On Fri, Feb 08, 2013 at 12:07:36AM -0200, Marcelo Tosatti wrote: > On Thu, Feb 07, 2013 at 03:52:24PM +0200, Gleb Natapov wrote: > > > Its not a bad idea to have a new KVM_REQ_ bit for PIR processing (just > > > as the current patches do). > > Without the numbers I do not see why. > > KVM_REQ_EVENT already means... counting... many things. Its a well Exactly my point. KVM_REQ_EVENT means many things, all of them event injection related. It can be split may be to 2/3 smaller request, but it will complicate the code and why would we do that without actually able to show performance improvements that split provides? > defined request, to sync PIR->VIRR, don't see your point about > performance. And it is just one more things that needs to be done during event injection. Without providing a good reason no need to handle it specially. Performance is convincing enough reason, but I what to see numbers. -- Gleb. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
