From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759741AbYHSUg1 (ORCPT ); Tue, 19 Aug 2008 16:36:27 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1758853AbYHSUfy (ORCPT ); Tue, 19 Aug 2008 16:35:54 -0400 Received: from mx1.redhat.com ([66.187.233.31]:58772 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758334AbYHSUfw (ORCPT ); Tue, 19 Aug 2008 16:35:52 -0400 Date: Tue, 19 Aug 2008 16:33:45 -0400 From: Rik van Riel To: Pavel Machek Cc: Alan Cox , Eric Paris , Arjan van de Ven , Jan Harkes , "Press, Jonathan" , peterz@infradead.org, linux-kernel@vger.kernel.org, malware-list@lists.printk.net, hch@infradead.org, andi@firstfloor.org, viro@ZenIV.linux.org.uk Subject: Re: HSM (was Re: [malware-list] TALPA - a threat model? well sorta.) Message-ID: <20080819163345.196961f9@riellaptop.surriel.com> In-Reply-To: <20080819192027.GA6484@ucw.cz> References: <20080813205906.559d3f37@lxorguk.ukuu.org.uk> <2629CC4E1D22A64593B02C43E855530304AE4BC2@USILMS12.ca.com> <20080813173529.7069b5f1@cuia.bos.redhat.com> <20080815201622.GD31584@cs.cmu.edu> <20080815150509.20ffb91d@infradead.org> <1219015177.27389.1.camel@localhost.localdomain> <20080818163313.72927af7@lxorguk.ukuu.org.uk> <20080818124339.4b65b6c0@riellaptop.surriel.com> <20080819071416.GA14731@elf.ucw.cz> <20080819121032.02632c09@riellaptop.surriel.com> <20080819192027.GA6484@ucw.cz> Organization: Red Hat, Inc X-Mailer: Claws Mail 3.5.0 (GTK+ 2.12.11; i386-redhat-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 19 Aug 2008 21:20:27 +0200 Pavel Machek wrote: > > If you think FUSE+unionfs is a cleaner solution than one > > hook in the VFS, I've got a bridge to sell you. > > If you can do it with one clean enough hook, I'll buy that bridge. > > [If you want to do 'list directory before files are there' -- and you > seem to want to from description above -- fuse seems like a way to > go.] For a poor man's HSM, it may be a possibility to simply restore all the files as sparse files, fetching the content on demand. The files can be marked as HSM-empty files with some kind of bit flag or xattr. Restoring sparse files takes very little time. This scheme could even work for directories, so you can start using the filesystem before the whole directory tree has been restored. All it needs is block-on-open when the file has a special flag or xattr set. -- All rights reversed.