Subject: Gateway Problems
To: None <netbsd-users@NetBSD.ORG>
From: src <steve@scurvy.spc.uchicago.edu>
List: amiga
Date: 09/13/1995 22:25:09
	I trying to set up an Amiga 3000 running 1.0 as a gateway to a
subnet, but having extreme difficulty.  I'm not sure if the problem is 
with the way I am configuring things, or a problem with netbsd (or the routed
that comes with netbsd).

	The gateway has an ethernet interface (128.135.252.227) on the 
subnet (128.135.252.224) with a netmask of 255.255.255.240.  When needed,
I also add the SLIP link to the rest of the world, interface sl0, IP 
128.135.252.225, dest 128.135.252.10, netmask 255.255.255.0.

	The SLIP link works fine, but noone else inside the ethernet subnet 
can reach either side of the link (pinging 128.135.252.225 from another
host on the ethernet is futile).  What's really odd is that pinging
128.135.252.225 from the gateway results in "host is down"!!  I can't
even reach my own interface?  I'm starting to think that there may be 
something I don't know about PtoP links or something wrong with the
system (GATEWAY is configured in the kernel).

	Another problem I face is with routed.  I figured that I should run
an active router to let people on the rest of the net reach the machines
inside my subnet...  But, routed will not monitor for new interfaces unless
the interface is configured down when routed is started (at least that's
what I gather from a quick look through the sources).  So, I tried to 
set up /etc/hostname.sl0, with the option "down" included.  Well, all of
a sudden I start getting "arpresolv: can't allocate llinfo" on the tty, just
as soon as routed starts - and 'ifconfig sl0' reveals that the interface is 
marked as up!
	So I removed hostname.sl0, and rebooted, but I was still getting
the error from arpresolv until I killed off routed.

	Now I'm back where I started, except that I can't run routed until
I start the slip connection (i'm guessing it's related to the fact that 
/etc/gateways now includes my SLIP server, why that keeps it from allocating a 
llinfo is a good question, though).

	What's really odd about all this is netstat -r:

root@herb:/home/steve [253/8]$ netstat -r
Routing tables

Internet:
Destination      Gateway            Flags     Refs     Use  Interface
default          oedipus.spc.uchica UG          2       91  sl0
localhost        localhost          UH          1       40  lo0
128.135.252      oedipus.spc.uchica UG          0        0  sl0
oedipus.spc.uchi kuru-gw.spc.uchica UH          4      221  sl0
kuru-net.spc.uch link#1             UC          0        0  le0
kuru-gw.spc.uchi link#1             UHRL        0       10  le0
lupus            0:80:10:0:15:32    UHL         2      298  le0
herb             localhost          UGHS        1       70  lo0
128.135.252.239  link#1             UHL         1       40  le0

oedipus is the slip server, kuru-net is the subnet, lupus is a machine on the 
subnet, and kuru-gw is my slip interface.

	I'm really not sure what all the "link#1"'s are, but I know
that kuru-gw is _not_ on le0 - it is sl0.

	Does anyone have any idea what I am doing wrong, or what is wrong?
I'd really appreciate any help or even just clues.

-- 
//  Stephen Champion                   Freelance Programmer ;^)
//  schampion@hampshire.edu            Chicago, IL  USA
//  steve@kuru.com                     Amiga 3000/25    NetBSD 1.0
#define Disclaimer "This product is completely without warranty."