Bug 97440 - GPU hang on Skylake whilst playing 3D games
Summary: GPU hang on Skylake whilst playing 3D games
Status: RESOLVED MOVED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Ian Romanick
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-08-22 19:51 UTC by Dan Dart
Modified: 2019-09-25 18:57 UTC (History)
1 user (show)

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


Attachments
/sys/class/drm/card0/error (562.16 KB, text/plain)
2016-08-22 19:51 UTC, Dan Dart
Details
dmesg | grep drm (2.85 KB, text/x-log)
2016-08-22 19:52 UTC, Dan Dart
Details
lshw output (22.01 KB, text/x-log)
2016-08-22 19:52 UTC, Dan Dart
Details

Description Dan Dart 2016-08-22 19:51:44 UTC
Created attachment 125956 [details]
/sys/class/drm/card0/error

Hello

I have a GPU hang very often when playing 3D games, which makes it jog and occasionally just completely crash out.

I have an Intel Skylake CPU/GPU stack (6200u).

Attached dmesg | grep drm and /sys/class/drm/card0/error file, with a lshw output if you need it.

I've had jogs and crashes on the ubuntu oibaf graphics drivers (gl 4.4 from git) but have reverted to the standard ones (gl 3.3) in Ubuntu 16.04.
Comment 1 Dan Dart 2016-08-22 19:52:27 UTC
Created attachment 125957 [details]
dmesg | grep drm
Comment 2 Dan Dart 2016-08-22 19:52:44 UTC
Created attachment 125958 [details]
lshw output
Comment 3 yann 2016-08-30 09:31:45 UTC
Assigning to Mesa product (please let me know if I am mistaken with this GPU Hang).

From this error dump, hung is happening in render ring batch with active head at 0xfd7d06e0, with 0x7a000004 (PIPE_CONTROL) as IPEHR.

Batch extract (around 0xfd7d06e0):

0xfd7d06a0:      0x78490001: 3D UNKNOWN: 3d_965 opcode = 0x7849
0xfd7d06a4:      0x00000004: MI_NOOP
0xfd7d06a8:      0x00000000: MI_NOOP
Bad length 7 in (null), expected 6-6
0xfd7d06ac:      0x7b000005: 3DPRIMITIVE: fail sequential
0xfd7d06b0:      0x00000104:    vertex count
0xfd7d06b4:      0x00000a74:    start vertex
0xfd7d06b8:      0x00000000:    instance count
0xfd7d06bc:      0x00000001:    start instance
0xfd7d06c0:      0x00000000:    index bias
0xfd7d06c4:      0x00000000: MI_NOOP
Bad count in PIPE_CONTROL
0xfd7d06c8:      0x7a000004: PIPE_CONTROL: no write, no depth stall, no RC write flush, no inst flush
0xfd7d06cc:      0x00101001:    destination address
0xfd7d06d0:      0x00000000:    immediate dword low
0xfd7d06d4:      0x00000000:    immediate dword high
Bad count in PIPE_CONTROL
0xfd7d06e0:      0x7a000004: PIPE_CONTROL: no write, no depth stall, no RC write flush, no inst flush
0xfd7d06e4:      0x00000408:    destination address
0xfd7d06e8:      0x00000000:    immediate dword low
0xfd7d06ec:      0x00000000:    immediate dword high
0xfd7d06f8:      0x78150009: 3D UNKNOWN: 3d_965 opcode = 0x7815
Comment 4 Kenneth Graunke 2016-09-11 04:29:26 UTC
Which game?
Comment 5 Dan Dart 2016-09-11 04:43:54 UTC
In this case it was thestanley parable, it happens with most of that power.
Comment 6 Marco d'Itri 2017-05-28 11:53:05 UTC
I am experiencing the same problem with the same game on Skylake hardware.
I am running Debian stretch.
Comment 7 Elizabeth 2018-03-21 22:38:01 UTC
Hello, some important fixes for games hangs were include in latest mesa release 17.3.6. Could you confirm that this also fix you issue? If issue persist, is there a way to easily trigger the hang? Thank you.
Comment 8 Denis 2018-07-12 12:10:28 UTC
hi. The same game was mentioned on this ticket https://bugs.freedesktop.org/show_bug.cgi?id=98013

during checking it I found out that 18.2.0 fixes an issue 100% (possibly even 17.3.6 fixes them).
Comment 9 Denis 2018-09-19 13:14:38 UTC
found one more report on the same game
https://bugs.freedesktop.org/show_bug.cgi?id=104440
Stanley Parable
Comment 10 GitLab Migration User 2019-09-25 18:57:43 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/1534.


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.