【NewIR】add share_data_op (#57212)

* add share_data_op

* fix share_data op

Author: yangguohao

paddle/fluid/pir/dialect/operator/ir/ops.yaml

@@ -163,6 +163,13 @@
   args : (Tensor[] x, bool[] share_dims_and_dtype = {})
   output : Tensor[](out){x.size()}, Tensor[](xout){x.size()}
 
+- op : share_data
+  args : (Tensor x)
+  output : Tensor(out)
+  kernel:
+    func: share_data
+    param: [x]
+
 - op : write_to_array
   args : (Tensor i, Tensor x)
   output : Tensor[](out)

paddle/fluid/pir/dialect/operator/utils/utils.cc

@@ -34,7 +34,8 @@ const std::unordered_set<std::string> LegacyOpList = {
     "pd_op.c_reduce_sum_",
     "pd_op.c_allreduce_max_",
     "pd_op.c_allgather",
-    "pd_op.seed"};
+    "pd_op.seed",
+    "pd_op.share_data"};
 
 enum class AttrType {
   UNDEFINED = 0,

paddle/phi/api/yaml/op_compat.yaml

@@ -3191,6 +3191,12 @@
       data_type : int64_t
       tensors_name : StepsTensorList
 
+- op: share_data
+  inputs :
+    x : X
+  outputs :
+    out : Out
+
 - op: sigmoid_cross_entropy_with_logits
   backward: sigmoid_cross_entropy_with_logits_grad
   inputs :