Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platform-drivers-x86

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platform-drivers-x86: (32 commits)
  Move N014, N051 and CR620 dmi information to load scm dmi table
  drivers/platform/x86/eeepc-wmi.c: fix build warning
  X86 platfrom wmi: Add debug facility to dump WMI data in a readable way
  X86 platform wmi: Also log GUID string when an event happens and debug is set
  X86 platform wmi: Introduce debug param to log all WMI events
  Clean up all objects used by scm model when driver initial fail or exit
  msi-laptop: fix up some coding style issues found by checkpatch
  msi-laptop: Add i8042 filter to sync sw state with BIOS when function key pressed
  msi-laptop: Set rfkill init state when msi-laptop intiial
  msi-laptop: Add MSI CR620 notebook dmi information to scm models table
  msi-laptop: Add N014 N051 dmi information to scm models table
  drivers/platform/x86: Use kmemdup
  drivers/platform/x86: Use kzalloc
  drivers/platform/x86: Clarify the MRST IPC driver description slightly
  eeepc-wmi: depends on BACKLIGHT_CLASS_DEVICE
  IPC driver for Intel Mobile Internet Device (MID) platforms
  classmate-laptop: Add RFKILL support.
  thinkpad-acpi: document backlight level writeback at driver init
  thinkpad-acpi: clean up ACPI handles handling
  thinkpad-acpi: don't depend on led_path for led firmware type (v2)
  ...

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx:
  DMAENGINE: extend the control command to include an arg
  async_tx: trim dma_async_tx_descriptor in 'no channel switch' case
  DMAENGINE: DMA40 fix for allocation of logical channel 0
  DMAENGINE: DMA40 support paused channel status
  dmaengine: mpc512x: Use resource_size
  DMA ENGINE: Do not reset 'private' of channel
  ioat: Remove duplicated devm_kzalloc() calls for ioatdma_device
  ioat3: disable cacheline-unaligned transfers for raid operations
  ioat2,3: convert to producer/consumer locking
  ioat: convert to circ_buf
  DMAENGINE: Support for ST-Ericssons DMA40 block v3
  async_tx: use of kzalloc/kfree requires the include of slab.h
  dmaengine: provide helper for setting txstate
  DMAENGINE: generic channel status v2
  DMAENGINE: generic slave control v2
  dma: timb-dma: Update comment and fix compiler warning
  dma: Add timb-dma
  DMAENGINE: COH 901 318 fix bytesleft
  DMAENGINE: COH 901 318 rename confusing vars

* 'for-linus' of git://neil.brown.name/md: (45 commits)
  md: don't insist on valid event count for spare devices.
  md: simplify updating of event count to sometimes avoid updating spares.
  md/raid6: Fix raid-6 read-error correction in degraded state
  md: restore ability of spare drives to spin down.
  md: Fix read balancing in RAID1 and RAID10 on drives > 2TB
  md/linear: standardise all printk messages
  md/raid0: tidy up printk messages.
  md/raid10: tidy up printk messages.
  md/raid1: improve printk messages
  md/raid5: improve consistency of error messages.
  md: remove EXPERIMENTAL designation from RAID10
  md: allow integers to be passed to md/level
  md: notify mdstat waiters of level change
  md/raid4: permit raid0 takeover
  md/raid1: delay reads that could overtake behind-writes.
  md/raid1: fix confusing 'redirect sector' message.
  md: don't unregister the thread in mddev_suspend
  md: factor out init code for an mddev
  md: pass mddev to make_request functions rather than request_queue
  md: call md_stop_writes from md_stop
  ...

Conflicts:
	drivers/md/md.c

- Resolved conflict in md_update_sb
- Added extra 'NULL' arg to new instance of sysfs_get_dirent.
Signed-off-by: NNeilBrown <neilb@suse.de>

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (577 commits)
  Staging: ramzswap: Handler for swap slot free callback
  swap: Add swap slot free callback to block_device_operations
  swap: Add flag to identify block swap devices
  Staging: vt6655: use ETH_FRAME_LEN macro instead of custom one
  Staging: vt6655: use ETH_DATA_LEN macro instead of custom one
  Staging: vt6655: use ETH_FCS_LEN macro instead of custom one
  Staging: vt6656: use ETH_HLEN macro instead of custom one
  Staging: comedi: quatech_daqp_cs.c Replace eos semaphore with a completion.
  Staging: dt3155v4l: remove private memory allocator
  Staging: crystalhd: Remove typedefs from driver
  Staging: winbond: Fix for pointer name format issue in mds.c
  Staging: vt6656: removed custom UCHAR/USHORT/UINT/ULONG/ULONGLONG typedefs
  Staging: vt6656: removed custom CHAR/SHORT/INT/LONG typedefs
  Staging: comedi: Altered the way printk is used in 8255.c
  staging: iio: adis16350 and similar IMU driver
  Staging: iio: max1363 Fix two bugs in single_channel_from_ring
  Staging: iio: adis16220 extract bin_attribute structures from state
  Staging: iio: adis16220 vibration sensor driver
  Staging: comedi: Kconfig dependancy fixes
  Staging: comedi: fix up build error from last Kconfig changes
  ...

* 'for-2.6.35' of git://git.kernel.dk/linux-2.6-block: (86 commits)
  pipe: set lower and upper limit on max pages in the pipe page array
  pipe: add support for shrinking and growing pipes
  drbd: This is now equivalent to drbd release 8.3.8rc1
  drbd: Do not free p_uuid early, this is done in the exit code of the receiver
  drbd: Null pointer deref fix to the large "multi bio rewrite"
  drbd: Fix: Do not detach, if a bio with a barrier fails
  drbd: Ensure to not trigger late-new-UUID creation multiple times
  drbd: Do not Oops when C_STANDALONE when uuid gets generated
  writeback: fix mixed up arguments to bdi_start_writeback()
  writeback: fix problem with !CONFIG_BLOCK compilation
  block: improve automatic native capacity unlocking
  block: use struct parsed_partitions *state universally in partition check code
  block,ide: simplify bdops->set_capacity() to ->unlock_native_capacity()
  block: restart partition scan after resizing a device
  buffer: make invalidate_bdev() drain all percpu LRU add caches
  block: remove all rcu head initializations
  writeback: fixups for !dirty_writeback_centisecs
  writeback: bdi_writeback_task() must set task state before calling schedule()
  writeback: ensure that WB_SYNC_NONE writeback with sb pinned is sync
  drivers/block/drbd: Use kzalloc
  ...

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus: (25 commits)
  MIPS: Use GCC __builtin_prefetch() to implement prefetch().
  MIPS: Octeon: Serial port fixes for OCTEON simulator.
  MIPS: Octeon: Get rid of early serial.
  MIPS: AR7: prevent race between clock initialization and devices registration
  MIPS: AR7: use ar7_has_high_vlynq() to determine watchdog base address
  MIPS: BCM63xx: Avoid namespace clash on GPIO_DIR_{IN,OUT}
  MIPS: MTX-1: Update defconfig
  MIPS: BCM47xx: Update defconfig
  MIPS: RB532: Update defconfig
  MIPS: AR7: Update defconfig
  RTC: rtc-cmos: Fix binary mode support
  MIPS: Oprofile: Loongson: Cleanup the comments
  MIPS: Oprofile: Loongson: Cleanup of the macros
  MIPS: Oprofile: Loongson: Remove unused variable from loongson2_cpu_setup()
  MIPS: Oprofile: Loongson: Remove useless parentheses
  MIPS: Oprofile: Loongson: Unify macro for setting events
  MIPS: nofpu and nodsp only affect CPU0
  MIPS: Clean up tables for bootmem allocation
  MIPS: Coding style cleanups of access of FCSR rounding mode bits
  MIPS: Loongson 2F: Add gpio/gpioilb support
  ...

Fix kernel-doc warnings, kernel-doc special characters, and
typos in recent kernel/sysctl.c additions.
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Cc: Amerigo Wang <amwang@redhat.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

* 'for-linus' of git://gitorious.org/linux-omap-dss2/linux: (23 commits)
  OMAP: DSS2: Fix omap_dss_probe() error path
  OMAP: DSS2: omap_dss_probe() conditional compilation cleanup
  board-omap3-beagle: add DSS2 support
  OMAP2: DSS: Add missing line for update bg color
  OMAP3630: DSS2: Updating MAX divider value
  OMAP: RX51: Update board defconfig
  OMAP: DSS2: Add ACX565AKM Panel Driver
  OMAP: RX51: Add Touch Controller in SPI board info
  OMAP: RX51: Add LCD Panel support
  OMAP: DSS2: TPO-TD03MTEA1: fix Kconfig dependency
  OMAP: LCD LS037V7DW01: Add Backlight driver support
  OMAP: DSS2: Taal: Fix DSI bus locking problem
  OMAP: DSS2: Taal: add mutex to protect panel data
  OMAP: DSS2: Make partial update width even
  OMAP: DSS2: Fix device disable when driver is not loaded
  OMAP: DSS2: VENC: don't call platform_enable/disable() twice
  OMAP: DSS2: check lock_fb_info() return value
  OMAP: DSS2: fix lock_fb_info() and omapfb_lock() locking order
  OMAP: DSS2: Use vdds_sdi regulator supply in SDI
  OMAP: DSS2: Remove redundant enable/disable calls from SDI
  ...

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (46 commits)
  random: simplify fips mode
  crypto: authenc - Fix cryptlen calculation
  crypto: talitos - add support for sha224
  crypto: talitos - add hash algorithms
  crypto: talitos - second prepare step for adding ahash algorithms
  crypto: talitos - prepare for adding ahash algorithms
  crypto: n2 - Add Niagara2 crypto driver
  crypto: skcipher - Add ablkcipher_walk interfaces
  crypto: testmgr - Add testing for async hashing and update/final
  crypto: tcrypt - Add speed tests for async hashing
  crypto: scatterwalk - Fix scatterwalk_done() test
  crypto: hifn_795x - Rename ablkcipher_walk to hifn_cipher_walk
  padata: Use get_online_cpus/put_online_cpus in padata_free
  padata: Add some code comments
  padata: Flush the padata queues actively
  padata: Use a timer to handle remaining objects in the reorder queues
  crypto: shash - Remove usage of CRYPTO_MINALIGN
  crypto: mv_cesa - Use resource_size
  crypto: omap - OMAP macros corrected
  padata: Use get_online_cpus/put_online_cpus
  ...

Fix up conflicts in arch/arm/mach-omap2/devices.c

* 'msm-video' of git://codeaurora.org/quic/kernel/dwalker/linux-msm:
  drivers: video: msm: add include msm_mdp.h
  drivers: video: msm: default to no

* 'msm-mmc_sdcc' of git://codeaurora.org/quic/kernel/dwalker/linux-msm:
  drivers: mmc: msm_sdcc: Add EMBEDDED_SDIO support
  mmc: msm_sdcc: Fix issue where clocks could be disabled mid transaction
  mmc: msm_sdcc: Fix the dma exec function to use the proper delays
  mmc: msm_sdcc: Don't set host->curr.mrq until after we're sure the busclk timer won't fire
  mmc: msm_sdcc: Enable busclk idle timer for power savings
  mmc: msm_sdcc: Don't disable interrupts while suspending
  mmc: msm_sdcc: Fix issue where we might not end a sucessfull request
  mmc: msm_sdcc: Featurize busclock power save and disable it by default
  mmc: msm_sdcc: Fix bug where busclk expiry timer was not properly disabled
  mmc: msm_sdcc: Reduce command timeouts and improve reliability.
  mmc: msm_sdcc: Schedule clock disable after probe
  mmc: msm_sdcc: Wrap readl/writel calls with appropriate clk delays
  mmc: msm_sdcc: Driver clocking/irq improvements
  msm: Add 'execute' datamover callback
  mmc: msm_sdcc: Snoop SDIO_CCCR_ABORT register
  mmc: msm_sdcc: Clean up clock management and add a 10us delay after enabling clocks

GCC's __builtin_prefetch() was introduced a long time ago, all
supported GCC versions have it.  Lets do what the big boys up in
linux/prefetch.h do, except we use '1' as the third parameter to
provoke 'PREF 0,...'  and 'PREF 1,...' instead of other prefetch
hints.

This allows for better code generation.  In theory the existing
embedded asm could be optimized, but the compiler has these builtins,
so there is really no point.
Signed-off-by: NDavid Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1235/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

For the simulator, fake a slow clock to get fast output.

In prom_putchar we have to mask the value so the simulator doesn't
ASSERT when printing non-ASCII characters.
Signed-off-by: NDavid Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1255/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Get rid of early_serial_setup, we use CONFIG_EARLY_PRINTK instead.
Signed-off-by: NDavid Daney <ddaney@caviumnetworks.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1254/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

ar7_regiser_devices needs ar7_clocks_init to have been called first,
however clock.o is currently linked later due to its order in the Makefile,
therefore ar7_clocks_init always gets called later than ar7_register_devices
because both have the same initcall level. Fix this by moving
ar7_register_devices to the right initcall level.
Reported-by: NMichael J. Evans <mjevans1983@gmail.com>
Signed-off-by: NFlorian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Cc: Ralf Baechle <ralf@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/1212/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Instead of doing yet another switch/case on the chip_id, use existing
inline function to set the watchdog base address.
Signed-off-by: NFlorian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1211/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

This is too generic a name, so prefix it with BCM63XX_ to avoid potential
namespace clashes when including <asm/gpio.h>.
Signed-off-by: NFlorian Fainelli <ffainelli@freebox.fr>
To:     linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1171/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Updates the 4G Systems MTX-1 defconfig, and remove systems unavailable
on this platform.  Changes include:

- built-in MTD char/block support
- GPIO sysfs support
- MTX-1 watchdog driver
- I2C over GPIO driver as a module
- SquashFS support
Signed-off-by: NFlorian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1168/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Update Broadcom BCM47xx defconfig with:

- tiny RCU
- BCM47xx watchdog driver
- b43/b43legacy wireless as modules
- SSB Gigabit Ethernet driver
- disabled IRDA

6igned-off-by: Florian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1167/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

This updates the Mikrotik RB532 defconfig with:

- tiny RCU
- RB532 input buttons driver
- RB532 LED driver
- RC32434 watchdog driver
- GPIO sysfs support
- Wireless support
- SquashFS support
- more LED triggers
Signed-off-by: NFlorian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1166/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Update Texas Instruments AR7 defconfig with:

- tiny RCU
- LEDs GPIO
- disable SSB
- enable zboot support
- enable GPIO sysfs support
Signed-off-by: NFlorian Fainelli <florian@openwrt.org>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1165/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

As a follow-up to the thread about RTC support for some Loongson 2E/2F
boards, this patch tries to address the "REVISIT"/"FIXME" comments about
rtc binary mode handling and allow rtc to work with rtc in binary mode.
I've also raised the message about 24-h mode not supported to warning
otherwise, one may end up with no rtc without any message in the kernel
log.
Signed-off-by: NArnaud Patard <apatard@mandriva.com>
To: linux-mips@linux-mips.org
To: rtc-linux@googlegroups.com
Cc: david-b@pacbell.net
Cc: a.zummo@towertech.it
Cc: akpm@linux-foundation.org
Patchwork: http://patchwork.linux-mips.org/patch/1158/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Removes some out-of-date comments and empty lines.
Signed-off-by: NWu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1204/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The _EXL, _KERNEL etc. bits are in the performance control register so
use _PERFCTRL prefix instead of _PERFCNT.  While at it make the macro
more readable, use _ENABLE instead of _INT_EN suffix to describe the
interrupt enable bit.
Signed-off-by: NWu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1203/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NWu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1202/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NWu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1201/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Unified macro for counter0 and counter1 to set the event in the control
register.  This will be needed by Perf.
Signed-off-by: NWu Zhangjin <wuzhangjin@gmail.com>
Cc: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1200/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

The "nofpu" and "nodsp" kernel command line options currently do not
affect CPUs that are brought online later in the boot process or
hotplugged at runtime.  It is desirable to apply the nofpu/nodsp options
to all CPUs in the system, so that surprising results are not seen when
a process migrates from one CPU to another.

[Ralf: Moved definitions of mips_fpu_disabled, fpu_disable,
mips_dsp_disabled and dsp_disable from setup.c to cpu-probe.c to allow
making mips_fpu_disabled and mips_dsp_disabled static.]
Signed-off-by: NKevin Cernekee <cernekee@gmail.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: http://patchwork.linux-mips.org/patch/1169/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Modifications to the boot memory allocation structures to make them easier
to read and maintain. Note that this will not pass checkpatch because
it wants a structure element initializer to be enclosed in a
do {...} while(...), which is obvious nonsensical.
Signed-off-by: NDavid VomLehn <dvomlehn@cisco.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1207/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Replaces references to the magic number 0x3 with constants and macros
indicating the real purpose of those bits.  They are the rounding mode
bits of the FCSR register.
Signed-off-by: NShane McDonald <mcdonald.shane@gmail.com>
To: anemo@mba.ocn.ne.jp
To: kevink@paralogos.com
To: linux-mips@linux-mips.org
To: sshtylyov@mvista.com
Patchwork: http://patchwork.linux-mips.org/patch/1206/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Signed-off-by: NArnaud Patard <apatard@mandriva.com>
To: linux-mips@linux-mips.org
Patchwork: http://patchwork.linux-mips.org/patch/1163/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Add a sysdev for DBDMA PM.
Signed-off-by: NManuel Lauss <manuel.lauss@gmail.com>
To: Linux-MIPS <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/1119/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Use a sysdev to implement PM methods for the Au1000 interrupt controllers.
Signed-off-by: NManuel Lauss <manuel.lauss@gmail.com>
To: Linux-MIPS <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/1114/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Clean out stray unused board_reset() calls in pb1x boards, the PB1000 is
different from the rest and gets private methods.

(Cleanup after 32fd6901)
Signed-off-by: NManuel Lauss <manuel.lauss@gmail.com>
To: Linux-MIPS <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/1085/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Loongson doesn't support MIPSR2, therefore, MIPSR2 vectored interrupts
(cpu_has_vint) and MIPSR2 external interrupt controller mode
(cpu_has_veic) are 0.
Signed-off-by: NWu Zhangjin <wuzhangjin@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Patchwork: http://patchwork.linux-mips.org/patch/1112/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

From: Julia Lawall <julia@diku.dk>

Use set_cpus_allowed_ptr rather than set_cpus_allowed.

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression E1,E2;
@@

- set_cpus_allowed(E1, cpumask_of_cpu(E2))
+ set_cpus_allowed_ptr(E1, cpumask_of(E2))

@@
expression E;
identifier I;
@@

- set_cpus_allowed(E, I)
+ set_cpus_allowed_ptr(E, &I)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
To: peterz@infradead.org
To: mingo@elte.hu
To: tglx@linutronix.de
To: oleg@redhat.com
To: linux-mips@linux-mips.org
To: linux-kernel@vger.kernel.org
To: kernel-janitors@vger.kernel.org
Patchwork: http://patchwork.linux-mips.org/patch/1087/Signed-off-by: NRalf Baechle <ralf@linux-mips.org>

Conflicts:
	drivers/staging/arlan/arlan-main.c
	drivers/staging/comedi/drivers/cb_das16_cs.c
	drivers/staging/cx25821/cx25821-alsa.c
	drivers/staging/dt3155/dt3155_drv.c
	drivers/staging/hv/hv.c
	drivers/staging/netwave/netwave_cs.c
	drivers/staging/wavelan/wavelan.c
	drivers/staging/wavelan/wavelan_cs.c
	drivers/staging/wlags49_h2/wl_cs.c

This required a bit of hand merging due to the conflicts
that happened in the later .34-rc releases, as well as
some staging driver changing coming in through other trees
(v4l and pcmcia).
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Conflicts:
	fs/ext3/fsync.c
Signed-off-by: NJens Axboe <jens.axboe@oracle.com>

We need at least two to guarantee proper POSIX behaviour, so
never allow a smaller limit than that.

Also expose a /proc/sys/fs/pipe-max-pages sysctl file that allows
root to define a sane upper limit. Make it default to 16 times the
default size, which is 16 pages.
Signed-off-by: NJens Axboe <jens.axboe@oracle.com>