add a default limit for the amount of data xmlrpclib.gzip_decode will return (closes #16043)

3 files changed, 32 insertions, 3 deletions

diff --git a/Lib/test/test_xmlrpc.py b/Lib/test/test_xmlrpc.py
index 09235fd9c9..2bb3978b84 100644
--- a/Lib/test/test_xmlrpc.py
+++ b/Lib/test/test_xmlrpc.py
@@ -737,7 +737,7 @@ class GzipServerTestCase(BaseServerTestCase):
         with cm:
             p.pow(6, 8)
 
-    def test_gsip_response(self):
+    def test_gzip_response(self):
         t = self.Transport()
         p = xmlrpclib.ServerProxy(URL, transport=t)
         old = self.requestHandler.encode_threshold
@@ -750,6 +750,23 @@ class GzipServerTestCase(BaseServerTestCase):
         self.requestHandler.encode_threshold = old
         self.assertTrue(a>b)
 
+    def test_gzip_decode_limit(self):
+        max_gzip_decode = 20 * 1024 * 1024
+        data = '\0' * max_gzip_decode
+        encoded = xmlrpclib.gzip_encode(data)
+        decoded = xmlrpclib.gzip_decode(encoded)
+        self.assertEqual(len(decoded), max_gzip_decode)
+
+        data = '\0' * (max_gzip_decode + 1)
+        encoded = xmlrpclib.gzip_encode(data)
+
+        with self.assertRaisesRegexp(ValueError,
+                                     "max gzipped payload length exceeded"):
+            xmlrpclib.gzip_decode(encoded)
+
+        xmlrpclib.gzip_decode(encoded, max_decode=-1)
+
+
 #Test special attributes of the ServerProxy object
 class ServerProxyTestCase(unittest.TestCase):
     def setUp(self):
diff --git a/Lib/xmlrpclib.py b/Lib/xmlrpclib.py
index 5354ec246a..340dc51eb6 100644
--- a/Lib/xmlrpclib.py
+++ b/Lib/xmlrpclib.py
@@ -49,6 +49,7 @@
 # 2003-07-12 gp  Correct marshalling of Faults
 # 2003-10-31 mvl Add multicall support
 # 2004-08-20 mvl Bump minimum supported Python version to 2.1
+# 2014-12-02 ch/doko  Add workaround for gzip bomb vulnerability
 #
 # Copyright (c) 1999-2002 by Secret Labs AB.
 # Copyright (c) 1999-2002 by Fredrik Lundh.
@@ -1165,10 +1166,13 @@ def gzip_encode(data):
 # in the HTTP header, as described in RFC 1952
 #
 # @param data The encoded data
+# @keyparam max_decode Maximum bytes to decode (20MB default), use negative
+#    values for unlimited decoding
 # @return the unencoded data
 # @raises ValueError if data is not correctly coded.
+# @raises ValueError if max gzipped payload length exceeded
 
-def gzip_decode(data):
+def gzip_decode(data, max_decode=20971520):
     """gzip encoded data -> unencoded data
 
     Decode data using the gzip content encoding as described in RFC 1952
@@ -1178,11 +1182,16 @@ def gzip_decode(data):
     f = StringIO.StringIO(data)
     gzf = gzip.GzipFile(mode="rb", fileobj=f)
     try:
-        decoded = gzf.read()
+        if max_decode < 0: # no limit
+            decoded = gzf.read()
+        else:
+            decoded = gzf.read(max_decode + 1)
     except IOError:
         raise ValueError("invalid data")
     f.close()
     gzf.close()
+    if max_decode >= 0 and len(decoded) > max_decode:
+        raise ValueError("max gzipped payload length exceeded")
     return decoded
 
 ##
diff --git a/Misc/NEWS b/Misc/NEWS
index ff68b451f4..dea29e3a97 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -10,6 +10,9 @@ What's New in Python 2.7.9?
 Library
 -------
 
+- Issue #16043: Add a default limit for the amount of data xmlrpclib.gzip_decode
+  will return. This resolves CVE-2013-1753.
+
 - Issue #16042: CVE-2013-1752: smtplib: Limit amount of data read by limiting
   the call to readline().  Original patch by Christian Heimes.