From owner-freebsd-questions  Mon May 31  7:14:30 1999
Delivered-To: freebsd-questions@freebsd.org
Received: from gate.volvo.se (gate.volvo.se [192.138.110.253])
	by hub.freebsd.org (Postfix) with ESMTP id 40E0114E50
	for <freebsd-questions@freebsd.org>; Mon, 31 May 1999 07:14:27 -0700 (PDT)
	(envelope-from john@nike.volvo.se)
Received: by gate.volvo.se; id QAA05386; Mon, 31 May 1999 16:14:04 +0200 (MET DST)
Received: from unknown(131.97.124.70) by gate.volvo.se via smap (4.1)
	id xma004883; Mon, 31 May 99 16:13:38 +0200
Received: by nike.volvo.se with  (8.6.12/1.37)
	id QAA20593; Mon, 31 May 1999 16:13:38 +0200
Date: Mon, 31 May 1999 16:13:37 +0200 (MET DST)
From: John K <john@volvo.se>
To: slava <sl@zeus.dnt.md>
Cc: freebsd-questions@freebsd.org
Subject: Re: [off-topic] DF bit and IP 
In-Reply-To: <Pine.BSF.4.05.9905311019300.61220-100000@zeus.dnt.md>
Message-ID: <Pine.BSD/.3.91.990531160115.20110B-100000@nike.volvo.se>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Content-Transfer-Encoding: QUOTED-PRINTABLE
Sender: owner-freebsd-questions@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG


Hi,=20
A dont fragment bit should be set only when the application "at the other=
=20
end" isn=B4t able to puzzle the packets together, so AFAIK , setting DF=20
bits is an application specific matter. Also, you shouldn=B4t write a=20
program that cannot=20
defragment packets, as it behaves wired on routed network with differnet=20
MTU=B4s (eg, T/R-> Eth-> T/R )

I dont think TCP=B4s will be resent if there isn=B4t even an ICMP answer.

When a packet is to be sent out on a router interface with a smaller MTU
than the inbound interface, it is dropped by default and an ICMP packet is=
=20
sent back. (like in the example above, if a DF bit would have been set.)

hope this helped some,=20

br
John

On Mon, 31 May 1999, slava wrote:

>=20
> Hi
>=20
> Does TCP always gets encapsulated in an IP header with DF bit set?
> I know this is needed for path MTU discovery to make tcp more efficient
> but is this implemented on all OSes?
>=20
> What if something in the way is blocking the icmp packet-too-big type
> to the initiator of a TCP connection and it never finds out about a small
> MTU in the path? Will it retry with a smaller MTU itself?
>=20
> thanks
> slava
>=20


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message