Hi Andi, On Mon, 13 May 2013 16:11:17 +0200, Andi Kleen wrote: > On Mon, May 13, 2013 at 05:41:42PM +0900, Namhyung Kim wrote: >> From: Namhyung Kim <[email protected]> >> >> The --percent-limit option is for not showing small overheaded entries >> in the output. Maybe we want to set a certain default value like 0.1. >> >> Cc: Andi Kleen <[email protected]> > > I like the option, but 0 is not a good default. > Make it 1 or something.
Thanks. I just wanted to reserve same behavior since I didn't know what's the reasonable default value. We might use a config variable for this to customize the default value. > > Perhaps a better heuristic would be > percent < threshold || samples < sample-threshold > with sample-threshold being several hundred at least. > > A very small number of hits on something is usually quite much useless, > unless your original event was extremly rare in the beginning > (but then it's unclear if the profile is any useful) Looks reasonable, but I know there's also some users who use the perf tools lightly. A several hundreds sample-threshold looks too harsh to them IMHO. In addition, it should handle tracepoints. Thanks, Namhyung -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
