how to lock uaysnc server while a task is in use

[GM-Script-Writer-62850]

I have a W5500 PCIO controller, and i have core 1 doing things at time, if i have activity on core 1 when i get input on the server things go south without any error, when i have core 1 active i have a uasyncio.Lock(), how can i have this respect the lock (when do i acquire/release it?)

async def webAPI(reader, writer):
	request_line = await reader.readline()
	print("\nClient connected:",request_line)
	if(request_line == b""):
		writer.write('HTTP/1.0 204 No Content')# is this right?
		await writer.drain()
		await writer.wait_closed()
		return print("Disconnect")

	length=0
	header=await reader.readline()
	while header != b"\r\n":
		if(header != b""):
			line=header.decode("utf-8")
			if line[0:15] == "Content-Length:":
				length=int(line[15:])
				print("Content length:",length)
			else:
				pass#print('Ignore header:',line)
		header=await reader.readline()

	request=request_line.decode("utf-8").split(' ')
	if request[0]=="POST":
		data=await reader.readexactly(length)
		data=data.decode("utf-8")
		data=json_decode(data)
		print(data)
		if data["action"] == "buzzer":
			uasyncio.create_task(buzzer(data["value"]))
		elif data["action"] == "blink":
			if not data["value"]:
				notifyControlActive=0
			elif not notifyControlActive:
				notifyControlActive=1
				uasyncio.create_task(notifyControl())
		writer.write('HTTP/1.0 200 OK\r\nContent-type: text/plain\r\nContent-Length: 2\r\nCache-Control: no-cache\r\n\r\nOK')
	else:
		writer.write("")
	try:
		await writer.drain()
		await writer.wait_closed()
		print("Client disconnected")
	except Exception as e:
		print("Error:",e)
uasyncio.create_task(uasyncio.start_server(webAPI, "0.0.0.0", 80))

[peterhinch]

asyncio.Lock is not threadsafe. You need to use _thread.allocate_lock for a threasafe lock. You may want to read this doc which lists some of the issues linking asyncio code and threaded code.

[bixb922]

Yes, that's true, and asyncio can run many, many tasks on that single core 1. It's not easy to tame that second core, so personally I'd use the second core only when running out of cpu on the first core, or for example for long running cpu intensive tasks.

[GM-Script-Writer-62850]

hence the reason i run urequest on the second core as it take a relatively long time period

i have used my second core for stuff like polling or directly driving a 7 segment display in software and then peterhinch informed me of a way easier way i can deal with polling and switch bounce and not need my second core (in other words i can use it for something else) i would prefer to not poll my server for data with core 1 like a child saying 'are we there yet' (just need to get data sent from another controller on the network)

[bixb922]

OK, now I understand the need

[GM-Script-Writer-62850]

i did manage to find this while looking around just now, maybe i could use that on core 0

i should probably make a bug report since i can't find one... micropython/micropython-lib#700

[peterhinch]

That async HTTP client was last updated four years ago and the docs say

Development of the uasyncio library for MicroPython seems to be fragmented and in flux a this time. This project is tested against the uasyncio from https://pypi.org/.

That observation is very out of date as is the uasyncio version on PyPi. The author claims it works on CPython so it should work on current uasyncio. I would definitely avoid the one on PyPi.