From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx2.redhat.com ([66.187.237.31]:40145 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752041AbZBWRNG (ORCPT ); Mon, 23 Feb 2009 12:13:06 -0500 Date: Mon, 23 Feb 2009 12:10:32 -0500 From: "Frank Ch. Eigler" Subject: Re: [PATCH] tracing/markers: make markers select tracepoints Message-ID: <20090223171032.GA2566@redhat.com> References: <499edf47.1818d00a.060b.2b8d@mx.google.com> <499EE162.4050008@oracle.com> <20090220172241.GF24538@elte.hu> <1235302980.4632.8.camel@laptop> <1235304876.4632.10.camel@laptop> <20090222122424.GA9951@redhat.com> <1235387487.4645.28.camel@laptop> <20090223154404.GA30147@redhat.com> <1235406159.4645.366.camel@laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1235406159.4645.366.camel@laptop> Sender: linux-kbuild-owner@vger.kernel.org List-ID: To: Peter Zijlstra Cc: Ingo Molnar , Randy Dunlap , Frederic Weisbecker , Steven Rostedt , linux-kernel@vger.kernel.org, zippel@linux-m68k.org, linux-kbuild@vger.kernel.org On Mon, Feb 23, 2009 at 05:22:39PM +0100, Peter Zijlstra wrote: > [...] > > > Not so. In both cases the regular stuff (NMI trace, OOPS, > > > function/graph/sched trace, etc) is not enough and you wish to > > > augment its output. > > > > Sorry, I don't see how that relates. If the general function tracing > > widgetry is insufficient for some subsystem/purpose, some sort of > > static instrumentation is needed. Whether that instrumentation is > > done by markers (with a thin glue to ftrace) or by tracepoints (with a > > thick glue to ftrace) doesn't change the need for "augmentation". > > I'm not arguing against static instrumentation per-se (although > expanding the coverage of dynamic/automatic instrumentation is much more > profitable IMHO). Much prior discussion (incl. at the kernel summit) indicates that we need both. > What I'm arguing is that trace_mark()s one distinguishing feature over > tracepoints is only suited for quick debug like work. I see where you're coming from, but one may also caricaturize the other alternative as requiring make-work glue code to pack & unpack all the same inforation. > Furthermore, trace_mark() exposes that crap like an ABI, now suppose > some distro goes and declares that stable for some daft reason, > imagine the poor sod having to fix something littered with > trace_mark(). The impression that this is somehow different with tracepoints is mistaken. Tracepoints are *exactly* as "ABI-like" as markers. > [...] presenting that information in big bloated blobs is beyond > that scope. Do you have some specific bloated blobs in mind? It's not as if the rendered text is necessarily much bigger than a struct containing all the same parameters. Consider all the fields rounded up to 4 or 8 bytes each. - FChE