Merge branch 'pythongh-115999-tlbc-call' into pythongh-115999-tlbc-in…

…tegration

Include/internal/pycore_list.h

@@ -37,6 +37,9 @@ _PyList_AppendTakeRef(PyListObject *self, PyObject *newitem)
     return _PyList_AppendTakeRefListResize(self, newitem);
 }
 
+// Like _PyList_AppendTakeRef, but locks self in free-threaded builds.
+extern int _PyList_AppendTakeRefAndLock(PyListObject *self, PyObject *newitem);
+
 // Repeat the bytes of a buffer in place
 static inline void
 _Py_memory_repeat(char* dest, Py_ssize_t len_dest, Py_ssize_t len_src)

Include/internal/pycore_object.h

@@ -775,6 +775,10 @@ extern int _PyObject_StoreInstanceAttribute(PyObject *obj,
                                             PyObject *name, PyObject *value);
 extern bool _PyObject_TryGetInstanceAttribute(PyObject *obj, PyObject *name,
                                               PyObject **attr);
+extern PyObject *_PyType_LookupRefAndVersion(PyTypeObject *, PyObject *,
+                                             unsigned int *);
+extern int _PyType_CacheInitForSpecialization(PyTypeObject *, PyObject *,
+                                              unsigned int);
 
 #ifdef Py_GIL_DISABLED
 #  define MANAGED_DICT_OFFSET    (((Py_ssize_t)sizeof(PyObject *))*-1)

Objects/listobject.c

@@ -490,6 +490,16 @@ PyList_Append(PyObject *op, PyObject *newitem)
     return -1;
 }
 
+int
+_PyList_AppendTakeRefAndLock(PyListObject *self, PyObject *newitem)
+{
+    int ret;
+    Py_BEGIN_CRITICAL_SECTION(self);
+    ret = _PyList_AppendTakeRef((PyListObject *)self, newitem);
+    Py_END_CRITICAL_SECTION();
+    return ret;
+}
+
 /* Methods */
 
 static void

Objects/typeobject.c

@@ -5512,9 +5512,12 @@ _PyTypes_AfterFork(void)
 }
 
 /* Internal API to look for a name through the MRO.
-   This returns a borrowed reference, and doesn't set an exception! */
+   This returns a strong reference, and doesn't set an exception!
+   If nonzero, version is set to the value of type->tp_version at the time of
+   the lookup.
+*/
 PyObject *
-_PyType_LookupRef(PyTypeObject *type, PyObject *name)
+_PyType_LookupRefAndVersion(PyTypeObject *type, PyObject *name, unsigned int *version)
 {
     PyObject *res;
     int error;
@@ -5537,6 +5540,9 @@ _PyType_LookupRef(PyTypeObject *type, PyObject *name)
             // If the sequence is still valid then we're done
             if (value == NULL || _Py_TryIncref(value)) {
                 if (_PySeqLock_EndRead(&entry->sequence, sequence)) {
+                    if (version != NULL) {
+                        *version = entry_version;
+                    }
                     return value;
                 }
                 Py_XDECREF(value);
@@ -5558,6 +5564,9 @@ _PyType_LookupRef(PyTypeObject *type, PyObject *name)
         OBJECT_STAT_INC_COND(type_cache_hits, !is_dunder_name(name));
         OBJECT_STAT_INC_COND(type_cache_dunder_hits, is_dunder_name(name));
         Py_XINCREF(entry->value);
+        if (version != NULL) {
+            *version = entry->version;
+        }
         return entry->value;
     }
 #endif
@@ -5571,12 +5580,12 @@ _PyType_LookupRef(PyTypeObject *type, PyObject *name)
     // anyone else can modify our mro or mutate the type.
 
     int has_version = 0;
-    int version = 0;
+    unsigned int assigned_version = 0;
     BEGIN_TYPE_LOCK();
     res = find_name_in_mro(type, name, &error);
     if (MCACHE_CACHEABLE_NAME(name)) {
         has_version = assign_version_tag(interp, type);
-        version = type->tp_version_tag;
+        assigned_version = type->tp_version_tag;
     }
     END_TYPE_LOCK();
 
@@ -5593,28 +5602,68 @@ _PyType_LookupRef(PyTypeObject *type, PyObject *name)
         if (error == -1) {
             PyErr_Clear();
         }
+        if (version != NULL) {
+            // 0 is not a valid version
+            *version = 0;
+        }
         return NULL;
     }
 
     if (has_version) {
 #if Py_GIL_DISABLED
-        update_cache_gil_disabled(entry, name, version, res);
+        update_cache_gil_disabled(entry, name, assigned_version, res);
 #else
-        PyObject *old_value = update_cache(entry, name, version, res);
+        PyObject *old_value = update_cache(entry, name, assigned_version, res);
         Py_DECREF(old_value);
 #endif
     }
+    if (version != NULL) {
+        // 0 is not a valid version
+        *version = has_version ? assigned_version : 0;
+    }
     return res;
 }
 
+/* Internal API to look for a name through the MRO.
+   This returns a strong reference, and doesn't set an exception!
+*/
+PyObject *
+_PyType_LookupRef(PyTypeObject *type, PyObject *name)
+{
+    return _PyType_LookupRefAndVersion(type, name, NULL);
+}
+
+/* Internal API to look for a name through the MRO.
+   This returns a borrowed reference, and doesn't set an exception! */
 PyObject *
 _PyType_Lookup(PyTypeObject *type, PyObject *name)
 {
-    PyObject *res = _PyType_LookupRef(type, name);
+    PyObject *res = _PyType_LookupRefAndVersion(type, name, NULL);
     Py_XDECREF(res);
     return res;
 }
 
+
+int
+_PyType_CacheInitForSpecialization(PyTypeObject *type, PyObject *init,
+                                   unsigned int tp_version)
+{
+    if (!init || !tp_version) {
+        return 0;
+    }
+    int can_cache = type->tp_version_tag == tp_version;
+    BEGIN_TYPE_LOCK();
+    #ifdef Py_GIL_DISABLED
+    can_cache = can_cache && _PyObject_HasDeferredRefcount(init);
+    #endif
+    if (can_cache) {
+        PyHeapTypeObject *ht = (PyHeapTypeObject*) type;
+        FT_ATOMIC_STORE_PTR_RELAXED(ht->_spec_cache.init, init);
+    }
+    END_TYPE_LOCK();
+    return can_cache;
+}
+
 static void
 set_flags(PyTypeObject *self, unsigned long mask, unsigned long flags)
 {

Python/bytecodes.c

@@ -3316,15 +3316,15 @@ dummy_func(
         };
 
         specializing op(_SPECIALIZE_CALL, (counter/1, callable[1], self_or_null[1], args[oparg] -- callable[1], self_or_null[1], args[oparg])) {
-            #if ENABLE_SPECIALIZATION
+            #if ENABLE_SPECIALIZATION_FT
             if (ADAPTIVE_COUNTER_TRIGGERS(counter)) {
                 next_instr = this_instr;
                 _Py_Specialize_Call(callable[0], next_instr, oparg + !PyStackRef_IsNull(self_or_null[0]));
                 DISPATCH_SAME_OPARG();
             }
             OPCODE_DEFERRED_INC(CALL);
             ADVANCE_ADAPTIVE_COUNTER(this_instr[1].counter);
-            #endif  /* ENABLE_SPECIALIZATION */
+            #endif  /* ENABLE_SPECIALIZATION_FT */
         }
 
         op(_MAYBE_EXPAND_METHOD, (callable[1], self_or_null[1], args[oparg] -- func[1], maybe_self[1], args[oparg])) {
@@ -3662,11 +3662,12 @@ dummy_func(
             DEOPT_IF(!PyStackRef_IsNull(null));
             DEOPT_IF(callable_o != (PyObject *)&PyUnicode_Type);
             STAT_INC(CALL, hit);
-            res = PyStackRef_FromPyObjectSteal(PyObject_Str(arg_o));
+            PyObject *res_o = PyObject_Str(arg_o);
             DEAD(null);
             DEAD(callable);
             PyStackRef_CLOSE(arg);
-            ERROR_IF(PyStackRef_IsNull(res), error);
+            ERROR_IF(res_o == NULL, error);
+            res = PyStackRef_FromPyObjectSteal(res_o);
         }
 
         macro(CALL_STR_1) =
@@ -3683,11 +3684,12 @@ dummy_func(
             DEOPT_IF(!PyStackRef_IsNull(null));
             DEOPT_IF(callable_o != (PyObject *)&PyTuple_Type);
             STAT_INC(CALL, hit);
-            res = PyStackRef_FromPyObjectSteal(PySequence_Tuple(arg_o));
+            PyObject *res_o = PySequence_Tuple(arg_o);
             DEAD(null);
             DEAD(callable);
             PyStackRef_CLOSE(arg);
-            ERROR_IF(PyStackRef_IsNull(res), error);
+            ERROR_IF(res_o == NULL, error);
+            res = PyStackRef_FromPyObjectSteal(res_o);
         }
 
         macro(CALL_TUPLE_1) =
@@ -3701,10 +3703,10 @@ dummy_func(
             DEOPT_IF(!PyStackRef_IsNull(null[0]));
             DEOPT_IF(!PyType_Check(callable_o));
             PyTypeObject *tp = (PyTypeObject *)callable_o;
-            DEOPT_IF(tp->tp_version_tag != type_version);
+            DEOPT_IF(FT_ATOMIC_LOAD_UINT32_RELAXED(tp->tp_version_tag) != type_version);
             assert(tp->tp_flags & Py_TPFLAGS_INLINE_VALUES);
             PyHeapTypeObject *cls = (PyHeapTypeObject *)callable_o;
-            PyFunctionObject *init_func = (PyFunctionObject *)cls->_spec_cache.init;
+            PyFunctionObject *init_func = (PyFunctionObject *)FT_ATOMIC_LOAD_PTR_RELAXED(cls->_spec_cache.init);
             PyCodeObject *code = (PyCodeObject *)init_func->func_code;
             DEOPT_IF(!_PyThreadState_HasStackSpace(tstate, code->co_framesize + _Py_InitCleanup.co_framesize));
             STAT_INC(CALL, hit);
@@ -3980,7 +3982,12 @@ dummy_func(
             assert(self_o != NULL);
             DEOPT_IF(!PyList_Check(self_o));
             STAT_INC(CALL, hit);
+        #ifdef Py_GIL_DISABLED
+            int err;
+            err = _PyList_AppendTakeRefAndLock((PyListObject *)self_o, PyStackRef_AsPyObjectSteal(arg));
+        #else
             int err = _PyList_AppendTakeRef((PyListObject *)self_o, PyStackRef_AsPyObjectSteal(arg));
+        #endif
             PyStackRef_CLOSE(self);
             PyStackRef_CLOSE(callable);
             ERROR_IF(err, error);