From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ipmail06.adl2.internode.on.net ([150.101.137.129]:36815 "EHLO ipmail06.adl2.internode.on.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1031527AbeBNWn2 (ORCPT ); Wed, 14 Feb 2018 17:43:28 -0500 Date: Thu, 15 Feb 2018 09:43:26 +1100 From: Dave Chinner Subject: Re: [PATCH] xfs: rewrite the fdatasync optimization Message-ID: <20180214224326.GL7000@dastard> References: <20180205073933.17065-1-hch@lst.de> <20180205221726.c2hdvaairki27gxf@destitution> <20180206072356.GB1649@lst.de> <20180208231725.GH20266@dastard> <20180213150402.GA15030@lst.de> <20180213232640.GG6778@dastard> <20180214155343.GA4472@lst.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180214155343.GA4472@lst.de> Sender: linux-xfs-owner@vger.kernel.org List-ID: List-Id: xfs To: Christoph Hellwig Cc: linux-xfs@vger.kernel.org, Jeff Layton On Wed, Feb 14, 2018 at 04:53:43PM +0100, Christoph Hellwig wrote: > Ok, for the case where the same file also accessed throught NFS is > locally written and fdatasync we'll run into an issue. > > But we already have the exact same issue on NFSv3 which always uses > ctime and never looks at iversion. The NFSv3 server won't be querying the iversion field, so on upstream kernels it will never change or need to be logged on local fs accesses. Hence I don't think we need to care about NFSv3 at all here.... Cheers, Dave. -- Dave Chinner david@fromorbit.com