/* Main routine of bgpd. * Copyright (C) 1996, 97, 98, 1999 Kunihiro Ishiguro * * This file is part of GNU Zebra. * * GNU Zebra 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, or (at your option) any * later version. * * GNU Zebra 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; see the file COPYING; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include #include #include "vector.h" #include "command.h" #include "getopt.h" #include "thread.h" #include #include "memory.h" #include "prefix.h" #include "log.h" #include "privs.h" #include "sigevent.h" #include "zclient.h" #include "routemap.h" #include "filter.h" #include "plist.h" #include "stream.h" #include "queue.h" #include "vrf.h" #include "bfd.h" #include "libfrr.h" #include "ns.h" #include "bgpd/bgpd.h" #include "bgpd/bgp_attr.h" #include "bgpd/bgp_route.h" #include "bgpd/bgp_mplsvpn.h" #include "bgpd/bgp_aspath.h" #include "bgpd/bgp_dump.h" #include "bgpd/bgp_route.h" #include "bgpd/bgp_nexthop.h" #include "bgpd/bgp_regex.h" #include "bgpd/bgp_clist.h" #include "bgpd/bgp_debug.h" #include "bgpd/bgp_errors.h" #include "bgpd/bgp_filter.h" #include "bgpd/bgp_zebra.h" #include "bgpd/bgp_packet.h" #include "bgpd/bgp_keepalives.h" #include "bgpd/bgp_network.h" #include "bgpd/bgp_errors.h" #include "bgpd/bgp_script.h" #include "lib/routing_nb.h" #include "bgpd/bgp_nb.h" #include "bgpd/bgp_evpn_mh.h" #include "bgpd/bgp_nht.h" #include "bgpd/bgp_routemap_nb.h" #include "bgpd/bgp_community_alias.h" #ifdef ENABLE_BGP_VNC #include "bgpd/rfapi/rfapi_backend.h" #endif /* bgpd options, we use GNU getopt library. */ static const struct option longopts[] = { {"bgp_port", required_argument, NULL, 'p'}, {"listenon", required_argument, NULL, 'l'}, {"no_kernel", no_argument, NULL, 'n'}, {"skip_runas", no_argument, NULL, 'S'}, {"ecmp", required_argument, NULL, 'e'}, {"int_num", required_argument, NULL, 'I'}, {"no_zebra", no_argument, NULL, 'Z'}, {"socket_size", required_argument, NULL, 's'}, {0}}; /* signal definitions */ void sighup(void); void sigint(void); void sigusr1(void); static void bgp_exit(int); static void bgp_vrf_terminate(void); static struct quagga_signal_t bgp_signals[] = { { .signal = SIGHUP, .handler = &sighup, }, { .signal = SIGUSR1, .handler = &sigusr1, }, { .signal = SIGINT, .handler = &sigint, }, { .signal = SIGTERM, .handler = &sigint, }, }; /* privileges */ static zebra_capabilities_t _caps_p[] = {ZCAP_BIND, ZCAP_NET_RAW, ZCAP_NET_ADMIN, ZCAP_SYS_ADMIN}; struct zebra_privs_t bgpd_privs = { #if defined(FRR_USER) && defined(FRR_GROUP) .user = FRR_USER, .group = FRR_GROUP, #endif #ifdef VTY_GROUP .vty_group = VTY_GROUP, #endif .caps_p = _caps_p, .cap_num_p = array_size(_caps_p), .cap_num_i = 0, }; static struct frr_daemon_info bgpd_di; /* SIGHUP handler. */ void sighup(void) { zlog_info("SIGHUP received, ignoring"); return; /* * This is turned off for the moment. There is all * sorts of config turned off by bgp_terminate * that is not setup properly again in bgp_reset. * I see no easy way to do this nor do I see that * this is a desirable way to reload config * given the yang work. */ /* Terminate all thread. */ /* * bgp_terminate(); * bgp_reset(); * zlog_info("bgpd restarting!"); * Reload config file. * vty_read_config(NULL, bgpd_di.config_file, config_default); */ /* Try to return to normal operation. */ } /* SIGINT handler. */ __attribute__((__noreturn__)) void sigint(void) { zlog_notice("Terminating on signal"); assert(bm->terminating == false); bm->terminating = true; /* global flag that shutting down */ /* Disable BFD events to avoid wasting processing. */ bfd_protocol_integration_set_shutdown(true); bgp_terminate(); bgp_exit(0); exit(0); } /* SIGUSR1 handler. */ void sigusr1(void) { zlog_rotate(); } /* Try to free up allocations we know about so that diagnostic tools such as valgrind are able to better illuminate leaks. Zebra route removal and protocol teardown are not meant to be done here. For example, "retain_mode" may be set. */ static __attribute__((__noreturn__)) void bgp_exit(int status) { struct bgp *bgp, *bgp_default, *bgp_evpn; struct listnode *node, *nnode; /* it only makes sense for this to be called on a clean exit */ assert(status == 0); frr_early_fini(); bgp_close(); bgp_default = bgp_get_default(); bgp_evpn = bgp_get_evpn(); /* reverse bgp_master_init */ for (ALL_LIST_ELEMENTS(bm->bgp, node, nnode, bgp)) { if (bgp_default == bgp || bgp_evpn == bgp) continue; bgp_delete(bgp); } if (bgp_evpn && bgp_evpn != bgp_default) bgp_delete(bgp_evpn); if (bgp_default) bgp_delete(bgp_default); bgp_evpn_mh_finish(); bgp_l3nhg_finish(); /* reverse bgp_dump_init */ bgp_dump_finish(); /* BGP community aliases */ bgp_community_alias_finish(); /* reverse bgp_route_init */ bgp_route_finish(); /* cleanup route maps */ bgp_route_map_terminate(); /* reverse bgp_attr_init */ bgp_attr_finish(); /* stop pthreads */ bgp_pthreads_finish(); /* reverse access_list_init */ access_list_add_hook(NULL); access_list_delete_hook(NULL); access_list_reset(); /* reverse bgp_filter_init */ as_list_add_hook(NULL); as_list_delete_hook(NULL); bgp_filter_reset(); /* reverse prefix_list_init */ prefix_list_add_hook(NULL); prefix_list_delete_hook(NULL); prefix_list_reset(); /* reverse community_list_init */ community_list_terminate(bgp_clist); bgp_vrf_terminate(); #ifdef ENABLE_BGP_VNC vnc_zebra_destroy(); #endif bgp_zebra_destroy(); bf_free(bm->rd_idspace); list_delete(&bm->bgp); list_delete(&bm->addresses); bgp_lp_finish(); memset(bm, 0, sizeof(*bm)); frr_fini(); exit(status); } static int bgp_vrf_new(struct vrf *vrf) { if (BGP_DEBUG(zebra, ZEBRA)) zlog_debug("VRF Created: %s(%u)", vrf->name, vrf->vrf_id); return 0; } static int bgp_vrf_delete(struct vrf *vrf) { if (BGP_DEBUG(zebra, ZEBRA)) zlog_debug("VRF Deletion: %s(%u)", vrf->name, vrf->vrf_id); return 0; } static int bgp_vrf_enable(struct vrf *vrf) { struct bgp *bgp; vrf_id_t old_vrf_id; if (BGP_DEBUG(zebra, ZEBRA)) zlog_debug("VRF enable add %s id %u", vrf->name, vrf->vrf_id); bgp = bgp_lookup_by_name(vrf->name); if (bgp && bgp->vrf_id != vrf->vrf_id) { if (bgp->name && strmatch(vrf->name, VRF_DEFAULT_NAME)) { XFREE(MTYPE_BGP, bgp->name); XFREE(MTYPE_BGP, bgp->name_pretty); bgp->name_pretty = XSTRDUP(MTYPE_BGP, "VRF default"); bgp->inst_type = BGP_INSTANCE_TYPE_DEFAULT; #ifdef ENABLE_BGP_VNC if (!bgp->rfapi) { bgp->rfapi = bgp_rfapi_new(bgp); assert(bgp->rfapi); assert(bgp->rfapi_cfg); } #endif /* ENABLE_BGP_VNC */ } old_vrf_id = bgp->vrf_id; /* We have instance configured, link to VRF and make it "up". */ bgp_vrf_link(bgp, vrf); bgp_handle_socket(bgp, vrf, old_vrf_id, true); bgp_instance_up(bgp); vpn_leak_zebra_vrf_label_update(bgp, AFI_IP); vpn_leak_zebra_vrf_label_update(bgp, AFI_IP6); vpn_leak_postchange(BGP_VPN_POLICY_DIR_TOVPN, AFI_IP, bgp_get_default(), bgp); vpn_leak_postchange(BGP_VPN_POLICY_DIR_FROMVPN, AFI_IP, bgp_get_default(), bgp); vpn_leak_postchange(BGP_VPN_POLICY_DIR_TOVPN, AFI_IP6, bgp_get_default(), bgp); vpn_leak_postchange(BGP_VPN_POLICY_DIR_FROMVPN, AFI_IP6, bgp_get_default(), bgp); } return 0; } static int bgp_vrf_disable(struct vrf *vrf) { struct bgp *bgp; if (vrf->vrf_id == VRF_DEFAULT) return 0; if (BGP_DEBUG(zebra, ZEBRA)) zlog_debug("VRF disable %s id %d", vrf->name, vrf->vrf_id); bgp = bgp_lookup_by_name(vrf->name); if (bgp) { vpn_leak_zebra_vrf_label_withdraw(bgp, AFI_IP); vpn_leak_zebra_vrf_label_withdraw(bgp, AFI_IP6); vpn_leak_prechange(BGP_VPN_POLICY_DIR_TOVPN, AFI_IP, bgp_get_default(), bgp); vpn_leak_prechange(BGP_VPN_POLICY_DIR_FROMVPN, AFI_IP, bgp_get_default(), bgp); vpn_leak_prechange(BGP_VPN_POLICY_DIR_TOVPN, AFI_IP6, bgp_get_default(), bgp); vpn_leak_prechange(BGP_VPN_POLICY_DIR_FROMVPN, AFI_IP6, bgp_get_default(), bgp); bgp_handle_socket(bgp, vrf, VRF_UNKNOWN, false); /* We have instance configured, unlink from VRF and make it * "down". */ bgp_instance_down(bgp); bgp_vrf_unlink(bgp, vrf); } /* Note: This is a callback, the VRF will be deleted by the caller. */ return 0; } static void bgp_vrf_init(void) { vrf_init(bgp_vrf_new, bgp_vrf_enable, bgp_vrf_disable, bgp_vrf_delete, bgp_vrf_enable); } static void bgp_vrf_terminate(void) { vrf_terminate(); } static const struct frr_yang_module_info *const bgpd_yang_modules[] = { &frr_bgp_info, &frr_filter_info, &frr_interface_info, &frr_route_map_info, &frr_routing_info, &frr_vrf_info, &frr_bgp_route_map_info, }; FRR_DAEMON_INFO(bgpd, BGP, .vty_port = BGP_VTY_PORT, .proghelp = "Implementation of the BGP routing protocol.", .signals = bgp_signals, .n_signals = array_size(bgp_signals), .privs = &bgpd_privs, .yang_modules = bgpd_yang_modules, .n_yang_modules = array_size(bgpd_yang_modules), ); #define DEPRECATED_OPTIONS "" /* Main routine of bgpd. Treatment of argument and start bgp finite state machine is handled at here. */ int main(int argc, char **argv) { int opt; int tmp_port; int bgp_port = BGP_PORT_DEFAULT; struct list *addresses = list_new(); int no_fib_flag = 0; int no_zebra_flag = 0; int skip_runas = 0; int instance = 0; int buffer_size = BGP_SOCKET_SNDBUF_SIZE; char *address; struct listnode *node; addresses->cmp = (int (*)(void *, void *))strcmp; frr_preinit(&bgpd_di, argc, argv); frr_opt_add( "p:l:SnZe:I:s:" DEPRECATED_OPTIONS, longopts, " -p, --bgp_port Set BGP listen port number (0 means do not listen).\n" " -l, --listenon Listen on specified address (implies -n)\n" " -n, --no_kernel Do not install route to kernel.\n" " -Z, --no_zebra Do not communicate with Zebra.\n" " -S, --skip_runas Skip capabilities checks, and changing user and group IDs.\n" " -e, --ecmp Specify ECMP to use.\n" " -I, --int_num Set instance number (label-manager)\n" " -s, --socket_size Set BGP peer socket send buffer size\n"); /* Command line argument treatment. */ while (1) { opt = frr_getopt(argc, argv, 0); if (opt && opt < 128 && strchr(DEPRECATED_OPTIONS, opt)) { fprintf(stderr, "The -%c option no longer exists.\nPlease refer to the manual.\n", opt); continue; } if (opt == EOF) break; switch (opt) { case 0: break; case 'p': tmp_port = atoi(optarg); if (tmp_port < 0 || tmp_port > 0xffff) bgp_port = BGP_PORT_DEFAULT; else bgp_port = tmp_port; break; case 'e': { unsigned long int parsed_multipath = strtoul(optarg, NULL, 10); if (parsed_multipath == 0 || parsed_multipath > MULTIPATH_NUM || parsed_multipath > UINT_MAX) { flog_err( EC_BGP_MULTIPATH, "Multipath Number specified must be less than %u and greater than 0", MULTIPATH_NUM); return 1; } multipath_num = parsed_multipath; break; } case 'l': listnode_add_sort_nodup(addresses, optarg); /* listenon implies -n */ /* fallthru */ case 'n': no_fib_flag = 1; break; case 'Z': no_zebra_flag = 1; break; case 'S': skip_runas = 1; break; case 'I': instance = atoi(optarg); if (instance > (unsigned short)-1) zlog_err("Instance %i out of range (0..%u)", instance, (unsigned short)-1); break; case 's': buffer_size = atoi(optarg); break; default: frr_help_exit(1); break; } } if (skip_runas) memset(&bgpd_privs, 0, sizeof(bgpd_privs)); /* BGP master init. */ bgp_master_init(frr_init(), buffer_size, addresses); bm->port = bgp_port; if (bgp_port == 0) bgp_option_set(BGP_OPT_NO_LISTEN); if (no_fib_flag || no_zebra_flag) bgp_option_set(BGP_OPT_NO_FIB); if (no_zebra_flag) bgp_option_set(BGP_OPT_NO_ZEBRA); bgp_error_init(); /* Initializations. */ bgp_vrf_init(); #ifdef HAVE_SCRIPTING bgp_script_init(); #endif hook_register(routing_conf_event, routing_control_plane_protocols_name_validate); /* BGP related initialization. */ bgp_init((unsigned short)instance); if (list_isempty(bm->addresses)) { snprintf(bgpd_di.startinfo, sizeof(bgpd_di.startinfo), ", bgp@:%d", bm->port); } else { for (ALL_LIST_ELEMENTS_RO(bm->addresses, node, address)) snprintf(bgpd_di.startinfo + strlen(bgpd_di.startinfo), sizeof(bgpd_di.startinfo) - strlen(bgpd_di.startinfo), ", bgp@%s:%d", address, bm->port); } frr_config_fork(); /* must be called after fork() */ bgp_gr_apply_running_config(); bgp_pthreads_run(); frr_run(bm->master); /* Not reached. */ return 0; }