[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAP-5=fVuB02yxjQu2qvEVy9WHbSvVaAgPCpgSY8bXrSHGkr3rg@mail.gmail.com>
Date: Tue, 28 Oct 2025 12:55:56 -0700
From: Ian Rogers <irogers@...gle.com>
To: Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
Namhyung Kim <namhyung@...nel.org>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...nel.org>,
Adrian Hunter <adrian.hunter@...el.com>, linux-perf-users@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] perf test workload: Add thread count argument to thloop
On Tue, Oct 28, 2025 at 12:13 PM Arnaldo Carvalho de Melo
<acme@...nel.org> wrote:
>
> On Tue, Oct 28, 2025 at 08:38:20AM -0700, Ian Rogers wrote:
> > Allow the number of threads for the thloop workload to be increased
> > beyond the normal 2. Add error checking to the parsed time and thread
> > count values.
> >
> > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > ---
> > v2: Perform the pthread_join unconditionally and ensure started
> > threads terminate.
> > ---
> > tools/perf/tests/workloads/thloop.c | 45 ++++++++++++++++++++++++-----
> > 1 file changed, 38 insertions(+), 7 deletions(-)
> >
> > diff --git a/tools/perf/tests/workloads/thloop.c b/tools/perf/tests/workloads/thloop.c
> > index 457b29f91c3e..bd8168f883fb 100644
> > --- a/tools/perf/tests/workloads/thloop.c
> > +++ b/tools/perf/tests/workloads/thloop.c
> > @@ -31,21 +31,52 @@ static void *thfunc(void *arg)
> >
> > static int thloop(int argc, const char **argv)
> > {
> > - int sec = 1;
> > - pthread_t th;
> > + int nt = 2, sec = 1, err = 1;
> > + pthread_t *thread_list = NULL;
> >
> > if (argc > 0)
> > sec = atoi(argv[0]);
> >
> > + if (sec <= 0) {
> > + fprintf(stderr, "Error: seconds (%d) must be >= 1\n", sec);
> > + return 1;
> > + }
> > +
> > + if (argc > 1)
> > + nt = atoi(argv[1]);
> > +
> > + if (nt <= 0) {
> > + fprintf(stderr, "Error: thread count (%d) must be >= 1\n", nt);
> > + return 1;
> > + }
> > +
> > signal(SIGINT, sighandler);
> > signal(SIGALRM, sighandler);
> > - alarm(sec);
> >
> > - pthread_create(&th, NULL, thfunc, test_loop);
> > - test_loop();
> > - pthread_join(th, NULL);
> > + thread_list = calloc(nt, sizeof(pthread_t));
> > + if (thread_list == NULL) {
> > + fprintf(stderr, "Error: malloc failed for %d threads\n", nt);
> > + goto out;
> > + }
> > + for (int i = 1; i < nt; i++) {
>
> Why do you start at 1? What goes in thread[0]? calloc() leaves it at
> NULL, then if you have two threads you'll allocate thread[1], i gets
> incremented, 2 < 2 fails, you get just one thread created, when two were
> asked?
>
> Oh, I see, you use the main thread to run, that test_loop() just before
> the err = 0, its just that you allocate thread[0] for nothing, that
> confused me.
Agreed. It was so I could avoid doing "nt-1" as I mentioned before.
Fwiw, pthread_self will sometimes return 0 for the main thread and so
it isn't entirely inconsistent to do things this way. Not that the
main thread should try to do a pthread_join with itself.
Thanks,
Ian
> - Arnaldo
>
> > + int ret = pthread_create(&thread_list[i], NULL, thfunc, test_loop);
> >
> > - return 0;
> > + if (ret) {
> > + fprintf(stderr, "Error: failed to create thread %d\n", i);
> > + done = 1; // Ensure started threads terminate.
> > + goto out;
> > + }
> > + }
> > + alarm(sec);
> > + test_loop();
> > + err = 0;
> > +out:
> > + for (int i = 1; i < nt; i++) {
> > + if (thread_list && thread_list[i])
> > + pthread_join(thread_list[i], /*retval=*/NULL);
> > + }
> > + free(thread_list);
> > + return err;
> > }
> >
> > DEFINE_WORKLOAD(thloop);
> > --
> > 2.51.1.851.g4ebd6896fd-goog
Powered by blists - more mailing lists