10 Jul, 2016
1 commit
-
This function compiles to 2554 bytes of machine code.
In C, the function is almost 200 lines long.It has only one callsite, but forced inlining that much code
makes gcc generate significantly worse code. Let gcc itself decide
what to do.Signed-off-by: Denys Vlasenko
CC: David Woodhouse
CC: Dan Carpenter
CC: Artem Bityutskiy
CC: linux-mtd@lists.infradead.org
CC: linux-kernel@vger.kernel.org
Signed-off-by: Brian Norris
04 Apr, 2016
1 commit
-
When map_word gets too large, we use a lot of kernel stack, and for
MTD_MAP_BANK_WIDTH_32, this means we use more than the recommended
1024 bytes in a number of functions:drivers/mtd/chips/cfi_cmdset_0020.c: In function 'cfi_staa_write_buffers':
drivers/mtd/chips/cfi_cmdset_0020.c:651:1: warning: the frame size of 1336 bytes is larger than 1024 bytes [-Wframe-larger-than=]
drivers/mtd/chips/cfi_cmdset_0020.c: In function 'cfi_staa_erase_varsize':
drivers/mtd/chips/cfi_cmdset_0020.c:972:1: warning: the frame size of 1208 bytes is larger than 1024 bytes [-Wframe-larger-than=]
drivers/mtd/chips/cfi_cmdset_0001.c: In function 'do_write_buffer':
drivers/mtd/chips/cfi_cmdset_0001.c:1835:1: warning: the frame size of 1240 bytes is larger than 1024 bytes [-Wframe-larger-than=]This can be avoided if all operations on the map word are done
indirectly and the stack gets reused between the calls. We can
mostly achieve this by selecting MTD_COMPLEX_MAPPINGS whenever
MTD_MAP_BANK_WIDTH_32 is set, but for the case that no other
bank width is enabled, we also need to use a non-constant
map_bankwidth() to convince the compiler to use less stack.Signed-off-by: Arnd Bergmann
[Brian: this patch mostly achieves its goal by forcing
MTD_COMPLEX_MAPPINGS (and the accompanying indirection) for 256-bit
mappings; the rest of the change is mostly a wash, though it helps
reduce stack size slightly. If we really care about supporting
256-bit mappings though, we should consider rewriting some of this
code to avoid keeping and assigning so many 256-bit objects on the
stack.]
Signed-off-by: Brian Norris
07 Jan, 2016
1 commit
-
kmalloc needs to be handled when failing in memory pressure.
Also, it has memory leak in error routine.Signed-off-by: Insu Yun
Signed-off-by: Brian Norris
19 Dec, 2015
1 commit
-
Use kernel.h macro definition.
Thanks to Julia Lawall for Coccinelle scripting support.
Signed-off-by: Fabian Frederick
Signed-off-by: Brian Norris
01 Dec, 2015
1 commit
-
MTD allows compile-time configuration of the possible CFI geometry
settings that are allowed by the kernel, but that includes a couple of
invalid configurations, where no bank width or no interleave setting
is allowed. These are then caught with a compile-time warning:include/linux/mtd/cfi.h:76:2: warning: #warning No CONFIG_MTD_CFI_Ix selected. No NOR chip support can work.
include/linux/mtd/map.h:145:2: warning: #warning "No CONFIG_MTD_MAP_BANK_WIDTH_xx selected. No NOR chip support can work"This is a bit annoying for randconfig tests, and can be avoided if
we change the Kconfig logic to always select the simplest configuration
when no other one is enabled.Signed-off-by: Arnd Bergmann
Signed-off-by: Brian Norris
29 May, 2015
1 commit
-
Commit 4612c715a6ea ("mtd: cfi: deinline large functions") moved some
code into the cfi_util library without creating a new dependency. So we
can get build failures like the following, when CONFIG_MTD_JEDECPROBE=y
and CONFIG_MTD_CFI_UTIL=m.drivers/built-in.o: In function `jedec_read_id':
>> jedec_probe.c:(.text+0x187ed8): undefined reference to `cfi_build_cmd_addr'
drivers/built-in.o: In function `jedec_read_mfr':
>> jedec_probe.c:(.text+0x187f4a): undefined reference to `cfi_build_cmd_addr'
drivers/built-in.o: In function `jedec_reset':
>> jedec_probe.c:(.text+0x187fe0): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x188004): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x18802b): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x18804e): undefined reference to `cfi_send_gen_cmd'
drivers/built-in.o: In function `jedec_probe_chip':
>> jedec_probe.c:(.text+0x188130): undefined reference to `cfi_build_cmd_addr'
>> jedec_probe.c:(.text+0x18814d): undefined reference to `cfi_build_cmd_addr'
>> jedec_probe.c:(.text+0x1881dc): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x188203): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x18822d): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x1884c0): undefined reference to `cfi_send_gen_cmd'
>> jedec_probe.c:(.text+0x1884e7): undefined reference to `cfi_send_gen_cmd'
drivers/built-in.o:jedec_probe.c:(.text+0x188511): more undefined references to `cfi_send_gen_cmd' follow
drivers/built-in.o: In function `jedec_probe_chip':
>> jedec_probe.c:(.text+0x188618): undefined reference to `cfi_build_cmd'So let's express the dependency properly.
Reported-by: kbuild test robot
Signed-off-by: Brian Norris
28 May, 2015
2 commits
-
In do_otp_write we must initialize the variable datum before calling
map_word_load_partial. Otherwise the upper bits of datum may be undefined,
which later causes problems in chip_good called by do_write_oneword.Signed-off-by: Christian Riesch
Signed-off-by: Brian Norris -
With this .config: http://busybox.net/~vda/kernel_config,
after uninlining these functions have sizes and callsite counts
as follows:cfi_udelay(): 74 bytes, 26 callsites
cfi_send_gen_cmd(): 153 bytes, 95 callsites
cfi_build_cmd(): 274 bytes, 123 callsites
cfi_build_cmd_addr(): 49 bytes, 15 callsites
cfi_merge_status(): 230 bytes, 3 callsitesReduction in code size is about 50,000:
text data bss dec hex filename
85842882 22294584 20627456 128764922 7accbfa vmlinux.before
85789648 22294616 20627456 128711720 7abfc28 vmlinuxSigned-off-by: Denys Vlasenko
CC: Dan Carpenter
CC: Jingoo Han
CC: Brian Norris
CC: Aaron Sierra
CC: Artem Bityutskiy
CC: David Woodhouse
CC: linux-mtd@lists.infradead.org
CC: linux-kernel@vger.kernel.org
Signed-off-by: Brian Norris
31 Mar, 2015
1 commit
-
These lines were all indented one tab more than they should be.
Signed-off-by: Dan Carpenter
Signed-off-by: Brian Norris
10 Jan, 2015
2 commits
-
UBI needs to know the physical erase block size, even on read-only
devices, since it defines the on-device layout. Use a device-tree
provided value to support previously written UBI on read-only NOR.UBI also needs a non-zero writebufsize, so we set it to one.
Note: This was implemented because hardware write-protected CFI
NOR cannot be probed for the physical erase block size.Signed-off-by: Joe Schultz
Signed-off-by: Aaron Sierra
[Brian: removed unneeded #ifdef, note 'optional' erase-size property]
Signed-off-by: Brian Norris -
Signed-off-by: Alessio Igor Bogani
Signed-off-by: Brian Norris
26 Nov, 2014
1 commit
-
The functions kfree() and pci_dev_put() test whether their argument is NULL
and then return immediately. Thus the test around the call is not needed.This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring
Signed-off-by: Brian Norris
30 Oct, 2014
1 commit
-
After '#echo mem > /sys/power/state' some devices can not be properly resumed
because apparently the MTD Partition Configuration Register has been reset
to default thus the rootfs cannot be mounted cleanly on resume.
An example of this can be found in the SA-1100 Developer's Manual at 9.5.3.3
where the second step of the Sleep Shutdown Sequence is described:
"An internal reset is applied to the SA-1100. All units are reset...".As workaround we refresh the PCR value as done initially on chip setup.
This behavior and the fix are confirmed by our tests done on 2 different Zaurus
collie units with kernel 3.17.Fixes: 812c5fa82bae: ("mtd: cfi_cmdset_0001.c: add support for Sharp LH28F640BF NOR")
Signed-off-by: Dmitry Eremin-Solenikov
Signed-off-by: Andrea Adami
Cc: # 3.16+
Signed-off-by: Brian Norris
20 Aug, 2014
2 commits
-
The variable 'retries' is never modified, so if the reset operation
never is going to complete, we'll get stuck in an infinite loop.It looks like the intention was to decrement 'retries' on every loop.
Untested.Caught by Coverity.
Signed-off-by: Brian Norris
16 Aug, 2014
1 commit
-
Coverity CID 1230633
Signed-off-by: Brian Norris
Tested-by: Christian Riesch
21 Jul, 2014
1 commit
-
Linux 3.16-rc6
17 Jul, 2014
1 commit
-
For some NOR flashes, the size of the buffer program has been increased
from 256 bytes to 512 bytes, and so 2ms maximum timeout can may not be
sufficient for all different vendor's NOR flash. There is maximum
timeout information in the CFI area, so we instead of picking a fixed
value, we can calculate this according to the standard CFI parameters
parsed at probe time. If we haven't probed this information, or it is
smaller than 2000us, then specify a minimum value 2000us.Tested with Micron JS28F512M29EWx and Micron MT28EW512ABA flash devices.
Signed-off-by: Bean Huo
[Brian: fix up comments, use 'max()']
Signed-off-by: Brian Norris
14 Jul, 2014
1 commit
-
This family of chips was long ago supported by the pre-cfi driver.
CFI code tested on several Zaurus SL-5500 (Collie) 2x16 on 32 bit bus.Function is_LH28F640BF() mimics is_m29ew() from cmdset_0002.c
Buffer write fixes as seen in 2007 patch c/o
Anti Sullin artecdesign.ee>
http://comments.gmane.org/gmane.linux.ports.arm.kernel/36733[Brian: this patch is semi-urgent, because the following patch switches
to using CFI detection for a chip which (until now) is unsupported by
the CFI driver9218310 ARM: 8084/1: sa1100: collie: revert back to cfi_probe
]Signed-off-by: Andrea Adami
Signed-off-by: Brian Norris
12 Jul, 2014
4 commits
-
This patch adds support for the locking of the one time
programmable (OTP) memory of Micron M29EW devices.Signed-off-by: Christian Riesch
Signed-off-by: Brian Norris -
This patch adds support for writing the one time programmable (OTP)
memory of Micron M29EW devices.Signed-off-by: Christian Riesch
Signed-off-by: Brian Norris -
When the one time programmable (OTP) memory region is entered by
issuing the 0xaa/0x55/0x88 command, the OTP memory occupies the
addresses which are normally used by the first sector of the regular
flash memory. This patch therefore invalidates cache for this
addresses after entering/exiting OTP memory.This patch also moves the code into separate functions.
Signed-off-by: Christian Riesch
Signed-off-by: Brian Norris -
The Micron M29EW has a 256 byte one time programmable (OTP) memory.
This patch adds support for reading this memory. This support will be
extended for locking and writing in subsequent patches.Signed-off-by: Christian Riesch
Signed-off-by: Brian Norris
21 May, 2014
2 commits
-
Enhances the help for the CFI command set choices.
Signed-off-by: Philippe De Muyter
Signed-off-by: Brian Norris -
The break statements should be indented another tab.
Signed-off-by: Dan Carpenter
Reviewed-by: Jingoo Han
Signed-off-by: Brian Norris
11 Mar, 2014
4 commits
-
Signed-off-by: Christian Riesch
Cc: Artem Bityutskiy
Signed-off-by: Brian Norris -
The site-specific OOM messages are unnecessary, because they
duplicate the MM subsystem generic OOM message.Signed-off-by: Jingoo Han
Signed-off-by: Brian Norris -
The site-specific OOM messages are unnecessary, because they
duplicate the MM subsystem generic OOM message.Signed-off-by: Jingoo Han
Signed-off-by: Brian Norris -
None of these files are actually using any __init type directives
and hence don't need to include . Most are just a
left over from __devinit and __cpuinit removal, or simply due to
code getting copied from one driver to the next.Cc: David Woodhouse
Cc: Brian Norris
Cc: linux-mtd@lists.infradead.org
Signed-off-by: Paul Gortmaker
[Brian: dropped one incorrect hunk]
Signed-off-by: Brian Norris
06 Aug, 2013
3 commits
-
Zaurus 5500 contains 2 LH28F640BFHE-PTTL90 (64M 4Mx16) and
the LH28F640BFHE-PTTL90.pdf datasheet available on the net shows
the exact erasesize and the OTP support.
At the moment only jedec_probe can discover the chip and
the NOR is mounted read only probably because of wrong vpp.Signed-off-by: Andrea Adami
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
This reduces the size of the stack frame when calling request_module().
Performing the sprintf before the call is not needed.Signed-off-by: Kees Cook
Acked-by: Rusty Russell
Reviewed-by: James Hogan
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse -
Print out the @adr when the write timeout occurs.
This is useful to check if the write timeouts occur at the
same address.Signed-off-by: Huang Shijie
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse
10 May, 2013
1 commit
-
Pull MTD update from David Woodhouse:
- Lots of cleanups from Artem, including deletion of some obsolete
drivers- Support partitions larger than 4GiB in device tree
- Support for new SPI chips
* tag 'for-linus-20130509' of git://git.infradead.org/linux-mtd: (83 commits)
mtd: omap2: Use module_platform_driver()
mtd: bf5xx_nand: Use module_platform_driver()
mtd: denali_dt: Remove redundant use of of_match_ptr
mtd: denali_dt: Change return value to fix smatch warning
mtd: denali_dt: Use module_platform_driver()
mtd: denali_dt: Fix incorrect error check
mtd: nand: subpage write support for hardware based ECC schemes
mtd: omap2: use msecs_to_jiffies()
mtd: nand_ids: use size macros
mtd: nand_ids: improve LEGACY_ID_NAND macro a bit
mtd: add 4 Toshiba nand chips for the full-id case
mtd: add the support to parse out the full-id nand type
mtd: add new fields to nand_flash_dev{}
mtd: sh_flctl: Use of_match_ptr() macro
mtd: gpio: Use of_match_ptr() macro
mtd: gpio: Use devm_kzalloc()
mtd: davinci_nand: Use of_match_ptr()
mtd: dataflash: Use of_match_ptr() macro
mtd: remove h720x flash support
mtd: onenand: remove OneNAND simulator
...
05 Apr, 2013
1 commit
-
The MTD subsystem has historically tried to be as configurable as possible. The
side-effect of this is that its configuration menu is rather large, and we are
gradually shrinking it. For example, we recently merged partitions support with
the mtdcore.This patch does the next step - it merges the mtdchar module to mtdcore. And in
this case this is not only about eliminating too fine-grained separation and
simplifying the configuration menu. This is also about eliminating seemingly
useless kernel module.Indeed, mtdchar is a module that allows user-space making use of MTD devices
via /dev/mtd* character devices. If users do not enable it, they simply cannot
use MTD devices at all. They cannot read or write the flash contents. Is it a
sane and useful setup? I believe not. And everyone just enables mtdchar.Having mtdchar separate is also a little bit harmful. People sometimes miss the
fact that they need to enable an additional configuration option to have
user-space MTD interfaces, and then they wonder why on earth the kernel does
not allow using the flash? They spend time asking around.Thus, let's just get rid of this module and make it part of mtd core.
Note, mtdchar had additional configuration option to enable OTP interfaces,
which are present on some flashes. I removed that option as well - it saves a
really tiny amount space.[dwmw2: Strictly speaking, you can mount file systems on MTD devices just
fine without the mtdchar (or mtdblock) devices; you just can't do
other manipulations directly on the underlying device. But still I
agree that it makes sense to make this unconditional. And Yay! we
get to kill off an instance of checking CONFIG_foo_MODULE, which is
an abomination that should never happen.]Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse
15 Mar, 2013
1 commit
-
We have CONFIG_SYMBOL_PREFIX, which three archs define to the string
"_". But Al Viro broke this in "consolidate cond_syscall and
SYSCALL_ALIAS declarations" (in linux-next), and he's not the first to
do so.Using CONFIG_SYMBOL_PREFIX is awkward, since we usually just want to
prefix it so something. So various places define helpers which are
defined to nothing if CONFIG_SYMBOL_PREFIX isn't set:1) include/asm-generic/unistd.h defines __SYMBOL_PREFIX.
2) include/asm-generic/vmlinux.lds.h defines VMLINUX_SYMBOL(sym)
3) include/linux/export.h defines MODULE_SYMBOL_PREFIX.
4) include/linux/kernel.h defines SYMBOL_PREFIX (which differs from #7)
5) kernel/modsign_certificate.S defines ASM_SYMBOL(sym)
6) scripts/modpost.c defines MODULE_SYMBOL_PREFIX
7) scripts/Makefile.lib defines SYMBOL_PREFIX on the commandline if
CONFIG_SYMBOL_PREFIX is set, so that we have a non-string version
for pasting.(arch/h8300/include/asm/linkage.h defines SYMBOL_NAME(), too).
Let's solve this properly:
1) No more generic prefix, just CONFIG_HAVE_UNDERSCORE_SYMBOL_PREFIX.
2) Make linux/export.h usable from asm.
3) Define VMLINUX_SYMBOL() and VMLINUX_SYMBOL_STR().
4) Make everyone use them.Signed-off-by: Rusty Russell
Reviewed-by: James Hogan
Tested-by: James Hogan (metag)
03 Mar, 2013
1 commit
-
Pull MTD update from David Woodhouse:
"Fairly unexciting MTD merge for 3.9:- misc clean-ups in the MTD command-line partitioning parser
(cmdlinepart)
- add flash locking support for STmicro chips serial flash chips, as
well as for CFI command set 2 chips.
- new driver for the ELM error correction HW module found in various
TI chips, enable the OMAP NAND driver to use the ELM HW error
correction
- added number of new serial flash IDs
- various fixes and improvements in the gpmi NAND driver
- bcm47xx NAND driver improvements
- make the mtdpart module actually removable"* tag 'for-linus-20130301' of git://git.infradead.org/linux-mtd: (45 commits)
mtd: map: BUG() in non handled cases
mtd: bcm47xxnflash: use pr_fmt for module prefix in messages
mtd: davinci_nand: Use managed resources
mtd: mtd_torturetest can cause stack overflows
mtd: physmap_of: Convert device allocation to managed devm_kzalloc()
mtd: at91: atmel_nand: for PMECC, add code to check the ONFI parameter ECC requirement.
mtd: atmel_nand: make pmecc-cap, pmecc-sector-size in dts is optional.
mtd: atmel_nand: avoid to report an error when lookup table offset is 0.
mtd: bcm47xxsflash: adjust names of bus-specific functions
mtd: bcm47xxpart: improve probing of nvram partition
mtd: bcm47xxpart: add support for other erase sizes
mtd: bcm47xxnflash: register this as normal driver
mtd: bcm47xxnflash: fix message
mtd: bcm47xxsflash: register this as normal driver
mtd: bcm47xxsflash: write number of written bytes
mtd: gpmi: add sanity check for the ECC
mtd: gpmi: set the Golois Field bit for mx6q's BCH
mtd: devices: elm: Removes literals in elm DT node
mtd: gpmi: fix a dereferencing freed memory error
mtd: fix the wrong timeo for panic_nand_wait()
...
04 Feb, 2013
1 commit
-
Currently cfi_cmdset_0002.c does not support PPB locking of sectors. This
patch adds support for this locking/unlocking mechanism. It is needed on
some platforms, since newer U-Boot versions do support this PPB locking
and protect for example their environment sector(s) this way.This PPB locking/unlocking will be enabled for all devices supported by
cfi_cmdset_0002 reporting 8 in the CFI word 0x49 (Sector Protect/Unprotect
scheme).Please note that PPB locking does support sector-by-sector locking. But
the whole chip can only be unlocked together. So unlocking one sector
will automatically unlock all sectors of this device. Because of this
chip limitation, the PPB unlocking function saves the current locking
status of all sectors before unlocking the whole device. After unlocking
the saved locking status is re-configured. This way only the addressed
sectors will be unlocked.To selectively enable this advanced sector protection mechanism, the
device-tree property "use-advanced-sector-protection" has been created.
To enable support for this locking this property needs to be present in the
flash DT node. E.g.:nor_flash@0,0 {
compatible = "amd,s29gl256n", "cfi-flash";
bank-width = ;
use-advanced-sector-protection;
...Tested with Spansion S29GL512S10THI and Micron JS28F512M29EWx flash
devices.Signed-off-by: Stefan Roese
Tested-by: Holger Brunck
Signed-off-by: Artem Bityutskiy
12 Jan, 2013
1 commit
-
The CONFIG_EXPERIMENTAL config item has not carried much meaning for a
while now and is almost always enabled by default. As agreed during the
Linux kernel summit, remove it from any "depends on" lines in Kconfigs.CC: David Woodhouse
Signed-off-by: Kees Cook
Acked-by: Artem Bityutskiy
03 Dec, 2012
1 commit
-
When working on a problem with some flash chips that lock up during
write-buffer operations, I think there may be a bug in the linux
handling of chips using cfi_cmdset_0002.c.The datasheets I have found for a number of these chips all specify that
when aborting a write-buffer command, it is not enough to use the
standard reset. Rather a "write-to-buffer-reset command" is needed.
This command is quite similar for all chips, the main variance seem to
be if the final 0xF0 can go to any address or must go to addr_unlock1.The bug is then in the recovery handling when timing out at the end of
do_write_buffer, where using the normal reset command is not sufficient.Without this change, if the write-buffer command fails then any
following operations on the flash also fail.Signed-off-by: Harald Nordgard-Hansen
Signed-off-by: Artem Bityutskiy
29 Sep, 2012
1 commit
-
Unlocking may take up to 1.4 seconds on some Intel flashes. So
lets use a max. of 1.5 seconds (1500ms) as timeout.See "Clear Block Lock-Bits Time" on page 40 in
"3 Volt Intel StrataFlash Memory" 28F128J3,28F640J3,28F320J3 manual
from February 2003This patch also fixes some other problems with this timeout:
- Don't use HZ in timeout "calculation"!
While testing we noticed that an unlocking timeout occured with
HZ=1000 and didn't occur with HZ=300. This was because the
timeout parameter was calculated differently depending on the
HZ value. Now a fixed value of 1500ms is used.- The last parameter of WAIT_TIMEOUT (defined to
inval_cache_and_wait_for_operation) has to be passed in
micro-seconds. So multiply the ms value with 1000 and not 100
to calculate this value.- Use variable name "mdelay" instead of misleading "udelay".
Signed-off-by: Stefan Roese
Tested-by: Stephan Gatzka
Signed-off-by: Artem Bityutskiy
Signed-off-by: David Woodhouse