From mboxrd@z Thu Jan 1 00:00:00 1970 From: Viresh Kumar Subject: Re: [PATCH] thermal/cpu_cooling: remove local cooling state variable Date: Sat, 1 Aug 2015 17:04:05 +0530 Message-ID: <20150801113405.GL899@linux> References: <1437516835-198750-1-git-send-email-radivoje.jovanovic@linux.intel.com> <9hhegjxbmqz.fsf@e105922-lin.cambridge.arm.com> <20150724100905.7e2e53f4@radivoje-desk2> <9hhoaiuap3m.fsf@e105922-lin.cambridge.arm.com> <20150730080541.GD31351@linux> <20150730132130.3c957267@radivoje-desk2> <20150731031841.GH17794@linux> <20150731083003.2f47ca5f@radivoje-desk2> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mail-pa0-f42.google.com ([209.85.220.42]:35894 "EHLO mail-pa0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751207AbbHALeK (ORCPT ); Sat, 1 Aug 2015 07:34:10 -0400 Received: by pachj5 with SMTP id hj5so55641164pac.3 for ; Sat, 01 Aug 2015 04:34:10 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20150731083003.2f47ca5f@radivoje-desk2> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Radivoje Jovanovic Cc: Punit Agrawal , rjw@rjwysocki.net, LKML , Linux PM , Zhang Rui , Eduardo Valentin , Radivoje Jovanovic On 31-07-15, 08:30, Radivoje Jovanovic wrote: > I just looked over the notifier in the current upstream (my patch was > made on our production kernel which is 3.14 and has old notifier > implementation with notifier_device in place) and I see your point. That's disappointing. You were expected to check if the same problem exists in mainline. > I agree with you that this patch is trivial for the current > implementation since the notifier, as it is currently, will enforce > cpu_cooling policy change at every CPUFREQ_ADJUST which would cause > problems in our current implementation. In our implementation there is > a cpufreq driver that will also change policies during CPUFREQ_ADJUST, > once the request comes from the underlying FW so there would be a fight > who gets there first since cpu_cooling will change the policy in > CPUFREQ_ADJUST notifier_chain and the driver would do the same thing. > It seems to me that better implementation of the cpu_cooling notifer > would be to keep the flag and change the policy in CPUFREQ_ADJUST only > when the change was requested by cpu_cooling, and update the current > state of cpufreq_cooling_device during CPUFREQ_NOTIFY event. > What do you think? I think the way cpu-cooling is written today, is an *ugly* hack. We hack the notifier to change policy->max and no one is notified for it. That's crap. I would rather get some help from cpufreq core on that. Which can provide some APIs to take care of thermal considerations. Okay, I push that to my todo list. Will keep you all posted. -- viresh