[Vega10][bisected] Vega56 VM_L2_PROTECTION_FAULT when logging into KDE with kernel 5.3.0-rc1 and newer

Submitted by Andreas

Assigned to Default DRI bug account

Link to original bug (#111948)

Description

Created attachment 145692
dmesg for 5.3.0-rc1

Hi,

I'm getting this problem when I log into KDE (5.16.5) if I use kernel 5.3.0-rc1 or newer including 5.4.0-rc2, it happens every time, the splash show for a few seconds and then the screen goes black and it comes back garbled.

I've done a bisect between 5.2 and 5.3.0-rc1:

52d2d44eee8091e740d0d275df1311fb8373c9a9 is the first bad commit

This is a merge commit though...

[ 19.840654] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:144 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840656] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a021000 from 27
[ 19.840657] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00101121
[ 19.840661] amdgpu 0000:0a:00.0: [gfxhub] retry page fault (src_id:0 ring:0 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840662] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a036000 from 27
[ 19.840662] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840665] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840666] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a037000 from 27
[ 19.840666] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840670] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840670] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a035000 from 27
[ 19.840671] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840674] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840675] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a035000 from 27
[ 19.840675] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840679] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840679] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a035000 from 27
[ 19.840680] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840683] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:128 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840683] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a025000 from 27
[ 19.840684] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840690] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840690] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a05b000 from 27
[ 19.840691] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840694] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840695] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a05b000 from 27
[ 19.840695] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 19.840698] amdgpu 0000:0a:00.0: [gfxhub] no-retry page fault (src_id:0 ring:157 vmid:1 pasid:32769, for process X pid 4118 thread X:cs0 pid 4169)
[ 19.840699] amdgpu 0000:0a:00.0: in page starting at address 0x000080010a058000 from 27
[ 19.840699] amdgpu 0000:0a:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
[ 25.066367] [drm:amdgpu_dm_atomic_commit_tail] ERROR Waiting for fences timed out or interrupted!
[ 25.066369] [drm:amdgpu_dm_atomic_commit_tail] ERROR Waiting for fences timed out or interrupted!
[ 30.194884] [drm:amdgpu_job_timedout] ERROR ring sdma0 timeout, signaled seq=491, emitted seq=493
[ 30.194886] [drm:amdgpu_job_timedout] ERROR Process information: process ksplashqml pid 4333 thread ksplashqml:cs0 pid 4336
[ 30.194889] amdgpu 0000:0a:00.0: GPU reset begin!
[ 30.195169] [drm:amdgpu_job_timedout] ERROR ring gfx timeout, but soft recovered

I've attached the dmesg log.

Attachment 145692, "dmesg for 5.3.0-rc1":
dmesg_5.3.0_rc1.log