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: <e3b30905-4497-29b4-4636-a313283dbc56@i-love.sakura.ne.jp> <20200528065603.3596-1-penguin-kernel@I-love.SAKURA.ne.jp>
In-Reply-To: <20200528065603.3596-1-penguin-kernel@I-love.SAKURA.ne.jp>
From:   Geert Uytterhoeven <geert@linux-m68k.org>
Date:   Mon, 8 Jun 2020 18:39:53 +0200
Message-ID: <CAMuHMdWOjCtV9W7GtHrhKCJwJ6wNSpd_RQTBaoG7sVrb4oYSDA@mail.gmail.com>
Subject: Re: [PATCH v2] twist: allow converting pr_devel()/pr_debug() into snprintf()
To:     Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Dmitry Vyukov <dvyukov@google.com>,
        Ondrej Mosnacek <omosnace@redhat.com>,
        Petr Mladek <pmladek@suse.com>,
        Sergey Senozhatsky <sergey.senozhatsky@gmail.com>,
        Steven Rostedt <rostedt@goodmis.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Tetsuo,

On Thu, May 28, 2020 at 8:57 AM Tetsuo Handa
<penguin-kernel@i-love.sakura.ne.jp> wrote:
> syzbot found a NULL pointer dereference bug inside mptcp_recvmsg() due to
> ssock == NULL, but this bug manifested inside selinux_socket_recvmsg()
> because pr_debug() was no-op [1].
>
>   pr_debug("fallback-read subflow=%p",
>            mptcp_subflow_ctx(ssock->sk));
>   copied = sock_recvmsg(ssock, msg, flags);
>
> Thus, let's allow fuzzers to always evaluate pr_devel()/pr_debug()
> messages, by redirecting no-op pr_devel()/pr_debug() calls to snprintf().
>
> [1] https://syzkaller.appspot.com/bug?id=12be9aa373be9d8727cdd172f190de39528a413a
>
> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>

Thanks for your patch!

> --- a/lib/Kconfig.twist
> +++ b/lib/Kconfig.twist
> @@ -12,10 +12,22 @@ if TWIST_KERNEL_BEHAVIOR
>
>  config TWIST_FOR_SYZKALLER_TESTING
>         bool "Select all twist options suitable for syzkaller testing"
> +       select TWIST_ALWAYS_EVALUATE_PRINTK_ARGUMENTS
>         select TWIST_DISABLE_KBD_K_SPEC_HANDLER
>         help
>          Say N unless you are building kernels for syzkaller's testing.
>
> +config TWIST_ALWAYS_EVALUATE_PRINTK_ARGUMENTS
> +       bool "Always evaluate printk() arguments"
> +       help
> +        Currently, only format string of printk() arguments is checked
> +        by compiler if pr_devel()/pr_debug() are disabled. Therefore,
> +        fuzz testing cannot catch runtime bugs (e.g. NULL pointer
> +        dereference, use-after-free/out-of-bounds/uninitialized read)
> +        in disabled printk() calls. This option redirects disabled
> +        printk(...) to snprintf(NULL, 0, ...) in order to evaluate
> +        arguments without printing.
> +
>  config TWIST_DISABLE_KBD_K_SPEC_HANDLER
>         bool "Disable k_spec() function in drivers/tty/vt/keyboard.c"
>         help

Can't you just enable CONFIG_DYNAMIC_DEBUG in your fuzzer
config instead?

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds