Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1761641AbXEPL1r (ORCPT ); Wed, 16 May 2007 07:27:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1760902AbXEPL1j (ORCPT ); Wed, 16 May 2007 07:27:39 -0400 Received: from lazybastard.de ([212.112.238.170]:39949 "EHLO longford.lazybastard.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756182AbXEPL1i (ORCPT ); Wed, 16 May 2007 07:27:38 -0400 Date: Wed, 16 May 2007 13:23:17 +0200 From: =?utf-8?B?SsO2cm4=?= Engel To: Willy Tarreau Cc: Andrew Morton , linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mtd@lists.infradead.org, Albert Cahalan , Thomas Gleixner , Jan Engelhardt , Evgeniy Polyakov , Pekka Enberg , Greg KH , Ingo Oeser Subject: Re: [PATCH] LogFS take three Message-ID: <20070516112317.GB5472@lazybastard.org> References: <20070515151919.GA32510@lazybastard.org> <20070515133759.9ee434a2.akpm@linux-foundation.org> <20070516000630.GD1220@lazybastard.org> <20070516052253.GI943@1wt.eu> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <20070516052253.GI943@1wt.eu> User-Agent: Mutt/1.5.9i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1647 Lines: 38 On Wed, 16 May 2007 07:22:54 +0200, Willy Tarreau wrote: > > On hard disks, yes, but as you suggested, there are lots of other flash > devices interfaced as block devices. CompactFlash comes to mind, USB > keys too. And on these ones, the most important is to reduce the number > of writes and to support large sizes. I already see LogFS as an interesting > alternative to JFFS2 on such devices, eventhough it does not (yet?) support > compression. It does support compression now. That part was finished, erm, about two weeks ago. Logfs on Smartmedia format (I assume most CompactFlash, USB, SD and MMC media use that) will currently cause a nasty interaction. Unlike JFFS2, logfs will alternately write to several eraseblocks. On real flash, alternating between half-finished blocks is free. Smartmedia standard requires to finish the last block before moving to the next. So the logfs write pattern will cause it to add an erase/write overhead of - depending on erase size and compression - 4x to 512x. Pretty bad. If Smartmedia was really stupid, there would be no overhead at all. Any manucaturers listening should seriously think about allowing raw flash access to their devices. ;) Jörn -- To recognize individual spam features you have to try to get into the mind of the spammer, and frankly I want to spend as little time inside the minds of spammers as possible. -- Paul Graham - 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/