Bug 99184 - [SNB] GPU HANG: ecode 6:0:0xfeffffff, reason: Hang on render ring, action: reset (on drm-tip)
Summary: [SNB] GPU HANG: ecode 6:0:0xfeffffff, reason: Hang on render ring, action: re...
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: DRI git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-12-22 15:40 UTC by Erik Quaeghebeur
Modified: 2017-04-11 13:33 UTC (History)
2 users (show)

See Also:
i915 platform: SNB
i915 features: display/Other, GPU hang


Attachments
/sys/class/drm/card0/error (14.51 KB, text/plain)
2016-12-22 15:40 UTC, Erik Quaeghebeur
no flags Details

Description Erik Quaeghebeur 2016-12-22 15:40:05 UTC
Created attachment 128637 [details]
/sys/class/drm/card0/error

drm-tip (4.9.0+, commit 86a26727690faef9018c6585e0094e701e0e9b0c)

    dec 20 22:23:27 <hostname> kernel: [drm] GPU HANG: ecode 6:-1:0x00000000, reason: Kicking stuck semaphore on render ring, action: continue
    dec 20 22:23:27 <hostname> kernel: [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
    dec 20 22:23:27 <hostname> kernel: [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
    dec 20 22:23:27 <hostname> kernel: [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
    dec 20 22:23:27 <hostname> kernel: [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
    dec 20 22:23:27 <hostname> kernel: [drm] GPU crash dump saved to /sys/class/drm/card0/error
Comment 1 yann 2017-01-20 17:37:22 UTC
Erik Quaeghebeur, please update your kernel with latest on and confirm whether or not you are reproducing the issue
Comment 2 Erik Quaeghebeur 2017-01-26 08:32:00 UTC
(In reply to yann from comment #1)
> Erik Quaeghebeur, please update your kernel with latest on and confirm
> whether or not you are reproducing the issue

I've tested 4.10.0-rc3+, i.e., drm-tip aa012aa081f6a6d2dd5a1df0f3c3736017df0d56 and didn't reproduce this specific issue, but my compositor crashed, which I've not had happen in a long time (years?). So I'm back to my standard 4.8.17, whose problem isn't fixed (see Bug 98516), but neither does the drm-tip I tested. (Since an upgrade of the desktop environment, the original issue does seem to be occurring much less).
Comment 3 Chris Wilson 2017-01-26 09:46:12 UTC
86a26727690faef9018c6585e0094e701e0e9b0c dies trying to do a flip before it has reacquired the outputs. Latest drm-tip will not do that as it no longer uses CS flips.
Comment 4 Erik Quaeghebeur 2017-01-26 10:14:47 UTC
(In reply to Chris Wilson from comment #3)
> 86a26727690faef9018c6585e0094e701e0e9b0c dies trying to do a flip before it
> has reacquired the outputs. Latest drm-tip will not do that as it no longer
> uses CS flips.
So to be clear: I should update and try again?
Comment 5 Ricardo 2017-03-03 17:00:00 UTC
(In reply to Erik Quaeghebeur from comment #4)
> (In reply to Chris Wilson from comment #3)
> > 86a26727690faef9018c6585e0094e701e0e9b0c dies trying to do a flip before it
> > has reacquired the outputs. Latest drm-tip will not do that as it no longer
> > uses CS flips.
> So to be clear: I should update and try again?

yes you should retry using latest from https://cgit.freedesktop.org/drm-tip
Comment 6 Ricardo 2017-03-14 15:12:36 UTC
Looks like the issue was fixed with DRM-Tip, please let us know. 

If the problem persist attach update logs, and change the bug to reopen status
if the problem is fixed change the bug to resolved
Comment 7 yann 2017-04-11 13:13:06 UTC
(In reply to Ricardo from comment #6)
> Looks like the issue was fixed with DRM-Tip, please let us know. 
> 
> If the problem persist attach update logs, and change the bug to reopen
> status
> if the problem is fixed change the bug to resolved

Timeout - assuming resolved+fixed.

If problem still persist with the latest kernels (preferable drm-tip from git://anongit.freedesktop.org/git/drm-tip), reopen this bug with latest logs as attachments.
Comment 8 Erik Quaeghebeur 2017-04-11 13:33:22 UTC
(In reply to yann from comment #7)
> Timeout - assuming resolved+fixed.

Fair enough. (Did not have time to investigate.)
 
> If problem still persist with the latest kernels (preferable drm-tip from
> git://anongit.freedesktop.org/git/drm-tip), reopen this bug with latest logs
> as attachments.

Ok.


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.