From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757131AbZCZWaX (ORCPT ); Thu, 26 Mar 2009 18:30:23 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753427AbZCZWaH (ORCPT ); Thu, 26 Mar 2009 18:30:07 -0400 Received: from mx2.mail.elte.hu ([157.181.151.9]:40219 "EHLO mx2.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752848AbZCZWaG (ORCPT ); Thu, 26 Mar 2009 18:30:06 -0400 Date: Thu, 26 Mar 2009 23:29:24 +0100 From: Ingo Molnar To: Stephen Rothwell Cc: Steven Rostedt , linux-kernel@vger.kernel.org, Andrew Morton , Thomas Gleixner , Peter Zijlstra , Frederic Weisbecker , Mike Galbraith , Arjan van de Ven , Steven Rostedt Subject: Re: [PATCH 2/2] tracing: add average time in function to function profiler Message-ID: <20090326222924.GA28755@elte.hu> References: <20090326010737.611522983@goodmis.org> <20090326010811.510051401@goodmis.org> <20090327091507.f133257e.sfr@canb.auug.org.au> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090327091507.f133257e.sfr@canb.auug.org.au> 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 * Stephen Rothwell wrote: > Hi Steve, > > On Wed, 25 Mar 2009 21:07:39 -0400 Steven Rostedt wrote: > > > > From: Steven Rostedt > > > > Show the average time in the function (Time / Hit) > > Just wondering why we need to do this in the kernel ... there is > already enough information for a user to figure these averages > out. Well, this argument can be applied to a lot of information exposed in /proc and /sys. For example in theory there's no need for /proc/interrupts at all, because the IRQ mappings are present in the syslog and in /sys enumerations and the IRQ counts are present in /proc/stat already. I still find /proc/interrupts very handy, despite purist arguments that it is redundant and that it thus should not be there. So we provide such things when it's easy to do, when it's a meaningful piece of information and it is convenient to users and developers. Note that the function profiler is a double-non-default item in /debug (you need to enable two non-default options in the .config to get it), so you cannot stumble on it accidentally. Ingo