1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-25 23:21:54 +03:00
samba-mirror/source3/auth/auth_util.c

1653 lines
48 KiB
C
Raw Normal View History

/*
Unix SMB/CIFS implementation.
Authentication utility functions
Copyright (C) Andrew Tridgell 1992-1998
Copyright (C) Andrew Bartlett 2001
Copyright (C) Jeremy Allison 2000-2001
Copyright (C) Rafal Szczesniak 2002
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 02139, USA.
*/
#include "includes.h"
#undef DBGC_CLASS
#define DBGC_CLASS DBGC_AUTH
/****************************************************************************
Create a UNIX user on demand.
****************************************************************************/
static int smb_create_user(const char *domain, const char *unix_username, const char *homedir)
{
pstring add_script;
int ret;
pstrcpy(add_script, lp_adduser_script());
if (! *add_script)
return -1;
all_string_sub(add_script, "%u", unix_username, sizeof(pstring));
if (domain)
all_string_sub(add_script, "%D", domain, sizeof(pstring));
if (homedir)
all_string_sub(add_script, "%H", homedir, sizeof(pstring));
ret = smbrun(add_script,NULL);
flush_pwnam_cache();
DEBUG(ret ? 0 : 3,("smb_create_user: Running the command `%s' gave %d\n",add_script,ret));
return ret;
}
/****************************************************************************
Create an auth_usersupplied_data structure
****************************************************************************/
static NTSTATUS make_user_info(auth_usersupplied_info **user_info,
const char *smb_name,
const char *internal_username,
const char *client_domain,
const char *domain,
const char *wksta_name,
DATA_BLOB *lm_pwd, DATA_BLOB *nt_pwd,
DATA_BLOB *lm_interactive_pwd, DATA_BLOB *nt_interactive_pwd,
DATA_BLOB *plaintext,
BOOL encrypted)
{
DEBUG(5,("attempting to make a user_info for %s (%s)\n", internal_username, smb_name));
*user_info = SMB_MALLOC_P(auth_usersupplied_info);
if (!user_info) {
DEBUG(0,("malloc failed for user_info (size %lu)\n", (unsigned long)sizeof(*user_info)));
return NT_STATUS_NO_MEMORY;
}
ZERO_STRUCTP(*user_info);
DEBUG(5,("making strings for %s's user_info struct\n", internal_username));
(*user_info)->smb_name.str = SMB_STRDUP(smb_name);
if ((*user_info)->smb_name.str) {
(*user_info)->smb_name.len = strlen(smb_name);
} else {
free_user_info(user_info);
return NT_STATUS_NO_MEMORY;
}
(*user_info)->internal_username.str = SMB_STRDUP(internal_username);
if ((*user_info)->internal_username.str) {
(*user_info)->internal_username.len = strlen(internal_username);
} else {
free_user_info(user_info);
return NT_STATUS_NO_MEMORY;
}
(*user_info)->domain.str = SMB_STRDUP(domain);
if ((*user_info)->domain.str) {
(*user_info)->domain.len = strlen(domain);
} else {
free_user_info(user_info);
return NT_STATUS_NO_MEMORY;
}
(*user_info)->client_domain.str = SMB_STRDUP(client_domain);
if ((*user_info)->client_domain.str) {
(*user_info)->client_domain.len = strlen(client_domain);
} else {
free_user_info(user_info);
return NT_STATUS_NO_MEMORY;
}
(*user_info)->wksta_name.str = SMB_STRDUP(wksta_name);
if ((*user_info)->wksta_name.str) {
(*user_info)->wksta_name.len = strlen(wksta_name);
} else {
free_user_info(user_info);
return NT_STATUS_NO_MEMORY;
}
DEBUG(5,("making blobs for %s's user_info struct\n", internal_username));
if (lm_pwd)
(*user_info)->lm_resp = data_blob(lm_pwd->data, lm_pwd->length);
if (nt_pwd)
(*user_info)->nt_resp = data_blob(nt_pwd->data, nt_pwd->length);
if (lm_interactive_pwd)
(*user_info)->lm_interactive_pwd = data_blob(lm_interactive_pwd->data, lm_interactive_pwd->length);
if (nt_interactive_pwd)
(*user_info)->nt_interactive_pwd = data_blob(nt_interactive_pwd->data, nt_interactive_pwd->length);
if (plaintext)
(*user_info)->plaintext_password = data_blob(plaintext->data, plaintext->length);
(*user_info)->encrypted = encrypted;
(*user_info)->logon_parameters = 0;
DEBUG(10,("made an %sencrypted user_info for %s (%s)\n", encrypted ? "":"un" , internal_username, smb_name));
return NT_STATUS_OK;
}
/****************************************************************************
Create an auth_usersupplied_data structure after appropriate mapping.
****************************************************************************/
NTSTATUS make_user_info_map(auth_usersupplied_info **user_info,
const char *smb_name,
const char *client_domain,
const char *wksta_name,
DATA_BLOB *lm_pwd, DATA_BLOB *nt_pwd,
DATA_BLOB *lm_interactive_pwd, DATA_BLOB *nt_interactive_pwd,
DATA_BLOB *plaintext,
BOOL encrypted)
{
const char *domain;
fstring internal_username;
fstrcpy(internal_username, smb_name);
map_username(internal_username);
DEBUG(5, ("make_user_info_map: Mapping user [%s]\\[%s] from workstation [%s]\n",
client_domain, smb_name, wksta_name));
/* don't allow "" as a domain, fixes a Win9X bug
where it doens't supply a domain for logon script
'net use' commands. */
if ( *client_domain )
domain = client_domain;
else
domain = lp_workgroup();
/* do what win2k does. Always map unknown domains to our own
and let the "passdb backend" handle unknown users. */
if ( !is_trusted_domain(domain) && !strequal(domain, get_global_sam_name()) )
domain = get_default_sam_name();
/* we know that it is a trusted domain (and we are allowing them) or it is our domain */
return make_user_info(user_info, smb_name, internal_username,
client_domain, domain, wksta_name,
lm_pwd, nt_pwd,
lm_interactive_pwd, nt_interactive_pwd,
plaintext, encrypted);
}
/****************************************************************************
Create an auth_usersupplied_data, making the DATA_BLOBs here.
Decrypt and encrypt the passwords.
****************************************************************************/
BOOL make_user_info_netlogon_network(auth_usersupplied_info **user_info,
const char *smb_name,
const char *client_domain,
const char *wksta_name,
uint32 logon_parameters,
const uchar *lm_network_pwd, int lm_pwd_len,
const uchar *nt_network_pwd, int nt_pwd_len)
{
BOOL ret;
NTSTATUS nt_status;
DATA_BLOB lm_blob = data_blob(lm_network_pwd, lm_pwd_len);
DATA_BLOB nt_blob = data_blob(nt_network_pwd, nt_pwd_len);
nt_status = make_user_info_map(user_info,
smb_name, client_domain,
wksta_name,
lm_pwd_len ? &lm_blob : NULL,
nt_pwd_len ? &nt_blob : NULL,
NULL, NULL, NULL,
True);
if (NT_STATUS_IS_OK(nt_status)) {
(*user_info)->logon_parameters = logon_parameters;
}
ret = NT_STATUS_IS_OK(nt_status) ? True : False;
data_blob_free(&lm_blob);
data_blob_free(&nt_blob);
return ret;
}
/****************************************************************************
Create an auth_usersupplied_data, making the DATA_BLOBs here.
Decrypt and encrypt the passwords.
****************************************************************************/
BOOL make_user_info_netlogon_interactive(auth_usersupplied_info **user_info,
const char *smb_name,
const char *client_domain,
const char *wksta_name,
uint32 logon_parameters,
const uchar chal[8],
const uchar lm_interactive_pwd[16],
const uchar nt_interactive_pwd[16],
const uchar *dc_sess_key)
{
char lm_pwd[16];
char nt_pwd[16];
unsigned char local_lm_response[24];
unsigned char local_nt_response[24];
unsigned char key[16];
ZERO_STRUCT(key);
memcpy(key, dc_sess_key, 8);
if (lm_interactive_pwd) memcpy(lm_pwd, lm_interactive_pwd, sizeof(lm_pwd));
if (nt_interactive_pwd) memcpy(nt_pwd, nt_interactive_pwd, sizeof(nt_pwd));
#ifdef DEBUG_PASSWORD
DEBUG(100,("key:"));
dump_data(100, (char *)key, sizeof(key));
DEBUG(100,("lm owf password:"));
dump_data(100, lm_pwd, sizeof(lm_pwd));
DEBUG(100,("nt owf password:"));
dump_data(100, nt_pwd, sizeof(nt_pwd));
#endif
if (lm_interactive_pwd)
SamOEMhash((uchar *)lm_pwd, key, sizeof(lm_pwd));
if (nt_interactive_pwd)
SamOEMhash((uchar *)nt_pwd, key, sizeof(nt_pwd));
#ifdef DEBUG_PASSWORD
DEBUG(100,("decrypt of lm owf password:"));
dump_data(100, lm_pwd, sizeof(lm_pwd));
DEBUG(100,("decrypt of nt owf password:"));
dump_data(100, nt_pwd, sizeof(nt_pwd));
#endif
if (lm_interactive_pwd)
SMBOWFencrypt((const unsigned char *)lm_pwd, chal, local_lm_response);
if (nt_interactive_pwd)
SMBOWFencrypt((const unsigned char *)nt_pwd, chal, local_nt_response);
/* Password info paranoia */
ZERO_STRUCT(key);
{
BOOL ret;
NTSTATUS nt_status;
DATA_BLOB local_lm_blob;
DATA_BLOB local_nt_blob;
DATA_BLOB lm_interactive_blob;
DATA_BLOB nt_interactive_blob;
if (lm_interactive_pwd) {
local_lm_blob = data_blob(local_lm_response, sizeof(local_lm_response));
lm_interactive_blob = data_blob(lm_pwd, sizeof(lm_pwd));
ZERO_STRUCT(lm_pwd);
}
if (nt_interactive_pwd) {
local_nt_blob = data_blob(local_nt_response, sizeof(local_nt_response));
nt_interactive_blob = data_blob(nt_pwd, sizeof(nt_pwd));
ZERO_STRUCT(nt_pwd);
}
nt_status = make_user_info_map(user_info,
smb_name, client_domain,
wksta_name,
lm_interactive_pwd ? &local_lm_blob : NULL,
nt_interactive_pwd ? &local_nt_blob : NULL,
lm_interactive_pwd ? &lm_interactive_blob : NULL,
nt_interactive_pwd ? &nt_interactive_blob : NULL,
NULL,
True);
if (NT_STATUS_IS_OK(nt_status)) {
(*user_info)->logon_parameters = logon_parameters;
}
ret = NT_STATUS_IS_OK(nt_status) ? True : False;
data_blob_free(&local_lm_blob);
data_blob_free(&local_nt_blob);
data_blob_free(&lm_interactive_blob);
data_blob_free(&nt_interactive_blob);
return ret;
}
}
/****************************************************************************
Create an auth_usersupplied_data structure
****************************************************************************/
BOOL make_user_info_for_reply(auth_usersupplied_info **user_info,
const char *smb_name,
const char *client_domain,
const uint8 chal[8],
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
DATA_BLOB plaintext_password)
{
DATA_BLOB local_lm_blob;
DATA_BLOB local_nt_blob;
NTSTATUS ret = NT_STATUS_UNSUCCESSFUL;
/*
* Not encrypted - do so.
*/
This is another rather major change to the samba authenticaion subystem. The particular aim is to modularized the interface - so that we can have arbitrary password back-ends. This code adds one such back-end, a 'winbind' module to authenticate against the winbind_auth_crap functionality. While fully-functional this code is mainly useful as a demonstration, because we don't get back the info3 as we would for direct ntdomain authentication. This commit introduced the new 'auth methods' parameter, in the spirit of the 'auth order' discussed on the lists. It is renamed because not all the methods may be consulted, even if previous methods fail - they may not have a suitable challenge for example. Also, we have a 'local' authentication method, for old-style 'unix if plaintext, sam if encrypted' authentication and a 'guest' module to handle guest logins in a single place. While this current design is not ideal, I feel that it does provide a better infrastructure than the current design, and can be built upon. The following parameters have changed: - use rhosts = This has been replaced by the 'rhosts' authentication method, and can be specified like 'auth methods = guest rhosts' - hosts equiv = This needs both this parameter and an 'auth methods' entry to be effective. (auth methods = guest hostsequiv ....) - plaintext to smbpasswd = This is replaced by specifying 'sam' rather than 'local' in the auth methods. The security = parameter is unchanged, and now provides defaults for the 'auth methods' parameter. The available auth methods are: guest rhosts hostsequiv sam (passdb direct hash access) unix (PAM, crypt() etc) local (the combination of the above, based on encryption) smbserver (old security=server) ntdomain (old security=domain) winbind (use winbind to cache DC connections) Assistance in testing, or the production of new and interesting authentication modules is always appreciated. Andrew Bartlett (This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
DEBUG(5,("make_user_info_for_reply: User passwords not in encrypted format.\n"));
if (plaintext_password.data) {
unsigned char local_lm_response[24];
This is another rather major change to the samba authenticaion subystem. The particular aim is to modularized the interface - so that we can have arbitrary password back-ends. This code adds one such back-end, a 'winbind' module to authenticate against the winbind_auth_crap functionality. While fully-functional this code is mainly useful as a demonstration, because we don't get back the info3 as we would for direct ntdomain authentication. This commit introduced the new 'auth methods' parameter, in the spirit of the 'auth order' discussed on the lists. It is renamed because not all the methods may be consulted, even if previous methods fail - they may not have a suitable challenge for example. Also, we have a 'local' authentication method, for old-style 'unix if plaintext, sam if encrypted' authentication and a 'guest' module to handle guest logins in a single place. While this current design is not ideal, I feel that it does provide a better infrastructure than the current design, and can be built upon. The following parameters have changed: - use rhosts = This has been replaced by the 'rhosts' authentication method, and can be specified like 'auth methods = guest rhosts' - hosts equiv = This needs both this parameter and an 'auth methods' entry to be effective. (auth methods = guest hostsequiv ....) - plaintext to smbpasswd = This is replaced by specifying 'sam' rather than 'local' in the auth methods. The security = parameter is unchanged, and now provides defaults for the 'auth methods' parameter. The available auth methods are: guest rhosts hostsequiv sam (passdb direct hash access) unix (PAM, crypt() etc) local (the combination of the above, based on encryption) smbserver (old security=server) ntdomain (old security=domain) winbind (use winbind to cache DC connections) Assistance in testing, or the production of new and interesting authentication modules is always appreciated. Andrew Bartlett (This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
#ifdef DEBUG_PASSWORD
DEBUG(10,("Unencrypted password (len %d):\n",(int)plaintext_password.length));
dump_data(100, (const char *)plaintext_password.data, plaintext_password.length);
This is another rather major change to the samba authenticaion subystem. The particular aim is to modularized the interface - so that we can have arbitrary password back-ends. This code adds one such back-end, a 'winbind' module to authenticate against the winbind_auth_crap functionality. While fully-functional this code is mainly useful as a demonstration, because we don't get back the info3 as we would for direct ntdomain authentication. This commit introduced the new 'auth methods' parameter, in the spirit of the 'auth order' discussed on the lists. It is renamed because not all the methods may be consulted, even if previous methods fail - they may not have a suitable challenge for example. Also, we have a 'local' authentication method, for old-style 'unix if plaintext, sam if encrypted' authentication and a 'guest' module to handle guest logins in a single place. While this current design is not ideal, I feel that it does provide a better infrastructure than the current design, and can be built upon. The following parameters have changed: - use rhosts = This has been replaced by the 'rhosts' authentication method, and can be specified like 'auth methods = guest rhosts' - hosts equiv = This needs both this parameter and an 'auth methods' entry to be effective. (auth methods = guest hostsequiv ....) - plaintext to smbpasswd = This is replaced by specifying 'sam' rather than 'local' in the auth methods. The security = parameter is unchanged, and now provides defaults for the 'auth methods' parameter. The available auth methods are: guest rhosts hostsequiv sam (passdb direct hash access) unix (PAM, crypt() etc) local (the combination of the above, based on encryption) smbserver (old security=server) ntdomain (old security=domain) winbind (use winbind to cache DC connections) Assistance in testing, or the production of new and interesting authentication modules is always appreciated. Andrew Bartlett (This used to be commit 8d31eae52a9757739711dbb82035a4dfe6b40c99)
2001-11-24 15:12:38 +03:00
#endif
SMBencrypt( (const char *)plaintext_password.data, (const uchar*)chal, local_lm_response);
local_lm_blob = data_blob(local_lm_response, 24);
/* We can't do an NT hash here, as the password needs to be
case insensitive */
local_nt_blob = data_blob(NULL, 0);
} else {
local_lm_blob = data_blob(NULL, 0);
local_nt_blob = data_blob(NULL, 0);
}
ret = make_user_info_map(user_info, smb_name,
client_domain,
get_remote_machine_name(),
local_lm_blob.data ? &local_lm_blob : NULL,
local_nt_blob.data ? &local_nt_blob : NULL,
NULL, NULL,
plaintext_password.data ? &plaintext_password : NULL,
False);
data_blob_free(&local_lm_blob);
return NT_STATUS_IS_OK(ret) ? True : False;
}
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
/****************************************************************************
Create an auth_usersupplied_data structure
****************************************************************************/
NTSTATUS make_user_info_for_reply_enc(auth_usersupplied_info **user_info,
const char *smb_name,
const char *client_domain,
DATA_BLOB lm_resp, DATA_BLOB nt_resp)
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 make_user_info_map(user_info, smb_name,
client_domain,
get_remote_machine_name(),
lm_resp.data ? &lm_resp : NULL,
nt_resp.data ? &nt_resp : NULL,
NULL, NULL, NULL,
True);
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
}
/****************************************************************************
Create a guest user_info blob, for anonymous authenticaion.
****************************************************************************/
BOOL make_user_info_guest(auth_usersupplied_info **user_info)
{
NTSTATUS nt_status;
nt_status = make_user_info(user_info,
"","",
"","",
"",
NULL, NULL,
NULL, NULL,
NULL,
True);
return NT_STATUS_IS_OK(nt_status) ? True : False;
}
/****************************************************************************
prints a NT_USER_TOKEN to debug output.
****************************************************************************/
void debug_nt_user_token(int dbg_class, int dbg_lev, NT_USER_TOKEN *token)
{
fstring sid_str;
size_t i;
if (!token) {
DEBUGC(dbg_class, dbg_lev, ("NT user token: (NULL)\n"));
return;
}
DEBUGC(dbg_class, dbg_lev, ("NT user token of user %s\n",
sid_to_string(sid_str, &token->user_sids[0]) ));
DEBUGADDC(dbg_class, dbg_lev, ("contains %lu SIDs\n", (unsigned long)token->num_sids));
for (i = 0; i < token->num_sids; i++)
DEBUGADDC(dbg_class, dbg_lev, ("SID[%3lu]: %s\n", (unsigned long)i,
sid_to_string(sid_str, &token->user_sids[i])));
dump_se_priv( dbg_class, dbg_lev, &token->privileges );
}
/****************************************************************************
prints a UNIX 'token' to debug output.
****************************************************************************/
void debug_unix_user_token(int dbg_class, int dbg_lev, uid_t uid, gid_t gid, int n_groups, gid_t *groups)
{
int i;
DEBUGC(dbg_class, dbg_lev, ("UNIX token of user %ld\n", (long int)uid));
DEBUGADDC(dbg_class, dbg_lev, ("Primary group is %ld and contains %i supplementary groups\n", (long int)gid, n_groups));
for (i = 0; i < n_groups; i++)
DEBUGADDC(dbg_class, dbg_lev, ("Group[%3i]: %ld\n", i,
(long int)groups[i]));
}
/****************************************************************************
Create the SID list for this user.
****************************************************************************/
static NTSTATUS create_nt_user_token(const DOM_SID *user_sid, const DOM_SID *group_sid,
int n_groupSIDs, DOM_SID *groupSIDs,
BOOL is_guest, NT_USER_TOKEN **token)
{
NTSTATUS nt_status = NT_STATUS_OK;
NT_USER_TOKEN *ptoken;
int i;
int sid_ndx;
if ((ptoken = SMB_MALLOC_P(NT_USER_TOKEN)) == NULL) {
DEBUG(0, ("create_nt_token: Out of memory allocating token\n"));
nt_status = NT_STATUS_NO_MEMORY;
return nt_status;
}
ZERO_STRUCTP(ptoken);
ptoken->num_sids = n_groupSIDs + 5;
if ((ptoken->user_sids = SMB_MALLOC_ARRAY( DOM_SID, ptoken->num_sids )) == NULL) {
DEBUG(0, ("create_nt_token: Out of memory allocating SIDs\n"));
nt_status = NT_STATUS_NO_MEMORY;
return nt_status;
}
memset((char*)ptoken->user_sids,0,sizeof(DOM_SID) * ptoken->num_sids);
/*
* Note - user SID *MUST* be first in token !
* se_access_check depends on this.
*
* Primary group SID is second in token. Convention.
*/
sid_copy(&ptoken->user_sids[PRIMARY_USER_SID_INDEX], user_sid);
if (group_sid)
sid_copy(&ptoken->user_sids[PRIMARY_GROUP_SID_INDEX], group_sid);
/*
* Finally add the "standard" SIDs.
* The only difference between guest and "anonymous" (which we
* don't really support) is the addition of Authenticated_Users.
*/
sid_copy(&ptoken->user_sids[2], &global_sid_World);
sid_copy(&ptoken->user_sids[3], &global_sid_Network);
if (is_guest)
sid_copy(&ptoken->user_sids[4], &global_sid_Builtin_Guests);
else
sid_copy(&ptoken->user_sids[4], &global_sid_Authenticated_Users);
sid_ndx = 5; /* next available spot */
for (i = 0; i < n_groupSIDs; i++) {
size_t check_sid_idx;
for (check_sid_idx = 1; check_sid_idx < ptoken->num_sids; check_sid_idx++) {
if (sid_equal(&ptoken->user_sids[check_sid_idx],
&groupSIDs[i])) {
break;
}
}
if (check_sid_idx >= ptoken->num_sids) /* Not found already */ {
sid_copy(&ptoken->user_sids[sid_ndx++], &groupSIDs[i]);
} else {
ptoken->num_sids--;
}
}
/* add privileges assigned to this user */
get_privileges_for_sids( &ptoken->privileges, ptoken->user_sids, ptoken->num_sids );
debug_nt_user_token(DBGC_AUTH, 10, ptoken);
if ((lp_log_nt_token_command() != NULL) &&
(strlen(lp_log_nt_token_command()) > 0)) {
TALLOC_CTX *mem_ctx;
char *command;
fstring sidstr;
char *user_sidstr, *group_sidstr;
mem_ctx = talloc_init("setnttoken");
if (mem_ctx == NULL)
return NT_STATUS_NO_MEMORY;
sid_to_string(sidstr, &ptoken->user_sids[0]);
user_sidstr = talloc_strdup(mem_ctx, sidstr);
group_sidstr = talloc_strdup(mem_ctx, "");
for (i=1; i<ptoken->num_sids; i++) {
sid_to_string(sidstr, &ptoken->user_sids[i]);
group_sidstr = talloc_asprintf(mem_ctx, "%s %s",
group_sidstr, sidstr);
}
command = SMB_STRDUP(lp_log_nt_token_command());
command = realloc_string_sub(command, "%s", user_sidstr);
command = realloc_string_sub(command, "%t", group_sidstr);
DEBUG(8, ("running command: [%s]\n", command));
if (smbrun(command, NULL) != 0) {
DEBUG(0, ("Could not log NT token\n"));
nt_status = NT_STATUS_ACCESS_DENIED;
}
talloc_destroy(mem_ctx);
SAFE_FREE(command);
}
*token = ptoken;
return nt_status;
}
/****************************************************************************
Create the SID list for this user.
****************************************************************************/
NT_USER_TOKEN *create_nt_token(uid_t uid, gid_t gid, int ngroups, gid_t *groups, BOOL is_guest)
{
DOM_SID user_sid;
DOM_SID group_sid;
DOM_SID *group_sids;
NT_USER_TOKEN *token;
int i;
if (!NT_STATUS_IS_OK(uid_to_sid(&user_sid, uid))) {
return NULL;
}
if (!NT_STATUS_IS_OK(gid_to_sid(&group_sid, gid))) {
return NULL;
}
group_sids = SMB_MALLOC_ARRAY(DOM_SID, ngroups);
if (!group_sids) {
DEBUG(0, ("create_nt_token: malloc() failed for DOM_SID list!\n"));
return NULL;
}
for (i = 0; i < ngroups; i++) {
if (!NT_STATUS_IS_OK(gid_to_sid(&(group_sids)[i], (groups)[i]))) {
DEBUG(1, ("create_nt_token: failed to convert gid %ld to a sid!\n", (long int)groups[i]));
SAFE_FREE(group_sids);
return NULL;
}
}
if (!NT_STATUS_IS_OK(create_nt_user_token(&user_sid, &group_sid,
ngroups, group_sids, is_guest, &token))) {
SAFE_FREE(group_sids);
return NULL;
}
SAFE_FREE(group_sids);
return token;
}
/******************************************************************************
Create a token for the root user to be used internally by smbd.
This is similar to running under the context of the LOCAL_SYSTEM account
in Windows. This is a read-only token. Do not modify it or free() it.
Create a copy if your need to change it.
******************************************************************************/
NT_USER_TOKEN *get_root_nt_token( void )
{
static NT_USER_TOKEN *token = NULL;
DOM_SID u_sid, g_sid;
DOM_SID g_sids[1];
struct passwd *pw;
NTSTATUS result;
if ( token )
return token;
if ( !(pw = getpwnam( "root" )) ) {
DEBUG(0,("create_root_nt_token: getpwnam\"root\") failed!\n"));
return NULL;
}
/* get the user and primary group SIDs; although the
BUILTIN\Administrators SId is really the one that matters here */
if ( !NT_STATUS_IS_OK(uid_to_sid(&u_sid, pw->pw_uid)) )
return NULL;
if ( !NT_STATUS_IS_OK(gid_to_sid(&g_sid, pw->pw_gid)) )
return NULL;
sid_copy( &g_sids[0], &global_sid_Builtin_Administrators );
result = create_nt_user_token( &u_sid, &g_sid, 1, g_sids, False, &token);
return NT_STATUS_IS_OK(result) ? token : NULL;
}
/******************************************************************************
* this function returns the groups (SIDs) of the local SAM the user is in.
* If this samba server is a DC of the domain the user belongs to, it returns
* both domain groups and local / builtin groups. If the user is in a trusted
* domain, or samba is a member server of a domain, then this function returns
* local and builtin groups the user is a member of.
*
* currently this is a hack, as there is no sam implementation that is capable
* of groups.
*
* NOTE!! This function will fail if you pass in a winbind user without
* the domain --jerry
******************************************************************************/
static NTSTATUS get_user_groups(const char *username, uid_t uid, gid_t gid,
size_t *n_groups, DOM_SID **groups, gid_t **unix_groups)
{
int n_unix_groups;
int i;
*n_groups = 0;
*groups = NULL;
if (strchr(username, *lp_winbind_separator()) == NULL) {
NTSTATUS result;
become_root();
result = pdb_enum_group_memberships(username, gid, groups,
unix_groups, n_groups);
unbecome_root();
return result;
}
/* We have the separator, this must be winbind */
n_unix_groups = winbind_getgroups( username, unix_groups );
DEBUG(10,("get_user_groups: winbind_getgroups(%s): result = %s\n",
username, n_unix_groups == -1 ? "FAIL" : "SUCCESS"));
if ( n_unix_groups == -1 )
return NT_STATUS_NO_SUCH_USER; /* what should this return
* value be? */
debug_unix_user_token(DBGC_CLASS, 5, uid, gid, n_unix_groups, *unix_groups);
/* now setup the space for storing the SIDS */
if (n_unix_groups > 0) {
*groups = SMB_MALLOC_ARRAY(DOM_SID, n_unix_groups);
if (!*groups) {
DEBUG(0, ("get_user_group: malloc() failed for DOM_SID list!\n"));
SAFE_FREE(*unix_groups);
return NT_STATUS_NO_MEMORY;
}
}
*n_groups = n_unix_groups;
for (i = 0; i < *n_groups; i++) {
if (!NT_STATUS_IS_OK(gid_to_sid(&(*groups)[i], (*unix_groups)[i]))) {
DEBUG(1, ("get_user_groups: failed to convert gid %ld to a sid!\n",
(long int)(*unix_groups)[i+1]));
SAFE_FREE(*groups);
SAFE_FREE(*unix_groups);
return NT_STATUS_NO_SUCH_USER;
}
}
return NT_STATUS_OK;
}
/***************************************************************************
Make a user_info struct
***************************************************************************/
static NTSTATUS make_server_info(auth_serversupplied_info **server_info)
{
*server_info = SMB_MALLOC_P(auth_serversupplied_info);
if (!*server_info) {
DEBUG(0,("make_server_info: malloc failed!\n"));
return NT_STATUS_NO_MEMORY;
}
ZERO_STRUCTP(*server_info);
/* Initialise the uid and gid values to something non-zero
which may save us from giving away root access if there
is a bug in allocating these fields. */
(*server_info)->uid = -1;
(*server_info)->gid = -1;
return NT_STATUS_OK;
}
/***************************************************************************
Fill a server_info struct from a SAM_ACCOUNT with their groups
***************************************************************************/
static NTSTATUS add_user_groups(auth_serversupplied_info **server_info,
const char * unix_username,
SAM_ACCOUNT *sampass,
uid_t uid, gid_t gid)
{
NTSTATUS nt_status;
const DOM_SID *user_sid = pdb_get_user_sid(sampass);
const DOM_SID *group_sid = pdb_get_group_sid(sampass);
size_t n_groupSIDs = 0;
DOM_SID *groupSIDs = NULL;
gid_t *unix_groups = NULL;
NT_USER_TOKEN *token;
BOOL is_guest;
uint32 rid;
nt_status = get_user_groups(unix_username, uid, gid,
&n_groupSIDs, &groupSIDs, &unix_groups);
if (!NT_STATUS_IS_OK(nt_status)) {
DEBUG(4,("get_user_groups_from_local_sam failed\n"));
free_server_info(server_info);
return nt_status;
}
is_guest = (sid_peek_rid(user_sid, &rid) && rid == DOMAIN_USER_RID_GUEST);
if (!NT_STATUS_IS_OK(nt_status = create_nt_user_token(user_sid, group_sid,
n_groupSIDs, groupSIDs, is_guest,
&token)))
{
DEBUG(4,("create_nt_user_token failed\n"));
SAFE_FREE(groupSIDs);
SAFE_FREE(unix_groups);
free_server_info(server_info);
return nt_status;
}
SAFE_FREE(groupSIDs);
(*server_info)->n_groups = n_groupSIDs;
(*server_info)->groups = unix_groups;
(*server_info)->ptok = token;
return nt_status;
}
/***************************************************************************
Make (and fill) a user_info struct from a SAM_ACCOUNT
***************************************************************************/
NTSTATUS make_server_info_sam(auth_serversupplied_info **server_info,
SAM_ACCOUNT *sampass)
{
NTSTATUS nt_status;
struct passwd *pwd;
if (!NT_STATUS_IS_OK(nt_status = make_server_info(server_info)))
return nt_status;
(*server_info)->sam_account = sampass;
if ( !(pwd = getpwnam_alloc(pdb_get_username(sampass))) ) {
DEBUG(1, ("User %s in passdb, but getpwnam() fails!\n",
pdb_get_username(sampass)));
free_server_info(server_info);
return NT_STATUS_NO_SUCH_USER;
}
(*server_info)->unix_name = smb_xstrdup(pwd->pw_name);
(*server_info)->gid = pwd->pw_gid;
(*server_info)->uid = pwd->pw_uid;
passwd_free(&pwd);
if (!NT_STATUS_IS_OK(nt_status = add_user_groups(server_info, pdb_get_username(sampass),
sampass,
(*server_info)->uid,
(*server_info)->gid)))
{
free_server_info(server_info);
return nt_status;
}
(*server_info)->sam_fill_level = SAM_FILL_ALL;
DEBUG(5,("make_server_info_sam: made server info for user %s -> %s\n",
pdb_get_username(sampass),
(*server_info)->unix_name));
return nt_status;
}
/***************************************************************************
Make (and fill) a user_info struct from a Kerberos PAC logon_info by conversion
to a SAM_ACCOUNT
***************************************************************************/
NTSTATUS make_server_info_pac(auth_serversupplied_info **server_info,
char *unix_username,
struct passwd *pwd,
PAC_LOGON_INFO *logon_info)
{
NTSTATUS nt_status;
SAM_ACCOUNT *sampass = NULL;
DOM_SID user_sid, group_sid;
fstring dom_name;
if (!NT_STATUS_IS_OK(nt_status = pdb_init_sam_pw(&sampass, pwd))) {
return nt_status;
}
if (!NT_STATUS_IS_OK(nt_status = make_server_info(server_info))) {
return nt_status;
}
/* only copy user_sid, group_sid and domain name out of the PAC for
* now, we will benefit from more later - Guenther */
sid_copy(&user_sid, &logon_info->info3.dom_sid.sid);
sid_append_rid(&user_sid, logon_info->info3.user_rid);
pdb_set_user_sid(sampass, &user_sid, PDB_SET);
sid_copy(&group_sid, &logon_info->info3.dom_sid.sid);
sid_append_rid(&group_sid, logon_info->info3.group_rid);
pdb_set_group_sid(sampass, &group_sid, PDB_SET);
unistr2_to_ascii(dom_name, &logon_info->info3.uni_logon_dom, -1);
pdb_set_domain(sampass, dom_name, PDB_SET);
pdb_set_logon_count(sampass, logon_info->info3.logon_count, PDB_SET);
(*server_info)->sam_account = sampass;
if (!NT_STATUS_IS_OK(nt_status = add_user_groups(server_info, unix_username,
sampass, pwd->pw_uid, pwd->pw_gid)))
{
return nt_status;
}
(*server_info)->unix_name = smb_xstrdup(unix_username);
(*server_info)->sam_fill_level = SAM_FILL_ALL;
(*server_info)->uid = pwd->pw_uid;
(*server_info)->gid = pwd->pw_gid;
return nt_status;
}
/***************************************************************************
Make (and fill) a user_info struct from a 'struct passwd' by conversion
to a SAM_ACCOUNT
***************************************************************************/
NTSTATUS make_server_info_pw(auth_serversupplied_info **server_info,
char *unix_username,
struct passwd *pwd)
{
NTSTATUS nt_status;
SAM_ACCOUNT *sampass = NULL;
if (!NT_STATUS_IS_OK(nt_status = pdb_init_sam_pw(&sampass, pwd))) {
return nt_status;
}
if (!NT_STATUS_IS_OK(nt_status = make_server_info(server_info))) {
return nt_status;
}
(*server_info)->sam_account = sampass;
if (!NT_STATUS_IS_OK(nt_status = add_user_groups(server_info, unix_username,
sampass, pwd->pw_uid, pwd->pw_gid)))
{
return nt_status;
}
(*server_info)->unix_name = smb_xstrdup(unix_username);
(*server_info)->sam_fill_level = SAM_FILL_ALL;
(*server_info)->uid = pwd->pw_uid;
(*server_info)->gid = pwd->pw_gid;
return nt_status;
}
/***************************************************************************
Make (and fill) a user_info struct for a guest login.
***************************************************************************/
static NTSTATUS make_new_server_info_guest(auth_serversupplied_info **server_info)
{
NTSTATUS nt_status;
SAM_ACCOUNT *sampass = NULL;
DOM_SID guest_sid;
if (!NT_STATUS_IS_OK(nt_status = pdb_init_sam(&sampass))) {
return nt_status;
}
sid_copy(&guest_sid, get_global_sam_sid());
sid_append_rid(&guest_sid, DOMAIN_USER_RID_GUEST);
become_root();
if (!pdb_getsampwsid(sampass, &guest_sid)) {
unbecome_root();
return NT_STATUS_NO_SUCH_USER;
}
unbecome_root();
nt_status = make_server_info_sam(server_info, sampass);
if (NT_STATUS_IS_OK(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
static const char zeros[16];
(*server_info)->guest = True;
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
/* annoying, but the Guest really does have a session key,
and it is all zeros! */
(*server_info)->user_session_key = data_blob(zeros, sizeof(zeros));
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)->lm_session_key = data_blob(zeros, sizeof(zeros));
}
return nt_status;
}
static auth_serversupplied_info *copy_serverinfo(auth_serversupplied_info *src)
{
auth_serversupplied_info *dst;
if (!NT_STATUS_IS_OK(make_server_info(&dst)))
return NULL;
dst->guest = src->guest;
dst->uid = src->uid;
dst->gid = src->gid;
dst->n_groups = src->n_groups;
if (src->n_groups != 0)
dst->groups = memdup(src->groups, sizeof(gid_t)*dst->n_groups);
else
dst->groups = NULL;
dst->ptok = dup_nt_token(src->ptok);
dst->user_session_key = data_blob(src->user_session_key.data,
src->user_session_key.length);
dst->lm_session_key = data_blob(src->lm_session_key.data,
src->lm_session_key.length);
pdb_copy_sam_account(src->sam_account, &dst->sam_account);
dst->pam_handle = NULL;
dst->unix_name = smb_xstrdup(src->unix_name);
return dst;
}
static auth_serversupplied_info *guest_info = NULL;
BOOL init_guest_info(void)
{
if (guest_info != NULL)
return True;
return NT_STATUS_IS_OK(make_new_server_info_guest(&guest_info));
}
NTSTATUS make_server_info_guest(auth_serversupplied_info **server_info)
{
*server_info = copy_serverinfo(guest_info);
return (*server_info != NULL) ? NT_STATUS_OK : NT_STATUS_NO_MEMORY;
}
/***************************************************************************
Purely internal function for make_server_info_info3
Fill the sam account from getpwnam
***************************************************************************/
static NTSTATUS fill_sam_account(TALLOC_CTX *mem_ctx,
const char *domain,
const char *username,
char **found_username,
uid_t *uid, gid_t *gid,
SAM_ACCOUNT **sam_account)
{
NTSTATUS nt_status;
fstring dom_user, lower_username;
fstring real_username;
struct passwd *passwd;
fstrcpy( lower_username, username );
strlower_m( lower_username );
fstr_sprintf(dom_user, "%s%c%s", domain, *lp_winbind_separator(),
lower_username);
/* get the passwd struct but don't create the user if he/she
does not exist. We were explicitly called from a following
a winbindd authentication request so we should assume that
nss_winbindd is working */
map_username( dom_user );
if ( !(passwd = smb_getpwnam( dom_user, real_username, True )) )
return NT_STATUS_NO_SUCH_USER;
*uid = passwd->pw_uid;
*gid = passwd->pw_gid;
/* This is pointless -- there is no suport for differing
unix and windows names. Make sure to always store the
one we actually looked up and succeeded. Have I mentioned
why I hate the 'winbind use default domain' parameter?
--jerry */
*found_username = talloc_strdup( mem_ctx, real_username );
DEBUG(5,("fill_sam_account: located username was [%s]\n",
*found_username));
nt_status = pdb_init_sam_pw(sam_account, passwd);
passwd_free(&passwd);
return nt_status;
}
/****************************************************************************
Wrapper to allow the getpwnam() call to strip the domain name and
try again in case a local UNIX user is already there. Also run through
the username if we fallback to the username only.
****************************************************************************/
struct passwd *smb_getpwnam( char *domuser, fstring save_username, BOOL create )
{
struct passwd *pw = NULL;
char *p;
fstring username;
/* we only save a copy of the username it has been mangled
by winbindd use default domain */
save_username[0] = '\0';
/* don't call map_username() here since it has to be done higher
up the stack so we don't call it mutliple times */
fstrcpy( username, domuser );
p = strchr_m( username, *lp_winbind_separator() );
/* code for a DOMAIN\user string */
if ( p ) {
fstring strip_username;
pw = Get_Pwnam_alloc( domuser );
if ( pw ) {
/* make sure we get the case of the username correct */
/* work around 'winbind use default domain = yes' */
if ( !strchr_m( pw->pw_name, *lp_winbind_separator() ) ) {
char *domain;
/* split the domain and username into 2 strings */
*p = '\0';
domain = username;
fstr_sprintf(save_username, "%s%c%s", domain, *lp_winbind_separator(), pw->pw_name);
}
else
fstrcpy( save_username, pw->pw_name );
/* whew -- done! */
return pw;
}
/* setup for lookup of just the username */
/* remember that p and username are overlapping memory */
p++;
fstrcpy( strip_username, p );
fstrcpy( username, strip_username );
}
/* just lookup a plain username */
pw = Get_Pwnam_alloc(username);
/* Create local user if requested. */
if ( !pw && create ) {
/* Don't add a machine account. */
if (username[strlen(username)-1] == '$')
return NULL;
smb_create_user(NULL, username, NULL);
pw = Get_Pwnam_alloc(username);
}
/* one last check for a valid passwd struct */
if ( pw )
fstrcpy( save_username, pw->pw_name );
return pw;
}
/***************************************************************************
Make a server_info struct from the info3 returned by a domain logon
***************************************************************************/
NTSTATUS make_server_info_info3(TALLOC_CTX *mem_ctx,
const char *internal_username,
const char *sent_nt_username,
const char *domain,
auth_serversupplied_info **server_info,
NET_USER_INFO_3 *info3)
{
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
static const char zeros[16];
NTSTATUS nt_status = NT_STATUS_OK;
char *found_username;
const char *nt_domain;
const char *nt_username;
SAM_ACCOUNT *sam_account = NULL;
DOM_SID user_sid;
DOM_SID group_sid;
uid_t uid;
gid_t gid;
size_t n_lgroupSIDs;
DOM_SID *lgroupSIDs = NULL;
gid_t *unix_groups = NULL;
NT_USER_TOKEN *token;
DOM_SID *all_group_SIDs;
size_t i;
/*
Here is where we should check the list of
trusted domains, and verify that the SID
matches.
*/
sid_copy(&user_sid, &info3->dom_sid.sid);
if (!sid_append_rid(&user_sid, info3->user_rid)) {
return NT_STATUS_INVALID_PARAMETER;
}
sid_copy(&group_sid, &info3->dom_sid.sid);
if (!sid_append_rid(&group_sid, info3->group_rid)) {
return NT_STATUS_INVALID_PARAMETER;
}
if (!(nt_username = unistr2_tdup(mem_ctx, &(info3->uni_user_name)))) {
/* If the server didn't give us one, just use the one we sent them */
nt_username = sent_nt_username;
}
if (!(nt_domain = unistr2_tdup(mem_ctx, &(info3->uni_logon_dom)))) {
/* If the server didn't give us one, just use the one we sent them */
nt_domain = domain;
}
/* try to fill the SAM account.. If getpwnam() fails, then try the
add user script (2.2.x behavior).
We use the _unmapped_ username here in an attempt to provide
consistent username mapping behavior between kerberos and NTLM[SSP]
authentication in domain mode security. I.E. Username mapping should
be applied to the fully qualified username (e.g. DOMAIN\user) and
no just the login name. Yes this mean swe called map_username()
unnecessarily in make_user_info_map() but that is how the current
code is designed. Making the change here is the least disruptive
place. -- jerry */
nt_status = fill_sam_account(mem_ctx, nt_domain, sent_nt_username,
&found_username, &uid, &gid, &sam_account);
if (NT_STATUS_EQUAL(nt_status, NT_STATUS_NO_SUCH_USER)) {
DEBUG(3,("User %s does not exist, trying to add it\n", internal_username));
smb_create_user( nt_domain, sent_nt_username, NULL);
nt_status = fill_sam_account( mem_ctx, nt_domain, sent_nt_username,
&found_username, &uid, &gid, &sam_account );
}
/* if we still don't have a valid unix account check for
'map to gues = bad uid' */
if (!NT_STATUS_IS_OK(nt_status)) {
if ( lp_map_to_guest() == MAP_TO_GUEST_ON_BAD_UID ) {
make_server_info_guest(server_info);
return NT_STATUS_OK;
}
DEBUG(0, ("make_server_info_info3: pdb_init_sam failed!\n"));
return nt_status;
}
if (!pdb_set_nt_username(sam_account, nt_username, PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_username(sam_account, nt_username, PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_domain(sam_account, nt_domain, PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_user_sid(sam_account, &user_sid, PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_UNSUCCESSFUL;
}
if (!pdb_set_group_sid(sam_account, &group_sid, PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_UNSUCCESSFUL;
}
if (!pdb_set_fullname(sam_account, unistr2_static(&(info3->uni_full_name)),
PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_logon_script(sam_account, unistr2_static(&(info3->uni_logon_script)), PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_profile_path(sam_account, unistr2_static(&(info3->uni_profile_path)), PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_homedir(sam_account, unistr2_static(&(info3->uni_home_dir)), PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!pdb_set_dir_drive(sam_account, unistr2_static(&(info3->uni_dir_drive)), PDB_CHANGED)) {
pdb_free_sam(&sam_account);
return NT_STATUS_NO_MEMORY;
}
if (!NT_STATUS_IS_OK(nt_status = make_server_info(server_info))) {
DEBUG(4, ("make_server_info failed!\n"));
pdb_free_sam(&sam_account);
return nt_status;
}
Here's the code to make winbindd work on a Samba DC to handle domain trusts. Jeremy and I talked about this and it's going in as working code. It keeps winbind clean and solves the trust problem with minimal changes. To summarize, there are 2 basic cases where the deadlock would occur. (1) lookuping up secondary groups for a user, and (2) get[gr|pw]nam() calls that fall through the NSS layer because they don't exist anywhere. o To handle case #1, we bypass winbindd in sys_getgrouplist() unless the username includes the 'winbind separator'. o Case #2 is handled by adding checks in winbindd to return failure if we are a DC and the domain matches our own. This code has been tested using basic share connections, domain logons, and with pam_winbind (both with and without 'winbind use default domain'). The 'trustdomain' auth module should work as well if an admin wants to manually create UNIX users for acounts in the trusted domains. Other misc fixes: * we need to fix check_ntlm_password() to be able to determine if an auth module is authoritative over a user (NT_STATUS_WRONG_PASSWORD, etc...). I worked around my specific situation, but this needs to be fixed. the winbindd auth module was causing delays. * fix named server mutex deadlock between trust domain auth module and winbindd looking up a uid * make sure SAM_ACCOUNT gets stored in the server_info struct for the _net_sam_logon() reply. Configuration details: The recommended method for supporting trusts is to use winbind. The gets us around some of the server mutex issues as well. * set 'files winbind' for passwd: and group: in /etc/nsswitch.conf * create domain trusts like normal * join winbind on the pdc to the Samba domain using 'net rpc join' * add normal parameters to smb.conf for winbind * set 'auth method = guest sam winbind' * start smbd, nmbd, & winbindd Problems that remain: * join a Windows 2k/XP box to a Samba domain. * create a 2-way trust between the Samba domain and an NT domain * logon to the windows client as a user from theh trusted domain * try to browse server in the trusted domain (or other workstations). an NT client seems to work ok, but 2k and XP either prompt for passwords or fail with errors. apparanently this never got tested since no one has ever been able to logon as a trusted user to a Samba domain from a Windows client. (This used to be commit f804b590f9dbf1f0147c06a0a2f12e221ae6fc3b)
2003-06-29 07:39:50 +04:00
/* save this here to _net_sam_logon() doesn't fail (it assumes a
valid SAM_ACCOUNT) */
(*server_info)->sam_account = sam_account;
(*server_info)->unix_name = smb_xstrdup(found_username);
/* Fill in the unix info we found on the way */
(*server_info)->sam_fill_level = SAM_FILL_ALL;
(*server_info)->uid = uid;
(*server_info)->gid = gid;
/* Store the user group information in the server_info
returned to the caller. */
nt_status = get_user_groups((*server_info)->unix_name,
uid, gid, &n_lgroupSIDs, &lgroupSIDs, &unix_groups);
if ( !NT_STATUS_IS_OK(nt_status) ) {
DEBUG(4,("get_user_groups failed\n"));
return nt_status;
}
(*server_info)->groups = unix_groups;
(*server_info)->n_groups = n_lgroupSIDs;
/* Create a 'combined' list of all SIDs we might want in the SD */
all_group_SIDs = SMB_MALLOC_ARRAY(DOM_SID,info3->num_groups2 + info3->num_other_sids + n_lgroupSIDs);
if (!all_group_SIDs) {
DEBUG(0, ("malloc() failed for DOM_SID list!\n"));
SAFE_FREE(lgroupSIDs);
free_server_info(server_info);
return NT_STATUS_NO_MEMORY;
}
/* and create (by appending rids) the 'domain' sids */
for (i = 0; i < info3->num_groups2; i++) {
sid_copy(&all_group_SIDs[i], &(info3->dom_sid.sid));
if (!sid_append_rid(&all_group_SIDs[i], info3->gids[i].g_rid)) {
nt_status = NT_STATUS_INVALID_PARAMETER;
DEBUG(3,("could not append additional group rid 0x%x\n",
info3->gids[i].g_rid));
SAFE_FREE(lgroupSIDs);
SAFE_FREE(all_group_SIDs);
free_server_info(server_info);
return nt_status;
}
}
/* Copy 'other' sids. We need to do sid filtering here to
prevent possible elevation of privileges. See:
http://www.microsoft.com/windows2000/techinfo/administration/security/sidfilter.asp
*/
for (i = 0; i < info3->num_other_sids; i++) {
sid_copy(&all_group_SIDs[info3->num_groups2 + i],
&info3->other_sids[i].sid);
}
/* add local alias sids */
for (i = 0; i < n_lgroupSIDs; i++) {
sid_copy(&all_group_SIDs[info3->num_groups2 +
info3->num_other_sids + i],
&lgroupSIDs[i]);
}
/* Where are the 'global' sids... */
/* can the user be guest? if yes, where is it stored? */
nt_status = create_nt_user_token(&user_sid, &group_sid,
info3->num_groups2 + info3->num_other_sids + n_lgroupSIDs,
all_group_SIDs, False, &token);
if ( !NT_STATUS_IS_OK(nt_status) ) {
DEBUG(4,("create_nt_user_token failed\n"));
SAFE_FREE(lgroupSIDs);
SAFE_FREE(all_group_SIDs);
free_server_info(server_info);
return nt_status;
}
(*server_info)->login_server = unistr2_tdup(mem_ctx,
&(info3->uni_logon_srv));
(*server_info)->ptok = token;
SAFE_FREE(lgroupSIDs);
SAFE_FREE(all_group_SIDs);
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
/* ensure we are never given NULL session keys */
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
if (memcmp(info3->user_sess_key, zeros, sizeof(zeros)) == 0) {
(*server_info)->user_session_key = data_blob(NULL, 0);
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
} else {
(*server_info)->user_session_key = data_blob(info3->user_sess_key, sizeof(info3->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
}
2003-03-24 12:54:13 +03:00
if (memcmp(info3->lm_sess_key, zeros, 8) == 0) {
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)->lm_session_key = data_blob(NULL, 0);
} else {
(*server_info)->lm_session_key = data_blob(info3->lm_sess_key, sizeof(info3->lm_sess_key));
}
return NT_STATUS_OK;
}
/***************************************************************************
Free a user_info struct
***************************************************************************/
void free_user_info(auth_usersupplied_info **user_info)
{
DEBUG(5,("attempting to free (and zero) a user_info structure\n"));
if (*user_info != NULL) {
if ((*user_info)->smb_name.str) {
DEBUG(10,("structure was created for %s\n", (*user_info)->smb_name.str));
}
SAFE_FREE((*user_info)->smb_name.str);
SAFE_FREE((*user_info)->internal_username.str);
SAFE_FREE((*user_info)->client_domain.str);
SAFE_FREE((*user_info)->domain.str);
SAFE_FREE((*user_info)->wksta_name.str);
data_blob_free(&(*user_info)->lm_resp);
data_blob_free(&(*user_info)->nt_resp);
data_blob_clear_free(&(*user_info)->lm_interactive_pwd);
data_blob_clear_free(&(*user_info)->nt_interactive_pwd);
data_blob_clear_free(&(*user_info)->plaintext_password);
ZERO_STRUCT(**user_info);
}
SAFE_FREE(*user_info);
}
/***************************************************************************
Clear out a server_info struct that has been allocated
***************************************************************************/
void free_server_info(auth_serversupplied_info **server_info)
{
DEBUG(5,("attempting to free (and zero) a server_info structure\n"));
if (*server_info != NULL) {
pdb_free_sam(&(*server_info)->sam_account);
/* call pam_end here, unless we know we are keeping it */
delete_nt_token( &(*server_info)->ptok );
SAFE_FREE((*server_info)->groups);
SAFE_FREE((*server_info)->unix_name);
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_free(&(*server_info)->lm_session_key);
data_blob_free(&(*server_info)->user_session_key);
ZERO_STRUCT(**server_info);
}
SAFE_FREE(*server_info);
}
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
/***************************************************************************
Make an auth_methods struct
***************************************************************************/
BOOL make_auth_methods(struct auth_context *auth_context, 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
{
if (!auth_context) {
smb_panic("no auth_context supplied to make_auth_methods()!\n");
}
if (!auth_method) {
smb_panic("make_auth_methods: pointer to auth_method pointer is NULL!\n");
}
*auth_method = TALLOC_P(auth_context->mem_ctx, auth_methods);
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 (!*auth_method) {
DEBUG(0,("make_auth_method: malloc failed!\n"));
return False;
}
ZERO_STRUCTP(*auth_method);
return True;
}
/****************************************************************************
Delete a SID token.
****************************************************************************/
void delete_nt_token(NT_USER_TOKEN **pptoken)
{
if (*pptoken) {
NT_USER_TOKEN *ptoken = *pptoken;
SAFE_FREE( ptoken->user_sids );
ZERO_STRUCTP(ptoken);
}
SAFE_FREE(*pptoken);
}
/****************************************************************************
Duplicate a SID token.
****************************************************************************/
NT_USER_TOKEN *dup_nt_token(NT_USER_TOKEN *ptoken)
{
NT_USER_TOKEN *token;
if (!ptoken)
return NULL;
if ((token = SMB_MALLOC_P(NT_USER_TOKEN)) == NULL)
return NULL;
ZERO_STRUCTP(token);
token->user_sids = (DOM_SID *)memdup( ptoken->user_sids, sizeof(DOM_SID) * ptoken->num_sids );
if ( !token ) {
SAFE_FREE(token);
return NULL;
}
token->num_sids = ptoken->num_sids;
/* copy the privileges; don't consider failure to be critical here */
if ( !se_priv_copy( &token->privileges, &ptoken->privileges ) ) {
DEBUG(0,("dup_nt_token: Failure to copy SE_PRIV!. Continuing with 0 privileges assigned.\n"));
}
return token;
}
/****************************************************************************
Check for a SID in an NT_USER_TOKEN
****************************************************************************/
static BOOL nt_token_check_sid ( DOM_SID *sid, NT_USER_TOKEN *token )
{
int i;
if ( !sid || !token )
return False;
for ( i=0; i<token->num_sids; i++ ) {
if ( sid_equal( sid, &token->user_sids[i] ) )
return True;
}
return False;
}
BOOL nt_token_check_domain_rid( NT_USER_TOKEN *token, uint32 rid )
{
DOM_SID domain_sid;
/* if we are a domain member, the get the domain SID, else for
a DC or standalone server, use our own SID */
if ( lp_server_role() == ROLE_DOMAIN_MEMBER ) {
if ( !secrets_fetch_domain_sid( lp_workgroup(), &domain_sid ) ) {
DEBUG(1,("nt_token_check_domain_rid: Cannot lookup SID for domain [%s]\n",
lp_workgroup()));
return False;
}
}
else
sid_copy( &domain_sid, get_global_sam_sid() );
sid_append_rid( &domain_sid, rid );
return nt_token_check_sid( &domain_sid, token );\
}
/**
* Verify whether or not given domain is trusted.
*
* @param domain_name name of the domain to be verified
* @return true if domain is one of the trusted once or
* false if otherwise
**/
BOOL is_trusted_domain(const char* dom_name)
{
DOM_SID trustdom_sid;
BOOL ret;
/* no trusted domains for a standalone server */
if ( lp_server_role() == ROLE_STANDALONE )
return False;
/* if we are a DC, then check for a direct trust relationships */
if ( IS_DC ) {
become_root();
DEBUG (5,("is_trusted_domain: Checking for domain trust with [%s]\n",
dom_name ));
ret = secrets_fetch_trusted_domain_password(dom_name, NULL, NULL, NULL);
unbecome_root();
if (ret)
return True;
}
else {
NSS_STATUS result;
/* If winbind is around, ask it */
result = wb_is_trusted_domain(dom_name);
if (result == NSS_STATUS_SUCCESS) {
return True;
}
if (result == NSS_STATUS_NOTFOUND) {
/* winbind could not find the domain */
return False;
}
/* The only other possible result is that winbind is not up
and running. We need to update the trustdom_cache
ourselves */
update_trustdom_cache();
}
/* now the trustdom cache should be available a DC could still
* have a transitive trust so fall back to the cache of trusted
* domains (like a domain member would use */
if ( trustdom_cache_fetch(dom_name, &trustdom_sid) ) {
return True;
}
return False;
}