From c55479556db015f48fc8bbca17f64d3e65598559 Mon Sep 17 00:00:00 2001
From: "Miss Islington (bot)"
 <31488909+miss-islington@users.noreply.github.com>
Date: Wed, 15 Jul 2020 05:30:53 -0700
Subject: [PATCH] [3.8] bpo-39017: Avoid infinite loop in the tarfile module
 (GH-21454) (GH-21483)

Avoid infinite loop when reading specially crafted TAR files using the tarfile module
(CVE-2019-20907).
(cherry picked from commit 5a8d121a1f3ef5ad7c105ee378cc79a3eac0c7d4)


Co-authored-by: Rishi <rishi_devan@mail.com>

Automerge-Triggered-By: @encukou
---
 Lib/tarfile.py                                    |   2 ++
 Lib/test/recursion.tar                            | Bin 0 -> 516 bytes
 Lib/test/test_tarfile.py                          |   7 +++++++
 .../2020-07-12-22-16-58.bpo-39017.x3Cg-9.rst      |   1 +
 4 files changed, 10 insertions(+)
 create mode 100644 Lib/test/recursion.tar
 create mode 100644 Misc/NEWS.d/next/Library/2020-07-12-22-16-58.bpo-39017.x3Cg-9.rst

diff --git a/Lib/tarfile.py b/Lib/tarfile.py
index d31b9cbb51d..7a69e1b1aa5 100755
--- a/Lib/tarfile.py
+++ b/Lib/tarfile.py
@@ -1241,6 +1241,8 @@ class TarInfo(object):
 
             length, keyword = match.groups()
             length = int(length)
+            if length == 0:
+                raise InvalidHeaderError("invalid header")
             value = buf[match.end(2) + 1:match.start(1) + length - 1]
 
             # Normally, we could just use "utf-8" as the encoding and "strict"
diff --git a/Lib/test/recursion.tar b/Lib/test/recursion.tar
new file mode 100644
index 0000000000000000000000000000000000000000..b8237251964983f54ed1966297e887636cd0c5f4
GIT binary patch
literal 516
zcmYdFPRz+kEn=W0Fn}74P8%Xw3X=l~85kIuo0>8xq$A1Gm}!7)KUsFc41m#O8A5+e
I1_}|j06>QaCIA2c

literal 0
HcmV?d00001

diff --git a/Lib/test/test_tarfile.py b/Lib/test/test_tarfile.py
index 15324a4e488..b512168d6ea 100644
--- a/Lib/test/test_tarfile.py
+++ b/Lib/test/test_tarfile.py
@@ -397,6 +397,13 @@ class CommonReadTest(ReadTest):
                 with self.assertRaisesRegex(tarfile.ReadError, "unexpected end of data"):
                     tar.extractfile(t).read()
 
+    def test_length_zero_header(self):
+        # bpo-39017 (CVE-2019-20907): reading a zero-length header should fail
+        # with an exception
+        with self.assertRaisesRegex(tarfile.ReadError, "file could not be opened successfully"):
+            with tarfile.open(support.findfile('recursion.tar')) as tar:
+                pass
+
 class MiscReadTestBase(CommonReadTest):
     def requires_name_attribute(self):
         pass
diff --git a/Misc/NEWS.d/next/Library/2020-07-12-22-16-58.bpo-39017.x3Cg-9.rst b/Misc/NEWS.d/next/Library/2020-07-12-22-16-58.bpo-39017.x3Cg-9.rst
new file mode 100644
index 00000000000..ad26676f8b8
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2020-07-12-22-16-58.bpo-39017.x3Cg-9.rst
@@ -0,0 +1 @@
+Avoid infinite loop when reading specially crafted TAR files using the tarfile module (CVE-2019-20907).