From: Arjan van de Ven Subject: Re: [GIT PULL] Ext3 latency fixes Date: Sun, 5 Apr 2009 11:56:29 -0700 Message-ID: <20090405115629.521057fc@infradead.org> References: <20090404135719.GA9812@mit.edu> <20090404151649.GE5178@kernel.dk> <20090404173412.GF5178@kernel.dk> <20090404180108.GH5178@kernel.dk> <20090404232222.GA7480@mit.edu> <20090404163349.20df1208@infradead.org> <20090405001005.GA7553@mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: Theodore Tso , Jens Axboe , Linux Kernel Developers List , Ext4 Developers List To: Linus Torvalds Return-path: Received: from casper.infradead.org ([85.118.1.10]:38588 "EHLO casper.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751224AbZDESzI (ORCPT ); Sun, 5 Apr 2009 14:55:08 -0400 In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On Sun, 5 Apr 2009 10:01:06 -0700 (PDT) Linus Torvalds wrote: > Also, one of the issues seems to literally be that the higher-level > request handling doesn't care AT ALL about the priority. Allocating > the request itself does keep reads and writes separated, but if the > request is a SYNCIO request, and non-sync writes have filled up th > write requests, we'll have to wait for the background writes to free > up requests. > > That is quite possibly the longest wait we have in the system. it often is; latencytop tends to point that out. > > See get_request(): our default number of requests is so low that we very regularly hit the limit. In addition to setting kjournald to higher priority, I tend to set the number of requests to 4096 or so to improve interactive performance on my own systems. That way at least the elevator has a chance to see the requests ;-) -- Arjan van de Ven Intel Open Source Technology Centre For development, discussion and tips for power savings, visit http://www.lesswatts.org