Re: [TLS] DTLS Handshake race condition

Michael Tuexen <Michael.Tuexen@lurchi.franken.de> Mon, 12 August 2013 11:39 UTC

Return-Path: <Michael.Tuexen@lurchi.franken.de>
X-Original-To: tls@ietfa.amsl.com
Delivered-To: tls@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 6FDA321F8C66 for <tls@ietfa.amsl.com>; Mon, 12 Aug 2013 04:39:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level:
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ooaDPqO6zgfr for <tls@ietfa.amsl.com>; Mon, 12 Aug 2013 04:39:18 -0700 (PDT)
Received: from mail-n.franken.de (drew.ipv6.franken.de [IPv6:2001:638:a02:a001:20e:cff:fe4a:feaa]) by ietfa.amsl.com (Postfix) with ESMTP id ACCFB21F9B53 for <tls@ietf.org>; Mon, 12 Aug 2013 03:58:21 -0700 (PDT)
Received: from [192.168.1.200] (p508F2769.dip0.t-ipconnect.de [80.143.39.105]) (Authenticated sender: macmic) by mail-n.franken.de (Postfix) with ESMTP id 317FF1C0C069F; Mon, 12 Aug 2013 12:58:05 +0200 (CEST)
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 6.5 \(1508\))
From: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>
In-Reply-To: <A34A16A6-6D79-4A8B-A4A9-CF05B5C4C5F7@lurchi.franken.de>
Date: Mon, 12 Aug 2013 12:57:42 +0200
Content-Transfer-Encoding: quoted-printable
Message-Id: <DE729DF9-0F66-4758-8A89-891297222C0A@lurchi.franken.de>
References: <1CBCCCAF-163A-474B-8DD0-6634460644C1@lurchi.franken.de> <CAL2p+8QvmiH-L0WYWDdtAirgQ8i_VoaJQUDfNw4dXDZ1xOzq=w@mail.gmail.com> <1F5455F5-439B-4215-9D92-1FDC2FDFBDE7@lurchi.franken.de> <CAL2p+8QiEdGD9inixnPH4U1MjNa-errq6Um5VMHtB1UkjSgBcA@mail.gmail.com> <A34A16A6-6D79-4A8B-A4A9-CF05B5C4C5F7@lurchi.franken.de>
To: Andy Wilson <andrewgwilson@gmail.com>
X-Mailer: Apple Mail (2.1508)
Cc: tls@ietf.org
Subject: Re: [TLS] DTLS Handshake race condition
X-BeenThere: tls@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: "This is the mailing list for the Transport Layer Security working group of the IETF." <tls.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/tls>, <mailto:tls-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/tls>
List-Post: <mailto:tls@ietf.org>
List-Help: <mailto:tls-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/tls>, <mailto:tls-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 12 Aug 2013 11:39:19 -0000

On Aug 12, 2013, at 12:46 PM, Michael Tuexen <Michael.Tuexen@lurchi.franken.de> wrote:

> 
> On Aug 12, 2013, at 12:28 PM, Andy Wilson <andrewgwilson@gmail.com> wrote:
> 
>> Section 4.2.2 states:
>> 
>> The first message each side transmits in each handshake always has
>>   message_seq = 0.  Whenever each new message is generated, the
>>   message_seq value is incremented by one.  Note that in the case of a
>>   rehandshake, this implies that the HelloRequest will have message_seq
>>   = 0 and the ServerHello will have message_seq = 1
>> 
>> This would imply that the client WOULD process the ServerHello with seq==1 as this is a re-handshake.
>> 
>> That's the way i'm reading the spec..
> Me too.
> 
> I think the RFC does cover the race condition I'm referring to...
... I meant "does NOT cover"
> What if the HelloRequest(message_seq=0) is lost and the client sends
> a ClientHello(message_seq=0) on its own (since the local user initates
> a re-handshake). The server accepts the ClientHello and responds with
> a ServerHello(message_seq=1). The client however expects a
> ServerHello(message_seq=0), since it never saw the HelloRequest.
> The HelloRequest is also not retransmitted, since the server considers
> it acked by the ClientHello.
> 
> It is only the collision case I'm considering and I think which is
> not covered by the RFC.
> 
> Best regards
> Michael
>> 
>> 
>> 
>> 
>> On 12 August 2013 05:08, Michael Tuexen <Michael.Tuexen@lurchi.franken.de> wrote:
>> On Aug 11, 2013, at 4:19 PM, Andy Wilson <andrewgwilson@gmail.com> wrote:
>> 
>>> After looking over the RFC a bit, wouldn't the client be expecting a HelloVerifyRequest after its ClientHello?
>> I don't think this is necessary, since the client and the server have already
>> a relation. So there is no need to use the cookie mechanism.
>> 
>> However, the question is the same. In your case
>> * The server sends a HelloRequest(MsgSeqNo = 0) and starts the retransmission
>>  timer, since this is a flight.
>> * The HelloRequest is dropped by the network.
>> * The client sends a ClientHello(MsgSeqNo = 0) and start a retransmission timer,
>>  since it is its first flight.
>> * The server sends a HelloVerifyRequest(MsgSeqNo = 1)
>> * The client doesn't process the ServerHello, since it expects the
>>  MsgSeqNo == 0.
>> 
>> Therefore, the server retransmits the flight consisting of the HelloVerifyRequest
>> and the client retransmits the flight containing the ClientHello.
>> So the solution would be that the client accepts
>> * ServerHellos with MsgSeqNo=0 and MsgSeqNo=1.
>> * HelloVerifyRequest with MsgSeqNo=0 and MsgSeqNo=1.
>> 
>> Am I missing something?
>> 
>> Best regards
>> Michael
>>> 
>>> On 12 August 2013 01:17, Michael Tuexen <Michael.Tuexen@lurchi.franken.de> wrote:
>>> Dear all,
>>> 
>>> while fixing a bug in OpenSSL regarding the DTLS handshake, I thought about
>>> the following scenario (both sides decide to renegotiate at about the same
>>> time):
>>> 
>>> * A DTLS connection is established.
>>> * The server sends a HelloRequest(MsgSeqNo = 0) and starts the retransmission
>>>  timer, since this is a flight.
>>> * The HelloRequest is dropped by the network.
>>> * The client sends a ClientHello(MsgSeqNo = 0) and start a retransmission timer,
>>>  since it is its first flight.
>>> * The server receives the ClientHello, stops the retransmission timer
>>>  and sends the next flight starting with ServerHello(MsgSeqNo = 1)
>>>  since it considers the received ClientHello as an ack for the flight.
>>> * The client doesn't process the ServerHello, since it expects the
>>>  MsgSeqNo == 0.
>>> 
>>> Therefore the client retransmits its ClientHello and the server retransmits
>>> its flight containing the ServerHello. Am I missing something?
>>> 
>>> The problem is that the server has no way to figure out if the received
>>> ClientHello is a reaction to a HelloRequest or not.
>>> The only way out I see is that the client accepts ServerHellos with
>>> MsgSeqNo=0 and MsgSeqNo=1.
>>> I don't think this is covered in http://tools.ietf.org/html/rfc6347
>>> 
>>> Any opinions?
>>> 
>>> Best regards
>>> Michael
>>> _______________________________________________
>>> TLS mailing list
>>> TLS@ietf.org
>>> https://www.ietf.org/mailman/listinfo/tls
>>> 
>>> 
>>> 
>>> --
>>> Regards
>>> 
>>> Andy
>> 
>> 
>> 
>> 
>> -- 
>> Regards
>> 
>> Andy
>