Issue #28618: Make hot functions using __attribute__((hot))

When Python is not compiled with PGO, the performance of Python on call_simple
and call_method microbenchmarks depend highly on the code placement. In the
worst case, the performance slowdown can be up to 70%.

The GCC __attribute__((hot)) attribute helps to keep hot code close to reduce
the risk of such major slowdown. This attribute is ignored when Python is
compiled with PGO.

The following functions are considered as hot according to statistics collected
by perf record/perf report:

* _PyEval_EvalFrameDefault()
* call_function()
* _PyFunction_FastCall()
* PyFrame_New()
* frame_dealloc()
* PyErr_Occurred()

Python/ceval.c

@@ -718,7 +718,7 @@ PyEval_EvalFrameEx(PyFrameObject *f, int throwflag)
     return tstate->interp->eval_frame(f, throwflag);
 }
 
-PyObject *
+PyObject* _Py_HOT_FUNCTION
 _PyEval_EvalFrameDefault(PyFrameObject *f, int throwflag)
 {
 #ifdef DXPAIRS
@@ -4771,7 +4771,7 @@ if (tstate->use_tracing && tstate->c_profilefunc) { \
     x = call; \
     }
 
-static PyObject *
+static PyObject* _Py_HOT_FUNCTION
 call_function(PyObject ***pp_stack, Py_ssize_t oparg, PyObject *kwnames)
 {
     PyObject **pfunc = (*pp_stack) - oparg - 1;
@@ -4844,7 +4844,7 @@ call_function(PyObject ***pp_stack, Py_ssize_t oparg, PyObject *kwnames)
    done before evaluating the frame.
 */
 
-static PyObject*
+static PyObject* _Py_HOT_FUNCTION
 _PyFunction_FastCall(PyCodeObject *co, PyObject **args, Py_ssize_t nargs,
                      PyObject *globals)
 {

Python/errors.c

@@ -158,7 +158,7 @@ PyErr_SetString(PyObject *exception, const char *string)
 }
 
 
-PyObject *
+PyObject* _Py_HOT_FUNCTION
 PyErr_Occurred(void)
 {
     PyThreadState *tstate = PyThreadState_GET();