[TRITON-MLIR][BACKEND]support atomic_cas (#914)

1. support atomics-cas
2. add xchg support in atomic_rmw

Co-authored-by: dongdongl <dongdongl@nvidia.com>

python/tests/test_core.py

@@ -700,6 +700,16 @@ def test_tensor_atomic_rmw(axis, device="cuda"):
 #     serialized_add[(64,)](data, Lock)
 #     triton.testing.assert_almost_equal(data, ref)
 
+def test_simple_atomic_cas():
+    # 1. make sure that atomic_cas changes the original value (Lock)
+    @triton.jit
+    def change_value(Lock):
+        tl.atomic_cas(Lock, 0, 1)
+
+    Lock = torch.zeros((1,), device='cuda', dtype=torch.int32)
+    change_value[(1,)](Lock)
+
+    assert (Lock[0] == 1)
 
 # # ---------------
 # # test cast