Please bisect.

Bisect shows: 8232644ccf099548710843e97360a3fcd6d28e04 is the first bad commit
commit 8232644ccf099548710843e97360a3fcd6d28e04
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Mar 5 12:00:39 2014 +0000

    drm/i915: Convert the forcewake worker into a timer func

    We don't want to suffer scheduling delay when turning off the GPU after
    waking it up to touch registers. Ideally, we only want to keep the GPU
    awake for the register access sequence, with a single forcewake dance on
    the first access and release immediately after the last. We set a timer
    on the first access so that we only dance once and on the next scheduler
    tick, we drop the forcewake again.

    This moves the cleanup routine from the common i915 workqueue to a timer
    func so that we don't anger powertop, and drop the forcewake again
    quicker.

    v2: Enable the deferred force_wake_put for regular register reads as
        well.
    v3: Beautification and make sure we disable forcewake when shutting
        down.

    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Cc: Ben Widawsky <ben@bwidawsk.net>
    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>

Can you please attach the output (trace.dat) of:

$ trace-cmd record -e i915 ./gem_dummy_reloc_loop --run-subtest blt

(In reply to comment #3)
> Can you please attach the output (trace.dat) of:
> 
> $ trace-cmd record -e i915 ./gem_dummy_reloc_loop --run-subtest blt

/sys/kernel/debug/tracing/events/i915/filter
/sys/kernel/debug/tracing/events/*/i915/filter
IGT-Version: 1.6-gcde058a (x86_64) (Linux: 3.14.0-rc7_drm-intel-nightly_2c0d38_2
running dummy loop on render
dummy loop run on render completed
Subtest render: SUCCESS
running dummy loop on bsd
dummy loop run on bsd completed
Subtest bsd: SUCCESS
running dummy loop on blt
dummy loop run on blt completed
Subtest blt: SUCCESS
running dummy loop on vebox
Connection to x-bdw02 closed by remote host.
Connection to x-bdw02 closed.
[root@x-pk2 ~]# ssh x-bdw02
Last login: Thu Dec 12 00:01:05 2013 from luhua.sh.intel.com
[root@x-bdw02 ~]# uname -a
Linux x-bdw02 3.13.0_drm-intel-next-queued_823264_20140314_debug+ #960 SMP Fri Mar 14 02:31:27 CST 2014 x86_64 x86_64 x86_64 GNU/Linux
[root@x-bdw02 ~]# cd /GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests/
[root@x-bdw02 tests]# trace-cmd record -e i915 ./gem_dummy_reloc_loop
/sys/kernel/debug/tracing/events/i915/filter
/sys/kernel/debug/tracing/events/*/i915/filter
IGT-Version: 1.6-gcde058a (x86_64) (Linux: 3.13.0_drm-intel-next-queued_823264_20140314_debug+ x86_64)
running dummy loop on render
dummy loop run on render completed
Subtest render: SUCCESS
running dummy loop on bsd
dummy loop run on bsd completed
Subtest bsd: SUCCESS
running dummy loop on blt
dummy loop run on blt completed
Subtest blt: SUCCESS
running dummy loop on vebox
dummy loop run on vebox completed
Subtest vebox: SUCCESS
running dummy loop on random rings
dummy loop run on random rings completed
Subtest mixed: SUCCESS
Kernel buffer statistics:
  Note: "entries" are the entries left in the kernel ring buffer and are not
        recorded in the trace data. They should all be zero.

CPU: 0
entries: 0
overrun: 0
commit overrun: 0
bytes: 3868
oldest event ts:   522.073751
now ts:   522.154184
dropped events: 0
read events: 68516225

CPU: 1
entries: 0
overrun: 0
commit overrun: 0
bytes: 1800
oldest event ts:   521.657562
now ts:   522.154479
dropped events: 0
read events: 6934134

CPU: 2
entries: 0
overrun: 0
commit overrun: 0
bytes: 304
oldest event ts:   520.351796
now ts:   522.154750
dropped events: 0
read events: 15421150

CPU: 3
entries: 0
overrun: 0
commit overrun: 0
bytes: 780
oldest event ts:   506.852074
now ts:   522.155017
dropped events: 0
read events: 2452128

CPU0 data recorded at offset=0x33a000
    1883295744 bytes in size
CPU1 data recorded at offset=0x70747000
    190619648 bytes in size
CPU2 data recorded at offset=0x7bd11000
    423923712 bytes in size
CPU3 data recorded at offset=0x9515a000
    67411968 bytes in size

Attach the trace.dat

(In reply to comment #5)
> Attach the trace.dat

The file is too large. I will give you file location via mail.

Chris, any update on this issue?

Nope. The debug perturbed the system enough to hide the issue. So far it looks like a genuine missed interrupt.

I can't reproduce, still an issue?

Test it on latest -nightly kernel, It works well.

Verified.Fixed.

Closing old verified.