Comment 43 for bug 2039868

I have a similar problem, a cascade of errors that typically starts with one of these:
[drm:amdgpu_job_timedout [amdgpu]] *ERROR* ring gfx timeout, signaled seq=1093546, emitted seq=1093548

This used to occur only when playing Dauntless, and only after my MSI Radeon RX580 ran hot for a while. Warframe never crashed. Totally different methods of running the games (Dauntless=Lutris and Epic Games Store, Warframe = Steam and Proton). Something then changed after one of the updates within the last month, and now it crashes on both Warframe and Dauntless well before the card is at a high temp. Basically can't run more than about 5 minutes.

I was running Ubuntu 18.04, so I figured maybe a newer kernel would fix this, but updating to 20.10 did nothing but waste a couple of days of reloading everything.

System: Ryzen 5 3600 on Gigabyte x570 UD with a MSI Radeon RX580 8GB

I'm willing to work with whoever sending whatever info/logs are necessary to get this fixed.