From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756606AbZDTT4L (ORCPT ); Mon, 20 Apr 2009 15:56:11 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752595AbZDTTz5 (ORCPT ); Mon, 20 Apr 2009 15:55:57 -0400 Received: from hrndva-omtalb.mail.rr.com ([71.74.56.122]:49240 "EHLO hrndva-omtalb.mail.rr.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752085AbZDTTz4 (ORCPT ); Mon, 20 Apr 2009 15:55:56 -0400 Date: Mon, 20 Apr 2009 15:55:55 -0400 (EDT) From: Steven Rostedt X-X-Sender: rostedt@gandalf.stny.rr.com To: Ingo Molnar cc: =?ISO-8859-15?Q?Fr=E9d=E9ric_Weisbecker?= , linux-kernel@vger.kernel.org, Andrew Morton Subject: Re: [PATCH 0/4] [GIT PULL] tracing: recursion and compile fixes In-Reply-To: Message-ID: References: <20090420173819.957332585@goodmis.org> <20090420175533.GA22449@elte.hu> <20090420192917.GB25629@elte.hu> User-Agent: Alpine 2.00 (DEB 1167 2008-08-23) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 20 Apr 2009, Steven Rostedt wrote: > > On Mon, 20 Apr 2009, Ingo Molnar wrote: > > Running tests on all trace events: > > Testing all events: <4>Tracing recursion: HC[0]:SC[0]:NMI[0] > > Recursion is wrong :-/ > > > ------------[ cut here ]------------ > > WARNING: at kernel/trace/ring_buffer.c:1510 ring_buffer_lock_reserve+0xf7/0x1ce() > > Hardware name: To Be Filled By O.E.M. > > Pid: 1, comm: swapper Not tainted 2.6.30-rc2-tip-01538-ge8201b9-dirty #37740 > > Call Trace: > > [] warn_slowpath+0xcc/0x114 > > [] ? trace_hardirqs_on_caller+0x11d/0x179 > > [] ? mark_held_locks+0x60/0x95 > > [] ring_buffer_lock_reserve+0xf7/0x1ce > > [] trace_buffer_lock_reserve+0x34/0x7e > > [] ? do_IRQ+0xf/0xe8 > > [] trace_current_buffer_lock_reserve+0x36/0x4c > > [] function_test_events_call+0x7e/0xe2 > > [] ? ret_from_intr+0x0/0x16 > > [] ftrace_call+0x5/0x2b > > [] ? common_interrupt+0xe/0x13 > > [] ? do_IRQ+0x14/0xe8 > > The function tracer was called before we updated the hardirq_count :-/ > > I liked the fact that we can detect same level recursion. Perhaps I'll > just make a depth test instead. Have a max of 16? times we can recurse > (should be plenty) and then bug if we hit it? > The other option is not to trace any interrupt function that is called before "irq_enter". That would mean we would need to add notrace to do_IRQ and friends. -- Steve