1
0
mirror of https://github.com/samba-team/samba.git synced 2024-12-31 17:18:04 +03:00
samba-mirror/source4/libcli/finddcs.c

273 lines
8.5 KiB
C
Raw Normal View History

r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
/*
Unix SMB/CIFS implementation.
a composite API for finding a DC and its name
Copyright (C) Volker Lendecke 2005
Copyright (C) Andrew Bartlett <abartlet@samba.org> 2006
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
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
(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, see <http://www.gnu.org/licenses/>.
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
*/
#include "include/includes.h"
#include "lib/messaging/irpc.h"
#include "librpc/gen_ndr/ndr_irpc.h"
#include "librpc/gen_ndr/samr.h"
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
#include "libcli/composite/composite.h"
#include "libcli/libcli.h"
#include "libcli/resolve/resolve.h"
#include "libcli/finddcs.h"
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
struct finddcs_state {
struct composite_context *ctx;
struct messaging_context *msg_ctx;
const char *my_netbios_name;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
const char *domain_name;
struct dom_sid *domain_sid;
struct nbtd_getdcname r;
struct nbt_name_status node_status;
int num_dcs;
struct nbt_dc_name *dcs;
uint16_t nbt_port;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
};
static void finddcs_name_resolved(struct composite_context *ctx);
static void finddcs_getdc_replied(struct irpc_request *ireq);
static void fallback_node_status(struct finddcs_state *state);
static void fallback_node_status_replied(struct nbt_name_request *name_req);
/*
* Setup and send off the a normal name resolution for the target name.
*
* The domain_sid parameter is optional, and is used in the subsequent getdc request.
*
* This will try a GetDC request, but this may not work. It will try
* a node status as a fallback, then return no name (but still include
* the IP)
*/
struct composite_context *finddcs_send(TALLOC_CTX *mem_ctx,
const char *my_netbios_name,
uint16_t nbt_port,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
const char *domain_name,
int name_type,
struct dom_sid *domain_sid,
struct resolve_context *resolve_ctx,
struct tevent_context *event_ctx,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
struct messaging_context *msg_ctx)
{
struct composite_context *c, *creq;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
struct finddcs_state *state;
struct nbt_name name;
c = composite_create(mem_ctx, event_ctx);
if (c == NULL) return NULL;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state = talloc(c, struct finddcs_state);
if (composite_nomem(state, c)) return c;
c->private_data = state;
state->ctx = c;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->nbt_port = nbt_port;
state->my_netbios_name = talloc_strdup(state, my_netbios_name);
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->domain_name = talloc_strdup(state, domain_name);
if (composite_nomem(state->domain_name, c)) return c;
if (domain_sid) {
state->domain_sid = talloc_reference(state, domain_sid);
if (composite_nomem(state->domain_sid, c)) return c;
} else {
state->domain_sid = NULL;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
}
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->msg_ctx = msg_ctx;
make_nbt_name(&name, state->domain_name, name_type);
creq = resolve_name_send(resolve_ctx, state, &name, event_ctx);
composite_continue(c, creq, finddcs_name_resolved, state);
return c;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
}
/* Having got an name query answer, fire off a GetDC request, so we
* can find the target's all-important name. (Kerberos and some
* netlogon operations are quite picky about names)
*
* The name is a courtesy, if we don't find it, don't completely fail.
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
*
* However, if the nbt server is down, fall back to a node status
* request
*/
static void finddcs_name_resolved(struct composite_context *ctx)
{
struct finddcs_state *state =
talloc_get_type(ctx->async.private_data, struct finddcs_state);
struct irpc_request *ireq;
struct server_id *nbt_servers;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
const char *address;
state->ctx->status = resolve_name_recv(ctx, state, &address);
if (!composite_is_ok(state->ctx)) return;
/* TODO: This should try and find all the DCs, and give the
* caller them in the order they responded */
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->num_dcs = 1;
state->dcs = talloc_array(state, struct nbt_dc_name, state->num_dcs);
if (composite_nomem(state->dcs, state->ctx)) return;
state->dcs[0].address = talloc_steal(state->dcs, address);
/* Try and find the nbt server. Fallback to a node status
* request if we can't make this happen The nbt server just
* might not be running, or we may not have a messaging
* context (not root etc) */
if (!state->msg_ctx) {
fallback_node_status(state);
return;
}
nbt_servers = irpc_servers_byname(state->msg_ctx, state, "nbt_server");
if ((nbt_servers == NULL) || (nbt_servers[0].id == 0)) {
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
fallback_node_status(state);
return;
}
state->r.in.domainname = state->domain_name;
state->r.in.ip_address = state->dcs[0].address;
state->r.in.my_computername = state->my_netbios_name;
state->r.in.my_accountname = talloc_asprintf(state, "%s$", state->my_netbios_name);
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
if (composite_nomem(state->r.in.my_accountname, state->ctx)) return;
state->r.in.account_control = ACB_WSTRUST;
state->r.in.domain_sid = state->domain_sid;
if (state->r.in.domain_sid == NULL) {
state->r.in.domain_sid = talloc_zero(state, struct dom_sid);
}
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
ireq = irpc_call_send(state->msg_ctx, nbt_servers[0],
&ndr_table_irpc, NDR_NBTD_GETDCNAME,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
&state->r, state);
if (!ireq) {
fallback_node_status(state);
return;
}
composite_continue_irpc(state->ctx, ireq, finddcs_getdc_replied, state);
}
/* Called when the GetDC request returns */
static void finddcs_getdc_replied(struct irpc_request *ireq)
{
struct finddcs_state *state =
talloc_get_type(ireq->async.private_data, struct finddcs_state);
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->ctx->status = irpc_call_recv(ireq);
if (!composite_is_ok(state->ctx)) return;
state->dcs[0].name = talloc_steal(state->dcs, state->r.out.dcname);
composite_done(state->ctx);
}
/* The GetDC request might not be available (such as occours when the
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
* NBT server is down). Fallback to a node status. It is the best
* hope we have... */
static void fallback_node_status(struct finddcs_state *state)
{
struct nbt_name_socket *nbtsock;
struct nbt_name_request *name_req;
state->node_status.in.name.name = "*";
state->node_status.in.name.type = NBT_NAME_CLIENT;
state->node_status.in.name.scope = NULL;
state->node_status.in.dest_addr = state->dcs[0].address;
state->node_status.in.dest_port = state->nbt_port;
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->node_status.in.timeout = 1;
state->node_status.in.retries = 2;
nbtsock = nbt_name_socket_init(state, state->ctx->event_ctx);
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
if (composite_nomem(nbtsock, state->ctx)) return;
name_req = nbt_name_status_send(nbtsock, &state->node_status);
if (composite_nomem(name_req, state->ctx)) return;
composite_continue_nbt(state->ctx,
name_req,
fallback_node_status_replied,
state);
}
/* We have a node status reply (or perhaps a timeout) */
static void fallback_node_status_replied(struct nbt_name_request *name_req)
{
int i;
struct finddcs_state *state = talloc_get_type(name_req->async.private_data, struct finddcs_state);
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
state->ctx->status = nbt_name_status_recv(name_req, state, &state->node_status);
if (!composite_is_ok(state->ctx)) return;
for (i=0; i < state->node_status.out.status.num_names; i++) {
int j;
if (state->node_status.out.status.names[i].type == NBT_NAME_SERVER) {
char *name = talloc_strndup(state->dcs, state->node_status.out.status.names[0].name, 15);
/* Strip space padding */
if (name) {
j = MIN(strlen(name), 15);
for (; j > 0 && name[j - 1] == ' '; j--) {
name[j - 1] = '\0';
}
}
state->dcs[0].name = name;
composite_done(state->ctx);
return;
}
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
}
composite_error(state->ctx, NT_STATUS_NO_LOGON_SERVERS);
}
NTSTATUS finddcs_recv(struct composite_context *c, TALLOC_CTX *mem_ctx,
int *num_dcs, struct nbt_dc_name **dcs)
{
NTSTATUS status = composite_wait(c);
if (NT_STATUS_IS_OK(status)) {
struct finddcs_state *state =
talloc_get_type(c->private_data, struct finddcs_state);
*num_dcs = state->num_dcs;
*dcs = talloc_steal(mem_ctx, state->dcs);
}
talloc_free(c);
return status;
}
NTSTATUS finddcs(TALLOC_CTX *mem_ctx,
const char *my_netbios_name,
uint16_t nbt_port,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
const char *domain_name, int name_type,
struct dom_sid *domain_sid,
struct resolve_context *resolve_ctx,
struct tevent_context *event_ctx,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
struct messaging_context *msg_ctx,
int *num_dcs, struct nbt_dc_name **dcs)
{
struct composite_context *c = finddcs_send(mem_ctx,
my_netbios_name,
nbt_port,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
domain_name, name_type,
domain_sid,
resolve_ctx,
r12858: This moves the libnet_LookupPdc code to use a GetDC request to find the remote server's name, or in the absence of a local nbt_server to communicate with (or without root access), a node status request. The result is that we are in a better position to use kerberos, as well as to remove the 'password server' mandatory parameter for the samsync and samdump commands. (I need this to put these into SWAT). The only problem I have is that I must create a messaging context, which requires a server ID. As a client process, I don't expect to get messages, but it is currently required for replies, so I generate a random() number. We probably need the servers to accept connections on streamed sockets too, for client-only tasks that want IRPC. Because I wanted to test this code, I have put the NET-API-* tests into our test scripts, to ensure they pass and keep passing. They are good frontends onto the libnet system, and I see no reason not to test them. In doing so the NET-API-RPCCONNECT test was simplified to take a binding string on the command line, removing duplicate code, and testing the combinations in the scripts instead. (I have done a bit of work on the list shares code in libnet_share.c to make it pass 'make test') In the future, I would like to extend the libcli/findds.c code (based off volker's winbind/wb_async_helpers.c, which is why it shows up a bit odd in the patch) to handle getting multiple name replies, sending a getdc request to each in turn. (posted to samba-technical for review, and I'll happily update with any comments) Andrew Bartlett (This used to be commit 7ccddfd3515fc2c0d6f447c768ccbf7a220c3380)
2006-01-12 06:02:00 +03:00
event_ctx, msg_ctx);
return finddcs_recv(c, mem_ctx, num_dcs, dcs);
}