Bug 99497 - Hang on null pointer dereference during gfx stack restart after GPU hang
Summary: Hang on null pointer dereference during gfx stack restart after GPU hang
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-01-22 22:44 UTC by Christopher Brodt
Modified: 2017-04-11 21:33 UTC (History)
2 users (show)

See Also:
i915 platform: KBL
i915 features: GPU hang


Attachments
1st boot log of the bug occuring (1.51 MB, text/x-log)
2017-01-22 22:44 UTC, Christopher Brodt
no flags Details
2nd log of the error occuring. (555.71 KB, text/x-log)
2017-01-22 22:44 UTC, Christopher Brodt
no flags Details
issue dmesg output w/debug parameters enabled (869.71 KB, application/gzip)
2017-01-22 23:46 UTC, Christopher Brodt
no flags Details

Description Christopher Brodt 2017-01-22 22:44:12 UTC
Created attachment 129098 [details]
1st boot log of the bug occuring

I seem to get this bug an indeterminate amount of time after a boot. It's happened twice when I'm away. Unfortunately, I don't know how to reproduce it yet as a result. I thought it was suspend, but that seems to work fine when I invoke it manually. I'll attach both boot logs. I have the debug kernel options added and will attach that log when it happens again.
Comment 1 Christopher Brodt 2017-01-22 22:44:46 UTC
Created attachment 129099 [details]
2nd log of the error occuring.
Comment 2 Christopher Brodt 2017-01-22 22:46:49 UTC
This is my kernel version: 4.9.3-1-default

I'm using OpenSUSE Tumbleweed, with this weeks latest updates.
Comment 3 Christopher Brodt 2017-01-22 23:46:21 UTC
Created attachment 129100 [details]
issue dmesg output w/debug parameters enabled
Comment 4 Christopher Brodt 2017-01-23 00:35:30 UTC
The issue may not be with suspend but with screen blanking. Disabling screen blanking in GNOME and seeing if that prevents the issue from occuring
Comment 5 Christopher Brodt 2017-01-23 00:35:30 UTC
The issue may not be with suspend but with screen blanking. Disabling screen blanking in GNOME and seeing if that prevents the issue from occuring
Comment 6 Chris Wilson 2017-01-25 17:55:51 UTC
Just landed a very minor tweak to the execlists dispatch in https://cgit.freedesktop.org/drm-tip that will be very useful in checking.
Comment 7 Christopher Brodt 2017-01-26 15:27:08 UTC
Great! I'll check it out and report back.
Comment 8 Jari Tahvanainen 2017-04-11 10:46:03 UTC
Christopher - any feedback related to comment 6?
Comment 9 Jari Tahvanainen 2017-04-11 10:50:06 UTC
Related to gpu hang, note the instructions on https://01.org/linuxgraphics/documentation/how-report-bugs.
Comment 10 Chris Wilson 2017-04-11 10:55:46 UTC
I have high confidence that this is now fixed,

commit fe3288b5da2c1286a7aac1fb1b2234caa752a81b
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Sun Feb 12 17:20:01 2017 +0000

    drm/i915: Park the breadcrumbs signaler across a GPU reset

commit 1f7b847d72c3583df5048d83bd945d0c2c524c28
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Feb 8 14:30:33 2017 +0000

    drm/i915: Disable engine->irq_tasklet around resets
Comment 11 Christopher Brodt 2017-04-11 21:33:54 UTC
Just an update, this has been working since the 4.10 release. Thanks


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.