Hello Geert, On Thu, Dec 05, 2024 at 09:52:28AM +0100, Geert Uytterhoeven wrote: > On Thu, Dec 5, 2024 at 9:28 AM Uwe Kleine-König <ukleinek@xxxxxxxxxx> wrote: > > Well, the policy is to pick the highest possible period not bigger than > > the requested period. So if B is asked to be set to 5ms and 5ns is the > > highest currently possible value, that's it. > > Really? So overclocking is preferred over underclocking? It really depends on your use case if 8ms or 10 ms is better in a situation where 9 ms is the optimal value. Note that with the new waveform abstraction there is also a way to round a waveform W to the waveform W' that would actually be configured if W was passed to the apply function. That allows a consumer (in the future probably with the help of some functions provided by the pwm framework) to determine a request to get the smallest period >= 9 ms. Best regards Uwe
Attachment:
signature.asc
Description: PGP signature
