Re: [PATCH RFC V4 1/6] perf tools: lock to protect namespaces and comm list

From: Arnaldo Carvalho de Melo
Date: Mon Oct 02 2017 - 15:35:00 EST


Em Mon, Oct 02, 2017 at 07:14:10PM +0000, Liang, Kan escreveu:
> > Em Fri, Sep 29, 2017 at 07:47:52AM -0700, kan.liang@xxxxxxxxx escreveu:
> > > From: Kan Liang <kan.liang@xxxxxxxxx>
> > >
> > > Add two locks to protect namespaces_list and comm_list.
> > >
> > > The lock is only needed for multithreaded code, so using mutex
> > > wrappers provided by perf tool.
> > >
> > > Not all the comm_list/namespaces_list accessing are protected, e.g.
> > > thread__exec_comm. Because the multithread code for perf top event
> > > synthesizing does not touch them. They don't need a lock.
> >
> > They don't need a lock _now_, ok I think we can proceed that way, this cset
> > should serve as a warning to people working in further multithreading perf.
> >
>
> Do I need to add some comments in the code to give a warning?

Maybe, but then one can use 'git blame' and see the comments on the
patch introducing the locking.

BTW, I'm applied and tested it, pushed to acme/perf/core.

- Arnaldo