From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758358AbZKRTjd (ORCPT ); Wed, 18 Nov 2009 14:39:33 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1758023AbZKRTjc (ORCPT ); Wed, 18 Nov 2009 14:39:32 -0500 Received: from mx2.mail.elte.hu ([157.181.151.9]:39722 "EHLO mx2.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754776AbZKRTjb (ORCPT ); Wed, 18 Nov 2009 14:39:31 -0500 Date: Wed, 18 Nov 2009 20:39:27 +0100 From: Ingo Molnar To: Peter Zijlstra Cc: "Luck, Tony" , Paul Mackerras , "linux-kernel@vger.kernel.org" Subject: Re: [PATCH] ia64 support for tools/perf/ Message-ID: <20091118193927.GA30470@elte.hu> References: <4b02d7f43514327a@agluck-desktop.sc.intel.com> <20091117172139.GB5967@elte.hu> <57C9024A16AD2D4C97DC78E552063EA3E3876A5B@orsmsx505.amr.corp.intel.com> <1258563396.3918.698.camel@laptop> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1258563396.3918.698.camel@laptop> User-Agent: Mutt/1.5.20 (2009-08-17) X-ELTE-SpamScore: -2.0 X-ELTE-SpamLevel: X-ELTE-SpamCheck: no X-ELTE-SpamVersion: ELTE 2.0 X-ELTE-SpamCheck-Details: score=-2.0 required=5.9 tests=BAYES_00 autolearn=no SpamAssassin version=3.2.5 -2.0 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 * Peter Zijlstra wrote: > On Tue, 2009-11-17 at 09:25 -0800, Luck, Tony wrote: > > > Btw., are you planning HAVE_PERF_EVENTS support for IA64? (It's nice > > > stuff, i can only recommend it! ;-) > > > > My first goal is getting the s/w events hooked up and working. I'm > > still wondering about how to get the h/w counters to co-exist with > > perfmon (which has been part of ia64 API for many years). > > A first approach could be to make them mutually exclusive at runtime. > When there's a perf_event user, perfmon users get FAIL and vs. > > This is basically the state for x86 perf_event vs oprofile atm. Yep, this is probably the least intrusive approach. Mixing the implementations beyond trivial exclusion would probably be pain. Ingo