From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx102.postini.com [74.125.245.102]) by kanga.kvack.org (Postfix) with SMTP id B76106B0031 for ; Wed, 31 Jul 2013 05:13:34 -0400 (EDT) Date: Wed, 31 Jul 2013 10:13:30 +0100 From: Mel Gorman Subject: Re: [PATCH 13/18] mm: numa: Scan pages with elevated page_mapcount Message-ID: <20130731091330.GK2296@suse.de> References: <1373901620-2021-1-git-send-email-mgorman@suse.de> <1373901620-2021-14-git-send-email-mgorman@suse.de> <51E62A0E.1070102@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <51E62A0E.1070102@gmail.com> Sender: owner-linux-mm@kvack.org List-ID: To: Sam Ben Cc: Peter Zijlstra , Srikar Dronamraju , Ingo Molnar , Andrea Arcangeli , Johannes Weiner , Linux-MM , LKML On Wed, Jul 17, 2013 at 01:22:22PM +0800, Sam Ben wrote: > On 07/15/2013 11:20 PM, Mel Gorman wrote: > >Currently automatic NUMA balancing is unable to distinguish between false > >shared versus private pages except by ignoring pages with an elevated > > What's the meaning of false shared? > Two tasks may be operating on a shared buffer that is not aligned. It is expected that will at least cache align to avoid CPU cache line bouncing but the buffers are not necessarily page aligned. A page is the minimum granularity we can track NUMA hinting faults so two tasks sharing such a page will appear to be sharing data when in fact they are not. -- Mel Gorman SUSE Labs -- 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