[Bug 67243] [ILK]igt/kms_render/gpu-blit randomly causes system hang

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=67243

lu hua <huax.lu at intel.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEEDINFO                    |NEW

--- Comment #14 from lu hua <huax.lu at intel.com> ---
(In reply to comment #13)
> Created attachment 84929 [details] [review]
> fix ilk ring flush workaround
> 
> The only notable from your .config was that you're running on 32 bit kernel,
> but I couldn't reproduce the hang even on that.
> 
> I'm still not sure that it's not a ring flushing issue, since in your last
> reply you didn't provide separate results for --run-subtest gpu-blit and
> direct-render. Anyway, I noticed that Chris' infinite __wait_seqno patch
> makes a clear improvement in gem_quiescent_gpu(), by getting rid of
> occasional stalls in it, so it would be worth trying it (and I just realized
> he also suggested this earlier on IRC to me). Please give a go to the
> following:
> 
> https://patchwork.kernel.org/patch/2849600/
> 
> I also found an igt regression in flush_on_ring_common() and since that is
> related to ilk, it's also a possible cause. I attached the fix for it, could
> you please try that too?


I test with these two patches, It happens in the fourth cycle.
output:
Beginning test gpu-blit with 1366x768 @ 60Hz / RGB565 on pipe A, encoder TMDS,
connector eDP
Test gpu-blit with 1366x768 @ 60Hz / RGB565 on pipe A, encoder TMDS, connector
eDP: PASSED
Beginning test gpu-blit with 1366x768 @ 60Hz / RGB888 on pipe A, encoder TMDS,
connector eDP
Test gpu-blit with 1366x768 @ 60Hz / RGB888 on pipe A, encoder TMDS, connector
eDP: SKIPPED
Beginning test gpu-blit with 1366x768 @ 60Hz / XRGB8888 on pipe A, encoder
TMDS, connector eDP
Test gpu-blit with 1366x768 @ 60Hz / XRGB8888 on pipe A, encoder TMDS,
connector eDP: PASSED
Beginning test gpu-blit with 1366x768 @ 60Hz / XRGB2101010 on pipe A, encoder
TMDS, connector eDP
Test gpu-blit with 1366x768 @ 60Hz / XRGB2101010 on pipe A, encoder TMDS,
connector eDP: PASSED
Beginning test gpu-blit with 1366x768 @ 60Hz / ARGB8888 on pipe A, encoder
TMDS, connector eDP
Test gpu-blit with 1366x768 @ 60Hz / ARGB8888 on pipe A, encoder TMDS,
connector eDP: PASSED
Beginning test gpu-blit with 1366x768 @ 40Hz / RGB565 on pipe A, encoder TMDS,
connector eDP
Test gpu-blit with 1366x768 @ 40Hz / RGB565 on pipe A, encoder TMDS, connector
eDP: PASSED
Beginning test gpu-blit with 1366x768 @ 40Hz / RGB888 on pipe A, encoder TMDS,
connector eDP
Test gpu-blit with 1366x768 @ 40Hz / RGB888 on pipe A, encoder TMDS, connector
eDP: SKIPPED
Beginning test gpu-blit with 1366x768 @ 40Hz / XRGB8888 on pipe A, encoder
TMDS, connector eDP
Test gpu-blit with 1366x768 @ 40Hz / XRGB8888 on pipe A, encoder TMDS,
connector eDP: PASSED

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20130902/d655cab7/attachment.html>