Re: [PATCH bpf 2/2] selftests/bpf: Add tests for allowed helpers

From: Dmitrii Banshchikov <me@ubique.spb.ru>
To: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Cc: bpf <bpf@vger.kernel.org>, Alexei Starovoitov <ast@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Andrii Nakryiko <andrii@kernel.org>, Martin Lau <kafai@fb.com>,
	Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
	john fastabend <john.fastabend@gmail.com>,
	KP Singh <kpsingh@kernel.org>,
	Networking <netdev@vger.kernel.org>, Andrey Ignatov <rdna@fb.com>,
	john.stultz@linaro.org, sboyd@kernel.org,
	Peter Ziljstra <peterz@infradead.org>,
	Mark Rutland <mark.rutland@arm.com>,
	rosted@goodmis.org
Subject: Re: [PATCH bpf 2/2] selftests/bpf: Add tests for allowed helpers
Date: Wed, 10 Nov 2021 11:32:35 +0400	[thread overview]
Message-ID: <20211110073235.4cwxqxeit3hgdluf@amnesia> (raw)
In-Reply-To: <CAEf4BzbaFSwSA9R1FgeD=CXdOi3iWW1QR7cF0jEnRmw6tZpiAQ@mail.gmail.com>

On Tue, Nov 09, 2021 at 05:16:14PM -0800, Andrii Nakryiko wrote:
> On Mon, Nov 8, 2021 at 10:48 PM Dmitrii Banshchikov <me@ubique.spb.ru> wrote:
> >
> > On Mon, Nov 08, 2021 at 08:46:20PM +0400, Dmitrii Banshchikov wrote:
> > > This patch adds tests that bpf_ktime_get_coarse_ns() and bpf_timer_* and
> > > bpf_spin_lock()/bpf_spin_unlock() helpers are forbidden in tracing
> > > progs as it may result in various locking issues.
> > >
> > > Signed-off-by: Dmitrii Banshchikov <me@ubique.spb.ru>
> > > ---
> > >  tools/testing/selftests/bpf/test_verifier.c   |  36 +++-
> > >  .../selftests/bpf/verifier/helper_allowed.c   | 196 ++++++++++++++++++
> > >  2 files changed, 231 insertions(+), 1 deletion(-)
> > >  create mode 100644 tools/testing/selftests/bpf/verifier/helper_allowed.c
> > >
> > > diff --git a/tools/testing/selftests/bpf/test_verifier.c b/tools/testing/selftests/bpf/test_verifier.c
> > > index 25afe423b3f0..e16eab6fc3a9 100644
> > > --- a/tools/testing/selftests/bpf/test_verifier.c
> > > +++ b/tools/testing/selftests/bpf/test_verifier.c
> > > @@ -92,6 +92,7 @@ struct bpf_test {
> > >       int fixup_map_event_output[MAX_FIXUPS];
> > >       int fixup_map_reuseport_array[MAX_FIXUPS];
> > >       int fixup_map_ringbuf[MAX_FIXUPS];
> > > +     int fixup_map_timer[MAX_FIXUPS];
> > >       /* Expected verifier log output for result REJECT or VERBOSE_ACCEPT.
> > >        * Can be a tab-separated sequence of expected strings. An empty string
> > >        * means no log verification.
> > > @@ -605,7 +606,7 @@ static int create_cgroup_storage(bool percpu)
> > >   *   struct bpf_spin_lock l;
> > >   * };
> > >   */
> > > -static const char btf_str_sec[] = "\0bpf_spin_lock\0val\0cnt\0l";
> > > +static const char btf_str_sec[] = "\0bpf_spin_lock\0val\0cnt\0l\0bpf_timer\0";
> >
> > There is extra null byte at the end.
> 
> Won't hurt, probably. But I wonder if it will be much easier to add
> all those programs as C code and test from test_progs? Instead of all
> this assembly.
> 
> You can put all of them into a single file and have loop that disabled
> all but one program at a time (using bpf_program__set_autoload()) and
> loading it and validating that the load failed. WDYT?

Will give it a try, thanks.

-- 

Dmitrii Banshchikov