"Works on my machine"

The following configuration works on my 2021 M1 Max laptop running MacOS Sequoia 15.6.1.

ARM Architecture

Most of the guides I found were focused on emulating amd64/x86-64 architectures. This guide instead targets the aarch64/arm64 architecture (same as the host system) to minimize emulation overhead.

Steps

These steps should get you a Fedora Server VM that you can SSH into from your host OS.

Install QEMU

$ brew install qemu

Download Fedora Server filesystem image

Thankfully, Fedora provides ready-to-go filesystem images on their downloads page. Download the QEMU qcow2 filesystem image for the latest server version (42 as of writing).

Resize the Fedora Server filesystem image

The Fedora Server image downloaded in the previous step is only large enough to fit the base installation and will likely need to be resized for your needs. You can use qemu-img to resize it like so:

$ qemu-img resize Fedora-Server-Guest-Generic-42-1.1.aarch64.qcow2 15G

Where 15G is the target size.

Create a startup script start.sh

Create a file start.sh that will contain the qemu invocation for your new VM.

I found it helpful to check this script into git while iterating on different qemu arguments so I could rollback to working versions.

The script:

qemu-system-aarch64 \
    -name "Fedora" \
    -machine type=virt,iommu=smmuv3 \
    -accel hvf \
    -cpu cortex-a57 \
    -smp 4 \
    -m 2048 \
    -rtc base=utc \
    -drive if=pflash,format=raw,file="/opt/homebrew/share/qemu/edk2-aarch64-code.fd",readonly=on \
    -drive if=pflash,format=raw,file="/opt/homebrew/share/qemu/edk2-arm-vars.fd" \
    -drive file="Fedora-Server-Guest-Generic-42-1.1.aarch64.qcow2",if=virtio,format=qcow2 \
    -boot menu=on \
    -display default \
    -device virtio-gpu \
    -nic user,id=NAT,model=virtio-net-pci,mac=02:00:00:00:00:01,hostfwd=tcp::2222-:22 \
    -device qemu-xhci \
    -device usb-kbd \
    -device usb-tablet

Argument Explanation

Some of these arguments are worth paying attention to, I'll list the important ones below.

• accel hvf: This sets the acceleration framework to the use native MacOS Hypervisor Framework.

• m 2048: Give the VM 2GB of RAM.

• smp 4: Give the VM 4 CPUs.

• -nic user,id=NAT,model=virtio-net-pci,mac=02:00:00:00:00:01,hostfwd=tcp::2222-:22: The hostfwd option forwards port 2222 on the host OS to port 22 (ssh) on the guest OS.

Start the VM

Boot up the VM using start.sh

$ sh start.sh

You should see a window open up like so:

Then you'll see it say "Display output is not active.", this is okay, just wait a bit, eventually you will see the machine boot.

Configure Fedora Server

The first time you boot Fedora Server there is a clunky setup text UI thing for setting up a non-root user and giving it a password. Once this is done and you can hit c to continue, your VM is good to go.

SSH from Host OS

You should now be able to SSH into the Guest OS using the username and password created in the previous step by using port 2222, like so:

$ ssh capotej@localhost -P 2222

Grow Guest filesystem to use new size

In order for the Guest OS to take advantage of new size from the qemu-resize above, we need to grow the root partition as well as extend our logical volume using the commands below.

Extend the partition

$ sudo growpart /dev/vda 3

Extend the physical volume

$ sudo pvresize /dev/vda3

Extend the logical volume

$ sudo lvextend -l +100%FREE /dev/systemVG/LVRoot

Resize the filesystem

$ sudo xfs_growfs /

Verify extra space

You can verify that / has more space by running df like so

$ df -h /

Shutting down the VM

You can use the standard shutdown command to shutdown the VM while logged into ssh:

$ sudo shutdown -h now

Headless Mode

Once you have your VM setup, you can edit start.sh and change display to :

-display none -serial null \

Then when you run start.sh it will run the VM in the foreground without a display. You can use Ctrl-C to stop the VM from the host OS.

Your VM should now be ready for experimentation, happy VMing!