From: Nick Piggin <npiggin@suse.de>
Subject: Re: [PATCH, RFC] add a vfs_fsync helper
Date: Mon, 22 Dec 2008 14:00:55 +0100
Message-ID: <20081222130055.GF13406@wotan.suse.de>
References: <20081222090357.GA27399@lst.de> <20081222123534.GE13406@wotan.suse.de> <20081222125622.GA3019@lst.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: viro@zeniv.linux.org.uk, dbrownell@users.sourceforge.net,
	agl@us.ibm.com, ebiederm@xmission.com,
	linux-fsdevel@vger.kernel.org, ecryptfs-devel@lists.launchpad.net,
	jaharkes@cs.cmu.edu
To: Christoph Hellwig <hch@lst.de>
Content-Disposition: inline
In-Reply-To: <20081222125622.GA3019@lst.de>
Sender: linux-fsdevel-owner@vger.kernel.org

On Mon, Dec 22, 2008 at 01:56:22PM +0100, Christoph Hellwig wrote:
> On Mon, Dec 22, 2008 at 01:35:34PM +0100, Nick Piggin wrote:
> > Seems like an improvement.. I've wanted to know, though, what do
> > we need i_mutex for? Is it just convention, or is there some good
> > reason to have it in generic code?
> 
> At least XFS doesn't need it.  Same for the filemap_fdatawrite /
> filemap_fdatawait which at least for filesystems that want to provide
> integrity guarantees is in the wrong place.
> 
> This patch is a first one out my work to refactor fsync, and I'm trying
> to feed it to mainline in pieces.  The next one will be to make sure

OK, I gues that's fair enough. The generic pagecache layer as you know
shouldn't require i_mutex... so long as you have a plan in mind to
decouple that requirement from generic code I'm happy with it.


> nfsd always has a struct file available when calling fsync, but I need
> to do some extensive benchmarking.  After that we can change the
> fsync prototype to drop the dentry paramters, and move the
> filemap_fdatawrite / filemap_fdatawait aswell as the i_mutex locking
> into the actual methods.

They all sound like good steps I think. Well I have no really helpful
comments for your RFC except that I think it is a good idea. Probably
some filesystem-side people should have a look at it...