fdivs
Integer division by a constant got strength-reduced into shifts. Floating-point division has no such trick — fdivs is a genuine hardware divide and the compiler emits it directly:
fdivs f1, f1, f2 # f1 = a / b
blr
Be aware: dividing a float by a constant (say x / 2.0f) is usually turned into a multiply by the reciprocal (x * 0.5f) instead, which is cheaper. You only get fdivs when the divisor is a runtime value.
Your task
Write div_f to match the fdivs above.