Reimplement wakeup()

[caternuson]

This is a potential solution for: adafruit/Adafruit_Blinka#262

With the current implementation, the MCP2221 apparently does not output something that works, so the subsequent I2C comms fail. See logic traces in issue thread above. This reimplementation takes a different approach for generating the ATECC wakeup signal. Instead of using a scan it tries to address something at 0x00 and relies on 8 clock cycles of a 100kHz bus. So this hack will not work for I2C clock frequencies higher than 100kHz.

NOTE: the other big change is that wakeup() no longer returns a boolean. But that appears to have only been needed by the previous implementation. It's not used anywhere else in the driver.

It now works with MCP2221:

(blinka) user@linux:~$ export BLINKA_MCP2221=1
(blinka) user@linux:~$ python3 atecc_simpletest.py 
ATECC Serial:  0123A3B16C97546FEE
Random Value:  1016
ATECC Counter #1 Value:  bytearray(b'\x01\x00\x00\x00')
Appending to the digest...
Appending to the digest...
SHA Digest:  bytearray(b'\x03\x1e\xdd}Ae\x15\x93\xc5\xfe\\\x00o\xa5u+7\xfd\xdf\xf7\xbcN\x84:\xa6\xaf\x0c\x95\x0fK\x94\x06')

and still works with native CP:

Adafruit CircuitPython 6.0.1 on 2020-12-28; Adafruit ItsyBitsy M4 Express with samd51g19
>>> import atecc_simpletest
ATECC Serial:  0123A3B16C97546FEE
Random Value:  1016
ATECC Counter #1 Value:  bytearray(b'\x03\x00\x00\x00')
Appending to the digest...
Appending to the digest...
SHA Digest:  bytearray(b'\x03\x1e\xdd}Ae\x15\x93\xc5\xfe\\\x00o\xa5u+7\xfd\xdf\xf7\xbcN\x84:\xa6\xaf\x0c\x95\x0fK\x94\x06')
>>> 

[brentru]

@caternuson before you merge - CI is still failing.

[caternuson]

Thanks. Yah, looks like new stuff I'm not familiar with yet. So trying to trace in and figure it all out.

[caternuson]

Adding a trace of what the new wakeup() function ends up looking like on the MCP2221:

Note that the contents, also 0x00 but could be anything, never actually get written out since it NACKS on the write start. But it's the ~100us SDA low time that the ATECC sees and wakes up. It's the use of 0x00 for the address that's important. At least that's my theory :)

[caternuson]

Looks like #17 will fix most of these. Still not sure why I'm failing black, since it's passing locally.