Problem description:
525.85.05 causes lags and freezes in both Wayland and X11. At worse, this prevents me from logging into gnome-shell. At best, this prevents me from using the nvidia card.
I suspect the issue lies within the driver since:
- Downgrading to 525.78.01 (everything else being equal) fixes all issues.
- The issue occurs with both stable and lts kernels.
- Starting from a downgraded system (i.e. 525.78.01), re-installing 525.85.05 with DKMS triggers the same issue.
The reported error is:
kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 525.85.05 Sat Jan 14 00:40:03 UTC 2023
kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 1
gnome-shell[995]: Added device '/dev/dri/card1' (nvidia-drm) using atomic mode setting.
kernel: NVRM: GPU at PCI:0000:01:00: GPU-14bb0cac-de77-533b-00cd-01c88961c2b7
kernel: NVRM: Xid (PCI:0000:01:00): 79, pid='<unknown>', name=<unknown>, GPU has fallen off the bus.
kernel: NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
Let me know if I can do anything to help. I will test new drivers as they are published and report back if I notice a return to normal.
Steps to reproduce:
Reproducible on my laptop with both kernel 6.1.8 (stable) and 5.15.90 (lts).
- Boot and log into gnome-shell (X11 or Wayland)
- If the desktop has not frozen just after login, run
vkcube. - Note that the desktop has frozen.
- Reboot and note that the nvidia card disappeared (e.g. nvidia-smi never returns).
- Reboot again and note that the nvidia card is back.
In a nutshell: any use of the nvidia card freezes the desktop and requires two reboots to recover from.
Workaround
The only consistent workaround I have found is to satisfy these three conditions simultaneously:
- Boot with
nvidia-drm.modeset=0.
(This prevents freezing at login by not initializing the nvidia card.) - Use X11.
(Wayland is unusable with 525.85.05 on my laptop.) - Set powermizer to maximum performance before any use of the nvidia card.
(Any deviation from triggers a freeze each when the nvidia card is used.)
Then the nvidia card can be used as usual.
Material specs
Laptop model: ASUS ROG GL552VL
Intel Core i7 6700HQ
Intel HM170 Chipset
Integrated Intel HD Graphics 530 (connected to both Laptop Display and HDMI)
NVIDIA GeForce GTX 965M with 4GB GDDR5 VRAM (PCI Express x16 Gen3)
Bug report
The .log.gz I managed to capture is attached. In that instance the desktop recovered after the freeze and I was able to run nvidia-bug-report.sh.
nvidia-bug-report.log.gz (655.0 KB)