Bug 85157 - [IVB] GPU hang after S3 (3.13)
Summary: [IVB] GPU hang after S3 (3.13)
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
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: 2014-10-17 20:02 UTC by lharrington
Modified: 2017-07-24 22:51 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments
contents of /sys/class/drm/card0/error after bug occurred (2.17 MB, application/octet-stream)
2014-10-17 20:02 UTC, lharrington
no flags Details

Description lharrington 2014-10-17 20:02:50 UTC
Created attachment 108005 [details]
contents of /sys/class/drm/card0/error after bug occurred

Whenever I open my laptop, resuming from suspend, if my HDMI connected external display is powered on, both displays flicker until I restart X. 

I am using the current video drivers available from Intel from their PPA:
  /etc/apt/sources.list.d/intellinuxgraphics.list:
    deb https://download.01.org/gfx/ubuntu/14.04/main trusty main #Intel Graphics drivers

lsb info:
Distributor ID:	Ubuntu
Description:	Ubuntu 14.04.1 LTS
Release:	14.04
Codename:	trusty

relevant dmesg showing before & after:
[ 2267.549136] Watchdog[4645]: segfault at 0 ip 00007f143b948d98 sp 00007f142bf21710 error 6 in libcontent.so[7f143b113000+1170000]
[ 2278.795343] Watchdog[14362]: segfault at 0 ip 00007f9336bffd98 sp 00007f93271d8710 error 6 in libcontent.so[7f93363ca000+1170000]
[ 4038.327596] [drm] stuck on render ring
[ 4038.328237] [drm] GPU HANG: ecode 0:0xf5f7fffe, in ShCrOpenGL [5127], reason: Ring hung, action: reset
[ 4038.328239] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[ 4038.328240] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[ 4038.328241] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[ 4038.328242] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[ 4038.328243] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[ 4040.329884] [drm] Enabling RC6 states: RC6 on, RC6p on, RC6pp off
[ 4788.457392] NET: Unregistered protocol family 15
[ 4788.524731] NET: Registered protocol family 15
Comment 1 Paulo Zanoni 2014-10-20 12:57:55 UTC
Can you reproduce this on every suspend/resume cycle?

Can you please boot with drm.debug=0xe as a Kernel parameter - you can pass it form Grub -, reproduce the problem, then attach the output of "dmesg" here?
Comment 2 Chris Wilson 2014-10-20 13:05:43 UTC
The hang as such isn't particularly interesting (bug 77104). The flickering if it is at subsecond intervals is something different and novel.
Comment 3 lharrington 2014-10-25 16:27:33 UTC
(In reply to Chris Wilson from comment #2)
> The hang as such isn't particularly interesting (bug 77104). The flickering
> if it is at subsecond intervals is something different and novel.

Flickering is definitely sub-second.  Is a video of it helpful?  I'll continue to profile the issue and report back with my findings, after setting the variable suggested by Paulo.

Thank you gentlemen
Comment 4 lharrington 2014-10-25 16:29:00 UTC
(In reply to Paulo Zanoni from comment #1)
> Can you reproduce this on every suspend/resume cycle?
> 
> Can you please boot with drm.debug=0xe as a Kernel parameter - you can pass
> it form Grub -, reproduce the problem, then attach the output of "dmesg"
> here?

I can reproduce every time I resume from suspend with HDMI attached.  I will add the kernel flag and report back with findings.

Thank you
Comment 5 lharrington 2014-12-13 15:31:20 UTC
Had disk failure and ended up with Ubuntu 14.10, which the driver package does not yet support.  How should the bug be closed?  If the Q4 2014 release ends up acting the same I'll open another if necessary.  

Thank you for your help.
Comment 6 Jesse Barnes 2015-03-30 20:42:22 UTC
Ok, please re-open if this is still an issue with current kernels, along with a current crash dump (assuming the crash dump doesn't crash itself!).


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.