Bug 96281 - [drm] GPU HANG: ecode 9:0:0x85dffffb, in csgo_linux64 [3821], reason: Engine(s) hung, action: reset (NUC6i7KYK)
Summary: [drm] GPU HANG: ecode 9:0:0x85dffffb, in csgo_linux64 [3821], reason: Engine(...
Status: CLOSED WORKSFORME
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
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: 2016-05-30 22:15 UTC by Pontus Lundkvist
Modified: 2017-07-24 22:41 UTC (History)
4 users (show)

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


Attachments
/sys/class/drm/card0/error (87.18 KB, application/gzip)
2016-05-30 22:15 UTC, Pontus Lundkvist
no flags Details
dmesg (195.38 KB, application/gzip)
2016-05-30 22:16 UTC, Pontus Lundkvist
no flags Details
dmesg entries for the error. (1.08 KB, text/plain)
2016-07-08 20:16 UTC, angiolucci
no flags Details
/sys/class/drm/card0/error (549.34 KB, text/plain)
2016-07-08 20:17 UTC, angiolucci
no flags Details

Description Pontus Lundkvist 2016-05-30 22:15:11 UTC
Created attachment 124191 [details]
/sys/class/drm/card0/error

Hardware: NUC6i7KYK (i7-6770HQ)
OS: Debian stretch

Same problem with default kernel in stretch (linux-image-4.5.0-2-amd64 4.5.4-1), 4.6 from experimental (linux-image-4.6.0-trunk-amd64 4.6-1~exp1) as well as what I'm doing this bugreport with; drm-intel-nightly: 2016y-05m-30d-17h-51m-33s, built from git.

Running mesa 11.2.2-1 and xserver-xorg-video-intel 2:2.99.917+git20160522-1, all from stretch.

...
[  160.897606] [drm] RC6 on
[  169.886045] [drm] stuck on render ring
[  169.888450] [drm] GPU HANG: ecode 9:0:0xfffffffe, in csgo_linux64 [1172], reason: Engine(s) hung, action: reset
[  169.888516] [drm:i915_reset_and_wakeup] resetting chip
[  169.890965] drm/i915: Resetting chip after gpu hang
...

Attaching full kernel output (*very* verbose from drm.debug=0x1e) as well as card0/error output.
Comment 1 Pontus Lundkvist 2016-05-30 22:16:05 UTC
Created attachment 124192 [details]
dmesg
Comment 2 Pontus Lundkvist 2016-05-30 22:30:58 UTC
With i915.enable_rc6=0 the result is the same (with dmesg saying [drm] RC6 off).
Comment 3 Nick Maniscalco 2016-07-07 02:52:22 UTC
FWIW, I get the same hang on my NUC6i7KYK with TF2 and Kerbal Space Program.  With TF2 it happens reliably within ~10 seconds of entering the game.  Sometimes it brings down the whole machine (unresponsive until reboot).

I'm running Linux 4.7.0-994-generic from http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-nightly/2016-07-05-yakkety/

I'd be happy to file a separate bug report if that would help someone troubleshoot this.
Comment 4 angiolucci 2016-07-08 20:15:26 UTC
Same issue here. 
GPU hang while in CSGO_LINUX64. It do not happens always, so I can't provide steps to reproduce.
Last time it happened, I was watching (in-game) a live from CSGO championship.
Comment 5 angiolucci 2016-07-08 20:16:19 UTC
Created attachment 124960 [details]
dmesg entries for the error.
Comment 6 angiolucci 2016-07-08 20:17:23 UTC
Created attachment 124961 [details]
/sys/class/drm/card0/error
Comment 7 mwa 2016-08-11 11:46:35 UTC
There may be a chance that this is a manifestation of https://bugs.freedesktop.org/show_bug.cgi?id=96175.

Does upgrading your Mesa version do anything for you?
Comment 8 Pontus Lundkvist 2016-08-11 21:36:11 UTC
Upgrading to Mesa 12.0.1 (The 12.0.1-3 packages from Debian experimental) seems to have fixed the problem stated in my initial post at least, yes - Great!

Will test some other programs and maybe newer kernel than 4.6-series again, will report back if I get it to crash with the same error as before.
But if not, consider the issue resolved for me at least.

Thank you very much.
Comment 9 Pontus Lundkvist 2016-08-11 21:59:07 UTC
Tested a few other heavy games, as well as running on Linux 4.7 again, all works great.

I'm unsure about what Status you want, I'd call it RESOLVED and WORKSFORME, pardon me if that is incorrect.


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.