Add a safe_minidom_parse_string function.

Adds a new utils.safe_minidom_parse_string function and
updates external API facing Nova modules to use it.
This ensures we have safe defaults on our incoming API XML parsing.

Internally safe_minidom_parse_string uses a ProtectedExpatParser
class to disable DTDs and entities from being parsed when using
minidom.

Fixes LP Bug #1100282.

Change-Id: Ib90d6379320ff1d007f8a661f7ddaa286ba6918e

7 files changed, 22 insertions, 24 deletions

diff --git a/nova/api/openstack/common.py b/nova/api/openstack/common.py
index b2583588d..1f479b82e 100644
--- a/nova/api/openstack/common.py
+++ b/nova/api/openstack/common.py
@@ -21,7 +21,6 @@ import re
 import urlparse
 
 import webob
-from xml.dom import minidom
 
 from nova.api.openstack import wsgi
 from nova.api.openstack import xmlutil
@@ -32,6 +31,7 @@ from nova import exception
 from nova.openstack.common import cfg
 from nova.openstack.common import log as logging
 from nova import quota
+from nova import utils
 
 osapi_opts = [
     cfg.IntOpt('osapi_max_limit',
@@ -356,7 +356,7 @@ def raise_http_conflict_for_instance_invalid_state(exc, action):
 
 class MetadataDeserializer(wsgi.MetadataXMLDeserializer):
     def deserialize(self, text):
-        dom = minidom.parseString(text)
+        dom = utils.safe_minidom_parse_string(text)
         metadata_node = self.find_first_child_named(dom, "metadata")
         metadata = self.extract_metadata(metadata_node)
         return {'body': {'metadata': metadata}}
@@ -364,7 +364,7 @@ class MetadataDeserializer(wsgi.MetadataXMLDeserializer):
 
 class MetaItemDeserializer(wsgi.MetadataXMLDeserializer):
     def deserialize(self, text):
-        dom = minidom.parseString(text)
+        dom = utils.safe_minidom_parse_string(text)
         metadata_item = self.extract_metadata(dom)
         return {'body': {'meta': metadata_item}}
 
@@ -382,7 +382,7 @@ class MetadataXMLDeserializer(wsgi.XMLDeserializer):
         return metadata
 
     def _extract_metadata_container(self, datastring):
-        dom = minidom.parseString(datastring)
+        dom = utils.safe_minidom_parse_string(datastring)
         metadata_node = self.find_first_child_named(dom, "metadata")
         metadata = self.extract_metadata(metadata_node)
         return {'body': {'metadata': metadata}}
@@ -394,7 +394,7 @@ class MetadataXMLDeserializer(wsgi.XMLDeserializer):
         return self._extract_metadata_container(datastring)
 
     def update(self, datastring):
-        dom = minidom.parseString(datastring)
+        dom = utils.safe_minidom_parse_string(datastring)
         metadata_item = self.extract_metadata(dom)
         return {'body': {'meta': metadata_item}}
 
diff --git a/nova/api/openstack/compute/contrib/cells.py b/nova/api/openstack/compute/contrib/cells.py
index 03e2e4ca2..29a9ffcc4 100644
--- a/nova/api/openstack/compute/contrib/cells.py
+++ b/nova/api/openstack/compute/contrib/cells.py
@@ -16,7 +16,6 @@
 #    under the License.
 
 """The cells extension."""
-from xml.dom import minidom
 from xml.parsers import expat
 
 from webob import exc
@@ -32,6 +31,7 @@ from nova import exception
 from nova.openstack.common import cfg
 from nova.openstack.common import log as logging
 from nova.openstack.common import timeutils
+from nova import utils
 
 
 LOG = logging.getLogger(__name__)
@@ -99,7 +99,7 @@ class CellDeserializer(wsgi.XMLDeserializer):
     def default(self, string):
         """Deserialize an xml-formatted cell create request."""
         try:
-            node = minidom.parseString(string)
+            node = utils.safe_minidom_parse_string(string)
         except expat.ExpatError:
             msg = _("cannot understand XML")
             raise exception.MalformedRequestBody(reason=msg)
diff --git a/nova/api/openstack/compute/contrib/hosts.py b/nova/api/openstack/compute/contrib/hosts.py
index 9ce278900..a3cfd229c 100644
--- a/nova/api/openstack/compute/contrib/hosts.py
+++ b/nova/api/openstack/compute/contrib/hosts.py
@@ -16,7 +16,6 @@
 """The hosts admin extension."""
 
 import webob.exc
-from xml.dom import minidom
 from xml.parsers import expat
 
 from nova.api.openstack import extensions
@@ -25,6 +24,7 @@ from nova.api.openstack import xmlutil
 from nova import compute
 from nova import exception
 from nova.openstack.common import log as logging
+from nova import utils
 
 LOG = logging.getLogger(__name__)
 authorize = extensions.extension_authorizer('compute', 'hosts')
@@ -72,7 +72,7 @@ class HostShowTemplate(xmlutil.TemplateBuilder):
 class HostUpdateDeserializer(wsgi.XMLDeserializer):
     def default(self, string):
         try:
-            node = minidom.parseString(string)
+            node = utils.safe_minidom_parse_string(string)
         except expat.ExpatError:
             msg = _("cannot understand XML")
             raise exception.MalformedRequestBody(reason=msg)
diff --git a/nova/api/openstack/compute/contrib/security_groups.py b/nova/api/openstack/compute/contrib/security_groups.py
index c49e7af70..d42dc1b0a 100644
--- a/nova/api/openstack/compute/contrib/security_groups.py
+++ b/nova/api/openstack/compute/contrib/security_groups.py
@@ -16,8 +16,6 @@
 
 """The security groups extension."""
 
-from xml.dom import minidom
-
 import webob
 from webob import exc
 
@@ -30,6 +28,7 @@ from nova.compute import api as compute_api
 from nova import db
 from nova import exception
 from nova.openstack.common import log as logging
+from nova import utils
 from nova.virt import netutils
 
 LOG = logging.getLogger(__name__)
@@ -109,7 +108,7 @@ class SecurityGroupXMLDeserializer(wsgi.MetadataXMLDeserializer):
     """
     def default(self, string):
         """Deserialize an xml-formatted security group create request."""
-        dom = minidom.parseString(string)
+        dom = utils.safe_minidom_parse_string(string)
         security_group = {}
         sg_node = self.find_first_child_named(dom,
                                                'security_group')
@@ -130,7 +129,7 @@ class SecurityGroupRulesXMLDeserializer(wsgi.MetadataXMLDeserializer):
 
     def default(self, string):
         """Deserialize an xml-formatted security group create request."""
-        dom = minidom.parseString(string)
+        dom = utils.safe_minidom_parse_string(string)
         security_group_rule = self._extract_security_group_rule(dom)
         return {'body': {'security_group_rule': security_group_rule}}
 
diff --git a/nova/api/openstack/compute/contrib/volumes.py b/nova/api/openstack/compute/contrib/volumes.py
index 0ab93db35..760dc953a 100644
--- a/nova/api/openstack/compute/contrib/volumes.py
+++ b/nova/api/openstack/compute/contrib/volumes.py
@@ -17,7 +17,6 @@
 
 import webob
 from webob import exc
-from xml.dom import minidom
 
 from nova.api.openstack import common
 from nova.api.openstack import extensions
@@ -155,7 +154,7 @@ class CreateDeserializer(CommonDeserializer):
 
     def default(self, string):
         """Deserialize an xml-formatted volume create request."""
-        dom = minidom.parseString(string)
+        dom = utils.safe_minidom_parse_string(string)
         vol = self._extract_volume(dom)
         return {'body': {'volume': vol}}
 
diff --git a/nova/api/openstack/compute/servers.py b/nova/api/openstack/compute/servers.py
index 729754ea6..3b8eb506d 100644
--- a/nova/api/openstack/compute/servers.py
+++ b/nova/api/openstack/compute/servers.py
@@ -21,7 +21,6 @@ import socket
 
 import webob
 from webob import exc
-from xml.dom import minidom
 
 from nova.api.openstack import common
 from nova.api.openstack.compute import ips
@@ -312,7 +311,7 @@ class ActionDeserializer(CommonDeserializer):
     """
 
     def default(self, string):
-        dom = minidom.parseString(string)
+        dom = utils.safe_minidom_parse_string(string)
         action_node = dom.childNodes[0]
         action_name = action_node.tagName
 
@@ -419,7 +418,7 @@ class CreateDeserializer(CommonDeserializer):
 
     def default(self, string):
         """Deserialize an xml-formatted server create request."""
-        dom = minidom.parseString(string)
+        dom = utils.safe_minidom_parse_string(string)
         server = self._extract_server(dom)
         return {'body': {'server': server}}
 
diff --git a/nova/api/openstack/wsgi.py b/nova/api/openstack/wsgi.py
index f68eff2a7..8c77f1c9c 100644
--- a/nova/api/openstack/wsgi.py
+++ b/nova/api/openstack/wsgi.py
@@ -27,6 +27,7 @@ import webob
 from nova import exception
 from nova.openstack.common import jsonutils
 from nova.openstack.common import log as logging
+from nova import utils
 from nova import wsgi
 
 
@@ -217,7 +218,7 @@ class XMLDeserializer(TextDeserializer):
         plurals = set(self.metadata.get('plurals', {}))
 
         try:
-            node = minidom.parseString(datastring).childNodes[0]
+            node = utils.safe_minidom_parse_string(datastring).childNodes[0]
             return {node.nodeName: self._from_xml_node(node, plurals)}
         except expat.ExpatError:
             msg = _("cannot understand XML")
@@ -268,11 +269,11 @@ class XMLDeserializer(TextDeserializer):
 
     def extract_text(self, node):
         """Get the text field contained by the given node."""
-        if len(node.childNodes) == 1:
-            child = node.childNodes[0]
+        ret_val = ""
+        for child in node.childNodes:
             if child.nodeType == child.TEXT_NODE:
-                return child.nodeValue
-        return ""
+                ret_val += child.nodeValue
+        return ret_val
 
     def extract_elements(self, node):
         """Get only Element type childs from node."""
@@ -633,7 +634,7 @@ def action_peek_json(body):
 def action_peek_xml(body):
     """Determine action to invoke."""
 
-    dom = minidom.parseString(body)
+    dom = utils.safe_minidom_parse_string(body)
     action_node = dom.childNodes[0]
 
     return action_node.tagName