Created attachment 123131 [details] The GPU crash dump I get a cpu hang on intel. The hang occurs using an OpenSceneGraph based OpenGL application. The system is a fedora 23 x86_64 on the current update level. The desktop used is kde. The kernel used is drm-intel-nightly as of some minutes ago at 9dabb0053 but the hang also occurs with the regular distribution kernel. I will attach the GPU crash dump as mandated by the kernel message and the kernel messages. thanks in advance Mathias
Created attachment 123132 [details] dmesg up to the GPU hang
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 0x017ea974, with 0x7a000003 (PIPE_CONTROL) as IPEHR. Batch extract (around 0x017ea974): 0x017ea93c: 0x7b000005: 3DPRIMITIVE: 0x017ea940: 0x00000001: point list sequential 0x017ea944: 0x00002388: vertex count 0x017ea948: 0x00000000: start vertex 0x017ea94c: 0x00000001: instance count 0x017ea950: 0x00000000: start instance 0x017ea954: 0x00000000: index bias 0x017ea958: 0x78250000: 3DSTATE_DEPTH_STENCIL_STATE_POINTERS 0x017ea95c: 0x00007b01: pointer to DEPTH_STENCIL_STATE at 0x00007b00 (changed) 0x017ea960: 0x7a000003: PIPE_CONTROL 0x017ea964: 0x00006000: qword write, depth stall, 0x017ea968: 0x7e40f000: destination address 0x017ea96c: 0x00000000: immediate dword low 0x017ea970: 0x00000000: immediate dword high 0x017ea974: 0x78150005: 3DSTATE_CONSTANT_VS 0x017ea978: 0x00000004: len 0 = 4, len 1 = 0 0x017ea97c: 0x00000000: len 2 = 0, len 3 = 0 0x017ea980: 0x00007a81: pointer to constbuf 0 0x017ea984: 0x00000000: pointer to constbuf 1 0x017ea988: 0x00000000: pointer to constbuf 2 0x017ea98c: 0x00000000: pointer to constbuf 3
Please test a new version of Mesa (12 or 13) and mark as REOPENED if you can reproduce and RESOLVED/* if you cannot reproduce. If you can reproduce, please capture and upload an apitrace (https://github.com/apitrace/apitrace) so that we can easily reproduce as well.
Using Fedora 24 it seems to work. So one of the components kernel/mesa on the way there has fixed this.
(In reply to Mathias Fröhlich from comment #4) > Using Fedora 24 it seems to work. > So one of the components kernel/mesa on the way there has fixed this. thanks Mathias, closing bug then
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.