From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753837AbZEIQcp (ORCPT ); Sat, 9 May 2009 12:32:45 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754095AbZEIQc3 (ORCPT ); Sat, 9 May 2009 12:32:29 -0400 Received: from tomts10.bellnexxia.net ([209.226.175.54]:63317 "EHLO tomts10-srv.bellnexxia.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754849AbZEIQc2 (ORCPT ); Sat, 9 May 2009 12:32:28 -0400 X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: Ah8GAKdIBUpMQW1W/2dsb2JhbACBUM0Kg34F Date: Sat, 9 May 2009 12:32:16 -0400 From: Mathieu Desnoyers To: "Frank Ch. Eigler" Cc: Ingo Molnar , Jason Baron , Tom Zanussi , linux-kernel@vger.kernel.org, fweisbec@gmail.com, laijs@cn.fujitsu.com, rostedt@goodmis.org, peterz@infradead.org, jiayingz@google.com, mbligh@google.com, roland@redhat.com Subject: Re: [RFC] convert ftrace syscall tracer to TRACE_EVENT() Message-ID: <20090509163216.GA7548@Krystal> References: <20090508210347.GA3121@redhat.com> <20090509083737.GE3656@elte.hu> <20090509153629.GC29911@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline In-Reply-To: <20090509153629.GC29911@redhat.com> X-Editor: vi X-Info: http://krystal.dyndns.org:8080 X-Operating-System: Linux/2.6.21.3-grsec (i686) X-Uptime: 12:30:30 up 70 days, 12:56, 4 users, load average: 0.51, 0.52, 0.53 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 * Frank Ch. Eigler (fche@redhat.com) wrote: > Hi - > > On Sat, May 09, 2009 at 10:37:37AM +0200, Ingo Molnar wrote: > > [...] > > Firstly, it adds two new tracepoints to every system call. That is > > unnecessary - we already have the TIF flag based callbacks, and we > > can use the existing syscall attributes table to get to tracepoints > > - without slow down (or impacting) the fast path in any way. > > [...] > > However, as the TIF_SYSCALL_FTRACE and/or TIF_SYSCALL_TRACE flags are > tied to a single consumer, it would limit the usefulness of the > naturally multiple-client tracepoints, if they were made conditional > on them. Would you be interested in a proper reference-counting API? > Hrm, actually, looking at void start_ftrace_syscalls(void) { unsigned long flags; struct task_struct *g, *t; mutex_lock(&syscall_trace_lock); /* Don't enable the flag on the tasks twice */ if (++refcount != 1) goto unlock; shows me that there is in fact some refcounting done there. But I'll send the TIF_KERNEL_TRACE patchset anyway, given it supports much more architectures and has a more generic thread flag name. Mathieu > - FChE -- Mathieu Desnoyers OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68