finding an esp32 wakeup source

[krismc]

I normally use machine.reset_cause to log deepsleep wakeups but my current project also has a esp32.wake_on_ext0 so I can't tell the difference between a time expired wake from deepsleep & a pin triggered wakeup.
Googling produced some mention of a esp32.wake_reason() but I can't find it the upython docs & it's not on my esp32. I guess I could log the time difference between going to deepsleep & waking to tell if it's pin triggered or not but it feels clumsy. Does anyone know a better way to pick the difference?

[robert-hh]

There is as well machine.wake_reason(), which can tell the difference between WLAN_WAKE, PIN_WAKE and RTC_WAKE.

[krismc]

Tnx Rob, that works. I'm countin' myself lucky since it's the only one that does!

>>> machine.PIN_WAKE
2
>>> machine.WLAN_WAKE
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute 'WLAN_WAKE'
>>> machine.RTC_WAKE
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute 'RTC_WAKE'

[scruss]

cross-platform wakeup handling is kind of inconsistent. The ESP32-S3 in front of me (v1.25.0) knows about (machine.): EXT0_WAKE EXT1_WAKE PIN_WAKE TIMER_WAKE TOUCHPAD_WAKE ULP_WAKE

On rp2, the test function is machine.reset_cause(), with machine.PWRON_RESET and machine.WDT_RESET being the returned values.

[krismc]

Yeah the constants seem fairly variable. The esp32 I'm using returns 0,1,2 depending on the wakeup source & I think I've figured out now which is which. I don't really get the idea of names for constants when the chip is returning numbers. I just use a dict to get a name from the returned number. The fact the numbers have different meanings in the repl than when it's headless don't make it any easier!

[scruss]

don't think they should have different meanings in the REPL from any other application. Are you sure?

On this S3, I see:

2 machine.EXT0_WAKE
2 machine.PIN_WAKE
3 machine.EXT1_WAKE
4 machine.TIMER_WAKE
5 machine.TOUCHPAD_WAKE
6 machine.ULP_WAKE

I think it's supposed to be used something like:

if machine.wake_reason()==machine.TOUCHPAD_WAKE: ...

but I agree, it's hard to use when you're not sure what reset_cause() is likely to return. It took me a while to understand the rp2's weirdness.

[krismc]

Well I thought I was sure, so much so that I had 2 different lookup dics, one for when I was running from the cmd line in the repl & the other for when I was running a program from the esp32. However after multiple cycles of

from machine import deepsleep,reset_cause,RTC; rtc=RTC(); print(rtc.memory()); rtc.memory(str(reset_cause())); deepsleep(1000) 

both ways I can't force the same anomaly. Will report back if it occurs again organically.