Subject: Re: perhaps time to check our TCP against spec?
To: Kevin M. Lahey <firstname.lastname@example.org>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
Date: 04/06/1998 21:05:53
>>sizeof(struct tcpiphdr). "Where's the problem?"
>The key issue here is the difference between the advertised MSS
>and the largest packet to be sent.
Yup. Now, try this one on for size:
--FDDI ring A---
-------------------------- Ethernet ---- router R
--FDDI ring B---
Again, assume no PMTU, and that the two FDDI rings are disjoint.
With in_maxmtu, and without PMTU, as far as I can tell, A and B
negotiate FDDD-sized MSS. That means traffic from B to A ends up
getting fragmented at the router R.
Doing this the traditional way, A advertises an Ethernet-sized MSS, so
TCP connections between A and B have A's MTU as an upper bound on MSS.
Call me oldfashioned, but I tend to see this as a bug.
I do understand how people using HIPPI and giving the `other'
interface address might want this to behave differently. I'd suggest
maybe those people should turn on PMTU, and not break things for the
rest of us with 100Mbit or Ethernet NetBSD hosts talking to old FDDI
Sure, this specific scenario may be farfetched, but I have colleagues
here at Stanford who regularly do something very similar with their
mobile-IP setup. _That_ is not far-fetched at all.