/** * A client-side 802.1x implementation supporting EAP/TLS * * This code is released under both the GPL version 2 and BSD licenses. * Either license may be used.  The respective licenses are found below. *  * Copyright (C) 2002 Bryan D. Payne & Nick L. Petroni Jr. * All Rights Reserved * * --- GPL Version 2 License --- * 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 2 * 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, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA. * * --- BSD License --- * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are met: * *  - Redistributions of source code must retain the above copyright notice, *    this list of conditions and the following disclaimer. *  - Redistributions in binary form must reproduce the above copyright *    notice, this list of conditions and the following disclaimer in the *    documentation and/or other materials provided with the distribution. *  - All advertising materials mentioning features or use of this software *    must display the following acknowledgement: *       This product includes software developed by the University of *       Maryland at College Park and its contributors. *  - Neither the name of the University nor the names of its contributors *    may be used to endorse or promote products derived from this software *    without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. *//******************************************************************* * EAPTTLS Function implementations *  * File: eapttls.c * * Authors: Chris.Hessing@utah.edu * * $Id: eapttls.c,v 1.25 2004/04/14 21:09:33 chessing Exp $ * $Date: 2004/04/14 21:09:33 $ * $Log: eapttls.c,v $ * Revision 1.25  2004/04/14 21:09:33  chessing * * Finished up extra error checking code.  Added ability to have passwords removed from memory on an authentication failure, so that a new password can be entered.  However, this feature has been disabled at this point due to a few small issues.  It will probably show up in 1.1. ;)  (It just isn't stable enough right now.) * * Revision 1.24  2004/04/13 22:13:46  chessing * * Additional error checking in all eap methods. * * Revision 1.23  2004/04/06 20:31:27  chessing * * PEAP NOW WORKS WITH IAS!!!!!! (Thanks to help from Matthew Gast!! (We love you! ;))  Also, added patches from yesterday's testing at iLabs, including some keying fixes, some segfault fixes, and a few other misc. issues.  iLabs testing has been worth it! * * Revision 1.22  2004/03/28 06:07:17  chessing * Added failure call to EAP methods to enable context resets for TLS based authentication protocols.  The resets are needed if an authentiction attempt fails, and we have session resumption enabled.  However, resetting it when we aren't using session resumption won't hurt anything, and probably isn't a bad idea.  The new failure handler can also be used to destroy passwords after a failed attempt, which will then cause xsupplicant to request another password from any listening GUIs. TLS session resumption is enabled (and works) for TLS and TTLS.  PEAP loops forever, and needs to be reviewed. * * Revision 1.21  2004/03/27 01:40:46  chessing * * Lots of small updates to free memory that wasn't getting freed, add some additional debug output, and fix a couple of memory leaks. * * Revision 1.20  2004/03/26 03:52:52  chessing * * Fixed a bug in xsup_debug that would cause config-parse to crash.  Added new key word for session resumption.  Added code to attempt session resumption.  So far, testing has not succeeded, but it is attempting resume. (Four TTLS packets are exchanged, and then we get a failure.)  More testing is needed. * * Revision 1.19  2004/03/19 23:43:57  chessing * * Lots of changes.  Changed the password prompting code to no longer require the EAP methods to maintain their own stale frame buffer.  (Frame buffer pointers should be moved out of generic_eap_data before a final release.)  Instead, EAP methods should set need_password in generic_eap_data to 1, along with the variables that identify the eap type being used, and the challenge data (if any -- only interesting to OTP/GTC at this point).  Also fixed up xsup_set_pwd.c, and got it back in CVS.  (For some reason, it was in limbo.)  Added xsup_monitor under gui_tools/cli.  xsup_monitor will eventually be a cli program that will monitor XSupplicant (running as a daemon) and display status information, and request passwords when they are not in the config. * * Revision 1.18  2004/03/17 21:21:41  chessing * * Hopefully xsup_set_pwd is in the right place now. ;)  Added the functions needed for xsupplicant to request a password from a GUI client.  (Still needs to be tested.)  Updated TTLS and PEAP to support password prompting.  Fixed up curState change in statemachine.c, so it doesn't print [ALL] in front of the current state. * * Revision 1.17  2004/03/15 16:23:25  chessing * * Added some checks to TLS using EAP types to make sure the root certificate isn't set to NULL.  (If it is, we can't authenticate, so we bail out.)  Changed the user certificate settings in the config file to all start with user_.  So, "cert" is now "user_cert", "key" is now "user_key", and "key_pass" is now "user_key_pass".  The structures and other related variables were also updated to reflect this change.  THIS WILL PROBABLY BREAK CONFIG FILES FOR SOME USERS!  (Be prepared for complaints on the list!)  ;) * * Revision 1.16  2004/03/05 23:58:45  chessing * * Added CN (sometimes called server name) checking to TTLS and PEAP.  This resulted in two new config options in the eap-ttls, and eap-peap blocks.  cncheck should be the name (or partial name) to match in the CN.  cnexact should be yes/no depending on if we want to match the CN exactly, or just see if our substring is in the CN. * * Revision 1.15  2004/03/02 01:03:53  chessing * * Added Jari Ahonen's SSL verification callback.  Added support to PEAP and TTLS to turn off certificate validation checking by setting the root_cert variable in the config to NONE.  (Case sensative!)  We will also display a warning when running in this mode.  Added initial hooks to support certificate CN checking. * * Revision 1.14  2004/02/28 01:26:39  chessing * * Several critical updates.  Fixed the HMAC failure on some keys. (This was due to a lot more than just an off-by-one.)  Fixed up the key decryption routine to identify key packets with no encrypted key, and use the peer key instead.  When using the peer key, we also can handle packets that are padded funny.  (Our Cisco AP1200 has two null pad bytes at the end of some key frames.)  Changed the response ID function to not add a 00 to the end of the ID.  The 00 byte shouldn't have been seen by the RADIUS server unless they were not paying attention to the EAP-Length.  So, this wasn't really a bug fix.  Started to add support for CN checking for TLS based protocols. * * Revision 1.13  2004/02/16 05:06:01  chessing * * Added support for the deny_interfaces, and allow_interfaces options in the config file.  (We should now have support for *EVERYTHING* in the new config file format!)  Updated EAP types other than SIM to use the new password prompt/delay code.  Phase 2 of TTLS still needs to be completed, along with the code that actually notifies the GUI. * * Revision 1.12  2004/01/17 21:16:16  chessing * * Various segfault fixes.  PEAP now works correctly again.  Some new error checking in the tls handlers.  Fixes for the way we determine if we have changed ESSIDs.  We now quit when we don't have a config, or when the config is bad. Added code to check and see if a frame is in the queue, and don't sleep if there is.  "Fixed" ID issue by inheriting the ID from the parent where needed.  However, assigning an ID inside of a handler will override the parent ID.  This could cause problems with some EAP types.  We should add a "username" field to PEAP to allow configuration of the inner EAP identity. * * Revision 1.11  2004/01/15 23:45:12  chessing * * Fixed a segfault when looking for wireless interfaces when all we had was a wired interface.  Fixed external command execution so that junk doesn't end up in the processed string anymore.  Changed the state machine to call txRspAuth even if there isn't a frame to process.  This will enable EAP methods to request information from a GUI interface (such as passwords, or supply challenge information that might be needed to generate passwords).  EAP methods now must decide what to do when they are handed NULL for the pointer to the in frame.  If they don't need any more data, they should quietly exit. * * Revision 1.10  2004/01/15 01:12:45  chessing * * Fixed a keying problem (keying material wasn't being generated correctly).  Added support for global counter variables from the config file. (Such as auth_period)  Added support for executing command defined in the config file based on different events.  (Things such as what to do on reauth.)  Added the ability to roam to a different SSID.  We now check to make sure our BSSID hasn't changed, and we follow it, if it has.  Fixed a sefault when the program was terminated in certain states.  Added attempt at better garbage collection on program termination. Various small code cleanups. * * Revision 1.9  2004/01/14 22:07:25  chessing * * Fixes that were needed in order to allow us to authenticate correctly.  We should now be able to authenticate using only information provided by the config file! * * Revision 1.8  2004/01/14 05:44:48  chessing * * Added pid file support. (Very basic for now, needs to be improved a little.)  Attempted to add setup of global variables. (Need to figure out why it is segfaulting.)  Added more groundwork for IPC. * * Revision 1.7  2004/01/13 01:55:56  chessing * * Major changes to EAP related code.  We no longer pass in an interface_data struct to EAP handlers.  Instead, we hand in a generic_eap_data struct which containsnon-interface specific information.  This will allow EAP types to be reused as phase 2 type easier.  However, this new code may create issues with EAP types that make use of the identity in the eap type.  Somehow, the identity value needs to propigate down to the EAP method.  It currently does not.  This should be any easy fix, but more testing will be needed. * * Revision 1.6  2004/01/06 23:35:08  chessing * * Fixed a couple known bugs in SIM.  Config file support should now be in place!!! But, because of the changes, PEAP is probably broken.  We will need to reconsider how the phase 2 piece of PEAP works. * * Revision 1.5  2003/12/14 06:11:03  chessing * * Fixed some stuff with SIM in relation to the new config structures.  Cleaned out CR/LF from LEAP source files.  Added user certificate support to TTLS and PEAP. Some additions to the IPC code. (Not tested yet.) * * Revision 1.4  2003/11/24 02:14:08  chessing * * Added EAP-SIM (draft 11 still needs work), various small changes to eap calls, new hex dump code including ASCII dump (used mostly for dumping frames) * * Revision 1.3  2003/11/22 06:10:39  chessing * * Changes to the eap type process calls, to remove a pointless parameter. * * Revision 1.2  2003/11/21 05:09:47  chessing * * PEAP now works! * * Revision 1.1.1.1  2003/11/19 04:13:27  chessing * New source tree * * *******************************************************************/#include <string.h>#include "profile.h"#include "config.h"#include "xsup_debug.h"#include "xsup_err.h"#include "frame_structs.h"#include "eap_types/tls/eaptls.h"#include "eap_types/ttls/eapttls.h"#include "eap_types/ttls/ttlsphase2.h"#include "eap_types/tls/tls_funcs.h"#include "eap.h"#include "interactive.h"int eapttls_setup(struct generic_eap_data *thisint){  struct tls_vars *mytls_vars;  struct config_eap_ttls *userdata;  if (!thisint)    {      debug_printf(DEBUG_NORMAL, "Invalid interface structure passed to eapttls_setup()!\n");      return XEMALLOC;    }  userdata = (struct config_eap_ttls *)thisint->eap_conf_data;  if (userdata == NULL)    {      debug_printf(DEBUG_NORMAL, "EAP-TTLS setup was passed NULL userdata!  We cannot continue with this attempt at authentication!\n");      return XENOUSERDATA;    }  // First, set up the structure to hold all of our instance specific  // variables.  thisint->eap_data = (char *)malloc(sizeof(struct tls_vars));  if (thisint->eap_data == NULL)    {      debug_printf(DEBUG_NORMAL, "Unable to allocate memory for thisint->eapdata in eapttls_setup()!\n");      return XEMALLOC;    }  mytls_vars = (struct tls_vars *)thisint->eap_data;  // Set our variables to NULL.  mytls_vars->ctx = NULL;  mytls_vars->ssl = NULL;  mytls_vars->ssl_in = NULL;  mytls_vars->ssl_out = NULL;  mytls_vars->tlsoutdata = NULL;