Created attachment 104234 [details]
/sys/class/drm/card0/error

I was scrolling in gvim on Thinkpad x220 and everything became slow. New X windows are completely black. It happened only 1 time so far but I upgraded to the kernel to 3.16-rc6-amd64 (from Debian packages) only yesterday. It looks similar to Bug #79248 but it's a different ecode.

dmesg:

[22130.610086] [drm] stuck on render ring
[22130.610095] [drm] stuck on blitter ring
[22130.610624] [drm] GPU HANG: ecode 0:0xf4e9fffe, in Xorg [1507], reason: Ring hung, action: reset
[22130.610628] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[22130.610630] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[22130.610633] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[22130.610635] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[22130.610636] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[22130.611134] [drm:i915_context_is_banned] *ERROR* gpu hanging too fast, banning!
[22132.610790] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off

/sys/class/drm/card0/error:

GPU HANG: ecode 0:0xf4e9fffe, in Xorg [1507], reason: Ring hung, action: reset
Time: 1407423560 s 878412 us
Kernel: 3.16-rc6-amd64
Active process (on ring render): Xorg [1507]
Active process (on ring blt): Xorg [1507]
(etc., see attached file)