Since I upgraded to Fedora 35, I am seeing graphic corruption upon resume from suspend:
options nvidia NVreg_PreserveVideoMemoryAllocations=1
options nvidia NVreg_TemporaryFilePath=/var/tmp
Bug report log attached: nvidia-bug-report.log.gz (504.3 KB)
Encountering the same issue on Arch with Gnome 41, xorg-server (21.1.1), and nvidia (495.44).
I’ve rolled back and pinned at Gnome 40 for now. Haven’t seen this tracked anywhere else yet.
I can confirm I have the same issue with Fedora 35 and Gnome 41 on Xorg
I also have this issue, with the following configuration:
Restarting the Gnome shell (ALT+F2, then “r” in the command box) removes the artefacting, however it does return after the next suspend.
I can confirm the issue exists for me as well:
Same here with 495.44 driver (plus some other issues too): 495.44 driver regression: long delay resuming from suspend
470.82 does not have this issue.
I also started seeing this problem recently (1/6/22) after upgrading to the 495.46 driver on a Fedora 34 system (Lenovo P52 laptop with the Quadro P1000).
Any fix available yet?
Same issue with 495.46 on Fedora 35 (Acer G3-571 with GTX 1060), sometimes it corrupts the entire screen but most of the time just black and pink squares over the screen.
Just updated to 510.47.03 on Fedora 35, issue is still present
I haven’t seen this issue prior to update to 510.47. I’m not sure what have triggered it though.
I’ll have to try to boot with previous kernel to see if issue persist.
Same issue with Kernel 5.16 and driver 510.47.
I tried all the powermanagement options but no dice.
I hope a fix is coming soon, it’s been since kernel 5.15 and driver 495 that this bug is present.
I originally started seeing this problem on Fedora 34 using the 495.46 driver.
Yesterday I upgraded to the 5.16.7-100 kernel and 510.47 driver.
Still see the same corrupted graphics after a resume. Doing an Alt-F2/r restart of the window manager will fix, but this is annoying! :)
Anyone have any info on when a fix might be coming???
Thanks! You saved me from unnecessary restarts!
Same issue with F35, Gnome 41, kernel 5.17.5-200 and driver 510.60.02. As previously suggested, an Alt-F2/r restart solves the problem. Lock screen is also affected.
We are working on this issue and will keep you updated on it.
We have also filed a bug 3641108 internally for tracking purpose.
I also observed similar behavior on my test system and issue disappeared once I restart gdm service.
We have root caused the issue and fix will be incorporated in future driver release.
Please test with latest beta driver release and share test results.
https://us.download.nvidia.com/XFree86/Linux-x86_64/515.43.04/NVIDIA-Linux-x86_64-515.43.04.run
Can confirm with 515.43.04 resume from standby works without corrupted graphics.
@jurgenxiv Thanks for the test results.
I’m still experiencing this issue using extra/nvidia 515.43.04-4 on Arch Linux.
I’ll add that I’m using Wayland with nvidia-drm.modeset=1.
Is using NVreg_PreserveVideoMemoryAllocations=1 required? This seems to trigger OOM Killer on suspend for me, no matter which fs I point NVreg_TemporaryFilePath to.