Merge remote-tracking branch 'ibft/master'

Merge remote-tracking branch 'swiotlb/master'

Merge remote-tracking branch 'libata/NEXT'

Merge remote-tracking branch 'kconfig/for-next'

Merge remote-tracking branch 'kbuild/for-next'

Merge remote-tracking branch 'docs/docs-move'

Merge branch 'quilt/kernel-doc'

Merge remote-tracking branch 'hwmon-staging/hwmon-next'

Merge branch 'quilt/jdelvare-hwmon'

Merge remote-tracking branch 'bjdooks-i2c/next-i2c'

Merge branch 'quilt/i2c'

Merge remote-tracking branch 'hid/for-next'

Conflicts:
drivers/hid/hid-core.c
drivers/hid/hid-ids.h

Merge remote-tracking branch 'xfs/master'

Conflicts:
fs/xfs/xfs_super.c

Merge remote-tracking branch 'v9fs/for-next'

Conflicts:
net/9p/trans_virtio.c

Merge remote-tracking branch 'ocfs2/linux-next'

Merge remote-tracking branch 'nfsd/nfsd-next'

Merge remote-tracking branch 'nfs/linux-next'

Merge remote-tracking branch 'logfs/master'

Merge remote-tracking branch 'gfs2/master'

Merge remote-tracking branch 'ext4/dev'

Merge remote-tracking branch 'cifs/master'

Merge remote-tracking branch 'ceph/for-next'

Merge remote-tracking branch 'xtensa/master'

Merge remote-tracking branch 'unicore32/unicore32'

Merge remote-tracking branch 'tile/master'

Merge remote-tracking branch 'rmobile/rmobile-latest'

Merge remote-tracking branch 'sh/sh-latest'

Merge remote-tracking branch 's390/features'

Merge remote-tracking branch '52xx-and-virtex/powerpc/next'

Merge remote-tracking branch 'powerpc/next'

Merge remote-tracking branch 'mips/mips-for-linux-next'

Merge remote-tracking branch 'm68knommu/for-next'

Merge remote-tracking branch 'm68k/for-next'

Merge remote-tracking branch 'ia64/test'

Merge branch 'quilt/hexagon'

Merge remote-tracking branch 'cris/for-next'

Merge remote-tracking branch 'blackfin/for-linus'

Merge remote-tracking branch 'ux500-core/ux500-core'

Merge remote-tracking branch 's5p/for-next'

Conflicts:
arch/arm/mach-exynos4/mach-smdkc210.c

Merge remote-tracking branch 'i.MX/for-next'

Merge remote-tracking branch 'at91/at91-next'

Merge remote-tracking branch 'arm-soc/for-next'

Conflicts:
arch/arm/include/asm/hardware/cache-l2x0.h
arch/arm/mach-at91/board-usb-a9260.c
arch/arm/mach-msm/board-msm8x60.c
arch/arm/mach-mxs/include/mach/gpio.h
arch/arm/mach-u300/Makefile.boot
arch/arm/plat-mxc/include/mach/gpio.h

Merge remote-tracking branch 'arm-lpae/for-next'

Conflicts:
arch/arm/include/asm/page.h
arch/arm/include/asm/pgalloc.h
arch/arm/include/asm/pgtable-hwdef.h
arch/arm/include/asm/pgtable.h
arch/arm/include/asm/tlb.h
arch/arm/kernel/head.S
arch/arm/kernel/sleep.S
arch/arm/mm/dma-mapping.c
arch/arm/mm/mmu.c

Merge remote-tracking branch 'arm/for-next'

Merge remote-tracking branch 'input-current/for-linus'

Merge remote-tracking branch 'wireless-current/master'

Merge remote-tracking branch 'pci-current/for-linus'

Merge remote-tracking branch 'net-current/master'

Merge remote-tracking branch 'scsi-rc-fixes/master'

Add extra arch overrides to asm-generic/checksum.h

There are plausible reasons for architectures to provide their own
versions of csum_partial_copy_nocheck and csum_tcpudp_magic.
By protecting these, the architecture can still re-use the
asm-generic checksum.h, instead of copying it.

Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add self to MAINTAINERS

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>

Hexagon: Add basic stacktrace functionality for Hexagon architecture.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add configuration and makefiles for the Hexagon architecture.

Removed CONFIG_ARCH_HAS_KMAP

Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Comet platform support

Used the shortened version of the config; other platform stuff moved out until
it's ready.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: kgdb support files

Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add page-fault support.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add page table header files & etc.

Removed highmem more highmem stuff.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>

Hexagon: Add ioremap support

Removed redundant flush_cache_vmap.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>

Hexagon: Provide DMA implementation

Basic DMA implementation for our architecture.

Added sync_single_for_* routines; hexagon_map_page() still seems to be a
convenient place to flush/invalidate.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Implement basic TLB management routines for Hexagon.

Mostly all stubs, as the TLB is managed by the hypervisor.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Implement basic cache-flush support

We have separate I/D caches.  Data caches are physically indexed.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Provide basic implementation and/or stubs for I/O routines.

Filled in the ins/outs; leaving them in for now.

Change _IO_BASE and IO_SPACE_LIMIT for safety.

I have compiled this with sparse and it builds; I've also picked over
various files related to the HAS_IOPORT_MAP from arm-soc.git and
tried them out, and they work, so I guess we'll just wait until
they're merged.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>

Hexagon: Add user access functions

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add locking types and functions

Moved rwsem to asm-generic.  Will submit a subsequent patch for PPC
use it from the new location.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>

Hexagon: Add SMP support

Changed smp.c to use generic IPI handler functions.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>

Hexagon: Provide basic debugging and system trap support.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add ptrace support

Will look at the more generic ptrace support in the future.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>

Hexagon: Add time and timer functions

Cleaned up iomem and register declarations, unnecessary
SMP ifdefs, cpu_relax() in udelay.

Platform stuff is in development; platform defines will be cleaned up.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>

Hexagon: Add interrupts

Removed init_IRQ wrapper and fixed up some comments.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add startup code

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add init_task and process functions

Fixed potential race-to-sleep condition in cpu_idle() and calling schedule
with irq's disabled.  I still have to have that extra enable/disbale in the
while() loop else the interrupt never gets serviced, and thus need_resched()
will always be false.

If called with interrupts disabled, our __vmwait() actually just returns
without servicing it.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add signal functions

Added vDSO support; signal trampoline is no longer on the stack
so it no longer needs to be executable.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>

Hexagon: Support dynamic module loading.

Modules should be compiled as ordinary .o's; shared objects are not
supported.

Deleted module functions that already have generic equivalents.

Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Export ksyms defined in assembly files.

Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add hypervisor interface

The port of Linux for Hexagon runs on top of a hypervisor layer,
which presents an interface for memory management, events, hardware
control, etc.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add memcpy and memset accelerated functions

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add checksum functions

Removed most non-assembly functions and use the generic ones instead.

Made a few more minor changes to use the generic header file.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add delay functions

Added read_current_timer elsewhere.

udelay is based off of __vmgettime() which returns processor ticks.

Removed __delay and CONFIG_GENERIC_CALIBRATE_DELAY.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>

Hexagon: Add threadinfo

Removed some cruft and changed to allow kernel stack to be smaller than a page
size.

I left the thread_info alignment alone for the time being; may revisit that in
the future.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add processor and system headers

Adds Hexagon information for processor/system type stuff
like context switching, register access, and irqflags.

Changed cpu_relax() to call __vmyield().

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add syscalls

Removed sys_mmap* and replaced with sys_mmap_pgoff()

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>

Hexagon: Add atomic ops support

Our architecture uses load locked/store conditional type semantics for atomic
ops.

Added inline assembly version of atomic_add_unless.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add bitops support

Merged sub-files into bitops.h; fixed return types and volatiles.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Core arch-specific header files

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Signed-off-by: Linas Vepstas <linas@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

Hexagon: Add generic headers

Reworked to use the Kbuild generic headers mechanism.

Signed-off-by: Richard Kuo <rkuo@codeaurora.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>

lib/sha1.c: quiet sparse noise about symbol not declared

Include <linux/cryptohash.h> to pickup the declarations for sha_transform
and sha_init to quite the sparse noise:

  warning: symbol 'sha_transform' was not declared. Should it be static?
  warning: symbol 'sha_init' was not declared. Should it be static?

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: Mandeep Singh Baines <msb@chromium.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

pci: Don't crash when reading mpss from root complex

In pcie_find_smpss(), we have the following statement:

  if (dev->is_hotplug_bridge && (!list_is_singular(&dev->bus->devices) ||
    dev->bus->self->pcie_type != PCI_EXP_TYPE_ROOT_PORT))

The problem is that at least on my machine, this gets called for the
root complex (virtual P2P bridge), and dev->bus->self is NULL since
the parent bus for this is not itself anchor to a PCI device.

This adds the necessary NULL check.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Jon Mason <mason@myri.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

nfs: Do not allow multiple mounts on same mountpoint when using -o noac

Do not allow multiple mounts on same mountpoint when using -o noac

When you normally attempt to mount a share twice on the same mountpoint,
a check in do_add_mount causes it to return an error

# mount localhost:/nfsv3 /mnt
# mount localhost:/nfsv3 /mnt
mount.nfs: /mnt is already mounted or busy

However when using the option 'noac', the user is able to mount the same
share on the same mountpoint multiple times. This happens because a
share mounted with the noac option is automatically assigned the 'sync'
flag MS_SYNCHRONOUS in nfs_initialise_sb(). This flag is set after the
check for already existing superblocks is done in sget(). The check for
the mount flags in nfs_compare_mount_options() does not take into
account the 'sync' flag applied later on in the code path. This means
that when using 'noac', a new superblock structure is assigned for every
new mount of the same share and multiple shares on the same mountpoint
are allowed.

ie.
# mount -onoac localhost:/nfsv3 /mnt
can be run multiple times.

The patch checks for noac and assigns the sync flag before sget() is
called to obtain an already existing superblock structure.

Signed-off-by: Sachin Prabhu <sprabhu@redhat.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

NFS: Fix a typo in nfs_flush_multi

Fix a typo which causes an Oops in the RPC layer, when using wsize < 4k.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Tested-by: Sricharan R <r.sricharan@ti.com>

libata-eh: ata_eh_followup_srst_needed() does not need 'classes' parameter

... since it does not use it.

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

sata_sil24: Use const

Reduce data by using const.

$ size drivers/ata/sata_sil24.o*
   text    data     bss     dec     hex filename
  12764     614    2688   16066    3ec2 drivers/ata/sata_sil24.o.new
  12320    1058    2688   16066    3ec2 drivers/ata/sata_sil24.o.old

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>

hwmon/f71882fg: Make the decision wether to register fan attr. per fan

Before this patch the f71882fg driver completely fails to initialize
on systems which have reserved settings in the pwm enable register, and
it disables all auto pwm sysfs attributes if any fan is controlled by
a digital sensor reading.

This patch changes the fail to initialize into don't register any attributes
for the fan for which there are reserved settings in the pwm enable register
and also makes the not registering of auto pwm sysfs attributes a per fan
thing.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

hwmon/f71882fg: Add a f71882fg_create_fan_sysfs_files helper function

This is a preparation patch for not registering fan/pwm attributes for
some fans (rather then register them for all or for none).

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

hwmon/f71882fg: Make all fan/pwm attr tables 2 dimensional

This is a preparation patch for not registering fan/pwm attributes for
some fans (rather then register them for all or for none).

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>

hwmon: (pmbus) Add support for Intersil power management chips

Add support for Intersil / Zilker Labs ZL2004, ZL2006, ZL2008, ZL2105, ZL2106,
ZL6100, and ZL6105.

Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>
Reviewed-by: Robert Coulson <robert.coulson@ericsson.com>

hwmon: (pmbus) Always call _pmbus_read_byte in core driver

Always call _pmbus_read_byte() instead of pmbus_read_byte() in PMBus core
driver. With this change, device specific read functions can be implemented for
all registers.

Since the device specific read_byte function is now always called, we need to be
more careful with page validations. Only fail if the passed page number is larger
than 0, since -1 means "current page".

Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>
Reviewed-by: Robert Coulson <robert.coulson@ericsson.com>

iwlagn: fix command queue timeout

If the command queue is constantly busy,
which can happen in P2P, the hangcheck
timer will frequently find a command in
it and will eventually reset the device
because nothing sets the timestamp for
this queue when commands are processed.

Fix this by setting the timestamp when
a command completes.

Cc: stable@kernel.org #2.6.39, #3.0.0 #3.1.0
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
SIgned-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: fix stack corruption

Alexander reported a strange crash in iwlagn that
Meenakshi and Wey couldn't reproduce. I just ran
into the same issue and tracked it down to stack
corruption. This fixes it.

The problem was introduced in
commit 4b8b99b6e650d0527f3a123744b7459976581d14
Author: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Date:   Fri Jul 8 14:29:48 2011 -0700

    iwlagn: radio sensor offset in le16 format

Cc: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Cc: Meenakshi Venkataraman <meenakshi.venkataraman@intel.com>
Reported-by: Alexander Diewald <alex@diewald.cc>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Fix kernel panic on unplugging the device

when the device is yanked out ath_pci_remove starts doing the cleanups,
unregistering the hardware etc. so we should bail out immediately when
we get drv_flush callback from mac80211 when the card is being unplugged.
the panic occurs after we had associated to an AP.

EIP: 0060:[<fb315b00>] EFLAGS: 00010246 CPU: 0
EIP is at ath_reset+0xa0/0x1c0 [ath9k]
EAX: 00000000 EBX: 000697c0 ECX: 00000002 EDX: f3c3ccf0
ESI: 00000000 EDI: 00000000 EBP: f43e7b78 ESP: f43e7b50
  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
  Process kworker/u:2 (pid: 182, ti=f43e6000 task=f3c3c7c0
task.ti=f43e6000)
  Stack:
  0000002a 00000000 00000000 003e7b78 0000000f eaaa8500
ffffffea eaaa97c0
  eaaaa000 00000001 f43e7ba8 fb315d23 f99e7721 ecece680
eaaac738 eaaa8500
  eaaaa020 000000c8 000000c8 00000000 eaaa8d58 eaaa8500
f43e7bd0 fb080b29
  Call Trace:
[<fb315d23>] ath9k_flush+0x103/0x170 [ath9k]
[<fb080b29>] __ieee80211_recalc_idle+0x2c9/0x400
[mac80211]
[<fb080c8e>] ieee80211_recalc_idle+0x2e/0x60 [mac80211]
[<fb07aa73>] ieee80211_mgd_deauth+0x173/0x210 [mac80211]
[<fb084559>] ieee80211_deauth+0x19/0x20 [mac80211]
[<f99dda53>] __cfg80211_mlme_deauth+0xf3/0x140
[cfg80211]
[<c0633d00>] ? __mutex_lock_common+0x1f0/0x380
[<f99e1b5d>] __cfg80211_disconnect+0x18d/0x1f0
[cfg80211]
[<f99c8199>] cfg80211_netdev_notifier_call+0x159/0x5c0
[cfg80211]
[<c0608a64>] ? packet_notifier+0x174/0x1f0
[<c0639202>] notifier_call_chain+0x82/0xb0
[<c0170d8f>] raw_notifier_call_chain+0x1f/0x30
[<c053b86c>] call_netdevice_notifiers+0x2c/0x60
[<c0182184>] ? trace_hardirqs_on_caller+0xf4/0x180
[<c053b8ec>] __dev_close_many+0x4c/0xd0
[<c053ba2d>] dev_close_many+0x6d/0xc0
[<c053bb53>] rollback_registered_many+0x93/0x1c0
[<c018221b>] ? trace_hardirqs_on+0xb/0x10
[<c053bc95>] unregister_netdevice_many+0x15/0x50
[<fb07f83b>] ieee80211_remove_interfaces+0x7b/0xb0
[mac80211]
[<fb06a14b>] ieee80211_unregister_hw+0x4b/0x110
[mac80211]
[<fb311a4a>] ath9k_deinit_device+0x3a/0x60 [ath9k]
[<fb31eed6>] ath_pci_remove+0x46/0x90 [ath9k]
[<c03b4ac4>] pci_device_remove+0x44/0x100
[<c043eb54>] __device_release_driver+0x64/0xb0
[<c043ec67>] device_release_driver+0x27/0x40
[<c043deeb>] bus_remove_device+0x7b/0xa0
[<c043c491>] device_del+0xf1/0x180
[<c043c530>] device_unregister+0x10/0x20
[<c03afafe>] pci_stop_bus_device+0x6e/0x80
[<c03afb72>] pci_remove_bus_device+0x12/0xa0
[<c03c2f29>] pciehp_unconfigure_device+0x89/0x180
[<c0181e54>] ? mark_held_locks+0x64/0x100
[<c063390f>] ? __mutex_unlock_slowpath+0xaf/0x140
[<c03c1f84>] pciehp_disable_slot+0x64/0x1b0
[<c03c2850>] pciehp_power_thread+0xd0/0x100
[<c0164ad0>] ? process_one_work+0x100/0x4d0
[<c0164b4c>] process_one_work+0x17c/0x4d0
[<c0164ad0>] ? process_one_work+0x100/0x4d0
[<c03c2780>] ? queue_interrupt_event+0xa0/0xa0
[<c01662bb>] worker_thread+0x13b/0x320
[<c018221b>] ? trace_hardirqs_on+0xb/0x10
[<c0166180>] ? manage_workers+0x1e0/0x1e0
[<c016a654>] kthread+0x84/0x90
[<c016a5d0>] ? __init_kthread_worker+0x60/0x60
[<c063d106>] kernel_thread_helper+0x6/0x10

Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: fix missing sta_lock in __sta_info_destroy

Since my commit 34e895075e21be3e21e71d6317440d1ee7969ad0
("mac80211: allow station add/remove to sleep") there is
a race in mac80211 when it clears the TIM bit because a
sleeping station disconnected, the spinlock isn't held
around the relevant code any more. Use the right API to
acquire the spinlock correctly.

Cc: stable@kernel.org [2.6.34+]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>