Created attachment 117754 [details]
Kernel log of lockup
Using Kubuntu vivid with linux kernel 4.2-rc7 (from ubuntu mainline repo), and mesa/xorg from the oibaf ppa with a self-compiled xf86-video-amdgpu driver I've been getting random GPU lockups. When the lockup happens the computer seems to get stuck for a few seconds, then I get a black screen and a few seconds later the screen reappears, and I am able to move the mouse, but can't click anything. If I go to a tty and restart X it seems to resort to using llvmpipe, instead of amdgpu, and that works fine (without any acceleration).
If it's relevant, I am using a multi-monitor setup.
I have attached a kernel log of the lockup, and it looks to me like what's causing the lockup is this:
[ 2481.663794] amdgpu 0000:01:00.0: ring 9 stalled for more than 10000msec
[ 2481.663798] amdgpu 0000:01:00.0: GPU lockup (waiting for 0x0000000000013658 last fence id 0x0000000000013657 on ring 9)
[ 2481.663804] amdgpu 0000:01:00.0: failed to sync rings (-35)
Haven't been able to reproduce this using latest mesa and drm from git, and drm-next-4.3-wip (commit 9066b0c318589f47b754a3def4fe8ec4688dc21a) kernel, so I'm going to mark this as resolved.