AirLift support for wifi, socketpool, and ssl modules

[dhalbert]

Implement wifi/socketpool/ssl for Airlift (NINA-FW) co-processors.

Starting this as a draft, for testing purposes. Tested with Adafruit NINA-FW 3.4.0.

• Web workflow not yet implemented.
• HTTP and HTTPS client work. UDP client and server work.
• HTTP server gets an error after a minute or so of continuous polling, alternating GET_STATE_TCP_CMD and AVAIL_DATA_TCP_CMD. The failure is that the ready line from the AirLift is in the wrong state. This may be some failure or race condition on the AirLift side. I have some other ideas for investigating this as well. Adding some delays did not help. Error is repeated:

Traceback (most recent call last):
  File "adafruit_httpserver/server.py", line 482, in poll
TimeoutError: timeout waiting for ready False

• Main implementation is in devices/airlift.
• wifi.radio.init_airlift() added to specify pins to use.
• CIRCUITPY_{WIFI,SSL,SOCKETPOOL} further refined as CIRCUITPY_*_NATIVE and CIRCUITPY_*_AIRLIFT. Compiled guards changed appropriately.
• Fixed common_hal_socketpool_socket_bind() return type from size_t to int.
• Fixed some socketpool function names which were not named canonically.
• Added common_hal_wifi_radio_get_version(().
• common_hal_socketpool_socket_recv_into() and socketpool_socket_recv_into() buffer args were const and should not have been.
• Use -flto=auto instead of -flto=jobserver to fix spurious message about jobserver.
• Simplify arg validation in Freeverb,c (found while consolidating some error messages).
• Check for deinited SPI in SPI try_lock().
• Add some internal IPv4 address conversion functions in IPv4Address. Simplify some conversion code.
• Be more precise with uses of "host", "address", and "port" in socket documentaiton.
• Add an elem_print_helper() routines in shared_bindings/util.c to make printing objects easier. Used to give more info when printing a Network object.
• Add port_malloc_check() and port_realloc_check() routines. These are not used now, but were added when it looked like I needed to do port mallocs for talking to the AirLift.

There were a lot more commits with some dead ends and snapshot commits but I rebased to only a few.