From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from merlin.infradead.org ([205.233.59.134]:57490 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751512Ab2LCSdq (ORCPT ); Mon, 3 Dec 2012 13:33:46 -0500 Message-ID: <50BCF086.8040105@kernel.dk> Date: Mon, 03 Dec 2012 19:33:42 +0100 From: Jens Axboe MIME-Version: 1.0 Subject: Re: I/O is issued twice at scsi level References: <1582502590.159531.1354372015482.JavaMail.root@thomas-krenn.com> <50BCAC6C.6010009@kernel.dk> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: fio-owner@vger.kernel.org List-Id: fio@vger.kernel.org To: Hiroyuki Yamada Cc: Andrey Kuzmin , =?ISO-8859-1?Q?Georg_Sch?= =?ISO-8859-1?Q?=F6nberger?= , fio@vger.kernel.org On 2012-12-03 18:09, Hiroyuki Yamada wrote: > Sorry for update. > I figured it out. > It's from ext3's indirect block mapping. > There is one mapping block in each 1024 blocks, > so accessing a large file randomly has high chance of accessing the > mapping block and the requested block. > > Ext4 has different and more efficient addressing called extent (tree), and > we can avoid the issue. Ah, yes that makes sense. I thought we were talking about extra writes, but it seems I misread the log since it clearly states FROM_DEVICE transfers. -- Jens Axboe