[3.14] gh-148393: Use acquire load for _ma_watcher_tag in dict notify event (gh-148509) (#148512)

The watcher-bits read in _PyDict_NotifyEvent needs to use acquire to
synchronize with the release from PyDict_Watch so that the callback
publication is visible before the callback is invoked.
(cherry picked from commit 19f96f9)

Co-authored-by: Sam Gross <colesbury@gmail.com>

Author: miss-islington

Include/internal/pycore_dict.h

@@ -286,7 +286,7 @@ _PyDict_NotifyEvent(PyInterpreterState *interp,
                     PyObject *value)
 {
     assert(Py_REFCNT((PyObject*)mp) > 0);
-    int watcher_bits = FT_ATOMIC_LOAD_UINT64_RELAXED(mp->_ma_watcher_tag) & DICT_WATCHER_MASK;
+    int watcher_bits = FT_ATOMIC_LOAD_UINT64_ACQUIRE(mp->_ma_watcher_tag) & DICT_WATCHER_MASK;
     if (watcher_bits) {
         RARE_EVENT_STAT_INC(watched_dict_modification);
         _PyDict_SendEvent(watcher_bits, event, mp, key, value);

Include/internal/pycore_pyatomic_ft_wrappers.h

@@ -47,6 +47,8 @@ extern "C" {
     _Py_atomic_load_uint16_relaxed(&value)
 #define FT_ATOMIC_LOAD_UINT32_RELAXED(value) \
     _Py_atomic_load_uint32_relaxed(&value)
+#define FT_ATOMIC_LOAD_UINT64_ACQUIRE(value) \
+    _Py_atomic_load_uint64_acquire(&value)
 #define FT_ATOMIC_LOAD_UINT64_RELAXED(value) \
     _Py_atomic_load_uint64_relaxed(&value)
 #define FT_ATOMIC_LOAD_ULONG_RELAXED(value) \
@@ -147,6 +149,7 @@ extern "C" {
 #define FT_ATOMIC_LOAD_UINT8_RELAXED(value) value
 #define FT_ATOMIC_LOAD_UINT16_RELAXED(value) value
 #define FT_ATOMIC_LOAD_UINT32_RELAXED(value) value
+#define FT_ATOMIC_LOAD_UINT64_ACQUIRE(value) value
 #define FT_ATOMIC_LOAD_UINT64_RELAXED(value) value
 #define FT_ATOMIC_LOAD_ULONG_RELAXED(value) value
 #define FT_ATOMIC_STORE_PTR_RELAXED(value, new_value) value = new_value