Skip to content

Commit

Permalink
Fixed CVE-2024-39614 -- Mitigated potential DoS in get_supported_lang…
Browse files Browse the repository at this point in the history
…uage_variant().

Language codes are now parsed with a maximum length limit of 500 chars.

Thanks to MProgrammer for the report.
  • Loading branch information
sarahboyce authored and nessita committed Jul 9, 2024
1 parent fe4a0bb commit 9e97922
Show file tree
Hide file tree
Showing 5 changed files with 71 additions and 5 deletions.
25 changes: 20 additions & 5 deletions django/utils/translation/trans_real.py
Original file line number Diff line number Diff line change
Expand Up @@ -32,9 +32,10 @@
CONTEXT_SEPARATOR = "\x04"

# Maximum number of characters that will be parsed from the Accept-Language
# header to prevent possible denial of service or memory exhaustion attacks.
# About 10x longer than the longest value shown on MDN’s Accept-Language page.
ACCEPT_LANGUAGE_HEADER_MAX_LENGTH = 500
# header or cookie to prevent possible denial of service or memory exhaustion
# attacks. About 10x longer than the longest value shown on MDN’s
# Accept-Language page.
LANGUAGE_CODE_MAX_LENGTH = 500

# Format of Accept-Language header values. From RFC 9110 Sections 12.4.2 and
# 12.5.4, and RFC 5646 Section 2.1.
Expand Down Expand Up @@ -498,11 +499,25 @@ def get_supported_language_variant(lang_code, strict=False):
If `strict` is False (the default), look for a country-specific variant
when neither the language code nor its generic variant is found.
The language code is truncated to a maximum length to avoid potential
denial of service attacks.
lru_cache should have a maxsize to prevent from memory exhaustion attacks,
as the provided language codes are taken from the HTTP request. See also
<https://www.djangoproject.com/weblog/2007/oct/26/security-fix/>.
"""
if lang_code:
# Truncate the language code to a maximum length to avoid potential
# denial of service attacks.
if len(lang_code) > LANGUAGE_CODE_MAX_LENGTH:
if (
not strict
and (index := lang_code.rfind("-", 0, LANGUAGE_CODE_MAX_LENGTH)) > 0
):
# There is a generic variant under the maximum length accepted length.
lang_code = lang_code[:index]
else:
raise ValueError("'lang_code' exceeds the maximum accepted length")
# If 'zh-hant-tw' is not supported, try special fallback or subsequent
# language codes i.e. 'zh-hant' and 'zh'.
possible_lang_codes = [lang_code]
Expand Down Expand Up @@ -626,13 +641,13 @@ def parse_accept_lang_header(lang_string):
functools.lru_cache() to avoid repetitive parsing of common header values.
"""
# If the header value doesn't exceed the maximum allowed length, parse it.
if len(lang_string) <= ACCEPT_LANGUAGE_HEADER_MAX_LENGTH:
if len(lang_string) <= LANGUAGE_CODE_MAX_LENGTH:
return _parse_accept_lang_header(lang_string)

# If there is at least one comma in the value, parse up to the last comma
# before the max length, skipping any truncated parts at the end of the
# header value.
if (index := lang_string.rfind(",", 0, ACCEPT_LANGUAGE_HEADER_MAX_LENGTH)) > 0:
if (index := lang_string.rfind(",", 0, LANGUAGE_CODE_MAX_LENGTH)) > 0:
return _parse_accept_lang_header(lang_string[:index])

# Don't attempt to parse if there is only one language-range value which is
Expand Down
10 changes: 10 additions & 0 deletions docs/ref/utils.txt
Original file line number Diff line number Diff line change
Expand Up @@ -1147,6 +1147,11 @@ For a complete discussion on the usage of the following see the
``lang_code`` is ``'es-ar'`` and ``'es'`` is in :setting:`LANGUAGES` but
``'es-ar'`` isn't.

``lang_code`` has a maximum accepted length of 500 characters. A
:exc:`ValueError` is raised if ``lang_code`` exceeds this limit and
``strict`` is ``True``, or if there is no generic variant and ``strict``
is ``False``.

If ``strict`` is ``False`` (the default), a country-specific variant may
be returned when neither the language code nor its generic variant is found.
For example, if only ``'es-co'`` is in :setting:`LANGUAGES`, that's
Expand All @@ -1155,6 +1160,11 @@ For a complete discussion on the usage of the following see the

Raises :exc:`LookupError` if nothing is found.

.. versionchanged:: 4.2.14

In older versions, ``lang_code`` values over 500 characters were
processed without raising a :exc:`ValueError`.

.. function:: to_locale(language)

Turns a language name (en-us) into a locale name (en_US).
Expand Down
15 changes: 15 additions & 0 deletions docs/releases/4.2.14.txt
Original file line number Diff line number Diff line change
Expand Up @@ -32,3 +32,18 @@ directory-traversal via certain inputs when calling :meth:`save()
<django.core.files.storage.Storage.save()>`.

Built-in ``Storage`` sub-classes were not affected by this vulnerability.

CVE-2024-39614: Potential denial-of-service vulnerability in ``get_supported_language_variant()``
=================================================================================================

:meth:`~django.utils.translation.get_supported_language_variant` was subject to
a potential denial-of-service attack when used with very long strings
containing specific characters.

To mitigate this vulnerability, the language code provided to
:meth:`~django.utils.translation.get_supported_language_variant` is now parsed
up to a maximum length of 500 characters.

When the language code is over 500 characters, a :exc:`ValueError` will now be
raised if ``strict`` is ``True``, or if there is no generic variant and
``strict`` is ``False``.
15 changes: 15 additions & 0 deletions docs/releases/5.0.7.txt
Original file line number Diff line number Diff line change
Expand Up @@ -33,6 +33,21 @@ directory-traversal via certain inputs when calling :meth:`save()

Built-in ``Storage`` sub-classes were not affected by this vulnerability.

CVE-2024-39614: Potential denial-of-service vulnerability in ``get_supported_language_variant()``
=================================================================================================

:meth:`~django.utils.translation.get_supported_language_variant` was subject to
a potential denial-of-service attack when used with very long strings
containing specific characters.

To mitigate this vulnerability, the language code provided to
:meth:`~django.utils.translation.get_supported_language_variant` is now parsed
up to a maximum length of 500 characters.

When the language code is over 500 characters, a :exc:`ValueError` will now be
raised if ``strict`` is ``True``, or if there is no generic variant and
``strict`` is ``False``.

Bugfixes
========

Expand Down
11 changes: 11 additions & 0 deletions tests/i18n/tests.py
Original file line number Diff line number Diff line change
Expand Up @@ -58,6 +58,7 @@
translation_file_changed,
watch_for_translation_changes,
)
from django.utils.translation.trans_real import LANGUAGE_CODE_MAX_LENGTH

from .forms import CompanyForm, I18nForm, SelectDateForm
from .models import Company, TestModel
Expand Down Expand Up @@ -1672,6 +1673,16 @@ def test_get_supported_language_variant_real(self):
g("xyz")
with self.assertRaises(LookupError):
g("xy-zz")
msg = "'lang_code' exceeds the maximum accepted length"
with self.assertRaises(LookupError):
g("x" * LANGUAGE_CODE_MAX_LENGTH)
with self.assertRaisesMessage(ValueError, msg):
g("x" * (LANGUAGE_CODE_MAX_LENGTH + 1))
# 167 * 3 = 501 which is LANGUAGE_CODE_MAX_LENGTH + 1.
self.assertEqual(g("en-" * 167), "en")
with self.assertRaisesMessage(ValueError, msg):
g("en-" * 167, strict=True)
self.assertEqual(g("en-" * 30000), "en") # catastrophic test

def test_get_supported_language_variant_null(self):
g = trans_null.get_supported_language_variant
Expand Down

0 comments on commit 9e97922

Please sign in to comment.