Bug 88777

Summary:

GPU hang (ring hung) with chromium and OpenGL composite manager on Intel 965GM

Product:

DRI

Reporter:

Gustavo Sverzut Barbieri <barbieri>

Component:

DRM/Intel

Assignee:

Intel GFX Bugs mailing list <intel-gfx-bugs>

Status:

CLOSED DUPLICATE

QA Contact:

Intel GFX Bugs mailing list <intel-gfx-bugs>

Severity:

normal

Priority:

medium

CC:

intel-gfx-bugs

Version:

unspecified

Hardware:

x86-64 (AMD64)

OS:

Linux (All)

Whiteboard:

i915 platform:

i915 features:

Attachments:

Description	Flags
dmesg.txt of kernel with drm.debug=0xe	none
/sys/class/drm/card0/error	none
Xorg.log of a working system (before crash)	none

Description Gustavo Sverzut Barbieri 2015-01-24 15:04:25 UTC

Created attachment 112778 [details]
dmesg.txt of kernel with drm.debug=0xe

I've noticed this bug to be reproducible on my system even after recent updates to xorg and mesa.

My system:
 - MacBook 4,1 (Intel(R) Core(TM)2 Duo T8300, video card is 965GM)
 - ArchLinux (kernel 3.18.2-2-ARCH x86_64 SMP PREEMPT)
 - Package versions:
    * xorg-server: 1.16.3-2
    * xf86-video-intel: 2.99.917-1
    * libva-intel-driver: 1.5.0-1
    * libva-mesa-driver: 10.4.2-1 
    * mesa: 10.4.2-1
    * libdrm: 2.4.59-1
    * enlightenment: 0.19.2-1 
    * slim: 1.3.6-5
 - Window Manager: Enlightenment 0.19.2 running OpenGL compositor (also happens with GNOME3, but that often crashes earlier)
 - Login Manager: slim (shouldn't matter but...)

Steps to reproduce:
 1. Login to enlightenment using its OpenGL composite manager;
 2. Visit Chrome Themes page https://chrome.google.com/webstore/category/themes?hl=en
 3. GPU hungs

See attached logs:
 * dmesg.txt: kernel booted with drm.debug=0xe
 * drm_card0_error.txt: /sys/class/drm/card0/error, same as /sys/kernel/debug/dri/0/i915_error_state

Comment 1 Gustavo Sverzut Barbieri 2015-01-24 15:05:20 UTC

Created attachment 112779 [details]
/sys/class/drm/card0/error

Comment 2 Gustavo Sverzut Barbieri 2015-01-24 15:05:54 UTC

Created attachment 112780 [details]
Xorg.log of a working system (before crash)

Comment 3 Chris Wilson 2015-02-03 12:45:59 UTC

Presuming this is a dupe of 88809:

This is likely fixed in Mesa 10.4.3 or master.  Can you try one of those?

 You may also be able to work around the problem by running with
 "always_flush_cache=true".

*** This bug has been marked as a duplicate of bug 88809 ***

Comment 4 Gustavo Sverzut Barbieri 2015-02-03 17:54:28 UTC

I'm traveling, will test next week.

Thanks!

Comment 5 Gustavo Sverzut Barbieri 2015-02-07 20:35:03 UTC

confirmed as solved (mesa 10.4.4)

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.