Re: [PATCH bpf-next 1/5] selftests/bpf: add remaining ASSERT_xxx() variants

From: Andrii Nakryiko <andrii.nakryiko@gmail.com>
To: "Toke Høiland-Jørgensen" <toke@redhat.com>
Cc: Lorenz Bauer <lmb@cloudflare.com>,
	Andrii Nakryiko <andrii@kernel.org>, bpf <bpf@vger.kernel.org>,
	Networking <netdev@vger.kernel.org>,
	Alexei Starovoitov <ast@fb.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Kernel Team <kernel-team@fb.com>
Subject: Re: [PATCH bpf-next 1/5] selftests/bpf: add remaining ASSERT_xxx() variants
Date: Mon, 26 Apr 2021 09:15:07 -0700	[thread overview]
Message-ID: <CAEf4BzYixzoqzE_c+sd7QoQDg8dGaKf_UBf06AqTmCdUagoJvg@mail.gmail.com> (raw)
In-Reply-To: <875z09ca0p.fsf@toke.dk>

On Mon, Apr 26, 2021 at 8:59 AM Toke Høiland-Jørgensen <toke@redhat.com> wrote:
>
> Andrii Nakryiko <andrii.nakryiko@gmail.com> writes:
>
> > On Mon, Apr 26, 2021 at 1:06 AM Lorenz Bauer <lmb@cloudflare.com> wrote:
> >>
> >> On Sat, 24 Apr 2021 at 00:36, Andrii Nakryiko <andrii@kernel.org> wrote:
> >> >
> >> > Add ASSERT_TRUE/ASSERT_FALSE for conditions calculated with custom logic to
> >> > true/false. Also add remaining arithmetical assertions:
> >> >   - ASSERT_LE -- less than or equal;
> >> >   - ASSERT_GT -- greater than;
> >> >   - ASSERT_GE -- greater than or equal.
> >> > This should cover most scenarios where people fall back to error-prone
> >> > CHECK()s.
> >> >
> >> > Also extend ASSERT_ERR() to print out errno, in addition to direct error.
> >> >
> >> > Also convert few CHECK() instances to ensure new ASSERT_xxx() variants work as
> >> > expected. Subsequent patch will also use ASSERT_TRUE/ASSERT_FALSE more
> >> > extensively.
> >> >
> >> > Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
> >> > ---
> >> >  .../selftests/bpf/prog_tests/btf_dump.c       |  2 +-
> >> >  .../selftests/bpf/prog_tests/btf_endian.c     |  4 +-
> >> >  .../selftests/bpf/prog_tests/cgroup_link.c    |  2 +-
> >> >  .../selftests/bpf/prog_tests/kfree_skb.c      |  2 +-
> >> >  .../selftests/bpf/prog_tests/resolve_btfids.c |  7 +--
> >> >  .../selftests/bpf/prog_tests/snprintf_btf.c   |  4 +-
> >> >  tools/testing/selftests/bpf/test_progs.h      | 50 ++++++++++++++++++-
> >> >  7 files changed, 56 insertions(+), 15 deletions(-)
> >> >
> >> > diff --git a/tools/testing/selftests/bpf/prog_tests/btf_dump.c b/tools/testing/selftests/bpf/prog_tests/btf_dump.c
> >> > index c60091ee8a21..5e129dc2073c 100644
> >> > --- a/tools/testing/selftests/bpf/prog_tests/btf_dump.c
> >> > +++ b/tools/testing/selftests/bpf/prog_tests/btf_dump.c
> >> > @@ -77,7 +77,7 @@ static int test_btf_dump_case(int n, struct btf_dump_test_case *t)
> >> >
> >> >         snprintf(out_file, sizeof(out_file), "/tmp/%s.output.XXXXXX", t->file);
> >> >         fd = mkstemp(out_file);
> >> > -       if (CHECK(fd < 0, "create_tmp", "failed to create file: %d\n", fd)) {
> >> > +       if (!ASSERT_GE(fd, 0, "create_tmp")) {
> >>
> >> Nit: I would find ASSERT_LE easier to read here. Inverting boolean
> >> conditions is easy to get wrong.
> >
> > You mean if (ASSERT_LE(fd, -1, "create_tmp")) { err = fd; goto done; } ?
> >
> > That will mark the test failing if fd >= 0, which is exactly opposite
> > to what we wan't. It's confusing because CHECK() checks invalid
> > conditions and returns "true" if it holds. But ASSERT_xxx() checks
> > *valid* condition and returns whether valid condition holds. So the
> > pattern is always
>
> There's already an ASSERT_OK_PTR(), so maybe a corresponding
> ASSERT_OK_FD() would be handy?

I honestly don't see the point. OK_PTR is special, it checks NULL and
the special ERR_PTR() variants, which is a lot of hassle to check
manually. While for FD doing ASSERT_GE(fd, 0) seems to be fine and
just mostly natural. Also for some APIs valid FD is > 0 and for other
cases valid FD is plain >= 0, so that just adds to the confusion.

>
> -Toke
>