Add low_power_set_pins_low_leakage function #3049

2026-06-30T09:33:12-05:00

will-v-pi commented

2026-06-30 09:33:12 -05:00

(Migrated from github.com)

This sets the pins to what I think is the lowest leakage state (pads_bank0 have pulls & input disabled, io_bank0 has output override set to disabled)

Wren6991 (Migrated from github.com) reviewed

2026-06-30 09:33:12 -05:00

lurch commented

2026-06-30 09:44:38 -05:00

(Migrated from github.com)

Naive question, but do the GPIO isolation latches on RP2350 help at all?

peterharperuk (Migrated from github.com) reviewed

2026-06-30 09:44:42 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -224,6 +224,27 @@ int low_power_set_external_clock_source(uint src_hz, uint gpio_pin) {

				}

I wonder whether that neede to be (1ull << i)

will-v-pi (Migrated from github.com) reviewed

2026-06-30 09:45:53 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -224,6 +224,27 @@ int low_power_set_external_clock_source(uint src_hz, uint gpio_pin) {

				}

Yes, it probably does

        if (exclude_mask & (1ull << i)) continue;

Yes, it probably does ```suggestion if (exclude_mask & (1ull << i)) continue; ```

will-v-pi commented

2026-06-30 09:57:33 -05:00

(Migrated from github.com)

Naive question, but do the GPIO isolation latches on RP2350 help at all?

As far as I understand it, the isolation latches just isolate the pads logic from the switched core until they are configured - isolation is enabled anyway when the switched core is powered off (9.7. Pad isolation latches in the datasheet)

> Naive question, but do the GPIO isolation latches on RP2350 help at all? As far as I understand it, the isolation latches just isolate the pads logic from the switched core until they are configured - isolation is enabled anyway when the switched core is powered off (9.7. Pad isolation latches in the datasheet)

lurch (Migrated from github.com) reviewed

2026-06-30 10:09:17 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -226,1 +226,4 @@

				void low_power_set_pins_low_leakage_exclude_mask(uint32_t exclude_mask) {

				    for (uint i=0; i < NUM_BANK0_GPIOS; i++) {

				        if (exclude_mask & (1u << i)) continue;

I guess you could also add

if (! (include_mask >> i) ) break;

but it's probably not worth it?

I guess you _could_ also add ```c if (! (include_mask >> i) ) break; ``` but it's probably not worth it?

will-v-pi (Migrated from github.com) reviewed

2026-06-30 11:06:21 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -226,1 +226,4 @@

				void low_power_set_pins_low_leakage_exclude_mask(uint32_t exclude_mask) {

				    for (uint i=0; i < NUM_BANK0_GPIOS; i++) {

				        if (exclude_mask & (1u << i)) continue;

I've removed include_mask now anyway, so not worth adding it back in - generally I'd expect this function to be operating on lots of pins anyway, as opposed to the masked functions in hardware_gpio which I'd expect to be operating on a smaller set of pins

I've removed `include_mask` now anyway, so not worth adding it back in - generally I'd expect this function to be operating on lots of pins anyway, as opposed to the masked functions in `hardware_gpio` which I'd expect to be operating on a smaller set of pins

lurch (Migrated from github.com) reviewed

2026-06-30 11:56:18 -05:00

src/rp2_common/pico_low_power/include/pico/low_power.h

					
				@ -130,6 +130,32 @@ typedef enum {

				typedef void (*low_power_pstate_resume_func)(pstate_bitset_t *pstate);

I guess the GPIO_FUNC_SIO comment here needs to be updated to reflect one of the later commits?

I guess the `GPIO_FUNC_SIO` comment here needs to be updated to reflect one of the later commits?

will-v-pi (Migrated from github.com) reviewed

2026-06-30 11:58:48 -05:00

src/rp2_common/pico_low_power/include/pico/low_power.h

					
				@ -130,6 +130,32 @@ typedef enum {

				typedef void (*low_power_pstate_resume_func)(pstate_bitset_t *pstate);

 * Disables pulls & inputs on the pads, and disables the IO output

```suggestion * Disables pulls & inputs on the pads, and disables the IO output ```

lurch (Migrated from github.com) reviewed

2026-06-30 11:59:38 -05:00

src/rp2_common/pico_low_power/include/pico/low_power.h

					
				@ -130,6 +130,32 @@ typedef enum {

				typedef void (*low_power_pstate_resume_func)(pstate_bitset_t *pstate);

/*! \brief  Set all pins to a low leakage state (64-bit mask version)

(just to clarify which part of it is "64-bit"?)

```suggestion /*! \brief Set all pins to a low leakage state (64-bit mask version) ``` (just to clarify which part of it is "64-bit"?)

lurch (Migrated from github.com) reviewed

2026-07-01 10:09:00 -05:00

src/rp2_common/pico_low_power/include/pico/low_power.h

					
				@ -130,6 +130,32 @@ typedef enum {

				typedef void (*low_power_pstate_resume_func)(pstate_bitset_t *pstate);

 * Does not change the state of pins in the exclude_mask.

```suggestion * Does not change the state of pins in the exclude_mask. ```

lurch (Migrated from github.com) reviewed

2026-07-01 10:12:58 -05:00

test/pico_low_power_test/low_power_test_gpio.c

					
				@ -127,6 +127,7 @@ int main() {

				    init_powman_ext_ctrl();

When reading isolated code like this (i.e. without referring to the doxygen), one might naively assume that it's the pins in USED_PIN_MASK that are being set to low-leakage? 🤔

low_power_set_all_pins_low_leakage_except(USED_PIN_MASK);

is maybe a bit too wordy though 😉

When reading isolated code like this (i.e. without referring to the doxygen), one might naively assume that it's the pins in `USED_PIN_MASK` that are being set to low-leakage? :thinking: ```c low_power_set_all_pins_low_leakage_except(USED_PIN_MASK); ``` is maybe a bit too wordy though :wink:

will-v-pi (Migrated from github.com) reviewed

2026-07-01 10:44:58 -05:00

test/pico_low_power_test/low_power_test_gpio.c

					
				@ -127,6 +127,7 @@ int main() {

				    init_powman_ext_ctrl();

The function naming is up for debate, but I think that might be a bit too long - maybe low_power_set_all_pins_low_leakage instead?

The function naming is up for debate, but I think that might be a bit too long - maybe `low_power_set_all_pins_low_leakage` instead?

lurch (Migrated from github.com) reviewed

2026-07-01 10:49:10 -05:00

test/pico_low_power_test/low_power_test_gpio.c

					
				@ -127,6 +127,7 @@ int main() {

				    init_powman_ext_ctrl();

Yeah, that's probably better than low_power_set_most_pins_low_leakage 🤣

Yeah, that's probably better than `low_power_set_most_pins_low_leakage` :rofl:

kilograham (Migrated from github.com) reviewed

2026-07-01 11:07:20 -05:00

kilograham (Migrated from github.com) reviewed

2026-07-01 11:16:22 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -224,6 +224,27 @@ int low_power_set_external_clock_source(uint src_hz, uint gpio_pin) {

				}

i think perhaps low_power_set_pins_low_leakage_exclude_mask whilst a mouthful is probably the most consistent with other naming

i think perhaps `low_power_set_pins_low_leakage_exclude_mask` whilst a mouthful is probably the most consistent with other naming

will-v-pi (Migrated from github.com) reviewed

2026-07-01 11:19:41 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -224,6 +224,27 @@ int low_power_set_external_clock_source(uint src_hz, uint gpio_pin) {

				}

I will change it

lurch (Migrated from github.com) reviewed

2026-07-01 11:45:36 -05:00

src/rp2_common/pico_low_power/low_power.c

					
				@ -224,6 +224,27 @@ int low_power_set_external_clock_source(uint src_hz, uint gpio_pin) {

				}

OTOH, it might sound like "low leakage exclude mask" is a thing, and that's what this function is setting? 😉 But I'll stop bike-shedding and just leave it as-is.

OTOH, it might sound like "low leakage exclude mask" is a thing, and that's what this function is setting? :wink: But I'll stop bike-shedding and just leave it as-is.