I found something in the dmesg log right after the crash of FG
- Code: Select all
kernel: [drm] PCIE gen 3 link speeds already enabled
kernel: amdgpu 0000:01:00.0: amdgpu: PCIE GART of 256M enabled (table at 0x000000F400000000).
kernel: i915 0000:00:02.0: cannot be used for peer-to-peer DMA as the client and provider (0000:01:00.0) do not share an upstream bridge or whitelisted host bridge
kernel: amdgpu 0000:01:00.0: amdgpu: GPU fault detected: 146 0x0006200c
kernel: amdgpu 0000:01:00.0: amdgpu: VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x00000000
kernel: amdgpu 0000:01:00.0: amdgpu: VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0602000C
kernel: amdgpu 0000:01:00.0: amdgpu: VM fault (0x0c, vmid 3) at page 0, read from '' (0x00000000) (32)
kernel: i915 0000:00:02.0: cannot be used for peer-to-peer DMA as the client and provider (0000:01:00.0) do not share an upstream bridge or whitelisted host bridge
kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=77, emitted seq=78
kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process fgfs pid 901 thread fgfs:cs0 pid 907
kernel: amdgpu 0000:01:00.0: amdgpu: GPU recovery disabled.
Hooray wrote in Fri Apr 23, 2021 7:46 pm:To see if it's OSG specific, try any of the OSG examples/demos (or osgviewer/fgviewer standalone).
However, first of all, ensure that other OpenGL software works properly.
There's also dedicated OpenGL benchmarks that you could try, i.e. other than just running glxgears
good idea. chose gfxbench because I've already used it and it is quickly installed for me via AUR. starting a openGL benchmark works on the intel side. using DRI_PRIME=1 ends in a freeze similar to FG.
so a driver problem after all? openGL is working with my intel card ? i am confused what to do next.