Date: Tue, 2 Jan 96 20:22:54 +1000 From: Andrew MacIntyre <andymac@bullseye.apana.org.au> To: joerg_wunsch@uriah.heep.sax.de Cc: bugs@freebsd.org Subject: Re: 2.1.0 PPP hassles... Message-ID: <30e9077f.bullseye@bullseye.apana.org.au> In-Reply-To: <199512310852.JAA16764@uriah.heep.sax.de> from "J Wunsch" at Dec 31 95 9:52 am
next in thread | previous in thread | raw e-mail | index | archive | help
Thanks for your reply.
> > 1. The dial chat script behaviour appears to have changed, seemingly
> > requiring a terminating send string, whereas all the samples and the
> > working config from the above mentioned SNAP-950726 system don't
> > have/need it. I've used the string "\\c" to fill this for now.
>
> Hmm, i'm not claiming that this hasn't changed, but i) i haven't
> noticed any change myself, and ii) the only diff between 2.0.5 and
> 2.1's chat.c is:
{...}
This is a puzzler. If I don't have that "null" send string, it succeeds
in hanging up the modem - as near as I can figure by sending \n before
the text following the CONNECT is completely received.
> > 2. Using the dial command, after the connect has been made, but before the
> > login has been executed, PPP core dumps with an "illegal instruction" error.
> > Upon restarting PPP, with the modem link still up, I can use "term" and
> > manually log in.
>
> That is very strange. The only know occasions for ``illegal
> instruction'' traps by now are broken CPUs. No normal code should
> ever be able to cause illegal instructions, since the compiler and
> assembler do not generate them.
>
> > 3. Using a manual login, after PPP reports "Packet mode.", it then
> > reports "SIOCAIFADDR: File exists" netstat -r shows that the link is
> > in place, but as far as the rest of the systems is concerned, the link is
> > down.
>
> This means that some of the routes PPP was going to install did still
> exist, perhaps from a previous crashed invocation of PPP. (The ``File
> exists'' must be translated into ``Route exists'' as far as the
> routing code is concerned.)
CPU is an AMD - not quite sure of vintage.
I went back to scratch with the network configuration, and have got
past #3, and even got the dial to work - only abt 3-5 mins later PPP
crashed with a signal 4 (illegal inst) while I was ftping a file :-(
This time, the tunnel driver was left open, so a reboot was required to
try again...
> Something is very weird with your machine, but since very many people
> are using iijppp, and nobody has ever reported anything resembling
> your description, i would first suspect hardware.
I think you're right to suspect hardware. I vaguely recall a few
years ago there being known problems with UMC chipsets and Unix systems
(in particular Linux). I have a replacement SIS471 chipset motherboard
(my SNAP-950726 system is running very nicely on a slightly different
SIS471 mb) for this system which I've held off installing for lack of
RAM (current mb is 30pin, new mb is 72pin).
Crazy thing is, I spent hours online with ppp installing 2.1 without
any sign of this, although as neither the NE2000 nic or Sony CDU31A
CD-ROM were found by the boot floppy kernel, maybe something is a
little uncomfortable now that they're active.
I did try booting the GENERIC kernel but had the same problems as with
my custom kernel.
I'll be getting the 2.1 CD-ROM in the very near future, so may attempt
a re-install.
--
Andrew I MacIntyre "These thoughts are mine alone..."
E-mail: andrew.macintyre@aba.gov.au (work) | Snail: PO Box 370
andymac@bullseye.apana.org.au (play) | Belconnen ACT 2616
Fido: Andrew MacIntyre, 3:620/243.18 | Australia
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?30e9077f.bullseye>