Andi Kleen wrote:

quoted

In the Xen case, 
they may want to run a dom-0 hypervisor which is compiled for an actual 
hardware sub-arch, such as Summit or ES7000. 
    

There is no reason Summit or es7000 or any other subarchitecture 
would need to do different  virtualization. In fact these subarchitectures 
are pretty much obsolete by the generic subarchitecture and could be fully
done by runtime switching.
  

For privileged domains that have hardware privileges and need to send 
IPIs or something it might make sense.  Othewsie, there is no issue.

quoted

I would expect to see these new sub-architectures 
begin to grow like a rash. 
    

I hope not. The i386 subarchitecture setup is pretty bad already
and mostly obsolete for modern systems.
  

Yes, I hope not too.

quoted

I'm now talking lightyears into the future
    

tststs - please watch your units.
  

I realized after I wrote it ;)

I don't fully agree to move everything into paravirt ops. IMHO
it should be only done for stuff which is performance critical
or cannot be virtualized.

Yes, this is all just a crazy idea, not an actual proposal.

And it's unlikely PCI will be ever a good fit for a Quantum computer @)
  

Hmm, a quantum bus would only allow one reader of each quantum bit.  So 
you couldn't broadcast without daisy chaining everything.  Could be an 
issue.

quoted

Maybe someday Xen and VMware can share the same ABI interface and both 
use a VMI like layer. 
    

The problem with VMI is that while it allows hypervisor side evolution
it doesn't really allow Linux side evolution with its fixed spec.
  

It doesn't stop Linux from using the provided primitives in any way is 
sees fit.  So it doesn't top evolution in that sense.  What it does stop 
is having the Linux hypervisor interface grow antlers and have new 
hooves grafted onto it.  What it sorely needed in the interface is a way 
to probe and detect optional features that allow it to grow independent 
of one particular hypervisor vendor.

Zach