Bug 93465 - skylake gpu hang when playing company of heroes 2
Summary: skylake gpu hang when playing company of heroes 2
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Ian Romanick
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-12-21 10:44 UTC by mattenklicker
Modified: 2018-07-05 15:28 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
dmesg output (8.36 KB, text/plain)
2015-12-21 10:44 UTC, mattenklicker
Details
gpu error dump (561.91 KB, text/plain)
2015-12-21 10:46 UTC, mattenklicker
Details

Description mattenklicker 2015-12-21 10:44:13 UTC
Created attachment 120622 [details]
dmesg output

kernel: intel-drm-nightly 4.4 from 20151220
mesa: 11.1
GPU: i5-6600K iGPU
Resolution: 1920x1080 via HDMI

Steps to reproduce: Start a campain in CoH2. Crash happens always in the first scripted game scene with heavy explosions and smoke.
Comment 1 mattenklicker 2015-12-21 10:46:34 UTC
Created attachment 120623 [details]
gpu error dump
Comment 2 Chris Wilson 2015-12-21 10:56:52 UTC
Let's assume for a moment this is a mesa bug rather than a missing w/a. The first step will be to generate an apitrace so that the trigger can be narrowed down.
Comment 3 Denis 2018-07-05 12:15:22 UTC
Hi. This issue can be closed as fixed somewhere between 10.6.0 and 18.1.3 (release).


__________
Hang was reproduced on 10.6.0 (exactly on mentioned place - when the bomb is hitting boat):

[687866.977207] Watchdog[8183]: segfault at 0 ip 00007f4cbcecc3a6 sp 00007f4caac0bd60 error 6 in libcef.so[7f4cb9e65000+3cae000]
[688467.141132] CompanyOfHeroes[8147]: segfault at 21 ip 0000000000000021 sp 00007fff1a2f2808 error 14 in CompanyOfHeroes2[400000+4515000]
[688613.861152] Watchdog[8550]: segfault at 0 ip 00007f7c98afd3a6 sp 00007f7c8538ad60 error 6 in libcef.so[7f7c95a96000+3cae000]
[688796.106810] [drm] GPU HANG: ecode 9:0:0x84df7cfc, in CompanyOfHeroes [8513], reason: Hang on rcs0, action: reset
[688796.106811] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[688796.106812] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[688796.106812] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[688796.106813] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[688796.106813] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[688796.106826] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688805.095709] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688814.088209] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688823.111753] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688850.088083] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688859.111630] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688868.104132] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688877.095641] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688888.103609] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688897.095673] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688906.088134] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688915.111688] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688924.104116] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688933.095578] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[688939.791454] CompanyOfHeroes[8513]: segfault at 0 ip 0000000000000000 sp 00007fffc9b610d8 error 14 in CompanyOfHeroes2[400000+4515000]
____________________________________________________________________

Hang wasn't reproduced on 18.1.3 release.


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.