GitHub/moto-9609/android_kernel_motorola_exynos9610.git
14 years agodrm/radeon/kms: enable unmappable vram for evergreen
Alex Deucher [Tue, 19 Oct 2010 04:36:19 +0000 (00:36 -0400)]
drm/radeon/kms: enable unmappable vram for evergreen

Evergreen now has blit support, but unmappable vram support
was disabled in c919b371cb734f42b1130e706ecee262f8d9261d
(drm/radeon/kms: avoid corner case issue with unmappable vram V2)
due to merge ordering.  This re-enables unmappable vram on
evergreen.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix tiled db height calculation on 6xx/7xx
Alex Deucher [Wed, 27 Oct 2010 05:44:35 +0000 (01:44 -0400)]
drm/radeon/kms: fix tiled db height calculation on 6xx/7xx

Calculate height based on the slice bitfield rather than the size.
Same as Dave's CB fix.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix handling of tex lookup disable in cs checker on r2xx
Alex Deucher [Wed, 27 Oct 2010 05:02:35 +0000 (01:02 -0400)]
drm/radeon/kms: fix handling of tex lookup disable in cs checker on r2xx

There are cases when multiple texture units have to be enabled,
but not actually used to sample.  This patch checks to see if
the lookup_disable bit is set and if so, skips the texture check.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=25544

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agovmwgfx: Implement a proper GMR eviction mechanism
Thomas Hellstrom [Tue, 26 Oct 2010 19:21:47 +0000 (21:21 +0200)]
vmwgfx: Implement a proper GMR eviction mechanism

Use Ben's new range manager hooks to implement a manager for
GMRs that manages ids rather than ranges.
This means we can use the standard TTM code for binding, unbinding and
eviction.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix r6xx/7xx 1D tiling CS checker v2
Alex Deucher [Wed, 27 Oct 2010 00:22:42 +0000 (20:22 -0400)]
drm/radeon/kms: fix r6xx/7xx 1D tiling CS checker v2

broken by:
drm/radeon/r600: fix tiling issues in CS checker.

v2: only apply it to 1D tiling case.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: properly compute group_size on 6xx/7xx
Alex Deucher [Tue, 19 Oct 2010 03:54:56 +0000 (23:54 -0400)]
drm/radeon/kms: properly compute group_size on 6xx/7xx

Needed for tiled surfaces.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: fix 2D tile height alignment in the r600 CS checker
Alex Deucher [Tue, 19 Oct 2010 03:45:39 +0000 (23:45 -0400)]
drm/radeon/kms: fix 2D tile height alignment in the r600 CS checker

macro tile heights are aligned to num channels, not num banks.

Noticed by Dave Airlie.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms/evergreen: set the clear state to the blit state
Alex Deucher [Thu, 21 Oct 2010 17:31:38 +0000 (13:31 -0400)]
drm/radeon/kms/evergreen: set the clear state to the blit state

The hw stores a default clear state for registers in the context
range that can be initialized when the CP is set up.  Set the
blit state as the default clear state and use the CLEAR_STATE
packet to load the blit state rather than loading it from an IB.
This reduces overhead when doing bo moves using the 3D engine.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: don't poll dac load detect.
Dave Airlie [Tue, 26 Oct 2010 02:55:52 +0000 (12:55 +1000)]
drm/radeon/kms: don't poll dac load detect.

This is slightly destructive, cpu intensive and can cause lockups.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agogpu: Add Intel GMA500(Poulsbo) Stub Driver
Lee, Chun-Yi [Wed, 20 Oct 2010 05:51:19 +0000 (13:51 +0800)]
gpu: Add Intel GMA500(Poulsbo) Stub Driver

Currently, there have no GMA500(Poulsbo) native video driver to support
intel opregion. So, use this stub driver to enable the acpi backlight
control sysfs entry files by requrest acpi_video_register.

[airlied: fix warnings]

Signed-off-by: Lee, Chun-Yi <jlee@novell.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: MC vram map needs to be >= pci aperture size
Alex Deucher [Mon, 25 Oct 2010 23:44:00 +0000 (19:44 -0400)]
drm/radeon/kms: MC vram map needs to be >= pci aperture size

The vram map in the radeon memory controller needs to be
>= the pci aperture size.  Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=28402

The problematic cards in the above bug have 64 MB of vram,
but the pci aperture is 128 MB and the MC vram map was only
64 MB.  This can lead to hangs.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms: implement display watermark support for evergreen
Alex Deucher [Fri, 22 Oct 2010 06:51:05 +0000 (02:51 -0400)]
drm/radeon/kms: implement display watermark support for evergreen

Improper display watermarks can result in underflow to the display
controllers which can cause flickering or other artifacts.
This patch implements display watermark support and line buffer
allocation for evergreen asics.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/kms/evergreen: add some additional safe regs v2
Alex Deucher [Thu, 21 Oct 2010 17:45:30 +0000 (13:45 -0400)]
drm/radeon/kms/evergreen: add some additional safe regs v2

These are needed for enabling dynamic GPR allocation in the shaders
in the userspace acceleration drivers.

v2: fix typo in reg name

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/radeon/r600: fix tiling issues in CS checker.
Dave Airlie [Thu, 21 Oct 2010 03:55:40 +0000 (13:55 +1000)]
drm/radeon/r600: fix tiling issues in CS checker.

The CS checker had some incorrect alignment requirements for 2D surfaces,
this made rendering to mipmap levels that were 2D broken.

Also the CB height was being worked out from the BO size, this doesn't work
at all when rendering mipmap levels, instead we work out what height userspace
wanted from slice max and use that to check it fits inside the BO, however
the DDX send the wrong slice max for an unaligned buffer so we have to workaround
for that even though its a userspace bug.

Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoMerge remote branch 'intel/drm-intel-next' of ../drm-next into drm-core-next
Dave Airlie [Mon, 25 Oct 2010 23:23:22 +0000 (09:23 +1000)]
Merge remote branch 'intel/drm-intel-next' of ../drm-next into drm-core-next

* 'intel/drm-intel-next' of ../drm-next: (63 commits)
  drm/i915: Move gpu_write_list to per-ring
  drm/i915: Invalidate the to-ring, flush the old-ring when updating domains
  drm/i915/ringbuffer: Write the value passed in to the tail register
  agp/intel: Restore valid PTE bit for Sandybridge after bdd3072
  drm/i915: Fix flushing regression from 9af90d19f
  drm/i915/sdvo: Remove unused encoding member
  i915: enable AVI infoframe for intel_hdmi.c [v4]
  drm/i915: Fix current fb blocking for page flip
  drm/i915: IS_IRONLAKE is synonymous with gen == 5
  drm/i915: Enable SandyBridge blitter ring
  drm/i915/ringbuffer: Remove broken intel_fill_struct()
  drm/i915/ringbuffer: Fix emit batch buffer regression from 8187a2b
  drm/i915: Copy the updated reloc->presumed_offset back to the user
  drm/i915: Track objects in global active list (as well as per-ring)
  drm/i915: Simplify most HAS_BSD() checks
  drm/i915: cache the last object lookup during pin_and_relocate()
  drm/i915: Do interrupible mutex lock first to avoid locking for unreference
  drivers: gpu: drm: i915: Fix a typo.
  agp/intel: Also add B43.1 to list of supported devices
  drm/i915: rearrange mutex acquisition for pread
  ...

14 years agodrm/i915: Move gpu_write_list to per-ring
Chris Wilson [Sun, 24 Oct 2010 11:38:05 +0000 (12:38 +0100)]
drm/i915: Move gpu_write_list to per-ring

... to prevent flush processing of an idle (or even absent) ring.

This fixes a regression during suspend from 87acb0a5.

Reported-and-tested-by: Alexey Fisher <bug-track@fisher-privat.net>
Tested-by: Peter Clifton <pcjc2@cam.ac.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Invalidate the to-ring, flush the old-ring when updating domains
Chris Wilson [Sat, 23 Oct 2010 09:15:06 +0000 (10:15 +0100)]
drm/i915: Invalidate the to-ring, flush the old-ring when updating domains

When the object has been written to by the gpu it remains on the ring
until its flush has been retired. However, when the object is moving to
the ring and the associated cache needs to be invalidated, we need to
perform the flush on the target ring, not the one it came from (which is
NULL in the reported case and so the flush was entirely absent).

Reported-by: Peter Clifton <pcjc2@cam.ac.uk>
Reported-and-tested-by: Alexey Fisher <bug-track@fisher-privat.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/ringbuffer: Write the value passed in to the tail register
Chris Wilson [Fri, 22 Oct 2010 16:02:41 +0000 (17:02 +0100)]
drm/i915/ringbuffer: Write the value passed in to the tail register

This should fix the error along the reset path were we tried to clear the
tail register by setting it to 0, but were in fact setting it to the
current value and complaining when it did not reset to 0.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoagp/intel: Restore valid PTE bit for Sandybridge after bdd3072
Chris Wilson [Fri, 22 Oct 2010 13:59:29 +0000 (14:59 +0100)]
agp/intel: Restore valid PTE bit for Sandybridge after bdd3072

In cleaning up the mask functions in bdd3072, the setting of the PTE
valid bit was dropped for Sandybridge.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Fix flushing regression from 9af90d19f
Chris Wilson [Fri, 22 Oct 2010 09:48:12 +0000 (10:48 +0100)]
drm/i915: Fix flushing regression from 9af90d19f

Whilst moving the code around in 9af90d19f, I dropped the or'ing in of
new write domains which would zero out the write domain for a render
target if later reused as a source later in the batch. This meant that
we might drop a required flush before reading from the render target.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=31043
Reported-by: xunx.fang@intel.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/sdvo: Remove unused encoding member
Chris Wilson [Fri, 22 Oct 2010 08:15:22 +0000 (09:15 +0100)]
drm/i915/sdvo: Remove unused encoding member

This block is only used when detecting whether the connector is HDMI and
never again, so scope the variable to the detection routine.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoi915: enable AVI infoframe for intel_hdmi.c [v4]
David Härdeman [Fri, 24 Sep 2010 19:44:32 +0000 (21:44 +0200)]
i915: enable AVI infoframe for intel_hdmi.c [v4]

This patch enables the sending of AVI infoframes in
drivers/gpu/drm/i915/intel_hdmi.c.

My receiver currently loses sync when the HDMI output on my computer
(DG45FC motherboard) is switched from 800x600 (the BIOS resolution) to
1920x1080 as part of the boot. Fixable by switching inputs on the receiver
a couple of times.

With this patch, my receiver has not lost sync yet (> 40 tries).

Fourth version, now based on drm-intel-next from:
git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel.git

Two questions still remain:

I'm assuming that the sdvo hardware also stores a header ECC byte in
the MSB of the first dword - is this correct?

Does the SDVOB and SDVOC handling in intel_hdmi_set_avi_infoframe()
look correct?

Signed-off-by: David Härdeman <david@hardeman.nu>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Fix current fb blocking for page flip
Chia-I Wu [Thu, 21 Oct 2010 18:35:54 +0000 (19:35 +0100)]
drm/i915: Fix current fb blocking for page flip

Block execbuffer for the fb to be flipped away, not the one that is to
be flipped in.

[ickle: rewritten for -next]
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: Kristian Høgsberg <krh@bitplanet.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: IS_IRONLAKE is synonymous with gen == 5
Chris Wilson [Thu, 21 Oct 2010 13:57:17 +0000 (14:57 +0100)]
drm/i915: IS_IRONLAKE is synonymous with gen == 5

So remove the redundant bit in the capabilities block and
s/IS_IRONLAKE/IS_GEN5/.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Enable SandyBridge blitter ring
Chris Wilson [Tue, 19 Oct 2010 10:19:32 +0000 (11:19 +0100)]
drm/i915: Enable SandyBridge blitter ring

Based on an original patch by Zhenyu Wang, this initializes the BLT ring for
SandyBridge and enables support for user execbuffers.

Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/ringbuffer: Remove broken intel_fill_struct()
Chris Wilson [Thu, 21 Oct 2010 18:00:02 +0000 (19:00 +0100)]
drm/i915/ringbuffer: Remove broken intel_fill_struct()

... before someone tries to use it. The code both calls
intel_ring_begin/advance() and open-codes the bookkeeping performed by
those two functions.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/ringbuffer: Fix emit batch buffer regression from 8187a2b
Chris Wilson [Thu, 21 Oct 2010 17:51:09 +0000 (18:51 +0100)]
drm/i915/ringbuffer: Fix emit batch buffer regression from 8187a2b

In commit 8187a2b, the number of dwords used in the ringbuffer for
executing the batch buffer was erroneously changed from 2 to 4.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/ttm: Optimize delayed buffer destruction
Thomas Hellstrom [Tue, 19 Oct 2010 07:01:01 +0000 (09:01 +0200)]
drm/ttm: Optimize delayed buffer destruction

This commit replaces the ttm_bo_cleanup_ref function with two new functions.
One for the case where the bo is not yet on the delayed destroy list, and
one for the case where the bo was on the delayed destroy list, at least at
the time of call. This makes it possible to optimize the two cases somewhat.

It also enables the possibility to directly destroy buffers on the
delayed delete list when they are about to be evicted or swapped out.
Currently they were only evicted / swapped and destruction was left for the
delayed buffer destruction thread.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: Avoid using the ttm_mem_type_manager::put_locked function
Thomas Hellstrom [Tue, 19 Oct 2010 07:01:00 +0000 (09:01 +0200)]
drm/ttm: Avoid using the ttm_mem_type_manager::put_locked function

Release the lru spinlock early.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/i915: Copy the updated reloc->presumed_offset back to the user
Chris Wilson [Wed, 20 Oct 2010 19:59:57 +0000 (20:59 +0100)]
drm/i915: Copy the updated reloc->presumed_offset back to the user

If the userspace driver is using a constant relocation array with a
static buffer, they will pass the same relocation array back to the
kernel. So we *do* need to update the presumed offset value in those
relocations to reflect the current object so that they remain correct
with future batchbuffers and we avoid the necessity of having to suspend
execution and perform redundant relocations.

Fixes the regression introduced by 12f889c for applications using
absolute addressing on trees of buffer (i.e. the current consumers of
libdrm_intel.so).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=30996
Reported-by: Wang, Jinjin <jinjin.wang@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Track objects in global active list (as well as per-ring)
Chris Wilson [Tue, 19 Oct 2010 09:36:51 +0000 (10:36 +0100)]
drm/i915: Track objects in global active list (as well as per-ring)

To handle retirements, we need per-ring tracking of active objects.
To handle evictions, we need global tracking of active objects.

As we enable more rings, rebuilding the global list from the individual
per-ring lists quickly grows tiresome and overly complicated. Tracking the
active objects in two lists is the lesser of two evils.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Simplify most HAS_BSD() checks
Chris Wilson [Tue, 19 Oct 2010 09:13:00 +0000 (10:13 +0100)]
drm/i915: Simplify most HAS_BSD() checks

... by always initialising the empty ringbuffer it is always then safe
to check whether it is active.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: cache the last object lookup during pin_and_relocate()
Chris Wilson [Sun, 17 Oct 2010 09:01:56 +0000 (10:01 +0100)]
drm/i915: cache the last object lookup during pin_and_relocate()

The most frequent relocation within a batchbuffer is a contiguous sequence
of vertex buffer relocations, for which we can virtually eliminate the
drm_gem_object_lookup() overhead by caching the last handle to object
translation.

In doing so we refactor the pin and relocate retry loop out of
do_execbuffer into its own helper function and so improve the error
paths.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Do interrupible mutex lock first to avoid locking for unreference
Chris Wilson [Sun, 17 Oct 2010 08:45:41 +0000 (09:45 +0100)]
drm/i915: Do interrupible mutex lock first to avoid locking for unreference

One of the primarily consumers of the i915 driver is X, a large signal
driven application. Frequently when writing into the buffers, there is a
pending signal which causes us not to take the interruptible lock but
then we need to take that same lock around the object unreference. By
rearranging the code to do the interruptible lock as the first check, we
can avoid the frequent additional locking around the unreference.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrivers: gpu: drm: i915: Fix a typo.
Andrea Gelmini [Fri, 15 Oct 2010 15:14:33 +0000 (17:14 +0200)]
drivers: gpu: drm: i915: Fix a typo.

"userpace" -> "userspace"

Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoagp/intel: Also add B43.1 to list of supported devices
Chris Wilson [Thu, 14 Oct 2010 15:30:41 +0000 (16:30 +0100)]
agp/intel: Also add B43.1 to list of supported devices

This was a missing piece from 41a5142 that dropped recognition of the
AGP module for the second B43 variant.

Reported-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org
14 years agodrm/i915: rearrange mutex acquisition for pread
Chris Wilson [Thu, 14 Oct 2010 14:26:45 +0000 (15:26 +0100)]
drm/i915: rearrange mutex acquisition for pread

... to avoid the double acquisition along fast[er] paths.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Rearrange acquisition of mutex during pwrite
Chris Wilson [Thu, 14 Oct 2010 14:03:58 +0000 (15:03 +0100)]
drm/i915: Rearrange acquisition of mutex during pwrite

... to avoid reacquiring it to drop the object reference count on
exit. Note we have to make sure we now drop (and reacquire) the lock
around acquiring the mm semaphore on the slow paths.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Attempt to prefault user pages for pread/pwrite
Chris Wilson [Thu, 14 Oct 2010 12:47:43 +0000 (13:47 +0100)]
drm/i915: Attempt to prefault user pages for pread/pwrite

... in the hope that it makes the atomic fast paths more likely.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Avoid taking the mutex for dropping the refcnt upon creation
Chris Wilson [Thu, 14 Oct 2010 12:20:40 +0000 (13:20 +0100)]
drm/i915: Avoid taking the mutex for dropping the refcnt upon creation

After allocation a handle for the fresh object, we know that we can
safely drop the refcnt without triggering a free so we do not need the
mutex. Strangely, this mutex acquisition is the one that appears on
driver profiles.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Perform relocations in CPU domain [if in CPU domain]
Chris Wilson [Thu, 14 Oct 2010 11:44:48 +0000 (12:44 +0100)]
drm/i915: Perform relocations in CPU domain [if in CPU domain]

Avoid an early eviction of the batch buffer into the uncached GTT
domain, and so do the relocation fixup in cacheable memory.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Avoid vmallocing a buffer for the relocations
Chris Wilson [Thu, 14 Oct 2010 11:10:41 +0000 (12:10 +0100)]
drm/i915: Avoid vmallocing a buffer for the relocations

... perform an access validation check up front instead and copy them in
on-demand, during i915_gem_object_pin_and_relocate(). As around 20% of
the CPU overhead may be spent inside vmalloc for the relocation entries
when submitting an execbuffer [for x11perf -aa10text], the savings are
considerable and result in around a 10% throughput increase [for glyphs].

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/hdmi: Add 'force_audio' property
Chris Wilson [Sun, 19 Sep 2010 08:29:33 +0000 (09:29 +0100)]
drm/i915/hdmi: Add 'force_audio' property

Allow the user to override the detection of the sink's audio capabilities
from EDID. Not all sinks support the required EDID level to specify
whether they handle audio over the display connection, so allow the user
to enable it manually.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/sdvo: Add 'force_audio' property
Chris Wilson [Sun, 19 Sep 2010 08:29:33 +0000 (09:29 +0100)]
drm/i915/sdvo: Add 'force_audio' property

Allow the user to override the detection of the sink's audio capabilities
from EDID. Not all sinks support the required EDID level to specify
whether they handle audio over the display connection, so allow the user
to enable it manually.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/dp: Add 'force_audio' property
Chris Wilson [Sun, 19 Sep 2010 08:29:33 +0000 (09:29 +0100)]
drm/i915/dp: Add 'force_audio' property

Allow the user to override the detection of the sink's audio capabilities
from EDID. Not all sinks support the required EDID level to specify
whether they handle audio over the display connection, so allow the user
to enable it manually.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Enable HDMI audio for monitor with audio support
Zhenyu Wang [Fri, 10 Sep 2010 02:39:40 +0000 (10:39 +0800)]
drm/i915: Enable HDMI audio for monitor with audio support

Rely on monitor's audio capability to turn on audio output for HDMI.

Tested-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Enable DisplayPort audio
Zhenyu Wang [Sun, 19 Sep 2010 05:09:06 +0000 (13:09 +0800)]
drm/i915: Enable DisplayPort audio

This will turn on DP audio output by checking monitor's audio
capability.

Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
[ickle: rebase onto recent changes and rearranged for clarity]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/edid: add helper function to detect monitor audio capability
Zhenyu Wang [Sun, 19 Sep 2010 06:27:28 +0000 (14:27 +0800)]
drm/edid: add helper function to detect monitor audio capability

To help to determine if digital display port needs to enable
audio output or not. This one adds a helper to get monitor's
audio capability via EDID CEA extension block.

Tested-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Initialize panel timing registers if VBIOS did not
Bryan Freed [Thu, 14 Oct 2010 08:14:51 +0000 (09:14 +0100)]
drm/i915: Initialize panel timing registers if VBIOS did not

The time between start of the pixel clock and backlight enable is a basic
panel timing constraint.  If the Panel Power On/Off registers are found
to be 0, assume we are booting without VBIOS initialization and set these
registers to something reasonable.

Change-Id: Ibed6cc10d46bf52fd92e0beb25ae3525b5eef99d
Signed-off-by: Bryan Freed <bfreed@chromium.org>
[ickle: rearranged into a separate function to distinguish its role from
simply parsing the VBIOS tables.]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Sleep whilst waiting for the ring
Chris Wilson [Wed, 13 Oct 2010 09:09:14 +0000 (10:09 +0100)]
drm/i915: Sleep whilst waiting for the ring

If userspace is submitting so many long running batches that the ring
becomes full, throttle by sleeping for a 1ms before checking for free
space. Simply yielding was causing excessive scheduler overhead whilst
making no progress.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Fix GPIO pin to register mapping
Zhenyu Wang [Wed, 13 Oct 2010 08:40:12 +0000 (16:40 +0800)]
drm/i915: Fix GPIO pin to register mapping

In i2c GPIO fallback, index 6 is reserved for nothing.

Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: restore fixed FDI link rate on Sandybridge
Chris Wilson [Wed, 13 Oct 2010 08:59:17 +0000 (09:59 +0100)]
drm/i915: restore fixed FDI link rate on Sandybridge

FDI_PLL_BIOS_0 register is for Ironlake only, don't apply to
Sandybridge.

Original-patch-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915/dp: down the DP link even if the reg indicates it's already down
Jesse Barnes [Fri, 8 Oct 2010 17:35:55 +0000 (10:35 -0700)]
drm/i915/dp: down the DP link even if the reg indicates it's already down

Since the PLL may still be on, and the training pattern may not be
correct.  Fixes suspend/resume on my PCH eDP test system.

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[ickle: minor merge conflict and silence the compiler]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Fix oops on HWS unload
Daniel Vetter [Sun, 10 Oct 2010 17:54:08 +0000 (18:54 +0100)]
drm/i915: Fix oops on HWS unload

Freeing the Hardware Status Page was writing to the HWS register in
order to disable the GPU writing to the HWS page. Unfortunately, we were
writing to the mmio register after unmapping the register space, hence
the oops.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agodrm/i915: Remove duplicate set of ADPA definitions
Chris Wilson [Sat, 9 Oct 2010 09:33:26 +0000 (10:33 +0100)]
drm/i915: Remove duplicate set of ADPA definitions

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
14 years agoRevert "drm/i915: Prevent module unload to avoid random memory corruption"
Chris Wilson [Fri, 8 Oct 2010 15:28:27 +0000 (16:28 +0100)]
Revert "drm/i915: Prevent module unload to avoid random memory corruption"

This reverts commit 6939a5aca7cfada279a24c307e772f33104fca20.

Daniel Vetter supplied a set of fixes for all the module unload bugs he
could trigger on his machines, so let the fun recommence!

14 years agoMerge remote branch 'airlied/drm-core-next' into tmp
Chris Wilson [Tue, 19 Oct 2010 08:13:00 +0000 (09:13 +0100)]
Merge remote branch 'airlied/drm-core-next' into tmp

14 years agodrm, kdb, kms: Change mode_set_base_atomic() enter argument to be an enum
Jason Wessel [Wed, 13 Oct 2010 19:09:44 +0000 (14:09 -0500)]
drm, kdb, kms: Change mode_set_base_atomic() enter argument to be an enum

The enter argument as implemented by commit 413d45d3627 (drm, kdb, kms:
Add an enter argument to mode_set_base_atomic() API) should be more
descriptive as to what it does vs just passing 1 and 0 around.

There is no runtime behavior change as a result of this patch.

Reported-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
CC: David Airlie <airlied@linux.ie>
CC: dri-devel@lists.freedesktop.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agokdb, kms: Save and restore the LUT on atomic KMS enter/exit
Jason Wessel [Wed, 13 Oct 2010 19:09:43 +0000 (14:09 -0500)]
kdb, kms: Save and restore the LUT on atomic KMS enter/exit

When changing VTs non-atomically the kernel works in conjunction with
the Xserver in user space and receives the LUT information from the
Xserver via a system call.  When changing modes atomically for kdb,
this information must be saved and restored without disturbing user
space as if nothing ever happened.

There is a short cut used by this patch where gamma_store is used as
the save space.  If this turns out to be a problem in the future a
pre-allocated chunk of memory will be required for each crtc to save
and restore the LUT information.

Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
CC: Jesse Barnes <jbarnes@virtuousgeek.org>
CC: David Airlie <airlied@linux.ie>
CC: dri-devel@lists.freedesktop.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoRevert "radeon, kdb, kms: Save and restore the LUT on atomic KMS enter/exit"
Jason Wessel [Wed, 13 Oct 2010 19:09:42 +0000 (14:09 -0500)]
Revert "radeon, kdb, kms: Save and restore the LUT on atomic KMS enter/exit"

This reverts commit ff773714dd30b802c336064109c535d8b2774e2f.

A generic solution is needed to save and retore the LUT information.

CC: Jesse Barnes <jbarnes@virtuousgeek.org>
CC: dri-devel@lists.freedesktop.org
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoagp/amd-k7: Allow binding user memory to the AGP GART.
Francisco Jerez [Sat, 16 Oct 2010 00:45:15 +0000 (00:45 +0000)]
agp/amd-k7: Allow binding user memory to the AGP GART.

TTM-based DRM drivers need to be able to bind user memory to the AGP
aperture. This patch fixes the "[TTM] AGP Bind memory failed." errors
and the subsequent fallout seen with the nouveau driver.

Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Tested-by: Grzesiek Sójka <pld@pfu.pl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: Simplify ttm_bo_wait_unreserved
Jean Delvare [Sat, 9 Oct 2010 12:36:45 +0000 (12:36 +0000)]
drm/ttm: Simplify ttm_bo_wait_unreserved

Function ttm_bo_wait_unreserved can be slightly simplified.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Thomas Hellstrom <thellstrom@vmware.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoRevert "drm/radeon/kms: remove some pll algo flags"
Dave Airlie [Tue, 19 Oct 2010 00:36:47 +0000 (10:36 +1000)]
Revert "drm/radeon/kms: remove some pll algo flags"

This reverts commit f28488c282d8916b9b6190cc41714815bbaf97d5.

On my rv610 test machine the monitor failed to light up after this.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agodrm/ttm: add unlocked variant of new manager put node.
Dave Airlie [Thu, 7 Oct 2010 22:57:10 +0000 (08:57 +1000)]
drm/ttm: add unlocked variant of new manager put node.

We need the unlocked variant for the new codepath introduced to fix the
race condition in master recently.

Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoMerge branch 'drm-fixes' of /home/airlied/kernel/linux-2.6 into drm-core-next
Dave Airlie [Mon, 18 Oct 2010 23:48:34 +0000 (09:48 +1000)]
Merge branch 'drm-fixes' of /home/airlied/kernel/linux-2.6 into drm-core-next

Conflicts:
drivers/gpu/drm/i915/intel_fb.c
drivers/gpu/drm/radeon/r600_blit_kms.c
drivers/gpu/drm/ttm/ttm_bo.c

14 years agodrm/radeon/kms: avivo cursor workaround applies to evergreen as well
Alex Deucher [Thu, 14 Oct 2010 21:14:57 +0000 (17:14 -0400)]
drm/radeon/kms: avivo cursor workaround applies to evergreen as well

Fixes cursor corruption in certain cases.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Cc: stable@kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
14 years agoMerge branch 'fix/misc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Sun, 17 Oct 2010 16:38:08 +0000 (09:38 -0700)]
Merge branch 'fix/misc' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'fix/misc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: rawmidi: fix oops (use after free) when unloading a driver module

14 years agoALSA: rawmidi: fix oops (use after free) when unloading a driver module
Clemens Ladisch [Fri, 15 Oct 2010 10:06:18 +0000 (12:06 +0200)]
ALSA: rawmidi: fix oops (use after free) when unloading a driver module

When a driver module is unloaded and the last still open file is a raw
MIDI device, the card and its devices will be actually freed in the
snd_card_file_remove() call when that file is closed.  Afterwards, rmidi
and rmidi->card point into freed memory, so the module pointer is likely
to be garbage.
(This was introduced by commit 9a1b64caac82aa02cb74587ffc798e6f42c6170a.)

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Reported-by: Krzysztof Foltman <wdev@foltman.com>
Cc: 2.6.30-2.6.35 <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
14 years agom32r: test __LITTLE_ENDIAN__ instead of __LITTLE_ENDIAN
Kyle McMartin [Sat, 16 Oct 2010 01:17:16 +0000 (21:17 -0400)]
m32r: test __LITTLE_ENDIAN__ instead of __LITTLE_ENDIAN

Fixes build for me... these are what's tested in byteorder.h...

Signed-off-by: Kyle McMartin <kyle@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom32r: add kernel/.gitignore and ignore vmlinux.lds
Kyle McMartin [Sat, 16 Oct 2010 01:17:13 +0000 (21:17 -0400)]
m32r: add kernel/.gitignore and ignore vmlinux.lds

Signed-off-by: Kyle McMartin <kyle@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom32r: get_user takes an lvalue, not a pointer
Kyle McMartin [Sat, 16 Oct 2010 01:17:09 +0000 (21:17 -0400)]
m32r: get_user takes an lvalue, not a pointer

Signed-off-by: Kyle McMartin <kyle@redhat.com>
Acked-by: Al "my fuckup" Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom32r: restore _BLOCKABLE
Kyle McMartin [Sat, 16 Oct 2010 01:16:45 +0000 (21:16 -0400)]
m32r: restore _BLOCKABLE

Commit a7f8388e accidentally removed it... Al explains:

  "Sorry, reordering breakage.  In the signals tree here I have

   static inline void sig_set_blocked(struct sigset_t *set)
   ...

   and it's used all over the place (including quite a few places where
   we currently have sigprocmask(SIG_SETMASK, set, NULL), which is what
   it's equivalent to).  With that done, m32r doesn't use _BLOCKABLE
   anywhere, so it got removed.  And that chunk got picked when I'd been
   reordering the queue to pull the arch-specific fixes in front.
   Sorry."

Signed-off-by: Kyle McMartin <kyle@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agotypes.h: define __aligned_u64 and expose to userspace
Eric Paris [Fri, 15 Oct 2010 21:34:14 +0000 (14:34 -0700)]
types.h: define __aligned_u64 and expose to userspace

We currently have a kernel internal type called aligned_u64 which aligns
__u64's on 8 bytes boundaries even on systems which would normally align
them on 4 byte boundaries.  This patch creates a new type __aligned_u64
which does the same thing but which is exposed to userspace rather than
being kernel internal.

[akpm: merge early as both the net and audit trees want this]

[akpm@linux-foundation.org: enhance the comment describing the reasons for using aligned_u64.  Via Andreas and Andi.]
Based-on-patch-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Eric Paris <eparis@redhat.com>
Cc: Jan Engelhardt <jengelh@medozas.de>
Cc: David Miller <davem@davemloft.net>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agouml: fix build
FUJITA Tomonori [Fri, 15 Oct 2010 21:34:13 +0000 (14:34 -0700)]
uml: fix build

Fix a build error introduced by d6d1b650ae6acce73d55dd024 ("param: simple
locking for sysfs-writable charp parameters").

    CC      arch/um/kernel/trap.o
  arch/um/drivers/hostaudio_kern.c: In function 'hostaudio_open':
  arch/um/drivers/hostaudio_kern.c:204: error: '__param_dsp' undeclared (first use in this function)
  arch/um/drivers/hostaudio_kern.c:204: error: (Each undeclared identifier is reported only once
  arch/um/drivers/hostaudio_kern.c:204: error: for each function it appears in.)
  arch/um/drivers/hostaudio_kern.c: In function 'hostmixer_open_mixdev':
  arch/um/drivers/hostaudio_kern.c:265: error: '__param_mixer' undeclared (first use in this function)
  arch/um/drivers/hostaudio_kern.c:272: error: '__param_dsp' undeclared (first use in this function)

Reported-by: Toralf Förster <toralf.foerster@gmx.de>
Tested-by: Toralf Förster <toralf.foerster@gmx.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Takashi Iwai <tiwai@suse.de>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agosysctl: min/max bounds are optional
Eric Dumazet [Fri, 15 Oct 2010 21:34:12 +0000 (14:34 -0700)]
sysctl: min/max bounds are optional

sysctl check complains with a WARN() when proc_doulongvec_minmax() or
proc_doulongvec_ms_jiffies_minmax() are used by a vector of longs (with
more than one element), with no min or max value specified.

This is unexpected, given we had a bug on this min/max handling :)

Reported-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: David Miller <davem@davemloft.net>
Acked-by: WANG Cong <xiyou.wangcong@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc
Linus Torvalds [Fri, 15 Oct 2010 17:18:36 +0000 (10:18 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/cjb/mmc

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
  mmc: sdio: fix SDIO suspend/resume regression

14 years agommc: sdio: fix SDIO suspend/resume regression
Ohad Ben-Cohen [Wed, 13 Oct 2010 07:31:56 +0000 (09:31 +0200)]
mmc: sdio: fix SDIO suspend/resume regression

Fix SDIO suspend/resume regression introduced by 4c2ef25fe0b "mmc: fix
all hangs related to mmc/sd card insert/removal during suspend/resume":

  PM: Syncing filesystems ... done.
  Freezing user space processes ... (elapsed 0.01 seconds) done.
  Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done.
  Suspending console(s) (use no_console_suspend to debug)
  pm_op(): platform_pm_suspend+0x0/0x5c returns -38
  PM: Device pxa2xx-mci.0 failed to suspend: error -38
  PM: Some devices failed to suspend

4c2ef25fe0b moved the card removal/insertion mechanism out of MMC's
suspend/resume path and into pm notifiers (mmc_pm_notify), and that
broke SDIO's expectation that mmc_suspend_host() will remove the card,
and squash the error, in case -ENOSYS is returned from the bus suspend
handler (mmc_sdio_suspend() in this case).

mmc_sdio_suspend() is using this whenever at least one of the card's SDIO
function drivers does not have suspend/resume handlers - in that case
it is agreed to force removal of the entire card.

This patch fixes this regression by trivially bringing back that part of
mmc_suspend_host(), which was removed by 4c2ef25fe0b.

Reported-and-tested-by: Sven Neumann <s.neumann@raumfeld.com>
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Cc: Maxim Levitsky <maximlevitsky@gmail.com>
Cc: <stable@kernel.org>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Chris Ball <cjb@laptop.org>
14 years agoMerge branch 'timers-for-linus-urgent' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 15 Oct 2010 16:50:03 +0000 (09:50 -0700)]
Merge branch 'timers-for-linus-urgent' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'timers-for-linus-urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  hrtimer: Preserve timer state in remove_hrtimer()

14 years agoMerge branch 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Fri, 15 Oct 2010 16:49:43 +0000 (09:49 -0700)]
Merge branch 'upstream-fixes' of git://git./linux/kernel/git/jikos/hid

* 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: Add Cando touch screen 15.6-inch product id
  HID: Add MULTI_INPUT quirk for turbox/mosart touchscreen
  HID: hidraw, fix a NULL pointer dereference in hidraw_write
  HID: hidraw, fix a NULL pointer dereference in hidraw_ioctl

14 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Fri, 15 Oct 2010 16:49:16 +0000 (09:49 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  ubd: fix incorrect sector handling during request restart
  ps3disk: passing wrong variable to bvec_kunmap_irq()

14 years agoubd: fix incorrect sector handling during request restart
Tejun Heo [Fri, 15 Oct 2010 10:56:21 +0000 (12:56 +0200)]
ubd: fix incorrect sector handling during request restart

Commit f81f2f7c (ubd: drop unnecessary rq->sector manipulation)
dropped request->sector manipulation in preparation for global request
handling cleanup; unfortunately, it incorrectly assumed that the
updated sector wasn't being used.

ubd tries to issue as many requests as possible to io_thread.  When
issuing fails due to memory pressure or other reasons, the device is
put on the restart list and issuing stops.  On IO completion, devices
on the restart list are scanned and IO issuing is restarted.

ubd issues IOs sg-by-sg and issuing can be stopped in the middle of a
request, so each device on the restart queue needs to remember where
to restart in its current request.  ubd needs to keep track of the
issue position itself because,

* blk_rq_pos(req) is now updated by the block layer to keep track of
  _completion_ position.

* Multiple io_req's for the current request may be in flight, so it's
  difficult to tell where blk_rq_pos(req) currently is.

Add ubd->rq_pos to keep track of the issue position and use it to
correctly restart io_req issue.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Richard Weinberger <richard@nod.at>
Tested-by: Richard Weinberger <richard@nod.at>
Tested-by: Chris Frey <cdfrey@foursquare.net>
Cc: stable@kernel.org
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
14 years agoExport dump_{write,seek} to binary loader modules
Linus Torvalds [Fri, 15 Oct 2010 02:15:28 +0000 (19:15 -0700)]
Export dump_{write,seek} to binary loader modules

If you build aout support as a module, you'll want these exported.

Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoLinux 2.6.36-rc8
Linus Torvalds [Thu, 14 Oct 2010 23:26:43 +0000 (16:26 -0700)]
Linux 2.6.36-rc8

14 years agoUn-inline the core-dump helper functions
Linus Torvalds [Thu, 14 Oct 2010 21:32:06 +0000 (14:32 -0700)]
Un-inline the core-dump helper functions

Tony Luck reports that the addition of the access_ok() check in commit
0eead9ab41da ("Don't dump task struct in a.out core-dumps") broke the
ia64 compile due to missing the necessary header file includes.

Rather than add yet another include (<asm/unistd.h>) to make everything
happy, just uninline the silly core dump helper functions and move the
bodies to fs/exec.c where they make a lot more sense.

dump_seek() in particular was too big to be an inline function anyway,
and none of them are in any way performance-critical.  And we really
don't need to mess up our include file headers more than they already
are.

Reported-and-tested-by: Tony Luck <tony.luck@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Thu, 14 Oct 2010 18:19:44 +0000 (11:19 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  ehea: Fix a checksum issue on the receive path
  net: allow FEC driver to use fixed PHY support
  tg3: restore rx_dropped accounting
  b44: fix carrier detection on bind
  net: clear heap allocations for privileged ethtool actions
  NET: wimax, fix use after free
  ATM: iphase, remove sleep-inside-atomic
  ATM: mpc, fix use after free
  ATM: solos-pci, remove use after free
  net/fec: carrier off initially to avoid root mount failure
  r8169: use device model DMA API
  r8169: allocate with GFP_KERNEL flag when able to sleep

14 years agoDon't dump task struct in a.out core-dumps
Linus Torvalds [Thu, 14 Oct 2010 17:57:40 +0000 (10:57 -0700)]
Don't dump task struct in a.out core-dumps

akiphie points out that a.out core-dumps have that odd task struct
dumping that was never used and was never really a good idea (it goes
back into the mists of history, probably the original core-dumping
code).  Just remove it.

Also do the access_ok() check on dump_write().  It probably doesn't
matter (since normal filesystems all seem to do it anyway), but he
points out that it's normally done by the VFS layer, so ...

[ I suspect that we should possibly do "vfs_write()" instead of
  calling ->write directly.  That also does the whole fsnotify and write
  statistics thing, which may or may not be a good idea. ]

And just to be anal, do this all for the x86-64 32-bit a.out emulation
code too, even though it's not enabled (and won't currently even
compile)

Reported-by: akiphie <akiphie@lavabit.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agohrtimer: Preserve timer state in remove_hrtimer()
Salman Qazi [Tue, 12 Oct 2010 14:25:19 +0000 (07:25 -0700)]
hrtimer: Preserve timer state in remove_hrtimer()

The race is described as follows:

CPU X                                 CPU Y
remove_hrtimer
// state & QUEUED == 0
timer->state = CALLBACK
unlock timer base
timer->f(n) //very long
                                  hrtimer_start
                                    lock timer base
                                    remove_hrtimer // no effect
                                    hrtimer_enqueue
                                    timer->state = CALLBACK |
                                                   QUEUED
                                    unlock timer base
                                  hrtimer_start
                                    lock timer base
                                    remove_hrtimer
                                        mode = INACTIVE
                                        // CALLBACK bit lost!
                                    switch_hrtimer_base
                                            CALLBACK bit not set:
                                                    timer->base
                                                    changes to a
                                                    different CPU.
lock this CPU's timer base

The bug was introduced with commit ca109491f (hrtimer: removing all ur
callback modes) in 2.6.29

[ tglx: Feed new state via local variable and add a comment. ]

Signed-off-by: Salman Qazi <sqazi@google.com>
Cc: akpm@linux-foundation.org
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20101012142351.8485.21823.stgit@dungbeetle.mtv.corp.google.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@kernel.org
14 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx
Linus Torvalds [Wed, 13 Oct 2010 23:51:59 +0000 (16:51 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/djbw/async_tx

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx:
  ioat2: fix performance regression

14 years agoMerge branch 'for-2.6.36' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Wed, 13 Oct 2010 23:51:29 +0000 (16:51 -0700)]
Merge branch 'for-2.6.36' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.36' of git://linux-nfs.org/~bfields/linux:
  nfsd: fix BUG at fs/nfsd/nfsfh.h:199 on unlink

14 years agoMerge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 13 Oct 2010 23:50:23 +0000 (16:50 -0700)]
Merge branch 'perf-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  ring-buffer: Fix typo of time extends per page
  perf, MIPS: Support cross compiling of tools/perf for MIPS
  perf: Fix incorrect copy_from_user() usage

14 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Wed, 13 Oct 2010 23:35:33 +0000 (16:35 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm:
  ARM: relax ioremap prohibition (309caa9) for -final and -stable
  ARM: 6440/1: ep93xx: DMA: fix channel_disable
  cpuimx27: fix i2c bus selection
  cpuimx27: fix compile when ULPI is selected
  ARM: 6435/1: Fix HWCAP_TLS flag for ARM11MPCore/Cortex-A9
  ARM: 6436/1: AT91: Fix power-saving in idle-mode on 926T processors
  ARM: fix section mismatch warnings in Versatile Express
  ARM: 6412/1: kprobes-decode: add support for MOVW instruction
  ARM: 6419/1: mmu: Fix MT_MEMORY and MT_MEMORY_NONCACHED pte flags
  ARM: 6416/1: errata: faulty hazard checking in the Store Buffer may lead to data corruption

14 years agoMerge branch 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 13 Oct 2010 23:35:05 +0000 (16:35 -0700)]
Merge branch 'omap-fixes-for-linus' of git://git./linux/kernel/git/tmlind/linux-omap-2.6

* 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6:
  omap: iommu-load cam register before flushing the entry

14 years agoMerge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Wed, 13 Oct 2010 23:34:46 +0000 (16:34 -0700)]
Merge branch 'drm-fixes' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm/radeon/kms: Silent spurious error message
  drm/radeon/kms: fix bad cast/shift in evergreen.c
  drm/radeon/kms: make TV/DFP table info less verbose
  drm/radeon/kms: leave certain CP int bits enabled
  drm/radeon/kms: avoid corner case issue with unmappable vram V2

14 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 13 Oct 2010 23:34:23 +0000 (16:34 -0700)]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86, numa: For each node, register the memory blocks actually used
  x86, AMD, MCE thresholding: Fix the MCi_MISCj iteration order
  x86, mce, therm_throt.c: Fix missing curly braces in error handling logic

14 years agoioat2: fix performance regression
Dan Williams [Wed, 13 Oct 2010 22:43:10 +0000 (15:43 -0700)]
ioat2: fix performance regression

Commit 0793448 "DMAENGINE: generic channel status v2" changed the interface for
how dma channel progress is retrieved.  It inadvertently exported an internal
helper function ioat_tx_status() instead of ioat_dma_tx_status().  The latter
polls the hardware to get the latest completion state, while the helper just
evaluates the current state without touching hardware.  The effect is that we
end up waiting for completion timeouts or descriptor allocation errors before
the completion state is updated.

iperf (before fix):
[SUM]  0.0-41.3 sec   364 MBytes  73.9 Mbits/sec

iperf (after fix):
[SUM]  0.0- 4.5 sec   499 MBytes   940 Mbits/sec

This is a regression starting with 2.6.35.

Cc: <stable@kernel.org>
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Cc: Linus Walleij <linus.walleij@stericsson.com>
Cc: Maciej Sosnowski <maciej.sosnowski@intel.com>
Reported-by: Richard Scobie <richard@sauce.co.nz>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
14 years agoehea: Fix a checksum issue on the receive path
Breno Leitao [Thu, 7 Oct 2010 13:17:33 +0000 (13:17 +0000)]
ehea: Fix a checksum issue on the receive path

Currently we set all skbs with CHECKSUM_UNNECESSARY, even
those whose protocol we don't know. This patch just
add the CHECKSUM_COMPLETE tag for non TCP/UDP packets.

Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Breno Leitao <leitao@linux.vnet.ibm.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agonfsd: fix BUG at fs/nfsd/nfsfh.h:199 on unlink
J. Bruce Fields [Wed, 13 Oct 2010 18:46:17 +0000 (14:46 -0400)]
nfsd: fix BUG at fs/nfsd/nfsfh.h:199 on unlink

As of commit 43a9aa64a2f4330a9cb59aaf5c5636566bce067c "NFSD:
Fill in WCC data for REMOVE, RMDIR, MKNOD, and MKDIR", we sometimes call
fh_unlock on a filehandle that isn't fully initialized.

We should fix up the callers, but as a quick fix it is also sufficient
just to remove this assertion.

Reported-by: Marius Tolzmann <tolzmann@molgen.mpg.de>
Cc: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
14 years agonet: allow FEC driver to use fixed PHY support
Greg Ungerer [Mon, 11 Oct 2010 21:03:05 +0000 (21:03 +0000)]
net: allow FEC driver to use fixed PHY support

At least one board using the FEC driver does not have a conventional
PHY attached to it, it is directly connected to a somewhat simple
ethernet switch (the board is the SnapGear/LITE, and the attached
4-port ethernet switch is a RealTek RTL8305). This switch does not
present the usual register interface of a PHY, it presents nothing.
So a PHY scan will find nothing - it finds ID's of 0 for each PHY
on the attached MII bus.

After the FEC driver was changed to use phylib for supporting PHYs
it no longer works on this particular board/switch setup.

Add code support to use a fixed phy if no PHY is found on the MII bus.
This is based on the way the cpmac.c driver solved this same problem.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years agoHID: Add Cando touch screen 15.6-inch product id
François Jaouen [Wed, 13 Oct 2010 08:47:32 +0000 (10:47 +0200)]
HID: Add Cando touch screen 15.6-inch product id

This add the product id of the touch screen found on ACER Aspire 5738PZ.  Works
with hid-cando driver.

Signed-off-by: Francois Jaouen<francois.jaouen@laposte.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
14 years agoARM: relax ioremap prohibition (309caa9) for -final and -stable
Russell King [Tue, 12 Oct 2010 23:15:25 +0000 (00:15 +0100)]
ARM: relax ioremap prohibition (309caa9) for -final and -stable

... but produce a big warning about the problem as encouragement
for people to fix their drivers.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>