From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757178AbZJGGon (ORCPT ); Wed, 7 Oct 2009 02:44:43 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756622AbZJGGon (ORCPT ); Wed, 7 Oct 2009 02:44:43 -0400 Received: from mx3.mail.elte.hu ([157.181.1.138]:48095 "EHLO mx3.mail.elte.hu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756415AbZJGGom (ORCPT ); Wed, 7 Oct 2009 02:44:42 -0400 Date: Wed, 7 Oct 2009 08:43:55 +0200 From: Ingo Molnar To: Frederic Weisbecker Cc: Peter Zijlstra , Arnaldo Carvalho de Melo , Paul Mackerras , Mike Galbraith , LKML Subject: Re: [RFC][PATCH] perf tools: Merge trace.info content into perf.data Message-ID: <20091007064355.GA3283@elte.hu> References: <20091006213643.GA5343@nowhere> <20091007062204.GE21673@elte.hu> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20091007062204.GE21673@elte.hu> User-Agent: Mutt/1.5.18 (2008-05-17) 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=3.2.5 -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 * Ingo Molnar wrote: > * Frederic Weisbecker wrote: > > > Hi, > > > > Here is an attempt to remove the trace.info file. It works well for > > me, the reason for it to be an RFC is that I have doubts about the > > backward compatibility. > > > > A file created by perf after his patch is unsupported by previous > > version because the size of the headers have increased. > > That's OK i think in terms of trace.info - trace.info was a temporary > hack to begin with. > > > That said, it's two new fields that have been added in the end of the > > headers, and those could be ignored by previous versions if they just > > handled the dynamic header size and then ignore the unknow part. The > > offsets guarantee the compatibility. > > Yes, that's how it should work. > > > But previous versions handle the header size using its static size, > > not dynamic, then it's not backward compatible. > > > > Anyway, I'm not sure exactly how to handle that. > > That's a bug in the previous version - mind doing a standalone fix for > that so we can mark it Cc: stable? Meanwhile i've applied your two patches - they are clear steps forward. We'll need the -stable fix so that the new perf.data can be read. (and even that only affects -R afaics - so normal perf record / perf report is compatible.) Btw., we also need a patch for new perf to read older perf.data files [non-trace.info ones], as those are not working either: $ perf report Fatal: incompatible file format ( We dont want to push it - i.e. i dont think we need to support old perf.data + trace.info combos. ) Ingo