Bug 76763 - GPU hang and display corruption after mesa 10.1 upgrade
Summary: GPU hang and display corruption after mesa 10.1 upgrade
Status: RESOLVED DUPLICATE of bug 77207
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: 10.1
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Ian Romanick
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-03-28 23:44 UTC by Coucouf
Modified: 2014-04-13 17:18 UTC (History)
2 users (show)

See Also:
i915 platform:
i915 features:


Attachments
/sys/class/drm/card0/error (2.14 MB, text/plain)
2014-03-28 23:47 UTC, Coucouf
Details
Display corruption screenshot (68.16 KB, image/png)
2014-03-28 23:51 UTC, Coucouf
Details

Description Coucouf 2014-03-28 23:44:59 UTC
Looks a bit like #75724 with additional display corruption, feel free to close if duplicate.

Hardware : Dell XPS 13 / i7-3537U
On Debian Jessie upgrading from mesa 9.2.2 to 10.1.0 leads to display corruption in plasma-desktop/kwin, kwin crashes and GPU hangs reported by dmesg.

Error dump attached.
Comment 1 Coucouf 2014-03-28 23:47:06 UTC
Created attachment 96573 [details]
/sys/class/drm/card0/error

Error dump
Comment 2 Coucouf 2014-03-28 23:51:22 UTC
Created attachment 96574 [details]
Display corruption screenshot
Comment 3 Chris Wilson 2014-03-29 07:26:54 UTC
Batch clobbered by wild write. Looks like the framebuffer was as well and the context elsewhere.
Comment 4 Coucouf 2014-04-03 13:48:17 UTC
In fact I also have occasional hangs on Mesa 9.2.2, but not that serious (no application crash, only 3D performance goes down), and without display corruption.

Dmesg goes :
[11010.821533] [drm] stuck on render ring
[11010.821537] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[11010.821538] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[11010.821539] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[11010.821540] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[11010.821540] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[11010.824949] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0xf985000 ctx 1) at 0xf985e00
[11016.849418] [drm] stuck on render ring
[11016.849459] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0x10e35000 ctx 1) at 0x10e35e00
[11016.849461] [drm:i915_context_is_banned] *ERROR* context hanging too fast, declaring banned!


I can post an error dump for that if it’s of any help.

Earlier I got :
[ 9058.287646] [drm:intel_dp_start_link_train] *ERROR* too many full retries, give up

but I would guess it’s about the display pannel and not related to my hangs and crashes.
Comment 5 Coucouf 2014-04-13 11:17:02 UTC
I can’t reproduce that bug after the last updates.
I waited a few weeks before retrying as it was really broken on my machine, so several other packages were updated in the meantime.

The display corruption used to happen immediately after login, and I retried for a few days with both kernels 3.13 and 3.14 and both work fine.
Comment 6 Kenneth Graunke 2014-04-13 17:18:54 UTC
You may want to try one of the workarounds I posted in bug #77207.

*** This bug has been marked as a duplicate of bug 77207 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.