Oh boy,

I forgot the "."'s after "acc". And the way I wrote it relies on it that the accu is not affected by the "work" subroutines, which is probably not going to be the case, so do this:

mov tempp1,p1
bit0: jnb tempp1.0,bit1
call work0
bit1: jnb tempp1.1,bit2
call work1
...

or test the bits of p1 directly.

What is best for your case depends on the nature of your project.
By the way, this is a classic case of what is called "polling".
And remember: the sitation on the real p1 may chenge while the CPU is busy doing some "work". The CPU can only execute 1 instruction at a time.