Merge tag 'topic/devm' of git://git./linux/kernel/git/broonie/regulator into HEAD

Provide managed versions of regulator_get() and regulator_bulk_get(),
saving code in error handling and cleanup paths by ensuring that that
the regulators will be automatically unregistered when the device is
unregistered.

ASoC: soc-pcm: msbits constraint: Drop 8 and 16 bit sample sizes

As per discussion we can safely ignore the 8 and 16 bit sample
sizes when applying the msbits constraint.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8962: Optimise power consumption for IN4 DC measurement usage

When the hardware is configured with one or both of the IN4 inputs used
for DC measurement (with no DC blocking capacitor connected) then we can
improve power consumption slightly in idle modes by applying a register
write sequence. Provide platform data to enable this, implemented using
a regmap patch.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm5100: Fix warnings from recent patches

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert the WM5100 revision A updates to a regmap patch

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

Merge tag 'regmap-patch-initial' of git://git./linux/kernel/git/broonie/regmap into for-3.4

regmap: Inital export of the patching implementation

An initial export of the regmap register patch implementation, mostly
for integration into ASoC so it can be used by drivers there which have
conflicting updates already.

ASoC: wm5100: Fix mismerge of IRQ frees

We only want them at the device level, not at the CODEC level.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8731 to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

regmap: Support register patch sets

Device manufacturers frequently provide register sequences, usually not
fully documented, to be run at startup in order to provide better defaults
for devices (for example, improving performance in the light of silicon
evaluation). Support such updates by allowing drivers to register update
sets with the core. These updates will be written to the device immediately
and will also be rewritten when the cache is synced.

The assumption is that the reason for resyncing the cache will always be
that the device has been powered off. If this turns out to not be the case
then a separate operation can be provided.

Currently the implementation only allows a single set of updates to be
specified for a device, this could be extended in future.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Add __devinit/__devexit annotations at necessary places

Fix below build warning when CONFIG_HOTPLUG is not set.

  CC      sound/soc/codecs/alc5623.o
sound/soc/codecs/alc5623.c:1062: warning: ‘alc5623_i2c_remove’ defined but not used
  CC      sound/soc/codecs/alc5632.o
sound/soc/codecs/alc5632.c:1112: warning: ‘alc5632_i2c_remove’ defined but not used

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Leon Romanovsky <leon@leon.nu>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: sgtl5000: It's sgtl5000 not sgtl500

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: sgtl5000: Convert to table based DAPM and control init

Convert to table based DAPM and control init.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

regulator: Provide devm_regulator_bulk_get()

Allow drivers to benefit from both the bulk APIs and managed resources
simultaneously.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

regulator: Implement devm_regulator_free()

Allow consumers to free regulators allocated using devm_regulator_get()
if they need to. This will not normally be required.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8988: Convert to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8988: Convert to table based DAPM and control init

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8985: Convert to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8985 Don't directly reference the cache data structure

In preparation for conversion to regmap.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8985: Use standard cache sync implementation

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8985: Convert to table based DAPM and control init

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8985: Convert to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Add external amplifier controls for Visstrim_M10.

Visstrim_M10 has an external class D amplifier.
This patch provides support for controlling the 4
possible gain levels and per channel muting.

Signed-off-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Fix build of tlv320dac33

The problem was introduced due to the obscure formatting some of the
older drivers use.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Use regmap update bits operation for drivers using regmap

If a driver is using regmap directly ensure that we're coherent with
non-ASoC register updates by using the regmap API directly to do our
read/modify/write cycles. This will bypass the ASoC cache but drivers
using regmap directly should not be using the ASoC cache.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Don't tell applications about msbits unless we're ignoring input

On the off chance that an application both pays attention and gets
confused.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Push WM5100 interrupt request into I2C probe

This is more what the device model wants us to do and will allow use by
non-audio functions before the audio part of the device has come up.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make WM5100 interrupt path use regmap directly

This will allow us to move the interrupt allocation out of the ASoC part
of the driver and simplifies the locking by removing any reliance in the
bulk of the interrupt path on the big CODEC lock.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Remove redundant set_bias_level() from WM5100 remove()

The framework should bring the device down before it calls the driver.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: omap-mcpdm: Set 24msbits constraint

McPDM internal FIFO is 24 bit wide. From the 32 bit sample
8 bit is discarded. Let application know about this
via msbits constraint.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: omap-dmic: Use core to set the msbits constraint

Core can set the msbits constraint in behalf of the dai.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: twl4030: Use core to set the msbits constraint

Core can set the msbits constraint in behalf of the dai.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: tlv320dac33: Use core to set the msbits constraint

Core can set the msbits constraint in behalf of the dai.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: sgtl5000: Print revision number in hex

Throughout the sgtl5000 driver source code and also in the sgtl5000 datasheet
the revision code is shown in hexadecimal.

Print it hex format, for consistency.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Say we can't read WM5100 ID register

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: When releasing WM5100 put /RESET into reset

Reset is active low, make sure we leave it asserted when release the
device.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: 24 bits are significant on wm_hubs DAIs

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make WM8993 I2C usage unconditional

The WM8993 only supports I2C so don't ifdef the I2C support in the driver.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Implement basic WM8993 interrupt support

If an interrupt is supplied then use it for thermal warning and FLL lock
notifications. When using the interrupt raise the timeout for the FLL lock
substantially to reduce the chances of spurious warnings.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: 24 bits are significant on the WM8996 audio interfaces

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Allow drivers to specify how many bits are significant on a DAI

Most devices accept data in formats that don't correspond directly to
their internal format. ALSA allows us to set a msbits constraint which
tells userspace about this in case it finds it useful (for example, in
order to avoid wasting effort dithering bits that will be ignored when
raising the sample size of data) so provide a mechanism for drivers to
specify the number of bits that are actually significant on a DAI and
add the appropriate constraints along with all the others.

This is done slightly awkwardly as the constraint is specified per sample
size - we loop over every possible sample size, including ones that the
device doesn't support and including ones that have fewer bits than are
actually used, but this is harmless as the upper layers do the right thing
in these cases.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Liam Girdwood <lrg@ti.com>

ASoC: Add WM8962 DAC and ADC L/R swap support

Simple switches since there's no per-channel control.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: wm8983: Remove useless snd_kcontrol

This must be a leftover from a previous driver.

Signed-off-by: Dimitris Papastamos <dp@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: jz4740: Convert qi_lb60 to use snd_soc_register_card()

Use snd_soc_register_card() instead of creating a "soc-audio" platform device.

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8962 register access map to modern style

Much more compact, both in terms of source and especially in terms of
RAM used at runtime.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Route Mic Bias in Visstrim_M10 board.

Visstrim_M10 board uses an external microphone
that can be enabled/disabled by the user

Signed-off-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ep93xx: Don't use system controller defines in audio drivers

Both the Snapper CL15 and EDB93xx audio drivers set the same audio
configuration in ep93xx_i2s_acquire. Remove the arguments to
ep93xx_i2s_acquire so that the audio drivers no longer need the
EP93XX_SYSCON defines exported.

Cc: Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Mika Westerberg <mika.westerberg@iki.fi>
Cc: Liam Girdwood <lrg@ti.com>
Cc: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Ryan Mallon <rmallon@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make WM8962 I2C usage unconditional

We only support I2C so no need to ifdef.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: CQ93VC: remove machine specific header file inclusion from codec driver

remove unnecessary inclusion of machine specific header
file mach/dm365.h from cq93vc.c voice codec driver
which comes in the way of platform code consolidation.

Signed-off-by: Manjunath Hadli <manjunath.hadli@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: utils - Add support for a dummy codec driver.

This is useful to create dummy codec devices where we need to have some
DAI links without a real Codec. e.g. could be used to represent dumb FM,
MODEM, etc

This is also used by dynamic PCM for DAI links that have no codec.

Signed-off-by: Liam Girdwood <lrg@ti.com>
[Fixed the indentation -- broonie]
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: twl6040 - add method to query HS DC offset step size in mV

Provide a method for mach drivers to query the HS DC offset step size in mV.

Signed-off-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Optimise performance of WM8904 ADC 128fs OSR mode

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: core - Improve card registration error messaging for large DAI links.

Print out the offending DAI link entry when a naming error occurs. Makes
thing easier to debug for machines with a large number of DAI links.

Signed-off-by: Liam Girdwood <lrg@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: replace 0xffffffff with DMA_BIT_MASK macro

Signed-off-by: Joachim Eastwood <joachim.eastwood@jotron.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: use proper defines for stream directions in pcm engines

Signed-off-by: Joachim Eastwood <joachim.eastwood@jotron.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8955 to table based DAPM and control init

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8955 to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Use standard register cache sync for WM8955

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8955 to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make I2C usage unconditional in WM8955

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8940 to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make I2C usage unconditional in WM8940

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8904 to direct regmap API usage

The device has a very sparse register map so should benefit from using a
rbtree cache.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Use standard cache sync for WM8904

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8904 to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make WM8904 I2C usage unconditional

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8804 to direct regmap API usage

The register map for this device is actually fairly sparse so the rbtree
should be beneficial.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8804 to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Use standard cache sync for wm8804

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Remove wm8974 private data

It's only ever referenced when being allocated and freed.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8974 to table based DAPM and control init

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make wm8974 I2C usage unconditional

The driver only supports I2C at present.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8971 to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8971 to table based DAPM and control init

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make WM8971 I2C usage unconditional

The driver only supports I2C so no need to worry about SPI only systems.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM8753 to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Move WM8993 resource acquisition and device reset to bus probe

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8993 to direct regmap API usage

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Move WM9090 device identification and reset to I2C probe

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert WM9090 to use regmap directly

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Push wm8978 reset down into the I2C probe

Ensures that we get control of the CODEC earlier and don't try to probe
the card at all if register I/O isn't working.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8978 to direct regmap API usage

Helps push the register cache code down out of ASoC and improves resume
times by using the more efficient regmap cache sync code.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Use standard cache sync for WM8978

Saves a bit of code and supports further refactoring.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8978 to devm_kzalloc()

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Convert wm8978 to table based DAPM and control init

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Remove unused control type from wm8978 driver

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Make WM8978 I2C usage unconditional

The driver only supports I2C.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: Disable register synchronisation for low frequency WM8996 SYSCLK

With a low frequency SYSCLK and a fast I2C clock register synchronisation
may occasionally take too long to take effect, causing I/O issues. Disable
synchronisation in order to avoid any issues.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@kernel.org

ASoC: Don't go through cache when applying WM5100 rev A updates

These are all to either uncached registers or fixes to register defaults,
in the former case the cache won't do anything and in the latter case
we're fixing things so the cache sync will do the right thing.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@kernel.org

ASoC: Mark WM5100 register map cache only when going into BIAS_OFF

Writing to the registers won't work if we do actually manage to hit a fully
powered off state.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Cc: stable@kernel.org

ASoC: tlv320aic32x4: always enable analouge block

Register LDOCTLEN must always be initialized to clear the analog power
control bit, otherwise the analog block will stay deactivated.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: tlv320aic32x4: always enable dividers

Dividers (such as MDAC) are always needed, independent of the codec
being I2S master or slave. Needed on a custom board where the codec has
to be slave.

Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Acked-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

ASoC: sgtl5000: Fix wrong register name in restore

Correct SGTL5000_CHIP_CLK_CTRL to SGTL5000_CHIP_REF_CTRL in
sgtl5000_restore_regs(), and add comment to explain the
restore order.

Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Zeng Zhaoming <zengzm.kernel@gmail.com>
Acked-by: Dong Aisheng <dong.aisheng@linaro.org>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

regulator: Add devm_regulator_get()

Add a resource managed regulator_get() to simplify regulator
usage in drivers. This allows driver authors to "get and forget"
about their regulators by automatically calling regulator_put()
when the driver is detached.

[Fixed up a couple of coding style issues -- broonie]

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

Linux 3.3-rc1

Merge branches 'sched-urgent-for-linus', 'perf-urgent-for-linus' and 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/accounting, proc: Fix /proc/stat interrupts sum

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tracepoints/module: Fix disabling tracepoints with taint CRAP or OOT
  x86/kprobes: Add arch/x86/tools/insn_sanity to .gitignore
  x86/kprobes: Fix typo transferred from Intel manual

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86, syscall: Need __ARCH_WANT_SYS_IPC for 32 bits
  x86, tsc: Fix SMI induced variation in quick_pit_calibrate()
  x86, opcode: ANDN and Group 17 in x86-opcode-map.txt
  x86/kconfig: Move the ZONE_DMA entry under a menu
  x86/UV2: Add accounting for BAU strong nacks
  x86/UV2: Ack BAU interrupt earlier
  x86/UV2: Remove stale no-resources test for UV2 BAU
  x86/UV2: Work around BAU bug
  x86/UV2: Fix BAU destination timeout initialization
  x86/UV2: Fix new UV2 hardware by using native UV2 broadcast mode
  x86: Get rid of dubious one-bit signed bitfield

Merge tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6

gpio bug fixes for v3.3

* tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  gpio: tps65910: Use correct offset for gpio initialization
  gpio/it8761e: Restrict it8761e gpio driver to x86.
  gpio-ml-ioh: cleanup __iomem annotation usage
  gpio-ml-ioh: cleanup NULL pointer checking
  gpio-pch: cleanup __iomem annotation usage
  gpio-pch: cleanup NULL pointer checking

Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  qnx4: don't leak ->BitMap on late failure exits
  qnx4: reduce the insane nesting in qnx4_checkroot()
  qnx4: di_fname is an array, for crying out loud...
  vfs: remove printk from set_nlink()
  wake up s_wait_unfrozen when ->freeze_fs fails

x86, syscall: Need __ARCH_WANT_SYS_IPC for 32 bits

In checkin

  303395ac3bf3 x86: Generate system call tables and unistd_*.h from tables

the feature macros in <asm/unistd.h> were unified between 32 and 64
bits.  Unfortunately 32 bits requires __ARCH_WANT_SYS_IPC and this was
inadvertently dropped.

Reported-by: Dmitry Kasatkin <dmitry.kasatkin@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Link: http://lkml.kernel.org/r/CALLzPKbeXN5gdngo8uYYU8mAow=XhrwBFBhKfG811f37BubQOg@mail.gmail.com

Merge remote-tracking branch 'linus/master' into x86/urgent

Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/linux-security

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security:
  KEYS: Permit key_serial() to be called with a const key pointer
  keys: fix user_defined key sparse messages
  ima: fix cred sparse warning
  MPILIB: Add a missing ENOMEM check