https://bugzilla.kernel.org/show_bug.cgi?id=221297
--- Comment #2 from lobsterman ([email protected]) --- Reporting results of kernel testing as requested. Hardware: ASUS Turbo Radeon AI Pro R9700 32GB vBIOS: 115-G287BP00-100 OS: Ubuntu 24.04 ROCm: 7.2.1 AMDGPU driver: 6.16.13 Current kernel: 6.17.0-19 SMU MISMATCH (persists across all tested configurations): amdgpu 0000:2d:00.0: amdgpu: smu driver if version = 0x0000002e (46) amdgpu 0000:2d:00.0: amdgpu: smu fw if version = 0x00000032 (50) amdgpu 0000:2d:00.0: amdgpu: smu fw version = 0x00684b00 (104.75.0) amdgpu 0000:2d:00.0: amdgpu: SMU driver if version not matched KERNEL TESTING RESULTS: Kernel 6.17.0-19: - SMU mismatch confirmed present - Fan does not spin under any load - GPU reached 109C and thermally throttled during AI training with fan physically stationary Kernel 6.18.20 mainline: - AMDGPU 6.16.13 DKMS fails to build against this kernel - NVIDIA 535 DKMS fails to build against this kernel - Kernel panics on boot: VFS: Unable to mount root fs on unknown-block(0,0) - NVMe storage drivers absent from mainline build - Testing blocked — system unbootable Kernel 6.19.10 mainline: - Same DKMS build failures as 6.18.20 - Kernel panic on boot - Testing blocked — system unbootable CONCLUSION: Mainline kernel testing on 6.18.20 and 6.19.10 is not feasible on this hardware due to VFS/NVMe boot failures in the minimal mainline builds. The SMU interface version mismatch (driver if version 46 vs firmware if version 50) persists on kernel 6.17 with AMDGPU 6.16.13 under ROCm 7.2.1. A fix in the AMDGPU driver to support SMU interface version 50 is required. Mainline kernel testing cannot be used to validate this fix on this system. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
