Use u32 in Type API (#4280)

Move from passing and returning u8 and u16 values to u32 in many of
the functions. This removes a number of type conversions and gives
a small compilation time speedup, around ~0.7% on my aarch64 machine.

Copyright (c) 2022, Arm Limited.

cranelift/codegen/src/machinst/helpers.rs

@@ -7,7 +7,7 @@ use std::ops::{Add, BitAnd, Not, Sub};
 
 /// Returns the size (in bits) of a given type.
 pub fn ty_bits(ty: Type) -> usize {
-    usize::from(ty.bits())
+    ty.bits() as usize
 }
 
 /// Is the type represented by an integer (not float) at the machine level?