I have an endogenous variable P . In my specification I also have interaction of P with X (X is an exogenous dummy variable). My model looks like

Y=P+PX+U
where P is endogenous and X is exogenous

I need to instrument X with Z. Based on posts on the forum I understood that I need to use

ivregrress Y (P PX= Z ZX)

My doubt lies in interpretation of the two first stage least square regressions I get. I am not able to interpret the two first stage regressions I get:

P=B1Z+B2.ZX
P.X=B3Z+B4.ZX

Any help is deeply appreciated.