From mboxrd@z Thu Jan  1 00:00:00 1970
From: Arnd Bergmann <arnd@arndb.de>
Subject: Re: Proposal to improve filesystem/block snapshot interaction
Date: Tue, 30 Oct 2007 08:43:07 +0100
Message-ID: <200710300843.08940.arnd@arndb.de>
References: <20070927063113.GD2989@sgi.com> <18214.45062.754722.885137@notabene.brown> <20071030051250.GE28875@sgi.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Neil Brown <neilb@suse.de>,
	Linux Filesystem Mailing List <linux-fsdevel@vger.kernel.org>,
	David Chinner <dgc@melbourne.sgi.com>,
	Donald Douwsma <donaldd@melbourne.sgi.com>,
	Christoph Hellwig <hch@infradead.org>,
	Roger Strassburg <rls@sgi.com>, Mark Goodwin <markgw@sgi.com>,
	Brett Jon Grandbois <brettg@melbourne.sgi.com>
To: Greg Banks <gnb@sgi.com>
Return-path: <linux-fsdevel-owner@vger.kernel.org>
Received: from moutng.kundenserver.de ([212.227.126.171]:50897 "EHLO
	moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1753056AbXJ3Hnn convert rfc822-to-8bit (ORCPT
	<rfc822;linux-fsdevel@vger.kernel.org>);
	Tue, 30 Oct 2007 03:43:43 -0400
In-Reply-To: <20071030051250.GE28875@sgi.com>
Content-Disposition: inline
Sender: linux-fsdevel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

On Tuesday 30 October 2007, Greg Banks wrote:
>=20
> > If the allocation unit of the storage device (e.g. a few MB) does n=
ot
> > match the allocation unit of the filesystem (e.g. a few KB) then fo=
r
> > this to be useful either the storage device must start recording ti=
ny
> > allocations, or the filesystem should re-release areas as they grow=
=2E
> > i.e. when releasing a range of a device, look in the filesystem's u=
sage
> > records for the largest surrounding free space, and release all of =
that.
>=20
> Good point. =A0I was planning on ignoring this problem :-/ Given that
> current snapshot implementations waste *all* the blocks in deleted
> files, it would be an improvement to scavenge the blocks in large
> extents. =A0This is especially true for XFS which goes to some effort
> to achieve large linear extents.
>=20

Ah, this is an important difference to my idea about an erase operation
on the block device. For erase to be meaningful, you need to know the
erase block size at the file system or user space, so it would be
encoded in the struct block_device, and the user has to issue erase
requests at erase block grenularity.

	Arnd <><
-
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel=
" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html