[GPIO] Added “Wake on Button Press” functionality. by chromoxdor · Pull Request #5519 · letscontrolit/ESPEasy

chromoxdor · 2026-03-20T23:29:08Z

I tried to integrate the feature, though I am not sure if the structure is fine with you since I more or less guestimated some things, but I am happy to receive your suggestions.
It also would help to do more testing on e.g. ESP32-S2,ESP32-C3,...

Add a checkbox in the switch plugin if the selected GPIO can wake the device from deep sleep.
tested on ESP32-S3 and ESP32-C6

ToDo:

More testing
Documentation

I tried to integrate the feature, though I am not sure if the structure is fine with you since I more or less guestimated some things, but I am happy to receive your suggestions. It also would help to do more testing on e.g. ESP32-S2,ESP32-C3,... - Add a checkbox in the switch plugin if the selected GPIO can wake the device from deep sleep. - tested on ESP32-S3 and ESP32-C6

TD-er · 2026-03-20T23:45:16Z

I have to look at it tomorrow, as I was just about to go to bed.
I had not yet thought about those wake features a lot as there is more to it than just 'wake up'.
For example you also must define whether a state of a GPIO must remain locked (e.g. output pin remain high/low during deep sleep) and/or what power domains must remain active during deep sleep, but also deepsleep is typically intended for reducing power consumption.
If a pin must remain 'active' during deep sleep, will it then be pulled high or low? And does this pull resistor cause the current draw to increase significantly?
Is it possible to set on what edge the ESP must boot?

And maybe not the least... how must ESPEasy react on wake-up by GPIO?
Afterall the settings are not yet loaded when ESPEasy 'boots', so I guess we also need to store stuff in RTC???

Just some ideas that pop up about this.
So this begs the question whether it is a feature for the GPIO plugin or should it be part of the 'Hardware' tab where we now also set the boot states of a pin.

chromoxdor · 2026-03-20T23:50:19Z

So this begs the question whether it is a feature for the GPIO plugin or should it be part of the 'Hardware' tab where we now also set the boot states of a pin.

Exactly my thought. But for simplicity, I was doing it in P001.
To the rest of your thoughts, I think I have some answers, but I also need to go to bed. It’s gotten too late... Good night

src/_P001_Switch.ino

TD-er · 2026-03-21T00:13:52Z

Oh and another idea...
You can see what the 'boot reason' is. One of them is when it is booted due to a GPIO pin and I think you can even derive which pin caused it.
Then there should be at least an event with the boot reason. Analog to what bootstate pins were active during boot (yes there is an event for it these days as I needed it for the Bhutan project :) )

src/src/Helpers/_Plugin_Helper_GPIO.cpp

src/_P001_Switch.ino

- added "System#BootCause" and "System#GPIOWake" events - moved gpio-wake related functions to _Plugin_Helper

chromoxdor · 2026-03-22T22:54:54Z

All seems to work fine, but I really need help with storing the bitmask to SettingsStruct.
This is above my pay grade 🙂 Help is really appreciated here.

src/src/WebServer/HardwarePage.cpp

chromoxdor · 2026-03-24T10:13:57Z

All seems to work fine, but I really need help with storing the bitmask to SettingsStruct. This is above my pay grade 🙂 Help is really appreciated here.

@TD-er, can you help me with that when you have some time to spare?

TD-er · 2026-03-24T10:36:02Z

And you think someone with his last diploma literally being one of tying his shoe laces will be able to? ;)

Anyway, given the ESP's with the largest amount of GPIO pins are already close to 64 pins, I guess we should at least add a 64bit uint to the settings struct.

This should be done at the end of the SettingsStruct, so other members won't shift and render the struct incompatible with stored ones.

Then you should add a get/set function. Use these as inspiration:

  bool getNetworkEnabled(ESPEasy::net::networkIndex_t index) const;

  void setNetworkEnabled(ESPEasy::net::networkIndex_t index, bool enabled);

The Arduino bitRead and bitWrite macros do not work on 64-bit ints, so you can choose to either add 2 uint32_t members to the settings, or use our own macros bitReadULL and bitWriteULL (see src/src/Helpers/Misc.h )

tonhuisman · 2026-03-24T11:02:42Z

Instead of adding extra bytes to the settings struct, you could reuse a couple of bytes from the OLD_TaskDeviceID array, where each item is 4 bytes in size.

chromoxdor · 2026-03-24T11:03:05Z

Then you should add a get/set function. Use these as inspiration:

Look at SettingsStruct.cpp and SettingsStruct.h. I already prepared everything but commented it for testing.

This should be done at the end of the SettingsStruct, so other members won't shift and render the struct incompatible with stored ones.

That a good starting point

TD-er · 2026-03-24T11:07:05Z

Ton's argument is also a good one as OLD_TaskDeviceID is indeed a vector of 32-bit ints and right now not all are being used.
Also it is an ESP32-only feature, so you could make it a bit more like this:

#ifdef ESP32
  uint32_t wakePin_bitmask_lo{};
  uint32_t wakePin_bitmask_hi{};
  unsigned int  OLD_TaskDeviceID[N_TASKS - 10] = {0};  // UNUSED: this can be reused
#else
  unsigned int  OLD_TaskDeviceID[N_TASKS - 8] = {0};  // UNUSED: this can be reused
#endif

Also we can expect those values to be already set to 0 in existing settings.
When appending to the struct, this may not be guaranteed.

tonhuisman · 2026-03-24T11:44:18Z

Should we reserve 2 bits per GPIO, so we can accommodate low and high activation?
There are functions to get and set 2 (and 3, 4, 8 and 16) bits to/from an UL (uint32_t) for convenience.

TD-er · 2026-03-24T11:50:32Z

Should we reserve 2 bits per GPIO, so we can accommodate low and high activation? There are functions to get and set 2 (and 3, 4, 8 and 16) bits to/from an UL (uint32_t) for convenience.

There are 8 uin32_t left, so we can go upto 4 bits per GPIO.
If the code doesn't get too complex, we could combine it, but maybe it is better readable if we add it when it is actually being implemented.

Edit:
Hmm there are 22 left (N_TASKS - 10) so even more reason to keep the bitmask per uint32_t and not try to combine bitmasks and try to wrap them in multiple uint32's

chromoxdor · 2026-03-24T12:31:31Z

Oh, oh, did I hear someone saying "low and high activation„? I will see how far I am coming. For now, I need to keep it as simple as possible, as I will probably very soon have no time at all for some months.

TD-er · 2026-03-24T12:32:25Z

Oh, oh, did I hear someone saying "low and high activation„? I will see how far I am coming. For now, I need to keep it as simple as possible, as I will probably very soon have no time at all for some months.

Make sure you will get enough sleep before that will happen :)

chromoxdor · 2026-03-28T19:21:43Z

#ifdef ESP32
  uint32_t wakePin_bitmask_lo{};
  uint32_t wakePin_bitmask_hi{};
  unsigned int  OLD_TaskDeviceID[N_TASKS - 10] = {0};  // UNUSED: this can be reused
#else
  unsigned int  OLD_TaskDeviceID[N_TASKS - 8] = {0};  // UNUSED: this can be reused
#endif

I tried this but getting:

src/src/Helpers/ESPEasy_checks.cpp: In function 'void run_compiletime_checks()':
src/src/Helpers/ESPEasy_checks.cpp:181:35: error: static assertion failed
  181 |   static_assert((232 + TASKS_MAX) == offsetof(SettingsStruct, OLD_TaskDeviceID), ""); // 32-bit alignment, so offset of 2 bytes.
      |                                   ^
src/src/Helpers/ESPEasy_checks.cpp:181:35: note: the comparison reduces to '(264 == 272)'

Any idea what I am doing wrong?

tonhuisman · 2026-03-28T20:44:03Z

By conditionally inserting 2 variables before the address of the OLD_TaskDeviceID array the offset will change, so you would also have to make the assertion for that address conditional, adding the missing 8 bytes to that check (232 -> 240).

Alternatively, you can also make (only) these variables not conditional, but also include them for ESP8266, as the size of the SettingsStruct doesn't change it won't have an impact on the build-size. Then you can just adjust the offset assertion as noted above, and it will fit either build-type.

src/src/DataStructs/SettingsStruct.h

- using bitmask instead of bool for wakeonhigh - refining docs

chromoxdor · 2026-03-29T16:48:52Z

FIY, both submenus are now collapsed by default:

src/src/DataStructs_templ/SettingsStruct.cpp

TD-er · 2026-03-29T17:17:15Z

Is wake-on-high vs. -low a global flag for all wake-up pin settings?
So it is either wake-up-on-X for all pins ?

Can you set multiple wake-up pins?

chromoxdor · 2026-03-29T17:25:44Z

Is wake-on-high vs. -low a global flag for all wake-up pin settings? So it is either wake-up-on-X for all pins ?

Yes, as stated in the documentation I wrote ;)
For EXT1-Wakeup, this is not possible to set it individually.
For devices with GPIO-Wakeup, it is possible, but not implemented as it gets a bit too complex for now… at least for me.

Can you set multiple wake-up pins?

You can check any GPIO that is supported. Just added a picture in the first post.

src/src/DataStructs/SettingsStruct.h

src/_Plugin_Helper.cpp

src/src/WebServer/Markup_Forms.cpp

src/src/WebServer/HardwarePage.cpp

- adding all what was suggested....

src/src/DataStructs_templ/SettingsStruct.cpp

src/_Plugin_Helper.cpp

src/src/WebServer/HardwarePage.cpp

- cleanup + refinements

TD-er · 2026-03-30T07:03:06Z

@chromoxdor
If you're going to use the VariousBits_3.unused_00 position, I will be using the bit position for VariousBits_3.unused_01 for a checkbox to enable/disable mDNS

chromoxdor · 2026-03-30T07:08:17Z

Be my guest. :) I am done here (or at least I think I am)

chromoxdor · 2026-03-30T07:56:19Z

BTW: Should I remove the Note for ESP32 devices? It seem to be a bit outdated

TD-er · 2026-03-30T08:33:58Z

BTW: Should I remove the Note for ESP32 devices? It seem to be a bit outdated

It is still valid for some ESP32-boards.
I guess we should check per ESP32-xx board to see what is the onboard LED pin and then act on it.

chromoxdor · 2026-03-30T08:54:27Z

It is still valid for some ESP32-boards. I guess we should check per ESP32-xx board to see what is the onboard LED pin and then act on it.

But to keep track of this multitude of different boards, it would need an extra database since there is no existing one. And an automatic detection of the specific board is impossible.

Edit: And not everyone wants to use the onboard LED for that...

- added a warning to the docs - added missing monitor keywords

chromoxdor · 2026-03-30T09:15:09Z

@chromoxdor If you're going to use the VariousBits_3.unused_00 position, I will be using the bit position for VariousBits_3.unused_01 for a checkbox to enable/disable mDNS

When I am thinking about it, I would add at some point a checkbox for disabling internal pull-resistors to give the option to use exclusively external ones. But this will not happen in this PR anymore. The problem I had with the Espressif documentation is that it is sometimes inconclusive, not up to date, or hard to understand. So I had to research other sources as well until I got all the info I needed.

But you never know... It also depends on how long it takes until this PR gets merged :)
😉

- added a "disable pull resistors" checkbox for EXT1 Wakeup

chromoxdor · 2026-03-30T10:12:57Z

@TD-er
Ok, I used VariousBits_3.unused_01 to disable internal RTC pull-resistors globally for EXT1 wakeup (ESP32 classic, S2,S3, C6).
For others like the C3 with only the GPIO wakeup option, it seems you can only disable this at compile time as I found here: https://forum.arduino.cc/t/esp32-c3-super-mini-deepsleep-doorsensor/1401895/35

- refinements

tonhuisman · 2026-03-30T13:18:53Z

src/src/DataStructs/SettingsStruct.h

-      uint32_t unused_00                        : 1; // Bit 0
-      uint32_t unused_01                        : 1; // Bit 1
+      uint32_t wakeOnHigh_ckd                   : 1; // Bit 0 //HardwarePage: used for wake on high or low
+      uint32_t diableWakePulls                  : 1; // Bit 1 //HardwarePage: used for for disabling pulls in general


Is that a typo? disableWakePulls ?

Wasn't bit 1 already claimed? (~~not sure~~) (here)

- fixed typos

tonhuisman · 2026-03-30T15:24:34Z

src/src/DataStructs/SettingsStruct.h

    struct {
      uint32_t wakeOnHigh_ckd                   : 1; // Bit 0 //HardwarePage: used for wake on high or low
-      uint32_t diableWakePulls                  : 1; // Bit 1 //HardwarePage: used for for disabling pulls in general
+      uint32_t disableWakePulls                 : 1; // Bit 1 //HardwarePage: used for disabling internal pulls in general


It's still using bit 1... 🤔

Well I already changed it in my code, so leave it.

TD-er reviewed Mar 21, 2026

View reviewed changes

src/_P001_Switch.ino Outdated Show resolved Hide resolved

TD-er reviewed Mar 21, 2026

View reviewed changes

src/_P001_Switch.ino Outdated Show resolved Hide resolved

tonhuisman reviewed Mar 21, 2026

View reviewed changes

src/src/Helpers/_Plugin_Helper_GPIO.cpp Outdated Show resolved Hide resolved

chromoxdor added 2 commits March 21, 2026 14:21

[P001] Switched to EXT1 for all devices for lowest power consumption

b1abf4a

[P001] changed a line according to Tons suggestion

cedc78d

tonhuisman reviewed Mar 21, 2026

View reviewed changes

src/_P001_Switch.ino Outdated Show resolved Hide resolved

[GPIO] Setting moved to Hardware tab

6f3544a

- added "System#BootCause" and "System#GPIOWake" events - moved gpio-wake related functions to _Plugin_Helper

chromoxdor changed the title ~~[P001] Added “Wake on Button Press” functionality.~~ [GPIO] Added “Wake on Button Press” functionality. Mar 22, 2026

chromoxdor added 3 commits March 22, 2026 23:57

[GPIO] Setting moved to Hardware tab (removes unnecessary functions)

9cdac4d

[GPIO] add check for gpiowake event

3b5d8c7

[GPIO] fixed a c&p mistake

ac6ece0

tonhuisman reviewed Mar 23, 2026

View reviewed changes

src/src/WebServer/HardwarePage.cpp Outdated Show resolved Hide resolved

tonhuisman reviewed Mar 23, 2026

View reviewed changes

src/src/WebServer/HardwarePage.cpp Outdated Show resolved Hide resolved

[GPIO] adding details expansion + rules-keywords

2365dff

TD-er reviewed Mar 29, 2026

View reviewed changes

src/src/DataStructs/SettingsStruct.h Outdated Show resolved Hide resolved

[GPIO] adding all the missing bits and pieces part 4

e21f2ca

- using bitmask instead of bool for wakeonhigh - refining docs

TD-er reviewed Mar 29, 2026

View reviewed changes

src/src/DataStructs_templ/SettingsStruct.cpp Outdated Show resolved Hide resolved