Created attachment 113655 [details] Contents of /sys/class/drm/card0/error The DRM driver in the kernel hangs at gnome-shell startup, resets the chip & then seems to carry on fine after that. $ dmesg ... [ 31.021308] [drm] stuck on render ring [ 31.025961] [drm] GPU HANG: ecode 5:0:0xfdffffdd, in gnome-shell [2833], reason: Ring hung, action: reset [ 31.025965] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 31.025966] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 31.025967] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 31.025968] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 31.025969] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 31.041341] drm/i915: Resetting chip after gpu hang I see this bug both with the stock Debian kernel (a patched 3.16) and with 3.19. $ uname -a Linux bill 3.19.0-trunk-amd64 #1 SMP Debian 3.19-1~exp1 (2015-02-12) x86_64 GNU/Linux Upgrading to the latest userspace driver doesn't seem to have helped, although I haven't tried compiling from git yet: $ dpkg -s xserver-xorg-video-intel Version: 2:2.99.917-1~exp1
The actual cause of hang is a little dubious as there is no command barrier between the last batch and the dieing MI_FLUSH i.e. it may either be the batch or the flip itself causing the hang. However, since the batch itself is not recorded (the seqno convinced hangcheck that it is not at fault - be suspicious!) all we have to ponder is whether the flip could have died.
There were improvements pushed in kernel and Mesa that will benefit to your system, so please re-test with latest kernel & Mesa to see if this issue is still occurring.
(In reply to yann from comment #2) > There were improvements pushed in kernel and Mesa that will benefit to your > system, so please re-test with latest kernel & Mesa to see if this issue is > still occurring. Timeout. Assuming that it is fixed by now. If this is not the case, please re-test with latest kernel & Mesa (12-13) to see if this issue is still occurring since there were improvements pushed in kernel and Mesa that will benefit to your system, and fill a new bug.
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.