Add examples to atmega-hal documentation

[armandas]

Part of #590

[Rahix]

Very nice, thanks a lot!

[Rahix]

A minor thing, but because people like copying example code for production use, let's talk about it:

I think this should be using a saturating_add() instead of wrapping_add() so the value doesn't cycle back to zero after 256 boots. The print statement should then display a different message on overflow. E.g.

Suggested change

	let mut boot_count = eeprom.read_byte(BOOT_COUNT_OFFSET);
	boot_count = boot_count.wrapping_add(1);
	eeprom.write_byte(BOOT_COUNT_OFFSET, boot_count);
	ufmt::uwriteln!(&mut serial, "Boot count: {}", boot_count).unwrap();
	let mut boot_count = eeprom.read_byte(BOOT_COUNT_OFFSET);
	boot_count = boot_count.saturating_add(1);
	eeprom.write_byte(BOOT_COUNT_OFFSET, boot_count);
	if boot_count == u8::MAX {
	ufmt::uwriteln!(&mut serial, "Boot count: >={} (overflow)", boot_count).unwrap();
	} else {
	ufmt::uwriteln!(&mut serial, "Boot count: {}", boot_count).unwrap();
	}

[armandas]

I agree, examples are how people learn, so we should try to show the best code.

The issue with EEPROM is that the data is 0xFF in the default/erased state, so we need to "initialize" it on first boot. A couple of possible solutions would be:

• Check and stop counting at u8::MAX - 1
• Use a larger integer (u32 should be enough for everyone™)

[Rahix]

Hmm, I see it is getting even more difficult... Tell you what, let's merge it as is and then maybe think about improvements in the future. I smell the solution being a custom "driver" to manage a bootcount, to contain all this complexity...