Ubuntu 24.04 - Unable to change power state from D3cold to D0, device inaccessible

[lx_b]

I have the nvidia drivers installed, dkms reports them as ok, i used latest nvidia drivers recommended by ubunt and install via apt (not nouveau which is blacklisted).

I have a 2080ti in a sonnet 750ex TB3 enclosure chained to a OWC Helios 3s to a Zimacube Pro USB4/TB4 compliant system. I need to chain this way due to limited number of ports.

Driver fails to load with repeated entries below. I tried adding pcie_aspm=force to my grub, to no effect.

What next?

[ 3153.892536] nvidia 0000:1f:00.0: Unable to change power state from D3cold to D0, device inaccessible
[ 3153.892665] nvidia 0000:1f:00.0: vgaarb: VGA decodes changed: olddecodes=none,decodes=none:owns=none
[ 3153.892693] NVRM: The NVIDIA GPU 0000:1f:00.0
               NVRM: (PCI ID: 10de:1e07) installed in this system has
               NVRM: fallen off the bus and is not responding to commands.
[ 3153.892751] nvidia: probe of 0000:1f:00.0 failed with error -1
[ 3153.892771] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 3153.892772] NVRM: None of the NVIDIA devices were initialized.
[ 3153.893004] nvidia-nvlink: Unregistered Nvlink Core, major device number 509

[lx_b]

nvidia-bug-report.log.gz (757.9 KB)

[lx_b]

This was fixed by upgrading to kernel 6.8.4-060804-generic

(promox users had reported the need to downgrade to this kernel - seems drivers are suer sensitive to kernel versions)

Commands i used to downgrade

wget https://raw.githubusercontent.com/pimlie/ubuntu-mainline-kernel.sh/master/ubuntu-mainline-kernel.sh

chmod +x ubuntu-mainline-kernel.sh

sudo ./ubuntu-mainline-kernel.sh -i v6.8.4

sudo reboot now