[SOLVED] Error - GPU Passthrough for VM (Nvidia GPU GT 1030 2GB)

bhdschmidt

New Member
Jan 7, 2023
5
1
3
Hi there,

I try to get the GPU Passthrough working for an Ubuntu VM. I used the following guides:

https://www.reddit.com/r/homelab/comments/b5xpua/the_ultimate_beginners_guide_to_gpu_passthrough/

and

https://pve.proxmox.com/wiki/Pci_passthrough.

With this guides I verified that IOMMU and IOMMU Isolation is enabled. In the system there are two GPUs:
1) Intel® HD Graphics 4600 (integrated in the CPU, used for Proxmox)
2) Nvidia GPU GT 1030 2 GB which shall be used for a Ubuntu 20.04 VM

(CPU is i5-4690K, I use Proxmox VE 7.3-3).

When I try to start the Ubuntu VM with the added PCI Devices, I get the following error message:

kvm: -device vfio-pci,host=0000:01:00.0,id=hostpci0,bus=pci.0,addr=0x10: vfio 0000:01:00.0: failed to setup container for group 11: No available IOMMU models
TASK ERROR: start failed: QEMU exited with code 1

I wasn't able to find so much information about the error "No available IOMMU models", but it seems that it's probably a general problem. I tried a bunch of different configuration but I always get the same error. Do you have any suggestions how to proceed?

Thanks in advanced,
Bernhard
 
did you enable VT-d in your bios?

what does find /sys/kernel/iommu_groups/ -type l output?
 
Last edited: