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;
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable
From:   Qian Cai <cai@lca.pw>
Mime-Version: 1.0 (1.0)
Subject: Re: UBSAN: array-index-out-of-bounds in kernel/bpf/arraymap.c:177
Date:   Tue, 19 May 2020 21:55:45 -0400
Message-Id: <FE7742FF-713E-4310-95E7-9B217662E53E@lca.pw>
References: <CAEf4BzZKCh7+2TL8GVetxrOKYCoL0U7jTGsO5CbDExs7Px+bYQ@mail.gmail.com>
Cc:     Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Martin KaFai Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
        Andrii Nakryiko <andriin@fb.com>,
        John Fastabend <john.fastabend@gmail.com>,
        KP Singh <kpsingh@chromium.org>,
        Linux Netdev List <netdev@vger.kernel.org>,
        bpf <bpf@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        clang-built-linux <clang-built-linux@googlegroups.com>,
        Kees Cook <keescook@chromium.org>
In-Reply-To: <CAEf4BzZKCh7+2TL8GVetxrOKYCoL0U7jTGsO5CbDExs7Px+bYQ@mail.gmail.com>
To:     Andrii Nakryiko <andrii.nakryiko@gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On May 19, 2020, at 7:23 PM, Andrii Nakryiko <andrii.nakryiko@gmail.com> w=
rote:
>=20
> I agree, it's bad to have this noise. But again, there is nothing
> wrong with the way it's used in BPF code base. We'd gladly use
> flexible array, if we could. But given we can't, I'd say the proper
> solution (in order of my preference) would be:
>=20
>  - don't trigger false error, if zero-sized array is the member of union;
>  - or have some sort of annotation at field declaration site (not a
> field access site).
>=20
> Is that possible?

I am not a compiler expert, but with my experience with all those compiler i=
nstrumental technology like KCSAN, KASAN and UBSAN, it seems both options yo=
u prop need to modify compilers, i.e., -fsanitize=3Dundefined=