I'm hitting this for the second time today.  Not sure what the trigger is - but I was using firefox heavily and for certain reasons I had to disable addblocking so there was unusually much flash activity than usual and flash in my experience has the potencial of causing all sorts of ill side effect including the plague.

Sympthom are that new application windows, pulldown menus in firefox or other apps are rendered all black, same for the URL completion widget in firefox. An existing gnome-terminal window was getting rendered black with the bar at the top being rendered transparent after I tried to resize the window. There are the following kernel messages:


Aug 26 12:16:12 h7.dl5rb.org.uk kernel: perf interrupt took too long (2532 > 2500), lowering kernel.perf_event_max_sample_rate to 50000
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] stuck on render ring
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] stuck on blitter ring
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] GPU HANG: ecode 0:0xf4e9fffe, in Xorg [861], reason: Ring hung, action: reset
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm] GPU crash dump saved to /sys/class/drm/card0/error
Aug 26 12:25:19 h7.dl5rb.org.uk kernel: [drm:i915_context_is_banned] *ERROR* gpu hanging too fast, banning!
Aug 26 12:25:21 h7.dl5rb.org.uk kernel: [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off

Will attach the dump of /sys/class/drm/card0/error later - the all black file widget in firefox makes that impossible right now.