Bug 81710 - after Kernel Update to 3.15.6-40.2.gfdb2dde i915 G850 GPU HANG: ecode 0:0xf4e9fffe, in Xorg [698], reason: Ring hung, action: reset
Summary: after Kernel Update to 3.15.6-40.2.gfdb2dde i915 G850 GPU HANG: ecode 0:0xf4e...
Status: CLOSED DUPLICATE of bug 54226
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Daniel Vetter
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-07-24 10:11 UTC by f.st
Modified: 2017-07-24 22:52 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
enabling Option Tearfree and SNA (153 bytes, text/plain)
2014-07-24 10:25 UTC, f.st
no flags Details
/var/log/messages (2.45 MB, text/plain)
2014-07-24 11:00 UTC, f.st
no flags Details
cat /sys/class/drm/card0/error >/tmp/dumpfile (2.10 MB, text/plain)
2014-07-24 13:30 UTC, f.st
no flags Details
current messages (122.11 KB, text/plain)
2014-07-24 13:32 UTC, f.st
no flags Details

Description f.st 2014-07-24 10:11:19 UTC
OpenSUSE 13.1 with Tumbleweed Kernel 3.15.6-40.2.gfdb2dde
Kernel Update was made to get hibernate/resume to work correctly. See Bug 888666 in bugzilla.novell.com.

The GPU Hang documented here happens after a clean cold start, so it's not because of PM-events.

2014-07-24T10:18:36.913598+02:00 peacy kernel: [ 1499.114478] [drm] stuck on render ring
2014-07-24T10:18:36.913613+02:00 peacy kernel: [ 1499.114486] [drm] stuck on blitter ring
2014-07-24T10:18:36.914147+02:00 peacy kernel: [ 1499.115033] [drm] GPU HANG: ecode 0:0xf4e9fffe, in Xorg [698], reason: Ring hung, action: reset
2014-07-24T10:18:36.914155+02:00 peacy kernel: [ 1499.115036] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
2014-07-24T10:18:36.914156+02:00 peacy kernel: [ 1499.115038] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
2014-07-24T10:18:36.914157+02:00 peacy kernel: [ 1499.115039] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
2014-07-24T10:18:36.914157+02:00 peacy kernel: [ 1499.115040] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
2014-07-24T10:18:36.914158+02:00 peacy kernel: [ 1499.115042] [drm] GPU crash dump saved to /sys/class/drm/card0/error
2014-07-24T10:18:36.914478+02:00 peacy kernel: [ 1499.115087] [drm:i915_context_is_banned] *ERROR* gpu hanging too fast, banning!
2014-07-24T10:18:38.914461+02:00 peacy kernel: [ 1501.115165] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off
Comment 1 f.st 2014-07-24 10:25:45 UTC
Created attachment 103395 [details]
enabling Option Tearfree and SNA
Comment 2 f.st 2014-07-24 10:28:01 UTC
Trying booting now with i915.semaphores=1

linux	/boot/vmlinuz-3.15.6-40.gfdb2dde-desktop root=UUID=fc2823aa-e3b-4fd4-8111-aa16ef2538de   resume=/dev/disk/by-id/ata-OCZ-VERTEX4_OCZ-CPQ92H5W3DKY9T58-part2 splash=silent quiet showopts i915.semaphores=1
Comment 3 f.st 2014-07-24 10:45:52 UTC
booting with i915.semaphores=1 doesn't solve the problem
Comment 4 f.st 2014-07-24 10:56:01 UTC
trying change from SNA to UXA now



"/sys/class/drm/card0/error" dump file is always 0-length so now upload possible




2014-07-24T12:32:42.914130+02:00 peacy kernel: [  837.937346] [drm] GPU HANG: ecode -1:0x00000000, reason: Kicking stuck semaphore on render ring, action: continue
2014-07-24T12:32:42.914143+02:00 peacy kernel: [  837.937354] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
2014-07-24T12:32:42.914144+02:00 peacy kernel: [  837.937356] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
2014-07-24T12:32:42.914145+02:00 peacy kernel: [  837.937357] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
2014-07-24T12:32:42.914145+02:00 peacy kernel: [  837.937358] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
2014-07-24T12:32:42.914146+02:00 peacy kernel: [  837.937360] [drm] GPU crash dump saved to /sys/class/drm/card0/error


2014-07-24T12:40:26.910209+02:00 peacy kernel: [ 1302.056521] [drm] stuck on render ring
2014-07-24T12:40:26.910221+02:00 peacy kernel: [ 1302.056529] [drm] stuck on blitter ring
2014-07-24T12:40:26.910222+02:00 peacy kernel: [ 1302.057067] [drm] GPU HANG: ecode 0:0xf4e9fffe, in Xorg [680], reason: Ring hung, action: reset
2014-07-24T12:40:26.910725+02:00 peacy kernel: [ 1302.057269] [drm:i915_context_is_banned] *ERROR* gpu hanging too fast, banning!
2014-07-24T12:40:28.910542+02:00 peacy kernel: [ 1304.057239] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off
Comment 5 f.st 2014-07-24 11:00:08 UTC
Created attachment 103396 [details]
/var/log/messages
Comment 6 Chris Wilson 2014-07-24 11:13:59 UTC
(In reply to comment #3)
> booting with i915.semaphores=1 doesn't solve the problem

No, i915.semaphores=0 would. Well that's a guess since you didn't follow:

[drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[drm] GPU crash dump saved to /sys/class/drm/card0/error
Comment 7 f.st 2014-07-24 11:32:31 UTC
please see above, the mentioned dump is empty (0-length)

UXA seems more stable, but slow and tearing.

I'll try SNA and i915.semaphores=0 now.
Comment 8 f.st 2014-07-24 12:02:36 UTC
in the short time now, it seems, that SNA/Tearfree and i915.semaphores=0 resolved the problem. I'll get back to you in case the problem occurs again.
Thank you, Chris !
Frank
Comment 9 Chris Wilson 2014-07-24 12:35:12 UTC
Just cat /sys/class/drm/card0/error - it's a virtual file so its size is always reported as zero but it does have content!
Comment 10 f.st 2014-07-24 12:38:32 UTC
OK done that :

peacy:~ # cat /sys/class/drm/card0/error
no error state collected
Comment 11 f.st 2014-07-24 12:42:05 UTC
I'll do it again immediately after the error ... wait
Comment 12 f.st 2014-07-24 13:30:10 UTC
Created attachment 103399 [details]
cat /sys/class/drm/card0/error >/tmp/dumpfile
Comment 13 f.st 2014-07-24 13:32:53 UTC
Created attachment 103400 [details]
current messages
Comment 14 f.st 2014-07-24 13:36:11 UTC
Hi Chris,
now I was successful collection the dump.
Thank you
Frank
Comment 15 Chris Wilson 2014-07-24 13:45:27 UTC
As expected it is the semaphore hang.

*** This bug has been marked as a duplicate of bug 54226 ***


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.