Alexander Graf wrote:
Ah, so we don't sync hflags with the userspace on every transition but
only request it on save/restore?
Yes. That's how all of the other state is synced.
Do we really want to support storing the state when within the
Level2 VM? Can't we just #VMEXIT as soon as we switch to userspace
or so? That would probably simplify things a lot.
Great idea. What do we give as the exit reason?
I thought of INTR here. That should be a nop in every VMM. Even if
not, worst case it would end in an hlt that still fires on the next
timer.
Yes. Luckily svm doesn't support capturing the vector number like vmx does.
There's only one case where this doesn't work, AFAICT: when the INTR
intercept isn't enabled. Once can think of a microkernel hypervisor
that forwards all interrupts to a privileged guest (and when that guest
is running, it can disable INTR intercepts). I don't know of any
hypervisor which does this (Xen doesn't and can't), so I don't think
this is an objection to using INTR.
--
error compiling committee.c: too many arguments to function
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
