aarch64: Implement ireduce for i128 values

cranelift/filetests/filetests/runtests/i128-reduce.clif

@@ -0,0 +1,43 @@
+test run
+target aarch64
+target x86_64 machinst
+
+function %ireduce_128_64(i64, i64) -> i64 {
+block0(v0: i64, v1: i64):
+    v2 = iconcat v0, v1
+    v3 = ireduce.i64 v2
+    return v3
+}
+; run: %ireduce_128_64(0, 0) == 0
+; run: %ireduce_128_64(-1, -1) == -1
+; run: %ireduce_128_64(0xC0FFEEEE_DECAFFFF, 0xDECAFFFF_C0FFEEEE) == 0xC0FFEEEE_DECAFFFF
+
+function %ireduce_128_32(i64, i64) -> i32 {
+block0(v0: i64, v1: i64):
+    v2 = iconcat v0, v1
+    v3 = ireduce.i32 v2
+    return v3
+}
+; run: %ireduce_128_32(0, 0) == 0
+; run: %ireduce_128_32(-1, -1) == -1
+; run: %ireduce_128_32(0xC0FFEEEE_DECAFFFF, 0xDECAFFFF_C0FFEEEE) == 0xDECAFFFF
+
+function %ireduce_128_16(i64, i64) -> i16 {
+block0(v0: i64, v1: i64):
+    v2 = iconcat v0, v1
+    v3 = ireduce.i16 v2
+    return v3
+}
+; run: %ireduce_128_16(0, 0) == 0
+; run: %ireduce_128_16(-1, -1) == -1
+; run: %ireduce_128_16(0xC0FFEEEE_DECAFFFF, 0xDECAFFFF_C0FFEEEE) == 0xFFFF
+
+function %ireduce_128_8(i64, i64) -> i8 {
+block0(v0: i64, v1: i64):
+    v2 = iconcat v0, v1
+    v3 = ireduce.i8 v2
+    return v3
+}
+; run: %ireduce_128_8(0, 0) == 0
+; run: %ireduce_128_8(-1, -1) == -1
+; run: %ireduce_128_8(0xC0FFEEEE_DECAFFFF, 0xDECAFFFF_C0FFEEEE) == 0xFF