// SPDX-License-Identifier: GPL-2.0-or-later /* * pim_bsm.c: PIM BSM handling routines * * Copyright (C) 2018-19 Vmware, Inc. * Saravanan K */ #ifdef HAVE_CONFIG_H #include "config.h" #endif #include #include #include #include #include #include #include #include #include #include #include "if.h" #include "pimd.h" #include "pim_iface.h" #include "pim_instance.h" #include "pim_neighbor.h" #include "pim_rpf.h" #include "pim_hello.h" #include "pim_pim.h" #include "pim_nht.h" #include "pim_bsm.h" #include "pim_time.h" #include "pim_zebra.h" #include "pim_util.h" #include "pim_sock.h" /* Functions forward declaration */ static void pim_bs_timer_start(struct bsm_scope *scope, int bs_timeout); static void pim_g2rp_timer_start(struct bsm_rpinfo *bsrp, int hold_time); static inline void pim_g2rp_timer_restart(struct bsm_rpinfo *bsrp, int hold_time); static void pim_bsm_accept_any(struct bsm_scope *scope); static void pim_cand_bsr_trigger(struct bsm_scope *scope, bool verbose); static void pim_cand_bsr_pending(struct bsm_scope *scope); /* Memory Types */ DEFINE_MTYPE_STATIC(PIMD, PIM_BSGRP_NODE, "PIM BSR advertised grp info"); DEFINE_MTYPE_STATIC(PIMD, PIM_BSRP_INFO, "PIM BSR advertised RP info"); DEFINE_MTYPE(PIMD, PIM_BSM_FRAG, "PIM BSM fragment"); DEFINE_MTYPE_STATIC(PIMD, PIM_BSM_PKT_VAR_MEM, "PIM BSM Packet"); DEFINE_MTYPE_STATIC(PIMD, PIM_CAND_RP_GRP, "PIM Candidate RP group"); static int cand_rp_group_cmp(const struct cand_rp_group *a, const struct cand_rp_group *b) { return prefix_cmp(&a->p, &b->p); } DECLARE_RBTREE_UNIQ(cand_rp_groups, struct cand_rp_group, item, cand_rp_group_cmp); /* All bsm packets forwarded shall be fit within ip mtu less iphdr(max) */ #define MAX_IP_HDR_LEN 24 /* pim_bsm_write_config - Write the interface pim bsm configuration.*/ void pim_bsm_write_config(struct vty *vty, struct interface *ifp) { struct pim_interface *pim_ifp = ifp->info; if (pim_ifp) { if (!pim_ifp->bsm_enable) vty_out(vty, " no " PIM_AF_NAME " pim bsm\n"); if (!pim_ifp->ucast_bsm_accept) vty_out(vty, " no " PIM_AF_NAME " pim unicast-bsm\n"); } } static void pim_bsm_rpinfo_free(struct bsm_rpinfo *bsrp_info) { EVENT_OFF(bsrp_info->g2rp_timer); XFREE(MTYPE_PIM_BSRP_INFO, bsrp_info); } static void pim_bsm_rpinfos_free(struct bsm_rpinfos_head *head) { struct bsm_rpinfo *bsrp_info; while ((bsrp_info = bsm_rpinfos_pop(head))) pim_bsm_rpinfo_free(bsrp_info); } static void pim_free_bsgrp_data(struct bsgrp_node *bsgrp_node) { pim_bsm_rpinfos_free(bsgrp_node->bsrp_list); pim_bsm_rpinfos_free(bsgrp_node->partial_bsrp_list); XFREE(MTYPE_PIM_BSGRP_NODE, bsgrp_node); } static void pim_free_bsgrp_node(struct route_table *rt, struct prefix *grp) { struct route_node *rn; rn = route_node_lookup(rt, grp); if (rn) { rn->info = NULL; route_unlock_node(rn); route_unlock_node(rn); } } static void pim_bsm_frag_free(struct bsm_frag *bsfrag) { XFREE(MTYPE_PIM_BSM_FRAG, bsfrag); } void pim_bsm_frags_free(struct bsm_scope *scope) { struct bsm_frag *bsfrag; while ((bsfrag = bsm_frags_pop(scope->bsm_frags))) pim_bsm_frag_free(bsfrag); } int pim_bsm_rpinfo_cmp(const struct bsm_rpinfo *node1, const struct bsm_rpinfo *node2) { /* RP election Algo : * Step-1 : Loweset Rp priority will have higher precedance. * Step-2 : If priority same then higher hash val will have * higher precedance. * Step-3 : If Hash val is same then highest rp address will * become elected RP. */ if (node1->rp_prio < node2->rp_prio) return -1; if (node1->rp_prio > node2->rp_prio) return 1; if (node1->hash < node2->hash) return 1; if (node1->hash > node2->hash) return -1; return pim_addr_cmp(node2->rp_address, node1->rp_address); } static struct bsgrp_node *pim_bsm_new_bsgrp_node(struct route_table *rt, struct prefix *grp) { struct route_node *rn; struct bsgrp_node *bsgrp; rn = route_node_get(rt, grp); if (!rn) { zlog_warn("%s: route node creation failed", __func__); return NULL; } bsgrp = XCALLOC(MTYPE_PIM_BSGRP_NODE, sizeof(struct bsgrp_node)); rn->info = bsgrp; bsm_rpinfos_init(bsgrp->bsrp_list); bsm_rpinfos_init(bsgrp->partial_bsrp_list); prefix_copy(&bsgrp->group, grp); return bsgrp; } /* BS timer for NO_INFO, ACCEPT_ANY & ACCEPT_PREFERRED. * Candidate BSR handling is separate further below */ static void pim_on_bs_timer(struct event *t) { struct bsm_scope *scope; scope = EVENT_ARG(t); EVENT_OFF(scope->bs_timer); if (PIM_DEBUG_BSM) zlog_debug("%s: Bootstrap Timer expired for scope: %d", __func__, scope->sz_id); assertf(scope->state <= ACCEPT_PREFERRED, "state=%d", scope->state); pim_nht_bsr_del(scope->pim, scope->current_bsr); pim_bsm_accept_any(scope); } static void pim_bsm_accept_any(struct bsm_scope *scope) { struct route_node *rn; struct bsgrp_node *bsgrp_node; struct bsm_rpinfo *bsrp; EVENT_OFF(scope->t_ebsr_regen_bsm); /* Reset scope zone data */ scope->state = ACCEPT_ANY; scope->current_bsr = PIMADDR_ANY; scope->current_bsr_prio = 0; scope->current_bsr_first_ts = 0; scope->current_bsr_last_ts = 0; scope->bsm_frag_tag = 0; pim_bsm_frags_free(scope); for (rn = route_top(scope->bsrp_table); rn; rn = route_next(rn)) { bsgrp_node = (struct bsgrp_node *)rn->info; if (!bsgrp_node) { if (PIM_DEBUG_BSM) zlog_debug("%s: bsgrp_node is null", __func__); continue; } /* Give grace time for rp to continue for another hold time */ bsrp = bsm_rpinfos_first(bsgrp_node->bsrp_list); if (bsrp) pim_g2rp_timer_restart(bsrp, bsrp->rp_holdtime); /* clear pending list */ pim_bsm_rpinfos_free(bsgrp_node->partial_bsrp_list); bsgrp_node->pend_rp_cnt = 0; } /* we're leaving ACCEPT_PREFERRED, which doubles as C-BSR if we're * configured to be a Candidate BSR. See if we're P-BSR now. */ pim_cand_bsr_trigger(scope, false); } static void pim_bs_timer_stop(struct bsm_scope *scope) { if (PIM_DEBUG_BSM) zlog_debug("%s : BS timer being stopped of sz: %d", __func__, scope->sz_id); EVENT_OFF(scope->bs_timer); } static void pim_bs_timer_start(struct bsm_scope *scope, int bs_timeout) { if (!scope) { if (PIM_DEBUG_BSM) zlog_debug("%s : Invalid scope(NULL).", __func__); return; } EVENT_OFF(scope->bs_timer); if (PIM_DEBUG_BSM) zlog_debug( "%s : starting bs timer for scope %d with timeout %d secs", __func__, scope->sz_id, bs_timeout); event_add_timer(router->master, pim_on_bs_timer, scope, bs_timeout, &scope->bs_timer); } static inline void pim_bs_timer_restart(struct bsm_scope *scope, int bs_timeout) { pim_bs_timer_start(scope, bs_timeout); } static void bsm_unicast_sock_read(struct event *t) { struct bsm_scope *scope = EVENT_ARG(t); pim_sock_read_helper(scope->unicast_sock, scope->pim, false); event_add_read(router->master, bsm_unicast_sock_read, scope, scope->unicast_sock, &scope->unicast_read); } void pim_bsm_proc_init(struct pim_instance *pim) { struct bsm_scope *scope = &pim->global_scope; memset(scope, 0, sizeof(*scope)); scope->sz_id = PIM_GBL_SZ_ID; scope->bsrp_table = route_table_init(); scope->accept_nofwd_bsm = true; scope->state = NO_INFO; scope->pim = pim; bsm_frags_init(scope->bsm_frags); pim_bs_timer_start(scope, PIM_BS_TIME); scope->cand_rp_interval = PIM_CRP_ADV_INTERVAL; cand_rp_groups_init(scope->cand_rp_groups); scope->unicast_sock = pim_socket_raw(IPPROTO_PIM); set_nonblocking(scope->unicast_sock); sockopt_reuseaddr(scope->unicast_sock); if (setsockopt_ifindex(PIM_AF, scope->unicast_sock, 1) == -1) zlog_warn("%s: Without IP_PKTINFO, src interface can't be determined", __func__); pim_socket_ip_hdr(scope->unicast_sock); frr_with_privs (&pimd_privs) { vrf_bind(pim->vrf->vrf_id, scope->unicast_sock, NULL); } event_add_read(router->master, bsm_unicast_sock_read, scope, scope->unicast_sock, &scope->unicast_read); } void pim_bsm_proc_free(struct pim_instance *pim) { struct bsm_scope *scope = &pim->global_scope; struct route_node *rn; struct bsgrp_node *bsgrp; struct cand_rp_group *crpgrp; EVENT_OFF(scope->unicast_read); close(scope->unicast_sock); pim_bs_timer_stop(scope); pim_bsm_frags_free(scope); for (rn = route_top(scope->bsrp_table); rn; rn = route_next(rn)) { bsgrp = rn->info; if (!bsgrp) continue; pim_free_bsgrp_data(bsgrp); } while ((crpgrp = cand_rp_groups_pop(scope->cand_rp_groups))) XFREE(MTYPE_PIM_CAND_RP_GRP, crpgrp); cand_rp_groups_fini(scope->cand_rp_groups); route_table_finish(scope->bsrp_table); } static bool is_hold_time_elapsed(void *data) { struct bsm_rpinfo *bsrp; bsrp = data; if (bsrp->elapse_time < bsrp->rp_holdtime) return false; else return true; } static void pim_on_g2rp_timer(struct event *t) { struct bsm_rpinfo *bsrp; struct bsm_rpinfo *bsrp_node; struct bsgrp_node *bsgrp_node; struct pim_instance *pim; struct rp_info *rp_info; struct route_node *rn; uint16_t elapse; pim_addr bsrp_addr; bsrp = EVENT_ARG(t); EVENT_OFF(bsrp->g2rp_timer); bsgrp_node = bsrp->bsgrp_node; /* elapse time is the hold time of expired node */ elapse = bsrp->rp_holdtime; bsrp_addr = bsrp->rp_address; /* update elapse for all bsrp nodes */ frr_each_safe (bsm_rpinfos, bsgrp_node->bsrp_list, bsrp_node) { bsrp_node->elapse_time += elapse; if (is_hold_time_elapsed(bsrp_node)) { bsm_rpinfos_del(bsgrp_node->bsrp_list, bsrp_node); pim_bsm_rpinfo_free(bsrp_node); } } /* Get the next elected rp node */ bsrp = bsm_rpinfos_first(bsgrp_node->bsrp_list); pim = bsgrp_node->scope->pim; rn = route_node_lookup(pim->rp_table, &bsgrp_node->group); if (!rn) { zlog_warn("%s: Route node doesn't exist", __func__); return; } rp_info = (struct rp_info *)rn->info; if (!rp_info) { route_unlock_node(rn); return; } if (rp_info->rp_src != RP_SRC_STATIC) { /* If new rp available, change it else delete the existing */ if (bsrp) { pim_g2rp_timer_start( bsrp, (bsrp->rp_holdtime - bsrp->elapse_time)); pim_rp_change(pim, bsrp->rp_address, bsgrp_node->group, RP_SRC_BSR); } else { pim_rp_del(pim, bsrp_addr, bsgrp_node->group, NULL, RP_SRC_BSR); } } if (!bsm_rpinfos_count(bsgrp_node->bsrp_list) && !bsm_rpinfos_count(bsgrp_node->partial_bsrp_list)) { pim_free_bsgrp_node(pim->global_scope.bsrp_table, &bsgrp_node->group); pim_free_bsgrp_data(bsgrp_node); } } static void pim_g2rp_timer_start(struct bsm_rpinfo *bsrp, int hold_time) { if (!bsrp) { if (PIM_DEBUG_BSM) zlog_debug("%s : Invalid brsp(NULL).", __func__); return; } EVENT_OFF(bsrp->g2rp_timer); if (PIM_DEBUG_BSM) zlog_debug( "%s : starting g2rp timer for grp: %pFX - rp: %pPAs with timeout %d secs(Actual Hold time : %d secs)", __func__, &bsrp->bsgrp_node->group, &bsrp->rp_address, hold_time, bsrp->rp_holdtime); event_add_timer(router->master, pim_on_g2rp_timer, bsrp, hold_time, &bsrp->g2rp_timer); } static inline void pim_g2rp_timer_restart(struct bsm_rpinfo *bsrp, int hold_time) { pim_g2rp_timer_start(bsrp, hold_time); } static void pim_g2rp_timer_stop(struct bsm_rpinfo *bsrp) { if (!bsrp) return; if (PIM_DEBUG_BSM) zlog_debug("%s : stopping g2rp timer for grp: %pFX - rp: %pPAs", __func__, &bsrp->bsgrp_node->group, &bsrp->rp_address); EVENT_OFF(bsrp->g2rp_timer); } static bool is_hold_time_zero(void *data) { struct bsm_rpinfo *bsrp; bsrp = data; if (bsrp->rp_holdtime) return false; else return true; } static void pim_instate_pend_list(struct bsgrp_node *bsgrp_node) { struct bsm_rpinfo *active; struct bsm_rpinfo *pend; struct rp_info *rp_info; struct route_node *rn; struct pim_instance *pim; struct rp_info *rp_all; struct prefix group_all; bool had_rp_node = true; pim = bsgrp_node->scope->pim; active = bsm_rpinfos_first(bsgrp_node->bsrp_list); /* Remove nodes with hold time 0 & check if list still has a head */ frr_each_safe (bsm_rpinfos, bsgrp_node->partial_bsrp_list, pend) { if (is_hold_time_zero(pend)) { bsm_rpinfos_del(bsgrp_node->partial_bsrp_list, pend); pim_bsm_rpinfo_free(pend); } } pend = bsm_rpinfos_first(bsgrp_node->partial_bsrp_list); if (!pim_get_all_mcast_group(&group_all)) return; rp_all = pim_rp_find_match_group(pim, &group_all); rn = route_node_lookup(pim->rp_table, &bsgrp_node->group); if (pend) pim_g2rp_timer_start(pend, pend->rp_holdtime); /* if rp node doesn't exist or exist but not configured(rp_all), * install the rp from head(if exists) of partial list. List is * is sorted such that head is the elected RP for the group. */ if (!rn || (prefix_same(&rp_all->group, &bsgrp_node->group) && pim_rpf_addr_is_inaddr_any(&rp_all->rp))) { if (PIM_DEBUG_BSM) zlog_debug("%s: Route node doesn't exist", __func__); if (pend) pim_rp_new(pim, pend->rp_address, bsgrp_node->group, NULL, RP_SRC_BSR); had_rp_node = false; } else { rp_info = (struct rp_info *)rn->info; if (!rp_info) { route_unlock_node(rn); if (pend) pim_rp_new(pim, pend->rp_address, bsgrp_node->group, NULL, RP_SRC_BSR); had_rp_node = false; } } /* We didn't have rp node and pending list is empty(unlikely), cleanup*/ if ((!had_rp_node) && (!pend)) { pim_free_bsgrp_node(bsgrp_node->scope->bsrp_table, &bsgrp_node->group); pim_free_bsgrp_data(bsgrp_node); return; } if ((had_rp_node) && (rp_info->rp_src != RP_SRC_STATIC)) { /* This means we searched and got rp node, needs unlock */ route_unlock_node(rn); if (active && pend) { if (pim_addr_cmp(active->rp_address, pend->rp_address)) pim_rp_change(pim, pend->rp_address, bsgrp_node->group, RP_SRC_BSR); } /* Possible when the first BSM has group with 0 rp count */ if ((!active) && (!pend)) { if (PIM_DEBUG_BSM) { zlog_debug( "%s: Both bsrp and partial list are empty", __func__); } pim_free_bsgrp_node(bsgrp_node->scope->bsrp_table, &bsgrp_node->group); pim_free_bsgrp_data(bsgrp_node); return; } /* Possible when a group with 0 rp count received in BSM */ if ((active) && (!pend)) { pim_rp_del(pim, active->rp_address, bsgrp_node->group, NULL, RP_SRC_BSR); pim_free_bsgrp_node(bsgrp_node->scope->bsrp_table, &bsgrp_node->group); if (PIM_DEBUG_BSM) { zlog_debug("%s:Pend List is null,del grp node", __func__); } pim_free_bsgrp_data(bsgrp_node); return; } } if ((had_rp_node) && (rp_info->rp_src == RP_SRC_STATIC)) { /* We need to unlock rn this case */ route_unlock_node(rn); /* there is a chance that static rp exist and bsrp cleaned * so clean bsgrp node if pending list empty */ if (!pend) { if (PIM_DEBUG_BSM) zlog_debug( "%s: Partial list is empty, static rp exists", __func__); pim_free_bsgrp_node(bsgrp_node->scope->bsrp_table, &bsgrp_node->group); pim_free_bsgrp_data(bsgrp_node); return; } } /* swap the list & delete all nodes in partial list (old bsrp_list) * before swap * active is head of bsrp list * pend is head of partial list * After swap * active is head of partial list * pend is head of bsrp list * So check appriate head after swap and clean the new partial list */ bsm_rpinfos_swap_all(bsgrp_node->bsrp_list, bsgrp_node->partial_bsrp_list); if (active) pim_g2rp_timer_stop(active); pim_bsm_rpinfos_free(bsgrp_node->partial_bsrp_list); } static bool is_preferred_bsr(struct pim_instance *pim, pim_addr bsr, uint32_t bsr_prio) { if (bsr_prio > pim->global_scope.current_bsr_prio) return true; else if (bsr_prio == pim->global_scope.current_bsr_prio) { if (pim_addr_cmp(bsr, pim->global_scope.current_bsr) >= 0) return true; else return false; } else if (!pim_addr_cmp(bsr, pim->global_scope.current_bsr)) { /* BSR config changed, lower prio now. local BSR check * is handled separately in pim_bsm_update() */ return true; } else return false; } static void pim_bsm_update(struct pim_instance *pim, pim_addr bsr, uint32_t bsr_prio) { pim->global_scope.current_bsr_prio = bsr_prio; pim->global_scope.current_bsr_last_ts = pim_time_monotonic_sec(); if (pim->global_scope.bsr_addrsel.run && pim->global_scope.cand_bsr_prio > bsr_prio && pim->global_scope.state < BSR_PENDING) { /* current BSR is now less preferred than ourselves */ pim_cand_bsr_pending(&pim->global_scope); return; } if (!pim_addr_cmp(bsr, pim->global_scope.current_bsr)) return; switch (pim->global_scope.state) { case BSR_PENDING: if (PIM_DEBUG_BSM) zlog_debug("Candidate BSR dropping out of BSR election, better BSR (%u, %pPA)", bsr_prio, &bsr); break; case BSR_ELECTED: if (PIM_DEBUG_BSM) zlog_debug("Lost BSR status, better BSR (%u, %pPA)", bsr_prio, &bsr); break; case NO_INFO: case ACCEPT_ANY: case ACCEPT_PREFERRED: break; } EVENT_OFF(pim->global_scope.t_ebsr_regen_bsm); if (pim->global_scope.state == BSR_ELECTED) pim_crp_db_clear(&pim->global_scope); else pim_nht_bsr_del(pim, pim->global_scope.current_bsr); pim_nht_bsr_add(pim, bsr); pim->global_scope.current_bsr = bsr; pim->global_scope.current_bsr_first_ts = pim_time_monotonic_sec(); pim->global_scope.state = ACCEPT_PREFERRED; pim_cand_rp_trigger(&pim->global_scope); } void pim_bsm_clear(struct pim_instance *pim) { struct route_node *rn; struct route_node *rpnode; struct bsgrp_node *bsgrp; pim_addr nht_p; struct prefix g_all; struct rp_info *rp_all; struct pim_upstream *up; struct rp_info *rp_info; bool upstream_updated = false; EVENT_OFF(pim->global_scope.t_ebsr_regen_bsm); if (pim->global_scope.state == BSR_ELECTED) pim_crp_db_clear(&pim->global_scope); else pim_nht_bsr_del(pim, pim->global_scope.current_bsr); /* Reset scope zone data */ pim->global_scope.accept_nofwd_bsm = false; pim->global_scope.state = ACCEPT_ANY; pim->global_scope.current_bsr = PIMADDR_ANY; pim->global_scope.current_bsr_prio = 0; pim->global_scope.current_bsr_first_ts = 0; pim->global_scope.current_bsr_last_ts = 0; pim->global_scope.bsm_frag_tag = 0; pim_bsm_frags_free(&pim->global_scope); pim_bs_timer_stop(&pim->global_scope); for (rn = route_top(pim->global_scope.bsrp_table); rn; rn = route_next(rn)) { bsgrp = rn->info; if (!bsgrp) continue; rpnode = route_node_lookup(pim->rp_table, &bsgrp->group); if (!rpnode) { pim_free_bsgrp_node(bsgrp->scope->bsrp_table, &bsgrp->group); pim_free_bsgrp_data(bsgrp); continue; } rp_info = (struct rp_info *)rpnode->info; if ((!rp_info) || (rp_info->rp_src != RP_SRC_BSR)) { pim_free_bsgrp_node(bsgrp->scope->bsrp_table, &bsgrp->group); pim_free_bsgrp_data(bsgrp); continue; } /* Deregister addr with Zebra NHT */ nht_p = rp_info->rp.rpf_addr; if (PIM_DEBUG_PIM_NHT_RP) { zlog_debug("%s: Deregister RP addr %pPA with Zebra ", __func__, &nht_p); } pim_delete_tracked_nexthop(pim, nht_p, NULL, rp_info); if (!pim_get_all_mcast_group(&g_all)) return; rp_all = pim_rp_find_match_group(pim, &g_all); if (rp_all == rp_info) { rp_all->rp.rpf_addr = PIMADDR_ANY; rp_all->i_am_rp = 0; } else { /* Delete the rp_info from rp-list */ listnode_delete(pim->rp_list, rp_info); /* Delete the rp node from rp_table */ rpnode->info = NULL; route_unlock_node(rpnode); route_unlock_node(rpnode); XFREE(MTYPE_PIM_RP, rp_info); } pim_free_bsgrp_node(bsgrp->scope->bsrp_table, &bsgrp->group); pim_free_bsgrp_data(bsgrp); } pim_rp_refresh_group_to_rp_mapping(pim); frr_each (rb_pim_upstream, &pim->upstream_head, up) { /* Find the upstream (*, G) whose upstream address is same as * the RP */ if (!pim_addr_is_any(up->sg.src)) continue; struct prefix grp; struct rp_info *trp_info; pim_addr_to_prefix(&grp, up->sg.grp); trp_info = pim_rp_find_match_group(pim, &grp); /* RP not found for the group grp */ if (pim_rpf_addr_is_inaddr_any(&trp_info->rp)) { pim_upstream_rpf_clear(pim, up); pim_rp_set_upstream_addr(pim, &up->upstream_addr, up->sg.src, up->sg.grp); } else { /* RP found for the group grp */ pim_upstream_update(pim, up); upstream_updated = true; } } if (upstream_updated) pim_zebra_update_all_interfaces(pim); } static bool pim_bsm_send_intf(uint8_t *buf, int len, struct interface *ifp, pim_addr dst_addr) { struct pim_interface *pim_ifp; pim_ifp = ifp->info; if (!pim_ifp) { if (PIM_DEBUG_BSM) zlog_debug("%s: Pim interface not available for %s", __func__, ifp->name); return false; } if (pim_ifp->pim_sock_fd == -1) { if (PIM_DEBUG_BSM) zlog_debug("%s: Pim sock not available for %s", __func__, ifp->name); return false; } if (pim_msg_send(pim_ifp->pim_sock_fd, pim_ifp->primary_address, dst_addr, buf, len, ifp)) { zlog_warn("%s: Could not send BSM message on interface: %s", __func__, ifp->name); return false; } if (!pim_ifp->pim_passive_enable) pim_ifp->pim_ifstat_bsm_tx++; pim_ifp->pim->bsm_sent++; return true; } static bool pim_bsm_frag_send(uint8_t *buf, uint32_t len, struct interface *ifp, uint32_t pim_mtu, pim_addr dst_addr, bool no_fwd) { struct pim_interface *pim_ifp = ifp->info; struct bsmmsg_grpinfo *grpinfo, *curgrp; uint8_t *firstgrp_ptr; uint8_t *pkt; uint8_t *pak_start; uint32_t parsed_len = 0; uint32_t this_pkt_rem; uint32_t copy_byte_count; uint32_t this_pkt_len; uint8_t total_rp_cnt; uint8_t this_rp_cnt; uint8_t frag_rp_cnt; uint8_t rp_fit_cnt; bool pak_pending = false; /* MTU passed here is PIM MTU (IP MTU less IP Hdr) */ if (pim_mtu < (PIM_MIN_BSM_LEN)) { zlog_warn( "%s: mtu(pim mtu: %d) size less than minimum bootstrap len", __func__, pim_mtu); if (PIM_DEBUG_BSM) zlog_debug( "%s: mtu (pim mtu:%d) less than minimum bootstrap len", __func__, pim_mtu); return false; } pak_start = XCALLOC(MTYPE_PIM_BSM_PKT_VAR_MEM, pim_mtu); pkt = pak_start; /* Fill PIM header later before sending packet to calc checksum */ pkt += PIM_MSG_HEADER_LEN; buf += PIM_MSG_HEADER_LEN; /* copy bsm header to new packet at offset of pim hdr */ memcpy(pkt, buf, PIM_BSM_HDR_LEN); pkt += PIM_BSM_HDR_LEN; buf += PIM_BSM_HDR_LEN; parsed_len += (PIM_MSG_HEADER_LEN + PIM_BSM_HDR_LEN); /* Store the position of first grp ptr, which can be reused for * next packet to start filling group. old bsm header and pim hdr * remains. So need not be filled again for next packet onwards. */ firstgrp_ptr = pkt; /* we received mtu excluding IP hdr len as param * now this_pkt_rem is mtu excluding * PIM_BSM_HDR_LEN + PIM_MSG_HEADER_LEN */ this_pkt_rem = pim_mtu - (PIM_BSM_HDR_LEN + PIM_MSG_HEADER_LEN); /* For each group till the packet length parsed */ while (parsed_len < len) { /* pkt ---> fragment's current pointer * buf ---> input buffer's current pointer * mtu ---> size of the pim packet - PIM header * curgrp ---> current group on the fragment * grpinfo ---> current group on the input buffer * this_pkt_rem ---> bytes remaing on the current fragment * rp_fit_cnt ---> num of rp for current grp that * fits this frag * total_rp_cnt ---> total rp present for the group in the buf * frag_rp_cnt ---> no of rp for the group to be fit in * the frag * this_rp_cnt ---> how many rp have we parsed */ grpinfo = (struct bsmmsg_grpinfo *)buf; memcpy(pkt, buf, PIM_BSM_GRP_LEN); curgrp = (struct bsmmsg_grpinfo *)pkt; parsed_len += PIM_BSM_GRP_LEN; pkt += PIM_BSM_GRP_LEN; buf += PIM_BSM_GRP_LEN; this_pkt_rem -= PIM_BSM_GRP_LEN; /* initialize rp count and total_rp_cnt before the rp loop */ this_rp_cnt = 0; total_rp_cnt = grpinfo->frag_rp_count; /* Loop till all RPs for the group parsed */ while (this_rp_cnt < total_rp_cnt) { /* All RP from a group processed here. * group is pointed by grpinfo. * At this point make sure buf pointing to a RP * within a group */ rp_fit_cnt = this_pkt_rem / PIM_BSM_RP_LEN; /* calculate how many rp am i going to copy in * this frag */ if (rp_fit_cnt > (total_rp_cnt - this_rp_cnt)) frag_rp_cnt = total_rp_cnt - this_rp_cnt; else frag_rp_cnt = rp_fit_cnt; /* populate the frag rp count for the current grp */ curgrp->frag_rp_count = frag_rp_cnt; copy_byte_count = frag_rp_cnt * PIM_BSM_RP_LEN; /* copy all the rp that we are fitting in this * frag for the grp */ memcpy(pkt, buf, copy_byte_count); this_rp_cnt += frag_rp_cnt; buf += copy_byte_count; pkt += copy_byte_count; parsed_len += copy_byte_count; this_pkt_rem -= copy_byte_count; /* Either we couldn't fit all rp for the group or the * mtu reached */ if ((this_rp_cnt < total_rp_cnt) || (this_pkt_rem < (PIM_BSM_GRP_LEN + PIM_BSM_RP_LEN))) { /* No space to fit in more rp, send this pkt */ this_pkt_len = pim_mtu - this_pkt_rem; pim_msg_build_header( pim_ifp->primary_address, dst_addr, pak_start, this_pkt_len, PIM_MSG_TYPE_BOOTSTRAP, no_fwd); pim_bsm_send_intf(pak_start, this_pkt_len, ifp, dst_addr); /* Construct next fragment. Reuse old packet */ pkt = firstgrp_ptr; this_pkt_rem = pim_mtu - (PIM_BSM_HDR_LEN + PIM_MSG_HEADER_LEN); /* If pkt can't accommodate next group + at * least one rp, we must break out of this inner * loop and process next RP */ if (total_rp_cnt == this_rp_cnt) break; /* If some more RPs for the same group pending, * fill grp hdr */ memcpy(pkt, (uint8_t *)grpinfo, PIM_BSM_GRP_LEN); curgrp = (struct bsmmsg_grpinfo *)pkt; pkt += PIM_BSM_GRP_LEN; this_pkt_rem -= PIM_BSM_GRP_LEN; pak_pending = false; } else { /* We filled something but not yet sent out */ pak_pending = true; } } /* while RP count */ } /*while parsed len */ /* Send if we have any unsent packet */ if (pak_pending) { this_pkt_len = pim_mtu - this_pkt_rem; pim_msg_build_header(pim_ifp->primary_address, dst_addr, pak_start, this_pkt_len, PIM_MSG_TYPE_BOOTSTRAP, no_fwd); pim_bsm_send_intf(pak_start, (pim_mtu - this_pkt_rem), ifp, dst_addr); } XFREE(MTYPE_PIM_BSM_PKT_VAR_MEM, pak_start); return true; } static void pim_bsm_fwd_whole_sz(struct pim_instance *pim, uint8_t *buf, uint32_t len, int sz) { struct interface *ifp; struct pim_interface *pim_ifp; pim_addr dst_addr; uint32_t pim_mtu; bool no_fwd = false; bool ret = false; /* For now only global scope zone is supported, so send on all * pim interfaces in the vrf */ dst_addr = qpim_all_pim_routers_addr; FOR_ALL_INTERFACES (pim->vrf, ifp) { pim_ifp = ifp->info; if ((!pim_ifp) || (!pim_ifp->bsm_enable)) continue; /* * RFC 5059 Sec 3.4: * When a Bootstrap message is forwarded, it is forwarded out * of every multicast-capable interface that has PIM neighbors. * * So skipping pim interfaces with no neighbors. */ if (listcount(pim_ifp->pim_neighbor_list) == 0) continue; pim_hello_require(ifp); pim_mtu = ifp->mtu - MAX_IP_HDR_LEN; if (pim_mtu < len) { ret = pim_bsm_frag_send(buf, len, ifp, pim_mtu, dst_addr, no_fwd); if (PIM_DEBUG_BSM) zlog_debug("%s: pim_bsm_frag_send returned %s", __func__, ret ? "TRUE" : "FALSE"); } else { pim_msg_build_header(pim_ifp->primary_address, dst_addr, buf, len, PIM_MSG_TYPE_BOOTSTRAP, no_fwd); if (!pim_bsm_send_intf(buf, len, ifp, dst_addr)) { if (PIM_DEBUG_BSM) zlog_debug( "%s: pim_bsm_send_intf returned false", __func__); } } } } bool pim_bsm_new_nbr_fwd(struct pim_neighbor *neigh, struct interface *ifp) { pim_addr dst_addr; struct pim_interface *pim_ifp; struct bsm_scope *scope; struct bsm_frag *bsfrag; uint32_t pim_mtu; bool no_fwd = true; bool ret = false; if (PIM_DEBUG_BSM) zlog_debug("%s: New neighbor %pPA seen on %s", __func__, &neigh->source_addr, ifp->name); pim_ifp = ifp->info; /* DR only forwards BSM packet */ if (!pim_addr_cmp(pim_ifp->pim_dr_addr, pim_ifp->primary_address)) { if (PIM_DEBUG_BSM) zlog_debug( "%s: It is not DR, so don't forward BSM packet", __func__); } if (!pim_ifp->bsm_enable) { if (PIM_DEBUG_BSM) zlog_debug("%s: BSM proc not enabled on %s", __func__, ifp->name); return ret; } scope = &pim_ifp->pim->global_scope; if (!bsm_frags_count(scope->bsm_frags)) { if (PIM_DEBUG_BSM) zlog_debug("%s: BSM list for the scope is empty", __func__); return ret; } if (!pim_ifp->ucast_bsm_accept) { dst_addr = qpim_all_pim_routers_addr; if (PIM_DEBUG_BSM) zlog_debug("%s: Sending BSM mcast to %pPA", __func__, &neigh->source_addr); } else { dst_addr = neigh->source_addr; if (PIM_DEBUG_BSM) zlog_debug("%s: Sending BSM ucast to %pPA", __func__, &neigh->source_addr); } pim_mtu = ifp->mtu - MAX_IP_HDR_LEN; pim_hello_require(ifp); frr_each (bsm_frags, scope->bsm_frags, bsfrag) { if (pim_mtu < bsfrag->size) { ret = pim_bsm_frag_send(bsfrag->data, bsfrag->size, ifp, pim_mtu, dst_addr, no_fwd); if (!ret) { if (PIM_DEBUG_BSM) zlog_debug( "%s: pim_bsm_frag_send failed", __func__); } } else { /* Pim header needs to be constructed */ pim_msg_build_header(pim_ifp->primary_address, dst_addr, bsfrag->data, bsfrag->size, PIM_MSG_TYPE_BOOTSTRAP, no_fwd); ret = pim_bsm_send_intf(bsfrag->data, bsfrag->size, ifp, dst_addr); if (!ret) { if (PIM_DEBUG_BSM) zlog_debug( "%s: pim_bsm_frag_send failed", __func__); } } } return ret; } struct bsgrp_node *pim_bsm_get_bsgrp_node(struct bsm_scope *scope, struct prefix *grp) { struct route_node *rn; struct bsgrp_node *bsgrp; rn = route_node_lookup(scope->bsrp_table, grp); if (!rn) { if (PIM_DEBUG_BSM) zlog_debug("%s: Route node doesn't exist for the group", __func__); return NULL; } bsgrp = rn->info; route_unlock_node(rn); return bsgrp; } static uint32_t hash_calc_on_grp_rp(struct prefix group, pim_addr rp, uint8_t hashmasklen) { uint64_t temp; uint32_t hash; uint32_t grpaddr; uint32_t rp_add; uint32_t mask = 0xffffffff; /* mask to be made zero if hashmasklen is 0 because mask << 32 * may not give 0. hashmasklen can be 0 to 32. */ if (hashmasklen == 0) mask = 0; /* in_addr stores ip in big endian, hence network byte order * convert to uint32 before processing hash */ #if PIM_IPV == 4 grpaddr = ntohl(group.u.prefix4.s_addr); #else grpaddr = group.u.prefix6.s6_addr32[0] ^ group.u.prefix6.s6_addr32[1] ^ group.u.prefix6.s6_addr32[2] ^ group.u.prefix6.s6_addr32[3]; #endif /* Avoid shifting by 32 bit on a 32 bit register */ if (hashmasklen) grpaddr = grpaddr & ((mask << (32 - hashmasklen))); else grpaddr = grpaddr & mask; #if PIM_IPV == 4 rp_add = ntohl(rp.s_addr); #else rp_add = rp.s6_addr32[0] ^ rp.s6_addr32[1] ^ rp.s6_addr32[2] ^ rp.s6_addr32[3]; #endif temp = 1103515245 * ((1103515245 * (uint64_t)grpaddr + 12345) ^ rp_add) + 12345; hash = temp & (0x7fffffff); return hash; } static bool pim_install_bsm_grp_rp(struct pim_instance *pim, struct bsgrp_node *grpnode, struct bsmmsg_rpinfo *rp) { struct bsm_rpinfo *bsm_rpinfo; uint8_t hashMask_len = pim->global_scope.hashMasklen; /*memory allocation for bsm_rpinfo */ bsm_rpinfo = XCALLOC(MTYPE_PIM_BSRP_INFO, sizeof(*bsm_rpinfo)); bsm_rpinfo->rp_prio = rp->rp_pri; bsm_rpinfo->rp_holdtime = rp->rp_holdtime; bsm_rpinfo->rp_address = rp->rpaddr.addr; bsm_rpinfo->elapse_time = 0; /* Back pointer to the group node. */ bsm_rpinfo->bsgrp_node = grpnode; /* update hash for this rp node */ bsm_rpinfo->hash = hash_calc_on_grp_rp(grpnode->group, rp->rpaddr.addr, hashMask_len); if (bsm_rpinfos_add(grpnode->partial_bsrp_list, bsm_rpinfo) == NULL) { if (PIM_DEBUG_BSM) zlog_debug( "%s, bs_rpinfo node added to the partial bs_rplist.", __func__); return true; } if (PIM_DEBUG_BSM) zlog_debug("%s: list node not added", __func__); XFREE(MTYPE_PIM_BSRP_INFO, bsm_rpinfo); return false; } static void pim_update_pending_rp_cnt(struct bsm_scope *sz, struct bsgrp_node *bsgrp, uint16_t bsm_frag_tag, uint32_t total_rp_count) { if (bsgrp->pend_rp_cnt) { /* received bsm is different packet , * it is not same fragment. */ if (bsm_frag_tag != bsgrp->frag_tag) { if (PIM_DEBUG_BSM) zlog_debug( "%s,Received a new BSM ,so clear the pending bs_rpinfo list.", __func__); pim_bsm_rpinfos_free(bsgrp->partial_bsrp_list); bsgrp->pend_rp_cnt = total_rp_count; } } else bsgrp->pend_rp_cnt = total_rp_count; bsgrp->frag_tag = bsm_frag_tag; } /* Parsing BSR packet and adding to partial list of corresponding bsgrp node */ bool pim_bsm_parse_install_g2rp(struct bsm_scope *scope, uint8_t *buf, int buflen, uint16_t bsm_frag_tag) { struct bsmmsg_grpinfo grpinfo; struct bsmmsg_rpinfo rpinfo; struct prefix group; struct bsgrp_node *bsgrp = NULL; int frag_rp_cnt = 0; int offset = 0; int ins_count = 0; pim_addr grp_addr; while (buflen > offset) { if (offset + (int)sizeof(struct bsmmsg_grpinfo) > buflen) { if (PIM_DEBUG_BSM) zlog_debug( "%s: buflen received %d is less than the internal data structure of the packet would suggest", __func__, buflen); return false; } /* Extract Group tlv from BSM */ memcpy(&grpinfo, buf, sizeof(struct bsmmsg_grpinfo)); grp_addr = grpinfo.group.addr; if (PIM_DEBUG_BSM) zlog_debug( "%s, Group %pPAs Rpcount:%d Fragment-Rp-count:%d", __func__, &grp_addr, grpinfo.rp_count, grpinfo.frag_rp_count); buf += sizeof(struct bsmmsg_grpinfo); offset += sizeof(struct bsmmsg_grpinfo); group.family = PIM_AF; if (grpinfo.group.mask > PIM_MAX_BITLEN) { if (PIM_DEBUG_BSM) zlog_debug( "%s, prefix length specified: %d is too long", __func__, grpinfo.group.mask); return false; } pim_addr_to_prefix(&group, grp_addr); group.prefixlen = grpinfo.group.mask; /* Get the Group node for the BSM rp table */ bsgrp = pim_bsm_get_bsgrp_node(scope, &group); if (grpinfo.rp_count == 0) { struct bsm_rpinfo *old_rpinfo; /* BSR explicitly no longer has RPs for this group */ if (!bsgrp) continue; if (PIM_DEBUG_BSM) zlog_debug( "%s, Rp count is zero for group: %pPAs", __func__, &grp_addr); old_rpinfo = bsm_rpinfos_first(bsgrp->bsrp_list); if (old_rpinfo) pim_rp_del(scope->pim, old_rpinfo->rp_address, group, NULL, RP_SRC_BSR); pim_free_bsgrp_node(scope->bsrp_table, &bsgrp->group); pim_free_bsgrp_data(bsgrp); continue; } if (!bsgrp) { if (PIM_DEBUG_BSM) zlog_debug("%s, Create new BSM Group node.", __func__); /* create a new node to be added to the tree. */ bsgrp = pim_bsm_new_bsgrp_node(scope->bsrp_table, &group); if (!bsgrp) { zlog_debug( "%s, Failed to get the BSM group node.", __func__); continue; } bsgrp->scope = scope; } pim_update_pending_rp_cnt(scope, bsgrp, bsm_frag_tag, grpinfo.rp_count); frag_rp_cnt = grpinfo.frag_rp_count; ins_count = 0; while (frag_rp_cnt--) { if (offset + (int)sizeof(struct bsmmsg_rpinfo) > buflen) { if (PIM_DEBUG_BSM) zlog_debug( "%s, buflen received: %u is less than the internal data structure of the packet would suggest", __func__, buflen); return false; } /* Extract RP address tlv from BSM */ memcpy(&rpinfo, buf, sizeof(struct bsmmsg_rpinfo)); rpinfo.rp_holdtime = ntohs(rpinfo.rp_holdtime); buf += sizeof(struct bsmmsg_rpinfo); offset += sizeof(struct bsmmsg_rpinfo); if (PIM_DEBUG_BSM) { pim_addr rp_addr; rp_addr = rpinfo.rpaddr.addr; zlog_debug( "%s, Rp address - %pPAs; pri:%d hold:%d", __func__, &rp_addr, rpinfo.rp_pri, rpinfo.rp_holdtime); } /* Call Install api to update grp-rp mappings */ if (pim_install_bsm_grp_rp(scope->pim, bsgrp, &rpinfo)) ins_count++; } bsgrp->pend_rp_cnt -= ins_count; if (!bsgrp->pend_rp_cnt) { if (PIM_DEBUG_BSM) zlog_debug( "%s, Recvd all the rps for this group, so bsrp list with penidng rp list.", __func__); /* replace the bsrp_list with pending list */ pim_instate_pend_list(bsgrp); } } return true; } int pim_bsm_process(struct interface *ifp, pim_sgaddr *sg, uint8_t *buf, uint32_t buf_size, bool no_fwd) { struct bsm_hdr *bshdr; int sz = PIM_GBL_SZ_ID; struct bsmmsg_grpinfo *msg_grp; struct pim_interface *pim_ifp = NULL; struct bsm_frag *bsfrag; struct pim_instance *pim; uint16_t frag_tag; pim_addr bsr_addr; bool empty_bsm = false; /* BSM Packet acceptance validation */ pim_ifp = ifp->info; if (!pim_ifp) { if (PIM_DEBUG_BSM) zlog_debug("%s: multicast not enabled on interface %s", __func__, ifp->name); return -1; } if (pim_ifp->pim_passive_enable) { if (PIM_DEBUG_PIM_PACKETS) zlog_debug( "skip receiving PIM message on passive interface %s", ifp->name); return 0; } pim_ifp->pim_ifstat_bsm_rx++; pim = pim_ifp->pim; pim->bsm_rcvd++; /* Drop if bsm processing is disabled on interface */ if (!pim_ifp->bsm_enable) { zlog_warn("%s: BSM not enabled on interface %s", __func__, ifp->name); pim_ifp->pim_ifstat_bsm_cfg_miss++; pim->bsm_dropped++; return -1; } if (buf_size < (PIM_MSG_HEADER_LEN + sizeof(struct bsm_hdr))) { if (PIM_DEBUG_BSM) zlog_debug( "%s: received buffer length of %d which is too small to properly decode", __func__, buf_size); return -1; } bshdr = (struct bsm_hdr *)(buf + PIM_MSG_HEADER_LEN); if (bshdr->hm_len > PIM_MAX_BITLEN) { zlog_warn( "Bad hashmask length for %s; got %hhu, expected value in range 0-32", PIM_AF_NAME, bshdr->hm_len); pim->bsm_dropped++; return -1; } pim->global_scope.hashMasklen = bshdr->hm_len; frag_tag = ntohs(bshdr->frag_tag); /* NB: bshdr->bsr_addr.addr is packed/unaligned => memcpy */ memcpy(&bsr_addr, &bshdr->bsr_addr.addr, sizeof(bsr_addr)); /* Identify empty BSM */ if ((buf_size - PIM_BSM_HDR_LEN - PIM_MSG_HEADER_LEN) < PIM_BSM_GRP_LEN) empty_bsm = true; if (!empty_bsm) { msg_grp = (struct bsmmsg_grpinfo *)(buf + PIM_MSG_HEADER_LEN + PIM_BSM_HDR_LEN); /* Currently we don't support scope zoned BSM */ if (msg_grp->group.sz) { if (PIM_DEBUG_BSM) zlog_debug( "%s : Administratively scoped range BSM received", __func__); pim_ifp->pim_ifstat_bsm_invalid_sz++; pim->bsm_dropped++; return -1; } } if (!pim_addr_cmp(sg->grp, qpim_all_pim_routers_addr)) { /* Multicast BSMs are only accepted if source interface & IP * match RPF towards the BSR's IP address, or they have * no-forward set */ if (!no_fwd && !pim_nht_bsr_rpf_check(pim, bsr_addr, ifp, sg->src)) { if (PIM_DEBUG_BSM) zlog_debug( "BSM check: RPF to BSR %pPAs is not %pPA%%%s", &bsr_addr, &sg->src, ifp->name); pim->bsm_dropped++; return -1; } } else if (if_address_is_local(&sg->grp, PIM_AF, pim->vrf->vrf_id)) { /* Unicast BSM received - if ucast bsm not enabled on * the interface, drop it */ if (!pim_ifp->ucast_bsm_accept) { if (PIM_DEBUG_BSM) zlog_debug( "%s : Unicast BSM not enabled on interface %s", __func__, ifp->name); pim_ifp->pim_ifstat_ucast_bsm_cfg_miss++; pim->bsm_dropped++; return -1; } } else { if (PIM_DEBUG_BSM) zlog_debug("%s : Invalid destination address", __func__); pim->bsm_dropped++; return -1; } /* when the BSR restarts, it can get its own BSR advertisement thrown * back at it, and without this we'll go into ACCEPT_PREFERRED with * ourselves as the BSR when we should be in BSR_ELECTED. */ if (if_address_is_local(&bshdr->bsr_addr.addr, PIM_AF, pim->vrf->vrf_id)) { if (PIM_DEBUG_BSM) zlog_debug("%s : Dropping BSM from ourselves", __func__); pim->bsm_dropped++; return -1; } /* Drop if bsr is not preferred bsr */ if (!is_preferred_bsr(pim, bsr_addr, bshdr->bsr_prio)) { if (pim->global_scope.state == BSR_PENDING && !no_fwd) { /* in P-BSR state, non-preferred BSMs are forwarded, but * content is ignored. */ if (PIM_DEBUG_BSM) zlog_debug("%s : Forwarding non-preferred BSM during Pending-BSR state", __func__); pim_bsm_fwd_whole_sz(pim_ifp->pim, buf, buf_size, sz); return -1; } if (PIM_DEBUG_BSM) zlog_debug("%s : Received a non-preferred BSM", __func__); pim->bsm_dropped++; return -1; } if (no_fwd) { /* only accept no-forward BSM if quick refresh on startup */ if ((pim->global_scope.accept_nofwd_bsm) || (frag_tag == pim->global_scope.bsm_frag_tag)) { pim->global_scope.accept_nofwd_bsm = false; } else { if (PIM_DEBUG_BSM) zlog_debug("%s : nofwd_bsm received on %pPAs when accpt_nofwd_bsm false", __func__, &bsr_addr); pim->bsm_dropped++; pim_ifp->pim_ifstat_ucast_bsm_cfg_miss++; return -1; } } /* BSM packet is seen, so resetting accept_nofwd_bsm to false */ if (pim->global_scope.accept_nofwd_bsm) pim->global_scope.accept_nofwd_bsm = false; if (empty_bsm) { if (PIM_DEBUG_BSM) zlog_debug("%s : Empty Pref BSM received", __func__); } /* Parse Update bsm rp table and install/uninstall rp if required */ if (!pim_bsm_parse_install_g2rp( &pim_ifp->pim->global_scope, (buf + PIM_BSM_HDR_LEN + PIM_MSG_HEADER_LEN), (buf_size - PIM_BSM_HDR_LEN - PIM_MSG_HEADER_LEN), frag_tag)) { zlog_warn("BSM from %pPA failed to parse", (pim_addr *)&bshdr->bsr_addr.addr); pim->bsm_dropped++; return -1; } /* Restart the bootstrap timer */ pim_bs_timer_restart(&pim_ifp->pim->global_scope, PIM_BSR_DEFAULT_TIMEOUT); /* If new BSM received, clear the old bsm database */ if (pim_ifp->pim->global_scope.bsm_frag_tag != frag_tag) { if (PIM_DEBUG_BSM) { zlog_debug("%s: Current frag tag: %d Frag teg rcvd: %d", __func__, pim_ifp->pim->global_scope.bsm_frag_tag, frag_tag); } pim_bsm_frags_free(&pim_ifp->pim->global_scope); pim_ifp->pim->global_scope.bsm_frag_tag = frag_tag; } /* update the scope information from bsm */ pim_bsm_update(pim, bsr_addr, bshdr->bsr_prio); if (!no_fwd) { pim_bsm_fwd_whole_sz(pim_ifp->pim, buf, buf_size, sz); bsfrag = XCALLOC(MTYPE_PIM_BSM_FRAG, sizeof(struct bsm_frag) + buf_size); bsfrag->size = buf_size; memcpy(bsfrag->data, buf, buf_size); bsm_frags_add_tail(pim_ifp->pim->global_scope.bsm_frags, bsfrag); } return 0; } static void pim_elec_bsr_timer(struct event *t) { struct bsm_scope *scope = EVENT_ARG(t); struct bsm_frag *frag; struct bsm_hdr *hdr; assert(scope->state == BSR_ELECTED); scope->bsm_frag_tag++; frag = bsm_frags_first(scope->bsm_frags); assert(frag); hdr = (struct bsm_hdr *)(frag->data + PIM_MSG_HEADER_LEN); hdr->frag_tag = htons(scope->bsm_frag_tag); unsigned int timer = PIM_BS_TIME; if (scope->changed_bsm_trigger) { if (PIM_DEBUG_BSM) zlog_debug("Sending triggered BSM"); scope->changed_bsm_trigger--; timer = 5; } else { if (PIM_DEBUG_BSM) zlog_debug("Sending scheduled BSM"); pim_bsm_sent(scope); } pim_bsm_fwd_whole_sz(scope->pim, frag->data, frag->size, scope->sz_id); scope->current_bsr_last_ts = pim_time_monotonic_sec(); event_add_timer(router->master, pim_elec_bsr_timer, scope, timer, &scope->bs_timer); } void pim_bsm_changed(struct bsm_scope *scope) { struct event t; EVENT_OFF(scope->bs_timer); scope->changed_bsm_trigger = 2; t.arg = scope; pim_elec_bsr_timer(&t); } static void pim_cand_bsr_pending_expire(struct event *t) { struct bsm_scope *scope = EVENT_ARG(t); assertf(scope->state == BSR_PENDING, "state=%d", scope->state); if (!pim_addr_is_any(scope->current_bsr)) { assertf(scope->cand_bsr_prio >= scope->current_bsr_prio, "cand_bsr %pPA prio %u is less than current_bsr %pPA prio %u", &scope->bsr_addrsel.run_addr, scope->current_bsr_prio, &scope->current_bsr, scope->cand_bsr_prio); if (scope->cand_bsr_prio == scope->current_bsr_prio) assertf(pim_addr_cmp(scope->bsr_addrsel.run_addr, scope->current_bsr) > 0, "cand_bsr %pPA < current_bsr %pPA", &scope->bsr_addrsel.run_addr, &scope->current_bsr); } if (PIM_DEBUG_BSM) zlog_debug("Elected BSR, wait expired without preferable BSMs"); scope->state = BSR_ELECTED; scope->current_bsr_prio = scope->cand_bsr_prio; scope->current_bsr = scope->bsr_addrsel.run_addr; scope->bsm_frag_tag = frr_weak_random(); scope->current_bsr_first_ts = pim_time_monotonic_sec(); pim_cand_rp_trigger(scope); pim_bsm_generate(scope); } #if PIM_IPV == 6 static float bsr_addr_delay(pim_addr best, pim_addr local) { unsigned int pos; uint32_t best_4b, local_4b; float delay_log; for (pos = 0; pos < 12; pos++) { if (best.s6_addr[pos] != local.s6_addr[pos]) break; } memcpy(&best_4b, &best.s6_addr[pos], 4); memcpy(&local_4b, &local.s6_addr[pos], 4); delay_log = log2(1 + ntohl(best_4b) - ntohl(local_4b)); delay_log += (12 - pos) * 8; return delay_log / 64.; } #endif static void pim_cand_bsr_pending(struct bsm_scope *scope) { unsigned int bs_rand_override; uint8_t best_prio; pim_addr best_addr; float prio_delay, addr_delay; EVENT_OFF(scope->bs_timer); EVENT_OFF(scope->t_ebsr_regen_bsm); scope->state = BSR_PENDING; best_prio = MAX(scope->cand_bsr_prio, scope->current_bsr_prio); best_addr = pim_addr_cmp(scope->bsr_addrsel.run_addr, scope->current_bsr) > 0 ? scope->bsr_addrsel.run_addr : scope->current_bsr; /* RFC5059 sec.5 */ #if PIM_IPV == 4 if (scope->cand_bsr_prio == best_prio) { prio_delay = 0.; /* log2(1) = 0 */ addr_delay = log2(1 + ntohl(best_addr.s_addr) - ntohl(scope->bsr_addrsel.run_addr.s_addr)) / 16.; } else { prio_delay = 2. * log2(1 + best_prio - scope->cand_bsr_prio); addr_delay = 2 - (ntohl(scope->bsr_addrsel.run_addr.s_addr) / (float)(1 << 31)); } #else if (scope->cand_bsr_prio == best_prio) { prio_delay = 0.; /* log2(1) = 0 */ addr_delay = bsr_addr_delay(best_addr, scope->bsr_addrsel.run_addr); } else { prio_delay = 2. * log2(1 + best_prio - scope->cand_bsr_prio); addr_delay = 2 - (ntohl(scope->bsr_addrsel.run_addr.s6_addr32[0]) / (float)(1 << 31)); } #endif bs_rand_override = 5000 + (int)((prio_delay + addr_delay) * 1000.); if (PIM_DEBUG_BSM) zlog_debug("Pending-BSR (%u, %pPA), waiting %ums", scope->cand_bsr_prio, &scope->bsr_addrsel.run_addr, bs_rand_override); event_add_timer_msec(router->master, pim_cand_bsr_pending_expire, scope, bs_rand_override, &scope->bs_timer); } static inline pim_addr if_highest_addr(pim_addr cur, struct interface *ifp) { struct connected *connected; frr_each (if_connected, ifp->connected, connected) { pim_addr conn_addr; if (connected->address->family != PIM_AF) continue; conn_addr = pim_addr_from_prefix(connected->address); /* highest address */ if (pim_addr_cmp(conn_addr, cur) > 0) cur = conn_addr; } return cur; } static void cand_addrsel_clear(struct cand_addrsel *asel) { asel->run = false; asel->run_addr = PIMADDR_ANY; } /* returns whether address or active changed */ static bool cand_addrsel_update(struct cand_addrsel *asel, struct vrf *vrf) { bool is_any = false, prev_run = asel->run; struct interface *ifp = NULL; pim_addr new_addr = PIMADDR_ANY; if (!asel->cfg_enable) goto out_disable; switch (asel->cfg_mode) { case CAND_ADDR_EXPLICIT: new_addr = asel->cfg_addr; ifp = if_lookup_address_local(&asel->cfg_addr, PIM_AF, vrf->vrf_id); break; case CAND_ADDR_IFACE: ifp = if_lookup_by_name_vrf(asel->cfg_ifname, vrf); if (ifp) new_addr = if_highest_addr(PIMADDR_ANY, ifp); break; case CAND_ADDR_ANY: is_any = true; /* fallthru */ case CAND_ADDR_LO: FOR_ALL_INTERFACES (vrf, ifp) { if (!if_is_up(ifp)) continue; if (is_any || if_is_loopback(ifp) || if_is_vrf(ifp)) new_addr = if_highest_addr(new_addr, ifp); } break; } if (ifp && !if_is_up(ifp)) goto out_disable; if (pim_addr_is_any(new_addr)) goto out_disable; /* nothing changed re. address (don't care about interface changes) */ if (asel->run && !pim_addr_cmp(asel->run_addr, new_addr)) return !prev_run; asel->run = true; asel->run_addr = new_addr; return true; out_disable: asel->run = false; asel->run_addr = PIMADDR_ANY; return prev_run; } static void pim_cand_bsr_stop(struct bsm_scope *scope, bool verbose) { cand_addrsel_clear(&scope->bsr_addrsel); switch (scope->state) { case NO_INFO: case ACCEPT_ANY: case ACCEPT_PREFERRED: return; case BSR_PENDING: case BSR_ELECTED: break; } if (PIM_DEBUG_BSM) zlog_debug("Candidate BSR ceasing operation"); EVENT_OFF(scope->t_ebsr_regen_bsm); EVENT_OFF(scope->bs_timer); pim_crp_db_clear(scope); pim_bsm_accept_any(scope); } static void pim_cand_bsr_trigger(struct bsm_scope *scope, bool verbose) { /* this is called on all state changes even if we aren't configured * to be C-BSR at all. */ if (!scope->bsr_addrsel.run) return; if (scope->current_bsr_prio > scope->cand_bsr_prio) { assert(scope->state == ACCEPT_PREFERRED); if (!verbose) return; if (PIM_DEBUG_BSM) zlog_debug("Candidate BSR: known better BSR %pPA (higher priority %u > %u)", &scope->current_bsr, scope->current_bsr_prio, scope->cand_bsr_prio); return; } else if (scope->current_bsr_prio == scope->cand_bsr_prio && pim_addr_cmp(scope->current_bsr, scope->bsr_addrsel.run_addr) > 0) { assert(scope->state == ACCEPT_PREFERRED); if (!verbose) return; if (PIM_DEBUG_BSM) zlog_debug("Candidate BSR: known better BSR %pPA (higher address > %pPA)", &scope->current_bsr, &scope->bsr_addrsel.run_addr); return; } if (!pim_addr_cmp(scope->current_bsr, scope->bsr_addrsel.run_addr)) return; pim_cand_bsr_pending(scope); } void pim_cand_bsr_apply(struct bsm_scope *scope) { if (!cand_addrsel_update(&scope->bsr_addrsel, scope->pim->vrf)) return; if (!scope->bsr_addrsel.run) { pim_cand_bsr_stop(scope, true); return; } if (PIM_DEBUG_BSM) zlog_debug("Candidate BSR: %pPA, priority %u", &scope->bsr_addrsel.run_addr, scope->cand_bsr_prio); pim_cand_bsr_trigger(scope, true); } static void pim_cand_rp_adv_stop_maybe(struct bsm_scope *scope) { /* actual check whether stop should be sent - covers address * changes as well as run_addr = 0.0.0.0 (C-RP shutdown) */ if (pim_addr_is_any(scope->cand_rp_prev_addr) || !pim_addr_cmp(scope->cand_rp_prev_addr, scope->cand_rp_addrsel.run_addr)) return; switch (scope->state) { case ACCEPT_PREFERRED: case BSR_ELECTED: break; case NO_INFO: case ACCEPT_ANY: case BSR_PENDING: default: return; } if (PIM_DEBUG_BSM) zlog_debug("Candidate-RP (-, %pPA) deregistering self to %pPA", &scope->cand_rp_prev_addr, &scope->current_bsr); struct cand_rp_msg *msg; uint8_t buf[PIM_MSG_HEADER_LEN + sizeof(*msg) + sizeof(pim_encoded_group)]; msg = (struct cand_rp_msg *)(&buf[PIM_MSG_HEADER_LEN]); msg->prefix_cnt = 0; msg->rp_prio = 255; msg->rp_holdtime = 0; msg->rp_addr.family = PIM_IANA_AFI; msg->rp_addr.reserved = 0; msg->rp_addr.addr = scope->cand_rp_prev_addr; pim_msg_build_header(PIMADDR_ANY, scope->current_bsr, buf, sizeof(buf), PIM_MSG_TYPE_CANDIDATE, false); if (pim_msg_send(scope->unicast_sock, PIMADDR_ANY, scope->current_bsr, buf, sizeof(buf), NULL)) { zlog_warn("failed to send Cand-RP message: %m"); } scope->cand_rp_prev_addr = PIMADDR_ANY; } static void pim_cand_rp_adv(struct event *t) { struct bsm_scope *scope = EVENT_ARG(t); int next_msec; pim_cand_rp_adv_stop_maybe(scope); if (!scope->cand_rp_addrsel.run) { scope->cand_rp_adv_trigger = 0; return; } switch (scope->state) { case ACCEPT_PREFERRED: case BSR_ELECTED: break; case ACCEPT_ANY: case BSR_PENDING: case NO_INFO: default: /* state change will retrigger */ scope->cand_rp_adv_trigger = 0; zlog_warn("Candidate-RP advertisement not sent in state %d", scope->state); return; } if (PIM_DEBUG_BSM) zlog_debug("Candidate-RP (%u, %pPA) advertising %zu groups to %pPA", scope->cand_rp_prio, &scope->cand_rp_addrsel.run_addr, cand_rp_groups_count(scope->cand_rp_groups), &scope->current_bsr); struct cand_rp_group *grp; struct cand_rp_msg *msg; uint8_t buf[PIM_MSG_HEADER_LEN + sizeof(*msg) + sizeof(pim_encoded_group) * cand_rp_groups_count(scope->cand_rp_groups)]; size_t i = 0; msg = (struct cand_rp_msg *)(&buf[PIM_MSG_HEADER_LEN]); msg->prefix_cnt = cand_rp_groups_count(scope->cand_rp_groups); msg->rp_prio = scope->cand_rp_prio; msg->rp_holdtime = htons(MAX(151, (scope->cand_rp_interval * 5 + 1) / 2)); msg->rp_addr.family = PIM_IANA_AFI; msg->rp_addr.reserved = 0; msg->rp_addr.addr = scope->cand_rp_addrsel.run_addr; frr_each (cand_rp_groups, scope->cand_rp_groups, grp) { memset(&msg->groups[i], 0, sizeof(msg->groups[i])); msg->groups[i].family = PIM_IANA_AFI; msg->groups[i].mask = grp->p.prefixlen; msg->groups[i].addr = grp->p.prefix; i++; } scope->cand_rp_prev_addr = scope->cand_rp_addrsel.run_addr; pim_msg_build_header(scope->cand_rp_addrsel.run_addr, scope->current_bsr, buf, sizeof(buf), PIM_MSG_TYPE_CANDIDATE, false); if (pim_msg_send(scope->unicast_sock, scope->cand_rp_addrsel.run_addr, scope->current_bsr, buf, sizeof(buf), NULL)) { zlog_warn("failed to send Cand-RP message: %m"); } /* -1s...+1s */ next_msec = (frr_weak_random() & 2047) - 1024; if (scope->cand_rp_adv_trigger) { scope->cand_rp_adv_trigger--; next_msec += 2000; } else next_msec += scope->cand_rp_interval * 1000; event_add_timer_msec(router->master, pim_cand_rp_adv, scope, next_msec, &scope->cand_rp_adv_timer); } void pim_cand_rp_trigger(struct bsm_scope *scope) { if (scope->cand_rp_adv_trigger && scope->cand_rp_addrsel.run) { scope->cand_rp_adv_trigger = PIM_CRP_ADV_TRIGCOUNT; /* already scheduled to send triggered advertisements, don't * reschedule so burst changes don't result in an advertisement * burst */ return; } EVENT_OFF(scope->cand_rp_adv_timer); if (!scope->cand_rp_addrsel.run) return; scope->cand_rp_adv_trigger = PIM_CRP_ADV_TRIGCOUNT; struct event t; t.arg = scope; pim_cand_rp_adv(&t); } void pim_cand_rp_apply(struct bsm_scope *scope) { if (!cand_addrsel_update(&scope->cand_rp_addrsel, scope->pim->vrf)) return; if (!scope->cand_rp_addrsel.run) { if (PIM_DEBUG_BSM) zlog_debug("Candidate RP ceasing operation"); cand_addrsel_clear(&scope->cand_rp_addrsel); EVENT_OFF(scope->cand_rp_adv_timer); pim_cand_rp_adv_stop_maybe(scope); scope->cand_rp_adv_trigger = 0; return; } if (PIM_DEBUG_BSM) zlog_debug("Candidate RP: %pPA, priority %u", &scope->cand_rp_addrsel.run_addr, scope->cand_rp_prio); pim_cand_rp_trigger(scope); } void pim_cand_rp_grp_add(struct bsm_scope *scope, const prefix_pim *p) { struct cand_rp_group *grp, ref; ref.p = *p; grp = cand_rp_groups_find(scope->cand_rp_groups, &ref); if (grp) return; grp = XCALLOC(MTYPE_PIM_CAND_RP_GRP, sizeof(*grp)); grp->p = *p; cand_rp_groups_add(scope->cand_rp_groups, grp); pim_cand_rp_trigger(scope); } void pim_cand_rp_grp_del(struct bsm_scope *scope, const prefix_pim *p) { struct cand_rp_group *grp, ref; ref.p = *p; grp = cand_rp_groups_find(scope->cand_rp_groups, &ref); if (!grp) return; cand_rp_groups_del(scope->cand_rp_groups, grp); XFREE(MTYPE_PIM_CAND_RP_GRP, grp); pim_cand_rp_trigger(scope); } static struct event *t_cand_addrs_reapply; static void pim_cand_addrs_reapply(struct event *t) { struct vrf *vrf; RB_FOREACH (vrf, vrf_id_head, &vrfs_by_id) { struct pim_instance *pi = vrf->info; if (!pi) continue; /* these call cand_addrsel_update() and apply changes */ pim_cand_bsr_apply(&pi->global_scope); pim_cand_rp_apply(&pi->global_scope); } } void pim_cand_addrs_changed(void) { EVENT_OFF(t_cand_addrs_reapply); event_add_timer_msec(router->master, pim_cand_addrs_reapply, NULL, 1, &t_cand_addrs_reapply); } static void cand_addrsel_config_write(struct vty *vty, struct cand_addrsel *addrsel) { switch (addrsel->cfg_mode) { case CAND_ADDR_LO: break; case CAND_ADDR_ANY: vty_out(vty, " source any"); break; case CAND_ADDR_IFACE: vty_out(vty, " source interface %s", addrsel->cfg_ifname); break; case CAND_ADDR_EXPLICIT: vty_out(vty, " source address %pPA", &addrsel->cfg_addr); break; } } int pim_cand_config_write(struct pim_instance *pim, struct vty *vty) { struct bsm_scope *scope = &pim->global_scope; int ret = 0; if (scope->cand_rp_addrsel.cfg_enable) { vty_out(vty, " bsr candidate-rp"); if (scope->cand_rp_prio != 192) vty_out(vty, " priority %u", scope->cand_rp_prio); if (scope->cand_rp_interval != PIM_CRP_ADV_INTERVAL) vty_out(vty, " interval %u", scope->cand_rp_interval); cand_addrsel_config_write(vty, &scope->cand_rp_addrsel); vty_out(vty, "\n"); ret++; struct cand_rp_group *group; frr_each (cand_rp_groups, scope->cand_rp_groups, group) { vty_out(vty, " bsr candidate-rp group %pFX\n", &group->p); ret++; } } if (scope->bsr_addrsel.cfg_enable) { vty_out(vty, " bsr candidate-bsr"); if (scope->cand_bsr_prio != 64) vty_out(vty, " priority %u", scope->cand_bsr_prio); cand_addrsel_config_write(vty, &scope->bsr_addrsel); vty_out(vty, "\n"); ret++; } return ret; }