On 15/12/06 18:46, "Ky Srinivasan" <ksrinivasan@xxxxxxxxxx> wrote:
> Currently, only a very small subset of the hypercalls are exposed to HVM
> guests. Was this set arrived at based on what was required to host
> para-virtualized drivers? Or are there other considerations that governed
> this decision. Specifically, would there be any objections to opening up more
> hypercalls to HVM guests.
They need auditing one-by-one, and also we have to worry about 32-on-64
compatibility issues (which will be helped by the PV PAE-on-64 work by Jan
Beulich). We do plan to broaden the supported set of hypercalls (memory_op
for example).
 -- Keir
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel