I mean, sure, this specific work is in qemu, not the kernel. Some hypercalls are trapped by the kernel, some trap out to a vmm like qemu. Support for pv devic

I mean, sure, this specific work is in qemu, not the kernel. Some hypercalls are trapped by the kernel, some trap out to a vmm like qemu. Support for pv devices mainly lives in the vmm. But A) even the hypercalls the kernel handles needs to be manually enabled by the vmm, and B) the pv device support ultimately is exposed as in guest memory tables and hypercalls (now partially emulated by qemu).