amdgpu randomly hanging, sometimes not all monitors
Hello! I've been trying to get to the root of this problem for several months yet had little luck with knowing how to resolve this due to my lack of knowledge on handling system faults on Linux systems. This issue doesn't seem to happen on a single OS or GPU as I've switched through some during this time. I'll start by going back a bit to see if anyone may notice any similarities or solutions:
The first time this happened was when I was running Garuda for about half a year on my old R9 Fury X GPU. I booted up a new game (Atlyss) and within a couple minutes I had noticed the monitor that was playing the game had completely frozen; yet the game logic and audio were still playing fine. I then realised my second monitor was responding fine. I had no clue how to restart my GPU drivers or even diagnose the issue so I just rebooted and that was that. This problem occurred on a few other titles as well (such as Palworld and Slime Rancher 2); sometimes the first monitor would hang, sometimes both, sometimes the system would completely hang. Regardless. All issues lead to a forced reboot to resolve it. It was around this time I tried to look into how to diagnose the problem though my usual method of "just look up Event Viewer in Windows!" wasn't possible and the closest I could find was the systemd journal which spat out a log I couldn't really find any concrete answers to (possibly due to my lack of understanding.)
Eventually I noticed the issue had also spread to Youtube videos (both in-site and embedded on platforms such as Discord) at random moments. Sometimes the screen I was watching on would freeze, sometimes the whole system would hang for about 20 seconds. Which then also became an issue with watching ANY video (i.e mp4 files in Discord, mkv recordings in VLC, etc.)
I had since upgraded to a new GPU (Radeon 9070XT) and even re-installed the OS so I knew the issue wasn't directly related to the hardware if two different GPUs were experiencing the same problem and would have thought a fresh OS would help alleviate any user error though the problem still persisted. Eventually that (alongside some other arch-based hiccups) lead me to give Bazzite a try and on a VM and everything seemed alright so today I decided I'd have a bit of fun and install Bazzite to bare metal which also went great (albeit the partitioning tool in the installer felt a bit confusing compared to disk management tools), yet I noticed when watching a Youtube video someone sent to me on Discord, around 15 minutes in my screen had frozen again with the second monitor being fine. So it seems the problem persists even on different distros using different bases (I couldn't test this theory on a Debian/Ubuntu-based distro as my GPU is too new for those to function properly.) I decided to look in KDE's Discover app to see if there were any logging tools more suited to my way of troubleshooting and did find GNOME's Logs tool Which spewed out the following error at a time that matched when my monitor froze:
For me this at least lets me know the issue is related to amdgpu which I had I guess it was related to already though I have no clue how to resolve this. Would anyone more knowledgeable with Linux systems be able to point me in the right direction? I would also note that this is my first time using a Fedora-based OS so I would have a beginner level of knowledge on the structure of the OS or commands to use in the terminal. Thanks to anyone who can help! :cocohello:
The first time this happened was when I was running Garuda for about half a year on my old R9 Fury X GPU. I booted up a new game (Atlyss) and within a couple minutes I had noticed the monitor that was playing the game had completely frozen; yet the game logic and audio were still playing fine. I then realised my second monitor was responding fine. I had no clue how to restart my GPU drivers or even diagnose the issue so I just rebooted and that was that. This problem occurred on a few other titles as well (such as Palworld and Slime Rancher 2); sometimes the first monitor would hang, sometimes both, sometimes the system would completely hang. Regardless. All issues lead to a forced reboot to resolve it. It was around this time I tried to look into how to diagnose the problem though my usual method of "just look up Event Viewer in Windows!" wasn't possible and the closest I could find was the systemd journal which spat out a log I couldn't really find any concrete answers to (possibly due to my lack of understanding.)
Eventually I noticed the issue had also spread to Youtube videos (both in-site and embedded on platforms such as Discord) at random moments. Sometimes the screen I was watching on would freeze, sometimes the whole system would hang for about 20 seconds. Which then also became an issue with watching ANY video (i.e mp4 files in Discord, mkv recordings in VLC, etc.)
I had since upgraded to a new GPU (Radeon 9070XT) and even re-installed the OS so I knew the issue wasn't directly related to the hardware if two different GPUs were experiencing the same problem and would have thought a fresh OS would help alleviate any user error though the problem still persisted. Eventually that (alongside some other arch-based hiccups) lead me to give Bazzite a try and on a VM and everything seemed alright so today I decided I'd have a bit of fun and install Bazzite to bare metal which also went great (albeit the partitioning tool in the installer felt a bit confusing compared to disk management tools), yet I noticed when watching a Youtube video someone sent to me on Discord, around 15 minutes in my screen had frozen again with the second monitor being fine. So it seems the problem persists even on different distros using different bases (I couldn't test this theory on a Debian/Ubuntu-based distro as my GPU is too new for those to function properly.) I decided to look in KDE's Discover app to see if there were any logging tools more suited to my way of troubleshooting and did find GNOME's Logs tool Which spewed out the following error at a time that matched when my monitor froze:
amdgpu 0000:0e:00.0: [drm] *ERROR* [CRTC:93:crtc-0] flip_done timed outFor me this at least lets me know the issue is related to amdgpu which I had I guess it was related to already though I have no clue how to resolve this. Would anyone more knowledgeable with Linux systems be able to point me in the right direction? I would also note that this is my first time using a Fedora-based OS so I would have a beginner level of knowledge on the structure of the OS or commands to use in the terminal. Thanks to anyone who can help! :cocohello:
