From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756374Ab0EGPaj (ORCPT ); Fri, 7 May 2010 11:30:39 -0400 Received: from mail.openrapids.net ([64.15.138.104]:56615 "EHLO blackscsi.openrapids.net" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1755080Ab0EGPah (ORCPT ); Fri, 7 May 2010 11:30:37 -0400 Date: Fri, 7 May 2010 11:30:35 -0400 From: Mathieu Desnoyers To: Steven Rostedt Cc: Frederic Weisbecker , linux-kernel@vger.kernel.org, Ingo Molnar , Andrew Morton , Thomas Gleixner , Peter Zijlstra , Arnaldo Carvalho de Melo , Lai Jiangshan , Li Zefan , Masami Hiramatsu , Christoph Hellwig Subject: Re: [PATCH 2/9 - v2][RFC] tracing: Let tracepoints have data?passed to tracepoint callbacks Message-ID: <20100507153035.GA15267@Krystal> References: <7a22ce9b-4c74-4818-9521-7fbccd1b8b1d@email.android.com> <20100507143901.GA18408@Krystal> <1273244123.22438.140.camel@gandalf.stny.rr.com> <20100507150838.GB30356@Krystal> <1273245338.22438.168.camel@gandalf.stny.rr.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1273245338.22438.168.camel@gandalf.stny.rr.com> X-Editor: vi X-Info: http://www.efficios.com X-Operating-System: Linux/2.6.26-2-686 (i686) X-Uptime: 11:23:25 up 104 days, 18:00, 7 users, load average: 0.05, 0.09, 0.09 User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Steven Rostedt (rostedt@goodmis.org) wrote: > On Fri, 2010-05-07 at 11:08 -0400, Mathieu Desnoyers wrote: > > > > > Can you show me where the C standard says it is safe to do so ? > > > > > > No, but it seems safe in the kernel ;-) > > > > The use of "seems" here does not give me a warm feeling of safety. ;) > > Right, which is why I added the below. > > > > > > > > > But that said. There is another option that will conform to this, and > > > that is to add flags to registering tracepoints. I already wrote a patch > > > for this in trying to do some other work (that I threw away). > > > > > > > > > So here's the proposal. > > > > > > Change struct tracepoint_func to... > > > > > > struct tracepoint_func { > > > void *func; > > > void *data; > > > unsigned int flags; > > > }; > > > > > > > > > The flags is set when registered. If a function is registered with data, > > > then the flags field will be set. Then the calling of the function can > > > be: > > > > > > if ((it_func_ptr)->flags & TP_FL_DATA) > > > ((void(*)(proto, void *))(it_func)(args, __data); > > > else > > > ((void(*)(proto))(it_func)(args); > > > > > > This would comply with the C standard. > > > > This would also add a branch on the tracing fast path, which I would like to > > avoid. Why can't we simply change all prototypes to take an extra void *__data > > parameter instead ? > > I'm fine with making the data parameter mandatory with all tracers. Thus > the call back must require it. I would then move the data parameter from > the end to the beginning. > > So a tracepoint with proto, will have a callback: > > void callback(void *data, proto); > > I'm fine with forcing all callbacks to include a data parameter if you > are. This would also make the changes simpler. Yes, I am all for it. As for the extra type checking, it is basically just trying to force you to generate matching caller-callee prototypes in your CPP macros. The goal is really to check that the data parameter type match in both the caller and callee. I see that as a mean to make sure nobody is going to try to take shortcuts by playing with the callback types in the "undefined behavior" zone of the C standard in future TRACE_EVENT() modifications. Thanks, Mathieu > > -- Steve > > -- Mathieu Desnoyers Operating System Efficiency R&D Consultant EfficiOS Inc. http://www.efficios.com