Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220930110900.75492-1-asavkov@redhat.com> <CAEf4BzZpkgXi9Y6x-_-6mDDW12GvTj0Y_e7cpQMqF3dtiBBhpA@mail.gmail.com>
 <YzqHmHRjxAc4Nndc@samus.usersys.redhat.com>
In-Reply-To: <YzqHmHRjxAc4Nndc@samus.usersys.redhat.com>
From:   Andrii Nakryiko <andrii.nakryiko@gmail.com>
Date:   Mon, 3 Oct 2022 17:03:18 -0700
Message-ID: <CAEf4BzZaGvXM7Vquc=SEM3-cD=s_gfX1jadm4TsGxHnsLG4daw@mail.gmail.com>
Subject: Re: [PATCH bpf-next] selftests/bpf: make libbpf_probe_prog_types
 testcase aware of kernel configuration
To:     Andrii Nakryiko <andrii.nakryiko@gmail.com>,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Andrii Nakryiko <andrii@kernel.org>, bpf@vger.kernel.org,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        jbenc@redhat.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sun, Oct 2, 2022 at 11:56 PM Artem Savkov <asavkov@redhat.com> wrote:
>
> On Fri, Sep 30, 2022 at 04:06:41PM -0700, Andrii Nakryiko wrote:
> > On Fri, Sep 30, 2022 at 4:09 AM Artem Savkov <asavkov@redhat.com> wrote:
> > >
> > > At the moment libbpf_probe_prog_types test iterates over all available
> > > BPF_PROG_TYPE regardless of kernel configuration which can exclude some
> > > of those. Unfortunately there is no direct way to tell which types are
> > > available, but we can look at struct bpf_ctx_onvert to tell which ones
> > > are available.
> > >
> > > Signed-off-by: Artem Savkov <asavkov@redhat.com>
> > > ---
> >
> > Many selftests assume correct kernel configuration which is encoded in
> > config and config.<arch> files. So it seems fair to assume that all
> > defined program types are available on kernel-under-test.
>
> Ok. Wasn't sure if this is the assumption being made.
>
> > If someone is running selftests under custom more minimal kernel they
> > can use denylist to ignore specific prog type subtests?
>
> Thanks for the suggestion. Denylist is a bit too broad in this case as
> it means we'll be disabling the whole libbpf_probe_prog_types test while
> only a single type is a problem. Looks like we'll have to live with a
> downstream-only patch in this case.

Allow/deny lists allow to specify subtests as well, so you can have
very granular control. E.g.,

[vmuser@archvm bpf]$ sudo ./test_progs -a 'libbpf_probe_prog_types/*SK*'
Failed to load bpf_testmod.ko into the kernel: -22
WARNING! Selftests relying on bpf_testmod.ko will be skipped.
#96/8    libbpf_probe_prog_types/BPF_PROG_TYPE_CGROUP_SKB:OK
#96/14   libbpf_probe_prog_types/BPF_PROG_TYPE_SK_SKB:OK
#96/16   libbpf_probe_prog_types/BPF_PROG_TYPE_SK_MSG:OK
#96/21   libbpf_probe_prog_types/BPF_PROG_TYPE_SK_REUSEPORT:OK
#96/30   libbpf_probe_prog_types/BPF_PROG_TYPE_SK_LOOKUP:OK
#96      libbpf_probe_prog_types:OK
Summary: 1/5 PASSED, 0 SKIPPED, 0 FAILED


As you can see each program type is a subtest, so you can pick and
choose which ones to run.

>
> --
>  Artem
>