If a kthread happens to use get_user_pages() on an mm (as KSM does),
there's a chance that it will end up trying to read in a swap page, then
oops in grab_swap_token() because the kthread has no mm: GUP passes down
the right mm, so grab_swap_token() ought to be using it.

We have not identified a stronger case than KSM's daemon (not yet in
mainline), but the issue must have come up before, since RHEL has included
a fix for this for years (though a different fix, they just back out of
grab_swap_token if current->mm is unset: which is what we first proposed,
but using the right mm here seems more correct).
Reported-by: Izik Eidus <ieidus@redhat.com>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6:
  sh: Fix up HAVE_PERF_COUNTERS typo.
  sh: Fix up more dma-mapping fallout.
  sh: SH7786 SMP support.
  sh: Wire up the uncached fixmap on sh64 as well.
  sh: Use local TLB flush in set_pte_phys().
  sh: Provide cpu_idle_wait() to fix up cpuidle/SMP build.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA: Add __init/__exit macros to addr.c and cma.c
  IB/ehca: Bump version number
  mlx4_core: Fix dma_sync_single_for_cpu() with matching for_device() calls
  IB/mthca: Replace dma_sync_single() use with proper functions
  RDMA/nes: Fix FIN state handling under error conditions
  RDMA/nes: Fix max_qp_init_rd_atom returned from query device
  IB/ehca: Ensure that guid_entry index is not negative
  IB/ehca: Tolerate dynamic memory operations before driver load

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (38 commits)
  fusion: mptsas, fix lock imbalance
  [SCSI] scsi_transport_fc: replace BUS_ID_SIZE by fixed count
  sd, sr: fix Driver 'sd' needs updating message
  scsi_transport_iscsi: return -EOVERFLOW for Too many iscsi targets
  fc_transport: Selective return value from BSG timeout function
  fc_transport: The softirq_done function registration for BSG request
  sym53c8xx: ratelimit parity errors
  explain the hidden scsi_wait_scan Kconfig variable
  ibmvfc: Fix endless PRLI loop in discovery
  ibmvfc: Process async events before command responses
  libfc: Add runtime debugging with debug_logging module parameter
  libfcoe: Add runtime debugging with module param debug_logging
  fcoe: Add runtime debug logging with module parameter debug_logging
  scsi_debug: Add support for physical block exponent and alignment
  cnic: add NETDEV_1000 and NETDEVICES to Kconfig select
  cnic: Fix __symbol_get() build error.
  Revert "[SCSI] cnic: fix error: implicit declaration of function ‘__symbol_get’"
  ipr: differentiate pci-x and pci-e based adapters
  ipr: add test for MSI interrupt support
  scsi_transport_spi: Blacklist Ultrium-3 tape for IU transfers
  ...

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic:
  asm-generic: add dummy pgprot_noncached()
  lib/checksum.c: fix endianess bug
  asm-generic: hook up new system calls
  asm-generic: list Arnd as asm-generic maintainer
  asm-generic: drop HARDIRQ_BITS definition from hardirq.h
  asm-generic: uaccess: fix up local access_ok() usage
  asm-generic: uaccess: add missing access_ok() check to strnlen_user()

* 'kmemleak' of git://linux-arm.org/linux-2.6:
  kmemleak: Do not force the slab debugging Kconfig options
  kmemleak: use pr_fmt

Indeed FOLL_WRITE matches FAULT_FLAG_WRITE, matches GUP_FLAGS_WRITE,
and it's tempting to devise a set of Grand Unified Paging flags;
but not today.  So until then, let's rely upon the compiler to spot
the coincidence, "rather than have that subtle dependency and a
comment for it" - as you remarked in another context yesterday.
Signed-off-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Acked-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

handle_mm_fault() is now passing fault flags rather than write_access
down to hugetlb_fault(), so better recognize that in hugetlb_fault(),
and in hugetlb_no_page().
Signed-off-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Acked-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

* 'for-linus' of git://www.jni.nu/cris:
  CRISv10: remove redundant tests on unsigned
  CRISv32: irq.c - Move end brace outside #endif
  CRISv32: Fix potential null reference in cryptocop driver.
  CRISv32: Add arch optimized strcmp.
  CRIS: assignment/is equal confusion

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: hda - Make jack-plug notification selectable
  ALSA: ctxfi - Add PM support
  sound: seq_midi_event: fix decoding of (N)RPN events
  ALSA: hda - Add digital-mic support to ALC262 auto model
  ALSA: hda - Fix check of input source type for realtek codecs
  ALSA: hda - Add quirk for Sony VAIO Z21MN
  ALSA: hda - Get back Input Source for ALC262 toshiba-s06 model
  ALSA: hda - Fix unsigned comparison in patch_sigmatel.c
  ALSA: via82xx: add option to disable 500ms delay in snd_via82xx_codec_wait
  sound: fix check for return value in snd_pcm_hw_refine
  ALSA: ctxfi - Allow unknown PCI SSIDs
  ASoC: Blackfin: update the bf5xx_i2s_resume parameters
  ASoC: Blackfin: keep better track of SPORT configuration state

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-2.6-watchdog:
  [WATCHDOG] wdt_pci.c: remove #ifdef CONFIG_WDT_501_PCI
  [WATCHDOG] hpwdt: Add NMI priority option
  [WATCHDOG] OMAP fixes: enable clock in probe, trigger timer reload
  [WATCHDOG] add bcm47xx watchdog driver
  [WATCHDOG] Freescale STMP: watchdog driver
  [WATCHDOG] twl4030 watchdog driver
  [WATCHDOG] U300 COH 901 327 watchdog driver
  [WATCHDOG] Add pnx833x_wdt

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6: (62 commits)
  V4L/DVB (12131): BUGFIX: An incorrect Carrier Recovery Loop optimization table was being
  V4L/DVB (12130): Fix a redundant compiler warning
  V4L/DVB (12003): v4l2: Move bounding code outside I2C ifdef block
  V4L/DVB (11913): cx231xx: TRY_FMT should not actually set anything
  V4L/DVB (11912): em28xx: Use v4l bounding/alignment function
  V4L/DVB (11911): cx231xx: Use v4l bounding/alignment function
  V4L/DVB (11910): mt9: Use v4l bounding/alignment function
  V4L/DVB (11909): cx23885: Use v4l bounding/alignment function
  V4L/DVB (11908): w8968cf: Use v4l bounding/alignment function
  V4L/DVB (11907): cx88: Use v4l bounding/alignment function
  V4L/DVB (11906): saa7134: Use v4l bounding/alignment function
  V4L/DVB (11905): vivi: Use v4l bounding/alignment function
  V4L/DVB (11904): zoran: Use v4l bounding/alignment functiob
  V4L/DVB (11903): sh_mobile_ceu_camera: Use v4l bounding/alignment function
  V4L/DVB (11902): pxa-camera: Use v4l bounding/alignment function
  V4L/DVB (11901): v4l2: Create helper function for bounding and aligning images
  V4L/DVB (12128): v4l2: update framework documentation.
  V4L/DVB (12125): v4l2: add new s_config subdev ops and v4l2_i2c_new_subdev_cfg/board calls
  V4L/DVB (12122): pvrusb2: De-obfuscate code which handles routing schemes
  V4L/DVB (12121): pvrusb2: Improve handling of routing schemes
  ...

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  libata: don't set IORDY for reset
  sata_fsl: Add power mgmt support
  [libata] PATA driver for CF interface on AT91SAM9260 SoC
  [libata] beautify module parameters

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/ide-2.6:
  ide cmd64x: Remove serialize setting.
  ide: Take over as maintainer.

Add __init and __exit annotations to the module_init/module_exit
functions from drivers/infiniband/core/addr.c and cma.c.
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Acked-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

Increment version number for DMEM toleration.
Signed-off-by: Alexander Schmidt <alexs@linux.vnet.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

The isolated page is "cursor_page" not "page".

This could cause LRU list corruption under memory pressure, caught by
CONFIG_DEBUG_LIST.
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

That's HAVE_PERF_COUNTERS not HAVE_PERF_COUNTER. This was right
initially but I seem to have screwed it up while re-typing it out
by hand on another machine when I checked it in. Hmph.
Reported-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

Selecting DEBUG_SLAB or SLUB_DEBUG by the KMEMLEAK menu entry may cause
issues with other dependencies (KMEMCHECK). These configuration options
aren't strictly needed by kmemleak but they may increase the chances of
finding leaks. This patch also updates the KMEMLEAK config entry help
text.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Pekka Enberg <penberg@cs.helsinki.fi>

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>

Most architectures now provide a pgprot_noncached(), the
remaining ones can simply use an dummy default implementation,
except for cris and xtensa, which should override the
default appropriately.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Jesper Nilsson <jesper.nilsson@axis.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Magnus Damm <magnus.damm@gmail.com>

commit dbe6f186
("dma-mapping: mark dma_sync_single and dma_sync_sg as deprecated"
conveniently broke every single SH build.

In the future it would be great if people could at least bother
figuring out how to use grep.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

SH7786 is roughly identical to SH-X3 proto SMP, though there are only 2
CPUs. This just wraps in to the existing SH-X3 SMP code with some minor
changes for SH7786, including wiring up the IPIs properly, enabling
IRQ_PER_CPU, and so forth.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

Now that sh64 also can use the uncached section, wire up the fixmap for
it as well.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

set_pte_phys() presently uses the global flush_tlb_one(), which locks on
SMP trying to do the IPI. As we have not even initialized the other CPUs
at this point, switch to the local_ variant so the flush happens on the
boot CPU.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

Crib the x86 cpu_idle_wait() implementation and shove it in with the
idle code, subsequently enabling ARCH_HAS_CPU_IDLE_WAIT.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>

Since dmanr is unsigned, negatives are wrapped and caught by the other test.
Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Acked-by: Jesper Nilsson <jesper.nilsson@axis.com>

Change the wdt_pci.c watchdog driver so that the code is the same for
both the PCI-WDT500 as the PCI-WDT501 card. The selection of the card
is now being done via the module parameter: 'type' instead of the
config option CONFIG_WDT_501_PCI.
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>

Add a priority option so that the user can choose if we do the NMI
first or last.
Signed-off-by: Thomas Mingarelli <thomas.mingarelli@hp.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>

This patch contains two fixes:

1)In omap_wdt_probe() the watchdog is reset and disabled. This
requires register access and the clks needs to be enabled temporarily

2)In omap_wdt_open() the timer register needs to be reloaded
to trigger a new timer value (the default of 60s)

Tested on OMAP34xx platform (Zoom1)
Reviewed-by: Kevin Hilman <khilman@ti.deeprootsystems.com>
Signed-off-by: Ulrik Bech Hald <ubh@ti.com>
Signed-off-by: Wim Van Sebroeck <wim@iguana.be>

loaded for a given chip version. This would cause the optimization in
tuning not to be applied and thus a failed expectation, in tuning speed
increment. The patch swaps the tables in use. It also fixes a possible
one in a million condition where state->dev_ver implies an older Cut
(Cut < 2.0, eventhough the driver doesn't attach to any Cut older than
2.0) or even negative (due to a bad I2C bus master driver) for the card
combination.

Thanks to Mauro Carvalho Chehab <mchehab@infradead.org> for pointing
out the issue at large.
Signed-off-by: Manu Abraham <manu@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

drivers/media/dvb/frontends/stv090x.c: In function ‘stv090x_optimize_carloop_short’:
drivers/media/dvb/frontends/stv090x.c:2677: warning: ‘short_crl’ may be used uninitialized in this function
Signed-off-by: Manu Abraham <manu@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

On Fri, 12 Jun 2009, Randy Dunlap wrote:
> From: Randy Dunlap <randy.dunlap@oracle.com>
>
> Move v4l_bound_align_image() outside of an #ifdef CONFIG_I2C block
> so that it is always built.  Fixes a build error:

clamp_align() should be moved as well, since it's only used by
v4l_bound_align_image().  I'm attaching an alternate version that fixes
this.  Labeled the endif too.
Reported-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

In the TRY_FMT handler the function get_scale() is called to find what the
scaler hardware will produce for a requested size.

The problem is that get_scale(struct cx231xx *dev, ..., unsigned int *vscale,
unsigned int *hscale) saves the calculated scale values into both the
pointer arguments and into dev's hscale and vscale fields.  TRY_FMT shouldn't
actually change anything in the device state.

The code to in get_scale() that writes to dev->[hv]scale can just be
deleted.  In all cases when dev's fields should be modified, get_scale()
was called with get_scale(dev, ..., &dev->hscale, &dev->vscale), so dev was
getting updated anyway.

This didn't actually cause a problem because nothing ever actually made use
of the hscale and vscale fields.  I changed cx231xx_resolution_set() to use
those fields rather than re-calculate them with a call to get_scale().

Updating [hv]scale in cx231xx_resolution_set() isn't necessary because
every call of cx231xx_resolution_set() was already preceded by a call to
get_scale() or setting the [hv]scale fields, so they will be always be
up-to-date w.r.t. width and height.

Removing the call to get_scale() from cx231xx_resolution_set() allowed
making get_scale() a static function, which is a good thing for something
with such a short name.  There is already another function with the same
name in the em28xx driver, but that one is static.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

The v4l function has a better algorithm for aligning image size.

It appears that the em2800 can only scale by 50% or 100%, i.e. the only
heights supported might be 240 and 480.  In that case the old code would
set any height other than 240 to 480.  Request 240 get 240, but request 239
and then you get 480.  Change it to round to the nearest supported value.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

The v4l function has a better algorithm for aligning image size.

Cc: Srinivasa Deevi <srinivasa.deevi@conexant.com>
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

The v4l function has a better algorithm for aligning image size.

Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

The v4l function has a better algorithm for aligning image size.

The existing code was casting pointers to u32 and to unsigned int into
pointers to u16.  This could mess up if someone passed in an image size
greater than 65,535 and on big-endian platforms it won't work at all.

The existing bounding code would shrink an image if it was too big, but
returned ERANGE if it was too small.  The code will not shrink or expand as
necessary.
Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>