AMDGPU randomly crashes my pc

Ask Fedora
, ,
1

Hey guys and girls,

i encountered some problems with my current system after swapping my Distro from Manjaro to Fedora.

Overall im really happy that i took the time to swap distros and feeling very comfortable with Fedora.
But since a few weeks ago now i had some really annoying crashes on my system, related to my amdgpu driver i guess.
I’m playing some Games on Steam and some with Lutris but never had problems.
My game that caused these problems is “Ring of Pain” on Steam. So far its the only Game/Program that can be related to this crash and never had any problems with other titles so far.

It’s does not matter if i swap out my proton versions to a far older build or using custom ones like the GE Version, it still happens on a frequent period of time.

My Error Code :

Apr 18 08:05:33 fedora kernel: [drm:amdgpu_dm_atomic_commit_tail [amdgpu]] *ERROR* Waiting for fences timed out!
Apr 18 08:05:33 fedora kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx_0.0.0 timeout, signaled seq=23000376, emitted seq=23000378
Apr 18 08:05:33 fedora kernel: [drm:amdgpu_job_timedout [amdgpu]] *ERROR* Process information: process Ring Of Pain.ex pid 21151 thread dxvk-submit pid 21222
Apr 18 08:05:33 fedora kernel: amdgpu 0000:09:00.0: amdgpu: GPU reset begin!
Apr 18 08:05:33 fedora kernel: amdgpu 0000:09:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_2.1.0 test failed (-110)
Apr 18 08:05:33 fedora kernel: [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* KGQ disable failed
Apr 18 08:05:34 fedora kernel: amdgpu 0000:09:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_2.1.0 test failed (-110)
Apr 18 08:05:34 fedora kernel: [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* KCQ disable failed
Apr 18 08:05:34 fedora kernel: [drm:gfx_v10_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
Apr 18 08:05:34 fedora kernel: [drm] free PSP TMR buffer
Apr 18 08:05:34 fedora kernel: amdgpu 0000:09:00.0: amdgpu: BACO reset
Apr 18 08:05:34 fedora kernel: amdgpu 0000:09:00.0: amdgpu: SMU: response:0xFFFFFFFF for index:25 param:0x00000000 message:ExitBaco?
Apr 18 08:05:34 fedora kernel: amdgpu 0000:09:00.0: amdgpu: Failed to exit BACO state!
Apr 18 08:05:34 fedora kernel: amdgpu 0000:09:00.0: amdgpu: ASIC reset failed with error, -121 for drm dev, 0000:09:00.0

What happens exactly :

After some time in the game, its freezing randomly, sound is missing or running in a loop and my system is not responsive any longer. If im lucky my system will reboot on his on, or i have to reset it manually to reboot.
Even if i left my Pc idle while this game is running and nothing happens, it crashes but at a total random moment. Its not related to a animation, process or event that happens in game its just crashes without any reason.

I really hope someone can help me out. Thanks a lot for your time and i will answer as soon as possible.

2

Since this is happening with steam and a particular game it seems like it is not a fedora issue but rather a steam issue. While it may be related to how steam acts within fedora, we cannot know that unless the same happens (or not) with steam and other distros.

I suggest you contact steam about this specific issue.

It also may be related to how you installed steam. Did you install the flatpak? or did you install it from the rpmfusion-nonfree-steam repo?

1 Like
3

In fact i tried to compared my old Manjaro Distro and PoP OS! to get something to work with, but my Game isn’t crashing at any moment there. I trying to solve this problem for like 4 weeks now so i tried a lot of things which didn’t worked out so far.

I also tried both ways, either Flatpak or the Discovery version and both gave me the same result. Both are crashing my system with the same error code.
Also i searched a lot with google, viewed a lot of threads with compliments about the amddriver which is shipping in fedora since you cant install amdpro driver that easy, if at all.

That’s where my concerns came from, i think its an issue with the amd driver in fedora.

Sadly i already contacted the Steam support, but they could not help me in particular and they said i should update or change my driver version. As well, contacting the Fedora Team.

4

This looks like *ERROR* ring gfx_0.0.0 timeout when using firefox, chrome or icaclient when dpm performance level = auto (#1974) · Issues · drm / amd · GitLab. I have the same problem with my AMD GPU on Fedora, but I don’t think its Fedora-specific.