cpu/esp32/stdio_usb_serial_jtag: Fix to ESP32 stdio usb serial hanging if it receives data too quickly

[IsikcanYilmaz]

Contribution description

When the stdio serial usb jtag module of the ESP32 receives data at (relatively) high speeds, there's a chance its tx fifo won't be flushed and/or the correct irq flag won't be cleared. This leads to nothing being printed through the usb stdio anymore, and (most instances) data being unable to be received by the device, rendering the shell unusable.

This PR clears the IRQ flags at the end of the ISR and flushes the fifo. With this change I have not been able to repro the bug in question.

Testing procedure

• Use an ESP32 that has stdio_usb_serial_jtag support. (I've been using a seeedstudio xiao esp32s3)
• Load up a basic example that uses the shell module, like examples/default
• Wail on the shell using the following python code, or anything equivalent:

#!/usr/bin/env python3
import serial, time
ser = serial.Serial("/dev/tty.usbmodem111301")
count = 10
while(True):
    ser.write("help\n".encode())
    time.sleep(0.001)
    count -= 1
    if (count == 0):
        break
print("Writer done")

• Observe that the shell is no longer responsive
• This PR fixes this issue

Issues/PRs references

[riot-ci]

Murdock results

✔️ PASSED

a21d302 cpu/esp32/stdio_usb_serial_jtag: clear all usb serial jtag flags and flush the tx fifo at the end of usb serial isrs

Success	Failures	Total	Runtime
10215	0	10215	19m:11s

Artifacts

• Documentation preview

[mguetschow]

Thanks @IsikcanYilmaz for your PR and welcome to the community! :)

Unfortunately, I don't have hardware at hand to test this, but I'm sure someone else will.

[benpicco]

Thank you for the fix, I've stumbled upon this several times before!
Works like a charm now.