Configurable token hashing algorithm

The user's authentication token was hashed using the MD5 algorithm.
The MD5 algorithm shouldn't be used because of the potential for
hash collisions. Some security standards mandate a SHA2 algorithm
or better must be used.

With this change the algorithm to use for hashing tokens can be
configured by setting the OPENSTACK_TOKEN_HASH_ALGORITHM
configuration option to a hash algorithm supported by Python's
hashlib library[1]. For example, a deployer could set the option to
'sha256' to meet a SHA2 security standard.

The algorithm chosen must match the hash algorithm that the
identity server is configured to use (Keystone and the auth_token
middleware can be configured to use any hash algorithm supported by
hashlib).

This is for security hardening.

[1] https://docs.python.org/2/library/hashlib.html

DocImpact
SecurityImpact

Change-Id: I9e3eba7e0a12ae40a08d0ed851ea916ec6591bcc
Closes-Bug: #1174499

openstack_auth/user.py

@@ -68,8 +68,12 @@ class Token(object):
 
         # Token-related attributes
         self.id = auth_ref.auth_token
-        if len(self.id) > 32:
-            self.id = hashlib.md5(self.id).hexdigest()
+        if len(self.id) > 64:
+            algorithm = getattr(settings, 'OPENSTACK_TOKEN_HASH_ALGORITHM',
+                                'md5')
+            hasher = hashlib.new(algorithm)
+            hasher.update(self.id)
+            self.id = hasher.hexdigest()
         self.expires = auth_ref.expires
 
         # Project-related attributes