Bug 111888 - [SKL] GPU hang when watching video using GNOME MPV using VAAPI
Summary: [SKL] GPU hang when watching video using GNOME MPV using VAAPI
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: DRI git
Hardware: Other All
: medium major
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
Whiteboard: Triaged
Depends on:
Reported: 2019-10-03 02:02 UTC by leozinho29_eu
Modified: 2019-11-29 19:37 UTC (History)
1 user (show)

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

Logs and MPV configuration (22.29 KB, application/gzip)
2019-10-03 02:02 UTC, leozinho29_eu
no flags Details
card0error (4.81 KB, text/x-log)
2019-10-04 22:05 UTC, leozinho29_eu
no flags Details

Description leozinho29_eu 2019-10-03 02:02:13 UTC
Created attachment 145617 [details]
Logs and MPV configuration

When watching videos with GNOME MPV on which VAAPI can be used to decode them, multiple GPU hangs may happen. As the GPU hangs never stop, it's required to use another system, start a SSH session and restart LightDM to make the frozen computer work again. Sometimes it may kept hanging for a long time until the GPU hangs stop.

Stesps to reproduce:

1) Have a video that can be decoded using VAAPI;
2) Have GNOME MPV (probably MPV works too) configured to decode using VAAPI;
3) Play the video using GNOME MPV;
4) Let the video play until the GPU hangs start.

It is very complicated to recover the frozen machine without another computer, but it is possible.

The attached file contain dmesg, the card0/error and MPV configuration that reproduces this.

Processor: Intel Core i3-6100U;
Video: Intel HD Graphics 520;
Architecture: amd64;
Mesa:  19.3.0-devel (git-5a747306ce);
Kernel version: drm-tip (5.4.0-rc1-9300459553e8);
Distribution: Xubuntu 18.04.3 amd64.
Comment 1 Lakshmi 2019-10-04 15:07:21 UTC
[  692.065533] i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0
[  692.065535] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[  692.065535] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[  692.065535] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[  692.065536] The GPU crash dump is required to analyze GPU hangs, so please always attach it.
[  692.065537] GPU crash dump saved to /sys/class/drm/card0/error

Can you please attach the error file?
Comment 2 leozinho29_eu 2019-10-04 22:05:47 UTC
Created attachment 145652 [details]

The card0error was in the first attachment together with the other files. Here is it uncompressed.
Comment 3 Martin Peres 2019-11-29 19:37:09 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/intel/issues/477.

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.