Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20211101124310.3947887-1-yangyingliang@huawei.com>
 <CAADnVQJS_2St=iaqHU+zasy_0A0bidJN=STnkHrNcSNL5vO1Dg@mail.gmail.com>
 <90518c5d-36ea-ec97-9f14-0687fdd6074f@fb.com> <CAADnVQ+aeAnBEN=dp92q0RBXT+Um1ha4_F=sQ7fr08Sa3qauLQ@mail.gmail.com>
In-Reply-To: <CAADnVQ+aeAnBEN=dp92q0RBXT+Um1ha4_F=sQ7fr08Sa3qauLQ@mail.gmail.com>
From:   Andrii Nakryiko <andrii.nakryiko@gmail.com>
Date:   Mon, 1 Nov 2021 19:46:56 -0700
Message-ID: <CAEf4BzaCQecivGZuXaVJyERZPg-T6+ewPRKBEvpJ366_ZNe+2A@mail.gmail.com>
Subject: Re: [PATCH -next v2] bpf/benchs: Fix return value check of bpf_program__attach()
To:     Alexei Starovoitov <alexei.starovoitov@gmail.com>
Cc:     Yonghong Song <yhs@fb.com>,
        Yang Yingliang <yangyingliang@huawei.com>,
        Joanne Koong <joannekoong@fb.com>,
        Andrii Nakryiko <andrii@kernel.org>,
        LKML <linux-kernel@vger.kernel.org>, bpf <bpf@vger.kernel.org>,
        "open list:KERNEL SELFTEST FRAMEWORK" 
        <linux-kselftest@vger.kernel.org>, Shuah Khan <shuah@kernel.org>,
        Alexei Starovoitov <ast@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Nov 1, 2021 at 3:30 PM Alexei Starovoitov
<alexei.starovoitov@gmail.com> wrote:
>
> On Mon, Nov 1, 2021 at 3:21 PM Yonghong Song <yhs@fb.com> wrote:
> >
> >
> >
> > On 11/1/21 3:00 PM, Alexei Starovoitov wrote:
> > > On Mon, Nov 1, 2021 at 5:35 AM Yang Yingliang <yangyingliang@huawei.com> wrote:
> > >>
> > >> If bpf_program__attach() fails, it never returns NULL,
> > >> we should use libbpf_get_error() to check the return value.
> > >>
> > >> Reported-by: Hulk Robot <hulkci@huawei.com>
> > >> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
> > >> Acked-by: Yonghong Song <yhs@fb.com>
> > >> ---
> > >> v2:
> > >>    don't use 'int err'
> > >> ---
> > >>   .../selftests/bpf/benchs/bench_bloom_filter_map.c      | 10 +++++-----
> > >>   1 file changed, 5 insertions(+), 5 deletions(-)
> > >>
> > >> diff --git a/tools/testing/selftests/bpf/benchs/bench_bloom_filter_map.c b/tools/testing/selftests/bpf/benchs/bench_bloom_filter_map.c
> > >> index 6eeeed2913e6..4afaa4adb327 100644
> > >> --- a/tools/testing/selftests/bpf/benchs/bench_bloom_filter_map.c
> > >> +++ b/tools/testing/selftests/bpf/benchs/bench_bloom_filter_map.c
> > >> @@ -304,7 +304,7 @@ static void bloom_lookup_setup(void)
> > >>          populate_maps();
> > >>
> > >>          link = bpf_program__attach(ctx.skel->progs.bloom_lookup);
> > >> -       if (!link) {
> > >> +       if (libbpf_get_error(link)) {
> > >
> > > Please use ASSERT_OK_PTR() instead.
> > > See how other tests are doing it.
> >
> > I actually looked at this. ASSERT_OK_PTR() is defined in test_progs.h
> > and test_progs.h is ONLY included in files which eventually linked to
> > test_progs. That is why I didn't recommend to use ASSERT_OK_PTR().
> >
> > Maybe it is okay to include test_progs.h in benchs/*.c. Or we may
> > want to refactor to a separate header file to contain these macros
> > which can be used for test_progs.h and other applications.
>
> hmm.
> Looks like bench_ringbufs.c has the same issue doing:
> if (!link)
> and bench_rename.c too.

bench.c does:

libbpf_set_strict_mode(LIBBPF_STRICT_ALL);

and so on error all the pointers will be NULL. So it's ok to check if
(!link) and not use libbpf_get_error() at all.

>
> Probably would be good to fix in all bench-s.
>
> If test_progs.h cannot be included directly
> copy-pasting ASSERT_OK_PTR in a reduced form into bench.h
> is probably cleaner than open coding libbpf_get_error.