Bug 104162

Possibly:

commit 16c8619a7c53fe05526c31d4758be0eeabd16364
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Dec 19 22:09:16 2017 +0000

    drm/i915: Avoid context dereference inside execlists_submission_tasklet
    
    A lesson that has to be relearnt over and over again is that the request
    does not keep a reference to the context and so we cannot freely
    dereference the context from inside the execlists_submission_tasklet. In
    particular, we try to do so in the new GEM_TRACE() so convert those over
    to the port->context_id we keep for GEM debugging. This means the
    tracing now depends on DRM_I915_GEM_DEBUG.
    
    Fixes: bccd3b831185 ("drm/i915: Use trace_printk to provide a death rattle for GEM")
    References: https://bugs.freedesktop.org/show_bug.cgi?id=104066
    References: https://bugs.freedesktop.org/show_bug.cgi?id=104162
    References: https://bugs.freedesktop.org/show_bug.cgi?id=104242
    References: https://bugs.freedesktop.org/show_bug.cgi?id=104310
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
    Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
    Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
    Reviewed-by: Michel Thierry <michel.thierry@intel.com>
    Link: https://patchwork.freedesktop.org/patch/msgid/20171219220916.30882-1-chris@chris-wilson.co.uk

Unfortunately reproduced on:

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3602/fi-bdw-5557u/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-c.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3605/fi-bdw-5557u/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b.html

https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4141/fi-bdw-5557u/igt@gem_exec_suspend@basic-s4-devices.html

pstore is just a bunch of:
<0>[  308.200127] gem_exec-2999    1..s1 223027120us : execlists_submission_tasklet: rcs0 in[0]:  ctx=7.1, seqno=821

This is a Meta bug for incompletes on BDW.

https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4352/fi-bdw-5557u/igt@gem_exec_suspend@basic-s3.html

run.log:
pass: igt/gem_exec_suspend/basic-s3

[108/288] skip: 3, pass: 105 \
FATAL: command execution failed
...
Completed CI_IGT_test CI_DRM_3916/fi-bdw-5557u/0 : FAILURE
CI_IGT_test runtime 328 seconds
Rebooting fi-bdw-5557u

Last dmesg:
<7>[  194.262113] [IGT] gem_exec_suspend: executing
<4>[  194.271639] Setting dangerous option reset - tainting kernel
<7>[  194.275272] [IGT] gem_exec_suspend: starting subtest basic-S3
Followed by "stray"

pstore just shows a sysrq-trigger backtrace unfortunately the reset is overwritten by ftrace.

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_3/fi-bdw-5557u/igt@kms_cursor_crc@cursor-256x256-suspend.html

run.log:
running: igt/kms_cursor_crc/cursor-256x256-suspend

[74/97] skip: 37, pass: 37 -                      
FATAL: command execution failed
...
Completed CI_IGT_test drmtip_3/fi-bdw-5557u/27 : FAILURE
CI_IGT_test runtime 333 seconds
Rebooting fi-bdw-5557u

Last dmesg:
<7>[  136.677451] [drm:verify_connector_state.isra.78 [i915]] [CONNECTOR:76:HDMI-A-2]
<7>[  136.677522] [drm:intel_atomic_commit_tail [i915]] [CRTC:51:pipe B]
<7>[  136.677600] [drm:verify_single_dpll_state.isra.79 [i915]] WRPLL 1

pstore overwritten by ftrace

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_3/fi-bdw-5557u/igt@gem_ctx_isolation@bcs0-s3.html

run.log:
pass: igt/gem_ctx_isolation/bcs0-s3

[15/97] skip: 8, pass: 7 -
FATAL: command execution failed
...
Completed CI_IGT_test drmtip_3/fi-bdw-5557u/34 : FAILURE
CI_IGT_test runtime 240 seconds
Rebooting fi-bdw-5557u

last dmesg:
<4>[   40.493107] Setting dangerous option reset - tainting kernel
<7>[   40.497149] [IGT] gem_ctx_isolation: starting subtest bcs0-S3
<6>[   40.567914] PM: suspend entry (deep)

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_4/fi-bdw-gvtdvm/igt@gem_ctx_switch@basic-all-heavy.html

run.log:
running: igt/gem_ctx_switch/basic-all-heavy

[39/97] skip: 20, pass: 19 \               
FATAL: command execution failed
...
Completed CI_IGT_test drmtip_4/fi-bdw-gvtdvm/25 : FAILURE
CI_IGT_test runtime 600 seconds
Rebooting fi-bdw-gvtdvm

Last dmesg:
<7>[   87.178419] [IGT] gem_ctx_switch: starting subtest basic-all-heavy
<6>[  186.052216] perf: interrupt took too long (8296 > 8205), lowering kernel.perf_event_max_sample_rate to 24000
<6>[  346.791638] perf: interrupt took too long (10381 > 10370), lowering kernel.perf_event_max_sample_rate to 19000

https://intel-gfx-ci.01.org/tree/drm-tip/drmtip_3/fi-bdw-gvtdvm/igt@gem_exec_reuse@contexts.html

run.log:
running: igt/gem_exec_reuse/contexts

[06/97] skip: 2, pass: 3, fail: 1 - 
FATAL: command execution failed
...
Completed CI_IGT_test drmtip_3/fi-bdw-gvtdvm/11 : FAILURE
CI_IGT_test runtime 462 seconds
Rebooting fi-bdw-gvtdvm

Last dmesg:
<7>[   47.105282] [IGT] gem_exec_reuse: executing
<4>[   47.156940] Setting dangerous option reset - tainting kernel
<6>[   47.222774] perf: interrupt took too long (5310 > 5207), lowering kernel.perf_event_max_sample_rate to 37000
<7>[   48.749247] [IGT] gem_exec_reuse: starting subtest contexts
<6>[   56.215270] perf: interrupt took too long (6707 > 6637), lowering kernel.perf_event_max_sample_rate to 29000
<6>[   67.716713] perf: interrupt took too long (8462 > 8383), lowering kernel.perf_event_max_sample_rate to 23000

I guess can be closed?

Not seen on CI lately.