gh-102304: Move _Py_RefTotal to _PyRuntimeState (gh-102543)

The essentially eliminates the global variable, with the associated benefits. This is also a precursor to isolating this bit of state to PyInterpreterState.

Folks that currently read _Py_RefTotal directly would have to start using _Py_GetGlobalRefTotal() instead.

https://github.com/python/cpython/issues/102304

Python/pylifecycle.c

@@ -1930,6 +1930,7 @@ Py_FinalizeEx(void)
     if (show_ref_count) {
         _PyDebug_PrintTotalRefs();
     }
+    _Py_FinalizeRefTotal(runtime);
 #endif
 
 #ifdef Py_TRACE_REFS

Python/pystate.c

@@ -482,6 +482,9 @@ _PyRuntimeState_Init(_PyRuntimeState *runtime)
 void
 _PyRuntimeState_Fini(_PyRuntimeState *runtime)
 {
+    /* The reftotal is cleared by _Py_FinalizeRefTotal(). */
+    assert(runtime->object_state.reftotal == 0);
+
     if (gilstate_tss_initialized(runtime)) {
         gilstate_tss_fini(runtime);
     }

Python/sysmodule.c

@@ -1854,7 +1854,7 @@ static Py_ssize_t
 sys_gettotalrefcount_impl(PyObject *module)
 /*[clinic end generated code: output=4103886cf17c25bc input=53b744faa5d2e4f6]*/
 {
-    return _Py_GetRefTotal();
+    return _Py_GetGlobalRefTotal();
 }
 
 #endif /* Py_REF_DEBUG */