Hi Jörg, I am seeing that amdgpu uses amd_iommu_v2 kernel-module. To me this is the last puzzle piece that was missing to explain the dependency of that bug in amdgpu to the iommu.
[cid:[email protected]] Best regards Edgar From: Merger, Edgar [AUTOSOL/MAS/AUGS] Sent: Freitag, 30. Oktober 2020 15:26 To: [email protected] Cc: [email protected] Subject: amdgpu error whenever IOMMU is enabled Hello Jörg, We have developed a Board "mCOM10L1900" that can be populated with an AMD R1305G Ryzen CPU but also with other CPUs from AMD´s R1000 and V1000 Series. Please see attached datasheet. With one board we have a boot-problem that is reproducible at every ~50 boot. The system is accessible via ssh and works fine except for the Graphics. The graphics is off. We don´t see a screen. Please see attached "dmesg.log". From [52.772273] onwards the kernel reports drm/amdgpu errors. It even tries to reset the GPU but that fails too. I tried to reset amdgpu also by command "sudo cat /sys/kernel/debug/dri/N/amdgpu_gpu_recover". That did not help either. There is a similar error reported here https://bugzilla.kernel.org/show_bug.cgi?id=204241. However the applied patch should have already been in the Linux-kernel we use, which is "5.4.0-47-generic". Also found that "amdgpu_info" shows different versions at "SMC firmware version". Please see attachment. It is still unclear what role the IOMMU plays here. Whenever we turn it off, the error does not show up anymore on the failing board. Best regards Edgar
_______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
