Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753777AbXKSWW3 (ORCPT ); Mon, 19 Nov 2007 17:22:29 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756200AbXKSWWK (ORCPT ); Mon, 19 Nov 2007 17:22:10 -0500 Received: from wa-out-1112.google.com ([209.85.146.180]:57166 "EHLO wa-out-1112.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756074AbXKSWWH (ORCPT ); Mon, 19 Nov 2007 17:22:07 -0500 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=C6gzDU4zKHDC0MjeZrdKR2u1cloVZjM1CuXuY1CY3n3PF5fOXIuzfzjRoANu5gTcaHrWyz+n2AzxyFEuLqZ8CloP+QWeJHgytA5/KB88ueCIi69Q7IZMO9PTP7+g24fROzcCFrorVcT6gDSwWV31NP9vk89LlNuMOrmDqSRxRwg= Message-ID: Date: Mon, 19 Nov 2007 23:22:06 +0100 From: "Dmitry Adamushko" To: "Micah Dowty" Subject: Re: High priority tasks break SMP balancer? Cc: "Ingo Molnar" , "Christoph Lameter" , "Kyle Moffett" , "Cyrus Massoumi" , "LKML Kernel" , "Andrew Morton" , "Mike Galbraith" , "Paul Menage" , "Peter Williams" In-Reply-To: <20071119185116.GA28173@vmware.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <20071116024408.GA20322@vmware.com> <20071116060700.GD16273@elte.hu> <20071116221404.GC31527@vmware.com> <20071117010352.GA13666@vmware.com> <20071119185116.GA28173@vmware.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2211 Lines: 55 On 19/11/2007, Micah Dowty wrote: > [ ... ] > > micah@micah-64:~$ cat /proc/schedstat > version 14 > timestamp 4366722208 > cpu0 0 0 0 785868111 0 828020771 12621812 22703872 14096041 9504937730116 23603703739504 815398959 > domain0 03 8766051 8760123 95 14544377 6374 0 13 8760110 7369 6771 0 20835105 605 0 1 6770 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7026998 0 24826 > cpu1 0 0 0 783505304 0 860700497 26129151 41776419 34749418 10682815341474 1753703225423 834571346 > domain0 03 8685595 8677710 224 7037051 8127 0 76 8677634 7500 7320 7 471762 179 0 8 7312 ===> 0 0 0 0 0 0 0 0 <=== 1 0 1 0 0 0 0 0 0 8607831 0 36753 > micah@micah-64:~$ cat /proc/schedstat > version 14 > timestamp 4366723231 > cpu0 0 0 0 785868111 0 828021128 12621812 22703965 14096130 9509029340381 23603703795868 815399316 > domain0 03 8766051 8760123 95 14544377 6374 0 13 8760110 7374 6776 0 20835105 605 0 1 6775 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7027010 0 24826 > cpu1 0 0 0 783505304 0 860708521 26132612 41780521 34753508 10683770853336 1753758309667 834575909 > domain0 03 8685718 8677833 224 7037051 8127 0 76 8677757 7500 7320 7 471762 179 0 8 7312 ===> 0 0 0 0 0 0 0 0 <=== 1 0 1 0 0 0 0 0 0 8607835 0 36753 You seem to have a configuration with domains which don't have SD_BALANCE_NEWIDLE on (CONFIG_NUMA?) as there are no events (all zeros above) for CPU_NEWLY_IDLE. this one is being triggered whenever a cpu becomes idle (schedule() --> idle_balance() --> load_balance_newidle()). (this flag is a bit #1 == 2) cat /proc/sys/kernel/sched_domain/cpu0/domain0/flags ? it can be changed with "echo new_value > ..." does a change of it make any difference? moreover, /proc/sys/kernel/sched_domain/cpu1/domain0/newidle_idx seems to be responsible for a source of the load for calculating the busiest group. e.g. with newidle_idx == 0, the current load on the queue is used instead of cpu_load[]. > > Thanks, > --Micah > -- Best regards, Dmitry Adamushko - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/