| Message ID | 20211006185619.364369-14-fallentree@fb.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | BPF |
| Headers | show |
| Series | selftests/bpf: Add parallelism to test_progs | expand |
On Wed, Oct 6, 2021 at 11:56 AM Yucong Sun <fallentree@fb.com> wrote: > > From: Yucong Sun <sunyucong@gmail.com> > > This make this test more likely to succeed. > > Signed-off-by: Yucong Sun <sunyucong@gmail.com> > --- 100 million iterations seems a bit excessive. Why one million loops doesn't cause a single perf event? Can we make it more robust in some other way that is not as slow? I've dropped it for now while we discuss. > tools/testing/selftests/bpf/prog_tests/perf_branches.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/tools/testing/selftests/bpf/prog_tests/perf_branches.c b/tools/testing/selftests/bpf/prog_tests/perf_branches.c > index 6b2e3dced619..d7e88b2c5f36 100644 > --- a/tools/testing/selftests/bpf/prog_tests/perf_branches.c > +++ b/tools/testing/selftests/bpf/prog_tests/perf_branches.c > @@ -16,7 +16,7 @@ static void check_good_sample(struct test_perf_branches *skel) > int duration = 0; > > if (CHECK(!skel->bss->valid, "output not valid", > - "no valid sample from prog")) > + "no valid sample from prog\n")) > return; > > /* > @@ -46,7 +46,7 @@ static void check_bad_sample(struct test_perf_branches *skel) > int duration = 0; > > if (CHECK(!skel->bss->valid, "output not valid", > - "no valid sample from prog")) > + "no valid sample from prog\n")) > return; > > CHECK((required_size != -EINVAL && required_size != -ENOENT), > @@ -84,7 +84,7 @@ static void test_perf_branches_common(int perf_fd, > if (CHECK(err, "set_affinity", "cpu #0, err %d\n", err)) > goto out_destroy; > /* spin the loop for a while (random high number) */ > - for (i = 0; i < 1000000; ++i) > + for (i = 0; i < 100000000; ++i) > ++j; > > test_perf_branches__detach(skel); > -- > 2.30.2 >
On Fri, Oct 8, 2021 at 3:27 PM Andrii Nakryiko <andrii.nakryiko@gmail.com> wrote: > > On Wed, Oct 6, 2021 at 11:56 AM Yucong Sun <fallentree@fb.com> wrote: > > > > From: Yucong Sun <sunyucong@gmail.com> > > > > This make this test more likely to succeed. > > > > Signed-off-by: Yucong Sun <sunyucong@gmail.com> > > --- > > 100 million iterations seems a bit excessive. Why one million loops > doesn't cause a single perf event? Can we make it more robust in some > other way that is not as slow? I've dropped it for now while we > discuss. I don't know, without this patch the test constantly fails for me regardless of serial or parallel mode. I think it could be something related to compiler optimizations or hardware? > > > > tools/testing/selftests/bpf/prog_tests/perf_branches.c | 6 +++--- > > 1 file changed, 3 insertions(+), 3 deletions(-) > > > > diff --git a/tools/testing/selftests/bpf/prog_tests/perf_branches.c b/tools/testing/selftests/bpf/prog_tests/perf_branches.c > > index 6b2e3dced619..d7e88b2c5f36 100644 > > --- a/tools/testing/selftests/bpf/prog_tests/perf_branches.c > > +++ b/tools/testing/selftests/bpf/prog_tests/perf_branches.c > > @@ -16,7 +16,7 @@ static void check_good_sample(struct test_perf_branches *skel) > > int duration = 0; > > > > if (CHECK(!skel->bss->valid, "output not valid", > > - "no valid sample from prog")) > > + "no valid sample from prog\n")) > > return; > > > > /* > > @@ -46,7 +46,7 @@ static void check_bad_sample(struct test_perf_branches *skel) > > int duration = 0; > > > > if (CHECK(!skel->bss->valid, "output not valid", > > - "no valid sample from prog")) > > + "no valid sample from prog\n")) > > return; > > > > CHECK((required_size != -EINVAL && required_size != -ENOENT), > > @@ -84,7 +84,7 @@ static void test_perf_branches_common(int perf_fd, > > if (CHECK(err, "set_affinity", "cpu #0, err %d\n", err)) > > goto out_destroy; > > /* spin the loop for a while (random high number) */ > > - for (i = 0; i < 1000000; ++i) > > + for (i = 0; i < 100000000; ++i) > > ++j; > > > > test_perf_branches__detach(skel); > > -- > > 2.30.2 > >
diff --git a/tools/testing/selftests/bpf/prog_tests/perf_branches.c b/tools/testing/selftests/bpf/prog_tests/perf_branches.c index 6b2e3dced619..d7e88b2c5f36 100644 --- a/tools/testing/selftests/bpf/prog_tests/perf_branches.c +++ b/tools/testing/selftests/bpf/prog_tests/perf_branches.c @@ -16,7 +16,7 @@ static void check_good_sample(struct test_perf_branches *skel) int duration = 0; if (CHECK(!skel->bss->valid, "output not valid", - "no valid sample from prog")) + "no valid sample from prog\n")) return; /* @@ -46,7 +46,7 @@ static void check_bad_sample(struct test_perf_branches *skel) int duration = 0; if (CHECK(!skel->bss->valid, "output not valid", - "no valid sample from prog")) + "no valid sample from prog\n")) return; CHECK((required_size != -EINVAL && required_size != -ENOENT), @@ -84,7 +84,7 @@ static void test_perf_branches_common(int perf_fd, if (CHECK(err, "set_affinity", "cpu #0, err %d\n", err)) goto out_destroy; /* spin the loop for a while (random high number) */ - for (i = 0; i < 1000000; ++i) + for (i = 0; i < 100000000; ++i) ++j; test_perf_branches__detach(skel);