You are not logged in.
Hello,
I am experiencing (semi-)random graphical glitches and crashes on my PC running Arch Linux and would appreciate any help with it.
The system is using an ASUS B650-PLUS WIFI mainboard with an AMD Ryzen 5 7600X processor, using the CPU's integrated graphics with no dedicated graphics card¹.
Two kinds of issues occur sporadically:
1. Graphical glitches across the entire screen. This issue usually only occurs after using specific software, e.g. watching video content in Chromium. However, this does not always happen. Passing the flags --disable-gpu --disable-gpu-driver-bug-workarounds when starting Chromium seems to prevent these from occuring. After this problem is triggered, it affects the entire screen including other applications, even after the "triggering" application is closed, and persists until reboot; restarting applications and Xorg does not help. When this issue occurs, there do not appear to be any unusual entries in either dmesg, journalctl, or the application's output. Some screenshots of this issue in a browser: https://0x0.st/8htJ.png and glxgears: http://0x0.st/8htt.png
2. Freezing of the screen for a few seconds followed by a crash of the amdgpu driver (resulting in an Xorg crash/restart). As far as I can tell this issue occurs entirely at random (and indepdently of the first issue), I have not found any specific activities to trigger it. When this occurs, messages like the following are logged in dmesg:
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: Dumping IP State
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: Dumping IP State Completed
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=7645, emitted seq=7647
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: Process information: process glxgears pid 2357 thread glxgears:cs0 pid 2358
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: Starting gfx_0.0.0 ring reset
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: Ring gfx_0.0.0 reset failure
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: GPU reset begin!
Aug 06 23:16:05 sdesktop kernel: amdgpu 0000:0b:00.0: amdgpu: MODE2 reset
[...]A more complete log example can be found here: https://0x0.st/8htQ.txt
I have tried multiple solutions I found in the wiki, in this forum, in issue trackers and on Reddit that mention "gfx_0.0.0 ring reset" issues (even though most of them appear to be with dedicated AMD graphics cards rather than iGPU), including adding various parameters like amdgpu.ppfeaturemask=0xf7fff and others to the kernel parameters, but none of them seem to have helped.
If anyone would be able to help, I would very much appreciate it.
¹ A separate graphics card was installed for use by a VM, but was not used by the host system and the issues occur even if the graphics card is physically removed.
Offline