11 Aug, 2010
40 commits
-
After 97e7449a7ad: "autofs4: fix indirect mount pending expire race" we no
longer assumed that "ino" can be null. The other null checks got removed
but this was one was missed.Signed-off-by: Dan Carpenter
Cc: Ian Kent
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Check return value of put_user() and return -EFAULT if it failed.
Original comment "We did a get user...so assuming mem is ok...is this
bad?" is incorrect because memory can be read only.Signed-off-by: Kulikov Vasiliy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If (len > reslen) we must not call copy_to_user() since kernel buffer is
smaller than we want to copy. Similar code in this file is correct, so
this bug was a typo.Signed-off-by: Kulikov Vasiliy
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add a mutex_unlock missing on the error path.
The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)//
@@
expression E1;
@@* mutex_lock(E1,...);
* mutex_unlock(E1,...);
//Signed-off-by: Julia Lawall
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Pablo Castillo
Signed-off-by: Stanislaw Gruszka
Tested-by: Gregg Lebovitz
Cc: Maxim Levitsky
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
P4080 ESDHC controller does not support 1.8V and 3.0V voltage. but the
host controller capabilities register wrongly set the bits. This patch
adds the workaround to correct the weird voltage setting bits. Only 3.3V
voltage is supported for P4080 ESDHC controller.Signed-off-by: Roy Zang
Cc: Jerry Huang
Cc: Benjamin Herrenschmidt
Cc: Kumar Gala
Cc: Grant Likely
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Roy Zang
Cc: Jerry Huang
Cc: Benjamin Herrenschmidt
Cc: Kumar Gala
Cc: Grant Likely
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add auto CMD12 command support for eSDHC driver. This is needed by P4080
and P1022 for block read/write. Manual asynchronous CMD12 abort operation
causes protocol violations on these silicons.Signed-off-by: Jerry Huang
Signed-off-by: Roy Zang
Cc: Benjamin Herrenschmidt
Cc: Kumar Gala
Cc: Grant Likely
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch adds support for regulator API to sdhci core driver.
Regulators can be used to disable power in suspended state to reduce
dissipated energy.Signed-off-by: Marek Szyprowski
Signed-off-by: Kyungmin Park
Acked-by: Mark Brown
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
On some Samsung SoCs not all SDHCI controllers have card detect (CD) line.
For some embedded designs it is not even needed, because ususally the
device (like SDIO flash memory or wifi controller) is permanently wired to
the controller. There are also systems which have a card detect line
connected to some of the external interrupt lines or the presence of the
card depends on some other actions (like enabling a power regulator).This patch adds support for all these cases. The following card detection
methods are possible:1. internal sdhci host card detect line
2. external event
3. external gpio interrupt
4. no card detect line, controller will poll for the card
5. no card detect line, card is permanently wired to the controller
(once detected host won't poll it any more)By default, all existing code would use method #1, what is compatible with
the previous version of the driver.In case of external event, two callbacks must be provided in platdata:
ext_cd_init and ext_cd_cleanup. Both of them get a callback to a function
that notifies the s3c-sdhci host contoller as their argument. That
callback function should be called from the even dispatcher to let host
notice the card insertion/removal.In case of external gpio interrupt, a gpio pin number must be provided in
platdata (ext_cd_gpio parameter), as well as the information about the
polarity of that gpio pin (ext_cd_gpio_invert). By default
(ext_cd_gpio_invert == 0) gpio value 0 means 'card has been removed', but
this can be changed to 'card has been removed' when ext_cd_gpio_invert ==
1.This patch adds all required changes to sdhci-s3c driver.
Signed-off-by: Marek Szyprowski
Signed-off-by: Kyungmin Park
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch enables SDHCI_QUIRK_NO_HISPD_BIT on Samsung SDHCI driver. This
solves detection problems with some external SD cards. This change has
been tested on S5PC100 and S5PC110. It has no inpact on driver speed.Signed-off-by: Kyungmin Park
Signed-off-by: Marek Szyprowski
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
S3C SDHCI host controller can change the source for generating mmc clock.
By default host bus clock is used, what causes some problems on machines
with 133MHz bus, because the SDHCI divider cannot be as high get proper
clock value for identification mode. This is not a problem for the
controller, because it can generate lower frequencies from other clock
sources. This patch changes sdhci driver to use get_min_clock() call if
it has been provided.This fixes the flood of the following warnings on Samsung S5PV210 SoCs:
mmc0: Minimum clock frequency too high for identification modeSigned-off-by: Marek Szyprowski
Signed-off-by: Kyungmin Park
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
On Samsung's SDMMC hosts the timeout clock is derivied from the SD Clock
which is set dynamically. So checked SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK
quirk and removed 'sdhci_s3c_get_timeout_clk' callback which doesn't need
any more.Signed-off-by: Hyuk Lee
Signed-off-by: Kukjin Kim
Cc: Ben Dooks
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If other informative interrupts are enabled for the DMA channel used by
hsmmc, those are incorrectly treated as block completion. This patch lets
only the block completion interrupt to be processed.Signed-off-by: Venkatraman S
Acked-by: Madhusudhan Chikkature
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
CONFIG_MMC_MSM7X00A_RESUME_IN_WQ and CONFIG_MMC_EMBEDDED_SDIO don't exist
in Kconfig and is never defined anywhere else, therefore removing all
references for it from the source code.Signed-off-by: Christian Dietrich
Acked-by: David Brown
Cc: Daniel Walker
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Pandora has TI WL1251 attached on MMC3, which is non-standard SDIO chip.
Make use MMC_QUIRK_NONSTD_SDIO to tell SDIO core about it.Signed-off-by: Grazvydas Ignotas
Cc: Adrian Hunter
Cc: Tony Lindgren
Cc: Bob Copeland
Cc: Kalle Valo
Cc: Madhusudhan Chikkature
Cc: Kishore Kadiyala
Cc: Russell King
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This will allow us to set up special cards in machine drivers just after
they are detected by MMC core.Signed-off-by: Grazvydas Ignotas
Cc: Adrian Hunter
Cc: Tony Lindgren
Cc: Bob Copeland
Cc: Kalle Valo
Cc: Madhusudhan Chikkature
Cc: Kishore Kadiyala
Cc: Russell King
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
There are some chips (like TI WL12xx series) that can be interfaced over
SDIO but don't support the SDIO specification, meaning that they are
missing CIA (Common I/O Area) with all it's registers. Current Linux SDIO
implementation relies on those registers to identify and configure the
card, so non-standard cards can not function and cause lots of warnings
from the core when it reads invalid data from non-existent registers.After this patch, init_card() host callback can now set new quirk
MMC_QUIRK_NONSTD_SDIO, which means that SDIO core should not try to access
any standard SDIO registers and rely on init_card() to fill all SDIO
structures instead. As those cards are usually embedded chips, all the
required information can be obtained from machine board files by the host
driver when it's called through init_card() callback.Signed-off-by: Grazvydas Ignotas
Cc: Adrian Hunter
Cc: Tony Lindgren
Cc: Bob Copeland
Cc: Kalle Valo
Cc: Madhusudhan Chikkature
Cc: Kishore Kadiyala
Cc: Russell King
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
There's nothing special, just SoC-specific ops and quirks.
Signed-off-by: Anton Vorontsov
Cc: Ben Dooks
Cc: Richard R?jfors
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Due to build system limitations, intermediate and final objects can't have
the same names. And as we're going to start building SoC-specific
objects, let's rename the module to sdhci-platform, into which we'll link
sdhci-pltfm and SoC-specifc objects.There should be no issue in renaming as the driver uses modalias
mechanism.This is exactly the same approach as in sdhci-of driver.
Signed-off-by: Anton Vorontsov
Cc: Ben Dooks
Cc: Richard R?jfors
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Sometimes want to place SoC-specific parts alongside with the generic
driver, and to do so, we have to switch the driver over to the module
device table matching.Note that drivers/mmc/host/sdhci-pltfm.h is so far empty, but it'll hold
SoC-specific driver data handlers soon.Signed-off-by: Anton Vorontsov
Cc: Ben Dooks
Cc: Richard R?jfors
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This patch improves low speeds for SD cards.
OMAP-MMC controller's can support maximum bus width of '8'. when bus
width is mentioned as "8" in controller data,the SD stack will check
whether bus width is "4" and if not it will set bus width to "1" and there
by degrading performance. This patch fixes the issue and improves the
performance of SD cards.Signed-off-by: Kishore Kadiyala
Signed-off-by: Venkatraman S
Signed-off-by: Nishanth Menon
Acked-by: Madhusudhan Chikkature
Tested-by: Jarkko Nikula
Cc: Adrian Hunter
Cc: Matt Fleming
Cc: Tony Lindgren
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
A non-zero value of SEC_COUNT does not indicate that the card is sector
addressed. According to the MMC specification, cards with a density
greater than 2GiB are sector addressed.Acked-by: Linus Walleij
Signed-off-by: Hanumath Prasad
Signed-off-by: Rabin Vincent
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some host controllers can set mmc->caps before sdhci_add_host().
Signed-off-by: Kyungmin Park
Cc: Grant Likely
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some host controllers such as s5pc110 support the WIDE8 feature.
Signed-off-by: Kyungmin Park
Cc: Grant Likely
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
When using QUIRK_NONSTANDARD_CLOCK, it checks the set_clock() function
which is not used actually. So delete it.Signed-off-by: Kyungmin Park
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The current way of disabling it is not well tested by vendor and has all
kinds of bugs that show up on resume from ram/disk. A very good example
is a dead SDHCI controller.Old way of disabling is still supported by continuing to use
CONFIG_MMC_RICOH_MMC.Based on 'http://list.drzeus.cx/pipermail/sdhci-devel/2007-December/002085.html'
Therefore most of the credit for this goes to Andrew de QuinceySigned-off-by: Maxim Levitsky
Cc: Andrew de Quincey
Acked-by: Philip Langdale
Cc: "Rafael J. Wysocki"
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
If you don't use CONFIG_MMC_UNSAFE_RESUME, as soon as you attempt to
suspend, the card will be removed, therefore this patch doesn't change the
behavior of this option.However the removal will be done by pm notifier, which runs while
userspace is still not frozen and thus can freely use del_gendisk, without
the risk of deadlock which would happen otherwise.Card detect workqueue is now disabled while userspace is frozen, Therefore
if you do use CONFIG_MMC_UNSAFE_RESUME, and remove the card during
suspend, the removal will be detected as soon as userspace is unfrozen,
again at the moment it is safe to call del_gendisk.Tested with and without CONFIG_MMC_UNSAFE_RESUME with suspend and hibernate.
[akpm@linux-foundation.org: clean up function prototype]
[akpm@linux-foundation.org: fix CONFIG_PM-n linkage, small cleanups]
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Maxim Levitsky
Cc: David Brownell
Cc: Alan Stern
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Michal Miroslaw
Cc: Adrian Hunter
Cc: Chris Ball
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
This series adds support for SD combo cards to MMC/SD driver stack.
SD combo consists of SD memory and SDIO parts in one package. Since the
parts have a separate SD command sets, after initialization, they can be
treated as independent cards on one bus.Changes are divided into two patches. First is just moving initialization
code around so that SD memory part init can be called from SDIO init.
Second patch is a proper change enabling SD memory along SDIO. I tried to
move as much no-op changes to the first patch so that it's easier to
follow the required changes to initialization flow for SDIO cards.This is based on Simplified SDIO spec v.2.00. The init sequence is
slightly modified to follow current SD memory init implementation.
Command sequences, assuming SD memory and SDIO indeed ignore unknown
commands, are the same as before for both parts.This patch:
Prepare for SD-combo (IO+mem) support by splitting SD memory
card init and related functions.Signed-off-by: Michal Miroslaw
Cc: Adrian Hunter
Cc: Chris Ball
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The eMMC spec 4.4 and 4.3 + additional feature chips has CSD structure
version 3 and version 3 have to check the CSD_STRUCTURE byte in the
EXT_CSD register.Also fix EXT_CSD revision message.
[akpm@linux-foundation.org: fix comment, per Chris Ball]
Signed-off-by: Kyungmin Park
Cc: Adrian Hunter
Cc: Chris Ball
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Use kmalloc() to allocate fdmem if possible.
vmalloc() is used as a fallback solution for fdmem allocation. A new
helper function __free_fdtable() is introduced to reduce the lines of
code.A potential bug, vfree() a memory allocated by kmalloc(), is fixed.
[akpm@linux-foundation.org: use __GFP_NOWARN, uninline alloc_fdmem() and free_fdmem()]
Signed-off-by: Changli Gao
Cc: Alexander Viro
Cc: Jiri Slaby
Cc: "Paul E. McKenney"
Cc: Alexey Dobriyan
Cc: Ingo Molnar
Cc: Peter Zijlstra
Cc: Avi Kivity
Cc: Tetsuo Handa
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Dmitry Torokhov
Acked-by: Arnd Bergmann
Acked-by: John Kacur
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The O_* bit numbers are defined in 20+ arch/*, and can silently overlap.
Add a compile time check to ensure the uniqueness as suggested by David
Miller.Signed-off-by: Wu Fengguang
Cc: David Miller
Cc: Stephen Rothwell
Cc: Al Viro
Cc: Christoph Hellwig
Cc: Eric Paris
Cc: Roland Dreier
Cc: Jamie Lokier
Cc: Andreas Schwab
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Improve the description of fget_light(), which is currently incorrect
about needing a prior refcnt (judging by the way it is actually used).Signed-off-by: Tony Battersby
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add to so that types are explicitly
defined:
linux/virtio_9p.h:15: found __[us]{8,16,32,64} type without #includeSigned-off-by: Randy Dunlap
Cc: Eric Van Hensbergen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Andy Shevchenko
Cc: Adaptec OEM Raid Solutions
Cc: "James E.J. Bottomley"
Cc: James Smart
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Use memdup_user when user data is immediately copied into the
allocated region.The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)//
@@
expression from,to,size,flag;
position p;
identifier l1,l2;
@@- to = \(kmalloc@p\|kzalloc@p\)(size,flag);
+ to = memdup_user(from,size);
if (
- to==NULL
+ IS_ERR(to)
|| ...) {
}
- if (copy_from_user(to, from, size) != 0) {
-
- }
//Signed-off-by: Julia Lawall
Cc: Doug Gilbert
Cc: Boaz Harrosh
Cc: James Bottomley
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
A couple of scsi drivers define a BIT() macro, duplicating the one in
bitops.h.Cc: Jing Huang
Cc: Robert Love
Cc: James Bottomley
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
bnx2i should depend on NET since it selects SCSI_ISCSI_ATTRS, which
depends on NET.Also move the dependencies together.
The "depends" change fixes multiple build errors when CONFIG_NET is
not enabled:
ERROR: "skb_trim" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "netlink_kernel_create" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "netlink_kernel_release" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "skb_pull" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "init_net" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "__alloc_skb" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "netlink_broadcast" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "kfree_skb" [drivers/scsi/scsi_transport_iscsi.ko] undefined!
ERROR: "skb_put" [drivers/scsi/scsi_transport_iscsi.ko] undefined!Signed-off-by: Randy Dunlap
Cc: Anil Veerabhadrappa
Cc: James Bottomley
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
