How’s everyone weekend going? I’ve been dealing with this issue for a little while so I thought I would finally look into it and make a post about it.
Sometimes when I wake my PC from sleep, it fails to resume correctly. I’m able to log in from the lock screen just fine, but when I get to my desktop there’s usually some graphical glitches like medium-sized black boxes spanning across different application windows. The next thing I’ll notice is that if I try to close windows or interact with the desktop, it will lock up to the point I can’t use my mouse or keyboard. Trying to switch to another TTY terminal doesn’t even work at that point. If I switch to a TTY quickly before it locks up then I can use the reboot command normally as a fix.
If it does lock up however, I’m still able to REISUB in order to reset. I’ve found all the times that I’ve done that recently in my syslog and linked it to this pastebin; I tried to keep anything that seemed relevant from when the PC goes to sleep until when I had to reset it manually.
Looking through my syslogs, I notice lines that seem to indicate the GPU isn’t waking correctly. (ie codium.desktop[19525]: [19525:0616/173717.963456:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=512, and GpuWatchdog[84807]: segfault at 0 ip 00007fa297f929a6 sp 00007fa28c3fd370 error 6 in libcef.so[7fa293aef000+7770000] likely on CPU 3 (core 3, socket 0) ) I’m not sure if that means its a mesa driver issue, a CPU sleep state issue, or maybe just the current kernel update. I was hoping someone with more experience could look through the pastebin link and see if you notice anything I didn’t. I separated the different crashes with white space.
My desktop is an Asus ROG Crosshair VIII Formula with a Ryzen 9 5950X CPU, and a XFX Radeon 6700x GPU
Thank you for you time and help, let me know if I left anything important out.
i think it’s a three-way problem between asus bios, ryzen, and linux kernel. i had a similar problem that was pinning my PN51 to kernel 5.19.16 (in my case, though, i didn’t get a garbage screen, i just got a hard freeze).
i am deathly afraid of bios upgrades, but i swallowed hard and took it - and then the sleep problem went away, and i went back to floating on kernel 6.
but i just got a fanless enclosure, and once i flipped off the fan monitoring, sleep stopped working again - now it hard reboots by itself without any obvious clue in the logs.
the fanless life is more valuable to me than sleep - i have an ssd and don’t run persistent services, so i don’t care enough to keep fighting it. but there’s definitely a tussle in there.
I do get those graphical glitches once in a while but I restart immediately whenever I do and it seems to be okay
It may be related to this: https://bbs.archlinux.org/viewtopic.php?id=284076&p=2
If so, then this could be resolved by our next kernel update.
I think you could get around this by disabling hardware acceleration in VS Code. Open the command palette (
Ctrl + Shift + P), search forPreferences: Configure Runtime Arguments, and add"disable-hardware-acceleration": true. You might need to do similar for other applications triggering this behavior.I think that fixed it, haven’t run into it since.
How did you do it¿? I also have an all amd setup and I am asking in case I have to deal with it in the future
I added that line into codium’s config file using the instructions mmstick wrote out. What are you having issues with?
Edit: and if you’re not noticing any issues you probably aren’t affected then.
I think that fixed it
Damn I read it ‘I fixed it’ instead of ‘that fixed it’. My bad.
