mirror of
https://github.com/samba-team/samba.git
synced 2024-12-28 07:21:54 +03:00
fcbfc7ad06
- 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 f3bbc87b0d
)
300 lines
7.7 KiB
C
300 lines
7.7 KiB
C
/*
|
|
Unix SMB/CIFS implementation.
|
|
simple kerberos5/SPNEGO routines
|
|
Copyright (C) Andrew Tridgell 2001
|
|
Copyright (C) Jim McDonough <jmcd@us.ibm.com> 2002
|
|
Copyright (C) Andrew Bartlett 2002-2003
|
|
|
|
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"
|
|
|
|
/*
|
|
this is a tiny msrpc packet generator. I am only using this to
|
|
avoid tying this code to a particular varient of our rpc code. This
|
|
generator is not general enough for all our rpc needs, its just
|
|
enough for the spnego/ntlmssp code
|
|
|
|
format specifiers are:
|
|
|
|
U = unicode string (input is unix string)
|
|
a = address (input is char *unix_string)
|
|
(1 byte type, 1 byte length, unicode/ASCII string, all inline)
|
|
A = ASCII string (input is unix string)
|
|
B = data blob (pointer + length)
|
|
b = data blob in header (pointer + length)
|
|
D
|
|
d = word (4 bytes)
|
|
C = constant ascii string
|
|
*/
|
|
BOOL msrpc_gen(DATA_BLOB *blob,
|
|
const char *format, ...)
|
|
{
|
|
int i, n;
|
|
va_list ap;
|
|
char *s;
|
|
uint8 *b;
|
|
int head_size=0, data_size=0;
|
|
int head_ofs, data_ofs;
|
|
|
|
/* first scan the format to work out the header and body size */
|
|
va_start(ap, format);
|
|
for (i=0; format[i]; i++) {
|
|
switch (format[i]) {
|
|
case 'U':
|
|
s = va_arg(ap, char *);
|
|
head_size += 8;
|
|
data_size += str_charnum(s) * 2;
|
|
break;
|
|
case 'A':
|
|
s = va_arg(ap, char *);
|
|
head_size += 8;
|
|
data_size += str_ascii_charnum(s);
|
|
break;
|
|
case 'a':
|
|
n = va_arg(ap, int);
|
|
s = va_arg(ap, char *);
|
|
data_size += (str_charnum(s) * 2) + 4;
|
|
break;
|
|
case 'B':
|
|
b = va_arg(ap, uint8 *);
|
|
head_size += 8;
|
|
data_size += va_arg(ap, int);
|
|
break;
|
|
case 'b':
|
|
b = va_arg(ap, uint8 *);
|
|
head_size += va_arg(ap, int);
|
|
break;
|
|
case 'd':
|
|
n = va_arg(ap, int);
|
|
head_size += 4;
|
|
break;
|
|
case 'C':
|
|
s = va_arg(ap, char *);
|
|
head_size += str_charnum(s) + 1;
|
|
break;
|
|
}
|
|
}
|
|
va_end(ap);
|
|
|
|
/* allocate the space, then scan the format again to fill in the values */
|
|
*blob = data_blob(NULL, head_size + data_size);
|
|
|
|
head_ofs = 0;
|
|
data_ofs = head_size;
|
|
|
|
va_start(ap, format);
|
|
for (i=0; format[i]; i++) {
|
|
switch (format[i]) {
|
|
case 'U':
|
|
s = va_arg(ap, char *);
|
|
n = str_charnum(s);
|
|
SSVAL(blob->data, head_ofs, n*2); head_ofs += 2;
|
|
SSVAL(blob->data, head_ofs, n*2); head_ofs += 2;
|
|
SIVAL(blob->data, head_ofs, data_ofs); head_ofs += 4;
|
|
push_string(NULL, blob->data+data_ofs, s, n*2, STR_UNICODE|STR_NOALIGN);
|
|
data_ofs += n*2;
|
|
break;
|
|
case 'A':
|
|
s = va_arg(ap, char *);
|
|
n = str_ascii_charnum(s);
|
|
SSVAL(blob->data, head_ofs, n); head_ofs += 2;
|
|
SSVAL(blob->data, head_ofs, n); head_ofs += 2;
|
|
SIVAL(blob->data, head_ofs, data_ofs); head_ofs += 4;
|
|
push_string(NULL, blob->data+data_ofs, s, n, STR_ASCII|STR_NOALIGN);
|
|
data_ofs += n;
|
|
break;
|
|
case 'a':
|
|
n = va_arg(ap, int);
|
|
SSVAL(blob->data, data_ofs, n); data_ofs += 2;
|
|
s = va_arg(ap, char *);
|
|
n = str_charnum(s);
|
|
SSVAL(blob->data, data_ofs, n*2); data_ofs += 2;
|
|
if (0 < n) {
|
|
push_string(NULL, blob->data+data_ofs, s, n*2,
|
|
STR_UNICODE|STR_NOALIGN);
|
|
}
|
|
data_ofs += n*2;
|
|
break;
|
|
|
|
case 'B':
|
|
b = va_arg(ap, uint8 *);
|
|
n = va_arg(ap, int);
|
|
SSVAL(blob->data, head_ofs, n); head_ofs += 2;
|
|
SSVAL(blob->data, head_ofs, n); head_ofs += 2;
|
|
SIVAL(blob->data, head_ofs, data_ofs); head_ofs += 4;
|
|
if (n && b) /* don't follow null pointers... */
|
|
memcpy(blob->data+data_ofs, b, n);
|
|
data_ofs += n;
|
|
break;
|
|
case 'd':
|
|
n = va_arg(ap, int);
|
|
SIVAL(blob->data, head_ofs, n); head_ofs += 4;
|
|
break;
|
|
case 'b':
|
|
b = va_arg(ap, uint8 *);
|
|
n = va_arg(ap, int);
|
|
memcpy(blob->data + head_ofs, b, n);
|
|
head_ofs += n;
|
|
break;
|
|
case 'C':
|
|
s = va_arg(ap, char *);
|
|
head_ofs += push_string(NULL, blob->data+head_ofs, s, -1,
|
|
STR_ASCII|STR_TERMINATE);
|
|
break;
|
|
}
|
|
}
|
|
va_end(ap);
|
|
|
|
return True;
|
|
}
|
|
|
|
|
|
/* a helpful macro to avoid running over the end of our blob */
|
|
#define NEED_DATA(amount) \
|
|
if ((head_ofs + amount) > blob->length) { \
|
|
return False; \
|
|
}
|
|
|
|
/*
|
|
this is a tiny msrpc packet parser. This the the partner of msrpc_gen
|
|
|
|
format specifiers are:
|
|
|
|
U = unicode string (output is unix string)
|
|
A = ascii string
|
|
B = data blob
|
|
b = data blob in header
|
|
d = word (4 bytes)
|
|
C = constant ascii string
|
|
*/
|
|
|
|
BOOL msrpc_parse(const DATA_BLOB *blob,
|
|
const char *format, ...)
|
|
{
|
|
int i;
|
|
va_list ap;
|
|
char **ps, *s;
|
|
DATA_BLOB *b;
|
|
size_t head_ofs = 0;
|
|
uint16 len1, len2;
|
|
uint32 ptr;
|
|
uint32 *v;
|
|
pstring p;
|
|
|
|
va_start(ap, format);
|
|
for (i=0; format[i]; i++) {
|
|
switch (format[i]) {
|
|
case 'U':
|
|
NEED_DATA(8);
|
|
len1 = SVAL(blob->data, head_ofs); head_ofs += 2;
|
|
len2 = SVAL(blob->data, head_ofs); head_ofs += 2;
|
|
ptr = IVAL(blob->data, head_ofs); head_ofs += 4;
|
|
|
|
ps = va_arg(ap, char **);
|
|
if (len1 == 0 && len2 == 0) {
|
|
*ps = smb_xstrdup("");
|
|
} else {
|
|
/* make sure its in the right format - be strict */
|
|
if ((len1 != len2) || (ptr + len1 < ptr) || (ptr + len1 < len1) || (ptr + len1 > blob->length)) {
|
|
return False;
|
|
}
|
|
if (len1 & 1) {
|
|
/* if odd length and unicode */
|
|
return False;
|
|
}
|
|
|
|
if (0 < len1) {
|
|
pull_string(NULL, p, blob->data + ptr, sizeof(p),
|
|
len1,
|
|
STR_UNICODE|STR_NOALIGN);
|
|
(*ps) = smb_xstrdup(p);
|
|
} else {
|
|
(*ps) = smb_xstrdup("");
|
|
}
|
|
}
|
|
break;
|
|
case 'A':
|
|
NEED_DATA(8);
|
|
len1 = SVAL(blob->data, head_ofs); head_ofs += 2;
|
|
len2 = SVAL(blob->data, head_ofs); head_ofs += 2;
|
|
ptr = IVAL(blob->data, head_ofs); head_ofs += 4;
|
|
|
|
ps = va_arg(ap, char **);
|
|
/* make sure its in the right format - be strict */
|
|
if (len1 == 0 && len2 == 0) {
|
|
*ps = smb_xstrdup("");
|
|
} else {
|
|
if ((len1 != len2) || (ptr + len1 < ptr) || (ptr + len1 < len1) || (ptr + len1 > blob->length)) {
|
|
return False;
|
|
}
|
|
|
|
if (0 < len1) {
|
|
pull_string(NULL, p, blob->data + ptr, sizeof(p),
|
|
len1,
|
|
STR_ASCII|STR_NOALIGN);
|
|
(*ps) = smb_xstrdup(p);
|
|
} else {
|
|
(*ps) = smb_xstrdup("");
|
|
}
|
|
}
|
|
break;
|
|
case 'B':
|
|
NEED_DATA(8);
|
|
len1 = SVAL(blob->data, head_ofs); head_ofs += 2;
|
|
len2 = SVAL(blob->data, head_ofs); head_ofs += 2;
|
|
ptr = IVAL(blob->data, head_ofs); head_ofs += 4;
|
|
|
|
b = (DATA_BLOB *)va_arg(ap, void *);
|
|
if (len1 == 0 && len2 == 0) {
|
|
*b = data_blob(NULL, 0);
|
|
} else {
|
|
/* make sure its in the right format - be strict */
|
|
if ((len1 != len2) || (ptr + len1 < ptr) || (ptr + len1 < len1) || (ptr + len1 > blob->length)) {
|
|
return False;
|
|
}
|
|
*b = data_blob(blob->data + ptr, len1);
|
|
}
|
|
break;
|
|
case 'b':
|
|
b = (DATA_BLOB *)va_arg(ap, void *);
|
|
len1 = va_arg(ap, unsigned);
|
|
/* make sure its in the right format - be strict */
|
|
NEED_DATA(len1);
|
|
*b = data_blob(blob->data + head_ofs, len1);
|
|
head_ofs += len1;
|
|
break;
|
|
case 'd':
|
|
v = va_arg(ap, uint32 *);
|
|
NEED_DATA(4);
|
|
*v = IVAL(blob->data, head_ofs); head_ofs += 4;
|
|
break;
|
|
case 'C':
|
|
s = va_arg(ap, char *);
|
|
head_ofs += pull_string(NULL, p, blob->data+head_ofs, sizeof(p),
|
|
blob->length - head_ofs,
|
|
STR_ASCII|STR_TERMINATE);
|
|
if (strcmp(s, p) != 0) {
|
|
return False;
|
|
}
|
|
break;
|
|
}
|
|
}
|
|
va_end(ap);
|
|
|
|
return True;
|
|
}
|