Raven ridge, *ERROR* ring vcn_dec timeout

Submitted by ojab @ojab

Assigned to Default DRI bug account

Link to original bug (#106773)

Description

Created attachment 139954
dmesg

Raven Ridge APU (2400G), no external GPU, linux-4.17-rc6, libdrm-2.4.91, mesa-18.1.0-rc1, xorg-server-1.19.6, xf86-video-amdgpu-18.0.1, libva-2.1.0, mpv-0.28.2.


while playing video using mpv --vo=vaapi --hwdec=vaapi:
[119339.517978] amdgpu 0000:08:00.0: [mmhub] VMC page fault (src_id:0 ring:174 vmid:3 pasid:32770)
[119339.517987] amdgpu 0000:08:00.0: at page 0x0000000115730000 from 18
[119339.517991] amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x0034055D
[119339.517999] amdgpu 0000:08:00.0: [mmhub] VMC page fault (src_id:0 ring:174 vmid:3 pasid:32770)
[119339.518003] amdgpu 0000:08:00.0: at page 0x0000000115730000 from 18
[119339.518007] amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[119339.518015] amdgpu 0000:08:00.0: [mmhub] VMC page fault (src_id:0 ring:169 vmid:0 pasid:0)
[119339.518019] amdgpu 0000:08:00.0: at page 0x0000000000000000 from 18
[119339.518022] amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[119339.518029] amdgpu 0000:08:00.0: [mmhub] VMC page fault (src_id:0 ring:153 vmid:0 pasid:0)
[119339.518032] amdgpu 0000:08:00.0: at page 0x0000000000000000 from 18
[119339.518035] amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[119349.744979] [drm:amdgpu_job_timedout [amdgpu]] ERROR ring vcn_dec timeout, last signaled seq=5760, last emitted seq=5762
[119349.744987] [drm] No hardware hang detected. Did some blocks stall?

nothing graphical except X/mpv is started. I could exit X via Ctrl+C in console and fbdev console becomes accessible, but running X is unkillable and another X cannot be started.

Attachment 139954, "dmesg":
dmegs.txt