Update on "move has_torch_function to C++, and make a special case ob…

…ject_has_torch_function"


This PR pulls `__torch_function__` checking entirely into C++, and adds a special `object_has_torch_function` method for ops which only have one arg as this lets us skip tuple construction and unpacking. We can now also do away with the Python side fast bailout for `Tensor` (e.g. `if any(type(t) is not Tensor for t in tensors) and has_torch_function(tensors)`) because they're actually slower than checking with the Python C API.

Test plan: Existing unit tests. Benchmarks are in #48966

[ghstack-poisoned]

torch/csrc/utils/disable_torch_function.cpp

@@ -160,11 +160,10 @@ static bool is_basic_python_type(PyTypeObject *tp)
 }
 
 inline bool has_torch_function_attr(PyObject* obj) {
-  auto attr = PyObject_GetAttrString(obj, "__torch_function__");
+  auto attr = PyObject_FastGetAttrString(obj, "__torch_function__");
   return (
-    // NOLINTNEXTLINE(modernize-use-nullptr)
-    attr != NULL &&
-    attr != torch::disabled_torch_function);
+    attr.ptr() != nullptr &&
+    attr.ptr() != torch::disabled_torch_function);
 }
 
 inline bool sequence_has_torch_function(PyObject* args) {