It would be good to test exact same (physically) cell both in phone, and (2nd time) via external charger with such function. Maybe something is wrong with calculation of mAh from raw values, reported by chip?

But, bnf (and bq27200.sh) are using i2c to access raw values reported by chip, then, bnf calculates mAh internally, always doing it the same way. It doesn't depend on newest charging/measuring kernel modules by Pali, so updates to kernel-power shouldn't affect it (in fact, it doesn't require KP at all).

So, no idea what could have changed. For sure, our internal chips didn't, scripts too... WTF?

/Estel