solve the bug of res error when axis or shift is a number

docs/api/api_python/ops/mindspore.ops.func_roll.rst

@@ -8,11 +8,11 @@ mindspore.ops.roll
     参数：
         - **x** (Tensor) - 输入Tensor。
         - **shifts** (Union[list(int), tuple(int), int]) - 指定元素移动方式，如果为正数，则元素沿指定维度正向移动（朝向较大的索引）的位置数。负偏移将向相反的方向滚动元素。
-        - **dims** (Union[list(int), tuple(int), int], optional) - 指定需移动维度的轴。默认值：None。
+        - **dims** (Union[list(int), tuple(int), int], optional) - 指定需移动维度的轴。默认值：None。如果 `dims` 为None，则会将输入Tensor展平后再进行roll计算，然后将计算结果reshape为输入的shape。
 
     返回：
         Tensor，shape和数据类型与输入 `x` 相同。
 
     异常：
         - **TypeError** - `shifts` 不是int、tuple或list。
-        - **TypeError** - `dims` 不是int、tuple或list。
+        - **TypeError** - `dims` 不是int、tuple或list。

mindspore/python/mindspore/ops/function/math_func.py

@@ -6133,7 +6133,8 @@ def roll(x, shifts, dims=None):
             positively (towards larger indices) along the specified dimension. Negative shifts will roll the elements
             in the opposite direction.
         dims (Union[list(int), tuple(int), int], optional): Specifies the dimension indexes of shape to be rolled.
-            Default: None.
+            Default: None. If `dims` is None, the Tensor will be flattened before rolling and then restored to the
+            original shape.
 
     Returns:
         Tensor, has the same shape and type as `x`.
@@ -6155,7 +6156,10 @@ def roll(x, shifts, dims=None):
         >>> print(output)
         [3. 4. 0. 1. 2.]
     """
-    dims = dims if dims is not None else 0
+    _shape = x.shape
+    if dims is None:
+        flatten_x = x.reshape(-1)
+        return Roll(shifts, 0)(flatten_x).reshape(_shape)
     return Roll(shifts, dims)(x)
 
 

mindspore/python/mindspore/ops/operations/math_ops.py

@@ -7797,7 +7797,11 @@ class Roll(Primitive):
         """Initialize Roll"""
         if context.get_context("device_target") == "GPU":
             validator.check_value_type("shift", shift, [int, tuple, list], self.name)
+            if not isinstance(shift, (list, tuple)):
+                self.add_prim_attr('shift', [shift])
             validator.check_value_type("axis", axis, [int, tuple, list], self.name)
+            if not isinstance(axis, (list, tuple)):
+                self.add_prim_attr('axis', [axis])
         else:
             if isinstance(shift, (tuple, list)) and isinstance(axis, (tuple, list)):
                 validator.check_equal_int(len(shift), 1, "shift size", self.name)