2023-06-30 13:06:01 +03:00
/*
2002-01-30 09:08:46 +03:00
Unix SMB / CIFS implementation .
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
Password and authentication handling
2002-01-05 07:55:41 +03:00
Copyright ( C ) Andrew Bartlett 2001 - 2002
2009-02-21 15:20:08 +03:00
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
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
2007-07-09 23:25:36 +04:00
the Free Software Foundation ; either version 3 of the License , or
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
( at your option ) any later version .
2009-02-21 15:20:08 +03:00
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
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 .
2009-02-21 15:20:08 +03:00
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
You should have received a copy of the GNU General Public License
2007-07-10 04:52:41 +04:00
along with this program . If not , see < http : //www.gnu.org/licenses/>.
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
*/
# include "includes.h"
2011-03-25 04:28:05 +03:00
# include "auth.h"
2011-06-08 20:55:37 +04:00
# include "../lib/tsocket/tsocket.h"
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
2019-01-28 05:30:23 +03:00
# include "param/param.h"
# include "../lib/messaging/messaging.h"
2021-01-03 23:53:49 +03:00
# include "lib/global_contexts.h"
2019-01-28 05:30:23 +03:00
2002-07-15 14:35:28 +04:00
# undef DBGC_CLASS
# define DBGC_CLASS DBGC_AUTH
2002-01-05 07:55:41 +03:00
2006-04-26 19:41:25 +04:00
static_decl_auth ;
2010-08-18 19:31:39 +04:00
static struct auth_init_function_entry * auth_backends = NULL ;
2003-04-16 16:13:07 +04:00
2003-04-28 21:48:48 +04:00
static struct auth_init_function_entry * auth_find_backend_entry ( const char * name ) ;
2003-05-01 03:06:44 +04:00
NTSTATUS smb_register_auth ( int version , const char * name , auth_init_function init )
2003-04-16 16:13:07 +04:00
{
2019-09-13 16:57:16 +03:00
struct auth_init_function_entry * entry = NULL ;
2003-04-16 16:13:07 +04:00
2003-04-28 21:48:48 +04:00
if ( version ! = AUTH_INTERFACE_VERSION ) {
DEBUG ( 0 , ( " Can't register auth_method! \n "
" You tried to register an auth module with AUTH_INTERFACE_VERSION %d, while this version of samba uses %d \n " ,
version , AUTH_INTERFACE_VERSION ) ) ;
return NT_STATUS_OBJECT_TYPE_MISMATCH ;
}
if ( ! name | | ! init ) {
return NT_STATUS_INVALID_PARAMETER ;
}
2003-04-16 16:13:07 +04:00
DEBUG ( 5 , ( " Attempting to register auth backend %s \n " , name ) ) ;
2003-04-28 21:48:48 +04:00
if ( auth_find_backend_entry ( name ) ) {
DEBUG ( 0 , ( " There already is an auth method registered with the name %s! \n " , name ) ) ;
return NT_STATUS_OBJECT_NAME_COLLISION ;
2003-04-16 16:13:07 +04:00
}
2009-02-21 15:20:08 +03:00
2004-12-07 21:25:53 +03:00
entry = SMB_XMALLOC_P ( struct auth_init_function_entry ) ;
2003-04-16 16:13:07 +04:00
entry - > name = smb_xstrdup ( name ) ;
entry - > init = init ;
2010-08-18 19:31:39 +04:00
DLIST_ADD ( auth_backends , entry ) ;
2003-04-28 21:48:48 +04:00
DEBUG ( 5 , ( " Successfully added auth method '%s' \n " , name ) ) ;
return NT_STATUS_OK ;
2003-04-16 16:13:07 +04:00
}
static struct auth_init_function_entry * auth_find_backend_entry ( const char * name )
{
2010-08-18 19:31:39 +04:00
struct auth_init_function_entry * entry = auth_backends ;
2003-04-16 16:13:07 +04:00
while ( entry ) {
2003-05-16 10:20:57 +04:00
if ( strcmp ( entry - > name , name ) = = 0 ) return entry ;
2003-04-16 16:13:07 +04:00
entry = entry - > next ;
}
2009-02-21 15:20:08 +03:00
2003-04-16 16:13:07 +04:00
return NULL ;
}
2002-01-05 07:55:41 +03:00
/****************************************************************************
2002-07-15 14:35:28 +04:00
Try to get a challenge out of the various authentication modules .
2002-01-05 07:55:41 +03:00
Returns a const char of length 8 bytes .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2012-02-03 14:58:44 +04:00
NTSTATUS auth_get_ntlm_challenge ( struct auth_context * auth_context ,
uint8_t chal [ 8 ] )
2002-01-05 07:55:41 +03:00
{
if ( auth_context - > challenge . length ) {
2023-07-03 12:19:33 +03:00
DBG_INFO ( " get_ntlm_challenge (auth subsystem): returning "
" previous challenge by module %s (normal) \n " ,
auth_context - > challenge_set_by ) ;
2009-02-20 01:41:48 +03:00
memcpy ( chal , auth_context - > challenge . data , 8 ) ;
2009-12-14 11:43:59 +03:00
return NT_STATUS_OK ;
2002-01-05 07:55:41 +03:00
}
2020-01-03 14:51:04 +03:00
auth_context - > challenge = data_blob_talloc ( auth_context , NULL , 8 ) ;
2020-01-02 14:41:16 +03:00
if ( auth_context - > challenge . data = = NULL ) {
DBG_WARNING ( " data_blob_talloc failed \n " ) ;
return NT_STATUS_NO_MEMORY ;
}
2020-01-03 14:51:04 +03:00
generate_random_buffer (
auth_context - > challenge . data , auth_context - > challenge . length ) ;
2002-01-05 07:55:41 +03:00
2012-06-30 12:30:57 +04:00
auth_context - > challenge_set_by = " random " ;
2002-01-05 07:55:41 +03:00
2009-02-20 01:41:48 +03:00
memcpy ( chal , auth_context - > challenge . data , 8 ) ;
2009-12-14 11:43:59 +03:00
return NT_STATUS_OK ;
2002-01-05 07:55:41 +03:00
}
2001-12-30 13:54:58 +03:00
/**
* Check user is in correct domain ( if required )
*
* @ param user Only used to fill in the debug message
2023-06-30 13:06:01 +03:00
*
2001-12-30 13:54:58 +03:00
* @ param domain The domain to be verified
*
2023-06-30 13:06:01 +03:00
* @ return True if the user can connect with that domain ,
2001-12-30 13:54:58 +03:00
* False otherwise .
* */
2023-06-30 13:06:01 +03:00
static bool check_domain_match ( const char * user , const char * domain )
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
{
2001-12-19 12:53:30 +03:00
/*
* If we aren ' t serving to trusted domains , we must make sure that
* the validation request comes from an account in the same domain
* as the Samba server
*/
if ( ! lp_allow_trusted_domains ( ) & &
2023-06-30 13:06:01 +03:00
! ( strequal ( " " , domain ) | |
strequal ( lp_workgroup ( ) , domain ) | |
2002-11-13 02:20:50 +03:00
is_myname ( domain ) ) ) {
2001-12-19 12:53:30 +03:00
DEBUG ( 1 , ( " check_domain_match: Attempt to connect as user %s from domain %s denied. \n " , user , domain ) ) ;
return False ;
} else {
return True ;
}
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
}
2001-12-30 13:54:58 +03:00
/**
* Check a user ' s Plaintext , LM or NTLM password .
*
* Check a user ' s password , as given in the user_info struct and return various
* interesting details in the server_info struct .
*
* This function does NOT need to be in a become_root ( ) / unbecome_root ( ) pair
* as it makes the calls itself when needed .
*
2023-06-30 13:06:01 +03:00
* The return value takes precedence over the contents of the server_info
* struct . When the return is other than NT_STATUS_OK the contents
2001-12-30 13:54:58 +03:00
* of that structure is undefined .
*
* @ param user_info Contains the user supplied components , including the passwords .
* Must be created with make_user_info ( ) or one of its wrappers .
*
2023-06-30 13:06:01 +03:00
* @ param auth_context Supplies the challenges and some other data .
* Must be created with make_auth_context ( ) , and the challenges should be
* filled in , either at creation or by calling the challenge generation
* function auth_get_challenge ( ) .
2001-12-30 13:54:58 +03:00
*
2017-03-17 11:43:59 +03:00
* @ param pserver_info If successful , contains information about the authentication ,
* including a struct samu struct describing the user .
*
* @ param pauthoritative Indicates if the result should be treated as final
* result .
2001-12-30 13:54:58 +03:00
*
* @ return An NTSTATUS with NT_STATUS_OK or an appropriate error .
*
* */
2014-02-18 13:19:57 +04:00
NTSTATUS auth_check_ntlm_password ( TALLOC_CTX * mem_ctx ,
const struct auth_context * auth_context ,
const struct auth_usersupplied_info * user_info ,
2017-03-17 11:43:59 +03:00
struct auth_serversupplied_info * * pserver_info ,
uint8_t * pauthoritative )
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
{
2017-02-11 13:26:09 +03:00
TALLOC_CTX * frame ;
2017-02-11 17:44:01 +03:00
const char * auth_method_name = " " ;
2003-07-03 18:36:42 +04:00
/* if all the modules say 'not for me' this is reasonable */
2017-03-17 11:43:59 +03:00
NTSTATUS nt_status = NT_STATUS_NOT_IMPLEMENTED ;
2003-07-05 12:05:06 +04:00
const char * unix_username ;
2020-01-03 15:47:14 +03:00
struct auth_methods * auth_method ;
2017-02-23 03:50:14 +03:00
struct auth_serversupplied_info * server_info = NULL ;
struct dom_sid sid = { 0 } ;
2019-01-28 05:30:23 +03:00
struct imessaging_context * msg_ctx = NULL ;
struct loadparm_context * lp_ctx = NULL ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
2014-02-18 13:19:57 +04:00
if ( user_info = = NULL | | auth_context = = NULL | | pserver_info = = NULL ) {
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
return NT_STATUS_LOGON_FAILURE ;
2014-02-18 13:19:57 +04:00
}
2001-10-31 13:46:25 +03:00
2017-02-11 13:26:09 +03:00
frame = talloc_stackframe ( ) ;
2019-01-28 05:30:23 +03:00
if ( lp_auth_event_notification ( ) ) {
lp_ctx = loadparm_init_s3 ( frame , loadparm_s3_helpers ( ) ) ;
msg_ctx = imessaging_client_init (
frame , lp_ctx , global_event_context ( ) ) ;
}
2017-03-17 11:43:59 +03:00
* pauthoritative = 1 ;
2023-07-03 12:19:33 +03:00
DBG_NOTICE ( " check_ntlm_password: Checking password for unmapped user "
" [%s] \\ [%s]@[%s] with the new password interface \n " ,
user_info - > client . domain_name ,
user_info - > client . account_name ,
user_info - > workstation_name ) ;
2001-11-09 01:19:01 +03:00
2023-07-03 12:19:33 +03:00
DBG_NOTICE ( " check_ntlm_password: mapped user is: [%s] \\ [%s]@[%s] \n " ,
user_info - > mapped . domain_name ,
user_info - > mapped . account_name ,
user_info - > workstation_name ) ;
2003-01-28 15:07:02 +03:00
if ( auth_context - > challenge . length ! = 8 ) {
DEBUG ( 0 , ( " check_ntlm_password: Invalid challenge stored for this auth context - cannot continue \n " ) ) ;
2017-02-11 13:24:22 +03:00
nt_status = NT_STATUS_LOGON_FAILURE ;
goto fail ;
2003-01-28 15:07:02 +03:00
}
2002-12-12 02:54:40 +03:00
if ( auth_context - > challenge_set_by )
DEBUG ( 10 , ( " check_ntlm_password: auth_context challenge created by %s \n " ,
auth_context - > challenge_set_by ) ) ;
2001-11-26 07:05:28 +03:00
DEBUG ( 10 , ( " challenge is: \n " ) ) ;
2007-03-28 17:34:59 +04:00
dump_data ( 5 , auth_context - > challenge . data , auth_context - > challenge . length ) ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
# ifdef DEBUG_PASSWORD
2023-06-30 13:06:01 +03:00
DEBUG ( 100 , ( " user_info has passwords of length %d and %d \n " ,
2010-06-01 15:52:01 +04:00
( int ) user_info - > password . response . lanman . length , ( int ) user_info - > password . response . nt . length ) ) ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
DEBUG ( 100 , ( " lm: \n " ) ) ;
2010-06-01 15:52:01 +04:00
dump_data ( 100 , user_info - > password . response . lanman . data , user_info - > password . response . lanman . length ) ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
DEBUG ( 100 , ( " nt: \n " ) ) ;
2010-06-01 15:52:01 +04:00
dump_data ( 100 , user_info - > password . response . nt . data , user_info - > password . response . nt . length ) ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
# endif
2001-12-30 13:54:58 +03:00
/* This needs to be sorted: If it doesn't match, what should we do? */
2017-02-11 13:24:22 +03:00
if ( ! check_domain_match ( user_info - > client . account_name ,
user_info - > mapped . domain_name ) ) {
nt_status = NT_STATUS_LOGON_FAILURE ;
goto fail ;
}
2001-12-30 13:54:58 +03:00
2002-12-12 02:54:40 +03:00
for ( auth_method = auth_context - > auth_method_list ; auth_method ; auth_method = auth_method - > next ) {
2017-02-11 17:44:01 +03:00
auth_method_name = auth_method - > name ;
2009-02-21 15:20:08 +03:00
2017-02-11 17:44:01 +03:00
nt_status = auth_method - > auth ( auth_context ,
auth_method - > private_data ,
talloc_tos ( ) ,
user_info ,
& server_info ) ;
2003-06-28 12:29:42 +04:00
2017-02-11 17:44:01 +03:00
if ( ! NT_STATUS_EQUAL ( nt_status , NT_STATUS_NOT_IMPLEMENTED ) ) {
break ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
}
2002-01-01 06:10:32 +03:00
2017-02-11 17:44:01 +03:00
DBG_DEBUG ( " %s had nothing to say \n " , auth_method - > name ) ;
2001-11-09 01:19:01 +03:00
}
2001-10-31 13:46:25 +03:00
2017-03-17 11:43:59 +03:00
if ( NT_STATUS_EQUAL ( nt_status , NT_STATUS_NOT_IMPLEMENTED ) ) {
* pauthoritative = 0 ;
2017-02-11 17:44:01 +03:00
nt_status = NT_STATUS_NO_SUCH_USER ;
}
2009-02-21 15:20:08 +03:00
2017-02-11 13:38:56 +03:00
if ( ! NT_STATUS_IS_OK ( nt_status ) ) {
2017-02-11 17:44:01 +03:00
DBG_INFO ( " %s authentication for user [%s] FAILED with "
2017-03-17 11:43:59 +03:00
" error %s, authoritative=%u \n " ,
2017-02-11 17:44:01 +03:00
auth_method_name ,
user_info - > client . account_name ,
2017-03-17 11:43:59 +03:00
nt_errstr ( nt_status ) ,
* pauthoritative ) ;
2017-02-11 13:38:56 +03:00
goto fail ;
}
2011-06-08 20:55:37 +04:00
2017-02-11 17:44:01 +03:00
DBG_NOTICE ( " %s authentication for user [%s] succeeded \n " ,
auth_method_name , user_info - > client . account_name ) ;
2017-02-11 13:38:56 +03:00
unix_username = server_info - > unix_name ;
/* We skip doing this step if the caller asked us not to */
if ( ! ( user_info - > flags & USER_INFO_INFO3_AND_NO_AUTHZ )
& & ! ( server_info - > guest ) ) {
const char * rhost ;
if ( tsocket_address_is_inet ( user_info - > remote_host , " ip " ) ) {
rhost = tsocket_address_inet_addr_string (
user_info - > remote_host , talloc_tos ( ) ) ;
if ( rhost = = NULL ) {
nt_status = NT_STATUS_NO_MEMORY ;
goto fail ;
}
} else {
rhost = " 127.0.0.1 " ;
2001-11-09 01:19:01 +03:00
}
2009-02-21 15:20:08 +03:00
2017-02-11 13:38:56 +03:00
/* We might not be root if we are an RPC call */
become_root ( ) ;
nt_status = smb_pam_accountcheck ( unix_username , rhost ) ;
unbecome_root ( ) ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
if ( NT_STATUS_IS_OK ( nt_status ) ) {
2017-02-11 13:38:56 +03:00
DEBUG ( 5 , ( " check_ntlm_password: PAM Account for user [%s] "
" succeeded \n " , unix_username ) ) ;
} else {
DEBUG ( 3 , ( " check_ntlm_password: PAM Account for user [%s] "
" FAILED with error %s \n " ,
unix_username , nt_errstr ( nt_status ) ) ) ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
}
2017-02-11 13:38:56 +03:00
}
2009-02-21 15:20:08 +03:00
2017-02-11 13:38:56 +03:00
if ( ! NT_STATUS_IS_OK ( nt_status ) ) {
goto fail ;
2001-10-31 13:46:25 +03:00
}
2009-02-21 15:20:08 +03:00
2017-02-23 03:50:14 +03:00
nt_status = get_user_sid_info3_and_extra ( server_info - > info3 ,
& server_info - > extra ,
& sid ) ;
if ( ! NT_STATUS_IS_OK ( nt_status ) ) {
sid = ( struct dom_sid ) { 0 } ;
}
2019-01-28 05:30:23 +03:00
log_authentication_event ( msg_ctx ,
lp_ctx ,
2018-06-10 14:00:34 +03:00
& auth_context - > start_time ,
user_info ,
nt_status ,
2017-02-23 03:50:14 +03:00
server_info - > info3 - > base . logon_domain . string ,
server_info - > info3 - > base . account_name . string ,
2023-06-15 08:07:05 +03:00
& sid ,
NULL /* client_audit_info */ ,
NULL /* server_audit_info */ ) ;
2017-02-23 03:50:14 +03:00
2017-02-11 13:38:56 +03:00
DEBUG ( server_info - > guest ? 5 : 2 ,
( " check_ntlm_password: %sauthentication for user "
" [%s] -> [%s] -> [%s] succeeded \n " ,
server_info - > guest ? " guest " : " " ,
user_info - > client . account_name ,
user_info - > mapped . account_name ,
unix_username ) ) ;
* pserver_info = talloc_move ( mem_ctx , & server_info ) ;
2017-02-11 13:38:56 +03:00
TALLOC_FREE ( frame ) ;
2017-02-11 13:38:56 +03:00
return NT_STATUS_OK ;
2017-02-11 13:38:56 +03:00
2017-02-11 13:24:22 +03:00
fail :
2005-06-08 18:23:49 +04:00
/* failed authentication; check for guest lapping */
2009-02-21 15:20:08 +03:00
2017-02-20 05:55:34 +03:00
/*
* Please try not to change this string , it is probably in use
* in audit logging tools
*/
2017-03-17 11:43:59 +03:00
DEBUG ( 2 , ( " check_ntlm_password: Authentication for user "
" [%s] -> [%s] FAILED with error %s, authoritative=%u \n " ,
2010-06-01 14:30:56 +04:00
user_info - > client . account_name , user_info - > mapped . account_name ,
2017-03-17 11:43:59 +03:00
nt_errstr ( nt_status ) , * pauthoritative ) ) ;
2017-02-23 03:50:14 +03:00
2019-01-28 05:30:23 +03:00
log_authentication_event ( msg_ctx ,
lp_ctx ,
2018-06-10 14:00:34 +03:00
& auth_context - > start_time ,
user_info ,
nt_status ,
NULL ,
NULL ,
2023-06-15 08:07:05 +03:00
NULL ,
NULL /* client_audit_info */ ,
NULL /* server_audit_info */ ) ;
2017-02-23 03:50:14 +03:00
2014-02-18 13:19:57 +04:00
ZERO_STRUCTP ( pserver_info ) ;
2009-02-21 15:20:08 +03:00
2017-02-11 13:26:09 +03:00
TALLOC_FREE ( frame ) ;
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
return nt_status ;
}
2002-01-05 07:55:41 +03:00
/***************************************************************************
Clear out a auth_context , and destroy the attached TALLOC_CTX
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
2010-07-17 02:23:55 +04:00
static int auth_context_destructor ( void * ptr )
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
{
2010-07-17 02:23:55 +04:00
struct auth_context * ctx = talloc_get_type ( ptr , struct auth_context ) ;
struct auth_methods * am ;
2003-11-23 03:16:54 +03:00
2010-07-17 02:23:55 +04:00
/* Free private data of context's authentication methods */
for ( am = ctx - > auth_method_list ; am ; am = am - > next ) {
TALLOC_FREE ( am - > private_data ) ;
2003-11-23 03:16:54 +03:00
}
2010-07-17 02:23:55 +04:00
return 0 ;
2001-10-31 13:46:25 +03:00
}
2001-09-29 17:08:26 +04:00
2002-01-05 07:55:41 +03:00
/***************************************************************************
Make a auth_info struct
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
2010-09-26 01:50:46 +04:00
static NTSTATUS make_auth_context ( TALLOC_CTX * mem_ctx ,
struct auth_context * * auth_context )
2002-01-05 07:55:41 +03:00
{
2010-07-17 02:23:55 +04:00
struct auth_context * ctx ;
2010-09-26 01:50:46 +04:00
ctx = talloc_zero ( mem_ctx , struct auth_context ) ;
2010-07-17 02:23:55 +04:00
if ( ! ctx ) {
2002-01-05 07:55:41 +03:00
DEBUG ( 0 , ( " make_auth_context: talloc failed! \n " ) ) ;
return NT_STATUS_NO_MEMORY ;
}
2001-10-31 09:20:58 +03:00
2018-06-10 14:00:34 +03:00
ctx - > start_time = timeval_current ( ) ;
2010-07-17 02:23:55 +04:00
talloc_set_destructor ( ( TALLOC_CTX * ) ctx , auth_context_destructor ) ;
2009-02-21 15:20:08 +03:00
2010-07-17 02:23:55 +04:00
* auth_context = ctx ;
2002-01-05 07:55:41 +03:00
return NT_STATUS_OK ;
}
2001-10-31 09:20:58 +03:00
2021-04-14 13:33:09 +03:00
static bool load_auth_module (
struct auth_context * auth_context ,
const char * module ,
struct auth_methods * * ret )
2003-04-24 15:56:09 +04:00
{
2007-10-19 04:40:25 +04:00
static bool initialised_static_modules = False ;
2003-04-24 15:56:09 +04:00
struct auth_init_function_entry * entry ;
char * module_name = smb_xstrdup ( module ) ;
char * module_params = NULL ;
char * p ;
2007-10-19 04:40:25 +04:00
bool good = False ;
2003-04-24 15:56:09 +04:00
/* Initialise static modules if not done so yet */
if ( ! initialised_static_modules ) {
2017-04-20 22:24:43 +03:00
static_init_auth ( NULL ) ;
2003-04-24 15:56:09 +04:00
initialised_static_modules = True ;
}
2009-02-21 15:20:08 +03:00
2003-04-24 15:56:09 +04:00
DEBUG ( 5 , ( " load_auth_module: Attempting to find an auth method to match %s \n " ,
module ) ) ;
2009-02-21 15:20:08 +03:00
2003-04-24 15:56:09 +04:00
p = strchr ( module_name , ' : ' ) ;
if ( p ) {
* p = 0 ;
module_params = p + 1 ;
2003-09-05 23:59:55 +04:00
trim_char ( module_params , ' ' , ' ' ) ;
2003-04-24 15:56:09 +04:00
}
2009-02-21 15:20:08 +03:00
2003-09-05 23:59:55 +04:00
trim_char ( module_name , ' ' , ' ' ) ;
2009-02-21 15:20:08 +03:00
2003-04-24 15:56:09 +04:00
entry = auth_find_backend_entry ( module_name ) ;
2009-02-21 15:20:08 +03:00
2003-05-15 12:47:28 +04:00
if ( entry = = NULL ) {
if ( NT_STATUS_IS_OK ( smb_probe_module ( " auth " , module_name ) ) ) {
entry = auth_find_backend_entry ( module_name ) ;
}
}
if ( entry ! = NULL ) {
if ( ! NT_STATUS_IS_OK ( entry - > init ( auth_context , module_params , ret ) ) ) {
DEBUG ( 0 , ( " load_auth_module: auth method %s did not correctly init \n " ,
module_name ) ) ;
} else {
DEBUG ( 5 , ( " load_auth_module: auth method %s has a valid init \n " ,
module_name ) ) ;
good = True ;
}
2003-04-24 15:56:09 +04:00
} else {
2003-05-15 12:47:28 +04:00
DEBUG ( 0 , ( " load_auth_module: can't find auth method %s! \n " , module_name ) ) ;
2003-04-24 15:56:09 +04:00
}
2003-05-15 12:47:28 +04:00
2003-04-24 15:56:09 +04:00
SAFE_FREE ( module_name ) ;
return good ;
}
2002-01-05 07:55:41 +03:00
/***************************************************************************
Make a auth_info struct for the auth subsystem
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
2010-09-26 01:51:45 +04:00
static NTSTATUS make_auth_context_text_list ( TALLOC_CTX * mem_ctx ,
struct auth_context * * auth_context ,
char * * text_list )
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
{
2020-01-03 15:47:14 +03:00
struct auth_methods * list = NULL ;
struct auth_methods * t , * method = NULL ;
2001-10-31 13:46:25 +03:00
NTSTATUS nt_status ;
This is another rather major change to the samba authenticaion
subystem.
The particular aim is to modularized the interface - so that we
can have arbitrary password back-ends.
This code adds one such back-end, a 'winbind' module to authenticate
against the winbind_auth_crap functionality. While fully-functional
this code is mainly useful as a demonstration, because we don't get
back the info3 as we would for direct ntdomain authentication.
This commit introduced the new 'auth methods' parameter, in the
spirit of the 'auth order' discussed on the lists. It is renamed
because not all the methods may be consulted, even if previous
methods fail - they may not have a suitable challenge for example.
Also, we have a 'local' authentication method, for old-style
'unix if plaintext, sam if encrypted' authentication and a
'guest' module to handle guest logins in a single place.
While this current design is not ideal, I feel that it does
provide a better infrastructure than the current design, and can
be built upon.
The following parameters have changed:
- use rhosts =
This has been replaced by the 'rhosts' authentication method,
and can be specified like 'auth methods = guest rhosts'
- hosts equiv =
This needs both this parameter and an 'auth methods' entry
to be effective. (auth methods = guest hostsequiv ....)
- plaintext to smbpasswd =
This is replaced by specifying 'sam' rather than 'local'
in the auth methods.
The security = parameter is unchanged, and now provides defaults
for the 'auth methods' parameter.
The available auth methods are:
guest
rhosts
hostsequiv
sam (passdb direct hash access)
unix (PAM, crypt() etc)
local (the combination of the above, based on encryption)
smbserver (old security=server)
ntdomain (old security=domain)
winbind (use winbind to cache DC connections)
Assistance in testing, or the production of new and interesting
authentication modules is always appreciated.
Andrew Bartlett
(This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
2002-01-05 07:55:41 +03:00
if ( ! text_list ) {
2002-12-12 02:54:40 +03:00
DEBUG ( 2 , ( " make_auth_context_text_list: No auth method list!? \n " ) ) ;
2002-01-05 07:55:41 +03:00
return NT_STATUS_UNSUCCESSFUL ;
}
2009-02-21 15:20:08 +03:00
2010-09-26 01:51:45 +04:00
nt_status = make_auth_context ( mem_ctx , auth_context ) ;
2010-09-26 01:50:46 +04:00
if ( ! NT_STATUS_IS_OK ( nt_status ) ) {
2002-01-05 07:55:41 +03:00
return nt_status ;
2010-09-26 01:50:46 +04:00
}
2003-04-16 16:13:07 +04:00
2023-06-30 13:06:01 +03:00
for ( ; * text_list ; text_list + + ) {
2003-04-24 15:56:09 +04:00
if ( load_auth_module ( * auth_context , * text_list , & t ) ) {
2016-02-05 13:32:18 +03:00
DLIST_ADD_END ( list , t ) ;
2003-04-24 15:56:09 +04:00
}
2002-01-05 07:55:41 +03:00
}
2009-02-21 15:20:08 +03:00
2002-01-05 07:55:41 +03:00
( * auth_context ) - > auth_method_list = list ;
2009-02-21 15:20:08 +03:00
2012-02-03 09:14:42 +04:00
/* Look for the first module to provide a prepare_gensec and
* make_auth4_context hook , and set that if provided */
2011-07-21 08:27:00 +04:00
for ( method = ( * auth_context ) - > auth_method_list ; method ; method = method - > next ) {
2012-02-03 09:14:42 +04:00
if ( method - > prepare_gensec & & method - > make_auth4_context ) {
2011-07-26 04:01:39 +04:00
( * auth_context ) - > prepare_gensec = method - > prepare_gensec ;
2012-02-03 09:14:42 +04:00
( * auth_context ) - > make_auth4_context = method - > make_auth4_context ;
2011-07-21 08:27:00 +04:00
break ;
}
}
return NT_STATUS_OK ;
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
}
2017-03-13 10:58:43 +03:00
static NTSTATUS make_auth_context_specific ( TALLOC_CTX * mem_ctx ,
struct auth_context * * auth_context ,
const char * methods )
{
char * * method_list ;
NTSTATUS status ;
method_list = str_list_make_v3 ( talloc_tos ( ) , methods , NULL ) ;
if ( method_list = = NULL ) {
return NT_STATUS_NO_MEMORY ;
}
status = make_auth_context_text_list (
mem_ctx , auth_context , method_list ) ;
TALLOC_FREE ( method_list ) ;
return status ;
}
2002-01-05 07:55:41 +03:00
/***************************************************************************
Make a auth_context struct for the auth subsystem
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2017-03-17 18:53:27 +03:00
NTSTATUS make_auth3_context_for_ntlm ( TALLOC_CTX * mem_ctx ,
struct auth_context * * auth_context )
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
{
2017-03-13 10:58:43 +03:00
const char * methods = NULL ;
2021-04-13 18:14:01 +03:00
const char * role = NULL ;
This is my 'Authentication Rewrite' version 1.01, mostly as submitted to
samba-technical a few weeks ago.
The idea here is to standardize the checking of user names and passwords,
thereby ensuring that all authtentications pass the same standards. The
interface currently implemented in as
nt_status = check_password(user_info, server_info)
where user_info contains (mostly) the authentication data, and server_info
contains things like the user-id they got, and their resolved user name.
The current ugliness with the way the structures are created will be killed
the next revision, when they will be created and malloced by creator functions.
This patch also includes the first implementation of NTLMv2 in HEAD, but which
needs some more testing. We also add a hack to allow plaintext passwords to be
compared with smbpasswd, not the system password database.
Finally, this patch probably reintroduces the PAM accounts bug we had in
2.2.0, I'll fix that once this hits the tree. (I've just finished testing
it on a wide variety of platforms, so I want to get this patch in).
(This used to be commit b30b6202f31d339b48d51c0d38174cafd1cfcd42)
2001-08-03 17:09:23 +04:00
2017-03-23 14:54:40 +03:00
switch ( lp_server_role ( ) ) {
case ROLE_ACTIVE_DIRECTORY_DC :
2021-04-13 18:14:01 +03:00
role = " 'active directory domain controller' " ;
2017-06-19 11:55:35 +03:00
methods = " samba4 " ;
2017-03-23 14:54:40 +03:00
break ;
2017-03-13 10:43:06 +03:00
case ROLE_DOMAIN_MEMBER :
2021-04-13 18:14:01 +03:00
role = " 'domain member' " ;
2018-03-01 17:04:17 +03:00
methods = " anonymous sam winbind sam_ignoredomain " ;
2017-03-13 10:43:06 +03:00
break ;
case ROLE_DOMAIN_BDC :
case ROLE_DOMAIN_PDC :
2020-11-11 19:50:45 +03:00
case ROLE_IPA_DC :
2021-04-13 18:14:01 +03:00
role = " 'DC' " ;
2018-03-01 17:04:17 +03:00
methods = " anonymous sam winbind sam_ignoredomain " ;
2017-03-13 10:43:06 +03:00
break ;
case ROLE_STANDALONE :
if ( lp_encrypt_passwords ( ) ) {
2021-04-13 18:14:01 +03:00
role = " 'standalone server', encrypt passwords = yes " ;
2018-03-01 17:04:17 +03:00
methods = " anonymous sam_ignoredomain " ;
2017-03-13 10:43:06 +03:00
} else {
2021-04-13 18:14:01 +03:00
role = " 'standalone server', encrypt passwords = no " ;
2018-03-01 17:04:17 +03:00
methods = " anonymous unix " ;
2001-10-31 13:46:25 +03:00
}
2017-03-13 10:43:06 +03:00
break ;
default :
DEBUG ( 5 , ( " Unknown auth method! \n " ) ) ;
return NT_STATUS_UNSUCCESSFUL ;
2002-01-05 07:55:41 +03:00
}
2009-02-21 15:20:08 +03:00
2021-04-13 18:14:01 +03:00
DBG_INFO ( " Making default auth method list for server role = %s \n " ,
role ) ;
2017-03-13 10:58:43 +03:00
return make_auth_context_specific ( mem_ctx , auth_context , methods ) ;
2002-01-05 07:55:41 +03:00
}
2017-03-17 11:13:02 +03:00
NTSTATUS make_auth3_context_for_netlogon ( TALLOC_CTX * mem_ctx ,
struct auth_context * * auth_context )
{
2017-03-17 18:51:45 +03:00
const char * methods = NULL ;
switch ( lp_server_role ( ) ) {
case ROLE_DOMAIN_BDC :
case ROLE_DOMAIN_PDC :
2020-11-11 19:50:45 +03:00
case ROLE_IPA_DC :
2018-12-09 00:48:33 +03:00
methods = " sam_netlogon3 winbind " ;
2017-03-17 18:51:45 +03:00
break ;
default :
DBG_ERR ( " Invalid server role! \n " ) ;
return NT_STATUS_INVALID_SERVER_STATE ;
}
return make_auth_context_specific ( mem_ctx , auth_context , methods ) ;
2017-03-17 11:13:02 +03:00
}
NTSTATUS make_auth3_context_for_winbind ( TALLOC_CTX * mem_ctx ,
struct auth_context * * auth_context )
{
2017-03-17 18:46:38 +03:00
const char * methods = NULL ;
switch ( lp_server_role ( ) ) {
case ROLE_STANDALONE :
case ROLE_DOMAIN_MEMBER :
case ROLE_DOMAIN_BDC :
case ROLE_DOMAIN_PDC :
2020-11-11 19:50:45 +03:00
case ROLE_IPA_DC :
2017-03-17 18:46:38 +03:00
methods = " sam " ;
break ;
case ROLE_ACTIVE_DIRECTORY_DC :
methods = " samba4:sam " ;
break ;
default :
DEBUG ( 5 , ( " Unknown auth method! \n " ) ) ;
return NT_STATUS_UNSUCCESSFUL ;
}
return make_auth_context_specific ( mem_ctx , auth_context , methods ) ;
2017-03-17 11:13:02 +03:00
}
2021-04-21 13:28:12 +03:00
bool auth3_context_set_challenge (
struct auth_context * ctx ,
const uint8_t chal [ 8 ] ,
const char * challenge_set_by )
2017-03-13 10:08:44 +03:00
{
ctx - > challenge = data_blob_talloc ( ctx , chal , 8 ) ;
if ( ctx - > challenge . data = = NULL ) {
return false ;
}
ctx - > challenge_set_by = talloc_strdup ( ctx , challenge_set_by ) ;
if ( ctx - > challenge_set_by = = NULL ) {
return false ;
}
return true ;
}