From mboxrd@z Thu Jan 1 00:00:00 1970 From: Juergen Gross Subject: Re: [PATCH 4 of 4] Support new xl command cpupool-numa-split Date: Thu, 09 Dec 2010 10:45:11 +0100 Message-ID: <4D00A527.7070505@ts.fujitsu.com> References: <1291806978.13966.4529.camel@zakaz.uk.xensource.com> <4CFF7812.4050506@ts.fujitsu.com> <1291813922.13966.4553.camel@zakaz.uk.xensource.com> <4CFF8569.4010104@ts.fujitsu.com> <1291815484.13966.4555.camel@zakaz.uk.xensource.com> <4CFF8B04.8080305@ts.fujitsu.com> <1291817544.13966.4557.camel@zakaz.uk.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1291817544.13966.4557.camel@zakaz.uk.xensource.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Ian Campbell Cc: "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org On 12/08/10 15:12, Ian Campbell wrote: > On Wed, 2010-12-08 at 13:41 +0000, Juergen Gross wrote: >> On 12/08/10 14:38, Ian Campbell wrote: >>> On Wed, 2010-12-08 at 13:17 +0000, Juergen Gross wrote: >>>> On 12/08/10 14:12, Ian Campbell wrote: >>>>> On Wed, 2010-12-08 at 12:20 +0000, Juergen Gross wrote: >>>>>> On 12/08/10 12:16, Ian Campbell wrote: >>>>>>> Can this loop be merged with the preceding loop, with the body being the >>>>>>> else case of the if? >>>>>> >>>>>> No. I have to add new cpus first to avoid a cpupool without cpus in between. >>>>> >>>>> ok. >>>>> >>>>> I was thinking that because this function only gets here if there is a >>>>> single pool that all CPUs must be in that pool -- but that's not >>>>> actually true is it? Even if that were the common case there's nothing >>>>> to enforce that. >>>> >>>> Perhaps I should add a comment to avoid a problem later... >>> >>> That would certainly help. >>> >>> The alternative would be to bail out if all cpus are not associated with >>> Pool-0, not just when there are> 1 pools. That would be consistent with >>> the function only acting on the default configuration. >> >> I suspect NUMA systems are subject to cpu hot plug... > > And hotplugged CPUs don't automatically go into Pool-0? I just checked it: they ARE added to Pool-0. This rises another problem: it would be nice to adjust the NUMA splitting after a hot-plug of a cpu... An "update" option to xl cpupool-numa-split would be appropriate, I think. It should move cpus from Pool 0 to the correct node pools and create new node pools if necessary. I'll prepare another patch later... Juergen -- Juergen Gross Principal Developer Operating Systems TSP ES&S SWE OS6 Telephone: +49 (0) 89 3222 2967 Fujitsu Technology Solutions e-mail: juergen.gross@ts.fujitsu.com Domagkstr. 28 Internet: ts.fujitsu.com D-80807 Muenchen Company details: ts.fujitsu.com/imprint.html