Re: [PATCH 10/28] perf report: Cache cumulative callchains

[Jiri Olsa <jolsa@redhat.com>]

On Tue, Jan 14, 2014 at 08:55:50AM +0900, Namhyung Kim wrote:
> Hi Jiri,
> 
> Thanks for reminding me. :)
> 
> On Thu, 9 Jan 2014 19:06:28 +0100, Jiri Olsa wrote:
> > On Wed, Jan 08, 2014 at 05:46:15PM +0900, Namhyung Kim wrote:
> 
> [SNIP]
> >> +	/*
> >> +	 * Check if there's duplicate entries in the callchain.
> >> +	 * It's possible that it has cycles or recursive calls.
> >> +	 */
> >> +	for (i = 0; i < iter->curr; i++) {
> >> +		if (hist_entry__cmp(he_cache[i], &he_tmp) == 0)
> >
> > we need here:
> > 			iter->he = he_cache[i];
> >
> >> +			return 0;
> >> +	}
> >
> > otherwise iter->he and al are not in sync and 
> > hist_entry__inc_addr_samples fails in hist_iter_cb
> 
> Right.  But the point of the he_cache is to skip duplicate entries.  So
> I'd like to change it like setting it to NULL and check it before the
> callback function:
> 
> 	while (iter->next_entry(iter, al)) {
> 		err = iter->add_next_entry(iter, al);
>                 if (err)
> 			break;
> 
> 		if (iter->he && iter->add_entry_cb) {
> 			err = iter->add_entry_cb(iter, al, ...);
> 			if (err)
> 				break;
> 		}
>         }
> 
> What do you think?

yep, better ;-) I saw you sent v6.. I'll check that

jirka