Convert fma, valltrue & vanytrue to ISLE (AArch64) (#4608)

* Convert `fma`, `valltrue` & `vanytrue` to ISLE (AArch64)

Ported the existing implementations of the following opcodes to ISLE on
AArch64:
- `fma`
  - Introduced missing support for `fma` on vector values, as per the
    docs.
- `valltrue`
- `vanytrue`

Also fixed `fcmp` on scalar values in the interpreter, and enabled
interpreter tests in `simd-fma.clif`.

This introduces the `FMLA` machine instruction.

Copyright (c) 2022 Arm Limited

* Add comments for `Fmla` and `Bsl`

Copyright (c) 2022 Arm Limited

cranelift/codegen/src/machinst/isle.rs

@@ -357,6 +357,15 @@ macro_rules! isle_prelude_methods {
             }
         }
 
+        #[inline]
+        fn ty_vec64(&mut self, ty: Type) -> Option<Type> {
+            if ty.is_vector() && ty.bits() == 64 {
+                Some(ty)
+            } else {
+                None
+            }
+        }
+
         #[inline]
         fn ty_vec128(&mut self, ty: Type) -> Option<Type> {
             if ty.is_vector() && ty.bits() == 128 {
@@ -588,6 +597,14 @@ macro_rules! isle_prelude_methods {
             }
         }
 
+        fn not_vec32x2(&mut self, ty: Type) -> Option<Type> {
+            if ty.lane_bits() == 32 && ty.lane_count() == 2 {
+                None
+            } else {
+                Some(ty)
+            }
+        }
+
         fn not_i64x2(&mut self, ty: Type) -> Option<()> {
             if ty == I64X2 {
                 None