Created attachment 80931 [details]
dmesg

System Environment:
--------------------------
Platform:         Pineview/GM45
Kernel:     (drm-intel-next-queued)cab8b5862acd55019fbeede6940d1a601912d6b8

Bug detailed description:
-------------------------
It segfault on Pineview and GM45 with drm-intel-next-queued kernel.It works well on drm-intel-fixes kernel.

Bisect shows:a89cdd3cfcd66d932d4ae2617c5e89279515f14d is the first bad commit.
commit a89cdd3cfcd66d932d4ae2617c5e89279515f14d
Author:     Mika Kuoppala <mika.kuoppala@linux.intel.com>
AuthorDate: Wed Jun 12 12:35:34 2013 +0300
Commit:     Daniel Vetter <daniel.vetter@ffwll.ch>
CommitDate: Thu Jun 13 17:42:18 2013 +0200

    drm/i915: refuse to submit more batchbuffers from guilty context

    If context has recently submitted a faulty batchbuffers guilty of
    gpu hang and decides to keep submitting more crap, ban it permanently.

    Note: This is just a refinment of the existing policy to declare the
    gpu wedged when it's hanging too often. We have that in place to
    increase the odds of a system surviving gpu hang storms (and so
    getting a useful bug report with the error state).

    With this patch here we reduce the risk to take down unrelated
    applications, since especially modern compositors tend to die when
    mesa can't submit batches any more. So for the somewhat common case of
    a specific gl workload causing a gpu hang flurry we should now have
    less grumpy users since we don't take down their entire deskopt
    enviroment any more.

    v2: Store guilty ban status bool in gpu_error instead of pointers
        that might become danling before hang is declared.

    v3: Use return value for banned status instead of stashing state
        into gpu_error (Chris Wilson)

    Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
    Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
    [danvet: Added note to commit message to explain why we want this,
    which addresses a concern Ben raised.]
    Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>


output:
Running synchronized to the vertical refresh.  The framerate should be
approximately the same as the monitor refresh rate.
Segmentation fault (core dumped)


Reproduce steps:
----------------
1. xinit
2. glxgears