From: Andrew Morton Subject: Re: [EXT4 set 4][PATCH 1/5] i_version:64 bit inode version Date: Tue, 10 Jul 2007 18:22:37 -0700 Message-ID: <20070710182237.e2f88bf3.akpm@linux-foundation.org> References: <1183275424.4010.126.camel@localhost.localdomain> <20070710163038.ceb2ae94.akpm@linux-foundation.org> <1184105380.3759.65.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, linux-ext4@vger.kernel.org, nfsv4@linux-nfs.org To: cmm@us.ibm.com Return-path: In-Reply-To: <1184105380.3759.65.camel@localhost.localdomain> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Tue, 10 Jul 2007 18:09:40 -0400 Mingming Cao wrote: > On Tue, 2007-07-10 at 16:30 -0700, Andrew Morton wrote: > > On Sun, 01 Jul 2007 03:37:04 -0400 > > Mingming Cao wrote: > > > > > This patch converts the 32-bit i_version in the generic inode to a 64-bit > > > i_version field. > > > > > > > That's obvious from the patch. But what was the reason for making this > > (unrelated to ext4) change? > > > > The need is came from NFSv4 > > On Fri, 2007-05-25 at 18:25 +0200, Jean noel Cordenner wrote: > > Hi, > > > > This is an update of the i_version patch. > > The i_version field is a 64bit counter that is set on every inode > > creation and that is incremented every time the inode data is modified > > (similarly to the "ctime" time-stamp). > > The aim is to fulfill a NFSv4 requirement for rfc3530: > > "5.5. Mandatory Attributes - Definitions > > Name # DataType Access Description > > ___________________________________________________________________ > > change 3 uint64 READ A value created by the > > server that the client can use to determine if file > > data, directory contents or attributes of the object > > have been modified. The servermay return the object's > > time_metadata attribute for this attribute's value but > > only if the filesystem object can not be updated more > > frequently than the resolution of time_metadata. > > " > > > > > Please update the changelog for this. > > > > Is above description clear to you? > Yes, thanks. It doesn't actually tell us why we want to implement this attribute and it doesn't tell us what the implications of failing to do so are, but I guess we can take that on trust from the NFS guys. But I suspect the ext4 implementation doesn't actually do this. afaict we won't update i_version for file overwrites (especially if s_time_gran can indeed be 1,000,000,000) and of course for MAP_SHARED modifications. What would be the implications of this? And how does the NFS server know that the filesystem implements i_version? Will a zero-value of i_version have special significance, telling the server to not send this attribute, perhaps?