deliverable/linux.git
13 years agoMerge branch 'for-2.6.38/core' into for-next
Jens Axboe [Mon, 20 Dec 2010 08:17:35 +0000 (09:17 +0100)] 
Merge branch 'for-2.6.38/core' into for-next

13 years agoFix compile warnings due to missing removal of a 'ret' variable
Jens Axboe [Mon, 20 Dec 2010 08:15:19 +0000 (09:15 +0100)] 
Fix compile warnings due to missing removal of a 'ret' variable

Commit a8adbe3 forgot to remove the return variable, kill it.

drivers/block/loop.c: In function 'lo_splice_actor':
drivers/block/loop.c:398: warning: unused variable 'ret'
[...]
fs/nfsd/vfs.c: In function 'nfsd_splice_actor':
fs/nfsd/vfs.c:848: warning: unused variable 'ret'

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoMerge branch 'for-2.6.38/core' into for-next
Jens Axboe [Fri, 17 Dec 2010 08:00:46 +0000 (09:00 +0100)] 
Merge branch 'for-2.6.38/core' into for-next

13 years agofs/block: type signature of major_to_index(int) to major_to_index(unsigned)
Yang Zhang [Fri, 17 Dec 2010 08:00:18 +0000 (09:00 +0100)] 
fs/block: type signature of major_to_index(int) to major_to_index(unsigned)

The major/minor device numbers are always defined and used as `unsigned'.

Signed-off-by: Yang Zhang <kthreadd@gmail.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoblock: convert !IS_ERR(p) && p to !IS_ERR_NOR_NULL(p)
Yang Zhang [Fri, 17 Dec 2010 07:58:36 +0000 (08:58 +0100)] 
block: convert !IS_ERR(p) && p to !IS_ERR_NOR_NULL(p)

Signed-off-by: Yang Zhang <kthreadd@gmail.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agocfq-iosched: don't check cfqg in choose_service_tree()
Gui Jianfeng [Fri, 17 Dec 2010 07:57:14 +0000 (08:57 +0100)] 
cfq-iosched: don't check cfqg in choose_service_tree()

When cfq_choose_cfqg() is called in select_queue(), there must be at least one
backlogged CFQ queue waiting for dispatching, hence there must be at least one
backlogged CFQ group on service tree. So we never call choose_service_tree()
with cfqg == NULL.

Signed-off-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agofs/splice: Pull buf->ops->confirm() from splice_from_pipe actors
Michał Mirosław [Fri, 17 Dec 2010 07:56:44 +0000 (08:56 +0100)] 
fs/splice: Pull buf->ops->confirm() from splice_from_pipe actors

This patch pulls calls to buf->ops->confirm() from all actors passed
(also indirectly) to splice_from_pipe_feed().

Is avoiding the call to buf->ops->confirm() while splice()ing to
/dev/null is an intentional optimization? No other user does that
and this will remove this special case.

Against current linux.git 6313e3c21743cc88bb5bd8aa72948ee1e83937b6.

Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoMerge branch 'for-2.6.38/event-handling' into for-next
Jens Axboe [Thu, 16 Dec 2010 16:59:39 +0000 (17:59 +0100)] 
Merge branch 'for-2.6.38/event-handling' into for-next

13 years agocdrom: export cdrom_check_events()
Jens Axboe [Thu, 16 Dec 2010 16:58:50 +0000 (17:58 +0100)] 
cdrom: export cdrom_check_events()

It's used by sr, so we need to export it.

Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoMerge branch 'for-2.6.38/event-handling' into for-next
Jens Axboe [Thu, 16 Dec 2010 16:54:38 +0000 (17:54 +0100)] 
Merge branch 'for-2.6.38/event-handling' into for-next

13 years agoMerge branch 'for-2.6.38/core' into for-next
Jens Axboe [Thu, 16 Dec 2010 16:54:32 +0000 (17:54 +0100)] 
Merge branch 'for-2.6.38/core' into for-next

13 years agosd: implement sd_check_events()
Tejun Heo [Wed, 8 Dec 2010 19:57:42 +0000 (20:57 +0100)] 
sd: implement sd_check_events()

Replace sd_media_change() with sd_check_events().  sd used to set the
changed state whenever the device is not ready, which can cause event
loop while the device is not ready.  Media presence handling code is
changed such that the changed state is set iff the media presence
actually changes.  UA still always sets the changed state and
NOT_READY always (at least where it used to set ->changed) clears
media presence, so no event is lost.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agosr: implement sr_check_events()
Tejun Heo [Thu, 16 Dec 2010 16:52:17 +0000 (17:52 +0100)] 
sr: implement sr_check_events()

Replace sr_media_change() with sr_check_events().  It normally only
uses GET_EVENT_STATUS_NOTIFICATION to check both media change and
eject request.  If @clearing includes DISK_EVENT_MEDIA_CHANGE, it
issues TUR and compares whether media presence has changed.  The SCSI
specific media change uevent is kept for compatibility.

sr_media_change() was doing both media change check and revalidation.
The revalidation part is split into sr_block_revalidate_disk().

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoscsi: replace sr_test_unit_ready() with scsi_test_unit_ready()
Tejun Heo [Wed, 8 Dec 2010 19:57:40 +0000 (20:57 +0100)] 
scsi: replace sr_test_unit_ready() with scsi_test_unit_ready()

The usage of TUR has been confusing involving several different
commits updating different parts over time.  Currently, the only
differences between scsi_test_unit_ready() and sr_test_unit_ready()
are,

* scsi_test_unit_ready() also sets sdev->changed on NOT_READY.

* scsi_test_unit_ready() returns 0 if TUR ended with UNIT_ATTENTION or
  NOT_READY.

Due to the above two differences, sr is using its own
sr_test_unit_ready(), but sd - the sole user of the above extra
handling - doesn't even need them.

Where scsi_test_unit_ready() is used in sd_media_changed(), the code
is looking for device ready w/ media present state which is true iff
TUR succeeds w/o sense data or UA, and when the device is not ready
for whatever reason sd_media_changed() explicitly marks media as
missing so there's no reason to set sdev->changed automatically from
scsi_test_unit_ready() on NOT_READY.

Drop both special handlings from scsi_test_unit_ready(), which makes
it equivalant to sr_test_unit_ready(), and replace
sr_test_unit_ready() with scsi_test_unit_ready().  Also, drop the
unnecessary explicit NOT_READY check from sd_media_changed().
Checking return value is enough for testing device readiness.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoscsi: fix TUR error handling in sr_media_change()
Tejun Heo [Thu, 9 Dec 2010 10:18:42 +0000 (11:18 +0100)] 
scsi: fix TUR error handling in sr_media_change()

sr_test_unit_ready() returns 0 iff TUR succeeded - IOW, when media is
present and the device is actually ready, so the return value wouldn't
be zero when TUR ends with sense data. sr_media_change() incorrectly
tests (retval || (scsi_sense_valid(sshdr)...)) when it tries to test
whether TUR failed without sense data or with sense data indicating
media-not-present.

Fix the test using scsi_status_is_good() and update comments.

- Fixed a comment typo spotted by Eike.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agocdrom: add ->check_events() support
Tejun Heo [Wed, 8 Dec 2010 19:57:38 +0000 (20:57 +0100)] 
cdrom: add ->check_events() support

In principle, cdrom just needs to pass through ->check_events() but
CDROM_MEDIA_CHANGED ioctl makes things a bit more complex.  Just as
with ->media_changed() support, cdrom code needs to buffer the events
and serve them to ioctl and vfs as requested.

As the code has to deal with both ->check_events() and
->media_changed(), and vfs and ioctl event buffering, this patch adds
check_events caching on top of the existing cdi->mc_flags buffering.

It may be a good idea to deprecate CDROM_MEDIA_CHANGED ioctl and
remove all this mess.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoimplement in-kernel gendisk events handling
Tejun Heo [Wed, 8 Dec 2010 19:57:37 +0000 (20:57 +0100)] 
implement in-kernel gendisk events handling

Currently, media presence polling for removeable block devices is done
from userland.  There are several issues with this.

* Polling is done by periodically opening the device.  For SCSI
  devices, the command sequence generated by such action involves a
  few different commands including TEST_UNIT_READY.  This behavior,
  while perfectly legal, is different from Windows which only issues
  single command, GET_EVENT_STATUS_NOTIFICATION.  Unfortunately, some
  ATAPI devices lock up after being periodically queried such command
  sequences.

* There is no reliable and unintrusive way for a userland program to
  tell whether the target device is safe for media presence polling.
  For example, polling for media presence during an on-going burning
  session can make it fail.  The polling program can avoid this by
  opening the device with O_EXCL but then it risks making a valid
  exclusive user of the device fail w/ -EBUSY.

* Userland polling is unnecessarily heavy and in-kernel implementation
  is lighter and better coordinated (workqueue, timer slack).

This patch implements framework for in-kernel disk event handling,
which includes media presence polling.

* bdops->check_events() is added, which supercedes ->media_changed().
  It should check whether there's any pending event and return if so.
  Currently, two events are defined - DISK_EVENT_MEDIA_CHANGE and
  DISK_EVENT_EJECT_REQUEST.  ->check_events() is guaranteed not to be
  called parallelly.

* gendisk->events and ->async_events are added.  These should be
  initialized by block driver before passing the device to add_disk().
  The former contains the mask of all supported events and the latter
  the mask of all events which the device can report without polling.
  /sys/block/*/events[_async] export these to userland.

* Kernel parameter block.events_dfl_poll_msecs controls the system
  polling interval (default is 0 which means disable) and
  /sys/block/*/events_poll_msecs control polling intervals for
  individual devices (default is -1 meaning use system setting).  Note
  that if a device can report all supported events asynchronously and
  its polling interval isn't explicitly set, the device won't be
  polled regardless of the system polling interval.

* If a device is opened exclusively with write access, event checking
  is automatically disabled until all write exclusive accesses are
  released.

* There are event 'clearing' events.  For example, both of currently
  defined events are cleared after the device has been successfully
  opened.  This information is passed to ->check_events() callback
  using @clearing argument as a hint.

* Event checking is always performed from system_nrt_wq and timer
  slack is set to 25% for polling.

* Nothing changes for drivers which implement ->media_changed() but
  not ->check_events().  Going forward, all drivers will be converted
  to ->check_events() and ->media_change() will be dropped.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoblock: move register_disk() and del_gendisk() to block/genhd.c
Tejun Heo [Wed, 8 Dec 2010 19:57:36 +0000 (20:57 +0100)] 
block: move register_disk() and del_gendisk() to block/genhd.c

There's no reason for register_disk() and del_gendisk() to be in
fs/partitions/check.c.  Move both to genhd.c.  While at it, collapse
unlink_gendisk(), which was artificially in a separate function due to
genhd.c / check.c split, into del_gendisk().

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoblock: kill genhd_media_change_notify()
Tejun Heo [Wed, 8 Dec 2010 19:57:35 +0000 (20:57 +0100)] 
block: kill genhd_media_change_notify()

There's no user of the facility.  Kill it.

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoblock cfq: select new workload if priority changed
Shaohua Li writes [Mon, 13 Dec 2010 13:32:22 +0000 (14:32 +0100)] 
block cfq: select new workload if priority changed

If priority is changed, continuing to check workload_expires and service tree
count of the previous workload does not make sense. We should always choose
the workload with lowest key of new priority in such case.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agocfq-iosched: Get rid of on_st flag
Gui Jianfeng [Tue, 30 Nov 2010 19:52:47 +0000 (20:52 +0100)] 
cfq-iosched: Get rid of on_st flag

It's able to check whether a CFQ group on a service tree by
checking "cfqg->rb_node". There's no need to maintain an
extra flag here.

Signed-off-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agocfq-iosched: Get rid of st->active
Gui Jianfeng [Tue, 30 Nov 2010 19:52:46 +0000 (20:52 +0100)] 
cfq-iosched: Get rid of st->active

When a cfq group is running, it won't be dequeued from service tree, so
there's no need to store the active one in st->active. Just gid rid of it.

Signed-off-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
13 years agoMerge branch 'cleanup-bd_claim' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jens Axboe [Sat, 27 Nov 2010 18:49:18 +0000 (19:49 +0100)] 
Merge branch 'cleanup-bd_claim' of git://git./linux/kernel/git/tj/misc into for-2.6.38/core

14 years agoblock: Rename "block_remap" tracepoint to "block_bio_remap" to clarify the event.
Mike Snitzer [Tue, 16 Nov 2010 11:52:38 +0000 (12:52 +0100)] 
block: Rename "block_remap" tracepoint to "block_bio_remap" to clarify the event.

Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
14 years agoMerge branch 'for-2.6.38/rc2-holder' into for-2.6.38/core
Jens Axboe [Tue, 16 Nov 2010 09:10:12 +0000 (10:10 +0100)] 
Merge branch 'for-2.6.38/rc2-holder' into for-2.6.38/core

14 years agoMerge branch 'v2.6.37-rc2' into for-2.6.38/core
Jens Axboe [Tue, 16 Nov 2010 09:09:42 +0000 (10:09 +0100)] 
Merge branch 'v2.6.37-rc2' into for-2.6.38/core

14 years agoLinux 2.6.37-rc2
Linus Torvalds [Tue, 16 Nov 2010 02:31:02 +0000 (18:31 -0800)] 
Linux 2.6.37-rc2

14 years agocapabilities/syslog: open code cap_syslog logic to fix build failure
Eric Paris [Mon, 15 Nov 2010 23:36:29 +0000 (18:36 -0500)] 
capabilities/syslog: open code cap_syslog logic to fix build failure

The addition of CONFIG_SECURITY_DMESG_RESTRICT resulted in a build
failure when CONFIG_PRINTK=n.  This is because the capabilities code
which used the new option was built even though the variable in question
didn't exist.

The patch here fixes this by moving the capabilities checks out of the
LSM and into the caller.  All (known) LSMs should have been calling the
capabilities hook already so it actually makes the code organization
better to eliminate the hook altogether.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: James Morris <jmorris@namei.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Mon, 15 Nov 2010 22:06:11 +0000 (14:06 -0800)] 
Merge branch 'omap-fixes-for-linus' of git://git./linux/kernel/git/tmlind/linux-omap-2.6

* 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6:
  arm: omap1: devices: need to return with a value
  OMAP1: camera.h: add missing include
  omap: dma: Add read-back to DMA interrupt handler to avoid spuriousinterrupts
  OMAP2: Devkit8000: Fix mmc regulator failure

14 years agoMerge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelv...
Linus Torvalds [Mon, 15 Nov 2010 22:05:44 +0000 (14:05 -0800)] 
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/jdelvare/staging

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  hwmon: (w83795) Check for BEEP pin availability
  hwmon: (w83795) Clear intrusion alarm immediately
  hwmon: (w83795) Read the intrusion state properly
  hwmon: (w83795) Print the actual temperature channels as sources
  hwmon: (w83795) List all usable temperature sources
  hwmon: (w83795) Expose fan control method
  hwmon: (w83795) Fix fan control mode attributes
  hwmon: (lm95241) Check validity of input values
  hwmon: Change mail address of Hans J. Koch

14 years agoMerge branch 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvar...
Linus Torvalds [Mon, 15 Nov 2010 22:03:17 +0000 (14:03 -0800)] 
Merge branch 'i2c-for-linus' of git://git./linux/kernel/git/jdelvare/staging

* 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  i2c: Sanity checks on adapter registration
  i2c: Mark i2c_adapter.id as deprecated
  i2c: Drivers shouldn't include <linux/i2c-id.h>
  i2c: Delete unused adapter IDs
  i2c: Remove obsolete cleanup for clientdata

14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes...
Linus Torvalds [Mon, 15 Nov 2010 22:01:33 +0000 (14:01 -0800)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/jbarnes/pci-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
  PCI: sysfs: fix printk warnings
  PCI: fix pci_bus_alloc_resource() hang, prefer positive decode
  PCI: read current power state at enable time
  PCI: fix size checks for mmap() on /proc/bus/pci files
  x86/PCI: coalesce overlapping host bridge windows
  PCI hotplug: ibmphp: Add check to prevent reading beyond mapped area

14 years agoi2c: Sanity checks on adapter registration
Jean Delvare [Mon, 15 Nov 2010 21:40:38 +0000 (22:40 +0100)] 
i2c: Sanity checks on adapter registration

Make sure I2C adapters being registered have the required struct
fields set. If they don't, problems will happen later.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agoi2c: Mark i2c_adapter.id as deprecated
Jean Delvare [Mon, 15 Nov 2010 21:40:38 +0000 (22:40 +0100)] 
i2c: Mark i2c_adapter.id as deprecated

It's about time to make it clear that i2c_adapter.id is deprecated.
Hopefully this will remind the last user to move over to a different
strategy.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Jarod Wilson <jarod@redhat.com>
Acked-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Acked-by: Hans Verkuil <hverkuil@xs4all.nl>
14 years agoi2c: Drivers shouldn't include <linux/i2c-id.h>
Jean Delvare [Mon, 15 Nov 2010 21:40:38 +0000 (22:40 +0100)] 
i2c: Drivers shouldn't include <linux/i2c-id.h>

Drivers don't need to include <linux/i2c-id.h>, especially not when
they don't use anything that header file provides.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Michael Hunold <michael@mihu.de>
Acked-by: Mauro Carvalho Chehab <mchehab@redhat.com>
14 years agoi2c: Delete unused adapter IDs
Jean Delvare [Mon, 15 Nov 2010 21:40:38 +0000 (22:40 +0100)] 
i2c: Delete unused adapter IDs

Delete unused I2C adapter IDs. Special cases are:

* I2C_HW_B_RIVA was still set in driver rivafb, however no other
  driver is ever looking for this value, so we can safely remove it.
* I2C_HW_B_HDPVR is used in staging driver lirc_zilog, however no
  adapter ID is ever set to this value, so the code in question never
  runs. As the code additionally expects that I2C_HW_B_HDPVR may not
  be defined, we can delete it now and let the lirc_zilog driver
  maintainer rewrite this piece of code.

Big thanks for Hans Verkuil for doing all the hard work :)

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Jarod Wilson <jarod@redhat.com>
Acked-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Acked-by: Hans Verkuil <hverkuil@xs4all.nl>
14 years agoi2c: Remove obsolete cleanup for clientdata
Wolfram Sang [Mon, 15 Nov 2010 21:40:38 +0000 (22:40 +0100)] 
i2c: Remove obsolete cleanup for clientdata

A few new i2c-drivers came into the kernel which clear the clientdata-pointer
on exit. This is obsolete meanwhile, so fix it and hope the word will spread.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Alan Cox <alan@linux.intel.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agoinclude/linux/kernel.h: Move logging bits to include/linux/printk.h
Linus Torvalds [Mon, 15 Nov 2010 21:37:37 +0000 (13:37 -0800)] 
include/linux/kernel.h: Move logging bits to include/linux/printk.h

Move the logging bits from kernel.h into printk.h so that
there is a bit more logical separation of the generic from
the printk logging specific parts.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoFix gcc 4.5.1 miscompiling drivers/char/i8k.c (again)
Jim Bos [Mon, 15 Nov 2010 20:22:37 +0000 (21:22 +0100)] 
Fix gcc 4.5.1 miscompiling drivers/char/i8k.c (again)

The fix in commit 6b4e81db2552 ("i8k: Tell gcc that *regs gets
clobbered") to work around the gcc miscompiling i8k.c to add "+m
(*regs)" caused register pressure problems and a build failure.

Changing the 'asm' statement to 'asm volatile' instead should prevent
that and works around the gcc bug as well, so we can remove the "+m".

[ Background on the gcc bug: a memory clobber fails to mark the function
  the asm resides in as non-pure (aka "__attribute__((const))"), so if
  the function does nothing else that triggers the non-pure logic, gcc
  will think that that function has no side effects at all. As a result,
  callers will be mis-compiled.

  Adding the "+m" made gcc see that it's not a pure function, and so
  does "asm volatile". The problem was never really the need to mark
  "*regs" as changed, since the memory clobber did that part - the
  problem was just a bug in the gcc "pure" function analysis  - Linus ]

Signed-off-by: Jim Bos <jim876@xs4all.nl>
Acked-by: Jakub Jelinek <jakub@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agohwmon: (w83795) Check for BEEP pin availability
Jean Delvare [Mon, 15 Nov 2010 20:38:57 +0000 (21:38 +0100)] 
hwmon: (w83795) Check for BEEP pin availability

On the W83795ADG, there's a single pin for BEEP and OVT#, so you
can't have both. Check the configuration and don't create beep
attributes when BEEP pin is not available.

The W83795G has a dedicated BEEP pin so the functionality is always
available there.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (w83795) Clear intrusion alarm immediately
Jean Delvare [Mon, 15 Nov 2010 20:38:57 +0000 (21:38 +0100)] 
hwmon: (w83795) Clear intrusion alarm immediately

When asked to clear the intrusion alarm, do so immediately. We have to
invalidate the cache to make sure the new status will be read. But we
also have to read from the status register once to clear the pending
alarm, as writing to CLR_CHS surprising won't clear it automatically.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (w83795) Read the intrusion state properly
Jean Delvare [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: (w83795) Read the intrusion state properly

We can't read the intrusion state from the real-time alarm registers
as we do for all other alarm flags, because real-time alarm bits don't
stick (by definition) and the intrusion state has to stick until
explicitly cleared (otherwise it has little value.)

So we have to use the interrupt status register instead, which is read
from the same address but with a configuration bit flipped in another
register.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (w83795) Print the actual temperature channels as sources
Jean Delvare [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: (w83795) Print the actual temperature channels as sources

Don't expose raw register values to user-space. Decode and encode
temperature channels selected as temperature sources as needed.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (w83795) List all usable temperature sources
Jean Delvare [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: (w83795) List all usable temperature sources

Temperature sources are not correlated directly with temperature
channels. A look-up table is required to find out which temperature
sources can be used depending on which temperature channels (both
analog and digital) are enabled.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (w83795) Expose fan control method
Jean Delvare [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: (w83795) Expose fan control method

Expose fan control method (DC vs. PWM) using the standard sysfs
attributes. I've made it read-only as the board should be wired for
a given mode, the BIOS should have set up the chip for this mode, and
you shouldn't have to change it. But it would be easy enough to make
it changeable if someone comes up with a use case.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (w83795) Fix fan control mode attributes
Jean Delvare [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: (w83795) Fix fan control mode attributes

There were two bugs:
* Speed cruise mode was improperly reported for all fans but fan1.
* Fan control method (PWM vs. DC) was mixed with the control mode.
  It will be added back as a separate attribute, as per the standard
  sysfs interface.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
14 years agohwmon: (lm95241) Check validity of input values
Jean Delvare [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: (lm95241) Check validity of input values

This clears the following build-time warnings I was seeing:

drivers/hwmon/lm95241.c: In function "set_interval":
drivers/hwmon/lm95241.c:132:15: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result
drivers/hwmon/lm95241.c: In function "set_max2":
drivers/hwmon/lm95241.c:278:1: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result
drivers/hwmon/lm95241.c: In function "set_max1":
drivers/hwmon/lm95241.c:277:1: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result
drivers/hwmon/lm95241.c: In function "set_min2":
drivers/hwmon/lm95241.c:249:1: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result
drivers/hwmon/lm95241.c: In function "set_min1":
drivers/hwmon/lm95241.c:248:1: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result
drivers/hwmon/lm95241.c: In function "set_type2":
drivers/hwmon/lm95241.c:220:1: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result
drivers/hwmon/lm95241.c: In function "set_type1":
drivers/hwmon/lm95241.c:219:1: warning: ignoring return value of "strict_strtol", declared with attribute warn_unused_result

This also fixes a small race in set_interval() as a side effect: by
working with a temporary local variable we prevent data->interval from
being accessed at a time it contains the interval value in the wrong
unit.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Davide Rizzo <elpa.rizzo@gmail.com>
14 years agohwmon: Change mail address of Hans J. Koch
Hans J. Koch [Mon, 15 Nov 2010 20:38:56 +0000 (21:38 +0100)] 
hwmon: Change mail address of Hans J. Koch

My old mail address doesn't exist anymore. This changes all occurrences
to my new address.

Signed-off-by: Hans J. Koch <hjk@hansjkoch.de>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
14 years agoblk-cgroup: Allow creation of hierarchical cgroups
Vivek Goyal [Mon, 15 Nov 2010 18:37:36 +0000 (19:37 +0100)] 
blk-cgroup: Allow creation of hierarchical cgroups

o Allow hierarchical cgroup creation for blkio controller

o Currently we disallow it as both the io controller policies (throttling
  as well as proportion bandwidth) do not support hierarhical accounting
  and control. But the flip side is that blkio controller can not be used with
  libvirt as libvirt creates a cgroup hierarchy deeper than 1 level.

  <top-level-cgroup-dir>/<controller>/libvirt/qemu/<virtual-machine-groups>

o So this patch will allow creation of cgroup hierarhcy but at the backend
  everything will be treated as flat. So if somebody created a an hierarchy
  like as follows.

root
/  \
     test1 test2
|
     test3

  CFQ and throttling will practically treat all groups at same level.

pivot
     /  |   \  \
root  test1 test2  test3

o Once we have actual support for hierarchical accounting and control
  then we can introduce another cgroup tunable file "blkio.use_hierarchy"
  which will be 0 by default but if user wants to enforce hierarhical
  control then it can be set to 1. This way there should not be any
  ABI problems down the line.

o The only not so pretty part is introduction of extra file "use_hierarchy"
  down the line. Kame-san had mentioned that hierarhical accounting is
  expensive in memory controller hence they keep it off by default. I
  suspect same will be the case for IO controller also as for each IO
  completion we shall have to account IO through hierarchy up to the root.
  if yes, then it probably is not a very bad idea to introduce this extra
  file so that it will be used only when somebody needs it and some people
  might enable hierarchy only in part of the hierarchy.

o This is how basically memory controller also uses "use_hierarhcy" and
  they also allowed creation of hierarchies when actual backend support
  was not available.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Acked-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Reviewed-by: Gui Jianfeng <guijianfeng@cn.fujitsu.com>
Reviewed-by: Ciju Rajan K <ciju@linux.vnet.ibm.com>
Tested-by: Ciju Rajan K <ciju@linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
14 years agoPCI: sysfs: fix printk warnings
Randy Dunlap [Sat, 13 Nov 2010 16:44:33 +0000 (08:44 -0800)] 
PCI: sysfs: fix printk warnings

Cast pci_resource_start() and pci_resource_len() to u64 for printk.

drivers/pci/pci-sysfs.c:753: warning: format '%16Lx' expects type 'long long unsigned int', but argument 9 has type 'resource_size_t'
drivers/pci/pci-sysfs.c:753: warning: format '%16Lx' expects type 'long long unsigned int', but argument 10 has type 'resource_size_t'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-2.6-fixes
Linus Torvalds [Mon, 15 Nov 2010 16:43:29 +0000 (08:43 -0800)] 
Merge git://git./linux/kernel/git/steve/gfs2-2.6-fixes

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-2.6-fixes:
  GFS2: Fix inode deallocation race

14 years agoMerge branch 's5p-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 15 Nov 2010 16:43:04 +0000 (08:43 -0800)] 
Merge branch 's5p-fixes-for-linus' of git://git./linux/kernel/git/kgene/linux-samsung

* 's5p-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: SAMSUNG: Fix HAVE_S3C_RTC warnings
  ARM: SAMSUNG: Fix HAVE_S3C2410_I2C warnings
  ARM: SAMSUNG: Fix HAVE_S3C2410_WATCHDOG warnings

14 years agoMerge branch 'fbdev-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Mon, 15 Nov 2010 16:42:07 +0000 (08:42 -0800)] 
Merge branch 'fbdev-fixes-for-linus' of git://git./linux/kernel/git/lethal/fbdev-2.6

* 'fbdev-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/fbdev-2.6:
  fsl-diu-fb: drop dead ioctl define
  MAINTAINERS: Add an fbdev git tree entry.
  OMAP: DSS: Fix documentation regarding 'vram' kernel parameter
  OMAP: VRAM: Fix boot-time memory allocation
  OMAP: VRAM: improve VRAM error prints
  sisfb: limit POST memory test according to PCI resource length
  fbdev: sh_mobile_lcdc: use correct number of modes, when using the default
  fbdev: sh_mobile_lcdc: use the standard CEA-861 720p timing
  fbdev: sh_mobile_hdmi: properly clean up modedb on monitor unplug

14 years agoMerge branches 'sh-fixes-for-linus' and 'rmobile-fixes-for-linus' of git://git.kernel...
Linus Torvalds [Mon, 15 Nov 2010 16:41:30 +0000 (08:41 -0800)] 
Merge branches 'sh-fixes-for-linus' and 'rmobile-fixes-for-linus' of git://git./linux/kernel/git/lethal/sh-2.6

* 'sh-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6:
  sh: intc: Fix up build failure introduced by radix tree changes.
  MAINTAINERS: update the sh git tree entry.
  sh: clkfwk: fix up compiler warnings.
  sh: intc: Fix up initializers for gcc 4.5.
  rtc: rtc-sh - fix a memory leak

* 'rmobile-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6:
  ARM: mach-shmobile: ap4evb: add fsib 44100Hz rate
  MAINTAINERS: update the ARM SH-Mobile git tree entry.
  ARM: mach-shmobile: ap4evb: Mark NOR boot loader partitions read-only.
  ARM: mach-shmobile: intc-sh7372: fix interrupt number

14 years agoGFS2: Fix inode deallocation race
Steven Whitehouse [Wed, 3 Nov 2010 20:01:07 +0000 (20:01 +0000)] 
GFS2: Fix inode deallocation race

This area of the code has always been a bit delicate due to the
subtleties of lock ordering. The problem is that for "normal"
alloc/dealloc, we always grab the inode locks first and the rgrp lock
later.

In order to ensure no races in looking up the unlinked, but still
allocated inodes, we need to hold the rgrp lock when we do the lookup,
which means that we can't take the inode glock.

The solution is to borrow the technique already used by NFS to solve
what is essentially the same problem (given an inode number, look up
the inode carefully, checking that it really is in the expected
state).

We cannot do that directly from the allocation code (lock ordering
again) so we give the job to the pre-existing delete workqueue and
carry on with the allocation as normal.

If we find there is no space, we do a journal flush (required anyway
if space from a deallocation is to be released) which should block
against the pending deallocations, so we should always get the space
back.

Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
14 years agoMerge branch 'for-linus' into for-next
Jens Axboe [Mon, 15 Nov 2010 09:25:49 +0000 (10:25 +0100)] 
Merge branch 'for-linus' into for-next

14 years agoioprio: grab rcu_read_lock in sys_ioprio_{set,get}()
Greg Thelen [Mon, 15 Nov 2010 09:20:52 +0000 (10:20 +0100)] 
ioprio: grab rcu_read_lock in sys_ioprio_{set,get}()

Using:
- CONFIG_LOCKUP_DETECTOR=y
- CONFIG_PREEMPT=y
- CONFIG_LOCKDEP=y
- CONFIG_PROVE_LOCKING=y
- CONFIG_PROVE_RCU=y
found a missing rcu lock during boot on a 512 MiB x86_64 ubuntu vm:
  ===================================================
  [ INFO: suspicious rcu_dereference_check() usage. ]
  ---------------------------------------------------
  kernel/pid.c:419 invoked rcu_dereference_check() without protection!

  other info that might help us debug this:

  rcu_scheduler_active = 1, debug_locks = 0
  1 lock held by ureadahead/1355:
   #0:  (tasklist_lock){.+.+..}, at: [<ffffffff8115bc09>] sys_ioprio_set+0x7f/0x29e

  stack backtrace:
  Pid: 1355, comm: ureadahead Not tainted 2.6.37-dbg-DEV #1
  Call Trace:
   [<ffffffff8109c10c>] lockdep_rcu_dereference+0xaa/0xb3
   [<ffffffff81088cbf>] find_task_by_pid_ns+0x44/0x5d
   [<ffffffff81088cfa>] find_task_by_vpid+0x22/0x24
   [<ffffffff8115bc3e>] sys_ioprio_set+0xb4/0x29e
   [<ffffffff8147cf21>] ? trace_hardirqs_off_thunk+0x3a/0x3c
   [<ffffffff8105c409>] sysenter_dispatch+0x7/0x2c
   [<ffffffff8147cee2>] ? trace_hardirqs_on_thunk+0x3a/0x3f

The fix is to:
a) grab rcu lock in sys_ioprio_{set,get}() and
b) avoid grabbing tasklist_lock.
Discussion in: http://marc.info/?l=linux-kernel&m=128951324702889

Signed-off-by: Greg Thelen <gthelen@google.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Modified by Jens to remove the now redundant inner rcu lock and
unlock since they are now protected by the outer lock.

Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
14 years agoMerge branch 'rmobile/core' into rmobile-fixes-for-linus
Paul Mundt [Mon, 15 Nov 2010 05:47:16 +0000 (14:47 +0900)] 
Merge branch 'rmobile/core' into rmobile-fixes-for-linus

14 years agoARM: mach-shmobile: ap4evb: add fsib 44100Hz rate
Kuninori Morimoto [Mon, 15 Nov 2010 03:11:09 +0000 (03:11 +0000)] 
ARM: mach-shmobile: ap4evb: add fsib 44100Hz rate

Tested-by: Tony SIM <chinyeow.sim.xt@renesas.com>
Tested-by: TAKEI Mitsuharu <takei.andr@gmail.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
14 years agosh: intc: Fix up build failure introduced by radix tree changes.
Paul Mundt [Mon, 15 Nov 2010 05:30:30 +0000 (14:30 +0900)] 
sh: intc: Fix up build failure introduced by radix tree changes.

The radix tree retry logic got a bit of an overhaul and subsequently
broke the virtual IRQ subgroup build. Simply switch over to
radix_tree_deref_retry() as per the filemap changes, which the virq
lookup logic was modelled after in the first place.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
14 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux...
Paul Mundt [Mon, 15 Nov 2010 04:54:00 +0000 (13:54 +0900)] 
Merge branch 'master' of git://git./linux/kernel/git/torvalds/linux-2.6 into sh/urgent

14 years agofsl-diu-fb: drop dead ioctl define
Mike Frysinger [Sat, 13 Nov 2010 07:06:27 +0000 (02:06 -0500)] 
fsl-diu-fb: drop dead ioctl define

The fsl-diu-fb driver no longer uses this define, and we have a common one
to cover this already (FBIO_WAITFORVSYNC).

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
14 years agoMAINTAINERS: Add an fbdev git tree entry.
Paul Mundt [Mon, 15 Nov 2010 04:25:31 +0000 (13:25 +0900)] 
MAINTAINERS: Add an fbdev git tree entry.

Now that there's an fbdev git tree (this is also what is pulled in to
-next), stub it in to the MAINTAINERS entry.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg...
Linus Torvalds [Sun, 14 Nov 2010 21:06:37 +0000 (13:06 -0800)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/penberg/slab-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  slub: Fix slub_lock down/up imbalance

14 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec...
Linus Torvalds [Sun, 14 Nov 2010 21:04:53 +0000 (13:04 -0800)] 
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jlbec/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2:
  ocfs2: Change some lock status member in ocfs2_lock_res to char.

14 years agoslub: Fix slub_lock down/up imbalance
Pavel Emelyanov [Thu, 28 Oct 2010 09:50:37 +0000 (13:50 +0400)] 
slub: Fix slub_lock down/up imbalance

There are two places, that do not release the slub_lock.

Respective bugs were introduced by sysfs changes ab4d5ed5 (slub: Enable
sysfs support for !CONFIG_SLUB_DEBUG) and 2bce6485 ( slub: Allow removal
of slab caches during boot).

Acked-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
14 years agoMerge branch 'urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6
Linus Torvalds [Sat, 13 Nov 2010 18:00:15 +0000 (10:00 -0800)] 
Merge branch 'urgent' of git://git./linux/kernel/git/brodo/pcmcia-2.6

* 'urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6:
  drivers/pcmcia/soc_common.c: Use printf extension %pV
  pcmcia: fix warning in synclink driver
  pcmcia/sa1100: don't put machine specific init functions in .init.text
  pcmcia/cm4000: fix error code
  pd6729: Coding Style fixes

14 years agoRevert "8250: Fix tcsetattr to avoid ioctl(TIOCMIWAIT) hang"
Ingo Molnar [Sat, 13 Nov 2010 15:21:58 +0000 (16:21 +0100)] 
Revert "8250: Fix tcsetattr to avoid ioctl(TIOCMIWAIT) hang"

This reverts commit 47d3904fe40d62deee8cd46e79ca784e7a548acd.

Crashes any x86 serial console bootup:

  Console: colour VGA+ 80x25
  BUG: unable to handle kernel NULL pointer dereference at 0000000000000158
  IP: [<ffffffff811ebcb4>] serial8250_do_set_termios+0x1d4/0x430
  ...

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Greg KH <gregkh@suse.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Sat, 13 Nov 2010 17:55:56 +0000 (09:55 -0800)] 
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: padlock - Fix AES-CBC handling on odd-block-sized input
  crypto: n2 - dubious error check

14 years agoMerge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Sat, 13 Nov 2010 17:55:19 +0000 (09:55 -0800)] 
Merge branch 'v4l_for_linus' of git://git./linux/kernel/git/mchehab/linux-2.6

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6:
  [media] soc-camera: Compile fixes for mx2-camera
  [media] SoC Camera: ov6650: minor cleanups
  [media] SOC Camera: OMAP1: typo fix
  [media] SoC Camera: OMAP1: update for recent videobuf changes
  [media] SoC Camera: OMAP1: update for recent framework changes
  [media] ARM mx3_camera: check for DMA engine type
  [media] tm6000: bugfix set tv standards
  [media] cafe_ccic: fix subdev configuration
  [media] saa7134: Fix autodetect for Behold A7 and H7 TV cards
  [media] v4l: kill the BKL
  [media] BZ#22292: dibx000_common: Restore i2c algo pointer

14 years agoi8k: Tell gcc that *regs gets clobbered
Jim Bos [Sat, 13 Nov 2010 11:13:53 +0000 (12:13 +0100)] 
i8k: Tell gcc that *regs gets clobbered

More recent GCC caused the i8k driver to stop working, on Slackware
compiler was upgraded from gcc-4.4.4 to gcc-4.5.1 after which it didn't
work anymore, meaning the driver didn't load or gave total nonsensical
output.

As it turned out the asm(..) statement forgot to mention it modifies the
*regs variable.

Credits to Andi Kleen and Andreas Schwab for providing the fix.

Signed-off-by: Jim Bos <jim876@xs4all.nl>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoACPI: debugfs custom_method open to non-root
Dave Jones [Sat, 13 Nov 2010 05:58:54 +0000 (00:58 -0500)] 
ACPI: debugfs custom_method open to non-root

Currently we have:

  --w--w--w-. 1 root root 0 2010-11-11 14:56 /sys/kernel/debug/acpi/custom_method

which is just crazy. Change this to --w-------.

Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Cc: stable@kernel.org (for 2.6.36)
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoocfs2: Change some lock status member in ocfs2_lock_res to char.
Tao Ma [Sat, 13 Nov 2010 08:22:02 +0000 (16:22 +0800)] 
ocfs2: Change some lock status member in ocfs2_lock_res to char.

Commit 83fd9c7 changes l_level, l_requested and l_blocking of
ocfs2_lock_res from int to unsigned char. But actually it is
initially as -1(ocfs2_lock_res_init_common) which
correspoding to 255 for unsigned char. So the whole dlm lock
mechanism doesn't work now which means a disaster to ocfs2.

Cc: Goldwyn Rodrigues <rgoldwyn@suse.de>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Joel Becker <joel.becker@oracle.com>
14 years agoblock: clean up blkdev_get() wrappers and their users
Tejun Heo [Sat, 13 Nov 2010 10:55:18 +0000 (11:55 +0100)] 
block: clean up blkdev_get() wrappers and their users

After recent blkdev_get() modifications, open_by_devnum() and
open_bdev_exclusive() are simple wrappers around blkdev_get().
Replace them with blkdev_get_by_dev() and blkdev_get_by_path().

blkdev_get_by_dev() is identical to open_by_devnum().
blkdev_get_by_path() is slightly different in that it doesn't
automatically add %FMODE_EXCL to @mode.

All users are converted.  Most conversions are mechanical and don't
introduce any behavior difference.  There are several exceptions.

* btrfs now sets FMODE_EXCL in btrfs_device->mode, so there's no
  reason to OR it explicitly on blkdev_put().

* gfs2, nilfs2 and the generic mount_bdev() now set FMODE_EXCL in
  sb->s_mode.

* With the above changes, sb->s_mode now always should contain
  FMODE_EXCL.  WARN_ON_ONCE() added to kill_block_super() to detect
  errors.

The new blkdev_get_*() functions are with proper docbook comments.
While at it, add function description to blkdev_get() too.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Neil Brown <neilb@suse.de>
Cc: Mike Snitzer <snitzer@redhat.com>
Cc: Joern Engel <joern@lazybastard.org>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Jan Kara <jack@suse.cz>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp>
Cc: reiserfs-devel@vger.kernel.org
Cc: xfs-masters@oss.sgi.com
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
14 years agoblock: check bdev_read_only() from blkdev_get()
Tejun Heo [Sat, 13 Nov 2010 10:55:17 +0000 (11:55 +0100)] 
block: check bdev_read_only() from blkdev_get()

bdev read-only status can be queried using bdev_read_only() and may
change while the device is being opened.  Enforce it by checking it
from blkdev_get() after open succeeds.

This makes bdev_read_only() check in open_bdev_exclusive() and
fsg_lun_open() unnecessary.  Drop them.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: David Brownell <dbrownell@users.sourceforge.net>
Cc: linux-usb@vger.kernel.org
14 years agoblock: reorganize claim/release implementation
Tejun Heo [Sat, 13 Nov 2010 10:55:17 +0000 (11:55 +0100)] 
block: reorganize claim/release implementation

With claim/release rolled into blkdev_get/put(), there's no reason to
keep bd_abort/finish_claim(), __bd_claim() and bd_release() as
separate functions.  It only makes the code difficult to follow.
Collapse them into blkdev_get/put().  This will ease future changes
around claim/release.

Signed-off-by: Tejun Heo <tj@kernel.org>
14 years agoblock: make blkdev_get/put() handle exclusive access
Tejun Heo [Sat, 13 Nov 2010 10:55:17 +0000 (11:55 +0100)] 
block: make blkdev_get/put() handle exclusive access

Over time, block layer has accumulated a set of APIs dealing with bdev
open, close, claim and release.

* blkdev_get/put() are the primary open and close functions.

* bd_claim/release() deal with exclusive open.

* open/close_bdev_exclusive() are combination of open and claim and
  the other way around, respectively.

* bd_link/unlink_disk_holder() to create and remove holder/slave
  symlinks.

* open_by_devnum() wraps bdget() + blkdev_get().

The interface is a bit confusing and the decoupling of open and claim
makes it impossible to properly guarantee exclusive access as
in-kernel open + claim sequence can disturb the existing exclusive
open even before the block layer knows the current open if for another
exclusive access.  Reorganize the interface such that,

* blkdev_get() is extended to include exclusive access management.
  @holder argument is added and, if is @FMODE_EXCL specified, it will
  gain exclusive access atomically w.r.t. other exclusive accesses.

* blkdev_put() is similarly extended.  It now takes @mode argument and
  if @FMODE_EXCL is set, it releases an exclusive access.  Also, when
  the last exclusive claim is released, the holder/slave symlinks are
  removed automatically.

* bd_claim/release() and close_bdev_exclusive() are no longer
  necessary and either made static or removed.

* bd_link_disk_holder() remains the same but bd_unlink_disk_holder()
  is no longer necessary and removed.

* open_bdev_exclusive() becomes a simple wrapper around lookup_bdev()
  and blkdev_get().  It also has an unexpected extra bdev_read_only()
  test which probably should be moved into blkdev_get().

* open_by_devnum() is modified to take @holder argument and pass it to
  blkdev_get().

Most of bdev open/close operations are unified into blkdev_get/put()
and most exclusive accesses are tested atomically at the open time (as
it should).  This cleans up code and removes some, both valid and
invalid, but unnecessary all the same, corner cases.

open_bdev_exclusive() and open_by_devnum() can use further cleanup -
rename to blkdev_get_by_path() and blkdev_get_by_devt() and drop
special features.  Well, let's leave them for another day.

Most conversions are straight-forward.  drbd conversion is a bit more
involved as there was some reordering, but the logic should stay the
same.

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Neil Brown <neilb@suse.de>
Acked-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Acked-by: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Peter Osterlund <petero2@telia.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andreas Dilger <adilger.kernel@dilger.ca>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <joel.becker@oracle.com>
Cc: Alex Elder <aelder@sgi.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: dm-devel@redhat.com
Cc: drbd-dev@lists.linbit.com
Cc: Leo Chen <leochen@broadcom.com>
Cc: Scott Branden <sbranden@broadcom.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Cc: Joern Engel <joern@logfs.org>
Cc: reiserfs-devel@vger.kernel.org
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
14 years agoblock: simplify holder symlink handling
Tejun Heo [Sat, 13 Nov 2010 10:55:17 +0000 (11:55 +0100)] 
block: simplify holder symlink handling

Code to manage symlinks in /sys/block/*/{holders|slaves} are overly
complex with multiple holder considerations, redundant extra
references to all involved kobjects, unused generic kobject holder
support and unnecessary mixup with bd_claim/release functionalities.

Strip it down to what's necessary (single gendisk holder) and make it
use a separate interface.  This is a step for cleaning up
bd_claim/release.  This patch makes dm-table slightly more complex but
it will be simplified again with further changes.

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Neil Brown <neilb@suse.de>
Acked-by: Mike Snitzer <snitzer@redhat.com>
Cc: dm-devel@redhat.com
14 years agobtrfs: close_bdev_exclusive() should use the same @flags as the matching open_bdev_ex...
Tejun Heo [Sat, 13 Nov 2010 10:55:17 +0000 (11:55 +0100)] 
btrfs: close_bdev_exclusive() should use the same @flags as the matching open_bdev_exclusive()

In the failure path of __btrfs_open_devices(), close_bdev_exclusive()
is called with @flags which doesn't match the one used during
open_bdev_exclusive().  Fix it.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Chris Mason <chris.mason@oracle.com>
14 years agomtd: fix bdev exclusive open bugs in block2mtd::add_device()
Tejun Heo [Sat, 13 Nov 2010 10:55:16 +0000 (11:55 +0100)] 
mtd: fix bdev exclusive open bugs in block2mtd::add_device()

There are two bdev exclusive open bugs.

* open_bdev_exclusive() must not be called with NULL holder.  Use dev
  as the holder.

* open_by_devnum() doesn't open the bdev exclusively but
  block2mtd_free_device() always assumes it.  Explicitly claim the
  bdev.

The latter is rather clumsy but will be simplified with future
blkdev_get/put() cleanups.

- Updated to use local variable @mode to cache FMODE_* masks as
  suggested by Artem Bityutskiy.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: linux-mtd@lists.infradead.org
Cc: Artem Bityutskiy <dedekind1@gmail.com>
14 years agoARM: SAMSUNG: Fix HAVE_S3C_RTC warnings
Kukjin Kim [Sat, 13 Nov 2010 07:11:46 +0000 (16:11 +0900)] 
ARM: SAMSUNG: Fix HAVE_S3C_RTC warnings

This patch fixes followng build warnings.
warning: (ARCH_S5P64X0 && <choice> || ARCH_S5PC100 && <choice> ||
          ARCH_S5PV210 && <choice> || ARCH_S5PV310 && <choice>)
         selects HAVE_S3C_RTC which has unmet direct dependencies (RTC_CLASS)

Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
14 years agoARM: SAMSUNG: Fix HAVE_S3C2410_I2C warnings
Kukjin Kim [Sat, 13 Nov 2010 07:08:32 +0000 (16:08 +0900)] 
ARM: SAMSUNG: Fix HAVE_S3C2410_I2C warnings

This patch fixes following warnings.
warning: (ARCH_S3C2410 && <choice> || ARCH_S3C64XX && <choice> ||
          ARCH_S5P64X0 && <choice> || ARCH_S5PC100 && <choice> ||
          ARCH_S5PV210 && <choice> || ARCH_S5PV310 && <choice>)
         selects HAVE_S3C2410_I2C which has unmet direct dependencies (I2C)

Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
14 years agoARM: SAMSUNG: Fix HAVE_S3C2410_WATCHDOG warnings
Kyungmin Park [Sat, 13 Nov 2010 07:01:59 +0000 (16:01 +0900)] 
ARM: SAMSUNG: Fix HAVE_S3C2410_WATCHDOG warnings

Fix build warnings
warning: (ARCH_S3C64XX && <choice> && WATCHDOG || ARCH_S5P64X0 && <choice> && WATCHDOG ||
          ARCH_S5P6442 && <choice> && WATCHDOG || ARCH_S5PC100 && <choice> ||
          ARCH_S5PV210 && <choice> || ARCH_S5PV310 && <choice> || MACH_SMDK6410 && ARCH_S3C64XX)
         selects HAVE_S3C2410_WATCHDOG which has unmet direct dependencies (WATCHDOG)

Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
[kgene.kim@samsung.com: Added fix same warning(mach-s3c64xx/Kconfig)]
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Sat, 13 Nov 2010 01:17:55 +0000 (17:17 -0800)] 
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (66 commits)
  can-bcm: fix minor heap overflow
  gianfar: Do not call device_set_wakeup_enable() under a spinlock
  ipv6: Warn users if maximum number of routes is reached.
  docs: Add neigh/gc_thresh3 and route/max_size documentation.
  axnet_cs: fix resume problem for some Ax88790 chip
  ipv6: addrconf: don't remove address state on ifdown if the address is being kept
  tcp: Don't change unlocked socket state in tcp_v4_err().
  x25: Prevent crashing when parsing bad X.25 facilities
  cxgb4vf: add call to Firmware to reset VF State.
  cxgb4vf: Fail open if link_start() fails.
  cxgb4vf: flesh out PCI Device ID Table ...
  cxgb4vf: fix some errors in Gather List to skb conversion
  cxgb4vf: fix bug in Generic Receive Offload
  cxgb4vf: don't implement trivial (and incorrect) ndo_select_queue()
  ixgbe: Look inside vlan when determining offload protocol.
  bnx2x: Look inside vlan when determining checksum proto.
  vlan: Add function to retrieve EtherType from vlan packets.
  virtio-net: init link state correctly
  ucc_geth: Fix deadlock
  ucc_geth: Do not bring the whole IF down when TX failure.
  ...

14 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzi...
Linus Torvalds [Sat, 13 Nov 2010 01:17:09 +0000 (17:17 -0800)] 
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  drivers/ata/pata_octeon_cf.c: delete double assignment
  pata_legacy: fix CONFIG_PATA_WINBOND_VLB_MODULE test
  libata: fix NULL sdev dereference race in atapi_qc_complete()

14 years agoMerge branch 'staging-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 13 Nov 2010 01:14:20 +0000 (17:14 -0800)] 
Merge branch 'staging-linus' of git://git./linux/kernel/git/gregkh/staging-2.6

* 'staging-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (38 commits)
  Revert "staging: tidspbridge: replace iommu custom for opensource implementation"
  Revert "staging: tidspbridge - move shared memory iommu maps to tiomap3430.c"
  Revert "staging: tidspbridge - rename bridge_brd_mem_map/unmap to a proper name"
  Revert "staging: tidspbridge - remove custom mmu code from tiomap3430.c"
  Revert "staging: tidspbridge - fix mmufault support"
  Revert "staging: tidspbridge - remove hw directory"
  Revert "staging: tidspbridge - move all iommu related code to a new file"
  Revert "staging: tidspbridge: remove dw_dmmu_base from cfg_hostres struct"
  Revert "staging: tidspbridge - remove reserved memory clean up"
  Revert "staging: tidspbridge - deprecate reserve/unreserve_memory funtions"
  Revert "staging: tidspbridge - remove dmm custom module"
  Revert "staging: tidspbridge - update Kconfig to select IOMMU module"
  staging: tidspbridge: hardcode SCM macros while fix is upstreamed
  Staging: keucr driver: fix uninitialized variable & proper memset length
  omap: dsp: remove shm from normal memory
  Staging: wlan-ng: Fix wrong #ifdef #endif sequence
  Staging: Update parameters for cfg80211 key management operation
  Staging: ath6kl: Fix pointer casts on 64-bit architectures
  Staging: batman-adv: suppress false warning when changing the mac address
  Staging: batman-adv: fix interface alternating and bonding reggression
  ...

14 years agoMerge branch 'usb-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 13 Nov 2010 01:13:28 +0000 (17:13 -0800)] 
Merge branch 'usb-linus' of git://git./linux/kernel/git/gregkh/usb-2.6

* 'usb-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (28 commits)
  Revert "USB: xhci: Use GFP_ATOMIC under spin_lock"
  USB: ohci-jz4740: Fix spelling in MODULE_ALIAS
  UWB: Return UWB_RSV_ALLOC_NOT_FOUND rather than crashing on NULL dereference if kzalloc fails
  usb: core: fix information leak to userland
  usb: misc: iowarrior: fix information leak to userland
  usb: misc: sisusbvga: fix information leak to userland
  usb: subtle increased memory usage in u_serial
  USB: option: fix when the driver is loaded incorrectly for some Huawei devices.
  USB: xhci: Use GFP_ATOMIC under spin_lock
  usb: gadget: goku_udc: add registered flag bit, fixing build
  USB: ehci/mxc: compile fix
  USB: Fix FSL USB driver on non Open Firmware systems
  USB: the development of the usb tree is now in git
  usb: musb: fail unaligned DMA transfers on v1.8 and above
  USB: ftdi_sio: add device IDs for Milkymist One JTAG/serial
  usb.h: fix ioctl kernel-doc info
  usb: musb: gadget: kill duplicate code in musb_gadget_queue()
  usb: musb: Fix handling of spurious SESSREQ
  usb: musb: fix kernel oops when loading musb_hdrc module for the 2nd time
  USB: musb: blackfin: push clkin value to platform resources
  ...

14 years agoMerge branch 'tty-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 13 Nov 2010 00:02:30 +0000 (16:02 -0800)] 
Merge branch 'tty-linus' of git://git./linux/kernel/git/gregkh/tty-2.6

* 'tty-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6:
  n_gsm: Fix length handling
  n_gsm: Copy n2 over when configuring via ioctl interface
  serial: bfin_5xx: grab port lock before making port termios changes
  serial: bfin_5xx: disable CON_PRINTBUFFER for consoles
  serial: bfin_5xx: remove redundant SSYNC to improve TX speed
  serial: bfin_5xx: always include DMA headers
  vcs: make proper usage of the poll flags
  amiserial: Remove unused variable icount
  8250: Fix tcsetattr to avoid ioctl(TIOCMIWAIT) hang
  tty_ldisc: Fix BUG() on hangup
  TTY: restore tty_ldisc_wait_idle
  SERIAL: blacklist si3052 chip
  drivers/serial/bfin_5xx.c: Fix line continuation defects
  tty: prevent DOS in the flush_to_ldisc
  8250: add support for Kouwell KW-L221N-2
  nozomi: Fix warning from the previous TIOCGCOUNT changes
  tty: fix warning in synclink driver
  tty: Fix formatting in tty.h
  tty: the development tree is now done in git

14 years agoMerge branch 'upstream/core' of git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen
Linus Torvalds [Sat, 13 Nov 2010 00:01:55 +0000 (16:01 -0800)] 
Merge branch 'upstream/core' of git://git./linux/kernel/git/jeremy/xen

* 'upstream/core' of git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen:
  xen: do not release any memory under 1M in domain 0
  xen: events: do not unmask event channels on resume
  xen: correct size of level2_kernel_pgt

14 years agoMerge branch 'stable/xen-pcifront-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 12 Nov 2010 23:54:39 +0000 (15:54 -0800)] 
Merge branch 'stable/xen-pcifront-fixes' of git://git./linux/kernel/git/konrad/xen

* 'stable/xen-pcifront-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen:
  MAINTAINERS: Mark XEN lists as moderated
  xen-pcifront: fix PCI reference leak
  xen-pcifront: Remove duplicate inclusion of headers.
  xen: fix memory leak in Xen PCI MSI/MSI-X allocator.
  MAINTAINERS: Update mailing list name for Xen pieces.

14 years agodrivers/ata/pata_octeon_cf.c: delete double assignment
Julia Lawall [Tue, 26 Oct 2010 10:25:43 +0000 (12:25 +0200)] 
drivers/ata/pata_octeon_cf.c: delete double assignment

Delete successive assignments to the same location.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@@
expression i;
@@

*i = ...;
 i = ...;
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agopata_legacy: fix CONFIG_PATA_WINBOND_VLB_MODULE test
Tejun Heo [Fri, 12 Nov 2010 11:01:41 +0000 (12:01 +0100)] 
pata_legacy: fix CONFIG_PATA_WINBOND_VLB_MODULE test

pata_legacy is incorrectly testing PATA_WINBOND_VLB_MODULE instead of
CONFIG_PATA_WINBOND_VLB_MODULE.  Fix it.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: "Robert P. J. Day" <rpjday@crashcourse.ca>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agolibata: fix NULL sdev dereference race in atapi_qc_complete()
Tejun Heo [Mon, 1 Nov 2010 10:39:19 +0000 (11:39 +0100)] 
libata: fix NULL sdev dereference race in atapi_qc_complete()

SCSI commands may be issued between __scsi_add_device() and dev->sdev
assignment, so it's unsafe for ata_qc_complete() to dereference
dev->sdev->locked without checking whether it's NULL or not.  Fix it.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
14 years agocan-bcm: fix minor heap overflow
Oliver Hartkopp [Wed, 10 Nov 2010 12:10:30 +0000 (12:10 +0000)] 
can-bcm: fix minor heap overflow

On 64-bit platforms the ASCII representation of a pointer may be up to 17
bytes long. This patch increases the length of the buffer accordingly.

http://marc.info/?l=linux-netdev&m=128872251418192&w=2

Reported-by: Dan Rosenberg <drosenberg@vsecurity.com>
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
CC: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agogianfar: Do not call device_set_wakeup_enable() under a spinlock
Rafael J. Wysocki [Tue, 9 Nov 2010 11:54:19 +0000 (11:54 +0000)] 
gianfar: Do not call device_set_wakeup_enable() under a spinlock

The gianfar driver calls device_set_wakeup_enable() under a spinlock,
which causes a problem to happen after the recent core power
management changes, because this function can sleep now.  Fix this
by moving the device_set_wakeup_enable() call out of the
spinlock-protected area.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoipv6: Warn users if maximum number of routes is reached.
Ben Greear [Mon, 8 Nov 2010 12:33:48 +0000 (12:33 +0000)] 
ipv6: Warn users if maximum number of routes is reached.

This gives users at least some clue as to what the problem
might be and how to go about fixing it.

Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agodocs: Add neigh/gc_thresh3 and route/max_size documentation.
Ben Greear [Mon, 8 Nov 2010 09:13:48 +0000 (09:13 +0000)] 
docs: Add neigh/gc_thresh3 and route/max_size documentation.

Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoaxnet_cs: fix resume problem for some Ax88790 chip
Ken Kawasaki [Sat, 6 Nov 2010 05:11:24 +0000 (05:11 +0000)] 
axnet_cs: fix resume problem for some Ax88790 chip

axnet_cs:
    Some Ax88790 chip need to reinitialize the CISREG_CCSR register
    after resume.

Signed-off-by: Ken Kawasaki <ken_kawasaki@spring.nifty.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoipv6: addrconf: don't remove address state on ifdown if the address is being kept
Lorenzo Colitti [Wed, 27 Oct 2010 18:16:49 +0000 (18:16 +0000)] 
ipv6: addrconf: don't remove address state on ifdown if the address is being kept

Currently, addrconf_ifdown does not delete statically configured IPv6
addresses when the interface is brought down. The intent is that when
the interface comes back up the address will be usable again. However,
this doesn't actually work, because the system stops listening on the
corresponding solicited-node multicast address, so the address cannot
respond to neighbor solicitations and thus receive traffic. Also, the
code notifies the rest of the system that the address is being deleted
(e.g, RTM_DELADDR), even though it is not. Fix it so that none of this
state is updated if the address is being kept on the interface.

Tested: Added a statically configured IPv6 address to an interface,
started ping, brought link down, brought link up again. When link came
up ping kept on going and "ip -6 maddr" showed that the host was still
subscribed to there

Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agotcp: Don't change unlocked socket state in tcp_v4_err().
David S. Miller [Fri, 12 Nov 2010 21:35:00 +0000 (13:35 -0800)] 
tcp: Don't change unlocked socket state in tcp_v4_err().

Alexey Kuznetsov noticed a regression introduced by
commit f1ecd5d9e7366609d640ff4040304ea197fbc618
("Revert Backoff [v3]: Revert RTO on ICMP destination unreachable")

The RTO and timer modification code added to tcp_v4_err()
doesn't check sock_owned_by_user(), which if true means we
don't have exclusive access to the socket and therefore cannot
modify it's critical state.

Just skip this new code block if sock_owned_by_user() is true
and eliminate the now superfluous sock_owned_by_user() code
block contained within.

Reported-by: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
CC: Damian Lukowski <damian@tvk.rwth-aachen.de>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
This page took 0.049366 seconds and 5 git commands to generate.