Subject: Pppd under -current panics remote system!
To: NetBSD/pmax Discussion List <port-pmax@netbsd.org>
From: Paul Mather <paul@gromit.dlib.vt.edu>
List: port-pmax
Date: 09/23/1999 09:22:00
Yesterday, I finished doing a "make build" and making a new kernel for
my DECstation 3100.  My remote PPP server is a DECstation 5000/240
running NetBSD 1.4 release, with a 1.4.1 release kernel.

When I did a reboot this morning to try the new kernel, the first thing
I noticed is that pppd would not start.  It complained about not being
able to authenticate the remote link (I used "demand" in my peer file).
Adding "noauth" to the file for my peer at least allowed pppd to run.
(My remote PPP server uses neither PAP nor CHAP---just plain login-style
authentication.)

However, it still wouldn't connect.  Mysteriously, there were long
delays in retries, too.  I discovered the reason for that was that my
3100 was causing a panic on the remote machine, and causing it to reboot
every time!

Here is a snippet of /var/log/messages from the remote machine:

>>>>>
Sep 23 08:39:11 wendolene pppd[11418]: pppd 2.3.5 started by ppp, uid 1126
Sep 23 08:39:11 wendolene pppd[11418]: Using interface ppp0
Sep 23 08:39:11 wendolene pppd[11418]: Connect: ppp0 <--> /dev/ttyp1
Sep 23 08:39:12 wendolene pppd[11418]: Could not determine remote IP address
Sep 23 08:39:13 wendolene pppd[11418]: Connection terminated.
Sep 23 08:39:13 wendolene pppd[11418]: Exit.
Sep 23 08:42:31 wendolene syslogd: restart
Sep 23 08:42:32 wendolene /netbsd: panic: rtfree 2
Sep 23 08:42:32 wendolene /netbsd: syncing disks... 23 23 21 17 8 done
<<<<<

Here is the /etc/ppp/options file from the PPP server:

>>>>>
proxyarp
nodetach
lcp-echo-interval 30
lcp-echo-failure 6
local
asyncmap 200a0000
escape 7e,7f,fe,ff,93
<<<<<

As of last night, the remote system could figure out both the local and
remote IP addresses just fine, so I presume some incompatibility has
been introduced on the 3100's end.

Is the "solution" to upgrade the 5000/240 to -current (which I'm loathe
to do right now because of a pressing lack of time), or is there
something I can do at the 3100 end to make it behave nicely and stop it
rebooting the remote system? ;-)

A quick fix would be most appreciated, as PPP is the only way I have to
get data on or off the 3100. :-(

Cheers,

Paul.

e-mail: paul@gromit.dlib.vt.edu

"I don't live today; maybe tomorrow..."
	--- James Marshall Hendrix