Thomas Gleixner wrote:

Once you are there, you are near the point where you created a virtual
architecture, which could run on any real architecture which gets
supported by a hypervisor backend.

I'd love that :)
  

Sure.  But not even hypervisors.  Once we sort out pv_ops's SMP support,
it will be this >< close to covering everything in the subarch
interface.  So we can drop all that goo in favour of paravirt_ops, and
make a single kernel that will boot on everything from voyager to
numa-q!  How's that for world peace?

I know it is tricky to combine this with the upcoming hardware
virtualization support. But it's at least a worthwhile thought
experiment.
  

Well, in many ways that's a step backwards.  The upside is that its
easier to get away with simply emulating the some particular piece
hardware, but it does lose a lot of opportunities for interesting
flexibility and optimisations.

But I anticipate we'll get a xen-hvm pv_ops backend, for running under
Xen with a virtualizing cpu.  It will probably look a lot like kvm's
pv_ops backend.

    J