Received: from SOUTH-STATION-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA12355; Fri, 1 Dec 95 17:44:43 EST
Received: from LOLA-GRANOLA.MIT.EDU by MIT.EDU with SMTP
	id AA07600; Fri, 1 Dec 95 17:42:43 EST
Received: (from ghudson@localhost) by lola-granola.MIT.EDU (8.6.11/8.6.11) id RAA18621; Fri, 1 Dec 1995 17:43:48 -0500
Date: Fri, 1 Dec 1995 17:43:48 -0500
From: Greg Hudson <ghudson@MIT.EDU>
Message-Id: <199512012243.RAA18621@lola-granola.MIT.EDU>
To: mhbraun@MIT.EDU
Cc: ghudson@MIT.EDU, pc-dialup@MIT.EDU
In-Reply-To: "[34] in pc-dialup"
Subject: Re: Added gunner

> The reason why we have write_message in gunner instead of an echo
> like you do is that sometimes the echo fails to exit do to some
> weird state of the tty and we end up with scores of gunner processes
> sitting in wait.  It is possible this is not a problem on
> netbsd...we should keep an eye on it.

It's a problem under NetBSD (that is, we had scores of gunner process
blocking on writing to ttyp2).  I've fixed gunner to use write_tty.c
to avoid this.

The problem occurred due to a login from kretch from tunafish.mit.edu.
The login process does not respond to HUP or KILL, and is in state
IEs+, indicating an idle, exiting session leader process in the
foreground:

	ringworld# ps tp2
	  PID TT  STAT      TIME COMMAND
	 3019 p2- IEs+   0:00.23 (login)
	 3031 p2- Z+     0:00.00 (tcsh)
	 3454 p2- IE     0:02.91 /usr/athena/bin/zwgc 

I don't know what could cause this to happen.  Right now I'm going to
try to figure out why login left /etc/ptmp around.

