[3482] in linux-net channel archive
Re: Speaking of MTU problems
daemon@ATHENA.MIT.EDU (Alan Cox)
Tue Jun 25 06:41:10 1996
From: Alan Cox <alan@cymru.net>
To: dennis@etinc.com (Dennis)
Date: Tue, 25 Jun 1996 10:15:51 +0100 (BST)
Cc: inr-linux-net@ms2.inr.ac.ru, linux-net@vger.rutgers.edu
In-Reply-To: <199606242143.RAA17425@etinc.com> from "Dennis" at Jun 24, 96 05:43:48 pm
> for example has only a 2 byte header, but IETF requires an additional
> 2. So with a 1500 byte mtu does that mean that only 1498 bytes of IP
> data can be sent with IETF encapsulation in force? The encapsulation,
> although a header, is really data for the protocol decoder, not the
> physical medium.
The interface MTU listed and used by all the config structures is precisely
the same as everywhere else for IP, that is its the largest IP packet
including IP headers you could send fragmented. It includes no device
headers/padding/preamble/etc of any kind.
Alan
