gh-135871: Fix needless spinning in _PyMutex_LockTimed (timeout==0)

* Move the timeout == 0 guard outside the else so a non-blocking call returns immediately after a failed CAS instead of entering the spin loop. * Reload v on every spin iteration, allowing timed/blocking callers to notice an unlock promptly. No-GIL builds now honor the semantics of non-blocking attempts and avoid wasted CPU; GIL builds are unaffected (MAX_SPIN_COUNT == 0).
python · colesbury · Jun 25, 2025 · Jun 24, 2025 · Jun 24, 2025 · Jun 24, 2025
commit 7296a7d133c078f1b9d6294f64dc653c9b03eb56
diff --git a/Misc/NEWS.d/next/Core and Builtins/2025-06-23-18-08-32.gh-issue-135871.50C528.rst b/Misc/NEWS.d/next/Core and Builtins/2025-06-23-18-08-32.gh-issue-135871.50C528.rst
@@ -0,0 +1 @@
+gh-135871: _PyMutex_LockTimed() no longer spins or fails for timeout==0 in no-GIL builds; spin loop now reloads lock state.
diff --git a/Python/lock.c b/Python/lock.c
@@ -58,7 +58,7 @@ _PyMutex_LockTimed(PyMutex *m, PyTime_t timeout, _PyLockFlags flags)
             return PY_LOCK_ACQUIRED;
         }
     }
-    else if (timeout == 0) {
+    if (timeout == 0) {
         return PY_LOCK_FAILURE;
     }
 
@@ -89,6 +89,7 @@ _PyMutex_LockTimed(PyMutex *m, PyTime_t timeout, _PyLockFlags flags)
             // Spin for a bit.
             _Py_yield();
             spin_count++;
+            v = _Py_atomic_load_uint8_relaxed(&m->_bits);
             continue;
         }
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		gh-135871: _PyMutex_LockTimed() no longer spins or fails for timeout==0 in no-GIL builds; spin loop now reloads lock state.
jtibbertsma marked this conversation as resolved. Show resolved Hide resolved