pio.c needs 32u to be changed to NUM_BANK0_GPIOS on line 347 to support RP2350B

[peterk268]

Findings found here:
https://picopalgbc.wordpress.com/2025/08/06/97-swap-pclk-and-den-port-code-to-pico-pal/

Line 347 in src/rp2_common/hardware_pio/pio.c currently uses 32u as the check for the gpio pin number threshold but this doesn't support the RP2350B with 48 GPIO so this would return the error code -5 or invalid arg from this line.

invalid_params_if_and_return(PIO, pin >= 32u, PICO_ERROR_INVALID_ARG);

Thus the fixed line is here:
invalid_params_if_and_return(PIO, pin >= NUM_BANK0_GPIOS, PICO_ERROR_INVALID_ARG);

[lurch]

It looks like you missed the pin -= pio_get_gpio_base(pio); on line 346 ? 😉
Each PIO can only talk to 32 pins at a time, so if you're using one of the "higher" GPIOs on a RP2350B, pio_get_gpio_base(pio) will return 16, which means that pio_sm_set_consecutive_pindirs will be expecting a pin parameter between 16 and 47 inclusive.

@kilograham Perhaps there ought to be a

invalid_params_if_and_return(PIO, pin < pio_get_gpio_base(pio), PICO_ERROR_INVALID_ARG);

before line 346?

[peterk268]

Ah that's my bad. I guess once I kept going with fixing the error codes and then I set the base to 16 after making these changes it lead me to believe that this was an issue.
As I only realized after getting the bad alignment error code that setting the base was needed. That was something I thought scanvideo.c would handle but I understand as you said the pico-extras repo is experimental.

[lurch]

Can this pico-sdk issue be closed now then @peterk268 ?

[peterk268]

Yes thank you.