I bailed on the stock dtoverlay pretty quickly since it doesn't seem to do anything much at all.
The following - I think - enables rp1_pwm0 and sets up each of the pins into the correct alt modes (confirmed after boot with pinctrl), but still no dice:
Is there some missing clock config or other confounding factor. PWM must be working in some capacity, or the cooling fan wouldn't work, right?
Bonus weirdness, PWM0 reports four PWMs but only lets me used three channels:
This is especially interesting since that's the channel the fan uses...
AAH!
I was assuming the newly added pwmchip2 would be PWM0, but it's actually PWM1... *sigh*... is there actually any way to know this at runtime!?
Looks like it can kinda be...
The following - I think - enables rp1_pwm0 and sets up each of the pins into the correct alt modes (confirmed after boot with pinctrl), but still no dice:
Code:
/dts-v1/;/plugin/;/{compatible = "brcm,bcm2712";fragment@0 {target = <&rp1_gpio>;__overlay__ {pwm_pins: pwm_pins {pins = "gpio12", "gpio13", "gpio18", "gpio19";function = "pwm0", "pwm0", "pwm0", "pwm0";};};};fragment@1 {target = <&rp1_pwm0>;frag1: __overlay__ {pinctrl-names = "default";pinctrl-0 = <&pwm_pins>;status = "okay";};};};Bonus weirdness, PWM0 reports four PWMs but only lets me used three channels:
Code:
$ cat /sys/class/pwm/pwmchip2/npwm 4$ echo "3" | sudo tee -a /sys/class/pwm/pwmchip2/export 3tee: /sys/class/pwm/pwmchip2/export: Device or resource busy$ ls /sys/class/pwm/pwmchip2/device export npwm power pwm0 pwm1 pwm2 subsystem uevent unexportAAH!
I was assuming the newly added pwmchip2 would be PWM0, but it's actually PWM1... *sigh*... is there actually any way to know this at runtime!?
Looks like it can kinda be...
Code:
$ ls /sys/class/pwm/pwmchip2/device/consumer:platform:cooling_fan driver driver_override modalias of_node power pwm subsystem supplier:platform:1f00018000.clocks ueventStatistics: Posted by gadgetoid — Wed Mar 06, 2024 1:35 pm
