From owner-freebsd-bugs  Fri Oct 11 11:00:04 1996
Return-Path: owner-bugs
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id LAA14673
          for bugs-outgoing; Fri, 11 Oct 1996 11:00:04 -0700 (PDT)
Received: (from gnats@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id LAA14661;
          Fri, 11 Oct 1996 11:00:03 -0700 (PDT)
Resent-Date: Fri, 11 Oct 1996 11:00:03 -0700 (PDT)
Resent-Message-Id: <199610111800.LAA14661@freefall.freebsd.org>
Resent-From: gnats (GNATS Management)
Resent-To: freebsd-bugs
Resent-Reply-To: FreeBSD-gnats@freefall.FreeBSD.org,
        anderson@hawaii.conterra.com
Received: from hawaii.conterra.com (hawaii.conterra.com [206.30.180.3])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id KAA14350
          for <FreeBSD-gnats-submit@freebsd.org>; Fri, 11 Oct 1996 10:53:47 -0700 (PDT)
Received: (from anderson@localhost) by hawaii.conterra.com (8.7.5/8.6.9) id NAA08025; Fri, 11 Oct 1996 13:47:55 -0400 (EDT)
Message-Id: <199610111747.NAA08025@hawaii.conterra.com>
Date: Fri, 11 Oct 1996 13:47:55 -0400 (EDT)
From: "Stuart R. Anderson" <anderson@hawaii.conterra.com>
Reply-To: anderson@hawaii.conterra.com
To: FreeBSD-gnats-submit@freebsd.org
X-Send-Pr-Version: 3.2
Subject: bin/1774: telnet hangup bug
Sender: owner-bugs@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk


>Number:         1774
>Category:       bin
>Synopsis:       telnet spins when killed before network is closed
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Oct 11 11:00:02 PDT 1996
>Last-Modified:
>Originator:     Stuart R. Anderson
>Organization:
Stuart Anderson                  anderson@conterra.com
Conterra Communications
>Release:        FreeBSD 2.1-STABLE i386
>Environment:

	Main server runnning ISP business. Pentium/166.

>Description:

	telnet can go into a run-bound loop if it is killed during the shutdown
	process. telnet keeps going into deadpeer().

>How-To-Repeat:

	We are using pmwho (a portmaster related utility) to create a web
	page displaying who is logged in. Pmwho uses telnet to communicate
	with the Portmaster. When pmwho exits, telnet receives a signal which
	begins the looping.

	During the last incident, we managed to get a load avarage of 100+.

>Fix:

	Reset the signal handler to avoid the loop. There is probably one or
	two other places where this could be placed, but this one works for us.
	
diff -c sys_bsd.c.orig sys_bsd.c
*** sys_bsd.c.orig      Fri Oct 11 13:33:49 1996
--- sys_bsd.c   Fri Oct 11 13:33:59 1996
***************
*** 789,794 ****
--- 789,795 ----
  NetClose(fd)
      int       fd;
  {
+     (void) signal(SIGPIPE, SIG_DFL);
      return close(fd);
  }
  


>Audit-Trail:
>Unformatted: