[Bug 88727] [gen4] Batch buffer overwritten

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=88727

Chris Wilson <chris at chris-wilson.co.uk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[drm] GPU HANG ecode        |[gen4] Batch buffer
                   |0:0x00ababab, in            |overwritten
                   |systemd-logind [248],       |
                   |reason: Ring hung, action:  |
                   |reset                       |

--- Comment #1 from Chris Wilson <chris at chris-wilson.co.uk> ---
There are a few cachelines worth of pixel data inside the batchbuffer - which
is what the GPU hung on. The writes within the batches, up to the hang at
least, are all contained within their respective target surfaces and shouldn't
be the source of the scribbling. The batch doesn't overlap a fence register. It
is not clear how we could end up with corruption in the middle of the batch.

All I can ask is that you keep filing the error state in case one captures a
clear violation, and if you could build debug packages of X, xf86-video-intel,
mesa, the additional self-checks may find the culprit.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are on the CC list for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20150123/abbc39ab/attachment.html>