This crash was detected running Ubuntu 3.13.6-031306-generic and nightly i915 X driver: # apt-cache show xserver-xorg-video-intel Package: xserver-xorg-video-intel Priority: optional Section: x11 Installed-Size: 2734 Maintainer: Ubuntu Developers <ubuntu-devel-discuss@lists.ubuntu.com> Architecture: amd64 Version: 2:2.99.910+git1403181930.15be6b~gd~s Provides: xorg-driver-video ....
Created attachment 96066 [details] Xorg log X also crashed, please see the attached X.org.log
Created attachment 96067 [details] /sys/class/drm/card0/error dump This is the dump of /sys/class/drm/card0/error and the logs in dmesg: [24510.522243] [drm] stuck on bsd ring [24510.522252] [drm] GPU crash dump saved to /sys/class/drm/card0/error [24510.522254] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [24510.522255] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [24510.522256] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [24510.522259] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [24510.526322] [drm:i915_set_reset_status] *ERROR* bsd ring hung flushing bo (0x1a0dc000 ctx 0) at 0x20947c [24541.536417] [drm] stuck on bsd ring [24541.536506] [drm:i915_set_reset_status] *ERROR* bsd ring hung inside bo (0x1baed000 ctx 0) at 0x1baed354 [24547.527156] [drm] stuck on render ring [24547.527214] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0x7248000 ctx 1) at 0x72482a4 [24553.553904] [drm] stuck on render ring [24553.553955] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0xf296000 ctx 1) at 0xf296270 [24553.553958] [drm:i915_context_is_banned] *ERROR* context hanging too fast, declaring banned! [24555.213923] [drm:intel_dp_start_link_train] *ERROR* too many voltage retries, give up The computer is a Asus UX-31A and X is running with glamor acceleration: Section "Device" Identifier "Intel Graphics" Driver "intel" Option "AccelMethod" "glamor" Option "TearFree" "true" EndSection
Please let me know if more details are needed.
Actually looks like libva. Don't use glamor if you want X to not crash.
Do you use hardware accelerated video decoding, encoding or video processing ?
I general I have libva decoding enabled but in this case no, I was trying to get HTML5 + gstreamer working in Firefox Aurora. Aurora does not enable the vaapi sink yet. With sna enabled in X11 I get weird corruption with HTML5 videos and Firefox, but glamor works fine. In this case the video was fullscreen and it just finished. I general the system has been quite stable, this was the only crash I have seen in the last month.
It is weird. According to your dmesg, libva is involved. Could you double check the issue can be reproduced withou libva (e.g. remove libva package totally).
Unfortunately I've tried to reproduce this but I never got the crash again. I'm pretty sure Firefox doesn't yet use vaapi (https://bugzilla.mozilla.org/show_bug.cgi?id=894372) which makes the libva crash really weird.
Ok, closing as wtf?, thanks for your reporting and please reopen if it pops up again. Usually these bugs do that right after ignoring them like this ;-)
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.