[nos-bbs] Telnet forwarding with JNOS

[Michael Fox - N6MEF]

Maiko,

I'm not enough of a telnet guru to know if that approach is "out to lunch"
or not.  I would think maybe you'd want to only use that during forwarding,
and not during a "normal", user initiated telnet session?  Anyway, I'm
trying to get some network connections finished and distributed to our JNOS
network before a deadline later this month.   But I'll try to fit it in.

But as I pointed out in my earlier message, there seems to also be a mistake
in the FBB 1 option.  I know others have told me it doesn't work right and
I've seen the FBB BBS complain about protocol error when I have FBB 1
configured.  And when using FBB 2, I occasionally see a "protocole error
(3)" go by (my spelling is correct).  (When JNOS is talking to an FBB
station and the FBB station says there's an error in the FBB protocol, I
tend to believe the FBB station.)  ;-)  So I'm thinking the FBB code itself
may need some work.

Also, I have a rule in my Linux firewall set to capture attempts to open a
new TCP session without the SYN flag.  It's catching such packets coming
from JNOS.  So there seems to be some telnet issue, too.  

I can capture some .pcap traces if you'd like to see them.  You can use
"wireshark" to view them.  If you haven't tried wireshark, it's stupid
simple to use and gives you full protocol decode.  (I don't think it
understands FBB, but it does understand telnet).   If you want to contact me
off list we can discuss what to capture and where to deposit the files.

Michael



-----Original Message-----
From: nos-bbs-bounces at tapr.org [mailto:nos-bbs-bounces at tapr.org] On Behalf
Of Maiko Langelaar
Sent: Monday, November 19, 2012 7:29 AM
To: TAPR xNOS Mailing List
Subject: [nos-bbs] Telnet forwarding with JNOS

Greetings all,

I've heard people complain about this from time to time, but nobody has ever
really pressed this issue on me, so it never got alot of attention.

Several years ago it came up, the solution was simply to not use telnet
forwarding and setup an AXIP or AXUDP interface instead. I actually took a
look at the code a few years ago, and here is my technical assessment of why
telnet forwarding with JNOS is *inconsistent* - please read it :

The receive process of any telnet connection to JNOS looks for what is
called the Interpret As Command (IAC) Character, which incidently is the
value 255 (0xff). If JNOS sees the IAC in any telnet connection, it will
insert an extra 2 bytes of control characters type of thing.

Now that Michael has brought this up once again, perhaps it's time to just
force the telnet stream to switch off telnet options (as is done for any
AX25 or NETROM streams). If you look at the code in mboxgate.c, look for the
gw_input () function around line number 266, and you will see a bit further
down a 'int no_tel_opts = 0'.

I'm going to create an option in a subsequent release, but in the mean time
(Michael this is for you), edit your code, change the 0 to 1, then try your
telnet forward. I'm sure it will work fine !

This usually just happens with compressed (FBB) modes, chances of a 0xff
character are very high when using any compress forwarding. The chances of
this happening for ascii forwarding are next to none, unless someone some
how inserted a next ascii character in the mail message (using a editor),
and buy some fluke it got in there.

So why didn't I fix it back then ? Well, some people thought my IAC theory
was out to lunch, and the conversation kinda fell off the radar scope, and I
just told myself, okay let the 'know it alls' have their way :)

Have a great day :)

Maiko

_______________________________________________
nos-bbs mailing list
nos-bbs at tapr.org
https://www.tapr.org/cgi-bin/mailman/listinfo/nos-bbs