/* * BIRD -- OSPF * * (c) 1999--2004 Ondrej Filip * * Can be freely distributed and used under the terms of the GNU GPL. */ /** * DOC: Open Shortest Path First (OSPF) * * The OSPF protocol is quite complicated and its complex implemenation is * split to many files. In |ospf.c|, you will find mainly the interface * for communication with the core (e.g., reconfiguration hooks, shutdown * and initialisation and so on). In |packet.c|, you will find various * functions for sending and receiving generic OSPF packets. There are * also routines for authentication and checksumming. File |iface.c| contains * the interface state machine and functions for allocation and deallocation of OSPF's * interface data structures. Source |neighbor.c| includes the neighbor state * machine and functions for election of Designated Router and Backup * Designated router. In |hello.c|, there are routines for sending * and receiving of hello packets as well as functions for maintaining * wait times and the inactivity timer. Files |lsreq.c|, |lsack.c|, |dbdes.c| * contain functions for sending and receiving of link-state requests, * link-state acknowledgements and database descriptions respectively. * In |lsupd.c|, there are functions for sending and receiving * of link-state updates and also the flooding algorithm. Source |topology.c| is * a place where routines for searching LSAs in the link-state database, * adding and deleting them reside, there also are functions for originating * of various types of LSAs (router LSA, net LSA, external LSA). File |rt.c| * contains routines for calculating the routing table. |lsalib.c| is a set * of various functions for working with the LSAs (endianity conversions, * calculation of checksum etc.). * * One instance of the protocol is able to hold LSA databases for * multiple OSPF areas, to exchange routing information between * multiple neighbors and to calculate the routing tables. The core * structure is &proto_ospf to which multiple &ospf_area and * &ospf_iface structures are connected. &ospf_area is also connected to * &top_hash_graph which is a dynamic hashing structure that * describes the link-state database. It allows fast search, addition * and deletion. Each LSA is kept in two pieces: header and body. Both of them are * kept in the endianity of the CPU. * * In OSPFv2 specification, it is implied that there is one IP prefix * for each physical network/interface (unless it is an ptp link). But * in modern systems, there might be more independent IP prefixes * associated with an interface. To handle this situation, we have * one &ospf_iface for each active IP prefix (instead for each active * iface); This behaves like virtual interface for the purpose of OSPF. * If we receive packet, we associate it with a proper virtual interface * mainly according to its source address. * * OSPF keeps one socket per &ospf_iface. This allows us (compared to * one socket approach) to evade problems with a limit of multicast * groups per socket and with sending multicast packets to appropriate * interface in a portable way. The socket is associated with * underlying physical iface and should not receive packets received * on other ifaces (unfortunately, this is not true on * BSD). Generally, one packet can be received by more sockets (for * example, if there are more &ospf_iface on one physical iface), * therefore we explicitly filter received packets according to * src/dst IP address and received iface. * * Vlinks are implemented using particularly degenerate form of * &ospf_iface, which has several exceptions: it does not have its * iface or socket (it copies these from 'parent' &ospf_iface) and it * is present in iface list even when down (it is not freed in * ospf_iface_down()). * * The heart beat of ospf is ospf_disp(). It is called at regular intervals * (&proto_ospf->tick). It is responsible for aging and flushing of LSAs in * the database, for routing table calculaction and it call area_disp() of every * ospf_area. * * The function area_disp() is * responsible for late originating of router LSA and network LSA * and for cleanup before routing table calculation process in * the area. * To every &ospf_iface, we connect one or more * &ospf_neighbor's -- a structure containing many timers and queues * for building adjacency and for exchange of routing messages. * * BIRD's OSPF implementation respects RFC2328 in every detail, but * some of internal algorithms do differ. The RFC recommends making a snapshot * of the link-state database when a new adjacency is forming and sending * the database description packets based on the information in this * snapshot. The database can be quite large in some networks, so * rather we walk through a &slist structure which allows us to * continue even if the actual LSA we were working with is deleted. New * LSAs are added at the tail of this &slist. * * We also don't keep a separate OSPF routing table, because the core * helps us by being able to recognize when a route is updated * to an identical one and it suppresses the update automatically. * Due to this, we can flush all the routes we've recalculated and * also those we've deleted to the core's routing table and the * core will take care of the rest. This simplifies the process * and conserves memory. */ #include #include "ospf.h" static int ospf_reload_routes(struct proto *p); static void ospf_rt_notify(struct proto *p, struct rtable *table UNUSED, net * n, rte * new, rte * old UNUSED, ea_list * attrs); static int ospf_rte_better(struct rte *new, struct rte *old); static int ospf_rte_same(struct rte *new, struct rte *old); static void ospf_disp(timer *timer); static void ospf_area_initfib(struct fib_node *fn) { struct area_net *an = (struct area_net *) fn; an->hidden = 0; an->active = -1; /* Force to regenerate summary lsa */ /* ac->oldactive will be rewritten by ospf_rt_spf() */ } static void add_area_nets(struct ospf_area *oa, struct ospf_area_config *ac) { struct proto_ospf *po = oa->po; struct area_net_config *anc; struct area_net *an; fib_init(&oa->net_fib, po->proto.pool, sizeof(struct area_net), 0, ospf_area_initfib); fib_init(&oa->enet_fib, po->proto.pool, sizeof(struct area_net), 0, ospf_area_initfib); WALK_LIST(anc, ac->net_list) { an = (struct area_net *) fib_get(&oa->net_fib, &anc->px.addr, anc->px.len); an->hidden = anc->hidden; } WALK_LIST(anc, ac->enet_list) { an = (struct area_net *) fib_get(&oa->enet_fib, &anc->px.addr, anc->px.len); an->hidden = anc->hidden; an->tag = anc->tag; } } static void ospf_area_add(struct proto_ospf *po, struct ospf_area_config *ac, int reconf) { struct proto *p = &po->proto; struct ospf_area *oa; OSPF_TRACE(D_EVENTS, "Adding area %R", ac->areaid); oa = mb_allocz(p->pool, sizeof(struct ospf_area)); add_tail(&po->area_list, NODE oa); po->areano++; oa->ac = ac; oa->areaid = ac->areaid; oa->rt = NULL; oa->po = po; fib_init(&oa->rtr, p->pool, sizeof(ort), 0, ospf_rt_initort); add_area_nets(oa, ac); if (oa->areaid == 0) po->backbone = oa; #ifdef OSPFv2 oa->options = ac->type; #else /* OSPFv3 */ oa->options = OPT_R | ac->type | OPT_V6; #endif if (reconf) ospf_ifaces_reconfigure(oa, ac); } static void ospf_area_remove(struct ospf_area *oa) { struct proto *p = &oa->po->proto; OSPF_TRACE(D_EVENTS, "Removing area %R", oa->areaid); /* We suppose that interfaces are already removed */ ospf_flush_area(oa->po, oa->areaid); fib_free(&oa->rtr); fib_free(&oa->net_fib); fib_free(&oa->enet_fib); if (oa->translator_timer) rfree(oa->translator_timer); oa->po->areano--; rem_node(NODE oa); mb_free(oa); } struct ospf_area * ospf_find_area(struct proto_ospf *po, u32 aid) { struct ospf_area *oa; WALK_LIST(oa, po->area_list) if (((struct ospf_area *) oa)->areaid == aid) return oa; return NULL; } static struct ospf_iface * ospf_find_vlink(struct proto_ospf *po, u32 voa, u32 vid) { struct ospf_iface *ifa; WALK_LIST(ifa, po->iface_list) if ((ifa->type == OSPF_IT_VLINK) && (ifa->voa->areaid == voa) && (ifa->vid == vid)) return ifa; return NULL; } static int ospf_start(struct proto *p) { struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_config *c = (struct ospf_config *) (p->cf); struct ospf_area_config *ac; po->router_id = proto_get_router_id(p->cf); po->rfc1583 = c->rfc1583; po->ebit = 0; po->ecmp = c->ecmp; po->tick = c->tick; po->disp_timer = tm_new(p->pool); po->disp_timer->data = po; po->disp_timer->randomize = 0; po->disp_timer->hook = ospf_disp; po->disp_timer->recurrent = po->tick; tm_start(po->disp_timer, 1); po->lsab_size = 256; po->lsab_used = 0; po->lsab = mb_alloc(p->pool, po->lsab_size); po->nhpool = lp_new(p->pool, 12*sizeof(struct mpnh)); init_list(&(po->iface_list)); init_list(&(po->area_list)); fib_init(&po->rtf, p->pool, sizeof(ort), 0, ospf_rt_initort); po->areano = 0; po->gr = ospf_top_new(p->pool); s_init_list(&(po->lsal)); WALK_LIST(ac, c->area_list) ospf_area_add(po, ac, 0); /* Add all virtual links */ struct ospf_iface_patt *ic; WALK_LIST(ic, c->vlink_list) ospf_iface_new(po->backbone, NULL, ic); return PS_UP; } static void ospf_dump(struct proto *p) { struct ospf_iface *ifa; struct ospf_neighbor *n; struct proto_ospf *po = (struct proto_ospf *) p; OSPF_TRACE(D_EVENTS, "Area number: %d", po->areano); WALK_LIST(ifa, po->iface_list) { OSPF_TRACE(D_EVENTS, "Interface: %s", (ifa->iface ? ifa->iface->name : "(null)")); OSPF_TRACE(D_EVENTS, "state: %u", ifa->state); OSPF_TRACE(D_EVENTS, "DR: %R", ifa->drid); OSPF_TRACE(D_EVENTS, "BDR: %R", ifa->bdrid); WALK_LIST(n, ifa->neigh_list) { OSPF_TRACE(D_EVENTS, " neighbor %R in state %u", n->rid, n->state); } } /* OSPF_TRACE(D_EVENTS, "LSA graph dump start:"); ospf_top_dump(po->gr, p); OSPF_TRACE(D_EVENTS, "LSA graph dump finished"); */ neigh_dump_all(); } static struct proto * ospf_init(struct proto_config *c) { struct proto *p = proto_new(c, sizeof(struct proto_ospf)); p->make_tmp_attrs = ospf_make_tmp_attrs; p->store_tmp_attrs = ospf_store_tmp_attrs; p->import_control = ospf_import_control; p->reload_routes = ospf_reload_routes; p->accept_ra_types = RA_OPTIMAL; p->rt_notify = ospf_rt_notify; p->if_notify = ospf_if_notify; p->ifa_notify = ospf_ifa_notify; p->rte_better = ospf_rte_better; p->rte_same = ospf_rte_same; return p; } /* If new is better return 1 */ static int ospf_rte_better(struct rte *new, struct rte *old) { if (new->u.ospf.metric1 == LSINFINITY) return 0; if(new->attrs->source < old->attrs->source) return 1; if(new->attrs->source > old->attrs->source) return 0; if(new->attrs->source == RTS_OSPF_EXT2) { if(new->u.ospf.metric2 < old->u.ospf.metric2) return 1; if(new->u.ospf.metric2 > old->u.ospf.metric2) return 0; } if (new->u.ospf.metric1 < old->u.ospf.metric1) return 1; return 0; /* Old is shorter or same */ } static int ospf_rte_same(struct rte *new, struct rte *old) { /* new->attrs == old->attrs always */ return new->u.ospf.metric1 == old->u.ospf.metric1 && new->u.ospf.metric2 == old->u.ospf.metric2 && new->u.ospf.tag == old->u.ospf.tag && new->u.ospf.router_id == old->u.ospf.router_id; } static ea_list * ospf_build_attrs(ea_list * next, struct linpool *pool, u32 m1, u32 m2, u32 tag, u32 rid) { struct ea_list *l = lp_alloc(pool, sizeof(struct ea_list) + 4 * sizeof(eattr)); l->next = next; l->flags = EALF_SORTED; l->count = 4; l->attrs[0].id = EA_OSPF_METRIC1; l->attrs[0].flags = 0; l->attrs[0].type = EAF_TYPE_INT | EAF_TEMP; l->attrs[0].u.data = m1; l->attrs[1].id = EA_OSPF_METRIC2; l->attrs[1].flags = 0; l->attrs[1].type = EAF_TYPE_INT | EAF_TEMP; l->attrs[1].u.data = m2; l->attrs[2].id = EA_OSPF_TAG; l->attrs[2].flags = 0; l->attrs[2].type = EAF_TYPE_INT | EAF_TEMP; l->attrs[2].u.data = tag; l->attrs[3].id = EA_OSPF_ROUTER_ID; l->attrs[3].flags = 0; l->attrs[3].type = EAF_TYPE_ROUTER_ID | EAF_TEMP; l->attrs[3].u.data = rid; return l; } void schedule_net_lsa(struct ospf_iface *ifa) { struct proto *p = &ifa->oa->po->proto; OSPF_TRACE(D_EVENTS, "Scheduling network-LSA origination for iface %s", ifa->iface->name); ifa->orignet = 1; } #ifdef OSPFv3 void schedule_link_lsa(struct ospf_iface *ifa) { struct proto *p = &ifa->oa->po->proto; OSPF_TRACE(D_EVENTS, "Scheduling link-LSA origination for iface %s", ifa->iface->name); ifa->origlink = 1; } #endif void schedule_rt_lsa(struct ospf_area *oa) { struct proto *p = &oa->po->proto; OSPF_TRACE(D_EVENTS, "Scheduling router-LSA origination for area %R", oa->areaid); oa->origrt = 1; } void schedule_rtcalc(struct proto_ospf *po) { struct proto *p = &po->proto; if (po->calcrt) return; OSPF_TRACE(D_EVENTS, "Scheduling routing table calculation"); po->calcrt = 1; } static int ospf_reload_routes(struct proto *p) { struct proto_ospf *po = (struct proto_ospf *) p; if (po->calcrt != 2) OSPF_TRACE(D_EVENTS, "Scheduling routing table calculation with route reload"); po->calcrt = 2; return 1; } /** * area_disp - invokes origination of * router LSA and routing table cleanup * @oa: ospf area * * It invokes aging and when @ospf_area->origrt is set to 1, start * function for origination of router, network LSAs. */ void area_disp(struct ospf_area *oa) { struct proto_ospf *po = oa->po; struct ospf_iface *ifa; /* Now try to originage rt_lsa */ if (oa->origrt) update_rt_lsa(oa); /* Now try to originate network LSA's */ WALK_LIST(ifa, po->iface_list) { #ifdef OSPFv3 /* Link LSA should be originated before Network LSA */ if (ifa->origlink && (ifa->oa == oa)) update_link_lsa(ifa); #endif if (ifa->orignet && (ifa->oa == oa)) update_net_lsa(ifa); } } /** * ospf_disp - invokes routing table calctulation, aging and also area_disp() * @timer: timer usually called every @proto_ospf->tick second, @timer->data * point to @proto_ospf */ void ospf_disp(timer * timer) { struct proto_ospf *po = timer->data; struct ospf_area *oa; WALK_LIST(oa, po->area_list) area_disp(oa); /* Age LSA DB */ ospf_age(po); /* Calculate routing table */ if (po->calcrt) ospf_rt_spf(po); } /** * ospf_import_control - accept or reject new route from nest's routing table * @p: current instance of protocol * @new: the new route * @attrs: list of attributes * @pool: pool for allocation of attributes * * Its quite simple. It does not accept our own routes and leaves the decision on * import to the filters. */ int ospf_import_control(struct proto *p, rte ** new, ea_list ** attrs, struct linpool *pool) { struct ospf_area *oa = ospf_main_area((struct proto_ospf *) p); rte *e = *new; if (p == e->attrs->proto) return -1; /* Reject our own routes */ if (oa_is_stub(oa)) return -1; /* Do not export routes to stub areas */ eattr *ea = ea_find(e->attrs->eattrs, EA_GEN_IGP_METRIC); u32 m1 = (ea && (ea->u.data < LSINFINITY)) ? ea->u.data : LSINFINITY; *attrs = ospf_build_attrs(*attrs, pool, m1, 10000, 0, 0); return 0; /* Leave decision to the filters */ } struct ea_list * ospf_make_tmp_attrs(struct rte *rt, struct linpool *pool) { return ospf_build_attrs(NULL, pool, rt->u.ospf.metric1, rt->u.ospf.metric2, rt->u.ospf.tag, rt->u.ospf.router_id); } void ospf_store_tmp_attrs(struct rte *rt, struct ea_list *attrs) { rt->u.ospf.metric1 = ea_get_int(attrs, EA_OSPF_METRIC1, LSINFINITY); rt->u.ospf.metric2 = ea_get_int(attrs, EA_OSPF_METRIC2, 10000); rt->u.ospf.tag = ea_get_int(attrs, EA_OSPF_TAG, 0); rt->u.ospf.router_id = ea_get_int(attrs, EA_OSPF_ROUTER_ID, 0); } /** * ospf_shutdown - Finish of OSPF instance * @p: current instance of protocol * * RFC does not define any action that should be taken before router * shutdown. To make my neighbors react as fast as possible, I send * them hello packet with empty neighbor list. They should start * their neighbor state machine with event %NEIGHBOR_1WAY. */ static int ospf_shutdown(struct proto *p) { struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_iface *ifa; OSPF_TRACE(D_EVENTS, "Shutdown requested"); /* And send to all my neighbors 1WAY */ WALK_LIST(ifa, po->iface_list) ospf_iface_shutdown(ifa); /* Cleanup locked rta entries */ FIB_WALK(&po->rtf, nftmp) { rta_free(((ort *) nftmp)->old_rta); } FIB_WALK_END; return PS_DOWN; } static void ospf_rt_notify(struct proto *p, rtable *tbl UNUSED, net * n, rte * new, rte * old UNUSED, ea_list * attrs) { struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_area *oa = ospf_main_area(po); ort *nf = (ort *) fib_get(&po->rtf, &n->n.prefix, n->n.pxlen); struct fib_node *fn = &nf->fn; if (!new) { if (fn->x1 != EXT_EXPORT) return; flush_ext_lsa(oa, fn); /* Old external route might blocked some NSSA translation */ if (po->areano > 1) schedule_rtcalc(po); } /* Get route attributes */ u32 m1 = ea_get_int(attrs, EA_OSPF_METRIC1, LSINFINITY); u32 m2 = ea_get_int(attrs, EA_OSPF_METRIC2, 10000); u32 metric = (m1 != LSINFINITY) ? m1 : (m2 | LSA_EXT_EBIT); u32 tag = ea_get_int(attrs, EA_OSPF_TAG, 0); ip_addr gw = IPA_NONE; // FIXME check for gw should be per ifa, not per iface if ((new->attrs->dest == RTD_ROUTER) && ipa_nonzero(new->attrs->gw) && !ipa_has_link_scope(new->attrs->gw) && (ospf_iface_find((struct proto_ospf *) p, new->attrs->iface) != NULL)) gw = new->attrs->gw; originate_ext_lsa(oa, fn, EXT_EXPORT, metric, gw, tag); } static void ospf_get_status(struct proto *p, byte * buf) { struct proto_ospf *po = (struct proto_ospf *) p; if (p->proto_state == PS_DOWN) buf[0] = 0; else { struct ospf_iface *ifa; struct ospf_neighbor *n; int adj = 0; WALK_LIST(ifa, po->iface_list) WALK_LIST(n, ifa->neigh_list) if (n->state == NEIGHBOR_FULL) adj = 1; if (adj == 0) strcpy(buf, "Alone"); else strcpy(buf, "Running"); } } static void ospf_get_route_info(rte * rte, byte * buf, ea_list * attrs UNUSED) { char *type = ""; switch(rte->attrs->source) { case RTS_OSPF: type = "I"; break; case RTS_OSPF_IA: type = "IA"; break; case RTS_OSPF_EXT1: type = "E1"; break; case RTS_OSPF_EXT2: type = "E2"; break; } buf += bsprintf(buf, " %s", type); buf += bsprintf(buf, " (%d/%d", rte->pref, rte->u.ospf.metric1); if (rte->attrs->source == RTS_OSPF_EXT2) buf += bsprintf(buf, "/%d", rte->u.ospf.metric2); buf += bsprintf(buf, ")"); if ((rte->attrs->source == RTS_OSPF_EXT1 || rte->attrs->source == RTS_OSPF_EXT2) && rte->u.ospf.tag) { buf += bsprintf(buf, " [%x]", rte->u.ospf.tag); } if (rte->u.ospf.router_id) buf += bsprintf(buf, " [%R]", rte->u.ospf.router_id); } static int ospf_get_attr(eattr * a, byte * buf, int buflen UNUSED) { switch (a->id) { case EA_OSPF_METRIC1: bsprintf(buf, "metric1"); return GA_NAME; case EA_OSPF_METRIC2: bsprintf(buf, "metric2"); return GA_NAME; case EA_OSPF_TAG: bsprintf(buf, "tag: 0x%08x", a->u.data); return GA_FULL; case EA_OSPF_ROUTER_ID: bsprintf(buf, "router_id"); return GA_NAME; default: return GA_UNKNOWN; } } static void ospf_area_reconfigure(struct ospf_area *oa, struct ospf_area_config *nac) { oa->ac = nac; // FIXME NSSA check type ospf_ifaces_reconfigure(oa, nac); /* Handle net_list */ fib_free(&oa->net_fib); fib_free(&oa->enet_fib); add_area_nets(oa, nac); /* No need to handle stubnet_list */ oa->marked = 0; schedule_rt_lsa(oa); } /** * ospf_reconfigure - reconfiguration hook * @p: current instance of protocol (with old configuration) * @c: new configuration requested by user * * This hook tries to be a little bit intelligent. Instance of OSPF * will survive change of many constants like hello interval, * password change, addition or deletion of some neighbor on * nonbroadcast network, cost of interface, etc. */ static int ospf_reconfigure(struct proto *p, struct proto_config *c) { struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_config *old = (struct ospf_config *) (p->cf); struct ospf_config *new = (struct ospf_config *) c; struct ospf_area_config *nac; struct ospf_area *oa, *oax; struct ospf_iface *ifa, *ifx; struct ospf_iface_patt *ip; if (po->rfc1583 != new->rfc1583) return 0; if (old->abr != new->abr) return 0; po->ecmp = new->ecmp; po->tick = new->tick; po->disp_timer->recurrent = po->tick; tm_start(po->disp_timer, 1); /* Mark all areas and ifaces */ WALK_LIST(oa, po->area_list) oa->marked = 1; WALK_LIST(ifa, po->iface_list) ifa->marked = 1; /* Add and update areas */ WALK_LIST(nac, new->area_list) { oa = ospf_find_area(po, nac->areaid); if (oa) ospf_area_reconfigure(oa, nac); else ospf_area_add(po, nac, 1); } /* Add and update vlinks */ WALK_LIST(ip, new->vlink_list) { ifa = ospf_find_vlink(po, ip->voa, ip->vid); if (ifa) ospf_iface_reconfigure(ifa, ip); else ospf_iface_new(po->backbone, NULL, ip); } /* Delete remaining ifaces and areas */ WALK_LIST_DELSAFE(ifa, ifx, po->iface_list) if (ifa->marked) { ospf_iface_shutdown(ifa); ospf_iface_remove(ifa); } WALK_LIST_DELSAFE(oa, oax, po->area_list) if (oa->marked) ospf_area_remove(oa); schedule_rtcalc(po); return 1; } void ospf_sh_neigh(struct proto *p, char *iff) { struct ospf_iface *ifa = NULL; struct ospf_neighbor *n; struct proto_ospf *po = (struct proto_ospf *) p; if (p->proto_state != PS_UP) { cli_msg(-1013, "%s: is not up", p->name); cli_msg(0, ""); return; } cli_msg(-1013, "%s:", p->name); cli_msg(-1013, "%-12s\t%3s\t%-15s\t%-5s\t%-10s %-12s", "Router ID", "Pri", " State", "DTime", "Interface", "Router IP"); WALK_LIST(ifa, po->iface_list) if ((iff == NULL) || patmatch(iff, ifa->iface->name)) WALK_LIST(n, ifa->neigh_list) ospf_sh_neigh_info(n); cli_msg(0, ""); } void ospf_sh(struct proto *p) { struct ospf_area *oa; struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_iface *ifa; struct ospf_neighbor *n; int ifano, nno, adjno, firstfib; struct area_net *anet; if (p->proto_state != PS_UP) { cli_msg(-1014, "%s: is not up", p->name); cli_msg(0, ""); return; } cli_msg(-1014, "%s:", p->name); cli_msg(-1014, "RFC1583 compatibility: %s", (po->rfc1583 ? "enable" : "disabled")); cli_msg(-1014, "RT scheduler tick: %d", po->tick); cli_msg(-1014, "Number of areas: %u", po->areano); cli_msg(-1014, "Number of LSAs in DB:\t%u", po->gr->hash_entries); WALK_LIST(oa, po->area_list) { cli_msg(-1014, "\tArea: %R (%u) %s", oa->areaid, oa->areaid, oa->areaid == 0 ? "[BACKBONE]" : ""); ifano = 0; nno = 0; adjno = 0; WALK_LIST(ifa, po->iface_list) { if (oa == ifa->oa) { ifano++; WALK_LIST(n, ifa->neigh_list) { nno++; if (n->state == NEIGHBOR_FULL) adjno++; } } } cli_msg(-1014, "\t\tStub:\t%s", oa_is_stub(oa) ? "Yes" : "No"); cli_msg(-1014, "\t\tNSSA:\t%s", oa_is_nssa(oa) ? "Yes" : "No"); cli_msg(-1014, "\t\tTransit:\t%s", oa->trcap ? "Yes" : "No"); if (oa_is_nssa(oa)) cli_msg(-1014, "\t\tNSSA translation:\t%s%s", oa->translate ? "Yes" : "No", oa->translate == TRANS_WAIT ? " (run down)" : ""); cli_msg(-1014, "\t\tNumber of interfaces:\t%u", ifano); cli_msg(-1014, "\t\tNumber of neighbors:\t%u", nno); cli_msg(-1014, "\t\tNumber of adjacent neighbors:\t%u", adjno); firstfib = 1; FIB_WALK(&oa->net_fib, nftmp) { anet = (struct area_net *) nftmp; if(firstfib) { cli_msg(-1014, "\t\tArea networks:"); firstfib = 0; } cli_msg(-1014, "\t\t\t%1I/%u\t%s\t%s", anet->fn.prefix, anet->fn.pxlen, anet->hidden ? "Hidden" : "Advertise", anet->active ? "Active" : ""); } FIB_WALK_END; firstfib = 1; FIB_WALK(&oa->enet_fib, nftmp) { anet = (struct area_net *) nftmp; if(firstfib) { cli_msg(-1014, "\t\tArea external networks:"); firstfib = 0; } cli_msg(-1014, "\t\t\t%1I/%u\t%s\t%s", anet->fn.prefix, anet->fn.pxlen, anet->hidden ? "Hidden" : "Advertise", anet->active ? "Active" : ""); } FIB_WALK_END; } cli_msg(0, ""); } void ospf_sh_iface(struct proto *p, char *iff) { struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_iface *ifa = NULL; if (p->proto_state != PS_UP) { cli_msg(-1015, "%s: is not up", p->name); cli_msg(0, ""); return; } cli_msg(-1015, "%s:", p->name); WALK_LIST(ifa, po->iface_list) if ((iff == NULL) || patmatch(iff, ifa->iface->name)) ospf_iface_info(ifa); cli_msg(0, ""); } /* lsa_compare_for_state() - Compare function for 'show ospf state' * * First we want to separate network-LSAs and other LSAs (because network-LSAs * will be presented as network nodes and other LSAs together as router nodes) * Network-LSAs are sorted according to network prefix, other LSAs are sorted * according to originating router id (to get all LSA needed to represent one * router node together). Then, according to LSA type, ID and age. * * For OSPFv3, we have to handle also Prefix-LSAs. We would like to put each * immediately after the referenced LSA. We will make faked LSA based on ref_ * values */ #ifdef OSPFv3 static struct ospf_lsa_header * fake_lsa_from_prefix_lsa(struct ospf_lsa_header *dst, struct ospf_lsa_header *src, struct ospf_lsa_prefix *px) { dst->age = src->age; dst->type = px->ref_type; dst->id = px->ref_id; dst->rt = px->ref_rt; dst->sn = src->sn; return dst; } #endif static int lsa_compare_for_state(const void *p1, const void *p2) { struct top_hash_entry * he1 = * (struct top_hash_entry **) p1; struct top_hash_entry * he2 = * (struct top_hash_entry **) p2; struct ospf_lsa_header *lsa1 = &(he1->lsa); struct ospf_lsa_header *lsa2 = &(he2->lsa); if (he1->domain != he2->domain) return he1->domain - he2->domain; #ifdef OSPFv3 struct ospf_lsa_header lsatmp1, lsatmp2; int px1 = (lsa1->type == LSA_T_PREFIX); int px2 = (lsa2->type == LSA_T_PREFIX); if (px1) lsa1 = fake_lsa_from_prefix_lsa(&lsatmp1, lsa1, he1->lsa_body); if (px2) lsa2 = fake_lsa_from_prefix_lsa(&lsatmp2, lsa2, he2->lsa_body); #endif int nt1 = (lsa1->type == LSA_T_NET); int nt2 = (lsa2->type == LSA_T_NET); if (nt1 != nt2) return nt1 - nt2; if (nt1) { #ifdef OSPFv3 /* In OSPFv3, neworks are named base on ID of DR */ if (lsa1->rt != lsa2->rt) return lsa1->rt - lsa2->rt; #endif /* For OSPFv2, this is IP of the network, for OSPFv3, this is interface ID */ if (lsa1->id != lsa2->id) return lsa1->id - lsa2->id; #ifdef OSPFv3 if (px1 != px2) return px1 - px2; #endif return lsa1->sn - lsa2->sn; } else { if (lsa1->rt != lsa2->rt) return lsa1->rt - lsa2->rt; if (lsa1->type != lsa2->type) return lsa1->type - lsa2->type; if (lsa1->id != lsa2->id) return lsa1->id - lsa2->id; #ifdef OSPFv3 if (px1 != px2) return px1 - px2; #endif return lsa1->sn - lsa2->sn; } } static int ext_compare_for_state(const void *p1, const void *p2) { struct top_hash_entry * he1 = * (struct top_hash_entry **) p1; struct top_hash_entry * he2 = * (struct top_hash_entry **) p2; struct ospf_lsa_header *lsa1 = &(he1->lsa); struct ospf_lsa_header *lsa2 = &(he2->lsa); if (lsa1->rt != lsa2->rt) return lsa1->rt - lsa2->rt; if (lsa1->id != lsa2->id) return lsa1->id - lsa2->id; return lsa1->sn - lsa2->sn; } static inline void show_lsa_distance(struct top_hash_entry *he) { if (he->color == INSPF) cli_msg(-1016, "\t\tdistance %u", he->dist); else cli_msg(-1016, "\t\tunreachable"); } static inline void show_lsa_router(struct proto_ospf *po, struct top_hash_entry *he, int first, int verbose) { struct ospf_lsa_header *lsa = &(he->lsa); struct ospf_lsa_rt *rt = he->lsa_body; struct ospf_lsa_rt_link *rr = (struct ospf_lsa_rt_link *) (rt + 1); int max = lsa_rt_count(lsa); int i; if (first) { cli_msg(-1016, ""); cli_msg(-1016, "\trouter %R", he->lsa.rt); show_lsa_distance(he); } for (i = 0; i < max; i++) if (rr[i].type == LSART_VLNK) cli_msg(-1016, "\t\tvlink %R metric %u", rr[i].id, rr[i].metric); for (i = 0; i < max; i++) if (rr[i].type == LSART_PTP) cli_msg(-1016, "\t\trouter %R metric %u", rr[i].id, rr[i].metric); for (i = 0; i < max; i++) if (rr[i].type == LSART_NET) { #ifdef OSPFv2 struct top_hash_entry *net_he = ospf_hash_find_net(po->gr, he->domain, rr[i].id); if (net_he) { struct ospf_lsa_header *net_lsa = &(net_he->lsa); struct ospf_lsa_net *net_ln = net_he->lsa_body; cli_msg(-1016, "\t\tnetwork %I/%d metric %u", ipa_and(ipa_from_u32(net_lsa->id), net_ln->netmask), ipa_mklen(net_ln->netmask), rr[i].metric); } else cli_msg(-1016, "\t\tnetwork [%R] metric %u", rr[i].id, rr[i].metric); #else /* OSPFv3 */ cli_msg(-1016, "\t\tnetwork [%R-%u] metric %u", rr[i].id, rr[i].nif, rr[i].metric); #endif } #ifdef OSPFv2 if (!verbose) return; for (i = 0; i < max; i++) if (rr[i].type == LSART_STUB) cli_msg(-1016, "\t\tstubnet %I/%d metric %u", ipa_from_u32(rr[i].id), ipa_mklen(ipa_from_u32(rr[i].data)), rr[i].metric); #endif } static inline void show_lsa_network(struct top_hash_entry *he) { struct ospf_lsa_header *lsa = &(he->lsa); struct ospf_lsa_net *ln = he->lsa_body; u32 i; #ifdef OSPFv2 cli_msg(-1016, ""); cli_msg(-1016, "\tnetwork %I/%d", ipa_and(ipa_from_u32(lsa->id), ln->netmask), ipa_mklen(ln->netmask)); cli_msg(-1016, "\t\tdr %R", lsa->rt); #else /* OSPFv3 */ cli_msg(-1016, ""); cli_msg(-1016, "\tnetwork [%R-%u]", lsa->rt, lsa->id); #endif show_lsa_distance(he); for (i = 0; i < lsa_net_count(lsa); i++) cli_msg(-1016, "\t\trouter %R", ln->routers[i]); } static inline void show_lsa_sum_net(struct top_hash_entry *he) { ip_addr ip; int pxlen; #ifdef OSPFv2 struct ospf_lsa_sum *ls = he->lsa_body; pxlen = ipa_mklen(ls->netmask); ip = ipa_and(ipa_from_u32(he->lsa.id), ls->netmask); #else /* OSPFv3 */ u8 pxopts; u16 rest; struct ospf_lsa_sum_net *ls = he->lsa_body; lsa_get_ipv6_prefix(ls->prefix, &ip, &pxlen, &pxopts, &rest); #endif cli_msg(-1016, "\t\txnetwork %I/%d metric %u", ip, pxlen, ls->metric); } static inline void show_lsa_sum_rt(struct top_hash_entry *he) { u32 dst_rid, options; #ifdef OSPFv2 struct ospf_lsa_sum *ls = he->lsa_body; dst_rid = he->lsa.id; options = 0; #else /* OSPFv3 */ struct ospf_lsa_sum_rt *ls = he->lsa_body; dst_rid = ls->drid; options = ls->options & OPTIONS_MASK; #endif cli_msg(-1016, "\t\txrouter %R metric %u", dst_rid, ls->metric); } static inline void show_lsa_external(struct top_hash_entry *he) { struct ospf_lsa_ext *ext = he->lsa_body; char str_via[STD_ADDRESS_P_LENGTH + 8] = ""; char str_tag[16] = ""; ip_addr ip, rt_fwaddr; int pxlen, ebit, rt_fwaddr_valid; u32 rt_tag, rt_metric; if (he->lsa.type == LSA_T_EXT) he->domain = 0; /* Unmark the LSA */ rt_metric = ext->metric & METRIC_MASK; ebit = ext->metric & LSA_EXT_EBIT; #ifdef OSPFv2 ip = ipa_and(ipa_from_u32(he->lsa.id), ext->netmask); pxlen = ipa_mklen(ext->netmask); rt_fwaddr = ext->fwaddr; rt_fwaddr_valid = !ipa_equal(rt_fwaddr, IPA_NONE); rt_tag = ext->tag; #else /* OSPFv3 */ u8 pxopts; u16 rest; u32 *buf = ext->rest; buf = lsa_get_ipv6_prefix(buf, &ip, &pxlen, &pxopts, &rest); rt_fwaddr_valid = ext->metric & LSA_EXT_FBIT; if (rt_fwaddr_valid) buf = lsa_get_ipv6_addr(buf, &rt_fwaddr); else rt_fwaddr = IPA_NONE; if (ext->metric & LSA_EXT_TBIT) rt_tag = *buf++; else rt_tag = 0; #endif if (rt_fwaddr_valid) bsprintf(str_via, " via %I", rt_fwaddr); if (rt_tag) bsprintf(str_tag, " tag %08x", rt_tag); cli_msg(-1016, "\t\t%s %I/%d metric%s %u%s%s", (he->lsa.type == LSA_T_NSSA) ? "nssa-ext" : "external", ip, pxlen, ebit ? "2" : "", rt_metric, str_via, str_tag); } #ifdef OSPFv3 static inline void show_lsa_prefix(struct top_hash_entry *he, struct ospf_lsa_header *cnode) { struct ospf_lsa_prefix *px = he->lsa_body; ip_addr pxa; int pxlen; u8 pxopts; u16 metric; u32 *buf; int i; /* We check whether given prefix-LSA is related to the current node */ if ((px->ref_type != cnode->type) || (px->ref_rt != cnode->rt)) return; if ((px->ref_type == LSA_T_RT) && (px->ref_id != 0)) return; if ((px->ref_type == LSA_T_NET) && (px->ref_id != cnode->id)) return; buf = px->rest; for (i = 0; i < px->pxcount; i++) { buf = lsa_get_ipv6_prefix(buf, &pxa, &pxlen, &pxopts, &metric); if (px->ref_type == LSA_T_RT) cli_msg(-1016, "\t\tstubnet %I/%d metric %u", pxa, pxlen, metric); else cli_msg(-1016, "\t\taddress %I/%d", pxa, pxlen); } } #endif void ospf_sh_state(struct proto *p, int verbose, int reachable) { struct proto_ospf *po = (struct proto_ospf *) p; struct ospf_lsa_header *cnode = NULL; int num = po->gr->hash_entries; unsigned int i, ix, j1, j2, jx; u32 last_area = 0xFFFFFFFF; if (p->proto_state != PS_UP) { cli_msg(-1016, "%s: is not up", p->name); cli_msg(0, ""); return; } /* We store interesting area-scoped LSAs in array hea and global-scoped (LSA_T_EXT) LSAs in array hex */ struct top_hash_entry *hea[num]; struct top_hash_entry *hex[verbose ? num : 0]; struct top_hash_entry *he; j1 = j2 = jx = 0; WALK_SLIST(he, po->lsal) { int accept; switch (he->lsa.type) { case LSA_T_RT: case LSA_T_NET: accept = 1; break; case LSA_T_SUM_NET: case LSA_T_SUM_RT: case LSA_T_NSSA: #ifdef OSPFv3 case LSA_T_PREFIX: #endif accept = verbose; break; case LSA_T_EXT: if (verbose) { he->domain = 1; /* Abuse domain field to mark the LSA */ hex[jx++] = he; } default: accept = 0; } if (accept) hea[j1++] = he; else j2++; } if ((j1 + j2) != num) die("Fatal mismatch"); qsort(hea, j1, sizeof(struct top_hash_entry *), lsa_compare_for_state); qsort(hex, jx, sizeof(struct top_hash_entry *), ext_compare_for_state); /* * This code is a bit tricky, we have a primary LSAs (router and * network) that are presented as a node, and secondary LSAs that * are presented as a part of a primary node. cnode represents an * currently opened node (whose header was presented). The LSAs are * sorted to get secondary LSAs just after related primary LSA (if * available). We present secondary LSAs only when related primary * LSA is opened. * * AS-external LSAs are stored separately as they might be presented * several times (for each area when related ASBR is opened). When * the node is closed, related external routes are presented. We * also have to take into account that in OSPFv3, there might be * more router-LSAs and only the first should be considered as a * primary. This is handled by not closing old router-LSA when next * one is processed (which is not opened because there is already * one opened). */ ix = 0; for (i = 0; i < j1; i++) { he = hea[i]; /* If there is no opened node, we open the LSA (if appropriate) or skip to the next one */ if (!cnode) { if (((he->lsa.type == LSA_T_RT) || (he->lsa.type == LSA_T_NET)) && ((he->color == INSPF) || !reachable)) { cnode = &(he->lsa); if (he->domain != last_area) { cli_msg(-1016, ""); cli_msg(-1016, "area %R", he->domain); last_area = he->domain; ix = 0; } } else continue; } ASSERT(cnode && (he->domain == last_area) && (he->lsa.rt == cnode->rt)); switch (he->lsa.type) { case LSA_T_RT: show_lsa_router(po, he, he->lsa.id == cnode->id, verbose); break; case LSA_T_NET: show_lsa_network(he); break; case LSA_T_SUM_NET: if (cnode->type == LSA_T_RT) show_lsa_sum_net(he); break; case LSA_T_SUM_RT: if (cnode->type == LSA_T_RT) show_lsa_sum_rt(he); break; #ifdef OSPFv3 case LSA_T_PREFIX: show_lsa_prefix(he, cnode); break; #endif case LSA_T_EXT: case LSA_T_NSSA: show_lsa_external(he); break; } /* In these cases, we close the current node */ if ((i+1 == j1) || (hea[i+1]->domain != last_area) || (hea[i+1]->lsa.rt != cnode->rt) || (hea[i+1]->lsa.type == LSA_T_NET)) { while ((ix < jx) && (hex[ix]->lsa.rt < cnode->rt)) ix++; while ((ix < jx) && (hex[ix]->lsa.rt == cnode->rt)) show_lsa_external(hex[ix++]); cnode = NULL; } } int hdr = 0; u32 last_rt = 0xFFFFFFFF; for (ix = 0; ix < jx; ix++) { he = hex[ix]; /* If it is still marked, we show it now. */ if (he->domain) { he->domain = 0; if ((he->color != INSPF) && reachable) continue; if (!hdr) { cli_msg(-1016, ""); cli_msg(-1016, "other ASBRs"); hdr = 1; } if (he->lsa.rt != last_rt) { cli_msg(-1016, ""); cli_msg(-1016, "\trouter %R", he->lsa.rt); last_rt = he->lsa.rt; } show_lsa_external(he); } } cli_msg(0, ""); } static int lsa_compare_for_lsadb(const void *p1, const void *p2) { struct top_hash_entry * he1 = * (struct top_hash_entry **) p1; struct top_hash_entry * he2 = * (struct top_hash_entry **) p2; struct ospf_lsa_header *lsa1 = &(he1->lsa); struct ospf_lsa_header *lsa2 = &(he2->lsa); int sc1 = LSA_SCOPE(lsa1); int sc2 = LSA_SCOPE(lsa2); if (sc1 != sc2) return sc2 - sc1; if (he1->domain != he2->domain) return he1->domain - he2->domain; if (lsa1->rt != lsa2->rt) return lsa1->rt - lsa2->rt; if (lsa1->id != lsa2->id) return lsa1->id - lsa2->id; if (lsa1->type != lsa2->type) return lsa1->type - lsa2->type; return lsa1->sn - lsa2->sn; } void ospf_sh_lsadb(struct proto *p) { struct proto_ospf *po = (struct proto_ospf *) p; int num = po->gr->hash_entries; unsigned int i, j; int last_dscope = -1; u32 last_domain = 0; if (p->proto_state != PS_UP) { cli_msg(-1017, "%s: is not up", p->name); cli_msg(0, ""); return; } struct top_hash_entry *hea[num]; struct top_hash_entry *he; j = 0; WALK_SLIST(he, po->lsal) hea[j++] = he; if (j != num) die("Fatal mismatch"); qsort(hea, j, sizeof(struct top_hash_entry *), lsa_compare_for_lsadb); for (i = 0; i < j; i++) { struct ospf_lsa_header *lsa = &(hea[i]->lsa); int dscope = LSA_SCOPE(lsa); if ((dscope != last_dscope) || (hea[i]->domain != last_domain)) { cli_msg(-1017, ""); switch (dscope) { case LSA_SCOPE_AS: cli_msg(-1017, "Global"); break; case LSA_SCOPE_AREA: cli_msg(-1017, "Area %R", hea[i]->domain); break; #ifdef OSPFv3 case LSA_SCOPE_LINK: { struct iface *ifa = if_find_by_index(hea[i]->domain); cli_msg(-1017, "Link %s", (ifa != NULL) ? ifa->name : "?"); } break; #endif } cli_msg(-1017, ""); cli_msg(-1017," Type LS ID Router Age Sequence Checksum"); last_dscope = dscope; last_domain = hea[i]->domain; } cli_msg(-1017," %04x %-15R %-15R %5u %08x %04x", lsa->type, lsa->id, lsa->rt, lsa->age, lsa->sn, lsa->checksum); } cli_msg(0, ""); } struct protocol proto_ospf = { name: "OSPF", template: "ospf%d", attr_class: EAP_OSPF, init: ospf_init, dump: ospf_dump, start: ospf_start, shutdown: ospf_shutdown, reconfigure: ospf_reconfigure, get_status: ospf_get_status, get_attr: ospf_get_attr, get_route_info: ospf_get_route_info // show_proto_info: ospf_sh };