From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [RCF] [PATCH] unprivileged mount/umount Date: Wed, 11 May 2005 09:51:54 +0100 Message-ID: <20050511085154.GB24495@infradead.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: ericvh@gmail.com, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, smfrench@austin.rr.com, hch@infradead.org Return-path: Received: from pentafluge.infradead.org ([213.146.154.40]:51626 "EHLO pentafluge.infradead.org") by vger.kernel.org with ESMTP id S261932AbVEKIvz (ORCPT ); Wed, 11 May 2005 04:51:55 -0400 To: Miklos Szeredi Content-Disposition: inline In-Reply-To: Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Wed, May 04, 2005 at 04:21:23PM +0200, Miklos Szeredi wrote: > Yes, I see your point. However the problem of malicious filesystem > "traps" applies to private namespaces as well (because of suid > programs). > > So if a user creates a private namespace, it should have the choice of: > > 1) Giving up all suid rights (i.e. all mounts are cloned and > propagated with nosuid) > > 2) Not giving up suid for cloned and propagated mounts, but having > extra limitations (suid/sgid programs cannot access unprivileged > "synthetic" mounts) Although I hate special cases I think that we might need 2) to avoid too much trouble tripping over the global namespace. OTOH that should also accelarate adoption of giving each user a separate namespace on login in the various distribution, which is a good thing ;-)