machine.reset_cause 2

[konomikitten]

I keep getting reset_cause of 2 from one of my boards, does anyone know what this means? It's not defined in any of the reset cause constants.

MicroPython v1.24.1-1.g50c8864e7f on 2024-12-18; LOLIN D1 MINI with ESP8266

[Josverl]

The reset_cause value of 2 on your LOLIN D1 MINI with ESP8266 typically indicates a watchdog reset.
edit : this is not correct
This happens when the system's watchdog timer detects that the program is not running as expected, possibly due to an infinite loop, blocking code, or other issues causing the system to hang.

To troubleshoot:

1. Check your code for any long-running loops or blocking operations that might prevent the watchdog timer from being reset.
2. Use machine.WDT() in MicroPython to manage the watchdog timer explicitly, if needed.

[konomikitten]

That's useful to know, but why isn't it set as a constant? There's already a watchdog constant defined but it's a value of 1 (machine.WDT_RESET) not 2?

[Josverl]

not sure ,
I note that you are using a custom or preview build (...-g50c8864e7f) ; is that the same build as the other boards ?
do you see the same with one of the released firmwares ? ( v1.24.1 / v1.25.0) ?
what code is running before the reset ?

[Josverl]

I think these are the values from : xtensa-lx106-elf/xtensa-lx106-elf/sysroot/usr/include/user_interface.h

enum rst_reason {
    REASON_DEFAULT_RST        = 0,
    REASON_WDT_RST            = 1,
    REASON_EXCEPTION_RST    = 2,
    REASON_SOFT_WDT_RST       = 3,
    REASON_SOFT_RESTART     = 4,
    REASON_DEEP_SLEEP_AWAKE    = 5,
    REASON_EXT_SYS_RST      = 6
};

(some) additional information in this esp8266_reset_causes_and_common_fatal_exception_causes pdf

[konomikitten]

Looks like it's missing from the build https://github.com/micropython/micropython/blob/master/ports/esp8266/modmachine.c#L55-L61. There seems to be a pull to add extra reset reasons but it wasn't merged? #2994

[Josverl]

I think the challenge is striking the right balance between a common consistent API across ports and boards with dissimilar error and return codes, while keeping the firmware small.

There is nothing stopping you from defining

REASON_EXCEPTION_RST = const(2)

in your code, and using that to compare
Or add it as a comment.

It may be useful to add a link to the doc above to the esp8266 quickref.