On Wed, Jul 30, 2008 at 10:42:12AM -0500, Christoph Lameter wrote: > Andrea Arcangeli wrote: > > > I think the current implementation is fine for the long run, it can > > provide the fastest performance when armed, and each invalidate either > > requires IPIs or it may may need to access the southbridge, so when > > freeing large areas of memory it's good being able to do a single > > invalidate. > > Right. A couple of months ago we had this discussion and agreed that the begin / end was the way to go. I still support that decision. That's OK. We don't have to make decisions just by people supporting one way or the other, because I'll come up with some competing patches and if they turn out to be significantly simpler to the core VM without having a significant negative impact on performance then naturally everybody should be happy to merge them, so nobody has to argue with handwaving. -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html