From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cuda.sgi.com (cuda2.sgi.com [192.48.176.25]) by oss.sgi.com (8.14.3/8.14.3/SuSE Linux 0.8) with ESMTP id n847JO1r183045 for ; Fri, 4 Sep 2009 02:19:35 -0500 Received: from kernel.dk (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id 7C72A427849 for ; Fri, 4 Sep 2009 00:20:16 -0700 (PDT) Received: from kernel.dk (brick.kernel.dk [93.163.65.50]) by cuda.sgi.com with ESMTP id Nx0WC8pNEZFjnNh0 for ; Fri, 04 Sep 2009 00:20:16 -0700 (PDT) Date: Fri, 4 Sep 2009 09:20:13 +0200 From: Jens Axboe Subject: Re: ext4 writepages is making tiny bios? Message-ID: <20090904072013.GR18599@kernel.dk> References: <20090901184450.GB7885@think> <20090901205744.GE6996@mit.edu> <20090901212740.GA9930@infradead.org> <20090903055201.GA7146@discord.disaster> <20090903164209.GA28384@infradead.org> <20090904001545.GA30759@mit.edu> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <20090904001545.GA30759@mit.edu> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: xfs-bounces@oss.sgi.com Errors-To: xfs-bounces@oss.sgi.com To: Theodore Tso Cc: xfs@oss.sgi.com, Christoph Hellwig , linux-fsdevel@vger.kernel.org, linux-ext4@vger.kernel.org, Chris Mason On Thu, Sep 03 2009, Theodore Tso wrote: > On Thu, Sep 03, 2009 at 12:42:09PM -0400, Christoph Hellwig wrote: > > > Careful: > > > > > > - tloff = min(tlast, startpage->index + 64); > > > + tloff = min(tlast, startpage->index + 8192); > > > > > > That will cause 64k page machines to try to write back 512MB at a > > > time. This will re-introduce similar to the behaviour in sles9 where > > > writeback would only terminate at the end of an extent (because the > > > mapping end wasn't capped like above). > > > > Pretty good point, any applies to all the different things we discussed > > recently. Ted, should be maybe introduce a max_writeback_mb instead of > > the max_writeback_pages in the VM, too? > > Good point. > > Jens, maybe we should replace my patch with this one, which makes the > tunable in terms of megabytes instead of pages? That is probably a better metric than 'pages', lets update it. -- Jens Axboe _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs