Re: [PATCH 3/8] perf test: Use a test program in perf record tests

[Namhyung Kim <namhyung@kernel.org>]

On Mon, Oct 24, 2022 at 4:33 AM Arnaldo Carvalho de Melo
<acme@kernel.org> wrote:
>
> Em Thu, Oct 20, 2022 at 04:52:14PM -0700, Ian Rogers escreveu:
> > On Thu, Oct 20, 2022 at 10:26 AM Namhyung Kim <namhyung@kernel.org> wrote:
>
> > > If the system has cc it could build a test program with two threads
> > > and then use it for more detailed testing.  Also it accepts an option
> > > to run a thread forever to ensure multi-thread runs.
> > >
> > > If cc is not found, it falls back to use the default value 'true'.
> > >
> > > Signed-off-by: Namhyung Kim <namhyung@kernel.org>
> >
> > Acked-by: Ian Rogers <irogers@google.com>
> >
> > I wonder if these utilities should just be built into perf to avoid
> > the cc dependency. Perhaps we can have a hidden option built into perf
> > test.
>
> Agreed, not depending on a compiler makes 'perf test' usable in more
> systems, particularly production ones where we may want to check if perf
> is passing all tests applicable to that system.

Good idea.

I think we can add an option (like -w/--workload) to run the specified
workload rather than executing the tests.  Like below (assuming we
have a workload called 'noploop').  Thoughts?

  $ perf stat -- perf test -w noploop

Thanks,
Namhyung