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
/*
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
Copyright ( C ) Andrew Tridgell 1992 - 2000
Copyright ( C ) Luke Kenneth Casson Leighton 1996 - 2000
Copyright ( C ) Andrew Bartlett 2001
2003-06-30 21:24:59 +04:00
Copyright ( C ) Gerald Carter 2003
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
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 . , 675 Mass Ave , Cambridge , MA 0213 9 , USA .
*/
# include "includes.h"
2002-07-15 14:35:28 +04:00
# undef DBGC_CLASS
# define DBGC_CLASS DBGC_AUTH
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
/****************************************************************************
2003-09-09 00:42:33 +04:00
Core of smb password checking routine .
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
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2003-09-09 00:42:33 +04:00
2003-05-09 19:02:11 +04:00
static BOOL smb_pwd_check_ntlmv1 ( const DATA_BLOB * nt_response ,
2002-01-05 07:55:41 +03:00
const uchar * part_passwd ,
2003-05-09 19:02:11 +04:00
const DATA_BLOB * sec_blob ,
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
DATA_BLOB * user_sess_key )
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
{
2001-10-31 13:46:25 +03:00
/* Finish the encryption of part_passwd. */
uchar p24 [ 24 ] ;
if ( part_passwd = = NULL ) {
DEBUG ( 10 , ( " No password set - DISALLOWING access \n " ) ) ;
/* No password set - always false ! */
return False ;
}
2003-05-09 19:02:11 +04:00
if ( sec_blob - > length ! = 8 ) {
2003-07-25 08:24:40 +04:00
DEBUG ( 0 , ( " smb_pwd_check_ntlmv1: incorrect challenge size (%lu) \n " , ( unsigned long ) sec_blob - > length ) ) ;
2001-10-31 13:46:25 +03:00
return False ;
}
2003-05-09 19:02:11 +04:00
if ( nt_response - > length ! = 24 ) {
2003-07-25 08:24:40 +04:00
DEBUG ( 0 , ( " smb_pwd_check_ntlmv1: incorrect password length (%lu) \n " , ( unsigned long ) nt_response - > length ) ) ;
2001-10-31 13:46:25 +03:00
return False ;
}
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
2003-05-09 19:02:11 +04:00
SMBOWFencrypt ( part_passwd , sec_blob - > data , p24 ) ;
2003-09-09 00:42:33 +04:00
if ( user_sess_key ! = NULL ) {
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
* user_sess_key = data_blob ( NULL , 16 ) ;
SMBsesskeygen_ntv1 ( part_passwd , NULL , user_sess_key - > data ) ;
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
}
2001-10-31 13:46:25 +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
# if DEBUG_PASSWORD
2003-05-09 19:02:11 +04:00
DEBUG ( 100 , ( " Part password (P16) was | \n " ) ) ;
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
dump_data ( 100 , part_passwd , 16 ) ;
2003-05-09 19:02:11 +04:00
DEBUGADD ( 100 , ( " Password from client was | \n " ) ) ;
dump_data ( 100 , nt_response - > data , nt_response - > length ) ;
DEBUGADD ( 100 , ( " Given challenge was | \n " ) ) ;
dump_data ( 100 , sec_blob - > data , sec_blob - > length ) ;
DEBUGADD ( 100 , ( " Value from encryption was | \n " ) ) ;
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
dump_data ( 100 , p24 , 24 ) ;
# endif
2003-05-09 19:02:11 +04:00
return ( memcmp ( p24 , nt_response - > data , 24 ) = = 0 ) ;
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
}
/****************************************************************************
2003-09-09 00:42:33 +04:00
Core of smb password checking routine . ( NTLMv2 , LMv2 )
Note : The same code works with both NTLMv2 and LMv2 .
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
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2003-09-09 00:42:33 +04:00
2003-05-09 19:02:11 +04:00
static BOOL smb_pwd_check_ntlmv2 ( const DATA_BLOB * ntv2_response ,
2002-01-05 07:55:41 +03:00
const uchar * part_passwd ,
2003-05-09 19:02:11 +04:00
const DATA_BLOB * sec_blob ,
2002-01-05 07:55:41 +03:00
const char * user , const char * domain ,
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
DATA_BLOB * user_sess_key )
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
{
/* Finish the encryption of part_passwd. */
uchar kr [ 16 ] ;
2001-10-31 13:46:25 +03:00
uchar value_from_encryption [ 16 ] ;
uchar client_response [ 16 ] ;
DATA_BLOB client_key_data ;
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
2003-09-09 00:42:33 +04:00
if ( part_passwd = = 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
DEBUG ( 10 , ( " No password set - DISALLOWING access \n " ) ) ;
/* No password set - always False */
return False ;
}
2003-05-09 19:02:11 +04:00
if ( ntv2_response - > length < 24 ) {
2001-10-31 13:46:25 +03:00
/* We MUST have more than 16 bytes, or the stuff below will go
2003-05-09 13:41:08 +04:00
crazy . No known implementation sends less than the 24 bytes
for LMv2 , let alone NTLMv2 . */
2003-07-25 08:24:40 +04:00
DEBUG ( 0 , ( " smb_pwd_check_ntlmv2: incorrect password length (%lu) \n " ,
( unsigned long ) ntv2_response - > length ) ) ;
2001-10-31 13:46:25 +03:00
return False ;
}
2003-05-09 19:02:11 +04:00
client_key_data = data_blob ( ntv2_response - > data + 16 , ntv2_response - > length - 16 ) ;
2003-01-05 11:09:16 +03:00
/*
todo : should we be checking this for anything ? We can ' t for LMv2 ,
but for NTLMv2 it is meant to contain the current time etc .
*/
2003-05-09 19:02:11 +04:00
memcpy ( client_response , ntv2_response - > data , sizeof ( client_response ) ) ;
2001-10-31 13:46:25 +03:00
2002-09-25 19:19:00 +04:00
if ( ! ntv2_owf_gen ( part_passwd , user , domain , kr ) ) {
return False ;
}
2003-05-09 19:02:11 +04:00
SMBOWFencrypt_ntv2 ( kr , sec_blob , & client_key_data , value_from_encryption ) ;
2003-09-09 00:42:33 +04:00
if ( user_sess_key ! = NULL ) {
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
* user_sess_key = data_blob ( NULL , 16 ) ;
SMBsesskeygen_ntv2 ( kr , value_from_encryption , user_sess_key - > data ) ;
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
}
# if DEBUG_PASSWORD
2003-05-09 19:02:11 +04:00
DEBUG ( 100 , ( " Part password (P16) was | \n " ) ) ;
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
dump_data ( 100 , part_passwd , 16 ) ;
2003-05-09 19:02:11 +04:00
DEBUGADD ( 100 , ( " Password from client was | \n " ) ) ;
dump_data ( 100 , ntv2_response - > data , ntv2_response - > length ) ;
DEBUGADD ( 100 , ( " Variable data from client was | \n " ) ) ;
2001-11-03 03:20:31 +03:00
dump_data ( 100 , client_key_data . data , client_key_data . length ) ;
2003-05-09 19:02:11 +04:00
DEBUGADD ( 100 , ( " Given challenge was | \n " ) ) ;
dump_data ( 100 , sec_blob - > data , sec_blob - > length ) ;
DEBUGADD ( 100 , ( " Value from encryption was | \n " ) ) ;
2001-10-31 13:46:25 +03:00
dump_data ( 100 , value_from_encryption , 16 ) ;
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
# endif
2001-10-31 13:46:25 +03:00
data_blob_clear_free ( & client_key_data ) ;
return ( memcmp ( value_from_encryption , client_response , 16 ) = = 0 ) ;
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
}
2003-12-30 08:02:32 +03:00
/**
* Check a challenge - response password against the value of the NT or
* LM password hash .
*
* @ param mem_ctx talloc context
* @ param challenge 8 - byte challenge . If all zero , forces plaintext comparison
* @ param nt_response ' unicode ' NT response to the challenge , or unicode password
* @ param lm_response ASCII or LANMAN response to the challenge , or password in DOS code page
* @ param username internal Samba username , for log messages
* @ param client_username username the client used
* @ param client_domain domain name the client used ( may be mapped )
* @ param nt_pw MD4 unicode password from our passdb or similar
* @ param lm_pw LANMAN ASCII password from our passdb or similar
* @ param user_sess_key User session key
* @ param lm_sess_key LM session key ( first 8 bytes of the LM hash )
*/
2003-09-09 00:42:33 +04:00
2003-12-30 08:02:32 +03:00
static NTSTATUS ntlm_password_check ( TALLOC_CTX * mem_ctx ,
const DATA_BLOB * challenge ,
const DATA_BLOB * lm_response ,
const DATA_BLOB * nt_response ,
const char * username ,
const char * client_username ,
const char * client_domain ,
const uint8 * lm_pw , const uint8 * nt_pw ,
DATA_BLOB * user_sess_key ,
DATA_BLOB * lm_sess_key )
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
{
2003-12-30 08:02:32 +03:00
static const unsigned char zeros [ 8 ] ;
if ( nt_pw = = NULL ) {
DEBUG ( 3 , ( " sam_password_ok: NO NT password stored for user %s. \n " ,
username ) ) ;
2001-10-06 05:51:57 +04:00
}
2003-12-30 08:02:32 +03:00
/* Check for cleartext netlogon. Used by Exchange 5.5. */
if ( challenge - > length = = sizeof ( zeros ) & &
( memcmp ( challenge - > data , zeros , challenge - > length ) = = 0 ) ) {
2001-10-31 13:46:25 +03:00
2003-12-30 08:02:32 +03:00
DEBUG ( 4 , ( " sam_password_ok: checking plaintext passwords for user %s \n " ,
username ) ) ;
if ( nt_pw & & nt_response - > length ) {
unsigned char pwhash [ 16 ] ;
mdfour ( pwhash , nt_response - > data , nt_response - > length ) ;
if ( memcmp ( pwhash , nt_pw , sizeof ( pwhash ) ) = = 0 ) {
return NT_STATUS_OK ;
} else {
DEBUG ( 3 , ( " sam_password_ok: NT (Unicode) plaintext password check failed for user %s \n " ,
username ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
2003-10-24 05:18:56 +04:00
}
2003-12-30 08:02:32 +03:00
} else if ( ! lp_lanman_auth ( ) ) {
DEBUG ( 3 , ( " sam_password_ok: (plaintext password check) LANMAN passwords NOT PERMITTED for user %s \n " ,
username ) ) ;
} else if ( lm_pw & & lm_response - > length ) {
uchar dospwd [ 14 ] ;
uchar p16 [ 16 ] ;
ZERO_STRUCT ( dospwd ) ;
DEBUG ( 100 , ( " DOS password: %s \n " ) ) ;
memcpy ( dospwd , lm_response - > data , MIN ( lm_response - > length , sizeof ( dospwd ) ) ) ;
/* Only the fisrt 14 chars are considered, password need not be null terminated. */
E_P16 ( ( const unsigned char * ) dospwd , p16 ) ;
dump_data_pw ( " DOS password (first buffer) \n " , dospwd , 14 ) ;
dump_data_pw ( " DOS password (wire DES hash) \n " , p16 , 16 ) ;
dump_data_pw ( " DOS password (passdb DES hash) \n " , lm_pw , 16 ) ;
if ( memcmp ( p16 , lm_pw , sizeof ( p16 ) ) = = 0 ) {
return NT_STATUS_OK ;
} else {
DEBUG ( 3 , ( " sam_password_ok: LANMAN (ASCII) plaintext password check failed for user %s \n " ,
username ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
}
} else {
DEBUG ( 3 , ( " Plaintext authentication for user %s attempted, but neither NT nor LM passwords available \n " , username ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
2003-10-24 05:18:56 +04:00
}
2001-10-31 13:46:25 +03:00
}
2003-12-30 08:02:32 +03:00
if ( nt_response - > length ! = 0 & & nt_response - > length < 24 ) {
DEBUG ( 2 , ( " sam_password_ok: invalid NT password length (%lu) for user %s \n " ,
( unsigned long ) nt_response - > length , username ) ) ;
}
2001-11-11 14:15:28 +03:00
2003-12-30 08:02:32 +03:00
if ( nt_response - > length > = 24 & & nt_pw ) {
if ( nt_response - > length > 24 ) {
/* We have the NT MD4 hash challenge available - see if we can
use it ( ie . does it exist in the smbpasswd file ) .
*/
DEBUG ( 4 , ( " sam_password_ok: Checking NTLMv2 password with domain [%s] \n " , client_domain ) ) ;
if ( smb_pwd_check_ntlmv2 ( nt_response ,
nt_pw , challenge ,
client_username ,
client_domain ,
user_sess_key ) ) {
return NT_STATUS_OK ;
}
DEBUG ( 4 , ( " sam_password_ok: Checking NTLMv2 password without a domain \n " ) ) ;
if ( smb_pwd_check_ntlmv2 ( nt_response ,
nt_pw , challenge ,
client_username ,
" " ,
user_sess_key ) ) {
return NT_STATUS_OK ;
} else {
DEBUG ( 3 , ( " sam_password_ok: NTLMv2 password check failed \n " ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
}
2003-02-17 01:23:33 +03:00
}
2002-11-09 02:08:59 +03:00
if ( lp_ntlm_auth ( ) ) {
2001-10-06 05:51:57 +04:00
/* We have the NT MD4 hash challenge available - see if we can
use it ( ie . does it exist in the smbpasswd file ) .
*/
2002-01-03 02:28:55 +03:00
DEBUG ( 4 , ( " sam_password_ok: Checking NT MD4 password \n " ) ) ;
2003-12-30 08:02:32 +03:00
if ( smb_pwd_check_ntlmv1 ( nt_response ,
nt_pw , challenge ,
2003-09-09 00:42:33 +04:00
user_sess_key ) ) {
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
/* The LM session key for this response is not very secure,
so use it only if we otherwise allow LM authentication */
if ( lp_lanman_auth ( ) & & lm_pw ) {
uint8 first_8_lm_hash [ 16 ] ;
memcpy ( first_8_lm_hash , lm_pw , 8 ) ;
memset ( first_8_lm_hash + 8 , ' \0 ' , 8 ) ;
* lm_sess_key = data_blob ( first_8_lm_hash , 16 ) ;
}
2001-10-06 05:51:57 +04:00
return NT_STATUS_OK ;
} else {
2003-12-30 08:02:32 +03:00
DEBUG ( 3 , ( " sam_password_ok: NT MD4 password check failed for user %s \n " ,
username ) ) ;
2001-10-31 13:46:25 +03:00
return NT_STATUS_WRONG_PASSWORD ;
2001-10-06 05:51:57 +04:00
}
2001-11-11 14:15:28 +03:00
} else {
2003-12-30 08:02:32 +03:00
DEBUG ( 2 , ( " sam_password_ok: NTLMv1 passwords NOT PERMITTED for user %s \n " ,
username ) ) ;
2003-05-26 10:53:25 +04:00
/* no return, becouse we might pick up LMv2 in the LM field */
2001-11-11 14:15:28 +03:00
}
}
2003-12-30 08:02:32 +03:00
if ( lm_response - > length = = 0 ) {
DEBUG ( 3 , ( " sam_password_ok: NEITHER LanMan nor NT password supplied for user %s \n " ,
username ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
}
if ( lm_response - > length < 24 ) {
DEBUG ( 2 , ( " sam_password_ok: invalid LanMan password length (%lu) for user %s \n " ,
( unsigned long ) nt_response - > length , username ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
}
2001-10-31 13:46:25 +03:00
2003-12-30 08:02:32 +03:00
if ( ! lp_lanman_auth ( ) ) {
DEBUG ( 3 , ( " sam_password_ok: Lanman passwords NOT PERMITTED for user %s \n " ,
username ) ) ;
} else if ( ! lm_pw ) {
DEBUG ( 3 , ( " sam_password_ok: NO LanMan password set for user %s (and no NT password supplied) \n " ,
username ) ) ;
} else {
DEBUG ( 4 , ( " sam_password_ok: Checking LM password \n " ) ) ;
if ( smb_pwd_check_ntlmv1 ( lm_response ,
lm_pw , challenge ,
NULL ) ) {
uint8 first_8_lm_hash [ 16 ] ;
memcpy ( first_8_lm_hash , lm_pw , 8 ) ;
memset ( first_8_lm_hash + 8 , ' \0 ' , 8 ) ;
* user_sess_key = data_blob ( first_8_lm_hash , 16 ) ;
* lm_sess_key = data_blob ( first_8_lm_hash , 16 ) ;
return NT_STATUS_OK ;
}
}
if ( ! nt_pw ) {
DEBUG ( 4 , ( " sam_password_ok: LM password check failed for user, no NT password %s \n " , username ) ) ;
return NT_STATUS_WRONG_PASSWORD ;
}
/* This is for 'LMv2' authentication. almost NTLMv2 but limited to 24 bytes.
- related to Win9X , legacy NAS pass - though authentication
*/
DEBUG ( 4 , ( " sam_password_ok: Checking LMv2 password with domain %s \n " , client_domain ) ) ;
if ( smb_pwd_check_ntlmv2 ( lm_response ,
nt_pw , challenge ,
client_username ,
client_domain ,
NULL ) ) {
return NT_STATUS_OK ;
}
DEBUG ( 4 , ( " sam_password_ok: Checking LMv2 password without a domain \n " ) ) ;
if ( smb_pwd_check_ntlmv2 ( lm_response ,
nt_pw , challenge ,
client_username ,
" " ,
NULL ) ) {
return NT_STATUS_OK ;
}
/* Apparently NT accepts NT responses in the LM field
- I think this is related to Win9X pass - though authentication
*/
DEBUG ( 4 , ( " sam_password_ok: Checking NT MD4 password in LM field \n " ) ) ;
if ( lp_ntlm_auth ( ) ) {
if ( smb_pwd_check_ntlmv1 ( lm_response ,
nt_pw , challenge ,
NULL ) ) {
/* The session key for this response is still very odd.
It not very secure , so use it only if we otherwise
allow LM authentication */
if ( lp_lanman_auth ( ) & & lm_pw ) {
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
uint8 first_8_lm_hash [ 16 ] ;
memcpy ( first_8_lm_hash , lm_pw , 8 ) ;
memset ( first_8_lm_hash + 8 , ' \0 ' , 8 ) ;
* user_sess_key = data_blob ( first_8_lm_hash , 16 ) ;
* lm_sess_key = data_blob ( first_8_lm_hash , 16 ) ;
2003-01-05 11:09:16 +03:00
}
2001-10-31 13:46:25 +03:00
return NT_STATUS_OK ;
2003-01-05 11:09:16 +03:00
}
2003-12-30 08:02:32 +03:00
DEBUG ( 3 , ( " sam_password_ok: LM password, NT MD4 password in LM field and LMv2 failed for user %s \n " , username ) ) ;
} else {
DEBUG ( 3 , ( " sam_password_ok: LM password and LMv2 failed for user %s, and NT MD4 password in LM field not permitted \n " , username ) ) ;
}
return NT_STATUS_WRONG_PASSWORD ;
}
2003-01-05 11:09:16 +03:00
2003-12-30 08:02:32 +03:00
/****************************************************************************
Do a specific test for an smb password being correct , given a smb_password and
the lanman and NT responses .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2003-02-17 01:23:33 +03:00
2003-12-30 08:02:32 +03:00
static NTSTATUS sam_password_ok ( const struct auth_context * auth_context ,
TALLOC_CTX * mem_ctx ,
SAM_ACCOUNT * sampass ,
const auth_usersupplied_info * user_info ,
DATA_BLOB * user_sess_key ,
DATA_BLOB * lm_sess_key )
{
uint16 acct_ctrl ;
const uint8 * lm_pw , * nt_pw ;
const char * username = pdb_get_username ( sampass ) ;
acct_ctrl = pdb_get_acct_ctrl ( sampass ) ;
if ( acct_ctrl & ACB_PWNOTREQ ) {
if ( lp_null_passwords ( ) ) {
DEBUG ( 3 , ( " Account for user '%s' has no password and null passwords are allowed. \n " , username ) ) ;
return NT_STATUS_OK ;
2003-01-05 11:09:16 +03:00
} else {
2003-12-30 08:02:32 +03:00
DEBUG ( 3 , ( " Account for user '%s' has no password and null passwords are NOT allowed. \n " , username ) ) ;
return NT_STATUS_LOGON_FAILURE ;
}
2001-09-26 16:28:46 +04:00
}
2003-12-30 08:02:32 +03:00
lm_pw = pdb_get_lanman_passwd ( sampass ) ;
nt_pw = pdb_get_nt_passwd ( sampass ) ;
return ntlm_password_check ( mem_ctx , & auth_context - > challenge ,
& user_info - > lm_resp , & user_info - > nt_resp ,
username ,
user_info - > smb_name . str ,
user_info - > client_domain . str ,
lm_pw , nt_pw , user_sess_key , lm_sess_key ) ;
2001-09-26 16:28:46 +04:00
}
2003-12-30 08:02:32 +03:00
2001-09-26 16:28:46 +04:00
/****************************************************************************
Do a specific test for a SAM_ACCOUNT being vaild for this connection
( ie not disabled , expired and the like ) .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2003-09-09 00:42:33 +04:00
2002-01-01 06:10:32 +03:00
static NTSTATUS sam_account_ok ( TALLOC_CTX * mem_ctx ,
SAM_ACCOUNT * sampass ,
const auth_usersupplied_info * user_info )
2001-09-26 16:28:46 +04:00
{
uint16 acct_ctrl = pdb_get_acct_ctrl ( sampass ) ;
2001-09-16 10:35:35 +04:00
char * workstation_list ;
time_t kickoff_time ;
2001-08-17 10:06:56 +04:00
2002-01-03 02:28:55 +03:00
DEBUG ( 4 , ( " sam_account_ok: Checking SMB password for user %s \n " , pdb_get_username ( sampass ) ) ) ;
2001-09-26 16:28:46 +04:00
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
/* Quit if the account was disabled. */
2001-10-01 14:54:11 +04:00
if ( acct_ctrl & ACB_DISABLED ) {
2003-09-09 00:42:33 +04:00
DEBUG ( 1 , ( " sam_account_ok: Account for user '%s' was disabled. \n " , pdb_get_username ( sampass ) ) ) ;
2001-09-26 16:28:46 +04:00
return NT_STATUS_ACCOUNT_DISABLED ;
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
}
2003-09-09 00:42:33 +04:00
/* Quit if the account was locked out. */
if ( acct_ctrl & ACB_AUTOLOCK ) {
DEBUG ( 1 , ( " sam_account_ok: Account for user %s was locked out. \n " , pdb_get_username ( sampass ) ) ) ;
return NT_STATUS_ACCOUNT_LOCKED_OUT ;
}
2001-09-16 10:35:35 +04:00
/* Test account expire time */
kickoff_time = pdb_get_kickoff_time ( sampass ) ;
2001-10-01 14:54:11 +04:00
if ( kickoff_time ! = 0 & & time ( NULL ) > kickoff_time ) {
2003-09-09 00:42:33 +04:00
DEBUG ( 1 , ( " sam_account_ok: Account for user '%s' has expired. \n " , pdb_get_username ( sampass ) ) ) ;
DEBUG ( 3 , ( " sam_account_ok: Account expired at '%ld' unix time. \n " , ( long ) kickoff_time ) ) ;
2001-10-01 14:54:11 +04:00
return NT_STATUS_ACCOUNT_EXPIRED ;
2001-09-16 10:35:35 +04:00
}
2001-09-26 17:11:22 +04:00
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 ( ! ( pdb_get_acct_ctrl ( sampass ) & ACB_PWNOEXP ) ) {
time_t must_change_time = pdb_get_pass_must_change_time ( sampass ) ;
time_t last_set_time = pdb_get_pass_last_set_time ( sampass ) ;
/* check for immediate expiry "must change at next logon" */
if ( must_change_time = = 0 & & last_set_time ! = 0 ) {
2003-09-09 00:42:33 +04:00
DEBUG ( 1 , ( " sam_account_ok: Account for user '%s' password must change!. \n " , pdb_get_username ( sampass ) ) ) ;
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_PASSWORD_MUST_CHANGE ;
}
/* check for expired password */
if ( must_change_time < time ( NULL ) & & must_change_time ! = 0 ) {
2003-09-09 00:42:33 +04:00
DEBUG ( 1 , ( " sam_account_ok: Account for user '%s' password expired!. \n " , pdb_get_username ( sampass ) ) ) ;
DEBUG ( 1 , ( " sam_account_ok: Password expired at '%s' (%ld) unix time. \n " , http_timestring ( must_change_time ) , ( long ) must_change_time ) ) ;
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_PASSWORD_EXPIRED ;
}
}
2001-09-16 10:35:35 +04:00
/* Test workstation. Workstation list is comma separated. */
2001-09-26 17:11:22 +04:00
2002-01-01 06:10:32 +03:00
workstation_list = talloc_strdup ( mem_ctx , pdb_get_workstations ( sampass ) ) ;
2003-09-09 00:42:33 +04:00
if ( ! workstation_list )
return NT_STATUS_NO_MEMORY ;
2001-10-01 14:54:11 +04:00
if ( * workstation_list ) {
BOOL invalid_ws = True ;
2002-11-13 02:20:50 +03:00
const char * s = workstation_list ;
2001-09-16 10:35:35 +04:00
2001-10-01 14:54:11 +04:00
fstring tok ;
2001-09-16 10:35:35 +04:00
2001-10-01 14:54:11 +04:00
while ( next_token ( & s , tok , " , " , sizeof ( tok ) ) ) {
2003-09-09 00:42:33 +04:00
DEBUG ( 10 , ( " sam_account_ok: checking for workstation match %s and %s (len=%d) \n " ,
2001-10-01 14:54:11 +04:00
tok , user_info - > wksta_name . str , user_info - > wksta_name . len ) ) ;
if ( strequal ( tok , user_info - > wksta_name . str ) ) {
invalid_ws = False ;
break ;
2001-09-16 10:35:35 +04:00
}
}
2001-10-01 14:54:11 +04:00
if ( invalid_ws )
return NT_STATUS_INVALID_WORKSTATION ;
2001-09-16 10:35:35 +04:00
}
2001-10-01 14:54:11 +04:00
2001-09-26 17:11:22 +04:00
if ( acct_ctrl & ACB_DOMTRUST ) {
2002-01-03 02:28:55 +03:00
DEBUG ( 2 , ( " sam_account_ok: Domain trust account %s denied by server \n " , pdb_get_username ( sampass ) ) ) ;
2001-09-26 17:11:22 +04:00
return NT_STATUS_NOLOGON_INTERDOMAIN_TRUST_ACCOUNT ;
}
if ( acct_ctrl & ACB_SVRTRUST ) {
2002-01-03 02:28:55 +03:00
DEBUG ( 2 , ( " sam_account_ok: Server trust account %s denied by server \n " , pdb_get_username ( sampass ) ) ) ;
2001-09-26 17:11:22 +04:00
return NT_STATUS_NOLOGON_SERVER_TRUST_ACCOUNT ;
}
if ( acct_ctrl & ACB_WSTRUST ) {
2002-01-03 02:28:55 +03:00
DEBUG ( 4 , ( " sam_account_ok: Wksta trust account %s denied by server \n " , pdb_get_username ( sampass ) ) ) ;
2001-09-26 17:11:22 +04:00
return NT_STATUS_NOLOGON_WORKSTATION_TRUST_ACCOUNT ;
}
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
2001-09-26 16:28:46 +04:00
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
}
/****************************************************************************
check if a username / password is OK assuming the password is a 24 byte
2001-08-17 10:06:56 +04:00
SMB hash supplied in the user_info structure
return an NT_STATUS constant .
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
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-01-05 07:55:41 +03:00
static NTSTATUS check_sam_security ( const struct auth_context * auth_context ,
void * my_private_data ,
2002-01-01 06:10:32 +03:00
TALLOC_CTX * mem_ctx ,
const auth_usersupplied_info * user_info ,
auth_serversupplied_info * * server_info )
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
{
SAM_ACCOUNT * sampass = NULL ;
BOOL ret ;
2001-09-04 11:13:01 +04:00
NTSTATUS nt_status ;
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
DATA_BLOB user_sess_key = data_blob ( NULL , 0 ) ;
DATA_BLOB lm_sess_key = data_blob ( NULL , 0 ) ;
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
2002-01-05 07:55:41 +03:00
if ( ! user_info | | ! auth_context ) {
return NT_STATUS_UNSUCCESSFUL ;
2001-11-11 14:18:45 +03:00
}
2003-05-14 04:46:43 +04:00
/* Can't use the talloc version here, because the returned struct gets
2002-01-15 04:14:58 +03:00
kept on the server_info */
if ( ! NT_STATUS_IS_OK ( nt_status = pdb_init_sam ( & sampass ) ) ) {
return nt_status ;
2001-11-11 14:18:45 +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
/* get the account information */
2001-08-12 15:19:57 +04:00
become_root ( ) ;
2001-10-31 13:46:25 +03:00
ret = pdb_getsampwnam ( sampass , user_info - > internal_username . str ) ;
2001-08-12 15:19:57 +04:00
unbecome_root ( ) ;
2003-09-09 00:42:33 +04:00
if ( ret = = False ) {
DEBUG ( 3 , ( " check_sam_security: Couldn't find user '%s' in passdb file. \n " , user_info - > internal_username . str ) ) ;
2001-09-29 17:08:26 +04:00
pdb_free_sam ( & sampass ) ;
2001-09-04 11:13:01 +04:00
return NT_STATUS_NO_SUCH_USER ;
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
}
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
nt_status = sam_password_ok ( auth_context , mem_ctx , sampass ,
user_info , & user_sess_key , & lm_sess_key ) ;
2002-11-16 00:43:57 +03:00
2001-10-31 13:46:25 +03:00
if ( ! NT_STATUS_IS_OK ( nt_status ) ) {
pdb_free_sam ( & sampass ) ;
return nt_status ;
2001-09-26 16:28:46 +04:00
}
2003-05-11 17:20:27 +04:00
nt_status = sam_account_ok ( mem_ctx , sampass , user_info ) ;
2003-02-04 13:08:45 +03:00
2001-10-31 13:46:25 +03:00
if ( ! NT_STATUS_IS_OK ( nt_status ) ) {
pdb_free_sam ( & sampass ) ;
return nt_status ;
}
2002-09-25 19:19:00 +04:00
if ( ! NT_STATUS_IS_OK ( nt_status = make_server_info_sam ( server_info , sampass ) ) ) {
DEBUG ( 0 , ( " check_sam_security: make_server_info_sam() failed with '%s' \n " , nt_errstr ( nt_status ) ) ) ;
return nt_status ;
2001-10-31 13:46:25 +03:00
}
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
( * server_info ) - > nt_session_key = user_sess_key ;
( * server_info ) - > lm_session_key = lm_sess_key ;
2001-10-31 13:46:25 +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
return nt_status ;
}
2001-10-31 13:46:25 +03:00
2002-01-05 07:55:41 +03:00
/* module initialisation */
2003-06-30 21:24:59 +04:00
static NTSTATUS auth_init_sam_ignoredomain ( struct auth_context * auth_context , const char * param , auth_methods * * auth_method )
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 ( ! make_auth_methods ( auth_context , auth_method ) ) {
2002-07-15 14:35:28 +04:00
return NT_STATUS_NO_MEMORY ;
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
}
2002-07-15 14:35:28 +04:00
( * auth_method ) - > auth = check_sam_security ;
2003-06-30 21:24:59 +04:00
( * auth_method ) - > name = " sam_ignoredomain " ;
2002-07-15 14:35:28 +04:00
return NT_STATUS_OK ;
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
}
2001-10-31 13:46:25 +03:00
2002-01-05 07:55:41 +03:00
2001-11-26 09:47:04 +03:00
/****************************************************************************
2002-01-05 07:55:41 +03:00
Check SAM security ( above ) but with a few extra checks .
2001-11-26 09:47:04 +03:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-01-05 07:55:41 +03:00
static NTSTATUS check_samstrict_security ( const struct auth_context * auth_context ,
void * my_private_data ,
2002-01-01 06:10:32 +03:00
TALLOC_CTX * mem_ctx ,
const auth_usersupplied_info * user_info ,
auth_serversupplied_info * * server_info )
2001-11-26 09:47:04 +03:00
{
2003-06-30 21:24:59 +04:00
BOOL is_local_name , is_my_domain ;
2001-11-26 09:47:04 +03:00
2002-01-05 07:55:41 +03:00
if ( ! user_info | | ! auth_context ) {
2001-11-26 09:47:04 +03:00
return NT_STATUS_LOGON_FAILURE ;
}
2003-06-30 21:24:59 +04:00
is_local_name = is_myname ( user_info - > domain . str ) ;
is_my_domain = strequal ( user_info - > domain . str , lp_workgroup ( ) ) ;
/* check whether or not we service this domain/workgroup name */
2001-11-26 09:47:04 +03:00
2003-06-30 21:24:59 +04:00
switch ( lp_server_role ( ) ) {
case ROLE_STANDALONE :
case ROLE_DOMAIN_MEMBER :
if ( ! is_local_name ) {
DEBUG ( 6 , ( " check_samstrict_security: %s is not one of my local names (%s) \n " ,
user_info - > domain . str , ( lp_server_role ( ) = = ROLE_DOMAIN_MEMBER
? " ROLE_DOMAIN_MEMBER " : " ROLE_STANDALONE " ) ) ) ;
return NT_STATUS_NOT_IMPLEMENTED ;
}
case ROLE_DOMAIN_PDC :
case ROLE_DOMAIN_BDC :
if ( ! is_local_name & & ! is_my_domain ) {
DEBUG ( 6 , ( " check_samstrict_security: %s is not one of my local names or domain name (DC) \n " ,
user_info - > domain . str ) ) ;
return NT_STATUS_NOT_IMPLEMENTED ;
}
default : /* name is ok */
break ;
2001-11-26 09:47:04 +03:00
}
2002-01-05 07:55:41 +03:00
return check_sam_security ( auth_context , my_private_data , mem_ctx , user_info , server_info ) ;
2001-11-26 09:47:04 +03:00
}
2002-01-05 07:55:41 +03:00
/* module initialisation */
2003-06-30 21:24:59 +04:00
static NTSTATUS auth_init_sam ( struct auth_context * auth_context , const char * param , auth_methods * * auth_method )
2001-11-26 09:47:04 +03:00
{
2002-01-05 07:55:41 +03:00
if ( ! make_auth_methods ( auth_context , auth_method ) ) {
2002-07-15 14:35:28 +04:00
return NT_STATUS_NO_MEMORY ;
2001-11-26 09:47:04 +03:00
}
2001-10-31 13:46:25 +03:00
2001-11-26 09:47:04 +03:00
( * auth_method ) - > auth = check_samstrict_security ;
2003-06-30 21:24:59 +04:00
( * auth_method ) - > name = " sam " ;
2003-05-26 06:04:23 +04:00
return NT_STATUS_OK ;
}
2003-04-28 21:48:48 +04:00
NTSTATUS auth_sam_init ( void )
2003-04-16 16:13:07 +04:00
{
2003-04-28 21:48:48 +04:00
smb_register_auth ( AUTH_INTERFACE_VERSION , " sam " , auth_init_sam ) ;
2003-06-30 21:24:59 +04:00
smb_register_auth ( AUTH_INTERFACE_VERSION , " sam_ignoredomain " , auth_init_sam_ignoredomain ) ;
2003-04-28 21:48:48 +04:00
return NT_STATUS_OK ;
2003-04-16 16:13:07 +04:00
}