From mboxrd@z Thu Jan 1 00:00:00 1970 From: rigoberto perez Subject: Problem with TCP connections Date: Wed, 11 Jun 2008 13:49:41 -0700 (PDT) Message-ID: <917672.36526.qm@web33207.mail.mud.yahoo.com> Reply-To: rigo666beast@yahoo.com Mime-Version: 1.0 Return-path: Sender: netfilter-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: netfilter In a topology like: windows1 --- linux router1 --- linux router2 --- windows2 I have a netfilter module in every linux router with 3 hooks: localin localout forward and the module ipconntrack is loaded. In forward hook i change the outcoming packets in this way: padd 1 byte at the end of a packet save the protocol of the ip header in the padded byte modify the protocol in ip header by 250 When the forward hook receives an ip packet with the protocol 250 I do the inverse function: restore the original protocol trim the last byte of the packet This works OK for ICMP packets - even large packets to force fragmentation, but, When I use TCP the connections, for example FTP, - with packets of 1500 bytes - stops in the middle of the transfer. If I change the MSS in the server to 1499 everything works OK but I dont want to do this, I wish resolve the problem in the router. I need some help, thanks.