1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-24 21:34:56 +03:00
samba-mirror/source3/passdb/pdb_interface.c

2061 lines
51 KiB
C
Raw Normal View History

This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
/*
Unix SMB/CIFS implementation.
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
Password and authentication handling
Copyright (C) Andrew Bartlett 2002
Copyright (C) Jelmer Vernooij 2002
Copyright (C) Simo Sorce 2003
Copyright (C) Volker Lendecke 2006
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
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 3 of the License, or
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
(at your option) any later version.
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
*/
#include "includes.h"
#undef DBGC_CLASS
#define DBGC_CLASS DBGC_PASSDB
static_decl_pdb;
static struct pdb_init_function_entry *backends = NULL;
static void lazy_initialize_passdb(void)
{
static bool initialized = False;
if(initialized) {
return;
}
static_init_pdb;
initialized = True;
}
static bool lookup_global_sam_rid(TALLOC_CTX *mem_ctx, uint32 rid,
const char **name,
enum lsa_SidType *psid_name_use,
union unid_t *unix_id);
NTSTATUS smb_register_passdb(int version, const char *name, pdb_init_function init)
{
struct pdb_init_function_entry *entry = backends;
if(version != PASSDB_INTERFACE_VERSION) {
DEBUG(0,("Can't register passdb backend!\n"
"You tried to register a passdb module with PASSDB_INTERFACE_VERSION %d, "
"while this version of samba uses version %d\n",
version,PASSDB_INTERFACE_VERSION));
return NT_STATUS_OBJECT_TYPE_MISMATCH;
}
if (!name || !init) {
return NT_STATUS_INVALID_PARAMETER;
}
DEBUG(5,("Attempting to register passdb backend %s\n", name));
/* Check for duplicates */
if (pdb_find_backend_entry(name)) {
DEBUG(0,("There already is a passdb backend registered with the name %s!\n", name));
return NT_STATUS_OBJECT_NAME_COLLISION;
}
entry = SMB_XMALLOC_P(struct pdb_init_function_entry);
entry->name = smb_xstrdup(name);
entry->init = init;
DLIST_ADD(backends, entry);
DEBUG(5,("Successfully added passdb backend '%s'\n", name));
return NT_STATUS_OK;
}
struct pdb_init_function_entry *pdb_find_backend_entry(const char *name)
{
struct pdb_init_function_entry *entry = backends;
while(entry) {
if (strcmp(entry->name, name)==0) return entry;
entry = entry->next;
}
return NULL;
}
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
/*
* The event context for the passdb backend. I know this is a bad hack and yet
* another static variable, but our pdb API is a global thing per
* definition. The first use for this is the LDAP idle function, more might be
* added later.
*
* I don't feel too bad about this static variable, it replaces the
* smb_idle_event_list that used to exist in lib/module.c. -- VL
*/
static struct event_context *pdb_event_ctx;
struct event_context *pdb_get_event_context(void)
{
return pdb_event_ctx;
}
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
/******************************************************************
Make a pdb_methods from scratch
*******************************************************************/
NTSTATUS make_pdb_method_name(struct pdb_methods **methods, const char *selected)
{
char *module_name = smb_xstrdup(selected);
char *module_location = NULL, *p;
struct pdb_init_function_entry *entry;
NTSTATUS nt_status = NT_STATUS_UNSUCCESSFUL;
lazy_initialize_passdb();
p = strchr(module_name, ':');
if (p) {
*p = 0;
module_location = p+1;
trim_char(module_location, ' ', ' ');
}
trim_char(module_name, ' ', ' ');
DEBUG(5,("Attempting to find a passdb backend to match %s (%s)\n", selected, module_name));
entry = pdb_find_backend_entry(module_name);
/* Try to find a module that contains this module */
if (!entry) {
DEBUG(2,("No builtin backend found, trying to load plugin\n"));
if(NT_STATUS_IS_OK(smb_probe_module("pdb", module_name)) && !(entry = pdb_find_backend_entry(module_name))) {
DEBUG(0,("Plugin is available, but doesn't register passdb backend %s\n", module_name));
SAFE_FREE(module_name);
return NT_STATUS_UNSUCCESSFUL;
}
}
/* No such backend found */
if(!entry) {
DEBUG(0,("No builtin nor plugin backend for %s found\n", module_name));
SAFE_FREE(module_name);
return NT_STATUS_INVALID_PARAMETER;
}
DEBUG(5,("Found pdb backend %s\n", module_name));
if ( !NT_STATUS_IS_OK( nt_status = entry->init(methods, module_location) ) ) {
DEBUG(0,("pdb backend %s did not correctly init (error was %s)\n",
selected, nt_errstr(nt_status)));
SAFE_FREE(module_name);
return nt_status;
}
SAFE_FREE(module_name);
DEBUG(5,("pdb backend %s has a valid init\n", selected));
return nt_status;
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
/******************************************************************
Return an already initialized pdb_methods structure
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
*******************************************************************/
static struct pdb_methods *pdb_get_methods_reload( bool reload )
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
{
static struct pdb_methods *pdb = NULL;
if ( pdb && reload ) {
pdb->free_private_data( &(pdb->private_data) );
if ( !NT_STATUS_IS_OK( make_pdb_method_name( &pdb, lp_passdb_backend() ) ) ) {
char *msg = NULL;
if (asprintf(&msg, "pdb_get_methods_reload: "
"failed to get pdb methods for backend %s\n",
lp_passdb_backend()) > 0) {
smb_panic(msg);
} else {
smb_panic("pdb_get_methods_reload");
}
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
}
if ( !pdb ) {
if ( !NT_STATUS_IS_OK( make_pdb_method_name( &pdb, lp_passdb_backend() ) ) ) {
char *msg = NULL;
if (asprintf(&msg, "pdb_get_methods_reload: "
"failed to get pdb methods for backend %s\n",
lp_passdb_backend()) > 0) {
smb_panic(msg);
} else {
smb_panic("pdb_get_methods_reload");
}
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
}
return pdb;
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
static struct pdb_methods *pdb_get_methods(void)
{
return pdb_get_methods_reload(False);
}
bool pdb_getsampwnam(struct samu *sam_acct, const char *username)
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
{
struct pdb_methods *pdb = pdb_get_methods();
struct samu *for_cache;
const struct dom_sid *user_sid;
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
if (!NT_STATUS_IS_OK(pdb->getsampwnam(pdb, sam_acct, username))) {
return False;
}
for_cache = samu_new(NULL);
if (for_cache == NULL) {
return False;
}
if (!pdb_copy_sam_account(for_cache, sam_acct)) {
TALLOC_FREE(for_cache);
return False;
}
user_sid = pdb_get_user_sid(for_cache);
memcache_add_talloc(NULL, PDB_GETPWSID_CACHE,
data_blob_const(user_sid, sizeof(*user_sid)),
&for_cache);
return True;
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
/**********************************************************************
**********************************************************************/
bool guest_user_info( struct samu *user )
{
struct passwd *pwd;
NTSTATUS result;
const char *guestname = lp_guestaccount();
if ( !(pwd = getpwnam_alloc(talloc_autofree_context(), guestname ) ) ) {
DEBUG(0,("guest_user_info: Unable to locate guest account [%s]!\n",
guestname));
return False;
}
result = samu_set_unix(user, pwd );
TALLOC_FREE( pwd );
return NT_STATUS_IS_OK( result );
}
/**********************************************************************
**********************************************************************/
bool pdb_getsampwsid(struct samu *sam_acct, const DOM_SID *sid)
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
{
struct pdb_methods *pdb = pdb_get_methods();
uint32 rid;
void *cache_data;
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
/* hard code the Guest RID of 501 */
if ( !sid_peek_check_rid( get_global_sam_sid(), sid, &rid ) )
return False;
if ( rid == DOMAIN_USER_RID_GUEST ) {
DEBUG(6,("pdb_getsampwsid: Building guest account\n"));
return guest_user_info( sam_acct );
}
/* check the cache first */
cache_data = memcache_lookup_talloc(
NULL, PDB_GETPWSID_CACHE, data_blob_const(sid, sizeof(*sid)));
if (cache_data != NULL) {
struct samu *cache_copy = talloc_get_type_abort(
cache_data, struct samu);
return pdb_copy_sam_account(sam_acct, cache_copy);
}
return NT_STATUS_IS_OK(pdb->getsampwsid(pdb, sam_acct, sid));
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
static NTSTATUS pdb_default_create_user(struct pdb_methods *methods,
TALLOC_CTX *tmp_ctx, const char *name,
uint32 acb_info, uint32 *rid)
{
struct samu *sam_pass;
NTSTATUS status;
struct passwd *pwd;
if ((sam_pass = samu_new(tmp_ctx)) == NULL) {
return NT_STATUS_NO_MEMORY;
}
if ( !(pwd = Get_Pwnam_alloc(tmp_ctx, name)) ) {
char *add_script = NULL;
int add_ret;
fstring name2;
if ((acb_info & ACB_NORMAL) && name[strlen(name)-1] != '$') {
add_script = talloc_strdup(tmp_ctx,
lp_adduser_script());
} else {
add_script = talloc_strdup(tmp_ctx,
lp_addmachine_script());
}
if (!add_script || add_script[0] == '\0') {
DEBUG(3, ("Could not find user %s and no add script "
"defined\n", name));
return NT_STATUS_NO_SUCH_USER;
}
/* lowercase the username before creating the Unix account for
compatibility with previous Samba releases */
fstrcpy( name2, name );
strlower_m( name2 );
add_script = talloc_all_string_sub(tmp_ctx,
add_script,
"%u",
name2);
if (!add_script) {
return NT_STATUS_NO_MEMORY;
}
add_ret = smbrun(add_script,NULL);
DEBUG(add_ret ? 0 : 3, ("_samr_create_user: Running the command `%s' gave %d\n",
add_script, add_ret));
if (add_ret == 0) {
smb_nscd_flush_user_cache();
}
flush_pwnam_cache();
pwd = Get_Pwnam_alloc(tmp_ctx, name);
}
/* we have a valid SID coming out of this call */
status = samu_alloc_rid_unix( sam_pass, pwd );
TALLOC_FREE( pwd );
if (!NT_STATUS_IS_OK(status)) {
DEBUG(3, ("pdb_default_create_user: failed to create a new user structure: %s\n", nt_errstr(status)));
return status;
}
if (!sid_peek_check_rid(get_global_sam_sid(),
pdb_get_user_sid(sam_pass), rid)) {
DEBUG(0, ("Could not get RID of fresh user\n"));
return NT_STATUS_INTERNAL_ERROR;
}
/* Use the username case specified in the original request */
pdb_set_username( sam_pass, name, PDB_SET );
/* Disable the account on creation, it does not have a reasonable password yet. */
acb_info |= ACB_DISABLED;
pdb_set_acct_ctrl(sam_pass, acb_info, PDB_CHANGED);
status = pdb_add_sam_account(sam_pass);
TALLOC_FREE(sam_pass);
return status;
}
NTSTATUS pdb_create_user(TALLOC_CTX *mem_ctx, const char *name, uint32 flags,
uint32 *rid)
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->create_user(pdb, mem_ctx, name, flags, rid);
}
/****************************************************************************
Delete a UNIX user on demand.
****************************************************************************/
static int smb_delete_user(const char *unix_user)
{
char *del_script = NULL;
int ret;
/* safety check */
if ( strequal( unix_user, "root" ) ) {
DEBUG(0,("smb_delete_user: Refusing to delete local system root account!\n"));
return -1;
}
del_script = talloc_strdup(talloc_tos(), lp_deluser_script());
if (!del_script || !*del_script) {
return -1;
}
del_script = talloc_all_string_sub(talloc_tos(),
del_script,
"%u",
unix_user);
if (!del_script) {
return -1;
}
ret = smbrun(del_script,NULL);
flush_pwnam_cache();
if (ret == 0) {
smb_nscd_flush_user_cache();
}
DEBUG(ret ? 0 : 3,("smb_delete_user: Running the command `%s' gave %d\n",del_script,ret));
return ret;
}
static NTSTATUS pdb_default_delete_user(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
struct samu *sam_acct)
{
NTSTATUS status;
fstring username;
status = pdb_delete_sam_account(sam_acct);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
/*
* Now delete the unix side ....
* note: we don't check if the delete really happened as the script is
* not necessary present and maybe the sysadmin doesn't want to delete
* the unix side
*/
/* always lower case the username before handing it off to
external scripts */
fstrcpy( username, pdb_get_username(sam_acct) );
strlower_m( username );
smb_delete_user( username );
return status;
}
NTSTATUS pdb_delete_user(TALLOC_CTX *mem_ctx, struct samu *sam_acct)
{
struct pdb_methods *pdb = pdb_get_methods();
uid_t uid = -1;
/* sanity check to make sure we don't delete root */
if ( !sid_to_uid( pdb_get_user_sid(sam_acct), &uid ) ) {
return NT_STATUS_NO_SUCH_USER;
}
if ( uid == 0 ) {
return NT_STATUS_ACCESS_DENIED;
}
return pdb->delete_user(pdb, mem_ctx, sam_acct);
}
NTSTATUS pdb_add_sam_account(struct samu *sam_acct)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->add_sam_account(pdb, sam_acct);
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
NTSTATUS pdb_update_sam_account(struct samu *sam_acct)
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
{
struct pdb_methods *pdb = pdb_get_methods();
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
memcache_flush(NULL, PDB_GETPWSID_CACHE);
return pdb->update_sam_account(pdb, sam_acct);
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
NTSTATUS pdb_delete_sam_account(struct samu *sam_acct)
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
{
struct pdb_methods *pdb = pdb_get_methods();
memcache_flush(NULL, PDB_GETPWSID_CACHE);
return pdb->delete_sam_account(pdb, sam_acct);
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
NTSTATUS pdb_rename_sam_account(struct samu *oldname, const char *newname)
{
struct pdb_methods *pdb = pdb_get_methods();
uid_t uid;
NTSTATUS status;
memcache_flush(NULL, PDB_GETPWSID_CACHE);
/* sanity check to make sure we don't rename root */
if ( !sid_to_uid( pdb_get_user_sid(oldname), &uid ) ) {
return NT_STATUS_NO_SUCH_USER;
}
if ( uid == 0 ) {
return NT_STATUS_ACCESS_DENIED;
}
status = pdb->rename_sam_account(pdb, oldname, newname);
/* always flush the cache here just to be safe */
flush_pwnam_cache();
return status;
}
NTSTATUS pdb_update_login_attempts(struct samu *sam_acct, bool success)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->update_login_attempts(pdb, sam_acct, success);
}
bool pdb_getgrsid(GROUP_MAP *map, DOM_SID sid)
{
struct pdb_methods *pdb = pdb_get_methods();
return NT_STATUS_IS_OK(pdb->getgrsid(pdb, map, sid));
}
bool pdb_getgrgid(GROUP_MAP *map, gid_t gid)
{
struct pdb_methods *pdb = pdb_get_methods();
return NT_STATUS_IS_OK(pdb->getgrgid(pdb, map, gid));
}
bool pdb_getgrnam(GROUP_MAP *map, const char *name)
{
struct pdb_methods *pdb = pdb_get_methods();
return NT_STATUS_IS_OK(pdb->getgrnam(pdb, map, name));
}
static NTSTATUS pdb_default_create_dom_group(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
const char *name,
uint32 *rid)
{
DOM_SID group_sid;
struct group *grp;
fstring tmp;
grp = getgrnam(name);
if (grp == NULL) {
gid_t gid;
if (smb_create_group(name, &gid) != 0) {
return NT_STATUS_ACCESS_DENIED;
}
grp = getgrgid(gid);
}
if (grp == NULL) {
return NT_STATUS_ACCESS_DENIED;
}
if (pdb_rid_algorithm()) {
*rid = algorithmic_pdb_gid_to_group_rid( grp->gr_gid );
} else {
if (!pdb_new_rid(rid)) {
return NT_STATUS_ACCESS_DENIED;
}
}
sid_compose(&group_sid, get_global_sam_sid(), *rid);
return add_initial_entry(grp->gr_gid, sid_to_fstring(tmp, &group_sid),
SID_NAME_DOM_GRP, name, NULL);
}
NTSTATUS pdb_create_dom_group(TALLOC_CTX *mem_ctx, const char *name,
uint32 *rid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->create_dom_group(pdb, mem_ctx, name, rid);
}
static NTSTATUS pdb_default_delete_dom_group(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
uint32 rid)
{
DOM_SID group_sid;
GROUP_MAP map;
NTSTATUS status;
struct group *grp;
const char *grp_name;
/* coverity */
map.gid = (gid_t) -1;
sid_compose(&group_sid, get_global_sam_sid(), rid);
if (!get_domain_group_from_sid(group_sid, &map)) {
DEBUG(10, ("Could not find group for rid %d\n", rid));
return NT_STATUS_NO_SUCH_GROUP;
}
/* We need the group name for the smb_delete_group later on */
if (map.gid == (gid_t)-1) {
return NT_STATUS_NO_SUCH_GROUP;
}
grp = getgrgid(map.gid);
if (grp == NULL) {
return NT_STATUS_NO_SUCH_GROUP;
}
/* Copy the name, no idea what pdb_delete_group_mapping_entry does.. */
grp_name = talloc_strdup(mem_ctx, grp->gr_name);
if (grp_name == NULL) {
return NT_STATUS_NO_MEMORY;
}
status = pdb_delete_group_mapping_entry(group_sid);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
/* Don't check the result of smb_delete_group */
smb_delete_group(grp_name);
return NT_STATUS_OK;
}
NTSTATUS pdb_delete_dom_group(TALLOC_CTX *mem_ctx, uint32 rid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->delete_dom_group(pdb, mem_ctx, rid);
}
NTSTATUS pdb_add_group_mapping_entry(GROUP_MAP *map)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->add_group_mapping_entry(pdb, map);
}
NTSTATUS pdb_update_group_mapping_entry(GROUP_MAP *map)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->update_group_mapping_entry(pdb, map);
}
NTSTATUS pdb_delete_group_mapping_entry(DOM_SID sid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->delete_group_mapping_entry(pdb, sid);
}
bool pdb_enum_group_mapping(const DOM_SID *sid, enum lsa_SidType sid_name_use, GROUP_MAP **pp_rmap,
size_t *p_num_entries, bool unix_only)
{
struct pdb_methods *pdb = pdb_get_methods();
return NT_STATUS_IS_OK(pdb-> enum_group_mapping(pdb, sid, sid_name_use,
pp_rmap, p_num_entries, unix_only));
}
NTSTATUS pdb_enum_group_members(TALLOC_CTX *mem_ctx,
const DOM_SID *sid,
uint32 **pp_member_rids,
size_t *p_num_members)
{
struct pdb_methods *pdb = pdb_get_methods();
NTSTATUS result;
result = pdb->enum_group_members(pdb, mem_ctx,
sid, pp_member_rids, p_num_members);
/* special check for rid 513 */
if ( !NT_STATUS_IS_OK( result ) ) {
uint32 rid;
sid_peek_rid( sid, &rid );
if ( rid == DOMAIN_GROUP_RID_USERS ) {
*p_num_members = 0;
*pp_member_rids = NULL;
return NT_STATUS_OK;
}
}
return result;
}
NTSTATUS pdb_enum_group_memberships(TALLOC_CTX *mem_ctx, struct samu *user,
DOM_SID **pp_sids, gid_t **pp_gids,
size_t *p_num_groups)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->enum_group_memberships(
pdb, mem_ctx, user,
pp_sids, pp_gids, p_num_groups);
}
static NTSTATUS pdb_default_set_unix_primary_group(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
struct samu *sampass)
{
struct group *grp;
gid_t gid;
if (!sid_to_gid(pdb_get_group_sid(sampass), &gid) ||
(grp = getgrgid(gid)) == NULL) {
return NT_STATUS_INVALID_PRIMARY_GROUP;
}
if (smb_set_primary_group(grp->gr_name,
pdb_get_username(sampass)) != 0) {
return NT_STATUS_ACCESS_DENIED;
}
return NT_STATUS_OK;
}
NTSTATUS pdb_set_unix_primary_group(TALLOC_CTX *mem_ctx, struct samu *user)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->set_unix_primary_group(pdb, mem_ctx, user);
}
/*
* Helper function to see whether a user is in a group. We can't use
* user_in_group_sid here because this creates dependencies only smbd can
* fulfil.
*/
static bool pdb_user_in_group(TALLOC_CTX *mem_ctx, struct samu *account,
const DOM_SID *group_sid)
{
DOM_SID *sids;
gid_t *gids;
size_t i, num_groups;
if (!NT_STATUS_IS_OK(pdb_enum_group_memberships(mem_ctx, account,
&sids, &gids,
&num_groups))) {
return False;
}
for (i=0; i<num_groups; i++) {
if (sid_equal(group_sid, &sids[i])) {
return True;
}
}
return False;
}
static NTSTATUS pdb_default_add_groupmem(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
uint32 group_rid,
uint32 member_rid)
{
DOM_SID group_sid, member_sid;
struct samu *account = NULL;
GROUP_MAP map;
struct group *grp;
struct passwd *pwd;
const char *group_name;
uid_t uid;
/* coverity */
map.gid = (gid_t) -1;
sid_compose(&group_sid, get_global_sam_sid(), group_rid);
sid_compose(&member_sid, get_global_sam_sid(), member_rid);
if (!get_domain_group_from_sid(group_sid, &map) ||
(map.gid == (gid_t)-1) ||
((grp = getgrgid(map.gid)) == NULL)) {
return NT_STATUS_NO_SUCH_GROUP;
}
group_name = talloc_strdup(mem_ctx, grp->gr_name);
if (group_name == NULL) {
return NT_STATUS_NO_MEMORY;
}
if ( !(account = samu_new( NULL )) ) {
return NT_STATUS_NO_MEMORY;
}
if (!pdb_getsampwsid(account, &member_sid) ||
!sid_to_uid(&member_sid, &uid) ||
((pwd = getpwuid_alloc(mem_ctx, uid)) == NULL)) {
return NT_STATUS_NO_SUCH_USER;
}
if (pdb_user_in_group(mem_ctx, account, &group_sid)) {
return NT_STATUS_MEMBER_IN_GROUP;
}
/*
* ok, the group exist, the user exist, the user is not in the group,
* we can (finally) add it to the group !
*/
smb_add_user_group(group_name, pwd->pw_name);
if (!pdb_user_in_group(mem_ctx, account, &group_sid)) {
return NT_STATUS_ACCESS_DENIED;
}
return NT_STATUS_OK;
}
NTSTATUS pdb_add_groupmem(TALLOC_CTX *mem_ctx, uint32 group_rid,
uint32 member_rid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->add_groupmem(pdb, mem_ctx, group_rid, member_rid);
}
static NTSTATUS pdb_default_del_groupmem(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
uint32 group_rid,
uint32 member_rid)
{
DOM_SID group_sid, member_sid;
struct samu *account = NULL;
GROUP_MAP map;
struct group *grp;
struct passwd *pwd;
const char *group_name;
uid_t uid;
sid_compose(&group_sid, get_global_sam_sid(), group_rid);
sid_compose(&member_sid, get_global_sam_sid(), member_rid);
if (!get_domain_group_from_sid(group_sid, &map) ||
(map.gid == (gid_t)-1) ||
((grp = getgrgid(map.gid)) == NULL)) {
return NT_STATUS_NO_SUCH_GROUP;
}
group_name = talloc_strdup(mem_ctx, grp->gr_name);
if (group_name == NULL) {
return NT_STATUS_NO_MEMORY;
}
if ( !(account = samu_new( NULL )) ) {
return NT_STATUS_NO_MEMORY;
}
if (!pdb_getsampwsid(account, &member_sid) ||
!sid_to_uid(&member_sid, &uid) ||
((pwd = getpwuid_alloc(mem_ctx, uid)) == NULL)) {
return NT_STATUS_NO_SUCH_USER;
}
if (!pdb_user_in_group(mem_ctx, account, &group_sid)) {
return NT_STATUS_MEMBER_NOT_IN_GROUP;
}
/*
* ok, the group exist, the user exist, the user is in the group,
* we can (finally) delete it from the group!
*/
smb_delete_user_group(group_name, pwd->pw_name);
if (pdb_user_in_group(mem_ctx, account, &group_sid)) {
return NT_STATUS_ACCESS_DENIED;
}
return NT_STATUS_OK;
}
NTSTATUS pdb_del_groupmem(TALLOC_CTX *mem_ctx, uint32 group_rid,
uint32 member_rid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->del_groupmem(pdb, mem_ctx, group_rid, member_rid);
}
NTSTATUS pdb_create_alias(const char *name, uint32 *rid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->create_alias(pdb, name, rid);
}
NTSTATUS pdb_delete_alias(const DOM_SID *sid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->delete_alias(pdb, sid);
}
NTSTATUS pdb_get_aliasinfo(const DOM_SID *sid, struct acct_info *info)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->get_aliasinfo(pdb, sid, info);
}
NTSTATUS pdb_set_aliasinfo(const DOM_SID *sid, struct acct_info *info)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->set_aliasinfo(pdb, sid, info);
}
NTSTATUS pdb_add_aliasmem(const DOM_SID *alias, const DOM_SID *member)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->add_aliasmem(pdb, alias, member);
}
NTSTATUS pdb_del_aliasmem(const DOM_SID *alias, const DOM_SID *member)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->del_aliasmem(pdb, alias, member);
}
NTSTATUS pdb_enum_aliasmem(const DOM_SID *alias,
DOM_SID **pp_members, size_t *p_num_members)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->enum_aliasmem(pdb, alias, pp_members, p_num_members);
}
NTSTATUS pdb_enum_alias_memberships(TALLOC_CTX *mem_ctx,
const DOM_SID *domain_sid,
const DOM_SID *members, size_t num_members,
uint32 **pp_alias_rids,
size_t *p_num_alias_rids)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->enum_alias_memberships(pdb, mem_ctx,
domain_sid,
members, num_members,
pp_alias_rids,
p_num_alias_rids);
}
NTSTATUS pdb_lookup_rids(const DOM_SID *domain_sid,
int num_rids,
uint32 *rids,
const char **names,
enum lsa_SidType *attrs)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->lookup_rids(pdb, domain_sid, num_rids, rids, names, attrs);
}
/*
* NOTE: pdb_lookup_names is currently (2007-01-12) not used anywhere
* in the samba code.
* Unlike _lsa_lookup_sids and _samr_lookup_rids, which eventually
* also ask pdb_lookup_rids, thus looking up a bunch of rids at a time,
* the pdb_ calls _lsa_lookup_names and _samr_lookup_names come
* down to are pdb_getsampwnam and pdb_getgrnam instead of
* pdb_lookup_names.
* But in principle, it the call belongs to the API and might get
* used in this context some day.
*/
#if 0
NTSTATUS pdb_lookup_names(const DOM_SID *domain_sid,
int num_names,
const char **names,
uint32 *rids,
enum lsa_SidType *attrs)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->lookup_names(pdb, domain_sid, num_names, names, rids, attrs);
}
#endif
bool pdb_get_account_policy(int policy_index, uint32 *value)
{
struct pdb_methods *pdb = pdb_get_methods();
NTSTATUS status;
become_root();
status = pdb->get_account_policy(pdb, policy_index, value);
unbecome_root();
return NT_STATUS_IS_OK(status);
}
bool pdb_set_account_policy(int policy_index, uint32 value)
{
struct pdb_methods *pdb = pdb_get_methods();
NTSTATUS status;
become_root();
status = pdb->set_account_policy(pdb, policy_index, value);
unbecome_root();
return NT_STATUS_IS_OK(status);
}
bool pdb_get_seq_num(time_t *seq_num)
{
struct pdb_methods *pdb = pdb_get_methods();
return NT_STATUS_IS_OK(pdb->get_seq_num(pdb, seq_num));
}
bool pdb_uid_to_rid(uid_t uid, uint32 *rid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->uid_to_rid(pdb, uid, rid);
}
bool pdb_uid_to_sid(uid_t uid, DOM_SID *sid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->uid_to_sid(pdb, uid, sid);
}
bool pdb_gid_to_sid(gid_t gid, DOM_SID *sid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->gid_to_sid(pdb, gid, sid);
}
bool pdb_sid_to_id(const DOM_SID *sid, union unid_t *id,
enum lsa_SidType *type)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->sid_to_id(pdb, sid, id, type);
}
bool pdb_rid_algorithm(void)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->rid_algorithm(pdb);
}
/********************************************************************
Allocate a new RID from the passdb backend. Verify that it is free
by calling lookup_global_sam_rid() to verify that the RID is not
in use. This handles servers that have existing users or groups
with add RIDs (assigned from previous algorithmic mappings)
********************************************************************/
bool pdb_new_rid(uint32 *rid)
{
struct pdb_methods *pdb = pdb_get_methods();
const char *name = NULL;
enum lsa_SidType type;
uint32 allocated_rid = 0;
int i;
TALLOC_CTX *ctx;
if (pdb_rid_algorithm()) {
DEBUG(0, ("Trying to allocate a RID when algorithmic RIDs "
"are active\n"));
return False;
}
if (algorithmic_rid_base() != BASE_RID) {
DEBUG(0, ("'algorithmic rid base' is set but a passdb backend "
"without algorithmic RIDs is chosen.\n"));
DEBUGADD(0, ("Please map all used groups using 'net groupmap "
"add', set the maximum used RID\n"));
DEBUGADD(0, ("and remove the parameter\n"));
return False;
}
if ( (ctx = talloc_init("pdb_new_rid")) == NULL ) {
DEBUG(0,("pdb_new_rid: Talloc initialization failure\n"));
return False;
}
/* Attempt to get an unused RID (max tires is 250...yes that it is
and arbitrary number I pulkled out of my head). -- jerry */
for ( i=0; allocated_rid==0 && i<250; i++ ) {
/* get a new RID */
if ( !pdb->new_rid(pdb, &allocated_rid) ) {
return False;
}
/* validate that the RID is not in use */
if ( lookup_global_sam_rid( ctx, allocated_rid, &name, &type, NULL ) ) {
allocated_rid = 0;
}
}
TALLOC_FREE( ctx );
if ( allocated_rid == 0 ) {
DEBUG(0,("pdb_new_rid: Failed to find unused RID\n"));
return False;
}
*rid = allocated_rid;
return True;
}
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
/***************************************************************
Initialize the static context (at smbd startup etc).
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
If uninitialised, context will auto-init on first use.
***************************************************************/
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
bool initialize_password_db(bool reload, struct event_context *event_ctx)
{
pdb_event_ctx = event_ctx;
return (pdb_get_methods_reload(reload) != NULL);
This is another *BIG* change... Samba now features a pluggable passdb interface, along the same lines as the one in use in the auth subsystem. In this case, only one backend may be active at a time by the 'normal' interface, and only one backend per passdb_context is permitted outside that. This pluggable interface is designed to allow any number of passdb backends to be compiled in, with the selection at runtime. The 'passdb backend' paramater has been created (and documented!) to support this. As such, configure has been modfied to allow (for example) --with-ldap and the old smbpasswd to be selected at the same time. This patch also introduces two new backends: smbpasswd_nua and tdbsam_nua. These two backends accept 'non unix accounts', where the user does *not* exist in /etc/passwd. These accounts' don't have UIDs in the unix sense, but to avoid conflicts in the algroitmic mapping of RIDs, they use the values specified in the 'non unix account range' paramter - in the same way as the winbind ranges are specifed. While I was at it, I cleaned up some of the code in pdb_tdb (code copied directly from smbpasswd and not really considered properly). Most of this was to do with % macro expansion on stored data. It isn't easy to get the macros into the tdb, and the first password change will 'expand' them. tdbsam needs to use a similar system to pdb_ldap in this regard. This patch only makes minor adjustments to pdb_nisplus and pdb_ldap, becouse I don't have the test facilities for these. I plan to incoroprate at least pdb_ldap into this scheme after consultation with Jerry. Each (converted) passdb module now no longer has any 'static' variables, and only exports 1 init function outside its .c file. The non-unix-account support in this patch has been proven! It is now possible to join a win2k machine to a Samba PDC without an account in /etc/passwd! Other changes: Minor interface adjustments: pdb_delete_sam_account() now takes a SAM_ACCOUNT, not a char*. pdb_update_sam_account() no longer takes the 'override' argument that was being ignored so often (every other passdb backend). Extra checks have been added in some places. Minor code changes: smbpasswd no longer attempts to initialise the passdb at startup, this is now done on first use. pdbedit has lost some of its 'machine account' logic, as this behaviour is now controlled by the passdb subsystem directly. The samr subsystem no longer calls 'local password change', but does the pdb interactions directly. This allow the ACB_ flags specifed to be transferred direct to the backend, without interference. Doco: I've updated the doco to reflect some of the changes, and removed some paramters no longer applicable to HEAD. (This used to be commit ff354c99c585068af6dc1ff35a1f109a806b326b)
2002-01-20 17:30:58 +03:00
}
/***************************************************************************
Default implementations of some functions.
****************************************************************************/
static NTSTATUS pdb_default_getsampwnam (struct pdb_methods *methods, struct samu *user, const char *sname)
{
return NT_STATUS_NO_SUCH_USER;
}
static NTSTATUS pdb_default_getsampwsid(struct pdb_methods *my_methods, struct samu * user, const DOM_SID *sid)
{
return NT_STATUS_NO_SUCH_USER;
}
static NTSTATUS pdb_default_add_sam_account (struct pdb_methods *methods, struct samu *newpwd)
{
return NT_STATUS_NOT_IMPLEMENTED;
}
static NTSTATUS pdb_default_update_sam_account (struct pdb_methods *methods, struct samu *newpwd)
{
return NT_STATUS_NOT_IMPLEMENTED;
}
static NTSTATUS pdb_default_delete_sam_account (struct pdb_methods *methods, struct samu *pwd)
{
return NT_STATUS_NOT_IMPLEMENTED;
}
static NTSTATUS pdb_default_rename_sam_account (struct pdb_methods *methods, struct samu *pwd, const char *newname)
{
return NT_STATUS_NOT_IMPLEMENTED;
}
static NTSTATUS pdb_default_update_login_attempts (struct pdb_methods *methods, struct samu *newpwd, bool success)
{
/* Only the pdb_nds backend implements this, by
* default just return ok. */
return NT_STATUS_OK;
}
static NTSTATUS pdb_default_get_account_policy(struct pdb_methods *methods, int policy_index, uint32 *value)
{
return account_policy_get(policy_index, value) ? NT_STATUS_OK : NT_STATUS_UNSUCCESSFUL;
}
static NTSTATUS pdb_default_set_account_policy(struct pdb_methods *methods, int policy_index, uint32 value)
{
return account_policy_set(policy_index, value) ? NT_STATUS_OK : NT_STATUS_UNSUCCESSFUL;
}
static NTSTATUS pdb_default_get_seq_num(struct pdb_methods *methods, time_t *seq_num)
{
*seq_num = time(NULL);
return NT_STATUS_OK;
}
static bool pdb_default_uid_to_sid(struct pdb_methods *methods, uid_t uid,
DOM_SID *sid)
{
struct samu *sampw = NULL;
struct passwd *unix_pw;
bool ret;
unix_pw = sys_getpwuid( uid );
if ( !unix_pw ) {
DEBUG(4,("pdb_default_uid_to_rid: host has no idea of uid "
"%lu\n", (unsigned long)uid));
return False;
}
if ( !(sampw = samu_new( NULL )) ) {
DEBUG(0,("pdb_default_uid_to_rid: samu_new() failed!\n"));
return False;
}
become_root();
ret = NT_STATUS_IS_OK(
methods->getsampwnam(methods, sampw, unix_pw->pw_name ));
unbecome_root();
if (!ret) {
DEBUG(5, ("pdb_default_uid_to_rid: Did not find user "
"%s (%u)\n", unix_pw->pw_name, (unsigned int)uid));
TALLOC_FREE(sampw);
return False;
}
sid_copy(sid, pdb_get_user_sid(sampw));
TALLOC_FREE(sampw);
return True;
}
static bool pdb_default_uid_to_rid(struct pdb_methods *methods, uid_t uid,
uint32 *rid)
{
DOM_SID sid;
bool ret;
ret = pdb_default_uid_to_sid(methods, uid, &sid);
if (!ret) {
return ret;
}
ret = sid_peek_check_rid(get_global_sam_sid(), &sid, rid);
if (!ret) {
DEBUG(1, ("Could not peek rid out of sid %s\n",
sid_string_dbg(&sid)));
}
return ret;
}
static bool pdb_default_gid_to_sid(struct pdb_methods *methods, gid_t gid,
DOM_SID *sid)
{
GROUP_MAP map;
if (!NT_STATUS_IS_OK(methods->getgrgid(methods, &map, gid))) {
return False;
}
sid_copy(sid, &map.sid);
return True;
}
static bool pdb_default_sid_to_id(struct pdb_methods *methods,
const DOM_SID *sid,
union unid_t *id, enum lsa_SidType *type)
{
TALLOC_CTX *mem_ctx;
bool ret = False;
const char *name;
uint32 rid;
mem_ctx = talloc_new(NULL);
if (mem_ctx == NULL) {
DEBUG(0, ("talloc_new failed\n"));
return False;
}
if (sid_peek_check_rid(get_global_sam_sid(), sid, &rid)) {
/* Here we might have users as well as groups and aliases */
ret = lookup_global_sam_rid(mem_ctx, rid, &name, type, id);
goto done;
}
/* check for "Unix User" */
if ( sid_peek_check_rid(&global_sid_Unix_Users, sid, &rid) ) {
id->uid = rid;
*type = SID_NAME_USER;
ret = True;
goto done;
}
/* check for "Unix Group" */
if ( sid_peek_check_rid(&global_sid_Unix_Groups, sid, &rid) ) {
id->gid = rid;
*type = SID_NAME_ALIAS;
ret = True;
goto done;
}
/* BUILTIN */
if (sid_check_is_in_builtin(sid) ||
sid_check_is_in_wellknown_domain(sid)) {
/* Here we only have aliases */
GROUP_MAP map;
if (!NT_STATUS_IS_OK(methods->getgrsid(methods, &map, *sid))) {
DEBUG(10, ("Could not find map for sid %s\n",
sid_string_dbg(sid)));
goto done;
}
if ((map.sid_name_use != SID_NAME_ALIAS) &&
(map.sid_name_use != SID_NAME_WKN_GRP)) {
DEBUG(10, ("Map for sid %s is a %s, expected an "
"alias\n", sid_string_dbg(sid),
sid_type_lookup(map.sid_name_use)));
goto done;
}
id->gid = map.gid;
*type = SID_NAME_ALIAS;
ret = True;
goto done;
}
DEBUG(5, ("Sid %s is neither ours, a Unix SID, nor builtin\n",
sid_string_dbg(sid)));
done:
TALLOC_FREE(mem_ctx);
return ret;
}
static bool get_memberuids(TALLOC_CTX *mem_ctx, gid_t gid, uid_t **pp_uids, size_t *p_num)
{
struct group *grp;
char **gr;
struct passwd *pwd;
bool winbind_env;
bool ret = False;
*pp_uids = NULL;
*p_num = 0;
/* We only look at our own sam, so don't care about imported stuff */
winbind_env = winbind_env_set();
(void)winbind_off();
if ((grp = getgrgid(gid)) == NULL) {
/* allow winbindd lookups, but only if they weren't already disabled */
goto done;
}
/* Primary group members */
setpwent();
while ((pwd = getpwent()) != NULL) {
if (pwd->pw_gid == gid) {
if (!add_uid_to_array_unique(mem_ctx, pwd->pw_uid,
pp_uids, p_num)) {
goto done;
}
}
}
endpwent();
/* Secondary group members */
for (gr = grp->gr_mem; (*gr != NULL) && ((*gr)[0] != '\0'); gr += 1) {
struct passwd *pw = getpwnam(*gr);
if (pw == NULL)
continue;
if (!add_uid_to_array_unique(mem_ctx, pw->pw_uid, pp_uids, p_num)) {
goto done;
}
}
ret = True;
done:
/* allow winbindd lookups, but only if they weren't already disabled */
if (!winbind_env) {
(void)winbind_on();
}
return ret;
}
static NTSTATUS pdb_default_enum_group_members(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
const DOM_SID *group,
uint32 **pp_member_rids,
size_t *p_num_members)
{
gid_t gid;
uid_t *uids;
size_t i, num_uids;
*pp_member_rids = NULL;
*p_num_members = 0;
if (!sid_to_gid(group, &gid))
return NT_STATUS_NO_SUCH_GROUP;
if(!get_memberuids(mem_ctx, gid, &uids, &num_uids))
return NT_STATUS_NO_SUCH_GROUP;
if (num_uids == 0)
return NT_STATUS_OK;
*pp_member_rids = TALLOC_ZERO_ARRAY(mem_ctx, uint32, num_uids);
for (i=0; i<num_uids; i++) {
DOM_SID sid;
uid_to_sid(&sid, uids[i]);
if (!sid_check_is_in_our_domain(&sid)) {
DEBUG(5, ("Inconsistent SAM -- group member uid not "
"in our domain\n"));
continue;
}
sid_peek_rid(&sid, &(*pp_member_rids)[*p_num_members]);
*p_num_members += 1;
}
return NT_STATUS_OK;
}
static NTSTATUS pdb_default_enum_group_memberships(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
struct samu *user,
DOM_SID **pp_sids,
gid_t **pp_gids,
size_t *p_num_groups)
{
size_t i;
gid_t gid;
struct passwd *pw;
const char *username = pdb_get_username(user);
/* Ignore the primary group SID. Honor the real Unix primary group.
The primary group SID is only of real use to Windows clients */
if ( !(pw = getpwnam_alloc(mem_ctx, username)) ) {
return NT_STATUS_NO_SUCH_USER;
}
gid = pw->pw_gid;
TALLOC_FREE( pw );
if (!getgroups_unix_user(mem_ctx, username, gid, pp_gids, p_num_groups)) {
return NT_STATUS_NO_SUCH_USER;
}
if (*p_num_groups == 0) {
smb_panic("primary group missing");
}
*pp_sids = TALLOC_ARRAY(mem_ctx, DOM_SID, *p_num_groups);
if (*pp_sids == NULL) {
TALLOC_FREE(*pp_gids);
return NT_STATUS_NO_MEMORY;
}
for (i=0; i<*p_num_groups; i++) {
gid_to_sid(&(*pp_sids)[i], (*pp_gids)[i]);
}
return NT_STATUS_OK;
}
/*******************************************************************
Look up a rid in the SAM we're responsible for (i.e. passdb)
********************************************************************/
static bool lookup_global_sam_rid(TALLOC_CTX *mem_ctx, uint32 rid,
const char **name,
enum lsa_SidType *psid_name_use,
union unid_t *unix_id)
{
struct samu *sam_account = NULL;
GROUP_MAP map;
bool ret;
DOM_SID sid;
*psid_name_use = SID_NAME_UNKNOWN;
DEBUG(5,("lookup_global_sam_rid: looking up RID %u.\n",
(unsigned int)rid));
sid_copy(&sid, get_global_sam_sid());
sid_append_rid(&sid, rid);
/* see if the passdb can help us with the name of the user */
if ( !(sam_account = samu_new( NULL )) ) {
return False;
}
/* BEING ROOT BLLOCK */
become_root();
if (pdb_getsampwsid(sam_account, &sid)) {
struct passwd *pw;
unbecome_root(); /* -----> EXIT BECOME_ROOT() */
*name = talloc_strdup(mem_ctx, pdb_get_username(sam_account));
if (!*name) {
TALLOC_FREE(sam_account);
return False;
}
*psid_name_use = SID_NAME_USER;
TALLOC_FREE(sam_account);
if (unix_id == NULL) {
return True;
}
pw = Get_Pwnam_alloc(talloc_tos(), *name);
if (pw == NULL) {
return False;
}
unix_id->uid = pw->pw_uid;
TALLOC_FREE(pw);
return True;
}
TALLOC_FREE(sam_account);
ret = pdb_getgrsid(&map, sid);
unbecome_root();
/* END BECOME_ROOT BLOCK */
/* do not resolve SIDs to a name unless there is a valid
gid associated with it */
if ( ret && (map.gid != (gid_t)-1) ) {
*name = talloc_strdup(mem_ctx, map.nt_name);
*psid_name_use = map.sid_name_use;
if ( unix_id ) {
unix_id->gid = map.gid;
}
return True;
}
/* Windows will always map RID 513 to something. On a non-domain
controller, this gets mapped to SERVER\None. */
if ( unix_id ) {
DEBUG(5, ("Can't find a unix id for an unmapped group\n"));
return False;
}
if ( rid == DOMAIN_GROUP_RID_USERS ) {
*name = talloc_strdup(mem_ctx, "None" );
*psid_name_use = SID_NAME_DOM_GRP;
return True;
}
return False;
}
static NTSTATUS pdb_default_lookup_rids(struct pdb_methods *methods,
const DOM_SID *domain_sid,
int num_rids,
uint32 *rids,
const char **names,
enum lsa_SidType *attrs)
{
int i;
NTSTATUS result;
bool have_mapped = False;
bool have_unmapped = False;
if (sid_check_is_builtin(domain_sid)) {
for (i=0; i<num_rids; i++) {
const char *name;
if (lookup_builtin_rid(names, rids[i], &name)) {
attrs[i] = SID_NAME_ALIAS;
names[i] = name;
DEBUG(5,("lookup_rids: %s:%d\n",
names[i], attrs[i]));
have_mapped = True;
} else {
have_unmapped = True;
attrs[i] = SID_NAME_UNKNOWN;
}
}
goto done;
}
/* Should not happen, but better check once too many */
if (!sid_check_is_domain(domain_sid)) {
return NT_STATUS_INVALID_HANDLE;
}
for (i = 0; i < num_rids; i++) {
const char *name;
if (lookup_global_sam_rid(names, rids[i], &name, &attrs[i],
NULL)) {
if (name == NULL) {
return NT_STATUS_NO_MEMORY;
}
names[i] = name;
DEBUG(5,("lookup_rids: %s:%d\n", names[i], attrs[i]));
have_mapped = True;
} else {
have_unmapped = True;
attrs[i] = SID_NAME_UNKNOWN;
}
}
done:
result = NT_STATUS_NONE_MAPPED;
if (have_mapped)
result = have_unmapped ? STATUS_SOME_UNMAPPED : NT_STATUS_OK;
return result;
}
#if 0
static NTSTATUS pdb_default_lookup_names(struct pdb_methods *methods,
const DOM_SID *domain_sid,
int num_names,
const char **names,
uint32 *rids,
enum lsa_SidType *attrs)
{
int i;
NTSTATUS result;
bool have_mapped = False;
bool have_unmapped = False;
if (sid_check_is_builtin(domain_sid)) {
for (i=0; i<num_names; i++) {
uint32 rid;
if (lookup_builtin_name(names[i], &rid)) {
attrs[i] = SID_NAME_ALIAS;
rids[i] = rid;
DEBUG(5,("lookup_rids: %s:%d\n",
names[i], attrs[i]));
have_mapped = True;
} else {
have_unmapped = True;
attrs[i] = SID_NAME_UNKNOWN;
}
}
goto done;
}
/* Should not happen, but better check once too many */
if (!sid_check_is_domain(domain_sid)) {
return NT_STATUS_INVALID_HANDLE;
}
for (i = 0; i < num_names; i++) {
if (lookup_global_sam_name(names[i], 0, &rids[i], &attrs[i])) {
DEBUG(5,("lookup_names: %s-> %d:%d\n", names[i],
rids[i], attrs[i]));
have_mapped = True;
} else {
have_unmapped = True;
attrs[i] = SID_NAME_UNKNOWN;
}
}
done:
result = NT_STATUS_NONE_MAPPED;
if (have_mapped)
result = have_unmapped ? STATUS_SOME_UNMAPPED : NT_STATUS_OK;
return result;
}
#endif
static int pdb_search_destructor(struct pdb_search *search)
{
if ((!search->search_ended) && (search->search_end != NULL)) {
search->search_end(search);
}
return 0;
}
struct pdb_search *pdb_search_init(TALLOC_CTX *mem_ctx,
enum pdb_search_type type)
{
struct pdb_search *result;
result = talloc(mem_ctx, struct pdb_search);
if (result == NULL) {
DEBUG(0, ("talloc failed\n"));
return NULL;
}
result->type = type;
result->cache = NULL;
result->num_entries = 0;
result->cache_size = 0;
result->search_ended = False;
result->search_end = NULL;
/* Segfault appropriately if not initialized */
result->next_entry = NULL;
result->search_end = NULL;
talloc_set_destructor(result, pdb_search_destructor);
return result;
}
static void fill_displayentry(TALLOC_CTX *mem_ctx, uint32 rid,
uint16 acct_flags,
const char *account_name,
const char *fullname,
const char *description,
struct samr_displayentry *entry)
{
entry->rid = rid;
entry->acct_flags = acct_flags;
if (account_name != NULL)
entry->account_name = talloc_strdup(mem_ctx, account_name);
else
entry->account_name = "";
if (fullname != NULL)
entry->fullname = talloc_strdup(mem_ctx, fullname);
else
entry->fullname = "";
if (description != NULL)
entry->description = talloc_strdup(mem_ctx, description);
else
entry->description = "";
}
struct group_search {
GROUP_MAP *groups;
size_t num_groups, current_group;
};
static bool next_entry_groups(struct pdb_search *s,
struct samr_displayentry *entry)
{
struct group_search *state = (struct group_search *)s->private_data;
uint32 rid;
GROUP_MAP *map = &state->groups[state->current_group];
if (state->current_group == state->num_groups)
return False;
sid_peek_rid(&map->sid, &rid);
fill_displayentry(s, rid, 0, map->nt_name, NULL, map->comment, entry);
state->current_group += 1;
return True;
}
static void search_end_groups(struct pdb_search *search)
{
struct group_search *state =
(struct group_search *)search->private_data;
SAFE_FREE(state->groups);
}
static bool pdb_search_grouptype(struct pdb_search *search,
const DOM_SID *sid, enum lsa_SidType type)
{
struct group_search *state;
state = talloc(search, struct group_search);
if (state == NULL) {
DEBUG(0, ("talloc failed\n"));
return False;
}
if (!pdb_enum_group_mapping(sid, type, &state->groups, &state->num_groups,
True)) {
DEBUG(0, ("Could not enum groups\n"));
return False;
}
state->current_group = 0;
search->private_data = state;
search->next_entry = next_entry_groups;
search->search_end = search_end_groups;
return True;
}
static bool pdb_default_search_groups(struct pdb_methods *methods,
struct pdb_search *search)
{
return pdb_search_grouptype(search, get_global_sam_sid(), SID_NAME_DOM_GRP);
}
static bool pdb_default_search_aliases(struct pdb_methods *methods,
struct pdb_search *search,
const DOM_SID *sid)
{
return pdb_search_grouptype(search, sid, SID_NAME_ALIAS);
}
static struct samr_displayentry *pdb_search_getentry(struct pdb_search *search,
uint32 idx)
{
if (idx < search->num_entries)
return &search->cache[idx];
if (search->search_ended)
return NULL;
while (idx >= search->num_entries) {
struct samr_displayentry entry;
if (!search->next_entry(search, &entry)) {
search->search_end(search);
search->search_ended = True;
break;
}
ADD_TO_LARGE_ARRAY(search, struct samr_displayentry,
entry, &search->cache, &search->num_entries,
&search->cache_size);
}
return (search->num_entries > idx) ? &search->cache[idx] : NULL;
}
struct pdb_search *pdb_search_users(TALLOC_CTX *mem_ctx, uint32 acct_flags)
{
struct pdb_methods *pdb = pdb_get_methods();
struct pdb_search *result;
result = pdb_search_init(mem_ctx, PDB_USER_SEARCH);
if (result == NULL) {
return NULL;
}
if (!pdb->search_users(pdb, result, acct_flags)) {
TALLOC_FREE(result);
return NULL;
}
return result;
}
struct pdb_search *pdb_search_groups(TALLOC_CTX *mem_ctx)
{
struct pdb_methods *pdb = pdb_get_methods();
struct pdb_search *result;
result = pdb_search_init(mem_ctx, PDB_GROUP_SEARCH);
if (result == NULL) {
return NULL;
}
if (!pdb->search_groups(pdb, result)) {
TALLOC_FREE(result);
return NULL;
}
return result;
}
struct pdb_search *pdb_search_aliases(TALLOC_CTX *mem_ctx, const DOM_SID *sid)
{
struct pdb_methods *pdb = pdb_get_methods();
struct pdb_search *result;
if (pdb == NULL) return NULL;
result = pdb_search_init(mem_ctx, PDB_ALIAS_SEARCH);
if (result == NULL) {
return NULL;
}
if (!pdb->search_aliases(pdb, result, sid)) {
TALLOC_FREE(result);
return NULL;
}
return result;
}
uint32 pdb_search_entries(struct pdb_search *search,
uint32 start_idx, uint32 max_entries,
struct samr_displayentry **result)
{
struct samr_displayentry *end_entry;
uint32 end_idx = start_idx+max_entries-1;
/* The first entry needs to be searched after the last. Otherwise the
* first entry might have moved due to a realloc during the search for
* the last entry. */
end_entry = pdb_search_getentry(search, end_idx);
*result = pdb_search_getentry(search, start_idx);
if (end_entry != NULL)
return max_entries;
if (start_idx >= search->num_entries)
return 0;
return search->num_entries - start_idx;
}
/*******************************************************************
2009-02-12 18:21:45 +03:00
trustdom methods
*******************************************************************/
bool pdb_get_trusteddom_pw(const char *domain, char** pwd, DOM_SID *sid,
time_t *pass_last_set_time)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->get_trusteddom_pw(pdb, domain, pwd, sid,
pass_last_set_time);
}
bool pdb_set_trusteddom_pw(const char* domain, const char* pwd,
const DOM_SID *sid)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->set_trusteddom_pw(pdb, domain, pwd, sid);
}
bool pdb_del_trusteddom_pw(const char *domain)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->del_trusteddom_pw(pdb, domain);
}
NTSTATUS pdb_enum_trusteddoms(TALLOC_CTX *mem_ctx, uint32 *num_domains,
struct trustdom_info ***domains)
{
struct pdb_methods *pdb = pdb_get_methods();
return pdb->enum_trusteddoms(pdb, mem_ctx, num_domains, domains);
}
/*******************************************************************
the defaults for trustdom methods:
these simply call the original passdb/secrets.c actions,
to be replaced by pdb_ldap.
*******************************************************************/
static bool pdb_default_get_trusteddom_pw(struct pdb_methods *methods,
const char *domain,
char** pwd,
DOM_SID *sid,
time_t *pass_last_set_time)
{
return secrets_fetch_trusted_domain_password(domain, pwd,
sid, pass_last_set_time);
}
static bool pdb_default_set_trusteddom_pw(struct pdb_methods *methods,
const char* domain,
const char* pwd,
const DOM_SID *sid)
{
return secrets_store_trusted_domain_password(domain, pwd, sid);
}
static bool pdb_default_del_trusteddom_pw(struct pdb_methods *methods,
const char *domain)
{
return trusted_domain_password_delete(domain);
}
static NTSTATUS pdb_default_enum_trusteddoms(struct pdb_methods *methods,
TALLOC_CTX *mem_ctx,
uint32 *num_domains,
struct trustdom_info ***domains)
{
return secrets_trusted_domains(mem_ctx, num_domains, domains);
}
/*******************************************************************
Create a pdb_methods structure and initialize it with the default
operations. In this way a passdb module can simply implement
the functionality it cares about. However, normally this is done
in groups of related functions.
*******************************************************************/
NTSTATUS make_pdb_method( struct pdb_methods **methods )
{
/* allocate memory for the structure as its own talloc CTX */
2009-05-27 20:40:31 +04:00
*methods = talloc_zero(talloc_autofree_context(), struct pdb_methods);
if (*methods == NULL) {
return NT_STATUS_NO_MEMORY;
}
(*methods)->getsampwnam = pdb_default_getsampwnam;
(*methods)->getsampwsid = pdb_default_getsampwsid;
(*methods)->create_user = pdb_default_create_user;
(*methods)->delete_user = pdb_default_delete_user;
(*methods)->add_sam_account = pdb_default_add_sam_account;
(*methods)->update_sam_account = pdb_default_update_sam_account;
(*methods)->delete_sam_account = pdb_default_delete_sam_account;
(*methods)->rename_sam_account = pdb_default_rename_sam_account;
(*methods)->update_login_attempts = pdb_default_update_login_attempts;
(*methods)->getgrsid = pdb_default_getgrsid;
(*methods)->getgrgid = pdb_default_getgrgid;
(*methods)->getgrnam = pdb_default_getgrnam;
(*methods)->create_dom_group = pdb_default_create_dom_group;
(*methods)->delete_dom_group = pdb_default_delete_dom_group;
(*methods)->add_group_mapping_entry = pdb_default_add_group_mapping_entry;
(*methods)->update_group_mapping_entry = pdb_default_update_group_mapping_entry;
(*methods)->delete_group_mapping_entry = pdb_default_delete_group_mapping_entry;
(*methods)->enum_group_mapping = pdb_default_enum_group_mapping;
(*methods)->enum_group_members = pdb_default_enum_group_members;
(*methods)->enum_group_memberships = pdb_default_enum_group_memberships;
(*methods)->set_unix_primary_group = pdb_default_set_unix_primary_group;
(*methods)->add_groupmem = pdb_default_add_groupmem;
(*methods)->del_groupmem = pdb_default_del_groupmem;
(*methods)->create_alias = pdb_default_create_alias;
(*methods)->delete_alias = pdb_default_delete_alias;
(*methods)->get_aliasinfo = pdb_default_get_aliasinfo;
(*methods)->set_aliasinfo = pdb_default_set_aliasinfo;
(*methods)->add_aliasmem = pdb_default_add_aliasmem;
(*methods)->del_aliasmem = pdb_default_del_aliasmem;
(*methods)->enum_aliasmem = pdb_default_enum_aliasmem;
(*methods)->enum_alias_memberships = pdb_default_alias_memberships;
(*methods)->lookup_rids = pdb_default_lookup_rids;
(*methods)->get_account_policy = pdb_default_get_account_policy;
(*methods)->set_account_policy = pdb_default_set_account_policy;
(*methods)->get_seq_num = pdb_default_get_seq_num;
(*methods)->uid_to_rid = pdb_default_uid_to_rid;
(*methods)->uid_to_sid = pdb_default_uid_to_sid;
(*methods)->gid_to_sid = pdb_default_gid_to_sid;
(*methods)->sid_to_id = pdb_default_sid_to_id;
(*methods)->search_groups = pdb_default_search_groups;
(*methods)->search_aliases = pdb_default_search_aliases;
(*methods)->get_trusteddom_pw = pdb_default_get_trusteddom_pw;
(*methods)->set_trusteddom_pw = pdb_default_set_trusteddom_pw;
(*methods)->del_trusteddom_pw = pdb_default_del_trusteddom_pw;
(*methods)->enum_trusteddoms = pdb_default_enum_trusteddoms;
return NT_STATUS_OK;
}