LM22671 based power supply output voltage swing issue, please help

This looks like high frequency noise from the switching operation.  Most likely you started to see it when the part went from discontinuous to continuous conduction mode.

The noise comes from two places.

1)  The fast rise time of the switch node couples through the parasitic capacitance of the inductor.  This noise then rings on the inductance of the output capacitors.  You can improve this by: 1: Using and inductor with a higher self resonant frequency for the same inductance value. 2: Improve the routing of your output capacitors so that they have a lower inductance path to ground 3) Adding a ground plane to return the circulating currents. 4) Adding some small 0603  COG 0.047uf capacitors.  The small case size has low inductance and can bypass much higher frequencies

2) Layout of the input capacitor and the recirculating diode.  During the switching operation the current changes between the high side fet and the diode every switching cycle.  If not bypassed very closely this will cause serious ground bounce in your circuit which will be seen every switching cycle.  To improve this make sure that the loop created by the Cin capacitor to the input of the 22671, then from switch node to the diode then back to the ground of the Cin capacitor is as small as possible.  Ideally you want less than 1mm from the capacitor ground to the diode anode.

I