PCI Passthrough

Introduction

PCI passthrough allows you to use a physical PCI device (graphic card, network card) inside a VM (KVM virtualization only) If you "PCI passthrough" a device, the device is not available in the host anymore.

Note:

PCI passthrough is a experimental features in Proxmox VE

Intel CPU

edit:

#vi /etc/default/grub

change

GRUB_CMDLINE_LINUX_DEFAULT="quiet"

to

GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=on"

then

# update-grub

Then run "dmesg | grep -e DMAR -e IOMMU" from the command line.  If there is no output, then something is wrong.

AMD CPU

Edit:

# vi /etc/default/grub

Change:

GRUB_CMDLINE_LINUX_DEFAULT="quiet"

To:

GRUB_CMDLINE_LINUX_DEFAULT="quiet amd_iommu=on"

Modules loading

add to /etc/modules

vfio
vfio_iommu_type1
vfio_pci
vfio_virqfd

IOMMU interrupt remapping

it will not be possible to use PCI passthrough without interrupt remapping.

Device assignment will fail with a 'Failed to assign device "[device name]" : Operation not permitted' error for users of KVM, and a 'Interrupt Remapping hardware not found, passing devices to unprivileged domains is insecure.

• All systems using an AMD processor and chipset that have AMD I/O Virtualization (AMD-Vi) support. Such hardware has interrupt remapping support; however, the software support is not yet available upstream.

• All systems using an Intel processor and chipset that have support for Intel Virtualization Technology for Directed I/O (VT-d), but do not have support for interrupt remapping. Interrupt remapping support is provided in newer processors and chipsets. To identify if your system has support for interrupt remapping:

1) Run the "dmesg | grep ecap" command.

2) On the IOMMU lines, the hexadecimal value after "ecap" indicates whether interrupt remapping is supported. If the last character of this value is an 8, 9, a, b, c, d, e, or an f, interrupt remapping is supported. For example, "ecap 1000" indicates there is no interrupt remapping support. "ecap 10207f" indicates interrupt remapping support, as the last character is an "f".

Interrupt remapping will only be enabled if every IOMMU supports it.

Alternatively, run the following script to determine if your system has interrupt remapping support:

#!/bin/sh
if [ $(dmesg | grep ecap | wc -l) -eq 0 ]; then
  echo "No interrupt remapping support found"
  exit 1
fi

for i in $(dmesg | grep ecap | awk '{print $NF}'); do
  if [ $(( (0x$i & 0xf) >> 3 )) -ne 1 ]; then
    echo "Interrupt remapping not supported"
    exit 1
  fi
done

If your system don't support interrupt remapping,

you can allow unsafe interrupts with:

echo "options vfio_iommu_type1 allow_unsafe_interrupts=1" > /etc/modprobe.d/iommu_unsafe_interrupts.conf

Verify IOMMU isolation

To have pci passthrough working fine, you need dedicated iommu group for your pci devices

You should have something like

# find /sys/kernel/iommu_groups/ -type l
/sys/kernel/iommu_groups/0/devices/0000:00:00.0
/sys/kernel/iommu_groups/1/devices/0000:00:01.0
/sys/kernel/iommu_groups/1/devices/0000:01:00.0
/sys/kernel/iommu_groups/1/devices/0000:01:00.1
/sys/kernel/iommu_groups/2/devices/0000:00:02.0
/sys/kernel/iommu_groups/3/devices/0000:00:16.0
/sys/kernel/iommu_groups/4/devices/0000:00:1a.0
/sys/kernel/iommu_groups/5/devices/0000:00:1b.0
/sys/kernel/iommu_groups/6/devices/0000:00:1c.0
/sys/kernel/iommu_groups/7/devices/0000:00:1c.5
/sys/kernel/iommu_groups/8/devices/0000:00:1c.6
/sys/kernel/iommu_groups/9/devices/0000:00:1c.7
/sys/kernel/iommu_groups/9/devices/0000:05:00.0
/sys/kernel/iommu_groups/10/devices/0000:00:1d.0
/sys/kernel/iommu_groups/11/devices/0000:00:1f.0
/sys/kernel/iommu_groups/11/devices/0000:00:1f.2
/sys/kernel/iommu_groups/11/devices/0000:00:1f.3
/sys/kernel/iommu_groups/12/devices/0000:02:00.0
/sys/kernel/iommu_groups/12/devices/0000:02:00.1
/sys/kernel/iommu_groups/13/devices/0000:03:00.0
/sys/kernel/iommu_groups/14/devices/0000:04:00.0

To have separate iommu, your processor need to have support for a feature called ACS (Access Control Services).

All Xeon processor support them (E3,E5) excluding Xeon E3-1200

For intel Core , it's different, only some processors support ACS

Haswell-E (LGA2011-v3)
i7-5960X (8-core, 3/3.5GHz)
i7-5930K (6-core, 3.2/3.8GHz)
i7-5820K (6-core, 3.3/3.6GHz)

Ivy Bridge-E (LGA2011)
i7-4960X (6-core, 3.6/4GHz)
i7-4930K (6-core, 3.4/3.6GHz)
i7-4820K (4-core, 3.7/3.9GHz)

Sandy Bridge-E (LGA2011)
i7-3960X (6-core, 3.3/3.9GHz)
i7-3970X (6-core, 3.5/4GHz)
i7-3930K (6-core, 3.2/3.8GHz)
i7-3820 (4-core, 3.6/3.8GHz)

UPDATE ME : AMD processors ?

If you don't have dedicated iommu, your can try :

1) move the card to another pci slot

2) add "pcie_acs_override=downstream" to grub options, which can help on some setup with bad ACS implementation.

More infos :

http://vfio.blogspot.be/2015/10/intel-processors-with-acs-support.html

http://vfio.blogspot.be/2014/08/iommu-groups-inside-and-out.html

Determine your PCI card address, and configure your VM

Locate your card using "lspci".  The address should be in the form of: 04:00.0

Manually edit the node.conf file.  It can be located at: /etc/pve/qemu-server/vmid.conf.

Add this line to the end of the file:

hostpci0: 04:00.0

If you have a multi-function device (like a vga card with embedded audio chipset),

you can pass both with removing the ".0" in pci address.

hostpci0: 04:00

PCI EXPRESS PASSTHROUGH

/etc/pve/qemuserver/<vmid>.cfg

simple pci-express passthrough

machine: q35
hostpci0: 04:00.0,pcie=1

GPU PASSTHROUGH

• MD RADEON 5xxx, 6xxx, 7xxx and NVIDIA GEFORCE 7, 8, 4xx, 5xx, 6xx, 7xx have been reported working.
• Maybe you'll need to load some specific options in grub.cfg or other tuning values,
• Here a good forum thread of archlinux: https://bbs.archlinux.org/viewtopic.php?id=162768

For GPU, it's good that host don't try to use the GPU, and avoids issues with the host driver unbinding and re-binding to the device.

First, find the device and vendor id of your vga card

$ lspci -n -s 1:
01:00.0 0300: 10de:1381 (rev a2)
01:00.1 0403: 10de:0fbc (rev a1)

The Vendor:Device IDs for my GPU and audio functions are therefore 10de:1381, 10de:0fbc.

Then, create a file

echo "options vfio-pci ids=10de:1381,10de:0fbc" > /etc/modprobe.d/vfio.conf

For VM configuration, They are 4 configuration possible:

GPU Seabios PCI PASSTHROUGH

hostpci0: 04:00.0,x-vga=on

GPU Seabios PCI EXPRESS PASSTHROUGH

machine: q35
hostpci0: 04:00.0,pcie=1,x-vga=on

GPU OVMF PCI PASSTHROUGH

OVMF replace bios by UEFI boot. You need to install your guest OS with uefi support,

and you need to your graphic card have a uefi bootable rom

This is the recommended mode

bios: ovmf
scsihw: virtio-scsi-pci
bootdisk: scsi0
scsi0: .....
hostpci0: 04:00.0,x-vga=on

GPU OVMF PCI EXPRESS PASSTHROUGH

OVMF replace bios by UEFI boot.

You need to install your guest OS with uefi support,

and you need to your graphic card have a uefi bootable rom

bios: ovmf
scsihw: virtio-scsi-pci
bootdisk: scsi0
scsi0: .....
machine: q35
hostpci0: 04:00.0,pcie=1,x-vga=on

Verify Operation

Start the VM from the UI.

Enter the qm monitor.  "qm monitor vmnumber"

Verify that your card is listed here: "info pci"

Then install drivers on your guest OS.  

NOTE: Card support might be limited to 2 or 3 devices.

NOTE: This process will remove the card from the proxmox host OS.  

Editorial Note: Using PCI passthrough to present drives direct to a ZFS (FreeNAS, Openfiler, OmniOS) virtual machine is OK for testing, but not recommended for production use. Specific FreeNAS warnings can be found here: http://forums.freenas.org/threads/absolutely-must-virtualize-freenas-a-guide-to-not-completely-losing-your-data.12714/

WORKING NVIDIA SETUP

I've been able to get this working with an NVIDIA GTX 750 Ti card using driver version 344.75 (newer versions inconsistently cause Code 43 Errors) by using the following setup:

Install pve-kernel-3.10.0-5-pve with:

Add to /etc/modules:
pci_stub
vfio
vfio_iommu_type1
vfio_pci
vfio_virqfd
kvm
kvm_intel

Add the following options to /etc/default/grub on the GRUB_CMDLINE_LINUX_DEFAULT line:

intel_iommu=on vfio_iommu_type1.allow_unsafe_interrupts=1 rootdelay=10 scsi_mod.scan=sync

Run: update-grub

Add the following to <tt>/etc/initramfs-tools/modules</tt> (find the PCI stub IDs for your card by running <tt>lspci -nn | grep NVIDIA</tt>):
pci_stub ids=10de:0f02,10de:0bea

Run: update-initramfs -u and then reboot into the new kernel.

Boot into the VM using the Proxmox web interface, and install the OS (I had better luck with Windows 8.1 - can't remember specifics though).

Add the following options to /etc/pve/host/qemu-server/vmid.conf (get the PCI address from lspci command, and I added the USB device address for my Avocent KVM DSRIQ USB module, you can do the same for a physical keyboard and mouse):

hostpci0: 05:00,x-vga=on,pcie=1
machine: q35
usb0: host=0624:0307