From owner-freebsd-hackers@FreeBSD.ORG Wed May 25 17:55:46 2005 Return-Path: X-Original-To: freebsd-hackers@freebsd.org Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id D94D616A41C for ; Wed, 25 May 2005 17:55:46 +0000 (GMT) (envelope-from dga+@cs.cmu.edu) Received: from smtp.andrew.cmu.edu (smtp.andrew.cmu.edu [128.2.10.81]) by mx1.FreeBSD.org (Postfix) with ESMTP id 82D9243D1D for ; Wed, 25 May 2005 17:55:46 +0000 (GMT) (envelope-from dga+@cs.cmu.edu) Received: from [192.168.15.50] (dsl093-060-134.pit1.dsl.speakeasy.net [66.93.60.134]) (user=dga2 mech=PLAIN (0 bits)) by smtp.andrew.cmu.edu (8.12.10/8.12.10) with ESMTP id j4PHtkLG023176 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NOT); Wed, 25 May 2005 13:55:46 -0400 In-Reply-To: References: Mime-Version: 1.0 (Apple Message framework v622) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: Content-Transfer-Encoding: 7bit From: David Andersen Date: Wed, 25 May 2005 13:55:41 -0400 To: David Andersen X-Mailer: Apple Mail (2.622) Cc: freebsd-hackers@freebsd.org Subject: Re: Opening raw disk while mounted in 5.x? X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 25 May 2005 17:55:47 -0000 To answer my own question - this appears to be done in the geom subsystem now. It can be avoided by enabling the magic foot-shooting debug flag: sysctl -w kern.geom.debugflags=16 if you're inclined to shoot yourself in the foot. -Dave On May 25, 2005, at 1:10 PM, David Andersen wrote: > Hoping someone knows the quick answer to this - in 4.x, it was > possible to open /dev/ad0 while a filesystem one one of its slices was > mounted. This no longer appears possible under 5.x. Could someone > point me to the spot in the code where I'd need to disable a > permissions check (or a sysctl, or anything) to permit this behavior > again? > > (The context: mounting slice 2 of a disk and using it to store a > compressed filesystem image. Then opening the primary disk device to > directly write the filesystem image onto slice 1. The kernel seems to > muck with the write calls if we try to do the write onto slice 1 > instead of the raw disk). This is using the very cool > imagezip/imageunzip utilities from the Utah Emulab project. > > Thanks! > > -Dave >