summaryrefslogtreecommitdiffstats
path: root/meta/recipes-devtools/python/python3/CVE-2024-6232.patch
diff options
context:
space:
mode:
authorDivya Chellam <divya.chellam@windriver.com>2024-09-18 06:40:30 +0000
committerSteve Sakoman <steve@sakoman.com>2024-09-27 05:57:35 -0700
commit487e8cdf1df6feba6d88fa29e11791f4ebaaa362 (patch)
tree1840e3774e8af68dbe0248a463055af51fd25dd7 /meta/recipes-devtools/python/python3/CVE-2024-6232.patch
parent8190d9c754c9c3a1962123e1e86d99de96c1224c (diff)
downloadpoky-487e8cdf1df6feba6d88fa29e11791f4ebaaa362.tar.gz
python3: Upgrade 3.10.14 -> 3.10.15
Includes security fixes for CVE-2024-4030, CVE-2024-7592, CVE-2024-4032, CVE-2024-8088 CVE-2024-6232, CVE-2024-6923, CVE-2023-27043 and other bug fixes. Removed below patches, as the fixes included in 3.10.15 upgrade: 1. CVE-2023-27043.patch 2. CVE-2024-6232.patch 3. CVE-2024-7592.patch 4. CVE-2024-8088.patch Release Notes: https://www.python.org/downloads/release/python-31015/ (From OE-Core rev: e64ead97ae3d680f97bf85422f777cd77ae7c434) Signed-off-by: Divya Chellam <divya.chellam@windriver.com> Signed-off-by: Steve Sakoman <steve@sakoman.com>
Diffstat (limited to 'meta/recipes-devtools/python/python3/CVE-2024-6232.patch')
-rw-r--r--meta/recipes-devtools/python/python3/CVE-2024-6232.patch251
1 files changed, 0 insertions, 251 deletions
diff --git a/meta/recipes-devtools/python/python3/CVE-2024-6232.patch b/meta/recipes-devtools/python/python3/CVE-2024-6232.patch
deleted file mode 100644
index 874cbfe40c..0000000000
--- a/meta/recipes-devtools/python/python3/CVE-2024-6232.patch
+++ /dev/null
@@ -1,251 +0,0 @@
1From 3a22dc1079be5a75750d24dc6992956e7b84b5a0 Mon Sep 17 00:00:00 2001
2From: Seth Michael Larson <seth@python.org>
3Date: Tue, 3 Sep 2024 10:07:53 -0500
4Subject: [PATCH 2/2] [3.10] gh-121285: Remove backtracking when parsing
5 tarfile headers (GH-121286) (#123640)
6
7* Remove backtracking when parsing tarfile headers
8* Rewrite PAX header parsing to be stricter
9* Optimize parsing of GNU extended sparse headers v0.0
10
11(cherry picked from commit 34ddb64d088dd7ccc321f6103d23153256caa5d4)
12
13Upstream-Status: Backport from https://github.com/python/cpython/commit/743acbe872485dc18df4d8ab2dc7895187f062c4
14CVE: CVE-2024-6232
15
16Co-authored-by: Kirill Podoprigora <kirill.bast9@mail.ru>
17Co-authored-by: Gregory P. Smith <greg@krypto.org>
18Signed-off-by: Hugo SIMELIERE <hsimeliere.opensource@witekio.com>
19---
20 Lib/tarfile.py | 105 +++++++++++-------
21 Lib/test/test_tarfile.py | 42 +++++++
22 ...-07-02-13-39-20.gh-issue-121285.hrl-yI.rst | 2 +
23 3 files changed, 111 insertions(+), 38 deletions(-)
24 create mode 100644 Misc/NEWS.d/next/Security/2024-07-02-13-39-20.gh-issue-121285.hrl-yI.rst
25
26diff --git a/Lib/tarfile.py b/Lib/tarfile.py
27index 495349f08f9..3ab6811d633 100755
28--- a/Lib/tarfile.py
29+++ b/Lib/tarfile.py
30@@ -841,6 +841,9 @@ def data_filter(member, dest_path):
31 # Sentinel for replace() defaults, meaning "don't change the attribute"
32 _KEEP = object()
33
34+# Header length is digits followed by a space.
35+_header_length_prefix_re = re.compile(br"([0-9]{1,20}) ")
36+
37 class TarInfo(object):
38 """Informational class which holds the details about an
39 archive member given by a tar header block.
40@@ -1410,41 +1413,59 @@ def _proc_pax(self, tarfile):
41 else:
42 pax_headers = tarfile.pax_headers.copy()
43
44- # Check if the pax header contains a hdrcharset field. This tells us
45- # the encoding of the path, linkpath, uname and gname fields. Normally,
46- # these fields are UTF-8 encoded but since POSIX.1-2008 tar
47- # implementations are allowed to store them as raw binary strings if
48- # the translation to UTF-8 fails.
49- match = re.search(br"\d+ hdrcharset=([^\n]+)\n", buf)
50- if match is not None:
51- pax_headers["hdrcharset"] = match.group(1).decode("utf-8")
52-
53- # For the time being, we don't care about anything other than "BINARY".
54- # The only other value that is currently allowed by the standard is
55- # "ISO-IR 10646 2000 UTF-8" in other words UTF-8.
56- hdrcharset = pax_headers.get("hdrcharset")
57- if hdrcharset == "BINARY":
58- encoding = tarfile.encoding
59- else:
60- encoding = "utf-8"
61-
62 # Parse pax header information. A record looks like that:
63 # "%d %s=%s\n" % (length, keyword, value). length is the size
64 # of the complete record including the length field itself and
65- # the newline. keyword and value are both UTF-8 encoded strings.
66- regex = re.compile(br"(\d+) ([^=]+)=")
67+ # the newline.
68 pos = 0
69- while True:
70- match = regex.match(buf, pos)
71- if not match:
72- break
73+ encoding = None
74+ raw_headers = []
75+ while len(buf) > pos and buf[pos] != 0x00:
76+ if not (match := _header_length_prefix_re.match(buf, pos)):
77+ raise InvalidHeaderError("invalid header")
78+ try:
79+ length = int(match.group(1))
80+ except ValueError:
81+ raise InvalidHeaderError("invalid header")
82+ # Headers must be at least 5 bytes, shortest being '5 x=\n'.
83+ # Value is allowed to be empty.
84+ if length < 5:
85+ raise InvalidHeaderError("invalid header")
86+ if pos + length > len(buf):
87+ raise InvalidHeaderError("invalid header")
88
89- length, keyword = match.groups()
90- length = int(length)
91- if length == 0:
92+ header_value_end_offset = match.start(1) + length - 1 # Last byte of the header
93+ keyword_and_value = buf[match.end(1) + 1:header_value_end_offset]
94+ raw_keyword, equals, raw_value = keyword_and_value.partition(b"=")
95+
96+ # Check the framing of the header. The last character must be '\n' (0x0A)
97+ if not raw_keyword or equals != b"=" or buf[header_value_end_offset] != 0x0A:
98 raise InvalidHeaderError("invalid header")
99- value = buf[match.end(2) + 1:match.start(1) + length - 1]
100+ raw_headers.append((length, raw_keyword, raw_value))
101+
102+ # Check if the pax header contains a hdrcharset field. This tells us
103+ # the encoding of the path, linkpath, uname and gname fields. Normally,
104+ # these fields are UTF-8 encoded but since POSIX.1-2008 tar
105+ # implementations are allowed to store them as raw binary strings if
106+ # the translation to UTF-8 fails. For the time being, we don't care about
107+ # anything other than "BINARY". The only other value that is currently
108+ # allowed by the standard is "ISO-IR 10646 2000 UTF-8" in other words UTF-8.
109+ # Note that we only follow the initial 'hdrcharset' setting to preserve
110+ # the initial behavior of the 'tarfile' module.
111+ if raw_keyword == b"hdrcharset" and encoding is None:
112+ if raw_value == b"BINARY":
113+ encoding = tarfile.encoding
114+ else: # This branch ensures only the first 'hdrcharset' header is used.
115+ encoding = "utf-8"
116+
117+ pos += length
118
119+ # If no explicit hdrcharset is set, we use UTF-8 as a default.
120+ if encoding is None:
121+ encoding = "utf-8"
122+
123+ # After parsing the raw headers we can decode them to text.
124+ for length, raw_keyword, raw_value in raw_headers:
125 # Normally, we could just use "utf-8" as the encoding and "strict"
126 # as the error handler, but we better not take the risk. For
127 # example, GNU tar <= 1.23 is known to store filenames it cannot
128@@ -1452,17 +1473,16 @@ def _proc_pax(self, tarfile):
129 # hdrcharset=BINARY header).
130 # We first try the strict standard encoding, and if that fails we
131 # fall back on the user's encoding and error handler.
132- keyword = self._decode_pax_field(keyword, "utf-8", "utf-8",
133+ keyword = self._decode_pax_field(raw_keyword, "utf-8", "utf-8",
134 tarfile.errors)
135 if keyword in PAX_NAME_FIELDS:
136- value = self._decode_pax_field(value, encoding, tarfile.encoding,
137+ value = self._decode_pax_field(raw_value, encoding, tarfile.encoding,
138 tarfile.errors)
139 else:
140- value = self._decode_pax_field(value, "utf-8", "utf-8",
141+ value = self._decode_pax_field(raw_value, "utf-8", "utf-8",
142 tarfile.errors)
143
144 pax_headers[keyword] = value
145- pos += length
146
147 # Fetch the next header.
148 try:
149@@ -1477,7 +1497,7 @@ def _proc_pax(self, tarfile):
150
151 elif "GNU.sparse.size" in pax_headers:
152 # GNU extended sparse format version 0.0.
153- self._proc_gnusparse_00(next, pax_headers, buf)
154+ self._proc_gnusparse_00(next, raw_headers)
155
156 elif pax_headers.get("GNU.sparse.major") == "1" and pax_headers.get("GNU.sparse.minor") == "0":
157 # GNU extended sparse format version 1.0.
158@@ -1499,15 +1519,24 @@ def _proc_pax(self, tarfile):
159
160 return next
161
162- def _proc_gnusparse_00(self, next, pax_headers, buf):
163+ def _proc_gnusparse_00(self, next, raw_headers):
164 """Process a GNU tar extended sparse header, version 0.0.
165 """
166 offsets = []
167- for match in re.finditer(br"\d+ GNU.sparse.offset=(\d+)\n", buf):
168- offsets.append(int(match.group(1)))
169 numbytes = []
170- for match in re.finditer(br"\d+ GNU.sparse.numbytes=(\d+)\n", buf):
171- numbytes.append(int(match.group(1)))
172+ for _, keyword, value in raw_headers:
173+ if keyword == b"GNU.sparse.offset":
174+ try:
175+ offsets.append(int(value.decode()))
176+ except ValueError:
177+ raise InvalidHeaderError("invalid header")
178+
179+ elif keyword == b"GNU.sparse.numbytes":
180+ try:
181+ numbytes.append(int(value.decode()))
182+ except ValueError:
183+ raise InvalidHeaderError("invalid header")
184+
185 next.sparse = list(zip(offsets, numbytes))
186
187 def _proc_gnusparse_01(self, next, pax_headers):
188diff --git a/Lib/test/test_tarfile.py b/Lib/test/test_tarfile.py
189index cfc13bccb20..007c3e94acb 100644
190--- a/Lib/test/test_tarfile.py
191+++ b/Lib/test/test_tarfile.py
192@@ -1139,6 +1139,48 @@ def test_pax_number_fields(self):
193 finally:
194 tar.close()
195
196+ def test_pax_header_bad_formats(self):
197+ # The fields from the pax header have priority over the
198+ # TarInfo.
199+ pax_header_replacements = (
200+ b" foo=bar\n",
201+ b"0 \n",
202+ b"1 \n",
203+ b"2 \n",
204+ b"3 =\n",
205+ b"4 =a\n",
206+ b"1000000 foo=bar\n",
207+ b"0 foo=bar\n",
208+ b"-12 foo=bar\n",
209+ b"000000000000000000000000036 foo=bar\n",
210+ )
211+ pax_headers = {"foo": "bar"}
212+
213+ for replacement in pax_header_replacements:
214+ with self.subTest(header=replacement):
215+ tar = tarfile.open(tmpname, "w", format=tarfile.PAX_FORMAT,
216+ encoding="iso8859-1")
217+ try:
218+ t = tarfile.TarInfo()
219+ t.name = "pax" # non-ASCII
220+ t.uid = 1
221+ t.pax_headers = pax_headers
222+ tar.addfile(t)
223+ finally:
224+ tar.close()
225+
226+ with open(tmpname, "rb") as f:
227+ data = f.read()
228+ self.assertIn(b"11 foo=bar\n", data)
229+ data = data.replace(b"11 foo=bar\n", replacement)
230+
231+ with open(tmpname, "wb") as f:
232+ f.truncate()
233+ f.write(data)
234+
235+ with self.assertRaisesRegex(tarfile.ReadError, r"method tar: ReadError\('invalid header'\)"):
236+ tarfile.open(tmpname, encoding="iso8859-1")
237+
238
239 class WriteTestBase(TarTest):
240 # Put all write tests in here that are supposed to be tested
241diff --git a/Misc/NEWS.d/next/Security/2024-07-02-13-39-20.gh-issue-121285.hrl-yI.rst b/Misc/NEWS.d/next/Security/2024-07-02-13-39-20.gh-issue-121285.hrl-yI.rst
242new file mode 100644
243index 00000000000..81f918bfe2b
244--- /dev/null
245+++ b/Misc/NEWS.d/next/Security/2024-07-02-13-39-20.gh-issue-121285.hrl-yI.rst
246@@ -0,0 +1,2 @@
247+Remove backtracking from tarfile header parsing for ``hdrcharset``, PAX, and
248+GNU sparse headers.
249--
2502.46.0
251