AMDGPU crashes and system freezes w/ hardware accelerated video playback

[nveeee]

Using hardware acceleration with SVP and MPV causes MPV to crash and sometimes my entire system freezes forcing a reboot. Switching to the LTS kernel resolves the issue. I'm using an RX 9070 XT.

Here is what journalctl logs look like during the crash:

Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: Dumping IP State Completed
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: ring comp_1.1.0 timeout, signaled seq=3281, emitted seq=3282
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:  Process mpv pid 16419 thread vo pid 16437
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: Starting comp_1.1.0 ring reset
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: reset compute queue (1:1:0)
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:32 vmid:0 pasid:0)
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000000000000000 from client 10
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00040A40
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: CPC (0x5)
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x0
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x4
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu:          RW: 0x1
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: amdgpu: Ring comp_1.1.0 reset succeeded
Dec 07 21:38:26 cachyos kernel: amdgpu 0000:03:00.0: [drm] device wedged, but recovered through reset

[1Naim]

Hi, can you reproduce this on:

1. the latest kernel version (6.18.3)
2. Arch Linux's kernel package

Thanks!

[nveeee]

As of 6.18.3-2 (not sure when exactly it was resolved) this issue no longer occurs.