Just putting this here for others to find in the future. Updated to Fedora 42 a few days ago. I don’t use the machine, but my son does and he kept saying it was “crashing”. Looked at random errors saying FFMPEG was crashing. Then Freetube caused the desktop to crash. Then steam. Caught this in the logs:

Jun 10 18:42:19 desktop kernel: [drm] scheduler comp_1.0.1 is not ready, skipping  
Jun 10 18:42:19 desktop kernel: [drm] scheduler comp_1.0.4 is not ready, skipping  
Jun 10 18:42:20 desktop kernel: [drm] scheduler comp_1.0.6 is not ready, skipping  

Then saw this:

Jun 11 18:29:53 desktop kernel: amdgpu 0000:01:00.0: amdgpu: ring uvd timeout, signaled seq=77082, emitted seq=77085  
Jun 11 18:29:53 desktop kernel: amdgpu 0000:01:00.0: amdgpu: Process information: process ffmpeg pid 576748 thread ffmpeg:cs0 pid 5767>  
Jun 11 18:29:53 desktop kernel: amdgpu 0000:01:00.0: amdgpu: GPU reset begin!  
Jun 11 18:29:53 desktop kernel: amdgpu 0000:01:00.0: amdgpu: BACO reset  
Jun 11 18:29:54 desktop kernel: amdgpu 0000:01:00.0: amdgpu: GPU reset succeeded, trying to resume  
Jun 11 18:29:54 desktop kernel: [drm] PCIE GART of 256M enabled (table at 0x000000F400900000).  
Jun 11 18:29:54 desktop kernel: [drm] VRAM is lost due to GPU reset!  

Tracked that amdgpu ring uvd timeout to: https://wiki.gentoo.org/wiki/AMDGPU “Flickering and white screens” section.

It recommended checking GPU clock speeds at: https://www.techpowerup.com/gpu-specs/

Then check what the current clock and memory clock were:

cat /sys/class/drm/card0/device/pp_dpm_sclk
cat /sys/class/drm/card0/device/pp_dpm_mclk

Sure enough the GPU was clocked to 1196Mhz, but the max clock was 1175Mkhz. I used LACT to change the clock speed and I’m back to a stable system.