On Mon, Oct 25, 2010 at 04:04:26PM +0900, Isaku Yamahata wrote: > On Mon, Oct 25, 2010 at 08:46:38AM +0200, Michael S. Tsirkin wrote: > > On Mon, Oct 25, 2010 at 03:44:01PM +0900, Isaku Yamahata wrote: > > > On Mon, Oct 25, 2010 at 07:49:41AM +0200, Michael S. Tsirkin wrote: > > > > Simplify logic for hotplug notification, by tracking state of the > > > > logical interrupt condition. We then simply use this variable to make > > > > the interrupt decision, according to spec. > > > > > > > > API is made cleaner as we no longer force users to pass in > > > > old slot control value. > > > > > > Thank you for looking into it. > > > Some comments below. > > > > Thanks! Care fixing up the remaining issues? > > I will fold the fixup in. > > Will do.
One other thing: I think when we validate data on load need to take w1c mask into account, not just wmask. > yamahata