Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <1559117459-27353-1-git-send-email-92siuyang@gmail.com> <CANn89iLxxiX+4E7EURNKb=xRkk97rPaKTkpSc6Yu7fZbiwPT6w@mail.gmail.com>
In-Reply-To: <CANn89iLxxiX+4E7EURNKb=xRkk97rPaKTkpSc6Yu7fZbiwPT6w@mail.gmail.com>
From:   Yang Xiao <92siuyang@gmail.com>
Date:   Wed, 29 May 2019 23:11:17 +0800
Message-ID: <CAKgHYH2uW=iSUM1j5pLhaQXpm35XK2rWq45M2Yih1-Dn=es0SA@mail.gmail.com>
Subject: Re: [PATCH] ipv4: tcp_input: fix stack out of bounds when parsing TCP options.
To:     Eric Dumazet <edumazet@google.com>
Cc:     David Miller <davem@davemloft.net>,
        Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
        netdev <netdev@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Indeed, condition opsize < 2 and opsize > length can deduce that length >= 2.
However, before the condition (if opsize < 2), there may be one-byte
out-of-bound access in line 12.
I'm not sure whether I have put it very clearly.

On Wed, May 29, 2019 at 10:20 PM Eric Dumazet <edumazet@google.com> wrote:
>
> On Wed, May 29, 2019 at 1:10 AM Young Xiao <92siuyang@gmail.com> wrote:
> >
> > The TCP option parsing routines in tcp_parse_options function could
> > read one byte out of the buffer of the TCP options.
> >
> > 1         while (length > 0) {
> > 2                 int opcode = *ptr++;
> > 3                 int opsize;
> > 4
> > 5                 switch (opcode) {
> > 6                 case TCPOPT_EOL:
> > 7                         return;
> > 8                 case TCPOPT_NOP:        /* Ref: RFC 793 section 3.1 */
> > 9                         length--;
> > 10                        continue;
> > 11                default:
> > 12                        opsize = *ptr++; //out of bound access
> >
> > If length = 1, then there is an access in line2.
> > And another access is occurred in line 12.
> > This would lead to out-of-bound access.
> >
> > Therefore, in the patch we check that the available data length is
> > larger enough to pase both TCP option code and size.
> >
> > Signed-off-by: Young Xiao <92siuyang@gmail.com>
> > ---
> >  net/ipv4/tcp_input.c | 2 ++
> >  1 file changed, 2 insertions(+)
> >
> > diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
> > index 20f6fac..9775825 100644
> > --- a/net/ipv4/tcp_input.c
> > +++ b/net/ipv4/tcp_input.c
> > @@ -3791,6 +3791,8 @@ void tcp_parse_options(const struct net *net,
> >                         length--;
> >                         continue;
> >                 default:
> > +                       if (length < 2)
> > +                               return;
> >                         opsize = *ptr++;
> >                         if (opsize < 2) /* "silly options" */
> >                                 return;
>
> In practice we are good, since we have at least 320 bytes of room there,
> and the test done later catches silly options.
>
> if (opsize < 2) /* "silly options" */
>     return;
> if (opsize > length)   /* remember, opsize >= 2 here */
>      return; /* don't parse partial options */
>
> I guess adding yet another conditional will make this code obviously
> correct for all eyes
> and various tools.
>
> Thanks.
>
> Signed-off-by: Eric Dumazet <edumazet@google.com>


-- 
Best regards!

Young
-----------------------------------------------------------