From owner-freebsd-stable  Tue Mar 23 10:38:30 1999
Delivered-To: freebsd-stable@freebsd.org
Received: from ensta.ensta.fr (ensta.ensta.fr [147.250.1.1])
	by hub.freebsd.org (Postfix) with ESMTP id 00A6615378
	for <freebsd-stable@FreeBSD.ORG>; Tue, 23 Mar 1999 10:38:26 -0800 (PST)
	(envelope-from loyer@quickstep.ensta.fr)
Received: from quickstep.ensta.fr (quickstep.ensta.fr [147.250.61.100])
	by ensta.ensta.fr (8.9.1/8.9.1) with ESMTP id TAA08094;
	Tue, 23 Mar 1999 19:38:05 +0100 (CET)
Received: (from loyer@localhost)
	by quickstep.ensta.fr (8.9.2/8.8.5) id UAA20478;
	Mon, 22 Mar 1999 20:04:08 +0100 (CET)
From: Frederic LOYER <loyer@ensta.fr>
Message-Id: <199903221904.UAA20478@quickstep.ensta.fr>
Subject: Re: Strange Error Message
In-Reply-To: <000001be7402$9766b5c0$0402030a@yiff.bunnynet.org> from Matt White at "Mar 21, 1999  8:23:28 pm"
To: mwhite@donet.com (Matt White)
Date: Mon, 22 Mar 1999 20:04:07 +0100 (CET)
Cc: freebsd-stable@FreeBSD.ORG
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-stable@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

> Hello...
> 
> Today while copying a large number of files I started getting this error
> message:
> 
> inetd in realloc(): warning: junk pointer, too low to make sense.
> inetd in free(): warning: junk pointer, too low to make sense.
> 
> This is a -STABLE system, last cvsupped and built on the 20th of this month.
> 
> Any idea what this means?  It prevented me from logging in, whenever I'd try
> to start up a telnet session it would toss those two lines to me and close
> the session. (I was still able to ssh in, which is not started via inetd.)
> 
> Thanks,
> Matt

I've found the same problem. This means there is a bug in "inetd" or one
of the called libraries. Trying to work with a debugger is difficult
since inetd crash after the fork(): you can't login but the "main"
inetd process is still alive to answer the following connection...
and crash again.

It seems to happen in malloc() called from login_getclass()... but
this doesn't tell us where the memory is screwed: there must be something
wrong before the fork() since this error comes on *every* connection after
the first error.

It could be triggered by the folowing lines:

comsat  dgram   udp     wait    tty:tty /usr/libexec/comsat     comsat
ntalk   dgram   udp     wait    tty:tty /usr/libexec/ntalkd     ntalkd

tty:tty is quite recent on -STABLE, and I don't remember seeing this
error before updating /etc/inetd.conf.  Anyway this error is too rare to
be debugged easily (I can't reproduce it).

--
Frederic Loyer <loyer@ensta.fr>


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message