1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-31 17:18:04 +03:00
samba-mirror/source3/rpc_server/srv_samr_nt.c

3715 lines
102 KiB
C
Raw Normal View History

/*
* Unix SMB/CIFS implementation.
* RPC Pipe client / server routines
* Copyright (C) Andrew Tridgell 1992-1997,
* Copyright (C) Luke Kenneth Casson Leighton 1996-1997,
* Copyright (C) Paul Ashton 1997.
* Copyright (C) Marc Jacobsen 1999.
* Copyright (C) Jeremy Allison 2001-2002.
* Copyright (C) Jean Fran<EFBFBD>ois Micouleau 1998-2001.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
/*
* This is the implementation of the SAMR code.
*/
#include "includes.h"
extern fstring global_myworkgroup;
extern pstring global_myname;
extern DOM_SID global_sam_sid;
extern DOM_SID global_sid_Builtin;
extern rid_name domain_group_rids[];
extern rid_name domain_alias_rids[];
extern rid_name builtin_alias_rids[];
typedef struct _disp_info {
BOOL user_dbloaded;
uint32 num_user_account;
DISP_USER_INFO *disp_user_info;
BOOL group_dbloaded;
uint32 num_group_account;
DISP_GROUP_INFO *disp_group_info;
} DISP_INFO;
struct samr_info {
/* for use by the \PIPE\samr policy */
DOM_SID sid;
uint32 status; /* some sort of flag. best to record it. comes from opnum 0x39 */
DISP_INFO disp_info;
};
/*******************************************************************
Create a samr_info struct.
********************************************************************/
static struct samr_info *get_samr_info_by_sid(DOM_SID *psid)
{
struct samr_info *info;
fstring sid_str;
if ((info = (struct samr_info *)malloc(sizeof(struct samr_info))) == NULL)
return NULL;
ZERO_STRUCTP(info);
if (psid) {
DEBUG(10,("get_samr_info_by_sid: created new info for sid %s\n", sid_to_string(sid_str, psid) ));
sid_copy( &info->sid, psid);
} else {
DEBUG(10,("get_samr_info_by_sid: created new info for NULL sid.\n"));
}
return info;
}
/*******************************************************************
Function to free the per handle data.
********************************************************************/
static void free_samr_db(struct samr_info *info)
{
int i;
if (info->disp_info.group_dbloaded) {
for (i=0; i<info->disp_info.num_group_account; i++)
SAFE_FREE(info->disp_info.disp_group_info[i].grp);
SAFE_FREE(info->disp_info.disp_group_info);
}
if (info->disp_info.user_dbloaded){
for (i=0; i<info->disp_info.num_user_account; i++)
pdb_free_sam(&info->disp_info.disp_user_info[i].sam);
SAFE_FREE(info->disp_info.disp_user_info);
}
info->disp_info.user_dbloaded=False;
info->disp_info.group_dbloaded=False;
info->disp_info.num_group_account=0;
info->disp_info.num_user_account=0;
}
static void free_samr_info(void *ptr)
{
struct samr_info *info=(struct samr_info *) ptr;
free_samr_db(info);
SAFE_FREE(info);
}
/*******************************************************************
Ensure password info is never given out. Paranioa... JRA.
********************************************************************/
static void samr_clear_passwd_fields( SAM_USER_INFO_21 *pass, int num_entries)
{
int i;
if (!pass)
return;
for (i = 0; i < num_entries; i++) {
memset(&pass[i].lm_pwd, '\0', sizeof(pass[i].lm_pwd));
memset(&pass[i].nt_pwd, '\0', sizeof(pass[i].nt_pwd));
}
}
static void samr_clear_sam_passwd(SAM_ACCOUNT *sam_pass)
{
if (!sam_pass)
return;
/* These now zero out the old password */
pdb_set_lanman_passwd(sam_pass, NULL);
pdb_set_nt_passwd(sam_pass, NULL);
}
static NTSTATUS load_sampwd_entries(struct samr_info *info, uint16 acb_mask)
{
SAM_ACCOUNT *pwd = NULL;
DISP_USER_INFO *pwd_array = NULL;
DEBUG(10,("load_sampwd_entries\n"));
/* if the snapshoot is already loaded, return */
if (info->disp_info.user_dbloaded==True) {
DEBUG(10,("load_sampwd_entries: already in memory\n"));
return NT_STATUS_OK;
}
if (!pdb_setsampwent(False)) {
DEBUG(0, ("load_sampwd_entries: Unable to open passdb.\n"));
return NT_STATUS_ACCESS_DENIED;
}
for (pdb_init_sam(&pwd); pdb_getsampwent(pwd) == True; pwd=NULL, pdb_init_sam(&pwd) ) {
if (acb_mask != 0 && !(pdb_get_acct_ctrl(pwd) & acb_mask)) {
pdb_free_sam(&pwd);
DEBUG(5,(" acb_mask %x reject\n", acb_mask));
continue;
}
/* Realloc some memory for the array of ptr to the SAM_ACCOUNT structs */
if (info->disp_info.num_user_account % MAX_SAM_ENTRIES == 0) {
DEBUG(10,("load_sampwd_entries: allocating more memory\n"));
pwd_array=(DISP_USER_INFO *)Realloc(info->disp_info.disp_user_info,
(info->disp_info.num_user_account+MAX_SAM_ENTRIES)*sizeof(DISP_USER_INFO));
if (pwd_array==NULL)
return NT_STATUS_NO_MEMORY;
info->disp_info.disp_user_info=pwd_array;
}
/* link the SAM_ACCOUNT to the array */
info->disp_info.disp_user_info[info->disp_info.num_user_account].sam=pwd;
DEBUG(10,("load_sampwd_entries: entry: %d\n", info->disp_info.num_user_account));
info->disp_info.num_user_account++;
}
pdb_endsampwent();
/* the snapshoot is in memory, we're ready to enumerate fast */
info->disp_info.user_dbloaded=True;
DEBUG(12,("load_sampwd_entries: done\n"));
return NT_STATUS_OK;
}
static NTSTATUS load_group_domain_entries(struct samr_info *info, DOM_SID *sid)
{
GROUP_MAP *map=NULL;
DISP_GROUP_INFO *grp_array = NULL;
uint32 group_entries = 0;
uint32 i;
DEBUG(10,("load_group_domain_entries\n"));
/* if the snapshoot is already loaded, return */
if (info->disp_info.group_dbloaded==True) {
DEBUG(10,("load_group_domain_entries: already in memory\n"));
return NT_STATUS_OK;
}
enum_group_mapping(SID_NAME_DOM_GRP, &map, (int *)&group_entries, ENUM_ONLY_MAPPED, MAPPING_WITHOUT_PRIV);
info->disp_info.num_group_account=group_entries;
grp_array=(DISP_GROUP_INFO *)malloc(info->disp_info.num_group_account*sizeof(DISP_GROUP_INFO));
if (group_entries!=0 && grp_array==NULL) {
SAFE_FREE(map);
return NT_STATUS_NO_MEMORY;
}
info->disp_info.disp_group_info=grp_array;
for (i=0; i<group_entries; i++) {
grp_array[i].grp=(DOMAIN_GRP *)malloc(sizeof(DOMAIN_GRP));
fstrcpy(grp_array[i].grp->name, map[i].nt_name);
fstrcpy(grp_array[i].grp->comment, map[i].comment);
sid_split_rid(&map[i].sid, &grp_array[i].grp->rid);
grp_array[i].grp->attr=SID_NAME_DOM_GRP;
}
SAFE_FREE(map);
/* the snapshoot is in memory, we're ready to enumerate fast */
info->disp_info.group_dbloaded=True;
DEBUG(12,("load_group_domain_entries: done\n"));
return NT_STATUS_OK;
}
/*******************************************************************
This next function should be replaced with something that
dynamically returns the correct user info..... JRA.
********************************************************************/
static NTSTATUS get_sampwd_entries(SAM_USER_INFO_21 *pw_buf, int start_idx,
int *total_entries, int *num_entries,
int max_num_entries, uint16 acb_mask)
{
SAM_ACCOUNT *pwd = NULL;
BOOL not_finished = True;
(*num_entries) = 0;
(*total_entries) = 0;
if (pw_buf == NULL)
return NT_STATUS_NO_MEMORY;
pdb_init_sam(&pwd);
if (!pdb_setsampwent(False)) {
DEBUG(0, ("get_sampwd_entries: Unable to open passdb.\n"));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return NT_STATUS_ACCESS_DENIED;
}
while (((not_finished = pdb_getsampwent(pwd)) != False)
&& (*num_entries) < max_num_entries)
{
int user_name_len;
if (start_idx > 0) {
pdb_reset_sam(pwd);
/* skip the requested number of entries.
not very efficient, but hey... */
start_idx--;
continue;
}
user_name_len = strlen(pdb_get_username(pwd))+1;
init_unistr2(&pw_buf[(*num_entries)].uni_user_name, pdb_get_username(pwd), user_name_len);
init_uni_hdr(&pw_buf[(*num_entries)].hdr_user_name, user_name_len);
pw_buf[(*num_entries)].user_rid = pdb_get_user_rid(pwd);
memset((char *)pw_buf[(*num_entries)].nt_pwd, '\0', 16);
/* Now check if the NT compatible password is available. */
if (pdb_get_nt_passwd(pwd))
memcpy( pw_buf[(*num_entries)].nt_pwd , pdb_get_nt_passwd(pwd), 16);
pw_buf[(*num_entries)].acb_info = pdb_get_acct_ctrl(pwd);
DEBUG(5, ("entry idx: %d user %s, rid 0x%x, acb %x",
(*num_entries), pdb_get_username(pwd), pdb_get_user_rid(pwd), pdb_get_acct_ctrl(pwd) ));
if (acb_mask == 0 || (pdb_get_acct_ctrl(pwd) & acb_mask)) {
DEBUG(5,(" acb_mask %x accepts\n", acb_mask));
(*num_entries)++;
} else {
DEBUG(5,(" acb_mask %x rejects\n", acb_mask));
}
(*total_entries)++;
pdb_reset_sam(pwd);
}
pdb_endsampwent();
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
if (not_finished)
return STATUS_MORE_ENTRIES;
else
return NT_STATUS_OK;
}
/*******************************************************************
_samr_close_hnd
********************************************************************/
NTSTATUS _samr_close_hnd(pipes_struct *p, SAMR_Q_CLOSE_HND *q_u, SAMR_R_CLOSE_HND *r_u)
{
r_u->status = NT_STATUS_OK;
/* close the policy handle */
if (!close_policy_hnd(p, &q_u->pol))
return NT_STATUS_OBJECT_NAME_INVALID;
DEBUG(5,("samr_reply_close_hnd: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
samr_reply_open_domain
********************************************************************/
NTSTATUS _samr_open_domain(pipes_struct *p, SAMR_Q_OPEN_DOMAIN *q_u, SAMR_R_OPEN_DOMAIN *r_u)
{
struct samr_info *info;
r_u->status = NT_STATUS_OK;
/* find the connection policy handle. */
if (!find_policy_by_hnd(p, &q_u->pol, NULL))
return NT_STATUS_INVALID_HANDLE;
/* associate the domain SID with the (unique) handle. */
if ((info = get_samr_info_by_sid(&q_u->dom_sid.sid))==NULL)
return NT_STATUS_NO_MEMORY;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, &r_u->domain_pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
DEBUG(5,("samr_open_domain: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
_samr_get_usrdom_pwinfo
********************************************************************/
NTSTATUS _samr_get_usrdom_pwinfo(pipes_struct *p, SAMR_Q_GET_USRDOM_PWINFO *q_u, SAMR_R_GET_USRDOM_PWINFO *r_u)
{
struct samr_info *info = NULL;
r_u->status = NT_STATUS_OK;
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->user_pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
if (!sid_check_is_in_our_domain(&info->sid))
return NT_STATUS_OBJECT_TYPE_MISMATCH;
init_samr_r_get_usrdom_pwinfo(r_u, NT_STATUS_OK);
DEBUG(5,("_samr_get_usrdom_pwinfo: %d\n", __LINE__));
/*
* NT sometimes return NT_STATUS_ACCESS_DENIED
* I don't know yet why.
*/
return r_u->status;
}
/*******************************************************************
samr_make_usr_obj_sd
********************************************************************/
static NTSTATUS samr_make_usr_obj_sd(TALLOC_CTX *ctx, SEC_DESC_BUF **buf, DOM_SID *usr_sid)
{
extern DOM_SID global_sid_World;
DOM_SID adm_sid;
DOM_SID act_sid;
SEC_ACE ace[4];
SEC_ACCESS mask;
SEC_ACL *psa = NULL;
SEC_DESC *psd = NULL;
size_t sd_size;
sid_copy(&adm_sid, &global_sid_Builtin);
sid_append_rid(&adm_sid, BUILTIN_ALIAS_RID_ADMINS);
sid_copy(&act_sid, &global_sid_Builtin);
sid_append_rid(&act_sid, BUILTIN_ALIAS_RID_ACCOUNT_OPS);
init_sec_access(&mask, 0x2035b);
init_sec_ace(&ace[0], &global_sid_World, SEC_ACE_TYPE_ACCESS_ALLOWED, mask, 0);
init_sec_access(&mask, 0xf07ff);
init_sec_ace(&ace[1], &adm_sid, SEC_ACE_TYPE_ACCESS_ALLOWED, mask, 0);
init_sec_ace(&ace[2], &act_sid, SEC_ACE_TYPE_ACCESS_ALLOWED, mask, 0);
init_sec_access(&mask,0x20044);
init_sec_ace(&ace[3], usr_sid, SEC_ACE_TYPE_ACCESS_ALLOWED, mask, 0);
if((psa = make_sec_acl(ctx, NT4_ACL_REVISION, 4, ace)) == NULL)
return NT_STATUS_NO_MEMORY;
if((psd = make_sec_desc(ctx, SEC_DESC_REVISION, NULL, NULL, NULL, psa, &sd_size)) == NULL)
return NT_STATUS_NO_MEMORY;
if((*buf = make_sec_desc_buf(ctx, sd_size, psd)) == NULL)
return NT_STATUS_NO_MEMORY;
return NT_STATUS_OK;
}
static BOOL get_lsa_policy_samr_sid(pipes_struct *p, POLICY_HND *pol, DOM_SID *sid)
{
struct samr_info *info = NULL;
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, pol, (void **)&info))
return False;
if (!info)
return False;
*sid = info->sid;
return True;
}
/*******************************************************************
_samr_query_sec_obj
********************************************************************/
NTSTATUS _samr_query_sec_obj(pipes_struct *p, SAMR_Q_QUERY_SEC_OBJ *q_u, SAMR_R_QUERY_SEC_OBJ *r_u)
{
DOM_SID pol_sid;
fstring str_sid;
r_u->status = NT_STATUS_OK;
/* Get the SID. */
if (!get_lsa_policy_samr_sid(p, &q_u->user_pol, &pol_sid))
return NT_STATUS_INVALID_HANDLE;
DEBUG(10,("_samr_query_sec_obj: querying security on SID: %s\n", sid_to_string(str_sid, &pol_sid)));
r_u->status = samr_make_usr_obj_sd(p->mem_ctx, &r_u->buf, &pol_sid);
if (NT_STATUS_IS_OK(r_u->status))
r_u->ptr = 1;
return r_u->status;
}
/*******************************************************************
makes a SAM_ENTRY / UNISTR2* structure from a user list.
********************************************************************/
static void make_user_sam_entry_list(TALLOC_CTX *ctx, SAM_ENTRY **sam_pp, UNISTR2 **uni_name_pp,
uint32 num_sam_entries, SAM_USER_INFO_21 *pass)
{
uint32 i;
SAM_ENTRY *sam;
UNISTR2 *uni_name;
*sam_pp = NULL;
*uni_name_pp = NULL;
if (num_sam_entries == 0)
return;
sam = (SAM_ENTRY *)talloc_zero(ctx, sizeof(SAM_ENTRY)*num_sam_entries);
uni_name = (UNISTR2 *)talloc_zero(ctx, sizeof(UNISTR2)*num_sam_entries);
if (sam == NULL || uni_name == NULL) {
DEBUG(0, ("NULL pointers in SAMR_R_QUERY_DISPINFO\n"));
return;
}
ZERO_STRUCTP(sam);
ZERO_STRUCTP(uni_name);
for (i = 0; i < num_sam_entries; i++) {
int len = pass[i].uni_user_name.uni_str_len;
init_sam_entry(&sam[i], len, pass[i].user_rid);
copy_unistr2(&uni_name[i], &pass[i].uni_user_name);
}
*sam_pp = sam;
*uni_name_pp = uni_name;
}
/*******************************************************************
samr_reply_enum_dom_users
********************************************************************/
NTSTATUS _samr_enum_dom_users(pipes_struct *p, SAMR_Q_ENUM_DOM_USERS *q_u, SAMR_R_ENUM_DOM_USERS *r_u)
{
SAM_USER_INFO_21 pass[MAX_SAM_ENTRIES];
int num_entries = 0;
int total_entries = 0;
r_u->status = NT_STATUS_OK;
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->pol, NULL))
return NT_STATUS_INVALID_HANDLE;
DEBUG(5,("_samr_enum_dom_users: %d\n", __LINE__));
become_root();
r_u->status = get_sampwd_entries(pass, q_u->start_idx, &total_entries, &num_entries,
MAX_SAM_ENTRIES, q_u->acb_mask);
unbecome_root();
if (NT_STATUS_IS_ERR(r_u->status))
return r_u->status;
samr_clear_passwd_fields(pass, num_entries);
/*
* Note from JRA. total_entries is not being used here. Currently if there is a
* large user base then it looks like NT will enumerate until get_sampwd_entries
* returns False due to num_entries being zero. This will cause an access denied
* return. I don't think this is right and needs further investigation. Note that
* this is also the same in the TNG code (I don't think that has been tested with
* a very large user list as MAX_SAM_ENTRIES is set to 600).
*
* I also think that one of the 'num_entries' return parameters is probably
* the "max entries" parameter - but in the TNG code they're all currently set to the same
* value (again I think this is wrong).
*/
make_user_sam_entry_list(p->mem_ctx, &r_u->sam, &r_u->uni_acct_name, num_entries, pass);
init_samr_r_enum_dom_users(r_u, q_u->start_idx + num_entries, num_entries);
DEBUG(5,("_samr_enum_dom_users: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
makes a SAM_ENTRY / UNISTR2* structure from a group list.
********************************************************************/
static void make_group_sam_entry_list(TALLOC_CTX *ctx, SAM_ENTRY **sam_pp, UNISTR2 **uni_name_pp,
uint32 num_sam_entries, DOMAIN_GRP *grp)
{
uint32 i;
SAM_ENTRY *sam;
UNISTR2 *uni_name;
*sam_pp = NULL;
*uni_name_pp = NULL;
if (num_sam_entries == 0)
return;
sam = (SAM_ENTRY *)talloc_zero(ctx, sizeof(SAM_ENTRY)*num_sam_entries);
uni_name = (UNISTR2 *)talloc_zero(ctx, sizeof(UNISTR2)*num_sam_entries);
if (sam == NULL || uni_name == NULL) {
DEBUG(0, ("NULL pointers in SAMR_R_QUERY_DISPINFO\n"));
return;
}
for (i = 0; i < num_sam_entries; i++) {
/*
* JRA. I think this should include the null. TNG does not.
*/
int len = strlen(grp[i].name)+1;
init_sam_entry(&sam[i], len, grp[i].rid);
init_unistr2(&uni_name[i], grp[i].name, len);
}
*sam_pp = sam;
*uni_name_pp = uni_name;
}
/*******************************************************************
Get the group entries - similar to get_sampwd_entries().
********************************************************************/
static NTSTATUS get_group_alias_entries(TALLOC_CTX *ctx, DOMAIN_GRP **d_grp, DOM_SID *sid, uint32 start_idx,
uint32 *p_num_entries, uint32 max_entries)
{
fstring sid_str;
uint32 num_entries = 0;
int i;
GROUP_MAP smap;
GROUP_MAP *map;
sid_to_string(sid_str, sid);
DEBUG(5, ("get_group_alias_entries: enumerating aliases on SID: %s\n", sid_str));
*p_num_entries = 0;
/* well-known aliases */
if (sid_equal(sid, &global_sid_Builtin) && !lp_hide_local_users()) {
enum_group_mapping(SID_NAME_ALIAS, &map, (int *)&num_entries, ENUM_ONLY_MAPPED, MAPPING_WITHOUT_PRIV);
if (num_entries != 0) {
*d_grp=(DOMAIN_GRP *)talloc_zero(ctx, num_entries*sizeof(DOMAIN_GRP));
if (*d_grp==NULL)
return NT_STATUS_NO_MEMORY;
for(i=0; i<num_entries && i<max_entries; i++) {
fstrcpy((*d_grp)[i].name, map[i+start_idx].nt_name);
sid_split_rid(&map[i+start_idx].sid, &(*d_grp)[i].rid);
}
}
SAFE_FREE(map);
} else if (sid_equal(sid, &global_sam_sid) && !lp_hide_local_users()) {
struct sys_grent *glist;
struct sys_grent *grp;
A nice *big* change to the fundemental way we do things. Samba (ab)uses the returns from getpwnam() a lot - in particular it keeps them around for a long time - often past the next call... This adds a getpwnam_alloc and a getpwuid_alloc to the collection. These function as expected, returning a malloced structure that can be free()ed with passwd_free(&passwd). This patch also cuts down on the number of calls to getpwnam - mostly by taking advantage of the fact that the passdb interface is already case-insensiteve. With this patch most of the recursive cases have been removed (that I know of) and the problems are reduced further by not using the sys_ interface in the new code. This means that pointers to the cache won't be affected. (This is a tempoary HACK, I intend to kill the password cache entirly). The only change I'm a little worried about is the change to rpc_server/srv_samr_nt.c for private groups. In this case we are getting groups from the new group mapping DB. Do we still need to check for private groups? I've toned down the check to a case sensitve match with the new code, but we might be able to kill it entirly. I've also added a make_modifyable_passwd() function, that copies a passwd struct into the form that the old sys_getpw* code provided. As far as I can tell this is only actually used in the pass_check.c crazies, where I moved the final 'special case' for shadow passwords (out of _Get_Pwnam()). The matching case for getpwent() is dealt with already, in lib/util_getent.c Also included in here is a small change to register the [homes] share at vuid creation rather than just in one varient of the session setup. (This picks up the SPNEGO cases). The home directory is now stored on the vuid, and I am hoping this might provide a saner way to do %H substitions. TODO: Kill off remaining Get_Pwnam_Modify calls (they are not needed), change the remaining sys_getpwnam() callers to use getpwnam_alloc() and move Get_Pwnam to return an allocated struct. Andrew Bartlett (This used to be commit 1d86c7f94230bc53daebd4d2cd829da6292e05da)
2002-01-17 11:45:58 +03:00
struct passwd *pw;
gid_t winbind_gid_low, winbind_gid_high;
lp_winbind_gid(&winbind_gid_low, &winbind_gid_high);
/* local aliases */
/* we return the UNIX groups here. This seems to be the right */
/* thing to do, since NT member servers return their local */
/* groups in the same situation. */
/* use getgrent_list() to retrieve the list of groups to avoid
* problems with getgrent possible infinite loop by internal
* libc grent structures overwrites by called functions */
grp = glist = getgrent_list();
if (grp == NULL)
return NT_STATUS_NO_MEMORY;
for (; (num_entries < max_entries) && (grp != NULL); grp = grp->next) {
uint32 trid;
if(!get_group_from_gid(grp->gr_gid, &smap, MAPPING_WITHOUT_PRIV))
continue;
if (smap.sid_name_use!=SID_NAME_ALIAS) {
continue;
}
sid_split_rid(&smap.sid, &trid);
if (!sid_equal(sid, &smap.sid))
continue;
/* Don't return winbind groups as they are not local! */
if ((grp->gr_gid >= winbind_gid_low)&&(grp->gr_gid <= winbind_gid_high)) {
DEBUG(10,("get_group_alias_entries: not returing %s, not local.\n", smap.nt_name ));
continue;
}
/* Don't return user private groups... */
A nice *big* change to the fundemental way we do things. Samba (ab)uses the returns from getpwnam() a lot - in particular it keeps them around for a long time - often past the next call... This adds a getpwnam_alloc and a getpwuid_alloc to the collection. These function as expected, returning a malloced structure that can be free()ed with passwd_free(&passwd). This patch also cuts down on the number of calls to getpwnam - mostly by taking advantage of the fact that the passdb interface is already case-insensiteve. With this patch most of the recursive cases have been removed (that I know of) and the problems are reduced further by not using the sys_ interface in the new code. This means that pointers to the cache won't be affected. (This is a tempoary HACK, I intend to kill the password cache entirly). The only change I'm a little worried about is the change to rpc_server/srv_samr_nt.c for private groups. In this case we are getting groups from the new group mapping DB. Do we still need to check for private groups? I've toned down the check to a case sensitve match with the new code, but we might be able to kill it entirly. I've also added a make_modifyable_passwd() function, that copies a passwd struct into the form that the old sys_getpw* code provided. As far as I can tell this is only actually used in the pass_check.c crazies, where I moved the final 'special case' for shadow passwords (out of _Get_Pwnam()). The matching case for getpwent() is dealt with already, in lib/util_getent.c Also included in here is a small change to register the [homes] share at vuid creation rather than just in one varient of the session setup. (This picks up the SPNEGO cases). The home directory is now stored on the vuid, and I am hoping this might provide a saner way to do %H substitions. TODO: Kill off remaining Get_Pwnam_Modify calls (they are not needed), change the remaining sys_getpwnam() callers to use getpwnam_alloc() and move Get_Pwnam to return an allocated struct. Andrew Bartlett (This used to be commit 1d86c7f94230bc53daebd4d2cd829da6292e05da)
2002-01-17 11:45:58 +03:00
/*
* We used to do a Get_Pwnam() here, but this has been
* trimmed back to the common case for private groups
* to save lookups and to use the _alloc interface.
*
* This also matches the group mapping code
*/
if ((pw = getpwnam_alloc(smap.nt_name)) != 0) {
DEBUG(10,("get_group_alias_entries: not returing %s, clashes with user.\n", smap.nt_name ));
A nice *big* change to the fundemental way we do things. Samba (ab)uses the returns from getpwnam() a lot - in particular it keeps them around for a long time - often past the next call... This adds a getpwnam_alloc and a getpwuid_alloc to the collection. These function as expected, returning a malloced structure that can be free()ed with passwd_free(&passwd). This patch also cuts down on the number of calls to getpwnam - mostly by taking advantage of the fact that the passdb interface is already case-insensiteve. With this patch most of the recursive cases have been removed (that I know of) and the problems are reduced further by not using the sys_ interface in the new code. This means that pointers to the cache won't be affected. (This is a tempoary HACK, I intend to kill the password cache entirly). The only change I'm a little worried about is the change to rpc_server/srv_samr_nt.c for private groups. In this case we are getting groups from the new group mapping DB. Do we still need to check for private groups? I've toned down the check to a case sensitve match with the new code, but we might be able to kill it entirly. I've also added a make_modifyable_passwd() function, that copies a passwd struct into the form that the old sys_getpw* code provided. As far as I can tell this is only actually used in the pass_check.c crazies, where I moved the final 'special case' for shadow passwords (out of _Get_Pwnam()). The matching case for getpwent() is dealt with already, in lib/util_getent.c Also included in here is a small change to register the [homes] share at vuid creation rather than just in one varient of the session setup. (This picks up the SPNEGO cases). The home directory is now stored on the vuid, and I am hoping this might provide a saner way to do %H substitions. TODO: Kill off remaining Get_Pwnam_Modify calls (they are not needed), change the remaining sys_getpwnam() callers to use getpwnam_alloc() and move Get_Pwnam to return an allocated struct. Andrew Bartlett (This used to be commit 1d86c7f94230bc53daebd4d2cd829da6292e05da)
2002-01-17 11:45:58 +03:00
passwd_free(&pw);
continue;
}
for( i = 0; i < num_entries; i++)
if ( (*d_grp)[i].rid == trid )
break;
if ( i < num_entries ) {
continue; /* rid was there, dup! */
}
/* JRA - added this for large group db enumeration... */
if (start_idx > 0) {
/* skip the requested number of entries.
not very efficient, but hey...
*/
start_idx--;
continue;
}
*d_grp=talloc_realloc(ctx,*d_grp, (num_entries+1)*sizeof(DOMAIN_GRP));
if (*d_grp==NULL) {
grent_free(glist);
return NT_STATUS_NO_MEMORY;
}
fstrcpy((*d_grp)[num_entries].name, smap.nt_name);
(*d_grp)[num_entries].rid = trid;
num_entries++;
DEBUG(10,("get_group_alias_entries: added entry %d, rid:%d\n", num_entries, trid));
}
grent_free(glist);
}
*p_num_entries = num_entries;
DEBUG(10,("get_group_alias_entries: returning %d entries\n", *p_num_entries));
if (num_entries >= max_entries)
return STATUS_MORE_ENTRIES;
return NT_STATUS_OK;
}
/*******************************************************************
Get the group entries - similar to get_sampwd_entries().
********************************************************************/
static NTSTATUS get_group_domain_entries(TALLOC_CTX *ctx, DOMAIN_GRP **d_grp, DOM_SID *sid, uint32 start_idx,
uint32 *p_num_entries, uint32 max_entries)
{
GROUP_MAP *map=NULL;
int i;
uint32 group_entries = 0;
uint32 num_entries = 0;
*p_num_entries = 0;
enum_group_mapping(SID_NAME_DOM_GRP, &map, (int *)&group_entries, ENUM_ONLY_MAPPED, MAPPING_WITHOUT_PRIV);
num_entries=group_entries-start_idx;
/* limit the number of entries */
if (num_entries>max_entries) {
DEBUG(5,("Limiting to %d entries\n", max_entries));
num_entries=max_entries;
}
*d_grp=(DOMAIN_GRP *)talloc_zero(ctx, num_entries*sizeof(DOMAIN_GRP));
if (num_entries!=0 && *d_grp==NULL){
SAFE_FREE(map);
return NT_STATUS_NO_MEMORY;
}
for (i=0; i<num_entries; i++) {
fstrcpy((*d_grp)[i].name, map[i+start_idx].nt_name);
fstrcpy((*d_grp)[i].comment, map[i+start_idx].comment);
sid_split_rid(&map[i+start_idx].sid, &(*d_grp)[i].rid);
(*d_grp)[i].attr=SID_NAME_DOM_GRP;
}
SAFE_FREE(map);
*p_num_entries = num_entries;
return NT_STATUS_OK;
}
/*******************************************************************
samr_reply_enum_dom_groups
Only reply with one group - domain admins. This must be fixed for
a real PDC. JRA.
********************************************************************/
NTSTATUS _samr_enum_dom_groups(pipes_struct *p, SAMR_Q_ENUM_DOM_GROUPS *q_u, SAMR_R_ENUM_DOM_GROUPS *r_u)
{
DOMAIN_GRP *grp=NULL;
uint32 num_entries;
DOM_SID sid;
r_u->status = NT_STATUS_OK;
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &sid))
return NT_STATUS_INVALID_HANDLE;
DEBUG(5,("samr_reply_enum_dom_groups: %d\n", __LINE__));
/* the domain group array is being allocated in the function below */
get_group_domain_entries(p->mem_ctx, &grp, &sid, q_u->start_idx, &num_entries, MAX_SAM_ENTRIES);
make_group_sam_entry_list(p->mem_ctx, &r_u->sam, &r_u->uni_grp_name, num_entries, grp);
init_samr_r_enum_dom_groups(r_u, q_u->start_idx, num_entries);
DEBUG(5,("samr_enum_dom_groups: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
samr_reply_enum_dom_aliases
********************************************************************/
NTSTATUS _samr_enum_dom_aliases(pipes_struct *p, SAMR_Q_ENUM_DOM_ALIASES *q_u, SAMR_R_ENUM_DOM_ALIASES *r_u)
{
DOMAIN_GRP *grp=NULL;
uint32 num_entries = 0;
fstring sid_str;
DOM_SID sid;
NTSTATUS status;
r_u->status = NT_STATUS_OK;
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &sid))
return NT_STATUS_INVALID_HANDLE;
sid_to_string(sid_str, &sid);
DEBUG(5,("samr_reply_enum_dom_aliases: sid %s\n", sid_str));
status = get_group_alias_entries(p->mem_ctx, &grp, &sid, q_u->start_idx,
&num_entries, MAX_SAM_ENTRIES);
if (NT_STATUS_IS_ERR(status)) return status;
make_group_sam_entry_list(p->mem_ctx, &r_u->sam, &r_u->uni_grp_name, num_entries, grp);
/*safe_free(grp);*/
init_samr_r_enum_dom_aliases(r_u, q_u->start_idx + num_entries, num_entries);
DEBUG(5,("samr_enum_dom_aliases: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
samr_reply_query_dispinfo
********************************************************************/
NTSTATUS _samr_query_dispinfo(pipes_struct *p, SAMR_Q_QUERY_DISPINFO *q_u, SAMR_R_QUERY_DISPINFO *r_u)
{
struct samr_info *info = NULL;
uint32 struct_size=0x20; /* W2K always reply that, client doesn't care */
uint16 acb_mask;
uint32 max_entries=q_u->max_entries;
uint32 enum_context=q_u->start_idx;
uint32 max_size=q_u->max_size;
SAM_DISPINFO_CTR *ctr;
uint32 temp_size=0, total_data_size=0;
NTSTATUS disp_ret;
uint32 num_account = 0;
enum remote_arch_types ra_type = get_remote_arch();
int max_sam_entries;
max_sam_entries = (ra_type == RA_WIN95) ? MAX_SAM_ENTRIES_W95 : MAX_SAM_ENTRIES_W2K;
DEBUG(5, ("samr_reply_query_dispinfo: %d\n", __LINE__));
r_u->status = NT_STATUS_OK;
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->domain_pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
/*
* calculate how many entries we will return.
* based on
* - the number of entries the client asked
* - our limit on that
* - the starting point (enumeration context)
* - the buffer size the client will accept
*/
/*
* We are a lot more like W2K. Instead of reading the SAM
* each time to find the records we need to send back,
* we read it once and link that copy to the sam handle.
* For large user list (over the MAX_SAM_ENTRIES)
* it's a definitive win.
* second point to notice: between enumerations
* our sam is now the same as it's a snapshoot.
* third point: got rid of the static SAM_USER_21 struct
* no more intermediate.
* con: it uses much more memory, as a full copy is stored
* in memory.
*
* If you want to change it, think twice and think
* of the second point , that's really important.
*
* JFM, 12/20/2001
*/
/* Get what we need from the password database */
if (q_u->switch_level==2)
acb_mask = ACB_WSTRUST;
else
acb_mask = ACB_NORMAL;
/* Get what we need from the password database */
switch (q_u->switch_level) {
case 0x1:
case 0x2:
case 0x4:
become_root();
r_u->status=load_sampwd_entries(info, acb_mask);
unbecome_root();
if (NT_STATUS_IS_ERR(r_u->status)) {
DEBUG(5, ("_samr_query_dispinfo: load_sampwd_entries failed\n"));
return r_u->status;
}
num_account = info->disp_info.num_user_account;
break;
case 0x3:
case 0x5:
r_u->status = load_group_domain_entries(info, &info->sid);
if (NT_STATUS_IS_ERR(r_u->status))
return r_u->status;
num_account = info->disp_info.num_group_account;
break;
default:
DEBUG(0,("_samr_query_dispinfo: Unknown info level (%u)\n", (unsigned int)q_u->switch_level ));
return NT_STATUS_INVALID_INFO_CLASS;
}
/* first limit the number of entries we will return */
if(max_entries > max_sam_entries) {
DEBUG(5, ("samr_reply_query_dispinfo: client requested %d entries, limiting to %d\n", max_entries, max_sam_entries));
max_entries = max_sam_entries;
}
if (enum_context > num_account) {
DEBUG(5, ("samr_reply_query_dispinfo: enumeration handle over total entries\n"));
return NT_STATUS_OK;
}
/* verify we won't overflow */
if (max_entries > num_account-enum_context) {
max_entries = num_account-enum_context;
DEBUG(5, ("samr_reply_query_dispinfo: only %d entries to return\n", max_entries));
}
/* calculate the size and limit on the number of entries we will return */
temp_size=(enum_context+max_entries)*struct_size;
if (temp_size>max_size) {
max_entries=max_size/struct_size;
DEBUG(5, ("samr_reply_query_dispinfo: buffer size limits to only %d entries\n", max_entries));
}
if (!(ctr = (SAM_DISPINFO_CTR *)talloc_zero(p->mem_ctx,sizeof(SAM_DISPINFO_CTR))))
return NT_STATUS_NO_MEMORY;
ZERO_STRUCTP(ctr);
/* Now create reply structure */
switch (q_u->switch_level) {
case 0x1:
if (max_entries) {
if (!(ctr->sam.info1 = (SAM_DISPINFO_1 *)talloc_zero(p->mem_ctx,max_entries*sizeof(SAM_DISPINFO_1))))
return NT_STATUS_NO_MEMORY;
}
disp_ret = init_sam_dispinfo_1(p->mem_ctx, ctr->sam.info1, max_entries, enum_context, info->disp_info.disp_user_info);
if (NT_STATUS_IS_ERR(disp_ret))
return disp_ret;
break;
case 0x2:
if (max_entries) {
if (!(ctr->sam.info2 = (SAM_DISPINFO_2 *)talloc_zero(p->mem_ctx,max_entries*sizeof(SAM_DISPINFO_2))))
return NT_STATUS_NO_MEMORY;
}
disp_ret = init_sam_dispinfo_2(p->mem_ctx, ctr->sam.info2, max_entries, enum_context, info->disp_info.disp_user_info);
if (NT_STATUS_IS_ERR(disp_ret))
return disp_ret;
break;
case 0x3:
if (max_entries) {
if (!(ctr->sam.info3 = (SAM_DISPINFO_3 *)talloc_zero(p->mem_ctx,max_entries*sizeof(SAM_DISPINFO_3))))
return NT_STATUS_NO_MEMORY;
}
disp_ret = init_sam_dispinfo_3(p->mem_ctx, ctr->sam.info3, max_entries, enum_context, info->disp_info.disp_group_info);
if (NT_STATUS_IS_ERR(disp_ret))
return disp_ret;
break;
case 0x4:
if (max_entries) {
if (!(ctr->sam.info4 = (SAM_DISPINFO_4 *)talloc_zero(p->mem_ctx,max_entries*sizeof(SAM_DISPINFO_4))))
return NT_STATUS_NO_MEMORY;
}
disp_ret = init_sam_dispinfo_4(p->mem_ctx, ctr->sam.info4, max_entries, enum_context, info->disp_info.disp_user_info);
if (NT_STATUS_IS_ERR(disp_ret))
return disp_ret;
break;
case 0x5:
if (max_entries) {
if (!(ctr->sam.info5 = (SAM_DISPINFO_5 *)talloc_zero(p->mem_ctx,max_entries*sizeof(SAM_DISPINFO_5))))
return NT_STATUS_NO_MEMORY;
}
disp_ret = init_sam_dispinfo_5(p->mem_ctx, ctr->sam.info5, max_entries, enum_context, info->disp_info.disp_group_info);
if (NT_STATUS_IS_ERR(disp_ret))
return disp_ret;
break;
default:
ctr->sam.info = NULL;
return NT_STATUS_INVALID_INFO_CLASS;
}
/* calculate the total size */
total_data_size=num_account*struct_size;
if (enum_context+max_entries < num_account)
r_u->status = STATUS_MORE_ENTRIES;
DEBUG(5, ("_samr_query_dispinfo: %d\n", __LINE__));
init_samr_r_query_dispinfo(r_u, max_entries, total_data_size, temp_size, q_u->switch_level, ctr, r_u->status);
return r_u->status;
}
/*******************************************************************
samr_reply_query_aliasinfo
********************************************************************/
NTSTATUS _samr_query_aliasinfo(pipes_struct *p, SAMR_Q_QUERY_ALIASINFO *q_u, SAMR_R_QUERY_ALIASINFO *r_u)
{
struct samr_info *info = NULL;
GROUP_MAP map;
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_query_aliasinfo: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
if (!sid_check_is_in_our_domain(&info->sid) &&
!sid_check_is_in_builtin(&info->sid))
return NT_STATUS_OBJECT_TYPE_MISMATCH;
if(!get_local_group_from_sid(info->sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
switch (q_u->switch_level) {
case 1:
r_u->ptr = 1;
r_u->ctr.switch_value1 = 1;
init_samr_alias_info1(&r_u->ctr.alias.info1, map.nt_name, 1, map.comment);
break;
case 3:
r_u->ptr = 1;
r_u->ctr.switch_value1 = 3;
init_samr_alias_info3(&r_u->ctr.alias.info3, map.comment);
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
DEBUG(5,("_samr_query_aliasinfo: %d\n", __LINE__));
return r_u->status;
}
#if 0
/*******************************************************************
samr_reply_lookup_ids
********************************************************************/
uint32 _samr_lookup_ids(pipes_struct *p, SAMR_Q_LOOKUP_IDS *q_u, SAMR_R_LOOKUP_IDS *r_u)
{
uint32 rid[MAX_SAM_ENTRIES];
int num_rids = q_u->num_sids1;
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_lookup_ids: %d\n", __LINE__));
if (num_rids > MAX_SAM_ENTRIES) {
num_rids = MAX_SAM_ENTRIES;
DEBUG(5,("_samr_lookup_ids: truncating entries to %d\n", num_rids));
}
#if 0
int i;
SMB_ASSERT_ARRAY(q_u->uni_user_name, num_rids);
for (i = 0; i < num_rids && status == 0; i++)
{
struct sam_passwd *sam_pass;
fstring user_name;
fstrcpy(user_name, unistrn2(q_u->uni_user_name[i].buffer,
q_u->uni_user_name[i].uni_str_len));
/* find the user account */
become_root();
sam_pass = get_smb21pwd_entry(user_name, 0);
unbecome_root();
if (sam_pass == NULL)
{
status = 0xC0000000 | NT_STATUS_NO_SUCH_USER;
rid[i] = 0;
}
else
{
rid[i] = sam_pass->user_rid;
}
}
#endif
num_rids = 1;
rid[0] = BUILTIN_ALIAS_RID_USERS;
init_samr_r_lookup_ids(&r_u, num_rids, rid, NT_STATUS_OK);
DEBUG(5,("_samr_lookup_ids: %d\n", __LINE__));
return r_u->status;
}
#endif
/*******************************************************************
_samr_lookup_names
********************************************************************/
NTSTATUS _samr_lookup_names(pipes_struct *p, SAMR_Q_LOOKUP_NAMES *q_u, SAMR_R_LOOKUP_NAMES *r_u)
{
uint32 rid[MAX_SAM_ENTRIES];
uint32 local_rid;
enum SID_NAME_USE type[MAX_SAM_ENTRIES];
enum SID_NAME_USE local_type;
int i;
int num_rids = q_u->num_names2;
DOM_SID pol_sid;
fstring sid_str;
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_lookup_names: %d\n", __LINE__));
ZERO_ARRAY(rid);
ZERO_ARRAY(type);
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &pol_sid)) {
init_samr_r_lookup_names(p->mem_ctx, r_u, 0, NULL, NULL, NT_STATUS_OBJECT_TYPE_MISMATCH);
return r_u->status;
}
if (num_rids > MAX_SAM_ENTRIES) {
num_rids = MAX_SAM_ENTRIES;
DEBUG(5,("_samr_lookup_names: truncating entries to %d\n", num_rids));
}
DEBUG(5,("_samr_lookup_names: looking name on SID %s\n", sid_to_string(sid_str, &pol_sid)));
become_root(); /* local_lookup_name can require root privs */
for (i = 0; i < num_rids; i++) {
fstring name;
DOM_SID sid;
r_u->status = NT_STATUS_NONE_MAPPED;
rid [i] = 0xffffffff;
type[i] = SID_NAME_UNKNOWN;
rpcstr_pull(name, q_u->uni_name[i].buffer, sizeof(name), q_u->uni_name[i].uni_str_len*2, 0);
/*
* we are only looking for a name
* the SID we get back can be outside
* the scope of the pol_sid
*
* in clear: it prevents to reply to domain\group: yes
* when only builtin\group exists.
*
* a cleaner code is to add the sid of the domain we're looking in
* to the local_lookup_name function.
*/
if(local_lookup_name(name, &sid, &local_type)) {
sid_split_rid(&sid, &local_rid);
if (sid_equal(&sid, &pol_sid)) {
rid[i]=local_rid;
type[i]=local_type;
r_u->status = NT_STATUS_OK;
}
}
}
unbecome_root();
init_samr_r_lookup_names(p->mem_ctx, r_u, num_rids, rid, (uint32 *)type, r_u->status);
DEBUG(5,("_samr_lookup_names: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
_samr_chgpasswd_user
********************************************************************/
NTSTATUS _samr_chgpasswd_user(pipes_struct *p, SAMR_Q_CHGPASSWD_USER *q_u, SAMR_R_CHGPASSWD_USER *r_u)
{
fstring user_name;
fstring wks;
DEBUG(5,("_samr_chgpasswd_user: %d\n", __LINE__));
r_u->status = NT_STATUS_OK;
rpcstr_pull(user_name, q_u->uni_user_name.buffer, sizeof(user_name), q_u->uni_user_name.uni_str_len*2, 0);
rpcstr_pull(wks, q_u->uni_dest_host.buffer, sizeof(wks), q_u->uni_dest_host.uni_str_len*2,0);
DEBUG(5,("samr_chgpasswd_user: user: %s wks: %s\n", user_name, wks));
/*
* Pass the user through the NT -> unix user mapping
* function.
*/
(void)map_username(user_name);
/*
A nice *big* change to the fundemental way we do things. Samba (ab)uses the returns from getpwnam() a lot - in particular it keeps them around for a long time - often past the next call... This adds a getpwnam_alloc and a getpwuid_alloc to the collection. These function as expected, returning a malloced structure that can be free()ed with passwd_free(&passwd). This patch also cuts down on the number of calls to getpwnam - mostly by taking advantage of the fact that the passdb interface is already case-insensiteve. With this patch most of the recursive cases have been removed (that I know of) and the problems are reduced further by not using the sys_ interface in the new code. This means that pointers to the cache won't be affected. (This is a tempoary HACK, I intend to kill the password cache entirly). The only change I'm a little worried about is the change to rpc_server/srv_samr_nt.c for private groups. In this case we are getting groups from the new group mapping DB. Do we still need to check for private groups? I've toned down the check to a case sensitve match with the new code, but we might be able to kill it entirly. I've also added a make_modifyable_passwd() function, that copies a passwd struct into the form that the old sys_getpw* code provided. As far as I can tell this is only actually used in the pass_check.c crazies, where I moved the final 'special case' for shadow passwords (out of _Get_Pwnam()). The matching case for getpwent() is dealt with already, in lib/util_getent.c Also included in here is a small change to register the [homes] share at vuid creation rather than just in one varient of the session setup. (This picks up the SPNEGO cases). The home directory is now stored on the vuid, and I am hoping this might provide a saner way to do %H substitions. TODO: Kill off remaining Get_Pwnam_Modify calls (they are not needed), change the remaining sys_getpwnam() callers to use getpwnam_alloc() and move Get_Pwnam to return an allocated struct. Andrew Bartlett (This used to be commit 1d86c7f94230bc53daebd4d2cd829da6292e05da)
2002-01-17 11:45:58 +03:00
* UNIX username case mangling not required, pass_oem_change
* is case insensitive.
*/
if (!pass_oem_change(user_name, q_u->lm_newpass.pass, q_u->lm_oldhash.hash,
q_u->nt_newpass.pass, q_u->nt_oldhash.hash))
r_u->status = NT_STATUS_WRONG_PASSWORD;
init_samr_r_chgpasswd_user(r_u, r_u->status);
DEBUG(5,("_samr_chgpasswd_user: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
makes a SAMR_R_LOOKUP_RIDS structure.
********************************************************************/
static BOOL make_samr_lookup_rids(TALLOC_CTX *ctx, uint32 num_names, fstring names[],
UNIHDR **pp_hdr_name, UNISTR2 **pp_uni_name)
{
uint32 i;
UNIHDR *hdr_name=NULL;
UNISTR2 *uni_name=NULL;
*pp_uni_name = NULL;
*pp_hdr_name = NULL;
if (num_names != 0) {
hdr_name = (UNIHDR *)talloc_zero(ctx, sizeof(UNIHDR)*num_names);
if (hdr_name == NULL)
return False;
uni_name = (UNISTR2 *)talloc_zero(ctx,sizeof(UNISTR2)*num_names);
if (uni_name == NULL)
return False;
}
for (i = 0; i < num_names; i++) {
int len = names[i] != NULL ? strlen(names[i]) : 0;
DEBUG(10, ("names[%d]:%s\n", i, names[i]));
init_uni_hdr(&hdr_name[i], len);
init_unistr2(&uni_name[i], names[i], len);
}
*pp_uni_name = uni_name;
*pp_hdr_name = hdr_name;
return True;
}
/*******************************************************************
_samr_lookup_rids
********************************************************************/
NTSTATUS _samr_lookup_rids(pipes_struct *p, SAMR_Q_LOOKUP_RIDS *q_u, SAMR_R_LOOKUP_RIDS *r_u)
{
fstring group_names[MAX_SAM_ENTRIES];
uint32 *group_attrs = NULL;
UNIHDR *hdr_name = NULL;
UNISTR2 *uni_name = NULL;
DOM_SID pol_sid;
int num_rids = q_u->num_rids1;
int i;
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_lookup_rids: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &pol_sid))
return NT_STATUS_INVALID_HANDLE;
if (num_rids > MAX_SAM_ENTRIES) {
num_rids = MAX_SAM_ENTRIES;
DEBUG(5,("_samr_lookup_rids: truncating entries to %d\n", num_rids));
}
if (num_rids) {
if ((group_attrs = (uint32 *)talloc_zero(p->mem_ctx, num_rids * sizeof(uint32))) == NULL)
return NT_STATUS_NO_MEMORY;
}
r_u->status = NT_STATUS_NONE_MAPPED;
become_root(); /* lookup_sid can require root privs */
for (i = 0; i < num_rids; i++) {
fstring tmpname;
fstring domname;
DOM_SID sid;
enum SID_NAME_USE type;
group_attrs[i] = SID_NAME_UNKNOWN;
*group_names[i] = '\0';
if (sid_equal(&pol_sid, &global_sam_sid)) {
sid_copy(&sid, &pol_sid);
sid_append_rid(&sid, q_u->rid[i]);
if (lookup_sid(&sid, domname, tmpname, &type)) {
r_u->status = NT_STATUS_OK;
group_attrs[i] = (uint32)type;
fstrcpy(group_names[i],tmpname);
DEBUG(5,("_samr_lookup_rids: %s:%d\n", group_names[i], group_attrs[i]));
}
}
}
unbecome_root();
if(!make_samr_lookup_rids(p->mem_ctx, num_rids, group_names, &hdr_name, &uni_name))
return NT_STATUS_NO_MEMORY;
init_samr_r_lookup_rids(r_u, num_rids, hdr_name, uni_name, group_attrs);
DEBUG(5,("_samr_lookup_rids: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
_api_samr_open_user. Safe - gives out no passwd info.
********************************************************************/
NTSTATUS _api_samr_open_user(pipes_struct *p, SAMR_Q_OPEN_USER *q_u, SAMR_R_OPEN_USER *r_u)
{
SAM_ACCOUNT *sampass=NULL;
DOM_SID sid;
POLICY_HND domain_pol = q_u->domain_pol;
uint32 user_rid = q_u->user_rid;
POLICY_HND *user_pol = &r_u->user_pol;
struct samr_info *info = NULL;
BOOL ret;
r_u->status = NT_STATUS_OK;
/* find the domain policy handle. */
if (!find_policy_by_hnd(p, &domain_pol, NULL))
return NT_STATUS_INVALID_HANDLE;
pdb_init_sam(&sampass);
become_root();
ret=pdb_getsampwrid(sampass, user_rid);
unbecome_root();
/* check that the RID exists in our domain. */
if (ret == False) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sampass);
return NT_STATUS_NO_SUCH_USER;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sampass);
/* Get the domain SID stored in the domain policy */
if(!get_lsa_policy_samr_sid(p, &domain_pol, &sid))
return NT_STATUS_INVALID_HANDLE;
/* append the user's RID to it */
if(!sid_append_rid(&sid, user_rid))
return NT_STATUS_NO_SUCH_USER;
/* associate the user's SID with the new handle. */
if ((info = get_samr_info_by_sid(&sid)) == NULL)
return NT_STATUS_NO_MEMORY;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, user_pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
return r_u->status;
}
/*************************************************************************
get_user_info_10. Safe. Only gives out acb bits.
*************************************************************************/
static BOOL get_user_info_10(SAM_USER_INFO_10 *id10, uint32 user_rid)
{
SAM_ACCOUNT *smbpass=NULL;
BOOL ret;
if (!pdb_rid_is_user(user_rid)) {
DEBUG(4,("RID 0x%x is not a user RID\n", user_rid));
return False;
}
pdb_init_sam(&smbpass);
become_root();
ret = pdb_getsampwrid(smbpass, user_rid);
unbecome_root();
if (ret==False) {
DEBUG(4,("User 0x%x not found\n", user_rid));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&smbpass);
return False;
}
DEBUG(3,("User:[%s]\n", pdb_get_username(smbpass) ));
ZERO_STRUCTP(id10);
init_sam_user_info10(id10, pdb_get_acct_ctrl(smbpass) );
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&smbpass);
return True;
}
/*************************************************************************
get_user_info_12. OK - this is the killer as it gives out password info.
Ensure that this is only allowed on an encrypted connection with a root
user. JRA.
*************************************************************************/
static NTSTATUS get_user_info_12(pipes_struct *p, SAM_USER_INFO_12 * id12, uint32 user_rid)
{
SAM_ACCOUNT *smbpass=NULL;
BOOL ret;
if (!p->ntlmssp_auth_validated)
return NT_STATUS_ACCESS_DENIED;
if (!(p->ntlmssp_chal_flags & NTLMSSP_NEGOTIATE_SIGN) || !(p->ntlmssp_chal_flags & NTLMSSP_NEGOTIATE_SEAL))
return NT_STATUS_ACCESS_DENIED;
/*
* Do *NOT* do become_root()/unbecome_root() here ! JRA.
*/
pdb_init_sam(&smbpass);
ret = pdb_getsampwrid(smbpass, user_rid);
if (ret == False) {
DEBUG(4, ("User 0x%x not found\n", user_rid));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&smbpass);
return (geteuid() == (uid_t)0) ? NT_STATUS_NO_SUCH_USER : NT_STATUS_ACCESS_DENIED;
}
DEBUG(3,("User:[%s] 0x%x\n", pdb_get_username(smbpass), pdb_get_acct_ctrl(smbpass) ));
if ( pdb_get_acct_ctrl(smbpass) & ACB_DISABLED) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&smbpass);
return NT_STATUS_ACCOUNT_DISABLED;
}
ZERO_STRUCTP(id12);
init_sam_user_info12(id12, pdb_get_lanman_passwd(smbpass), pdb_get_nt_passwd(smbpass));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&smbpass);
return NT_STATUS_OK;
}
/*************************************************************************
get_user_info_20
*************************************************************************/
static BOOL get_user_info_20(SAM_USER_INFO_20 *id20, uint32 user_rid)
{
SAM_ACCOUNT *sampass=NULL;
BOOL ret;
if (!pdb_rid_is_user(user_rid)) {
DEBUG(4,("RID 0x%x is not a user RID\n", user_rid));
return False;
}
pdb_init_sam(&sampass);
become_root();
ret = pdb_getsampwrid(sampass, user_rid);
unbecome_root();
if (ret == False) {
DEBUG(4,("User 0x%x not found\n", user_rid));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sampass);
return False;
}
samr_clear_sam_passwd(sampass);
DEBUG(3,("User:[%s]\n", pdb_get_username(sampass) ));
ZERO_STRUCTP(id20);
init_sam_user_info20A(id20, sampass);
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sampass);
return True;
}
/*************************************************************************
get_user_info_21
*************************************************************************/
static BOOL get_user_info_21(SAM_USER_INFO_21 *id21, uint32 user_rid)
{
SAM_ACCOUNT *sampass=NULL;
BOOL ret;
if (!pdb_rid_is_user(user_rid)) {
DEBUG(4,("RID 0x%x is not a user RID\n", user_rid));
return False;
}
pdb_init_sam(&sampass);
become_root();
ret = pdb_getsampwrid(sampass, user_rid);
unbecome_root();
if (ret == False) {
DEBUG(4,("User 0x%x not found\n", user_rid));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sampass);
return False;
}
samr_clear_sam_passwd(sampass);
DEBUG(3,("User:[%s]\n", pdb_get_username(sampass) ));
ZERO_STRUCTP(id21);
init_sam_user_info21A(id21, sampass);
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sampass);
return True;
}
/*******************************************************************
_samr_query_userinfo
********************************************************************/
NTSTATUS _samr_query_userinfo(pipes_struct *p, SAMR_Q_QUERY_USERINFO *q_u, SAMR_R_QUERY_USERINFO *r_u)
{
SAM_USERINFO_CTR *ctr;
uint32 rid = 0;
struct samr_info *info = NULL;
r_u->status=NT_STATUS_OK;
/* search for the handle */
if (!find_policy_by_hnd(p, &q_u->pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
if (!sid_check_is_in_our_domain(&info->sid))
return NT_STATUS_OBJECT_TYPE_MISMATCH;
sid_peek_rid(&info->sid, &rid);
DEBUG(5,("_samr_query_userinfo: rid:0x%x\n", rid));
ctr = (SAM_USERINFO_CTR *)talloc_zero(p->mem_ctx, sizeof(SAM_USERINFO_CTR));
if (!ctr)
return NT_STATUS_NO_MEMORY;
ZERO_STRUCTP(ctr);
/* ok! user info levels (lots: see MSDEV help), off we go... */
ctr->switch_value = q_u->switch_value;
switch (q_u->switch_value) {
case 0x10:
ctr->info.id10 = (SAM_USER_INFO_10 *)talloc_zero(p->mem_ctx, sizeof(SAM_USER_INFO_10));
if (ctr->info.id10 == NULL)
return NT_STATUS_NO_MEMORY;
if (!get_user_info_10(ctr->info.id10, rid))
return NT_STATUS_NO_SUCH_USER;
break;
#if 0
/* whoops - got this wrong. i think. or don't understand what's happening. */
case 0x11:
{
NTTIME expire;
info = (void *)&id11;
expire.low = 0xffffffff;
expire.high = 0x7fffffff;
ctr->info.id = (SAM_USER_INFO_11 *)talloc_zero(p->mem_ctx,
sizeof
(*ctr->
info.
id11));
ZERO_STRUCTP(ctr->info.id11);
init_sam_user_info11(ctr->info.id11, &expire,
"BROOKFIELDS$", /* name */
0x03ef, /* user rid */
0x201, /* group rid */
0x0080); /* acb info */
break;
}
#endif
case 0x12:
ctr->info.id12 = (SAM_USER_INFO_12 *)talloc_zero(p->mem_ctx, sizeof(SAM_USER_INFO_12));
if (ctr->info.id12 == NULL)
return NT_STATUS_NO_MEMORY;
if (NT_STATUS_IS_ERR(r_u->status = get_user_info_12(p, ctr->info.id12, rid)))
return r_u->status;
break;
case 20:
ctr->info.id20 = (SAM_USER_INFO_20 *)talloc_zero(p->mem_ctx,sizeof(SAM_USER_INFO_20));
if (ctr->info.id20 == NULL)
return NT_STATUS_NO_MEMORY;
if (!get_user_info_20(ctr->info.id20, rid))
return NT_STATUS_NO_SUCH_USER;
break;
case 21:
ctr->info.id21 = (SAM_USER_INFO_21 *)talloc_zero(p->mem_ctx,sizeof(SAM_USER_INFO_21));
if (ctr->info.id21 == NULL)
return NT_STATUS_NO_MEMORY;
if (!get_user_info_21(ctr->info.id21, rid))
return NT_STATUS_NO_SUCH_USER;
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
init_samr_r_query_userinfo(r_u, ctr, r_u->status);
DEBUG(5,("_samr_query_userinfo: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
samr_reply_query_usergroups
********************************************************************/
NTSTATUS _samr_query_usergroups(pipes_struct *p, SAMR_Q_QUERY_USERGROUPS *q_u, SAMR_R_QUERY_USERGROUPS *r_u)
{
SAM_ACCOUNT *sam_pass=NULL;
DOM_GID *gids = NULL;
int num_groups = 0;
uint32 rid;
struct samr_info *info = NULL;
BOOL ret;
/*
* from the SID in the request:
* we should send back the list of DOMAIN GROUPS
* the user is a member of
*
* and only the DOMAIN GROUPS
* no ALIASES !!! neither aliases of the domain
* nor aliases of the builtin SID
*
* JFM, 12/2/2001
*/
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_query_usergroups: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
if (!sid_check_is_in_our_domain(&info->sid))
return NT_STATUS_OBJECT_TYPE_MISMATCH;
sid_peek_rid(&info->sid, &rid);
pdb_init_sam(&sam_pass);
become_root();
ret = pdb_getsampwrid(sam_pass, rid);
unbecome_root();
if (ret == False) {
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_SUCH_USER;
}
if(!get_domain_user_groups(p->mem_ctx, &num_groups, &gids, sam_pass)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_SUCH_GROUP;
}
/* construct the response. lkclXXXX: gids are not copied! */
init_samr_r_query_usergroups(r_u, num_groups, gids, r_u->status);
DEBUG(5,("_samr_query_usergroups: %d\n", __LINE__));
pdb_free_sam(&sam_pass);
return r_u->status;
}
/*******************************************************************
_samr_query_dom_info
********************************************************************/
NTSTATUS _samr_query_dom_info(pipes_struct *p, SAMR_Q_QUERY_DOMAIN_INFO *q_u, SAMR_R_QUERY_DOMAIN_INFO *r_u)
{
struct samr_info *info = NULL;
SAM_UNK_CTR *ctr;
uint32 min_pass_len,pass_hist,flag;
time_t u_expire, u_min_age;
NTTIME nt_expire, nt_min_age;
time_t u_lock_duration, u_reset_time;
NTTIME nt_lock_duration, nt_reset_time;
uint32 lockout;
time_t u_logout;
NTTIME nt_logout;
uint32 num_users=0, num_groups=0, num_aliases=0;
if ((ctr = (SAM_UNK_CTR *)talloc_zero(p->mem_ctx, sizeof(SAM_UNK_CTR))) == NULL)
return NT_STATUS_NO_MEMORY;
ZERO_STRUCTP(ctr);
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_query_dom_info: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->domain_pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
switch (q_u->switch_value) {
case 0x01:
account_policy_get(AP_MIN_PASSWORD_LEN, &min_pass_len);
account_policy_get(AP_PASSWORD_HISTORY, &pass_hist);
account_policy_get(AP_USER_MUST_LOGON_TO_CHG_PASS, &flag);
account_policy_get(AP_MAX_PASSWORD_AGE, (int *)&u_expire);
account_policy_get(AP_MIN_PASSWORD_AGE, (int *)&u_min_age);
unix_to_nt_time_abs(&nt_expire, u_expire);
unix_to_nt_time_abs(&nt_min_age, u_min_age);
init_unk_info1(&ctr->info.inf1, (uint16)min_pass_len, (uint16)pass_hist,
flag, nt_expire, nt_min_age);
break;
case 0x02:
become_root();
r_u->status=load_sampwd_entries(info, ACB_NORMAL);
unbecome_root();
if (NT_STATUS_IS_ERR(r_u->status)) {
DEBUG(5, ("_samr_query_dispinfo: load_sampwd_entries failed\n"));
return r_u->status;
}
num_users=info->disp_info.num_user_account;
free_samr_db(info);
r_u->status=load_group_domain_entries(info, &global_sam_sid);
if (NT_STATUS_IS_ERR(r_u->status)) {
DEBUG(5, ("_samr_query_dispinfo: load_group_domain_entries failed\n"));
return r_u->status;
}
num_groups=info->disp_info.num_group_account;
free_samr_db(info);
/* The time call below is to get a sequence number for the sam. FIXME !!! JRA. */
init_unk_info2(&ctr->info.inf2, global_myworkgroup, global_myname, (uint32) time(NULL),
num_users, num_groups, num_aliases);
break;
case 0x03:
account_policy_get(AP_TIME_TO_LOGOUT, (int *)&u_logout);
unix_to_nt_time_abs(&nt_logout, u_logout);
init_unk_info3(&ctr->info.inf3, nt_logout);
break;
case 0x05:
init_unk_info5(&ctr->info.inf5, global_myname);
break;
case 0x06:
init_unk_info6(&ctr->info.inf6);
break;
case 0x07:
init_unk_info7(&ctr->info.inf7);
break;
case 0x0c:
account_policy_get(AP_LOCK_ACCOUNT_DURATION, (int *)&u_lock_duration);
account_policy_get(AP_RESET_COUNT_TIME, (int *)&u_reset_time);
account_policy_get(AP_BAD_ATTEMPT_LOCKOUT, &lockout);
unix_to_nt_time_abs(&nt_lock_duration, u_lock_duration);
unix_to_nt_time_abs(&nt_reset_time, u_reset_time);
init_unk_info12(&ctr->info.inf12, nt_lock_duration, nt_reset_time, (uint16)lockout);
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
init_samr_r_query_dom_info(r_u, q_u->switch_value, ctr, NT_STATUS_OK);
DEBUG(5,("_samr_query_dom_info: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
_api_samr_create_user
Create an account, can be either a normal user or a machine.
This funcion will need to be updated for bdc/domain trusts.
********************************************************************/
NTSTATUS _api_samr_create_user(pipes_struct *p, SAMR_Q_CREATE_USER *q_u, SAMR_R_CREATE_USER *r_u)
{
SAM_ACCOUNT *sam_pass=NULL;
fstring account;
DOM_SID sid;
pstring add_script;
POLICY_HND dom_pol = q_u->domain_pol;
UNISTR2 user_account = q_u->uni_name;
uint16 acb_info = q_u->acb_info;
POLICY_HND *user_pol = &r_u->user_pol;
struct samr_info *info = NULL;
BOOL ret;
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 nt_status;
struct passwd *pw;
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &dom_pol, NULL))
return NT_STATUS_INVALID_HANDLE;
/* find the account: tell the caller if it exists.
lkclXXXX i have *no* idea if this is a problem or not
or even if you are supposed to construct a different
reply if the account already exists...
*/
rpcstr_pull(account, user_account.buffer, sizeof(account), user_account.uni_str_len*2, 0);
strlower(account);
pdb_init_sam(&sam_pass);
become_root();
ret = pdb_getsampwnam(sam_pass, account);
unbecome_root();
if (ret == True) {
/* this account exists: say so */
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sam_pass);
return NT_STATUS_USER_EXISTS;
}
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
pdb_free_sam(&sam_pass);
/*
* NB. VERY IMPORTANT ! This call must be done as the current pipe user,
* *NOT* surrounded by a become_root()/unbecome_root() call. This ensures
* that only people with write access to the smbpasswd file will be able
* to create a user. JRA.
*/
/*
* add the user in the /etc/passwd file or the unix authority system.
* We don't check if the smb_create_user() function succed or not for 2 reasons:
* a) local_password_change() checks for us if the /etc/passwd account really exists
* b) smb_create_user() would return an error if the account already exists
* and as it could return an error also if it can't create the account, it would be tricky.
*
* So we go the easy way, only check after if the account exists.
* JFM (2/3/2001), to clear any possible bad understanding (-:
*
* We now have seperate script paramaters for adding users/machines so we
* now have some sainity-checking to match.
*/
DEBUG(10,("checking account %s at pos %d for $ termination\n",account, strlen(account)-1));
#if 0
if ((acb_info & ACB_WSTRUST) && (account[strlen(account)-1] == '$')) {
pstrcpy(add_script, lp_addmachine_script());
} else if ((!(acb_info & ACB_WSTRUST)) && (account[strlen(account)-1] != '$')) {
pstrcpy(add_script, lp_adduser_script());
} else {
DEBUG(0, ("_api_samr_create_user: mismatch between trust flags and $ termination\n"));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sam_pass);
return NT_STATUS_UNSUCCESSFUL;
}
#endif
/*
* we can't check both the ending $ and the acb_info.
*
* UserManager creates trust accounts (ending in $,
* normal that hidden accounts) with the acb_info equals to ACB_NORMAL.
* JFM, 11/29/2001
*/
if (account[strlen(account)-1] == '$')
pstrcpy(add_script, lp_addmachine_script());
else
pstrcpy(add_script, lp_adduser_script());
if(*add_script) {
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
int add_ret;
all_string_sub(add_script, "%u", account, sizeof(account));
add_ret = smbrun(add_script,NULL);
DEBUG(3,("_api_samr_create_user: Running the command `%s' gave %d\n", add_script, add_ret));
}
pw = getpwnam_alloc(account);
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 (pw) {
if (!NT_STATUS_IS_OK(nt_status = pdb_init_sam_pw(&sam_pass, pw))) {
passwd_free(&pw);
return nt_status;
}
passwd_free(&pw); /* done with this now */
} else {
DEBUG(3,("attempting to create non-unix account %s\n", account));
if (!NT_STATUS_IS_OK(nt_status = pdb_init_sam(&sam_pass))) {
return nt_status;
}
if (!pdb_set_username(sam_pass, account)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_MEMORY;
}
}
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
pdb_set_acct_ctrl(sam_pass, acb_info);
if (!pdb_add_sam_account(sam_pass)) {
pdb_free_sam(&sam_pass);
DEBUG(0, ("could not add user/computer %s to passdb. Check permissions?\n",
account));
return NT_STATUS_ACCESS_DENIED;
}
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
pdb_reset_sam(sam_pass);
if (!pdb_getsampwnam(sam_pass, account)) {
pdb_free_sam(&sam_pass);
DEBUG(0, ("could not find user/computer %s just added to passdb?!?\n",
account));
return NT_STATUS_ACCESS_DENIED;
}
/* Get the domain SID stored in the domain policy */
if(!get_lsa_policy_samr_sid(p, &dom_pol, &sid)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_INVALID_HANDLE;
}
/* append the user's RID to it */
if(!sid_append_rid(&sid, pdb_get_user_rid(sam_pass) )) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_SUCH_USER;
}
/* associate the user's SID with the new handle. */
if ((info = get_samr_info_by_sid(&sid)) == NULL) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_MEMORY;
}
ZERO_STRUCTP(info);
info->sid = sid;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, user_pol, free_samr_info, (void *)info)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sam_pass);
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
}
r_u->user_rid=pdb_get_user_rid(sam_pass);
r_u->unknown_0 = 0x000703ff;
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&sam_pass);
return NT_STATUS_OK;
}
/*******************************************************************
samr_reply_connect_anon
********************************************************************/
NTSTATUS _samr_connect_anon(pipes_struct *p, SAMR_Q_CONNECT_ANON *q_u, SAMR_R_CONNECT_ANON *r_u)
{
struct samr_info *info = NULL;
/* set up the SAMR connect_anon response */
r_u->status = NT_STATUS_OK;
/* associate the user's SID with the new handle. */
if ((info = get_samr_info_by_sid(NULL)) == NULL)
return NT_STATUS_NO_MEMORY;
info->status = q_u->unknown_0;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, &r_u->connect_pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
return r_u->status;
}
/*******************************************************************
samr_reply_connect
********************************************************************/
NTSTATUS _samr_connect(pipes_struct *p, SAMR_Q_CONNECT *q_u, SAMR_R_CONNECT *r_u)
{
struct samr_info *info = NULL;
DEBUG(5,("_samr_connect: %d\n", __LINE__));
r_u->status = NT_STATUS_OK;
/* associate the user's SID with the new handle. */
if ((info = get_samr_info_by_sid(NULL)) == NULL)
return NT_STATUS_NO_MEMORY;
info->status = q_u->access_mask;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, &r_u->connect_pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
DEBUG(5,("_samr_connect: %d\n", __LINE__));
return r_u->status;
}
/**********************************************************************
api_samr_lookup_domain
**********************************************************************/
NTSTATUS _samr_lookup_domain(pipes_struct *p, SAMR_Q_LOOKUP_DOMAIN *q_u, SAMR_R_LOOKUP_DOMAIN *r_u)
{
fstring domain_name;
DOM_SID sid;
r_u->status = NT_STATUS_OK;
if (!find_policy_by_hnd(p, &q_u->connect_pol, NULL))
return NT_STATUS_INVALID_HANDLE;
rpcstr_pull(domain_name, q_u->uni_domain.buffer, sizeof(domain_name), q_u->uni_domain.uni_str_len*2, 0);
ZERO_STRUCT(sid);
if (!secrets_fetch_domain_sid(domain_name, &sid)) {
r_u->status = NT_STATUS_NO_SUCH_DOMAIN;
}
DEBUG(2,("Returning domain sid for domain %s -> %s\n", domain_name, sid_string_static(&sid)));
init_samr_r_lookup_domain(r_u, &sid, r_u->status);
return r_u->status;
}
/******************************************************************
makes a SAMR_R_ENUM_DOMAINS structure.
********************************************************************/
static BOOL make_enum_domains(TALLOC_CTX *ctx, SAM_ENTRY **pp_sam,
UNISTR2 **pp_uni_name, uint32 num_sam_entries, fstring doms[])
{
uint32 i;
SAM_ENTRY *sam;
UNISTR2 *uni_name;
DEBUG(5, ("make_enum_domains\n"));
*pp_sam = NULL;
*pp_uni_name = NULL;
if (num_sam_entries == 0)
return True;
sam = (SAM_ENTRY *)talloc_zero(ctx, sizeof(SAM_ENTRY)*num_sam_entries);
uni_name = (UNISTR2 *)talloc_zero(ctx, sizeof(UNISTR2)*num_sam_entries);
if (sam == NULL || uni_name == NULL)
return False;
for (i = 0; i < num_sam_entries; i++) {
int len = doms[i] != NULL ? strlen(doms[i]) : 0;
init_sam_entry(&sam[i], len, 0);
init_unistr2(&uni_name[i], doms[i], len);
}
*pp_sam = sam;
*pp_uni_name = uni_name;
return True;
}
/**********************************************************************
api_samr_enum_domains
**********************************************************************/
NTSTATUS _samr_enum_domains(pipes_struct *p, SAMR_Q_ENUM_DOMAINS *q_u, SAMR_R_ENUM_DOMAINS *r_u)
{
uint32 num_entries = 2;
fstring dom[2];
char *name;
r_u->status = NT_STATUS_OK;
switch (lp_server_role()) {
case ROLE_DOMAIN_PDC:
case ROLE_DOMAIN_BDC:
name = global_myworkgroup;
break;
default:
name = global_myname;
}
fstrcpy(dom[0],name);
strupper(dom[0]);
fstrcpy(dom[1],"Builtin");
if (!make_enum_domains(p->mem_ctx, &r_u->sam, &r_u->uni_dom_name, num_entries, dom))
return NT_STATUS_NO_MEMORY;
init_samr_r_enum_domains(r_u, q_u->start_idx + num_entries, num_entries);
return r_u->status;
}
/*******************************************************************
api_samr_open_alias
********************************************************************/
NTSTATUS _api_samr_open_alias(pipes_struct *p, SAMR_Q_OPEN_ALIAS *q_u, SAMR_R_OPEN_ALIAS *r_u)
{
DOM_SID sid;
POLICY_HND domain_pol = q_u->dom_pol;
uint32 alias_rid = q_u->rid_alias;
POLICY_HND *alias_pol = &r_u->pol;
struct samr_info *info = NULL;
r_u->status = NT_STATUS_OK;
/* get the domain policy. */
if (!find_policy_by_hnd(p, &domain_pol, NULL))
return NT_STATUS_INVALID_HANDLE;
/* Get the domain SID stored in the domain policy */
if(!get_lsa_policy_samr_sid(p, &domain_pol, &sid))
return NT_STATUS_INVALID_HANDLE;
/* append the alias' RID to it */
if(!sid_append_rid(&sid, alias_rid))
return NT_STATUS_NO_SUCH_USER;
/*
* we should check if the rid really exist !!!
* JFM.
*/
/* associate the user's SID with the new handle. */
if ((info = get_samr_info_by_sid(&sid)) == NULL)
return NT_STATUS_NO_MEMORY;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, alias_pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
return r_u->status;
}
/*******************************************************************
set_user_info_10
********************************************************************/
static BOOL set_user_info_10(const SAM_USER_INFO_10 *id10, uint32 rid)
{
SAM_ACCOUNT *pwd =NULL;
BOOL ret;
pdb_init_sam(&pwd);
ret = pdb_getsampwrid(pwd, rid);
if(ret==False) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
if (id10 == NULL) {
DEBUG(5, ("set_user_info_10: NULL id10\n"));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
if (!pdb_set_acct_ctrl(pwd, id10->acb_info)) {
pdb_free_sam(&pwd);
return False;
}
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_update_sam_account(pwd)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return True;
}
/*******************************************************************
set_user_info_12
********************************************************************/
static BOOL set_user_info_12(SAM_USER_INFO_12 *id12, uint32 rid)
{
SAM_ACCOUNT *pwd = NULL;
pdb_init_sam(&pwd);
if(!pdb_getsampwrid(pwd, rid)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
if (id12 == NULL) {
DEBUG(2, ("set_user_info_12: id12 is NULL\n"));
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
if (!pdb_set_lanman_passwd (pwd, id12->lm_pwd)) {
pdb_free_sam(&pwd);
return False;
}
if (!pdb_set_nt_passwd (pwd, id12->nt_pwd)) {
pdb_free_sam(&pwd);
return False;
}
if (!pdb_set_pass_changed_now (pwd)) {
pdb_free_sam(&pwd);
return False;
}
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_update_sam_account(pwd)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return True;
}
/*******************************************************************
set_user_info_21
********************************************************************/
static BOOL set_user_info_21(SAM_USER_INFO_21 *id21, uint32 rid)
{
SAM_ACCOUNT *pwd = NULL;
if (id21 == NULL) {
DEBUG(5, ("set_user_info_21: NULL id21\n"));
return False;
}
pdb_init_sam(&pwd);
if (!pdb_getsampwrid(pwd, rid)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
copy_id21_to_sam_passwd(pwd, id21);
/*
* The funny part about the previous two calls is
* that pwd still has the password hashes from the
* passdb entry. These have not been updated from
* id21. I don't know if they need to be set. --jerry
*/
/* write the change out */
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_update_sam_account(pwd)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return True;
}
/*******************************************************************
set_user_info_23
********************************************************************/
static BOOL set_user_info_23(SAM_USER_INFO_23 *id23, uint32 rid)
{
SAM_ACCOUNT *pwd = NULL;
pstring plaintext_buf;
uint32 len;
uint16 acct_ctrl;
if (id23 == NULL) {
DEBUG(5, ("set_user_info_23: NULL id23\n"));
return False;
}
pdb_init_sam(&pwd);
if (!pdb_getsampwrid(pwd, rid)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
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
DEBUG(5, ("Attempting administrator password change (level 23) for user %s\n",
pdb_get_username(pwd)));
acct_ctrl = pdb_get_acct_ctrl(pwd);
copy_id23_to_sam_passwd(pwd, id23);
if (!decode_pw_buffer((char*)id23->pass, plaintext_buf, 256, &len)) {
pdb_free_sam(&pwd);
return False;
}
if (!pdb_set_plaintext_passwd (pwd, plaintext_buf)) {
pdb_free_sam(&pwd);
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
return False;
}
/* if it's a trust account, don't update /etc/passwd */
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 ( (!IS_SAM_UNIX_USER(pwd)) ||
( (acct_ctrl & ACB_DOMTRUST) == ACB_DOMTRUST ) ||
( (acct_ctrl & ACB_WSTRUST) == ACB_WSTRUST) ||
( (acct_ctrl & ACB_SVRTRUST) == ACB_SVRTRUST) ) {
DEBUG(5, ("Changing trust account or non-unix-user password, not updating /etc/passwd\n"));
} else {
/* update the UNIX password */
if (lp_unix_password_sync() )
if(!chgpasswd(pdb_get_username(pwd), "", plaintext_buf, True)) {
pdb_free_sam(&pwd);
return False;
}
}
ZERO_STRUCT(plaintext_buf);
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_update_sam_account(pwd)) {
pdb_free_sam(&pwd);
return False;
}
pdb_free_sam(&pwd);
return True;
}
/*******************************************************************
set_user_info_pw
********************************************************************/
static BOOL set_user_info_pw(char *pass, uint32 rid)
{
SAM_ACCOUNT *pwd = NULL;
uint32 len;
pstring plaintext_buf;
uint16 acct_ctrl;
pdb_init_sam(&pwd);
if (!pdb_getsampwrid(pwd, rid)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
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
DEBUG(5, ("Attempting administrator password change for user %s\n",
pdb_get_username(pwd)));
acct_ctrl = pdb_get_acct_ctrl(pwd);
ZERO_STRUCT(plaintext_buf);
if (!decode_pw_buffer(pass, plaintext_buf, 256, &len)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
if (!pdb_set_plaintext_passwd (pwd, plaintext_buf)) {
pdb_free_sam(&pwd);
return False;
}
/* if it's a trust account, don't update /etc/passwd */
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 ( (!IS_SAM_UNIX_USER(pwd)) ||
( (acct_ctrl & ACB_DOMTRUST) == ACB_DOMTRUST ) ||
( (acct_ctrl & ACB_WSTRUST) == ACB_WSTRUST) ||
( (acct_ctrl & ACB_SVRTRUST) == ACB_SVRTRUST) ) {
DEBUG(5, ("Changing trust account or non-unix-user password, not updating /etc/passwd\n"));
} else {
/* update the UNIX password */
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 (lp_unix_password_sync()) {
if(!chgpasswd(pdb_get_username(pwd), "", plaintext_buf, True)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
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
}
}
ZERO_STRUCT(plaintext_buf);
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
DEBUG(5,("set_user_info_pw: pdb_update_pwd()\n"));
/* update the SAMBA password */
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_update_sam_account(pwd)) {
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return False;
}
Fix up a number of intertwined issues: The big one is a global change to allow us to NULLify the free'ed pointer to a former passdb object. This was done to allow idra's SAFE_FREE() macro to do its magic, and to satisfy the input test in pdb_init_sam() for a NULL pointer to start with. This NULL pointer test was what was breaking the adding of accounts up until now, and this code has been reworked to avoid duplicating work - I hope this will avoid a similar mess-up in future. Finally, I fixed a few nasty bugs where the pdb_ fuctions's return codes were being ignored. Some of these functions malloc() and are permitted to fail. Also, this caught a nasty bug where pdb_set_lanman_password(sam, NULL) acheived precisely didilly-squat, just returning False. Now that we check the returns this bug was spotted. This could allow different LM and NT passwords. - the pdbedit code needs to start checking these too, but I havn't had a chance to fix it. I have also fixed up where some of the password changing code was using the pdb_set functions to store *internal* data. I assume this is from a previous lot of mass conversion work... Most likally (and going on past experience) I have missed somthing, probably in the LanMan password change code which I havn't yet been able to test, but this lot is in much better shape than it was before. If all this is too much to swallow (particularly for 2.2.2) then just adding a sam_pass = NULL to the particular line of passdb.c should do the trick for the ovbious bug. Andrew Bartlett (This used to be commit 762c8758a7869809d89b4da9c2a5249678942930)
2001-09-29 17:08:26 +04:00
pdb_free_sam(&pwd);
return True;
}
/*******************************************************************
samr_reply_set_userinfo
********************************************************************/
NTSTATUS _samr_set_userinfo(pipes_struct *p, SAMR_Q_SET_USERINFO *q_u, SAMR_R_SET_USERINFO *r_u)
{
uint32 rid = 0x0;
DOM_SID sid;
POLICY_HND *pol = &q_u->pol;
uint16 switch_value = q_u->switch_value;
SAM_USERINFO_CTR *ctr = q_u->ctr;
DEBUG(5, ("_samr_set_userinfo: %d\n", __LINE__));
r_u->status = NT_STATUS_OK;
/* find the policy handle. open a policy on it. */
if (!get_lsa_policy_samr_sid(p, pol, &sid))
return NT_STATUS_INVALID_HANDLE;
sid_split_rid(&sid, &rid);
DEBUG(5, ("_samr_set_userinfo: rid:0x%x, level:%d\n", rid, switch_value));
if (ctr == NULL) {
DEBUG(5, ("_samr_set_userinfo: NULL info level\n"));
return NT_STATUS_INVALID_INFO_CLASS;
}
/* ok! user info levels (lots: see MSDEV help), off we go... */
switch (switch_value) {
case 0x12:
if (!set_user_info_12(ctr->info.id12, rid))
return NT_STATUS_ACCESS_DENIED;
break;
case 24:
SamOEMhash(ctr->info.id24->pass, p->session_key, 516);
dump_data(100, (char *)ctr->info.id24->pass, 516);
if (!set_user_info_pw((char *)ctr->info.id24->pass, rid))
return NT_STATUS_ACCESS_DENIED;
break;
case 25:
#if 0
/*
* Currently we don't really know how to unmarshall
* the level 25 struct, and the password encryption
* is different. This is a placeholder for when we
* do understand it. In the meantime just return INVALID
* info level and W2K SP2 drops down to level 23... JRA.
*/
SamOEMhash(ctr->info.id25->pass, p->session_key, 532);
dump_data(100, (char *)ctr->info.id25->pass, 532);
if (!set_user_info_pw(ctr->info.id25->pass, rid))
return NT_STATUS_ACCESS_DENIED;
break;
#endif
return NT_STATUS_INVALID_INFO_CLASS;
case 23:
SamOEMhash(ctr->info.id23->pass, p->session_key, 516);
dump_data(100, (char *)ctr->info.id23->pass, 516);
if (!set_user_info_23(ctr->info.id23, rid))
return NT_STATUS_ACCESS_DENIED;
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
return r_u->status;
}
/*******************************************************************
samr_reply_set_userinfo2
********************************************************************/
NTSTATUS _samr_set_userinfo2(pipes_struct *p, SAMR_Q_SET_USERINFO2 *q_u, SAMR_R_SET_USERINFO2 *r_u)
{
DOM_SID sid;
uint32 rid = 0x0;
SAM_USERINFO_CTR *ctr = q_u->ctr;
POLICY_HND *pol = &q_u->pol;
uint16 switch_value = q_u->switch_value;
DEBUG(5, ("samr_reply_set_userinfo2: %d\n", __LINE__));
r_u->status = NT_STATUS_OK;
/* find the policy handle. open a policy on it. */
if (!get_lsa_policy_samr_sid(p, pol, &sid))
return NT_STATUS_INVALID_HANDLE;
sid_split_rid(&sid, &rid);
DEBUG(5, ("samr_reply_set_userinfo2: rid:0x%x\n", rid));
if (ctr == NULL) {
DEBUG(5, ("samr_reply_set_userinfo2: NULL info level\n"));
return NT_STATUS_INVALID_INFO_CLASS;
}
switch_value=ctr->switch_value;
/* ok! user info levels (lots: see MSDEV help), off we go... */
switch (switch_value) {
case 21:
if (!set_user_info_21(ctr->info.id21, rid))
return NT_STATUS_ACCESS_DENIED;
break;
case 16:
if (!set_user_info_10(ctr->info.id10, rid))
return NT_STATUS_ACCESS_DENIED;
break;
case 18:
/* Used by AS/U JRA. */
if (!set_user_info_12(ctr->info.id12, rid))
return NT_STATUS_ACCESS_DENIED;
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
return r_u->status;
}
/*********************************************************************
_samr_query_aliasmem
*********************************************************************/
NTSTATUS _samr_query_useraliases(pipes_struct *p, SAMR_Q_QUERY_USERALIASES *q_u, SAMR_R_QUERY_USERALIASES *r_u)
{
int num_groups = 0, tmp_num_groups=0;
uint32 *rids=NULL, *new_rids=NULL, *tmp_rids=NULL;
struct samr_info *info = NULL;
int i,j;
/* until i see a real useraliases query, we fack one up */
/* I have seen one, JFM 2/12/2001 */
/*
* Explanation of what this call does:
* for all the SID given in the request:
* return a list of alias (local groups)
* that have those SID as members.
*
* and that's the alias in the domain specified
* in the policy_handle
*
* if the policy handle is on an incorrect sid
* for example a user's sid
* we should reply NT_STATUS_OBJECT_TYPE_MISMATCH
*/
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_query_useraliases: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
if (!sid_check_is_domain(&info->sid) &&
!sid_check_is_builtin(&info->sid))
return NT_STATUS_OBJECT_TYPE_MISMATCH;
for (i=0; i<q_u->num_sids1; i++) {
r_u->status=get_alias_user_groups(p->mem_ctx, &info->sid, &tmp_num_groups, &tmp_rids, &(q_u->sid[i].sid));
/*
* if there is an error, we just continue as
* it can be an unfound user or group
*/
if (NT_STATUS_IS_ERR(r_u->status)) {
DEBUG(10,("_samr_query_useraliases: an error occured while getting groups\n"));
continue;
}
if (tmp_num_groups==0) {
DEBUG(10,("_samr_query_useraliases: no groups found\n"));
continue;
}
new_rids=(uint32 *)talloc_realloc(p->mem_ctx, rids, (num_groups+tmp_num_groups)*sizeof(uint32));
if (new_rids==NULL) {
DEBUG(0,("_samr_query_useraliases: could not realloc memory\n"));
return NT_STATUS_NO_MEMORY;
}
rids=new_rids;
for (j=0; j<tmp_num_groups; j++)
rids[j+num_groups]=tmp_rids[j];
safe_free(tmp_rids);
num_groups+=tmp_num_groups;
}
init_samr_r_query_useraliases(r_u, num_groups, rids, NT_STATUS_OK);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_query_aliasmem
*********************************************************************/
NTSTATUS _samr_query_aliasmem(pipes_struct *p, SAMR_Q_QUERY_ALIASMEM *q_u, SAMR_R_QUERY_ALIASMEM *r_u)
{
int i;
GROUP_MAP map;
int num_uids = 0;
DOM_SID2 *sid;
uid_t *uid=NULL;
DOM_SID alias_sid;
DOM_SID als_sid;
uint32 alias_rid;
fstring alias_sid_str;
DOM_SID temp_sid;
/* find the policy handle. open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->alias_pol, &alias_sid))
return NT_STATUS_INVALID_HANDLE;
sid_copy(&als_sid, &alias_sid);
sid_to_string(alias_sid_str, &alias_sid);
sid_split_rid(&alias_sid, &alias_rid);
DEBUG(10, ("sid is %s\n", alias_sid_str));
if (sid_equal(&alias_sid, &global_sid_Builtin)) {
DEBUG(10, ("lookup on Builtin SID (S-1-5-32)\n"));
if(!get_local_group_from_sid(als_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
} else {
if (sid_equal(&alias_sid, &global_sam_sid)) {
DEBUG(10, ("lookup on Server SID\n"));
if(!get_local_group_from_sid(als_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
}
}
if(!get_uid_list_of_group(map.gid, &uid, &num_uids))
return NT_STATUS_NO_SUCH_ALIAS;
DEBUG(10, ("sid is %s\n", alias_sid_str));
sid = (DOM_SID2 *)talloc_zero(p->mem_ctx, sizeof(DOM_SID2) * num_uids);
if (num_uids!=0 && sid == NULL)
return NT_STATUS_NO_MEMORY;
for (i = 0; i < num_uids; i++) {
sid_copy(&temp_sid, &global_sam_sid);
sid_append_rid(&temp_sid, pdb_uid_to_user_rid(uid[i]));
init_dom_sid2(&sid[i], &temp_sid);
}
DEBUG(10, ("sid is %s\n", alias_sid_str));
init_samr_r_query_aliasmem(r_u, num_uids, sid, NT_STATUS_OK);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_query_groupmem
*********************************************************************/
NTSTATUS _samr_query_groupmem(pipes_struct *p, SAMR_Q_QUERY_GROUPMEM *q_u, SAMR_R_QUERY_GROUPMEM *r_u)
{
int num_uids = 0;
int i;
DOM_SID group_sid;
uint32 group_rid;
fstring group_sid_str;
uid_t *uid=NULL;
GROUP_MAP map;
uint32 *rid=NULL;
uint32 *attr=NULL;
/* find the policy handle. open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->group_pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
/* todo: change to use sid_compare_front */
sid_split_rid(&group_sid, &group_rid);
sid_to_string(group_sid_str, &group_sid);
DEBUG(10, ("sid is %s\n", group_sid_str));
/* can we get a query for an SID outside our domain ? */
if (!sid_equal(&group_sid, &global_sam_sid))
return NT_STATUS_NO_SUCH_GROUP;
sid_append_rid(&group_sid, group_rid);
DEBUG(10, ("lookup on Domain SID\n"));
if(!get_domain_group_from_sid(group_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
if(!get_uid_list_of_group(map.gid, &uid, &num_uids))
return NT_STATUS_NO_SUCH_GROUP;
rid=talloc_zero(p->mem_ctx, sizeof(uint32)*num_uids);
attr=talloc_zero(p->mem_ctx, sizeof(uint32)*num_uids);
if (num_uids!=0 && (rid==NULL || attr==NULL))
return NT_STATUS_NO_MEMORY;
for (i=0; i<num_uids; i++) {
rid[i]=pdb_uid_to_user_rid(uid[i]);
attr[i] = SID_NAME_USER;
}
init_samr_r_query_groupmem(r_u, num_uids, rid, attr, NT_STATUS_OK);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_add_aliasmem
*********************************************************************/
NTSTATUS _samr_add_aliasmem(pipes_struct *p, SAMR_Q_ADD_ALIASMEM *q_u, SAMR_R_ADD_ALIASMEM *r_u)
{
DOM_SID alias_sid;
fstring alias_sid_str;
uid_t uid;
struct passwd *pwd;
struct group *grp;
fstring grp_name;
uint32 rid;
GROUP_MAP map;
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->alias_pol, &alias_sid))
return NT_STATUS_INVALID_HANDLE;
sid_to_string(alias_sid_str, &alias_sid);
DEBUG(10, ("sid is %s\n", alias_sid_str));
if (sid_compare(&alias_sid, &global_sam_sid)>0) {
DEBUG(10, ("adding member on Server SID\n"));
if(!get_local_group_from_sid(alias_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
} else {
if (sid_compare(&alias_sid, &global_sid_Builtin)>0) {
DEBUG(10, ("adding member on BUILTIN SID\n"));
if( !get_local_group_from_sid(alias_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
} else
return NT_STATUS_NO_SUCH_ALIAS;
}
sid_split_rid(&q_u->sid.sid, &rid);
uid=pdb_user_rid_to_uid(rid);
if ((pwd=getpwuid(uid)) == NULL)
return NT_STATUS_NO_SUCH_USER;
if ((grp=getgrgid(map.gid)) == NULL)
return NT_STATUS_NO_SUCH_ALIAS;
/* we need to copy the name otherwise it's overloaded in user_in_group_list */
fstrcpy(grp_name, grp->gr_name);
/* if the user is already in the group */
if(user_in_group_list(pwd->pw_name, grp_name))
return NT_STATUS_MEMBER_IN_ALIAS;
/*
* 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(grp_name, pwd->pw_name);
/* check if the user has been added then ... */
if(!user_in_group_list(pwd->pw_name, grp_name))
return NT_STATUS_MEMBER_NOT_IN_ALIAS; /* don't know what to reply else */
return NT_STATUS_OK;
}
/*********************************************************************
_samr_del_aliasmem
*********************************************************************/
NTSTATUS _samr_del_aliasmem(pipes_struct *p, SAMR_Q_DEL_ALIASMEM *q_u, SAMR_R_DEL_ALIASMEM *r_u)
{
DOM_SID alias_sid;
fstring alias_sid_str;
struct group *grp;
fstring grp_name;
uint32 rid;
GROUP_MAP map;
SAM_ACCOUNT *sam_pass=NULL;
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->alias_pol, &alias_sid))
return NT_STATUS_INVALID_HANDLE;
sid_to_string(alias_sid_str, &alias_sid);
DEBUG(10, ("_samr_del_aliasmem:sid is %s\n", alias_sid_str));
if (!sid_check_is_in_our_domain(&alias_sid) &&
!sid_check_is_in_builtin(&alias_sid)) {
DEBUG(10, ("_samr_del_aliasmem:invalid alias group\n"));
return NT_STATUS_NO_SUCH_ALIAS;
}
if( !get_local_group_from_sid(alias_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
if ((grp=getgrgid(map.gid)) == NULL)
return NT_STATUS_NO_SUCH_ALIAS;
/* we need to copy the name otherwise it's overloaded in user_in_group_list */
fstrcpy(grp_name, grp->gr_name);
sid_peek_rid(&q_u->sid.sid, &rid);
/* check if the user exists before trying to remove it from the group */
pdb_init_sam(&sam_pass);
if(!pdb_getsampwrid(sam_pass, rid)) {
DEBUG(5,("_samr_del_aliasmem:User %s doesn't exist.\n", pdb_get_username(sam_pass)));
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_SUCH_USER;
}
/* if the user is not in the group */
if(!user_in_group_list(pdb_get_username(sam_pass), grp_name)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_MEMBER_IN_ALIAS;
}
smb_delete_user_group(grp_name, pdb_get_username(sam_pass));
/* check if the user has been removed then ... */
if(user_in_group_list(pdb_get_username(sam_pass), grp_name)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_MEMBER_NOT_IN_ALIAS; /* don't know what to reply else */
}
pdb_free_sam(&sam_pass);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_add_groupmem
*********************************************************************/
NTSTATUS _samr_add_groupmem(pipes_struct *p, SAMR_Q_ADD_GROUPMEM *q_u, SAMR_R_ADD_GROUPMEM *r_u)
{
DOM_SID group_sid;
fstring group_sid_str;
struct passwd *pwd;
struct group *grp;
fstring grp_name;
GROUP_MAP map;
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
sid_to_string(group_sid_str, &group_sid);
DEBUG(10, ("sid is %s\n", group_sid_str));
if (sid_compare(&group_sid, &global_sam_sid)<=0)
return NT_STATUS_NO_SUCH_GROUP;
DEBUG(10, ("lookup on Domain SID\n"));
if(!get_domain_group_from_sid(group_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
if ((pwd=getpwuid(pdb_user_rid_to_uid(q_u->rid))) ==NULL)
return NT_STATUS_NO_SUCH_USER;
if ((grp=getgrgid(map.gid)) == NULL)
return NT_STATUS_NO_SUCH_GROUP;
/* we need to copy the name otherwise it's overloaded in user_in_group_list */
fstrcpy(grp_name, grp->gr_name);
/* if the user is already in the group */
if(user_in_group_list(pwd->pw_name, grp_name))
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(grp_name, pwd->pw_name);
/* check if the user has been added then ... */
if(!user_in_group_list(pwd->pw_name, grp_name))
return NT_STATUS_MEMBER_NOT_IN_GROUP; /* don't know what to reply else */
return NT_STATUS_OK;
}
/*********************************************************************
_samr_del_groupmem
*********************************************************************/
NTSTATUS _samr_del_groupmem(pipes_struct *p, SAMR_Q_DEL_GROUPMEM *q_u, SAMR_R_DEL_GROUPMEM *r_u)
{
DOM_SID group_sid;
SAM_ACCOUNT *sam_pass=NULL;
uint32 rid;
GROUP_MAP map;
fstring grp_name;
struct group *grp;
/*
* delete the group member named q_u->rid
* who is a member of the sid associated with the handle
* the rid is a user's rid as the group is a domain group.
*/
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
if(!sid_check_is_in_our_domain(&group_sid))
return NT_STATUS_NO_SUCH_GROUP;
rid=q_u->rid;
if(!get_domain_group_from_sid(group_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
if ((grp=getgrgid(map.gid)) == NULL)
return NT_STATUS_NO_SUCH_GROUP;
/* we need to copy the name otherwise it's overloaded in user_in_group_list */
fstrcpy(grp_name, grp->gr_name);
/* check if the user exists before trying to remove it from the group */
pdb_init_sam(&sam_pass);
if(!pdb_getsampwrid(sam_pass, rid)) {
DEBUG(5,("User %s doesn't exist.\n", pdb_get_username(sam_pass)));
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_SUCH_USER;
}
/* if the user is not in the group */
if(!user_in_group_list(pdb_get_username(sam_pass), grp_name)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_MEMBER_NOT_IN_GROUP;
}
smb_delete_user_group(grp_name, pdb_get_username(sam_pass));
/* check if the user has been removed then ... */
if(user_in_group_list(pdb_get_username(sam_pass), grp_name)) {
pdb_free_sam(&sam_pass);
return NT_STATUS_ACCESS_DENIED; /* don't know what to reply else */
}
pdb_free_sam(&sam_pass);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_delete_dom_user
*********************************************************************/
NTSTATUS _samr_delete_dom_user(pipes_struct *p, SAMR_Q_DELETE_DOM_USER *q_u, SAMR_R_DELETE_DOM_USER *r_u )
{
DOM_SID user_sid;
SAM_ACCOUNT *sam_pass=NULL;
uint32 rid;
DEBUG(5, ("_samr_delete_dom_user: %d\n", __LINE__));
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->user_pol, &user_sid))
return NT_STATUS_INVALID_HANDLE;
if (!sid_check_is_in_our_domain(&user_sid))
return NT_STATUS_CANNOT_DELETE;
sid_peek_rid(&user_sid, &rid);
/* check if the user exists before trying to delete */
pdb_init_sam(&sam_pass);
if(!pdb_getsampwrid(sam_pass, rid)) {
DEBUG(5,("_samr_delete_dom_user:User %s doesn't exist.\n", pdb_get_username(sam_pass)));
pdb_free_sam(&sam_pass);
return NT_STATUS_NO_SUCH_USER;
}
/* 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
*/
smb_delete_user(pdb_get_username(sam_pass));
/* and delete the samba side */
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_delete_sam_account(sam_pass)) {
DEBUG(5,("_samr_delete_dom_user:Failed to delete entry for user %s.\n", pdb_get_username(sam_pass)));
pdb_free_sam(&sam_pass);
return NT_STATUS_CANNOT_DELETE;
}
pdb_free_sam(&sam_pass);
if (!close_policy_hnd(p, &q_u->user_pol))
return NT_STATUS_OBJECT_NAME_INVALID;
return NT_STATUS_OK;
}
/*********************************************************************
_samr_delete_dom_group
*********************************************************************/
NTSTATUS _samr_delete_dom_group(pipes_struct *p, SAMR_Q_DELETE_DOM_GROUP *q_u, SAMR_R_DELETE_DOM_GROUP *r_u)
{
DOM_SID group_sid;
DOM_SID dom_sid;
uint32 group_rid;
fstring group_sid_str;
gid_t gid;
struct group *grp;
GROUP_MAP map;
DEBUG(5, ("samr_delete_dom_group: %d\n", __LINE__));
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->group_pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
sid_copy(&dom_sid, &group_sid);
sid_to_string(group_sid_str, &dom_sid);
sid_split_rid(&dom_sid, &group_rid);
DEBUG(10, ("sid is %s\n", group_sid_str));
/* we check if it's our SID before deleting */
if (!sid_equal(&dom_sid, &global_sam_sid))
return NT_STATUS_NO_SUCH_GROUP;
DEBUG(10, ("lookup on Domain SID\n"));
if(!get_domain_group_from_sid(group_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
gid=map.gid;
/* check if group really exists */
if ( (grp=getgrgid(gid)) == NULL)
return NT_STATUS_NO_SUCH_GROUP;
/* we can delete the UNIX group */
smb_delete_group(grp->gr_name);
/* check if the group has been successfully deleted */
if ( (grp=getgrgid(gid)) != NULL)
return NT_STATUS_ACCESS_DENIED;
if(!group_map_remove(group_sid))
return NT_STATUS_ACCESS_DENIED;
if (!close_policy_hnd(p, &q_u->group_pol))
return NT_STATUS_OBJECT_NAME_INVALID;
return NT_STATUS_OK;
}
/*********************************************************************
_samr_delete_dom_alias
*********************************************************************/
NTSTATUS _samr_delete_dom_alias(pipes_struct *p, SAMR_Q_DELETE_DOM_ALIAS *q_u, SAMR_R_DELETE_DOM_ALIAS *r_u)
{
DOM_SID alias_sid;
DOM_SID dom_sid;
uint32 alias_rid;
fstring alias_sid_str;
gid_t gid;
struct group *grp;
GROUP_MAP map;
DEBUG(5, ("_samr_delete_dom_alias: %d\n", __LINE__));
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->alias_pol, &alias_sid))
return NT_STATUS_INVALID_HANDLE;
sid_copy(&dom_sid, &alias_sid);
sid_to_string(alias_sid_str, &dom_sid);
sid_split_rid(&dom_sid, &alias_rid);
DEBUG(10, ("sid is %s\n", alias_sid_str));
/* we check if it's our SID before deleting */
if (!sid_equal(&dom_sid, &global_sam_sid))
return NT_STATUS_NO_SUCH_ALIAS;
DEBUG(10, ("lookup on Local SID\n"));
if(!get_local_group_from_sid(alias_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_ALIAS;
gid=map.gid;
/* check if group really exists */
if ( (grp=getgrgid(gid)) == NULL)
return NT_STATUS_NO_SUCH_ALIAS;
/* we can delete the UNIX group */
smb_delete_group(grp->gr_name);
/* check if the group has been successfully deleted */
if ( (grp=getgrgid(gid)) != NULL)
return NT_STATUS_ACCESS_DENIED;
/* don't check if we removed it as it could be an un-mapped group */
group_map_remove(alias_sid);
if (!close_policy_hnd(p, &q_u->alias_pol))
return NT_STATUS_OBJECT_NAME_INVALID;
return NT_STATUS_OK;
}
/*********************************************************************
_samr_create_dom_group
*********************************************************************/
NTSTATUS _samr_create_dom_group(pipes_struct *p, SAMR_Q_CREATE_DOM_GROUP *q_u, SAMR_R_CREATE_DOM_GROUP *r_u)
{
DOM_SID dom_sid;
DOM_SID info_sid;
fstring name;
fstring sid_string;
struct group *grp;
struct samr_info *info;
PRIVILEGE_SET priv_set;
init_privilege(&priv_set);
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &dom_sid))
return NT_STATUS_INVALID_HANDLE;
if (!sid_equal(&dom_sid, &global_sam_sid))
return NT_STATUS_ACCESS_DENIED;
/* TODO: check if allowed to create group and add a become_root/unbecome_root pair.*/
unistr2_to_ascii(name, &q_u->uni_acct_desc, sizeof(name)-1);
/* check if group already exist */
if ((grp=getgrnam(name)) != NULL)
return NT_STATUS_GROUP_EXISTS;
/* we can create the UNIX group */
smb_create_group(name);
/* check if the group has been successfully created */
if ((grp=getgrnam(name)) == NULL)
return NT_STATUS_ACCESS_DENIED;
r_u->rid=pdb_gid_to_group_rid(grp->gr_gid);
/* add the group to the mapping table */
sid_copy(&info_sid, &global_sam_sid);
sid_append_rid(&info_sid, r_u->rid);
sid_to_string(sid_string, &info_sid);
if(!add_initial_entry(grp->gr_gid, sid_string, SID_NAME_DOM_GRP, name, NULL, priv_set, PR_ACCESS_FROM_NETWORK))
return NT_STATUS_ACCESS_DENIED;
if ((info = get_samr_info_by_sid(&info_sid)) == NULL)
return NT_STATUS_NO_MEMORY;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, &r_u->pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
return NT_STATUS_OK;
}
/*********************************************************************
_samr_create_dom_alias
*********************************************************************/
NTSTATUS _samr_create_dom_alias(pipes_struct *p, SAMR_Q_CREATE_DOM_ALIAS *q_u, SAMR_R_CREATE_DOM_ALIAS *r_u)
{
DOM_SID dom_sid;
DOM_SID info_sid;
fstring name;
fstring sid_string;
struct group *grp;
struct samr_info *info;
PRIVILEGE_SET priv_set;
init_privilege(&priv_set);
/* Find the policy handle. Open a policy on it. */
if (!get_lsa_policy_samr_sid(p, &q_u->dom_pol, &dom_sid))
return NT_STATUS_INVALID_HANDLE;
if (!sid_equal(&dom_sid, &global_sam_sid))
return NT_STATUS_ACCESS_DENIED;
/* TODO: check if allowed to create group and add a become_root/unbecome_root pair.*/
unistr2_to_ascii(name, &q_u->uni_acct_desc, sizeof(name)-1);
/* check if group already exists */
if ( (grp=getgrnam(name)) != NULL)
return NT_STATUS_GROUP_EXISTS;
/* we can create the UNIX group */
smb_create_group(name);
/* check if the group has been successfully created */
if ((grp=getgrnam(name)) == NULL)
return NT_STATUS_ACCESS_DENIED;
r_u->rid=pdb_gid_to_group_rid(grp->gr_gid);
sid_copy(&info_sid, &global_sam_sid);
sid_append_rid(&info_sid, r_u->rid);
sid_to_string(sid_string, &info_sid);
/* add the group to the mapping table */
if(!add_initial_entry(grp->gr_gid, sid_string, SID_NAME_ALIAS, name, NULL, priv_set, PR_ACCESS_FROM_NETWORK))
return NT_STATUS_ACCESS_DENIED;
if ((info = get_samr_info_by_sid(&info_sid)) == NULL)
return NT_STATUS_NO_MEMORY;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, &r_u->alias_pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
return NT_STATUS_OK;
}
/*********************************************************************
_samr_query_groupinfo
sends the name/comment pair of a domain group
level 1 send also the number of users of that group
*********************************************************************/
NTSTATUS _samr_query_groupinfo(pipes_struct *p, SAMR_Q_QUERY_GROUPINFO *q_u, SAMR_R_QUERY_GROUPINFO *r_u)
{
DOM_SID group_sid;
GROUP_MAP map;
uid_t *uid=NULL;
int num_uids=0;
GROUP_INFO_CTR *ctr;
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
if (!get_domain_group_from_sid(group_sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_INVALID_HANDLE;
ctr=(GROUP_INFO_CTR *)talloc_zero(p->mem_ctx, sizeof(GROUP_INFO_CTR));
if (ctr==NULL)
return NT_STATUS_NO_MEMORY;
switch (q_u->switch_level) {
case 1:
ctr->switch_value1 = 1;
if(!get_uid_list_of_group(map.gid, &uid, &num_uids))
return NT_STATUS_NO_SUCH_GROUP;
init_samr_group_info1(&ctr->group.info1, map.nt_name, map.comment, num_uids);
SAFE_FREE(uid);
break;
case 3:
ctr->switch_value1 = 3;
init_samr_group_info3(&ctr->group.info3);
break;
case 4:
ctr->switch_value1 = 4;
init_samr_group_info4(&ctr->group.info4, map.comment);
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
init_samr_r_query_groupinfo(r_u, ctr, NT_STATUS_OK);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_set_groupinfo
update a domain group's comment.
*********************************************************************/
NTSTATUS _samr_set_groupinfo(pipes_struct *p, SAMR_Q_SET_GROUPINFO *q_u, SAMR_R_SET_GROUPINFO *r_u)
{
DOM_SID group_sid;
GROUP_MAP map;
GROUP_INFO_CTR *ctr;
if (!get_lsa_policy_samr_sid(p, &q_u->pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
if (!get_domain_group_from_sid(group_sid, &map, MAPPING_WITH_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
ctr=q_u->ctr;
switch (ctr->switch_value1) {
case 1:
unistr2_to_ascii(map.comment, &(ctr->group.info1.uni_acct_desc), sizeof(map.comment)-1);
break;
case 4:
unistr2_to_ascii(map.comment, &(ctr->group.info4.uni_acct_desc), sizeof(map.comment)-1);
break;
default:
free_privilege(&map.priv_set);
return NT_STATUS_INVALID_INFO_CLASS;
}
if(!add_mapping_entry(&map, TDB_REPLACE)) {
free_privilege(&map.priv_set);
return NT_STATUS_NO_SUCH_GROUP;
}
free_privilege(&map.priv_set);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_set_groupinfo
update a domain group's comment.
*********************************************************************/
NTSTATUS _samr_set_aliasinfo(pipes_struct *p, SAMR_Q_SET_ALIASINFO *q_u, SAMR_R_SET_ALIASINFO *r_u)
{
DOM_SID group_sid;
GROUP_MAP map;
ALIAS_INFO_CTR *ctr;
if (!get_lsa_policy_samr_sid(p, &q_u->alias_pol, &group_sid))
return NT_STATUS_INVALID_HANDLE;
if (!get_local_group_from_sid(group_sid, &map, MAPPING_WITH_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
ctr=&q_u->ctr;
switch (ctr->switch_value1) {
case 3:
unistr2_to_ascii(map.comment, &(ctr->alias.info3.uni_acct_desc), sizeof(map.comment)-1);
break;
default:
free_privilege(&map.priv_set);
return NT_STATUS_INVALID_INFO_CLASS;
}
if(!add_mapping_entry(&map, TDB_REPLACE)) {
free_privilege(&map.priv_set);
return NT_STATUS_NO_SUCH_GROUP;
}
free_privilege(&map.priv_set);
return NT_STATUS_OK;
}
/*********************************************************************
_samr_get_dom_pwinfo
*********************************************************************/
NTSTATUS _samr_get_dom_pwinfo(pipes_struct *p, SAMR_Q_GET_DOM_PWINFO *q_u, SAMR_R_GET_DOM_PWINFO *r_u)
{
/* Actually, returning zeros here works quite well :-). */
return NT_STATUS_OK;
}
/*********************************************************************
_samr_open_group
*********************************************************************/
NTSTATUS _samr_open_group(pipes_struct *p, SAMR_Q_OPEN_GROUP *q_u, SAMR_R_OPEN_GROUP *r_u)
{
DOM_SID sid;
DOM_SID info_sid;
GROUP_MAP map;
struct samr_info *info;
fstring sid_string;
if (!get_lsa_policy_samr_sid(p, &q_u->domain_pol, &sid))
return NT_STATUS_INVALID_HANDLE;
/* this should not be hard-coded like this */
if (!sid_equal(&sid, &global_sam_sid))
return NT_STATUS_ACCESS_DENIED;
sid_copy(&info_sid, &global_sam_sid);
sid_append_rid(&info_sid, q_u->rid_group);
sid_to_string(sid_string, &info_sid);
if ((info = get_samr_info_by_sid(&info_sid)) == NULL)
return NT_STATUS_NO_MEMORY;
DEBUG(10, ("_samr_open_group:Opening SID: %s\n", sid_string));
/* check if that group really exists */
if (!get_domain_group_from_sid(info->sid, &map, MAPPING_WITHOUT_PRIV))
return NT_STATUS_NO_SUCH_GROUP;
/* get a (unique) handle. open a policy on it. */
if (!create_policy_hnd(p, &r_u->pol, free_samr_info, (void *)info))
return NT_STATUS_OBJECT_NAME_NOT_FOUND;
return NT_STATUS_OK;
}
/*********************************************************************
_samr_unknown_2d
*********************************************************************/
NTSTATUS _samr_unknown_2d(pipes_struct *p, SAMR_Q_UNKNOWN_2D *q_u, SAMR_R_UNKNOWN_2D *r_u)
{
DEBUG(0,("_samr_unknown_2d: Not yet implemented.\n"));
return NT_STATUS_NOT_IMPLEMENTED;
}
/*******************************************************************
_samr_unknown_2e
********************************************************************/
NTSTATUS _samr_unknown_2e(pipes_struct *p, SAMR_Q_UNKNOWN_2E *q_u, SAMR_R_UNKNOWN_2E *r_u)
{
struct samr_info *info = NULL;
SAM_UNK_CTR *ctr;
uint32 min_pass_len,pass_hist,flag;
time_t u_expire, u_min_age;
NTTIME nt_expire, nt_min_age;
time_t u_lock_duration, u_reset_time;
NTTIME nt_lock_duration, nt_reset_time;
uint32 lockout;
time_t u_logout;
NTTIME nt_logout;
uint32 num_users=0, num_groups=0, num_aliases=0;
if ((ctr = (SAM_UNK_CTR *)talloc_zero(p->mem_ctx, sizeof(SAM_UNK_CTR))) == NULL)
return NT_STATUS_NO_MEMORY;
ZERO_STRUCTP(ctr);
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_unknown_2e: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->domain_pol, (void **)&info))
return NT_STATUS_INVALID_HANDLE;
switch (q_u->switch_value) {
case 0x01:
account_policy_get(AP_MIN_PASSWORD_LEN, &min_pass_len);
account_policy_get(AP_PASSWORD_HISTORY, &pass_hist);
account_policy_get(AP_USER_MUST_LOGON_TO_CHG_PASS, &flag);
account_policy_get(AP_MAX_PASSWORD_AGE, (int *)&u_expire);
account_policy_get(AP_MIN_PASSWORD_AGE, (int *)&u_min_age);
unix_to_nt_time_abs(&nt_expire, u_expire);
unix_to_nt_time_abs(&nt_min_age, u_min_age);
init_unk_info1(&ctr->info.inf1, (uint16)min_pass_len, (uint16)pass_hist,
flag, nt_expire, nt_min_age);
break;
case 0x02:
become_root();
r_u->status=load_sampwd_entries(info, ACB_NORMAL);
unbecome_root();
if (NT_STATUS_IS_ERR(r_u->status)) {
DEBUG(5, ("_samr_query_dispinfo: load_sampwd_entries failed\n"));
return r_u->status;
}
num_users=info->disp_info.num_user_account;
free_samr_db(info);
r_u->status=load_group_domain_entries(info, &global_sam_sid);
if (NT_STATUS_IS_ERR(r_u->status)) {
DEBUG(5, ("_samr_query_dispinfo: load_group_domain_entries failed\n"));
return r_u->status;
}
num_groups=info->disp_info.num_group_account;
free_samr_db(info);
/* The time call below is to get a sequence number for the sam. FIXME !!! JRA. */
init_unk_info2(&ctr->info.inf2, global_myworkgroup, global_myname, (uint32) time(NULL),
num_users, num_groups, num_aliases);
break;
case 0x03:
account_policy_get(AP_TIME_TO_LOGOUT, (int *)&u_logout);
unix_to_nt_time_abs(&nt_logout, u_logout);
init_unk_info3(&ctr->info.inf3, nt_logout);
break;
case 0x05:
init_unk_info5(&ctr->info.inf5, global_myname);
break;
case 0x06:
init_unk_info6(&ctr->info.inf6);
break;
case 0x07:
init_unk_info7(&ctr->info.inf7);
break;
case 0x0c:
account_policy_get(AP_LOCK_ACCOUNT_DURATION, (int *)&u_lock_duration);
account_policy_get(AP_RESET_COUNT_TIME, (int *)&u_reset_time);
account_policy_get(AP_BAD_ATTEMPT_LOCKOUT, &lockout);
unix_to_nt_time_abs(&nt_lock_duration, u_lock_duration);
unix_to_nt_time_abs(&nt_reset_time, u_reset_time);
init_unk_info12(&ctr->info.inf12, nt_lock_duration, nt_reset_time, (uint16)lockout);
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
init_samr_r_samr_unknown_2e(r_u, q_u->switch_value, ctr, NT_STATUS_OK);
DEBUG(5,("_samr_unknown_2e: %d\n", __LINE__));
return r_u->status;
}
/*******************************************************************
_samr_
********************************************************************/
NTSTATUS _samr_set_dom_info(pipes_struct *p, SAMR_Q_SET_DOMAIN_INFO *q_u, SAMR_R_SET_DOMAIN_INFO *r_u)
{
time_t u_expire, u_min_age;
time_t u_logout;
time_t u_lock_duration, u_reset_time;
r_u->status = NT_STATUS_OK;
DEBUG(5,("_samr_set_dom_info: %d\n", __LINE__));
/* find the policy handle. open a policy on it. */
if (!find_policy_by_hnd(p, &q_u->domain_pol, NULL))
return NT_STATUS_INVALID_HANDLE;
DEBUG(5,("_samr_set_dom_info: switch_value: %d\n", q_u->switch_value));
switch (q_u->switch_value) {
case 0x01:
u_expire=nt_time_to_unix_abs(&q_u->ctr->info.inf1.expire);
u_min_age=nt_time_to_unix_abs(&q_u->ctr->info.inf1.min_passwordage);
account_policy_set(AP_MIN_PASSWORD_LEN, (uint32)q_u->ctr->info.inf1.min_length_password);
account_policy_set(AP_PASSWORD_HISTORY, (uint32)q_u->ctr->info.inf1.password_history);
account_policy_set(AP_USER_MUST_LOGON_TO_CHG_PASS, (uint32)q_u->ctr->info.inf1.flag);
account_policy_set(AP_MAX_PASSWORD_AGE, (int)u_expire);
account_policy_set(AP_MIN_PASSWORD_AGE, (int)u_min_age);
break;
case 0x02:
break;
case 0x03:
u_logout=nt_time_to_unix_abs(&q_u->ctr->info.inf3.logout);
account_policy_set(AP_TIME_TO_LOGOUT, (int)u_logout);
break;
case 0x05:
break;
case 0x06:
break;
case 0x07:
break;
case 0x0c:
u_lock_duration=nt_time_to_unix_abs(&q_u->ctr->info.inf12.duration);
u_reset_time=nt_time_to_unix_abs(&q_u->ctr->info.inf12.reset_count);
account_policy_set(AP_LOCK_ACCOUNT_DURATION, (int)u_lock_duration);
account_policy_set(AP_RESET_COUNT_TIME, (int)u_reset_time);
account_policy_set(AP_BAD_ATTEMPT_LOCKOUT, (uint32)q_u->ctr->info.inf12.bad_attempt_lockout);
break;
default:
return NT_STATUS_INVALID_INFO_CLASS;
}
init_samr_r_set_domain_info(r_u, NT_STATUS_OK);
DEBUG(5,("_samr_set_dom_info: %d\n", __LINE__));
return r_u->status;
}