Implement mpr121 touch sensor capability to GPIO Expanders

[caelor]

This is an updated PR cleanly against the v4.3 branch, as #190 was raised against the wrong branch.

Commit 396e754 from #190 could be cherry picked for a REAME.md update (as the v4.3 README.md isn't a small update to match)

This pull request adds an I2C based MP121 chip as a GPIO Expander. This chip provides 12 capacitive touch sensing pins, which are implemented here as inputs suitable for buttons.

The chip also can alternatively provide 8 GPIOs if correctly configured - this pull request does not implement this functionality (there seem to be more cost effective i2c GPIO expanders).

To use, set the gpio_exp_config nvs parameter to model=mpr121,addr=90,count=12,intr=<n> (assuming on default i2c address). GPIOs 40-51 can then be used as BUTTON_HIGH buttons.

The chip init parameters have been taken from those used by https://github.com/BareConductive/mpr121 with touch and release thresholds used by the SimpleTouch example there. Those values work for me.

[philippe44]

Can you put the initialization in a table instead so so many calls to I2c_write?

[caelor]

Can you put the initialization in a table instead so so many calls to I2c_write?

I agree that would likely be tidier, but also beyond my abilities these days - I've not routinely written C for over 20 years and I'm more than rusty at it.

I'd be happy to test a table based alternative to confirm correct initialisation though.

[philippe44]

Something like

static const struct {
	uint8_t addr;
	uint8_t data;
} mpr121_init_table[] = { { 0x2b, 0x01 }, { 0x2c, 0x01 }, { 0x2d, 0x10 }, {0x2e, 0x20}, {0, 0} };

for (int i = 0; mpr121_init_table[i].addr; i++) {
	err |= i2c_write(self->phy.port, self->phy.addr, mpr121_init_table[i].addr, mpr121_init_table[i].data, 1);
}

[caelor]

Thanks, I'm currently away on business but I'll try that when I return.

[caelor]

I've attempted to implement a tabular initialisation approach (see gist https://gist.github.com/caelor/db098ec88741630ab662f3b8fc4cdcd8).

Unfortunately, I've not been able to test as my docker container now seems to fail with IDF errors. I've spent several hours trying to get to the bottom of it without any luck. I'll revisit if I get chance around other commitments.

[philippe44]

Looks like the right implementation to me. I'll take it and push an update