1
0
mirror of https://github.com/samba-team/samba.git synced 2025-01-08 21:18:16 +03:00
samba-mirror/source3/passdb/smbpass.c

1256 lines
36 KiB
C
Raw Normal View History

/*
* Unix SMB/Netbios implementation. Version 1.9. SMB parameters and setup
* Copyright (C) Andrew Tridgell 1992-1998 Modified by Jeremy Allison 1995.
*
* This program is free software; you can redistribute it and/or modify it under
* the terms of the GNU General Public License as published by the Free
* Software Foundation; either version 2 of the License, or (at your option)
* any later version.
*
* This program is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
* more details.
*
* You should have received a copy of the GNU General Public License along with
* this program; if not, write to the Free Software Foundation, Inc., 675
* Mass Ave, Cambridge, MA 02139, USA.
*/
#include "includes.h"
#ifdef USE_SMBPASS_DB
extern int DEBUGLEVEL;
extern pstring samlogon_user;
extern BOOL sam_logon_in_ssb;
static int pw_file_lock_depth;
enum pwf_access_type { PWF_READ, PWF_UPDATE, PWF_CREATE };
/***************************************************************
Internal fn to enumerate the smbpasswd list. Returns a void pointer
to ensure no modification outside this module. Checks for atomic
rename of smbpasswd file on update or create once the lock has
been granted to prevent race conditions. JRA.
****************************************************************/
static void *startsmbfilepwent_internal(const char *pfile, enum pwf_access_type type, int *lock_depth)
{
FILE *fp = NULL;
const char *open_mode = NULL;
int race_loop = 0;
int lock_type = F_RDLCK;
if (!*pfile) {
DEBUG(0, ("startsmbfilepwent: No SMB password file set\n"));
return (NULL);
}
switch(type) {
case PWF_READ:
open_mode = "rb";
lock_type = F_RDLCK;
break;
case PWF_UPDATE:
open_mode = "r+b";
lock_type = F_WRLCK;
break;
case PWF_CREATE:
/*
* Ensure atomic file creation.
*/
{
int i, fd = -1;
for(i = 0; i < 5; i++) {
if((fd = sys_open(pfile, O_CREAT|O_TRUNC|O_EXCL|O_RDWR, 0600))!=-1)
break;
sys_usleep(200); /* Spin, spin... */
}
if(fd == -1) {
DEBUG(0,("startsmbfilepwent_internal: too many race conditions creating file %s\n", pfile));
return NULL;
}
close(fd);
open_mode = "r+b";
lock_type = F_WRLCK;
break;
}
}
for(race_loop = 0; race_loop < 5; race_loop++) {
DEBUG(10, ("startsmbfilepwent_internal: opening file %s\n", pfile));
if((fp = sys_fopen(pfile, open_mode)) == NULL) {
DEBUG(0, ("startsmbfilepwent_internal: unable to open file %s. Error was %s\n", pfile, strerror(errno) ));
return NULL;
}
if (!pw_file_lock(fileno(fp), lock_type, 5, lock_depth)) {
DEBUG(0, ("startsmbfilepwent_internal: unable to lock file %s. Error was %s\n", pfile, strerror(errno) ));
fclose(fp);
return NULL;
}
/*
* Only check for replacement races on update or create.
* For read we don't mind if the data is one record out of date.
*/
if(type == PWF_READ) {
break;
} else {
SMB_STRUCT_STAT sbuf1, sbuf2;
/*
* Avoid the potential race condition between the open and the lock
* by doing a stat on the filename and an fstat on the fd. If the
* two inodes differ then someone did a rename between the open and
* the lock. Back off and try the open again. Only do this 5 times to
* prevent infinate loops. JRA.
*/
if (sys_stat(pfile,&sbuf1) != 0) {
DEBUG(0, ("startsmbfilepwent_internal: unable to stat file %s. Error was %s\n", pfile, strerror(errno)));
pw_file_unlock(fileno(fp), lock_depth);
fclose(fp);
return NULL;
}
if (sys_fstat(fileno(fp),&sbuf2) != 0) {
DEBUG(0, ("startsmbfilepwent_internal: unable to fstat file %s. Error was %s\n", pfile, strerror(errno)));
pw_file_unlock(fileno(fp), lock_depth);
fclose(fp);
return NULL;
}
if( sbuf1.st_ino == sbuf2.st_ino) {
/* No race. */
break;
}
/*
* Race occurred - back off and try again...
*/
pw_file_unlock(fileno(fp), lock_depth);
fclose(fp);
}
}
if(race_loop == 5) {
DEBUG(0, ("startsmbfilepwent_internal: too many race conditions opening file %s\n", pfile));
return NULL;
}
/* Set a buffer to do more efficient reads */
setvbuf(fp, (char *)NULL, _IOFBF, 1024);
/* Make sure it is only rw by the owner */
if(fchmod(fileno(fp), S_IRUSR|S_IWUSR) == -1) {
DEBUG(0, ("startsmbfilepwent_internal: failed to set 0600 permissions on password file %s. \
Error was %s\n.", pfile, strerror(errno) ));
pw_file_unlock(fileno(fp), lock_depth);
fclose(fp);
return NULL;
}
/* We have a lock on the file. */
return (void *)fp;
}
/***************************************************************
Start to enumerate the smbpasswd list. Returns a void pointer
to ensure no modification outside this module.
****************************************************************/
static void *startsmbfilepwent(BOOL update)
{
return startsmbfilepwent_internal(lp_smb_passwd_file(), update ? PWF_UPDATE : PWF_READ, &pw_file_lock_depth);
}
/***************************************************************
End enumeration of the smbpasswd list.
****************************************************************/
static void endsmbfilepwent_internal(void *vp, int *lock_depth)
{
FILE *fp = (FILE *)vp;
pw_file_unlock(fileno(fp), lock_depth);
fclose(fp);
DEBUG(7, ("endsmbfilepwent_internal: closed password file.\n"));
}
/***************************************************************
End enumeration of the smbpasswd list - operate on the default
lock_depth.
****************************************************************/
static void endsmbfilepwent(void *vp)
{
endsmbfilepwent_internal(vp, &pw_file_lock_depth);
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
}
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
/*************************************************************************
Routine to return the next entry in the smbpasswd list.
*************************************************************************/
static struct smb_passwd *getsmbfilepwent(void *vp)
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
{
/* Static buffers we will return. */
static struct smb_passwd pw_buf;
static pstring user_name;
static unsigned char smbpwd[16];
static unsigned char smbntpwd[16];
FILE *fp = (FILE *)vp;
char linebuf[256];
unsigned char c;
unsigned char *p;
long uidval;
size_t linebuf_len;
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
if(fp == NULL) {
DEBUG(0,("getsmbfilepwent: Bad password file pointer.\n"));
return NULL;
}
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
pdb_init_smb(&pw_buf);
pw_buf.acct_ctrl = ACB_NORMAL;
/*
* Scan the file, a line at a time and check if the name matches.
*/
while (!feof(fp)) {
linebuf[0] = '\0';
fgets(linebuf, 256, fp);
if (ferror(fp)) {
return NULL;
}
/*
* Check if the string is terminated with a newline - if not
* then we must keep reading and discard until we get one.
*/
linebuf_len = strlen(linebuf);
if (linebuf[linebuf_len - 1] != '\n') {
c = '\0';
while (!ferror(fp) && !feof(fp)) {
c = fgetc(fp);
if (c == '\n')
break;
}
} else
linebuf[linebuf_len - 1] = '\0';
#ifdef DEBUG_PASSWORD
DEBUG(100, ("getsmbfilepwent: got line |%s|\n", linebuf));
#endif
if ((linebuf[0] == 0) && feof(fp)) {
DEBUG(4, ("getsmbfilepwent: end of file reached\n"));
break;
}
/*
* The line we have should be of the form :-
*
* username:uid:32hex bytes:[Account type]:LCT-12345678....other flags presently
* ignored....
*
* or,
*
* username:uid:32hex bytes:32hex bytes:[Account type]:LCT-12345678....ignored....
*
* if Windows NT compatible passwords are also present.
* [Account type] is an ascii encoding of the type of account.
* LCT-(8 hex digits) is the time_t value of the last change time.
*/
if (linebuf[0] == '#' || linebuf[0] == '\0') {
DEBUG(6, ("getsmbfilepwent: skipping comment or blank line\n"));
continue;
}
p = (unsigned char *) strchr(linebuf, ':');
if (p == NULL) {
DEBUG(0, ("getsmbfilepwent: malformed password entry (no :)\n"));
continue;
}
/*
* As 256 is shorter than a pstring we don't need to check
* length here - if this ever changes....
*/
strncpy(user_name, linebuf, PTR_DIFF(p, linebuf));
user_name[PTR_DIFF(p, linebuf)] = '\0';
/* Get smb uid. */
p++; /* Go past ':' */
if(*p == '-') {
DEBUG(0, ("getsmbfilepwent: uids in the smbpasswd file must not be negative.\n"));
continue;
}
if (!isdigit(*p)) {
DEBUG(0, ("getsmbfilepwent: malformed password entry (uid not number)\n"));
continue;
}
uidval = atoi((char *) p);
while (*p && isdigit(*p))
p++;
if (*p != ':') {
DEBUG(0, ("getsmbfilepwent: malformed password entry (no : after uid)\n"));
continue;
}
pw_buf.smb_name = user_name;
pw_buf.smb_userid = uidval;
/*
* Now get the password value - this should be 32 hex digits
* which are the ascii representations of a 16 byte string.
* Get two at a time and put them into the password.
*/
/* Skip the ':' */
p++;
if (*p == '*' || *p == 'X') {
/* Password deliberately invalid - end here. */
DEBUG(10, ("getsmbfilepwent: entry invalidated for user %s\n", user_name));
pw_buf.smb_nt_passwd = NULL;
pw_buf.smb_passwd = NULL;
pw_buf.acct_ctrl |= ACB_DISABLED;
return &pw_buf;
}
if (linebuf_len < (PTR_DIFF(p, linebuf) + 33)) {
DEBUG(0, ("getsmbfilepwent: malformed password entry (passwd too short)\n"));
continue;
}
if (p[32] != ':') {
DEBUG(0, ("getsmbfilepwent: malformed password entry (no terminating :)\n"));
continue;
}
if (!strncasecmp((char *) p, "NO PASSWORD", 11)) {
pw_buf.smb_passwd = NULL;
pw_buf.acct_ctrl |= ACB_PWNOTREQ;
} else {
if (!pdb_gethexpwd((char *)p, smbpwd)) {
DEBUG(0, ("getsmbfilepwent: Malformed Lanman password entry (non hex chars)\n"));
continue;
}
pw_buf.smb_passwd = smbpwd;
}
/*
* Now check if the NT compatible password is
* available.
*/
pw_buf.smb_nt_passwd = NULL;
p += 33; /* Move to the first character of the line after
the lanman password. */
if ((linebuf_len >= (PTR_DIFF(p, linebuf) + 33)) && (p[32] == ':')) {
if (*p != '*' && *p != 'X') {
if(pdb_gethexpwd((char *)p,smbntpwd))
pw_buf.smb_nt_passwd = smbntpwd;
}
p += 33; /* Move to the first character of the line after
the NT password. */
}
DEBUG(5,("getsmbfilepwent: returning passwd entry for user %s, uid %ld\n",
user_name, uidval));
if (*p == '[')
{
unsigned char *end_p = (unsigned char *)strchr((char *)p, ']');
pw_buf.acct_ctrl = pdb_decode_acct_ctrl((char*)p);
/* Must have some account type set. */
if(pw_buf.acct_ctrl == 0)
pw_buf.acct_ctrl = ACB_NORMAL;
/* Now try and get the last change time. */
if(end_p)
p = end_p + 1;
if(*p == ':') {
p++;
if(*p && (StrnCaseCmp((char *)p, "LCT-", 4)==0)) {
int i;
p += 4;
for(i = 0; i < 8; i++) {
if(p[i] == '\0' || !isxdigit(p[i]))
break;
}
if(i == 8) {
/*
* p points at 8 characters of hex digits -
* read into a time_t as the seconds since
* 1970 that the password was last changed.
*/
pw_buf.pass_last_set_time = (time_t)strtol((char *)p, NULL, 16);
}
}
}
} else {
/* 'Old' style file. Fake up based on user name. */
/*
* Currently trust accounts are kept in the same
* password file as 'normal accounts'. If this changes
* we will have to fix this code. JRA.
*/
if(pw_buf.smb_name[strlen(pw_buf.smb_name) - 1] == '$') {
pw_buf.acct_ctrl &= ~ACB_NORMAL;
pw_buf.acct_ctrl |= ACB_WSTRUST;
}
}
return &pw_buf;
}
DEBUG(5,("getsmbfilepwent: end of file reached.\n"));
return NULL;
}
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
/*************************************************************************
Routine to return the next entry in the smbpasswd list.
this function is a nice, messy combination of reading:
- the smbpasswd file
- the unix password database
- smb.conf options (not done at present).
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
*************************************************************************/
static struct sam_passwd *getsmbfile21pwent(void *vp)
{
struct smb_passwd *pw_buf = getsmbfilepwent(vp);
static struct sam_passwd user;
struct passwd *pwfile;
static pstring full_name;
static pstring home_dir;
static pstring home_drive;
static pstring logon_script;
static pstring profile_path;
static pstring acct_desc;
static pstring workstations;
DEBUG(5,("getsmbfile21pwent\n"));
if (pw_buf == NULL) return NULL;
pwfile = sys_getpwnam(pw_buf->smb_name);
if (pwfile == NULL)
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
{
DEBUG(0,("getsmbfile21pwent: smbpasswd database is corrupt!\n"));
DEBUG(0,("getsmbfile21pwent: username %s not in unix passwd database!\n", pw_buf->smb_name));
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
return NULL;
}
pdb_init_sam(&user);
pstrcpy(samlogon_user, pw_buf->smb_name);
if (samlogon_user[strlen(samlogon_user)-1] != '$')
{
/* XXXX hack to get standard_sub_basic() to use sam logon username */
/* possibly a better way would be to do a become_user() call */
sam_logon_in_ssb = True;
user.smb_userid = pw_buf->smb_userid;
user.smb_grpid = pwfile->pw_gid;
user.user_rid = pdb_uid_to_user_rid (user.smb_userid);
user.group_rid = pdb_gid_to_group_rid(user.smb_grpid );
pstrcpy(full_name , pwfile->pw_gecos );
pstrcpy(logon_script , lp_logon_script ());
pstrcpy(profile_path , lp_logon_path ());
pstrcpy(home_drive , lp_logon_drive ());
pstrcpy(home_dir , lp_logon_home ());
pstrcpy(acct_desc , "");
pstrcpy(workstations , "");
sam_logon_in_ssb = False;
}
else
ldap back-end database development Makefile: created PASSBD_OBJ group includes.h: added #ifdef USE_LDAP to #include <ldap> headers ldap.c: - renamed "_machine" to "_trust" everywhere. - added sam_passwd support routines - removed get_ldappwd_entry function: replaced with get_sampwd_entry - removed getldappwnam/uid: replaced with getsampwnam/uid - other messing about bits which are probably going to annoy the hell out of jean-francois (sorry!) mkproto.awk: - added stuff to wrap ldap.c protos with #ifdef USE_LDAP - added uid_t and gid_t return results to the prototype generation passdb.c: - created getsam21pwent, add_sam21pwd_entry, mod_sam21pwd_entry. - modified getsampwnam/uid and created getsam21pwnam/rid functions to replace the local get_smbpwd_entry() and get_ldappwd_entry() functions, which jeremy didn't like anyway because they were dual-purpose. - added utility routines which are or may be useful to all the password database routines. password.c: - renamed "machine_" to "trust_" everywhere. smbpass.c: - removed get_smbpwd_entry function: replaced it with get_sampwd_entry functions in passdb.c - moved code that decoded acct_ctrl into passdb.c - moved encode_acct_ctrl into passdb.c - removed getsmbpwnam/uid: replaced with getsampwnam/uid - renamed "machine_" to "trust_" everywhere. smbpasswd.c: - renamed "machine_" to "trust_" everywhere. util.c: - moved gethexpwd function into passdb.c lib/rpc/server/srv_util.c: - moved user_rid_to_uid, group_rid_to_rid etc etc into passdb.c (This used to be commit 673ab50c4c2c25db355d90efde3a6bfbb4d8369e)
1998-05-11 19:56:01 +04:00
{
user.smb_userid = pw_buf->smb_userid;
user.smb_grpid = pwfile->pw_gid;
user.user_rid = pdb_uid_to_user_rid (user.smb_userid);
user.group_rid = DOMAIN_GROUP_RID_USERS; /* lkclXXXX this is OBSERVED behaviour by NT PDCs, enforced here. */
pstrcpy(full_name , "");
pstrcpy(logon_script , "");
pstrcpy(profile_path , "");
pstrcpy(home_drive , "");
pstrcpy(home_dir , "");
pstrcpy(acct_desc , "");
pstrcpy(workstations , "");
- group database API. oops and oh dear, the threat has been carried out: the pre-alpha "domain group" etc parameters have disappeared. - interactive debug detection - re-added mem_man (andrew's memory management, detects memory corruption) - american spellings of "initialise" replaced with english spelling of "initialise". - started on "lookup_name()" and "lookup_sid()" functions. proper ones. - moved lots of functions around. created some modules of commonly used code. e.g the password file locking code, which is used in groupfile.c and aliasfile.c and smbpass.c - moved RID_TYPE_MASK up another bit. this is really unfortunate, but there is no other "fast" way to identify users from groups from aliases. i do not believe that this code saves us anything (the multipliers) and puts us at a disadvantage (reduces the useable rid space). the designers of NT aren't silly: if they can get away with a user- interface-speed LsaLookupNames / LsaLookupSids, then so can we. i spoke with isaac at the cifs conference, the only time for example that they do a security context check is on file create. certainly not on individual file reads / writes, which would drastically hit their performance and ours, too. - renamed myworkgroup to global_sam_name, amongst other things, when used in the rpc code. there is also a global_member_name, as we are always responsible for a SAM database, the scope of which is limited by the role of the machine (e.g if a member of a workgroup, your SAM is for _local_ logins only, and its name is the name of your server. you even still have a SID. see LsaQueryInfoPolicy, levels 3 and 5). - updated functionality of groupname.c to be able to cope with names like DOMAIN\group and SERVER\alias. used this code to be able to do aliases as well as groups. this code may actually be better off being used in username mapping, too. - created a connect to serverlist function in clientgen.c and used it in password.c - initialisation in server.c depends on the role of the server. well, it does now. - rpctorture. smbtorture. EXERCISE EXTREME CAUTION. (This used to be commit 0d21e1e6090b933f396c764af535ca3388a562db)
1998-11-17 19:19:04 +03:00
}
user.smb_name = pw_buf->smb_name;
user.full_name = full_name;
user.home_dir = home_dir;
user.dir_drive = home_drive;
user.logon_script = logon_script;
user.profile_path = profile_path;
user.acct_desc = acct_desc;
user.workstations = workstations;
user.unknown_str = NULL; /* don't know, yet! */
user.munged_dial = NULL; /* "munged" dial-back telephone number */
user.smb_nt_passwd = pw_buf->smb_nt_passwd;
user.smb_passwd = pw_buf->smb_passwd;
user.acct_ctrl = pw_buf->acct_ctrl;
user.unknown_3 = 0xffffff; /* don't know */
user.logon_divs = 168; /* hours per week */
user.hours_len = 21; /* 21 times 8 bits = 168 */
memset(user.hours, 0xff, user.hours_len); /* available at all hours */
user.unknown_5 = 0x00020000; /* don't know */
user.unknown_5 = 0x000004ec; /* don't know */
return &user;
}
/*************************************************************************
Return the current position in the smbpasswd list as an SMB_BIG_UINT.
This must be treated as an opaque token.
*************************************************************************/
static SMB_BIG_UINT getsmbfilepwpos(void *vp)
{
return (SMB_BIG_UINT)sys_ftell((FILE *)vp);
}
/*************************************************************************
Set the current position in the smbpasswd list from an SMB_BIG_UINT.
This must be treated as an opaque token.
*************************************************************************/
static BOOL setsmbfilepwpos(void *vp, SMB_BIG_UINT tok)
{
return !sys_fseek((FILE *)vp, (SMB_OFF_T)tok, SEEK_SET);
}
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
/************************************************************************
Create a new smbpasswd entry - malloced space returned.
*************************************************************************/
char *format_new_smbpasswd_entry(struct smb_passwd *newpwd)
{
int new_entry_length;
char *new_entry;
char *p;
int i;
new_entry_length = strlen(newpwd->smb_name) + 1 + 15 + 1 + 32 + 1 + 32 + 1 + NEW_PW_FORMAT_SPACE_PADDED_LEN + 1 + 13 + 2;
if((new_entry = (char *)malloc( new_entry_length )) == NULL) {
DEBUG(0, ("format_new_smbpasswd_entry: Malloc failed adding entry for user %s.\n", newpwd->smb_name ));
return NULL;
}
slprintf(new_entry, new_entry_length - 1, "%s:%u:", newpwd->smb_name, (unsigned)newpwd->smb_userid);
p = &new_entry[strlen(new_entry)];
if(newpwd->smb_passwd != NULL) {
for( i = 0; i < 16; i++) {
slprintf((char *)&p[i*2], new_entry_length - (p - new_entry) - 1, "%02X", newpwd->smb_passwd[i]);
}
} else {
i=0;
if(newpwd->acct_ctrl & ACB_PWNOTREQ)
safe_strcpy((char *)p, "NO PASSWORDXXXXXXXXXXXXXXXXXXXXX", new_entry_length - 1 - (p - new_entry));
else
safe_strcpy((char *)p, "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX", new_entry_length - 1 - (p - new_entry));
}
p += 32;
*p++ = ':';
if(newpwd->smb_nt_passwd != NULL) {
for( i = 0; i < 16; i++) {
slprintf((char *)&p[i*2], new_entry_length - 1 - (p - new_entry), "%02X", newpwd->smb_nt_passwd[i]);
}
} else {
if(newpwd->acct_ctrl & ACB_PWNOTREQ)
safe_strcpy((char *)p, "NO PASSWORDXXXXXXXXXXXXXXXXXXXXX", new_entry_length - 1 - (p - new_entry));
else
safe_strcpy((char *)p, "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX", new_entry_length - 1 - (p - new_entry));
}
p += 32;
*p++ = ':';
/* Add the account encoding and the last change time. */
slprintf((char *)p, new_entry_length - 1 - (p - new_entry), "%s:LCT-%08X:\n",
pdb_encode_acct_ctrl(newpwd->acct_ctrl, NEW_PW_FORMAT_SPACE_PADDED_LEN),
(uint32)newpwd->pass_last_set_time);
return new_entry;
}
/************************************************************************
Routine to add an entry to the smbpasswd file.
*************************************************************************/
static BOOL add_smbfilepwd_entry(struct smb_passwd *newpwd)
{
char *pfile = lp_smb_passwd_file();
struct smb_passwd *pwd = NULL;
FILE *fp = NULL;
int wr_len;
int fd;
size_t new_entry_length;
char *new_entry;
SMB_OFF_T offpos;
/* Open the smbpassword file - for update. */
fp = startsmbfilepwent(True);
if (fp == NULL) {
DEBUG(0, ("add_smbfilepwd_entry: unable to open file.\n"));
return False;
}
/*
* Scan the file, a line at a time and check if the name matches.
*/
while ((pwd = getsmbfilepwent(fp)) != NULL) {
if (strequal(newpwd->smb_name, pwd->smb_name)) {
DEBUG(0, ("add_smbfilepwd_entry: entry with name %s already exists\n", pwd->smb_name));
endsmbfilepwent(fp);
return False;
}
}
/* Ok - entry doesn't exist. We can add it */
/* Create a new smb passwd entry and set it to the given password. */
/*
* The add user write needs to be atomic - so get the fd from
* the fp and do a raw write() call.
*/
fd = fileno(fp);
if((offpos = sys_lseek(fd, 0, SEEK_END)) == -1) {
DEBUG(0, ("add_smbfilepwd_entry(sys_lseek): Failed to add entry for user %s to file %s. \
Error was %s\n", newpwd->smb_name, pfile, strerror(errno)));
endsmbfilepwent(fp);
return False;
}
if((new_entry = format_new_smbpasswd_entry(newpwd)) == NULL) {
DEBUG(0, ("add_smbfilepwd_entry(malloc): Failed to add entry for user %s to file %s. \
Error was %s\n", newpwd->smb_name, pfile, strerror(errno)));
endsmbfilepwent(fp);
return False;
}
new_entry_length = strlen(new_entry);
#ifdef DEBUG_PASSWORD
DEBUG(100, ("add_smbfilepwd_entry(%d): new_entry_len %d made line |%s|",
fd, new_entry_length, new_entry));
#endif
if ((wr_len = write(fd, new_entry, new_entry_length)) != new_entry_length) {
DEBUG(0, ("add_smbfilepwd_entry(write): %d Failed to add entry for user %s to file %s. \
Error was %s\n", wr_len, newpwd->smb_name, pfile, strerror(errno)));
/* Remove the entry we just wrote. */
if(sys_ftruncate(fd, offpos) == -1) {
DEBUG(0, ("add_smbfilepwd_entry: ERROR failed to ftruncate file %s. \
Error was %s. Password file may be corrupt ! Please examine by hand !\n",
newpwd->smb_name, strerror(errno)));
}
endsmbfilepwent(fp);
free(new_entry);
return False;
}
free(new_entry);
endsmbfilepwent(fp);
return True;
}
/************************************************************************
Routine to search the smbpasswd file for an entry matching the username.
and then modify its password entry. We can't use the startsmbpwent()/
getsmbpwent()/endsmbpwent() interfaces here as we depend on looking
in the actual file to decide how much room we have to write data.
override = False, normal
override = True, override XXXXXXXX'd out password or NO PASS
************************************************************************/
static BOOL mod_smbfilepwd_entry(struct smb_passwd* pwd, BOOL override)
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
{
/* Static buffers we will return. */
static pstring user_name;
char linebuf[256];
char readbuf[1024];
unsigned char c;
fstring ascii_p16;
fstring encode_bits;
unsigned char *p = NULL;
size_t linebuf_len = 0;
FILE *fp;
int lockfd;
char *pfile = lp_smb_passwd_file();
BOOL found_entry = False;
BOOL got_pass_last_set_time = False;
SMB_OFF_T pwd_seekpos = 0;
int i;
int wr_len;
int fd;
if (!*pfile) {
DEBUG(0, ("No SMB password file set\n"));
return False;
}
DEBUG(10, ("mod_smbfilepwd_entry: opening file %s\n", pfile));
fp = sys_fopen(pfile, "r+");
if (fp == NULL) {
DEBUG(0, ("mod_smbfilepwd_entry: unable to open file %s\n", pfile));
return False;
}
/* Set a buffer to do more efficient reads */
setvbuf(fp, readbuf, _IOFBF, sizeof(readbuf));
lockfd = fileno(fp);
if (!pw_file_lock(lockfd, F_WRLCK, 5, &pw_file_lock_depth)) {
DEBUG(0, ("mod_smbfilepwd_entry: unable to lock file %s\n", pfile));
fclose(fp);
return False;
}
/* Make sure it is only rw by the owner */
chmod(pfile, 0600);
/* We have a write lock on the file. */
/*
* Scan the file, a line at a time and check if the name matches.
*/
while (!feof(fp)) {
pwd_seekpos = sys_ftell(fp);
linebuf[0] = '\0';
fgets(linebuf, sizeof(linebuf), fp);
if (ferror(fp)) {
pw_file_unlock(lockfd, &pw_file_lock_depth);
fclose(fp);
return False;
}
/*
* Check if the string is terminated with a newline - if not
* then we must keep reading and discard until we get one.
*/
linebuf_len = strlen(linebuf);
if (linebuf[linebuf_len - 1] != '\n') {
c = '\0';
while (!ferror(fp) && !feof(fp)) {
c = fgetc(fp);
if (c == '\n') {
break;
}
}
} else {
linebuf[linebuf_len - 1] = '\0';
}
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
#ifdef DEBUG_PASSWORD
DEBUG(100, ("mod_smbfilepwd_entry: got line |%s|\n", linebuf));
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
#endif
if ((linebuf[0] == 0) && feof(fp)) {
DEBUG(4, ("mod_smbfilepwd_entry: end of file reached\n"));
break;
}
/*
* The line we have should be of the form :-
*
* username:uid:[32hex bytes]:....other flags presently
* ignored....
*
* or,
*
* username:uid:[32hex bytes]:[32hex bytes]:[attributes]:LCT-XXXXXXXX:...ignored.
*
* if Windows NT compatible passwords are also present.
*/
if (linebuf[0] == '#' || linebuf[0] == '\0') {
DEBUG(6, ("mod_smbfilepwd_entry: skipping comment or blank line\n"));
continue;
}
p = (unsigned char *) strchr(linebuf, ':');
if (p == NULL) {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (no :)\n"));
continue;
}
/*
* As 256 is shorter than a pstring we don't need to check
* length here - if this ever changes....
*/
strncpy(user_name, linebuf, PTR_DIFF(p, linebuf));
user_name[PTR_DIFF(p, linebuf)] = '\0';
if (strequal(user_name, pwd->smb_name)) {
found_entry = True;
break;
}
}
if (!found_entry) {
pw_file_unlock(lockfd, &pw_file_lock_depth);
fclose(fp);
return False;
}
DEBUG(6, ("mod_smbfilepwd_entry: entry exists\n"));
/* User name matches - get uid and password */
p++; /* Go past ':' */
if (!isdigit(*p)) {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (uid not number)\n"));
pw_file_unlock(lockfd, &pw_file_lock_depth);
fclose(fp);
return False;
}
while (*p && isdigit(*p))
p++;
if (*p != ':') {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (no : after uid)\n"));
pw_file_unlock(lockfd, &pw_file_lock_depth);
fclose(fp);
return False;
}
/*
* Now get the password value - this should be 32 hex digits
* which are the ascii representations of a 16 byte string.
* Get two at a time and put them into the password.
*/
p++;
/* Record exact password position */
pwd_seekpos += PTR_DIFF(p, linebuf);
if (!override && (*p == '*' || *p == 'X')) {
/* Password deliberately invalid - end here. */
DEBUG(10, ("mod_smbfilepwd_entry: entry invalidated for user %s\n", user_name));
pw_file_unlock(lockfd, &pw_file_lock_depth);
fclose(fp);
return False;
}
if (linebuf_len < (PTR_DIFF(p, linebuf) + 33)) {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (passwd too short)\n"));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return (False);
}
if (p[32] != ':') {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (no terminating :)\n"));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
if (!override && (*p == '*' || *p == 'X')) {
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
/* Now check if the NT compatible password is
available. */
p += 33; /* Move to the first character of the line after
the lanman password. */
if (linebuf_len < (PTR_DIFF(p, linebuf) + 33)) {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (passwd too short)\n"));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return (False);
}
if (p[32] != ':') {
DEBUG(0, ("mod_smbfilepwd_entry: malformed password entry (no terminating :)\n"));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
/*
* Now check if the account info and the password last
* change time is available.
*/
p += 33; /* Move to the first character of the line after
the NT password. */
Makefile.in: Fixed bug with continuation line causing proto to fail. Added $(PROGS) $(SPROGS) as targets for make clean. acconfig.h: Added HAVE_IRIX_SPECIFIC_CAPABILITIES. configure.in: Added sys/capability.h header check. Added function checks for srandom random srand rand. Added HAVE_IRIX_SPECIFIC_CAPABILITIES test. includes.h: Added #include <sys/capability.h>. ntdomain.h: Moved struct acct_info into here from smb.h smb.h: Added KERNEL_OPLOCK_CAPABILITY define. Moved enum action_type into rpcclient.h Moved struct cli_state into client.h Moved struct nt_client_info, struct tar_client_info, struct client_info into rpcclient.h lib/genrand.c: Changed to use sys_random() & friends. lib/smbrun.c: Lose capabilities after fork. lib/system.c: Added set_process_capability(), set_inherited_process_capability() sys_random(), sys_srandom(). lib/util.c: Added Ander's EFBIG lock check to fcntl_lock for 64 bit access to an 32 bit mounted NFS filesystem. nmbd/nmbd.c: Changed to use sys_random() & friends. nmbd/nmbd_browsesync.c: Changed to use sys_random() & friends. passdb/ldap.c: Missed one pdb_encode_acct_ctrl call. passdb/passdb.c: Changed to Ander's code for ' ' characters. passdb/smbpass.c: Added Ander's code to reset ACB_PWNOTREQ. script/mkproto.awk: Added 'long' to prototypes. smbd/chgpasswd.c: Lose capabilities after fork. smbd/open.c: Do the mmap *after* the kernel oplock. smbd/oplock.c: Removed stub code from kernel oplock path. Added set_process_capability(), set_inherited_process_capability() calls. smbd/reply.c: Initialize count = 0, offset = 0. smbd/server.c: Added set_process_capability(), set_inherited_process_capability() calls. tests/summary.c: Ensure we have RANDOM or RAND. utils/smbpasswd.c: Added Ander's code to reset ACB_PWNOTREQ. utils/torture.c: Changed to use sys_random() & friends. Jeremy. (This used to be commit e8be306f23963ac00b1a383ebe0cc1421529fb02)
1998-09-26 03:40:49 +04:00
/*
* If both NT and lanman passwords are provided - reset password
* not required flag.
*/
if(pwd->smb_passwd != NULL || pwd->smb_nt_passwd != NULL) {
/* Reqiure password in the future (should ACB_DISABLED also be reset?) */
pwd->acct_ctrl &= ~(ACB_PWNOTREQ);
}
if (*p == '[') {
i = 0;
encode_bits[i++] = *p++;
while((linebuf_len > PTR_DIFF(p, linebuf)) && (*p != ']'))
encode_bits[i++] = *p++;
encode_bits[i++] = ']';
encode_bits[i++] = '\0';
if(i == NEW_PW_FORMAT_SPACE_PADDED_LEN) {
/*
* We are using a new format, space padded
* acct ctrl field. Encode the given acct ctrl
* bits into it.
*/
fstrcpy(encode_bits, pdb_encode_acct_ctrl(pwd->acct_ctrl, NEW_PW_FORMAT_SPACE_PADDED_LEN));
} else {
/*
* If using the old format and the ACB_DISABLED or
* ACB_PWNOTREQ are set then set the lanman and NT passwords to NULL
* here as we have no space to encode the change.
*/
if(pwd->acct_ctrl & (ACB_DISABLED|ACB_PWNOTREQ)) {
pwd->smb_passwd = NULL;
pwd->smb_nt_passwd = NULL;
}
}
/* Go past the ']' */
if(linebuf_len > PTR_DIFF(p, linebuf))
p++;
if((linebuf_len > PTR_DIFF(p, linebuf)) && (*p == ':')) {
p++;
/* We should be pointing at the LCT entry. */
if((linebuf_len > (PTR_DIFF(p, linebuf) + 13)) && (StrnCaseCmp((char *)p, "LCT-", 4) == 0)) {
p += 4;
for(i = 0; i < 8; i++) {
if(p[i] == '\0' || !isxdigit(p[i]))
break;
}
if(i == 8) {
/*
* p points at 8 characters of hex digits -
* read into a time_t as the seconds since
* 1970 that the password was last changed.
*/
got_pass_last_set_time = True;
} /* i == 8 */
} /* *p && StrnCaseCmp() */
} /* p == ':' */
} /* p == '[' */
/* Entry is correctly formed. */
/* Create the 32 byte representation of the new p16 */
if(pwd->smb_passwd != NULL) {
for (i = 0; i < 16; i++) {
slprintf(&ascii_p16[i*2], sizeof(fstring) - 1, "%02X", (uchar) pwd->smb_passwd[i]);
}
} else {
if(pwd->acct_ctrl & ACB_PWNOTREQ)
fstrcpy(ascii_p16, "NO PASSWORDXXXXXXXXXXXXXXXXXXXXX");
else
fstrcpy(ascii_p16, "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX");
}
/* Add on the NT md4 hash */
ascii_p16[32] = ':';
wr_len = 66;
if (pwd->smb_nt_passwd != NULL) {
for (i = 0; i < 16; i++) {
slprintf(&ascii_p16[(i*2)+33], sizeof(fstring) - 1, "%02X", (uchar) pwd->smb_nt_passwd[i]);
}
} else {
if(pwd->acct_ctrl & ACB_PWNOTREQ)
fstrcpy(&ascii_p16[33], "NO PASSWORDXXXXXXXXXXXXXXXXXXXXX");
else
fstrcpy(&ascii_p16[33], "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX");
}
ascii_p16[65] = ':';
ascii_p16[66] = '\0'; /* null-terminate the string so that strlen works */
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
/* Add on the account info bits and the time of last
password change. */
pwd->pass_last_set_time = time(NULL);
if(got_pass_last_set_time) {
slprintf(&ascii_p16[strlen(ascii_p16)],
sizeof(ascii_p16)-(strlen(ascii_p16)+1),
"%s:LCT-%08X:",
encode_bits, (uint32)pwd->pass_last_set_time );
wr_len = strlen(ascii_p16);
}
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
#ifdef DEBUG_PASSWORD
DEBUG(100,("mod_smbfilepwd_entry: "));
dump_data(100, ascii_p16, wr_len);
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
#endif
if(wr_len > sizeof(linebuf)) {
DEBUG(0, ("mod_smbfilepwd_entry: line to write (%d) is too long.\n", wr_len+1));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return (False);
}
/*
* Do an atomic write into the file at the position defined by
* seekpos.
*/
/* The mod user write needs to be atomic - so get the fd from
the fp and do a raw write() call.
*/
fd = fileno(fp);
if (sys_lseek(fd, pwd_seekpos - 1, SEEK_SET) != pwd_seekpos - 1) {
DEBUG(0, ("mod_smbfilepwd_entry: seek fail on file %s.\n", pfile));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
/* Sanity check - ensure the areas we are writing are framed by ':' */
if (read(fd, linebuf, wr_len+1) != wr_len+1) {
DEBUG(0, ("mod_smbfilepwd_entry: read fail on file %s.\n", pfile));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
if ((linebuf[0] != ':') || (linebuf[wr_len] != ':')) {
DEBUG(0, ("mod_smbfilepwd_entry: check on passwd file %s failed.\n", pfile));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
if (sys_lseek(fd, pwd_seekpos, SEEK_SET) != pwd_seekpos) {
DEBUG(0, ("mod_smbfilepwd_entry: seek fail on file %s.\n", pfile));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
if (write(fd, ascii_p16, wr_len) != wr_len) {
DEBUG(0, ("mod_smbfilepwd_entry: write failed in passwd file %s\n", pfile));
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return False;
}
pw_file_unlock(lockfd,&pw_file_lock_depth);
fclose(fp);
return True;
following a cvs error, i am rewriting this monster-commit. with bad grace. Modified Files: --------------- Makefile: adding extra files ipc.c : send_trans_reply() - alignment issue. this makes the alignment the same as that in NT. this should be looked at by people who understand the SMB stuff better than i. api_fd_commands[] - added samr and wkssvc pipes. loadparm.c : lp_domain_controller() changed to mean "samba is a domain controller". it's a "yes/no" parameter, now. no, it isn't used _anywhere_. namedbwork.c nameelect.c : if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the host _and_ workgroup announcements. yes, you must do both: nt does. namelogon.c : important NETLOGON bug in SAMLOGON request parsing, which may be the source of some people's problems with logging on to the Samba PDC. password.c : get_smbpwnam() renamed to get_smbpwd_entry(). pipes.c : added samr and wkssvc pipes. proto.h : usual. can we actually _remove_ proto.h from the cvs tree, and have it as one of the Makefile dependencies, or something? reply.c : get_smbpwnam() renamed to get_smbpwd_entry() - also changed response error code when logging in from a WORKSTATION$ account. yes, paul is right: we need to know when to return the right error code, and why. server.c : added call to reset_chain_pnum(). #ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif. jeremy, you'd be proud: i did a compile without NTDOMAIN, and caught a link error for this function. smb.h : defines and structures for samr and wkssvc pipes. smbpass.c : modified get_smbpwnam() to get_smbpwd_entry() and it now takes two arguments. one for the name; if this is null, it looks up by smb_userid instead. oh, by the way, smb_userids are actually domain relative ids (RIDs). concatenate a RID with the domain SID, and you have an internet globally unique way of identifying a user. we're using RIDs in the wrong way.... added mod_smbpwnam() function. this was based on code in smbpasswd.c rpc_pipes/lsaparse.c : added enum trusted domain parsing. this is incomplete: i need a packet trace to write it properly. rpc_pipes/pipe_hnd.c : added reset_chain_pnum() function. rpc_pipes/pipenetlog.c : get_smbpwnam() function renamed to get_smbpwd_entry(). arcfour() issues. removed capability of get_md4pw() function to automatically add workstation accounts. this should either be done using smbpasswd -add MACHINE$, or by using \PIPE\samr. rpc_pipes/pipe_util.c : create_pol_hnd() - creates a unique LSA Policy Handle. overkill function: uses a 64 bit sequence number; current unix time and the smbd pid. rpc_pipes/smbparse.c : arcfour() issues. smb_io_unistr2() should advance by uni_str_len not uni_max_len. smb_io_smb_hdr_rb() - request bind uses uint16 for the context id, and uint8 for the num_syntaxes. oops, i put these both as uint32s. Added Files: ------------ rpc_pipes/lsa_hnd.c : on the samr pipe, allocate and associate an LSA Policy Handle with a SID. you receive queries with the LSA Policy Handle, and have to turn this back into a SID in order to answer the query... rpc_pipes/pipesamr.c rpc_pipes/samrparse.c \PIPE\samr processing. samr i presume is the SAM Replication pipe. rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c \PIPE\wkssvc processing. the Workstation Service pipe? holy cow. (This used to be commit 1bd084b3e690eb26a1006d616075e53d711ecd2f)
1997-11-07 02:03:58 +03:00
}
/************************************************************************
Routine to delete an entry in the smbpasswd file by name.
*************************************************************************/
static BOOL del_smbfilepwd_entry(const char *name)
{
char *pfile = lp_smb_passwd_file();
pstring pfile2;
struct smb_passwd *pwd = NULL;
FILE *fp = NULL;
FILE *fp_write = NULL;
int pfile2_lockdepth = 0;
slprintf(pfile2, sizeof(pfile2)-1, "%s.%u", pfile, (unsigned)sys_getpid() );
/*
* Open the smbpassword file - for update. It needs to be update
* as we need any other processes to wait until we have replaced
* it.
*/
if((fp = startsmbfilepwent(True)) == NULL) {
DEBUG(0, ("del_smbfilepwd_entry: unable to open file %s.\n", pfile));
return False;
}
/*
* Create the replacement password file.
*/
if((fp_write = startsmbfilepwent_internal(pfile2, PWF_CREATE, &pfile2_lockdepth)) == NULL) {
DEBUG(0, ("del_smbfilepwd_entry: unable to open file %s.\n", pfile));
endsmbfilepwent(fp);
return False;
}
/*
* Scan the file, a line at a time and check if the name matches.
*/
while ((pwd = getsmbfilepwent(fp)) != NULL) {
char *new_entry;
size_t new_entry_length;
if (strequal(name, pwd->smb_name)) {
DEBUG(10, ("add_smbfilepwd_entry: found entry with name %s - deleting it.\n", name));
continue;
}
/*
* We need to copy the entry out into the second file.
*/
if((new_entry = format_new_smbpasswd_entry(pwd)) == NULL) {
DEBUG(0, ("del_smbfilepwd_entry(malloc): Failed to copy entry for user %s to file %s. \
Error was %s\n", pwd->smb_name, pfile2, strerror(errno)));
unlink(pfile2);
endsmbfilepwent(fp);
endsmbfilepwent_internal(fp_write,&pfile2_lockdepth);
return False;
}
new_entry_length = strlen(new_entry);
if(fwrite(new_entry, 1, new_entry_length, fp_write) != new_entry_length) {
DEBUG(0, ("del_smbfilepwd_entry(write): Failed to copy entry for user %s to file %s. \
Error was %s\n", pwd->smb_name, pfile2, strerror(errno)));
unlink(pfile2);
endsmbfilepwent(fp);
endsmbfilepwent_internal(fp_write,&pfile2_lockdepth);
free(new_entry);
return False;
}
free(new_entry);
}
/*
* Ensure pfile2 is flushed before rename.
*/
if(fflush(fp_write) != 0) {
DEBUG(0, ("del_smbfilepwd_entry: Failed to flush file %s. Error was %s\n", pfile2, strerror(errno)));
endsmbfilepwent(fp);
endsmbfilepwent_internal(fp_write,&pfile2_lockdepth);
return False;
}
/*
* Do an atomic rename - then release the locks.
*/
if(rename(pfile2,pfile) != 0) {
unlink(pfile2);
}
endsmbfilepwent(fp);
endsmbfilepwent_internal(fp_write,&pfile2_lockdepth);
return True;
}
/*
* Stub functions - implemented in terms of others.
*/
static BOOL mod_smbfile21pwd_entry(struct sam_passwd* pwd, BOOL override)
{
return mod_smbfilepwd_entry(pdb_sam_to_smb(pwd), override);
}
static BOOL add_smbfile21pwd_entry(struct sam_passwd *newpwd)
{
return add_smbfilepwd_entry(pdb_sam_to_smb(newpwd));
}
static struct sam_disp_info *getsmbfiledispnam(char *name)
{
return pdb_sam_to_dispinfo(getsam21pwnam(name));
}
static struct sam_disp_info *getsmbfiledisprid(uint32 rid)
{
return pdb_sam_to_dispinfo(getsam21pwrid(rid));
}
static struct sam_disp_info *getsmbfiledispent(void *vp)
{
return pdb_sam_to_dispinfo(getsam21pwent(vp));
}
static struct passdb_ops file_ops = {
startsmbfilepwent,
endsmbfilepwent,
getsmbfilepwpos,
setsmbfilepwpos,
iterate_getsmbpwnam, /* In passdb.c */
iterate_getsmbpwuid, /* In passdb.c */
iterate_getsmbpwrid, /* In passdb.c */
getsmbfilepwent,
add_smbfilepwd_entry,
mod_smbfilepwd_entry,
del_smbfilepwd_entry,
getsmbfile21pwent,
iterate_getsam21pwnam,
iterate_getsam21pwuid,
iterate_getsam21pwrid,
add_smbfile21pwd_entry,
mod_smbfile21pwd_entry,
getsmbfiledispnam,
getsmbfiledisprid,
getsmbfiledispent
};
struct passdb_ops *file_initialize_password_db(void)
{
return &file_ops;
}
#else
/* Do *NOT* make this function static. It breaks the compile on gcc. JRA */
void smbpass_dummy_function(void) { } /* stop some compilers complaining */
#endif /* USE_SMBPASS_DB */