From mboxrd@z Thu Jan 1 00:00:00 1970 From: Quentin Perret Subject: Re: [PATCH v6 10/14] sched/cpufreq: Refactor the utilization aggregation method Date: Mon, 10 Sep 2018 11:07:21 +0100 Message-ID: <20180910100719.uan6gskxpgu25dj5@queper01-lin> References: <20180820094420.26590-1-quentin.perret@arm.com> <20180820094420.26590-11-quentin.perret@arm.com> <3619149.Cl0Cp2apok@aspire.rjw.lan> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <3619149.Cl0Cp2apok@aspire.rjw.lan> Sender: linux-kernel-owner@vger.kernel.org To: "Rafael J. Wysocki" Cc: peterz@infradead.org, linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org, gregkh@linuxfoundation.org, mingo@redhat.com, dietmar.eggemann@arm.com, morten.rasmussen@arm.com, chris.redpath@arm.com, patrick.bellasi@arm.com, valentin.schneider@arm.com, vincent.guittot@linaro.org, thara.gopinath@linaro.org, viresh.kumar@linaro.org, tkjos@google.com, joel@joelfernandes.org, smuckle@google.com, adharmap@codeaurora.org, skannan@codeaurora.org, pkondeti@codeaurora.org, juri.lelli@redhat.com, edubezval@gmail.com, srinivas.pandruvada@linux.intel.com, currojerez@riseup.net, javi.merino@kernel.org List-Id: linux-pm@vger.kernel.org On Monday 10 Sep 2018 at 11:53:58 (+0200), Rafael J. Wysocki wrote: > On Monday, August 20, 2018 11:44:16 AM CEST Quentin Perret wrote: > > Schedutil aggregates the PELT signals of CFS, RT, DL and IRQ in order > > to decide which frequency to request. Energy Aware Scheduling (EAS) > > needs to be able to predict those requests to assess the energy impact > > of scheduling decisions. However, the PELT signals aggregation is only > > done in schedutil for now, hence making it hard to synchronize it with > > EAS. > > > > To address this issue, introduce schedutil_freq_util() to perform the > > aforementioned aggregation and make it available to other parts of the > > scheduler. Since frequency selection and energy estimation still need > > to deal with RT and DL signals slightly differently, schedutil_freq_util() > > is called with a different 'type' parameter in those two contexts, and > > returns an aggregated utilization signal accordingly. > > This is complementary to patch [02/14] IMO. > > schedutil_freq_util() and map_util_freq() introduced by that patch should > always be used together as they are two parts of one algorithm in my view. I agree. > Would it be possible to make that clearer? I could squash the two at the beginning of the series in a preparatory patch that refactors schedutil for EAS all in one go, with a clear mention of what we intend to do (make EAS depend on sugov) in the commit message. Would that work ? Thanks, Quentin