core: arduino: add ShiftIn/ShiftOut implementation

[KurtE]

resolves: #345

Note: I simply copied in the MBED version of this. The APIS were already contained within the header files.

Verified that the sketch within the issue built.
Modified it to verify that worked at all:

void setup() {
  // put your setup code here, to run once:
  pinMode(2, INPUT);
  pinMode(3, OUTPUT);
  pinMode(4, OUTPUT);
  pinMode(5, OUTPUT);
  //shiftIn(2, 3, MSBFIRST);
  //shiftOut(4, 5, LSBFIRST, 42);
  pinMode(LED_BUILTIN, OUTPUT);

}

uint8_t loop_count = 0;
void loop() {
  // put your main code here, to run repeatedly:
  loop_count++;
  digitalWrite(LED_BUILTIN, loop_count & 1);
  shiftOut(4, 5, LSBFIRST, loop_count);
  delay(1);
}

Built and ran on UnoQ with current sources as of this morning.

It has the same possible issue as the MBED version in that the faster the processor, the narrower the pulses are:

I know in the case of Teensy 4.x boards, they put in possible delays, but that could be looked at later if needed.

[pennam]

Thanks @KurtE I've force pushed your changes to make the ci happy.

[KurtE]

Thanks

[pillo79]

Thanks @KurtE, but we would like to avoid including LGPL licensed code directly in the main core. Even though the final license of any Arduino build is always LGPL, the "core" itself is Apache2 and this would be the first incompatible bit.

The implementation of shiftIn/shiftOut should be rewritten from scratch to make this viable for inclusion. Would you like to give it a try? 🤗

[KurtE]

Thanks @KurtE, but we would like to avoid including LGPL licensed code directly in the main core. Even though the final license of any Arduino build is always LGPL, the "core" itself is Apache2 and this would be the first incompatible bit.

Sorry, not sure how things like that would apply, as for example my hands are dirty, so not sure what writing from
scratch would mean.

Maybe you could just copy in the version from renesas_uno, which does not say anything? So no idea what license that code has.

The complete shift.cpp

#include <Arduino.h>

uint8_t shiftIn(pin_size_t dataPin, uint8_t clockPin, BitOrder bitOrder) {
	uint8_t value = 0;
	uint8_t i;

	for (i = 0; i < 8; ++i) {
		digitalWrite(clockPin, HIGH);
		if (bitOrder == LSBFIRST)
			value |= digitalRead(dataPin) << i;
		else
			value |= digitalRead(dataPin) << (7 - i);
		digitalWrite(clockPin, LOW);
	}
	return value;
}

void shiftOut(pin_size_t dataPin, uint8_t clockPin, BitOrder bitOrder, uint8_t val)
{
	uint8_t i;

	for (i = 0; i < 8; i++)  {
		if (bitOrder == LSBFIRST)
			digitalWrite(dataPin, !!(val & (1 << i)) ? HIGH : LOW);
		else	
			digitalWrite(dataPin, !!(val & (1 << (7 - i))) ? HIGH : LOW);
			
		digitalWrite(clockPin, HIGH);
		digitalWrite(clockPin, LOW);		
	}
}

EDIT: Looks like that whole project has MIT License... Only one of your cores which are Apache2 are your
two (or three) Zephyr boards setup... Did that mean that for your zephyr cores all of the core had to be rewritten
from scratch?

[CLAassistant]

All committers have signed the CLA.

[KurtE]

I will see about hacking up a different version... Although hard to say I can come up with an original version 😉

[KurtE]

I hacked in a different version of shiftIn and shiftOut,

The code is semi-similar to Teensy version, although the Teensy version is more advanced in that for example:
That is each of these two functions are made up of 3 or 4 routines. One for MSBFIRST one for LSBFirst, a version
in the header file that checks to see if what is passed in is a constant and not variable for (MSBFIRST/LSBFirst) and then it compiles it to call the appropriate ORDER function and the rest is optimized out, else it has full function that checks and calls the appropriate one. They also have checks in their code to see if it will be too fast and add in delays...
Actually for the output function, they only fully implement it for LSB order for MSB order they use the arm operator to reverse the bits and then call the LSB version...

This version does none of this, it simply checks MSB or LSB and has a for loop difference of either:
for (mask=1; mask; mask <<=1) or for (mask=-0x80; mask; mask >>=1)

Edit:
For testing of ShiftOut: I used the following sketch:

void setup() {
  pinMode(2, OUTPUT);
  pinMode(3, OUTPUT);
  pinMode(4, OUTPUT);
  pinMode(5, OUTPUT);
  pinMode(LED_BUILTIN, OUTPUT);
}

void loop() {
  digitalWrite(LED_BUILTIN, !digitalRead(LED_BUILTIN));
  for (uint16_t i = 0; i <= 0xff; i++) {
    shiftOut(2, 3, LSBFIRST, i);
    shiftOut(4, 5, MSBFIRST, i);
  }
  delay(250);
}

And checked with logic analyzer

For ShiftIn, I used the sketch:

//#define TEST_MSBFIRST
void setup() {
  Serial.begin(115200);
  pinMode(2, INPUT);
  pinMode(3, OUTPUT);
  pinMode(4, OUTPUT);

  pinMode(LED_BUILTIN, OUTPUT);

  attachInterrupt(3, &shiftOut_int, RISING);
}

uint8_t shiftOut_val;
void shiftOut_int() {
#ifdef TEST_MSBFIRST
  digitalWrite(4, (shiftOut_val & 0x80)? HIGH : LOW);
  shiftOut_val <<= 1;
#else  
  digitalWrite(4, (shiftOut_val & 0x01)? HIGH : LOW);
  shiftOut_val >>= 1;
#endif
}


void loop() {
  digitalWrite(LED_BUILTIN, !digitalRead(LED_BUILTIN));
  for (uint16_t i = 0; i <= 0xff; i++) {
    shiftOut_val = i;
#ifdef TEST_MSBFIRST
    uint8_t val = shiftIn(2, 3, MSBFIRST);
#else
    uint8_t val = shiftIn(2, 3, LSBFIRST);
#endif
    if (val != i) {
      printk("%x != %x\n", i, val);
      delay(250);
    }

  }
  delay(250);
}

With this I ran a jumper from D4 to D2 and checked for output Serial
And ran this on an Arduino GIGA

[pillo79]

Thank you for the code and test! I suggest one trivial fix, but LGTM! 🙂

[pillo79]

Use your name (or alias, if you prefer)... it's your work! 😉

[KurtE]

Could do, but is there any way to mark it as Copyright, everyone who has ever shifted in or out a byte ;)

[KurtE]

But pushed up the change

[KurtE]

However now the test for licenses fail...

[pillo79]

It seems the name is too short for it to be reliably detected by scancode. Black magic.
Can you please try this alternative? (the tag is case sensitive)

Suggested change

	* Copyright (c) KurtE
	* SPDX-FileCopyrightText: Copyright (c) KurtE

(or try adding the year?).

[pillo79]

Did a few tests locally. It appears if you add SPDX-FileCopyrightText: then you do NOT have to add Copyright. But it does require the year. So confusing! 🤷‍♂️

I have tested this works:

 * Copyright (c) 2026 KurtE

(there is also * SPDX-FileCopyrightText: 2026 KurtE, but it isn't used anywhere else in the repo, so I'd prefer the above one).
Sorry about this flurry of changes!

[KurtE]

Updated, fingers crossed

[github-actions]

Built 0.54.1-0.dev+62a97010

CI run PASSED 🟢

Artifact	Board	Core	Tests	RAM	Sketches	Warnings	Errors
✅ zephyr_contrib	ek_ra8d1	📗	✅	11.9%	2	-	-
	frdm_mcxn947	3 🏷️	✅	58.0%	2	-	-
	frdm_rw612	1 🏷️	✅	83.0%	2	-	-
✔️* zephyr_main	giga	4 🏷️	✅*	54.5%	44	8	-
	nano33ble	1 🏷️	✅*	78.7%	22	8	-
	nano_matter	📗	✔️*	⚠️ 85.7%	20	8	(2*)
	niclasense	2 🏷️	✅*	⚠️ 87.3%	20	8	-
	opta	4 🏷️	✔️*	46.7%	54	8	(2*)
	portentac33	3 🏷️	✔️*	⚠️ 95.1%	56	8	(8*)
	portentah7	3 🏷️	✔️*	47.3%	58	8	(2*)
✅* zephyr_unoq	unoq	📗	✅*	26.4%	62	8	-

Legend

Board	Test	Status description
🔥	🔥	Test run failed to complete.
❌	🔴	Test completed with unexpected errors.
✔️*	🚫	Test completed with errors, but all are known/expected.
✅*	🟡	Test completed with some warnings; no errors detected.
✅	🟢	Test passed successfully, with no warnings or errors.
🌑	🌑	Test was skipped.