Porting mpn_bdiv_q_1

[Niels Fröhling]

Niels Fröhling wrote:

> Maybe I don't understand the conditions of the computation: "binvert_limb" works
> for all odd divisors?

  I believe I found the algorithm of "Improved division by invariant integers" 
to be in "mpn_divrem_1", is that correct? That's at least what I was looking for.

  I understand the "exact"-functions are for divisions with resulting remainder 
0, what are the "b"-functions for (like "mpn_bdiv_q_1")?

  The number of different functions is pretty confusing. :)

  Thanks
   Niels