From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755612AbZDDUS6 (ORCPT ); Sat, 4 Apr 2009 16:18:58 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753532AbZDDUSq (ORCPT ); Sat, 4 Apr 2009 16:18:46 -0400 Received: from mx2.mail.elte.hu ([157.181.151.9]:40069 "EHLO mx2.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753197AbZDDUSp (ORCPT ); Sat, 4 Apr 2009 16:18:45 -0400 Date: Sat, 4 Apr 2009 22:18:20 +0200 From: Ingo Molnar To: Jens Axboe Cc: Linus Torvalds , Theodore Tso , Linux Kernel Developers List , Ext4 Developers List Subject: Re: [GIT PULL] Ext3 latency fixes Message-ID: <20090404201820.GA18067@elte.hu> References: <1238742067-30814-1-git-send-email-tytso@mit.edu> <20090404135719.GA9812@mit.edu> <20090404151649.GE5178@kernel.dk> <20090404173412.GF5178@kernel.dk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090404173412.GF5178@kernel.dk> User-Agent: Mutt/1.5.18 (2008-05-17) X-ELTE-VirusStatus: clean X-ELTE-SpamScore: -1.5 X-ELTE-SpamLevel: X-ELTE-SpamCheck: no X-ELTE-SpamVersion: ELTE 2.0 X-ELTE-SpamCheck-Details: score=-1.5 required=5.9 tests=BAYES_00 autolearn=no SpamAssassin version=3.2.3 -1.5 BAYES_00 BODY: Bayesian spam probability is 0 to 1% [score: 0.0000] Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Jens Axboe wrote: > > Quite frankly, the fact that I can see _seconds_ of latencies > > with a really good SSD is not acceptable. The fact that it is by > > design is even less so. > > Agree, multi-second latencies is not acceptable. btw., just to insert some hard data: usability studies place the acceptance threshold for delays to around 300 milliseconds. If the computer does not 'respond' (in a real or a fake, visible or audible way) to their input within 0.3 seconds they get annoyed emotionally. It does not matter how complex it is for the kernel to solve this problem, it does not matter how far away and difficult to access the data is. If we are not absolutely latency centric in Linux, if we spuriously delay apps that do supposedly simple-looking things the user _will_ get annoyed and _will_ pick another OS. All things considered it is certainly a combined kernel and app problem space to get there, but not being completely aware of its importance in the kernel kills any chance of us ever having a long-term solution. Ingo