Currently, I_SYNC can never be set when evict_inode() (and thus
end_writeback()) is called because flusher thread holds inode reference while
inode is under writeback. As a result inode_sync_wait() in those places
currently does nothing. However that is going to change and unveils problems
with calling inode_sync_wait() from end_writeback(). Several filesystems call
end_writeback() after they have deleted the inode (btrfs, gfs2, ...) and other
filesystems (ext3, ext4, reiserfs, ...) can deadlock when waiting for I_SYNC
because they call end_writeback() from within a transaction.

To avoid these issues, we move inode_sync_wait() into evict_inode() before
calling ->evict_inode(). That way we preserve the current property that
->evict_inode() and writeback never run in parallel and all filesystems are
safe.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

The code in writeback_single_inode() is relatively complex. The list requeing
logic makes sense only for flusher thread but not really for sync_inode() or
write_inode_now() callers. Also when we want to get rid of inode references
held by flusher thread, we will need a special I_SYNC handling there.

So separate part of writeback_single_inode() which does the real writeback work
into __writeback_single_inode() and make writeback_single_inode() do only stuff
necessary for callers writing only one inode, moving the special list handling
into writeback_sb_inodes(). As a sideeffect this fixes a possible race where we
could skip some inode during sync(2) because other writer refiled it from b_io
to b_dirty list. Also I_SYNC handling is moved into the callers of
__writeback_single_inode() to make locking easier.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

writeback_single_inode() doesn't need wb->list_lock for anything on entry now.
So remove the requirement. This makes locking of writeback_single_inode()
temporarily awkward (entering with i_lock, returning with i_lock and
wb->list_lock) but it will be sanitized in the next patch.

Also inode_wait_for_writeback() doesn't need wb->list_lock for anything. It was
just taking it to make usage convenient for callers but with
writeback_single_inode() changing it's not very convenient anymore. So remove
the lock from that function.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

Move inode requeueing after inode has been written out into a separate
function.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

Instead of clearing I_DIRTY_PAGES and resetting it when we didn't succeed in
writing them all, just clear the bit only when we succeeded writing all the
pages. We also move the clearing of the bit close to other i_state handling to
separate it from writeback list handling. This is desirable because list
handling will differ for flusher thread and other writeback_single_inode()
callers in future. No filesystem plays any tricks with I_DIRTY_PAGES (like
checking it in ->writepages or ->write_inode implementation) so this movement
is safe.
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

When writeback_single_inode() is called on inode which has I_SYNC already
set while doing WB_SYNC_NONE, inode is moved to b_more_io list. However
this makes sense only if the caller is flusher thread. For other callers of
writeback_single_inode() it doesn't really make sense and may be even wrong
- flusher thread may be doing WB_SYNC_ALL writeback in parallel.

So we move requeueing from writeback_single_inode() to writeback_sb_inodes().
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

Move clearing of I_SYNC into inode_sync_complete().  It is more logical to have
clearing of I_SYNC bit and waking of waiters in one place. Also later we will
have two places needing to clear I_SYNC and wake up waiters so this allows them
to use the common helper. Moving of I_SYNC clearing to a later stage of
writeback_single_inode() is safe since we hold i_lock all the time.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

This prevents global_dirty_limit from remaining 0 (the initial value)
for long time, since it's only updated in update_dirty_limit() when
above the dirty freerun area.

It will avoid unexpected consequences when some random code use it as a
convenient approximation of the global dirty threshold.
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

Reorder structure writeback_control to remove 8 bytes of padding on 64
bit builds, this shrinks its size from 48 to 40 bytes.

This structure is always on the stack and uses C99 named initialisation,
so should be safe and have a small impact on stack usage.
Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

The function global_dirtyable_memory is only referenced in this file and
should be marked static to prevent it from being exposed globally.

This quiets the sparse warning:

warning: symbol 'global_dirtyable_memory' was not declared. Should it be static?
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>

Pull system.h fixups for less common arch's from Paul Gortmaker:
 "Here is what is hopefully the last of the system.h related fixups.

  The fixes for Alpha and ia64 are code relocations consistent with what
  was done for the more mainstream architectures.  Note that the
  diffstat lines removed vs lines added are not the same since I've
  fixed some of the whitespace issues in the relocated code blocks.
  However they are functionally the same.  Compile tested locally, plus
  these two have been in linux-next for a while.

  There is also a trivial one line system.h related fix for the Tilera
  arch from Chris Metcalf to fix an implict include.."

* 'systemh-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux:
  irq_work: fix compile failure on tile from missing include
  ia64: populate the cmpxchg header with appropriate code
  alpha: fix build failures from system.h dismemberment

Pull fbdev fixes from Florian Tobias Schandinat:
 - a compile fix for au1*fb
 - a fix to make kyrofb usable on x86_64
 - a fix for uvesafb to prevent an oops due to NX-protection

 "The fix for kyrofb is a bit large but it's just replacing "unsigned
  long" by "u32" for 64 bit compatibility."

* tag 'fbdev-fixes-for-3.4-1' of git://github.com/schandinat/linux-2.6:
  video:uvesafb: Fix oops that uvesafb try to execute NX-protected page
  fbdev: fix au1*fb builds
  kyrofb: fix on x86_64

Pull the minimal btrfs branch from Chris Mason:
 "We have a use-after-free in there, along with errors when mount -o
  discard is enabled, and a BUG_ON(we should compile with UP more
  often)."

* 'for-linus-min' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: use commit root when loading free space cache
  Btrfs: fix use-after-free in __btrfs_end_transaction
  Btrfs: check return value of bio_alloc() properly
  Btrfs: remove lock assert from get_restripe_target()
  Btrfs: fix eof while discarding extents
  Btrfs: fix uninit variable in repair_eb_io_failure
  Revert "Btrfs: increase the global block reserve estimates"

Pull block driver bits from Jens Axboe:

 - A series of fixes for mtip32xx.  Most from Asai at Micron, but also
   one from Greg, getting rid of the dependency on PCIE_HOTPLUG.

 - A few bug fixes for xen-blkfront, and blkback.

 - A virtio-blk fix for Vivek, making resize actually work.

 - Two fixes from Stephen, making larger transfers possible on cciss.
   This is needed for tape drive support.

* 'for-3.4/drivers' of git://git.kernel.dk/linux-block:
  block: mtip32xx: remove HOTPLUG_PCI_PCIE dependancy
  mtip32xx: dump tagmap on failure
  mtip32xx: fix handling of commands in various scenarios
  mtip32xx: Shorten macro names
  mtip32xx: misc changes
  mtip32xx: Add new sysfs entry 'status'
  mtip32xx: make setting comp_time as common
  mtip32xx: Add new bitwise flag 'dd_flag'
  mtip32xx: fix error handling in mtip_init()
  virtio-blk: Call revalidate_disk() upon online disk resize
  xen/blkback: Make optional features be really optional.
  xen/blkback: Squash the discard support for 'file' and 'phy' type.
  mtip32xx: fix incorrect value set for drv_cleanup_done, and re-initialize and start port in mtip_restart_port()
  cciss: Fix scsi tape io with more than 255 scatter gather elements
  cciss: Initialize scsi host max_sectors for tape drive support
  xen-blkfront: make blkif_io_lock spinlock per-device
  xen/blkfront: don't put bdev right after getting it
  xen-blkfront: use bitmap_set() and bitmap_clear()
  xen/blkback: Enable blkback on HVM guests
  xen/blkback: use grant-table.c hypercall wrappers

Pull block core bits from Jens Axboe:
 "It's a nice and quiet round this time, since most of the tricky stuff
  has been pushed to 3.5 to give it more time to mature.  After a few
  hectic block IO core changes for 3.3 and 3.2, I'm quite happy with a
  slow round.

  Really minor stuff in here, the only real functional change is making
  the auto-unplug threshold a per-queue entity.  The threshold is set so
  that it's low enough that we don't hold off IO for too long, but still
  big enough to get a nice benefit from the batched insert (and hence
  queue lock cost reduction).  For raid configurations, this currently
  breaks down."

* 'for-3.4/core' of git://git.kernel.dk/linux-block:
  block: make auto block plug flush threshold per-disk based
  Documentation: Add sysfs ABI change for cfq's target latency.
  block: Make cfq_target_latency tunable through sysfs.
  block: use lockdep_assert_held for queue locking
  block: blk_alloc_queue_node(): use caller's GFP flags instead of GFP_KERNEL

The OMAP driver needs a 'depends on ARCH_OMAP2PLUS' since it only
builds for OMAP2+ platforms.

This 'depends on' was in the original patch from Russell King, but was
erroneously removed by me when making this option user-selectable in
commit b09db45c (cpufreq: OMAP driver depends CPUfreq tables.)  This
patch remedies that.

Apologies to Russell King for breaking his originally working patch.

Also, thanks to Grazvydas Ignotas for reporting the same problem.

Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Grazvydas Ignotas <notasas@gmail.com>
Signed-off-by: Kevin Hilman <khilman@ti.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull sparc fixes from David Miller.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc64: Eliminate obsolete __handle_softirq() function
  sparc64: Fix bootup crash on sun4v.

Pull hwmon patches from Guenter Roeck:
 "Fix build warnings in four drivers"

* tag 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (pmbus_core) Fix compiler warning
  hwmon: (smsc47m1) Fix compiler warning
  hwmon: (acpi_power_meter) Fix compiler warning seen in some configurations
  hwmon: (smsc47b397) Fix compiler warning

Pull GPIO bug fixes from Grant Likely:
 "Miscellaneous bug fixes to GPIO drivers and for a corner case in the
  gpio device tree parsing code."

* tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  gpio/exynos: Fix compiler warning in gpio-samsung.c file
  gpio: Fix range check in of_gpio_simple_xlate()
  gpio: Fix uninitialized variable bit in adp5588_irq_handler
  gpio/sodaville: Convert sodaville driver to new irqdomain API

Pull SPI bug fixes from Grant Likely:
 "Miscellaneous driver bug fixes.  No major changes in this branch."

* tag 'spi-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  spi/imx: prevent NULL pointer dereference in spi_imx_probe()
  spi/imx: mark base member in spi_imx_data as __iomem
  spi/mpc83xx: fix NULL pdata dereference bug
  spi/davinci: Fix DMA API usage in davinci
  spi/pL022: include types.h to remove compilation warnings

The invocation of softirq is now handled by irq_exit(), so there is no
need for sparc64 to invoke it on the trap-return path.  In fact, doing so
is a bug because if the trap occurred in the idle loop, this invocation
can result in lockdep-RCU failures.  The problem is that RCU ignores idle
CPUs, and the sparc64 trap-return path to the softirq handlers fails to
tell RCU that the CPU must be considered non-idle while those handlers
are executing.  This means that RCU is ignoring any RCU read-side critical
sections in those handlers, which in turn means that RCU-protected data
can be yanked out from under those read-side critical sections.

The shiny new lockdep-RCU ability to detect RCU read-side critical sections
that RCU is ignoring located this problem.

The fix is straightforward: Make sparc64 stop manually invoking the
softirq handlers.
Reported-by: Meelis Roos <mroos@linux.ee>
Suggested-by: David Miller <davem@davemloft.net>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Meelis Roos <mroos@linux.ee>
Cc: stable@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

The DS driver registers as a subsys_initcall() but this can be too
early, in particular this risks registering before we've had a chance
to allocate and setup module_kset in kernel/params.c which is
performed also as a subsyts_initcall().

Register DS using device_initcall() insteal.
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: stable@vger.kernel.org

Building with IRQ_WORK configured results in

kernel/irq_work.c: In function ‘irq_work_run’:
kernel/irq_work.c:110: error: implicit declaration of function ‘irqs_disabled’

The appropriate header just needs to be included.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

commit 93f37888

    "Fix ia64 build errors (fallout from system.h disintegration)"

introduced arch/ia64/include/asm/cmpxchg.h as a temporary
build fix and stated:

    "... leave the migration of xchg() and cmpxchg() to this new
     header file for a future patch."

Migrate the appropriate chunks from asm/intrinsics.h and fix
the whitespace issues in the migrated chunk.

Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: David Howells <dhowells@redhat.com>
Acked-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

commit ec221208

    "Disintegrate asm/system.h for Alpha"

combined with commit b4816afa

    "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h"

introduced the concept of asm/cmpxchg.h but the alpha arch
never got one.  Fork the cmpxchg content out of the asm/atomic.h
file to create one.

Some minor whitespace fixups were done on the block of code that
created the new file.

Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Acked-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

Pull infiniband fix from Roland Dreier:
 "Add a fix for a bug hit by Alexey Shvetsov in ib_srtp that hits on
  non-mlx4 hardware."

* tag 'srpt-srq-type' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/srpt: Set srq_type to IB_SRQT_BASIC

We've now fixed IS_ENABLED() and friends to not require any special
"__enabled_" prefixed versions of the normal Kconfig options, so delete
the last traces of them being generated.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

This reverts commit 953742c8.

Dumping two lines into autoconf.h for all existing Kconfig options
results in a giant file (~16k lines) we have to process each time we
compile something.  We've weaned IS_ENABLED() and similar off of
requiring the __enabled_ definitions so now we can revert the change
which caused all the extra lines.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Using IS_ENABLED() within C (vs.  within CPP #if statements) in its
current form requires us to actually define every possible bool/tristate
Kconfig option twice (__enabled_* and __enabled_*_MODULE variants).

This results in a huge autoconf.h file, on the order of 16k lines for a
x86_64 defconfig.

Fixing IS_ENABLED to be able to work on the smaller subset of just
things that we really have defined is step one to fixing this.  Which
means it has to not choke when fed non-enabled options, such as:

  include/linux/netdevice.h:964:1: warning: "__enabled_CONFIG_FCOE_MODULE" is not defined [-Wundef]

The original prototype of how to implement a C and preprocessor
compatible way of doing this came from the Google+ user "comex ." in
response to Linus' crowdsourcing challenge for a possible improvement on
his earlier C specific solution:

	#define config_enabled(x)       (__stringify(x)[0] == '1')

In this implementation, I've chosen variable names that hopefully make
how it works more understandable.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

A user reported that booting his box up with btrfs root on 3.4 was way
slower than on 3.3 because I removed the ideal caching code.  It turns out
that we don't load the free space cache if we're in a commit for deadlock
reasons, but since we're reading the cache and it hasn't changed yet we are
safe reading the inode and free space item from the commit root, so do that
and remove all of the deadlock checks so we don't unnecessarily skip loading
the free space cache.  The user reported this fixed the slowness.  Thanks,
Tested-by: Calvin Walton <calvin.walton@kepstin.ca>
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

Pull USB fixes from Greg KH:
 "Here are a number of fixes for the USB core and drivers for 3.4-rc2

  Lots of tiny xhci fixes here, a few usb-serial driver fixes and new
  device ids, and a smattering of other minor fixes in different USB
  drivers.

  Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"

* tag 'usb-3.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (30 commits)
  USB: update usbtmc api documentation
  xHCI: Correct the #define XHCI_LEGACY_DISABLE_SMI
  xHCI: use gfp flags from caller instead of GFP_ATOMIC
  xHCI: add XHCI_RESET_ON_RESUME quirk for VIA xHCI host
  USB: fix bug of device descriptor got from superspeed device
  xhci: Fix register save/restore order.
  xhci: Restore event ring dequeue pointer on resume.
  xhci: Don't write zeroed pointers to xHC registers.
  xhci: Warn when hosts don't halt.
  xhci: don't re-enable IE constantly
  usb: xhci: fix section mismatch in linux-next
  xHCI: correct to print the true HSEE of USBCMD
  USB: serial: fix race between probe and open
  UHCI: hub_status_data should indicate if ports are resuming
  EHCI: keep track of ports being resumed and indicate in hub_status_data
  USB: fix race between root-hub suspend and remote wakeup
  USB: sierra: add support for Sierra Wireless MC7710
  USB: ftdi_sio: fix race condition in TIOCMIWAIT, and abort of TIOCMIWAIT when the device is removed
  USB: ftdi_sio: fix status line change handling for TIOCMIWAIT and TIOCGICOUNT
  USB: don't ignore suspend errors for root hubs
  ...

Pull tty and serial fixes from Greg KH:
 "Here are some tty and serial fixes for 3.4-rc2.

  Most important here is the pl011 fix, which has been reported by about
  100 different people, which means more people use it than I expected
  :)

  There are also some 8250 driver reverts due to some problems reported
  by them.  And other minor fixes as well.

  Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"

* tag 'tty-3.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  pch_uart: Add Kontron COMe-mTT10 uart clock quirk
  pch_uart: Fix MSI setting issue
  serial/8250_pci: add a "force background timer" flag and use it for the "kt" serial port
  Revert "serial/8250_pci: setup-quirk workaround for the kt serial controller"
  Revert "serial/8250_pci: init-quirk msi support for kt serial controller"
  tty/serial/omap: console can only be built-in
  serial: samsung: fix omission initialize ulcon in reset port fn()
  printk(): add KERN_CONT where needed in hpet and vt code
  tty/serial: atmel_serial: fix RS485 half-duplex problem
  tty: serial: altera_uart: Check for NULL platform_data in probe.
  isdn/gigaset: use gig_dbg() for debugging output
  omap-serial: Fix the error handling in the omap_serial probe
  serial: PL011: move interrupt clearing

Pull staging tree fixes from Greg KH:
 "Here are a number of bugfixes for the drivers/staging/ portion of the
  kernel that have been reported recently.

  Nothing major here, with maybe the exception of the ramster code can
  now be built so it is enabled in the build again, and lots of memory
  leaks that people like to have fixed on their systems.

  Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"

* tag 'staging-3.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: android: fix mem leaks in __persistent_ram_init()
  staging: vt6656: Don't leak memory in drivers/staging/vt6656/ioctl.c::private_ioctl()
  staging: iio: hmc5843: Fix crash in probe function.
  staging/xgifb: fix display on XGI Volari Z11m cards
  Staging: android: timed_gpio: Fix resource leak in timed_gpio_probe error paths
  android: make persistent_ram based drivers depend on HAVE_MEMBLOCK
  staging: iio: ak8975: Remove i2c client data corruption
  staging: drm/omap: move where DMM driver is registered
  staging: zsmalloc: fix memory leak
  Staging: rts_pstor: off by one in for loop
  staging: ozwpan: Added new maintainer for ozwpan
  staging:rts_pstor:Avoid "Bad target number" message when probing driver
  staging:rts_pstor:Fix possible panic by NULL pointer dereference
  Staging: vt6655-6: check keysize before memcpy()
  staging/media/as102: Don't call release_firmware() on uninitialized variable
  staging:iio:core add missing increment of loop index in iio_map_array_unregister()
  staging: ramster: unbreak my heart
  staging/vme: Fix module parameters
  staging: sep: Fix sign of error

Pull driver core and kobject fixes from Greg KH:
 "Here are some minor fixes for the driver core and kobjects that people
  have reported recently.

  Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>"

* tag 'driver-core-3.4-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  kobject: provide more diagnostic info for kobject_add_internal() failures
  sysfs: handle 'parent deleted before child added'
  sysfs: Prevent crash on unset sysfs group attributes
  sysfs: Update the name hash for an entry after changing the namespace
  drivers/base: fix compiler warning in SoC export driver - idr should be ida
  drivers/base: Remove unneeded spin_lock_init call for soc_lock

Pull a fix for the recent irqdomain bug fixes from Grant Likely:
 "I flubbed one patch in the last pull request which broke a format
  string on 64 bit platforms.  Here's the fix."

* tag 'irqdomain-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  irq_domain: fix type mismatch in debugfs output format

sizeof(void*) returns an unsigned long, but it was being used as a width parameter to a "%-*s" format string which requires an int.  On 64 bit platforms this causes a type mismatch:

    linux/kernel/irq/irqdomain.c:575: warning: field width should have type
    'int', but argument 6 has type 'long unsigned int'

This change casts the size to an int so printf gets the right data type.
Reported-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: David Daney <david.daney@cavium.com>

Pull trivial perf build failure fix from Thomas Gleixner.

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf tools: Fix getrusage() related build failure on glibc trunk

Pull timer fixes from Thomas Gleixner:
 "The itimer removal one is not strictly a fix, but I really wanted to
  avoid a rebase of the urgent ones."

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  Revert "clocksource: Load the ACPI PM clocksource asynchronously"
  clockevents: tTack broadcast device mode change in tick_broadcast_switch_to_oneshot()
  itimer: Use printk_once instead of WARN_ONCE
  nohz: Fix stale jiffies update in tick_nohz_restart()
  tick: Document TICK_ONESHOT config option
  proc: stats: Use arch_idle_time for idle and iowait times if available
  itimer: Schedule silent NULL pointer fixup in setitimer() for removal

Pull x86 fixes from Thomas Gleixner.

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86: Use correct byte-sized register constraint in __add()
  x86: Use correct byte-sized register constraint in __xchg_op()
  x86: vsyscall: Use NULL instead 0 for a pointer argument

If, in __persistent_ram_init(), the call to
persistent_ram_buffer_init() fails or the call to
persistent_ram_init_ecc() fails then we fail to free the memory we
allocated to 'prz' with kzalloc() - thus leaking it.

To prevent the leaks I consolidated all error exits from the function
at a 'err:' label at the end and made all error cases jump to that
label where we can then make sure we always free 'prz'. This is safe
since all the situations where the code bails out happen before 'prz'
has been stored anywhere and although we'll do a redundant kfree(NULL)
call in the case of kzalloc() itself failing that's OK since kfree()
deals gracefully with NULL pointers and I felt it was more important
to keep all error exits at a single location than to avoid that one
harmless/redundant kfree() on a error path.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Acked-by: Colin Cross <ccross@android.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>