Created attachment 123846 [details] /sys/class/drm/card0/error From time to time, usually every day or so, I have a GPU hang. I can notice, since for some second, I can't do anything. Inspecting dmesg's output indeed indicates that they're GPU hangs. This is one example. It's (32 bit) Ubuntu 15.10. It seems - according to dmesg - always in the 'compiz' process.
Created attachment 123847 [details] dmesg output
Assigning to Mesa product (please let me know if I am mistaken with this GPU Hang). From this error dump, hung is happening in render ring batch with active head at 0x294c46f8, with 0x7a000003 (PIPE_CONTROL) as IPEHR. Batch extract (around 0x294c46f8): 0x294c46b8: 0x78090003: 3DSTATE_VERTEX_ELEMENTS 0x294c46bc: 0x02000000: buffer 0: valid, type 0x0000, src offset 0x0000 bytes 0x294c46c0: 0x11110000: (X, Y, Z, W), dst offset 0x00 bytes 0x294c46c4: 0x06000000: buffer 1: valid, type 0x0000, src offset 0x0000 bytes 0x294c46c8: 0x11110000: (X, Y, Z, W), dst offset 0x00 bytes 0x294c46cc: 0x7b001c04: 3DPRIMITIVE: quad list sequential 0x294c46d0: 0x00000004: vertex count 0x294c46d4: 0x00000000: start vertex 0x294c46d8: 0x00000001: instance count 0x294c46dc: 0x00000000: start instance 0x294c46e0: 0x00000000: index bias 0x294c46e4: 0x7a000003: PIPE_CONTROL 0x294c46e8: 0x00100002: no write, cs stall, stall at scoreboard, 0x294c46ec: 0x00000000: destination address 0x294c46f0: 0x00000000: immediate dword low 0x294c46f4: 0x00000000: immediate dword high 0x294c46f8: 0x7a000003: PIPE_CONTROL 0x294c46fc: 0x00004000: qword write, 0x294c4700: 0x068fa004: destination address 0x294c4704: 0x00000000: immediate dword low 0x294c4708: 0x00000000: immediate dword high 0x294c470c: 0x78021302: 3DSTATE_SAMPLER_STATE_POINTERS: VS mod 1, GS mod 1, PS mod 1 0x294c4710: 0x00000000: VS sampler state 0x294c4714: 0x00000000: GS sampler state 0x294c4718: 0x00007000: WM sampler state
*** Bug 97591 has been marked as a duplicate of this bug. ***
Please test a new version of Mesa (12 or 13) and mark as REOPENED if you can reproduce and RESOLVED/* if you cannot reproduce.
(In reply to yann from comment #4) > Please test a new version of Mesa (12 or 13) and mark as REOPENED > if you can reproduce and RESOLVED/* if you cannot reproduce. Timeout, so assuming this is fixed by now. According to mesa engineers, mesa only emits 3DSTATE_VERTEX_ELEMENTS on-demand right before 3DPRIMITIVE. Chris has changed SNA to emit a dummy primitive between VertexElements in 4acd4a7d3d2f41227022fa7581cfb85a0b124eae in xf86-video-intel (thanks to https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=4acd4a7d3d2f41227022fa7581cfb85a0b124eae) So if this is still occuring, I recommend to update to latest version of xf86-video-intel and reopen this bug if it is still present then. An alternate solution is to use Glamor/modesetting.
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.