1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-27 03:21:53 +03:00
samba-mirror/source3/libsmb/ntlmssp.c

796 lines
23 KiB
C
Raw Normal View History

/*
Unix SMB/Netbios implementation.
Version 3.0
handle NLTMSSP, server side
Copyright (C) Andrew Tridgell 2001
Copyright (C) Andrew Bartlett 2001-2010
Copyright (C) Stefan Metzmacher 2005
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3 of the License, or
(at your option) any later version.
2008-11-30 14:43:12 +03:00
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
2008-11-30 14:43:12 +03:00
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "includes.h"
#include "../auth/ntlmssp/ntlmssp.h"
#include "../auth/ntlmssp/ntlmssp_private.h"
#include "../libcli/auth/libcli_auth.h"
#include "../librpc/gen_ndr/ndr_ntlmssp.h"
#include "../auth/ntlmssp/ntlmssp_ndr.h"
#include "../lib/crypto/md5.h"
#include "../lib/crypto/arcfour.h"
#include "../lib/crypto/hmacmd5.h"
#include "../nsswitch/libwbclient/wbclient.h"
static NTSTATUS ntlmssp_client_initial(struct ntlmssp_state *ntlmssp_state,
TALLOC_CTX *out_mem_ctx, /* Unused at this time */
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 reply, DATA_BLOB *next_request);
static NTSTATUS ntlmssp_client_challenge(struct ntlmssp_state *ntlmssp_state,
TALLOC_CTX *out_mem_ctx, /* Unused at this time */
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
const DATA_BLOB reply, DATA_BLOB *next_request);
/**
* Callbacks for NTLMSSP - for both client and server operating modes
*
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
*/
static const struct ntlmssp_callbacks {
enum ntlmssp_role role;
enum ntlmssp_message_type ntlmssp_command;
NTSTATUS (*fn)(struct ntlmssp_state *ntlmssp_state,
TALLOC_CTX *out_mem_ctx,
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 in, DATA_BLOB *out);
} ntlmssp_callbacks[] = {
{NTLMSSP_CLIENT, NTLMSSP_INITIAL, ntlmssp_client_initial},
{NTLMSSP_SERVER, NTLMSSP_NEGOTIATE, ntlmssp_server_negotiate},
{NTLMSSP_CLIENT, NTLMSSP_CHALLENGE, ntlmssp_client_challenge},
{NTLMSSP_SERVER, NTLMSSP_AUTH, ntlmssp_server_auth},
{NTLMSSP_CLIENT, NTLMSSP_UNKNOWN, NULL},
{NTLMSSP_SERVER, NTLMSSP_UNKNOWN, NULL}
};
/**
* Default challenge generation code.
*
*/
2008-11-30 14:43:12 +03:00
static NTSTATUS get_challenge(const struct ntlmssp_state *ntlmssp_state,
uint8_t chal[8])
{
generate_random_buffer(chal, 8);
return NT_STATUS_OK;
}
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
/**
* Default 'we can set the challenge to anything we like' implementation
*
*/
2008-11-30 14:43:12 +03:00
static bool may_set_challenge(const struct ntlmssp_state *ntlmssp_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
{
return True;
}
/**
* Default 'we can set the challenge to anything we like' implementation
*
* Does not actually do anything, as the value is always in the structure anyway.
*
*/
2008-11-30 14:43:12 +03:00
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
static NTSTATUS set_challenge(struct ntlmssp_state *ntlmssp_state, DATA_BLOB *challenge)
{
SMB_ASSERT(challenge->length == 8);
return NT_STATUS_OK;
}
/**
* Set a username on an NTLMSSP context - ensures it is talloc()ed
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
*
*/
NTSTATUS ntlmssp_set_username(struct ntlmssp_state *ntlmssp_state, const char *user)
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_state->user = talloc_strdup(ntlmssp_state, user ? user : "" );
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
if (!ntlmssp_state->user) {
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}
/**
* Store NT and LM hashes on an NTLMSSP context - ensures they are talloc()ed
*
*/
NTSTATUS ntlmssp_set_hashes(struct ntlmssp_state *ntlmssp_state,
const uint8_t lm_hash[16],
const uint8_t nt_hash[16])
{
ntlmssp_state->lm_hash = (uint8_t *)
talloc_memdup(ntlmssp_state, lm_hash, 16);
ntlmssp_state->nt_hash = (uint8_t *)
talloc_memdup(ntlmssp_state, nt_hash, 16);
if (!ntlmssp_state->lm_hash || !ntlmssp_state->nt_hash) {
TALLOC_FREE(ntlmssp_state->lm_hash);
TALLOC_FREE(ntlmssp_state->nt_hash);
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}
/**
* Converts a password to the hashes on an NTLMSSP context.
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
*
*/
NTSTATUS ntlmssp_set_password(struct ntlmssp_state *ntlmssp_state, const char *password)
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
{
if (!password) {
ntlmssp_state->lm_hash = NULL;
ntlmssp_state->nt_hash = NULL;
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
} else {
uint8_t lm_hash[16];
uint8_t nt_hash[16];
E_deshash(password, lm_hash);
E_md4hash(password, nt_hash);
return ntlmssp_set_hashes(ntlmssp_state, lm_hash, nt_hash);
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
}
return NT_STATUS_OK;
}
/**
* Set a domain on an NTLMSSP context - ensures it is talloc()ed
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
*
*/
NTSTATUS ntlmssp_set_domain(struct ntlmssp_state *ntlmssp_state, const char *domain)
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_state->domain = talloc_strdup(ntlmssp_state,
domain ? domain : "" );
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
if (!ntlmssp_state->domain) {
return NT_STATUS_NO_MEMORY;
}
return NT_STATUS_OK;
}
/**
* Request features for the NTLMSSP negotiation
*
* @param ntlmssp_state NTLMSSP state
* @param feature_list List of space seperated features requested from NTLMSSP.
*/
void ntlmssp_want_feature_list(struct ntlmssp_state *ntlmssp_state, char *feature_list)
{
/*
* We need to set this to allow a later SetPassword
* via the SAMR pipe to succeed. Strange.... We could
* also add NTLMSSP_NEGOTIATE_SEAL here. JRA.
*/
if (in_list("NTLMSSP_FEATURE_SESSION_KEY", feature_list, True)) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SIGN;
}
if (in_list("NTLMSSP_FEATURE_SIGN", feature_list, True)) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SIGN;
}
if(in_list("NTLMSSP_FEATURE_SEAL", feature_list, True)) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SEAL;
}
if (in_list("NTLMSSP_FEATURE_CCACHE", feature_list, true)) {
ntlmssp_state->use_ccache = true;
}
}
/**
* Request a feature for the NTLMSSP negotiation
*
* @param ntlmssp_state NTLMSSP state
* @param feature Bit flag specifying the requested feature
*/
void ntlmssp_want_feature(struct ntlmssp_state *ntlmssp_state, uint32_t feature)
{
/* As per JRA's comment above */
if (feature & NTLMSSP_FEATURE_SESSION_KEY) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SIGN;
}
if (feature & NTLMSSP_FEATURE_SIGN) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SIGN;
}
if (feature & NTLMSSP_FEATURE_SEAL) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SIGN;
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_SEAL;
}
if (feature & NTLMSSP_FEATURE_CCACHE) {
ntlmssp_state->use_ccache = true;
}
}
2008-11-30 14:43:12 +03:00
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
/**
* Next state function for the NTLMSSP state machine
*
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
* @param ntlmssp_state NTLMSSP State
* @param in The packet in from the NTLMSSP partner, as a DATA_BLOB
* @param out The reply, as an allocated DATA_BLOB, caller to free.
* @return Errors, NT_STATUS_MORE_PROCESSING_REQUIRED or NT_STATUS_OK.
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
*/
NTSTATUS ntlmssp_update(struct ntlmssp_state *ntlmssp_state,
const DATA_BLOB input, DATA_BLOB *out)
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
{
uint32_t ntlmssp_command;
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
int i;
if (ntlmssp_state->expected_state == NTLMSSP_DONE) {
/* Called update after negotiations finished. */
DEBUG(1, ("Called NTLMSSP after state machine was 'done'\n"));
return NT_STATUS_INVALID_PARAMETER;
}
*out = data_blob_null;
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
if (!input.length) {
switch (ntlmssp_state->role) {
case NTLMSSP_CLIENT:
ntlmssp_command = NTLMSSP_INITIAL;
break;
case NTLMSSP_SERVER:
/* 'datagram' mode - no neg packet */
ntlmssp_command = NTLMSSP_NEGOTIATE;
break;
2010-08-14 12:06:55 +04:00
default:
DEBUG(1, ("Invalid role: %d\n", ntlmssp_state->role));
return NT_STATUS_INVALID_PARAMETER;
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
}
} else {
if (!msrpc_parse(ntlmssp_state, &input, "Cd",
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",
&ntlmssp_command)) {
DEBUG(1, ("Failed to parse NTLMSSP packet, could not extract NTLMSSP command\n"));
dump_data(2, input.data, input.length);
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
return NT_STATUS_INVALID_PARAMETER;
}
}
if (ntlmssp_command != ntlmssp_state->expected_state) {
DEBUG(1, ("got NTLMSSP command %u, expected %u\n", ntlmssp_command, ntlmssp_state->expected_state));
return NT_STATUS_INVALID_PARAMETER;
}
for (i=0; ntlmssp_callbacks[i].fn; i++) {
if (ntlmssp_callbacks[i].role == ntlmssp_state->role
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_callbacks[i].ntlmssp_command == ntlmssp_command) {
return ntlmssp_callbacks[i].fn(ntlmssp_state, ntlmssp_state, input, out);
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
}
}
DEBUG(1, ("failed to find NTLMSSP callback for NTLMSSP mode %u, command %u\n",
ntlmssp_state->role, ntlmssp_command));
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
return NT_STATUS_INVALID_PARAMETER;
}
/**
* Create an NTLMSSP state machine
*
* @param ntlmssp_state NTLMSSP State, allocated by this function
*/
NTSTATUS ntlmssp_server_start(TALLOC_CTX *mem_ctx,
bool is_standalone,
const char *netbios_name,
const char *netbios_domain,
const char *dns_name,
const char *dns_domain,
struct ntlmssp_state **_ntlmssp_state)
{
struct ntlmssp_state *ntlmssp_state;
if (!netbios_name) {
netbios_name = "";
}
if (!netbios_domain) {
netbios_domain = "";
}
if (!dns_domain) {
dns_domain = "";
}
if (!dns_name) {
dns_name = "";
}
ntlmssp_state = talloc_zero(mem_ctx, struct ntlmssp_state);
if (!ntlmssp_state) {
return NT_STATUS_NO_MEMORY;
}
ntlmssp_state->role = NTLMSSP_SERVER;
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_state->get_challenge = get_challenge;
ntlmssp_state->set_challenge = set_challenge;
ntlmssp_state->may_set_challenge = may_set_challenge;
ntlmssp_state->server.is_standalone = is_standalone;
ntlmssp_state->expected_state = NTLMSSP_NEGOTIATE;
ntlmssp_state->allow_lm_key = lp_lanman_auth();
ntlmssp_state->neg_flags =
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_NEGOTIATE_128 |
NTLMSSP_NEGOTIATE_56 |
NTLMSSP_NEGOTIATE_VERSION |
NTLMSSP_NEGOTIATE_ALWAYS_SIGN |
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_NEGOTIATE_NTLM |
NTLMSSP_NEGOTIATE_NTLM2 |
NTLMSSP_NEGOTIATE_KEY_EXCH;
ntlmssp_state->server.netbios_name = talloc_strdup(ntlmssp_state, netbios_name);
if (!ntlmssp_state->server.netbios_name) {
talloc_free(ntlmssp_state);
return NT_STATUS_NO_MEMORY;
}
ntlmssp_state->server.netbios_domain = talloc_strdup(ntlmssp_state, netbios_domain);
if (!ntlmssp_state->server.netbios_domain) {
talloc_free(ntlmssp_state);
return NT_STATUS_NO_MEMORY;
}
ntlmssp_state->server.dns_name = talloc_strdup(ntlmssp_state, dns_name);
if (!ntlmssp_state->server.dns_name) {
talloc_free(ntlmssp_state);
return NT_STATUS_NO_MEMORY;
}
ntlmssp_state->server.dns_domain = talloc_strdup(ntlmssp_state, dns_domain);
if (!ntlmssp_state->server.dns_domain) {
talloc_free(ntlmssp_state);
return NT_STATUS_NO_MEMORY;
}
*_ntlmssp_state = ntlmssp_state;
return NT_STATUS_OK;
}
/*********************************************************************
Client side NTLMSSP
*********************************************************************/
/**
* Next state function for the Initial packet
*
* @param ntlmssp_state NTLMSSP State
* @param request The request, as a DATA_BLOB. reply.data must be NULL
* @param request The reply, as an allocated DATA_BLOB, caller to free.
* @return Errors or NT_STATUS_OK.
*/
static NTSTATUS ntlmssp_client_initial(struct ntlmssp_state *ntlmssp_state,
TALLOC_CTX *out_mem_ctx, /* Unused at this time */
DATA_BLOB reply, DATA_BLOB *next_request)
{
NTSTATUS status;
if (ntlmssp_state->unicode) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_UNICODE;
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
} else {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_OEM;
}
2008-11-30 14:43:12 +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
if (ntlmssp_state->use_ntlmv2) {
ntlmssp_state->neg_flags |= NTLMSSP_NEGOTIATE_NTLM2;
}
/* generate the ntlmssp negotiate packet */
status = msrpc_gen(ntlmssp_state, next_request, "CddAA",
"NTLMSSP",
NTLMSSP_NEGOTIATE,
ntlmssp_state->neg_flags,
ntlmssp_state->client.netbios_domain,
ntlmssp_state->client.netbios_name);
if (!NT_STATUS_IS_OK(status)) {
DEBUG(0, ("ntlmssp_client_initial: failed to generate "
"ntlmssp negotiate packet\n"));
return status;
}
if (DEBUGLEVEL >= 10) {
struct NEGOTIATE_MESSAGE *negotiate = talloc(
talloc_tos(), struct NEGOTIATE_MESSAGE);
if (negotiate != NULL) {
status = ntlmssp_pull_NEGOTIATE_MESSAGE(
next_request, negotiate, negotiate);
if (NT_STATUS_IS_OK(status)) {
NDR_PRINT_DEBUG(NEGOTIATE_MESSAGE,
negotiate);
}
TALLOC_FREE(negotiate);
}
}
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_state->expected_state = NTLMSSP_CHALLENGE;
return NT_STATUS_MORE_PROCESSING_REQUIRED;
}
/**
* Next state function for the Challenge Packet. Generate an auth packet.
*
* @param ntlmssp_state NTLMSSP State
* @param request The request, as a DATA_BLOB. reply.data must be NULL
* @param request The reply, as an allocated DATA_BLOB, caller to free.
* @return Errors or NT_STATUS_OK.
*/
static NTSTATUS ntlmssp_client_challenge(struct ntlmssp_state *ntlmssp_state,
TALLOC_CTX *out_mem_ctx, /* Unused at this time */
const DATA_BLOB reply, DATA_BLOB *next_request)
{
uint32_t chal_flags, ntlmssp_command, unkn1, unkn2;
DATA_BLOB server_domain_blob;
DATA_BLOB challenge_blob;
DATA_BLOB struct_blob = data_blob_null;
char *server_domain;
const char *chal_parse_string;
const char *auth_gen_string;
DATA_BLOB lm_response = data_blob_null;
DATA_BLOB nt_response = data_blob_null;
DATA_BLOB session_key = data_blob_null;
DATA_BLOB encrypted_session_key = data_blob_null;
NTSTATUS nt_status = NT_STATUS_OK;
if (ntlmssp_state->use_ccache) {
struct wbcCredentialCacheParams params;
struct wbcCredentialCacheInfo *info = NULL;
struct wbcAuthErrorInfo *error = NULL;
struct wbcNamedBlob auth_blob;
struct wbcBlob *wbc_next = NULL;
struct wbcBlob *wbc_session_key = NULL;
wbcErr wbc_status;
int i;
params.account_name = ntlmssp_state->user;
params.domain_name = ntlmssp_state->domain;
params.level = WBC_CREDENTIAL_CACHE_LEVEL_NTLMSSP;
auth_blob.name = "challenge_blob";
auth_blob.flags = 0;
auth_blob.blob.data = reply.data;
auth_blob.blob.length = reply.length;
params.num_blobs = 1;
params.blobs = &auth_blob;
wbc_status = wbcCredentialCache(&params, &info, &error);
wbcFreeMemory(error);
if (!WBC_ERROR_IS_OK(wbc_status)) {
goto noccache;
}
for (i=0; i<info->num_blobs; i++) {
if (strequal(info->blobs[i].name, "auth_blob")) {
wbc_next = &info->blobs[i].blob;
}
if (strequal(info->blobs[i].name, "session_key")) {
wbc_session_key = &info->blobs[i].blob;
}
}
if ((wbc_next == NULL) || (wbc_session_key == NULL)) {
wbcFreeMemory(info);
goto noccache;
}
*next_request = data_blob(wbc_next->data, wbc_next->length);
ntlmssp_state->session_key = data_blob(
wbc_session_key->data, wbc_session_key->length);
wbcFreeMemory(info);
goto done;
}
noccache:
if (!msrpc_parse(ntlmssp_state, &reply, "CdBd",
"NTLMSSP",
&ntlmssp_command,
&server_domain_blob,
&chal_flags)) {
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(1, ("Failed to parse the NTLMSSP Challenge: (#1)\n"));
dump_data(2, reply.data, reply.length);
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
return NT_STATUS_INVALID_PARAMETER;
}
2008-11-30 14:43:12 +03:00
if (DEBUGLEVEL >= 10) {
struct CHALLENGE_MESSAGE *challenge = talloc(
talloc_tos(), struct CHALLENGE_MESSAGE);
if (challenge != NULL) {
NTSTATUS status;
challenge->NegotiateFlags = chal_flags;
status = ntlmssp_pull_CHALLENGE_MESSAGE(
&reply, challenge, challenge);
if (NT_STATUS_IS_OK(status)) {
NDR_PRINT_DEBUG(CHALLENGE_MESSAGE,
challenge);
}
TALLOC_FREE(challenge);
}
}
data_blob_free(&server_domain_blob);
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(3, ("Got challenge flags:\n"));
debug_ntlmssp_flags(chal_flags);
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_handle_neg_flags(ntlmssp_state, chal_flags, lp_client_lanman_auth());
if (ntlmssp_state->unicode) {
if (chal_flags & NTLMSSP_NEGOTIATE_TARGET_INFO) {
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
chal_parse_string = "CdUdbddB";
} else {
chal_parse_string = "CdUdbdd";
}
auth_gen_string = "CdBBUUUBd";
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
} else {
if (chal_flags & NTLMSSP_NEGOTIATE_TARGET_INFO) {
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
chal_parse_string = "CdAdbddB";
} else {
chal_parse_string = "CdAdbdd";
}
auth_gen_string = "CdBBAAABd";
}
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(3, ("NTLMSSP: Set final flags:\n"));
debug_ntlmssp_flags(ntlmssp_state->neg_flags);
if (!msrpc_parse(ntlmssp_state, &reply, chal_parse_string,
"NTLMSSP",
&ntlmssp_command,
&server_domain,
&chal_flags,
&challenge_blob, 8,
&unkn1, &unkn2,
&struct_blob)) {
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(1, ("Failed to parse the NTLMSSP Challenge: (#2)\n"));
dump_data(2, reply.data, reply.length);
return NT_STATUS_INVALID_PARAMETER;
}
if (chal_flags & NTLMSSP_TARGET_TYPE_SERVER) {
ntlmssp_state->server.is_standalone = true;
} else {
ntlmssp_state->server.is_standalone = false;
}
/* TODO: parse struct_blob and fill in the rest */
ntlmssp_state->server.netbios_name = "";
ntlmssp_state->server.netbios_domain = server_domain;
ntlmssp_state->server.dns_name = "";
ntlmssp_state->server.dns_domain = "";
if (challenge_blob.length != 8) {
data_blob_free(&struct_blob);
return NT_STATUS_INVALID_PARAMETER;
}
if (!ntlmssp_state->nt_hash || !ntlmssp_state->lm_hash) {
static const uint8_t zeros[16] = {0, };
/* do nothing - blobs are zero length */
/* session key is all zeros */
session_key = data_blob_talloc(ntlmssp_state, zeros, 16);
2008-11-30 14:43:12 +03:00
/* not doing NLTM2 without a password */
ntlmssp_state->neg_flags &= ~NTLMSSP_NEGOTIATE_NTLM2;
} else if (ntlmssp_state->use_ntlmv2) {
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 (!struct_blob.length) {
/* be lazy, match win2k - we can't do NTLMv2 without it */
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
DEBUG(1, ("Server did not provide 'target information', required for NTLMv2\n"));
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
return NT_STATUS_INVALID_PARAMETER;
}
/* TODO: if the remote server is standalone, then we should replace 'domain'
with the server name as supplied above */
2008-11-30 14:43:12 +03:00
if (!SMBNTLMv2encrypt_hash(ntlmssp_state,
ntlmssp_state->user,
ntlmssp_state->domain,
ntlmssp_state->nt_hash, &challenge_blob,
&struct_blob,
&lm_response, &nt_response, NULL,
&session_key)) {
data_blob_free(&challenge_blob);
data_blob_free(&struct_blob);
return NT_STATUS_NO_MEMORY;
}
} else if (ntlmssp_state->neg_flags & NTLMSSP_NEGOTIATE_NTLM2) {
struct MD5Context md5_session_nonce_ctx;
uint8_t session_nonce[16];
uint8_t session_nonce_hash[16];
uint8_t user_session_key[16];
2008-11-30 14:43:12 +03:00
lm_response = data_blob_talloc(ntlmssp_state, NULL, 24);
generate_random_buffer(lm_response.data, 8);
memset(lm_response.data+8, 0, 16);
memcpy(session_nonce, challenge_blob.data, 8);
memcpy(&session_nonce[8], lm_response.data, 8);
2008-11-30 14:43:12 +03:00
MD5Init(&md5_session_nonce_ctx);
MD5Update(&md5_session_nonce_ctx, challenge_blob.data, 8);
MD5Update(&md5_session_nonce_ctx, lm_response.data, 8);
MD5Final(session_nonce_hash, &md5_session_nonce_ctx);
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
DEBUG(5, ("NTLMSSP challenge set by NTLM2\n"));
DEBUG(5, ("challenge is: \n"));
dump_data(5, session_nonce_hash, 8);
2008-11-30 14:43:12 +03:00
nt_response = data_blob_talloc(ntlmssp_state, NULL, 24);
SMBNTencrypt_hash(ntlmssp_state->nt_hash,
session_nonce_hash,
nt_response.data);
session_key = data_blob_talloc(ntlmssp_state, NULL, 16);
SMBsesskeygen_ntv1(ntlmssp_state->nt_hash, user_session_key);
hmac_md5(user_session_key, session_nonce, sizeof(session_nonce), session_key.data);
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
dump_data_pw("NTLM2 session key:\n", session_key.data, session_key.length);
} else {
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
/* lanman auth is insecure, it may be disabled */
if (lp_client_lanman_auth()) {
lm_response = data_blob_talloc(ntlmssp_state,
NULL, 24);
SMBencrypt_hash(ntlmssp_state->lm_hash,challenge_blob.data,
lm_response.data);
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
}
2008-11-30 14:43:12 +03:00
nt_response = data_blob_talloc(ntlmssp_state, NULL, 24);
SMBNTencrypt_hash(ntlmssp_state->nt_hash,challenge_blob.data,
nt_response.data);
2008-11-30 14:43:12 +03:00
session_key = data_blob_talloc(ntlmssp_state, NULL, 16);
if ((ntlmssp_state->neg_flags & NTLMSSP_NEGOTIATE_LM_KEY)
&& lp_client_lanman_auth()) {
SMBsesskeygen_lm_sess_key(ntlmssp_state->lm_hash, lm_response.data,
session_key.data);
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
dump_data_pw("LM session key\n", session_key.data, session_key.length);
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
} else {
SMBsesskeygen_ntv1(ntlmssp_state->nt_hash, session_key.data);
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
dump_data_pw("NT session key:\n", session_key.data, session_key.length);
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
}
}
data_blob_free(&struct_blob);
/* Key exchange encryptes a new client-generated session key with
the password-derived key */
if (ntlmssp_state->neg_flags & NTLMSSP_NEGOTIATE_KEY_EXCH) {
/* Make up a new session key */
uint8_t client_session_key[16];
generate_random_buffer(client_session_key, sizeof(client_session_key));
/* Encrypt the new session key with the old one */
encrypted_session_key = data_blob(client_session_key, sizeof(client_session_key));
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
dump_data_pw("KEY_EXCH session key:\n", encrypted_session_key.data, encrypted_session_key.length);
arcfour_crypt_blob(encrypted_session_key.data, encrypted_session_key.length, &session_key);
dump_data_pw("KEY_EXCH session key (enc):\n", encrypted_session_key.data, encrypted_session_key.length);
/* Mark the new session key as the 'real' session key */
data_blob_free(&session_key);
session_key = data_blob_talloc(ntlmssp_state,
client_session_key,
sizeof(client_session_key));
}
/* this generates the actual auth packet */
nt_status = msrpc_gen(ntlmssp_state, next_request, auth_gen_string,
"NTLMSSP",
NTLMSSP_AUTH,
lm_response.data, lm_response.length,
nt_response.data, nt_response.length,
ntlmssp_state->domain,
ntlmssp_state->user,
ntlmssp_state->client.netbios_name,
encrypted_session_key.data, encrypted_session_key.length,
ntlmssp_state->neg_flags);
2008-11-30 14:43:12 +03:00
if (!NT_STATUS_IS_OK(nt_status)) {
return NT_STATUS_NO_MEMORY;
}
if (DEBUGLEVEL >= 10) {
struct AUTHENTICATE_MESSAGE *authenticate = talloc(
talloc_tos(), struct AUTHENTICATE_MESSAGE);
if (authenticate != NULL) {
NTSTATUS status;
authenticate->NegotiateFlags =
ntlmssp_state->neg_flags;
status = ntlmssp_pull_AUTHENTICATE_MESSAGE(
next_request, authenticate, authenticate);
if (NT_STATUS_IS_OK(status)) {
NDR_PRINT_DEBUG(AUTHENTICATE_MESSAGE,
authenticate);
}
TALLOC_FREE(authenticate);
}
}
data_blob_free(&encrypted_session_key);
data_blob_free(&ntlmssp_state->chal);
ntlmssp_state->session_key = session_key;
ntlmssp_state->chal = challenge_blob;
ntlmssp_state->lm_resp = lm_response;
ntlmssp_state->nt_resp = nt_response;
done:
ntlmssp_state->expected_state = NTLMSSP_DONE;
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
if (!NT_STATUS_IS_OK(nt_status = ntlmssp_sign_init(ntlmssp_state))) {
DEBUG(1, ("Could not setup NTLMSSP signing/sealing system (error was: %s)\n", nt_errstr(nt_status)));
}
return nt_status;
}
NTSTATUS ntlmssp_client_start(TALLOC_CTX *mem_ctx,
const char *netbios_name,
const char *netbios_domain,
bool use_ntlmv2,
struct ntlmssp_state **_ntlmssp_state)
{
struct ntlmssp_state *ntlmssp_state;
if (!netbios_name) {
netbios_name = "";
}
if (!netbios_domain) {
netbios_domain = "";
}
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_state = talloc_zero(mem_ctx, struct ntlmssp_state);
if (!ntlmssp_state) {
return NT_STATUS_NO_MEMORY;
}
ntlmssp_state->role = NTLMSSP_CLIENT;
ntlmssp_state->unicode = True;
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
ntlmssp_state->use_ntlmv2 = use_ntlmv2;
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_state->expected_state = NTLMSSP_INITIAL;
ntlmssp_state->neg_flags =
NTLMSSP_NEGOTIATE_128 |
NTLMSSP_NEGOTIATE_ALWAYS_SIGN |
NTLMSSP_NEGOTIATE_NTLM |
NTLMSSP_NEGOTIATE_NTLM2 |
NTLMSSP_NEGOTIATE_KEY_EXCH |
NTLMSSP_REQUEST_TARGET;
ntlmssp_state->client.netbios_name = talloc_strdup(ntlmssp_state, netbios_name);
if (!ntlmssp_state->client.netbios_name) {
talloc_free(ntlmssp_state);
return NT_STATUS_NO_MEMORY;
}
ntlmssp_state->client.netbios_domain = talloc_strdup(ntlmssp_state, netbios_domain);
if (!ntlmssp_state->client.netbios_domain) {
talloc_free(ntlmssp_state);
return NT_STATUS_NO_MEMORY;
}
*_ntlmssp_state = ntlmssp_state;
return NT_STATUS_OK;
}