99536 – [SKL] GPU HANG: ecode 9:0:0x85dffffb, in Xorg [956], reason: Hang on render ring, action: reset

Bug 99536 - [SKL] GPU HANG: ecode 9:0:0x85dffffb, in Xorg [956], reason: Hang on render ring, action: reset

Summary: [SKL] GPU HANG: ecode 9:0:0x85dffffb, in Xorg [956], reason: Hang on render r...

Status:	RESOLVED MOVED

Alias:	None

Product:	Mesa
Classification:	Unclassified
Component:	Drivers/DRI/i965 (show other bugs)
Version:	unspecified
Hardware:	x86-64 (AMD64) Linux (All)

Importance:	medium normal
Assignee:	vladimir.campos
QA Contact:	Intel 3D Bugs Mailing List

URL:
Whiteboard:
Keywords:

Depends on:
Blocks:

Reported:	2017-01-25 17:04 UTC by Andreas Metzler
Modified:	2019-09-25 19:00 UTC (History)
CC List:	1 user (show)

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

Attachments
/sys/class/drm/card0/error (112.11 KB, application/gzip) 2017-01-25 17:04 UTC, Andreas Metzler	Details
View All

Description Andreas Metzler 2017-01-25 17:04:18 UTC

Created attachment 129144 [details]
/sys/class/drm/card0/error

Hello,

Shortly after startup my xserver hangs, and shows some artifacts. After a couple of seconds it starts usually working again. Today it seem to have crashed the xserver.

Jän 25 17:38:06 argenau kernel: [drm] GPU HANG: ecode 9:0:0x85dffffb, in Xorg [956], reason: Hang on render ring, action: reset
Jän 25 17:38:06 argenau kernel: [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Jän 25 17:38:06 argenau kernel: [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Jän 25 17:38:06 argenau kernel: [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Jän 25 17:38:06 argenau kernel: [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Jän 25 17:38:06 argenau kernel: [drm] GPU crash dump saved to /sys/class/drm/card0/error
Jän 25 17:38:06 argenau kernel: drm/i915: Resetting chip after gpu hang
Jän 25 17:38:06 argenau kernel: [drm] RC6 on
Jän 25 17:38:06 argenau kernel: [drm] GuC firmware load skipped
Jän 25 17:38:17 argenau kernel: drm/i915: Resetting chip after gpu hang
Jän 25 17:38:17 argenau kernel: [drm] RC6 on
Jän 25 17:38:17 argenau kernel: [drm] GuC firmware load skipped

This is on Debian/testing on Skylake (Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz)
ametzler@argenau:~$ uname -a
Linux argenau 4.9.0-1-amd64 #1 SMP Debian 4.9.2-2 (2017-01-12) x86_64 GNU/Linux
libdrm-intel1:amd64          2.4.74-1
libegl1-mesa:amd64           13.0.3-1
libgl1-mesa-dri:amd64        13.0.3-1
libgl1-mesa-glx:amd64        13.0.3-1
libglapi-mesa:amd64          13.0.3-1
libglu1-mesa:amd64           9.0.0-2.1
libwayland-egl1-mesa:amd64   13.0.3-1
mesa-utils                   8.3.0-3
mesa-vdpau-drivers:amd64     13.0.3-1
xorg                         1:7.7+18
xorg-docs-core               1:1.7.1-1
xserver-xorg                 1:7.7+18
xserver-xorg-core            2:1.19.0-3
xserver-xorg-input-evdev     1:2.10.4-1+b1
xserver-xorg-video-intel     2:2.99.917+git20161206-1

Comment 1 Elizabeth 2018-03-21 21:52:21 UTC

Hello Andreas, Mesa 13 is quite old. If this is still reproducible, you could try new mesa release 17.3.6.

Comment 2 Andreas Metzler 2018-03-22 14:56:48 UTC

I occasionally still see this, now with mesa 17.3.6, kernel 4.14.17 and xserver-xorg-video-intel 2.99.917+git20171229. Things have improved though, with the latest microcode updates.

Comment 3 Elizabeth 2018-03-22 15:12:45 UTC

Thanks for the update, if you find a way to reliably reproduce/trigger this hang, please let us know.

Comment 4 GitLab Migration User 2019-09-25 19:00:06 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/mesa/mesa/issues/1567.

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.