From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail191.messagelabs.com (mail191.messagelabs.com [216.82.242.19]) by kanga.kvack.org (Postfix) with SMTP id 4AF336B01B1 for ; Fri, 21 May 2010 14:07:07 -0400 (EDT) Date: Fri, 21 May 2010 13:03:51 -0500 (CDT) From: Christoph Lameter Subject: RE: [PATCH] slub: move kmem_cache_node into it's own cacheline In-Reply-To: <6E3BC7F7C9A4BF4286DD4C043110F30B0B5969081B@shsmsx502.ccr.corp.intel.com> Message-ID: References: <20100520234714.6633.75614.stgit@gitlad.jf.intel.com> <6E3BC7F7C9A4BF4286DD4C043110F30B0B5969081B@shsmsx502.ccr.corp.intel.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org To: "Shi, Alex" Cc: Pekka Enberg , "Duyck, Alexander H" , "linux-mm@kvack.org" , Zhang Yanmin , "Chen, Tim C" List-ID: Yes right. The cacheline that also contains local_node is dirtied by the locking in the SMP case and will evict the cacheline used to lookup the per cpu vector and other important information. The per cpu patches aggravated that problem by making more use of the fields that are evicted with the cacheline. Acked-by: Christoph Lameter -- 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