From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S267238AbUH3Kbt (ORCPT ); Mon, 30 Aug 2004 06:31:49 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S267626AbUH3Kbt (ORCPT ); Mon, 30 Aug 2004 06:31:49 -0400 Received: from [195.23.16.24] ([195.23.16.24]:47276 "EHLO bipbip.comserver-pie.com") by vger.kernel.org with ESMTP id S267238AbUH3Kbq (ORCPT ); Mon, 30 Aug 2004 06:31:46 -0400 Message-ID: <4133020F.1060306@grupopie.com> Date: Mon, 30 Aug 2004 11:31:43 +0100 From: Paulo Marques Organization: Grupo PIE User-Agent: Mozilla Thunderbird 0.7.1 (X11/20040626) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Albert Cahalan Cc: Roger Luethi , William Lee Irwin III , linux-kernel mailing list , Paul Jackson Subject: Re: [BENCHMARK] nproc: netlink access to /proc information References: <20040827122412.GA20052@k3.hellgate.ch> <20040827162308.GP2793@holomorphy.com> <20040828194546.GA25523@k3.hellgate.ch> <20040828195647.GP5492@holomorphy.com> <20040828201435.GB25523@k3.hellgate.ch> <20040829160542.GF5492@holomorphy.com> <20040829170247.GA9841@k3.hellgate.ch> <20040829172022.GL5492@holomorphy.com> <20040829175245.GA32117@k3.hellgate.ch> <20040829181627.GR5492@holomorphy.com> <20040829190050.GA31641@k3.hellgate.ch> <1093810645.434.6859.camel@cube> In-Reply-To: <1093810645.434.6859.camel@cube> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-AntiVirus: checked by Vexira MailArmor (version: 2.0.1.16; VAE: 6.27.0.6; VDF: 6.27.0.37; host: bipbip) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Albert Cahalan wrote: >... > > This is crummy. It's done for wchan, since that is so horribly > expensive, but I'm not liking the larger race condition window. > Remember that PIDs get reused. There isn't a generation counter > or UUID that can be checked. I just wanted to call your attention to the kallsyms speedup patch that is now on the -mm tree. It should improve wchan speed. My benchmarks for kallsyms_lookup (the function that was responsible for the wchan time) went from 1340us to 0.5us. So maybe this is enough not to make wchan a special case anymore... -- Paulo Marques - www.grupopie.com To err is human, but to really foul things up requires a computer. Farmers' Almanac, 1978