Subject: Oddball network misbehavior
To: None <>
From: John F. Woods <>
List: current-users
Date: 12/09/1994 19:34:12
It's possible, even probable, that NetBSD itself isn't to blame for this, but
perhaps someone will instantly see the problem in any case.

I've just gotten an Internet link for my home system, a PPP link through
UltraNet Communications.  I've been happily grabbing stuff from systems all
over the Internet -- but I can't reach MIT.  If I try to ping systems at MIT,
nothing happens (I don't even see the modem lights flash, which is interesting,
but possibly meaningless) (note that I am able to correctly look up their
IP addresses).  If I telnet out to some site I can reach, and telnet from there
to a machine at MIT, and then ping my home system, it also fails to work, BUT:
I do, in that case, seem to see regular modem activity!  (The MIT site is able
to ping the gateway on the other end of the PPP link.)

netstat -pip seemed to indicate that there were a lot of packets being
discarded as being for "no protocol"; I put a printf in raw_ip.c and found
that there were a lot of packets with destination address of {my IP address},
a source address of (!), and a protocol of 1 (ICMP).  netstat -picmp
also seemed to indicate that it was generating lots of ICMP error packets,
I *think* mostly "time exceeded" packets.

Is NetBSD just being excessively fussy about packet ages?  Is MIT unwisely
generating packets with excessively small TTL values?  Does anyone have a clue
as to what I should look for next?

Oh, I'm currently running stock NetBSD1.0, but will be upgrading to current
this weekend (I'll even be able to run sup at last :-).

John Woods,