During forward bias : An electron from the negative terminal of battery enters the conduction band of $n$ side, runs toward the junction, waits for a hole for $\tau$ seconds, when it finds one, it falls into the hole in the valence band of $p$ side, then the electron hops along the $p$ side holes and finally reaches the positive terminal of the battery.
Question1 : Why does the electron in the conduction band fall into the hole in the valence band ? Is it because the valence band has lower energy ? If so, kindly keep this in mind when looking into my second question.
During reverse bias, during recovery phase : An electron from the negative terminal of battery enters the $p$ side valence band, runs toward the junction, then what ? There are no empty energy states available in the valence band of $n$ side, so it seems the electron must climb up to the conduction band to complete the loop. ( Note that if we abruptly change from forward bias to reverse, initially, during recovery phase, there exist many diffused electrons on the $p$ side near the junction. )
Question2 : Does the electron in the valence band really climb up to the conduction band at the junction ?