Messages in this thread | | | Date | Tue, 7 Oct 2014 10:32:36 -0300 | From | Arnaldo Carvalho de Melo <> | Subject | Re: [PATCH 1/1] rasd: Use perf_evlist__open() instead of open coded |
| |
Em Tue, Oct 07, 2014 at 10:45:14AM +0200, Jean Pihet escreveu: > > --- a/src/rasd.c > > - evlist__for_each(evlist, c) { > > - /* On all online cpus by default, system wide tracing */ > > - if (perf_evsel__open(c, evlist->cpus, NULL) < 0) > > - err("opening tracepoint, are you root?"); > > - } > > - perf_evlist__set_id_pos(evlist); > > + /* On all online cpus by default, system wide tracing */ > > + if (perf_evlist__open(evlist) < 0) > > + err("opening tracepoint, are you root?");
> That way the system wide tracing does not work. perf_evlist__open uses > a non-NULL thread mapping and so it only traces the events generated > by the daemon itself.
> perf top uses a similar approach but uses a lot of code in machine.c > (machine__synthesize_threads) and util/event.c > (perf_event__synthesize_threads) to synthesize the threads etc., which > we want to avoid in the minimalistic approach of rasd.
Humm, I will look into making it support this usecase, the point of evsel/evlist is to try to abstract away as much as possible, leaving just a few methods to be used by tools.
I think the default for perf_evlist__open() should be the most useful for the majority of tools, which I thought would serve rasd well. It is not the case, so I'll into how the tools that currently use perf_evlist__open() behave and try to get a sane default in place.
Hopefully we will completely remove the need to set up any thread or cpu map, as what you want is syswide tracing, right?
- Arnaldo
> Maybe I missed something about the system wide tracing, any suggestion > is welcome. > > Thanks for looking! > > Jean > > > > > /* mmap buffers */ > > if (perf_evlist__mmap(evlist, 4 /* opts->mmap_pages */, false) < 0) > > -- > > 1.9.3 > >
| |