From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Lezcano Subject: Re: [PATCH V5] powercap/drivers/idle_injection: Add an idle injection framework Date: Thu, 7 Jun 2018 10:18:27 +0200 Message-ID: References: <1528190208-22915-1-git-send-email-daniel.lezcano@linaro.org> <20180606122357.GN12258@hirez.programming.kicks-ass.net> <22f5cf0b-049e-7938-55f6-4b4b154f8389@linaro.org> <20180606150203.GE12180@hirez.programming.kicks-ass.net> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Return-path: In-Reply-To: <20180606150203.GE12180@hirez.programming.kicks-ass.net> Content-Language: en-US Sender: linux-kernel-owner@vger.kernel.org To: Peter Zijlstra Cc: rjw@rjwysocki.net, linux-kernel@vger.kernel.org, Viresh Kumar , Eduardo Valentin , Javi Merino , Leo Yan , Kevin Wangtao , Vincent Guittot , Rui Zhang , Daniel Thompson , "open list:POWER MANAGEMENT CORE" List-Id: linux-pm@vger.kernel.org On 06/06/2018 17:02, Peter Zijlstra wrote: > On Wed, Jun 06, 2018 at 03:42:08PM +0200, Daniel Lezcano wrote: >> On 06/06/2018 14:23, Peter Zijlstra wrote: >>> On Tue, Jun 05, 2018 at 11:16:40AM +0200, Daniel Lezcano wrote: >>>> + atomic_t idle_duration_ms; >>>> + atomic_t run_duration_ms; >>> >>>> + idle_duration_ms = atomic_read(&ii_dev->idle_duration_ms); >>> >>>> + run_duration_ms = atomic_read(&ii_dev->run_duration_ms); >>> >>>> + atomic_set(&ii_dev->run_duration_ms, run_duration_ms); >>>> + atomic_set(&ii_dev->idle_duration_ms, idle_duration_ms); >>> >>>> + *run_duration_ms = atomic_read(&ii_dev->run_duration_ms); >>>> + *idle_duration_ms = atomic_read(&ii_dev->idle_duration_ms); >>> >>>> + if (!atomic_read(&ii_dev->idle_duration_ms)) >>> >>>> + if (!atomic_read(&ii_dev->run_duration_ms)) >>> >>> What is the point of atomic_t here ?! >> >> idle_duration and run_duration can be changed from different places at >> the same time. The atomic is here to ensure the read/write are consistent. >> >> Do you think it is pointless ? > > Yes, atomic_read() / atomic_set() are no more atomic than READ_ONCE() / > WRITE_ONCE(). So IIUC, neither atomic or WRITE|READ_ONCE are necessary in this code because of the wake_up_process() barrier is enough, right ? -- Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog