From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lan Tianyu Subject: Re: [PATCH V2] Cpufreq: Make governor data on nonboot cpus across system suspend/resume Date: Fri, 22 Nov 2013 16:19:23 +0800 Message-ID: <528F138B.1090107@intel.com> References: <9847309.KdKOG5y1Zx@vostro.rjw.lan> <1384616184-6197-1-git-send-email-tianyu.lan@intel.com> <5288425A.6000501@linaro.org> <528F0C9E.8020409@linaro.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <528F0C9E.8020409@linaro.org> Sender: cpufreq-owner@vger.kernel.org To: viresh kumar Cc: "Rafael J. Wysocki" , "cpufreq@vger.kernel.org" , "linux-pm@vger.kernel.org" , Linux Kernel Mailing List , Nishanth Menon List-Id: linux-pm@vger.kernel.org On 2013=E5=B9=B411=E6=9C=8822=E6=97=A5 15:49, viresh kumar wrote: > On Sunday 17 November 2013 09:43 AM, viresh kumar wrote: >> On 16 November 2013 21:06, Lan Tianyu wrote: >=20 >> But I don't really like the solution here. You are handling frozen f= or EXIT in >> cpufreq-core and for INIT in governor. That doesn't look like the ri= ght >> approach. There are out of tree governors too (I know we don't care = about them >> :)), and those also need to adapt with some policy made at cpufreq-c= ore level. >> >> I told you that I had another solution for this problem, pretty simi= lar to >> yours. It looked like this: >=20 > Hi Lan, >=20 Hi Viresh: > There is some confusion going on here :) >=20 I think you also are in the Cc list and replied the mail.:) https://lkml.org/lkml/2013/11/21/273 > There were few problems in the approach in your patch, which I have m= entioned > above, and Rafael agreed to them.. I only saw the out-of-tree governor issue your mentioned but where they are? How upstream kernel cares them? >=20 >> But after the PM notifiers patch I even don't want this to go.. I wi= ll make sure >> that that patch goes in, in one form or another :) >=20 > And I was still trying to get a better solution in place of these cha= nges. And > was going on the suggestions you gave about calling cpufreq callbacks= from > dpm_{suspend|resume}_noirq() calls.. And I am on my way to get things= fixed that > way. And so we don't actually need this patch anymore (I just saw tha= t you have > sent another version of it, probably because Rafael asked? Don't know= what > happened there :)).. >=20 > So, I will try to get something working soon for you and Nishanth.. >=20 --=20 Best regards Tianyu Lan