RFC: Avoid calling XXX_skip_prologue for assembly code

[Fred Fish]

> Does anyone know the motivation behind the patch?

I ran into a target for which the skip_prologue code did not behave
correctly for assembly language functions.  The details are hazy now
but as I recall the skip prologue function always returned PC+2 in
such cases, instead of just returning PC.  In looking at some of the
other XXX_skip_prologue functions it appeared to me that they would
have the same problems under similar conditions.  Ideally we would
just fix all the skip_prologue functions to do the right thing, but then
I began to question why we were even calling skip_prologue anyway on
functions that were already known to have been written in assembly
language.

-Fred