On Thu, Mar 12, 2020 at 01:35:32PM +0530, Lokesh Vutla wrote: > On 12/03/20 12:10 PM, Uwe Kleine-König wrote: > > On Thu, Mar 12, 2020 at 09:52:09AM +0530, Lokesh Vutla wrote: > >> Only the Timer control register(TCLR) cannot be updated when the timer > >> is running. Registers like Counter register(TCRR), loader register(TLDR), > >> match register(TMAR) can be updated when the counter is running. Since > >> TCLR is not updated in pwm_omap_dmtimer_config(), do not stop the > >> timer for period/duty_cycle update. > > > > I'm not sure what is sensible here. Stopping the PWM for a short period > > is bad, but maybe emitting a wrong period isn't better. You can however > > optimise it if only one of period or duty_cycle changes. > > > > @Thierry, what is your position here? I tend to say a short stop is > > preferable. > > Short stop has side effects especially in the case where 1PPS is generated using > this PWM. In this case where PWM period is continuously synced with PTP clock, > cannot expect any breaks in PWM. This doesn't fall in the above limitations as > well. as duty_cycle is not a worry and only the rising edge is all that matters. > > Also any specific reason why you wanted to stop rather than having the mentioned > limitation? it is just a corner anyway and doesn't happen all the time. I'm a bit torn here. Which of the two steps out of line is worse depends on what is driven by the PWM in question. And also I think ignoring "just corner cases" is a reliable way into trouble. The usual PWM contributer (understandably) cares mostly about their own problem they have to solve. If however you take a step back and care about the PWM framework as a whole to be capable to solve problems in general, such that any consumer just has to know that there is a PWM and start requesting specific settings for their work to get done, it gets obvious that you want some kind of uniform behaviour of each hardware driver. And then a short inactive break between two periods is more common and better understandable than a mixed period. Also being a corner case that only happens (say) once in 100000 cases isn't a clear upside. This just results in a machine leaving the development department, passing the production test and then behave unexpected once per year in the field. Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | https://www.pengutronix.de/ |