gh-114312: Collect stats for unlikely events (GH-114493)

11 files changed, 199 insertions, 1 deletions

diff --git a/Include/cpython/pystats.h b/Include/cpython/pystats.h
index ba67eef..bf0cfe4 100644
--- a/Include/cpython/pystats.h
+++ b/Include/cpython/pystats.h
@@ -122,11 +122,25 @@ typedef struct _optimization_stats {
     uint64_t optimized_trace_length_hist[_Py_UOP_HIST_SIZE];
 } OptimizationStats;
 
+typedef struct _rare_event_stats {
+    /* Setting an object's class, obj.__class__ = ... */
+    uint64_t set_class;
+    /* Setting the bases of a class, cls.__bases__ = ... */
+    uint64_t set_bases;
+    /* Setting the PEP 523 frame eval function, _PyInterpreterState_SetFrameEvalFunc() */
+    uint64_t set_eval_frame_func;
+    /* Modifying the builtins,  __builtins__.__dict__[var] = ... */
+    uint64_t builtin_dict;
+    /* Modifying a function, e.g. func.__defaults__ = ..., etc. */
+    uint64_t func_modification;
+} RareEventStats;
+
 typedef struct _stats {
     OpcodeStats opcode_stats[256];
     CallStats call_stats;
     ObjectStats object_stats;
     OptimizationStats optimization_stats;
+    RareEventStats rare_event_stats;
     GCStats *gc_stats;
 } PyStats;
 
diff --git a/Include/internal/pycore_code.h b/Include/internal/pycore_code.h
index 73df6c3..fdd5918 100644
--- a/Include/internal/pycore_code.h
+++ b/Include/internal/pycore_code.h
@@ -295,6 +295,7 @@ extern int _PyStaticCode_Init(PyCodeObject *co);
             _Py_stats->optimization_stats.name[bucket]++; \
         } \
     } while (0)
+#define RARE_EVENT_STAT_INC(name) do { if (_Py_stats) _Py_stats->rare_event_stats.name++; } while (0)
 
 // Export for '_opcode' shared extension
 PyAPI_FUNC(PyObject*) _Py_GetSpecializationStats(void);
@@ -313,6 +314,7 @@ PyAPI_FUNC(PyObject*) _Py_GetSpecializationStats(void);
 #define UOP_STAT_INC(opname, name) ((void)0)
 #define OPT_UNSUPPORTED_OPCODE(opname) ((void)0)
 #define OPT_HIST(length, name) ((void)0)
+#define RARE_EVENT_STAT_INC(name) ((void)0)
 #endif  // !Py_STATS
 
 // Utility functions for reading/writing 32/64-bit values in the inline caches.
diff --git a/Include/internal/pycore_interp.h b/Include/internal/pycore_interp.h
index f953b84..662a18d 100644
--- a/Include/internal/pycore_interp.h
+++ b/Include/internal/pycore_interp.h
@@ -60,6 +60,21 @@ struct _stoptheworld_state {
 
 /* cross-interpreter data registry */
 
+/* Tracks some rare events per-interpreter, used by the optimizer to turn on/off
+   specific optimizations. */
+typedef struct _rare_events {
+    /* Setting an object's class, obj.__class__ = ... */
+    uint8_t set_class;
+    /* Setting the bases of a class, cls.__bases__ = ... */
+    uint8_t set_bases;
+    /* Setting the PEP 523 frame eval function, _PyInterpreterState_SetFrameEvalFunc() */
+    uint8_t set_eval_frame_func;
+    /* Modifying the builtins,  __builtins__.__dict__[var] = ... */
+    uint8_t builtin_dict;
+    int builtins_dict_watcher_id;
+    /* Modifying a function, e.g. func.__defaults__ = ..., etc. */
+    uint8_t func_modification;
+} _rare_events;
 
 /* interpreter state */
 
@@ -217,6 +232,7 @@ struct _is {
     uint16_t optimizer_resume_threshold;
     uint16_t optimizer_backedge_threshold;
     uint32_t next_func_version;
+    _rare_events rare_events;
 
     _Py_GlobalMonitors monitors;
     bool sys_profile_initialized;
@@ -347,6 +363,19 @@ PyAPI_FUNC(PyStatus) _PyInterpreterState_New(
     PyInterpreterState **pinterp);
 
 
+#define RARE_EVENT_INTERP_INC(interp, name) \
+    do { \
+        /* saturating add */ \
+        if (interp->rare_events.name < UINT8_MAX) interp->rare_events.name++; \
+        RARE_EVENT_STAT_INC(name); \
+    } while (0); \
+
+#define RARE_EVENT_INC(name) \
+    do { \
+        PyInterpreterState *interp = PyInterpreterState_Get(); \
+        RARE_EVENT_INTERP_INC(interp, name); \
+    } while (0); \
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/Lib/test/test_optimizer.py b/Lib/test/test_optimizer.py
new file mode 100644
index 0000000..b56bf3c
--- /dev/null
+++ b/Lib/test/test_optimizer.py
@@ -0,0 +1,75 @@
+import _testinternalcapi
+import unittest
+import types
+
+
+class TestRareEventCounters(unittest.TestCase):
+    def test_set_class(self):
+        class A:
+            pass
+        class B:
+            pass
+        a = A()
+
+        orig_counter = _testinternalcapi.get_rare_event_counters()["set_class"]
+        a.__class__ = B
+        self.assertEqual(
+            orig_counter + 1,
+            _testinternalcapi.get_rare_event_counters()["set_class"]
+        )
+
+    def test_set_bases(self):
+        class A:
+            pass
+        class B:
+            pass
+        class C(B):
+            pass
+
+        orig_counter = _testinternalcapi.get_rare_event_counters()["set_bases"]
+        C.__bases__ = (A,)
+        self.assertEqual(
+            orig_counter + 1,
+            _testinternalcapi.get_rare_event_counters()["set_bases"]
+        )
+
+    def test_set_eval_frame_func(self):
+        orig_counter = _testinternalcapi.get_rare_event_counters()["set_eval_frame_func"]
+        _testinternalcapi.set_eval_frame_record([])
+        self.assertEqual(
+            orig_counter + 1,
+            _testinternalcapi.get_rare_event_counters()["set_eval_frame_func"]
+        )
+        _testinternalcapi.set_eval_frame_default()
+
+    def test_builtin_dict(self):
+        orig_counter = _testinternalcapi.get_rare_event_counters()["builtin_dict"]
+        if isinstance(__builtins__, types.ModuleType):
+            builtins = __builtins__.__dict__
+        else:
+            builtins = __builtins__
+        builtins["FOO"] = 42
+        self.assertEqual(
+            orig_counter + 1,
+            _testinternalcapi.get_rare_event_counters()["builtin_dict"]
+        )
+        del builtins["FOO"]
+
+    def test_func_modification(self):
+        def func(x=0):
+            pass
+
+        for attribute in (
+            "__code__",
+            "__defaults__",
+            "__kwdefaults__"
+        ):
+            orig_counter = _testinternalcapi.get_rare_event_counters()["func_modification"]
+            setattr(func, attribute, getattr(func, attribute))
+            self.assertEqual(
+                orig_counter + 1,
+                _testinternalcapi.get_rare_event_counters()["func_modification"]
+            )
+
+if __name__ == "__main__":
+    unittest.main()
diff --git a/Modules/_testinternalcapi.c b/Modules/_testinternalcapi.c
index 7d277df..2c32c69 100644
--- a/Modules/_testinternalcapi.c
+++ b/Modules/_testinternalcapi.c
@@ -1635,6 +1635,21 @@ get_type_module_name(PyObject *self, PyObject *type)
     return _PyType_GetModuleName((PyTypeObject *)type);
 }
 
+static PyObject *
+get_rare_event_counters(PyObject *self, PyObject *type)
+{
+    PyInterpreterState *interp = PyInterpreterState_Get();
+
+    return Py_BuildValue(
+        "{sksksksksk}",
+        "set_class", interp->rare_events.set_class,
+        "set_bases", interp->rare_events.set_bases,
+        "set_eval_frame_func", interp->rare_events.set_eval_frame_func,
+        "builtin_dict", interp->rare_events.builtin_dict,
+        "func_modification", interp->rare_events.func_modification
+    );
+}
+
 
 #ifdef Py_GIL_DISABLED
 static PyObject *
@@ -1711,6 +1726,7 @@ static PyMethodDef module_functions[] = {
     {"restore_crossinterp_data", restore_crossinterp_data,       METH_VARARGS},
     _TESTINTERNALCAPI_TEST_LONG_NUMBITS_METHODDEF
     {"get_type_module_name",    get_type_module_name,            METH_O},
+    {"get_rare_event_counters", get_rare_event_counters, METH_NOARGS},
 #ifdef Py_GIL_DISABLED
     {"py_thread_id", get_py_thread_id, METH_NOARGS},
 #endif
diff --git a/Objects/funcobject.c b/Objects/funcobject.c
index 2620dc6..08b2823 100644
--- a/Objects/funcobject.c
+++ b/Objects/funcobject.c
@@ -53,6 +53,15 @@ handle_func_event(PyFunction_WatchEvent event, PyFunctionObject *func,
     if (interp->active_func_watchers) {
         notify_func_watchers(interp, event, func, new_value);
     }
+    switch (event) {
+        case PyFunction_EVENT_MODIFY_CODE:
+        case PyFunction_EVENT_MODIFY_DEFAULTS:
+        case PyFunction_EVENT_MODIFY_KWDEFAULTS:
+            RARE_EVENT_INTERP_INC(interp, func_modification);
+            break;
+        default:
+            break;
+    }
 }
 
 int
diff --git a/Objects/typeobject.c b/Objects/typeobject.c
index 3a35a5b..a8c3b88 100644
--- a/Objects/typeobject.c
+++ b/Objects/typeobject.c
@@ -1371,6 +1371,7 @@ type_set_bases(PyTypeObject *type, PyObject *new_bases, void *context)
         res = 0;
     }
 
+    RARE_EVENT_INC(set_bases);
     Py_DECREF(old_bases);
     Py_DECREF(old_base);
 
@@ -5842,6 +5843,8 @@ object_set_class(PyObject *self, PyObject *value, void *closure)
         Py_SET_TYPE(self, newto);
         if (oldto->tp_flags & Py_TPFLAGS_HEAPTYPE)
             Py_DECREF(oldto);
+
+        RARE_EVENT_INC(set_class);
         return 0;
     }
     else {
diff --git a/Python/pylifecycle.c b/Python/pylifecycle.c
index 0d5eec0..261622a 100644
--- a/Python/pylifecycle.c
+++ b/Python/pylifecycle.c
@@ -605,6 +605,12 @@ init_interp_create_gil(PyThreadState *tstate, int gil)
     _PyEval_InitGIL(tstate, own_gil);
 }
 
+static int
+builtins_dict_watcher(PyDict_WatchEvent event, PyObject *dict, PyObject *key, PyObject *new_value)
+{
+    RARE_EVENT_INC(builtin_dict);
+    return 0;
+}
 
 static PyStatus
 pycore_create_interpreter(_PyRuntimeState *runtime,
@@ -1266,6 +1272,14 @@ init_interp_main(PyThreadState *tstate)
         }
     }
 
+    if ((interp->rare_events.builtins_dict_watcher_id = PyDict_AddWatcher(&builtins_dict_watcher)) == -1) {
+        return _PyStatus_ERR("failed to add builtin dict watcher");
+    }
+
+    if (PyDict_Watch(interp->rare_events.builtins_dict_watcher_id, interp->builtins) != 0) {
+        return _PyStatus_ERR("failed to set builtin dict watcher");
+    }
+
     assert(!_PyErr_Occurred(tstate));
 
     return _PyStatus_OK();
@@ -1592,6 +1606,10 @@ static void
 finalize_modules(PyThreadState *tstate)
 {
     PyInterpreterState *interp = tstate->interp;
+
+    // Stop collecting stats on __builtin__ modifications during teardown
+    PyDict_Unwatch(interp->rare_events.builtins_dict_watcher_id, interp->builtins);
+
     PyObject *modules = _PyImport_GetModules(interp);
     if (modules == NULL) {
         // Already done
diff --git a/Python/pystate.c b/Python/pystate.c
index 548c77b..c9b5213 100644
--- a/Python/pystate.c
+++ b/Python/pystate.c
@@ -2616,6 +2616,7 @@ _PyInterpreterState_SetEvalFrameFunc(PyInterpreterState *interp,
     if (eval_frame != NULL) {
         _Py_Executors_InvalidateAll(interp);
     }
+    RARE_EVENT_INC(set_eval_frame_func);
     interp->eval_frame = eval_frame;
 }
 
diff --git a/Python/specialize.c b/Python/specialize.c
index 13e0440..a9efbe0 100644
--- a/Python/specialize.c
+++ b/Python/specialize.c
@@ -268,6 +268,16 @@ print_optimization_stats(FILE *out, OptimizationStats *stats)
 }
 
 static void
+print_rare_event_stats(FILE *out, RareEventStats *stats)
+{
+    fprintf(out, "Rare event (set_class): %" PRIu64 "\n", stats->set_class);
+    fprintf(out, "Rare event (set_bases): %" PRIu64 "\n", stats->set_bases);
+    fprintf(out, "Rare event (set_eval_frame_func): %" PRIu64 "\n", stats->set_eval_frame_func);
+    fprintf(out, "Rare event (builtin_dict): %" PRIu64 "\n", stats->builtin_dict);
+    fprintf(out, "Rare event (func_modification): %" PRIu64 "\n", stats->func_modification);
+}
+
+static void
 print_stats(FILE *out, PyStats *stats)
 {
     print_spec_stats(out, stats->opcode_stats);
@@ -275,6 +285,7 @@ print_stats(FILE *out, PyStats *stats)
     print_object_stats(out, &stats->object_stats);
     print_gc_stats(out, stats->gc_stats);
     print_optimization_stats(out, &stats->optimization_stats);
+    print_rare_event_stats(out, &stats->rare_event_stats);
 }
 
 void
diff --git a/Tools/scripts/summarize_stats.py b/Tools/scripts/summarize_stats.py
index 1e9dc07..9b7e7b9 100644
--- a/Tools/scripts/summarize_stats.py
+++ b/Tools/scripts/summarize_stats.py
@@ -412,6 +412,14 @@ class Stats:
         rows.sort()
         return rows
 
+    def get_rare_events(self) -> list[tuple[str, int]]:
+        prefix = "Rare event "
+        return [
+            (key[len(prefix) + 1:-1], val)
+            for key, val in self._data.items()
+            if key.startswith(prefix)
+        ]
+
 
 class Count(int):
     def markdown(self) -> str:
@@ -1064,6 +1072,17 @@ def optimization_section() -> Section:
     )
 
 
+def rare_event_section() -> Section:
+    def calc_rare_event_table(stats: Stats) -> Table:
+        return [(x, Count(y)) for x, y in stats.get_rare_events()]
+
+    return Section(
+        "Rare events",
+        "Counts of rare/unlikely events",
+        [Table(("Event", "Count:"), calc_rare_event_table, JoinMode.CHANGE)],
+    )
+
+
 def meta_stats_section() -> Section:
     def calc_rows(stats: Stats) -> Rows:
         return [("Number of data files", Count(stats.get("__nfiles__")))]
@@ -1085,6 +1104,7 @@ LAYOUT = [
     object_stats_section(),
     gc_stats_section(),
     optimization_section(),
+    rare_event_section(),
     meta_stats_section(),
 ]
 
@@ -1162,7 +1182,7 @@ def output_stats(inputs: list[Path], json_output=str | None):
         case 1:
             data = load_raw_data(Path(inputs[0]))
             if json_output is not None:
-                with open(json_output, 'w', encoding='utf-8') as f:
+                with open(json_output, "w", encoding="utf-8") as f:
                     save_raw_data(data, f)  # type: ignore
             stats = Stats(data)
             output_markdown(sys.stdout, LAYOUT, stats)