Use PyObject_TypeCheck in type_call instead of IsSubtype equality test

Summary:
In D14351113, the equality check in IsSubtype was introduced as an optimization for the hot path of instance creation.

When upstreaming it ([bpo-45697](https://bugs.python.org/issue45697)) it was pointed out that it's preferable to use the `PyObject_TypeCheck` in the hot callsite
(which performs an inline equality check before calling to IsSubtype) instead of adding a test to all paths
(including those that already use `PyObject_TypeCheck`, hence would hit the equality check twice).

This diff ports the [merged upstream PR](python/cpython#29392) to cinder.

Test Plan:
`make testcinder` passes

and sandcastle

Reviewers: mpage, carljm, #python_efficiency, #cinder

Reviewed By: carljm

Subscribers: orvid

Differential Revision: https://phabricator.intern.facebook.com/D32175230

Tasks: T104755769

Signature: 32175230:1636152684:21f153f68470c360b16a4a567fa52cb49c375148

Objects/typeobject.c

@@ -1137,7 +1137,7 @@ type_call(PyTypeObject *type, PyObject *args, PyObject *kwds)
 
     /* If the returned object is not an instance of type,
        it won't be initialized. */
-    if (!PyType_IsSubtype(Py_TYPE(obj), type))
+    if (!PyObject_TypeCheck(obj, type))
         return obj;
 
     type = Py_TYPE(obj);
@@ -1531,10 +1531,6 @@ PyType_IsSubtype(PyTypeObject *a, PyTypeObject *b)
 {
     PyObject *mro;
 
-    if (a == b) {
-        return 1;
-    }
-
     mro = a->tp_mro;
     if (mro != NULL) {
         /* Deal with multiple inheritance without recursion