On Thu, 28 Apr 2011, Tejun Heo wrote: > On Thu, Apr 28, 2011 at 05:31:01PM +0200, Tejun Heo wrote: > > > The hugely expensive _sum() is IMHO pretty useless given the above. It is > > > a function that is called with the *hope* of getting a more accurate > > > result. > ... > > Christoph, I'm sorry but I really can't explain it any better and am > > out of this thread. If you still wanna proceed, you'll need to route > > the patches yourself. Please keep me cc'd. > > Oh, another way to proceed would be, if _sum() is as useless as you > suggested above, remove _sum() first and see how that goes. If that > flies, there will be no reason to argue over anything, right? There is no point of arguing anymore since you do not accept that the percpu counters do not have atomic_t consistency properties. The relaxing of the consistency requirements in order to increase performance was the intend when these schemes were created. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxxx For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>