gh-115999: Enable specialization of CALL instructions in free-threaded builds

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

@@ -835,6 +835,20 @@ 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 *);
+
+// Cache the provided init method in the specialization cache of type if the
+// provided type version matches the current version of the type.
+//
+// The cached value is borrowed and is only valid if guarded by a type
+// version check. In free-threaded builds the init method must also use
+// deferred reference counting.
+//
+// Returns 1 if the value was cached or 0 otherwise.
+extern int _PyType_CacheInitForSpecialization(PyHeapTypeObject *type,
+                                              PyObject *init,
+                                              unsigned int tp_version);
 
 #ifdef Py_GIL_DISABLED
 #  define MANAGED_DICT_OFFSET    (((Py_ssize_t)sizeof(PyObject *))*-1)

Lib/test/test_monitoring.py

@@ -11,7 +11,7 @@
 import unittest
 
 import test.support
-from test.support import requires_specialization, script_helper
+from test.support import requires_specialization_ft, script_helper
 from test.support.import_helper import import_module
 
 _testcapi = test.support.import_helper.import_module("_testcapi")
@@ -850,6 +850,13 @@ def __init__(self, events):
     def __call__(self, code, offset, val):
         self.events.append(("return", code.co_name, val))
 
+# gh-127274: CALL_ALLOC_AND_ENTER_INIT will only cache __init__ methods that
+# are deferred. We only defer functions defined at the top-level.
+class ValueErrorRaiser:
+    def __init__(self):
+        raise ValueError()
+
+
 class ExceptionMonitoringTest(CheckEvents):
 
     exception_recorders = (
@@ -1045,16 +1052,12 @@ def func():
         )
         self.assertEqual(events[0], ("throw", IndexError))
 
-    @requires_specialization
+    @requires_specialization_ft
     def test_no_unwind_for_shim_frame(self):
 
-        class B:
-            def __init__(self):
-                raise ValueError()
-
         def f():
             try:
-                return B()
+                return ValueErrorRaiser()
             except ValueError:
                 pass
 

Lib/test/test_opcache.py

@@ -493,6 +493,18 @@ def f():
             self.assertFalse(f())
 
 
+# gh-127274: CALL_ALLOC_AND_ENTER_INIT will only cache __init__ methods that
+# are deferred. We only defer functions defined at the top-level.
+class MyClass:
+    def __init__(self):
+        pass
+
+
+class InitTakesArg:
+    def __init__(self, arg):
+        self.arg = arg
+
+
 class TestCallCache(TestBase):
     def test_too_many_defaults_0(self):
         def f():
@@ -522,12 +534,8 @@ def f(x, y):
             f()
 
     @disabling_optimizer
-    @requires_specialization
+    @requires_specialization_ft
     def test_assign_init_code(self):
-        class MyClass:
-            def __init__(self):
-                pass
-
         def instantiate():
             return MyClass()
 
@@ -544,6 +552,20 @@ def count_args(self, *args):
         MyClass.__init__.__code__ = count_args.__code__
         instantiate()
 
+    @disabling_optimizer
+    @requires_specialization_ft
+    def test_push_init_frame_fails(self):
+        def instantiate():
+            return InitTakesArg()
+
+        for _ in range(2):
+            with self.assertRaises(TypeError):
+                instantiate()
+        self.assert_specialized(instantiate, "CALL_ALLOC_AND_ENTER_INIT")
+
+        with self.assertRaises(TypeError):
+            instantiate()
+
 
 @threading_helper.requires_working_threading()
 class TestRacesDoNotCrash(TestBase):

Lib/test/test_type_cache.py

@@ -2,7 +2,7 @@
 import unittest
 import dis
 from test import support
-from test.support import import_helper, requires_specialization
+from test.support import import_helper, requires_specialization, requires_specialization_ft
 try:
     from sys import _clear_type_cache
 except ImportError:
@@ -110,7 +110,6 @@ class HolderSub(Holder):
             HolderSub.value
 
 @support.cpython_only
-@requires_specialization
 class TypeCacheWithSpecializationTests(unittest.TestCase):
     def tearDown(self):
         _clear_type_cache()
@@ -140,6 +139,7 @@ def _check_specialization(self, func, arg, opname, *, should_specialize):
         else:
             self.assertIn(opname, self._all_opnames(func))
 
+    @requires_specialization
     def test_class_load_attr_specialization_user_type(self):
         class A:
             def foo(self):
@@ -160,6 +160,7 @@ def load_foo_2(type_):
 
         self._check_specialization(load_foo_2, A, "LOAD_ATTR", should_specialize=False)
 
+    @requires_specialization
     def test_class_load_attr_specialization_static_type(self):
         self.assertNotEqual(type_get_version(str), 0)
         self.assertNotEqual(type_get_version(bytes), 0)
@@ -171,6 +172,7 @@ def get_capitalize_1(type_):
         self.assertEqual(get_capitalize_1(str)('hello'), 'Hello')
         self.assertEqual(get_capitalize_1(bytes)(b'hello'), b'Hello')
 
+    @requires_specialization
     def test_property_load_attr_specialization_user_type(self):
         class G:
             @property
@@ -192,6 +194,7 @@ def load_x_2(instance):
 
         self._check_specialization(load_x_2, G(), "LOAD_ATTR", should_specialize=False)
 
+    @requires_specialization
     def test_store_attr_specialization_user_type(self):
         class B:
             __slots__ = ("bar",)
@@ -211,6 +214,7 @@ def store_bar_2(type_):
 
         self._check_specialization(store_bar_2, B(), "STORE_ATTR", should_specialize=False)
 
+    @requires_specialization_ft
     def test_class_call_specialization_user_type(self):
         class F:
             def __init__(self):
@@ -231,6 +235,7 @@ def call_class_2(type_):
 
         self._check_specialization(call_class_2, F, "CALL", should_specialize=False)
 
+    @requires_specialization
     def test_to_bool_specialization_user_type(self):
         class H:
             pass

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

@@ -5527,9 +5527,12 @@ _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;
@@ -5552,6 +5555,9 @@ _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);
@@ -5573,6 +5579,9 @@ _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
@@ -5586,12 +5595,12 @@ _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();
 
@@ -5608,28 +5617,67 @@ _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(PyHeapTypeObject *type, PyObject *init,
+                                   unsigned int tp_version)
+{
+    if (!init || !tp_version) {
+        return 0;
+    }
+    int can_cache;
+    BEGIN_TYPE_LOCK();
+    can_cache = ((PyTypeObject*)type)->tp_version_tag == tp_version;
+    #ifdef Py_GIL_DISABLED
+    can_cache = can_cache && _PyObject_HasDeferredRefcount(init);
+    #endif
+    if (can_cache) {
+        FT_ATOMIC_STORE_PTR_RELAXED(type->_spec_cache.init, init);
+    }
+    END_TYPE_LOCK();
+    return can_cache;
+}
+
 static void
 set_flags(PyTypeObject *self, unsigned long mask, unsigned long flags)
 {