Make good use of the dual cores of the RP2350(PICO2) to assist IRQ

[MasterYuan418]

Port: Raspberry Pico 2 W(RP2350)

Hi there,

I'm working on a project which is about receiving data from UART. I use UART triggers to receive data, but it seems like that RP2350 appears to be lagging (REPL responds very slowly to my interrupt commands).

Now I'm considering if I can manually set IRQ handlers running on a specific core, in order to reduce the lag(maybe?). I googled, and seems like Pico SDK can disable/enable a core for IRQ. But I cannot find anything tutorials about how to choose a core manually.

Is there are something wrong in my idea or I missed something in document?

Thanks.

[GitHubsSilverBullet]

It seems as if the UART.IRQ_RXIDLE happens about three chartimes after the last data has been received.
I don't believe that using the other core will make any difference,
with the possible exception when you write your own UART-RX system.

[robert-hh]

responds very slowly to my interrupt commands

What is the response time you notice?

[MasterYuan418]

I apologize, but I may be wrong about the UART IRQ. it may be that I wrote the code in question all the way into main.py (using the timer to poll for data). I guess this causes the REPL to respond slowly.

For the response time, it might have been well over 5s (estimated, but pico are slow to respond to Interrput)

Here is my code(Simplified & Abridged)

// in main.py
def receive_msg():
    pass
def msg_queue(t: Timer):
    // receive something from another device by UART
if __name == "__main__":
    device.uart.irq(msg_queue, UART.RX_IDLE, hard=False)
    timer.init(period=1000, mode=Timer.PERIODIC, callback=msg_queue)
    while True:
        do_something()
        // do something else, pretty slow

Anyway, thank for the reply!

[peterhinch]

It would help us to offer suggestions if you told us more about the data source:

• What is the baudrate?
• How frequently do messages arrive?
• Do they arrive in response to an outgoing message or are they unsolicited?
• Are messages fixed length or variable length with delimiter (e.g. newline).
• What is the maximum message length.

[MasterYuan418]

It would help us to offer suggestions if you told us more about the data source:

* What is the baudrate?

* How frequently do messages arrive?

* Do they arrive in response to an outgoing message or are they unsolicited?

* Are messages fixed length or variable length with delimiter (e.g. newline).

* What is the maximum message length.

• The data comes from a LoRA Module, while the sender send message every second(both receiver and sender are in the same room).
• Baudrate: 9600
• Yes, it comes by what micropython sent(also for every second, for example, micropython: PING will cause a message PONG reply)
• With \r\n
• less than 80 bytes.

Here is a real example(M means micropython, and R means receiver reply. There are some modifications, but that's about it):

M: STATUS,0,0,0,0,0,1.00,OK@
M: CMD,<......>,OK@
R: REPLY,STATUS,OK@
M: HEARTBEAT,1755782387,1@
R: REPLY,HEARTBEAT,OK@

As I write this, I carefully reviewed the code, and this may be caused by delays between multiple modules working together (such as GNSS reception and LORA half-duplex transmission). Perhaps my question is silly, but thank you very much for your reply :)

[peterhinch]

OK, this is quite undemanding.

The easy way to do this is to forget about interrupts and threading. Use asyncio.StreamReader, specifically the readline() method. See this example: this has a Pyboard concurrently sending and receiving messages. See tutorial and docs for general information about asyncio.

There is a bit of a learning curve, but asyncio really is the easiest way to interface UARTS and to have other code run concurrently.

[MasterYuan418]

Ok I'll try again later. Thank for guide!