cranelift: Align Scalar and SIMD shift semantics (#4520)

* cranelift: Reorganize test suite

Group some SIMD operations by instruction.

* cranelift: Deduplicate some shift tests

Also, new tests with the mod behaviour

* aarch64: Lower shifts with mod behaviour

* x64: Lower shifts with mod behaviour

* wasmtime: Don't mask SIMD shifts

cranelift/codegen/src/isa/aarch64/lower/isle.rs

@@ -335,7 +335,9 @@ where
     }
 
     fn shift_mask(&mut self, ty: Type) -> ImmLogic {
-        let mask = (ty.bits() - 1) as u64;
+        debug_assert!(ty.lane_bits().is_power_of_two());
+
+        let mask = (ty.lane_bits() - 1) as u64;
         ImmLogic::maybe_from_u64(mask, I32).unwrap()
     }