Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754252AbaJUC6U (ORCPT ); Mon, 20 Oct 2014 22:58:20 -0400 Received: from fgwmail6.fujitsu.co.jp ([192.51.44.36]:60095 "EHLO fgwmail6.fujitsu.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754158AbaJUC6S (ORCPT ); Mon, 20 Oct 2014 22:58:18 -0400 X-SecurityPolicyCheck: OK by SHieldMailChecker v2.0.1 X-SHieldMailCheckerPolicyVersion: FJ-ISEC-20120718-3 Message-ID: <5445CB94.1090302@jp.fujitsu.com> Date: Tue, 21 Oct 2014 11:57:24 +0900 From: Yasuaki Ishimatsu User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:24.0) Gecko/20100101 Thunderbird/24.6.0 MIME-Version: 1.0 To: Dan Streetman , Neil Zhang CC: Greg KH , "linux-kernel@vger.kernel.org" Subject: Re: [PATCH] drivers: base: update cpu offline info when do hotplug References: <1413775748-22478-1-git-send-email-zhangwm@marvell.com> <20141020044340.GA9589@kroah.com> <9034CBD80F070943B59700D7F8149ED9024EB81CAF@SC-VEXCH4.marvell.com> <20141020064826.GA4064@kroah.com> <9034CBD80F070943B59700D7F8149ED9024EB81CC6@SC-VEXCH4.marvell.com> In-Reply-To: Content-Type: text/plain; charset="UTF-8"; format=flowed Content-Transfer-Encoding: 8bit X-SecurityPolicyCheck-GC: OK by FENCE-Mail Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Neil and Dan, (2014/10/21 2:02), Dan Streetman wrote: > On Mon, Oct 20, 2014 at 3:40 AM, Neil Zhang wrote: >> Greg, >> >> >> -----Original Message----- >> From: Greg KH [mailto:gregkh@linuxfoundation.org] >> Sent: 2014年10月20日 14:48 >> To: Neil Zhang >> Cc: linux-kernel@vger.kernel.org >> Subject: Re: [PATCH] drivers: base: update cpu offline info when do hotplug >> >> On Sun, Oct 19, 2014 at 11:39:23PM -0700, Neil Zhang wrote: >>>> How much noise is this going to cause on a big/little system that >>>> constantly hot unplug/plugs processors all of the time? >>> >>> Can you explain more what kind of noise will be introduced on a big/little system? >> >> Have you tested this on such a machine? >> >> I didn't have such kind of machine on hand. >> Can anyone has such machine to verify it? >> Thanks! > > I tested this on a ppc PowerVM system, using dlpar operations to > remove/add cpus. > > Without this patch the cpu online nodes get out of sync with the main > online node (and the actual state of the cpus), because they aren't > updated as the cpus are brought up/down: > > [root@br10p02 cpu]$ pwd > /sys/devices/system/cpu > [root@br10p02 cpu]$ cat online > 0-39 > [root@br10p02 cpu]$ for n in {0..47} ; do test $( cat cpu$n/online ) > -eq 1 && echo -n "$n " ; done ; echo "" > 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 > 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 How is the issue reproduced. Here is a result on my x86 box with linux-3.18-rc1. - before offline CPU # cd /sys/devices/system/cpu/ # cat online 0-59 # for n in {0..59} ; do test $( cat cpu$n/online ) -eq 1 && echo -n "$n " ; done ; echo "" 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 - after offline CPU{1..59} # for n in {1..59} ; do echo 0 > cpu$n/online; done # cat online 0 # for n in {0..59} ; do test $( cat cpu$n/online ) -eq 1 && echo -n "$n " ; done ; echo "" 0 It seems that dev->offline is set to correct valute. Thanks, Yasuaki Ishimatsu > > > While with the patch, the cpu online nodes are kept up to date as the > cpus are brought up/down: > > [root@br10p02 cpu]$ pwd > /sys/devices/system/cpu > [root@br10p02 cpu]$ cat online > 0-39 > [root@br10p02 cpu]$ for n in {0..47} ; do test $( cat cpu$n/online ) > -eq 1 && echo -n "$n " ; done ; echo "" > 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 > 27 28 29 30 31 32 33 34 35 36 37 38 39 > > > Feel free to add > > Tested-by: Dan Streetman > >> >>> As I know IKS on arm will use cpu_suspend way to power down a core. >> >> Are you sure that it also doesn't use that same functionality to drop a processor to save power? >> >> As I know it use cpu_suspend to switch out a processor in IKS and there is no cpu hotplug notifier in this procedure. >> >> >> Why do you need/want this notification? What are you going to do with this information that you don't already have? >> >> The offline won't be updated if an in kernel hotplug governor plug in / out a core which cause the sysfs interface report a wrong status. >> >> >> thanks, >> >> greg k-h >> >> >> Best Regards, >> Neil Zhang > -- > 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/ > -- 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/