From 06fd745dd91c757b643f69847cf1234609968132 Mon Sep 17 00:00:00 2001 From: "Miss Islington (bot)" <31488909+miss-islington@users.noreply.github.com> Date: Mon, 1 Jul 2024 21:40:28 +0200 Subject: [PATCH] [3.13] gh-117657: Fix data races reported by TSAN in some set methods (GH-120914) (#121240) Refactor the fast Unicode hash check into `_PyObject_HashFast` and use relaxed atomic loads in the free-threaded build. After this change, the TSAN doesn't report data races for this method. (cherry picked from commit 294e72496439da984cb8dba9100d3613c8cc8a6d) Co-authored-by: AN Long --- Include/internal/pycore_object.h | 14 +++ Modules/_collectionsmodule.c | 9 +- Objects/dictobject.c | 135 ++++++++------------- Objects/setobject.c | 30 ++--- Objects/typeobject.c | 13 +- Tools/tsan/suppressions_free_threading.txt | 1 - 6 files changed, 81 insertions(+), 121 deletions(-) diff --git a/Include/internal/pycore_object.h b/Include/internal/pycore_object.h index d1e2773a247..e016ec5e12c 100644 --- a/Include/internal/pycore_object.h +++ b/Include/internal/pycore_object.h @@ -628,6 +628,20 @@ _PyObject_IS_GC(PyObject *obj) && (type->tp_is_gc == NULL || type->tp_is_gc(obj))); } +// Fast inlined version of PyObject_Hash() +static inline Py_hash_t +_PyObject_HashFast(PyObject *op) +{ + if (PyUnicode_CheckExact(op)) { + Py_hash_t hash = FT_ATOMIC_LOAD_SSIZE_RELAXED( + _PyASCIIObject_CAST(op)->hash); + if (hash != -1) { + return hash; + } + } + return PyObject_Hash(op); +} + // Fast inlined version of PyType_IS_GC() #define _PyType_IS_GC(t) _PyType_HasFeature((t), Py_TPFLAGS_HAVE_GC) diff --git a/Modules/_collectionsmodule.c b/Modules/_collectionsmodule.c index 644a90a8c71..c821f18b170 100644 --- a/Modules/_collectionsmodule.c +++ b/Modules/_collectionsmodule.c @@ -2537,12 +2537,9 @@ _collections__count_elements_impl(PyObject *module, PyObject *mapping, if (key == NULL) break; - if (!PyUnicode_CheckExact(key) || - (hash = _PyASCIIObject_CAST(key)->hash) == -1) - { - hash = PyObject_Hash(key); - if (hash == -1) - goto done; + hash = _PyObject_HashFast(key); + if (hash == -1) { + goto done; } oldval = _PyDict_GetItem_KnownHash(mapping, key, hash); diff --git a/Objects/dictobject.c b/Objects/dictobject.c index 27da631a574..47cd561f243 100644 --- a/Objects/dictobject.c +++ b/Objects/dictobject.c @@ -427,7 +427,7 @@ static inline Py_hash_t unicode_get_hash(PyObject *o) { assert(PyUnicode_CheckExact(o)); - return _PyASCIIObject_CAST(o)->hash; + return FT_ATOMIC_LOAD_SSIZE_RELAXED(_PyASCIIObject_CAST(o)->hash); } /* Print summary info about the state of the optimized allocator */ @@ -2170,13 +2170,10 @@ dict_getitem(PyObject *op, PyObject *key, const char *warnmsg) } PyDictObject *mp = (PyDictObject *)op; - Py_hash_t hash; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) { - PyErr_FormatUnraisable(warnmsg); - return NULL; - } + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + PyErr_FormatUnraisable(warnmsg); + return NULL; } PyThreadState *tstate = _PyThreadState_GET(); @@ -2225,12 +2222,9 @@ _PyDict_LookupIndex(PyDictObject *mp, PyObject *key) assert(PyDict_CheckExact((PyObject*)mp)); assert(PyUnicode_CheckExact(key)); - Py_hash_t hash = unicode_get_hash(key); + Py_hash_t hash = _PyObject_HashFast(key); if (hash == -1) { - hash = PyObject_Hash(key); - if (hash == -1) { - return -1; - } + return -1; } return _Py_dict_lookup(mp, key, hash, &value); @@ -2301,14 +2295,10 @@ PyDict_GetItemRef(PyObject *op, PyObject *key, PyObject **result) return -1; } - Py_hash_t hash; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) - { - hash = PyObject_Hash(key); - if (hash == -1) { - *result = NULL; - return -1; - } + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + *result = NULL; + return -1; } return _PyDict_GetItemRef_KnownHash((PyDictObject *)op, key, hash, result); @@ -2320,13 +2310,10 @@ _PyDict_GetItemRef_Unicode_LockHeld(PyDictObject *op, PyObject *key, PyObject ** ASSERT_DICT_LOCKED(op); assert(PyUnicode_CheckExact(key)); - Py_hash_t hash; - if ((hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) { - *result = NULL; - return -1; - } + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + *result = NULL; + return -1; } PyObject *value; @@ -2360,12 +2347,9 @@ PyDict_GetItemWithError(PyObject *op, PyObject *key) PyErr_BadInternalCall(); return NULL; } - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) - { - hash = PyObject_Hash(key); - if (hash == -1) { - return NULL; - } + hash = _PyObject_HashFast(key); + if (hash == -1) { + return NULL; } #ifdef Py_GIL_DISABLED @@ -2433,10 +2417,9 @@ _PyDict_LoadGlobal(PyDictObject *globals, PyDictObject *builtins, PyObject *key) Py_hash_t hash; PyObject *value; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return NULL; + hash = _PyObject_HashFast(key); + if (hash == -1) { + return NULL; } /* namespace 1: globals */ @@ -2461,14 +2444,11 @@ setitem_take2_lock_held(PyDictObject *mp, PyObject *key, PyObject *value) assert(key); assert(value); assert(PyDict_Check(mp)); - Py_hash_t hash; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) { - Py_DECREF(key); - Py_DECREF(value); - return -1; - } + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + Py_DECREF(key); + Py_DECREF(value); + return -1; } PyInterpreterState *interp = _PyInterpreterState_GET(); @@ -2617,12 +2597,10 @@ delitem_common(PyDictObject *mp, Py_hash_t hash, Py_ssize_t ix, int PyDict_DelItem(PyObject *op, PyObject *key) { - Py_hash_t hash; assert(key); - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return -1; + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + return -1; } return _PyDict_DelItem_KnownHash(op, key, hash); @@ -2946,15 +2924,12 @@ pop_lock_held(PyObject *op, PyObject *key, PyObject **result) return 0; } - Py_hash_t hash; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) { - if (result) { - *result = NULL; - } - return -1; + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + if (result) { + *result = NULL; } + return -1; } return _PyDict_Pop_KnownHash(dict, key, hash, result); } @@ -3285,10 +3260,9 @@ dict_subscript(PyObject *self, PyObject *key) Py_hash_t hash; PyObject *value; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return NULL; + hash = _PyObject_HashFast(key); + if (hash == -1) { + return NULL; } ix = _Py_dict_lookup_threadsafe(mp, key, hash, &value); if (ix == DKIX_ERROR) @@ -4172,10 +4146,9 @@ dict_get_impl(PyDictObject *self, PyObject *key, PyObject *default_value) Py_hash_t hash; Py_ssize_t ix; - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return NULL; + hash = _PyObject_HashFast(key); + if (hash == -1) { + return NULL; } ix = _Py_dict_lookup_threadsafe(self, key, hash, &val); if (ix == DKIX_ERROR) @@ -4205,14 +4178,12 @@ dict_setdefault_ref_lock_held(PyObject *d, PyObject *key, PyObject *default_valu return -1; } - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) { - if (result) { - *result = NULL; - } - return -1; + hash = _PyObject_HashFast(key); + if (hash == -1) { + if (result) { + *result = NULL; } + return -1; } if (mp->ma_keys == Py_EMPTY_KEYS) { @@ -4642,12 +4613,10 @@ static PyMethodDef mapp_methods[] = { int PyDict_Contains(PyObject *op, PyObject *key) { - Py_hash_t hash; + Py_hash_t hash = _PyObject_HashFast(key); - if (!PyUnicode_CheckExact(key) || (hash = unicode_get_hash(key)) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return -1; + if (hash == -1) { + return -1; } return _PyDict_Contains_KnownHash(op, key, hash); @@ -6727,11 +6696,9 @@ int _PyDict_SetItem_LockHeld(PyDictObject *dict, PyObject *name, PyObject *value) { if (value == NULL) { - Py_hash_t hash; - if (!PyUnicode_CheckExact(name) || (hash = unicode_get_hash(name)) == -1) { - hash = PyObject_Hash(name); - if (hash == -1) - return -1; + Py_hash_t hash = _PyObject_HashFast(name); + if (hash == -1) { + return -1; } return delitem_knownhash_lock_held((PyObject *)dict, name, hash); } else { diff --git a/Objects/setobject.c b/Objects/setobject.c index 68986bb6a6b..eb0c404bf6b 100644 --- a/Objects/setobject.c +++ b/Objects/setobject.c @@ -365,13 +365,9 @@ set_discard_entry(PySetObject *so, PyObject *key, Py_hash_t hash) static int set_add_key(PySetObject *so, PyObject *key) { - Py_hash_t hash; - - if (!PyUnicode_CheckExact(key) || - (hash = _PyASCIIObject_CAST(key)->hash) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return -1; + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + return -1; } return set_add_entry(so, key, hash); } @@ -379,13 +375,9 @@ set_add_key(PySetObject *so, PyObject *key) static int set_contains_key(PySetObject *so, PyObject *key) { - Py_hash_t hash; - - if (!PyUnicode_CheckExact(key) || - (hash = _PyASCIIObject_CAST(key)->hash) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return -1; + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + return -1; } return set_contains_entry(so, key, hash); } @@ -393,13 +385,9 @@ set_contains_key(PySetObject *so, PyObject *key) static int set_discard_key(PySetObject *so, PyObject *key) { - Py_hash_t hash; - - if (!PyUnicode_CheckExact(key) || - (hash = _PyASCIIObject_CAST(key)->hash) == -1) { - hash = PyObject_Hash(key); - if (hash == -1) - return -1; + Py_hash_t hash = _PyObject_HashFast(key); + if (hash == -1) { + return -1; } return set_discard_entry(so, key, hash); } diff --git a/Objects/typeobject.c b/Objects/typeobject.c index 63902ebc1f9..13eac2efb98 100644 --- a/Objects/typeobject.c +++ b/Objects/typeobject.c @@ -5069,15 +5069,10 @@ find_name_in_mro(PyTypeObject *type, PyObject *name, int *error) { ASSERT_TYPE_LOCK_HELD(); - Py_hash_t hash; - if (!PyUnicode_CheckExact(name) || - (hash = _PyASCIIObject_CAST(name)->hash) == -1) - { - hash = PyObject_Hash(name); - if (hash == -1) { - *error = -1; - return NULL; - } + Py_hash_t hash = _PyObject_HashFast(name); + if (hash == -1) { + *error = -1; + return NULL; } /* Look in tp_dict of types in MRO */ diff --git a/Tools/tsan/suppressions_free_threading.txt b/Tools/tsan/suppressions_free_threading.txt index 2986efe6774..d6e439883e2 100644 --- a/Tools/tsan/suppressions_free_threading.txt +++ b/Tools/tsan/suppressions_free_threading.txt @@ -30,7 +30,6 @@ race_top:assign_version_tag race_top:insertdict race_top:lookup_tp_dict race_top:new_reference -race_top:set_contains_key # https://gist.github.com/colesbury/d13d033f413b4ad07929d044bed86c35 race_top:set_discard_entry race_top:_PyDict_CheckConsistency