servo not updating on slave (after ~49.7days)

Issue described in wr-switch-sw#240 (closed).

When the switch uptime in miliseconds overflows, ppsi will hit the bug described in #31, wr-switch-sw#240 (closed) and lost sync. Usually, the sync is back after ~1 minute. However, sometimes switch can stuck in a TRACK_PHASE state, with the update counter not increasing.

The problem is that the function(s) returning the delay to the next timeout (pp_next_delay_*) can return wrong value when the timeout's timestamp is after the overflow of the uptime counter. In the described scenario pp_next_delay_* function returns 0 instead of time to the closest timeout.