Fix analogWrite()

[facchinm]

Implement DAC and fix the dtoverlay for Giga

Fixes #38 and #19

[facchinm]

@KurtE can you try giving it a spin?
I also noticed in your last PRs that you are still targeting the old zephyr branch. To update, just run west update once in the venv

[KurtE]

Will try it out.

Running the west update now.

For some reason I assumed that using the bootstrap.sh would update everything...

[KurtE]

I verified that at least most of the PWM pins work, a few I did not test as they were hooked up to my lcd display..., but tried,
2, 3, 4,5, 7

I also tried the DAC and my simple sketch worked:

enum dacPins dp = DAC0;
#define toggle_pin 31
void setup() {
    // put your setup code here, to run once:
    Serial.begin(115200);
    //while (!Serial) {}
    delay(500);

    Serial.println("Start of PWM test");
    pinMode(toggle_pin, OUTPUT);
    delay(1000);
}

void maybe_pause() {
    if (Serial.available()) {
        while (Serial.read() != -1) {}
        Serial.println("Paused");
        while (Serial.read() == -1) {}
        while (Serial.read() != -1) {}

        dp = (dp == DAC0)? DAC1 : DAC0;
    }
}

void loop() {
    // put your main code here, to run repeatedly:
    digitalWrite(toggle_pin, !digitalRead(toggle_pin));
    digitalWrite(toggle_pin, !digitalRead(toggle_pin));
    for (int i = 0; i < 256; i += 32) {
        analogWrite(dp, i);
        digitalWrite(toggle_pin, !digitalRead(toggle_pin));
        maybe_pause();
        delay(100);
    }

    digitalWrite(toggle_pin, !digitalRead(toggle_pin));
    digitalWrite(toggle_pin, !digitalRead(toggle_pin));
    for (int i = 255; i >= 0; i -= 32) {
        analogWrite(dp, i);
        digitalWrite(toggle_pin, !digitalRead(toggle_pin));
        maybe_pause();
        delay(100);
    }
    delay(50);
}
![image](https://github.com/user-attachments/assets/d7ed5c01-803d-4a29-86d9-6b588838325c)

[KurtE]

Side note:

I wish that some of these apis did not have two forms of their API.
In this case, having two APIs for analogWrite(pin, value);
does not really impact me, as my usages for DAC versus digital PWM is different.

However for analogRead where the A0-A7 is done by one API with one data type and A8-A11 are done by another API with
different type for selecting pin, feels wrong, as I am using them for the same thing. And for example if I have libraries who take parameters for which Analog pin to use, they won't probably work on the two different types...

But that is probably just me. Glad more things are working!

[facchinm]

The big problem about A8-A11 on the Giga is that the pins can't be used as digital, so either we pollute the whole core for a single use case (which feels very wrong) or we implement something as we did (which indeed creates problems when wrapping the function call, but keeps the core clean)