Re: [tcpm] SYN in SYN_RCVD state

On Mon, Oct 22, 2007 at 12:37:23PM -0700, Vishal Study wrote:
> Hi Ted:
> 
> The server would retransmit SYN+ACK when server times out in SYN-RECEIVED state.
> I agree and have seen this.
> 
> But, when the server hasn't timed out in SYN-RECEIVED state and it
> receives retransmitted SYN from client (due to client's timeout), what
> is the expected behavior on server side? Is server supposed to send
> SYN+ACK?

The spec says no.   RFC793 p 69 gives the sequence number test and the
ACK requirement:

----
   first check sequence number

      SYN-RECEIVED STATE
      ESTABLISHED STATE
      FIN-WAIT-1 STATE
      FIN-WAIT-2 STATE
      CLOSE-WAIT STATE
      CLOSING STATE
      LAST-ACK STATE
      TIME-WAIT STATE

        Segments are processed in sequence.  Initial tests on arrival
        are used to discard old duplicates, but further processing is
        done in SEG.SEQ order.  If a segment's contents straddle the
        boundary between old and new, only the new parts should be
        processed.

[table and some text snipped.  None of it applies; the segment is
outside the window and unacceptable. ]

        If an incoming segment is not acceptable, an acknowledgment
        should be sent in reply (unless the RST bit is set, if so drop
        the segment and return):

          <SEQ=SND.NXT><ACK=RCV.NXT><CTL=ACK>

        After sending the acknowledgment, drop the unacceptable segment
        and return.
----

I can imagine an implementor sending a SYN+ACK here, but it would be
outside the spec.  I assume you have no basis for sending one.

-- 
Ted Faber
http://www.isi.edu/~faber           PGP: http://www.isi.edu/~faber/pubkeys.asc
Unexpected attachment on this mail? See http://www.isi.edu/~faber/FAQ.html#SIG