PCIe passthrough VFIO prep

PCIe passthrough lets a VM own a physical device (GPU, HBA, NIC) directly. The setup is at the host level before any VM config.

Enable IOMMU

Edit /etc/default/grub and add the appropriate kernel parameter:

# Intel
GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=on iommu=pt"

# AMD
GRUB_CMDLINE_LINUX_DEFAULT="quiet amd_iommu=on iommu=pt"

iommu=pt (passthrough mode) is required for stability on most hardware — see the LSI passthrough entry for why.

update-grub

reboot

Verify after reboot:

dmesg | grep -e IOMMU -e iommu

Load VFIO modules

Add to /etc/modules:

vfio
vfio_iommu_type1
vfio_pci
vfio_virqfd

update-initramfs -u -k all

Find the device and its IOMMU group

List IOMMU groups:

find /sys/kernel/iommu_groups -type l | sort -V

Find PCI IDs for your device (GPU example):

lspci -nn | grep -i nvidia

All devices in the same IOMMU group must be passed through together. If a GPU shares a group with an unrelated device, you may need the ACS override kernel patch — but that’s a security tradeoff.

Bind the device to VFIO

Add to /etc/modprobe.d/vfio.conf using the IDs from lspci -nn:

options vfio-pci ids=10de:2204,10de:1aef

Blacklist the host driver in /etc/modprobe.d/blacklist.conf:

blacklist nouveau
blacklist nvidia

update-initramfs -u -k all

reboot

Assign to a VM

In the Proxmox UI: VM → Hardware → Add → PCI Device. Select the device. Enable ROM-Bar and PCI-Express for GPUs.

Verify inside the VM

lspci

nvidia-smi