[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAG_fn=X-ednJduk3k6JEA9mMw+DrT3uSJ=+WY269gY6p45fJ0w@mail.gmail.com>
Date: Wed, 3 Jun 2020 18:33:40 +0200
From: Alexander Potapenko <glider@...gle.com>
To: Edward Cree <ecree@...arflare.com>
Cc: Alexei Starovoitov <alexei.starovoitov@...il.com>,
Daniel Borkmann <daniel@...earbox.net>,
Michal Kubecek <mkubecek@...e.cz>,
Alexei Starovoitov <ast@...nel.org>,
Dmitriy Vyukov <dvyukov@...gle.com>,
Networking <netdev@...r.kernel.org>
Subject: Re: Self-XORing BPF registers is undefined behavior
On Wed, Jun 3, 2020 at 5:37 PM Edward Cree <ecree@...arflare.com> wrote:
>
> On 02/06/2020 18:32, Alexei Starovoitov wrote:
> > The target for bpf codegen is JITs.
> > bpf interpreter is simulating hw.
> > For now if you want UB fuzzer running in your environment please add
> > _out_of_tree_ patch that inits all interpreter registers to zero.
> +1 to all the above.
Noted, thank you.
> Also, note that you can still fuzz BPF JITs by building the kernel
> without the interpreter: CONFIG_BPF_JIT_ALWAYS_ON.
Unfortunately KMSAN doesn't play well with JITed code. To be able to
detect uninit bugs in JIT, we'll need to instrument the generated code
as well.
--
Alexander Potapenko
Software Engineer
Google Germany GmbH
Erika-Mann-Straße, 33
80636 München
Geschäftsführer: Paul Manicle, Halimah DeLaine Prado
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Powered by blists - more mailing lists