Date: Fri, 12 Aug 2005 17:14:59 -0400 From: Chuck Swiger <cswiger@mac.com> To: Miku Jha <jha_miku@yahoo.com> Cc: jha miku <jha0147@yahoo.com>, freebsd-net@freebsd.org Subject: Re: Question regd timestamp option Message-ID: <42FD1153.50202@mac.com> In-Reply-To: <20050812205736.26235.qmail@web50201.mail.yahoo.com> References: <20050812205736.26235.qmail@web50201.mail.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Miku Jha wrote: [ ... ] > The situation is that if the client crashes, the server eventually sends a > RST (10.39.53) Following this RST, the client comes back in lets say around > 2-3 minutes. Now when the client sends a SYN(10.42.23), there is no > timestamp option. If the client opens a connection and both sides exchange packets with timestamps, you'll probably end up seeing "NNT" in all packets during the first session. This right? Now if you open a second connection, while things are still OK, do you see the SYN packet contain all options as normal? I assume the client is opening connections to the server? And it is a FreeBSD box...? Showing tcpdump data (or putting on a website somewhere) would help understand the issue... > Is there some requirement that RST needs to be ACKED > or RST flag will remain set for some time window > within which if SYN is send, timestamp option will not > be set. A RST to a closed or listening socket will be ignored (dropped), a RST which matches an established connection will flush and close that connection but will not be ACKed itself. -- -Chuck
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42FD1153.50202>