It's been about 2 weeks of a couple of hours a day trying to get this to work
after so long, my experience tells me that you know what it is but it has just not registered. Here is what I would do if I had foght such a problem for too long modified for 'no emulator'(I guess you do not have an emulator, you have not stated so)

I assume you gave checked in a simulator and found no difference.

This looks like "grabbing for straws" but the message is somewhere.

a) do a global search through your code for P1 and all derivatives thereof.
b) do a global search through your code for P3 and all derivatives thereof - did you iss changing one
c) run the code in the P3 version with a temporary jmp $/while (1); just before the first 1 wire access of the pin, what is the pin level - then do same for the p1 version. I know you "know" that will be the same, but experience tell me that you should check anyway.

often, when changing a port pin some change is missed, typically a mov Px,#.. may be missed because all searches was for px.n

d) change the pullup resistor. I have never had reason to think so, but maybe, on your chip, the internal pullup is different for the 2 ports (the tolerance is something like -50+100%)
e) go to your favorite locale, absorb some golden, foaming liquid, have a good nights sleep, come in the next morning and try everything again.

Erik