | ret = btrfs_read_buffer(root_eb, root_gen); | 7618 | ret = btrfs_read_buffer(root_eb, root_gen); |
| 7619 | if (ret) | 7619 | if (ret) |
| 7620 | goto out; | 7620 | goto out; |
| 7621 | } | 7621 | } |
| 7622 | 7622 | ||
| 7623 | if (root_level == 0) { | 7623 | if (root_level == 0) { |
| 7624 | ret = account_leaf_items(trans, root, root_eb); | 7624 | ret = account_leaf_items(trans, root, root_eb); |
| 7625 | goto out; | 7625 | goto out; |
| 7626 | } | 7626 | } |
| 7627 | 7627 | ||
| 7628 | path = btrfs_alloc_path(); | 7628 | path = btrfs_alloc_path(); |
| 7629 | if (!path) | 7629 | if (!path) |
| 7630 | return -ENOMEM; | 7630 | return -ENOMEM; |
| 7631 | 7631 | ||
| 7632 | /* | 7632 | /* |
| 7633 | * Walk down the tree. Missing extent blocks are filled in as | 7633 | * Walk down the tree. Missing extent blocks are filled in as |
| 7634 | * we go. Metadata is accounted every time we read a new | 7634 | * we go. Metadata is accounted every time we read a new |
| 7635 | * extent block. | 7635 | * extent block. |
| 7636 | * | 7636 | * |
| 7637 | * When we reach a leaf, we account for file extent items in it, | 7637 | * When we reach a leaf, we account for file extent items in it, |
| 7638 | * walk back up the tree (adjusting slot pointers as we go) | 7638 | * walk back up the tree (adjusting slot pointers as we go) |
| 7639 | * and restart the search process. | 7639 | * and restart the search process. |
| 7640 | */ | 7640 | */ |
| 7641 | extent_buffer_get(root_eb); /* For path */ | 7641 | extent_buffer_get(root_eb); /* For path */ |
| 7642 | path->nodes[root_level] = root_eb; | 7642 | path->nodes[root_level] = root_eb; |
| 7643 | path->slots[root_level] = 0; | 7643 | path->slots[root_level] = 0; |
| 7644 | path->locks[root_level] = 0; /* so release_path doesn't try to unlock */ | 7644 | path->locks[root_level] = 0; /* so release_path doesn't try to unlock */ |
| 7645 | walk_down: | 7645 | walk_down: |
| 7646 | level = root_level; | 7646 | level = root_level; |
| 7647 | while (level >= 0) { | 7647 | while (level >= 0) { |
| 7648 | if (path->nodes[level] == NULL) { | 7648 | if (path->nodes[level] == NULL) { |
| 7649 | int parent_slot; | 7649 | int parent_slot; |
| 7650 | u64 child_gen; | 7650 | u64 child_gen; |
| 7651 | u64 child_bytenr; | 7651 | u64 child_bytenr; |
| 7652 | 7652 | ||
| 7653 | /* We need to get child blockptr/gen from | 7653 | /* We need to get child blockptr/gen from |
| 7654 | * parent before we can read it. */ | 7654 | * parent before we can read it. */ |
| 7655 | eb = path->nodes[level + 1]; | 7655 | eb = path->nodes[level + 1]; |
| 7656 | parent_slot = path->slots[level + 1]; | 7656 | parent_slot = path->slots[level + 1]; |
| 7657 | child_bytenr = btrfs_node_blockptr(eb, parent_slot); | 7657 | child_bytenr = btrfs_node_blockptr(eb, parent_slot); |
| 7658 | child_gen = btrfs_node_ptr_generation(eb, parent_slot); | 7658 | child_gen = btrfs_node_ptr_generation(eb, parent_slot); |
| 7659 | 7659 | ||
| 7660 | eb = read_tree_block(root, child_bytenr, child_gen); | 7660 | eb = read_tree_block(root, child_bytenr, child_gen); |
| 7661 | if (!eb || !extent_buffer_uptodate(eb)) { | 7661 | if (!eb || !extent_buffer_uptodate(eb)) { |
| 7662 | ret = -EIO; | 7662 | ret = -EIO; |
| 7663 | goto out; | 7663 | goto out; |
| 7664 | } | 7664 | } |
| 7665 | 7665 | ||
| 7666 | path->nodes[level] = eb; | 7666 | path->nodes[level] = eb; |
| 7667 | path->slots[level] = 0; | 7667 | path->slots[level] = 0; |
| 7668 | 7668 | ||
| 7669 | btrfs_tree_read_lock(eb); | 7669 | btrfs_tree_read_lock(eb); |
| 7670 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); | 7670 | btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK); |
| 7671 | path->locks[level] = BTRFS_READ_LOCK_BLOCKING; | 7671 | path->locks[level] = BTRFS_READ_LOCK_BLOCKING; |
| 7672 | 7672 | ||
| 7673 | ret = btrfs_qgroup_record_ref(trans, root->fs_info, | 7673 | ret = btrfs_qgroup_record_ref(trans, root->fs_info, |
| 7674 | root->objectid, | 7674 | root->objectid, |
| 7675 | child_bytenr, | 7675 | child_bytenr, |
| 7676 | root->nodesize, | 7676 | root->nodesize, |
| 7677 | BTRFS_QGROUP_OPER_SUB_SUBTREE, | 7677 | BTRFS_QGROUP_OPER_SUB_SUBTREE, |
| 7678 | 0); | 7678 | 0); |
| 7679 | if (ret) | 7679 | if (ret) |
| 7680 | goto out; | 7680 | goto out; |
| 7681 | 7681 | ||
| 7682 | } | 7682 | } |
| 7683 | 7683 | ||
| 7684 | if (level == 0) { | 7684 | if (level == 0) { |
| 7685 | ret = account_leaf_items(trans, root, path->nodes[level]); | 7685 | ret = account_leaf_items(trans, root, path->nodes[level]); |
| 7686 | if (ret) | 7686 | if (ret) |
| 7687 | goto out; | 7687 | goto out; |
| 7688 | 7688 | ||
| 7689 | /* Nonzero return here means we completed our search */ | 7689 | /* Nonzero return here means we completed our search */ |
| 7690 | ret = adjust_slots_upwards(root, path, root_level); | 7690 | ret = adjust_slots_upwards(root, path, root_level); |
| 7691 | if (ret) | 7691 | if (ret) |
| 7692 | break; | 7692 | break; |
| 7693 | 7693 | ||
| 7694 | /* Restart search with new slots */ | 7694 | /* Restart search with new slots */ |
| 7695 | goto walk_down; | 7695 | goto walk_down; |
| 7696 | } | 7696 | } |
| 7697 | 7697 | ||
| 7698 | level--; | 7698 | level--; |
| 7699 | } | 7699 | } |
| 7700 | 7700 | ||
| 7701 | ret = 0; | 7701 | ret = 0; |
| 7702 | out: | 7702 | out: |
| 7703 | btrfs_free_path(path); | 7703 | btrfs_free_path(path); |
| 7704 | 7704 | ||
| 7705 | return ret; | 7705 | return ret; |
| 7706 | } | 7706 | } |
| 7707 | 7707 | ||
| 7708 | /* | 7708 | /* |
| 7709 | * helper to process tree block while walking down the tree. | 7709 | * helper to process tree block while walking down the tree. |
| 7710 | * | 7710 | * |
| 7711 | * when wc->stage == UPDATE_BACKREF, this function updates | 7711 | * when wc->stage == UPDATE_BACKREF, this function updates |
| 7712 | * back refs for pointers in the block. | 7712 | * back refs for pointers in the block. |
| 7713 | * | 7713 | * |
| 7714 | * NOTE: return value 1 means we should stop walking down. | 7714 | * NOTE: return value 1 means we should stop walking down. |
| 7715 | */ | 7715 | */ |
| 7716 | static noinline int walk_down_proc(struct btrfs_trans_handle *trans, | 7716 | static noinline int walk_down_proc(struct btrfs_trans_handle *trans, |
| 7717 | struct btrfs_root *root, | 7717 | struct btrfs_root *root, |
| 7718 | struct btrfs_path *path, | 7718 | struct btrfs_path *path, |
| 7719 | struct walk_control *wc, int lookup_info) | 7719 | struct walk_control *wc, int lookup_info) |
| 7720 | { | 7720 | { |
| 7721 | int level = wc->level; | 7721 | int level = wc->level; |
| 7722 | struct extent_buffer *eb = path->nodes[level]; | 7722 | struct extent_buffer *eb = path->nodes[level]; |
| 7723 | u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF; | 7723 | u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF; |
| 7724 | int ret; | 7724 | int ret; |
| 7725 | 7725 | ||
| 7726 | if (wc->stage == UPDATE_BACKREF && | 7726 | if (wc->stage == UPDATE_BACKREF && |
| 7727 | btrfs_header_owner(eb) != root->root_key.objectid) | 7727 | btrfs_header_owner(eb) != root->root_key.objectid) |
| 7728 | return 1; | 7728 | return 1; |
| 7729 | 7729 | ||
| 7730 | /* | 7730 | /* |
| 7731 | * when reference count of tree block is 1, it won't increase | 7731 | * when reference count of tree block is 1, it won't increase |
| 7732 | * again. once full backref flag is set, we never clear it. | 7732 | * again. once full backref flag is set, we never clear it. |
| 7733 | */ | 7733 | */ |
| 7734 | if (lookup_info && | 7734 | if (lookup_info && |
| 7735 | ((wc->stage == DROP_REFERENCE && wc->refs[level] != 1) || | 7735 | ((wc->stage == DROP_REFERENCE && wc->refs[level] != 1) || |
| 7736 | (wc->stage == UPDATE_BACKREF && !(wc->flags[level] & flag)))) { | 7736 | (wc->stage == UPDATE_BACKREF && !(wc->flags[level] & flag)))) { |
| 7737 | BUG_ON(!path->locks[level]); | 7737 | BUG_ON(!path->locks[level]); |
| 7738 | ret = btrfs_lookup_extent_info(trans, root, | 7738 | ret = btrfs_lookup_extent_info(trans, root, |
| 7739 | eb->start, level, 1, | 7739 | eb->start, level, 1, |
| 7740 | &wc->refs[level], | 7740 | &wc->refs[level], |
| 7741 | &wc->flags[level]); | 7741 | &wc->flags[level]); |
| 7742 | BUG_ON(ret == -ENOMEM); | 7742 | BUG_ON(ret == -ENOMEM); |
| 7743 | if (ret) | 7743 | if (ret) |
| 7744 | return ret; | 7744 | return ret; |
| 7745 | BUG_ON(wc->refs[level] == 0); | 7745 | BUG_ON(wc->refs[level] == 0); |
| 7746 | } | 7746 | } |
| 7747 | 7747 | ||
| 7748 | if (wc->stage == DROP_REFERENCE) { | 7748 | if (wc->stage == DROP_REFERENCE) { |
| 7749 | if (wc->refs[level] > 1) | 7749 | if (wc->refs[level] > 1) |
| 7750 | return 1; | 7750 | return 1; |
| 7751 | 7751 | ||
| 7752 | if (path->locks[level] && !wc->keep_locks) { | 7752 | if (path->locks[level] && !wc->keep_locks) { |
| 7753 | btrfs_tree_unlock_rw(eb, path->locks[level]); | 7753 | btrfs_tree_unlock_rw(eb, path->locks[level]); |
| 7754 | path->locks[level] = 0; | 7754 | path->locks[level] = 0; |
| 7755 | } | 7755 | } |
| 7756 | return 0; | 7756 | return 0; |
| 7757 | } | 7757 | } |
| 7758 | 7758 | ||
| 7759 | /* wc->stage == UPDATE_BACKREF */ | 7759 | /* wc->stage == UPDATE_BACKREF */ |
| 7760 | if (!(wc->flags[level] & flag)) { | 7760 | if (!(wc->flags[level] & flag)) { |
| 7761 | BUG_ON(!path->locks[level]); | 7761 | BUG_ON(!path->locks[level]); |
| 7762 | ret = btrfs_inc_ref(trans, root, eb, 1); | 7762 | ret = btrfs_inc_ref(trans, root, eb, 1); |
| 7763 | BUG_ON(ret); /* -ENOMEM */ | 7763 | BUG_ON(ret); /* -ENOMEM */ |
| 7764 | ret = btrfs_dec_ref(trans, root, eb, 0); | 7764 | ret = btrfs_dec_ref(trans, root, eb, 0); |
| 7765 | BUG_ON(ret); /* -ENOMEM */ | 7765 | BUG_ON(ret); /* -ENOMEM */ |
| 7766 | ret = btrfs_set_disk_extent_flags(trans, root, eb->start, | 7766 | ret = btrfs_set_disk_extent_flags(trans, root, eb->start, |
| 7767 | eb->len, flag, | 7767 | eb->len, flag, |
| 7768 | btrfs_header_level(eb), 0); | 7768 | btrfs_header_level(eb), 0); |
| 7769 | BUG_ON(ret); /* -ENOMEM */ | 7769 | BUG_ON(ret); /* -ENOMEM */ |
| 7770 | wc->flags[level] |= flag; | 7770 | wc->flags[level] |= flag; |
| 7771 | } | 7771 | } |
| 7772 | 7772 | ||
| 7773 | /* | 7773 | /* |
| 7774 | * the block is shared by multiple trees, so it's not good to | 7774 | * the block is shared by multiple trees, so it's not good to |
| 7775 | * keep the tree lock | 7775 | * keep the tree lock |
| 7776 | */ | 7776 | */ |
| 7777 | if (path->locks[level] && level > 0) { | 7777 | if (path->locks[level] && level > 0) { |
| 7778 | btrfs_tree_unlock_rw(eb, path->locks[level]); | 7778 | btrfs_tree_unlock_rw(eb, path->locks[level]); |
| 7779 | path->locks[level] = 0; | 7779 | path->locks[level] = 0; |
| 7780 | } | 7780 | } |
| 7781 | return 0; | 7781 | return 0; |
| 7782 | } | 7782 | } |
| 7783 | 7783 | ||
| 7784 | /* | 7784 | /* |
| 7785 | * helper to process tree block pointer. | 7785 | * helper to process tree block pointer. |
| 7786 | * | 7786 | * |
| 7787 | * when wc->stage == DROP_REFERENCE, this function checks | 7787 | * when wc->stage == DROP_REFERENCE, this function checks |
| 7788 | * reference count of the block pointed to. if the block | 7788 | * reference count of the block pointed to. if the block |
| 7789 | * is shared and we need update back refs for the subtree | 7789 | * is shared and we need update back refs for the subtree |
| 7790 | * rooted at the block, this function changes wc->stage to | 7790 | * rooted at the block, this function changes wc->stage to |
| 7791 | * UPDATE_BACKREF. if the block is shared and there is no | 7791 | * UPDATE_BACKREF. if the block is shared and there is no |
| 7792 | * need to update back, this function drops the reference | 7792 | * need to update back, this function drops the reference |
| 7793 | * to the block. | 7793 | * to the block. |
| 7794 | * | 7794 | * |
| 7795 | * NOTE: return value 1 means we should stop walking down. | 7795 | * NOTE: return value 1 means we should stop walking down. |
| 7796 | */ | 7796 | */ |
| 7797 | static noinline int do_walk_down(struct btrfs_trans_handle *trans, | 7797 | static noinline int do_walk_down(struct btrfs_trans_handle *trans, |
| 7798 | struct btrfs_root *root, | 7798 | struct btrfs_root *root, |
| 7799 | struct btrfs_path *path, | 7799 | struct btrfs_path *path, |
| 7800 | struct walk_control *wc, int *lookup_info) | 7800 | struct walk_control *wc, int *lookup_info) |
| 7801 | { | 7801 | { |
| 7802 | u64 bytenr; | 7802 | u64 bytenr; |
| 7803 | u64 generation; | 7803 | u64 generation; |
| 7804 | u64 parent; | 7804 | u64 parent; |
| 7805 | u32 blocksize; | 7805 | u32 blocksize; |
| 7806 | struct btrfs_key key; | 7806 | struct btrfs_key key; |
| 7807 | struct extent_buffer *next; | 7807 | struct extent_buffer *next; |
| 7808 | int level = wc->level; | 7808 | int level = wc->level; |
| 7809 | int reada = 0; | 7809 | int reada = 0; |
| 7810 | int ret = 0; | 7810 | int ret = 0; |
| 7811 | bool need_account = false; | 7811 | bool need_account = false; |
| 7812 | 7812 | ||
| 7813 | generation = btrfs_node_ptr_generation(path->nodes[level], | 7813 | generation = btrfs_node_ptr_generation(path->nodes[level], |
| 7814 | path->slots[level]); | 7814 | path->slots[level]); |
| 7815 | /* | 7815 | /* |
| 7816 | * if the lower level block was created before the snapshot | 7816 | * if the lower level block was created before the snapshot |
| 7817 | * was created, we know there is no need to update back refs | 7817 | * was created, we know there is no need to update back refs |
| 7818 | * for the subtree | 7818 | * for the subtree |
| 7819 | */ | 7819 | */ |
| 7820 | if (wc->stage == UPDATE_BACKREF && | 7820 | if (wc->stage == UPDATE_BACKREF && |
| 7821 | generation <= root->root_key.offset) { | 7821 | generation <= root->root_key.offset) { |
| 7822 | *lookup_info = 1; | 7822 | *lookup_info = 1; |
| 7823 | return 1; | 7823 | return 1; |
| 7824 | } | 7824 | } |
| 7825 | 7825 | ||
| 7826 | bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]); | 7826 | bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]); |
| 7827 | blocksize = root->nodesize; | 7827 | blocksize = root->nodesize; |
| 7828 | 7828 | ||
| 7829 | next = btrfs_find_tree_block(root, bytenr); | 7829 | next = btrfs_find_tree_block(root, bytenr); |
| 7830 | if (!next) { | 7830 | if (!next) { |
| 7831 | next = btrfs_find_create_tree_block(root, bytenr, blocksize); | 7831 | next = btrfs_find_create_tree_block(root, bytenr, blocksize); |
| 7832 | if (!next) | 7832 | if (!next) |
| 7833 | return -ENOMEM; | 7833 | return -ENOMEM; |
| 7834 | btrfs_set_buffer_lockdep_class(root->root_key.objectid, next, | 7834 | btrfs_set_buffer_lockdep_class(root->root_key.objectid, next, |
| 7835 | level - 1); | 7835 | level - 1); |
| 7836 | reada = 1; | 7836 | reada = 1; |
| 7837 | } | 7837 | } |
| 7838 | btrfs_tree_lock(next); | 7838 | btrfs_tree_lock(next); |
| 7839 | btrfs_set_lock_blocking(next); | 7839 | btrfs_set_lock_blocking(next); |
| 7840 | 7840 | ||
| 7841 | ret = btrfs_lookup_extent_info(trans, root, bytenr, level - 1, 1, | 7841 | ret = btrfs_lookup_extent_info(trans, root, bytenr, level - 1, 1, |
| 7842 | &wc->refs[level - 1], | 7842 | &wc->refs[level - 1], |
| 7843 | &wc->flags[level - 1]); | 7843 | &wc->flags[level - 1]); |
| 7844 | if (ret < 0) { | 7844 | if (ret < 0) { |
| 7845 | btrfs_tree_unlock(next); | 7845 | btrfs_tree_unlock(next); |
| 7846 | return ret; | 7846 | return ret; |
| 7847 | } | 7847 | } |
| 7848 | 7848 | ||
| 7849 | if (unlikely(wc->refs[level - 1] == 0)) { | 7849 | if (unlikely(wc->refs[level - 1] == 0)) { |
| 7850 | btrfs_err(root->fs_info, "Missing references."); | 7850 | btrfs_err(root->fs_info, "Missing references."); |
| 7851 | BUG(); | 7851 | BUG(); |
| 7852 | } | 7852 | } |
| 7853 | *lookup_info = 0; | 7853 | *lookup_info = 0; |
| 7854 | 7854 | ||
| 7855 | if (wc->stage == DROP_REFERENCE) { | 7855 | if (wc->stage == DROP_REFERENCE) { |
| 7856 | if (wc->refs[level - 1] > 1) { | 7856 | if (wc->refs[level - 1] > 1) { |
| 7857 | need_account = true; | 7857 | need_account = true; |
| 7858 | if (level == 1 && | 7858 | if (level == 1 && |
| 7859 | (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF)) | 7859 | (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF)) |
| 7860 | goto skip; | 7860 | goto skip; |
| 7861 | 7861 | ||
| 7862 | if (!wc->update_ref || | 7862 | if (!wc->update_ref || |
| 7863 | generation <= root->root_key.offset) | 7863 | generation <= root->root_key.offset) |
| 7864 | goto skip; | 7864 | goto skip; |
| 7865 | 7865 | ||
| 7866 | btrfs_node_key_to_cpu(path->nodes[level], &key, | 7866 | btrfs_node_key_to_cpu(path->nodes[level], &key, |
| 7867 | path->slots[level]); | 7867 | path->slots[level]); |
| 7868 | ret = btrfs_comp_cpu_keys(&key, &wc->update_progress); | 7868 | ret = btrfs_comp_cpu_keys(&key, &wc->update_progress); |
| 7869 | if (ret < 0) | 7869 | if (ret < 0) |
| 7870 | goto skip; | 7870 | goto skip; |
| 7871 | 7871 | ||
| 7872 | wc->stage = UPDATE_BACKREF; | 7872 | wc->stage = UPDATE_BACKREF; |
| 7873 | wc->shared_level = level - 1; | 7873 | wc->shared_level = level - 1; |
| 7874 | } | 7874 | } |
| 7875 | } else { | 7875 | } else { |
| 7876 | if (level == 1 && | 7876 | if (level == 1 && |
| 7877 | (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF)) | 7877 | (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF)) |
| 7878 | goto skip; | 7878 | goto skip; |
| 7879 | } | 7879 | } |
| 7880 | 7880 | ||
| 7881 | if (!btrfs_buffer_uptodate(next, generation, 0)) { | 7881 | if (!btrfs_buffer_uptodate(next, generation, 0)) { |
| 7882 | btrfs_tree_unlock(next); | 7882 | btrfs_tree_unlock(next); |
| 7883 | free_extent_buffer(next); | 7883 | free_extent_buffer(next); |
| 7884 | next = NULL; | 7884 | next = NULL; |
| 7885 | *lookup_info = 1; | 7885 | *lookup_info = 1; |
| 7886 | } | 7886 | } |
| 7887 | 7887 | ||
| 7888 | if (!next) { | 7888 | if (!next) { |
| 7889 | if (reada && level == 1) | 7889 | if (reada && level == 1) |
| 7890 | reada_walk_down(trans, root, wc, path); | 7890 | reada_walk_down(trans, root, wc, path); |
| 7891 | next = read_tree_block(root, bytenr, generation); | 7891 | next = read_tree_block(root, bytenr, generation); |
| 7892 | if (!next || !extent_buffer_uptodate(next)) { | 7892 | if (!next || !extent_buffer_uptodate(next)) { |
| 7893 | free_extent_buffer(next); | 7893 | free_extent_buffer(next); |
| 7894 | return -EIO; | 7894 | return -EIO; |
| 7895 | } | 7895 | } |
| 7896 | btrfs_tree_lock(next); | 7896 | btrfs_tree_lock(next); |
| 7897 | btrfs_set_lock_blocking(next); | 7897 | btrfs_set_lock_blocking(next); |
| 7898 | } | 7898 | } |
| 7899 | 7899 | ||
| 7900 | level--; | 7900 | level--; |
| 7901 | BUG_ON(level != btrfs_header_level(next)); | 7901 | BUG_ON(level != btrfs_header_level(next)); |
| 7902 | path->nodes[level] = next; | 7902 | path->nodes[level] = next; |
| 7903 | path->slots[level] = 0; | 7903 | path->slots[level] = 0; |
| 7904 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; | 7904 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; |
| 7905 | wc->level = level; | 7905 | wc->level = level; |
| 7906 | if (wc->level == 1) | 7906 | if (wc->level == 1) |
| 7907 | wc->reada_slot = 0; | 7907 | wc->reada_slot = 0; |
| 7908 | return 0; | 7908 | return 0; |
| 7909 | skip: | 7909 | skip: |
| 7910 | wc->refs[level - 1] = 0; | 7910 | wc->refs[level - 1] = 0; |
| 7911 | wc->flags[level - 1] = 0; | 7911 | wc->flags[level - 1] = 0; |
| 7912 | if (wc->stage == DROP_REFERENCE) { | 7912 | if (wc->stage == DROP_REFERENCE) { |
| 7913 | if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) { | 7913 | if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) { |
| 7914 | parent = path->nodes[level]->start; | 7914 | parent = path->nodes[level]->start; |
| 7915 | } else { | 7915 | } else { |
| 7916 | BUG_ON(root->root_key.objectid != | 7916 | BUG_ON(root->root_key.objectid != |
| 7917 | btrfs_header_owner(path->nodes[level])); | 7917 | btrfs_header_owner(path->nodes[level])); |
| 7918 | parent = 0; | 7918 | parent = 0; |
| 7919 | } | 7919 | } |
| 7920 | 7920 | ||
| 7921 | if (need_account) { | 7921 | if (need_account) { |
| 7922 | ret = account_shared_subtree(trans, root, next, | 7922 | ret = account_shared_subtree(trans, root, next, |
| 7923 | generation, level - 1); | 7923 | generation, level - 1); |
| 7924 | if (ret) { | 7924 | if (ret) { |
| 7925 | printk_ratelimited(KERN_ERR "BTRFS: %s Error " | 7925 | printk_ratelimited(KERN_ERR "BTRFS: %s Error " |
| 7926 | "%d accounting shared subtree. Quota " | 7926 | "%d accounting shared subtree. Quota " |
| 7927 | "is out of sync, rescan required.\n", | 7927 | "is out of sync, rescan required.\n", |
| 7928 | root->fs_info->sb->s_id, ret); | 7928 | root->fs_info->sb->s_id, ret); |
| 7929 | } | 7929 | } |
| 7930 | } | 7930 | } |
| 7931 | ret = btrfs_free_extent(trans, root, bytenr, blocksize, parent, | 7931 | ret = btrfs_free_extent(trans, root, bytenr, blocksize, parent, |
| 7932 | root->root_key.objectid, level - 1, 0, 0); | 7932 | root->root_key.objectid, level - 1, 0, 0); |
| 7933 | BUG_ON(ret); /* -ENOMEM */ | 7933 | BUG_ON(ret); /* -ENOMEM */ |
| 7934 | } | 7934 | } |
| 7935 | btrfs_tree_unlock(next); | 7935 | btrfs_tree_unlock(next); |
| 7936 | free_extent_buffer(next); | 7936 | free_extent_buffer(next); |
| 7937 | *lookup_info = 1; | 7937 | *lookup_info = 1; |
| 7938 | return 1; | 7938 | return 1; |
| 7939 | } | 7939 | } |
| 7940 | 7940 | ||
| 7941 | /* | 7941 | /* |
| 7942 | * helper to process tree block while walking up the tree. | 7942 | * helper to process tree block while walking up the tree. |
| 7943 | * | 7943 | * |
| 7944 | * when wc->stage == DROP_REFERENCE, this function drops | 7944 | * when wc->stage == DROP_REFERENCE, this function drops |
| 7945 | * reference count on the block. | 7945 | * reference count on the block. |
| 7946 | * | 7946 | * |
| 7947 | * when wc->stage == UPDATE_BACKREF, this function changes | 7947 | * when wc->stage == UPDATE_BACKREF, this function changes |
| 7948 | * wc->stage back to DROP_REFERENCE if we changed wc->stage | 7948 | * wc->stage back to DROP_REFERENCE if we changed wc->stage |
| 7949 | * to UPDATE_BACKREF previously while processing the block. | 7949 | * to UPDATE_BACKREF previously while processing the block. |
| 7950 | * | 7950 | * |
| 7951 | * NOTE: return value 1 means we should stop walking up. | 7951 | * NOTE: return value 1 means we should stop walking up. |
| 7952 | */ | 7952 | */ |
| 7953 | static noinline int walk_up_proc(struct btrfs_trans_handle *trans, | 7953 | static noinline int walk_up_proc(struct btrfs_trans_handle *trans, |
| 7954 | struct btrfs_root *root, | 7954 | struct btrfs_root *root, |
| 7955 | struct btrfs_path *path, | 7955 | struct btrfs_path *path, |
| 7956 | struct walk_control *wc) | 7956 | struct walk_control *wc) |
| 7957 | { | 7957 | { |
| 7958 | int ret; | 7958 | int ret; |
| 7959 | int level = wc->level; | 7959 | int level = wc->level; |
| 7960 | struct extent_buffer *eb = path->nodes[level]; | 7960 | struct extent_buffer *eb = path->nodes[level]; |
| 7961 | u64 parent = 0; | 7961 | u64 parent = 0; |
| 7962 | 7962 | ||
| 7963 | if (wc->stage == UPDATE_BACKREF) { | 7963 | if (wc->stage == UPDATE_BACKREF) { |
| 7964 | BUG_ON(wc->shared_level < level); | 7964 | BUG_ON(wc->shared_level < level); |
| 7965 | if (level < wc->shared_level) | 7965 | if (level < wc->shared_level) |
| 7966 | goto out; | 7966 | goto out; |
| 7967 | 7967 | ||
| 7968 | ret = find_next_key(path, level + 1, &wc->update_progress); | 7968 | ret = find_next_key(path, level + 1, &wc->update_progress); |
| 7969 | if (ret > 0) | 7969 | if (ret > 0) |
| 7970 | wc->update_ref = 0; | 7970 | wc->update_ref = 0; |
| 7971 | 7971 | ||
| 7972 | wc->stage = DROP_REFERENCE; | 7972 | wc->stage = DROP_REFERENCE; |
| 7973 | wc->shared_level = -1; | 7973 | wc->shared_level = -1; |
| 7974 | path->slots[level] = 0; | 7974 | path->slots[level] = 0; |
| 7975 | 7975 | ||
| 7976 | /* | 7976 | /* |
| 7977 | * check reference count again if the block isn't locked. | 7977 | * check reference count again if the block isn't locked. |
| 7978 | * we should start walking down the tree again if reference | 7978 | * we should start walking down the tree again if reference |
| 7979 | * count is one. | 7979 | * count is one. |
| 7980 | */ | 7980 | */ |
| 7981 | if (!path->locks[level]) { | 7981 | if (!path->locks[level]) { |
| 7982 | BUG_ON(level == 0); | 7982 | BUG_ON(level == 0); |
| 7983 | btrfs_tree_lock(eb); | 7983 | btrfs_tree_lock(eb); |
| 7984 | btrfs_set_lock_blocking(eb); | 7984 | btrfs_set_lock_blocking(eb); |
| 7985 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; | 7985 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; |
| 7986 | 7986 | ||
| 7987 | ret = btrfs_lookup_extent_info(trans, root, | 7987 | ret = btrfs_lookup_extent_info(trans, root, |
| 7988 | eb->start, level, 1, | 7988 | eb->start, level, 1, |
| 7989 | &wc->refs[level], | 7989 | &wc->refs[level], |
| 7990 | &wc->flags[level]); | 7990 | &wc->flags[level]); |
| 7991 | if (ret < 0) { | 7991 | if (ret < 0) { |
| 7992 | btrfs_tree_unlock_rw(eb, path->locks[level]); | 7992 | btrfs_tree_unlock_rw(eb, path->locks[level]); |
| 7993 | path->locks[level] = 0; | 7993 | path->locks[level] = 0; |
| 7994 | return ret; | 7994 | return ret; |
| 7995 | } | 7995 | } |
| 7996 | BUG_ON(wc->refs[level] == 0); | 7996 | BUG_ON(wc->refs[level] == 0); |
| 7997 | if (wc->refs[level] == 1) { | 7997 | if (wc->refs[level] == 1) { |
| 7998 | btrfs_tree_unlock_rw(eb, path->locks[level]); | 7998 | btrfs_tree_unlock_rw(eb, path->locks[level]); |
| 7999 | path->locks[level] = 0; | 7999 | path->locks[level] = 0; |
| 8000 | return 1; | 8000 | return 1; |
| 8001 | } | 8001 | } |
| 8002 | } | 8002 | } |
| 8003 | } | 8003 | } |
| 8004 | 8004 | ||
| 8005 | /* wc->stage == DROP_REFERENCE */ | 8005 | /* wc->stage == DROP_REFERENCE */ |
| 8006 | BUG_ON(wc->refs[level] > 1 && !path->locks[level]); | 8006 | BUG_ON(wc->refs[level] > 1 && !path->locks[level]); |
| 8007 | 8007 | ||
| 8008 | if (wc->refs[level] == 1) { | 8008 | if (wc->refs[level] == 1) { |
| 8009 | if (level == 0) { | 8009 | if (level == 0) { |
| 8010 | if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) | 8010 | if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) |
| 8011 | ret = btrfs_dec_ref(trans, root, eb, 1); | 8011 | ret = btrfs_dec_ref(trans, root, eb, 1); |
| 8012 | else | 8012 | else |
| 8013 | ret = btrfs_dec_ref(trans, root, eb, 0); | 8013 | ret = btrfs_dec_ref(trans, root, eb, 0); |
| 8014 | BUG_ON(ret); /* -ENOMEM */ | 8014 | BUG_ON(ret); /* -ENOMEM */ |
| 8015 | ret = account_leaf_items(trans, root, eb); | 8015 | ret = account_leaf_items(trans, root, eb); |
| 8016 | if (ret) { | 8016 | if (ret) { |
| 8017 | printk_ratelimited(KERN_ERR "BTRFS: %s Error " | 8017 | printk_ratelimited(KERN_ERR "BTRFS: %s Error " |
| 8018 | "%d accounting leaf items. Quota " | 8018 | "%d accounting leaf items. Quota " |
| 8019 | "is out of sync, rescan required.\n", | 8019 | "is out of sync, rescan required.\n", |
| 8020 | root->fs_info->sb->s_id, ret); | 8020 | root->fs_info->sb->s_id, ret); |
| 8021 | } | 8021 | } |
| 8022 | } | 8022 | } |
| 8023 | /* make block locked assertion in clean_tree_block happy */ | 8023 | /* make block locked assertion in clean_tree_block happy */ |
| 8024 | if (!path->locks[level] && | 8024 | if (!path->locks[level] && |
| 8025 | btrfs_header_generation(eb) == trans->transid) { | 8025 | btrfs_header_generation(eb) == trans->transid) { |
| 8026 | btrfs_tree_lock(eb); | 8026 | btrfs_tree_lock(eb); |
| 8027 | btrfs_set_lock_blocking(eb); | 8027 | btrfs_set_lock_blocking(eb); |
| 8028 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; | 8028 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; |
| 8029 | } | 8029 | } |
| 8030 | clean_tree_block(trans, root, eb); | 8030 | clean_tree_block(trans, root, eb); |
| 8031 | } | 8031 | } |
| 8032 | 8032 | ||
| 8033 | if (eb == root->node) { | 8033 | if (eb == root->node) { |
| 8034 | if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) | 8034 | if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) |
| 8035 | parent = eb->start; | 8035 | parent = eb->start; |
| 8036 | else | 8036 | else |
| 8037 | BUG_ON(root->root_key.objectid != | 8037 | BUG_ON(root->root_key.objectid != |
| 8038 | btrfs_header_owner(eb)); | 8038 | btrfs_header_owner(eb)); |
| 8039 | } else { | 8039 | } else { |
| 8040 | if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF) | 8040 | if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF) |
| 8041 | parent = path->nodes[level + 1]->start; | 8041 | parent = path->nodes[level + 1]->start; |
| 8042 | else | 8042 | else |
| 8043 | BUG_ON(root->root_key.objectid != | 8043 | BUG_ON(root->root_key.objectid != |
| 8044 | btrfs_header_owner(path->nodes[level + 1])); | 8044 | btrfs_header_owner(path->nodes[level + 1])); |
| 8045 | } | 8045 | } |
| 8046 | 8046 | ||
| 8047 | btrfs_free_tree_block(trans, root, eb, parent, wc->refs[level] == 1); | 8047 | btrfs_free_tree_block(trans, root, eb, parent, wc->refs[level] == 1); |
| 8048 | out: | 8048 | out: |
| 8049 | wc->refs[level] = 0; | 8049 | wc->refs[level] = 0; |
| 8050 | wc->flags[level] = 0; | 8050 | wc->flags[level] = 0; |
| 8051 | return 0; | 8051 | return 0; |
| 8052 | } | 8052 | } |
| 8053 | 8053 | ||
| 8054 | static noinline int walk_down_tree(struct btrfs_trans_handle *trans, | 8054 | static noinline int walk_down_tree(struct btrfs_trans_handle *trans, |
| 8055 | struct btrfs_root *root, | 8055 | struct btrfs_root *root, |
| 8056 | struct btrfs_path *path, | 8056 | struct btrfs_path *path, |
| 8057 | struct walk_control *wc) | 8057 | struct walk_control *wc) |
| 8058 | { | 8058 | { |
| 8059 | int level = wc->level; | 8059 | int level = wc->level; |
| 8060 | int lookup_info = 1; | 8060 | int lookup_info = 1; |
| 8061 | int ret; | 8061 | int ret; |
| 8062 | 8062 | ||
| 8063 | while (level >= 0) { | 8063 | while (level >= 0) { |
| 8064 | ret = walk_down_proc(trans, root, path, wc, lookup_info); | 8064 | ret = walk_down_proc(trans, root, path, wc, lookup_info); |
| 8065 | if (ret > 0) | 8065 | if (ret > 0) |
| 8066 | break; | 8066 | break; |
| 8067 | 8067 | ||
| 8068 | if (level == 0) | 8068 | if (level == 0) |
| 8069 | break; | 8069 | break; |
| 8070 | 8070 | ||
| 8071 | if (path->slots[level] >= | 8071 | if (path->slots[level] >= |
| 8072 | btrfs_header_nritems(path->nodes[level])) | 8072 | btrfs_header_nritems(path->nodes[level])) |
| 8073 | break; | 8073 | break; |
| 8074 | 8074 | ||
| 8075 | ret = do_walk_down(trans, root, path, wc, &lookup_info); | 8075 | ret = do_walk_down(trans, root, path, wc, &lookup_info); |
| 8076 | if (ret > 0) { | 8076 | if (ret > 0) { |
| 8077 | path->slots[level]++; | 8077 | path->slots[level]++; |
| 8078 | continue; | 8078 | continue; |
| 8079 | } else if (ret < 0) | 8079 | } else if (ret < 0) |
| 8080 | return ret; | 8080 | return ret; |
| 8081 | level = wc->level; | 8081 | level = wc->level; |
| 8082 | } | 8082 | } |
| 8083 | return 0; | 8083 | return 0; |
| 8084 | } | 8084 | } |
| 8085 | 8085 | ||
| 8086 | static noinline int walk_up_tree(struct btrfs_trans_handle *trans, | 8086 | static noinline int walk_up_tree(struct btrfs_trans_handle *trans, |
| 8087 | struct btrfs_root *root, | 8087 | struct btrfs_root *root, |
| 8088 | struct btrfs_path *path, | 8088 | struct btrfs_path *path, |
| 8089 | struct walk_control *wc, int max_level) | 8089 | struct walk_control *wc, int max_level) |
| 8090 | { | 8090 | { |
| 8091 | int level = wc->level; | 8091 | int level = wc->level; |
| 8092 | int ret; | 8092 | int ret; |
| 8093 | 8093 | ||
| 8094 | path->slots[level] = btrfs_header_nritems(path->nodes[level]); | 8094 | path->slots[level] = btrfs_header_nritems(path->nodes[level]); |
| 8095 | while (level < max_level && path->nodes[level]) { | 8095 | while (level < max_level && path->nodes[level]) { |
| 8096 | wc->level = level; | 8096 | wc->level = level; |
| 8097 | if (path->slots[level] + 1 < | 8097 | if (path->slots[level] + 1 < |
| 8098 | btrfs_header_nritems(path->nodes[level])) { | 8098 | btrfs_header_nritems(path->nodes[level])) { |
| 8099 | path->slots[level]++; | 8099 | path->slots[level]++; |
| 8100 | return 0; | 8100 | return 0; |
| 8101 | } else { | 8101 | } else { |
| 8102 | ret = walk_up_proc(trans, root, path, wc); | 8102 | ret = walk_up_proc(trans, root, path, wc); |
| 8103 | if (ret > 0) | 8103 | if (ret > 0) |
| 8104 | return 0; | 8104 | return 0; |
| 8105 | 8105 | ||
| 8106 | if (path->locks[level]) { | 8106 | if (path->locks[level]) { |
| 8107 | btrfs_tree_unlock_rw(path->nodes[level], | 8107 | btrfs_tree_unlock_rw(path->nodes[level], |
| 8108 | path->locks[level]); | 8108 | path->locks[level]); |
| 8109 | path->locks[level] = 0; | 8109 | path->locks[level] = 0; |
| 8110 | } | 8110 | } |
| 8111 | free_extent_buffer(path->nodes[level]); | 8111 | free_extent_buffer(path->nodes[level]); |
| 8112 | path->nodes[level] = NULL; | 8112 | path->nodes[level] = NULL; |
| 8113 | level++; | 8113 | level++; |
| 8114 | } | 8114 | } |
| 8115 | } | 8115 | } |
| 8116 | return 1; | 8116 | return 1; |
| 8117 | } | 8117 | } |
| 8118 | 8118 | ||
| 8119 | /* | 8119 | /* |
| 8120 | * drop a subvolume tree. | 8120 | * drop a subvolume tree. |
| 8121 | * | 8121 | * |
| 8122 | * this function traverses the tree freeing any blocks that only | 8122 | * this function traverses the tree freeing any blocks that only |
| 8123 | * referenced by the tree. | 8123 | * referenced by the tree. |
| 8124 | * | 8124 | * |
| 8125 | * when a shared tree block is found. this function decreases its | 8125 | * when a shared tree block is found. this function decreases its |
| 8126 | * reference count by one. if update_ref is true, this function | 8126 | * reference count by one. if update_ref is true, this function |
| 8127 | * also make sure backrefs for the shared block and all lower level | 8127 | * also make sure backrefs for the shared block and all lower level |
| 8128 | * blocks are properly updated. | 8128 | * blocks are properly updated. |
| 8129 | * | 8129 | * |
| 8130 | * If called with for_reloc == 0, may exit early with -EAGAIN | 8130 | * If called with for_reloc == 0, may exit early with -EAGAIN |
| 8131 | */ | 8131 | */ |
| 8132 | int btrfs_drop_snapshot(struct btrfs_root *root, | 8132 | int btrfs_drop_snapshot(struct btrfs_root *root, |
| 8133 | struct btrfs_block_rsv *block_rsv, int update_ref, | 8133 | struct btrfs_block_rsv *block_rsv, int update_ref, |
| 8134 | int for_reloc) | 8134 | int for_reloc) |
| 8135 | { | 8135 | { |
| 8136 | struct btrfs_path *path; | 8136 | struct btrfs_path *path; |
| 8137 | struct btrfs_trans_handle *trans; | 8137 | struct btrfs_trans_handle *trans; |
| 8138 | struct btrfs_root *tree_root = root->fs_info->tree_root; | 8138 | struct btrfs_root *tree_root = root->fs_info->tree_root; |
| 8139 | struct btrfs_root_item *root_item = &root->root_item; | 8139 | struct btrfs_root_item *root_item = &root->root_item; |
| 8140 | struct walk_control *wc; | 8140 | struct walk_control *wc; |
| 8141 | struct btrfs_key key; | 8141 | struct btrfs_key key; |
| 8142 | int err = 0; | 8142 | int err = 0; |
| 8143 | int ret; | 8143 | int ret; |
| 8144 | int level; | 8144 | int level; |
| 8145 | bool root_dropped = false; | 8145 | bool root_dropped = false; |
| 8146 | 8146 | ||
| 8147 | btrfs_debug(root->fs_info, "Drop subvolume %llu", root->objectid); | 8147 | btrfs_debug(root->fs_info, "Drop subvolume %llu", root->objectid); |
| 8148 | 8148 | ||
| 8149 | path = btrfs_alloc_path(); | 8149 | path = btrfs_alloc_path(); |
| 8150 | if (!path) { | 8150 | if (!path) { |
| 8151 | err = -ENOMEM; | 8151 | err = -ENOMEM; |
| 8152 | goto out; | 8152 | goto out; |
| 8153 | } | 8153 | } |
| 8154 | 8154 | ||
| 8155 | wc = kzalloc(sizeof(*wc), GFP_NOFS); | 8155 | wc = kzalloc(sizeof(*wc), GFP_NOFS); |
| 8156 | if (!wc) { | 8156 | if (!wc) { |
| 8157 | btrfs_free_path(path); | 8157 | btrfs_free_path(path); |
| 8158 | err = -ENOMEM; | 8158 | err = -ENOMEM; |
| 8159 | goto out; | 8159 | goto out; |
| 8160 | } | 8160 | } |
| 8161 | 8161 | ||
| 8162 | trans = btrfs_start_transaction(tree_root, 0); | 8162 | trans = btrfs_start_transaction(tree_root, 0); |
| 8163 | if (IS_ERR(trans)) { | 8163 | if (IS_ERR(trans)) { |
| 8164 | err = PTR_ERR(trans); | 8164 | err = PTR_ERR(trans); |
| 8165 | goto out_free; | 8165 | goto out_free; |
| 8166 | } | 8166 | } |
| 8167 | 8167 | ||
| 8168 | if (block_rsv) | 8168 | if (block_rsv) |
| 8169 | trans->block_rsv = block_rsv; | 8169 | trans->block_rsv = block_rsv; |
| 8170 | 8170 | ||
| 8171 | if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) { | 8171 | if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) { |
| 8172 | level = btrfs_header_level(root->node); | 8172 | level = btrfs_header_level(root->node); |
| 8173 | path->nodes[level] = btrfs_lock_root_node(root); | 8173 | path->nodes[level] = btrfs_lock_root_node(root); |
| 8174 | btrfs_set_lock_blocking(path->nodes[level]); | 8174 | btrfs_set_lock_blocking(path->nodes[level]); |
| 8175 | path->slots[level] = 0; | 8175 | path->slots[level] = 0; |
| 8176 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; | 8176 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; |
| 8177 | memset(&wc->update_progress, 0, | 8177 | memset(&wc->update_progress, 0, |
| 8178 | sizeof(wc->update_progress)); | 8178 | sizeof(wc->update_progress)); |
| 8179 | } else { | 8179 | } else { |
| 8180 | btrfs_disk_key_to_cpu(&key, &root_item->drop_progress); | 8180 | btrfs_disk_key_to_cpu(&key, &root_item->drop_progress); |
| 8181 | memcpy(&wc->update_progress, &key, | 8181 | memcpy(&wc->update_progress, &key, |
| 8182 | sizeof(wc->update_progress)); | 8182 | sizeof(wc->update_progress)); |
| 8183 | 8183 | ||
| 8184 | level = root_item->drop_level; | 8184 | level = root_item->drop_level; |
| 8185 | BUG_ON(level == 0); | 8185 | BUG_ON(level == 0); |
| 8186 | path->lowest_level = level; | 8186 | path->lowest_level = level; |
| 8187 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 8187 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 8188 | path->lowest_level = 0; | 8188 | path->lowest_level = 0; |
| 8189 | if (ret < 0) { | 8189 | if (ret < 0) { |
| 8190 | err = ret; | 8190 | err = ret; |
| 8191 | goto out_end_trans; | 8191 | goto out_end_trans; |
| 8192 | } | 8192 | } |
| 8193 | WARN_ON(ret > 0); | 8193 | WARN_ON(ret > 0); |
| 8194 | 8194 | ||
| 8195 | /* | 8195 | /* |
| 8196 | * unlock our path, this is safe because only this | 8196 | * unlock our path, this is safe because only this |
| 8197 | * function is allowed to delete this snapshot | 8197 | * function is allowed to delete this snapshot |
| 8198 | */ | 8198 | */ |
| 8199 | btrfs_unlock_up_safe(path, 0); | 8199 | btrfs_unlock_up_safe(path, 0); |
| 8200 | 8200 | ||
| 8201 | level = btrfs_header_level(root->node); | 8201 | level = btrfs_header_level(root->node); |
| 8202 | while (1) { | 8202 | while (1) { |
| 8203 | btrfs_tree_lock(path->nodes[level]); | 8203 | btrfs_tree_lock(path->nodes[level]); |
| 8204 | btrfs_set_lock_blocking(path->nodes[level]); | 8204 | btrfs_set_lock_blocking(path->nodes[level]); |
| 8205 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; | 8205 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; |
| 8206 | 8206 | ||
| 8207 | ret = btrfs_lookup_extent_info(trans, root, | 8207 | ret = btrfs_lookup_extent_info(trans, root, |
| 8208 | path->nodes[level]->start, | 8208 | path->nodes[level]->start, |
| 8209 | level, 1, &wc->refs[level], | 8209 | level, 1, &wc->refs[level], |
| 8210 | &wc->flags[level]); | 8210 | &wc->flags[level]); |
| 8211 | if (ret < 0) { | 8211 | if (ret < 0) { |
| 8212 | err = ret; | 8212 | err = ret; |
| 8213 | goto out_end_trans; | 8213 | goto out_end_trans; |
| 8214 | } | 8214 | } |
| 8215 | BUG_ON(wc->refs[level] == 0); | 8215 | BUG_ON(wc->refs[level] == 0); |
| 8216 | 8216 | ||
| 8217 | if (level == root_item->drop_level) | 8217 | if (level == root_item->drop_level) |
| 8218 | break; | 8218 | break; |
| 8219 | 8219 | ||
| 8220 | btrfs_tree_unlock(path->nodes[level]); | 8220 | btrfs_tree_unlock(path->nodes[level]); |
| 8221 | path->locks[level] = 0; | 8221 | path->locks[level] = 0; |
| 8222 | WARN_ON(wc->refs[level] != 1); | 8222 | WARN_ON(wc->refs[level] != 1); |
| 8223 | level--; | 8223 | level--; |
| 8224 | } | 8224 | } |
| 8225 | } | 8225 | } |
| 8226 | 8226 | ||
| 8227 | wc->level = level; | 8227 | wc->level = level; |
| 8228 | wc->shared_level = -1; | 8228 | wc->shared_level = -1; |
| 8229 | wc->stage = DROP_REFERENCE; | 8229 | wc->stage = DROP_REFERENCE; |
| 8230 | wc->update_ref = update_ref; | 8230 | wc->update_ref = update_ref; |
| 8231 | wc->keep_locks = 0; | 8231 | wc->keep_locks = 0; |
| 8232 | wc->for_reloc = for_reloc; | 8232 | wc->for_reloc = for_reloc; |
| 8233 | wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root); | 8233 | wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root); |
| 8234 | 8234 | ||
| 8235 | while (1) { | 8235 | while (1) { |
| 8236 | 8236 | ||
| 8237 | ret = walk_down_tree(trans, root, path, wc); | 8237 | ret = walk_down_tree(trans, root, path, wc); |
| 8238 | if (ret < 0) { | 8238 | if (ret < 0) { |
| 8239 | err = ret; | 8239 | err = ret; |
| 8240 | break; | 8240 | break; |
| 8241 | } | 8241 | } |
| 8242 | 8242 | ||
| 8243 | ret = walk_up_tree(trans, root, path, wc, BTRFS_MAX_LEVEL); | 8243 | ret = walk_up_tree(trans, root, path, wc, BTRFS_MAX_LEVEL); |
| 8244 | if (ret < 0) { | 8244 | if (ret < 0) { |
| 8245 | err = ret; | 8245 | err = ret; |
| 8246 | break; | 8246 | break; |
| 8247 | } | 8247 | } |
| 8248 | 8248 | ||
| 8249 | if (ret > 0) { | 8249 | if (ret > 0) { |
| 8250 | BUG_ON(wc->stage != DROP_REFERENCE); | 8250 | BUG_ON(wc->stage != DROP_REFERENCE); |
| 8251 | break; | 8251 | break; |
| 8252 | } | 8252 | } |
| 8253 | 8253 | ||
| 8254 | if (wc->stage == DROP_REFERENCE) { | 8254 | if (wc->stage == DROP_REFERENCE) { |
| 8255 | level = wc->level; | 8255 | level = wc->level; |
| 8256 | btrfs_node_key(path->nodes[level], | 8256 | btrfs_node_key(path->nodes[level], |
| 8257 | &root_item->drop_progress, | 8257 | &root_item->drop_progress, |
| 8258 | path->slots[level]); | 8258 | path->slots[level]); |
| 8259 | root_item->drop_level = level; | 8259 | root_item->drop_level = level; |
| 8260 | } | 8260 | } |
| 8261 | 8261 | ||
| 8262 | BUG_ON(wc->level == 0); | 8262 | BUG_ON(wc->level == 0); |
| 8263 | if (btrfs_should_end_transaction(trans, tree_root) || | 8263 | if (btrfs_should_end_transaction(trans, tree_root) || |
| 8264 | (!for_reloc && btrfs_need_cleaner_sleep(root))) { | 8264 | (!for_reloc && btrfs_need_cleaner_sleep(root))) { |
| 8265 | ret = btrfs_update_root(trans, tree_root, | 8265 | ret = btrfs_update_root(trans, tree_root, |
| 8266 | &root->root_key, | 8266 | &root->root_key, |
| 8267 | root_item); | 8267 | root_item); |
| 8268 | if (ret) { | 8268 | if (ret) { |
| 8269 | btrfs_abort_transaction(trans, tree_root, ret); | 8269 | btrfs_abort_transaction(trans, tree_root, ret); |
| 8270 | err = ret; | 8270 | err = ret; |
| 8271 | goto out_end_trans; | 8271 | goto out_end_trans; |
| 8272 | } | 8272 | } |
| 8273 | 8273 | ||
| 8274 | /* | 8274 | /* |
| 8275 | * Qgroup update accounting is run from | 8275 | * Qgroup update accounting is run from |
| 8276 | * delayed ref handling. This usually works | 8276 | * delayed ref handling. This usually works |
| 8277 | * out because delayed refs are normally the | 8277 | * out because delayed refs are normally the |
| 8278 | * only way qgroup updates are added. However, | 8278 | * only way qgroup updates are added. However, |
| 8279 | * we may have added updates during our tree | 8279 | * we may have added updates during our tree |
| 8280 | * walk so run qgroups here to make sure we | 8280 | * walk so run qgroups here to make sure we |
| 8281 | * don't lose any updates. | 8281 | * don't lose any updates. |
| 8282 | */ | 8282 | */ |
| 8283 | ret = btrfs_delayed_qgroup_accounting(trans, | 8283 | ret = btrfs_delayed_qgroup_accounting(trans, |
| 8284 | root->fs_info); | 8284 | root->fs_info); |
| 8285 | if (ret) | 8285 | if (ret) |
| 8286 | printk_ratelimited(KERN_ERR "BTRFS: Failure %d " | 8286 | printk_ratelimited(KERN_ERR "BTRFS: Failure %d " |
| 8287 | "running qgroup updates " | 8287 | "running qgroup updates " |
| 8288 | "during snapshot delete. " | 8288 | "during snapshot delete. " |
| 8289 | "Quota is out of sync, " | 8289 | "Quota is out of sync, " |
| 8290 | "rescan required.\n", ret); | 8290 | "rescan required.\n", ret); |
| 8291 | 8291 | ||
| 8292 | btrfs_end_transaction_throttle(trans, tree_root); | 8292 | btrfs_end_transaction_throttle(trans, tree_root); |
| 8293 | if (!for_reloc && btrfs_need_cleaner_sleep(root)) { | 8293 | if (!for_reloc && btrfs_need_cleaner_sleep(root)) { |
| 8294 | pr_debug("BTRFS: drop snapshot early exit\n"); | 8294 | pr_debug("BTRFS: drop snapshot early exit\n"); |
| 8295 | err = -EAGAIN; | 8295 | err = -EAGAIN; |
| 8296 | goto out_free; | 8296 | goto out_free; |
| 8297 | } | 8297 | } |
| 8298 | 8298 | ||
| 8299 | trans = btrfs_start_transaction(tree_root, 0); | 8299 | trans = btrfs_start_transaction(tree_root, 0); |
| 8300 | if (IS_ERR(trans)) { | 8300 | if (IS_ERR(trans)) { |
| 8301 | err = PTR_ERR(trans); | 8301 | err = PTR_ERR(trans); |
| 8302 | goto out_free; | 8302 | goto out_free; |
| 8303 | } | 8303 | } |
| 8304 | if (block_rsv) | 8304 | if (block_rsv) |
| 8305 | trans->block_rsv = block_rsv; | 8305 | trans->block_rsv = block_rsv; |
| 8306 | } | 8306 | } |
| 8307 | } | 8307 | } |
| 8308 | btrfs_release_path(path); | 8308 | btrfs_release_path(path); |
| 8309 | if (err) | 8309 | if (err) |
| 8310 | goto out_end_trans; | 8310 | goto out_end_trans; |
| 8311 | 8311 | ||
| 8312 | ret = btrfs_del_root(trans, tree_root, &root->root_key); | 8312 | ret = btrfs_del_root(trans, tree_root, &root->root_key); |
| 8313 | if (ret) { | 8313 | if (ret) { |
| 8314 | btrfs_abort_transaction(trans, tree_root, ret); | 8314 | btrfs_abort_transaction(trans, tree_root, ret); |
| 8315 | goto out_end_trans; | 8315 | goto out_end_trans; |
| 8316 | } | 8316 | } |
| 8317 | 8317 | ||
| 8318 | if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) { | 8318 | if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) { |
| 8319 | ret = btrfs_find_root(tree_root, &root->root_key, path, | 8319 | ret = btrfs_find_root(tree_root, &root->root_key, path, |
| 8320 | NULL, NULL); | 8320 | NULL, NULL); |
| 8321 | if (ret < 0) { | 8321 | if (ret < 0) { |
| 8322 | btrfs_abort_transaction(trans, tree_root, ret); | 8322 | btrfs_abort_transaction(trans, tree_root, ret); |
| 8323 | err = ret; | 8323 | err = ret; |
| 8324 | goto out_end_trans; | 8324 | goto out_end_trans; |
| 8325 | } else if (ret > 0) { | 8325 | } else if (ret > 0) { |
| 8326 | /* if we fail to delete the orphan item this time | 8326 | /* if we fail to delete the orphan item this time |
| 8327 | * around, it'll get picked up the next time. | 8327 | * around, it'll get picked up the next time. |
| 8328 | * | 8328 | * |
| 8329 | * The most common failure here is just -ENOENT. | 8329 | * The most common failure here is just -ENOENT. |
| 8330 | */ | 8330 | */ |
| 8331 | btrfs_del_orphan_item(trans, tree_root, | 8331 | btrfs_del_orphan_item(trans, tree_root, |
| 8332 | root->root_key.objectid); | 8332 | root->root_key.objectid); |
| 8333 | } | 8333 | } |
| 8334 | } | 8334 | } |
| 8335 | 8335 | ||
| 8336 | if (test_bit(BTRFS_ROOT_IN_RADIX, &root->state)) { | 8336 | if (test_bit(BTRFS_ROOT_IN_RADIX, &root->state)) { |
| 8337 | btrfs_drop_and_free_fs_root(tree_root->fs_info, root); | 8337 | btrfs_drop_and_free_fs_root(tree_root->fs_info, root); |
| 8338 | } else { | 8338 | } else { |
| 8339 | free_extent_buffer(root->node); | 8339 | free_extent_buffer(root->node); |
| 8340 | free_extent_buffer(root->commit_root); | 8340 | free_extent_buffer(root->commit_root); |
| 8341 | btrfs_put_fs_root(root); | 8341 | btrfs_put_fs_root(root); |
| 8342 | } | 8342 | } |
| 8343 | root_dropped = true; | 8343 | root_dropped = true; |
| 8344 | out_end_trans: | 8344 | out_end_trans: |
| 8345 | ret = btrfs_delayed_qgroup_accounting(trans, tree_root->fs_info); | 8345 | ret = btrfs_delayed_qgroup_accounting(trans, tree_root->fs_info); |
| 8346 | if (ret) | 8346 | if (ret) |
| 8347 | printk_ratelimited(KERN_ERR "BTRFS: Failure %d " | 8347 | printk_ratelimited(KERN_ERR "BTRFS: Failure %d " |
| 8348 | "running qgroup updates " | 8348 | "running qgroup updates " |
| 8349 | "during snapshot delete. " | 8349 | "during snapshot delete. " |
| 8350 | "Quota is out of sync, " | 8350 | "Quota is out of sync, " |
| 8351 | "rescan required.\n", ret); | 8351 | "rescan required.\n", ret); |
| 8352 | 8352 | ||
| 8353 | btrfs_end_transaction_throttle(trans, tree_root); | 8353 | btrfs_end_transaction_throttle(trans, tree_root); |
| 8354 | out_free: | 8354 | out_free: |
| 8355 | kfree(wc); | 8355 | kfree(wc); |
| 8356 | btrfs_free_path(path); | 8356 | btrfs_free_path(path); |
| 8357 | out: | 8357 | out: |
| 8358 | /* | 8358 | /* |
| 8359 | * So if we need to stop dropping the snapshot for whatever reason we | 8359 | * So if we need to stop dropping the snapshot for whatever reason we |
| 8360 | * need to make sure to add it back to the dead root list so that we | 8360 | * need to make sure to add it back to the dead root list so that we |
| 8361 | * keep trying to do the work later. This also cleans up roots if we | 8361 | * keep trying to do the work later. This also cleans up roots if we |
| 8362 | * don't have it in the radix (like when we recover after a power fail | 8362 | * don't have it in the radix (like when we recover after a power fail |
| 8363 | * or unmount) so we don't leak memory. | 8363 | * or unmount) so we don't leak memory. |
| 8364 | */ | 8364 | */ |
| 8365 | if (!for_reloc && root_dropped == false) | 8365 | if (!for_reloc && root_dropped == false) |
| 8366 | btrfs_add_dead_root(root); | 8366 | btrfs_add_dead_root(root); |
| 8367 | if (err && err != -EAGAIN) | 8367 | if (err && err != -EAGAIN) |
| 8368 | btrfs_std_error(root->fs_info, err); | 8368 | btrfs_std_error(root->fs_info, err); |
| 8369 | return err; | 8369 | return err; |
| 8370 | } | 8370 | } |
| 8371 | 8371 | ||
| 8372 | /* | 8372 | /* |
| 8373 | * drop subtree rooted at tree block 'node'. | 8373 | * drop subtree rooted at tree block 'node'. |
| 8374 | * | 8374 | * |
| 8375 | * NOTE: this function will unlock and release tree block 'node' | 8375 | * NOTE: this function will unlock and release tree block 'node' |
| 8376 | * only used by relocation code | 8376 | * only used by relocation code |
| 8377 | */ | 8377 | */ |
| 8378 | int btrfs_drop_subtree(struct btrfs_trans_handle *trans, | 8378 | int btrfs_drop_subtree(struct btrfs_trans_handle *trans, |
| 8379 | struct btrfs_root *root, | 8379 | struct btrfs_root *root, |
| 8380 | struct extent_buffer *node, | 8380 | struct extent_buffer *node, |
| 8381 | struct extent_buffer *parent) | 8381 | struct extent_buffer *parent) |
| 8382 | { | 8382 | { |
| 8383 | struct btrfs_path *path; | 8383 | struct btrfs_path *path; |
| 8384 | struct walk_control *wc; | 8384 | struct walk_control *wc; |
| 8385 | int level; | 8385 | int level; |
| 8386 | int parent_level; | 8386 | int parent_level; |
| 8387 | int ret = 0; | 8387 | int ret = 0; |
| 8388 | int wret; | 8388 | int wret; |
| 8389 | 8389 | ||
| 8390 | BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID); | 8390 | BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID); |
| 8391 | 8391 | ||
| 8392 | path = btrfs_alloc_path(); | 8392 | path = btrfs_alloc_path(); |
| 8393 | if (!path) | 8393 | if (!path) |
| 8394 | return -ENOMEM; | 8394 | return -ENOMEM; |
| 8395 | 8395 | ||
| 8396 | wc = kzalloc(sizeof(*wc), GFP_NOFS); | 8396 | wc = kzalloc(sizeof(*wc), GFP_NOFS); |
| 8397 | if (!wc) { | 8397 | if (!wc) { |
| 8398 | btrfs_free_path(path); | 8398 | btrfs_free_path(path); |
| 8399 | return -ENOMEM; | 8399 | return -ENOMEM; |
| 8400 | } | 8400 | } |
| 8401 | 8401 | ||
| 8402 | btrfs_assert_tree_locked(parent); | 8402 | btrfs_assert_tree_locked(parent); |
| 8403 | parent_level = btrfs_header_level(parent); | 8403 | parent_level = btrfs_header_level(parent); |
| 8404 | extent_buffer_get(parent); | 8404 | extent_buffer_get(parent); |
| 8405 | path->nodes[parent_level] = parent; | 8405 | path->nodes[parent_level] = parent; |
| 8406 | path->slots[parent_level] = btrfs_header_nritems(parent); | 8406 | path->slots[parent_level] = btrfs_header_nritems(parent); |
| 8407 | 8407 | ||
| 8408 | btrfs_assert_tree_locked(node); | 8408 | btrfs_assert_tree_locked(node); |
| 8409 | level = btrfs_header_level(node); | 8409 | level = btrfs_header_level(node); |
| 8410 | path->nodes[level] = node; | 8410 | path->nodes[level] = node; |
| 8411 | path->slots[level] = 0; | 8411 | path->slots[level] = 0; |
| 8412 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; | 8412 | path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING; |
| 8413 | 8413 | ||
| 8414 | wc->refs[parent_level] = 1; | 8414 | wc->refs[parent_level] = 1; |
| 8415 | wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF; | 8415 | wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF; |
| 8416 | wc->level = level; | 8416 | wc->level = level; |
| 8417 | wc->shared_level = -1; | 8417 | wc->shared_level = -1; |
| 8418 | wc->stage = DROP_REFERENCE; | 8418 | wc->stage = DROP_REFERENCE; |
| 8419 | wc->update_ref = 0; | 8419 | wc->update_ref = 0; |
| 8420 | wc->keep_locks = 1; | 8420 | wc->keep_locks = 1; |
| 8421 | wc->for_reloc = 1; | 8421 | wc->for_reloc = 1; |
| 8422 | wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root); | 8422 | wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(root); |
| 8423 | 8423 | ||
| 8424 | while (1) { | 8424 | while (1) { |
| 8425 | wret = walk_down_tree(trans, root, path, wc); | 8425 | wret = walk_down_tree(trans, root, path, wc); |
| 8426 | if (wret < 0) { | 8426 | if (wret < 0) { |
| 8427 | ret = wret; | 8427 | ret = wret; |
| 8428 | break; | 8428 | break; |
| 8429 | } | 8429 | } |
| 8430 | 8430 | ||
| 8431 | wret = walk_up_tree(trans, root, path, wc, parent_level); | 8431 | wret = walk_up_tree(trans, root, path, wc, parent_level); |
| 8432 | if (wret < 0) | 8432 | if (wret < 0) |
| 8433 | ret = wret; | 8433 | ret = wret; |
| 8434 | if (wret != 0) | 8434 | if (wret != 0) |
| 8435 | break; | 8435 | break; |
| 8436 | } | 8436 | } |
| 8437 | 8437 | ||
| 8438 | kfree(wc); | 8438 | kfree(wc); |
| 8439 | btrfs_free_path(path); | 8439 | btrfs_free_path(path); |
| 8440 | return ret; | 8440 | return ret; |
| 8441 | } | 8441 | } |
| 8442 | 8442 | ||
| 8443 | static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) | 8443 | static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) |
| 8444 | { | 8444 | { |
| 8445 | u64 num_devices; | 8445 | u64 num_devices; |
| 8446 | u64 stripped; | 8446 | u64 stripped; |
| 8447 | 8447 | ||
| 8448 | /* | 8448 | /* |
| 8449 | * if restripe for this chunk_type is on pick target profile and | 8449 | * if restripe for this chunk_type is on pick target profile and |
| 8450 | * return, otherwise do the usual balance | 8450 | * return, otherwise do the usual balance |
| 8451 | */ | 8451 | */ |
| 8452 | stripped = get_restripe_target(root->fs_info, flags); | 8452 | stripped = get_restripe_target(root->fs_info, flags); |
| 8453 | if (stripped) | 8453 | if (stripped) |
| 8454 | return extended_to_chunk(stripped); | 8454 | return extended_to_chunk(stripped); |
| 8455 | 8455 | ||
| 8456 | num_devices = root->fs_info->fs_devices->rw_devices; | 8456 | num_devices = root->fs_info->fs_devices->rw_devices; |
| 8457 | 8457 | ||
| 8458 | stripped = BTRFS_BLOCK_GROUP_RAID0 | | 8458 | stripped = BTRFS_BLOCK_GROUP_RAID0 | |
| 8459 | BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6 | | 8459 | BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6 | |
| 8460 | BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10; | 8460 | BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10; |
| 8461 | 8461 | ||
| 8462 | if (num_devices == 1) { | 8462 | if (num_devices == 1) { |
| 8463 | stripped |= BTRFS_BLOCK_GROUP_DUP; | 8463 | stripped |= BTRFS_BLOCK_GROUP_DUP; |
| 8464 | stripped = flags & ~stripped; | 8464 | stripped = flags & ~stripped; |
| 8465 | 8465 | ||
| 8466 | /* turn raid0 into single device chunks */ | 8466 | /* turn raid0 into single device chunks */ |
| 8467 | if (flags & BTRFS_BLOCK_GROUP_RAID0) | 8467 | if (flags & BTRFS_BLOCK_GROUP_RAID0) |
| 8468 | return stripped; | 8468 | return stripped; |
| 8469 | 8469 | ||
| 8470 | /* turn mirroring into duplication */ | 8470 | /* turn mirroring into duplication */ |
| 8471 | if (flags & (BTRFS_BLOCK_GROUP_RAID1 | | 8471 | if (flags & (BTRFS_BLOCK_GROUP_RAID1 | |
| 8472 | BTRFS_BLOCK_GROUP_RAID10)) | 8472 | BTRFS_BLOCK_GROUP_RAID10)) |
| 8473 | return stripped | BTRFS_BLOCK_GROUP_DUP; | 8473 | return stripped | BTRFS_BLOCK_GROUP_DUP; |
| 8474 | } else { | 8474 | } else { |
| 8475 | /* they already had raid on here, just return */ | 8475 | /* they already had raid on here, just return */ |
| 8476 | if (flags & stripped) | 8476 | if (flags & stripped) |
| 8477 | return flags; | 8477 | return flags; |
| 8478 | 8478 | ||
| 8479 | stripped |= BTRFS_BLOCK_GROUP_DUP; | 8479 | stripped |= BTRFS_BLOCK_GROUP_DUP; |
| 8480 | stripped = flags & ~stripped; | 8480 | stripped = flags & ~stripped; |
| 8481 | 8481 | ||
| 8482 | /* switch duplicated blocks with raid1 */ | 8482 | /* switch duplicated blocks with raid1 */ |
| 8483 | if (flags & BTRFS_BLOCK_GROUP_DUP) | 8483 | if (flags & BTRFS_BLOCK_GROUP_DUP) |
| 8484 | return stripped | BTRFS_BLOCK_GROUP_RAID1; | 8484 | return stripped | BTRFS_BLOCK_GROUP_RAID1; |
| 8485 | 8485 | ||
| 8486 | /* this is drive concat, leave it alone */ | 8486 | /* this is drive concat, leave it alone */ |
| 8487 | } | 8487 | } |
| 8488 | 8488 | ||
| 8489 | return flags; | 8489 | return flags; |
| 8490 | } | 8490 | } |
| 8491 | 8491 | ||
| 8492 | static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force) | 8492 | static int set_block_group_ro(struct btrfs_block_group_cache *cache, int force) |
| 8493 | { | 8493 | { |
| 8494 | struct btrfs_space_info *sinfo = cache->space_info; | 8494 | struct btrfs_space_info *sinfo = cache->space_info; |
| 8495 | u64 num_bytes; | 8495 | u64 num_bytes; |
| 8496 | u64 min_allocable_bytes; | 8496 | u64 min_allocable_bytes; |
| 8497 | int ret = -ENOSPC; | 8497 | int ret = -ENOSPC; |
| 8498 | 8498 | ||
| 8499 | 8499 | ||
| 8500 | /* | 8500 | /* |
| 8501 | * We need some metadata space and system metadata space for | 8501 | * We need some metadata space and system metadata space for |
| 8502 | * allocating chunks in some corner cases until we force to set | 8502 | * allocating chunks in some corner cases until we force to set |
| 8503 | * it to be readonly. | 8503 | * it to be readonly. |
| 8504 | */ | 8504 | */ |
| 8505 | if ((sinfo->flags & | 8505 | if ((sinfo->flags & |
| 8506 | (BTRFS_BLOCK_GROUP_SYSTEM | BTRFS_BLOCK_GROUP_METADATA)) && | 8506 | (BTRFS_BLOCK_GROUP_SYSTEM | BTRFS_BLOCK_GROUP_METADATA)) && |
| 8507 | !force) | 8507 | !force) |
| 8508 | min_allocable_bytes = 1 * 1024 * 1024; | 8508 | min_allocable_bytes = 1 * 1024 * 1024; |
| 8509 | else | 8509 | else |
| 8510 | min_allocable_bytes = 0; | 8510 | min_allocable_bytes = 0; |
| 8511 | 8511 | ||
| 8512 | spin_lock(&sinfo->lock); | 8512 | spin_lock(&sinfo->lock); |
| 8513 | spin_lock(&cache->lock); | 8513 | spin_lock(&cache->lock); |
| 8514 | 8514 | ||
| 8515 | if (cache->ro) { | 8515 | if (cache->ro) { |
| 8516 | ret = 0; | 8516 | ret = 0; |
| 8517 | goto out; | 8517 | goto out; |
| 8518 | } | 8518 | } |
| 8519 | 8519 | ||
| 8520 | num_bytes = cache->key.offset - cache->reserved - cache->pinned - | 8520 | num_bytes = cache->key.offset - cache->reserved - cache->pinned - |
| 8521 | cache->bytes_super - btrfs_block_group_used(&cache->item); | 8521 | cache->bytes_super - btrfs_block_group_used(&cache->item); |
| 8522 | 8522 | ||
| 8523 | if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned + | 8523 | if (sinfo->bytes_used + sinfo->bytes_reserved + sinfo->bytes_pinned + |
| 8524 | sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes + | 8524 | sinfo->bytes_may_use + sinfo->bytes_readonly + num_bytes + |
| 8525 | min_allocable_bytes <= sinfo->total_bytes) { | 8525 | min_allocable_bytes <= sinfo->total_bytes) { |
| 8526 | sinfo->bytes_readonly += num_bytes; | 8526 | sinfo->bytes_readonly += num_bytes; |
| 8527 | cache->ro = 1; | 8527 | cache->ro = 1; |
| 8528 | list_add_tail(&cache->ro_list, &sinfo->ro_bgs); | 8528 | list_add_tail(&cache->ro_list, &sinfo->ro_bgs); |
| 8529 | ret = 0; | 8529 | ret = 0; |
| 8530 | } | 8530 | } |
| 8531 | out: | 8531 | out: |
| 8532 | spin_unlock(&cache->lock); | 8532 | spin_unlock(&cache->lock); |
| 8533 | spin_unlock(&sinfo->lock); | 8533 | spin_unlock(&sinfo->lock); |
| 8534 | return ret; | 8534 | return ret; |
| 8535 | } | 8535 | } |
| 8536 | 8536 | ||
| 8537 | int btrfs_set_block_group_ro(struct btrfs_root *root, | 8537 | int btrfs_set_block_group_ro(struct btrfs_root *root, |
| 8538 | struct btrfs_block_group_cache *cache) | 8538 | struct btrfs_block_group_cache *cache) |
| 8539 | 8539 | ||
| 8540 | { | 8540 | { |
| 8541 | struct btrfs_trans_handle *trans; | 8541 | struct btrfs_trans_handle *trans; |
| 8542 | u64 alloc_flags; | 8542 | u64 alloc_flags; |
| 8543 | int ret; | 8543 | int ret; |
| 8544 | 8544 | ||
| 8545 | BUG_ON(cache->ro); | 8545 | BUG_ON(cache->ro); |
| 8546 | 8546 | ||
| 8547 | trans = btrfs_join_transaction(root); | 8547 | trans = btrfs_join_transaction(root); |
| 8548 | if (IS_ERR(trans)) | 8548 | if (IS_ERR(trans)) |
| 8549 | return PTR_ERR(trans); | 8549 | return PTR_ERR(trans); |
| 8550 | 8550 | ||
| 8551 | alloc_flags = update_block_group_flags(root, cache->flags); | 8551 | alloc_flags = update_block_group_flags(root, cache->flags); |
| 8552 | if (alloc_flags != cache->flags) { | 8552 | if (alloc_flags != cache->flags) { |
| 8553 | ret = do_chunk_alloc(trans, root, alloc_flags, | 8553 | ret = do_chunk_alloc(trans, root, alloc_flags, |
| 8554 | CHUNK_ALLOC_FORCE); | 8554 | CHUNK_ALLOC_FORCE); |
| 8555 | if (ret < 0) | 8555 | if (ret < 0) |
| 8556 | goto out; | 8556 | goto out; |
| 8557 | } | 8557 | } |
| 8558 | 8558 | ||
| 8559 | ret = set_block_group_ro(cache, 0); | 8559 | ret = set_block_group_ro(cache, 0); |
| 8560 | if (!ret) | 8560 | if (!ret) |
| 8561 | goto out; | 8561 | goto out; |
| 8562 | alloc_flags = get_alloc_profile(root, cache->space_info->flags); | 8562 | alloc_flags = get_alloc_profile(root, cache->space_info->flags); |
| 8563 | ret = do_chunk_alloc(trans, root, alloc_flags, | 8563 | ret = do_chunk_alloc(trans, root, alloc_flags, |
| 8564 | CHUNK_ALLOC_FORCE); | 8564 | CHUNK_ALLOC_FORCE); |
| 8565 | if (ret < 0) | 8565 | if (ret < 0) |
| 8566 | goto out; | 8566 | goto out; |
| 8567 | ret = set_block_group_ro(cache, 0); | 8567 | ret = set_block_group_ro(cache, 0); |
| 8568 | out: | 8568 | out: |
| 8569 | btrfs_end_transaction(trans, root); | 8569 | btrfs_end_transaction(trans, root); |
| 8570 | return ret; | 8570 | return ret; |
| 8571 | } | 8571 | } |
| 8572 | 8572 | ||
| 8573 | int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, | 8573 | int btrfs_force_chunk_alloc(struct btrfs_trans_handle *trans, |
| 8574 | struct btrfs_root *root, u64 type) | 8574 | struct btrfs_root *root, u64 type) |
| 8575 | { | 8575 | { |
| 8576 | u64 alloc_flags = get_alloc_profile(root, type); | 8576 | u64 alloc_flags = get_alloc_profile(root, type); |
| 8577 | return do_chunk_alloc(trans, root, alloc_flags, | 8577 | return do_chunk_alloc(trans, root, alloc_flags, |
| 8578 | CHUNK_ALLOC_FORCE); | 8578 | CHUNK_ALLOC_FORCE); |
| 8579 | } | 8579 | } |
| 8580 | 8580 | ||
| 8581 | /* | 8581 | /* |
| 8582 | * helper to account the unused space of all the readonly block group in the | 8582 | * helper to account the unused space of all the readonly block group in the |
| 8583 | * space_info. takes mirrors into account. | 8583 | * space_info. takes mirrors into account. |
| 8584 | */ | 8584 | */ |
| 8585 | u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo) | 8585 | u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo) |
| 8586 | { | 8586 | { |
| 8587 | struct btrfs_block_group_cache *block_group; | 8587 | struct btrfs_block_group_cache *block_group; |
| 8588 | u64 free_bytes = 0; | 8588 | u64 free_bytes = 0; |
| 8589 | int factor; | 8589 | int factor; |
| 8590 | 8590 | ||
| 8591 | /* It's df, we don't care if it's racey */ | 8591 | /* It's df, we don't care if it's racey */ |
| 8592 | if (list_empty(&sinfo->ro_bgs)) | 8592 | if (list_empty(&sinfo->ro_bgs)) |
| 8593 | return 0; | 8593 | return 0; |
| 8594 | 8594 | ||
| 8595 | spin_lock(&sinfo->lock); | 8595 | spin_lock(&sinfo->lock); |
| 8596 | list_for_each_entry(block_group, &sinfo->ro_bgs, ro_list) { | 8596 | list_for_each_entry(block_group, &sinfo->ro_bgs, ro_list) { |
| 8597 | spin_lock(&block_group->lock); | 8597 | spin_lock(&block_group->lock); |
| 8598 | 8598 | ||
| 8599 | if (!block_group->ro) { | 8599 | if (!block_group->ro) { |
| 8600 | spin_unlock(&block_group->lock); | 8600 | spin_unlock(&block_group->lock); |
| 8601 | continue; | 8601 | continue; |
| 8602 | } | 8602 | } |
| 8603 | 8603 | ||
| 8604 | if (block_group->flags & (BTRFS_BLOCK_GROUP_RAID1 | | 8604 | if (block_group->flags & (BTRFS_BLOCK_GROUP_RAID1 | |
| 8605 | BTRFS_BLOCK_GROUP_RAID10 | | 8605 | BTRFS_BLOCK_GROUP_RAID10 | |
| 8606 | BTRFS_BLOCK_GROUP_DUP)) | 8606 | BTRFS_BLOCK_GROUP_DUP)) |
| 8607 | factor = 2; | 8607 | factor = 2; |
| 8608 | else | 8608 | else |
| 8609 | factor = 1; | 8609 | factor = 1; |
| 8610 | 8610 | ||
| 8611 | free_bytes += (block_group->key.offset - | 8611 | free_bytes += (block_group->key.offset - |
| 8612 | btrfs_block_group_used(&block_group->item)) * | 8612 | btrfs_block_group_used(&block_group->item)) * |
| 8613 | factor; | 8613 | factor; |
| 8614 | 8614 | ||
| 8615 | spin_unlock(&block_group->lock); | 8615 | spin_unlock(&block_group->lock); |
| 8616 | } | 8616 | } |
| 8617 | spin_unlock(&sinfo->lock); | 8617 | spin_unlock(&sinfo->lock); |
| 8618 | 8618 | ||
| 8619 | return free_bytes; | 8619 | return free_bytes; |
| 8620 | } | 8620 | } |
| 8621 | 8621 | ||
| 8622 | void btrfs_set_block_group_rw(struct btrfs_root *root, | 8622 | void btrfs_set_block_group_rw(struct btrfs_root *root, |
| 8623 | struct btrfs_block_group_cache *cache) | 8623 | struct btrfs_block_group_cache *cache) |
| 8624 | { | 8624 | { |
| 8625 | struct btrfs_space_info *sinfo = cache->space_info; | 8625 | struct btrfs_space_info *sinfo = cache->space_info; |
| 8626 | u64 num_bytes; | 8626 | u64 num_bytes; |
| 8627 | 8627 | ||
| 8628 | BUG_ON(!cache->ro); | 8628 | BUG_ON(!cache->ro); |
| 8629 | 8629 | ||
| 8630 | spin_lock(&sinfo->lock); | 8630 | spin_lock(&sinfo->lock); |
| 8631 | spin_lock(&cache->lock); | 8631 | spin_lock(&cache->lock); |
| 8632 | num_bytes = cache->key.offset - cache->reserved - cache->pinned - | 8632 | num_bytes = cache->key.offset - cache->reserved - cache->pinned - |
| 8633 | cache->bytes_super - btrfs_block_group_used(&cache->item); | 8633 | cache->bytes_super - btrfs_block_group_used(&cache->item); |
| 8634 | sinfo->bytes_readonly -= num_bytes; | 8634 | sinfo->bytes_readonly -= num_bytes; |
| 8635 | cache->ro = 0; | 8635 | cache->ro = 0; |
| 8636 | list_del_init(&cache->ro_list); | 8636 | list_del_init(&cache->ro_list); |
| 8637 | spin_unlock(&cache->lock); | 8637 | spin_unlock(&cache->lock); |
| 8638 | spin_unlock(&sinfo->lock); | 8638 | spin_unlock(&sinfo->lock); |
| 8639 | } | 8639 | } |
| 8640 | 8640 | ||
| 8641 | /* | 8641 | /* |
| 8642 | * checks to see if its even possible to relocate this block group. | 8642 | * checks to see if its even possible to relocate this block group. |
| 8643 | * | 8643 | * |
| 8644 | * @return - -1 if it's not a good idea to relocate this block group, 0 if its | 8644 | * @return - -1 if it's not a good idea to relocate this block group, 0 if its |
| 8645 | * ok to go ahead and try. | 8645 | * ok to go ahead and try. |
| 8646 | */ | 8646 | */ |
| 8647 | int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) | 8647 | int btrfs_can_relocate(struct btrfs_root *root, u64 bytenr) |
| 8648 | { | 8648 | { |
| 8649 | struct btrfs_block_group_cache *block_group; | 8649 | struct btrfs_block_group_cache *block_group; |
| 8650 | struct btrfs_space_info *space_info; | 8650 | struct btrfs_space_info *space_info; |
| 8651 | struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; | 8651 | struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; |
| 8652 | struct btrfs_device *device; | 8652 | struct btrfs_device *device; |
| 8653 | struct btrfs_trans_handle *trans; | 8653 | struct btrfs_trans_handle *trans; |
| 8654 | u64 min_free; | 8654 | u64 min_free; |
| 8655 | u64 dev_min = 1; | 8655 | u64 dev_min = 1; |
| 8656 | u64 dev_nr = 0; | 8656 | u64 dev_nr = 0; |
| 8657 | u64 target; | 8657 | u64 target; |
| 8658 | int index; | 8658 | int index; |
| 8659 | int full = 0; | 8659 | int full = 0; |
| 8660 | int ret = 0; | 8660 | int ret = 0; |
| 8661 | 8661 | ||
| 8662 | block_group = btrfs_lookup_block_group(root->fs_info, bytenr); | 8662 | block_group = btrfs_lookup_block_group(root->fs_info, bytenr); |
| 8663 | 8663 | ||
| 8664 | /* odd, couldn't find the block group, leave it alone */ | 8664 | /* odd, couldn't find the block group, leave it alone */ |
| 8665 | if (!block_group) | 8665 | if (!block_group) |
| 8666 | return -1; | 8666 | return -1; |
| 8667 | 8667 | ||
| 8668 | min_free = btrfs_block_group_used(&block_group->item); | 8668 | min_free = btrfs_block_group_used(&block_group->item); |
| 8669 | 8669 | ||
| 8670 | /* no bytes used, we're good */ | 8670 | /* no bytes used, we're good */ |
| 8671 | if (!min_free) | 8671 | if (!min_free) |
| 8672 | goto out; | 8672 | goto out; |
| 8673 | 8673 | ||
| 8674 | space_info = block_group->space_info; | 8674 | space_info = block_group->space_info; |
| 8675 | spin_lock(&space_info->lock); | 8675 | spin_lock(&space_info->lock); |
| 8676 | 8676 | ||
| 8677 | full = space_info->full; | 8677 | full = space_info->full; |
| 8678 | 8678 | ||
| 8679 | /* | 8679 | /* |
| 8680 | * if this is the last block group we have in this space, we can't | 8680 | * if this is the last block group we have in this space, we can't |
| 8681 | * relocate it unless we're able to allocate a new chunk below. | 8681 | * relocate it unless we're able to allocate a new chunk below. |
| 8682 | * | 8682 | * |
| 8683 | * Otherwise, we need to make sure we have room in the space to handle | 8683 | * Otherwise, we need to make sure we have room in the space to handle |
| 8684 | * all of the extents from this block group. If we can, we're good | 8684 | * all of the extents from this block group. If we can, we're good |
| 8685 | */ | 8685 | */ |
| 8686 | if ((space_info->total_bytes != block_group->key.offset) && | 8686 | if ((space_info->total_bytes != block_group->key.offset) && |
| 8687 | (space_info->bytes_used + space_info->bytes_reserved + | 8687 | (space_info->bytes_used + space_info->bytes_reserved + |
| 8688 | space_info->bytes_pinned + space_info->bytes_readonly + | 8688 | space_info->bytes_pinned + space_info->bytes_readonly + |
| 8689 | min_free < space_info->total_bytes)) { | 8689 | min_free < space_info->total_bytes)) { |
| 8690 | spin_unlock(&space_info->lock); | 8690 | spin_unlock(&space_info->lock); |
| 8691 | goto out; | 8691 | goto out; |
| 8692 | } | 8692 | } |
| 8693 | spin_unlock(&space_info->lock); | 8693 | spin_unlock(&space_info->lock); |
| 8694 | 8694 | ||
| 8695 | /* | 8695 | /* |
| 8696 | * ok we don't have enough space, but maybe we have free space on our | 8696 | * ok we don't have enough space, but maybe we have free space on our |
| 8697 | * devices to allocate new chunks for relocation, so loop through our | 8697 | * devices to allocate new chunks for relocation, so loop through our |
| 8698 | * alloc devices and guess if we have enough space. if this block | 8698 | * alloc devices and guess if we have enough space. if this block |
| 8699 | * group is going to be restriped, run checks against the target | 8699 | * group is going to be restriped, run checks against the target |
| 8700 | * profile instead of the current one. | 8700 | * profile instead of the current one. |
| 8701 | */ | 8701 | */ |
| 8702 | ret = -1; | 8702 | ret = -1; |
| 8703 | 8703 | ||
| 8704 | /* | 8704 | /* |
| 8705 | * index: | 8705 | * index: |
| 8706 | * 0: raid10 | 8706 | * 0: raid10 |
| 8707 | * 1: raid1 | 8707 | * 1: raid1 |
| 8708 | * 2: dup | 8708 | * 2: dup |
| 8709 | * 3: raid0 | 8709 | * 3: raid0 |
| 8710 | * 4: single | 8710 | * 4: single |
| 8711 | */ | 8711 | */ |
| 8712 | target = get_restripe_target(root->fs_info, block_group->flags); | 8712 | target = get_restripe_target(root->fs_info, block_group->flags); |
| 8713 | if (target) { | 8713 | if (target) { |
| 8714 | index = __get_raid_index(extended_to_chunk(target)); | 8714 | index = __get_raid_index(extended_to_chunk(target)); |
| 8715 | } else { | 8715 | } else { |
| 8716 | /* | 8716 | /* |
| 8717 | * this is just a balance, so if we were marked as full | 8717 | * this is just a balance, so if we were marked as full |
| 8718 | * we know there is no space for a new chunk | 8718 | * we know there is no space for a new chunk |
| 8719 | */ | 8719 | */ |
| 8720 | if (full) | 8720 | if (full) |
| 8721 | goto out; | 8721 | goto out; |
| 8722 | 8722 | ||
| 8723 | index = get_block_group_index(block_group); | 8723 | index = get_block_group_index(block_group); |
| 8724 | } | 8724 | } |
| 8725 | 8725 | ||
| 8726 | if (index == BTRFS_RAID_RAID10) { | 8726 | if (index == BTRFS_RAID_RAID10) { |
| 8727 | dev_min = 4; | 8727 | dev_min = 4; |
| 8728 | /* Divide by 2 */ | 8728 | /* Divide by 2 */ |
| 8729 | min_free >>= 1; | 8729 | min_free >>= 1; |
| 8730 | } else if (index == BTRFS_RAID_RAID1) { | 8730 | } else if (index == BTRFS_RAID_RAID1) { |
| 8731 | dev_min = 2; | 8731 | dev_min = 2; |
| 8732 | } else if (index == BTRFS_RAID_DUP) { | 8732 | } else if (index == BTRFS_RAID_DUP) { |
| 8733 | /* Multiply by 2 */ | 8733 | /* Multiply by 2 */ |
| 8734 | min_free <<= 1; | 8734 | min_free <<= 1; |
| 8735 | } else if (index == BTRFS_RAID_RAID0) { | 8735 | } else if (index == BTRFS_RAID_RAID0) { |
| 8736 | dev_min = fs_devices->rw_devices; | 8736 | dev_min = fs_devices->rw_devices; |
| 8737 | do_div(min_free, dev_min); | 8737 | do_div(min_free, dev_min); |
| 8738 | } | 8738 | } |
| 8739 | 8739 | ||
| 8740 | /* We need to do this so that we can look at pending chunks */ | 8740 | /* We need to do this so that we can look at pending chunks */ |
| 8741 | trans = btrfs_join_transaction(root); | 8741 | trans = btrfs_join_transaction(root); |
| 8742 | if (IS_ERR(trans)) { | 8742 | if (IS_ERR(trans)) { |
| 8743 | ret = PTR_ERR(trans); | 8743 | ret = PTR_ERR(trans); |
| 8744 | goto out; | 8744 | goto out; |
| 8745 | } | 8745 | } |
| 8746 | 8746 | ||
| 8747 | mutex_lock(&root->fs_info->chunk_mutex); | 8747 | mutex_lock(&root->fs_info->chunk_mutex); |
| 8748 | list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) { | 8748 | list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) { |
| 8749 | u64 dev_offset; | 8749 | u64 dev_offset; |
| 8750 | 8750 | ||
| 8751 | /* | 8751 | /* |
| 8752 | * check to make sure we can actually find a chunk with enough | 8752 | * check to make sure we can actually find a chunk with enough |
| 8753 | * space to fit our block group in. | 8753 | * space to fit our block group in. |
| 8754 | */ | 8754 | */ |
| 8755 | if (device->total_bytes > device->bytes_used + min_free && | 8755 | if (device->total_bytes > device->bytes_used + min_free && |
| 8756 | !device->is_tgtdev_for_dev_replace) { | 8756 | !device->is_tgtdev_for_dev_replace) { |
| 8757 | ret = find_free_dev_extent(trans, device, min_free, | 8757 | ret = find_free_dev_extent(trans, device, min_free, |
| 8758 | &dev_offset, NULL); | 8758 | &dev_offset, NULL); |
| 8759 | if (!ret) | 8759 | if (!ret) |
| 8760 | dev_nr++; | 8760 | dev_nr++; |
| 8761 | 8761 | ||
| 8762 | if (dev_nr >= dev_min) | 8762 | if (dev_nr >= dev_min) |
| 8763 | break; | 8763 | break; |
| 8764 | 8764 | ||
| 8765 | ret = -1; | 8765 | ret = -1; |
| 8766 | } | 8766 | } |
| 8767 | } | 8767 | } |
| 8768 | mutex_unlock(&root->fs_info->chunk_mutex); | 8768 | mutex_unlock(&root->fs_info->chunk_mutex); |
| 8769 | btrfs_end_transaction(trans, root); | 8769 | btrfs_end_transaction(trans, root); |
| 8770 | out: | 8770 | out: |
| 8771 | btrfs_put_block_group(block_group); | 8771 | btrfs_put_block_group(block_group); |
| 8772 | return ret; | 8772 | return ret; |
| 8773 | } | 8773 | } |
| 8774 | 8774 | ||
| 8775 | static int find_first_block_group(struct btrfs_root *root, | 8775 | static int find_first_block_group(struct btrfs_root *root, |
| 8776 | struct btrfs_path *path, struct btrfs_key *key) | 8776 | struct btrfs_path *path, struct btrfs_key *key) |
| 8777 | { | 8777 | { |
| 8778 | int ret = 0; | 8778 | int ret = 0; |
| 8779 | struct btrfs_key found_key; | 8779 | struct btrfs_key found_key; |
| 8780 | struct extent_buffer *leaf; | 8780 | struct extent_buffer *leaf; |
| 8781 | int slot; | 8781 | int slot; |
| 8782 | 8782 | ||
| 8783 | ret = btrfs_search_slot(NULL, root, key, path, 0, 0); | 8783 | ret = btrfs_search_slot(NULL, root, key, path, 0, 0); |
| 8784 | if (ret < 0) | 8784 | if (ret < 0) |
| 8785 | goto out; | 8785 | goto out; |
| 8786 | 8786 | ||
| 8787 | while (1) { | 8787 | while (1) { |
| 8788 | slot = path->slots[0]; | 8788 | slot = path->slots[0]; |
| 8789 | leaf = path->nodes[0]; | 8789 | leaf = path->nodes[0]; |
| 8790 | if (slot >= btrfs_header_nritems(leaf)) { | 8790 | if (slot >= btrfs_header_nritems(leaf)) { |
| 8791 | ret = btrfs_next_leaf(root, path); | 8791 | ret = btrfs_next_leaf(root, path); |
| 8792 | if (ret == 0) | 8792 | if (ret == 0) |
| 8793 | continue; | 8793 | continue; |
| 8794 | if (ret < 0) | 8794 | if (ret < 0) |
| 8795 | goto out; | 8795 | goto out; |
| 8796 | break; | 8796 | break; |
| 8797 | } | 8797 | } |
| 8798 | btrfs_item_key_to_cpu(leaf, &found_key, slot); | 8798 | btrfs_item_key_to_cpu(leaf, &found_key, slot); |
| 8799 | 8799 | ||
| 8800 | if (found_key.objectid >= key->objectid && | 8800 | if (found_key.objectid >= key->objectid && |
| 8801 | found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) { | 8801 | found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) { |
| 8802 | ret = 0; | 8802 | ret = 0; |
| 8803 | goto out; | 8803 | goto out; |
| 8804 | } | 8804 | } |
| 8805 | path->slots[0]++; | 8805 | path->slots[0]++; |
| 8806 | } | 8806 | } |
| 8807 | out: | 8807 | out: |
| 8808 | return ret; | 8808 | return ret; |
| 8809 | } | 8809 | } |
| 8810 | 8810 | ||
| 8811 | void btrfs_put_block_group_cache(struct btrfs_fs_info *info) | 8811 | void btrfs_put_block_group_cache(struct btrfs_fs_info *info) |
| 8812 | { | 8812 | { |
| 8813 | struct btrfs_block_group_cache *block_group; | 8813 | struct btrfs_block_group_cache *block_group; |
| 8814 | u64 last = 0; | 8814 | u64 last = 0; |
| 8815 | 8815 | ||
| 8816 | while (1) { | 8816 | while (1) { |
| 8817 | struct inode *inode; | 8817 | struct inode *inode; |
| 8818 | 8818 | ||
| 8819 | block_group = btrfs_lookup_first_block_group(info, last); | 8819 | block_group = btrfs_lookup_first_block_group(info, last); |
| 8820 | while (block_group) { | 8820 | while (block_group) { |
| 8821 | spin_lock(&block_group->lock); | 8821 | spin_lock(&block_group->lock); |
| 8822 | if (block_group->iref) | 8822 | if (block_group->iref) |
| 8823 | break; | 8823 | break; |
| 8824 | spin_unlock(&block_group->lock); | 8824 | spin_unlock(&block_group->lock); |
| 8825 | block_group = next_block_group(info->tree_root, | 8825 | block_group = next_block_group(info->tree_root, |
| 8826 | block_group); | 8826 | block_group); |
| 8827 | } | 8827 | } |
| 8828 | if (!block_group) { | 8828 | if (!block_group) { |
| 8829 | if (last == 0) | 8829 | if (last == 0) |
| 8830 | break; | 8830 | break; |
| 8831 | last = 0; | 8831 | last = 0; |
| 8832 | continue; | 8832 | continue; |
| 8833 | } | 8833 | } |
| 8834 | 8834 | ||
| 8835 | inode = block_group->inode; | 8835 | inode = block_group->inode; |
| 8836 | block_group->iref = 0; | 8836 | block_group->iref = 0; |
| 8837 | block_group->inode = NULL; | 8837 | block_group->inode = NULL; |
| 8838 | spin_unlock(&block_group->lock); | 8838 | spin_unlock(&block_group->lock); |
| 8839 | iput(inode); | 8839 | iput(inode); |
| 8840 | last = block_group->key.objectid + block_group->key.offset; | 8840 | last = block_group->key.objectid + block_group->key.offset; |
| 8841 | btrfs_put_block_group(block_group); | 8841 | btrfs_put_block_group(block_group); |
| 8842 | } | 8842 | } |
| 8843 | } | 8843 | } |
| 8844 | 8844 | ||
| 8845 | int btrfs_free_block_groups(struct btrfs_fs_info *info) | 8845 | int btrfs_free_block_groups(struct btrfs_fs_info *info) |
| 8846 | { | 8846 | { |
| 8847 | struct btrfs_block_group_cache *block_group; | 8847 | struct btrfs_block_group_cache *block_group; |
| 8848 | struct btrfs_space_info *space_info; | 8848 | struct btrfs_space_info *space_info; |
| 8849 | struct btrfs_caching_control *caching_ctl; | 8849 | struct btrfs_caching_control *caching_ctl; |
| 8850 | struct rb_node *n; | 8850 | struct rb_node *n; |
| 8851 | 8851 | ||
| 8852 | down_write(&info->commit_root_sem); | 8852 | down_write(&info->commit_root_sem); |
| 8853 | while (!list_empty(&info->caching_block_groups)) { | 8853 | while (!list_empty(&info->caching_block_groups)) { |
| 8854 | caching_ctl = list_entry(info->caching_block_groups.next, | 8854 | caching_ctl = list_entry(info->caching_block_groups.next, |
| 8855 | struct btrfs_caching_control, list); | 8855 | struct btrfs_caching_control, list); |
| 8856 | list_del(&caching_ctl->list); | 8856 | list_del(&caching_ctl->list); |
| 8857 | put_caching_control(caching_ctl); | 8857 | put_caching_control(caching_ctl); |
| 8858 | } | 8858 | } |
| 8859 | up_write(&info->commit_root_sem); | 8859 | up_write(&info->commit_root_sem); |
| 8860 | 8860 | ||
| 8861 | spin_lock(&info->unused_bgs_lock); | 8861 | spin_lock(&info->unused_bgs_lock); |
| 8862 | while (!list_empty(&info->unused_bgs)) { | 8862 | while (!list_empty(&info->unused_bgs)) { |
| 8863 | block_group = list_first_entry(&info->unused_bgs, | 8863 | block_group = list_first_entry(&info->unused_bgs, |
| 8864 | struct btrfs_block_group_cache, | 8864 | struct btrfs_block_group_cache, |
| 8865 | bg_list); | 8865 | bg_list); |
| 8866 | list_del_init(&block_group->bg_list); | 8866 | list_del_init(&block_group->bg_list); |
| 8867 | btrfs_put_block_group(block_group); | 8867 | btrfs_put_block_group(block_group); |
| 8868 | } | 8868 | } |
| 8869 | spin_unlock(&info->unused_bgs_lock); | 8869 | spin_unlock(&info->unused_bgs_lock); |
| 8870 | 8870 | ||
| 8871 | spin_lock(&info->block_group_cache_lock); | 8871 | spin_lock(&info->block_group_cache_lock); |
| 8872 | while ((n = rb_last(&info->block_group_cache_tree)) != NULL) { | 8872 | while ((n = rb_last(&info->block_group_cache_tree)) != NULL) { |
| 8873 | block_group = rb_entry(n, struct btrfs_block_group_cache, | 8873 | block_group = rb_entry(n, struct btrfs_block_group_cache, |
| 8874 | cache_node); | 8874 | cache_node); |
| 8875 | rb_erase(&block_group->cache_node, | 8875 | rb_erase(&block_group->cache_node, |
| 8876 | &info->block_group_cache_tree); | 8876 | &info->block_group_cache_tree); |
| 8877 | RB_CLEAR_NODE(&block_group->cache_node); | 8877 | RB_CLEAR_NODE(&block_group->cache_node); |
| 8878 | spin_unlock(&info->block_group_cache_lock); | 8878 | spin_unlock(&info->block_group_cache_lock); |
| 8879 | 8879 | ||
| 8880 | down_write(&block_group->space_info->groups_sem); | 8880 | down_write(&block_group->space_info->groups_sem); |
| 8881 | list_del(&block_group->list); | 8881 | list_del(&block_group->list); |
| 8882 | up_write(&block_group->space_info->groups_sem); | 8882 | up_write(&block_group->space_info->groups_sem); |
| 8883 | 8883 | ||
| 8884 | if (block_group->cached == BTRFS_CACHE_STARTED) | 8884 | if (block_group->cached == BTRFS_CACHE_STARTED) |
| 8885 | wait_block_group_cache_done(block_group); | 8885 | wait_block_group_cache_done(block_group); |
| 8886 | 8886 | ||
| 8887 | /* | 8887 | /* |
| 8888 | * We haven't cached this block group, which means we could | 8888 | * We haven't cached this block group, which means we could |
| 8889 | * possibly have excluded extents on this block group. | 8889 | * possibly have excluded extents on this block group. |
| 8890 | */ | 8890 | */ |
| 8891 | if (block_group->cached == BTRFS_CACHE_NO || | 8891 | if (block_group->cached == BTRFS_CACHE_NO || |
| 8892 | block_group->cached == BTRFS_CACHE_ERROR) | 8892 | block_group->cached == BTRFS_CACHE_ERROR) |
| 8893 | free_excluded_extents(info->extent_root, block_group); | 8893 | free_excluded_extents(info->extent_root, block_group); |
| 8894 | 8894 | ||
| 8895 | btrfs_remove_free_space_cache(block_group); | 8895 | btrfs_remove_free_space_cache(block_group); |
| 8896 | btrfs_put_block_group(block_group); | 8896 | btrfs_put_block_group(block_group); |
| 8897 | 8897 | ||
| 8898 | spin_lock(&info->block_group_cache_lock); | 8898 | spin_lock(&info->block_group_cache_lock); |
| 8899 | } | 8899 | } |
| 8900 | spin_unlock(&info->block_group_cache_lock); | 8900 | spin_unlock(&info->block_group_cache_lock); |
| 8901 | 8901 | ||
| 8902 | /* now that all the block groups are freed, go through and | 8902 | /* now that all the block groups are freed, go through and |
| 8903 | * free all the space_info structs. This is only called during | 8903 | * free all the space_info structs. This is only called during |
| 8904 | * the final stages of unmount, and so we know nobody is | 8904 | * the final stages of unmount, and so we know nobody is |
| 8905 | * using them. We call synchronize_rcu() once before we start, | 8905 | * using them. We call synchronize_rcu() once before we start, |
| 8906 | * just to be on the safe side. | 8906 | * just to be on the safe side. |
| 8907 | */ | 8907 | */ |
| 8908 | synchronize_rcu(); | 8908 | synchronize_rcu(); |
| 8909 | 8909 | ||
| 8910 | release_global_block_rsv(info); | 8910 | release_global_block_rsv(info); |
| 8911 | 8911 | ||
| 8912 | while (!list_empty(&info->space_info)) { | 8912 | while (!list_empty(&info->space_info)) { |
| 8913 | int i; | 8913 | int i; |
| 8914 | 8914 | ||
| 8915 | space_info = list_entry(info->space_info.next, | 8915 | space_info = list_entry(info->space_info.next, |
| 8916 | struct btrfs_space_info, | 8916 | struct btrfs_space_info, |
| 8917 | list); | 8917 | list); |
| 8918 | if (btrfs_test_opt(info->tree_root, ENOSPC_DEBUG)) { | 8918 | if (btrfs_test_opt(info->tree_root, ENOSPC_DEBUG)) { |
| 8919 | if (WARN_ON(space_info->bytes_pinned > 0 || | 8919 | if (WARN_ON(space_info->bytes_pinned > 0 || |
| 8920 | space_info->bytes_reserved > 0 || | 8920 | space_info->bytes_reserved > 0 || |
| 8921 | space_info->bytes_may_use > 0)) { | 8921 | space_info->bytes_may_use > 0)) { |
| 8922 | dump_space_info(space_info, 0, 0); | 8922 | dump_space_info(space_info, 0, 0); |
| 8923 | } | 8923 | } |
| 8924 | } | 8924 | } |
| 8925 | list_del(&space_info->list); | 8925 | list_del(&space_info->list); |
| 8926 | for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) { | 8926 | for (i = 0; i < BTRFS_NR_RAID_TYPES; i++) { |
| 8927 | struct kobject *kobj; | 8927 | struct kobject *kobj; |
| 8928 | kobj = space_info->block_group_kobjs[i]; | 8928 | kobj = space_info->block_group_kobjs[i]; |
| 8929 | space_info->block_group_kobjs[i] = NULL; | 8929 | space_info->block_group_kobjs[i] = NULL; |
| 8930 | if (kobj) { | 8930 | if (kobj) { |
| 8931 | kobject_del(kobj); | 8931 | kobject_del(kobj); |
| 8932 | kobject_put(kobj); | 8932 | kobject_put(kobj); |
| 8933 | } | 8933 | } |
| 8934 | } | 8934 | } |
| 8935 | kobject_del(&space_info->kobj); | 8935 | kobject_del(&space_info->kobj); |
| 8936 | kobject_put(&space_info->kobj); | 8936 | kobject_put(&space_info->kobj); |
| 8937 | } | 8937 | } |
| 8938 | return 0; | 8938 | return 0; |
| 8939 | } | 8939 | } |
| 8940 | 8940 | ||
| 8941 | static void __link_block_group(struct btrfs_space_info *space_info, | 8941 | static void __link_block_group(struct btrfs_space_info *space_info, |
| 8942 | struct btrfs_block_group_cache *cache) | 8942 | struct btrfs_block_group_cache *cache) |
| 8943 | { | 8943 | { |
| 8944 | int index = get_block_group_index(cache); | 8944 | int index = get_block_group_index(cache); |
| 8945 | bool first = false; | 8945 | bool first = false; |
| 8946 | 8946 | ||
| 8947 | down_write(&space_info->groups_sem); | 8947 | down_write(&space_info->groups_sem); |
| 8948 | if (list_empty(&space_info->block_groups[index])) | 8948 | if (list_empty(&space_info->block_groups[index])) |
| 8949 | first = true; | 8949 | first = true; |
| 8950 | list_add_tail(&cache->list, &space_info->block_groups[index]); | 8950 | list_add_tail(&cache->list, &space_info->block_groups[index]); |
| 8951 | up_write(&space_info->groups_sem); | 8951 | up_write(&space_info->groups_sem); |
| 8952 | 8952 | ||
| 8953 | if (first) { | 8953 | if (first) { |
| 8954 | struct raid_kobject *rkobj; | 8954 | struct raid_kobject *rkobj; |
| 8955 | int ret; | 8955 | int ret; |
| 8956 | 8956 | ||
| 8957 | rkobj = kzalloc(sizeof(*rkobj), GFP_NOFS); | 8957 | rkobj = kzalloc(sizeof(*rkobj), GFP_NOFS); |
| 8958 | if (!rkobj) | 8958 | if (!rkobj) |
| 8959 | goto out_err; | 8959 | goto out_err; |
| 8960 | rkobj->raid_type = index; | 8960 | rkobj->raid_type = index; |
| 8961 | kobject_init(&rkobj->kobj, &btrfs_raid_ktype); | 8961 | kobject_init(&rkobj->kobj, &btrfs_raid_ktype); |
| 8962 | ret = kobject_add(&rkobj->kobj, &space_info->kobj, | 8962 | ret = kobject_add(&rkobj->kobj, &space_info->kobj, |
| 8963 | "%s", get_raid_name(index)); | 8963 | "%s", get_raid_name(index)); |
| 8964 | if (ret) { | 8964 | if (ret) { |
| 8965 | kobject_put(&rkobj->kobj); | 8965 | kobject_put(&rkobj->kobj); |
| 8966 | goto out_err; | 8966 | goto out_err; |
| 8967 | } | 8967 | } |
| 8968 | space_info->block_group_kobjs[index] = &rkobj->kobj; | 8968 | space_info->block_group_kobjs[index] = &rkobj->kobj; |
| 8969 | } | 8969 | } |
| 8970 | 8970 | ||
| 8971 | return; | 8971 | return; |
| 8972 | out_err: | 8972 | out_err: |
| 8973 | pr_warn("BTRFS: failed to add kobject for block cache. ignoring.\n"); | 8973 | pr_warn("BTRFS: failed to add kobject for block cache. ignoring.\n"); |
| 8974 | } | 8974 | } |
| 8975 | 8975 | ||
| 8976 | static struct btrfs_block_group_cache * | 8976 | static struct btrfs_block_group_cache * |
| 8977 | btrfs_create_block_group_cache(struct btrfs_root *root, u64 start, u64 size) | 8977 | btrfs_create_block_group_cache(struct btrfs_root *root, u64 start, u64 size) |
| 8978 | { | 8978 | { |
| 8979 | struct btrfs_block_group_cache *cache; | 8979 | struct btrfs_block_group_cache *cache; |
| 8980 | 8980 | ||
| 8981 | cache = kzalloc(sizeof(*cache), GFP_NOFS); | 8981 | cache = kzalloc(sizeof(*cache), GFP_NOFS); |
| 8982 | if (!cache) | 8982 | if (!cache) |
| 8983 | return NULL; | 8983 | return NULL; |
| 8984 | 8984 | ||
| 8985 | cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl), | 8985 | cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl), |
| 8986 | GFP_NOFS); | 8986 | GFP_NOFS); |
| 8987 | if (!cache->free_space_ctl) { | 8987 | if (!cache->free_space_ctl) { |
| 8988 | kfree(cache); | 8988 | kfree(cache); |
| 8989 | return NULL; | 8989 | return NULL; |
| 8990 | } | 8990 | } |
| 8991 | 8991 | ||
| 8992 | cache->key.objectid = start; | 8992 | cache->key.objectid = start; |
| 8993 | cache->key.offset = size; | 8993 | cache->key.offset = size; |
| 8994 | cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY; | 8994 | cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY; |
| 8995 | 8995 | ||
| 8996 | cache->sectorsize = root->sectorsize; | 8996 | cache->sectorsize = root->sectorsize; |
| 8997 | cache->fs_info = root->fs_info; | 8997 | cache->fs_info = root->fs_info; |
| 8998 | cache->full_stripe_len = btrfs_full_stripe_len(root, | 8998 | cache->full_stripe_len = btrfs_full_stripe_len(root, |
| 8999 | &root->fs_info->mapping_tree, | 8999 | &root->fs_info->mapping_tree, |
| 9000 | start); | 9000 | start); |
| 9001 | atomic_set(&cache->count, 1); | 9001 | atomic_set(&cache->count, 1); |
| 9002 | spin_lock_init(&cache->lock); | 9002 | spin_lock_init(&cache->lock); |
| 9003 | init_rwsem(&cache->data_rwsem); | 9003 | init_rwsem(&cache->data_rwsem); |
| 9004 | INIT_LIST_HEAD(&cache->list); | 9004 | INIT_LIST_HEAD(&cache->list); |
| 9005 | INIT_LIST_HEAD(&cache->cluster_list); | 9005 | INIT_LIST_HEAD(&cache->cluster_list); |
| 9006 | INIT_LIST_HEAD(&cache->bg_list); | 9006 | INIT_LIST_HEAD(&cache->bg_list); |
| 9007 | INIT_LIST_HEAD(&cache->ro_list); | 9007 | INIT_LIST_HEAD(&cache->ro_list); |
| 9008 | btrfs_init_free_space_ctl(cache); | 9008 | btrfs_init_free_space_ctl(cache); |
| 9009 | atomic_set(&cache->trimming, 0); | 9009 | atomic_set(&cache->trimming, 0); |
| 9010 | 9010 | ||
| 9011 | return cache; | 9011 | return cache; |
| 9012 | } | 9012 | } |
| 9013 | 9013 | ||
| 9014 | int btrfs_read_block_groups(struct btrfs_root *root) | 9014 | int btrfs_read_block_groups(struct btrfs_root *root) |
| 9015 | { | 9015 | { |
| 9016 | struct btrfs_path *path; | 9016 | struct btrfs_path *path; |
| 9017 | int ret; | 9017 | int ret; |
| 9018 | struct btrfs_block_group_cache *cache; | 9018 | struct btrfs_block_group_cache *cache; |
| 9019 | struct btrfs_fs_info *info = root->fs_info; | 9019 | struct btrfs_fs_info *info = root->fs_info; |
| 9020 | struct btrfs_space_info *space_info; | 9020 | struct btrfs_space_info *space_info; |
| 9021 | struct btrfs_key key; | 9021 | struct btrfs_key key; |
| 9022 | struct btrfs_key found_key; | 9022 | struct btrfs_key found_key; |
| 9023 | struct extent_buffer *leaf; | 9023 | struct extent_buffer *leaf; |
| 9024 | int need_clear = 0; | 9024 | int need_clear = 0; |
| 9025 | u64 cache_gen; | 9025 | u64 cache_gen; |
| 9026 | 9026 | ||
| 9027 | root = info->extent_root; | 9027 | root = info->extent_root; |
| 9028 | key.objectid = 0; | 9028 | key.objectid = 0; |
| 9029 | key.offset = 0; | 9029 | key.offset = 0; |
| 9030 | key.type = BTRFS_BLOCK_GROUP_ITEM_KEY; | 9030 | key.type = BTRFS_BLOCK_GROUP_ITEM_KEY; |
| 9031 | path = btrfs_alloc_path(); | 9031 | path = btrfs_alloc_path(); |
| 9032 | if (!path) | 9032 | if (!path) |
| 9033 | return -ENOMEM; | 9033 | return -ENOMEM; |
| 9034 | path->reada = 1; | 9034 | path->reada = 1; |
| 9035 | 9035 | ||
| 9036 | cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); | 9036 | cache_gen = btrfs_super_cache_generation(root->fs_info->super_copy); |
| 9037 | if (btrfs_test_opt(root, SPACE_CACHE) && | 9037 | if (btrfs_test_opt(root, SPACE_CACHE) && |
| 9038 | btrfs_super_generation(root->fs_info->super_copy) != cache_gen) | 9038 | btrfs_super_generation(root->fs_info->super_copy) != cache_gen) |
| 9039 | need_clear = 1; | 9039 | need_clear = 1; |
| 9040 | if (btrfs_test_opt(root, CLEAR_CACHE)) | 9040 | if (btrfs_test_opt(root, CLEAR_CACHE)) |
| 9041 | need_clear = 1; | 9041 | need_clear = 1; |
| 9042 | 9042 | ||
| 9043 | while (1) { | 9043 | while (1) { |
| 9044 | ret = find_first_block_group(root, path, &key); | 9044 | ret = find_first_block_group(root, path, &key); |
| 9045 | if (ret > 0) | 9045 | if (ret > 0) |
| 9046 | break; | 9046 | break; |
| 9047 | if (ret != 0) | 9047 | if (ret != 0) |
| 9048 | goto error; | 9048 | goto error; |
| 9049 | 9049 | ||
| 9050 | leaf = path->nodes[0]; | 9050 | leaf = path->nodes[0]; |
| 9051 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 9051 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 9052 | 9052 | ||
| 9053 | cache = btrfs_create_block_group_cache(root, found_key.objectid, | 9053 | cache = btrfs_create_block_group_cache(root, found_key.objectid, |
| 9054 | found_key.offset); | 9054 | found_key.offset); |
| 9055 | if (!cache) { | 9055 | if (!cache) { |
| 9056 | ret = -ENOMEM; | 9056 | ret = -ENOMEM; |
| 9057 | goto error; | 9057 | goto error; |
| 9058 | } | 9058 | } |
| 9059 | 9059 | ||
| 9060 | if (need_clear) { | 9060 | if (need_clear) { |
| 9061 | /* | 9061 | /* |
| 9062 | * When we mount with old space cache, we need to | 9062 | * When we mount with old space cache, we need to |
| 9063 | * set BTRFS_DC_CLEAR and set dirty flag. | 9063 | * set BTRFS_DC_CLEAR and set dirty flag. |
| 9064 | * | 9064 | * |
| 9065 | * a) Setting 'BTRFS_DC_CLEAR' makes sure that we | 9065 | * a) Setting 'BTRFS_DC_CLEAR' makes sure that we |
| 9066 | * truncate the old free space cache inode and | 9066 | * truncate the old free space cache inode and |
| 9067 | * setup a new one. | 9067 | * setup a new one. |
| 9068 | * b) Setting 'dirty flag' makes sure that we flush | 9068 | * b) Setting 'dirty flag' makes sure that we flush |
| 9069 | * the new space cache info onto disk. | 9069 | * the new space cache info onto disk. |
| 9070 | */ | 9070 | */ |
| 9071 | cache->disk_cache_state = BTRFS_DC_CLEAR; | 9071 | cache->disk_cache_state = BTRFS_DC_CLEAR; |
| 9072 | if (btrfs_test_opt(root, SPACE_CACHE)) | 9072 | if (btrfs_test_opt(root, SPACE_CACHE)) |
| 9073 | cache->dirty = 1; | 9073 | cache->dirty = 1; |
| 9074 | } | 9074 | } |
| 9075 | 9075 | ||
| 9076 | read_extent_buffer(leaf, &cache->item, | 9076 | read_extent_buffer(leaf, &cache->item, |
| 9077 | btrfs_item_ptr_offset(leaf, path->slots[0]), | 9077 | btrfs_item_ptr_offset(leaf, path->slots[0]), |
| 9078 | sizeof(cache->item)); | 9078 | sizeof(cache->item)); |
| 9079 | cache->flags = btrfs_block_group_flags(&cache->item); | 9079 | cache->flags = btrfs_block_group_flags(&cache->item); |
| 9080 | 9080 | ||
| 9081 | key.objectid = found_key.objectid + found_key.offset; | 9081 | key.objectid = found_key.objectid + found_key.offset; |
| 9082 | btrfs_release_path(path); | 9082 | btrfs_release_path(path); |
| 9083 | 9083 | ||
| 9084 | /* | 9084 | /* |
| 9085 | * We need to exclude the super stripes now so that the space | 9085 | * We need to exclude the super stripes now so that the space |
| 9086 | * info has super bytes accounted for, otherwise we'll think | 9086 | * info has super bytes accounted for, otherwise we'll think |
| 9087 | * we have more space than we actually do. | 9087 | * we have more space than we actually do. |
| 9088 | */ | 9088 | */ |
| 9089 | ret = exclude_super_stripes(root, cache); | 9089 | ret = exclude_super_stripes(root, cache); |
| 9090 | if (ret) { | 9090 | if (ret) { |
| 9091 | /* | 9091 | /* |
| 9092 | * We may have excluded something, so call this just in | 9092 | * We may have excluded something, so call this just in |
| 9093 | * case. | 9093 | * case. |
| 9094 | */ | 9094 | */ |
| 9095 | free_excluded_extents(root, cache); | 9095 | free_excluded_extents(root, cache); |
| 9096 | btrfs_put_block_group(cache); | 9096 | btrfs_put_block_group(cache); |
| 9097 | goto error; | 9097 | goto error; |
| 9098 | } | 9098 | } |
| 9099 | 9099 | ||
| 9100 | /* | 9100 | /* |
| 9101 | * check for two cases, either we are full, and therefore | 9101 | * check for two cases, either we are full, and therefore |
| 9102 | * don't need to bother with the caching work since we won't | 9102 | * don't need to bother with the caching work since we won't |
| 9103 | * find any space, or we are empty, and we can just add all | 9103 | * find any space, or we are empty, and we can just add all |
| 9104 | * the space in and be done with it. This saves us _alot_ of | 9104 | * the space in and be done with it. This saves us _alot_ of |
| 9105 | * time, particularly in the full case. | 9105 | * time, particularly in the full case. |
| 9106 | */ | 9106 | */ |
| 9107 | if (found_key.offset == btrfs_block_group_used(&cache->item)) { | 9107 | if (found_key.offset == btrfs_block_group_used(&cache->item)) { |
| 9108 | cache->last_byte_to_unpin = (u64)-1; | 9108 | cache->last_byte_to_unpin = (u64)-1; |
| 9109 | cache->cached = BTRFS_CACHE_FINISHED; | 9109 | cache->cached = BTRFS_CACHE_FINISHED; |
| 9110 | free_excluded_extents(root, cache); | 9110 | free_excluded_extents(root, cache); |
| 9111 | } else if (btrfs_block_group_used(&cache->item) == 0) { | 9111 | } else if (btrfs_block_group_used(&cache->item) == 0) { |
| 9112 | cache->last_byte_to_unpin = (u64)-1; | 9112 | cache->last_byte_to_unpin = (u64)-1; |
| 9113 | cache->cached = BTRFS_CACHE_FINISHED; | 9113 | cache->cached = BTRFS_CACHE_FINISHED; |
| 9114 | add_new_free_space(cache, root->fs_info, | 9114 | add_new_free_space(cache, root->fs_info, |
| 9115 | found_key.objectid, | 9115 | found_key.objectid, |
| 9116 | found_key.objectid + | 9116 | found_key.objectid + |
| 9117 | found_key.offset); | 9117 | found_key.offset); |
| 9118 | free_excluded_extents(root, cache); | 9118 | free_excluded_extents(root, cache); |
| 9119 | } | 9119 | } |
| 9120 | 9120 | ||
| 9121 | ret = btrfs_add_block_group_cache(root->fs_info, cache); | 9121 | ret = btrfs_add_block_group_cache(root->fs_info, cache); |
| 9122 | if (ret) { | 9122 | if (ret) { |
| 9123 | btrfs_remove_free_space_cache(cache); | 9123 | btrfs_remove_free_space_cache(cache); |
| 9124 | btrfs_put_block_group(cache); | 9124 | btrfs_put_block_group(cache); |
| 9125 | goto error; | 9125 | goto error; |
| 9126 | } | 9126 | } |
| 9127 | 9127 | ||
| 9128 | ret = update_space_info(info, cache->flags, found_key.offset, | 9128 | ret = update_space_info(info, cache->flags, found_key.offset, |
| 9129 | btrfs_block_group_used(&cache->item), | 9129 | btrfs_block_group_used(&cache->item), |
| 9130 | &space_info); | 9130 | &space_info); |
| 9131 | if (ret) { | 9131 | if (ret) { |
| 9132 | btrfs_remove_free_space_cache(cache); | 9132 | btrfs_remove_free_space_cache(cache); |
| 9133 | spin_lock(&info->block_group_cache_lock); | 9133 | spin_lock(&info->block_group_cache_lock); |
| 9134 | rb_erase(&cache->cache_node, | 9134 | rb_erase(&cache->cache_node, |
| 9135 | &info->block_group_cache_tree); | 9135 | &info->block_group_cache_tree); |
| 9136 | RB_CLEAR_NODE(&cache->cache_node); | 9136 | RB_CLEAR_NODE(&cache->cache_node); |
| 9137 | spin_unlock(&info->block_group_cache_lock); | 9137 | spin_unlock(&info->block_group_cache_lock); |
| 9138 | btrfs_put_block_group(cache); | 9138 | btrfs_put_block_group(cache); |
| 9139 | goto error; | 9139 | goto error; |
| 9140 | } | 9140 | } |
| 9141 | 9141 | ||
| 9142 | cache->space_info = space_info; | 9142 | cache->space_info = space_info; |
| 9143 | spin_lock(&cache->space_info->lock); | 9143 | spin_lock(&cache->space_info->lock); |
| 9144 | cache->space_info->bytes_readonly += cache->bytes_super; | 9144 | cache->space_info->bytes_readonly += cache->bytes_super; |
| 9145 | spin_unlock(&cache->space_info->lock); | 9145 | spin_unlock(&cache->space_info->lock); |
| 9146 | 9146 | ||
| 9147 | __link_block_group(space_info, cache); | 9147 | __link_block_group(space_info, cache); |
| 9148 | 9148 | ||
| 9149 | set_avail_alloc_bits(root->fs_info, cache->flags); | 9149 | set_avail_alloc_bits(root->fs_info, cache->flags); |
| 9150 | if (btrfs_chunk_readonly(root, cache->key.objectid)) { | 9150 | if (btrfs_chunk_readonly(root, cache->key.objectid)) { |
| 9151 | set_block_group_ro(cache, 1); | 9151 | set_block_group_ro(cache, 1); |
| 9152 | } else if (btrfs_block_group_used(&cache->item) == 0) { | 9152 | } else if (btrfs_block_group_used(&cache->item) == 0) { |
| 9153 | spin_lock(&info->unused_bgs_lock); | 9153 | spin_lock(&info->unused_bgs_lock); |
| 9154 | /* Should always be true but just in case. */ | 9154 | /* Should always be true but just in case. */ |
| 9155 | if (list_empty(&cache->bg_list)) { | 9155 | if (list_empty(&cache->bg_list)) { |
| 9156 | btrfs_get_block_group(cache); | 9156 | btrfs_get_block_group(cache); |
| 9157 | list_add_tail(&cache->bg_list, | 9157 | list_add_tail(&cache->bg_list, |
| 9158 | &info->unused_bgs); | 9158 | &info->unused_bgs); |
| 9159 | } | 9159 | } |
| 9160 | spin_unlock(&info->unused_bgs_lock); | 9160 | spin_unlock(&info->unused_bgs_lock); |
| 9161 | } | 9161 | } |
| 9162 | } | 9162 | } |
| 9163 | 9163 | ||
| 9164 | list_for_each_entry_rcu(space_info, &root->fs_info->space_info, list) { | 9164 | list_for_each_entry_rcu(space_info, &root->fs_info->space_info, list) { |
| 9165 | if (!(get_alloc_profile(root, space_info->flags) & | 9165 | if (!(get_alloc_profile(root, space_info->flags) & |
| 9166 | (BTRFS_BLOCK_GROUP_RAID10 | | 9166 | (BTRFS_BLOCK_GROUP_RAID10 | |
| 9167 | BTRFS_BLOCK_GROUP_RAID1 | | 9167 | BTRFS_BLOCK_GROUP_RAID1 | |
| 9168 | BTRFS_BLOCK_GROUP_RAID5 | | 9168 | BTRFS_BLOCK_GROUP_RAID5 | |
| 9169 | BTRFS_BLOCK_GROUP_RAID6 | | 9169 | BTRFS_BLOCK_GROUP_RAID6 | |
| 9170 | BTRFS_BLOCK_GROUP_DUP))) | 9170 | BTRFS_BLOCK_GROUP_DUP))) |
| 9171 | continue; | 9171 | continue; |
| 9172 | /* | 9172 | /* |
| 9173 | * avoid allocating from un-mirrored block group if there are | 9173 | * avoid allocating from un-mirrored block group if there are |
| 9174 | * mirrored block groups. | 9174 | * mirrored block groups. |
| 9175 | */ | 9175 | */ |
| 9176 | list_for_each_entry(cache, | 9176 | list_for_each_entry(cache, |
| 9177 | &space_info->block_groups[BTRFS_RAID_RAID0], | 9177 | &space_info->block_groups[BTRFS_RAID_RAID0], |
| 9178 | list) | 9178 | list) |
| 9179 | set_block_group_ro(cache, 1); | 9179 | set_block_group_ro(cache, 1); |
| 9180 | list_for_each_entry(cache, | 9180 | list_for_each_entry(cache, |
| 9181 | &space_info->block_groups[BTRFS_RAID_SINGLE], | 9181 | &space_info->block_groups[BTRFS_RAID_SINGLE], |
| 9182 | list) | 9182 | list) |
| 9183 | set_block_group_ro(cache, 1); | 9183 | set_block_group_ro(cache, 1); |
| 9184 | } | 9184 | } |
| 9185 | 9185 | ||
| 9186 | init_global_block_rsv(info); | 9186 | init_global_block_rsv(info); |
| 9187 | ret = 0; | 9187 | ret = 0; |
| 9188 | error: | 9188 | error: |
| 9189 | btrfs_free_path(path); | 9189 | btrfs_free_path(path); |
| 9190 | return ret; | 9190 | return ret; |
| 9191 | } | 9191 | } |
| 9192 | 9192 | ||
| 9193 | void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans, | 9193 | void btrfs_create_pending_block_groups(struct btrfs_trans_handle *trans, |
| 9194 | struct btrfs_root *root) | 9194 | struct btrfs_root *root) |
| 9195 | { | 9195 | { |
| 9196 | struct btrfs_block_group_cache *block_group, *tmp; | 9196 | struct btrfs_block_group_cache *block_group, *tmp; |
| 9197 | struct btrfs_root *extent_root = root->fs_info->extent_root; | 9197 | struct btrfs_root *extent_root = root->fs_info->extent_root; |
| 9198 | struct btrfs_block_group_item item; | 9198 | struct btrfs_block_group_item item; |
| 9199 | struct btrfs_key key; | 9199 | struct btrfs_key key; |
| 9200 | int ret = 0; | 9200 | int ret = 0; |
| 9201 | 9201 | ||
| 9202 | list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) { | 9202 | list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) { |
| 9203 | if (ret) | 9203 | if (ret) |
| 9204 | goto next; | 9204 | goto next; |
| 9205 | 9205 | ||
| 9206 | spin_lock(&block_group->lock); | 9206 | spin_lock(&block_group->lock); |
| 9207 | memcpy(&item, &block_group->item, sizeof(item)); | 9207 | memcpy(&item, &block_group->item, sizeof(item)); |
| 9208 | memcpy(&key, &block_group->key, sizeof(key)); | 9208 | memcpy(&key, &block_group->key, sizeof(key)); |
| 9209 | spin_unlock(&block_group->lock); | 9209 | spin_unlock(&block_group->lock); |
| 9210 | 9210 | ||
| 9211 | ret = btrfs_insert_item(trans, extent_root, &key, &item, | 9211 | ret = btrfs_insert_item(trans, extent_root, &key, &item, |
| 9212 | sizeof(item)); | 9212 | sizeof(item)); |
| 9213 | if (ret) | 9213 | if (ret) |
| 9214 | btrfs_abort_transaction(trans, extent_root, ret); | 9214 | btrfs_abort_transaction(trans, extent_root, ret); |
| 9215 | ret = btrfs_finish_chunk_alloc(trans, extent_root, | 9215 | ret = btrfs_finish_chunk_alloc(trans, extent_root, |
| 9216 | key.objectid, key.offset); | 9216 | key.objectid, key.offset); |
| 9217 | if (ret) | 9217 | if (ret) |
| 9218 | btrfs_abort_transaction(trans, extent_root, ret); | 9218 | btrfs_abort_transaction(trans, extent_root, ret); |
| 9219 | next: | 9219 | next: |
| 9220 | list_del_init(&block_group->bg_list); | 9220 | list_del_init(&block_group->bg_list); |
| 9221 | } | 9221 | } |
| 9222 | } | 9222 | } |
| 9223 | 9223 | ||
| 9224 | int btrfs_make_block_group(struct btrfs_trans_handle *trans, | 9224 | int btrfs_make_block_group(struct btrfs_trans_handle *trans, |
| 9225 | struct btrfs_root *root, u64 bytes_used, | 9225 | struct btrfs_root *root, u64 bytes_used, |
| 9226 | u64 type, u64 chunk_objectid, u64 chunk_offset, | 9226 | u64 type, u64 chunk_objectid, u64 chunk_offset, |
| 9227 | u64 size) | 9227 | u64 size) |
| 9228 | { | 9228 | { |
| 9229 | int ret; | 9229 | int ret; |
| 9230 | struct btrfs_root *extent_root; | 9230 | struct btrfs_root *extent_root; |
| 9231 | struct btrfs_block_group_cache *cache; | 9231 | struct btrfs_block_group_cache *cache; |
| 9232 | 9232 | ||
| 9233 | extent_root = root->fs_info->extent_root; | 9233 | extent_root = root->fs_info->extent_root; |
| 9234 | 9234 | ||
| 9235 | btrfs_set_log_full_commit(root->fs_info, trans); | 9235 | btrfs_set_log_full_commit(root->fs_info, trans); |
| 9236 | 9236 | ||
| 9237 | cache = btrfs_create_block_group_cache(root, chunk_offset, size); | 9237 | cache = btrfs_create_block_group_cache(root, chunk_offset, size); |
| 9238 | if (!cache) | 9238 | if (!cache) |
| 9239 | return -ENOMEM; | 9239 | return -ENOMEM; |
| 9240 | 9240 | ||
| 9241 | btrfs_set_block_group_used(&cache->item, bytes_used); | 9241 | btrfs_set_block_group_used(&cache->item, bytes_used); |
| 9242 | btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid); | 9242 | btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid); |
| 9243 | btrfs_set_block_group_flags(&cache->item, type); | 9243 | btrfs_set_block_group_flags(&cache->item, type); |
| 9244 | 9244 | ||
| 9245 | cache->flags = type; | 9245 | cache->flags = type; |
| 9246 | cache->last_byte_to_unpin = (u64)-1; | 9246 | cache->last_byte_to_unpin = (u64)-1; |
| 9247 | cache->cached = BTRFS_CACHE_FINISHED; | 9247 | cache->cached = BTRFS_CACHE_FINISHED; |
| 9248 | ret = exclude_super_stripes(root, cache); | 9248 | ret = exclude_super_stripes(root, cache); |
| 9249 | if (ret) { | 9249 | if (ret) { |
| 9250 | /* | 9250 | /* |
| 9251 | * We may have excluded something, so call this just in | 9251 | * We may have excluded something, so call this just in |
| 9252 | * case. | 9252 | * case. |
| 9253 | */ | 9253 | */ |
| 9254 | free_excluded_extents(root, cache); | 9254 | free_excluded_extents(root, cache); |
| 9255 | btrfs_put_block_group(cache); | 9255 | btrfs_put_block_group(cache); |
| 9256 | return ret; | 9256 | return ret; |
| 9257 | } | 9257 | } |
| 9258 | 9258 | ||
| 9259 | add_new_free_space(cache, root->fs_info, chunk_offset, | 9259 | add_new_free_space(cache, root->fs_info, chunk_offset, |
| 9260 | chunk_offset + size); | 9260 | chunk_offset + size); |
| 9261 | 9261 | ||
| 9262 | free_excluded_extents(root, cache); | 9262 | free_excluded_extents(root, cache); |
| 9263 | 9263 | ||
| 9264 | ret = btrfs_add_block_group_cache(root->fs_info, cache); | 9264 | ret = btrfs_add_block_group_cache(root->fs_info, cache); |
| 9265 | if (ret) { | 9265 | if (ret) { |
| 9266 | btrfs_remove_free_space_cache(cache); | 9266 | btrfs_remove_free_space_cache(cache); |
| 9267 | btrfs_put_block_group(cache); | 9267 | btrfs_put_block_group(cache); |
| 9268 | return ret; | 9268 | return ret; |
| 9269 | } | 9269 | } |
| 9270 | 9270 | ||
| 9271 | ret = update_space_info(root->fs_info, cache->flags, size, bytes_used, | 9271 | ret = update_space_info(root->fs_info, cache->flags, size, bytes_used, |
| 9272 | &cache->space_info); | 9272 | &cache->space_info); |
| 9273 | if (ret) { | 9273 | if (ret) { |
| 9274 | btrfs_remove_free_space_cache(cache); | 9274 | btrfs_remove_free_space_cache(cache); |
| 9275 | spin_lock(&root->fs_info->block_group_cache_lock); | 9275 | spin_lock(&root->fs_info->block_group_cache_lock); |
| 9276 | rb_erase(&cache->cache_node, | 9276 | rb_erase(&cache->cache_node, |
| 9277 | &root->fs_info->block_group_cache_tree); | 9277 | &root->fs_info->block_group_cache_tree); |
| 9278 | RB_CLEAR_NODE(&cache->cache_node); | 9278 | RB_CLEAR_NODE(&cache->cache_node); |
| 9279 | spin_unlock(&root->fs_info->block_group_cache_lock); | 9279 | spin_unlock(&root->fs_info->block_group_cache_lock); |
| 9280 | btrfs_put_block_group(cache); | 9280 | btrfs_put_block_group(cache); |
| 9281 | return ret; | 9281 | return ret; |
| 9282 | } | 9282 | } |
| 9283 | update_global_block_rsv(root->fs_info); | 9283 | update_global_block_rsv(root->fs_info); |
| 9284 | 9284 | ||
| 9285 | spin_lock(&cache->space_info->lock); | 9285 | spin_lock(&cache->space_info->lock); |
| 9286 | cache->space_info->bytes_readonly += cache->bytes_super; | 9286 | cache->space_info->bytes_readonly += cache->bytes_super; |
| 9287 | spin_unlock(&cache->space_info->lock); | 9287 | spin_unlock(&cache->space_info->lock); |
| 9288 | 9288 | ||
| 9289 | __link_block_group(cache->space_info, cache); | 9289 | __link_block_group(cache->space_info, cache); |
| 9290 | 9290 | ||
| 9291 | list_add_tail(&cache->bg_list, &trans->new_bgs); | 9291 | list_add_tail(&cache->bg_list, &trans->new_bgs); |
| 9292 | 9292 | ||
| 9293 | set_avail_alloc_bits(extent_root->fs_info, type); | 9293 | set_avail_alloc_bits(extent_root->fs_info, type); |
| 9294 | 9294 | ||
| 9295 | return 0; | 9295 | return 0; |
| 9296 | } | 9296 | } |
| 9297 | 9297 | ||
| 9298 | static void clear_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) | 9298 | static void clear_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags) |
| 9299 | { | 9299 | { |
| 9300 | u64 extra_flags = chunk_to_extended(flags) & | 9300 | u64 extra_flags = chunk_to_extended(flags) & |
| 9301 | BTRFS_EXTENDED_PROFILE_MASK; | 9301 | BTRFS_EXTENDED_PROFILE_MASK; |
| 9302 | 9302 | ||
| 9303 | write_seqlock(&fs_info->profiles_lock); | 9303 | write_seqlock(&fs_info->profiles_lock); |
| 9304 | if (flags & BTRFS_BLOCK_GROUP_DATA) | 9304 | if (flags & BTRFS_BLOCK_GROUP_DATA) |
| 9305 | fs_info->avail_data_alloc_bits &= ~extra_flags; | 9305 | fs_info->avail_data_alloc_bits &= ~extra_flags; |
| 9306 | if (flags & BTRFS_BLOCK_GROUP_METADATA) | 9306 | if (flags & BTRFS_BLOCK_GROUP_METADATA) |
| 9307 | fs_info->avail_metadata_alloc_bits &= ~extra_flags; | 9307 | fs_info->avail_metadata_alloc_bits &= ~extra_flags; |
| 9308 | if (flags & BTRFS_BLOCK_GROUP_SYSTEM) | 9308 | if (flags & BTRFS_BLOCK_GROUP_SYSTEM) |
| 9309 | fs_info->avail_system_alloc_bits &= ~extra_flags; | 9309 | fs_info->avail_system_alloc_bits &= ~extra_flags; |
| 9310 | write_sequnlock(&fs_info->profiles_lock); | 9310 | write_sequnlock(&fs_info->profiles_lock); |
| 9311 | } | 9311 | } |
| 9312 | 9312 | ||
| 9313 | int btrfs_remove_block_group(struct btrfs_trans_handle *trans, | 9313 | int btrfs_remove_block_group(struct btrfs_trans_handle *trans, |
| 9314 | struct btrfs_root *root, u64 group_start, | 9314 | struct btrfs_root *root, u64 group_start, |
| 9315 | struct extent_map *em) | 9315 | struct extent_map *em) |
| 9316 | { | 9316 | { |
| 9317 | struct btrfs_path *path; | 9317 | struct btrfs_path *path; |
| 9318 | struct btrfs_block_group_cache *block_group; | 9318 | struct btrfs_block_group_cache *block_group; |
| 9319 | struct btrfs_free_cluster *cluster; | 9319 | struct btrfs_free_cluster *cluster; |
| 9320 | struct btrfs_root *tree_root = root->fs_info->tree_root; | 9320 | struct btrfs_root *tree_root = root->fs_info->tree_root; |
| 9321 | struct btrfs_key key; | 9321 | struct btrfs_key key; |
| 9322 | struct inode *inode; | 9322 | struct inode *inode; |
| 9323 | struct kobject *kobj = NULL; | 9323 | struct kobject *kobj = NULL; |
| 9324 | int ret; | 9324 | int ret; |
| 9325 | int index; | 9325 | int index; |
| 9326 | int factor; | 9326 | int factor; |
| 9327 | struct btrfs_caching_control *caching_ctl = NULL; | 9327 | struct btrfs_caching_control *caching_ctl = NULL; |
| 9328 | bool remove_em; | 9328 | bool remove_em; |
| 9329 | 9329 | ||
| 9330 | root = root->fs_info->extent_root; | 9330 | root = root->fs_info->extent_root; |
| 9331 | 9331 | ||
| 9332 | block_group = btrfs_lookup_block_group(root->fs_info, group_start); | 9332 | block_group = btrfs_lookup_block_group(root->fs_info, group_start); |
| 9333 | BUG_ON(!block_group); | 9333 | BUG_ON(!block_group); |
| 9334 | BUG_ON(!block_group->ro); | 9334 | BUG_ON(!block_group->ro); |
| 9335 | 9335 | ||
| 9336 | /* | 9336 | /* |
| 9337 | * Free the reserved super bytes from this block group before | 9337 | * Free the reserved super bytes from this block group before |
| 9338 | * remove it. | 9338 | * remove it. |
| 9339 | */ | 9339 | */ |
| 9340 | free_excluded_extents(root, block_group); | 9340 | free_excluded_extents(root, block_group); |
| 9341 | 9341 | ||
| 9342 | memcpy(&key, &block_group->key, sizeof(key)); | 9342 | memcpy(&key, &block_group->key, sizeof(key)); |
| 9343 | index = get_block_group_index(block_group); | 9343 | index = get_block_group_index(block_group); |
| 9344 | if (block_group->flags & (BTRFS_BLOCK_GROUP_DUP | | 9344 | if (block_group->flags & (BTRFS_BLOCK_GROUP_DUP | |
| 9345 | BTRFS_BLOCK_GROUP_RAID1 | | 9345 | BTRFS_BLOCK_GROUP_RAID1 | |
| 9346 | BTRFS_BLOCK_GROUP_RAID10)) | 9346 | BTRFS_BLOCK_GROUP_RAID10)) |
| 9347 | factor = 2; | 9347 | factor = 2; |
| 9348 | else | 9348 | else |
| 9349 | factor = 1; | 9349 | factor = 1; |
| 9350 | 9350 | ||
| 9351 | /* make sure this block group isn't part of an allocation cluster */ | 9351 | /* make sure this block group isn't part of an allocation cluster */ |
| 9352 | cluster = &root->fs_info->data_alloc_cluster; | 9352 | cluster = &root->fs_info->data_alloc_cluster; |
| 9353 | spin_lock(&cluster->refill_lock); | 9353 | spin_lock(&cluster->refill_lock); |
| 9354 | btrfs_return_cluster_to_free_space(block_group, cluster); | 9354 | btrfs_return_cluster_to_free_space(block_group, cluster); |
| 9355 | spin_unlock(&cluster->refill_lock); | 9355 | spin_unlock(&cluster->refill_lock); |
| 9356 | 9356 | ||
| 9357 | /* | 9357 | /* |
| 9358 | * make sure this block group isn't part of a metadata | 9358 | * make sure this block group isn't part of a metadata |
| 9359 | * allocation cluster | 9359 | * allocation cluster |
| 9360 | */ | 9360 | */ |
| 9361 | cluster = &root->fs_info->meta_alloc_cluster; | 9361 | cluster = &root->fs_info->meta_alloc_cluster; |
| 9362 | spin_lock(&cluster->refill_lock); | 9362 | spin_lock(&cluster->refill_lock); |
| 9363 | btrfs_return_cluster_to_free_space(block_group, cluster); | 9363 | btrfs_return_cluster_to_free_space(block_group, cluster); |
| 9364 | spin_unlock(&cluster->refill_lock); | 9364 | spin_unlock(&cluster->refill_lock); |
| 9365 | 9365 | ||
| 9366 | path = btrfs_alloc_path(); | 9366 | path = btrfs_alloc_path(); |
| 9367 | if (!path) { | 9367 | if (!path) { |
| 9368 | ret = -ENOMEM; | 9368 | ret = -ENOMEM; |
| 9369 | goto out; | 9369 | goto out; |
| 9370 | } | 9370 | } |
| 9371 | 9371 | ||
| 9372 | inode = lookup_free_space_inode(tree_root, block_group, path); | 9372 | inode = lookup_free_space_inode(tree_root, block_group, path); |
| 9373 | if (!IS_ERR(inode)) { | 9373 | if (!IS_ERR(inode)) { |
| 9374 | ret = btrfs_orphan_add(trans, inode); | 9374 | ret = btrfs_orphan_add(trans, inode); |
| 9375 | if (ret) { | 9375 | if (ret) { |
| 9376 | btrfs_add_delayed_iput(inode); | 9376 | btrfs_add_delayed_iput(inode); |
| 9377 | goto out; | 9377 | goto out; |
| 9378 | } | 9378 | } |
| 9379 | clear_nlink(inode); | 9379 | clear_nlink(inode); |
| 9380 | /* One for the block groups ref */ | 9380 | /* One for the block groups ref */ |
| 9381 | spin_lock(&block_group->lock); | 9381 | spin_lock(&block_group->lock); |
| 9382 | if (block_group->iref) { | 9382 | if (block_group->iref) { |
| 9383 | block_group->iref = 0; | 9383 | block_group->iref = 0; |
| 9384 | block_group->inode = NULL; | 9384 | block_group->inode = NULL; |
| 9385 | spin_unlock(&block_group->lock); | 9385 | spin_unlock(&block_group->lock); |
| 9386 | iput(inode); | 9386 | iput(inode); |
| 9387 | } else { | 9387 | } else { |
| 9388 | spin_unlock(&block_group->lock); | 9388 | spin_unlock(&block_group->lock); |
| 9389 | } | 9389 | } |
| 9390 | /* One for our lookup ref */ | 9390 | /* One for our lookup ref */ |
| 9391 | btrfs_add_delayed_iput(inode); | 9391 | btrfs_add_delayed_iput(inode); |
| 9392 | } | 9392 | } |
| 9393 | 9393 | ||
| 9394 | key.objectid = BTRFS_FREE_SPACE_OBJECTID; | 9394 | key.objectid = BTRFS_FREE_SPACE_OBJECTID; |
| 9395 | key.offset = block_group->key.objectid; | 9395 | key.offset = block_group->key.objectid; |
| 9396 | key.type = 0; | 9396 | key.type = 0; |
| 9397 | 9397 | ||
| 9398 | ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1); | 9398 | ret = btrfs_search_slot(trans, tree_root, &key, path, -1, 1); |
| 9399 | if (ret < 0) | 9399 | if (ret < 0) |
| 9400 | goto out; | 9400 | goto out; |
| 9401 | if (ret > 0) | 9401 | if (ret > 0) |
| 9402 | btrfs_release_path(path); | 9402 | btrfs_release_path(path); |
| 9403 | if (ret == 0) { | 9403 | if (ret == 0) { |
| 9404 | ret = btrfs_del_item(trans, tree_root, path); | 9404 | ret = btrfs_del_item(trans, tree_root, path); |
| 9405 | if (ret) | 9405 | if (ret) |
| 9406 | goto out; | 9406 | goto out; |
| 9407 | btrfs_release_path(path); | 9407 | btrfs_release_path(path); |
| 9408 | } | 9408 | } |
| 9409 | 9409 | ||
| 9410 | spin_lock(&root->fs_info->block_group_cache_lock); | 9410 | spin_lock(&root->fs_info->block_group_cache_lock); |
| 9411 | rb_erase(&block_group->cache_node, | 9411 | rb_erase(&block_group->cache_node, |
| 9412 | &root->fs_info->block_group_cache_tree); | 9412 | &root->fs_info->block_group_cache_tree); |
| 9413 | RB_CLEAR_NODE(&block_group->cache_node); | 9413 | RB_CLEAR_NODE(&block_group->cache_node); |
| 9414 | 9414 | ||
| 9415 | if (root->fs_info->first_logical_byte == block_group->key.objectid) | 9415 | if (root->fs_info->first_logical_byte == block_group->key.objectid) |
| 9416 | root->fs_info->first_logical_byte = (u64)-1; | 9416 | root->fs_info->first_logical_byte = (u64)-1; |
| 9417 | spin_unlock(&root->fs_info->block_group_cache_lock); | 9417 | spin_unlock(&root->fs_info->block_group_cache_lock); |
| 9418 | 9418 | ||
| 9419 | down_write(&block_group->space_info->groups_sem); | 9419 | down_write(&block_group->space_info->groups_sem); |
| 9420 | /* | 9420 | /* |
| 9421 | * we must use list_del_init so people can check to see if they | 9421 | * we must use list_del_init so people can check to see if they |
| 9422 | * are still on the list after taking the semaphore | 9422 | * are still on the list after taking the semaphore |
| 9423 | */ | 9423 | */ |
| 9424 | list_del_init(&block_group->list); | 9424 | list_del_init(&block_group->list); |
| 9425 | list_del_init(&block_group->ro_list); | 9425 | list_del_init(&block_group->ro_list); |
| 9426 | if (list_empty(&block_group->space_info->block_groups[index])) { | 9426 | if (list_empty(&block_group->space_info->block_groups[index])) { |
| 9427 | kobj = block_group->space_info->block_group_kobjs[index]; | 9427 | kobj = block_group->space_info->block_group_kobjs[index]; |
| 9428 | block_group->space_info->block_group_kobjs[index] = NULL; | 9428 | block_group->space_info->block_group_kobjs[index] = NULL; |
| 9429 | clear_avail_alloc_bits(root->fs_info, block_group->flags); | 9429 | clear_avail_alloc_bits(root->fs_info, block_group->flags); |
| 9430 | } | 9430 | } |
| 9431 | up_write(&block_group->space_info->groups_sem); | 9431 | up_write(&block_group->space_info->groups_sem); |
| 9432 | if (kobj) { | 9432 | if (kobj) { |
| 9433 | kobject_del(kobj); | 9433 | kobject_del(kobj); |
| 9434 | kobject_put(kobj); | 9434 | kobject_put(kobj); |
| 9435 | } | 9435 | } |
| 9436 | 9436 | ||
| 9437 | if (block_group->has_caching_ctl) | 9437 | if (block_group->has_caching_ctl) |
| 9438 | caching_ctl = get_caching_control(block_group); | 9438 | caching_ctl = get_caching_control(block_group); |
| 9439 | if (block_group->cached == BTRFS_CACHE_STARTED) | 9439 | if (block_group->cached == BTRFS_CACHE_STARTED) |
| 9440 | wait_block_group_cache_done(block_group); | 9440 | wait_block_group_cache_done(block_group); |
| 9441 | if (block_group->has_caching_ctl) { | 9441 | if (block_group->has_caching_ctl) { |
| 9442 | down_write(&root->fs_info->commit_root_sem); | 9442 | down_write(&root->fs_info->commit_root_sem); |
| 9443 | if (!caching_ctl) { | 9443 | if (!caching_ctl) { |
| 9444 | struct btrfs_caching_control *ctl; | 9444 | struct btrfs_caching_control *ctl; |
| 9445 | 9445 | ||
| 9446 | list_for_each_entry(ctl, | 9446 | list_for_each_entry(ctl, |
| 9447 | &root->fs_info->caching_block_groups, list) | 9447 | &root->fs_info->caching_block_groups, list) |
| 9448 | if (ctl->block_group == block_group) { | 9448 | if (ctl->block_group == block_group) { |
| 9449 | caching_ctl = ctl; | 9449 | caching_ctl = ctl; |
| 9450 | atomic_inc(&caching_ctl->count); | 9450 | atomic_inc(&caching_ctl->count); |
| 9451 | break; | 9451 | break; |
| 9452 | } | 9452 | } |
| 9453 | } | 9453 | } |
| 9454 | if (caching_ctl) | 9454 | if (caching_ctl) |
| 9455 | list_del_init(&caching_ctl->list); | 9455 | list_del_init(&caching_ctl->list); |
| 9456 | up_write(&root->fs_info->commit_root_sem); | 9456 | up_write(&root->fs_info->commit_root_sem); |
| 9457 | if (caching_ctl) { | 9457 | if (caching_ctl) { |
| 9458 | /* Once for the caching bgs list and once for us. */ | 9458 | /* Once for the caching bgs list and once for us. */ |
| 9459 | put_caching_control(caching_ctl); | 9459 | put_caching_control(caching_ctl); |
| 9460 | put_caching_control(caching_ctl); | 9460 | put_caching_control(caching_ctl); |
| 9461 | } | 9461 | } |
| 9462 | } | 9462 | } |
| 9463 | 9463 | ||
| 9464 | btrfs_remove_free_space_cache(block_group); | 9464 | btrfs_remove_free_space_cache(block_group); |
| 9465 | 9465 | ||
| 9466 | spin_lock(&block_group->space_info->lock); | 9466 | spin_lock(&block_group->space_info->lock); |
| 9467 | block_group->space_info->total_bytes -= block_group->key.offset; | 9467 | block_group->space_info->total_bytes -= block_group->key.offset; |
| 9468 | block_group->space_info->bytes_readonly -= block_group->key.offset; | 9468 | block_group->space_info->bytes_readonly -= block_group->key.offset; |
| 9469 | block_group->space_info->disk_total -= block_group->key.offset * factor; | 9469 | block_group->space_info->disk_total -= block_group->key.offset * factor; |
| 9470 | spin_unlock(&block_group->space_info->lock); | 9470 | spin_unlock(&block_group->space_info->lock); |
| 9471 | 9471 | ||
| 9472 | memcpy(&key, &block_group->key, sizeof(key)); | 9472 | memcpy(&key, &block_group->key, sizeof(key)); |
| 9473 | 9473 | ||
| 9474 | lock_chunks(root); | 9474 | lock_chunks(root); |
| 9475 | if (!list_empty(&em->list)) { | 9475 | if (!list_empty(&em->list)) { |
| 9476 | /* We're in the transaction->pending_chunks list. */ | 9476 | /* We're in the transaction->pending_chunks list. */ |
| 9477 | free_extent_map(em); | 9477 | free_extent_map(em); |
| 9478 | } | 9478 | } |
| 9479 | spin_lock(&block_group->lock); | 9479 | spin_lock(&block_group->lock); |
| 9480 | block_group->removed = 1; | 9480 | block_group->removed = 1; |
| 9481 | /* | 9481 | /* |
| 9482 | * At this point trimming can't start on this block group, because we | 9482 | * At this point trimming can't start on this block group, because we |
| 9483 | * removed the block group from the tree fs_info->block_group_cache_tree | 9483 | * removed the block group from the tree fs_info->block_group_cache_tree |
| 9484 | * so no one can't find it anymore and even if someone already got this | 9484 | * so no one can't find it anymore and even if someone already got this |
| 9485 | * block group before we removed it from the rbtree, they have already | 9485 | * block group before we removed it from the rbtree, they have already |
| 9486 | * incremented block_group->trimming - if they didn't, they won't find | 9486 | * incremented block_group->trimming - if they didn't, they won't find |
| 9487 | * any free space entries because we already removed them all when we | 9487 | * any free space entries because we already removed them all when we |
| 9488 | * called btrfs_remove_free_space_cache(). | 9488 | * called btrfs_remove_free_space_cache(). |
| 9489 | * | 9489 | * |
| 9490 | * And we must not remove the extent map from the fs_info->mapping_tree | 9490 | * And we must not remove the extent map from the fs_info->mapping_tree |
| 9491 | * to prevent the same logical address range and physical device space | 9491 | * to prevent the same logical address range and physical device space |
| 9492 | * ranges from being reused for a new block group. This is because our | 9492 | * ranges from being reused for a new block group. This is because our |
| 9493 | * fs trim operation (btrfs_trim_fs() / btrfs_ioctl_fitrim()) is | 9493 | * fs trim operation (btrfs_trim_fs() / btrfs_ioctl_fitrim()) is |
| 9494 | * completely transactionless, so while it is trimming a range the | 9494 | * completely transactionless, so while it is trimming a range the |
| 9495 | * currently running transaction might finish and a new one start, | 9495 | * currently running transaction might finish and a new one start, |
| 9496 | * allowing for new block groups to be created that can reuse the same | 9496 | * allowing for new block groups to be created that can reuse the same |
| 9497 | * physical device locations unless we take this special care. | 9497 | * physical device locations unless we take this special care. |
| 9498 | */ | 9498 | */ |
| 9499 | remove_em = (atomic_read(&block_group->trimming) == 0); | 9499 | remove_em = (atomic_read(&block_group->trimming) == 0); |
| 9500 | /* | 9500 | /* |
| 9501 | * Make sure a trimmer task always sees the em in the pinned_chunks list | 9501 | * Make sure a trimmer task always sees the em in the pinned_chunks list |
| 9502 | * if it sees block_group->removed == 1 (needs to lock block_group->lock | 9502 | * if it sees block_group->removed == 1 (needs to lock block_group->lock |
| 9503 | * before checking block_group->removed). | 9503 | * before checking block_group->removed). |
| 9504 | */ | 9504 | */ |
| 9505 | if (!remove_em) { | 9505 | if (!remove_em) { |
| 9506 | /* | 9506 | /* |
| 9507 | * Our em might be in trans->transaction->pending_chunks which | 9507 | * Our em might be in trans->transaction->pending_chunks which |
| 9508 | * is protected by fs_info->chunk_mutex ([lock|unlock]_chunks), | 9508 | * is protected by fs_info->chunk_mutex ([lock|unlock]_chunks), |
| 9509 | * and so is the fs_info->pinned_chunks list. | 9509 | * and so is the fs_info->pinned_chunks list. |
| 9510 | * | 9510 | * |
| 9511 | * So at this point we must be holding the chunk_mutex to avoid | 9511 | * So at this point we must be holding the chunk_mutex to avoid |
| 9512 | * any races with chunk allocation (more specifically at | 9512 | * any races with chunk allocation (more specifically at |
| 9513 | * volumes.c:contains_pending_extent()), to ensure it always | 9513 | * volumes.c:contains_pending_extent()), to ensure it always |
| 9514 | * sees the em, either in the pending_chunks list or in the | 9514 | * sees the em, either in the pending_chunks list or in the |
| 9515 | * pinned_chunks list. | 9515 | * pinned_chunks list. |
| 9516 | */ | 9516 | */ |
| 9517 | list_move_tail(&em->list, &root->fs_info->pinned_chunks); | 9517 | list_move_tail(&em->list, &root->fs_info->pinned_chunks); |
| 9518 | } | 9518 | } |
| 9519 | spin_unlock(&block_group->lock); | 9519 | spin_unlock(&block_group->lock); |
| 9520 | 9520 | ||
| 9521 | if (remove_em) { | 9521 | if (remove_em) { |
| 9522 | struct extent_map_tree *em_tree; | 9522 | struct extent_map_tree *em_tree; |
| 9523 | 9523 | ||
| 9524 | em_tree = &root->fs_info->mapping_tree.map_tree; | 9524 | em_tree = &root->fs_info->mapping_tree.map_tree; |
| 9525 | write_lock(&em_tree->lock); | 9525 | write_lock(&em_tree->lock); |
| 9526 | /* | 9526 | /* |
| 9527 | * The em might be in the pending_chunks list, so make sure the | 9527 | * The em might be in the pending_chunks list, so make sure the |
| 9528 | * chunk mutex is locked, since remove_extent_mapping() will | 9528 | * chunk mutex is locked, since remove_extent_mapping() will |
| 9529 | * delete us from that list. | 9529 | * delete us from that list. |
| 9530 | */ | 9530 | */ |
| 9531 | remove_extent_mapping(em_tree, em); | 9531 | remove_extent_mapping(em_tree, em); |
| 9532 | write_unlock(&em_tree->lock); | 9532 | write_unlock(&em_tree->lock); |
| 9533 | /* once for the tree */ | 9533 | /* once for the tree */ |
| 9534 | free_extent_map(em); | 9534 | free_extent_map(em); |
| 9535 | } | 9535 | } |
| 9536 | 9536 | ||
| 9537 | unlock_chunks(root); | 9537 | unlock_chunks(root); |
| 9538 | 9538 | ||
| 9539 | btrfs_put_block_group(block_group); | 9539 | btrfs_put_block_group(block_group); |
| 9540 | btrfs_put_block_group(block_group); | 9540 | btrfs_put_block_group(block_group); |
| 9541 | 9541 | ||
| 9542 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 9542 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 9543 | if (ret > 0) | 9543 | if (ret > 0) |
| 9544 | ret = -EIO; | 9544 | ret = -EIO; |
| 9545 | if (ret < 0) | 9545 | if (ret < 0) |
| 9546 | goto out; | 9546 | goto out; |
| 9547 | 9547 | ||
| 9548 | ret = btrfs_del_item(trans, root, path); | 9548 | ret = btrfs_del_item(trans, root, path); |
| 9549 | out: | 9549 | out: |
| 9550 | btrfs_free_path(path); | 9550 | btrfs_free_path(path); |
| 9551 | return ret; | 9551 | return ret; |
| 9552 | } | 9552 | } |
| 9553 | 9553 | ||
| 9554 | /* | 9554 | /* |
| 9555 | * Process the unused_bgs list and remove any that don't have any allocated | 9555 | * Process the unused_bgs list and remove any that don't have any allocated |
| 9556 | * space inside of them. | 9556 | * space inside of them. |
| 9557 | */ | 9557 | */ |
| 9558 | void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info) | 9558 | void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info) |
| 9559 | { | 9559 | { |
| 9560 | struct btrfs_block_group_cache *block_group; | 9560 | struct btrfs_block_group_cache *block_group; |
| 9561 | struct btrfs_space_info *space_info; | 9561 | struct btrfs_space_info *space_info; |
| 9562 | struct btrfs_root *root = fs_info->extent_root; | 9562 | struct btrfs_root *root = fs_info->extent_root; |
| 9563 | struct btrfs_trans_handle *trans; | 9563 | struct btrfs_trans_handle *trans; |
| 9564 | int ret = 0; | 9564 | int ret = 0; |
| 9565 | 9565 | ||
| 9566 | if (!fs_info->open) | 9566 | if (!fs_info->open) |
| 9567 | return; | 9567 | return; |
| 9568 | 9568 | ||
| 9569 | spin_lock(&fs_info->unused_bgs_lock); | 9569 | spin_lock(&fs_info->unused_bgs_lock); |
| 9570 | while (!list_empty(&fs_info->unused_bgs)) { | 9570 | while (!list_empty(&fs_info->unused_bgs)) { |
| 9571 | u64 start, end; | 9571 | u64 start, end; |
| 9572 | 9572 | ||
| 9573 | block_group = list_first_entry(&fs_info->unused_bgs, | 9573 | block_group = list_first_entry(&fs_info->unused_bgs, |
| 9574 | struct btrfs_block_group_cache, | 9574 | struct btrfs_block_group_cache, |
| 9575 | bg_list); | 9575 | bg_list); |
| 9576 | space_info = block_group->space_info; | 9576 | space_info = block_group->space_info; |
| 9577 | list_del_init(&block_group->bg_list); | 9577 | list_del_init(&block_group->bg_list); |
| 9578 | if (ret || btrfs_mixed_space_info(space_info)) { | 9578 | if (ret || btrfs_mixed_space_info(space_info)) { |
| 9579 | btrfs_put_block_group(block_group); | 9579 | btrfs_put_block_group(block_group); |
| 9580 | continue; | 9580 | continue; |
| 9581 | } | 9581 | } |
| 9582 | spin_unlock(&fs_info->unused_bgs_lock); | 9582 | spin_unlock(&fs_info->unused_bgs_lock); |
| 9583 | 9583 | ||
| 9584 | /* Don't want to race with allocators so take the groups_sem */ | 9584 | /* Don't want to race with allocators so take the groups_sem */ |
| 9585 | down_write(&space_info->groups_sem); | 9585 | down_write(&space_info->groups_sem); |
| 9586 | spin_lock(&block_group->lock); | 9586 | spin_lock(&block_group->lock); |
| 9587 | if (block_group->reserved || | 9587 | if (block_group->reserved || |
| 9588 | btrfs_block_group_used(&block_group->item) || | 9588 | btrfs_block_group_used(&block_group->item) || |
| 9589 | block_group->ro) { | 9589 | block_group->ro) { |
| 9590 | /* | 9590 | /* |
| 9591 | * We want to bail if we made new allocations or have | 9591 | * We want to bail if we made new allocations or have |
| 9592 | * outstanding allocations in this block group. We do | 9592 | * outstanding allocations in this block group. We do |
| 9593 | * the ro check in case balance is currently acting on | 9593 | * the ro check in case balance is currently acting on |
| 9594 | * this block group. | 9594 | * this block group. |
| 9595 | */ | 9595 | */ |
| 9596 | spin_unlock(&block_group->lock); | 9596 | spin_unlock(&block_group->lock); |
| 9597 | up_write(&space_info->groups_sem); | 9597 | up_write(&space_info->groups_sem); |
| 9598 | goto next; | 9598 | goto next; |
| 9599 | } | 9599 | } |
| 9600 | spin_unlock(&block_group->lock); | 9600 | spin_unlock(&block_group->lock); |
| 9601 | 9601 | ||
| 9602 | /* We don't want to force the issue, only flip if it's ok. */ | 9602 | /* We don't want to force the issue, only flip if it's ok. */ |
| 9603 | ret = set_block_group_ro(block_group, 0); | 9603 | ret = set_block_group_ro(block_group, 0); |
| 9604 | up_write(&space_info->groups_sem); | 9604 | up_write(&space_info->groups_sem); |
| 9605 | if (ret < 0) { | 9605 | if (ret < 0) { |
| 9606 | ret = 0; | 9606 | ret = 0; |
| 9607 | goto next; | 9607 | goto next; |
| 9608 | } | 9608 | } |
| 9609 | 9609 | ||
| 9610 | /* | 9610 | /* |
| 9611 | * Want to do this before we do anything else so we can recover | 9611 | * Want to do this before we do anything else so we can recover |
| 9612 | * properly if we fail to join the transaction. | 9612 | * properly if we fail to join the transaction. |
| 9613 | */ | 9613 | */ |
| 9614 | trans = btrfs_join_transaction(root); | 9614 | trans = btrfs_join_transaction(root); |
| 9615 | if (IS_ERR(trans)) { | 9615 | if (IS_ERR(trans)) { |
| 9616 | btrfs_set_block_group_rw(root, block_group); | 9616 | btrfs_set_block_group_rw(root, block_group); |
| 9617 | ret = PTR_ERR(trans); | 9617 | ret = PTR_ERR(trans); |
| 9618 | goto next; | 9618 | goto next; |
| 9619 | } | 9619 | } |
| 9620 | 9620 | ||
| 9621 | /* | 9621 | /* |
| 9622 | * We could have pending pinned extents for this block group, | 9622 | * We could have pending pinned extents for this block group, |
| 9623 | * just delete them, we don't care about them anymore. | 9623 | * just delete them, we don't care about them anymore. |
| 9624 | */ | 9624 | */ |
| 9625 | start = block_group->key.objectid; | 9625 | start = block_group->key.objectid; |
| 9626 | end = start + block_group->key.offset - 1; | 9626 | end = start + block_group->key.offset - 1; |
| 9627 | ret = clear_extent_bits(&fs_info->freed_extents[0], start, end, | 9627 | ret = clear_extent_bits(&fs_info->freed_extents[0], start, end, |
| 9628 | EXTENT_DIRTY, GFP_NOFS); | 9628 | EXTENT_DIRTY, GFP_NOFS); |
| 9629 | if (ret) { | 9629 | if (ret) { |
| 9630 | btrfs_set_block_group_rw(root, block_group); | 9630 | btrfs_set_block_group_rw(root, block_group); |
| 9631 | goto end_trans; | 9631 | goto end_trans; |
| 9632 | } | 9632 | } |
| 9633 | ret = clear_extent_bits(&fs_info->freed_extents[1], start, end, | 9633 | ret = clear_extent_bits(&fs_info->freed_extents[1], start, end, |
| 9634 | EXTENT_DIRTY, GFP_NOFS); | 9634 | EXTENT_DIRTY, GFP_NOFS); |
| 9635 | if (ret) { | 9635 | if (ret) { |
| 9636 | btrfs_set_block_group_rw(root, block_group); | 9636 | btrfs_set_block_group_rw(root, block_group); |
| 9637 | goto end_trans; | 9637 | goto end_trans; |
| 9638 | } | 9638 | } |
| 9639 | 9639 | ||
| 9640 | /* Reset pinned so btrfs_put_block_group doesn't complain */ | 9640 | /* Reset pinned so btrfs_put_block_group doesn't complain */ |
| 9641 | block_group->pinned = 0; | 9641 | block_group->pinned = 0; |
| 9642 | 9642 | ||
| 9643 | /* | 9643 | /* |
| 9644 | * Btrfs_remove_chunk will abort the transaction if things go | 9644 | * Btrfs_remove_chunk will abort the transaction if things go |
| 9645 | * horribly wrong. | 9645 | * horribly wrong. |
| 9646 | */ | 9646 | */ |
| 9647 | ret = btrfs_remove_chunk(trans, root, | 9647 | ret = btrfs_remove_chunk(trans, root, |
| 9648 | block_group->key.objectid); | 9648 | block_group->key.objectid); |
| 9649 | end_trans: | 9649 | end_trans: |
| 9650 | btrfs_end_transaction(trans, root); | 9650 | btrfs_end_transaction(trans, root); |
| 9651 | next: | 9651 | next: |
| 9652 | btrfs_put_block_group(block_group); | 9652 | btrfs_put_block_group(block_group); |
| 9653 | spin_lock(&fs_info->unused_bgs_lock); | 9653 | spin_lock(&fs_info->unused_bgs_lock); |
| 9654 | } | 9654 | } |
| 9655 | spin_unlock(&fs_info->unused_bgs_lock); | 9655 | spin_unlock(&fs_info->unused_bgs_lock); |
| 9656 | } | 9656 | } |
| 9657 | 9657 | ||
| 9658 | int btrfs_init_space_info(struct btrfs_fs_info *fs_info) | 9658 | int btrfs_init_space_info(struct btrfs_fs_info *fs_info) |
| 9659 | { | 9659 | { |
| 9660 | struct btrfs_space_info *space_info; | 9660 | struct btrfs_space_info *space_info; |
| 9661 | struct btrfs_super_block *disk_super; | 9661 | struct btrfs_super_block *disk_super; |
| 9662 | u64 features; | 9662 | u64 features; |
| 9663 | u64 flags; | 9663 | u64 flags; |
| 9664 | int mixed = 0; | 9664 | int mixed = 0; |
| 9665 | int ret; | 9665 | int ret; |
| 9666 | 9666 | ||
| 9667 | disk_super = fs_info->super_copy; | 9667 | disk_super = fs_info->super_copy; |
| 9668 | if (!btrfs_super_root(disk_super)) | 9668 | if (!btrfs_super_root(disk_super)) |
| 9669 | return 1; | 9669 | return 1; |
| 9670 | 9670 | ||
| 9671 | features = btrfs_super_incompat_flags(disk_super); | 9671 | features = btrfs_super_incompat_flags(disk_super); |
| 9672 | if (features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) | 9672 | if (features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) |
| 9673 | mixed = 1; | 9673 | mixed = 1; |
| 9674 | 9674 | ||
| 9675 | flags = BTRFS_BLOCK_GROUP_SYSTEM; | 9675 | flags = BTRFS_BLOCK_GROUP_SYSTEM; |
| 9676 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); | 9676 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); |
| 9677 | if (ret) | 9677 | if (ret) |
| 9678 | goto out; | 9678 | goto out; |
| 9679 | 9679 | ||
| 9680 | if (mixed) { | 9680 | if (mixed) { |
| 9681 | flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA; | 9681 | flags = BTRFS_BLOCK_GROUP_METADATA | BTRFS_BLOCK_GROUP_DATA; |
| 9682 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); | 9682 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); |
| 9683 | } else { | 9683 | } else { |
| 9684 | flags = BTRFS_BLOCK_GROUP_METADATA; | 9684 | flags = BTRFS_BLOCK_GROUP_METADATA; |
| 9685 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); | 9685 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); |
| 9686 | if (ret) | 9686 | if (ret) |
| 9687 | goto out; | 9687 | goto out; |
| 9688 | 9688 | ||
| 9689 | flags = BTRFS_BLOCK_GROUP_DATA; | 9689 | flags = BTRFS_BLOCK_GROUP_DATA; |
| 9690 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); | 9690 | ret = update_space_info(fs_info, flags, 0, 0, &space_info); |
| 9691 | } | 9691 | } |
| 9692 | out: | 9692 | out: |
| 9693 | return ret; | 9693 | return ret; |
| 9694 | } | 9694 | } |
| 9695 | 9695 | ||
| 9696 | int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) | 9696 | int btrfs_error_unpin_extent_range(struct btrfs_root *root, u64 start, u64 end) |
| 9697 | { | 9697 | { |
| 9698 | return unpin_extent_range(root, start, end, false); | 9698 | return unpin_extent_range(root, start, end, false); |
| 9699 | } | 9699 | } |
| 9700 | 9700 | ||
| 9701 | int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range) | 9701 | int btrfs_trim_fs(struct btrfs_root *root, struct fstrim_range *range) |
| 9702 | { | 9702 | { |
| 9703 | struct btrfs_fs_info *fs_info = root->fs_info; | 9703 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 9704 | struct btrfs_block_group_cache *cache = NULL; | 9704 | struct btrfs_block_group_cache *cache = NULL; |
| 9705 | u64 group_trimmed; | 9705 | u64 group_trimmed; |
| 9706 | u64 start; | 9706 | u64 start; |
| 9707 | u64 end; | 9707 | u64 end; |
| 9708 | u64 trimmed = 0; | 9708 | u64 trimmed = 0; |
| 9709 | u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy); | 9709 | u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy); |
| 9710 | int ret = 0; | 9710 | int ret = 0; |
| 9711 | 9711 | ||
| 9712 | /* | 9712 | /* |
| 9713 | * try to trim all FS space, our block group may start from non-zero. | 9713 | * try to trim all FS space, our block group may start from non-zero. |
| 9714 | */ | 9714 | */ |
| 9715 | if (range->len == total_bytes) | 9715 | if (range->len == total_bytes) |
| 9716 | cache = btrfs_lookup_first_block_group(fs_info, range->start); | 9716 | cache = btrfs_lookup_first_block_group(fs_info, range->start); |
| 9717 | else | 9717 | else |
| 9718 | cache = btrfs_lookup_block_group(fs_info, range->start); | 9718 | cache = btrfs_lookup_block_group(fs_info, range->start); |
| 9719 | 9719 | ||
| 9720 | while (cache) { | 9720 | while (cache) { |
| 9721 | if (cache->key.objectid >= (range->start + range->len)) { | 9721 | if (cache->key.objectid >= (range->start + range->len)) { |
| 9722 | btrfs_put_block_group(cache); | 9722 | btrfs_put_block_group(cache); |
| 9723 | break; | 9723 | break; |
| 9724 | } | 9724 | } |
| 9725 | 9725 | ||
| 9726 | start = max(range->start, cache->key.objectid); | 9726 | start = max(range->start, cache->key.objectid); |
| 9727 | end = min(range->start + range->len, | 9727 | end = min(range->start + range->len, |
| 9728 | cache->key.objectid + cache->key.offset); | 9728 | cache->key.objectid + cache->key.offset); |
| 9729 | 9729 | ||
| 9730 | if (end - start >= range->minlen) { | 9730 | if (end - start >= range->minlen) { |
| 9731 | if (!block_group_cache_done(cache)) { | 9731 | if (!block_group_cache_done(cache)) { |
| 9732 | ret = cache_block_group(cache, 0); | 9732 | ret = cache_block_group(cache, 0); |
| 9733 | if (ret) { | 9733 | if (ret) { |
| 9734 | btrfs_put_block_group(cache); | 9734 | btrfs_put_block_group(cache); |
| 9735 | break; | 9735 | break; |
| 9736 | } | 9736 | } |
| 9737 | ret = wait_block_group_cache_done(cache); | 9737 | ret = wait_block_group_cache_done(cache); |
| 9738 | if (ret) { | 9738 | if (ret) { |
| 9739 | btrfs_put_block_group(cache); | 9739 | btrfs_put_block_group(cache); |
| 9740 | break; | 9740 | break; |
| 9741 | } | 9741 | } |
| 9742 | } | 9742 | } |
| 9743 | ret = btrfs_trim_block_group(cache, | 9743 | ret = btrfs_trim_block_group(cache, |
| 9744 | &group_trimmed, | 9744 | &group_trimmed, |
| 9745 | start, | 9745 | start, |
| 9746 | end, | 9746 | end, |
| 9747 | range->minlen); | 9747 | range->minlen); |
| 9748 | 9748 | ||
| 9749 | trimmed += group_trimmed; | 9749 | trimmed += group_trimmed; |
| 9750 | if (ret) { | 9750 | if (ret) { |
| 9751 | btrfs_put_block_group(cache); | 9751 | btrfs_put_block_group(cache); |
| 9752 | break; | 9752 | break; |
| 9753 | } | 9753 | } |
| 9754 | } | 9754 | } |
| 9755 | 9755 | ||
| 9756 | cache = next_block_group(fs_info->tree_root, cache); | 9756 | cache = next_block_group(fs_info->tree_root, cache); |
| 9757 | } | 9757 | } |
| 9758 | 9758 | ||
| 9759 | range->len = trimmed; | 9759 | range->len = trimmed; |
| 9760 | return ret; | 9760 | return ret; |
| 9761 | } | 9761 | } |
| 9762 | 9762 | ||
| 9763 | /* | 9763 | /* |
| 9764 | * btrfs_{start,end}_write_no_snapshoting() are similar to | 9764 | * btrfs_{start,end}_write_no_snapshoting() are similar to |
| 9765 | * mnt_{want,drop}_write(), they are used to prevent some tasks from writing | 9765 | * mnt_{want,drop}_write(), they are used to prevent some tasks from writing |
| 9766 | * data into the page cache through nocow before the subvolume is snapshoted, | 9766 | * data into the page cache through nocow before the subvolume is snapshoted, |
| 9767 | * but flush the data into disk after the snapshot creation, or to prevent | 9767 | * but flush the data into disk after the snapshot creation, or to prevent |
| 9768 | * operations while snapshoting is ongoing and that cause the snapshot to be | 9768 | * operations while snapshoting is ongoing and that cause the snapshot to be |
| 9769 | * inconsistent (writes followed by expanding truncates for example). | 9769 | * inconsistent (writes followed by expanding truncates for example). |
| 9770 | */ | 9770 | */ |
| 9771 | void btrfs_end_write_no_snapshoting(struct btrfs_root *root) | 9771 | void btrfs_end_write_no_snapshoting(struct btrfs_root *root) |
| 9772 | { | 9772 | { |
| 9773 | percpu_counter_dec(&root->subv_writers->counter); | 9773 | percpu_counter_dec(&root->subv_writers->counter); |
| 9774 | /* | 9774 | /* |
| 9775 | * Make sure counter is updated before we wake up | 9775 | * Make sure counter is updated before we wake up |
| 9776 | * waiters. | 9776 | * waiters. |
| 9777 | */ | 9777 | */ |
| 9778 | smp_mb(); | 9778 | smp_mb(); |
| 9779 | if (waitqueue_active(&root->subv_writers->wait)) | 9779 | if (waitqueue_active(&root->subv_writers->wait)) |
| 9780 | wake_up(&root->subv_writers->wait); | 9780 | wake_up(&root->subv_writers->wait); |
| 9781 | } | 9781 | } |
| 9782 | 9782 | ||
| 9783 | int btrfs_start_write_no_snapshoting(struct btrfs_root *root) | 9783 | int btrfs_start_write_no_snapshoting(struct btrfs_root *root) |
| 9784 | { | 9784 | { |
| 9785 | if (atomic_read(&root->will_be_snapshoted)) | 9785 | if (atomic_read(&root->will_be_snapshoted)) |
| 9786 | return 0; | 9786 | return 0; |
| 9787 | 9787 | ||
| 9788 | percpu_counter_inc(&root->subv_writers->counter); | 9788 | percpu_counter_inc(&root->subv_writers->counter); |
| 9789 | /* | 9789 | /* |
| 9790 | * Make sure counter is updated before we check for snapshot creation. | 9790 | * Make sure counter is updated before we check for snapshot creation. |
| 9791 | */ | 9791 | */ |
| 9792 | smp_mb(); | 9792 | smp_mb(); |
| 9793 | if (atomic_read(&root->will_be_snapshoted)) { | 9793 | if (atomic_read(&root->will_be_snapshoted)) { |
| 9794 | btrfs_end_write_no_snapshoting(root); | 9794 | btrfs_end_write_no_snapshoting(root); |
| 9795 | return 0; | 9795 | return 0; |
| 9796 | } | 9796 | } |
| 9797 | return 1; | 9797 | return 1; |
fs/btrfs/inode.c
| 1 | /* | 1 | /* |
| 2 | * Copyright (C) 2007 Oracle. All rights reserved. | 2 | * Copyright (C) 2007 Oracle. All rights reserved. |
| 3 | * | 3 | * |
| 4 | * This program is free software; you can redistribute it and/or | 4 | * This program is free software; you can redistribute it and/or |
| 5 | * modify it under the terms of the GNU General Public | 5 | * modify it under the terms of the GNU General Public |
| 6 | * License v2 as published by the Free Software Foundation. | 6 | * License v2 as published by the Free Software Foundation. |
| 7 | * | 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, | 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 11 | * General Public License for more details. | 11 | * General Public License for more details. |
| 12 | * | 12 | * |
| 13 | * You should have received a copy of the GNU General Public | 13 | * You should have received a copy of the GNU General Public |
| 14 | * License along with this program; if not, write to the | 14 | * License along with this program; if not, write to the |
| 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 16 | * Boston, MA 021110-1307, USA. | 16 | * Boston, MA 021110-1307, USA. |
| 17 | */ | 17 | */ |
| 18 | 18 | ||
| 19 | #include <linux/kernel.h> | 19 | #include <linux/kernel.h> |
| 20 | #include <linux/bio.h> | 20 | #include <linux/bio.h> |
| 21 | #include <linux/buffer_head.h> | 21 | #include <linux/buffer_head.h> |
| 22 | #include <linux/file.h> | 22 | #include <linux/file.h> |
| 23 | #include <linux/fs.h> | 23 | #include <linux/fs.h> |
| 24 | #include <linux/pagemap.h> | 24 | #include <linux/pagemap.h> |
| 25 | #include <linux/highmem.h> | 25 | #include <linux/highmem.h> |
| 26 | #include <linux/time.h> | 26 | #include <linux/time.h> |
| 27 | #include <linux/init.h> | 27 | #include <linux/init.h> |
| 28 | #include <linux/string.h> | 28 | #include <linux/string.h> |
| 29 | #include <linux/backing-dev.h> | 29 | #include <linux/backing-dev.h> |
| 30 | #include <linux/mpage.h> | 30 | #include <linux/mpage.h> |
| 31 | #include <linux/swap.h> | 31 | #include <linux/swap.h> |
| 32 | #include <linux/writeback.h> | 32 | #include <linux/writeback.h> |
| 33 | #include <linux/statfs.h> | 33 | #include <linux/statfs.h> |
| 34 | #include <linux/compat.h> | 34 | #include <linux/compat.h> |
| 35 | #include <linux/aio.h> | 35 | #include <linux/aio.h> |
| 36 | #include <linux/bit_spinlock.h> | 36 | #include <linux/bit_spinlock.h> |
| 37 | #include <linux/xattr.h> | 37 | #include <linux/xattr.h> |
| 38 | #include <linux/posix_acl.h> | 38 | #include <linux/posix_acl.h> |
| 39 | #include <linux/falloc.h> | 39 | #include <linux/falloc.h> |
| 40 | #include <linux/slab.h> | 40 | #include <linux/slab.h> |
| 41 | #include <linux/ratelimit.h> | 41 | #include <linux/ratelimit.h> |
| 42 | #include <linux/mount.h> | 42 | #include <linux/mount.h> |
| 43 | #include <linux/btrfs.h> | 43 | #include <linux/btrfs.h> |
| 44 | #include <linux/blkdev.h> | 44 | #include <linux/blkdev.h> |
| 45 | #include <linux/posix_acl_xattr.h> | 45 | #include <linux/posix_acl_xattr.h> |
| 46 | #include "ctree.h" | 46 | #include "ctree.h" |
| 47 | #include "disk-io.h" | 47 | #include "disk-io.h" |
| 48 | #include "transaction.h" | 48 | #include "transaction.h" |
| 49 | #include "btrfs_inode.h" | 49 | #include "btrfs_inode.h" |
| 50 | #include "print-tree.h" | 50 | #include "print-tree.h" |
| 51 | #include "ordered-data.h" | 51 | #include "ordered-data.h" |
| 52 | #include "xattr.h" | 52 | #include "xattr.h" |
| 53 | #include "tree-log.h" | 53 | #include "tree-log.h" |
| 54 | #include "volumes.h" | 54 | #include "volumes.h" |
| 55 | #include "compression.h" | 55 | #include "compression.h" |
| 56 | #include "locking.h" | 56 | #include "locking.h" |
| 57 | #include "free-space-cache.h" | 57 | #include "free-space-cache.h" |
| 58 | #include "inode-map.h" | 58 | #include "inode-map.h" |
| 59 | #include "backref.h" | 59 | #include "backref.h" |
| 60 | #include "hash.h" | 60 | #include "hash.h" |
| 61 | #include "props.h" | 61 | #include "props.h" |
| 62 | 62 | ||
| 63 | struct btrfs_iget_args { | 63 | struct btrfs_iget_args { |
| 64 | struct btrfs_key *location; | 64 | struct btrfs_key *location; |
| 65 | struct btrfs_root *root; | 65 | struct btrfs_root *root; |
| 66 | }; | 66 | }; |
| 67 | 67 | ||
| 68 | static const struct inode_operations btrfs_dir_inode_operations; | 68 | static const struct inode_operations btrfs_dir_inode_operations; |
| 69 | static const struct inode_operations btrfs_symlink_inode_operations; | 69 | static const struct inode_operations btrfs_symlink_inode_operations; |
| 70 | static const struct inode_operations btrfs_dir_ro_inode_operations; | 70 | static const struct inode_operations btrfs_dir_ro_inode_operations; |
| 71 | static const struct inode_operations btrfs_special_inode_operations; | 71 | static const struct inode_operations btrfs_special_inode_operations; |
| 72 | static const struct inode_operations btrfs_file_inode_operations; | 72 | static const struct inode_operations btrfs_file_inode_operations; |
| 73 | static const struct address_space_operations btrfs_aops; | 73 | static const struct address_space_operations btrfs_aops; |
| 74 | static const struct address_space_operations btrfs_symlink_aops; | 74 | static const struct address_space_operations btrfs_symlink_aops; |
| 75 | static const struct file_operations btrfs_dir_file_operations; | 75 | static const struct file_operations btrfs_dir_file_operations; |
| 76 | static struct extent_io_ops btrfs_extent_io_ops; | 76 | static struct extent_io_ops btrfs_extent_io_ops; |
| 77 | 77 | ||
| 78 | static struct kmem_cache *btrfs_inode_cachep; | 78 | static struct kmem_cache *btrfs_inode_cachep; |
| 79 | static struct kmem_cache *btrfs_delalloc_work_cachep; | 79 | static struct kmem_cache *btrfs_delalloc_work_cachep; |
| 80 | struct kmem_cache *btrfs_trans_handle_cachep; | 80 | struct kmem_cache *btrfs_trans_handle_cachep; |
| 81 | struct kmem_cache *btrfs_transaction_cachep; | 81 | struct kmem_cache *btrfs_transaction_cachep; |
| 82 | struct kmem_cache *btrfs_path_cachep; | 82 | struct kmem_cache *btrfs_path_cachep; |
| 83 | struct kmem_cache *btrfs_free_space_cachep; | 83 | struct kmem_cache *btrfs_free_space_cachep; |
| 84 | 84 | ||
| 85 | #define S_SHIFT 12 | 85 | #define S_SHIFT 12 |
| 86 | static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { | 86 | static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { |
| 87 | [S_IFREG >> S_SHIFT] = BTRFS_FT_REG_FILE, | 87 | [S_IFREG >> S_SHIFT] = BTRFS_FT_REG_FILE, |
| 88 | [S_IFDIR >> S_SHIFT] = BTRFS_FT_DIR, | 88 | [S_IFDIR >> S_SHIFT] = BTRFS_FT_DIR, |
| 89 | [S_IFCHR >> S_SHIFT] = BTRFS_FT_CHRDEV, | 89 | [S_IFCHR >> S_SHIFT] = BTRFS_FT_CHRDEV, |
| 90 | [S_IFBLK >> S_SHIFT] = BTRFS_FT_BLKDEV, | 90 | [S_IFBLK >> S_SHIFT] = BTRFS_FT_BLKDEV, |
| 91 | [S_IFIFO >> S_SHIFT] = BTRFS_FT_FIFO, | 91 | [S_IFIFO >> S_SHIFT] = BTRFS_FT_FIFO, |
| 92 | [S_IFSOCK >> S_SHIFT] = BTRFS_FT_SOCK, | 92 | [S_IFSOCK >> S_SHIFT] = BTRFS_FT_SOCK, |
| 93 | [S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK, | 93 | [S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK, |
| 94 | }; | 94 | }; |
| 95 | 95 | ||
| 96 | static int btrfs_setsize(struct inode *inode, struct iattr *attr); | 96 | static int btrfs_setsize(struct inode *inode, struct iattr *attr); |
| 97 | static int btrfs_truncate(struct inode *inode); | 97 | static int btrfs_truncate(struct inode *inode); |
| 98 | static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent); | 98 | static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent); |
| 99 | static noinline int cow_file_range(struct inode *inode, | 99 | static noinline int cow_file_range(struct inode *inode, |
| 100 | struct page *locked_page, | 100 | struct page *locked_page, |
| 101 | u64 start, u64 end, int *page_started, | 101 | u64 start, u64 end, int *page_started, |
| 102 | unsigned long *nr_written, int unlock); | 102 | unsigned long *nr_written, int unlock); |
| 103 | static struct extent_map *create_pinned_em(struct inode *inode, u64 start, | 103 | static struct extent_map *create_pinned_em(struct inode *inode, u64 start, |
| 104 | u64 len, u64 orig_start, | 104 | u64 len, u64 orig_start, |
| 105 | u64 block_start, u64 block_len, | 105 | u64 block_start, u64 block_len, |
| 106 | u64 orig_block_len, u64 ram_bytes, | 106 | u64 orig_block_len, u64 ram_bytes, |
| 107 | int type); | 107 | int type); |
| 108 | 108 | ||
| 109 | static int btrfs_dirty_inode(struct inode *inode); | 109 | static int btrfs_dirty_inode(struct inode *inode); |
| 110 | 110 | ||
| 111 | static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, | 111 | static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, |
| 112 | struct inode *inode, struct inode *dir, | 112 | struct inode *inode, struct inode *dir, |
| 113 | const struct qstr *qstr) | 113 | const struct qstr *qstr) |
| 114 | { | 114 | { |
| 115 | int err; | 115 | int err; |
| 116 | 116 | ||
| 117 | err = btrfs_init_acl(trans, inode, dir); | 117 | err = btrfs_init_acl(trans, inode, dir); |
| 118 | if (!err) | 118 | if (!err) |
| 119 | err = btrfs_xattr_security_init(trans, inode, dir, qstr); | 119 | err = btrfs_xattr_security_init(trans, inode, dir, qstr); |
| 120 | return err; | 120 | return err; |
| 121 | } | 121 | } |
| 122 | 122 | ||
| 123 | /* | 123 | /* |
| 124 | * this does all the hard work for inserting an inline extent into | 124 | * this does all the hard work for inserting an inline extent into |
| 125 | * the btree. The caller should have done a btrfs_drop_extents so that | 125 | * the btree. The caller should have done a btrfs_drop_extents so that |
| 126 | * no overlapping inline items exist in the btree | 126 | * no overlapping inline items exist in the btree |
| 127 | */ | 127 | */ |
| 128 | static int insert_inline_extent(struct btrfs_trans_handle *trans, | 128 | static int insert_inline_extent(struct btrfs_trans_handle *trans, |
| 129 | struct btrfs_path *path, int extent_inserted, | 129 | struct btrfs_path *path, int extent_inserted, |
| 130 | struct btrfs_root *root, struct inode *inode, | 130 | struct btrfs_root *root, struct inode *inode, |
| 131 | u64 start, size_t size, size_t compressed_size, | 131 | u64 start, size_t size, size_t compressed_size, |
| 132 | int compress_type, | 132 | int compress_type, |
| 133 | struct page **compressed_pages) | 133 | struct page **compressed_pages) |
| 134 | { | 134 | { |
| 135 | struct extent_buffer *leaf; | 135 | struct extent_buffer *leaf; |
| 136 | struct page *page = NULL; | 136 | struct page *page = NULL; |
| 137 | char *kaddr; | 137 | char *kaddr; |
| 138 | unsigned long ptr; | 138 | unsigned long ptr; |
| 139 | struct btrfs_file_extent_item *ei; | 139 | struct btrfs_file_extent_item *ei; |
| 140 | int err = 0; | 140 | int err = 0; |
| 141 | int ret; | 141 | int ret; |
| 142 | size_t cur_size = size; | 142 | size_t cur_size = size; |
| 143 | unsigned long offset; | 143 | unsigned long offset; |
| 144 | 144 | ||
| 145 | if (compressed_size && compressed_pages) | 145 | if (compressed_size && compressed_pages) |
| 146 | cur_size = compressed_size; | 146 | cur_size = compressed_size; |
| 147 | 147 | ||
| 148 | inode_add_bytes(inode, size); | 148 | inode_add_bytes(inode, size); |
| 149 | 149 | ||
| 150 | if (!extent_inserted) { | 150 | if (!extent_inserted) { |
| 151 | struct btrfs_key key; | 151 | struct btrfs_key key; |
| 152 | size_t datasize; | 152 | size_t datasize; |
| 153 | 153 | ||
| 154 | key.objectid = btrfs_ino(inode); | 154 | key.objectid = btrfs_ino(inode); |
| 155 | key.offset = start; | 155 | key.offset = start; |
| 156 | key.type = BTRFS_EXTENT_DATA_KEY; | 156 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 157 | 157 | ||
| 158 | datasize = btrfs_file_extent_calc_inline_size(cur_size); | 158 | datasize = btrfs_file_extent_calc_inline_size(cur_size); |
| 159 | path->leave_spinning = 1; | 159 | path->leave_spinning = 1; |
| 160 | ret = btrfs_insert_empty_item(trans, root, path, &key, | 160 | ret = btrfs_insert_empty_item(trans, root, path, &key, |
| 161 | datasize); | 161 | datasize); |
| 162 | if (ret) { | 162 | if (ret) { |
| 163 | err = ret; | 163 | err = ret; |
| 164 | goto fail; | 164 | goto fail; |
| 165 | } | 165 | } |
| 166 | } | 166 | } |
| 167 | leaf = path->nodes[0]; | 167 | leaf = path->nodes[0]; |
| 168 | ei = btrfs_item_ptr(leaf, path->slots[0], | 168 | ei = btrfs_item_ptr(leaf, path->slots[0], |
| 169 | struct btrfs_file_extent_item); | 169 | struct btrfs_file_extent_item); |
| 170 | btrfs_set_file_extent_generation(leaf, ei, trans->transid); | 170 | btrfs_set_file_extent_generation(leaf, ei, trans->transid); |
| 171 | btrfs_set_file_extent_type(leaf, ei, BTRFS_FILE_EXTENT_INLINE); | 171 | btrfs_set_file_extent_type(leaf, ei, BTRFS_FILE_EXTENT_INLINE); |
| 172 | btrfs_set_file_extent_encryption(leaf, ei, 0); | 172 | btrfs_set_file_extent_encryption(leaf, ei, 0); |
| 173 | btrfs_set_file_extent_other_encoding(leaf, ei, 0); | 173 | btrfs_set_file_extent_other_encoding(leaf, ei, 0); |
| 174 | btrfs_set_file_extent_ram_bytes(leaf, ei, size); | 174 | btrfs_set_file_extent_ram_bytes(leaf, ei, size); |
| 175 | ptr = btrfs_file_extent_inline_start(ei); | 175 | ptr = btrfs_file_extent_inline_start(ei); |
| 176 | 176 | ||
| 177 | if (compress_type != BTRFS_COMPRESS_NONE) { | 177 | if (compress_type != BTRFS_COMPRESS_NONE) { |
| 178 | struct page *cpage; | 178 | struct page *cpage; |
| 179 | int i = 0; | 179 | int i = 0; |
| 180 | while (compressed_size > 0) { | 180 | while (compressed_size > 0) { |
| 181 | cpage = compressed_pages[i]; | 181 | cpage = compressed_pages[i]; |
| 182 | cur_size = min_t(unsigned long, compressed_size, | 182 | cur_size = min_t(unsigned long, compressed_size, |
| 183 | PAGE_CACHE_SIZE); | 183 | PAGE_CACHE_SIZE); |
| 184 | 184 | ||
| 185 | kaddr = kmap_atomic(cpage); | 185 | kaddr = kmap_atomic(cpage); |
| 186 | write_extent_buffer(leaf, kaddr, ptr, cur_size); | 186 | write_extent_buffer(leaf, kaddr, ptr, cur_size); |
| 187 | kunmap_atomic(kaddr); | 187 | kunmap_atomic(kaddr); |
| 188 | 188 | ||
| 189 | i++; | 189 | i++; |
| 190 | ptr += cur_size; | 190 | ptr += cur_size; |
| 191 | compressed_size -= cur_size; | 191 | compressed_size -= cur_size; |
| 192 | } | 192 | } |
| 193 | btrfs_set_file_extent_compression(leaf, ei, | 193 | btrfs_set_file_extent_compression(leaf, ei, |
| 194 | compress_type); | 194 | compress_type); |
| 195 | } else { | 195 | } else { |
| 196 | page = find_get_page(inode->i_mapping, | 196 | page = find_get_page(inode->i_mapping, |
| 197 | start >> PAGE_CACHE_SHIFT); | 197 | start >> PAGE_CACHE_SHIFT); |
| 198 | btrfs_set_file_extent_compression(leaf, ei, 0); | 198 | btrfs_set_file_extent_compression(leaf, ei, 0); |
| 199 | kaddr = kmap_atomic(page); | 199 | kaddr = kmap_atomic(page); |
| 200 | offset = start & (PAGE_CACHE_SIZE - 1); | 200 | offset = start & (PAGE_CACHE_SIZE - 1); |
| 201 | write_extent_buffer(leaf, kaddr + offset, ptr, size); | 201 | write_extent_buffer(leaf, kaddr + offset, ptr, size); |
| 202 | kunmap_atomic(kaddr); | 202 | kunmap_atomic(kaddr); |
| 203 | page_cache_release(page); | 203 | page_cache_release(page); |
| 204 | } | 204 | } |
| 205 | btrfs_mark_buffer_dirty(leaf); | 205 | btrfs_mark_buffer_dirty(leaf); |
| 206 | btrfs_release_path(path); | 206 | btrfs_release_path(path); |
| 207 | 207 | ||
| 208 | /* | 208 | /* |
| 209 | * we're an inline extent, so nobody can | 209 | * we're an inline extent, so nobody can |
| 210 | * extend the file past i_size without locking | 210 | * extend the file past i_size without locking |
| 211 | * a page we already have locked. | 211 | * a page we already have locked. |
| 212 | * | 212 | * |
| 213 | * We must do any isize and inode updates | 213 | * We must do any isize and inode updates |
| 214 | * before we unlock the pages. Otherwise we | 214 | * before we unlock the pages. Otherwise we |
| 215 | * could end up racing with unlink. | 215 | * could end up racing with unlink. |
| 216 | */ | 216 | */ |
| 217 | BTRFS_I(inode)->disk_i_size = inode->i_size; | 217 | BTRFS_I(inode)->disk_i_size = inode->i_size; |
| 218 | ret = btrfs_update_inode(trans, root, inode); | 218 | ret = btrfs_update_inode(trans, root, inode); |
| 219 | 219 | ||
| 220 | return ret; | 220 | return ret; |
| 221 | fail: | 221 | fail: |
| 222 | return err; | 222 | return err; |
| 223 | } | 223 | } |
| 224 | 224 | ||
| 225 | 225 | ||
| 226 | /* | 226 | /* |
| 227 | * conditionally insert an inline extent into the file. This | 227 | * conditionally insert an inline extent into the file. This |
| 228 | * does the checks required to make sure the data is small enough | 228 | * does the checks required to make sure the data is small enough |
| 229 | * to fit as an inline extent. | 229 | * to fit as an inline extent. |
| 230 | */ | 230 | */ |
| 231 | static noinline int cow_file_range_inline(struct btrfs_root *root, | 231 | static noinline int cow_file_range_inline(struct btrfs_root *root, |
| 232 | struct inode *inode, u64 start, | 232 | struct inode *inode, u64 start, |
| 233 | u64 end, size_t compressed_size, | 233 | u64 end, size_t compressed_size, |
| 234 | int compress_type, | 234 | int compress_type, |
| 235 | struct page **compressed_pages) | 235 | struct page **compressed_pages) |
| 236 | { | 236 | { |
| 237 | struct btrfs_trans_handle *trans; | 237 | struct btrfs_trans_handle *trans; |
| 238 | u64 isize = i_size_read(inode); | 238 | u64 isize = i_size_read(inode); |
| 239 | u64 actual_end = min(end + 1, isize); | 239 | u64 actual_end = min(end + 1, isize); |
| 240 | u64 inline_len = actual_end - start; | 240 | u64 inline_len = actual_end - start; |
| 241 | u64 aligned_end = ALIGN(end, root->sectorsize); | 241 | u64 aligned_end = ALIGN(end, root->sectorsize); |
| 242 | u64 data_len = inline_len; | 242 | u64 data_len = inline_len; |
| 243 | int ret; | 243 | int ret; |
| 244 | struct btrfs_path *path; | 244 | struct btrfs_path *path; |
| 245 | int extent_inserted = 0; | 245 | int extent_inserted = 0; |
| 246 | u32 extent_item_size; | 246 | u32 extent_item_size; |
| 247 | 247 | ||
| 248 | if (compressed_size) | 248 | if (compressed_size) |
| 249 | data_len = compressed_size; | 249 | data_len = compressed_size; |
| 250 | 250 | ||
| 251 | if (start > 0 || | 251 | if (start > 0 || |
| 252 | actual_end > PAGE_CACHE_SIZE || | 252 | actual_end > PAGE_CACHE_SIZE || |
| 253 | data_len > BTRFS_MAX_INLINE_DATA_SIZE(root) || | 253 | data_len > BTRFS_MAX_INLINE_DATA_SIZE(root) || |
| 254 | (!compressed_size && | 254 | (!compressed_size && |
| 255 | (actual_end & (root->sectorsize - 1)) == 0) || | 255 | (actual_end & (root->sectorsize - 1)) == 0) || |
| 256 | end + 1 < isize || | 256 | end + 1 < isize || |
| 257 | data_len > root->fs_info->max_inline) { | 257 | data_len > root->fs_info->max_inline) { |
| 258 | return 1; | 258 | return 1; |
| 259 | } | 259 | } |
| 260 | 260 | ||
| 261 | path = btrfs_alloc_path(); | 261 | path = btrfs_alloc_path(); |
| 262 | if (!path) | 262 | if (!path) |
| 263 | return -ENOMEM; | 263 | return -ENOMEM; |
| 264 | 264 | ||
| 265 | trans = btrfs_join_transaction(root); | 265 | trans = btrfs_join_transaction(root); |
| 266 | if (IS_ERR(trans)) { | 266 | if (IS_ERR(trans)) { |
| 267 | btrfs_free_path(path); | 267 | btrfs_free_path(path); |
| 268 | return PTR_ERR(trans); | 268 | return PTR_ERR(trans); |
| 269 | } | 269 | } |
| 270 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; | 270 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; |
| 271 | 271 | ||
| 272 | if (compressed_size && compressed_pages) | 272 | if (compressed_size && compressed_pages) |
| 273 | extent_item_size = btrfs_file_extent_calc_inline_size( | 273 | extent_item_size = btrfs_file_extent_calc_inline_size( |
| 274 | compressed_size); | 274 | compressed_size); |
| 275 | else | 275 | else |
| 276 | extent_item_size = btrfs_file_extent_calc_inline_size( | 276 | extent_item_size = btrfs_file_extent_calc_inline_size( |
| 277 | inline_len); | 277 | inline_len); |
| 278 | 278 | ||
| 279 | ret = __btrfs_drop_extents(trans, root, inode, path, | 279 | ret = __btrfs_drop_extents(trans, root, inode, path, |
| 280 | start, aligned_end, NULL, | 280 | start, aligned_end, NULL, |
| 281 | 1, 1, extent_item_size, &extent_inserted); | 281 | 1, 1, extent_item_size, &extent_inserted); |
| 282 | if (ret) { | 282 | if (ret) { |
| 283 | btrfs_abort_transaction(trans, root, ret); | 283 | btrfs_abort_transaction(trans, root, ret); |
| 284 | goto out; | 284 | goto out; |
| 285 | } | 285 | } |
| 286 | 286 | ||
| 287 | if (isize > actual_end) | 287 | if (isize > actual_end) |
| 288 | inline_len = min_t(u64, isize, actual_end); | 288 | inline_len = min_t(u64, isize, actual_end); |
| 289 | ret = insert_inline_extent(trans, path, extent_inserted, | 289 | ret = insert_inline_extent(trans, path, extent_inserted, |
| 290 | root, inode, start, | 290 | root, inode, start, |
| 291 | inline_len, compressed_size, | 291 | inline_len, compressed_size, |
| 292 | compress_type, compressed_pages); | 292 | compress_type, compressed_pages); |
| 293 | if (ret && ret != -ENOSPC) { | 293 | if (ret && ret != -ENOSPC) { |
| 294 | btrfs_abort_transaction(trans, root, ret); | 294 | btrfs_abort_transaction(trans, root, ret); |
| 295 | goto out; | 295 | goto out; |
| 296 | } else if (ret == -ENOSPC) { | 296 | } else if (ret == -ENOSPC) { |
| 297 | ret = 1; | 297 | ret = 1; |
| 298 | goto out; | 298 | goto out; |
| 299 | } | 299 | } |
| 300 | 300 | ||
| 301 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags); | 301 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags); |
| 302 | btrfs_delalloc_release_metadata(inode, end + 1 - start); | 302 | btrfs_delalloc_release_metadata(inode, end + 1 - start); |
| 303 | btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0); | 303 | btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0); |
| 304 | out: | 304 | out: |
| 305 | btrfs_free_path(path); | 305 | btrfs_free_path(path); |
| 306 | btrfs_end_transaction(trans, root); | 306 | btrfs_end_transaction(trans, root); |
| 307 | return ret; | 307 | return ret; |
| 308 | } | 308 | } |
| 309 | 309 | ||
| 310 | struct async_extent { | 310 | struct async_extent { |
| 311 | u64 start; | 311 | u64 start; |
| 312 | u64 ram_size; | 312 | u64 ram_size; |
| 313 | u64 compressed_size; | 313 | u64 compressed_size; |
| 314 | struct page **pages; | 314 | struct page **pages; |
| 315 | unsigned long nr_pages; | 315 | unsigned long nr_pages; |
| 316 | int compress_type; | 316 | int compress_type; |
| 317 | struct list_head list; | 317 | struct list_head list; |
| 318 | }; | 318 | }; |
| 319 | 319 | ||
| 320 | struct async_cow { | 320 | struct async_cow { |
| 321 | struct inode *inode; | 321 | struct inode *inode; |
| 322 | struct btrfs_root *root; | 322 | struct btrfs_root *root; |
| 323 | struct page *locked_page; | 323 | struct page *locked_page; |
| 324 | u64 start; | 324 | u64 start; |
| 325 | u64 end; | 325 | u64 end; |
| 326 | struct list_head extents; | 326 | struct list_head extents; |
| 327 | struct btrfs_work work; | 327 | struct btrfs_work work; |
| 328 | }; | 328 | }; |
| 329 | 329 | ||
| 330 | static noinline int add_async_extent(struct async_cow *cow, | 330 | static noinline int add_async_extent(struct async_cow *cow, |
| 331 | u64 start, u64 ram_size, | 331 | u64 start, u64 ram_size, |
| 332 | u64 compressed_size, | 332 | u64 compressed_size, |
| 333 | struct page **pages, | 333 | struct page **pages, |
| 334 | unsigned long nr_pages, | 334 | unsigned long nr_pages, |
| 335 | int compress_type) | 335 | int compress_type) |
| 336 | { | 336 | { |
| 337 | struct async_extent *async_extent; | 337 | struct async_extent *async_extent; |
| 338 | 338 | ||
| 339 | async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS); | 339 | async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS); |
| 340 | BUG_ON(!async_extent); /* -ENOMEM */ | 340 | BUG_ON(!async_extent); /* -ENOMEM */ |
| 341 | async_extent->start = start; | 341 | async_extent->start = start; |
| 342 | async_extent->ram_size = ram_size; | 342 | async_extent->ram_size = ram_size; |
| 343 | async_extent->compressed_size = compressed_size; | 343 | async_extent->compressed_size = compressed_size; |
| 344 | async_extent->pages = pages; | 344 | async_extent->pages = pages; |
| 345 | async_extent->nr_pages = nr_pages; | 345 | async_extent->nr_pages = nr_pages; |
| 346 | async_extent->compress_type = compress_type; | 346 | async_extent->compress_type = compress_type; |
| 347 | list_add_tail(&async_extent->list, &cow->extents); | 347 | list_add_tail(&async_extent->list, &cow->extents); |
| 348 | return 0; | 348 | return 0; |
| 349 | } | 349 | } |
| 350 | 350 | ||
| 351 | static inline int inode_need_compress(struct inode *inode) | 351 | static inline int inode_need_compress(struct inode *inode) |
| 352 | { | 352 | { |
| 353 | struct btrfs_root *root = BTRFS_I(inode)->root; | 353 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 354 | 354 | ||
| 355 | /* force compress */ | 355 | /* force compress */ |
| 356 | if (btrfs_test_opt(root, FORCE_COMPRESS)) | 356 | if (btrfs_test_opt(root, FORCE_COMPRESS)) |
| 357 | return 1; | 357 | return 1; |
| 358 | /* bad compression ratios */ | 358 | /* bad compression ratios */ |
| 359 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) | 359 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) |
| 360 | return 0; | 360 | return 0; |
| 361 | if (btrfs_test_opt(root, COMPRESS) || | 361 | if (btrfs_test_opt(root, COMPRESS) || |
| 362 | BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS || | 362 | BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS || |
| 363 | BTRFS_I(inode)->force_compress) | 363 | BTRFS_I(inode)->force_compress) |
| 364 | return 1; | 364 | return 1; |
| 365 | return 0; | 365 | return 0; |
| 366 | } | 366 | } |
| 367 | 367 | ||
| 368 | /* | 368 | /* |
| 369 | * we create compressed extents in two phases. The first | 369 | * we create compressed extents in two phases. The first |
| 370 | * phase compresses a range of pages that have already been | 370 | * phase compresses a range of pages that have already been |
| 371 | * locked (both pages and state bits are locked). | 371 | * locked (both pages and state bits are locked). |
| 372 | * | 372 | * |
| 373 | * This is done inside an ordered work queue, and the compression | 373 | * This is done inside an ordered work queue, and the compression |
| 374 | * is spread across many cpus. The actual IO submission is step | 374 | * is spread across many cpus. The actual IO submission is step |
| 375 | * two, and the ordered work queue takes care of making sure that | 375 | * two, and the ordered work queue takes care of making sure that |
| 376 | * happens in the same order things were put onto the queue by | 376 | * happens in the same order things were put onto the queue by |
| 377 | * writepages and friends. | 377 | * writepages and friends. |
| 378 | * | 378 | * |
| 379 | * If this code finds it can't get good compression, it puts an | 379 | * If this code finds it can't get good compression, it puts an |
| 380 | * entry onto the work queue to write the uncompressed bytes. This | 380 | * entry onto the work queue to write the uncompressed bytes. This |
| 381 | * makes sure that both compressed inodes and uncompressed inodes | 381 | * makes sure that both compressed inodes and uncompressed inodes |
| 382 | * are written in the same order that the flusher thread sent them | 382 | * are written in the same order that the flusher thread sent them |
| 383 | * down. | 383 | * down. |
| 384 | */ | 384 | */ |
| 385 | static noinline void compress_file_range(struct inode *inode, | 385 | static noinline void compress_file_range(struct inode *inode, |
| 386 | struct page *locked_page, | 386 | struct page *locked_page, |
| 387 | u64 start, u64 end, | 387 | u64 start, u64 end, |
| 388 | struct async_cow *async_cow, | 388 | struct async_cow *async_cow, |
| 389 | int *num_added) | 389 | int *num_added) |
| 390 | { | 390 | { |
| 391 | struct btrfs_root *root = BTRFS_I(inode)->root; | 391 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 392 | u64 num_bytes; | 392 | u64 num_bytes; |
| 393 | u64 blocksize = root->sectorsize; | 393 | u64 blocksize = root->sectorsize; |
| 394 | u64 actual_end; | 394 | u64 actual_end; |
| 395 | u64 isize = i_size_read(inode); | 395 | u64 isize = i_size_read(inode); |
| 396 | int ret = 0; | 396 | int ret = 0; |
| 397 | struct page **pages = NULL; | 397 | struct page **pages = NULL; |
| 398 | unsigned long nr_pages; | 398 | unsigned long nr_pages; |
| 399 | unsigned long nr_pages_ret = 0; | 399 | unsigned long nr_pages_ret = 0; |
| 400 | unsigned long total_compressed = 0; | 400 | unsigned long total_compressed = 0; |
| 401 | unsigned long total_in = 0; | 401 | unsigned long total_in = 0; |
| 402 | unsigned long max_compressed = 128 * 1024; | 402 | unsigned long max_compressed = 128 * 1024; |
| 403 | unsigned long max_uncompressed = 128 * 1024; | 403 | unsigned long max_uncompressed = 128 * 1024; |
| 404 | int i; | 404 | int i; |
| 405 | int will_compress; | 405 | int will_compress; |
| 406 | int compress_type = root->fs_info->compress_type; | 406 | int compress_type = root->fs_info->compress_type; |
| 407 | int redirty = 0; | 407 | int redirty = 0; |
| 408 | 408 | ||
| 409 | /* if this is a small write inside eof, kick off a defrag */ | 409 | /* if this is a small write inside eof, kick off a defrag */ |
| 410 | if ((end - start + 1) < 16 * 1024 && | 410 | if ((end - start + 1) < 16 * 1024 && |
| 411 | (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) | 411 | (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) |
| 412 | btrfs_add_inode_defrag(NULL, inode); | 412 | btrfs_add_inode_defrag(NULL, inode); |
| 413 | 413 | ||
| 414 | actual_end = min_t(u64, isize, end + 1); | 414 | actual_end = min_t(u64, isize, end + 1); |
| 415 | again: | 415 | again: |
| 416 | will_compress = 0; | 416 | will_compress = 0; |
| 417 | nr_pages = (end >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT) + 1; | 417 | nr_pages = (end >> PAGE_CACHE_SHIFT) - (start >> PAGE_CACHE_SHIFT) + 1; |
| 418 | nr_pages = min(nr_pages, (128 * 1024UL) / PAGE_CACHE_SIZE); | 418 | nr_pages = min(nr_pages, (128 * 1024UL) / PAGE_CACHE_SIZE); |
| 419 | 419 | ||
| 420 | /* | 420 | /* |
| 421 | * we don't want to send crud past the end of i_size through | 421 | * we don't want to send crud past the end of i_size through |
| 422 | * compression, that's just a waste of CPU time. So, if the | 422 | * compression, that's just a waste of CPU time. So, if the |
| 423 | * end of the file is before the start of our current | 423 | * end of the file is before the start of our current |
| 424 | * requested range of bytes, we bail out to the uncompressed | 424 | * requested range of bytes, we bail out to the uncompressed |
| 425 | * cleanup code that can deal with all of this. | 425 | * cleanup code that can deal with all of this. |
| 426 | * | 426 | * |
| 427 | * It isn't really the fastest way to fix things, but this is a | 427 | * It isn't really the fastest way to fix things, but this is a |
| 428 | * very uncommon corner. | 428 | * very uncommon corner. |
| 429 | */ | 429 | */ |
| 430 | if (actual_end <= start) | 430 | if (actual_end <= start) |
| 431 | goto cleanup_and_bail_uncompressed; | 431 | goto cleanup_and_bail_uncompressed; |
| 432 | 432 | ||
| 433 | total_compressed = actual_end - start; | 433 | total_compressed = actual_end - start; |
| 434 | 434 | ||
| 435 | /* | 435 | /* |
| 436 | * skip compression for a small file range(<=blocksize) that | 436 | * skip compression for a small file range(<=blocksize) that |
| 437 | * isn't an inline extent, since it dosen't save disk space at all. | 437 | * isn't an inline extent, since it dosen't save disk space at all. |
| 438 | */ | 438 | */ |
| 439 | if (total_compressed <= blocksize && | 439 | if (total_compressed <= blocksize && |
| 440 | (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) | 440 | (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) |
| 441 | goto cleanup_and_bail_uncompressed; | 441 | goto cleanup_and_bail_uncompressed; |
| 442 | 442 | ||
| 443 | /* we want to make sure that amount of ram required to uncompress | 443 | /* we want to make sure that amount of ram required to uncompress |
| 444 | * an extent is reasonable, so we limit the total size in ram | 444 | * an extent is reasonable, so we limit the total size in ram |
| 445 | * of a compressed extent to 128k. This is a crucial number | 445 | * of a compressed extent to 128k. This is a crucial number |
| 446 | * because it also controls how easily we can spread reads across | 446 | * because it also controls how easily we can spread reads across |
| 447 | * cpus for decompression. | 447 | * cpus for decompression. |
| 448 | * | 448 | * |
| 449 | * We also want to make sure the amount of IO required to do | 449 | * We also want to make sure the amount of IO required to do |
| 450 | * a random read is reasonably small, so we limit the size of | 450 | * a random read is reasonably small, so we limit the size of |
| 451 | * a compressed extent to 128k. | 451 | * a compressed extent to 128k. |
| 452 | */ | 452 | */ |
| 453 | total_compressed = min(total_compressed, max_uncompressed); | 453 | total_compressed = min(total_compressed, max_uncompressed); |
| 454 | num_bytes = ALIGN(end - start + 1, blocksize); | 454 | num_bytes = ALIGN(end - start + 1, blocksize); |
| 455 | num_bytes = max(blocksize, num_bytes); | 455 | num_bytes = max(blocksize, num_bytes); |
| 456 | total_in = 0; | 456 | total_in = 0; |
| 457 | ret = 0; | 457 | ret = 0; |
| 458 | 458 | ||
| 459 | /* | 459 | /* |
| 460 | * we do compression for mount -o compress and when the | 460 | * we do compression for mount -o compress and when the |
| 461 | * inode has not been flagged as nocompress. This flag can | 461 | * inode has not been flagged as nocompress. This flag can |
| 462 | * change at any time if we discover bad compression ratios. | 462 | * change at any time if we discover bad compression ratios. |
| 463 | */ | 463 | */ |
| 464 | if (inode_need_compress(inode)) { | 464 | if (inode_need_compress(inode)) { |
| 465 | WARN_ON(pages); | 465 | WARN_ON(pages); |
| 466 | pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); | 466 | pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); |
| 467 | if (!pages) { | 467 | if (!pages) { |
| 468 | /* just bail out to the uncompressed code */ | 468 | /* just bail out to the uncompressed code */ |
| 469 | goto cont; | 469 | goto cont; |
| 470 | } | 470 | } |
| 471 | 471 | ||
| 472 | if (BTRFS_I(inode)->force_compress) | 472 | if (BTRFS_I(inode)->force_compress) |
| 473 | compress_type = BTRFS_I(inode)->force_compress; | 473 | compress_type = BTRFS_I(inode)->force_compress; |
| 474 | 474 | ||
| 475 | /* | 475 | /* |
| 476 | * we need to call clear_page_dirty_for_io on each | 476 | * we need to call clear_page_dirty_for_io on each |
| 477 | * page in the range. Otherwise applications with the file | 477 | * page in the range. Otherwise applications with the file |
| 478 | * mmap'd can wander in and change the page contents while | 478 | * mmap'd can wander in and change the page contents while |
| 479 | * we are compressing them. | 479 | * we are compressing them. |
| 480 | * | 480 | * |
| 481 | * If the compression fails for any reason, we set the pages | 481 | * If the compression fails for any reason, we set the pages |
| 482 | * dirty again later on. | 482 | * dirty again later on. |
| 483 | */ | 483 | */ |
| 484 | extent_range_clear_dirty_for_io(inode, start, end); | 484 | extent_range_clear_dirty_for_io(inode, start, end); |
| 485 | redirty = 1; | 485 | redirty = 1; |
| 486 | ret = btrfs_compress_pages(compress_type, | 486 | ret = btrfs_compress_pages(compress_type, |
| 487 | inode->i_mapping, start, | 487 | inode->i_mapping, start, |
| 488 | total_compressed, pages, | 488 | total_compressed, pages, |
| 489 | nr_pages, &nr_pages_ret, | 489 | nr_pages, &nr_pages_ret, |
| 490 | &total_in, | 490 | &total_in, |
| 491 | &total_compressed, | 491 | &total_compressed, |
| 492 | max_compressed); | 492 | max_compressed); |
| 493 | 493 | ||
| 494 | if (!ret) { | 494 | if (!ret) { |
| 495 | unsigned long offset = total_compressed & | 495 | unsigned long offset = total_compressed & |
| 496 | (PAGE_CACHE_SIZE - 1); | 496 | (PAGE_CACHE_SIZE - 1); |
| 497 | struct page *page = pages[nr_pages_ret - 1]; | 497 | struct page *page = pages[nr_pages_ret - 1]; |
| 498 | char *kaddr; | 498 | char *kaddr; |
| 499 | 499 | ||
| 500 | /* zero the tail end of the last page, we might be | 500 | /* zero the tail end of the last page, we might be |
| 501 | * sending it down to disk | 501 | * sending it down to disk |
| 502 | */ | 502 | */ |
| 503 | if (offset) { | 503 | if (offset) { |
| 504 | kaddr = kmap_atomic(page); | 504 | kaddr = kmap_atomic(page); |
| 505 | memset(kaddr + offset, 0, | 505 | memset(kaddr + offset, 0, |
| 506 | PAGE_CACHE_SIZE - offset); | 506 | PAGE_CACHE_SIZE - offset); |
| 507 | kunmap_atomic(kaddr); | 507 | kunmap_atomic(kaddr); |
| 508 | } | 508 | } |
| 509 | will_compress = 1; | 509 | will_compress = 1; |
| 510 | } | 510 | } |
| 511 | } | 511 | } |
| 512 | cont: | 512 | cont: |
| 513 | if (start == 0) { | 513 | if (start == 0) { |
| 514 | /* lets try to make an inline extent */ | 514 | /* lets try to make an inline extent */ |
| 515 | if (ret || total_in < (actual_end - start)) { | 515 | if (ret || total_in < (actual_end - start)) { |
| 516 | /* we didn't compress the entire range, try | 516 | /* we didn't compress the entire range, try |
| 517 | * to make an uncompressed inline extent. | 517 | * to make an uncompressed inline extent. |
| 518 | */ | 518 | */ |
| 519 | ret = cow_file_range_inline(root, inode, start, end, | 519 | ret = cow_file_range_inline(root, inode, start, end, |
| 520 | 0, 0, NULL); | 520 | 0, 0, NULL); |
| 521 | } else { | 521 | } else { |
| 522 | /* try making a compressed inline extent */ | 522 | /* try making a compressed inline extent */ |
| 523 | ret = cow_file_range_inline(root, inode, start, end, | 523 | ret = cow_file_range_inline(root, inode, start, end, |
| 524 | total_compressed, | 524 | total_compressed, |
| 525 | compress_type, pages); | 525 | compress_type, pages); |
| 526 | } | 526 | } |
| 527 | if (ret <= 0) { | 527 | if (ret <= 0) { |
| 528 | unsigned long clear_flags = EXTENT_DELALLOC | | 528 | unsigned long clear_flags = EXTENT_DELALLOC | |
| 529 | EXTENT_DEFRAG; | 529 | EXTENT_DEFRAG; |
| 530 | unsigned long page_error_op; | 530 | unsigned long page_error_op; |
| 531 | 531 | ||
| 532 | clear_flags |= (ret < 0) ? EXTENT_DO_ACCOUNTING : 0; | 532 | clear_flags |= (ret < 0) ? EXTENT_DO_ACCOUNTING : 0; |
| 533 | page_error_op = ret < 0 ? PAGE_SET_ERROR : 0; | 533 | page_error_op = ret < 0 ? PAGE_SET_ERROR : 0; |
| 534 | 534 | ||
| 535 | /* | 535 | /* |
| 536 | * inline extent creation worked or returned error, | 536 | * inline extent creation worked or returned error, |
| 537 | * we don't need to create any more async work items. | 537 | * we don't need to create any more async work items. |
| 538 | * Unlock and free up our temp pages. | 538 | * Unlock and free up our temp pages. |
| 539 | */ | 539 | */ |
| 540 | extent_clear_unlock_delalloc(inode, start, end, NULL, | 540 | extent_clear_unlock_delalloc(inode, start, end, NULL, |
| 541 | clear_flags, PAGE_UNLOCK | | 541 | clear_flags, PAGE_UNLOCK | |
| 542 | PAGE_CLEAR_DIRTY | | 542 | PAGE_CLEAR_DIRTY | |
| 543 | PAGE_SET_WRITEBACK | | 543 | PAGE_SET_WRITEBACK | |
| 544 | page_error_op | | 544 | page_error_op | |
| 545 | PAGE_END_WRITEBACK); | 545 | PAGE_END_WRITEBACK); |
| 546 | goto free_pages_out; | 546 | goto free_pages_out; |
| 547 | } | 547 | } |
| 548 | } | 548 | } |
| 549 | 549 | ||
| 550 | if (will_compress) { | 550 | if (will_compress) { |
| 551 | /* | 551 | /* |
| 552 | * we aren't doing an inline extent round the compressed size | 552 | * we aren't doing an inline extent round the compressed size |
| 553 | * up to a block size boundary so the allocator does sane | 553 | * up to a block size boundary so the allocator does sane |
| 554 | * things | 554 | * things |
| 555 | */ | 555 | */ |
| 556 | total_compressed = ALIGN(total_compressed, blocksize); | 556 | total_compressed = ALIGN(total_compressed, blocksize); |
| 557 | 557 | ||
| 558 | /* | 558 | /* |
| 559 | * one last check to make sure the compression is really a | 559 | * one last check to make sure the compression is really a |
| 560 | * win, compare the page count read with the blocks on disk | 560 | * win, compare the page count read with the blocks on disk |
| 561 | */ | 561 | */ |
| 562 | total_in = ALIGN(total_in, PAGE_CACHE_SIZE); | 562 | total_in = ALIGN(total_in, PAGE_CACHE_SIZE); |
| 563 | if (total_compressed >= total_in) { | 563 | if (total_compressed >= total_in) { |
| 564 | will_compress = 0; | 564 | will_compress = 0; |
| 565 | } else { | 565 | } else { |
| 566 | num_bytes = total_in; | 566 | num_bytes = total_in; |
| 567 | } | 567 | } |
| 568 | } | 568 | } |
| 569 | if (!will_compress && pages) { | 569 | if (!will_compress && pages) { |
| 570 | /* | 570 | /* |
| 571 | * the compression code ran but failed to make things smaller, | 571 | * the compression code ran but failed to make things smaller, |
| 572 | * free any pages it allocated and our page pointer array | 572 | * free any pages it allocated and our page pointer array |
| 573 | */ | 573 | */ |
| 574 | for (i = 0; i < nr_pages_ret; i++) { | 574 | for (i = 0; i < nr_pages_ret; i++) { |
| 575 | WARN_ON(pages[i]->mapping); | 575 | WARN_ON(pages[i]->mapping); |
| 576 | page_cache_release(pages[i]); | 576 | page_cache_release(pages[i]); |
| 577 | } | 577 | } |
| 578 | kfree(pages); | 578 | kfree(pages); |
| 579 | pages = NULL; | 579 | pages = NULL; |
| 580 | total_compressed = 0; | 580 | total_compressed = 0; |
| 581 | nr_pages_ret = 0; | 581 | nr_pages_ret = 0; |
| 582 | 582 | ||
| 583 | /* flag the file so we don't compress in the future */ | 583 | /* flag the file so we don't compress in the future */ |
| 584 | if (!btrfs_test_opt(root, FORCE_COMPRESS) && | 584 | if (!btrfs_test_opt(root, FORCE_COMPRESS) && |
| 585 | !(BTRFS_I(inode)->force_compress)) { | 585 | !(BTRFS_I(inode)->force_compress)) { |
| 586 | BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS; | 586 | BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS; |
| 587 | } | 587 | } |
| 588 | } | 588 | } |
| 589 | if (will_compress) { | 589 | if (will_compress) { |
| 590 | *num_added += 1; | 590 | *num_added += 1; |
| 591 | 591 | ||
| 592 | /* the async work queues will take care of doing actual | 592 | /* the async work queues will take care of doing actual |
| 593 | * allocation on disk for these compressed pages, | 593 | * allocation on disk for these compressed pages, |
| 594 | * and will submit them to the elevator. | 594 | * and will submit them to the elevator. |
| 595 | */ | 595 | */ |
| 596 | add_async_extent(async_cow, start, num_bytes, | 596 | add_async_extent(async_cow, start, num_bytes, |
| 597 | total_compressed, pages, nr_pages_ret, | 597 | total_compressed, pages, nr_pages_ret, |
| 598 | compress_type); | 598 | compress_type); |
| 599 | 599 | ||
| 600 | if (start + num_bytes < end) { | 600 | if (start + num_bytes < end) { |
| 601 | start += num_bytes; | 601 | start += num_bytes; |
| 602 | pages = NULL; | 602 | pages = NULL; |
| 603 | cond_resched(); | 603 | cond_resched(); |
| 604 | goto again; | 604 | goto again; |
| 605 | } | 605 | } |
| 606 | } else { | 606 | } else { |
| 607 | cleanup_and_bail_uncompressed: | 607 | cleanup_and_bail_uncompressed: |
| 608 | /* | 608 | /* |
| 609 | * No compression, but we still need to write the pages in | 609 | * No compression, but we still need to write the pages in |
| 610 | * the file we've been given so far. redirty the locked | 610 | * the file we've been given so far. redirty the locked |
| 611 | * page if it corresponds to our extent and set things up | 611 | * page if it corresponds to our extent and set things up |
| 612 | * for the async work queue to run cow_file_range to do | 612 | * for the async work queue to run cow_file_range to do |
| 613 | * the normal delalloc dance | 613 | * the normal delalloc dance |
| 614 | */ | 614 | */ |
| 615 | if (page_offset(locked_page) >= start && | 615 | if (page_offset(locked_page) >= start && |
| 616 | page_offset(locked_page) <= end) { | 616 | page_offset(locked_page) <= end) { |
| 617 | __set_page_dirty_nobuffers(locked_page); | 617 | __set_page_dirty_nobuffers(locked_page); |
| 618 | /* unlocked later on in the async handlers */ | 618 | /* unlocked later on in the async handlers */ |
| 619 | } | 619 | } |
| 620 | if (redirty) | 620 | if (redirty) |
| 621 | extent_range_redirty_for_io(inode, start, end); | 621 | extent_range_redirty_for_io(inode, start, end); |
| 622 | add_async_extent(async_cow, start, end - start + 1, | 622 | add_async_extent(async_cow, start, end - start + 1, |
| 623 | 0, NULL, 0, BTRFS_COMPRESS_NONE); | 623 | 0, NULL, 0, BTRFS_COMPRESS_NONE); |
| 624 | *num_added += 1; | 624 | *num_added += 1; |
| 625 | } | 625 | } |
| 626 | 626 | ||
| 627 | return; | 627 | return; |
| 628 | 628 | ||
| 629 | free_pages_out: | 629 | free_pages_out: |
| 630 | for (i = 0; i < nr_pages_ret; i++) { | 630 | for (i = 0; i < nr_pages_ret; i++) { |
| 631 | WARN_ON(pages[i]->mapping); | 631 | WARN_ON(pages[i]->mapping); |
| 632 | page_cache_release(pages[i]); | 632 | page_cache_release(pages[i]); |
| 633 | } | 633 | } |
| 634 | kfree(pages); | 634 | kfree(pages); |
| 635 | } | 635 | } |
| 636 | 636 | ||
| 637 | static void free_async_extent_pages(struct async_extent *async_extent) | 637 | static void free_async_extent_pages(struct async_extent *async_extent) |
| 638 | { | 638 | { |
| 639 | int i; | 639 | int i; |
| 640 | 640 | ||
| 641 | if (!async_extent->pages) | 641 | if (!async_extent->pages) |
| 642 | return; | 642 | return; |
| 643 | 643 | ||
| 644 | for (i = 0; i < async_extent->nr_pages; i++) { | 644 | for (i = 0; i < async_extent->nr_pages; i++) { |
| 645 | WARN_ON(async_extent->pages[i]->mapping); | 645 | WARN_ON(async_extent->pages[i]->mapping); |
| 646 | page_cache_release(async_extent->pages[i]); | 646 | page_cache_release(async_extent->pages[i]); |
| 647 | } | 647 | } |
| 648 | kfree(async_extent->pages); | 648 | kfree(async_extent->pages); |
| 649 | async_extent->nr_pages = 0; | 649 | async_extent->nr_pages = 0; |
| 650 | async_extent->pages = NULL; | 650 | async_extent->pages = NULL; |
| 651 | } | 651 | } |
| 652 | 652 | ||
| 653 | /* | 653 | /* |
| 654 | * phase two of compressed writeback. This is the ordered portion | 654 | * phase two of compressed writeback. This is the ordered portion |
| 655 | * of the code, which only gets called in the order the work was | 655 | * of the code, which only gets called in the order the work was |
| 656 | * queued. We walk all the async extents created by compress_file_range | 656 | * queued. We walk all the async extents created by compress_file_range |
| 657 | * and send them down to the disk. | 657 | * and send them down to the disk. |
| 658 | */ | 658 | */ |
| 659 | static noinline void submit_compressed_extents(struct inode *inode, | 659 | static noinline void submit_compressed_extents(struct inode *inode, |
| 660 | struct async_cow *async_cow) | 660 | struct async_cow *async_cow) |
| 661 | { | 661 | { |
| 662 | struct async_extent *async_extent; | 662 | struct async_extent *async_extent; |
| 663 | u64 alloc_hint = 0; | 663 | u64 alloc_hint = 0; |
| 664 | struct btrfs_key ins; | 664 | struct btrfs_key ins; |
| 665 | struct extent_map *em; | 665 | struct extent_map *em; |
| 666 | struct btrfs_root *root = BTRFS_I(inode)->root; | 666 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 667 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; | 667 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; |
| 668 | struct extent_io_tree *io_tree; | 668 | struct extent_io_tree *io_tree; |
| 669 | int ret = 0; | 669 | int ret = 0; |
| 670 | 670 | ||
| 671 | again: | 671 | again: |
| 672 | while (!list_empty(&async_cow->extents)) { | 672 | while (!list_empty(&async_cow->extents)) { |
| 673 | async_extent = list_entry(async_cow->extents.next, | 673 | async_extent = list_entry(async_cow->extents.next, |
| 674 | struct async_extent, list); | 674 | struct async_extent, list); |
| 675 | list_del(&async_extent->list); | 675 | list_del(&async_extent->list); |
| 676 | 676 | ||
| 677 | io_tree = &BTRFS_I(inode)->io_tree; | 677 | io_tree = &BTRFS_I(inode)->io_tree; |
| 678 | 678 | ||
| 679 | retry: | 679 | retry: |
| 680 | /* did the compression code fall back to uncompressed IO? */ | 680 | /* did the compression code fall back to uncompressed IO? */ |
| 681 | if (!async_extent->pages) { | 681 | if (!async_extent->pages) { |
| 682 | int page_started = 0; | 682 | int page_started = 0; |
| 683 | unsigned long nr_written = 0; | 683 | unsigned long nr_written = 0; |
| 684 | 684 | ||
| 685 | lock_extent(io_tree, async_extent->start, | 685 | lock_extent(io_tree, async_extent->start, |
| 686 | async_extent->start + | 686 | async_extent->start + |
| 687 | async_extent->ram_size - 1); | 687 | async_extent->ram_size - 1); |
| 688 | 688 | ||
| 689 | /* allocate blocks */ | 689 | /* allocate blocks */ |
| 690 | ret = cow_file_range(inode, async_cow->locked_page, | 690 | ret = cow_file_range(inode, async_cow->locked_page, |
| 691 | async_extent->start, | 691 | async_extent->start, |
| 692 | async_extent->start + | 692 | async_extent->start + |
| 693 | async_extent->ram_size - 1, | 693 | async_extent->ram_size - 1, |
| 694 | &page_started, &nr_written, 0); | 694 | &page_started, &nr_written, 0); |
| 695 | 695 | ||
| 696 | /* JDM XXX */ | 696 | /* JDM XXX */ |
| 697 | 697 | ||
| 698 | /* | 698 | /* |
| 699 | * if page_started, cow_file_range inserted an | 699 | * if page_started, cow_file_range inserted an |
| 700 | * inline extent and took care of all the unlocking | 700 | * inline extent and took care of all the unlocking |
| 701 | * and IO for us. Otherwise, we need to submit | 701 | * and IO for us. Otherwise, we need to submit |
| 702 | * all those pages down to the drive. | 702 | * all those pages down to the drive. |
| 703 | */ | 703 | */ |
| 704 | if (!page_started && !ret) | 704 | if (!page_started && !ret) |
| 705 | extent_write_locked_range(io_tree, | 705 | extent_write_locked_range(io_tree, |
| 706 | inode, async_extent->start, | 706 | inode, async_extent->start, |
| 707 | async_extent->start + | 707 | async_extent->start + |
| 708 | async_extent->ram_size - 1, | 708 | async_extent->ram_size - 1, |
| 709 | btrfs_get_extent, | 709 | btrfs_get_extent, |
| 710 | WB_SYNC_ALL); | 710 | WB_SYNC_ALL); |
| 711 | else if (ret) | 711 | else if (ret) |
| 712 | unlock_page(async_cow->locked_page); | 712 | unlock_page(async_cow->locked_page); |
| 713 | kfree(async_extent); | 713 | kfree(async_extent); |
| 714 | cond_resched(); | 714 | cond_resched(); |
| 715 | continue; | 715 | continue; |
| 716 | } | 716 | } |
| 717 | 717 | ||
| 718 | lock_extent(io_tree, async_extent->start, | 718 | lock_extent(io_tree, async_extent->start, |
| 719 | async_extent->start + async_extent->ram_size - 1); | 719 | async_extent->start + async_extent->ram_size - 1); |
| 720 | 720 | ||
| 721 | ret = btrfs_reserve_extent(root, | 721 | ret = btrfs_reserve_extent(root, |
| 722 | async_extent->compressed_size, | 722 | async_extent->compressed_size, |
| 723 | async_extent->compressed_size, | 723 | async_extent->compressed_size, |
| 724 | 0, alloc_hint, &ins, 1, 1); | 724 | 0, alloc_hint, &ins, 1, 1); |
| 725 | if (ret) { | 725 | if (ret) { |
| 726 | free_async_extent_pages(async_extent); | 726 | free_async_extent_pages(async_extent); |
| 727 | 727 | ||
| 728 | if (ret == -ENOSPC) { | 728 | if (ret == -ENOSPC) { |
| 729 | unlock_extent(io_tree, async_extent->start, | 729 | unlock_extent(io_tree, async_extent->start, |
| 730 | async_extent->start + | 730 | async_extent->start + |
| 731 | async_extent->ram_size - 1); | 731 | async_extent->ram_size - 1); |
| 732 | 732 | ||
| 733 | /* | 733 | /* |
| 734 | * we need to redirty the pages if we decide to | 734 | * we need to redirty the pages if we decide to |
| 735 | * fallback to uncompressed IO, otherwise we | 735 | * fallback to uncompressed IO, otherwise we |
| 736 | * will not submit these pages down to lower | 736 | * will not submit these pages down to lower |
| 737 | * layers. | 737 | * layers. |
| 738 | */ | 738 | */ |
| 739 | extent_range_redirty_for_io(inode, | 739 | extent_range_redirty_for_io(inode, |
| 740 | async_extent->start, | 740 | async_extent->start, |
| 741 | async_extent->start + | 741 | async_extent->start + |
| 742 | async_extent->ram_size - 1); | 742 | async_extent->ram_size - 1); |
| 743 | 743 | ||
| 744 | goto retry; | 744 | goto retry; |
| 745 | } | 745 | } |
| 746 | goto out_free; | 746 | goto out_free; |
| 747 | } | 747 | } |
| 748 | 748 | ||
| 749 | /* | 749 | /* |
| 750 | * here we're doing allocation and writeback of the | 750 | * here we're doing allocation and writeback of the |
| 751 | * compressed pages | 751 | * compressed pages |
| 752 | */ | 752 | */ |
| 753 | btrfs_drop_extent_cache(inode, async_extent->start, | 753 | btrfs_drop_extent_cache(inode, async_extent->start, |
| 754 | async_extent->start + | 754 | async_extent->start + |
| 755 | async_extent->ram_size - 1, 0); | 755 | async_extent->ram_size - 1, 0); |
| 756 | 756 | ||
| 757 | em = alloc_extent_map(); | 757 | em = alloc_extent_map(); |
| 758 | if (!em) { | 758 | if (!em) { |
| 759 | ret = -ENOMEM; | 759 | ret = -ENOMEM; |
| 760 | goto out_free_reserve; | 760 | goto out_free_reserve; |
| 761 | } | 761 | } |
| 762 | em->start = async_extent->start; | 762 | em->start = async_extent->start; |
| 763 | em->len = async_extent->ram_size; | 763 | em->len = async_extent->ram_size; |
| 764 | em->orig_start = em->start; | 764 | em->orig_start = em->start; |
| 765 | em->mod_start = em->start; | 765 | em->mod_start = em->start; |
| 766 | em->mod_len = em->len; | 766 | em->mod_len = em->len; |
| 767 | 767 | ||
| 768 | em->block_start = ins.objectid; | 768 | em->block_start = ins.objectid; |
| 769 | em->block_len = ins.offset; | 769 | em->block_len = ins.offset; |
| 770 | em->orig_block_len = ins.offset; | 770 | em->orig_block_len = ins.offset; |
| 771 | em->ram_bytes = async_extent->ram_size; | 771 | em->ram_bytes = async_extent->ram_size; |
| 772 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 772 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 773 | em->compress_type = async_extent->compress_type; | 773 | em->compress_type = async_extent->compress_type; |
| 774 | set_bit(EXTENT_FLAG_PINNED, &em->flags); | 774 | set_bit(EXTENT_FLAG_PINNED, &em->flags); |
| 775 | set_bit(EXTENT_FLAG_COMPRESSED, &em->flags); | 775 | set_bit(EXTENT_FLAG_COMPRESSED, &em->flags); |
| 776 | em->generation = -1; | 776 | em->generation = -1; |
| 777 | 777 | ||
| 778 | while (1) { | 778 | while (1) { |
| 779 | write_lock(&em_tree->lock); | 779 | write_lock(&em_tree->lock); |
| 780 | ret = add_extent_mapping(em_tree, em, 1); | 780 | ret = add_extent_mapping(em_tree, em, 1); |
| 781 | write_unlock(&em_tree->lock); | 781 | write_unlock(&em_tree->lock); |
| 782 | if (ret != -EEXIST) { | 782 | if (ret != -EEXIST) { |
| 783 | free_extent_map(em); | 783 | free_extent_map(em); |
| 784 | break; | 784 | break; |
| 785 | } | 785 | } |
| 786 | btrfs_drop_extent_cache(inode, async_extent->start, | 786 | btrfs_drop_extent_cache(inode, async_extent->start, |
| 787 | async_extent->start + | 787 | async_extent->start + |
| 788 | async_extent->ram_size - 1, 0); | 788 | async_extent->ram_size - 1, 0); |
| 789 | } | 789 | } |
| 790 | 790 | ||
| 791 | if (ret) | 791 | if (ret) |
| 792 | goto out_free_reserve; | 792 | goto out_free_reserve; |
| 793 | 793 | ||
| 794 | ret = btrfs_add_ordered_extent_compress(inode, | 794 | ret = btrfs_add_ordered_extent_compress(inode, |
| 795 | async_extent->start, | 795 | async_extent->start, |
| 796 | ins.objectid, | 796 | ins.objectid, |
| 797 | async_extent->ram_size, | 797 | async_extent->ram_size, |
| 798 | ins.offset, | 798 | ins.offset, |
| 799 | BTRFS_ORDERED_COMPRESSED, | 799 | BTRFS_ORDERED_COMPRESSED, |
| 800 | async_extent->compress_type); | 800 | async_extent->compress_type); |
| 801 | if (ret) { | 801 | if (ret) { |
| 802 | btrfs_drop_extent_cache(inode, async_extent->start, | 802 | btrfs_drop_extent_cache(inode, async_extent->start, |
| 803 | async_extent->start + | 803 | async_extent->start + |
| 804 | async_extent->ram_size - 1, 0); | 804 | async_extent->ram_size - 1, 0); |
| 805 | goto out_free_reserve; | 805 | goto out_free_reserve; |
| 806 | } | 806 | } |
| 807 | 807 | ||
| 808 | /* | 808 | /* |
| 809 | * clear dirty, set writeback and unlock the pages. | 809 | * clear dirty, set writeback and unlock the pages. |
| 810 | */ | 810 | */ |
| 811 | extent_clear_unlock_delalloc(inode, async_extent->start, | 811 | extent_clear_unlock_delalloc(inode, async_extent->start, |
| 812 | async_extent->start + | 812 | async_extent->start + |
| 813 | async_extent->ram_size - 1, | 813 | async_extent->ram_size - 1, |
| 814 | NULL, EXTENT_LOCKED | EXTENT_DELALLOC, | 814 | NULL, EXTENT_LOCKED | EXTENT_DELALLOC, |
| 815 | PAGE_UNLOCK | PAGE_CLEAR_DIRTY | | 815 | PAGE_UNLOCK | PAGE_CLEAR_DIRTY | |
| 816 | PAGE_SET_WRITEBACK); | 816 | PAGE_SET_WRITEBACK); |
| 817 | ret = btrfs_submit_compressed_write(inode, | 817 | ret = btrfs_submit_compressed_write(inode, |
| 818 | async_extent->start, | 818 | async_extent->start, |
| 819 | async_extent->ram_size, | 819 | async_extent->ram_size, |
| 820 | ins.objectid, | 820 | ins.objectid, |
| 821 | ins.offset, async_extent->pages, | 821 | ins.offset, async_extent->pages, |
| 822 | async_extent->nr_pages); | 822 | async_extent->nr_pages); |
| 823 | if (ret) { | 823 | if (ret) { |
| 824 | struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; | 824 | struct extent_io_tree *tree = &BTRFS_I(inode)->io_tree; |
| 825 | struct page *p = async_extent->pages[0]; | 825 | struct page *p = async_extent->pages[0]; |
| 826 | const u64 start = async_extent->start; | 826 | const u64 start = async_extent->start; |
| 827 | const u64 end = start + async_extent->ram_size - 1; | 827 | const u64 end = start + async_extent->ram_size - 1; |
| 828 | 828 | ||
| 829 | p->mapping = inode->i_mapping; | 829 | p->mapping = inode->i_mapping; |
| 830 | tree->ops->writepage_end_io_hook(p, start, end, | 830 | tree->ops->writepage_end_io_hook(p, start, end, |
| 831 | NULL, 0); | 831 | NULL, 0); |
| 832 | p->mapping = NULL; | 832 | p->mapping = NULL; |
| 833 | extent_clear_unlock_delalloc(inode, start, end, NULL, 0, | 833 | extent_clear_unlock_delalloc(inode, start, end, NULL, 0, |
| 834 | PAGE_END_WRITEBACK | | 834 | PAGE_END_WRITEBACK | |
| 835 | PAGE_SET_ERROR); | 835 | PAGE_SET_ERROR); |
| 836 | free_async_extent_pages(async_extent); | 836 | free_async_extent_pages(async_extent); |
| 837 | } | 837 | } |
| 838 | alloc_hint = ins.objectid + ins.offset; | 838 | alloc_hint = ins.objectid + ins.offset; |
| 839 | kfree(async_extent); | 839 | kfree(async_extent); |
| 840 | cond_resched(); | 840 | cond_resched(); |
| 841 | } | 841 | } |
| 842 | return; | 842 | return; |
| 843 | out_free_reserve: | 843 | out_free_reserve: |
| 844 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); | 844 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); |
| 845 | out_free: | 845 | out_free: |
| 846 | extent_clear_unlock_delalloc(inode, async_extent->start, | 846 | extent_clear_unlock_delalloc(inode, async_extent->start, |
| 847 | async_extent->start + | 847 | async_extent->start + |
| 848 | async_extent->ram_size - 1, | 848 | async_extent->ram_size - 1, |
| 849 | NULL, EXTENT_LOCKED | EXTENT_DELALLOC | | 849 | NULL, EXTENT_LOCKED | EXTENT_DELALLOC | |
| 850 | EXTENT_DEFRAG | EXTENT_DO_ACCOUNTING, | 850 | EXTENT_DEFRAG | EXTENT_DO_ACCOUNTING, |
| 851 | PAGE_UNLOCK | PAGE_CLEAR_DIRTY | | 851 | PAGE_UNLOCK | PAGE_CLEAR_DIRTY | |
| 852 | PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK | | 852 | PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK | |
| 853 | PAGE_SET_ERROR); | 853 | PAGE_SET_ERROR); |
| 854 | free_async_extent_pages(async_extent); | 854 | free_async_extent_pages(async_extent); |
| 855 | kfree(async_extent); | 855 | kfree(async_extent); |
| 856 | goto again; | 856 | goto again; |
| 857 | } | 857 | } |
| 858 | 858 | ||
| 859 | static u64 get_extent_allocation_hint(struct inode *inode, u64 start, | 859 | static u64 get_extent_allocation_hint(struct inode *inode, u64 start, |
| 860 | u64 num_bytes) | 860 | u64 num_bytes) |
| 861 | { | 861 | { |
| 862 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; | 862 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; |
| 863 | struct extent_map *em; | 863 | struct extent_map *em; |
| 864 | u64 alloc_hint = 0; | 864 | u64 alloc_hint = 0; |
| 865 | 865 | ||
| 866 | read_lock(&em_tree->lock); | 866 | read_lock(&em_tree->lock); |
| 867 | em = search_extent_mapping(em_tree, start, num_bytes); | 867 | em = search_extent_mapping(em_tree, start, num_bytes); |
| 868 | if (em) { | 868 | if (em) { |
| 869 | /* | 869 | /* |
| 870 | * if block start isn't an actual block number then find the | 870 | * if block start isn't an actual block number then find the |
| 871 | * first block in this inode and use that as a hint. If that | 871 | * first block in this inode and use that as a hint. If that |
| 872 | * block is also bogus then just don't worry about it. | 872 | * block is also bogus then just don't worry about it. |
| 873 | */ | 873 | */ |
| 874 | if (em->block_start >= EXTENT_MAP_LAST_BYTE) { | 874 | if (em->block_start >= EXTENT_MAP_LAST_BYTE) { |
| 875 | free_extent_map(em); | 875 | free_extent_map(em); |
| 876 | em = search_extent_mapping(em_tree, 0, 0); | 876 | em = search_extent_mapping(em_tree, 0, 0); |
| 877 | if (em && em->block_start < EXTENT_MAP_LAST_BYTE) | 877 | if (em && em->block_start < EXTENT_MAP_LAST_BYTE) |
| 878 | alloc_hint = em->block_start; | 878 | alloc_hint = em->block_start; |
| 879 | if (em) | 879 | if (em) |
| 880 | free_extent_map(em); | 880 | free_extent_map(em); |
| 881 | } else { | 881 | } else { |
| 882 | alloc_hint = em->block_start; | 882 | alloc_hint = em->block_start; |
| 883 | free_extent_map(em); | 883 | free_extent_map(em); |
| 884 | } | 884 | } |
| 885 | } | 885 | } |
| 886 | read_unlock(&em_tree->lock); | 886 | read_unlock(&em_tree->lock); |
| 887 | 887 | ||
| 888 | return alloc_hint; | 888 | return alloc_hint; |
| 889 | } | 889 | } |
| 890 | 890 | ||
| 891 | /* | 891 | /* |
| 892 | * when extent_io.c finds a delayed allocation range in the file, | 892 | * when extent_io.c finds a delayed allocation range in the file, |
| 893 | * the call backs end up in this code. The basic idea is to | 893 | * the call backs end up in this code. The basic idea is to |
| 894 | * allocate extents on disk for the range, and create ordered data structs | 894 | * allocate extents on disk for the range, and create ordered data structs |
| 895 | * in ram to track those extents. | 895 | * in ram to track those extents. |
| 896 | * | 896 | * |
| 897 | * locked_page is the page that writepage had locked already. We use | 897 | * locked_page is the page that writepage had locked already. We use |
| 898 | * it to make sure we don't do extra locks or unlocks. | 898 | * it to make sure we don't do extra locks or unlocks. |
| 899 | * | 899 | * |
| 900 | * *page_started is set to one if we unlock locked_page and do everything | 900 | * *page_started is set to one if we unlock locked_page and do everything |
| 901 | * required to start IO on it. It may be clean and already done with | 901 | * required to start IO on it. It may be clean and already done with |
| 902 | * IO when we return. | 902 | * IO when we return. |
| 903 | */ | 903 | */ |
| 904 | static noinline int cow_file_range(struct inode *inode, | 904 | static noinline int cow_file_range(struct inode *inode, |
| 905 | struct page *locked_page, | 905 | struct page *locked_page, |
| 906 | u64 start, u64 end, int *page_started, | 906 | u64 start, u64 end, int *page_started, |
| 907 | unsigned long *nr_written, | 907 | unsigned long *nr_written, |
| 908 | int unlock) | 908 | int unlock) |
| 909 | { | 909 | { |
| 910 | struct btrfs_root *root = BTRFS_I(inode)->root; | 910 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 911 | u64 alloc_hint = 0; | 911 | u64 alloc_hint = 0; |
| 912 | u64 num_bytes; | 912 | u64 num_bytes; |
| 913 | unsigned long ram_size; | 913 | unsigned long ram_size; |
| 914 | u64 disk_num_bytes; | 914 | u64 disk_num_bytes; |
| 915 | u64 cur_alloc_size; | 915 | u64 cur_alloc_size; |
| 916 | u64 blocksize = root->sectorsize; | 916 | u64 blocksize = root->sectorsize; |
| 917 | struct btrfs_key ins; | 917 | struct btrfs_key ins; |
| 918 | struct extent_map *em; | 918 | struct extent_map *em; |
| 919 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; | 919 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; |
| 920 | int ret = 0; | 920 | int ret = 0; |
| 921 | 921 | ||
| 922 | if (btrfs_is_free_space_inode(inode)) { | 922 | if (btrfs_is_free_space_inode(inode)) { |
| 923 | WARN_ON_ONCE(1); | 923 | WARN_ON_ONCE(1); |
| 924 | ret = -EINVAL; | 924 | ret = -EINVAL; |
| 925 | goto out_unlock; | 925 | goto out_unlock; |
| 926 | } | 926 | } |
| 927 | 927 | ||
| 928 | num_bytes = ALIGN(end - start + 1, blocksize); | 928 | num_bytes = ALIGN(end - start + 1, blocksize); |
| 929 | num_bytes = max(blocksize, num_bytes); | 929 | num_bytes = max(blocksize, num_bytes); |
| 930 | disk_num_bytes = num_bytes; | 930 | disk_num_bytes = num_bytes; |
| 931 | 931 | ||
| 932 | /* if this is a small write inside eof, kick off defrag */ | 932 | /* if this is a small write inside eof, kick off defrag */ |
| 933 | if (num_bytes < 64 * 1024 && | 933 | if (num_bytes < 64 * 1024 && |
| 934 | (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) | 934 | (start > 0 || end + 1 < BTRFS_I(inode)->disk_i_size)) |
| 935 | btrfs_add_inode_defrag(NULL, inode); | 935 | btrfs_add_inode_defrag(NULL, inode); |
| 936 | 936 | ||
| 937 | if (start == 0) { | 937 | if (start == 0) { |
| 938 | /* lets try to make an inline extent */ | 938 | /* lets try to make an inline extent */ |
| 939 | ret = cow_file_range_inline(root, inode, start, end, 0, 0, | 939 | ret = cow_file_range_inline(root, inode, start, end, 0, 0, |
| 940 | NULL); | 940 | NULL); |
| 941 | if (ret == 0) { | 941 | if (ret == 0) { |
| 942 | extent_clear_unlock_delalloc(inode, start, end, NULL, | 942 | extent_clear_unlock_delalloc(inode, start, end, NULL, |
| 943 | EXTENT_LOCKED | EXTENT_DELALLOC | | 943 | EXTENT_LOCKED | EXTENT_DELALLOC | |
| 944 | EXTENT_DEFRAG, PAGE_UNLOCK | | 944 | EXTENT_DEFRAG, PAGE_UNLOCK | |
| 945 | PAGE_CLEAR_DIRTY | PAGE_SET_WRITEBACK | | 945 | PAGE_CLEAR_DIRTY | PAGE_SET_WRITEBACK | |
| 946 | PAGE_END_WRITEBACK); | 946 | PAGE_END_WRITEBACK); |
| 947 | 947 | ||
| 948 | *nr_written = *nr_written + | 948 | *nr_written = *nr_written + |
| 949 | (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE; | 949 | (end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE; |
| 950 | *page_started = 1; | 950 | *page_started = 1; |
| 951 | goto out; | 951 | goto out; |
| 952 | } else if (ret < 0) { | 952 | } else if (ret < 0) { |
| 953 | goto out_unlock; | 953 | goto out_unlock; |
| 954 | } | 954 | } |
| 955 | } | 955 | } |
| 956 | 956 | ||
| 957 | BUG_ON(disk_num_bytes > | 957 | BUG_ON(disk_num_bytes > |
| 958 | btrfs_super_total_bytes(root->fs_info->super_copy)); | 958 | btrfs_super_total_bytes(root->fs_info->super_copy)); |
| 959 | 959 | ||
| 960 | alloc_hint = get_extent_allocation_hint(inode, start, num_bytes); | 960 | alloc_hint = get_extent_allocation_hint(inode, start, num_bytes); |
| 961 | btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0); | 961 | btrfs_drop_extent_cache(inode, start, start + num_bytes - 1, 0); |
| 962 | 962 | ||
| 963 | while (disk_num_bytes > 0) { | 963 | while (disk_num_bytes > 0) { |
| 964 | unsigned long op; | 964 | unsigned long op; |
| 965 | 965 | ||
| 966 | cur_alloc_size = disk_num_bytes; | 966 | cur_alloc_size = disk_num_bytes; |
| 967 | ret = btrfs_reserve_extent(root, cur_alloc_size, | 967 | ret = btrfs_reserve_extent(root, cur_alloc_size, |
| 968 | root->sectorsize, 0, alloc_hint, | 968 | root->sectorsize, 0, alloc_hint, |
| 969 | &ins, 1, 1); | 969 | &ins, 1, 1); |
| 970 | if (ret < 0) | 970 | if (ret < 0) |
| 971 | goto out_unlock; | 971 | goto out_unlock; |
| 972 | 972 | ||
| 973 | em = alloc_extent_map(); | 973 | em = alloc_extent_map(); |
| 974 | if (!em) { | 974 | if (!em) { |
| 975 | ret = -ENOMEM; | 975 | ret = -ENOMEM; |
| 976 | goto out_reserve; | 976 | goto out_reserve; |
| 977 | } | 977 | } |
| 978 | em->start = start; | 978 | em->start = start; |
| 979 | em->orig_start = em->start; | 979 | em->orig_start = em->start; |
| 980 | ram_size = ins.offset; | 980 | ram_size = ins.offset; |
| 981 | em->len = ins.offset; | 981 | em->len = ins.offset; |
| 982 | em->mod_start = em->start; | 982 | em->mod_start = em->start; |
| 983 | em->mod_len = em->len; | 983 | em->mod_len = em->len; |
| 984 | 984 | ||
| 985 | em->block_start = ins.objectid; | 985 | em->block_start = ins.objectid; |
| 986 | em->block_len = ins.offset; | 986 | em->block_len = ins.offset; |
| 987 | em->orig_block_len = ins.offset; | 987 | em->orig_block_len = ins.offset; |
| 988 | em->ram_bytes = ram_size; | 988 | em->ram_bytes = ram_size; |
| 989 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 989 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 990 | set_bit(EXTENT_FLAG_PINNED, &em->flags); | 990 | set_bit(EXTENT_FLAG_PINNED, &em->flags); |
| 991 | em->generation = -1; | 991 | em->generation = -1; |
| 992 | 992 | ||
| 993 | while (1) { | 993 | while (1) { |
| 994 | write_lock(&em_tree->lock); | 994 | write_lock(&em_tree->lock); |
| 995 | ret = add_extent_mapping(em_tree, em, 1); | 995 | ret = add_extent_mapping(em_tree, em, 1); |
| 996 | write_unlock(&em_tree->lock); | 996 | write_unlock(&em_tree->lock); |
| 997 | if (ret != -EEXIST) { | 997 | if (ret != -EEXIST) { |
| 998 | free_extent_map(em); | 998 | free_extent_map(em); |
| 999 | break; | 999 | break; |
| 1000 | } | 1000 | } |
| 1001 | btrfs_drop_extent_cache(inode, start, | 1001 | btrfs_drop_extent_cache(inode, start, |
| 1002 | start + ram_size - 1, 0); | 1002 | start + ram_size - 1, 0); |
| 1003 | } | 1003 | } |
| 1004 | if (ret) | 1004 | if (ret) |
| 1005 | goto out_reserve; | 1005 | goto out_reserve; |
| 1006 | 1006 | ||
| 1007 | cur_alloc_size = ins.offset; | 1007 | cur_alloc_size = ins.offset; |
| 1008 | ret = btrfs_add_ordered_extent(inode, start, ins.objectid, | 1008 | ret = btrfs_add_ordered_extent(inode, start, ins.objectid, |
| 1009 | ram_size, cur_alloc_size, 0); | 1009 | ram_size, cur_alloc_size, 0); |
| 1010 | if (ret) | 1010 | if (ret) |
| 1011 | goto out_drop_extent_cache; | 1011 | goto out_drop_extent_cache; |
| 1012 | 1012 | ||
| 1013 | if (root->root_key.objectid == | 1013 | if (root->root_key.objectid == |
| 1014 | BTRFS_DATA_RELOC_TREE_OBJECTID) { | 1014 | BTRFS_DATA_RELOC_TREE_OBJECTID) { |
| 1015 | ret = btrfs_reloc_clone_csums(inode, start, | 1015 | ret = btrfs_reloc_clone_csums(inode, start, |
| 1016 | cur_alloc_size); | 1016 | cur_alloc_size); |
| 1017 | if (ret) | 1017 | if (ret) |
| 1018 | goto out_drop_extent_cache; | 1018 | goto out_drop_extent_cache; |
| 1019 | } | 1019 | } |
| 1020 | 1020 | ||
| 1021 | if (disk_num_bytes < cur_alloc_size) | 1021 | if (disk_num_bytes < cur_alloc_size) |
| 1022 | break; | 1022 | break; |
| 1023 | 1023 | ||
| 1024 | /* we're not doing compressed IO, don't unlock the first | 1024 | /* we're not doing compressed IO, don't unlock the first |
| 1025 | * page (which the caller expects to stay locked), don't | 1025 | * page (which the caller expects to stay locked), don't |
| 1026 | * clear any dirty bits and don't set any writeback bits | 1026 | * clear any dirty bits and don't set any writeback bits |
| 1027 | * | 1027 | * |
| 1028 | * Do set the Private2 bit so we know this page was properly | 1028 | * Do set the Private2 bit so we know this page was properly |
| 1029 | * setup for writepage | 1029 | * setup for writepage |
| 1030 | */ | 1030 | */ |
| 1031 | op = unlock ? PAGE_UNLOCK : 0; | 1031 | op = unlock ? PAGE_UNLOCK : 0; |
| 1032 | op |= PAGE_SET_PRIVATE2; | 1032 | op |= PAGE_SET_PRIVATE2; |
| 1033 | 1033 | ||
| 1034 | extent_clear_unlock_delalloc(inode, start, | 1034 | extent_clear_unlock_delalloc(inode, start, |
| 1035 | start + ram_size - 1, locked_page, | 1035 | start + ram_size - 1, locked_page, |
| 1036 | EXTENT_LOCKED | EXTENT_DELALLOC, | 1036 | EXTENT_LOCKED | EXTENT_DELALLOC, |
| 1037 | op); | 1037 | op); |
| 1038 | disk_num_bytes -= cur_alloc_size; | 1038 | disk_num_bytes -= cur_alloc_size; |
| 1039 | num_bytes -= cur_alloc_size; | 1039 | num_bytes -= cur_alloc_size; |
| 1040 | alloc_hint = ins.objectid + ins.offset; | 1040 | alloc_hint = ins.objectid + ins.offset; |
| 1041 | start += cur_alloc_size; | 1041 | start += cur_alloc_size; |
| 1042 | } | 1042 | } |
| 1043 | out: | 1043 | out: |
| 1044 | return ret; | 1044 | return ret; |
| 1045 | 1045 | ||
| 1046 | out_drop_extent_cache: | 1046 | out_drop_extent_cache: |
| 1047 | btrfs_drop_extent_cache(inode, start, start + ram_size - 1, 0); | 1047 | btrfs_drop_extent_cache(inode, start, start + ram_size - 1, 0); |
| 1048 | out_reserve: | 1048 | out_reserve: |
| 1049 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); | 1049 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); |
| 1050 | out_unlock: | 1050 | out_unlock: |
| 1051 | extent_clear_unlock_delalloc(inode, start, end, locked_page, | 1051 | extent_clear_unlock_delalloc(inode, start, end, locked_page, |
| 1052 | EXTENT_LOCKED | EXTENT_DO_ACCOUNTING | | 1052 | EXTENT_LOCKED | EXTENT_DO_ACCOUNTING | |
| 1053 | EXTENT_DELALLOC | EXTENT_DEFRAG, | 1053 | EXTENT_DELALLOC | EXTENT_DEFRAG, |
| 1054 | PAGE_UNLOCK | PAGE_CLEAR_DIRTY | | 1054 | PAGE_UNLOCK | PAGE_CLEAR_DIRTY | |
| 1055 | PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK); | 1055 | PAGE_SET_WRITEBACK | PAGE_END_WRITEBACK); |
| 1056 | goto out; | 1056 | goto out; |
| 1057 | } | 1057 | } |
| 1058 | 1058 | ||
| 1059 | /* | 1059 | /* |
| 1060 | * work queue call back to started compression on a file and pages | 1060 | * work queue call back to started compression on a file and pages |
| 1061 | */ | 1061 | */ |
| 1062 | static noinline void async_cow_start(struct btrfs_work *work) | 1062 | static noinline void async_cow_start(struct btrfs_work *work) |
| 1063 | { | 1063 | { |
| 1064 | struct async_cow *async_cow; | 1064 | struct async_cow *async_cow; |
| 1065 | int num_added = 0; | 1065 | int num_added = 0; |
| 1066 | async_cow = container_of(work, struct async_cow, work); | 1066 | async_cow = container_of(work, struct async_cow, work); |
| 1067 | 1067 | ||
| 1068 | compress_file_range(async_cow->inode, async_cow->locked_page, | 1068 | compress_file_range(async_cow->inode, async_cow->locked_page, |
| 1069 | async_cow->start, async_cow->end, async_cow, | 1069 | async_cow->start, async_cow->end, async_cow, |
| 1070 | &num_added); | 1070 | &num_added); |
| 1071 | if (num_added == 0) { | 1071 | if (num_added == 0) { |
| 1072 | btrfs_add_delayed_iput(async_cow->inode); | 1072 | btrfs_add_delayed_iput(async_cow->inode); |
| 1073 | async_cow->inode = NULL; | 1073 | async_cow->inode = NULL; |
| 1074 | } | 1074 | } |
| 1075 | } | 1075 | } |
| 1076 | 1076 | ||
| 1077 | /* | 1077 | /* |
| 1078 | * work queue call back to submit previously compressed pages | 1078 | * work queue call back to submit previously compressed pages |
| 1079 | */ | 1079 | */ |
| 1080 | static noinline void async_cow_submit(struct btrfs_work *work) | 1080 | static noinline void async_cow_submit(struct btrfs_work *work) |
| 1081 | { | 1081 | { |
| 1082 | struct async_cow *async_cow; | 1082 | struct async_cow *async_cow; |
| 1083 | struct btrfs_root *root; | 1083 | struct btrfs_root *root; |
| 1084 | unsigned long nr_pages; | 1084 | unsigned long nr_pages; |
| 1085 | 1085 | ||
| 1086 | async_cow = container_of(work, struct async_cow, work); | 1086 | async_cow = container_of(work, struct async_cow, work); |
| 1087 | 1087 | ||
| 1088 | root = async_cow->root; | 1088 | root = async_cow->root; |
| 1089 | nr_pages = (async_cow->end - async_cow->start + PAGE_CACHE_SIZE) >> | 1089 | nr_pages = (async_cow->end - async_cow->start + PAGE_CACHE_SIZE) >> |
| 1090 | PAGE_CACHE_SHIFT; | 1090 | PAGE_CACHE_SHIFT; |
| 1091 | 1091 | ||
| 1092 | if (atomic_sub_return(nr_pages, &root->fs_info->async_delalloc_pages) < | 1092 | if (atomic_sub_return(nr_pages, &root->fs_info->async_delalloc_pages) < |
| 1093 | 5 * 1024 * 1024 && | 1093 | 5 * 1024 * 1024 && |
| 1094 | waitqueue_active(&root->fs_info->async_submit_wait)) | 1094 | waitqueue_active(&root->fs_info->async_submit_wait)) |
| 1095 | wake_up(&root->fs_info->async_submit_wait); | 1095 | wake_up(&root->fs_info->async_submit_wait); |
| 1096 | 1096 | ||
| 1097 | if (async_cow->inode) | 1097 | if (async_cow->inode) |
| 1098 | submit_compressed_extents(async_cow->inode, async_cow); | 1098 | submit_compressed_extents(async_cow->inode, async_cow); |
| 1099 | } | 1099 | } |
| 1100 | 1100 | ||
| 1101 | static noinline void async_cow_free(struct btrfs_work *work) | 1101 | static noinline void async_cow_free(struct btrfs_work *work) |
| 1102 | { | 1102 | { |
| 1103 | struct async_cow *async_cow; | 1103 | struct async_cow *async_cow; |
| 1104 | async_cow = container_of(work, struct async_cow, work); | 1104 | async_cow = container_of(work, struct async_cow, work); |
| 1105 | if (async_cow->inode) | 1105 | if (async_cow->inode) |
| 1106 | btrfs_add_delayed_iput(async_cow->inode); | 1106 | btrfs_add_delayed_iput(async_cow->inode); |
| 1107 | kfree(async_cow); | 1107 | kfree(async_cow); |
| 1108 | } | 1108 | } |
| 1109 | 1109 | ||
| 1110 | static int cow_file_range_async(struct inode *inode, struct page *locked_page, | 1110 | static int cow_file_range_async(struct inode *inode, struct page *locked_page, |
| 1111 | u64 start, u64 end, int *page_started, | 1111 | u64 start, u64 end, int *page_started, |
| 1112 | unsigned long *nr_written) | 1112 | unsigned long *nr_written) |
| 1113 | { | 1113 | { |
| 1114 | struct async_cow *async_cow; | 1114 | struct async_cow *async_cow; |
| 1115 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1115 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1116 | unsigned long nr_pages; | 1116 | unsigned long nr_pages; |
| 1117 | u64 cur_end; | 1117 | u64 cur_end; |
| 1118 | int limit = 10 * 1024 * 1024; | 1118 | int limit = 10 * 1024 * 1024; |
| 1119 | 1119 | ||
| 1120 | clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED, | 1120 | clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED, |
| 1121 | 1, 0, NULL, GFP_NOFS); | 1121 | 1, 0, NULL, GFP_NOFS); |
| 1122 | while (start < end) { | 1122 | while (start < end) { |
| 1123 | async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS); | 1123 | async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS); |
| 1124 | BUG_ON(!async_cow); /* -ENOMEM */ | 1124 | BUG_ON(!async_cow); /* -ENOMEM */ |
| 1125 | async_cow->inode = igrab(inode); | 1125 | async_cow->inode = igrab(inode); |
| 1126 | async_cow->root = root; | 1126 | async_cow->root = root; |
| 1127 | async_cow->locked_page = locked_page; | 1127 | async_cow->locked_page = locked_page; |
| 1128 | async_cow->start = start; | 1128 | async_cow->start = start; |
| 1129 | 1129 | ||
| 1130 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS && | 1130 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS && |
| 1131 | !btrfs_test_opt(root, FORCE_COMPRESS)) | 1131 | !btrfs_test_opt(root, FORCE_COMPRESS)) |
| 1132 | cur_end = end; | 1132 | cur_end = end; |
| 1133 | else | 1133 | else |
| 1134 | cur_end = min(end, start + 512 * 1024 - 1); | 1134 | cur_end = min(end, start + 512 * 1024 - 1); |
| 1135 | 1135 | ||
| 1136 | async_cow->end = cur_end; | 1136 | async_cow->end = cur_end; |
| 1137 | INIT_LIST_HEAD(&async_cow->extents); | 1137 | INIT_LIST_HEAD(&async_cow->extents); |
| 1138 | 1138 | ||
| 1139 | btrfs_init_work(&async_cow->work, | 1139 | btrfs_init_work(&async_cow->work, |
| 1140 | btrfs_delalloc_helper, | 1140 | btrfs_delalloc_helper, |
| 1141 | async_cow_start, async_cow_submit, | 1141 | async_cow_start, async_cow_submit, |
| 1142 | async_cow_free); | 1142 | async_cow_free); |
| 1143 | 1143 | ||
| 1144 | nr_pages = (cur_end - start + PAGE_CACHE_SIZE) >> | 1144 | nr_pages = (cur_end - start + PAGE_CACHE_SIZE) >> |
| 1145 | PAGE_CACHE_SHIFT; | 1145 | PAGE_CACHE_SHIFT; |
| 1146 | atomic_add(nr_pages, &root->fs_info->async_delalloc_pages); | 1146 | atomic_add(nr_pages, &root->fs_info->async_delalloc_pages); |
| 1147 | 1147 | ||
| 1148 | btrfs_queue_work(root->fs_info->delalloc_workers, | 1148 | btrfs_queue_work(root->fs_info->delalloc_workers, |
| 1149 | &async_cow->work); | 1149 | &async_cow->work); |
| 1150 | 1150 | ||
| 1151 | if (atomic_read(&root->fs_info->async_delalloc_pages) > limit) { | 1151 | if (atomic_read(&root->fs_info->async_delalloc_pages) > limit) { |
| 1152 | wait_event(root->fs_info->async_submit_wait, | 1152 | wait_event(root->fs_info->async_submit_wait, |
| 1153 | (atomic_read(&root->fs_info->async_delalloc_pages) < | 1153 | (atomic_read(&root->fs_info->async_delalloc_pages) < |
| 1154 | limit)); | 1154 | limit)); |
| 1155 | } | 1155 | } |
| 1156 | 1156 | ||
| 1157 | while (atomic_read(&root->fs_info->async_submit_draining) && | 1157 | while (atomic_read(&root->fs_info->async_submit_draining) && |
| 1158 | atomic_read(&root->fs_info->async_delalloc_pages)) { | 1158 | atomic_read(&root->fs_info->async_delalloc_pages)) { |
| 1159 | wait_event(root->fs_info->async_submit_wait, | 1159 | wait_event(root->fs_info->async_submit_wait, |
| 1160 | (atomic_read(&root->fs_info->async_delalloc_pages) == | 1160 | (atomic_read(&root->fs_info->async_delalloc_pages) == |
| 1161 | 0)); | 1161 | 0)); |
| 1162 | } | 1162 | } |
| 1163 | 1163 | ||
| 1164 | *nr_written += nr_pages; | 1164 | *nr_written += nr_pages; |
| 1165 | start = cur_end + 1; | 1165 | start = cur_end + 1; |
| 1166 | } | 1166 | } |
| 1167 | *page_started = 1; | 1167 | *page_started = 1; |
| 1168 | return 0; | 1168 | return 0; |
| 1169 | } | 1169 | } |
| 1170 | 1170 | ||
| 1171 | static noinline int csum_exist_in_range(struct btrfs_root *root, | 1171 | static noinline int csum_exist_in_range(struct btrfs_root *root, |
| 1172 | u64 bytenr, u64 num_bytes) | 1172 | u64 bytenr, u64 num_bytes) |
| 1173 | { | 1173 | { |
| 1174 | int ret; | 1174 | int ret; |
| 1175 | struct btrfs_ordered_sum *sums; | 1175 | struct btrfs_ordered_sum *sums; |
| 1176 | LIST_HEAD(list); | 1176 | LIST_HEAD(list); |
| 1177 | 1177 | ||
| 1178 | ret = btrfs_lookup_csums_range(root->fs_info->csum_root, bytenr, | 1178 | ret = btrfs_lookup_csums_range(root->fs_info->csum_root, bytenr, |
| 1179 | bytenr + num_bytes - 1, &list, 0); | 1179 | bytenr + num_bytes - 1, &list, 0); |
| 1180 | if (ret == 0 && list_empty(&list)) | 1180 | if (ret == 0 && list_empty(&list)) |
| 1181 | return 0; | 1181 | return 0; |
| 1182 | 1182 | ||
| 1183 | while (!list_empty(&list)) { | 1183 | while (!list_empty(&list)) { |
| 1184 | sums = list_entry(list.next, struct btrfs_ordered_sum, list); | 1184 | sums = list_entry(list.next, struct btrfs_ordered_sum, list); |
| 1185 | list_del(&sums->list); | 1185 | list_del(&sums->list); |
| 1186 | kfree(sums); | 1186 | kfree(sums); |
| 1187 | } | 1187 | } |
| 1188 | return 1; | 1188 | return 1; |
| 1189 | } | 1189 | } |
| 1190 | 1190 | ||
| 1191 | /* | 1191 | /* |
| 1192 | * when nowcow writeback call back. This checks for snapshots or COW copies | 1192 | * when nowcow writeback call back. This checks for snapshots or COW copies |
| 1193 | * of the extents that exist in the file, and COWs the file as required. | 1193 | * of the extents that exist in the file, and COWs the file as required. |
| 1194 | * | 1194 | * |
| 1195 | * If no cow copies or snapshots exist, we write directly to the existing | 1195 | * If no cow copies or snapshots exist, we write directly to the existing |
| 1196 | * blocks on disk | 1196 | * blocks on disk |
| 1197 | */ | 1197 | */ |
| 1198 | static noinline int run_delalloc_nocow(struct inode *inode, | 1198 | static noinline int run_delalloc_nocow(struct inode *inode, |
| 1199 | struct page *locked_page, | 1199 | struct page *locked_page, |
| 1200 | u64 start, u64 end, int *page_started, int force, | 1200 | u64 start, u64 end, int *page_started, int force, |
| 1201 | unsigned long *nr_written) | 1201 | unsigned long *nr_written) |
| 1202 | { | 1202 | { |
| 1203 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1203 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1204 | struct btrfs_trans_handle *trans; | 1204 | struct btrfs_trans_handle *trans; |
| 1205 | struct extent_buffer *leaf; | 1205 | struct extent_buffer *leaf; |
| 1206 | struct btrfs_path *path; | 1206 | struct btrfs_path *path; |
| 1207 | struct btrfs_file_extent_item *fi; | 1207 | struct btrfs_file_extent_item *fi; |
| 1208 | struct btrfs_key found_key; | 1208 | struct btrfs_key found_key; |
| 1209 | u64 cow_start; | 1209 | u64 cow_start; |
| 1210 | u64 cur_offset; | 1210 | u64 cur_offset; |
| 1211 | u64 extent_end; | 1211 | u64 extent_end; |
| 1212 | u64 extent_offset; | 1212 | u64 extent_offset; |
| 1213 | u64 disk_bytenr; | 1213 | u64 disk_bytenr; |
| 1214 | u64 num_bytes; | 1214 | u64 num_bytes; |
| 1215 | u64 disk_num_bytes; | 1215 | u64 disk_num_bytes; |
| 1216 | u64 ram_bytes; | 1216 | u64 ram_bytes; |
| 1217 | int extent_type; | 1217 | int extent_type; |
| 1218 | int ret, err; | 1218 | int ret, err; |
| 1219 | int type; | 1219 | int type; |
| 1220 | int nocow; | 1220 | int nocow; |
| 1221 | int check_prev = 1; | 1221 | int check_prev = 1; |
| 1222 | bool nolock; | 1222 | bool nolock; |
| 1223 | u64 ino = btrfs_ino(inode); | 1223 | u64 ino = btrfs_ino(inode); |
| 1224 | 1224 | ||
| 1225 | path = btrfs_alloc_path(); | 1225 | path = btrfs_alloc_path(); |
| 1226 | if (!path) { | 1226 | if (!path) { |
| 1227 | extent_clear_unlock_delalloc(inode, start, end, locked_page, | 1227 | extent_clear_unlock_delalloc(inode, start, end, locked_page, |
| 1228 | EXTENT_LOCKED | EXTENT_DELALLOC | | 1228 | EXTENT_LOCKED | EXTENT_DELALLOC | |
| 1229 | EXTENT_DO_ACCOUNTING | | 1229 | EXTENT_DO_ACCOUNTING | |
| 1230 | EXTENT_DEFRAG, PAGE_UNLOCK | | 1230 | EXTENT_DEFRAG, PAGE_UNLOCK | |
| 1231 | PAGE_CLEAR_DIRTY | | 1231 | PAGE_CLEAR_DIRTY | |
| 1232 | PAGE_SET_WRITEBACK | | 1232 | PAGE_SET_WRITEBACK | |
| 1233 | PAGE_END_WRITEBACK); | 1233 | PAGE_END_WRITEBACK); |
| 1234 | return -ENOMEM; | 1234 | return -ENOMEM; |
| 1235 | } | 1235 | } |
| 1236 | 1236 | ||
| 1237 | nolock = btrfs_is_free_space_inode(inode); | 1237 | nolock = btrfs_is_free_space_inode(inode); |
| 1238 | 1238 | ||
| 1239 | if (nolock) | 1239 | if (nolock) |
| 1240 | trans = btrfs_join_transaction_nolock(root); | 1240 | trans = btrfs_join_transaction_nolock(root); |
| 1241 | else | 1241 | else |
| 1242 | trans = btrfs_join_transaction(root); | 1242 | trans = btrfs_join_transaction(root); |
| 1243 | 1243 | ||
| 1244 | if (IS_ERR(trans)) { | 1244 | if (IS_ERR(trans)) { |
| 1245 | extent_clear_unlock_delalloc(inode, start, end, locked_page, | 1245 | extent_clear_unlock_delalloc(inode, start, end, locked_page, |
| 1246 | EXTENT_LOCKED | EXTENT_DELALLOC | | 1246 | EXTENT_LOCKED | EXTENT_DELALLOC | |
| 1247 | EXTENT_DO_ACCOUNTING | | 1247 | EXTENT_DO_ACCOUNTING | |
| 1248 | EXTENT_DEFRAG, PAGE_UNLOCK | | 1248 | EXTENT_DEFRAG, PAGE_UNLOCK | |
| 1249 | PAGE_CLEAR_DIRTY | | 1249 | PAGE_CLEAR_DIRTY | |
| 1250 | PAGE_SET_WRITEBACK | | 1250 | PAGE_SET_WRITEBACK | |
| 1251 | PAGE_END_WRITEBACK); | 1251 | PAGE_END_WRITEBACK); |
| 1252 | btrfs_free_path(path); | 1252 | btrfs_free_path(path); |
| 1253 | return PTR_ERR(trans); | 1253 | return PTR_ERR(trans); |
| 1254 | } | 1254 | } |
| 1255 | 1255 | ||
| 1256 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; | 1256 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; |
| 1257 | 1257 | ||
| 1258 | cow_start = (u64)-1; | 1258 | cow_start = (u64)-1; |
| 1259 | cur_offset = start; | 1259 | cur_offset = start; |
| 1260 | while (1) { | 1260 | while (1) { |
| 1261 | ret = btrfs_lookup_file_extent(trans, root, path, ino, | 1261 | ret = btrfs_lookup_file_extent(trans, root, path, ino, |
| 1262 | cur_offset, 0); | 1262 | cur_offset, 0); |
| 1263 | if (ret < 0) | 1263 | if (ret < 0) |
| 1264 | goto error; | 1264 | goto error; |
| 1265 | if (ret > 0 && path->slots[0] > 0 && check_prev) { | 1265 | if (ret > 0 && path->slots[0] > 0 && check_prev) { |
| 1266 | leaf = path->nodes[0]; | 1266 | leaf = path->nodes[0]; |
| 1267 | btrfs_item_key_to_cpu(leaf, &found_key, | 1267 | btrfs_item_key_to_cpu(leaf, &found_key, |
| 1268 | path->slots[0] - 1); | 1268 | path->slots[0] - 1); |
| 1269 | if (found_key.objectid == ino && | 1269 | if (found_key.objectid == ino && |
| 1270 | found_key.type == BTRFS_EXTENT_DATA_KEY) | 1270 | found_key.type == BTRFS_EXTENT_DATA_KEY) |
| 1271 | path->slots[0]--; | 1271 | path->slots[0]--; |
| 1272 | } | 1272 | } |
| 1273 | check_prev = 0; | 1273 | check_prev = 0; |
| 1274 | next_slot: | 1274 | next_slot: |
| 1275 | leaf = path->nodes[0]; | 1275 | leaf = path->nodes[0]; |
| 1276 | if (path->slots[0] >= btrfs_header_nritems(leaf)) { | 1276 | if (path->slots[0] >= btrfs_header_nritems(leaf)) { |
| 1277 | ret = btrfs_next_leaf(root, path); | 1277 | ret = btrfs_next_leaf(root, path); |
| 1278 | if (ret < 0) | 1278 | if (ret < 0) |
| 1279 | goto error; | 1279 | goto error; |
| 1280 | if (ret > 0) | 1280 | if (ret > 0) |
| 1281 | break; | 1281 | break; |
| 1282 | leaf = path->nodes[0]; | 1282 | leaf = path->nodes[0]; |
| 1283 | } | 1283 | } |
| 1284 | 1284 | ||
| 1285 | nocow = 0; | 1285 | nocow = 0; |
| 1286 | disk_bytenr = 0; | 1286 | disk_bytenr = 0; |
| 1287 | num_bytes = 0; | 1287 | num_bytes = 0; |
| 1288 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 1288 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 1289 | 1289 | ||
| 1290 | if (found_key.objectid > ino || | 1290 | if (found_key.objectid > ino || |
| 1291 | found_key.type > BTRFS_EXTENT_DATA_KEY || | 1291 | found_key.type > BTRFS_EXTENT_DATA_KEY || |
| 1292 | found_key.offset > end) | 1292 | found_key.offset > end) |
| 1293 | break; | 1293 | break; |
| 1294 | 1294 | ||
| 1295 | if (found_key.offset > cur_offset) { | 1295 | if (found_key.offset > cur_offset) { |
| 1296 | extent_end = found_key.offset; | 1296 | extent_end = found_key.offset; |
| 1297 | extent_type = 0; | 1297 | extent_type = 0; |
| 1298 | goto out_check; | 1298 | goto out_check; |
| 1299 | } | 1299 | } |
| 1300 | 1300 | ||
| 1301 | fi = btrfs_item_ptr(leaf, path->slots[0], | 1301 | fi = btrfs_item_ptr(leaf, path->slots[0], |
| 1302 | struct btrfs_file_extent_item); | 1302 | struct btrfs_file_extent_item); |
| 1303 | extent_type = btrfs_file_extent_type(leaf, fi); | 1303 | extent_type = btrfs_file_extent_type(leaf, fi); |
| 1304 | 1304 | ||
| 1305 | ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi); | 1305 | ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi); |
| 1306 | if (extent_type == BTRFS_FILE_EXTENT_REG || | 1306 | if (extent_type == BTRFS_FILE_EXTENT_REG || |
| 1307 | extent_type == BTRFS_FILE_EXTENT_PREALLOC) { | 1307 | extent_type == BTRFS_FILE_EXTENT_PREALLOC) { |
| 1308 | disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi); | 1308 | disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi); |
| 1309 | extent_offset = btrfs_file_extent_offset(leaf, fi); | 1309 | extent_offset = btrfs_file_extent_offset(leaf, fi); |
| 1310 | extent_end = found_key.offset + | 1310 | extent_end = found_key.offset + |
| 1311 | btrfs_file_extent_num_bytes(leaf, fi); | 1311 | btrfs_file_extent_num_bytes(leaf, fi); |
| 1312 | disk_num_bytes = | 1312 | disk_num_bytes = |
| 1313 | btrfs_file_extent_disk_num_bytes(leaf, fi); | 1313 | btrfs_file_extent_disk_num_bytes(leaf, fi); |
| 1314 | if (extent_end <= start) { | 1314 | if (extent_end <= start) { |
| 1315 | path->slots[0]++; | 1315 | path->slots[0]++; |
| 1316 | goto next_slot; | 1316 | goto next_slot; |
| 1317 | } | 1317 | } |
| 1318 | if (disk_bytenr == 0) | 1318 | if (disk_bytenr == 0) |
| 1319 | goto out_check; | 1319 | goto out_check; |
| 1320 | if (btrfs_file_extent_compression(leaf, fi) || | 1320 | if (btrfs_file_extent_compression(leaf, fi) || |
| 1321 | btrfs_file_extent_encryption(leaf, fi) || | 1321 | btrfs_file_extent_encryption(leaf, fi) || |
| 1322 | btrfs_file_extent_other_encoding(leaf, fi)) | 1322 | btrfs_file_extent_other_encoding(leaf, fi)) |
| 1323 | goto out_check; | 1323 | goto out_check; |
| 1324 | if (extent_type == BTRFS_FILE_EXTENT_REG && !force) | 1324 | if (extent_type == BTRFS_FILE_EXTENT_REG && !force) |
| 1325 | goto out_check; | 1325 | goto out_check; |
| 1326 | if (btrfs_extent_readonly(root, disk_bytenr)) | 1326 | if (btrfs_extent_readonly(root, disk_bytenr)) |
| 1327 | goto out_check; | 1327 | goto out_check; |
| 1328 | if (btrfs_cross_ref_exist(trans, root, ino, | 1328 | if (btrfs_cross_ref_exist(trans, root, ino, |
| 1329 | found_key.offset - | 1329 | found_key.offset - |
| 1330 | extent_offset, disk_bytenr)) | 1330 | extent_offset, disk_bytenr)) |
| 1331 | goto out_check; | 1331 | goto out_check; |
| 1332 | disk_bytenr += extent_offset; | 1332 | disk_bytenr += extent_offset; |
| 1333 | disk_bytenr += cur_offset - found_key.offset; | 1333 | disk_bytenr += cur_offset - found_key.offset; |
| 1334 | num_bytes = min(end + 1, extent_end) - cur_offset; | 1334 | num_bytes = min(end + 1, extent_end) - cur_offset; |
| 1335 | /* | 1335 | /* |
| 1336 | * if there are pending snapshots for this root, | 1336 | * if there are pending snapshots for this root, |
| 1337 | * we fall into common COW way. | 1337 | * we fall into common COW way. |
| 1338 | */ | 1338 | */ |
| 1339 | if (!nolock) { | 1339 | if (!nolock) { |
| 1340 | err = btrfs_start_write_no_snapshoting(root); | 1340 | err = btrfs_start_write_no_snapshoting(root); |
| 1341 | if (!err) | 1341 | if (!err) |
| 1342 | goto out_check; | 1342 | goto out_check; |
| 1343 | } | 1343 | } |
| 1344 | /* | 1344 | /* |
| 1345 | * force cow if csum exists in the range. | 1345 | * force cow if csum exists in the range. |
| 1346 | * this ensure that csum for a given extent are | 1346 | * this ensure that csum for a given extent are |
| 1347 | * either valid or do not exist. | 1347 | * either valid or do not exist. |
| 1348 | */ | 1348 | */ |
| 1349 | if (csum_exist_in_range(root, disk_bytenr, num_bytes)) | 1349 | if (csum_exist_in_range(root, disk_bytenr, num_bytes)) |
| 1350 | goto out_check; | 1350 | goto out_check; |
| 1351 | nocow = 1; | 1351 | nocow = 1; |
| 1352 | } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { | 1352 | } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { |
| 1353 | extent_end = found_key.offset + | 1353 | extent_end = found_key.offset + |
| 1354 | btrfs_file_extent_inline_len(leaf, | 1354 | btrfs_file_extent_inline_len(leaf, |
| 1355 | path->slots[0], fi); | 1355 | path->slots[0], fi); |
| 1356 | extent_end = ALIGN(extent_end, root->sectorsize); | 1356 | extent_end = ALIGN(extent_end, root->sectorsize); |
| 1357 | } else { | 1357 | } else { |
| 1358 | BUG_ON(1); | 1358 | BUG_ON(1); |
| 1359 | } | 1359 | } |
| 1360 | out_check: | 1360 | out_check: |
| 1361 | if (extent_end <= start) { | 1361 | if (extent_end <= start) { |
| 1362 | path->slots[0]++; | 1362 | path->slots[0]++; |
| 1363 | if (!nolock && nocow) | 1363 | if (!nolock && nocow) |
| 1364 | btrfs_end_write_no_snapshoting(root); | 1364 | btrfs_end_write_no_snapshoting(root); |
| 1365 | goto next_slot; | 1365 | goto next_slot; |
| 1366 | } | 1366 | } |
| 1367 | if (!nocow) { | 1367 | if (!nocow) { |
| 1368 | if (cow_start == (u64)-1) | 1368 | if (cow_start == (u64)-1) |
| 1369 | cow_start = cur_offset; | 1369 | cow_start = cur_offset; |
| 1370 | cur_offset = extent_end; | 1370 | cur_offset = extent_end; |
| 1371 | if (cur_offset > end) | 1371 | if (cur_offset > end) |
| 1372 | break; | 1372 | break; |
| 1373 | path->slots[0]++; | 1373 | path->slots[0]++; |
| 1374 | goto next_slot; | 1374 | goto next_slot; |
| 1375 | } | 1375 | } |
| 1376 | 1376 | ||
| 1377 | btrfs_release_path(path); | 1377 | btrfs_release_path(path); |
| 1378 | if (cow_start != (u64)-1) { | 1378 | if (cow_start != (u64)-1) { |
| 1379 | ret = cow_file_range(inode, locked_page, | 1379 | ret = cow_file_range(inode, locked_page, |
| 1380 | cow_start, found_key.offset - 1, | 1380 | cow_start, found_key.offset - 1, |
| 1381 | page_started, nr_written, 1); | 1381 | page_started, nr_written, 1); |
| 1382 | if (ret) { | 1382 | if (ret) { |
| 1383 | if (!nolock && nocow) | 1383 | if (!nolock && nocow) |
| 1384 | btrfs_end_write_no_snapshoting(root); | 1384 | btrfs_end_write_no_snapshoting(root); |
| 1385 | goto error; | 1385 | goto error; |
| 1386 | } | 1386 | } |
| 1387 | cow_start = (u64)-1; | 1387 | cow_start = (u64)-1; |
| 1388 | } | 1388 | } |
| 1389 | 1389 | ||
| 1390 | if (extent_type == BTRFS_FILE_EXTENT_PREALLOC) { | 1390 | if (extent_type == BTRFS_FILE_EXTENT_PREALLOC) { |
| 1391 | struct extent_map *em; | 1391 | struct extent_map *em; |
| 1392 | struct extent_map_tree *em_tree; | 1392 | struct extent_map_tree *em_tree; |
| 1393 | em_tree = &BTRFS_I(inode)->extent_tree; | 1393 | em_tree = &BTRFS_I(inode)->extent_tree; |
| 1394 | em = alloc_extent_map(); | 1394 | em = alloc_extent_map(); |
| 1395 | BUG_ON(!em); /* -ENOMEM */ | 1395 | BUG_ON(!em); /* -ENOMEM */ |
| 1396 | em->start = cur_offset; | 1396 | em->start = cur_offset; |
| 1397 | em->orig_start = found_key.offset - extent_offset; | 1397 | em->orig_start = found_key.offset - extent_offset; |
| 1398 | em->len = num_bytes; | 1398 | em->len = num_bytes; |
| 1399 | em->block_len = num_bytes; | 1399 | em->block_len = num_bytes; |
| 1400 | em->block_start = disk_bytenr; | 1400 | em->block_start = disk_bytenr; |
| 1401 | em->orig_block_len = disk_num_bytes; | 1401 | em->orig_block_len = disk_num_bytes; |
| 1402 | em->ram_bytes = ram_bytes; | 1402 | em->ram_bytes = ram_bytes; |
| 1403 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 1403 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 1404 | em->mod_start = em->start; | 1404 | em->mod_start = em->start; |
| 1405 | em->mod_len = em->len; | 1405 | em->mod_len = em->len; |
| 1406 | set_bit(EXTENT_FLAG_PINNED, &em->flags); | 1406 | set_bit(EXTENT_FLAG_PINNED, &em->flags); |
| 1407 | set_bit(EXTENT_FLAG_FILLING, &em->flags); | 1407 | set_bit(EXTENT_FLAG_FILLING, &em->flags); |
| 1408 | em->generation = -1; | 1408 | em->generation = -1; |
| 1409 | while (1) { | 1409 | while (1) { |
| 1410 | write_lock(&em_tree->lock); | 1410 | write_lock(&em_tree->lock); |
| 1411 | ret = add_extent_mapping(em_tree, em, 1); | 1411 | ret = add_extent_mapping(em_tree, em, 1); |
| 1412 | write_unlock(&em_tree->lock); | 1412 | write_unlock(&em_tree->lock); |
| 1413 | if (ret != -EEXIST) { | 1413 | if (ret != -EEXIST) { |
| 1414 | free_extent_map(em); | 1414 | free_extent_map(em); |
| 1415 | break; | 1415 | break; |
| 1416 | } | 1416 | } |
| 1417 | btrfs_drop_extent_cache(inode, em->start, | 1417 | btrfs_drop_extent_cache(inode, em->start, |
| 1418 | em->start + em->len - 1, 0); | 1418 | em->start + em->len - 1, 0); |
| 1419 | } | 1419 | } |
| 1420 | type = BTRFS_ORDERED_PREALLOC; | 1420 | type = BTRFS_ORDERED_PREALLOC; |
| 1421 | } else { | 1421 | } else { |
| 1422 | type = BTRFS_ORDERED_NOCOW; | 1422 | type = BTRFS_ORDERED_NOCOW; |
| 1423 | } | 1423 | } |
| 1424 | 1424 | ||
| 1425 | ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr, | 1425 | ret = btrfs_add_ordered_extent(inode, cur_offset, disk_bytenr, |
| 1426 | num_bytes, num_bytes, type); | 1426 | num_bytes, num_bytes, type); |
| 1427 | BUG_ON(ret); /* -ENOMEM */ | 1427 | BUG_ON(ret); /* -ENOMEM */ |
| 1428 | 1428 | ||
| 1429 | if (root->root_key.objectid == | 1429 | if (root->root_key.objectid == |
| 1430 | BTRFS_DATA_RELOC_TREE_OBJECTID) { | 1430 | BTRFS_DATA_RELOC_TREE_OBJECTID) { |
| 1431 | ret = btrfs_reloc_clone_csums(inode, cur_offset, | 1431 | ret = btrfs_reloc_clone_csums(inode, cur_offset, |
| 1432 | num_bytes); | 1432 | num_bytes); |
| 1433 | if (ret) { | 1433 | if (ret) { |
| 1434 | if (!nolock && nocow) | 1434 | if (!nolock && nocow) |
| 1435 | btrfs_end_write_no_snapshoting(root); | 1435 | btrfs_end_write_no_snapshoting(root); |
| 1436 | goto error; | 1436 | goto error; |
| 1437 | } | 1437 | } |
| 1438 | } | 1438 | } |
| 1439 | 1439 | ||
| 1440 | extent_clear_unlock_delalloc(inode, cur_offset, | 1440 | extent_clear_unlock_delalloc(inode, cur_offset, |
| 1441 | cur_offset + num_bytes - 1, | 1441 | cur_offset + num_bytes - 1, |
| 1442 | locked_page, EXTENT_LOCKED | | 1442 | locked_page, EXTENT_LOCKED | |
| 1443 | EXTENT_DELALLOC, PAGE_UNLOCK | | 1443 | EXTENT_DELALLOC, PAGE_UNLOCK | |
| 1444 | PAGE_SET_PRIVATE2); | 1444 | PAGE_SET_PRIVATE2); |
| 1445 | if (!nolock && nocow) | 1445 | if (!nolock && nocow) |
| 1446 | btrfs_end_write_no_snapshoting(root); | 1446 | btrfs_end_write_no_snapshoting(root); |
| 1447 | cur_offset = extent_end; | 1447 | cur_offset = extent_end; |
| 1448 | if (cur_offset > end) | 1448 | if (cur_offset > end) |
| 1449 | break; | 1449 | break; |
| 1450 | } | 1450 | } |
| 1451 | btrfs_release_path(path); | 1451 | btrfs_release_path(path); |
| 1452 | 1452 | ||
| 1453 | if (cur_offset <= end && cow_start == (u64)-1) { | 1453 | if (cur_offset <= end && cow_start == (u64)-1) { |
| 1454 | cow_start = cur_offset; | 1454 | cow_start = cur_offset; |
| 1455 | cur_offset = end; | 1455 | cur_offset = end; |
| 1456 | } | 1456 | } |
| 1457 | 1457 | ||
| 1458 | if (cow_start != (u64)-1) { | 1458 | if (cow_start != (u64)-1) { |
| 1459 | ret = cow_file_range(inode, locked_page, cow_start, end, | 1459 | ret = cow_file_range(inode, locked_page, cow_start, end, |
| 1460 | page_started, nr_written, 1); | 1460 | page_started, nr_written, 1); |
| 1461 | if (ret) | 1461 | if (ret) |
| 1462 | goto error; | 1462 | goto error; |
| 1463 | } | 1463 | } |
| 1464 | 1464 | ||
| 1465 | error: | 1465 | error: |
| 1466 | err = btrfs_end_transaction(trans, root); | 1466 | err = btrfs_end_transaction(trans, root); |
| 1467 | if (!ret) | 1467 | if (!ret) |
| 1468 | ret = err; | 1468 | ret = err; |
| 1469 | 1469 | ||
| 1470 | if (ret && cur_offset < end) | 1470 | if (ret && cur_offset < end) |
| 1471 | extent_clear_unlock_delalloc(inode, cur_offset, end, | 1471 | extent_clear_unlock_delalloc(inode, cur_offset, end, |
| 1472 | locked_page, EXTENT_LOCKED | | 1472 | locked_page, EXTENT_LOCKED | |
| 1473 | EXTENT_DELALLOC | EXTENT_DEFRAG | | 1473 | EXTENT_DELALLOC | EXTENT_DEFRAG | |
| 1474 | EXTENT_DO_ACCOUNTING, PAGE_UNLOCK | | 1474 | EXTENT_DO_ACCOUNTING, PAGE_UNLOCK | |
| 1475 | PAGE_CLEAR_DIRTY | | 1475 | PAGE_CLEAR_DIRTY | |
| 1476 | PAGE_SET_WRITEBACK | | 1476 | PAGE_SET_WRITEBACK | |
| 1477 | PAGE_END_WRITEBACK); | 1477 | PAGE_END_WRITEBACK); |
| 1478 | btrfs_free_path(path); | 1478 | btrfs_free_path(path); |
| 1479 | return ret; | 1479 | return ret; |
| 1480 | } | 1480 | } |
| 1481 | 1481 | ||
| 1482 | static inline int need_force_cow(struct inode *inode, u64 start, u64 end) | 1482 | static inline int need_force_cow(struct inode *inode, u64 start, u64 end) |
| 1483 | { | 1483 | { |
| 1484 | 1484 | ||
| 1485 | if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) && | 1485 | if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) && |
| 1486 | !(BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC)) | 1486 | !(BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC)) |
| 1487 | return 0; | 1487 | return 0; |
| 1488 | 1488 | ||
| 1489 | /* | 1489 | /* |
| 1490 | * @defrag_bytes is a hint value, no spinlock held here, | 1490 | * @defrag_bytes is a hint value, no spinlock held here, |
| 1491 | * if is not zero, it means the file is defragging. | 1491 | * if is not zero, it means the file is defragging. |
| 1492 | * Force cow if given extent needs to be defragged. | 1492 | * Force cow if given extent needs to be defragged. |
| 1493 | */ | 1493 | */ |
| 1494 | if (BTRFS_I(inode)->defrag_bytes && | 1494 | if (BTRFS_I(inode)->defrag_bytes && |
| 1495 | test_range_bit(&BTRFS_I(inode)->io_tree, start, end, | 1495 | test_range_bit(&BTRFS_I(inode)->io_tree, start, end, |
| 1496 | EXTENT_DEFRAG, 0, NULL)) | 1496 | EXTENT_DEFRAG, 0, NULL)) |
| 1497 | return 1; | 1497 | return 1; |
| 1498 | 1498 | ||
| 1499 | return 0; | 1499 | return 0; |
| 1500 | } | 1500 | } |
| 1501 | 1501 | ||
| 1502 | /* | 1502 | /* |
| 1503 | * extent_io.c call back to do delayed allocation processing | 1503 | * extent_io.c call back to do delayed allocation processing |
| 1504 | */ | 1504 | */ |
| 1505 | static int run_delalloc_range(struct inode *inode, struct page *locked_page, | 1505 | static int run_delalloc_range(struct inode *inode, struct page *locked_page, |
| 1506 | u64 start, u64 end, int *page_started, | 1506 | u64 start, u64 end, int *page_started, |
| 1507 | unsigned long *nr_written) | 1507 | unsigned long *nr_written) |
| 1508 | { | 1508 | { |
| 1509 | int ret; | 1509 | int ret; |
| 1510 | int force_cow = need_force_cow(inode, start, end); | 1510 | int force_cow = need_force_cow(inode, start, end); |
| 1511 | 1511 | ||
| 1512 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW && !force_cow) { | 1512 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW && !force_cow) { |
| 1513 | ret = run_delalloc_nocow(inode, locked_page, start, end, | 1513 | ret = run_delalloc_nocow(inode, locked_page, start, end, |
| 1514 | page_started, 1, nr_written); | 1514 | page_started, 1, nr_written); |
| 1515 | } else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC && !force_cow) { | 1515 | } else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC && !force_cow) { |
| 1516 | ret = run_delalloc_nocow(inode, locked_page, start, end, | 1516 | ret = run_delalloc_nocow(inode, locked_page, start, end, |
| 1517 | page_started, 0, nr_written); | 1517 | page_started, 0, nr_written); |
| 1518 | } else if (!inode_need_compress(inode)) { | 1518 | } else if (!inode_need_compress(inode)) { |
| 1519 | ret = cow_file_range(inode, locked_page, start, end, | 1519 | ret = cow_file_range(inode, locked_page, start, end, |
| 1520 | page_started, nr_written, 1); | 1520 | page_started, nr_written, 1); |
| 1521 | } else { | 1521 | } else { |
| 1522 | set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, | 1522 | set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, |
| 1523 | &BTRFS_I(inode)->runtime_flags); | 1523 | &BTRFS_I(inode)->runtime_flags); |
| 1524 | ret = cow_file_range_async(inode, locked_page, start, end, | 1524 | ret = cow_file_range_async(inode, locked_page, start, end, |
| 1525 | page_started, nr_written); | 1525 | page_started, nr_written); |
| 1526 | } | 1526 | } |
| 1527 | return ret; | 1527 | return ret; |
| 1528 | } | 1528 | } |
| 1529 | 1529 | ||
| 1530 | static void btrfs_split_extent_hook(struct inode *inode, | 1530 | static void btrfs_split_extent_hook(struct inode *inode, |
| 1531 | struct extent_state *orig, u64 split) | 1531 | struct extent_state *orig, u64 split) |
| 1532 | { | 1532 | { |
| 1533 | /* not delalloc, ignore it */ | 1533 | /* not delalloc, ignore it */ |
| 1534 | if (!(orig->state & EXTENT_DELALLOC)) | 1534 | if (!(orig->state & EXTENT_DELALLOC)) |
| 1535 | return; | 1535 | return; |
| 1536 | 1536 | ||
| 1537 | spin_lock(&BTRFS_I(inode)->lock); | 1537 | spin_lock(&BTRFS_I(inode)->lock); |
| 1538 | BTRFS_I(inode)->outstanding_extents++; | 1538 | BTRFS_I(inode)->outstanding_extents++; |
| 1539 | spin_unlock(&BTRFS_I(inode)->lock); | 1539 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1540 | } | 1540 | } |
| 1541 | 1541 | ||
| 1542 | /* | 1542 | /* |
| 1543 | * extent_io.c merge_extent_hook, used to track merged delayed allocation | 1543 | * extent_io.c merge_extent_hook, used to track merged delayed allocation |
| 1544 | * extents so we can keep track of new extents that are just merged onto old | 1544 | * extents so we can keep track of new extents that are just merged onto old |
| 1545 | * extents, such as when we are doing sequential writes, so we can properly | 1545 | * extents, such as when we are doing sequential writes, so we can properly |
| 1546 | * account for the metadata space we'll need. | 1546 | * account for the metadata space we'll need. |
| 1547 | */ | 1547 | */ |
| 1548 | static void btrfs_merge_extent_hook(struct inode *inode, | 1548 | static void btrfs_merge_extent_hook(struct inode *inode, |
| 1549 | struct extent_state *new, | 1549 | struct extent_state *new, |
| 1550 | struct extent_state *other) | 1550 | struct extent_state *other) |
| 1551 | { | 1551 | { |
| 1552 | /* not delalloc, ignore it */ | 1552 | /* not delalloc, ignore it */ |
| 1553 | if (!(other->state & EXTENT_DELALLOC)) | 1553 | if (!(other->state & EXTENT_DELALLOC)) |
| 1554 | return; | 1554 | return; |
| 1555 | 1555 | ||
| 1556 | spin_lock(&BTRFS_I(inode)->lock); | 1556 | spin_lock(&BTRFS_I(inode)->lock); |
| 1557 | BTRFS_I(inode)->outstanding_extents--; | 1557 | BTRFS_I(inode)->outstanding_extents--; |
| 1558 | spin_unlock(&BTRFS_I(inode)->lock); | 1558 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1559 | } | 1559 | } |
| 1560 | 1560 | ||
| 1561 | static void btrfs_add_delalloc_inodes(struct btrfs_root *root, | 1561 | static void btrfs_add_delalloc_inodes(struct btrfs_root *root, |
| 1562 | struct inode *inode) | 1562 | struct inode *inode) |
| 1563 | { | 1563 | { |
| 1564 | spin_lock(&root->delalloc_lock); | 1564 | spin_lock(&root->delalloc_lock); |
| 1565 | if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) { | 1565 | if (list_empty(&BTRFS_I(inode)->delalloc_inodes)) { |
| 1566 | list_add_tail(&BTRFS_I(inode)->delalloc_inodes, | 1566 | list_add_tail(&BTRFS_I(inode)->delalloc_inodes, |
| 1567 | &root->delalloc_inodes); | 1567 | &root->delalloc_inodes); |
| 1568 | set_bit(BTRFS_INODE_IN_DELALLOC_LIST, | 1568 | set_bit(BTRFS_INODE_IN_DELALLOC_LIST, |
| 1569 | &BTRFS_I(inode)->runtime_flags); | 1569 | &BTRFS_I(inode)->runtime_flags); |
| 1570 | root->nr_delalloc_inodes++; | 1570 | root->nr_delalloc_inodes++; |
| 1571 | if (root->nr_delalloc_inodes == 1) { | 1571 | if (root->nr_delalloc_inodes == 1) { |
| 1572 | spin_lock(&root->fs_info->delalloc_root_lock); | 1572 | spin_lock(&root->fs_info->delalloc_root_lock); |
| 1573 | BUG_ON(!list_empty(&root->delalloc_root)); | 1573 | BUG_ON(!list_empty(&root->delalloc_root)); |
| 1574 | list_add_tail(&root->delalloc_root, | 1574 | list_add_tail(&root->delalloc_root, |
| 1575 | &root->fs_info->delalloc_roots); | 1575 | &root->fs_info->delalloc_roots); |
| 1576 | spin_unlock(&root->fs_info->delalloc_root_lock); | 1576 | spin_unlock(&root->fs_info->delalloc_root_lock); |
| 1577 | } | 1577 | } |
| 1578 | } | 1578 | } |
| 1579 | spin_unlock(&root->delalloc_lock); | 1579 | spin_unlock(&root->delalloc_lock); |
| 1580 | } | 1580 | } |
| 1581 | 1581 | ||
| 1582 | static void btrfs_del_delalloc_inode(struct btrfs_root *root, | 1582 | static void btrfs_del_delalloc_inode(struct btrfs_root *root, |
| 1583 | struct inode *inode) | 1583 | struct inode *inode) |
| 1584 | { | 1584 | { |
| 1585 | spin_lock(&root->delalloc_lock); | 1585 | spin_lock(&root->delalloc_lock); |
| 1586 | if (!list_empty(&BTRFS_I(inode)->delalloc_inodes)) { | 1586 | if (!list_empty(&BTRFS_I(inode)->delalloc_inodes)) { |
| 1587 | list_del_init(&BTRFS_I(inode)->delalloc_inodes); | 1587 | list_del_init(&BTRFS_I(inode)->delalloc_inodes); |
| 1588 | clear_bit(BTRFS_INODE_IN_DELALLOC_LIST, | 1588 | clear_bit(BTRFS_INODE_IN_DELALLOC_LIST, |
| 1589 | &BTRFS_I(inode)->runtime_flags); | 1589 | &BTRFS_I(inode)->runtime_flags); |
| 1590 | root->nr_delalloc_inodes--; | 1590 | root->nr_delalloc_inodes--; |
| 1591 | if (!root->nr_delalloc_inodes) { | 1591 | if (!root->nr_delalloc_inodes) { |
| 1592 | spin_lock(&root->fs_info->delalloc_root_lock); | 1592 | spin_lock(&root->fs_info->delalloc_root_lock); |
| 1593 | BUG_ON(list_empty(&root->delalloc_root)); | 1593 | BUG_ON(list_empty(&root->delalloc_root)); |
| 1594 | list_del_init(&root->delalloc_root); | 1594 | list_del_init(&root->delalloc_root); |
| 1595 | spin_unlock(&root->fs_info->delalloc_root_lock); | 1595 | spin_unlock(&root->fs_info->delalloc_root_lock); |
| 1596 | } | 1596 | } |
| 1597 | } | 1597 | } |
| 1598 | spin_unlock(&root->delalloc_lock); | 1598 | spin_unlock(&root->delalloc_lock); |
| 1599 | } | 1599 | } |
| 1600 | 1600 | ||
| 1601 | /* | 1601 | /* |
| 1602 | * extent_io.c set_bit_hook, used to track delayed allocation | 1602 | * extent_io.c set_bit_hook, used to track delayed allocation |
| 1603 | * bytes in this file, and to maintain the list of inodes that | 1603 | * bytes in this file, and to maintain the list of inodes that |
| 1604 | * have pending delalloc work to be done. | 1604 | * have pending delalloc work to be done. |
| 1605 | */ | 1605 | */ |
| 1606 | static void btrfs_set_bit_hook(struct inode *inode, | 1606 | static void btrfs_set_bit_hook(struct inode *inode, |
| 1607 | struct extent_state *state, unsigned long *bits) | 1607 | struct extent_state *state, unsigned long *bits) |
| 1608 | { | 1608 | { |
| 1609 | 1609 | ||
| 1610 | if ((*bits & EXTENT_DEFRAG) && !(*bits & EXTENT_DELALLOC)) | 1610 | if ((*bits & EXTENT_DEFRAG) && !(*bits & EXTENT_DELALLOC)) |
| 1611 | WARN_ON(1); | 1611 | WARN_ON(1); |
| 1612 | /* | 1612 | /* |
| 1613 | * set_bit and clear bit hooks normally require _irqsave/restore | 1613 | * set_bit and clear bit hooks normally require _irqsave/restore |
| 1614 | * but in this case, we are only testing for the DELALLOC | 1614 | * but in this case, we are only testing for the DELALLOC |
| 1615 | * bit, which is only set or cleared with irqs on | 1615 | * bit, which is only set or cleared with irqs on |
| 1616 | */ | 1616 | */ |
| 1617 | if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { | 1617 | if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { |
| 1618 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1618 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1619 | u64 len = state->end + 1 - state->start; | 1619 | u64 len = state->end + 1 - state->start; |
| 1620 | bool do_list = !btrfs_is_free_space_inode(inode); | 1620 | bool do_list = !btrfs_is_free_space_inode(inode); |
| 1621 | 1621 | ||
| 1622 | if (*bits & EXTENT_FIRST_DELALLOC) { | 1622 | if (*bits & EXTENT_FIRST_DELALLOC) { |
| 1623 | *bits &= ~EXTENT_FIRST_DELALLOC; | 1623 | *bits &= ~EXTENT_FIRST_DELALLOC; |
| 1624 | } else { | 1624 | } else { |
| 1625 | spin_lock(&BTRFS_I(inode)->lock); | 1625 | spin_lock(&BTRFS_I(inode)->lock); |
| 1626 | BTRFS_I(inode)->outstanding_extents++; | 1626 | BTRFS_I(inode)->outstanding_extents++; |
| 1627 | spin_unlock(&BTRFS_I(inode)->lock); | 1627 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1628 | } | 1628 | } |
| 1629 | 1629 | ||
| 1630 | __percpu_counter_add(&root->fs_info->delalloc_bytes, len, | 1630 | __percpu_counter_add(&root->fs_info->delalloc_bytes, len, |
| 1631 | root->fs_info->delalloc_batch); | 1631 | root->fs_info->delalloc_batch); |
| 1632 | spin_lock(&BTRFS_I(inode)->lock); | 1632 | spin_lock(&BTRFS_I(inode)->lock); |
| 1633 | BTRFS_I(inode)->delalloc_bytes += len; | 1633 | BTRFS_I(inode)->delalloc_bytes += len; |
| 1634 | if (*bits & EXTENT_DEFRAG) | 1634 | if (*bits & EXTENT_DEFRAG) |
| 1635 | BTRFS_I(inode)->defrag_bytes += len; | 1635 | BTRFS_I(inode)->defrag_bytes += len; |
| 1636 | if (do_list && !test_bit(BTRFS_INODE_IN_DELALLOC_LIST, | 1636 | if (do_list && !test_bit(BTRFS_INODE_IN_DELALLOC_LIST, |
| 1637 | &BTRFS_I(inode)->runtime_flags)) | 1637 | &BTRFS_I(inode)->runtime_flags)) |
| 1638 | btrfs_add_delalloc_inodes(root, inode); | 1638 | btrfs_add_delalloc_inodes(root, inode); |
| 1639 | spin_unlock(&BTRFS_I(inode)->lock); | 1639 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1640 | } | 1640 | } |
| 1641 | } | 1641 | } |
| 1642 | 1642 | ||
| 1643 | /* | 1643 | /* |
| 1644 | * extent_io.c clear_bit_hook, see set_bit_hook for why | 1644 | * extent_io.c clear_bit_hook, see set_bit_hook for why |
| 1645 | */ | 1645 | */ |
| 1646 | static void btrfs_clear_bit_hook(struct inode *inode, | 1646 | static void btrfs_clear_bit_hook(struct inode *inode, |
| 1647 | struct extent_state *state, | 1647 | struct extent_state *state, |
| 1648 | unsigned long *bits) | 1648 | unsigned long *bits) |
| 1649 | { | 1649 | { |
| 1650 | u64 len = state->end + 1 - state->start; | 1650 | u64 len = state->end + 1 - state->start; |
| 1651 | 1651 | ||
| 1652 | spin_lock(&BTRFS_I(inode)->lock); | 1652 | spin_lock(&BTRFS_I(inode)->lock); |
| 1653 | if ((state->state & EXTENT_DEFRAG) && (*bits & EXTENT_DEFRAG)) | 1653 | if ((state->state & EXTENT_DEFRAG) && (*bits & EXTENT_DEFRAG)) |
| 1654 | BTRFS_I(inode)->defrag_bytes -= len; | 1654 | BTRFS_I(inode)->defrag_bytes -= len; |
| 1655 | spin_unlock(&BTRFS_I(inode)->lock); | 1655 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1656 | 1656 | ||
| 1657 | /* | 1657 | /* |
| 1658 | * set_bit and clear bit hooks normally require _irqsave/restore | 1658 | * set_bit and clear bit hooks normally require _irqsave/restore |
| 1659 | * but in this case, we are only testing for the DELALLOC | 1659 | * but in this case, we are only testing for the DELALLOC |
| 1660 | * bit, which is only set or cleared with irqs on | 1660 | * bit, which is only set or cleared with irqs on |
| 1661 | */ | 1661 | */ |
| 1662 | if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { | 1662 | if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { |
| 1663 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1663 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1664 | bool do_list = !btrfs_is_free_space_inode(inode); | 1664 | bool do_list = !btrfs_is_free_space_inode(inode); |
| 1665 | 1665 | ||
| 1666 | if (*bits & EXTENT_FIRST_DELALLOC) { | 1666 | if (*bits & EXTENT_FIRST_DELALLOC) { |
| 1667 | *bits &= ~EXTENT_FIRST_DELALLOC; | 1667 | *bits &= ~EXTENT_FIRST_DELALLOC; |
| 1668 | } else if (!(*bits & EXTENT_DO_ACCOUNTING)) { | 1668 | } else if (!(*bits & EXTENT_DO_ACCOUNTING)) { |
| 1669 | spin_lock(&BTRFS_I(inode)->lock); | 1669 | spin_lock(&BTRFS_I(inode)->lock); |
| 1670 | BTRFS_I(inode)->outstanding_extents--; | 1670 | BTRFS_I(inode)->outstanding_extents--; |
| 1671 | spin_unlock(&BTRFS_I(inode)->lock); | 1671 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1672 | } | 1672 | } |
| 1673 | 1673 | ||
| 1674 | /* | 1674 | /* |
| 1675 | * We don't reserve metadata space for space cache inodes so we | 1675 | * We don't reserve metadata space for space cache inodes so we |
| 1676 | * don't need to call dellalloc_release_metadata if there is an | 1676 | * don't need to call dellalloc_release_metadata if there is an |
| 1677 | * error. | 1677 | * error. |
| 1678 | */ | 1678 | */ |
| 1679 | if (*bits & EXTENT_DO_ACCOUNTING && | 1679 | if (*bits & EXTENT_DO_ACCOUNTING && |
| 1680 | root != root->fs_info->tree_root) | 1680 | root != root->fs_info->tree_root) |
| 1681 | btrfs_delalloc_release_metadata(inode, len); | 1681 | btrfs_delalloc_release_metadata(inode, len); |
| 1682 | 1682 | ||
| 1683 | if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID | 1683 | if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID |
| 1684 | && do_list && !(state->state & EXTENT_NORESERVE)) | 1684 | && do_list && !(state->state & EXTENT_NORESERVE)) |
| 1685 | btrfs_free_reserved_data_space(inode, len); | 1685 | btrfs_free_reserved_data_space(inode, len); |
| 1686 | 1686 | ||
| 1687 | __percpu_counter_add(&root->fs_info->delalloc_bytes, -len, | 1687 | __percpu_counter_add(&root->fs_info->delalloc_bytes, -len, |
| 1688 | root->fs_info->delalloc_batch); | 1688 | root->fs_info->delalloc_batch); |
| 1689 | spin_lock(&BTRFS_I(inode)->lock); | 1689 | spin_lock(&BTRFS_I(inode)->lock); |
| 1690 | BTRFS_I(inode)->delalloc_bytes -= len; | 1690 | BTRFS_I(inode)->delalloc_bytes -= len; |
| 1691 | if (do_list && BTRFS_I(inode)->delalloc_bytes == 0 && | 1691 | if (do_list && BTRFS_I(inode)->delalloc_bytes == 0 && |
| 1692 | test_bit(BTRFS_INODE_IN_DELALLOC_LIST, | 1692 | test_bit(BTRFS_INODE_IN_DELALLOC_LIST, |
| 1693 | &BTRFS_I(inode)->runtime_flags)) | 1693 | &BTRFS_I(inode)->runtime_flags)) |
| 1694 | btrfs_del_delalloc_inode(root, inode); | 1694 | btrfs_del_delalloc_inode(root, inode); |
| 1695 | spin_unlock(&BTRFS_I(inode)->lock); | 1695 | spin_unlock(&BTRFS_I(inode)->lock); |
| 1696 | } | 1696 | } |
| 1697 | } | 1697 | } |
| 1698 | 1698 | ||
| 1699 | /* | 1699 | /* |
| 1700 | * extent_io.c merge_bio_hook, this must check the chunk tree to make sure | 1700 | * extent_io.c merge_bio_hook, this must check the chunk tree to make sure |
| 1701 | * we don't create bios that span stripes or chunks | 1701 | * we don't create bios that span stripes or chunks |
| 1702 | */ | 1702 | */ |
| 1703 | int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset, | 1703 | int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset, |
| 1704 | size_t size, struct bio *bio, | 1704 | size_t size, struct bio *bio, |
| 1705 | unsigned long bio_flags) | 1705 | unsigned long bio_flags) |
| 1706 | { | 1706 | { |
| 1707 | struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; | 1707 | struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; |
| 1708 | u64 logical = (u64)bio->bi_iter.bi_sector << 9; | 1708 | u64 logical = (u64)bio->bi_iter.bi_sector << 9; |
| 1709 | u64 length = 0; | 1709 | u64 length = 0; |
| 1710 | u64 map_length; | 1710 | u64 map_length; |
| 1711 | int ret; | 1711 | int ret; |
| 1712 | 1712 | ||
| 1713 | if (bio_flags & EXTENT_BIO_COMPRESSED) | 1713 | if (bio_flags & EXTENT_BIO_COMPRESSED) |
| 1714 | return 0; | 1714 | return 0; |
| 1715 | 1715 | ||
| 1716 | length = bio->bi_iter.bi_size; | 1716 | length = bio->bi_iter.bi_size; |
| 1717 | map_length = length; | 1717 | map_length = length; |
| 1718 | ret = btrfs_map_block(root->fs_info, rw, logical, | 1718 | ret = btrfs_map_block(root->fs_info, rw, logical, |
| 1719 | &map_length, NULL, 0); | 1719 | &map_length, NULL, 0); |
| 1720 | /* Will always return 0 with map_multi == NULL */ | 1720 | /* Will always return 0 with map_multi == NULL */ |
| 1721 | BUG_ON(ret < 0); | 1721 | BUG_ON(ret < 0); |
| 1722 | if (map_length < length + size) | 1722 | if (map_length < length + size) |
| 1723 | return 1; | 1723 | return 1; |
| 1724 | return 0; | 1724 | return 0; |
| 1725 | } | 1725 | } |
| 1726 | 1726 | ||
| 1727 | /* | 1727 | /* |
| 1728 | * in order to insert checksums into the metadata in large chunks, | 1728 | * in order to insert checksums into the metadata in large chunks, |
| 1729 | * we wait until bio submission time. All the pages in the bio are | 1729 | * we wait until bio submission time. All the pages in the bio are |
| 1730 | * checksummed and sums are attached onto the ordered extent record. | 1730 | * checksummed and sums are attached onto the ordered extent record. |
| 1731 | * | 1731 | * |
| 1732 | * At IO completion time the cums attached on the ordered extent record | 1732 | * At IO completion time the cums attached on the ordered extent record |
| 1733 | * are inserted into the btree | 1733 | * are inserted into the btree |
| 1734 | */ | 1734 | */ |
| 1735 | static int __btrfs_submit_bio_start(struct inode *inode, int rw, | 1735 | static int __btrfs_submit_bio_start(struct inode *inode, int rw, |
| 1736 | struct bio *bio, int mirror_num, | 1736 | struct bio *bio, int mirror_num, |
| 1737 | unsigned long bio_flags, | 1737 | unsigned long bio_flags, |
| 1738 | u64 bio_offset) | 1738 | u64 bio_offset) |
| 1739 | { | 1739 | { |
| 1740 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1740 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1741 | int ret = 0; | 1741 | int ret = 0; |
| 1742 | 1742 | ||
| 1743 | ret = btrfs_csum_one_bio(root, inode, bio, 0, 0); | 1743 | ret = btrfs_csum_one_bio(root, inode, bio, 0, 0); |
| 1744 | BUG_ON(ret); /* -ENOMEM */ | 1744 | BUG_ON(ret); /* -ENOMEM */ |
| 1745 | return 0; | 1745 | return 0; |
| 1746 | } | 1746 | } |
| 1747 | 1747 | ||
| 1748 | /* | 1748 | /* |
| 1749 | * in order to insert checksums into the metadata in large chunks, | 1749 | * in order to insert checksums into the metadata in large chunks, |
| 1750 | * we wait until bio submission time. All the pages in the bio are | 1750 | * we wait until bio submission time. All the pages in the bio are |
| 1751 | * checksummed and sums are attached onto the ordered extent record. | 1751 | * checksummed and sums are attached onto the ordered extent record. |
| 1752 | * | 1752 | * |
| 1753 | * At IO completion time the cums attached on the ordered extent record | 1753 | * At IO completion time the cums attached on the ordered extent record |
| 1754 | * are inserted into the btree | 1754 | * are inserted into the btree |
| 1755 | */ | 1755 | */ |
| 1756 | static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio, | 1756 | static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio, |
| 1757 | int mirror_num, unsigned long bio_flags, | 1757 | int mirror_num, unsigned long bio_flags, |
| 1758 | u64 bio_offset) | 1758 | u64 bio_offset) |
| 1759 | { | 1759 | { |
| 1760 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1760 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1761 | int ret; | 1761 | int ret; |
| 1762 | 1762 | ||
| 1763 | ret = btrfs_map_bio(root, rw, bio, mirror_num, 1); | 1763 | ret = btrfs_map_bio(root, rw, bio, mirror_num, 1); |
| 1764 | if (ret) | 1764 | if (ret) |
| 1765 | bio_endio(bio, ret); | 1765 | bio_endio(bio, ret); |
| 1766 | return ret; | 1766 | return ret; |
| 1767 | } | 1767 | } |
| 1768 | 1768 | ||
| 1769 | /* | 1769 | /* |
| 1770 | * extent_io.c submission hook. This does the right thing for csum calculation | 1770 | * extent_io.c submission hook. This does the right thing for csum calculation |
| 1771 | * on write, or reading the csums from the tree before a read | 1771 | * on write, or reading the csums from the tree before a read |
| 1772 | */ | 1772 | */ |
| 1773 | static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, | 1773 | static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, |
| 1774 | int mirror_num, unsigned long bio_flags, | 1774 | int mirror_num, unsigned long bio_flags, |
| 1775 | u64 bio_offset) | 1775 | u64 bio_offset) |
| 1776 | { | 1776 | { |
| 1777 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1777 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1778 | int ret = 0; | 1778 | int ret = 0; |
| 1779 | int skip_sum; | 1779 | int skip_sum; |
| 1780 | int metadata = 0; | 1780 | int metadata = 0; |
| 1781 | int async = !atomic_read(&BTRFS_I(inode)->sync_writers); | 1781 | int async = !atomic_read(&BTRFS_I(inode)->sync_writers); |
| 1782 | 1782 | ||
| 1783 | skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; | 1783 | skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; |
| 1784 | 1784 | ||
| 1785 | if (btrfs_is_free_space_inode(inode)) | 1785 | if (btrfs_is_free_space_inode(inode)) |
| 1786 | metadata = 2; | 1786 | metadata = 2; |
| 1787 | 1787 | ||
| 1788 | if (!(rw & REQ_WRITE)) { | 1788 | if (!(rw & REQ_WRITE)) { |
| 1789 | ret = btrfs_bio_wq_end_io(root->fs_info, bio, metadata); | 1789 | ret = btrfs_bio_wq_end_io(root->fs_info, bio, metadata); |
| 1790 | if (ret) | 1790 | if (ret) |
| 1791 | goto out; | 1791 | goto out; |
| 1792 | 1792 | ||
| 1793 | if (bio_flags & EXTENT_BIO_COMPRESSED) { | 1793 | if (bio_flags & EXTENT_BIO_COMPRESSED) { |
| 1794 | ret = btrfs_submit_compressed_read(inode, bio, | 1794 | ret = btrfs_submit_compressed_read(inode, bio, |
| 1795 | mirror_num, | 1795 | mirror_num, |
| 1796 | bio_flags); | 1796 | bio_flags); |
| 1797 | goto out; | 1797 | goto out; |
| 1798 | } else if (!skip_sum) { | 1798 | } else if (!skip_sum) { |
| 1799 | ret = btrfs_lookup_bio_sums(root, inode, bio, NULL); | 1799 | ret = btrfs_lookup_bio_sums(root, inode, bio, NULL); |
| 1800 | if (ret) | 1800 | if (ret) |
| 1801 | goto out; | 1801 | goto out; |
| 1802 | } | 1802 | } |
| 1803 | goto mapit; | 1803 | goto mapit; |
| 1804 | } else if (async && !skip_sum) { | 1804 | } else if (async && !skip_sum) { |
| 1805 | /* csum items have already been cloned */ | 1805 | /* csum items have already been cloned */ |
| 1806 | if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) | 1806 | if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) |
| 1807 | goto mapit; | 1807 | goto mapit; |
| 1808 | /* we're doing a write, do the async checksumming */ | 1808 | /* we're doing a write, do the async checksumming */ |
| 1809 | ret = btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info, | 1809 | ret = btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info, |
| 1810 | inode, rw, bio, mirror_num, | 1810 | inode, rw, bio, mirror_num, |
| 1811 | bio_flags, bio_offset, | 1811 | bio_flags, bio_offset, |
| 1812 | __btrfs_submit_bio_start, | 1812 | __btrfs_submit_bio_start, |
| 1813 | __btrfs_submit_bio_done); | 1813 | __btrfs_submit_bio_done); |
| 1814 | goto out; | 1814 | goto out; |
| 1815 | } else if (!skip_sum) { | 1815 | } else if (!skip_sum) { |
| 1816 | ret = btrfs_csum_one_bio(root, inode, bio, 0, 0); | 1816 | ret = btrfs_csum_one_bio(root, inode, bio, 0, 0); |
| 1817 | if (ret) | 1817 | if (ret) |
| 1818 | goto out; | 1818 | goto out; |
| 1819 | } | 1819 | } |
| 1820 | 1820 | ||
| 1821 | mapit: | 1821 | mapit: |
| 1822 | ret = btrfs_map_bio(root, rw, bio, mirror_num, 0); | 1822 | ret = btrfs_map_bio(root, rw, bio, mirror_num, 0); |
| 1823 | 1823 | ||
| 1824 | out: | 1824 | out: |
| 1825 | if (ret < 0) | 1825 | if (ret < 0) |
| 1826 | bio_endio(bio, ret); | 1826 | bio_endio(bio, ret); |
| 1827 | return ret; | 1827 | return ret; |
| 1828 | } | 1828 | } |
| 1829 | 1829 | ||
| 1830 | /* | 1830 | /* |
| 1831 | * given a list of ordered sums record them in the inode. This happens | 1831 | * given a list of ordered sums record them in the inode. This happens |
| 1832 | * at IO completion time based on sums calculated at bio submission time. | 1832 | * at IO completion time based on sums calculated at bio submission time. |
| 1833 | */ | 1833 | */ |
| 1834 | static noinline int add_pending_csums(struct btrfs_trans_handle *trans, | 1834 | static noinline int add_pending_csums(struct btrfs_trans_handle *trans, |
| 1835 | struct inode *inode, u64 file_offset, | 1835 | struct inode *inode, u64 file_offset, |
| 1836 | struct list_head *list) | 1836 | struct list_head *list) |
| 1837 | { | 1837 | { |
| 1838 | struct btrfs_ordered_sum *sum; | 1838 | struct btrfs_ordered_sum *sum; |
| 1839 | 1839 | ||
| 1840 | list_for_each_entry(sum, list, list) { | 1840 | list_for_each_entry(sum, list, list) { |
| 1841 | trans->adding_csums = 1; | 1841 | trans->adding_csums = 1; |
| 1842 | btrfs_csum_file_blocks(trans, | 1842 | btrfs_csum_file_blocks(trans, |
| 1843 | BTRFS_I(inode)->root->fs_info->csum_root, sum); | 1843 | BTRFS_I(inode)->root->fs_info->csum_root, sum); |
| 1844 | trans->adding_csums = 0; | 1844 | trans->adding_csums = 0; |
| 1845 | } | 1845 | } |
| 1846 | return 0; | 1846 | return 0; |
| 1847 | } | 1847 | } |
| 1848 | 1848 | ||
| 1849 | int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end, | 1849 | int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end, |
| 1850 | struct extent_state **cached_state) | 1850 | struct extent_state **cached_state) |
| 1851 | { | 1851 | { |
| 1852 | WARN_ON((end & (PAGE_CACHE_SIZE - 1)) == 0); | 1852 | WARN_ON((end & (PAGE_CACHE_SIZE - 1)) == 0); |
| 1853 | return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end, | 1853 | return set_extent_delalloc(&BTRFS_I(inode)->io_tree, start, end, |
| 1854 | cached_state, GFP_NOFS); | 1854 | cached_state, GFP_NOFS); |
| 1855 | } | 1855 | } |
| 1856 | 1856 | ||
| 1857 | /* see btrfs_writepage_start_hook for details on why this is required */ | 1857 | /* see btrfs_writepage_start_hook for details on why this is required */ |
| 1858 | struct btrfs_writepage_fixup { | 1858 | struct btrfs_writepage_fixup { |
| 1859 | struct page *page; | 1859 | struct page *page; |
| 1860 | struct btrfs_work work; | 1860 | struct btrfs_work work; |
| 1861 | }; | 1861 | }; |
| 1862 | 1862 | ||
| 1863 | static void btrfs_writepage_fixup_worker(struct btrfs_work *work) | 1863 | static void btrfs_writepage_fixup_worker(struct btrfs_work *work) |
| 1864 | { | 1864 | { |
| 1865 | struct btrfs_writepage_fixup *fixup; | 1865 | struct btrfs_writepage_fixup *fixup; |
| 1866 | struct btrfs_ordered_extent *ordered; | 1866 | struct btrfs_ordered_extent *ordered; |
| 1867 | struct extent_state *cached_state = NULL; | 1867 | struct extent_state *cached_state = NULL; |
| 1868 | struct page *page; | 1868 | struct page *page; |
| 1869 | struct inode *inode; | 1869 | struct inode *inode; |
| 1870 | u64 page_start; | 1870 | u64 page_start; |
| 1871 | u64 page_end; | 1871 | u64 page_end; |
| 1872 | int ret; | 1872 | int ret; |
| 1873 | 1873 | ||
| 1874 | fixup = container_of(work, struct btrfs_writepage_fixup, work); | 1874 | fixup = container_of(work, struct btrfs_writepage_fixup, work); |
| 1875 | page = fixup->page; | 1875 | page = fixup->page; |
| 1876 | again: | 1876 | again: |
| 1877 | lock_page(page); | 1877 | lock_page(page); |
| 1878 | if (!page->mapping || !PageDirty(page) || !PageChecked(page)) { | 1878 | if (!page->mapping || !PageDirty(page) || !PageChecked(page)) { |
| 1879 | ClearPageChecked(page); | 1879 | ClearPageChecked(page); |
| 1880 | goto out_page; | 1880 | goto out_page; |
| 1881 | } | 1881 | } |
| 1882 | 1882 | ||
| 1883 | inode = page->mapping->host; | 1883 | inode = page->mapping->host; |
| 1884 | page_start = page_offset(page); | 1884 | page_start = page_offset(page); |
| 1885 | page_end = page_offset(page) + PAGE_CACHE_SIZE - 1; | 1885 | page_end = page_offset(page) + PAGE_CACHE_SIZE - 1; |
| 1886 | 1886 | ||
| 1887 | lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, 0, | 1887 | lock_extent_bits(&BTRFS_I(inode)->io_tree, page_start, page_end, 0, |
| 1888 | &cached_state); | 1888 | &cached_state); |
| 1889 | 1889 | ||
| 1890 | /* already ordered? We're done */ | 1890 | /* already ordered? We're done */ |
| 1891 | if (PagePrivate2(page)) | 1891 | if (PagePrivate2(page)) |
| 1892 | goto out; | 1892 | goto out; |
| 1893 | 1893 | ||
| 1894 | ordered = btrfs_lookup_ordered_extent(inode, page_start); | 1894 | ordered = btrfs_lookup_ordered_extent(inode, page_start); |
| 1895 | if (ordered) { | 1895 | if (ordered) { |
| 1896 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, | 1896 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, |
| 1897 | page_end, &cached_state, GFP_NOFS); | 1897 | page_end, &cached_state, GFP_NOFS); |
| 1898 | unlock_page(page); | 1898 | unlock_page(page); |
| 1899 | btrfs_start_ordered_extent(inode, ordered, 1); | 1899 | btrfs_start_ordered_extent(inode, ordered, 1); |
| 1900 | btrfs_put_ordered_extent(ordered); | 1900 | btrfs_put_ordered_extent(ordered); |
| 1901 | goto again; | 1901 | goto again; |
| 1902 | } | 1902 | } |
| 1903 | 1903 | ||
| 1904 | ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); | 1904 | ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); |
| 1905 | if (ret) { | 1905 | if (ret) { |
| 1906 | mapping_set_error(page->mapping, ret); | 1906 | mapping_set_error(page->mapping, ret); |
| 1907 | end_extent_writepage(page, ret, page_start, page_end); | 1907 | end_extent_writepage(page, ret, page_start, page_end); |
| 1908 | ClearPageChecked(page); | 1908 | ClearPageChecked(page); |
| 1909 | goto out; | 1909 | goto out; |
| 1910 | } | 1910 | } |
| 1911 | 1911 | ||
| 1912 | btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state); | 1912 | btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state); |
| 1913 | ClearPageChecked(page); | 1913 | ClearPageChecked(page); |
| 1914 | set_page_dirty(page); | 1914 | set_page_dirty(page); |
| 1915 | out: | 1915 | out: |
| 1916 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end, | 1916 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, page_start, page_end, |
| 1917 | &cached_state, GFP_NOFS); | 1917 | &cached_state, GFP_NOFS); |
| 1918 | out_page: | 1918 | out_page: |
| 1919 | unlock_page(page); | 1919 | unlock_page(page); |
| 1920 | page_cache_release(page); | 1920 | page_cache_release(page); |
| 1921 | kfree(fixup); | 1921 | kfree(fixup); |
| 1922 | } | 1922 | } |
| 1923 | 1923 | ||
| 1924 | /* | 1924 | /* |
| 1925 | * There are a few paths in the higher layers of the kernel that directly | 1925 | * There are a few paths in the higher layers of the kernel that directly |
| 1926 | * set the page dirty bit without asking the filesystem if it is a | 1926 | * set the page dirty bit without asking the filesystem if it is a |
| 1927 | * good idea. This causes problems because we want to make sure COW | 1927 | * good idea. This causes problems because we want to make sure COW |
| 1928 | * properly happens and the data=ordered rules are followed. | 1928 | * properly happens and the data=ordered rules are followed. |
| 1929 | * | 1929 | * |
| 1930 | * In our case any range that doesn't have the ORDERED bit set | 1930 | * In our case any range that doesn't have the ORDERED bit set |
| 1931 | * hasn't been properly setup for IO. We kick off an async process | 1931 | * hasn't been properly setup for IO. We kick off an async process |
| 1932 | * to fix it up. The async helper will wait for ordered extents, set | 1932 | * to fix it up. The async helper will wait for ordered extents, set |
| 1933 | * the delalloc bit and make it safe to write the page. | 1933 | * the delalloc bit and make it safe to write the page. |
| 1934 | */ | 1934 | */ |
| 1935 | static int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end) | 1935 | static int btrfs_writepage_start_hook(struct page *page, u64 start, u64 end) |
| 1936 | { | 1936 | { |
| 1937 | struct inode *inode = page->mapping->host; | 1937 | struct inode *inode = page->mapping->host; |
| 1938 | struct btrfs_writepage_fixup *fixup; | 1938 | struct btrfs_writepage_fixup *fixup; |
| 1939 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1939 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1940 | 1940 | ||
| 1941 | /* this page is properly in the ordered list */ | 1941 | /* this page is properly in the ordered list */ |
| 1942 | if (TestClearPagePrivate2(page)) | 1942 | if (TestClearPagePrivate2(page)) |
| 1943 | return 0; | 1943 | return 0; |
| 1944 | 1944 | ||
| 1945 | if (PageChecked(page)) | 1945 | if (PageChecked(page)) |
| 1946 | return -EAGAIN; | 1946 | return -EAGAIN; |
| 1947 | 1947 | ||
| 1948 | fixup = kzalloc(sizeof(*fixup), GFP_NOFS); | 1948 | fixup = kzalloc(sizeof(*fixup), GFP_NOFS); |
| 1949 | if (!fixup) | 1949 | if (!fixup) |
| 1950 | return -EAGAIN; | 1950 | return -EAGAIN; |
| 1951 | 1951 | ||
| 1952 | SetPageChecked(page); | 1952 | SetPageChecked(page); |
| 1953 | page_cache_get(page); | 1953 | page_cache_get(page); |
| 1954 | btrfs_init_work(&fixup->work, btrfs_fixup_helper, | 1954 | btrfs_init_work(&fixup->work, btrfs_fixup_helper, |
| 1955 | btrfs_writepage_fixup_worker, NULL, NULL); | 1955 | btrfs_writepage_fixup_worker, NULL, NULL); |
| 1956 | fixup->page = page; | 1956 | fixup->page = page; |
| 1957 | btrfs_queue_work(root->fs_info->fixup_workers, &fixup->work); | 1957 | btrfs_queue_work(root->fs_info->fixup_workers, &fixup->work); |
| 1958 | return -EBUSY; | 1958 | return -EBUSY; |
| 1959 | } | 1959 | } |
| 1960 | 1960 | ||
| 1961 | static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, | 1961 | static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, |
| 1962 | struct inode *inode, u64 file_pos, | 1962 | struct inode *inode, u64 file_pos, |
| 1963 | u64 disk_bytenr, u64 disk_num_bytes, | 1963 | u64 disk_bytenr, u64 disk_num_bytes, |
| 1964 | u64 num_bytes, u64 ram_bytes, | 1964 | u64 num_bytes, u64 ram_bytes, |
| 1965 | u8 compression, u8 encryption, | 1965 | u8 compression, u8 encryption, |
| 1966 | u16 other_encoding, int extent_type) | 1966 | u16 other_encoding, int extent_type) |
| 1967 | { | 1967 | { |
| 1968 | struct btrfs_root *root = BTRFS_I(inode)->root; | 1968 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 1969 | struct btrfs_file_extent_item *fi; | 1969 | struct btrfs_file_extent_item *fi; |
| 1970 | struct btrfs_path *path; | 1970 | struct btrfs_path *path; |
| 1971 | struct extent_buffer *leaf; | 1971 | struct extent_buffer *leaf; |
| 1972 | struct btrfs_key ins; | 1972 | struct btrfs_key ins; |
| 1973 | int extent_inserted = 0; | 1973 | int extent_inserted = 0; |
| 1974 | int ret; | 1974 | int ret; |
| 1975 | 1975 | ||
| 1976 | path = btrfs_alloc_path(); | 1976 | path = btrfs_alloc_path(); |
| 1977 | if (!path) | 1977 | if (!path) |
| 1978 | return -ENOMEM; | 1978 | return -ENOMEM; |
| 1979 | 1979 | ||
| 1980 | /* | 1980 | /* |
| 1981 | * we may be replacing one extent in the tree with another. | 1981 | * we may be replacing one extent in the tree with another. |
| 1982 | * The new extent is pinned in the extent map, and we don't want | 1982 | * The new extent is pinned in the extent map, and we don't want |
| 1983 | * to drop it from the cache until it is completely in the btree. | 1983 | * to drop it from the cache until it is completely in the btree. |
| 1984 | * | 1984 | * |
| 1985 | * So, tell btrfs_drop_extents to leave this extent in the cache. | 1985 | * So, tell btrfs_drop_extents to leave this extent in the cache. |
| 1986 | * the caller is expected to unpin it and allow it to be merged | 1986 | * the caller is expected to unpin it and allow it to be merged |
| 1987 | * with the others. | 1987 | * with the others. |
| 1988 | */ | 1988 | */ |
| 1989 | ret = __btrfs_drop_extents(trans, root, inode, path, file_pos, | 1989 | ret = __btrfs_drop_extents(trans, root, inode, path, file_pos, |
| 1990 | file_pos + num_bytes, NULL, 0, | 1990 | file_pos + num_bytes, NULL, 0, |
| 1991 | 1, sizeof(*fi), &extent_inserted); | 1991 | 1, sizeof(*fi), &extent_inserted); |
| 1992 | if (ret) | 1992 | if (ret) |
| 1993 | goto out; | 1993 | goto out; |
| 1994 | 1994 | ||
| 1995 | if (!extent_inserted) { | 1995 | if (!extent_inserted) { |
| 1996 | ins.objectid = btrfs_ino(inode); | 1996 | ins.objectid = btrfs_ino(inode); |
| 1997 | ins.offset = file_pos; | 1997 | ins.offset = file_pos; |
| 1998 | ins.type = BTRFS_EXTENT_DATA_KEY; | 1998 | ins.type = BTRFS_EXTENT_DATA_KEY; |
| 1999 | 1999 | ||
| 2000 | path->leave_spinning = 1; | 2000 | path->leave_spinning = 1; |
| 2001 | ret = btrfs_insert_empty_item(trans, root, path, &ins, | 2001 | ret = btrfs_insert_empty_item(trans, root, path, &ins, |
| 2002 | sizeof(*fi)); | 2002 | sizeof(*fi)); |
| 2003 | if (ret) | 2003 | if (ret) |
| 2004 | goto out; | 2004 | goto out; |
| 2005 | } | 2005 | } |
| 2006 | leaf = path->nodes[0]; | 2006 | leaf = path->nodes[0]; |
| 2007 | fi = btrfs_item_ptr(leaf, path->slots[0], | 2007 | fi = btrfs_item_ptr(leaf, path->slots[0], |
| 2008 | struct btrfs_file_extent_item); | 2008 | struct btrfs_file_extent_item); |
| 2009 | btrfs_set_file_extent_generation(leaf, fi, trans->transid); | 2009 | btrfs_set_file_extent_generation(leaf, fi, trans->transid); |
| 2010 | btrfs_set_file_extent_type(leaf, fi, extent_type); | 2010 | btrfs_set_file_extent_type(leaf, fi, extent_type); |
| 2011 | btrfs_set_file_extent_disk_bytenr(leaf, fi, disk_bytenr); | 2011 | btrfs_set_file_extent_disk_bytenr(leaf, fi, disk_bytenr); |
| 2012 | btrfs_set_file_extent_disk_num_bytes(leaf, fi, disk_num_bytes); | 2012 | btrfs_set_file_extent_disk_num_bytes(leaf, fi, disk_num_bytes); |
| 2013 | btrfs_set_file_extent_offset(leaf, fi, 0); | 2013 | btrfs_set_file_extent_offset(leaf, fi, 0); |
| 2014 | btrfs_set_file_extent_num_bytes(leaf, fi, num_bytes); | 2014 | btrfs_set_file_extent_num_bytes(leaf, fi, num_bytes); |
| 2015 | btrfs_set_file_extent_ram_bytes(leaf, fi, ram_bytes); | 2015 | btrfs_set_file_extent_ram_bytes(leaf, fi, ram_bytes); |
| 2016 | btrfs_set_file_extent_compression(leaf, fi, compression); | 2016 | btrfs_set_file_extent_compression(leaf, fi, compression); |
| 2017 | btrfs_set_file_extent_encryption(leaf, fi, encryption); | 2017 | btrfs_set_file_extent_encryption(leaf, fi, encryption); |
| 2018 | btrfs_set_file_extent_other_encoding(leaf, fi, other_encoding); | 2018 | btrfs_set_file_extent_other_encoding(leaf, fi, other_encoding); |
| 2019 | 2019 | ||
| 2020 | btrfs_mark_buffer_dirty(leaf); | 2020 | btrfs_mark_buffer_dirty(leaf); |
| 2021 | btrfs_release_path(path); | 2021 | btrfs_release_path(path); |
| 2022 | 2022 | ||
| 2023 | inode_add_bytes(inode, num_bytes); | 2023 | inode_add_bytes(inode, num_bytes); |
| 2024 | 2024 | ||
| 2025 | ins.objectid = disk_bytenr; | 2025 | ins.objectid = disk_bytenr; |
| 2026 | ins.offset = disk_num_bytes; | 2026 | ins.offset = disk_num_bytes; |
| 2027 | ins.type = BTRFS_EXTENT_ITEM_KEY; | 2027 | ins.type = BTRFS_EXTENT_ITEM_KEY; |
| 2028 | ret = btrfs_alloc_reserved_file_extent(trans, root, | 2028 | ret = btrfs_alloc_reserved_file_extent(trans, root, |
| 2029 | root->root_key.objectid, | 2029 | root->root_key.objectid, |
| 2030 | btrfs_ino(inode), file_pos, &ins); | 2030 | btrfs_ino(inode), file_pos, &ins); |
| 2031 | out: | 2031 | out: |
| 2032 | btrfs_free_path(path); | 2032 | btrfs_free_path(path); |
| 2033 | 2033 | ||
| 2034 | return ret; | 2034 | return ret; |
| 2035 | } | 2035 | } |
| 2036 | 2036 | ||
| 2037 | /* snapshot-aware defrag */ | 2037 | /* snapshot-aware defrag */ |
| 2038 | struct sa_defrag_extent_backref { | 2038 | struct sa_defrag_extent_backref { |
| 2039 | struct rb_node node; | 2039 | struct rb_node node; |
| 2040 | struct old_sa_defrag_extent *old; | 2040 | struct old_sa_defrag_extent *old; |
| 2041 | u64 root_id; | 2041 | u64 root_id; |
| 2042 | u64 inum; | 2042 | u64 inum; |
| 2043 | u64 file_pos; | 2043 | u64 file_pos; |
| 2044 | u64 extent_offset; | 2044 | u64 extent_offset; |
| 2045 | u64 num_bytes; | 2045 | u64 num_bytes; |
| 2046 | u64 generation; | 2046 | u64 generation; |
| 2047 | }; | 2047 | }; |
| 2048 | 2048 | ||
| 2049 | struct old_sa_defrag_extent { | 2049 | struct old_sa_defrag_extent { |
| 2050 | struct list_head list; | 2050 | struct list_head list; |
| 2051 | struct new_sa_defrag_extent *new; | 2051 | struct new_sa_defrag_extent *new; |
| 2052 | 2052 | ||
| 2053 | u64 extent_offset; | 2053 | u64 extent_offset; |
| 2054 | u64 bytenr; | 2054 | u64 bytenr; |
| 2055 | u64 offset; | 2055 | u64 offset; |
| 2056 | u64 len; | 2056 | u64 len; |
| 2057 | int count; | 2057 | int count; |
| 2058 | }; | 2058 | }; |
| 2059 | 2059 | ||
| 2060 | struct new_sa_defrag_extent { | 2060 | struct new_sa_defrag_extent { |
| 2061 | struct rb_root root; | 2061 | struct rb_root root; |
| 2062 | struct list_head head; | 2062 | struct list_head head; |
| 2063 | struct btrfs_path *path; | 2063 | struct btrfs_path *path; |
| 2064 | struct inode *inode; | 2064 | struct inode *inode; |
| 2065 | u64 file_pos; | 2065 | u64 file_pos; |
| 2066 | u64 len; | 2066 | u64 len; |
| 2067 | u64 bytenr; | 2067 | u64 bytenr; |
| 2068 | u64 disk_len; | 2068 | u64 disk_len; |
| 2069 | u8 compress_type; | 2069 | u8 compress_type; |
| 2070 | }; | 2070 | }; |
| 2071 | 2071 | ||
| 2072 | static int backref_comp(struct sa_defrag_extent_backref *b1, | 2072 | static int backref_comp(struct sa_defrag_extent_backref *b1, |
| 2073 | struct sa_defrag_extent_backref *b2) | 2073 | struct sa_defrag_extent_backref *b2) |
| 2074 | { | 2074 | { |
| 2075 | if (b1->root_id < b2->root_id) | 2075 | if (b1->root_id < b2->root_id) |
| 2076 | return -1; | 2076 | return -1; |
| 2077 | else if (b1->root_id > b2->root_id) | 2077 | else if (b1->root_id > b2->root_id) |
| 2078 | return 1; | 2078 | return 1; |
| 2079 | 2079 | ||
| 2080 | if (b1->inum < b2->inum) | 2080 | if (b1->inum < b2->inum) |
| 2081 | return -1; | 2081 | return -1; |
| 2082 | else if (b1->inum > b2->inum) | 2082 | else if (b1->inum > b2->inum) |
| 2083 | return 1; | 2083 | return 1; |
| 2084 | 2084 | ||
| 2085 | if (b1->file_pos < b2->file_pos) | 2085 | if (b1->file_pos < b2->file_pos) |
| 2086 | return -1; | 2086 | return -1; |
| 2087 | else if (b1->file_pos > b2->file_pos) | 2087 | else if (b1->file_pos > b2->file_pos) |
| 2088 | return 1; | 2088 | return 1; |
| 2089 | 2089 | ||
| 2090 | /* | 2090 | /* |
| 2091 | * [------------------------------] ===> (a range of space) | 2091 | * [------------------------------] ===> (a range of space) |
| 2092 | * |<--->| |<---->| =============> (fs/file tree A) | 2092 | * |<--->| |<---->| =============> (fs/file tree A) |
| 2093 | * |<---------------------------->| ===> (fs/file tree B) | 2093 | * |<---------------------------->| ===> (fs/file tree B) |
| 2094 | * | 2094 | * |
| 2095 | * A range of space can refer to two file extents in one tree while | 2095 | * A range of space can refer to two file extents in one tree while |
| 2096 | * refer to only one file extent in another tree. | 2096 | * refer to only one file extent in another tree. |
| 2097 | * | 2097 | * |
| 2098 | * So we may process a disk offset more than one time(two extents in A) | 2098 | * So we may process a disk offset more than one time(two extents in A) |
| 2099 | * and locate at the same extent(one extent in B), then insert two same | 2099 | * and locate at the same extent(one extent in B), then insert two same |
| 2100 | * backrefs(both refer to the extent in B). | 2100 | * backrefs(both refer to the extent in B). |
| 2101 | */ | 2101 | */ |
| 2102 | return 0; | 2102 | return 0; |
| 2103 | } | 2103 | } |
| 2104 | 2104 | ||
| 2105 | static void backref_insert(struct rb_root *root, | 2105 | static void backref_insert(struct rb_root *root, |
| 2106 | struct sa_defrag_extent_backref *backref) | 2106 | struct sa_defrag_extent_backref *backref) |
| 2107 | { | 2107 | { |
| 2108 | struct rb_node **p = &root->rb_node; | 2108 | struct rb_node **p = &root->rb_node; |
| 2109 | struct rb_node *parent = NULL; | 2109 | struct rb_node *parent = NULL; |
| 2110 | struct sa_defrag_extent_backref *entry; | 2110 | struct sa_defrag_extent_backref *entry; |
| 2111 | int ret; | 2111 | int ret; |
| 2112 | 2112 | ||
| 2113 | while (*p) { | 2113 | while (*p) { |
| 2114 | parent = *p; | 2114 | parent = *p; |
| 2115 | entry = rb_entry(parent, struct sa_defrag_extent_backref, node); | 2115 | entry = rb_entry(parent, struct sa_defrag_extent_backref, node); |
| 2116 | 2116 | ||
| 2117 | ret = backref_comp(backref, entry); | 2117 | ret = backref_comp(backref, entry); |
| 2118 | if (ret < 0) | 2118 | if (ret < 0) |
| 2119 | p = &(*p)->rb_left; | 2119 | p = &(*p)->rb_left; |
| 2120 | else | 2120 | else |
| 2121 | p = &(*p)->rb_right; | 2121 | p = &(*p)->rb_right; |
| 2122 | } | 2122 | } |
| 2123 | 2123 | ||
| 2124 | rb_link_node(&backref->node, parent, p); | 2124 | rb_link_node(&backref->node, parent, p); |
| 2125 | rb_insert_color(&backref->node, root); | 2125 | rb_insert_color(&backref->node, root); |
| 2126 | } | 2126 | } |
| 2127 | 2127 | ||
| 2128 | /* | 2128 | /* |
| 2129 | * Note the backref might has changed, and in this case we just return 0. | 2129 | * Note the backref might has changed, and in this case we just return 0. |
| 2130 | */ | 2130 | */ |
| 2131 | static noinline int record_one_backref(u64 inum, u64 offset, u64 root_id, | 2131 | static noinline int record_one_backref(u64 inum, u64 offset, u64 root_id, |
| 2132 | void *ctx) | 2132 | void *ctx) |
| 2133 | { | 2133 | { |
| 2134 | struct btrfs_file_extent_item *extent; | 2134 | struct btrfs_file_extent_item *extent; |
| 2135 | struct btrfs_fs_info *fs_info; | 2135 | struct btrfs_fs_info *fs_info; |
| 2136 | struct old_sa_defrag_extent *old = ctx; | 2136 | struct old_sa_defrag_extent *old = ctx; |
| 2137 | struct new_sa_defrag_extent *new = old->new; | 2137 | struct new_sa_defrag_extent *new = old->new; |
| 2138 | struct btrfs_path *path = new->path; | 2138 | struct btrfs_path *path = new->path; |
| 2139 | struct btrfs_key key; | 2139 | struct btrfs_key key; |
| 2140 | struct btrfs_root *root; | 2140 | struct btrfs_root *root; |
| 2141 | struct sa_defrag_extent_backref *backref; | 2141 | struct sa_defrag_extent_backref *backref; |
| 2142 | struct extent_buffer *leaf; | 2142 | struct extent_buffer *leaf; |
| 2143 | struct inode *inode = new->inode; | 2143 | struct inode *inode = new->inode; |
| 2144 | int slot; | 2144 | int slot; |
| 2145 | int ret; | 2145 | int ret; |
| 2146 | u64 extent_offset; | 2146 | u64 extent_offset; |
| 2147 | u64 num_bytes; | 2147 | u64 num_bytes; |
| 2148 | 2148 | ||
| 2149 | if (BTRFS_I(inode)->root->root_key.objectid == root_id && | 2149 | if (BTRFS_I(inode)->root->root_key.objectid == root_id && |
| 2150 | inum == btrfs_ino(inode)) | 2150 | inum == btrfs_ino(inode)) |
| 2151 | return 0; | 2151 | return 0; |
| 2152 | 2152 | ||
| 2153 | key.objectid = root_id; | 2153 | key.objectid = root_id; |
| 2154 | key.type = BTRFS_ROOT_ITEM_KEY; | 2154 | key.type = BTRFS_ROOT_ITEM_KEY; |
| 2155 | key.offset = (u64)-1; | 2155 | key.offset = (u64)-1; |
| 2156 | 2156 | ||
| 2157 | fs_info = BTRFS_I(inode)->root->fs_info; | 2157 | fs_info = BTRFS_I(inode)->root->fs_info; |
| 2158 | root = btrfs_read_fs_root_no_name(fs_info, &key); | 2158 | root = btrfs_read_fs_root_no_name(fs_info, &key); |
| 2159 | if (IS_ERR(root)) { | 2159 | if (IS_ERR(root)) { |
| 2160 | if (PTR_ERR(root) == -ENOENT) | 2160 | if (PTR_ERR(root) == -ENOENT) |
| 2161 | return 0; | 2161 | return 0; |
| 2162 | WARN_ON(1); | 2162 | WARN_ON(1); |
| 2163 | pr_debug("inum=%llu, offset=%llu, root_id=%llu\n", | 2163 | pr_debug("inum=%llu, offset=%llu, root_id=%llu\n", |
| 2164 | inum, offset, root_id); | 2164 | inum, offset, root_id); |
| 2165 | return PTR_ERR(root); | 2165 | return PTR_ERR(root); |
| 2166 | } | 2166 | } |
| 2167 | 2167 | ||
| 2168 | key.objectid = inum; | 2168 | key.objectid = inum; |
| 2169 | key.type = BTRFS_EXTENT_DATA_KEY; | 2169 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 2170 | if (offset > (u64)-1 << 32) | 2170 | if (offset > (u64)-1 << 32) |
| 2171 | key.offset = 0; | 2171 | key.offset = 0; |
| 2172 | else | 2172 | else |
| 2173 | key.offset = offset; | 2173 | key.offset = offset; |
| 2174 | 2174 | ||
| 2175 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 2175 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 2176 | if (WARN_ON(ret < 0)) | 2176 | if (WARN_ON(ret < 0)) |
| 2177 | return ret; | 2177 | return ret; |
| 2178 | ret = 0; | 2178 | ret = 0; |
| 2179 | 2179 | ||
| 2180 | while (1) { | 2180 | while (1) { |
| 2181 | cond_resched(); | 2181 | cond_resched(); |
| 2182 | 2182 | ||
| 2183 | leaf = path->nodes[0]; | 2183 | leaf = path->nodes[0]; |
| 2184 | slot = path->slots[0]; | 2184 | slot = path->slots[0]; |
| 2185 | 2185 | ||
| 2186 | if (slot >= btrfs_header_nritems(leaf)) { | 2186 | if (slot >= btrfs_header_nritems(leaf)) { |
| 2187 | ret = btrfs_next_leaf(root, path); | 2187 | ret = btrfs_next_leaf(root, path); |
| 2188 | if (ret < 0) { | 2188 | if (ret < 0) { |
| 2189 | goto out; | 2189 | goto out; |
| 2190 | } else if (ret > 0) { | 2190 | } else if (ret > 0) { |
| 2191 | ret = 0; | 2191 | ret = 0; |
| 2192 | goto out; | 2192 | goto out; |
| 2193 | } | 2193 | } |
| 2194 | continue; | 2194 | continue; |
| 2195 | } | 2195 | } |
| 2196 | 2196 | ||
| 2197 | path->slots[0]++; | 2197 | path->slots[0]++; |
| 2198 | 2198 | ||
| 2199 | btrfs_item_key_to_cpu(leaf, &key, slot); | 2199 | btrfs_item_key_to_cpu(leaf, &key, slot); |
| 2200 | 2200 | ||
| 2201 | if (key.objectid > inum) | 2201 | if (key.objectid > inum) |
| 2202 | goto out; | 2202 | goto out; |
| 2203 | 2203 | ||
| 2204 | if (key.objectid < inum || key.type != BTRFS_EXTENT_DATA_KEY) | 2204 | if (key.objectid < inum || key.type != BTRFS_EXTENT_DATA_KEY) |
| 2205 | continue; | 2205 | continue; |
| 2206 | 2206 | ||
| 2207 | extent = btrfs_item_ptr(leaf, slot, | 2207 | extent = btrfs_item_ptr(leaf, slot, |
| 2208 | struct btrfs_file_extent_item); | 2208 | struct btrfs_file_extent_item); |
| 2209 | 2209 | ||
| 2210 | if (btrfs_file_extent_disk_bytenr(leaf, extent) != old->bytenr) | 2210 | if (btrfs_file_extent_disk_bytenr(leaf, extent) != old->bytenr) |
| 2211 | continue; | 2211 | continue; |
| 2212 | 2212 | ||
| 2213 | /* | 2213 | /* |
| 2214 | * 'offset' refers to the exact key.offset, | 2214 | * 'offset' refers to the exact key.offset, |
| 2215 | * NOT the 'offset' field in btrfs_extent_data_ref, ie. | 2215 | * NOT the 'offset' field in btrfs_extent_data_ref, ie. |
| 2216 | * (key.offset - extent_offset). | 2216 | * (key.offset - extent_offset). |
| 2217 | */ | 2217 | */ |
| 2218 | if (key.offset != offset) | 2218 | if (key.offset != offset) |
| 2219 | continue; | 2219 | continue; |
| 2220 | 2220 | ||
| 2221 | extent_offset = btrfs_file_extent_offset(leaf, extent); | 2221 | extent_offset = btrfs_file_extent_offset(leaf, extent); |
| 2222 | num_bytes = btrfs_file_extent_num_bytes(leaf, extent); | 2222 | num_bytes = btrfs_file_extent_num_bytes(leaf, extent); |
| 2223 | 2223 | ||
| 2224 | if (extent_offset >= old->extent_offset + old->offset + | 2224 | if (extent_offset >= old->extent_offset + old->offset + |
| 2225 | old->len || extent_offset + num_bytes <= | 2225 | old->len || extent_offset + num_bytes <= |
| 2226 | old->extent_offset + old->offset) | 2226 | old->extent_offset + old->offset) |
| 2227 | continue; | 2227 | continue; |
| 2228 | break; | 2228 | break; |
| 2229 | } | 2229 | } |
| 2230 | 2230 | ||
| 2231 | backref = kmalloc(sizeof(*backref), GFP_NOFS); | 2231 | backref = kmalloc(sizeof(*backref), GFP_NOFS); |
| 2232 | if (!backref) { | 2232 | if (!backref) { |
| 2233 | ret = -ENOENT; | 2233 | ret = -ENOENT; |
| 2234 | goto out; | 2234 | goto out; |
| 2235 | } | 2235 | } |
| 2236 | 2236 | ||
| 2237 | backref->root_id = root_id; | 2237 | backref->root_id = root_id; |
| 2238 | backref->inum = inum; | 2238 | backref->inum = inum; |
| 2239 | backref->file_pos = offset; | 2239 | backref->file_pos = offset; |
| 2240 | backref->num_bytes = num_bytes; | 2240 | backref->num_bytes = num_bytes; |
| 2241 | backref->extent_offset = extent_offset; | 2241 | backref->extent_offset = extent_offset; |
| 2242 | backref->generation = btrfs_file_extent_generation(leaf, extent); | 2242 | backref->generation = btrfs_file_extent_generation(leaf, extent); |
| 2243 | backref->old = old; | 2243 | backref->old = old; |
| 2244 | backref_insert(&new->root, backref); | 2244 | backref_insert(&new->root, backref); |
| 2245 | old->count++; | 2245 | old->count++; |
| 2246 | out: | 2246 | out: |
| 2247 | btrfs_release_path(path); | 2247 | btrfs_release_path(path); |
| 2248 | WARN_ON(ret); | 2248 | WARN_ON(ret); |
| 2249 | return ret; | 2249 | return ret; |
| 2250 | } | 2250 | } |
| 2251 | 2251 | ||
| 2252 | static noinline bool record_extent_backrefs(struct btrfs_path *path, | 2252 | static noinline bool record_extent_backrefs(struct btrfs_path *path, |
| 2253 | struct new_sa_defrag_extent *new) | 2253 | struct new_sa_defrag_extent *new) |
| 2254 | { | 2254 | { |
| 2255 | struct btrfs_fs_info *fs_info = BTRFS_I(new->inode)->root->fs_info; | 2255 | struct btrfs_fs_info *fs_info = BTRFS_I(new->inode)->root->fs_info; |
| 2256 | struct old_sa_defrag_extent *old, *tmp; | 2256 | struct old_sa_defrag_extent *old, *tmp; |
| 2257 | int ret; | 2257 | int ret; |
| 2258 | 2258 | ||
| 2259 | new->path = path; | 2259 | new->path = path; |
| 2260 | 2260 | ||
| 2261 | list_for_each_entry_safe(old, tmp, &new->head, list) { | 2261 | list_for_each_entry_safe(old, tmp, &new->head, list) { |
| 2262 | ret = iterate_inodes_from_logical(old->bytenr + | 2262 | ret = iterate_inodes_from_logical(old->bytenr + |
| 2263 | old->extent_offset, fs_info, | 2263 | old->extent_offset, fs_info, |
| 2264 | path, record_one_backref, | 2264 | path, record_one_backref, |
| 2265 | old); | 2265 | old); |
| 2266 | if (ret < 0 && ret != -ENOENT) | 2266 | if (ret < 0 && ret != -ENOENT) |
| 2267 | return false; | 2267 | return false; |
| 2268 | 2268 | ||
| 2269 | /* no backref to be processed for this extent */ | 2269 | /* no backref to be processed for this extent */ |
| 2270 | if (!old->count) { | 2270 | if (!old->count) { |
| 2271 | list_del(&old->list); | 2271 | list_del(&old->list); |
| 2272 | kfree(old); | 2272 | kfree(old); |
| 2273 | } | 2273 | } |
| 2274 | } | 2274 | } |
| 2275 | 2275 | ||
| 2276 | if (list_empty(&new->head)) | 2276 | if (list_empty(&new->head)) |
| 2277 | return false; | 2277 | return false; |
| 2278 | 2278 | ||
| 2279 | return true; | 2279 | return true; |
| 2280 | } | 2280 | } |
| 2281 | 2281 | ||
| 2282 | static int relink_is_mergable(struct extent_buffer *leaf, | 2282 | static int relink_is_mergable(struct extent_buffer *leaf, |
| 2283 | struct btrfs_file_extent_item *fi, | 2283 | struct btrfs_file_extent_item *fi, |
| 2284 | struct new_sa_defrag_extent *new) | 2284 | struct new_sa_defrag_extent *new) |
| 2285 | { | 2285 | { |
| 2286 | if (btrfs_file_extent_disk_bytenr(leaf, fi) != new->bytenr) | 2286 | if (btrfs_file_extent_disk_bytenr(leaf, fi) != new->bytenr) |
| 2287 | return 0; | 2287 | return 0; |
| 2288 | 2288 | ||
| 2289 | if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG) | 2289 | if (btrfs_file_extent_type(leaf, fi) != BTRFS_FILE_EXTENT_REG) |
| 2290 | return 0; | 2290 | return 0; |
| 2291 | 2291 | ||
| 2292 | if (btrfs_file_extent_compression(leaf, fi) != new->compress_type) | 2292 | if (btrfs_file_extent_compression(leaf, fi) != new->compress_type) |
| 2293 | return 0; | 2293 | return 0; |
| 2294 | 2294 | ||
| 2295 | if (btrfs_file_extent_encryption(leaf, fi) || | 2295 | if (btrfs_file_extent_encryption(leaf, fi) || |
| 2296 | btrfs_file_extent_other_encoding(leaf, fi)) | 2296 | btrfs_file_extent_other_encoding(leaf, fi)) |
| 2297 | return 0; | 2297 | return 0; |
| 2298 | 2298 | ||
| 2299 | return 1; | 2299 | return 1; |
| 2300 | } | 2300 | } |
| 2301 | 2301 | ||
| 2302 | /* | 2302 | /* |
| 2303 | * Note the backref might has changed, and in this case we just return 0. | 2303 | * Note the backref might has changed, and in this case we just return 0. |
| 2304 | */ | 2304 | */ |
| 2305 | static noinline int relink_extent_backref(struct btrfs_path *path, | 2305 | static noinline int relink_extent_backref(struct btrfs_path *path, |
| 2306 | struct sa_defrag_extent_backref *prev, | 2306 | struct sa_defrag_extent_backref *prev, |
| 2307 | struct sa_defrag_extent_backref *backref) | 2307 | struct sa_defrag_extent_backref *backref) |
| 2308 | { | 2308 | { |
| 2309 | struct btrfs_file_extent_item *extent; | 2309 | struct btrfs_file_extent_item *extent; |
| 2310 | struct btrfs_file_extent_item *item; | 2310 | struct btrfs_file_extent_item *item; |
| 2311 | struct btrfs_ordered_extent *ordered; | 2311 | struct btrfs_ordered_extent *ordered; |
| 2312 | struct btrfs_trans_handle *trans; | 2312 | struct btrfs_trans_handle *trans; |
| 2313 | struct btrfs_fs_info *fs_info; | 2313 | struct btrfs_fs_info *fs_info; |
| 2314 | struct btrfs_root *root; | 2314 | struct btrfs_root *root; |
| 2315 | struct btrfs_key key; | 2315 | struct btrfs_key key; |
| 2316 | struct extent_buffer *leaf; | 2316 | struct extent_buffer *leaf; |
| 2317 | struct old_sa_defrag_extent *old = backref->old; | 2317 | struct old_sa_defrag_extent *old = backref->old; |
| 2318 | struct new_sa_defrag_extent *new = old->new; | 2318 | struct new_sa_defrag_extent *new = old->new; |
| 2319 | struct inode *src_inode = new->inode; | 2319 | struct inode *src_inode = new->inode; |
| 2320 | struct inode *inode; | 2320 | struct inode *inode; |
| 2321 | struct extent_state *cached = NULL; | 2321 | struct extent_state *cached = NULL; |
| 2322 | int ret = 0; | 2322 | int ret = 0; |
| 2323 | u64 start; | 2323 | u64 start; |
| 2324 | u64 len; | 2324 | u64 len; |
| 2325 | u64 lock_start; | 2325 | u64 lock_start; |
| 2326 | u64 lock_end; | 2326 | u64 lock_end; |
| 2327 | bool merge = false; | 2327 | bool merge = false; |
| 2328 | int index; | 2328 | int index; |
| 2329 | 2329 | ||
| 2330 | if (prev && prev->root_id == backref->root_id && | 2330 | if (prev && prev->root_id == backref->root_id && |
| 2331 | prev->inum == backref->inum && | 2331 | prev->inum == backref->inum && |
| 2332 | prev->file_pos + prev->num_bytes == backref->file_pos) | 2332 | prev->file_pos + prev->num_bytes == backref->file_pos) |
| 2333 | merge = true; | 2333 | merge = true; |
| 2334 | 2334 | ||
| 2335 | /* step 1: get root */ | 2335 | /* step 1: get root */ |
| 2336 | key.objectid = backref->root_id; | 2336 | key.objectid = backref->root_id; |
| 2337 | key.type = BTRFS_ROOT_ITEM_KEY; | 2337 | key.type = BTRFS_ROOT_ITEM_KEY; |
| 2338 | key.offset = (u64)-1; | 2338 | key.offset = (u64)-1; |
| 2339 | 2339 | ||
| 2340 | fs_info = BTRFS_I(src_inode)->root->fs_info; | 2340 | fs_info = BTRFS_I(src_inode)->root->fs_info; |
| 2341 | index = srcu_read_lock(&fs_info->subvol_srcu); | 2341 | index = srcu_read_lock(&fs_info->subvol_srcu); |
| 2342 | 2342 | ||
| 2343 | root = btrfs_read_fs_root_no_name(fs_info, &key); | 2343 | root = btrfs_read_fs_root_no_name(fs_info, &key); |
| 2344 | if (IS_ERR(root)) { | 2344 | if (IS_ERR(root)) { |
| 2345 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 2345 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 2346 | if (PTR_ERR(root) == -ENOENT) | 2346 | if (PTR_ERR(root) == -ENOENT) |
| 2347 | return 0; | 2347 | return 0; |
| 2348 | return PTR_ERR(root); | 2348 | return PTR_ERR(root); |
| 2349 | } | 2349 | } |
| 2350 | 2350 | ||
| 2351 | if (btrfs_root_readonly(root)) { | 2351 | if (btrfs_root_readonly(root)) { |
| 2352 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 2352 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 2353 | return 0; | 2353 | return 0; |
| 2354 | } | 2354 | } |
| 2355 | 2355 | ||
| 2356 | /* step 2: get inode */ | 2356 | /* step 2: get inode */ |
| 2357 | key.objectid = backref->inum; | 2357 | key.objectid = backref->inum; |
| 2358 | key.type = BTRFS_INODE_ITEM_KEY; | 2358 | key.type = BTRFS_INODE_ITEM_KEY; |
| 2359 | key.offset = 0; | 2359 | key.offset = 0; |
| 2360 | 2360 | ||
| 2361 | inode = btrfs_iget(fs_info->sb, &key, root, NULL); | 2361 | inode = btrfs_iget(fs_info->sb, &key, root, NULL); |
| 2362 | if (IS_ERR(inode)) { | 2362 | if (IS_ERR(inode)) { |
| 2363 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 2363 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 2364 | return 0; | 2364 | return 0; |
| 2365 | } | 2365 | } |
| 2366 | 2366 | ||
| 2367 | srcu_read_unlock(&fs_info->subvol_srcu, index); | 2367 | srcu_read_unlock(&fs_info->subvol_srcu, index); |
| 2368 | 2368 | ||
| 2369 | /* step 3: relink backref */ | 2369 | /* step 3: relink backref */ |
| 2370 | lock_start = backref->file_pos; | 2370 | lock_start = backref->file_pos; |
| 2371 | lock_end = backref->file_pos + backref->num_bytes - 1; | 2371 | lock_end = backref->file_pos + backref->num_bytes - 1; |
| 2372 | lock_extent_bits(&BTRFS_I(inode)->io_tree, lock_start, lock_end, | 2372 | lock_extent_bits(&BTRFS_I(inode)->io_tree, lock_start, lock_end, |
| 2373 | 0, &cached); | 2373 | 0, &cached); |
| 2374 | 2374 | ||
| 2375 | ordered = btrfs_lookup_first_ordered_extent(inode, lock_end); | 2375 | ordered = btrfs_lookup_first_ordered_extent(inode, lock_end); |
| 2376 | if (ordered) { | 2376 | if (ordered) { |
| 2377 | btrfs_put_ordered_extent(ordered); | 2377 | btrfs_put_ordered_extent(ordered); |
| 2378 | goto out_unlock; | 2378 | goto out_unlock; |
| 2379 | } | 2379 | } |
| 2380 | 2380 | ||
| 2381 | trans = btrfs_join_transaction(root); | 2381 | trans = btrfs_join_transaction(root); |
| 2382 | if (IS_ERR(trans)) { | 2382 | if (IS_ERR(trans)) { |
| 2383 | ret = PTR_ERR(trans); | 2383 | ret = PTR_ERR(trans); |
| 2384 | goto out_unlock; | 2384 | goto out_unlock; |
| 2385 | } | 2385 | } |
| 2386 | 2386 | ||
| 2387 | key.objectid = backref->inum; | 2387 | key.objectid = backref->inum; |
| 2388 | key.type = BTRFS_EXTENT_DATA_KEY; | 2388 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 2389 | key.offset = backref->file_pos; | 2389 | key.offset = backref->file_pos; |
| 2390 | 2390 | ||
| 2391 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 2391 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 2392 | if (ret < 0) { | 2392 | if (ret < 0) { |
| 2393 | goto out_free_path; | 2393 | goto out_free_path; |
| 2394 | } else if (ret > 0) { | 2394 | } else if (ret > 0) { |
| 2395 | ret = 0; | 2395 | ret = 0; |
| 2396 | goto out_free_path; | 2396 | goto out_free_path; |
| 2397 | } | 2397 | } |
| 2398 | 2398 | ||
| 2399 | extent = btrfs_item_ptr(path->nodes[0], path->slots[0], | 2399 | extent = btrfs_item_ptr(path->nodes[0], path->slots[0], |
| 2400 | struct btrfs_file_extent_item); | 2400 | struct btrfs_file_extent_item); |
| 2401 | 2401 | ||
| 2402 | if (btrfs_file_extent_generation(path->nodes[0], extent) != | 2402 | if (btrfs_file_extent_generation(path->nodes[0], extent) != |
| 2403 | backref->generation) | 2403 | backref->generation) |
| 2404 | goto out_free_path; | 2404 | goto out_free_path; |
| 2405 | 2405 | ||
| 2406 | btrfs_release_path(path); | 2406 | btrfs_release_path(path); |
| 2407 | 2407 | ||
| 2408 | start = backref->file_pos; | 2408 | start = backref->file_pos; |
| 2409 | if (backref->extent_offset < old->extent_offset + old->offset) | 2409 | if (backref->extent_offset < old->extent_offset + old->offset) |
| 2410 | start += old->extent_offset + old->offset - | 2410 | start += old->extent_offset + old->offset - |
| 2411 | backref->extent_offset; | 2411 | backref->extent_offset; |
| 2412 | 2412 | ||
| 2413 | len = min(backref->extent_offset + backref->num_bytes, | 2413 | len = min(backref->extent_offset + backref->num_bytes, |
| 2414 | old->extent_offset + old->offset + old->len); | 2414 | old->extent_offset + old->offset + old->len); |
| 2415 | len -= max(backref->extent_offset, old->extent_offset + old->offset); | 2415 | len -= max(backref->extent_offset, old->extent_offset + old->offset); |
| 2416 | 2416 | ||
| 2417 | ret = btrfs_drop_extents(trans, root, inode, start, | 2417 | ret = btrfs_drop_extents(trans, root, inode, start, |
| 2418 | start + len, 1); | 2418 | start + len, 1); |
| 2419 | if (ret) | 2419 | if (ret) |
| 2420 | goto out_free_path; | 2420 | goto out_free_path; |
| 2421 | again: | 2421 | again: |
| 2422 | key.objectid = btrfs_ino(inode); | 2422 | key.objectid = btrfs_ino(inode); |
| 2423 | key.type = BTRFS_EXTENT_DATA_KEY; | 2423 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 2424 | key.offset = start; | 2424 | key.offset = start; |
| 2425 | 2425 | ||
| 2426 | path->leave_spinning = 1; | 2426 | path->leave_spinning = 1; |
| 2427 | if (merge) { | 2427 | if (merge) { |
| 2428 | struct btrfs_file_extent_item *fi; | 2428 | struct btrfs_file_extent_item *fi; |
| 2429 | u64 extent_len; | 2429 | u64 extent_len; |
| 2430 | struct btrfs_key found_key; | 2430 | struct btrfs_key found_key; |
| 2431 | 2431 | ||
| 2432 | ret = btrfs_search_slot(trans, root, &key, path, 0, 1); | 2432 | ret = btrfs_search_slot(trans, root, &key, path, 0, 1); |
| 2433 | if (ret < 0) | 2433 | if (ret < 0) |
| 2434 | goto out_free_path; | 2434 | goto out_free_path; |
| 2435 | 2435 | ||
| 2436 | path->slots[0]--; | 2436 | path->slots[0]--; |
| 2437 | leaf = path->nodes[0]; | 2437 | leaf = path->nodes[0]; |
| 2438 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 2438 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 2439 | 2439 | ||
| 2440 | fi = btrfs_item_ptr(leaf, path->slots[0], | 2440 | fi = btrfs_item_ptr(leaf, path->slots[0], |
| 2441 | struct btrfs_file_extent_item); | 2441 | struct btrfs_file_extent_item); |
| 2442 | extent_len = btrfs_file_extent_num_bytes(leaf, fi); | 2442 | extent_len = btrfs_file_extent_num_bytes(leaf, fi); |
| 2443 | 2443 | ||
| 2444 | if (extent_len + found_key.offset == start && | 2444 | if (extent_len + found_key.offset == start && |
| 2445 | relink_is_mergable(leaf, fi, new)) { | 2445 | relink_is_mergable(leaf, fi, new)) { |
| 2446 | btrfs_set_file_extent_num_bytes(leaf, fi, | 2446 | btrfs_set_file_extent_num_bytes(leaf, fi, |
| 2447 | extent_len + len); | 2447 | extent_len + len); |
| 2448 | btrfs_mark_buffer_dirty(leaf); | 2448 | btrfs_mark_buffer_dirty(leaf); |
| 2449 | inode_add_bytes(inode, len); | 2449 | inode_add_bytes(inode, len); |
| 2450 | 2450 | ||
| 2451 | ret = 1; | 2451 | ret = 1; |
| 2452 | goto out_free_path; | 2452 | goto out_free_path; |
| 2453 | } else { | 2453 | } else { |
| 2454 | merge = false; | 2454 | merge = false; |
| 2455 | btrfs_release_path(path); | 2455 | btrfs_release_path(path); |
| 2456 | goto again; | 2456 | goto again; |
| 2457 | } | 2457 | } |
| 2458 | } | 2458 | } |
| 2459 | 2459 | ||
| 2460 | ret = btrfs_insert_empty_item(trans, root, path, &key, | 2460 | ret = btrfs_insert_empty_item(trans, root, path, &key, |
| 2461 | sizeof(*extent)); | 2461 | sizeof(*extent)); |
| 2462 | if (ret) { | 2462 | if (ret) { |
| 2463 | btrfs_abort_transaction(trans, root, ret); | 2463 | btrfs_abort_transaction(trans, root, ret); |
| 2464 | goto out_free_path; | 2464 | goto out_free_path; |
| 2465 | } | 2465 | } |
| 2466 | 2466 | ||
| 2467 | leaf = path->nodes[0]; | 2467 | leaf = path->nodes[0]; |
| 2468 | item = btrfs_item_ptr(leaf, path->slots[0], | 2468 | item = btrfs_item_ptr(leaf, path->slots[0], |
| 2469 | struct btrfs_file_extent_item); | 2469 | struct btrfs_file_extent_item); |
| 2470 | btrfs_set_file_extent_disk_bytenr(leaf, item, new->bytenr); | 2470 | btrfs_set_file_extent_disk_bytenr(leaf, item, new->bytenr); |
| 2471 | btrfs_set_file_extent_disk_num_bytes(leaf, item, new->disk_len); | 2471 | btrfs_set_file_extent_disk_num_bytes(leaf, item, new->disk_len); |
| 2472 | btrfs_set_file_extent_offset(leaf, item, start - new->file_pos); | 2472 | btrfs_set_file_extent_offset(leaf, item, start - new->file_pos); |
| 2473 | btrfs_set_file_extent_num_bytes(leaf, item, len); | 2473 | btrfs_set_file_extent_num_bytes(leaf, item, len); |
| 2474 | btrfs_set_file_extent_ram_bytes(leaf, item, new->len); | 2474 | btrfs_set_file_extent_ram_bytes(leaf, item, new->len); |
| 2475 | btrfs_set_file_extent_generation(leaf, item, trans->transid); | 2475 | btrfs_set_file_extent_generation(leaf, item, trans->transid); |
| 2476 | btrfs_set_file_extent_type(leaf, item, BTRFS_FILE_EXTENT_REG); | 2476 | btrfs_set_file_extent_type(leaf, item, BTRFS_FILE_EXTENT_REG); |
| 2477 | btrfs_set_file_extent_compression(leaf, item, new->compress_type); | 2477 | btrfs_set_file_extent_compression(leaf, item, new->compress_type); |
| 2478 | btrfs_set_file_extent_encryption(leaf, item, 0); | 2478 | btrfs_set_file_extent_encryption(leaf, item, 0); |
| 2479 | btrfs_set_file_extent_other_encoding(leaf, item, 0); | 2479 | btrfs_set_file_extent_other_encoding(leaf, item, 0); |
| 2480 | 2480 | ||
| 2481 | btrfs_mark_buffer_dirty(leaf); | 2481 | btrfs_mark_buffer_dirty(leaf); |
| 2482 | inode_add_bytes(inode, len); | 2482 | inode_add_bytes(inode, len); |
| 2483 | btrfs_release_path(path); | 2483 | btrfs_release_path(path); |
| 2484 | 2484 | ||
| 2485 | ret = btrfs_inc_extent_ref(trans, root, new->bytenr, | 2485 | ret = btrfs_inc_extent_ref(trans, root, new->bytenr, |
| 2486 | new->disk_len, 0, | 2486 | new->disk_len, 0, |
| 2487 | backref->root_id, backref->inum, | 2487 | backref->root_id, backref->inum, |
| 2488 | new->file_pos, 0); /* start - extent_offset */ | 2488 | new->file_pos, 0); /* start - extent_offset */ |
| 2489 | if (ret) { | 2489 | if (ret) { |
| 2490 | btrfs_abort_transaction(trans, root, ret); | 2490 | btrfs_abort_transaction(trans, root, ret); |
| 2491 | goto out_free_path; | 2491 | goto out_free_path; |
| 2492 | } | 2492 | } |
| 2493 | 2493 | ||
| 2494 | ret = 1; | 2494 | ret = 1; |
| 2495 | out_free_path: | 2495 | out_free_path: |
| 2496 | btrfs_release_path(path); | 2496 | btrfs_release_path(path); |
| 2497 | path->leave_spinning = 0; | 2497 | path->leave_spinning = 0; |
| 2498 | btrfs_end_transaction(trans, root); | 2498 | btrfs_end_transaction(trans, root); |
| 2499 | out_unlock: | 2499 | out_unlock: |
| 2500 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, lock_start, lock_end, | 2500 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, lock_start, lock_end, |
| 2501 | &cached, GFP_NOFS); | 2501 | &cached, GFP_NOFS); |
| 2502 | iput(inode); | 2502 | iput(inode); |
| 2503 | return ret; | 2503 | return ret; |
| 2504 | } | 2504 | } |
| 2505 | 2505 | ||
| 2506 | static void free_sa_defrag_extent(struct new_sa_defrag_extent *new) | 2506 | static void free_sa_defrag_extent(struct new_sa_defrag_extent *new) |
| 2507 | { | 2507 | { |
| 2508 | struct old_sa_defrag_extent *old, *tmp; | 2508 | struct old_sa_defrag_extent *old, *tmp; |
| 2509 | 2509 | ||
| 2510 | if (!new) | 2510 | if (!new) |
| 2511 | return; | 2511 | return; |
| 2512 | 2512 | ||
| 2513 | list_for_each_entry_safe(old, tmp, &new->head, list) { | 2513 | list_for_each_entry_safe(old, tmp, &new->head, list) { |
| 2514 | list_del(&old->list); | 2514 | list_del(&old->list); |
| 2515 | kfree(old); | 2515 | kfree(old); |
| 2516 | } | 2516 | } |
| 2517 | kfree(new); | 2517 | kfree(new); |
| 2518 | } | 2518 | } |
| 2519 | 2519 | ||
| 2520 | static void relink_file_extents(struct new_sa_defrag_extent *new) | 2520 | static void relink_file_extents(struct new_sa_defrag_extent *new) |
| 2521 | { | 2521 | { |
| 2522 | struct btrfs_path *path; | 2522 | struct btrfs_path *path; |
| 2523 | struct sa_defrag_extent_backref *backref; | 2523 | struct sa_defrag_extent_backref *backref; |
| 2524 | struct sa_defrag_extent_backref *prev = NULL; | 2524 | struct sa_defrag_extent_backref *prev = NULL; |
| 2525 | struct inode *inode; | 2525 | struct inode *inode; |
| 2526 | struct btrfs_root *root; | 2526 | struct btrfs_root *root; |
| 2527 | struct rb_node *node; | 2527 | struct rb_node *node; |
| 2528 | int ret; | 2528 | int ret; |
| 2529 | 2529 | ||
| 2530 | inode = new->inode; | 2530 | inode = new->inode; |
| 2531 | root = BTRFS_I(inode)->root; | 2531 | root = BTRFS_I(inode)->root; |
| 2532 | 2532 | ||
| 2533 | path = btrfs_alloc_path(); | 2533 | path = btrfs_alloc_path(); |
| 2534 | if (!path) | 2534 | if (!path) |
| 2535 | return; | 2535 | return; |
| 2536 | 2536 | ||
| 2537 | if (!record_extent_backrefs(path, new)) { | 2537 | if (!record_extent_backrefs(path, new)) { |
| 2538 | btrfs_free_path(path); | 2538 | btrfs_free_path(path); |
| 2539 | goto out; | 2539 | goto out; |
| 2540 | } | 2540 | } |
| 2541 | btrfs_release_path(path); | 2541 | btrfs_release_path(path); |
| 2542 | 2542 | ||
| 2543 | while (1) { | 2543 | while (1) { |
| 2544 | node = rb_first(&new->root); | 2544 | node = rb_first(&new->root); |
| 2545 | if (!node) | 2545 | if (!node) |
| 2546 | break; | 2546 | break; |
| 2547 | rb_erase(node, &new->root); | 2547 | rb_erase(node, &new->root); |
| 2548 | 2548 | ||
| 2549 | backref = rb_entry(node, struct sa_defrag_extent_backref, node); | 2549 | backref = rb_entry(node, struct sa_defrag_extent_backref, node); |
| 2550 | 2550 | ||
| 2551 | ret = relink_extent_backref(path, prev, backref); | 2551 | ret = relink_extent_backref(path, prev, backref); |
| 2552 | WARN_ON(ret < 0); | 2552 | WARN_ON(ret < 0); |
| 2553 | 2553 | ||
| 2554 | kfree(prev); | 2554 | kfree(prev); |
| 2555 | 2555 | ||
| 2556 | if (ret == 1) | 2556 | if (ret == 1) |
| 2557 | prev = backref; | 2557 | prev = backref; |
| 2558 | else | 2558 | else |
| 2559 | prev = NULL; | 2559 | prev = NULL; |
| 2560 | cond_resched(); | 2560 | cond_resched(); |
| 2561 | } | 2561 | } |
| 2562 | kfree(prev); | 2562 | kfree(prev); |
| 2563 | 2563 | ||
| 2564 | btrfs_free_path(path); | 2564 | btrfs_free_path(path); |
| 2565 | out: | 2565 | out: |
| 2566 | free_sa_defrag_extent(new); | 2566 | free_sa_defrag_extent(new); |
| 2567 | 2567 | ||
| 2568 | atomic_dec(&root->fs_info->defrag_running); | 2568 | atomic_dec(&root->fs_info->defrag_running); |
| 2569 | wake_up(&root->fs_info->transaction_wait); | 2569 | wake_up(&root->fs_info->transaction_wait); |
| 2570 | } | 2570 | } |
| 2571 | 2571 | ||
| 2572 | static struct new_sa_defrag_extent * | 2572 | static struct new_sa_defrag_extent * |
| 2573 | record_old_file_extents(struct inode *inode, | 2573 | record_old_file_extents(struct inode *inode, |
| 2574 | struct btrfs_ordered_extent *ordered) | 2574 | struct btrfs_ordered_extent *ordered) |
| 2575 | { | 2575 | { |
| 2576 | struct btrfs_root *root = BTRFS_I(inode)->root; | 2576 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 2577 | struct btrfs_path *path; | 2577 | struct btrfs_path *path; |
| 2578 | struct btrfs_key key; | 2578 | struct btrfs_key key; |
| 2579 | struct old_sa_defrag_extent *old; | 2579 | struct old_sa_defrag_extent *old; |
| 2580 | struct new_sa_defrag_extent *new; | 2580 | struct new_sa_defrag_extent *new; |
| 2581 | int ret; | 2581 | int ret; |
| 2582 | 2582 | ||
| 2583 | new = kmalloc(sizeof(*new), GFP_NOFS); | 2583 | new = kmalloc(sizeof(*new), GFP_NOFS); |
| 2584 | if (!new) | 2584 | if (!new) |
| 2585 | return NULL; | 2585 | return NULL; |
| 2586 | 2586 | ||
| 2587 | new->inode = inode; | 2587 | new->inode = inode; |
| 2588 | new->file_pos = ordered->file_offset; | 2588 | new->file_pos = ordered->file_offset; |
| 2589 | new->len = ordered->len; | 2589 | new->len = ordered->len; |
| 2590 | new->bytenr = ordered->start; | 2590 | new->bytenr = ordered->start; |
| 2591 | new->disk_len = ordered->disk_len; | 2591 | new->disk_len = ordered->disk_len; |
| 2592 | new->compress_type = ordered->compress_type; | 2592 | new->compress_type = ordered->compress_type; |
| 2593 | new->root = RB_ROOT; | 2593 | new->root = RB_ROOT; |
| 2594 | INIT_LIST_HEAD(&new->head); | 2594 | INIT_LIST_HEAD(&new->head); |
| 2595 | 2595 | ||
| 2596 | path = btrfs_alloc_path(); | 2596 | path = btrfs_alloc_path(); |
| 2597 | if (!path) | 2597 | if (!path) |
| 2598 | goto out_kfree; | 2598 | goto out_kfree; |
| 2599 | 2599 | ||
| 2600 | key.objectid = btrfs_ino(inode); | 2600 | key.objectid = btrfs_ino(inode); |
| 2601 | key.type = BTRFS_EXTENT_DATA_KEY; | 2601 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 2602 | key.offset = new->file_pos; | 2602 | key.offset = new->file_pos; |
| 2603 | 2603 | ||
| 2604 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 2604 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 2605 | if (ret < 0) | 2605 | if (ret < 0) |
| 2606 | goto out_free_path; | 2606 | goto out_free_path; |
| 2607 | if (ret > 0 && path->slots[0] > 0) | 2607 | if (ret > 0 && path->slots[0] > 0) |
| 2608 | path->slots[0]--; | 2608 | path->slots[0]--; |
| 2609 | 2609 | ||
| 2610 | /* find out all the old extents for the file range */ | 2610 | /* find out all the old extents for the file range */ |
| 2611 | while (1) { | 2611 | while (1) { |
| 2612 | struct btrfs_file_extent_item *extent; | 2612 | struct btrfs_file_extent_item *extent; |
| 2613 | struct extent_buffer *l; | 2613 | struct extent_buffer *l; |
| 2614 | int slot; | 2614 | int slot; |
| 2615 | u64 num_bytes; | 2615 | u64 num_bytes; |
| 2616 | u64 offset; | 2616 | u64 offset; |
| 2617 | u64 end; | 2617 | u64 end; |
| 2618 | u64 disk_bytenr; | 2618 | u64 disk_bytenr; |
| 2619 | u64 extent_offset; | 2619 | u64 extent_offset; |
| 2620 | 2620 | ||
| 2621 | l = path->nodes[0]; | 2621 | l = path->nodes[0]; |
| 2622 | slot = path->slots[0]; | 2622 | slot = path->slots[0]; |
| 2623 | 2623 | ||
| 2624 | if (slot >= btrfs_header_nritems(l)) { | 2624 | if (slot >= btrfs_header_nritems(l)) { |
| 2625 | ret = btrfs_next_leaf(root, path); | 2625 | ret = btrfs_next_leaf(root, path); |
| 2626 | if (ret < 0) | 2626 | if (ret < 0) |
| 2627 | goto out_free_path; | 2627 | goto out_free_path; |
| 2628 | else if (ret > 0) | 2628 | else if (ret > 0) |
| 2629 | break; | 2629 | break; |
| 2630 | continue; | 2630 | continue; |
| 2631 | } | 2631 | } |
| 2632 | 2632 | ||
| 2633 | btrfs_item_key_to_cpu(l, &key, slot); | 2633 | btrfs_item_key_to_cpu(l, &key, slot); |
| 2634 | 2634 | ||
| 2635 | if (key.objectid != btrfs_ino(inode)) | 2635 | if (key.objectid != btrfs_ino(inode)) |
| 2636 | break; | 2636 | break; |
| 2637 | if (key.type != BTRFS_EXTENT_DATA_KEY) | 2637 | if (key.type != BTRFS_EXTENT_DATA_KEY) |
| 2638 | break; | 2638 | break; |
| 2639 | if (key.offset >= new->file_pos + new->len) | 2639 | if (key.offset >= new->file_pos + new->len) |
| 2640 | break; | 2640 | break; |
| 2641 | 2641 | ||
| 2642 | extent = btrfs_item_ptr(l, slot, struct btrfs_file_extent_item); | 2642 | extent = btrfs_item_ptr(l, slot, struct btrfs_file_extent_item); |
| 2643 | 2643 | ||
| 2644 | num_bytes = btrfs_file_extent_num_bytes(l, extent); | 2644 | num_bytes = btrfs_file_extent_num_bytes(l, extent); |
| 2645 | if (key.offset + num_bytes < new->file_pos) | 2645 | if (key.offset + num_bytes < new->file_pos) |
| 2646 | goto next; | 2646 | goto next; |
| 2647 | 2647 | ||
| 2648 | disk_bytenr = btrfs_file_extent_disk_bytenr(l, extent); | 2648 | disk_bytenr = btrfs_file_extent_disk_bytenr(l, extent); |
| 2649 | if (!disk_bytenr) | 2649 | if (!disk_bytenr) |
| 2650 | goto next; | 2650 | goto next; |
| 2651 | 2651 | ||
| 2652 | extent_offset = btrfs_file_extent_offset(l, extent); | 2652 | extent_offset = btrfs_file_extent_offset(l, extent); |
| 2653 | 2653 | ||
| 2654 | old = kmalloc(sizeof(*old), GFP_NOFS); | 2654 | old = kmalloc(sizeof(*old), GFP_NOFS); |
| 2655 | if (!old) | 2655 | if (!old) |
| 2656 | goto out_free_path; | 2656 | goto out_free_path; |
| 2657 | 2657 | ||
| 2658 | offset = max(new->file_pos, key.offset); | 2658 | offset = max(new->file_pos, key.offset); |
| 2659 | end = min(new->file_pos + new->len, key.offset + num_bytes); | 2659 | end = min(new->file_pos + new->len, key.offset + num_bytes); |
| 2660 | 2660 | ||
| 2661 | old->bytenr = disk_bytenr; | 2661 | old->bytenr = disk_bytenr; |
| 2662 | old->extent_offset = extent_offset; | 2662 | old->extent_offset = extent_offset; |
| 2663 | old->offset = offset - key.offset; | 2663 | old->offset = offset - key.offset; |
| 2664 | old->len = end - offset; | 2664 | old->len = end - offset; |
| 2665 | old->new = new; | 2665 | old->new = new; |
| 2666 | old->count = 0; | 2666 | old->count = 0; |
| 2667 | list_add_tail(&old->list, &new->head); | 2667 | list_add_tail(&old->list, &new->head); |
| 2668 | next: | 2668 | next: |
| 2669 | path->slots[0]++; | 2669 | path->slots[0]++; |
| 2670 | cond_resched(); | 2670 | cond_resched(); |
| 2671 | } | 2671 | } |
| 2672 | 2672 | ||
| 2673 | btrfs_free_path(path); | 2673 | btrfs_free_path(path); |
| 2674 | atomic_inc(&root->fs_info->defrag_running); | 2674 | atomic_inc(&root->fs_info->defrag_running); |
| 2675 | 2675 | ||
| 2676 | return new; | 2676 | return new; |
| 2677 | 2677 | ||
| 2678 | out_free_path: | 2678 | out_free_path: |
| 2679 | btrfs_free_path(path); | 2679 | btrfs_free_path(path); |
| 2680 | out_kfree: | 2680 | out_kfree: |
| 2681 | free_sa_defrag_extent(new); | 2681 | free_sa_defrag_extent(new); |
| 2682 | return NULL; | 2682 | return NULL; |
| 2683 | } | 2683 | } |
| 2684 | 2684 | ||
| 2685 | static void btrfs_release_delalloc_bytes(struct btrfs_root *root, | 2685 | static void btrfs_release_delalloc_bytes(struct btrfs_root *root, |
| 2686 | u64 start, u64 len) | 2686 | u64 start, u64 len) |
| 2687 | { | 2687 | { |
| 2688 | struct btrfs_block_group_cache *cache; | 2688 | struct btrfs_block_group_cache *cache; |
| 2689 | 2689 | ||
| 2690 | cache = btrfs_lookup_block_group(root->fs_info, start); | 2690 | cache = btrfs_lookup_block_group(root->fs_info, start); |
| 2691 | ASSERT(cache); | 2691 | ASSERT(cache); |
| 2692 | 2692 | ||
| 2693 | spin_lock(&cache->lock); | 2693 | spin_lock(&cache->lock); |
| 2694 | cache->delalloc_bytes -= len; | 2694 | cache->delalloc_bytes -= len; |
| 2695 | spin_unlock(&cache->lock); | 2695 | spin_unlock(&cache->lock); |
| 2696 | 2696 | ||
| 2697 | btrfs_put_block_group(cache); | 2697 | btrfs_put_block_group(cache); |
| 2698 | } | 2698 | } |
| 2699 | 2699 | ||
| 2700 | /* as ordered data IO finishes, this gets called so we can finish | 2700 | /* as ordered data IO finishes, this gets called so we can finish |
| 2701 | * an ordered extent if the range of bytes in the file it covers are | 2701 | * an ordered extent if the range of bytes in the file it covers are |
| 2702 | * fully written. | 2702 | * fully written. |
| 2703 | */ | 2703 | */ |
| 2704 | static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) | 2704 | static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent) |
| 2705 | { | 2705 | { |
| 2706 | struct inode *inode = ordered_extent->inode; | 2706 | struct inode *inode = ordered_extent->inode; |
| 2707 | struct btrfs_root *root = BTRFS_I(inode)->root; | 2707 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 2708 | struct btrfs_trans_handle *trans = NULL; | 2708 | struct btrfs_trans_handle *trans = NULL; |
| 2709 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 2709 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 2710 | struct extent_state *cached_state = NULL; | 2710 | struct extent_state *cached_state = NULL; |
| 2711 | struct new_sa_defrag_extent *new = NULL; | 2711 | struct new_sa_defrag_extent *new = NULL; |
| 2712 | int compress_type = 0; | 2712 | int compress_type = 0; |
| 2713 | int ret = 0; | 2713 | int ret = 0; |
| 2714 | u64 logical_len = ordered_extent->len; | 2714 | u64 logical_len = ordered_extent->len; |
| 2715 | bool nolock; | 2715 | bool nolock; |
| 2716 | bool truncated = false; | 2716 | bool truncated = false; |
| 2717 | 2717 | ||
| 2718 | nolock = btrfs_is_free_space_inode(inode); | 2718 | nolock = btrfs_is_free_space_inode(inode); |
| 2719 | 2719 | ||
| 2720 | if (test_bit(BTRFS_ORDERED_IOERR, &ordered_extent->flags)) { | 2720 | if (test_bit(BTRFS_ORDERED_IOERR, &ordered_extent->flags)) { |
| 2721 | ret = -EIO; | 2721 | ret = -EIO; |
| 2722 | goto out; | 2722 | goto out; |
| 2723 | } | 2723 | } |
| 2724 | 2724 | ||
| 2725 | btrfs_free_io_failure_record(inode, ordered_extent->file_offset, | 2725 | btrfs_free_io_failure_record(inode, ordered_extent->file_offset, |
| 2726 | ordered_extent->file_offset + | 2726 | ordered_extent->file_offset + |
| 2727 | ordered_extent->len - 1); | 2727 | ordered_extent->len - 1); |
| 2728 | 2728 | ||
| 2729 | if (test_bit(BTRFS_ORDERED_TRUNCATED, &ordered_extent->flags)) { | 2729 | if (test_bit(BTRFS_ORDERED_TRUNCATED, &ordered_extent->flags)) { |
| 2730 | truncated = true; | 2730 | truncated = true; |
| 2731 | logical_len = ordered_extent->truncated_len; | 2731 | logical_len = ordered_extent->truncated_len; |
| 2732 | /* Truncated the entire extent, don't bother adding */ | 2732 | /* Truncated the entire extent, don't bother adding */ |
| 2733 | if (!logical_len) | 2733 | if (!logical_len) |
| 2734 | goto out; | 2734 | goto out; |
| 2735 | } | 2735 | } |
| 2736 | 2736 | ||
| 2737 | if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) { | 2737 | if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) { |
| 2738 | BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */ | 2738 | BUG_ON(!list_empty(&ordered_extent->list)); /* Logic error */ |
| 2739 | btrfs_ordered_update_i_size(inode, 0, ordered_extent); | 2739 | btrfs_ordered_update_i_size(inode, 0, ordered_extent); |
| 2740 | if (nolock) | 2740 | if (nolock) |
| 2741 | trans = btrfs_join_transaction_nolock(root); | 2741 | trans = btrfs_join_transaction_nolock(root); |
| 2742 | else | 2742 | else |
| 2743 | trans = btrfs_join_transaction(root); | 2743 | trans = btrfs_join_transaction(root); |
| 2744 | if (IS_ERR(trans)) { | 2744 | if (IS_ERR(trans)) { |
| 2745 | ret = PTR_ERR(trans); | 2745 | ret = PTR_ERR(trans); |
| 2746 | trans = NULL; | 2746 | trans = NULL; |
| 2747 | goto out; | 2747 | goto out; |
| 2748 | } | 2748 | } |
| 2749 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; | 2749 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; |
| 2750 | ret = btrfs_update_inode_fallback(trans, root, inode); | 2750 | ret = btrfs_update_inode_fallback(trans, root, inode); |
| 2751 | if (ret) /* -ENOMEM or corruption */ | 2751 | if (ret) /* -ENOMEM or corruption */ |
| 2752 | btrfs_abort_transaction(trans, root, ret); | 2752 | btrfs_abort_transaction(trans, root, ret); |
| 2753 | goto out; | 2753 | goto out; |
| 2754 | } | 2754 | } |
| 2755 | 2755 | ||
| 2756 | lock_extent_bits(io_tree, ordered_extent->file_offset, | 2756 | lock_extent_bits(io_tree, ordered_extent->file_offset, |
| 2757 | ordered_extent->file_offset + ordered_extent->len - 1, | 2757 | ordered_extent->file_offset + ordered_extent->len - 1, |
| 2758 | 0, &cached_state); | 2758 | 0, &cached_state); |
| 2759 | 2759 | ||
| 2760 | ret = test_range_bit(io_tree, ordered_extent->file_offset, | 2760 | ret = test_range_bit(io_tree, ordered_extent->file_offset, |
| 2761 | ordered_extent->file_offset + ordered_extent->len - 1, | 2761 | ordered_extent->file_offset + ordered_extent->len - 1, |
| 2762 | EXTENT_DEFRAG, 1, cached_state); | 2762 | EXTENT_DEFRAG, 1, cached_state); |
| 2763 | if (ret) { | 2763 | if (ret) { |
| 2764 | u64 last_snapshot = btrfs_root_last_snapshot(&root->root_item); | 2764 | u64 last_snapshot = btrfs_root_last_snapshot(&root->root_item); |
| 2765 | if (0 && last_snapshot >= BTRFS_I(inode)->generation) | 2765 | if (0 && last_snapshot >= BTRFS_I(inode)->generation) |
| 2766 | /* the inode is shared */ | 2766 | /* the inode is shared */ |
| 2767 | new = record_old_file_extents(inode, ordered_extent); | 2767 | new = record_old_file_extents(inode, ordered_extent); |
| 2768 | 2768 | ||
| 2769 | clear_extent_bit(io_tree, ordered_extent->file_offset, | 2769 | clear_extent_bit(io_tree, ordered_extent->file_offset, |
| 2770 | ordered_extent->file_offset + ordered_extent->len - 1, | 2770 | ordered_extent->file_offset + ordered_extent->len - 1, |
| 2771 | EXTENT_DEFRAG, 0, 0, &cached_state, GFP_NOFS); | 2771 | EXTENT_DEFRAG, 0, 0, &cached_state, GFP_NOFS); |
| 2772 | } | 2772 | } |
| 2773 | 2773 | ||
| 2774 | if (nolock) | 2774 | if (nolock) |
| 2775 | trans = btrfs_join_transaction_nolock(root); | 2775 | trans = btrfs_join_transaction_nolock(root); |
| 2776 | else | 2776 | else |
| 2777 | trans = btrfs_join_transaction(root); | 2777 | trans = btrfs_join_transaction(root); |
| 2778 | if (IS_ERR(trans)) { | 2778 | if (IS_ERR(trans)) { |
| 2779 | ret = PTR_ERR(trans); | 2779 | ret = PTR_ERR(trans); |
| 2780 | trans = NULL; | 2780 | trans = NULL; |
| 2781 | goto out_unlock; | 2781 | goto out_unlock; |
| 2782 | } | 2782 | } |
| 2783 | 2783 | ||
| 2784 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; | 2784 | trans->block_rsv = &root->fs_info->delalloc_block_rsv; |
| 2785 | 2785 | ||
| 2786 | if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags)) | 2786 | if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags)) |
| 2787 | compress_type = ordered_extent->compress_type; | 2787 | compress_type = ordered_extent->compress_type; |
| 2788 | if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) { | 2788 | if (test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) { |
| 2789 | BUG_ON(compress_type); | 2789 | BUG_ON(compress_type); |
| 2790 | ret = btrfs_mark_extent_written(trans, inode, | 2790 | ret = btrfs_mark_extent_written(trans, inode, |
| 2791 | ordered_extent->file_offset, | 2791 | ordered_extent->file_offset, |
| 2792 | ordered_extent->file_offset + | 2792 | ordered_extent->file_offset + |
| 2793 | logical_len); | 2793 | logical_len); |
| 2794 | } else { | 2794 | } else { |
| 2795 | BUG_ON(root == root->fs_info->tree_root); | 2795 | BUG_ON(root == root->fs_info->tree_root); |
| 2796 | ret = insert_reserved_file_extent(trans, inode, | 2796 | ret = insert_reserved_file_extent(trans, inode, |
| 2797 | ordered_extent->file_offset, | 2797 | ordered_extent->file_offset, |
| 2798 | ordered_extent->start, | 2798 | ordered_extent->start, |
| 2799 | ordered_extent->disk_len, | 2799 | ordered_extent->disk_len, |
| 2800 | logical_len, logical_len, | 2800 | logical_len, logical_len, |
| 2801 | compress_type, 0, 0, | 2801 | compress_type, 0, 0, |
| 2802 | BTRFS_FILE_EXTENT_REG); | 2802 | BTRFS_FILE_EXTENT_REG); |
| 2803 | if (!ret) | 2803 | if (!ret) |
| 2804 | btrfs_release_delalloc_bytes(root, | 2804 | btrfs_release_delalloc_bytes(root, |
| 2805 | ordered_extent->start, | 2805 | ordered_extent->start, |
| 2806 | ordered_extent->disk_len); | 2806 | ordered_extent->disk_len); |
| 2807 | } | 2807 | } |
| 2808 | unpin_extent_cache(&BTRFS_I(inode)->extent_tree, | 2808 | unpin_extent_cache(&BTRFS_I(inode)->extent_tree, |
| 2809 | ordered_extent->file_offset, ordered_extent->len, | 2809 | ordered_extent->file_offset, ordered_extent->len, |
| 2810 | trans->transid); | 2810 | trans->transid); |
| 2811 | if (ret < 0) { | 2811 | if (ret < 0) { |
| 2812 | btrfs_abort_transaction(trans, root, ret); | 2812 | btrfs_abort_transaction(trans, root, ret); |
| 2813 | goto out_unlock; | 2813 | goto out_unlock; |
| 2814 | } | 2814 | } |
| 2815 | 2815 | ||
| 2816 | add_pending_csums(trans, inode, ordered_extent->file_offset, | 2816 | add_pending_csums(trans, inode, ordered_extent->file_offset, |
| 2817 | &ordered_extent->list); | 2817 | &ordered_extent->list); |
| 2818 | 2818 | ||
| 2819 | btrfs_ordered_update_i_size(inode, 0, ordered_extent); | 2819 | btrfs_ordered_update_i_size(inode, 0, ordered_extent); |
| 2820 | ret = btrfs_update_inode_fallback(trans, root, inode); | 2820 | ret = btrfs_update_inode_fallback(trans, root, inode); |
| 2821 | if (ret) { /* -ENOMEM or corruption */ | 2821 | if (ret) { /* -ENOMEM or corruption */ |
| 2822 | btrfs_abort_transaction(trans, root, ret); | 2822 | btrfs_abort_transaction(trans, root, ret); |
| 2823 | goto out_unlock; | 2823 | goto out_unlock; |
| 2824 | } | 2824 | } |
| 2825 | ret = 0; | 2825 | ret = 0; |
| 2826 | out_unlock: | 2826 | out_unlock: |
| 2827 | unlock_extent_cached(io_tree, ordered_extent->file_offset, | 2827 | unlock_extent_cached(io_tree, ordered_extent->file_offset, |
| 2828 | ordered_extent->file_offset + | 2828 | ordered_extent->file_offset + |
| 2829 | ordered_extent->len - 1, &cached_state, GFP_NOFS); | 2829 | ordered_extent->len - 1, &cached_state, GFP_NOFS); |
| 2830 | out: | 2830 | out: |
| 2831 | if (root != root->fs_info->tree_root) | 2831 | if (root != root->fs_info->tree_root) |
| 2832 | btrfs_delalloc_release_metadata(inode, ordered_extent->len); | 2832 | btrfs_delalloc_release_metadata(inode, ordered_extent->len); |
| 2833 | if (trans) | 2833 | if (trans) |
| 2834 | btrfs_end_transaction(trans, root); | 2834 | btrfs_end_transaction(trans, root); |
| 2835 | 2835 | ||
| 2836 | if (ret || truncated) { | 2836 | if (ret || truncated) { |
| 2837 | u64 start, end; | 2837 | u64 start, end; |
| 2838 | 2838 | ||
| 2839 | if (truncated) | 2839 | if (truncated) |
| 2840 | start = ordered_extent->file_offset + logical_len; | 2840 | start = ordered_extent->file_offset + logical_len; |
| 2841 | else | 2841 | else |
| 2842 | start = ordered_extent->file_offset; | 2842 | start = ordered_extent->file_offset; |
| 2843 | end = ordered_extent->file_offset + ordered_extent->len - 1; | 2843 | end = ordered_extent->file_offset + ordered_extent->len - 1; |
| 2844 | clear_extent_uptodate(io_tree, start, end, NULL, GFP_NOFS); | 2844 | clear_extent_uptodate(io_tree, start, end, NULL, GFP_NOFS); |
| 2845 | 2845 | ||
| 2846 | /* Drop the cache for the part of the extent we didn't write. */ | 2846 | /* Drop the cache for the part of the extent we didn't write. */ |
| 2847 | btrfs_drop_extent_cache(inode, start, end, 0); | 2847 | btrfs_drop_extent_cache(inode, start, end, 0); |
| 2848 | 2848 | ||
| 2849 | /* | 2849 | /* |
| 2850 | * If the ordered extent had an IOERR or something else went | 2850 | * If the ordered extent had an IOERR or something else went |
| 2851 | * wrong we need to return the space for this ordered extent | 2851 | * wrong we need to return the space for this ordered extent |
| 2852 | * back to the allocator. We only free the extent in the | 2852 | * back to the allocator. We only free the extent in the |
| 2853 | * truncated case if we didn't write out the extent at all. | 2853 | * truncated case if we didn't write out the extent at all. |
| 2854 | */ | 2854 | */ |
| 2855 | if ((ret || !logical_len) && | 2855 | if ((ret || !logical_len) && |
| 2856 | !test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags) && | 2856 | !test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags) && |
| 2857 | !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) | 2857 | !test_bit(BTRFS_ORDERED_PREALLOC, &ordered_extent->flags)) |
| 2858 | btrfs_free_reserved_extent(root, ordered_extent->start, | 2858 | btrfs_free_reserved_extent(root, ordered_extent->start, |
| 2859 | ordered_extent->disk_len, 1); | 2859 | ordered_extent->disk_len, 1); |
| 2860 | } | 2860 | } |
| 2861 | 2861 | ||
| 2862 | 2862 | ||
| 2863 | /* | 2863 | /* |
| 2864 | * This needs to be done to make sure anybody waiting knows we are done | 2864 | * This needs to be done to make sure anybody waiting knows we are done |
| 2865 | * updating everything for this ordered extent. | 2865 | * updating everything for this ordered extent. |
| 2866 | */ | 2866 | */ |
| 2867 | btrfs_remove_ordered_extent(inode, ordered_extent); | 2867 | btrfs_remove_ordered_extent(inode, ordered_extent); |
| 2868 | 2868 | ||
| 2869 | /* for snapshot-aware defrag */ | 2869 | /* for snapshot-aware defrag */ |
| 2870 | if (new) { | 2870 | if (new) { |
| 2871 | if (ret) { | 2871 | if (ret) { |
| 2872 | free_sa_defrag_extent(new); | 2872 | free_sa_defrag_extent(new); |
| 2873 | atomic_dec(&root->fs_info->defrag_running); | 2873 | atomic_dec(&root->fs_info->defrag_running); |
| 2874 | } else { | 2874 | } else { |
| 2875 | relink_file_extents(new); | 2875 | relink_file_extents(new); |
| 2876 | } | 2876 | } |
| 2877 | } | 2877 | } |
| 2878 | 2878 | ||
| 2879 | /* once for us */ | 2879 | /* once for us */ |
| 2880 | btrfs_put_ordered_extent(ordered_extent); | 2880 | btrfs_put_ordered_extent(ordered_extent); |
| 2881 | /* once for the tree */ | 2881 | /* once for the tree */ |
| 2882 | btrfs_put_ordered_extent(ordered_extent); | 2882 | btrfs_put_ordered_extent(ordered_extent); |
| 2883 | 2883 | ||
| 2884 | return ret; | 2884 | return ret; |
| 2885 | } | 2885 | } |
| 2886 | 2886 | ||
| 2887 | static void finish_ordered_fn(struct btrfs_work *work) | 2887 | static void finish_ordered_fn(struct btrfs_work *work) |
| 2888 | { | 2888 | { |
| 2889 | struct btrfs_ordered_extent *ordered_extent; | 2889 | struct btrfs_ordered_extent *ordered_extent; |
| 2890 | ordered_extent = container_of(work, struct btrfs_ordered_extent, work); | 2890 | ordered_extent = container_of(work, struct btrfs_ordered_extent, work); |
| 2891 | btrfs_finish_ordered_io(ordered_extent); | 2891 | btrfs_finish_ordered_io(ordered_extent); |
| 2892 | } | 2892 | } |
| 2893 | 2893 | ||
| 2894 | static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, | 2894 | static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, |
| 2895 | struct extent_state *state, int uptodate) | 2895 | struct extent_state *state, int uptodate) |
| 2896 | { | 2896 | { |
| 2897 | struct inode *inode = page->mapping->host; | 2897 | struct inode *inode = page->mapping->host; |
| 2898 | struct btrfs_root *root = BTRFS_I(inode)->root; | 2898 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 2899 | struct btrfs_ordered_extent *ordered_extent = NULL; | 2899 | struct btrfs_ordered_extent *ordered_extent = NULL; |
| 2900 | struct btrfs_workqueue *wq; | 2900 | struct btrfs_workqueue *wq; |
| 2901 | btrfs_work_func_t func; | 2901 | btrfs_work_func_t func; |
| 2902 | 2902 | ||
| 2903 | trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); | 2903 | trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); |
| 2904 | 2904 | ||
| 2905 | ClearPagePrivate2(page); | 2905 | ClearPagePrivate2(page); |
| 2906 | if (!btrfs_dec_test_ordered_pending(inode, &ordered_extent, start, | 2906 | if (!btrfs_dec_test_ordered_pending(inode, &ordered_extent, start, |
| 2907 | end - start + 1, uptodate)) | 2907 | end - start + 1, uptodate)) |
| 2908 | return 0; | 2908 | return 0; |
| 2909 | 2909 | ||
| 2910 | if (btrfs_is_free_space_inode(inode)) { | 2910 | if (btrfs_is_free_space_inode(inode)) { |
| 2911 | wq = root->fs_info->endio_freespace_worker; | 2911 | wq = root->fs_info->endio_freespace_worker; |
| 2912 | func = btrfs_freespace_write_helper; | 2912 | func = btrfs_freespace_write_helper; |
| 2913 | } else { | 2913 | } else { |
| 2914 | wq = root->fs_info->endio_write_workers; | 2914 | wq = root->fs_info->endio_write_workers; |
| 2915 | func = btrfs_endio_write_helper; | 2915 | func = btrfs_endio_write_helper; |
| 2916 | } | 2916 | } |
| 2917 | 2917 | ||
| 2918 | btrfs_init_work(&ordered_extent->work, func, finish_ordered_fn, NULL, | 2918 | btrfs_init_work(&ordered_extent->work, func, finish_ordered_fn, NULL, |
| 2919 | NULL); | 2919 | NULL); |
| 2920 | btrfs_queue_work(wq, &ordered_extent->work); | 2920 | btrfs_queue_work(wq, &ordered_extent->work); |
| 2921 | 2921 | ||
| 2922 | return 0; | 2922 | return 0; |
| 2923 | } | 2923 | } |
| 2924 | 2924 | ||
| 2925 | static int __readpage_endio_check(struct inode *inode, | 2925 | static int __readpage_endio_check(struct inode *inode, |
| 2926 | struct btrfs_io_bio *io_bio, | 2926 | struct btrfs_io_bio *io_bio, |
| 2927 | int icsum, struct page *page, | 2927 | int icsum, struct page *page, |
| 2928 | int pgoff, u64 start, size_t len) | 2928 | int pgoff, u64 start, size_t len) |
| 2929 | { | 2929 | { |
| 2930 | char *kaddr; | 2930 | char *kaddr; |
| 2931 | u32 csum_expected; | 2931 | u32 csum_expected; |
| 2932 | u32 csum = ~(u32)0; | 2932 | u32 csum = ~(u32)0; |
| 2933 | static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, | 2933 | static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, |
| 2934 | DEFAULT_RATELIMIT_BURST); | 2934 | DEFAULT_RATELIMIT_BURST); |
| 2935 | 2935 | ||
| 2936 | csum_expected = *(((u32 *)io_bio->csum) + icsum); | 2936 | csum_expected = *(((u32 *)io_bio->csum) + icsum); |
| 2937 | 2937 | ||
| 2938 | kaddr = kmap_atomic(page); | 2938 | kaddr = kmap_atomic(page); |
| 2939 | csum = btrfs_csum_data(kaddr + pgoff, csum, len); | 2939 | csum = btrfs_csum_data(kaddr + pgoff, csum, len); |
| 2940 | btrfs_csum_final(csum, (char *)&csum); | 2940 | btrfs_csum_final(csum, (char *)&csum); |
| 2941 | if (csum != csum_expected) | 2941 | if (csum != csum_expected) |
| 2942 | goto zeroit; | 2942 | goto zeroit; |
| 2943 | 2943 | ||
| 2944 | kunmap_atomic(kaddr); | 2944 | kunmap_atomic(kaddr); |
| 2945 | return 0; | 2945 | return 0; |
| 2946 | zeroit: | 2946 | zeroit: |
| 2947 | if (__ratelimit(&_rs)) | 2947 | if (__ratelimit(&_rs)) |
| 2948 | btrfs_info(BTRFS_I(inode)->root->fs_info, | 2948 | btrfs_info(BTRFS_I(inode)->root->fs_info, |
| 2949 | "csum failed ino %llu off %llu csum %u expected csum %u", | 2949 | "csum failed ino %llu off %llu csum %u expected csum %u", |
| 2950 | btrfs_ino(inode), start, csum, csum_expected); | 2950 | btrfs_ino(inode), start, csum, csum_expected); |
| 2951 | memset(kaddr + pgoff, 1, len); | 2951 | memset(kaddr + pgoff, 1, len); |
| 2952 | flush_dcache_page(page); | 2952 | flush_dcache_page(page); |
| 2953 | kunmap_atomic(kaddr); | 2953 | kunmap_atomic(kaddr); |
| 2954 | if (csum_expected == 0) | 2954 | if (csum_expected == 0) |
| 2955 | return 0; | 2955 | return 0; |
| 2956 | return -EIO; | 2956 | return -EIO; |
| 2957 | } | 2957 | } |
| 2958 | 2958 | ||
| 2959 | /* | 2959 | /* |
| 2960 | * when reads are done, we need to check csums to verify the data is correct | 2960 | * when reads are done, we need to check csums to verify the data is correct |
| 2961 | * if there's a match, we allow the bio to finish. If not, the code in | 2961 | * if there's a match, we allow the bio to finish. If not, the code in |
| 2962 | * extent_io.c will try to find good copies for us. | 2962 | * extent_io.c will try to find good copies for us. |
| 2963 | */ | 2963 | */ |
| 2964 | static int btrfs_readpage_end_io_hook(struct btrfs_io_bio *io_bio, | 2964 | static int btrfs_readpage_end_io_hook(struct btrfs_io_bio *io_bio, |
| 2965 | u64 phy_offset, struct page *page, | 2965 | u64 phy_offset, struct page *page, |
| 2966 | u64 start, u64 end, int mirror) | 2966 | u64 start, u64 end, int mirror) |
| 2967 | { | 2967 | { |
| 2968 | size_t offset = start - page_offset(page); | 2968 | size_t offset = start - page_offset(page); |
| 2969 | struct inode *inode = page->mapping->host; | 2969 | struct inode *inode = page->mapping->host; |
| 2970 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 2970 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 2971 | struct btrfs_root *root = BTRFS_I(inode)->root; | 2971 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 2972 | 2972 | ||
| 2973 | if (PageChecked(page)) { | 2973 | if (PageChecked(page)) { |
| 2974 | ClearPageChecked(page); | 2974 | ClearPageChecked(page); |
| 2975 | return 0; | 2975 | return 0; |
| 2976 | } | 2976 | } |
| 2977 | 2977 | ||
| 2978 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) | 2978 | if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) |
| 2979 | return 0; | 2979 | return 0; |
| 2980 | 2980 | ||
| 2981 | if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID && | 2981 | if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID && |
| 2982 | test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) { | 2982 | test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) { |
| 2983 | clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM, | 2983 | clear_extent_bits(io_tree, start, end, EXTENT_NODATASUM, |
| 2984 | GFP_NOFS); | 2984 | GFP_NOFS); |
| 2985 | return 0; | 2985 | return 0; |
| 2986 | } | 2986 | } |
| 2987 | 2987 | ||
| 2988 | phy_offset >>= inode->i_sb->s_blocksize_bits; | 2988 | phy_offset >>= inode->i_sb->s_blocksize_bits; |
| 2989 | return __readpage_endio_check(inode, io_bio, phy_offset, page, offset, | 2989 | return __readpage_endio_check(inode, io_bio, phy_offset, page, offset, |
| 2990 | start, (size_t)(end - start + 1)); | 2990 | start, (size_t)(end - start + 1)); |
| 2991 | } | 2991 | } |
| 2992 | 2992 | ||
| 2993 | struct delayed_iput { | 2993 | struct delayed_iput { |
| 2994 | struct list_head list; | 2994 | struct list_head list; |
| 2995 | struct inode *inode; | 2995 | struct inode *inode; |
| 2996 | }; | 2996 | }; |
| 2997 | 2997 | ||
| 2998 | /* JDM: If this is fs-wide, why can't we add a pointer to | 2998 | /* JDM: If this is fs-wide, why can't we add a pointer to |
| 2999 | * btrfs_inode instead and avoid the allocation? */ | 2999 | * btrfs_inode instead and avoid the allocation? */ |
| 3000 | void btrfs_add_delayed_iput(struct inode *inode) | 3000 | void btrfs_add_delayed_iput(struct inode *inode) |
| 3001 | { | 3001 | { |
| 3002 | struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info; | 3002 | struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info; |
| 3003 | struct delayed_iput *delayed; | 3003 | struct delayed_iput *delayed; |
| 3004 | 3004 | ||
| 3005 | if (atomic_add_unless(&inode->i_count, -1, 1)) | 3005 | if (atomic_add_unless(&inode->i_count, -1, 1)) |
| 3006 | return; | 3006 | return; |
| 3007 | 3007 | ||
| 3008 | delayed = kmalloc(sizeof(*delayed), GFP_NOFS | __GFP_NOFAIL); | 3008 | delayed = kmalloc(sizeof(*delayed), GFP_NOFS | __GFP_NOFAIL); |
| 3009 | delayed->inode = inode; | 3009 | delayed->inode = inode; |
| 3010 | 3010 | ||
| 3011 | spin_lock(&fs_info->delayed_iput_lock); | 3011 | spin_lock(&fs_info->delayed_iput_lock); |
| 3012 | list_add_tail(&delayed->list, &fs_info->delayed_iputs); | 3012 | list_add_tail(&delayed->list, &fs_info->delayed_iputs); |
| 3013 | spin_unlock(&fs_info->delayed_iput_lock); | 3013 | spin_unlock(&fs_info->delayed_iput_lock); |
| 3014 | } | 3014 | } |
| 3015 | 3015 | ||
| 3016 | void btrfs_run_delayed_iputs(struct btrfs_root *root) | 3016 | void btrfs_run_delayed_iputs(struct btrfs_root *root) |
| 3017 | { | 3017 | { |
| 3018 | LIST_HEAD(list); | 3018 | LIST_HEAD(list); |
| 3019 | struct btrfs_fs_info *fs_info = root->fs_info; | 3019 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 3020 | struct delayed_iput *delayed; | 3020 | struct delayed_iput *delayed; |
| 3021 | int empty; | 3021 | int empty; |
| 3022 | 3022 | ||
| 3023 | spin_lock(&fs_info->delayed_iput_lock); | 3023 | spin_lock(&fs_info->delayed_iput_lock); |
| 3024 | empty = list_empty(&fs_info->delayed_iputs); | 3024 | empty = list_empty(&fs_info->delayed_iputs); |
| 3025 | spin_unlock(&fs_info->delayed_iput_lock); | 3025 | spin_unlock(&fs_info->delayed_iput_lock); |
| 3026 | if (empty) | 3026 | if (empty) |
| 3027 | return; | 3027 | return; |
| 3028 | 3028 | ||
| 3029 | spin_lock(&fs_info->delayed_iput_lock); | 3029 | spin_lock(&fs_info->delayed_iput_lock); |
| 3030 | list_splice_init(&fs_info->delayed_iputs, &list); | 3030 | list_splice_init(&fs_info->delayed_iputs, &list); |
| 3031 | spin_unlock(&fs_info->delayed_iput_lock); | 3031 | spin_unlock(&fs_info->delayed_iput_lock); |
| 3032 | 3032 | ||
| 3033 | while (!list_empty(&list)) { | 3033 | while (!list_empty(&list)) { |
| 3034 | delayed = list_entry(list.next, struct delayed_iput, list); | 3034 | delayed = list_entry(list.next, struct delayed_iput, list); |
| 3035 | list_del(&delayed->list); | 3035 | list_del(&delayed->list); |
| 3036 | iput(delayed->inode); | 3036 | iput(delayed->inode); |
| 3037 | kfree(delayed); | 3037 | kfree(delayed); |
| 3038 | } | 3038 | } |
| 3039 | } | 3039 | } |
| 3040 | 3040 | ||
| 3041 | /* | 3041 | /* |
| 3042 | * This is called in transaction commit time. If there are no orphan | 3042 | * This is called in transaction commit time. If there are no orphan |
| 3043 | * files in the subvolume, it removes orphan item and frees block_rsv | 3043 | * files in the subvolume, it removes orphan item and frees block_rsv |
| 3044 | * structure. | 3044 | * structure. |
| 3045 | */ | 3045 | */ |
| 3046 | void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, | 3046 | void btrfs_orphan_commit_root(struct btrfs_trans_handle *trans, |
| 3047 | struct btrfs_root *root) | 3047 | struct btrfs_root *root) |
| 3048 | { | 3048 | { |
| 3049 | struct btrfs_block_rsv *block_rsv; | 3049 | struct btrfs_block_rsv *block_rsv; |
| 3050 | int ret; | 3050 | int ret; |
| 3051 | 3051 | ||
| 3052 | if (atomic_read(&root->orphan_inodes) || | 3052 | if (atomic_read(&root->orphan_inodes) || |
| 3053 | root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) | 3053 | root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) |
| 3054 | return; | 3054 | return; |
| 3055 | 3055 | ||
| 3056 | spin_lock(&root->orphan_lock); | 3056 | spin_lock(&root->orphan_lock); |
| 3057 | if (atomic_read(&root->orphan_inodes)) { | 3057 | if (atomic_read(&root->orphan_inodes)) { |
| 3058 | spin_unlock(&root->orphan_lock); | 3058 | spin_unlock(&root->orphan_lock); |
| 3059 | return; | 3059 | return; |
| 3060 | } | 3060 | } |
| 3061 | 3061 | ||
| 3062 | if (root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) { | 3062 | if (root->orphan_cleanup_state != ORPHAN_CLEANUP_DONE) { |
| 3063 | spin_unlock(&root->orphan_lock); | 3063 | spin_unlock(&root->orphan_lock); |
| 3064 | return; | 3064 | return; |
| 3065 | } | 3065 | } |
| 3066 | 3066 | ||
| 3067 | block_rsv = root->orphan_block_rsv; | 3067 | block_rsv = root->orphan_block_rsv; |
| 3068 | root->orphan_block_rsv = NULL; | 3068 | root->orphan_block_rsv = NULL; |
| 3069 | spin_unlock(&root->orphan_lock); | 3069 | spin_unlock(&root->orphan_lock); |
| 3070 | 3070 | ||
| 3071 | if (test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state) && | 3071 | if (test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state) && |
| 3072 | btrfs_root_refs(&root->root_item) > 0) { | 3072 | btrfs_root_refs(&root->root_item) > 0) { |
| 3073 | ret = btrfs_del_orphan_item(trans, root->fs_info->tree_root, | 3073 | ret = btrfs_del_orphan_item(trans, root->fs_info->tree_root, |
| 3074 | root->root_key.objectid); | 3074 | root->root_key.objectid); |
| 3075 | if (ret) | 3075 | if (ret) |
| 3076 | btrfs_abort_transaction(trans, root, ret); | 3076 | btrfs_abort_transaction(trans, root, ret); |
| 3077 | else | 3077 | else |
| 3078 | clear_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, | 3078 | clear_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, |
| 3079 | &root->state); | 3079 | &root->state); |
| 3080 | } | 3080 | } |
| 3081 | 3081 | ||
| 3082 | if (block_rsv) { | 3082 | if (block_rsv) { |
| 3083 | WARN_ON(block_rsv->size > 0); | 3083 | WARN_ON(block_rsv->size > 0); |
| 3084 | btrfs_free_block_rsv(root, block_rsv); | 3084 | btrfs_free_block_rsv(root, block_rsv); |
| 3085 | } | 3085 | } |
| 3086 | } | 3086 | } |
| 3087 | 3087 | ||
| 3088 | /* | 3088 | /* |
| 3089 | * This creates an orphan entry for the given inode in case something goes | 3089 | * This creates an orphan entry for the given inode in case something goes |
| 3090 | * wrong in the middle of an unlink/truncate. | 3090 | * wrong in the middle of an unlink/truncate. |
| 3091 | * | 3091 | * |
| 3092 | * NOTE: caller of this function should reserve 5 units of metadata for | 3092 | * NOTE: caller of this function should reserve 5 units of metadata for |
| 3093 | * this function. | 3093 | * this function. |
| 3094 | */ | 3094 | */ |
| 3095 | int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) | 3095 | int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) |
| 3096 | { | 3096 | { |
| 3097 | struct btrfs_root *root = BTRFS_I(inode)->root; | 3097 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 3098 | struct btrfs_block_rsv *block_rsv = NULL; | 3098 | struct btrfs_block_rsv *block_rsv = NULL; |
| 3099 | int reserve = 0; | 3099 | int reserve = 0; |
| 3100 | int insert = 0; | 3100 | int insert = 0; |
| 3101 | int ret; | 3101 | int ret; |
| 3102 | 3102 | ||
| 3103 | if (!root->orphan_block_rsv) { | 3103 | if (!root->orphan_block_rsv) { |
| 3104 | block_rsv = btrfs_alloc_block_rsv(root, BTRFS_BLOCK_RSV_TEMP); | 3104 | block_rsv = btrfs_alloc_block_rsv(root, BTRFS_BLOCK_RSV_TEMP); |
| 3105 | if (!block_rsv) | 3105 | if (!block_rsv) |
| 3106 | return -ENOMEM; | 3106 | return -ENOMEM; |
| 3107 | } | 3107 | } |
| 3108 | 3108 | ||
| 3109 | spin_lock(&root->orphan_lock); | 3109 | spin_lock(&root->orphan_lock); |
| 3110 | if (!root->orphan_block_rsv) { | 3110 | if (!root->orphan_block_rsv) { |
| 3111 | root->orphan_block_rsv = block_rsv; | 3111 | root->orphan_block_rsv = block_rsv; |
| 3112 | } else if (block_rsv) { | 3112 | } else if (block_rsv) { |
| 3113 | btrfs_free_block_rsv(root, block_rsv); | 3113 | btrfs_free_block_rsv(root, block_rsv); |
| 3114 | block_rsv = NULL; | 3114 | block_rsv = NULL; |
| 3115 | } | 3115 | } |
| 3116 | 3116 | ||
| 3117 | if (!test_and_set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, | 3117 | if (!test_and_set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, |
| 3118 | &BTRFS_I(inode)->runtime_flags)) { | 3118 | &BTRFS_I(inode)->runtime_flags)) { |
| 3119 | #if 0 | 3119 | #if 0 |
| 3120 | /* | 3120 | /* |
| 3121 | * For proper ENOSPC handling, we should do orphan | 3121 | * For proper ENOSPC handling, we should do orphan |
| 3122 | * cleanup when mounting. But this introduces backward | 3122 | * cleanup when mounting. But this introduces backward |
| 3123 | * compatibility issue. | 3123 | * compatibility issue. |
| 3124 | */ | 3124 | */ |
| 3125 | if (!xchg(&root->orphan_item_inserted, 1)) | 3125 | if (!xchg(&root->orphan_item_inserted, 1)) |
| 3126 | insert = 2; | 3126 | insert = 2; |
| 3127 | else | 3127 | else |
| 3128 | insert = 1; | 3128 | insert = 1; |
| 3129 | #endif | 3129 | #endif |
| 3130 | insert = 1; | 3130 | insert = 1; |
| 3131 | atomic_inc(&root->orphan_inodes); | 3131 | atomic_inc(&root->orphan_inodes); |
| 3132 | } | 3132 | } |
| 3133 | 3133 | ||
| 3134 | if (!test_and_set_bit(BTRFS_INODE_ORPHAN_META_RESERVED, | 3134 | if (!test_and_set_bit(BTRFS_INODE_ORPHAN_META_RESERVED, |
| 3135 | &BTRFS_I(inode)->runtime_flags)) | 3135 | &BTRFS_I(inode)->runtime_flags)) |
| 3136 | reserve = 1; | 3136 | reserve = 1; |
| 3137 | spin_unlock(&root->orphan_lock); | 3137 | spin_unlock(&root->orphan_lock); |
| 3138 | 3138 | ||
| 3139 | /* grab metadata reservation from transaction handle */ | 3139 | /* grab metadata reservation from transaction handle */ |
| 3140 | if (reserve) { | 3140 | if (reserve) { |
| 3141 | ret = btrfs_orphan_reserve_metadata(trans, inode); | 3141 | ret = btrfs_orphan_reserve_metadata(trans, inode); |
| 3142 | BUG_ON(ret); /* -ENOSPC in reservation; Logic error? JDM */ | 3142 | BUG_ON(ret); /* -ENOSPC in reservation; Logic error? JDM */ |
| 3143 | } | 3143 | } |
| 3144 | 3144 | ||
| 3145 | /* insert an orphan item to track this unlinked/truncated file */ | 3145 | /* insert an orphan item to track this unlinked/truncated file */ |
| 3146 | if (insert >= 1) { | 3146 | if (insert >= 1) { |
| 3147 | ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); | 3147 | ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); |
| 3148 | if (ret) { | 3148 | if (ret) { |
| 3149 | atomic_dec(&root->orphan_inodes); | 3149 | atomic_dec(&root->orphan_inodes); |
| 3150 | if (reserve) { | 3150 | if (reserve) { |
| 3151 | clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, | 3151 | clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, |
| 3152 | &BTRFS_I(inode)->runtime_flags); | 3152 | &BTRFS_I(inode)->runtime_flags); |
| 3153 | btrfs_orphan_release_metadata(inode); | 3153 | btrfs_orphan_release_metadata(inode); |
| 3154 | } | 3154 | } |
| 3155 | if (ret != -EEXIST) { | 3155 | if (ret != -EEXIST) { |
| 3156 | clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, | 3156 | clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, |
| 3157 | &BTRFS_I(inode)->runtime_flags); | 3157 | &BTRFS_I(inode)->runtime_flags); |
| 3158 | btrfs_abort_transaction(trans, root, ret); | 3158 | btrfs_abort_transaction(trans, root, ret); |
| 3159 | return ret; | 3159 | return ret; |
| 3160 | } | 3160 | } |
| 3161 | } | 3161 | } |
| 3162 | ret = 0; | 3162 | ret = 0; |
| 3163 | } | 3163 | } |
| 3164 | 3164 | ||
| 3165 | /* insert an orphan item to track subvolume contains orphan files */ | 3165 | /* insert an orphan item to track subvolume contains orphan files */ |
| 3166 | if (insert >= 2) { | 3166 | if (insert >= 2) { |
| 3167 | ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, | 3167 | ret = btrfs_insert_orphan_item(trans, root->fs_info->tree_root, |
| 3168 | root->root_key.objectid); | 3168 | root->root_key.objectid); |
| 3169 | if (ret && ret != -EEXIST) { | 3169 | if (ret && ret != -EEXIST) { |
| 3170 | btrfs_abort_transaction(trans, root, ret); | 3170 | btrfs_abort_transaction(trans, root, ret); |
| 3171 | return ret; | 3171 | return ret; |
| 3172 | } | 3172 | } |
| 3173 | } | 3173 | } |
| 3174 | return 0; | 3174 | return 0; |
| 3175 | } | 3175 | } |
| 3176 | 3176 | ||
| 3177 | /* | 3177 | /* |
| 3178 | * We have done the truncate/delete so we can go ahead and remove the orphan | 3178 | * We have done the truncate/delete so we can go ahead and remove the orphan |
| 3179 | * item for this particular inode. | 3179 | * item for this particular inode. |
| 3180 | */ | 3180 | */ |
| 3181 | static int btrfs_orphan_del(struct btrfs_trans_handle *trans, | 3181 | static int btrfs_orphan_del(struct btrfs_trans_handle *trans, |
| 3182 | struct inode *inode) | 3182 | struct inode *inode) |
| 3183 | { | 3183 | { |
| 3184 | struct btrfs_root *root = BTRFS_I(inode)->root; | 3184 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 3185 | int delete_item = 0; | 3185 | int delete_item = 0; |
| 3186 | int release_rsv = 0; | 3186 | int release_rsv = 0; |
| 3187 | int ret = 0; | 3187 | int ret = 0; |
| 3188 | 3188 | ||
| 3189 | spin_lock(&root->orphan_lock); | 3189 | spin_lock(&root->orphan_lock); |
| 3190 | if (test_and_clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, | 3190 | if (test_and_clear_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, |
| 3191 | &BTRFS_I(inode)->runtime_flags)) | 3191 | &BTRFS_I(inode)->runtime_flags)) |
| 3192 | delete_item = 1; | 3192 | delete_item = 1; |
| 3193 | 3193 | ||
| 3194 | if (test_and_clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, | 3194 | if (test_and_clear_bit(BTRFS_INODE_ORPHAN_META_RESERVED, |
| 3195 | &BTRFS_I(inode)->runtime_flags)) | 3195 | &BTRFS_I(inode)->runtime_flags)) |
| 3196 | release_rsv = 1; | 3196 | release_rsv = 1; |
| 3197 | spin_unlock(&root->orphan_lock); | 3197 | spin_unlock(&root->orphan_lock); |
| 3198 | 3198 | ||
| 3199 | if (delete_item) { | 3199 | if (delete_item) { |
| 3200 | atomic_dec(&root->orphan_inodes); | 3200 | atomic_dec(&root->orphan_inodes); |
| 3201 | if (trans) | 3201 | if (trans) |
| 3202 | ret = btrfs_del_orphan_item(trans, root, | 3202 | ret = btrfs_del_orphan_item(trans, root, |
| 3203 | btrfs_ino(inode)); | 3203 | btrfs_ino(inode)); |
| 3204 | } | 3204 | } |
| 3205 | 3205 | ||
| 3206 | if (release_rsv) | 3206 | if (release_rsv) |
| 3207 | btrfs_orphan_release_metadata(inode); | 3207 | btrfs_orphan_release_metadata(inode); |
| 3208 | 3208 | ||
| 3209 | return ret; | 3209 | return ret; |
| 3210 | } | 3210 | } |
| 3211 | 3211 | ||
| 3212 | /* | 3212 | /* |
| 3213 | * this cleans up any orphans that may be left on the list from the last use | 3213 | * this cleans up any orphans that may be left on the list from the last use |
| 3214 | * of this root. | 3214 | * of this root. |
| 3215 | */ | 3215 | */ |
| 3216 | int btrfs_orphan_cleanup(struct btrfs_root *root) | 3216 | int btrfs_orphan_cleanup(struct btrfs_root *root) |
| 3217 | { | 3217 | { |
| 3218 | struct btrfs_path *path; | 3218 | struct btrfs_path *path; |
| 3219 | struct extent_buffer *leaf; | 3219 | struct extent_buffer *leaf; |
| 3220 | struct btrfs_key key, found_key; | 3220 | struct btrfs_key key, found_key; |
| 3221 | struct btrfs_trans_handle *trans; | 3221 | struct btrfs_trans_handle *trans; |
| 3222 | struct inode *inode; | 3222 | struct inode *inode; |
| 3223 | u64 last_objectid = 0; | 3223 | u64 last_objectid = 0; |
| 3224 | int ret = 0, nr_unlink = 0, nr_truncate = 0; | 3224 | int ret = 0, nr_unlink = 0, nr_truncate = 0; |
| 3225 | 3225 | ||
| 3226 | if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) | 3226 | if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) |
| 3227 | return 0; | 3227 | return 0; |
| 3228 | 3228 | ||
| 3229 | path = btrfs_alloc_path(); | 3229 | path = btrfs_alloc_path(); |
| 3230 | if (!path) { | 3230 | if (!path) { |
| 3231 | ret = -ENOMEM; | 3231 | ret = -ENOMEM; |
| 3232 | goto out; | 3232 | goto out; |
| 3233 | } | 3233 | } |
| 3234 | path->reada = -1; | 3234 | path->reada = -1; |
| 3235 | 3235 | ||
| 3236 | key.objectid = BTRFS_ORPHAN_OBJECTID; | 3236 | key.objectid = BTRFS_ORPHAN_OBJECTID; |
| 3237 | key.type = BTRFS_ORPHAN_ITEM_KEY; | 3237 | key.type = BTRFS_ORPHAN_ITEM_KEY; |
| 3238 | key.offset = (u64)-1; | 3238 | key.offset = (u64)-1; |
| 3239 | 3239 | ||
| 3240 | while (1) { | 3240 | while (1) { |
| 3241 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 3241 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 3242 | if (ret < 0) | 3242 | if (ret < 0) |
| 3243 | goto out; | 3243 | goto out; |
| 3244 | 3244 | ||
| 3245 | /* | 3245 | /* |
| 3246 | * if ret == 0 means we found what we were searching for, which | 3246 | * if ret == 0 means we found what we were searching for, which |
| 3247 | * is weird, but possible, so only screw with path if we didn't | 3247 | * is weird, but possible, so only screw with path if we didn't |
| 3248 | * find the key and see if we have stuff that matches | 3248 | * find the key and see if we have stuff that matches |
| 3249 | */ | 3249 | */ |
| 3250 | if (ret > 0) { | 3250 | if (ret > 0) { |
| 3251 | ret = 0; | 3251 | ret = 0; |
| 3252 | if (path->slots[0] == 0) | 3252 | if (path->slots[0] == 0) |
| 3253 | break; | 3253 | break; |
| 3254 | path->slots[0]--; | 3254 | path->slots[0]--; |
| 3255 | } | 3255 | } |
| 3256 | 3256 | ||
| 3257 | /* pull out the item */ | 3257 | /* pull out the item */ |
| 3258 | leaf = path->nodes[0]; | 3258 | leaf = path->nodes[0]; |
| 3259 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 3259 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 3260 | 3260 | ||
| 3261 | /* make sure the item matches what we want */ | 3261 | /* make sure the item matches what we want */ |
| 3262 | if (found_key.objectid != BTRFS_ORPHAN_OBJECTID) | 3262 | if (found_key.objectid != BTRFS_ORPHAN_OBJECTID) |
| 3263 | break; | 3263 | break; |
| 3264 | if (found_key.type != BTRFS_ORPHAN_ITEM_KEY) | 3264 | if (found_key.type != BTRFS_ORPHAN_ITEM_KEY) |
| 3265 | break; | 3265 | break; |
| 3266 | 3266 | ||
| 3267 | /* release the path since we're done with it */ | 3267 | /* release the path since we're done with it */ |
| 3268 | btrfs_release_path(path); | 3268 | btrfs_release_path(path); |
| 3269 | 3269 | ||
| 3270 | /* | 3270 | /* |
| 3271 | * this is where we are basically btrfs_lookup, without the | 3271 | * this is where we are basically btrfs_lookup, without the |
| 3272 | * crossing root thing. we store the inode number in the | 3272 | * crossing root thing. we store the inode number in the |
| 3273 | * offset of the orphan item. | 3273 | * offset of the orphan item. |
| 3274 | */ | 3274 | */ |
| 3275 | 3275 | ||
| 3276 | if (found_key.offset == last_objectid) { | 3276 | if (found_key.offset == last_objectid) { |
| 3277 | btrfs_err(root->fs_info, | 3277 | btrfs_err(root->fs_info, |
| 3278 | "Error removing orphan entry, stopping orphan cleanup"); | 3278 | "Error removing orphan entry, stopping orphan cleanup"); |
| 3279 | ret = -EINVAL; | 3279 | ret = -EINVAL; |
| 3280 | goto out; | 3280 | goto out; |
| 3281 | } | 3281 | } |
| 3282 | 3282 | ||
| 3283 | last_objectid = found_key.offset; | 3283 | last_objectid = found_key.offset; |
| 3284 | 3284 | ||
| 3285 | found_key.objectid = found_key.offset; | 3285 | found_key.objectid = found_key.offset; |
| 3286 | found_key.type = BTRFS_INODE_ITEM_KEY; | 3286 | found_key.type = BTRFS_INODE_ITEM_KEY; |
| 3287 | found_key.offset = 0; | 3287 | found_key.offset = 0; |
| 3288 | inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); | 3288 | inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); |
| 3289 | ret = PTR_ERR_OR_ZERO(inode); | 3289 | ret = PTR_ERR_OR_ZERO(inode); |
| 3290 | if (ret && ret != -ESTALE) | 3290 | if (ret && ret != -ESTALE) |
| 3291 | goto out; | 3291 | goto out; |
| 3292 | 3292 | ||
| 3293 | if (ret == -ESTALE && root == root->fs_info->tree_root) { | 3293 | if (ret == -ESTALE && root == root->fs_info->tree_root) { |
| 3294 | struct btrfs_root *dead_root; | 3294 | struct btrfs_root *dead_root; |
| 3295 | struct btrfs_fs_info *fs_info = root->fs_info; | 3295 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 3296 | int is_dead_root = 0; | 3296 | int is_dead_root = 0; |
| 3297 | 3297 | ||
| 3298 | /* | 3298 | /* |
| 3299 | * this is an orphan in the tree root. Currently these | 3299 | * this is an orphan in the tree root. Currently these |
| 3300 | * could come from 2 sources: | 3300 | * could come from 2 sources: |
| 3301 | * a) a snapshot deletion in progress | 3301 | * a) a snapshot deletion in progress |
| 3302 | * b) a free space cache inode | 3302 | * b) a free space cache inode |
| 3303 | * We need to distinguish those two, as the snapshot | 3303 | * We need to distinguish those two, as the snapshot |
| 3304 | * orphan must not get deleted. | 3304 | * orphan must not get deleted. |
| 3305 | * find_dead_roots already ran before us, so if this | 3305 | * find_dead_roots already ran before us, so if this |
| 3306 | * is a snapshot deletion, we should find the root | 3306 | * is a snapshot deletion, we should find the root |
| 3307 | * in the dead_roots list | 3307 | * in the dead_roots list |
| 3308 | */ | 3308 | */ |
| 3309 | spin_lock(&fs_info->trans_lock); | 3309 | spin_lock(&fs_info->trans_lock); |
| 3310 | list_for_each_entry(dead_root, &fs_info->dead_roots, | 3310 | list_for_each_entry(dead_root, &fs_info->dead_roots, |
| 3311 | root_list) { | 3311 | root_list) { |
| 3312 | if (dead_root->root_key.objectid == | 3312 | if (dead_root->root_key.objectid == |
| 3313 | found_key.objectid) { | 3313 | found_key.objectid) { |
| 3314 | is_dead_root = 1; | 3314 | is_dead_root = 1; |
| 3315 | break; | 3315 | break; |
| 3316 | } | 3316 | } |
| 3317 | } | 3317 | } |
| 3318 | spin_unlock(&fs_info->trans_lock); | 3318 | spin_unlock(&fs_info->trans_lock); |
| 3319 | if (is_dead_root) { | 3319 | if (is_dead_root) { |
| 3320 | /* prevent this orphan from being found again */ | 3320 | /* prevent this orphan from being found again */ |
| 3321 | key.offset = found_key.objectid - 1; | 3321 | key.offset = found_key.objectid - 1; |
| 3322 | continue; | 3322 | continue; |
| 3323 | } | 3323 | } |
| 3324 | } | 3324 | } |
| 3325 | /* | 3325 | /* |
| 3326 | * Inode is already gone but the orphan item is still there, | 3326 | * Inode is already gone but the orphan item is still there, |
| 3327 | * kill the orphan item. | 3327 | * kill the orphan item. |
| 3328 | */ | 3328 | */ |
| 3329 | if (ret == -ESTALE) { | 3329 | if (ret == -ESTALE) { |
| 3330 | trans = btrfs_start_transaction(root, 1); | 3330 | trans = btrfs_start_transaction(root, 1); |
| 3331 | if (IS_ERR(trans)) { | 3331 | if (IS_ERR(trans)) { |
| 3332 | ret = PTR_ERR(trans); | 3332 | ret = PTR_ERR(trans); |
| 3333 | goto out; | 3333 | goto out; |
| 3334 | } | 3334 | } |
| 3335 | btrfs_debug(root->fs_info, "auto deleting %Lu", | 3335 | btrfs_debug(root->fs_info, "auto deleting %Lu", |
| 3336 | found_key.objectid); | 3336 | found_key.objectid); |
| 3337 | ret = btrfs_del_orphan_item(trans, root, | 3337 | ret = btrfs_del_orphan_item(trans, root, |
| 3338 | found_key.objectid); | 3338 | found_key.objectid); |
| 3339 | btrfs_end_transaction(trans, root); | 3339 | btrfs_end_transaction(trans, root); |
| 3340 | if (ret) | 3340 | if (ret) |
| 3341 | goto out; | 3341 | goto out; |
| 3342 | continue; | 3342 | continue; |
| 3343 | } | 3343 | } |
| 3344 | 3344 | ||
| 3345 | /* | 3345 | /* |
| 3346 | * add this inode to the orphan list so btrfs_orphan_del does | 3346 | * add this inode to the orphan list so btrfs_orphan_del does |
| 3347 | * the proper thing when we hit it | 3347 | * the proper thing when we hit it |
| 3348 | */ | 3348 | */ |
| 3349 | set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, | 3349 | set_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, |
| 3350 | &BTRFS_I(inode)->runtime_flags); | 3350 | &BTRFS_I(inode)->runtime_flags); |
| 3351 | atomic_inc(&root->orphan_inodes); | 3351 | atomic_inc(&root->orphan_inodes); |
| 3352 | 3352 | ||
| 3353 | /* if we have links, this was a truncate, lets do that */ | 3353 | /* if we have links, this was a truncate, lets do that */ |
| 3354 | if (inode->i_nlink) { | 3354 | if (inode->i_nlink) { |
| 3355 | if (WARN_ON(!S_ISREG(inode->i_mode))) { | 3355 | if (WARN_ON(!S_ISREG(inode->i_mode))) { |
| 3356 | iput(inode); | 3356 | iput(inode); |
| 3357 | continue; | 3357 | continue; |
| 3358 | } | 3358 | } |
| 3359 | nr_truncate++; | 3359 | nr_truncate++; |
| 3360 | 3360 | ||
| 3361 | /* 1 for the orphan item deletion. */ | 3361 | /* 1 for the orphan item deletion. */ |
| 3362 | trans = btrfs_start_transaction(root, 1); | 3362 | trans = btrfs_start_transaction(root, 1); |
| 3363 | if (IS_ERR(trans)) { | 3363 | if (IS_ERR(trans)) { |
| 3364 | iput(inode); | 3364 | iput(inode); |
| 3365 | ret = PTR_ERR(trans); | 3365 | ret = PTR_ERR(trans); |
| 3366 | goto out; | 3366 | goto out; |
| 3367 | } | 3367 | } |
| 3368 | ret = btrfs_orphan_add(trans, inode); | 3368 | ret = btrfs_orphan_add(trans, inode); |
| 3369 | btrfs_end_transaction(trans, root); | 3369 | btrfs_end_transaction(trans, root); |
| 3370 | if (ret) { | 3370 | if (ret) { |
| 3371 | iput(inode); | 3371 | iput(inode); |
| 3372 | goto out; | 3372 | goto out; |
| 3373 | } | 3373 | } |
| 3374 | 3374 | ||
| 3375 | ret = btrfs_truncate(inode); | 3375 | ret = btrfs_truncate(inode); |
| 3376 | if (ret) | 3376 | if (ret) |
| 3377 | btrfs_orphan_del(NULL, inode); | 3377 | btrfs_orphan_del(NULL, inode); |
| 3378 | } else { | 3378 | } else { |
| 3379 | nr_unlink++; | 3379 | nr_unlink++; |
| 3380 | } | 3380 | } |
| 3381 | 3381 | ||
| 3382 | /* this will do delete_inode and everything for us */ | 3382 | /* this will do delete_inode and everything for us */ |
| 3383 | iput(inode); | 3383 | iput(inode); |
| 3384 | if (ret) | 3384 | if (ret) |
| 3385 | goto out; | 3385 | goto out; |
| 3386 | } | 3386 | } |
| 3387 | /* release the path since we're done with it */ | 3387 | /* release the path since we're done with it */ |
| 3388 | btrfs_release_path(path); | 3388 | btrfs_release_path(path); |
| 3389 | 3389 | ||
| 3390 | root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; | 3390 | root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; |
| 3391 | 3391 | ||
| 3392 | if (root->orphan_block_rsv) | 3392 | if (root->orphan_block_rsv) |
| 3393 | btrfs_block_rsv_release(root, root->orphan_block_rsv, | 3393 | btrfs_block_rsv_release(root, root->orphan_block_rsv, |
| 3394 | (u64)-1); | 3394 | (u64)-1); |
| 3395 | 3395 | ||
| 3396 | if (root->orphan_block_rsv || | 3396 | if (root->orphan_block_rsv || |
| 3397 | test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state)) { | 3397 | test_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state)) { |
| 3398 | trans = btrfs_join_transaction(root); | 3398 | trans = btrfs_join_transaction(root); |
| 3399 | if (!IS_ERR(trans)) | 3399 | if (!IS_ERR(trans)) |
| 3400 | btrfs_end_transaction(trans, root); | 3400 | btrfs_end_transaction(trans, root); |
| 3401 | } | 3401 | } |
| 3402 | 3402 | ||
| 3403 | if (nr_unlink) | 3403 | if (nr_unlink) |
| 3404 | btrfs_debug(root->fs_info, "unlinked %d orphans", nr_unlink); | 3404 | btrfs_debug(root->fs_info, "unlinked %d orphans", nr_unlink); |
| 3405 | if (nr_truncate) | 3405 | if (nr_truncate) |
| 3406 | btrfs_debug(root->fs_info, "truncated %d orphans", nr_truncate); | 3406 | btrfs_debug(root->fs_info, "truncated %d orphans", nr_truncate); |
| 3407 | 3407 | ||
| 3408 | out: | 3408 | out: |
| 3409 | if (ret) | 3409 | if (ret) |
| 3410 | btrfs_crit(root->fs_info, | 3410 | btrfs_crit(root->fs_info, |
| 3411 | "could not do orphan cleanup %d", ret); | 3411 | "could not do orphan cleanup %d", ret); |
| 3412 | btrfs_free_path(path); | 3412 | btrfs_free_path(path); |
| 3413 | return ret; | 3413 | return ret; |
| 3414 | } | 3414 | } |
| 3415 | 3415 | ||
| 3416 | /* | 3416 | /* |
| 3417 | * very simple check to peek ahead in the leaf looking for xattrs. If we | 3417 | * very simple check to peek ahead in the leaf looking for xattrs. If we |
| 3418 | * don't find any xattrs, we know there can't be any acls. | 3418 | * don't find any xattrs, we know there can't be any acls. |
| 3419 | * | 3419 | * |
| 3420 | * slot is the slot the inode is in, objectid is the objectid of the inode | 3420 | * slot is the slot the inode is in, objectid is the objectid of the inode |
| 3421 | */ | 3421 | */ |
| 3422 | static noinline int acls_after_inode_item(struct extent_buffer *leaf, | 3422 | static noinline int acls_after_inode_item(struct extent_buffer *leaf, |
| 3423 | int slot, u64 objectid, | 3423 | int slot, u64 objectid, |
| 3424 | int *first_xattr_slot) | 3424 | int *first_xattr_slot) |
| 3425 | { | 3425 | { |
| 3426 | u32 nritems = btrfs_header_nritems(leaf); | 3426 | u32 nritems = btrfs_header_nritems(leaf); |
| 3427 | struct btrfs_key found_key; | 3427 | struct btrfs_key found_key; |
| 3428 | static u64 xattr_access = 0; | 3428 | static u64 xattr_access = 0; |
| 3429 | static u64 xattr_default = 0; | 3429 | static u64 xattr_default = 0; |
| 3430 | int scanned = 0; | 3430 | int scanned = 0; |
| 3431 | 3431 | ||
| 3432 | if (!xattr_access) { | 3432 | if (!xattr_access) { |
| 3433 | xattr_access = btrfs_name_hash(POSIX_ACL_XATTR_ACCESS, | 3433 | xattr_access = btrfs_name_hash(POSIX_ACL_XATTR_ACCESS, |
| 3434 | strlen(POSIX_ACL_XATTR_ACCESS)); | 3434 | strlen(POSIX_ACL_XATTR_ACCESS)); |
| 3435 | xattr_default = btrfs_name_hash(POSIX_ACL_XATTR_DEFAULT, | 3435 | xattr_default = btrfs_name_hash(POSIX_ACL_XATTR_DEFAULT, |
| 3436 | strlen(POSIX_ACL_XATTR_DEFAULT)); | 3436 | strlen(POSIX_ACL_XATTR_DEFAULT)); |
| 3437 | } | 3437 | } |
| 3438 | 3438 | ||
| 3439 | slot++; | 3439 | slot++; |
| 3440 | *first_xattr_slot = -1; | 3440 | *first_xattr_slot = -1; |
| 3441 | while (slot < nritems) { | 3441 | while (slot < nritems) { |
| 3442 | btrfs_item_key_to_cpu(leaf, &found_key, slot); | 3442 | btrfs_item_key_to_cpu(leaf, &found_key, slot); |
| 3443 | 3443 | ||
| 3444 | /* we found a different objectid, there must not be acls */ | 3444 | /* we found a different objectid, there must not be acls */ |
| 3445 | if (found_key.objectid != objectid) | 3445 | if (found_key.objectid != objectid) |
| 3446 | return 0; | 3446 | return 0; |
| 3447 | 3447 | ||
| 3448 | /* we found an xattr, assume we've got an acl */ | 3448 | /* we found an xattr, assume we've got an acl */ |
| 3449 | if (found_key.type == BTRFS_XATTR_ITEM_KEY) { | 3449 | if (found_key.type == BTRFS_XATTR_ITEM_KEY) { |
| 3450 | if (*first_xattr_slot == -1) | 3450 | if (*first_xattr_slot == -1) |
| 3451 | *first_xattr_slot = slot; | 3451 | *first_xattr_slot = slot; |
| 3452 | if (found_key.offset == xattr_access || | 3452 | if (found_key.offset == xattr_access || |
| 3453 | found_key.offset == xattr_default) | 3453 | found_key.offset == xattr_default) |
| 3454 | return 1; | 3454 | return 1; |
| 3455 | } | 3455 | } |
| 3456 | 3456 | ||
| 3457 | /* | 3457 | /* |
| 3458 | * we found a key greater than an xattr key, there can't | 3458 | * we found a key greater than an xattr key, there can't |
| 3459 | * be any acls later on | 3459 | * be any acls later on |
| 3460 | */ | 3460 | */ |
| 3461 | if (found_key.type > BTRFS_XATTR_ITEM_KEY) | 3461 | if (found_key.type > BTRFS_XATTR_ITEM_KEY) |
| 3462 | return 0; | 3462 | return 0; |
| 3463 | 3463 | ||
| 3464 | slot++; | 3464 | slot++; |
| 3465 | scanned++; | 3465 | scanned++; |
| 3466 | 3466 | ||
| 3467 | /* | 3467 | /* |
| 3468 | * it goes inode, inode backrefs, xattrs, extents, | 3468 | * it goes inode, inode backrefs, xattrs, extents, |
| 3469 | * so if there are a ton of hard links to an inode there can | 3469 | * so if there are a ton of hard links to an inode there can |
| 3470 | * be a lot of backrefs. Don't waste time searching too hard, | 3470 | * be a lot of backrefs. Don't waste time searching too hard, |
| 3471 | * this is just an optimization | 3471 | * this is just an optimization |
| 3472 | */ | 3472 | */ |
| 3473 | if (scanned >= 8) | 3473 | if (scanned >= 8) |
| 3474 | break; | 3474 | break; |
| 3475 | } | 3475 | } |
| 3476 | /* we hit the end of the leaf before we found an xattr or | 3476 | /* we hit the end of the leaf before we found an xattr or |
| 3477 | * something larger than an xattr. We have to assume the inode | 3477 | * something larger than an xattr. We have to assume the inode |
| 3478 | * has acls | 3478 | * has acls |
| 3479 | */ | 3479 | */ |
| 3480 | if (*first_xattr_slot == -1) | 3480 | if (*first_xattr_slot == -1) |
| 3481 | *first_xattr_slot = slot; | 3481 | *first_xattr_slot = slot; |
| 3482 | return 1; | 3482 | return 1; |
| 3483 | } | 3483 | } |
| 3484 | 3484 | ||
| 3485 | /* | 3485 | /* |
| 3486 | * read an inode from the btree into the in-memory inode | 3486 | * read an inode from the btree into the in-memory inode |
| 3487 | */ | 3487 | */ |
| 3488 | static void btrfs_read_locked_inode(struct inode *inode) | 3488 | static void btrfs_read_locked_inode(struct inode *inode) |
| 3489 | { | 3489 | { |
| 3490 | struct btrfs_path *path; | 3490 | struct btrfs_path *path; |
| 3491 | struct extent_buffer *leaf; | 3491 | struct extent_buffer *leaf; |
| 3492 | struct btrfs_inode_item *inode_item; | 3492 | struct btrfs_inode_item *inode_item; |
| 3493 | struct btrfs_timespec *tspec; | 3493 | struct btrfs_timespec *tspec; |
| 3494 | struct btrfs_root *root = BTRFS_I(inode)->root; | 3494 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 3495 | struct btrfs_key location; | 3495 | struct btrfs_key location; |
| 3496 | unsigned long ptr; | 3496 | unsigned long ptr; |
| 3497 | int maybe_acls; | 3497 | int maybe_acls; |
| 3498 | u32 rdev; | 3498 | u32 rdev; |
| 3499 | int ret; | 3499 | int ret; |
| 3500 | bool filled = false; | 3500 | bool filled = false; |
| 3501 | int first_xattr_slot; | 3501 | int first_xattr_slot; |
| 3502 | 3502 | ||
| 3503 | ret = btrfs_fill_inode(inode, &rdev); | 3503 | ret = btrfs_fill_inode(inode, &rdev); |
| 3504 | if (!ret) | 3504 | if (!ret) |
| 3505 | filled = true; | 3505 | filled = true; |
| 3506 | 3506 | ||
| 3507 | path = btrfs_alloc_path(); | 3507 | path = btrfs_alloc_path(); |
| 3508 | if (!path) | 3508 | if (!path) |
| 3509 | goto make_bad; | 3509 | goto make_bad; |
| 3510 | 3510 | ||
| 3511 | memcpy(&location, &BTRFS_I(inode)->location, sizeof(location)); | 3511 | memcpy(&location, &BTRFS_I(inode)->location, sizeof(location)); |
| 3512 | 3512 | ||
| 3513 | ret = btrfs_lookup_inode(NULL, root, path, &location, 0); | 3513 | ret = btrfs_lookup_inode(NULL, root, path, &location, 0); |
| 3514 | if (ret) | 3514 | if (ret) |
| 3515 | goto make_bad; | 3515 | goto make_bad; |
| 3516 | 3516 | ||
| 3517 | leaf = path->nodes[0]; | 3517 | leaf = path->nodes[0]; |
| 3518 | 3518 | ||
| 3519 | if (filled) | 3519 | if (filled) |
| 3520 | goto cache_index; | 3520 | goto cache_index; |
| 3521 | 3521 | ||
| 3522 | inode_item = btrfs_item_ptr(leaf, path->slots[0], | 3522 | inode_item = btrfs_item_ptr(leaf, path->slots[0], |
| 3523 | struct btrfs_inode_item); | 3523 | struct btrfs_inode_item); |
| 3524 | inode->i_mode = btrfs_inode_mode(leaf, inode_item); | 3524 | inode->i_mode = btrfs_inode_mode(leaf, inode_item); |
| 3525 | set_nlink(inode, btrfs_inode_nlink(leaf, inode_item)); | 3525 | set_nlink(inode, btrfs_inode_nlink(leaf, inode_item)); |
| 3526 | i_uid_write(inode, btrfs_inode_uid(leaf, inode_item)); | 3526 | i_uid_write(inode, btrfs_inode_uid(leaf, inode_item)); |
| 3527 | i_gid_write(inode, btrfs_inode_gid(leaf, inode_item)); | 3527 | i_gid_write(inode, btrfs_inode_gid(leaf, inode_item)); |
| 3528 | btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item)); | 3528 | btrfs_i_size_write(inode, btrfs_inode_size(leaf, inode_item)); |
| 3529 | 3529 | ||
| 3530 | tspec = btrfs_inode_atime(inode_item); | 3530 | tspec = btrfs_inode_atime(inode_item); |
| 3531 | inode->i_atime.tv_sec = btrfs_timespec_sec(leaf, tspec); | 3531 | inode->i_atime.tv_sec = btrfs_timespec_sec(leaf, tspec); |
| 3532 | inode->i_atime.tv_nsec = btrfs_timespec_nsec(leaf, tspec); | 3532 | inode->i_atime.tv_nsec = btrfs_timespec_nsec(leaf, tspec); |
| 3533 | 3533 | ||
| 3534 | tspec = btrfs_inode_mtime(inode_item); | 3534 | tspec = btrfs_inode_mtime(inode_item); |
| 3535 | inode->i_mtime.tv_sec = btrfs_timespec_sec(leaf, tspec); | 3535 | inode->i_mtime.tv_sec = btrfs_timespec_sec(leaf, tspec); |
| 3536 | inode->i_mtime.tv_nsec = btrfs_timespec_nsec(leaf, tspec); | 3536 | inode->i_mtime.tv_nsec = btrfs_timespec_nsec(leaf, tspec); |
| 3537 | 3537 | ||
| 3538 | tspec = btrfs_inode_ctime(inode_item); | 3538 | tspec = btrfs_inode_ctime(inode_item); |
| 3539 | inode->i_ctime.tv_sec = btrfs_timespec_sec(leaf, tspec); | 3539 | inode->i_ctime.tv_sec = btrfs_timespec_sec(leaf, tspec); |
| 3540 | inode->i_ctime.tv_nsec = btrfs_timespec_nsec(leaf, tspec); | 3540 | inode->i_ctime.tv_nsec = btrfs_timespec_nsec(leaf, tspec); |
| 3541 | 3541 | ||
| 3542 | inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item)); | 3542 | inode_set_bytes(inode, btrfs_inode_nbytes(leaf, inode_item)); |
| 3543 | BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item); | 3543 | BTRFS_I(inode)->generation = btrfs_inode_generation(leaf, inode_item); |
| 3544 | BTRFS_I(inode)->last_trans = btrfs_inode_transid(leaf, inode_item); | 3544 | BTRFS_I(inode)->last_trans = btrfs_inode_transid(leaf, inode_item); |
| 3545 | 3545 | ||
| 3546 | /* | 3546 | /* |
| 3547 | * If we were modified in the current generation and evicted from memory | 3547 | * If we were modified in the current generation and evicted from memory |
| 3548 | * and then re-read we need to do a full sync since we don't have any | 3548 | * and then re-read we need to do a full sync since we don't have any |
| 3549 | * idea about which extents were modified before we were evicted from | 3549 | * idea about which extents were modified before we were evicted from |
| 3550 | * cache. | 3550 | * cache. |
| 3551 | */ | 3551 | */ |
| 3552 | if (BTRFS_I(inode)->last_trans == root->fs_info->generation) | 3552 | if (BTRFS_I(inode)->last_trans == root->fs_info->generation) |
| 3553 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, | 3553 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, |
| 3554 | &BTRFS_I(inode)->runtime_flags); | 3554 | &BTRFS_I(inode)->runtime_flags); |
| 3555 | 3555 | ||
| 3556 | inode->i_version = btrfs_inode_sequence(leaf, inode_item); | 3556 | inode->i_version = btrfs_inode_sequence(leaf, inode_item); |
| 3557 | inode->i_generation = BTRFS_I(inode)->generation; | 3557 | inode->i_generation = BTRFS_I(inode)->generation; |
| 3558 | inode->i_rdev = 0; | 3558 | inode->i_rdev = 0; |
| 3559 | rdev = btrfs_inode_rdev(leaf, inode_item); | 3559 | rdev = btrfs_inode_rdev(leaf, inode_item); |
| 3560 | 3560 | ||
| 3561 | BTRFS_I(inode)->index_cnt = (u64)-1; | 3561 | BTRFS_I(inode)->index_cnt = (u64)-1; |
| 3562 | BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item); | 3562 | BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item); |
| 3563 | 3563 | ||
| 3564 | cache_index: | 3564 | cache_index: |
| 3565 | path->slots[0]++; | 3565 | path->slots[0]++; |
| 3566 | if (inode->i_nlink != 1 || | 3566 | if (inode->i_nlink != 1 || |
| 3567 | path->slots[0] >= btrfs_header_nritems(leaf)) | 3567 | path->slots[0] >= btrfs_header_nritems(leaf)) |
| 3568 | goto cache_acl; | 3568 | goto cache_acl; |
| 3569 | 3569 | ||
| 3570 | btrfs_item_key_to_cpu(leaf, &location, path->slots[0]); | 3570 | btrfs_item_key_to_cpu(leaf, &location, path->slots[0]); |
| 3571 | if (location.objectid != btrfs_ino(inode)) | 3571 | if (location.objectid != btrfs_ino(inode)) |
| 3572 | goto cache_acl; | 3572 | goto cache_acl; |
| 3573 | 3573 | ||
| 3574 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); | 3574 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); |
| 3575 | if (location.type == BTRFS_INODE_REF_KEY) { | 3575 | if (location.type == BTRFS_INODE_REF_KEY) { |
| 3576 | struct btrfs_inode_ref *ref; | 3576 | struct btrfs_inode_ref *ref; |
| 3577 | 3577 | ||
| 3578 | ref = (struct btrfs_inode_ref *)ptr; | 3578 | ref = (struct btrfs_inode_ref *)ptr; |
| 3579 | BTRFS_I(inode)->dir_index = btrfs_inode_ref_index(leaf, ref); | 3579 | BTRFS_I(inode)->dir_index = btrfs_inode_ref_index(leaf, ref); |
| 3580 | } else if (location.type == BTRFS_INODE_EXTREF_KEY) { | 3580 | } else if (location.type == BTRFS_INODE_EXTREF_KEY) { |
| 3581 | struct btrfs_inode_extref *extref; | 3581 | struct btrfs_inode_extref *extref; |
| 3582 | 3582 | ||
| 3583 | extref = (struct btrfs_inode_extref *)ptr; | 3583 | extref = (struct btrfs_inode_extref *)ptr; |
| 3584 | BTRFS_I(inode)->dir_index = btrfs_inode_extref_index(leaf, | 3584 | BTRFS_I(inode)->dir_index = btrfs_inode_extref_index(leaf, |
| 3585 | extref); | 3585 | extref); |
| 3586 | } | 3586 | } |
| 3587 | cache_acl: | 3587 | cache_acl: |
| 3588 | /* | 3588 | /* |
| 3589 | * try to precache a NULL acl entry for files that don't have | 3589 | * try to precache a NULL acl entry for files that don't have |
| 3590 | * any xattrs or acls | 3590 | * any xattrs or acls |
| 3591 | */ | 3591 | */ |
| 3592 | maybe_acls = acls_after_inode_item(leaf, path->slots[0], | 3592 | maybe_acls = acls_after_inode_item(leaf, path->slots[0], |
| 3593 | btrfs_ino(inode), &first_xattr_slot); | 3593 | btrfs_ino(inode), &first_xattr_slot); |
| 3594 | if (first_xattr_slot != -1) { | 3594 | if (first_xattr_slot != -1) { |
| 3595 | path->slots[0] = first_xattr_slot; | 3595 | path->slots[0] = first_xattr_slot; |
| 3596 | ret = btrfs_load_inode_props(inode, path); | 3596 | ret = btrfs_load_inode_props(inode, path); |
| 3597 | if (ret) | 3597 | if (ret) |
| 3598 | btrfs_err(root->fs_info, | 3598 | btrfs_err(root->fs_info, |
| 3599 | "error loading props for ino %llu (root %llu): %d", | 3599 | "error loading props for ino %llu (root %llu): %d", |
| 3600 | btrfs_ino(inode), | 3600 | btrfs_ino(inode), |
| 3601 | root->root_key.objectid, ret); | 3601 | root->root_key.objectid, ret); |
| 3602 | } | 3602 | } |
| 3603 | btrfs_free_path(path); | 3603 | btrfs_free_path(path); |
| 3604 | 3604 | ||
| 3605 | if (!maybe_acls) | 3605 | if (!maybe_acls) |
| 3606 | cache_no_acl(inode); | 3606 | cache_no_acl(inode); |
| 3607 | 3607 | ||
| 3608 | switch (inode->i_mode & S_IFMT) { | 3608 | switch (inode->i_mode & S_IFMT) { |
| 3609 | case S_IFREG: | 3609 | case S_IFREG: |
| 3610 | inode->i_mapping->a_ops = &btrfs_aops; | 3610 | inode->i_mapping->a_ops = &btrfs_aops; |
| 3611 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; | 3611 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; |
| 3612 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; | 3612 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; |
| 3613 | inode->i_fop = &btrfs_file_operations; | 3613 | inode->i_fop = &btrfs_file_operations; |
| 3614 | inode->i_op = &btrfs_file_inode_operations; | 3614 | inode->i_op = &btrfs_file_inode_operations; |
| 3615 | break; | 3615 | break; |
| 3616 | case S_IFDIR: | 3616 | case S_IFDIR: |
| 3617 | inode->i_fop = &btrfs_dir_file_operations; | 3617 | inode->i_fop = &btrfs_dir_file_operations; |
| 3618 | if (root == root->fs_info->tree_root) | 3618 | if (root == root->fs_info->tree_root) |
| 3619 | inode->i_op = &btrfs_dir_ro_inode_operations; | 3619 | inode->i_op = &btrfs_dir_ro_inode_operations; |
| 3620 | else | 3620 | else |
| 3621 | inode->i_op = &btrfs_dir_inode_operations; | 3621 | inode->i_op = &btrfs_dir_inode_operations; |
| 3622 | break; | 3622 | break; |
| 3623 | case S_IFLNK: | 3623 | case S_IFLNK: |
| 3624 | inode->i_op = &btrfs_symlink_inode_operations; | 3624 | inode->i_op = &btrfs_symlink_inode_operations; |
| 3625 | inode->i_mapping->a_ops = &btrfs_symlink_aops; | 3625 | inode->i_mapping->a_ops = &btrfs_symlink_aops; |
| 3626 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; | 3626 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; |
| 3627 | break; | 3627 | break; |
| 3628 | default: | 3628 | default: |
| 3629 | inode->i_op = &btrfs_special_inode_operations; | 3629 | inode->i_op = &btrfs_special_inode_operations; |
| 3630 | init_special_inode(inode, inode->i_mode, rdev); | 3630 | init_special_inode(inode, inode->i_mode, rdev); |
| 3631 | break; | 3631 | break; |
| 3632 | } | 3632 | } |
| 3633 | 3633 | ||
| 3634 | btrfs_update_iflags(inode); | 3634 | btrfs_update_iflags(inode); |
| 3635 | return; | 3635 | return; |
| 3636 | 3636 | ||
| 3637 | make_bad: | 3637 | make_bad: |
| 3638 | btrfs_free_path(path); | 3638 | btrfs_free_path(path); |
| 3639 | make_bad_inode(inode); | 3639 | make_bad_inode(inode); |
| 3640 | } | 3640 | } |
| 3641 | 3641 | ||
| 3642 | /* | 3642 | /* |
| 3643 | * given a leaf and an inode, copy the inode fields into the leaf | 3643 | * given a leaf and an inode, copy the inode fields into the leaf |
| 3644 | */ | 3644 | */ |
| 3645 | static void fill_inode_item(struct btrfs_trans_handle *trans, | 3645 | static void fill_inode_item(struct btrfs_trans_handle *trans, |
| 3646 | struct extent_buffer *leaf, | 3646 | struct extent_buffer *leaf, |
| 3647 | struct btrfs_inode_item *item, | 3647 | struct btrfs_inode_item *item, |
| 3648 | struct inode *inode) | 3648 | struct inode *inode) |
| 3649 | { | 3649 | { |
| 3650 | struct btrfs_map_token token; | 3650 | struct btrfs_map_token token; |
| 3651 | 3651 | ||
| 3652 | btrfs_init_map_token(&token); | 3652 | btrfs_init_map_token(&token); |
| 3653 | 3653 | ||
| 3654 | btrfs_set_token_inode_uid(leaf, item, i_uid_read(inode), &token); | 3654 | btrfs_set_token_inode_uid(leaf, item, i_uid_read(inode), &token); |
| 3655 | btrfs_set_token_inode_gid(leaf, item, i_gid_read(inode), &token); | 3655 | btrfs_set_token_inode_gid(leaf, item, i_gid_read(inode), &token); |
| 3656 | btrfs_set_token_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size, | 3656 | btrfs_set_token_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size, |
| 3657 | &token); | 3657 | &token); |
| 3658 | btrfs_set_token_inode_mode(leaf, item, inode->i_mode, &token); | 3658 | btrfs_set_token_inode_mode(leaf, item, inode->i_mode, &token); |
| 3659 | btrfs_set_token_inode_nlink(leaf, item, inode->i_nlink, &token); | 3659 | btrfs_set_token_inode_nlink(leaf, item, inode->i_nlink, &token); |
| 3660 | 3660 | ||
| 3661 | btrfs_set_token_timespec_sec(leaf, btrfs_inode_atime(item), | 3661 | btrfs_set_token_timespec_sec(leaf, btrfs_inode_atime(item), |
| 3662 | inode->i_atime.tv_sec, &token); | 3662 | inode->i_atime.tv_sec, &token); |
| 3663 | btrfs_set_token_timespec_nsec(leaf, btrfs_inode_atime(item), | 3663 | btrfs_set_token_timespec_nsec(leaf, btrfs_inode_atime(item), |
| 3664 | inode->i_atime.tv_nsec, &token); | 3664 | inode->i_atime.tv_nsec, &token); |
| 3665 | 3665 | ||
| 3666 | btrfs_set_token_timespec_sec(leaf, btrfs_inode_mtime(item), | 3666 | btrfs_set_token_timespec_sec(leaf, btrfs_inode_mtime(item), |
| 3667 | inode->i_mtime.tv_sec, &token); | 3667 | inode->i_mtime.tv_sec, &token); |
| 3668 | btrfs_set_token_timespec_nsec(leaf, btrfs_inode_mtime(item), | 3668 | btrfs_set_token_timespec_nsec(leaf, btrfs_inode_mtime(item), |
| 3669 | inode->i_mtime.tv_nsec, &token); | 3669 | inode->i_mtime.tv_nsec, &token); |
| 3670 | 3670 | ||
| 3671 | btrfs_set_token_timespec_sec(leaf, btrfs_inode_ctime(item), | 3671 | btrfs_set_token_timespec_sec(leaf, btrfs_inode_ctime(item), |
| 3672 | inode->i_ctime.tv_sec, &token); | 3672 | inode->i_ctime.tv_sec, &token); |
| 3673 | btrfs_set_token_timespec_nsec(leaf, btrfs_inode_ctime(item), | 3673 | btrfs_set_token_timespec_nsec(leaf, btrfs_inode_ctime(item), |
| 3674 | inode->i_ctime.tv_nsec, &token); | 3674 | inode->i_ctime.tv_nsec, &token); |
| 3675 | 3675 | ||
| 3676 | btrfs_set_token_inode_nbytes(leaf, item, inode_get_bytes(inode), | 3676 | btrfs_set_token_inode_nbytes(leaf, item, inode_get_bytes(inode), |
| 3677 | &token); | 3677 | &token); |
| 3678 | btrfs_set_token_inode_generation(leaf, item, BTRFS_I(inode)->generation, | 3678 | btrfs_set_token_inode_generation(leaf, item, BTRFS_I(inode)->generation, |
| 3679 | &token); | 3679 | &token); |
| 3680 | btrfs_set_token_inode_sequence(leaf, item, inode->i_version, &token); | 3680 | btrfs_set_token_inode_sequence(leaf, item, inode->i_version, &token); |
| 3681 | btrfs_set_token_inode_transid(leaf, item, trans->transid, &token); | 3681 | btrfs_set_token_inode_transid(leaf, item, trans->transid, &token); |
| 3682 | btrfs_set_token_inode_rdev(leaf, item, inode->i_rdev, &token); | 3682 | btrfs_set_token_inode_rdev(leaf, item, inode->i_rdev, &token); |
| 3683 | btrfs_set_token_inode_flags(leaf, item, BTRFS_I(inode)->flags, &token); | 3683 | btrfs_set_token_inode_flags(leaf, item, BTRFS_I(inode)->flags, &token); |
| 3684 | btrfs_set_token_inode_block_group(leaf, item, 0, &token); | 3684 | btrfs_set_token_inode_block_group(leaf, item, 0, &token); |
| 3685 | } | 3685 | } |
| 3686 | 3686 | ||
| 3687 | /* | 3687 | /* |
| 3688 | * copy everything in the in-memory inode into the btree. | 3688 | * copy everything in the in-memory inode into the btree. |
| 3689 | */ | 3689 | */ |
| 3690 | static noinline int btrfs_update_inode_item(struct btrfs_trans_handle *trans, | 3690 | static noinline int btrfs_update_inode_item(struct btrfs_trans_handle *trans, |
| 3691 | struct btrfs_root *root, struct inode *inode) | 3691 | struct btrfs_root *root, struct inode *inode) |
| 3692 | { | 3692 | { |
| 3693 | struct btrfs_inode_item *inode_item; | 3693 | struct btrfs_inode_item *inode_item; |
| 3694 | struct btrfs_path *path; | 3694 | struct btrfs_path *path; |
| 3695 | struct extent_buffer *leaf; | 3695 | struct extent_buffer *leaf; |
| 3696 | int ret; | 3696 | int ret; |
| 3697 | 3697 | ||
| 3698 | path = btrfs_alloc_path(); | 3698 | path = btrfs_alloc_path(); |
| 3699 | if (!path) | 3699 | if (!path) |
| 3700 | return -ENOMEM; | 3700 | return -ENOMEM; |
| 3701 | 3701 | ||
| 3702 | path->leave_spinning = 1; | 3702 | path->leave_spinning = 1; |
| 3703 | ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location, | 3703 | ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location, |
| 3704 | 1); | 3704 | 1); |
| 3705 | if (ret) { | 3705 | if (ret) { |
| 3706 | if (ret > 0) | 3706 | if (ret > 0) |
| 3707 | ret = -ENOENT; | 3707 | ret = -ENOENT; |
| 3708 | goto failed; | 3708 | goto failed; |
| 3709 | } | 3709 | } |
| 3710 | 3710 | ||
| 3711 | leaf = path->nodes[0]; | 3711 | leaf = path->nodes[0]; |
| 3712 | inode_item = btrfs_item_ptr(leaf, path->slots[0], | 3712 | inode_item = btrfs_item_ptr(leaf, path->slots[0], |
| 3713 | struct btrfs_inode_item); | 3713 | struct btrfs_inode_item); |
| 3714 | 3714 | ||
| 3715 | fill_inode_item(trans, leaf, inode_item, inode); | 3715 | fill_inode_item(trans, leaf, inode_item, inode); |
| 3716 | btrfs_mark_buffer_dirty(leaf); | 3716 | btrfs_mark_buffer_dirty(leaf); |
| 3717 | btrfs_set_inode_last_trans(trans, inode); | 3717 | btrfs_set_inode_last_trans(trans, inode); |
| 3718 | ret = 0; | 3718 | ret = 0; |
| 3719 | failed: | 3719 | failed: |
| 3720 | btrfs_free_path(path); | 3720 | btrfs_free_path(path); |
| 3721 | return ret; | 3721 | return ret; |
| 3722 | } | 3722 | } |
| 3723 | 3723 | ||
| 3724 | /* | 3724 | /* |
| 3725 | * copy everything in the in-memory inode into the btree. | 3725 | * copy everything in the in-memory inode into the btree. |
| 3726 | */ | 3726 | */ |
| 3727 | noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, | 3727 | noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, |
| 3728 | struct btrfs_root *root, struct inode *inode) | 3728 | struct btrfs_root *root, struct inode *inode) |
| 3729 | { | 3729 | { |
| 3730 | int ret; | 3730 | int ret; |
| 3731 | 3731 | ||
| 3732 | /* | 3732 | /* |
| 3733 | * If the inode is a free space inode, we can deadlock during commit | 3733 | * If the inode is a free space inode, we can deadlock during commit |
| 3734 | * if we put it into the delayed code. | 3734 | * if we put it into the delayed code. |
| 3735 | * | 3735 | * |
| 3736 | * The data relocation inode should also be directly updated | 3736 | * The data relocation inode should also be directly updated |
| 3737 | * without delay | 3737 | * without delay |
| 3738 | */ | 3738 | */ |
| 3739 | if (!btrfs_is_free_space_inode(inode) | 3739 | if (!btrfs_is_free_space_inode(inode) |
| 3740 | && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID | 3740 | && root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID |
| 3741 | && !root->fs_info->log_root_recovering) { | 3741 | && !root->fs_info->log_root_recovering) { |
| 3742 | btrfs_update_root_times(trans, root); | 3742 | btrfs_update_root_times(trans, root); |
| 3743 | 3743 | ||
| 3744 | ret = btrfs_delayed_update_inode(trans, root, inode); | 3744 | ret = btrfs_delayed_update_inode(trans, root, inode); |
| 3745 | if (!ret) | 3745 | if (!ret) |
| 3746 | btrfs_set_inode_last_trans(trans, inode); | 3746 | btrfs_set_inode_last_trans(trans, inode); |
| 3747 | return ret; | 3747 | return ret; |
| 3748 | } | 3748 | } |
| 3749 | 3749 | ||
| 3750 | return btrfs_update_inode_item(trans, root, inode); | 3750 | return btrfs_update_inode_item(trans, root, inode); |
| 3751 | } | 3751 | } |
| 3752 | 3752 | ||
| 3753 | noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, | 3753 | noinline int btrfs_update_inode_fallback(struct btrfs_trans_handle *trans, |
| 3754 | struct btrfs_root *root, | 3754 | struct btrfs_root *root, |
| 3755 | struct inode *inode) | 3755 | struct inode *inode) |
| 3756 | { | 3756 | { |
| 3757 | int ret; | 3757 | int ret; |
| 3758 | 3758 | ||
| 3759 | ret = btrfs_update_inode(trans, root, inode); | 3759 | ret = btrfs_update_inode(trans, root, inode); |
| 3760 | if (ret == -ENOSPC) | 3760 | if (ret == -ENOSPC) |
| 3761 | return btrfs_update_inode_item(trans, root, inode); | 3761 | return btrfs_update_inode_item(trans, root, inode); |
| 3762 | return ret; | 3762 | return ret; |
| 3763 | } | 3763 | } |
| 3764 | 3764 | ||
| 3765 | /* | 3765 | /* |
| 3766 | * unlink helper that gets used here in inode.c and in the tree logging | 3766 | * unlink helper that gets used here in inode.c and in the tree logging |
| 3767 | * recovery code. It remove a link in a directory with a given name, and | 3767 | * recovery code. It remove a link in a directory with a given name, and |
| 3768 | * also drops the back refs in the inode to the directory | 3768 | * also drops the back refs in the inode to the directory |
| 3769 | */ | 3769 | */ |
| 3770 | static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, | 3770 | static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, |
| 3771 | struct btrfs_root *root, | 3771 | struct btrfs_root *root, |
| 3772 | struct inode *dir, struct inode *inode, | 3772 | struct inode *dir, struct inode *inode, |
| 3773 | const char *name, int name_len) | 3773 | const char *name, int name_len) |
| 3774 | { | 3774 | { |
| 3775 | struct btrfs_path *path; | 3775 | struct btrfs_path *path; |
| 3776 | int ret = 0; | 3776 | int ret = 0; |
| 3777 | struct extent_buffer *leaf; | 3777 | struct extent_buffer *leaf; |
| 3778 | struct btrfs_dir_item *di; | 3778 | struct btrfs_dir_item *di; |
| 3779 | struct btrfs_key key; | 3779 | struct btrfs_key key; |
| 3780 | u64 index; | 3780 | u64 index; |
| 3781 | u64 ino = btrfs_ino(inode); | 3781 | u64 ino = btrfs_ino(inode); |
| 3782 | u64 dir_ino = btrfs_ino(dir); | 3782 | u64 dir_ino = btrfs_ino(dir); |
| 3783 | 3783 | ||
| 3784 | path = btrfs_alloc_path(); | 3784 | path = btrfs_alloc_path(); |
| 3785 | if (!path) { | 3785 | if (!path) { |
| 3786 | ret = -ENOMEM; | 3786 | ret = -ENOMEM; |
| 3787 | goto out; | 3787 | goto out; |
| 3788 | } | 3788 | } |
| 3789 | 3789 | ||
| 3790 | path->leave_spinning = 1; | 3790 | path->leave_spinning = 1; |
| 3791 | di = btrfs_lookup_dir_item(trans, root, path, dir_ino, | 3791 | di = btrfs_lookup_dir_item(trans, root, path, dir_ino, |
| 3792 | name, name_len, -1); | 3792 | name, name_len, -1); |
| 3793 | if (IS_ERR(di)) { | 3793 | if (IS_ERR(di)) { |
| 3794 | ret = PTR_ERR(di); | 3794 | ret = PTR_ERR(di); |
| 3795 | goto err; | 3795 | goto err; |
| 3796 | } | 3796 | } |
| 3797 | if (!di) { | 3797 | if (!di) { |
| 3798 | ret = -ENOENT; | 3798 | ret = -ENOENT; |
| 3799 | goto err; | 3799 | goto err; |
| 3800 | } | 3800 | } |
| 3801 | leaf = path->nodes[0]; | 3801 | leaf = path->nodes[0]; |
| 3802 | btrfs_dir_item_key_to_cpu(leaf, di, &key); | 3802 | btrfs_dir_item_key_to_cpu(leaf, di, &key); |
| 3803 | ret = btrfs_delete_one_dir_name(trans, root, path, di); | 3803 | ret = btrfs_delete_one_dir_name(trans, root, path, di); |
| 3804 | if (ret) | 3804 | if (ret) |
| 3805 | goto err; | 3805 | goto err; |
| 3806 | btrfs_release_path(path); | 3806 | btrfs_release_path(path); |
| 3807 | 3807 | ||
| 3808 | /* | 3808 | /* |
| 3809 | * If we don't have dir index, we have to get it by looking up | 3809 | * If we don't have dir index, we have to get it by looking up |
| 3810 | * the inode ref, since we get the inode ref, remove it directly, | 3810 | * the inode ref, since we get the inode ref, remove it directly, |
| 3811 | * it is unnecessary to do delayed deletion. | 3811 | * it is unnecessary to do delayed deletion. |
| 3812 | * | 3812 | * |
| 3813 | * But if we have dir index, needn't search inode ref to get it. | 3813 | * But if we have dir index, needn't search inode ref to get it. |
| 3814 | * Since the inode ref is close to the inode item, it is better | 3814 | * Since the inode ref is close to the inode item, it is better |
| 3815 | * that we delay to delete it, and just do this deletion when | 3815 | * that we delay to delete it, and just do this deletion when |
| 3816 | * we update the inode item. | 3816 | * we update the inode item. |
| 3817 | */ | 3817 | */ |
| 3818 | if (BTRFS_I(inode)->dir_index) { | 3818 | if (BTRFS_I(inode)->dir_index) { |
| 3819 | ret = btrfs_delayed_delete_inode_ref(inode); | 3819 | ret = btrfs_delayed_delete_inode_ref(inode); |
| 3820 | if (!ret) { | 3820 | if (!ret) { |
| 3821 | index = BTRFS_I(inode)->dir_index; | 3821 | index = BTRFS_I(inode)->dir_index; |
| 3822 | goto skip_backref; | 3822 | goto skip_backref; |
| 3823 | } | 3823 | } |
| 3824 | } | 3824 | } |
| 3825 | 3825 | ||
| 3826 | ret = btrfs_del_inode_ref(trans, root, name, name_len, ino, | 3826 | ret = btrfs_del_inode_ref(trans, root, name, name_len, ino, |
| 3827 | dir_ino, &index); | 3827 | dir_ino, &index); |
| 3828 | if (ret) { | 3828 | if (ret) { |
| 3829 | btrfs_info(root->fs_info, | 3829 | btrfs_info(root->fs_info, |
| 3830 | "failed to delete reference to %.*s, inode %llu parent %llu", | 3830 | "failed to delete reference to %.*s, inode %llu parent %llu", |
| 3831 | name_len, name, ino, dir_ino); | 3831 | name_len, name, ino, dir_ino); |
| 3832 | btrfs_abort_transaction(trans, root, ret); | 3832 | btrfs_abort_transaction(trans, root, ret); |
| 3833 | goto err; | 3833 | goto err; |
| 3834 | } | 3834 | } |
| 3835 | skip_backref: | 3835 | skip_backref: |
| 3836 | ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); | 3836 | ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); |
| 3837 | if (ret) { | 3837 | if (ret) { |
| 3838 | btrfs_abort_transaction(trans, root, ret); | 3838 | btrfs_abort_transaction(trans, root, ret); |
| 3839 | goto err; | 3839 | goto err; |
| 3840 | } | 3840 | } |
| 3841 | 3841 | ||
| 3842 | ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, | 3842 | ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, |
| 3843 | inode, dir_ino); | 3843 | inode, dir_ino); |
| 3844 | if (ret != 0 && ret != -ENOENT) { | 3844 | if (ret != 0 && ret != -ENOENT) { |
| 3845 | btrfs_abort_transaction(trans, root, ret); | 3845 | btrfs_abort_transaction(trans, root, ret); |
| 3846 | goto err; | 3846 | goto err; |
| 3847 | } | 3847 | } |
| 3848 | 3848 | ||
| 3849 | ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len, | 3849 | ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len, |
| 3850 | dir, index); | 3850 | dir, index); |
| 3851 | if (ret == -ENOENT) | 3851 | if (ret == -ENOENT) |
| 3852 | ret = 0; | 3852 | ret = 0; |
| 3853 | else if (ret) | 3853 | else if (ret) |
| 3854 | btrfs_abort_transaction(trans, root, ret); | 3854 | btrfs_abort_transaction(trans, root, ret); |
| 3855 | err: | 3855 | err: |
| 3856 | btrfs_free_path(path); | 3856 | btrfs_free_path(path); |
| 3857 | if (ret) | 3857 | if (ret) |
| 3858 | goto out; | 3858 | goto out; |
| 3859 | 3859 | ||
| 3860 | btrfs_i_size_write(dir, dir->i_size - name_len * 2); | 3860 | btrfs_i_size_write(dir, dir->i_size - name_len * 2); |
| 3861 | inode_inc_iversion(inode); | 3861 | inode_inc_iversion(inode); |
| 3862 | inode_inc_iversion(dir); | 3862 | inode_inc_iversion(dir); |
| 3863 | inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME; | 3863 | inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME; |
| 3864 | ret = btrfs_update_inode(trans, root, dir); | 3864 | ret = btrfs_update_inode(trans, root, dir); |
| 3865 | out: | 3865 | out: |
| 3866 | return ret; | 3866 | return ret; |
| 3867 | } | 3867 | } |
| 3868 | 3868 | ||
| 3869 | int btrfs_unlink_inode(struct btrfs_trans_handle *trans, | 3869 | int btrfs_unlink_inode(struct btrfs_trans_handle *trans, |
| 3870 | struct btrfs_root *root, | 3870 | struct btrfs_root *root, |
| 3871 | struct inode *dir, struct inode *inode, | 3871 | struct inode *dir, struct inode *inode, |
| 3872 | const char *name, int name_len) | 3872 | const char *name, int name_len) |
| 3873 | { | 3873 | { |
| 3874 | int ret; | 3874 | int ret; |
| 3875 | ret = __btrfs_unlink_inode(trans, root, dir, inode, name, name_len); | 3875 | ret = __btrfs_unlink_inode(trans, root, dir, inode, name, name_len); |
| 3876 | if (!ret) { | 3876 | if (!ret) { |
| 3877 | drop_nlink(inode); | 3877 | drop_nlink(inode); |
| 3878 | ret = btrfs_update_inode(trans, root, inode); | 3878 | ret = btrfs_update_inode(trans, root, inode); |
| 3879 | } | 3879 | } |
| 3880 | return ret; | 3880 | return ret; |
| 3881 | } | 3881 | } |
| 3882 | 3882 | ||
| 3883 | /* | 3883 | /* |
| 3884 | * helper to start transaction for unlink and rmdir. | 3884 | * helper to start transaction for unlink and rmdir. |
| 3885 | * | 3885 | * |
| 3886 | * unlink and rmdir are special in btrfs, they do not always free space, so | 3886 | * unlink and rmdir are special in btrfs, they do not always free space, so |
| 3887 | * if we cannot make our reservations the normal way try and see if there is | 3887 | * if we cannot make our reservations the normal way try and see if there is |
| 3888 | * plenty of slack room in the global reserve to migrate, otherwise we cannot | 3888 | * plenty of slack room in the global reserve to migrate, otherwise we cannot |
| 3889 | * allow the unlink to occur. | 3889 | * allow the unlink to occur. |
| 3890 | */ | 3890 | */ |
| 3891 | static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir) | 3891 | static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir) |
| 3892 | { | 3892 | { |
| 3893 | struct btrfs_trans_handle *trans; | 3893 | struct btrfs_trans_handle *trans; |
| 3894 | struct btrfs_root *root = BTRFS_I(dir)->root; | 3894 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 3895 | int ret; | 3895 | int ret; |
| 3896 | 3896 | ||
| 3897 | /* | 3897 | /* |
| 3898 | * 1 for the possible orphan item | 3898 | * 1 for the possible orphan item |
| 3899 | * 1 for the dir item | 3899 | * 1 for the dir item |
| 3900 | * 1 for the dir index | 3900 | * 1 for the dir index |
| 3901 | * 1 for the inode ref | 3901 | * 1 for the inode ref |
| 3902 | * 1 for the inode | 3902 | * 1 for the inode |
| 3903 | */ | 3903 | */ |
| 3904 | trans = btrfs_start_transaction(root, 5); | 3904 | trans = btrfs_start_transaction(root, 5); |
| 3905 | if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) | 3905 | if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) |
| 3906 | return trans; | 3906 | return trans; |
| 3907 | 3907 | ||
| 3908 | if (PTR_ERR(trans) == -ENOSPC) { | 3908 | if (PTR_ERR(trans) == -ENOSPC) { |
| 3909 | u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5); | 3909 | u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5); |
| 3910 | 3910 | ||
| 3911 | trans = btrfs_start_transaction(root, 0); | 3911 | trans = btrfs_start_transaction(root, 0); |
| 3912 | if (IS_ERR(trans)) | 3912 | if (IS_ERR(trans)) |
| 3913 | return trans; | 3913 | return trans; |
| 3914 | ret = btrfs_cond_migrate_bytes(root->fs_info, | 3914 | ret = btrfs_cond_migrate_bytes(root->fs_info, |
| 3915 | &root->fs_info->trans_block_rsv, | 3915 | &root->fs_info->trans_block_rsv, |
| 3916 | num_bytes, 5); | 3916 | num_bytes, 5); |
| 3917 | if (ret) { | 3917 | if (ret) { |
| 3918 | btrfs_end_transaction(trans, root); | 3918 | btrfs_end_transaction(trans, root); |
| 3919 | return ERR_PTR(ret); | 3919 | return ERR_PTR(ret); |
| 3920 | } | 3920 | } |
| 3921 | trans->block_rsv = &root->fs_info->trans_block_rsv; | 3921 | trans->block_rsv = &root->fs_info->trans_block_rsv; |
| 3922 | trans->bytes_reserved = num_bytes; | 3922 | trans->bytes_reserved = num_bytes; |
| 3923 | } | 3923 | } |
| 3924 | return trans; | 3924 | return trans; |
| 3925 | } | 3925 | } |
| 3926 | 3926 | ||
| 3927 | static int btrfs_unlink(struct inode *dir, struct dentry *dentry) | 3927 | static int btrfs_unlink(struct inode *dir, struct dentry *dentry) |
| 3928 | { | 3928 | { |
| 3929 | struct btrfs_root *root = BTRFS_I(dir)->root; | 3929 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 3930 | struct btrfs_trans_handle *trans; | 3930 | struct btrfs_trans_handle *trans; |
| 3931 | struct inode *inode = dentry->d_inode; | 3931 | struct inode *inode = dentry->d_inode; |
| 3932 | int ret; | 3932 | int ret; |
| 3933 | 3933 | ||
| 3934 | trans = __unlink_start_trans(dir); | 3934 | trans = __unlink_start_trans(dir); |
| 3935 | if (IS_ERR(trans)) | 3935 | if (IS_ERR(trans)) |
| 3936 | return PTR_ERR(trans); | 3936 | return PTR_ERR(trans); |
| 3937 | 3937 | ||
| 3938 | btrfs_record_unlink_dir(trans, dir, dentry->d_inode, 0); | 3938 | btrfs_record_unlink_dir(trans, dir, dentry->d_inode, 0); |
| 3939 | 3939 | ||
| 3940 | ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode, | 3940 | ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode, |
| 3941 | dentry->d_name.name, dentry->d_name.len); | 3941 | dentry->d_name.name, dentry->d_name.len); |
| 3942 | if (ret) | 3942 | if (ret) |
| 3943 | goto out; | 3943 | goto out; |
| 3944 | 3944 | ||
| 3945 | if (inode->i_nlink == 0) { | 3945 | if (inode->i_nlink == 0) { |
| 3946 | ret = btrfs_orphan_add(trans, inode); | 3946 | ret = btrfs_orphan_add(trans, inode); |
| 3947 | if (ret) | 3947 | if (ret) |
| 3948 | goto out; | 3948 | goto out; |
| 3949 | } | 3949 | } |
| 3950 | 3950 | ||
| 3951 | out: | 3951 | out: |
| 3952 | btrfs_end_transaction(trans, root); | 3952 | btrfs_end_transaction(trans, root); |
| 3953 | btrfs_btree_balance_dirty(root); | 3953 | btrfs_btree_balance_dirty(root); |
| 3954 | return ret; | 3954 | return ret; |
| 3955 | } | 3955 | } |
| 3956 | 3956 | ||
| 3957 | int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, | 3957 | int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, |
| 3958 | struct btrfs_root *root, | 3958 | struct btrfs_root *root, |
| 3959 | struct inode *dir, u64 objectid, | 3959 | struct inode *dir, u64 objectid, |
| 3960 | const char *name, int name_len) | 3960 | const char *name, int name_len) |
| 3961 | { | 3961 | { |
| 3962 | struct btrfs_path *path; | 3962 | struct btrfs_path *path; |
| 3963 | struct extent_buffer *leaf; | 3963 | struct extent_buffer *leaf; |
| 3964 | struct btrfs_dir_item *di; | 3964 | struct btrfs_dir_item *di; |
| 3965 | struct btrfs_key key; | 3965 | struct btrfs_key key; |
| 3966 | u64 index; | 3966 | u64 index; |
| 3967 | int ret; | 3967 | int ret; |
| 3968 | u64 dir_ino = btrfs_ino(dir); | 3968 | u64 dir_ino = btrfs_ino(dir); |
| 3969 | 3969 | ||
| 3970 | path = btrfs_alloc_path(); | 3970 | path = btrfs_alloc_path(); |
| 3971 | if (!path) | 3971 | if (!path) |
| 3972 | return -ENOMEM; | 3972 | return -ENOMEM; |
| 3973 | 3973 | ||
| 3974 | di = btrfs_lookup_dir_item(trans, root, path, dir_ino, | 3974 | di = btrfs_lookup_dir_item(trans, root, path, dir_ino, |
| 3975 | name, name_len, -1); | 3975 | name, name_len, -1); |
| 3976 | if (IS_ERR_OR_NULL(di)) { | 3976 | if (IS_ERR_OR_NULL(di)) { |
| 3977 | if (!di) | 3977 | if (!di) |
| 3978 | ret = -ENOENT; | 3978 | ret = -ENOENT; |
| 3979 | else | 3979 | else |
| 3980 | ret = PTR_ERR(di); | 3980 | ret = PTR_ERR(di); |
| 3981 | goto out; | 3981 | goto out; |
| 3982 | } | 3982 | } |
| 3983 | 3983 | ||
| 3984 | leaf = path->nodes[0]; | 3984 | leaf = path->nodes[0]; |
| 3985 | btrfs_dir_item_key_to_cpu(leaf, di, &key); | 3985 | btrfs_dir_item_key_to_cpu(leaf, di, &key); |
| 3986 | WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); | 3986 | WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); |
| 3987 | ret = btrfs_delete_one_dir_name(trans, root, path, di); | 3987 | ret = btrfs_delete_one_dir_name(trans, root, path, di); |
| 3988 | if (ret) { | 3988 | if (ret) { |
| 3989 | btrfs_abort_transaction(trans, root, ret); | 3989 | btrfs_abort_transaction(trans, root, ret); |
| 3990 | goto out; | 3990 | goto out; |
| 3991 | } | 3991 | } |
| 3992 | btrfs_release_path(path); | 3992 | btrfs_release_path(path); |
| 3993 | 3993 | ||
| 3994 | ret = btrfs_del_root_ref(trans, root->fs_info->tree_root, | 3994 | ret = btrfs_del_root_ref(trans, root->fs_info->tree_root, |
| 3995 | objectid, root->root_key.objectid, | 3995 | objectid, root->root_key.objectid, |
| 3996 | dir_ino, &index, name, name_len); | 3996 | dir_ino, &index, name, name_len); |
| 3997 | if (ret < 0) { | 3997 | if (ret < 0) { |
| 3998 | if (ret != -ENOENT) { | 3998 | if (ret != -ENOENT) { |
| 3999 | btrfs_abort_transaction(trans, root, ret); | 3999 | btrfs_abort_transaction(trans, root, ret); |
| 4000 | goto out; | 4000 | goto out; |
| 4001 | } | 4001 | } |
| 4002 | di = btrfs_search_dir_index_item(root, path, dir_ino, | 4002 | di = btrfs_search_dir_index_item(root, path, dir_ino, |
| 4003 | name, name_len); | 4003 | name, name_len); |
| 4004 | if (IS_ERR_OR_NULL(di)) { | 4004 | if (IS_ERR_OR_NULL(di)) { |
| 4005 | if (!di) | 4005 | if (!di) |
| 4006 | ret = -ENOENT; | 4006 | ret = -ENOENT; |
| 4007 | else | 4007 | else |
| 4008 | ret = PTR_ERR(di); | 4008 | ret = PTR_ERR(di); |
| 4009 | btrfs_abort_transaction(trans, root, ret); | 4009 | btrfs_abort_transaction(trans, root, ret); |
| 4010 | goto out; | 4010 | goto out; |
| 4011 | } | 4011 | } |
| 4012 | 4012 | ||
| 4013 | leaf = path->nodes[0]; | 4013 | leaf = path->nodes[0]; |
| 4014 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); | 4014 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
| 4015 | btrfs_release_path(path); | 4015 | btrfs_release_path(path); |
| 4016 | index = key.offset; | 4016 | index = key.offset; |
| 4017 | } | 4017 | } |
| 4018 | btrfs_release_path(path); | 4018 | btrfs_release_path(path); |
| 4019 | 4019 | ||
| 4020 | ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); | 4020 | ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); |
| 4021 | if (ret) { | 4021 | if (ret) { |
| 4022 | btrfs_abort_transaction(trans, root, ret); | 4022 | btrfs_abort_transaction(trans, root, ret); |
| 4023 | goto out; | 4023 | goto out; |
| 4024 | } | 4024 | } |
| 4025 | 4025 | ||
| 4026 | btrfs_i_size_write(dir, dir->i_size - name_len * 2); | 4026 | btrfs_i_size_write(dir, dir->i_size - name_len * 2); |
| 4027 | inode_inc_iversion(dir); | 4027 | inode_inc_iversion(dir); |
| 4028 | dir->i_mtime = dir->i_ctime = CURRENT_TIME; | 4028 | dir->i_mtime = dir->i_ctime = CURRENT_TIME; |
| 4029 | ret = btrfs_update_inode_fallback(trans, root, dir); | 4029 | ret = btrfs_update_inode_fallback(trans, root, dir); |
| 4030 | if (ret) | 4030 | if (ret) |
| 4031 | btrfs_abort_transaction(trans, root, ret); | 4031 | btrfs_abort_transaction(trans, root, ret); |
| 4032 | out: | 4032 | out: |
| 4033 | btrfs_free_path(path); | 4033 | btrfs_free_path(path); |
| 4034 | return ret; | 4034 | return ret; |
| 4035 | } | 4035 | } |
| 4036 | 4036 | ||
| 4037 | static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) | 4037 | static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) |
| 4038 | { | 4038 | { |
| 4039 | struct inode *inode = dentry->d_inode; | 4039 | struct inode *inode = dentry->d_inode; |
| 4040 | int err = 0; | 4040 | int err = 0; |
| 4041 | struct btrfs_root *root = BTRFS_I(dir)->root; | 4041 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 4042 | struct btrfs_trans_handle *trans; | 4042 | struct btrfs_trans_handle *trans; |
| 4043 | 4043 | ||
| 4044 | if (inode->i_size > BTRFS_EMPTY_DIR_SIZE) | 4044 | if (inode->i_size > BTRFS_EMPTY_DIR_SIZE) |
| 4045 | return -ENOTEMPTY; | 4045 | return -ENOTEMPTY; |
| 4046 | if (btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID) | 4046 | if (btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID) |
| 4047 | return -EPERM; | 4047 | return -EPERM; |
| 4048 | 4048 | ||
| 4049 | trans = __unlink_start_trans(dir); | 4049 | trans = __unlink_start_trans(dir); |
| 4050 | if (IS_ERR(trans)) | 4050 | if (IS_ERR(trans)) |
| 4051 | return PTR_ERR(trans); | 4051 | return PTR_ERR(trans); |
| 4052 | 4052 | ||
| 4053 | if (unlikely(btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { | 4053 | if (unlikely(btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { |
| 4054 | err = btrfs_unlink_subvol(trans, root, dir, | 4054 | err = btrfs_unlink_subvol(trans, root, dir, |
| 4055 | BTRFS_I(inode)->location.objectid, | 4055 | BTRFS_I(inode)->location.objectid, |
| 4056 | dentry->d_name.name, | 4056 | dentry->d_name.name, |
| 4057 | dentry->d_name.len); | 4057 | dentry->d_name.len); |
| 4058 | goto out; | 4058 | goto out; |
| 4059 | } | 4059 | } |
| 4060 | 4060 | ||
| 4061 | err = btrfs_orphan_add(trans, inode); | 4061 | err = btrfs_orphan_add(trans, inode); |
| 4062 | if (err) | 4062 | if (err) |
| 4063 | goto out; | 4063 | goto out; |
| 4064 | 4064 | ||
| 4065 | /* now the directory is empty */ | 4065 | /* now the directory is empty */ |
| 4066 | err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode, | 4066 | err = btrfs_unlink_inode(trans, root, dir, dentry->d_inode, |
| 4067 | dentry->d_name.name, dentry->d_name.len); | 4067 | dentry->d_name.name, dentry->d_name.len); |
| 4068 | if (!err) | 4068 | if (!err) |
| 4069 | btrfs_i_size_write(inode, 0); | 4069 | btrfs_i_size_write(inode, 0); |
| 4070 | out: | 4070 | out: |
| 4071 | btrfs_end_transaction(trans, root); | 4071 | btrfs_end_transaction(trans, root); |
| 4072 | btrfs_btree_balance_dirty(root); | 4072 | btrfs_btree_balance_dirty(root); |
| 4073 | 4073 | ||
| 4074 | return err; | 4074 | return err; |
| 4075 | } | 4075 | } |
| 4076 | 4076 | ||
| 4077 | /* | 4077 | /* |
| 4078 | * this can truncate away extent items, csum items and directory items. | 4078 | * this can truncate away extent items, csum items and directory items. |
| 4079 | * It starts at a high offset and removes keys until it can't find | 4079 | * It starts at a high offset and removes keys until it can't find |
| 4080 | * any higher than new_size | 4080 | * any higher than new_size |
| 4081 | * | 4081 | * |
| 4082 | * csum items that cross the new i_size are truncated to the new size | 4082 | * csum items that cross the new i_size are truncated to the new size |
| 4083 | * as well. | 4083 | * as well. |
| 4084 | * | 4084 | * |
| 4085 | * min_type is the minimum key type to truncate down to. If set to 0, this | 4085 | * min_type is the minimum key type to truncate down to. If set to 0, this |
| 4086 | * will kill all the items on this inode, including the INODE_ITEM_KEY. | 4086 | * will kill all the items on this inode, including the INODE_ITEM_KEY. |
| 4087 | */ | 4087 | */ |
| 4088 | int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, | 4088 | int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, |
| 4089 | struct btrfs_root *root, | 4089 | struct btrfs_root *root, |
| 4090 | struct inode *inode, | 4090 | struct inode *inode, |
| 4091 | u64 new_size, u32 min_type) | 4091 | u64 new_size, u32 min_type) |
| 4092 | { | 4092 | { |
| 4093 | struct btrfs_path *path; | 4093 | struct btrfs_path *path; |
| 4094 | struct extent_buffer *leaf; | 4094 | struct extent_buffer *leaf; |
| 4095 | struct btrfs_file_extent_item *fi; | 4095 | struct btrfs_file_extent_item *fi; |
| 4096 | struct btrfs_key key; | 4096 | struct btrfs_key key; |
| 4097 | struct btrfs_key found_key; | 4097 | struct btrfs_key found_key; |
| 4098 | u64 extent_start = 0; | 4098 | u64 extent_start = 0; |
| 4099 | u64 extent_num_bytes = 0; | 4099 | u64 extent_num_bytes = 0; |
| 4100 | u64 extent_offset = 0; | 4100 | u64 extent_offset = 0; |
| 4101 | u64 item_end = 0; | 4101 | u64 item_end = 0; |
| 4102 | u64 last_size = (u64)-1; | 4102 | u64 last_size = (u64)-1; |
| 4103 | u32 found_type = (u8)-1; | 4103 | u32 found_type = (u8)-1; |
| 4104 | int found_extent; | 4104 | int found_extent; |
| 4105 | int del_item; | 4105 | int del_item; |
| 4106 | int pending_del_nr = 0; | 4106 | int pending_del_nr = 0; |
| 4107 | int pending_del_slot = 0; | 4107 | int pending_del_slot = 0; |
| 4108 | int extent_type = -1; | 4108 | int extent_type = -1; |
| 4109 | int ret; | 4109 | int ret; |
| 4110 | int err = 0; | 4110 | int err = 0; |
| 4111 | u64 ino = btrfs_ino(inode); | 4111 | u64 ino = btrfs_ino(inode); |
| 4112 | 4112 | ||
| 4113 | BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY); | 4113 | BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY); |
| 4114 | 4114 | ||
| 4115 | path = btrfs_alloc_path(); | 4115 | path = btrfs_alloc_path(); |
| 4116 | if (!path) | 4116 | if (!path) |
| 4117 | return -ENOMEM; | 4117 | return -ENOMEM; |
| 4118 | path->reada = -1; | 4118 | path->reada = -1; |
| 4119 | 4119 | ||
| 4120 | /* | 4120 | /* |
| 4121 | * We want to drop from the next block forward in case this new size is | 4121 | * We want to drop from the next block forward in case this new size is |
| 4122 | * not block aligned since we will be keeping the last block of the | 4122 | * not block aligned since we will be keeping the last block of the |
| 4123 | * extent just the way it is. | 4123 | * extent just the way it is. |
| 4124 | */ | 4124 | */ |
| 4125 | if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) || | 4125 | if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) || |
| 4126 | root == root->fs_info->tree_root) | 4126 | root == root->fs_info->tree_root) |
| 4127 | btrfs_drop_extent_cache(inode, ALIGN(new_size, | 4127 | btrfs_drop_extent_cache(inode, ALIGN(new_size, |
| 4128 | root->sectorsize), (u64)-1, 0); | 4128 | root->sectorsize), (u64)-1, 0); |
| 4129 | 4129 | ||
| 4130 | /* | 4130 | /* |
| 4131 | * This function is also used to drop the items in the log tree before | 4131 | * This function is also used to drop the items in the log tree before |
| 4132 | * we relog the inode, so if root != BTRFS_I(inode)->root, it means | 4132 | * we relog the inode, so if root != BTRFS_I(inode)->root, it means |
| 4133 | * it is used to drop the loged items. So we shouldn't kill the delayed | 4133 | * it is used to drop the loged items. So we shouldn't kill the delayed |
| 4134 | * items. | 4134 | * items. |
| 4135 | */ | 4135 | */ |
| 4136 | if (min_type == 0 && root == BTRFS_I(inode)->root) | 4136 | if (min_type == 0 && root == BTRFS_I(inode)->root) |
| 4137 | btrfs_kill_delayed_inode_items(inode); | 4137 | btrfs_kill_delayed_inode_items(inode); |
| 4138 | 4138 | ||
| 4139 | key.objectid = ino; | 4139 | key.objectid = ino; |
| 4140 | key.offset = (u64)-1; | 4140 | key.offset = (u64)-1; |
| 4141 | key.type = (u8)-1; | 4141 | key.type = (u8)-1; |
| 4142 | 4142 | ||
| 4143 | search_again: | 4143 | search_again: |
| 4144 | path->leave_spinning = 1; | 4144 | path->leave_spinning = 1; |
| 4145 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); | 4145 | ret = btrfs_search_slot(trans, root, &key, path, -1, 1); |
| 4146 | if (ret < 0) { | 4146 | if (ret < 0) { |
| 4147 | err = ret; | 4147 | err = ret; |
| 4148 | goto out; | 4148 | goto out; |
| 4149 | } | 4149 | } |
| 4150 | 4150 | ||
| 4151 | if (ret > 0) { | 4151 | if (ret > 0) { |
| 4152 | /* there are no items in the tree for us to truncate, we're | 4152 | /* there are no items in the tree for us to truncate, we're |
| 4153 | * done | 4153 | * done |
| 4154 | */ | 4154 | */ |
| 4155 | if (path->slots[0] == 0) | 4155 | if (path->slots[0] == 0) |
| 4156 | goto out; | 4156 | goto out; |
| 4157 | path->slots[0]--; | 4157 | path->slots[0]--; |
| 4158 | } | 4158 | } |
| 4159 | 4159 | ||
| 4160 | while (1) { | 4160 | while (1) { |
| 4161 | fi = NULL; | 4161 | fi = NULL; |
| 4162 | leaf = path->nodes[0]; | 4162 | leaf = path->nodes[0]; |
| 4163 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 4163 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 4164 | found_type = found_key.type; | 4164 | found_type = found_key.type; |
| 4165 | 4165 | ||
| 4166 | if (found_key.objectid != ino) | 4166 | if (found_key.objectid != ino) |
| 4167 | break; | 4167 | break; |
| 4168 | 4168 | ||
| 4169 | if (found_type < min_type) | 4169 | if (found_type < min_type) |
| 4170 | break; | 4170 | break; |
| 4171 | 4171 | ||
| 4172 | item_end = found_key.offset; | 4172 | item_end = found_key.offset; |
| 4173 | if (found_type == BTRFS_EXTENT_DATA_KEY) { | 4173 | if (found_type == BTRFS_EXTENT_DATA_KEY) { |
| 4174 | fi = btrfs_item_ptr(leaf, path->slots[0], | 4174 | fi = btrfs_item_ptr(leaf, path->slots[0], |
| 4175 | struct btrfs_file_extent_item); | 4175 | struct btrfs_file_extent_item); |
| 4176 | extent_type = btrfs_file_extent_type(leaf, fi); | 4176 | extent_type = btrfs_file_extent_type(leaf, fi); |
| 4177 | if (extent_type != BTRFS_FILE_EXTENT_INLINE) { | 4177 | if (extent_type != BTRFS_FILE_EXTENT_INLINE) { |
| 4178 | item_end += | 4178 | item_end += |
| 4179 | btrfs_file_extent_num_bytes(leaf, fi); | 4179 | btrfs_file_extent_num_bytes(leaf, fi); |
| 4180 | } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { | 4180 | } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { |
| 4181 | item_end += btrfs_file_extent_inline_len(leaf, | 4181 | item_end += btrfs_file_extent_inline_len(leaf, |
| 4182 | path->slots[0], fi); | 4182 | path->slots[0], fi); |
| 4183 | } | 4183 | } |
| 4184 | item_end--; | 4184 | item_end--; |
| 4185 | } | 4185 | } |
| 4186 | if (found_type > min_type) { | 4186 | if (found_type > min_type) { |
| 4187 | del_item = 1; | 4187 | del_item = 1; |
| 4188 | } else { | 4188 | } else { |
| 4189 | if (item_end < new_size) | 4189 | if (item_end < new_size) |
| 4190 | break; | 4190 | break; |
| 4191 | if (found_key.offset >= new_size) | 4191 | if (found_key.offset >= new_size) |
| 4192 | del_item = 1; | 4192 | del_item = 1; |
| 4193 | else | 4193 | else |
| 4194 | del_item = 0; | 4194 | del_item = 0; |
| 4195 | } | 4195 | } |
| 4196 | found_extent = 0; | 4196 | found_extent = 0; |
| 4197 | /* FIXME, shrink the extent if the ref count is only 1 */ | 4197 | /* FIXME, shrink the extent if the ref count is only 1 */ |
| 4198 | if (found_type != BTRFS_EXTENT_DATA_KEY) | 4198 | if (found_type != BTRFS_EXTENT_DATA_KEY) |
| 4199 | goto delete; | 4199 | goto delete; |
| 4200 | 4200 | ||
| 4201 | if (del_item) | 4201 | if (del_item) |
| 4202 | last_size = found_key.offset; | 4202 | last_size = found_key.offset; |
| 4203 | else | 4203 | else |
| 4204 | last_size = new_size; | 4204 | last_size = new_size; |
| 4205 | 4205 | ||
| 4206 | if (extent_type != BTRFS_FILE_EXTENT_INLINE) { | 4206 | if (extent_type != BTRFS_FILE_EXTENT_INLINE) { |
| 4207 | u64 num_dec; | 4207 | u64 num_dec; |
| 4208 | extent_start = btrfs_file_extent_disk_bytenr(leaf, fi); | 4208 | extent_start = btrfs_file_extent_disk_bytenr(leaf, fi); |
| 4209 | if (!del_item) { | 4209 | if (!del_item) { |
| 4210 | u64 orig_num_bytes = | 4210 | u64 orig_num_bytes = |
| 4211 | btrfs_file_extent_num_bytes(leaf, fi); | 4211 | btrfs_file_extent_num_bytes(leaf, fi); |
| 4212 | extent_num_bytes = ALIGN(new_size - | 4212 | extent_num_bytes = ALIGN(new_size - |
| 4213 | found_key.offset, | 4213 | found_key.offset, |
| 4214 | root->sectorsize); | 4214 | root->sectorsize); |
| 4215 | btrfs_set_file_extent_num_bytes(leaf, fi, | 4215 | btrfs_set_file_extent_num_bytes(leaf, fi, |
| 4216 | extent_num_bytes); | 4216 | extent_num_bytes); |
| 4217 | num_dec = (orig_num_bytes - | 4217 | num_dec = (orig_num_bytes - |
| 4218 | extent_num_bytes); | 4218 | extent_num_bytes); |
| 4219 | if (test_bit(BTRFS_ROOT_REF_COWS, | 4219 | if (test_bit(BTRFS_ROOT_REF_COWS, |
| 4220 | &root->state) && | 4220 | &root->state) && |
| 4221 | extent_start != 0) | 4221 | extent_start != 0) |
| 4222 | inode_sub_bytes(inode, num_dec); | 4222 | inode_sub_bytes(inode, num_dec); |
| 4223 | btrfs_mark_buffer_dirty(leaf); | 4223 | btrfs_mark_buffer_dirty(leaf); |
| 4224 | } else { | 4224 | } else { |
| 4225 | extent_num_bytes = | 4225 | extent_num_bytes = |
| 4226 | btrfs_file_extent_disk_num_bytes(leaf, | 4226 | btrfs_file_extent_disk_num_bytes(leaf, |
| 4227 | fi); | 4227 | fi); |
| 4228 | extent_offset = found_key.offset - | 4228 | extent_offset = found_key.offset - |
| 4229 | btrfs_file_extent_offset(leaf, fi); | 4229 | btrfs_file_extent_offset(leaf, fi); |
| 4230 | 4230 | ||
| 4231 | /* FIXME blocksize != 4096 */ | 4231 | /* FIXME blocksize != 4096 */ |
| 4232 | num_dec = btrfs_file_extent_num_bytes(leaf, fi); | 4232 | num_dec = btrfs_file_extent_num_bytes(leaf, fi); |
| 4233 | if (extent_start != 0) { | 4233 | if (extent_start != 0) { |
| 4234 | found_extent = 1; | 4234 | found_extent = 1; |
| 4235 | if (test_bit(BTRFS_ROOT_REF_COWS, | 4235 | if (test_bit(BTRFS_ROOT_REF_COWS, |
| 4236 | &root->state)) | 4236 | &root->state)) |
| 4237 | inode_sub_bytes(inode, num_dec); | 4237 | inode_sub_bytes(inode, num_dec); |
| 4238 | } | 4238 | } |
| 4239 | } | 4239 | } |
| 4240 | } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { | 4240 | } else if (extent_type == BTRFS_FILE_EXTENT_INLINE) { |
| 4241 | /* | 4241 | /* |
| 4242 | * we can't truncate inline items that have had | 4242 | * we can't truncate inline items that have had |
| 4243 | * special encodings | 4243 | * special encodings |
| 4244 | */ | 4244 | */ |
| 4245 | if (!del_item && | 4245 | if (!del_item && |
| 4246 | btrfs_file_extent_compression(leaf, fi) == 0 && | 4246 | btrfs_file_extent_compression(leaf, fi) == 0 && |
| 4247 | btrfs_file_extent_encryption(leaf, fi) == 0 && | 4247 | btrfs_file_extent_encryption(leaf, fi) == 0 && |
| 4248 | btrfs_file_extent_other_encoding(leaf, fi) == 0) { | 4248 | btrfs_file_extent_other_encoding(leaf, fi) == 0) { |
| 4249 | u32 size = new_size - found_key.offset; | 4249 | u32 size = new_size - found_key.offset; |
| 4250 | 4250 | ||
| 4251 | if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) | 4251 | if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) |
| 4252 | inode_sub_bytes(inode, item_end + 1 - | 4252 | inode_sub_bytes(inode, item_end + 1 - |
| 4253 | new_size); | 4253 | new_size); |
| 4254 | 4254 | ||
| 4255 | /* | 4255 | /* |
| 4256 | * update the ram bytes to properly reflect | 4256 | * update the ram bytes to properly reflect |
| 4257 | * the new size of our item | 4257 | * the new size of our item |
| 4258 | */ | 4258 | */ |
| 4259 | btrfs_set_file_extent_ram_bytes(leaf, fi, size); | 4259 | btrfs_set_file_extent_ram_bytes(leaf, fi, size); |
| 4260 | size = | 4260 | size = |
| 4261 | btrfs_file_extent_calc_inline_size(size); | 4261 | btrfs_file_extent_calc_inline_size(size); |
| 4262 | btrfs_truncate_item(root, path, size, 1); | 4262 | btrfs_truncate_item(root, path, size, 1); |
| 4263 | } else if (test_bit(BTRFS_ROOT_REF_COWS, | 4263 | } else if (test_bit(BTRFS_ROOT_REF_COWS, |
| 4264 | &root->state)) { | 4264 | &root->state)) { |
| 4265 | inode_sub_bytes(inode, item_end + 1 - | 4265 | inode_sub_bytes(inode, item_end + 1 - |
| 4266 | found_key.offset); | 4266 | found_key.offset); |
| 4267 | } | 4267 | } |
| 4268 | } | 4268 | } |
| 4269 | delete: | 4269 | delete: |
| 4270 | if (del_item) { | 4270 | if (del_item) { |
| 4271 | if (!pending_del_nr) { | 4271 | if (!pending_del_nr) { |
| 4272 | /* no pending yet, add ourselves */ | 4272 | /* no pending yet, add ourselves */ |
| 4273 | pending_del_slot = path->slots[0]; | 4273 | pending_del_slot = path->slots[0]; |
| 4274 | pending_del_nr = 1; | 4274 | pending_del_nr = 1; |
| 4275 | } else if (pending_del_nr && | 4275 | } else if (pending_del_nr && |
| 4276 | path->slots[0] + 1 == pending_del_slot) { | 4276 | path->slots[0] + 1 == pending_del_slot) { |
| 4277 | /* hop on the pending chunk */ | 4277 | /* hop on the pending chunk */ |
| 4278 | pending_del_nr++; | 4278 | pending_del_nr++; |
| 4279 | pending_del_slot = path->slots[0]; | 4279 | pending_del_slot = path->slots[0]; |
| 4280 | } else { | 4280 | } else { |
| 4281 | BUG(); | 4281 | BUG(); |
| 4282 | } | 4282 | } |
| 4283 | } else { | 4283 | } else { |
| 4284 | break; | 4284 | break; |
| 4285 | } | 4285 | } |
| 4286 | if (found_extent && | 4286 | if (found_extent && |
| 4287 | (test_bit(BTRFS_ROOT_REF_COWS, &root->state) || | 4287 | (test_bit(BTRFS_ROOT_REF_COWS, &root->state) || |
| 4288 | root == root->fs_info->tree_root)) { | 4288 | root == root->fs_info->tree_root)) { |
| 4289 | btrfs_set_path_blocking(path); | 4289 | btrfs_set_path_blocking(path); |
| 4290 | ret = btrfs_free_extent(trans, root, extent_start, | 4290 | ret = btrfs_free_extent(trans, root, extent_start, |
| 4291 | extent_num_bytes, 0, | 4291 | extent_num_bytes, 0, |
| 4292 | btrfs_header_owner(leaf), | 4292 | btrfs_header_owner(leaf), |
| 4293 | ino, extent_offset, 0); | 4293 | ino, extent_offset, 0); |
| 4294 | BUG_ON(ret); | 4294 | BUG_ON(ret); |
| 4295 | } | 4295 | } |
| 4296 | 4296 | ||
| 4297 | if (found_type == BTRFS_INODE_ITEM_KEY) | 4297 | if (found_type == BTRFS_INODE_ITEM_KEY) |
| 4298 | break; | 4298 | break; |
| 4299 | 4299 | ||
| 4300 | if (path->slots[0] == 0 || | 4300 | if (path->slots[0] == 0 || |
| 4301 | path->slots[0] != pending_del_slot) { | 4301 | path->slots[0] != pending_del_slot) { |
| 4302 | if (pending_del_nr) { | 4302 | if (pending_del_nr) { |
| 4303 | ret = btrfs_del_items(trans, root, path, | 4303 | ret = btrfs_del_items(trans, root, path, |
| 4304 | pending_del_slot, | 4304 | pending_del_slot, |
| 4305 | pending_del_nr); | 4305 | pending_del_nr); |
| 4306 | if (ret) { | 4306 | if (ret) { |
| 4307 | btrfs_abort_transaction(trans, | 4307 | btrfs_abort_transaction(trans, |
| 4308 | root, ret); | 4308 | root, ret); |
| 4309 | goto error; | 4309 | goto error; |
| 4310 | } | 4310 | } |
| 4311 | pending_del_nr = 0; | 4311 | pending_del_nr = 0; |
| 4312 | } | 4312 | } |
| 4313 | btrfs_release_path(path); | 4313 | btrfs_release_path(path); |
| 4314 | goto search_again; | 4314 | goto search_again; |
| 4315 | } else { | 4315 | } else { |
| 4316 | path->slots[0]--; | 4316 | path->slots[0]--; |
| 4317 | } | 4317 | } |
| 4318 | } | 4318 | } |
| 4319 | out: | 4319 | out: |
| 4320 | if (pending_del_nr) { | 4320 | if (pending_del_nr) { |
| 4321 | ret = btrfs_del_items(trans, root, path, pending_del_slot, | 4321 | ret = btrfs_del_items(trans, root, path, pending_del_slot, |
| 4322 | pending_del_nr); | 4322 | pending_del_nr); |
| 4323 | if (ret) | 4323 | if (ret) |
| 4324 | btrfs_abort_transaction(trans, root, ret); | 4324 | btrfs_abort_transaction(trans, root, ret); |
| 4325 | } | 4325 | } |
| 4326 | error: | 4326 | error: |
| 4327 | if (last_size != (u64)-1 && | 4327 | if (last_size != (u64)-1 && |
| 4328 | root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) | 4328 | root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) |
| 4329 | btrfs_ordered_update_i_size(inode, last_size, NULL); | 4329 | btrfs_ordered_update_i_size(inode, last_size, NULL); |
| 4330 | btrfs_free_path(path); | 4330 | btrfs_free_path(path); |
| 4331 | return err; | 4331 | return err; |
| 4332 | } | 4332 | } |
| 4333 | 4333 | ||
| 4334 | /* | 4334 | /* |
| 4335 | * btrfs_truncate_page - read, zero a chunk and write a page | 4335 | * btrfs_truncate_page - read, zero a chunk and write a page |
| 4336 | * @inode - inode that we're zeroing | 4336 | * @inode - inode that we're zeroing |
| 4337 | * @from - the offset to start zeroing | 4337 | * @from - the offset to start zeroing |
| 4338 | * @len - the length to zero, 0 to zero the entire range respective to the | 4338 | * @len - the length to zero, 0 to zero the entire range respective to the |
| 4339 | * offset | 4339 | * offset |
| 4340 | * @front - zero up to the offset instead of from the offset on | 4340 | * @front - zero up to the offset instead of from the offset on |
| 4341 | * | 4341 | * |
| 4342 | * This will find the page for the "from" offset and cow the page and zero the | 4342 | * This will find the page for the "from" offset and cow the page and zero the |
| 4343 | * part we want to zero. This is used with truncate and hole punching. | 4343 | * part we want to zero. This is used with truncate and hole punching. |
| 4344 | */ | 4344 | */ |
| 4345 | int btrfs_truncate_page(struct inode *inode, loff_t from, loff_t len, | 4345 | int btrfs_truncate_page(struct inode *inode, loff_t from, loff_t len, |
| 4346 | int front) | 4346 | int front) |
| 4347 | { | 4347 | { |
| 4348 | struct address_space *mapping = inode->i_mapping; | 4348 | struct address_space *mapping = inode->i_mapping; |
| 4349 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4349 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4350 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 4350 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 4351 | struct btrfs_ordered_extent *ordered; | 4351 | struct btrfs_ordered_extent *ordered; |
| 4352 | struct extent_state *cached_state = NULL; | 4352 | struct extent_state *cached_state = NULL; |
| 4353 | char *kaddr; | 4353 | char *kaddr; |
| 4354 | u32 blocksize = root->sectorsize; | 4354 | u32 blocksize = root->sectorsize; |
| 4355 | pgoff_t index = from >> PAGE_CACHE_SHIFT; | 4355 | pgoff_t index = from >> PAGE_CACHE_SHIFT; |
| 4356 | unsigned offset = from & (PAGE_CACHE_SIZE-1); | 4356 | unsigned offset = from & (PAGE_CACHE_SIZE-1); |
| 4357 | struct page *page; | 4357 | struct page *page; |
| 4358 | gfp_t mask = btrfs_alloc_write_mask(mapping); | 4358 | gfp_t mask = btrfs_alloc_write_mask(mapping); |
| 4359 | int ret = 0; | 4359 | int ret = 0; |
| 4360 | u64 page_start; | 4360 | u64 page_start; |
| 4361 | u64 page_end; | 4361 | u64 page_end; |
| 4362 | 4362 | ||
| 4363 | if ((offset & (blocksize - 1)) == 0 && | 4363 | if ((offset & (blocksize - 1)) == 0 && |
| 4364 | (!len || ((len & (blocksize - 1)) == 0))) | 4364 | (!len || ((len & (blocksize - 1)) == 0))) |
| 4365 | goto out; | 4365 | goto out; |
| 4366 | ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); | 4366 | ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); |
| 4367 | if (ret) | 4367 | if (ret) |
| 4368 | goto out; | 4368 | goto out; |
| 4369 | 4369 | ||
| 4370 | again: | 4370 | again: |
| 4371 | page = find_or_create_page(mapping, index, mask); | 4371 | page = find_or_create_page(mapping, index, mask); |
| 4372 | if (!page) { | 4372 | if (!page) { |
| 4373 | btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); | 4373 | btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); |
| 4374 | ret = -ENOMEM; | 4374 | ret = -ENOMEM; |
| 4375 | goto out; | 4375 | goto out; |
| 4376 | } | 4376 | } |
| 4377 | 4377 | ||
| 4378 | page_start = page_offset(page); | 4378 | page_start = page_offset(page); |
| 4379 | page_end = page_start + PAGE_CACHE_SIZE - 1; | 4379 | page_end = page_start + PAGE_CACHE_SIZE - 1; |
| 4380 | 4380 | ||
| 4381 | if (!PageUptodate(page)) { | 4381 | if (!PageUptodate(page)) { |
| 4382 | ret = btrfs_readpage(NULL, page); | 4382 | ret = btrfs_readpage(NULL, page); |
| 4383 | lock_page(page); | 4383 | lock_page(page); |
| 4384 | if (page->mapping != mapping) { | 4384 | if (page->mapping != mapping) { |
| 4385 | unlock_page(page); | 4385 | unlock_page(page); |
| 4386 | page_cache_release(page); | 4386 | page_cache_release(page); |
| 4387 | goto again; | 4387 | goto again; |
| 4388 | } | 4388 | } |
| 4389 | if (!PageUptodate(page)) { | 4389 | if (!PageUptodate(page)) { |
| 4390 | ret = -EIO; | 4390 | ret = -EIO; |
| 4391 | goto out_unlock; | 4391 | goto out_unlock; |
| 4392 | } | 4392 | } |
| 4393 | } | 4393 | } |
| 4394 | wait_on_page_writeback(page); | 4394 | wait_on_page_writeback(page); |
| 4395 | 4395 | ||
| 4396 | lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state); | 4396 | lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state); |
| 4397 | set_page_extent_mapped(page); | 4397 | set_page_extent_mapped(page); |
| 4398 | 4398 | ||
| 4399 | ordered = btrfs_lookup_ordered_extent(inode, page_start); | 4399 | ordered = btrfs_lookup_ordered_extent(inode, page_start); |
| 4400 | if (ordered) { | 4400 | if (ordered) { |
| 4401 | unlock_extent_cached(io_tree, page_start, page_end, | 4401 | unlock_extent_cached(io_tree, page_start, page_end, |
| 4402 | &cached_state, GFP_NOFS); | 4402 | &cached_state, GFP_NOFS); |
| 4403 | unlock_page(page); | 4403 | unlock_page(page); |
| 4404 | page_cache_release(page); | 4404 | page_cache_release(page); |
| 4405 | btrfs_start_ordered_extent(inode, ordered, 1); | 4405 | btrfs_start_ordered_extent(inode, ordered, 1); |
| 4406 | btrfs_put_ordered_extent(ordered); | 4406 | btrfs_put_ordered_extent(ordered); |
| 4407 | goto again; | 4407 | goto again; |
| 4408 | } | 4408 | } |
| 4409 | 4409 | ||
| 4410 | clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end, | 4410 | clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end, |
| 4411 | EXTENT_DIRTY | EXTENT_DELALLOC | | 4411 | EXTENT_DIRTY | EXTENT_DELALLOC | |
| 4412 | EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, | 4412 | EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, |
| 4413 | 0, 0, &cached_state, GFP_NOFS); | 4413 | 0, 0, &cached_state, GFP_NOFS); |
| 4414 | 4414 | ||
| 4415 | ret = btrfs_set_extent_delalloc(inode, page_start, page_end, | 4415 | ret = btrfs_set_extent_delalloc(inode, page_start, page_end, |
| 4416 | &cached_state); | 4416 | &cached_state); |
| 4417 | if (ret) { | 4417 | if (ret) { |
| 4418 | unlock_extent_cached(io_tree, page_start, page_end, | 4418 | unlock_extent_cached(io_tree, page_start, page_end, |
| 4419 | &cached_state, GFP_NOFS); | 4419 | &cached_state, GFP_NOFS); |
| 4420 | goto out_unlock; | 4420 | goto out_unlock; |
| 4421 | } | 4421 | } |
| 4422 | 4422 | ||
| 4423 | if (offset != PAGE_CACHE_SIZE) { | 4423 | if (offset != PAGE_CACHE_SIZE) { |
| 4424 | if (!len) | 4424 | if (!len) |
| 4425 | len = PAGE_CACHE_SIZE - offset; | 4425 | len = PAGE_CACHE_SIZE - offset; |
| 4426 | kaddr = kmap(page); | 4426 | kaddr = kmap(page); |
| 4427 | if (front) | 4427 | if (front) |
| 4428 | memset(kaddr, 0, offset); | 4428 | memset(kaddr, 0, offset); |
| 4429 | else | 4429 | else |
| 4430 | memset(kaddr + offset, 0, len); | 4430 | memset(kaddr + offset, 0, len); |
| 4431 | flush_dcache_page(page); | 4431 | flush_dcache_page(page); |
| 4432 | kunmap(page); | 4432 | kunmap(page); |
| 4433 | } | 4433 | } |
| 4434 | ClearPageChecked(page); | 4434 | ClearPageChecked(page); |
| 4435 | set_page_dirty(page); | 4435 | set_page_dirty(page); |
| 4436 | unlock_extent_cached(io_tree, page_start, page_end, &cached_state, | 4436 | unlock_extent_cached(io_tree, page_start, page_end, &cached_state, |
| 4437 | GFP_NOFS); | 4437 | GFP_NOFS); |
| 4438 | 4438 | ||
| 4439 | out_unlock: | 4439 | out_unlock: |
| 4440 | if (ret) | 4440 | if (ret) |
| 4441 | btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); | 4441 | btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); |
| 4442 | unlock_page(page); | 4442 | unlock_page(page); |
| 4443 | page_cache_release(page); | 4443 | page_cache_release(page); |
| 4444 | out: | 4444 | out: |
| 4445 | return ret; | 4445 | return ret; |
| 4446 | } | 4446 | } |
| 4447 | 4447 | ||
| 4448 | static int maybe_insert_hole(struct btrfs_root *root, struct inode *inode, | 4448 | static int maybe_insert_hole(struct btrfs_root *root, struct inode *inode, |
| 4449 | u64 offset, u64 len) | 4449 | u64 offset, u64 len) |
| 4450 | { | 4450 | { |
| 4451 | struct btrfs_trans_handle *trans; | 4451 | struct btrfs_trans_handle *trans; |
| 4452 | int ret; | 4452 | int ret; |
| 4453 | 4453 | ||
| 4454 | /* | 4454 | /* |
| 4455 | * Still need to make sure the inode looks like it's been updated so | 4455 | * Still need to make sure the inode looks like it's been updated so |
| 4456 | * that any holes get logged if we fsync. | 4456 | * that any holes get logged if we fsync. |
| 4457 | */ | 4457 | */ |
| 4458 | if (btrfs_fs_incompat(root->fs_info, NO_HOLES)) { | 4458 | if (btrfs_fs_incompat(root->fs_info, NO_HOLES)) { |
| 4459 | BTRFS_I(inode)->last_trans = root->fs_info->generation; | 4459 | BTRFS_I(inode)->last_trans = root->fs_info->generation; |
| 4460 | BTRFS_I(inode)->last_sub_trans = root->log_transid; | 4460 | BTRFS_I(inode)->last_sub_trans = root->log_transid; |
| 4461 | BTRFS_I(inode)->last_log_commit = root->last_log_commit; | 4461 | BTRFS_I(inode)->last_log_commit = root->last_log_commit; |
| 4462 | return 0; | 4462 | return 0; |
| 4463 | } | 4463 | } |
| 4464 | 4464 | ||
| 4465 | /* | 4465 | /* |
| 4466 | * 1 - for the one we're dropping | 4466 | * 1 - for the one we're dropping |
| 4467 | * 1 - for the one we're adding | 4467 | * 1 - for the one we're adding |
| 4468 | * 1 - for updating the inode. | 4468 | * 1 - for updating the inode. |
| 4469 | */ | 4469 | */ |
| 4470 | trans = btrfs_start_transaction(root, 3); | 4470 | trans = btrfs_start_transaction(root, 3); |
| 4471 | if (IS_ERR(trans)) | 4471 | if (IS_ERR(trans)) |
| 4472 | return PTR_ERR(trans); | 4472 | return PTR_ERR(trans); |
| 4473 | 4473 | ||
| 4474 | ret = btrfs_drop_extents(trans, root, inode, offset, offset + len, 1); | 4474 | ret = btrfs_drop_extents(trans, root, inode, offset, offset + len, 1); |
| 4475 | if (ret) { | 4475 | if (ret) { |
| 4476 | btrfs_abort_transaction(trans, root, ret); | 4476 | btrfs_abort_transaction(trans, root, ret); |
| 4477 | btrfs_end_transaction(trans, root); | 4477 | btrfs_end_transaction(trans, root); |
| 4478 | return ret; | 4478 | return ret; |
| 4479 | } | 4479 | } |
| 4480 | 4480 | ||
| 4481 | ret = btrfs_insert_file_extent(trans, root, btrfs_ino(inode), offset, | 4481 | ret = btrfs_insert_file_extent(trans, root, btrfs_ino(inode), offset, |
| 4482 | 0, 0, len, 0, len, 0, 0, 0); | 4482 | 0, 0, len, 0, len, 0, 0, 0); |
| 4483 | if (ret) | 4483 | if (ret) |
| 4484 | btrfs_abort_transaction(trans, root, ret); | 4484 | btrfs_abort_transaction(trans, root, ret); |
| 4485 | else | 4485 | else |
| 4486 | btrfs_update_inode(trans, root, inode); | 4486 | btrfs_update_inode(trans, root, inode); |
| 4487 | btrfs_end_transaction(trans, root); | 4487 | btrfs_end_transaction(trans, root); |
| 4488 | return ret; | 4488 | return ret; |
| 4489 | } | 4489 | } |
| 4490 | 4490 | ||
| 4491 | /* | 4491 | /* |
| 4492 | * This function puts in dummy file extents for the area we're creating a hole | 4492 | * This function puts in dummy file extents for the area we're creating a hole |
| 4493 | * for. So if we are truncating this file to a larger size we need to insert | 4493 | * for. So if we are truncating this file to a larger size we need to insert |
| 4494 | * these file extents so that btrfs_get_extent will return a EXTENT_MAP_HOLE for | 4494 | * these file extents so that btrfs_get_extent will return a EXTENT_MAP_HOLE for |
| 4495 | * the range between oldsize and size | 4495 | * the range between oldsize and size |
| 4496 | */ | 4496 | */ |
| 4497 | int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) | 4497 | int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) |
| 4498 | { | 4498 | { |
| 4499 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4499 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4500 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 4500 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 4501 | struct extent_map *em = NULL; | 4501 | struct extent_map *em = NULL; |
| 4502 | struct extent_state *cached_state = NULL; | 4502 | struct extent_state *cached_state = NULL; |
| 4503 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; | 4503 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; |
| 4504 | u64 hole_start = ALIGN(oldsize, root->sectorsize); | 4504 | u64 hole_start = ALIGN(oldsize, root->sectorsize); |
| 4505 | u64 block_end = ALIGN(size, root->sectorsize); | 4505 | u64 block_end = ALIGN(size, root->sectorsize); |
| 4506 | u64 last_byte; | 4506 | u64 last_byte; |
| 4507 | u64 cur_offset; | 4507 | u64 cur_offset; |
| 4508 | u64 hole_size; | 4508 | u64 hole_size; |
| 4509 | int err = 0; | 4509 | int err = 0; |
| 4510 | 4510 | ||
| 4511 | /* | 4511 | /* |
| 4512 | * If our size started in the middle of a page we need to zero out the | 4512 | * If our size started in the middle of a page we need to zero out the |
| 4513 | * rest of the page before we expand the i_size, otherwise we could | 4513 | * rest of the page before we expand the i_size, otherwise we could |
| 4514 | * expose stale data. | 4514 | * expose stale data. |
| 4515 | */ | 4515 | */ |
| 4516 | err = btrfs_truncate_page(inode, oldsize, 0, 0); | 4516 | err = btrfs_truncate_page(inode, oldsize, 0, 0); |
| 4517 | if (err) | 4517 | if (err) |
| 4518 | return err; | 4518 | return err; |
| 4519 | 4519 | ||
| 4520 | if (size <= hole_start) | 4520 | if (size <= hole_start) |
| 4521 | return 0; | 4521 | return 0; |
| 4522 | 4522 | ||
| 4523 | while (1) { | 4523 | while (1) { |
| 4524 | struct btrfs_ordered_extent *ordered; | 4524 | struct btrfs_ordered_extent *ordered; |
| 4525 | 4525 | ||
| 4526 | lock_extent_bits(io_tree, hole_start, block_end - 1, 0, | 4526 | lock_extent_bits(io_tree, hole_start, block_end - 1, 0, |
| 4527 | &cached_state); | 4527 | &cached_state); |
| 4528 | ordered = btrfs_lookup_ordered_range(inode, hole_start, | 4528 | ordered = btrfs_lookup_ordered_range(inode, hole_start, |
| 4529 | block_end - hole_start); | 4529 | block_end - hole_start); |
| 4530 | if (!ordered) | 4530 | if (!ordered) |
| 4531 | break; | 4531 | break; |
| 4532 | unlock_extent_cached(io_tree, hole_start, block_end - 1, | 4532 | unlock_extent_cached(io_tree, hole_start, block_end - 1, |
| 4533 | &cached_state, GFP_NOFS); | 4533 | &cached_state, GFP_NOFS); |
| 4534 | btrfs_start_ordered_extent(inode, ordered, 1); | 4534 | btrfs_start_ordered_extent(inode, ordered, 1); |
| 4535 | btrfs_put_ordered_extent(ordered); | 4535 | btrfs_put_ordered_extent(ordered); |
| 4536 | } | 4536 | } |
| 4537 | 4537 | ||
| 4538 | cur_offset = hole_start; | 4538 | cur_offset = hole_start; |
| 4539 | while (1) { | 4539 | while (1) { |
| 4540 | em = btrfs_get_extent(inode, NULL, 0, cur_offset, | 4540 | em = btrfs_get_extent(inode, NULL, 0, cur_offset, |
| 4541 | block_end - cur_offset, 0); | 4541 | block_end - cur_offset, 0); |
| 4542 | if (IS_ERR(em)) { | 4542 | if (IS_ERR(em)) { |
| 4543 | err = PTR_ERR(em); | 4543 | err = PTR_ERR(em); |
| 4544 | em = NULL; | 4544 | em = NULL; |
| 4545 | break; | 4545 | break; |
| 4546 | } | 4546 | } |
| 4547 | last_byte = min(extent_map_end(em), block_end); | 4547 | last_byte = min(extent_map_end(em), block_end); |
| 4548 | last_byte = ALIGN(last_byte , root->sectorsize); | 4548 | last_byte = ALIGN(last_byte , root->sectorsize); |
| 4549 | if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) { | 4549 | if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) { |
| 4550 | struct extent_map *hole_em; | 4550 | struct extent_map *hole_em; |
| 4551 | hole_size = last_byte - cur_offset; | 4551 | hole_size = last_byte - cur_offset; |
| 4552 | 4552 | ||
| 4553 | err = maybe_insert_hole(root, inode, cur_offset, | 4553 | err = maybe_insert_hole(root, inode, cur_offset, |
| 4554 | hole_size); | 4554 | hole_size); |
| 4555 | if (err) | 4555 | if (err) |
| 4556 | break; | 4556 | break; |
| 4557 | btrfs_drop_extent_cache(inode, cur_offset, | 4557 | btrfs_drop_extent_cache(inode, cur_offset, |
| 4558 | cur_offset + hole_size - 1, 0); | 4558 | cur_offset + hole_size - 1, 0); |
| 4559 | hole_em = alloc_extent_map(); | 4559 | hole_em = alloc_extent_map(); |
| 4560 | if (!hole_em) { | 4560 | if (!hole_em) { |
| 4561 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, | 4561 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, |
| 4562 | &BTRFS_I(inode)->runtime_flags); | 4562 | &BTRFS_I(inode)->runtime_flags); |
| 4563 | goto next; | 4563 | goto next; |
| 4564 | } | 4564 | } |
| 4565 | hole_em->start = cur_offset; | 4565 | hole_em->start = cur_offset; |
| 4566 | hole_em->len = hole_size; | 4566 | hole_em->len = hole_size; |
| 4567 | hole_em->orig_start = cur_offset; | 4567 | hole_em->orig_start = cur_offset; |
| 4568 | 4568 | ||
| 4569 | hole_em->block_start = EXTENT_MAP_HOLE; | 4569 | hole_em->block_start = EXTENT_MAP_HOLE; |
| 4570 | hole_em->block_len = 0; | 4570 | hole_em->block_len = 0; |
| 4571 | hole_em->orig_block_len = 0; | 4571 | hole_em->orig_block_len = 0; |
| 4572 | hole_em->ram_bytes = hole_size; | 4572 | hole_em->ram_bytes = hole_size; |
| 4573 | hole_em->bdev = root->fs_info->fs_devices->latest_bdev; | 4573 | hole_em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 4574 | hole_em->compress_type = BTRFS_COMPRESS_NONE; | 4574 | hole_em->compress_type = BTRFS_COMPRESS_NONE; |
| 4575 | hole_em->generation = root->fs_info->generation; | 4575 | hole_em->generation = root->fs_info->generation; |
| 4576 | 4576 | ||
| 4577 | while (1) { | 4577 | while (1) { |
| 4578 | write_lock(&em_tree->lock); | 4578 | write_lock(&em_tree->lock); |
| 4579 | err = add_extent_mapping(em_tree, hole_em, 1); | 4579 | err = add_extent_mapping(em_tree, hole_em, 1); |
| 4580 | write_unlock(&em_tree->lock); | 4580 | write_unlock(&em_tree->lock); |
| 4581 | if (err != -EEXIST) | 4581 | if (err != -EEXIST) |
| 4582 | break; | 4582 | break; |
| 4583 | btrfs_drop_extent_cache(inode, cur_offset, | 4583 | btrfs_drop_extent_cache(inode, cur_offset, |
| 4584 | cur_offset + | 4584 | cur_offset + |
| 4585 | hole_size - 1, 0); | 4585 | hole_size - 1, 0); |
| 4586 | } | 4586 | } |
| 4587 | free_extent_map(hole_em); | 4587 | free_extent_map(hole_em); |
| 4588 | } | 4588 | } |
| 4589 | next: | 4589 | next: |
| 4590 | free_extent_map(em); | 4590 | free_extent_map(em); |
| 4591 | em = NULL; | 4591 | em = NULL; |
| 4592 | cur_offset = last_byte; | 4592 | cur_offset = last_byte; |
| 4593 | if (cur_offset >= block_end) | 4593 | if (cur_offset >= block_end) |
| 4594 | break; | 4594 | break; |
| 4595 | } | 4595 | } |
| 4596 | free_extent_map(em); | 4596 | free_extent_map(em); |
| 4597 | unlock_extent_cached(io_tree, hole_start, block_end - 1, &cached_state, | 4597 | unlock_extent_cached(io_tree, hole_start, block_end - 1, &cached_state, |
| 4598 | GFP_NOFS); | 4598 | GFP_NOFS); |
| 4599 | return err; | 4599 | return err; |
| 4600 | } | 4600 | } |
| 4601 | 4601 | ||
| 4602 | static int wait_snapshoting_atomic_t(atomic_t *a) | 4602 | static int wait_snapshoting_atomic_t(atomic_t *a) |
| 4603 | { | 4603 | { |
| 4604 | schedule(); | 4604 | schedule(); |
| 4605 | return 0; | 4605 | return 0; |
| 4606 | } | 4606 | } |
| 4607 | 4607 | ||
| 4608 | static void wait_for_snapshot_creation(struct btrfs_root *root) | 4608 | static void wait_for_snapshot_creation(struct btrfs_root *root) |
| 4609 | { | 4609 | { |
| 4610 | while (true) { | 4610 | while (true) { |
| 4611 | int ret; | 4611 | int ret; |
| 4612 | 4612 | ||
| 4613 | ret = btrfs_start_write_no_snapshoting(root); | 4613 | ret = btrfs_start_write_no_snapshoting(root); |
| 4614 | if (ret) | 4614 | if (ret) |
| 4615 | break; | 4615 | break; |
| 4616 | wait_on_atomic_t(&root->will_be_snapshoted, | 4616 | wait_on_atomic_t(&root->will_be_snapshoted, |
| 4617 | wait_snapshoting_atomic_t, | 4617 | wait_snapshoting_atomic_t, |
| 4618 | TASK_UNINTERRUPTIBLE); | 4618 | TASK_UNINTERRUPTIBLE); |
| 4619 | } | 4619 | } |
| 4620 | } | 4620 | } |
| 4621 | 4621 | ||
| 4622 | static int btrfs_setsize(struct inode *inode, struct iattr *attr) | 4622 | static int btrfs_setsize(struct inode *inode, struct iattr *attr) |
| 4623 | { | 4623 | { |
| 4624 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4624 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4625 | struct btrfs_trans_handle *trans; | 4625 | struct btrfs_trans_handle *trans; |
| 4626 | loff_t oldsize = i_size_read(inode); | 4626 | loff_t oldsize = i_size_read(inode); |
| 4627 | loff_t newsize = attr->ia_size; | 4627 | loff_t newsize = attr->ia_size; |
| 4628 | int mask = attr->ia_valid; | 4628 | int mask = attr->ia_valid; |
| 4629 | int ret; | 4629 | int ret; |
| 4630 | 4630 | ||
| 4631 | /* | 4631 | /* |
| 4632 | * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a | 4632 | * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a |
| 4633 | * special case where we need to update the times despite not having | 4633 | * special case where we need to update the times despite not having |
| 4634 | * these flags set. For all other operations the VFS set these flags | 4634 | * these flags set. For all other operations the VFS set these flags |
| 4635 | * explicitly if it wants a timestamp update. | 4635 | * explicitly if it wants a timestamp update. |
| 4636 | */ | 4636 | */ |
| 4637 | if (newsize != oldsize) { | 4637 | if (newsize != oldsize) { |
| 4638 | inode_inc_iversion(inode); | 4638 | inode_inc_iversion(inode); |
| 4639 | if (!(mask & (ATTR_CTIME | ATTR_MTIME))) | 4639 | if (!(mask & (ATTR_CTIME | ATTR_MTIME))) |
| 4640 | inode->i_ctime = inode->i_mtime = | 4640 | inode->i_ctime = inode->i_mtime = |
| 4641 | current_fs_time(inode->i_sb); | 4641 | current_fs_time(inode->i_sb); |
| 4642 | } | 4642 | } |
| 4643 | 4643 | ||
| 4644 | if (newsize > oldsize) { | 4644 | if (newsize > oldsize) { |
| 4645 | truncate_pagecache(inode, newsize); | 4645 | truncate_pagecache(inode, newsize); |
| 4646 | /* | 4646 | /* |
| 4647 | * Don't do an expanding truncate while snapshoting is ongoing. | 4647 | * Don't do an expanding truncate while snapshoting is ongoing. |
| 4648 | * This is to ensure the snapshot captures a fully consistent | 4648 | * This is to ensure the snapshot captures a fully consistent |
| 4649 | * state of this file - if the snapshot captures this expanding | 4649 | * state of this file - if the snapshot captures this expanding |
| 4650 | * truncation, it must capture all writes that happened before | 4650 | * truncation, it must capture all writes that happened before |
| 4651 | * this truncation. | 4651 | * this truncation. |
| 4652 | */ | 4652 | */ |
| 4653 | wait_for_snapshot_creation(root); | 4653 | wait_for_snapshot_creation(root); |
| 4654 | ret = btrfs_cont_expand(inode, oldsize, newsize); | 4654 | ret = btrfs_cont_expand(inode, oldsize, newsize); |
| 4655 | if (ret) { | 4655 | if (ret) { |
| 4656 | btrfs_end_write_no_snapshoting(root); | 4656 | btrfs_end_write_no_snapshoting(root); |
| 4657 | return ret; | 4657 | return ret; |
| 4658 | } | 4658 | } |
| 4659 | 4659 | ||
| 4660 | trans = btrfs_start_transaction(root, 1); | 4660 | trans = btrfs_start_transaction(root, 1); |
| 4661 | if (IS_ERR(trans)) { | 4661 | if (IS_ERR(trans)) { |
| 4662 | btrfs_end_write_no_snapshoting(root); | 4662 | btrfs_end_write_no_snapshoting(root); |
| 4663 | return PTR_ERR(trans); | 4663 | return PTR_ERR(trans); |
| 4664 | } | 4664 | } |
| 4665 | 4665 | ||
| 4666 | i_size_write(inode, newsize); | 4666 | i_size_write(inode, newsize); |
| 4667 | btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); | 4667 | btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); |
| 4668 | ret = btrfs_update_inode(trans, root, inode); | 4668 | ret = btrfs_update_inode(trans, root, inode); |
| 4669 | btrfs_end_write_no_snapshoting(root); | 4669 | btrfs_end_write_no_snapshoting(root); |
| 4670 | btrfs_end_transaction(trans, root); | 4670 | btrfs_end_transaction(trans, root); |
| 4671 | } else { | 4671 | } else { |
| 4672 | 4672 | ||
| 4673 | /* | 4673 | /* |
| 4674 | * We're truncating a file that used to have good data down to | 4674 | * We're truncating a file that used to have good data down to |
| 4675 | * zero. Make sure it gets into the ordered flush list so that | 4675 | * zero. Make sure it gets into the ordered flush list so that |
| 4676 | * any new writes get down to disk quickly. | 4676 | * any new writes get down to disk quickly. |
| 4677 | */ | 4677 | */ |
| 4678 | if (newsize == 0) | 4678 | if (newsize == 0) |
| 4679 | set_bit(BTRFS_INODE_ORDERED_DATA_CLOSE, | 4679 | set_bit(BTRFS_INODE_ORDERED_DATA_CLOSE, |
| 4680 | &BTRFS_I(inode)->runtime_flags); | 4680 | &BTRFS_I(inode)->runtime_flags); |
| 4681 | 4681 | ||
| 4682 | /* | 4682 | /* |
| 4683 | * 1 for the orphan item we're going to add | 4683 | * 1 for the orphan item we're going to add |
| 4684 | * 1 for the orphan item deletion. | 4684 | * 1 for the orphan item deletion. |
| 4685 | */ | 4685 | */ |
| 4686 | trans = btrfs_start_transaction(root, 2); | 4686 | trans = btrfs_start_transaction(root, 2); |
| 4687 | if (IS_ERR(trans)) | 4687 | if (IS_ERR(trans)) |
| 4688 | return PTR_ERR(trans); | 4688 | return PTR_ERR(trans); |
| 4689 | 4689 | ||
| 4690 | /* | 4690 | /* |
| 4691 | * We need to do this in case we fail at _any_ point during the | 4691 | * We need to do this in case we fail at _any_ point during the |
| 4692 | * actual truncate. Once we do the truncate_setsize we could | 4692 | * actual truncate. Once we do the truncate_setsize we could |
| 4693 | * invalidate pages which forces any outstanding ordered io to | 4693 | * invalidate pages which forces any outstanding ordered io to |
| 4694 | * be instantly completed which will give us extents that need | 4694 | * be instantly completed which will give us extents that need |
| 4695 | * to be truncated. If we fail to get an orphan inode down we | 4695 | * to be truncated. If we fail to get an orphan inode down we |
| 4696 | * could have left over extents that were never meant to live, | 4696 | * could have left over extents that were never meant to live, |
| 4697 | * so we need to garuntee from this point on that everything | 4697 | * so we need to garuntee from this point on that everything |
| 4698 | * will be consistent. | 4698 | * will be consistent. |
| 4699 | */ | 4699 | */ |
| 4700 | ret = btrfs_orphan_add(trans, inode); | 4700 | ret = btrfs_orphan_add(trans, inode); |
| 4701 | btrfs_end_transaction(trans, root); | 4701 | btrfs_end_transaction(trans, root); |
| 4702 | if (ret) | 4702 | if (ret) |
| 4703 | return ret; | 4703 | return ret; |
| 4704 | 4704 | ||
| 4705 | /* we don't support swapfiles, so vmtruncate shouldn't fail */ | 4705 | /* we don't support swapfiles, so vmtruncate shouldn't fail */ |
| 4706 | truncate_setsize(inode, newsize); | 4706 | truncate_setsize(inode, newsize); |
| 4707 | 4707 | ||
| 4708 | /* Disable nonlocked read DIO to avoid the end less truncate */ | 4708 | /* Disable nonlocked read DIO to avoid the end less truncate */ |
| 4709 | btrfs_inode_block_unlocked_dio(inode); | 4709 | btrfs_inode_block_unlocked_dio(inode); |
| 4710 | inode_dio_wait(inode); | 4710 | inode_dio_wait(inode); |
| 4711 | btrfs_inode_resume_unlocked_dio(inode); | 4711 | btrfs_inode_resume_unlocked_dio(inode); |
| 4712 | 4712 | ||
| 4713 | ret = btrfs_truncate(inode); | 4713 | ret = btrfs_truncate(inode); |
| 4714 | if (ret && inode->i_nlink) { | 4714 | if (ret && inode->i_nlink) { |
| 4715 | int err; | 4715 | int err; |
| 4716 | 4716 | ||
| 4717 | /* | 4717 | /* |
| 4718 | * failed to truncate, disk_i_size is only adjusted down | 4718 | * failed to truncate, disk_i_size is only adjusted down |
| 4719 | * as we remove extents, so it should represent the true | 4719 | * as we remove extents, so it should represent the true |
| 4720 | * size of the inode, so reset the in memory size and | 4720 | * size of the inode, so reset the in memory size and |
| 4721 | * delete our orphan entry. | 4721 | * delete our orphan entry. |
| 4722 | */ | 4722 | */ |
| 4723 | trans = btrfs_join_transaction(root); | 4723 | trans = btrfs_join_transaction(root); |
| 4724 | if (IS_ERR(trans)) { | 4724 | if (IS_ERR(trans)) { |
| 4725 | btrfs_orphan_del(NULL, inode); | 4725 | btrfs_orphan_del(NULL, inode); |
| 4726 | return ret; | 4726 | return ret; |
| 4727 | } | 4727 | } |
| 4728 | i_size_write(inode, BTRFS_I(inode)->disk_i_size); | 4728 | i_size_write(inode, BTRFS_I(inode)->disk_i_size); |
| 4729 | err = btrfs_orphan_del(trans, inode); | 4729 | err = btrfs_orphan_del(trans, inode); |
| 4730 | if (err) | 4730 | if (err) |
| 4731 | btrfs_abort_transaction(trans, root, err); | 4731 | btrfs_abort_transaction(trans, root, err); |
| 4732 | btrfs_end_transaction(trans, root); | 4732 | btrfs_end_transaction(trans, root); |
| 4733 | } | 4733 | } |
| 4734 | } | 4734 | } |
| 4735 | 4735 | ||
| 4736 | return ret; | 4736 | return ret; |
| 4737 | } | 4737 | } |
| 4738 | 4738 | ||
| 4739 | static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) | 4739 | static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) |
| 4740 | { | 4740 | { |
| 4741 | struct inode *inode = dentry->d_inode; | 4741 | struct inode *inode = dentry->d_inode; |
| 4742 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4742 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4743 | int err; | 4743 | int err; |
| 4744 | 4744 | ||
| 4745 | if (btrfs_root_readonly(root)) | 4745 | if (btrfs_root_readonly(root)) |
| 4746 | return -EROFS; | 4746 | return -EROFS; |
| 4747 | 4747 | ||
| 4748 | err = inode_change_ok(inode, attr); | 4748 | err = inode_change_ok(inode, attr); |
| 4749 | if (err) | 4749 | if (err) |
| 4750 | return err; | 4750 | return err; |
| 4751 | 4751 | ||
| 4752 | if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) { | 4752 | if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) { |
| 4753 | err = btrfs_setsize(inode, attr); | 4753 | err = btrfs_setsize(inode, attr); |
| 4754 | if (err) | 4754 | if (err) |
| 4755 | return err; | 4755 | return err; |
| 4756 | } | 4756 | } |
| 4757 | 4757 | ||
| 4758 | if (attr->ia_valid) { | 4758 | if (attr->ia_valid) { |
| 4759 | setattr_copy(inode, attr); | 4759 | setattr_copy(inode, attr); |
| 4760 | inode_inc_iversion(inode); | 4760 | inode_inc_iversion(inode); |
| 4761 | err = btrfs_dirty_inode(inode); | 4761 | err = btrfs_dirty_inode(inode); |
| 4762 | 4762 | ||
| 4763 | if (!err && attr->ia_valid & ATTR_MODE) | 4763 | if (!err && attr->ia_valid & ATTR_MODE) |
| 4764 | err = posix_acl_chmod(inode, inode->i_mode); | 4764 | err = posix_acl_chmod(inode, inode->i_mode); |
| 4765 | } | 4765 | } |
| 4766 | 4766 | ||
| 4767 | return err; | 4767 | return err; |
| 4768 | } | 4768 | } |
| 4769 | 4769 | ||
| 4770 | /* | 4770 | /* |
| 4771 | * While truncating the inode pages during eviction, we get the VFS calling | 4771 | * While truncating the inode pages during eviction, we get the VFS calling |
| 4772 | * btrfs_invalidatepage() against each page of the inode. This is slow because | 4772 | * btrfs_invalidatepage() against each page of the inode. This is slow because |
| 4773 | * the calls to btrfs_invalidatepage() result in a huge amount of calls to | 4773 | * the calls to btrfs_invalidatepage() result in a huge amount of calls to |
| 4774 | * lock_extent_bits() and clear_extent_bit(), which keep merging and splitting | 4774 | * lock_extent_bits() and clear_extent_bit(), which keep merging and splitting |
| 4775 | * extent_state structures over and over, wasting lots of time. | 4775 | * extent_state structures over and over, wasting lots of time. |
| 4776 | * | 4776 | * |
| 4777 | * Therefore if the inode is being evicted, let btrfs_invalidatepage() skip all | 4777 | * Therefore if the inode is being evicted, let btrfs_invalidatepage() skip all |
| 4778 | * those expensive operations on a per page basis and do only the ordered io | 4778 | * those expensive operations on a per page basis and do only the ordered io |
| 4779 | * finishing, while we release here the extent_map and extent_state structures, | 4779 | * finishing, while we release here the extent_map and extent_state structures, |
| 4780 | * without the excessive merging and splitting. | 4780 | * without the excessive merging and splitting. |
| 4781 | */ | 4781 | */ |
| 4782 | static void evict_inode_truncate_pages(struct inode *inode) | 4782 | static void evict_inode_truncate_pages(struct inode *inode) |
| 4783 | { | 4783 | { |
| 4784 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 4784 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 4785 | struct extent_map_tree *map_tree = &BTRFS_I(inode)->extent_tree; | 4785 | struct extent_map_tree *map_tree = &BTRFS_I(inode)->extent_tree; |
| 4786 | struct rb_node *node; | 4786 | struct rb_node *node; |
| 4787 | 4787 | ||
| 4788 | ASSERT(inode->i_state & I_FREEING); | 4788 | ASSERT(inode->i_state & I_FREEING); |
| 4789 | truncate_inode_pages_final(&inode->i_data); | 4789 | truncate_inode_pages_final(&inode->i_data); |
| 4790 | 4790 | ||
| 4791 | write_lock(&map_tree->lock); | 4791 | write_lock(&map_tree->lock); |
| 4792 | while (!RB_EMPTY_ROOT(&map_tree->map)) { | 4792 | while (!RB_EMPTY_ROOT(&map_tree->map)) { |
| 4793 | struct extent_map *em; | 4793 | struct extent_map *em; |
| 4794 | 4794 | ||
| 4795 | node = rb_first(&map_tree->map); | 4795 | node = rb_first(&map_tree->map); |
| 4796 | em = rb_entry(node, struct extent_map, rb_node); | 4796 | em = rb_entry(node, struct extent_map, rb_node); |
| 4797 | clear_bit(EXTENT_FLAG_PINNED, &em->flags); | 4797 | clear_bit(EXTENT_FLAG_PINNED, &em->flags); |
| 4798 | clear_bit(EXTENT_FLAG_LOGGING, &em->flags); | 4798 | clear_bit(EXTENT_FLAG_LOGGING, &em->flags); |
| 4799 | remove_extent_mapping(map_tree, em); | 4799 | remove_extent_mapping(map_tree, em); |
| 4800 | free_extent_map(em); | 4800 | free_extent_map(em); |
| 4801 | if (need_resched()) { | 4801 | if (need_resched()) { |
| 4802 | write_unlock(&map_tree->lock); | 4802 | write_unlock(&map_tree->lock); |
| 4803 | cond_resched(); | 4803 | cond_resched(); |
| 4804 | write_lock(&map_tree->lock); | 4804 | write_lock(&map_tree->lock); |
| 4805 | } | 4805 | } |
| 4806 | } | 4806 | } |
| 4807 | write_unlock(&map_tree->lock); | 4807 | write_unlock(&map_tree->lock); |
| 4808 | 4808 | ||
| 4809 | spin_lock(&io_tree->lock); | 4809 | spin_lock(&io_tree->lock); |
| 4810 | while (!RB_EMPTY_ROOT(&io_tree->state)) { | 4810 | while (!RB_EMPTY_ROOT(&io_tree->state)) { |
| 4811 | struct extent_state *state; | 4811 | struct extent_state *state; |
| 4812 | struct extent_state *cached_state = NULL; | 4812 | struct extent_state *cached_state = NULL; |
| 4813 | 4813 | ||
| 4814 | node = rb_first(&io_tree->state); | 4814 | node = rb_first(&io_tree->state); |
| 4815 | state = rb_entry(node, struct extent_state, rb_node); | 4815 | state = rb_entry(node, struct extent_state, rb_node); |
| 4816 | atomic_inc(&state->refs); | 4816 | atomic_inc(&state->refs); |
| 4817 | spin_unlock(&io_tree->lock); | 4817 | spin_unlock(&io_tree->lock); |
| 4818 | 4818 | ||
| 4819 | lock_extent_bits(io_tree, state->start, state->end, | 4819 | lock_extent_bits(io_tree, state->start, state->end, |
| 4820 | 0, &cached_state); | 4820 | 0, &cached_state); |
| 4821 | clear_extent_bit(io_tree, state->start, state->end, | 4821 | clear_extent_bit(io_tree, state->start, state->end, |
| 4822 | EXTENT_LOCKED | EXTENT_DIRTY | | 4822 | EXTENT_LOCKED | EXTENT_DIRTY | |
| 4823 | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | | 4823 | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | |
| 4824 | EXTENT_DEFRAG, 1, 1, | 4824 | EXTENT_DEFRAG, 1, 1, |
| 4825 | &cached_state, GFP_NOFS); | 4825 | &cached_state, GFP_NOFS); |
| 4826 | free_extent_state(state); | 4826 | free_extent_state(state); |
| 4827 | 4827 | ||
| 4828 | cond_resched(); | 4828 | cond_resched(); |
| 4829 | spin_lock(&io_tree->lock); | 4829 | spin_lock(&io_tree->lock); |
| 4830 | } | 4830 | } |
| 4831 | spin_unlock(&io_tree->lock); | 4831 | spin_unlock(&io_tree->lock); |
| 4832 | } | 4832 | } |
| 4833 | 4833 | ||
| 4834 | void btrfs_evict_inode(struct inode *inode) | 4834 | void btrfs_evict_inode(struct inode *inode) |
| 4835 | { | 4835 | { |
| 4836 | struct btrfs_trans_handle *trans; | 4836 | struct btrfs_trans_handle *trans; |
| 4837 | struct btrfs_root *root = BTRFS_I(inode)->root; | 4837 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 4838 | struct btrfs_block_rsv *rsv, *global_rsv; | 4838 | struct btrfs_block_rsv *rsv, *global_rsv; |
| 4839 | u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); | 4839 | u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); |
| 4840 | int ret; | 4840 | int ret; |
| 4841 | 4841 | ||
| 4842 | trace_btrfs_inode_evict(inode); | 4842 | trace_btrfs_inode_evict(inode); |
| 4843 | 4843 | ||
| 4844 | evict_inode_truncate_pages(inode); | 4844 | evict_inode_truncate_pages(inode); |
| 4845 | 4845 | ||
| 4846 | if (inode->i_nlink && | 4846 | if (inode->i_nlink && |
| 4847 | ((btrfs_root_refs(&root->root_item) != 0 && | 4847 | ((btrfs_root_refs(&root->root_item) != 0 && |
| 4848 | root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID) || | 4848 | root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID) || |
| 4849 | btrfs_is_free_space_inode(inode))) | 4849 | btrfs_is_free_space_inode(inode))) |
| 4850 | goto no_delete; | 4850 | goto no_delete; |
| 4851 | 4851 | ||
| 4852 | if (is_bad_inode(inode)) { | 4852 | if (is_bad_inode(inode)) { |
| 4853 | btrfs_orphan_del(NULL, inode); | 4853 | btrfs_orphan_del(NULL, inode); |
| 4854 | goto no_delete; | 4854 | goto no_delete; |
| 4855 | } | 4855 | } |
| 4856 | /* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */ | 4856 | /* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */ |
| 4857 | btrfs_wait_ordered_range(inode, 0, (u64)-1); | 4857 | btrfs_wait_ordered_range(inode, 0, (u64)-1); |
| 4858 | 4858 | ||
| 4859 | btrfs_free_io_failure_record(inode, 0, (u64)-1); | 4859 | btrfs_free_io_failure_record(inode, 0, (u64)-1); |
| 4860 | 4860 | ||
| 4861 | if (root->fs_info->log_root_recovering) { | 4861 | if (root->fs_info->log_root_recovering) { |
| 4862 | BUG_ON(test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, | 4862 | BUG_ON(test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, |
| 4863 | &BTRFS_I(inode)->runtime_flags)); | 4863 | &BTRFS_I(inode)->runtime_flags)); |
| 4864 | goto no_delete; | 4864 | goto no_delete; |
| 4865 | } | 4865 | } |
| 4866 | 4866 | ||
| 4867 | if (inode->i_nlink > 0) { | 4867 | if (inode->i_nlink > 0) { |
| 4868 | BUG_ON(btrfs_root_refs(&root->root_item) != 0 && | 4868 | BUG_ON(btrfs_root_refs(&root->root_item) != 0 && |
| 4869 | root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID); | 4869 | root->root_key.objectid != BTRFS_ROOT_TREE_OBJECTID); |
| 4870 | goto no_delete; | 4870 | goto no_delete; |
| 4871 | } | 4871 | } |
| 4872 | 4872 | ||
| 4873 | ret = btrfs_commit_inode_delayed_inode(inode); | 4873 | ret = btrfs_commit_inode_delayed_inode(inode); |
| 4874 | if (ret) { | 4874 | if (ret) { |
| 4875 | btrfs_orphan_del(NULL, inode); | 4875 | btrfs_orphan_del(NULL, inode); |
| 4876 | goto no_delete; | 4876 | goto no_delete; |
| 4877 | } | 4877 | } |
| 4878 | 4878 | ||
| 4879 | rsv = btrfs_alloc_block_rsv(root, BTRFS_BLOCK_RSV_TEMP); | 4879 | rsv = btrfs_alloc_block_rsv(root, BTRFS_BLOCK_RSV_TEMP); |
| 4880 | if (!rsv) { | 4880 | if (!rsv) { |
| 4881 | btrfs_orphan_del(NULL, inode); | 4881 | btrfs_orphan_del(NULL, inode); |
| 4882 | goto no_delete; | 4882 | goto no_delete; |
| 4883 | } | 4883 | } |
| 4884 | rsv->size = min_size; | 4884 | rsv->size = min_size; |
| 4885 | rsv->failfast = 1; | 4885 | rsv->failfast = 1; |
| 4886 | global_rsv = &root->fs_info->global_block_rsv; | 4886 | global_rsv = &root->fs_info->global_block_rsv; |
| 4887 | 4887 | ||
| 4888 | btrfs_i_size_write(inode, 0); | 4888 | btrfs_i_size_write(inode, 0); |
| 4889 | 4889 | ||
| 4890 | /* | 4890 | /* |
| 4891 | * This is a bit simpler than btrfs_truncate since we've already | 4891 | * This is a bit simpler than btrfs_truncate since we've already |
| 4892 | * reserved our space for our orphan item in the unlink, so we just | 4892 | * reserved our space for our orphan item in the unlink, so we just |
| 4893 | * need to reserve some slack space in case we add bytes and update | 4893 | * need to reserve some slack space in case we add bytes and update |
| 4894 | * inode item when doing the truncate. | 4894 | * inode item when doing the truncate. |
| 4895 | */ | 4895 | */ |
| 4896 | while (1) { | 4896 | while (1) { |
| 4897 | ret = btrfs_block_rsv_refill(root, rsv, min_size, | 4897 | ret = btrfs_block_rsv_refill(root, rsv, min_size, |
| 4898 | BTRFS_RESERVE_FLUSH_LIMIT); | 4898 | BTRFS_RESERVE_FLUSH_LIMIT); |
| 4899 | 4899 | ||
| 4900 | /* | 4900 | /* |
| 4901 | * Try and steal from the global reserve since we will | 4901 | * Try and steal from the global reserve since we will |
| 4902 | * likely not use this space anyway, we want to try as | 4902 | * likely not use this space anyway, we want to try as |
| 4903 | * hard as possible to get this to work. | 4903 | * hard as possible to get this to work. |
| 4904 | */ | 4904 | */ |
| 4905 | if (ret) | 4905 | if (ret) |
| 4906 | ret = btrfs_block_rsv_migrate(global_rsv, rsv, min_size); | 4906 | ret = btrfs_block_rsv_migrate(global_rsv, rsv, min_size); |
| 4907 | 4907 | ||
| 4908 | if (ret) { | 4908 | if (ret) { |
| 4909 | btrfs_warn(root->fs_info, | 4909 | btrfs_warn(root->fs_info, |
| 4910 | "Could not get space for a delete, will truncate on mount %d", | 4910 | "Could not get space for a delete, will truncate on mount %d", |
| 4911 | ret); | 4911 | ret); |
| 4912 | btrfs_orphan_del(NULL, inode); | 4912 | btrfs_orphan_del(NULL, inode); |
| 4913 | btrfs_free_block_rsv(root, rsv); | 4913 | btrfs_free_block_rsv(root, rsv); |
| 4914 | goto no_delete; | 4914 | goto no_delete; |
| 4915 | } | 4915 | } |
| 4916 | 4916 | ||
| 4917 | trans = btrfs_join_transaction(root); | 4917 | trans = btrfs_join_transaction(root); |
| 4918 | if (IS_ERR(trans)) { | 4918 | if (IS_ERR(trans)) { |
| 4919 | btrfs_orphan_del(NULL, inode); | 4919 | btrfs_orphan_del(NULL, inode); |
| 4920 | btrfs_free_block_rsv(root, rsv); | 4920 | btrfs_free_block_rsv(root, rsv); |
| 4921 | goto no_delete; | 4921 | goto no_delete; |
| 4922 | } | 4922 | } |
| 4923 | 4923 | ||
| 4924 | trans->block_rsv = rsv; | 4924 | trans->block_rsv = rsv; |
| 4925 | 4925 | ||
| 4926 | ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0); | 4926 | ret = btrfs_truncate_inode_items(trans, root, inode, 0, 0); |
| 4927 | if (ret != -ENOSPC) | 4927 | if (ret != -ENOSPC) |
| 4928 | break; | 4928 | break; |
| 4929 | 4929 | ||
| 4930 | trans->block_rsv = &root->fs_info->trans_block_rsv; | 4930 | trans->block_rsv = &root->fs_info->trans_block_rsv; |
| 4931 | btrfs_end_transaction(trans, root); | 4931 | btrfs_end_transaction(trans, root); |
| 4932 | trans = NULL; | 4932 | trans = NULL; |
| 4933 | btrfs_btree_balance_dirty(root); | 4933 | btrfs_btree_balance_dirty(root); |
| 4934 | } | 4934 | } |
| 4935 | 4935 | ||
| 4936 | btrfs_free_block_rsv(root, rsv); | 4936 | btrfs_free_block_rsv(root, rsv); |
| 4937 | 4937 | ||
| 4938 | /* | 4938 | /* |
| 4939 | * Errors here aren't a big deal, it just means we leave orphan items | 4939 | * Errors here aren't a big deal, it just means we leave orphan items |
| 4940 | * in the tree. They will be cleaned up on the next mount. | 4940 | * in the tree. They will be cleaned up on the next mount. |
| 4941 | */ | 4941 | */ |
| 4942 | if (ret == 0) { | 4942 | if (ret == 0) { |
| 4943 | trans->block_rsv = root->orphan_block_rsv; | 4943 | trans->block_rsv = root->orphan_block_rsv; |
| 4944 | btrfs_orphan_del(trans, inode); | 4944 | btrfs_orphan_del(trans, inode); |
| 4945 | } else { | 4945 | } else { |
| 4946 | btrfs_orphan_del(NULL, inode); | 4946 | btrfs_orphan_del(NULL, inode); |
| 4947 | } | 4947 | } |
| 4948 | 4948 | ||
| 4949 | trans->block_rsv = &root->fs_info->trans_block_rsv; | 4949 | trans->block_rsv = &root->fs_info->trans_block_rsv; |
| 4950 | if (!(root == root->fs_info->tree_root || | 4950 | if (!(root == root->fs_info->tree_root || |
| 4951 | root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) | 4951 | root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) |
| 4952 | btrfs_return_ino(root, btrfs_ino(inode)); | 4952 | btrfs_return_ino(root, btrfs_ino(inode)); |
| 4953 | 4953 | ||
| 4954 | btrfs_end_transaction(trans, root); | 4954 | btrfs_end_transaction(trans, root); |
| 4955 | btrfs_btree_balance_dirty(root); | 4955 | btrfs_btree_balance_dirty(root); |
| 4956 | no_delete: | 4956 | no_delete: |
| 4957 | btrfs_remove_delayed_node(inode); | 4957 | btrfs_remove_delayed_node(inode); |
| 4958 | clear_inode(inode); | 4958 | clear_inode(inode); |
| 4959 | return; | 4959 | return; |
| 4960 | } | 4960 | } |
| 4961 | 4961 | ||
| 4962 | /* | 4962 | /* |
| 4963 | * this returns the key found in the dir entry in the location pointer. | 4963 | * this returns the key found in the dir entry in the location pointer. |
| 4964 | * If no dir entries were found, location->objectid is 0. | 4964 | * If no dir entries were found, location->objectid is 0. |
| 4965 | */ | 4965 | */ |
| 4966 | static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry, | 4966 | static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry, |
| 4967 | struct btrfs_key *location) | 4967 | struct btrfs_key *location) |
| 4968 | { | 4968 | { |
| 4969 | const char *name = dentry->d_name.name; | 4969 | const char *name = dentry->d_name.name; |
| 4970 | int namelen = dentry->d_name.len; | 4970 | int namelen = dentry->d_name.len; |
| 4971 | struct btrfs_dir_item *di; | 4971 | struct btrfs_dir_item *di; |
| 4972 | struct btrfs_path *path; | 4972 | struct btrfs_path *path; |
| 4973 | struct btrfs_root *root = BTRFS_I(dir)->root; | 4973 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 4974 | int ret = 0; | 4974 | int ret = 0; |
| 4975 | 4975 | ||
| 4976 | path = btrfs_alloc_path(); | 4976 | path = btrfs_alloc_path(); |
| 4977 | if (!path) | 4977 | if (!path) |
| 4978 | return -ENOMEM; | 4978 | return -ENOMEM; |
| 4979 | 4979 | ||
| 4980 | di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(dir), name, | 4980 | di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(dir), name, |
| 4981 | namelen, 0); | 4981 | namelen, 0); |
| 4982 | if (IS_ERR(di)) | 4982 | if (IS_ERR(di)) |
| 4983 | ret = PTR_ERR(di); | 4983 | ret = PTR_ERR(di); |
| 4984 | 4984 | ||
| 4985 | if (IS_ERR_OR_NULL(di)) | 4985 | if (IS_ERR_OR_NULL(di)) |
| 4986 | goto out_err; | 4986 | goto out_err; |
| 4987 | 4987 | ||
| 4988 | btrfs_dir_item_key_to_cpu(path->nodes[0], di, location); | 4988 | btrfs_dir_item_key_to_cpu(path->nodes[0], di, location); |
| 4989 | out: | 4989 | out: |
| 4990 | btrfs_free_path(path); | 4990 | btrfs_free_path(path); |
| 4991 | return ret; | 4991 | return ret; |
| 4992 | out_err: | 4992 | out_err: |
| 4993 | location->objectid = 0; | 4993 | location->objectid = 0; |
| 4994 | goto out; | 4994 | goto out; |
| 4995 | } | 4995 | } |
| 4996 | 4996 | ||
| 4997 | /* | 4997 | /* |
| 4998 | * when we hit a tree root in a directory, the btrfs part of the inode | 4998 | * when we hit a tree root in a directory, the btrfs part of the inode |
| 4999 | * needs to be changed to reflect the root directory of the tree root. This | 4999 | * needs to be changed to reflect the root directory of the tree root. This |
| 5000 | * is kind of like crossing a mount point. | 5000 | * is kind of like crossing a mount point. |
| 5001 | */ | 5001 | */ |
| 5002 | static int fixup_tree_root_location(struct btrfs_root *root, | 5002 | static int fixup_tree_root_location(struct btrfs_root *root, |
| 5003 | struct inode *dir, | 5003 | struct inode *dir, |
| 5004 | struct dentry *dentry, | 5004 | struct dentry *dentry, |
| 5005 | struct btrfs_key *location, | 5005 | struct btrfs_key *location, |
| 5006 | struct btrfs_root **sub_root) | 5006 | struct btrfs_root **sub_root) |
| 5007 | { | 5007 | { |
| 5008 | struct btrfs_path *path; | 5008 | struct btrfs_path *path; |
| 5009 | struct btrfs_root *new_root; | 5009 | struct btrfs_root *new_root; |
| 5010 | struct btrfs_root_ref *ref; | 5010 | struct btrfs_root_ref *ref; |
| 5011 | struct extent_buffer *leaf; | 5011 | struct extent_buffer *leaf; |
| 5012 | int ret; | 5012 | int ret; |
| 5013 | int err = 0; | 5013 | int err = 0; |
| 5014 | 5014 | ||
| 5015 | path = btrfs_alloc_path(); | 5015 | path = btrfs_alloc_path(); |
| 5016 | if (!path) { | 5016 | if (!path) { |
| 5017 | err = -ENOMEM; | 5017 | err = -ENOMEM; |
| 5018 | goto out; | 5018 | goto out; |
| 5019 | } | 5019 | } |
| 5020 | 5020 | ||
| 5021 | err = -ENOENT; | 5021 | err = -ENOENT; |
| 5022 | ret = btrfs_find_item(root->fs_info->tree_root, path, | 5022 | ret = btrfs_find_item(root->fs_info->tree_root, path, |
| 5023 | BTRFS_I(dir)->root->root_key.objectid, | 5023 | BTRFS_I(dir)->root->root_key.objectid, |
| 5024 | location->objectid, BTRFS_ROOT_REF_KEY, NULL); | 5024 | location->objectid, BTRFS_ROOT_REF_KEY, NULL); |
| 5025 | if (ret) { | 5025 | if (ret) { |
| 5026 | if (ret < 0) | 5026 | if (ret < 0) |
| 5027 | err = ret; | 5027 | err = ret; |
| 5028 | goto out; | 5028 | goto out; |
| 5029 | } | 5029 | } |
| 5030 | 5030 | ||
| 5031 | leaf = path->nodes[0]; | 5031 | leaf = path->nodes[0]; |
| 5032 | ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); | 5032 | ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); |
| 5033 | if (btrfs_root_ref_dirid(leaf, ref) != btrfs_ino(dir) || | 5033 | if (btrfs_root_ref_dirid(leaf, ref) != btrfs_ino(dir) || |
| 5034 | btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len) | 5034 | btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len) |
| 5035 | goto out; | 5035 | goto out; |
| 5036 | 5036 | ||
| 5037 | ret = memcmp_extent_buffer(leaf, dentry->d_name.name, | 5037 | ret = memcmp_extent_buffer(leaf, dentry->d_name.name, |
| 5038 | (unsigned long)(ref + 1), | 5038 | (unsigned long)(ref + 1), |
| 5039 | dentry->d_name.len); | 5039 | dentry->d_name.len); |
| 5040 | if (ret) | 5040 | if (ret) |
| 5041 | goto out; | 5041 | goto out; |
| 5042 | 5042 | ||
| 5043 | btrfs_release_path(path); | 5043 | btrfs_release_path(path); |
| 5044 | 5044 | ||
| 5045 | new_root = btrfs_read_fs_root_no_name(root->fs_info, location); | 5045 | new_root = btrfs_read_fs_root_no_name(root->fs_info, location); |
| 5046 | if (IS_ERR(new_root)) { | 5046 | if (IS_ERR(new_root)) { |
| 5047 | err = PTR_ERR(new_root); | 5047 | err = PTR_ERR(new_root); |
| 5048 | goto out; | 5048 | goto out; |
| 5049 | } | 5049 | } |
| 5050 | 5050 | ||
| 5051 | *sub_root = new_root; | 5051 | *sub_root = new_root; |
| 5052 | location->objectid = btrfs_root_dirid(&new_root->root_item); | 5052 | location->objectid = btrfs_root_dirid(&new_root->root_item); |
| 5053 | location->type = BTRFS_INODE_ITEM_KEY; | 5053 | location->type = BTRFS_INODE_ITEM_KEY; |
| 5054 | location->offset = 0; | 5054 | location->offset = 0; |
| 5055 | err = 0; | 5055 | err = 0; |
| 5056 | out: | 5056 | out: |
| 5057 | btrfs_free_path(path); | 5057 | btrfs_free_path(path); |
| 5058 | return err; | 5058 | return err; |
| 5059 | } | 5059 | } |
| 5060 | 5060 | ||
| 5061 | static void inode_tree_add(struct inode *inode) | 5061 | static void inode_tree_add(struct inode *inode) |
| 5062 | { | 5062 | { |
| 5063 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5063 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5064 | struct btrfs_inode *entry; | 5064 | struct btrfs_inode *entry; |
| 5065 | struct rb_node **p; | 5065 | struct rb_node **p; |
| 5066 | struct rb_node *parent; | 5066 | struct rb_node *parent; |
| 5067 | struct rb_node *new = &BTRFS_I(inode)->rb_node; | 5067 | struct rb_node *new = &BTRFS_I(inode)->rb_node; |
| 5068 | u64 ino = btrfs_ino(inode); | 5068 | u64 ino = btrfs_ino(inode); |
| 5069 | 5069 | ||
| 5070 | if (inode_unhashed(inode)) | 5070 | if (inode_unhashed(inode)) |
| 5071 | return; | 5071 | return; |
| 5072 | parent = NULL; | 5072 | parent = NULL; |
| 5073 | spin_lock(&root->inode_lock); | 5073 | spin_lock(&root->inode_lock); |
| 5074 | p = &root->inode_tree.rb_node; | 5074 | p = &root->inode_tree.rb_node; |
| 5075 | while (*p) { | 5075 | while (*p) { |
| 5076 | parent = *p; | 5076 | parent = *p; |
| 5077 | entry = rb_entry(parent, struct btrfs_inode, rb_node); | 5077 | entry = rb_entry(parent, struct btrfs_inode, rb_node); |
| 5078 | 5078 | ||
| 5079 | if (ino < btrfs_ino(&entry->vfs_inode)) | 5079 | if (ino < btrfs_ino(&entry->vfs_inode)) |
| 5080 | p = &parent->rb_left; | 5080 | p = &parent->rb_left; |
| 5081 | else if (ino > btrfs_ino(&entry->vfs_inode)) | 5081 | else if (ino > btrfs_ino(&entry->vfs_inode)) |
| 5082 | p = &parent->rb_right; | 5082 | p = &parent->rb_right; |
| 5083 | else { | 5083 | else { |
| 5084 | WARN_ON(!(entry->vfs_inode.i_state & | 5084 | WARN_ON(!(entry->vfs_inode.i_state & |
| 5085 | (I_WILL_FREE | I_FREEING))); | 5085 | (I_WILL_FREE | I_FREEING))); |
| 5086 | rb_replace_node(parent, new, &root->inode_tree); | 5086 | rb_replace_node(parent, new, &root->inode_tree); |
| 5087 | RB_CLEAR_NODE(parent); | 5087 | RB_CLEAR_NODE(parent); |
| 5088 | spin_unlock(&root->inode_lock); | 5088 | spin_unlock(&root->inode_lock); |
| 5089 | return; | 5089 | return; |
| 5090 | } | 5090 | } |
| 5091 | } | 5091 | } |
| 5092 | rb_link_node(new, parent, p); | 5092 | rb_link_node(new, parent, p); |
| 5093 | rb_insert_color(new, &root->inode_tree); | 5093 | rb_insert_color(new, &root->inode_tree); |
| 5094 | spin_unlock(&root->inode_lock); | 5094 | spin_unlock(&root->inode_lock); |
| 5095 | } | 5095 | } |
| 5096 | 5096 | ||
| 5097 | static void inode_tree_del(struct inode *inode) | 5097 | static void inode_tree_del(struct inode *inode) |
| 5098 | { | 5098 | { |
| 5099 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5099 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5100 | int empty = 0; | 5100 | int empty = 0; |
| 5101 | 5101 | ||
| 5102 | spin_lock(&root->inode_lock); | 5102 | spin_lock(&root->inode_lock); |
| 5103 | if (!RB_EMPTY_NODE(&BTRFS_I(inode)->rb_node)) { | 5103 | if (!RB_EMPTY_NODE(&BTRFS_I(inode)->rb_node)) { |
| 5104 | rb_erase(&BTRFS_I(inode)->rb_node, &root->inode_tree); | 5104 | rb_erase(&BTRFS_I(inode)->rb_node, &root->inode_tree); |
| 5105 | RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node); | 5105 | RB_CLEAR_NODE(&BTRFS_I(inode)->rb_node); |
| 5106 | empty = RB_EMPTY_ROOT(&root->inode_tree); | 5106 | empty = RB_EMPTY_ROOT(&root->inode_tree); |
| 5107 | } | 5107 | } |
| 5108 | spin_unlock(&root->inode_lock); | 5108 | spin_unlock(&root->inode_lock); |
| 5109 | 5109 | ||
| 5110 | if (empty && btrfs_root_refs(&root->root_item) == 0) { | 5110 | if (empty && btrfs_root_refs(&root->root_item) == 0) { |
| 5111 | synchronize_srcu(&root->fs_info->subvol_srcu); | 5111 | synchronize_srcu(&root->fs_info->subvol_srcu); |
| 5112 | spin_lock(&root->inode_lock); | 5112 | spin_lock(&root->inode_lock); |
| 5113 | empty = RB_EMPTY_ROOT(&root->inode_tree); | 5113 | empty = RB_EMPTY_ROOT(&root->inode_tree); |
| 5114 | spin_unlock(&root->inode_lock); | 5114 | spin_unlock(&root->inode_lock); |
| 5115 | if (empty) | 5115 | if (empty) |
| 5116 | btrfs_add_dead_root(root); | 5116 | btrfs_add_dead_root(root); |
| 5117 | } | 5117 | } |
| 5118 | } | 5118 | } |
| 5119 | 5119 | ||
| 5120 | void btrfs_invalidate_inodes(struct btrfs_root *root) | 5120 | void btrfs_invalidate_inodes(struct btrfs_root *root) |
| 5121 | { | 5121 | { |
| 5122 | struct rb_node *node; | 5122 | struct rb_node *node; |
| 5123 | struct rb_node *prev; | 5123 | struct rb_node *prev; |
| 5124 | struct btrfs_inode *entry; | 5124 | struct btrfs_inode *entry; |
| 5125 | struct inode *inode; | 5125 | struct inode *inode; |
| 5126 | u64 objectid = 0; | 5126 | u64 objectid = 0; |
| 5127 | 5127 | ||
| 5128 | if (!test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) | 5128 | if (!test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) |
| 5129 | WARN_ON(btrfs_root_refs(&root->root_item) != 0); | 5129 | WARN_ON(btrfs_root_refs(&root->root_item) != 0); |
| 5130 | 5130 | ||
| 5131 | spin_lock(&root->inode_lock); | 5131 | spin_lock(&root->inode_lock); |
| 5132 | again: | 5132 | again: |
| 5133 | node = root->inode_tree.rb_node; | 5133 | node = root->inode_tree.rb_node; |
| 5134 | prev = NULL; | 5134 | prev = NULL; |
| 5135 | while (node) { | 5135 | while (node) { |
| 5136 | prev = node; | 5136 | prev = node; |
| 5137 | entry = rb_entry(node, struct btrfs_inode, rb_node); | 5137 | entry = rb_entry(node, struct btrfs_inode, rb_node); |
| 5138 | 5138 | ||
| 5139 | if (objectid < btrfs_ino(&entry->vfs_inode)) | 5139 | if (objectid < btrfs_ino(&entry->vfs_inode)) |
| 5140 | node = node->rb_left; | 5140 | node = node->rb_left; |
| 5141 | else if (objectid > btrfs_ino(&entry->vfs_inode)) | 5141 | else if (objectid > btrfs_ino(&entry->vfs_inode)) |
| 5142 | node = node->rb_right; | 5142 | node = node->rb_right; |
| 5143 | else | 5143 | else |
| 5144 | break; | 5144 | break; |
| 5145 | } | 5145 | } |
| 5146 | if (!node) { | 5146 | if (!node) { |
| 5147 | while (prev) { | 5147 | while (prev) { |
| 5148 | entry = rb_entry(prev, struct btrfs_inode, rb_node); | 5148 | entry = rb_entry(prev, struct btrfs_inode, rb_node); |
| 5149 | if (objectid <= btrfs_ino(&entry->vfs_inode)) { | 5149 | if (objectid <= btrfs_ino(&entry->vfs_inode)) { |
| 5150 | node = prev; | 5150 | node = prev; |
| 5151 | break; | 5151 | break; |
| 5152 | } | 5152 | } |
| 5153 | prev = rb_next(prev); | 5153 | prev = rb_next(prev); |
| 5154 | } | 5154 | } |
| 5155 | } | 5155 | } |
| 5156 | while (node) { | 5156 | while (node) { |
| 5157 | entry = rb_entry(node, struct btrfs_inode, rb_node); | 5157 | entry = rb_entry(node, struct btrfs_inode, rb_node); |
| 5158 | objectid = btrfs_ino(&entry->vfs_inode) + 1; | 5158 | objectid = btrfs_ino(&entry->vfs_inode) + 1; |
| 5159 | inode = igrab(&entry->vfs_inode); | 5159 | inode = igrab(&entry->vfs_inode); |
| 5160 | if (inode) { | 5160 | if (inode) { |
| 5161 | spin_unlock(&root->inode_lock); | 5161 | spin_unlock(&root->inode_lock); |
| 5162 | if (atomic_read(&inode->i_count) > 1) | 5162 | if (atomic_read(&inode->i_count) > 1) |
| 5163 | d_prune_aliases(inode); | 5163 | d_prune_aliases(inode); |
| 5164 | /* | 5164 | /* |
| 5165 | * btrfs_drop_inode will have it removed from | 5165 | * btrfs_drop_inode will have it removed from |
| 5166 | * the inode cache when its usage count | 5166 | * the inode cache when its usage count |
| 5167 | * hits zero. | 5167 | * hits zero. |
| 5168 | */ | 5168 | */ |
| 5169 | iput(inode); | 5169 | iput(inode); |
| 5170 | cond_resched(); | 5170 | cond_resched(); |
| 5171 | spin_lock(&root->inode_lock); | 5171 | spin_lock(&root->inode_lock); |
| 5172 | goto again; | 5172 | goto again; |
| 5173 | } | 5173 | } |
| 5174 | 5174 | ||
| 5175 | if (cond_resched_lock(&root->inode_lock)) | 5175 | if (cond_resched_lock(&root->inode_lock)) |
| 5176 | goto again; | 5176 | goto again; |
| 5177 | 5177 | ||
| 5178 | node = rb_next(node); | 5178 | node = rb_next(node); |
| 5179 | } | 5179 | } |
| 5180 | spin_unlock(&root->inode_lock); | 5180 | spin_unlock(&root->inode_lock); |
| 5181 | } | 5181 | } |
| 5182 | 5182 | ||
| 5183 | static int btrfs_init_locked_inode(struct inode *inode, void *p) | 5183 | static int btrfs_init_locked_inode(struct inode *inode, void *p) |
| 5184 | { | 5184 | { |
| 5185 | struct btrfs_iget_args *args = p; | 5185 | struct btrfs_iget_args *args = p; |
| 5186 | inode->i_ino = args->location->objectid; | 5186 | inode->i_ino = args->location->objectid; |
| 5187 | memcpy(&BTRFS_I(inode)->location, args->location, | 5187 | memcpy(&BTRFS_I(inode)->location, args->location, |
| 5188 | sizeof(*args->location)); | 5188 | sizeof(*args->location)); |
| 5189 | BTRFS_I(inode)->root = args->root; | 5189 | BTRFS_I(inode)->root = args->root; |
| 5190 | return 0; | 5190 | return 0; |
| 5191 | } | 5191 | } |
| 5192 | 5192 | ||
| 5193 | static int btrfs_find_actor(struct inode *inode, void *opaque) | 5193 | static int btrfs_find_actor(struct inode *inode, void *opaque) |
| 5194 | { | 5194 | { |
| 5195 | struct btrfs_iget_args *args = opaque; | 5195 | struct btrfs_iget_args *args = opaque; |
| 5196 | return args->location->objectid == BTRFS_I(inode)->location.objectid && | 5196 | return args->location->objectid == BTRFS_I(inode)->location.objectid && |
| 5197 | args->root == BTRFS_I(inode)->root; | 5197 | args->root == BTRFS_I(inode)->root; |
| 5198 | } | 5198 | } |
| 5199 | 5199 | ||
| 5200 | static struct inode *btrfs_iget_locked(struct super_block *s, | 5200 | static struct inode *btrfs_iget_locked(struct super_block *s, |
| 5201 | struct btrfs_key *location, | 5201 | struct btrfs_key *location, |
| 5202 | struct btrfs_root *root) | 5202 | struct btrfs_root *root) |
| 5203 | { | 5203 | { |
| 5204 | struct inode *inode; | 5204 | struct inode *inode; |
| 5205 | struct btrfs_iget_args args; | 5205 | struct btrfs_iget_args args; |
| 5206 | unsigned long hashval = btrfs_inode_hash(location->objectid, root); | 5206 | unsigned long hashval = btrfs_inode_hash(location->objectid, root); |
| 5207 | 5207 | ||
| 5208 | args.location = location; | 5208 | args.location = location; |
| 5209 | args.root = root; | 5209 | args.root = root; |
| 5210 | 5210 | ||
| 5211 | inode = iget5_locked(s, hashval, btrfs_find_actor, | 5211 | inode = iget5_locked(s, hashval, btrfs_find_actor, |
| 5212 | btrfs_init_locked_inode, | 5212 | btrfs_init_locked_inode, |
| 5213 | (void *)&args); | 5213 | (void *)&args); |
| 5214 | return inode; | 5214 | return inode; |
| 5215 | } | 5215 | } |
| 5216 | 5216 | ||
| 5217 | /* Get an inode object given its location and corresponding root. | 5217 | /* Get an inode object given its location and corresponding root. |
| 5218 | * Returns in *is_new if the inode was read from disk | 5218 | * Returns in *is_new if the inode was read from disk |
| 5219 | */ | 5219 | */ |
| 5220 | struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, | 5220 | struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, |
| 5221 | struct btrfs_root *root, int *new) | 5221 | struct btrfs_root *root, int *new) |
| 5222 | { | 5222 | { |
| 5223 | struct inode *inode; | 5223 | struct inode *inode; |
| 5224 | 5224 | ||
| 5225 | inode = btrfs_iget_locked(s, location, root); | 5225 | inode = btrfs_iget_locked(s, location, root); |
| 5226 | if (!inode) | 5226 | if (!inode) |
| 5227 | return ERR_PTR(-ENOMEM); | 5227 | return ERR_PTR(-ENOMEM); |
| 5228 | 5228 | ||
| 5229 | if (inode->i_state & I_NEW) { | 5229 | if (inode->i_state & I_NEW) { |
| 5230 | btrfs_read_locked_inode(inode); | 5230 | btrfs_read_locked_inode(inode); |
| 5231 | if (!is_bad_inode(inode)) { | 5231 | if (!is_bad_inode(inode)) { |
| 5232 | inode_tree_add(inode); | 5232 | inode_tree_add(inode); |
| 5233 | unlock_new_inode(inode); | 5233 | unlock_new_inode(inode); |
| 5234 | if (new) | 5234 | if (new) |
| 5235 | *new = 1; | 5235 | *new = 1; |
| 5236 | } else { | 5236 | } else { |
| 5237 | unlock_new_inode(inode); | 5237 | unlock_new_inode(inode); |
| 5238 | iput(inode); | 5238 | iput(inode); |
| 5239 | inode = ERR_PTR(-ESTALE); | 5239 | inode = ERR_PTR(-ESTALE); |
| 5240 | } | 5240 | } |
| 5241 | } | 5241 | } |
| 5242 | 5242 | ||
| 5243 | return inode; | 5243 | return inode; |
| 5244 | } | 5244 | } |
| 5245 | 5245 | ||
| 5246 | static struct inode *new_simple_dir(struct super_block *s, | 5246 | static struct inode *new_simple_dir(struct super_block *s, |
| 5247 | struct btrfs_key *key, | 5247 | struct btrfs_key *key, |
| 5248 | struct btrfs_root *root) | 5248 | struct btrfs_root *root) |
| 5249 | { | 5249 | { |
| 5250 | struct inode *inode = new_inode(s); | 5250 | struct inode *inode = new_inode(s); |
| 5251 | 5251 | ||
| 5252 | if (!inode) | 5252 | if (!inode) |
| 5253 | return ERR_PTR(-ENOMEM); | 5253 | return ERR_PTR(-ENOMEM); |
| 5254 | 5254 | ||
| 5255 | BTRFS_I(inode)->root = root; | 5255 | BTRFS_I(inode)->root = root; |
| 5256 | memcpy(&BTRFS_I(inode)->location, key, sizeof(*key)); | 5256 | memcpy(&BTRFS_I(inode)->location, key, sizeof(*key)); |
| 5257 | set_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags); | 5257 | set_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags); |
| 5258 | 5258 | ||
| 5259 | inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID; | 5259 | inode->i_ino = BTRFS_EMPTY_SUBVOL_DIR_OBJECTID; |
| 5260 | inode->i_op = &btrfs_dir_ro_inode_operations; | 5260 | inode->i_op = &btrfs_dir_ro_inode_operations; |
| 5261 | inode->i_fop = &simple_dir_operations; | 5261 | inode->i_fop = &simple_dir_operations; |
| 5262 | inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO; | 5262 | inode->i_mode = S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO; |
| 5263 | inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; | 5263 | inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; |
| 5264 | 5264 | ||
| 5265 | return inode; | 5265 | return inode; |
| 5266 | } | 5266 | } |
| 5267 | 5267 | ||
| 5268 | struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) | 5268 | struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) |
| 5269 | { | 5269 | { |
| 5270 | struct inode *inode; | 5270 | struct inode *inode; |
| 5271 | struct btrfs_root *root = BTRFS_I(dir)->root; | 5271 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 5272 | struct btrfs_root *sub_root = root; | 5272 | struct btrfs_root *sub_root = root; |
| 5273 | struct btrfs_key location; | 5273 | struct btrfs_key location; |
| 5274 | int index; | 5274 | int index; |
| 5275 | int ret = 0; | 5275 | int ret = 0; |
| 5276 | 5276 | ||
| 5277 | if (dentry->d_name.len > BTRFS_NAME_LEN) | 5277 | if (dentry->d_name.len > BTRFS_NAME_LEN) |
| 5278 | return ERR_PTR(-ENAMETOOLONG); | 5278 | return ERR_PTR(-ENAMETOOLONG); |
| 5279 | 5279 | ||
| 5280 | ret = btrfs_inode_by_name(dir, dentry, &location); | 5280 | ret = btrfs_inode_by_name(dir, dentry, &location); |
| 5281 | if (ret < 0) | 5281 | if (ret < 0) |
| 5282 | return ERR_PTR(ret); | 5282 | return ERR_PTR(ret); |
| 5283 | 5283 | ||
| 5284 | if (location.objectid == 0) | 5284 | if (location.objectid == 0) |
| 5285 | return ERR_PTR(-ENOENT); | 5285 | return ERR_PTR(-ENOENT); |
| 5286 | 5286 | ||
| 5287 | if (location.type == BTRFS_INODE_ITEM_KEY) { | 5287 | if (location.type == BTRFS_INODE_ITEM_KEY) { |
| 5288 | inode = btrfs_iget(dir->i_sb, &location, root, NULL); | 5288 | inode = btrfs_iget(dir->i_sb, &location, root, NULL); |
| 5289 | return inode; | 5289 | return inode; |
| 5290 | } | 5290 | } |
| 5291 | 5291 | ||
| 5292 | BUG_ON(location.type != BTRFS_ROOT_ITEM_KEY); | 5292 | BUG_ON(location.type != BTRFS_ROOT_ITEM_KEY); |
| 5293 | 5293 | ||
| 5294 | index = srcu_read_lock(&root->fs_info->subvol_srcu); | 5294 | index = srcu_read_lock(&root->fs_info->subvol_srcu); |
| 5295 | ret = fixup_tree_root_location(root, dir, dentry, | 5295 | ret = fixup_tree_root_location(root, dir, dentry, |
| 5296 | &location, &sub_root); | 5296 | &location, &sub_root); |
| 5297 | if (ret < 0) { | 5297 | if (ret < 0) { |
| 5298 | if (ret != -ENOENT) | 5298 | if (ret != -ENOENT) |
| 5299 | inode = ERR_PTR(ret); | 5299 | inode = ERR_PTR(ret); |
| 5300 | else | 5300 | else |
| 5301 | inode = new_simple_dir(dir->i_sb, &location, sub_root); | 5301 | inode = new_simple_dir(dir->i_sb, &location, sub_root); |
| 5302 | } else { | 5302 | } else { |
| 5303 | inode = btrfs_iget(dir->i_sb, &location, sub_root, NULL); | 5303 | inode = btrfs_iget(dir->i_sb, &location, sub_root, NULL); |
| 5304 | } | 5304 | } |
| 5305 | srcu_read_unlock(&root->fs_info->subvol_srcu, index); | 5305 | srcu_read_unlock(&root->fs_info->subvol_srcu, index); |
| 5306 | 5306 | ||
| 5307 | if (!IS_ERR(inode) && root != sub_root) { | 5307 | if (!IS_ERR(inode) && root != sub_root) { |
| 5308 | down_read(&root->fs_info->cleanup_work_sem); | 5308 | down_read(&root->fs_info->cleanup_work_sem); |
| 5309 | if (!(inode->i_sb->s_flags & MS_RDONLY)) | 5309 | if (!(inode->i_sb->s_flags & MS_RDONLY)) |
| 5310 | ret = btrfs_orphan_cleanup(sub_root); | 5310 | ret = btrfs_orphan_cleanup(sub_root); |
| 5311 | up_read(&root->fs_info->cleanup_work_sem); | 5311 | up_read(&root->fs_info->cleanup_work_sem); |
| 5312 | if (ret) { | 5312 | if (ret) { |
| 5313 | iput(inode); | 5313 | iput(inode); |
| 5314 | inode = ERR_PTR(ret); | 5314 | inode = ERR_PTR(ret); |
| 5315 | } | 5315 | } |
| 5316 | } | 5316 | } |
| 5317 | 5317 | ||
| 5318 | return inode; | 5318 | return inode; |
| 5319 | } | 5319 | } |
| 5320 | 5320 | ||
| 5321 | static int btrfs_dentry_delete(const struct dentry *dentry) | 5321 | static int btrfs_dentry_delete(const struct dentry *dentry) |
| 5322 | { | 5322 | { |
| 5323 | struct btrfs_root *root; | 5323 | struct btrfs_root *root; |
| 5324 | struct inode *inode = dentry->d_inode; | 5324 | struct inode *inode = dentry->d_inode; |
| 5325 | 5325 | ||
| 5326 | if (!inode && !IS_ROOT(dentry)) | 5326 | if (!inode && !IS_ROOT(dentry)) |
| 5327 | inode = dentry->d_parent->d_inode; | 5327 | inode = dentry->d_parent->d_inode; |
| 5328 | 5328 | ||
| 5329 | if (inode) { | 5329 | if (inode) { |
| 5330 | root = BTRFS_I(inode)->root; | 5330 | root = BTRFS_I(inode)->root; |
| 5331 | if (btrfs_root_refs(&root->root_item) == 0) | 5331 | if (btrfs_root_refs(&root->root_item) == 0) |
| 5332 | return 1; | 5332 | return 1; |
| 5333 | 5333 | ||
| 5334 | if (btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) | 5334 | if (btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) |
| 5335 | return 1; | 5335 | return 1; |
| 5336 | } | 5336 | } |
| 5337 | return 0; | 5337 | return 0; |
| 5338 | } | 5338 | } |
| 5339 | 5339 | ||
| 5340 | static void btrfs_dentry_release(struct dentry *dentry) | 5340 | static void btrfs_dentry_release(struct dentry *dentry) |
| 5341 | { | 5341 | { |
| 5342 | kfree(dentry->d_fsdata); | 5342 | kfree(dentry->d_fsdata); |
| 5343 | } | 5343 | } |
| 5344 | 5344 | ||
| 5345 | static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, | 5345 | static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, |
| 5346 | unsigned int flags) | 5346 | unsigned int flags) |
| 5347 | { | 5347 | { |
| 5348 | struct inode *inode; | 5348 | struct inode *inode; |
| 5349 | 5349 | ||
| 5350 | inode = btrfs_lookup_dentry(dir, dentry); | 5350 | inode = btrfs_lookup_dentry(dir, dentry); |
| 5351 | if (IS_ERR(inode)) { | 5351 | if (IS_ERR(inode)) { |
| 5352 | if (PTR_ERR(inode) == -ENOENT) | 5352 | if (PTR_ERR(inode) == -ENOENT) |
| 5353 | inode = NULL; | 5353 | inode = NULL; |
| 5354 | else | 5354 | else |
| 5355 | return ERR_CAST(inode); | 5355 | return ERR_CAST(inode); |
| 5356 | } | 5356 | } |
| 5357 | 5357 | ||
| 5358 | return d_splice_alias(inode, dentry); | 5358 | return d_splice_alias(inode, dentry); |
| 5359 | } | 5359 | } |
| 5360 | 5360 | ||
| 5361 | unsigned char btrfs_filetype_table[] = { | 5361 | unsigned char btrfs_filetype_table[] = { |
| 5362 | DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK | 5362 | DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK |
| 5363 | }; | 5363 | }; |
| 5364 | 5364 | ||
| 5365 | static int btrfs_real_readdir(struct file *file, struct dir_context *ctx) | 5365 | static int btrfs_real_readdir(struct file *file, struct dir_context *ctx) |
| 5366 | { | 5366 | { |
| 5367 | struct inode *inode = file_inode(file); | 5367 | struct inode *inode = file_inode(file); |
| 5368 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5368 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5369 | struct btrfs_item *item; | 5369 | struct btrfs_item *item; |
| 5370 | struct btrfs_dir_item *di; | 5370 | struct btrfs_dir_item *di; |
| 5371 | struct btrfs_key key; | 5371 | struct btrfs_key key; |
| 5372 | struct btrfs_key found_key; | 5372 | struct btrfs_key found_key; |
| 5373 | struct btrfs_path *path; | 5373 | struct btrfs_path *path; |
| 5374 | struct list_head ins_list; | 5374 | struct list_head ins_list; |
| 5375 | struct list_head del_list; | 5375 | struct list_head del_list; |
| 5376 | int ret; | 5376 | int ret; |
| 5377 | struct extent_buffer *leaf; | 5377 | struct extent_buffer *leaf; |
| 5378 | int slot; | 5378 | int slot; |
| 5379 | unsigned char d_type; | 5379 | unsigned char d_type; |
| 5380 | int over = 0; | 5380 | int over = 0; |
| 5381 | u32 di_cur; | 5381 | u32 di_cur; |
| 5382 | u32 di_total; | 5382 | u32 di_total; |
| 5383 | u32 di_len; | 5383 | u32 di_len; |
| 5384 | int key_type = BTRFS_DIR_INDEX_KEY; | 5384 | int key_type = BTRFS_DIR_INDEX_KEY; |
| 5385 | char tmp_name[32]; | 5385 | char tmp_name[32]; |
| 5386 | char *name_ptr; | 5386 | char *name_ptr; |
| 5387 | int name_len; | 5387 | int name_len; |
| 5388 | int is_curr = 0; /* ctx->pos points to the current index? */ | 5388 | int is_curr = 0; /* ctx->pos points to the current index? */ |
| 5389 | 5389 | ||
| 5390 | /* FIXME, use a real flag for deciding about the key type */ | 5390 | /* FIXME, use a real flag for deciding about the key type */ |
| 5391 | if (root->fs_info->tree_root == root) | 5391 | if (root->fs_info->tree_root == root) |
| 5392 | key_type = BTRFS_DIR_ITEM_KEY; | 5392 | key_type = BTRFS_DIR_ITEM_KEY; |
| 5393 | 5393 | ||
| 5394 | if (!dir_emit_dots(file, ctx)) | 5394 | if (!dir_emit_dots(file, ctx)) |
| 5395 | return 0; | 5395 | return 0; |
| 5396 | 5396 | ||
| 5397 | path = btrfs_alloc_path(); | 5397 | path = btrfs_alloc_path(); |
| 5398 | if (!path) | 5398 | if (!path) |
| 5399 | return -ENOMEM; | 5399 | return -ENOMEM; |
| 5400 | 5400 | ||
| 5401 | path->reada = 1; | 5401 | path->reada = 1; |
| 5402 | 5402 | ||
| 5403 | if (key_type == BTRFS_DIR_INDEX_KEY) { | 5403 | if (key_type == BTRFS_DIR_INDEX_KEY) { |
| 5404 | INIT_LIST_HEAD(&ins_list); | 5404 | INIT_LIST_HEAD(&ins_list); |
| 5405 | INIT_LIST_HEAD(&del_list); | 5405 | INIT_LIST_HEAD(&del_list); |
| 5406 | btrfs_get_delayed_items(inode, &ins_list, &del_list); | 5406 | btrfs_get_delayed_items(inode, &ins_list, &del_list); |
| 5407 | } | 5407 | } |
| 5408 | 5408 | ||
| 5409 | key.type = key_type; | 5409 | key.type = key_type; |
| 5410 | key.offset = ctx->pos; | 5410 | key.offset = ctx->pos; |
| 5411 | key.objectid = btrfs_ino(inode); | 5411 | key.objectid = btrfs_ino(inode); |
| 5412 | 5412 | ||
| 5413 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 5413 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 5414 | if (ret < 0) | 5414 | if (ret < 0) |
| 5415 | goto err; | 5415 | goto err; |
| 5416 | 5416 | ||
| 5417 | while (1) { | 5417 | while (1) { |
| 5418 | leaf = path->nodes[0]; | 5418 | leaf = path->nodes[0]; |
| 5419 | slot = path->slots[0]; | 5419 | slot = path->slots[0]; |
| 5420 | if (slot >= btrfs_header_nritems(leaf)) { | 5420 | if (slot >= btrfs_header_nritems(leaf)) { |
| 5421 | ret = btrfs_next_leaf(root, path); | 5421 | ret = btrfs_next_leaf(root, path); |
| 5422 | if (ret < 0) | 5422 | if (ret < 0) |
| 5423 | goto err; | 5423 | goto err; |
| 5424 | else if (ret > 0) | 5424 | else if (ret > 0) |
| 5425 | break; | 5425 | break; |
| 5426 | continue; | 5426 | continue; |
| 5427 | } | 5427 | } |
| 5428 | 5428 | ||
| 5429 | item = btrfs_item_nr(slot); | 5429 | item = btrfs_item_nr(slot); |
| 5430 | btrfs_item_key_to_cpu(leaf, &found_key, slot); | 5430 | btrfs_item_key_to_cpu(leaf, &found_key, slot); |
| 5431 | 5431 | ||
| 5432 | if (found_key.objectid != key.objectid) | 5432 | if (found_key.objectid != key.objectid) |
| 5433 | break; | 5433 | break; |
| 5434 | if (found_key.type != key_type) | 5434 | if (found_key.type != key_type) |
| 5435 | break; | 5435 | break; |
| 5436 | if (found_key.offset < ctx->pos) | 5436 | if (found_key.offset < ctx->pos) |
| 5437 | goto next; | 5437 | goto next; |
| 5438 | if (key_type == BTRFS_DIR_INDEX_KEY && | 5438 | if (key_type == BTRFS_DIR_INDEX_KEY && |
| 5439 | btrfs_should_delete_dir_index(&del_list, | 5439 | btrfs_should_delete_dir_index(&del_list, |
| 5440 | found_key.offset)) | 5440 | found_key.offset)) |
| 5441 | goto next; | 5441 | goto next; |
| 5442 | 5442 | ||
| 5443 | ctx->pos = found_key.offset; | 5443 | ctx->pos = found_key.offset; |
| 5444 | is_curr = 1; | 5444 | is_curr = 1; |
| 5445 | 5445 | ||
| 5446 | di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item); | 5446 | di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item); |
| 5447 | di_cur = 0; | 5447 | di_cur = 0; |
| 5448 | di_total = btrfs_item_size(leaf, item); | 5448 | di_total = btrfs_item_size(leaf, item); |
| 5449 | 5449 | ||
| 5450 | while (di_cur < di_total) { | 5450 | while (di_cur < di_total) { |
| 5451 | struct btrfs_key location; | 5451 | struct btrfs_key location; |
| 5452 | 5452 | ||
| 5453 | if (verify_dir_item(root, leaf, di)) | 5453 | if (verify_dir_item(root, leaf, di)) |
| 5454 | break; | 5454 | break; |
| 5455 | 5455 | ||
| 5456 | name_len = btrfs_dir_name_len(leaf, di); | 5456 | name_len = btrfs_dir_name_len(leaf, di); |
| 5457 | if (name_len <= sizeof(tmp_name)) { | 5457 | if (name_len <= sizeof(tmp_name)) { |
| 5458 | name_ptr = tmp_name; | 5458 | name_ptr = tmp_name; |
| 5459 | } else { | 5459 | } else { |
| 5460 | name_ptr = kmalloc(name_len, GFP_NOFS); | 5460 | name_ptr = kmalloc(name_len, GFP_NOFS); |
| 5461 | if (!name_ptr) { | 5461 | if (!name_ptr) { |
| 5462 | ret = -ENOMEM; | 5462 | ret = -ENOMEM; |
| 5463 | goto err; | 5463 | goto err; |
| 5464 | } | 5464 | } |
| 5465 | } | 5465 | } |
| 5466 | read_extent_buffer(leaf, name_ptr, | 5466 | read_extent_buffer(leaf, name_ptr, |
| 5467 | (unsigned long)(di + 1), name_len); | 5467 | (unsigned long)(di + 1), name_len); |
| 5468 | 5468 | ||
| 5469 | d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)]; | 5469 | d_type = btrfs_filetype_table[btrfs_dir_type(leaf, di)]; |
| 5470 | btrfs_dir_item_key_to_cpu(leaf, di, &location); | 5470 | btrfs_dir_item_key_to_cpu(leaf, di, &location); |
| 5471 | 5471 | ||
| 5472 | 5472 | ||
| 5473 | /* is this a reference to our own snapshot? If so | 5473 | /* is this a reference to our own snapshot? If so |
| 5474 | * skip it. | 5474 | * skip it. |
| 5475 | * | 5475 | * |
| 5476 | * In contrast to old kernels, we insert the snapshot's | 5476 | * In contrast to old kernels, we insert the snapshot's |
| 5477 | * dir item and dir index after it has been created, so | 5477 | * dir item and dir index after it has been created, so |
| 5478 | * we won't find a reference to our own snapshot. We | 5478 | * we won't find a reference to our own snapshot. We |
| 5479 | * still keep the following code for backward | 5479 | * still keep the following code for backward |
| 5480 | * compatibility. | 5480 | * compatibility. |
| 5481 | */ | 5481 | */ |
| 5482 | if (location.type == BTRFS_ROOT_ITEM_KEY && | 5482 | if (location.type == BTRFS_ROOT_ITEM_KEY && |
| 5483 | location.objectid == root->root_key.objectid) { | 5483 | location.objectid == root->root_key.objectid) { |
| 5484 | over = 0; | 5484 | over = 0; |
| 5485 | goto skip; | 5485 | goto skip; |
| 5486 | } | 5486 | } |
| 5487 | over = !dir_emit(ctx, name_ptr, name_len, | 5487 | over = !dir_emit(ctx, name_ptr, name_len, |
| 5488 | location.objectid, d_type); | 5488 | location.objectid, d_type); |
| 5489 | 5489 | ||
| 5490 | skip: | 5490 | skip: |
| 5491 | if (name_ptr != tmp_name) | 5491 | if (name_ptr != tmp_name) |
| 5492 | kfree(name_ptr); | 5492 | kfree(name_ptr); |
| 5493 | 5493 | ||
| 5494 | if (over) | 5494 | if (over) |
| 5495 | goto nopos; | 5495 | goto nopos; |
| 5496 | di_len = btrfs_dir_name_len(leaf, di) + | 5496 | di_len = btrfs_dir_name_len(leaf, di) + |
| 5497 | btrfs_dir_data_len(leaf, di) + sizeof(*di); | 5497 | btrfs_dir_data_len(leaf, di) + sizeof(*di); |
| 5498 | di_cur += di_len; | 5498 | di_cur += di_len; |
| 5499 | di = (struct btrfs_dir_item *)((char *)di + di_len); | 5499 | di = (struct btrfs_dir_item *)((char *)di + di_len); |
| 5500 | } | 5500 | } |
| 5501 | next: | 5501 | next: |
| 5502 | path->slots[0]++; | 5502 | path->slots[0]++; |
| 5503 | } | 5503 | } |
| 5504 | 5504 | ||
| 5505 | if (key_type == BTRFS_DIR_INDEX_KEY) { | 5505 | if (key_type == BTRFS_DIR_INDEX_KEY) { |
| 5506 | if (is_curr) | 5506 | if (is_curr) |
| 5507 | ctx->pos++; | 5507 | ctx->pos++; |
| 5508 | ret = btrfs_readdir_delayed_dir_index(ctx, &ins_list); | 5508 | ret = btrfs_readdir_delayed_dir_index(ctx, &ins_list); |
| 5509 | if (ret) | 5509 | if (ret) |
| 5510 | goto nopos; | 5510 | goto nopos; |
| 5511 | } | 5511 | } |
| 5512 | 5512 | ||
| 5513 | /* Reached end of directory/root. Bump pos past the last item. */ | 5513 | /* Reached end of directory/root. Bump pos past the last item. */ |
| 5514 | ctx->pos++; | 5514 | ctx->pos++; |
| 5515 | 5515 | ||
| 5516 | /* | 5516 | /* |
| 5517 | * Stop new entries from being returned after we return the last | 5517 | * Stop new entries from being returned after we return the last |
| 5518 | * entry. | 5518 | * entry. |
| 5519 | * | 5519 | * |
| 5520 | * New directory entries are assigned a strictly increasing | 5520 | * New directory entries are assigned a strictly increasing |
| 5521 | * offset. This means that new entries created during readdir | 5521 | * offset. This means that new entries created during readdir |
| 5522 | * are *guaranteed* to be seen in the future by that readdir. | 5522 | * are *guaranteed* to be seen in the future by that readdir. |
| 5523 | * This has broken buggy programs which operate on names as | 5523 | * This has broken buggy programs which operate on names as |
| 5524 | * they're returned by readdir. Until we re-use freed offsets | 5524 | * they're returned by readdir. Until we re-use freed offsets |
| 5525 | * we have this hack to stop new entries from being returned | 5525 | * we have this hack to stop new entries from being returned |
| 5526 | * under the assumption that they'll never reach this huge | 5526 | * under the assumption that they'll never reach this huge |
| 5527 | * offset. | 5527 | * offset. |
| 5528 | * | 5528 | * |
| 5529 | * This is being careful not to overflow 32bit loff_t unless the | 5529 | * This is being careful not to overflow 32bit loff_t unless the |
| 5530 | * last entry requires it because doing so has broken 32bit apps | 5530 | * last entry requires it because doing so has broken 32bit apps |
| 5531 | * in the past. | 5531 | * in the past. |
| 5532 | */ | 5532 | */ |
| 5533 | if (key_type == BTRFS_DIR_INDEX_KEY) { | 5533 | if (key_type == BTRFS_DIR_INDEX_KEY) { |
| 5534 | if (ctx->pos >= INT_MAX) | 5534 | if (ctx->pos >= INT_MAX) |
| 5535 | ctx->pos = LLONG_MAX; | 5535 | ctx->pos = LLONG_MAX; |
| 5536 | else | 5536 | else |
| 5537 | ctx->pos = INT_MAX; | 5537 | ctx->pos = INT_MAX; |
| 5538 | } | 5538 | } |
| 5539 | nopos: | 5539 | nopos: |
| 5540 | ret = 0; | 5540 | ret = 0; |
| 5541 | err: | 5541 | err: |
| 5542 | if (key_type == BTRFS_DIR_INDEX_KEY) | 5542 | if (key_type == BTRFS_DIR_INDEX_KEY) |
| 5543 | btrfs_put_delayed_items(&ins_list, &del_list); | 5543 | btrfs_put_delayed_items(&ins_list, &del_list); |
| 5544 | btrfs_free_path(path); | 5544 | btrfs_free_path(path); |
| 5545 | return ret; | 5545 | return ret; |
| 5546 | } | 5546 | } |
| 5547 | 5547 | ||
| 5548 | int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) | 5548 | int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) |
| 5549 | { | 5549 | { |
| 5550 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5550 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5551 | struct btrfs_trans_handle *trans; | 5551 | struct btrfs_trans_handle *trans; |
| 5552 | int ret = 0; | 5552 | int ret = 0; |
| 5553 | bool nolock = false; | 5553 | bool nolock = false; |
| 5554 | 5554 | ||
| 5555 | if (test_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags)) | 5555 | if (test_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags)) |
| 5556 | return 0; | 5556 | return 0; |
| 5557 | 5557 | ||
| 5558 | if (btrfs_fs_closing(root->fs_info) && btrfs_is_free_space_inode(inode)) | 5558 | if (btrfs_fs_closing(root->fs_info) && btrfs_is_free_space_inode(inode)) |
| 5559 | nolock = true; | 5559 | nolock = true; |
| 5560 | 5560 | ||
| 5561 | if (wbc->sync_mode == WB_SYNC_ALL) { | 5561 | if (wbc->sync_mode == WB_SYNC_ALL) { |
| 5562 | if (nolock) | 5562 | if (nolock) |
| 5563 | trans = btrfs_join_transaction_nolock(root); | 5563 | trans = btrfs_join_transaction_nolock(root); |
| 5564 | else | 5564 | else |
| 5565 | trans = btrfs_join_transaction(root); | 5565 | trans = btrfs_join_transaction(root); |
| 5566 | if (IS_ERR(trans)) | 5566 | if (IS_ERR(trans)) |
| 5567 | return PTR_ERR(trans); | 5567 | return PTR_ERR(trans); |
| 5568 | ret = btrfs_commit_transaction(trans, root); | 5568 | ret = btrfs_commit_transaction(trans, root); |
| 5569 | } | 5569 | } |
| 5570 | return ret; | 5570 | return ret; |
| 5571 | } | 5571 | } |
| 5572 | 5572 | ||
| 5573 | /* | 5573 | /* |
| 5574 | * This is somewhat expensive, updating the tree every time the | 5574 | * This is somewhat expensive, updating the tree every time the |
| 5575 | * inode changes. But, it is most likely to find the inode in cache. | 5575 | * inode changes. But, it is most likely to find the inode in cache. |
| 5576 | * FIXME, needs more benchmarking...there are no reasons other than performance | 5576 | * FIXME, needs more benchmarking...there are no reasons other than performance |
| 5577 | * to keep or drop this code. | 5577 | * to keep or drop this code. |
| 5578 | */ | 5578 | */ |
| 5579 | static int btrfs_dirty_inode(struct inode *inode) | 5579 | static int btrfs_dirty_inode(struct inode *inode) |
| 5580 | { | 5580 | { |
| 5581 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5581 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5582 | struct btrfs_trans_handle *trans; | 5582 | struct btrfs_trans_handle *trans; |
| 5583 | int ret; | 5583 | int ret; |
| 5584 | 5584 | ||
| 5585 | if (test_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags)) | 5585 | if (test_bit(BTRFS_INODE_DUMMY, &BTRFS_I(inode)->runtime_flags)) |
| 5586 | return 0; | 5586 | return 0; |
| 5587 | 5587 | ||
| 5588 | trans = btrfs_join_transaction(root); | 5588 | trans = btrfs_join_transaction(root); |
| 5589 | if (IS_ERR(trans)) | 5589 | if (IS_ERR(trans)) |
| 5590 | return PTR_ERR(trans); | 5590 | return PTR_ERR(trans); |
| 5591 | 5591 | ||
| 5592 | ret = btrfs_update_inode(trans, root, inode); | 5592 | ret = btrfs_update_inode(trans, root, inode); |
| 5593 | if (ret && ret == -ENOSPC) { | 5593 | if (ret && ret == -ENOSPC) { |
| 5594 | /* whoops, lets try again with the full transaction */ | 5594 | /* whoops, lets try again with the full transaction */ |
| 5595 | btrfs_end_transaction(trans, root); | 5595 | btrfs_end_transaction(trans, root); |
| 5596 | trans = btrfs_start_transaction(root, 1); | 5596 | trans = btrfs_start_transaction(root, 1); |
| 5597 | if (IS_ERR(trans)) | 5597 | if (IS_ERR(trans)) |
| 5598 | return PTR_ERR(trans); | 5598 | return PTR_ERR(trans); |
| 5599 | 5599 | ||
| 5600 | ret = btrfs_update_inode(trans, root, inode); | 5600 | ret = btrfs_update_inode(trans, root, inode); |
| 5601 | } | 5601 | } |
| 5602 | btrfs_end_transaction(trans, root); | 5602 | btrfs_end_transaction(trans, root); |
| 5603 | if (BTRFS_I(inode)->delayed_node) | 5603 | if (BTRFS_I(inode)->delayed_node) |
| 5604 | btrfs_balance_delayed_items(root); | 5604 | btrfs_balance_delayed_items(root); |
| 5605 | 5605 | ||
| 5606 | return ret; | 5606 | return ret; |
| 5607 | } | 5607 | } |
| 5608 | 5608 | ||
| 5609 | /* | 5609 | /* |
| 5610 | * This is a copy of file_update_time. We need this so we can return error on | 5610 | * This is a copy of file_update_time. We need this so we can return error on |
| 5611 | * ENOSPC for updating the inode in the case of file write and mmap writes. | 5611 | * ENOSPC for updating the inode in the case of file write and mmap writes. |
| 5612 | */ | 5612 | */ |
| 5613 | static int btrfs_update_time(struct inode *inode, struct timespec *now, | 5613 | static int btrfs_update_time(struct inode *inode, struct timespec *now, |
| 5614 | int flags) | 5614 | int flags) |
| 5615 | { | 5615 | { |
| 5616 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5616 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5617 | 5617 | ||
| 5618 | if (btrfs_root_readonly(root)) | 5618 | if (btrfs_root_readonly(root)) |
| 5619 | return -EROFS; | 5619 | return -EROFS; |
| 5620 | 5620 | ||
| 5621 | if (flags & S_VERSION) | 5621 | if (flags & S_VERSION) |
| 5622 | inode_inc_iversion(inode); | 5622 | inode_inc_iversion(inode); |
| 5623 | if (flags & S_CTIME) | 5623 | if (flags & S_CTIME) |
| 5624 | inode->i_ctime = *now; | 5624 | inode->i_ctime = *now; |
| 5625 | if (flags & S_MTIME) | 5625 | if (flags & S_MTIME) |
| 5626 | inode->i_mtime = *now; | 5626 | inode->i_mtime = *now; |
| 5627 | if (flags & S_ATIME) | 5627 | if (flags & S_ATIME) |
| 5628 | inode->i_atime = *now; | 5628 | inode->i_atime = *now; |
| 5629 | return btrfs_dirty_inode(inode); | 5629 | return btrfs_dirty_inode(inode); |
| 5630 | } | 5630 | } |
| 5631 | 5631 | ||
| 5632 | /* | 5632 | /* |
| 5633 | * find the highest existing sequence number in a directory | 5633 | * find the highest existing sequence number in a directory |
| 5634 | * and then set the in-memory index_cnt variable to reflect | 5634 | * and then set the in-memory index_cnt variable to reflect |
| 5635 | * free sequence numbers | 5635 | * free sequence numbers |
| 5636 | */ | 5636 | */ |
| 5637 | static int btrfs_set_inode_index_count(struct inode *inode) | 5637 | static int btrfs_set_inode_index_count(struct inode *inode) |
| 5638 | { | 5638 | { |
| 5639 | struct btrfs_root *root = BTRFS_I(inode)->root; | 5639 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 5640 | struct btrfs_key key, found_key; | 5640 | struct btrfs_key key, found_key; |
| 5641 | struct btrfs_path *path; | 5641 | struct btrfs_path *path; |
| 5642 | struct extent_buffer *leaf; | 5642 | struct extent_buffer *leaf; |
| 5643 | int ret; | 5643 | int ret; |
| 5644 | 5644 | ||
| 5645 | key.objectid = btrfs_ino(inode); | 5645 | key.objectid = btrfs_ino(inode); |
| 5646 | key.type = BTRFS_DIR_INDEX_KEY; | 5646 | key.type = BTRFS_DIR_INDEX_KEY; |
| 5647 | key.offset = (u64)-1; | 5647 | key.offset = (u64)-1; |
| 5648 | 5648 | ||
| 5649 | path = btrfs_alloc_path(); | 5649 | path = btrfs_alloc_path(); |
| 5650 | if (!path) | 5650 | if (!path) |
| 5651 | return -ENOMEM; | 5651 | return -ENOMEM; |
| 5652 | 5652 | ||
| 5653 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 5653 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 5654 | if (ret < 0) | 5654 | if (ret < 0) |
| 5655 | goto out; | 5655 | goto out; |
| 5656 | /* FIXME: we should be able to handle this */ | 5656 | /* FIXME: we should be able to handle this */ |
| 5657 | if (ret == 0) | 5657 | if (ret == 0) |
| 5658 | goto out; | 5658 | goto out; |
| 5659 | ret = 0; | 5659 | ret = 0; |
| 5660 | 5660 | ||
| 5661 | /* | 5661 | /* |
| 5662 | * MAGIC NUMBER EXPLANATION: | 5662 | * MAGIC NUMBER EXPLANATION: |
| 5663 | * since we search a directory based on f_pos we have to start at 2 | 5663 | * since we search a directory based on f_pos we have to start at 2 |
| 5664 | * since '.' and '..' have f_pos of 0 and 1 respectively, so everybody | 5664 | * since '.' and '..' have f_pos of 0 and 1 respectively, so everybody |
| 5665 | * else has to start at 2 | 5665 | * else has to start at 2 |
| 5666 | */ | 5666 | */ |
| 5667 | if (path->slots[0] == 0) { | 5667 | if (path->slots[0] == 0) { |
| 5668 | BTRFS_I(inode)->index_cnt = 2; | 5668 | BTRFS_I(inode)->index_cnt = 2; |
| 5669 | goto out; | 5669 | goto out; |
| 5670 | } | 5670 | } |
| 5671 | 5671 | ||
| 5672 | path->slots[0]--; | 5672 | path->slots[0]--; |
| 5673 | 5673 | ||
| 5674 | leaf = path->nodes[0]; | 5674 | leaf = path->nodes[0]; |
| 5675 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 5675 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 5676 | 5676 | ||
| 5677 | if (found_key.objectid != btrfs_ino(inode) || | 5677 | if (found_key.objectid != btrfs_ino(inode) || |
| 5678 | found_key.type != BTRFS_DIR_INDEX_KEY) { | 5678 | found_key.type != BTRFS_DIR_INDEX_KEY) { |
| 5679 | BTRFS_I(inode)->index_cnt = 2; | 5679 | BTRFS_I(inode)->index_cnt = 2; |
| 5680 | goto out; | 5680 | goto out; |
| 5681 | } | 5681 | } |
| 5682 | 5682 | ||
| 5683 | BTRFS_I(inode)->index_cnt = found_key.offset + 1; | 5683 | BTRFS_I(inode)->index_cnt = found_key.offset + 1; |
| 5684 | out: | 5684 | out: |
| 5685 | btrfs_free_path(path); | 5685 | btrfs_free_path(path); |
| 5686 | return ret; | 5686 | return ret; |
| 5687 | } | 5687 | } |
| 5688 | 5688 | ||
| 5689 | /* | 5689 | /* |
| 5690 | * helper to find a free sequence number in a given directory. This current | 5690 | * helper to find a free sequence number in a given directory. This current |
| 5691 | * code is very simple, later versions will do smarter things in the btree | 5691 | * code is very simple, later versions will do smarter things in the btree |
| 5692 | */ | 5692 | */ |
| 5693 | int btrfs_set_inode_index(struct inode *dir, u64 *index) | 5693 | int btrfs_set_inode_index(struct inode *dir, u64 *index) |
| 5694 | { | 5694 | { |
| 5695 | int ret = 0; | 5695 | int ret = 0; |
| 5696 | 5696 | ||
| 5697 | if (BTRFS_I(dir)->index_cnt == (u64)-1) { | 5697 | if (BTRFS_I(dir)->index_cnt == (u64)-1) { |
| 5698 | ret = btrfs_inode_delayed_dir_index_count(dir); | 5698 | ret = btrfs_inode_delayed_dir_index_count(dir); |
| 5699 | if (ret) { | 5699 | if (ret) { |
| 5700 | ret = btrfs_set_inode_index_count(dir); | 5700 | ret = btrfs_set_inode_index_count(dir); |
| 5701 | if (ret) | 5701 | if (ret) |
| 5702 | return ret; | 5702 | return ret; |
| 5703 | } | 5703 | } |
| 5704 | } | 5704 | } |
| 5705 | 5705 | ||
| 5706 | *index = BTRFS_I(dir)->index_cnt; | 5706 | *index = BTRFS_I(dir)->index_cnt; |
| 5707 | BTRFS_I(dir)->index_cnt++; | 5707 | BTRFS_I(dir)->index_cnt++; |
| 5708 | 5708 | ||
| 5709 | return ret; | 5709 | return ret; |
| 5710 | } | 5710 | } |
| 5711 | 5711 | ||
| 5712 | static int btrfs_insert_inode_locked(struct inode *inode) | 5712 | static int btrfs_insert_inode_locked(struct inode *inode) |
| 5713 | { | 5713 | { |
| 5714 | struct btrfs_iget_args args; | 5714 | struct btrfs_iget_args args; |
| 5715 | args.location = &BTRFS_I(inode)->location; | 5715 | args.location = &BTRFS_I(inode)->location; |
| 5716 | args.root = BTRFS_I(inode)->root; | 5716 | args.root = BTRFS_I(inode)->root; |
| 5717 | 5717 | ||
| 5718 | return insert_inode_locked4(inode, | 5718 | return insert_inode_locked4(inode, |
| 5719 | btrfs_inode_hash(inode->i_ino, BTRFS_I(inode)->root), | 5719 | btrfs_inode_hash(inode->i_ino, BTRFS_I(inode)->root), |
| 5720 | btrfs_find_actor, &args); | 5720 | btrfs_find_actor, &args); |
| 5721 | } | 5721 | } |
| 5722 | 5722 | ||
| 5723 | static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, | 5723 | static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, |
| 5724 | struct btrfs_root *root, | 5724 | struct btrfs_root *root, |
| 5725 | struct inode *dir, | 5725 | struct inode *dir, |
| 5726 | const char *name, int name_len, | 5726 | const char *name, int name_len, |
| 5727 | u64 ref_objectid, u64 objectid, | 5727 | u64 ref_objectid, u64 objectid, |
| 5728 | umode_t mode, u64 *index) | 5728 | umode_t mode, u64 *index) |
| 5729 | { | 5729 | { |
| 5730 | struct inode *inode; | 5730 | struct inode *inode; |
| 5731 | struct btrfs_inode_item *inode_item; | 5731 | struct btrfs_inode_item *inode_item; |
| 5732 | struct btrfs_key *location; | 5732 | struct btrfs_key *location; |
| 5733 | struct btrfs_path *path; | 5733 | struct btrfs_path *path; |
| 5734 | struct btrfs_inode_ref *ref; | 5734 | struct btrfs_inode_ref *ref; |
| 5735 | struct btrfs_key key[2]; | 5735 | struct btrfs_key key[2]; |
| 5736 | u32 sizes[2]; | 5736 | u32 sizes[2]; |
| 5737 | int nitems = name ? 2 : 1; | 5737 | int nitems = name ? 2 : 1; |
| 5738 | unsigned long ptr; | 5738 | unsigned long ptr; |
| 5739 | int ret; | 5739 | int ret; |
| 5740 | 5740 | ||
| 5741 | path = btrfs_alloc_path(); | 5741 | path = btrfs_alloc_path(); |
| 5742 | if (!path) | 5742 | if (!path) |
| 5743 | return ERR_PTR(-ENOMEM); | 5743 | return ERR_PTR(-ENOMEM); |
| 5744 | 5744 | ||
| 5745 | inode = new_inode(root->fs_info->sb); | 5745 | inode = new_inode(root->fs_info->sb); |
| 5746 | if (!inode) { | 5746 | if (!inode) { |
| 5747 | btrfs_free_path(path); | 5747 | btrfs_free_path(path); |
| 5748 | return ERR_PTR(-ENOMEM); | 5748 | return ERR_PTR(-ENOMEM); |
| 5749 | } | 5749 | } |
| 5750 | 5750 | ||
| 5751 | /* | 5751 | /* |
| 5752 | * O_TMPFILE, set link count to 0, so that after this point, | 5752 | * O_TMPFILE, set link count to 0, so that after this point, |
| 5753 | * we fill in an inode item with the correct link count. | 5753 | * we fill in an inode item with the correct link count. |
| 5754 | */ | 5754 | */ |
| 5755 | if (!name) | 5755 | if (!name) |
| 5756 | set_nlink(inode, 0); | 5756 | set_nlink(inode, 0); |
| 5757 | 5757 | ||
| 5758 | /* | 5758 | /* |
| 5759 | * we have to initialize this early, so we can reclaim the inode | 5759 | * we have to initialize this early, so we can reclaim the inode |
| 5760 | * number if we fail afterwards in this function. | 5760 | * number if we fail afterwards in this function. |
| 5761 | */ | 5761 | */ |
| 5762 | inode->i_ino = objectid; | 5762 | inode->i_ino = objectid; |
| 5763 | 5763 | ||
| 5764 | if (dir && name) { | 5764 | if (dir && name) { |
| 5765 | trace_btrfs_inode_request(dir); | 5765 | trace_btrfs_inode_request(dir); |
| 5766 | 5766 | ||
| 5767 | ret = btrfs_set_inode_index(dir, index); | 5767 | ret = btrfs_set_inode_index(dir, index); |
| 5768 | if (ret) { | 5768 | if (ret) { |
| 5769 | btrfs_free_path(path); | 5769 | btrfs_free_path(path); |
| 5770 | iput(inode); | 5770 | iput(inode); |
| 5771 | return ERR_PTR(ret); | 5771 | return ERR_PTR(ret); |
| 5772 | } | 5772 | } |
| 5773 | } else if (dir) { | 5773 | } else if (dir) { |
| 5774 | *index = 0; | 5774 | *index = 0; |
| 5775 | } | 5775 | } |
| 5776 | /* | 5776 | /* |
| 5777 | * index_cnt is ignored for everything but a dir, | 5777 | * index_cnt is ignored for everything but a dir, |
| 5778 | * btrfs_get_inode_index_count has an explanation for the magic | 5778 | * btrfs_get_inode_index_count has an explanation for the magic |
| 5779 | * number | 5779 | * number |
| 5780 | */ | 5780 | */ |
| 5781 | BTRFS_I(inode)->index_cnt = 2; | 5781 | BTRFS_I(inode)->index_cnt = 2; |
| 5782 | BTRFS_I(inode)->dir_index = *index; | 5782 | BTRFS_I(inode)->dir_index = *index; |
| 5783 | BTRFS_I(inode)->root = root; | 5783 | BTRFS_I(inode)->root = root; |
| 5784 | BTRFS_I(inode)->generation = trans->transid; | 5784 | BTRFS_I(inode)->generation = trans->transid; |
| 5785 | inode->i_generation = BTRFS_I(inode)->generation; | 5785 | inode->i_generation = BTRFS_I(inode)->generation; |
| 5786 | 5786 | ||
| 5787 | /* | 5787 | /* |
| 5788 | * We could have gotten an inode number from somebody who was fsynced | 5788 | * We could have gotten an inode number from somebody who was fsynced |
| 5789 | * and then removed in this same transaction, so let's just set full | 5789 | * and then removed in this same transaction, so let's just set full |
| 5790 | * sync since it will be a full sync anyway and this will blow away the | 5790 | * sync since it will be a full sync anyway and this will blow away the |
| 5791 | * old info in the log. | 5791 | * old info in the log. |
| 5792 | */ | 5792 | */ |
| 5793 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags); | 5793 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags); |
| 5794 | 5794 | ||
| 5795 | key[0].objectid = objectid; | 5795 | key[0].objectid = objectid; |
| 5796 | key[0].type = BTRFS_INODE_ITEM_KEY; | 5796 | key[0].type = BTRFS_INODE_ITEM_KEY; |
| 5797 | key[0].offset = 0; | 5797 | key[0].offset = 0; |
| 5798 | 5798 | ||
| 5799 | sizes[0] = sizeof(struct btrfs_inode_item); | 5799 | sizes[0] = sizeof(struct btrfs_inode_item); |
| 5800 | 5800 | ||
| 5801 | if (name) { | 5801 | if (name) { |
| 5802 | /* | 5802 | /* |
| 5803 | * Start new inodes with an inode_ref. This is slightly more | 5803 | * Start new inodes with an inode_ref. This is slightly more |
| 5804 | * efficient for small numbers of hard links since they will | 5804 | * efficient for small numbers of hard links since they will |
| 5805 | * be packed into one item. Extended refs will kick in if we | 5805 | * be packed into one item. Extended refs will kick in if we |
| 5806 | * add more hard links than can fit in the ref item. | 5806 | * add more hard links than can fit in the ref item. |
| 5807 | */ | 5807 | */ |
| 5808 | key[1].objectid = objectid; | 5808 | key[1].objectid = objectid; |
| 5809 | key[1].type = BTRFS_INODE_REF_KEY; | 5809 | key[1].type = BTRFS_INODE_REF_KEY; |
| 5810 | key[1].offset = ref_objectid; | 5810 | key[1].offset = ref_objectid; |
| 5811 | 5811 | ||
| 5812 | sizes[1] = name_len + sizeof(*ref); | 5812 | sizes[1] = name_len + sizeof(*ref); |
| 5813 | } | 5813 | } |
| 5814 | 5814 | ||
| 5815 | location = &BTRFS_I(inode)->location; | 5815 | location = &BTRFS_I(inode)->location; |
| 5816 | location->objectid = objectid; | 5816 | location->objectid = objectid; |
| 5817 | location->offset = 0; | 5817 | location->offset = 0; |
| 5818 | location->type = BTRFS_INODE_ITEM_KEY; | 5818 | location->type = BTRFS_INODE_ITEM_KEY; |
| 5819 | 5819 | ||
| 5820 | ret = btrfs_insert_inode_locked(inode); | 5820 | ret = btrfs_insert_inode_locked(inode); |
| 5821 | if (ret < 0) | 5821 | if (ret < 0) |
| 5822 | goto fail; | 5822 | goto fail; |
| 5823 | 5823 | ||
| 5824 | path->leave_spinning = 1; | 5824 | path->leave_spinning = 1; |
| 5825 | ret = btrfs_insert_empty_items(trans, root, path, key, sizes, nitems); | 5825 | ret = btrfs_insert_empty_items(trans, root, path, key, sizes, nitems); |
| 5826 | if (ret != 0) | 5826 | if (ret != 0) |
| 5827 | goto fail_unlock; | 5827 | goto fail_unlock; |
| 5828 | 5828 | ||
| 5829 | inode_init_owner(inode, dir, mode); | 5829 | inode_init_owner(inode, dir, mode); |
| 5830 | inode_set_bytes(inode, 0); | 5830 | inode_set_bytes(inode, 0); |
| 5831 | inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; | 5831 | inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; |
| 5832 | inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0], | 5832 | inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0], |
| 5833 | struct btrfs_inode_item); | 5833 | struct btrfs_inode_item); |
| 5834 | memset_extent_buffer(path->nodes[0], 0, (unsigned long)inode_item, | 5834 | memset_extent_buffer(path->nodes[0], 0, (unsigned long)inode_item, |
| 5835 | sizeof(*inode_item)); | 5835 | sizeof(*inode_item)); |
| 5836 | fill_inode_item(trans, path->nodes[0], inode_item, inode); | 5836 | fill_inode_item(trans, path->nodes[0], inode_item, inode); |
| 5837 | 5837 | ||
| 5838 | if (name) { | 5838 | if (name) { |
| 5839 | ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1, | 5839 | ref = btrfs_item_ptr(path->nodes[0], path->slots[0] + 1, |
| 5840 | struct btrfs_inode_ref); | 5840 | struct btrfs_inode_ref); |
| 5841 | btrfs_set_inode_ref_name_len(path->nodes[0], ref, name_len); | 5841 | btrfs_set_inode_ref_name_len(path->nodes[0], ref, name_len); |
| 5842 | btrfs_set_inode_ref_index(path->nodes[0], ref, *index); | 5842 | btrfs_set_inode_ref_index(path->nodes[0], ref, *index); |
| 5843 | ptr = (unsigned long)(ref + 1); | 5843 | ptr = (unsigned long)(ref + 1); |
| 5844 | write_extent_buffer(path->nodes[0], name, ptr, name_len); | 5844 | write_extent_buffer(path->nodes[0], name, ptr, name_len); |
| 5845 | } | 5845 | } |
| 5846 | 5846 | ||
| 5847 | btrfs_mark_buffer_dirty(path->nodes[0]); | 5847 | btrfs_mark_buffer_dirty(path->nodes[0]); |
| 5848 | btrfs_free_path(path); | 5848 | btrfs_free_path(path); |
| 5849 | 5849 | ||
| 5850 | btrfs_inherit_iflags(inode, dir); | 5850 | btrfs_inherit_iflags(inode, dir); |
| 5851 | 5851 | ||
| 5852 | if (S_ISREG(mode)) { | 5852 | if (S_ISREG(mode)) { |
| 5853 | if (btrfs_test_opt(root, NODATASUM)) | 5853 | if (btrfs_test_opt(root, NODATASUM)) |
| 5854 | BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; | 5854 | BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; |
| 5855 | if (btrfs_test_opt(root, NODATACOW)) | 5855 | if (btrfs_test_opt(root, NODATACOW)) |
| 5856 | BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW | | 5856 | BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW | |
| 5857 | BTRFS_INODE_NODATASUM; | 5857 | BTRFS_INODE_NODATASUM; |
| 5858 | } | 5858 | } |
| 5859 | 5859 | ||
| 5860 | inode_tree_add(inode); | 5860 | inode_tree_add(inode); |
| 5861 | 5861 | ||
| 5862 | trace_btrfs_inode_new(inode); | 5862 | trace_btrfs_inode_new(inode); |
| 5863 | btrfs_set_inode_last_trans(trans, inode); | 5863 | btrfs_set_inode_last_trans(trans, inode); |
| 5864 | 5864 | ||
| 5865 | btrfs_update_root_times(trans, root); | 5865 | btrfs_update_root_times(trans, root); |
| 5866 | 5866 | ||
| 5867 | ret = btrfs_inode_inherit_props(trans, inode, dir); | 5867 | ret = btrfs_inode_inherit_props(trans, inode, dir); |
| 5868 | if (ret) | 5868 | if (ret) |
| 5869 | btrfs_err(root->fs_info, | 5869 | btrfs_err(root->fs_info, |
| 5870 | "error inheriting props for ino %llu (root %llu): %d", | 5870 | "error inheriting props for ino %llu (root %llu): %d", |
| 5871 | btrfs_ino(inode), root->root_key.objectid, ret); | 5871 | btrfs_ino(inode), root->root_key.objectid, ret); |
| 5872 | 5872 | ||
| 5873 | return inode; | 5873 | return inode; |
| 5874 | 5874 | ||
| 5875 | fail_unlock: | 5875 | fail_unlock: |
| 5876 | unlock_new_inode(inode); | 5876 | unlock_new_inode(inode); |
| 5877 | fail: | 5877 | fail: |
| 5878 | if (dir && name) | 5878 | if (dir && name) |
| 5879 | BTRFS_I(dir)->index_cnt--; | 5879 | BTRFS_I(dir)->index_cnt--; |
| 5880 | btrfs_free_path(path); | 5880 | btrfs_free_path(path); |
| 5881 | iput(inode); | 5881 | iput(inode); |
| 5882 | return ERR_PTR(ret); | 5882 | return ERR_PTR(ret); |
| 5883 | } | 5883 | } |
| 5884 | 5884 | ||
| 5885 | static inline u8 btrfs_inode_type(struct inode *inode) | 5885 | static inline u8 btrfs_inode_type(struct inode *inode) |
| 5886 | { | 5886 | { |
| 5887 | return btrfs_type_by_mode[(inode->i_mode & S_IFMT) >> S_SHIFT]; | 5887 | return btrfs_type_by_mode[(inode->i_mode & S_IFMT) >> S_SHIFT]; |
| 5888 | } | 5888 | } |
| 5889 | 5889 | ||
| 5890 | /* | 5890 | /* |
| 5891 | * utility function to add 'inode' into 'parent_inode' with | 5891 | * utility function to add 'inode' into 'parent_inode' with |
| 5892 | * a give name and a given sequence number. | 5892 | * a give name and a given sequence number. |
| 5893 | * if 'add_backref' is true, also insert a backref from the | 5893 | * if 'add_backref' is true, also insert a backref from the |
| 5894 | * inode to the parent directory. | 5894 | * inode to the parent directory. |
| 5895 | */ | 5895 | */ |
| 5896 | int btrfs_add_link(struct btrfs_trans_handle *trans, | 5896 | int btrfs_add_link(struct btrfs_trans_handle *trans, |
| 5897 | struct inode *parent_inode, struct inode *inode, | 5897 | struct inode *parent_inode, struct inode *inode, |
| 5898 | const char *name, int name_len, int add_backref, u64 index) | 5898 | const char *name, int name_len, int add_backref, u64 index) |
| 5899 | { | 5899 | { |
| 5900 | int ret = 0; | 5900 | int ret = 0; |
| 5901 | struct btrfs_key key; | 5901 | struct btrfs_key key; |
| 5902 | struct btrfs_root *root = BTRFS_I(parent_inode)->root; | 5902 | struct btrfs_root *root = BTRFS_I(parent_inode)->root; |
| 5903 | u64 ino = btrfs_ino(inode); | 5903 | u64 ino = btrfs_ino(inode); |
| 5904 | u64 parent_ino = btrfs_ino(parent_inode); | 5904 | u64 parent_ino = btrfs_ino(parent_inode); |
| 5905 | 5905 | ||
| 5906 | if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { | 5906 | if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { |
| 5907 | memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key)); | 5907 | memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key)); |
| 5908 | } else { | 5908 | } else { |
| 5909 | key.objectid = ino; | 5909 | key.objectid = ino; |
| 5910 | key.type = BTRFS_INODE_ITEM_KEY; | 5910 | key.type = BTRFS_INODE_ITEM_KEY; |
| 5911 | key.offset = 0; | 5911 | key.offset = 0; |
| 5912 | } | 5912 | } |
| 5913 | 5913 | ||
| 5914 | if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { | 5914 | if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { |
| 5915 | ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, | 5915 | ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, |
| 5916 | key.objectid, root->root_key.objectid, | 5916 | key.objectid, root->root_key.objectid, |
| 5917 | parent_ino, index, name, name_len); | 5917 | parent_ino, index, name, name_len); |
| 5918 | } else if (add_backref) { | 5918 | } else if (add_backref) { |
| 5919 | ret = btrfs_insert_inode_ref(trans, root, name, name_len, ino, | 5919 | ret = btrfs_insert_inode_ref(trans, root, name, name_len, ino, |
| 5920 | parent_ino, index); | 5920 | parent_ino, index); |
| 5921 | } | 5921 | } |
| 5922 | 5922 | ||
| 5923 | /* Nothing to clean up yet */ | 5923 | /* Nothing to clean up yet */ |
| 5924 | if (ret) | 5924 | if (ret) |
| 5925 | return ret; | 5925 | return ret; |
| 5926 | 5926 | ||
| 5927 | ret = btrfs_insert_dir_item(trans, root, name, name_len, | 5927 | ret = btrfs_insert_dir_item(trans, root, name, name_len, |
| 5928 | parent_inode, &key, | 5928 | parent_inode, &key, |
| 5929 | btrfs_inode_type(inode), index); | 5929 | btrfs_inode_type(inode), index); |
| 5930 | if (ret == -EEXIST || ret == -EOVERFLOW) | 5930 | if (ret == -EEXIST || ret == -EOVERFLOW) |
| 5931 | goto fail_dir_item; | 5931 | goto fail_dir_item; |
| 5932 | else if (ret) { | 5932 | else if (ret) { |
| 5933 | btrfs_abort_transaction(trans, root, ret); | 5933 | btrfs_abort_transaction(trans, root, ret); |
| 5934 | return ret; | 5934 | return ret; |
| 5935 | } | 5935 | } |
| 5936 | 5936 | ||
| 5937 | btrfs_i_size_write(parent_inode, parent_inode->i_size + | 5937 | btrfs_i_size_write(parent_inode, parent_inode->i_size + |
| 5938 | name_len * 2); | 5938 | name_len * 2); |
| 5939 | inode_inc_iversion(parent_inode); | 5939 | inode_inc_iversion(parent_inode); |
| 5940 | parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME; | 5940 | parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME; |
| 5941 | ret = btrfs_update_inode(trans, root, parent_inode); | 5941 | ret = btrfs_update_inode(trans, root, parent_inode); |
| 5942 | if (ret) | 5942 | if (ret) |
| 5943 | btrfs_abort_transaction(trans, root, ret); | 5943 | btrfs_abort_transaction(trans, root, ret); |
| 5944 | return ret; | 5944 | return ret; |
| 5945 | 5945 | ||
| 5946 | fail_dir_item: | 5946 | fail_dir_item: |
| 5947 | if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { | 5947 | if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { |
| 5948 | u64 local_index; | 5948 | u64 local_index; |
| 5949 | int err; | 5949 | int err; |
| 5950 | err = btrfs_del_root_ref(trans, root->fs_info->tree_root, | 5950 | err = btrfs_del_root_ref(trans, root->fs_info->tree_root, |
| 5951 | key.objectid, root->root_key.objectid, | 5951 | key.objectid, root->root_key.objectid, |
| 5952 | parent_ino, &local_index, name, name_len); | 5952 | parent_ino, &local_index, name, name_len); |
| 5953 | 5953 | ||
| 5954 | } else if (add_backref) { | 5954 | } else if (add_backref) { |
| 5955 | u64 local_index; | 5955 | u64 local_index; |
| 5956 | int err; | 5956 | int err; |
| 5957 | 5957 | ||
| 5958 | err = btrfs_del_inode_ref(trans, root, name, name_len, | 5958 | err = btrfs_del_inode_ref(trans, root, name, name_len, |
| 5959 | ino, parent_ino, &local_index); | 5959 | ino, parent_ino, &local_index); |
| 5960 | } | 5960 | } |
| 5961 | return ret; | 5961 | return ret; |
| 5962 | } | 5962 | } |
| 5963 | 5963 | ||
| 5964 | static int btrfs_add_nondir(struct btrfs_trans_handle *trans, | 5964 | static int btrfs_add_nondir(struct btrfs_trans_handle *trans, |
| 5965 | struct inode *dir, struct dentry *dentry, | 5965 | struct inode *dir, struct dentry *dentry, |
| 5966 | struct inode *inode, int backref, u64 index) | 5966 | struct inode *inode, int backref, u64 index) |
| 5967 | { | 5967 | { |
| 5968 | int err = btrfs_add_link(trans, dir, inode, | 5968 | int err = btrfs_add_link(trans, dir, inode, |
| 5969 | dentry->d_name.name, dentry->d_name.len, | 5969 | dentry->d_name.name, dentry->d_name.len, |
| 5970 | backref, index); | 5970 | backref, index); |
| 5971 | if (err > 0) | 5971 | if (err > 0) |
| 5972 | err = -EEXIST; | 5972 | err = -EEXIST; |
| 5973 | return err; | 5973 | return err; |
| 5974 | } | 5974 | } |
| 5975 | 5975 | ||
| 5976 | static int btrfs_mknod(struct inode *dir, struct dentry *dentry, | 5976 | static int btrfs_mknod(struct inode *dir, struct dentry *dentry, |
| 5977 | umode_t mode, dev_t rdev) | 5977 | umode_t mode, dev_t rdev) |
| 5978 | { | 5978 | { |
| 5979 | struct btrfs_trans_handle *trans; | 5979 | struct btrfs_trans_handle *trans; |
| 5980 | struct btrfs_root *root = BTRFS_I(dir)->root; | 5980 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 5981 | struct inode *inode = NULL; | 5981 | struct inode *inode = NULL; |
| 5982 | int err; | 5982 | int err; |
| 5983 | int drop_inode = 0; | 5983 | int drop_inode = 0; |
| 5984 | u64 objectid; | 5984 | u64 objectid; |
| 5985 | u64 index = 0; | 5985 | u64 index = 0; |
| 5986 | 5986 | ||
| 5987 | if (!new_valid_dev(rdev)) | 5987 | if (!new_valid_dev(rdev)) |
| 5988 | return -EINVAL; | 5988 | return -EINVAL; |
| 5989 | 5989 | ||
| 5990 | /* | 5990 | /* |
| 5991 | * 2 for inode item and ref | 5991 | * 2 for inode item and ref |
| 5992 | * 2 for dir items | 5992 | * 2 for dir items |
| 5993 | * 1 for xattr if selinux is on | 5993 | * 1 for xattr if selinux is on |
| 5994 | */ | 5994 | */ |
| 5995 | trans = btrfs_start_transaction(root, 5); | 5995 | trans = btrfs_start_transaction(root, 5); |
| 5996 | if (IS_ERR(trans)) | 5996 | if (IS_ERR(trans)) |
| 5997 | return PTR_ERR(trans); | 5997 | return PTR_ERR(trans); |
| 5998 | 5998 | ||
| 5999 | err = btrfs_find_free_ino(root, &objectid); | 5999 | err = btrfs_find_free_ino(root, &objectid); |
| 6000 | if (err) | 6000 | if (err) |
| 6001 | goto out_unlock; | 6001 | goto out_unlock; |
| 6002 | 6002 | ||
| 6003 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, | 6003 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, |
| 6004 | dentry->d_name.len, btrfs_ino(dir), objectid, | 6004 | dentry->d_name.len, btrfs_ino(dir), objectid, |
| 6005 | mode, &index); | 6005 | mode, &index); |
| 6006 | if (IS_ERR(inode)) { | 6006 | if (IS_ERR(inode)) { |
| 6007 | err = PTR_ERR(inode); | 6007 | err = PTR_ERR(inode); |
| 6008 | goto out_unlock; | 6008 | goto out_unlock; |
| 6009 | } | 6009 | } |
| 6010 | 6010 | ||
| 6011 | /* | 6011 | /* |
| 6012 | * If the active LSM wants to access the inode during | 6012 | * If the active LSM wants to access the inode during |
| 6013 | * d_instantiate it needs these. Smack checks to see | 6013 | * d_instantiate it needs these. Smack checks to see |
| 6014 | * if the filesystem supports xattrs by looking at the | 6014 | * if the filesystem supports xattrs by looking at the |
| 6015 | * ops vector. | 6015 | * ops vector. |
| 6016 | */ | 6016 | */ |
| 6017 | inode->i_op = &btrfs_special_inode_operations; | 6017 | inode->i_op = &btrfs_special_inode_operations; |
| 6018 | init_special_inode(inode, inode->i_mode, rdev); | 6018 | init_special_inode(inode, inode->i_mode, rdev); |
| 6019 | 6019 | ||
| 6020 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); | 6020 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); |
| 6021 | if (err) | 6021 | if (err) |
| 6022 | goto out_unlock_inode; | 6022 | goto out_unlock_inode; |
| 6023 | 6023 | ||
| 6024 | err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); | 6024 | err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); |
| 6025 | if (err) { | 6025 | if (err) { |
| 6026 | goto out_unlock_inode; | 6026 | goto out_unlock_inode; |
| 6027 | } else { | 6027 | } else { |
| 6028 | btrfs_update_inode(trans, root, inode); | 6028 | btrfs_update_inode(trans, root, inode); |
| 6029 | unlock_new_inode(inode); | 6029 | unlock_new_inode(inode); |
| 6030 | d_instantiate(dentry, inode); | 6030 | d_instantiate(dentry, inode); |
| 6031 | } | 6031 | } |
| 6032 | 6032 | ||
| 6033 | out_unlock: | 6033 | out_unlock: |
| 6034 | btrfs_end_transaction(trans, root); | 6034 | btrfs_end_transaction(trans, root); |
| 6035 | btrfs_balance_delayed_items(root); | 6035 | btrfs_balance_delayed_items(root); |
| 6036 | btrfs_btree_balance_dirty(root); | 6036 | btrfs_btree_balance_dirty(root); |
| 6037 | if (drop_inode) { | 6037 | if (drop_inode) { |
| 6038 | inode_dec_link_count(inode); | 6038 | inode_dec_link_count(inode); |
| 6039 | iput(inode); | 6039 | iput(inode); |
| 6040 | } | 6040 | } |
| 6041 | return err; | 6041 | return err; |
| 6042 | 6042 | ||
| 6043 | out_unlock_inode: | 6043 | out_unlock_inode: |
| 6044 | drop_inode = 1; | 6044 | drop_inode = 1; |
| 6045 | unlock_new_inode(inode); | 6045 | unlock_new_inode(inode); |
| 6046 | goto out_unlock; | 6046 | goto out_unlock; |
| 6047 | 6047 | ||
| 6048 | } | 6048 | } |
| 6049 | 6049 | ||
| 6050 | static int btrfs_create(struct inode *dir, struct dentry *dentry, | 6050 | static int btrfs_create(struct inode *dir, struct dentry *dentry, |
| 6051 | umode_t mode, bool excl) | 6051 | umode_t mode, bool excl) |
| 6052 | { | 6052 | { |
| 6053 | struct btrfs_trans_handle *trans; | 6053 | struct btrfs_trans_handle *trans; |
| 6054 | struct btrfs_root *root = BTRFS_I(dir)->root; | 6054 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 6055 | struct inode *inode = NULL; | 6055 | struct inode *inode = NULL; |
| 6056 | int drop_inode_on_err = 0; | 6056 | int drop_inode_on_err = 0; |
| 6057 | int err; | 6057 | int err; |
| 6058 | u64 objectid; | 6058 | u64 objectid; |
| 6059 | u64 index = 0; | 6059 | u64 index = 0; |
| 6060 | 6060 | ||
| 6061 | /* | 6061 | /* |
| 6062 | * 2 for inode item and ref | 6062 | * 2 for inode item and ref |
| 6063 | * 2 for dir items | 6063 | * 2 for dir items |
| 6064 | * 1 for xattr if selinux is on | 6064 | * 1 for xattr if selinux is on |
| 6065 | */ | 6065 | */ |
| 6066 | trans = btrfs_start_transaction(root, 5); | 6066 | trans = btrfs_start_transaction(root, 5); |
| 6067 | if (IS_ERR(trans)) | 6067 | if (IS_ERR(trans)) |
| 6068 | return PTR_ERR(trans); | 6068 | return PTR_ERR(trans); |
| 6069 | 6069 | ||
| 6070 | err = btrfs_find_free_ino(root, &objectid); | 6070 | err = btrfs_find_free_ino(root, &objectid); |
| 6071 | if (err) | 6071 | if (err) |
| 6072 | goto out_unlock; | 6072 | goto out_unlock; |
| 6073 | 6073 | ||
| 6074 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, | 6074 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, |
| 6075 | dentry->d_name.len, btrfs_ino(dir), objectid, | 6075 | dentry->d_name.len, btrfs_ino(dir), objectid, |
| 6076 | mode, &index); | 6076 | mode, &index); |
| 6077 | if (IS_ERR(inode)) { | 6077 | if (IS_ERR(inode)) { |
| 6078 | err = PTR_ERR(inode); | 6078 | err = PTR_ERR(inode); |
| 6079 | goto out_unlock; | 6079 | goto out_unlock; |
| 6080 | } | 6080 | } |
| 6081 | drop_inode_on_err = 1; | 6081 | drop_inode_on_err = 1; |
| 6082 | /* | 6082 | /* |
| 6083 | * If the active LSM wants to access the inode during | 6083 | * If the active LSM wants to access the inode during |
| 6084 | * d_instantiate it needs these. Smack checks to see | 6084 | * d_instantiate it needs these. Smack checks to see |
| 6085 | * if the filesystem supports xattrs by looking at the | 6085 | * if the filesystem supports xattrs by looking at the |
| 6086 | * ops vector. | 6086 | * ops vector. |
| 6087 | */ | 6087 | */ |
| 6088 | inode->i_fop = &btrfs_file_operations; | 6088 | inode->i_fop = &btrfs_file_operations; |
| 6089 | inode->i_op = &btrfs_file_inode_operations; | 6089 | inode->i_op = &btrfs_file_inode_operations; |
| 6090 | inode->i_mapping->a_ops = &btrfs_aops; | 6090 | inode->i_mapping->a_ops = &btrfs_aops; |
| 6091 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; | 6091 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; |
| 6092 | 6092 | ||
| 6093 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); | 6093 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); |
| 6094 | if (err) | 6094 | if (err) |
| 6095 | goto out_unlock_inode; | 6095 | goto out_unlock_inode; |
| 6096 | 6096 | ||
| 6097 | err = btrfs_update_inode(trans, root, inode); | 6097 | err = btrfs_update_inode(trans, root, inode); |
| 6098 | if (err) | 6098 | if (err) |
| 6099 | goto out_unlock_inode; | 6099 | goto out_unlock_inode; |
| 6100 | 6100 | ||
| 6101 | err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); | 6101 | err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); |
| 6102 | if (err) | 6102 | if (err) |
| 6103 | goto out_unlock_inode; | 6103 | goto out_unlock_inode; |
| 6104 | 6104 | ||
| 6105 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; | 6105 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; |
| 6106 | unlock_new_inode(inode); | 6106 | unlock_new_inode(inode); |
| 6107 | d_instantiate(dentry, inode); | 6107 | d_instantiate(dentry, inode); |
| 6108 | 6108 | ||
| 6109 | out_unlock: | 6109 | out_unlock: |
| 6110 | btrfs_end_transaction(trans, root); | 6110 | btrfs_end_transaction(trans, root); |
| 6111 | if (err && drop_inode_on_err) { | 6111 | if (err && drop_inode_on_err) { |
| 6112 | inode_dec_link_count(inode); | 6112 | inode_dec_link_count(inode); |
| 6113 | iput(inode); | 6113 | iput(inode); |
| 6114 | } | 6114 | } |
| 6115 | btrfs_balance_delayed_items(root); | 6115 | btrfs_balance_delayed_items(root); |
| 6116 | btrfs_btree_balance_dirty(root); | 6116 | btrfs_btree_balance_dirty(root); |
| 6117 | return err; | 6117 | return err; |
| 6118 | 6118 | ||
| 6119 | out_unlock_inode: | 6119 | out_unlock_inode: |
| 6120 | unlock_new_inode(inode); | 6120 | unlock_new_inode(inode); |
| 6121 | goto out_unlock; | 6121 | goto out_unlock; |
| 6122 | 6122 | ||
| 6123 | } | 6123 | } |
| 6124 | 6124 | ||
| 6125 | static int btrfs_link(struct dentry *old_dentry, struct inode *dir, | 6125 | static int btrfs_link(struct dentry *old_dentry, struct inode *dir, |
| 6126 | struct dentry *dentry) | 6126 | struct dentry *dentry) |
| 6127 | { | 6127 | { |
| 6128 | struct btrfs_trans_handle *trans; | 6128 | struct btrfs_trans_handle *trans; |
| 6129 | struct btrfs_root *root = BTRFS_I(dir)->root; | 6129 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 6130 | struct inode *inode = old_dentry->d_inode; | 6130 | struct inode *inode = old_dentry->d_inode; |
| 6131 | u64 index; | 6131 | u64 index; |
| 6132 | int err; | 6132 | int err; |
| 6133 | int drop_inode = 0; | 6133 | int drop_inode = 0; |
| 6134 | 6134 | ||
| 6135 | /* do not allow sys_link's with other subvols of the same device */ | 6135 | /* do not allow sys_link's with other subvols of the same device */ |
| 6136 | if (root->objectid != BTRFS_I(inode)->root->objectid) | 6136 | if (root->objectid != BTRFS_I(inode)->root->objectid) |
| 6137 | return -EXDEV; | 6137 | return -EXDEV; |
| 6138 | 6138 | ||
| 6139 | if (inode->i_nlink >= BTRFS_LINK_MAX) | 6139 | if (inode->i_nlink >= BTRFS_LINK_MAX) |
| 6140 | return -EMLINK; | 6140 | return -EMLINK; |
| 6141 | 6141 | ||
| 6142 | err = btrfs_set_inode_index(dir, &index); | 6142 | err = btrfs_set_inode_index(dir, &index); |
| 6143 | if (err) | 6143 | if (err) |
| 6144 | goto fail; | 6144 | goto fail; |
| 6145 | 6145 | ||
| 6146 | /* | 6146 | /* |
| 6147 | * 2 items for inode and inode ref | 6147 | * 2 items for inode and inode ref |
| 6148 | * 2 items for dir items | 6148 | * 2 items for dir items |
| 6149 | * 1 item for parent inode | 6149 | * 1 item for parent inode |
| 6150 | */ | 6150 | */ |
| 6151 | trans = btrfs_start_transaction(root, 5); | 6151 | trans = btrfs_start_transaction(root, 5); |
| 6152 | if (IS_ERR(trans)) { | 6152 | if (IS_ERR(trans)) { |
| 6153 | err = PTR_ERR(trans); | 6153 | err = PTR_ERR(trans); |
| 6154 | goto fail; | 6154 | goto fail; |
| 6155 | } | 6155 | } |
| 6156 | 6156 | ||
| 6157 | /* There are several dir indexes for this inode, clear the cache. */ | 6157 | /* There are several dir indexes for this inode, clear the cache. */ |
| 6158 | BTRFS_I(inode)->dir_index = 0ULL; | 6158 | BTRFS_I(inode)->dir_index = 0ULL; |
| 6159 | inc_nlink(inode); | 6159 | inc_nlink(inode); |
| 6160 | inode_inc_iversion(inode); | 6160 | inode_inc_iversion(inode); |
| 6161 | inode->i_ctime = CURRENT_TIME; | 6161 | inode->i_ctime = CURRENT_TIME; |
| 6162 | ihold(inode); | 6162 | ihold(inode); |
| 6163 | set_bit(BTRFS_INODE_COPY_EVERYTHING, &BTRFS_I(inode)->runtime_flags); | 6163 | set_bit(BTRFS_INODE_COPY_EVERYTHING, &BTRFS_I(inode)->runtime_flags); |
| 6164 | 6164 | ||
| 6165 | err = btrfs_add_nondir(trans, dir, dentry, inode, 1, index); | 6165 | err = btrfs_add_nondir(trans, dir, dentry, inode, 1, index); |
| 6166 | 6166 | ||
| 6167 | if (err) { | 6167 | if (err) { |
| 6168 | drop_inode = 1; | 6168 | drop_inode = 1; |
| 6169 | } else { | 6169 | } else { |
| 6170 | struct dentry *parent = dentry->d_parent; | 6170 | struct dentry *parent = dentry->d_parent; |
| 6171 | err = btrfs_update_inode(trans, root, inode); | 6171 | err = btrfs_update_inode(trans, root, inode); |
| 6172 | if (err) | 6172 | if (err) |
| 6173 | goto fail; | 6173 | goto fail; |
| 6174 | if (inode->i_nlink == 1) { | 6174 | if (inode->i_nlink == 1) { |
| 6175 | /* | 6175 | /* |
| 6176 | * If new hard link count is 1, it's a file created | 6176 | * If new hard link count is 1, it's a file created |
| 6177 | * with open(2) O_TMPFILE flag. | 6177 | * with open(2) O_TMPFILE flag. |
| 6178 | */ | 6178 | */ |
| 6179 | err = btrfs_orphan_del(trans, inode); | 6179 | err = btrfs_orphan_del(trans, inode); |
| 6180 | if (err) | 6180 | if (err) |
| 6181 | goto fail; | 6181 | goto fail; |
| 6182 | } | 6182 | } |
| 6183 | d_instantiate(dentry, inode); | 6183 | d_instantiate(dentry, inode); |
| 6184 | btrfs_log_new_name(trans, inode, NULL, parent); | 6184 | btrfs_log_new_name(trans, inode, NULL, parent); |
| 6185 | } | 6185 | } |
| 6186 | 6186 | ||
| 6187 | btrfs_end_transaction(trans, root); | 6187 | btrfs_end_transaction(trans, root); |
| 6188 | btrfs_balance_delayed_items(root); | 6188 | btrfs_balance_delayed_items(root); |
| 6189 | fail: | 6189 | fail: |
| 6190 | if (drop_inode) { | 6190 | if (drop_inode) { |
| 6191 | inode_dec_link_count(inode); | 6191 | inode_dec_link_count(inode); |
| 6192 | iput(inode); | 6192 | iput(inode); |
| 6193 | } | 6193 | } |
| 6194 | btrfs_btree_balance_dirty(root); | 6194 | btrfs_btree_balance_dirty(root); |
| 6195 | return err; | 6195 | return err; |
| 6196 | } | 6196 | } |
| 6197 | 6197 | ||
| 6198 | static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) | 6198 | static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) |
| 6199 | { | 6199 | { |
| 6200 | struct inode *inode = NULL; | 6200 | struct inode *inode = NULL; |
| 6201 | struct btrfs_trans_handle *trans; | 6201 | struct btrfs_trans_handle *trans; |
| 6202 | struct btrfs_root *root = BTRFS_I(dir)->root; | 6202 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 6203 | int err = 0; | 6203 | int err = 0; |
| 6204 | int drop_on_err = 0; | 6204 | int drop_on_err = 0; |
| 6205 | u64 objectid = 0; | 6205 | u64 objectid = 0; |
| 6206 | u64 index = 0; | 6206 | u64 index = 0; |
| 6207 | 6207 | ||
| 6208 | /* | 6208 | /* |
| 6209 | * 2 items for inode and ref | 6209 | * 2 items for inode and ref |
| 6210 | * 2 items for dir items | 6210 | * 2 items for dir items |
| 6211 | * 1 for xattr if selinux is on | 6211 | * 1 for xattr if selinux is on |
| 6212 | */ | 6212 | */ |
| 6213 | trans = btrfs_start_transaction(root, 5); | 6213 | trans = btrfs_start_transaction(root, 5); |
| 6214 | if (IS_ERR(trans)) | 6214 | if (IS_ERR(trans)) |
| 6215 | return PTR_ERR(trans); | 6215 | return PTR_ERR(trans); |
| 6216 | 6216 | ||
| 6217 | err = btrfs_find_free_ino(root, &objectid); | 6217 | err = btrfs_find_free_ino(root, &objectid); |
| 6218 | if (err) | 6218 | if (err) |
| 6219 | goto out_fail; | 6219 | goto out_fail; |
| 6220 | 6220 | ||
| 6221 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, | 6221 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, |
| 6222 | dentry->d_name.len, btrfs_ino(dir), objectid, | 6222 | dentry->d_name.len, btrfs_ino(dir), objectid, |
| 6223 | S_IFDIR | mode, &index); | 6223 | S_IFDIR | mode, &index); |
| 6224 | if (IS_ERR(inode)) { | 6224 | if (IS_ERR(inode)) { |
| 6225 | err = PTR_ERR(inode); | 6225 | err = PTR_ERR(inode); |
| 6226 | goto out_fail; | 6226 | goto out_fail; |
| 6227 | } | 6227 | } |
| 6228 | 6228 | ||
| 6229 | drop_on_err = 1; | 6229 | drop_on_err = 1; |
| 6230 | /* these must be set before we unlock the inode */ | 6230 | /* these must be set before we unlock the inode */ |
| 6231 | inode->i_op = &btrfs_dir_inode_operations; | 6231 | inode->i_op = &btrfs_dir_inode_operations; |
| 6232 | inode->i_fop = &btrfs_dir_file_operations; | 6232 | inode->i_fop = &btrfs_dir_file_operations; |
| 6233 | 6233 | ||
| 6234 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); | 6234 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); |
| 6235 | if (err) | 6235 | if (err) |
| 6236 | goto out_fail_inode; | 6236 | goto out_fail_inode; |
| 6237 | 6237 | ||
| 6238 | btrfs_i_size_write(inode, 0); | 6238 | btrfs_i_size_write(inode, 0); |
| 6239 | err = btrfs_update_inode(trans, root, inode); | 6239 | err = btrfs_update_inode(trans, root, inode); |
| 6240 | if (err) | 6240 | if (err) |
| 6241 | goto out_fail_inode; | 6241 | goto out_fail_inode; |
| 6242 | 6242 | ||
| 6243 | err = btrfs_add_link(trans, dir, inode, dentry->d_name.name, | 6243 | err = btrfs_add_link(trans, dir, inode, dentry->d_name.name, |
| 6244 | dentry->d_name.len, 0, index); | 6244 | dentry->d_name.len, 0, index); |
| 6245 | if (err) | 6245 | if (err) |
| 6246 | goto out_fail_inode; | 6246 | goto out_fail_inode; |
| 6247 | 6247 | ||
| 6248 | d_instantiate(dentry, inode); | 6248 | d_instantiate(dentry, inode); |
| 6249 | /* | 6249 | /* |
| 6250 | * mkdir is special. We're unlocking after we call d_instantiate | 6250 | * mkdir is special. We're unlocking after we call d_instantiate |
| 6251 | * to avoid a race with nfsd calling d_instantiate. | 6251 | * to avoid a race with nfsd calling d_instantiate. |
| 6252 | */ | 6252 | */ |
| 6253 | unlock_new_inode(inode); | 6253 | unlock_new_inode(inode); |
| 6254 | drop_on_err = 0; | 6254 | drop_on_err = 0; |
| 6255 | 6255 | ||
| 6256 | out_fail: | 6256 | out_fail: |
| 6257 | btrfs_end_transaction(trans, root); | 6257 | btrfs_end_transaction(trans, root); |
| 6258 | if (drop_on_err) | 6258 | if (drop_on_err) { |
| 6259 | inode_dec_link_count(inode); | ||
| 6259 | iput(inode); | 6260 | iput(inode); |
| 6261 | } | ||
| 6260 | btrfs_balance_delayed_items(root); | 6262 | btrfs_balance_delayed_items(root); |
| 6261 | btrfs_btree_balance_dirty(root); | 6263 | btrfs_btree_balance_dirty(root); |
| 6262 | return err; | 6264 | return err; |
| 6263 | 6265 | ||
| 6264 | out_fail_inode: | 6266 | out_fail_inode: |
| 6265 | unlock_new_inode(inode); | 6267 | unlock_new_inode(inode); |
| 6266 | goto out_fail; | 6268 | goto out_fail; |
| 6267 | } | 6269 | } |
| 6268 | 6270 | ||
| 6269 | /* Find next extent map of a given extent map, caller needs to ensure locks */ | 6271 | /* Find next extent map of a given extent map, caller needs to ensure locks */ |
| 6270 | static struct extent_map *next_extent_map(struct extent_map *em) | 6272 | static struct extent_map *next_extent_map(struct extent_map *em) |
| 6271 | { | 6273 | { |
| 6272 | struct rb_node *next; | 6274 | struct rb_node *next; |
| 6273 | 6275 | ||
| 6274 | next = rb_next(&em->rb_node); | 6276 | next = rb_next(&em->rb_node); |
| 6275 | if (!next) | 6277 | if (!next) |
| 6276 | return NULL; | 6278 | return NULL; |
| 6277 | return container_of(next, struct extent_map, rb_node); | 6279 | return container_of(next, struct extent_map, rb_node); |
| 6278 | } | 6280 | } |
| 6279 | 6281 | ||
| 6280 | static struct extent_map *prev_extent_map(struct extent_map *em) | 6282 | static struct extent_map *prev_extent_map(struct extent_map *em) |
| 6281 | { | 6283 | { |
| 6282 | struct rb_node *prev; | 6284 | struct rb_node *prev; |
| 6283 | 6285 | ||
| 6284 | prev = rb_prev(&em->rb_node); | 6286 | prev = rb_prev(&em->rb_node); |
| 6285 | if (!prev) | 6287 | if (!prev) |
| 6286 | return NULL; | 6288 | return NULL; |
| 6287 | return container_of(prev, struct extent_map, rb_node); | 6289 | return container_of(prev, struct extent_map, rb_node); |
| 6288 | } | 6290 | } |
| 6289 | 6291 | ||
| 6290 | /* helper for btfs_get_extent. Given an existing extent in the tree, | 6292 | /* helper for btfs_get_extent. Given an existing extent in the tree, |
| 6291 | * the existing extent is the nearest extent to map_start, | 6293 | * the existing extent is the nearest extent to map_start, |
| 6292 | * and an extent that you want to insert, deal with overlap and insert | 6294 | * and an extent that you want to insert, deal with overlap and insert |
| 6293 | * the best fitted new extent into the tree. | 6295 | * the best fitted new extent into the tree. |
| 6294 | */ | 6296 | */ |
| 6295 | static int merge_extent_mapping(struct extent_map_tree *em_tree, | 6297 | static int merge_extent_mapping(struct extent_map_tree *em_tree, |
| 6296 | struct extent_map *existing, | 6298 | struct extent_map *existing, |
| 6297 | struct extent_map *em, | 6299 | struct extent_map *em, |
| 6298 | u64 map_start) | 6300 | u64 map_start) |
| 6299 | { | 6301 | { |
| 6300 | struct extent_map *prev; | 6302 | struct extent_map *prev; |
| 6301 | struct extent_map *next; | 6303 | struct extent_map *next; |
| 6302 | u64 start; | 6304 | u64 start; |
| 6303 | u64 end; | 6305 | u64 end; |
| 6304 | u64 start_diff; | 6306 | u64 start_diff; |
| 6305 | 6307 | ||
| 6306 | BUG_ON(map_start < em->start || map_start >= extent_map_end(em)); | 6308 | BUG_ON(map_start < em->start || map_start >= extent_map_end(em)); |
| 6307 | 6309 | ||
| 6308 | if (existing->start > map_start) { | 6310 | if (existing->start > map_start) { |
| 6309 | next = existing; | 6311 | next = existing; |
| 6310 | prev = prev_extent_map(next); | 6312 | prev = prev_extent_map(next); |
| 6311 | } else { | 6313 | } else { |
| 6312 | prev = existing; | 6314 | prev = existing; |
| 6313 | next = next_extent_map(prev); | 6315 | next = next_extent_map(prev); |
| 6314 | } | 6316 | } |
| 6315 | 6317 | ||
| 6316 | start = prev ? extent_map_end(prev) : em->start; | 6318 | start = prev ? extent_map_end(prev) : em->start; |
| 6317 | start = max_t(u64, start, em->start); | 6319 | start = max_t(u64, start, em->start); |
| 6318 | end = next ? next->start : extent_map_end(em); | 6320 | end = next ? next->start : extent_map_end(em); |
| 6319 | end = min_t(u64, end, extent_map_end(em)); | 6321 | end = min_t(u64, end, extent_map_end(em)); |
| 6320 | start_diff = start - em->start; | 6322 | start_diff = start - em->start; |
| 6321 | em->start = start; | 6323 | em->start = start; |
| 6322 | em->len = end - start; | 6324 | em->len = end - start; |
| 6323 | if (em->block_start < EXTENT_MAP_LAST_BYTE && | 6325 | if (em->block_start < EXTENT_MAP_LAST_BYTE && |
| 6324 | !test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { | 6326 | !test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) { |
| 6325 | em->block_start += start_diff; | 6327 | em->block_start += start_diff; |
| 6326 | em->block_len -= start_diff; | 6328 | em->block_len -= start_diff; |
| 6327 | } | 6329 | } |
| 6328 | return add_extent_mapping(em_tree, em, 0); | 6330 | return add_extent_mapping(em_tree, em, 0); |
| 6329 | } | 6331 | } |
| 6330 | 6332 | ||
| 6331 | static noinline int uncompress_inline(struct btrfs_path *path, | 6333 | static noinline int uncompress_inline(struct btrfs_path *path, |
| 6332 | struct inode *inode, struct page *page, | 6334 | struct inode *inode, struct page *page, |
| 6333 | size_t pg_offset, u64 extent_offset, | 6335 | size_t pg_offset, u64 extent_offset, |
| 6334 | struct btrfs_file_extent_item *item) | 6336 | struct btrfs_file_extent_item *item) |
| 6335 | { | 6337 | { |
| 6336 | int ret; | 6338 | int ret; |
| 6337 | struct extent_buffer *leaf = path->nodes[0]; | 6339 | struct extent_buffer *leaf = path->nodes[0]; |
| 6338 | char *tmp; | 6340 | char *tmp; |
| 6339 | size_t max_size; | 6341 | size_t max_size; |
| 6340 | unsigned long inline_size; | 6342 | unsigned long inline_size; |
| 6341 | unsigned long ptr; | 6343 | unsigned long ptr; |
| 6342 | int compress_type; | 6344 | int compress_type; |
| 6343 | 6345 | ||
| 6344 | WARN_ON(pg_offset != 0); | 6346 | WARN_ON(pg_offset != 0); |
| 6345 | compress_type = btrfs_file_extent_compression(leaf, item); | 6347 | compress_type = btrfs_file_extent_compression(leaf, item); |
| 6346 | max_size = btrfs_file_extent_ram_bytes(leaf, item); | 6348 | max_size = btrfs_file_extent_ram_bytes(leaf, item); |
| 6347 | inline_size = btrfs_file_extent_inline_item_len(leaf, | 6349 | inline_size = btrfs_file_extent_inline_item_len(leaf, |
| 6348 | btrfs_item_nr(path->slots[0])); | 6350 | btrfs_item_nr(path->slots[0])); |
| 6349 | tmp = kmalloc(inline_size, GFP_NOFS); | 6351 | tmp = kmalloc(inline_size, GFP_NOFS); |
| 6350 | if (!tmp) | 6352 | if (!tmp) |
| 6351 | return -ENOMEM; | 6353 | return -ENOMEM; |
| 6352 | ptr = btrfs_file_extent_inline_start(item); | 6354 | ptr = btrfs_file_extent_inline_start(item); |
| 6353 | 6355 | ||
| 6354 | read_extent_buffer(leaf, tmp, ptr, inline_size); | 6356 | read_extent_buffer(leaf, tmp, ptr, inline_size); |
| 6355 | 6357 | ||
| 6356 | max_size = min_t(unsigned long, PAGE_CACHE_SIZE, max_size); | 6358 | max_size = min_t(unsigned long, PAGE_CACHE_SIZE, max_size); |
| 6357 | ret = btrfs_decompress(compress_type, tmp, page, | 6359 | ret = btrfs_decompress(compress_type, tmp, page, |
| 6358 | extent_offset, inline_size, max_size); | 6360 | extent_offset, inline_size, max_size); |
| 6359 | kfree(tmp); | 6361 | kfree(tmp); |
| 6360 | return ret; | 6362 | return ret; |
| 6361 | } | 6363 | } |
| 6362 | 6364 | ||
| 6363 | /* | 6365 | /* |
| 6364 | * a bit scary, this does extent mapping from logical file offset to the disk. | 6366 | * a bit scary, this does extent mapping from logical file offset to the disk. |
| 6365 | * the ugly parts come from merging extents from the disk with the in-ram | 6367 | * the ugly parts come from merging extents from the disk with the in-ram |
| 6366 | * representation. This gets more complex because of the data=ordered code, | 6368 | * representation. This gets more complex because of the data=ordered code, |
| 6367 | * where the in-ram extents might be locked pending data=ordered completion. | 6369 | * where the in-ram extents might be locked pending data=ordered completion. |
| 6368 | * | 6370 | * |
| 6369 | * This also copies inline extents directly into the page. | 6371 | * This also copies inline extents directly into the page. |
| 6370 | */ | 6372 | */ |
| 6371 | 6373 | ||
| 6372 | struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page, | 6374 | struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page, |
| 6373 | size_t pg_offset, u64 start, u64 len, | 6375 | size_t pg_offset, u64 start, u64 len, |
| 6374 | int create) | 6376 | int create) |
| 6375 | { | 6377 | { |
| 6376 | int ret; | 6378 | int ret; |
| 6377 | int err = 0; | 6379 | int err = 0; |
| 6378 | u64 extent_start = 0; | 6380 | u64 extent_start = 0; |
| 6379 | u64 extent_end = 0; | 6381 | u64 extent_end = 0; |
| 6380 | u64 objectid = btrfs_ino(inode); | 6382 | u64 objectid = btrfs_ino(inode); |
| 6381 | u32 found_type; | 6383 | u32 found_type; |
| 6382 | struct btrfs_path *path = NULL; | 6384 | struct btrfs_path *path = NULL; |
| 6383 | struct btrfs_root *root = BTRFS_I(inode)->root; | 6385 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 6384 | struct btrfs_file_extent_item *item; | 6386 | struct btrfs_file_extent_item *item; |
| 6385 | struct extent_buffer *leaf; | 6387 | struct extent_buffer *leaf; |
| 6386 | struct btrfs_key found_key; | 6388 | struct btrfs_key found_key; |
| 6387 | struct extent_map *em = NULL; | 6389 | struct extent_map *em = NULL; |
| 6388 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; | 6390 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; |
| 6389 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 6391 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 6390 | struct btrfs_trans_handle *trans = NULL; | 6392 | struct btrfs_trans_handle *trans = NULL; |
| 6391 | const bool new_inline = !page || create; | 6393 | const bool new_inline = !page || create; |
| 6392 | 6394 | ||
| 6393 | again: | 6395 | again: |
| 6394 | read_lock(&em_tree->lock); | 6396 | read_lock(&em_tree->lock); |
| 6395 | em = lookup_extent_mapping(em_tree, start, len); | 6397 | em = lookup_extent_mapping(em_tree, start, len); |
| 6396 | if (em) | 6398 | if (em) |
| 6397 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 6399 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 6398 | read_unlock(&em_tree->lock); | 6400 | read_unlock(&em_tree->lock); |
| 6399 | 6401 | ||
| 6400 | if (em) { | 6402 | if (em) { |
| 6401 | if (em->start > start || em->start + em->len <= start) | 6403 | if (em->start > start || em->start + em->len <= start) |
| 6402 | free_extent_map(em); | 6404 | free_extent_map(em); |
| 6403 | else if (em->block_start == EXTENT_MAP_INLINE && page) | 6405 | else if (em->block_start == EXTENT_MAP_INLINE && page) |
| 6404 | free_extent_map(em); | 6406 | free_extent_map(em); |
| 6405 | else | 6407 | else |
| 6406 | goto out; | 6408 | goto out; |
| 6407 | } | 6409 | } |
| 6408 | em = alloc_extent_map(); | 6410 | em = alloc_extent_map(); |
| 6409 | if (!em) { | 6411 | if (!em) { |
| 6410 | err = -ENOMEM; | 6412 | err = -ENOMEM; |
| 6411 | goto out; | 6413 | goto out; |
| 6412 | } | 6414 | } |
| 6413 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 6415 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 6414 | em->start = EXTENT_MAP_HOLE; | 6416 | em->start = EXTENT_MAP_HOLE; |
| 6415 | em->orig_start = EXTENT_MAP_HOLE; | 6417 | em->orig_start = EXTENT_MAP_HOLE; |
| 6416 | em->len = (u64)-1; | 6418 | em->len = (u64)-1; |
| 6417 | em->block_len = (u64)-1; | 6419 | em->block_len = (u64)-1; |
| 6418 | 6420 | ||
| 6419 | if (!path) { | 6421 | if (!path) { |
| 6420 | path = btrfs_alloc_path(); | 6422 | path = btrfs_alloc_path(); |
| 6421 | if (!path) { | 6423 | if (!path) { |
| 6422 | err = -ENOMEM; | 6424 | err = -ENOMEM; |
| 6423 | goto out; | 6425 | goto out; |
| 6424 | } | 6426 | } |
| 6425 | /* | 6427 | /* |
| 6426 | * Chances are we'll be called again, so go ahead and do | 6428 | * Chances are we'll be called again, so go ahead and do |
| 6427 | * readahead | 6429 | * readahead |
| 6428 | */ | 6430 | */ |
| 6429 | path->reada = 1; | 6431 | path->reada = 1; |
| 6430 | } | 6432 | } |
| 6431 | 6433 | ||
| 6432 | ret = btrfs_lookup_file_extent(trans, root, path, | 6434 | ret = btrfs_lookup_file_extent(trans, root, path, |
| 6433 | objectid, start, trans != NULL); | 6435 | objectid, start, trans != NULL); |
| 6434 | if (ret < 0) { | 6436 | if (ret < 0) { |
| 6435 | err = ret; | 6437 | err = ret; |
| 6436 | goto out; | 6438 | goto out; |
| 6437 | } | 6439 | } |
| 6438 | 6440 | ||
| 6439 | if (ret != 0) { | 6441 | if (ret != 0) { |
| 6440 | if (path->slots[0] == 0) | 6442 | if (path->slots[0] == 0) |
| 6441 | goto not_found; | 6443 | goto not_found; |
| 6442 | path->slots[0]--; | 6444 | path->slots[0]--; |
| 6443 | } | 6445 | } |
| 6444 | 6446 | ||
| 6445 | leaf = path->nodes[0]; | 6447 | leaf = path->nodes[0]; |
| 6446 | item = btrfs_item_ptr(leaf, path->slots[0], | 6448 | item = btrfs_item_ptr(leaf, path->slots[0], |
| 6447 | struct btrfs_file_extent_item); | 6449 | struct btrfs_file_extent_item); |
| 6448 | /* are we inside the extent that was found? */ | 6450 | /* are we inside the extent that was found? */ |
| 6449 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 6451 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 6450 | found_type = found_key.type; | 6452 | found_type = found_key.type; |
| 6451 | if (found_key.objectid != objectid || | 6453 | if (found_key.objectid != objectid || |
| 6452 | found_type != BTRFS_EXTENT_DATA_KEY) { | 6454 | found_type != BTRFS_EXTENT_DATA_KEY) { |
| 6453 | /* | 6455 | /* |
| 6454 | * If we backup past the first extent we want to move forward | 6456 | * If we backup past the first extent we want to move forward |
| 6455 | * and see if there is an extent in front of us, otherwise we'll | 6457 | * and see if there is an extent in front of us, otherwise we'll |
| 6456 | * say there is a hole for our whole search range which can | 6458 | * say there is a hole for our whole search range which can |
| 6457 | * cause problems. | 6459 | * cause problems. |
| 6458 | */ | 6460 | */ |
| 6459 | extent_end = start; | 6461 | extent_end = start; |
| 6460 | goto next; | 6462 | goto next; |
| 6461 | } | 6463 | } |
| 6462 | 6464 | ||
| 6463 | found_type = btrfs_file_extent_type(leaf, item); | 6465 | found_type = btrfs_file_extent_type(leaf, item); |
| 6464 | extent_start = found_key.offset; | 6466 | extent_start = found_key.offset; |
| 6465 | if (found_type == BTRFS_FILE_EXTENT_REG || | 6467 | if (found_type == BTRFS_FILE_EXTENT_REG || |
| 6466 | found_type == BTRFS_FILE_EXTENT_PREALLOC) { | 6468 | found_type == BTRFS_FILE_EXTENT_PREALLOC) { |
| 6467 | extent_end = extent_start + | 6469 | extent_end = extent_start + |
| 6468 | btrfs_file_extent_num_bytes(leaf, item); | 6470 | btrfs_file_extent_num_bytes(leaf, item); |
| 6469 | } else if (found_type == BTRFS_FILE_EXTENT_INLINE) { | 6471 | } else if (found_type == BTRFS_FILE_EXTENT_INLINE) { |
| 6470 | size_t size; | 6472 | size_t size; |
| 6471 | size = btrfs_file_extent_inline_len(leaf, path->slots[0], item); | 6473 | size = btrfs_file_extent_inline_len(leaf, path->slots[0], item); |
| 6472 | extent_end = ALIGN(extent_start + size, root->sectorsize); | 6474 | extent_end = ALIGN(extent_start + size, root->sectorsize); |
| 6473 | } | 6475 | } |
| 6474 | next: | 6476 | next: |
| 6475 | if (start >= extent_end) { | 6477 | if (start >= extent_end) { |
| 6476 | path->slots[0]++; | 6478 | path->slots[0]++; |
| 6477 | if (path->slots[0] >= btrfs_header_nritems(leaf)) { | 6479 | if (path->slots[0] >= btrfs_header_nritems(leaf)) { |
| 6478 | ret = btrfs_next_leaf(root, path); | 6480 | ret = btrfs_next_leaf(root, path); |
| 6479 | if (ret < 0) { | 6481 | if (ret < 0) { |
| 6480 | err = ret; | 6482 | err = ret; |
| 6481 | goto out; | 6483 | goto out; |
| 6482 | } | 6484 | } |
| 6483 | if (ret > 0) | 6485 | if (ret > 0) |
| 6484 | goto not_found; | 6486 | goto not_found; |
| 6485 | leaf = path->nodes[0]; | 6487 | leaf = path->nodes[0]; |
| 6486 | } | 6488 | } |
| 6487 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); | 6489 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
| 6488 | if (found_key.objectid != objectid || | 6490 | if (found_key.objectid != objectid || |
| 6489 | found_key.type != BTRFS_EXTENT_DATA_KEY) | 6491 | found_key.type != BTRFS_EXTENT_DATA_KEY) |
| 6490 | goto not_found; | 6492 | goto not_found; |
| 6491 | if (start + len <= found_key.offset) | 6493 | if (start + len <= found_key.offset) |
| 6492 | goto not_found; | 6494 | goto not_found; |
| 6493 | if (start > found_key.offset) | 6495 | if (start > found_key.offset) |
| 6494 | goto next; | 6496 | goto next; |
| 6495 | em->start = start; | 6497 | em->start = start; |
| 6496 | em->orig_start = start; | 6498 | em->orig_start = start; |
| 6497 | em->len = found_key.offset - start; | 6499 | em->len = found_key.offset - start; |
| 6498 | goto not_found_em; | 6500 | goto not_found_em; |
| 6499 | } | 6501 | } |
| 6500 | 6502 | ||
| 6501 | btrfs_extent_item_to_extent_map(inode, path, item, new_inline, em); | 6503 | btrfs_extent_item_to_extent_map(inode, path, item, new_inline, em); |
| 6502 | 6504 | ||
| 6503 | if (found_type == BTRFS_FILE_EXTENT_REG || | 6505 | if (found_type == BTRFS_FILE_EXTENT_REG || |
| 6504 | found_type == BTRFS_FILE_EXTENT_PREALLOC) { | 6506 | found_type == BTRFS_FILE_EXTENT_PREALLOC) { |
| 6505 | goto insert; | 6507 | goto insert; |
| 6506 | } else if (found_type == BTRFS_FILE_EXTENT_INLINE) { | 6508 | } else if (found_type == BTRFS_FILE_EXTENT_INLINE) { |
| 6507 | unsigned long ptr; | 6509 | unsigned long ptr; |
| 6508 | char *map; | 6510 | char *map; |
| 6509 | size_t size; | 6511 | size_t size; |
| 6510 | size_t extent_offset; | 6512 | size_t extent_offset; |
| 6511 | size_t copy_size; | 6513 | size_t copy_size; |
| 6512 | 6514 | ||
| 6513 | if (new_inline) | 6515 | if (new_inline) |
| 6514 | goto out; | 6516 | goto out; |
| 6515 | 6517 | ||
| 6516 | size = btrfs_file_extent_inline_len(leaf, path->slots[0], item); | 6518 | size = btrfs_file_extent_inline_len(leaf, path->slots[0], item); |
| 6517 | extent_offset = page_offset(page) + pg_offset - extent_start; | 6519 | extent_offset = page_offset(page) + pg_offset - extent_start; |
| 6518 | copy_size = min_t(u64, PAGE_CACHE_SIZE - pg_offset, | 6520 | copy_size = min_t(u64, PAGE_CACHE_SIZE - pg_offset, |
| 6519 | size - extent_offset); | 6521 | size - extent_offset); |
| 6520 | em->start = extent_start + extent_offset; | 6522 | em->start = extent_start + extent_offset; |
| 6521 | em->len = ALIGN(copy_size, root->sectorsize); | 6523 | em->len = ALIGN(copy_size, root->sectorsize); |
| 6522 | em->orig_block_len = em->len; | 6524 | em->orig_block_len = em->len; |
| 6523 | em->orig_start = em->start; | 6525 | em->orig_start = em->start; |
| 6524 | ptr = btrfs_file_extent_inline_start(item) + extent_offset; | 6526 | ptr = btrfs_file_extent_inline_start(item) + extent_offset; |
| 6525 | if (create == 0 && !PageUptodate(page)) { | 6527 | if (create == 0 && !PageUptodate(page)) { |
| 6526 | if (btrfs_file_extent_compression(leaf, item) != | 6528 | if (btrfs_file_extent_compression(leaf, item) != |
| 6527 | BTRFS_COMPRESS_NONE) { | 6529 | BTRFS_COMPRESS_NONE) { |
| 6528 | ret = uncompress_inline(path, inode, page, | 6530 | ret = uncompress_inline(path, inode, page, |
| 6529 | pg_offset, | 6531 | pg_offset, |
| 6530 | extent_offset, item); | 6532 | extent_offset, item); |
| 6531 | if (ret) { | 6533 | if (ret) { |
| 6532 | err = ret; | 6534 | err = ret; |
| 6533 | goto out; | 6535 | goto out; |
| 6534 | } | 6536 | } |
| 6535 | } else { | 6537 | } else { |
| 6536 | map = kmap(page); | 6538 | map = kmap(page); |
| 6537 | read_extent_buffer(leaf, map + pg_offset, ptr, | 6539 | read_extent_buffer(leaf, map + pg_offset, ptr, |
| 6538 | copy_size); | 6540 | copy_size); |
| 6539 | if (pg_offset + copy_size < PAGE_CACHE_SIZE) { | 6541 | if (pg_offset + copy_size < PAGE_CACHE_SIZE) { |
| 6540 | memset(map + pg_offset + copy_size, 0, | 6542 | memset(map + pg_offset + copy_size, 0, |
| 6541 | PAGE_CACHE_SIZE - pg_offset - | 6543 | PAGE_CACHE_SIZE - pg_offset - |
| 6542 | copy_size); | 6544 | copy_size); |
| 6543 | } | 6545 | } |
| 6544 | kunmap(page); | 6546 | kunmap(page); |
| 6545 | } | 6547 | } |
| 6546 | flush_dcache_page(page); | 6548 | flush_dcache_page(page); |
| 6547 | } else if (create && PageUptodate(page)) { | 6549 | } else if (create && PageUptodate(page)) { |
| 6548 | BUG(); | 6550 | BUG(); |
| 6549 | if (!trans) { | 6551 | if (!trans) { |
| 6550 | kunmap(page); | 6552 | kunmap(page); |
| 6551 | free_extent_map(em); | 6553 | free_extent_map(em); |
| 6552 | em = NULL; | 6554 | em = NULL; |
| 6553 | 6555 | ||
| 6554 | btrfs_release_path(path); | 6556 | btrfs_release_path(path); |
| 6555 | trans = btrfs_join_transaction(root); | 6557 | trans = btrfs_join_transaction(root); |
| 6556 | 6558 | ||
| 6557 | if (IS_ERR(trans)) | 6559 | if (IS_ERR(trans)) |
| 6558 | return ERR_CAST(trans); | 6560 | return ERR_CAST(trans); |
| 6559 | goto again; | 6561 | goto again; |
| 6560 | } | 6562 | } |
| 6561 | map = kmap(page); | 6563 | map = kmap(page); |
| 6562 | write_extent_buffer(leaf, map + pg_offset, ptr, | 6564 | write_extent_buffer(leaf, map + pg_offset, ptr, |
| 6563 | copy_size); | 6565 | copy_size); |
| 6564 | kunmap(page); | 6566 | kunmap(page); |
| 6565 | btrfs_mark_buffer_dirty(leaf); | 6567 | btrfs_mark_buffer_dirty(leaf); |
| 6566 | } | 6568 | } |
| 6567 | set_extent_uptodate(io_tree, em->start, | 6569 | set_extent_uptodate(io_tree, em->start, |
| 6568 | extent_map_end(em) - 1, NULL, GFP_NOFS); | 6570 | extent_map_end(em) - 1, NULL, GFP_NOFS); |
| 6569 | goto insert; | 6571 | goto insert; |
| 6570 | } | 6572 | } |
| 6571 | not_found: | 6573 | not_found: |
| 6572 | em->start = start; | 6574 | em->start = start; |
| 6573 | em->orig_start = start; | 6575 | em->orig_start = start; |
| 6574 | em->len = len; | 6576 | em->len = len; |
| 6575 | not_found_em: | 6577 | not_found_em: |
| 6576 | em->block_start = EXTENT_MAP_HOLE; | 6578 | em->block_start = EXTENT_MAP_HOLE; |
| 6577 | set_bit(EXTENT_FLAG_VACANCY, &em->flags); | 6579 | set_bit(EXTENT_FLAG_VACANCY, &em->flags); |
| 6578 | insert: | 6580 | insert: |
| 6579 | btrfs_release_path(path); | 6581 | btrfs_release_path(path); |
| 6580 | if (em->start > start || extent_map_end(em) <= start) { | 6582 | if (em->start > start || extent_map_end(em) <= start) { |
| 6581 | btrfs_err(root->fs_info, "bad extent! em: [%llu %llu] passed [%llu %llu]", | 6583 | btrfs_err(root->fs_info, "bad extent! em: [%llu %llu] passed [%llu %llu]", |
| 6582 | em->start, em->len, start, len); | 6584 | em->start, em->len, start, len); |
| 6583 | err = -EIO; | 6585 | err = -EIO; |
| 6584 | goto out; | 6586 | goto out; |
| 6585 | } | 6587 | } |
| 6586 | 6588 | ||
| 6587 | err = 0; | 6589 | err = 0; |
| 6588 | write_lock(&em_tree->lock); | 6590 | write_lock(&em_tree->lock); |
| 6589 | ret = add_extent_mapping(em_tree, em, 0); | 6591 | ret = add_extent_mapping(em_tree, em, 0); |
| 6590 | /* it is possible that someone inserted the extent into the tree | 6592 | /* it is possible that someone inserted the extent into the tree |
| 6591 | * while we had the lock dropped. It is also possible that | 6593 | * while we had the lock dropped. It is also possible that |
| 6592 | * an overlapping map exists in the tree | 6594 | * an overlapping map exists in the tree |
| 6593 | */ | 6595 | */ |
| 6594 | if (ret == -EEXIST) { | 6596 | if (ret == -EEXIST) { |
| 6595 | struct extent_map *existing; | 6597 | struct extent_map *existing; |
| 6596 | 6598 | ||
| 6597 | ret = 0; | 6599 | ret = 0; |
| 6598 | 6600 | ||
| 6599 | existing = search_extent_mapping(em_tree, start, len); | 6601 | existing = search_extent_mapping(em_tree, start, len); |
| 6600 | /* | 6602 | /* |
| 6601 | * existing will always be non-NULL, since there must be | 6603 | * existing will always be non-NULL, since there must be |
| 6602 | * extent causing the -EEXIST. | 6604 | * extent causing the -EEXIST. |
| 6603 | */ | 6605 | */ |
| 6604 | if (start >= extent_map_end(existing) || | 6606 | if (start >= extent_map_end(existing) || |
| 6605 | start <= existing->start) { | 6607 | start <= existing->start) { |
| 6606 | /* | 6608 | /* |
| 6607 | * The existing extent map is the one nearest to | 6609 | * The existing extent map is the one nearest to |
| 6608 | * the [start, start + len) range which overlaps | 6610 | * the [start, start + len) range which overlaps |
| 6609 | */ | 6611 | */ |
| 6610 | err = merge_extent_mapping(em_tree, existing, | 6612 | err = merge_extent_mapping(em_tree, existing, |
| 6611 | em, start); | 6613 | em, start); |
| 6612 | free_extent_map(existing); | 6614 | free_extent_map(existing); |
| 6613 | if (err) { | 6615 | if (err) { |
| 6614 | free_extent_map(em); | 6616 | free_extent_map(em); |
| 6615 | em = NULL; | 6617 | em = NULL; |
| 6616 | } | 6618 | } |
| 6617 | } else { | 6619 | } else { |
| 6618 | free_extent_map(em); | 6620 | free_extent_map(em); |
| 6619 | em = existing; | 6621 | em = existing; |
| 6620 | err = 0; | 6622 | err = 0; |
| 6621 | } | 6623 | } |
| 6622 | } | 6624 | } |
| 6623 | write_unlock(&em_tree->lock); | 6625 | write_unlock(&em_tree->lock); |
| 6624 | out: | 6626 | out: |
| 6625 | 6627 | ||
| 6626 | trace_btrfs_get_extent(root, em); | 6628 | trace_btrfs_get_extent(root, em); |
| 6627 | 6629 | ||
| 6628 | if (path) | 6630 | if (path) |
| 6629 | btrfs_free_path(path); | 6631 | btrfs_free_path(path); |
| 6630 | if (trans) { | 6632 | if (trans) { |
| 6631 | ret = btrfs_end_transaction(trans, root); | 6633 | ret = btrfs_end_transaction(trans, root); |
| 6632 | if (!err) | 6634 | if (!err) |
| 6633 | err = ret; | 6635 | err = ret; |
| 6634 | } | 6636 | } |
| 6635 | if (err) { | 6637 | if (err) { |
| 6636 | free_extent_map(em); | 6638 | free_extent_map(em); |
| 6637 | return ERR_PTR(err); | 6639 | return ERR_PTR(err); |
| 6638 | } | 6640 | } |
| 6639 | BUG_ON(!em); /* Error is always set */ | 6641 | BUG_ON(!em); /* Error is always set */ |
| 6640 | return em; | 6642 | return em; |
| 6641 | } | 6643 | } |
| 6642 | 6644 | ||
| 6643 | struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *page, | 6645 | struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *page, |
| 6644 | size_t pg_offset, u64 start, u64 len, | 6646 | size_t pg_offset, u64 start, u64 len, |
| 6645 | int create) | 6647 | int create) |
| 6646 | { | 6648 | { |
| 6647 | struct extent_map *em; | 6649 | struct extent_map *em; |
| 6648 | struct extent_map *hole_em = NULL; | 6650 | struct extent_map *hole_em = NULL; |
| 6649 | u64 range_start = start; | 6651 | u64 range_start = start; |
| 6650 | u64 end; | 6652 | u64 end; |
| 6651 | u64 found; | 6653 | u64 found; |
| 6652 | u64 found_end; | 6654 | u64 found_end; |
| 6653 | int err = 0; | 6655 | int err = 0; |
| 6654 | 6656 | ||
| 6655 | em = btrfs_get_extent(inode, page, pg_offset, start, len, create); | 6657 | em = btrfs_get_extent(inode, page, pg_offset, start, len, create); |
| 6656 | if (IS_ERR(em)) | 6658 | if (IS_ERR(em)) |
| 6657 | return em; | 6659 | return em; |
| 6658 | if (em) { | 6660 | if (em) { |
| 6659 | /* | 6661 | /* |
| 6660 | * if our em maps to | 6662 | * if our em maps to |
| 6661 | * - a hole or | 6663 | * - a hole or |
| 6662 | * - a pre-alloc extent, | 6664 | * - a pre-alloc extent, |
| 6663 | * there might actually be delalloc bytes behind it. | 6665 | * there might actually be delalloc bytes behind it. |
| 6664 | */ | 6666 | */ |
| 6665 | if (em->block_start != EXTENT_MAP_HOLE && | 6667 | if (em->block_start != EXTENT_MAP_HOLE && |
| 6666 | !test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) | 6668 | !test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) |
| 6667 | return em; | 6669 | return em; |
| 6668 | else | 6670 | else |
| 6669 | hole_em = em; | 6671 | hole_em = em; |
| 6670 | } | 6672 | } |
| 6671 | 6673 | ||
| 6672 | /* check to see if we've wrapped (len == -1 or similar) */ | 6674 | /* check to see if we've wrapped (len == -1 or similar) */ |
| 6673 | end = start + len; | 6675 | end = start + len; |
| 6674 | if (end < start) | 6676 | if (end < start) |
| 6675 | end = (u64)-1; | 6677 | end = (u64)-1; |
| 6676 | else | 6678 | else |
| 6677 | end -= 1; | 6679 | end -= 1; |
| 6678 | 6680 | ||
| 6679 | em = NULL; | 6681 | em = NULL; |
| 6680 | 6682 | ||
| 6681 | /* ok, we didn't find anything, lets look for delalloc */ | 6683 | /* ok, we didn't find anything, lets look for delalloc */ |
| 6682 | found = count_range_bits(&BTRFS_I(inode)->io_tree, &range_start, | 6684 | found = count_range_bits(&BTRFS_I(inode)->io_tree, &range_start, |
| 6683 | end, len, EXTENT_DELALLOC, 1); | 6685 | end, len, EXTENT_DELALLOC, 1); |
| 6684 | found_end = range_start + found; | 6686 | found_end = range_start + found; |
| 6685 | if (found_end < range_start) | 6687 | if (found_end < range_start) |
| 6686 | found_end = (u64)-1; | 6688 | found_end = (u64)-1; |
| 6687 | 6689 | ||
| 6688 | /* | 6690 | /* |
| 6689 | * we didn't find anything useful, return | 6691 | * we didn't find anything useful, return |
| 6690 | * the original results from get_extent() | 6692 | * the original results from get_extent() |
| 6691 | */ | 6693 | */ |
| 6692 | if (range_start > end || found_end <= start) { | 6694 | if (range_start > end || found_end <= start) { |
| 6693 | em = hole_em; | 6695 | em = hole_em; |
| 6694 | hole_em = NULL; | 6696 | hole_em = NULL; |
| 6695 | goto out; | 6697 | goto out; |
| 6696 | } | 6698 | } |
| 6697 | 6699 | ||
| 6698 | /* adjust the range_start to make sure it doesn't | 6700 | /* adjust the range_start to make sure it doesn't |
| 6699 | * go backwards from the start they passed in | 6701 | * go backwards from the start they passed in |
| 6700 | */ | 6702 | */ |
| 6701 | range_start = max(start, range_start); | 6703 | range_start = max(start, range_start); |
| 6702 | found = found_end - range_start; | 6704 | found = found_end - range_start; |
| 6703 | 6705 | ||
| 6704 | if (found > 0) { | 6706 | if (found > 0) { |
| 6705 | u64 hole_start = start; | 6707 | u64 hole_start = start; |
| 6706 | u64 hole_len = len; | 6708 | u64 hole_len = len; |
| 6707 | 6709 | ||
| 6708 | em = alloc_extent_map(); | 6710 | em = alloc_extent_map(); |
| 6709 | if (!em) { | 6711 | if (!em) { |
| 6710 | err = -ENOMEM; | 6712 | err = -ENOMEM; |
| 6711 | goto out; | 6713 | goto out; |
| 6712 | } | 6714 | } |
| 6713 | /* | 6715 | /* |
| 6714 | * when btrfs_get_extent can't find anything it | 6716 | * when btrfs_get_extent can't find anything it |
| 6715 | * returns one huge hole | 6717 | * returns one huge hole |
| 6716 | * | 6718 | * |
| 6717 | * make sure what it found really fits our range, and | 6719 | * make sure what it found really fits our range, and |
| 6718 | * adjust to make sure it is based on the start from | 6720 | * adjust to make sure it is based on the start from |
| 6719 | * the caller | 6721 | * the caller |
| 6720 | */ | 6722 | */ |
| 6721 | if (hole_em) { | 6723 | if (hole_em) { |
| 6722 | u64 calc_end = extent_map_end(hole_em); | 6724 | u64 calc_end = extent_map_end(hole_em); |
| 6723 | 6725 | ||
| 6724 | if (calc_end <= start || (hole_em->start > end)) { | 6726 | if (calc_end <= start || (hole_em->start > end)) { |
| 6725 | free_extent_map(hole_em); | 6727 | free_extent_map(hole_em); |
| 6726 | hole_em = NULL; | 6728 | hole_em = NULL; |
| 6727 | } else { | 6729 | } else { |
| 6728 | hole_start = max(hole_em->start, start); | 6730 | hole_start = max(hole_em->start, start); |
| 6729 | hole_len = calc_end - hole_start; | 6731 | hole_len = calc_end - hole_start; |
| 6730 | } | 6732 | } |
| 6731 | } | 6733 | } |
| 6732 | em->bdev = NULL; | 6734 | em->bdev = NULL; |
| 6733 | if (hole_em && range_start > hole_start) { | 6735 | if (hole_em && range_start > hole_start) { |
| 6734 | /* our hole starts before our delalloc, so we | 6736 | /* our hole starts before our delalloc, so we |
| 6735 | * have to return just the parts of the hole | 6737 | * have to return just the parts of the hole |
| 6736 | * that go until the delalloc starts | 6738 | * that go until the delalloc starts |
| 6737 | */ | 6739 | */ |
| 6738 | em->len = min(hole_len, | 6740 | em->len = min(hole_len, |
| 6739 | range_start - hole_start); | 6741 | range_start - hole_start); |
| 6740 | em->start = hole_start; | 6742 | em->start = hole_start; |
| 6741 | em->orig_start = hole_start; | 6743 | em->orig_start = hole_start; |
| 6742 | /* | 6744 | /* |
| 6743 | * don't adjust block start at all, | 6745 | * don't adjust block start at all, |
| 6744 | * it is fixed at EXTENT_MAP_HOLE | 6746 | * it is fixed at EXTENT_MAP_HOLE |
| 6745 | */ | 6747 | */ |
| 6746 | em->block_start = hole_em->block_start; | 6748 | em->block_start = hole_em->block_start; |
| 6747 | em->block_len = hole_len; | 6749 | em->block_len = hole_len; |
| 6748 | if (test_bit(EXTENT_FLAG_PREALLOC, &hole_em->flags)) | 6750 | if (test_bit(EXTENT_FLAG_PREALLOC, &hole_em->flags)) |
| 6749 | set_bit(EXTENT_FLAG_PREALLOC, &em->flags); | 6751 | set_bit(EXTENT_FLAG_PREALLOC, &em->flags); |
| 6750 | } else { | 6752 | } else { |
| 6751 | em->start = range_start; | 6753 | em->start = range_start; |
| 6752 | em->len = found; | 6754 | em->len = found; |
| 6753 | em->orig_start = range_start; | 6755 | em->orig_start = range_start; |
| 6754 | em->block_start = EXTENT_MAP_DELALLOC; | 6756 | em->block_start = EXTENT_MAP_DELALLOC; |
| 6755 | em->block_len = found; | 6757 | em->block_len = found; |
| 6756 | } | 6758 | } |
| 6757 | } else if (hole_em) { | 6759 | } else if (hole_em) { |
| 6758 | return hole_em; | 6760 | return hole_em; |
| 6759 | } | 6761 | } |
| 6760 | out: | 6762 | out: |
| 6761 | 6763 | ||
| 6762 | free_extent_map(hole_em); | 6764 | free_extent_map(hole_em); |
| 6763 | if (err) { | 6765 | if (err) { |
| 6764 | free_extent_map(em); | 6766 | free_extent_map(em); |
| 6765 | return ERR_PTR(err); | 6767 | return ERR_PTR(err); |
| 6766 | } | 6768 | } |
| 6767 | return em; | 6769 | return em; |
| 6768 | } | 6770 | } |
| 6769 | 6771 | ||
| 6770 | static struct extent_map *btrfs_new_extent_direct(struct inode *inode, | 6772 | static struct extent_map *btrfs_new_extent_direct(struct inode *inode, |
| 6771 | u64 start, u64 len) | 6773 | u64 start, u64 len) |
| 6772 | { | 6774 | { |
| 6773 | struct btrfs_root *root = BTRFS_I(inode)->root; | 6775 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 6774 | struct extent_map *em; | 6776 | struct extent_map *em; |
| 6775 | struct btrfs_key ins; | 6777 | struct btrfs_key ins; |
| 6776 | u64 alloc_hint; | 6778 | u64 alloc_hint; |
| 6777 | int ret; | 6779 | int ret; |
| 6778 | 6780 | ||
| 6779 | alloc_hint = get_extent_allocation_hint(inode, start, len); | 6781 | alloc_hint = get_extent_allocation_hint(inode, start, len); |
| 6780 | ret = btrfs_reserve_extent(root, len, root->sectorsize, 0, | 6782 | ret = btrfs_reserve_extent(root, len, root->sectorsize, 0, |
| 6781 | alloc_hint, &ins, 1, 1); | 6783 | alloc_hint, &ins, 1, 1); |
| 6782 | if (ret) | 6784 | if (ret) |
| 6783 | return ERR_PTR(ret); | 6785 | return ERR_PTR(ret); |
| 6784 | 6786 | ||
| 6785 | em = create_pinned_em(inode, start, ins.offset, start, ins.objectid, | 6787 | em = create_pinned_em(inode, start, ins.offset, start, ins.objectid, |
| 6786 | ins.offset, ins.offset, ins.offset, 0); | 6788 | ins.offset, ins.offset, ins.offset, 0); |
| 6787 | if (IS_ERR(em)) { | 6789 | if (IS_ERR(em)) { |
| 6788 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); | 6790 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); |
| 6789 | return em; | 6791 | return em; |
| 6790 | } | 6792 | } |
| 6791 | 6793 | ||
| 6792 | ret = btrfs_add_ordered_extent_dio(inode, start, ins.objectid, | 6794 | ret = btrfs_add_ordered_extent_dio(inode, start, ins.objectid, |
| 6793 | ins.offset, ins.offset, 0); | 6795 | ins.offset, ins.offset, 0); |
| 6794 | if (ret) { | 6796 | if (ret) { |
| 6795 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); | 6797 | btrfs_free_reserved_extent(root, ins.objectid, ins.offset, 1); |
| 6796 | free_extent_map(em); | 6798 | free_extent_map(em); |
| 6797 | return ERR_PTR(ret); | 6799 | return ERR_PTR(ret); |
| 6798 | } | 6800 | } |
| 6799 | 6801 | ||
| 6800 | return em; | 6802 | return em; |
| 6801 | } | 6803 | } |
| 6802 | 6804 | ||
| 6803 | /* | 6805 | /* |
| 6804 | * returns 1 when the nocow is safe, < 1 on error, 0 if the | 6806 | * returns 1 when the nocow is safe, < 1 on error, 0 if the |
| 6805 | * block must be cow'd | 6807 | * block must be cow'd |
| 6806 | */ | 6808 | */ |
| 6807 | noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len, | 6809 | noinline int can_nocow_extent(struct inode *inode, u64 offset, u64 *len, |
| 6808 | u64 *orig_start, u64 *orig_block_len, | 6810 | u64 *orig_start, u64 *orig_block_len, |
| 6809 | u64 *ram_bytes) | 6811 | u64 *ram_bytes) |
| 6810 | { | 6812 | { |
| 6811 | struct btrfs_trans_handle *trans; | 6813 | struct btrfs_trans_handle *trans; |
| 6812 | struct btrfs_path *path; | 6814 | struct btrfs_path *path; |
| 6813 | int ret; | 6815 | int ret; |
| 6814 | struct extent_buffer *leaf; | 6816 | struct extent_buffer *leaf; |
| 6815 | struct btrfs_root *root = BTRFS_I(inode)->root; | 6817 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 6816 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 6818 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 6817 | struct btrfs_file_extent_item *fi; | 6819 | struct btrfs_file_extent_item *fi; |
| 6818 | struct btrfs_key key; | 6820 | struct btrfs_key key; |
| 6819 | u64 disk_bytenr; | 6821 | u64 disk_bytenr; |
| 6820 | u64 backref_offset; | 6822 | u64 backref_offset; |
| 6821 | u64 extent_end; | 6823 | u64 extent_end; |
| 6822 | u64 num_bytes; | 6824 | u64 num_bytes; |
| 6823 | int slot; | 6825 | int slot; |
| 6824 | int found_type; | 6826 | int found_type; |
| 6825 | bool nocow = (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW); | 6827 | bool nocow = (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW); |
| 6826 | 6828 | ||
| 6827 | path = btrfs_alloc_path(); | 6829 | path = btrfs_alloc_path(); |
| 6828 | if (!path) | 6830 | if (!path) |
| 6829 | return -ENOMEM; | 6831 | return -ENOMEM; |
| 6830 | 6832 | ||
| 6831 | ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(inode), | 6833 | ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(inode), |
| 6832 | offset, 0); | 6834 | offset, 0); |
| 6833 | if (ret < 0) | 6835 | if (ret < 0) |
| 6834 | goto out; | 6836 | goto out; |
| 6835 | 6837 | ||
| 6836 | slot = path->slots[0]; | 6838 | slot = path->slots[0]; |
| 6837 | if (ret == 1) { | 6839 | if (ret == 1) { |
| 6838 | if (slot == 0) { | 6840 | if (slot == 0) { |
| 6839 | /* can't find the item, must cow */ | 6841 | /* can't find the item, must cow */ |
| 6840 | ret = 0; | 6842 | ret = 0; |
| 6841 | goto out; | 6843 | goto out; |
| 6842 | } | 6844 | } |
| 6843 | slot--; | 6845 | slot--; |
| 6844 | } | 6846 | } |
| 6845 | ret = 0; | 6847 | ret = 0; |
| 6846 | leaf = path->nodes[0]; | 6848 | leaf = path->nodes[0]; |
| 6847 | btrfs_item_key_to_cpu(leaf, &key, slot); | 6849 | btrfs_item_key_to_cpu(leaf, &key, slot); |
| 6848 | if (key.objectid != btrfs_ino(inode) || | 6850 | if (key.objectid != btrfs_ino(inode) || |
| 6849 | key.type != BTRFS_EXTENT_DATA_KEY) { | 6851 | key.type != BTRFS_EXTENT_DATA_KEY) { |
| 6850 | /* not our file or wrong item type, must cow */ | 6852 | /* not our file or wrong item type, must cow */ |
| 6851 | goto out; | 6853 | goto out; |
| 6852 | } | 6854 | } |
| 6853 | 6855 | ||
| 6854 | if (key.offset > offset) { | 6856 | if (key.offset > offset) { |
| 6855 | /* Wrong offset, must cow */ | 6857 | /* Wrong offset, must cow */ |
| 6856 | goto out; | 6858 | goto out; |
| 6857 | } | 6859 | } |
| 6858 | 6860 | ||
| 6859 | fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item); | 6861 | fi = btrfs_item_ptr(leaf, slot, struct btrfs_file_extent_item); |
| 6860 | found_type = btrfs_file_extent_type(leaf, fi); | 6862 | found_type = btrfs_file_extent_type(leaf, fi); |
| 6861 | if (found_type != BTRFS_FILE_EXTENT_REG && | 6863 | if (found_type != BTRFS_FILE_EXTENT_REG && |
| 6862 | found_type != BTRFS_FILE_EXTENT_PREALLOC) { | 6864 | found_type != BTRFS_FILE_EXTENT_PREALLOC) { |
| 6863 | /* not a regular extent, must cow */ | 6865 | /* not a regular extent, must cow */ |
| 6864 | goto out; | 6866 | goto out; |
| 6865 | } | 6867 | } |
| 6866 | 6868 | ||
| 6867 | if (!nocow && found_type == BTRFS_FILE_EXTENT_REG) | 6869 | if (!nocow && found_type == BTRFS_FILE_EXTENT_REG) |
| 6868 | goto out; | 6870 | goto out; |
| 6869 | 6871 | ||
| 6870 | extent_end = key.offset + btrfs_file_extent_num_bytes(leaf, fi); | 6872 | extent_end = key.offset + btrfs_file_extent_num_bytes(leaf, fi); |
| 6871 | if (extent_end <= offset) | 6873 | if (extent_end <= offset) |
| 6872 | goto out; | 6874 | goto out; |
| 6873 | 6875 | ||
| 6874 | disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi); | 6876 | disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi); |
| 6875 | if (disk_bytenr == 0) | 6877 | if (disk_bytenr == 0) |
| 6876 | goto out; | 6878 | goto out; |
| 6877 | 6879 | ||
| 6878 | if (btrfs_file_extent_compression(leaf, fi) || | 6880 | if (btrfs_file_extent_compression(leaf, fi) || |
| 6879 | btrfs_file_extent_encryption(leaf, fi) || | 6881 | btrfs_file_extent_encryption(leaf, fi) || |
| 6880 | btrfs_file_extent_other_encoding(leaf, fi)) | 6882 | btrfs_file_extent_other_encoding(leaf, fi)) |
| 6881 | goto out; | 6883 | goto out; |
| 6882 | 6884 | ||
| 6883 | backref_offset = btrfs_file_extent_offset(leaf, fi); | 6885 | backref_offset = btrfs_file_extent_offset(leaf, fi); |
| 6884 | 6886 | ||
| 6885 | if (orig_start) { | 6887 | if (orig_start) { |
| 6886 | *orig_start = key.offset - backref_offset; | 6888 | *orig_start = key.offset - backref_offset; |
| 6887 | *orig_block_len = btrfs_file_extent_disk_num_bytes(leaf, fi); | 6889 | *orig_block_len = btrfs_file_extent_disk_num_bytes(leaf, fi); |
| 6888 | *ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi); | 6890 | *ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi); |
| 6889 | } | 6891 | } |
| 6890 | 6892 | ||
| 6891 | if (btrfs_extent_readonly(root, disk_bytenr)) | 6893 | if (btrfs_extent_readonly(root, disk_bytenr)) |
| 6892 | goto out; | 6894 | goto out; |
| 6893 | 6895 | ||
| 6894 | num_bytes = min(offset + *len, extent_end) - offset; | 6896 | num_bytes = min(offset + *len, extent_end) - offset; |
| 6895 | if (!nocow && found_type == BTRFS_FILE_EXTENT_PREALLOC) { | 6897 | if (!nocow && found_type == BTRFS_FILE_EXTENT_PREALLOC) { |
| 6896 | u64 range_end; | 6898 | u64 range_end; |
| 6897 | 6899 | ||
| 6898 | range_end = round_up(offset + num_bytes, root->sectorsize) - 1; | 6900 | range_end = round_up(offset + num_bytes, root->sectorsize) - 1; |
| 6899 | ret = test_range_bit(io_tree, offset, range_end, | 6901 | ret = test_range_bit(io_tree, offset, range_end, |
| 6900 | EXTENT_DELALLOC, 0, NULL); | 6902 | EXTENT_DELALLOC, 0, NULL); |
| 6901 | if (ret) { | 6903 | if (ret) { |
| 6902 | ret = -EAGAIN; | 6904 | ret = -EAGAIN; |
| 6903 | goto out; | 6905 | goto out; |
| 6904 | } | 6906 | } |
| 6905 | } | 6907 | } |
| 6906 | 6908 | ||
| 6907 | btrfs_release_path(path); | 6909 | btrfs_release_path(path); |
| 6908 | 6910 | ||
| 6909 | /* | 6911 | /* |
| 6910 | * look for other files referencing this extent, if we | 6912 | * look for other files referencing this extent, if we |
| 6911 | * find any we must cow | 6913 | * find any we must cow |
| 6912 | */ | 6914 | */ |
| 6913 | trans = btrfs_join_transaction(root); | 6915 | trans = btrfs_join_transaction(root); |
| 6914 | if (IS_ERR(trans)) { | 6916 | if (IS_ERR(trans)) { |
| 6915 | ret = 0; | 6917 | ret = 0; |
| 6916 | goto out; | 6918 | goto out; |
| 6917 | } | 6919 | } |
| 6918 | 6920 | ||
| 6919 | ret = btrfs_cross_ref_exist(trans, root, btrfs_ino(inode), | 6921 | ret = btrfs_cross_ref_exist(trans, root, btrfs_ino(inode), |
| 6920 | key.offset - backref_offset, disk_bytenr); | 6922 | key.offset - backref_offset, disk_bytenr); |
| 6921 | btrfs_end_transaction(trans, root); | 6923 | btrfs_end_transaction(trans, root); |
| 6922 | if (ret) { | 6924 | if (ret) { |
| 6923 | ret = 0; | 6925 | ret = 0; |
| 6924 | goto out; | 6926 | goto out; |
| 6925 | } | 6927 | } |
| 6926 | 6928 | ||
| 6927 | /* | 6929 | /* |
| 6928 | * adjust disk_bytenr and num_bytes to cover just the bytes | 6930 | * adjust disk_bytenr and num_bytes to cover just the bytes |
| 6929 | * in this extent we are about to write. If there | 6931 | * in this extent we are about to write. If there |
| 6930 | * are any csums in that range we have to cow in order | 6932 | * are any csums in that range we have to cow in order |
| 6931 | * to keep the csums correct | 6933 | * to keep the csums correct |
| 6932 | */ | 6934 | */ |
| 6933 | disk_bytenr += backref_offset; | 6935 | disk_bytenr += backref_offset; |
| 6934 | disk_bytenr += offset - key.offset; | 6936 | disk_bytenr += offset - key.offset; |
| 6935 | if (csum_exist_in_range(root, disk_bytenr, num_bytes)) | 6937 | if (csum_exist_in_range(root, disk_bytenr, num_bytes)) |
| 6936 | goto out; | 6938 | goto out; |
| 6937 | /* | 6939 | /* |
| 6938 | * all of the above have passed, it is safe to overwrite this extent | 6940 | * all of the above have passed, it is safe to overwrite this extent |
| 6939 | * without cow | 6941 | * without cow |
| 6940 | */ | 6942 | */ |
| 6941 | *len = num_bytes; | 6943 | *len = num_bytes; |
| 6942 | ret = 1; | 6944 | ret = 1; |
| 6943 | out: | 6945 | out: |
| 6944 | btrfs_free_path(path); | 6946 | btrfs_free_path(path); |
| 6945 | return ret; | 6947 | return ret; |
| 6946 | } | 6948 | } |
| 6947 | 6949 | ||
| 6948 | bool btrfs_page_exists_in_range(struct inode *inode, loff_t start, loff_t end) | 6950 | bool btrfs_page_exists_in_range(struct inode *inode, loff_t start, loff_t end) |
| 6949 | { | 6951 | { |
| 6950 | struct radix_tree_root *root = &inode->i_mapping->page_tree; | 6952 | struct radix_tree_root *root = &inode->i_mapping->page_tree; |
| 6951 | int found = false; | 6953 | int found = false; |
| 6952 | void **pagep = NULL; | 6954 | void **pagep = NULL; |
| 6953 | struct page *page = NULL; | 6955 | struct page *page = NULL; |
| 6954 | int start_idx; | 6956 | int start_idx; |
| 6955 | int end_idx; | 6957 | int end_idx; |
| 6956 | 6958 | ||
| 6957 | start_idx = start >> PAGE_CACHE_SHIFT; | 6959 | start_idx = start >> PAGE_CACHE_SHIFT; |
| 6958 | 6960 | ||
| 6959 | /* | 6961 | /* |
| 6960 | * end is the last byte in the last page. end == start is legal | 6962 | * end is the last byte in the last page. end == start is legal |
| 6961 | */ | 6963 | */ |
| 6962 | end_idx = end >> PAGE_CACHE_SHIFT; | 6964 | end_idx = end >> PAGE_CACHE_SHIFT; |
| 6963 | 6965 | ||
| 6964 | rcu_read_lock(); | 6966 | rcu_read_lock(); |
| 6965 | 6967 | ||
| 6966 | /* Most of the code in this while loop is lifted from | 6968 | /* Most of the code in this while loop is lifted from |
| 6967 | * find_get_page. It's been modified to begin searching from a | 6969 | * find_get_page. It's been modified to begin searching from a |
| 6968 | * page and return just the first page found in that range. If the | 6970 | * page and return just the first page found in that range. If the |
| 6969 | * found idx is less than or equal to the end idx then we know that | 6971 | * found idx is less than or equal to the end idx then we know that |
| 6970 | * a page exists. If no pages are found or if those pages are | 6972 | * a page exists. If no pages are found or if those pages are |
| 6971 | * outside of the range then we're fine (yay!) */ | 6973 | * outside of the range then we're fine (yay!) */ |
| 6972 | while (page == NULL && | 6974 | while (page == NULL && |
| 6973 | radix_tree_gang_lookup_slot(root, &pagep, NULL, start_idx, 1)) { | 6975 | radix_tree_gang_lookup_slot(root, &pagep, NULL, start_idx, 1)) { |
| 6974 | page = radix_tree_deref_slot(pagep); | 6976 | page = radix_tree_deref_slot(pagep); |
| 6975 | if (unlikely(!page)) | 6977 | if (unlikely(!page)) |
| 6976 | break; | 6978 | break; |
| 6977 | 6979 | ||
| 6978 | if (radix_tree_exception(page)) { | 6980 | if (radix_tree_exception(page)) { |
| 6979 | if (radix_tree_deref_retry(page)) { | 6981 | if (radix_tree_deref_retry(page)) { |
| 6980 | page = NULL; | 6982 | page = NULL; |
| 6981 | continue; | 6983 | continue; |
| 6982 | } | 6984 | } |
| 6983 | /* | 6985 | /* |
| 6984 | * Otherwise, shmem/tmpfs must be storing a swap entry | 6986 | * Otherwise, shmem/tmpfs must be storing a swap entry |
| 6985 | * here as an exceptional entry: so return it without | 6987 | * here as an exceptional entry: so return it without |
| 6986 | * attempting to raise page count. | 6988 | * attempting to raise page count. |
| 6987 | */ | 6989 | */ |
| 6988 | page = NULL; | 6990 | page = NULL; |
| 6989 | break; /* TODO: Is this relevant for this use case? */ | 6991 | break; /* TODO: Is this relevant for this use case? */ |
| 6990 | } | 6992 | } |
| 6991 | 6993 | ||
| 6992 | if (!page_cache_get_speculative(page)) { | 6994 | if (!page_cache_get_speculative(page)) { |
| 6993 | page = NULL; | 6995 | page = NULL; |
| 6994 | continue; | 6996 | continue; |
| 6995 | } | 6997 | } |
| 6996 | 6998 | ||
| 6997 | /* | 6999 | /* |
| 6998 | * Has the page moved? | 7000 | * Has the page moved? |
| 6999 | * This is part of the lockless pagecache protocol. See | 7001 | * This is part of the lockless pagecache protocol. See |
| 7000 | * include/linux/pagemap.h for details. | 7002 | * include/linux/pagemap.h for details. |
| 7001 | */ | 7003 | */ |
| 7002 | if (unlikely(page != *pagep)) { | 7004 | if (unlikely(page != *pagep)) { |
| 7003 | page_cache_release(page); | 7005 | page_cache_release(page); |
| 7004 | page = NULL; | 7006 | page = NULL; |
| 7005 | } | 7007 | } |
| 7006 | } | 7008 | } |
| 7007 | 7009 | ||
| 7008 | if (page) { | 7010 | if (page) { |
| 7009 | if (page->index <= end_idx) | 7011 | if (page->index <= end_idx) |
| 7010 | found = true; | 7012 | found = true; |
| 7011 | page_cache_release(page); | 7013 | page_cache_release(page); |
| 7012 | } | 7014 | } |
| 7013 | 7015 | ||
| 7014 | rcu_read_unlock(); | 7016 | rcu_read_unlock(); |
| 7015 | return found; | 7017 | return found; |
| 7016 | } | 7018 | } |
| 7017 | 7019 | ||
| 7018 | static int lock_extent_direct(struct inode *inode, u64 lockstart, u64 lockend, | 7020 | static int lock_extent_direct(struct inode *inode, u64 lockstart, u64 lockend, |
| 7019 | struct extent_state **cached_state, int writing) | 7021 | struct extent_state **cached_state, int writing) |
| 7020 | { | 7022 | { |
| 7021 | struct btrfs_ordered_extent *ordered; | 7023 | struct btrfs_ordered_extent *ordered; |
| 7022 | int ret = 0; | 7024 | int ret = 0; |
| 7023 | 7025 | ||
| 7024 | while (1) { | 7026 | while (1) { |
| 7025 | lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, | 7027 | lock_extent_bits(&BTRFS_I(inode)->io_tree, lockstart, lockend, |
| 7026 | 0, cached_state); | 7028 | 0, cached_state); |
| 7027 | /* | 7029 | /* |
| 7028 | * We're concerned with the entire range that we're going to be | 7030 | * We're concerned with the entire range that we're going to be |
| 7029 | * doing DIO to, so we need to make sure theres no ordered | 7031 | * doing DIO to, so we need to make sure theres no ordered |
| 7030 | * extents in this range. | 7032 | * extents in this range. |
| 7031 | */ | 7033 | */ |
| 7032 | ordered = btrfs_lookup_ordered_range(inode, lockstart, | 7034 | ordered = btrfs_lookup_ordered_range(inode, lockstart, |
| 7033 | lockend - lockstart + 1); | 7035 | lockend - lockstart + 1); |
| 7034 | 7036 | ||
| 7035 | /* | 7037 | /* |
| 7036 | * We need to make sure there are no buffered pages in this | 7038 | * We need to make sure there are no buffered pages in this |
| 7037 | * range either, we could have raced between the invalidate in | 7039 | * range either, we could have raced between the invalidate in |
| 7038 | * generic_file_direct_write and locking the extent. The | 7040 | * generic_file_direct_write and locking the extent. The |
| 7039 | * invalidate needs to happen so that reads after a write do not | 7041 | * invalidate needs to happen so that reads after a write do not |
| 7040 | * get stale data. | 7042 | * get stale data. |
| 7041 | */ | 7043 | */ |
| 7042 | if (!ordered && | 7044 | if (!ordered && |
| 7043 | (!writing || | 7045 | (!writing || |
| 7044 | !btrfs_page_exists_in_range(inode, lockstart, lockend))) | 7046 | !btrfs_page_exists_in_range(inode, lockstart, lockend))) |
| 7045 | break; | 7047 | break; |
| 7046 | 7048 | ||
| 7047 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend, | 7049 | unlock_extent_cached(&BTRFS_I(inode)->io_tree, lockstart, lockend, |
| 7048 | cached_state, GFP_NOFS); | 7050 | cached_state, GFP_NOFS); |
| 7049 | 7051 | ||
| 7050 | if (ordered) { | 7052 | if (ordered) { |
| 7051 | btrfs_start_ordered_extent(inode, ordered, 1); | 7053 | btrfs_start_ordered_extent(inode, ordered, 1); |
| 7052 | btrfs_put_ordered_extent(ordered); | 7054 | btrfs_put_ordered_extent(ordered); |
| 7053 | } else { | 7055 | } else { |
| 7054 | /* Screw you mmap */ | 7056 | /* Screw you mmap */ |
| 7055 | ret = btrfs_fdatawrite_range(inode, lockstart, lockend); | 7057 | ret = btrfs_fdatawrite_range(inode, lockstart, lockend); |
| 7056 | if (ret) | 7058 | if (ret) |
| 7057 | break; | 7059 | break; |
| 7058 | ret = filemap_fdatawait_range(inode->i_mapping, | 7060 | ret = filemap_fdatawait_range(inode->i_mapping, |
| 7059 | lockstart, | 7061 | lockstart, |
| 7060 | lockend); | 7062 | lockend); |
| 7061 | if (ret) | 7063 | if (ret) |
| 7062 | break; | 7064 | break; |
| 7063 | 7065 | ||
| 7064 | /* | 7066 | /* |
| 7065 | * If we found a page that couldn't be invalidated just | 7067 | * If we found a page that couldn't be invalidated just |
| 7066 | * fall back to buffered. | 7068 | * fall back to buffered. |
| 7067 | */ | 7069 | */ |
| 7068 | ret = invalidate_inode_pages2_range(inode->i_mapping, | 7070 | ret = invalidate_inode_pages2_range(inode->i_mapping, |
| 7069 | lockstart >> PAGE_CACHE_SHIFT, | 7071 | lockstart >> PAGE_CACHE_SHIFT, |
| 7070 | lockend >> PAGE_CACHE_SHIFT); | 7072 | lockend >> PAGE_CACHE_SHIFT); |
| 7071 | if (ret) | 7073 | if (ret) |
| 7072 | break; | 7074 | break; |
| 7073 | } | 7075 | } |
| 7074 | 7076 | ||
| 7075 | cond_resched(); | 7077 | cond_resched(); |
| 7076 | } | 7078 | } |
| 7077 | 7079 | ||
| 7078 | return ret; | 7080 | return ret; |
| 7079 | } | 7081 | } |
| 7080 | 7082 | ||
| 7081 | static struct extent_map *create_pinned_em(struct inode *inode, u64 start, | 7083 | static struct extent_map *create_pinned_em(struct inode *inode, u64 start, |
| 7082 | u64 len, u64 orig_start, | 7084 | u64 len, u64 orig_start, |
| 7083 | u64 block_start, u64 block_len, | 7085 | u64 block_start, u64 block_len, |
| 7084 | u64 orig_block_len, u64 ram_bytes, | 7086 | u64 orig_block_len, u64 ram_bytes, |
| 7085 | int type) | 7087 | int type) |
| 7086 | { | 7088 | { |
| 7087 | struct extent_map_tree *em_tree; | 7089 | struct extent_map_tree *em_tree; |
| 7088 | struct extent_map *em; | 7090 | struct extent_map *em; |
| 7089 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7091 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7090 | int ret; | 7092 | int ret; |
| 7091 | 7093 | ||
| 7092 | em_tree = &BTRFS_I(inode)->extent_tree; | 7094 | em_tree = &BTRFS_I(inode)->extent_tree; |
| 7093 | em = alloc_extent_map(); | 7095 | em = alloc_extent_map(); |
| 7094 | if (!em) | 7096 | if (!em) |
| 7095 | return ERR_PTR(-ENOMEM); | 7097 | return ERR_PTR(-ENOMEM); |
| 7096 | 7098 | ||
| 7097 | em->start = start; | 7099 | em->start = start; |
| 7098 | em->orig_start = orig_start; | 7100 | em->orig_start = orig_start; |
| 7099 | em->mod_start = start; | 7101 | em->mod_start = start; |
| 7100 | em->mod_len = len; | 7102 | em->mod_len = len; |
| 7101 | em->len = len; | 7103 | em->len = len; |
| 7102 | em->block_len = block_len; | 7104 | em->block_len = block_len; |
| 7103 | em->block_start = block_start; | 7105 | em->block_start = block_start; |
| 7104 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 7106 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 7105 | em->orig_block_len = orig_block_len; | 7107 | em->orig_block_len = orig_block_len; |
| 7106 | em->ram_bytes = ram_bytes; | 7108 | em->ram_bytes = ram_bytes; |
| 7107 | em->generation = -1; | 7109 | em->generation = -1; |
| 7108 | set_bit(EXTENT_FLAG_PINNED, &em->flags); | 7110 | set_bit(EXTENT_FLAG_PINNED, &em->flags); |
| 7109 | if (type == BTRFS_ORDERED_PREALLOC) | 7111 | if (type == BTRFS_ORDERED_PREALLOC) |
| 7110 | set_bit(EXTENT_FLAG_FILLING, &em->flags); | 7112 | set_bit(EXTENT_FLAG_FILLING, &em->flags); |
| 7111 | 7113 | ||
| 7112 | do { | 7114 | do { |
| 7113 | btrfs_drop_extent_cache(inode, em->start, | 7115 | btrfs_drop_extent_cache(inode, em->start, |
| 7114 | em->start + em->len - 1, 0); | 7116 | em->start + em->len - 1, 0); |
| 7115 | write_lock(&em_tree->lock); | 7117 | write_lock(&em_tree->lock); |
| 7116 | ret = add_extent_mapping(em_tree, em, 1); | 7118 | ret = add_extent_mapping(em_tree, em, 1); |
| 7117 | write_unlock(&em_tree->lock); | 7119 | write_unlock(&em_tree->lock); |
| 7118 | } while (ret == -EEXIST); | 7120 | } while (ret == -EEXIST); |
| 7119 | 7121 | ||
| 7120 | if (ret) { | 7122 | if (ret) { |
| 7121 | free_extent_map(em); | 7123 | free_extent_map(em); |
| 7122 | return ERR_PTR(ret); | 7124 | return ERR_PTR(ret); |
| 7123 | } | 7125 | } |
| 7124 | 7126 | ||
| 7125 | return em; | 7127 | return em; |
| 7126 | } | 7128 | } |
| 7127 | 7129 | ||
| 7128 | 7130 | ||
| 7129 | static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, | 7131 | static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, |
| 7130 | struct buffer_head *bh_result, int create) | 7132 | struct buffer_head *bh_result, int create) |
| 7131 | { | 7133 | { |
| 7132 | struct extent_map *em; | 7134 | struct extent_map *em; |
| 7133 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7135 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7134 | struct extent_state *cached_state = NULL; | 7136 | struct extent_state *cached_state = NULL; |
| 7135 | u64 start = iblock << inode->i_blkbits; | 7137 | u64 start = iblock << inode->i_blkbits; |
| 7136 | u64 lockstart, lockend; | 7138 | u64 lockstart, lockend; |
| 7137 | u64 len = bh_result->b_size; | 7139 | u64 len = bh_result->b_size; |
| 7138 | int unlock_bits = EXTENT_LOCKED; | 7140 | int unlock_bits = EXTENT_LOCKED; |
| 7139 | int ret = 0; | 7141 | int ret = 0; |
| 7140 | 7142 | ||
| 7141 | if (create) | 7143 | if (create) |
| 7142 | unlock_bits |= EXTENT_DELALLOC | EXTENT_DIRTY; | 7144 | unlock_bits |= EXTENT_DELALLOC | EXTENT_DIRTY; |
| 7143 | else | 7145 | else |
| 7144 | len = min_t(u64, len, root->sectorsize); | 7146 | len = min_t(u64, len, root->sectorsize); |
| 7145 | 7147 | ||
| 7146 | lockstart = start; | 7148 | lockstart = start; |
| 7147 | lockend = start + len - 1; | 7149 | lockend = start + len - 1; |
| 7148 | 7150 | ||
| 7149 | /* | 7151 | /* |
| 7150 | * If this errors out it's because we couldn't invalidate pagecache for | 7152 | * If this errors out it's because we couldn't invalidate pagecache for |
| 7151 | * this range and we need to fallback to buffered. | 7153 | * this range and we need to fallback to buffered. |
| 7152 | */ | 7154 | */ |
| 7153 | if (lock_extent_direct(inode, lockstart, lockend, &cached_state, create)) | 7155 | if (lock_extent_direct(inode, lockstart, lockend, &cached_state, create)) |
| 7154 | return -ENOTBLK; | 7156 | return -ENOTBLK; |
| 7155 | 7157 | ||
| 7156 | em = btrfs_get_extent(inode, NULL, 0, start, len, 0); | 7158 | em = btrfs_get_extent(inode, NULL, 0, start, len, 0); |
| 7157 | if (IS_ERR(em)) { | 7159 | if (IS_ERR(em)) { |
| 7158 | ret = PTR_ERR(em); | 7160 | ret = PTR_ERR(em); |
| 7159 | goto unlock_err; | 7161 | goto unlock_err; |
| 7160 | } | 7162 | } |
| 7161 | 7163 | ||
| 7162 | /* | 7164 | /* |
| 7163 | * Ok for INLINE and COMPRESSED extents we need to fallback on buffered | 7165 | * Ok for INLINE and COMPRESSED extents we need to fallback on buffered |
| 7164 | * io. INLINE is special, and we could probably kludge it in here, but | 7166 | * io. INLINE is special, and we could probably kludge it in here, but |
| 7165 | * it's still buffered so for safety lets just fall back to the generic | 7167 | * it's still buffered so for safety lets just fall back to the generic |
| 7166 | * buffered path. | 7168 | * buffered path. |
| 7167 | * | 7169 | * |
| 7168 | * For COMPRESSED we _have_ to read the entire extent in so we can | 7170 | * For COMPRESSED we _have_ to read the entire extent in so we can |
| 7169 | * decompress it, so there will be buffering required no matter what we | 7171 | * decompress it, so there will be buffering required no matter what we |
| 7170 | * do, so go ahead and fallback to buffered. | 7172 | * do, so go ahead and fallback to buffered. |
| 7171 | * | 7173 | * |
| 7172 | * We return -ENOTBLK because thats what makes DIO go ahead and go back | 7174 | * We return -ENOTBLK because thats what makes DIO go ahead and go back |
| 7173 | * to buffered IO. Don't blame me, this is the price we pay for using | 7175 | * to buffered IO. Don't blame me, this is the price we pay for using |
| 7174 | * the generic code. | 7176 | * the generic code. |
| 7175 | */ | 7177 | */ |
| 7176 | if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) || | 7178 | if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) || |
| 7177 | em->block_start == EXTENT_MAP_INLINE) { | 7179 | em->block_start == EXTENT_MAP_INLINE) { |
| 7178 | free_extent_map(em); | 7180 | free_extent_map(em); |
| 7179 | ret = -ENOTBLK; | 7181 | ret = -ENOTBLK; |
| 7180 | goto unlock_err; | 7182 | goto unlock_err; |
| 7181 | } | 7183 | } |
| 7182 | 7184 | ||
| 7183 | /* Just a good old fashioned hole, return */ | 7185 | /* Just a good old fashioned hole, return */ |
| 7184 | if (!create && (em->block_start == EXTENT_MAP_HOLE || | 7186 | if (!create && (em->block_start == EXTENT_MAP_HOLE || |
| 7185 | test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) { | 7187 | test_bit(EXTENT_FLAG_PREALLOC, &em->flags))) { |
| 7186 | free_extent_map(em); | 7188 | free_extent_map(em); |
| 7187 | goto unlock_err; | 7189 | goto unlock_err; |
| 7188 | } | 7190 | } |
| 7189 | 7191 | ||
| 7190 | /* | 7192 | /* |
| 7191 | * We don't allocate a new extent in the following cases | 7193 | * We don't allocate a new extent in the following cases |
| 7192 | * | 7194 | * |
| 7193 | * 1) The inode is marked as NODATACOW. In this case we'll just use the | 7195 | * 1) The inode is marked as NODATACOW. In this case we'll just use the |
| 7194 | * existing extent. | 7196 | * existing extent. |
| 7195 | * 2) The extent is marked as PREALLOC. We're good to go here and can | 7197 | * 2) The extent is marked as PREALLOC. We're good to go here and can |
| 7196 | * just use the extent. | 7198 | * just use the extent. |
| 7197 | * | 7199 | * |
| 7198 | */ | 7200 | */ |
| 7199 | if (!create) { | 7201 | if (!create) { |
| 7200 | len = min(len, em->len - (start - em->start)); | 7202 | len = min(len, em->len - (start - em->start)); |
| 7201 | lockstart = start + len; | 7203 | lockstart = start + len; |
| 7202 | goto unlock; | 7204 | goto unlock; |
| 7203 | } | 7205 | } |
| 7204 | 7206 | ||
| 7205 | if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) || | 7207 | if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) || |
| 7206 | ((BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) && | 7208 | ((BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW) && |
| 7207 | em->block_start != EXTENT_MAP_HOLE)) { | 7209 | em->block_start != EXTENT_MAP_HOLE)) { |
| 7208 | int type; | 7210 | int type; |
| 7209 | int ret; | 7211 | int ret; |
| 7210 | u64 block_start, orig_start, orig_block_len, ram_bytes; | 7212 | u64 block_start, orig_start, orig_block_len, ram_bytes; |
| 7211 | 7213 | ||
| 7212 | if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) | 7214 | if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) |
| 7213 | type = BTRFS_ORDERED_PREALLOC; | 7215 | type = BTRFS_ORDERED_PREALLOC; |
| 7214 | else | 7216 | else |
| 7215 | type = BTRFS_ORDERED_NOCOW; | 7217 | type = BTRFS_ORDERED_NOCOW; |
| 7216 | len = min(len, em->len - (start - em->start)); | 7218 | len = min(len, em->len - (start - em->start)); |
| 7217 | block_start = em->block_start + (start - em->start); | 7219 | block_start = em->block_start + (start - em->start); |
| 7218 | 7220 | ||
| 7219 | if (can_nocow_extent(inode, start, &len, &orig_start, | 7221 | if (can_nocow_extent(inode, start, &len, &orig_start, |
| 7220 | &orig_block_len, &ram_bytes) == 1) { | 7222 | &orig_block_len, &ram_bytes) == 1) { |
| 7221 | if (type == BTRFS_ORDERED_PREALLOC) { | 7223 | if (type == BTRFS_ORDERED_PREALLOC) { |
| 7222 | free_extent_map(em); | 7224 | free_extent_map(em); |
| 7223 | em = create_pinned_em(inode, start, len, | 7225 | em = create_pinned_em(inode, start, len, |
| 7224 | orig_start, | 7226 | orig_start, |
| 7225 | block_start, len, | 7227 | block_start, len, |
| 7226 | orig_block_len, | 7228 | orig_block_len, |
| 7227 | ram_bytes, type); | 7229 | ram_bytes, type); |
| 7228 | if (IS_ERR(em)) { | 7230 | if (IS_ERR(em)) { |
| 7229 | ret = PTR_ERR(em); | 7231 | ret = PTR_ERR(em); |
| 7230 | goto unlock_err; | 7232 | goto unlock_err; |
| 7231 | } | 7233 | } |
| 7232 | } | 7234 | } |
| 7233 | 7235 | ||
| 7234 | ret = btrfs_add_ordered_extent_dio(inode, start, | 7236 | ret = btrfs_add_ordered_extent_dio(inode, start, |
| 7235 | block_start, len, len, type); | 7237 | block_start, len, len, type); |
| 7236 | if (ret) { | 7238 | if (ret) { |
| 7237 | free_extent_map(em); | 7239 | free_extent_map(em); |
| 7238 | goto unlock_err; | 7240 | goto unlock_err; |
| 7239 | } | 7241 | } |
| 7240 | goto unlock; | 7242 | goto unlock; |
| 7241 | } | 7243 | } |
| 7242 | } | 7244 | } |
| 7243 | 7245 | ||
| 7244 | /* | 7246 | /* |
| 7245 | * this will cow the extent, reset the len in case we changed | 7247 | * this will cow the extent, reset the len in case we changed |
| 7246 | * it above | 7248 | * it above |
| 7247 | */ | 7249 | */ |
| 7248 | len = bh_result->b_size; | 7250 | len = bh_result->b_size; |
| 7249 | free_extent_map(em); | 7251 | free_extent_map(em); |
| 7250 | em = btrfs_new_extent_direct(inode, start, len); | 7252 | em = btrfs_new_extent_direct(inode, start, len); |
| 7251 | if (IS_ERR(em)) { | 7253 | if (IS_ERR(em)) { |
| 7252 | ret = PTR_ERR(em); | 7254 | ret = PTR_ERR(em); |
| 7253 | goto unlock_err; | 7255 | goto unlock_err; |
| 7254 | } | 7256 | } |
| 7255 | len = min(len, em->len - (start - em->start)); | 7257 | len = min(len, em->len - (start - em->start)); |
| 7256 | unlock: | 7258 | unlock: |
| 7257 | bh_result->b_blocknr = (em->block_start + (start - em->start)) >> | 7259 | bh_result->b_blocknr = (em->block_start + (start - em->start)) >> |
| 7258 | inode->i_blkbits; | 7260 | inode->i_blkbits; |
| 7259 | bh_result->b_size = len; | 7261 | bh_result->b_size = len; |
| 7260 | bh_result->b_bdev = em->bdev; | 7262 | bh_result->b_bdev = em->bdev; |
| 7261 | set_buffer_mapped(bh_result); | 7263 | set_buffer_mapped(bh_result); |
| 7262 | if (create) { | 7264 | if (create) { |
| 7263 | if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) | 7265 | if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) |
| 7264 | set_buffer_new(bh_result); | 7266 | set_buffer_new(bh_result); |
| 7265 | 7267 | ||
| 7266 | /* | 7268 | /* |
| 7267 | * Need to update the i_size under the extent lock so buffered | 7269 | * Need to update the i_size under the extent lock so buffered |
| 7268 | * readers will get the updated i_size when we unlock. | 7270 | * readers will get the updated i_size when we unlock. |
| 7269 | */ | 7271 | */ |
| 7270 | if (start + len > i_size_read(inode)) | 7272 | if (start + len > i_size_read(inode)) |
| 7271 | i_size_write(inode, start + len); | 7273 | i_size_write(inode, start + len); |
| 7272 | 7274 | ||
| 7273 | spin_lock(&BTRFS_I(inode)->lock); | 7275 | spin_lock(&BTRFS_I(inode)->lock); |
| 7274 | BTRFS_I(inode)->outstanding_extents++; | 7276 | BTRFS_I(inode)->outstanding_extents++; |
| 7275 | spin_unlock(&BTRFS_I(inode)->lock); | 7277 | spin_unlock(&BTRFS_I(inode)->lock); |
| 7276 | 7278 | ||
| 7277 | ret = set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, | 7279 | ret = set_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, |
| 7278 | lockstart + len - 1, EXTENT_DELALLOC, NULL, | 7280 | lockstart + len - 1, EXTENT_DELALLOC, NULL, |
| 7279 | &cached_state, GFP_NOFS); | 7281 | &cached_state, GFP_NOFS); |
| 7280 | BUG_ON(ret); | 7282 | BUG_ON(ret); |
| 7281 | } | 7283 | } |
| 7282 | 7284 | ||
| 7283 | /* | 7285 | /* |
| 7284 | * In the case of write we need to clear and unlock the entire range, | 7286 | * In the case of write we need to clear and unlock the entire range, |
| 7285 | * in the case of read we need to unlock only the end area that we | 7287 | * in the case of read we need to unlock only the end area that we |
| 7286 | * aren't using if there is any left over space. | 7288 | * aren't using if there is any left over space. |
| 7287 | */ | 7289 | */ |
| 7288 | if (lockstart < lockend) { | 7290 | if (lockstart < lockend) { |
| 7289 | clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, | 7291 | clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, |
| 7290 | lockend, unlock_bits, 1, 0, | 7292 | lockend, unlock_bits, 1, 0, |
| 7291 | &cached_state, GFP_NOFS); | 7293 | &cached_state, GFP_NOFS); |
| 7292 | } else { | 7294 | } else { |
| 7293 | free_extent_state(cached_state); | 7295 | free_extent_state(cached_state); |
| 7294 | } | 7296 | } |
| 7295 | 7297 | ||
| 7296 | free_extent_map(em); | 7298 | free_extent_map(em); |
| 7297 | 7299 | ||
| 7298 | return 0; | 7300 | return 0; |
| 7299 | 7301 | ||
| 7300 | unlock_err: | 7302 | unlock_err: |
| 7301 | clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, | 7303 | clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend, |
| 7302 | unlock_bits, 1, 0, &cached_state, GFP_NOFS); | 7304 | unlock_bits, 1, 0, &cached_state, GFP_NOFS); |
| 7303 | return ret; | 7305 | return ret; |
| 7304 | } | 7306 | } |
| 7305 | 7307 | ||
| 7306 | static inline int submit_dio_repair_bio(struct inode *inode, struct bio *bio, | 7308 | static inline int submit_dio_repair_bio(struct inode *inode, struct bio *bio, |
| 7307 | int rw, int mirror_num) | 7309 | int rw, int mirror_num) |
| 7308 | { | 7310 | { |
| 7309 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7311 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7310 | int ret; | 7312 | int ret; |
| 7311 | 7313 | ||
| 7312 | BUG_ON(rw & REQ_WRITE); | 7314 | BUG_ON(rw & REQ_WRITE); |
| 7313 | 7315 | ||
| 7314 | bio_get(bio); | 7316 | bio_get(bio); |
| 7315 | 7317 | ||
| 7316 | ret = btrfs_bio_wq_end_io(root->fs_info, bio, | 7318 | ret = btrfs_bio_wq_end_io(root->fs_info, bio, |
| 7317 | BTRFS_WQ_ENDIO_DIO_REPAIR); | 7319 | BTRFS_WQ_ENDIO_DIO_REPAIR); |
| 7318 | if (ret) | 7320 | if (ret) |
| 7319 | goto err; | 7321 | goto err; |
| 7320 | 7322 | ||
| 7321 | ret = btrfs_map_bio(root, rw, bio, mirror_num, 0); | 7323 | ret = btrfs_map_bio(root, rw, bio, mirror_num, 0); |
| 7322 | err: | 7324 | err: |
| 7323 | bio_put(bio); | 7325 | bio_put(bio); |
| 7324 | return ret; | 7326 | return ret; |
| 7325 | } | 7327 | } |
| 7326 | 7328 | ||
| 7327 | static int btrfs_check_dio_repairable(struct inode *inode, | 7329 | static int btrfs_check_dio_repairable(struct inode *inode, |
| 7328 | struct bio *failed_bio, | 7330 | struct bio *failed_bio, |
| 7329 | struct io_failure_record *failrec, | 7331 | struct io_failure_record *failrec, |
| 7330 | int failed_mirror) | 7332 | int failed_mirror) |
| 7331 | { | 7333 | { |
| 7332 | int num_copies; | 7334 | int num_copies; |
| 7333 | 7335 | ||
| 7334 | num_copies = btrfs_num_copies(BTRFS_I(inode)->root->fs_info, | 7336 | num_copies = btrfs_num_copies(BTRFS_I(inode)->root->fs_info, |
| 7335 | failrec->logical, failrec->len); | 7337 | failrec->logical, failrec->len); |
| 7336 | if (num_copies == 1) { | 7338 | if (num_copies == 1) { |
| 7337 | /* | 7339 | /* |
| 7338 | * we only have a single copy of the data, so don't bother with | 7340 | * we only have a single copy of the data, so don't bother with |
| 7339 | * all the retry and error correction code that follows. no | 7341 | * all the retry and error correction code that follows. no |
| 7340 | * matter what the error is, it is very likely to persist. | 7342 | * matter what the error is, it is very likely to persist. |
| 7341 | */ | 7343 | */ |
| 7342 | pr_debug("Check DIO Repairable: cannot repair, num_copies=%d, next_mirror %d, failed_mirror %d\n", | 7344 | pr_debug("Check DIO Repairable: cannot repair, num_copies=%d, next_mirror %d, failed_mirror %d\n", |
| 7343 | num_copies, failrec->this_mirror, failed_mirror); | 7345 | num_copies, failrec->this_mirror, failed_mirror); |
| 7344 | return 0; | 7346 | return 0; |
| 7345 | } | 7347 | } |
| 7346 | 7348 | ||
| 7347 | failrec->failed_mirror = failed_mirror; | 7349 | failrec->failed_mirror = failed_mirror; |
| 7348 | failrec->this_mirror++; | 7350 | failrec->this_mirror++; |
| 7349 | if (failrec->this_mirror == failed_mirror) | 7351 | if (failrec->this_mirror == failed_mirror) |
| 7350 | failrec->this_mirror++; | 7352 | failrec->this_mirror++; |
| 7351 | 7353 | ||
| 7352 | if (failrec->this_mirror > num_copies) { | 7354 | if (failrec->this_mirror > num_copies) { |
| 7353 | pr_debug("Check DIO Repairable: (fail) num_copies=%d, next_mirror %d, failed_mirror %d\n", | 7355 | pr_debug("Check DIO Repairable: (fail) num_copies=%d, next_mirror %d, failed_mirror %d\n", |
| 7354 | num_copies, failrec->this_mirror, failed_mirror); | 7356 | num_copies, failrec->this_mirror, failed_mirror); |
| 7355 | return 0; | 7357 | return 0; |
| 7356 | } | 7358 | } |
| 7357 | 7359 | ||
| 7358 | return 1; | 7360 | return 1; |
| 7359 | } | 7361 | } |
| 7360 | 7362 | ||
| 7361 | static int dio_read_error(struct inode *inode, struct bio *failed_bio, | 7363 | static int dio_read_error(struct inode *inode, struct bio *failed_bio, |
| 7362 | struct page *page, u64 start, u64 end, | 7364 | struct page *page, u64 start, u64 end, |
| 7363 | int failed_mirror, bio_end_io_t *repair_endio, | 7365 | int failed_mirror, bio_end_io_t *repair_endio, |
| 7364 | void *repair_arg) | 7366 | void *repair_arg) |
| 7365 | { | 7367 | { |
| 7366 | struct io_failure_record *failrec; | 7368 | struct io_failure_record *failrec; |
| 7367 | struct bio *bio; | 7369 | struct bio *bio; |
| 7368 | int isector; | 7370 | int isector; |
| 7369 | int read_mode; | 7371 | int read_mode; |
| 7370 | int ret; | 7372 | int ret; |
| 7371 | 7373 | ||
| 7372 | BUG_ON(failed_bio->bi_rw & REQ_WRITE); | 7374 | BUG_ON(failed_bio->bi_rw & REQ_WRITE); |
| 7373 | 7375 | ||
| 7374 | ret = btrfs_get_io_failure_record(inode, start, end, &failrec); | 7376 | ret = btrfs_get_io_failure_record(inode, start, end, &failrec); |
| 7375 | if (ret) | 7377 | if (ret) |
| 7376 | return ret; | 7378 | return ret; |
| 7377 | 7379 | ||
| 7378 | ret = btrfs_check_dio_repairable(inode, failed_bio, failrec, | 7380 | ret = btrfs_check_dio_repairable(inode, failed_bio, failrec, |
| 7379 | failed_mirror); | 7381 | failed_mirror); |
| 7380 | if (!ret) { | 7382 | if (!ret) { |
| 7381 | free_io_failure(inode, failrec); | 7383 | free_io_failure(inode, failrec); |
| 7382 | return -EIO; | 7384 | return -EIO; |
| 7383 | } | 7385 | } |
| 7384 | 7386 | ||
| 7385 | if (failed_bio->bi_vcnt > 1) | 7387 | if (failed_bio->bi_vcnt > 1) |
| 7386 | read_mode = READ_SYNC | REQ_FAILFAST_DEV; | 7388 | read_mode = READ_SYNC | REQ_FAILFAST_DEV; |
| 7387 | else | 7389 | else |
| 7388 | read_mode = READ_SYNC; | 7390 | read_mode = READ_SYNC; |
| 7389 | 7391 | ||
| 7390 | isector = start - btrfs_io_bio(failed_bio)->logical; | 7392 | isector = start - btrfs_io_bio(failed_bio)->logical; |
| 7391 | isector >>= inode->i_sb->s_blocksize_bits; | 7393 | isector >>= inode->i_sb->s_blocksize_bits; |
| 7392 | bio = btrfs_create_repair_bio(inode, failed_bio, failrec, page, | 7394 | bio = btrfs_create_repair_bio(inode, failed_bio, failrec, page, |
| 7393 | 0, isector, repair_endio, repair_arg); | 7395 | 0, isector, repair_endio, repair_arg); |
| 7394 | if (!bio) { | 7396 | if (!bio) { |
| 7395 | free_io_failure(inode, failrec); | 7397 | free_io_failure(inode, failrec); |
| 7396 | return -EIO; | 7398 | return -EIO; |
| 7397 | } | 7399 | } |
| 7398 | 7400 | ||
| 7399 | btrfs_debug(BTRFS_I(inode)->root->fs_info, | 7401 | btrfs_debug(BTRFS_I(inode)->root->fs_info, |
| 7400 | "Repair DIO Read Error: submitting new dio read[%#x] to this_mirror=%d, in_validation=%d\n", | 7402 | "Repair DIO Read Error: submitting new dio read[%#x] to this_mirror=%d, in_validation=%d\n", |
| 7401 | read_mode, failrec->this_mirror, failrec->in_validation); | 7403 | read_mode, failrec->this_mirror, failrec->in_validation); |
| 7402 | 7404 | ||
| 7403 | ret = submit_dio_repair_bio(inode, bio, read_mode, | 7405 | ret = submit_dio_repair_bio(inode, bio, read_mode, |
| 7404 | failrec->this_mirror); | 7406 | failrec->this_mirror); |
| 7405 | if (ret) { | 7407 | if (ret) { |
| 7406 | free_io_failure(inode, failrec); | 7408 | free_io_failure(inode, failrec); |
| 7407 | bio_put(bio); | 7409 | bio_put(bio); |
| 7408 | } | 7410 | } |
| 7409 | 7411 | ||
| 7410 | return ret; | 7412 | return ret; |
| 7411 | } | 7413 | } |
| 7412 | 7414 | ||
| 7413 | struct btrfs_retry_complete { | 7415 | struct btrfs_retry_complete { |
| 7414 | struct completion done; | 7416 | struct completion done; |
| 7415 | struct inode *inode; | 7417 | struct inode *inode; |
| 7416 | u64 start; | 7418 | u64 start; |
| 7417 | int uptodate; | 7419 | int uptodate; |
| 7418 | }; | 7420 | }; |
| 7419 | 7421 | ||
| 7420 | static void btrfs_retry_endio_nocsum(struct bio *bio, int err) | 7422 | static void btrfs_retry_endio_nocsum(struct bio *bio, int err) |
| 7421 | { | 7423 | { |
| 7422 | struct btrfs_retry_complete *done = bio->bi_private; | 7424 | struct btrfs_retry_complete *done = bio->bi_private; |
| 7423 | struct bio_vec *bvec; | 7425 | struct bio_vec *bvec; |
| 7424 | int i; | 7426 | int i; |
| 7425 | 7427 | ||
| 7426 | if (err) | 7428 | if (err) |
| 7427 | goto end; | 7429 | goto end; |
| 7428 | 7430 | ||
| 7429 | done->uptodate = 1; | 7431 | done->uptodate = 1; |
| 7430 | bio_for_each_segment_all(bvec, bio, i) | 7432 | bio_for_each_segment_all(bvec, bio, i) |
| 7431 | clean_io_failure(done->inode, done->start, bvec->bv_page, 0); | 7433 | clean_io_failure(done->inode, done->start, bvec->bv_page, 0); |
| 7432 | end: | 7434 | end: |
| 7433 | complete(&done->done); | 7435 | complete(&done->done); |
| 7434 | bio_put(bio); | 7436 | bio_put(bio); |
| 7435 | } | 7437 | } |
| 7436 | 7438 | ||
| 7437 | static int __btrfs_correct_data_nocsum(struct inode *inode, | 7439 | static int __btrfs_correct_data_nocsum(struct inode *inode, |
| 7438 | struct btrfs_io_bio *io_bio) | 7440 | struct btrfs_io_bio *io_bio) |
| 7439 | { | 7441 | { |
| 7440 | struct bio_vec *bvec; | 7442 | struct bio_vec *bvec; |
| 7441 | struct btrfs_retry_complete done; | 7443 | struct btrfs_retry_complete done; |
| 7442 | u64 start; | 7444 | u64 start; |
| 7443 | int i; | 7445 | int i; |
| 7444 | int ret; | 7446 | int ret; |
| 7445 | 7447 | ||
| 7446 | start = io_bio->logical; | 7448 | start = io_bio->logical; |
| 7447 | done.inode = inode; | 7449 | done.inode = inode; |
| 7448 | 7450 | ||
| 7449 | bio_for_each_segment_all(bvec, &io_bio->bio, i) { | 7451 | bio_for_each_segment_all(bvec, &io_bio->bio, i) { |
| 7450 | try_again: | 7452 | try_again: |
| 7451 | done.uptodate = 0; | 7453 | done.uptodate = 0; |
| 7452 | done.start = start; | 7454 | done.start = start; |
| 7453 | init_completion(&done.done); | 7455 | init_completion(&done.done); |
| 7454 | 7456 | ||
| 7455 | ret = dio_read_error(inode, &io_bio->bio, bvec->bv_page, start, | 7457 | ret = dio_read_error(inode, &io_bio->bio, bvec->bv_page, start, |
| 7456 | start + bvec->bv_len - 1, | 7458 | start + bvec->bv_len - 1, |
| 7457 | io_bio->mirror_num, | 7459 | io_bio->mirror_num, |
| 7458 | btrfs_retry_endio_nocsum, &done); | 7460 | btrfs_retry_endio_nocsum, &done); |
| 7459 | if (ret) | 7461 | if (ret) |
| 7460 | return ret; | 7462 | return ret; |
| 7461 | 7463 | ||
| 7462 | wait_for_completion(&done.done); | 7464 | wait_for_completion(&done.done); |
| 7463 | 7465 | ||
| 7464 | if (!done.uptodate) { | 7466 | if (!done.uptodate) { |
| 7465 | /* We might have another mirror, so try again */ | 7467 | /* We might have another mirror, so try again */ |
| 7466 | goto try_again; | 7468 | goto try_again; |
| 7467 | } | 7469 | } |
| 7468 | 7470 | ||
| 7469 | start += bvec->bv_len; | 7471 | start += bvec->bv_len; |
| 7470 | } | 7472 | } |
| 7471 | 7473 | ||
| 7472 | return 0; | 7474 | return 0; |
| 7473 | } | 7475 | } |
| 7474 | 7476 | ||
| 7475 | static void btrfs_retry_endio(struct bio *bio, int err) | 7477 | static void btrfs_retry_endio(struct bio *bio, int err) |
| 7476 | { | 7478 | { |
| 7477 | struct btrfs_retry_complete *done = bio->bi_private; | 7479 | struct btrfs_retry_complete *done = bio->bi_private; |
| 7478 | struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); | 7480 | struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); |
| 7479 | struct bio_vec *bvec; | 7481 | struct bio_vec *bvec; |
| 7480 | int uptodate; | 7482 | int uptodate; |
| 7481 | int ret; | 7483 | int ret; |
| 7482 | int i; | 7484 | int i; |
| 7483 | 7485 | ||
| 7484 | if (err) | 7486 | if (err) |
| 7485 | goto end; | 7487 | goto end; |
| 7486 | 7488 | ||
| 7487 | uptodate = 1; | 7489 | uptodate = 1; |
| 7488 | bio_for_each_segment_all(bvec, bio, i) { | 7490 | bio_for_each_segment_all(bvec, bio, i) { |
| 7489 | ret = __readpage_endio_check(done->inode, io_bio, i, | 7491 | ret = __readpage_endio_check(done->inode, io_bio, i, |
| 7490 | bvec->bv_page, 0, | 7492 | bvec->bv_page, 0, |
| 7491 | done->start, bvec->bv_len); | 7493 | done->start, bvec->bv_len); |
| 7492 | if (!ret) | 7494 | if (!ret) |
| 7493 | clean_io_failure(done->inode, done->start, | 7495 | clean_io_failure(done->inode, done->start, |
| 7494 | bvec->bv_page, 0); | 7496 | bvec->bv_page, 0); |
| 7495 | else | 7497 | else |
| 7496 | uptodate = 0; | 7498 | uptodate = 0; |
| 7497 | } | 7499 | } |
| 7498 | 7500 | ||
| 7499 | done->uptodate = uptodate; | 7501 | done->uptodate = uptodate; |
| 7500 | end: | 7502 | end: |
| 7501 | complete(&done->done); | 7503 | complete(&done->done); |
| 7502 | bio_put(bio); | 7504 | bio_put(bio); |
| 7503 | } | 7505 | } |
| 7504 | 7506 | ||
| 7505 | static int __btrfs_subio_endio_read(struct inode *inode, | 7507 | static int __btrfs_subio_endio_read(struct inode *inode, |
| 7506 | struct btrfs_io_bio *io_bio, int err) | 7508 | struct btrfs_io_bio *io_bio, int err) |
| 7507 | { | 7509 | { |
| 7508 | struct bio_vec *bvec; | 7510 | struct bio_vec *bvec; |
| 7509 | struct btrfs_retry_complete done; | 7511 | struct btrfs_retry_complete done; |
| 7510 | u64 start; | 7512 | u64 start; |
| 7511 | u64 offset = 0; | 7513 | u64 offset = 0; |
| 7512 | int i; | 7514 | int i; |
| 7513 | int ret; | 7515 | int ret; |
| 7514 | 7516 | ||
| 7515 | err = 0; | 7517 | err = 0; |
| 7516 | start = io_bio->logical; | 7518 | start = io_bio->logical; |
| 7517 | done.inode = inode; | 7519 | done.inode = inode; |
| 7518 | 7520 | ||
| 7519 | bio_for_each_segment_all(bvec, &io_bio->bio, i) { | 7521 | bio_for_each_segment_all(bvec, &io_bio->bio, i) { |
| 7520 | ret = __readpage_endio_check(inode, io_bio, i, bvec->bv_page, | 7522 | ret = __readpage_endio_check(inode, io_bio, i, bvec->bv_page, |
| 7521 | 0, start, bvec->bv_len); | 7523 | 0, start, bvec->bv_len); |
| 7522 | if (likely(!ret)) | 7524 | if (likely(!ret)) |
| 7523 | goto next; | 7525 | goto next; |
| 7524 | try_again: | 7526 | try_again: |
| 7525 | done.uptodate = 0; | 7527 | done.uptodate = 0; |
| 7526 | done.start = start; | 7528 | done.start = start; |
| 7527 | init_completion(&done.done); | 7529 | init_completion(&done.done); |
| 7528 | 7530 | ||
| 7529 | ret = dio_read_error(inode, &io_bio->bio, bvec->bv_page, start, | 7531 | ret = dio_read_error(inode, &io_bio->bio, bvec->bv_page, start, |
| 7530 | start + bvec->bv_len - 1, | 7532 | start + bvec->bv_len - 1, |
| 7531 | io_bio->mirror_num, | 7533 | io_bio->mirror_num, |
| 7532 | btrfs_retry_endio, &done); | 7534 | btrfs_retry_endio, &done); |
| 7533 | if (ret) { | 7535 | if (ret) { |
| 7534 | err = ret; | 7536 | err = ret; |
| 7535 | goto next; | 7537 | goto next; |
| 7536 | } | 7538 | } |
| 7537 | 7539 | ||
| 7538 | wait_for_completion(&done.done); | 7540 | wait_for_completion(&done.done); |
| 7539 | 7541 | ||
| 7540 | if (!done.uptodate) { | 7542 | if (!done.uptodate) { |
| 7541 | /* We might have another mirror, so try again */ | 7543 | /* We might have another mirror, so try again */ |
| 7542 | goto try_again; | 7544 | goto try_again; |
| 7543 | } | 7545 | } |
| 7544 | next: | 7546 | next: |
| 7545 | offset += bvec->bv_len; | 7547 | offset += bvec->bv_len; |
| 7546 | start += bvec->bv_len; | 7548 | start += bvec->bv_len; |
| 7547 | } | 7549 | } |
| 7548 | 7550 | ||
| 7549 | return err; | 7551 | return err; |
| 7550 | } | 7552 | } |
| 7551 | 7553 | ||
| 7552 | static int btrfs_subio_endio_read(struct inode *inode, | 7554 | static int btrfs_subio_endio_read(struct inode *inode, |
| 7553 | struct btrfs_io_bio *io_bio, int err) | 7555 | struct btrfs_io_bio *io_bio, int err) |
| 7554 | { | 7556 | { |
| 7555 | bool skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; | 7557 | bool skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; |
| 7556 | 7558 | ||
| 7557 | if (skip_csum) { | 7559 | if (skip_csum) { |
| 7558 | if (unlikely(err)) | 7560 | if (unlikely(err)) |
| 7559 | return __btrfs_correct_data_nocsum(inode, io_bio); | 7561 | return __btrfs_correct_data_nocsum(inode, io_bio); |
| 7560 | else | 7562 | else |
| 7561 | return 0; | 7563 | return 0; |
| 7562 | } else { | 7564 | } else { |
| 7563 | return __btrfs_subio_endio_read(inode, io_bio, err); | 7565 | return __btrfs_subio_endio_read(inode, io_bio, err); |
| 7564 | } | 7566 | } |
| 7565 | } | 7567 | } |
| 7566 | 7568 | ||
| 7567 | static void btrfs_endio_direct_read(struct bio *bio, int err) | 7569 | static void btrfs_endio_direct_read(struct bio *bio, int err) |
| 7568 | { | 7570 | { |
| 7569 | struct btrfs_dio_private *dip = bio->bi_private; | 7571 | struct btrfs_dio_private *dip = bio->bi_private; |
| 7570 | struct inode *inode = dip->inode; | 7572 | struct inode *inode = dip->inode; |
| 7571 | struct bio *dio_bio; | 7573 | struct bio *dio_bio; |
| 7572 | struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); | 7574 | struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); |
| 7573 | 7575 | ||
| 7574 | if (dip->flags & BTRFS_DIO_ORIG_BIO_SUBMITTED) | 7576 | if (dip->flags & BTRFS_DIO_ORIG_BIO_SUBMITTED) |
| 7575 | err = btrfs_subio_endio_read(inode, io_bio, err); | 7577 | err = btrfs_subio_endio_read(inode, io_bio, err); |
| 7576 | 7578 | ||
| 7577 | unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset, | 7579 | unlock_extent(&BTRFS_I(inode)->io_tree, dip->logical_offset, |
| 7578 | dip->logical_offset + dip->bytes - 1); | 7580 | dip->logical_offset + dip->bytes - 1); |
| 7579 | dio_bio = dip->dio_bio; | 7581 | dio_bio = dip->dio_bio; |
| 7580 | 7582 | ||
| 7581 | kfree(dip); | 7583 | kfree(dip); |
| 7582 | 7584 | ||
| 7583 | /* If we had a csum failure make sure to clear the uptodate flag */ | 7585 | /* If we had a csum failure make sure to clear the uptodate flag */ |
| 7584 | if (err) | 7586 | if (err) |
| 7585 | clear_bit(BIO_UPTODATE, &dio_bio->bi_flags); | 7587 | clear_bit(BIO_UPTODATE, &dio_bio->bi_flags); |
| 7586 | dio_end_io(dio_bio, err); | 7588 | dio_end_io(dio_bio, err); |
| 7587 | 7589 | ||
| 7588 | if (io_bio->end_io) | 7590 | if (io_bio->end_io) |
| 7589 | io_bio->end_io(io_bio, err); | 7591 | io_bio->end_io(io_bio, err); |
| 7590 | bio_put(bio); | 7592 | bio_put(bio); |
| 7591 | } | 7593 | } |
| 7592 | 7594 | ||
| 7593 | static void btrfs_endio_direct_write(struct bio *bio, int err) | 7595 | static void btrfs_endio_direct_write(struct bio *bio, int err) |
| 7594 | { | 7596 | { |
| 7595 | struct btrfs_dio_private *dip = bio->bi_private; | 7597 | struct btrfs_dio_private *dip = bio->bi_private; |
| 7596 | struct inode *inode = dip->inode; | 7598 | struct inode *inode = dip->inode; |
| 7597 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7599 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7598 | struct btrfs_ordered_extent *ordered = NULL; | 7600 | struct btrfs_ordered_extent *ordered = NULL; |
| 7599 | u64 ordered_offset = dip->logical_offset; | 7601 | u64 ordered_offset = dip->logical_offset; |
| 7600 | u64 ordered_bytes = dip->bytes; | 7602 | u64 ordered_bytes = dip->bytes; |
| 7601 | struct bio *dio_bio; | 7603 | struct bio *dio_bio; |
| 7602 | int ret; | 7604 | int ret; |
| 7603 | 7605 | ||
| 7604 | if (err) | 7606 | if (err) |
| 7605 | goto out_done; | 7607 | goto out_done; |
| 7606 | again: | 7608 | again: |
| 7607 | ret = btrfs_dec_test_first_ordered_pending(inode, &ordered, | 7609 | ret = btrfs_dec_test_first_ordered_pending(inode, &ordered, |
| 7608 | &ordered_offset, | 7610 | &ordered_offset, |
| 7609 | ordered_bytes, !err); | 7611 | ordered_bytes, !err); |
| 7610 | if (!ret) | 7612 | if (!ret) |
| 7611 | goto out_test; | 7613 | goto out_test; |
| 7612 | 7614 | ||
| 7613 | btrfs_init_work(&ordered->work, btrfs_endio_write_helper, | 7615 | btrfs_init_work(&ordered->work, btrfs_endio_write_helper, |
| 7614 | finish_ordered_fn, NULL, NULL); | 7616 | finish_ordered_fn, NULL, NULL); |
| 7615 | btrfs_queue_work(root->fs_info->endio_write_workers, | 7617 | btrfs_queue_work(root->fs_info->endio_write_workers, |
| 7616 | &ordered->work); | 7618 | &ordered->work); |
| 7617 | out_test: | 7619 | out_test: |
| 7618 | /* | 7620 | /* |
| 7619 | * our bio might span multiple ordered extents. If we haven't | 7621 | * our bio might span multiple ordered extents. If we haven't |
| 7620 | * completed the accounting for the whole dio, go back and try again | 7622 | * completed the accounting for the whole dio, go back and try again |
| 7621 | */ | 7623 | */ |
| 7622 | if (ordered_offset < dip->logical_offset + dip->bytes) { | 7624 | if (ordered_offset < dip->logical_offset + dip->bytes) { |
| 7623 | ordered_bytes = dip->logical_offset + dip->bytes - | 7625 | ordered_bytes = dip->logical_offset + dip->bytes - |
| 7624 | ordered_offset; | 7626 | ordered_offset; |
| 7625 | ordered = NULL; | 7627 | ordered = NULL; |
| 7626 | goto again; | 7628 | goto again; |
| 7627 | } | 7629 | } |
| 7628 | out_done: | 7630 | out_done: |
| 7629 | dio_bio = dip->dio_bio; | 7631 | dio_bio = dip->dio_bio; |
| 7630 | 7632 | ||
| 7631 | kfree(dip); | 7633 | kfree(dip); |
| 7632 | 7634 | ||
| 7633 | /* If we had an error make sure to clear the uptodate flag */ | 7635 | /* If we had an error make sure to clear the uptodate flag */ |
| 7634 | if (err) | 7636 | if (err) |
| 7635 | clear_bit(BIO_UPTODATE, &dio_bio->bi_flags); | 7637 | clear_bit(BIO_UPTODATE, &dio_bio->bi_flags); |
| 7636 | dio_end_io(dio_bio, err); | 7638 | dio_end_io(dio_bio, err); |
| 7637 | bio_put(bio); | 7639 | bio_put(bio); |
| 7638 | } | 7640 | } |
| 7639 | 7641 | ||
| 7640 | static int __btrfs_submit_bio_start_direct_io(struct inode *inode, int rw, | 7642 | static int __btrfs_submit_bio_start_direct_io(struct inode *inode, int rw, |
| 7641 | struct bio *bio, int mirror_num, | 7643 | struct bio *bio, int mirror_num, |
| 7642 | unsigned long bio_flags, u64 offset) | 7644 | unsigned long bio_flags, u64 offset) |
| 7643 | { | 7645 | { |
| 7644 | int ret; | 7646 | int ret; |
| 7645 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7647 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7646 | ret = btrfs_csum_one_bio(root, inode, bio, offset, 1); | 7648 | ret = btrfs_csum_one_bio(root, inode, bio, offset, 1); |
| 7647 | BUG_ON(ret); /* -ENOMEM */ | 7649 | BUG_ON(ret); /* -ENOMEM */ |
| 7648 | return 0; | 7650 | return 0; |
| 7649 | } | 7651 | } |
| 7650 | 7652 | ||
| 7651 | static void btrfs_end_dio_bio(struct bio *bio, int err) | 7653 | static void btrfs_end_dio_bio(struct bio *bio, int err) |
| 7652 | { | 7654 | { |
| 7653 | struct btrfs_dio_private *dip = bio->bi_private; | 7655 | struct btrfs_dio_private *dip = bio->bi_private; |
| 7654 | 7656 | ||
| 7655 | if (err) | 7657 | if (err) |
| 7656 | btrfs_warn(BTRFS_I(dip->inode)->root->fs_info, | 7658 | btrfs_warn(BTRFS_I(dip->inode)->root->fs_info, |
| 7657 | "direct IO failed ino %llu rw %lu sector %#Lx len %u err no %d", | 7659 | "direct IO failed ino %llu rw %lu sector %#Lx len %u err no %d", |
| 7658 | btrfs_ino(dip->inode), bio->bi_rw, | 7660 | btrfs_ino(dip->inode), bio->bi_rw, |
| 7659 | (unsigned long long)bio->bi_iter.bi_sector, | 7661 | (unsigned long long)bio->bi_iter.bi_sector, |
| 7660 | bio->bi_iter.bi_size, err); | 7662 | bio->bi_iter.bi_size, err); |
| 7661 | 7663 | ||
| 7662 | if (dip->subio_endio) | 7664 | if (dip->subio_endio) |
| 7663 | err = dip->subio_endio(dip->inode, btrfs_io_bio(bio), err); | 7665 | err = dip->subio_endio(dip->inode, btrfs_io_bio(bio), err); |
| 7664 | 7666 | ||
| 7665 | if (err) { | 7667 | if (err) { |
| 7666 | dip->errors = 1; | 7668 | dip->errors = 1; |
| 7667 | 7669 | ||
| 7668 | /* | 7670 | /* |
| 7669 | * before atomic variable goto zero, we must make sure | 7671 | * before atomic variable goto zero, we must make sure |
| 7670 | * dip->errors is perceived to be set. | 7672 | * dip->errors is perceived to be set. |
| 7671 | */ | 7673 | */ |
| 7672 | smp_mb__before_atomic(); | 7674 | smp_mb__before_atomic(); |
| 7673 | } | 7675 | } |
| 7674 | 7676 | ||
| 7675 | /* if there are more bios still pending for this dio, just exit */ | 7677 | /* if there are more bios still pending for this dio, just exit */ |
| 7676 | if (!atomic_dec_and_test(&dip->pending_bios)) | 7678 | if (!atomic_dec_and_test(&dip->pending_bios)) |
| 7677 | goto out; | 7679 | goto out; |
| 7678 | 7680 | ||
| 7679 | if (dip->errors) { | 7681 | if (dip->errors) { |
| 7680 | bio_io_error(dip->orig_bio); | 7682 | bio_io_error(dip->orig_bio); |
| 7681 | } else { | 7683 | } else { |
| 7682 | set_bit(BIO_UPTODATE, &dip->dio_bio->bi_flags); | 7684 | set_bit(BIO_UPTODATE, &dip->dio_bio->bi_flags); |
| 7683 | bio_endio(dip->orig_bio, 0); | 7685 | bio_endio(dip->orig_bio, 0); |
| 7684 | } | 7686 | } |
| 7685 | out: | 7687 | out: |
| 7686 | bio_put(bio); | 7688 | bio_put(bio); |
| 7687 | } | 7689 | } |
| 7688 | 7690 | ||
| 7689 | static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev, | 7691 | static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev, |
| 7690 | u64 first_sector, gfp_t gfp_flags) | 7692 | u64 first_sector, gfp_t gfp_flags) |
| 7691 | { | 7693 | { |
| 7692 | int nr_vecs = bio_get_nr_vecs(bdev); | 7694 | int nr_vecs = bio_get_nr_vecs(bdev); |
| 7693 | return btrfs_bio_alloc(bdev, first_sector, nr_vecs, gfp_flags); | 7695 | return btrfs_bio_alloc(bdev, first_sector, nr_vecs, gfp_flags); |
| 7694 | } | 7696 | } |
| 7695 | 7697 | ||
| 7696 | static inline int btrfs_lookup_and_bind_dio_csum(struct btrfs_root *root, | 7698 | static inline int btrfs_lookup_and_bind_dio_csum(struct btrfs_root *root, |
| 7697 | struct inode *inode, | 7699 | struct inode *inode, |
| 7698 | struct btrfs_dio_private *dip, | 7700 | struct btrfs_dio_private *dip, |
| 7699 | struct bio *bio, | 7701 | struct bio *bio, |
| 7700 | u64 file_offset) | 7702 | u64 file_offset) |
| 7701 | { | 7703 | { |
| 7702 | struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); | 7704 | struct btrfs_io_bio *io_bio = btrfs_io_bio(bio); |
| 7703 | struct btrfs_io_bio *orig_io_bio = btrfs_io_bio(dip->orig_bio); | 7705 | struct btrfs_io_bio *orig_io_bio = btrfs_io_bio(dip->orig_bio); |
| 7704 | int ret; | 7706 | int ret; |
| 7705 | 7707 | ||
| 7706 | /* | 7708 | /* |
| 7707 | * We load all the csum data we need when we submit | 7709 | * We load all the csum data we need when we submit |
| 7708 | * the first bio to reduce the csum tree search and | 7710 | * the first bio to reduce the csum tree search and |
| 7709 | * contention. | 7711 | * contention. |
| 7710 | */ | 7712 | */ |
| 7711 | if (dip->logical_offset == file_offset) { | 7713 | if (dip->logical_offset == file_offset) { |
| 7712 | ret = btrfs_lookup_bio_sums_dio(root, inode, dip->orig_bio, | 7714 | ret = btrfs_lookup_bio_sums_dio(root, inode, dip->orig_bio, |
| 7713 | file_offset); | 7715 | file_offset); |
| 7714 | if (ret) | 7716 | if (ret) |
| 7715 | return ret; | 7717 | return ret; |
| 7716 | } | 7718 | } |
| 7717 | 7719 | ||
| 7718 | if (bio == dip->orig_bio) | 7720 | if (bio == dip->orig_bio) |
| 7719 | return 0; | 7721 | return 0; |
| 7720 | 7722 | ||
| 7721 | file_offset -= dip->logical_offset; | 7723 | file_offset -= dip->logical_offset; |
| 7722 | file_offset >>= inode->i_sb->s_blocksize_bits; | 7724 | file_offset >>= inode->i_sb->s_blocksize_bits; |
| 7723 | io_bio->csum = (u8 *)(((u32 *)orig_io_bio->csum) + file_offset); | 7725 | io_bio->csum = (u8 *)(((u32 *)orig_io_bio->csum) + file_offset); |
| 7724 | 7726 | ||
| 7725 | return 0; | 7727 | return 0; |
| 7726 | } | 7728 | } |
| 7727 | 7729 | ||
| 7728 | static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, | 7730 | static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, |
| 7729 | int rw, u64 file_offset, int skip_sum, | 7731 | int rw, u64 file_offset, int skip_sum, |
| 7730 | int async_submit) | 7732 | int async_submit) |
| 7731 | { | 7733 | { |
| 7732 | struct btrfs_dio_private *dip = bio->bi_private; | 7734 | struct btrfs_dio_private *dip = bio->bi_private; |
| 7733 | int write = rw & REQ_WRITE; | 7735 | int write = rw & REQ_WRITE; |
| 7734 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7736 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7735 | int ret; | 7737 | int ret; |
| 7736 | 7738 | ||
| 7737 | if (async_submit) | 7739 | if (async_submit) |
| 7738 | async_submit = !atomic_read(&BTRFS_I(inode)->sync_writers); | 7740 | async_submit = !atomic_read(&BTRFS_I(inode)->sync_writers); |
| 7739 | 7741 | ||
| 7740 | bio_get(bio); | 7742 | bio_get(bio); |
| 7741 | 7743 | ||
| 7742 | if (!write) { | 7744 | if (!write) { |
| 7743 | ret = btrfs_bio_wq_end_io(root->fs_info, bio, | 7745 | ret = btrfs_bio_wq_end_io(root->fs_info, bio, |
| 7744 | BTRFS_WQ_ENDIO_DATA); | 7746 | BTRFS_WQ_ENDIO_DATA); |
| 7745 | if (ret) | 7747 | if (ret) |
| 7746 | goto err; | 7748 | goto err; |
| 7747 | } | 7749 | } |
| 7748 | 7750 | ||
| 7749 | if (skip_sum) | 7751 | if (skip_sum) |
| 7750 | goto map; | 7752 | goto map; |
| 7751 | 7753 | ||
| 7752 | if (write && async_submit) { | 7754 | if (write && async_submit) { |
| 7753 | ret = btrfs_wq_submit_bio(root->fs_info, | 7755 | ret = btrfs_wq_submit_bio(root->fs_info, |
| 7754 | inode, rw, bio, 0, 0, | 7756 | inode, rw, bio, 0, 0, |
| 7755 | file_offset, | 7757 | file_offset, |
| 7756 | __btrfs_submit_bio_start_direct_io, | 7758 | __btrfs_submit_bio_start_direct_io, |
| 7757 | __btrfs_submit_bio_done); | 7759 | __btrfs_submit_bio_done); |
| 7758 | goto err; | 7760 | goto err; |
| 7759 | } else if (write) { | 7761 | } else if (write) { |
| 7760 | /* | 7762 | /* |
| 7761 | * If we aren't doing async submit, calculate the csum of the | 7763 | * If we aren't doing async submit, calculate the csum of the |
| 7762 | * bio now. | 7764 | * bio now. |
| 7763 | */ | 7765 | */ |
| 7764 | ret = btrfs_csum_one_bio(root, inode, bio, file_offset, 1); | 7766 | ret = btrfs_csum_one_bio(root, inode, bio, file_offset, 1); |
| 7765 | if (ret) | 7767 | if (ret) |
| 7766 | goto err; | 7768 | goto err; |
| 7767 | } else { | 7769 | } else { |
| 7768 | ret = btrfs_lookup_and_bind_dio_csum(root, inode, dip, bio, | 7770 | ret = btrfs_lookup_and_bind_dio_csum(root, inode, dip, bio, |
| 7769 | file_offset); | 7771 | file_offset); |
| 7770 | if (ret) | 7772 | if (ret) |
| 7771 | goto err; | 7773 | goto err; |
| 7772 | } | 7774 | } |
| 7773 | map: | 7775 | map: |
| 7774 | ret = btrfs_map_bio(root, rw, bio, 0, async_submit); | 7776 | ret = btrfs_map_bio(root, rw, bio, 0, async_submit); |
| 7775 | err: | 7777 | err: |
| 7776 | bio_put(bio); | 7778 | bio_put(bio); |
| 7777 | return ret; | 7779 | return ret; |
| 7778 | } | 7780 | } |
| 7779 | 7781 | ||
| 7780 | static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, | 7782 | static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, |
| 7781 | int skip_sum) | 7783 | int skip_sum) |
| 7782 | { | 7784 | { |
| 7783 | struct inode *inode = dip->inode; | 7785 | struct inode *inode = dip->inode; |
| 7784 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7786 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7785 | struct bio *bio; | 7787 | struct bio *bio; |
| 7786 | struct bio *orig_bio = dip->orig_bio; | 7788 | struct bio *orig_bio = dip->orig_bio; |
| 7787 | struct bio_vec *bvec = orig_bio->bi_io_vec; | 7789 | struct bio_vec *bvec = orig_bio->bi_io_vec; |
| 7788 | u64 start_sector = orig_bio->bi_iter.bi_sector; | 7790 | u64 start_sector = orig_bio->bi_iter.bi_sector; |
| 7789 | u64 file_offset = dip->logical_offset; | 7791 | u64 file_offset = dip->logical_offset; |
| 7790 | u64 submit_len = 0; | 7792 | u64 submit_len = 0; |
| 7791 | u64 map_length; | 7793 | u64 map_length; |
| 7792 | int nr_pages = 0; | 7794 | int nr_pages = 0; |
| 7793 | int ret; | 7795 | int ret; |
| 7794 | int async_submit = 0; | 7796 | int async_submit = 0; |
| 7795 | 7797 | ||
| 7796 | map_length = orig_bio->bi_iter.bi_size; | 7798 | map_length = orig_bio->bi_iter.bi_size; |
| 7797 | ret = btrfs_map_block(root->fs_info, rw, start_sector << 9, | 7799 | ret = btrfs_map_block(root->fs_info, rw, start_sector << 9, |
| 7798 | &map_length, NULL, 0); | 7800 | &map_length, NULL, 0); |
| 7799 | if (ret) | 7801 | if (ret) |
| 7800 | return -EIO; | 7802 | return -EIO; |
| 7801 | 7803 | ||
| 7802 | if (map_length >= orig_bio->bi_iter.bi_size) { | 7804 | if (map_length >= orig_bio->bi_iter.bi_size) { |
| 7803 | bio = orig_bio; | 7805 | bio = orig_bio; |
| 7804 | dip->flags |= BTRFS_DIO_ORIG_BIO_SUBMITTED; | 7806 | dip->flags |= BTRFS_DIO_ORIG_BIO_SUBMITTED; |
| 7805 | goto submit; | 7807 | goto submit; |
| 7806 | } | 7808 | } |
| 7807 | 7809 | ||
| 7808 | /* async crcs make it difficult to collect full stripe writes. */ | 7810 | /* async crcs make it difficult to collect full stripe writes. */ |
| 7809 | if (btrfs_get_alloc_profile(root, 1) & | 7811 | if (btrfs_get_alloc_profile(root, 1) & |
| 7810 | (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6)) | 7812 | (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6)) |
| 7811 | async_submit = 0; | 7813 | async_submit = 0; |
| 7812 | else | 7814 | else |
| 7813 | async_submit = 1; | 7815 | async_submit = 1; |
| 7814 | 7816 | ||
| 7815 | bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); | 7817 | bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); |
| 7816 | if (!bio) | 7818 | if (!bio) |
| 7817 | return -ENOMEM; | 7819 | return -ENOMEM; |
| 7818 | 7820 | ||
| 7819 | bio->bi_private = dip; | 7821 | bio->bi_private = dip; |
| 7820 | bio->bi_end_io = btrfs_end_dio_bio; | 7822 | bio->bi_end_io = btrfs_end_dio_bio; |
| 7821 | btrfs_io_bio(bio)->logical = file_offset; | 7823 | btrfs_io_bio(bio)->logical = file_offset; |
| 7822 | atomic_inc(&dip->pending_bios); | 7824 | atomic_inc(&dip->pending_bios); |
| 7823 | 7825 | ||
| 7824 | while (bvec <= (orig_bio->bi_io_vec + orig_bio->bi_vcnt - 1)) { | 7826 | while (bvec <= (orig_bio->bi_io_vec + orig_bio->bi_vcnt - 1)) { |
| 7825 | if (map_length < submit_len + bvec->bv_len || | 7827 | if (map_length < submit_len + bvec->bv_len || |
| 7826 | bio_add_page(bio, bvec->bv_page, bvec->bv_len, | 7828 | bio_add_page(bio, bvec->bv_page, bvec->bv_len, |
| 7827 | bvec->bv_offset) < bvec->bv_len) { | 7829 | bvec->bv_offset) < bvec->bv_len) { |
| 7828 | /* | 7830 | /* |
| 7829 | * inc the count before we submit the bio so | 7831 | * inc the count before we submit the bio so |
| 7830 | * we know the end IO handler won't happen before | 7832 | * we know the end IO handler won't happen before |
| 7831 | * we inc the count. Otherwise, the dip might get freed | 7833 | * we inc the count. Otherwise, the dip might get freed |
| 7832 | * before we're done setting it up | 7834 | * before we're done setting it up |
| 7833 | */ | 7835 | */ |
| 7834 | atomic_inc(&dip->pending_bios); | 7836 | atomic_inc(&dip->pending_bios); |
| 7835 | ret = __btrfs_submit_dio_bio(bio, inode, rw, | 7837 | ret = __btrfs_submit_dio_bio(bio, inode, rw, |
| 7836 | file_offset, skip_sum, | 7838 | file_offset, skip_sum, |
| 7837 | async_submit); | 7839 | async_submit); |
| 7838 | if (ret) { | 7840 | if (ret) { |
| 7839 | bio_put(bio); | 7841 | bio_put(bio); |
| 7840 | atomic_dec(&dip->pending_bios); | 7842 | atomic_dec(&dip->pending_bios); |
| 7841 | goto out_err; | 7843 | goto out_err; |
| 7842 | } | 7844 | } |
| 7843 | 7845 | ||
| 7844 | start_sector += submit_len >> 9; | 7846 | start_sector += submit_len >> 9; |
| 7845 | file_offset += submit_len; | 7847 | file_offset += submit_len; |
| 7846 | 7848 | ||
| 7847 | submit_len = 0; | 7849 | submit_len = 0; |
| 7848 | nr_pages = 0; | 7850 | nr_pages = 0; |
| 7849 | 7851 | ||
| 7850 | bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, | 7852 | bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, |
| 7851 | start_sector, GFP_NOFS); | 7853 | start_sector, GFP_NOFS); |
| 7852 | if (!bio) | 7854 | if (!bio) |
| 7853 | goto out_err; | 7855 | goto out_err; |
| 7854 | bio->bi_private = dip; | 7856 | bio->bi_private = dip; |
| 7855 | bio->bi_end_io = btrfs_end_dio_bio; | 7857 | bio->bi_end_io = btrfs_end_dio_bio; |
| 7856 | btrfs_io_bio(bio)->logical = file_offset; | 7858 | btrfs_io_bio(bio)->logical = file_offset; |
| 7857 | 7859 | ||
| 7858 | map_length = orig_bio->bi_iter.bi_size; | 7860 | map_length = orig_bio->bi_iter.bi_size; |
| 7859 | ret = btrfs_map_block(root->fs_info, rw, | 7861 | ret = btrfs_map_block(root->fs_info, rw, |
| 7860 | start_sector << 9, | 7862 | start_sector << 9, |
| 7861 | &map_length, NULL, 0); | 7863 | &map_length, NULL, 0); |
| 7862 | if (ret) { | 7864 | if (ret) { |
| 7863 | bio_put(bio); | 7865 | bio_put(bio); |
| 7864 | goto out_err; | 7866 | goto out_err; |
| 7865 | } | 7867 | } |
| 7866 | } else { | 7868 | } else { |
| 7867 | submit_len += bvec->bv_len; | 7869 | submit_len += bvec->bv_len; |
| 7868 | nr_pages++; | 7870 | nr_pages++; |
| 7869 | bvec++; | 7871 | bvec++; |
| 7870 | } | 7872 | } |
| 7871 | } | 7873 | } |
| 7872 | 7874 | ||
| 7873 | submit: | 7875 | submit: |
| 7874 | ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, | 7876 | ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, |
| 7875 | async_submit); | 7877 | async_submit); |
| 7876 | if (!ret) | 7878 | if (!ret) |
| 7877 | return 0; | 7879 | return 0; |
| 7878 | 7880 | ||
| 7879 | bio_put(bio); | 7881 | bio_put(bio); |
| 7880 | out_err: | 7882 | out_err: |
| 7881 | dip->errors = 1; | 7883 | dip->errors = 1; |
| 7882 | /* | 7884 | /* |
| 7883 | * before atomic variable goto zero, we must | 7885 | * before atomic variable goto zero, we must |
| 7884 | * make sure dip->errors is perceived to be set. | 7886 | * make sure dip->errors is perceived to be set. |
| 7885 | */ | 7887 | */ |
| 7886 | smp_mb__before_atomic(); | 7888 | smp_mb__before_atomic(); |
| 7887 | if (atomic_dec_and_test(&dip->pending_bios)) | 7889 | if (atomic_dec_and_test(&dip->pending_bios)) |
| 7888 | bio_io_error(dip->orig_bio); | 7890 | bio_io_error(dip->orig_bio); |
| 7889 | 7891 | ||
| 7890 | /* bio_end_io() will handle error, so we needn't return it */ | 7892 | /* bio_end_io() will handle error, so we needn't return it */ |
| 7891 | return 0; | 7893 | return 0; |
| 7892 | } | 7894 | } |
| 7893 | 7895 | ||
| 7894 | static void btrfs_submit_direct(int rw, struct bio *dio_bio, | 7896 | static void btrfs_submit_direct(int rw, struct bio *dio_bio, |
| 7895 | struct inode *inode, loff_t file_offset) | 7897 | struct inode *inode, loff_t file_offset) |
| 7896 | { | 7898 | { |
| 7897 | struct btrfs_root *root = BTRFS_I(inode)->root; | 7899 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 7898 | struct btrfs_dio_private *dip; | 7900 | struct btrfs_dio_private *dip; |
| 7899 | struct bio *io_bio; | 7901 | struct bio *io_bio; |
| 7900 | struct btrfs_io_bio *btrfs_bio; | 7902 | struct btrfs_io_bio *btrfs_bio; |
| 7901 | int skip_sum; | 7903 | int skip_sum; |
| 7902 | int write = rw & REQ_WRITE; | 7904 | int write = rw & REQ_WRITE; |
| 7903 | int ret = 0; | 7905 | int ret = 0; |
| 7904 | 7906 | ||
| 7905 | skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; | 7907 | skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; |
| 7906 | 7908 | ||
| 7907 | io_bio = btrfs_bio_clone(dio_bio, GFP_NOFS); | 7909 | io_bio = btrfs_bio_clone(dio_bio, GFP_NOFS); |
| 7908 | if (!io_bio) { | 7910 | if (!io_bio) { |
| 7909 | ret = -ENOMEM; | 7911 | ret = -ENOMEM; |
| 7910 | goto free_ordered; | 7912 | goto free_ordered; |
| 7911 | } | 7913 | } |
| 7912 | 7914 | ||
| 7913 | dip = kzalloc(sizeof(*dip), GFP_NOFS); | 7915 | dip = kzalloc(sizeof(*dip), GFP_NOFS); |
| 7914 | if (!dip) { | 7916 | if (!dip) { |
| 7915 | ret = -ENOMEM; | 7917 | ret = -ENOMEM; |
| 7916 | goto free_io_bio; | 7918 | goto free_io_bio; |
| 7917 | } | 7919 | } |
| 7918 | 7920 | ||
| 7919 | dip->private = dio_bio->bi_private; | 7921 | dip->private = dio_bio->bi_private; |
| 7920 | dip->inode = inode; | 7922 | dip->inode = inode; |
| 7921 | dip->logical_offset = file_offset; | 7923 | dip->logical_offset = file_offset; |
| 7922 | dip->bytes = dio_bio->bi_iter.bi_size; | 7924 | dip->bytes = dio_bio->bi_iter.bi_size; |
| 7923 | dip->disk_bytenr = (u64)dio_bio->bi_iter.bi_sector << 9; | 7925 | dip->disk_bytenr = (u64)dio_bio->bi_iter.bi_sector << 9; |
| 7924 | io_bio->bi_private = dip; | 7926 | io_bio->bi_private = dip; |
| 7925 | dip->orig_bio = io_bio; | 7927 | dip->orig_bio = io_bio; |
| 7926 | dip->dio_bio = dio_bio; | 7928 | dip->dio_bio = dio_bio; |
| 7927 | atomic_set(&dip->pending_bios, 0); | 7929 | atomic_set(&dip->pending_bios, 0); |
| 7928 | btrfs_bio = btrfs_io_bio(io_bio); | 7930 | btrfs_bio = btrfs_io_bio(io_bio); |
| 7929 | btrfs_bio->logical = file_offset; | 7931 | btrfs_bio->logical = file_offset; |
| 7930 | 7932 | ||
| 7931 | if (write) { | 7933 | if (write) { |
| 7932 | io_bio->bi_end_io = btrfs_endio_direct_write; | 7934 | io_bio->bi_end_io = btrfs_endio_direct_write; |
| 7933 | } else { | 7935 | } else { |
| 7934 | io_bio->bi_end_io = btrfs_endio_direct_read; | 7936 | io_bio->bi_end_io = btrfs_endio_direct_read; |
| 7935 | dip->subio_endio = btrfs_subio_endio_read; | 7937 | dip->subio_endio = btrfs_subio_endio_read; |
| 7936 | } | 7938 | } |
| 7937 | 7939 | ||
| 7938 | ret = btrfs_submit_direct_hook(rw, dip, skip_sum); | 7940 | ret = btrfs_submit_direct_hook(rw, dip, skip_sum); |
| 7939 | if (!ret) | 7941 | if (!ret) |
| 7940 | return; | 7942 | return; |
| 7941 | 7943 | ||
| 7942 | if (btrfs_bio->end_io) | 7944 | if (btrfs_bio->end_io) |
| 7943 | btrfs_bio->end_io(btrfs_bio, ret); | 7945 | btrfs_bio->end_io(btrfs_bio, ret); |
| 7944 | free_io_bio: | 7946 | free_io_bio: |
| 7945 | bio_put(io_bio); | 7947 | bio_put(io_bio); |
| 7946 | 7948 | ||
| 7947 | free_ordered: | 7949 | free_ordered: |
| 7948 | /* | 7950 | /* |
| 7949 | * If this is a write, we need to clean up the reserved space and kill | 7951 | * If this is a write, we need to clean up the reserved space and kill |
| 7950 | * the ordered extent. | 7952 | * the ordered extent. |
| 7951 | */ | 7953 | */ |
| 7952 | if (write) { | 7954 | if (write) { |
| 7953 | struct btrfs_ordered_extent *ordered; | 7955 | struct btrfs_ordered_extent *ordered; |
| 7954 | ordered = btrfs_lookup_ordered_extent(inode, file_offset); | 7956 | ordered = btrfs_lookup_ordered_extent(inode, file_offset); |
| 7955 | if (!test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags) && | 7957 | if (!test_bit(BTRFS_ORDERED_PREALLOC, &ordered->flags) && |
| 7956 | !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) | 7958 | !test_bit(BTRFS_ORDERED_NOCOW, &ordered->flags)) |
| 7957 | btrfs_free_reserved_extent(root, ordered->start, | 7959 | btrfs_free_reserved_extent(root, ordered->start, |
| 7958 | ordered->disk_len, 1); | 7960 | ordered->disk_len, 1); |
| 7959 | btrfs_put_ordered_extent(ordered); | 7961 | btrfs_put_ordered_extent(ordered); |
| 7960 | btrfs_put_ordered_extent(ordered); | 7962 | btrfs_put_ordered_extent(ordered); |
| 7961 | } | 7963 | } |
| 7962 | bio_endio(dio_bio, ret); | 7964 | bio_endio(dio_bio, ret); |
| 7963 | } | 7965 | } |
| 7964 | 7966 | ||
| 7965 | static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *iocb, | 7967 | static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *iocb, |
| 7966 | const struct iov_iter *iter, loff_t offset) | 7968 | const struct iov_iter *iter, loff_t offset) |
| 7967 | { | 7969 | { |
| 7968 | int seg; | 7970 | int seg; |
| 7969 | int i; | 7971 | int i; |
| 7970 | unsigned blocksize_mask = root->sectorsize - 1; | 7972 | unsigned blocksize_mask = root->sectorsize - 1; |
| 7971 | ssize_t retval = -EINVAL; | 7973 | ssize_t retval = -EINVAL; |
| 7972 | 7974 | ||
| 7973 | if (offset & blocksize_mask) | 7975 | if (offset & blocksize_mask) |
| 7974 | goto out; | 7976 | goto out; |
| 7975 | 7977 | ||
| 7976 | if (iov_iter_alignment(iter) & blocksize_mask) | 7978 | if (iov_iter_alignment(iter) & blocksize_mask) |
| 7977 | goto out; | 7979 | goto out; |
| 7978 | 7980 | ||
| 7979 | /* If this is a write we don't need to check anymore */ | 7981 | /* If this is a write we don't need to check anymore */ |
| 7980 | if (rw & WRITE) | 7982 | if (rw & WRITE) |
| 7981 | return 0; | 7983 | return 0; |
| 7982 | /* | 7984 | /* |
| 7983 | * Check to make sure we don't have duplicate iov_base's in this | 7985 | * Check to make sure we don't have duplicate iov_base's in this |
| 7984 | * iovec, if so return EINVAL, otherwise we'll get csum errors | 7986 | * iovec, if so return EINVAL, otherwise we'll get csum errors |
| 7985 | * when reading back. | 7987 | * when reading back. |
| 7986 | */ | 7988 | */ |
| 7987 | for (seg = 0; seg < iter->nr_segs; seg++) { | 7989 | for (seg = 0; seg < iter->nr_segs; seg++) { |
| 7988 | for (i = seg + 1; i < iter->nr_segs; i++) { | 7990 | for (i = seg + 1; i < iter->nr_segs; i++) { |
| 7989 | if (iter->iov[seg].iov_base == iter->iov[i].iov_base) | 7991 | if (iter->iov[seg].iov_base == iter->iov[i].iov_base) |
| 7990 | goto out; | 7992 | goto out; |
| 7991 | } | 7993 | } |
| 7992 | } | 7994 | } |
| 7993 | retval = 0; | 7995 | retval = 0; |
| 7994 | out: | 7996 | out: |
| 7995 | return retval; | 7997 | return retval; |
| 7996 | } | 7998 | } |
| 7997 | 7999 | ||
| 7998 | static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb, | 8000 | static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb, |
| 7999 | struct iov_iter *iter, loff_t offset) | 8001 | struct iov_iter *iter, loff_t offset) |
| 8000 | { | 8002 | { |
| 8001 | struct file *file = iocb->ki_filp; | 8003 | struct file *file = iocb->ki_filp; |
| 8002 | struct inode *inode = file->f_mapping->host; | 8004 | struct inode *inode = file->f_mapping->host; |
| 8003 | size_t count = 0; | 8005 | size_t count = 0; |
| 8004 | int flags = 0; | 8006 | int flags = 0; |
| 8005 | bool wakeup = true; | 8007 | bool wakeup = true; |
| 8006 | bool relock = false; | 8008 | bool relock = false; |
| 8007 | ssize_t ret; | 8009 | ssize_t ret; |
| 8008 | 8010 | ||
| 8009 | if (check_direct_IO(BTRFS_I(inode)->root, rw, iocb, iter, offset)) | 8011 | if (check_direct_IO(BTRFS_I(inode)->root, rw, iocb, iter, offset)) |
| 8010 | return 0; | 8012 | return 0; |
| 8011 | 8013 | ||
| 8012 | atomic_inc(&inode->i_dio_count); | 8014 | atomic_inc(&inode->i_dio_count); |
| 8013 | smp_mb__after_atomic(); | 8015 | smp_mb__after_atomic(); |
| 8014 | 8016 | ||
| 8015 | /* | 8017 | /* |
| 8016 | * The generic stuff only does filemap_write_and_wait_range, which | 8018 | * The generic stuff only does filemap_write_and_wait_range, which |
| 8017 | * isn't enough if we've written compressed pages to this area, so | 8019 | * isn't enough if we've written compressed pages to this area, so |
| 8018 | * we need to flush the dirty pages again to make absolutely sure | 8020 | * we need to flush the dirty pages again to make absolutely sure |
| 8019 | * that any outstanding dirty pages are on disk. | 8021 | * that any outstanding dirty pages are on disk. |
| 8020 | */ | 8022 | */ |
| 8021 | count = iov_iter_count(iter); | 8023 | count = iov_iter_count(iter); |
| 8022 | if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, | 8024 | if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, |
| 8023 | &BTRFS_I(inode)->runtime_flags)) | 8025 | &BTRFS_I(inode)->runtime_flags)) |
| 8024 | filemap_fdatawrite_range(inode->i_mapping, offset, | 8026 | filemap_fdatawrite_range(inode->i_mapping, offset, |
| 8025 | offset + count - 1); | 8027 | offset + count - 1); |
| 8026 | 8028 | ||
| 8027 | if (rw & WRITE) { | 8029 | if (rw & WRITE) { |
| 8028 | /* | 8030 | /* |
| 8029 | * If the write DIO is beyond the EOF, we need update | 8031 | * If the write DIO is beyond the EOF, we need update |
| 8030 | * the isize, but it is protected by i_mutex. So we can | 8032 | * the isize, but it is protected by i_mutex. So we can |
| 8031 | * not unlock the i_mutex at this case. | 8033 | * not unlock the i_mutex at this case. |
| 8032 | */ | 8034 | */ |
| 8033 | if (offset + count <= inode->i_size) { | 8035 | if (offset + count <= inode->i_size) { |
| 8034 | mutex_unlock(&inode->i_mutex); | 8036 | mutex_unlock(&inode->i_mutex); |
| 8035 | relock = true; | 8037 | relock = true; |
| 8036 | } | 8038 | } |
| 8037 | ret = btrfs_delalloc_reserve_space(inode, count); | 8039 | ret = btrfs_delalloc_reserve_space(inode, count); |
| 8038 | if (ret) | 8040 | if (ret) |
| 8039 | goto out; | 8041 | goto out; |
| 8040 | } else if (test_bit(BTRFS_INODE_READDIO_NEED_LOCK, | 8042 | } else if (test_bit(BTRFS_INODE_READDIO_NEED_LOCK, |
| 8041 | &BTRFS_I(inode)->runtime_flags)) { | 8043 | &BTRFS_I(inode)->runtime_flags)) { |
| 8042 | inode_dio_done(inode); | 8044 | inode_dio_done(inode); |
| 8043 | flags = DIO_LOCKING | DIO_SKIP_HOLES; | 8045 | flags = DIO_LOCKING | DIO_SKIP_HOLES; |
| 8044 | wakeup = false; | 8046 | wakeup = false; |
| 8045 | } | 8047 | } |
| 8046 | 8048 | ||
| 8047 | ret = __blockdev_direct_IO(rw, iocb, inode, | 8049 | ret = __blockdev_direct_IO(rw, iocb, inode, |
| 8048 | BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev, | 8050 | BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev, |
| 8049 | iter, offset, btrfs_get_blocks_direct, NULL, | 8051 | iter, offset, btrfs_get_blocks_direct, NULL, |
| 8050 | btrfs_submit_direct, flags); | 8052 | btrfs_submit_direct, flags); |
| 8051 | if (rw & WRITE) { | 8053 | if (rw & WRITE) { |
| 8052 | if (ret < 0 && ret != -EIOCBQUEUED) | 8054 | if (ret < 0 && ret != -EIOCBQUEUED) |
| 8053 | btrfs_delalloc_release_space(inode, count); | 8055 | btrfs_delalloc_release_space(inode, count); |
| 8054 | else if (ret >= 0 && (size_t)ret < count) | 8056 | else if (ret >= 0 && (size_t)ret < count) |
| 8055 | btrfs_delalloc_release_space(inode, | 8057 | btrfs_delalloc_release_space(inode, |
| 8056 | count - (size_t)ret); | 8058 | count - (size_t)ret); |
| 8057 | else | 8059 | else |
| 8058 | btrfs_delalloc_release_metadata(inode, 0); | 8060 | btrfs_delalloc_release_metadata(inode, 0); |
| 8059 | } | 8061 | } |
| 8060 | out: | 8062 | out: |
| 8061 | if (wakeup) | 8063 | if (wakeup) |
| 8062 | inode_dio_done(inode); | 8064 | inode_dio_done(inode); |
| 8063 | if (relock) | 8065 | if (relock) |
| 8064 | mutex_lock(&inode->i_mutex); | 8066 | mutex_lock(&inode->i_mutex); |
| 8065 | 8067 | ||
| 8066 | return ret; | 8068 | return ret; |
| 8067 | } | 8069 | } |
| 8068 | 8070 | ||
| 8069 | #define BTRFS_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC) | 8071 | #define BTRFS_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC) |
| 8070 | 8072 | ||
| 8071 | static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, | 8073 | static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, |
| 8072 | __u64 start, __u64 len) | 8074 | __u64 start, __u64 len) |
| 8073 | { | 8075 | { |
| 8074 | int ret; | 8076 | int ret; |
| 8075 | 8077 | ||
| 8076 | ret = fiemap_check_flags(fieinfo, BTRFS_FIEMAP_FLAGS); | 8078 | ret = fiemap_check_flags(fieinfo, BTRFS_FIEMAP_FLAGS); |
| 8077 | if (ret) | 8079 | if (ret) |
| 8078 | return ret; | 8080 | return ret; |
| 8079 | 8081 | ||
| 8080 | return extent_fiemap(inode, fieinfo, start, len, btrfs_get_extent_fiemap); | 8082 | return extent_fiemap(inode, fieinfo, start, len, btrfs_get_extent_fiemap); |
| 8081 | } | 8083 | } |
| 8082 | 8084 | ||
| 8083 | int btrfs_readpage(struct file *file, struct page *page) | 8085 | int btrfs_readpage(struct file *file, struct page *page) |
| 8084 | { | 8086 | { |
| 8085 | struct extent_io_tree *tree; | 8087 | struct extent_io_tree *tree; |
| 8086 | tree = &BTRFS_I(page->mapping->host)->io_tree; | 8088 | tree = &BTRFS_I(page->mapping->host)->io_tree; |
| 8087 | return extent_read_full_page(tree, page, btrfs_get_extent, 0); | 8089 | return extent_read_full_page(tree, page, btrfs_get_extent, 0); |
| 8088 | } | 8090 | } |
| 8089 | 8091 | ||
| 8090 | static int btrfs_writepage(struct page *page, struct writeback_control *wbc) | 8092 | static int btrfs_writepage(struct page *page, struct writeback_control *wbc) |
| 8091 | { | 8093 | { |
| 8092 | struct extent_io_tree *tree; | 8094 | struct extent_io_tree *tree; |
| 8093 | 8095 | ||
| 8094 | 8096 | ||
| 8095 | if (current->flags & PF_MEMALLOC) { | 8097 | if (current->flags & PF_MEMALLOC) { |
| 8096 | redirty_page_for_writepage(wbc, page); | 8098 | redirty_page_for_writepage(wbc, page); |
| 8097 | unlock_page(page); | 8099 | unlock_page(page); |
| 8098 | return 0; | 8100 | return 0; |
| 8099 | } | 8101 | } |
| 8100 | tree = &BTRFS_I(page->mapping->host)->io_tree; | 8102 | tree = &BTRFS_I(page->mapping->host)->io_tree; |
| 8101 | return extent_write_full_page(tree, page, btrfs_get_extent, wbc); | 8103 | return extent_write_full_page(tree, page, btrfs_get_extent, wbc); |
| 8102 | } | 8104 | } |
| 8103 | 8105 | ||
| 8104 | static int btrfs_writepages(struct address_space *mapping, | 8106 | static int btrfs_writepages(struct address_space *mapping, |
| 8105 | struct writeback_control *wbc) | 8107 | struct writeback_control *wbc) |
| 8106 | { | 8108 | { |
| 8107 | struct extent_io_tree *tree; | 8109 | struct extent_io_tree *tree; |
| 8108 | 8110 | ||
| 8109 | tree = &BTRFS_I(mapping->host)->io_tree; | 8111 | tree = &BTRFS_I(mapping->host)->io_tree; |
| 8110 | return extent_writepages(tree, mapping, btrfs_get_extent, wbc); | 8112 | return extent_writepages(tree, mapping, btrfs_get_extent, wbc); |
| 8111 | } | 8113 | } |
| 8112 | 8114 | ||
| 8113 | static int | 8115 | static int |
| 8114 | btrfs_readpages(struct file *file, struct address_space *mapping, | 8116 | btrfs_readpages(struct file *file, struct address_space *mapping, |
| 8115 | struct list_head *pages, unsigned nr_pages) | 8117 | struct list_head *pages, unsigned nr_pages) |
| 8116 | { | 8118 | { |
| 8117 | struct extent_io_tree *tree; | 8119 | struct extent_io_tree *tree; |
| 8118 | tree = &BTRFS_I(mapping->host)->io_tree; | 8120 | tree = &BTRFS_I(mapping->host)->io_tree; |
| 8119 | return extent_readpages(tree, mapping, pages, nr_pages, | 8121 | return extent_readpages(tree, mapping, pages, nr_pages, |
| 8120 | btrfs_get_extent); | 8122 | btrfs_get_extent); |
| 8121 | } | 8123 | } |
| 8122 | static int __btrfs_releasepage(struct page *page, gfp_t gfp_flags) | 8124 | static int __btrfs_releasepage(struct page *page, gfp_t gfp_flags) |
| 8123 | { | 8125 | { |
| 8124 | struct extent_io_tree *tree; | 8126 | struct extent_io_tree *tree; |
| 8125 | struct extent_map_tree *map; | 8127 | struct extent_map_tree *map; |
| 8126 | int ret; | 8128 | int ret; |
| 8127 | 8129 | ||
| 8128 | tree = &BTRFS_I(page->mapping->host)->io_tree; | 8130 | tree = &BTRFS_I(page->mapping->host)->io_tree; |
| 8129 | map = &BTRFS_I(page->mapping->host)->extent_tree; | 8131 | map = &BTRFS_I(page->mapping->host)->extent_tree; |
| 8130 | ret = try_release_extent_mapping(map, tree, page, gfp_flags); | 8132 | ret = try_release_extent_mapping(map, tree, page, gfp_flags); |
| 8131 | if (ret == 1) { | 8133 | if (ret == 1) { |
| 8132 | ClearPagePrivate(page); | 8134 | ClearPagePrivate(page); |
| 8133 | set_page_private(page, 0); | 8135 | set_page_private(page, 0); |
| 8134 | page_cache_release(page); | 8136 | page_cache_release(page); |
| 8135 | } | 8137 | } |
| 8136 | return ret; | 8138 | return ret; |
| 8137 | } | 8139 | } |
| 8138 | 8140 | ||
| 8139 | static int btrfs_releasepage(struct page *page, gfp_t gfp_flags) | 8141 | static int btrfs_releasepage(struct page *page, gfp_t gfp_flags) |
| 8140 | { | 8142 | { |
| 8141 | if (PageWriteback(page) || PageDirty(page)) | 8143 | if (PageWriteback(page) || PageDirty(page)) |
| 8142 | return 0; | 8144 | return 0; |
| 8143 | return __btrfs_releasepage(page, gfp_flags & GFP_NOFS); | 8145 | return __btrfs_releasepage(page, gfp_flags & GFP_NOFS); |
| 8144 | } | 8146 | } |
| 8145 | 8147 | ||
| 8146 | static void btrfs_invalidatepage(struct page *page, unsigned int offset, | 8148 | static void btrfs_invalidatepage(struct page *page, unsigned int offset, |
| 8147 | unsigned int length) | 8149 | unsigned int length) |
| 8148 | { | 8150 | { |
| 8149 | struct inode *inode = page->mapping->host; | 8151 | struct inode *inode = page->mapping->host; |
| 8150 | struct extent_io_tree *tree; | 8152 | struct extent_io_tree *tree; |
| 8151 | struct btrfs_ordered_extent *ordered; | 8153 | struct btrfs_ordered_extent *ordered; |
| 8152 | struct extent_state *cached_state = NULL; | 8154 | struct extent_state *cached_state = NULL; |
| 8153 | u64 page_start = page_offset(page); | 8155 | u64 page_start = page_offset(page); |
| 8154 | u64 page_end = page_start + PAGE_CACHE_SIZE - 1; | 8156 | u64 page_end = page_start + PAGE_CACHE_SIZE - 1; |
| 8155 | int inode_evicting = inode->i_state & I_FREEING; | 8157 | int inode_evicting = inode->i_state & I_FREEING; |
| 8156 | 8158 | ||
| 8157 | /* | 8159 | /* |
| 8158 | * we have the page locked, so new writeback can't start, | 8160 | * we have the page locked, so new writeback can't start, |
| 8159 | * and the dirty bit won't be cleared while we are here. | 8161 | * and the dirty bit won't be cleared while we are here. |
| 8160 | * | 8162 | * |
| 8161 | * Wait for IO on this page so that we can safely clear | 8163 | * Wait for IO on this page so that we can safely clear |
| 8162 | * the PagePrivate2 bit and do ordered accounting | 8164 | * the PagePrivate2 bit and do ordered accounting |
| 8163 | */ | 8165 | */ |
| 8164 | wait_on_page_writeback(page); | 8166 | wait_on_page_writeback(page); |
| 8165 | 8167 | ||
| 8166 | tree = &BTRFS_I(inode)->io_tree; | 8168 | tree = &BTRFS_I(inode)->io_tree; |
| 8167 | if (offset) { | 8169 | if (offset) { |
| 8168 | btrfs_releasepage(page, GFP_NOFS); | 8170 | btrfs_releasepage(page, GFP_NOFS); |
| 8169 | return; | 8171 | return; |
| 8170 | } | 8172 | } |
| 8171 | 8173 | ||
| 8172 | if (!inode_evicting) | 8174 | if (!inode_evicting) |
| 8173 | lock_extent_bits(tree, page_start, page_end, 0, &cached_state); | 8175 | lock_extent_bits(tree, page_start, page_end, 0, &cached_state); |
| 8174 | ordered = btrfs_lookup_ordered_extent(inode, page_start); | 8176 | ordered = btrfs_lookup_ordered_extent(inode, page_start); |
| 8175 | if (ordered) { | 8177 | if (ordered) { |
| 8176 | /* | 8178 | /* |
| 8177 | * IO on this page will never be started, so we need | 8179 | * IO on this page will never be started, so we need |
| 8178 | * to account for any ordered extents now | 8180 | * to account for any ordered extents now |
| 8179 | */ | 8181 | */ |
| 8180 | if (!inode_evicting) | 8182 | if (!inode_evicting) |
| 8181 | clear_extent_bit(tree, page_start, page_end, | 8183 | clear_extent_bit(tree, page_start, page_end, |
| 8182 | EXTENT_DIRTY | EXTENT_DELALLOC | | 8184 | EXTENT_DIRTY | EXTENT_DELALLOC | |
| 8183 | EXTENT_LOCKED | EXTENT_DO_ACCOUNTING | | 8185 | EXTENT_LOCKED | EXTENT_DO_ACCOUNTING | |
| 8184 | EXTENT_DEFRAG, 1, 0, &cached_state, | 8186 | EXTENT_DEFRAG, 1, 0, &cached_state, |
| 8185 | GFP_NOFS); | 8187 | GFP_NOFS); |
| 8186 | /* | 8188 | /* |
| 8187 | * whoever cleared the private bit is responsible | 8189 | * whoever cleared the private bit is responsible |
| 8188 | * for the finish_ordered_io | 8190 | * for the finish_ordered_io |
| 8189 | */ | 8191 | */ |
| 8190 | if (TestClearPagePrivate2(page)) { | 8192 | if (TestClearPagePrivate2(page)) { |
| 8191 | struct btrfs_ordered_inode_tree *tree; | 8193 | struct btrfs_ordered_inode_tree *tree; |
| 8192 | u64 new_len; | 8194 | u64 new_len; |
| 8193 | 8195 | ||
| 8194 | tree = &BTRFS_I(inode)->ordered_tree; | 8196 | tree = &BTRFS_I(inode)->ordered_tree; |
| 8195 | 8197 | ||
| 8196 | spin_lock_irq(&tree->lock); | 8198 | spin_lock_irq(&tree->lock); |
| 8197 | set_bit(BTRFS_ORDERED_TRUNCATED, &ordered->flags); | 8199 | set_bit(BTRFS_ORDERED_TRUNCATED, &ordered->flags); |
| 8198 | new_len = page_start - ordered->file_offset; | 8200 | new_len = page_start - ordered->file_offset; |
| 8199 | if (new_len < ordered->truncated_len) | 8201 | if (new_len < ordered->truncated_len) |
| 8200 | ordered->truncated_len = new_len; | 8202 | ordered->truncated_len = new_len; |
| 8201 | spin_unlock_irq(&tree->lock); | 8203 | spin_unlock_irq(&tree->lock); |
| 8202 | 8204 | ||
| 8203 | if (btrfs_dec_test_ordered_pending(inode, &ordered, | 8205 | if (btrfs_dec_test_ordered_pending(inode, &ordered, |
| 8204 | page_start, | 8206 | page_start, |
| 8205 | PAGE_CACHE_SIZE, 1)) | 8207 | PAGE_CACHE_SIZE, 1)) |
| 8206 | btrfs_finish_ordered_io(ordered); | 8208 | btrfs_finish_ordered_io(ordered); |
| 8207 | } | 8209 | } |
| 8208 | btrfs_put_ordered_extent(ordered); | 8210 | btrfs_put_ordered_extent(ordered); |
| 8209 | if (!inode_evicting) { | 8211 | if (!inode_evicting) { |
| 8210 | cached_state = NULL; | 8212 | cached_state = NULL; |
| 8211 | lock_extent_bits(tree, page_start, page_end, 0, | 8213 | lock_extent_bits(tree, page_start, page_end, 0, |
| 8212 | &cached_state); | 8214 | &cached_state); |
| 8213 | } | 8215 | } |
| 8214 | } | 8216 | } |
| 8215 | 8217 | ||
| 8216 | if (!inode_evicting) { | 8218 | if (!inode_evicting) { |
| 8217 | clear_extent_bit(tree, page_start, page_end, | 8219 | clear_extent_bit(tree, page_start, page_end, |
| 8218 | EXTENT_LOCKED | EXTENT_DIRTY | | 8220 | EXTENT_LOCKED | EXTENT_DIRTY | |
| 8219 | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | | 8221 | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING | |
| 8220 | EXTENT_DEFRAG, 1, 1, | 8222 | EXTENT_DEFRAG, 1, 1, |
| 8221 | &cached_state, GFP_NOFS); | 8223 | &cached_state, GFP_NOFS); |
| 8222 | 8224 | ||
| 8223 | __btrfs_releasepage(page, GFP_NOFS); | 8225 | __btrfs_releasepage(page, GFP_NOFS); |
| 8224 | } | 8226 | } |
| 8225 | 8227 | ||
| 8226 | ClearPageChecked(page); | 8228 | ClearPageChecked(page); |
| 8227 | if (PagePrivate(page)) { | 8229 | if (PagePrivate(page)) { |
| 8228 | ClearPagePrivate(page); | 8230 | ClearPagePrivate(page); |
| 8229 | set_page_private(page, 0); | 8231 | set_page_private(page, 0); |
| 8230 | page_cache_release(page); | 8232 | page_cache_release(page); |
| 8231 | } | 8233 | } |
| 8232 | } | 8234 | } |
| 8233 | 8235 | ||
| 8234 | /* | 8236 | /* |
| 8235 | * btrfs_page_mkwrite() is not allowed to change the file size as it gets | 8237 | * btrfs_page_mkwrite() is not allowed to change the file size as it gets |
| 8236 | * called from a page fault handler when a page is first dirtied. Hence we must | 8238 | * called from a page fault handler when a page is first dirtied. Hence we must |
| 8237 | * be careful to check for EOF conditions here. We set the page up correctly | 8239 | * be careful to check for EOF conditions here. We set the page up correctly |
| 8238 | * for a written page which means we get ENOSPC checking when writing into | 8240 | * for a written page which means we get ENOSPC checking when writing into |
| 8239 | * holes and correct delalloc and unwritten extent mapping on filesystems that | 8241 | * holes and correct delalloc and unwritten extent mapping on filesystems that |
| 8240 | * support these features. | 8242 | * support these features. |
| 8241 | * | 8243 | * |
| 8242 | * We are not allowed to take the i_mutex here so we have to play games to | 8244 | * We are not allowed to take the i_mutex here so we have to play games to |
| 8243 | * protect against truncate races as the page could now be beyond EOF. Because | 8245 | * protect against truncate races as the page could now be beyond EOF. Because |
| 8244 | * vmtruncate() writes the inode size before removing pages, once we have the | 8246 | * vmtruncate() writes the inode size before removing pages, once we have the |
| 8245 | * page lock we can determine safely if the page is beyond EOF. If it is not | 8247 | * page lock we can determine safely if the page is beyond EOF. If it is not |
| 8246 | * beyond EOF, then the page is guaranteed safe against truncation until we | 8248 | * beyond EOF, then the page is guaranteed safe against truncation until we |
| 8247 | * unlock the page. | 8249 | * unlock the page. |
| 8248 | */ | 8250 | */ |
| 8249 | int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) | 8251 | int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf) |
| 8250 | { | 8252 | { |
| 8251 | struct page *page = vmf->page; | 8253 | struct page *page = vmf->page; |
| 8252 | struct inode *inode = file_inode(vma->vm_file); | 8254 | struct inode *inode = file_inode(vma->vm_file); |
| 8253 | struct btrfs_root *root = BTRFS_I(inode)->root; | 8255 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 8254 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; | 8256 | struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; |
| 8255 | struct btrfs_ordered_extent *ordered; | 8257 | struct btrfs_ordered_extent *ordered; |
| 8256 | struct extent_state *cached_state = NULL; | 8258 | struct extent_state *cached_state = NULL; |
| 8257 | char *kaddr; | 8259 | char *kaddr; |
| 8258 | unsigned long zero_start; | 8260 | unsigned long zero_start; |
| 8259 | loff_t size; | 8261 | loff_t size; |
| 8260 | int ret; | 8262 | int ret; |
| 8261 | int reserved = 0; | 8263 | int reserved = 0; |
| 8262 | u64 page_start; | 8264 | u64 page_start; |
| 8263 | u64 page_end; | 8265 | u64 page_end; |
| 8264 | 8266 | ||
| 8265 | sb_start_pagefault(inode->i_sb); | 8267 | sb_start_pagefault(inode->i_sb); |
| 8266 | ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); | 8268 | ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); |
| 8267 | if (!ret) { | 8269 | if (!ret) { |
| 8268 | ret = file_update_time(vma->vm_file); | 8270 | ret = file_update_time(vma->vm_file); |
| 8269 | reserved = 1; | 8271 | reserved = 1; |
| 8270 | } | 8272 | } |
| 8271 | if (ret) { | 8273 | if (ret) { |
| 8272 | if (ret == -ENOMEM) | 8274 | if (ret == -ENOMEM) |
| 8273 | ret = VM_FAULT_OOM; | 8275 | ret = VM_FAULT_OOM; |
| 8274 | else /* -ENOSPC, -EIO, etc */ | 8276 | else /* -ENOSPC, -EIO, etc */ |
| 8275 | ret = VM_FAULT_SIGBUS; | 8277 | ret = VM_FAULT_SIGBUS; |
| 8276 | if (reserved) | 8278 | if (reserved) |
| 8277 | goto out; | 8279 | goto out; |
| 8278 | goto out_noreserve; | 8280 | goto out_noreserve; |
| 8279 | } | 8281 | } |
| 8280 | 8282 | ||
| 8281 | ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */ | 8283 | ret = VM_FAULT_NOPAGE; /* make the VM retry the fault */ |
| 8282 | again: | 8284 | again: |
| 8283 | lock_page(page); | 8285 | lock_page(page); |
| 8284 | size = i_size_read(inode); | 8286 | size = i_size_read(inode); |
| 8285 | page_start = page_offset(page); | 8287 | page_start = page_offset(page); |
| 8286 | page_end = page_start + PAGE_CACHE_SIZE - 1; | 8288 | page_end = page_start + PAGE_CACHE_SIZE - 1; |
| 8287 | 8289 | ||
| 8288 | if ((page->mapping != inode->i_mapping) || | 8290 | if ((page->mapping != inode->i_mapping) || |
| 8289 | (page_start >= size)) { | 8291 | (page_start >= size)) { |
| 8290 | /* page got truncated out from underneath us */ | 8292 | /* page got truncated out from underneath us */ |
| 8291 | goto out_unlock; | 8293 | goto out_unlock; |
| 8292 | } | 8294 | } |
| 8293 | wait_on_page_writeback(page); | 8295 | wait_on_page_writeback(page); |
| 8294 | 8296 | ||
| 8295 | lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state); | 8297 | lock_extent_bits(io_tree, page_start, page_end, 0, &cached_state); |
| 8296 | set_page_extent_mapped(page); | 8298 | set_page_extent_mapped(page); |
| 8297 | 8299 | ||
| 8298 | /* | 8300 | /* |
| 8299 | * we can't set the delalloc bits if there are pending ordered | 8301 | * we can't set the delalloc bits if there are pending ordered |
| 8300 | * extents. Drop our locks and wait for them to finish | 8302 | * extents. Drop our locks and wait for them to finish |
| 8301 | */ | 8303 | */ |
| 8302 | ordered = btrfs_lookup_ordered_extent(inode, page_start); | 8304 | ordered = btrfs_lookup_ordered_extent(inode, page_start); |
| 8303 | if (ordered) { | 8305 | if (ordered) { |
| 8304 | unlock_extent_cached(io_tree, page_start, page_end, | 8306 | unlock_extent_cached(io_tree, page_start, page_end, |
| 8305 | &cached_state, GFP_NOFS); | 8307 | &cached_state, GFP_NOFS); |
| 8306 | unlock_page(page); | 8308 | unlock_page(page); |
| 8307 | btrfs_start_ordered_extent(inode, ordered, 1); | 8309 | btrfs_start_ordered_extent(inode, ordered, 1); |
| 8308 | btrfs_put_ordered_extent(ordered); | 8310 | btrfs_put_ordered_extent(ordered); |
| 8309 | goto again; | 8311 | goto again; |
| 8310 | } | 8312 | } |
| 8311 | 8313 | ||
| 8312 | /* | 8314 | /* |
| 8313 | * XXX - page_mkwrite gets called every time the page is dirtied, even | 8315 | * XXX - page_mkwrite gets called every time the page is dirtied, even |
| 8314 | * if it was already dirty, so for space accounting reasons we need to | 8316 | * if it was already dirty, so for space accounting reasons we need to |
| 8315 | * clear any delalloc bits for the range we are fixing to save. There | 8317 | * clear any delalloc bits for the range we are fixing to save. There |
| 8316 | * is probably a better way to do this, but for now keep consistent with | 8318 | * is probably a better way to do this, but for now keep consistent with |
| 8317 | * prepare_pages in the normal write path. | 8319 | * prepare_pages in the normal write path. |
| 8318 | */ | 8320 | */ |
| 8319 | clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end, | 8321 | clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, page_end, |
| 8320 | EXTENT_DIRTY | EXTENT_DELALLOC | | 8322 | EXTENT_DIRTY | EXTENT_DELALLOC | |
| 8321 | EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, | 8323 | EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, |
| 8322 | 0, 0, &cached_state, GFP_NOFS); | 8324 | 0, 0, &cached_state, GFP_NOFS); |
| 8323 | 8325 | ||
| 8324 | ret = btrfs_set_extent_delalloc(inode, page_start, page_end, | 8326 | ret = btrfs_set_extent_delalloc(inode, page_start, page_end, |
| 8325 | &cached_state); | 8327 | &cached_state); |
| 8326 | if (ret) { | 8328 | if (ret) { |
| 8327 | unlock_extent_cached(io_tree, page_start, page_end, | 8329 | unlock_extent_cached(io_tree, page_start, page_end, |
| 8328 | &cached_state, GFP_NOFS); | 8330 | &cached_state, GFP_NOFS); |
| 8329 | ret = VM_FAULT_SIGBUS; | 8331 | ret = VM_FAULT_SIGBUS; |
| 8330 | goto out_unlock; | 8332 | goto out_unlock; |
| 8331 | } | 8333 | } |
| 8332 | ret = 0; | 8334 | ret = 0; |
| 8333 | 8335 | ||
| 8334 | /* page is wholly or partially inside EOF */ | 8336 | /* page is wholly or partially inside EOF */ |
| 8335 | if (page_start + PAGE_CACHE_SIZE > size) | 8337 | if (page_start + PAGE_CACHE_SIZE > size) |
| 8336 | zero_start = size & ~PAGE_CACHE_MASK; | 8338 | zero_start = size & ~PAGE_CACHE_MASK; |
| 8337 | else | 8339 | else |
| 8338 | zero_start = PAGE_CACHE_SIZE; | 8340 | zero_start = PAGE_CACHE_SIZE; |
| 8339 | 8341 | ||
| 8340 | if (zero_start != PAGE_CACHE_SIZE) { | 8342 | if (zero_start != PAGE_CACHE_SIZE) { |
| 8341 | kaddr = kmap(page); | 8343 | kaddr = kmap(page); |
| 8342 | memset(kaddr + zero_start, 0, PAGE_CACHE_SIZE - zero_start); | 8344 | memset(kaddr + zero_start, 0, PAGE_CACHE_SIZE - zero_start); |
| 8343 | flush_dcache_page(page); | 8345 | flush_dcache_page(page); |
| 8344 | kunmap(page); | 8346 | kunmap(page); |
| 8345 | } | 8347 | } |
| 8346 | ClearPageChecked(page); | 8348 | ClearPageChecked(page); |
| 8347 | set_page_dirty(page); | 8349 | set_page_dirty(page); |
| 8348 | SetPageUptodate(page); | 8350 | SetPageUptodate(page); |
| 8349 | 8351 | ||
| 8350 | BTRFS_I(inode)->last_trans = root->fs_info->generation; | 8352 | BTRFS_I(inode)->last_trans = root->fs_info->generation; |
| 8351 | BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid; | 8353 | BTRFS_I(inode)->last_sub_trans = BTRFS_I(inode)->root->log_transid; |
| 8352 | BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->root->last_log_commit; | 8354 | BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->root->last_log_commit; |
| 8353 | 8355 | ||
| 8354 | unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS); | 8356 | unlock_extent_cached(io_tree, page_start, page_end, &cached_state, GFP_NOFS); |
| 8355 | 8357 | ||
| 8356 | out_unlock: | 8358 | out_unlock: |
| 8357 | if (!ret) { | 8359 | if (!ret) { |
| 8358 | sb_end_pagefault(inode->i_sb); | 8360 | sb_end_pagefault(inode->i_sb); |
| 8359 | return VM_FAULT_LOCKED; | 8361 | return VM_FAULT_LOCKED; |
| 8360 | } | 8362 | } |
| 8361 | unlock_page(page); | 8363 | unlock_page(page); |
| 8362 | out: | 8364 | out: |
| 8363 | btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); | 8365 | btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); |
| 8364 | out_noreserve: | 8366 | out_noreserve: |
| 8365 | sb_end_pagefault(inode->i_sb); | 8367 | sb_end_pagefault(inode->i_sb); |
| 8366 | return ret; | 8368 | return ret; |
| 8367 | } | 8369 | } |
| 8368 | 8370 | ||
| 8369 | static int btrfs_truncate(struct inode *inode) | 8371 | static int btrfs_truncate(struct inode *inode) |
| 8370 | { | 8372 | { |
| 8371 | struct btrfs_root *root = BTRFS_I(inode)->root; | 8373 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 8372 | struct btrfs_block_rsv *rsv; | 8374 | struct btrfs_block_rsv *rsv; |
| 8373 | int ret = 0; | 8375 | int ret = 0; |
| 8374 | int err = 0; | 8376 | int err = 0; |
| 8375 | struct btrfs_trans_handle *trans; | 8377 | struct btrfs_trans_handle *trans; |
| 8376 | u64 mask = root->sectorsize - 1; | 8378 | u64 mask = root->sectorsize - 1; |
| 8377 | u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); | 8379 | u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); |
| 8378 | 8380 | ||
| 8379 | ret = btrfs_wait_ordered_range(inode, inode->i_size & (~mask), | 8381 | ret = btrfs_wait_ordered_range(inode, inode->i_size & (~mask), |
| 8380 | (u64)-1); | 8382 | (u64)-1); |
| 8381 | if (ret) | 8383 | if (ret) |
| 8382 | return ret; | 8384 | return ret; |
| 8383 | 8385 | ||
| 8384 | /* | 8386 | /* |
| 8385 | * Yes ladies and gentelment, this is indeed ugly. The fact is we have | 8387 | * Yes ladies and gentelment, this is indeed ugly. The fact is we have |
| 8386 | * 3 things going on here | 8388 | * 3 things going on here |
| 8387 | * | 8389 | * |
| 8388 | * 1) We need to reserve space for our orphan item and the space to | 8390 | * 1) We need to reserve space for our orphan item and the space to |
| 8389 | * delete our orphan item. Lord knows we don't want to have a dangling | 8391 | * delete our orphan item. Lord knows we don't want to have a dangling |
| 8390 | * orphan item because we didn't reserve space to remove it. | 8392 | * orphan item because we didn't reserve space to remove it. |
| 8391 | * | 8393 | * |
| 8392 | * 2) We need to reserve space to update our inode. | 8394 | * 2) We need to reserve space to update our inode. |
| 8393 | * | 8395 | * |
| 8394 | * 3) We need to have something to cache all the space that is going to | 8396 | * 3) We need to have something to cache all the space that is going to |
| 8395 | * be free'd up by the truncate operation, but also have some slack | 8397 | * be free'd up by the truncate operation, but also have some slack |
| 8396 | * space reserved in case it uses space during the truncate (thank you | 8398 | * space reserved in case it uses space during the truncate (thank you |
| 8397 | * very much snapshotting). | 8399 | * very much snapshotting). |
| 8398 | * | 8400 | * |
| 8399 | * And we need these to all be seperate. The fact is we can use alot of | 8401 | * And we need these to all be seperate. The fact is we can use alot of |
| 8400 | * space doing the truncate, and we have no earthly idea how much space | 8402 | * space doing the truncate, and we have no earthly idea how much space |
| 8401 | * we will use, so we need the truncate reservation to be seperate so it | 8403 | * we will use, so we need the truncate reservation to be seperate so it |
| 8402 | * doesn't end up using space reserved for updating the inode or | 8404 | * doesn't end up using space reserved for updating the inode or |
| 8403 | * removing the orphan item. We also need to be able to stop the | 8405 | * removing the orphan item. We also need to be able to stop the |
| 8404 | * transaction and start a new one, which means we need to be able to | 8406 | * transaction and start a new one, which means we need to be able to |
| 8405 | * update the inode several times, and we have no idea of knowing how | 8407 | * update the inode several times, and we have no idea of knowing how |
| 8406 | * many times that will be, so we can't just reserve 1 item for the | 8408 | * many times that will be, so we can't just reserve 1 item for the |
| 8407 | * entirety of the opration, so that has to be done seperately as well. | 8409 | * entirety of the opration, so that has to be done seperately as well. |
| 8408 | * Then there is the orphan item, which does indeed need to be held on | 8410 | * Then there is the orphan item, which does indeed need to be held on |
| 8409 | * to for the whole operation, and we need nobody to touch this reserved | 8411 | * to for the whole operation, and we need nobody to touch this reserved |
| 8410 | * space except the orphan code. | 8412 | * space except the orphan code. |
| 8411 | * | 8413 | * |
| 8412 | * So that leaves us with | 8414 | * So that leaves us with |
| 8413 | * | 8415 | * |
| 8414 | * 1) root->orphan_block_rsv - for the orphan deletion. | 8416 | * 1) root->orphan_block_rsv - for the orphan deletion. |
| 8415 | * 2) rsv - for the truncate reservation, which we will steal from the | 8417 | * 2) rsv - for the truncate reservation, which we will steal from the |
| 8416 | * transaction reservation. | 8418 | * transaction reservation. |
| 8417 | * 3) fs_info->trans_block_rsv - this will have 1 items worth left for | 8419 | * 3) fs_info->trans_block_rsv - this will have 1 items worth left for |
| 8418 | * updating the inode. | 8420 | * updating the inode. |
| 8419 | */ | 8421 | */ |
| 8420 | rsv = btrfs_alloc_block_rsv(root, BTRFS_BLOCK_RSV_TEMP); | 8422 | rsv = btrfs_alloc_block_rsv(root, BTRFS_BLOCK_RSV_TEMP); |
| 8421 | if (!rsv) | 8423 | if (!rsv) |
| 8422 | return -ENOMEM; | 8424 | return -ENOMEM; |
| 8423 | rsv->size = min_size; | 8425 | rsv->size = min_size; |
| 8424 | rsv->failfast = 1; | 8426 | rsv->failfast = 1; |
| 8425 | 8427 | ||
| 8426 | /* | 8428 | /* |
| 8427 | * 1 for the truncate slack space | 8429 | * 1 for the truncate slack space |
| 8428 | * 1 for updating the inode. | 8430 | * 1 for updating the inode. |
| 8429 | */ | 8431 | */ |
| 8430 | trans = btrfs_start_transaction(root, 2); | 8432 | trans = btrfs_start_transaction(root, 2); |
| 8431 | if (IS_ERR(trans)) { | 8433 | if (IS_ERR(trans)) { |
| 8432 | err = PTR_ERR(trans); | 8434 | err = PTR_ERR(trans); |
| 8433 | goto out; | 8435 | goto out; |
| 8434 | } | 8436 | } |
| 8435 | 8437 | ||
| 8436 | /* Migrate the slack space for the truncate to our reserve */ | 8438 | /* Migrate the slack space for the truncate to our reserve */ |
| 8437 | ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, | 8439 | ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, rsv, |
| 8438 | min_size); | 8440 | min_size); |
| 8439 | BUG_ON(ret); | 8441 | BUG_ON(ret); |
| 8440 | 8442 | ||
| 8441 | /* | 8443 | /* |
| 8442 | * So if we truncate and then write and fsync we normally would just | 8444 | * So if we truncate and then write and fsync we normally would just |
| 8443 | * write the extents that changed, which is a problem if we need to | 8445 | * write the extents that changed, which is a problem if we need to |
| 8444 | * first truncate that entire inode. So set this flag so we write out | 8446 | * first truncate that entire inode. So set this flag so we write out |
| 8445 | * all of the extents in the inode to the sync log so we're completely | 8447 | * all of the extents in the inode to the sync log so we're completely |
| 8446 | * safe. | 8448 | * safe. |
| 8447 | */ | 8449 | */ |
| 8448 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags); | 8450 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &BTRFS_I(inode)->runtime_flags); |
| 8449 | trans->block_rsv = rsv; | 8451 | trans->block_rsv = rsv; |
| 8450 | 8452 | ||
| 8451 | while (1) { | 8453 | while (1) { |
| 8452 | ret = btrfs_truncate_inode_items(trans, root, inode, | 8454 | ret = btrfs_truncate_inode_items(trans, root, inode, |
| 8453 | inode->i_size, | 8455 | inode->i_size, |
| 8454 | BTRFS_EXTENT_DATA_KEY); | 8456 | BTRFS_EXTENT_DATA_KEY); |
| 8455 | if (ret != -ENOSPC) { | 8457 | if (ret != -ENOSPC) { |
| 8456 | err = ret; | 8458 | err = ret; |
| 8457 | break; | 8459 | break; |
| 8458 | } | 8460 | } |
| 8459 | 8461 | ||
| 8460 | trans->block_rsv = &root->fs_info->trans_block_rsv; | 8462 | trans->block_rsv = &root->fs_info->trans_block_rsv; |
| 8461 | ret = btrfs_update_inode(trans, root, inode); | 8463 | ret = btrfs_update_inode(trans, root, inode); |
| 8462 | if (ret) { | 8464 | if (ret) { |
| 8463 | err = ret; | 8465 | err = ret; |
| 8464 | break; | 8466 | break; |
| 8465 | } | 8467 | } |
| 8466 | 8468 | ||
| 8467 | btrfs_end_transaction(trans, root); | 8469 | btrfs_end_transaction(trans, root); |
| 8468 | btrfs_btree_balance_dirty(root); | 8470 | btrfs_btree_balance_dirty(root); |
| 8469 | 8471 | ||
| 8470 | trans = btrfs_start_transaction(root, 2); | 8472 | trans = btrfs_start_transaction(root, 2); |
| 8471 | if (IS_ERR(trans)) { | 8473 | if (IS_ERR(trans)) { |
| 8472 | ret = err = PTR_ERR(trans); | 8474 | ret = err = PTR_ERR(trans); |
| 8473 | trans = NULL; | 8475 | trans = NULL; |
| 8474 | break; | 8476 | break; |
| 8475 | } | 8477 | } |
| 8476 | 8478 | ||
| 8477 | ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, | 8479 | ret = btrfs_block_rsv_migrate(&root->fs_info->trans_block_rsv, |
| 8478 | rsv, min_size); | 8480 | rsv, min_size); |
| 8479 | BUG_ON(ret); /* shouldn't happen */ | 8481 | BUG_ON(ret); /* shouldn't happen */ |
| 8480 | trans->block_rsv = rsv; | 8482 | trans->block_rsv = rsv; |
| 8481 | } | 8483 | } |
| 8482 | 8484 | ||
| 8483 | if (ret == 0 && inode->i_nlink > 0) { | 8485 | if (ret == 0 && inode->i_nlink > 0) { |
| 8484 | trans->block_rsv = root->orphan_block_rsv; | 8486 | trans->block_rsv = root->orphan_block_rsv; |
| 8485 | ret = btrfs_orphan_del(trans, inode); | 8487 | ret = btrfs_orphan_del(trans, inode); |
| 8486 | if (ret) | 8488 | if (ret) |
| 8487 | err = ret; | 8489 | err = ret; |
| 8488 | } | 8490 | } |
| 8489 | 8491 | ||
| 8490 | if (trans) { | 8492 | if (trans) { |
| 8491 | trans->block_rsv = &root->fs_info->trans_block_rsv; | 8493 | trans->block_rsv = &root->fs_info->trans_block_rsv; |
| 8492 | ret = btrfs_update_inode(trans, root, inode); | 8494 | ret = btrfs_update_inode(trans, root, inode); |
| 8493 | if (ret && !err) | 8495 | if (ret && !err) |
| 8494 | err = ret; | 8496 | err = ret; |
| 8495 | 8497 | ||
| 8496 | ret = btrfs_end_transaction(trans, root); | 8498 | ret = btrfs_end_transaction(trans, root); |
| 8497 | btrfs_btree_balance_dirty(root); | 8499 | btrfs_btree_balance_dirty(root); |
| 8498 | } | 8500 | } |
| 8499 | 8501 | ||
| 8500 | out: | 8502 | out: |
| 8501 | btrfs_free_block_rsv(root, rsv); | 8503 | btrfs_free_block_rsv(root, rsv); |
| 8502 | 8504 | ||
| 8503 | if (ret && !err) | 8505 | if (ret && !err) |
| 8504 | err = ret; | 8506 | err = ret; |
| 8505 | 8507 | ||
| 8506 | return err; | 8508 | return err; |
| 8507 | } | 8509 | } |
| 8508 | 8510 | ||
| 8509 | /* | 8511 | /* |
| 8510 | * create a new subvolume directory/inode (helper for the ioctl). | 8512 | * create a new subvolume directory/inode (helper for the ioctl). |
| 8511 | */ | 8513 | */ |
| 8512 | int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, | 8514 | int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, |
| 8513 | struct btrfs_root *new_root, | 8515 | struct btrfs_root *new_root, |
| 8514 | struct btrfs_root *parent_root, | 8516 | struct btrfs_root *parent_root, |
| 8515 | u64 new_dirid) | 8517 | u64 new_dirid) |
| 8516 | { | 8518 | { |
| 8517 | struct inode *inode; | 8519 | struct inode *inode; |
| 8518 | int err; | 8520 | int err; |
| 8519 | u64 index = 0; | 8521 | u64 index = 0; |
| 8520 | 8522 | ||
| 8521 | inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, | 8523 | inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, |
| 8522 | new_dirid, new_dirid, | 8524 | new_dirid, new_dirid, |
| 8523 | S_IFDIR | (~current_umask() & S_IRWXUGO), | 8525 | S_IFDIR | (~current_umask() & S_IRWXUGO), |
| 8524 | &index); | 8526 | &index); |
| 8525 | if (IS_ERR(inode)) | 8527 | if (IS_ERR(inode)) |
| 8526 | return PTR_ERR(inode); | 8528 | return PTR_ERR(inode); |
| 8527 | inode->i_op = &btrfs_dir_inode_operations; | 8529 | inode->i_op = &btrfs_dir_inode_operations; |
| 8528 | inode->i_fop = &btrfs_dir_file_operations; | 8530 | inode->i_fop = &btrfs_dir_file_operations; |
| 8529 | 8531 | ||
| 8530 | set_nlink(inode, 1); | 8532 | set_nlink(inode, 1); |
| 8531 | btrfs_i_size_write(inode, 0); | 8533 | btrfs_i_size_write(inode, 0); |
| 8532 | unlock_new_inode(inode); | 8534 | unlock_new_inode(inode); |
| 8533 | 8535 | ||
| 8534 | err = btrfs_subvol_inherit_props(trans, new_root, parent_root); | 8536 | err = btrfs_subvol_inherit_props(trans, new_root, parent_root); |
| 8535 | if (err) | 8537 | if (err) |
| 8536 | btrfs_err(new_root->fs_info, | 8538 | btrfs_err(new_root->fs_info, |
| 8537 | "error inheriting subvolume %llu properties: %d", | 8539 | "error inheriting subvolume %llu properties: %d", |
| 8538 | new_root->root_key.objectid, err); | 8540 | new_root->root_key.objectid, err); |
| 8539 | 8541 | ||
| 8540 | err = btrfs_update_inode(trans, new_root, inode); | 8542 | err = btrfs_update_inode(trans, new_root, inode); |
| 8541 | 8543 | ||
| 8542 | iput(inode); | 8544 | iput(inode); |
| 8543 | return err; | 8545 | return err; |
| 8544 | } | 8546 | } |
| 8545 | 8547 | ||
| 8546 | struct inode *btrfs_alloc_inode(struct super_block *sb) | 8548 | struct inode *btrfs_alloc_inode(struct super_block *sb) |
| 8547 | { | 8549 | { |
| 8548 | struct btrfs_inode *ei; | 8550 | struct btrfs_inode *ei; |
| 8549 | struct inode *inode; | 8551 | struct inode *inode; |
| 8550 | 8552 | ||
| 8551 | ei = kmem_cache_alloc(btrfs_inode_cachep, GFP_NOFS); | 8553 | ei = kmem_cache_alloc(btrfs_inode_cachep, GFP_NOFS); |
| 8552 | if (!ei) | 8554 | if (!ei) |
| 8553 | return NULL; | 8555 | return NULL; |
| 8554 | 8556 | ||
| 8555 | ei->root = NULL; | 8557 | ei->root = NULL; |
| 8556 | ei->generation = 0; | 8558 | ei->generation = 0; |
| 8557 | ei->last_trans = 0; | 8559 | ei->last_trans = 0; |
| 8558 | ei->last_sub_trans = 0; | 8560 | ei->last_sub_trans = 0; |
| 8559 | ei->logged_trans = 0; | 8561 | ei->logged_trans = 0; |
| 8560 | ei->delalloc_bytes = 0; | 8562 | ei->delalloc_bytes = 0; |
| 8561 | ei->defrag_bytes = 0; | 8563 | ei->defrag_bytes = 0; |
| 8562 | ei->disk_i_size = 0; | 8564 | ei->disk_i_size = 0; |
| 8563 | ei->flags = 0; | 8565 | ei->flags = 0; |
| 8564 | ei->csum_bytes = 0; | 8566 | ei->csum_bytes = 0; |
| 8565 | ei->index_cnt = (u64)-1; | 8567 | ei->index_cnt = (u64)-1; |
| 8566 | ei->dir_index = 0; | 8568 | ei->dir_index = 0; |
| 8567 | ei->last_unlink_trans = 0; | 8569 | ei->last_unlink_trans = 0; |
| 8568 | ei->last_log_commit = 0; | 8570 | ei->last_log_commit = 0; |
| 8569 | 8571 | ||
| 8570 | spin_lock_init(&ei->lock); | 8572 | spin_lock_init(&ei->lock); |
| 8571 | ei->outstanding_extents = 0; | 8573 | ei->outstanding_extents = 0; |
| 8572 | ei->reserved_extents = 0; | 8574 | ei->reserved_extents = 0; |
| 8573 | 8575 | ||
| 8574 | ei->runtime_flags = 0; | 8576 | ei->runtime_flags = 0; |
| 8575 | ei->force_compress = BTRFS_COMPRESS_NONE; | 8577 | ei->force_compress = BTRFS_COMPRESS_NONE; |
| 8576 | 8578 | ||
| 8577 | ei->delayed_node = NULL; | 8579 | ei->delayed_node = NULL; |
| 8578 | 8580 | ||
| 8579 | inode = &ei->vfs_inode; | 8581 | inode = &ei->vfs_inode; |
| 8580 | extent_map_tree_init(&ei->extent_tree); | 8582 | extent_map_tree_init(&ei->extent_tree); |
| 8581 | extent_io_tree_init(&ei->io_tree, &inode->i_data); | 8583 | extent_io_tree_init(&ei->io_tree, &inode->i_data); |
| 8582 | extent_io_tree_init(&ei->io_failure_tree, &inode->i_data); | 8584 | extent_io_tree_init(&ei->io_failure_tree, &inode->i_data); |
| 8583 | ei->io_tree.track_uptodate = 1; | 8585 | ei->io_tree.track_uptodate = 1; |
| 8584 | ei->io_failure_tree.track_uptodate = 1; | 8586 | ei->io_failure_tree.track_uptodate = 1; |
| 8585 | atomic_set(&ei->sync_writers, 0); | 8587 | atomic_set(&ei->sync_writers, 0); |
| 8586 | mutex_init(&ei->log_mutex); | 8588 | mutex_init(&ei->log_mutex); |
| 8587 | mutex_init(&ei->delalloc_mutex); | 8589 | mutex_init(&ei->delalloc_mutex); |
| 8588 | btrfs_ordered_inode_tree_init(&ei->ordered_tree); | 8590 | btrfs_ordered_inode_tree_init(&ei->ordered_tree); |
| 8589 | INIT_LIST_HEAD(&ei->delalloc_inodes); | 8591 | INIT_LIST_HEAD(&ei->delalloc_inodes); |
| 8590 | RB_CLEAR_NODE(&ei->rb_node); | 8592 | RB_CLEAR_NODE(&ei->rb_node); |
| 8591 | 8593 | ||
| 8592 | return inode; | 8594 | return inode; |
| 8593 | } | 8595 | } |
| 8594 | 8596 | ||
| 8595 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS | 8597 | #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS |
| 8596 | void btrfs_test_destroy_inode(struct inode *inode) | 8598 | void btrfs_test_destroy_inode(struct inode *inode) |
| 8597 | { | 8599 | { |
| 8598 | btrfs_drop_extent_cache(inode, 0, (u64)-1, 0); | 8600 | btrfs_drop_extent_cache(inode, 0, (u64)-1, 0); |
| 8599 | kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode)); | 8601 | kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode)); |
| 8600 | } | 8602 | } |
| 8601 | #endif | 8603 | #endif |
| 8602 | 8604 | ||
| 8603 | static void btrfs_i_callback(struct rcu_head *head) | 8605 | static void btrfs_i_callback(struct rcu_head *head) |
| 8604 | { | 8606 | { |
| 8605 | struct inode *inode = container_of(head, struct inode, i_rcu); | 8607 | struct inode *inode = container_of(head, struct inode, i_rcu); |
| 8606 | kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode)); | 8608 | kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode)); |
| 8607 | } | 8609 | } |
| 8608 | 8610 | ||
| 8609 | void btrfs_destroy_inode(struct inode *inode) | 8611 | void btrfs_destroy_inode(struct inode *inode) |
| 8610 | { | 8612 | { |
| 8611 | struct btrfs_ordered_extent *ordered; | 8613 | struct btrfs_ordered_extent *ordered; |
| 8612 | struct btrfs_root *root = BTRFS_I(inode)->root; | 8614 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 8613 | 8615 | ||
| 8614 | WARN_ON(!hlist_empty(&inode->i_dentry)); | 8616 | WARN_ON(!hlist_empty(&inode->i_dentry)); |
| 8615 | WARN_ON(inode->i_data.nrpages); | 8617 | WARN_ON(inode->i_data.nrpages); |
| 8616 | WARN_ON(BTRFS_I(inode)->outstanding_extents); | 8618 | WARN_ON(BTRFS_I(inode)->outstanding_extents); |
| 8617 | WARN_ON(BTRFS_I(inode)->reserved_extents); | 8619 | WARN_ON(BTRFS_I(inode)->reserved_extents); |
| 8618 | WARN_ON(BTRFS_I(inode)->delalloc_bytes); | 8620 | WARN_ON(BTRFS_I(inode)->delalloc_bytes); |
| 8619 | WARN_ON(BTRFS_I(inode)->csum_bytes); | 8621 | WARN_ON(BTRFS_I(inode)->csum_bytes); |
| 8620 | WARN_ON(BTRFS_I(inode)->defrag_bytes); | 8622 | WARN_ON(BTRFS_I(inode)->defrag_bytes); |
| 8621 | 8623 | ||
| 8622 | /* | 8624 | /* |
| 8623 | * This can happen where we create an inode, but somebody else also | 8625 | * This can happen where we create an inode, but somebody else also |
| 8624 | * created the same inode and we need to destroy the one we already | 8626 | * created the same inode and we need to destroy the one we already |
| 8625 | * created. | 8627 | * created. |
| 8626 | */ | 8628 | */ |
| 8627 | if (!root) | 8629 | if (!root) |
| 8628 | goto free; | 8630 | goto free; |
| 8629 | 8631 | ||
| 8630 | if (test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, | 8632 | if (test_bit(BTRFS_INODE_HAS_ORPHAN_ITEM, |
| 8631 | &BTRFS_I(inode)->runtime_flags)) { | 8633 | &BTRFS_I(inode)->runtime_flags)) { |
| 8632 | btrfs_info(root->fs_info, "inode %llu still on the orphan list", | 8634 | btrfs_info(root->fs_info, "inode %llu still on the orphan list", |
| 8633 | btrfs_ino(inode)); | 8635 | btrfs_ino(inode)); |
| 8634 | atomic_dec(&root->orphan_inodes); | 8636 | atomic_dec(&root->orphan_inodes); |
| 8635 | } | 8637 | } |
| 8636 | 8638 | ||
| 8637 | while (1) { | 8639 | while (1) { |
| 8638 | ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1); | 8640 | ordered = btrfs_lookup_first_ordered_extent(inode, (u64)-1); |
| 8639 | if (!ordered) | 8641 | if (!ordered) |
| 8640 | break; | 8642 | break; |
| 8641 | else { | 8643 | else { |
| 8642 | btrfs_err(root->fs_info, "found ordered extent %llu %llu on inode cleanup", | 8644 | btrfs_err(root->fs_info, "found ordered extent %llu %llu on inode cleanup", |
| 8643 | ordered->file_offset, ordered->len); | 8645 | ordered->file_offset, ordered->len); |
| 8644 | btrfs_remove_ordered_extent(inode, ordered); | 8646 | btrfs_remove_ordered_extent(inode, ordered); |
| 8645 | btrfs_put_ordered_extent(ordered); | 8647 | btrfs_put_ordered_extent(ordered); |
| 8646 | btrfs_put_ordered_extent(ordered); | 8648 | btrfs_put_ordered_extent(ordered); |
| 8647 | } | 8649 | } |
| 8648 | } | 8650 | } |
| 8649 | inode_tree_del(inode); | 8651 | inode_tree_del(inode); |
| 8650 | btrfs_drop_extent_cache(inode, 0, (u64)-1, 0); | 8652 | btrfs_drop_extent_cache(inode, 0, (u64)-1, 0); |
| 8651 | free: | 8653 | free: |
| 8652 | call_rcu(&inode->i_rcu, btrfs_i_callback); | 8654 | call_rcu(&inode->i_rcu, btrfs_i_callback); |
| 8653 | } | 8655 | } |
| 8654 | 8656 | ||
| 8655 | int btrfs_drop_inode(struct inode *inode) | 8657 | int btrfs_drop_inode(struct inode *inode) |
| 8656 | { | 8658 | { |
| 8657 | struct btrfs_root *root = BTRFS_I(inode)->root; | 8659 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 8658 | 8660 | ||
| 8659 | if (root == NULL) | 8661 | if (root == NULL) |
| 8660 | return 1; | 8662 | return 1; |
| 8661 | 8663 | ||
| 8662 | /* the snap/subvol tree is on deleting */ | 8664 | /* the snap/subvol tree is on deleting */ |
| 8663 | if (btrfs_root_refs(&root->root_item) == 0) | 8665 | if (btrfs_root_refs(&root->root_item) == 0) |
| 8664 | return 1; | 8666 | return 1; |
| 8665 | else | 8667 | else |
| 8666 | return generic_drop_inode(inode); | 8668 | return generic_drop_inode(inode); |
| 8667 | } | 8669 | } |
| 8668 | 8670 | ||
| 8669 | static void init_once(void *foo) | 8671 | static void init_once(void *foo) |
| 8670 | { | 8672 | { |
| 8671 | struct btrfs_inode *ei = (struct btrfs_inode *) foo; | 8673 | struct btrfs_inode *ei = (struct btrfs_inode *) foo; |
| 8672 | 8674 | ||
| 8673 | inode_init_once(&ei->vfs_inode); | 8675 | inode_init_once(&ei->vfs_inode); |
| 8674 | } | 8676 | } |
| 8675 | 8677 | ||
| 8676 | void btrfs_destroy_cachep(void) | 8678 | void btrfs_destroy_cachep(void) |
| 8677 | { | 8679 | { |
| 8678 | /* | 8680 | /* |
| 8679 | * Make sure all delayed rcu free inodes are flushed before we | 8681 | * Make sure all delayed rcu free inodes are flushed before we |
| 8680 | * destroy cache. | 8682 | * destroy cache. |
| 8681 | */ | 8683 | */ |
| 8682 | rcu_barrier(); | 8684 | rcu_barrier(); |
| 8683 | if (btrfs_inode_cachep) | 8685 | if (btrfs_inode_cachep) |
| 8684 | kmem_cache_destroy(btrfs_inode_cachep); | 8686 | kmem_cache_destroy(btrfs_inode_cachep); |
| 8685 | if (btrfs_trans_handle_cachep) | 8687 | if (btrfs_trans_handle_cachep) |
| 8686 | kmem_cache_destroy(btrfs_trans_handle_cachep); | 8688 | kmem_cache_destroy(btrfs_trans_handle_cachep); |
| 8687 | if (btrfs_transaction_cachep) | 8689 | if (btrfs_transaction_cachep) |
| 8688 | kmem_cache_destroy(btrfs_transaction_cachep); | 8690 | kmem_cache_destroy(btrfs_transaction_cachep); |
| 8689 | if (btrfs_path_cachep) | 8691 | if (btrfs_path_cachep) |
| 8690 | kmem_cache_destroy(btrfs_path_cachep); | 8692 | kmem_cache_destroy(btrfs_path_cachep); |
| 8691 | if (btrfs_free_space_cachep) | 8693 | if (btrfs_free_space_cachep) |
| 8692 | kmem_cache_destroy(btrfs_free_space_cachep); | 8694 | kmem_cache_destroy(btrfs_free_space_cachep); |
| 8693 | if (btrfs_delalloc_work_cachep) | 8695 | if (btrfs_delalloc_work_cachep) |
| 8694 | kmem_cache_destroy(btrfs_delalloc_work_cachep); | 8696 | kmem_cache_destroy(btrfs_delalloc_work_cachep); |
| 8695 | } | 8697 | } |
| 8696 | 8698 | ||
| 8697 | int btrfs_init_cachep(void) | 8699 | int btrfs_init_cachep(void) |
| 8698 | { | 8700 | { |
| 8699 | btrfs_inode_cachep = kmem_cache_create("btrfs_inode", | 8701 | btrfs_inode_cachep = kmem_cache_create("btrfs_inode", |
| 8700 | sizeof(struct btrfs_inode), 0, | 8702 | sizeof(struct btrfs_inode), 0, |
| 8701 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, init_once); | 8703 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, init_once); |
| 8702 | if (!btrfs_inode_cachep) | 8704 | if (!btrfs_inode_cachep) |
| 8703 | goto fail; | 8705 | goto fail; |
| 8704 | 8706 | ||
| 8705 | btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle", | 8707 | btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle", |
| 8706 | sizeof(struct btrfs_trans_handle), 0, | 8708 | sizeof(struct btrfs_trans_handle), 0, |
| 8707 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); | 8709 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); |
| 8708 | if (!btrfs_trans_handle_cachep) | 8710 | if (!btrfs_trans_handle_cachep) |
| 8709 | goto fail; | 8711 | goto fail; |
| 8710 | 8712 | ||
| 8711 | btrfs_transaction_cachep = kmem_cache_create("btrfs_transaction", | 8713 | btrfs_transaction_cachep = kmem_cache_create("btrfs_transaction", |
| 8712 | sizeof(struct btrfs_transaction), 0, | 8714 | sizeof(struct btrfs_transaction), 0, |
| 8713 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); | 8715 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); |
| 8714 | if (!btrfs_transaction_cachep) | 8716 | if (!btrfs_transaction_cachep) |
| 8715 | goto fail; | 8717 | goto fail; |
| 8716 | 8718 | ||
| 8717 | btrfs_path_cachep = kmem_cache_create("btrfs_path", | 8719 | btrfs_path_cachep = kmem_cache_create("btrfs_path", |
| 8718 | sizeof(struct btrfs_path), 0, | 8720 | sizeof(struct btrfs_path), 0, |
| 8719 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); | 8721 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); |
| 8720 | if (!btrfs_path_cachep) | 8722 | if (!btrfs_path_cachep) |
| 8721 | goto fail; | 8723 | goto fail; |
| 8722 | 8724 | ||
| 8723 | btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space", | 8725 | btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space", |
| 8724 | sizeof(struct btrfs_free_space), 0, | 8726 | sizeof(struct btrfs_free_space), 0, |
| 8725 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); | 8727 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); |
| 8726 | if (!btrfs_free_space_cachep) | 8728 | if (!btrfs_free_space_cachep) |
| 8727 | goto fail; | 8729 | goto fail; |
| 8728 | 8730 | ||
| 8729 | btrfs_delalloc_work_cachep = kmem_cache_create("btrfs_delalloc_work", | 8731 | btrfs_delalloc_work_cachep = kmem_cache_create("btrfs_delalloc_work", |
| 8730 | sizeof(struct btrfs_delalloc_work), 0, | 8732 | sizeof(struct btrfs_delalloc_work), 0, |
| 8731 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, | 8733 | SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, |
| 8732 | NULL); | 8734 | NULL); |
| 8733 | if (!btrfs_delalloc_work_cachep) | 8735 | if (!btrfs_delalloc_work_cachep) |
| 8734 | goto fail; | 8736 | goto fail; |
| 8735 | 8737 | ||
| 8736 | return 0; | 8738 | return 0; |
| 8737 | fail: | 8739 | fail: |
| 8738 | btrfs_destroy_cachep(); | 8740 | btrfs_destroy_cachep(); |
| 8739 | return -ENOMEM; | 8741 | return -ENOMEM; |
| 8740 | } | 8742 | } |
| 8741 | 8743 | ||
| 8742 | static int btrfs_getattr(struct vfsmount *mnt, | 8744 | static int btrfs_getattr(struct vfsmount *mnt, |
| 8743 | struct dentry *dentry, struct kstat *stat) | 8745 | struct dentry *dentry, struct kstat *stat) |
| 8744 | { | 8746 | { |
| 8745 | u64 delalloc_bytes; | 8747 | u64 delalloc_bytes; |
| 8746 | struct inode *inode = dentry->d_inode; | 8748 | struct inode *inode = dentry->d_inode; |
| 8747 | u32 blocksize = inode->i_sb->s_blocksize; | 8749 | u32 blocksize = inode->i_sb->s_blocksize; |
| 8748 | 8750 | ||
| 8749 | generic_fillattr(inode, stat); | 8751 | generic_fillattr(inode, stat); |
| 8750 | stat->dev = BTRFS_I(inode)->root->anon_dev; | 8752 | stat->dev = BTRFS_I(inode)->root->anon_dev; |
| 8751 | stat->blksize = PAGE_CACHE_SIZE; | 8753 | stat->blksize = PAGE_CACHE_SIZE; |
| 8752 | 8754 | ||
| 8753 | spin_lock(&BTRFS_I(inode)->lock); | 8755 | spin_lock(&BTRFS_I(inode)->lock); |
| 8754 | delalloc_bytes = BTRFS_I(inode)->delalloc_bytes; | 8756 | delalloc_bytes = BTRFS_I(inode)->delalloc_bytes; |
| 8755 | spin_unlock(&BTRFS_I(inode)->lock); | 8757 | spin_unlock(&BTRFS_I(inode)->lock); |
| 8756 | stat->blocks = (ALIGN(inode_get_bytes(inode), blocksize) + | 8758 | stat->blocks = (ALIGN(inode_get_bytes(inode), blocksize) + |
| 8757 | ALIGN(delalloc_bytes, blocksize)) >> 9; | 8759 | ALIGN(delalloc_bytes, blocksize)) >> 9; |
| 8758 | return 0; | 8760 | return 0; |
| 8759 | } | 8761 | } |
| 8760 | 8762 | ||
| 8761 | static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, | 8763 | static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, |
| 8762 | struct inode *new_dir, struct dentry *new_dentry) | 8764 | struct inode *new_dir, struct dentry *new_dentry) |
| 8763 | { | 8765 | { |
| 8764 | struct btrfs_trans_handle *trans; | 8766 | struct btrfs_trans_handle *trans; |
| 8765 | struct btrfs_root *root = BTRFS_I(old_dir)->root; | 8767 | struct btrfs_root *root = BTRFS_I(old_dir)->root; |
| 8766 | struct btrfs_root *dest = BTRFS_I(new_dir)->root; | 8768 | struct btrfs_root *dest = BTRFS_I(new_dir)->root; |
| 8767 | struct inode *new_inode = new_dentry->d_inode; | 8769 | struct inode *new_inode = new_dentry->d_inode; |
| 8768 | struct inode *old_inode = old_dentry->d_inode; | 8770 | struct inode *old_inode = old_dentry->d_inode; |
| 8769 | struct timespec ctime = CURRENT_TIME; | 8771 | struct timespec ctime = CURRENT_TIME; |
| 8770 | u64 index = 0; | 8772 | u64 index = 0; |
| 8771 | u64 root_objectid; | 8773 | u64 root_objectid; |
| 8772 | int ret; | 8774 | int ret; |
| 8773 | u64 old_ino = btrfs_ino(old_inode); | 8775 | u64 old_ino = btrfs_ino(old_inode); |
| 8774 | 8776 | ||
| 8775 | if (btrfs_ino(new_dir) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) | 8777 | if (btrfs_ino(new_dir) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) |
| 8776 | return -EPERM; | 8778 | return -EPERM; |
| 8777 | 8779 | ||
| 8778 | /* we only allow rename subvolume link between subvolumes */ | 8780 | /* we only allow rename subvolume link between subvolumes */ |
| 8779 | if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) | 8781 | if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) |
| 8780 | return -EXDEV; | 8782 | return -EXDEV; |
| 8781 | 8783 | ||
| 8782 | if (old_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID || | 8784 | if (old_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID || |
| 8783 | (new_inode && btrfs_ino(new_inode) == BTRFS_FIRST_FREE_OBJECTID)) | 8785 | (new_inode && btrfs_ino(new_inode) == BTRFS_FIRST_FREE_OBJECTID)) |
| 8784 | return -ENOTEMPTY; | 8786 | return -ENOTEMPTY; |
| 8785 | 8787 | ||
| 8786 | if (S_ISDIR(old_inode->i_mode) && new_inode && | 8788 | if (S_ISDIR(old_inode->i_mode) && new_inode && |
| 8787 | new_inode->i_size > BTRFS_EMPTY_DIR_SIZE) | 8789 | new_inode->i_size > BTRFS_EMPTY_DIR_SIZE) |
| 8788 | return -ENOTEMPTY; | 8790 | return -ENOTEMPTY; |
| 8789 | 8791 | ||
| 8790 | 8792 | ||
| 8791 | /* check for collisions, even if the name isn't there */ | 8793 | /* check for collisions, even if the name isn't there */ |
| 8792 | ret = btrfs_check_dir_item_collision(dest, new_dir->i_ino, | 8794 | ret = btrfs_check_dir_item_collision(dest, new_dir->i_ino, |
| 8793 | new_dentry->d_name.name, | 8795 | new_dentry->d_name.name, |
| 8794 | new_dentry->d_name.len); | 8796 | new_dentry->d_name.len); |
| 8795 | 8797 | ||
| 8796 | if (ret) { | 8798 | if (ret) { |
| 8797 | if (ret == -EEXIST) { | 8799 | if (ret == -EEXIST) { |
| 8798 | /* we shouldn't get | 8800 | /* we shouldn't get |
| 8799 | * eexist without a new_inode */ | 8801 | * eexist without a new_inode */ |
| 8800 | if (WARN_ON(!new_inode)) { | 8802 | if (WARN_ON(!new_inode)) { |
| 8801 | return ret; | 8803 | return ret; |
| 8802 | } | 8804 | } |
| 8803 | } else { | 8805 | } else { |
| 8804 | /* maybe -EOVERFLOW */ | 8806 | /* maybe -EOVERFLOW */ |
| 8805 | return ret; | 8807 | return ret; |
| 8806 | } | 8808 | } |
| 8807 | } | 8809 | } |
| 8808 | ret = 0; | 8810 | ret = 0; |
| 8809 | 8811 | ||
| 8810 | /* | 8812 | /* |
| 8811 | * we're using rename to replace one file with another. Start IO on it | 8813 | * we're using rename to replace one file with another. Start IO on it |
| 8812 | * now so we don't add too much work to the end of the transaction | 8814 | * now so we don't add too much work to the end of the transaction |
| 8813 | */ | 8815 | */ |
| 8814 | if (new_inode && S_ISREG(old_inode->i_mode) && new_inode->i_size) | 8816 | if (new_inode && S_ISREG(old_inode->i_mode) && new_inode->i_size) |
| 8815 | filemap_flush(old_inode->i_mapping); | 8817 | filemap_flush(old_inode->i_mapping); |
| 8816 | 8818 | ||
| 8817 | /* close the racy window with snapshot create/destroy ioctl */ | 8819 | /* close the racy window with snapshot create/destroy ioctl */ |
| 8818 | if (old_ino == BTRFS_FIRST_FREE_OBJECTID) | 8820 | if (old_ino == BTRFS_FIRST_FREE_OBJECTID) |
| 8819 | down_read(&root->fs_info->subvol_sem); | 8821 | down_read(&root->fs_info->subvol_sem); |
| 8820 | /* | 8822 | /* |
| 8821 | * We want to reserve the absolute worst case amount of items. So if | 8823 | * We want to reserve the absolute worst case amount of items. So if |
| 8822 | * both inodes are subvols and we need to unlink them then that would | 8824 | * both inodes are subvols and we need to unlink them then that would |
| 8823 | * require 4 item modifications, but if they are both normal inodes it | 8825 | * require 4 item modifications, but if they are both normal inodes it |
| 8824 | * would require 5 item modifications, so we'll assume their normal | 8826 | * would require 5 item modifications, so we'll assume their normal |
| 8825 | * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items | 8827 | * inodes. So 5 * 2 is 10, plus 1 for the new link, so 11 total items |
| 8826 | * should cover the worst case number of items we'll modify. | 8828 | * should cover the worst case number of items we'll modify. |
| 8827 | */ | 8829 | */ |
| 8828 | trans = btrfs_start_transaction(root, 11); | 8830 | trans = btrfs_start_transaction(root, 11); |
| 8829 | if (IS_ERR(trans)) { | 8831 | if (IS_ERR(trans)) { |
| 8830 | ret = PTR_ERR(trans); | 8832 | ret = PTR_ERR(trans); |
| 8831 | goto out_notrans; | 8833 | goto out_notrans; |
| 8832 | } | 8834 | } |
| 8833 | 8835 | ||
| 8834 | if (dest != root) | 8836 | if (dest != root) |
| 8835 | btrfs_record_root_in_trans(trans, dest); | 8837 | btrfs_record_root_in_trans(trans, dest); |
| 8836 | 8838 | ||
| 8837 | ret = btrfs_set_inode_index(new_dir, &index); | 8839 | ret = btrfs_set_inode_index(new_dir, &index); |
| 8838 | if (ret) | 8840 | if (ret) |
| 8839 | goto out_fail; | 8841 | goto out_fail; |
| 8840 | 8842 | ||
| 8841 | BTRFS_I(old_inode)->dir_index = 0ULL; | 8843 | BTRFS_I(old_inode)->dir_index = 0ULL; |
| 8842 | if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { | 8844 | if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { |
| 8843 | /* force full log commit if subvolume involved. */ | 8845 | /* force full log commit if subvolume involved. */ |
| 8844 | btrfs_set_log_full_commit(root->fs_info, trans); | 8846 | btrfs_set_log_full_commit(root->fs_info, trans); |
| 8845 | } else { | 8847 | } else { |
| 8846 | ret = btrfs_insert_inode_ref(trans, dest, | 8848 | ret = btrfs_insert_inode_ref(trans, dest, |
| 8847 | new_dentry->d_name.name, | 8849 | new_dentry->d_name.name, |
| 8848 | new_dentry->d_name.len, | 8850 | new_dentry->d_name.len, |
| 8849 | old_ino, | 8851 | old_ino, |
| 8850 | btrfs_ino(new_dir), index); | 8852 | btrfs_ino(new_dir), index); |
| 8851 | if (ret) | 8853 | if (ret) |
| 8852 | goto out_fail; | 8854 | goto out_fail; |
| 8853 | /* | 8855 | /* |
| 8854 | * this is an ugly little race, but the rename is required | 8856 | * this is an ugly little race, but the rename is required |
| 8855 | * to make sure that if we crash, the inode is either at the | 8857 | * to make sure that if we crash, the inode is either at the |
| 8856 | * old name or the new one. pinning the log transaction lets | 8858 | * old name or the new one. pinning the log transaction lets |
| 8857 | * us make sure we don't allow a log commit to come in after | 8859 | * us make sure we don't allow a log commit to come in after |
| 8858 | * we unlink the name but before we add the new name back in. | 8860 | * we unlink the name but before we add the new name back in. |
| 8859 | */ | 8861 | */ |
| 8860 | btrfs_pin_log_trans(root); | 8862 | btrfs_pin_log_trans(root); |
| 8861 | } | 8863 | } |
| 8862 | 8864 | ||
| 8863 | inode_inc_iversion(old_dir); | 8865 | inode_inc_iversion(old_dir); |
| 8864 | inode_inc_iversion(new_dir); | 8866 | inode_inc_iversion(new_dir); |
| 8865 | inode_inc_iversion(old_inode); | 8867 | inode_inc_iversion(old_inode); |
| 8866 | old_dir->i_ctime = old_dir->i_mtime = ctime; | 8868 | old_dir->i_ctime = old_dir->i_mtime = ctime; |
| 8867 | new_dir->i_ctime = new_dir->i_mtime = ctime; | 8869 | new_dir->i_ctime = new_dir->i_mtime = ctime; |
| 8868 | old_inode->i_ctime = ctime; | 8870 | old_inode->i_ctime = ctime; |
| 8869 | 8871 | ||
| 8870 | if (old_dentry->d_parent != new_dentry->d_parent) | 8872 | if (old_dentry->d_parent != new_dentry->d_parent) |
| 8871 | btrfs_record_unlink_dir(trans, old_dir, old_inode, 1); | 8873 | btrfs_record_unlink_dir(trans, old_dir, old_inode, 1); |
| 8872 | 8874 | ||
| 8873 | if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { | 8875 | if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { |
| 8874 | root_objectid = BTRFS_I(old_inode)->root->root_key.objectid; | 8876 | root_objectid = BTRFS_I(old_inode)->root->root_key.objectid; |
| 8875 | ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid, | 8877 | ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid, |
| 8876 | old_dentry->d_name.name, | 8878 | old_dentry->d_name.name, |
| 8877 | old_dentry->d_name.len); | 8879 | old_dentry->d_name.len); |
| 8878 | } else { | 8880 | } else { |
| 8879 | ret = __btrfs_unlink_inode(trans, root, old_dir, | 8881 | ret = __btrfs_unlink_inode(trans, root, old_dir, |
| 8880 | old_dentry->d_inode, | 8882 | old_dentry->d_inode, |
| 8881 | old_dentry->d_name.name, | 8883 | old_dentry->d_name.name, |
| 8882 | old_dentry->d_name.len); | 8884 | old_dentry->d_name.len); |
| 8883 | if (!ret) | 8885 | if (!ret) |
| 8884 | ret = btrfs_update_inode(trans, root, old_inode); | 8886 | ret = btrfs_update_inode(trans, root, old_inode); |
| 8885 | } | 8887 | } |
| 8886 | if (ret) { | 8888 | if (ret) { |
| 8887 | btrfs_abort_transaction(trans, root, ret); | 8889 | btrfs_abort_transaction(trans, root, ret); |
| 8888 | goto out_fail; | 8890 | goto out_fail; |
| 8889 | } | 8891 | } |
| 8890 | 8892 | ||
| 8891 | if (new_inode) { | 8893 | if (new_inode) { |
| 8892 | inode_inc_iversion(new_inode); | 8894 | inode_inc_iversion(new_inode); |
| 8893 | new_inode->i_ctime = CURRENT_TIME; | 8895 | new_inode->i_ctime = CURRENT_TIME; |
| 8894 | if (unlikely(btrfs_ino(new_inode) == | 8896 | if (unlikely(btrfs_ino(new_inode) == |
| 8895 | BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { | 8897 | BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { |
| 8896 | root_objectid = BTRFS_I(new_inode)->location.objectid; | 8898 | root_objectid = BTRFS_I(new_inode)->location.objectid; |
| 8897 | ret = btrfs_unlink_subvol(trans, dest, new_dir, | 8899 | ret = btrfs_unlink_subvol(trans, dest, new_dir, |
| 8898 | root_objectid, | 8900 | root_objectid, |
| 8899 | new_dentry->d_name.name, | 8901 | new_dentry->d_name.name, |
| 8900 | new_dentry->d_name.len); | 8902 | new_dentry->d_name.len); |
| 8901 | BUG_ON(new_inode->i_nlink == 0); | 8903 | BUG_ON(new_inode->i_nlink == 0); |
| 8902 | } else { | 8904 | } else { |
| 8903 | ret = btrfs_unlink_inode(trans, dest, new_dir, | 8905 | ret = btrfs_unlink_inode(trans, dest, new_dir, |
| 8904 | new_dentry->d_inode, | 8906 | new_dentry->d_inode, |
| 8905 | new_dentry->d_name.name, | 8907 | new_dentry->d_name.name, |
| 8906 | new_dentry->d_name.len); | 8908 | new_dentry->d_name.len); |
| 8907 | } | 8909 | } |
| 8908 | if (!ret && new_inode->i_nlink == 0) | 8910 | if (!ret && new_inode->i_nlink == 0) |
| 8909 | ret = btrfs_orphan_add(trans, new_dentry->d_inode); | 8911 | ret = btrfs_orphan_add(trans, new_dentry->d_inode); |
| 8910 | if (ret) { | 8912 | if (ret) { |
| 8911 | btrfs_abort_transaction(trans, root, ret); | 8913 | btrfs_abort_transaction(trans, root, ret); |
| 8912 | goto out_fail; | 8914 | goto out_fail; |
| 8913 | } | 8915 | } |
| 8914 | } | 8916 | } |
| 8915 | 8917 | ||
| 8916 | ret = btrfs_add_link(trans, new_dir, old_inode, | 8918 | ret = btrfs_add_link(trans, new_dir, old_inode, |
| 8917 | new_dentry->d_name.name, | 8919 | new_dentry->d_name.name, |
| 8918 | new_dentry->d_name.len, 0, index); | 8920 | new_dentry->d_name.len, 0, index); |
| 8919 | if (ret) { | 8921 | if (ret) { |
| 8920 | btrfs_abort_transaction(trans, root, ret); | 8922 | btrfs_abort_transaction(trans, root, ret); |
| 8921 | goto out_fail; | 8923 | goto out_fail; |
| 8922 | } | 8924 | } |
| 8923 | 8925 | ||
| 8924 | if (old_inode->i_nlink == 1) | 8926 | if (old_inode->i_nlink == 1) |
| 8925 | BTRFS_I(old_inode)->dir_index = index; | 8927 | BTRFS_I(old_inode)->dir_index = index; |
| 8926 | 8928 | ||
| 8927 | if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { | 8929 | if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { |
| 8928 | struct dentry *parent = new_dentry->d_parent; | 8930 | struct dentry *parent = new_dentry->d_parent; |
| 8929 | btrfs_log_new_name(trans, old_inode, old_dir, parent); | 8931 | btrfs_log_new_name(trans, old_inode, old_dir, parent); |
| 8930 | btrfs_end_log_trans(root); | 8932 | btrfs_end_log_trans(root); |
| 8931 | } | 8933 | } |
| 8932 | out_fail: | 8934 | out_fail: |
| 8933 | btrfs_end_transaction(trans, root); | 8935 | btrfs_end_transaction(trans, root); |
| 8934 | out_notrans: | 8936 | out_notrans: |
| 8935 | if (old_ino == BTRFS_FIRST_FREE_OBJECTID) | 8937 | if (old_ino == BTRFS_FIRST_FREE_OBJECTID) |
| 8936 | up_read(&root->fs_info->subvol_sem); | 8938 | up_read(&root->fs_info->subvol_sem); |
| 8937 | 8939 | ||
| 8938 | return ret; | 8940 | return ret; |
| 8939 | } | 8941 | } |
| 8940 | 8942 | ||
| 8941 | static int btrfs_rename2(struct inode *old_dir, struct dentry *old_dentry, | 8943 | static int btrfs_rename2(struct inode *old_dir, struct dentry *old_dentry, |
| 8942 | struct inode *new_dir, struct dentry *new_dentry, | 8944 | struct inode *new_dir, struct dentry *new_dentry, |
| 8943 | unsigned int flags) | 8945 | unsigned int flags) |
| 8944 | { | 8946 | { |
| 8945 | if (flags & ~RENAME_NOREPLACE) | 8947 | if (flags & ~RENAME_NOREPLACE) |
| 8946 | return -EINVAL; | 8948 | return -EINVAL; |
| 8947 | 8949 | ||
| 8948 | return btrfs_rename(old_dir, old_dentry, new_dir, new_dentry); | 8950 | return btrfs_rename(old_dir, old_dentry, new_dir, new_dentry); |
| 8949 | } | 8951 | } |
| 8950 | 8952 | ||
| 8951 | static void btrfs_run_delalloc_work(struct btrfs_work *work) | 8953 | static void btrfs_run_delalloc_work(struct btrfs_work *work) |
| 8952 | { | 8954 | { |
| 8953 | struct btrfs_delalloc_work *delalloc_work; | 8955 | struct btrfs_delalloc_work *delalloc_work; |
| 8954 | struct inode *inode; | 8956 | struct inode *inode; |
| 8955 | 8957 | ||
| 8956 | delalloc_work = container_of(work, struct btrfs_delalloc_work, | 8958 | delalloc_work = container_of(work, struct btrfs_delalloc_work, |
| 8957 | work); | 8959 | work); |
| 8958 | inode = delalloc_work->inode; | 8960 | inode = delalloc_work->inode; |
| 8959 | if (delalloc_work->wait) { | 8961 | if (delalloc_work->wait) { |
| 8960 | btrfs_wait_ordered_range(inode, 0, (u64)-1); | 8962 | btrfs_wait_ordered_range(inode, 0, (u64)-1); |
| 8961 | } else { | 8963 | } else { |
| 8962 | filemap_flush(inode->i_mapping); | 8964 | filemap_flush(inode->i_mapping); |
| 8963 | if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, | 8965 | if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT, |
| 8964 | &BTRFS_I(inode)->runtime_flags)) | 8966 | &BTRFS_I(inode)->runtime_flags)) |
| 8965 | filemap_flush(inode->i_mapping); | 8967 | filemap_flush(inode->i_mapping); |
| 8966 | } | 8968 | } |
| 8967 | 8969 | ||
| 8968 | if (delalloc_work->delay_iput) | 8970 | if (delalloc_work->delay_iput) |
| 8969 | btrfs_add_delayed_iput(inode); | 8971 | btrfs_add_delayed_iput(inode); |
| 8970 | else | 8972 | else |
| 8971 | iput(inode); | 8973 | iput(inode); |
| 8972 | complete(&delalloc_work->completion); | 8974 | complete(&delalloc_work->completion); |
| 8973 | } | 8975 | } |
| 8974 | 8976 | ||
| 8975 | struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode, | 8977 | struct btrfs_delalloc_work *btrfs_alloc_delalloc_work(struct inode *inode, |
| 8976 | int wait, int delay_iput) | 8978 | int wait, int delay_iput) |
| 8977 | { | 8979 | { |
| 8978 | struct btrfs_delalloc_work *work; | 8980 | struct btrfs_delalloc_work *work; |
| 8979 | 8981 | ||
| 8980 | work = kmem_cache_zalloc(btrfs_delalloc_work_cachep, GFP_NOFS); | 8982 | work = kmem_cache_zalloc(btrfs_delalloc_work_cachep, GFP_NOFS); |
| 8981 | if (!work) | 8983 | if (!work) |
| 8982 | return NULL; | 8984 | return NULL; |
| 8983 | 8985 | ||
| 8984 | init_completion(&work->completion); | 8986 | init_completion(&work->completion); |
| 8985 | INIT_LIST_HEAD(&work->list); | 8987 | INIT_LIST_HEAD(&work->list); |
| 8986 | work->inode = inode; | 8988 | work->inode = inode; |
| 8987 | work->wait = wait; | 8989 | work->wait = wait; |
| 8988 | work->delay_iput = delay_iput; | 8990 | work->delay_iput = delay_iput; |
| 8989 | WARN_ON_ONCE(!inode); | 8991 | WARN_ON_ONCE(!inode); |
| 8990 | btrfs_init_work(&work->work, btrfs_flush_delalloc_helper, | 8992 | btrfs_init_work(&work->work, btrfs_flush_delalloc_helper, |
| 8991 | btrfs_run_delalloc_work, NULL, NULL); | 8993 | btrfs_run_delalloc_work, NULL, NULL); |
| 8992 | 8994 | ||
| 8993 | return work; | 8995 | return work; |
| 8994 | } | 8996 | } |
| 8995 | 8997 | ||
| 8996 | void btrfs_wait_and_free_delalloc_work(struct btrfs_delalloc_work *work) | 8998 | void btrfs_wait_and_free_delalloc_work(struct btrfs_delalloc_work *work) |
| 8997 | { | 8999 | { |
| 8998 | wait_for_completion(&work->completion); | 9000 | wait_for_completion(&work->completion); |
| 8999 | kmem_cache_free(btrfs_delalloc_work_cachep, work); | 9001 | kmem_cache_free(btrfs_delalloc_work_cachep, work); |
| 9000 | } | 9002 | } |
| 9001 | 9003 | ||
| 9002 | /* | 9004 | /* |
| 9003 | * some fairly slow code that needs optimization. This walks the list | 9005 | * some fairly slow code that needs optimization. This walks the list |
| 9004 | * of all the inodes with pending delalloc and forces them to disk. | 9006 | * of all the inodes with pending delalloc and forces them to disk. |
| 9005 | */ | 9007 | */ |
| 9006 | static int __start_delalloc_inodes(struct btrfs_root *root, int delay_iput, | 9008 | static int __start_delalloc_inodes(struct btrfs_root *root, int delay_iput, |
| 9007 | int nr) | 9009 | int nr) |
| 9008 | { | 9010 | { |
| 9009 | struct btrfs_inode *binode; | 9011 | struct btrfs_inode *binode; |
| 9010 | struct inode *inode; | 9012 | struct inode *inode; |
| 9011 | struct btrfs_delalloc_work *work, *next; | 9013 | struct btrfs_delalloc_work *work, *next; |
| 9012 | struct list_head works; | 9014 | struct list_head works; |
| 9013 | struct list_head splice; | 9015 | struct list_head splice; |
| 9014 | int ret = 0; | 9016 | int ret = 0; |
| 9015 | 9017 | ||
| 9016 | INIT_LIST_HEAD(&works); | 9018 | INIT_LIST_HEAD(&works); |
| 9017 | INIT_LIST_HEAD(&splice); | 9019 | INIT_LIST_HEAD(&splice); |
| 9018 | 9020 | ||
| 9019 | mutex_lock(&root->delalloc_mutex); | 9021 | mutex_lock(&root->delalloc_mutex); |
| 9020 | spin_lock(&root->delalloc_lock); | 9022 | spin_lock(&root->delalloc_lock); |
| 9021 | list_splice_init(&root->delalloc_inodes, &splice); | 9023 | list_splice_init(&root->delalloc_inodes, &splice); |
| 9022 | while (!list_empty(&splice)) { | 9024 | while (!list_empty(&splice)) { |
| 9023 | binode = list_entry(splice.next, struct btrfs_inode, | 9025 | binode = list_entry(splice.next, struct btrfs_inode, |
| 9024 | delalloc_inodes); | 9026 | delalloc_inodes); |
| 9025 | 9027 | ||
| 9026 | list_move_tail(&binode->delalloc_inodes, | 9028 | list_move_tail(&binode->delalloc_inodes, |
| 9027 | &root->delalloc_inodes); | 9029 | &root->delalloc_inodes); |
| 9028 | inode = igrab(&binode->vfs_inode); | 9030 | inode = igrab(&binode->vfs_inode); |
| 9029 | if (!inode) { | 9031 | if (!inode) { |
| 9030 | cond_resched_lock(&root->delalloc_lock); | 9032 | cond_resched_lock(&root->delalloc_lock); |
| 9031 | continue; | 9033 | continue; |
| 9032 | } | 9034 | } |
| 9033 | spin_unlock(&root->delalloc_lock); | 9035 | spin_unlock(&root->delalloc_lock); |
| 9034 | 9036 | ||
| 9035 | work = btrfs_alloc_delalloc_work(inode, 0, delay_iput); | 9037 | work = btrfs_alloc_delalloc_work(inode, 0, delay_iput); |
| 9036 | if (!work) { | 9038 | if (!work) { |
| 9037 | if (delay_iput) | 9039 | if (delay_iput) |
| 9038 | btrfs_add_delayed_iput(inode); | 9040 | btrfs_add_delayed_iput(inode); |
| 9039 | else | 9041 | else |
| 9040 | iput(inode); | 9042 | iput(inode); |
| 9041 | ret = -ENOMEM; | 9043 | ret = -ENOMEM; |
| 9042 | goto out; | 9044 | goto out; |
| 9043 | } | 9045 | } |
| 9044 | list_add_tail(&work->list, &works); | 9046 | list_add_tail(&work->list, &works); |
| 9045 | btrfs_queue_work(root->fs_info->flush_workers, | 9047 | btrfs_queue_work(root->fs_info->flush_workers, |
| 9046 | &work->work); | 9048 | &work->work); |
| 9047 | ret++; | 9049 | ret++; |
| 9048 | if (nr != -1 && ret >= nr) | 9050 | if (nr != -1 && ret >= nr) |
| 9049 | goto out; | 9051 | goto out; |
| 9050 | cond_resched(); | 9052 | cond_resched(); |
| 9051 | spin_lock(&root->delalloc_lock); | 9053 | spin_lock(&root->delalloc_lock); |
| 9052 | } | 9054 | } |
| 9053 | spin_unlock(&root->delalloc_lock); | 9055 | spin_unlock(&root->delalloc_lock); |
| 9054 | 9056 | ||
| 9055 | out: | 9057 | out: |
| 9056 | list_for_each_entry_safe(work, next, &works, list) { | 9058 | list_for_each_entry_safe(work, next, &works, list) { |
| 9057 | list_del_init(&work->list); | 9059 | list_del_init(&work->list); |
| 9058 | btrfs_wait_and_free_delalloc_work(work); | 9060 | btrfs_wait_and_free_delalloc_work(work); |
| 9059 | } | 9061 | } |
| 9060 | 9062 | ||
| 9061 | if (!list_empty_careful(&splice)) { | 9063 | if (!list_empty_careful(&splice)) { |
| 9062 | spin_lock(&root->delalloc_lock); | 9064 | spin_lock(&root->delalloc_lock); |
| 9063 | list_splice_tail(&splice, &root->delalloc_inodes); | 9065 | list_splice_tail(&splice, &root->delalloc_inodes); |
| 9064 | spin_unlock(&root->delalloc_lock); | 9066 | spin_unlock(&root->delalloc_lock); |
| 9065 | } | 9067 | } |
| 9066 | mutex_unlock(&root->delalloc_mutex); | 9068 | mutex_unlock(&root->delalloc_mutex); |
| 9067 | return ret; | 9069 | return ret; |
| 9068 | } | 9070 | } |
| 9069 | 9071 | ||
| 9070 | int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) | 9072 | int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) |
| 9071 | { | 9073 | { |
| 9072 | int ret; | 9074 | int ret; |
| 9073 | 9075 | ||
| 9074 | if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) | 9076 | if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) |
| 9075 | return -EROFS; | 9077 | return -EROFS; |
| 9076 | 9078 | ||
| 9077 | ret = __start_delalloc_inodes(root, delay_iput, -1); | 9079 | ret = __start_delalloc_inodes(root, delay_iput, -1); |
| 9078 | if (ret > 0) | 9080 | if (ret > 0) |
| 9079 | ret = 0; | 9081 | ret = 0; |
| 9080 | /* | 9082 | /* |
| 9081 | * the filemap_flush will queue IO into the worker threads, but | 9083 | * the filemap_flush will queue IO into the worker threads, but |
| 9082 | * we have to make sure the IO is actually started and that | 9084 | * we have to make sure the IO is actually started and that |
| 9083 | * ordered extents get created before we return | 9085 | * ordered extents get created before we return |
| 9084 | */ | 9086 | */ |
| 9085 | atomic_inc(&root->fs_info->async_submit_draining); | 9087 | atomic_inc(&root->fs_info->async_submit_draining); |
| 9086 | while (atomic_read(&root->fs_info->nr_async_submits) || | 9088 | while (atomic_read(&root->fs_info->nr_async_submits) || |
| 9087 | atomic_read(&root->fs_info->async_delalloc_pages)) { | 9089 | atomic_read(&root->fs_info->async_delalloc_pages)) { |
| 9088 | wait_event(root->fs_info->async_submit_wait, | 9090 | wait_event(root->fs_info->async_submit_wait, |
| 9089 | (atomic_read(&root->fs_info->nr_async_submits) == 0 && | 9091 | (atomic_read(&root->fs_info->nr_async_submits) == 0 && |
| 9090 | atomic_read(&root->fs_info->async_delalloc_pages) == 0)); | 9092 | atomic_read(&root->fs_info->async_delalloc_pages) == 0)); |
| 9091 | } | 9093 | } |
| 9092 | atomic_dec(&root->fs_info->async_submit_draining); | 9094 | atomic_dec(&root->fs_info->async_submit_draining); |
| 9093 | return ret; | 9095 | return ret; |
| 9094 | } | 9096 | } |
| 9095 | 9097 | ||
| 9096 | int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, | 9098 | int btrfs_start_delalloc_roots(struct btrfs_fs_info *fs_info, int delay_iput, |
| 9097 | int nr) | 9099 | int nr) |
| 9098 | { | 9100 | { |
| 9099 | struct btrfs_root *root; | 9101 | struct btrfs_root *root; |
| 9100 | struct list_head splice; | 9102 | struct list_head splice; |
| 9101 | int ret; | 9103 | int ret; |
| 9102 | 9104 | ||
| 9103 | if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) | 9105 | if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) |
| 9104 | return -EROFS; | 9106 | return -EROFS; |
| 9105 | 9107 | ||
| 9106 | INIT_LIST_HEAD(&splice); | 9108 | INIT_LIST_HEAD(&splice); |
| 9107 | 9109 | ||
| 9108 | mutex_lock(&fs_info->delalloc_root_mutex); | 9110 | mutex_lock(&fs_info->delalloc_root_mutex); |
| 9109 | spin_lock(&fs_info->delalloc_root_lock); | 9111 | spin_lock(&fs_info->delalloc_root_lock); |
| 9110 | list_splice_init(&fs_info->delalloc_roots, &splice); | 9112 | list_splice_init(&fs_info->delalloc_roots, &splice); |
| 9111 | while (!list_empty(&splice) && nr) { | 9113 | while (!list_empty(&splice) && nr) { |
| 9112 | root = list_first_entry(&splice, struct btrfs_root, | 9114 | root = list_first_entry(&splice, struct btrfs_root, |
| 9113 | delalloc_root); | 9115 | delalloc_root); |
| 9114 | root = btrfs_grab_fs_root(root); | 9116 | root = btrfs_grab_fs_root(root); |
| 9115 | BUG_ON(!root); | 9117 | BUG_ON(!root); |
| 9116 | list_move_tail(&root->delalloc_root, | 9118 | list_move_tail(&root->delalloc_root, |
| 9117 | &fs_info->delalloc_roots); | 9119 | &fs_info->delalloc_roots); |
| 9118 | spin_unlock(&fs_info->delalloc_root_lock); | 9120 | spin_unlock(&fs_info->delalloc_root_lock); |
| 9119 | 9121 | ||
| 9120 | ret = __start_delalloc_inodes(root, delay_iput, nr); | 9122 | ret = __start_delalloc_inodes(root, delay_iput, nr); |
| 9121 | btrfs_put_fs_root(root); | 9123 | btrfs_put_fs_root(root); |
| 9122 | if (ret < 0) | 9124 | if (ret < 0) |
| 9123 | goto out; | 9125 | goto out; |
| 9124 | 9126 | ||
| 9125 | if (nr != -1) { | 9127 | if (nr != -1) { |
| 9126 | nr -= ret; | 9128 | nr -= ret; |
| 9127 | WARN_ON(nr < 0); | 9129 | WARN_ON(nr < 0); |
| 9128 | } | 9130 | } |
| 9129 | spin_lock(&fs_info->delalloc_root_lock); | 9131 | spin_lock(&fs_info->delalloc_root_lock); |
| 9130 | } | 9132 | } |
| 9131 | spin_unlock(&fs_info->delalloc_root_lock); | 9133 | spin_unlock(&fs_info->delalloc_root_lock); |
| 9132 | 9134 | ||
| 9133 | ret = 0; | 9135 | ret = 0; |
| 9134 | atomic_inc(&fs_info->async_submit_draining); | 9136 | atomic_inc(&fs_info->async_submit_draining); |
| 9135 | while (atomic_read(&fs_info->nr_async_submits) || | 9137 | while (atomic_read(&fs_info->nr_async_submits) || |
| 9136 | atomic_read(&fs_info->async_delalloc_pages)) { | 9138 | atomic_read(&fs_info->async_delalloc_pages)) { |
| 9137 | wait_event(fs_info->async_submit_wait, | 9139 | wait_event(fs_info->async_submit_wait, |
| 9138 | (atomic_read(&fs_info->nr_async_submits) == 0 && | 9140 | (atomic_read(&fs_info->nr_async_submits) == 0 && |
| 9139 | atomic_read(&fs_info->async_delalloc_pages) == 0)); | 9141 | atomic_read(&fs_info->async_delalloc_pages) == 0)); |
| 9140 | } | 9142 | } |
| 9141 | atomic_dec(&fs_info->async_submit_draining); | 9143 | atomic_dec(&fs_info->async_submit_draining); |
| 9142 | out: | 9144 | out: |
| 9143 | if (!list_empty_careful(&splice)) { | 9145 | if (!list_empty_careful(&splice)) { |
| 9144 | spin_lock(&fs_info->delalloc_root_lock); | 9146 | spin_lock(&fs_info->delalloc_root_lock); |
| 9145 | list_splice_tail(&splice, &fs_info->delalloc_roots); | 9147 | list_splice_tail(&splice, &fs_info->delalloc_roots); |
| 9146 | spin_unlock(&fs_info->delalloc_root_lock); | 9148 | spin_unlock(&fs_info->delalloc_root_lock); |
| 9147 | } | 9149 | } |
| 9148 | mutex_unlock(&fs_info->delalloc_root_mutex); | 9150 | mutex_unlock(&fs_info->delalloc_root_mutex); |
| 9149 | return ret; | 9151 | return ret; |
| 9150 | } | 9152 | } |
| 9151 | 9153 | ||
| 9152 | static int btrfs_symlink(struct inode *dir, struct dentry *dentry, | 9154 | static int btrfs_symlink(struct inode *dir, struct dentry *dentry, |
| 9153 | const char *symname) | 9155 | const char *symname) |
| 9154 | { | 9156 | { |
| 9155 | struct btrfs_trans_handle *trans; | 9157 | struct btrfs_trans_handle *trans; |
| 9156 | struct btrfs_root *root = BTRFS_I(dir)->root; | 9158 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 9157 | struct btrfs_path *path; | 9159 | struct btrfs_path *path; |
| 9158 | struct btrfs_key key; | 9160 | struct btrfs_key key; |
| 9159 | struct inode *inode = NULL; | 9161 | struct inode *inode = NULL; |
| 9160 | int err; | 9162 | int err; |
| 9161 | int drop_inode = 0; | 9163 | int drop_inode = 0; |
| 9162 | u64 objectid; | 9164 | u64 objectid; |
| 9163 | u64 index = 0; | 9165 | u64 index = 0; |
| 9164 | int name_len; | 9166 | int name_len; |
| 9165 | int datasize; | 9167 | int datasize; |
| 9166 | unsigned long ptr; | 9168 | unsigned long ptr; |
| 9167 | struct btrfs_file_extent_item *ei; | 9169 | struct btrfs_file_extent_item *ei; |
| 9168 | struct extent_buffer *leaf; | 9170 | struct extent_buffer *leaf; |
| 9169 | 9171 | ||
| 9170 | name_len = strlen(symname); | 9172 | name_len = strlen(symname); |
| 9171 | if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root)) | 9173 | if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root)) |
| 9172 | return -ENAMETOOLONG; | 9174 | return -ENAMETOOLONG; |
| 9173 | 9175 | ||
| 9174 | /* | 9176 | /* |
| 9175 | * 2 items for inode item and ref | 9177 | * 2 items for inode item and ref |
| 9176 | * 2 items for dir items | 9178 | * 2 items for dir items |
| 9177 | * 1 item for xattr if selinux is on | 9179 | * 1 item for xattr if selinux is on |
| 9178 | */ | 9180 | */ |
| 9179 | trans = btrfs_start_transaction(root, 5); | 9181 | trans = btrfs_start_transaction(root, 5); |
| 9180 | if (IS_ERR(trans)) | 9182 | if (IS_ERR(trans)) |
| 9181 | return PTR_ERR(trans); | 9183 | return PTR_ERR(trans); |
| 9182 | 9184 | ||
| 9183 | err = btrfs_find_free_ino(root, &objectid); | 9185 | err = btrfs_find_free_ino(root, &objectid); |
| 9184 | if (err) | 9186 | if (err) |
| 9185 | goto out_unlock; | 9187 | goto out_unlock; |
| 9186 | 9188 | ||
| 9187 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, | 9189 | inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, |
| 9188 | dentry->d_name.len, btrfs_ino(dir), objectid, | 9190 | dentry->d_name.len, btrfs_ino(dir), objectid, |
| 9189 | S_IFLNK|S_IRWXUGO, &index); | 9191 | S_IFLNK|S_IRWXUGO, &index); |
| 9190 | if (IS_ERR(inode)) { | 9192 | if (IS_ERR(inode)) { |
| 9191 | err = PTR_ERR(inode); | 9193 | err = PTR_ERR(inode); |
| 9192 | goto out_unlock; | 9194 | goto out_unlock; |
| 9193 | } | 9195 | } |
| 9194 | 9196 | ||
| 9195 | /* | 9197 | /* |
| 9196 | * If the active LSM wants to access the inode during | 9198 | * If the active LSM wants to access the inode during |
| 9197 | * d_instantiate it needs these. Smack checks to see | 9199 | * d_instantiate it needs these. Smack checks to see |
| 9198 | * if the filesystem supports xattrs by looking at the | 9200 | * if the filesystem supports xattrs by looking at the |
| 9199 | * ops vector. | 9201 | * ops vector. |
| 9200 | */ | 9202 | */ |
| 9201 | inode->i_fop = &btrfs_file_operations; | 9203 | inode->i_fop = &btrfs_file_operations; |
| 9202 | inode->i_op = &btrfs_file_inode_operations; | 9204 | inode->i_op = &btrfs_file_inode_operations; |
| 9203 | inode->i_mapping->a_ops = &btrfs_aops; | 9205 | inode->i_mapping->a_ops = &btrfs_aops; |
| 9204 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; | 9206 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; |
| 9205 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; | 9207 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; |
| 9206 | 9208 | ||
| 9207 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); | 9209 | err = btrfs_init_inode_security(trans, inode, dir, &dentry->d_name); |
| 9208 | if (err) | 9210 | if (err) |
| 9209 | goto out_unlock_inode; | 9211 | goto out_unlock_inode; |
| 9210 | 9212 | ||
| 9211 | err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); | 9213 | err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); |
| 9212 | if (err) | 9214 | if (err) |
| 9213 | goto out_unlock_inode; | 9215 | goto out_unlock_inode; |
| 9214 | 9216 | ||
| 9215 | path = btrfs_alloc_path(); | 9217 | path = btrfs_alloc_path(); |
| 9216 | if (!path) { | 9218 | if (!path) { |
| 9217 | err = -ENOMEM; | 9219 | err = -ENOMEM; |
| 9218 | goto out_unlock_inode; | 9220 | goto out_unlock_inode; |
| 9219 | } | 9221 | } |
| 9220 | key.objectid = btrfs_ino(inode); | 9222 | key.objectid = btrfs_ino(inode); |
| 9221 | key.offset = 0; | 9223 | key.offset = 0; |
| 9222 | key.type = BTRFS_EXTENT_DATA_KEY; | 9224 | key.type = BTRFS_EXTENT_DATA_KEY; |
| 9223 | datasize = btrfs_file_extent_calc_inline_size(name_len); | 9225 | datasize = btrfs_file_extent_calc_inline_size(name_len); |
| 9224 | err = btrfs_insert_empty_item(trans, root, path, &key, | 9226 | err = btrfs_insert_empty_item(trans, root, path, &key, |
| 9225 | datasize); | 9227 | datasize); |
| 9226 | if (err) { | 9228 | if (err) { |
| 9227 | btrfs_free_path(path); | 9229 | btrfs_free_path(path); |
| 9228 | goto out_unlock_inode; | 9230 | goto out_unlock_inode; |
| 9229 | } | 9231 | } |
| 9230 | leaf = path->nodes[0]; | 9232 | leaf = path->nodes[0]; |
| 9231 | ei = btrfs_item_ptr(leaf, path->slots[0], | 9233 | ei = btrfs_item_ptr(leaf, path->slots[0], |
| 9232 | struct btrfs_file_extent_item); | 9234 | struct btrfs_file_extent_item); |
| 9233 | btrfs_set_file_extent_generation(leaf, ei, trans->transid); | 9235 | btrfs_set_file_extent_generation(leaf, ei, trans->transid); |
| 9234 | btrfs_set_file_extent_type(leaf, ei, | 9236 | btrfs_set_file_extent_type(leaf, ei, |
| 9235 | BTRFS_FILE_EXTENT_INLINE); | 9237 | BTRFS_FILE_EXTENT_INLINE); |
| 9236 | btrfs_set_file_extent_encryption(leaf, ei, 0); | 9238 | btrfs_set_file_extent_encryption(leaf, ei, 0); |
| 9237 | btrfs_set_file_extent_compression(leaf, ei, 0); | 9239 | btrfs_set_file_extent_compression(leaf, ei, 0); |
| 9238 | btrfs_set_file_extent_other_encoding(leaf, ei, 0); | 9240 | btrfs_set_file_extent_other_encoding(leaf, ei, 0); |
| 9239 | btrfs_set_file_extent_ram_bytes(leaf, ei, name_len); | 9241 | btrfs_set_file_extent_ram_bytes(leaf, ei, name_len); |
| 9240 | 9242 | ||
| 9241 | ptr = btrfs_file_extent_inline_start(ei); | 9243 | ptr = btrfs_file_extent_inline_start(ei); |
| 9242 | write_extent_buffer(leaf, symname, ptr, name_len); | 9244 | write_extent_buffer(leaf, symname, ptr, name_len); |
| 9243 | btrfs_mark_buffer_dirty(leaf); | 9245 | btrfs_mark_buffer_dirty(leaf); |
| 9244 | btrfs_free_path(path); | 9246 | btrfs_free_path(path); |
| 9245 | 9247 | ||
| 9246 | inode->i_op = &btrfs_symlink_inode_operations; | 9248 | inode->i_op = &btrfs_symlink_inode_operations; |
| 9247 | inode->i_mapping->a_ops = &btrfs_symlink_aops; | 9249 | inode->i_mapping->a_ops = &btrfs_symlink_aops; |
| 9248 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; | 9250 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; |
| 9249 | inode_set_bytes(inode, name_len); | 9251 | inode_set_bytes(inode, name_len); |
| 9250 | btrfs_i_size_write(inode, name_len); | 9252 | btrfs_i_size_write(inode, name_len); |
| 9251 | err = btrfs_update_inode(trans, root, inode); | 9253 | err = btrfs_update_inode(trans, root, inode); |
| 9252 | if (err) { | 9254 | if (err) { |
| 9253 | drop_inode = 1; | 9255 | drop_inode = 1; |
| 9254 | goto out_unlock_inode; | 9256 | goto out_unlock_inode; |
| 9255 | } | 9257 | } |
| 9256 | 9258 | ||
| 9257 | unlock_new_inode(inode); | 9259 | unlock_new_inode(inode); |
| 9258 | d_instantiate(dentry, inode); | 9260 | d_instantiate(dentry, inode); |
| 9259 | 9261 | ||
| 9260 | out_unlock: | 9262 | out_unlock: |
| 9261 | btrfs_end_transaction(trans, root); | 9263 | btrfs_end_transaction(trans, root); |
| 9262 | if (drop_inode) { | 9264 | if (drop_inode) { |
| 9263 | inode_dec_link_count(inode); | 9265 | inode_dec_link_count(inode); |
| 9264 | iput(inode); | 9266 | iput(inode); |
| 9265 | } | 9267 | } |
| 9266 | btrfs_btree_balance_dirty(root); | 9268 | btrfs_btree_balance_dirty(root); |
| 9267 | return err; | 9269 | return err; |
| 9268 | 9270 | ||
| 9269 | out_unlock_inode: | 9271 | out_unlock_inode: |
| 9270 | drop_inode = 1; | 9272 | drop_inode = 1; |
| 9271 | unlock_new_inode(inode); | 9273 | unlock_new_inode(inode); |
| 9272 | goto out_unlock; | 9274 | goto out_unlock; |
| 9273 | } | 9275 | } |
| 9274 | 9276 | ||
| 9275 | static int __btrfs_prealloc_file_range(struct inode *inode, int mode, | 9277 | static int __btrfs_prealloc_file_range(struct inode *inode, int mode, |
| 9276 | u64 start, u64 num_bytes, u64 min_size, | 9278 | u64 start, u64 num_bytes, u64 min_size, |
| 9277 | loff_t actual_len, u64 *alloc_hint, | 9279 | loff_t actual_len, u64 *alloc_hint, |
| 9278 | struct btrfs_trans_handle *trans) | 9280 | struct btrfs_trans_handle *trans) |
| 9279 | { | 9281 | { |
| 9280 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; | 9282 | struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; |
| 9281 | struct extent_map *em; | 9283 | struct extent_map *em; |
| 9282 | struct btrfs_root *root = BTRFS_I(inode)->root; | 9284 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 9283 | struct btrfs_key ins; | 9285 | struct btrfs_key ins; |
| 9284 | u64 cur_offset = start; | 9286 | u64 cur_offset = start; |
| 9285 | u64 i_size; | 9287 | u64 i_size; |
| 9286 | u64 cur_bytes; | 9288 | u64 cur_bytes; |
| 9287 | int ret = 0; | 9289 | int ret = 0; |
| 9288 | bool own_trans = true; | 9290 | bool own_trans = true; |
| 9289 | 9291 | ||
| 9290 | if (trans) | 9292 | if (trans) |
| 9291 | own_trans = false; | 9293 | own_trans = false; |
| 9292 | while (num_bytes > 0) { | 9294 | while (num_bytes > 0) { |
| 9293 | if (own_trans) { | 9295 | if (own_trans) { |
| 9294 | trans = btrfs_start_transaction(root, 3); | 9296 | trans = btrfs_start_transaction(root, 3); |
| 9295 | if (IS_ERR(trans)) { | 9297 | if (IS_ERR(trans)) { |
| 9296 | ret = PTR_ERR(trans); | 9298 | ret = PTR_ERR(trans); |
| 9297 | break; | 9299 | break; |
| 9298 | } | 9300 | } |
| 9299 | } | 9301 | } |
| 9300 | 9302 | ||
| 9301 | cur_bytes = min(num_bytes, 256ULL * 1024 * 1024); | 9303 | cur_bytes = min(num_bytes, 256ULL * 1024 * 1024); |
| 9302 | cur_bytes = max(cur_bytes, min_size); | 9304 | cur_bytes = max(cur_bytes, min_size); |
| 9303 | ret = btrfs_reserve_extent(root, cur_bytes, min_size, 0, | 9305 | ret = btrfs_reserve_extent(root, cur_bytes, min_size, 0, |
| 9304 | *alloc_hint, &ins, 1, 0); | 9306 | *alloc_hint, &ins, 1, 0); |
| 9305 | if (ret) { | 9307 | if (ret) { |
| 9306 | if (own_trans) | 9308 | if (own_trans) |
| 9307 | btrfs_end_transaction(trans, root); | 9309 | btrfs_end_transaction(trans, root); |
| 9308 | break; | 9310 | break; |
| 9309 | } | 9311 | } |
| 9310 | 9312 | ||
| 9311 | ret = insert_reserved_file_extent(trans, inode, | 9313 | ret = insert_reserved_file_extent(trans, inode, |
| 9312 | cur_offset, ins.objectid, | 9314 | cur_offset, ins.objectid, |
| 9313 | ins.offset, ins.offset, | 9315 | ins.offset, ins.offset, |
| 9314 | ins.offset, 0, 0, 0, | 9316 | ins.offset, 0, 0, 0, |
| 9315 | BTRFS_FILE_EXTENT_PREALLOC); | 9317 | BTRFS_FILE_EXTENT_PREALLOC); |
| 9316 | if (ret) { | 9318 | if (ret) { |
| 9317 | btrfs_free_reserved_extent(root, ins.objectid, | 9319 | btrfs_free_reserved_extent(root, ins.objectid, |
| 9318 | ins.offset, 0); | 9320 | ins.offset, 0); |
| 9319 | btrfs_abort_transaction(trans, root, ret); | 9321 | btrfs_abort_transaction(trans, root, ret); |
| 9320 | if (own_trans) | 9322 | if (own_trans) |
| 9321 | btrfs_end_transaction(trans, root); | 9323 | btrfs_end_transaction(trans, root); |
| 9322 | break; | 9324 | break; |
| 9323 | } | 9325 | } |
| 9324 | btrfs_drop_extent_cache(inode, cur_offset, | 9326 | btrfs_drop_extent_cache(inode, cur_offset, |
| 9325 | cur_offset + ins.offset -1, 0); | 9327 | cur_offset + ins.offset -1, 0); |
| 9326 | 9328 | ||
| 9327 | em = alloc_extent_map(); | 9329 | em = alloc_extent_map(); |
| 9328 | if (!em) { | 9330 | if (!em) { |
| 9329 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, | 9331 | set_bit(BTRFS_INODE_NEEDS_FULL_SYNC, |
| 9330 | &BTRFS_I(inode)->runtime_flags); | 9332 | &BTRFS_I(inode)->runtime_flags); |
| 9331 | goto next; | 9333 | goto next; |
| 9332 | } | 9334 | } |
| 9333 | 9335 | ||
| 9334 | em->start = cur_offset; | 9336 | em->start = cur_offset; |
| 9335 | em->orig_start = cur_offset; | 9337 | em->orig_start = cur_offset; |
| 9336 | em->len = ins.offset; | 9338 | em->len = ins.offset; |
| 9337 | em->block_start = ins.objectid; | 9339 | em->block_start = ins.objectid; |
| 9338 | em->block_len = ins.offset; | 9340 | em->block_len = ins.offset; |
| 9339 | em->orig_block_len = ins.offset; | 9341 | em->orig_block_len = ins.offset; |
| 9340 | em->ram_bytes = ins.offset; | 9342 | em->ram_bytes = ins.offset; |
| 9341 | em->bdev = root->fs_info->fs_devices->latest_bdev; | 9343 | em->bdev = root->fs_info->fs_devices->latest_bdev; |
| 9342 | set_bit(EXTENT_FLAG_PREALLOC, &em->flags); | 9344 | set_bit(EXTENT_FLAG_PREALLOC, &em->flags); |
| 9343 | em->generation = trans->transid; | 9345 | em->generation = trans->transid; |
| 9344 | 9346 | ||
| 9345 | while (1) { | 9347 | while (1) { |
| 9346 | write_lock(&em_tree->lock); | 9348 | write_lock(&em_tree->lock); |
| 9347 | ret = add_extent_mapping(em_tree, em, 1); | 9349 | ret = add_extent_mapping(em_tree, em, 1); |
| 9348 | write_unlock(&em_tree->lock); | 9350 | write_unlock(&em_tree->lock); |
| 9349 | if (ret != -EEXIST) | 9351 | if (ret != -EEXIST) |
| 9350 | break; | 9352 | break; |
| 9351 | btrfs_drop_extent_cache(inode, cur_offset, | 9353 | btrfs_drop_extent_cache(inode, cur_offset, |
| 9352 | cur_offset + ins.offset - 1, | 9354 | cur_offset + ins.offset - 1, |
| 9353 | 0); | 9355 | 0); |
| 9354 | } | 9356 | } |
| 9355 | free_extent_map(em); | 9357 | free_extent_map(em); |
| 9356 | next: | 9358 | next: |
| 9357 | num_bytes -= ins.offset; | 9359 | num_bytes -= ins.offset; |
| 9358 | cur_offset += ins.offset; | 9360 | cur_offset += ins.offset; |
| 9359 | *alloc_hint = ins.objectid + ins.offset; | 9361 | *alloc_hint = ins.objectid + ins.offset; |
| 9360 | 9362 | ||
| 9361 | inode_inc_iversion(inode); | 9363 | inode_inc_iversion(inode); |
| 9362 | inode->i_ctime = CURRENT_TIME; | 9364 | inode->i_ctime = CURRENT_TIME; |
| 9363 | BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC; | 9365 | BTRFS_I(inode)->flags |= BTRFS_INODE_PREALLOC; |
| 9364 | if (!(mode & FALLOC_FL_KEEP_SIZE) && | 9366 | if (!(mode & FALLOC_FL_KEEP_SIZE) && |
| 9365 | (actual_len > inode->i_size) && | 9367 | (actual_len > inode->i_size) && |
| 9366 | (cur_offset > inode->i_size)) { | 9368 | (cur_offset > inode->i_size)) { |
| 9367 | if (cur_offset > actual_len) | 9369 | if (cur_offset > actual_len) |
| 9368 | i_size = actual_len; | 9370 | i_size = actual_len; |
| 9369 | else | 9371 | else |
| 9370 | i_size = cur_offset; | 9372 | i_size = cur_offset; |
| 9371 | i_size_write(inode, i_size); | 9373 | i_size_write(inode, i_size); |
| 9372 | btrfs_ordered_update_i_size(inode, i_size, NULL); | 9374 | btrfs_ordered_update_i_size(inode, i_size, NULL); |
| 9373 | } | 9375 | } |
| 9374 | 9376 | ||
| 9375 | ret = btrfs_update_inode(trans, root, inode); | 9377 | ret = btrfs_update_inode(trans, root, inode); |
| 9376 | 9378 | ||
| 9377 | if (ret) { | 9379 | if (ret) { |
| 9378 | btrfs_abort_transaction(trans, root, ret); | 9380 | btrfs_abort_transaction(trans, root, ret); |
| 9379 | if (own_trans) | 9381 | if (own_trans) |
| 9380 | btrfs_end_transaction(trans, root); | 9382 | btrfs_end_transaction(trans, root); |
| 9381 | break; | 9383 | break; |
| 9382 | } | 9384 | } |
| 9383 | 9385 | ||
| 9384 | if (own_trans) | 9386 | if (own_trans) |
| 9385 | btrfs_end_transaction(trans, root); | 9387 | btrfs_end_transaction(trans, root); |
| 9386 | } | 9388 | } |
| 9387 | return ret; | 9389 | return ret; |
| 9388 | } | 9390 | } |
| 9389 | 9391 | ||
| 9390 | int btrfs_prealloc_file_range(struct inode *inode, int mode, | 9392 | int btrfs_prealloc_file_range(struct inode *inode, int mode, |
| 9391 | u64 start, u64 num_bytes, u64 min_size, | 9393 | u64 start, u64 num_bytes, u64 min_size, |
| 9392 | loff_t actual_len, u64 *alloc_hint) | 9394 | loff_t actual_len, u64 *alloc_hint) |
| 9393 | { | 9395 | { |
| 9394 | return __btrfs_prealloc_file_range(inode, mode, start, num_bytes, | 9396 | return __btrfs_prealloc_file_range(inode, mode, start, num_bytes, |
| 9395 | min_size, actual_len, alloc_hint, | 9397 | min_size, actual_len, alloc_hint, |
| 9396 | NULL); | 9398 | NULL); |
| 9397 | } | 9399 | } |
| 9398 | 9400 | ||
| 9399 | int btrfs_prealloc_file_range_trans(struct inode *inode, | 9401 | int btrfs_prealloc_file_range_trans(struct inode *inode, |
| 9400 | struct btrfs_trans_handle *trans, int mode, | 9402 | struct btrfs_trans_handle *trans, int mode, |
| 9401 | u64 start, u64 num_bytes, u64 min_size, | 9403 | u64 start, u64 num_bytes, u64 min_size, |
| 9402 | loff_t actual_len, u64 *alloc_hint) | 9404 | loff_t actual_len, u64 *alloc_hint) |
| 9403 | { | 9405 | { |
| 9404 | return __btrfs_prealloc_file_range(inode, mode, start, num_bytes, | 9406 | return __btrfs_prealloc_file_range(inode, mode, start, num_bytes, |
| 9405 | min_size, actual_len, alloc_hint, trans); | 9407 | min_size, actual_len, alloc_hint, trans); |
| 9406 | } | 9408 | } |
| 9407 | 9409 | ||
| 9408 | static int btrfs_set_page_dirty(struct page *page) | 9410 | static int btrfs_set_page_dirty(struct page *page) |
| 9409 | { | 9411 | { |
| 9410 | return __set_page_dirty_nobuffers(page); | 9412 | return __set_page_dirty_nobuffers(page); |
| 9411 | } | 9413 | } |
| 9412 | 9414 | ||
| 9413 | static int btrfs_permission(struct inode *inode, int mask) | 9415 | static int btrfs_permission(struct inode *inode, int mask) |
| 9414 | { | 9416 | { |
| 9415 | struct btrfs_root *root = BTRFS_I(inode)->root; | 9417 | struct btrfs_root *root = BTRFS_I(inode)->root; |
| 9416 | umode_t mode = inode->i_mode; | 9418 | umode_t mode = inode->i_mode; |
| 9417 | 9419 | ||
| 9418 | if (mask & MAY_WRITE && | 9420 | if (mask & MAY_WRITE && |
| 9419 | (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) { | 9421 | (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) { |
| 9420 | if (btrfs_root_readonly(root)) | 9422 | if (btrfs_root_readonly(root)) |
| 9421 | return -EROFS; | 9423 | return -EROFS; |
| 9422 | if (BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) | 9424 | if (BTRFS_I(inode)->flags & BTRFS_INODE_READONLY) |
| 9423 | return -EACCES; | 9425 | return -EACCES; |
| 9424 | } | 9426 | } |
| 9425 | return generic_permission(inode, mask); | 9427 | return generic_permission(inode, mask); |
| 9426 | } | 9428 | } |
| 9427 | 9429 | ||
| 9428 | static int btrfs_tmpfile(struct inode *dir, struct dentry *dentry, umode_t mode) | 9430 | static int btrfs_tmpfile(struct inode *dir, struct dentry *dentry, umode_t mode) |
| 9429 | { | 9431 | { |
| 9430 | struct btrfs_trans_handle *trans; | 9432 | struct btrfs_trans_handle *trans; |
| 9431 | struct btrfs_root *root = BTRFS_I(dir)->root; | 9433 | struct btrfs_root *root = BTRFS_I(dir)->root; |
| 9432 | struct inode *inode = NULL; | 9434 | struct inode *inode = NULL; |
| 9433 | u64 objectid; | 9435 | u64 objectid; |
| 9434 | u64 index; | 9436 | u64 index; |
| 9435 | int ret = 0; | 9437 | int ret = 0; |
| 9436 | 9438 | ||
| 9437 | /* | 9439 | /* |
| 9438 | * 5 units required for adding orphan entry | 9440 | * 5 units required for adding orphan entry |
| 9439 | */ | 9441 | */ |
| 9440 | trans = btrfs_start_transaction(root, 5); | 9442 | trans = btrfs_start_transaction(root, 5); |
| 9441 | if (IS_ERR(trans)) | 9443 | if (IS_ERR(trans)) |
| 9442 | return PTR_ERR(trans); | 9444 | return PTR_ERR(trans); |
| 9443 | 9445 | ||
| 9444 | ret = btrfs_find_free_ino(root, &objectid); | 9446 | ret = btrfs_find_free_ino(root, &objectid); |
| 9445 | if (ret) | 9447 | if (ret) |
| 9446 | goto out; | 9448 | goto out; |
| 9447 | 9449 | ||
| 9448 | inode = btrfs_new_inode(trans, root, dir, NULL, 0, | 9450 | inode = btrfs_new_inode(trans, root, dir, NULL, 0, |
| 9449 | btrfs_ino(dir), objectid, mode, &index); | 9451 | btrfs_ino(dir), objectid, mode, &index); |
| 9450 | if (IS_ERR(inode)) { | 9452 | if (IS_ERR(inode)) { |
| 9451 | ret = PTR_ERR(inode); | 9453 | ret = PTR_ERR(inode); |
| 9452 | inode = NULL; | 9454 | inode = NULL; |
| 9453 | goto out; | 9455 | goto out; |
| 9454 | } | 9456 | } |
| 9455 | 9457 | ||
| 9456 | inode->i_fop = &btrfs_file_operations; | 9458 | inode->i_fop = &btrfs_file_operations; |
| 9457 | inode->i_op = &btrfs_file_inode_operations; | 9459 | inode->i_op = &btrfs_file_inode_operations; |
| 9458 | 9460 | ||
| 9459 | inode->i_mapping->a_ops = &btrfs_aops; | 9461 | inode->i_mapping->a_ops = &btrfs_aops; |
| 9460 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; | 9462 | inode->i_mapping->backing_dev_info = &root->fs_info->bdi; |
| 9461 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; | 9463 | BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; |
| 9462 | 9464 | ||
| 9463 | ret = btrfs_init_inode_security(trans, inode, dir, NULL); | 9465 | ret = btrfs_init_inode_security(trans, inode, dir, NULL); |
| 9464 | if (ret) | 9466 | if (ret) |
| 9465 | goto out_inode; | 9467 | goto out_inode; |
| 9466 | 9468 | ||
| 9467 | ret = btrfs_update_inode(trans, root, inode); | 9469 | ret = btrfs_update_inode(trans, root, inode); |
| 9468 | if (ret) | 9470 | if (ret) |
| 9469 | goto out_inode; | 9471 | goto out_inode; |
| 9470 | ret = btrfs_orphan_add(trans, inode); | 9472 | ret = btrfs_orphan_add(trans, inode); |
| 9471 | if (ret) | 9473 | if (ret) |
| 9472 | goto out_inode; | 9474 | goto out_inode; |
| 9473 | 9475 | ||
| 9474 | /* | 9476 | /* |
| 9475 | * We set number of links to 0 in btrfs_new_inode(), and here we set | 9477 | * We set number of links to 0 in btrfs_new_inode(), and here we set |
| 9476 | * it to 1 because d_tmpfile() will issue a warning if the count is 0, | 9478 | * it to 1 because d_tmpfile() will issue a warning if the count is 0, |
| 9477 | * through: | 9479 | * through: |
| 9478 | * | 9480 | * |
| 9479 | * d_tmpfile() -> inode_dec_link_count() -> drop_nlink() | 9481 | * d_tmpfile() -> inode_dec_link_count() -> drop_nlink() |
| 9480 | */ | 9482 | */ |
| 9481 | set_nlink(inode, 1); | 9483 | set_nlink(inode, 1); |
| 9482 | unlock_new_inode(inode); | 9484 | unlock_new_inode(inode); |
| 9483 | d_tmpfile(dentry, inode); | 9485 | d_tmpfile(dentry, inode); |
| 9484 | mark_inode_dirty(inode); | 9486 | mark_inode_dirty(inode); |
| 9485 | 9487 | ||
| 9486 | out: | 9488 | out: |
| 9487 | btrfs_end_transaction(trans, root); | 9489 | btrfs_end_transaction(trans, root); |
| 9488 | if (ret) | 9490 | if (ret) |
| 9489 | iput(inode); | 9491 | iput(inode); |
| 9490 | btrfs_balance_delayed_items(root); | 9492 | btrfs_balance_delayed_items(root); |
| 9491 | btrfs_btree_balance_dirty(root); | 9493 | btrfs_btree_balance_dirty(root); |
| 9492 | return ret; | 9494 | return ret; |
| 9493 | 9495 | ||
| 9494 | out_inode: | 9496 | out_inode: |
| 9495 | unlock_new_inode(inode); | 9497 | unlock_new_inode(inode); |
| 9496 | goto out; | 9498 | goto out; |
| 9497 | 9499 | ||
| 9498 | } | 9500 | } |
| 9499 | 9501 | ||
| 9500 | /* Inspired by filemap_check_errors() */ | 9502 | /* Inspired by filemap_check_errors() */ |
| 9501 | int btrfs_inode_check_errors(struct inode *inode) | 9503 | int btrfs_inode_check_errors(struct inode *inode) |
| 9502 | { | 9504 | { |
| 9503 | int ret = 0; | 9505 | int ret = 0; |
| 9504 | 9506 | ||
| 9505 | if (test_bit(AS_ENOSPC, &inode->i_mapping->flags) && | 9507 | if (test_bit(AS_ENOSPC, &inode->i_mapping->flags) && |
| 9506 | test_and_clear_bit(AS_ENOSPC, &inode->i_mapping->flags)) | 9508 | test_and_clear_bit(AS_ENOSPC, &inode->i_mapping->flags)) |
| 9507 | ret = -ENOSPC; | 9509 | ret = -ENOSPC; |
| 9508 | if (test_bit(AS_EIO, &inode->i_mapping->flags) && | 9510 | if (test_bit(AS_EIO, &inode->i_mapping->flags) && |
| 9509 | test_and_clear_bit(AS_EIO, &inode->i_mapping->flags)) | 9511 | test_and_clear_bit(AS_EIO, &inode->i_mapping->flags)) |
| 9510 | ret = -EIO; | 9512 | ret = -EIO; |
| 9511 | 9513 | ||
| 9512 | return ret; | 9514 | return ret; |
| 9513 | } | 9515 | } |
| 9514 | 9516 | ||
| 9515 | static const struct inode_operations btrfs_dir_inode_operations = { | 9517 | static const struct inode_operations btrfs_dir_inode_operations = { |
| 9516 | .getattr = btrfs_getattr, | 9518 | .getattr = btrfs_getattr, |
| 9517 | .lookup = btrfs_lookup, | 9519 | .lookup = btrfs_lookup, |
| 9518 | .create = btrfs_create, | 9520 | .create = btrfs_create, |
| 9519 | .unlink = btrfs_unlink, | 9521 | .unlink = btrfs_unlink, |
| 9520 | .link = btrfs_link, | 9522 | .link = btrfs_link, |
| 9521 | .mkdir = btrfs_mkdir, | 9523 | .mkdir = btrfs_mkdir, |
| 9522 | .rmdir = btrfs_rmdir, | 9524 | .rmdir = btrfs_rmdir, |
| 9523 | .rename2 = btrfs_rename2, | 9525 | .rename2 = btrfs_rename2, |
| 9524 | .symlink = btrfs_symlink, | 9526 | .symlink = btrfs_symlink, |
| 9525 | .setattr = btrfs_setattr, | 9527 | .setattr = btrfs_setattr, |
| 9526 | .mknod = btrfs_mknod, | 9528 | .mknod = btrfs_mknod, |
| 9527 | .setxattr = btrfs_setxattr, | 9529 | .setxattr = btrfs_setxattr, |
| 9528 | .getxattr = btrfs_getxattr, | 9530 | .getxattr = btrfs_getxattr, |
| 9529 | .listxattr = btrfs_listxattr, | 9531 | .listxattr = btrfs_listxattr, |
| 9530 | .removexattr = btrfs_removexattr, | 9532 | .removexattr = btrfs_removexattr, |
| 9531 | .permission = btrfs_permission, | 9533 | .permission = btrfs_permission, |
| 9532 | .get_acl = btrfs_get_acl, | 9534 | .get_acl = btrfs_get_acl, |
| 9533 | .set_acl = btrfs_set_acl, | 9535 | .set_acl = btrfs_set_acl, |
| 9534 | .update_time = btrfs_update_time, | 9536 | .update_time = btrfs_update_time, |
| 9535 | .tmpfile = btrfs_tmpfile, | 9537 | .tmpfile = btrfs_tmpfile, |
| 9536 | }; | 9538 | }; |
| 9537 | static const struct inode_operations btrfs_dir_ro_inode_operations = { | 9539 | static const struct inode_operations btrfs_dir_ro_inode_operations = { |
| 9538 | .lookup = btrfs_lookup, | 9540 | .lookup = btrfs_lookup, |
| 9539 | .permission = btrfs_permission, | 9541 | .permission = btrfs_permission, |
| 9540 | .get_acl = btrfs_get_acl, | 9542 | .get_acl = btrfs_get_acl, |
| 9541 | .set_acl = btrfs_set_acl, | 9543 | .set_acl = btrfs_set_acl, |
| 9542 | .update_time = btrfs_update_time, | 9544 | .update_time = btrfs_update_time, |
| 9543 | }; | 9545 | }; |
| 9544 | 9546 | ||
| 9545 | static const struct file_operations btrfs_dir_file_operations = { | 9547 | static const struct file_operations btrfs_dir_file_operations = { |
| 9546 | .llseek = generic_file_llseek, | 9548 | .llseek = generic_file_llseek, |
| 9547 | .read = generic_read_dir, | 9549 | .read = generic_read_dir, |
| 9548 | .iterate = btrfs_real_readdir, | 9550 | .iterate = btrfs_real_readdir, |
| 9549 | .unlocked_ioctl = btrfs_ioctl, | 9551 | .unlocked_ioctl = btrfs_ioctl, |
| 9550 | #ifdef CONFIG_COMPAT | 9552 | #ifdef CONFIG_COMPAT |
| 9551 | .compat_ioctl = btrfs_ioctl, | 9553 | .compat_ioctl = btrfs_ioctl, |
| 9552 | #endif | 9554 | #endif |
| 9553 | .release = btrfs_release_file, | 9555 | .release = btrfs_release_file, |
| 9554 | .fsync = btrfs_sync_file, | 9556 | .fsync = btrfs_sync_file, |
| 9555 | }; | 9557 | }; |
| 9556 | 9558 | ||
| 9557 | static struct extent_io_ops btrfs_extent_io_ops = { | 9559 | static struct extent_io_ops btrfs_extent_io_ops = { |
| 9558 | .fill_delalloc = run_delalloc_range, | 9560 | .fill_delalloc = run_delalloc_range, |
| 9559 | .submit_bio_hook = btrfs_submit_bio_hook, | 9561 | .submit_bio_hook = btrfs_submit_bio_hook, |
| 9560 | .merge_bio_hook = btrfs_merge_bio_hook, | 9562 | .merge_bio_hook = btrfs_merge_bio_hook, |
| 9561 | .readpage_end_io_hook = btrfs_readpage_end_io_hook, | 9563 | .readpage_end_io_hook = btrfs_readpage_end_io_hook, |
| 9562 | .writepage_end_io_hook = btrfs_writepage_end_io_hook, | 9564 | .writepage_end_io_hook = btrfs_writepage_end_io_hook, |
| 9563 | .writepage_start_hook = btrfs_writepage_start_hook, | 9565 | .writepage_start_hook = btrfs_writepage_start_hook, |
| 9564 | .set_bit_hook = btrfs_set_bit_hook, | 9566 | .set_bit_hook = btrfs_set_bit_hook, |
| 9565 | .clear_bit_hook = btrfs_clear_bit_hook, | 9567 | .clear_bit_hook = btrfs_clear_bit_hook, |
| 9566 | .merge_extent_hook = btrfs_merge_extent_hook, | 9568 | .merge_extent_hook = btrfs_merge_extent_hook, |
| 9567 | .split_extent_hook = btrfs_split_extent_hook, | 9569 | .split_extent_hook = btrfs_split_extent_hook, |
| 9568 | }; | 9570 | }; |
| 9569 | 9571 | ||
| 9570 | /* | 9572 | /* |
| 9571 | * btrfs doesn't support the bmap operation because swapfiles | 9573 | * btrfs doesn't support the bmap operation because swapfiles |
| 9572 | * use bmap to make a mapping of extents in the file. They assume | 9574 | * use bmap to make a mapping of extents in the file. They assume |
| 9573 | * these extents won't change over the life of the file and they | 9575 | * these extents won't change over the life of the file and they |
| 9574 | * use the bmap result to do IO directly to the drive. | 9576 | * use the bmap result to do IO directly to the drive. |
| 9575 | * | 9577 | * |
| 9576 | * the btrfs bmap call would return logical addresses that aren't | 9578 | * the btrfs bmap call would return logical addresses that aren't |
| 9577 | * suitable for IO and they also will change frequently as COW | 9579 | * suitable for IO and they also will change frequently as COW |
| 9578 | * operations happen. So, swapfile + btrfs == corruption. | 9580 | * operations happen. So, swapfile + btrfs == corruption. |
| 9579 | * | 9581 | * |
| 9580 | * For now we're avoiding this by dropping bmap. | 9582 | * For now we're avoiding this by dropping bmap. |
| 9581 | */ | 9583 | */ |
| 9582 | static const struct address_space_operations btrfs_aops = { | 9584 | static const struct address_space_operations btrfs_aops = { |
| 9583 | .readpage = btrfs_readpage, | 9585 | .readpage = btrfs_readpage, |
| 9584 | .writepage = btrfs_writepage, | 9586 | .writepage = btrfs_writepage, |
| 9585 | .writepages = btrfs_writepages, | 9587 | .writepages = btrfs_writepages, |
| 9586 | .readpages = btrfs_readpages, | 9588 | .readpages = btrfs_readpages, |
| 9587 | .direct_IO = btrfs_direct_IO, | 9589 | .direct_IO = btrfs_direct_IO, |
| 9588 | .invalidatepage = btrfs_invalidatepage, | 9590 | .invalidatepage = btrfs_invalidatepage, |
| 9589 | .releasepage = btrfs_releasepage, | 9591 | .releasepage = btrfs_releasepage, |
| 9590 | .set_page_dirty = btrfs_set_page_dirty, | 9592 | .set_page_dirty = btrfs_set_page_dirty, |
| 9591 | .error_remove_page = generic_error_remove_page, | 9593 | .error_remove_page = generic_error_remove_page, |
| 9592 | }; | 9594 | }; |
| 9593 | 9595 | ||
| 9594 | static const struct address_space_operations btrfs_symlink_aops = { | 9596 | static const struct address_space_operations btrfs_symlink_aops = { |
| 9595 | .readpage = btrfs_readpage, | 9597 | .readpage = btrfs_readpage, |
| 9596 | .writepage = btrfs_writepage, | 9598 | .writepage = btrfs_writepage, |
| 9597 | .invalidatepage = btrfs_invalidatepage, | 9599 | .invalidatepage = btrfs_invalidatepage, |
| 9598 | .releasepage = btrfs_releasepage, | 9600 | .releasepage = btrfs_releasepage, |
| 9599 | }; | 9601 | }; |
| 9600 | 9602 | ||
| 9601 | static const struct inode_operations btrfs_file_inode_operations = { | 9603 | static const struct inode_operations btrfs_file_inode_operations = { |
| 9602 | .getattr = btrfs_getattr, | 9604 | .getattr = btrfs_getattr, |
| 9603 | .setattr = btrfs_setattr, | 9605 | .setattr = btrfs_setattr, |
| 9604 | .setxattr = btrfs_setxattr, | 9606 | .setxattr = btrfs_setxattr, |
| 9605 | .getxattr = btrfs_getxattr, | 9607 | .getxattr = btrfs_getxattr, |
| 9606 | .listxattr = btrfs_listxattr, | 9608 | .listxattr = btrfs_listxattr, |
| 9607 | .removexattr = btrfs_removexattr, | 9609 | .removexattr = btrfs_removexattr, |
| 9608 | .permission = btrfs_permission, | 9610 | .permission = btrfs_permission, |
| 9609 | .fiemap = btrfs_fiemap, | 9611 | .fiemap = btrfs_fiemap, |
| 9610 | .get_acl = btrfs_get_acl, | 9612 | .get_acl = btrfs_get_acl, |
| 9611 | .set_acl = btrfs_set_acl, | 9613 | .set_acl = btrfs_set_acl, |
| 9612 | .update_time = btrfs_update_time, | 9614 | .update_time = btrfs_update_time, |
| 9613 | }; | 9615 | }; |
| 9614 | static const struct inode_operations btrfs_special_inode_operations = { | 9616 | static const struct inode_operations btrfs_special_inode_operations = { |
| 9615 | .getattr = btrfs_getattr, | 9617 | .getattr = btrfs_getattr, |
| 9616 | .setattr = btrfs_setattr, | 9618 | .setattr = btrfs_setattr, |
| 9617 | .permission = btrfs_permission, | 9619 | .permission = btrfs_permission, |
| 9618 | .setxattr = btrfs_setxattr, | 9620 | .setxattr = btrfs_setxattr, |
| 9619 | .getxattr = btrfs_getxattr, | 9621 | .getxattr = btrfs_getxattr, |
| 9620 | .listxattr = btrfs_listxattr, | 9622 | .listxattr = btrfs_listxattr, |
| 9621 | .removexattr = btrfs_removexattr, | 9623 | .removexattr = btrfs_removexattr, |
| 9622 | .get_acl = btrfs_get_acl, | 9624 | .get_acl = btrfs_get_acl, |
| 9623 | .set_acl = btrfs_set_acl, | 9625 | .set_acl = btrfs_set_acl, |
| 9624 | .update_time = btrfs_update_time, | 9626 | .update_time = btrfs_update_time, |
| 9625 | }; | 9627 | }; |
| 9626 | static const struct inode_operations btrfs_symlink_inode_operations = { | 9628 | static const struct inode_operations btrfs_symlink_inode_operations = { |
| 9627 | .readlink = generic_readlink, | 9629 | .readlink = generic_readlink, |
| 9628 | .follow_link = page_follow_link_light, | 9630 | .follow_link = page_follow_link_light, |
| 9629 | .put_link = page_put_link, | 9631 | .put_link = page_put_link, |
| 9630 | .getattr = btrfs_getattr, | 9632 | .getattr = btrfs_getattr, |
| 9631 | .setattr = btrfs_setattr, | 9633 | .setattr = btrfs_setattr, |
| 9632 | .permission = btrfs_permission, | 9634 | .permission = btrfs_permission, |
| 9633 | .setxattr = btrfs_setxattr, | 9635 | .setxattr = btrfs_setxattr, |
| 9634 | .getxattr = btrfs_getxattr, | 9636 | .getxattr = btrfs_getxattr, |
| 9635 | .listxattr = btrfs_listxattr, | 9637 | .listxattr = btrfs_listxattr, |
| 9636 | .removexattr = btrfs_removexattr, | 9638 | .removexattr = btrfs_removexattr, |
| 9637 | .update_time = btrfs_update_time, | 9639 | .update_time = btrfs_update_time, |
| 9638 | }; | 9640 | }; |
| 9639 | 9641 | ||
| 9640 | const struct dentry_operations btrfs_dentry_operations = { | 9642 | const struct dentry_operations btrfs_dentry_operations = { |
| 9641 | .d_delete = btrfs_dentry_delete, | 9643 | .d_delete = btrfs_dentry_delete, |
| 9642 | .d_release = btrfs_dentry_release, | 9644 | .d_release = btrfs_dentry_release, |
| 9643 | }; | 9645 | }; |
| 9644 | 9646 |
fs/btrfs/scrub.c
| 1 | /* | 1 | /* |
| 2 | * Copyright (C) 2011, 2012 STRATO. All rights reserved. | 2 | * Copyright (C) 2011, 2012 STRATO. All rights reserved. |
| 3 | * | 3 | * |
| 4 | * This program is free software; you can redistribute it and/or | 4 | * This program is free software; you can redistribute it and/or |
| 5 | * modify it under the terms of the GNU General Public | 5 | * modify it under the terms of the GNU General Public |
| 6 | * License v2 as published by the Free Software Foundation. | 6 | * License v2 as published by the Free Software Foundation. |
| 7 | * | 7 | * |
| 8 | * This program is distributed in the hope that it will be useful, | 8 | * This program is distributed in the hope that it will be useful, |
| 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | 9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | 10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 11 | * General Public License for more details. | 11 | * General Public License for more details. |
| 12 | * | 12 | * |
| 13 | * You should have received a copy of the GNU General Public | 13 | * You should have received a copy of the GNU General Public |
| 14 | * License along with this program; if not, write to the | 14 | * License along with this program; if not, write to the |
| 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | 15 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, |
| 16 | * Boston, MA 021110-1307, USA. | 16 | * Boston, MA 021110-1307, USA. |
| 17 | */ | 17 | */ |
| 18 | 18 | ||
| 19 | #include <linux/blkdev.h> | 19 | #include <linux/blkdev.h> |
| 20 | #include <linux/ratelimit.h> | 20 | #include <linux/ratelimit.h> |
| 21 | #include "ctree.h" | 21 | #include "ctree.h" |
| 22 | #include "volumes.h" | 22 | #include "volumes.h" |
| 23 | #include "disk-io.h" | 23 | #include "disk-io.h" |
| 24 | #include "ordered-data.h" | 24 | #include "ordered-data.h" |
| 25 | #include "transaction.h" | 25 | #include "transaction.h" |
| 26 | #include "backref.h" | 26 | #include "backref.h" |
| 27 | #include "extent_io.h" | 27 | #include "extent_io.h" |
| 28 | #include "dev-replace.h" | 28 | #include "dev-replace.h" |
| 29 | #include "check-integrity.h" | 29 | #include "check-integrity.h" |
| 30 | #include "rcu-string.h" | 30 | #include "rcu-string.h" |
| 31 | #include "raid56.h" | 31 | #include "raid56.h" |
| 32 | 32 | ||
| 33 | /* | 33 | /* |
| 34 | * This is only the first step towards a full-features scrub. It reads all | 34 | * This is only the first step towards a full-features scrub. It reads all |
| 35 | * extent and super block and verifies the checksums. In case a bad checksum | 35 | * extent and super block and verifies the checksums. In case a bad checksum |
| 36 | * is found or the extent cannot be read, good data will be written back if | 36 | * is found or the extent cannot be read, good data will be written back if |
| 37 | * any can be found. | 37 | * any can be found. |
| 38 | * | 38 | * |
| 39 | * Future enhancements: | 39 | * Future enhancements: |
| 40 | * - In case an unrepairable extent is encountered, track which files are | 40 | * - In case an unrepairable extent is encountered, track which files are |
| 41 | * affected and report them | 41 | * affected and report them |
| 42 | * - track and record media errors, throw out bad devices | 42 | * - track and record media errors, throw out bad devices |
| 43 | * - add a mode to also read unallocated space | 43 | * - add a mode to also read unallocated space |
| 44 | */ | 44 | */ |
| 45 | 45 | ||
| 46 | struct scrub_block; | 46 | struct scrub_block; |
| 47 | struct scrub_ctx; | 47 | struct scrub_ctx; |
| 48 | 48 | ||
| 49 | /* | 49 | /* |
| 50 | * the following three values only influence the performance. | 50 | * the following three values only influence the performance. |
| 51 | * The last one configures the number of parallel and outstanding I/O | 51 | * The last one configures the number of parallel and outstanding I/O |
| 52 | * operations. The first two values configure an upper limit for the number | 52 | * operations. The first two values configure an upper limit for the number |
| 53 | * of (dynamically allocated) pages that are added to a bio. | 53 | * of (dynamically allocated) pages that are added to a bio. |
| 54 | */ | 54 | */ |
| 55 | #define SCRUB_PAGES_PER_RD_BIO 32 /* 128k per bio */ | 55 | #define SCRUB_PAGES_PER_RD_BIO 32 /* 128k per bio */ |
| 56 | #define SCRUB_PAGES_PER_WR_BIO 32 /* 128k per bio */ | 56 | #define SCRUB_PAGES_PER_WR_BIO 32 /* 128k per bio */ |
| 57 | #define SCRUB_BIOS_PER_SCTX 64 /* 8MB per device in flight */ | 57 | #define SCRUB_BIOS_PER_SCTX 64 /* 8MB per device in flight */ |
| 58 | 58 | ||
| 59 | /* | 59 | /* |
| 60 | * the following value times PAGE_SIZE needs to be large enough to match the | 60 | * the following value times PAGE_SIZE needs to be large enough to match the |
| 61 | * largest node/leaf/sector size that shall be supported. | 61 | * largest node/leaf/sector size that shall be supported. |
| 62 | * Values larger than BTRFS_STRIPE_LEN are not supported. | 62 | * Values larger than BTRFS_STRIPE_LEN are not supported. |
| 63 | */ | 63 | */ |
| 64 | #define SCRUB_MAX_PAGES_PER_BLOCK 16 /* 64k per node/leaf/sector */ | 64 | #define SCRUB_MAX_PAGES_PER_BLOCK 16 /* 64k per node/leaf/sector */ |
| 65 | 65 | ||
| 66 | struct scrub_recover { | 66 | struct scrub_recover { |
| 67 | atomic_t refs; | 67 | atomic_t refs; |
| 68 | struct btrfs_bio *bbio; | 68 | struct btrfs_bio *bbio; |
| 69 | u64 *raid_map; | 69 | u64 *raid_map; |
| 70 | u64 map_length; | 70 | u64 map_length; |
| 71 | }; | 71 | }; |
| 72 | 72 | ||
| 73 | struct scrub_page { | 73 | struct scrub_page { |
| 74 | struct scrub_block *sblock; | 74 | struct scrub_block *sblock; |
| 75 | struct page *page; | 75 | struct page *page; |
| 76 | struct btrfs_device *dev; | 76 | struct btrfs_device *dev; |
| 77 | struct list_head list; | 77 | struct list_head list; |
| 78 | u64 flags; /* extent flags */ | 78 | u64 flags; /* extent flags */ |
| 79 | u64 generation; | 79 | u64 generation; |
| 80 | u64 logical; | 80 | u64 logical; |
| 81 | u64 physical; | 81 | u64 physical; |
| 82 | u64 physical_for_dev_replace; | 82 | u64 physical_for_dev_replace; |
| 83 | atomic_t ref_count; | 83 | atomic_t ref_count; |
| 84 | struct { | 84 | struct { |
| 85 | unsigned int mirror_num:8; | 85 | unsigned int mirror_num:8; |
| 86 | unsigned int have_csum:1; | 86 | unsigned int have_csum:1; |
| 87 | unsigned int io_error:1; | 87 | unsigned int io_error:1; |
| 88 | }; | 88 | }; |
| 89 | u8 csum[BTRFS_CSUM_SIZE]; | 89 | u8 csum[BTRFS_CSUM_SIZE]; |
| 90 | 90 | ||
| 91 | struct scrub_recover *recover; | 91 | struct scrub_recover *recover; |
| 92 | }; | 92 | }; |
| 93 | 93 | ||
| 94 | struct scrub_bio { | 94 | struct scrub_bio { |
| 95 | int index; | 95 | int index; |
| 96 | struct scrub_ctx *sctx; | 96 | struct scrub_ctx *sctx; |
| 97 | struct btrfs_device *dev; | 97 | struct btrfs_device *dev; |
| 98 | struct bio *bio; | 98 | struct bio *bio; |
| 99 | int err; | 99 | int err; |
| 100 | u64 logical; | 100 | u64 logical; |
| 101 | u64 physical; | 101 | u64 physical; |
| 102 | #if SCRUB_PAGES_PER_WR_BIO >= SCRUB_PAGES_PER_RD_BIO | 102 | #if SCRUB_PAGES_PER_WR_BIO >= SCRUB_PAGES_PER_RD_BIO |
| 103 | struct scrub_page *pagev[SCRUB_PAGES_PER_WR_BIO]; | 103 | struct scrub_page *pagev[SCRUB_PAGES_PER_WR_BIO]; |
| 104 | #else | 104 | #else |
| 105 | struct scrub_page *pagev[SCRUB_PAGES_PER_RD_BIO]; | 105 | struct scrub_page *pagev[SCRUB_PAGES_PER_RD_BIO]; |
| 106 | #endif | 106 | #endif |
| 107 | int page_count; | 107 | int page_count; |
| 108 | int next_free; | 108 | int next_free; |
| 109 | struct btrfs_work work; | 109 | struct btrfs_work work; |
| 110 | }; | 110 | }; |
| 111 | 111 | ||
| 112 | struct scrub_block { | 112 | struct scrub_block { |
| 113 | struct scrub_page *pagev[SCRUB_MAX_PAGES_PER_BLOCK]; | 113 | struct scrub_page *pagev[SCRUB_MAX_PAGES_PER_BLOCK]; |
| 114 | int page_count; | 114 | int page_count; |
| 115 | atomic_t outstanding_pages; | 115 | atomic_t outstanding_pages; |
| 116 | atomic_t ref_count; /* free mem on transition to zero */ | 116 | atomic_t ref_count; /* free mem on transition to zero */ |
| 117 | struct scrub_ctx *sctx; | 117 | struct scrub_ctx *sctx; |
| 118 | struct scrub_parity *sparity; | 118 | struct scrub_parity *sparity; |
| 119 | struct { | 119 | struct { |
| 120 | unsigned int header_error:1; | 120 | unsigned int header_error:1; |
| 121 | unsigned int checksum_error:1; | 121 | unsigned int checksum_error:1; |
| 122 | unsigned int no_io_error_seen:1; | 122 | unsigned int no_io_error_seen:1; |
| 123 | unsigned int generation_error:1; /* also sets header_error */ | 123 | unsigned int generation_error:1; /* also sets header_error */ |
| 124 | 124 | ||
| 125 | /* The following is for the data used to check parity */ | 125 | /* The following is for the data used to check parity */ |
| 126 | /* It is for the data with checksum */ | 126 | /* It is for the data with checksum */ |
| 127 | unsigned int data_corrected:1; | 127 | unsigned int data_corrected:1; |
| 128 | }; | 128 | }; |
| 129 | }; | 129 | }; |
| 130 | 130 | ||
| 131 | /* Used for the chunks with parity stripe such RAID5/6 */ | 131 | /* Used for the chunks with parity stripe such RAID5/6 */ |
| 132 | struct scrub_parity { | 132 | struct scrub_parity { |
| 133 | struct scrub_ctx *sctx; | 133 | struct scrub_ctx *sctx; |
| 134 | 134 | ||
| 135 | struct btrfs_device *scrub_dev; | 135 | struct btrfs_device *scrub_dev; |
| 136 | 136 | ||
| 137 | u64 logic_start; | 137 | u64 logic_start; |
| 138 | 138 | ||
| 139 | u64 logic_end; | 139 | u64 logic_end; |
| 140 | 140 | ||
| 141 | int nsectors; | 141 | int nsectors; |
| 142 | 142 | ||
| 143 | int stripe_len; | 143 | int stripe_len; |
| 144 | 144 | ||
| 145 | atomic_t ref_count; | 145 | atomic_t ref_count; |
| 146 | 146 | ||
| 147 | struct list_head spages; | 147 | struct list_head spages; |
| 148 | 148 | ||
| 149 | /* Work of parity check and repair */ | 149 | /* Work of parity check and repair */ |
| 150 | struct btrfs_work work; | 150 | struct btrfs_work work; |
| 151 | 151 | ||
| 152 | /* Mark the parity blocks which have data */ | 152 | /* Mark the parity blocks which have data */ |
| 153 | unsigned long *dbitmap; | 153 | unsigned long *dbitmap; |
| 154 | 154 | ||
| 155 | /* | 155 | /* |
| 156 | * Mark the parity blocks which have data, but errors happen when | 156 | * Mark the parity blocks which have data, but errors happen when |
| 157 | * read data or check data | 157 | * read data or check data |
| 158 | */ | 158 | */ |
| 159 | unsigned long *ebitmap; | 159 | unsigned long *ebitmap; |
| 160 | 160 | ||
| 161 | unsigned long bitmap[0]; | 161 | unsigned long bitmap[0]; |
| 162 | }; | 162 | }; |
| 163 | 163 | ||
| 164 | struct scrub_wr_ctx { | 164 | struct scrub_wr_ctx { |
| 165 | struct scrub_bio *wr_curr_bio; | 165 | struct scrub_bio *wr_curr_bio; |
| 166 | struct btrfs_device *tgtdev; | 166 | struct btrfs_device *tgtdev; |
| 167 | int pages_per_wr_bio; /* <= SCRUB_PAGES_PER_WR_BIO */ | 167 | int pages_per_wr_bio; /* <= SCRUB_PAGES_PER_WR_BIO */ |
| 168 | atomic_t flush_all_writes; | 168 | atomic_t flush_all_writes; |
| 169 | struct mutex wr_lock; | 169 | struct mutex wr_lock; |
| 170 | }; | 170 | }; |
| 171 | 171 | ||
| 172 | struct scrub_ctx { | 172 | struct scrub_ctx { |
| 173 | struct scrub_bio *bios[SCRUB_BIOS_PER_SCTX]; | 173 | struct scrub_bio *bios[SCRUB_BIOS_PER_SCTX]; |
| 174 | struct btrfs_root *dev_root; | 174 | struct btrfs_root *dev_root; |
| 175 | int first_free; | 175 | int first_free; |
| 176 | int curr; | 176 | int curr; |
| 177 | atomic_t bios_in_flight; | 177 | atomic_t bios_in_flight; |
| 178 | atomic_t workers_pending; | 178 | atomic_t workers_pending; |
| 179 | spinlock_t list_lock; | 179 | spinlock_t list_lock; |
| 180 | wait_queue_head_t list_wait; | 180 | wait_queue_head_t list_wait; |
| 181 | u16 csum_size; | 181 | u16 csum_size; |
| 182 | struct list_head csum_list; | 182 | struct list_head csum_list; |
| 183 | atomic_t cancel_req; | 183 | atomic_t cancel_req; |
| 184 | int readonly; | 184 | int readonly; |
| 185 | int pages_per_rd_bio; | 185 | int pages_per_rd_bio; |
| 186 | u32 sectorsize; | 186 | u32 sectorsize; |
| 187 | u32 nodesize; | 187 | u32 nodesize; |
| 188 | 188 | ||
| 189 | int is_dev_replace; | 189 | int is_dev_replace; |
| 190 | struct scrub_wr_ctx wr_ctx; | 190 | struct scrub_wr_ctx wr_ctx; |
| 191 | 191 | ||
| 192 | /* | 192 | /* |
| 193 | * statistics | 193 | * statistics |
| 194 | */ | 194 | */ |
| 195 | struct btrfs_scrub_progress stat; | 195 | struct btrfs_scrub_progress stat; |
| 196 | spinlock_t stat_lock; | 196 | spinlock_t stat_lock; |
| 197 | }; | 197 | }; |
| 198 | 198 | ||
| 199 | struct scrub_fixup_nodatasum { | 199 | struct scrub_fixup_nodatasum { |
| 200 | struct scrub_ctx *sctx; | 200 | struct scrub_ctx *sctx; |
| 201 | struct btrfs_device *dev; | 201 | struct btrfs_device *dev; |
| 202 | u64 logical; | 202 | u64 logical; |
| 203 | struct btrfs_root *root; | 203 | struct btrfs_root *root; |
| 204 | struct btrfs_work work; | 204 | struct btrfs_work work; |
| 205 | int mirror_num; | 205 | int mirror_num; |
| 206 | }; | 206 | }; |
| 207 | 207 | ||
| 208 | struct scrub_nocow_inode { | 208 | struct scrub_nocow_inode { |
| 209 | u64 inum; | 209 | u64 inum; |
| 210 | u64 offset; | 210 | u64 offset; |
| 211 | u64 root; | 211 | u64 root; |
| 212 | struct list_head list; | 212 | struct list_head list; |
| 213 | }; | 213 | }; |
| 214 | 214 | ||
| 215 | struct scrub_copy_nocow_ctx { | 215 | struct scrub_copy_nocow_ctx { |
| 216 | struct scrub_ctx *sctx; | 216 | struct scrub_ctx *sctx; |
| 217 | u64 logical; | 217 | u64 logical; |
| 218 | u64 len; | 218 | u64 len; |
| 219 | int mirror_num; | 219 | int mirror_num; |
| 220 | u64 physical_for_dev_replace; | 220 | u64 physical_for_dev_replace; |
| 221 | struct list_head inodes; | 221 | struct list_head inodes; |
| 222 | struct btrfs_work work; | 222 | struct btrfs_work work; |
| 223 | }; | 223 | }; |
| 224 | 224 | ||
| 225 | struct scrub_warning { | 225 | struct scrub_warning { |
| 226 | struct btrfs_path *path; | 226 | struct btrfs_path *path; |
| 227 | u64 extent_item_size; | 227 | u64 extent_item_size; |
| 228 | const char *errstr; | 228 | const char *errstr; |
| 229 | sector_t sector; | 229 | sector_t sector; |
| 230 | u64 logical; | 230 | u64 logical; |
| 231 | struct btrfs_device *dev; | 231 | struct btrfs_device *dev; |
| 232 | }; | 232 | }; |
| 233 | 233 | ||
| 234 | static void scrub_pending_bio_inc(struct scrub_ctx *sctx); | 234 | static void scrub_pending_bio_inc(struct scrub_ctx *sctx); |
| 235 | static void scrub_pending_bio_dec(struct scrub_ctx *sctx); | 235 | static void scrub_pending_bio_dec(struct scrub_ctx *sctx); |
| 236 | static void scrub_pending_trans_workers_inc(struct scrub_ctx *sctx); | 236 | static void scrub_pending_trans_workers_inc(struct scrub_ctx *sctx); |
| 237 | static void scrub_pending_trans_workers_dec(struct scrub_ctx *sctx); | 237 | static void scrub_pending_trans_workers_dec(struct scrub_ctx *sctx); |
| 238 | static int scrub_handle_errored_block(struct scrub_block *sblock_to_check); | 238 | static int scrub_handle_errored_block(struct scrub_block *sblock_to_check); |
| 239 | static int scrub_setup_recheck_block(struct scrub_ctx *sctx, | 239 | static int scrub_setup_recheck_block(struct scrub_ctx *sctx, |
| 240 | struct btrfs_fs_info *fs_info, | 240 | struct btrfs_fs_info *fs_info, |
| 241 | struct scrub_block *original_sblock, | 241 | struct scrub_block *original_sblock, |
| 242 | u64 length, u64 logical, | 242 | u64 length, u64 logical, |
| 243 | struct scrub_block *sblocks_for_recheck); | 243 | struct scrub_block *sblocks_for_recheck); |
| 244 | static void scrub_recheck_block(struct btrfs_fs_info *fs_info, | 244 | static void scrub_recheck_block(struct btrfs_fs_info *fs_info, |
| 245 | struct scrub_block *sblock, int is_metadata, | 245 | struct scrub_block *sblock, int is_metadata, |
| 246 | int have_csum, u8 *csum, u64 generation, | 246 | int have_csum, u8 *csum, u64 generation, |
| 247 | u16 csum_size, int retry_failed_mirror); | 247 | u16 csum_size, int retry_failed_mirror); |
| 248 | static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info, | 248 | static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info, |
| 249 | struct scrub_block *sblock, | 249 | struct scrub_block *sblock, |
| 250 | int is_metadata, int have_csum, | 250 | int is_metadata, int have_csum, |
| 251 | const u8 *csum, u64 generation, | 251 | const u8 *csum, u64 generation, |
| 252 | u16 csum_size); | 252 | u16 csum_size); |
| 253 | static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad, | 253 | static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad, |
| 254 | struct scrub_block *sblock_good, | 254 | struct scrub_block *sblock_good, |
| 255 | int force_write); | 255 | int force_write); |
| 256 | static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad, | 256 | static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad, |
| 257 | struct scrub_block *sblock_good, | 257 | struct scrub_block *sblock_good, |
| 258 | int page_num, int force_write); | 258 | int page_num, int force_write); |
| 259 | static void scrub_write_block_to_dev_replace(struct scrub_block *sblock); | 259 | static void scrub_write_block_to_dev_replace(struct scrub_block *sblock); |
| 260 | static int scrub_write_page_to_dev_replace(struct scrub_block *sblock, | 260 | static int scrub_write_page_to_dev_replace(struct scrub_block *sblock, |
| 261 | int page_num); | 261 | int page_num); |
| 262 | static int scrub_checksum_data(struct scrub_block *sblock); | 262 | static int scrub_checksum_data(struct scrub_block *sblock); |
| 263 | static int scrub_checksum_tree_block(struct scrub_block *sblock); | 263 | static int scrub_checksum_tree_block(struct scrub_block *sblock); |
| 264 | static int scrub_checksum_super(struct scrub_block *sblock); | 264 | static int scrub_checksum_super(struct scrub_block *sblock); |
| 265 | static void scrub_block_get(struct scrub_block *sblock); | 265 | static void scrub_block_get(struct scrub_block *sblock); |
| 266 | static void scrub_block_put(struct scrub_block *sblock); | 266 | static void scrub_block_put(struct scrub_block *sblock); |
| 267 | static void scrub_page_get(struct scrub_page *spage); | 267 | static void scrub_page_get(struct scrub_page *spage); |
| 268 | static void scrub_page_put(struct scrub_page *spage); | 268 | static void scrub_page_put(struct scrub_page *spage); |
| 269 | static void scrub_parity_get(struct scrub_parity *sparity); | 269 | static void scrub_parity_get(struct scrub_parity *sparity); |
| 270 | static void scrub_parity_put(struct scrub_parity *sparity); | 270 | static void scrub_parity_put(struct scrub_parity *sparity); |
| 271 | static int scrub_add_page_to_rd_bio(struct scrub_ctx *sctx, | 271 | static int scrub_add_page_to_rd_bio(struct scrub_ctx *sctx, |
| 272 | struct scrub_page *spage); | 272 | struct scrub_page *spage); |
| 273 | static int scrub_pages(struct scrub_ctx *sctx, u64 logical, u64 len, | 273 | static int scrub_pages(struct scrub_ctx *sctx, u64 logical, u64 len, |
| 274 | u64 physical, struct btrfs_device *dev, u64 flags, | 274 | u64 physical, struct btrfs_device *dev, u64 flags, |
| 275 | u64 gen, int mirror_num, u8 *csum, int force, | 275 | u64 gen, int mirror_num, u8 *csum, int force, |
| 276 | u64 physical_for_dev_replace); | 276 | u64 physical_for_dev_replace); |
| 277 | static void scrub_bio_end_io(struct bio *bio, int err); | 277 | static void scrub_bio_end_io(struct bio *bio, int err); |
| 278 | static void scrub_bio_end_io_worker(struct btrfs_work *work); | 278 | static void scrub_bio_end_io_worker(struct btrfs_work *work); |
| 279 | static void scrub_block_complete(struct scrub_block *sblock); | 279 | static void scrub_block_complete(struct scrub_block *sblock); |
| 280 | static void scrub_remap_extent(struct btrfs_fs_info *fs_info, | 280 | static void scrub_remap_extent(struct btrfs_fs_info *fs_info, |
| 281 | u64 extent_logical, u64 extent_len, | 281 | u64 extent_logical, u64 extent_len, |
| 282 | u64 *extent_physical, | 282 | u64 *extent_physical, |
| 283 | struct btrfs_device **extent_dev, | 283 | struct btrfs_device **extent_dev, |
| 284 | int *extent_mirror_num); | 284 | int *extent_mirror_num); |
| 285 | static int scrub_setup_wr_ctx(struct scrub_ctx *sctx, | 285 | static int scrub_setup_wr_ctx(struct scrub_ctx *sctx, |
| 286 | struct scrub_wr_ctx *wr_ctx, | 286 | struct scrub_wr_ctx *wr_ctx, |
| 287 | struct btrfs_fs_info *fs_info, | 287 | struct btrfs_fs_info *fs_info, |
| 288 | struct btrfs_device *dev, | 288 | struct btrfs_device *dev, |
| 289 | int is_dev_replace); | 289 | int is_dev_replace); |
| 290 | static void scrub_free_wr_ctx(struct scrub_wr_ctx *wr_ctx); | 290 | static void scrub_free_wr_ctx(struct scrub_wr_ctx *wr_ctx); |
| 291 | static int scrub_add_page_to_wr_bio(struct scrub_ctx *sctx, | 291 | static int scrub_add_page_to_wr_bio(struct scrub_ctx *sctx, |
| 292 | struct scrub_page *spage); | 292 | struct scrub_page *spage); |
| 293 | static void scrub_wr_submit(struct scrub_ctx *sctx); | 293 | static void scrub_wr_submit(struct scrub_ctx *sctx); |
| 294 | static void scrub_wr_bio_end_io(struct bio *bio, int err); | 294 | static void scrub_wr_bio_end_io(struct bio *bio, int err); |
| 295 | static void scrub_wr_bio_end_io_worker(struct btrfs_work *work); | 295 | static void scrub_wr_bio_end_io_worker(struct btrfs_work *work); |
| 296 | static int write_page_nocow(struct scrub_ctx *sctx, | 296 | static int write_page_nocow(struct scrub_ctx *sctx, |
| 297 | u64 physical_for_dev_replace, struct page *page); | 297 | u64 physical_for_dev_replace, struct page *page); |
| 298 | static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, | 298 | static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, |
| 299 | struct scrub_copy_nocow_ctx *ctx); | 299 | struct scrub_copy_nocow_ctx *ctx); |
| 300 | static int copy_nocow_pages(struct scrub_ctx *sctx, u64 logical, u64 len, | 300 | static int copy_nocow_pages(struct scrub_ctx *sctx, u64 logical, u64 len, |
| 301 | int mirror_num, u64 physical_for_dev_replace); | 301 | int mirror_num, u64 physical_for_dev_replace); |
| 302 | static void copy_nocow_pages_worker(struct btrfs_work *work); | 302 | static void copy_nocow_pages_worker(struct btrfs_work *work); |
| 303 | static void __scrub_blocked_if_needed(struct btrfs_fs_info *fs_info); | 303 | static void __scrub_blocked_if_needed(struct btrfs_fs_info *fs_info); |
| 304 | static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info); | 304 | static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info); |
| 305 | 305 | ||
| 306 | 306 | ||
| 307 | static void scrub_pending_bio_inc(struct scrub_ctx *sctx) | 307 | static void scrub_pending_bio_inc(struct scrub_ctx *sctx) |
| 308 | { | 308 | { |
| 309 | atomic_inc(&sctx->bios_in_flight); | 309 | atomic_inc(&sctx->bios_in_flight); |
| 310 | } | 310 | } |
| 311 | 311 | ||
| 312 | static void scrub_pending_bio_dec(struct scrub_ctx *sctx) | 312 | static void scrub_pending_bio_dec(struct scrub_ctx *sctx) |
| 313 | { | 313 | { |
| 314 | atomic_dec(&sctx->bios_in_flight); | 314 | atomic_dec(&sctx->bios_in_flight); |
| 315 | wake_up(&sctx->list_wait); | 315 | wake_up(&sctx->list_wait); |
| 316 | } | 316 | } |
| 317 | 317 | ||
| 318 | static void __scrub_blocked_if_needed(struct btrfs_fs_info *fs_info) | 318 | static void __scrub_blocked_if_needed(struct btrfs_fs_info *fs_info) |
| 319 | { | 319 | { |
| 320 | while (atomic_read(&fs_info->scrub_pause_req)) { | 320 | while (atomic_read(&fs_info->scrub_pause_req)) { |
| 321 | mutex_unlock(&fs_info->scrub_lock); | 321 | mutex_unlock(&fs_info->scrub_lock); |
| 322 | wait_event(fs_info->scrub_pause_wait, | 322 | wait_event(fs_info->scrub_pause_wait, |
| 323 | atomic_read(&fs_info->scrub_pause_req) == 0); | 323 | atomic_read(&fs_info->scrub_pause_req) == 0); |
| 324 | mutex_lock(&fs_info->scrub_lock); | 324 | mutex_lock(&fs_info->scrub_lock); |
| 325 | } | 325 | } |
| 326 | } | 326 | } |
| 327 | 327 | ||
| 328 | static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info) | 328 | static void scrub_blocked_if_needed(struct btrfs_fs_info *fs_info) |
| 329 | { | 329 | { |
| 330 | atomic_inc(&fs_info->scrubs_paused); | 330 | atomic_inc(&fs_info->scrubs_paused); |
| 331 | wake_up(&fs_info->scrub_pause_wait); | 331 | wake_up(&fs_info->scrub_pause_wait); |
| 332 | 332 | ||
| 333 | mutex_lock(&fs_info->scrub_lock); | 333 | mutex_lock(&fs_info->scrub_lock); |
| 334 | __scrub_blocked_if_needed(fs_info); | 334 | __scrub_blocked_if_needed(fs_info); |
| 335 | atomic_dec(&fs_info->scrubs_paused); | 335 | atomic_dec(&fs_info->scrubs_paused); |
| 336 | mutex_unlock(&fs_info->scrub_lock); | 336 | mutex_unlock(&fs_info->scrub_lock); |
| 337 | 337 | ||
| 338 | wake_up(&fs_info->scrub_pause_wait); | 338 | wake_up(&fs_info->scrub_pause_wait); |
| 339 | } | 339 | } |
| 340 | 340 | ||
| 341 | /* | 341 | /* |
| 342 | * used for workers that require transaction commits (i.e., for the | 342 | * used for workers that require transaction commits (i.e., for the |
| 343 | * NOCOW case) | 343 | * NOCOW case) |
| 344 | */ | 344 | */ |
| 345 | static void scrub_pending_trans_workers_inc(struct scrub_ctx *sctx) | 345 | static void scrub_pending_trans_workers_inc(struct scrub_ctx *sctx) |
| 346 | { | 346 | { |
| 347 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; | 347 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; |
| 348 | 348 | ||
| 349 | /* | 349 | /* |
| 350 | * increment scrubs_running to prevent cancel requests from | 350 | * increment scrubs_running to prevent cancel requests from |
| 351 | * completing as long as a worker is running. we must also | 351 | * completing as long as a worker is running. we must also |
| 352 | * increment scrubs_paused to prevent deadlocking on pause | 352 | * increment scrubs_paused to prevent deadlocking on pause |
| 353 | * requests used for transactions commits (as the worker uses a | 353 | * requests used for transactions commits (as the worker uses a |
| 354 | * transaction context). it is safe to regard the worker | 354 | * transaction context). it is safe to regard the worker |
| 355 | * as paused for all matters practical. effectively, we only | 355 | * as paused for all matters practical. effectively, we only |
| 356 | * avoid cancellation requests from completing. | 356 | * avoid cancellation requests from completing. |
| 357 | */ | 357 | */ |
| 358 | mutex_lock(&fs_info->scrub_lock); | 358 | mutex_lock(&fs_info->scrub_lock); |
| 359 | atomic_inc(&fs_info->scrubs_running); | 359 | atomic_inc(&fs_info->scrubs_running); |
| 360 | atomic_inc(&fs_info->scrubs_paused); | 360 | atomic_inc(&fs_info->scrubs_paused); |
| 361 | mutex_unlock(&fs_info->scrub_lock); | 361 | mutex_unlock(&fs_info->scrub_lock); |
| 362 | 362 | ||
| 363 | /* | 363 | /* |
| 364 | * check if @scrubs_running=@scrubs_paused condition | 364 | * check if @scrubs_running=@scrubs_paused condition |
| 365 | * inside wait_event() is not an atomic operation. | 365 | * inside wait_event() is not an atomic operation. |
| 366 | * which means we may inc/dec @scrub_running/paused | 366 | * which means we may inc/dec @scrub_running/paused |
| 367 | * at any time. Let's wake up @scrub_pause_wait as | 367 | * at any time. Let's wake up @scrub_pause_wait as |
| 368 | * much as we can to let commit transaction blocked less. | 368 | * much as we can to let commit transaction blocked less. |
| 369 | */ | 369 | */ |
| 370 | wake_up(&fs_info->scrub_pause_wait); | 370 | wake_up(&fs_info->scrub_pause_wait); |
| 371 | 371 | ||
| 372 | atomic_inc(&sctx->workers_pending); | 372 | atomic_inc(&sctx->workers_pending); |
| 373 | } | 373 | } |
| 374 | 374 | ||
| 375 | /* used for workers that require transaction commits */ | 375 | /* used for workers that require transaction commits */ |
| 376 | static void scrub_pending_trans_workers_dec(struct scrub_ctx *sctx) | 376 | static void scrub_pending_trans_workers_dec(struct scrub_ctx *sctx) |
| 377 | { | 377 | { |
| 378 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; | 378 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; |
| 379 | 379 | ||
| 380 | /* | 380 | /* |
| 381 | * see scrub_pending_trans_workers_inc() why we're pretending | 381 | * see scrub_pending_trans_workers_inc() why we're pretending |
| 382 | * to be paused in the scrub counters | 382 | * to be paused in the scrub counters |
| 383 | */ | 383 | */ |
| 384 | mutex_lock(&fs_info->scrub_lock); | 384 | mutex_lock(&fs_info->scrub_lock); |
| 385 | atomic_dec(&fs_info->scrubs_running); | 385 | atomic_dec(&fs_info->scrubs_running); |
| 386 | atomic_dec(&fs_info->scrubs_paused); | 386 | atomic_dec(&fs_info->scrubs_paused); |
| 387 | mutex_unlock(&fs_info->scrub_lock); | 387 | mutex_unlock(&fs_info->scrub_lock); |
| 388 | atomic_dec(&sctx->workers_pending); | 388 | atomic_dec(&sctx->workers_pending); |
| 389 | wake_up(&fs_info->scrub_pause_wait); | 389 | wake_up(&fs_info->scrub_pause_wait); |
| 390 | wake_up(&sctx->list_wait); | 390 | wake_up(&sctx->list_wait); |
| 391 | } | 391 | } |
| 392 | 392 | ||
| 393 | static void scrub_free_csums(struct scrub_ctx *sctx) | 393 | static void scrub_free_csums(struct scrub_ctx *sctx) |
| 394 | { | 394 | { |
| 395 | while (!list_empty(&sctx->csum_list)) { | 395 | while (!list_empty(&sctx->csum_list)) { |
| 396 | struct btrfs_ordered_sum *sum; | 396 | struct btrfs_ordered_sum *sum; |
| 397 | sum = list_first_entry(&sctx->csum_list, | 397 | sum = list_first_entry(&sctx->csum_list, |
| 398 | struct btrfs_ordered_sum, list); | 398 | struct btrfs_ordered_sum, list); |
| 399 | list_del(&sum->list); | 399 | list_del(&sum->list); |
| 400 | kfree(sum); | 400 | kfree(sum); |
| 401 | } | 401 | } |
| 402 | } | 402 | } |
| 403 | 403 | ||
| 404 | static noinline_for_stack void scrub_free_ctx(struct scrub_ctx *sctx) | 404 | static noinline_for_stack void scrub_free_ctx(struct scrub_ctx *sctx) |
| 405 | { | 405 | { |
| 406 | int i; | 406 | int i; |
| 407 | 407 | ||
| 408 | if (!sctx) | 408 | if (!sctx) |
| 409 | return; | 409 | return; |
| 410 | 410 | ||
| 411 | scrub_free_wr_ctx(&sctx->wr_ctx); | 411 | scrub_free_wr_ctx(&sctx->wr_ctx); |
| 412 | 412 | ||
| 413 | /* this can happen when scrub is cancelled */ | 413 | /* this can happen when scrub is cancelled */ |
| 414 | if (sctx->curr != -1) { | 414 | if (sctx->curr != -1) { |
| 415 | struct scrub_bio *sbio = sctx->bios[sctx->curr]; | 415 | struct scrub_bio *sbio = sctx->bios[sctx->curr]; |
| 416 | 416 | ||
| 417 | for (i = 0; i < sbio->page_count; i++) { | 417 | for (i = 0; i < sbio->page_count; i++) { |
| 418 | WARN_ON(!sbio->pagev[i]->page); | 418 | WARN_ON(!sbio->pagev[i]->page); |
| 419 | scrub_block_put(sbio->pagev[i]->sblock); | 419 | scrub_block_put(sbio->pagev[i]->sblock); |
| 420 | } | 420 | } |
| 421 | bio_put(sbio->bio); | 421 | bio_put(sbio->bio); |
| 422 | } | 422 | } |
| 423 | 423 | ||
| 424 | for (i = 0; i < SCRUB_BIOS_PER_SCTX; ++i) { | 424 | for (i = 0; i < SCRUB_BIOS_PER_SCTX; ++i) { |
| 425 | struct scrub_bio *sbio = sctx->bios[i]; | 425 | struct scrub_bio *sbio = sctx->bios[i]; |
| 426 | 426 | ||
| 427 | if (!sbio) | 427 | if (!sbio) |
| 428 | break; | 428 | break; |
| 429 | kfree(sbio); | 429 | kfree(sbio); |
| 430 | } | 430 | } |
| 431 | 431 | ||
| 432 | scrub_free_csums(sctx); | 432 | scrub_free_csums(sctx); |
| 433 | kfree(sctx); | 433 | kfree(sctx); |
| 434 | } | 434 | } |
| 435 | 435 | ||
| 436 | static noinline_for_stack | 436 | static noinline_for_stack |
| 437 | struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev, int is_dev_replace) | 437 | struct scrub_ctx *scrub_setup_ctx(struct btrfs_device *dev, int is_dev_replace) |
| 438 | { | 438 | { |
| 439 | struct scrub_ctx *sctx; | 439 | struct scrub_ctx *sctx; |
| 440 | int i; | 440 | int i; |
| 441 | struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; | 441 | struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; |
| 442 | int pages_per_rd_bio; | 442 | int pages_per_rd_bio; |
| 443 | int ret; | 443 | int ret; |
| 444 | 444 | ||
| 445 | /* | 445 | /* |
| 446 | * the setting of pages_per_rd_bio is correct for scrub but might | 446 | * the setting of pages_per_rd_bio is correct for scrub but might |
| 447 | * be wrong for the dev_replace code where we might read from | 447 | * be wrong for the dev_replace code where we might read from |
| 448 | * different devices in the initial huge bios. However, that | 448 | * different devices in the initial huge bios. However, that |
| 449 | * code is able to correctly handle the case when adding a page | 449 | * code is able to correctly handle the case when adding a page |
| 450 | * to a bio fails. | 450 | * to a bio fails. |
| 451 | */ | 451 | */ |
| 452 | if (dev->bdev) | 452 | if (dev->bdev) |
| 453 | pages_per_rd_bio = min_t(int, SCRUB_PAGES_PER_RD_BIO, | 453 | pages_per_rd_bio = min_t(int, SCRUB_PAGES_PER_RD_BIO, |
| 454 | bio_get_nr_vecs(dev->bdev)); | 454 | bio_get_nr_vecs(dev->bdev)); |
| 455 | else | 455 | else |
| 456 | pages_per_rd_bio = SCRUB_PAGES_PER_RD_BIO; | 456 | pages_per_rd_bio = SCRUB_PAGES_PER_RD_BIO; |
| 457 | sctx = kzalloc(sizeof(*sctx), GFP_NOFS); | 457 | sctx = kzalloc(sizeof(*sctx), GFP_NOFS); |
| 458 | if (!sctx) | 458 | if (!sctx) |
| 459 | goto nomem; | 459 | goto nomem; |
| 460 | sctx->is_dev_replace = is_dev_replace; | 460 | sctx->is_dev_replace = is_dev_replace; |
| 461 | sctx->pages_per_rd_bio = pages_per_rd_bio; | 461 | sctx->pages_per_rd_bio = pages_per_rd_bio; |
| 462 | sctx->curr = -1; | 462 | sctx->curr = -1; |
| 463 | sctx->dev_root = dev->dev_root; | 463 | sctx->dev_root = dev->dev_root; |
| 464 | for (i = 0; i < SCRUB_BIOS_PER_SCTX; ++i) { | 464 | for (i = 0; i < SCRUB_BIOS_PER_SCTX; ++i) { |
| 465 | struct scrub_bio *sbio; | 465 | struct scrub_bio *sbio; |
| 466 | 466 | ||
| 467 | sbio = kzalloc(sizeof(*sbio), GFP_NOFS); | 467 | sbio = kzalloc(sizeof(*sbio), GFP_NOFS); |
| 468 | if (!sbio) | 468 | if (!sbio) |
| 469 | goto nomem; | 469 | goto nomem; |
| 470 | sctx->bios[i] = sbio; | 470 | sctx->bios[i] = sbio; |
| 471 | 471 | ||
| 472 | sbio->index = i; | 472 | sbio->index = i; |
| 473 | sbio->sctx = sctx; | 473 | sbio->sctx = sctx; |
| 474 | sbio->page_count = 0; | 474 | sbio->page_count = 0; |
| 475 | btrfs_init_work(&sbio->work, btrfs_scrub_helper, | 475 | btrfs_init_work(&sbio->work, btrfs_scrub_helper, |
| 476 | scrub_bio_end_io_worker, NULL, NULL); | 476 | scrub_bio_end_io_worker, NULL, NULL); |
| 477 | 477 | ||
| 478 | if (i != SCRUB_BIOS_PER_SCTX - 1) | 478 | if (i != SCRUB_BIOS_PER_SCTX - 1) |
| 479 | sctx->bios[i]->next_free = i + 1; | 479 | sctx->bios[i]->next_free = i + 1; |
| 480 | else | 480 | else |
| 481 | sctx->bios[i]->next_free = -1; | 481 | sctx->bios[i]->next_free = -1; |
| 482 | } | 482 | } |
| 483 | sctx->first_free = 0; | 483 | sctx->first_free = 0; |
| 484 | sctx->nodesize = dev->dev_root->nodesize; | 484 | sctx->nodesize = dev->dev_root->nodesize; |
| 485 | sctx->sectorsize = dev->dev_root->sectorsize; | 485 | sctx->sectorsize = dev->dev_root->sectorsize; |
| 486 | atomic_set(&sctx->bios_in_flight, 0); | 486 | atomic_set(&sctx->bios_in_flight, 0); |
| 487 | atomic_set(&sctx->workers_pending, 0); | 487 | atomic_set(&sctx->workers_pending, 0); |
| 488 | atomic_set(&sctx->cancel_req, 0); | 488 | atomic_set(&sctx->cancel_req, 0); |
| 489 | sctx->csum_size = btrfs_super_csum_size(fs_info->super_copy); | 489 | sctx->csum_size = btrfs_super_csum_size(fs_info->super_copy); |
| 490 | INIT_LIST_HEAD(&sctx->csum_list); | 490 | INIT_LIST_HEAD(&sctx->csum_list); |
| 491 | 491 | ||
| 492 | spin_lock_init(&sctx->list_lock); | 492 | spin_lock_init(&sctx->list_lock); |
| 493 | spin_lock_init(&sctx->stat_lock); | 493 | spin_lock_init(&sctx->stat_lock); |
| 494 | init_waitqueue_head(&sctx->list_wait); | 494 | init_waitqueue_head(&sctx->list_wait); |
| 495 | 495 | ||
| 496 | ret = scrub_setup_wr_ctx(sctx, &sctx->wr_ctx, fs_info, | 496 | ret = scrub_setup_wr_ctx(sctx, &sctx->wr_ctx, fs_info, |
| 497 | fs_info->dev_replace.tgtdev, is_dev_replace); | 497 | fs_info->dev_replace.tgtdev, is_dev_replace); |
| 498 | if (ret) { | 498 | if (ret) { |
| 499 | scrub_free_ctx(sctx); | 499 | scrub_free_ctx(sctx); |
| 500 | return ERR_PTR(ret); | 500 | return ERR_PTR(ret); |
| 501 | } | 501 | } |
| 502 | return sctx; | 502 | return sctx; |
| 503 | 503 | ||
| 504 | nomem: | 504 | nomem: |
| 505 | scrub_free_ctx(sctx); | 505 | scrub_free_ctx(sctx); |
| 506 | return ERR_PTR(-ENOMEM); | 506 | return ERR_PTR(-ENOMEM); |
| 507 | } | 507 | } |
| 508 | 508 | ||
| 509 | static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, | 509 | static int scrub_print_warning_inode(u64 inum, u64 offset, u64 root, |
| 510 | void *warn_ctx) | 510 | void *warn_ctx) |
| 511 | { | 511 | { |
| 512 | u64 isize; | 512 | u64 isize; |
| 513 | u32 nlink; | 513 | u32 nlink; |
| 514 | int ret; | 514 | int ret; |
| 515 | int i; | 515 | int i; |
| 516 | struct extent_buffer *eb; | 516 | struct extent_buffer *eb; |
| 517 | struct btrfs_inode_item *inode_item; | 517 | struct btrfs_inode_item *inode_item; |
| 518 | struct scrub_warning *swarn = warn_ctx; | 518 | struct scrub_warning *swarn = warn_ctx; |
| 519 | struct btrfs_fs_info *fs_info = swarn->dev->dev_root->fs_info; | 519 | struct btrfs_fs_info *fs_info = swarn->dev->dev_root->fs_info; |
| 520 | struct inode_fs_paths *ipath = NULL; | 520 | struct inode_fs_paths *ipath = NULL; |
| 521 | struct btrfs_root *local_root; | 521 | struct btrfs_root *local_root; |
| 522 | struct btrfs_key root_key; | 522 | struct btrfs_key root_key; |
| 523 | 523 | ||
| 524 | root_key.objectid = root; | 524 | root_key.objectid = root; |
| 525 | root_key.type = BTRFS_ROOT_ITEM_KEY; | 525 | root_key.type = BTRFS_ROOT_ITEM_KEY; |
| 526 | root_key.offset = (u64)-1; | 526 | root_key.offset = (u64)-1; |
| 527 | local_root = btrfs_read_fs_root_no_name(fs_info, &root_key); | 527 | local_root = btrfs_read_fs_root_no_name(fs_info, &root_key); |
| 528 | if (IS_ERR(local_root)) { | 528 | if (IS_ERR(local_root)) { |
| 529 | ret = PTR_ERR(local_root); | 529 | ret = PTR_ERR(local_root); |
| 530 | goto err; | 530 | goto err; |
| 531 | } | 531 | } |
| 532 | 532 | ||
| 533 | ret = inode_item_info(inum, 0, local_root, swarn->path); | 533 | ret = inode_item_info(inum, 0, local_root, swarn->path); |
| 534 | if (ret) { | 534 | if (ret) { |
| 535 | btrfs_release_path(swarn->path); | 535 | btrfs_release_path(swarn->path); |
| 536 | goto err; | 536 | goto err; |
| 537 | } | 537 | } |
| 538 | 538 | ||
| 539 | eb = swarn->path->nodes[0]; | 539 | eb = swarn->path->nodes[0]; |
| 540 | inode_item = btrfs_item_ptr(eb, swarn->path->slots[0], | 540 | inode_item = btrfs_item_ptr(eb, swarn->path->slots[0], |
| 541 | struct btrfs_inode_item); | 541 | struct btrfs_inode_item); |
| 542 | isize = btrfs_inode_size(eb, inode_item); | 542 | isize = btrfs_inode_size(eb, inode_item); |
| 543 | nlink = btrfs_inode_nlink(eb, inode_item); | 543 | nlink = btrfs_inode_nlink(eb, inode_item); |
| 544 | btrfs_release_path(swarn->path); | 544 | btrfs_release_path(swarn->path); |
| 545 | 545 | ||
| 546 | ipath = init_ipath(4096, local_root, swarn->path); | 546 | ipath = init_ipath(4096, local_root, swarn->path); |
| 547 | if (IS_ERR(ipath)) { | 547 | if (IS_ERR(ipath)) { |
| 548 | ret = PTR_ERR(ipath); | 548 | ret = PTR_ERR(ipath); |
| 549 | ipath = NULL; | 549 | ipath = NULL; |
| 550 | goto err; | 550 | goto err; |
| 551 | } | 551 | } |
| 552 | ret = paths_from_inode(inum, ipath); | 552 | ret = paths_from_inode(inum, ipath); |
| 553 | 553 | ||
| 554 | if (ret < 0) | 554 | if (ret < 0) |
| 555 | goto err; | 555 | goto err; |
| 556 | 556 | ||
| 557 | /* | 557 | /* |
| 558 | * we deliberately ignore the bit ipath might have been too small to | 558 | * we deliberately ignore the bit ipath might have been too small to |
| 559 | * hold all of the paths here | 559 | * hold all of the paths here |
| 560 | */ | 560 | */ |
| 561 | for (i = 0; i < ipath->fspath->elem_cnt; ++i) | 561 | for (i = 0; i < ipath->fspath->elem_cnt; ++i) |
| 562 | printk_in_rcu(KERN_WARNING "BTRFS: %s at logical %llu on dev " | 562 | printk_in_rcu(KERN_WARNING "BTRFS: %s at logical %llu on dev " |
| 563 | "%s, sector %llu, root %llu, inode %llu, offset %llu, " | 563 | "%s, sector %llu, root %llu, inode %llu, offset %llu, " |
| 564 | "length %llu, links %u (path: %s)\n", swarn->errstr, | 564 | "length %llu, links %u (path: %s)\n", swarn->errstr, |
| 565 | swarn->logical, rcu_str_deref(swarn->dev->name), | 565 | swarn->logical, rcu_str_deref(swarn->dev->name), |
| 566 | (unsigned long long)swarn->sector, root, inum, offset, | 566 | (unsigned long long)swarn->sector, root, inum, offset, |
| 567 | min(isize - offset, (u64)PAGE_SIZE), nlink, | 567 | min(isize - offset, (u64)PAGE_SIZE), nlink, |
| 568 | (char *)(unsigned long)ipath->fspath->val[i]); | 568 | (char *)(unsigned long)ipath->fspath->val[i]); |
| 569 | 569 | ||
| 570 | free_ipath(ipath); | 570 | free_ipath(ipath); |
| 571 | return 0; | 571 | return 0; |
| 572 | 572 | ||
| 573 | err: | 573 | err: |
| 574 | printk_in_rcu(KERN_WARNING "BTRFS: %s at logical %llu on dev " | 574 | printk_in_rcu(KERN_WARNING "BTRFS: %s at logical %llu on dev " |
| 575 | "%s, sector %llu, root %llu, inode %llu, offset %llu: path " | 575 | "%s, sector %llu, root %llu, inode %llu, offset %llu: path " |
| 576 | "resolving failed with ret=%d\n", swarn->errstr, | 576 | "resolving failed with ret=%d\n", swarn->errstr, |
| 577 | swarn->logical, rcu_str_deref(swarn->dev->name), | 577 | swarn->logical, rcu_str_deref(swarn->dev->name), |
| 578 | (unsigned long long)swarn->sector, root, inum, offset, ret); | 578 | (unsigned long long)swarn->sector, root, inum, offset, ret); |
| 579 | 579 | ||
| 580 | free_ipath(ipath); | 580 | free_ipath(ipath); |
| 581 | return 0; | 581 | return 0; |
| 582 | } | 582 | } |
| 583 | 583 | ||
| 584 | static void scrub_print_warning(const char *errstr, struct scrub_block *sblock) | 584 | static void scrub_print_warning(const char *errstr, struct scrub_block *sblock) |
| 585 | { | 585 | { |
| 586 | struct btrfs_device *dev; | 586 | struct btrfs_device *dev; |
| 587 | struct btrfs_fs_info *fs_info; | 587 | struct btrfs_fs_info *fs_info; |
| 588 | struct btrfs_path *path; | 588 | struct btrfs_path *path; |
| 589 | struct btrfs_key found_key; | 589 | struct btrfs_key found_key; |
| 590 | struct extent_buffer *eb; | 590 | struct extent_buffer *eb; |
| 591 | struct btrfs_extent_item *ei; | 591 | struct btrfs_extent_item *ei; |
| 592 | struct scrub_warning swarn; | 592 | struct scrub_warning swarn; |
| 593 | unsigned long ptr = 0; | 593 | unsigned long ptr = 0; |
| 594 | u64 extent_item_pos; | 594 | u64 extent_item_pos; |
| 595 | u64 flags = 0; | 595 | u64 flags = 0; |
| 596 | u64 ref_root; | 596 | u64 ref_root; |
| 597 | u32 item_size; | 597 | u32 item_size; |
| 598 | u8 ref_level; | 598 | u8 ref_level; |
| 599 | int ret; | 599 | int ret; |
| 600 | 600 | ||
| 601 | WARN_ON(sblock->page_count < 1); | 601 | WARN_ON(sblock->page_count < 1); |
| 602 | dev = sblock->pagev[0]->dev; | 602 | dev = sblock->pagev[0]->dev; |
| 603 | fs_info = sblock->sctx->dev_root->fs_info; | 603 | fs_info = sblock->sctx->dev_root->fs_info; |
| 604 | 604 | ||
| 605 | path = btrfs_alloc_path(); | 605 | path = btrfs_alloc_path(); |
| 606 | if (!path) | 606 | if (!path) |
| 607 | return; | 607 | return; |
| 608 | 608 | ||
| 609 | swarn.sector = (sblock->pagev[0]->physical) >> 9; | 609 | swarn.sector = (sblock->pagev[0]->physical) >> 9; |
| 610 | swarn.logical = sblock->pagev[0]->logical; | 610 | swarn.logical = sblock->pagev[0]->logical; |
| 611 | swarn.errstr = errstr; | 611 | swarn.errstr = errstr; |
| 612 | swarn.dev = NULL; | 612 | swarn.dev = NULL; |
| 613 | 613 | ||
| 614 | ret = extent_from_logical(fs_info, swarn.logical, path, &found_key, | 614 | ret = extent_from_logical(fs_info, swarn.logical, path, &found_key, |
| 615 | &flags); | 615 | &flags); |
| 616 | if (ret < 0) | 616 | if (ret < 0) |
| 617 | goto out; | 617 | goto out; |
| 618 | 618 | ||
| 619 | extent_item_pos = swarn.logical - found_key.objectid; | 619 | extent_item_pos = swarn.logical - found_key.objectid; |
| 620 | swarn.extent_item_size = found_key.offset; | 620 | swarn.extent_item_size = found_key.offset; |
| 621 | 621 | ||
| 622 | eb = path->nodes[0]; | 622 | eb = path->nodes[0]; |
| 623 | ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); | 623 | ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item); |
| 624 | item_size = btrfs_item_size_nr(eb, path->slots[0]); | 624 | item_size = btrfs_item_size_nr(eb, path->slots[0]); |
| 625 | 625 | ||
| 626 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { | 626 | if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { |
| 627 | do { | 627 | do { |
| 628 | ret = tree_backref_for_extent(&ptr, eb, &found_key, ei, | 628 | ret = tree_backref_for_extent(&ptr, eb, &found_key, ei, |
| 629 | item_size, &ref_root, | 629 | item_size, &ref_root, |
| 630 | &ref_level); | 630 | &ref_level); |
| 631 | printk_in_rcu(KERN_WARNING | 631 | printk_in_rcu(KERN_WARNING |
| 632 | "BTRFS: %s at logical %llu on dev %s, " | 632 | "BTRFS: %s at logical %llu on dev %s, " |
| 633 | "sector %llu: metadata %s (level %d) in tree " | 633 | "sector %llu: metadata %s (level %d) in tree " |
| 634 | "%llu\n", errstr, swarn.logical, | 634 | "%llu\n", errstr, swarn.logical, |
| 635 | rcu_str_deref(dev->name), | 635 | rcu_str_deref(dev->name), |
| 636 | (unsigned long long)swarn.sector, | 636 | (unsigned long long)swarn.sector, |
| 637 | ref_level ? "node" : "leaf", | 637 | ref_level ? "node" : "leaf", |
| 638 | ret < 0 ? -1 : ref_level, | 638 | ret < 0 ? -1 : ref_level, |
| 639 | ret < 0 ? -1 : ref_root); | 639 | ret < 0 ? -1 : ref_root); |
| 640 | } while (ret != 1); | 640 | } while (ret != 1); |
| 641 | btrfs_release_path(path); | 641 | btrfs_release_path(path); |
| 642 | } else { | 642 | } else { |
| 643 | btrfs_release_path(path); | 643 | btrfs_release_path(path); |
| 644 | swarn.path = path; | 644 | swarn.path = path; |
| 645 | swarn.dev = dev; | 645 | swarn.dev = dev; |
| 646 | iterate_extent_inodes(fs_info, found_key.objectid, | 646 | iterate_extent_inodes(fs_info, found_key.objectid, |
| 647 | extent_item_pos, 1, | 647 | extent_item_pos, 1, |
| 648 | scrub_print_warning_inode, &swarn); | 648 | scrub_print_warning_inode, &swarn); |
| 649 | } | 649 | } |
| 650 | 650 | ||
| 651 | out: | 651 | out: |
| 652 | btrfs_free_path(path); | 652 | btrfs_free_path(path); |
| 653 | } | 653 | } |
| 654 | 654 | ||
| 655 | static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *fixup_ctx) | 655 | static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *fixup_ctx) |
| 656 | { | 656 | { |
| 657 | struct page *page = NULL; | 657 | struct page *page = NULL; |
| 658 | unsigned long index; | 658 | unsigned long index; |
| 659 | struct scrub_fixup_nodatasum *fixup = fixup_ctx; | 659 | struct scrub_fixup_nodatasum *fixup = fixup_ctx; |
| 660 | int ret; | 660 | int ret; |
| 661 | int corrected = 0; | 661 | int corrected = 0; |
| 662 | struct btrfs_key key; | 662 | struct btrfs_key key; |
| 663 | struct inode *inode = NULL; | 663 | struct inode *inode = NULL; |
| 664 | struct btrfs_fs_info *fs_info; | 664 | struct btrfs_fs_info *fs_info; |
| 665 | u64 end = offset + PAGE_SIZE - 1; | 665 | u64 end = offset + PAGE_SIZE - 1; |
| 666 | struct btrfs_root *local_root; | 666 | struct btrfs_root *local_root; |
| 667 | int srcu_index; | 667 | int srcu_index; |
| 668 | 668 | ||
| 669 | key.objectid = root; | 669 | key.objectid = root; |
| 670 | key.type = BTRFS_ROOT_ITEM_KEY; | 670 | key.type = BTRFS_ROOT_ITEM_KEY; |
| 671 | key.offset = (u64)-1; | 671 | key.offset = (u64)-1; |
| 672 | 672 | ||
| 673 | fs_info = fixup->root->fs_info; | 673 | fs_info = fixup->root->fs_info; |
| 674 | srcu_index = srcu_read_lock(&fs_info->subvol_srcu); | 674 | srcu_index = srcu_read_lock(&fs_info->subvol_srcu); |
| 675 | 675 | ||
| 676 | local_root = btrfs_read_fs_root_no_name(fs_info, &key); | 676 | local_root = btrfs_read_fs_root_no_name(fs_info, &key); |
| 677 | if (IS_ERR(local_root)) { | 677 | if (IS_ERR(local_root)) { |
| 678 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); | 678 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); |
| 679 | return PTR_ERR(local_root); | 679 | return PTR_ERR(local_root); |
| 680 | } | 680 | } |
| 681 | 681 | ||
| 682 | key.type = BTRFS_INODE_ITEM_KEY; | 682 | key.type = BTRFS_INODE_ITEM_KEY; |
| 683 | key.objectid = inum; | 683 | key.objectid = inum; |
| 684 | key.offset = 0; | 684 | key.offset = 0; |
| 685 | inode = btrfs_iget(fs_info->sb, &key, local_root, NULL); | 685 | inode = btrfs_iget(fs_info->sb, &key, local_root, NULL); |
| 686 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); | 686 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); |
| 687 | if (IS_ERR(inode)) | 687 | if (IS_ERR(inode)) |
| 688 | return PTR_ERR(inode); | 688 | return PTR_ERR(inode); |
| 689 | 689 | ||
| 690 | index = offset >> PAGE_CACHE_SHIFT; | 690 | index = offset >> PAGE_CACHE_SHIFT; |
| 691 | 691 | ||
| 692 | page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); | 692 | page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); |
| 693 | if (!page) { | 693 | if (!page) { |
| 694 | ret = -ENOMEM; | 694 | ret = -ENOMEM; |
| 695 | goto out; | 695 | goto out; |
| 696 | } | 696 | } |
| 697 | 697 | ||
| 698 | if (PageUptodate(page)) { | 698 | if (PageUptodate(page)) { |
| 699 | if (PageDirty(page)) { | 699 | if (PageDirty(page)) { |
| 700 | /* | 700 | /* |
| 701 | * we need to write the data to the defect sector. the | 701 | * we need to write the data to the defect sector. the |
| 702 | * data that was in that sector is not in memory, | 702 | * data that was in that sector is not in memory, |
| 703 | * because the page was modified. we must not write the | 703 | * because the page was modified. we must not write the |
| 704 | * modified page to that sector. | 704 | * modified page to that sector. |
| 705 | * | 705 | * |
| 706 | * TODO: what could be done here: wait for the delalloc | 706 | * TODO: what could be done here: wait for the delalloc |
| 707 | * runner to write out that page (might involve | 707 | * runner to write out that page (might involve |
| 708 | * COW) and see whether the sector is still | 708 | * COW) and see whether the sector is still |
| 709 | * referenced afterwards. | 709 | * referenced afterwards. |
| 710 | * | 710 | * |
| 711 | * For the meantime, we'll treat this error | 711 | * For the meantime, we'll treat this error |
| 712 | * incorrectable, although there is a chance that a | 712 | * incorrectable, although there is a chance that a |
| 713 | * later scrub will find the bad sector again and that | 713 | * later scrub will find the bad sector again and that |
| 714 | * there's no dirty page in memory, then. | 714 | * there's no dirty page in memory, then. |
| 715 | */ | 715 | */ |
| 716 | ret = -EIO; | 716 | ret = -EIO; |
| 717 | goto out; | 717 | goto out; |
| 718 | } | 718 | } |
| 719 | ret = repair_io_failure(inode, offset, PAGE_SIZE, | 719 | ret = repair_io_failure(inode, offset, PAGE_SIZE, |
| 720 | fixup->logical, page, | 720 | fixup->logical, page, |
| 721 | offset - page_offset(page), | 721 | offset - page_offset(page), |
| 722 | fixup->mirror_num); | 722 | fixup->mirror_num); |
| 723 | unlock_page(page); | 723 | unlock_page(page); |
| 724 | corrected = !ret; | 724 | corrected = !ret; |
| 725 | } else { | 725 | } else { |
| 726 | /* | 726 | /* |
| 727 | * we need to get good data first. the general readpage path | 727 | * we need to get good data first. the general readpage path |
| 728 | * will call repair_io_failure for us, we just have to make | 728 | * will call repair_io_failure for us, we just have to make |
| 729 | * sure we read the bad mirror. | 729 | * sure we read the bad mirror. |
| 730 | */ | 730 | */ |
| 731 | ret = set_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, | 731 | ret = set_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, |
| 732 | EXTENT_DAMAGED, GFP_NOFS); | 732 | EXTENT_DAMAGED, GFP_NOFS); |
| 733 | if (ret) { | 733 | if (ret) { |
| 734 | /* set_extent_bits should give proper error */ | 734 | /* set_extent_bits should give proper error */ |
| 735 | WARN_ON(ret > 0); | 735 | WARN_ON(ret > 0); |
| 736 | if (ret > 0) | 736 | if (ret > 0) |
| 737 | ret = -EFAULT; | 737 | ret = -EFAULT; |
| 738 | goto out; | 738 | goto out; |
| 739 | } | 739 | } |
| 740 | 740 | ||
| 741 | ret = extent_read_full_page(&BTRFS_I(inode)->io_tree, page, | 741 | ret = extent_read_full_page(&BTRFS_I(inode)->io_tree, page, |
| 742 | btrfs_get_extent, | 742 | btrfs_get_extent, |
| 743 | fixup->mirror_num); | 743 | fixup->mirror_num); |
| 744 | wait_on_page_locked(page); | 744 | wait_on_page_locked(page); |
| 745 | 745 | ||
| 746 | corrected = !test_range_bit(&BTRFS_I(inode)->io_tree, offset, | 746 | corrected = !test_range_bit(&BTRFS_I(inode)->io_tree, offset, |
| 747 | end, EXTENT_DAMAGED, 0, NULL); | 747 | end, EXTENT_DAMAGED, 0, NULL); |
| 748 | if (!corrected) | 748 | if (!corrected) |
| 749 | clear_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, | 749 | clear_extent_bits(&BTRFS_I(inode)->io_tree, offset, end, |
| 750 | EXTENT_DAMAGED, GFP_NOFS); | 750 | EXTENT_DAMAGED, GFP_NOFS); |
| 751 | } | 751 | } |
| 752 | 752 | ||
| 753 | out: | 753 | out: |
| 754 | if (page) | 754 | if (page) |
| 755 | put_page(page); | 755 | put_page(page); |
| 756 | 756 | ||
| 757 | iput(inode); | 757 | iput(inode); |
| 758 | 758 | ||
| 759 | if (ret < 0) | 759 | if (ret < 0) |
| 760 | return ret; | 760 | return ret; |
| 761 | 761 | ||
| 762 | if (ret == 0 && corrected) { | 762 | if (ret == 0 && corrected) { |
| 763 | /* | 763 | /* |
| 764 | * we only need to call readpage for one of the inodes belonging | 764 | * we only need to call readpage for one of the inodes belonging |
| 765 | * to this extent. so make iterate_extent_inodes stop | 765 | * to this extent. so make iterate_extent_inodes stop |
| 766 | */ | 766 | */ |
| 767 | return 1; | 767 | return 1; |
| 768 | } | 768 | } |
| 769 | 769 | ||
| 770 | return -EIO; | 770 | return -EIO; |
| 771 | } | 771 | } |
| 772 | 772 | ||
| 773 | static void scrub_fixup_nodatasum(struct btrfs_work *work) | 773 | static void scrub_fixup_nodatasum(struct btrfs_work *work) |
| 774 | { | 774 | { |
| 775 | int ret; | 775 | int ret; |
| 776 | struct scrub_fixup_nodatasum *fixup; | 776 | struct scrub_fixup_nodatasum *fixup; |
| 777 | struct scrub_ctx *sctx; | 777 | struct scrub_ctx *sctx; |
| 778 | struct btrfs_trans_handle *trans = NULL; | 778 | struct btrfs_trans_handle *trans = NULL; |
| 779 | struct btrfs_path *path; | 779 | struct btrfs_path *path; |
| 780 | int uncorrectable = 0; | 780 | int uncorrectable = 0; |
| 781 | 781 | ||
| 782 | fixup = container_of(work, struct scrub_fixup_nodatasum, work); | 782 | fixup = container_of(work, struct scrub_fixup_nodatasum, work); |
| 783 | sctx = fixup->sctx; | 783 | sctx = fixup->sctx; |
| 784 | 784 | ||
| 785 | path = btrfs_alloc_path(); | 785 | path = btrfs_alloc_path(); |
| 786 | if (!path) { | 786 | if (!path) { |
| 787 | spin_lock(&sctx->stat_lock); | 787 | spin_lock(&sctx->stat_lock); |
| 788 | ++sctx->stat.malloc_errors; | 788 | ++sctx->stat.malloc_errors; |
| 789 | spin_unlock(&sctx->stat_lock); | 789 | spin_unlock(&sctx->stat_lock); |
| 790 | uncorrectable = 1; | 790 | uncorrectable = 1; |
| 791 | goto out; | 791 | goto out; |
| 792 | } | 792 | } |
| 793 | 793 | ||
| 794 | trans = btrfs_join_transaction(fixup->root); | 794 | trans = btrfs_join_transaction(fixup->root); |
| 795 | if (IS_ERR(trans)) { | 795 | if (IS_ERR(trans)) { |
| 796 | uncorrectable = 1; | 796 | uncorrectable = 1; |
| 797 | goto out; | 797 | goto out; |
| 798 | } | 798 | } |
| 799 | 799 | ||
| 800 | /* | 800 | /* |
| 801 | * the idea is to trigger a regular read through the standard path. we | 801 | * the idea is to trigger a regular read through the standard path. we |
| 802 | * read a page from the (failed) logical address by specifying the | 802 | * read a page from the (failed) logical address by specifying the |
| 803 | * corresponding copynum of the failed sector. thus, that readpage is | 803 | * corresponding copynum of the failed sector. thus, that readpage is |
| 804 | * expected to fail. | 804 | * expected to fail. |
| 805 | * that is the point where on-the-fly error correction will kick in | 805 | * that is the point where on-the-fly error correction will kick in |
| 806 | * (once it's finished) and rewrite the failed sector if a good copy | 806 | * (once it's finished) and rewrite the failed sector if a good copy |
| 807 | * can be found. | 807 | * can be found. |
| 808 | */ | 808 | */ |
| 809 | ret = iterate_inodes_from_logical(fixup->logical, fixup->root->fs_info, | 809 | ret = iterate_inodes_from_logical(fixup->logical, fixup->root->fs_info, |
| 810 | path, scrub_fixup_readpage, | 810 | path, scrub_fixup_readpage, |
| 811 | fixup); | 811 | fixup); |
| 812 | if (ret < 0) { | 812 | if (ret < 0) { |
| 813 | uncorrectable = 1; | 813 | uncorrectable = 1; |
| 814 | goto out; | 814 | goto out; |
| 815 | } | 815 | } |
| 816 | WARN_ON(ret != 1); | 816 | WARN_ON(ret != 1); |
| 817 | 817 | ||
| 818 | spin_lock(&sctx->stat_lock); | 818 | spin_lock(&sctx->stat_lock); |
| 819 | ++sctx->stat.corrected_errors; | 819 | ++sctx->stat.corrected_errors; |
| 820 | spin_unlock(&sctx->stat_lock); | 820 | spin_unlock(&sctx->stat_lock); |
| 821 | 821 | ||
| 822 | out: | 822 | out: |
| 823 | if (trans && !IS_ERR(trans)) | 823 | if (trans && !IS_ERR(trans)) |
| 824 | btrfs_end_transaction(trans, fixup->root); | 824 | btrfs_end_transaction(trans, fixup->root); |
| 825 | if (uncorrectable) { | 825 | if (uncorrectable) { |
| 826 | spin_lock(&sctx->stat_lock); | 826 | spin_lock(&sctx->stat_lock); |
| 827 | ++sctx->stat.uncorrectable_errors; | 827 | ++sctx->stat.uncorrectable_errors; |
| 828 | spin_unlock(&sctx->stat_lock); | 828 | spin_unlock(&sctx->stat_lock); |
| 829 | btrfs_dev_replace_stats_inc( | 829 | btrfs_dev_replace_stats_inc( |
| 830 | &sctx->dev_root->fs_info->dev_replace. | 830 | &sctx->dev_root->fs_info->dev_replace. |
| 831 | num_uncorrectable_read_errors); | 831 | num_uncorrectable_read_errors); |
| 832 | printk_ratelimited_in_rcu(KERN_ERR "BTRFS: " | 832 | printk_ratelimited_in_rcu(KERN_ERR "BTRFS: " |
| 833 | "unable to fixup (nodatasum) error at logical %llu on dev %s\n", | 833 | "unable to fixup (nodatasum) error at logical %llu on dev %s\n", |
| 834 | fixup->logical, rcu_str_deref(fixup->dev->name)); | 834 | fixup->logical, rcu_str_deref(fixup->dev->name)); |
| 835 | } | 835 | } |
| 836 | 836 | ||
| 837 | btrfs_free_path(path); | 837 | btrfs_free_path(path); |
| 838 | kfree(fixup); | 838 | kfree(fixup); |
| 839 | 839 | ||
| 840 | scrub_pending_trans_workers_dec(sctx); | 840 | scrub_pending_trans_workers_dec(sctx); |
| 841 | } | 841 | } |
| 842 | 842 | ||
| 843 | static inline void scrub_get_recover(struct scrub_recover *recover) | 843 | static inline void scrub_get_recover(struct scrub_recover *recover) |
| 844 | { | 844 | { |
| 845 | atomic_inc(&recover->refs); | 845 | atomic_inc(&recover->refs); |
| 846 | } | 846 | } |
| 847 | 847 | ||
| 848 | static inline void scrub_put_recover(struct scrub_recover *recover) | 848 | static inline void scrub_put_recover(struct scrub_recover *recover) |
| 849 | { | 849 | { |
| 850 | if (atomic_dec_and_test(&recover->refs)) { | 850 | if (atomic_dec_and_test(&recover->refs)) { |
| 851 | kfree(recover->bbio); | 851 | kfree(recover->bbio); |
| 852 | kfree(recover->raid_map); | 852 | kfree(recover->raid_map); |
| 853 | kfree(recover); | 853 | kfree(recover); |
| 854 | } | 854 | } |
| 855 | } | 855 | } |
| 856 | 856 | ||
| 857 | /* | 857 | /* |
| 858 | * scrub_handle_errored_block gets called when either verification of the | 858 | * scrub_handle_errored_block gets called when either verification of the |
| 859 | * pages failed or the bio failed to read, e.g. with EIO. In the latter | 859 | * pages failed or the bio failed to read, e.g. with EIO. In the latter |
| 860 | * case, this function handles all pages in the bio, even though only one | 860 | * case, this function handles all pages in the bio, even though only one |
| 861 | * may be bad. | 861 | * may be bad. |
| 862 | * The goal of this function is to repair the errored block by using the | 862 | * The goal of this function is to repair the errored block by using the |
| 863 | * contents of one of the mirrors. | 863 | * contents of one of the mirrors. |
| 864 | */ | 864 | */ |
| 865 | static int scrub_handle_errored_block(struct scrub_block *sblock_to_check) | 865 | static int scrub_handle_errored_block(struct scrub_block *sblock_to_check) |
| 866 | { | 866 | { |
| 867 | struct scrub_ctx *sctx = sblock_to_check->sctx; | 867 | struct scrub_ctx *sctx = sblock_to_check->sctx; |
| 868 | struct btrfs_device *dev; | 868 | struct btrfs_device *dev; |
| 869 | struct btrfs_fs_info *fs_info; | 869 | struct btrfs_fs_info *fs_info; |
| 870 | u64 length; | 870 | u64 length; |
| 871 | u64 logical; | 871 | u64 logical; |
| 872 | u64 generation; | 872 | u64 generation; |
| 873 | unsigned int failed_mirror_index; | 873 | unsigned int failed_mirror_index; |
| 874 | unsigned int is_metadata; | 874 | unsigned int is_metadata; |
| 875 | unsigned int have_csum; | 875 | unsigned int have_csum; |
| 876 | u8 *csum; | 876 | u8 *csum; |
| 877 | struct scrub_block *sblocks_for_recheck; /* holds one for each mirror */ | 877 | struct scrub_block *sblocks_for_recheck; /* holds one for each mirror */ |
| 878 | struct scrub_block *sblock_bad; | 878 | struct scrub_block *sblock_bad; |
| 879 | int ret; | 879 | int ret; |
| 880 | int mirror_index; | 880 | int mirror_index; |
| 881 | int page_num; | 881 | int page_num; |
| 882 | int success; | 882 | int success; |
| 883 | static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, | 883 | static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, |
| 884 | DEFAULT_RATELIMIT_BURST); | 884 | DEFAULT_RATELIMIT_BURST); |
| 885 | 885 | ||
| 886 | BUG_ON(sblock_to_check->page_count < 1); | 886 | BUG_ON(sblock_to_check->page_count < 1); |
| 887 | fs_info = sctx->dev_root->fs_info; | 887 | fs_info = sctx->dev_root->fs_info; |
| 888 | if (sblock_to_check->pagev[0]->flags & BTRFS_EXTENT_FLAG_SUPER) { | 888 | if (sblock_to_check->pagev[0]->flags & BTRFS_EXTENT_FLAG_SUPER) { |
| 889 | /* | 889 | /* |
| 890 | * if we find an error in a super block, we just report it. | 890 | * if we find an error in a super block, we just report it. |
| 891 | * They will get written with the next transaction commit | 891 | * They will get written with the next transaction commit |
| 892 | * anyway | 892 | * anyway |
| 893 | */ | 893 | */ |
| 894 | spin_lock(&sctx->stat_lock); | 894 | spin_lock(&sctx->stat_lock); |
| 895 | ++sctx->stat.super_errors; | 895 | ++sctx->stat.super_errors; |
| 896 | spin_unlock(&sctx->stat_lock); | 896 | spin_unlock(&sctx->stat_lock); |
| 897 | return 0; | 897 | return 0; |
| 898 | } | 898 | } |
| 899 | length = sblock_to_check->page_count * PAGE_SIZE; | 899 | length = sblock_to_check->page_count * PAGE_SIZE; |
| 900 | logical = sblock_to_check->pagev[0]->logical; | 900 | logical = sblock_to_check->pagev[0]->logical; |
| 901 | generation = sblock_to_check->pagev[0]->generation; | 901 | generation = sblock_to_check->pagev[0]->generation; |
| 902 | BUG_ON(sblock_to_check->pagev[0]->mirror_num < 1); | 902 | BUG_ON(sblock_to_check->pagev[0]->mirror_num < 1); |
| 903 | failed_mirror_index = sblock_to_check->pagev[0]->mirror_num - 1; | 903 | failed_mirror_index = sblock_to_check->pagev[0]->mirror_num - 1; |
| 904 | is_metadata = !(sblock_to_check->pagev[0]->flags & | 904 | is_metadata = !(sblock_to_check->pagev[0]->flags & |
| 905 | BTRFS_EXTENT_FLAG_DATA); | 905 | BTRFS_EXTENT_FLAG_DATA); |
| 906 | have_csum = sblock_to_check->pagev[0]->have_csum; | 906 | have_csum = sblock_to_check->pagev[0]->have_csum; |
| 907 | csum = sblock_to_check->pagev[0]->csum; | 907 | csum = sblock_to_check->pagev[0]->csum; |
| 908 | dev = sblock_to_check->pagev[0]->dev; | 908 | dev = sblock_to_check->pagev[0]->dev; |
| 909 | 909 | ||
| 910 | if (sctx->is_dev_replace && !is_metadata && !have_csum) { | 910 | if (sctx->is_dev_replace && !is_metadata && !have_csum) { |
| 911 | sblocks_for_recheck = NULL; | 911 | sblocks_for_recheck = NULL; |
| 912 | goto nodatasum_case; | 912 | goto nodatasum_case; |
| 913 | } | 913 | } |
| 914 | 914 | ||
| 915 | /* | 915 | /* |
| 916 | * read all mirrors one after the other. This includes to | 916 | * read all mirrors one after the other. This includes to |
| 917 | * re-read the extent or metadata block that failed (that was | 917 | * re-read the extent or metadata block that failed (that was |
| 918 | * the cause that this fixup code is called) another time, | 918 | * the cause that this fixup code is called) another time, |
| 919 | * page by page this time in order to know which pages | 919 | * page by page this time in order to know which pages |
| 920 | * caused I/O errors and which ones are good (for all mirrors). | 920 | * caused I/O errors and which ones are good (for all mirrors). |
| 921 | * It is the goal to handle the situation when more than one | 921 | * It is the goal to handle the situation when more than one |
| 922 | * mirror contains I/O errors, but the errors do not | 922 | * mirror contains I/O errors, but the errors do not |
| 923 | * overlap, i.e. the data can be repaired by selecting the | 923 | * overlap, i.e. the data can be repaired by selecting the |
| 924 | * pages from those mirrors without I/O error on the | 924 | * pages from those mirrors without I/O error on the |
| 925 | * particular pages. One example (with blocks >= 2 * PAGE_SIZE) | 925 | * particular pages. One example (with blocks >= 2 * PAGE_SIZE) |
| 926 | * would be that mirror #1 has an I/O error on the first page, | 926 | * would be that mirror #1 has an I/O error on the first page, |
| 927 | * the second page is good, and mirror #2 has an I/O error on | 927 | * the second page is good, and mirror #2 has an I/O error on |
| 928 | * the second page, but the first page is good. | 928 | * the second page, but the first page is good. |
| 929 | * Then the first page of the first mirror can be repaired by | 929 | * Then the first page of the first mirror can be repaired by |
| 930 | * taking the first page of the second mirror, and the | 930 | * taking the first page of the second mirror, and the |
| 931 | * second page of the second mirror can be repaired by | 931 | * second page of the second mirror can be repaired by |
| 932 | * copying the contents of the 2nd page of the 1st mirror. | 932 | * copying the contents of the 2nd page of the 1st mirror. |
| 933 | * One more note: if the pages of one mirror contain I/O | 933 | * One more note: if the pages of one mirror contain I/O |
| 934 | * errors, the checksum cannot be verified. In order to get | 934 | * errors, the checksum cannot be verified. In order to get |
| 935 | * the best data for repairing, the first attempt is to find | 935 | * the best data for repairing, the first attempt is to find |
| 936 | * a mirror without I/O errors and with a validated checksum. | 936 | * a mirror without I/O errors and with a validated checksum. |
| 937 | * Only if this is not possible, the pages are picked from | 937 | * Only if this is not possible, the pages are picked from |
| 938 | * mirrors with I/O errors without considering the checksum. | 938 | * mirrors with I/O errors without considering the checksum. |
| 939 | * If the latter is the case, at the end, the checksum of the | 939 | * If the latter is the case, at the end, the checksum of the |
| 940 | * repaired area is verified in order to correctly maintain | 940 | * repaired area is verified in order to correctly maintain |
| 941 | * the statistics. | 941 | * the statistics. |
| 942 | */ | 942 | */ |
| 943 | 943 | ||
| 944 | sblocks_for_recheck = kzalloc(BTRFS_MAX_MIRRORS * | 944 | sblocks_for_recheck = kzalloc(BTRFS_MAX_MIRRORS * |
| 945 | sizeof(*sblocks_for_recheck), | 945 | sizeof(*sblocks_for_recheck), |
| 946 | GFP_NOFS); | 946 | GFP_NOFS); |
| 947 | if (!sblocks_for_recheck) { | 947 | if (!sblocks_for_recheck) { |
| 948 | spin_lock(&sctx->stat_lock); | 948 | spin_lock(&sctx->stat_lock); |
| 949 | sctx->stat.malloc_errors++; | 949 | sctx->stat.malloc_errors++; |
| 950 | sctx->stat.read_errors++; | 950 | sctx->stat.read_errors++; |
| 951 | sctx->stat.uncorrectable_errors++; | 951 | sctx->stat.uncorrectable_errors++; |
| 952 | spin_unlock(&sctx->stat_lock); | 952 | spin_unlock(&sctx->stat_lock); |
| 953 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_READ_ERRS); | 953 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_READ_ERRS); |
| 954 | goto out; | 954 | goto out; |
| 955 | } | 955 | } |
| 956 | 956 | ||
| 957 | /* setup the context, map the logical blocks and alloc the pages */ | 957 | /* setup the context, map the logical blocks and alloc the pages */ |
| 958 | ret = scrub_setup_recheck_block(sctx, fs_info, sblock_to_check, length, | 958 | ret = scrub_setup_recheck_block(sctx, fs_info, sblock_to_check, length, |
| 959 | logical, sblocks_for_recheck); | 959 | logical, sblocks_for_recheck); |
| 960 | if (ret) { | 960 | if (ret) { |
| 961 | spin_lock(&sctx->stat_lock); | 961 | spin_lock(&sctx->stat_lock); |
| 962 | sctx->stat.read_errors++; | 962 | sctx->stat.read_errors++; |
| 963 | sctx->stat.uncorrectable_errors++; | 963 | sctx->stat.uncorrectable_errors++; |
| 964 | spin_unlock(&sctx->stat_lock); | 964 | spin_unlock(&sctx->stat_lock); |
| 965 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_READ_ERRS); | 965 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_READ_ERRS); |
| 966 | goto out; | 966 | goto out; |
| 967 | } | 967 | } |
| 968 | BUG_ON(failed_mirror_index >= BTRFS_MAX_MIRRORS); | 968 | BUG_ON(failed_mirror_index >= BTRFS_MAX_MIRRORS); |
| 969 | sblock_bad = sblocks_for_recheck + failed_mirror_index; | 969 | sblock_bad = sblocks_for_recheck + failed_mirror_index; |
| 970 | 970 | ||
| 971 | /* build and submit the bios for the failed mirror, check checksums */ | 971 | /* build and submit the bios for the failed mirror, check checksums */ |
| 972 | scrub_recheck_block(fs_info, sblock_bad, is_metadata, have_csum, | 972 | scrub_recheck_block(fs_info, sblock_bad, is_metadata, have_csum, |
| 973 | csum, generation, sctx->csum_size, 1); | 973 | csum, generation, sctx->csum_size, 1); |
| 974 | 974 | ||
| 975 | if (!sblock_bad->header_error && !sblock_bad->checksum_error && | 975 | if (!sblock_bad->header_error && !sblock_bad->checksum_error && |
| 976 | sblock_bad->no_io_error_seen) { | 976 | sblock_bad->no_io_error_seen) { |
| 977 | /* | 977 | /* |
| 978 | * the error disappeared after reading page by page, or | 978 | * the error disappeared after reading page by page, or |
| 979 | * the area was part of a huge bio and other parts of the | 979 | * the area was part of a huge bio and other parts of the |
| 980 | * bio caused I/O errors, or the block layer merged several | 980 | * bio caused I/O errors, or the block layer merged several |
| 981 | * read requests into one and the error is caused by a | 981 | * read requests into one and the error is caused by a |
| 982 | * different bio (usually one of the two latter cases is | 982 | * different bio (usually one of the two latter cases is |
| 983 | * the cause) | 983 | * the cause) |
| 984 | */ | 984 | */ |
| 985 | spin_lock(&sctx->stat_lock); | 985 | spin_lock(&sctx->stat_lock); |
| 986 | sctx->stat.unverified_errors++; | 986 | sctx->stat.unverified_errors++; |
| 987 | sblock_to_check->data_corrected = 1; | 987 | sblock_to_check->data_corrected = 1; |
| 988 | spin_unlock(&sctx->stat_lock); | 988 | spin_unlock(&sctx->stat_lock); |
| 989 | 989 | ||
| 990 | if (sctx->is_dev_replace) | 990 | if (sctx->is_dev_replace) |
| 991 | scrub_write_block_to_dev_replace(sblock_bad); | 991 | scrub_write_block_to_dev_replace(sblock_bad); |
| 992 | goto out; | 992 | goto out; |
| 993 | } | 993 | } |
| 994 | 994 | ||
| 995 | if (!sblock_bad->no_io_error_seen) { | 995 | if (!sblock_bad->no_io_error_seen) { |
| 996 | spin_lock(&sctx->stat_lock); | 996 | spin_lock(&sctx->stat_lock); |
| 997 | sctx->stat.read_errors++; | 997 | sctx->stat.read_errors++; |
| 998 | spin_unlock(&sctx->stat_lock); | 998 | spin_unlock(&sctx->stat_lock); |
| 999 | if (__ratelimit(&_rs)) | 999 | if (__ratelimit(&_rs)) |
| 1000 | scrub_print_warning("i/o error", sblock_to_check); | 1000 | scrub_print_warning("i/o error", sblock_to_check); |
| 1001 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_READ_ERRS); | 1001 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_READ_ERRS); |
| 1002 | } else if (sblock_bad->checksum_error) { | 1002 | } else if (sblock_bad->checksum_error) { |
| 1003 | spin_lock(&sctx->stat_lock); | 1003 | spin_lock(&sctx->stat_lock); |
| 1004 | sctx->stat.csum_errors++; | 1004 | sctx->stat.csum_errors++; |
| 1005 | spin_unlock(&sctx->stat_lock); | 1005 | spin_unlock(&sctx->stat_lock); |
| 1006 | if (__ratelimit(&_rs)) | 1006 | if (__ratelimit(&_rs)) |
| 1007 | scrub_print_warning("checksum error", sblock_to_check); | 1007 | scrub_print_warning("checksum error", sblock_to_check); |
| 1008 | btrfs_dev_stat_inc_and_print(dev, | 1008 | btrfs_dev_stat_inc_and_print(dev, |
| 1009 | BTRFS_DEV_STAT_CORRUPTION_ERRS); | 1009 | BTRFS_DEV_STAT_CORRUPTION_ERRS); |
| 1010 | } else if (sblock_bad->header_error) { | 1010 | } else if (sblock_bad->header_error) { |
| 1011 | spin_lock(&sctx->stat_lock); | 1011 | spin_lock(&sctx->stat_lock); |
| 1012 | sctx->stat.verify_errors++; | 1012 | sctx->stat.verify_errors++; |
| 1013 | spin_unlock(&sctx->stat_lock); | 1013 | spin_unlock(&sctx->stat_lock); |
| 1014 | if (__ratelimit(&_rs)) | 1014 | if (__ratelimit(&_rs)) |
| 1015 | scrub_print_warning("checksum/header error", | 1015 | scrub_print_warning("checksum/header error", |
| 1016 | sblock_to_check); | 1016 | sblock_to_check); |
| 1017 | if (sblock_bad->generation_error) | 1017 | if (sblock_bad->generation_error) |
| 1018 | btrfs_dev_stat_inc_and_print(dev, | 1018 | btrfs_dev_stat_inc_and_print(dev, |
| 1019 | BTRFS_DEV_STAT_GENERATION_ERRS); | 1019 | BTRFS_DEV_STAT_GENERATION_ERRS); |
| 1020 | else | 1020 | else |
| 1021 | btrfs_dev_stat_inc_and_print(dev, | 1021 | btrfs_dev_stat_inc_and_print(dev, |
| 1022 | BTRFS_DEV_STAT_CORRUPTION_ERRS); | 1022 | BTRFS_DEV_STAT_CORRUPTION_ERRS); |
| 1023 | } | 1023 | } |
| 1024 | 1024 | ||
| 1025 | if (sctx->readonly) { | 1025 | if (sctx->readonly) { |
| 1026 | ASSERT(!sctx->is_dev_replace); | 1026 | ASSERT(!sctx->is_dev_replace); |
| 1027 | goto out; | 1027 | goto out; |
| 1028 | } | 1028 | } |
| 1029 | 1029 | ||
| 1030 | if (!is_metadata && !have_csum) { | 1030 | if (!is_metadata && !have_csum) { |
| 1031 | struct scrub_fixup_nodatasum *fixup_nodatasum; | 1031 | struct scrub_fixup_nodatasum *fixup_nodatasum; |
| 1032 | 1032 | ||
| 1033 | nodatasum_case: | 1033 | nodatasum_case: |
| 1034 | WARN_ON(sctx->is_dev_replace); | 1034 | WARN_ON(sctx->is_dev_replace); |
| 1035 | 1035 | ||
| 1036 | /* | 1036 | /* |
| 1037 | * !is_metadata and !have_csum, this means that the data | 1037 | * !is_metadata and !have_csum, this means that the data |
| 1038 | * might not be COW'ed, that it might be modified | 1038 | * might not be COW'ed, that it might be modified |
| 1039 | * concurrently. The general strategy to work on the | 1039 | * concurrently. The general strategy to work on the |
| 1040 | * commit root does not help in the case when COW is not | 1040 | * commit root does not help in the case when COW is not |
| 1041 | * used. | 1041 | * used. |
| 1042 | */ | 1042 | */ |
| 1043 | fixup_nodatasum = kzalloc(sizeof(*fixup_nodatasum), GFP_NOFS); | 1043 | fixup_nodatasum = kzalloc(sizeof(*fixup_nodatasum), GFP_NOFS); |
| 1044 | if (!fixup_nodatasum) | 1044 | if (!fixup_nodatasum) |
| 1045 | goto did_not_correct_error; | 1045 | goto did_not_correct_error; |
| 1046 | fixup_nodatasum->sctx = sctx; | 1046 | fixup_nodatasum->sctx = sctx; |
| 1047 | fixup_nodatasum->dev = dev; | 1047 | fixup_nodatasum->dev = dev; |
| 1048 | fixup_nodatasum->logical = logical; | 1048 | fixup_nodatasum->logical = logical; |
| 1049 | fixup_nodatasum->root = fs_info->extent_root; | 1049 | fixup_nodatasum->root = fs_info->extent_root; |
| 1050 | fixup_nodatasum->mirror_num = failed_mirror_index + 1; | 1050 | fixup_nodatasum->mirror_num = failed_mirror_index + 1; |
| 1051 | scrub_pending_trans_workers_inc(sctx); | 1051 | scrub_pending_trans_workers_inc(sctx); |
| 1052 | btrfs_init_work(&fixup_nodatasum->work, btrfs_scrub_helper, | 1052 | btrfs_init_work(&fixup_nodatasum->work, btrfs_scrub_helper, |
| 1053 | scrub_fixup_nodatasum, NULL, NULL); | 1053 | scrub_fixup_nodatasum, NULL, NULL); |
| 1054 | btrfs_queue_work(fs_info->scrub_workers, | 1054 | btrfs_queue_work(fs_info->scrub_workers, |
| 1055 | &fixup_nodatasum->work); | 1055 | &fixup_nodatasum->work); |
| 1056 | goto out; | 1056 | goto out; |
| 1057 | } | 1057 | } |
| 1058 | 1058 | ||
| 1059 | /* | 1059 | /* |
| 1060 | * now build and submit the bios for the other mirrors, check | 1060 | * now build and submit the bios for the other mirrors, check |
| 1061 | * checksums. | 1061 | * checksums. |
| 1062 | * First try to pick the mirror which is completely without I/O | 1062 | * First try to pick the mirror which is completely without I/O |
| 1063 | * errors and also does not have a checksum error. | 1063 | * errors and also does not have a checksum error. |
| 1064 | * If one is found, and if a checksum is present, the full block | 1064 | * If one is found, and if a checksum is present, the full block |
| 1065 | * that is known to contain an error is rewritten. Afterwards | 1065 | * that is known to contain an error is rewritten. Afterwards |
| 1066 | * the block is known to be corrected. | 1066 | * the block is known to be corrected. |
| 1067 | * If a mirror is found which is completely correct, and no | 1067 | * If a mirror is found which is completely correct, and no |
| 1068 | * checksum is present, only those pages are rewritten that had | 1068 | * checksum is present, only those pages are rewritten that had |
| 1069 | * an I/O error in the block to be repaired, since it cannot be | 1069 | * an I/O error in the block to be repaired, since it cannot be |
| 1070 | * determined, which copy of the other pages is better (and it | 1070 | * determined, which copy of the other pages is better (and it |
| 1071 | * could happen otherwise that a correct page would be | 1071 | * could happen otherwise that a correct page would be |
| 1072 | * overwritten by a bad one). | 1072 | * overwritten by a bad one). |
| 1073 | */ | 1073 | */ |
| 1074 | for (mirror_index = 0; | 1074 | for (mirror_index = 0; |
| 1075 | mirror_index < BTRFS_MAX_MIRRORS && | 1075 | mirror_index < BTRFS_MAX_MIRRORS && |
| 1076 | sblocks_for_recheck[mirror_index].page_count > 0; | 1076 | sblocks_for_recheck[mirror_index].page_count > 0; |
| 1077 | mirror_index++) { | 1077 | mirror_index++) { |
| 1078 | struct scrub_block *sblock_other; | 1078 | struct scrub_block *sblock_other; |
| 1079 | 1079 | ||
| 1080 | if (mirror_index == failed_mirror_index) | 1080 | if (mirror_index == failed_mirror_index) |
| 1081 | continue; | 1081 | continue; |
| 1082 | sblock_other = sblocks_for_recheck + mirror_index; | 1082 | sblock_other = sblocks_for_recheck + mirror_index; |
| 1083 | 1083 | ||
| 1084 | /* build and submit the bios, check checksums */ | 1084 | /* build and submit the bios, check checksums */ |
| 1085 | scrub_recheck_block(fs_info, sblock_other, is_metadata, | 1085 | scrub_recheck_block(fs_info, sblock_other, is_metadata, |
| 1086 | have_csum, csum, generation, | 1086 | have_csum, csum, generation, |
| 1087 | sctx->csum_size, 0); | 1087 | sctx->csum_size, 0); |
| 1088 | 1088 | ||
| 1089 | if (!sblock_other->header_error && | 1089 | if (!sblock_other->header_error && |
| 1090 | !sblock_other->checksum_error && | 1090 | !sblock_other->checksum_error && |
| 1091 | sblock_other->no_io_error_seen) { | 1091 | sblock_other->no_io_error_seen) { |
| 1092 | if (sctx->is_dev_replace) { | 1092 | if (sctx->is_dev_replace) { |
| 1093 | scrub_write_block_to_dev_replace(sblock_other); | 1093 | scrub_write_block_to_dev_replace(sblock_other); |
| 1094 | } else { | 1094 | } else { |
| 1095 | int force_write = is_metadata || have_csum; | 1095 | int force_write = is_metadata || have_csum; |
| 1096 | 1096 | ||
| 1097 | ret = scrub_repair_block_from_good_copy( | 1097 | ret = scrub_repair_block_from_good_copy( |
| 1098 | sblock_bad, sblock_other, | 1098 | sblock_bad, sblock_other, |
| 1099 | force_write); | 1099 | force_write); |
| 1100 | } | 1100 | } |
| 1101 | if (0 == ret) | 1101 | if (0 == ret) |
| 1102 | goto corrected_error; | 1102 | goto corrected_error; |
| 1103 | } | 1103 | } |
| 1104 | } | 1104 | } |
| 1105 | 1105 | ||
| 1106 | /* | 1106 | /* |
| 1107 | * for dev_replace, pick good pages and write to the target device. | 1107 | * for dev_replace, pick good pages and write to the target device. |
| 1108 | */ | 1108 | */ |
| 1109 | if (sctx->is_dev_replace) { | 1109 | if (sctx->is_dev_replace) { |
| 1110 | success = 1; | 1110 | success = 1; |
| 1111 | for (page_num = 0; page_num < sblock_bad->page_count; | 1111 | for (page_num = 0; page_num < sblock_bad->page_count; |
| 1112 | page_num++) { | 1112 | page_num++) { |
| 1113 | int sub_success; | 1113 | int sub_success; |
| 1114 | 1114 | ||
| 1115 | sub_success = 0; | 1115 | sub_success = 0; |
| 1116 | for (mirror_index = 0; | 1116 | for (mirror_index = 0; |
| 1117 | mirror_index < BTRFS_MAX_MIRRORS && | 1117 | mirror_index < BTRFS_MAX_MIRRORS && |
| 1118 | sblocks_for_recheck[mirror_index].page_count > 0; | 1118 | sblocks_for_recheck[mirror_index].page_count > 0; |
| 1119 | mirror_index++) { | 1119 | mirror_index++) { |
| 1120 | struct scrub_block *sblock_other = | 1120 | struct scrub_block *sblock_other = |
| 1121 | sblocks_for_recheck + mirror_index; | 1121 | sblocks_for_recheck + mirror_index; |
| 1122 | struct scrub_page *page_other = | 1122 | struct scrub_page *page_other = |
| 1123 | sblock_other->pagev[page_num]; | 1123 | sblock_other->pagev[page_num]; |
| 1124 | 1124 | ||
| 1125 | if (!page_other->io_error) { | 1125 | if (!page_other->io_error) { |
| 1126 | ret = scrub_write_page_to_dev_replace( | 1126 | ret = scrub_write_page_to_dev_replace( |
| 1127 | sblock_other, page_num); | 1127 | sblock_other, page_num); |
| 1128 | if (ret == 0) { | 1128 | if (ret == 0) { |
| 1129 | /* succeeded for this page */ | 1129 | /* succeeded for this page */ |
| 1130 | sub_success = 1; | 1130 | sub_success = 1; |
| 1131 | break; | 1131 | break; |
| 1132 | } else { | 1132 | } else { |
| 1133 | btrfs_dev_replace_stats_inc( | 1133 | btrfs_dev_replace_stats_inc( |
| 1134 | &sctx->dev_root-> | 1134 | &sctx->dev_root-> |
| 1135 | fs_info->dev_replace. | 1135 | fs_info->dev_replace. |
| 1136 | num_write_errors); | 1136 | num_write_errors); |
| 1137 | } | 1137 | } |
| 1138 | } | 1138 | } |
| 1139 | } | 1139 | } |
| 1140 | 1140 | ||
| 1141 | if (!sub_success) { | 1141 | if (!sub_success) { |
| 1142 | /* | 1142 | /* |
| 1143 | * did not find a mirror to fetch the page | 1143 | * did not find a mirror to fetch the page |
| 1144 | * from. scrub_write_page_to_dev_replace() | 1144 | * from. scrub_write_page_to_dev_replace() |
| 1145 | * handles this case (page->io_error), by | 1145 | * handles this case (page->io_error), by |
| 1146 | * filling the block with zeros before | 1146 | * filling the block with zeros before |
| 1147 | * submitting the write request | 1147 | * submitting the write request |
| 1148 | */ | 1148 | */ |
| 1149 | success = 0; | 1149 | success = 0; |
| 1150 | ret = scrub_write_page_to_dev_replace( | 1150 | ret = scrub_write_page_to_dev_replace( |
| 1151 | sblock_bad, page_num); | 1151 | sblock_bad, page_num); |
| 1152 | if (ret) | 1152 | if (ret) |
| 1153 | btrfs_dev_replace_stats_inc( | 1153 | btrfs_dev_replace_stats_inc( |
| 1154 | &sctx->dev_root->fs_info-> | 1154 | &sctx->dev_root->fs_info-> |
| 1155 | dev_replace.num_write_errors); | 1155 | dev_replace.num_write_errors); |
| 1156 | } | 1156 | } |
| 1157 | } | 1157 | } |
| 1158 | 1158 | ||
| 1159 | goto out; | 1159 | goto out; |
| 1160 | } | 1160 | } |
| 1161 | 1161 | ||
| 1162 | /* | 1162 | /* |
| 1163 | * for regular scrub, repair those pages that are errored. | 1163 | * for regular scrub, repair those pages that are errored. |
| 1164 | * In case of I/O errors in the area that is supposed to be | 1164 | * In case of I/O errors in the area that is supposed to be |
| 1165 | * repaired, continue by picking good copies of those pages. | 1165 | * repaired, continue by picking good copies of those pages. |
| 1166 | * Select the good pages from mirrors to rewrite bad pages from | 1166 | * Select the good pages from mirrors to rewrite bad pages from |
| 1167 | * the area to fix. Afterwards verify the checksum of the block | 1167 | * the area to fix. Afterwards verify the checksum of the block |
| 1168 | * that is supposed to be repaired. This verification step is | 1168 | * that is supposed to be repaired. This verification step is |
| 1169 | * only done for the purpose of statistic counting and for the | 1169 | * only done for the purpose of statistic counting and for the |
| 1170 | * final scrub report, whether errors remain. | 1170 | * final scrub report, whether errors remain. |
| 1171 | * A perfect algorithm could make use of the checksum and try | 1171 | * A perfect algorithm could make use of the checksum and try |
| 1172 | * all possible combinations of pages from the different mirrors | 1172 | * all possible combinations of pages from the different mirrors |
| 1173 | * until the checksum verification succeeds. For example, when | 1173 | * until the checksum verification succeeds. For example, when |
| 1174 | * the 2nd page of mirror #1 faces I/O errors, and the 2nd page | 1174 | * the 2nd page of mirror #1 faces I/O errors, and the 2nd page |
| 1175 | * of mirror #2 is readable but the final checksum test fails, | 1175 | * of mirror #2 is readable but the final checksum test fails, |
| 1176 | * then the 2nd page of mirror #3 could be tried, whether now | 1176 | * then the 2nd page of mirror #3 could be tried, whether now |
| 1177 | * the final checksum succeedes. But this would be a rare | 1177 | * the final checksum succeedes. But this would be a rare |
| 1178 | * exception and is therefore not implemented. At least it is | 1178 | * exception and is therefore not implemented. At least it is |
| 1179 | * avoided that the good copy is overwritten. | 1179 | * avoided that the good copy is overwritten. |
| 1180 | * A more useful improvement would be to pick the sectors | 1180 | * A more useful improvement would be to pick the sectors |
| 1181 | * without I/O error based on sector sizes (512 bytes on legacy | 1181 | * without I/O error based on sector sizes (512 bytes on legacy |
| 1182 | * disks) instead of on PAGE_SIZE. Then maybe 512 byte of one | 1182 | * disks) instead of on PAGE_SIZE. Then maybe 512 byte of one |
| 1183 | * mirror could be repaired by taking 512 byte of a different | 1183 | * mirror could be repaired by taking 512 byte of a different |
| 1184 | * mirror, even if other 512 byte sectors in the same PAGE_SIZE | 1184 | * mirror, even if other 512 byte sectors in the same PAGE_SIZE |
| 1185 | * area are unreadable. | 1185 | * area are unreadable. |
| 1186 | */ | 1186 | */ |
| 1187 | 1187 | ||
| 1188 | /* can only fix I/O errors from here on */ | 1188 | /* can only fix I/O errors from here on */ |
| 1189 | if (sblock_bad->no_io_error_seen) | 1189 | if (sblock_bad->no_io_error_seen) |
| 1190 | goto did_not_correct_error; | 1190 | goto did_not_correct_error; |
| 1191 | 1191 | ||
| 1192 | success = 1; | 1192 | success = 1; |
| 1193 | for (page_num = 0; page_num < sblock_bad->page_count; page_num++) { | 1193 | for (page_num = 0; page_num < sblock_bad->page_count; page_num++) { |
| 1194 | struct scrub_page *page_bad = sblock_bad->pagev[page_num]; | 1194 | struct scrub_page *page_bad = sblock_bad->pagev[page_num]; |
| 1195 | 1195 | ||
| 1196 | if (!page_bad->io_error) | 1196 | if (!page_bad->io_error) |
| 1197 | continue; | 1197 | continue; |
| 1198 | 1198 | ||
| 1199 | for (mirror_index = 0; | 1199 | for (mirror_index = 0; |
| 1200 | mirror_index < BTRFS_MAX_MIRRORS && | 1200 | mirror_index < BTRFS_MAX_MIRRORS && |
| 1201 | sblocks_for_recheck[mirror_index].page_count > 0; | 1201 | sblocks_for_recheck[mirror_index].page_count > 0; |
| 1202 | mirror_index++) { | 1202 | mirror_index++) { |
| 1203 | struct scrub_block *sblock_other = sblocks_for_recheck + | 1203 | struct scrub_block *sblock_other = sblocks_for_recheck + |
| 1204 | mirror_index; | 1204 | mirror_index; |
| 1205 | struct scrub_page *page_other = sblock_other->pagev[ | 1205 | struct scrub_page *page_other = sblock_other->pagev[ |
| 1206 | page_num]; | 1206 | page_num]; |
| 1207 | 1207 | ||
| 1208 | if (!page_other->io_error) { | 1208 | if (!page_other->io_error) { |
| 1209 | ret = scrub_repair_page_from_good_copy( | 1209 | ret = scrub_repair_page_from_good_copy( |
| 1210 | sblock_bad, sblock_other, page_num, 0); | 1210 | sblock_bad, sblock_other, page_num, 0); |
| 1211 | if (0 == ret) { | 1211 | if (0 == ret) { |
| 1212 | page_bad->io_error = 0; | 1212 | page_bad->io_error = 0; |
| 1213 | break; /* succeeded for this page */ | 1213 | break; /* succeeded for this page */ |
| 1214 | } | 1214 | } |
| 1215 | } | 1215 | } |
| 1216 | } | 1216 | } |
| 1217 | 1217 | ||
| 1218 | if (page_bad->io_error) { | 1218 | if (page_bad->io_error) { |
| 1219 | /* did not find a mirror to copy the page from */ | 1219 | /* did not find a mirror to copy the page from */ |
| 1220 | success = 0; | 1220 | success = 0; |
| 1221 | } | 1221 | } |
| 1222 | } | 1222 | } |
| 1223 | 1223 | ||
| 1224 | if (success) { | 1224 | if (success) { |
| 1225 | if (is_metadata || have_csum) { | 1225 | if (is_metadata || have_csum) { |
| 1226 | /* | 1226 | /* |
| 1227 | * need to verify the checksum now that all | 1227 | * need to verify the checksum now that all |
| 1228 | * sectors on disk are repaired (the write | 1228 | * sectors on disk are repaired (the write |
| 1229 | * request for data to be repaired is on its way). | 1229 | * request for data to be repaired is on its way). |
| 1230 | * Just be lazy and use scrub_recheck_block() | 1230 | * Just be lazy and use scrub_recheck_block() |
| 1231 | * which re-reads the data before the checksum | 1231 | * which re-reads the data before the checksum |
| 1232 | * is verified, but most likely the data comes out | 1232 | * is verified, but most likely the data comes out |
| 1233 | * of the page cache. | 1233 | * of the page cache. |
| 1234 | */ | 1234 | */ |
| 1235 | scrub_recheck_block(fs_info, sblock_bad, | 1235 | scrub_recheck_block(fs_info, sblock_bad, |
| 1236 | is_metadata, have_csum, csum, | 1236 | is_metadata, have_csum, csum, |
| 1237 | generation, sctx->csum_size, 1); | 1237 | generation, sctx->csum_size, 1); |
| 1238 | if (!sblock_bad->header_error && | 1238 | if (!sblock_bad->header_error && |
| 1239 | !sblock_bad->checksum_error && | 1239 | !sblock_bad->checksum_error && |
| 1240 | sblock_bad->no_io_error_seen) | 1240 | sblock_bad->no_io_error_seen) |
| 1241 | goto corrected_error; | 1241 | goto corrected_error; |
| 1242 | else | 1242 | else |
| 1243 | goto did_not_correct_error; | 1243 | goto did_not_correct_error; |
| 1244 | } else { | 1244 | } else { |
| 1245 | corrected_error: | 1245 | corrected_error: |
| 1246 | spin_lock(&sctx->stat_lock); | 1246 | spin_lock(&sctx->stat_lock); |
| 1247 | sctx->stat.corrected_errors++; | 1247 | sctx->stat.corrected_errors++; |
| 1248 | sblock_to_check->data_corrected = 1; | 1248 | sblock_to_check->data_corrected = 1; |
| 1249 | spin_unlock(&sctx->stat_lock); | 1249 | spin_unlock(&sctx->stat_lock); |
| 1250 | printk_ratelimited_in_rcu(KERN_ERR | 1250 | printk_ratelimited_in_rcu(KERN_ERR |
| 1251 | "BTRFS: fixed up error at logical %llu on dev %s\n", | 1251 | "BTRFS: fixed up error at logical %llu on dev %s\n", |
| 1252 | logical, rcu_str_deref(dev->name)); | 1252 | logical, rcu_str_deref(dev->name)); |
| 1253 | } | 1253 | } |
| 1254 | } else { | 1254 | } else { |
| 1255 | did_not_correct_error: | 1255 | did_not_correct_error: |
| 1256 | spin_lock(&sctx->stat_lock); | 1256 | spin_lock(&sctx->stat_lock); |
| 1257 | sctx->stat.uncorrectable_errors++; | 1257 | sctx->stat.uncorrectable_errors++; |
| 1258 | spin_unlock(&sctx->stat_lock); | 1258 | spin_unlock(&sctx->stat_lock); |
| 1259 | printk_ratelimited_in_rcu(KERN_ERR | 1259 | printk_ratelimited_in_rcu(KERN_ERR |
| 1260 | "BTRFS: unable to fixup (regular) error at logical %llu on dev %s\n", | 1260 | "BTRFS: unable to fixup (regular) error at logical %llu on dev %s\n", |
| 1261 | logical, rcu_str_deref(dev->name)); | 1261 | logical, rcu_str_deref(dev->name)); |
| 1262 | } | 1262 | } |
| 1263 | 1263 | ||
| 1264 | out: | 1264 | out: |
| 1265 | if (sblocks_for_recheck) { | 1265 | if (sblocks_for_recheck) { |
| 1266 | for (mirror_index = 0; mirror_index < BTRFS_MAX_MIRRORS; | 1266 | for (mirror_index = 0; mirror_index < BTRFS_MAX_MIRRORS; |
| 1267 | mirror_index++) { | 1267 | mirror_index++) { |
| 1268 | struct scrub_block *sblock = sblocks_for_recheck + | 1268 | struct scrub_block *sblock = sblocks_for_recheck + |
| 1269 | mirror_index; | 1269 | mirror_index; |
| 1270 | struct scrub_recover *recover; | 1270 | struct scrub_recover *recover; |
| 1271 | int page_index; | 1271 | int page_index; |
| 1272 | 1272 | ||
| 1273 | for (page_index = 0; page_index < sblock->page_count; | 1273 | for (page_index = 0; page_index < sblock->page_count; |
| 1274 | page_index++) { | 1274 | page_index++) { |
| 1275 | sblock->pagev[page_index]->sblock = NULL; | 1275 | sblock->pagev[page_index]->sblock = NULL; |
| 1276 | recover = sblock->pagev[page_index]->recover; | 1276 | recover = sblock->pagev[page_index]->recover; |
| 1277 | if (recover) { | 1277 | if (recover) { |
| 1278 | scrub_put_recover(recover); | 1278 | scrub_put_recover(recover); |
| 1279 | sblock->pagev[page_index]->recover = | 1279 | sblock->pagev[page_index]->recover = |
| 1280 | NULL; | 1280 | NULL; |
| 1281 | } | 1281 | } |
| 1282 | scrub_page_put(sblock->pagev[page_index]); | 1282 | scrub_page_put(sblock->pagev[page_index]); |
| 1283 | } | 1283 | } |
| 1284 | } | 1284 | } |
| 1285 | kfree(sblocks_for_recheck); | 1285 | kfree(sblocks_for_recheck); |
| 1286 | } | 1286 | } |
| 1287 | 1287 | ||
| 1288 | return 0; | 1288 | return 0; |
| 1289 | } | 1289 | } |
| 1290 | 1290 | ||
| 1291 | static inline int scrub_nr_raid_mirrors(struct btrfs_bio *bbio, u64 *raid_map) | 1291 | static inline int scrub_nr_raid_mirrors(struct btrfs_bio *bbio, u64 *raid_map) |
| 1292 | { | 1292 | { |
| 1293 | if (raid_map) { | 1293 | if (raid_map) { |
| 1294 | if (raid_map[bbio->num_stripes - 1] == RAID6_Q_STRIPE) | 1294 | if (raid_map[bbio->num_stripes - 1] == RAID6_Q_STRIPE) |
| 1295 | return 3; | 1295 | return 3; |
| 1296 | else | 1296 | else |
| 1297 | return 2; | 1297 | return 2; |
| 1298 | } else { | 1298 | } else { |
| 1299 | return (int)bbio->num_stripes; | 1299 | return (int)bbio->num_stripes; |
| 1300 | } | 1300 | } |
| 1301 | } | 1301 | } |
| 1302 | 1302 | ||
| 1303 | static inline void scrub_stripe_index_and_offset(u64 logical, u64 *raid_map, | 1303 | static inline void scrub_stripe_index_and_offset(u64 logical, u64 *raid_map, |
| 1304 | u64 mapped_length, | 1304 | u64 mapped_length, |
| 1305 | int nstripes, int mirror, | 1305 | int nstripes, int mirror, |
| 1306 | int *stripe_index, | 1306 | int *stripe_index, |
| 1307 | u64 *stripe_offset) | 1307 | u64 *stripe_offset) |
| 1308 | { | 1308 | { |
| 1309 | int i; | 1309 | int i; |
| 1310 | 1310 | ||
| 1311 | if (raid_map) { | 1311 | if (raid_map) { |
| 1312 | /* RAID5/6 */ | 1312 | /* RAID5/6 */ |
| 1313 | for (i = 0; i < nstripes; i++) { | 1313 | for (i = 0; i < nstripes; i++) { |
| 1314 | if (raid_map[i] == RAID6_Q_STRIPE || | 1314 | if (raid_map[i] == RAID6_Q_STRIPE || |
| 1315 | raid_map[i] == RAID5_P_STRIPE) | 1315 | raid_map[i] == RAID5_P_STRIPE) |
| 1316 | continue; | 1316 | continue; |
| 1317 | 1317 | ||
| 1318 | if (logical >= raid_map[i] && | 1318 | if (logical >= raid_map[i] && |
| 1319 | logical < raid_map[i] + mapped_length) | 1319 | logical < raid_map[i] + mapped_length) |
| 1320 | break; | 1320 | break; |
| 1321 | } | 1321 | } |
| 1322 | 1322 | ||
| 1323 | *stripe_index = i; | 1323 | *stripe_index = i; |
| 1324 | *stripe_offset = logical - raid_map[i]; | 1324 | *stripe_offset = logical - raid_map[i]; |
| 1325 | } else { | 1325 | } else { |
| 1326 | /* The other RAID type */ | 1326 | /* The other RAID type */ |
| 1327 | *stripe_index = mirror; | 1327 | *stripe_index = mirror; |
| 1328 | *stripe_offset = 0; | 1328 | *stripe_offset = 0; |
| 1329 | } | 1329 | } |
| 1330 | } | 1330 | } |
| 1331 | 1331 | ||
| 1332 | static int scrub_setup_recheck_block(struct scrub_ctx *sctx, | 1332 | static int scrub_setup_recheck_block(struct scrub_ctx *sctx, |
| 1333 | struct btrfs_fs_info *fs_info, | 1333 | struct btrfs_fs_info *fs_info, |
| 1334 | struct scrub_block *original_sblock, | 1334 | struct scrub_block *original_sblock, |
| 1335 | u64 length, u64 logical, | 1335 | u64 length, u64 logical, |
| 1336 | struct scrub_block *sblocks_for_recheck) | 1336 | struct scrub_block *sblocks_for_recheck) |
| 1337 | { | 1337 | { |
| 1338 | struct scrub_recover *recover; | 1338 | struct scrub_recover *recover; |
| 1339 | struct btrfs_bio *bbio; | 1339 | struct btrfs_bio *bbio; |
| 1340 | u64 *raid_map; | 1340 | u64 *raid_map; |
| 1341 | u64 sublen; | 1341 | u64 sublen; |
| 1342 | u64 mapped_length; | 1342 | u64 mapped_length; |
| 1343 | u64 stripe_offset; | 1343 | u64 stripe_offset; |
| 1344 | int stripe_index; | 1344 | int stripe_index; |
| 1345 | int page_index; | 1345 | int page_index; |
| 1346 | int mirror_index; | 1346 | int mirror_index; |
| 1347 | int nmirrors; | 1347 | int nmirrors; |
| 1348 | int ret; | 1348 | int ret; |
| 1349 | 1349 | ||
| 1350 | /* | 1350 | /* |
| 1351 | * note: the two members ref_count and outstanding_pages | 1351 | * note: the two members ref_count and outstanding_pages |
| 1352 | * are not used (and not set) in the blocks that are used for | 1352 | * are not used (and not set) in the blocks that are used for |
| 1353 | * the recheck procedure | 1353 | * the recheck procedure |
| 1354 | */ | 1354 | */ |
| 1355 | 1355 | ||
| 1356 | page_index = 0; | 1356 | page_index = 0; |
| 1357 | while (length > 0) { | 1357 | while (length > 0) { |
| 1358 | sublen = min_t(u64, length, PAGE_SIZE); | 1358 | sublen = min_t(u64, length, PAGE_SIZE); |
| 1359 | mapped_length = sublen; | 1359 | mapped_length = sublen; |
| 1360 | bbio = NULL; | 1360 | bbio = NULL; |
| 1361 | raid_map = NULL; | 1361 | raid_map = NULL; |
| 1362 | 1362 | ||
| 1363 | /* | 1363 | /* |
| 1364 | * with a length of PAGE_SIZE, each returned stripe | 1364 | * with a length of PAGE_SIZE, each returned stripe |
| 1365 | * represents one mirror | 1365 | * represents one mirror |
| 1366 | */ | 1366 | */ |
| 1367 | ret = btrfs_map_sblock(fs_info, REQ_GET_READ_MIRRORS, logical, | 1367 | ret = btrfs_map_sblock(fs_info, REQ_GET_READ_MIRRORS, logical, |
| 1368 | &mapped_length, &bbio, 0, &raid_map); | 1368 | &mapped_length, &bbio, 0, &raid_map); |
| 1369 | if (ret || !bbio || mapped_length < sublen) { | 1369 | if (ret || !bbio || mapped_length < sublen) { |
| 1370 | kfree(bbio); | 1370 | kfree(bbio); |
| 1371 | kfree(raid_map); | 1371 | kfree(raid_map); |
| 1372 | return -EIO; | 1372 | return -EIO; |
| 1373 | } | 1373 | } |
| 1374 | 1374 | ||
| 1375 | recover = kzalloc(sizeof(struct scrub_recover), GFP_NOFS); | 1375 | recover = kzalloc(sizeof(struct scrub_recover), GFP_NOFS); |
| 1376 | if (!recover) { | 1376 | if (!recover) { |
| 1377 | kfree(bbio); | 1377 | kfree(bbio); |
| 1378 | kfree(raid_map); | 1378 | kfree(raid_map); |
| 1379 | return -ENOMEM; | 1379 | return -ENOMEM; |
| 1380 | } | 1380 | } |
| 1381 | 1381 | ||
| 1382 | atomic_set(&recover->refs, 1); | 1382 | atomic_set(&recover->refs, 1); |
| 1383 | recover->bbio = bbio; | 1383 | recover->bbio = bbio; |
| 1384 | recover->raid_map = raid_map; | 1384 | recover->raid_map = raid_map; |
| 1385 | recover->map_length = mapped_length; | 1385 | recover->map_length = mapped_length; |
| 1386 | 1386 | ||
| 1387 | BUG_ON(page_index >= SCRUB_PAGES_PER_RD_BIO); | 1387 | BUG_ON(page_index >= SCRUB_PAGES_PER_RD_BIO); |
| 1388 | 1388 | ||
| 1389 | nmirrors = scrub_nr_raid_mirrors(bbio, raid_map); | 1389 | nmirrors = scrub_nr_raid_mirrors(bbio, raid_map); |
| 1390 | for (mirror_index = 0; mirror_index < nmirrors; | 1390 | for (mirror_index = 0; mirror_index < nmirrors; |
| 1391 | mirror_index++) { | 1391 | mirror_index++) { |
| 1392 | struct scrub_block *sblock; | 1392 | struct scrub_block *sblock; |
| 1393 | struct scrub_page *page; | 1393 | struct scrub_page *page; |
| 1394 | 1394 | ||
| 1395 | if (mirror_index >= BTRFS_MAX_MIRRORS) | 1395 | if (mirror_index >= BTRFS_MAX_MIRRORS) |
| 1396 | continue; | 1396 | continue; |
| 1397 | 1397 | ||
| 1398 | sblock = sblocks_for_recheck + mirror_index; | 1398 | sblock = sblocks_for_recheck + mirror_index; |
| 1399 | sblock->sctx = sctx; | 1399 | sblock->sctx = sctx; |
| 1400 | page = kzalloc(sizeof(*page), GFP_NOFS); | 1400 | page = kzalloc(sizeof(*page), GFP_NOFS); |
| 1401 | if (!page) { | 1401 | if (!page) { |
| 1402 | leave_nomem: | 1402 | leave_nomem: |
| 1403 | spin_lock(&sctx->stat_lock); | 1403 | spin_lock(&sctx->stat_lock); |
| 1404 | sctx->stat.malloc_errors++; | 1404 | sctx->stat.malloc_errors++; |
| 1405 | spin_unlock(&sctx->stat_lock); | 1405 | spin_unlock(&sctx->stat_lock); |
| 1406 | scrub_put_recover(recover); | 1406 | scrub_put_recover(recover); |
| 1407 | return -ENOMEM; | 1407 | return -ENOMEM; |
| 1408 | } | 1408 | } |
| 1409 | scrub_page_get(page); | 1409 | scrub_page_get(page); |
| 1410 | sblock->pagev[page_index] = page; | 1410 | sblock->pagev[page_index] = page; |
| 1411 | page->logical = logical; | 1411 | page->logical = logical; |
| 1412 | 1412 | ||
| 1413 | scrub_stripe_index_and_offset(logical, raid_map, | 1413 | scrub_stripe_index_and_offset(logical, raid_map, |
| 1414 | mapped_length, | 1414 | mapped_length, |
| 1415 | bbio->num_stripes, | 1415 | bbio->num_stripes, |
| 1416 | mirror_index, | 1416 | mirror_index, |
| 1417 | &stripe_index, | 1417 | &stripe_index, |
| 1418 | &stripe_offset); | 1418 | &stripe_offset); |
| 1419 | page->physical = bbio->stripes[stripe_index].physical + | 1419 | page->physical = bbio->stripes[stripe_index].physical + |
| 1420 | stripe_offset; | 1420 | stripe_offset; |
| 1421 | page->dev = bbio->stripes[stripe_index].dev; | 1421 | page->dev = bbio->stripes[stripe_index].dev; |
| 1422 | 1422 | ||
| 1423 | BUG_ON(page_index >= original_sblock->page_count); | 1423 | BUG_ON(page_index >= original_sblock->page_count); |
| 1424 | page->physical_for_dev_replace = | 1424 | page->physical_for_dev_replace = |
| 1425 | original_sblock->pagev[page_index]-> | 1425 | original_sblock->pagev[page_index]-> |
| 1426 | physical_for_dev_replace; | 1426 | physical_for_dev_replace; |
| 1427 | /* for missing devices, dev->bdev is NULL */ | 1427 | /* for missing devices, dev->bdev is NULL */ |
| 1428 | page->mirror_num = mirror_index + 1; | 1428 | page->mirror_num = mirror_index + 1; |
| 1429 | sblock->page_count++; | 1429 | sblock->page_count++; |
| 1430 | page->page = alloc_page(GFP_NOFS); | 1430 | page->page = alloc_page(GFP_NOFS); |
| 1431 | if (!page->page) | 1431 | if (!page->page) |
| 1432 | goto leave_nomem; | 1432 | goto leave_nomem; |
| 1433 | 1433 | ||
| 1434 | scrub_get_recover(recover); | 1434 | scrub_get_recover(recover); |
| 1435 | page->recover = recover; | 1435 | page->recover = recover; |
| 1436 | } | 1436 | } |
| 1437 | scrub_put_recover(recover); | 1437 | scrub_put_recover(recover); |
| 1438 | length -= sublen; | 1438 | length -= sublen; |
| 1439 | logical += sublen; | 1439 | logical += sublen; |
| 1440 | page_index++; | 1440 | page_index++; |
| 1441 | } | 1441 | } |
| 1442 | 1442 | ||
| 1443 | return 0; | 1443 | return 0; |
| 1444 | } | 1444 | } |
| 1445 | 1445 | ||
| 1446 | struct scrub_bio_ret { | 1446 | struct scrub_bio_ret { |
| 1447 | struct completion event; | 1447 | struct completion event; |
| 1448 | int error; | 1448 | int error; |
| 1449 | }; | 1449 | }; |
| 1450 | 1450 | ||
| 1451 | static void scrub_bio_wait_endio(struct bio *bio, int error) | 1451 | static void scrub_bio_wait_endio(struct bio *bio, int error) |
| 1452 | { | 1452 | { |
| 1453 | struct scrub_bio_ret *ret = bio->bi_private; | 1453 | struct scrub_bio_ret *ret = bio->bi_private; |
| 1454 | 1454 | ||
| 1455 | ret->error = error; | 1455 | ret->error = error; |
| 1456 | complete(&ret->event); | 1456 | complete(&ret->event); |
| 1457 | } | 1457 | } |
| 1458 | 1458 | ||
| 1459 | static inline int scrub_is_page_on_raid56(struct scrub_page *page) | 1459 | static inline int scrub_is_page_on_raid56(struct scrub_page *page) |
| 1460 | { | 1460 | { |
| 1461 | return page->recover && page->recover->raid_map; | 1461 | return page->recover && page->recover->raid_map; |
| 1462 | } | 1462 | } |
| 1463 | 1463 | ||
| 1464 | static int scrub_submit_raid56_bio_wait(struct btrfs_fs_info *fs_info, | 1464 | static int scrub_submit_raid56_bio_wait(struct btrfs_fs_info *fs_info, |
| 1465 | struct bio *bio, | 1465 | struct bio *bio, |
| 1466 | struct scrub_page *page) | 1466 | struct scrub_page *page) |
| 1467 | { | 1467 | { |
| 1468 | struct scrub_bio_ret done; | 1468 | struct scrub_bio_ret done; |
| 1469 | int ret; | 1469 | int ret; |
| 1470 | 1470 | ||
| 1471 | init_completion(&done.event); | 1471 | init_completion(&done.event); |
| 1472 | done.error = 0; | 1472 | done.error = 0; |
| 1473 | bio->bi_iter.bi_sector = page->logical >> 9; | 1473 | bio->bi_iter.bi_sector = page->logical >> 9; |
| 1474 | bio->bi_private = &done; | 1474 | bio->bi_private = &done; |
| 1475 | bio->bi_end_io = scrub_bio_wait_endio; | 1475 | bio->bi_end_io = scrub_bio_wait_endio; |
| 1476 | 1476 | ||
| 1477 | ret = raid56_parity_recover(fs_info->fs_root, bio, page->recover->bbio, | 1477 | ret = raid56_parity_recover(fs_info->fs_root, bio, page->recover->bbio, |
| 1478 | page->recover->raid_map, | 1478 | page->recover->raid_map, |
| 1479 | page->recover->map_length, | 1479 | page->recover->map_length, |
| 1480 | page->mirror_num, 0); | 1480 | page->mirror_num, 0); |
| 1481 | if (ret) | 1481 | if (ret) |
| 1482 | return ret; | 1482 | return ret; |
| 1483 | 1483 | ||
| 1484 | wait_for_completion(&done.event); | 1484 | wait_for_completion(&done.event); |
| 1485 | if (done.error) | 1485 | if (done.error) |
| 1486 | return -EIO; | 1486 | return -EIO; |
| 1487 | 1487 | ||
| 1488 | return 0; | 1488 | return 0; |
| 1489 | } | 1489 | } |
| 1490 | 1490 | ||
| 1491 | /* | 1491 | /* |
| 1492 | * this function will check the on disk data for checksum errors, header | 1492 | * this function will check the on disk data for checksum errors, header |
| 1493 | * errors and read I/O errors. If any I/O errors happen, the exact pages | 1493 | * errors and read I/O errors. If any I/O errors happen, the exact pages |
| 1494 | * which are errored are marked as being bad. The goal is to enable scrub | 1494 | * which are errored are marked as being bad. The goal is to enable scrub |
| 1495 | * to take those pages that are not errored from all the mirrors so that | 1495 | * to take those pages that are not errored from all the mirrors so that |
| 1496 | * the pages that are errored in the just handled mirror can be repaired. | 1496 | * the pages that are errored in the just handled mirror can be repaired. |
| 1497 | */ | 1497 | */ |
| 1498 | static void scrub_recheck_block(struct btrfs_fs_info *fs_info, | 1498 | static void scrub_recheck_block(struct btrfs_fs_info *fs_info, |
| 1499 | struct scrub_block *sblock, int is_metadata, | 1499 | struct scrub_block *sblock, int is_metadata, |
| 1500 | int have_csum, u8 *csum, u64 generation, | 1500 | int have_csum, u8 *csum, u64 generation, |
| 1501 | u16 csum_size, int retry_failed_mirror) | 1501 | u16 csum_size, int retry_failed_mirror) |
| 1502 | { | 1502 | { |
| 1503 | int page_num; | 1503 | int page_num; |
| 1504 | 1504 | ||
| 1505 | sblock->no_io_error_seen = 1; | 1505 | sblock->no_io_error_seen = 1; |
| 1506 | sblock->header_error = 0; | 1506 | sblock->header_error = 0; |
| 1507 | sblock->checksum_error = 0; | 1507 | sblock->checksum_error = 0; |
| 1508 | 1508 | ||
| 1509 | for (page_num = 0; page_num < sblock->page_count; page_num++) { | 1509 | for (page_num = 0; page_num < sblock->page_count; page_num++) { |
| 1510 | struct bio *bio; | 1510 | struct bio *bio; |
| 1511 | struct scrub_page *page = sblock->pagev[page_num]; | 1511 | struct scrub_page *page = sblock->pagev[page_num]; |
| 1512 | 1512 | ||
| 1513 | if (page->dev->bdev == NULL) { | 1513 | if (page->dev->bdev == NULL) { |
| 1514 | page->io_error = 1; | 1514 | page->io_error = 1; |
| 1515 | sblock->no_io_error_seen = 0; | 1515 | sblock->no_io_error_seen = 0; |
| 1516 | continue; | 1516 | continue; |
| 1517 | } | 1517 | } |
| 1518 | 1518 | ||
| 1519 | WARN_ON(!page->page); | 1519 | WARN_ON(!page->page); |
| 1520 | bio = btrfs_io_bio_alloc(GFP_NOFS, 1); | 1520 | bio = btrfs_io_bio_alloc(GFP_NOFS, 1); |
| 1521 | if (!bio) { | 1521 | if (!bio) { |
| 1522 | page->io_error = 1; | 1522 | page->io_error = 1; |
| 1523 | sblock->no_io_error_seen = 0; | 1523 | sblock->no_io_error_seen = 0; |
| 1524 | continue; | 1524 | continue; |
| 1525 | } | 1525 | } |
| 1526 | bio->bi_bdev = page->dev->bdev; | 1526 | bio->bi_bdev = page->dev->bdev; |
| 1527 | 1527 | ||
| 1528 | bio_add_page(bio, page->page, PAGE_SIZE, 0); | 1528 | bio_add_page(bio, page->page, PAGE_SIZE, 0); |
| 1529 | if (!retry_failed_mirror && scrub_is_page_on_raid56(page)) { | 1529 | if (!retry_failed_mirror && scrub_is_page_on_raid56(page)) { |
| 1530 | if (scrub_submit_raid56_bio_wait(fs_info, bio, page)) | 1530 | if (scrub_submit_raid56_bio_wait(fs_info, bio, page)) |
| 1531 | sblock->no_io_error_seen = 0; | 1531 | sblock->no_io_error_seen = 0; |
| 1532 | } else { | 1532 | } else { |
| 1533 | bio->bi_iter.bi_sector = page->physical >> 9; | 1533 | bio->bi_iter.bi_sector = page->physical >> 9; |
| 1534 | 1534 | ||
| 1535 | if (btrfsic_submit_bio_wait(READ, bio)) | 1535 | if (btrfsic_submit_bio_wait(READ, bio)) |
| 1536 | sblock->no_io_error_seen = 0; | 1536 | sblock->no_io_error_seen = 0; |
| 1537 | } | 1537 | } |
| 1538 | 1538 | ||
| 1539 | bio_put(bio); | 1539 | bio_put(bio); |
| 1540 | } | 1540 | } |
| 1541 | 1541 | ||
| 1542 | if (sblock->no_io_error_seen) | 1542 | if (sblock->no_io_error_seen) |
| 1543 | scrub_recheck_block_checksum(fs_info, sblock, is_metadata, | 1543 | scrub_recheck_block_checksum(fs_info, sblock, is_metadata, |
| 1544 | have_csum, csum, generation, | 1544 | have_csum, csum, generation, |
| 1545 | csum_size); | 1545 | csum_size); |
| 1546 | 1546 | ||
| 1547 | return; | 1547 | return; |
| 1548 | } | 1548 | } |
| 1549 | 1549 | ||
| 1550 | static inline int scrub_check_fsid(u8 fsid[], | 1550 | static inline int scrub_check_fsid(u8 fsid[], |
| 1551 | struct scrub_page *spage) | 1551 | struct scrub_page *spage) |
| 1552 | { | 1552 | { |
| 1553 | struct btrfs_fs_devices *fs_devices = spage->dev->fs_devices; | 1553 | struct btrfs_fs_devices *fs_devices = spage->dev->fs_devices; |
| 1554 | int ret; | 1554 | int ret; |
| 1555 | 1555 | ||
| 1556 | ret = memcmp(fsid, fs_devices->fsid, BTRFS_UUID_SIZE); | 1556 | ret = memcmp(fsid, fs_devices->fsid, BTRFS_UUID_SIZE); |
| 1557 | return !ret; | 1557 | return !ret; |
| 1558 | } | 1558 | } |
| 1559 | 1559 | ||
| 1560 | static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info, | 1560 | static void scrub_recheck_block_checksum(struct btrfs_fs_info *fs_info, |
| 1561 | struct scrub_block *sblock, | 1561 | struct scrub_block *sblock, |
| 1562 | int is_metadata, int have_csum, | 1562 | int is_metadata, int have_csum, |
| 1563 | const u8 *csum, u64 generation, | 1563 | const u8 *csum, u64 generation, |
| 1564 | u16 csum_size) | 1564 | u16 csum_size) |
| 1565 | { | 1565 | { |
| 1566 | int page_num; | 1566 | int page_num; |
| 1567 | u8 calculated_csum[BTRFS_CSUM_SIZE]; | 1567 | u8 calculated_csum[BTRFS_CSUM_SIZE]; |
| 1568 | u32 crc = ~(u32)0; | 1568 | u32 crc = ~(u32)0; |
| 1569 | void *mapped_buffer; | 1569 | void *mapped_buffer; |
| 1570 | 1570 | ||
| 1571 | WARN_ON(!sblock->pagev[0]->page); | 1571 | WARN_ON(!sblock->pagev[0]->page); |
| 1572 | if (is_metadata) { | 1572 | if (is_metadata) { |
| 1573 | struct btrfs_header *h; | 1573 | struct btrfs_header *h; |
| 1574 | 1574 | ||
| 1575 | mapped_buffer = kmap_atomic(sblock->pagev[0]->page); | 1575 | mapped_buffer = kmap_atomic(sblock->pagev[0]->page); |
| 1576 | h = (struct btrfs_header *)mapped_buffer; | 1576 | h = (struct btrfs_header *)mapped_buffer; |
| 1577 | 1577 | ||
| 1578 | if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h) || | 1578 | if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h) || |
| 1579 | !scrub_check_fsid(h->fsid, sblock->pagev[0]) || | 1579 | !scrub_check_fsid(h->fsid, sblock->pagev[0]) || |
| 1580 | memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, | 1580 | memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, |
| 1581 | BTRFS_UUID_SIZE)) { | 1581 | BTRFS_UUID_SIZE)) { |
| 1582 | sblock->header_error = 1; | 1582 | sblock->header_error = 1; |
| 1583 | } else if (generation != btrfs_stack_header_generation(h)) { | 1583 | } else if (generation != btrfs_stack_header_generation(h)) { |
| 1584 | sblock->header_error = 1; | 1584 | sblock->header_error = 1; |
| 1585 | sblock->generation_error = 1; | 1585 | sblock->generation_error = 1; |
| 1586 | } | 1586 | } |
| 1587 | csum = h->csum; | 1587 | csum = h->csum; |
| 1588 | } else { | 1588 | } else { |
| 1589 | if (!have_csum) | 1589 | if (!have_csum) |
| 1590 | return; | 1590 | return; |
| 1591 | 1591 | ||
| 1592 | mapped_buffer = kmap_atomic(sblock->pagev[0]->page); | 1592 | mapped_buffer = kmap_atomic(sblock->pagev[0]->page); |
| 1593 | } | 1593 | } |
| 1594 | 1594 | ||
| 1595 | for (page_num = 0;;) { | 1595 | for (page_num = 0;;) { |
| 1596 | if (page_num == 0 && is_metadata) | 1596 | if (page_num == 0 && is_metadata) |
| 1597 | crc = btrfs_csum_data( | 1597 | crc = btrfs_csum_data( |
| 1598 | ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE, | 1598 | ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE, |
| 1599 | crc, PAGE_SIZE - BTRFS_CSUM_SIZE); | 1599 | crc, PAGE_SIZE - BTRFS_CSUM_SIZE); |
| 1600 | else | 1600 | else |
| 1601 | crc = btrfs_csum_data(mapped_buffer, crc, PAGE_SIZE); | 1601 | crc = btrfs_csum_data(mapped_buffer, crc, PAGE_SIZE); |
| 1602 | 1602 | ||
| 1603 | kunmap_atomic(mapped_buffer); | 1603 | kunmap_atomic(mapped_buffer); |
| 1604 | page_num++; | 1604 | page_num++; |
| 1605 | if (page_num >= sblock->page_count) | 1605 | if (page_num >= sblock->page_count) |
| 1606 | break; | 1606 | break; |
| 1607 | WARN_ON(!sblock->pagev[page_num]->page); | 1607 | WARN_ON(!sblock->pagev[page_num]->page); |
| 1608 | 1608 | ||
| 1609 | mapped_buffer = kmap_atomic(sblock->pagev[page_num]->page); | 1609 | mapped_buffer = kmap_atomic(sblock->pagev[page_num]->page); |
| 1610 | } | 1610 | } |
| 1611 | 1611 | ||
| 1612 | btrfs_csum_final(crc, calculated_csum); | 1612 | btrfs_csum_final(crc, calculated_csum); |
| 1613 | if (memcmp(calculated_csum, csum, csum_size)) | 1613 | if (memcmp(calculated_csum, csum, csum_size)) |
| 1614 | sblock->checksum_error = 1; | 1614 | sblock->checksum_error = 1; |
| 1615 | } | 1615 | } |
| 1616 | 1616 | ||
| 1617 | static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad, | 1617 | static int scrub_repair_block_from_good_copy(struct scrub_block *sblock_bad, |
| 1618 | struct scrub_block *sblock_good, | 1618 | struct scrub_block *sblock_good, |
| 1619 | int force_write) | 1619 | int force_write) |
| 1620 | { | 1620 | { |
| 1621 | int page_num; | 1621 | int page_num; |
| 1622 | int ret = 0; | 1622 | int ret = 0; |
| 1623 | 1623 | ||
| 1624 | for (page_num = 0; page_num < sblock_bad->page_count; page_num++) { | 1624 | for (page_num = 0; page_num < sblock_bad->page_count; page_num++) { |
| 1625 | int ret_sub; | 1625 | int ret_sub; |
| 1626 | 1626 | ||
| 1627 | ret_sub = scrub_repair_page_from_good_copy(sblock_bad, | 1627 | ret_sub = scrub_repair_page_from_good_copy(sblock_bad, |
| 1628 | sblock_good, | 1628 | sblock_good, |
| 1629 | page_num, | 1629 | page_num, |
| 1630 | force_write); | 1630 | force_write); |
| 1631 | if (ret_sub) | 1631 | if (ret_sub) |
| 1632 | ret = ret_sub; | 1632 | ret = ret_sub; |
| 1633 | } | 1633 | } |
| 1634 | 1634 | ||
| 1635 | return ret; | 1635 | return ret; |
| 1636 | } | 1636 | } |
| 1637 | 1637 | ||
| 1638 | static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad, | 1638 | static int scrub_repair_page_from_good_copy(struct scrub_block *sblock_bad, |
| 1639 | struct scrub_block *sblock_good, | 1639 | struct scrub_block *sblock_good, |
| 1640 | int page_num, int force_write) | 1640 | int page_num, int force_write) |
| 1641 | { | 1641 | { |
| 1642 | struct scrub_page *page_bad = sblock_bad->pagev[page_num]; | 1642 | struct scrub_page *page_bad = sblock_bad->pagev[page_num]; |
| 1643 | struct scrub_page *page_good = sblock_good->pagev[page_num]; | 1643 | struct scrub_page *page_good = sblock_good->pagev[page_num]; |
| 1644 | 1644 | ||
| 1645 | BUG_ON(page_bad->page == NULL); | 1645 | BUG_ON(page_bad->page == NULL); |
| 1646 | BUG_ON(page_good->page == NULL); | 1646 | BUG_ON(page_good->page == NULL); |
| 1647 | if (force_write || sblock_bad->header_error || | 1647 | if (force_write || sblock_bad->header_error || |
| 1648 | sblock_bad->checksum_error || page_bad->io_error) { | 1648 | sblock_bad->checksum_error || page_bad->io_error) { |
| 1649 | struct bio *bio; | 1649 | struct bio *bio; |
| 1650 | int ret; | 1650 | int ret; |
| 1651 | 1651 | ||
| 1652 | if (!page_bad->dev->bdev) { | 1652 | if (!page_bad->dev->bdev) { |
| 1653 | printk_ratelimited(KERN_WARNING "BTRFS: " | 1653 | printk_ratelimited(KERN_WARNING "BTRFS: " |
| 1654 | "scrub_repair_page_from_good_copy(bdev == NULL) " | 1654 | "scrub_repair_page_from_good_copy(bdev == NULL) " |
| 1655 | "is unexpected!\n"); | 1655 | "is unexpected!\n"); |
| 1656 | return -EIO; | 1656 | return -EIO; |
| 1657 | } | 1657 | } |
| 1658 | 1658 | ||
| 1659 | bio = btrfs_io_bio_alloc(GFP_NOFS, 1); | 1659 | bio = btrfs_io_bio_alloc(GFP_NOFS, 1); |
| 1660 | if (!bio) | 1660 | if (!bio) |
| 1661 | return -EIO; | 1661 | return -EIO; |
| 1662 | bio->bi_bdev = page_bad->dev->bdev; | 1662 | bio->bi_bdev = page_bad->dev->bdev; |
| 1663 | bio->bi_iter.bi_sector = page_bad->physical >> 9; | 1663 | bio->bi_iter.bi_sector = page_bad->physical >> 9; |
| 1664 | 1664 | ||
| 1665 | ret = bio_add_page(bio, page_good->page, PAGE_SIZE, 0); | 1665 | ret = bio_add_page(bio, page_good->page, PAGE_SIZE, 0); |
| 1666 | if (PAGE_SIZE != ret) { | 1666 | if (PAGE_SIZE != ret) { |
| 1667 | bio_put(bio); | 1667 | bio_put(bio); |
| 1668 | return -EIO; | 1668 | return -EIO; |
| 1669 | } | 1669 | } |
| 1670 | 1670 | ||
| 1671 | if (btrfsic_submit_bio_wait(WRITE, bio)) { | 1671 | if (btrfsic_submit_bio_wait(WRITE, bio)) { |
| 1672 | btrfs_dev_stat_inc_and_print(page_bad->dev, | 1672 | btrfs_dev_stat_inc_and_print(page_bad->dev, |
| 1673 | BTRFS_DEV_STAT_WRITE_ERRS); | 1673 | BTRFS_DEV_STAT_WRITE_ERRS); |
| 1674 | btrfs_dev_replace_stats_inc( | 1674 | btrfs_dev_replace_stats_inc( |
| 1675 | &sblock_bad->sctx->dev_root->fs_info-> | 1675 | &sblock_bad->sctx->dev_root->fs_info-> |
| 1676 | dev_replace.num_write_errors); | 1676 | dev_replace.num_write_errors); |
| 1677 | bio_put(bio); | 1677 | bio_put(bio); |
| 1678 | return -EIO; | 1678 | return -EIO; |
| 1679 | } | 1679 | } |
| 1680 | bio_put(bio); | 1680 | bio_put(bio); |
| 1681 | } | 1681 | } |
| 1682 | 1682 | ||
| 1683 | return 0; | 1683 | return 0; |
| 1684 | } | 1684 | } |
| 1685 | 1685 | ||
| 1686 | static void scrub_write_block_to_dev_replace(struct scrub_block *sblock) | 1686 | static void scrub_write_block_to_dev_replace(struct scrub_block *sblock) |
| 1687 | { | 1687 | { |
| 1688 | int page_num; | 1688 | int page_num; |
| 1689 | 1689 | ||
| 1690 | /* | 1690 | /* |
| 1691 | * This block is used for the check of the parity on the source device, | 1691 | * This block is used for the check of the parity on the source device, |
| 1692 | * so the data needn't be written into the destination device. | 1692 | * so the data needn't be written into the destination device. |
| 1693 | */ | 1693 | */ |
| 1694 | if (sblock->sparity) | 1694 | if (sblock->sparity) |
| 1695 | return; | 1695 | return; |
| 1696 | 1696 | ||
| 1697 | for (page_num = 0; page_num < sblock->page_count; page_num++) { | 1697 | for (page_num = 0; page_num < sblock->page_count; page_num++) { |
| 1698 | int ret; | 1698 | int ret; |
| 1699 | 1699 | ||
| 1700 | ret = scrub_write_page_to_dev_replace(sblock, page_num); | 1700 | ret = scrub_write_page_to_dev_replace(sblock, page_num); |
| 1701 | if (ret) | 1701 | if (ret) |
| 1702 | btrfs_dev_replace_stats_inc( | 1702 | btrfs_dev_replace_stats_inc( |
| 1703 | &sblock->sctx->dev_root->fs_info->dev_replace. | 1703 | &sblock->sctx->dev_root->fs_info->dev_replace. |
| 1704 | num_write_errors); | 1704 | num_write_errors); |
| 1705 | } | 1705 | } |
| 1706 | } | 1706 | } |
| 1707 | 1707 | ||
| 1708 | static int scrub_write_page_to_dev_replace(struct scrub_block *sblock, | 1708 | static int scrub_write_page_to_dev_replace(struct scrub_block *sblock, |
| 1709 | int page_num) | 1709 | int page_num) |
| 1710 | { | 1710 | { |
| 1711 | struct scrub_page *spage = sblock->pagev[page_num]; | 1711 | struct scrub_page *spage = sblock->pagev[page_num]; |
| 1712 | 1712 | ||
| 1713 | BUG_ON(spage->page == NULL); | 1713 | BUG_ON(spage->page == NULL); |
| 1714 | if (spage->io_error) { | 1714 | if (spage->io_error) { |
| 1715 | void *mapped_buffer = kmap_atomic(spage->page); | 1715 | void *mapped_buffer = kmap_atomic(spage->page); |
| 1716 | 1716 | ||
| 1717 | memset(mapped_buffer, 0, PAGE_CACHE_SIZE); | 1717 | memset(mapped_buffer, 0, PAGE_CACHE_SIZE); |
| 1718 | flush_dcache_page(spage->page); | 1718 | flush_dcache_page(spage->page); |
| 1719 | kunmap_atomic(mapped_buffer); | 1719 | kunmap_atomic(mapped_buffer); |
| 1720 | } | 1720 | } |
| 1721 | return scrub_add_page_to_wr_bio(sblock->sctx, spage); | 1721 | return scrub_add_page_to_wr_bio(sblock->sctx, spage); |
| 1722 | } | 1722 | } |
| 1723 | 1723 | ||
| 1724 | static int scrub_add_page_to_wr_bio(struct scrub_ctx *sctx, | 1724 | static int scrub_add_page_to_wr_bio(struct scrub_ctx *sctx, |
| 1725 | struct scrub_page *spage) | 1725 | struct scrub_page *spage) |
| 1726 | { | 1726 | { |
| 1727 | struct scrub_wr_ctx *wr_ctx = &sctx->wr_ctx; | 1727 | struct scrub_wr_ctx *wr_ctx = &sctx->wr_ctx; |
| 1728 | struct scrub_bio *sbio; | 1728 | struct scrub_bio *sbio; |
| 1729 | int ret; | 1729 | int ret; |
| 1730 | 1730 | ||
| 1731 | mutex_lock(&wr_ctx->wr_lock); | 1731 | mutex_lock(&wr_ctx->wr_lock); |
| 1732 | again: | 1732 | again: |
| 1733 | if (!wr_ctx->wr_curr_bio) { | 1733 | if (!wr_ctx->wr_curr_bio) { |
| 1734 | wr_ctx->wr_curr_bio = kzalloc(sizeof(*wr_ctx->wr_curr_bio), | 1734 | wr_ctx->wr_curr_bio = kzalloc(sizeof(*wr_ctx->wr_curr_bio), |
| 1735 | GFP_NOFS); | 1735 | GFP_NOFS); |
| 1736 | if (!wr_ctx->wr_curr_bio) { | 1736 | if (!wr_ctx->wr_curr_bio) { |
| 1737 | mutex_unlock(&wr_ctx->wr_lock); | 1737 | mutex_unlock(&wr_ctx->wr_lock); |
| 1738 | return -ENOMEM; | 1738 | return -ENOMEM; |
| 1739 | } | 1739 | } |
| 1740 | wr_ctx->wr_curr_bio->sctx = sctx; | 1740 | wr_ctx->wr_curr_bio->sctx = sctx; |
| 1741 | wr_ctx->wr_curr_bio->page_count = 0; | 1741 | wr_ctx->wr_curr_bio->page_count = 0; |
| 1742 | } | 1742 | } |
| 1743 | sbio = wr_ctx->wr_curr_bio; | 1743 | sbio = wr_ctx->wr_curr_bio; |
| 1744 | if (sbio->page_count == 0) { | 1744 | if (sbio->page_count == 0) { |
| 1745 | struct bio *bio; | 1745 | struct bio *bio; |
| 1746 | 1746 | ||
| 1747 | sbio->physical = spage->physical_for_dev_replace; | 1747 | sbio->physical = spage->physical_for_dev_replace; |
| 1748 | sbio->logical = spage->logical; | 1748 | sbio->logical = spage->logical; |
| 1749 | sbio->dev = wr_ctx->tgtdev; | 1749 | sbio->dev = wr_ctx->tgtdev; |
| 1750 | bio = sbio->bio; | 1750 | bio = sbio->bio; |
| 1751 | if (!bio) { | 1751 | if (!bio) { |
| 1752 | bio = btrfs_io_bio_alloc(GFP_NOFS, wr_ctx->pages_per_wr_bio); | 1752 | bio = btrfs_io_bio_alloc(GFP_NOFS, wr_ctx->pages_per_wr_bio); |
| 1753 | if (!bio) { | 1753 | if (!bio) { |
| 1754 | mutex_unlock(&wr_ctx->wr_lock); | 1754 | mutex_unlock(&wr_ctx->wr_lock); |
| 1755 | return -ENOMEM; | 1755 | return -ENOMEM; |
| 1756 | } | 1756 | } |
| 1757 | sbio->bio = bio; | 1757 | sbio->bio = bio; |
| 1758 | } | 1758 | } |
| 1759 | 1759 | ||
| 1760 | bio->bi_private = sbio; | 1760 | bio->bi_private = sbio; |
| 1761 | bio->bi_end_io = scrub_wr_bio_end_io; | 1761 | bio->bi_end_io = scrub_wr_bio_end_io; |
| 1762 | bio->bi_bdev = sbio->dev->bdev; | 1762 | bio->bi_bdev = sbio->dev->bdev; |
| 1763 | bio->bi_iter.bi_sector = sbio->physical >> 9; | 1763 | bio->bi_iter.bi_sector = sbio->physical >> 9; |
| 1764 | sbio->err = 0; | 1764 | sbio->err = 0; |
| 1765 | } else if (sbio->physical + sbio->page_count * PAGE_SIZE != | 1765 | } else if (sbio->physical + sbio->page_count * PAGE_SIZE != |
| 1766 | spage->physical_for_dev_replace || | 1766 | spage->physical_for_dev_replace || |
| 1767 | sbio->logical + sbio->page_count * PAGE_SIZE != | 1767 | sbio->logical + sbio->page_count * PAGE_SIZE != |
| 1768 | spage->logical) { | 1768 | spage->logical) { |
| 1769 | scrub_wr_submit(sctx); | 1769 | scrub_wr_submit(sctx); |
| 1770 | goto again; | 1770 | goto again; |
| 1771 | } | 1771 | } |
| 1772 | 1772 | ||
| 1773 | ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0); | 1773 | ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0); |
| 1774 | if (ret != PAGE_SIZE) { | 1774 | if (ret != PAGE_SIZE) { |
| 1775 | if (sbio->page_count < 1) { | 1775 | if (sbio->page_count < 1) { |
| 1776 | bio_put(sbio->bio); | 1776 | bio_put(sbio->bio); |
| 1777 | sbio->bio = NULL; | 1777 | sbio->bio = NULL; |
| 1778 | mutex_unlock(&wr_ctx->wr_lock); | 1778 | mutex_unlock(&wr_ctx->wr_lock); |
| 1779 | return -EIO; | 1779 | return -EIO; |
| 1780 | } | 1780 | } |
| 1781 | scrub_wr_submit(sctx); | 1781 | scrub_wr_submit(sctx); |
| 1782 | goto again; | 1782 | goto again; |
| 1783 | } | 1783 | } |
| 1784 | 1784 | ||
| 1785 | sbio->pagev[sbio->page_count] = spage; | 1785 | sbio->pagev[sbio->page_count] = spage; |
| 1786 | scrub_page_get(spage); | 1786 | scrub_page_get(spage); |
| 1787 | sbio->page_count++; | 1787 | sbio->page_count++; |
| 1788 | if (sbio->page_count == wr_ctx->pages_per_wr_bio) | 1788 | if (sbio->page_count == wr_ctx->pages_per_wr_bio) |
| 1789 | scrub_wr_submit(sctx); | 1789 | scrub_wr_submit(sctx); |
| 1790 | mutex_unlock(&wr_ctx->wr_lock); | 1790 | mutex_unlock(&wr_ctx->wr_lock); |
| 1791 | 1791 | ||
| 1792 | return 0; | 1792 | return 0; |
| 1793 | } | 1793 | } |
| 1794 | 1794 | ||
| 1795 | static void scrub_wr_submit(struct scrub_ctx *sctx) | 1795 | static void scrub_wr_submit(struct scrub_ctx *sctx) |
| 1796 | { | 1796 | { |
| 1797 | struct scrub_wr_ctx *wr_ctx = &sctx->wr_ctx; | 1797 | struct scrub_wr_ctx *wr_ctx = &sctx->wr_ctx; |
| 1798 | struct scrub_bio *sbio; | 1798 | struct scrub_bio *sbio; |
| 1799 | 1799 | ||
| 1800 | if (!wr_ctx->wr_curr_bio) | 1800 | if (!wr_ctx->wr_curr_bio) |
| 1801 | return; | 1801 | return; |
| 1802 | 1802 | ||
| 1803 | sbio = wr_ctx->wr_curr_bio; | 1803 | sbio = wr_ctx->wr_curr_bio; |
| 1804 | wr_ctx->wr_curr_bio = NULL; | 1804 | wr_ctx->wr_curr_bio = NULL; |
| 1805 | WARN_ON(!sbio->bio->bi_bdev); | 1805 | WARN_ON(!sbio->bio->bi_bdev); |
| 1806 | scrub_pending_bio_inc(sctx); | 1806 | scrub_pending_bio_inc(sctx); |
| 1807 | /* process all writes in a single worker thread. Then the block layer | 1807 | /* process all writes in a single worker thread. Then the block layer |
| 1808 | * orders the requests before sending them to the driver which | 1808 | * orders the requests before sending them to the driver which |
| 1809 | * doubled the write performance on spinning disks when measured | 1809 | * doubled the write performance on spinning disks when measured |
| 1810 | * with Linux 3.5 */ | 1810 | * with Linux 3.5 */ |
| 1811 | btrfsic_submit_bio(WRITE, sbio->bio); | 1811 | btrfsic_submit_bio(WRITE, sbio->bio); |
| 1812 | } | 1812 | } |
| 1813 | 1813 | ||
| 1814 | static void scrub_wr_bio_end_io(struct bio *bio, int err) | 1814 | static void scrub_wr_bio_end_io(struct bio *bio, int err) |
| 1815 | { | 1815 | { |
| 1816 | struct scrub_bio *sbio = bio->bi_private; | 1816 | struct scrub_bio *sbio = bio->bi_private; |
| 1817 | struct btrfs_fs_info *fs_info = sbio->dev->dev_root->fs_info; | 1817 | struct btrfs_fs_info *fs_info = sbio->dev->dev_root->fs_info; |
| 1818 | 1818 | ||
| 1819 | sbio->err = err; | 1819 | sbio->err = err; |
| 1820 | sbio->bio = bio; | 1820 | sbio->bio = bio; |
| 1821 | 1821 | ||
| 1822 | btrfs_init_work(&sbio->work, btrfs_scrubwrc_helper, | 1822 | btrfs_init_work(&sbio->work, btrfs_scrubwrc_helper, |
| 1823 | scrub_wr_bio_end_io_worker, NULL, NULL); | 1823 | scrub_wr_bio_end_io_worker, NULL, NULL); |
| 1824 | btrfs_queue_work(fs_info->scrub_wr_completion_workers, &sbio->work); | 1824 | btrfs_queue_work(fs_info->scrub_wr_completion_workers, &sbio->work); |
| 1825 | } | 1825 | } |
| 1826 | 1826 | ||
| 1827 | static void scrub_wr_bio_end_io_worker(struct btrfs_work *work) | 1827 | static void scrub_wr_bio_end_io_worker(struct btrfs_work *work) |
| 1828 | { | 1828 | { |
| 1829 | struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); | 1829 | struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); |
| 1830 | struct scrub_ctx *sctx = sbio->sctx; | 1830 | struct scrub_ctx *sctx = sbio->sctx; |
| 1831 | int i; | 1831 | int i; |
| 1832 | 1832 | ||
| 1833 | WARN_ON(sbio->page_count > SCRUB_PAGES_PER_WR_BIO); | 1833 | WARN_ON(sbio->page_count > SCRUB_PAGES_PER_WR_BIO); |
| 1834 | if (sbio->err) { | 1834 | if (sbio->err) { |
| 1835 | struct btrfs_dev_replace *dev_replace = | 1835 | struct btrfs_dev_replace *dev_replace = |
| 1836 | &sbio->sctx->dev_root->fs_info->dev_replace; | 1836 | &sbio->sctx->dev_root->fs_info->dev_replace; |
| 1837 | 1837 | ||
| 1838 | for (i = 0; i < sbio->page_count; i++) { | 1838 | for (i = 0; i < sbio->page_count; i++) { |
| 1839 | struct scrub_page *spage = sbio->pagev[i]; | 1839 | struct scrub_page *spage = sbio->pagev[i]; |
| 1840 | 1840 | ||
| 1841 | spage->io_error = 1; | 1841 | spage->io_error = 1; |
| 1842 | btrfs_dev_replace_stats_inc(&dev_replace-> | 1842 | btrfs_dev_replace_stats_inc(&dev_replace-> |
| 1843 | num_write_errors); | 1843 | num_write_errors); |
| 1844 | } | 1844 | } |
| 1845 | } | 1845 | } |
| 1846 | 1846 | ||
| 1847 | for (i = 0; i < sbio->page_count; i++) | 1847 | for (i = 0; i < sbio->page_count; i++) |
| 1848 | scrub_page_put(sbio->pagev[i]); | 1848 | scrub_page_put(sbio->pagev[i]); |
| 1849 | 1849 | ||
| 1850 | bio_put(sbio->bio); | 1850 | bio_put(sbio->bio); |
| 1851 | kfree(sbio); | 1851 | kfree(sbio); |
| 1852 | scrub_pending_bio_dec(sctx); | 1852 | scrub_pending_bio_dec(sctx); |
| 1853 | } | 1853 | } |
| 1854 | 1854 | ||
| 1855 | static int scrub_checksum(struct scrub_block *sblock) | 1855 | static int scrub_checksum(struct scrub_block *sblock) |
| 1856 | { | 1856 | { |
| 1857 | u64 flags; | 1857 | u64 flags; |
| 1858 | int ret; | 1858 | int ret; |
| 1859 | 1859 | ||
| 1860 | WARN_ON(sblock->page_count < 1); | 1860 | WARN_ON(sblock->page_count < 1); |
| 1861 | flags = sblock->pagev[0]->flags; | 1861 | flags = sblock->pagev[0]->flags; |
| 1862 | ret = 0; | 1862 | ret = 0; |
| 1863 | if (flags & BTRFS_EXTENT_FLAG_DATA) | 1863 | if (flags & BTRFS_EXTENT_FLAG_DATA) |
| 1864 | ret = scrub_checksum_data(sblock); | 1864 | ret = scrub_checksum_data(sblock); |
| 1865 | else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) | 1865 | else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) |
| 1866 | ret = scrub_checksum_tree_block(sblock); | 1866 | ret = scrub_checksum_tree_block(sblock); |
| 1867 | else if (flags & BTRFS_EXTENT_FLAG_SUPER) | 1867 | else if (flags & BTRFS_EXTENT_FLAG_SUPER) |
| 1868 | (void)scrub_checksum_super(sblock); | 1868 | (void)scrub_checksum_super(sblock); |
| 1869 | else | 1869 | else |
| 1870 | WARN_ON(1); | 1870 | WARN_ON(1); |
| 1871 | if (ret) | 1871 | if (ret) |
| 1872 | scrub_handle_errored_block(sblock); | 1872 | scrub_handle_errored_block(sblock); |
| 1873 | 1873 | ||
| 1874 | return ret; | 1874 | return ret; |
| 1875 | } | 1875 | } |
| 1876 | 1876 | ||
| 1877 | static int scrub_checksum_data(struct scrub_block *sblock) | 1877 | static int scrub_checksum_data(struct scrub_block *sblock) |
| 1878 | { | 1878 | { |
| 1879 | struct scrub_ctx *sctx = sblock->sctx; | 1879 | struct scrub_ctx *sctx = sblock->sctx; |
| 1880 | u8 csum[BTRFS_CSUM_SIZE]; | 1880 | u8 csum[BTRFS_CSUM_SIZE]; |
| 1881 | u8 *on_disk_csum; | 1881 | u8 *on_disk_csum; |
| 1882 | struct page *page; | 1882 | struct page *page; |
| 1883 | void *buffer; | 1883 | void *buffer; |
| 1884 | u32 crc = ~(u32)0; | 1884 | u32 crc = ~(u32)0; |
| 1885 | int fail = 0; | 1885 | int fail = 0; |
| 1886 | u64 len; | 1886 | u64 len; |
| 1887 | int index; | 1887 | int index; |
| 1888 | 1888 | ||
| 1889 | BUG_ON(sblock->page_count < 1); | 1889 | BUG_ON(sblock->page_count < 1); |
| 1890 | if (!sblock->pagev[0]->have_csum) | 1890 | if (!sblock->pagev[0]->have_csum) |
| 1891 | return 0; | 1891 | return 0; |
| 1892 | 1892 | ||
| 1893 | on_disk_csum = sblock->pagev[0]->csum; | 1893 | on_disk_csum = sblock->pagev[0]->csum; |
| 1894 | page = sblock->pagev[0]->page; | 1894 | page = sblock->pagev[0]->page; |
| 1895 | buffer = kmap_atomic(page); | 1895 | buffer = kmap_atomic(page); |
| 1896 | 1896 | ||
| 1897 | len = sctx->sectorsize; | 1897 | len = sctx->sectorsize; |
| 1898 | index = 0; | 1898 | index = 0; |
| 1899 | for (;;) { | 1899 | for (;;) { |
| 1900 | u64 l = min_t(u64, len, PAGE_SIZE); | 1900 | u64 l = min_t(u64, len, PAGE_SIZE); |
| 1901 | 1901 | ||
| 1902 | crc = btrfs_csum_data(buffer, crc, l); | 1902 | crc = btrfs_csum_data(buffer, crc, l); |
| 1903 | kunmap_atomic(buffer); | 1903 | kunmap_atomic(buffer); |
| 1904 | len -= l; | 1904 | len -= l; |
| 1905 | if (len == 0) | 1905 | if (len == 0) |
| 1906 | break; | 1906 | break; |
| 1907 | index++; | 1907 | index++; |
| 1908 | BUG_ON(index >= sblock->page_count); | 1908 | BUG_ON(index >= sblock->page_count); |
| 1909 | BUG_ON(!sblock->pagev[index]->page); | 1909 | BUG_ON(!sblock->pagev[index]->page); |
| 1910 | page = sblock->pagev[index]->page; | 1910 | page = sblock->pagev[index]->page; |
| 1911 | buffer = kmap_atomic(page); | 1911 | buffer = kmap_atomic(page); |
| 1912 | } | 1912 | } |
| 1913 | 1913 | ||
| 1914 | btrfs_csum_final(crc, csum); | 1914 | btrfs_csum_final(crc, csum); |
| 1915 | if (memcmp(csum, on_disk_csum, sctx->csum_size)) | 1915 | if (memcmp(csum, on_disk_csum, sctx->csum_size)) |
| 1916 | fail = 1; | 1916 | fail = 1; |
| 1917 | 1917 | ||
| 1918 | return fail; | 1918 | return fail; |
| 1919 | } | 1919 | } |
| 1920 | 1920 | ||
| 1921 | static int scrub_checksum_tree_block(struct scrub_block *sblock) | 1921 | static int scrub_checksum_tree_block(struct scrub_block *sblock) |
| 1922 | { | 1922 | { |
| 1923 | struct scrub_ctx *sctx = sblock->sctx; | 1923 | struct scrub_ctx *sctx = sblock->sctx; |
| 1924 | struct btrfs_header *h; | 1924 | struct btrfs_header *h; |
| 1925 | struct btrfs_root *root = sctx->dev_root; | 1925 | struct btrfs_root *root = sctx->dev_root; |
| 1926 | struct btrfs_fs_info *fs_info = root->fs_info; | 1926 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 1927 | u8 calculated_csum[BTRFS_CSUM_SIZE]; | 1927 | u8 calculated_csum[BTRFS_CSUM_SIZE]; |
| 1928 | u8 on_disk_csum[BTRFS_CSUM_SIZE]; | 1928 | u8 on_disk_csum[BTRFS_CSUM_SIZE]; |
| 1929 | struct page *page; | 1929 | struct page *page; |
| 1930 | void *mapped_buffer; | 1930 | void *mapped_buffer; |
| 1931 | u64 mapped_size; | 1931 | u64 mapped_size; |
| 1932 | void *p; | 1932 | void *p; |
| 1933 | u32 crc = ~(u32)0; | 1933 | u32 crc = ~(u32)0; |
| 1934 | int fail = 0; | 1934 | int fail = 0; |
| 1935 | int crc_fail = 0; | 1935 | int crc_fail = 0; |
| 1936 | u64 len; | 1936 | u64 len; |
| 1937 | int index; | 1937 | int index; |
| 1938 | 1938 | ||
| 1939 | BUG_ON(sblock->page_count < 1); | 1939 | BUG_ON(sblock->page_count < 1); |
| 1940 | page = sblock->pagev[0]->page; | 1940 | page = sblock->pagev[0]->page; |
| 1941 | mapped_buffer = kmap_atomic(page); | 1941 | mapped_buffer = kmap_atomic(page); |
| 1942 | h = (struct btrfs_header *)mapped_buffer; | 1942 | h = (struct btrfs_header *)mapped_buffer; |
| 1943 | memcpy(on_disk_csum, h->csum, sctx->csum_size); | 1943 | memcpy(on_disk_csum, h->csum, sctx->csum_size); |
| 1944 | 1944 | ||
| 1945 | /* | 1945 | /* |
| 1946 | * we don't use the getter functions here, as we | 1946 | * we don't use the getter functions here, as we |
| 1947 | * a) don't have an extent buffer and | 1947 | * a) don't have an extent buffer and |
| 1948 | * b) the page is already kmapped | 1948 | * b) the page is already kmapped |
| 1949 | */ | 1949 | */ |
| 1950 | 1950 | ||
| 1951 | if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h)) | 1951 | if (sblock->pagev[0]->logical != btrfs_stack_header_bytenr(h)) |
| 1952 | ++fail; | 1952 | ++fail; |
| 1953 | 1953 | ||
| 1954 | if (sblock->pagev[0]->generation != btrfs_stack_header_generation(h)) | 1954 | if (sblock->pagev[0]->generation != btrfs_stack_header_generation(h)) |
| 1955 | ++fail; | 1955 | ++fail; |
| 1956 | 1956 | ||
| 1957 | if (!scrub_check_fsid(h->fsid, sblock->pagev[0])) | 1957 | if (!scrub_check_fsid(h->fsid, sblock->pagev[0])) |
| 1958 | ++fail; | 1958 | ++fail; |
| 1959 | 1959 | ||
| 1960 | if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, | 1960 | if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, |
| 1961 | BTRFS_UUID_SIZE)) | 1961 | BTRFS_UUID_SIZE)) |
| 1962 | ++fail; | 1962 | ++fail; |
| 1963 | 1963 | ||
| 1964 | len = sctx->nodesize - BTRFS_CSUM_SIZE; | 1964 | len = sctx->nodesize - BTRFS_CSUM_SIZE; |
| 1965 | mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE; | 1965 | mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE; |
| 1966 | p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE; | 1966 | p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE; |
| 1967 | index = 0; | 1967 | index = 0; |
| 1968 | for (;;) { | 1968 | for (;;) { |
| 1969 | u64 l = min_t(u64, len, mapped_size); | 1969 | u64 l = min_t(u64, len, mapped_size); |
| 1970 | 1970 | ||
| 1971 | crc = btrfs_csum_data(p, crc, l); | 1971 | crc = btrfs_csum_data(p, crc, l); |
| 1972 | kunmap_atomic(mapped_buffer); | 1972 | kunmap_atomic(mapped_buffer); |
| 1973 | len -= l; | 1973 | len -= l; |
| 1974 | if (len == 0) | 1974 | if (len == 0) |
| 1975 | break; | 1975 | break; |
| 1976 | index++; | 1976 | index++; |
| 1977 | BUG_ON(index >= sblock->page_count); | 1977 | BUG_ON(index >= sblock->page_count); |
| 1978 | BUG_ON(!sblock->pagev[index]->page); | 1978 | BUG_ON(!sblock->pagev[index]->page); |
| 1979 | page = sblock->pagev[index]->page; | 1979 | page = sblock->pagev[index]->page; |
| 1980 | mapped_buffer = kmap_atomic(page); | 1980 | mapped_buffer = kmap_atomic(page); |
| 1981 | mapped_size = PAGE_SIZE; | 1981 | mapped_size = PAGE_SIZE; |
| 1982 | p = mapped_buffer; | 1982 | p = mapped_buffer; |
| 1983 | } | 1983 | } |
| 1984 | 1984 | ||
| 1985 | btrfs_csum_final(crc, calculated_csum); | 1985 | btrfs_csum_final(crc, calculated_csum); |
| 1986 | if (memcmp(calculated_csum, on_disk_csum, sctx->csum_size)) | 1986 | if (memcmp(calculated_csum, on_disk_csum, sctx->csum_size)) |
| 1987 | ++crc_fail; | 1987 | ++crc_fail; |
| 1988 | 1988 | ||
| 1989 | return fail || crc_fail; | 1989 | return fail || crc_fail; |
| 1990 | } | 1990 | } |
| 1991 | 1991 | ||
| 1992 | static int scrub_checksum_super(struct scrub_block *sblock) | 1992 | static int scrub_checksum_super(struct scrub_block *sblock) |
| 1993 | { | 1993 | { |
| 1994 | struct btrfs_super_block *s; | 1994 | struct btrfs_super_block *s; |
| 1995 | struct scrub_ctx *sctx = sblock->sctx; | 1995 | struct scrub_ctx *sctx = sblock->sctx; |
| 1996 | u8 calculated_csum[BTRFS_CSUM_SIZE]; | 1996 | u8 calculated_csum[BTRFS_CSUM_SIZE]; |
| 1997 | u8 on_disk_csum[BTRFS_CSUM_SIZE]; | 1997 | u8 on_disk_csum[BTRFS_CSUM_SIZE]; |
| 1998 | struct page *page; | 1998 | struct page *page; |
| 1999 | void *mapped_buffer; | 1999 | void *mapped_buffer; |
| 2000 | u64 mapped_size; | 2000 | u64 mapped_size; |
| 2001 | void *p; | 2001 | void *p; |
| 2002 | u32 crc = ~(u32)0; | 2002 | u32 crc = ~(u32)0; |
| 2003 | int fail_gen = 0; | 2003 | int fail_gen = 0; |
| 2004 | int fail_cor = 0; | 2004 | int fail_cor = 0; |
| 2005 | u64 len; | 2005 | u64 len; |
| 2006 | int index; | 2006 | int index; |
| 2007 | 2007 | ||
| 2008 | BUG_ON(sblock->page_count < 1); | 2008 | BUG_ON(sblock->page_count < 1); |
| 2009 | page = sblock->pagev[0]->page; | 2009 | page = sblock->pagev[0]->page; |
| 2010 | mapped_buffer = kmap_atomic(page); | 2010 | mapped_buffer = kmap_atomic(page); |
| 2011 | s = (struct btrfs_super_block *)mapped_buffer; | 2011 | s = (struct btrfs_super_block *)mapped_buffer; |
| 2012 | memcpy(on_disk_csum, s->csum, sctx->csum_size); | 2012 | memcpy(on_disk_csum, s->csum, sctx->csum_size); |
| 2013 | 2013 | ||
| 2014 | if (sblock->pagev[0]->logical != btrfs_super_bytenr(s)) | 2014 | if (sblock->pagev[0]->logical != btrfs_super_bytenr(s)) |
| 2015 | ++fail_cor; | 2015 | ++fail_cor; |
| 2016 | 2016 | ||
| 2017 | if (sblock->pagev[0]->generation != btrfs_super_generation(s)) | 2017 | if (sblock->pagev[0]->generation != btrfs_super_generation(s)) |
| 2018 | ++fail_gen; | 2018 | ++fail_gen; |
| 2019 | 2019 | ||
| 2020 | if (!scrub_check_fsid(s->fsid, sblock->pagev[0])) | 2020 | if (!scrub_check_fsid(s->fsid, sblock->pagev[0])) |
| 2021 | ++fail_cor; | 2021 | ++fail_cor; |
| 2022 | 2022 | ||
| 2023 | len = BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE; | 2023 | len = BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE; |
| 2024 | mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE; | 2024 | mapped_size = PAGE_SIZE - BTRFS_CSUM_SIZE; |
| 2025 | p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE; | 2025 | p = ((u8 *)mapped_buffer) + BTRFS_CSUM_SIZE; |
| 2026 | index = 0; | 2026 | index = 0; |
| 2027 | for (;;) { | 2027 | for (;;) { |
| 2028 | u64 l = min_t(u64, len, mapped_size); | 2028 | u64 l = min_t(u64, len, mapped_size); |
| 2029 | 2029 | ||
| 2030 | crc = btrfs_csum_data(p, crc, l); | 2030 | crc = btrfs_csum_data(p, crc, l); |
| 2031 | kunmap_atomic(mapped_buffer); | 2031 | kunmap_atomic(mapped_buffer); |
| 2032 | len -= l; | 2032 | len -= l; |
| 2033 | if (len == 0) | 2033 | if (len == 0) |
| 2034 | break; | 2034 | break; |
| 2035 | index++; | 2035 | index++; |
| 2036 | BUG_ON(index >= sblock->page_count); | 2036 | BUG_ON(index >= sblock->page_count); |
| 2037 | BUG_ON(!sblock->pagev[index]->page); | 2037 | BUG_ON(!sblock->pagev[index]->page); |
| 2038 | page = sblock->pagev[index]->page; | 2038 | page = sblock->pagev[index]->page; |
| 2039 | mapped_buffer = kmap_atomic(page); | 2039 | mapped_buffer = kmap_atomic(page); |
| 2040 | mapped_size = PAGE_SIZE; | 2040 | mapped_size = PAGE_SIZE; |
| 2041 | p = mapped_buffer; | 2041 | p = mapped_buffer; |
| 2042 | } | 2042 | } |
| 2043 | 2043 | ||
| 2044 | btrfs_csum_final(crc, calculated_csum); | 2044 | btrfs_csum_final(crc, calculated_csum); |
| 2045 | if (memcmp(calculated_csum, on_disk_csum, sctx->csum_size)) | 2045 | if (memcmp(calculated_csum, on_disk_csum, sctx->csum_size)) |
| 2046 | ++fail_cor; | 2046 | ++fail_cor; |
| 2047 | 2047 | ||
| 2048 | if (fail_cor + fail_gen) { | 2048 | if (fail_cor + fail_gen) { |
| 2049 | /* | 2049 | /* |
| 2050 | * if we find an error in a super block, we just report it. | 2050 | * if we find an error in a super block, we just report it. |
| 2051 | * They will get written with the next transaction commit | 2051 | * They will get written with the next transaction commit |
| 2052 | * anyway | 2052 | * anyway |
| 2053 | */ | 2053 | */ |
| 2054 | spin_lock(&sctx->stat_lock); | 2054 | spin_lock(&sctx->stat_lock); |
| 2055 | ++sctx->stat.super_errors; | 2055 | ++sctx->stat.super_errors; |
| 2056 | spin_unlock(&sctx->stat_lock); | 2056 | spin_unlock(&sctx->stat_lock); |
| 2057 | if (fail_cor) | 2057 | if (fail_cor) |
| 2058 | btrfs_dev_stat_inc_and_print(sblock->pagev[0]->dev, | 2058 | btrfs_dev_stat_inc_and_print(sblock->pagev[0]->dev, |
| 2059 | BTRFS_DEV_STAT_CORRUPTION_ERRS); | 2059 | BTRFS_DEV_STAT_CORRUPTION_ERRS); |
| 2060 | else | 2060 | else |
| 2061 | btrfs_dev_stat_inc_and_print(sblock->pagev[0]->dev, | 2061 | btrfs_dev_stat_inc_and_print(sblock->pagev[0]->dev, |
| 2062 | BTRFS_DEV_STAT_GENERATION_ERRS); | 2062 | BTRFS_DEV_STAT_GENERATION_ERRS); |
| 2063 | } | 2063 | } |
| 2064 | 2064 | ||
| 2065 | return fail_cor + fail_gen; | 2065 | return fail_cor + fail_gen; |
| 2066 | } | 2066 | } |
| 2067 | 2067 | ||
| 2068 | static void scrub_block_get(struct scrub_block *sblock) | 2068 | static void scrub_block_get(struct scrub_block *sblock) |
| 2069 | { | 2069 | { |
| 2070 | atomic_inc(&sblock->ref_count); | 2070 | atomic_inc(&sblock->ref_count); |
| 2071 | } | 2071 | } |
| 2072 | 2072 | ||
| 2073 | static void scrub_block_put(struct scrub_block *sblock) | 2073 | static void scrub_block_put(struct scrub_block *sblock) |
| 2074 | { | 2074 | { |
| 2075 | if (atomic_dec_and_test(&sblock->ref_count)) { | 2075 | if (atomic_dec_and_test(&sblock->ref_count)) { |
| 2076 | int i; | 2076 | int i; |
| 2077 | 2077 | ||
| 2078 | if (sblock->sparity) | 2078 | if (sblock->sparity) |
| 2079 | scrub_parity_put(sblock->sparity); | 2079 | scrub_parity_put(sblock->sparity); |
| 2080 | 2080 | ||
| 2081 | for (i = 0; i < sblock->page_count; i++) | 2081 | for (i = 0; i < sblock->page_count; i++) |
| 2082 | scrub_page_put(sblock->pagev[i]); | 2082 | scrub_page_put(sblock->pagev[i]); |
| 2083 | kfree(sblock); | 2083 | kfree(sblock); |
| 2084 | } | 2084 | } |
| 2085 | } | 2085 | } |
| 2086 | 2086 | ||
| 2087 | static void scrub_page_get(struct scrub_page *spage) | 2087 | static void scrub_page_get(struct scrub_page *spage) |
| 2088 | { | 2088 | { |
| 2089 | atomic_inc(&spage->ref_count); | 2089 | atomic_inc(&spage->ref_count); |
| 2090 | } | 2090 | } |
| 2091 | 2091 | ||
| 2092 | static void scrub_page_put(struct scrub_page *spage) | 2092 | static void scrub_page_put(struct scrub_page *spage) |
| 2093 | { | 2093 | { |
| 2094 | if (atomic_dec_and_test(&spage->ref_count)) { | 2094 | if (atomic_dec_and_test(&spage->ref_count)) { |
| 2095 | if (spage->page) | 2095 | if (spage->page) |
| 2096 | __free_page(spage->page); | 2096 | __free_page(spage->page); |
| 2097 | kfree(spage); | 2097 | kfree(spage); |
| 2098 | } | 2098 | } |
| 2099 | } | 2099 | } |
| 2100 | 2100 | ||
| 2101 | static void scrub_submit(struct scrub_ctx *sctx) | 2101 | static void scrub_submit(struct scrub_ctx *sctx) |
| 2102 | { | 2102 | { |
| 2103 | struct scrub_bio *sbio; | 2103 | struct scrub_bio *sbio; |
| 2104 | 2104 | ||
| 2105 | if (sctx->curr == -1) | 2105 | if (sctx->curr == -1) |
| 2106 | return; | 2106 | return; |
| 2107 | 2107 | ||
| 2108 | sbio = sctx->bios[sctx->curr]; | 2108 | sbio = sctx->bios[sctx->curr]; |
| 2109 | sctx->curr = -1; | 2109 | sctx->curr = -1; |
| 2110 | scrub_pending_bio_inc(sctx); | 2110 | scrub_pending_bio_inc(sctx); |
| 2111 | 2111 | ||
| 2112 | if (!sbio->bio->bi_bdev) { | 2112 | if (!sbio->bio->bi_bdev) { |
| 2113 | /* | 2113 | /* |
| 2114 | * this case should not happen. If btrfs_map_block() is | 2114 | * this case should not happen. If btrfs_map_block() is |
| 2115 | * wrong, it could happen for dev-replace operations on | 2115 | * wrong, it could happen for dev-replace operations on |
| 2116 | * missing devices when no mirrors are available, but in | 2116 | * missing devices when no mirrors are available, but in |
| 2117 | * this case it should already fail the mount. | 2117 | * this case it should already fail the mount. |
| 2118 | * This case is handled correctly (but _very_ slowly). | 2118 | * This case is handled correctly (but _very_ slowly). |
| 2119 | */ | 2119 | */ |
| 2120 | printk_ratelimited(KERN_WARNING | 2120 | printk_ratelimited(KERN_WARNING |
| 2121 | "BTRFS: scrub_submit(bio bdev == NULL) is unexpected!\n"); | 2121 | "BTRFS: scrub_submit(bio bdev == NULL) is unexpected!\n"); |
| 2122 | bio_endio(sbio->bio, -EIO); | 2122 | bio_endio(sbio->bio, -EIO); |
| 2123 | } else { | 2123 | } else { |
| 2124 | btrfsic_submit_bio(READ, sbio->bio); | 2124 | btrfsic_submit_bio(READ, sbio->bio); |
| 2125 | } | 2125 | } |
| 2126 | } | 2126 | } |
| 2127 | 2127 | ||
| 2128 | static int scrub_add_page_to_rd_bio(struct scrub_ctx *sctx, | 2128 | static int scrub_add_page_to_rd_bio(struct scrub_ctx *sctx, |
| 2129 | struct scrub_page *spage) | 2129 | struct scrub_page *spage) |
| 2130 | { | 2130 | { |
| 2131 | struct scrub_block *sblock = spage->sblock; | 2131 | struct scrub_block *sblock = spage->sblock; |
| 2132 | struct scrub_bio *sbio; | 2132 | struct scrub_bio *sbio; |
| 2133 | int ret; | 2133 | int ret; |
| 2134 | 2134 | ||
| 2135 | again: | 2135 | again: |
| 2136 | /* | 2136 | /* |
| 2137 | * grab a fresh bio or wait for one to become available | 2137 | * grab a fresh bio or wait for one to become available |
| 2138 | */ | 2138 | */ |
| 2139 | while (sctx->curr == -1) { | 2139 | while (sctx->curr == -1) { |
| 2140 | spin_lock(&sctx->list_lock); | 2140 | spin_lock(&sctx->list_lock); |
| 2141 | sctx->curr = sctx->first_free; | 2141 | sctx->curr = sctx->first_free; |
| 2142 | if (sctx->curr != -1) { | 2142 | if (sctx->curr != -1) { |
| 2143 | sctx->first_free = sctx->bios[sctx->curr]->next_free; | 2143 | sctx->first_free = sctx->bios[sctx->curr]->next_free; |
| 2144 | sctx->bios[sctx->curr]->next_free = -1; | 2144 | sctx->bios[sctx->curr]->next_free = -1; |
| 2145 | sctx->bios[sctx->curr]->page_count = 0; | 2145 | sctx->bios[sctx->curr]->page_count = 0; |
| 2146 | spin_unlock(&sctx->list_lock); | 2146 | spin_unlock(&sctx->list_lock); |
| 2147 | } else { | 2147 | } else { |
| 2148 | spin_unlock(&sctx->list_lock); | 2148 | spin_unlock(&sctx->list_lock); |
| 2149 | wait_event(sctx->list_wait, sctx->first_free != -1); | 2149 | wait_event(sctx->list_wait, sctx->first_free != -1); |
| 2150 | } | 2150 | } |
| 2151 | } | 2151 | } |
| 2152 | sbio = sctx->bios[sctx->curr]; | 2152 | sbio = sctx->bios[sctx->curr]; |
| 2153 | if (sbio->page_count == 0) { | 2153 | if (sbio->page_count == 0) { |
| 2154 | struct bio *bio; | 2154 | struct bio *bio; |
| 2155 | 2155 | ||
| 2156 | sbio->physical = spage->physical; | 2156 | sbio->physical = spage->physical; |
| 2157 | sbio->logical = spage->logical; | 2157 | sbio->logical = spage->logical; |
| 2158 | sbio->dev = spage->dev; | 2158 | sbio->dev = spage->dev; |
| 2159 | bio = sbio->bio; | 2159 | bio = sbio->bio; |
| 2160 | if (!bio) { | 2160 | if (!bio) { |
| 2161 | bio = btrfs_io_bio_alloc(GFP_NOFS, sctx->pages_per_rd_bio); | 2161 | bio = btrfs_io_bio_alloc(GFP_NOFS, sctx->pages_per_rd_bio); |
| 2162 | if (!bio) | 2162 | if (!bio) |
| 2163 | return -ENOMEM; | 2163 | return -ENOMEM; |
| 2164 | sbio->bio = bio; | 2164 | sbio->bio = bio; |
| 2165 | } | 2165 | } |
| 2166 | 2166 | ||
| 2167 | bio->bi_private = sbio; | 2167 | bio->bi_private = sbio; |
| 2168 | bio->bi_end_io = scrub_bio_end_io; | 2168 | bio->bi_end_io = scrub_bio_end_io; |
| 2169 | bio->bi_bdev = sbio->dev->bdev; | 2169 | bio->bi_bdev = sbio->dev->bdev; |
| 2170 | bio->bi_iter.bi_sector = sbio->physical >> 9; | 2170 | bio->bi_iter.bi_sector = sbio->physical >> 9; |
| 2171 | sbio->err = 0; | 2171 | sbio->err = 0; |
| 2172 | } else if (sbio->physical + sbio->page_count * PAGE_SIZE != | 2172 | } else if (sbio->physical + sbio->page_count * PAGE_SIZE != |
| 2173 | spage->physical || | 2173 | spage->physical || |
| 2174 | sbio->logical + sbio->page_count * PAGE_SIZE != | 2174 | sbio->logical + sbio->page_count * PAGE_SIZE != |
| 2175 | spage->logical || | 2175 | spage->logical || |
| 2176 | sbio->dev != spage->dev) { | 2176 | sbio->dev != spage->dev) { |
| 2177 | scrub_submit(sctx); | 2177 | scrub_submit(sctx); |
| 2178 | goto again; | 2178 | goto again; |
| 2179 | } | 2179 | } |
| 2180 | 2180 | ||
| 2181 | sbio->pagev[sbio->page_count] = spage; | 2181 | sbio->pagev[sbio->page_count] = spage; |
| 2182 | ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0); | 2182 | ret = bio_add_page(sbio->bio, spage->page, PAGE_SIZE, 0); |
| 2183 | if (ret != PAGE_SIZE) { | 2183 | if (ret != PAGE_SIZE) { |
| 2184 | if (sbio->page_count < 1) { | 2184 | if (sbio->page_count < 1) { |
| 2185 | bio_put(sbio->bio); | 2185 | bio_put(sbio->bio); |
| 2186 | sbio->bio = NULL; | 2186 | sbio->bio = NULL; |
| 2187 | return -EIO; | 2187 | return -EIO; |
| 2188 | } | 2188 | } |
| 2189 | scrub_submit(sctx); | 2189 | scrub_submit(sctx); |
| 2190 | goto again; | 2190 | goto again; |
| 2191 | } | 2191 | } |
| 2192 | 2192 | ||
| 2193 | scrub_block_get(sblock); /* one for the page added to the bio */ | 2193 | scrub_block_get(sblock); /* one for the page added to the bio */ |
| 2194 | atomic_inc(&sblock->outstanding_pages); | 2194 | atomic_inc(&sblock->outstanding_pages); |
| 2195 | sbio->page_count++; | 2195 | sbio->page_count++; |
| 2196 | if (sbio->page_count == sctx->pages_per_rd_bio) | 2196 | if (sbio->page_count == sctx->pages_per_rd_bio) |
| 2197 | scrub_submit(sctx); | 2197 | scrub_submit(sctx); |
| 2198 | 2198 | ||
| 2199 | return 0; | 2199 | return 0; |
| 2200 | } | 2200 | } |
| 2201 | 2201 | ||
| 2202 | static int scrub_pages(struct scrub_ctx *sctx, u64 logical, u64 len, | 2202 | static int scrub_pages(struct scrub_ctx *sctx, u64 logical, u64 len, |
| 2203 | u64 physical, struct btrfs_device *dev, u64 flags, | 2203 | u64 physical, struct btrfs_device *dev, u64 flags, |
| 2204 | u64 gen, int mirror_num, u8 *csum, int force, | 2204 | u64 gen, int mirror_num, u8 *csum, int force, |
| 2205 | u64 physical_for_dev_replace) | 2205 | u64 physical_for_dev_replace) |
| 2206 | { | 2206 | { |
| 2207 | struct scrub_block *sblock; | 2207 | struct scrub_block *sblock; |
| 2208 | int index; | 2208 | int index; |
| 2209 | 2209 | ||
| 2210 | sblock = kzalloc(sizeof(*sblock), GFP_NOFS); | 2210 | sblock = kzalloc(sizeof(*sblock), GFP_NOFS); |
| 2211 | if (!sblock) { | 2211 | if (!sblock) { |
| 2212 | spin_lock(&sctx->stat_lock); | 2212 | spin_lock(&sctx->stat_lock); |
| 2213 | sctx->stat.malloc_errors++; | 2213 | sctx->stat.malloc_errors++; |
| 2214 | spin_unlock(&sctx->stat_lock); | 2214 | spin_unlock(&sctx->stat_lock); |
| 2215 | return -ENOMEM; | 2215 | return -ENOMEM; |
| 2216 | } | 2216 | } |
| 2217 | 2217 | ||
| 2218 | /* one ref inside this function, plus one for each page added to | 2218 | /* one ref inside this function, plus one for each page added to |
| 2219 | * a bio later on */ | 2219 | * a bio later on */ |
| 2220 | atomic_set(&sblock->ref_count, 1); | 2220 | atomic_set(&sblock->ref_count, 1); |
| 2221 | sblock->sctx = sctx; | 2221 | sblock->sctx = sctx; |
| 2222 | sblock->no_io_error_seen = 1; | 2222 | sblock->no_io_error_seen = 1; |
| 2223 | 2223 | ||
| 2224 | for (index = 0; len > 0; index++) { | 2224 | for (index = 0; len > 0; index++) { |
| 2225 | struct scrub_page *spage; | 2225 | struct scrub_page *spage; |
| 2226 | u64 l = min_t(u64, len, PAGE_SIZE); | 2226 | u64 l = min_t(u64, len, PAGE_SIZE); |
| 2227 | 2227 | ||
| 2228 | spage = kzalloc(sizeof(*spage), GFP_NOFS); | 2228 | spage = kzalloc(sizeof(*spage), GFP_NOFS); |
| 2229 | if (!spage) { | 2229 | if (!spage) { |
| 2230 | leave_nomem: | 2230 | leave_nomem: |
| 2231 | spin_lock(&sctx->stat_lock); | 2231 | spin_lock(&sctx->stat_lock); |
| 2232 | sctx->stat.malloc_errors++; | 2232 | sctx->stat.malloc_errors++; |
| 2233 | spin_unlock(&sctx->stat_lock); | 2233 | spin_unlock(&sctx->stat_lock); |
| 2234 | scrub_block_put(sblock); | 2234 | scrub_block_put(sblock); |
| 2235 | return -ENOMEM; | 2235 | return -ENOMEM; |
| 2236 | } | 2236 | } |
| 2237 | BUG_ON(index >= SCRUB_MAX_PAGES_PER_BLOCK); | 2237 | BUG_ON(index >= SCRUB_MAX_PAGES_PER_BLOCK); |
| 2238 | scrub_page_get(spage); | 2238 | scrub_page_get(spage); |
| 2239 | sblock->pagev[index] = spage; | 2239 | sblock->pagev[index] = spage; |
| 2240 | spage->sblock = sblock; | 2240 | spage->sblock = sblock; |
| 2241 | spage->dev = dev; | 2241 | spage->dev = dev; |
| 2242 | spage->flags = flags; | 2242 | spage->flags = flags; |
| 2243 | spage->generation = gen; | 2243 | spage->generation = gen; |
| 2244 | spage->logical = logical; | 2244 | spage->logical = logical; |
| 2245 | spage->physical = physical; | 2245 | spage->physical = physical; |
| 2246 | spage->physical_for_dev_replace = physical_for_dev_replace; | 2246 | spage->physical_for_dev_replace = physical_for_dev_replace; |
| 2247 | spage->mirror_num = mirror_num; | 2247 | spage->mirror_num = mirror_num; |
| 2248 | if (csum) { | 2248 | if (csum) { |
| 2249 | spage->have_csum = 1; | 2249 | spage->have_csum = 1; |
| 2250 | memcpy(spage->csum, csum, sctx->csum_size); | 2250 | memcpy(spage->csum, csum, sctx->csum_size); |
| 2251 | } else { | 2251 | } else { |
| 2252 | spage->have_csum = 0; | 2252 | spage->have_csum = 0; |
| 2253 | } | 2253 | } |
| 2254 | sblock->page_count++; | 2254 | sblock->page_count++; |
| 2255 | spage->page = alloc_page(GFP_NOFS); | 2255 | spage->page = alloc_page(GFP_NOFS); |
| 2256 | if (!spage->page) | 2256 | if (!spage->page) |
| 2257 | goto leave_nomem; | 2257 | goto leave_nomem; |
| 2258 | len -= l; | 2258 | len -= l; |
| 2259 | logical += l; | 2259 | logical += l; |
| 2260 | physical += l; | 2260 | physical += l; |
| 2261 | physical_for_dev_replace += l; | 2261 | physical_for_dev_replace += l; |
| 2262 | } | 2262 | } |
| 2263 | 2263 | ||
| 2264 | WARN_ON(sblock->page_count == 0); | 2264 | WARN_ON(sblock->page_count == 0); |
| 2265 | for (index = 0; index < sblock->page_count; index++) { | 2265 | for (index = 0; index < sblock->page_count; index++) { |
| 2266 | struct scrub_page *spage = sblock->pagev[index]; | 2266 | struct scrub_page *spage = sblock->pagev[index]; |
| 2267 | int ret; | 2267 | int ret; |
| 2268 | 2268 | ||
| 2269 | ret = scrub_add_page_to_rd_bio(sctx, spage); | 2269 | ret = scrub_add_page_to_rd_bio(sctx, spage); |
| 2270 | if (ret) { | 2270 | if (ret) { |
| 2271 | scrub_block_put(sblock); | 2271 | scrub_block_put(sblock); |
| 2272 | return ret; | 2272 | return ret; |
| 2273 | } | 2273 | } |
| 2274 | } | 2274 | } |
| 2275 | 2275 | ||
| 2276 | if (force) | 2276 | if (force) |
| 2277 | scrub_submit(sctx); | 2277 | scrub_submit(sctx); |
| 2278 | 2278 | ||
| 2279 | /* last one frees, either here or in bio completion for last page */ | 2279 | /* last one frees, either here or in bio completion for last page */ |
| 2280 | scrub_block_put(sblock); | 2280 | scrub_block_put(sblock); |
| 2281 | return 0; | 2281 | return 0; |
| 2282 | } | 2282 | } |
| 2283 | 2283 | ||
| 2284 | static void scrub_bio_end_io(struct bio *bio, int err) | 2284 | static void scrub_bio_end_io(struct bio *bio, int err) |
| 2285 | { | 2285 | { |
| 2286 | struct scrub_bio *sbio = bio->bi_private; | 2286 | struct scrub_bio *sbio = bio->bi_private; |
| 2287 | struct btrfs_fs_info *fs_info = sbio->dev->dev_root->fs_info; | 2287 | struct btrfs_fs_info *fs_info = sbio->dev->dev_root->fs_info; |
| 2288 | 2288 | ||
| 2289 | sbio->err = err; | 2289 | sbio->err = err; |
| 2290 | sbio->bio = bio; | 2290 | sbio->bio = bio; |
| 2291 | 2291 | ||
| 2292 | btrfs_queue_work(fs_info->scrub_workers, &sbio->work); | 2292 | btrfs_queue_work(fs_info->scrub_workers, &sbio->work); |
| 2293 | } | 2293 | } |
| 2294 | 2294 | ||
| 2295 | static void scrub_bio_end_io_worker(struct btrfs_work *work) | 2295 | static void scrub_bio_end_io_worker(struct btrfs_work *work) |
| 2296 | { | 2296 | { |
| 2297 | struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); | 2297 | struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); |
| 2298 | struct scrub_ctx *sctx = sbio->sctx; | 2298 | struct scrub_ctx *sctx = sbio->sctx; |
| 2299 | int i; | 2299 | int i; |
| 2300 | 2300 | ||
| 2301 | BUG_ON(sbio->page_count > SCRUB_PAGES_PER_RD_BIO); | 2301 | BUG_ON(sbio->page_count > SCRUB_PAGES_PER_RD_BIO); |
| 2302 | if (sbio->err) { | 2302 | if (sbio->err) { |
| 2303 | for (i = 0; i < sbio->page_count; i++) { | 2303 | for (i = 0; i < sbio->page_count; i++) { |
| 2304 | struct scrub_page *spage = sbio->pagev[i]; | 2304 | struct scrub_page *spage = sbio->pagev[i]; |
| 2305 | 2305 | ||
| 2306 | spage->io_error = 1; | 2306 | spage->io_error = 1; |
| 2307 | spage->sblock->no_io_error_seen = 0; | 2307 | spage->sblock->no_io_error_seen = 0; |
| 2308 | } | 2308 | } |
| 2309 | } | 2309 | } |
| 2310 | 2310 | ||
| 2311 | /* now complete the scrub_block items that have all pages completed */ | 2311 | /* now complete the scrub_block items that have all pages completed */ |
| 2312 | for (i = 0; i < sbio->page_count; i++) { | 2312 | for (i = 0; i < sbio->page_count; i++) { |
| 2313 | struct scrub_page *spage = sbio->pagev[i]; | 2313 | struct scrub_page *spage = sbio->pagev[i]; |
| 2314 | struct scrub_block *sblock = spage->sblock; | 2314 | struct scrub_block *sblock = spage->sblock; |
| 2315 | 2315 | ||
| 2316 | if (atomic_dec_and_test(&sblock->outstanding_pages)) | 2316 | if (atomic_dec_and_test(&sblock->outstanding_pages)) |
| 2317 | scrub_block_complete(sblock); | 2317 | scrub_block_complete(sblock); |
| 2318 | scrub_block_put(sblock); | 2318 | scrub_block_put(sblock); |
| 2319 | } | 2319 | } |
| 2320 | 2320 | ||
| 2321 | bio_put(sbio->bio); | 2321 | bio_put(sbio->bio); |
| 2322 | sbio->bio = NULL; | 2322 | sbio->bio = NULL; |
| 2323 | spin_lock(&sctx->list_lock); | 2323 | spin_lock(&sctx->list_lock); |
| 2324 | sbio->next_free = sctx->first_free; | 2324 | sbio->next_free = sctx->first_free; |
| 2325 | sctx->first_free = sbio->index; | 2325 | sctx->first_free = sbio->index; |
| 2326 | spin_unlock(&sctx->list_lock); | 2326 | spin_unlock(&sctx->list_lock); |
| 2327 | 2327 | ||
| 2328 | if (sctx->is_dev_replace && | 2328 | if (sctx->is_dev_replace && |
| 2329 | atomic_read(&sctx->wr_ctx.flush_all_writes)) { | 2329 | atomic_read(&sctx->wr_ctx.flush_all_writes)) { |
| 2330 | mutex_lock(&sctx->wr_ctx.wr_lock); | 2330 | mutex_lock(&sctx->wr_ctx.wr_lock); |
| 2331 | scrub_wr_submit(sctx); | 2331 | scrub_wr_submit(sctx); |
| 2332 | mutex_unlock(&sctx->wr_ctx.wr_lock); | 2332 | mutex_unlock(&sctx->wr_ctx.wr_lock); |
| 2333 | } | 2333 | } |
| 2334 | 2334 | ||
| 2335 | scrub_pending_bio_dec(sctx); | 2335 | scrub_pending_bio_dec(sctx); |
| 2336 | } | 2336 | } |
| 2337 | 2337 | ||
| 2338 | static inline void __scrub_mark_bitmap(struct scrub_parity *sparity, | 2338 | static inline void __scrub_mark_bitmap(struct scrub_parity *sparity, |
| 2339 | unsigned long *bitmap, | 2339 | unsigned long *bitmap, |
| 2340 | u64 start, u64 len) | 2340 | u64 start, u64 len) |
| 2341 | { | 2341 | { |
| 2342 | int offset; | 2342 | int offset; |
| 2343 | int nsectors; | 2343 | int nsectors; |
| 2344 | int sectorsize = sparity->sctx->dev_root->sectorsize; | 2344 | int sectorsize = sparity->sctx->dev_root->sectorsize; |
| 2345 | 2345 | ||
| 2346 | if (len >= sparity->stripe_len) { | 2346 | if (len >= sparity->stripe_len) { |
| 2347 | bitmap_set(bitmap, 0, sparity->nsectors); | 2347 | bitmap_set(bitmap, 0, sparity->nsectors); |
| 2348 | return; | 2348 | return; |
| 2349 | } | 2349 | } |
| 2350 | 2350 | ||
| 2351 | start -= sparity->logic_start; | 2351 | start -= sparity->logic_start; |
| 2352 | offset = (int)do_div(start, sparity->stripe_len); | 2352 | offset = (int)do_div(start, sparity->stripe_len); |
| 2353 | offset /= sectorsize; | 2353 | offset /= sectorsize; |
| 2354 | nsectors = (int)len / sectorsize; | 2354 | nsectors = (int)len / sectorsize; |
| 2355 | 2355 | ||
| 2356 | if (offset + nsectors <= sparity->nsectors) { | 2356 | if (offset + nsectors <= sparity->nsectors) { |
| 2357 | bitmap_set(bitmap, offset, nsectors); | 2357 | bitmap_set(bitmap, offset, nsectors); |
| 2358 | return; | 2358 | return; |
| 2359 | } | 2359 | } |
| 2360 | 2360 | ||
| 2361 | bitmap_set(bitmap, offset, sparity->nsectors - offset); | 2361 | bitmap_set(bitmap, offset, sparity->nsectors - offset); |
| 2362 | bitmap_set(bitmap, 0, nsectors - (sparity->nsectors - offset)); | 2362 | bitmap_set(bitmap, 0, nsectors - (sparity->nsectors - offset)); |
| 2363 | } | 2363 | } |
| 2364 | 2364 | ||
| 2365 | static inline void scrub_parity_mark_sectors_error(struct scrub_parity *sparity, | 2365 | static inline void scrub_parity_mark_sectors_error(struct scrub_parity *sparity, |
| 2366 | u64 start, u64 len) | 2366 | u64 start, u64 len) |
| 2367 | { | 2367 | { |
| 2368 | __scrub_mark_bitmap(sparity, sparity->ebitmap, start, len); | 2368 | __scrub_mark_bitmap(sparity, sparity->ebitmap, start, len); |
| 2369 | } | 2369 | } |
| 2370 | 2370 | ||
| 2371 | static inline void scrub_parity_mark_sectors_data(struct scrub_parity *sparity, | 2371 | static inline void scrub_parity_mark_sectors_data(struct scrub_parity *sparity, |
| 2372 | u64 start, u64 len) | 2372 | u64 start, u64 len) |
| 2373 | { | 2373 | { |
| 2374 | __scrub_mark_bitmap(sparity, sparity->dbitmap, start, len); | 2374 | __scrub_mark_bitmap(sparity, sparity->dbitmap, start, len); |
| 2375 | } | 2375 | } |
| 2376 | 2376 | ||
| 2377 | static void scrub_block_complete(struct scrub_block *sblock) | 2377 | static void scrub_block_complete(struct scrub_block *sblock) |
| 2378 | { | 2378 | { |
| 2379 | int corrupted = 0; | 2379 | int corrupted = 0; |
| 2380 | 2380 | ||
| 2381 | if (!sblock->no_io_error_seen) { | 2381 | if (!sblock->no_io_error_seen) { |
| 2382 | corrupted = 1; | 2382 | corrupted = 1; |
| 2383 | scrub_handle_errored_block(sblock); | 2383 | scrub_handle_errored_block(sblock); |
| 2384 | } else { | 2384 | } else { |
| 2385 | /* | 2385 | /* |
| 2386 | * if has checksum error, write via repair mechanism in | 2386 | * if has checksum error, write via repair mechanism in |
| 2387 | * dev replace case, otherwise write here in dev replace | 2387 | * dev replace case, otherwise write here in dev replace |
| 2388 | * case. | 2388 | * case. |
| 2389 | */ | 2389 | */ |
| 2390 | corrupted = scrub_checksum(sblock); | 2390 | corrupted = scrub_checksum(sblock); |
| 2391 | if (!corrupted && sblock->sctx->is_dev_replace) | 2391 | if (!corrupted && sblock->sctx->is_dev_replace) |
| 2392 | scrub_write_block_to_dev_replace(sblock); | 2392 | scrub_write_block_to_dev_replace(sblock); |
| 2393 | } | 2393 | } |
| 2394 | 2394 | ||
| 2395 | if (sblock->sparity && corrupted && !sblock->data_corrected) { | 2395 | if (sblock->sparity && corrupted && !sblock->data_corrected) { |
| 2396 | u64 start = sblock->pagev[0]->logical; | 2396 | u64 start = sblock->pagev[0]->logical; |
| 2397 | u64 end = sblock->pagev[sblock->page_count - 1]->logical + | 2397 | u64 end = sblock->pagev[sblock->page_count - 1]->logical + |
| 2398 | PAGE_SIZE; | 2398 | PAGE_SIZE; |
| 2399 | 2399 | ||
| 2400 | scrub_parity_mark_sectors_error(sblock->sparity, | 2400 | scrub_parity_mark_sectors_error(sblock->sparity, |
| 2401 | start, end - start); | 2401 | start, end - start); |
| 2402 | } | 2402 | } |
| 2403 | } | 2403 | } |
| 2404 | 2404 | ||
| 2405 | static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u64 len, | 2405 | static int scrub_find_csum(struct scrub_ctx *sctx, u64 logical, u64 len, |
| 2406 | u8 *csum) | 2406 | u8 *csum) |
| 2407 | { | 2407 | { |
| 2408 | struct btrfs_ordered_sum *sum = NULL; | 2408 | struct btrfs_ordered_sum *sum = NULL; |
| 2409 | unsigned long index; | 2409 | unsigned long index; |
| 2410 | unsigned long num_sectors; | 2410 | unsigned long num_sectors; |
| 2411 | 2411 | ||
| 2412 | while (!list_empty(&sctx->csum_list)) { | 2412 | while (!list_empty(&sctx->csum_list)) { |
| 2413 | sum = list_first_entry(&sctx->csum_list, | 2413 | sum = list_first_entry(&sctx->csum_list, |
| 2414 | struct btrfs_ordered_sum, list); | 2414 | struct btrfs_ordered_sum, list); |
| 2415 | if (sum->bytenr > logical) | 2415 | if (sum->bytenr > logical) |
| 2416 | return 0; | 2416 | return 0; |
| 2417 | if (sum->bytenr + sum->len > logical) | 2417 | if (sum->bytenr + sum->len > logical) |
| 2418 | break; | 2418 | break; |
| 2419 | 2419 | ||
| 2420 | ++sctx->stat.csum_discards; | 2420 | ++sctx->stat.csum_discards; |
| 2421 | list_del(&sum->list); | 2421 | list_del(&sum->list); |
| 2422 | kfree(sum); | 2422 | kfree(sum); |
| 2423 | sum = NULL; | 2423 | sum = NULL; |
| 2424 | } | 2424 | } |
| 2425 | if (!sum) | 2425 | if (!sum) |
| 2426 | return 0; | 2426 | return 0; |
| 2427 | 2427 | ||
| 2428 | index = ((u32)(logical - sum->bytenr)) / sctx->sectorsize; | 2428 | index = ((u32)(logical - sum->bytenr)) / sctx->sectorsize; |
| 2429 | num_sectors = sum->len / sctx->sectorsize; | 2429 | num_sectors = sum->len / sctx->sectorsize; |
| 2430 | memcpy(csum, sum->sums + index, sctx->csum_size); | 2430 | memcpy(csum, sum->sums + index, sctx->csum_size); |
| 2431 | if (index == num_sectors - 1) { | 2431 | if (index == num_sectors - 1) { |
| 2432 | list_del(&sum->list); | 2432 | list_del(&sum->list); |
| 2433 | kfree(sum); | 2433 | kfree(sum); |
| 2434 | } | 2434 | } |
| 2435 | return 1; | 2435 | return 1; |
| 2436 | } | 2436 | } |
| 2437 | 2437 | ||
| 2438 | /* scrub extent tries to collect up to 64 kB for each bio */ | 2438 | /* scrub extent tries to collect up to 64 kB for each bio */ |
| 2439 | static int scrub_extent(struct scrub_ctx *sctx, u64 logical, u64 len, | 2439 | static int scrub_extent(struct scrub_ctx *sctx, u64 logical, u64 len, |
| 2440 | u64 physical, struct btrfs_device *dev, u64 flags, | 2440 | u64 physical, struct btrfs_device *dev, u64 flags, |
| 2441 | u64 gen, int mirror_num, u64 physical_for_dev_replace) | 2441 | u64 gen, int mirror_num, u64 physical_for_dev_replace) |
| 2442 | { | 2442 | { |
| 2443 | int ret; | 2443 | int ret; |
| 2444 | u8 csum[BTRFS_CSUM_SIZE]; | 2444 | u8 csum[BTRFS_CSUM_SIZE]; |
| 2445 | u32 blocksize; | 2445 | u32 blocksize; |
| 2446 | 2446 | ||
| 2447 | if (flags & BTRFS_EXTENT_FLAG_DATA) { | 2447 | if (flags & BTRFS_EXTENT_FLAG_DATA) { |
| 2448 | blocksize = sctx->sectorsize; | 2448 | blocksize = sctx->sectorsize; |
| 2449 | spin_lock(&sctx->stat_lock); | 2449 | spin_lock(&sctx->stat_lock); |
| 2450 | sctx->stat.data_extents_scrubbed++; | 2450 | sctx->stat.data_extents_scrubbed++; |
| 2451 | sctx->stat.data_bytes_scrubbed += len; | 2451 | sctx->stat.data_bytes_scrubbed += len; |
| 2452 | spin_unlock(&sctx->stat_lock); | 2452 | spin_unlock(&sctx->stat_lock); |
| 2453 | } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { | 2453 | } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { |
| 2454 | blocksize = sctx->nodesize; | 2454 | blocksize = sctx->nodesize; |
| 2455 | spin_lock(&sctx->stat_lock); | 2455 | spin_lock(&sctx->stat_lock); |
| 2456 | sctx->stat.tree_extents_scrubbed++; | 2456 | sctx->stat.tree_extents_scrubbed++; |
| 2457 | sctx->stat.tree_bytes_scrubbed += len; | 2457 | sctx->stat.tree_bytes_scrubbed += len; |
| 2458 | spin_unlock(&sctx->stat_lock); | 2458 | spin_unlock(&sctx->stat_lock); |
| 2459 | } else { | 2459 | } else { |
| 2460 | blocksize = sctx->sectorsize; | 2460 | blocksize = sctx->sectorsize; |
| 2461 | WARN_ON(1); | 2461 | WARN_ON(1); |
| 2462 | } | 2462 | } |
| 2463 | 2463 | ||
| 2464 | while (len) { | 2464 | while (len) { |
| 2465 | u64 l = min_t(u64, len, blocksize); | 2465 | u64 l = min_t(u64, len, blocksize); |
| 2466 | int have_csum = 0; | 2466 | int have_csum = 0; |
| 2467 | 2467 | ||
| 2468 | if (flags & BTRFS_EXTENT_FLAG_DATA) { | 2468 | if (flags & BTRFS_EXTENT_FLAG_DATA) { |
| 2469 | /* push csums to sbio */ | 2469 | /* push csums to sbio */ |
| 2470 | have_csum = scrub_find_csum(sctx, logical, l, csum); | 2470 | have_csum = scrub_find_csum(sctx, logical, l, csum); |
| 2471 | if (have_csum == 0) | 2471 | if (have_csum == 0) |
| 2472 | ++sctx->stat.no_csum; | 2472 | ++sctx->stat.no_csum; |
| 2473 | if (sctx->is_dev_replace && !have_csum) { | 2473 | if (sctx->is_dev_replace && !have_csum) { |
| 2474 | ret = copy_nocow_pages(sctx, logical, l, | 2474 | ret = copy_nocow_pages(sctx, logical, l, |
| 2475 | mirror_num, | 2475 | mirror_num, |
| 2476 | physical_for_dev_replace); | 2476 | physical_for_dev_replace); |
| 2477 | goto behind_scrub_pages; | 2477 | goto behind_scrub_pages; |
| 2478 | } | 2478 | } |
| 2479 | } | 2479 | } |
| 2480 | ret = scrub_pages(sctx, logical, l, physical, dev, flags, gen, | 2480 | ret = scrub_pages(sctx, logical, l, physical, dev, flags, gen, |
| 2481 | mirror_num, have_csum ? csum : NULL, 0, | 2481 | mirror_num, have_csum ? csum : NULL, 0, |
| 2482 | physical_for_dev_replace); | 2482 | physical_for_dev_replace); |
| 2483 | behind_scrub_pages: | 2483 | behind_scrub_pages: |
| 2484 | if (ret) | 2484 | if (ret) |
| 2485 | return ret; | 2485 | return ret; |
| 2486 | len -= l; | 2486 | len -= l; |
| 2487 | logical += l; | 2487 | logical += l; |
| 2488 | physical += l; | 2488 | physical += l; |
| 2489 | physical_for_dev_replace += l; | 2489 | physical_for_dev_replace += l; |
| 2490 | } | 2490 | } |
| 2491 | return 0; | 2491 | return 0; |
| 2492 | } | 2492 | } |
| 2493 | 2493 | ||
| 2494 | static int scrub_pages_for_parity(struct scrub_parity *sparity, | 2494 | static int scrub_pages_for_parity(struct scrub_parity *sparity, |
| 2495 | u64 logical, u64 len, | 2495 | u64 logical, u64 len, |
| 2496 | u64 physical, struct btrfs_device *dev, | 2496 | u64 physical, struct btrfs_device *dev, |
| 2497 | u64 flags, u64 gen, int mirror_num, u8 *csum) | 2497 | u64 flags, u64 gen, int mirror_num, u8 *csum) |
| 2498 | { | 2498 | { |
| 2499 | struct scrub_ctx *sctx = sparity->sctx; | 2499 | struct scrub_ctx *sctx = sparity->sctx; |
| 2500 | struct scrub_block *sblock; | 2500 | struct scrub_block *sblock; |
| 2501 | int index; | 2501 | int index; |
| 2502 | 2502 | ||
| 2503 | sblock = kzalloc(sizeof(*sblock), GFP_NOFS); | 2503 | sblock = kzalloc(sizeof(*sblock), GFP_NOFS); |
| 2504 | if (!sblock) { | 2504 | if (!sblock) { |
| 2505 | spin_lock(&sctx->stat_lock); | 2505 | spin_lock(&sctx->stat_lock); |
| 2506 | sctx->stat.malloc_errors++; | 2506 | sctx->stat.malloc_errors++; |
| 2507 | spin_unlock(&sctx->stat_lock); | 2507 | spin_unlock(&sctx->stat_lock); |
| 2508 | return -ENOMEM; | 2508 | return -ENOMEM; |
| 2509 | } | 2509 | } |
| 2510 | 2510 | ||
| 2511 | /* one ref inside this function, plus one for each page added to | 2511 | /* one ref inside this function, plus one for each page added to |
| 2512 | * a bio later on */ | 2512 | * a bio later on */ |
| 2513 | atomic_set(&sblock->ref_count, 1); | 2513 | atomic_set(&sblock->ref_count, 1); |
| 2514 | sblock->sctx = sctx; | 2514 | sblock->sctx = sctx; |
| 2515 | sblock->no_io_error_seen = 1; | 2515 | sblock->no_io_error_seen = 1; |
| 2516 | sblock->sparity = sparity; | 2516 | sblock->sparity = sparity; |
| 2517 | scrub_parity_get(sparity); | 2517 | scrub_parity_get(sparity); |
| 2518 | 2518 | ||
| 2519 | for (index = 0; len > 0; index++) { | 2519 | for (index = 0; len > 0; index++) { |
| 2520 | struct scrub_page *spage; | 2520 | struct scrub_page *spage; |
| 2521 | u64 l = min_t(u64, len, PAGE_SIZE); | 2521 | u64 l = min_t(u64, len, PAGE_SIZE); |
| 2522 | 2522 | ||
| 2523 | spage = kzalloc(sizeof(*spage), GFP_NOFS); | 2523 | spage = kzalloc(sizeof(*spage), GFP_NOFS); |
| 2524 | if (!spage) { | 2524 | if (!spage) { |
| 2525 | leave_nomem: | 2525 | leave_nomem: |
| 2526 | spin_lock(&sctx->stat_lock); | 2526 | spin_lock(&sctx->stat_lock); |
| 2527 | sctx->stat.malloc_errors++; | 2527 | sctx->stat.malloc_errors++; |
| 2528 | spin_unlock(&sctx->stat_lock); | 2528 | spin_unlock(&sctx->stat_lock); |
| 2529 | scrub_block_put(sblock); | 2529 | scrub_block_put(sblock); |
| 2530 | return -ENOMEM; | 2530 | return -ENOMEM; |
| 2531 | } | 2531 | } |
| 2532 | BUG_ON(index >= SCRUB_MAX_PAGES_PER_BLOCK); | 2532 | BUG_ON(index >= SCRUB_MAX_PAGES_PER_BLOCK); |
| 2533 | /* For scrub block */ | 2533 | /* For scrub block */ |
| 2534 | scrub_page_get(spage); | 2534 | scrub_page_get(spage); |
| 2535 | sblock->pagev[index] = spage; | 2535 | sblock->pagev[index] = spage; |
| 2536 | /* For scrub parity */ | 2536 | /* For scrub parity */ |
| 2537 | scrub_page_get(spage); | 2537 | scrub_page_get(spage); |
| 2538 | list_add_tail(&spage->list, &sparity->spages); | 2538 | list_add_tail(&spage->list, &sparity->spages); |
| 2539 | spage->sblock = sblock; | 2539 | spage->sblock = sblock; |
| 2540 | spage->dev = dev; | 2540 | spage->dev = dev; |
| 2541 | spage->flags = flags; | 2541 | spage->flags = flags; |
| 2542 | spage->generation = gen; | 2542 | spage->generation = gen; |
| 2543 | spage->logical = logical; | 2543 | spage->logical = logical; |
| 2544 | spage->physical = physical; | 2544 | spage->physical = physical; |
| 2545 | spage->mirror_num = mirror_num; | 2545 | spage->mirror_num = mirror_num; |
| 2546 | if (csum) { | 2546 | if (csum) { |
| 2547 | spage->have_csum = 1; | 2547 | spage->have_csum = 1; |
| 2548 | memcpy(spage->csum, csum, sctx->csum_size); | 2548 | memcpy(spage->csum, csum, sctx->csum_size); |
| 2549 | } else { | 2549 | } else { |
| 2550 | spage->have_csum = 0; | 2550 | spage->have_csum = 0; |
| 2551 | } | 2551 | } |
| 2552 | sblock->page_count++; | 2552 | sblock->page_count++; |
| 2553 | spage->page = alloc_page(GFP_NOFS); | 2553 | spage->page = alloc_page(GFP_NOFS); |
| 2554 | if (!spage->page) | 2554 | if (!spage->page) |
| 2555 | goto leave_nomem; | 2555 | goto leave_nomem; |
| 2556 | len -= l; | 2556 | len -= l; |
| 2557 | logical += l; | 2557 | logical += l; |
| 2558 | physical += l; | 2558 | physical += l; |
| 2559 | } | 2559 | } |
| 2560 | 2560 | ||
| 2561 | WARN_ON(sblock->page_count == 0); | 2561 | WARN_ON(sblock->page_count == 0); |
| 2562 | for (index = 0; index < sblock->page_count; index++) { | 2562 | for (index = 0; index < sblock->page_count; index++) { |
| 2563 | struct scrub_page *spage = sblock->pagev[index]; | 2563 | struct scrub_page *spage = sblock->pagev[index]; |
| 2564 | int ret; | 2564 | int ret; |
| 2565 | 2565 | ||
| 2566 | ret = scrub_add_page_to_rd_bio(sctx, spage); | 2566 | ret = scrub_add_page_to_rd_bio(sctx, spage); |
| 2567 | if (ret) { | 2567 | if (ret) { |
| 2568 | scrub_block_put(sblock); | 2568 | scrub_block_put(sblock); |
| 2569 | return ret; | 2569 | return ret; |
| 2570 | } | 2570 | } |
| 2571 | } | 2571 | } |
| 2572 | 2572 | ||
| 2573 | /* last one frees, either here or in bio completion for last page */ | 2573 | /* last one frees, either here or in bio completion for last page */ |
| 2574 | scrub_block_put(sblock); | 2574 | scrub_block_put(sblock); |
| 2575 | return 0; | 2575 | return 0; |
| 2576 | } | 2576 | } |
| 2577 | 2577 | ||
| 2578 | static int scrub_extent_for_parity(struct scrub_parity *sparity, | 2578 | static int scrub_extent_for_parity(struct scrub_parity *sparity, |
| 2579 | u64 logical, u64 len, | 2579 | u64 logical, u64 len, |
| 2580 | u64 physical, struct btrfs_device *dev, | 2580 | u64 physical, struct btrfs_device *dev, |
| 2581 | u64 flags, u64 gen, int mirror_num) | 2581 | u64 flags, u64 gen, int mirror_num) |
| 2582 | { | 2582 | { |
| 2583 | struct scrub_ctx *sctx = sparity->sctx; | 2583 | struct scrub_ctx *sctx = sparity->sctx; |
| 2584 | int ret; | 2584 | int ret; |
| 2585 | u8 csum[BTRFS_CSUM_SIZE]; | 2585 | u8 csum[BTRFS_CSUM_SIZE]; |
| 2586 | u32 blocksize; | 2586 | u32 blocksize; |
| 2587 | 2587 | ||
| 2588 | if (flags & BTRFS_EXTENT_FLAG_DATA) { | 2588 | if (flags & BTRFS_EXTENT_FLAG_DATA) { |
| 2589 | blocksize = sctx->sectorsize; | 2589 | blocksize = sctx->sectorsize; |
| 2590 | } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { | 2590 | } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { |
| 2591 | blocksize = sctx->nodesize; | 2591 | blocksize = sctx->nodesize; |
| 2592 | } else { | 2592 | } else { |
| 2593 | blocksize = sctx->sectorsize; | 2593 | blocksize = sctx->sectorsize; |
| 2594 | WARN_ON(1); | 2594 | WARN_ON(1); |
| 2595 | } | 2595 | } |
| 2596 | 2596 | ||
| 2597 | while (len) { | 2597 | while (len) { |
| 2598 | u64 l = min_t(u64, len, blocksize); | 2598 | u64 l = min_t(u64, len, blocksize); |
| 2599 | int have_csum = 0; | 2599 | int have_csum = 0; |
| 2600 | 2600 | ||
| 2601 | if (flags & BTRFS_EXTENT_FLAG_DATA) { | 2601 | if (flags & BTRFS_EXTENT_FLAG_DATA) { |
| 2602 | /* push csums to sbio */ | 2602 | /* push csums to sbio */ |
| 2603 | have_csum = scrub_find_csum(sctx, logical, l, csum); | 2603 | have_csum = scrub_find_csum(sctx, logical, l, csum); |
| 2604 | if (have_csum == 0) | 2604 | if (have_csum == 0) |
| 2605 | goto skip; | 2605 | goto skip; |
| 2606 | } | 2606 | } |
| 2607 | ret = scrub_pages_for_parity(sparity, logical, l, physical, dev, | 2607 | ret = scrub_pages_for_parity(sparity, logical, l, physical, dev, |
| 2608 | flags, gen, mirror_num, | 2608 | flags, gen, mirror_num, |
| 2609 | have_csum ? csum : NULL); | 2609 | have_csum ? csum : NULL); |
| 2610 | skip: | ||
| 2611 | if (ret) | 2610 | if (ret) |
| 2612 | return ret; | 2611 | return ret; |
| 2612 | skip: | ||
| 2613 | len -= l; | 2613 | len -= l; |
| 2614 | logical += l; | 2614 | logical += l; |
| 2615 | physical += l; | 2615 | physical += l; |
| 2616 | } | 2616 | } |
| 2617 | return 0; | 2617 | return 0; |
| 2618 | } | 2618 | } |
| 2619 | 2619 | ||
| 2620 | /* | 2620 | /* |
| 2621 | * Given a physical address, this will calculate it's | 2621 | * Given a physical address, this will calculate it's |
| 2622 | * logical offset. if this is a parity stripe, it will return | 2622 | * logical offset. if this is a parity stripe, it will return |
| 2623 | * the most left data stripe's logical offset. | 2623 | * the most left data stripe's logical offset. |
| 2624 | * | 2624 | * |
| 2625 | * return 0 if it is a data stripe, 1 means parity stripe. | 2625 | * return 0 if it is a data stripe, 1 means parity stripe. |
| 2626 | */ | 2626 | */ |
| 2627 | static int get_raid56_logic_offset(u64 physical, int num, | 2627 | static int get_raid56_logic_offset(u64 physical, int num, |
| 2628 | struct map_lookup *map, u64 *offset, | 2628 | struct map_lookup *map, u64 *offset, |
| 2629 | u64 *stripe_start) | 2629 | u64 *stripe_start) |
| 2630 | { | 2630 | { |
| 2631 | int i; | 2631 | int i; |
| 2632 | int j = 0; | 2632 | int j = 0; |
| 2633 | u64 stripe_nr; | 2633 | u64 stripe_nr; |
| 2634 | u64 last_offset; | 2634 | u64 last_offset; |
| 2635 | int stripe_index; | 2635 | int stripe_index; |
| 2636 | int rot; | 2636 | int rot; |
| 2637 | 2637 | ||
| 2638 | last_offset = (physical - map->stripes[num].physical) * | 2638 | last_offset = (physical - map->stripes[num].physical) * |
| 2639 | nr_data_stripes(map); | 2639 | nr_data_stripes(map); |
| 2640 | if (stripe_start) | 2640 | if (stripe_start) |
| 2641 | *stripe_start = last_offset; | 2641 | *stripe_start = last_offset; |
| 2642 | 2642 | ||
| 2643 | *offset = last_offset; | 2643 | *offset = last_offset; |
| 2644 | for (i = 0; i < nr_data_stripes(map); i++) { | 2644 | for (i = 0; i < nr_data_stripes(map); i++) { |
| 2645 | *offset = last_offset + i * map->stripe_len; | 2645 | *offset = last_offset + i * map->stripe_len; |
| 2646 | 2646 | ||
| 2647 | stripe_nr = *offset; | 2647 | stripe_nr = *offset; |
| 2648 | do_div(stripe_nr, map->stripe_len); | 2648 | do_div(stripe_nr, map->stripe_len); |
| 2649 | do_div(stripe_nr, nr_data_stripes(map)); | 2649 | do_div(stripe_nr, nr_data_stripes(map)); |
| 2650 | 2650 | ||
| 2651 | /* Work out the disk rotation on this stripe-set */ | 2651 | /* Work out the disk rotation on this stripe-set */ |
| 2652 | rot = do_div(stripe_nr, map->num_stripes); | 2652 | rot = do_div(stripe_nr, map->num_stripes); |
| 2653 | /* calculate which stripe this data locates */ | 2653 | /* calculate which stripe this data locates */ |
| 2654 | rot += i; | 2654 | rot += i; |
| 2655 | stripe_index = rot % map->num_stripes; | 2655 | stripe_index = rot % map->num_stripes; |
| 2656 | if (stripe_index == num) | 2656 | if (stripe_index == num) |
| 2657 | return 0; | 2657 | return 0; |
| 2658 | if (stripe_index < num) | 2658 | if (stripe_index < num) |
| 2659 | j++; | 2659 | j++; |
| 2660 | } | 2660 | } |
| 2661 | *offset = last_offset + j * map->stripe_len; | 2661 | *offset = last_offset + j * map->stripe_len; |
| 2662 | return 1; | 2662 | return 1; |
| 2663 | } | 2663 | } |
| 2664 | 2664 | ||
| 2665 | static void scrub_free_parity(struct scrub_parity *sparity) | 2665 | static void scrub_free_parity(struct scrub_parity *sparity) |
| 2666 | { | 2666 | { |
| 2667 | struct scrub_ctx *sctx = sparity->sctx; | 2667 | struct scrub_ctx *sctx = sparity->sctx; |
| 2668 | struct scrub_page *curr, *next; | 2668 | struct scrub_page *curr, *next; |
| 2669 | int nbits; | 2669 | int nbits; |
| 2670 | 2670 | ||
| 2671 | nbits = bitmap_weight(sparity->ebitmap, sparity->nsectors); | 2671 | nbits = bitmap_weight(sparity->ebitmap, sparity->nsectors); |
| 2672 | if (nbits) { | 2672 | if (nbits) { |
| 2673 | spin_lock(&sctx->stat_lock); | 2673 | spin_lock(&sctx->stat_lock); |
| 2674 | sctx->stat.read_errors += nbits; | 2674 | sctx->stat.read_errors += nbits; |
| 2675 | sctx->stat.uncorrectable_errors += nbits; | 2675 | sctx->stat.uncorrectable_errors += nbits; |
| 2676 | spin_unlock(&sctx->stat_lock); | 2676 | spin_unlock(&sctx->stat_lock); |
| 2677 | } | 2677 | } |
| 2678 | 2678 | ||
| 2679 | list_for_each_entry_safe(curr, next, &sparity->spages, list) { | 2679 | list_for_each_entry_safe(curr, next, &sparity->spages, list) { |
| 2680 | list_del_init(&curr->list); | 2680 | list_del_init(&curr->list); |
| 2681 | scrub_page_put(curr); | 2681 | scrub_page_put(curr); |
| 2682 | } | 2682 | } |
| 2683 | 2683 | ||
| 2684 | kfree(sparity); | 2684 | kfree(sparity); |
| 2685 | } | 2685 | } |
| 2686 | 2686 | ||
| 2687 | static void scrub_parity_bio_endio(struct bio *bio, int error) | 2687 | static void scrub_parity_bio_endio(struct bio *bio, int error) |
| 2688 | { | 2688 | { |
| 2689 | struct scrub_parity *sparity = (struct scrub_parity *)bio->bi_private; | 2689 | struct scrub_parity *sparity = (struct scrub_parity *)bio->bi_private; |
| 2690 | struct scrub_ctx *sctx = sparity->sctx; | 2690 | struct scrub_ctx *sctx = sparity->sctx; |
| 2691 | 2691 | ||
| 2692 | if (error) | 2692 | if (error) |
| 2693 | bitmap_or(sparity->ebitmap, sparity->ebitmap, sparity->dbitmap, | 2693 | bitmap_or(sparity->ebitmap, sparity->ebitmap, sparity->dbitmap, |
| 2694 | sparity->nsectors); | 2694 | sparity->nsectors); |
| 2695 | 2695 | ||
| 2696 | scrub_free_parity(sparity); | 2696 | scrub_free_parity(sparity); |
| 2697 | scrub_pending_bio_dec(sctx); | 2697 | scrub_pending_bio_dec(sctx); |
| 2698 | bio_put(bio); | 2698 | bio_put(bio); |
| 2699 | } | 2699 | } |
| 2700 | 2700 | ||
| 2701 | static void scrub_parity_check_and_repair(struct scrub_parity *sparity) | 2701 | static void scrub_parity_check_and_repair(struct scrub_parity *sparity) |
| 2702 | { | 2702 | { |
| 2703 | struct scrub_ctx *sctx = sparity->sctx; | 2703 | struct scrub_ctx *sctx = sparity->sctx; |
| 2704 | struct bio *bio; | 2704 | struct bio *bio; |
| 2705 | struct btrfs_raid_bio *rbio; | 2705 | struct btrfs_raid_bio *rbio; |
| 2706 | struct scrub_page *spage; | 2706 | struct scrub_page *spage; |
| 2707 | struct btrfs_bio *bbio = NULL; | 2707 | struct btrfs_bio *bbio = NULL; |
| 2708 | u64 *raid_map = NULL; | 2708 | u64 *raid_map = NULL; |
| 2709 | u64 length; | 2709 | u64 length; |
| 2710 | int ret; | 2710 | int ret; |
| 2711 | 2711 | ||
| 2712 | if (!bitmap_andnot(sparity->dbitmap, sparity->dbitmap, sparity->ebitmap, | 2712 | if (!bitmap_andnot(sparity->dbitmap, sparity->dbitmap, sparity->ebitmap, |
| 2713 | sparity->nsectors)) | 2713 | sparity->nsectors)) |
| 2714 | goto out; | 2714 | goto out; |
| 2715 | 2715 | ||
| 2716 | length = sparity->logic_end - sparity->logic_start + 1; | 2716 | length = sparity->logic_end - sparity->logic_start + 1; |
| 2717 | ret = btrfs_map_sblock(sctx->dev_root->fs_info, WRITE, | 2717 | ret = btrfs_map_sblock(sctx->dev_root->fs_info, WRITE, |
| 2718 | sparity->logic_start, | 2718 | sparity->logic_start, |
| 2719 | &length, &bbio, 0, &raid_map); | 2719 | &length, &bbio, 0, &raid_map); |
| 2720 | if (ret || !bbio || !raid_map) | 2720 | if (ret || !bbio || !raid_map) |
| 2721 | goto bbio_out; | 2721 | goto bbio_out; |
| 2722 | 2722 | ||
| 2723 | bio = btrfs_io_bio_alloc(GFP_NOFS, 0); | 2723 | bio = btrfs_io_bio_alloc(GFP_NOFS, 0); |
| 2724 | if (!bio) | 2724 | if (!bio) |
| 2725 | goto bbio_out; | 2725 | goto bbio_out; |
| 2726 | 2726 | ||
| 2727 | bio->bi_iter.bi_sector = sparity->logic_start >> 9; | 2727 | bio->bi_iter.bi_sector = sparity->logic_start >> 9; |
| 2728 | bio->bi_private = sparity; | 2728 | bio->bi_private = sparity; |
| 2729 | bio->bi_end_io = scrub_parity_bio_endio; | 2729 | bio->bi_end_io = scrub_parity_bio_endio; |
| 2730 | 2730 | ||
| 2731 | rbio = raid56_parity_alloc_scrub_rbio(sctx->dev_root, bio, bbio, | 2731 | rbio = raid56_parity_alloc_scrub_rbio(sctx->dev_root, bio, bbio, |
| 2732 | raid_map, length, | 2732 | raid_map, length, |
| 2733 | sparity->scrub_dev, | 2733 | sparity->scrub_dev, |
| 2734 | sparity->dbitmap, | 2734 | sparity->dbitmap, |
| 2735 | sparity->nsectors); | 2735 | sparity->nsectors); |
| 2736 | if (!rbio) | 2736 | if (!rbio) |
| 2737 | goto rbio_out; | 2737 | goto rbio_out; |
| 2738 | 2738 | ||
| 2739 | list_for_each_entry(spage, &sparity->spages, list) | 2739 | list_for_each_entry(spage, &sparity->spages, list) |
| 2740 | raid56_parity_add_scrub_pages(rbio, spage->page, | 2740 | raid56_parity_add_scrub_pages(rbio, spage->page, |
| 2741 | spage->logical); | 2741 | spage->logical); |
| 2742 | 2742 | ||
| 2743 | scrub_pending_bio_inc(sctx); | 2743 | scrub_pending_bio_inc(sctx); |
| 2744 | raid56_parity_submit_scrub_rbio(rbio); | 2744 | raid56_parity_submit_scrub_rbio(rbio); |
| 2745 | return; | 2745 | return; |
| 2746 | 2746 | ||
| 2747 | rbio_out: | 2747 | rbio_out: |
| 2748 | bio_put(bio); | 2748 | bio_put(bio); |
| 2749 | bbio_out: | 2749 | bbio_out: |
| 2750 | kfree(bbio); | 2750 | kfree(bbio); |
| 2751 | kfree(raid_map); | 2751 | kfree(raid_map); |
| 2752 | bitmap_or(sparity->ebitmap, sparity->ebitmap, sparity->dbitmap, | 2752 | bitmap_or(sparity->ebitmap, sparity->ebitmap, sparity->dbitmap, |
| 2753 | sparity->nsectors); | 2753 | sparity->nsectors); |
| 2754 | spin_lock(&sctx->stat_lock); | 2754 | spin_lock(&sctx->stat_lock); |
| 2755 | sctx->stat.malloc_errors++; | 2755 | sctx->stat.malloc_errors++; |
| 2756 | spin_unlock(&sctx->stat_lock); | 2756 | spin_unlock(&sctx->stat_lock); |
| 2757 | out: | 2757 | out: |
| 2758 | scrub_free_parity(sparity); | 2758 | scrub_free_parity(sparity); |
| 2759 | } | 2759 | } |
| 2760 | 2760 | ||
| 2761 | static inline int scrub_calc_parity_bitmap_len(int nsectors) | 2761 | static inline int scrub_calc_parity_bitmap_len(int nsectors) |
| 2762 | { | 2762 | { |
| 2763 | return DIV_ROUND_UP(nsectors, BITS_PER_LONG) * (BITS_PER_LONG / 8); | 2763 | return DIV_ROUND_UP(nsectors, BITS_PER_LONG) * (BITS_PER_LONG / 8); |
| 2764 | } | 2764 | } |
| 2765 | 2765 | ||
| 2766 | static void scrub_parity_get(struct scrub_parity *sparity) | 2766 | static void scrub_parity_get(struct scrub_parity *sparity) |
| 2767 | { | 2767 | { |
| 2768 | atomic_inc(&sparity->ref_count); | 2768 | atomic_inc(&sparity->ref_count); |
| 2769 | } | 2769 | } |
| 2770 | 2770 | ||
| 2771 | static void scrub_parity_put(struct scrub_parity *sparity) | 2771 | static void scrub_parity_put(struct scrub_parity *sparity) |
| 2772 | { | 2772 | { |
| 2773 | if (!atomic_dec_and_test(&sparity->ref_count)) | 2773 | if (!atomic_dec_and_test(&sparity->ref_count)) |
| 2774 | return; | 2774 | return; |
| 2775 | 2775 | ||
| 2776 | scrub_parity_check_and_repair(sparity); | 2776 | scrub_parity_check_and_repair(sparity); |
| 2777 | } | 2777 | } |
| 2778 | 2778 | ||
| 2779 | static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx, | 2779 | static noinline_for_stack int scrub_raid56_parity(struct scrub_ctx *sctx, |
| 2780 | struct map_lookup *map, | 2780 | struct map_lookup *map, |
| 2781 | struct btrfs_device *sdev, | 2781 | struct btrfs_device *sdev, |
| 2782 | struct btrfs_path *path, | 2782 | struct btrfs_path *path, |
| 2783 | u64 logic_start, | 2783 | u64 logic_start, |
| 2784 | u64 logic_end) | 2784 | u64 logic_end) |
| 2785 | { | 2785 | { |
| 2786 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; | 2786 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; |
| 2787 | struct btrfs_root *root = fs_info->extent_root; | 2787 | struct btrfs_root *root = fs_info->extent_root; |
| 2788 | struct btrfs_root *csum_root = fs_info->csum_root; | 2788 | struct btrfs_root *csum_root = fs_info->csum_root; |
| 2789 | struct btrfs_extent_item *extent; | 2789 | struct btrfs_extent_item *extent; |
| 2790 | u64 flags; | 2790 | u64 flags; |
| 2791 | int ret; | 2791 | int ret; |
| 2792 | int slot; | 2792 | int slot; |
| 2793 | struct extent_buffer *l; | 2793 | struct extent_buffer *l; |
| 2794 | struct btrfs_key key; | 2794 | struct btrfs_key key; |
| 2795 | u64 generation; | 2795 | u64 generation; |
| 2796 | u64 extent_logical; | 2796 | u64 extent_logical; |
| 2797 | u64 extent_physical; | 2797 | u64 extent_physical; |
| 2798 | u64 extent_len; | 2798 | u64 extent_len; |
| 2799 | struct btrfs_device *extent_dev; | 2799 | struct btrfs_device *extent_dev; |
| 2800 | struct scrub_parity *sparity; | 2800 | struct scrub_parity *sparity; |
| 2801 | int nsectors; | 2801 | int nsectors; |
| 2802 | int bitmap_len; | 2802 | int bitmap_len; |
| 2803 | int extent_mirror_num; | 2803 | int extent_mirror_num; |
| 2804 | int stop_loop = 0; | 2804 | int stop_loop = 0; |
| 2805 | 2805 | ||
| 2806 | nsectors = map->stripe_len / root->sectorsize; | 2806 | nsectors = map->stripe_len / root->sectorsize; |
| 2807 | bitmap_len = scrub_calc_parity_bitmap_len(nsectors); | 2807 | bitmap_len = scrub_calc_parity_bitmap_len(nsectors); |
| 2808 | sparity = kzalloc(sizeof(struct scrub_parity) + 2 * bitmap_len, | 2808 | sparity = kzalloc(sizeof(struct scrub_parity) + 2 * bitmap_len, |
| 2809 | GFP_NOFS); | 2809 | GFP_NOFS); |
| 2810 | if (!sparity) { | 2810 | if (!sparity) { |
| 2811 | spin_lock(&sctx->stat_lock); | 2811 | spin_lock(&sctx->stat_lock); |
| 2812 | sctx->stat.malloc_errors++; | 2812 | sctx->stat.malloc_errors++; |
| 2813 | spin_unlock(&sctx->stat_lock); | 2813 | spin_unlock(&sctx->stat_lock); |
| 2814 | return -ENOMEM; | 2814 | return -ENOMEM; |
| 2815 | } | 2815 | } |
| 2816 | 2816 | ||
| 2817 | sparity->stripe_len = map->stripe_len; | 2817 | sparity->stripe_len = map->stripe_len; |
| 2818 | sparity->nsectors = nsectors; | 2818 | sparity->nsectors = nsectors; |
| 2819 | sparity->sctx = sctx; | 2819 | sparity->sctx = sctx; |
| 2820 | sparity->scrub_dev = sdev; | 2820 | sparity->scrub_dev = sdev; |
| 2821 | sparity->logic_start = logic_start; | 2821 | sparity->logic_start = logic_start; |
| 2822 | sparity->logic_end = logic_end; | 2822 | sparity->logic_end = logic_end; |
| 2823 | atomic_set(&sparity->ref_count, 1); | 2823 | atomic_set(&sparity->ref_count, 1); |
| 2824 | INIT_LIST_HEAD(&sparity->spages); | 2824 | INIT_LIST_HEAD(&sparity->spages); |
| 2825 | sparity->dbitmap = sparity->bitmap; | 2825 | sparity->dbitmap = sparity->bitmap; |
| 2826 | sparity->ebitmap = (void *)sparity->bitmap + bitmap_len; | 2826 | sparity->ebitmap = (void *)sparity->bitmap + bitmap_len; |
| 2827 | 2827 | ||
| 2828 | ret = 0; | 2828 | ret = 0; |
| 2829 | while (logic_start < logic_end) { | 2829 | while (logic_start < logic_end) { |
| 2830 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) | 2830 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) |
| 2831 | key.type = BTRFS_METADATA_ITEM_KEY; | 2831 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 2832 | else | 2832 | else |
| 2833 | key.type = BTRFS_EXTENT_ITEM_KEY; | 2833 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 2834 | key.objectid = logic_start; | 2834 | key.objectid = logic_start; |
| 2835 | key.offset = (u64)-1; | 2835 | key.offset = (u64)-1; |
| 2836 | 2836 | ||
| 2837 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 2837 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 2838 | if (ret < 0) | 2838 | if (ret < 0) |
| 2839 | goto out; | 2839 | goto out; |
| 2840 | 2840 | ||
| 2841 | if (ret > 0) { | 2841 | if (ret > 0) { |
| 2842 | ret = btrfs_previous_extent_item(root, path, 0); | 2842 | ret = btrfs_previous_extent_item(root, path, 0); |
| 2843 | if (ret < 0) | 2843 | if (ret < 0) |
| 2844 | goto out; | 2844 | goto out; |
| 2845 | if (ret > 0) { | 2845 | if (ret > 0) { |
| 2846 | btrfs_release_path(path); | 2846 | btrfs_release_path(path); |
| 2847 | ret = btrfs_search_slot(NULL, root, &key, | 2847 | ret = btrfs_search_slot(NULL, root, &key, |
| 2848 | path, 0, 0); | 2848 | path, 0, 0); |
| 2849 | if (ret < 0) | 2849 | if (ret < 0) |
| 2850 | goto out; | 2850 | goto out; |
| 2851 | } | 2851 | } |
| 2852 | } | 2852 | } |
| 2853 | 2853 | ||
| 2854 | stop_loop = 0; | 2854 | stop_loop = 0; |
| 2855 | while (1) { | 2855 | while (1) { |
| 2856 | u64 bytes; | 2856 | u64 bytes; |
| 2857 | 2857 | ||
| 2858 | l = path->nodes[0]; | 2858 | l = path->nodes[0]; |
| 2859 | slot = path->slots[0]; | 2859 | slot = path->slots[0]; |
| 2860 | if (slot >= btrfs_header_nritems(l)) { | 2860 | if (slot >= btrfs_header_nritems(l)) { |
| 2861 | ret = btrfs_next_leaf(root, path); | 2861 | ret = btrfs_next_leaf(root, path); |
| 2862 | if (ret == 0) | 2862 | if (ret == 0) |
| 2863 | continue; | 2863 | continue; |
| 2864 | if (ret < 0) | 2864 | if (ret < 0) |
| 2865 | goto out; | 2865 | goto out; |
| 2866 | 2866 | ||
| 2867 | stop_loop = 1; | 2867 | stop_loop = 1; |
| 2868 | break; | 2868 | break; |
| 2869 | } | 2869 | } |
| 2870 | btrfs_item_key_to_cpu(l, &key, slot); | 2870 | btrfs_item_key_to_cpu(l, &key, slot); |
| 2871 | 2871 | ||
| 2872 | if (key.type == BTRFS_METADATA_ITEM_KEY) | 2872 | if (key.type == BTRFS_METADATA_ITEM_KEY) |
| 2873 | bytes = root->nodesize; | 2873 | bytes = root->nodesize; |
| 2874 | else | 2874 | else |
| 2875 | bytes = key.offset; | 2875 | bytes = key.offset; |
| 2876 | 2876 | ||
| 2877 | if (key.objectid + bytes <= logic_start) | 2877 | if (key.objectid + bytes <= logic_start) |
| 2878 | goto next; | 2878 | goto next; |
| 2879 | 2879 | ||
| 2880 | if (key.type != BTRFS_EXTENT_ITEM_KEY && | 2880 | if (key.type != BTRFS_EXTENT_ITEM_KEY && |
| 2881 | key.type != BTRFS_METADATA_ITEM_KEY) | 2881 | key.type != BTRFS_METADATA_ITEM_KEY) |
| 2882 | goto next; | 2882 | goto next; |
| 2883 | 2883 | ||
| 2884 | if (key.objectid > logic_end) { | 2884 | if (key.objectid > logic_end) { |
| 2885 | stop_loop = 1; | 2885 | stop_loop = 1; |
| 2886 | break; | 2886 | break; |
| 2887 | } | 2887 | } |
| 2888 | 2888 | ||
| 2889 | while (key.objectid >= logic_start + map->stripe_len) | 2889 | while (key.objectid >= logic_start + map->stripe_len) |
| 2890 | logic_start += map->stripe_len; | 2890 | logic_start += map->stripe_len; |
| 2891 | 2891 | ||
| 2892 | extent = btrfs_item_ptr(l, slot, | 2892 | extent = btrfs_item_ptr(l, slot, |
| 2893 | struct btrfs_extent_item); | 2893 | struct btrfs_extent_item); |
| 2894 | flags = btrfs_extent_flags(l, extent); | 2894 | flags = btrfs_extent_flags(l, extent); |
| 2895 | generation = btrfs_extent_generation(l, extent); | 2895 | generation = btrfs_extent_generation(l, extent); |
| 2896 | 2896 | ||
| 2897 | if (key.objectid < logic_start && | 2897 | if (key.objectid < logic_start && |
| 2898 | (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) { | 2898 | (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) { |
| 2899 | btrfs_err(fs_info, | 2899 | btrfs_err(fs_info, |
| 2900 | "scrub: tree block %llu spanning stripes, ignored. logical=%llu", | 2900 | "scrub: tree block %llu spanning stripes, ignored. logical=%llu", |
| 2901 | key.objectid, logic_start); | 2901 | key.objectid, logic_start); |
| 2902 | goto next; | 2902 | goto next; |
| 2903 | } | 2903 | } |
| 2904 | again: | 2904 | again: |
| 2905 | extent_logical = key.objectid; | 2905 | extent_logical = key.objectid; |
| 2906 | extent_len = bytes; | 2906 | extent_len = bytes; |
| 2907 | 2907 | ||
| 2908 | if (extent_logical < logic_start) { | 2908 | if (extent_logical < logic_start) { |
| 2909 | extent_len -= logic_start - extent_logical; | 2909 | extent_len -= logic_start - extent_logical; |
| 2910 | extent_logical = logic_start; | 2910 | extent_logical = logic_start; |
| 2911 | } | 2911 | } |
| 2912 | 2912 | ||
| 2913 | if (extent_logical + extent_len > | 2913 | if (extent_logical + extent_len > |
| 2914 | logic_start + map->stripe_len) | 2914 | logic_start + map->stripe_len) |
| 2915 | extent_len = logic_start + map->stripe_len - | 2915 | extent_len = logic_start + map->stripe_len - |
| 2916 | extent_logical; | 2916 | extent_logical; |
| 2917 | 2917 | ||
| 2918 | scrub_parity_mark_sectors_data(sparity, extent_logical, | 2918 | scrub_parity_mark_sectors_data(sparity, extent_logical, |
| 2919 | extent_len); | 2919 | extent_len); |
| 2920 | 2920 | ||
| 2921 | scrub_remap_extent(fs_info, extent_logical, | 2921 | scrub_remap_extent(fs_info, extent_logical, |
| 2922 | extent_len, &extent_physical, | 2922 | extent_len, &extent_physical, |
| 2923 | &extent_dev, | 2923 | &extent_dev, |
| 2924 | &extent_mirror_num); | 2924 | &extent_mirror_num); |
| 2925 | 2925 | ||
| 2926 | ret = btrfs_lookup_csums_range(csum_root, | 2926 | ret = btrfs_lookup_csums_range(csum_root, |
| 2927 | extent_logical, | 2927 | extent_logical, |
| 2928 | extent_logical + extent_len - 1, | 2928 | extent_logical + extent_len - 1, |
| 2929 | &sctx->csum_list, 1); | 2929 | &sctx->csum_list, 1); |
| 2930 | if (ret) | 2930 | if (ret) |
| 2931 | goto out; | 2931 | goto out; |
| 2932 | 2932 | ||
| 2933 | ret = scrub_extent_for_parity(sparity, extent_logical, | 2933 | ret = scrub_extent_for_parity(sparity, extent_logical, |
| 2934 | extent_len, | 2934 | extent_len, |
| 2935 | extent_physical, | 2935 | extent_physical, |
| 2936 | extent_dev, flags, | 2936 | extent_dev, flags, |
| 2937 | generation, | 2937 | generation, |
| 2938 | extent_mirror_num); | 2938 | extent_mirror_num); |
| 2939 | if (ret) | 2939 | if (ret) |
| 2940 | goto out; | 2940 | goto out; |
| 2941 | 2941 | ||
| 2942 | scrub_free_csums(sctx); | 2942 | scrub_free_csums(sctx); |
| 2943 | if (extent_logical + extent_len < | 2943 | if (extent_logical + extent_len < |
| 2944 | key.objectid + bytes) { | 2944 | key.objectid + bytes) { |
| 2945 | logic_start += map->stripe_len; | 2945 | logic_start += map->stripe_len; |
| 2946 | 2946 | ||
| 2947 | if (logic_start >= logic_end) { | 2947 | if (logic_start >= logic_end) { |
| 2948 | stop_loop = 1; | 2948 | stop_loop = 1; |
| 2949 | break; | 2949 | break; |
| 2950 | } | 2950 | } |
| 2951 | 2951 | ||
| 2952 | if (logic_start < key.objectid + bytes) { | 2952 | if (logic_start < key.objectid + bytes) { |
| 2953 | cond_resched(); | 2953 | cond_resched(); |
| 2954 | goto again; | 2954 | goto again; |
| 2955 | } | 2955 | } |
| 2956 | } | 2956 | } |
| 2957 | next: | 2957 | next: |
| 2958 | path->slots[0]++; | 2958 | path->slots[0]++; |
| 2959 | } | 2959 | } |
| 2960 | 2960 | ||
| 2961 | btrfs_release_path(path); | 2961 | btrfs_release_path(path); |
| 2962 | 2962 | ||
| 2963 | if (stop_loop) | 2963 | if (stop_loop) |
| 2964 | break; | 2964 | break; |
| 2965 | 2965 | ||
| 2966 | logic_start += map->stripe_len; | 2966 | logic_start += map->stripe_len; |
| 2967 | } | 2967 | } |
| 2968 | out: | 2968 | out: |
| 2969 | if (ret < 0) | 2969 | if (ret < 0) |
| 2970 | scrub_parity_mark_sectors_error(sparity, logic_start, | 2970 | scrub_parity_mark_sectors_error(sparity, logic_start, |
| 2971 | logic_end - logic_start + 1); | 2971 | logic_end - logic_start + 1); |
| 2972 | scrub_parity_put(sparity); | 2972 | scrub_parity_put(sparity); |
| 2973 | scrub_submit(sctx); | 2973 | scrub_submit(sctx); |
| 2974 | mutex_lock(&sctx->wr_ctx.wr_lock); | 2974 | mutex_lock(&sctx->wr_ctx.wr_lock); |
| 2975 | scrub_wr_submit(sctx); | 2975 | scrub_wr_submit(sctx); |
| 2976 | mutex_unlock(&sctx->wr_ctx.wr_lock); | 2976 | mutex_unlock(&sctx->wr_ctx.wr_lock); |
| 2977 | 2977 | ||
| 2978 | btrfs_release_path(path); | 2978 | btrfs_release_path(path); |
| 2979 | return ret < 0 ? ret : 0; | 2979 | return ret < 0 ? ret : 0; |
| 2980 | } | 2980 | } |
| 2981 | 2981 | ||
| 2982 | static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, | 2982 | static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, |
| 2983 | struct map_lookup *map, | 2983 | struct map_lookup *map, |
| 2984 | struct btrfs_device *scrub_dev, | 2984 | struct btrfs_device *scrub_dev, |
| 2985 | int num, u64 base, u64 length, | 2985 | int num, u64 base, u64 length, |
| 2986 | int is_dev_replace) | 2986 | int is_dev_replace) |
| 2987 | { | 2987 | { |
| 2988 | struct btrfs_path *path, *ppath; | 2988 | struct btrfs_path *path, *ppath; |
| 2989 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; | 2989 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; |
| 2990 | struct btrfs_root *root = fs_info->extent_root; | 2990 | struct btrfs_root *root = fs_info->extent_root; |
| 2991 | struct btrfs_root *csum_root = fs_info->csum_root; | 2991 | struct btrfs_root *csum_root = fs_info->csum_root; |
| 2992 | struct btrfs_extent_item *extent; | 2992 | struct btrfs_extent_item *extent; |
| 2993 | struct blk_plug plug; | 2993 | struct blk_plug plug; |
| 2994 | u64 flags; | 2994 | u64 flags; |
| 2995 | int ret; | 2995 | int ret; |
| 2996 | int slot; | 2996 | int slot; |
| 2997 | u64 nstripes; | 2997 | u64 nstripes; |
| 2998 | struct extent_buffer *l; | 2998 | struct extent_buffer *l; |
| 2999 | struct btrfs_key key; | 2999 | struct btrfs_key key; |
| 3000 | u64 physical; | 3000 | u64 physical; |
| 3001 | u64 logical; | 3001 | u64 logical; |
| 3002 | u64 logic_end; | 3002 | u64 logic_end; |
| 3003 | u64 physical_end; | 3003 | u64 physical_end; |
| 3004 | u64 generation; | 3004 | u64 generation; |
| 3005 | int mirror_num; | 3005 | int mirror_num; |
| 3006 | struct reada_control *reada1; | 3006 | struct reada_control *reada1; |
| 3007 | struct reada_control *reada2; | 3007 | struct reada_control *reada2; |
| 3008 | struct btrfs_key key_start; | 3008 | struct btrfs_key key_start; |
| 3009 | struct btrfs_key key_end; | 3009 | struct btrfs_key key_end; |
| 3010 | u64 increment = map->stripe_len; | 3010 | u64 increment = map->stripe_len; |
| 3011 | u64 offset; | 3011 | u64 offset; |
| 3012 | u64 extent_logical; | 3012 | u64 extent_logical; |
| 3013 | u64 extent_physical; | 3013 | u64 extent_physical; |
| 3014 | u64 extent_len; | 3014 | u64 extent_len; |
| 3015 | u64 stripe_logical; | 3015 | u64 stripe_logical; |
| 3016 | u64 stripe_end; | 3016 | u64 stripe_end; |
| 3017 | struct btrfs_device *extent_dev; | 3017 | struct btrfs_device *extent_dev; |
| 3018 | int extent_mirror_num; | 3018 | int extent_mirror_num; |
| 3019 | int stop_loop = 0; | 3019 | int stop_loop = 0; |
| 3020 | 3020 | ||
| 3021 | nstripes = length; | 3021 | nstripes = length; |
| 3022 | physical = map->stripes[num].physical; | 3022 | physical = map->stripes[num].physical; |
| 3023 | offset = 0; | 3023 | offset = 0; |
| 3024 | do_div(nstripes, map->stripe_len); | 3024 | do_div(nstripes, map->stripe_len); |
| 3025 | if (map->type & BTRFS_BLOCK_GROUP_RAID0) { | 3025 | if (map->type & BTRFS_BLOCK_GROUP_RAID0) { |
| 3026 | offset = map->stripe_len * num; | 3026 | offset = map->stripe_len * num; |
| 3027 | increment = map->stripe_len * map->num_stripes; | 3027 | increment = map->stripe_len * map->num_stripes; |
| 3028 | mirror_num = 1; | 3028 | mirror_num = 1; |
| 3029 | } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { | 3029 | } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { |
| 3030 | int factor = map->num_stripes / map->sub_stripes; | 3030 | int factor = map->num_stripes / map->sub_stripes; |
| 3031 | offset = map->stripe_len * (num / map->sub_stripes); | 3031 | offset = map->stripe_len * (num / map->sub_stripes); |
| 3032 | increment = map->stripe_len * factor; | 3032 | increment = map->stripe_len * factor; |
| 3033 | mirror_num = num % map->sub_stripes + 1; | 3033 | mirror_num = num % map->sub_stripes + 1; |
| 3034 | } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { | 3034 | } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { |
| 3035 | increment = map->stripe_len; | 3035 | increment = map->stripe_len; |
| 3036 | mirror_num = num % map->num_stripes + 1; | 3036 | mirror_num = num % map->num_stripes + 1; |
| 3037 | } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { | 3037 | } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { |
| 3038 | increment = map->stripe_len; | 3038 | increment = map->stripe_len; |
| 3039 | mirror_num = num % map->num_stripes + 1; | 3039 | mirror_num = num % map->num_stripes + 1; |
| 3040 | } else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | | 3040 | } else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | |
| 3041 | BTRFS_BLOCK_GROUP_RAID6)) { | 3041 | BTRFS_BLOCK_GROUP_RAID6)) { |
| 3042 | get_raid56_logic_offset(physical, num, map, &offset, NULL); | 3042 | get_raid56_logic_offset(physical, num, map, &offset, NULL); |
| 3043 | increment = map->stripe_len * nr_data_stripes(map); | 3043 | increment = map->stripe_len * nr_data_stripes(map); |
| 3044 | mirror_num = 1; | 3044 | mirror_num = 1; |
| 3045 | } else { | 3045 | } else { |
| 3046 | increment = map->stripe_len; | 3046 | increment = map->stripe_len; |
| 3047 | mirror_num = 1; | 3047 | mirror_num = 1; |
| 3048 | } | 3048 | } |
| 3049 | 3049 | ||
| 3050 | path = btrfs_alloc_path(); | 3050 | path = btrfs_alloc_path(); |
| 3051 | if (!path) | 3051 | if (!path) |
| 3052 | return -ENOMEM; | 3052 | return -ENOMEM; |
| 3053 | 3053 | ||
| 3054 | ppath = btrfs_alloc_path(); | 3054 | ppath = btrfs_alloc_path(); |
| 3055 | if (!ppath) { | 3055 | if (!ppath) { |
| 3056 | btrfs_free_path(ppath); | 3056 | btrfs_free_path(ppath); |
| 3057 | return -ENOMEM; | 3057 | return -ENOMEM; |
| 3058 | } | 3058 | } |
| 3059 | 3059 | ||
| 3060 | /* | 3060 | /* |
| 3061 | * work on commit root. The related disk blocks are static as | 3061 | * work on commit root. The related disk blocks are static as |
| 3062 | * long as COW is applied. This means, it is save to rewrite | 3062 | * long as COW is applied. This means, it is save to rewrite |
| 3063 | * them to repair disk errors without any race conditions | 3063 | * them to repair disk errors without any race conditions |
| 3064 | */ | 3064 | */ |
| 3065 | path->search_commit_root = 1; | 3065 | path->search_commit_root = 1; |
| 3066 | path->skip_locking = 1; | 3066 | path->skip_locking = 1; |
| 3067 | 3067 | ||
| 3068 | /* | 3068 | /* |
| 3069 | * trigger the readahead for extent tree csum tree and wait for | 3069 | * trigger the readahead for extent tree csum tree and wait for |
| 3070 | * completion. During readahead, the scrub is officially paused | 3070 | * completion. During readahead, the scrub is officially paused |
| 3071 | * to not hold off transaction commits | 3071 | * to not hold off transaction commits |
| 3072 | */ | 3072 | */ |
| 3073 | logical = base + offset; | 3073 | logical = base + offset; |
| 3074 | physical_end = physical + nstripes * map->stripe_len; | 3074 | physical_end = physical + nstripes * map->stripe_len; |
| 3075 | if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | | 3075 | if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | |
| 3076 | BTRFS_BLOCK_GROUP_RAID6)) { | 3076 | BTRFS_BLOCK_GROUP_RAID6)) { |
| 3077 | get_raid56_logic_offset(physical_end, num, | 3077 | get_raid56_logic_offset(physical_end, num, |
| 3078 | map, &logic_end, NULL); | 3078 | map, &logic_end, NULL); |
| 3079 | logic_end += base; | 3079 | logic_end += base; |
| 3080 | } else { | 3080 | } else { |
| 3081 | logic_end = logical + increment * nstripes; | 3081 | logic_end = logical + increment * nstripes; |
| 3082 | } | 3082 | } |
| 3083 | wait_event(sctx->list_wait, | 3083 | wait_event(sctx->list_wait, |
| 3084 | atomic_read(&sctx->bios_in_flight) == 0); | 3084 | atomic_read(&sctx->bios_in_flight) == 0); |
| 3085 | scrub_blocked_if_needed(fs_info); | 3085 | scrub_blocked_if_needed(fs_info); |
| 3086 | 3086 | ||
| 3087 | /* FIXME it might be better to start readahead at commit root */ | 3087 | /* FIXME it might be better to start readahead at commit root */ |
| 3088 | key_start.objectid = logical; | 3088 | key_start.objectid = logical; |
| 3089 | key_start.type = BTRFS_EXTENT_ITEM_KEY; | 3089 | key_start.type = BTRFS_EXTENT_ITEM_KEY; |
| 3090 | key_start.offset = (u64)0; | 3090 | key_start.offset = (u64)0; |
| 3091 | key_end.objectid = logic_end; | 3091 | key_end.objectid = logic_end; |
| 3092 | key_end.type = BTRFS_METADATA_ITEM_KEY; | 3092 | key_end.type = BTRFS_METADATA_ITEM_KEY; |
| 3093 | key_end.offset = (u64)-1; | 3093 | key_end.offset = (u64)-1; |
| 3094 | reada1 = btrfs_reada_add(root, &key_start, &key_end); | 3094 | reada1 = btrfs_reada_add(root, &key_start, &key_end); |
| 3095 | 3095 | ||
| 3096 | key_start.objectid = BTRFS_EXTENT_CSUM_OBJECTID; | 3096 | key_start.objectid = BTRFS_EXTENT_CSUM_OBJECTID; |
| 3097 | key_start.type = BTRFS_EXTENT_CSUM_KEY; | 3097 | key_start.type = BTRFS_EXTENT_CSUM_KEY; |
| 3098 | key_start.offset = logical; | 3098 | key_start.offset = logical; |
| 3099 | key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID; | 3099 | key_end.objectid = BTRFS_EXTENT_CSUM_OBJECTID; |
| 3100 | key_end.type = BTRFS_EXTENT_CSUM_KEY; | 3100 | key_end.type = BTRFS_EXTENT_CSUM_KEY; |
| 3101 | key_end.offset = logic_end; | 3101 | key_end.offset = logic_end; |
| 3102 | reada2 = btrfs_reada_add(csum_root, &key_start, &key_end); | 3102 | reada2 = btrfs_reada_add(csum_root, &key_start, &key_end); |
| 3103 | 3103 | ||
| 3104 | if (!IS_ERR(reada1)) | 3104 | if (!IS_ERR(reada1)) |
| 3105 | btrfs_reada_wait(reada1); | 3105 | btrfs_reada_wait(reada1); |
| 3106 | if (!IS_ERR(reada2)) | 3106 | if (!IS_ERR(reada2)) |
| 3107 | btrfs_reada_wait(reada2); | 3107 | btrfs_reada_wait(reada2); |
| 3108 | 3108 | ||
| 3109 | 3109 | ||
| 3110 | /* | 3110 | /* |
| 3111 | * collect all data csums for the stripe to avoid seeking during | 3111 | * collect all data csums for the stripe to avoid seeking during |
| 3112 | * the scrub. This might currently (crc32) end up to be about 1MB | 3112 | * the scrub. This might currently (crc32) end up to be about 1MB |
| 3113 | */ | 3113 | */ |
| 3114 | blk_start_plug(&plug); | 3114 | blk_start_plug(&plug); |
| 3115 | 3115 | ||
| 3116 | /* | 3116 | /* |
| 3117 | * now find all extents for each stripe and scrub them | 3117 | * now find all extents for each stripe and scrub them |
| 3118 | */ | 3118 | */ |
| 3119 | ret = 0; | 3119 | ret = 0; |
| 3120 | while (physical < physical_end) { | 3120 | while (physical < physical_end) { |
| 3121 | /* for raid56, we skip parity stripe */ | 3121 | /* for raid56, we skip parity stripe */ |
| 3122 | if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | | 3122 | if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | |
| 3123 | BTRFS_BLOCK_GROUP_RAID6)) { | 3123 | BTRFS_BLOCK_GROUP_RAID6)) { |
| 3124 | ret = get_raid56_logic_offset(physical, num, | 3124 | ret = get_raid56_logic_offset(physical, num, |
| 3125 | map, &logical, &stripe_logical); | 3125 | map, &logical, &stripe_logical); |
| 3126 | logical += base; | 3126 | logical += base; |
| 3127 | if (ret) { | 3127 | if (ret) { |
| 3128 | stripe_logical += base; | 3128 | stripe_logical += base; |
| 3129 | stripe_end = stripe_logical + increment - 1; | 3129 | stripe_end = stripe_logical + increment - 1; |
| 3130 | ret = scrub_raid56_parity(sctx, map, scrub_dev, | 3130 | ret = scrub_raid56_parity(sctx, map, scrub_dev, |
| 3131 | ppath, stripe_logical, | 3131 | ppath, stripe_logical, |
| 3132 | stripe_end); | 3132 | stripe_end); |
| 3133 | if (ret) | 3133 | if (ret) |
| 3134 | goto out; | 3134 | goto out; |
| 3135 | goto skip; | 3135 | goto skip; |
| 3136 | } | 3136 | } |
| 3137 | } | 3137 | } |
| 3138 | /* | 3138 | /* |
| 3139 | * canceled? | 3139 | * canceled? |
| 3140 | */ | 3140 | */ |
| 3141 | if (atomic_read(&fs_info->scrub_cancel_req) || | 3141 | if (atomic_read(&fs_info->scrub_cancel_req) || |
| 3142 | atomic_read(&sctx->cancel_req)) { | 3142 | atomic_read(&sctx->cancel_req)) { |
| 3143 | ret = -ECANCELED; | 3143 | ret = -ECANCELED; |
| 3144 | goto out; | 3144 | goto out; |
| 3145 | } | 3145 | } |
| 3146 | /* | 3146 | /* |
| 3147 | * check to see if we have to pause | 3147 | * check to see if we have to pause |
| 3148 | */ | 3148 | */ |
| 3149 | if (atomic_read(&fs_info->scrub_pause_req)) { | 3149 | if (atomic_read(&fs_info->scrub_pause_req)) { |
| 3150 | /* push queued extents */ | 3150 | /* push queued extents */ |
| 3151 | atomic_set(&sctx->wr_ctx.flush_all_writes, 1); | 3151 | atomic_set(&sctx->wr_ctx.flush_all_writes, 1); |
| 3152 | scrub_submit(sctx); | 3152 | scrub_submit(sctx); |
| 3153 | mutex_lock(&sctx->wr_ctx.wr_lock); | 3153 | mutex_lock(&sctx->wr_ctx.wr_lock); |
| 3154 | scrub_wr_submit(sctx); | 3154 | scrub_wr_submit(sctx); |
| 3155 | mutex_unlock(&sctx->wr_ctx.wr_lock); | 3155 | mutex_unlock(&sctx->wr_ctx.wr_lock); |
| 3156 | wait_event(sctx->list_wait, | 3156 | wait_event(sctx->list_wait, |
| 3157 | atomic_read(&sctx->bios_in_flight) == 0); | 3157 | atomic_read(&sctx->bios_in_flight) == 0); |
| 3158 | atomic_set(&sctx->wr_ctx.flush_all_writes, 0); | 3158 | atomic_set(&sctx->wr_ctx.flush_all_writes, 0); |
| 3159 | scrub_blocked_if_needed(fs_info); | 3159 | scrub_blocked_if_needed(fs_info); |
| 3160 | } | 3160 | } |
| 3161 | 3161 | ||
| 3162 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) | 3162 | if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) |
| 3163 | key.type = BTRFS_METADATA_ITEM_KEY; | 3163 | key.type = BTRFS_METADATA_ITEM_KEY; |
| 3164 | else | 3164 | else |
| 3165 | key.type = BTRFS_EXTENT_ITEM_KEY; | 3165 | key.type = BTRFS_EXTENT_ITEM_KEY; |
| 3166 | key.objectid = logical; | 3166 | key.objectid = logical; |
| 3167 | key.offset = (u64)-1; | 3167 | key.offset = (u64)-1; |
| 3168 | 3168 | ||
| 3169 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 3169 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 3170 | if (ret < 0) | 3170 | if (ret < 0) |
| 3171 | goto out; | 3171 | goto out; |
| 3172 | 3172 | ||
| 3173 | if (ret > 0) { | 3173 | if (ret > 0) { |
| 3174 | ret = btrfs_previous_extent_item(root, path, 0); | 3174 | ret = btrfs_previous_extent_item(root, path, 0); |
| 3175 | if (ret < 0) | 3175 | if (ret < 0) |
| 3176 | goto out; | 3176 | goto out; |
| 3177 | if (ret > 0) { | 3177 | if (ret > 0) { |
| 3178 | /* there's no smaller item, so stick with the | 3178 | /* there's no smaller item, so stick with the |
| 3179 | * larger one */ | 3179 | * larger one */ |
| 3180 | btrfs_release_path(path); | 3180 | btrfs_release_path(path); |
| 3181 | ret = btrfs_search_slot(NULL, root, &key, | 3181 | ret = btrfs_search_slot(NULL, root, &key, |
| 3182 | path, 0, 0); | 3182 | path, 0, 0); |
| 3183 | if (ret < 0) | 3183 | if (ret < 0) |
| 3184 | goto out; | 3184 | goto out; |
| 3185 | } | 3185 | } |
| 3186 | } | 3186 | } |
| 3187 | 3187 | ||
| 3188 | stop_loop = 0; | 3188 | stop_loop = 0; |
| 3189 | while (1) { | 3189 | while (1) { |
| 3190 | u64 bytes; | 3190 | u64 bytes; |
| 3191 | 3191 | ||
| 3192 | l = path->nodes[0]; | 3192 | l = path->nodes[0]; |
| 3193 | slot = path->slots[0]; | 3193 | slot = path->slots[0]; |
| 3194 | if (slot >= btrfs_header_nritems(l)) { | 3194 | if (slot >= btrfs_header_nritems(l)) { |
| 3195 | ret = btrfs_next_leaf(root, path); | 3195 | ret = btrfs_next_leaf(root, path); |
| 3196 | if (ret == 0) | 3196 | if (ret == 0) |
| 3197 | continue; | 3197 | continue; |
| 3198 | if (ret < 0) | 3198 | if (ret < 0) |
| 3199 | goto out; | 3199 | goto out; |
| 3200 | 3200 | ||
| 3201 | stop_loop = 1; | 3201 | stop_loop = 1; |
| 3202 | break; | 3202 | break; |
| 3203 | } | 3203 | } |
| 3204 | btrfs_item_key_to_cpu(l, &key, slot); | 3204 | btrfs_item_key_to_cpu(l, &key, slot); |
| 3205 | 3205 | ||
| 3206 | if (key.type == BTRFS_METADATA_ITEM_KEY) | 3206 | if (key.type == BTRFS_METADATA_ITEM_KEY) |
| 3207 | bytes = root->nodesize; | 3207 | bytes = root->nodesize; |
| 3208 | else | 3208 | else |
| 3209 | bytes = key.offset; | 3209 | bytes = key.offset; |
| 3210 | 3210 | ||
| 3211 | if (key.objectid + bytes <= logical) | 3211 | if (key.objectid + bytes <= logical) |
| 3212 | goto next; | 3212 | goto next; |
| 3213 | 3213 | ||
| 3214 | if (key.type != BTRFS_EXTENT_ITEM_KEY && | 3214 | if (key.type != BTRFS_EXTENT_ITEM_KEY && |
| 3215 | key.type != BTRFS_METADATA_ITEM_KEY) | 3215 | key.type != BTRFS_METADATA_ITEM_KEY) |
| 3216 | goto next; | 3216 | goto next; |
| 3217 | 3217 | ||
| 3218 | if (key.objectid >= logical + map->stripe_len) { | 3218 | if (key.objectid >= logical + map->stripe_len) { |
| 3219 | /* out of this device extent */ | 3219 | /* out of this device extent */ |
| 3220 | if (key.objectid >= logic_end) | 3220 | if (key.objectid >= logic_end) |
| 3221 | stop_loop = 1; | 3221 | stop_loop = 1; |
| 3222 | break; | 3222 | break; |
| 3223 | } | 3223 | } |
| 3224 | 3224 | ||
| 3225 | extent = btrfs_item_ptr(l, slot, | 3225 | extent = btrfs_item_ptr(l, slot, |
| 3226 | struct btrfs_extent_item); | 3226 | struct btrfs_extent_item); |
| 3227 | flags = btrfs_extent_flags(l, extent); | 3227 | flags = btrfs_extent_flags(l, extent); |
| 3228 | generation = btrfs_extent_generation(l, extent); | 3228 | generation = btrfs_extent_generation(l, extent); |
| 3229 | 3229 | ||
| 3230 | if (key.objectid < logical && | 3230 | if (key.objectid < logical && |
| 3231 | (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) { | 3231 | (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) { |
| 3232 | btrfs_err(fs_info, | 3232 | btrfs_err(fs_info, |
| 3233 | "scrub: tree block %llu spanning " | 3233 | "scrub: tree block %llu spanning " |
| 3234 | "stripes, ignored. logical=%llu", | 3234 | "stripes, ignored. logical=%llu", |
| 3235 | key.objectid, logical); | 3235 | key.objectid, logical); |
| 3236 | goto next; | 3236 | goto next; |
| 3237 | } | 3237 | } |
| 3238 | 3238 | ||
| 3239 | again: | 3239 | again: |
| 3240 | extent_logical = key.objectid; | 3240 | extent_logical = key.objectid; |
| 3241 | extent_len = bytes; | 3241 | extent_len = bytes; |
| 3242 | 3242 | ||
| 3243 | /* | 3243 | /* |
| 3244 | * trim extent to this stripe | 3244 | * trim extent to this stripe |
| 3245 | */ | 3245 | */ |
| 3246 | if (extent_logical < logical) { | 3246 | if (extent_logical < logical) { |
| 3247 | extent_len -= logical - extent_logical; | 3247 | extent_len -= logical - extent_logical; |
| 3248 | extent_logical = logical; | 3248 | extent_logical = logical; |
| 3249 | } | 3249 | } |
| 3250 | if (extent_logical + extent_len > | 3250 | if (extent_logical + extent_len > |
| 3251 | logical + map->stripe_len) { | 3251 | logical + map->stripe_len) { |
| 3252 | extent_len = logical + map->stripe_len - | 3252 | extent_len = logical + map->stripe_len - |
| 3253 | extent_logical; | 3253 | extent_logical; |
| 3254 | } | 3254 | } |
| 3255 | 3255 | ||
| 3256 | extent_physical = extent_logical - logical + physical; | 3256 | extent_physical = extent_logical - logical + physical; |
| 3257 | extent_dev = scrub_dev; | 3257 | extent_dev = scrub_dev; |
| 3258 | extent_mirror_num = mirror_num; | 3258 | extent_mirror_num = mirror_num; |
| 3259 | if (is_dev_replace) | 3259 | if (is_dev_replace) |
| 3260 | scrub_remap_extent(fs_info, extent_logical, | 3260 | scrub_remap_extent(fs_info, extent_logical, |
| 3261 | extent_len, &extent_physical, | 3261 | extent_len, &extent_physical, |
| 3262 | &extent_dev, | 3262 | &extent_dev, |
| 3263 | &extent_mirror_num); | 3263 | &extent_mirror_num); |
| 3264 | 3264 | ||
| 3265 | ret = btrfs_lookup_csums_range(csum_root, logical, | 3265 | ret = btrfs_lookup_csums_range(csum_root, logical, |
| 3266 | logical + map->stripe_len - 1, | 3266 | logical + map->stripe_len - 1, |
| 3267 | &sctx->csum_list, 1); | 3267 | &sctx->csum_list, 1); |
| 3268 | if (ret) | 3268 | if (ret) |
| 3269 | goto out; | 3269 | goto out; |
| 3270 | 3270 | ||
| 3271 | ret = scrub_extent(sctx, extent_logical, extent_len, | 3271 | ret = scrub_extent(sctx, extent_logical, extent_len, |
| 3272 | extent_physical, extent_dev, flags, | 3272 | extent_physical, extent_dev, flags, |
| 3273 | generation, extent_mirror_num, | 3273 | generation, extent_mirror_num, |
| 3274 | extent_logical - logical + physical); | 3274 | extent_logical - logical + physical); |
| 3275 | if (ret) | 3275 | if (ret) |
| 3276 | goto out; | 3276 | goto out; |
| 3277 | 3277 | ||
| 3278 | scrub_free_csums(sctx); | 3278 | scrub_free_csums(sctx); |
| 3279 | if (extent_logical + extent_len < | 3279 | if (extent_logical + extent_len < |
| 3280 | key.objectid + bytes) { | 3280 | key.objectid + bytes) { |
| 3281 | if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | | 3281 | if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | |
| 3282 | BTRFS_BLOCK_GROUP_RAID6)) { | 3282 | BTRFS_BLOCK_GROUP_RAID6)) { |
| 3283 | /* | 3283 | /* |
| 3284 | * loop until we find next data stripe | 3284 | * loop until we find next data stripe |
| 3285 | * or we have finished all stripes. | 3285 | * or we have finished all stripes. |
| 3286 | */ | 3286 | */ |
| 3287 | loop: | 3287 | loop: |
| 3288 | physical += map->stripe_len; | 3288 | physical += map->stripe_len; |
| 3289 | ret = get_raid56_logic_offset(physical, | 3289 | ret = get_raid56_logic_offset(physical, |
| 3290 | num, map, &logical, | 3290 | num, map, &logical, |
| 3291 | &stripe_logical); | 3291 | &stripe_logical); |
| 3292 | logical += base; | 3292 | logical += base; |
| 3293 | 3293 | ||
| 3294 | if (ret && physical < physical_end) { | 3294 | if (ret && physical < physical_end) { |
| 3295 | stripe_logical += base; | 3295 | stripe_logical += base; |
| 3296 | stripe_end = stripe_logical + | 3296 | stripe_end = stripe_logical + |
| 3297 | increment - 1; | 3297 | increment - 1; |
| 3298 | ret = scrub_raid56_parity(sctx, | 3298 | ret = scrub_raid56_parity(sctx, |
| 3299 | map, scrub_dev, ppath, | 3299 | map, scrub_dev, ppath, |
| 3300 | stripe_logical, | 3300 | stripe_logical, |
| 3301 | stripe_end); | 3301 | stripe_end); |
| 3302 | if (ret) | 3302 | if (ret) |
| 3303 | goto out; | 3303 | goto out; |
| 3304 | goto loop; | 3304 | goto loop; |
| 3305 | } | 3305 | } |
| 3306 | } else { | 3306 | } else { |
| 3307 | physical += map->stripe_len; | 3307 | physical += map->stripe_len; |
| 3308 | logical += increment; | 3308 | logical += increment; |
| 3309 | } | 3309 | } |
| 3310 | if (logical < key.objectid + bytes) { | 3310 | if (logical < key.objectid + bytes) { |
| 3311 | cond_resched(); | 3311 | cond_resched(); |
| 3312 | goto again; | 3312 | goto again; |
| 3313 | } | 3313 | } |
| 3314 | 3314 | ||
| 3315 | if (physical >= physical_end) { | 3315 | if (physical >= physical_end) { |
| 3316 | stop_loop = 1; | 3316 | stop_loop = 1; |
| 3317 | break; | 3317 | break; |
| 3318 | } | 3318 | } |
| 3319 | } | 3319 | } |
| 3320 | next: | 3320 | next: |
| 3321 | path->slots[0]++; | 3321 | path->slots[0]++; |
| 3322 | } | 3322 | } |
| 3323 | btrfs_release_path(path); | 3323 | btrfs_release_path(path); |
| 3324 | skip: | 3324 | skip: |
| 3325 | logical += increment; | 3325 | logical += increment; |
| 3326 | physical += map->stripe_len; | 3326 | physical += map->stripe_len; |
| 3327 | spin_lock(&sctx->stat_lock); | 3327 | spin_lock(&sctx->stat_lock); |
| 3328 | if (stop_loop) | 3328 | if (stop_loop) |
| 3329 | sctx->stat.last_physical = map->stripes[num].physical + | 3329 | sctx->stat.last_physical = map->stripes[num].physical + |
| 3330 | length; | 3330 | length; |
| 3331 | else | 3331 | else |
| 3332 | sctx->stat.last_physical = physical; | 3332 | sctx->stat.last_physical = physical; |
| 3333 | spin_unlock(&sctx->stat_lock); | 3333 | spin_unlock(&sctx->stat_lock); |
| 3334 | if (stop_loop) | 3334 | if (stop_loop) |
| 3335 | break; | 3335 | break; |
| 3336 | } | 3336 | } |
| 3337 | out: | 3337 | out: |
| 3338 | /* push queued extents */ | 3338 | /* push queued extents */ |
| 3339 | scrub_submit(sctx); | 3339 | scrub_submit(sctx); |
| 3340 | mutex_lock(&sctx->wr_ctx.wr_lock); | 3340 | mutex_lock(&sctx->wr_ctx.wr_lock); |
| 3341 | scrub_wr_submit(sctx); | 3341 | scrub_wr_submit(sctx); |
| 3342 | mutex_unlock(&sctx->wr_ctx.wr_lock); | 3342 | mutex_unlock(&sctx->wr_ctx.wr_lock); |
| 3343 | 3343 | ||
| 3344 | blk_finish_plug(&plug); | 3344 | blk_finish_plug(&plug); |
| 3345 | btrfs_free_path(path); | 3345 | btrfs_free_path(path); |
| 3346 | btrfs_free_path(ppath); | 3346 | btrfs_free_path(ppath); |
| 3347 | return ret < 0 ? ret : 0; | 3347 | return ret < 0 ? ret : 0; |
| 3348 | } | 3348 | } |
| 3349 | 3349 | ||
| 3350 | static noinline_for_stack int scrub_chunk(struct scrub_ctx *sctx, | 3350 | static noinline_for_stack int scrub_chunk(struct scrub_ctx *sctx, |
| 3351 | struct btrfs_device *scrub_dev, | 3351 | struct btrfs_device *scrub_dev, |
| 3352 | u64 chunk_tree, u64 chunk_objectid, | 3352 | u64 chunk_tree, u64 chunk_objectid, |
| 3353 | u64 chunk_offset, u64 length, | 3353 | u64 chunk_offset, u64 length, |
| 3354 | u64 dev_offset, int is_dev_replace) | 3354 | u64 dev_offset, int is_dev_replace) |
| 3355 | { | 3355 | { |
| 3356 | struct btrfs_mapping_tree *map_tree = | 3356 | struct btrfs_mapping_tree *map_tree = |
| 3357 | &sctx->dev_root->fs_info->mapping_tree; | 3357 | &sctx->dev_root->fs_info->mapping_tree; |
| 3358 | struct map_lookup *map; | 3358 | struct map_lookup *map; |
| 3359 | struct extent_map *em; | 3359 | struct extent_map *em; |
| 3360 | int i; | 3360 | int i; |
| 3361 | int ret = 0; | 3361 | int ret = 0; |
| 3362 | 3362 | ||
| 3363 | read_lock(&map_tree->map_tree.lock); | 3363 | read_lock(&map_tree->map_tree.lock); |
| 3364 | em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1); | 3364 | em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1); |
| 3365 | read_unlock(&map_tree->map_tree.lock); | 3365 | read_unlock(&map_tree->map_tree.lock); |
| 3366 | 3366 | ||
| 3367 | if (!em) | 3367 | if (!em) |
| 3368 | return -EINVAL; | 3368 | return -EINVAL; |
| 3369 | 3369 | ||
| 3370 | map = (struct map_lookup *)em->bdev; | 3370 | map = (struct map_lookup *)em->bdev; |
| 3371 | if (em->start != chunk_offset) | 3371 | if (em->start != chunk_offset) |
| 3372 | goto out; | 3372 | goto out; |
| 3373 | 3373 | ||
| 3374 | if (em->len < length) | 3374 | if (em->len < length) |
| 3375 | goto out; | 3375 | goto out; |
| 3376 | 3376 | ||
| 3377 | for (i = 0; i < map->num_stripes; ++i) { | 3377 | for (i = 0; i < map->num_stripes; ++i) { |
| 3378 | if (map->stripes[i].dev->bdev == scrub_dev->bdev && | 3378 | if (map->stripes[i].dev->bdev == scrub_dev->bdev && |
| 3379 | map->stripes[i].physical == dev_offset) { | 3379 | map->stripes[i].physical == dev_offset) { |
| 3380 | ret = scrub_stripe(sctx, map, scrub_dev, i, | 3380 | ret = scrub_stripe(sctx, map, scrub_dev, i, |
| 3381 | chunk_offset, length, | 3381 | chunk_offset, length, |
| 3382 | is_dev_replace); | 3382 | is_dev_replace); |
| 3383 | if (ret) | 3383 | if (ret) |
| 3384 | goto out; | 3384 | goto out; |
| 3385 | } | 3385 | } |
| 3386 | } | 3386 | } |
| 3387 | out: | 3387 | out: |
| 3388 | free_extent_map(em); | 3388 | free_extent_map(em); |
| 3389 | 3389 | ||
| 3390 | return ret; | 3390 | return ret; |
| 3391 | } | 3391 | } |
| 3392 | 3392 | ||
| 3393 | static noinline_for_stack | 3393 | static noinline_for_stack |
| 3394 | int scrub_enumerate_chunks(struct scrub_ctx *sctx, | 3394 | int scrub_enumerate_chunks(struct scrub_ctx *sctx, |
| 3395 | struct btrfs_device *scrub_dev, u64 start, u64 end, | 3395 | struct btrfs_device *scrub_dev, u64 start, u64 end, |
| 3396 | int is_dev_replace) | 3396 | int is_dev_replace) |
| 3397 | { | 3397 | { |
| 3398 | struct btrfs_dev_extent *dev_extent = NULL; | 3398 | struct btrfs_dev_extent *dev_extent = NULL; |
| 3399 | struct btrfs_path *path; | 3399 | struct btrfs_path *path; |
| 3400 | struct btrfs_root *root = sctx->dev_root; | 3400 | struct btrfs_root *root = sctx->dev_root; |
| 3401 | struct btrfs_fs_info *fs_info = root->fs_info; | 3401 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 3402 | u64 length; | 3402 | u64 length; |
| 3403 | u64 chunk_tree; | 3403 | u64 chunk_tree; |
| 3404 | u64 chunk_objectid; | 3404 | u64 chunk_objectid; |
| 3405 | u64 chunk_offset; | 3405 | u64 chunk_offset; |
| 3406 | int ret; | 3406 | int ret; |
| 3407 | int slot; | 3407 | int slot; |
| 3408 | struct extent_buffer *l; | 3408 | struct extent_buffer *l; |
| 3409 | struct btrfs_key key; | 3409 | struct btrfs_key key; |
| 3410 | struct btrfs_key found_key; | 3410 | struct btrfs_key found_key; |
| 3411 | struct btrfs_block_group_cache *cache; | 3411 | struct btrfs_block_group_cache *cache; |
| 3412 | struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace; | 3412 | struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace; |
| 3413 | 3413 | ||
| 3414 | path = btrfs_alloc_path(); | 3414 | path = btrfs_alloc_path(); |
| 3415 | if (!path) | 3415 | if (!path) |
| 3416 | return -ENOMEM; | 3416 | return -ENOMEM; |
| 3417 | 3417 | ||
| 3418 | path->reada = 2; | 3418 | path->reada = 2; |
| 3419 | path->search_commit_root = 1; | 3419 | path->search_commit_root = 1; |
| 3420 | path->skip_locking = 1; | 3420 | path->skip_locking = 1; |
| 3421 | 3421 | ||
| 3422 | key.objectid = scrub_dev->devid; | 3422 | key.objectid = scrub_dev->devid; |
| 3423 | key.offset = 0ull; | 3423 | key.offset = 0ull; |
| 3424 | key.type = BTRFS_DEV_EXTENT_KEY; | 3424 | key.type = BTRFS_DEV_EXTENT_KEY; |
| 3425 | 3425 | ||
| 3426 | while (1) { | 3426 | while (1) { |
| 3427 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | 3427 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
| 3428 | if (ret < 0) | 3428 | if (ret < 0) |
| 3429 | break; | 3429 | break; |
| 3430 | if (ret > 0) { | 3430 | if (ret > 0) { |
| 3431 | if (path->slots[0] >= | 3431 | if (path->slots[0] >= |
| 3432 | btrfs_header_nritems(path->nodes[0])) { | 3432 | btrfs_header_nritems(path->nodes[0])) { |
| 3433 | ret = btrfs_next_leaf(root, path); | 3433 | ret = btrfs_next_leaf(root, path); |
| 3434 | if (ret) | 3434 | if (ret) |
| 3435 | break; | 3435 | break; |
| 3436 | } | 3436 | } |
| 3437 | } | 3437 | } |
| 3438 | 3438 | ||
| 3439 | l = path->nodes[0]; | 3439 | l = path->nodes[0]; |
| 3440 | slot = path->slots[0]; | 3440 | slot = path->slots[0]; |
| 3441 | 3441 | ||
| 3442 | btrfs_item_key_to_cpu(l, &found_key, slot); | 3442 | btrfs_item_key_to_cpu(l, &found_key, slot); |
| 3443 | 3443 | ||
| 3444 | if (found_key.objectid != scrub_dev->devid) | 3444 | if (found_key.objectid != scrub_dev->devid) |
| 3445 | break; | 3445 | break; |
| 3446 | 3446 | ||
| 3447 | if (found_key.type != BTRFS_DEV_EXTENT_KEY) | 3447 | if (found_key.type != BTRFS_DEV_EXTENT_KEY) |
| 3448 | break; | 3448 | break; |
| 3449 | 3449 | ||
| 3450 | if (found_key.offset >= end) | 3450 | if (found_key.offset >= end) |
| 3451 | break; | 3451 | break; |
| 3452 | 3452 | ||
| 3453 | if (found_key.offset < key.offset) | 3453 | if (found_key.offset < key.offset) |
| 3454 | break; | 3454 | break; |
| 3455 | 3455 | ||
| 3456 | dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent); | 3456 | dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent); |
| 3457 | length = btrfs_dev_extent_length(l, dev_extent); | 3457 | length = btrfs_dev_extent_length(l, dev_extent); |
| 3458 | 3458 | ||
| 3459 | if (found_key.offset + length <= start) | 3459 | if (found_key.offset + length <= start) |
| 3460 | goto skip; | 3460 | goto skip; |
| 3461 | 3461 | ||
| 3462 | chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent); | 3462 | chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent); |
| 3463 | chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent); | 3463 | chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent); |
| 3464 | chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent); | 3464 | chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent); |
| 3465 | 3465 | ||
| 3466 | /* | 3466 | /* |
| 3467 | * get a reference on the corresponding block group to prevent | 3467 | * get a reference on the corresponding block group to prevent |
| 3468 | * the chunk from going away while we scrub it | 3468 | * the chunk from going away while we scrub it |
| 3469 | */ | 3469 | */ |
| 3470 | cache = btrfs_lookup_block_group(fs_info, chunk_offset); | 3470 | cache = btrfs_lookup_block_group(fs_info, chunk_offset); |
| 3471 | 3471 | ||
| 3472 | /* some chunks are removed but not committed to disk yet, | 3472 | /* some chunks are removed but not committed to disk yet, |
| 3473 | * continue scrubbing */ | 3473 | * continue scrubbing */ |
| 3474 | if (!cache) | 3474 | if (!cache) |
| 3475 | goto skip; | 3475 | goto skip; |
| 3476 | 3476 | ||
| 3477 | dev_replace->cursor_right = found_key.offset + length; | 3477 | dev_replace->cursor_right = found_key.offset + length; |
| 3478 | dev_replace->cursor_left = found_key.offset; | 3478 | dev_replace->cursor_left = found_key.offset; |
| 3479 | dev_replace->item_needs_writeback = 1; | 3479 | dev_replace->item_needs_writeback = 1; |
| 3480 | ret = scrub_chunk(sctx, scrub_dev, chunk_tree, chunk_objectid, | 3480 | ret = scrub_chunk(sctx, scrub_dev, chunk_tree, chunk_objectid, |
| 3481 | chunk_offset, length, found_key.offset, | 3481 | chunk_offset, length, found_key.offset, |
| 3482 | is_dev_replace); | 3482 | is_dev_replace); |
| 3483 | 3483 | ||
| 3484 | /* | 3484 | /* |
| 3485 | * flush, submit all pending read and write bios, afterwards | 3485 | * flush, submit all pending read and write bios, afterwards |
| 3486 | * wait for them. | 3486 | * wait for them. |
| 3487 | * Note that in the dev replace case, a read request causes | 3487 | * Note that in the dev replace case, a read request causes |
| 3488 | * write requests that are submitted in the read completion | 3488 | * write requests that are submitted in the read completion |
| 3489 | * worker. Therefore in the current situation, it is required | 3489 | * worker. Therefore in the current situation, it is required |
| 3490 | * that all write requests are flushed, so that all read and | 3490 | * that all write requests are flushed, so that all read and |
| 3491 | * write requests are really completed when bios_in_flight | 3491 | * write requests are really completed when bios_in_flight |
| 3492 | * changes to 0. | 3492 | * changes to 0. |
| 3493 | */ | 3493 | */ |
| 3494 | atomic_set(&sctx->wr_ctx.flush_all_writes, 1); | 3494 | atomic_set(&sctx->wr_ctx.flush_all_writes, 1); |
| 3495 | scrub_submit(sctx); | 3495 | scrub_submit(sctx); |
| 3496 | mutex_lock(&sctx->wr_ctx.wr_lock); | 3496 | mutex_lock(&sctx->wr_ctx.wr_lock); |
| 3497 | scrub_wr_submit(sctx); | 3497 | scrub_wr_submit(sctx); |
| 3498 | mutex_unlock(&sctx->wr_ctx.wr_lock); | 3498 | mutex_unlock(&sctx->wr_ctx.wr_lock); |
| 3499 | 3499 | ||
| 3500 | wait_event(sctx->list_wait, | 3500 | wait_event(sctx->list_wait, |
| 3501 | atomic_read(&sctx->bios_in_flight) == 0); | 3501 | atomic_read(&sctx->bios_in_flight) == 0); |
| 3502 | atomic_inc(&fs_info->scrubs_paused); | 3502 | atomic_inc(&fs_info->scrubs_paused); |
| 3503 | wake_up(&fs_info->scrub_pause_wait); | 3503 | wake_up(&fs_info->scrub_pause_wait); |
| 3504 | 3504 | ||
| 3505 | /* | 3505 | /* |
| 3506 | * must be called before we decrease @scrub_paused. | 3506 | * must be called before we decrease @scrub_paused. |
| 3507 | * make sure we don't block transaction commit while | 3507 | * make sure we don't block transaction commit while |
| 3508 | * we are waiting pending workers finished. | 3508 | * we are waiting pending workers finished. |
| 3509 | */ | 3509 | */ |
| 3510 | wait_event(sctx->list_wait, | 3510 | wait_event(sctx->list_wait, |
| 3511 | atomic_read(&sctx->workers_pending) == 0); | 3511 | atomic_read(&sctx->workers_pending) == 0); |
| 3512 | atomic_set(&sctx->wr_ctx.flush_all_writes, 0); | 3512 | atomic_set(&sctx->wr_ctx.flush_all_writes, 0); |
| 3513 | 3513 | ||
| 3514 | mutex_lock(&fs_info->scrub_lock); | 3514 | mutex_lock(&fs_info->scrub_lock); |
| 3515 | __scrub_blocked_if_needed(fs_info); | 3515 | __scrub_blocked_if_needed(fs_info); |
| 3516 | atomic_dec(&fs_info->scrubs_paused); | 3516 | atomic_dec(&fs_info->scrubs_paused); |
| 3517 | mutex_unlock(&fs_info->scrub_lock); | 3517 | mutex_unlock(&fs_info->scrub_lock); |
| 3518 | wake_up(&fs_info->scrub_pause_wait); | 3518 | wake_up(&fs_info->scrub_pause_wait); |
| 3519 | 3519 | ||
| 3520 | btrfs_put_block_group(cache); | 3520 | btrfs_put_block_group(cache); |
| 3521 | if (ret) | 3521 | if (ret) |
| 3522 | break; | 3522 | break; |
| 3523 | if (is_dev_replace && | 3523 | if (is_dev_replace && |
| 3524 | atomic64_read(&dev_replace->num_write_errors) > 0) { | 3524 | atomic64_read(&dev_replace->num_write_errors) > 0) { |
| 3525 | ret = -EIO; | 3525 | ret = -EIO; |
| 3526 | break; | 3526 | break; |
| 3527 | } | 3527 | } |
| 3528 | if (sctx->stat.malloc_errors > 0) { | 3528 | if (sctx->stat.malloc_errors > 0) { |
| 3529 | ret = -ENOMEM; | 3529 | ret = -ENOMEM; |
| 3530 | break; | 3530 | break; |
| 3531 | } | 3531 | } |
| 3532 | 3532 | ||
| 3533 | dev_replace->cursor_left = dev_replace->cursor_right; | 3533 | dev_replace->cursor_left = dev_replace->cursor_right; |
| 3534 | dev_replace->item_needs_writeback = 1; | 3534 | dev_replace->item_needs_writeback = 1; |
| 3535 | skip: | 3535 | skip: |
| 3536 | key.offset = found_key.offset + length; | 3536 | key.offset = found_key.offset + length; |
| 3537 | btrfs_release_path(path); | 3537 | btrfs_release_path(path); |
| 3538 | } | 3538 | } |
| 3539 | 3539 | ||
| 3540 | btrfs_free_path(path); | 3540 | btrfs_free_path(path); |
| 3541 | 3541 | ||
| 3542 | /* | 3542 | /* |
| 3543 | * ret can still be 1 from search_slot or next_leaf, | 3543 | * ret can still be 1 from search_slot or next_leaf, |
| 3544 | * that's not an error | 3544 | * that's not an error |
| 3545 | */ | 3545 | */ |
| 3546 | return ret < 0 ? ret : 0; | 3546 | return ret < 0 ? ret : 0; |
| 3547 | } | 3547 | } |
| 3548 | 3548 | ||
| 3549 | static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx, | 3549 | static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx, |
| 3550 | struct btrfs_device *scrub_dev) | 3550 | struct btrfs_device *scrub_dev) |
| 3551 | { | 3551 | { |
| 3552 | int i; | 3552 | int i; |
| 3553 | u64 bytenr; | 3553 | u64 bytenr; |
| 3554 | u64 gen; | 3554 | u64 gen; |
| 3555 | int ret; | 3555 | int ret; |
| 3556 | struct btrfs_root *root = sctx->dev_root; | 3556 | struct btrfs_root *root = sctx->dev_root; |
| 3557 | 3557 | ||
| 3558 | if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) | 3558 | if (test_bit(BTRFS_FS_STATE_ERROR, &root->fs_info->fs_state)) |
| 3559 | return -EIO; | 3559 | return -EIO; |
| 3560 | 3560 | ||
| 3561 | /* Seed devices of a new filesystem has their own generation. */ | 3561 | /* Seed devices of a new filesystem has their own generation. */ |
| 3562 | if (scrub_dev->fs_devices != root->fs_info->fs_devices) | 3562 | if (scrub_dev->fs_devices != root->fs_info->fs_devices) |
| 3563 | gen = scrub_dev->generation; | 3563 | gen = scrub_dev->generation; |
| 3564 | else | 3564 | else |
| 3565 | gen = root->fs_info->last_trans_committed; | 3565 | gen = root->fs_info->last_trans_committed; |
| 3566 | 3566 | ||
| 3567 | for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { | 3567 | for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { |
| 3568 | bytenr = btrfs_sb_offset(i); | 3568 | bytenr = btrfs_sb_offset(i); |
| 3569 | if (bytenr + BTRFS_SUPER_INFO_SIZE > | 3569 | if (bytenr + BTRFS_SUPER_INFO_SIZE > |
| 3570 | scrub_dev->commit_total_bytes) | 3570 | scrub_dev->commit_total_bytes) |
| 3571 | break; | 3571 | break; |
| 3572 | 3572 | ||
| 3573 | ret = scrub_pages(sctx, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr, | 3573 | ret = scrub_pages(sctx, bytenr, BTRFS_SUPER_INFO_SIZE, bytenr, |
| 3574 | scrub_dev, BTRFS_EXTENT_FLAG_SUPER, gen, i, | 3574 | scrub_dev, BTRFS_EXTENT_FLAG_SUPER, gen, i, |
| 3575 | NULL, 1, bytenr); | 3575 | NULL, 1, bytenr); |
| 3576 | if (ret) | 3576 | if (ret) |
| 3577 | return ret; | 3577 | return ret; |
| 3578 | } | 3578 | } |
| 3579 | wait_event(sctx->list_wait, atomic_read(&sctx->bios_in_flight) == 0); | 3579 | wait_event(sctx->list_wait, atomic_read(&sctx->bios_in_flight) == 0); |
| 3580 | 3580 | ||
| 3581 | return 0; | 3581 | return 0; |
| 3582 | } | 3582 | } |
| 3583 | 3583 | ||
| 3584 | /* | 3584 | /* |
| 3585 | * get a reference count on fs_info->scrub_workers. start worker if necessary | 3585 | * get a reference count on fs_info->scrub_workers. start worker if necessary |
| 3586 | */ | 3586 | */ |
| 3587 | static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info, | 3587 | static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info, |
| 3588 | int is_dev_replace) | 3588 | int is_dev_replace) |
| 3589 | { | 3589 | { |
| 3590 | int ret = 0; | 3590 | int ret = 0; |
| 3591 | int flags = WQ_FREEZABLE | WQ_UNBOUND; | 3591 | int flags = WQ_FREEZABLE | WQ_UNBOUND; |
| 3592 | int max_active = fs_info->thread_pool_size; | 3592 | int max_active = fs_info->thread_pool_size; |
| 3593 | 3593 | ||
| 3594 | if (fs_info->scrub_workers_refcnt == 0) { | 3594 | if (fs_info->scrub_workers_refcnt == 0) { |
| 3595 | if (is_dev_replace) | 3595 | if (is_dev_replace) |
| 3596 | fs_info->scrub_workers = | 3596 | fs_info->scrub_workers = |
| 3597 | btrfs_alloc_workqueue("btrfs-scrub", flags, | 3597 | btrfs_alloc_workqueue("btrfs-scrub", flags, |
| 3598 | 1, 4); | 3598 | 1, 4); |
| 3599 | else | 3599 | else |
| 3600 | fs_info->scrub_workers = | 3600 | fs_info->scrub_workers = |
| 3601 | btrfs_alloc_workqueue("btrfs-scrub", flags, | 3601 | btrfs_alloc_workqueue("btrfs-scrub", flags, |
| 3602 | max_active, 4); | 3602 | max_active, 4); |
| 3603 | if (!fs_info->scrub_workers) { | 3603 | if (!fs_info->scrub_workers) { |
| 3604 | ret = -ENOMEM; | 3604 | ret = -ENOMEM; |
| 3605 | goto out; | 3605 | goto out; |
| 3606 | } | 3606 | } |
| 3607 | fs_info->scrub_wr_completion_workers = | 3607 | fs_info->scrub_wr_completion_workers = |
| 3608 | btrfs_alloc_workqueue("btrfs-scrubwrc", flags, | 3608 | btrfs_alloc_workqueue("btrfs-scrubwrc", flags, |
| 3609 | max_active, 2); | 3609 | max_active, 2); |
| 3610 | if (!fs_info->scrub_wr_completion_workers) { | 3610 | if (!fs_info->scrub_wr_completion_workers) { |
| 3611 | ret = -ENOMEM; | 3611 | ret = -ENOMEM; |
| 3612 | goto out; | 3612 | goto out; |
| 3613 | } | 3613 | } |
| 3614 | fs_info->scrub_nocow_workers = | 3614 | fs_info->scrub_nocow_workers = |
| 3615 | btrfs_alloc_workqueue("btrfs-scrubnc", flags, 1, 0); | 3615 | btrfs_alloc_workqueue("btrfs-scrubnc", flags, 1, 0); |
| 3616 | if (!fs_info->scrub_nocow_workers) { | 3616 | if (!fs_info->scrub_nocow_workers) { |
| 3617 | ret = -ENOMEM; | 3617 | ret = -ENOMEM; |
| 3618 | goto out; | 3618 | goto out; |
| 3619 | } | 3619 | } |
| 3620 | } | 3620 | } |
| 3621 | ++fs_info->scrub_workers_refcnt; | 3621 | ++fs_info->scrub_workers_refcnt; |
| 3622 | out: | 3622 | out: |
| 3623 | return ret; | 3623 | return ret; |
| 3624 | } | 3624 | } |
| 3625 | 3625 | ||
| 3626 | static noinline_for_stack void scrub_workers_put(struct btrfs_fs_info *fs_info) | 3626 | static noinline_for_stack void scrub_workers_put(struct btrfs_fs_info *fs_info) |
| 3627 | { | 3627 | { |
| 3628 | if (--fs_info->scrub_workers_refcnt == 0) { | 3628 | if (--fs_info->scrub_workers_refcnt == 0) { |
| 3629 | btrfs_destroy_workqueue(fs_info->scrub_workers); | 3629 | btrfs_destroy_workqueue(fs_info->scrub_workers); |
| 3630 | btrfs_destroy_workqueue(fs_info->scrub_wr_completion_workers); | 3630 | btrfs_destroy_workqueue(fs_info->scrub_wr_completion_workers); |
| 3631 | btrfs_destroy_workqueue(fs_info->scrub_nocow_workers); | 3631 | btrfs_destroy_workqueue(fs_info->scrub_nocow_workers); |
| 3632 | } | 3632 | } |
| 3633 | WARN_ON(fs_info->scrub_workers_refcnt < 0); | 3633 | WARN_ON(fs_info->scrub_workers_refcnt < 0); |
| 3634 | } | 3634 | } |
| 3635 | 3635 | ||
| 3636 | int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, | 3636 | int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, |
| 3637 | u64 end, struct btrfs_scrub_progress *progress, | 3637 | u64 end, struct btrfs_scrub_progress *progress, |
| 3638 | int readonly, int is_dev_replace) | 3638 | int readonly, int is_dev_replace) |
| 3639 | { | 3639 | { |
| 3640 | struct scrub_ctx *sctx; | 3640 | struct scrub_ctx *sctx; |
| 3641 | int ret; | 3641 | int ret; |
| 3642 | struct btrfs_device *dev; | 3642 | struct btrfs_device *dev; |
| 3643 | struct rcu_string *name; | 3643 | struct rcu_string *name; |
| 3644 | 3644 | ||
| 3645 | if (btrfs_fs_closing(fs_info)) | 3645 | if (btrfs_fs_closing(fs_info)) |
| 3646 | return -EINVAL; | 3646 | return -EINVAL; |
| 3647 | 3647 | ||
| 3648 | if (fs_info->chunk_root->nodesize > BTRFS_STRIPE_LEN) { | 3648 | if (fs_info->chunk_root->nodesize > BTRFS_STRIPE_LEN) { |
| 3649 | /* | 3649 | /* |
| 3650 | * in this case scrub is unable to calculate the checksum | 3650 | * in this case scrub is unable to calculate the checksum |
| 3651 | * the way scrub is implemented. Do not handle this | 3651 | * the way scrub is implemented. Do not handle this |
| 3652 | * situation at all because it won't ever happen. | 3652 | * situation at all because it won't ever happen. |
| 3653 | */ | 3653 | */ |
| 3654 | btrfs_err(fs_info, | 3654 | btrfs_err(fs_info, |
| 3655 | "scrub: size assumption nodesize <= BTRFS_STRIPE_LEN (%d <= %d) fails", | 3655 | "scrub: size assumption nodesize <= BTRFS_STRIPE_LEN (%d <= %d) fails", |
| 3656 | fs_info->chunk_root->nodesize, BTRFS_STRIPE_LEN); | 3656 | fs_info->chunk_root->nodesize, BTRFS_STRIPE_LEN); |
| 3657 | return -EINVAL; | 3657 | return -EINVAL; |
| 3658 | } | 3658 | } |
| 3659 | 3659 | ||
| 3660 | if (fs_info->chunk_root->sectorsize != PAGE_SIZE) { | 3660 | if (fs_info->chunk_root->sectorsize != PAGE_SIZE) { |
| 3661 | /* not supported for data w/o checksums */ | 3661 | /* not supported for data w/o checksums */ |
| 3662 | btrfs_err(fs_info, | 3662 | btrfs_err(fs_info, |
| 3663 | "scrub: size assumption sectorsize != PAGE_SIZE " | 3663 | "scrub: size assumption sectorsize != PAGE_SIZE " |
| 3664 | "(%d != %lu) fails", | 3664 | "(%d != %lu) fails", |
| 3665 | fs_info->chunk_root->sectorsize, PAGE_SIZE); | 3665 | fs_info->chunk_root->sectorsize, PAGE_SIZE); |
| 3666 | return -EINVAL; | 3666 | return -EINVAL; |
| 3667 | } | 3667 | } |
| 3668 | 3668 | ||
| 3669 | if (fs_info->chunk_root->nodesize > | 3669 | if (fs_info->chunk_root->nodesize > |
| 3670 | PAGE_SIZE * SCRUB_MAX_PAGES_PER_BLOCK || | 3670 | PAGE_SIZE * SCRUB_MAX_PAGES_PER_BLOCK || |
| 3671 | fs_info->chunk_root->sectorsize > | 3671 | fs_info->chunk_root->sectorsize > |
| 3672 | PAGE_SIZE * SCRUB_MAX_PAGES_PER_BLOCK) { | 3672 | PAGE_SIZE * SCRUB_MAX_PAGES_PER_BLOCK) { |
| 3673 | /* | 3673 | /* |
| 3674 | * would exhaust the array bounds of pagev member in | 3674 | * would exhaust the array bounds of pagev member in |
| 3675 | * struct scrub_block | 3675 | * struct scrub_block |
| 3676 | */ | 3676 | */ |
| 3677 | btrfs_err(fs_info, "scrub: size assumption nodesize and sectorsize " | 3677 | btrfs_err(fs_info, "scrub: size assumption nodesize and sectorsize " |
| 3678 | "<= SCRUB_MAX_PAGES_PER_BLOCK (%d <= %d && %d <= %d) fails", | 3678 | "<= SCRUB_MAX_PAGES_PER_BLOCK (%d <= %d && %d <= %d) fails", |
| 3679 | fs_info->chunk_root->nodesize, | 3679 | fs_info->chunk_root->nodesize, |
| 3680 | SCRUB_MAX_PAGES_PER_BLOCK, | 3680 | SCRUB_MAX_PAGES_PER_BLOCK, |
| 3681 | fs_info->chunk_root->sectorsize, | 3681 | fs_info->chunk_root->sectorsize, |
| 3682 | SCRUB_MAX_PAGES_PER_BLOCK); | 3682 | SCRUB_MAX_PAGES_PER_BLOCK); |
| 3683 | return -EINVAL; | 3683 | return -EINVAL; |
| 3684 | } | 3684 | } |
| 3685 | 3685 | ||
| 3686 | 3686 | ||
| 3687 | mutex_lock(&fs_info->fs_devices->device_list_mutex); | 3687 | mutex_lock(&fs_info->fs_devices->device_list_mutex); |
| 3688 | dev = btrfs_find_device(fs_info, devid, NULL, NULL); | 3688 | dev = btrfs_find_device(fs_info, devid, NULL, NULL); |
| 3689 | if (!dev || (dev->missing && !is_dev_replace)) { | 3689 | if (!dev || (dev->missing && !is_dev_replace)) { |
| 3690 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3690 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3691 | return -ENODEV; | 3691 | return -ENODEV; |
| 3692 | } | 3692 | } |
| 3693 | 3693 | ||
| 3694 | if (!is_dev_replace && !readonly && !dev->writeable) { | 3694 | if (!is_dev_replace && !readonly && !dev->writeable) { |
| 3695 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3695 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3696 | rcu_read_lock(); | 3696 | rcu_read_lock(); |
| 3697 | name = rcu_dereference(dev->name); | 3697 | name = rcu_dereference(dev->name); |
| 3698 | btrfs_err(fs_info, "scrub: device %s is not writable", | 3698 | btrfs_err(fs_info, "scrub: device %s is not writable", |
| 3699 | name->str); | 3699 | name->str); |
| 3700 | rcu_read_unlock(); | 3700 | rcu_read_unlock(); |
| 3701 | return -EROFS; | 3701 | return -EROFS; |
| 3702 | } | 3702 | } |
| 3703 | 3703 | ||
| 3704 | mutex_lock(&fs_info->scrub_lock); | 3704 | mutex_lock(&fs_info->scrub_lock); |
| 3705 | if (!dev->in_fs_metadata || dev->is_tgtdev_for_dev_replace) { | 3705 | if (!dev->in_fs_metadata || dev->is_tgtdev_for_dev_replace) { |
| 3706 | mutex_unlock(&fs_info->scrub_lock); | 3706 | mutex_unlock(&fs_info->scrub_lock); |
| 3707 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3707 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3708 | return -EIO; | 3708 | return -EIO; |
| 3709 | } | 3709 | } |
| 3710 | 3710 | ||
| 3711 | btrfs_dev_replace_lock(&fs_info->dev_replace); | 3711 | btrfs_dev_replace_lock(&fs_info->dev_replace); |
| 3712 | if (dev->scrub_device || | 3712 | if (dev->scrub_device || |
| 3713 | (!is_dev_replace && | 3713 | (!is_dev_replace && |
| 3714 | btrfs_dev_replace_is_ongoing(&fs_info->dev_replace))) { | 3714 | btrfs_dev_replace_is_ongoing(&fs_info->dev_replace))) { |
| 3715 | btrfs_dev_replace_unlock(&fs_info->dev_replace); | 3715 | btrfs_dev_replace_unlock(&fs_info->dev_replace); |
| 3716 | mutex_unlock(&fs_info->scrub_lock); | 3716 | mutex_unlock(&fs_info->scrub_lock); |
| 3717 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3717 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3718 | return -EINPROGRESS; | 3718 | return -EINPROGRESS; |
| 3719 | } | 3719 | } |
| 3720 | btrfs_dev_replace_unlock(&fs_info->dev_replace); | 3720 | btrfs_dev_replace_unlock(&fs_info->dev_replace); |
| 3721 | 3721 | ||
| 3722 | ret = scrub_workers_get(fs_info, is_dev_replace); | 3722 | ret = scrub_workers_get(fs_info, is_dev_replace); |
| 3723 | if (ret) { | 3723 | if (ret) { |
| 3724 | mutex_unlock(&fs_info->scrub_lock); | 3724 | mutex_unlock(&fs_info->scrub_lock); |
| 3725 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3725 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3726 | return ret; | 3726 | return ret; |
| 3727 | } | 3727 | } |
| 3728 | 3728 | ||
| 3729 | sctx = scrub_setup_ctx(dev, is_dev_replace); | 3729 | sctx = scrub_setup_ctx(dev, is_dev_replace); |
| 3730 | if (IS_ERR(sctx)) { | 3730 | if (IS_ERR(sctx)) { |
| 3731 | mutex_unlock(&fs_info->scrub_lock); | 3731 | mutex_unlock(&fs_info->scrub_lock); |
| 3732 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3732 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3733 | scrub_workers_put(fs_info); | 3733 | scrub_workers_put(fs_info); |
| 3734 | return PTR_ERR(sctx); | 3734 | return PTR_ERR(sctx); |
| 3735 | } | 3735 | } |
| 3736 | sctx->readonly = readonly; | 3736 | sctx->readonly = readonly; |
| 3737 | dev->scrub_device = sctx; | 3737 | dev->scrub_device = sctx; |
| 3738 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3738 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3739 | 3739 | ||
| 3740 | /* | 3740 | /* |
| 3741 | * checking @scrub_pause_req here, we can avoid | 3741 | * checking @scrub_pause_req here, we can avoid |
| 3742 | * race between committing transaction and scrubbing. | 3742 | * race between committing transaction and scrubbing. |
| 3743 | */ | 3743 | */ |
| 3744 | __scrub_blocked_if_needed(fs_info); | 3744 | __scrub_blocked_if_needed(fs_info); |
| 3745 | atomic_inc(&fs_info->scrubs_running); | 3745 | atomic_inc(&fs_info->scrubs_running); |
| 3746 | mutex_unlock(&fs_info->scrub_lock); | 3746 | mutex_unlock(&fs_info->scrub_lock); |
| 3747 | 3747 | ||
| 3748 | if (!is_dev_replace) { | 3748 | if (!is_dev_replace) { |
| 3749 | /* | 3749 | /* |
| 3750 | * by holding device list mutex, we can | 3750 | * by holding device list mutex, we can |
| 3751 | * kick off writing super in log tree sync. | 3751 | * kick off writing super in log tree sync. |
| 3752 | */ | 3752 | */ |
| 3753 | mutex_lock(&fs_info->fs_devices->device_list_mutex); | 3753 | mutex_lock(&fs_info->fs_devices->device_list_mutex); |
| 3754 | ret = scrub_supers(sctx, dev); | 3754 | ret = scrub_supers(sctx, dev); |
| 3755 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); | 3755 | mutex_unlock(&fs_info->fs_devices->device_list_mutex); |
| 3756 | } | 3756 | } |
| 3757 | 3757 | ||
| 3758 | if (!ret) | 3758 | if (!ret) |
| 3759 | ret = scrub_enumerate_chunks(sctx, dev, start, end, | 3759 | ret = scrub_enumerate_chunks(sctx, dev, start, end, |
| 3760 | is_dev_replace); | 3760 | is_dev_replace); |
| 3761 | 3761 | ||
| 3762 | wait_event(sctx->list_wait, atomic_read(&sctx->bios_in_flight) == 0); | 3762 | wait_event(sctx->list_wait, atomic_read(&sctx->bios_in_flight) == 0); |
| 3763 | atomic_dec(&fs_info->scrubs_running); | 3763 | atomic_dec(&fs_info->scrubs_running); |
| 3764 | wake_up(&fs_info->scrub_pause_wait); | 3764 | wake_up(&fs_info->scrub_pause_wait); |
| 3765 | 3765 | ||
| 3766 | wait_event(sctx->list_wait, atomic_read(&sctx->workers_pending) == 0); | 3766 | wait_event(sctx->list_wait, atomic_read(&sctx->workers_pending) == 0); |
| 3767 | 3767 | ||
| 3768 | if (progress) | 3768 | if (progress) |
| 3769 | memcpy(progress, &sctx->stat, sizeof(*progress)); | 3769 | memcpy(progress, &sctx->stat, sizeof(*progress)); |
| 3770 | 3770 | ||
| 3771 | mutex_lock(&fs_info->scrub_lock); | 3771 | mutex_lock(&fs_info->scrub_lock); |
| 3772 | dev->scrub_device = NULL; | 3772 | dev->scrub_device = NULL; |
| 3773 | scrub_workers_put(fs_info); | 3773 | scrub_workers_put(fs_info); |
| 3774 | mutex_unlock(&fs_info->scrub_lock); | 3774 | mutex_unlock(&fs_info->scrub_lock); |
| 3775 | 3775 | ||
| 3776 | scrub_free_ctx(sctx); | 3776 | scrub_free_ctx(sctx); |
| 3777 | 3777 | ||
| 3778 | return ret; | 3778 | return ret; |
| 3779 | } | 3779 | } |
| 3780 | 3780 | ||
| 3781 | void btrfs_scrub_pause(struct btrfs_root *root) | 3781 | void btrfs_scrub_pause(struct btrfs_root *root) |
| 3782 | { | 3782 | { |
| 3783 | struct btrfs_fs_info *fs_info = root->fs_info; | 3783 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 3784 | 3784 | ||
| 3785 | mutex_lock(&fs_info->scrub_lock); | 3785 | mutex_lock(&fs_info->scrub_lock); |
| 3786 | atomic_inc(&fs_info->scrub_pause_req); | 3786 | atomic_inc(&fs_info->scrub_pause_req); |
| 3787 | while (atomic_read(&fs_info->scrubs_paused) != | 3787 | while (atomic_read(&fs_info->scrubs_paused) != |
| 3788 | atomic_read(&fs_info->scrubs_running)) { | 3788 | atomic_read(&fs_info->scrubs_running)) { |
| 3789 | mutex_unlock(&fs_info->scrub_lock); | 3789 | mutex_unlock(&fs_info->scrub_lock); |
| 3790 | wait_event(fs_info->scrub_pause_wait, | 3790 | wait_event(fs_info->scrub_pause_wait, |
| 3791 | atomic_read(&fs_info->scrubs_paused) == | 3791 | atomic_read(&fs_info->scrubs_paused) == |
| 3792 | atomic_read(&fs_info->scrubs_running)); | 3792 | atomic_read(&fs_info->scrubs_running)); |
| 3793 | mutex_lock(&fs_info->scrub_lock); | 3793 | mutex_lock(&fs_info->scrub_lock); |
| 3794 | } | 3794 | } |
| 3795 | mutex_unlock(&fs_info->scrub_lock); | 3795 | mutex_unlock(&fs_info->scrub_lock); |
| 3796 | } | 3796 | } |
| 3797 | 3797 | ||
| 3798 | void btrfs_scrub_continue(struct btrfs_root *root) | 3798 | void btrfs_scrub_continue(struct btrfs_root *root) |
| 3799 | { | 3799 | { |
| 3800 | struct btrfs_fs_info *fs_info = root->fs_info; | 3800 | struct btrfs_fs_info *fs_info = root->fs_info; |
| 3801 | 3801 | ||
| 3802 | atomic_dec(&fs_info->scrub_pause_req); | 3802 | atomic_dec(&fs_info->scrub_pause_req); |
| 3803 | wake_up(&fs_info->scrub_pause_wait); | 3803 | wake_up(&fs_info->scrub_pause_wait); |
| 3804 | } | 3804 | } |
| 3805 | 3805 | ||
| 3806 | int btrfs_scrub_cancel(struct btrfs_fs_info *fs_info) | 3806 | int btrfs_scrub_cancel(struct btrfs_fs_info *fs_info) |
| 3807 | { | 3807 | { |
| 3808 | mutex_lock(&fs_info->scrub_lock); | 3808 | mutex_lock(&fs_info->scrub_lock); |
| 3809 | if (!atomic_read(&fs_info->scrubs_running)) { | 3809 | if (!atomic_read(&fs_info->scrubs_running)) { |
| 3810 | mutex_unlock(&fs_info->scrub_lock); | 3810 | mutex_unlock(&fs_info->scrub_lock); |
| 3811 | return -ENOTCONN; | 3811 | return -ENOTCONN; |
| 3812 | } | 3812 | } |
| 3813 | 3813 | ||
| 3814 | atomic_inc(&fs_info->scrub_cancel_req); | 3814 | atomic_inc(&fs_info->scrub_cancel_req); |
| 3815 | while (atomic_read(&fs_info->scrubs_running)) { | 3815 | while (atomic_read(&fs_info->scrubs_running)) { |
| 3816 | mutex_unlock(&fs_info->scrub_lock); | 3816 | mutex_unlock(&fs_info->scrub_lock); |
| 3817 | wait_event(fs_info->scrub_pause_wait, | 3817 | wait_event(fs_info->scrub_pause_wait, |
| 3818 | atomic_read(&fs_info->scrubs_running) == 0); | 3818 | atomic_read(&fs_info->scrubs_running) == 0); |
| 3819 | mutex_lock(&fs_info->scrub_lock); | 3819 | mutex_lock(&fs_info->scrub_lock); |
| 3820 | } | 3820 | } |
| 3821 | atomic_dec(&fs_info->scrub_cancel_req); | 3821 | atomic_dec(&fs_info->scrub_cancel_req); |
| 3822 | mutex_unlock(&fs_info->scrub_lock); | 3822 | mutex_unlock(&fs_info->scrub_lock); |
| 3823 | 3823 | ||
| 3824 | return 0; | 3824 | return 0; |
| 3825 | } | 3825 | } |
| 3826 | 3826 | ||
| 3827 | int btrfs_scrub_cancel_dev(struct btrfs_fs_info *fs_info, | 3827 | int btrfs_scrub_cancel_dev(struct btrfs_fs_info *fs_info, |
| 3828 | struct btrfs_device *dev) | 3828 | struct btrfs_device *dev) |
| 3829 | { | 3829 | { |
| 3830 | struct scrub_ctx *sctx; | 3830 | struct scrub_ctx *sctx; |
| 3831 | 3831 | ||
| 3832 | mutex_lock(&fs_info->scrub_lock); | 3832 | mutex_lock(&fs_info->scrub_lock); |
| 3833 | sctx = dev->scrub_device; | 3833 | sctx = dev->scrub_device; |
| 3834 | if (!sctx) { | 3834 | if (!sctx) { |
| 3835 | mutex_unlock(&fs_info->scrub_lock); | 3835 | mutex_unlock(&fs_info->scrub_lock); |
| 3836 | return -ENOTCONN; | 3836 | return -ENOTCONN; |
| 3837 | } | 3837 | } |
| 3838 | atomic_inc(&sctx->cancel_req); | 3838 | atomic_inc(&sctx->cancel_req); |
| 3839 | while (dev->scrub_device) { | 3839 | while (dev->scrub_device) { |
| 3840 | mutex_unlock(&fs_info->scrub_lock); | 3840 | mutex_unlock(&fs_info->scrub_lock); |
| 3841 | wait_event(fs_info->scrub_pause_wait, | 3841 | wait_event(fs_info->scrub_pause_wait, |
| 3842 | dev->scrub_device == NULL); | 3842 | dev->scrub_device == NULL); |
| 3843 | mutex_lock(&fs_info->scrub_lock); | 3843 | mutex_lock(&fs_info->scrub_lock); |
| 3844 | } | 3844 | } |
| 3845 | mutex_unlock(&fs_info->scrub_lock); | 3845 | mutex_unlock(&fs_info->scrub_lock); |
| 3846 | 3846 | ||
| 3847 | return 0; | 3847 | return 0; |
| 3848 | } | 3848 | } |
| 3849 | 3849 | ||
| 3850 | int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, | 3850 | int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, |
| 3851 | struct btrfs_scrub_progress *progress) | 3851 | struct btrfs_scrub_progress *progress) |
| 3852 | { | 3852 | { |
| 3853 | struct btrfs_device *dev; | 3853 | struct btrfs_device *dev; |
| 3854 | struct scrub_ctx *sctx = NULL; | 3854 | struct scrub_ctx *sctx = NULL; |
| 3855 | 3855 | ||
| 3856 | mutex_lock(&root->fs_info->fs_devices->device_list_mutex); | 3856 | mutex_lock(&root->fs_info->fs_devices->device_list_mutex); |
| 3857 | dev = btrfs_find_device(root->fs_info, devid, NULL, NULL); | 3857 | dev = btrfs_find_device(root->fs_info, devid, NULL, NULL); |
| 3858 | if (dev) | 3858 | if (dev) |
| 3859 | sctx = dev->scrub_device; | 3859 | sctx = dev->scrub_device; |
| 3860 | if (sctx) | 3860 | if (sctx) |
| 3861 | memcpy(progress, &sctx->stat, sizeof(*progress)); | 3861 | memcpy(progress, &sctx->stat, sizeof(*progress)); |
| 3862 | mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); | 3862 | mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); |
| 3863 | 3863 | ||
| 3864 | return dev ? (sctx ? 0 : -ENOTCONN) : -ENODEV; | 3864 | return dev ? (sctx ? 0 : -ENOTCONN) : -ENODEV; |
| 3865 | } | 3865 | } |
| 3866 | 3866 | ||
| 3867 | static void scrub_remap_extent(struct btrfs_fs_info *fs_info, | 3867 | static void scrub_remap_extent(struct btrfs_fs_info *fs_info, |
| 3868 | u64 extent_logical, u64 extent_len, | 3868 | u64 extent_logical, u64 extent_len, |
| 3869 | u64 *extent_physical, | 3869 | u64 *extent_physical, |
| 3870 | struct btrfs_device **extent_dev, | 3870 | struct btrfs_device **extent_dev, |
| 3871 | int *extent_mirror_num) | 3871 | int *extent_mirror_num) |
| 3872 | { | 3872 | { |
| 3873 | u64 mapped_length; | 3873 | u64 mapped_length; |
| 3874 | struct btrfs_bio *bbio = NULL; | 3874 | struct btrfs_bio *bbio = NULL; |
| 3875 | int ret; | 3875 | int ret; |
| 3876 | 3876 | ||
| 3877 | mapped_length = extent_len; | 3877 | mapped_length = extent_len; |
| 3878 | ret = btrfs_map_block(fs_info, READ, extent_logical, | 3878 | ret = btrfs_map_block(fs_info, READ, extent_logical, |
| 3879 | &mapped_length, &bbio, 0); | 3879 | &mapped_length, &bbio, 0); |
| 3880 | if (ret || !bbio || mapped_length < extent_len || | 3880 | if (ret || !bbio || mapped_length < extent_len || |
| 3881 | !bbio->stripes[0].dev->bdev) { | 3881 | !bbio->stripes[0].dev->bdev) { |
| 3882 | kfree(bbio); | 3882 | kfree(bbio); |
| 3883 | return; | 3883 | return; |
| 3884 | } | 3884 | } |
| 3885 | 3885 | ||
| 3886 | *extent_physical = bbio->stripes[0].physical; | 3886 | *extent_physical = bbio->stripes[0].physical; |
| 3887 | *extent_mirror_num = bbio->mirror_num; | 3887 | *extent_mirror_num = bbio->mirror_num; |
| 3888 | *extent_dev = bbio->stripes[0].dev; | 3888 | *extent_dev = bbio->stripes[0].dev; |
| 3889 | kfree(bbio); | 3889 | kfree(bbio); |
| 3890 | } | 3890 | } |
| 3891 | 3891 | ||
| 3892 | static int scrub_setup_wr_ctx(struct scrub_ctx *sctx, | 3892 | static int scrub_setup_wr_ctx(struct scrub_ctx *sctx, |
| 3893 | struct scrub_wr_ctx *wr_ctx, | 3893 | struct scrub_wr_ctx *wr_ctx, |
| 3894 | struct btrfs_fs_info *fs_info, | 3894 | struct btrfs_fs_info *fs_info, |
| 3895 | struct btrfs_device *dev, | 3895 | struct btrfs_device *dev, |
| 3896 | int is_dev_replace) | 3896 | int is_dev_replace) |
| 3897 | { | 3897 | { |
| 3898 | WARN_ON(wr_ctx->wr_curr_bio != NULL); | 3898 | WARN_ON(wr_ctx->wr_curr_bio != NULL); |
| 3899 | 3899 | ||
| 3900 | mutex_init(&wr_ctx->wr_lock); | 3900 | mutex_init(&wr_ctx->wr_lock); |
| 3901 | wr_ctx->wr_curr_bio = NULL; | 3901 | wr_ctx->wr_curr_bio = NULL; |
| 3902 | if (!is_dev_replace) | 3902 | if (!is_dev_replace) |
| 3903 | return 0; | 3903 | return 0; |
| 3904 | 3904 | ||
| 3905 | WARN_ON(!dev->bdev); | 3905 | WARN_ON(!dev->bdev); |
| 3906 | wr_ctx->pages_per_wr_bio = min_t(int, SCRUB_PAGES_PER_WR_BIO, | 3906 | wr_ctx->pages_per_wr_bio = min_t(int, SCRUB_PAGES_PER_WR_BIO, |
| 3907 | bio_get_nr_vecs(dev->bdev)); | 3907 | bio_get_nr_vecs(dev->bdev)); |
| 3908 | wr_ctx->tgtdev = dev; | 3908 | wr_ctx->tgtdev = dev; |
| 3909 | atomic_set(&wr_ctx->flush_all_writes, 0); | 3909 | atomic_set(&wr_ctx->flush_all_writes, 0); |
| 3910 | return 0; | 3910 | return 0; |
| 3911 | } | 3911 | } |
| 3912 | 3912 | ||
| 3913 | static void scrub_free_wr_ctx(struct scrub_wr_ctx *wr_ctx) | 3913 | static void scrub_free_wr_ctx(struct scrub_wr_ctx *wr_ctx) |
| 3914 | { | 3914 | { |
| 3915 | mutex_lock(&wr_ctx->wr_lock); | 3915 | mutex_lock(&wr_ctx->wr_lock); |
| 3916 | kfree(wr_ctx->wr_curr_bio); | 3916 | kfree(wr_ctx->wr_curr_bio); |
| 3917 | wr_ctx->wr_curr_bio = NULL; | 3917 | wr_ctx->wr_curr_bio = NULL; |
| 3918 | mutex_unlock(&wr_ctx->wr_lock); | 3918 | mutex_unlock(&wr_ctx->wr_lock); |
| 3919 | } | 3919 | } |
| 3920 | 3920 | ||
| 3921 | static int copy_nocow_pages(struct scrub_ctx *sctx, u64 logical, u64 len, | 3921 | static int copy_nocow_pages(struct scrub_ctx *sctx, u64 logical, u64 len, |
| 3922 | int mirror_num, u64 physical_for_dev_replace) | 3922 | int mirror_num, u64 physical_for_dev_replace) |
| 3923 | { | 3923 | { |
| 3924 | struct scrub_copy_nocow_ctx *nocow_ctx; | 3924 | struct scrub_copy_nocow_ctx *nocow_ctx; |
| 3925 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; | 3925 | struct btrfs_fs_info *fs_info = sctx->dev_root->fs_info; |
| 3926 | 3926 | ||
| 3927 | nocow_ctx = kzalloc(sizeof(*nocow_ctx), GFP_NOFS); | 3927 | nocow_ctx = kzalloc(sizeof(*nocow_ctx), GFP_NOFS); |
| 3928 | if (!nocow_ctx) { | 3928 | if (!nocow_ctx) { |
| 3929 | spin_lock(&sctx->stat_lock); | 3929 | spin_lock(&sctx->stat_lock); |
| 3930 | sctx->stat.malloc_errors++; | 3930 | sctx->stat.malloc_errors++; |
| 3931 | spin_unlock(&sctx->stat_lock); | 3931 | spin_unlock(&sctx->stat_lock); |
| 3932 | return -ENOMEM; | 3932 | return -ENOMEM; |
| 3933 | } | 3933 | } |
| 3934 | 3934 | ||
| 3935 | scrub_pending_trans_workers_inc(sctx); | 3935 | scrub_pending_trans_workers_inc(sctx); |
| 3936 | 3936 | ||
| 3937 | nocow_ctx->sctx = sctx; | 3937 | nocow_ctx->sctx = sctx; |
| 3938 | nocow_ctx->logical = logical; | 3938 | nocow_ctx->logical = logical; |
| 3939 | nocow_ctx->len = len; | 3939 | nocow_ctx->len = len; |
| 3940 | nocow_ctx->mirror_num = mirror_num; | 3940 | nocow_ctx->mirror_num = mirror_num; |
| 3941 | nocow_ctx->physical_for_dev_replace = physical_for_dev_replace; | 3941 | nocow_ctx->physical_for_dev_replace = physical_for_dev_replace; |
| 3942 | btrfs_init_work(&nocow_ctx->work, btrfs_scrubnc_helper, | 3942 | btrfs_init_work(&nocow_ctx->work, btrfs_scrubnc_helper, |
| 3943 | copy_nocow_pages_worker, NULL, NULL); | 3943 | copy_nocow_pages_worker, NULL, NULL); |
| 3944 | INIT_LIST_HEAD(&nocow_ctx->inodes); | 3944 | INIT_LIST_HEAD(&nocow_ctx->inodes); |
| 3945 | btrfs_queue_work(fs_info->scrub_nocow_workers, | 3945 | btrfs_queue_work(fs_info->scrub_nocow_workers, |
| 3946 | &nocow_ctx->work); | 3946 | &nocow_ctx->work); |
| 3947 | 3947 | ||
| 3948 | return 0; | 3948 | return 0; |
| 3949 | } | 3949 | } |
| 3950 | 3950 | ||
| 3951 | static int record_inode_for_nocow(u64 inum, u64 offset, u64 root, void *ctx) | 3951 | static int record_inode_for_nocow(u64 inum, u64 offset, u64 root, void *ctx) |
| 3952 | { | 3952 | { |
| 3953 | struct scrub_copy_nocow_ctx *nocow_ctx = ctx; | 3953 | struct scrub_copy_nocow_ctx *nocow_ctx = ctx; |
| 3954 | struct scrub_nocow_inode *nocow_inode; | 3954 | struct scrub_nocow_inode *nocow_inode; |
| 3955 | 3955 | ||
| 3956 | nocow_inode = kzalloc(sizeof(*nocow_inode), GFP_NOFS); | 3956 | nocow_inode = kzalloc(sizeof(*nocow_inode), GFP_NOFS); |
| 3957 | if (!nocow_inode) | 3957 | if (!nocow_inode) |
| 3958 | return -ENOMEM; | 3958 | return -ENOMEM; |
| 3959 | nocow_inode->inum = inum; | 3959 | nocow_inode->inum = inum; |
| 3960 | nocow_inode->offset = offset; | 3960 | nocow_inode->offset = offset; |
| 3961 | nocow_inode->root = root; | 3961 | nocow_inode->root = root; |
| 3962 | list_add_tail(&nocow_inode->list, &nocow_ctx->inodes); | 3962 | list_add_tail(&nocow_inode->list, &nocow_ctx->inodes); |
| 3963 | return 0; | 3963 | return 0; |
| 3964 | } | 3964 | } |
| 3965 | 3965 | ||
| 3966 | #define COPY_COMPLETE 1 | 3966 | #define COPY_COMPLETE 1 |
| 3967 | 3967 | ||
| 3968 | static void copy_nocow_pages_worker(struct btrfs_work *work) | 3968 | static void copy_nocow_pages_worker(struct btrfs_work *work) |
| 3969 | { | 3969 | { |
| 3970 | struct scrub_copy_nocow_ctx *nocow_ctx = | 3970 | struct scrub_copy_nocow_ctx *nocow_ctx = |
| 3971 | container_of(work, struct scrub_copy_nocow_ctx, work); | 3971 | container_of(work, struct scrub_copy_nocow_ctx, work); |
| 3972 | struct scrub_ctx *sctx = nocow_ctx->sctx; | 3972 | struct scrub_ctx *sctx = nocow_ctx->sctx; |
| 3973 | u64 logical = nocow_ctx->logical; | 3973 | u64 logical = nocow_ctx->logical; |
| 3974 | u64 len = nocow_ctx->len; | 3974 | u64 len = nocow_ctx->len; |
| 3975 | int mirror_num = nocow_ctx->mirror_num; | 3975 | int mirror_num = nocow_ctx->mirror_num; |
| 3976 | u64 physical_for_dev_replace = nocow_ctx->physical_for_dev_replace; | 3976 | u64 physical_for_dev_replace = nocow_ctx->physical_for_dev_replace; |
| 3977 | int ret; | 3977 | int ret; |
| 3978 | struct btrfs_trans_handle *trans = NULL; | 3978 | struct btrfs_trans_handle *trans = NULL; |
| 3979 | struct btrfs_fs_info *fs_info; | 3979 | struct btrfs_fs_info *fs_info; |
| 3980 | struct btrfs_path *path; | 3980 | struct btrfs_path *path; |
| 3981 | struct btrfs_root *root; | 3981 | struct btrfs_root *root; |
| 3982 | int not_written = 0; | 3982 | int not_written = 0; |
| 3983 | 3983 | ||
| 3984 | fs_info = sctx->dev_root->fs_info; | 3984 | fs_info = sctx->dev_root->fs_info; |
| 3985 | root = fs_info->extent_root; | 3985 | root = fs_info->extent_root; |
| 3986 | 3986 | ||
| 3987 | path = btrfs_alloc_path(); | 3987 | path = btrfs_alloc_path(); |
| 3988 | if (!path) { | 3988 | if (!path) { |
| 3989 | spin_lock(&sctx->stat_lock); | 3989 | spin_lock(&sctx->stat_lock); |
| 3990 | sctx->stat.malloc_errors++; | 3990 | sctx->stat.malloc_errors++; |
| 3991 | spin_unlock(&sctx->stat_lock); | 3991 | spin_unlock(&sctx->stat_lock); |
| 3992 | not_written = 1; | 3992 | not_written = 1; |
| 3993 | goto out; | 3993 | goto out; |
| 3994 | } | 3994 | } |
| 3995 | 3995 | ||
| 3996 | trans = btrfs_join_transaction(root); | 3996 | trans = btrfs_join_transaction(root); |
| 3997 | if (IS_ERR(trans)) { | 3997 | if (IS_ERR(trans)) { |
| 3998 | not_written = 1; | 3998 | not_written = 1; |
| 3999 | goto out; | 3999 | goto out; |
| 4000 | } | 4000 | } |
| 4001 | 4001 | ||
| 4002 | ret = iterate_inodes_from_logical(logical, fs_info, path, | 4002 | ret = iterate_inodes_from_logical(logical, fs_info, path, |
| 4003 | record_inode_for_nocow, nocow_ctx); | 4003 | record_inode_for_nocow, nocow_ctx); |
| 4004 | if (ret != 0 && ret != -ENOENT) { | 4004 | if (ret != 0 && ret != -ENOENT) { |
| 4005 | btrfs_warn(fs_info, "iterate_inodes_from_logical() failed: log %llu, " | 4005 | btrfs_warn(fs_info, "iterate_inodes_from_logical() failed: log %llu, " |
| 4006 | "phys %llu, len %llu, mir %u, ret %d", | 4006 | "phys %llu, len %llu, mir %u, ret %d", |
| 4007 | logical, physical_for_dev_replace, len, mirror_num, | 4007 | logical, physical_for_dev_replace, len, mirror_num, |
| 4008 | ret); | 4008 | ret); |
| 4009 | not_written = 1; | 4009 | not_written = 1; |
| 4010 | goto out; | 4010 | goto out; |
| 4011 | } | 4011 | } |
| 4012 | 4012 | ||
| 4013 | btrfs_end_transaction(trans, root); | 4013 | btrfs_end_transaction(trans, root); |
| 4014 | trans = NULL; | 4014 | trans = NULL; |
| 4015 | while (!list_empty(&nocow_ctx->inodes)) { | 4015 | while (!list_empty(&nocow_ctx->inodes)) { |
| 4016 | struct scrub_nocow_inode *entry; | 4016 | struct scrub_nocow_inode *entry; |
| 4017 | entry = list_first_entry(&nocow_ctx->inodes, | 4017 | entry = list_first_entry(&nocow_ctx->inodes, |
| 4018 | struct scrub_nocow_inode, | 4018 | struct scrub_nocow_inode, |
| 4019 | list); | 4019 | list); |
| 4020 | list_del_init(&entry->list); | 4020 | list_del_init(&entry->list); |
| 4021 | ret = copy_nocow_pages_for_inode(entry->inum, entry->offset, | 4021 | ret = copy_nocow_pages_for_inode(entry->inum, entry->offset, |
| 4022 | entry->root, nocow_ctx); | 4022 | entry->root, nocow_ctx); |
| 4023 | kfree(entry); | 4023 | kfree(entry); |
| 4024 | if (ret == COPY_COMPLETE) { | 4024 | if (ret == COPY_COMPLETE) { |
| 4025 | ret = 0; | 4025 | ret = 0; |
| 4026 | break; | 4026 | break; |
| 4027 | } else if (ret) { | 4027 | } else if (ret) { |
| 4028 | break; | 4028 | break; |
| 4029 | } | 4029 | } |
| 4030 | } | 4030 | } |
| 4031 | out: | 4031 | out: |
| 4032 | while (!list_empty(&nocow_ctx->inodes)) { | 4032 | while (!list_empty(&nocow_ctx->inodes)) { |
| 4033 | struct scrub_nocow_inode *entry; | 4033 | struct scrub_nocow_inode *entry; |
| 4034 | entry = list_first_entry(&nocow_ctx->inodes, | 4034 | entry = list_first_entry(&nocow_ctx->inodes, |
| 4035 | struct scrub_nocow_inode, | 4035 | struct scrub_nocow_inode, |
| 4036 | list); | 4036 | list); |
| 4037 | list_del_init(&entry->list); | 4037 | list_del_init(&entry->list); |
| 4038 | kfree(entry); | 4038 | kfree(entry); |
| 4039 | } | 4039 | } |
| 4040 | if (trans && !IS_ERR(trans)) | 4040 | if (trans && !IS_ERR(trans)) |
| 4041 | btrfs_end_transaction(trans, root); | 4041 | btrfs_end_transaction(trans, root); |
| 4042 | if (not_written) | 4042 | if (not_written) |
| 4043 | btrfs_dev_replace_stats_inc(&fs_info->dev_replace. | 4043 | btrfs_dev_replace_stats_inc(&fs_info->dev_replace. |
| 4044 | num_uncorrectable_read_errors); | 4044 | num_uncorrectable_read_errors); |
| 4045 | 4045 | ||
| 4046 | btrfs_free_path(path); | 4046 | btrfs_free_path(path); |
| 4047 | kfree(nocow_ctx); | 4047 | kfree(nocow_ctx); |
| 4048 | 4048 | ||
| 4049 | scrub_pending_trans_workers_dec(sctx); | 4049 | scrub_pending_trans_workers_dec(sctx); |
| 4050 | } | 4050 | } |
| 4051 | 4051 | ||
| 4052 | static int check_extent_to_block(struct inode *inode, u64 start, u64 len, | 4052 | static int check_extent_to_block(struct inode *inode, u64 start, u64 len, |
| 4053 | u64 logical) | 4053 | u64 logical) |
| 4054 | { | 4054 | { |
| 4055 | struct extent_state *cached_state = NULL; | 4055 | struct extent_state *cached_state = NULL; |
| 4056 | struct btrfs_ordered_extent *ordered; | 4056 | struct btrfs_ordered_extent *ordered; |
| 4057 | struct extent_io_tree *io_tree; | 4057 | struct extent_io_tree *io_tree; |
| 4058 | struct extent_map *em; | 4058 | struct extent_map *em; |
| 4059 | u64 lockstart = start, lockend = start + len - 1; | 4059 | u64 lockstart = start, lockend = start + len - 1; |
| 4060 | int ret = 0; | 4060 | int ret = 0; |
| 4061 | 4061 | ||
| 4062 | io_tree = &BTRFS_I(inode)->io_tree; | 4062 | io_tree = &BTRFS_I(inode)->io_tree; |
| 4063 | 4063 | ||
| 4064 | lock_extent_bits(io_tree, lockstart, lockend, 0, &cached_state); | 4064 | lock_extent_bits(io_tree, lockstart, lockend, 0, &cached_state); |
| 4065 | ordered = btrfs_lookup_ordered_range(inode, lockstart, len); | 4065 | ordered = btrfs_lookup_ordered_range(inode, lockstart, len); |
| 4066 | if (ordered) { | 4066 | if (ordered) { |
| 4067 | btrfs_put_ordered_extent(ordered); | 4067 | btrfs_put_ordered_extent(ordered); |
| 4068 | ret = 1; | 4068 | ret = 1; |
| 4069 | goto out_unlock; | 4069 | goto out_unlock; |
| 4070 | } | 4070 | } |
| 4071 | 4071 | ||
| 4072 | em = btrfs_get_extent(inode, NULL, 0, start, len, 0); | 4072 | em = btrfs_get_extent(inode, NULL, 0, start, len, 0); |
| 4073 | if (IS_ERR(em)) { | 4073 | if (IS_ERR(em)) { |
| 4074 | ret = PTR_ERR(em); | 4074 | ret = PTR_ERR(em); |
| 4075 | goto out_unlock; | 4075 | goto out_unlock; |
| 4076 | } | 4076 | } |
| 4077 | 4077 | ||
| 4078 | /* | 4078 | /* |
| 4079 | * This extent does not actually cover the logical extent anymore, | 4079 | * This extent does not actually cover the logical extent anymore, |
| 4080 | * move on to the next inode. | 4080 | * move on to the next inode. |
| 4081 | */ | 4081 | */ |
| 4082 | if (em->block_start > logical || | 4082 | if (em->block_start > logical || |
| 4083 | em->block_start + em->block_len < logical + len) { | 4083 | em->block_start + em->block_len < logical + len) { |
| 4084 | free_extent_map(em); | 4084 | free_extent_map(em); |
| 4085 | ret = 1; | 4085 | ret = 1; |
| 4086 | goto out_unlock; | 4086 | goto out_unlock; |
| 4087 | } | 4087 | } |
| 4088 | free_extent_map(em); | 4088 | free_extent_map(em); |
| 4089 | 4089 | ||
| 4090 | out_unlock: | 4090 | out_unlock: |
| 4091 | unlock_extent_cached(io_tree, lockstart, lockend, &cached_state, | 4091 | unlock_extent_cached(io_tree, lockstart, lockend, &cached_state, |
| 4092 | GFP_NOFS); | 4092 | GFP_NOFS); |
| 4093 | return ret; | 4093 | return ret; |
| 4094 | } | 4094 | } |
| 4095 | 4095 | ||
| 4096 | static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, | 4096 | static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, |
| 4097 | struct scrub_copy_nocow_ctx *nocow_ctx) | 4097 | struct scrub_copy_nocow_ctx *nocow_ctx) |
| 4098 | { | 4098 | { |
| 4099 | struct btrfs_fs_info *fs_info = nocow_ctx->sctx->dev_root->fs_info; | 4099 | struct btrfs_fs_info *fs_info = nocow_ctx->sctx->dev_root->fs_info; |
| 4100 | struct btrfs_key key; | 4100 | struct btrfs_key key; |
| 4101 | struct inode *inode; | 4101 | struct inode *inode; |
| 4102 | struct page *page; | 4102 | struct page *page; |
| 4103 | struct btrfs_root *local_root; | 4103 | struct btrfs_root *local_root; |
| 4104 | struct extent_io_tree *io_tree; | 4104 | struct extent_io_tree *io_tree; |
| 4105 | u64 physical_for_dev_replace; | 4105 | u64 physical_for_dev_replace; |
| 4106 | u64 nocow_ctx_logical; | 4106 | u64 nocow_ctx_logical; |
| 4107 | u64 len = nocow_ctx->len; | 4107 | u64 len = nocow_ctx->len; |
| 4108 | unsigned long index; | 4108 | unsigned long index; |
| 4109 | int srcu_index; | 4109 | int srcu_index; |
| 4110 | int ret = 0; | 4110 | int ret = 0; |
| 4111 | int err = 0; | 4111 | int err = 0; |
| 4112 | 4112 | ||
| 4113 | key.objectid = root; | 4113 | key.objectid = root; |
| 4114 | key.type = BTRFS_ROOT_ITEM_KEY; | 4114 | key.type = BTRFS_ROOT_ITEM_KEY; |
| 4115 | key.offset = (u64)-1; | 4115 | key.offset = (u64)-1; |
| 4116 | 4116 | ||
| 4117 | srcu_index = srcu_read_lock(&fs_info->subvol_srcu); | 4117 | srcu_index = srcu_read_lock(&fs_info->subvol_srcu); |
| 4118 | 4118 | ||
| 4119 | local_root = btrfs_read_fs_root_no_name(fs_info, &key); | 4119 | local_root = btrfs_read_fs_root_no_name(fs_info, &key); |
| 4120 | if (IS_ERR(local_root)) { | 4120 | if (IS_ERR(local_root)) { |
| 4121 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); | 4121 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); |
| 4122 | return PTR_ERR(local_root); | 4122 | return PTR_ERR(local_root); |
| 4123 | } | 4123 | } |
| 4124 | 4124 | ||
| 4125 | key.type = BTRFS_INODE_ITEM_KEY; | 4125 | key.type = BTRFS_INODE_ITEM_KEY; |
| 4126 | key.objectid = inum; | 4126 | key.objectid = inum; |
| 4127 | key.offset = 0; | 4127 | key.offset = 0; |
| 4128 | inode = btrfs_iget(fs_info->sb, &key, local_root, NULL); | 4128 | inode = btrfs_iget(fs_info->sb, &key, local_root, NULL); |
| 4129 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); | 4129 | srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); |
| 4130 | if (IS_ERR(inode)) | 4130 | if (IS_ERR(inode)) |
| 4131 | return PTR_ERR(inode); | 4131 | return PTR_ERR(inode); |
| 4132 | 4132 | ||
| 4133 | /* Avoid truncate/dio/punch hole.. */ | 4133 | /* Avoid truncate/dio/punch hole.. */ |
| 4134 | mutex_lock(&inode->i_mutex); | 4134 | mutex_lock(&inode->i_mutex); |
| 4135 | inode_dio_wait(inode); | 4135 | inode_dio_wait(inode); |
| 4136 | 4136 | ||
| 4137 | physical_for_dev_replace = nocow_ctx->physical_for_dev_replace; | 4137 | physical_for_dev_replace = nocow_ctx->physical_for_dev_replace; |
| 4138 | io_tree = &BTRFS_I(inode)->io_tree; | 4138 | io_tree = &BTRFS_I(inode)->io_tree; |
| 4139 | nocow_ctx_logical = nocow_ctx->logical; | 4139 | nocow_ctx_logical = nocow_ctx->logical; |
| 4140 | 4140 | ||
| 4141 | ret = check_extent_to_block(inode, offset, len, nocow_ctx_logical); | 4141 | ret = check_extent_to_block(inode, offset, len, nocow_ctx_logical); |
| 4142 | if (ret) { | 4142 | if (ret) { |
| 4143 | ret = ret > 0 ? 0 : ret; | 4143 | ret = ret > 0 ? 0 : ret; |
| 4144 | goto out; | 4144 | goto out; |
| 4145 | } | 4145 | } |
| 4146 | 4146 | ||
| 4147 | while (len >= PAGE_CACHE_SIZE) { | 4147 | while (len >= PAGE_CACHE_SIZE) { |
| 4148 | index = offset >> PAGE_CACHE_SHIFT; | 4148 | index = offset >> PAGE_CACHE_SHIFT; |
| 4149 | again: | 4149 | again: |
| 4150 | page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); | 4150 | page = find_or_create_page(inode->i_mapping, index, GFP_NOFS); |
| 4151 | if (!page) { | 4151 | if (!page) { |
| 4152 | btrfs_err(fs_info, "find_or_create_page() failed"); | 4152 | btrfs_err(fs_info, "find_or_create_page() failed"); |
| 4153 | ret = -ENOMEM; | 4153 | ret = -ENOMEM; |
| 4154 | goto out; | 4154 | goto out; |
| 4155 | } | 4155 | } |
| 4156 | 4156 | ||
| 4157 | if (PageUptodate(page)) { | 4157 | if (PageUptodate(page)) { |
| 4158 | if (PageDirty(page)) | 4158 | if (PageDirty(page)) |
| 4159 | goto next_page; | 4159 | goto next_page; |
| 4160 | } else { | 4160 | } else { |
| 4161 | ClearPageError(page); | 4161 | ClearPageError(page); |
| 4162 | err = extent_read_full_page(io_tree, page, | 4162 | err = extent_read_full_page(io_tree, page, |
| 4163 | btrfs_get_extent, | 4163 | btrfs_get_extent, |
| 4164 | nocow_ctx->mirror_num); | 4164 | nocow_ctx->mirror_num); |
| 4165 | if (err) { | 4165 | if (err) { |
| 4166 | ret = err; | 4166 | ret = err; |
| 4167 | goto next_page; | 4167 | goto next_page; |
| 4168 | } | 4168 | } |
| 4169 | 4169 | ||
| 4170 | lock_page(page); | 4170 | lock_page(page); |
| 4171 | /* | 4171 | /* |
| 4172 | * If the page has been remove from the page cache, | 4172 | * If the page has been remove from the page cache, |
| 4173 | * the data on it is meaningless, because it may be | 4173 | * the data on it is meaningless, because it may be |
| 4174 | * old one, the new data may be written into the new | 4174 | * old one, the new data may be written into the new |
| 4175 | * page in the page cache. | 4175 | * page in the page cache. |
| 4176 | */ | 4176 | */ |
| 4177 | if (page->mapping != inode->i_mapping) { | 4177 | if (page->mapping != inode->i_mapping) { |
| 4178 | unlock_page(page); | 4178 | unlock_page(page); |
| 4179 | page_cache_release(page); | 4179 | page_cache_release(page); |
| 4180 | goto again; | 4180 | goto again; |
| 4181 | } | 4181 | } |
| 4182 | if (!PageUptodate(page)) { | 4182 | if (!PageUptodate(page)) { |
| 4183 | ret = -EIO; | 4183 | ret = -EIO; |
| 4184 | goto next_page; | 4184 | goto next_page; |
| 4185 | } | 4185 | } |
| 4186 | } | 4186 | } |
| 4187 | 4187 | ||
| 4188 | ret = check_extent_to_block(inode, offset, len, | 4188 | ret = check_extent_to_block(inode, offset, len, |
| 4189 | nocow_ctx_logical); | 4189 | nocow_ctx_logical); |
| 4190 | if (ret) { | 4190 | if (ret) { |
| 4191 | ret = ret > 0 ? 0 : ret; | 4191 | ret = ret > 0 ? 0 : ret; |
| 4192 | goto next_page; | 4192 | goto next_page; |
| 4193 | } | 4193 | } |
| 4194 | 4194 | ||
| 4195 | err = write_page_nocow(nocow_ctx->sctx, | 4195 | err = write_page_nocow(nocow_ctx->sctx, |
| 4196 | physical_for_dev_replace, page); | 4196 | physical_for_dev_replace, page); |
| 4197 | if (err) | 4197 | if (err) |
| 4198 | ret = err; | 4198 | ret = err; |
| 4199 | next_page: | 4199 | next_page: |
| 4200 | unlock_page(page); | 4200 | unlock_page(page); |
| 4201 | page_cache_release(page); | 4201 | page_cache_release(page); |
| 4202 | 4202 | ||
| 4203 | if (ret) | 4203 | if (ret) |
| 4204 | break; | 4204 | break; |
| 4205 | 4205 | ||
| 4206 | offset += PAGE_CACHE_SIZE; | 4206 | offset += PAGE_CACHE_SIZE; |
| 4207 | physical_for_dev_replace += PAGE_CACHE_SIZE; | 4207 | physical_for_dev_replace += PAGE_CACHE_SIZE; |
| 4208 | nocow_ctx_logical += PAGE_CACHE_SIZE; | 4208 | nocow_ctx_logical += PAGE_CACHE_SIZE; |
| 4209 | len -= PAGE_CACHE_SIZE; | 4209 | len -= PAGE_CACHE_SIZE; |
| 4210 | } | 4210 | } |
| 4211 | ret = COPY_COMPLETE; | 4211 | ret = COPY_COMPLETE; |
| 4212 | out: | 4212 | out: |
| 4213 | mutex_unlock(&inode->i_mutex); | 4213 | mutex_unlock(&inode->i_mutex); |
| 4214 | iput(inode); | 4214 | iput(inode); |
| 4215 | return ret; | 4215 | return ret; |
| 4216 | } | 4216 | } |
| 4217 | 4217 | ||
| 4218 | static int write_page_nocow(struct scrub_ctx *sctx, | 4218 | static int write_page_nocow(struct scrub_ctx *sctx, |
| 4219 | u64 physical_for_dev_replace, struct page *page) | 4219 | u64 physical_for_dev_replace, struct page *page) |
| 4220 | { | 4220 | { |
| 4221 | struct bio *bio; | 4221 | struct bio *bio; |
| 4222 | struct btrfs_device *dev; | 4222 | struct btrfs_device *dev; |
| 4223 | int ret; | 4223 | int ret; |
| 4224 | 4224 | ||
| 4225 | dev = sctx->wr_ctx.tgtdev; | 4225 | dev = sctx->wr_ctx.tgtdev; |
| 4226 | if (!dev) | 4226 | if (!dev) |
| 4227 | return -EIO; | 4227 | return -EIO; |
| 4228 | if (!dev->bdev) { | 4228 | if (!dev->bdev) { |
| 4229 | printk_ratelimited(KERN_WARNING | 4229 | printk_ratelimited(KERN_WARNING |
| 4230 | "BTRFS: scrub write_page_nocow(bdev == NULL) is unexpected!\n"); | 4230 | "BTRFS: scrub write_page_nocow(bdev == NULL) is unexpected!\n"); |
| 4231 | return -EIO; | 4231 | return -EIO; |
| 4232 | } | 4232 | } |
| 4233 | bio = btrfs_io_bio_alloc(GFP_NOFS, 1); | 4233 | bio = btrfs_io_bio_alloc(GFP_NOFS, 1); |
| 4234 | if (!bio) { | 4234 | if (!bio) { |
| 4235 | spin_lock(&sctx->stat_lock); | 4235 | spin_lock(&sctx->stat_lock); |
| 4236 | sctx->stat.malloc_errors++; | 4236 | sctx->stat.malloc_errors++; |
| 4237 | spin_unlock(&sctx->stat_lock); | 4237 | spin_unlock(&sctx->stat_lock); |
| 4238 | return -ENOMEM; | 4238 | return -ENOMEM; |
| 4239 | } | 4239 | } |
| 4240 | bio->bi_iter.bi_size = 0; | 4240 | bio->bi_iter.bi_size = 0; |
| 4241 | bio->bi_iter.bi_sector = physical_for_dev_replace >> 9; | 4241 | bio->bi_iter.bi_sector = physical_for_dev_replace >> 9; |
| 4242 | bio->bi_bdev = dev->bdev; | 4242 | bio->bi_bdev = dev->bdev; |
| 4243 | ret = bio_add_page(bio, page, PAGE_CACHE_SIZE, 0); | 4243 | ret = bio_add_page(bio, page, PAGE_CACHE_SIZE, 0); |
| 4244 | if (ret != PAGE_CACHE_SIZE) { | 4244 | if (ret != PAGE_CACHE_SIZE) { |
| 4245 | leave_with_eio: | 4245 | leave_with_eio: |
| 4246 | bio_put(bio); | 4246 | bio_put(bio); |
| 4247 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_WRITE_ERRS); | 4247 | btrfs_dev_stat_inc_and_print(dev, BTRFS_DEV_STAT_WRITE_ERRS); |
| 4248 | return -EIO; | 4248 | return -EIO; |
| 4249 | } | 4249 | } |
| 4250 | 4250 | ||
| 4251 | if (btrfsic_submit_bio_wait(WRITE_SYNC, bio)) | 4251 | if (btrfsic_submit_bio_wait(WRITE_SYNC, bio)) |
| 4252 | goto leave_with_eio; | 4252 | goto leave_with_eio; |
| 4253 | 4253 | ||
| 4254 | bio_put(bio); | 4254 | bio_put(bio); |
| 4255 | return 0; | 4255 | return 0; |
| 4256 | } | 4256 | } |