[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAFXGftLY1ywdRj79RMHW57oLa9_HNEEcUYRAiM=zs6YHHGVFzw@mail.gmail.com>
Date: Tue, 10 Jan 2017 00:31:01 +0800
From: Sun Paul <paulrbk@...il.com>
To: Neil Horman <nhorman@...driver.com>
Cc: David Laight <David.Laight@...lab.com>,
"linux-sctp@...r.kernel.org" <linux-sctp@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: Problem on SCTP
what kind of information do you need? the whole INIT packet?
On Mon, Jan 9, 2017 at 8:25 PM, Neil Horman <nhorman@...driver.com> wrote:
> On Mon, Jan 09, 2017 at 06:00:36PM +0800, Sun Paul wrote:
>> Hi
>>
>> the linux router just change the destination, so it can arrive on the
>> the SERVER.
>>
> Please post the relevant snippets from the client and server tcpdump
> operations
>
> Neil
>
>> On Mon, Jan 9, 2017 at 5:51 PM, David Laight <David.Laight@...lab.com> wrote:
>> > From: Sun Paul
>> >> Sent: 09 January 2017 02:08
>> >
>> >> >> I am setting up a lab where the SCTP traffics from client is passing
>> >> >> through a linux router before reaching to the SCTP server running
>> >> >> LKSCTP.
>> >> >>
>> >> >> The linux router did not change the source address of the client, so
>> >> >> when it arrived to the SCTP server, the source address is the oriingal
>> >> >> one.
>> >>
>> >> the INIT chunk arrive on the SERVER, but then no response. the
>> >> application that using in SERVER is the same as the other test.
>> >>
>> >> I noticed one thing in Ethernet frame of the incoming packet on the
>> >> SERVER compared to the one captured from the client is the LG bit on
>> >> the source address.
>> >>
>> >> The LG bit is set to 0 on the request packet received in the
>> >> SERVER,but it is 0 from the one originated on the client. willl it be
>> >> the root cause?
>> >
>> > Which addresses are you talking about, and what do you mean by the LG bit?
>> >
>> > Is your linux 'router' just routing (ie IP forwarding) or is it doing NAT?
>> > If it is changing the IP addresses then the addresses inside some SCTP
>> > chunks also need changing.
>> >
>> > David
>> >
>>
Powered by blists - more mailing lists