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 ? -- <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook | <http://twitter.com/#!/linaroorg> Twitter | <http://www.linaro.org/linaro-blog/> Blog
