[RFC] First (incomplete) cut of Xen paravirt binding

[chrisw at sous-sol.org (Chris Wright)]

* Jeremy Fitzhardinge (jeremy at goop.org) wrote:
> Andi Kleen wrote:
> >The other approach - check for ring != 0 - do generic paravirtualized
> >startup or normal one and then probe later in C seems certainly
> >cleaner.
> 
> Xen supports trusted ring 0 paravirtualized kernels.

That's only for the "single kernel runs on hv and native" AFAIK.  And we
certainly can't require a Xen hypervisor for running on native.  So, I
don't think that has any sane requirements that we should consider.

thanks,
-chris