Hardware acceleration with OpenMediaVault 8 and kernels newer than 6.12

Last night I have made the upgrade from OMV 7 to OMV 8. There were minor issues that were caused by services that I am using, but I was able to fix them. I have waited a little longer this time because I wanted drivers to catch up to the newer kernels that would come with Debian 13. This time I was not able to enable GPU hardware acceleration with kernel 6.17, no matter what I tried (even by activating Backports from OMV extras). Debian wiki page says that there should be a way to use Nvidia driver 550 with kernels up to 6.19 kernels, but I was not able to have the drivers compiled with Proxmox kernel 6.17. It is a shame that I had to use mainline Debian kernel 6.12 to have it working. I even tried to use a newer driver (580 was recommended by Nvidia) with kernel 6.17. I think I was using Proxmox kernel 6.11 with OMV 7 and Debian 12 before the upgrade.

I need it because I am running Jellyfin and I have a GTX 1060 for that exact purpose. With every major Debian upgrade I had to tweak and change how I install drivers and what I am installing. I am aware that OMV does provide kernels. It is mostly a heads-up for people trying to use the Proxmox kernels (like me) with GPU hardware transcoding/acceleration on Nvidia GPU.

These are commands I used. I also blacklisted nouveau drivers in the kernel. Maybe I will try again when I have some time.

Quote

Code

sudo apt purge *nvidia*
sudo apt autoremove
sudo apt autoclean
sudo cat /etc/modprobe.d/blacklist-nouveau.conf
blacklist nouveau
options nouveau modeset=0

sudo reboot
sudo apt install nvidia-kernel-dkms nvidia-driver nvidia-opencl-icd
sudo apt install nvidia-smi clinfo
lsmod | grep -i nouv
# Nouveau drivers shouldn't be loaded
sudo reboot
clinfo
nvidia-smi
lsmod | grep -i nvidia
# Nvidia drivers should be loaded

Display More

Original errors were that nvidia-kernel-dkms and dmks encountered errors, but unfortunately I didn't see any compilation errors from apt and I also didn't look for any in the system. There were quite a lot of packages updates today (104), although I remember clearly that I have checked and I have updated all packages in my system.

I installed Proxmox kernel 6.17 today (6.17.13-21-pve), I followed the steps listed above and this time there were no errors at all installing nvidia drivers on Debian 13. Unfortunately Nvidia drivers still are not loading at all. Maybe this driver has issues with my GTX 1060, I don't know what to say.I see that Nvidia driver version matches the one that Debian says it's available through backports, although I know for sure the backports repo is disabled (I have a file named openmediavault-kernel-backports.list.save in /etc/apt/sources.list.d.

Quote

mihai@nas:~$ sudo nvidia-smi

[sudo] password for mihai:

NVIDIA-SMI has failed because it couldn't communicate with the NVIDIA driver. Make sure that the latest NVIDIA driver is installed and running.

mihai@nas:~$ lsmod | grep -i nvidia

mihai@nas:~$ lsmod | grep -i nouv

mihai@nas:~$ sudo apt install nvidia-kernel-dkms nvidia-driver nvidia-opencl-icd

nvidia-kernel-dkms is already the newest version (550.163.01-2).

nvidia-driver is already the newest version (550.163.01-2).

nvidia-opencl-icd is already the newest version (550.163.01-2).

Summary:

Upgrading: 0, Installing: 0, Removing: 0, Not Upgrading: 0

mihai@nas:~$ uname -r

6.17.13-21-pve

mihai@nas:~$ lsmod | grep -i nv^C

mihai@nas:~$ sudo dmesg | grep -i error

[ 6.860144] RAS: Correctable Errors collector initialized.

[ 7.508231] ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.SAT1.SPT0._GTF.DSSP], AE_NOT_FOUND (20250404/psargs-332)

[ 7.508252] ACPI Error: Aborting method \_SB.PCI0.SAT1.SPT0._GTF due to previous error (AE_NOT_FOUND) (20250404/psparse-529)

[ 7.588858] ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.SAT1.SPT0._GTF.DSSP], AE_NOT_FOUND (20250404/psargs-332)

[ 7.588869] ACPI Error: Aborting method \_SB.PCI0.SAT1.SPT0._GTF due to previous error (AE_NOT_FOUND) (20250404/psparse-529)

[ 8.844690] faux_driver regulatory: Direct firmware load for regulatory.db failed with error -2

[ 9.183191] EDAC sbridge: Failed to register device with error -19.

mihai@nas:~$ sudo dkms status 'nvidia*'

nvidia-current/550.163.01, 6.12.107+deb13-amd64, x86_64: installed

mihai@nas:~$

Display More