From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail144.messagelabs.com (mail144.messagelabs.com [216.82.254.51]) by kanga.kvack.org (Postfix) with ESMTP id 526C56B0047 for ; Sat, 20 Feb 2010 04:29:39 -0500 (EST) Date: Sat, 20 Feb 2010 09:29:21 +0000 From: Mel Gorman Subject: Re: [PATCH 03/12] mm: Share the anon_vma ref counts between KSM and page migration Message-ID: <20100220092921.GH1445@csn.ul.ie> References: <1266516162-14154-1-git-send-email-mel@csn.ul.ie> <1266516162-14154-4-git-send-email-mel@csn.ul.ie> <4B7F05BA.4080903@redhat.com> <20100219215826.GF1445@csn.ul.ie> <4B7F29CD.1050703@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <4B7F29CD.1050703@redhat.com> Sender: owner-linux-mm@kvack.org To: Rik van Riel Cc: Andrea Arcangeli , Christoph Lameter , Adam Litke , Avi Kivity , David Rientjes , KOSAKI Motohiro , linux-kernel@vger.kernel.org, linux-mm@kvack.org List-ID: On Fri, Feb 19, 2010 at 07:16:13PM -0500, Rik van Riel wrote: > On 02/19/2010 04:58 PM, Mel Gorman wrote: > >> external_refcount is about as good as I can think of to explain what's >> going on :/ > > Sounds "good" to me. Much better than giving the wrong > impression that this is the only refcount for the anon_vma. > Have renamed it so. If/when this all gets merged, I'll look into what's required to make this a "real" refcount rather than the existing locking mechanism. I'm very wary though because even with your anon_vma changes to avoid excessive sharing, a refcount in there that is used in all paths might become a hotly contended cache line. i.e. it might look nice, but it might be a performance hit. It needs to be done carefully and as a separate series. -- Mel Gorman Part-time Phd Student Linux Technology Center University of Limerick IBM Dublin Software Lab -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org