On 05/06/2018 07:53, Viresh Kumar wrote: > On 05-06-18, 07:48, Daniel Lezcano wrote: >> As soon as we reach complete(), no timer can be set because of the >> condition before. > > Why not ? We aren't using any locks here and it is possible that > run_duration_ms > is set to 0 from idle_injection_stop() only after the first thread has > restarted > the hrtimer. Isn't it ?
If he restarted the timer, complete() won't be called and idle_injection_stop() will wait until the idle cycle injection is finished. -- <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
