From owner-freebsd-stable@FreeBSD.ORG  Tue Oct 14 03:29:27 2003
Return-Path: <owner-freebsd-stable@FreeBSD.ORG>
Delivered-To: freebsd-stable@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id D14CC16A4B3
	for <freebsd-stable@freebsd.org>;
	Tue, 14 Oct 2003 03:29:27 -0700 (PDT)
Received: from smithers.nildram.co.uk (smithers.nildram.co.uk [195.112.4.34])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 87BF543F75
	for <freebsd-stable@freebsd.org>;
	Tue, 14 Oct 2003 03:29:26 -0700 (PDT)
	(envelope-from aivanov@sigsegv.cx)
Received: from gondor (arivanov.gotadsl.co.uk [213.208.103.208])
	by smithers.nildram.co.uk (Postfix) with ESMTP id 9FDA324DBDC
	for <freebsd-stable@freebsd.org>;
	Tue, 14 Oct 2003 11:22:39 +0100 (BST)
Received: from [172.28.0.84] (helo=localhost.localdomain)
	by gondor with esmtp (Exim 3.35 #1 (Debian))
	id 1A9MK9-0000vD-00
	for <freebsd-stable@freebsd.org>; Tue, 14 Oct 2003 11:22:41 +0100
From: Anton Ivanov <aivanov@sigsegv.cx>
To: freebsd-stable@freebsd.org
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
X-Mailer: Ximian Evolution 1.0.5 
Date: 14 Oct 2003 11:22:11 +0100
Message-Id: <1066126961.1306.11.camel@via800-1>
Mime-Version: 1.0
Subject: Netgraph initialization race?
X-BeenThere: freebsd-stable@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: Production branch of FreeBSD source code
	<freebsd-stable.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>,
	<mailto:freebsd-stable-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-stable>
List-Post: <mailto:freebsd-stable@freebsd.org>
List-Help: <mailto:freebsd-stable-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-stable>,
	<mailto:freebsd-stable-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Oct 2003 10:29:27 -0000

Hi list,

I think I found a race condition somewhere in the netgraph subsystem

It can be reproduced around 80% of the time on any 4.8 through 4.9-RC to
5.1-RELENG, possibly earlier. 4.7 IFIRC did not have the problem.

If you start two mpd processes immediately one after another listening
on different IPs with different interface ranges (ng1-ng32) and
(ng33-ng64) you get a hang. Currently running processes continue to run.
New processes cannot be created. mpd instances cannot be killed. 

I have also noticed that if you start and stop an mpd now you get a ng0
interface left after mpd exits. 

Brgds,
A.