From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754668Ab0ETSgQ (ORCPT ); Thu, 20 May 2010 14:36:16 -0400 Received: from rcsinet10.oracle.com ([148.87.113.121]:41512 "EHLO rcsinet10.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751343Ab0ETSgP (ORCPT ); Thu, 20 May 2010 14:36:15 -0400 Message-ID: <4BF580ED.3080207@oracle.com> Date: Thu, 20 May 2010 11:35:25 -0700 From: Randy Dunlap Organization: Oracle Linux Engineering User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.5) Gecko/20091209 Fedora/3.0-3.fc11 Thunderbird/3.0 MIME-Version: 1.0 To: "Frank A. Kingswood" CC: linux-kernel@vger.kernel.org Subject: Re: NR_CPUS configuration option wrong References: <20100520103851.8e051a3f.randy.dunlap@oracle.com> <4BF58054.8060601@kingswood-consulting.co.uk> In-Reply-To: <4BF58054.8060601@kingswood-consulting.co.uk> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Auth-Type: Internal IP X-Source-IP: acsinet15.oracle.com [141.146.126.227] X-CT-RefId: str=0001.0A090201.4BF5811A.001D:SCFMA922111,ss=1,fgs=0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 05/20/10 11:32, Frank A. Kingswood wrote: > Randy Dunlap wrote: >> On Thu, 20 May 2010 18:34:32 +0100 Frank A. Kingswood wrote: >> >>> Hi! >>> >>> The configuration MAXSMP help text indicates it is used to enable >>> configuration of number of CPUs. However, NR_CPUS depends on !MAXSMP. >>> >>> arch/x86/Kconfig: >>> 758 config NR_CPUS >>> 759 int "Maximum number of CPUs" if SMP && !MAXSMP >> >> That line depends on !MAXSMP. Then below there, we find: >> >> default "4096" if MAXSMP > > Hah, weird. > I can see now it is possible to parse the MAXSMP help this way, but it > was not obvious. Ambiguity is fairly easy. :( > In menuconfig you see two questions: > > [ ] Configure Maximum number of SMP Processors and NUMA Nodes > [8] Maximum number of CPUs > > and setting the first one YES makes the second one go away. > > Perhaps restating the help as "Set number of SMP processors and NUMA > nodes to the maximum that the kernel supports" might help. Yes, that sounds better. Want to submit a patch for that? -- ~Randy *** Remember to use Documentation/SubmitChecklist when testing your code ***