Created attachment 115386 [details] GPU crash dump (First occasion with 4.0 kernel running 2 weeks) [drm] GPU HANG: ecode 3:0:0x0071e0c0, in Xorg [2106], reason: Ring hung, action: reset
The crash occurs on the instruction to set a colour buffer (with Y-tiling) that is currently being fenced with X. It is used earlier in the batch as a Y-tiled texture. The hang could be at the pipeline stall at the end of that sampling - hard to say for sure. At any rate, that we have a dirty fence register seems to be the most likely explanation.
*** This bug has been marked as a duplicate of bug 90841 ***
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.