If PULSECOUNT is enabled, current logic in stm32_pwm.c assumes that all PWM instances support PULSECOUNT which is not true. Timers without RCR register shouldn't use pwm_pulsecount_* functions. Probably that is the problem.
I guess you can do something like this: https://github.com/raiden00pl/incubator-nuttx/commit/dc30c50d161cc7d4cbea849f8d46dc35924e51fe I haven't tested it, haven't even compiled it, but I guess that's the way to fix the problem. niedz., 4 kwi 2021 o 10:13 <disruptivesolution...@gmail.com> napisał(a): > I Tried to show the difference (debug) in one image.... > > https://imgshare.io/image/image.p307iH > > I tried to state the differences... according the reference sheets and > application notes from the STM32F4xx family this should work "together".... > where you use say TIM1 (or TIM8) [ADVANCED TIMERS] with "COUNTS" and where > you say use TIM9 (or any other [GENERIC TIMER] in say "time bases" mode. > > What am I doing wrong? And how could I tackle this to get it working? > Is seems (according tot he debug output) that its a "driver" > implementation and that the CONFIG_PWM_PULSECOUNT KConfig is causing this > behaviour?? > > Thanks you for thinking with me... > Ben > >
