Jeremy Fitzhardinge escreveu:
Glauber de Oliveira Costa wrote:
Thanks for the explanation, Andi. I understand it much better now, and
agree with you.
As alternatives what we have now, we can either keep the paravirt_ops
as it is now for the native case, just hooking the vsmp functions in
place of the normal one, (there are just three ops anyway), refill the
paravirt_ops entirely in somewhere like vsmp.c, or similar (or maybe
even assigning paravirt_ops.fn = vsmp_fn on the fly, but early enough).
One thing to note is that current code assumes the IF flag is always in
bit 9, so if you paravirtualize this, you need to either a) make the
vsmp version copy AC into IF to satisfy the interface, or b) add a new
op meaning "tell me if this eflags has interrupts enabled or not". I
went for option a), and it seems to work OK (using bit 9 for "interrupt
enabled" is pretty arbitrary from a Xen perspective, but not very hard
to implement, and more localized than making all eflags tests a pvop).
J
It is implemented like a) in the latest patch I send, following chris'
suggestion.
_______________________________________________
Virtualization mailing list
Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/virtualization
