%1 = "tf.While"(%0) {
      body = @while_body, cond = @while_cond, device = "", is_stateless = false}
         : (tensor<*x!tf_type.resource<tensor<f32>>>) -> (tensor<*x!tf_type.resource<tensor<f32>>>)
    tf_device.return
  }) {cluster_attr = "cluster_attr"} : () -> ()
  func.return
}

  %2 = "tf.LogicalAnd"(%0, %1) {device = "/job:localhost/replica:0/task:0/device:CPU:0"} : (tensor<i1>, tensor<i1>) -> tensor<i1>
  return %2 : tensor<i1>
}

// CHECK-LABEL: map/while_body
func.func private @"map/while_body"(%arg0: tensor<i32>, %arg1: tensor<i32>, %arg2: tensor<!tf_type.variant<tensor<*xf32>>>, %arg3: tensor<?xf32>) -> (tensor<i32>, tensor<i32>, tensor<!tf_type.variant<tensor<*xf32>>>, tensor<?xf32>) {

  func.return
}
// CHECK: func @while_body(%[[BARG0:.*]]: tensor<!tf_type.resource<tensor<5x3xf32>>>, %[[BARG1:.*]]: tensor<i32>, %[[BARG2:.*]]: tensor<!tf_type.resource<tensor<5x3xf32>>>)
func.func @while_body(%arg0: tensor<!tf_type.resource>, %arg1: tensor<i32>) -> (tensor<!tf_type.resource>, tensor<i32>) {

variable ops, thus allowing them to run in parallel across iterations.

Before:

```mlir
!tf_res = type tensor<!tf_type.resource<tensor<f32>>>
func @while_body(%arg0: !tf_res, %arg1: !tf_res, %arg2: tensor<f32>, %arg3: tensor<f32>) -> (!tf_res, !tf_res, tensor<f32>, tensor<f32>) {
  %graph:4 = tf_executor.graph {

  // Use this to clone an op and lift it outside its parent function. The
  // original while body is unchanged. Example:
  // Original:
  //    %x = tf.while(%a, %b)
  //    ...
  //    while_body:
  //       call(f=@sc_fw, %arg0, %arg1)
  // Lifted:
  //    call(f=@sc_fw, %a, %b)
  //    %x = tf.while(%a, %b)
  //    ...
  func::FuncOp orig_parent_func =