riscv64: Add masking for small types when lowering select (#5504)

When lowering `select+icmp` we have an optimization that allows us to
avoid materializing the icmp result.

We were accidentally not masking the high bits for i8 and i16 in this case.

Issue #5498 reported this as an illegal instruction but what was happening
there was that the invalid select caused a division by zero.

cranelift/codegen/src/isa/riscv64/lower.isle

@@ -615,8 +615,10 @@
   (gen_select ty (truthy_to_reg cty (normalize_cmp_value cty c)) x y))
 
 (rule 1
-  (lower (has_type ty (select (icmp cc a b) x y)))
-  (gen_select_reg cc a b x y))
+  (lower (has_type ty (select (icmp cc a b @ (value_type in_ty)) x y)))
+  (let ((a Reg (normalize_cmp_value in_ty a))
+        (b Reg (normalize_cmp_value in_ty b)))
+    (gen_select_reg cc a b x y)))
 
 ;;;;;  Rules for `bitselect`;;;;;;;;;