From: =?iso-8859-1?Q?J=F6rn?= Engel Subject: Re: [PATCH 00/22][RFC] Unionfs: Stackable Namespace Unification Filesystem Date: Tue, 5 Sep 2006 09:01:10 +0200 Message-ID: <20060905070110.GA30923@wohnheim.fh-wedel.de> References: <20060901013512.GA5788@fsl.cs.sunysb.edu> <20060903110507.GD4884@ucw.cz> <20060904125744.GA1961@wohnheim.fh-wedel.de> <200609050746.44502.a1426z@gawab.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Josef Sipek , linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, hch@infradead.org, akpm@osdl.org, viro@ftp.linux.org.uk, Pavel Machek Return-path: Received: from wohnheim.fh-wedel.de ([213.39.233.138]:6530 "EHLO wohnheim.fh-wedel.de") by vger.kernel.org with ESMTP id S965183AbWIEHB0 (ORCPT ); Tue, 5 Sep 2006 03:01:26 -0400 To: Al Boldi Content-Disposition: inline In-Reply-To: <200609050746.44502.a1426z@gawab.com> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Tue, 5 September 2006 07:46:44 +0300, Al Boldi wrote: > J=F6rn Engel wrote: > > > > Direct modification of branches is similar to direct modification o= f > > block devices underneith a mounted filesystem. While I agree that > > such a thing _should_ not oops the kernel, I'd bet that you can eas= ily > > run a stresstest on a filesystem while randomly flipping bits in th= e > > block device and get just that. >=20 > Not really a fair comparison. The block level is conceptionally tota= lly=20 > different than the fs level, while a stackable fs is within the realm= s of=20 > the fs level. Well, I didn't realize that unionfs required its backing filesystems to be mounted. That's more like having the block device open in a text editor while mounting ext3. In the presence of such a design, an oops clearly is not acceptable. And this sort of design is just what I was talking about when I said: > > There are bigger problems in unionfs to worry about. J=F6rn --=20 You can't tell where a program is going to spend its time. Bottlenecks occur in surprising places, so don't try to second guess and put in a speed hack until you've proven that's where the bottleneck is. -- Rob Pike - To unsubscribe from this list: send the line "unsubscribe linux-fsdevel= " in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html