Created attachment 128353 [details] GPU crash dump from /sys/class/drm/card0/error Hi I found this after an hibernation wakeup on my machine, so here is the requested bug. Thanks [34699.708021] [drm] stuck on bsd ring [34699.709270] [drm] GPU HANG: ecode 4:1:0x01000000, reason: Ring hung, action: reset [34699.709273] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [34699.709274] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [34699.709275] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [34699.709276] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [34699.709277] [drm] GPU crash dump saved to /sys/class/drm/card0/error [34699.709348] drm/i915: Resetting chip after gpu hang
The order of the requests in BSD look reversed. Not much to say but please do test a later kernel - this could be any one of many bugs fixed wrt requests.
I had the same problem. Try kernel 4.8 or later. Kernels 4.3 to 4.7 don't seem to work properly with older Intel CPUs/GPUs, for instance Ironlake/Clarkdale (see: https://bugs.freedesktop.org/show_bug.cgi?id=97396). If you want to stick to 4.4 (and not update the CPU), you could try a patch that I have created and used with 4.4.22/38 for months without problems (https://bugs.freedesktop.org/show_bug.cgi?id=99737). I have applied it to 4.4.47 yesterday and it works there, too.
Marking this as resolved+fixed. On 4.3 with patch (see comment 2) and on 4.8 onward on drm-tip. Please reopen this if you disagree. Jean-Christophe, please mark this as verified and closed after checking fix with patch on 4.3 or by using the latest kernel from drm-tip.
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.