Lmic plus rfm95w staying awake for long periods

End Devices (Nodes)
37

I don’t know if rollover occurred yet; it seems the debug print is just thinking it’s a signed value, rather than an unsigned value. For an unsigned number, the value seems to have passed half its maximum then. But sure: at some point it will rollover.

Ah, forget the negative printed values, these are the lines that you’re referring to:

34282: engineUpdate, tx at 33716; now is: 34282
...
4294933165: engineUpdate, tx at 33981; now is: 4294933165
...
4294933470: engineUpdate, tx at 34328; now is: 4294933470
...
4294933777: engineUpdate, tx at 4294933170; now is: 4294933777

Are all times in each line divided by OSTICKS_PER_SEC?

Something I noticed which is a bit weird, but might not affect things, os_getTime is defined as:

ostime_t os_getTime () {
    return hal_ticks();
}

…with hal_ticks expected to return an unsigned 32 bits number:

u4_t hal_ticks () {
    ...
}

…but ostime_t being a signed 32 bits number:

typedef uint32_t           u4_t;
typedef int32_t            s4_t;
...
// the HAL needs to give us ticks, so it ought to know the right type.
typedef              s4_t  ostime_t;