Date: Mon, 15 Jun 1998 08:37:21 +0200 (CEST) From: hans@brandinnovators.com (Hans Zuidam) To: jdp@polstra.com (John Polstra) Cc: hackers@FreeBSD.ORG Subject: Re: TCP sequence numbers Message-ID: <199806150637.IAA11326@truk.brandinnovators.com> In-Reply-To: <199806121439.HAA11334@austin.polstra.com> from John Polstra at "Jun 12, 98 07:39:28 am"
next in thread | previous in thread | raw e-mail | index | archive | help
> John Polstra wrote: > In article <199806120831.KAA00529@truk.brandinnovators.com>, > Hans Zuidam <hans@brandinnovators.com> wrote: > > For a particularly nasty TCP problem I have with an embedded target > > I would like to have the ability to always use the same initial > > sequence number on a particular connection. It would make it easier > > to compare tcpdump traces. > Are you sure? Tcpdump prints TCP sequence numbers in relative form, > relative to the initial one, unless you give it the "-S" option. Your > traces should match already, except for the initial packet from each > side. But the other side cannot "tcpdump", it's an embedded system from which I can only get absolute sequence numbers ;-( Also, trpt(8) dumps it's sequence numbers in abosolute (hex) numbers, that's easy to change. It's probably easier to have TCP use the same initial sequence number instead of modifying other programs. To add predictable sequence numbers I am thinking to add a TCP option for setsockopt(3) to set an initial sequence number. Maybe a sysctl could be added to turn the feature globally on or off. Regards, Hans -- H. Zuidam E-Mail: hans@brandinnovators.com Brand Innovators B.V. P-Mail: P.O. Box 1377 de Pinckart 54 5602 BJ Eindhoven, The Netherlands 5674 CC Nuenen Tel. +31 40 2631134, Fax. +31 40 2831138 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806150637.IAA11326>