From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jamie Lokier Subject: Re: [RFC PATCH] Filesystem Journal Notifications Date: Sat, 20 Sep 2008 16:51:42 +0100 Message-ID: <20080920155142.GC6061@shareable.org> References: <20080912220649.GB26999@apaithan-desktop.sanmateo.corp.akamai.com> <20080915023126.GF4090@webber.adilger.int> <20080915053937.GG4090@webber.adilger.int> <20080915193648.GC31055@apaithan-desktop.sanmateo.corp.akamai.com> <20080915233740.GF3241@webber.adilger.int> <20080916230552.GB13446@apaithan-desktop.sanmateo.corp.akamai.com> <20080919203402.GI10950@webber.adilger.int> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Abhijit Paithankar , linux-fsdevel@vger.kernel.org, linux-ext4@vger.kernel.org, Dave Chinner , Joel Becker To: Andreas Dilger Return-path: Received: from mail2.shareable.org ([80.68.89.115]:40324 "EHLO mail2.shareable.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750780AbYITPvr (ORCPT ); Sat, 20 Sep 2008 11:51:47 -0400 Content-Disposition: inline In-Reply-To: <20080919203402.GI10950@webber.adilger.int> Sender: linux-ext4-owner@vger.kernel.org List-ID: Andreas Dilger wrote: > That assumes agreement between the applications that are using this > interface. It isn't at all desirable that applications have to know > the "mountpoint" of the filesystem in order to use inotify, and in > some cases (e.g. bind mount in a new namespace) there isn't even access > to the root inode. A filesystem's root inode needn't be mounted at all. You don't need a new namespace - bind mount is enough by itself. It has me wondering - how can an application even tell when it has the root inode of a filesystem? You can't tell from /etc/mtab or /proc/mounts, nor from traversing the filesystem itself - except for filesystems where you know the expected inode number of the root inode. -- Jamie