cpython/Modules/_xxtestfuzz
Victor Stinner 89f9875448
gh-106320: Move private _PyHash API to the internal C API (#107026)
* No longer export most private _PyHash symbols, only export the ones
  which are needed by shared extensions.
* Modules/_xxtestfuzz/fuzzer.c now uses the internal C API.
2023-07-22 13:49:37 +00:00
..
dictionaries bpo-29505: Add more fuzzing for re.compile, re.load and csv.reader (GH-14255) 2019-06-29 22:54:42 -07:00
fuzz_csv_reader_corpus bpo-29505: Add more fuzzing for re.compile, re.load and csv.reader (GH-14255) 2019-06-29 22:54:42 -07:00
fuzz_json_loads_corpus bpo-29505: Fuzz json module, enforce size limit on int(x) fuzz (GH-13991) 2019-06-11 21:30:34 -07:00
fuzz_sre_compile_corpus bpo-29505: Add more fuzzing for re.compile, re.load and csv.reader (GH-14255) 2019-06-29 22:54:42 -07:00
fuzz_struct_unpack_corpus Fuzz struct.unpack and catch RecursionError in re.compile (GH-18679) 2020-02-27 23:05:02 -08:00
README.rst bpo-29505: Fuzz json module, enforce size limit on int(x) fuzz (GH-13991) 2019-06-11 21:30:34 -07:00
_xxtestfuzz.c gh-104922: remove PY_SSIZE_T_CLEAN (#106315) 2023-07-02 15:07:46 +09:00
fuzz_tests.txt bpo-29505: Add fuzzer for ast.literal_eval (GH-28777) 2021-10-06 16:22:09 -07:00
fuzzer.c gh-106320: Move private _PyHash API to the internal C API (#107026) 2023-07-22 13:49:37 +00:00

README.rst

Fuzz Tests for CPython
======================

These fuzz tests are designed to be included in Google's `oss-fuzz`_ project.

oss-fuzz works against a library exposing a function of the form
``int LLVMFuzzerTestOneInput(const uint8_t* data, size_t length)``. We provide
that library (``fuzzer.c``), and include a ``_fuzz`` module for testing with
some toy values -- no fuzzing occurs in Python's test suite.

oss-fuzz will regularly pull from CPython, discover all the tests in
``fuzz_tests.txt``, and run them -- so adding a new test here means it will
automatically be run in oss-fuzz, while also being smoke-tested as part of
CPython's test suite.

Adding a new fuzz test
----------------------

Add the test name on a new line in ``fuzz_tests.txt``.

In ``fuzzer.c``, add a function to be run::

    int $test_name (const char* data, size_t size) {
        ...
        return 0;
    }


And invoke it from ``LLVMFuzzerTestOneInput``::

    #if _Py_FUZZ_YES(fuzz_builtin_float)
        rv |= _run_fuzz(data, size, fuzz_builtin_float);
    #endif

``LLVMFuzzerTestOneInput`` will run in oss-fuzz, with each test in
``fuzz_tests.txt`` run separately.

Seed data (corpus) for the test can be provided in a subfolder called
``<test_name>_corpus`` such as ``fuzz_json_loads_corpus``. A wide variety
of good input samples allows the fuzzer to more easily explore a diverse
set of paths and provides a better base to find buggy input from.

Dictionaries of tokens (see oss-fuzz documentation for more details) can
be placed in the ``dictionaries`` folder with the name of the test.
For example, ``dictionaries/fuzz_json_loads.dict`` contains JSON tokens
to guide the fuzzer.

What makes a good fuzz test
---------------------------

Libraries written in C that might handle untrusted data are worthwhile. The
more complex the logic (e.g. parsing), the more likely this is to be a useful
fuzz test. See the existing examples for reference, and refer to the
`oss-fuzz`_ docs.

.. _oss-fuzz: https://github.com/google/oss-fuzz