Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754017Ab1FPDH3 (ORCPT ); Wed, 15 Jun 2011 23:07:29 -0400 Received: from mail01-md.ns.itscom.net ([175.177.155.111]:38019 "EHLO mail01-md.ns.itscom.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752225Ab1FPDH1 (ORCPT ); Wed, 15 Jun 2011 23:07:27 -0400 From: "J. R. Okajima" Subject: Re: [PATCH 0/7] overlay filesystem: request for inclusion To: Erez Zadok Cc: Miklos Szeredi , Alan Cox , Valerie Aurora , Andrew Morton , NeilBrown , "viro@ZenIV.linux.org.uk" , "torvalds@linux-foundation.org" , "linux-fsdevel@vger.kernel.org" , "linux-kernel@vger.kernel.org" , "apw@canonical.com" , "nbd@openwrt.org" , "hramrach@centrum.cz" , "jordipujolp@gmail.com" In-Reply-To: References: <20110609125114.8dff08da.akpm@linux-foundation.org> <20110610100143.28037551@lxorguk.ukuu.org.uk> <8739jbjqa7.fsf@tucsk.pomaz.szeredi.hu> <11186.1308148376@jrobl> <803fd88dc28748428861b75afdee3575@HUBCAS1.cs.stonybrook.edu> Date: Thu, 16 Jun 2011 12:07:19 +0900 Message-ID: <18299.1308193639@jrobl> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1628 Lines: 36 Erez Zadok: > ... Asking = > overlayfs or other stackable file systems to solve this multi-layer = > coherency perfectly is somewhat ridiculous: we don't expect file systems = > like ext3 to detect and correctly handle changes to lower devices =97 = > i.e., if someone hand-edited direct blocks in /dev/sda1, do we? I agree with you if we discuss about union-type-mount, which handles a block device as its member. As long as the layered-fs handles a directory (mounted filesystem) as its member, it is obviously right that users expect the modification on the member fs (by-passing a union) is available. Of course I agree it brings complication to us, and I'd suggest three level options to support this issue. - detect the direct changes and reflect it to union (hardest option) - skip the detection, but verify the parent-child relationship or more at least. (this is something like overlayfs is trying to do) - skip both of the detection and verification (lowest option) this option depends how user sets up the union and its member. if user hides the members totally by over-mounting an empty dir on the member (or something), then he can specify this option. otherwise, this option is dangerous. also some symlinks may not work. # mkdir /hide # mount -o upper=/rw,lower=/ro none /union # mount -o bind /hide /rw # mount -o bind /hide /ro J. R. Okajima -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/