Path: utzoo!attcan!uunet!mcvax!ukc!eagle!icdoc!qmc-cs!liam
From: liam@cs.qmc.ac.uk (William Roberts)
Newsgroups: comp.protocols.appletalk
Subject: Re: EtherTalk vs. AppleTalk vs. ???
Summary: DYNIX has something that doesn't really work
Keywords: ethertalk, UNIX, dynix
Message-ID: <605@sequent.cs.qmc.ac.uk>
Date: 31 Aug 88 19:45:01 GMT
References: <1364@cayman.COM> <1367@cayman.COM> <927@cunixc.columbia.edu>
Reply-To: liam@cs.qmc.ac.uk (William Roberts)
Organization: Computer Science Dept, Queen Mary College, University of London, UK.
Lines: 31
Expires:
Sender:
Followup-To:
Distribution:

The DYNIX 3.0 operating system for the Sequent machines (a version of BSD)
has some limited support for EtherTalk: according to the
documentation you can say things like

        s = socket(AF_APPLETALK, SOCK_DGRAM, 0);

and it has descriptions of all the various LAP headers, DDP
headers etc.  The physical addresses of EtherTalk Macs are
acquired by arp and stored in the relevant arp tables (confuses
/etc/arp though).

The bad news is it doesn't really work. In particular it
doesn't understand longDDP headers and so can only work with
short ones, no software is provided for setting the AppleTalk
address of the DYNIX machine etc. It is just a peg for some
third party software from Kinetics - ironic that all our
ETherPort II Mac IIs have Kinetics software that only ever
sends longDDP packets :-)

The task of receiving EtherTalk packets on your UNIX box isn't
too evil: the Sun Network Interface Tap (NIT) code would allow
you to do it without kernel hacking, even.  If this message
doesn't help shame Sequent into fixing their implementation (I
have offered...) then watch this space for code: I need to
prevent our FastPath box from being a bottleneck and single point
of failure.
-- 

William Roberts         ARPA: liam@cs.qmc.ac.uk  (gw: cs.ucl.edu)
Queen Mary College      UUCP: liam@qmc-cs.UUCP
LONDON, UK              Tel:  01-975 5250