From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Fri, 22 Sep 2006 09:36:53 -0700 (PDT) From: Christoph Lameter Subject: Re: [PATCH] GFP_THISNODE for the slab allocator In-Reply-To: <20060922092631.ae24a777.pj@sgi.com> Message-ID: References: <20060914220011.2be9100a.akpm@osdl.org> <20060914234926.9b58fd77.pj@sgi.com> <20060915002325.bffe27d1.akpm@osdl.org> <20060915004402.88d462ff.pj@sgi.com> <20060915010622.0e3539d2.akpm@osdl.org> <20060917041707.28171868.pj@sgi.com> <20060917060358.ac16babf.pj@sgi.com> <20060917152723.5bb69b82.pj@sgi.com> <20060917192010.cc360ece.pj@sgi.com> <20060918093434.e66b8887.pj@sgi.com> <20060922092631.ae24a777.pj@sgi.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: Paul Jackson Cc: David Rientjes , akpm@osdl.org, linux-mm@kvack.org List-ID: On Fri, 22 Sep 2006, Paul Jackson wrote: > The topology.h header has: > > #define LOCAL_DISTANCE 10 > > though -no-one- uses it, why I don't know ... It is a SLIT table reference value. This is the distance to memory that is local to the processor and it is the lowest possible value. > This simple forcing of distances to 10 is probably good enough for your > setup, but if this gets serious, we'll need to handle multiple arch's, > and hybrid systems with both fake and real numa. That will take a bit > of work to get the SLIT table, node_distance and zonelist sorting > correct. Distance 10 is okay if the memory is on the node where the processor sits. -- 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