From: Li Zefan Subject: Re: [PATCH] vfs: Add a trace point in the mark_inode_dirty function Date: Mon, 16 Nov 2009 08:56:21 +0800 Message-ID: <4B00A335.9070900@cn.fujitsu.com> References: <20091025225342.007138f5@infradead.org> <4AFA2293.5080803@cn.fujitsu.com> <20091115110043.6cafc221@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, Christoph Hellwig , Al Viro , mingo@elte.hu, Frederic Weisbecker , auke-jan.h.kok@intel.com To: Arjan van de Ven Return-path: Received: from cn.fujitsu.com ([222.73.24.84]:58192 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1752152AbZKPA5L (ORCPT ); Sun, 15 Nov 2009 19:57:11 -0500 In-Reply-To: <20091115110043.6cafc221@infradead.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Arjan van de Ven wrote: > On Wed, 11 Nov 2009 10:33:55 +0800 > Li Zefan wrote: >> This will leave __entry->comm, __entry->file and __entry->dev >> uninitialized in the "else" case.. > > I would expect the struct to be zeroed before.... is it? > (the TP_ stuff is so obscure I find it hard to figure out where it ends > up) > No, it won't. Ring buffer won't zero the buffer before returning it, and TP_ stuff won't zero the buffer after getting it. See static void ftrace_raw_event_##call(proto) in include/event/ftrace.h