implementation detail

Gary Scott Malkin <gmalkin@xylogics.com> Fri, 12 February 1993 16:14 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa05928; 12 Feb 93 11:14 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa05924; 12 Feb 93 11:14 EST
Received: from atlas.xylogics.com by CNRI.Reston.VA.US id aa05783; 12 Feb 93 11:14 EST
Received: by atlas.xylogics.com id AA14739 (5.65c/UK-2.1-930202); Fri, 12 Feb 1993 11:12:30 -0500
Received: by atlas.xylogics.com id AA32551 (5.65c/UK-2.1-930202); Fri, 12 Feb 1993 11:12:28 -0500
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Gary Scott Malkin <gmalkin@xylogics.com>
Date: Fri, 12 Feb 1993 11:12:28 -0500
Message-Id: <32551.199302121612@atlas.xylogics.com>
To: ietf-rip@xylogics.com
Subject: implementation detail

Here's something I'd like to share with those who are planning
to implement RIP-2.  It's something that bit me in the butt and
I want to warn you about it.

Apparently, there are implementations of routed which equate a
pointer to a RIP entry with a pointer to a sockaddr.  That is,
when adding a route to the routing table, routed assumes that
the first 16 bytes of a RIP entry are a sockaddr with a family
and an address and all other fields zero.  However, a RIP-2
packet will have stuff in those zero fields.  Therefore, it is
important to mask out those fields to create a real sockaddr.

----------------------------------------------------------------------
Gary Malkin            "Felis Catus, is your taxonomic nomenclature an
(617) 272-8140         `exothermic quadruped, carnivorous by nature'?"