QEMU-KVM in Arch Linux

Check Virtualization Support

Check CPU has VM capabilities

lscpu | grep -i Virtualization

VT-x for Intel
AMD-Vi for AMD

Ensure that your kernel includes KVM modules

zgrep CONFIG_KVM /proc/config.gz

y = Yes (always installed)
m = Loadable module

Install QEMU, libvirt, viewers, and tools

sudo pacman -S qemu-full qemu-img libvirt virt-install virt-manager virt-viewer \
edk2-ovmf swtpm guestfs-tools libosinfo

yay -S tuned

qemu-full - user-space KVM emulator, manages communication between hosts and VMs
qemu-img - provides create, convert, modify, and snapshot, offline disk images
libvirt - an open-source API, daemon, and tool for managing platform virtualization
virt-install - CLI tool to create guest VMs
virt-manager - GUI tool to create and manage guest VMs
virt-viewer - GUI console to connect to running VMs
edk2-ovmf - enables UEFI support for VMs
swtpm - TPM (Trusted Platform Module) emulator for VMs
guestfs-tools - provides a set of extended CLI tools for managing VMs
libosinfo - a library for managing OS information for virtualization.
tuned - system tuning service for linux allows us to optimise the hypervisor for speed.

VirtIO Drivers for Windows Guests

Go to the Fedora People repository and download virtio-win.iso.

Save it anywhere on disk, and attach it to a CD-ROM it when creating Windows VM.

The default location on Debian/RedHat based is `/usr/share/virtio-win/

Enable the libvirt daemon

Here is the documentation detailing the difference between monolithic and modular daemons.
Choose between option 1 and 2 and then do a reboot.

Option 1: Enable the modular daemon.

for drv in qemu interface network nodedev nwfilter secret storage; do
    sudo systemctl enable virt${drv}d.service;
    sudo systemctl enable virt${drv}d{,-ro,-admin}.socket;
done

loop through virtualization systemd services necessary for the libvirt modular daemon.

Option 2: Enable the monolithic daemon.

sudo systemctl enable libvirtd.service

Verify Host Virtualization

sudo virt-host-validate qemu

If you receive warnings, proceed to their respective sections. Re-run the above command to check your changes.

Intel CPU IOMMU Support

WARN (IOMMU appears to be disabled in the kernel. Add intel_iommu=on to kernel cmdline arguments)

Enable IOMMU using GRUB

Open your GRUB config

sudo vim /etc/default/grub

Add the following kernel module entries

# /etc/default/grub
GRUB_CMDLINE_LINUX="... intel_iommu=on iommu=pt"

Regenerate your grub.cfg file

sudo grub-mkconfig -o /boot/grub/grub.cfg
sudo reboot

AMD SEV Support

For AMU CPUs with SEV feature, you might receive this warning:

WARN (AMD Secure Encrypted Virtualization appears to be disabled in kernel. Add kvm_amd.sev=1 to the kernel cmdline arguments)

If you are in Intel you may ignore the warning below, as this only affects AMD CPUs.

WARN (Unknown if this platform has Secure Guest support).

You may refer to this bug report:

https://bugzilla.redhat.com/show_bug.cgi?id=1850351#c5

Or this libvirt documentation:

https://libvirt.org/kbase/launch_security_sev.html

Option 1: Enable AMD SEV using modprobe

echo "options kvm_amd sev=1" >> /etc/modprobe.d/amd-sev.conf
sudo reboot

Option 2: Enable AMD SEV using GRUB

Open your GRUB config and add kernel modules

sudo vim /etc/default/grub

# /etc/default/grub
GRUB_CMDLINE_LINUX="... mem_encrypt=on kvm_amd.sev=1"

Regenerate your grub.cfg file

sudo grub-mkconfig -o /boot/grub/grub.cfg
sudo reboot

Optimise Host with TuneD

Enable TuneD daemon

sudo systemctl enable --now tuned.service

Check active TuneD profile

tuned-adm active

Current active profile: balanced

balanced - generic profile not specialised for KVM, we will change this.

List all TuneD profiles

tuned-adm list

Set profile to virtual-host

sudo tuned-adm profile virtual-host

Verify that TuneD profile

tuned-adm active

Current active profile: virtual-host

sudo tuned-adm verify

Verification succeeded, current system settings match the preset profile. See TuneD log file ('/var/log/tuned/tuned/log') for details.

KVM Networking (optional)

By default all virtual machines will connect to the built-in default NAT network.
To make VMs accessible via the LAN you must create a network bridge.
Keep in mind that network bridges won't work with hosts running on Wireless NICs.
All configuration steps below is done using NetworkManager. If you use a different program to manage networking use that instead.

Default NAT network XML dump:

<network>
    <uuid>...</uuid>
    <forward mode='nat'>
        <nat>
            <port start='1024' end='65535'/>
        </nat>
    </forward>
    <bridge name='virbr0' stp='on' delay='0'/>
    <mac address='AB:CD:EF:AB:CD:EF'/>
    <ip address='10.1.1.1' netmask='255.255.255.0'>
        <dhcp>
            <range start="10.1.1.2" end="10.1.1.254"/>
        </dhcp>
    </ip>
</network>

Sample Netfilter (nftables) rules for default NAT

The rules below is my personal config for my laptop machine. Which works nicely with the autogenerated libvirt network rules.

#!/usr/bin/nft -f

flush ruleset;

define qemu_iface = "virbr0";

table inet filter {
	chain input {
		type filter hook input priority filter; policy drop;

		ct state established,related accept;

		iifname "lo" accept comment "allow loopback";
		iifname $qemu_iface accept comment "allow qemu";

		tcp dport http accept comment "allow sending http";
		tcp dport https accept comment "allow sending https";
		udp dport 67 udp sport 68 accept comment "allow sending dhcp";
		tcp dport ssh accept comment "allow ssh";

		counter drop;
	}

	chain forward {
		type filter hook forward priority filter; policy drop;

		ct state established,related accept;

		iifname $qemu_iface accept comment "forward qemu input";
		oifname $qemu_iface accept comment "forward qemu output";

		counter drop;
	}
}

table ip nat {
	chain postrouting {
		type nat hook postrouting priority srcnat; policy accept;
		ip saddr 10.1.1.0/24 masquerade;
	}
}

Configure bridge interface

find the interface name of your ethernet connection.

sudo nmcli device status

DEVICE	TYPE	STATE	CONNECTION
enp2s0	ethernet	connected	Wired connection 1
lo	loopback	connected (externally)	lo
virbr0	bridge	connected (externally)	virbr0

create a bridge interface using nmcli

sudo nmcli connection add type bridge con-name bridge0 ifname bridge0

connect the ethernet interface to the bridge

sudo nmcli connection add type ethernet slave-type bridge con-name 'Bridge connection 1' \
ifname enp2s0 master bridge0

active the newly created connection

sudo nmcli connection up bridge 0

enable connection.autoconnect-slaves parameter.

sudo nmcli connection modify bridge0 connection.autoconnect-slaves 1

reactivate the bridge and verify connection.

sudo nmcli connection up bridge0
sudo nmcli device status

DEVICE	TYPE	STATE	CONNECTION
bridge0	bridge	connected	bridge0
lo	loopback	connected (externally)	lo
virbr0	bridge	connected (externally)	virbr0
enp2s0	ethernet	connected	Bridge connection 1

Configure bridge network

create an XML file called nwbridge.xml.

vim nwbridge.xml

post the following XML

<network>
    <name>nwbridge</name>
    <forward mode="bridge" />
    <bridge name="bridge0" />
</network>

define the bridge network

sudo virsh net-define nwbridge.xml

Network nwbridge defined from nwbridge.xml

start the bridge network

sudo virsh net-start nwbridge

auto-start bridge network on boot

sudo virsh net-autostart nwbridge

delete nwbridge.xml file

rm nwbridge.xml

verify that nwbridge network exists.

sudo virsh net-list --all

Name	State	Autostart	Persistent
default	active	yes	yes
nwbridge	active	yes	yes

Removing bridge network and interface

If you want to revert the changes to your network, do the following:

sudo virsh net-destroy nwbridge
sudo virsh net-undefine nwbridge
sudo nmcli connection up 'Wired connection 1'
sudo nmcli connection del bridge0
sudo nmcli connection del 'Bridge connection 1'

Libvirt connection modes

libvirt as two methods for connecting to the KVM Hypervisor.

Session Mode

In session mode, a regular user is connected to a per-user instance. Allowing each user to manage their own pool of virtual machines. This is also the default mode.

The advantage of this mode is, permissions are not an issue. As no root access is required.

The disadvantage is this mode uses QEMU User Networking (SLIRP). This is a user-space IP stack, which yields overhead resulting in poor networking performance.

And if you want to implement an option that requires root privileges. You will be unable to do so.

System Mode

In the system mode you are granted access to all system resources.

Granting system-wide access to regular user.

check current mode

sudo virsh uri

qemu:///session

add the current user to the libvirt group

sudo usermod -aG libvirt $USER

set env variable with the default uri and check

echo 'export LIBVIRT_DEFAULT_URI="qemu:///system"' >> ~/.bashrc
sudo virsh uri

Set ACL for the KVM images directory

check permissions on the images directory

sudo getfacl /var/lib/libvirt/images

getfacl: Removing leading '/' from absolute path names
# file : var/lib/libvirt/images/
# owner: root
# group: root
user::rwx
group::--x
other::--x

recursively remove existing ACL permissions

sudo setfacl -R -b /var/lib/libvirt/images/

recursively grant permission to the current user

sudo setfacl -R -m u:${USER}:rwX /var/lib/libvirt/images/

uppercase X states that execution permission only applied to child folders and not child files.

enable special permissions default ACL

sudo setfacl -m d:u:${USER}:rwx /var/lib/libvirt/images/

if this step is omitted, new dirs or files created within the images directory will not have this ACL set.

verify your ACL permissions within the images directory.

sudo getfacl /var/lib/libvirt/images/

getfacl: Removing leading '/' from absolute path names
# file : var/lib/libvirt/images/
# owner: root
# group: root
user::rwx
user:tatum:rwx
group::--x
mask::rwx
other::--x
default:user::rwx
default:user:tatum:rwx
default:group::--x
default:mask::rwx
default:other::--x

tatumroaquin/archlinux-qemu-kvm.md