David Hildenbrand <[email protected]> writes: > I have no idea how expensive would be bouncing writes (and reads?) > through the kernel. Did you ever experiment with that/evaluate that?
I would expect it to be quite expensive, as in virtio IO performance tanking. -Andi
