[912] in Kerberos_V5_Development
pty handling in krlogind
daemon@ATHENA.MIT.EDU (Sam Hartman)
Wed Jul 19 23:16:12 1995
Date: Wed, 19 Jul 1995 23:15:56 -0400
From: Sam Hartman <hartmans@MIT.EDU>
To: probe@MIT.EDU, eichin@MIT.EDU
Cc: krbdev@MIT.EDU
I notice that you were one of the last people to significantly
hack on the pty handling in krlogind.c, and there's something the code
does that I find confusing, so I was wondering if you might be able to
elucidate it's behavior.
Basically, under Ultrix, it calls vhangup when the terminal is
closed, but under SunOS, it calls vhangup initially. I looked at the
man page and can't really determine what would cause this difference
in behavior. I'm trying to pull out the operating system dependencies
and test for the specific case. In particular, I'm wondering if the
problem is just with Ultrix doing a vhangup on the wrong controlling
terminal, or if there's some deeper bug.
Apparently, bsd-4.3 rlogind and telnetd take the current
behavior, only doing a vhangup at the end. I looked at the 4.3 kernel
sources and don't see anything that would cause a major problem.
Also, I see that 4.3's getty does actually perform a vhangup at the
beginning.
I don't have access to Ultrix sources to make the same checks
on that platform, so I would appreciate any advice people have on why
this is the current behavior, and what breaks if it changes. I would
especially appreciate help that would allow me to be certain that this
either really is an Ultrix-specificthat there's some actual autoconf
test I can use.
--Sam