From owner-cvs-all Mon May 14 7: 1:30 2001 Delivered-To: cvs-all@freebsd.org Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by hub.freebsd.org (Postfix) with ESMTP id E057D37B423; Mon, 14 May 2001 07:01:21 -0700 (PDT) (envelope-from robert@fledge.watson.org) Received: from fledge.watson.org (robert@fledge.pr.watson.org [192.0.2.3]) by fledge.watson.org (8.11.3/8.11.3) with SMTP id f4EE18f70925; Mon, 14 May 2001 10:01:08 -0400 (EDT) (envelope-from robert@fledge.watson.org) Date: Mon, 14 May 2001 10:01:07 -0400 (EDT) From: Robert Watson X-Sender: robert@fledge.watson.org To: Poul-Henning Kamp Cc: Greg Lehey , Eric Melville , Kris Kennaway , cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: Destroying and remaking device nodes (was: cvs commit: src/sys/conf files options src/sys/alpha/conf GENERIC src/sys/i386/conf GENERIC NEWCARD NOTES src/sys/ia64/conf GENERIC src/sys/pc98/conf GENERIC src/sys/fs/devfs devfs_devs.c devfs_vfsops.c devfs_vnops.c) In-Reply-To: <20629.989848031@critter> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Mon, 14 May 2001, Poul-Henning Kamp wrote: > In message , Robe > rt Watson writes: > > >Personally, I'd like to see rm unmodified: part of the goal here is to > >allow /dev to act like a normal file system from the perspective of > >management tools (including things like graphical file managers). > >Modifying base system tools is going to hurt later. We should really just > >allow rm and rm -W to work as normal. They use documented APIs as they > >exist already. > > I have had a number of people ask for "mknod" instead of "rm -W" and > after thinking about it I guess that is the most POLA compliant > solution. > > I'll work that one tonight if I can. On the other hand, from the perspective of dynamically allocated device major numbers for kernel modules, using whiteout *also* makes sense. :-) You can't mknod a device without a priori knowledge of its device numbers. As long as we like that assertion, I'm fine with us using mknod. I recognize the jail issue as a problem, but think it would be fine for the vop_whiteout code to simply do a: if (jailed(p->p_ucred)) return (EPERM); > The mountflags issue persist in preventing the more powerfull > specification of policy... This is a continuing problem that I think Adrian Chadd agreed to fix, but who hasn't been seen fixing anything much lately... :-) Robert N M Watson FreeBSD Core Team, TrustedBSD Project robert@fledge.watson.org NAI Labs, Safeport Network Services To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message