From: Mikko Vainikainen Date: Fri, 8 Mar 2013 14:41:06 +0000 Message-ID: References: , <1CED409804E2164C8104F9E623B08B9013B8D6E33F@...1436...> In-Reply-To: <1CED409804E2164C8104F9E623B08B9013B8D6E33F@...1436...> Content-Language: en-US MIME-Version: 1.0 Subject: Re: [Openvpn-devel] [PATCH] Fix for bug #49 for openvpn 2.2.2 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable To: Joachim Schipper , "openvpn-devel@lists.sourceforge.net" Hi, it is absolutely true that it is necessary to do HMAC hash per every existi= ng session, as far as I know there's no other way to tell reliably if the p= acket really belongs to an existing session... I haven't measured the HMAC = performance myself, but I'd guess it would take a bit more than a few hundr= ed packets per second to exhaust a modern CPU. The for-loop part was originally (shamelessly) adapted from tls_pre_decrypt= () @ ssl.c, it could have been more clear I have to admit... I think the memcmp() is already constant-time as the HMAC hash has fixed si= ze per hash algorithm (as far as I can remember, please correct me if I'm w= rong). This was also shamelessly adopted from openvpn_decrypt() @ crypto.c.= =20 I don't think there are any timing related attack opportunities as acceptin= g the incoming packet does not trigger any immediate replies being sent bac= k to anyone; the accepted packet is just processed as if it was received fr= om the original address. As the data channel packet is being forwarded away= from the original sender to the trusted side of the tunnel it should be qu= ite hard for an attacker to make any timing calculations.. I wasn't aware the --float option could be specified per client... Best regards, and nice weekend! Mikko ________________________________________ From: Joachim Schipper [joachim.schipper@...1435...] Sent: Friday, March 08, 2013 2:58 PM To: openvpn-devel@lists.sourceforge.net Cc: Mikko Vainikainen; Steffan Karger Subject: RE: [PATCH] Fix for bug #49 for openvpn 2.2.2 > Hi, > > our setup needs openvpn UDP/TLS tunnels with dynamic client IP addresses,= so I implemented a fix for the bug #49 that has been open for over two yea= rs. > > The patch is for version 2.2.2 as I had trouble compiling the 2.3.x tarba= ll from openvpn.net. As the solution is rather simple (just two small utili= ty functions in mudp.c) I'd guess it could be comfortambly migrated to 2.3.= x. > > Basically what the fix does is the following: incoming data channel UDP p= ackets from an unknown IP are matched against existing UDP/TLS sessions, an= d if the packet passes the HMAC authentication against an existing TLS cont= ext we know the client IP has changed and the session state will be instant= ly updated accordingly. > > I have tested this fix to some extent, and the IP handover works impressi= vely smoothly in my test setup where I randomly switch between two routes f= rom client to server. > > Dynamic client IP's are enabled/disabled with --float in the server side. > > Please feel free to contact me for any questions etc. This is not a full review of the patch, just a few quick remarks. This patch seems to do (number of tunnels) HMAC's for any packet received f= rom an unknown IP. If this is correct, couldn't a server with a couple thou= sand tunnels be brought just by sending it a few hundred packets a second? authenticate_tls_packet() contains a for() loop, but I see no code path tha= t actually loops - everything seems to return immediately? memcmp() should be replaced by a constant-time function to prevent timing a= ttacks (probably only realistic on a server with a single tunnel); even the= n, this patch allows determining the (approximate) number of tunnels in use= by looking at processing time. I haven't looked at how this interacts with other features; it's at least n= oteworthy that the floating behavior can't be specified on a per-connection= basis. Joachim