No subf surprise here
Remember how integer subtraction used subf and reversed its operands? The floating-point unit has a real fsubs that computes fD = fA - fB in the natural order:
fsubs f1, f1, f2 # f1 = a - b
blr
Destination, then the operands left-to-right exactly as written in the C. The floating-point instruction set is generally more regular than the integer one — one fewer idiom to memorize.
Your task
Write sub_f to compile to the fsubs above.