From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753101Ab0EHAyo (ORCPT ); Fri, 7 May 2010 20:54:44 -0400 Received: from cdptpa-omtalb.mail.rr.com ([75.180.132.122]:64615 "EHLO cdptpa-omtalb.mail.rr.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752126Ab0EHAyn (ORCPT ); Fri, 7 May 2010 20:54:43 -0400 X-Authority-Analysis: v=1.1 cv=cbUT6B7wEcA8xSNNYfl1w4X1FOlyJIGpOcdyq2z1u20= c=1 sm=0 a=jbQAF9jGaHUA:10 a=IkcTkHD0fZMA:10 a=Ojb4KS2nAQbJgCTPCAnicg==:17 a=k-b6faDFtAYqCehWhMYA:9 a=vsr8YYNk1BBZ8GYIpIcmunC260wA:4 a=QEXdDO2ut3YA:10 a=Ojb4KS2nAQbJgCTPCAnicg==:117 X-Cloudmark-Score: 0 X-Originating-IP: 70.121.210.183 Subject: Re: unified page and buffer cache? From: Phillip Susi To: tytso@mit.edu Cc: Matthew Wilcox , linux-fsdevel@vger.kernel.org, Linux-kernel In-Reply-To: <20100508004617.GF18762@thunk.org> References: <4BCC7C05.8000803@cfl.rr.com> <4BE417CB.7000806@cfl.rr.com> <20100507135329.GA17259@parisc-linux.org> <4BE4359E.9020400@cfl.rr.com> <20100508004617.GF18762@thunk.org> Content-Type: text/plain; charset="UTF-8" Date: Fri, 07 May 2010 20:54:41 -0400 Message-ID: <1273280081.2444.5.camel@faldara> Mime-Version: 1.0 X-Mailer: Evolution 2.28.3 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Would it be possible to somehow keep the current buffer heads, but associate them with the inode such that readahead() on the directory would work? On Fri, 2010-05-07 at 20:46 -0400, tytso@mit.edu wrote: > Ext2 does use the page cache for directories. Ext3 and Ext4 access > directories via buffer heads because of the journaling requirement. > > In *theory* they could be modified to use the page cache, given that > we can do data journaling for files, and files live in the page cache > --- however, for cases where the PAGE_SIZE > FS_BLOCKSIZE, which will > happen if you are using 1k or 2k block filesystems, or on the Power > Architecture or on the Itanic where the page size is 16k, updates to > the directory will be much less efficient, since we journal changes to > data files on page granularity and not buffer granuality. > > Furthermore, someone would have to supply me with the patches; it's > pretty low on my priority list. And people on the Power and ia64 > platforms won't be happy.... > > - Ted