From owner-freebsd-current@FreeBSD.ORG  Thu Apr 10 02:17:48 2003
Return-Path: <owner-freebsd-current@FreeBSD.ORG>
Delivered-To: freebsd-current@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 9633F37B401
	for <current@freebsd.org>; Thu, 10 Apr 2003 02:17:48 -0700 (PDT)
Received: from mailhub.fokus.fraunhofer.de (mailhub.fokus.fraunhofer.de
	[193.174.154.14])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 630A443FAF
	for <current@freebsd.org>; Thu, 10 Apr 2003 02:17:47 -0700 (PDT)
	(envelope-from brandt@fokus.fraunhofer.de)
Received: from beagle (beagle [193.175.132.100])h3A9HgE07536;
	Thu, 10 Apr 2003 11:17:42 +0200 (MEST)
Date: Thu, 10 Apr 2003 11:17:42 +0200 (CEST)
From: Harti Brandt <brandt@fokus.fraunhofer.de>
To: "M. Warner Losh" <imp@bsdimp.com>
In-Reply-To: <20030409.233705.56936788.imp@bsdimp.com>
Message-ID: <20030410110656.M664@beagle.fokus.fraunhofer.de>
References: <20030409183521.D644@beagle.fokus.fraunhofer.de>
 <20030409.233705.56936788.imp@bsdimp.com>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
cc: current@freebsd.org
Subject: Re: LOR in if_detach
X-BeenThere: freebsd-current@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: Discussions about the use of FreeBSD-current
	<freebsd-current.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-current>,
	<mailto:freebsd-current-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-current>
List-Post: <mailto:freebsd-current@freebsd.org>
List-Help: <mailto:freebsd-current-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-current>,
	<mailto:freebsd-current-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Thu, 10 Apr 2003 09:17:48 -0000

On Wed, 9 Apr 2003, M. Warner Losh wrote:

MWL>How does the lock order get to be radix, driver in the first place?

It took me some time to find how this comes. It occures with drivers, that
don't use the default ifa_rtrequest, in my case in the midway driver.
In rtrequest1 the radix node head is locked. Later on it calls
ifa->ifa_rtrequest, which in turn calls the driver, where the driver lock
occures.

I suppose, that the radix head lock should not be held accross the
ifa_rtrequest call, but it is needed a few lines below. Unfortunately my
understanding of the routing code is limited, so I don't know whether this
could be rearranged.

harti
-- 
harti brandt,
http://www.fokus.fraunhofer.de/research/cc/cats/employees/hartmut.brandt/private
brandt@fokus.fraunhofer.de, harti@freebsd.org