Created attachment 65044 [details] gdb backtrace from gnome-shell Machine running Debian GNU/Linux sid with Linux 3.4. Video driver: xserver-xorg-video-intel 2:2.20.2-1 libdrm 2.4.33 Video chipset is Intel Graphics HD 4000 (Ivy Bridge GT2) on a Sony Vaio Z. xorg.conf is empty. xserver is 1.12.3. gnome-shell crashed with the attached backtrace just after I played with xrandr settings. I didn't manage to reproduce this so far. Original Debian bug is at http://bugs.debian.org/cgi-bin/bugreport.cgi?archive=yes&bug=683499
FWIW, Julien Cristau think it's caused by "[drm:i915_hangcheck_ring_idle] *ERROR* Hangcheck timer elapsed... render ring idle [waiting on 4043152, at 4043164], missed IRQ?" I have all over my dmesg. Checked /sys/kernel/debug/dri/0/i915_error_state but got "no error state collected" anyway. See http://bugs.debian.org/cgi-bin/bugreport.cgi?archive=yes&bug=683499#27
Kernel guys say: <danvet_> ok, for race-free reset handling you need 3.9 I tried to pin them down on whether the subset of calls made in the path that allocates miptrees would be safe when a reset is about to happen, and that's the best I got. Of course, what you' probably rather have is a fix for whatever's causing the reset in the first place. If xrandr is implicated, updating the 2D driver and kernel are your best bets (there have been many issues with the vblank syncing between those two during a modeset over time) If a current kernel/2D driver still have this problem, drop the NEEDINFO flag. Oh, and if you keep getting the dmesg missed IRQ warning there, the kernel guys really ought to know about that one.
Dear Reporter, This Mesa bug has been in the "NEEDINFO" status for over 60 days. I am closing this bug based on lack of response but feel free to reopen if resolution is still needed. Please ensure you're supplying the correct information as requested. Thank you.
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.