[18821] in bugtraq
Re: ICMP fragmentation required but DF set problems.
daemon@ATHENA.MIT.EDU (Felix von Leitner)
Thu Jan 25 16:05:06 2001
Mail-Followup-To: Felix von Leitner <leitner@vim.org>, BUGTRAQ@SECURITYFOCUS.COM
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Message-Id: <20010125195109.A6368@vim.org>
Date: Thu, 25 Jan 2001 19:51:09 +0100
Reply-To: Felix von Leitner <leitner@VIM.ORG>
From: Felix von Leitner <leitner@VIM.ORG>
To: BUGTRAQ@SECURITYFOCUS.COM
In-Reply-To: <200101232352.f0NNqRF96940@drugs.dv.isc.org>; from
Mark.Andrews@NOMINUM.COM on Wed, Jan 24, 2001 at 10:52:27AM +1100
Thus spake Mark.Andrews@NOMINUM.COM (Mark.Andrews@NOMINUM.COM):
> > IPv6 is another case though. Here you have mandatory PMTU for all
> > protocols.
> This is a myth. It is quite possible to have a IPv6
> implementation that does not require PMTU discovery. You
> just fragment all your traffic at the network MTU. If fact
> there are hooks in the BSD advansed API to allow you to do
> this on a per socket basis.
That is a myth, too.
If chose not to to PMTU in IPv6, you have to limit your MTU to ~1200
byts (the exact value is in the RFC). So it's not the network MTU you
are using.
> DNS servers, for one, need to be able to disable PMTU on
> the replies.
?
Felix
