From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Chinner Subject: Re: vm/fs meetup details Date: Fri, 6 Jul 2007 23:42:01 +1000 Message-ID: <20070706134201.GL31489@sgi.com> References: <20070705040138.GG32240@wotan.suse.de> <468D303E.4040902@redhat.com> <137D15F6-EABE-4EC1-A3AF-DAB0A22CF4E3@oracle.com> <20070705212757.GB12413810@sgi.com> <468D6569.6050606@redhat.com> <20070706022651.GG14215@wotan.suse.de> <20070706100110.GD12413810@sgi.com> <20070706102623.GA846@lazybastard.org> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Chinner , Nick Piggin , Rik van Riel , Zach Brown , Anton Altaparmakov , Suparna Bhattacharya , Christoph Hellwig , Hugh Dickins , Jared Hulbert , Chris Mason , "Martin J. Bligh" , Trond Myklebust , Neil Brown , Miklos Szeredi , Mingming Cao , Linux Memory Management List , linux-fsdevel@vger.kernel.org, Evgeniy Polyakov , Steven Whitehouse , Dave McCracken , Peter Zijlstra To: =?iso-8859-1?B?SsO2cm4=?= Engel Return-path: Received: from netops-testserver-4-out.sgi.com ([192.48.171.29]:42676 "EHLO relay.sgi.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752171AbXGFNn1 (ORCPT ); Fri, 6 Jul 2007 09:43:27 -0400 Content-Disposition: inline In-Reply-To: <20070706102623.GA846@lazybastard.org> Sender: linux-fsdevel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Fri, Jul 06, 2007 at 12:26:23PM +0200, J=C3=B6rn Engel wrote: > On Fri, 6 July 2007 20:01:10 +1000, David Chinner wrote: > > On Fri, Jul 06, 2007 at 04:26:51AM +0200, Nick Piggin wrote: > > >=20 > > > Keep in mind that the way to get the most out of this meeting is = for the > > > fs people to have topics of the form "we'd really like to do X, c= an we > > > get some help from the VM"? Or vice versa from vm people. > >=20 > > *nod* > >=20 > > But, surprisingly enough, the above work is relevent to this forum = because > > of two things: > >=20 > > - we've had to move to direct I/O and user space caching to work > > around deficiencies in kernel block device caching under memory > > pressure.... > >=20 > > - we've exploited techniques that XFS supports but the VM does not= =2E > > i.e. priority tagging of cached metadata so that less important > > metadata is tossed first (e.g. toss tree leaves before nodes and n= odes > > before roots) when under memory pressure. >=20 > And the latter is exactly what logfs needs as well. You certainly ha= ve me > interested. >=20 > I believe it applies to btrfs and any other cow-fs as well. The poin= t is > that higher levels get dirtied by writing lower layers. So perfect > behaviour for sync is to write leaves first, then nodes, then the roo= t. Any > other order will either cause sync not to sync or cause unnecessary w= rites > and cost performance. Hmmm - I guess you could use it for writeback ordering. I hadn't really thought about that. Doesn't seem a particularly efficient way of doing it, though. Why not just use multiple address spaces for this? i.e. one per level and flush in ascending order. Cheers, Dave. --=20 Dave Chinner Principal Engineer SGI Australian Software Group - 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