summaryrefslogtreecommitdiff
path: root/source3
diff options
context:
space:
mode:
Diffstat (limited to 'source3')
-rw-r--r--source3/Makefile.in1
-rw-r--r--source3/rpc_server/lsasd.c1034
2 files changed, 1035 insertions, 0 deletions
diff --git a/source3/Makefile.in b/source3/Makefile.in
index 41bd45db9e..197aa1d87e 100644
--- a/source3/Makefile.in
+++ b/source3/Makefile.in
@@ -912,6 +912,7 @@ SMBD_OBJ_SRV = smbd/server_reload.o \
smbd/posix_acls.o lib/sysacls.o \
smbd/process.o smbd/service.o param/service.o smbd/error.o \
rpc_server/epmd.o \
+ rpc_server/lsasd.o \
printing/printspoolss.o printing/spoolssd.o \
lib/sysquotas.o lib/sysquotas_linux.o \
lib/sysquotas_xfs.o lib/sysquotas_4A.o \
diff --git a/source3/rpc_server/lsasd.c b/source3/rpc_server/lsasd.c
new file mode 100644
index 0000000000..bb7fb789d5
--- /dev/null
+++ b/source3/rpc_server/lsasd.c
@@ -0,0 +1,1034 @@
+/*
+ * Unix SMB/CIFS implementation.
+ *
+ * LSA service daemon
+ *
+ * Copyright (c) 2011 Andreas Schneider <asn@samba.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * 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/>.
+ */
+
+#include "includes.h"
+#include "serverid.h"
+#include "messages.h"
+#include "ntdomain.h"
+
+#include "../lib/tsocket/tsocket.h"
+#include "lib/server_prefork.h"
+#include "librpc/rpc/dcerpc_ep.h"
+
+#include "rpc_server/rpc_server.h"
+#include "rpc_server/rpc_ep_register.h"
+#include "rpc_server/rpc_sock_helper.h"
+
+#include "librpc/gen_ndr/srv_lsa.h"
+#include "librpc/gen_ndr/srv_samr.h"
+#include "librpc/gen_ndr/srv_netlogon.h"
+
+#define DAEMON_NAME "lsasd"
+
+#define LSASD_MIN_CHILDREN 5
+#define LSASD_MAX_CHILDREN 25
+#define LSASD_SPAWN_RATE 5
+#define LSASD_MIN_LIFE 60 /* 1 minute minimum life time */
+
+#define LSASD_MAX_SOCKETS 64
+
+#define LSASD_ALL_FINE 0x00
+#define LSASD_NEW_MAX 0x01
+#define LSASD_ENOSPC 0x02
+
+static int lsasd_min_children;
+static int lsasd_max_children;
+static int lsasd_spawn_rate;
+static int lsasd_prefork_status;
+
+void start_lsasd(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx);
+
+static void lsasd_prefork_config(void)
+{
+ static int lsasd_prefork_config_init = false;
+ const char *prefork_str;
+ int min, max, rate;
+ bool use_defaults = false;
+ int ret;
+
+ if (!lsasd_prefork_config_init) {
+ lsasd_prefork_status = LSASD_ALL_FINE;
+ lsasd_min_children = 0;
+ lsasd_max_children = 0;
+ lsasd_spawn_rate = 0;
+ lsasd_prefork_config_init = true;
+ }
+
+ prefork_str = lp_parm_const_string(GLOBAL_SECTION_SNUM,
+ "lsasd", "prefork", "none");
+ if (strcmp(prefork_str, "none") == 0) {
+ use_defaults = true;
+ } else {
+ ret = sscanf(prefork_str, "%d:%d:%d", &min, &max, &rate);
+ if (ret != 3) {
+ DEBUG(0, ("invalid format for lsasd:prefork!\n"));
+ use_defaults = true;
+ }
+ }
+
+ if (use_defaults) {
+ min = LSASD_MIN_CHILDREN;
+ max = LSASD_MAX_CHILDREN;
+ rate = LSASD_SPAWN_RATE;
+ }
+
+ if (max > lsasd_max_children && lsasd_max_children != 0) {
+ lsasd_prefork_status |= LSASD_NEW_MAX;
+ }
+
+ lsasd_min_children = min;
+ lsasd_max_children = max;
+ lsasd_spawn_rate = rate;
+}
+
+static void lsasd_reopen_logs(int child_id)
+{
+ char *lfile = lp_logfile();
+ char *extension;
+ int rc;
+
+ if (child_id) {
+ rc = asprintf(&extension, "%s.%d", DAEMON_NAME, child_id);
+ } else {
+ rc = asprintf(&extension, "%s", DAEMON_NAME);
+ }
+ if (rc == -1) {
+ return;
+ }
+
+ rc = 0;
+ if (lfile == NULL || lfile[0] == '\0') {
+ rc = asprintf(&lfile, "%s/log.%s",
+ get_dyn_LOGFILEBASE(), extension);
+ } else {
+ if (strstr(lfile, extension) == NULL) {
+ if (child_id) {
+ rc = asprintf(&lfile, "%s.%d",
+ lp_logfile(),
+ child_id);
+ } else {
+ rc = asprintf(&lfile, "%s.%s",
+ lp_logfile(),
+ extension);
+ }
+ }
+ }
+
+ if (rc > 0) {
+ lp_set_logfile(lfile);
+ SAFE_FREE(lfile);
+ }
+
+ SAFE_FREE(extension);
+
+ reopen_logs();
+}
+
+static void lsasd_smb_conf_updated(struct messaging_context *msg,
+ void *private_data,
+ uint32_t msg_type,
+ struct server_id server_id,
+ DATA_BLOB *data)
+{
+ DEBUG(10, ("Got message saying smb.conf was updated. Reloading.\n"));
+ change_to_root_user();
+ lp_load(get_dyn_CONFIGFILE(), true, false, false, true);
+
+ lsasd_reopen_logs(0);
+ lsasd_prefork_config();
+}
+
+static void lsasd_sig_term_handler(struct tevent_context *ev,
+ struct tevent_signal *se,
+ int signum,
+ int count,
+ void *siginfo,
+ void *private_data)
+{
+ rpc_netlogon_shutdown();
+ rpc_samr_shutdown();
+ rpc_lsarpc_shutdown();
+
+ DEBUG(0, ("termination signal\n"));
+ exit(0);
+}
+
+static void lsasd_setup_sig_term_handler(struct tevent_context *ev_ctx)
+{
+ struct tevent_signal *se;
+
+ se = tevent_add_signal(ev_ctx,
+ ev_ctx,
+ SIGTERM, 0,
+ lsasd_sig_term_handler,
+ NULL);
+ if (!se) {
+ DEBUG(0, ("failed to setup SIGTERM handler\n"));
+ exit(1);
+ }
+}
+
+struct lsasd_hup_ctx {
+ struct messaging_context *msg_ctx;
+ struct prefork_pool *pfp;
+};
+
+static void lsasd_sig_hup_handler(struct tevent_context *ev,
+ struct tevent_signal *se,
+ int signum,
+ int count,
+ void *siginfo,
+ void *pvt)
+{
+ struct lsasd_hup_ctx *hup_ctx;
+
+ hup_ctx = talloc_get_type_abort(pvt, struct lsasd_hup_ctx);
+
+ change_to_root_user();
+ lp_load(get_dyn_CONFIGFILE(), true, false, false, true);
+
+ lsasd_reopen_logs(0);
+ lsasd_prefork_config();
+
+ /* relay to all children */
+ prefork_send_signal_to_all(hup_ctx->pfp, SIGHUP);
+}
+
+static void lsasd_setup_sig_hup_handler(struct tevent_context *ev_ctx,
+ struct prefork_pool *pfp,
+ struct messaging_context *msg_ctx)
+{
+ struct lsasd_hup_ctx *hup_ctx;
+ struct tevent_signal *se;
+
+ hup_ctx = talloc(ev_ctx, struct lsasd_hup_ctx);
+ if (!hup_ctx) {
+ DEBUG(0, ("failed to setup SIGHUP handler\n"));
+ exit(1);
+ }
+ hup_ctx->pfp = pfp;
+ hup_ctx->msg_ctx = msg_ctx;
+
+ se = tevent_add_signal(ev_ctx,
+ ev_ctx,
+ SIGHUP, 0,
+ lsasd_sig_hup_handler,
+ hup_ctx);
+ if (!se) {
+ DEBUG(0, ("failed to setup SIGHUP handler\n"));
+ exit(1);
+ }
+}
+
+/**********************************************************
+ * Children
+ **********************************************************/
+
+struct lsasd_chld_sig_hup_ctx {
+ struct messaging_context *msg_ctx;
+ struct pf_worker_data *pf;
+ int child_id;
+};
+
+static void lsasd_chld_sig_hup_handler(struct tevent_context *ev,
+ struct tevent_signal *se,
+ int signum,
+ int count,
+ void *siginfo,
+ void *pvt)
+{
+ struct lsasd_chld_sig_hup_ctx *shc;
+
+ shc = talloc_get_type_abort(pvt, struct lsasd_chld_sig_hup_ctx);
+
+ /* avoid wasting CPU cycles if we are going to exit soon anyways */
+ if (shc->pf != NULL &&
+ shc->pf->cmds == PF_SRV_MSG_EXIT) {
+ return;
+ }
+
+ change_to_root_user();
+ lsasd_reopen_logs(shc->child_id);
+}
+
+static bool lsasd_setup_chld_hup_handler(struct tevent_context *ev_ctx,
+ struct pf_worker_data *pf,
+ struct messaging_context *msg_ctx,
+ int child_id)
+{
+ struct lsasd_chld_sig_hup_ctx *shc;
+ struct tevent_signal *se;
+
+ shc = talloc(ev_ctx, struct lsasd_chld_sig_hup_ctx);
+ if (!shc) {
+ DEBUG(1, ("failed to setup SIGHUP handler"));
+ return false;
+ }
+ shc->child_id = child_id;
+ shc->pf = pf;
+ shc->msg_ctx = msg_ctx;
+
+ se = tevent_add_signal(ev_ctx,
+ ev_ctx,
+ SIGHUP, 0,
+ lsasd_chld_sig_hup_handler,
+ shc);
+ if (!se) {
+ DEBUG(1, ("failed to setup SIGHUP handler"));
+ return false;
+ }
+
+ return true;
+}
+
+static bool lsasd_child_init(struct tevent_context *ev_ctx,
+ int child_id,
+ struct pf_worker_data *pf)
+{
+ NTSTATUS status;
+ struct messaging_context *msg_ctx = server_messaging_context();
+ bool ok;
+
+ status = reinit_after_fork(msg_ctx, ev_ctx,
+ procid_self(), true);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0,("reinit_after_fork() failed\n"));
+ smb_panic("reinit_after_fork() failed");
+ }
+
+ lsasd_reopen_logs(child_id);
+
+ ok = lsasd_setup_chld_hup_handler(ev_ctx, pf, msg_ctx, child_id);
+ if (!ok) {
+ return false;
+ }
+
+ if (!serverid_register(procid_self(), FLAG_MSG_GENERAL)) {
+ return false;
+ }
+
+ messaging_register(msg_ctx, ev_ctx,
+ MSG_SMB_CONF_UPDATED, lsasd_smb_conf_updated);
+
+ status = rpc_lsarpc_init(NULL);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0, ("Failed to register lsarpc rpc inteface! (%s)\n",
+ nt_errstr(status)));
+ return false;
+ }
+
+ status = rpc_samr_init(NULL);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0, ("Failed to register samr rpc inteface! (%s)\n",
+ nt_errstr(status)));
+ return false;
+ }
+
+ status = rpc_netlogon_init(NULL);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0, ("Failed to register netlogon rpc inteface! (%s)\n",
+ nt_errstr(status)));
+ return false;
+ }
+
+ return true;
+}
+
+struct lsasd_children_data {
+ struct tevent_context *ev_ctx;
+ struct messaging_context *msg_ctx;
+ int child_id;
+ struct pf_worker_data *pf;
+ int listen_fd_size;
+ int *listen_fds;
+ int lock_fd;
+
+ bool listening;
+};
+
+static void lsasd_next_client(void *pvt);
+
+static int lsasd_children_main(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx,
+ struct pf_worker_data *pf,
+ int child_id,
+ int listen_fd_size,
+ int *listen_fds,
+ int lock_fd,
+ void *private_data)
+{
+ struct lsasd_children_data *data;
+ bool ok;
+ int ret;
+
+ ok = lsasd_child_init(ev_ctx, child_id, pf);
+ if (!ok) {
+ return 1;
+ }
+
+ data = talloc(ev_ctx, struct lsasd_children_data);
+ if (!data) {
+ return 1;
+ }
+ data->child_id = child_id;
+ data->pf = pf;
+ data->ev_ctx = ev_ctx;
+ data->msg_ctx = msg_ctx;
+ data->lock_fd = lock_fd;
+ data->listen_fd_size = listen_fd_size;
+ data->listen_fds = listen_fds;
+ data->listening = false;
+
+ /* loop until it is time to exit */
+ while (pf->status != PF_WORKER_EXITING) {
+ /* try to see if it is time to schedule the next client */
+ lsasd_next_client(data);
+
+ ret = tevent_loop_once(ev_ctx);
+ if (ret != 0) {
+ DEBUG(0, ("tevent_loop_once() exited with %d: %s\n",
+ ret, strerror(errno)));
+ pf->status = PF_WORKER_EXITING;
+ }
+ }
+
+ return ret;
+}
+
+static void lsasd_client_terminated(void *pvt)
+{
+ struct lsasd_children_data *data;
+
+ data = talloc_get_type_abort(pvt, struct lsasd_children_data);
+
+ if (data->pf->num_clients) {
+ data->pf->num_clients--;
+ } else {
+ DEBUG(2, ("Invalid num clients, aborting!\n"));
+ data->pf->status = PF_WORKER_EXITING;
+ return;
+ }
+
+ lsasd_next_client(pvt);
+}
+
+struct lsasd_new_client {
+ struct lsasd_children_data *data;
+};
+
+static void lsasd_handle_client(struct tevent_req *req);
+
+static void lsasd_next_client(void *pvt)
+{
+ struct tevent_req *req;
+ struct lsasd_children_data *data;
+ struct lsasd_new_client *next;
+
+ data = talloc_get_type_abort(pvt, struct lsasd_children_data);
+
+ if (data->pf->num_clients == 0) {
+ data->pf->status = PF_WORKER_IDLE;
+ }
+
+ if (data->pf->cmds == PF_SRV_MSG_EXIT) {
+ DEBUG(2, ("Parent process commands we terminate!\n"));
+ return;
+ }
+
+ if (data->listening ||
+ data->pf->num_clients >= data->pf->allowed_clients) {
+ /* nothing to do for now we are already listening
+ * or reached the number of clients we are allowed
+ * to handle in parallel */
+ return;
+ }
+
+ next = talloc_zero(data, struct lsasd_new_client);
+ if (!next) {
+ DEBUG(1, ("Out of memory!?\n"));
+ return;
+ }
+ next->data = data;
+
+ req = prefork_listen_send(next,
+ data->ev_ctx,
+ data->pf,
+ data->listen_fd_size,
+ data->listen_fds,
+ data->lock_fd);
+ if (!req) {
+ DEBUG(1, ("Failed to make listening request!?\n"));
+ talloc_free(next);
+ return;
+ }
+ tevent_req_set_callback(req, lsasd_handle_client, next);
+
+ data->listening = true;
+}
+
+static void lsasd_handle_client(struct tevent_req *req)
+{
+ struct lsasd_children_data *data;
+ struct lsasd_new_client *client;
+ int rc;
+ int sd;
+ TALLOC_CTX *tmp_ctx;
+ struct tsocket_address *srv_addr;
+ struct tsocket_address *cli_addr;
+
+ client = tevent_req_callback_data(req, struct lsasd_new_client);
+ data = client->data;
+
+ tmp_ctx = talloc_stackframe();
+ if (tmp_ctx == NULL) {
+ DEBUG(1, ("Failed to allocate stackframe!\n"));
+ return;
+ }
+
+ rc = prefork_listen_recv(req,
+ tmp_ctx,
+ &sd,
+ &srv_addr,
+ &cli_addr);
+
+ /* this will free the request too */
+ talloc_free(client);
+ /* we are done listening */
+ data->listening = false;
+
+ if (rc > 0) {
+ DEBUG(1, ("Failed to accept client connection!\n"));
+ /* bail out if we are not serving any other client */
+ if (data->pf->num_clients == 0) {
+ data->pf->status = PF_WORKER_EXITING;
+ }
+ return;
+ }
+
+ if (rc == -2) {
+ DEBUG(1, ("Server asks us to die!\n"));
+ data->pf->status = PF_WORKER_EXITING;
+ return;
+ }
+
+ DEBUG(2, ("LSASD preforked child %d got client connection!\n",
+ (int)(data->pf->pid)));
+
+ if (tsocket_address_is_inet(srv_addr, "ip")) {
+ DEBUG(3, ("Got a tcpip client connection from %s on inteface %s\n",
+ tsocket_address_string(cli_addr, tmp_ctx),
+ tsocket_address_string(srv_addr, tmp_ctx)));
+
+ dcerpc_ncacn_accept(data->ev_ctx,
+ data->msg_ctx,
+ NCACN_IP_TCP,
+ "IP",
+ cli_addr,
+ srv_addr,
+ sd,
+ NULL);
+ } else if (tsocket_address_is_unix(srv_addr)) {
+ char *p;
+
+ p = tsocket_address_unix_path(srv_addr, tmp_ctx);
+ if (p == NULL) {
+ talloc_free(tmp_ctx);
+ return;
+ }
+
+ if (strstr(p, "/np/")) {
+ p = basename(p);
+
+ named_pipe_accept_function(data->ev_ctx,
+ data->msg_ctx,
+ p,
+ sd,
+ lsasd_client_terminated,
+ data);
+ } else {
+ p = basename(p);
+
+ dcerpc_ncacn_accept(data->ev_ctx,
+ data->msg_ctx,
+ NCALRPC,
+ p,
+ cli_addr,
+ srv_addr,
+ sd,
+ NULL);
+ }
+ } else {
+ DEBUG(0, ("ERROR: Unsupported socket!\n"));
+ }
+
+ talloc_free(tmp_ctx);
+}
+
+/*
+ * MAIN
+ */
+
+static bool lsasd_schedule_check(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx,
+ struct prefork_pool *pfp,
+ struct timeval current_time);
+
+static void lsasd_check_children(struct tevent_context *ev_ctx,
+ struct tevent_timer *te,
+ struct timeval current_time,
+ void *pvt);
+
+static void lsasd_sigchld_handler(struct tevent_context *ev_ctx,
+ struct prefork_pool *pfp,
+ void *pvt)
+{
+ struct messaging_context *msg_ctx;
+ int active, total;
+ int n, r;
+
+ msg_ctx = talloc_get_type_abort(pvt, struct messaging_context);
+
+ /* now check we do not descend below the minimum */
+ active = prefork_count_active_children(pfp, &total);
+
+ n = 0;
+ if (total < lsasd_min_children) {
+ n = total - lsasd_min_children;
+ } else if (total - active < (total / 4)) {
+ n = lsasd_min_children;
+ }
+
+ if (n > 0) {
+ r = prefork_add_children(ev_ctx, msg_ctx, pfp, n);
+ if (r < n) {
+ DEBUG(10, ("Tried to start %d children but only,"
+ "%d were actually started.!\n", n, r));
+ }
+ }
+}
+
+static bool lsasd_setup_children_monitor(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx,
+ struct prefork_pool *pfp)
+{
+ bool ok;
+
+ /* add our oun sigchld callback */
+ prefork_set_sigchld_callback(pfp, lsasd_sigchld_handler, msg_ctx);
+
+ ok = lsasd_schedule_check(ev_ctx,
+ msg_ctx,
+ pfp,
+ tevent_timeval_current());
+
+ return ok;
+}
+
+struct schedule_check_state {
+ struct messaging_context *msg_ctx;
+ struct prefork_pool *pfp;
+};
+
+static bool lsasd_schedule_check(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx,
+ struct prefork_pool *pfp,
+ struct timeval current_time)
+{
+ struct tevent_timer *te;
+ struct timeval next_event;
+ struct schedule_check_state *state;
+
+ state = talloc(ev_ctx, struct schedule_check_state);
+ if (!state) {
+ DEBUG(0, ("Out of memory!\n"));
+ return false;
+ }
+ state->msg_ctx = msg_ctx;
+ state->pfp = pfp;
+
+ /* check situation again in 10 seconds */
+ next_event = tevent_timeval_current_ofs(10, 0);
+
+ /* TODO: check when the socket becomes readable, so that children
+ * are checked only when there is some activity ? */
+ te = tevent_add_timer(ev_ctx,
+ pfp,
+ next_event,
+ lsasd_check_children,
+ state);
+ if (!te) {
+ DEBUG(2, ("Failed to set up children monitoring!\n"));
+ talloc_free(state);
+ return false;
+ }
+ talloc_steal(te, state);
+
+ return true;
+}
+
+static void lsasd_check_children(struct tevent_context *ev_ctx,
+ struct tevent_timer *te,
+ struct timeval current_time,
+ void *pvt)
+{
+ struct schedule_check_state *state;
+ int active, total;
+ int rc, n;
+ bool ok;
+
+ state = talloc_get_type_abort(pvt, struct schedule_check_state);
+
+ if ((lsasd_prefork_status & LSASD_NEW_MAX) &&
+ !(lsasd_prefork_status & LSASD_ENOSPC)) {
+ rc = prefork_expand_pool(state->pfp, lsasd_max_children);
+ if (rc == ENOSPC) {
+ lsasd_prefork_status |= LSASD_ENOSPC;
+ }
+ lsasd_prefork_status &= ~LSASD_NEW_MAX;
+ }
+
+ active = prefork_count_active_children(state->pfp, &total);
+
+ if (total - active < lsasd_spawn_rate) {
+ n = prefork_add_children(ev_ctx,
+ state->msg_ctx,
+ state->pfp,
+ lsasd_spawn_rate);
+ if (n < lsasd_spawn_rate) {
+ DEBUG(10, ("Tried to start 5 children but only,"
+ "%d were actually started.!\n", n));
+ }
+ }
+
+ if (total - active > lsasd_min_children) {
+ if ((total - lsasd_min_children) >= lsasd_spawn_rate) {
+ prefork_retire_children(state->pfp,
+ lsasd_spawn_rate,
+ time(NULL) - LSASD_MIN_LIFE);
+ }
+ }
+
+ ok = lsasd_schedule_check(ev_ctx,
+ state->msg_ctx,
+ state->pfp,
+ current_time);
+}
+
+/*
+ * start it up
+ */
+
+static bool lsasd_create_sockets(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx,
+ int *listen_fd,
+ int *listen_fd_size)
+{
+ struct dcerpc_binding_vector *v, *v_orig;
+ TALLOC_CTX *tmp_ctx;
+ NTSTATUS status;
+ uint32_t i;
+ int fd;
+ int rc;
+ bool ok = true;
+
+ tmp_ctx = talloc_stackframe();
+ if (tmp_ctx == NULL) {
+ return false;
+ }
+
+ status = dcerpc_binding_vector_new(tmp_ctx, &v_orig);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ /* Create only one tcpip listener for all services */
+ status = rpc_create_tcpip_sockets(&ndr_table_lsarpc,
+ v_orig,
+ 0,
+ listen_fd,
+ listen_fd_size);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ /* Start to listen on tcpip sockets */
+ for (i = 0; i < *listen_fd_size; i++) {
+ rc = listen(listen_fd[i], lsasd_max_children);
+ if (rc == -1) {
+ DEBUG(0, ("Failed to listen on tcpip socket - %s\n",
+ strerror(errno)));
+ ok = false;
+ goto done;
+ }
+ }
+
+ /* LSARPC */
+ fd = create_named_pipe_socket("lsarpc");
+ if (fd < 0) {
+ ok = false;
+ goto done;
+ }
+ listen_fd[*listen_fd_size] = fd;
+ (*listen_fd_size)++;
+
+ rc = listen(fd, lsasd_max_children);
+ if (rc == -1) {
+ DEBUG(0, ("Failed to listen on lsarpc pipe - %s\n",
+ strerror(errno)));
+ ok = false;
+ goto done;
+ }
+
+ v = dcerpc_binding_vector_dup(tmp_ctx, v_orig);
+ if (v == NULL) {
+ ok = false;
+ goto done;
+ }
+
+ status = dcerpc_binding_vector_replace_iface(&ndr_table_lsarpc, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ return false;
+ }
+
+ status = dcerpc_binding_vector_add_np_default(&ndr_table_lsarpc, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ status = rpc_ep_register(ev_ctx, msg_ctx, &ndr_table_lsarpc, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ /* SAMR */
+ fd = create_named_pipe_socket("samr");
+ if (fd < 0) {
+ ok = false;
+ goto done;
+ }
+
+ rc = listen(fd, lsasd_max_children);
+ if (rc == -1) {
+ DEBUG(0, ("Failed to listen on samr pipe - %s\n",
+ strerror(errno)));
+ ok = false;
+ goto done;
+ }
+ listen_fd[*listen_fd_size] = fd;
+ (*listen_fd_size)++;
+
+ v = dcerpc_binding_vector_dup(tmp_ctx, v_orig);
+ if (v == NULL) {
+ ok = false;
+ goto done;
+ }
+
+ status = dcerpc_binding_vector_replace_iface(&ndr_table_samr, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ return false;
+ }
+
+ status = dcerpc_binding_vector_add_np_default(&ndr_table_samr, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ status = rpc_ep_register(ev_ctx, msg_ctx, &ndr_table_samr, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ /* NETLOGON */
+ fd = create_named_pipe_socket("netlogon");
+ if (fd < 0) {
+ ok = false;
+ goto done;
+ }
+
+ rc = listen(fd, lsasd_max_children);
+ if (rc == -1) {
+ DEBUG(0, ("Failed to listen on samr pipe - %s\n",
+ strerror(errno)));
+ ok = false;
+ goto done;
+ }
+ listen_fd[*listen_fd_size] = fd;
+ (*listen_fd_size)++;
+
+ v = dcerpc_binding_vector_dup(tmp_ctx, v_orig);
+ if (v == NULL) {
+ ok = false;
+ goto done;
+ }
+
+ status = dcerpc_binding_vector_replace_iface(&ndr_table_netlogon, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ return false;
+ }
+
+ status = dcerpc_binding_vector_add_np_default(&ndr_table_netlogon, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+ status = rpc_ep_register(ev_ctx, msg_ctx, &ndr_table_netlogon, v);
+ if (!NT_STATUS_IS_OK(status)) {
+ ok = false;
+ goto done;
+ }
+
+done:
+ talloc_free(tmp_ctx);
+ return ok;
+}
+
+void start_lsasd(struct tevent_context *ev_ctx,
+ struct messaging_context *msg_ctx)
+{
+ struct prefork_pool *pool;
+ NTSTATUS status;
+ int listen_fd[LSASD_MAX_SOCKETS];
+ int listen_fd_size = 0;
+ pid_t pid;
+ int rc;
+ bool ok;
+
+ DEBUG(1, ("Forking LSA Service Daemon\n"));
+
+ /*
+ * Block signals before forking child as it will have to
+ * set its own handlers. Child will re-enable SIGHUP as
+ * soon as the handlers are set up.
+ */
+ BlockSignals(true, SIGTERM);
+ BlockSignals(true, SIGHUP);
+
+ pid = sys_fork();
+ if (pid == -1) {
+ DEBUG(0, ("Failed to fork LSASD [%s], aborting ...\n",
+ strerror(errno)));
+ exit(1);
+ }
+
+ /* parent or error */
+ if (pid != 0) {
+
+ /* Re-enable SIGHUP before returnig */
+ BlockSignals(false, SIGTERM);
+ BlockSignals(false, SIGHUP);
+
+ return;
+ }
+
+ /* child */
+ close_low_fds(false);
+
+ status = reinit_after_fork(msg_ctx,
+ ev_ctx,
+ procid_self(), true);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0,("reinit_after_fork() failed\n"));
+ smb_panic("reinit_after_fork() failed");
+ }
+
+ lsasd_reopen_logs(0);
+ lsasd_prefork_config();
+
+ lsasd_setup_sig_term_handler(ev_ctx);
+ lsasd_setup_sig_hup_handler(ev_ctx, pool, msg_ctx);
+
+ BlockSignals(false, SIGTERM);
+ BlockSignals(false, SIGHUP);
+
+ ok = lsasd_create_sockets(ev_ctx, msg_ctx, listen_fd, &listen_fd_size);
+ if (!ok) {
+ exit(1);
+ }
+
+ /* start children before any more initialization is done */
+ ok = prefork_create_pool(ev_ctx, /* mem_ctx */
+ ev_ctx,
+ msg_ctx,
+ listen_fd_size,
+ listen_fd,
+ lsasd_min_children,
+ lsasd_max_children,
+ &lsasd_children_main,
+ NULL,
+ &pool);
+ if (!ok) {
+ exit(1);
+ }
+
+ if (!serverid_register(procid_self(), FLAG_MSG_GENERAL)) {
+ exit(1);
+ }
+
+ messaging_register(msg_ctx,
+ ev_ctx,
+ MSG_SMB_CONF_UPDATED,
+ lsasd_smb_conf_updated);
+
+ status = rpc_lsarpc_init(NULL);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0, ("Failed to register winreg rpc inteface! (%s)\n",
+ nt_errstr(status)));
+ exit(1);
+ }
+
+ status = rpc_samr_init(NULL);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0, ("Failed to register lsasd rpc inteface! (%s)\n",
+ nt_errstr(status)));
+ exit(1);
+ }
+
+ status = rpc_netlogon_init(NULL);
+ if (!NT_STATUS_IS_OK(status)) {
+ DEBUG(0, ("Failed to register lsasd rpc inteface! (%s)\n",
+ nt_errstr(status)));
+ exit(1);
+ }
+
+ ok = lsasd_setup_children_monitor(ev_ctx, msg_ctx, pool);
+ if (!ok) {
+ DEBUG(0, ("Failed to setup children monitoring!\n"));
+ exit(1);
+ }
+
+ DEBUG(1, ("LSASD Daemon Started (%d)\n", getpid()));
+
+ /* loop forever */
+ rc = tevent_loop_wait(ev_ctx);
+
+ /* should not be reached */
+ DEBUG(0,("background_queue: tevent_loop_wait() exited with %d - %s\n",
+ rc, (rc == 0) ? "out of events" : strerror(errno)));
+ exit(1);
+}