From mboxrd@z Thu Jan 1 00:00:00 1970 From: Neil Brown Subject: Re: Thoughts about cache consistency and directories in particular. Date: Sat, 21 Feb 2009 08:06:20 +1100 Message-ID: <18847.6988.418374.839185@notabene.brown> References: <18846.4842.625445.980681@notabene.brown> <1235156207.31066.21.camel@heimdal.trondhjem.org> <18847.2578.480148.216735@notabene.brown> <1235161966.31066.24.camel@heimdal.trondhjem.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "J. Bruce Fields" , linux-nfs@vger.kernel.org To: Trond Myklebust Return-path: Received: from cantor2.suse.de ([195.135.220.15]:40890 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752218AbZBTVG2 (ORCPT ); Fri, 20 Feb 2009 16:06:28 -0500 In-Reply-To: message from Trond Myklebust on Friday February 20 Sender: linux-nfs-owner@vger.kernel.org List-ID: On Friday February 20, trond.myklebust@fys.uio.no wrote: > On Sat, 2009-02-21 at 06:52 +1100, Neil Brown wrote: > > On Friday February 20, trond.myklebust@fys.uio.no wrote: > > > It should work fine. The ctime tracks the mtime in all cases except when > > > you setacl, setfattr, chown, chgrp, chmod, or touch the directory. Those > > > should be very rare operations for pretty much any workload... > > > > > > > Does that mean you'll take the patch ?? > > Not as it stands. You really want to be calling > nfs_force_lookup_revalidate() instead of invalidating the directory > contents. Surely it is 'aswell' rather than 'instead' ?? I need invalid |= NFS_INO_INVALID_DATA; to flush the readdir cache, and nfs_force_lookup_revalidate(inode) to flush the lookup cache ?? I'll respin the patches later today. Thanks, NeilBrown