Message-ID: <163dd35d-fee8-4e7f-a6b3-c34bd2c18cb3@...1227...> Date: Thu, 26 Jun 2025 00:43:41 +0200 MIME-Version: 1.0 References: <1e8ccfd451cbf9e6d090bd6a6886e366.squirrel@...2890...> <20250507074438.1326755-1-walter.openvpn@...2889...> <8ba925a1-85c9-4274-a512-32dcec35b147@...1227...> <2bb3487e-6762-40ea-b6ad-2622a879ca21@...2889...> <866e86f674f55d0d39cce796f3f7ed9b.squirrel@...2890...> Content-Language: en-GB From: Arne Schwabe In-Reply-To: <866e86f674f55d0d39cce796f3f7ed9b.squirrel@...2890...> Subject: Re: [Openvpn-devel] [PATCH] multi.c: Allow floating to a new IP right after connection setup List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit To: Walter Doekes Cc: openvpn-devel@lists.sourceforge.net Am 25.06.25 um 15:30 schrieb Walter Doekes: > Good. I backported the patch so it ran against the culprit version > (b3647114). > > I got these mesages: > > > SENT CONTROL [mycommonname]: 'PUSH_REPLY,route ... 255.255.255.255 > net_gateway,route-gateway 10.x.x.1,topology subnet,ping 15,ping-restart > 55,route 10.x.x.0 255.255.0.0 vpn_gateway,ifconfig 10.x.x.3 > 255.255.255.0,peer-id 4,cipher AES-256-GCM' (status=1) > > Packet with invalid or missing SID from [AF_INET]HOME_IP:33567 > > Float requested for peer 4 to HOME_IP:33567 > > peer 4 (mycommonname) floated from VPN_IP:33567 to [AF_INET]HOME_IP:33567 > > > > The "Packet with invalid or missing SID" is new to me. But other than > that, it works. Thanks for testing and confirming that it works. > I also tried it against 2.6-latest (0169b4ad). Also works. There the > message is: > > Packet (P_ACK_V1) with invalid or missing SID from [AF_INET]HOME_IP:46088 > > I can't tell if this new message is problematic or not. It doesn't > negatively impact my connection setup. And I (now) know when to expect it. > > > > As for your patch: there's a minor typo in your patch at ssl_pkt.h in the > signature: > > "bool check_session_id_hmac" should be "bool pkt_is_ack" Thanks will fix in the next revision. > Further, I would prefer if the commit message itself mentioned something > about "floating IPs and 60 second timeout after connect" instead of "rare > circumstances" which are not rare in 100% of my use cases. That might be > beneficial to the next person who runs into this. You need a connection that - starts on IP A - successfully floats to IP B by data packet - then has a control packet from IP A before any data packet can trigger the float back to IP A In this scenario we would trigger a new connection coming before while know we detect that this should not trigger a new connection as it is not a new connection attempt. So instead of creating a new connection, you see (at verb 4 or higher) the message Packet (P_ACK_V1) with invalid or missing SID instead with the patch. I will update the commit message when the patch is no longer in draft mode. I wanted to confirm that this is actually the problem/scenario we are fixing before finishing the patch. Arne