Hi. Trond. First sorry for noise. if goto is encountered by branch predictor with switch condition, CPU try to fetch instruction code after label to cpu cache. And Cpu is idiling and CPU clock is watsted. if branch prediction is fail, this cost is more big. but this patch is good as I said before. Thanks. 2012/3/29 Myklebust, Trond <Trond.Myklebust@xxxxxxxxxx>: > On Thu, 2012-03-29 at 08:55 +0900, Namjae Jeon wrote: >> Hi. Trond. >> >> Normally goto statement is not recommended(pipeline and branch >> prediction) it is only recommended as a last resort if there is no >> method to escape this loop. > > ????????????? A goto is an unconditional branch; there is no further > branch prediction involved once once the processor gets the value of > 'err' in that switch statement correct. Exactly how would adding an > extra set+test of exception.retry after the switch() help the branch > predictor in this situation? > > -- > Trond Myklebust > Linux NFS client maintainer > > NetApp > Trond.Myklebust@xxxxxxxxxx > www.netapp.com > -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
