From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Kent Subject: Re: [PATCH 1/2] vfs: make real_lookup do dentry revalidation with i_mutex held Date: Thu, 25 Jun 2009 15:21:13 +0800 Message-ID: <4A432569.7050404@themaw.net> References: <1237493790-5665-1-git-send-email-sage@newdream.net> <49C85E3E.7030505@themaw.net> <49C85F0A.9030409@themaw.net> <49C9CA5A.5080809@themaw.net> <49CA579B.5080307@themaw.net> <49CAE3DB.3030909@themaw.net> <1238063926.23853.15.camel@zeus.themaw.net> <49D56DD5.80707@themaw.net> <4A4023CF.7040203@themaw.net> <4A4040AA.5030108@zytor.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Sage Weil , linux-fsdevel , Christoph Hellwig , akpm@linux-foundation.org, Al Viro , Yehuda Sadeh , Jon Masters To: "H. Peter Anvin" Return-path: Received: from outbound.icp-qv1-irony-out3.iinet.net.au ([203.59.1.148]:4424 "EHLO outbound.icp-qv1-irony-out3.iinet.net.au" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751877AbZFYHVN (ORCPT ); Thu, 25 Jun 2009 03:21:13 -0400 In-Reply-To: <4A4040AA.5030108@zytor.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: H. Peter Anvin wrote: > Ian Kent wrote: >> Sorry, I haven't had time to do more on this. >> There is also the issue of what to do about removing the autofs module >> and renaming autofs4 to autofs, as this will break the autofs module. >> > > The autofs module is pretty much historic at this point. I say just > nuke it. But while I sort out the details of changing autofs4 to handle the change to path lookup locking there are a couple of things work out for the rename of autofs4 to autofs. It's true that removing autofs shouldn't be a big deal but existing users of autofs4 will have a problem. Particularly if people run a newer kernel on and older distribution. We probably have to accept that people doing insmod with a specific path are going to have a problem and they shouldn't use that method anyway. But, AFAICS using a MODULE_ALIAS() in the kernel module will allow for the name change but doesn't seem to take account of the directory name change. Is that correct? There are those that may have an alias in the module-init-tools configuration as well. But we need get rid of that practice as well as, for a long time now, the alias hasn't been needed. Anyone have any thoughts how we might better handle these difficulties? Ian