By complicated, I mean that you have to modify your the default behavior of your host OS (which by default will grab any hardware) and pick a GPU that is supported by the guest OS. People with Nvidia GPUs cannot do that for macOS. Ideally, out-of-the-box, with a few clicks, most people including non-tech savy ones should be able to enjoy 3D acceleration in their VM, regardless of their host configuration.

Besides, in the long term, Apple is likely to drop support for AMD or Intel GPU in their OS (but yes, they might also drop support for virtio devices).

Yes, evdev is currently the best way to share your inputs devices to a VM, thank you for pointing that out (alas, it is only available in distributions with very recent virtualization-related packages).

In my (perhaps narrow) view: virtualization should be about making you less reliant on the underlying hardware. I wish that GPU manufacturers would agree on a common standard to allow users to split their beefy GPU into smaller parts (Intel is apparently dropping support for vfio-mdev (Intel GVT-d), adopting SR/IOV instead for their latest offerings. But SR/IOV is only available on professional models for AMD and Nvidia. In summary, there won't be a standard any time soon, leaving virtio-gpu as the sole hardware-agnostic contender)

> Besides, in the long term, Apple is likely to drop support for AMD or Intel GPU in their OS

Off topic but I’m curious about this. On one hand I think it’s a very Apple like thing to do but on the other hand it would practically kill Mac for pro use cases, which is barely hanging on by a thread right now. Could you imagine buying a $10,000+ workstation and Apple telling you 1 8k display is enough?

I think it really hinges on what the next Mac Pro looks like. If it’s arm based and allows expansion then it’ll likely support AMD and intel gpus and support for those will stick around for years. If not , then the current intel macs and their egpu support will be the end of the line.

You correlate an Apple GPU with a hard limit on the amount of displays you can use, but honestly I don't see the point. Apple has bittersweet limits on the amount of displays one can connect right now, but it is no guarantee that this will continue in the future.

With what they unveiled in terms of GPU performance with the Macbook Pro, the writing's on the wall; Apple is not going to support third-party GPUs going forward. The Mac Pro will only offer Apple CPUs and GPUs, I'm sure of it.

IME you’ll also need to use isolcpus and affinity on the KVM threads to get stable GPU performance (with macOS or other OSes). Plus there can be some other fun issues depending on your setup, like having to flash a different BIOS on your card or having to deal with multiple devices on the same IOMMU domain (I had to pass through a USB controller with my card).

"you get 99% performance and 100% stability of a native Mac (not a single macOS/QEMU crash in years for me)"

I am curious how this is so stable for you. Are you running this on Intel? Just last week I installed this on my AMD Linux workstation to test some stuff with Xcode and qemu would crash randomly while compiling LLVM/Clang.

Yes, I run this on Intel i7, under Ubuntu 20.04, with QEMU 4.x, as "-machine q35".