1997-10-21 13:12:41 +04:00
/*
2002-01-30 09:08:46 +03:00
Unix SMB / CIFS implementation .
1997-10-21 13:12:41 +04:00
SMB client generic functions
1999-12-13 16:27:58 +03:00
Copyright ( C ) Andrew Tridgell 1994 - 1998
1997-10-21 13:12:41 +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 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 0213 9 , USA .
*/
1998-07-29 07:08:05 +04:00
# define NO_SYSLOG
1997-10-21 13:12:41 +04:00
# include "includes.h"
2002-10-16 01:30:06 +04:00
/****************************************************************************
Change the timeout ( in milliseconds ) .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
unsigned int cli_set_timeout ( struct cli_state * cli , unsigned int timeout )
{
unsigned int old_timeout = cli - > timeout ;
cli - > timeout = timeout ;
return old_timeout ;
}
2002-10-01 06:11:35 +04:00
/****************************************************************************
Change the port number used to call on .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
1999-01-25 04:46:14 +03:00
int cli_set_port ( struct cli_state * cli , int port )
{
2001-08-25 00:11:09 +04:00
cli - > port = port ;
return port ;
1999-01-25 04:46:14 +03:00
}
1998-10-19 06:48:57 +04:00
2002-07-15 14:35:28 +04:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Read an smb from a fd ignoring all keepalive packets . Note that the buffer
* MUST * be of size BUFFER_SIZE + SAFETY_MARGIN .
The timeout is in milliseconds
This is exactly the same as receive_smb except that it never returns
a session keepalive packet ( just as receive_smb used to do ) .
receive_smb was changed to return keepalives as the oplock processing means this call
should never go into a blocking read .
2002-07-15 14:35:28 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
static BOOL client_receive_smb ( int fd , char * buffer , unsigned int timeout )
{
2002-09-25 19:19:00 +04:00
BOOL ret ;
for ( ; ; ) {
2003-10-22 01:19:00 +04:00
ret = receive_smb_raw ( fd , buffer , timeout ) ;
2002-07-15 14:35:28 +04:00
2002-09-25 19:19:00 +04:00
if ( ! ret ) {
DEBUG ( 10 , ( " client_receive_smb failed \n " ) ) ;
show_msg ( buffer ) ;
return ret ;
}
/* Ignore session keepalive packets. */
if ( CVAL ( buffer , 0 ) ! = SMBkeepalive )
break ;
}
show_msg ( buffer ) ;
return ret ;
}
2002-07-15 14:35:28 +04:00
1999-10-08 02:10:29 +04:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Recv an smb .
1999-10-08 02:10:29 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
2000-04-25 18:04:06 +04:00
BOOL cli_receive_smb ( struct cli_state * cli )
1999-10-08 02:10:29 +04:00
{
2002-10-01 05:41:20 +04:00
extern int smb_read_error ;
1999-12-13 16:27:58 +03:00
BOOL ret ;
2001-11-21 14:04:49 +03:00
/* fd == -1 causes segfaults -- Tom (tom@ninja.nl) */
2002-09-25 19:19:00 +04:00
if ( cli - > fd = = - 1 )
return False ;
2001-11-21 14:04:49 +03:00
1999-12-13 16:27:58 +03:00
again :
ret = client_receive_smb ( cli - > fd , cli - > inbuf , cli - > timeout ) ;
if ( ret ) {
/* it might be an oplock break request */
2000-01-10 16:40:27 +03:00
if ( ! ( CVAL ( cli - > inbuf , smb_flg ) & FLAG_REPLY ) & &
CVAL ( cli - > inbuf , smb_com ) = = SMBlockingX & &
1999-12-13 16:27:58 +03:00
SVAL ( cli - > inbuf , smb_vwv6 ) = = 0 & &
SVAL ( cli - > inbuf , smb_vwv7 ) = = 0 ) {
2001-06-18 12:26:15 +04:00
if ( cli - > oplock_handler ) {
int fnum = SVAL ( cli - > inbuf , smb_vwv2 ) ;
unsigned char level = CVAL ( cli - > inbuf , smb_vwv3 + 1 ) ;
if ( ! cli - > oplock_handler ( cli , fnum , level ) ) return False ;
}
1999-12-13 16:27:58 +03:00
/* try to prevent loops */
2002-01-11 22:10:25 +03:00
SCVAL ( cli - > inbuf , smb_com , 0xFF ) ;
1999-12-13 16:27:58 +03:00
goto again ;
1999-10-08 02:10:29 +04:00
}
}
2002-10-01 06:11:35 +04:00
/* If the server is not responding, note that now */
2001-11-14 08:58:51 +03:00
2002-10-01 06:11:35 +04:00
if ( ! ret ) {
2002-10-01 10:50:38 +04:00
cli - > smb_rw_error = smb_read_error ;
2002-10-01 06:11:35 +04:00
close ( cli - > fd ) ;
cli - > fd = - 1 ;
2003-02-24 05:55:00 +03:00
return ret ;
2002-10-01 06:11:35 +04:00
}
2001-11-14 08:58:51 +03:00
2003-02-24 05:55:00 +03:00
if ( ! cli_check_sign_mac ( cli ) ) {
2003-04-21 17:00:39 +04:00
DEBUG ( 0 , ( " SMB Signature verification failed on incoming packet! \n " ) ) ;
cli - > smb_rw_error = READ_BAD_SIG ;
close ( cli - > fd ) ;
cli - > fd = - 1 ;
2003-02-24 05:55:00 +03:00
return False ;
} ;
return True ;
1998-11-09 06:45:49 +03:00
}
1998-10-19 06:48:57 +04:00
/****************************************************************************
2002-07-15 14:35:28 +04:00
Send an smb to a fd .
1998-10-19 06:48:57 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2001-11-14 07:15:36 +03:00
2000-04-25 18:04:06 +04:00
BOOL cli_send_smb ( struct cli_state * cli )
1998-10-19 06:48:57 +04:00
{
size_t len ;
size_t nwritten = 0 ;
ssize_t ret ;
2001-11-21 14:04:49 +03:00
/* fd == -1 causes segfaults -- Tom (tom@ninja.nl) */
2002-07-15 14:35:28 +04:00
if ( cli - > fd = = - 1 )
return False ;
2003-07-16 09:51:10 +04:00
cli_calculate_sign_mac ( cli ) ;
2001-11-21 14:04:49 +03:00
1999-08-19 00:10:12 +04:00
len = smb_len ( cli - > outbuf ) + 4 ;
1998-10-19 06:48:57 +04:00
while ( nwritten < len ) {
ret = write_socket ( cli - > fd , cli - > outbuf + nwritten , len - nwritten ) ;
if ( ret < = 0 ) {
2002-07-15 14:35:28 +04:00
close ( cli - > fd ) ;
cli - > fd = - 1 ;
2002-10-01 10:50:38 +04:00
cli - > smb_rw_error = WRITE_ERROR ;
2002-10-01 06:11:35 +04:00
DEBUG ( 0 , ( " Error writing %d bytes to client. %d (%s) \n " ,
( int ) len , ( int ) ret , strerror ( errno ) ) ) ;
2000-04-11 17:55:53 +04:00
return False ;
1998-10-19 06:48:57 +04:00
}
nwritten + = ret ;
}
2003-08-02 04:29:45 +04:00
/* Increment the mid so we can tell between responses. */
cli - > mid + + ;
if ( ! cli - > mid )
cli - > mid + + ;
1998-10-19 06:48:57 +04:00
return True ;
}
1999-12-13 16:27:58 +03:00
/****************************************************************************
2002-07-15 14:35:28 +04:00
Setup basics in a outgoing packet .
1999-12-13 16:27:58 +03:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-07-15 14:35:28 +04:00
2000-04-25 18:04:06 +04:00
void cli_setup_packet ( struct cli_state * cli )
1999-12-13 16:27:58 +03:00
{
2002-07-15 14:35:28 +04:00
cli - > rap_error = 0 ;
1999-12-13 16:27:58 +03:00
SSVAL ( cli - > outbuf , smb_pid , cli - > pid ) ;
SSVAL ( cli - > outbuf , smb_uid , cli - > vuid ) ;
SSVAL ( cli - > outbuf , smb_mid , cli - > mid ) ;
if ( cli - > protocol > PROTOCOL_CORE ) {
2001-02-20 11:09:06 +03:00
uint16 flags2 ;
1999-12-13 16:27:58 +03:00
SCVAL ( cli - > outbuf , smb_flg , 0x8 ) ;
2001-02-20 11:09:06 +03:00
flags2 = FLAGS2_LONG_PATH_COMPONENTS ;
2002-09-25 19:19:00 +04:00
if ( cli - > capabilities & CAP_UNICODE )
2001-02-20 11:09:06 +03:00
flags2 | = FLAGS2_UNICODE_STRINGS ;
2002-09-25 19:19:00 +04:00
if ( cli - > capabilities & CAP_STATUS32 )
2001-08-27 12:19:43 +04:00
flags2 | = FLAGS2_32_BIT_ERROR_CODES ;
2002-09-25 19:19:00 +04:00
if ( cli - > use_spnego )
2001-10-22 10:48:35 +04:00
flags2 | = FLAGS2_EXTENDED_SECURITY ;
2001-02-20 11:09:06 +03:00
SSVAL ( cli - > outbuf , smb_flg2 , flags2 ) ;
1999-12-13 16:27:58 +03:00
}
1999-06-24 22:58:08 +04:00
}
2001-02-20 15:45:50 +03:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Setup the bcc length of the packet from a pointer to the end of the data .
2001-02-20 15:45:50 +03:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
2001-02-20 15:45:50 +03:00
void cli_setup_bcc ( struct cli_state * cli , void * p )
{
set_message_bcc ( cli - > outbuf , PTR_DIFF ( p , smb_buf ( cli - > outbuf ) ) ) ;
}
1999-12-13 16:27:58 +03:00
2000-07-03 08:24:31 +04:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Initialise credentials of a client structure .
2000-07-03 08:24:31 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
2000-07-03 08:24:31 +04:00
void cli_init_creds ( struct cli_state * cli , const struct ntuser_creds * usr )
{
/* copy_nt_creds(&cli->usr, usr); */
2003-03-18 15:01:47 +03:00
fstrcpy ( cli - > domain , usr - > domain ) ;
fstrcpy ( cli - > user_name , usr - > user_name ) ;
2000-07-03 08:24:31 +04:00
memcpy ( & cli - > pwd , & usr - > pwd , sizeof ( usr - > pwd ) ) ;
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
DEBUG ( 10 , ( " cli_init_creds: user %s domain %s \n " ,
cli - > user_name , cli - > domain ) ) ;
2000-07-03 08:24:31 +04:00
}
2003-07-31 03:49:29 +04:00
/****************************************************************************
Set the signing state ( used from the command line ) .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
void cli_setup_signing_state ( struct cli_state * cli , int signing_state )
{
if ( signing_state = = Undefined )
return ;
if ( signing_state = = False ) {
cli - > sign_info . allow_smb_signing = False ;
cli - > sign_info . mandatory_signing = False ;
return ;
}
cli - > sign_info . allow_smb_signing = True ;
if ( signing_state = = Required )
cli - > sign_info . mandatory_signing = True ;
}
1998-04-24 00:12:17 +04:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Initialise a client structure .
1998-04-24 00:12:17 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
2000-04-25 18:04:06 +04:00
struct cli_state * cli_initialise ( struct cli_state * cli )
1998-04-24 00:12:17 +04:00
{
2001-08-07 03:29:25 +04:00
BOOL alloced_cli = False ;
2001-10-31 04:52:34 +03:00
/* Check the effective uid - make sure we are not setuid */
if ( is_setuid_root ( ) ) {
DEBUG ( 0 , ( " libsmb based programs must *NOT* be setuid root. \n " ) ) ;
return NULL ;
}
2000-04-25 18:04:06 +04:00
if ( ! cli ) {
cli = ( struct cli_state * ) malloc ( sizeof ( * cli ) ) ;
if ( ! cli )
return NULL ;
ZERO_STRUCTP ( cli ) ;
2001-08-07 03:29:25 +04:00
alloced_cli = True ;
Makefile.in :
- added srvsvc client files
clientgen.c :
- replaced cli_error(cli, int *cls, int *err) with
cli_error(cli, uint8 cls, uint32 *err). this version detects
32 bit status messages. the DOS error "MORE_DATA", the
equivalent of the 32 bit *warning* 0x8000 0005
(STATUS_BUFFER_OVERFLOW), was being processed as an error,
terminating the cli_receive_trans() call.
cli_pipe.c :
- replaced calls that had been incorrectly modified from
32 bit warnings (0x8000 0005 - STATUS_BUFFER_OVERFLOW)
to 8 bit DOS errors (0x01 0xEA - MORE_DATA).
the use of the old version of cli_error (DOS only)
instead of the new one (DOS and 32 bit) caused the
dce/rpc client code to fail.
- replaced 2 space indentation with tab indentation in all functions.
cli_srvsvc.c :
cmd_srvsvc.c :
- added these files back in, fixing them up to use jeremy's
modified versions of the dce/rpc client functions.
parse_srv.c :
- added back in some "unused" functions required by dce/rpc
client-side code. it would be helpful if all such "unused"
functions could be added back in.
rpcclient.c :
- added "session", "file", "share", "connection" enumeration
functions back in. these are equivalent to nt's "NetXXXXXEnum"
Win32 (MSDN) functions.
- added "srvinfo" function back in. this is equivalent to
nt's NetServerGetInfo Win32 (MSDN) function.
(This used to be commit bcf39ffdcc64e049bca2d70a394a99976291e81d)
1998-10-03 01:09:23 +04:00
}
1998-04-24 02:59:19 +04:00
2002-09-25 19:19:00 +04:00
if ( cli - > initialised )
cli_close_connection ( cli ) ;
1999-10-08 02:10:29 +04:00
2000-04-25 18:04:06 +04:00
ZERO_STRUCTP ( cli ) ;
1999-12-13 16:27:58 +03:00
2000-04-25 18:04:06 +04:00
cli - > port = 0 ;
cli - > fd = - 1 ;
cli - > cnum = - 1 ;
2000-05-02 06:23:41 +04:00
cli - > pid = ( uint16 ) sys_getpid ( ) ;
2000-04-25 18:04:06 +04:00
cli - > mid = 1 ;
cli - > vuid = UID_FIELD_INVALID ;
cli - > protocol = PROTOCOL_NT1 ;
cli - > timeout = 20000 ; /* Timeout is in milliseconds. */
cli - > bufsize = CLI_BUFFER_SIZE + 4 ;
cli - > max_xmit = cli - > bufsize ;
2003-03-18 15:02:51 +03:00
cli - > outbuf = ( char * ) malloc ( cli - > bufsize + SAFETY_MARGIN ) ;
cli - > inbuf = ( char * ) malloc ( cli - > bufsize + SAFETY_MARGIN ) ;
2001-06-18 12:26:15 +04:00
cli - > oplock_handler = cli_oplock_ack ;
2003-02-24 05:55:00 +03:00
cli - > use_spnego = lp_client_use_spnego ( ) ;
cli - > capabilities = CAP_UNICODE | CAP_STATUS32 ;
2001-06-18 12:26:15 +04:00
2001-11-25 05:35:37 +03:00
/* Set the CLI_FORCE_DOSERR environment variable to test
client routines using DOS errors instead of STATUS32
ones . This intended only as a temporary hack . */
2002-10-01 06:11:35 +04:00
if ( getenv ( " CLI_FORCE_DOSERR " ) )
2001-11-25 05:35:37 +03:00
cli - > force_dos_errors = True ;
2003-02-24 05:55:00 +03:00
if ( lp_client_signing ( ) )
cli - > sign_info . allow_smb_signing = True ;
2003-07-16 03:05:57 +04:00
if ( lp_client_signing ( ) = = Required )
cli - > sign_info . mandatory_signing = True ;
2002-09-25 19:19:00 +04:00
2000-04-25 18:04:06 +04:00
if ( ! cli - > outbuf | | ! cli - > inbuf )
2001-08-06 06:18:40 +04:00
goto error ;
1997-10-26 10:32:02 +03:00
2002-12-20 23:21:31 +03:00
if ( ( cli - > mem_ctx = talloc_init ( " cli based talloc " ) ) = = NULL )
2001-08-06 06:18:40 +04:00
goto error ;
2000-07-27 04:47:19 +04:00
2001-08-06 06:18:40 +04:00
memset ( cli - > outbuf , 0 , cli - > bufsize ) ;
memset ( cli - > inbuf , 0 , cli - > bufsize ) ;
1997-10-26 10:32:02 +03:00
2003-09-04 05:12:39 +04:00
# if defined(DEVELOPER)
2003-05-14 04:46:43 +04:00
/* just because we over-allocate, doesn't mean it's right to use it */
2003-04-21 17:00:39 +04:00
clobber_region ( FUNCTION_MACRO , __LINE__ , cli - > outbuf + cli - > bufsize , SAFETY_MARGIN ) ;
clobber_region ( FUNCTION_MACRO , __LINE__ , cli - > inbuf + cli - > bufsize , SAFETY_MARGIN ) ;
2003-09-04 05:12:39 +04:00
# endif
2003-04-21 17:00:39 +04:00
/* initialise signing */
cli_null_set_signing ( cli ) ;
2001-05-04 11:28:41 +04:00
cli - > nt_pipe_fnum = 0 ;
2003-04-09 19:54:17 +04:00
cli - > saved_netlogon_pipe_fnum = 0 ;
2001-05-04 11:28:41 +04:00
2000-04-25 18:04:06 +04:00
cli - > initialised = 1 ;
2001-11-27 06:29:20 +03:00
cli - > allocated = alloced_cli ;
1997-10-26 10:32:02 +03:00
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
cli - > pipe_idx = - 1 ;
2000-04-25 18:04:06 +04:00
return cli ;
2001-08-06 06:18:40 +04:00
/* Clean up after malloc() error */
error :
2001-09-17 07:33:37 +04:00
SAFE_FREE ( cli - > inbuf ) ;
SAFE_FREE ( cli - > outbuf ) ;
2001-08-06 06:18:40 +04:00
2001-08-07 03:29:25 +04:00
if ( alloced_cli )
2001-09-17 07:33:37 +04:00
SAFE_FREE ( cli ) ;
2001-08-07 03:29:25 +04:00
2001-08-06 06:18:40 +04:00
return NULL ;
1997-10-26 10:32:02 +03:00
}
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
/****************************************************************************
close the session
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
void cli_nt_session_close ( struct cli_state * cli )
{
if ( cli - > ntlmssp_pipe_state ) {
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
ntlmssp_end ( & cli - > ntlmssp_pipe_state ) ;
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
}
2003-08-14 00:26:24 +04:00
if ( cli - > nt_pipe_fnum ! = 0 )
cli_close ( cli , cli - > nt_pipe_fnum ) ;
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
cli - > nt_pipe_fnum = 0 ;
cli - > pipe_idx = - 1 ;
}
/****************************************************************************
close the NETLOGON session holding the session key for NETSEC
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
void cli_nt_netlogon_netsec_session_close ( struct cli_state * cli )
{
if ( cli - > saved_netlogon_pipe_fnum ! = 0 ) {
cli_close ( cli , cli - > saved_netlogon_pipe_fnum ) ;
cli - > saved_netlogon_pipe_fnum = 0 ;
}
}
1997-10-21 13:12:41 +04:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Close a client connection and free the memory without destroying cli itself .
1997-10-21 13:12:41 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
void cli_close_connection ( struct cli_state * cli )
1997-10-21 13:12:41 +04:00
{
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
cli_nt_session_close ( cli ) ;
cli_nt_netlogon_netsec_session_close ( cli ) ;
2001-09-17 07:33:37 +04:00
SAFE_FREE ( cli - > outbuf ) ;
SAFE_FREE ( cli - > inbuf ) ;
2000-07-27 04:47:19 +04:00
2003-03-18 15:01:47 +03:00
cli_free_signing_context ( cli ) ;
2001-10-17 12:54:19 +04:00
data_blob_free ( & cli - > secblob ) ;
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
data_blob_free ( & cli - > user_session_key ) ;
2001-10-11 11:42:52 +04:00
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
if ( cli - > ntlmssp_pipe_state )
Changes all over the shop, but all towards:
- NTLM2 support in the server
- KEY_EXCH support in the server
- variable length session keys.
In detail:
- NTLM2 is an extension of NTLMv1, that is compatible with existing
domain controllers (unlike NTLMv2, which requires a DC upgrade).
* This is known as 'NTLMv2 session security' *
(This is not yet implemented on the RPC pipes however, so there may
well still be issues for PDC setups, particuarly around password
changes. We do not fully understand the sign/seal implications of
NTLM2 on RPC pipes.)
This requires modifications to our authentication subsystem, as we
must handle the 'challege' input into the challenge-response algorithm
being changed. This also needs to be turned off for
'security=server', which does not support this.
- KEY_EXCH is another 'security' mechanism, whereby the session key
actually used by the server is sent by the client, rather than being
the shared-secret directly or indirectly.
- As both these methods change the session key, the auth subsystem
needed to be changed, to 'override' session keys provided by the
backend.
- There has also been a major overhaul of the NTLMSSP subsystem, to merge the 'client' and 'server' functions, so they both operate on a single structure. This should help the SPNEGO implementation.
- The 'names blob' in NTLMSSP is always in unicode - never in ascii.
Don't make an ascii version ever.
- The other big change is to allow variable length session keys. We
have always assumed that session keys are 16 bytes long - and padded
to this length if shorter. However, Kerberos session keys are 8 bytes
long, when the krb5 login uses DES.
* This fix allows SMB signging on machines not yet running MIT KRB5 1.3.1. *
- Add better DEBUG() messages to ntlm_auth, warning administrators of
misconfigurations that prevent access to the privileged pipe. This
should help reduce some of the 'it just doesn't work' issues.
- Fix data_blob_talloc() to behave the same way data_blob() does when
passed a NULL data pointer. (just allocate)
REMEMBER to make clean after this commit - I have changed plenty of data structures...
(This used to be commit f3bbc87b0dac63426cda6fac7a295d3aad810ecc)
2003-11-22 16:19:38 +03:00
ntlmssp_end ( & cli - > ntlmssp_pipe_state ) ;
Jeremy requested that I get my NTLMSSP patch into CVS. He didn't request
the schannel code, but I've included that anyway. :-)
This patch revives the client-side NTLMSSP support for RPC named pipes
in Samba, and cleans up the client and server schannel code. The use of the
new code is enabled by the 'sign', 'seal' and 'schannel' commands in
rpcclient.
The aim was to prove that our separate NTLMSSP client library actually
implements NTLMSSP signing and sealing as per Microsoft's NTLMv1 implementation,
in the hope that knowing this will assist us in correctly implementing
NTLMSSP signing for SMB packets. (Still not yet functional)
This patch replaces the NTLMSSP implementation in rpc_client/cli_pipe.c with
calls to libsmb/ntlmssp.c. In the process, we have gained the ability to
use the more secure NT password, and the ability to sign-only, instead of
having to seal the pipe connection. (Previously we were limited to sealing,
and could only use the LM-password derived key).
Our new client-side NTLMSSP code also needed alteration to cope with our
comparatively simple server-side implementation. A future step is to replace
it with calls to the same NTLMSSP library.
Also included in this patch is the schannel 'sign only' patch I submitted to
the team earlier. While not enabled (and not functional, at this stage) the
work in this patch makes the code paths *much* easier to follow. I have also
included similar hooks in rpccleint to allow the use of schannel on *any* pipe.
rpcclient now defaults to not using schannel (or any other extra per-pipe
authenticiation) for any connection. The 'schannel' command enables schannel
for all pipes until disabled.
This code is also much more secure than the previous code, as changes to our
cli_pipe routines ensure that the authentication footer cannot be removed
by an attacker, and more error states are correctly handled.
(The same needs to be done to our server)
Andrew Bartlett
(This used to be commit 5472ddc9eaf4e79c5b2e1c8ee8c7f190dc285f19)
2003-07-14 12:46:32 +04:00
2002-09-25 19:19:00 +04:00
if ( cli - > mem_ctx ) {
2000-07-27 04:47:19 +04:00
talloc_destroy ( cli - > mem_ctx ) ;
2002-09-25 19:19:00 +04:00
cli - > mem_ctx = NULL ;
}
2000-07-27 04:47:19 +04:00
2000-04-25 18:04:06 +04:00
if ( cli - > fd ! = - 1 )
2001-10-11 11:42:52 +04:00
close ( cli - > fd ) ;
2002-09-25 19:19:00 +04:00
cli - > fd = - 1 ;
2002-10-01 10:50:38 +04:00
cli - > smb_rw_error = 0 ;
2003-03-18 15:01:47 +03:00
2002-09-25 19:19:00 +04:00
}
/****************************************************************************
Shutdown a client structure .
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
void cli_shutdown ( struct cli_state * cli )
{
BOOL allocated = cli - > allocated ;
cli_close_connection ( cli ) ;
2001-11-27 06:29:20 +03:00
ZERO_STRUCTP ( cli ) ;
2002-10-01 06:11:35 +04:00
if ( allocated )
2001-11-27 06:29:20 +03:00
free ( cli ) ;
1997-10-21 13:12:41 +04:00
}
/****************************************************************************
2002-09-25 19:19:00 +04:00
Set socket options on a open connection .
1997-10-21 13:12:41 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
2003-01-03 11:28:12 +03:00
void cli_sockopt ( struct cli_state * cli , const char * options )
1997-10-21 13:12:41 +04:00
{
2000-04-25 18:04:06 +04:00
set_socket_options ( cli - > fd , options ) ;
1997-10-21 13:12:41 +04:00
}
1997-11-30 05:58:34 +03:00
/****************************************************************************
2002-09-25 19:19:00 +04:00
Set the PID to use for smb messages . Return the old pid .
1997-11-30 05:58:34 +03:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2002-09-25 19:19:00 +04:00
2000-04-25 18:04:06 +04:00
uint16 cli_setpid ( struct cli_state * cli , uint16 pid )
1997-11-30 05:58:34 +03:00
{
2000-04-25 18:04:06 +04:00
uint16 ret = cli - > pid ;
cli - > pid = pid ;
return ret ;
1997-11-30 05:58:34 +03:00
}
2002-09-25 19:19:00 +04:00
/****************************************************************************
Send a keepalive packet to the server
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
BOOL cli_send_keepalive ( struct cli_state * cli )
{
if ( cli - > fd = = - 1 ) {
DEBUG ( 3 , ( " cli_send_keepalive: fd == -1 \n " ) ) ;
return False ;
}
if ( ! send_keepalive ( cli - > fd ) ) {
close ( cli - > fd ) ;
cli - > fd = - 1 ;
DEBUG ( 0 , ( " Error sending keepalive packet to client. \n " ) ) ;
return False ;
}
return True ;
}