Computer freezing with artifacts after nvidia driver upgrade


I’m quite new on Linux and it’s hard to me to debug this problem.

I used kernel 513 with an Nvidia 3060 ti and after the last upgrade (today) my computer started freezing randomly (with screen artifacts) and sometimes when I reboot it does not load with a “unable to start lightdm” error message. In this screen, when I ctrl+alt+f2 and run “nvidia-smi” it says that there is no device (not finding the GPU).

As a solution I tried installing kernel410 (because its LTS) and using mhwd to downgrade the nvidia driver to the 470xx series which seemed to make it better, but still intermittent.

Can someone please help me to diagnose and solve this problem?

I’m also getting this before freezing.

kernel: NVRM: GPU at PCI:0000:07:00: GPU-5c835339-45df-c2fe-a127-cafd24e37f17

kernel: NVRM: Xid (PCI:0000:07:00): 31, pid=402, Ch 00000002, intr 00000000. MMU Fault: ENGINE GRAPHICS GPCCLIENT_PROP_0 faulted @ 0x4_00000000. Fault is of type FAULT_PDE ACCESS_TYPE>

kernel: NVRM: Xid (PCI:0000:07:00): 31, pid=1763, Ch 00000010, intr 00000000. MMU Fault: ENGINE HOST0 HUBCLIENT_ESC faulted @ 0x1_06122000. Fault is of type FAULT_PRIV_VIOLATION ACCES>

kernel: NVRM: Xid (PCI:0000:07:00): 62, pid=1763, 0000(0000) 00000000 00000000

kernel: sched: RT throttling activated

Do you mean 5.10?
In either case you may wish to check 5.15 which is expected to be LTS as well.

Yes, 510.

I’ll try 5.15

can you report

inxi -Fza