Messages in this thread | | | Date | Tue, 10 Dec 2013 13:24:49 +0100 | From | Ingo Molnar <> | Subject | Re: [GIT PULL 00/21] perf/core improvements and fixes |
| |
* Arnaldo Carvalho de Melo <acme@ghostprotocols.net> wrote:
> Em Tue, Dec 10, 2013 at 02:22:58PM +0200, Adrian Hunter escreveu: > > On 10/12/13 14:10, Arnaldo Carvalho de Melo wrote: > > > Em Tue, Dec 10, 2013 at 02:13:12PM +0200, Adrian Hunter escreveu: > > >> On 10/12/13 13:44, Arnaldo Carvalho de Melo wrote: > > >>> Em Tue, Dec 10, 2013 at 12:12:29PM +0100, Ingo Molnar escreveu: > > >>>> > > >>>> Hm, I've unpulled it because 'perf top' crashes on exit, in > > >>>> dso__delete(): > > >>> > > >>> 495 if (dso->sname_alloc) > > >>> 496 free((char *)dso->short_name) > > >>> > > >>> Yeah, must be that basename() patch from Stephane, I'll work on a fix > > >>> and resubmit this batch, thanks for the report. > > >> > > >> The problem is sname_alloc is not maintained. Perhaps it should be > > >> set in dso__set_short_name() e.g. > > > > > > Yeah, sounds better than having all callers manage that thing, quickie, > > > was this with Stephane's patch applied? > > > > Yes it was at > > e993d10caeb6dca690dbaf86e1981ba240d1414a > > perf symbols: fix bug in usage of the basename() function > > Yes, this is the buggy patch, my question was if Ingo did the > changes that streamlined the dso->sname_alloc management with > e993d10caeb6 applied to his working tree.
My current perf/core head is:
789790791ad2 tools/perf/build: Fix install dependency
which does not have e993d10c.
[ Btw., a small nit: the capitalization of the commit title looks inconsistent. ]
Thanks,
Ingo
| |