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

1267 lines
33 KiB
C
Raw Normal View History

/*
Unix SMB/CIFS implementation.
Samba utility functions
Copyright (C) Andrew Tridgell 1992-1998
Copyright (C) Andrew Bartlett 2001-2004
2010-04-16 16:16:40 +04:00
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3 of the License, or
(at your option) any later version.
2010-04-16 16:16:40 +04:00
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
2010-04-16 16:16:40 +04:00
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
/* These comments regard the code to change the user's unix password: */
/* fork a child process to exec passwd and write to its
* tty to change a users password. This is running as the
* user who is attempting to change the password.
*/
/*
* This code was copied/borrowed and stolen from various sources.
* The primary source was the poppasswd.c from the authors of POPMail. This software
* was included as a client to change passwords using the 'passwd' program
* on the remote machine.
*
* This code has been hacked by Bob Nance (nance@niehs.nih.gov) and Evan Patterson
* (patters2@niehs.nih.gov) at the National Institute of Environmental Health Sciences
* and rights to modify, distribute or incorporate this change to the CAP suite or
* using it for any other reason are granted, so long as this disclaimer is left intact.
*/
/*
This code was hacked considerably for inclusion in Samba, primarily
by Andrew.Tridgell@anu.edu.au. The biggest change was the addition
of the "password chat" option, which allows the easy runtime
specification of the expected sequence of events to change a
password.
*/
#include "includes.h"
#include "../libcli/auth/libcli_auth.h"
#include "../lib/crypto/arcfour.h"
static NTSTATUS check_oem_password(const char *user,
uchar password_encrypted_with_lm_hash[516],
const uchar old_lm_hash_encrypted[16],
uchar password_encrypted_with_nt_hash[516],
const uchar old_nt_hash_encrypted[16],
struct samu *sampass,
char **pp_new_passwd);
#if ALLOW_CHANGE_PASSWORD
static int findpty(char **slave)
{
int master = -1;
char *line = NULL;
SMB_STRUCT_DIR *dirp = NULL;
const char *dpname;
*slave = NULL;
#if defined(HAVE_GRANTPT)
/* Try to open /dev/ptmx. If that fails, fall through to old method. */
if ((master = sys_open("/dev/ptmx", O_RDWR, 0)) >= 0) {
grantpt(master);
unlockpt(master);
line = (char *)ptsname(master);
if (line) {
*slave = SMB_STRDUP(line);
}
if (*slave == NULL) {
DEBUG(0,
("findpty: Unable to create master/slave pty pair.\n"));
/* Stop fd leak on error. */
close(master);
return -1;
} else {
DEBUG(10,
("findpty: Allocated slave pty %s\n", *slave));
return (master);
}
}
#endif /* HAVE_GRANTPT */
line = SMB_STRDUP("/dev/ptyXX");
if (!line) {
return (-1);
}
dirp = sys_opendir("/dev");
if (!dirp) {
SAFE_FREE(line);
return (-1);
}
while ((dpname = readdirname(dirp)) != NULL) {
if (strncmp(dpname, "pty", 3) == 0 && strlen(dpname) == 5) {
DEBUG(3,
("pty: try to open %s, line was %s\n", dpname,
line));
line[8] = dpname[3];
line[9] = dpname[4];
if ((master = sys_open(line, O_RDWR, 0)) >= 0) {
DEBUG(3, ("pty: opened %s\n", line));
line[5] = 't';
*slave = line;
sys_closedir(dirp);
return (master);
}
}
}
sys_closedir(dirp);
SAFE_FREE(line);
return (-1);
}
static int dochild(int master, const char *slavedev, const struct passwd *pass,
const char *passwordprogram, bool as_root)
{
int slave;
struct termios stermios;
gid_t gid;
uid_t uid;
char * const eptrs[1] = { NULL };
if (pass == NULL)
{
DEBUG(0,
("dochild: user doesn't exist in the UNIX password database.\n"));
return False;
}
gid = pass->pw_gid;
uid = pass->pw_uid;
gain_root_privilege();
/* Start new session - gets rid of controlling terminal. */
if (setsid() < 0)
{
DEBUG(3,
("Weirdness, couldn't let go of controlling terminal\n"));
return (False);
}
/* Open slave pty and acquire as new controlling terminal. */
if ((slave = sys_open(slavedev, O_RDWR, 0)) < 0)
{
DEBUG(3, ("More weirdness, could not open %s\n", slavedev));
return (False);
}
#if defined(TIOCSCTTY) && !defined(SUNOS5)
/*
* On patched Solaris 10 TIOCSCTTY is defined but seems not to work,
* see the discussion under
* https://bugzilla.samba.org/show_bug.cgi?id=5366.
*/
if (ioctl(slave, TIOCSCTTY, 0) < 0)
{
DEBUG(3, ("Error in ioctl call for slave pty\n"));
/* return(False); */
}
#elif defined(I_PUSH) && defined(I_FIND)
if (ioctl(slave, I_FIND, "ptem") == 0) {
ioctl(slave, I_PUSH, "ptem");
}
if (ioctl(slave, I_FIND, "ldterm") == 0) {
ioctl(slave, I_PUSH, "ldterm");
}
#endif
/* Close master. */
close(master);
/* Make slave stdin/out/err of child. */
if (dup2(slave, STDIN_FILENO) != STDIN_FILENO)
{
DEBUG(3, ("Could not re-direct stdin\n"));
return (False);
}
if (dup2(slave, STDOUT_FILENO) != STDOUT_FILENO)
{
DEBUG(3, ("Could not re-direct stdout\n"));
return (False);
}
if (dup2(slave, STDERR_FILENO) != STDERR_FILENO)
{
DEBUG(3, ("Could not re-direct stderr\n"));
return (False);
}
if (slave > 2)
close(slave);
/* Set proper terminal attributes - no echo, canonical input processing,
no map NL to CR/NL on output. */
if (tcgetattr(0, &stermios) < 0)
{
DEBUG(3,
("could not read default terminal attributes on pty\n"));
return (False);
}
stermios.c_lflag &= ~(ECHO | ECHOE | ECHOK | ECHONL);
stermios.c_lflag |= ICANON;
#ifdef ONLCR
stermios.c_oflag &= ~(ONLCR);
#endif
if (tcsetattr(0, TCSANOW, &stermios) < 0)
{
DEBUG(3, ("could not set attributes of pty\n"));
return (False);
}
/* make us completely into the right uid */
if (!as_root)
{
become_user_permanently(uid, gid);
}
DEBUG(10,
("Invoking '%s' as password change program.\n",
passwordprogram));
/* execl() password-change application */
if (execle("/bin/sh", "sh", "-c", passwordprogram, NULL, eptrs) < 0)
{
DEBUG(3, ("Bad status returned from %s\n", passwordprogram));
return (False);
}
return (True);
}
static int expect(int master, char *issue, char *expected)
{
char buffer[1024];
int attempts, timeout, nread;
size_t len;
bool match = False;
for (attempts = 0; attempts < 2; attempts++) {
NTSTATUS status;
if (!strequal(issue, ".")) {
if (lp_passwd_chat_debug())
DEBUG(100, ("expect: sending [%s]\n", issue));
if ((len = sys_write(master, issue, strlen(issue))) != strlen(issue)) {
DEBUG(2,("expect: (short) write returned %d\n",
(int)len ));
return False;
}
}
if (strequal(expected, "."))
return True;
/* Initial timeout. */
timeout = lp_passwd_chat_timeout() * 1000;
nread = 0;
buffer[nread] = 0;
while (True) {
status = read_fd_with_timeout(
master, buffer + nread, 1,
sizeof(buffer) - nread - 1,
timeout, &len);
if (!NT_STATUS_IS_OK(status)) {
break;
}
nread += len;
buffer[nread] = 0;
{
/* Eat leading/trailing whitespace before match. */
char *str = SMB_STRDUP(buffer);
if (!str) {
DEBUG(2,("expect: ENOMEM\n"));
return False;
}
trim_char(str, ' ', ' ');
if ((match = unix_wild_match(expected, str)) == True) {
/* Now data has started to return, lower timeout. */
timeout = lp_passwd_chat_timeout() * 100;
}
SAFE_FREE(str);
}
}
if (lp_passwd_chat_debug())
DEBUG(100, ("expect: expected [%s] received [%s] match %s\n",
expected, buffer, match ? "yes" : "no" ));
if (match)
break;
if (!NT_STATUS_IS_OK(status)) {
DEBUG(2, ("expect: %s\n", nt_errstr(status)));
return False;
}
}
DEBUG(10,("expect: returning %s\n", match ? "True" : "False" ));
return match;
}
static void pwd_sub(char *buf)
{
all_string_sub(buf, "\\n", "\n", 0);
all_string_sub(buf, "\\r", "\r", 0);
all_string_sub(buf, "\\s", " ", 0);
all_string_sub(buf, "\\t", "\t", 0);
}
static int talktochild(int master, const char *seq)
{
TALLOC_CTX *frame = talloc_stackframe();
int count = 0;
char *issue;
char *expected;
issue = talloc_strdup(frame, ".");
if (!issue) {
TALLOC_FREE(frame);
return false;
}
while (next_token_talloc(frame, &seq, &expected, NULL)) {
pwd_sub(expected);
count++;
if (!expect(master, issue, expected)) {
DEBUG(3, ("Response %d incorrect\n", count));
TALLOC_FREE(frame);
return false;
}
if (!next_token_talloc(frame, &seq, &issue, NULL)) {
issue = talloc_strdup(frame, ".");
if (!issue) {
TALLOC_FREE(frame);
return false;
}
}
pwd_sub(issue);
}
if (!strequal(issue, ".")) {
/* we have one final issue to send */
expected = talloc_strdup(frame, ".");
if (!expected) {
TALLOC_FREE(frame);
return false;
}
if (!expect(master, issue, expected)) {
TALLOC_FREE(frame);
return False;
}
}
TALLOC_FREE(frame);
return (count > 0);
}
static bool chat_with_program(char *passwordprogram, const struct passwd *pass,
char *chatsequence, bool as_root)
{
char *slavedev = NULL;
int master;
pid_t pid, wpid;
int wstat;
bool chstat = False;
if (pass == NULL) {
DEBUG(0, ("chat_with_program: user doesn't exist in the UNIX password database.\n"));
return False;
}
/* allocate a pseudo-terminal device */
if ((master = findpty(&slavedev)) < 0) {
DEBUG(3, ("chat_with_program: Cannot Allocate pty for password change: %s\n", pass->pw_name));
return (False);
}
/*
* We need to temporarily stop CatchChild from eating
* SIGCLD signals as it also eats the exit status code. JRA.
*/
CatchChildLeaveStatus();
if ((pid = sys_fork()) < 0) {
DEBUG(3, ("chat_with_program: Cannot fork() child for password change: %s\n", pass->pw_name));
SAFE_FREE(slavedev);
close(master);
CatchChild();
return (False);
}
/* we now have a pty */
if (pid > 0) { /* This is the parent process */
/* Don't need this anymore in parent. */
SAFE_FREE(slavedev);
if ((chstat = talktochild(master, chatsequence)) == False) {
DEBUG(3, ("chat_with_program: Child failed to change password: %s\n", pass->pw_name));
kill(pid, SIGKILL); /* be sure to end this process */
}
while ((wpid = sys_waitpid(pid, &wstat, 0)) < 0) {
if (errno == EINTR) {
errno = 0;
continue;
}
break;
}
if (wpid < 0) {
DEBUG(3, ("chat_with_program: The process is no longer waiting!\n\n"));
close(master);
CatchChild();
return (False);
}
/*
* Go back to ignoring children.
*/
CatchChild();
close(master);
if (pid != wpid) {
DEBUG(3, ("chat_with_program: We were waiting for the wrong process ID\n"));
return (False);
}
if (WIFEXITED(wstat) && (WEXITSTATUS(wstat) != 0)) {
DEBUG(3, ("chat_with_program: The process exited with status %d \
while we were waiting\n", WEXITSTATUS(wstat)));
return (False);
}
#if defined(WIFSIGNALLED) && defined(WTERMSIG)
else if (WIFSIGNALLED(wstat)) {
DEBUG(3, ("chat_with_program: The process was killed by signal %d \
while we were waiting\n", WTERMSIG(wstat)));
return (False);
}
#endif
} else {
/* CHILD */
/*
* Lose any elevated privileges.
*/
drop_effective_capability(KERNEL_OPLOCK_CAPABILITY);
drop_effective_capability(DMAPI_ACCESS_CAPABILITY);
/* make sure it doesn't freeze */
alarm(20);
if (as_root)
become_root();
DEBUG(3, ("chat_with_program: Dochild for user %s (uid=%d,gid=%d) (as_root = %s)\n", pass->pw_name,
(int)getuid(), (int)getgid(), BOOLSTR(as_root) ));
chstat = dochild(master, slavedev, pass, passwordprogram, as_root);
if (as_root)
unbecome_root();
/*
* The child should never return from dochild() ....
*/
DEBUG(0, ("chat_with_program: Error: dochild() returned %d\n", chstat));
exit(1);
}
if (chstat)
DEBUG(3, ("chat_with_program: Password change %ssuccessful for user %s\n",
(chstat ? "" : "un"), pass->pw_name));
return (chstat);
}
bool chgpasswd(const char *name, const struct passwd *pass,
const char *oldpass, const char *newpass, bool as_root)
{
char *passwordprogram = NULL;
char *chatsequence = NULL;
size_t i;
size_t len;
TALLOC_CTX *ctx = talloc_tos();
if (!oldpass) {
oldpass = "";
}
DEBUG(3, ("chgpasswd: Password change (as_root=%s) for user: %s\n", BOOLSTR(as_root), name));
#ifdef DEBUG_PASSWORD
DEBUG(100, ("chgpasswd: Passwords: old=%s new=%s\n", oldpass, newpass));
#endif
/* Take the passed information and test it for minimum criteria */
/* Password is same as old password */
if (strcmp(oldpass, newpass) == 0) {
/* don't allow same password */
DEBUG(2, ("chgpasswd: Password Change: %s, New password is same as old\n", name)); /* log the attempt */
return (False); /* inform the user */
}
/*
* Check the old and new passwords don't contain any control
* characters.
*/
len = strlen(oldpass);
for (i = 0; i < len; i++) {
if (iscntrl((int)oldpass[i])) {
DEBUG(0, ("chgpasswd: oldpass contains control characters (disallowed).\n"));
return False;
}
}
len = strlen(newpass);
for (i = 0; i < len; i++) {
if (iscntrl((int)newpass[i])) {
DEBUG(0, ("chgpasswd: newpass contains control characters (disallowed).\n"));
return False;
}
}
#ifdef WITH_PAM
if (lp_pam_password_change()) {
bool ret;
#ifdef HAVE_SETLOCALE
const char *prevlocale = setlocale(LC_ALL, "C");
#endif
if (as_root)
become_root();
if (pass) {
ret = smb_pam_passchange(pass->pw_name, oldpass, newpass);
} else {
ret = smb_pam_passchange(name, oldpass, newpass);
}
if (as_root)
unbecome_root();
#ifdef HAVE_SETLOCALE
setlocale(LC_ALL, prevlocale);
#endif
return ret;
}
#endif
/* A non-PAM password change just doen't make sense without a valid local user */
if (pass == NULL) {
DEBUG(0, ("chgpasswd: user %s doesn't exist in the UNIX password database.\n", name));
return false;
}
passwordprogram = talloc_strdup(ctx, lp_passwd_program());
if (!passwordprogram || !*passwordprogram) {
DEBUG(2, ("chgpasswd: Null password program - no password changing\n"));
return false;
}
chatsequence = talloc_strdup(ctx, lp_passwd_chat());
if (!chatsequence || !*chatsequence) {
DEBUG(2, ("chgpasswd: Null chat sequence - no password changing\n"));
return false;
}
if (as_root) {
/* The password program *must* contain the user name to work. Fail if not. */
if (strstr_m(passwordprogram, "%u") == NULL) {
DEBUG(0,("chgpasswd: Running as root the 'passwd program' parameter *MUST* contain \
the string %%u, and the given string %s does not.\n", passwordprogram ));
return false;
}
}
passwordprogram = talloc_string_sub(ctx, passwordprogram, "%u", name);
if (!passwordprogram) {
return false;
}
/* note that we do NOT substitute the %o and %n in the password program
as this would open up a security hole where the user could use
a new password containing shell escape characters */
chatsequence = talloc_string_sub(ctx, chatsequence, "%u", name);
if (!chatsequence) {
return false;
}
chatsequence = talloc_all_string_sub(ctx,
chatsequence,
"%o",
oldpass);
if (!chatsequence) {
return false;
}
chatsequence = talloc_all_string_sub(ctx,
chatsequence,
"%n",
newpass);
return chat_with_program(passwordprogram,
pass,
chatsequence,
as_root);
}
#else /* ALLOW_CHANGE_PASSWORD */
bool chgpasswd(const char *name, const struct passwd *pass,
const char *oldpass, const char *newpass, bool as_root)
{
DEBUG(0, ("chgpasswd: Unix Password changing not compiled in (user=%s)\n", name));
return (False);
}
#endif /* ALLOW_CHANGE_PASSWORD */
/***********************************************************
Code to check the lanman hashed password.
************************************************************/
bool check_lanman_password(char *user, uchar * pass1,
uchar * pass2, struct samu **hnd)
{
uchar unenc_new_pw[16];
uchar unenc_old_pw[16];
struct samu *sampass = NULL;
uint32 acct_ctrl;
const uint8 *lanman_pw;
bool ret;
if ( !(sampass = samu_new(NULL)) ) {
DEBUG(0, ("samu_new() failed!\n"));
return False;
}
2010-04-16 16:16:40 +04:00
become_root();
ret = pdb_getsampwnam(sampass, user);
unbecome_root();
if (ret == False) {
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
DEBUG(0,("check_lanman_password: getsampwnam returned NULL\n"));
TALLOC_FREE(sampass);
return False;
}
2010-04-16 16:16:40 +04:00
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
acct_ctrl = pdb_get_acct_ctrl (sampass);
lanman_pw = pdb_get_lanman_passwd (sampass);
if (acct_ctrl & ACB_DISABLED) {
DEBUG(0,("check_lanman_password: account %s disabled.\n", user));
TALLOC_FREE(sampass);
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 (lanman_pw == NULL) {
if (acct_ctrl & ACB_PWNOTREQ) {
/* this saves the pointer for the caller */
*hnd = sampass;
return True;
} else {
DEBUG(0, ("check_lanman_password: no lanman password !\n"));
TALLOC_FREE(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
return False;
}
}
/* Get the new lanman hash. */
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
D_P16(lanman_pw, pass2, unenc_new_pw);
/* Use this to get the old lanman hash. */
D_P16(unenc_new_pw, pass1, unenc_old_pw);
/* Check that the two old passwords match. */
if (memcmp(lanman_pw, unenc_old_pw, 16)) {
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
DEBUG(0,("check_lanman_password: old password doesn't match.\n"));
TALLOC_FREE(sampass);
return False;
}
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
/* this saves the pointer for the caller */
*hnd = sampass;
return True;
}
/***********************************************************
Code to change the lanman hashed password.
It nulls out the NT hashed password as it will
no longer be valid.
NOTE this function is designed to be called as root. Check the old password
is correct before calling. JRA.
************************************************************/
bool change_lanman_password(struct samu *sampass, uchar *pass2)
{
uchar null_pw[16];
uchar unenc_new_pw[16];
bool ret;
uint32 acct_ctrl;
const uint8 *pwd;
if (sampass == NULL) {
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
DEBUG(0,("change_lanman_password: no smb password entry.\n"));
return False;
}
2010-04-16 16:16:40 +04:00
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
acct_ctrl = pdb_get_acct_ctrl(sampass);
pwd = pdb_get_lanman_passwd(sampass);
if (acct_ctrl & ACB_DISABLED) {
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
DEBUG(0,("change_lanman_password: account %s disabled.\n",
pdb_get_username(sampass)));
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 (pwd == NULL) {
if (acct_ctrl & ACB_PWNOTREQ) {
uchar no_pw[14];
ZERO_STRUCT(no_pw);
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
E_P16(no_pw, null_pw);
pwd = null_pw;
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
} else {
DEBUG(0,("change_lanman_password: no lanman password !\n"));
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
/* Get the new lanman hash. */
D_P16(pwd, pass2, unenc_new_pw);
if (!pdb_set_lanman_passwd(sampass, unenc_new_pw, PDB_CHANGED)) {
return False;
}
if (!pdb_set_nt_passwd (sampass, NULL, PDB_CHANGED)) {
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; /* We lose the NT hash. Sorry. */
}
if (!pdb_set_pass_last_set_time (sampass, time(NULL), PDB_CHANGED)) {
TALLOC_FREE(sampass);
/* Not quite sure what this one qualifies as, but this will do */
return False;
}
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
/* Now flush the sam_passwd struct to persistent storage */
ret = NT_STATUS_IS_OK(pdb_update_sam_account (sampass));
return ret;
}
/***********************************************************
Code to check and change the OEM hashed password.
************************************************************/
NTSTATUS pass_oem_change(char *user,
uchar password_encrypted_with_lm_hash[516],
const uchar old_lm_hash_encrypted[16],
uchar password_encrypted_with_nt_hash[516],
const uchar old_nt_hash_encrypted[16],
enum samPwdChangeReason *reject_reason)
{
char *new_passwd = NULL;
struct samu *sampass = NULL;
NTSTATUS nt_status;
bool ret = false;
if (!(sampass = samu_new(NULL))) {
return NT_STATUS_NO_MEMORY;
}
become_root();
ret = pdb_getsampwnam(sampass, user);
unbecome_root();
if (ret == false) {
DEBUG(0,("pass_oem_change: getsmbpwnam returned NULL\n"));
TALLOC_FREE(sampass);
return NT_STATUS_NO_SUCH_USER;
}
nt_status = check_oem_password(user,
password_encrypted_with_lm_hash,
old_lm_hash_encrypted,
password_encrypted_with_nt_hash,
old_nt_hash_encrypted,
sampass,
&new_passwd);
if (!NT_STATUS_IS_OK(nt_status)) {
TALLOC_FREE(sampass);
return nt_status;
}
/* We've already checked the old password here.... */
become_root();
nt_status = change_oem_password(sampass, NULL, new_passwd, True, reject_reason);
unbecome_root();
memset(new_passwd, 0, strlen(new_passwd));
TALLOC_FREE(sampass);
return nt_status;
}
/***********************************************************
Decrypt and verify a user password change.
The 516 byte long buffers are encrypted with the old NT and
old LM passwords, and if the NT passwords are present, both
buffers contain a unicode string.
After decrypting the buffers, check the password is correct by
matching the old hashed passwords with the passwords in the passdb.
************************************************************/
static NTSTATUS check_oem_password(const char *user,
uchar password_encrypted_with_lm_hash[516],
const uchar old_lm_hash_encrypted[16],
uchar password_encrypted_with_nt_hash[516],
const uchar old_nt_hash_encrypted[16],
struct samu *sampass,
char **pp_new_passwd)
{
uchar null_pw[16];
uchar null_ntpw[16];
uint8 *password_encrypted;
const uint8 *encryption_key;
const uint8 *lanman_pw, *nt_pw;
uint32 acct_ctrl;
size_t new_pw_len;
uchar new_nt_hash[16];
uchar new_lm_hash[16];
uchar verifier[16];
char no_pw[2];
bool nt_pass_set = (password_encrypted_with_nt_hash && old_nt_hash_encrypted);
bool lm_pass_set = (password_encrypted_with_lm_hash && old_lm_hash_encrypted);
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
acct_ctrl = pdb_get_acct_ctrl(sampass);
#if 0
/* I am convinced this check here is wrong, it is valid to
* change a password of a user that has a disabled account - gd */
if (acct_ctrl & ACB_DISABLED) {
DEBUG(2,("check_lanman_password: account %s disabled.\n", user));
return NT_STATUS_ACCOUNT_DISABLED;
}
#endif
if ((acct_ctrl & ACB_PWNOTREQ) && lp_null_passwords()) {
/* construct a null password (in case one is needed */
no_pw[0] = 0;
no_pw[1] = 0;
nt_lm_owf_gen(no_pw, null_ntpw, null_pw);
lanman_pw = null_pw;
nt_pw = null_pw;
} else {
/* save pointers to passwords so we don't have to keep looking them up */
if (lp_lanman_auth()) {
lanman_pw = pdb_get_lanman_passwd(sampass);
} else {
lanman_pw = NULL;
}
nt_pw = pdb_get_nt_passwd(sampass);
}
if (nt_pw && nt_pass_set) {
/* IDEAL Case: passwords are in unicode, and we can
* read use the password encrypted with the NT hash
*/
password_encrypted = password_encrypted_with_nt_hash;
encryption_key = nt_pw;
} else if (lanman_pw && lm_pass_set) {
/* password may still be in unicode, but use LM hash version */
password_encrypted = password_encrypted_with_lm_hash;
encryption_key = lanman_pw;
} else if (nt_pass_set) {
DEBUG(1, ("NT password change supplied for user %s, but we have no NT password to check it with\n",
user));
return NT_STATUS_WRONG_PASSWORD;
} else if (lm_pass_set) {
if (lp_lanman_auth()) {
DEBUG(1, ("LM password change supplied for user %s, but we have no LanMan password to check it with\n",
user));
} else {
DEBUG(1, ("LM password change supplied for user %s, but we have disabled LanMan authentication\n",
user));
}
return NT_STATUS_WRONG_PASSWORD;
} else {
DEBUG(1, ("password change requested for user %s, but no password supplied!\n",
user));
return NT_STATUS_WRONG_PASSWORD;
}
/*
* Decrypt the password with the key
*/
arcfour_crypt( password_encrypted, encryption_key, 516);
if (!decode_pw_buffer(talloc_tos(),
password_encrypted,
pp_new_passwd,
&new_pw_len,
nt_pass_set ? CH_UTF16 : CH_DOS)) {
return NT_STATUS_WRONG_PASSWORD;
}
/*
* To ensure we got the correct new password, hash it and
* use it as a key to test the passed old password.
*/
if (nt_pass_set) {
/* NT passwords, verify the NT hash. */
/* Calculate the MD4 hash (NT compatible) of the password */
memset(new_nt_hash, '\0', 16);
E_md4hash(*pp_new_passwd, new_nt_hash);
if (nt_pw) {
/*
* check the NT verifier
*/
E_old_pw_hash(new_nt_hash, nt_pw, verifier);
if (memcmp(verifier, old_nt_hash_encrypted, 16)) {
DEBUG(0, ("check_oem_password: old nt "
"password doesn't match.\n"));
return NT_STATUS_WRONG_PASSWORD;
}
/* We could check the LM password here, but there is
* little point, we already know the password is
* correct, and the LM password might not even be
* present. */
/* Further, LM hash generation algorithms
* differ with charset, so we could
* incorrectly fail a perfectly valid password
* change */
#ifdef DEBUG_PASSWORD
DEBUG(100,
("check_oem_password: password %s ok\n", *pp_new_passwd));
#endif
return NT_STATUS_OK;
}
if (lanman_pw) {
/*
* check the lm verifier
*/
E_old_pw_hash(new_nt_hash, lanman_pw, verifier);
if (memcmp(verifier, old_lm_hash_encrypted, 16)) {
DEBUG(0,("check_oem_password: old lm password doesn't match.\n"));
return NT_STATUS_WRONG_PASSWORD;
}
#ifdef DEBUG_PASSWORD
DEBUG(100,
("check_oem_password: password %s ok\n", *pp_new_passwd));
#endif
return NT_STATUS_OK;
}
}
if (lanman_pw && lm_pass_set) {
E_deshash(*pp_new_passwd, new_lm_hash);
/*
* check the lm verifier
*/
E_old_pw_hash(new_lm_hash, lanman_pw, verifier);
if (memcmp(verifier, old_lm_hash_encrypted, 16)) {
Large commit which restructures the local password storage API. Currently the only backend which works is smbpasswd (tdb, LDAP, and NIS+) are broken, but they were somewhat broken before. :) The following functions implement the storage manipulation interface /*The following definitions come from passdb/pdb_smbpasswd.c */ BOOL pdb_setsampwent (BOOL update); void pdb_endsampwent (void); SAM_ACCOUNT* pdb_getsampwent (void); SAM_ACCOUNT* pdb_getsampwnam (char *username); SAM_ACCOUNT* pdb_getsampwuid (uid_t uid); SAM_ACCOUNT* pdb_getsampwrid (uint32 rid); BOOL pdb_add_sam_account (SAM_ACCOUNT *sampass); BOOL pdb_update_sam_account (SAM_ACCOUNT *sampass, BOOL override); BOOL pdb_delete_sam_account (char* username); There is also a host of pdb_set..() and pdb_get..() functions for manipulating SAM_ACCOUNT struct members. Note that the struct passdb_ops {} has gone away. Also notice that struct smb_passwd (formally in smb.h) has been moved to passdb/pdb_smbpasswd.c and is not accessed outisde of static internal functions in this file. All local password searches should make use of the the SAM_ACCOUNT struct and the previously mentioned functions. I'll write some documentation for this later. The next step is to fix the TDB passdb backend, then work on spliting the backends out into share libraries, and finally get the LDAP backend going. What works and may not: o domain logons from Win9x works o domain logons from WinNT 4 works o user and group enumeration as implemented by Tim works o file and print access works o changing password from Win9x & NT ummm...i'll fix this tonight :) If I broke anything else, just yell and I'll fix it. I think it should be fairly quite. -- jerry (This used to be commit 0b92d0838ebdbe24f34f17e313ecbf61a0301389)
2000-11-14 02:03:34 +03:00
DEBUG(0,("check_oem_password: old lm password doesn't match.\n"));
return NT_STATUS_WRONG_PASSWORD;
}
#ifdef DEBUG_PASSWORD
DEBUG(100,
("check_oem_password: password %s ok\n", *pp_new_passwd));
#endif
return NT_STATUS_OK;
}
/* should not be reached */
return NT_STATUS_WRONG_PASSWORD;
}
bool password_in_history(uint8_t nt_pw[NT_HASH_LEN],
uint32_t pw_history_len,
const uint8_t *pw_history)
{
static const uint8_t zero_md5_nt_pw[SALTED_MD5_HASH_LEN] = { 0, };
int i;
dump_data(100, nt_pw, NT_HASH_LEN);
dump_data(100, pw_history, PW_HISTORY_ENTRY_LEN * pw_history_len);
for (i=0; i<pw_history_len; i++) {
uint8_t new_nt_pw_salted_md5_hash[SALTED_MD5_HASH_LEN];
const uint8_t *current_salt;
const uint8_t *old_nt_pw_salted_md5_hash;
current_salt = &pw_history[i*PW_HISTORY_ENTRY_LEN];
old_nt_pw_salted_md5_hash = current_salt + PW_HISTORY_SALT_LEN;
if (memcmp(zero_md5_nt_pw, old_nt_pw_salted_md5_hash,
SALTED_MD5_HASH_LEN) == 0) {
/* Ignore zero valued entries. */
continue;
}
if (memcmp(zero_md5_nt_pw, current_salt,
PW_HISTORY_SALT_LEN) == 0)
{
/*
* New format: zero salt and then plain nt hash.
* Directly compare the hashes.
*/
if (memcmp(nt_pw, old_nt_pw_salted_md5_hash,
SALTED_MD5_HASH_LEN) == 0)
{
return true;
}
} else {
/*
* Old format: md5sum of salted nt hash.
* Create salted version of new pw to compare.
*/
E_md5hash(current_salt, nt_pw, new_nt_pw_salted_md5_hash);
if (memcmp(new_nt_pw_salted_md5_hash,
old_nt_pw_salted_md5_hash,
SALTED_MD5_HASH_LEN) == 0) {
return true;
}
}
}
return false;
}
/***********************************************************
This routine takes the given password and checks it against
the password history. Returns True if this password has been
found in the history list.
************************************************************/
static bool check_passwd_history(struct samu *sampass, const char *plaintext)
{
uchar new_nt_p16[NT_HASH_LEN];
const uint8 *nt_pw;
const uint8 *pwhistory;
uint32 pwHisLen, curr_pwHisLen;
pdb_get_account_policy(PDB_POLICY_PASSWORD_HISTORY, &pwHisLen);
if (pwHisLen == 0) {
return False;
}
pwhistory = pdb_get_pw_history(sampass, &curr_pwHisLen);
if (!pwhistory || curr_pwHisLen == 0) {
return False;
}
/* Only examine the minimum of the current history len and
the stored history len. Avoids race conditions. */
pwHisLen = MIN(pwHisLen,curr_pwHisLen);
nt_pw = pdb_get_nt_passwd(sampass);
E_md4hash(plaintext, new_nt_p16);
if (!memcmp(nt_pw, new_nt_p16, NT_HASH_LEN)) {
DEBUG(10,("check_passwd_history: proposed new password for user %s is the same as the current password !\n",
pdb_get_username(sampass) ));
return True;
}
if (password_in_history(new_nt_p16, pwHisLen, pwhistory)) {
DEBUG(1,("check_passwd_history: proposed new password for "
"user %s found in history list !\n",
pdb_get_username(sampass) ));
return true;
}
return false;
}
/***********************************************************
************************************************************/
NTSTATUS check_password_complexity(const char *username,
const char *password,
enum samPwdChangeReason *samr_reject_reason)
{
TALLOC_CTX *tosctx = talloc_tos();
int check_ret;
char *cmd;
/* Use external script to check password complexity */
if ((lp_check_password_script() == NULL)
|| (*(lp_check_password_script()) == '\0')) {
return NT_STATUS_OK;
}
cmd = talloc_string_sub(tosctx, lp_check_password_script(), "%u",
username);
if (!cmd) {
return NT_STATUS_PASSWORD_RESTRICTION;
}
check_ret = smbrunsecret(cmd, password);
DEBUG(5,("check_password_complexity: check password script (%s) "
"returned [%d]\n", cmd, check_ret));
TALLOC_FREE(cmd);
if (check_ret != 0) {
DEBUG(1,("check_password_complexity: "
"check password script said new password is not good "
"enough!\n"));
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_NOT_COMPLEX;
}
return NT_STATUS_PASSWORD_RESTRICTION;
}
return NT_STATUS_OK;
}
/***********************************************************
Code to change the oem password. Changes both the lanman
and NT hashes. Old_passwd is almost always NULL.
NOTE this function is designed to be called as root. Check the old password
is correct before calling. JRA.
************************************************************/
NTSTATUS change_oem_password(struct samu *hnd, char *old_passwd, char *new_passwd, bool as_root, enum samPwdChangeReason *samr_reject_reason)
{
uint32 min_len;
uint32 refuse;
TALLOC_CTX *tosctx = talloc_tos();
struct passwd *pass = NULL;
const char *username = pdb_get_username(hnd);
time_t can_change_time = pdb_get_pass_can_change_time(hnd);
NTSTATUS status;
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_NO_ERROR;
}
/* check to see if the secdesc has previously been set to disallow */
if (!pdb_get_pass_can_change(hnd)) {
DEBUG(1, ("user %s does not have permissions to change password\n", username));
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_NO_ERROR;
}
return NT_STATUS_ACCOUNT_RESTRICTION;
}
/* check to see if it is a Machine account and if the policy
* denies machines to change the password. *
* Should we deny also SRVTRUST and/or DOMSTRUST ? .SSS. */
if (pdb_get_acct_ctrl(hnd) & ACB_WSTRUST) {
if (pdb_get_account_policy(PDB_POLICY_REFUSE_MACHINE_PW_CHANGE, &refuse) && refuse) {
DEBUG(1, ("Machine %s cannot change password now, "
"denied by Refuse Machine Password Change policy\n",
username));
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_NO_ERROR;
}
return NT_STATUS_ACCOUNT_RESTRICTION;
}
}
2009-12-14 19:51:39 +03:00
/* removed calculation here, because passdb now calculates
based on policy. jmcd */
if ((can_change_time != 0) && (time(NULL) < can_change_time)) {
DEBUG(1, ("user %s cannot change password now, must "
"wait until %s\n", username,
http_timestring(tosctx, can_change_time)));
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_NO_ERROR;
}
return NT_STATUS_ACCOUNT_RESTRICTION;
}
if (pdb_get_account_policy(PDB_POLICY_MIN_PASSWORD_LEN, &min_len) && (str_charnum(new_passwd) < min_len)) {
DEBUG(1, ("user %s cannot change password - password too short\n",
username));
DEBUGADD(1, (" account policy min password len = %d\n", min_len));
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_PASSWORD_TOO_SHORT;
}
return NT_STATUS_PASSWORD_RESTRICTION;
/* return NT_STATUS_PWD_TOO_SHORT; */
}
if (check_passwd_history(hnd,new_passwd)) {
if (samr_reject_reason) {
*samr_reject_reason = SAM_PWD_CHANGE_PWD_IN_HISTORY;
}
return NT_STATUS_PASSWORD_RESTRICTION;
}
pass = Get_Pwnam_alloc(tosctx, username);
if (!pass) {
DEBUG(1, ("change_oem_password: Username %s does not exist in system !?!\n", username));
return NT_STATUS_ACCESS_DENIED;
}
status = check_password_complexity(username, new_passwd, samr_reject_reason);
if (!NT_STATUS_IS_OK(status)) {
TALLOC_FREE(pass);
return status;
}
/*
* If unix password sync was requested, attempt to change
* the /etc/passwd database first. Return failure if this cannot
* be done.
*
* This occurs before the oem change, because we don't want to
* update it if chgpasswd failed.
*
* Conditional on lp_unix_password_sync() because we don't want
* to touch the unix db unless we have admin permission.
*/
2010-04-16 16:16:40 +04:00
if(lp_unix_password_sync() &&
!chgpasswd(username, pass, old_passwd, new_passwd, as_root)) {
TALLOC_FREE(pass);
return NT_STATUS_ACCESS_DENIED;
}
TALLOC_FREE(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
if (!pdb_set_plaintext_passwd (hnd, new_passwd)) {
return NT_STATUS_ACCESS_DENIED;
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
}
/* Now write it into the file. */
return pdb_update_sam_account (hnd);
}