Created attachment 99535 [details] crash dump dmesg asked me to file a bug report: [ 833.694844] [drm] stuck on render ring [ 833.694850] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 833.694852] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 833.694853] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 833.694855] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 833.694856] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
The batch is pure garbage.
It seems to have got stuck at MI_SET_CONTEXT before the batch. But I guess if the next batch is junk earlier batches could have been junk too. The context itself looks somewhat sane to me, but the kernel is buggy and only dumped the first 1KB of it. If you can reproduce this with a more recent kernel we might have a better look at the offending context.
For the batch to be garbage means that either the GPU is overwriting random chunks of memory, or that someone is scribbling asynchronously through the GTT. Either way, it is usually userspace that screwed up.
I guess we can't do anything with this unless there's more information. So I'll just close it.
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.