gh-149965: Improve thread-safety in ElementObjects within element_ass_subscr

Misc/NEWS.d/next/Library/2026-05-17-22-42-54.gh-issue-149965.OFW3ZD.rst

@@ -0,0 +1 @@
+WIthin ``Modules/_elementtree.c``, improve thread-safety in ElementObjects within ``element_ass_subscr``

Modules/_elementtree.c

@@ -1863,6 +1863,35 @@ element_subscr(PyObject *op, PyObject *item)
     }
 }
 
+// Pointer-by-pointer memmove for PyObject** arrays that is safe
+// for shared ElementObjects in Py_GIL_DISABLED builds.
+static void
+ptr_wise_atomic_memmove(ElementObject *a, PyObject **dest, PyObject **src, Py_ssize_t n)
+{
+#ifndef Py_GIL_DISABLED
+    memmove(dest, src, n * sizeof(PyObject *));
+#else
+    // XXX: maybe a critical section isn't needed for ElementObject?
+    // _Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(a);
+    if (_Py_IsOwnedByCurrentThread((PyObject *)a) && !_PyObject_GC_IS_SHARED(a)) {
+        // No other threads can read this list concurrently
+        memmove(dest, src, n * sizeof(PyObject *));
+        return;
+    }
+    if (dest < src) {
+        for (Py_ssize_t i = 0; i != n; i++) {
+            _Py_atomic_store_ptr_release(&dest[i], src[i]);
+        }
+    }
+    else {
+        // copy backwards to avoid overwriting src before it's read
+        for (Py_ssize_t i = n; i != 0; i--) {
+            _Py_atomic_store_ptr_release(&dest[i - 1], src[i - 1]);
+        }
+    }
+#endif
+}
+
 static int
 element_ass_subscr(PyObject *op, PyObject *item, PyObject *value)
 {
@@ -1938,19 +1967,21 @@ element_ass_subscr(PyObject *op, PyObject *item, PyObject *value)
 
                 PyList_SET_ITEM(recycle, i, self->extra->children[cur]);
 
-                memmove(
+                ptr_wise_atomic_memmove(
+                    self,
                     self->extra->children + cur - i,
                     self->extra->children + cur + 1,
-                    num_moved * sizeof(PyObject *));
+                    num_moved);
             }
 
             /* Leftover "tail" after the last removed child */
             cur = start + (size_t)slicelen * step;
             if (cur < (size_t)self->extra->length) {
-                memmove(
+                ptr_wise_atomic_memmove(
+                    self,
                     self->extra->children + cur - slicelen,
                     self->extra->children + cur,
-                    (self->extra->length - cur) * sizeof(PyObject *));
+                    self->extra->length - cur);
             }
 
             self->extra->length -= slicelen;