From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: with ECARTIS (v1.0.0; list xfs); Wed, 07 Mar 2007 09:19:14 -0800 (PST) Received: from mail.lst.de (verein.lst.de [213.95.11.210]) by oss.sgi.com (8.12.10/8.12.10/SuSE Linux 0.7) with ESMTP id l27HJ46p023039 for ; Wed, 7 Mar 2007 09:19:06 -0800 Date: Wed, 7 Mar 2007 18:15:47 +0100 From: Christoph Hellwig Subject: Re: [PATCH 2/2] xfs: stop using kmalloc in xfs_buf_get_noaddr Message-ID: <20070307171547.GA22641@lst.de> References: <20070307101324.GC30587@lst.de> <45EEF0B5.40905@agami.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <45EEF0B5.40905@agami.com> Sender: xfs-bounce@oss.sgi.com Errors-to: xfs-bounce@oss.sgi.com List-Id: xfs To: Michael Nishimoto Cc: Christoph Hellwig , xfs@oss.sgi.com, ecashin@coraid.com, akpm@osdl.org, linux-kernel@vger.kernel.org On Wed, Mar 07, 2007 at 09:04:53AM -0800, Michael Nishimoto wrote: > Incore log buffers are not always a power of two of the page size. > In particular, when xfs is running over software raid devices, the > log buffers are allocated to match the size of a stripe. > > However, they are always a multiple of PAGE_SIZE, so we are still safe. It's not actually about beeing safe - any allocation is still safe with this patch. The issue is just that we waste memory because we round up allocations to the next page size. The power of two bit is actually wrong in this mail, it was about another optimization I have that needs some more testing first.