[16] in SIPB_Linux_Development
re: kerberized rlogin on quiche
daemon@ATHENA.MIT.EDU (Mark W. Eichin)
Mon Jul 5 02:22:23 1993
Date: Sun, 30 May 93 20:46:04 -0400
From: "Mark W. Eichin" <eichin@athena.mit.edu>
To: yandros@mit.edu
Cc: linux-dev@mit.edu
In-Reply-To: Chad Phillip Brown's message of Sat, 29 May 93 23:48 EDT <m0nzeO6-00012TC@quiche.MIT.EDU>
Yes, I've seen that behaviour. I'll have to see if the latest kernel
release deals with it better. It may be enough to try the rlogin
again; part of the problem is that rlogin uses a completely
brain-damaged method to signal certain "out of band" information, and
often this screws up the encrypted channel. One solution is to log in
to rs/6000's (or any other machine where kerberos is compiled -DPOSIX,
possibly decstations as well) because they are broken in a different
way which makes this work.
The "read: Out of memory" is because the Cygnus kerberos code
actually *checks* for an error, as opposed to merely crashing (and
exiting silently) like the MIT code does...
_Mark_
