On 10.07.2013, at 12:48, Gleb Natapov wrote: > On Wed, Jul 10, 2013 at 12:45:59PM +0200, Alexander Graf wrote: >> >> On 10.07.2013, at 12:42, Gleb Natapov wrote: >> >>> On Wed, Jul 10, 2013 at 12:39:01PM +0200, Alexander Graf wrote: >>>> >>>> On 09.07.2013, at 18:01, Christian Borntraeger wrote: >>>> >>>>> On 09/07/13 15:56, Dominik Dingel wrote: >>>>>> By setting a Kconfig option, the architecture can control when >>>>>> guest notifications will be presented by the apf backend. >>>>>> So there is the default batch mechanism, working as before, where the vcpu thread >>>>>> should pull in this information. On the other hand there is now the direct >>>>>> mechanism, this will directly push the information to the guest. >>>>>> >>>>>> Still the vcpu thread should call check_completion to cleanup leftovers, >>>>>> that leaves most of the common code untouched. >>>>>> >>>>>> Signed-off-by: Dominik Dingel <dingel@xxxxxxxxxxxxxxxxxx> >>>>> >>>>> Acked-by: Christian Borntraeger <borntraeger@xxxxxxxxxx> >>>>> for the "why". We want to use the existing architectured interface. >>>> >>>> Shouldn't this be a runtime option? >>>> >>> Why? What is the advantage of using sync delivery when HW can do it >>> async? >> >> What's the advantage of having an option at all then? Who selects it? >> > x86 is stupid and cannot deliver the even asynchronously. Platform that > can do it select the option. We're in generic code. S390x enables it. X86 does not. That was the missing link! Thanks a lot and sorry for the fuss :). Alex -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href