gh-116738: Make _json module safe in the free-threading build

Lib/test/test_free_threading/test_json.py

@@ -0,0 +1,84 @@
+import unittest
+from threading import Barrier, Thread
+from test.test_json import CTest
+from test.support import threading_helper
+
+
+def encode_json_helper(json, worker, data, number_of_threads, number_of_json_encodings=100):
+    worker_threads = []
+    barrier = Barrier(number_of_threads)
+    for index in range(number_of_threads):
+        worker_threads.append(Thread(target=worker, args=[barrier, data, index]))
+    for t in worker_threads:
+        t.start()
+    for ii in range(number_of_json_encodings):
+        json.dumps(data)
+    data.clear()
+    for t in worker_threads:
+        t.join()
+
+
+class MyMapping(dict):
+    def __init__(self):
+        self.mapping = []
+
+    def items(self):
+        return self.mapping
+
+
+@threading_helper.reap_threads
+@threading_helper.requires_working_threading()
+class TestJsonEncoding(CTest):
+    # Test encoding json with multiple threads modifying the data cannot
+    # corrupt the interpreter
+
+    def test_json_mutating_list(self):
+
+        def worker(barrier, data, index):
+            barrier.wait()
+            while data:
+                for d in data:
+                    if len(d) > 5:
+                        d.clear()
+                    else:
+                        d.append(index)
+                        d.append(index)
+                        d.append(index)
+        encode_json_helper(self.json, worker,  [[], []], number_of_threads=16)
+
+    def test_json_mutating_dict(self):
+
+        def worker(barrier, data, index):
+            barrier.wait()
+            while data:
+                for d in data:
+                    if len(d) > 5:
+                        try:
+                            d.pop(list(d)[0])
+                        except (KeyError, IndexError):
+                            pass
+                    else:
+                        d[index] = index
+        encode_json_helper(self.json, worker, [{}, {}], number_of_threads=16)
+
+    def test_json_mutating_mapping(self):
+
+        def worker(barrier, data, index):
+            barrier.wait()
+            while data:
+                for d in data:
+                    if len(d.mapping) > 3:
+                        d.mapping.clear()
+                    else:
+                        d.mapping.append((index, index))
+        encode_json_helper(self.json,
+                           worker, [MyMapping(), MyMapping()], number_of_threads=16)
+
+
+if __name__ == "__main__":
+    import time
+
+    t0 = time.time()
+    unittest.main()
+    dt = time.time()-t0
+    print(f'Done: {dt:.2f}')

Misc/NEWS.d/next/Core_and_Builtins/2024-06-04-20-26-21.gh-issue-116738.q_hPYq.rst

@@ -0,0 +1 @@
+Make the module :mod:`json` safe to use under the free-theading build.

Modules/_json.c

@@ -10,6 +10,7 @@
 
 #include "Python.h"
 #include "pycore_ceval.h"         // _Py_EnterRecursiveCall()
+#include "pycore_critical_section.h"  // Py_BEGIN_CRITICAL_SECTION()
 #include "pycore_global_strings.h" // _Py_ID()
 #include "pycore_pyerrors.h"      // _PyErr_FormatNote
 #include "pycore_runtime.h"       // _PyRuntime
@@ -1636,15 +1637,58 @@ encoder_encode_key_value(PyEncoderObject *s, PyUnicodeWriter *writer, bool *firs
     return 0;
 }
 
+static inline int
+_encoder_iterate_mapping_lock_held(PyEncoderObject *s, PyUnicodeWriter *writer,
+                            bool *first, PyObject *dct, PyObject *items,
+                            Py_ssize_t indent_level, PyObject *indent_cache,
+                            PyObject *separator)
+{
+    PyObject *key, *value;
+    for (Py_ssize_t  i = 0; i < PyList_GET_SIZE(items); i++) {
+        PyObject *item = PyList_GET_ITEM(items, i);
+
+        if (!PyTuple_Check(item) || PyTuple_GET_SIZE(item) != 2) {
+            PyErr_SetString(PyExc_ValueError, "items must return 2-tuples");
+            return -1;
+        }
+
+        key = PyTuple_GET_ITEM(item, 0);
+        value = PyTuple_GET_ITEM(item, 1);
+        if (encoder_encode_key_value(s, writer, first, dct, key, value,
+                                     indent_level, indent_cache,
+                                     separator) < 0) {
+            return -1;
+        }
+    }
+
+    return 0;
+}
+
+static inline int
+_encoder_iterate_dict_lock_held(PyEncoderObject *s, PyUnicodeWriter *writer,
+                         bool *first, PyObject *dct, Py_ssize_t indent_level,
+                         PyObject *indent_cache, PyObject *separator)
+{
+    PyObject *key, *value;
+    Py_ssize_t pos = 0;
+    while (PyDict_Next(dct, &pos, &key, &value)) {
+        if (encoder_encode_key_value(s, writer, first, dct, key, value,
+                                    indent_level, indent_cache,
+                                    separator) < 0) {
+            return -1;
+        }
+    }
+    return 0;
+}
+
 static int
 encoder_listencode_dict(PyEncoderObject *s, PyUnicodeWriter *writer,
-                        PyObject *dct,
+                       PyObject *dct,
                        Py_ssize_t indent_level, PyObject *indent_cache)
 {
     /* Encode Python dict dct a JSON term */
     PyObject *ident = NULL;
     PyObject *items = NULL;
-    PyObject *key, *value;
     bool first = true;
 
     if (PyDict_GET_SIZE(dct) == 0) {
@@ -1682,33 +1726,29 @@ encoder_listencode_dict(PyEncoderObject *s, PyUnicodeWriter *writer,
 
     if (s->sort_keys || !PyDict_CheckExact(dct)) {
         items = PyMapping_Items(dct);
-        if (items == NULL || (s->sort_keys && PyList_Sort(items) < 0))
+        if (items == NULL || (s->sort_keys && PyList_Sort(items) < 0)) {
             goto bail;
-
-        for (Py_ssize_t  i = 0; i < PyList_GET_SIZE(items); i++) {
-            PyObject *item = PyList_GET_ITEM(items, i);
-
-            if (!PyTuple_Check(item) || PyTuple_GET_SIZE(item) != 2) {
-                PyErr_SetString(PyExc_ValueError, "items must return 2-tuples");
-                goto bail;
-            }
-
-            key = PyTuple_GET_ITEM(item, 0);
-            value = PyTuple_GET_ITEM(item, 1);
-            if (encoder_encode_key_value(s, writer, &first, dct, key, value,
-                                         indent_level, indent_cache,
-                                         separator) < 0)
-                goto bail;
         }
+
+        int result;
+        Py_BEGIN_CRITICAL_SECTION_SEQUENCE_FAST(items);
+        result = _encoder_iterate_mapping_lock_held(s, writer, &first, dct,
+                    items, indent_level, indent_cache, separator);
+        Py_END_CRITICAL_SECTION_SEQUENCE_FAST();
         Py_CLEAR(items);
+        if (result < 0) {
+            Py_XDECREF(items);
+            goto bail;
+        }
 
     } else {
-        Py_ssize_t pos = 0;
-        while (PyDict_Next(dct, &pos, &key, &value)) {
-            if (encoder_encode_key_value(s, writer, &first, dct, key, value,
-                                         indent_level, indent_cache,
-                                         separator) < 0)
-                goto bail;
+        int result;
+        Py_BEGIN_CRITICAL_SECTION(dct);
+        result = _encoder_iterate_dict_lock_held(s, writer, &first, dct,
+                            indent_level, indent_cache, separator);
+        Py_END_CRITICAL_SECTION();
+        if (result < 0) {
+            goto bail;
         }
     }
 
@@ -1735,17 +1775,35 @@ encoder_listencode_dict(PyEncoderObject *s, PyUnicodeWriter *writer,
     return -1;
 }
 
+static inline int
+_encoder_iterate_fast_seq_lock_held(PyEncoderObject *s, PyUnicodeWriter *writer,
+    PyObject *seq, PyObject *s_fast,
+    Py_ssize_t indent_level, PyObject *indent_cache, PyObject *separator)
+{
+    for (Py_ssize_t i = 0; i < PySequence_Fast_GET_SIZE(s_fast); i++) {
+        PyObject *obj = PySequence_Fast_GET_ITEM(s_fast, i);
+        if (i) {
+            if (PyUnicodeWriter_WriteStr(writer, separator) < 0) {
+                return -1;
+            }
+        }
+        if (encoder_listencode_obj(s, writer, obj, indent_level, indent_cache)) {
+            _PyErr_FormatNote("when serializing %T item %zd", seq, i);
+            return -1;
+        }
+    }
+    return 0;
+}
+
 static int
 encoder_listencode_list(PyEncoderObject *s, PyUnicodeWriter *writer,
                         PyObject *seq,
                         Py_ssize_t indent_level, PyObject *indent_cache)
 {
     PyObject *ident = NULL;
     PyObject *s_fast = NULL;
-    Py_ssize_t i;
 
-    ident = NULL;
-    s_fast = PySequence_Fast(seq, "_iterencode_list needs a sequence");
+    s_fast = PySequence_Fast(seq, "encoder_listencode_list needs a sequence");
     if (s_fast == NULL)
         return -1;
     if (PySequence_Fast_GET_SIZE(s_fast) == 0) {
@@ -1783,16 +1841,13 @@ encoder_listencode_list(PyEncoderObject *s, PyUnicodeWriter *writer,
             goto bail;
         }
     }
-    for (i = 0; i < PySequence_Fast_GET_SIZE(s_fast); i++) {
-        PyObject *obj = PySequence_Fast_GET_ITEM(s_fast, i);
-        if (i) {
-            if (PyUnicodeWriter_WriteStr(writer, separator) < 0)
-                goto bail;
-        }
-        if (encoder_listencode_obj(s, writer, obj, indent_level, indent_cache)) {
-            _PyErr_FormatNote("when serializing %T item %zd", seq, i);
-            goto bail;
-        }
+    int result;
+    Py_BEGIN_CRITICAL_SECTION_SEQUENCE_FAST(seq);
+    result = _encoder_iterate_fast_seq_lock_held(s, writer, seq, s_fast,
+                     indent_level, indent_cache, separator);
+    Py_END_CRITICAL_SECTION_SEQUENCE_FAST();
+    if (result < 0) {
+        goto bail;
     }
     if (ident != NULL) {
         if (PyDict_DelItem(s->markers, ident))