From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: [PATCH 11/35] lmb: Add find_memory_core_early() Date: Sat, 15 May 2010 08:34:48 +1000 Message-ID: <1273876488.21352.643.camel@pasglop> References: <1273804238.21352.389.camel@pasglop> <4BECF00D.6080800@oracle.com> <1273825187.21352.590.camel@pasglop> <20100514.013012.13730149.davem@davemloft.net> <4BED7DFE.2090108@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <4BED7DFE.2090108@oracle.com> Sender: linux-kernel-owner@vger.kernel.org To: Yinghai Lu Cc: David Miller , mingo@elte.hu, tglx@linutronix.de, hpa@zytor.com, akpm@linux-foundation.org, torvalds@linux-foundation.org, hannes@cmpxchg.org, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org List-Id: linux-arch.vger.kernel.org On Fri, 2010-05-14 at 09:44 -0700, Yinghai Lu wrote: > On 05/14/2010 01:30 AM, David Miller wrote: > > From: Benjamin Herrenschmidt > > Date: Fri, 14 May 2010 18:19:47 +1000 > > > > > >> OK, I'll have to let Davem deal with the fine point of the sparc bits, > >> but I think basically sparc has CONFIG_ARCH_POPULATES_NODE_MAP set, but > >> the way it's NUMA affinity works, the early_node_map[] is crap, you > >> cannot rely on the ranges in there. > >> > > > > > Right, we can't use early_node_map[] on sparc, because the NUMA > > mappings are far too granular to use that kind of representation. > > > good to know. > > early_node_map[] doesn't have enough slot? I think that's the case. sparc doesn't override MAX_ACTIVE_REGIONS, which means you get the default which is 256 or 50 per node depending on MAX_NUMNODES. My understanding is that this may not be enough. Dave, what does it look like in practice tho ? Cheers, Ben. From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org ([63.228.1.57]:50858 "EHLO gate.crashing.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757293Ab0ENWgM (ORCPT ); Fri, 14 May 2010 18:36:12 -0400 Subject: Re: [PATCH 11/35] lmb: Add find_memory_core_early() From: Benjamin Herrenschmidt In-Reply-To: <4BED7DFE.2090108@oracle.com> References: <1273804238.21352.389.camel@pasglop> <4BECF00D.6080800@oracle.com> <1273825187.21352.590.camel@pasglop> <20100514.013012.13730149.davem@davemloft.net> <4BED7DFE.2090108@oracle.com> Content-Type: text/plain; charset="UTF-8" Date: Sat, 15 May 2010 08:34:48 +1000 Message-ID: <1273876488.21352.643.camel@pasglop> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: Yinghai Lu Cc: David Miller , mingo@elte.hu, tglx@linutronix.de, hpa@zytor.com, akpm@linux-foundation.org, torvalds@linux-foundation.org, hannes@cmpxchg.org, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org Message-ID: <20100514223448.mseRQUhQZ3p80kVxC1s41tCRkh0ZAiDDTyDEf0Ft95k@z> On Fri, 2010-05-14 at 09:44 -0700, Yinghai Lu wrote: > On 05/14/2010 01:30 AM, David Miller wrote: > > From: Benjamin Herrenschmidt > > Date: Fri, 14 May 2010 18:19:47 +1000 > > > > > >> OK, I'll have to let Davem deal with the fine point of the sparc bits, > >> but I think basically sparc has CONFIG_ARCH_POPULATES_NODE_MAP set, but > >> the way it's NUMA affinity works, the early_node_map[] is crap, you > >> cannot rely on the ranges in there. > >> > > > > > Right, we can't use early_node_map[] on sparc, because the NUMA > > mappings are far too granular to use that kind of representation. > > > good to know. > > early_node_map[] doesn't have enough slot? I think that's the case. sparc doesn't override MAX_ACTIVE_REGIONS, which means you get the default which is 256 or 50 per node depending on MAX_NUMNODES. My understanding is that this may not be enough. Dave, what does it look like in practice tho ? Cheers, Ben.