[3.13] gh-117511: Make PyMutex public in the non-limited API (GH-117731) (#120800)

(cherry picked from commit 3af7263)

Author: colesbury

Doc/c-api/init.rst

@@ -55,6 +55,11 @@ The following functions can be safely called before Python is initialized:
   * :c:func:`PyMem_RawCalloc`
   * :c:func:`PyMem_RawFree`
 
+* Synchronization:
+
+  * :c:func:`PyMutex_Lock`
+  * :c:func:`PyMutex_Unlock`
+
 .. note::
 
    The following functions **should not be called** before
@@ -2152,3 +2157,41 @@ be used in new code.
 .. c:function:: void PyThread_delete_key_value(int key)
 .. c:function:: void PyThread_ReInitTLS()
 
+Synchronization Primitives
+==========================
+
+The C-API provides a basic mutual exclusion lock.
+
+.. c:type:: PyMutex
+
+   A mutual exclusion lock.  The :c:type:`!PyMutex` should be initialized to
+   zero to represent the unlocked state.  For example::
+
+      PyMutex mutex = {0};
+
+   Instances of :c:type:`!PyMutex` should not be copied or moved.  Both the
+   contents and address of a :c:type:`!PyMutex` are meaningful, and it must
+   remain at a fixed, writable location in memory.
+
+   .. note::
+
+      A :c:type:`!PyMutex` currently occupies one byte, but the size should be
+      considered unstable.  The size may change in future Python releases
+      without a deprecation period.
+
+   .. versionadded:: 3.13
+
+.. c:function:: void PyMutex_Lock(PyMutex *m)
+
+   Lock mutex *m*.  If another thread has already locked it, the calling
+   thread will block until the mutex is unlocked.  While blocked, the thread
+   will temporarily release the :term:`GIL` if it is held.
+
+   .. versionadded:: 3.13
+
+.. c:function:: void PyMutex_Unlock(PyMutex *m)
+
+   Unlock mutex *m*. The mutex must be locked --- otherwise, the function will
+   issue a fatal error.
+
+   .. versionadded:: 3.13

Doc/whatsnew/3.13.rst

@@ -2160,6 +2160,11 @@ New Features
   :c:func:`PyEval_GetLocals` return :term:`strong references <strong reference>`
   rather than borrowed references. (Added as part of :pep:`667`.)
 
+* Add :c:type:`PyMutex` API, a lightweight mutex that occupies a single byte.
+  The :c:func:`PyMutex_Lock` function will release the GIL (if currently held)
+  if the operation needs to block.
+  (Contributed by Sam Gross in :gh:`108724`.)
+
 Build Changes
 =============
 

Include/Python.h

@@ -64,6 +64,7 @@
 #include "pybuffer.h"
 #include "pystats.h"
 #include "pyatomic.h"
+#include "lock.h"
 #include "object.h"
 #include "objimpl.h"
 #include "typeslots.h"

Include/cpython/lock.h

@@ -0,0 +1,63 @@
+#ifndef Py_CPYTHON_LOCK_H
+#  error "this header file must not be included directly"
+#endif
+
+#define _Py_UNLOCKED    0
+#define _Py_LOCKED      1
+
+// A mutex that occupies one byte. The lock can be zero initialized to
+// represent the unlocked state.
+//
+// Typical initialization:
+//   PyMutex m = (PyMutex){0};
+//
+// Or initialize as global variables:
+//   static PyMutex m;
+//
+// Typical usage:
+//   PyMutex_Lock(&m);
+//   ...
+//   PyMutex_Unlock(&m);
+//
+// The contents of the PyMutex are not part of the public API, but are
+// described to aid in understanding the implementation and debugging. Only
+// the two least significant bits are used. The remaining bits are always zero:
+// 0b00: unlocked
+// 0b01: locked
+// 0b10: unlocked and has parked threads
+// 0b11: locked and has parked threads
+typedef struct PyMutex {
+    uint8_t _bits;  // (private)
+} PyMutex;
+
+// exported function for locking the mutex
+PyAPI_FUNC(void) PyMutex_Lock(PyMutex *m);
+
+// exported function for unlocking the mutex
+PyAPI_FUNC(void) PyMutex_Unlock(PyMutex *m);
+
+// Locks the mutex.
+//
+// If the mutex is currently locked, the calling thread will be parked until
+// the mutex is unlocked. If the current thread holds the GIL, then the GIL
+// will be released while the thread is parked.
+static inline void
+_PyMutex_Lock(PyMutex *m)
+{
+    uint8_t expected = _Py_UNLOCKED;
+    if (!_Py_atomic_compare_exchange_uint8(&m->_bits, &expected, _Py_LOCKED)) {
+        PyMutex_Lock(m);
+    }
+}
+#define PyMutex_Lock _PyMutex_Lock
+
+// Unlocks the mutex.
+static inline void
+_PyMutex_Unlock(PyMutex *m)
+{
+    uint8_t expected = _Py_LOCKED;
+    if (!_Py_atomic_compare_exchange_uint8(&m->_bits, &expected, _Py_UNLOCKED)) {
+        PyMutex_Unlock(m);
+    }
+}
+#define PyMutex_Unlock _PyMutex_Unlock

Include/cpython/weakrefobject.h

@@ -36,7 +36,7 @@ struct _PyWeakReference {
      * Normally this can be derived from wr_object, but in some cases we need
      * to lock after wr_object has been set to Py_None.
      */
-    struct _PyMutex *weakrefs_lock;
+    PyMutex *weakrefs_lock;
 #endif
 };
 

Include/internal/pycore_critical_section.h

@@ -202,7 +202,7 @@ _PyCriticalSection2_BeginSlow(_PyCriticalSection2 *c, PyMutex *m1, PyMutex *m2,
 static inline void
 _PyCriticalSection_Begin(_PyCriticalSection *c, PyMutex *m)
 {
-    if (PyMutex_LockFast(&m->v)) {
+    if (PyMutex_LockFast(&m->_bits)) {
         PyThreadState *tstate = _PyThreadState_GET();
         c->mutex = m;
         c->prev = tstate->critical_section;
@@ -255,8 +255,8 @@ _PyCriticalSection2_Begin(_PyCriticalSection2 *c, PyMutex *m1, PyMutex *m2)
         m2 = tmp;
     }
 
-    if (PyMutex_LockFast(&m1->v)) {
-        if (PyMutex_LockFast(&m2->v)) {
+    if (PyMutex_LockFast(&m1->_bits)) {
+        if (PyMutex_LockFast(&m2->_bits)) {
             PyThreadState *tstate = _PyThreadState_GET();
             c->base.mutex = m1;
             c->mutex2 = m2;

Include/internal/pycore_lock.h

@@ -13,84 +13,22 @@ extern "C" {
 #  error "this header requires Py_BUILD_CORE define"
 #endif
 
-
-// A mutex that occupies one byte. The lock can be zero initialized.
-//
-// Only the two least significant bits are used. The remaining bits should be
-// zero:
-// 0b00: unlocked
-// 0b01: locked
-// 0b10: unlocked and has parked threads
-// 0b11: locked and has parked threads
-//
-// Typical initialization:
-//   PyMutex m = (PyMutex){0};
-//
-// Or initialize as global variables:
-//   static PyMutex m;
-//
-// Typical usage:
-//   PyMutex_Lock(&m);
-//   ...
-//   PyMutex_Unlock(&m);
-
-// NOTE: In Py_GIL_DISABLED builds, `struct _PyMutex` is defined in Include/object.h.
-// The Py_GIL_DISABLED builds need the definition in Include/object.h for the
-// `ob_mutex` field in PyObject. For the default (non-free-threaded) build,
-// we define the struct here to avoid exposing it in the public API.
-#ifndef Py_GIL_DISABLED
-struct _PyMutex { uint8_t v; };
-#endif
-
-typedef struct _PyMutex PyMutex;
-
-#define _Py_UNLOCKED    0
-#define _Py_LOCKED      1
+//_Py_UNLOCKED is defined as 0 and _Py_LOCKED as 1 in Include/cpython/lock.h
 #define _Py_HAS_PARKED  2
 #define _Py_ONCE_INITIALIZED 4
 
-// (private) slow path for locking the mutex
-PyAPI_FUNC(void) _PyMutex_LockSlow(PyMutex *m);
-
-// (private) slow path for unlocking the mutex
-PyAPI_FUNC(void) _PyMutex_UnlockSlow(PyMutex *m);
-
 static inline int
 PyMutex_LockFast(uint8_t *lock_bits)
 {
     uint8_t expected = _Py_UNLOCKED;
     return _Py_atomic_compare_exchange_uint8(lock_bits, &expected, _Py_LOCKED);
 }
 
-// Locks the mutex.
-//
-// If the mutex is currently locked, the calling thread will be parked until
-// the mutex is unlocked. If the current thread holds the GIL, then the GIL
-// will be released while the thread is parked.
-static inline void
-PyMutex_Lock(PyMutex *m)
-{
-    uint8_t expected = _Py_UNLOCKED;
-    if (!_Py_atomic_compare_exchange_uint8(&m->v, &expected, _Py_LOCKED)) {
-        _PyMutex_LockSlow(m);
-    }
-}
-
-// Unlocks the mutex.
-static inline void
-PyMutex_Unlock(PyMutex *m)
-{
-    uint8_t expected = _Py_LOCKED;
-    if (!_Py_atomic_compare_exchange_uint8(&m->v, &expected, _Py_UNLOCKED)) {
-        _PyMutex_UnlockSlow(m);
-    }
-}
-
 // Checks if the mutex is currently locked.
 static inline int
 PyMutex_IsLocked(PyMutex *m)
 {
-    return (_Py_atomic_load_uint8(&m->v) & _Py_LOCKED) != 0;
+    return (_Py_atomic_load_uint8(&m->_bits) & _Py_LOCKED) != 0;
 }
 
 // Re-initializes the mutex after a fork to the unlocked state.
@@ -121,7 +59,7 @@ static inline void
 PyMutex_LockFlags(PyMutex *m, _PyLockFlags flags)
 {
     uint8_t expected = _Py_UNLOCKED;
-    if (!_Py_atomic_compare_exchange_uint8(&m->v, &expected, _Py_LOCKED)) {
+    if (!_Py_atomic_compare_exchange_uint8(&m->_bits, &expected, _Py_LOCKED)) {
         _PyMutex_LockTimed(m, -1, flags);
     }
 }

Include/internal/pycore_warnings.h

@@ -14,7 +14,7 @@ struct _warnings_runtime_state {
     PyObject *filters;  /* List */
     PyObject *once_registry;  /* Dict */
     PyObject *default_action; /* String */
-    struct _PyMutex mutex;
+    PyMutex mutex;
     long filters_version;
 };
 

Include/lock.h

@@ -0,0 +1,16 @@
+#ifndef Py_LOCK_H
+#define Py_LOCK_H
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#ifndef Py_LIMITED_API
+#  define Py_CPYTHON_LOCK_H
+#  include "cpython/lock.h"
+#  undef Py_CPYTHON_LOCK_H
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_LOCK_H */

Include/object.h

@@ -204,17 +204,13 @@ struct _object {
 // Create a shared field from a refcnt and desired flags
 #define _Py_REF_SHARED(refcnt, flags) (((refcnt) << _Py_REF_SHARED_SHIFT) + (flags))
 
-// NOTE: In non-free-threaded builds, `struct _PyMutex` is defined in
-// pycore_lock.h. See pycore_lock.h for more details.
-struct _PyMutex { uint8_t v; };
-
 struct _object {
     // ob_tid stores the thread id (or zero). It is also used by the GC and the
     // trashcan mechanism as a linked list pointer and by the GC to store the
     // computed "gc_refs" refcount.
     uintptr_t ob_tid;
     uint16_t _padding;
-    struct _PyMutex ob_mutex;   // per-object lock
+    PyMutex ob_mutex;           // per-object lock
     uint8_t ob_gc_bits;         // gc-related state
     uint32_t ob_ref_local;      // local reference count
     Py_ssize_t ob_ref_shared;   // shared (atomic) reference count