When the device was suspended and the pin level changed during
that time, the pin level of course isn't updated in the pins cb_data.
Once the device is resumed, this leads to potentially having
a wrong value in the pin state data leading to swallowing the first
event due to comparing the stored level vs. the new level before
reporting.
Also added some `const`s and deleted an unused struct element.
Signed-off-by: Marcel Krüger <marcel@mkgr.dev>
