1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-23 17:34:34 +03:00
samba-mirror/source4/libcli/smb_composite/sesssetup.c

432 lines
14 KiB
C
Raw Normal View History

/*
Unix SMB/CIFS implementation.
Copyright (C) Andrew Tridgell 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 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
/*
a composite API for making handling a generic async session setup
*/
#include "includes.h"
#include "libcli/raw/libcliraw.h"
#include "libcli/composite/composite.h"
#include "libcli/smb_composite/smb_composite.h"
#include "auth/auth.h"
#include "version.h"
struct sesssetup_state {
union smb_sesssetup setup;
NTSTATUS gensec_status;
struct smb_composite_sesssetup *io;
struct smbcli_request *req;
};
/*
store the user session key for a transport
*/
static void set_user_session_key(struct smbcli_session *session,
const DATA_BLOB *session_key)
{
session->user_session_key = data_blob_talloc(session,
session_key->data,
session_key->length);
}
/*
handler for completion of a smbcli_request sub-request
*/
static void request_handler(struct smbcli_request *req)
{
struct composite_context *c = req->async.private;
struct sesssetup_state *state = talloc_get_type(c->private_data, struct sesssetup_state);
struct smbcli_session *session = req->session;
DATA_BLOB session_key = data_blob(NULL, 0);
DATA_BLOB null_data_blob = data_blob(NULL, 0);
NTSTATUS session_key_err;
c->status = smb_raw_sesssetup_recv(req, state, &state->setup);
switch (state->setup.old.level) {
case RAW_SESSSETUP_OLD:
state->io->out.vuid = state->setup.old.out.vuid;
break;
case RAW_SESSSETUP_NT1:
state->io->out.vuid = state->setup.nt1.out.vuid;
break;
case RAW_SESSSETUP_SPNEGO:
session->vuid = state->io->out.vuid = state->setup.spnego.out.vuid;
if (!NT_STATUS_EQUAL(c->status, NT_STATUS_MORE_PROCESSING_REQUIRED) &&
!NT_STATUS_IS_OK(c->status)) {
break;
}
if (NT_STATUS_EQUAL(state->gensec_status, NT_STATUS_MORE_PROCESSING_REQUIRED)) {
/* The status value here, from the earlier pass at GENSEC is
* vital to the security of the system. Even if the other end
* accepts, if GENSEC claims 'MORE_PROCESSING_REQUIRED' then
* you must keep feeding it blobs, or else the remote
* host/attacker might avoid mutal authentication
* requirements */
state->gensec_status = gensec_update(session->gensec, state,
state->setup.spnego.out.secblob,
&state->setup.spnego.in.secblob);
c->status = state->gensec_status;
if (!NT_STATUS_EQUAL(c->status, NT_STATUS_MORE_PROCESSING_REQUIRED) &&
!NT_STATUS_IS_OK(c->status)) {
break;
}
} else {
state->setup.spnego.in.secblob = data_blob(NULL, 0);
}
/* we need to do another round of session setup. We keep going until both sides
are happy */
session_key_err = gensec_session_key(session->gensec, &session_key);
if (NT_STATUS_IS_OK(session_key_err)) {
set_user_session_key(session, &session_key);
smbcli_transport_simple_set_signing(session->transport, session_key, null_data_blob);
}
if (state->setup.spnego.in.secblob.length) {
state->req = smb_raw_sesssetup_send(session, &state->setup);
state->req->async.fn = request_handler;
state->req->async.private = c;
return;
}
}
/* enforce the local signing required flag */
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
if (NT_STATUS_IS_OK(c->status) && !cli_credentials_is_anonymous(state->io->in.credentials)) {
if (!session->transport->negotiate.sign_info.doing_signing
&& session->transport->negotiate.sign_info.mandatory_signing) {
DEBUG(0, ("SMB signing required, but server does not support it\n"));
c->status = NT_STATUS_ACCESS_DENIED;
}
}
if (NT_STATUS_IS_OK(c->status)) {
c->state = COMPOSITE_STATE_DONE;
} else {
c->state = COMPOSITE_STATE_ERROR;
}
if (c->async.fn) {
c->async.fn(c);
}
}
/*
send a nt1 style session setup
*/
static NTSTATUS session_setup_nt1(struct composite_context *c,
struct smbcli_session *session,
struct smb_composite_sesssetup *io,
struct smbcli_request **req)
{
NTSTATUS nt_status;
struct sesssetup_state *state = talloc_get_type(c->private_data, struct sesssetup_state);
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
const char *password = cli_credentials_get_password(io->in.credentials);
DATA_BLOB names_blob = NTLMv2_generate_names_blob(state, session->transport->socket->hostname, lp_workgroup());
DATA_BLOB session_key;
int flags = CLI_CRED_NTLM_AUTH;
if (lp_client_lanman_auth()) {
flags |= CLI_CRED_LANMAN_AUTH;
}
if (lp_client_ntlmv2_auth()) {
flags |= CLI_CRED_NTLMv2_AUTH;
}
state->setup.nt1.level = RAW_SESSSETUP_NT1;
state->setup.nt1.in.bufsize = session->transport->options.max_xmit;
state->setup.nt1.in.mpx_max = session->transport->options.max_mux;
state->setup.nt1.in.vc_num = 1;
state->setup.nt1.in.sesskey = io->in.sesskey;
state->setup.nt1.in.capabilities = io->in.capabilities;
state->setup.nt1.in.os = "Unix";
state->setup.nt1.in.lanman = talloc_asprintf(state, "Samba %s", SAMBA_VERSION_STRING);
cli_credentials_get_ntlm_username_domain(io->in.credentials, state,
&state->setup.nt1.in.user,
&state->setup.nt1.in.domain);
if (session->transport->negotiate.sec_mode & NEGOTIATE_SECURITY_CHALLENGE_RESPONSE) {
nt_status = cli_credentials_get_ntlm_response(io->in.credentials, state,
&flags,
session->transport->negotiate.secblob,
names_blob,
&state->setup.nt1.in.password1,
&state->setup.nt1.in.password2,
NULL, &session_key);
smbcli_transport_simple_set_signing(session->transport, session_key,
state->setup.nt1.in.password2);
set_user_session_key(session, &session_key);
data_blob_free(&session_key);
} else if (lp_client_plaintext_auth()) {
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
state->setup.nt1.in.password1 = data_blob_talloc(state, password, strlen(password));
state->setup.nt1.in.password2 = data_blob(NULL, 0);
} else {
/* could match windows client and return 'cannot logon from this workstation', but it just confuses everybody */
return NT_STATUS_INVALID_PARAMETER;
}
*req = smb_raw_sesssetup_send(session, &state->setup);
if (!*req) {
return NT_STATUS_NO_MEMORY;
}
return (*req)->status;
}
/*
old style session setup (pre NT1 protocol level)
*/
static NTSTATUS session_setup_old(struct composite_context *c,
struct smbcli_session *session,
struct smb_composite_sesssetup *io,
struct smbcli_request **req)
{
NTSTATUS nt_status;
struct sesssetup_state *state = talloc_get_type(c->private_data, struct sesssetup_state);
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
const char *password = cli_credentials_get_password(io->in.credentials);
DATA_BLOB names_blob = NTLMv2_generate_names_blob(state, session->transport->socket->hostname, lp_workgroup());
DATA_BLOB session_key;
int flags = 0;
if (lp_client_lanman_auth()) {
flags |= CLI_CRED_LANMAN_AUTH;
}
if (lp_client_ntlmv2_auth()) {
flags |= CLI_CRED_NTLMv2_AUTH;
}
state->setup.old.level = RAW_SESSSETUP_OLD;
state->setup.old.in.bufsize = session->transport->options.max_xmit;
state->setup.old.in.mpx_max = session->transport->options.max_mux;
state->setup.old.in.vc_num = 1;
state->setup.old.in.sesskey = io->in.sesskey;
state->setup.old.in.os = "Unix";
state->setup.old.in.lanman = talloc_asprintf(state, "Samba %s", SAMBA_VERSION_STRING);
cli_credentials_get_ntlm_username_domain(io->in.credentials, state,
&state->setup.old.in.user,
&state->setup.old.in.domain);
if (session->transport->negotiate.sec_mode & NEGOTIATE_SECURITY_CHALLENGE_RESPONSE) {
nt_status = cli_credentials_get_ntlm_response(io->in.credentials, state,
&flags,
session->transport->negotiate.secblob,
names_blob,
&state->setup.old.in.password,
NULL,
NULL, &session_key);
set_user_session_key(session, &session_key);
data_blob_free(&session_key);
} else if (lp_client_plaintext_auth()) {
state->setup.old.in.password = data_blob_talloc(state, password, strlen(password));
} else {
/* could match windows client and return 'cannot logon from this workstation', but it just confuses everybody */
return NT_STATUS_INVALID_PARAMETER;
}
*req = smb_raw_sesssetup_send(session, &state->setup);
if (!*req) {
return NT_STATUS_NO_MEMORY;
}
return (*req)->status;
}
/*
Modern, all singing, all dancing extended security (and possibly SPNEGO) request
*/
static NTSTATUS session_setup_spnego(struct composite_context *c,
struct smbcli_session *session,
struct smb_composite_sesssetup *io,
struct smbcli_request **req)
{
struct sesssetup_state *state = talloc_get_type(c->private_data, struct sesssetup_state);
NTSTATUS status, session_key_err;
DATA_BLOB session_key = data_blob(NULL, 0);
DATA_BLOB null_data_blob = data_blob(NULL, 0);
const char *chosen_oid = NULL;
state->setup.spnego.level = RAW_SESSSETUP_SPNEGO;
state->setup.spnego.in.bufsize = session->transport->options.max_xmit;
state->setup.spnego.in.mpx_max = session->transport->options.max_mux;
state->setup.spnego.in.vc_num = 1;
state->setup.spnego.in.sesskey = io->in.sesskey;
state->setup.spnego.in.capabilities = io->in.capabilities;
state->setup.spnego.in.os = "Unix";
state->setup.spnego.in.lanman = talloc_asprintf(state, "Samba %s", SAMBA_VERSION_STRING);
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
state->setup.spnego.in.workgroup = io->in.workgroup;
state->setup.spnego.out.vuid = session->vuid;
smbcli_temp_set_signing(session->transport);
status = gensec_client_start(session, &session->gensec, c->event_ctx);
if (!NT_STATUS_IS_OK(status)) {
DEBUG(1, ("Failed to start GENSEC client mode: %s\n", nt_errstr(status)));
return status;
}
gensec_want_feature(session->gensec, GENSEC_FEATURE_SESSION_KEY);
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
status = gensec_set_credentials(session->gensec, io->in.credentials);
if (!NT_STATUS_IS_OK(status)) {
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
DEBUG(1, ("Failed to start set GENSEC client credentails: %s\n",
nt_errstr(status)));
return status;
}
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
status = gensec_set_target_hostname(session->gensec, session->transport->socket->hostname);
if (!NT_STATUS_IS_OK(status)) {
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
DEBUG(1, ("Failed to start set GENSEC target hostname: %s\n",
nt_errstr(status)));
return status;
}
status = gensec_set_target_service(session->gensec, "cifs");
if (!NT_STATUS_IS_OK(status)) {
r6028: A MAJOR update to intergrate the new credentails system fully with GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'. GENSEC now no longer has it's own handling of 'set username' etc, instead it uses cli_credentials calls. In order to link the credentails code right though Samba, a lot of interfaces have changed to remove 'username, domain, password' arguments, and these have been replaced with a single 'struct cli_credentials'. In the session setup code, a new parameter 'workgroup' contains the client/server current workgroup, which seems unrelated to the authentication exchange (it was being filled in from the auth info). This allows in particular kerberos to only call back for passwords when it actually needs to perform the kinit. The kerberos code has been modified not to use the SPNEGO provided 'principal name' (in the mechListMIC), but to instead use the name the host was connected to as. This better matches Microsoft behaviour, is more secure and allows better use of standard kerberos functions. To achieve this, I made changes to our socket code so that the hostname (before name resolution) is now recorded on the socket. In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now in libcli/auth/schannel.c, and it looks much more like a standard GENSEC module. The actual sign/seal code moved to libcli/auth/schannel_sign.c in a previous commit. The schannel credentails structure is now merged with the rest of the credentails, as many of the values (username, workstation, domain) where already present there. This makes handling this in a generic manner much easier, as there is no longer a custom entry-point. The auth_domain module continues to be developed, but is now just as functional as auth_winbind. The changes here are consequential to the schannel changes. The only removed function at this point is the RPC-LOGIN test (simulating the load of a WinXP login), which needs much more work to clean it up (it contains copies of too much code from all over the torture suite, and I havn't been able to penetrate its 'structure'). Andrew Bartlett (This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
2005-03-24 07:14:06 +03:00
DEBUG(1, ("Failed to start set GENSEC target service: %s\n",
nt_errstr(status)));
return status;
}
if (session->transport->negotiate.secblob.length) {
chosen_oid = GENSEC_OID_SPNEGO;
} else {
/* without a sec blob, means raw NTLMSSP */
chosen_oid = GENSEC_OID_NTLMSSP;
}
status = gensec_start_mech_by_oid(session->gensec, chosen_oid);
if (!NT_STATUS_IS_OK(status)) {
DEBUG(1, ("Failed to start set GENSEC client mechanism %s: %s\n",
gensec_get_name_by_oid(chosen_oid), nt_errstr(status)));
return status;
}
status = gensec_update(session->gensec, state,
session->transport->negotiate.secblob,
&state->setup.spnego.in.secblob);
if (!NT_STATUS_EQUAL(status, NT_STATUS_MORE_PROCESSING_REQUIRED) &&
!NT_STATUS_IS_OK(status)) {
DEBUG(1, ("Failed initial gensec_update with mechanism %s: %s\n",
gensec_get_name_by_oid(chosen_oid), nt_errstr(status)));
return status;
}
state->gensec_status = status;
session_key_err = gensec_session_key(session->gensec, &session_key);
if (NT_STATUS_IS_OK(session_key_err)) {
smbcli_transport_simple_set_signing(session->transport, session_key, null_data_blob);
}
*req = smb_raw_sesssetup_send(session, &state->setup);
if (!*req) {
return NT_STATUS_NO_MEMORY;
}
return (*req)->status;
}
/*
composite session setup function that hides the details of all the
different session setup varients, including the multi-pass nature of
the spnego varient
*/
struct composite_context *smb_composite_sesssetup_send(struct smbcli_session *session,
struct smb_composite_sesssetup *io)
{
struct composite_context *c;
struct sesssetup_state *state;
NTSTATUS status;
c = talloc_zero(session, struct composite_context);
if (c == NULL) return NULL;
state = talloc(c, struct sesssetup_state);
if (state == NULL) {
talloc_free(c);
return NULL;
}
state->io = io;
c->state = COMPOSITE_STATE_IN_PROGRESS;
c->private_data = state;
c->event_ctx = session->transport->socket->event.ctx;
/* no session setup at all in earliest protocol varients */
if (session->transport->negotiate.protocol < PROTOCOL_LANMAN1) {
ZERO_STRUCT(io->out);
c->state = COMPOSITE_STATE_DONE;
return c;
}
/* see what session setup interface we will use */
if (session->transport->negotiate.protocol < PROTOCOL_NT1) {
status = session_setup_old(c, session, io, &state->req);
} else if (!session->transport->options.use_spnego ||
!(io->in.capabilities & CAP_EXTENDED_SECURITY)) {
status = session_setup_nt1(c, session, io, &state->req);
} else {
status = session_setup_spnego(c, session, io, &state->req);
}
if (NT_STATUS_EQUAL(status, NT_STATUS_MORE_PROCESSING_REQUIRED) ||
NT_STATUS_IS_OK(status)) {
state->req->async.fn = request_handler;
state->req->async.private = c;
return c;
}
c->state = COMPOSITE_STATE_ERROR;
c->status = status;
return c;
}
/*
receive a composite session setup reply
*/
NTSTATUS smb_composite_sesssetup_recv(struct composite_context *c)
{
NTSTATUS status;
status = composite_wait(c);
talloc_free(c);
return status;
}
/*
sync version of smb_composite_sesssetup
*/
NTSTATUS smb_composite_sesssetup(struct smbcli_session *session, struct smb_composite_sesssetup *io)
{
struct composite_context *c = smb_composite_sesssetup_send(session, io);
return smb_composite_sesssetup_recv(c);
}