bpo-40521: Make dict free lists per-interpreter (GH-20645)

Each interpreter now has its own dict free list:

* Move dict free lists into PyInterpreterState.
* Move PyDict_MAXFREELIST define to pycore_interp.h
* Add _Py_dict_state structure.
* Add tstate parameter to _PyDict_ClearFreeList() and _PyDict_Fini().
* In debug mode, ensure that the dict free lists are not used after
  _PyDict_Fini() is called.
* Remove "#ifdef EXPERIMENTAL_ISOLATED_SUBINTERPRETERS".

3 files changed, 26 insertions, 15 deletions

diff --git a/Include/internal/pycore_gc.h b/Include/internal/pycore_gc.h
index fd3fb7f..da202a1 100644
--- a/Include/internal/pycore_gc.h
+++ b/Include/internal/pycore_gc.h
@@ -169,7 +169,7 @@ extern void _PyFrame_ClearFreeList(PyThreadState *tstate);
 extern void _PyTuple_ClearFreeList(PyThreadState *tstate);
 extern void _PyFloat_ClearFreeList(PyThreadState *tstate);
 extern void _PyList_ClearFreeList(PyThreadState *tstate);
-extern void _PyDict_ClearFreeList(void);
+extern void _PyDict_ClearFreeList(PyThreadState *tstate);
 extern void _PyAsyncGen_ClearFreeLists(PyThreadState *tstate);
 extern void _PyContext_ClearFreeList(PyThreadState *tstate);
 
diff --git a/Include/internal/pycore_interp.h b/Include/internal/pycore_interp.h
index 981b733..3f64edc 100644
--- a/Include/internal/pycore_interp.h
+++ b/Include/internal/pycore_interp.h
@@ -69,6 +69,14 @@ struct _Py_unicode_state {
     struct _Py_unicode_fs_codec fs_codec;
 };
 
+struct _Py_float_state {
+    /* Special free list
+       free_list is a singly-linked list of available PyFloatObjects,
+       linked via abuse of their ob_type members. */
+    int numfree;
+    PyFloatObject *free_list;
+};
+
 /* Speed optimization to avoid frequent malloc/free of small tuples */
 #ifndef PyTuple_MAXSAVESIZE
    // Largest tuple to save on free list
@@ -99,12 +107,16 @@ struct _Py_list_state {
     int numfree;
 };
 
-struct _Py_float_state {
-    /* Special free list
-       free_list is a singly-linked list of available PyFloatObjects,
-       linked via abuse of their ob_type members. */
+#ifndef PyDict_MAXFREELIST
+#  define PyDict_MAXFREELIST 80
+#endif
+
+struct _Py_dict_state {
+    /* Dictionary reuse scheme to save calls to malloc and free */
+    PyDictObject *free_list[PyDict_MAXFREELIST];
     int numfree;
-    PyFloatObject *free_list;
+    PyDictKeysObject *keys_free_list[PyDict_MAXFREELIST];
+    int keys_numfree;
 };
 
 struct _Py_frame_state {
@@ -136,7 +148,6 @@ struct _Py_context_state {
 };
 
 
-
 /* interpreter state */
 
 #define _PY_NSMALLPOSINTS           257
@@ -182,8 +193,6 @@ struct _is {
     PyObject *codec_error_registry;
     int codecs_initialized;
 
-    struct _Py_unicode_state unicode;
-
     PyConfig config;
 #ifdef HAVE_DLOPEN
     int dlopenflags;
@@ -224,16 +233,18 @@ struct _is {
     */
     PyLongObject* small_ints[_PY_NSMALLNEGINTS + _PY_NSMALLPOSINTS];
 #endif
+    struct _Py_unicode_state unicode;
+    struct _Py_float_state float_state;
+    /* Using a cache is very effective since typically only a single slice is
+       created and then deleted again. */
+    PySliceObject *slice_cache;
+
     struct _Py_tuple_state tuple;
     struct _Py_list_state list;
-    struct _Py_float_state float_state;
+    struct _Py_dict_state dict_state;
     struct _Py_frame_state frame;
     struct _Py_async_gen_state async_gen;
     struct _Py_context_state context;
-
-    /* Using a cache is very effective since typically only a single slice is
-       created and then deleted again. */
-    PySliceObject *slice_cache;
 };
 
 /* Used by _PyImport_Cleanup() */
diff --git a/Include/internal/pycore_pylifecycle.h b/Include/internal/pycore_pylifecycle.h
index 3e36573..dc99737 100644
--- a/Include/internal/pycore_pylifecycle.h
+++ b/Include/internal/pycore_pylifecycle.h
@@ -59,7 +59,7 @@ extern PyStatus _PyGC_Init(PyThreadState *tstate);
 /* Various internal finalizers */
 
 extern void _PyFrame_Fini(PyThreadState *tstate);
-extern void _PyDict_Fini(void);
+extern void _PyDict_Fini(PyThreadState *tstate);
 extern void _PyTuple_Fini(PyThreadState *tstate);
 extern void _PyList_Fini(PyThreadState *tstate);
 extern void _PySet_Fini(void);