gh-142560: prevent use-after-free in search-like methods by exporting buffer in bytearray (#142938)

Author: fatelei

Lib/test/test_bytes.py

@@ -2060,6 +2060,37 @@ def __index__(self):
             self.assertEqual(instance.ba[0], ord("?"), "Assigned bytearray not altered")
             self.assertEqual(instance.new_ba, bytearray(0x180), "Wrong object altered")
 
+    def test_search_methods_reentrancy_raises_buffererror(self):
+        # gh-142560: Raise BufferError if buffer mutates during search arg conversion.
+        class Evil:
+            def __init__(self, ba):
+                self.ba = ba
+            def __buffer__(self, flags):
+                self.ba.clear()
+                return memoryview(self.ba)
+            def __release_buffer__(self, view: memoryview) -> None:
+                view.release()
+            def __index__(self):
+                self.ba.clear()
+                return ord("A")
+
+        def make_case():
+            ba = bytearray(b"A")
+            return ba, Evil(ba)
+
+        for name in ("find", "count", "index", "rindex", "rfind"):
+            ba, evil = make_case()
+            with self.subTest(name):
+                with self.assertRaises(BufferError):
+                    getattr(ba, name)(evil)
+
+        ba, evil = make_case()
+        with self.assertRaises(BufferError):
+            evil in ba
+        with self.assertRaises(BufferError):
+            ba.split(evil)
+        with self.assertRaises(BufferError):
+            ba.rsplit(evil)
 
 class AssortedBytesTest(unittest.TestCase):
     #

Misc/NEWS.d/next/Library/2025-12-11-22-59-33.gh-issue-142560.GkJrkk.rst

@@ -0,0 +1 @@
+Fix use-after-free in :class:`bytearray` search-like methods (:meth:`~bytearray.find`, :meth:`~bytearray.count`, :meth:`~bytearray.index`, :meth:`~bytearray.rindex`, and :meth:`~bytearray.rfind`) by marking the storage as exported which causes reallocation attempts to raise :exc:`BufferError`. For :func:`~operator.contains`, :meth:`~bytearray.split`, and :meth:`~bytearray.rsplit` the :ref:`buffer protocol <bufferobjects>` is used for this.

Objects/bytearrayobject.c

@@ -90,6 +90,25 @@ bytearray_releasebuffer(PyObject *self, Py_buffer *view)
     Py_END_CRITICAL_SECTION();
 }
 
+typedef PyObject* (*_ba_bytes_op)(const char *buf, Py_ssize_t len,
+                                  PyObject *sub, Py_ssize_t start,
+                                  Py_ssize_t end);
+
+static PyObject *
+_bytearray_with_buffer(PyByteArrayObject *self, _ba_bytes_op op, PyObject *sub,
+                       Py_ssize_t start, Py_ssize_t end)
+{
+    PyObject *res;
+
+    _Py_CRITICAL_SECTION_ASSERT_OBJECT_LOCKED(self);
+
+    /* Increase exports to prevent bytearray storage from changing during op. */
+    self->ob_exports++;
+    res = op(PyByteArray_AS_STRING(self), Py_SIZE(self), sub, start, end);
+    self->ob_exports--;
+    return res;
+}
+
 static int
 _canresize(PyByteArrayObject *self)
 {
@@ -1248,8 +1267,7 @@ bytearray_find_impl(PyByteArrayObject *self, PyObject *sub, Py_ssize_t start,
                     Py_ssize_t end)
 /*[clinic end generated code: output=413e1cab2ae87da0 input=df3aa94840d893a7]*/
 {
-    return _Py_bytes_find(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                          sub, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_find, sub, start, end);
 }
 
 /*[clinic input]
@@ -1265,8 +1283,7 @@ bytearray_count_impl(PyByteArrayObject *self, PyObject *sub,
                      Py_ssize_t start, Py_ssize_t end)
 /*[clinic end generated code: output=a21ee2692e4f1233 input=e8fcdca8272857e0]*/
 {
-    return _Py_bytes_count(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                           sub, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_count, sub, start, end);
 }
 
 /*[clinic input]
@@ -1314,8 +1331,7 @@ bytearray_index_impl(PyByteArrayObject *self, PyObject *sub,
                      Py_ssize_t start, Py_ssize_t end)
 /*[clinic end generated code: output=067a1e78efc672a7 input=c37f177cfee19fe4]*/
 {
-    return _Py_bytes_index(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                           sub, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_index, sub, start, end);
 }
 
 /*[clinic input]
@@ -1333,8 +1349,7 @@ bytearray_rfind_impl(PyByteArrayObject *self, PyObject *sub,
                      Py_ssize_t start, Py_ssize_t end)
 /*[clinic end generated code: output=51bf886f932b283c input=1265b11c437d2750]*/
 {
-    return _Py_bytes_rfind(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                           sub, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_rfind, sub, start, end);
 }
 
 /*[clinic input]
@@ -1352,18 +1367,22 @@ bytearray_rindex_impl(PyByteArrayObject *self, PyObject *sub,
                       Py_ssize_t start, Py_ssize_t end)
 /*[clinic end generated code: output=38e1cf66bafb08b9 input=7d198b3d6b0a62ce]*/
 {
-    return _Py_bytes_rindex(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                            sub, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_rindex, sub, start, end);
 }
 
 static int
 bytearray_contains(PyObject *self, PyObject *arg)
 {
-    int ret;
+    int ret = -1;
     Py_BEGIN_CRITICAL_SECTION(self);
-    ret = _Py_bytes_contains(PyByteArray_AS_STRING(self),
+    PyByteArrayObject *ba = _PyByteArray_CAST(self);
+
+    /* Increase exports to prevent bytearray storage from changing during _Py_bytes_contains(). */
+    ba->ob_exports++;
+    ret = _Py_bytes_contains(PyByteArray_AS_STRING(ba),
                              PyByteArray_GET_SIZE(self),
                              arg);
+    ba->ob_exports--;
     Py_END_CRITICAL_SECTION();
     return ret;
 }
@@ -1390,8 +1409,7 @@ bytearray_startswith_impl(PyByteArrayObject *self, PyObject *subobj,
                           Py_ssize_t start, Py_ssize_t end)
 /*[clinic end generated code: output=a3d9b6d44d3662a6 input=93f9ffee684f109a]*/
 {
-    return _Py_bytes_startswith(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                                subobj, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_startswith, subobj, start, end);
 }
 
 /*[clinic input]
@@ -1416,8 +1434,7 @@ bytearray_endswith_impl(PyByteArrayObject *self, PyObject *subobj,
                         Py_ssize_t start, Py_ssize_t end)
 /*[clinic end generated code: output=e75ea8c227954caa input=d158b030a11d0b06]*/
 {
-    return _Py_bytes_endswith(PyByteArray_AS_STRING(self), PyByteArray_GET_SIZE(self),
-                              subobj, start, end);
+    return _bytearray_with_buffer(self, _Py_bytes_endswith, subobj, start, end);
 }
 
 /*[clinic input]
@@ -1782,26 +1799,32 @@ bytearray_split_impl(PyByteArrayObject *self, PyObject *sep,
                      Py_ssize_t maxsplit)
 /*[clinic end generated code: output=833e2cf385d9a04d input=dd9f6e2910cc3a34]*/
 {
-    Py_ssize_t len = PyByteArray_GET_SIZE(self), n;
-    const char *s = PyByteArray_AS_STRING(self), *sub;
-    PyObject *list;
-    Py_buffer vsub;
+    PyObject *list = NULL;
+
+    /* Increase exports to prevent bytearray storage from changing during _Py_bytes_contains(). */
+    self->ob_exports++;
+    const char *sbuf = PyByteArray_AS_STRING(self);
+    Py_ssize_t slen = PyByteArray_GET_SIZE((PyObject *)self);
 
     if (maxsplit < 0)
         maxsplit = PY_SSIZE_T_MAX;
 
-    if (sep == Py_None)
-        return stringlib_split_whitespace((PyObject*) self, s, len, maxsplit);
+    if (sep == Py_None) {
+        list = stringlib_split_whitespace((PyObject*)self, sbuf, slen, maxsplit);
+        goto done;
+    }
 
-    if (PyObject_GetBuffer(sep, &vsub, PyBUF_SIMPLE) != 0)
-        return NULL;
-    sub = vsub.buf;
-    n = vsub.len;
+    Py_buffer vsub;
+    if (PyObject_GetBuffer(sep, &vsub, PyBUF_SIMPLE) != 0) {
+        goto done;
+    }
 
-    list = stringlib_split(
-        (PyObject*) self, s, len, sub, n, maxsplit
-        );
+    list = stringlib_split((PyObject*)self, sbuf, slen,
+                           (const char *)vsub.buf, vsub.len, maxsplit);
     PyBuffer_Release(&vsub);
+
+done:
+    self->ob_exports--;
     return list;
 }
 
@@ -1900,26 +1923,32 @@ bytearray_rsplit_impl(PyByteArrayObject *self, PyObject *sep,
                       Py_ssize_t maxsplit)
 /*[clinic end generated code: output=a55e0b5a03cb6190 input=60e9abf305128ff4]*/
 {
-    Py_ssize_t len = PyByteArray_GET_SIZE(self), n;
-    const char *s = PyByteArray_AS_STRING(self), *sub;
-    PyObject *list;
-    Py_buffer vsub;
+    PyObject *list = NULL;
+
+    /* Increase exports to prevent bytearray storage from changing during _Py_bytes_contains(). */
+    self->ob_exports++;
+    const char *sbuf = PyByteArray_AS_STRING(self);
+    Py_ssize_t slen = PyByteArray_GET_SIZE((PyObject *)self);
 
     if (maxsplit < 0)
         maxsplit = PY_SSIZE_T_MAX;
 
-    if (sep == Py_None)
-        return stringlib_rsplit_whitespace((PyObject*) self, s, len, maxsplit);
+    if (sep == Py_None) {
+        list = stringlib_rsplit_whitespace((PyObject*)self, sbuf, slen, maxsplit);
+        goto done;
+    }
 
-    if (PyObject_GetBuffer(sep, &vsub, PyBUF_SIMPLE) != 0)
-        return NULL;
-    sub = vsub.buf;
-    n = vsub.len;
+    Py_buffer vsub;
+    if (PyObject_GetBuffer(sep, &vsub, PyBUF_SIMPLE) != 0) {
+        goto done;
+    }
 
-    list = stringlib_rsplit(
-        (PyObject*) self, s, len, sub, n, maxsplit
-        );
+    list = stringlib_rsplit((PyObject*)self, sbuf, slen,
+                            (const char *)vsub.buf, vsub.len, maxsplit);
     PyBuffer_Release(&vsub);
+
+done:
+    self->ob_exports--;
     return list;
 }