The MPLS-OPS Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] RE: TCP state machine
To add to Ken's response: TCP state machine specifies that the only valid transitions out of "SYN_SENT" state are when you receive a SYN, an ACK, or when you timeout. Any other event or message should have no effect on this state. So, the fact that you received a FIN when you are in SYN_SENT state has no relevance. It has no meaning to this state. Hope this helps. Have a nice day, Santosh Kumar Ohio State University --- "Hull, Kenneth A." <KHULL@alleghenyenergy.com> wrote: > I would suggest you build a timer to take either end > of the session back to > the previous state if the an appropriate response is > not received in the > timer window, and be sure your software ignores any > extra messages > corresponding to move form the previous state it may > receive for the same > session (same source TCP port) once it moves to the > next state. > > -----Original Message----- > From: Mohammed [mailto:ipswitching@yahoo.com] > Sent: Wednesday, February 13, 2002 11:08 AM > To: mpls-ops@mplsrc.com > Cc: ahussain@eng.utoledo.edu > Subject: TCP state machine > > > Hi all, > > I was writing the code for TCP state machine so > that > i could use it for running LDP over TCP. This would > just be a component of the simulator that i am > working > on. > > I was looking at the state transition diagram given > in TCP/IP illustrated Vol 1 by Richard Stevens. > > Please consider the following scenario -- > > A TCP software on a peer working on passive open > mode > is in "SYN-RCVD" state. It is waiting for an 'ACK' > from its peer on the other end of the connection. > There are two things that might happen here ... > > (1) SYN + ACK sent before transition to this state > was lost, so the other side never received it and > hence never ACK. > > (2) The other side have received it and had ACk but > the ACk was lost. > > Anyways 'SYN-RCVD' i waiting for an ACK. But during > its wait period or even before its time oust and > sends > RST ... if it receives an appl:close interrupt then > it > is supposed to send a FIN to its peer. > > Now if the peer on the other side is in > > -- "ESTB" state then this would be okay. > > But if it is "SYN SENT" state .. then how am i > supposed to handle this FIN. > > Any pointers from TCP developers or any other > interested professionals are welcome. > > Thanks for your time and help!!! > Abrar. > > > ===== > Abrar Hussain. > Graduate Student/Network Designer. > The University of Toledo. > Toledo,OH. > > __________________________________________________ > Do You Yahoo!? > Send FREE Valentine eCards with Yahoo! Greetings! > http://greetings.yahoo.com > > ------- > The MPLS-OPS Mailing List > Subscribe/Unsubscribe: > http://www.mplsrc.com/mplsops.shtml > Archive: > http://www.mplsrc.com/mpls-ops_archive.shtml > > ------- > The MPLS-OPS Mailing List > Subscribe/Unsubscribe: > http://www.mplsrc.com/mplsops.shtml > Archive: http://www.mplsrc.com/mpls-ops_archive.shtml __________________________________________________ Do You Yahoo!? Yahoo! Sports - Coverage of the 2002 Olympic Games http://sports.yahoo.com ------- The MPLS-OPS Mailing List Subscribe/Unsubscribe: http://www.mplsrc.com/mplsops.shtml Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
|
|