1
0
mirror of https://github.com/samba-team/samba.git synced 2025-01-25 06:04:04 +03:00

613 lines
15 KiB
C
Raw Normal View History

/*
Unix SMB/CIFS implementation.
mapping routines for SID <-> unix uid/gid
Copyright (C) Andrew Tridgell 2004
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3 of the License, or
(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, see <http://www.gnu.org/licenses/>.
*/
#include "includes.h"
#include "system/passwd.h"
#include "../libds/common/flags.h"
#include "dsdb/samdb/samdb.h"
#include "auth/auth.h"
#include "libcli/ldap/ldap_ndr.h"
#include "lib/ldb/include/ldb.h"
#include "../lib/util/util_ldb.h"
#include "libcli/security/security.h"
#include "param/param.h"
/*
these are used for the fallback local uid/gid to sid mapping
code.
*/
#define SIDMAP_LOCAL_USER_BASE 0x80000000
#define SIDMAP_LOCAL_GROUP_BASE 0xC0000000
#define SIDMAP_MAX_LOCAL_UID 0x3fffffff
#define SIDMAP_MAX_LOCAL_GID 0x3fffffff
/*
private context for sid mapping routines
*/
struct sidmap_context {
struct ldb_context *samctx;
};
/*
open a sidmap context - use talloc_free to close
*/
struct sidmap_context *sidmap_open(TALLOC_CTX *mem_ctx, struct tevent_context *ev_ctx,
struct loadparm_context *lp_ctx)
{
struct sidmap_context *sidmap;
sidmap = talloc(mem_ctx, struct sidmap_context);
if (sidmap == NULL) {
return NULL;
}
sidmap->samctx = samdb_connect(sidmap, ev_ctx, lp_ctx, system_session(sidmap, lp_ctx));
if (sidmap->samctx == NULL) {
talloc_free(sidmap);
return NULL;
}
return sidmap;
}
/*
check the sAMAccountType field of a search result to see if
the account is a user account
*/
static bool is_user_account(struct ldb_message *res)
{
uint_t atype = samdb_result_uint(res, "sAMAccountType", 0);
if (atype && (!(atype & ATYPE_ACCOUNT))) {
return false;
}
return true;
}
/*
check the sAMAccountType field of a search result to see if
the account is a group account
*/
static bool is_group_account(struct ldb_message *res)
{
uint_t atype = samdb_result_uint(res, "sAMAccountType", 0);
if (atype && atype == ATYPE_NORMAL_ACCOUNT) {
return false;
}
return true;
}
/*
return the dom_sid of our primary domain
*/
static NTSTATUS sidmap_primary_domain_sid(struct sidmap_context *sidmap,
TALLOC_CTX *mem_ctx, struct dom_sid **sid)
{
const char *attrs[] = { "objectSid", NULL };
int ret;
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
struct ldb_message **res = NULL;
ret = gendb_search_dn(sidmap->samctx, mem_ctx, NULL, &res, attrs);
if (ret != 1) {
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
talloc_free(res);
return NT_STATUS_NO_SUCH_DOMAIN;
}
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
*sid = samdb_result_dom_sid(mem_ctx, res[0], "objectSid");
talloc_free(res);
if (*sid == NULL) {
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}
/*
map a sid to a unix uid
*/
NTSTATUS sidmap_sid_to_unixuid(struct sidmap_context *sidmap,
const struct dom_sid *sid, uid_t *uid)
{
const char *attrs[] = { "sAMAccountName", "uidNumber",
"sAMAccountType", "unixName", NULL };
int ret;
const char *s;
TALLOC_CTX *tmp_ctx;
struct ldb_message **res;
struct dom_sid *domain_sid;
NTSTATUS status;
tmp_ctx = talloc_new(sidmap);
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"objectSid=%s",
ldap_encode_ndr_dom_sid(tmp_ctx, sid));
if (ret != 1) {
goto allocated_sid;
}
/* make sure its a user, not a group */
if (!is_user_account(res[0])) {
DEBUG(0,("sid_to_unixuid: sid %s is not an account!\n",
dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_INVALID_SID;
}
/* first try to get the uid directly */
s = samdb_result_string(res[0], "uidNumber", NULL);
if (s != NULL) {
*uid = strtoul(s, NULL, 0);
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
/* next try via the UnixName attribute */
s = samdb_result_string(res[0], "unixName", NULL);
if (s != NULL) {
struct passwd *pwd = getpwnam(s);
if (!pwd) {
DEBUG(0,("unixName %s for sid %s does not exist as a local user\n", s,
dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_NO_SUCH_USER;
}
*uid = pwd->pw_uid;
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
/* finally try via the sAMAccountName attribute */
s = samdb_result_string(res[0], "sAMAccountName", NULL);
if (s != NULL) {
struct passwd *pwd = getpwnam(s);
if (!pwd) {
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
DEBUG(0,("sAMAccountName '%s' for sid %s does not exist as a local user\n",
s, dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_NO_SUCH_USER;
}
*uid = pwd->pw_uid;
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
allocated_sid:
status = sidmap_primary_domain_sid(sidmap, tmp_ctx, &domain_sid);
if (!NT_STATUS_IS_OK(status)) {
talloc_free(tmp_ctx);
return NT_STATUS_NO_SUCH_DOMAIN;
}
if (dom_sid_in_domain(domain_sid, sid)) {
uint32_t rid = sid->sub_auths[sid->num_auths-1];
if (rid >= SIDMAP_LOCAL_USER_BASE &&
rid < SIDMAP_LOCAL_GROUP_BASE) {
*uid = rid - SIDMAP_LOCAL_USER_BASE;
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
}
DEBUG(0,("sid_to_unixuid: no uidNumber, unixName or sAMAccountName for sid %s\n",
dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_NONE_MAPPED;
}
/*
see if a sid is a group - very inefficient!
*/
bool sidmap_sid_is_group(struct sidmap_context *sidmap, struct dom_sid *sid)
{
const char *attrs[] = { "sAMAccountType", NULL };
int ret;
TALLOC_CTX *tmp_ctx;
struct ldb_message **res;
NTSTATUS status;
struct dom_sid *domain_sid;
bool is_group;
tmp_ctx = talloc_new(sidmap);
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"objectSid=%s", ldap_encode_ndr_dom_sid(tmp_ctx, sid));
if (ret == 1) {
is_group = is_group_account(res[0]);
talloc_free(tmp_ctx);
return is_group;
}
status = sidmap_primary_domain_sid(sidmap, tmp_ctx, &domain_sid);
if (!NT_STATUS_IS_OK(status)) {
talloc_free(tmp_ctx);
return false;
}
if (dom_sid_in_domain(domain_sid, sid)) {
uint32_t rid = sid->sub_auths[sid->num_auths-1];
if (rid >= SIDMAP_LOCAL_GROUP_BASE) {
talloc_free(tmp_ctx);
return true;
}
}
talloc_free(tmp_ctx);
return false;
}
/*
map a sid to a unix gid
*/
NTSTATUS sidmap_sid_to_unixgid(struct sidmap_context *sidmap,
const struct dom_sid *sid, gid_t *gid)
{
const char *attrs[] = { "sAMAccountName", "gidNumber",
"unixName", "sAMAccountType", NULL };
int ret;
const char *s;
TALLOC_CTX *tmp_ctx;
struct ldb_message **res;
NTSTATUS status;
struct dom_sid *domain_sid;
tmp_ctx = talloc_new(sidmap);
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"objectSid=%s", ldap_encode_ndr_dom_sid(tmp_ctx, sid));
if (ret != 1) {
goto allocated_sid;
}
/* make sure its not a user */
if (!is_group_account(res[0])) {
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
DEBUG(0,("sid_to_unixgid: sid %s is a ATYPE_NORMAL_ACCOUNT\n",
dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_INVALID_SID;
}
/* first try to get the gid directly */
s = samdb_result_string(res[0], "gidNumber", NULL);
if (s != NULL) {
*gid = strtoul(s, NULL, 0);
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
/* next try via the UnixName attribute */
s = samdb_result_string(res[0], "unixName", NULL);
if (s != NULL) {
struct group *grp = getgrnam(s);
if (!grp) {
DEBUG(0,("unixName '%s' for sid %s does not exist as a local group\n",
s, dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_NO_SUCH_GROUP;
}
*gid = grp->gr_gid;
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
/* finally try via the sAMAccountName attribute */
s = samdb_result_string(res[0], "sAMAccountName", NULL);
if (s != NULL) {
struct group *grp = getgrnam(s);
if (!grp) {
DEBUG(0,("sAMAccountName '%s' for sid %s does not exist as a local group\n", s, dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_NO_SUCH_GROUP;
}
*gid = grp->gr_gid;
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
allocated_sid:
status = sidmap_primary_domain_sid(sidmap, tmp_ctx, &domain_sid);
if (!NT_STATUS_IS_OK(status)) {
talloc_free(tmp_ctx);
return NT_STATUS_NO_SUCH_DOMAIN;
}
if (dom_sid_in_domain(domain_sid, sid)) {
uint32_t rid = sid->sub_auths[sid->num_auths-1];
if (rid >= SIDMAP_LOCAL_GROUP_BASE) {
*gid = rid - SIDMAP_LOCAL_GROUP_BASE;
talloc_free(tmp_ctx);
return NT_STATUS_OK;
}
}
DEBUG(0,("sid_to_unixgid: no gidNumber, unixName or sAMAccountName for sid %s\n",
dom_sid_string(tmp_ctx, sid)));
talloc_free(tmp_ctx);
return NT_STATUS_NONE_MAPPED;
}
/*
map a unix uid to a dom_sid
the returned sid is allocated in the supplied mem_ctx
*/
NTSTATUS sidmap_uid_to_sid(struct sidmap_context *sidmap,
TALLOC_CTX *mem_ctx,
const uid_t uid, struct dom_sid **sid)
{
const char *attrs[] = { "sAMAccountName", "objectSid", "sAMAccountType", NULL };
int ret, i;
TALLOC_CTX *tmp_ctx;
struct ldb_message **res;
struct passwd *pwd;
struct dom_sid *domain_sid;
NTSTATUS status;
/*
we search for the mapping in the following order:
- check if the uid is in the dynamic uid range assigned for winbindd
use. If it is, then look in winbindd sid mapping
database (not implemented yet)
- look for a user account in samdb that has uidNumber set to the
given uid
- look for a user account in samdb that has unixName or
sAMAccountName set to the name given by getpwuid()
- assign a SID by adding the uid to SIDMAP_LOCAL_USER_BASE in the local
domain
*/
tmp_ctx = talloc_new(mem_ctx);
/*
step 2: look for a user account in samdb that has uidNumber set to the
given uid
*/
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"uidNumber=%u", (unsigned int)uid);
for (i=0;i<ret;i++) {
if (!is_user_account(res[i])) continue;
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
*sid = samdb_result_dom_sid(mem_ctx, res[i], "objectSid");
talloc_free(tmp_ctx);
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
NT_STATUS_HAVE_NO_MEMORY(*sid);
return NT_STATUS_OK;
}
/*
step 3: look for a user account in samdb that has unixName
or sAMAccountName set to the name given by getpwuid()
*/
pwd = getpwuid(uid);
if (pwd == NULL) {
goto allocate_sid;
}
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"(|(unixName=%s)(sAMAccountName=%s))",
pwd->pw_name, pwd->pw_name);
for (i=0;i<ret;i++) {
if (!is_user_account(res[i])) continue;
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
*sid = samdb_result_dom_sid(mem_ctx, res[i], "objectSid");
talloc_free(tmp_ctx);
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
NT_STATUS_HAVE_NO_MEMORY(*sid);
return NT_STATUS_OK;
}
/*
step 4: assign a SID by adding the uid to
SIDMAP_LOCAL_USER_BASE in the local domain
*/
allocate_sid:
if (uid > SIDMAP_MAX_LOCAL_UID) {
return NT_STATUS_NONE_MAPPED;
}
status = sidmap_primary_domain_sid(sidmap, tmp_ctx, &domain_sid);
if (!NT_STATUS_IS_OK(status)) {
talloc_free(tmp_ctx);
return status;
}
*sid = dom_sid_add_rid(mem_ctx, domain_sid, SIDMAP_LOCAL_USER_BASE + uid);
talloc_free(tmp_ctx);
if (*sid == NULL) {
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}
/*
map a unix gid to a dom_sid
the returned sid is allocated in the supplied mem_ctx
*/
NTSTATUS sidmap_gid_to_sid(struct sidmap_context *sidmap,
TALLOC_CTX *mem_ctx,
const gid_t gid, struct dom_sid **sid)
{
const char *attrs[] = { "sAMAccountName", "objectSid", "sAMAccountType", NULL };
int ret, i;
TALLOC_CTX *tmp_ctx;
struct ldb_message **res;
struct group *grp;
struct dom_sid *domain_sid;
NTSTATUS status;
/*
we search for the mapping in the following order:
- check if the gid is in the dynamic gid range assigned for winbindd
use. If it is, then look in winbindd sid mapping
database (not implemented yet)
- look for a group account in samdb that has gidNumber set to the
given gid
- look for a group account in samdb that has unixName or
sAMAccountName set to the name given by getgrgid()
- assign a SID by adding the gid to SIDMAP_LOCAL_GROUP_BASE in the local
domain
*/
tmp_ctx = talloc_new(sidmap);
/*
step 2: look for a group account in samdb that has gidNumber set to the
given gid
*/
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"gidNumber=%u", (unsigned int)gid);
for (i=0;i<ret;i++) {
if (!is_group_account(res[i])) continue;
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
*sid = samdb_result_dom_sid(mem_ctx, res[i], "objectSid");
talloc_free(tmp_ctx);
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
NT_STATUS_HAVE_NO_MEMORY(*sid);
return NT_STATUS_OK;
}
/*
step 3: look for a group account in samdb that has unixName
or sAMAccountName set to the name given by getgrgid()
*/
grp = getgrgid(gid);
if (grp == NULL) {
goto allocate_sid;
}
ret = gendb_search(sidmap->samctx, tmp_ctx, NULL, &res, attrs,
"(|(unixName=%s)(sAMAccountName=%s))",
grp->gr_name, grp->gr_name);
for (i=0;i<ret;i++) {
if (!is_group_account(res[i])) continue;
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
*sid = samdb_result_dom_sid(mem_ctx, res[i], "objectSid");
talloc_free(tmp_ctx);
r7860: switch our ldb storage format to use a NDR encoded objectSid. This is quite a large change as we had lots of code that assumed that objectSid was a string in S- format. metze and simo tried to convince me to use NDR format months ago, but I didn't listen, so its fair that I have the pain of fixing all the code now :-) This builds on the ldb_register_samba_handlers() and ldif handlers code I did earlier this week. There are still three parts of this conversion I have not finished: - the ltdb index records need to use the string form of the objectSid (to keep the DNs sane). Until that it done I have disabled indexing on objectSid, which is a big performance hit, but allows us to pass all our tests while I rejig the indexing system to use a externally supplied conversion function - I haven't yet put in place the code that allows client to use the "S-xxx-yyy" form for objectSid in ldap search expressions. w2k3 supports this, presumably by looking for the "S-" prefix to determine what type of objectSid form is being used by the client. I have been working on ways to handle this, but am not happy with them yet so they aren't part of this patch - I need to change pidl to generate push functions that take a "const void *" instead of a "void*" for the data pointer. That will fix the couple of new warnings this code generates. Luckily it many places the conversion to NDR formatted records actually simplified the code, as it means we no longer need as many calls to dom_sid_parse_talloc(). In some places it got more complex, but not many. (This used to be commit d40bc2fa8ddd43560315688eebdbe98bdd02756c)
2005-06-24 00:18:20 +00:00
NT_STATUS_HAVE_NO_MEMORY(*sid);
return NT_STATUS_OK;
}
/*
step 4: assign a SID by adding the gid to
SIDMAP_LOCAL_GROUP_BASE in the local domain
*/
allocate_sid:
if (gid > SIDMAP_MAX_LOCAL_GID) {
return NT_STATUS_NONE_MAPPED;
}
status = sidmap_primary_domain_sid(sidmap, tmp_ctx, &domain_sid);
if (!NT_STATUS_IS_OK(status)) {
talloc_free(tmp_ctx);
return status;
}
*sid = dom_sid_add_rid(mem_ctx, domain_sid, SIDMAP_LOCAL_GROUP_BASE + gid);
talloc_free(tmp_ctx);
if (*sid == NULL) {
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}
/*
check if a sid is in the range of auto-allocated SIDs from our primary domain,
and if it is, then return the name and atype
*/
NTSTATUS sidmap_allocated_sid_lookup(struct sidmap_context *sidmap,
TALLOC_CTX *mem_ctx,
const struct dom_sid *sid,
const char **name,
enum lsa_SidType *rtype)
{
NTSTATUS status;
struct dom_sid *domain_sid;
TALLOC_CTX *tmp_ctx = talloc_new(mem_ctx);
uint32_t rid, atype;
status = sidmap_primary_domain_sid(sidmap, tmp_ctx, &domain_sid);
if (!NT_STATUS_IS_OK(status)) {
return NT_STATUS_NO_SUCH_DOMAIN;
}
if (!dom_sid_in_domain(domain_sid, sid)) {
talloc_free(tmp_ctx);
return NT_STATUS_NONE_MAPPED;
}
talloc_free(tmp_ctx);
rid = sid->sub_auths[sid->num_auths-1];
if (rid < SIDMAP_LOCAL_USER_BASE) {
return NT_STATUS_NONE_MAPPED;
}
if (rid < SIDMAP_LOCAL_GROUP_BASE) {
struct passwd *pwd;
uid_t uid = rid - SIDMAP_LOCAL_USER_BASE;
atype = ATYPE_NORMAL_ACCOUNT;
*rtype = samdb_atype_map(atype);
pwd = getpwuid(uid);
if (pwd == NULL) {
*name = talloc_asprintf(mem_ctx, "uid%u", uid);
} else {
*name = talloc_strdup(mem_ctx, pwd->pw_name);
}
} else {
struct group *grp;
gid_t gid = rid - SIDMAP_LOCAL_GROUP_BASE;
atype = ATYPE_LOCAL_GROUP;
*rtype = samdb_atype_map(atype);
grp = getgrgid(gid);
if (grp == NULL) {
*name = talloc_asprintf(mem_ctx, "gid%u", gid);
} else {
*name = talloc_strdup(mem_ctx, grp->gr_name);
}
}
if (*name == NULL) {
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}