From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750712AbVHOEJE (ORCPT ); Mon, 15 Aug 2005 00:09:04 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750812AbVHOEJD (ORCPT ); Mon, 15 Aug 2005 00:09:03 -0400 Received: from 216-239-45-4.google.com ([216.239.45.4]:43326 "EHLO 216-239-45-4.google.com") by vger.kernel.org with ESMTP id S1750712AbVHOEJD (ORCPT ); Mon, 15 Aug 2005 00:09:03 -0400 Message-ID: <430014EA.4030404@google.com> Date: Sun, 14 Aug 2005 21:07:06 -0700 From: Mike Waychison User-Agent: Mozilla Thunderbird 1.0.6 (Macintosh/20050716) X-Accept-Language: en-us, en MIME-Version: 1.0 To: coywolf@sosdg.org CC: akpm@osdl.org, bunk@stusta.de, linux-kernel@vger.kernel.org Subject: Re: [patch] unexport __mntput() References: <20050815015357.GA16778@everest.sosdg.org> In-Reply-To: <20050815015357.GA16778@everest.sosdg.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Coywolf Qi Hunt wrote: > Hello, > > Unexport __mntput() was talked about two months ago. http://lkml.org/lkml/2005/6/9/69 > Modules should not call __mntput() directly. If autofs or nfsd does that, it's > being wrong. I think you missed the point in the last discussion. __mntput is called from mntput(), which autofs and nfsd call. Their use is correct given what they do: Autofs 3 and 4 use it for walking the vfsmount tree and determining if/when a mountpoint is ready to expire. Nfsd uses it to serve up nfs exports that don't cross mountpoints (or do, if "crossmnt" is specified in /etc/exports. Thanks, Mike Waychison > > Coywolf > > > Signed-off-by: Coywolf Qi Hunt > --- 2.6.13-rc6/fs/namespace.c~unexport-__mntput 2005-08-12 08:21:22.000000000 -0500 > +++ 2.6.13-rc6/fs/namespace.c 2005-08-14 20:32:01.000000000 -0500 > @@ -180,8 +180,6 @@ > deactivate_super(sb); > } > > -EXPORT_SYMBOL(__mntput); > - > /* iterator */ > static void *m_start(struct seq_file *m, loff_t *pos) > {