Yet another regression due to
commit 135c81b8c3
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun Jul 21 21:37:09 2013 +0200
drm/i915: clean up crtc timings computation
I'm starting to wonder whether this was worth it ...
v2: Actually make it compile.
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We've failed to properly clear out the flags when converting a dtd to
a drm mode. For more paranoia just memset the entire structure (and
drop the now redundant clears).
Also since
commit 135c81b8c3
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date: Sun Jul 21 21:37:09 2013 +0200
drm/i915: clean up crtc timings computation
we don't update the crtc timings any more properly, so do that again.
v2: Remove more redundant clearing, spotted by Ville.
v3: Actually make it compile. Oops.
v4: Use a temporary structure to fill in the mode and copy it over
with drm_mode_copy. This will ensure we don't clobber the mode list or
id. Suggested by Ville.
Cc: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
[danvet: Use the = {}; structure clearing instead of memset as
suggested by Ville.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Occasionally we seem to miss an IRQ from the ME (microengine). I'm not
entirely sure the root cause, but for now we can unwedge things by
retiring from the hangcheck timer.
Signed-off-by: Rob Clark <robdclark@gmail.com>
When we CPU_PREP a bo with NOSYNC flag (for example, to implement
PIPE_TRANSFER_DISCARD_WHOLE_RESOURCE), an -EBUSY return indicates to
userspace that the bo is still busy. Previously it was incorrectly
returning 0 in this case.
And while we're in there throw in an bit of extra sanity checking in
case userspace tries to wait for a bogus fence.
Signed-off-by: Rob Clark <robdclark@gmail.com>
In case of error, the function drm_prime_pages_to_sg() returns ERR_PTR()
and never returns NULL. The NULL test in the return value check should
be replaced with IS_ERR().
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
bapm is a pm feature for sharing the power budget between
the GPU and the CPU on APUs. It needs to be enabled or
disabled in certain circumstances. For now, disable it
when on battery and enable it when on AC power.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
bapm is a power management feature for handling the
power budget between the CPU and GPU on APUs. This
patch adds support for enabling or disabling it.
For now disable it by default. Enabling it properly
requires quite a bit more work and will be addressed
in a separate patch.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
bapm is a power management feature for handling the
power budget between the CPU and GPU on APUs. This
patch adds support for enabling or disabling it.
For now disable it by default. Enabling it properly
requires quite a bit more work and will be addressed
in a separate patch.
This patch fixes hangs on boot on certain trinity
laptops when the system is on battery power.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
The hwmon sysfs interface allows exposing temperature limits. The "max"
and "min" thresholds will be exposed as a critical high limit and its
hysteresis value, respectively. This gives the user a better idea of how
well cooling is doing and whether it is sufficient.
Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
You can get the driver data from struct device directly, there's no
need to get the PCI device first.
Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The OUTPUT_ENABLE action jumps past the point in the coder where
the data_offset is set on certain rs780 cards. This worked
previously because the OUTPUT_ENABLE action is always called
immediately after the ENABLE action so the data_offset remained
set. In 6f8bbaf568c7f2c497558bfd04654c0b9841ad57
(drm/radeon/atom: initialize more atom interpretor elements to 0),
we explictly reset data_offset to 0 between atom calls which then
caused this to fail. The fix is to just skip calling the
OUTPUT_ENABLE action on the problematic chipsets. The ENABLE
action does the same thing and more. Ultimately, we could
probably drop the OUTPUT_ENABLE action all together on DCE3
asics.
fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=60791
v2: only rs880 seems to be affected
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Some older 6xx-7xx boards didn't always fill in the
UVD clocks properly in the UVD power states. This
leads to the driver trying to set a 0 clock which
results in slow or broken UVD playback.
Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=69120
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Neither complete nor perfect, but solves my problem at hand
and might be useful in the future.
Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Certain r6xx boards use the same power state for both UVD
and other things. Since we don't support UVD on r6xx boards
at the moment, there was no callback installed for setting
the UVD clocks, however, on systems that use the same power
state, this leads to a NULL pointer dereference. Fill
in a stubbed out implementation for now to avoid the crash.
Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=66963
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: "3.11" <stable@vger.kernel.org>
This adds spinlocks to protect access to other
indirect register apertures. These indirect spaces are
used pretty infrequently and we haven't had an reported
problems, but better safe than sorry.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
smc registers are access indirectly via the main mmio aperture, so
there may be problems with concurrent access. This adds a spinlock
to protect access to this register space.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This updates dpm support for KV asics. Notably there
are some changes in acp handling and forcing performance
levels.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The problem here is that "unsigned i" is always greater than or equal to
zero. These loops mostly have a second check for "(i == 0)" so only the
last two are actually buggy. The rest is just cleanup.
Bug 1: kv_force_dpm_highest() doesn't have an "(i == 0)" check so it's
a potential forever loop.
Bug 2: In kv_get_sleep_divider_id_from_clock() there is a typo and the
test is reversed "<=" vs ">" so we never enter the loop. That means
normally we return KV_MAX_DEEPSLEEP_DIVIDER_ID (5). The return value
from here is saved in ->DeepSleepDivId and I wasn't able to determine
how that is used. This is a static checker fix and I have not tested
it.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
kfree() can accept NULL pointers so I have removed the checks. Also
I've used a pointer to shorten the lines.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This provides a connector property to enable/disable hdmi
audio on the fly. The default is disabled, but you can select
auto (let the driver detect an audio capable monitor and enable it)
or enabled (force audio enabled). This also enables audio by
default so you no longer need a module parameter to enable audio.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Make sure the audio pin is valid before accessing its members.
Noticed by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Allows you to limit the selected power levels via sysfs.
Force the feedback divider to select a power level.
v2: fix checking in rs780_force_fbdiv,
drop a duplicate divider structure in rs780_dpm_force_performance_level,
Force the voltage level too.
Signed-off-by: Anthoine Bourgeois <anthoine.bourgeois@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The internal cp interrupts need to be enabled and
disabled at specific times in order clockgating to
work properly. This patch changes the handling
of the CP_INT_CNTL register to respect the current
state of the internal CP interrupts when making
changes to the other interrupts in CP_INT_CNTL.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The internal cp interrupts need to be enabled and
disabled at specific times in order clockgating to
work properly. This patch changes the handling
of the CP_INT_CNTL register to respect the current
state of the internal CP interrupts when making
changes to the other interrupts in CP_INT_CNTL.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Need to check size+offset against bo size (duh!).. now we have a test
case to make sure I've done it right:
https://github.com/freedreno/msmtest/blob/master/submittest.c
Also, use DRM_ERROR() for error case traces, which makes debugging
userspace easier when enabling debug traces is too much.
Signed-off-by: Rob Clark <robdclark@gmail.com>
If gpu locks up with the rptr shortly beyond the wrap-around point in
the ringbuffer, because the rptr was not reset (but wptr is, by virtue
of resetting rb->cur), we could end up in a scenario where we think
there is not enough space in the ringbuffer for the next cmds. And
since the CP won't reset rptr until after processing an IB, this leaves
things in a sort of deadlock.
So reset rptr too. And a bit more spiffing up of hangcheck to make
things easier to debug.
Signed-off-by: Rob Clark <robdclark@gmail.com>
The userspace API already had everything needed to handle read vs write
synchronization. This patch actually bothers to hook it up properly, so
that we don't need to (for example) stall on userspace read access to a
buffer that gpu is also still reading.
Signed-off-by: Rob Clark <robdclark@gmail.com>
Instead of just a flag bit for each of the positive/negative sync
modes drm actually uses a separate flag for each ... This upsets the
modeset checker since the adjusted mode filled out at modeset time
doesn't match the one reconstructed at check time (since the
->get_config callback already gets this right).
Reported-by: Knut Petersen <Knut_Petersen@t-online.de>
Cc: Knut Petersen <Knut_Petersen@t-online.de>
References: http://www.gossamer-threads.com/lists/linux/kernel/1778688?do=post_view_threaded
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Replace "%8x" with "%08x".
The hex number should be shown with zero stuffed instead of spaces.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
if we have no crtcs we need to not call the display resume code.
Reported-by: Tobias Klausmann <tobias.johannes.klausmann@mni.thm.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>
- Early stolen mem reservation from Jesse in x86 boot code. Acked by Ingo
and hpa. This was ready much earlier but somehow I've thought it'd go
in through x86 trees, hence why this is late. Avoids the pci resource
code to plant mmiobars in the middle of stolen mem and other ugliness.
- vgaarb improvements from Alex Williamson plus the fix from Ville for the
vgacon->fbcon smooth transition "feature".
- Render pageflips on ivb/hsw to avoid stalls due to the ring switching
when only flipping on the blitter (Chris).
- Deadlock fixes around our flush_workqueue which crept back in - lockdep
isn't clever enough :(
- Shrinker recursion fix from Chris - this is the thing that blew the vma
patches from Ben I've taken out of 3.12.
- Fixup for the relocation refactoring. Also an igt testcase to make sure
we don't break this again.
- Pile of smaller fixups all over, shortlog has full details.
* tag 'drm-intel-fixes-2013-09-06' of git://people.freedesktop.org/~danvet/drm-intel: (29 commits)
drm/i915: Delay disabling of VGA memory until vgacon->fbcon handoff is done
drm/i915: try not to lose backlight CBLV precision
drm/i915: Confine page flips to BCS on Valleyview
drm/i915: Skip stolen region initialisation if none is reserved
drm/i915: fix gpu hang vs. flip stall deadlocks
drm/i915: Hold an object reference whilst we shrink it
drm/i915: fix i9xx_crtc_clock_get for multiplied pixels
drm/i915: handle sdvo input pixel multiplier correctly again
drm/i915: fix hpd work vs. flush_work in the pageflip code deadlock
drm/i915: fix up the relocate_entry refactoring
drm/i915: Fix pipe config warnings when dealing with LVDS fixed mode
drm/i915: Don't call sg_free_table() if sg_alloc_table() fails
i915: Update VGA arbiter support for newer devices
vgaarb: Fix VGA decodes changes
vgaarb: Don't disable resources that are not owned
drm/i915: Pin pages whilst mapping the dma-buf
drm/i915: enable trickle feed on Haswell
x86: add early quirk for reserving Intel graphics stolen memory v5
drm/i915: split PCI IDs out into i915_drm.h v4
i915_gem: Convert kmem_cache_alloc(...GFP_ZERO) to kmem_cache_zalloc
...
My g33 here seems to be shockingly good at hitting them all. This time
around kms_flip/flip-vs-panning-vs-hang blows up:
intel_crtc_wait_for_pending_flips correctly checks for gpu hangs and
if a gpu hang is pending aborts the wait for outstanding flips so that
the setcrtc call will succeed and release the crtc mutex. And the gpu
hang handler needs that lock in intel_display_handle_reset to be able
to complete outstanding flips.
The problem is that we can race in two ways:
- Waiters on the dev_priv->pending_flip_queue aren't woken up after
we've the reset as pending, but before we actually start the reset
work. This means that the waiter doesn't notice the pending reset
and hence will keep on hogging the locks.
Like with dev->struct_mutex and the ring->irq_queue wait queues we
there need to wake up everyone that potentially holds a lock which
the reset handler needs.
- intel_display_handle_reset was called _after_ we've already
signalled the completion of the reset work. Which means a waiter
could sneak in, grab the lock and never release it (since the
pageflips won't ever get released).
Similar to resetting the gem state all the reset work must complete
before we update the reset counter. Contrary to the gem reset we
don't need to have a second explicit wake up call since that will
have happened already when completing the pageflips. We also don't
have any issues that the completion happens while the reset state is
still pending - wait_for_pending_flips is only there to ensure we
display the right frame. After a gpu hang&reset events such
guarantees are out the window anyway. This is in contrast to the gem
code where too-early wake-up would result in unnecessary restarting
of ioctls.
Also, since we've gotten these various deadlocks and ordering
constraints wrong so often throw copious amounts of comments at the
code.
This deadlock regression has been introduced in the commit which added
the pageflip reset logic to the gpu hang work:
commit 96a02917a0
Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date: Mon Feb 18 19:08:49 2013 +0200
drm/i915: Finish page flips and update primary planes after a GPU reset
v2:
- Add comments to explain how the wake_up serves as memory barriers
for the atomic_t reset counter.
- Improve the comments a bit as suggested by Chris Wilson.
- Extract the wake_up calls before/after the reset into a little
i915_error_wake_up and unconditionally wake up the
pending_flip_queue waiters, again as suggested by Chris Wilson.
v3: Throw copious amounts of comments at i915_error_wake_up as
suggested by Chris Wilson.
Cc: stable@vger.kernel.org
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Detangle the additional state of whether or not the hw has the pfit
enabled from whether it has zero size. This allows us to cleanly
distinguish in the code when we expect the pfit to be enabled (for
Haswell pc8), and when the BIOS is confused and needs sanitizing.
Reported-by: shui yanwei <yangweix.shui@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68251
Tested-by: shui yanwei <yangweix.shui@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
When transitioning away from vgacon the system tries to save the
current contents of the VGA memory, so that it can be cleanly handed
off to fbcon (or whatever comes afterwards).
The recent change
commit 81b5c7bc8d
Author: Alex Williamson <alex.williamson@redhat.com>
Date: Wed Aug 28 09:39:08 2013 -0600
i915: Update VGA arbiter support for newer devices
caused i915 to disable VGA memory decode for the IGD when i915 is
initializing. Unfortunately that happens before the vgacon->fbcon
handoff so vgacon_save_screen() will read out all ones from the
VGA memory.
After the handoff fbcon will inherit the bogus state from vgacon,
and pre-fills the fb with matching contents. The end result is
a white rectangle in the top left corner of the screen, the size
of which matches the now inactive VGA console.
To remedy the situation delay the disabling of VGA memory until
the vgacon->fbcon handoff has happened.
Also rename i915_enable_vga to i915_enable_vga_mem to make
the relationship between these functions clearer.
Cc: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
ACPI has _BCM and _BQC methods to set and query the backlight
brightness, respectively. The ACPI opregion has variables BCLP and CBLV
to hold the requested and current backlight brightness, respectively.
The BCLP variable has range 0..255 while the others have range
0..100. This means the _BCM method has to scale the brightness for BCLP,
and the gfx driver has to scale the requested value back for CBLV. If
the _BQC method uses the CBLV variable (apparently some implementations
do, some don't) for current backlight level reporting, there's room for
rounding errors.
Use DIV_ROUND_UP for scaling back to CBLV to get back to the same values
that were passed to _BCM, presuming the _BCM simply uses bclp = (in *
255) / 100 for scaling to BCLP.
Reference: https://gist.github.com/aaronlu/6314920
Reported-by: Aaron Lu <aaron.lu@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Aaron Lu <aaron.lu@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Once again we find that Valleyview is ever so subtlety different from
the rest of its gen7 brethen. In this case, Valleyview has no support
for pageflipping from the RCS ring.
Fixes a regression from
commit ffe74d7550
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Mon Aug 26 20:58:12 2013 +0100
drm/i915: Use RCS flips on Ivybridge+
Reported-by: "Lee, Chon Ming" <chon.ming.lee@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68968
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Paulo reported that if he set the amount of reserved memory to 0, then
we emitted a warning about a conflict before disabling our use of stolen
memory. This was introduced with
commit eaba1b8f33
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date: Thu Jul 4 12:28:35 2013 +0100
drm/i915: Verify that our stolen memory doesn't conflict
and is simply fixed by checking for a no reservation first.
Reported-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Since we've started to clean up pending flips when the gpu hangs in
commit 96a02917a0
Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date: Mon Feb 18 19:08:49 2013 +0200
drm/i915: Finish page flips and update primary planes after a GPU reset
the gpu reset work now also grabs modeset locks. But since work items
on our private work queue are not allowed to do that due to the
flush_workqueue from the pageflip code this results in a neat
deadlock:
INFO: task kms_flip:14676 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
kms_flip D ffff88019283a5c0 0 14676 13344 0x00000004
ffff88018e62dbf8 0000000000000046 ffff88013bdb12e0 ffff88018e62dfd8
ffff88018e62dfd8 00000000001d3b00 ffff88019283a5c0 ffff88018ec21000
ffff88018f693f00 ffff88018eece000 ffff88018e62dd60 ffff88018eece898
Call Trace:
[<ffffffff8138ee7b>] schedule+0x60/0x62
[<ffffffffa046c0dd>] intel_crtc_wait_for_pending_flips+0xb2/0x114 [i915]
[<ffffffff81050ff4>] ? finish_wait+0x60/0x60
[<ffffffffa0478041>] intel_crtc_set_config+0x7f3/0x81e [i915]
[<ffffffffa031780a>] drm_mode_set_config_internal+0x4f/0xc6 [drm]
[<ffffffffa0319cf3>] drm_mode_setcrtc+0x44d/0x4f9 [drm]
[<ffffffff810e44da>] ? might_fault+0x38/0x86
[<ffffffffa030d51f>] drm_ioctl+0x2f9/0x447 [drm]
[<ffffffff8107a722>] ? trace_hardirqs_off+0xd/0xf
[<ffffffffa03198a6>] ? drm_mode_setplane+0x343/0x343 [drm]
[<ffffffff8112222f>] ? mntput_no_expire+0x3e/0x13d
[<ffffffff81117f33>] vfs_ioctl+0x18/0x34
[<ffffffff81118776>] do_vfs_ioctl+0x396/0x454
[<ffffffff81396b37>] ? sysret_check+0x1b/0x56
[<ffffffff81118886>] SyS_ioctl+0x52/0x7d
[<ffffffff81396b12>] system_call_fastpath+0x16/0x1b
2 locks held by kms_flip/14676:
#0: (&dev->mode_config.mutex){+.+.+.}, at: [<ffffffffa0316545>] drm_modeset_lock_all+0x22/0x59 [drm]
#1: (&crtc->mutex){+.+.+.}, at: [<ffffffffa031656b>] drm_modeset_lock_all+0x48/0x59 [drm]
INFO: task kworker/u8:4:175 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
kworker/u8:4 D ffff88018de9a5c0 0 175 2 0x00000000
Workqueue: i915 i915_error_work_func [i915]
ffff88018e37dc30 0000000000000046 ffff8801938ab8a0 ffff88018e37dfd8
ffff88018e37dfd8 00000000001d3b00 ffff88018de9a5c0 ffff88018ec21018
0000000000000246 ffff88018e37dca0 000000005a865a86 ffff88018de9a5c0
Call Trace:
[<ffffffff8138ee7b>] schedule+0x60/0x62
[<ffffffff8138f23d>] schedule_preempt_disabled+0x9/0xb
[<ffffffff8138d0cd>] mutex_lock_nested+0x205/0x3b1
[<ffffffffa0477094>] ? intel_display_handle_reset+0x7e/0xbd [i915]
[<ffffffffa0477094>] ? intel_display_handle_reset+0x7e/0xbd [i915]
[<ffffffffa0477094>] intel_display_handle_reset+0x7e/0xbd [i915]
[<ffffffffa044e0a2>] i915_error_work_func+0x128/0x147 [i915]
[<ffffffff8104a89a>] process_one_work+0x1d4/0x35a
[<ffffffff8104a821>] ? process_one_work+0x15b/0x35a
[<ffffffff8104b4a5>] worker_thread+0x144/0x1f0
[<ffffffff8104b361>] ? rescuer_thread+0x275/0x275
[<ffffffff8105076d>] kthread+0xac/0xb4
[<ffffffff81059d30>] ? finish_task_switch+0x3b/0xc0
[<ffffffff810506c1>] ? __kthread_parkme+0x60/0x60
[<ffffffff81396a6c>] ret_from_fork+0x7c/0xb0
[<ffffffff810506c1>] ? __kthread_parkme+0x60/0x60
3 locks held by kworker/u8:4/175:
#0: (i915){.+.+.+}, at: [<ffffffff8104a821>] process_one_work+0x15b/0x35a
#1: ((&dev_priv->gpu_error.work)){+.+.+.}, at: [<ffffffff8104a821>] process_one_work+0x15b/0x35a
#2: (&crtc->mutex){+.+.+.}, at: [<ffffffffa0477094>] intel_display_handle_reset+0x7e/0xbd [i915]
This blew up while running kms_flip/flip-vs-panning-vs-hang-interruptible
on one of my older machines.
Unfortunately (despite the proper lockdep annotations for
flush_workqueue) lockdep still doesn't detect this correctly, so we
need to rely on chance to discover these bugs.
Apply the usual bugfix and schedule the reset work on the system
workqueue to keep our own driver workqueue free of any modeset lock
grabbing.
Note that this is not a terribly serious regression since before the
offending commit we'd simply have stalled userspace forever due to
failing to abort all outstanding pageflips.
v2: Add a comment as requested by Chris.
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Whilst running the shrinker, we need to hold a reference as we unbind
the objects, or else we may end up waiting for and retiring requests,
which in turn may result in this object being freed.
This is very similar to the eviction code which also has to be very
careful to keep a reference to its objects as it retires and unbinds
them.
Another similarity, that Ben pointed out, is that as we may call
retire-requests, the unbound_list is outside of our control. We must
only process a single element of that list at a time, that is we can not
rely on the "safe" next pointer being valid after a call to
i915_vma_unbind().
BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
IP: [<ffffffffa0082892>] i915_gem_gtt_finish_object+0x68/0xbd [i915]
PGD 758d3067 PUD ac0d6067 PMD 0
Oops: 0000 [#1] SMP
Modules linked in: dm_mod snd_hda_codec_realtek iTCO_wdt iTCO_vendor_support pcspkr snd_hda_intel i2c_i801 snd_hda_codec snd_hwdep snd_pcm snd_page_alloc snd_timer snd lpc_ich mfd_core soundcore battery ac option usb_wwan usbserial uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core videodev i915 video button drm_kms_helper drm acpi_cpufreq mperf freq_table
CPU: 1 PID: 16835 Comm: fbo-maxsize Not tainted 3.11.0-rc7_nightlytop_8fdad4_20130902_+ #7977
task: ffff8800712106d0 ti: ffff880028e4a000 task.ti: ffff880028e4a000
RIP: 0010:[<ffffffffa0082892>] [<ffffffffa0082892>] i915_gem_gtt_finish_object+0x68/0xbd [i915]
RSP: 0018:ffff880028e4b9e8 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff880145734000 RCX: ffff880145735328
RDX: ffff8801457353fc RSI: 0000000000000000 RDI: ffff88007597cc00
RBP: ffff88007597cc00 R08: 0000000000000001 R09: ffff88014f257f00
R10: ffffea0001d65f00 R11: 0000000000bba60b R12: ffff880149e5b000
R13: ffff880145734001 R14: ffff88007597ccc8 R15: ffff88007597cc00
FS: 00007ff5bc919740(0000) GS:ffff88014f240000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000008 CR3: 0000000028f4c000 CR4: 00000000001407e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Stack:
0000000000000000 ffff88007597cc00 ffff8801440d6840 0000000000000000
ffff880145734000 ffffffffa007c854 0000000000000010 ffff88007597c900
0000000000018000 00000000004a1201 ffff88007597cc60 ffffffffa007d183
Call Trace:
[<ffffffffa007c854>] ? i915_vma_unbind+0xe2/0x1d1 [i915]
[<ffffffffa007d183>] ? __i915_gem_shrink+0xf1/0x162 [i915]
[<ffffffffa007d2ee>] ? i915_gem_object_get_pages_gtt+0xfa/0x303 [i915]
[<ffffffffa00795f4>] ? i915_gem_object_get_pages+0x54/0x89 [i915]
[<ffffffffa007cbda>] ? i915_gem_object_pin+0x238/0x5ce [i915]
[<ffffffff812cba5f>] ? __sg_page_iter_next+0x2b/0x58
[<ffffffffa0082056>] ? gen6_ppgtt_insert_entries+0xf2/0x114 [i915]
[<ffffffffa007fe4b>] ? i915_gem_execbuffer_reserve_vma.isra.13+0x79/0x18d [i915]
[<ffffffffa008017c>] ? i915_gem_execbuffer_reserve+0x21d/0x347 [i915]
[<ffffffffa0080bfb>] ? i915_gem_do_execbuffer.isra.17+0x4f3/0xe61 [i915]
[<ffffffffa00795f4>] ? i915_gem_object_get_pages+0x54/0x89 [i915]
[<ffffffffa007e405>] ? i915_gem_pwrite_ioctl+0x743/0x7a5 [i915]
[<ffffffffa0081a46>] ? i915_gem_execbuffer2+0x15e/0x1e4 [i915]
[<ffffffffa000e20d>] ? drm_ioctl+0x2a5/0x3c4 [drm]
[<ffffffffa00818e8>] ? i915_gem_execbuffer+0x37f/0x37f [i915]
[<ffffffff816f64c0>] ? __do_page_fault+0x3ab/0x449
[<ffffffff810be3da>] ? do_mmap_pgoff+0x2b2/0x341
[<ffffffff810e49be>] ? vfs_ioctl+0x1e/0x31
[<ffffffff810e5194>] ? do_vfs_ioctl+0x3ad/0x3ef
[<ffffffff810e5224>] ? SyS_ioctl+0x4e/0x7e
[<ffffffff816f88d2>] ? system_call_fastpath+0x16/0x1b
Code: 52 0c a0 48 c7 c6 22 30 0d a0 31 c0 e8 ef 00 f9 ff bf c6 a7 00 00 e8 90 5d 24 e1 f6 85 13 01 00 00 10 75 44 48 8b 85 18 01 00 00 <8b> 50 08 48 8b 30 49 8b 84 24 88 02 00 00 48 89 c7 48 81 c7 98
RIP [<ffffffffa0082892>] i915_gem_gtt_finish_object+0x68/0xbd [i915]
RSP <ffff880028e4b9e8>
CR2: 0000000000000008
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68171
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
[danvet: Bikeshed the comments a bit as discussed with Chris.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Summary:
- Consider fallback option to gem allocation fail
. try to allocate physically non-contiguous memory
if iommu is supported when physically contiguous memory allocation
failed.
- Add runtime pm support to g2d driver
- Add device tree support
. add device tree support to rotator driver, make fimd driver get
signal polarities from device tree.
- some fixups
. correct pixel format setting to fimd driver, and consider pixel
format checking to a particular window layer.
- some cleanups
. replace fb_videomode with videomode.
. remove non-DT support
* 'exynos-drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos: (21 commits)
drm/exynos: Fix build error with exynos_drm_connector.c
drm/exynos: Remove non-DT support in exynos_drm_fimd
drm/exynos: Remove non-DT support in exynos_hdmi
drm/exynos: Remove non-DT support in exynos_drm_g2d
drm/exynos: Remove non-DT support in exynos_hdmiphy
drm/exynos: Remove non-DT support in exynos_ddc
drm/exynos: Make Exynos DRM drivers depend on OF
drm/exynos: Consider fallback option to allocation fail
drm/exynos: fimd: move platform data parsing to separate function
drm/exynos: fimd: get signal polarities from device tree
drm/exynos: fimd: replace struct fb_videomode with videomode
drm/exynos: check a pixel format to a particular window layer
drm/exynos: fix fimd pixel format setting
drm/exynos: Add NULL pointer check
drm/exynos: Remove redundant error messages
drm/exynos: Add missing of.h header include
drm/exynos: Remove redundant NULL check in exynos_drm_buf
drm/exynos: add device tree support for rotator
drm/exynos: Add missing includes
drm/exynos: add runtime pm interfaces to g2d driver
...
exynos_drm_connector.c now uses videomode helper API. Hence select
VIDEOMODE_HELPERS at DRM_EXYNOS level itself instead of at
DRM_EXYNOS_FIMD to avoid the following compilation error when FIMD
is not selected (introduced by commit cf796235a6 "drm/exynos: fimd:
replace struct fb_videomode with videomode"):
drivers/built-in.o: In function `exynos_drm_connector_get_modes':
drivers/gpu/drm/exynos/exynos_drm_connector.c:86:
undefined reference to `drm_display_mode_from_videomode'
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Since commit 383ffda2fa ("ARM: EXYNOS: no more support non-DT
for EXYNOS SoCs"), Exynos platform is DT only. Hence remove
all the conditional macros and make the driver DT only.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Since commit 383ffda2fa ("ARM: EXYNOS: no more support non-DT
for EXYNOS SoCs"), Exynos platform is DT only. Hence remove
all the conditional macros and make the driver DT only.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Since commit 383ffda2fa ("ARM: EXYNOS: no more support non-DT
for EXYNOS SoCs"), Exynos platform is DT only. Hence remove
all the conditional macros and make the driver DT only.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Daniel Vetter