macOS's Virtualization.framework is a high-level framework for building virtual machines, with device emulation, graphics, etc. built in (as opposed to the lower-level Hypervisor.framework, which is analogous to /dev/kvm on Linux).
This repo is a small (< 200 lines of code) client application that can productively run Linux virtual machines. Part of the intent is that it's small enough that you don't have to trust me; you can just read all of the code. To this end, it builds with a short Makefile instead of an Xcode project. To compile, run make. There are two optional features so that you can read even fewer lines of code: CD boot (which requires EFI) and a GUI. If you want to compile these out, use make features= instead of make.
The CLI is intended to feel vaguely familiar to a qemu user, though also some configuration might be easiest by editing the source. This VMM does not support persistent hard disks (though it would be easy to add); it is designed for stateless VMs that store any files on the host in a shared directory.
./smolvm -kernel vmlinuz -initramfs initrd.gz -append 'console=hvc0'
You will need an uncompressed kernel; Virtualization.framework does not know how to deal with compressed kernels. If your file is vmlinuz, gunzip < vmlinuz > vmlinux should do the trick. (The initramfs can be compressed, though, because Linux loads it.)
The kernel also cannot be an EFI stub or UKI. In the Linux kernel repo is a mildly hacky shell script scripts/extract-vmlinux that can deal with both finding a kernel inside an EFI binary and uncompressing it.
Virtualization.framework supports only virtio devices, so you need virtio-pci to be available either as a module or compiled in, or you won't be able to interact with the VM at all. (Note this also means it usually takes several seconds to see anything from the VM, either on the display or on the serial port.)
./smolvm -cd livecd.iso
Virtualization.framework will set up a transient, unprivileged bridge which allows making connections in either direction. Traffic from inside the VM should just work with DHCP. You can also connect from your Mac to the VM. On my machine, VMs get addresses in 192.168.64/24.
When booting the VM, pass an argument like -share src ~/src. The first parameter is the virtio "tag" (which can be anything); the second is the path to share.
From inside the VM, mount the shared directory by giving the tag, e.g.,
mount -t virtiofs src /mnt
Note that for any serious coding work you probably want to make a case-sensitive filesystem (which is easy with APFS).
You get read-only output from the serial console (hvc0) on standard output. I have not yet made this read-write.
Unless you've disabled the ui feature or are running with -nographic, you'll get a window to interact with your VM. It defaults to 800x600 but it is resizable and will pass resize events onto the guest. (Note that the Linux console doesn't support being resized, but if you resize the window before the kernel boots, it will initialize with the right size.) Pressing the power button does a graceful shutdown; use Ctrl-C to force kill the VM.
Linux uses the last console= argument as the interactive one, so you can do something like -append 'console=hvc0 console=tty0' (or the equivalent in GRUB) to see boot messages on stdout but primarily use the graphical console.
Apple has perfectly good sample applications using Virtualization.framework:
One of those might be more to your taste. Of course there are many other good options, including UTM, which supports both Virtualization.framework and qemu with Hypervisor.framework.