protonscr

amdgpu: Page fault while starting steam

steamclosed AMD drivers
ValveSoftware/steam-for-linux#6268 · opened 2019-05-14 by pstn · updated 2019-06-27 · 4 comments · github
Ppstn 2019-05-14 github

Your system information

  • Steam client version (build number or date): 1556574584
  • Distribution (e.g. Ubuntu): nixos
  • Opted into Steam client beta?: No
  • Have you checked for system updates?: Yes

Please describe your issue in as much detail as possible:

When starting steam with amdgpu on linux 5.1.0 it crashes my xorg with a GPU reset.
Log:

May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x0020153D
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: [gfxhub] no-retry page fault (src_id:0 ring:158 vmid:2 pasid:32775, for process vulkandriverque pid 28248 thread vulkandriverque pid 28248)
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0:   in page starting at address 0x0000000000000000 from 27
May 14 18:57:30 tuxbase kernel: amdgpu 0000:08:00.0: VM_L2_PROTECTION_FAULT_STATUS:0x00000000
May 14 18:57:41 tuxbase kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=1818717, emitted seq=1818719
May 14 18:57:41 tuxbase kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process vulkandriverque pid 28248 thread vulkandriverque pid 28248
May 14 18:57:41 tuxbase kernel: amdgpu 0000:08:00.0: GPU reset begin!

It seems to happen during this phase of the steam launch (it's the last thing I see on the screen before the GPU crashes)

Fossilize INFO: Overriding serialization path: "/home/philipp/.local/share/Steam/shader_cache_temp_dir_vk_64/steamapprun_pipeline_cache".
amdgpu: radv_amdgpu_cs_query_fence_status failed.
Fatal : VkResult is "TIMEOUT" in /home/pgriffais/src/Vulkan/base/VulkanDevice.hpp at line 548
vulkandriverquery: /home/pgriffais/src/Vulkan/base/VulkanDevice.hpp:548: void vks::VulkanDevice::flushCommandBuffer(VkCommandBuffer, VkQueue, bool): Assertion `res == VK_SUCCESS' failed.

Steps for reproducing this issue:

  1. Install most recent nixos with kernel 5.1.0
  2. start steam
  3. crash
Kkisak-valve maintainer 2019-05-14 github

Hello @pstn, this reads like a video driver issue and should also be mentioned to your video driver vendor if it hasn't been already.

Ppstn 2019-05-14 github

Yes, I forgot to mention that. It indeed looks like a driver bug, but it also seems to have something to do with Fossilize and might be of interest to it's developers.

Ppstn 2019-05-26 github

It turned out to be 100% a driver issue without any weird invocations from steam.

Ppstn 2019-06-27 github

Since people were asking: This is an issue with mesa 18.x. Upgrade to mesa 19.x to fix.

Nothing extracted yet.