I was watching a movie using mplayer. At some point the whole screen froze. However, the audio was still working. I could switch to the text console and kill mplayer from there. When I switched back it was normal. I also saw following messages in dmesg: [80349.833880] [drm] stuck on render ring [80349.833881] [drm] GPU crash dump saved to /sys/class/drm/card0/error [80349.833882] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [80349.833882] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [80349.833882] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [80349.833882] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [80349.853689] [drm:i915_set_reset_status] *ERROR* render ring hung inside bo (0xddf1000 ctx 1) at 0xddf1004 This is Intel card in T440p laptop. I am using an external LCD monitor connected to the display port of the docking station. Kernel version: 3.13.6 Drivers: intel-dri 10.1.0 libva-intel-driver 1.2.2 xf86-video-intel 2.99.910
Created attachment 95773 [details] /sys/class/drm/card0/error
Hmm, somebody is overwriting memory with wild writes. This is the second report of that. Can you try downgrading to mesa 10.0?
I could. I would have to downgrade all of the following packages: intel-dri-10.0.3 mesa-10.0.3 mesa-libgl-10.0.3 However, I have had mesa 10.1 for the last week (although I don't think I rebooted until today) and this is the first time anything like that happened. So I may not be able to tell if this actually changes anything. Maybe I will just keep the new mesa and see if it happens again (and when) and then do some tests.
I kind of forgot about it. I kept having the same problem (as well as some weird glitches/artifacts on the screen from time to time) with mesa 10.1.0. So after a while I downgraded mesa, mesa-libgl and intel-dri to 10.0.3 version. I haven't seen this problem ever since. The only thing that looks weird (I'm pretty sure it was the same with 10.1.0) is "[drm:intel_dp_complete_link_train] *ERROR* failed to train DP, aborting" message in dmesg. I am not sure what that means but everything seems fine.
Ok, looks like the mesa upgrade resolved this bug, thanks for reporting back. The dmesg error is about DP, but as long as everything works out for you not really any need to file a new bug. But if you have issues with displays please do so.
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.