From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [PATCH 4/4] fs: take the ACL checks to common code Date: Sat, 23 Jul 2011 18:00:00 +0200 Message-ID: <20110723160000.GA11196@lst.de> References: <20110723153621.GA10832@lst.de> <20110723153731.GD10887@lst.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Christoph Hellwig , Al Viro , linux-fsdevel To: Linus Torvalds Return-path: Received: from verein.lst.de ([213.95.11.211]:44778 "EHLO newverein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750708Ab1GWQAA (ORCPT ); Sat, 23 Jul 2011 12:00:00 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Sat, Jul 23, 2011 at 08:56:12AM -0700, Linus Torvalds wrote: > On Sat, Jul 23, 2011 at 8:37 AM, Christoph Hellwig wrote: > > > > Replace the ->check_acl method with a ->get_acl method that simply reads an > > ACL from disk after having a cache miss. ?This means we can replace the ACL > > checking boilerplate code with a single implementation in namei.c. > > Hmm. If we just had a rule for locking (i_mutex?), we could make the > cache update be in namei.c too.. Yes, and if we add a ->set_acl we can take most of the existing boilerplate code completely into posix_acl.c. I'll see if I can do something like that for v3.2. > Things that want to use timeouts etc to make for more complicated > permissions rules than the straightforward acl cache have to use > ->permission anyway, so it would make sense to make the ->get_acl > method as trivial as possible for filesystems. At this point no one does that anyway, the only real complications are the cluster filesystems having remote cache invalidations.