[18262] in North American Network Operators' Group
Re: MTU problems with GRE tunnels (fwd)
daemon@ATHENA.MIT.EDU (Dennis Ferguson)
Tue Jul 7 15:37:29 1998
To: schweikh@noc.dfn.de (Jens Schweikhardt)
cc: bridge@ip-plus.net, horke@regio.net, nanog@merit.edu,
noc@noc.dfn.de (DFN NOC)
In-reply-to: Your message of "06 Jul 1998 11:39:04 GMT."
<199807061139.NAA20048@obsidian.noc.dfn.de>
Date: Tue, 07 Jul 1998 12:25:09 -0700
From: Dennis Ferguson <dennis@juniper.net>
> Here's something very strange I observe with GRE tunnels (the default
> tunnel mode). It looks like cisco routers send IP datagrams violating RFC 791
> [Internet Protocol] over GRE tunnels. In particular, the length field of
> the IP header is computed incorrectly to *not* include the size of the
> IP header. RFC 791 says about the length field:
[...]
> I have an application on my workstation that serves as one endpoint
> of a GRE tunnel. In fact, it's such a tiny perl program that I have
> appended it at the end of this mail.
[...]
> We note that the length as reported in the IP header is
> always 20 octets less than what we receive on the socket.
> This leads me to the question
>
> Do you cisco guys read RFCs? :-)
I can tell you for sure that the Cisco routers do send the packets (GRE
or IP protocol 4) with a length which includes the IP header, just like
the RFC. If you look I think you'll find that it is your kernel which is
subtracting out the IP header length before it hands the packet to you on
the raw socket.
Dennis Ferguson
