On Mon, 29 Mar 2010, Freddie Cash wrote:
PV DomU support is "fairly easy",
Dom0 support is hard, and requires a lot of changes to the kernel, as you are essentially adding a whole new CPU architecture (Xen).
Well, I come from NetBSD, so I don't really see the problem with doing that. Happens all the time. :-)
In the long-run, a KVM-like system where the DFly kernel becomes the hypervisor and VMs run as user processes is better than the Xen split-dom, "we're our own architecture, but we fake it using a real kernel" stuff.
It's really two different things entirely. Seems to me that with KVM, it is assumed that the host OS is given, and it provides a way to more easily virtualize guests; Xen puts itself in the center, and instead treats hosts and guests as exchangable. I like the idea that you can treat it as just another hardware architecture.
But as has been said before, KVM could benefit vkernels, and any other virtualization solution, so I can definitely see why that would be attractive.
MAgnus
