Created attachment 104424 [details] GPU crash dump Using Xubuntu 14.04 (3.13.0-32-generic #57-Ubuntu SMP Tue Jul 15 03:51:08 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux). Screen freezes and doesn't recover. Mouse pointer can be moved, but no action is possible. Switching to console (Ctrl+Alt+F1) takes some time but finally works. Syslog says: kernel: [41262.418736] [drm] stuck on render ring kernel: [41262.418745] [drm] GPU crash dump saved to /sys/class/drm/card0/error kernel: [41262.418747] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. kernel: [41262.418749] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel kernel: [41262.418751] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. kernel: [41262.418754] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. kernel: [41262.423397] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0x40c8000 ctx 0) at 0x40c961c kernel: [41268.398388] [drm] stuck on render ring kernel: [41274.390087] [drm] stuck on render ring kernel: [41280.405822] [drm] stuck on render ring ...
Please check that you have lastest mesa (10.2.y).
mesa is version 10.1.3-0ubuntu0.1 That's the version provided with Ubuntu 14.04 Are there any fixes in mesa 10.2.x that might by related to my problem?
I'm seeing more and more of these as well. Currently running 3.16.0-031600-generic. When they occur they are tied to the WM process. Let me know if I should open a separate bug report (seems a bit silly for an identical issue). There are also some more serious problems that are different from the one here which I will open new reports for.
Could you please reproduce with a newer version of kernel and preferably with latest drm-intel-nightly branch from cgit.freedesktop.org/drm-intel and attach updated logs and error state?
Sorry for the late comment. I've been testing this on 3.18rc1 (3.18.0-031800rc1-generic #201410192135) for about a week, and I am happy to report that I'm unable to reproduce the problem. I see quite a lot of drm/i915 fixes went into this version, and so far the system has been running much better than on 3.16/3.17. To my surprise I also notice that 23.976Hz refreshrates now appear to be smoother on Ivy Bridge (/HD4000) – although still not perfect, but an improvement nonetheless. (50Hz on one display where the other is running on 60 is still a little jerky though.) The only GPU-related errors in the syslog comes shortly after the system has been booted (presumably when XBMC is started on the second display.) I am including those here for reference, although they do not seem to be critical. Mon Oct 27 01:32:03 2014] [drm:ivybridge_set_fifo_underrun_reporting] *ERROR* uncleared fifo underrun on pipe A [Mon Oct 27 01:32:03 2014] [drm:ivb_err_int_handler] *ERROR* Pipe A FIFO underrun [Mon Oct 27 01:32:03 2014] [drm:cpt_set_fifo_underrun_reporting] *ERROR* uncleared pch fifo underrun on pch transcoder A [Mon Oct 27 01:32:03 2014] [drm:cpt_serr_int_handler] *ERROR* PCH transcoder A FIFO underrun So it appears the original issue has been fixed (hopefully the person who created the bug report can also verify this.)
Lately I was able to reproduce the issue by excessively resizing windows. The problem pretty much happened instantly. After upgrading to 3.14.24-031424-generic #201411141736 I'm unable to reproduce it anymore. So the problem seems to be fixed for me.
(In reply to Philipp Jacob from comment #6) > Lately I was able to reproduce the issue by excessively resizing windows. > The problem pretty much happened instantly. > > After upgrading to 3.14.24-031424-generic #201411141736 I'm unable to > reproduce it anymore. So the problem seems to be fixed for me. Thanks for reporting back, closing this one. To everyone else: If you still have gpu hangs (with latest kernel+mesa and all that) please file a new bug report, you likely have a different issue.
Closing.
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.