8 years agogeneric-ipi: Fix deadlock in __smp_call_function_single
Heiko Carstens [Fri, 10 Sep 2010 11:47:29 +0000]
generic-ipi: Fix deadlock in __smp_call_function_single

Just got my 6 way machine to a state where cpu 0 is in an
endless loop within __smp_call_function_single.
All other cpus are idle.

The call trace on cpu 0 looks like this:

 __smp_call_function_single
 scheduler_tick
 update_process_times
 tick_sched_timer
 __run_hrtimer
 hrtimer_interrupt
 clock_comparator_work
 do_extint
 ext_int_handler
 ----> timer irq
 cpu_idle

__smp_call_function_single() got called from nohz_balancer_kick()
(inlined) with the remote cpu being 1, wait being 0 and the per
cpu variable remote_sched_softirq_cb (call_single_data) of the
current cpu (0).

Then it loops forever when it tries to grab the lock of the
call_single_data, since it is already locked and enqueued on cpu 0.

My theory how this could have happened: for some reason the
scheduler decided to call __smp_call_function_single() on it's own
cpu, and sends an IPI to itself. The interrupt stays pending
since IRQs are disabled. If then the hypervisor schedules the
cpu away it might happen that upon rescheduling both the IPI and
the timer IRQ are pending. If then interrupts are enabled again
it depends which one gets scheduled first.
If the timer interrupt gets delivered first we end up with the
local deadlock as seen in the calltrace above.

Let's make __smp_call_function_single() check if the target cpu is
the current cpu and execute the function immediately just like
smp_call_function_single does. That should prevent at least the
scenario described here.

It might also be that the scheduler is not supposed to call
__smp_call_function_single with the remote cpu being the current
cpu, but that is a different issue.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Jens Axboe <jaxboe@fusionio.com>
Cc: Venkatesh Pallipadi <venki@google.com>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

8 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzi...
Linus Torvalds [Fri, 10 Sep 2010 03:28:19 +0000]
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:
  libata-sff: Reenable Port Multiplier after libata-sff remodeling.
  libata: skip EH autopsy and recovery during suspend
  ahci: AHCI and RAID mode SATA patch for Intel Patsburg DeviceIDs
  ata_piix: IDE Mode SATA patch for Intel Patsburg DeviceIDs
  libata,pata_via: revert ata_wait_idle() removal from ata_sff/via_tf_load()
  ahci: fix hang on failed softreset
  pata_artop: Fix device ID parity check

8 years agolibata-sff: Reenable Port Multiplier after libata-sff remodeling.
Gwendal Grignou [Tue, 31 Aug 2010 23:20:36 +0000]
libata-sff: Reenable Port Multiplier after libata-sff remodeling.

Keep track of the link on the which the current request is in progress.
It allows support of links behind port multiplier.

Not all libata-sff is PMP compliant. Code for native BMDMA controller
does not take in accound PMP.

Tested on Marvell 7042 and Sil7526.

Signed-off-by: Gwendal Grignou <gwendal@google.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agolibata: skip EH autopsy and recovery during suspend
Tejun Heo [Tue, 7 Sep 2010 12:05:31 +0000]
libata: skip EH autopsy and recovery during suspend

For some mysterious reason, certain hardware reacts badly to usual EH
actions while the system is going for suspend.  As the devices won't
be needed until the system is resumed, ask EH to skip usual autopsy
and recovery and proceed directly to suspend.

Signed-off-by: Tejun Heo <tj@kernel.org>
Tested-by: Stephan Diestelhorst <stephan.diestelhorst@amd.com>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agoahci: AHCI and RAID mode SATA patch for Intel Patsburg DeviceIDs
Seth Heasley [Thu, 9 Sep 2010 16:44:56 +0000]
ahci: AHCI and RAID mode SATA patch for Intel Patsburg DeviceIDs

This patch adds the Intel Patsburg (PCH) SATA AHCI and RAID Controller
DeviceIDs.

Signed-off-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agoata_piix: IDE Mode SATA patch for Intel Patsburg DeviceIDs
Seth Heasley [Thu, 9 Sep 2010 16:42:40 +0000]
ata_piix: IDE Mode SATA patch for Intel Patsburg DeviceIDs

This patch adds the Intel Patsburg (PCH) IDE mode SATA Controller DeviceIDs.

Signed-off-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agolibata,pata_via: revert ata_wait_idle() removal from ata_sff/via_tf_load()
Tejun Heo [Thu, 9 Sep 2010 15:13:31 +0000]
libata,pata_via: revert ata_wait_idle() removal from ata_sff/via_tf_load()

Commit 978c0666 (libata: Remove excess delay in the tf_load path)
removed ata_wait_idle() from ata_sff_tf_load() and via_tf_load().
This caused obscure detection problems in sata_sil.

  https://bugzilla.kernel.org/show_bug.cgi?id=16606

The commit was pure performance optimization.  Revert it for now.

Reported-by: Dieter Plaetinck <dieter@plaetinck.be>
Reported-by: Jan Beulich <JBeulich@novell.com>
Bisected-by: gianluca <gianluca@sottospazio.it>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agominix: fix regression in minix_mkdir()
Jorge Boncompte [DTI2] [Thu, 9 Sep 2010 23:38:19 +0000]
minix: fix regression in minix_mkdir()

Commit 9eed1fb721c ("minix: replace inode uid,gid,mode init with helper")
broke directory creation on minix filesystems.

Fix it by passing the needed mode flag to inode init helper.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Cc: Dmitry Monakhov <dmonakhov@openvz.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@kernel.org> [2.6.35.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: page allocator: drain per-cpu lists after direct reclaim allocation fails
Mel Gorman [Thu, 9 Sep 2010 23:38:18 +0000]
mm: page allocator: drain per-cpu lists after direct reclaim allocation fails

When under significant memory pressure, a process enters direct reclaim
and immediately afterwards tries to allocate a page.  If it fails and no
further progress is made, it's possible the system will go OOM.  However,
on systems with large amounts of memory, it's possible that a significant
number of pages are on per-cpu lists and inaccessible to the calling
process.  This leads to a process entering direct reclaim more often than
it should increasing the pressure on the system and compounding the
problem.

This patch notes that if direct reclaim is making progress but allocations
are still failing that the system is already under heavy pressure.  In
this case, it drains the per-cpu lists and tries the allocation a second
time before continuing.

Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: page allocator: calculate a better estimate of NR_FREE_PAGES when memory is low...
Christoph Lameter [Thu, 9 Sep 2010 23:38:17 +0000]
mm: page allocator: calculate a better estimate of NR_FREE_PAGES when memory is low and kswapd is awake

Ordinarily watermark checks are based on the vmstat NR_FREE_PAGES as it is
cheaper than scanning a number of lists.  To avoid synchronization
overhead, counter deltas are maintained on a per-cpu basis and drained
both periodically and when the delta is above a threshold.  On large CPU
systems, the difference between the estimated and real value of
NR_FREE_PAGES can be very high.  If NR_FREE_PAGES is much higher than
number of real free page in buddy, the VM can allocate pages below min
watermark, at worst reducing the real number of pages to zero.  Even if
the OOM killer kills some victim for freeing memory, it may not free
memory if the exit path requires a new page resulting in livelock.

This patch introduces a zone_page_state_snapshot() function (courtesy of
Christoph) that takes a slightly more accurate view of an arbitrary vmstat
counter.  It is used to read NR_FREE_PAGES while kswapd is awake to avoid
the watermark being accidentally broken.  The estimate is not perfect and
may result in cache line bounces but is expected to be lighter than the
IPI calls necessary to continually drain the per-cpu counters while kswapd
is awake.

Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: page allocator: update free page counters after pages are placed on the free...
Mel Gorman [Thu, 9 Sep 2010 23:38:16 +0000]
mm: page allocator: update free page counters after pages are placed on the free list

When allocating a page, the system uses NR_FREE_PAGES counters to
determine if watermarks would remain intact after the allocation was made.
This check is made without interrupts disabled or the zone lock held and
so is race-prone by nature.  Unfortunately, when pages are being freed in
batch, the counters are updated before the pages are added on the list.
During this window, the counters are misleading as the pages do not exist
yet.  When under significant pressure on systems with large numbers of
CPUs, it's possible for processes to make progress even though they should
have been stalled.  This is particularly problematic if a number of the
processes are using GFP_ATOMIC as the min watermark can be accidentally
breached and in extreme cases, the system can livelock.

This patch updates the counters after the pages have been added to the
list.  This makes the allocator more cautious with respect to preserving
the watermarks and mitigates livelock possibilities.

[akpm@linux-foundation.org: avoid modifying incoming args]
Signed-off-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovmstat: update zone stat threshold when onlining a cpu
KAMEZAWA Hiroyuki [Thu, 9 Sep 2010 23:38:14 +0000]
vmstat: update zone stat threshold when onlining a cpu

refresh_zone_stat_thresholds() calculates parameter based on the number of
online cpus.  It's called at cpu offlining but needs to be called at
onlining, too.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agovfs: take O_NONBLOCK out of the O_* uniqueness test
James Bottomley [Thu, 9 Sep 2010 23:38:12 +0000]
vfs: take O_NONBLOCK out of the O_* uniqueness test

O_NONBLOCK on parisc has a dual value:

#define O_NONBLOCK 000200004 /* HPUX has separate NDELAY & NONBLOCK */

It is caught by the O_* bits uniqueness check and leads to a parisc
compile error.  The fix would be to take O_NONBLOCK out.

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
Cc: Jamie Lokier <jamie@shareable.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoswap: discard while swapping only if SWAP_FLAG_DISCARD
Hugh Dickins [Thu, 9 Sep 2010 23:38:11 +0000]
swap: discard while swapping only if SWAP_FLAG_DISCARD

Tests with recent firmware on Intel X25-M 80GB and OCZ Vertex 60GB SSDs
show a shift since I last tested in December: in part because of firmware
updates, in part because of the necessary move from barriers to awaiting
completion at the block layer.  While discard at swapon still shows as
slightly beneficial on both, discarding 1MB swap cluster when allocating
is now disadvanteous: adds 25% overhead on Intel, adds 230% on OCZ (YMMV).

Surrender: discard as presently implemented is more hindrance than help
for swap; but might prove useful on other devices, or with improvements.
So continue to do the discard at swapon, but make discard while swapping
conditional on a SWAP_FLAG_DISCARD to sys_swapon() (which has been using
only the lower 16 bits of int flags).

We can add a --discard or -d to swapon(8), and a "discard" to swap in
/etc/fstab: matching the mount option for btrfs, ext4, fat, gfs2, nilfs2.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Nigel Cunningham <nigel@tuxonice.net>
Cc: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <jaxboe@fusionio.com>
Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoswap: do not send discards as barriers
Christoph Hellwig [Thu, 9 Sep 2010 23:38:10 +0000]
swap: do not send discards as barriers

The swap code already uses synchronous discards, no need to add I/O
barriers.

This fixes the worst of the terrible slowdown in swap allocation for
hibernation, reported on 2.6.35 by Nigel Cunningham; but does not entirely
eliminate that regression.

[tj@kernel.org: superflous newlines removed]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Nigel Cunningham <nigel@tuxonice.net>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoswap: prevent reuse during hibernation
Hugh Dickins [Thu, 9 Sep 2010 23:38:09 +0000]
swap: prevent reuse during hibernation

Move the hibernation check from scan_swap_map() into try_to_free_swap():
to catch not only the common case when hibernation's allocation itself
triggers swap reuse, but also the less likely case when concurrent page
reclaim (shrink_page_list) might happen to try_to_free_swap from a page.

Hibernation already clears __GFP_IO from the gfp_allowed_mask, to stop
reclaim from going to swap: check that to prevent swap reuse too.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Ondrej Zary <linux@rainbow-software.org>
Cc: Andrea Gelmini <andrea.gelmini@gmail.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Nigel Cunningham <nigel@tuxonice.net>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoswap: revert special hibernation allocation
Hugh Dickins [Thu, 9 Sep 2010 23:38:07 +0000]
swap: revert special hibernation allocation

Please revert 2.6.36-rc commit d2997b1042ec150616c1963b5e5e919ffd0b0ebf
"hibernation: freeze swap at hibernation".  It complicated matters by
adding a second swap allocation path, just for hibernation; without in any
way fixing the issue that it was intended to address - page reclaim after
fixing the hibernation image might free swap from a page already imaged as
swapcache, letting its swap be reallocated to store a different page of
the image: resulting in data corruption if the imaged page were freed as
clean then swapped back in.  Pages freed to si->swap_map were still in
danger of being reallocated by the alternative allocation path.

I guess it inadvertently fixed slow SSD swap allocation for hibernation,
as reported by Nigel Cunningham: by missing out the discards that occur on
the usual swap allocation path; but that was unintentional, and needs a
separate fix.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Ondrej Zary <linux@rainbow-software.org>
Cc: Andrea Gelmini <andrea.gelmini@gmail.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Nigel Cunningham <nigel@tuxonice.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agobounce: call flush_dcache_page() after bounce_copy_vec()
Gary King [Thu, 9 Sep 2010 23:38:05 +0000]
bounce: call flush_dcache_page() after bounce_copy_vec()

I have been seeing problems on Tegra 2 (ARMv7 SMP) systems with HIGHMEM
enabled on 2.6.35 (plus some patches targetted at 2.6.36 to perform cache
maintenance lazily), and the root cause appears to be that the mm bouncing
code is calling flush_dcache_page before it copies the bounce buffer into
the bio.

The bounced page needs to be flushed after data is copied into it, to
ensure that architecture implementations can synchronize instruction and
data caches if necessary.

Signed-off-by: Gary King <gking@nvidia.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Acked-by: Jens Axboe <axboe@kernel.dk>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agodrivers/rtc/rtc-pl031.c: do not mark PL031 IRQ as shared
Linus Walleij [Thu, 9 Sep 2010 23:38:04 +0000]
drivers/rtc/rtc-pl031.c: do not mark PL031 IRQ as shared

It was a mistake to mark the PL031 IRQ as shared (for the U8500),
we misread the datasheet. Get rid of this.

Signed-off-by: Linus Walleij <linus.walleij@stericsson.com>
Cc: Jonas Aberg <jonas.aberg@stericsson.com>
Cc: Mian Yousaf Kaukab <mian.yousaf.kaukab@stericsson.com>
Acked-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoMAINTAINERS: correct entry for legacy RTC-driver
Wolfram Sang [Thu, 9 Sep 2010 23:38:04 +0000]
MAINTAINERS: correct entry for legacy RTC-driver

Because no one dared to remove it so far, let's keep the entry correct, at
least.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Mike Frysinger <vapier@gentoo.org>
Cc: Paul Gortmaker <p_gortmaker@yahoo.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agogpio: doc updates
David Brownell [Thu, 9 Sep 2010 23:38:03 +0000]
gpio: doc updates

There's been some recent confusion about error checking GPIO numbers.
briefly, it should be handled mostly during setup, when gpio_request() is
called, and NEVER by expectig gpio_is_valid to report more than
never-usable GPIO numbers.

[akpm@linux-foundation.org: terminate unterminated comment]
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Cc: Eric Miao" <eric.y.miao@gmail.com>
Cc: "Ryan Mallon" <ryan@bluewatersys.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agogpio: sx150x: correct and refine reset-on-probe behavior
Gregory Bean [Thu, 9 Sep 2010 23:38:02 +0000]
gpio: sx150x: correct and refine reset-on-probe behavior

Replace the arbitrary software-reset call from the device-probe
method, because:

- It is defective.  To work correctly, it should be two byte writes,
  not a single word write.  As it stands, it does nothing.

- Some devices with sx150x expanders installed have their NRESET pins
  ganged on the same line, so resetting one causes the others to reset -
  not a nice thing to do arbitrarily!

- The probe, usually taking place at boot, implies a recent hard-reset,
  so a software reset at this point is just a waste of energy anyway.

Therefore, make it optional, defaulting to off, as this will match the
common case of probing at powerup and also matches the current broken
no-op behavior.

Signed-off-by: Gregory Bean <gbean@codeaurora.org>
Reviewed-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomemory hotplug: fix next block calculation in is_removable
KAMEZAWA Hiroyuki [Thu, 9 Sep 2010 23:38:01 +0000]
memory hotplug: fix next block calculation in is_removable

next_active_pageblock() is for finding next _used_ freeblock.  It skips
several blocks when it finds there are a chunk of free pages lager than
pageblock.  But it has 2 bugs.

  1. We have no lock. page_order(page) - pageblock_order can be minus.
  2. pageblocks_stride += is wrong. it should skip page_order(p) of pages.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: compaction: handle active and inactive fairly in too_many_isolated
Minchan Kim [Thu, 9 Sep 2010 23:38:00 +0000]
mm: compaction: handle active and inactive fairly in too_many_isolated

Iram reported that compaction's too_many_isolated() loops forever.
(http://www.spinics.net/lists/linux-mm/msg08123.html)

The meminfo when the situation happened was inactive anon is zero.  That's
because the system has no memory pressure until then.  While all anon
pages were in the active lru, compaction could select active lru as well
as inactive lru.  That's a different thing from vmscan's isolated.  So we
has been two too_many_isolated.

While compaction can isolate pages in both active and inactive, current
implementation of too_many_isolated only considers inactive.  It made
Iram's problem.

This patch handles active and inactive fairly.  That's because we can't
expect where from and how many compaction would isolated pages.

This patch changes (nr_isolated > nr_inactive) with
nr_isolated > (nr_active + nr_inactive) / 2.

Signed-off-by: Minchan Kim <minchan.kim@gmail.com>
Reported-by: Iram Shahzad <iram.shahzad@jp.fujitsu.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agokernel/groups.c: fix integer overflow in groups_search
Jerome Marchand [Thu, 9 Sep 2010 23:37:59 +0000]
kernel/groups.c: fix integer overflow in groups_search

gid_t is a unsigned int.  If group_info contains a gid greater than
MAX_INT, groups_search() function may look on the wrong side of the search
tree.

This solves some unfair "permission denied" problems.

Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoarch/powerpc/include/asm/fsldma.h needs slab.h
Ira W. Snyder [Thu, 9 Sep 2010 23:37:58 +0000]
arch/powerpc/include/asm/fsldma.h needs slab.h

The slab.h header is required to use the kmalloc() family of functions.
Due to recent kernel changes, this header must be directly included by
code that calls into the memory allocator.

Without this patch, any code which includes this header fails to build.

Signed-off-by: Ira W. Snyder <iws@ovro.caltech.edu>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agortc: m41t80: do not use rtc_valid_tm in m41t80_rtc_read_alarm
Atsushi Nemoto [Thu, 9 Sep 2010 23:37:56 +0000]
rtc: m41t80: do not use rtc_valid_tm in m41t80_rtc_read_alarm

Commit b485fe5ea ("rtc/m41t80: use rtc_valid_tm() to check returned tm")
added rtc_valid_tm to m41t80_rtc_read_alarm() but it was wrong while the
t->time does not contain complete date/time.

This patch also fixes a warning:
warning: passing argument 1 of 'rtc_valid_tm' from incompatible pointer type

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: Wan ZongShun <mcuos.com@gmail.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agobinfmt_misc: fix binfmt_misc priority
Jan Sembera [Thu, 9 Sep 2010 23:37:54 +0000]
binfmt_misc: fix binfmt_misc priority

Commit 74641f584da ("alpha: binfmt_aout fix") (May 2009) introduced a
regression - binfmt_misc is now consulted after binfmt_elf, which will
unfortunately break ia32el.  ia32 ELF binaries on ia64 used to be matched
using binfmt_misc and executed using wrapper.  As 32bit binaries are now
matched by binfmt_elf before bindmt_misc kicks in, the wrapper is ignored.

The fix increases precedence of binfmt_misc to the original state.

Signed-off-by: Jan Sembera <jsembera@suse.cz>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Richard Henderson <rth@twiddle.net
Cc: <stable@kernel.org> [2.6.everything.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: avoid warning when COMPACTION is selected
Andrea Arcangeli [Thu, 9 Sep 2010 23:37:53 +0000]
mm: avoid warning when COMPACTION is selected

COMPACTION enables MIGRATION, but MIGRATION spawns a warning if numa or
memhotplug aren't selected.  However MIGRATION doesn't depend on them.  I
guess it's just trying to be strict doing a double check on who's enabling
it, but it doesn't know that compaction also enables MIGRATION.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: fix swapin race condition
Andrea Arcangeli [Thu, 9 Sep 2010 23:37:52 +0000]
mm: fix swapin race condition

The pte_same check is reliable only if the swap entry remains pinned (by
the page lock on swapcache).  We've also to ensure the swapcache isn't
removed before we take the lock as try_to_free_swap won't care about the
page pin.

One of the possible impacts of this patch is that a KSM-shared page can
point to the anon_vma of another process, which could exit before the page
is freed.

This can leave a page with a pointer to a recycled anon_vma object, or
worse, a pointer to something that is no longer an anon_vma.

[riel@redhat.com: changelog help]
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Acked-by: Hugh Dickins <hughd@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agodrivers/mmc/host/imxmmc.c: adjust confusing if indentation
Julia Lawall [Thu, 9 Sep 2010 23:37:50 +0000]
drivers/mmc/host/imxmmc.c: adjust confusing if indentation

Move the second if (reg & ...) test into the branch indicated by its
indentation.  The test was previously always executed after the if
containing that branch, but it was always false unless the if branch was
taken.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@r disable braces4@
position p1,p2;
statement S1,S2;
@@

(
if (...) { ... }
|
if (...) S1@p1 S2@p2
)

@script:python@
p1 << r.p1;
p2 << r.p2;
@@

if (p1[0].column == p2[0].column):
  cocci.print_main("branch",p1)
  cocci.print_secs("after",p2)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Chris Ball <cjb@laptop.org>
Cc: Pavel Pisa <ppisa@pikron.com>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoomap hsmmc: fix a racing case between kmmcd and omap_hsmmc_suspend
Ethan Du [Thu, 9 Sep 2010 23:37:49 +0000]
omap hsmmc: fix a racing case between kmmcd and omap_hsmmc_suspend

If suspend called when kmmcd is doing host->ops->disable, as kmmcd already
increased host->en_dis_recurs to 1, the mmc_host_enable in suspend
function will return directly without increase the nesting_cnt, which will
cause the followed register access carried out to the disabled host.

mmc_suspend_host will enable host itself.  No need to enable host before
it.  Also works on kmmcd will get flushed in mmc_suspend_host, enable host
after it will be safe.  So make the mmc_host_enable after it.

[cjb: rebase against current Linus]
Signed-off-by: Ethan <ethan.too@gmail.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Acked-by: Adrian Hunter <adrian.hunter@nokia.com>
Acked-by: Madhusudhan Chikkature <madhu.cr@ti.com>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agommc: at91_mci: add missing linux/highmem.h include
Marc Kleine-Budde [Thu, 9 Sep 2010 23:37:48 +0000]
mmc: at91_mci: add missing linux/highmem.h include

Fix the following error:

at91_mci.c: In function 'at91_mci_sg_to_dma':
at91_mci.c:236: error: implicit declaration of function 'kmap_atomic'
at91_mci.c:236: error: 'KM_BIO_SRC_IRQ' undeclared (first use in this function)
at91_mci.c:236: error: (Each undeclared identifier is reported only once
at91_mci.c:236: error: for each function it appears in.)
at91_mci.c:236: warning: assignment makes pointer from integer without a cast
at91_mci.c:252: error: implicit declaration of function 'kunmap_atomic'
at91_mci.c: In function 'at91_mci_post_dma_read':
at91_mci.c:302: error: 'KM_BIO_SRC_IRQ' undeclared (first use in this function)
at91_mci.c:302: warning: assignment makes pointer from integer without a cast
at91_mci.c:317: error: implicit declaration of function 'flush_kernel_dcache_page'

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
Cc: Nicolas Ferre <nicolas.ferre@atmel.com>
Cc: Andrew Victor <avictor.za@gmail.com>
Cc: Wolfgang Muees <wolfgang.mues@auerswald.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoomap_hsmmc: remove unused local `state'
Sergio Aguirre [Thu, 9 Sep 2010 23:37:46 +0000]
omap_hsmmc: remove unused local `state'

This fixes the following warning:

drivers/mmc/host/omap_hsmmc.c: In function 'omap_hsmmc_suspend':
drivers/mmc/host/omap_hsmmc.c:2275: warning: unused variable 'state'

Introduced by commit ID:

  commit 1a13f8fa76c880be41d6b1e6a2b44404bcbfdf9e
  Author: Matt Fleming <matt@console-pimps.org>
  Date:   Wed May 26 14:42:08 2010 -0700

      mmc: remove the "state" argument to mmc_suspend_host()

The unique usage of this var was removed there, and missed
removing the respective declaration aswell.

Signed-off-by: Sergio Aguirre <saaguirre@ti.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
Acked-by: Matt Fleming <matt@console-pimps.org>
Cc: Madhusudhan Chikkature <madhu.cr@ti.com>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoARM: SAMSUNG: MMC: fix build error when both DMA and PIO mode selected
Jiri Pinkava [Tue, 25 May 2010 07:48:58 +0000]
ARM: SAMSUNG: MMC: fix build error when both DMA and PIO mode selected

[cjb: fix line-wrapped patch]
Signed-off-by: Jiri Pinkava <jiri.pinkava@vscht.cz>
Signed-off-by: Chris Ball <cjb@laptop.org>
Cc: Matt Fleming <matt@console-pimps.org>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Ben Dooks <ben-linux@fluff.org>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agommc: fix the use of kunmap_atomic() in tmio_mmc.h
Guennadi Liakhovetski [Thu, 9 Sep 2010 23:37:43 +0000]
mmc: fix the use of kunmap_atomic() in tmio_mmc.h

kunmap_atomic() takes the cookie, returned by the kmap_atomic() as its
argument and not the page address, used as an argument to kmap_atomic().
This patch fixes the compile error:

In file included from drivers/mmc/host/tmio_mmc.c:37:
drivers/mmc/host/tmio_mmc.h: In function 'tmio_mmc_kunmap_atomic':
drivers/mmc/host/tmio_mmc.h:192: error: negative width in bit-field '<anonymous>'

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: Eric Miao <eric.y.miao@gmail.com>
Tested-by: Magnus Damm <damm@opensource.se>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agotmio_mmc: don't clear unhandled pending interrupts
Yusuke Goda [Thu, 9 Sep 2010 23:37:39 +0000]
tmio_mmc: don't clear unhandled pending interrupts

Previously, it was possible for ack_mmc_irqs() to clear pending interrupt
bits in the CTL_STATUS register, even though the interrupt handler had not
been called.  This was because of a race that existed when doing a
read-modify-write sequence on CTL_STATUS.  After the read step in this
sequence, if an interrupt occurred (causing one of the bits in CTL_STATUS
to be set) the write step would inadvertently clear it.

Observed with the TMIO_STAT_RXRDY bit together with CMD53 on AR6002 and
BCM4318 SDIO cards in polled mode.

This patch eliminates this race by only writing to CTL_STATUS and clearing
the interrupts that were passed as an argument to ack_mmc_irqs()."

[matt@console-pimps.org: rewrote changelog]
Signed-off-by: Yusuke Goda <yusuke.goda.sx@renesas.com>
Acked-by: Magnus Damm <damm@opensource.se>"
Tested-by: Arnd Hannemann <arnd@arndnet.de>"
Acked-by: Ian Molton <ian@mnementh.co.uk>
Cc: Matt Fleming <matt@console-pimps.org>
Cc: Samuel Ortiz <sameo@linux.intel.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: <linux-mmc@vger.kernel.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agocgroups: fix API thinko
Michael S. Tsirkin [Thu, 9 Sep 2010 23:37:37 +0000]
cgroups: fix API thinko

Add cgroup_attach_task_all()

The existing cgroup_attach_task_current_cg() API is called by a thread to
attach another thread to all of its cgroups; this is unsuitable for cases
where a privileged task wants to attach itself to the cgroups of a less
privileged one, since the call must be made from the context of the target
task.

This patch adds a more generic cgroup_attach_task_all() API that allows
both the source task and to-be-moved task to be specified.
cgroup_attach_task_current_cg() becomes a specialization of the more
generic new function.

[menage@google.com: rewrote changelog]
[akpm@linux-foundation.org: address reviewer comments]
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Acked-by: Paul Menage <menage@google.com>
Cc: Li Zefan <lizf@cn.fujitsu.com>
Cc: Ben Blum <bblum@google.com>
Cc: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoproc: export uncached bit properly in /proc/kpageflags
Takashi Iwai [Thu, 9 Sep 2010 23:37:36 +0000]
proc: export uncached bit properly in /proc/kpageflags

Fix the left-over old ifdef for PG_uncached in /proc/kpageflags.  Now it's
used by x86, too.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agogcov: fix null-pointer dereference for certain module types
Peter Oberparleiter [Thu, 9 Sep 2010 23:37:35 +0000]
gcov: fix null-pointer dereference for certain module types

The gcov-kernel infrastructure expects that each object file is loaded
only once.  This may not be true, e.g.  when loading multiple kernel
modules which are linked to the same object file.  As a result, loading
such kernel modules will result in incorrect gcov results while unloading
will cause a null-pointer dereference.

This patch fixes these problems by changing the gcov-kernel infrastructure
so that multiple profiling data sets can be associated with one debugfs
entry.  It applies to 2.6.36-rc1.

Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reported-by: Werner Spies <werner.spies@thalesgroup.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoMAINTAINERS: kexec-tools has moved
Simon Horman [Thu, 9 Sep 2010 23:37:33 +0000]
MAINTAINERS: kexec-tools has moved

Kexec tools has been moved to http://kernel.org/pub/linux/utils/kernel/kexec/
as user-space code shouldn't be in /pub/linux/kernel

Signed-off-by: Simon Horman <horms@verge.net.au>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoO_DIRECT: fix the splitting up of contiguous I/O
Jeff Moyer [Thu, 9 Sep 2010 23:37:33 +0000]
O_DIRECT: fix the splitting up of contiguous I/O

commit c2c6ca4 (direct-io: do not merge logically non-contiguous requests)
introduced a bug whereby all O_DIRECT I/Os were submitted a page at a time
to the block layer.  The problem is that the code expected
dio->block_in_file to correspond to the current page in the dio.  In fact,
it corresponds to the previous page submitted via submit_page_section.
This was purely an oversight, as the dio->cur_page_fs_offset field was
introduced for just this purpose.  This patch simply uses the correct
variable when calculating whether there is a mismatch between contiguous
logical blocks and contiguous physical blocks (as described in the
comments).

I also switched the if conditional following this check to an else if, to
ensure that we never call dio_bio_submit twice for the same dio (in
theory, this should not happen, anyway).

I've tested this by running blktrace and verifying that a 64KB I/O was
submitted as a single I/O.  I also ran the patched kernel through
xfstests' aio tests using xfs, ext4 (with 1k and 4k block sizes) and btrfs
and verified that there were no regressions as compared to an unpatched
kernel.

Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Josef Bacik <jbacik@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: <stable@kernel.org> [2.6.35.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agohp_accel: add quirks for HP ProBook 532x and HP Mini 5102
Takashi Iwai [Thu, 9 Sep 2010 23:37:31 +0000]
hp_accel: add quirks for HP ProBook 532x and HP Mini 5102

Added missing axis-mapping for HP ProBook 532x and HP Mini 5102.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Cc: Eric Piel <eric.piel@tremplin-utc.net>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoMAINTAINERS: remove drivers/char/hpet.c
Jiri Slaby [Thu, 9 Sep 2010 23:37:30 +0000]
MAINTAINERS: remove drivers/char/hpet.c

bob.picco@hp.com doesn't work any more and Bob says that he's unlikely to
work on hpet.c in the future.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Acked-by: Bob Picco <bpicco@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agortc-bfin: fix state restoration when resuming
Mike Frysinger [Thu, 9 Sep 2010 23:37:29 +0000]
rtc-bfin: fix state restoration when resuming

Much (but not all) of the RTC state is kept in the RTC peripheral which
has its own power domain.  Periodically (1 HZ), that state is synced from
one power domain to the other (peripheral->core).  When we are resuming,
we need to wait for the sync to occur so that we don't get a mismatch of
reading undefined state in the rest of the driver.

Further, once the externally maintained bits have been synced back into
the core, we then need to restore the bits maintained in the core.  In our
particular case, that is just the write completion interrupt bit.

If we don't do any of this, working with the RTC causes ~5 second delays
from time to time after waking up due to the write completion interrupt
never firing.

Reported-by: Michael Dean <mdean@aeronix.com>
Reported-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agortc-bfin: fix inverted logic in suspend path
Mike Frysinger [Thu, 9 Sep 2010 23:37:27 +0000]
rtc-bfin: fix inverted logic in suspend path

The int_clear helper takes a bitmask of interrupts to keep, not to
disable.  When suspending without wakeup enabled, we want to disable
all interrupts, so use 0 (keep none) instead of -1 (keep all).

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agokfifo: add parenthesis for macro parameter reference
Huang Ying [Thu, 9 Sep 2010 23:37:26 +0000]
kfifo: add parenthesis for macro parameter reference

Some macro parameter references inside typeof() operator are not enclosed
with parenthesis.  It should be safer to add them.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Acked-by: Stefani Seibold <stefani@seibold.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agommc: avoid getting CID on SDIO-only cards
David Vrabel [Thu, 9 Sep 2010 23:37:24 +0000]
mmc: avoid getting CID on SDIO-only cards

The introduction of support for SD combo cards breaks the initialization
of all CSR SDIO chips.  The GO_IDLE (CMD0) in mmc_sd_get_cid() causes CSR
chips to be reset (this is non-standard behavior).

When initializing an SDIO card check for a combo card by using the memory
present bit in the R4 response to IO_SEND_OP_COND (CMD5).  This avoids the
call to mmc_sd_get_cid() on an SDIO-only card.

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Acked-by: Michal Mirolaw <mirq-linux@rere.qmqm.pl>
Cc: <linux-mmc@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Fri, 10 Sep 2010 01:31:34 +0000]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm: (30 commits)
  ARM: Update mach-types
  ARM: Partially revert "Auto calculate ZRELADDR and provide option for exceptions"
  ARM: Ensure PTE modifications via dma_alloc_coherent are visible
  ARM: 6359/1: ep93xx: move clock initialization earlier
  Revert "[ARM] pxa: remove now unnecessary dma_needs_bounce()"
  ARM: 6352/1: perf: fix event validation
  ARM: 6344/1: Mark CPU_32v6K as depended on CPU_V7
  ARM: 6343/1: wire up fanotify and prlimit64 syscalls on ARM
  ARM: 6330/1: perf: reword comments relating to perf_event_do_pending
  ARM: pxa168fb: fix section mismatch
  ARM: pxa: Make id const in pwm_probe()
  ARM: pxa: fix CI_HSYNC and CI_VSYNC MFP defines for pxa300
  ARM: pxa: remove __init from cpufreq_driver->init()
  ARM: imx: set cache line size to 64 bytes for i.MX5
  mx5/clock: fix clear bit fields issue in _clk_ccgr_disable function
  mxc/tzic: add base address when accessing TZIC registers
  ARM: mach-shmobile: ap4evb: fix write protect for SDHI1
  ARM: mach-shmobile: ap4evb: modify FSI2 ID
  ARM: mach-shmobile: do not enable the PLLC2 clock on init
  ARM: mach-shmobile: Clock framework comment fix
  ...

8 years agoARM: Update mach-types
Russell King [Thu, 9 Sep 2010 21:43:02 +0000]
ARM: Update mach-types

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

8 years agoARM: Partially revert "Auto calculate ZRELADDR and provide option for exceptions"
Russell King [Thu, 9 Sep 2010 21:39:41 +0000]
ARM: Partially revert "Auto calculate ZRELADDR and provide option for exceptions"

Partially revert e69edc7, which introduced automatic zreladdr
support.  The change in the way the manual definition is defined
seems to be error and conflict prone.  Go back to the original way
we were handling this for the time being, while keeping the automatic
zreladdr facility.

Acked-by: Uwe Kleine-K├Ânig <u.kleine-koenig@pengutronix.de
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

8 years agoMerge branch 'origin'
Russell King [Thu, 9 Sep 2010 21:38:43 +0000]
Merge branch 'origin'

8 years agoahci: fix hang on failed softreset
Tejun Heo [Fri, 27 Aug 2010 09:09:15 +0000]
ahci: fix hang on failed softreset

ahci_do_softreset() compared the current time and deadline in reverse
when calculating timeout for SRST issue.  The result is that if
@deadline is in future, SRST is issued with 0 timeout, which hasn't
caused any problem because it later waits for DRDY with the correct
timeout.  If deadline is already exceeded by the time SRST is about to
be issued, the timeout calculation underflows and if the device
doesn't respond, timeout doesn't trigger for a _very_ long time.

Reverse the incorrect comparison order.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Anssi Hannula <anssi.hannula@iki.fi>
Tested-by: Gwendal Grignou <gwendal@google.com>
Cc: stable@kernel.org
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agopata_artop: Fix device ID parity check
Jean Delvare [Mon, 30 Aug 2010 15:37:05 +0000]
pata_artop: Fix device ID parity check

x % 1 always evaluates to 0, which clearly isn't the intent. The
author probably had "% 2" or "& 1" in mind, and mispelled it.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: Alan Cox <alan@linux.intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

8 years agolglock: make lg_lock_global() actually lock globally
Jonathan Corbet [Wed, 8 Sep 2010 22:54:54 +0000]
lglock: make lg_lock_global() actually lock globally

lg_lock_global() currently only acquires spinlocks for online CPUs, but
it's meant to lock all possible CPUs.  Lglock-protected resources may be
associated with removed CPUs - and, indeed, that could happen with the
per-superblock open files lists.

At Nick's suggestion, change for_each_online_cpu() to
for_each_possible_cpu() to protect accesses to those resources.

Cc: Al Viro <viro@ZenIV.linux.org.uk>
Acked-by: Nick Piggin <npiggin@kernel.dk>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agomm: Move vma_stack_continue into mm.h
Stefan Bader [Tue, 31 Aug 2010 13:52:27 +0000]
mm: Move vma_stack_continue into mm.h

So it can be used by all that need to check for that.

Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Thu, 9 Sep 2010 15:58:52 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/nes: Fix hang with modified FIN handling on A0 cards
  RDMA/nes: Change state to closing after FIN
  RDMA/nes: Fix double CLOSE event indication crash
  RDMA/nes: Write correct register write to set TX pause param
  RDMA/cxgb3: Don't exceed the max HW CQ depth

8 years agoMerge branch 'fixes' of git://oss.oracle.com/git/tma/linux-2.6
Linus Torvalds [Thu, 9 Sep 2010 15:57:02 +0000]
Merge branch 'fixes' of git://oss.oracle.com/git/tma/linux-2.6

* 'fixes' of git://oss.oracle.com/git/tma/linux-2.6:
  ocfs2: Fix orphan add in ocfs2_create_inode_in_orphan
  ocfs2: split out ocfs2_prepare_orphan_dir() into locking and prep functions
  ocfs2: allow return of new inode block location before allocation of the inode
  ocfs2: use ocfs2_alloc_dinode_update_counts() instead of open coding
  ocfs2: split out inode alloc code from ocfs2_mknod_locked
  Ocfs2: Fix a regression bug from mainline commit(6b933c8e6f1a2f3118082c455eef25f9b1ac7b45).
  ocfs2: Fix deadlock when allocating page
  ocfs2: properly set and use inode group alloc hint
  ocfs2: Use the right group in nfs sync check.
  ocfs2: Flush drive's caches on fdatasync
  ocfs2: make __ocfs2_page_mkwrite handle file end properly.
  ocfs2: Fix incorrect checksum validation error
  ocfs2: Fix metaecc error messages

8 years agoMerge branches 'cxgb3' and 'nes' into for-linus
Roland Dreier [Wed, 8 Sep 2010 21:43:28 +0000]
Merge branches 'cxgb3' and 'nes' into for-linus

8 years agoRDMA/nes: Fix hang with modified FIN handling on A0 cards
Faisal Latif [Wed, 1 Sep 2010 15:43:11 +0000]
RDMA/nes: Fix hang with modified FIN handling on A0 cards

Changing state to CLOSING when FIN is received causes A0 cards to
hang.  Fix this by checking for A0 cards in FIN handling.

Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

8 years agoRDMA/nes: Change state to closing after FIN
Faisal Latif [Sat, 14 Aug 2010 21:05:04 +0000]
RDMA/nes: Change state to closing after FIN

When the driver receives an AE for FIN received, it closes the
connection without changing the state of the connection in the
hardware to closing.  By changing the state to closing, hardware will
do a normal close sequence.

Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

8 years agoRDMA/nes: Fix double CLOSE event indication crash
Faisal Latif [Sat, 14 Aug 2010 21:04:56 +0000]
RDMA/nes: Fix double CLOSE event indication crash

During a stress testing in a large cluster, multiple close event are
detected and BUG() is hit in the iWARP core.  The cause is that the
active node gave up while waiting for an MPA response from the peer
and tried to close the connection by sending RST.  The passive node
driver receives the RST but is waiting for MPA response from the user.
When the MPA accept is received, the driver offloads the connection
and sends a CLOSE event.  The driver gets an AE indicating RESET
received and also sends a CLOSE event, hitting a BUG().

Fix this by correcting RESET handling and sending CLOSE events.

Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

8 years agoRDMA/nes: Write correct register write to set TX pause param
Chien Tung [Tue, 7 Sep 2010 16:31:20 +0000]
RDMA/nes: Write correct register write to set TX pause param

Setting TX pause param writes to the wrong register location causing
the adapter to hang.  Correct the define used to write the reigster.

Addresses: https://bugs.openfabrics.org/show_bug.cgi?id=2116
Reported-by: Shiri Franchi <shirif@voltaire.com>
Signed-off-by: Chien Tung <chien.tin.tung@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Wed, 8 Sep 2010 18:19:18 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: i8042 - fix device removal on unload
  Input: bcm5974 - adjust major/minor to scale
  Input: MT - initialize slots to unused
  Input: use PIT_TICK_RATE in vt beep ioctl
  Input: wacom - fix mousewheel handling for old wacom tablets

8 years agoMerge branch 'semaphore-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 8 Sep 2010 18:15:51 +0000]
Merge branch 'semaphore-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'semaphore-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  semaphore: Add DEFINE_SEMAPHORE

8 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 8 Sep 2010 18:14:10 +0000]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86, mcheck: Avoid duplicate sysfs links/files for thresholding banks
  io-mapping: Fix the address space annotations
  x86: Fix the address space annotations of iomap_atomic_prot_pfn()
  x86, mm: Fix CONFIG_VMSPLIT_1G and 2G_OPT trampoline
  x86, hwmon: Fix unsafe smp_processor_id() in thermal_throttle_add_dev

8 years agoMerge branch 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 8 Sep 2010 18:13:42 +0000]
Merge branch 'core-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  gcc-4.6: kernel/*: Fix unused but set warnings
  mutex: Fix annotations to include it in kernel-locking docbook
  pid: make setpgid() system call use RCU read-side critical section
  MAINTAINERS: Add RCU's public git tree

8 years agoMerge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 8 Sep 2010 18:13:16 +0000]
Merge branch 'perf-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  perf, x86: Try to handle unknown nmis with an enabled PMU
  perf, x86: Fix handle_irq return values
  perf, x86: Fix accidentally ack'ing a second event on intel perf counter
  oprofile, x86: fix init_sysfs() function stub
  lockup_detector: Sync touch_*_watchdog back to old semantics
  tracing: Fix a race in function profile
  oprofile, x86: fix init_sysfs error handling
  perf_events: Fix time tracking for events with pid != -1 and cpu != -1
  perf: Initialize callchains roots's childen hits
  oprofile: fix crash when accessing freed task structs

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi...
Linus Torvalds [Wed, 8 Sep 2010 18:12:59 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/mszeredi/fuse

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
  fuse: fix lock annotations
  fuse: flush background queue on connection close

8 years agoARM: Ensure PTE modifications via dma_alloc_coherent are visible
Russell King [Wed, 8 Sep 2010 15:27:56 +0000]
ARM: Ensure PTE modifications via dma_alloc_coherent are visible

Dave Hylands reports:
| We've observed a problem with dma_alloc_writecombine when the system
| is under heavy load (heavy bus traffic).  We've managed to reduce the
| problem to the following snippet, which is run from a kthread in a
| continuous loop:
|
|   void *virtAddr;
|   dma_addr_t physAddr;
|   unsigned int numBytes = 256;
|
|   for (;;) {
|       virtAddr = dma_alloc_writecombine(NULL,
|             numBytes, &physAddr, GFP_KERNEL);
|       if (virtAddr == NULL) {
|          printk(KERN_ERR "Running out of memory\n");
|          break;
|       }
|
|       /* access DMA memory allocated */
|       tmp = virtAddr;
|       *tmp = 0x77;
|
|       /* free DMA memory */
|       dma_free_writecombine(NULL,
|             numBytes, virtAddr, physAddr);
|
|         ...sleep here...
|     }
|
| By itself, the code will run forever with no issues. However, as we
| increase our bus traffic (typically using DMA) then the *tmp = 0x77
| line will eventually cause a page fault. If we add a small delay (a
| few microseconds) before the *tmp = 0x77, then we don't see a page
| fault, even under heavy load.

A dsb() is required after modifying the PTE entries to ensure that they
will always be visible.  Add this dsb().

Reported-by: Dave Hylands <dhylands@gmail.com>
Tested-by: Dave Hylands <dhylands@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

8 years agosemaphore: Add DEFINE_SEMAPHORE
Thomas Gleixner [Tue, 7 Sep 2010 12:46:37 +0000]
semaphore: Add DEFINE_SEMAPHORE

The full cleanup of init_MUTEX[_LOCKED] and DECLARE_MUTEX has not been
done. Some of the users are real semaphores and we should name them as
such instead of confusing everyone with "MUTEX".

Provide the infrastructure to get finally rid of init_MUTEX[_LOCKED]
and DECLARE_MUTEX.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Christoph Hellwig <hch@infradead.org>
LKML-Reference: <20100907125054.795929962@linutronix.de>

8 years agoARM: 6359/1: ep93xx: move clock initialization earlier
Mika Westerberg [Fri, 3 Sep 2010 16:14:54 +0000]
ARM: 6359/1: ep93xx: move clock initialization earlier

Commit 7cfe24947 ("ARM: AMBA: Add pclk support to AMBA bus
infrastructure") changed AMBA bus to handle the PCLK automatically.
However, in EP93xx clock initialization is arch_initcall which is done
later than AMBA device identification. This causes
amba_get_enable_pclk() to fail resulting device where UARTs are not
functional.

So change ep93xx_clock_init() to be postcore_initcall.

Signed-off-by: Mika Westerberg <mika.westerberg@iki.fi>
Acked-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

8 years agoRevert "[ARM] pxa: remove now unnecessary dma_needs_bounce()"
Russell King [Sun, 5 Sep 2010 11:19:50 +0000]
Revert "[ARM] pxa: remove now unnecessary dma_needs_bounce()"

This reverts commit 4fa5518, which causes a compilation regression for
IXP4xx platforms.

Reported-by: Richard Cochran <richardcochran@gmail.com>
Acked-by: Eric Miao <eric.y.miao@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

8 years agoocfs2: Fix orphan add in ocfs2_create_inode_in_orphan
Mark Fasheh [Fri, 13 Aug 2010 22:15:19 +0000]
ocfs2: Fix orphan add in ocfs2_create_inode_in_orphan

ocfs2_create_inode_in_orphan() is used by reflink to create the newly
reflinked inode simultaneously in the orphan dir. This allows us to easily
handle partially-reflinked files during recovery cleanup.

We have a problem though - the orphan dir stringifies inode # to determine
a unique name under which the orphan entry dirent can be created. Since
ocfs2_create_inode_in_orphan() needs the space allocated in the orphan dir
before it can allocate the inode, we currently call into the orphan code:

       /*
        * We give the orphan dir the root blkno to fake an orphan name,
        * and allocate enough space for our insertion.
        */
       status = ocfs2_prepare_orphan_dir(osb, &orphan_dir,
                                         osb->root_blkno,
                                         orphan_name, &orphan_insert);

Using osb->root_blkno might work fine on unindexed directories, but the
orphan dir can have an index.  When it has that index, the above code fails
to allocate the proper index entry.  Later, when we try to remove the file
from the orphan dir (using the actual inode #), the reflink operation will
fail.

To fix this, I created a function ocfs2_alloc_orphaned_file() which uses the
newly split out orphan and inode alloc code to figure out what the inode
block number will be (once allocated) and then prepare the orphan dir from
that data.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: split out ocfs2_prepare_orphan_dir() into locking and prep functions
Mark Fasheh [Fri, 13 Aug 2010 22:15:18 +0000]
ocfs2: split out ocfs2_prepare_orphan_dir() into locking and prep functions

We do this because ocfs2_create_inode_in_orphan() wants to order locking of
the orphan dir with respect to locking of the inode allocator *before*
making any changes to the directory.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: allow return of new inode block location before allocation of the inode
Mark Fasheh [Fri, 13 Aug 2010 22:15:17 +0000]
ocfs2: allow return of new inode block location before allocation of the inode

This allows code which needs to know the eventual block number of an inode
but can't allocate it yet due to transaction or lock ordering. For example,
ocfs2_create_inode_in_orphan() currently gives a junk blkno for preparation
of the orphan dir because it can't yet know where the actual inode is placed
- that code is actually in ocfs2_mknod_locked. This is a problem when the
orphan dirs are indexed as the junk inode number will create an index entry
which goes unused (and fails the later removal from the orphan dir).  Now
with these interfaces, ocfs2_create_inode_in_orphan() can run the block
group search (and get back the inode block number) *before* any actual
allocation occurs.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: use ocfs2_alloc_dinode_update_counts() instead of open coding
Mark Fasheh [Fri, 13 Aug 2010 22:15:16 +0000]
ocfs2: use ocfs2_alloc_dinode_update_counts() instead of open coding

ocfs2_search_chain() makes the same updates as
ocfs2_alloc_dinode_update_counts to the alloc inode. Instead of open coding
the bitmap update, use our helper function.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: split out inode alloc code from ocfs2_mknod_locked
Mark Fasheh [Fri, 13 Aug 2010 22:15:15 +0000]
ocfs2: split out inode alloc code from ocfs2_mknod_locked

Do this by splitting the bulk of the function away from the inode allocation
code at the very tom of ocfs2_mknod_locked(). Existing callers don't need to
change and won't see any difference. The new function created,
__ocfs2_mknod_locked() will be used shortly.

Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoOcfs2: Fix a regression bug from mainline commit(6b933c8e6f1a2f3118082c455eef25f9b1ac...
Tristan Ye [Thu, 19 Aug 2010 07:15:00 +0000]
Ocfs2: Fix a regression bug from mainline commit(6b933c8e6f1a2f3118082c455eef25f9b1ac7b45).

The patch is to fix the regression bug brought from commit 6b933c8...( 'ocfs2:
Avoid direct write if we fall back to buffered I/O'):

http://oss.oracle.com/bugzilla/show_bug.cgi?id=1285

The commit 6b933c8e6f1a2f3118082c455eef25f9b1ac7b45 changed __generic_file_aio_write
to generic_file_buffered_write, which didn't call filemap_{write,wait}_range to  flush
the pagecaches when we were falling O_DIRECT writes back to buffered ones. it did hurt
the O_DIRECT semantics somehow in extented odirect writes.

This patch tries to guarantee O_DIRECT writes of 'fall back to buffered' to be correctly
flushed.

Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: Fix deadlock when allocating page
Jan Kara [Tue, 24 Aug 2010 12:28:03 +0000]
ocfs2: Fix deadlock when allocating page

We cannot call grab_cache_page() when holding filesystem locks or with
a transaction started as grab_cache_page() calls page allocation with
GFP_KERNEL flag and thus page reclaim can recurse back into the filesystem
causing deadlocks or various assertion failures. We have to use
find_or_create_page() instead and pass it GFP_NOFS as we do with other
allocations.

Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: properly set and use inode group alloc hint
Mark Fasheh [Thu, 26 Aug 2010 20:06:50 +0000]
ocfs2: properly set and use inode group alloc hint

We were setting ac->ac_last_group in ocfs2_claim_suballoc_bits from
res->sr_bg_blkno.  Unfortunately, res->sr_bg_blkno is going to be zero under
normal (non-fragmented) circumstances. The discontig block group patches
effectively turned off that feature. Fix this by correctly calculating what
the next group hint should be.

Acked-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
Tested-by: Goldwyn Rodrigues <rgoldwyn@suse.de>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: Use the right group in nfs sync check.
Tao Ma [Thu, 2 Sep 2010 05:10:10 +0000]
ocfs2: Use the right group in nfs sync check.

We have added discontig block group now, and now an inode
can be allocated in an discontig block group. So get
it in ocfs2_get_suballoc_slot_bit.

The old ocfs2_test_suballoc_bit gets group block no
from the allocation inode which is wrong. Fix it by
passing the right group.

Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: Flush drive's caches on fdatasync
Jan Kara [Thu, 5 Aug 2010 18:32:45 +0000]
ocfs2: Flush drive's caches on fdatasync

When 'barrier' mount option is specified, we have to issue a cache flush
during fdatasync(2). We have to do this even if inode doesn't have
I_DIRTY_DATASYNC set because we still have to get written *data* to disk so
that they are not lost in case of crash.

Acked-by: Tao Ma <tao.ma@oracle.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Singed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: make __ocfs2_page_mkwrite handle file end properly.
Tao Ma [Sat, 17 Jul 2010 13:45:49 +0000]
ocfs2: make __ocfs2_page_mkwrite handle file end properly.

__ocfs2_page_mkwrite now is broken in handling file end.
1. the last page should be the page contains i_size - 1.
2. the len in the last page is also calculated wrong.
So change them accordingly.

Acked-by: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: Fix incorrect checksum validation error
Sunil Mushran [Thu, 12 Aug 2010 23:24:26 +0000]
ocfs2: Fix incorrect checksum validation error

For local mounts, ocfs2_read_locked_inode() calls ocfs2_read_blocks_sync() to
read the inode off the disk. The latter first checks to see if that block is
cached in the journal, and, if so, returns that block. That is ok.

But ocfs2_read_locked_inode() goes wrong when it tries to validate the checksum
of such blocks. Blocks that are cached in the journal may not have had their
checksum computed as yet. We should not validate the checksums of such blocks.

Fixes ossbz#1282
http://oss.oracle.com/bugzilla/show_bug.cgi?id=1282

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Cc: stable@kernel.org
Singed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoocfs2: Fix metaecc error messages
Sunil Mushran [Thu, 12 Aug 2010 23:24:25 +0000]
ocfs2: Fix metaecc error messages

Like tools, the checksum validate function now prints the values in hex.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Singed-off-by: Tao Ma <tao.ma@oracle.com>

8 years agoMerge branch 'for-2.6.36' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Wed, 8 Sep 2010 02:21:02 +0000]
Merge branch 'for-2.6.36' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.36' of git://linux-nfs.org/~bfields/linux:
  nfsd4: mask out non-access bits in nfs4_access_to_omode

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Wed, 8 Sep 2010 02:13:06 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/security-testing-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6:
  ima: always maintain counters
  AppArmor: Fix locking from removal of profile namespace
  AppArmor: Fix splitting an fqname into separate namespace and profile names
  AppArmor: Fix security_task_setrlimit logic for 2.6.36 changes
  AppArmor: Drop hack to remove appended " (deleted)" string

8 years agoima: always maintain counters
Mimi Zohar [Tue, 31 Aug 2010 13:38:51 +0000]
ima: always maintain counters

commit 8262bb85da allocated the inode integrity struct (iint) before any
inodes were created. Only after IMA was initialized in late_initcall were
the counters updated. This patch updates the counters, whether or not IMA
has been initialized, to resolve 'imbalance' messages.

This patch fixes the bug as reported in bugzilla: 15673.  When the i915
is builtin, the ring_buffer is initialized before IMA, causing the
imbalance message on suspend.

Reported-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: Mimi Zohar <zohar@linux.vnet.ibm.com>
Tested-by: Thomas Meyer <thomas@m3y3r.de>
Tested-by: David Safford<safford@watson.ibm.com>
Cc: Stable Kernel <stable@kernel.org>
Signed-off-by: James Morris <jmorris@namei.org>

8 years agoAppArmor: Fix locking from removal of profile namespace
John Johansen [Sat, 28 Aug 2010 01:33:29 +0000]
AppArmor: Fix locking from removal of profile namespace

The locking for profile namespace removal is wrong, when removing a
profile namespace, it needs to be removed from its parent's list.
Lock the parent of namespace list instead of the namespace being removed.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>

8 years agoAppArmor: Fix splitting an fqname into separate namespace and profile names
John Johansen [Sat, 28 Aug 2010 01:33:28 +0000]
AppArmor: Fix splitting an fqname into separate namespace and profile names

As per Dan Carpenter <error27@gmail.com>
  If we have a ns name without a following profile then in the original
  code it did "*ns_name = &name[1];".  "name" is NULL so "*ns_name" is
  0x1.  That isn't useful and could cause an oops when this function is
  called from aa_remove_profiles().

Beyond this the assignment of the namespace name was wrong in the case
where the profile name was provided as it was being set to &name[1]
after name  = skip_spaces(split + 1);

Move the ns_name assignment before updating name for the split and
also add skip_spaces, making the interface more robust.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>

8 years agoAppArmor: Fix security_task_setrlimit logic for 2.6.36 changes
John Johansen [Mon, 6 Sep 2010 17:10:20 +0000]
AppArmor: Fix security_task_setrlimit logic for 2.6.36 changes

2.6.36 introduced the abilitiy to specify the task that is having its
rlimits set.  Update mediation to ensure that confined tasks can only
set their own group_leader as expected by current policy.

Add TODO note about extending policy to support setting other tasks
rlimits.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>

8 years agoAppArmor: Drop hack to remove appended " (deleted)" string
John Johansen [Sat, 28 Aug 2010 01:33:26 +0000]
AppArmor: Drop hack to remove appended " (deleted)" string

The 2.6.36 kernel has refactored __d_path() so that it no longer appends
" (deleted)" to unlinked paths.  So drop the hack that was used to detect
and remove the appended string.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: James Morris <jmorris@namei.org>

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes...
Linus Torvalds [Tue, 7 Sep 2010 23:00:17 +0000]
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: bus speed strings should be const
  PCI hotplug: Fix build with CONFIG_ACPI unset
  PCI: PCIe: Remove the port driver module exit routine
  PCI: PCIe: Move PCIe PME code to the pcie directory
  PCI: PCIe: Disable PCIe port services during port initialization
  PCI: PCIe: Ask BIOS for control of all native services at once
  ACPI/PCI: Negotiate _OSC control bits before requesting them
  ACPI/PCI: Do not preserve _OSC control bits returned by a query
  ACPI/PCI: Make acpi_pci_query_osc() return control bits
  ACPI/PCI: Reorder checks in acpi_pci_osc_control_set()
  PCI: PCIe: Introduce commad line switch for disabling port services
  PCI: PCIe AER: Introduce pci_aer_available()
  x86/PCI: only define pci_domain_nr if PCI and PCI_DOMAINS are set
  PCI: provide stub pci_domain_nr function for !CONFIG_PCI configs

8 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Tue, 7 Sep 2010 22:44:28 +0000]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: Make fiemap work with sparse files
  xfs: prevent 32bit overflow in space reservation
  xfs: Disallow 32bit project quota id
  xfs: improve buffer cache hash scalability

8 years agoMerge branch 'for-linus' of git://android.kernel.org/kernel/tegra
Linus Torvalds [Tue, 7 Sep 2010 21:48:44 +0000]
Merge branch 'for-linus' of git://android./kernel/tegra

* 'for-linus' of git://android.kernel.org/kernel/tegra:
  [ARM] tegra: Add ZRELADDR default for ARCH_TEGRA

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88...
Linus Torvalds [Tue, 7 Sep 2010 21:38:54 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/mattst88/alpha-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88/alpha-2.6:
  alpha: Fix printk format errors
  alpha: convert perf_event to use local_t
  Fix call to replaced SuperIO functions
  alpha: remove homegrown L1_CACHE_ALIGN macro

8 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh...
Linus Torvalds [Tue, 7 Sep 2010 21:38:21 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/ericvh/v9fs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ericvh/v9fs:
  9p: potential ERR_PTR() dereference

8 years agoMerge branch 'for-linus' of git://neil.brown.name/md
Linus Torvalds [Tue, 7 Sep 2010 21:37:34 +0000]
Merge branch 'for-linus' of git://neil.brown.name/md

* 'for-linus' of git://neil.brown.name/md:
  md: resolve confusion of MD_CHANGE_CLEAN
  md: don't clear MD_CHANGE_CLEAN in md_update_sb() for external arrays
  Move .gitignore from drivers/md to lib/raid6

8 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Tue, 7 Sep 2010 21:35:16 +0000]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  Revert "hwrng: n2-drv - remove casts from void*"
  crypto: testmgr - Default to no tests
  crypto: testmgr - Fix test disabling option
  crypto: hash - Fix handling of small unaligned buffers