freeipa/ipaserver/rpcserver.py

750 lines
25 KiB
Python
Raw Normal View History

# Authors:
# Jason Gerard DeRose <jderose@redhat.com>
#
# Copyright (C) 2008 Red Hat
# see file 'COPYING' for use and warranty information
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
"""
RPC server.
Also see the `ipalib.rpc` module.
"""
from cgi import parse_qs
2010-02-23 11:53:47 -06:00
from xml.sax.saxutils import escape
from xmlrpclib import Fault
from ipalib.backend import Executioner
from ipalib.errors import PublicError, InternalError, CommandError, JSONError, ConversionError, CCacheError, RefererError
2009-10-13 12:28:00 -05:00
from ipalib.request import context, Connection, destroy_context
from ipalib.rpc import xml_dumps, xml_loads
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
from ipalib.util import make_repr, parse_time_duration
from ipapython.compat import json
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
from ipalib.session import session_mgr, AuthManager, read_krbccache_file, store_krbccache_file, delete_krbccache_file, fmt_time, default_max_session_lifetime
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
from ipalib.backend import Backend
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
from ipalib.krb_utils import krb5_parse_ccache, KRB5_CCache, krb_ticket_expiration_threshold
2010-02-23 11:53:47 -06:00
from wsgiref.util import shift_path_info
from ipapython.version import VERSION
import base64
import os
import string
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
import datetime
from decimal import Decimal
2010-02-23 11:53:47 -06:00
_not_found_template = """<html>
<head>
<title>404 Not Found</title>
</head>
<body>
<h1>Not Found</h1>
<p>
The requested URL <strong>%(url)s</strong> was not found on this server.
</p>
</body>
</html>"""
def not_found(environ, start_response):
"""
Return a 404 Not Found error.
"""
status = '404 Not Found'
response_headers = [('Content-Type', 'text/html')]
start_response(status, response_headers)
output = _not_found_template % dict(
url=escape(environ['SCRIPT_NAME'] + environ['PATH_INFO'])
)
return [output]
2009-10-13 12:28:00 -05:00
def read_input(environ):
"""
Read the request body from environ['wsgi.input'].
"""
try:
length = int(environ.get('CONTENT_LENGTH'))
except (ValueError, TypeError):
return
return environ['wsgi.input'].read(length)
def params_2_args_options(params):
if len(params) == 0:
return (tuple(), dict())
2010-03-26 04:56:53 -05:00
if len(params) == 1:
return (params[0], dict())
return (params[0], params[1])
2009-10-13 12:28:00 -05:00
def nicify_query(query, encoding='utf-8'):
if not query:
return
for (key, value) in query.iteritems():
if len(value) == 0:
yield (key, None)
elif len(value) == 1:
yield (key, value[0].decode(encoding))
else:
yield (key, tuple(v.decode(encoding) for v in value))
def extract_query(environ):
"""
Return the query as a ``dict``, or ``None`` if no query is presest.
"""
qstr = None
if environ['REQUEST_METHOD'] == 'POST':
if environ['CONTENT_TYPE'] == 'application/x-www-form-urlencoded':
qstr = read_input(environ)
elif environ['REQUEST_METHOD'] == 'GET':
qstr = environ['QUERY_STRING']
if qstr:
query = dict(nicify_query(
parse_qs(qstr)#, keep_blank_values=True)
2009-10-13 12:28:00 -05:00
))
else:
query = {}
environ['wsgi.query'] = query
return query
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
class wsgi_dispatch(Executioner):
2010-02-23 11:53:47 -06:00
"""
WSGI routing middleware and entry point into IPA server.
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
The `wsgi_dispatch` plugin is the entry point into the IPA server.
It dispatchs the request to the appropriate wsgi application
handler which is specific to the authentication and RPC mechanism.
2010-02-23 11:53:47 -06:00
"""
def __init__(self):
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
super(wsgi_dispatch, self).__init__()
2010-02-23 11:53:47 -06:00
self.__apps = {}
def __iter__(self):
for key in sorted(self.__apps):
yield key
def __getitem__(self, key):
return self.__apps[key]
def __contains__(self, key):
return key in self.__apps
def __call__(self, environ, start_response):
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
self.debug('WSGI wsgi_dispatch.__call__:')
2010-02-23 11:53:47 -06:00
try:
return self.route(environ, start_response)
finally:
destroy_context()
def _on_finalize(self):
2010-02-23 11:53:47 -06:00
self.url = self.env['mount_ipa']
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
super(wsgi_dispatch, self)._on_finalize()
2010-02-23 11:53:47 -06:00
def route(self, environ, start_response):
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
key = environ.get('PATH_INFO')
2010-02-23 11:53:47 -06:00
if key in self.__apps:
app = self.__apps[key]
return app(environ, start_response)
return not_found(environ, start_response)
def mount(self, app, key):
"""
Mount the WSGI application *app* at *key*.
"""
# if self.__islocked__():
# raise StandardError('%s.mount(): locked, cannot mount %r at %r' % (
# self.name, app, key)
# )
if key in self.__apps:
raise StandardError('%s.mount(): cannot replace %r with %r at %r' % (
self.name, self.__apps[key], app, key)
)
self.debug('Mounting %r at %r', app, key)
2010-02-23 11:53:47 -06:00
self.__apps[key] = app
2009-10-13 12:28:00 -05:00
class WSGIExecutioner(Executioner):
"""
Base class for execution backends with a WSGI application interface.
"""
2011-04-21 03:13:06 -05:00
content_type = None
2010-02-23 11:53:47 -06:00
key = ''
def set_api(self, api):
super(WSGIExecutioner, self).set_api(api)
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
if 'wsgi_dispatch' in self.api.Backend:
self.api.Backend.wsgi_dispatch.mount(self, self.key)
2010-02-23 11:53:47 -06:00
def _on_finalize(self):
2010-02-23 11:53:47 -06:00
self.url = self.env.mount_ipa + self.key
super(WSGIExecutioner, self)._on_finalize()
2009-10-13 12:28:00 -05:00
def wsgi_execute(self, environ):
2009-10-13 12:28:00 -05:00
result = None
error = None
_id = None
lang = os.environ['LANG']
name = None
args = ()
options = {}
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
if not 'HTTP_REFERER' in environ:
return self.marshal(result, RefererError(referer='missing'), _id)
if not environ['HTTP_REFERER'].startswith('https://%s/ipa' % self.api.env.host) and not self.env.in_tree:
return self.marshal(result, RefererError(referer=environ['HTTP_REFERER']), _id)
2009-10-13 12:28:00 -05:00
try:
if ('HTTP_ACCEPT_LANGUAGE' in environ):
lang_reg_w_q = environ['HTTP_ACCEPT_LANGUAGE'].split(',')[0]
lang_reg = lang_reg_w_q.split(';')[0]
lang_ = lang_reg.split('-')[0]
if '-' in lang_reg:
reg = lang_reg.split('-')[1].upper();
else:
reg = lang_.upper()
os.environ['LANG'] = '%s_%s' % (lang_, reg)
2010-02-23 11:53:47 -06:00
if (
environ.get('CONTENT_TYPE', '').startswith(self.content_type)
and environ['REQUEST_METHOD'] == 'POST'
):
data = read_input(environ)
(name, args, options, _id) = self.unmarshal(data)
else:
(name, args, options, _id) = self.simple_unmarshal(environ)
if name not in self.Command:
raise CommandError(name=name)
result = self.Command[name](*args, **options)
except PublicError, e:
error = e
except StandardError, e:
self.exception(
'non-public: %s: %s', e.__class__.__name__, str(e)
)
error = InternalError()
finally:
os.environ['LANG'] = lang
if name:
try:
params = self.Command[name].args_options_2_params(*args, **options)
except Exception, e:
self.info(
'exception %s caught when converting options: %s', e.__class__.__name__, str(e)
)
# get at least some context of what is going on
params = options
if error:
self.info('%s: %s(%s): %s', context.principal, name, ', '.join(self.Command[name]._repr_iter(**params)), e.__class__.__name__)
else:
self.info('%s: %s(%s): SUCCESS', context.principal, name, ', '.join(self.Command[name]._repr_iter(**params)))
else:
self.info('%s: %s', context.principal, e.__class__.__name__)
2009-10-13 12:28:00 -05:00
return self.marshal(result, error, _id)
def simple_unmarshal(self, environ):
name = environ['PATH_INFO'].strip('/')
options = extract_query(environ)
return (name, tuple(), options, None)
def __call__(self, environ, start_response):
"""
WSGI application for execution.
"""
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
self.debug('WSGI WSGIExecutioner.__call__:')
2009-10-13 12:28:00 -05:00
try:
status = '200 OK'
response = self.wsgi_execute(environ)
2009-10-13 12:28:00 -05:00
headers = [('Content-Type', self.content_type + '; charset=utf-8')]
except StandardError, e:
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
self.exception('WSGI %s.__call__():', self.name)
2009-10-13 12:28:00 -05:00
status = '500 Internal Server Error'
response = status
headers = [('Content-Type', 'text/plain')]
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
session_data = getattr(context, 'session_data', None)
if session_data is not None:
# Send session cookie back and store session data
# FIXME: the URL path should be retreived from somewhere (but where?), not hardcoded
session_cookie = session_mgr.generate_cookie('/ipa', session_data['session_id'])
headers.append(('Set-Cookie', session_cookie))
2009-10-13 12:28:00 -05:00
start_response(status, headers)
return [response]
def unmarshal(self, data):
raise NotImplementedError('%s.unmarshal()' % self.fullname)
def marshal(self, result, error, _id=None):
raise NotImplementedError('%s.marshal()' % self.fullname)
class xmlserver(WSGIExecutioner):
"""
Execution backend plugin for XML-RPC server.
Also see the `ipalib.rpc.xmlclient` plugin.
"""
2009-10-13 12:28:00 -05:00
content_type = 'text/xml'
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
key = '/xml'
2009-10-13 12:28:00 -05:00
def _on_finalize(self):
self.__system = {
'system.listMethods': self.listMethods,
'system.methodSignature': self.methodSignature,
'system.methodHelp': self.methodHelp,
}
super(xmlserver, self)._on_finalize()
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
def __call__(self, environ, start_response):
'''
'''
self.debug('WSGI xmlserver.__call__:')
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
ccache=environ.get('KRB5CCNAME')
if ccache is None:
return self.marshal(None, CCacheError())
self.create_context(ccache=ccache)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
try:
self.create_context(ccache=environ.get('KRB5CCNAME'))
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
response = super(xmlserver, self).__call__(environ, start_response)
except PublicError, e:
status = '200 OK'
response = status
headers = [('Content-Type', 'text/plain')]
start_response(status, headers)
return self.marshal(None, e)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
finally:
destroy_context()
return response
def listMethods(self, *params):
return tuple(name.decode('UTF-8') for name in self.Command)
def methodSignature(self, *params):
return u'methodSignature not implemented'
def methodHelp(self, *params):
return u'methodHelp not implemented'
2009-10-13 12:28:00 -05:00
def unmarshal(self, data):
(params, name) = xml_loads(data)
(args, options) = params_2_args_options(params)
return (name, args, options, None)
def marshal(self, result, error, _id=None):
if error:
self.debug('response: %s: %s', error.__class__.__name__, str(error))
2009-10-13 12:28:00 -05:00
response = Fault(error.errno, error.strerror)
else:
if isinstance(result, dict):
self.debug('response: entries returned %d', result.get('count', 1))
2009-10-13 12:28:00 -05:00
response = (result,)
return xml_dumps(response, methodresponse=True)
def json_encode_binary(val):
'''
JSON cannot encode binary values. We encode binary values in Python str
objects and text in Python unicode objects. In order to allow a binary
object to be passed through JSON we base64 encode it thus converting it to
text which JSON can transport. To assure we recognize the value is a base64
encoded representation of the original binary value and not confuse it with
other text we convert the binary value to a dict in this form:
{'__base64__' : base64_encoding_of_binary_value}
This modification of the original input value cannot be done "in place" as
one might first assume (e.g. replacing any binary items in a container
(e.g. list, tuple, dict) with the base64 dict because the container might be
an immutable object (i.e. a tuple). Therefore this function returns a copy
of any container objects it encounters with tuples replaced by lists. This
is O.K. because the JSON encoding will map both lists and tuples to JSON
arrays.
'''
if isinstance(val, dict):
new_dict = {}
for k,v in val.items():
if isinstance(v, str):
new_dict[k] = {'__base64__' : base64.b64encode(v)}
else:
new_dict[k] = json_encode_binary(v)
del val
return new_dict
elif isinstance(val, (list, tuple)):
new_list = []
n = len(val)
i = 0
while i < n:
v = val[i]
if isinstance(v, str):
new_list.append({'__base64__' : base64.b64encode(v)})
else:
new_list.append(json_encode_binary(v))
i += 1
del val
return new_list
elif isinstance(val, str):
return {'__base64__' : base64.b64encode(val)}
elif isinstance(val, Decimal):
return {'__base64__' : base64.b64encode(str(val))}
else:
return val
def json_decode_binary(val):
'''
JSON cannot transport binary data. In order to transport binary data we
convert binary data to a form like this:
{'__base64__' : base64_encoding_of_binary_value}
see json_encode_binary()
After JSON had decoded the JSON stream back into a Python object we must
recursively scan the object looking for any dicts which might represent
binary values and replace the dict containing the base64 encoding of the
binary value with the decoded binary value. Unlike the encoding problem
where the input might consist of immutable object, all JSON decoded
container are mutable so the conversion could be done in place. However we
don't modify objects in place because of side effects which may be
dangerous. Thus we elect to spend a few more cycles and avoid the
possibility of unintended side effects in favor of robustness.
'''
if isinstance(val, dict):
if val.has_key('__base64__'):
return base64.b64decode(val['__base64__'])
else:
new_dict = {}
for k,v in val.items():
if isinstance(v, dict) and v.has_key('__base64__'):
new_dict[k] = base64.b64decode(v['__base64__'])
else:
new_dict[k] = json_decode_binary(v)
del val
return new_dict
elif isinstance(val, list):
new_list = []
n = len(val)
i = 0
while i < n:
v = val[i]
if isinstance(v, dict) and v.has_key('__base64__'):
binary_val = base64.b64decode(v['__base64__'])
new_list.append(binary_val)
else:
new_list.append(json_decode_binary(v))
i += 1
del val
return new_list
else:
if isinstance(val, basestring):
try:
return val.decode('utf-8')
except UnicodeDecodeError:
raise ConversionError(
name=val,
error='incorrect type'
)
else:
return val
2009-10-13 12:28:00 -05:00
class jsonserver(WSGIExecutioner):
"""
JSON RPC server.
For information on the JSON-RPC spec, see:
http://json-rpc.org/wiki/specification
"""
content_type = 'application/json'
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
def __call__(self, environ, start_response):
'''
'''
self.debug('WSGI jsonserver.__call__:')
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
response = super(jsonserver, self).__call__(environ, start_response)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
return response
2009-10-13 12:28:00 -05:00
def marshal(self, result, error, _id=None):
if error:
assert isinstance(error, PublicError)
error = dict(
code=error.errno,
message=error.strerror,
name=error.__class__.__name__,
)
response = dict(
result=result,
error=error,
id=_id,
principal=unicode(context.principal),
version=unicode(VERSION),
2009-10-13 12:28:00 -05:00
)
response = json_encode_binary(response)
2009-10-13 12:28:00 -05:00
return json.dumps(response, sort_keys=True, indent=4)
def unmarshal(self, data):
try:
d = json.loads(data)
except ValueError, e:
raise JSONError(error=e)
if not isinstance(d, dict):
raise JSONError(error='Request must be a dict')
if 'method' not in d:
raise JSONError(error='Request is missing "method"')
if 'params' not in d:
raise JSONError(error='Request is missing "params"')
d = json_decode_binary(d)
2009-10-13 12:28:00 -05:00
method = d['method']
params = d['params']
_id = d.get('id')
if not isinstance(params, (list, tuple)):
raise JSONError(error='params must be a list')
if len(params) != 2:
raise JSONError(
error='params must contain [args, options]'
)
args = params[0]
if not isinstance(args, (list, tuple)):
raise JSONError(
error='params[0] (aka args) must be a list'
)
options = params[1]
if not isinstance(options, dict):
raise JSONError(
error='params[1] (aka options) must be a dict'
)
options = dict((str(k), v) for (k, v) in options.iteritems())
2009-10-13 12:28:00 -05:00
return (method, args, options, _id)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
class AuthManagerKerb(AuthManager):
'''
Instances of the AuthManger class are used to handle
authentication events delivered by the SessionManager. This class
specifcally handles the management of Kerbeos credentials which
may be stored in the session.
'''
def __init__(self, name):
super(AuthManagerKerb, self).__init__(name)
def logout(self, session_data):
'''
The current user has requested to be logged out. To accomplish
this we remove the user's kerberos credentials from their
session. This does not destroy the session, it just prevents
it from being used for fast authentication. Because the
credentials are no longer in the session cache any future
attempt will require the acquisition of credentials using one
of the login mechanisms.
'''
if session_data.has_key('ccache_data'):
self.debug('AuthManager.logout.%s: deleting ccache_data', self.name)
del session_data['ccache_data']
else:
self.error('AuthManager.logout.%s: session_data does not contain ccache_data', self.name)
class jsonserver_session(jsonserver):
"""
JSON RPC server protected with session auth.
"""
key = '/session/json'
def __init__(self):
super(jsonserver_session, self).__init__()
auth_mgr = AuthManagerKerb(self.__class__.__name__)
session_mgr.auth_mgr.register(auth_mgr.name, auth_mgr)
def need_login(self, start_response):
status = '401 Unauthorized'
headers = []
response = ''
self.debug('jsonserver_session: %s', status)
start_response(status, headers)
return [response]
def __call__(self, environ, start_response):
'''
'''
self.debug('WSGI jsonserver_session.__call__:')
# Load the session data
session_data = session_mgr.load_session_data(environ.get('HTTP_COOKIE'))
session_id = session_data['session_id']
self.debug('jsonserver_session.__call__: session_id=%s start_timestamp=%s write_timestamp=%s expiration_timestamp=%s',
session_id,
fmt_time(session_data['session_start_timestamp']),
fmt_time(session_data['session_write_timestamp']),
fmt_time(session_data['session_expiration_timestamp']))
ccache_data = session_data.get('ccache_data')
# Redirect to login if no Kerberos credentials
if ccache_data is None:
self.debug('no ccache, need login')
return self.need_login(start_response)
krbccache_pathname = store_krbccache_file(ccache_data)
# Redirect to login if Kerberos credentials are expired
cc = KRB5_CCache(krbccache_pathname)
if not cc.valid(self.api.env.host, self.api.env.realm):
self.debug('ccache expired, deleting session, need login')
delete_krbccache_file(krbccache_pathname)
return self.need_login(start_response)
# Store the session data in the per-thread context
setattr(context, 'session_data', session_data)
self.create_context(ccache=krbccache_pathname)
try:
response = super(jsonserver_session, self).__call__(environ, start_response)
finally:
# Kerberos may have updated the ccache data during the
# execution of the command therefore we need refresh our
# copy of it in the session data so the next command sees
# the same state of the ccache.
#
# However we must be careful not to restore the ccache
# data in the session data if it was explicitly deleted
# during the execution of the command. For example the
# logout command removes the ccache data from the session
# data to invalidate the session credentials.
if session_data.has_key('ccache_data'):
session_data['ccache_data'] = read_krbccache_file(krbccache_pathname)
# Delete the temporary ccache file we used
delete_krbccache_file(krbccache_pathname)
# Store the session data.
session_mgr.store_session_data(session_data)
destroy_context()
return response
class jsonserver_kerb(jsonserver):
"""
JSON RPC server protected with kerberos auth.
"""
key = '/json'
def __call__(self, environ, start_response):
'''
'''
self.debug('WSGI jsonserver_kerb.__call__:')
ccache=environ.get('KRB5CCNAME')
if ccache is None:
return self.marshal(None, CCacheError())
self.create_context(ccache=ccache)
try:
response = super(jsonserver_kerb, self).__call__(environ, start_response)
finally:
destroy_context()
return response
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
class krblogin(Backend):
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
key = '/login'
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
def __init__(self):
super(krblogin, self).__init__()
def _on_finalize(self):
super(krblogin, self)._on_finalize()
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
self.api.Backend.wsgi_dispatch.mount(self, self.key)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
# Set the session expiration time
try:
seconds = parse_time_duration(self.api.env.session_auth_duration)
self.session_auth_duration = int(seconds)
self.debug("session_auth_duration: %s", datetime.timedelta(seconds=self.session_auth_duration))
except Exception, e:
self.session_auth_duration = default_max_session_lifetime
self.error('unable to parse session_auth_duration, defaulting to %d: %s',
self.session_auth_duration, e)
def __call__(self, environ, start_response):
headers = []
self.debug('WSGI krblogin.__call__:')
# Get the ccache created by mod_auth_kerb
ccache=environ.get('KRB5CCNAME')
if ccache is None:
status = '500 Internal Error'
response = 'KRB5CCNAME not defined'
start_response(status, headers)
return [response]
ccache_scheme, ccache_location = krb5_parse_ccache(ccache)
assert ccache_scheme == 'FILE'
# Retrieve the session data (or newly create)
session_data = session_mgr.load_session_data(environ.get('HTTP_COOKIE'))
session_id = session_data['session_id']
# Copy the ccache file contents into the session data
session_data['ccache_data'] = read_krbccache_file(ccache_location)
# Compute when the session will expire
cc = KRB5_CCache(ccache)
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
endtime = cc.endtime(self.api.env.host, self.api.env.realm)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
# Account for clock skew and/or give us some time leeway
krb_expiration = endtime - krb_ticket_expiration_threshold
# Set the session expiration time
session_mgr.set_session_expiration_time(session_data,
lifetime=self.session_auth_duration,
max_age=krb_expiration)
# Store the session data now that it's been updated with the ccache
session_mgr.store_session_data(session_data)
self.debug('krblogin: ccache="%s" session_id="%s" ccache="%s"',
ccache, session_id, ccache)
# Return success and set session cookie
status = '200 Success'
response = ''
Tweak the session auth to reflect developer consensus. * Increase the session ID from 48 random bits to 128. * Implement the sesison_logout RPC command. It permits the UI to send a command that destroys the users credentials in the current session. * Restores the original web URL's and their authentication protections. Adds a new URL for sessions /ipa/session/json. Restores the original Kerberos auth which was for /ipa and everything below. New /ipa/session/json URL is treated as an exception and turns all authenticaion off. Similar to how /ipa/ui is handled. * Refactor the RPC handlers in rpcserver.py such that there is one handler per URL, specifically one handler per RPC and AuthMechanism combination. * Reworked how the URL names are used to map a URL to a handler. Previously it only permitted one level in the URL path hierarchy. We now dispatch on more that one URL path component. * Renames the api.Backend.session object to wsgi_dispatch. The use of the name session was historical and is now confusing since we've implemented sessions in a different location than the api.Backend.session object, which is really a WSGI dispatcher, hence the new name wsgi_dispatch. * Bullet-proof the setting of the KRB5CCNAME environment variable. ldap2.connect already sets it via the create_context() call but just in case that's not called or not called early enough (we now have other things besides ldap which need the ccache) we explicitly set it early as soon as we know it. * Rework how we test for credential validity and expiration. The previous code did not work with s4u2proxy because it assumed the existance of a TGT. Now we first try ldap credentials and if we can't find those fallback to the TGT. This logic was moved to the KRB5_CCache object, it's an imperfect location for it but it's the only location that makes sense at the moment given some of the current code limitations. The new methods are KRB5_CCache.valid() and KRB5_CCache.endtime(). * Add two new classes to session.py AuthManager and SessionAuthManager. Their purpose is to emit authication events to interested listeners. At the moment the logout event is the only event, but the framework should support other events as they arise. * Add BuildRequires python-memcached to freeipa.spec.in * Removed the marshaled_dispatch method, it was cruft, no longer referenced. https://fedorahosted.org/freeipa/ticket/2362
2012-02-15 09:26:42 -06:00
session_cookie = session_mgr.generate_cookie('/ipa', session_id)
add session manager and cache krb auth This patch adds a session manager and support for caching authentication in the session. Major elements of the patch are: * Add a session manager to support cookie based sessions which stores session data in a memcached entry. * Add ipalib/krb_utils.py which contains functions to parse ccache names, format principals, format KRB timestamps, and a KRB_CCache class which reads ccache entry and allows one to extract information such as the principal, credentials, credential timestamps, etc. * Move krb constants defined in ipalib/rpc.py to ipa_krb_utils.py so that all kerberos items are co-located. * Modify javascript in ipa.js so that the IPA.command() RPC call checks for authentication needed error response and if it receives it sends a GET request to /ipa/login URL to refresh credentials. * Add session_auth_duration config item to constants.py, used to configure how long a session remains valid. * Add parse_time_duration utility to ipalib/util.py. Used to parse the session_auth_duration config item. * Update the default.conf.5 man page to document session_auth_duration config item (also added documentation for log_manager config items which had been inadvertantly omitted from a previous commit). * Add SessionError object to ipalib/errors.py * Move Kerberos protection in Apache config from /ipa to /ipa/xml and /ipa/login * Add SessionCCache class to session.py to manage temporary Kerberos ccache file in effect for the duration of an RPC command. * Adds a krblogin plugin used to implement the /ipa/login handler. login handler sets the session expiration time, currently 60 minutes or the expiration of the TGT, whichever is shorter. It also copies the ccache provied by mod_auth_kerb into the session data. The json handler will later extract and validate the ccache belonging to the session. * Refactored the WSGI handlers so that json and xlmrpc could have independent behavior, this also moves where create and destroy context occurs, now done in the individual handler rather than the parent class. * The json handler now looks up the session data, validates the ccache bound to the session, if it's expired replies with authenicated needed error. * Add documentation to session.py. Fully documents the entire process, got questions, read the doc. * Add exclusions to make-lint as needed.
2012-02-06 12:29:56 -06:00
headers.append(('Set-Cookie', session_cookie))
start_response(status, headers)
return [response]