// CHECK: %[[VAR_HANDLE_0:.*]] = "tf.VarHandleOp"()
// CHECK-SAME: {{.*shared_name = "var_0".*}}
// CHECK: %[[VAR_HANDLE_1:.*]] = "tf.VarHandleOp"()
// CHECK-SAME: {{.*shared_name = "var_1".*}}

// ReadVariableOps are inserted for each VarHandleOp to read the tensor values.
// CHECK-DAG: %[[READ_VARIABLE_0:.*]] = "tf.ReadVariableOp"(%[[VAR_HANDLE_0]]) : (tensor<!tf_type.resource<tensor<2xf32>>>) -> tensor<2xf32>

  if (auto pop_back = dyn_cast<TF::TensorListPopBackOp>(op)) {
    return InferShapeForTensorListPopBackOp(pop_back);
  }

  if (auto var_handle_op = dyn_cast<VarHandleOp>(op)) {
    return InferShapeForVarHandleOp(var_handle_op);
  }

  if (auto xla_reduce_window_op = dyn_cast<XlaReduceWindowOp>(op)) {
    return InferShapeForXlaReduceWindowOp(xla_reduce_window_op);
  }

      return signalPassFailure();
    }

    auto var_handle_ops =
        init_func_op.getBlocks().front().getOps<VarHandleOp>();
    llvm::SmallVector<VarHandleOp, 4> init_vars(var_handle_ops.begin(),
                                                var_handle_ops.end());
    RemoveVariables(init_vars);
  }
}

}  // namespace

std::unique_ptr<OperationPass<ModuleOp>>

shared_name: the name by which this variable is referred to.
dtype and shape: attributes representing the data type and shape held in the
  variable.

Example:
    resource_variable_ops.var_handle_op(
          dtype=dtypes.int32, shape=[8, 16], container="foo", shared_name="bar")
  returns a handle for a variable with name "bar" in container "foo", and the
  variable holds a tensor of shape [8, 16] and dtype int32.
  }];

                                           ImmediateTensorHandlePtr* handle) {
  ImmediateOpPtr varhandle_op(ctx->CreateOperation());

  TF_RETURN_IF_ERROR(varhandle_op->Reset("VarHandleOp", raw_device_name));
  TF_RETURN_IF_ERROR(varhandle_op->SetAttrType("dtype", dtype));

  // Note that if shape is unknown rank, shape.dim_sizes() will be empty, and
  // shape.dims() will be -1.

  attributes {tf_saved_model.exported_names = ["main"]} {
  // CHECK: tf.VarHandleOp
  %0 = "tf.VarHandleOp"() {device = "/device:CPU:0", container = "", shared_name = "variable1"} : () -> tensor<!tf_type.resource<tensor<1x3xf32>>>
  %1 = "tf.VarHandleOp"() {device = "/device:CPU:0", container = "", shared_name = "variable2"} : () -> tensor<!tf_type.resource<tensor<1x3xf32>>>

    %0 = "tf.VarHandleOp"() {_class = ["loc:@dense/kernel"], allowed_devices = [], container = "", device = "/job:worker/replica:0/task:1/device:CPU:0", shared_name = "var1"} : () -> tensor<!tf_type.resource<tensor<100x50xf32>>>
    %1 = "tf.ReadVariableOp"(%0) {device = ""} : (tensor<!tf_type.resource<tensor<100x50xf32>>>) -> tensor<100x50xf32>

    // CHECK: "tf.VarHandleOp"
    func.return
  }
}

// -----

// Test mutation detection propagates across function calls.

module {
  func.func @f() -> tensor<0xf32> {
    // CHECK: "tf.VarHandleOp"

// CHECK-NOT: "tf.VarHandleOp"
func.func @no_args() {
  %0 = "tf.VarHandleOp"() {container = "", shape = "tfshape$", shared_name = "x"} : () -> tensor<!tf_type.resource<tensor<f32>>>
  func.return
}

// CHECK-LABEL: func @some_args
// CHECK-SAME: (%arg0: tensor<i1>, %arg1: tensor<!tf_type.resource<tensor<f32>>> {tf.resource_name = "x"})
// CHECK-NOT: "tf.VarHandleOp"
func.func @some_args(%arg0: tensor<i1>) {

  attributes {tf_saved_model.exported_names = ["handles_ambiguous_var_handles"]} {
    // CHECK: VarHandleOp
    %0 = "tf.VarHandleOp"() {container = "", shared_name = "foo"} : () -> tensor<!tf_type.resource<tensor<10xf32>>>
    func.call @assign_resource(%0) : (tensor<!tf_type.resource<tensor<10xf32>>>) -> ()
    %1 = "tf.VarHandleOp"() {container = "", shared_name = "foo"} : () -> tensor<!tf_type.resource<tensor<10xf32>>>