Bug 95145

Summary: GPU HANG: ecode 9:0:0x84dffff8, in X [1481], reason: Ring hung, action: reset
Product: DRI Reporter: Robert <bugs.freedesktop.org>
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: awaters1, harry_x, intel-gfx-bugs, schlaganfall.patient
Version: XOrg git   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: SKL i915 features: GPU hang
Attachments:
Description Flags
Xorg.0.log
none
/sys/class/drm/card0/error
none
/sys/class/drm/card0/error none

Description Robert 2016-04-26 09:15:15 UTC 
Created attachment 123268 [details]
Xorg.0.log

Steps to reproduce for me:
1. Start X
2. Start Eclipse
3. About 10-20s after starting Eclipse the error occurs

Used software versions:
Gentoo Linux
Kernel 4.5.2 / x86_64 Intel(R) Core(TM) i5-6440HQ CPU @ 2.60GHz
xorg-server 1.18.3
Intel driver 2.99.917_p20160423
Nesa 11.2.1

dmesg output when error occurs:

[Di Apr 26 11:07:57 2016] [drm] stuck on render ring
[Di Apr 26 11:07:57 2016] [drm] GPU HANG: ecode 9:0:0x84dffff8, in X [1481], reason: Ring hung, action: reset
[Di Apr 26 11:07:57 2016] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[Di Apr 26 11:07:57 2016] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[Di Apr 26 11:07:57 2016] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[Di Apr 26 11:07:57 2016] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[Di Apr 26 11:07:57 2016] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[Di Apr 26 11:07:57 2016] drm/i915: Resetting chip after gpu hang
[Di Apr 26 11:07:59 2016] [drm] RC6 on
[Di Apr 26 11:08:03 2016] [drm] stuck on render ring
[Di Apr 26 11:08:03 2016] [drm] GPU HANG: ecode 9:0:0x84dffff8, in X [1481], reason: Ring hung, action: reset
[Di Apr 26 11:08:03 2016] [drm:i915_set_reset_status] *ERROR* gpu hanging too fast, banning!
[Di Apr 26 11:08:03 2016] drm/i915: Resetting chip after gpu hang
[Di Apr 26 11:08:05 2016] [drm] RC6 on


Relevant Xorg.0.log output:

[     8.693] (II) intel(0): EDID vendor "MEI", prod id 38562
[     8.693] (II) intel(0):     EDID quirk: Use maximum size instead of detailed timing sizes.
[     8.693] (II) intel(0): Using EDID range info for horizontal sync
[     8.693] (II) intel(0): Using EDID range info for vertical refresh
[     8.693] (II) intel(0): Printing DDC gathered Modelines:
[     8.693] (II) intel(0): Modeline "2560x1440"x0.0  257.54  2560 2648 2680 2800  1440 1468 1473 1533 +hsync +vsync (92.0 kHz eP)
[     8.693] (II) intel(0): Modeline "2560x1440"x0.0  257.51  2560 2648 2680 2800  1440 1468 1473 1916 +hsync +vsync (92.0 kHz e)
[     8.989] (II) intel(0): resizing framebuffer to 4480x1440
[     9.015] (II) intel(0): switch to mode 1920x1200@60.0 on DP3-1 using pipe 1, position (2560, 240), rotation normal, reflection none
[   108.709] (EE) intel(0): Failed to submit rendering commands (Input/output error), disabling acceleration.
[   108.709] (EE) intel(0): When reporting this, please include /sys/class/drm/card0/error and the full dmesg.


Full Xorg.0.log and /sys/class/drm/card0/error are attached.
Comment 1 Robert 2016-04-26 09:15:47 UTC 
Created attachment 123269 [details]
/sys/class/drm/card0/error
Comment 2 Robert 2016-04-26 09:17:45 UTC 
I can reproduce the problem with different driver configurations in xorg.conf:

Section "Device"
        Identifier  "Intel Graphics"
        Driver      "intel"
#       Option  "DRI"   "false" 
#       Option      "AccelMethod"  "uxa"
#       Option      "AccelMethod"  "SNA"
EndSection

The problem is reproducible with any of the 3 commented lines uncommented, and also with all 3 commented.
Comment 3 Anthony Waters 2016-05-02 21:40:04 UTC 
I had a similar issue and reverting back to Gentoo's x11-drivers/xf86-video-intel-2.99.917_p20160403 works.  If I have time I will try git bisect.
Comment 4 Thomas 2016-05-14 15:45:08 UTC 
Created attachment 123751 [details]
/sys/class/drm/card0/error

same here

distro: opensuse
uname -rm: 4.5.3-1-default x86_64
xf86-video-intel: 2.99.917.641_ge4ef6e9-1.1
Intel(R) Core(TM) i7-6500U CPU @ 2.50GHz
Comment 5 Thomas 2016-05-15 09:53:30 UTC 
no errors anymore, changed:

Device "intel"

to

Device "modesetting"

in xorg.conf
Comment 6 Robert 2016-05-15 13:38:42 UTC 
(In reply to Thomas from comment #5)
> no errors anymore, changed:
> 
> Device "intel"
> 
> to
> 
> Device "modesetting"
> 
> in xorg.conf

Did you mean 
Driver "modesetting"?
Comment 7 Thomas 2016-05-15 13:41:33 UTC 
sure, entry is

Section "Device"
        Identifier "Intel Graphics"
        Driver "modesetting"
EndSection
Comment 8 yann 2016-09-05 08:26:34 UTC 
*** Bug 95187 has been marked as a duplicate of this bug. ***
Comment 9 yann 2016-09-05 08:33:08 UTC 
(In reply to Robert from comment #6)
> (In reply to Thomas from comment #5)
> > no errors anymore, changed:
> > 
> > Device "intel"
> > 
> > to
> > 
> > Device "modesetting"
> > 
> > in xorg.conf
> 
> Did you mean 
> Driver "modesetting"?

Robert is it working with this change?

Another alternate could be to upgrade xf86-video-intel (repo is located here https://cgit.freedesktop.org/xorg/driver/xf86-video-intel)
Comment 10 Robert 2016-09-05 10:22:58 UTC 
Hasn't occurred anymore for a while. I'm on xf86-video-intel 2.99.917_p2016082 now.

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.