frr/isisd/isis_spf.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

2749 lines
72 KiB
C
Raw Normal View History

2003-12-23 09:09:43 +01:00
/*
* IS-IS Rout(e)ing protocol - isis_spf.c
* The SPT algorithm
*
* Copyright (C) 2001,2002 Sampo Saaristo
* Tampere University of Technology
* Institute of Communications Engineering
* Copyright (C) 2017 Christian Franke <chris@opensourcerouting.org>
2003-12-23 09:09:43 +01:00
*
* This program is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public Licenseas published by the Free
* Software Foundation; either version 2 of the License, or (at your option)
* any later version.
*
* This program is distributed in the hope that it will be useful,but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
* more details.
*
* You should have received a copy of the GNU General Public License along
* with this program; see the file COPYING; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
2003-12-23 09:09:43 +01:00
*/
#include <zebra.h>
#include "thread.h"
#include "linklist.h"
#include "vty.h"
#include "log.h"
#include "command.h"
#include "termtable.h"
2003-12-23 09:09:43 +01:00
#include "memory.h"
#include "prefix.h"
#include "filter.h"
2003-12-23 09:09:43 +01:00
#include "if.h"
#include "hash.h"
2003-12-23 09:09:43 +01:00
#include "table.h"
#include "spf_backoff.h"
#include "srcdest_table.h"
#include "vrf.h"
2003-12-23 09:09:43 +01:00
#include "isis_errors.h"
2003-12-23 09:09:43 +01:00
#include "isis_constants.h"
#include "isis_common.h"
2012-03-24 16:35:20 +01:00
#include "isis_flags.h"
2003-12-23 09:09:43 +01:00
#include "isisd.h"
#include "isis_misc.h"
#include "isis_adjacency.h"
#include "isis_circuit.h"
#include "isis_pdu.h"
#include "isis_lsp.h"
#include "isis_dynhn.h"
#include "isis_spf.h"
#include "isis_route.h"
#include "isis_csm.h"
#include "isis_mt.h"
#include "isis_tlvs.h"
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
#include "isis_zebra.h"
#include "fabricd.h"
#include "isis_spf_private.h"
2003-12-23 09:09:43 +01:00
DEFINE_MTYPE_STATIC(ISISD, ISIS_SPFTREE, "ISIS SPFtree");
DEFINE_MTYPE_STATIC(ISISD, ISIS_SPF_RUN, "ISIS SPF Run Info");
DEFINE_MTYPE_STATIC(ISISD, ISIS_SPF_ADJ, "ISIS SPF Adjacency");
DEFINE_MTYPE_STATIC(ISISD, ISIS_VERTEX, "ISIS vertex");
DEFINE_MTYPE_STATIC(ISISD, ISIS_VERTEX_ADJ, "ISIS SPF Vertex Adjacency");
static void spf_adj_list_parse_lsp(struct isis_spftree *spftree,
struct list *adj_list, struct isis_lsp *lsp,
const uint8_t *pseudo_nodeid,
uint32_t pseudo_metric);
/*
* supports the given af ?
*/
static bool speaks(uint8_t *protocols, uint8_t count, int family)
{
for (uint8_t i = 0; i < count; i++) {
if (family == AF_INET && protocols[i] == NLPID_IP)
return true;
if (family == AF_INET6 && protocols[i] == NLPID_IPV6)
return true;
}
return false;
}
struct isis_spf_run {
struct isis_area *area;
int level;
};
2003-12-23 09:09:43 +01:00
/* 7.2.7 */
static void remove_excess_adjs(struct list *adjs)
{
struct listnode *node, *excess = NULL;
struct isis_vertex_adj *vadj, *candidate = NULL;
2003-12-23 09:09:43 +01:00
int comp;
for (ALL_LIST_ELEMENTS_RO(adjs, node, vadj)) {
struct isis_adjacency *adj, *candidate_adj;
adj = vadj->sadj->adj;
assert(adj);
if (excess == NULL)
excess = node;
candidate = listgetdata(excess);
candidate_adj = candidate->sadj->adj;
if (candidate_adj->sys_type < adj->sys_type) {
excess = node;
continue;
}
if (candidate_adj->sys_type > adj->sys_type)
continue;
comp = memcmp(candidate_adj->sysid, adj->sysid,
ISIS_SYS_ID_LEN);
if (comp > 0) {
excess = node;
continue;
}
if (comp < 0)
continue;
2003-12-23 09:09:43 +01:00
if (candidate_adj->circuit->idx > adj->circuit->idx) {
excess = node;
continue;
}
if (candidate_adj->circuit->idx < adj->circuit->idx)
continue;
comp = memcmp(candidate_adj->snpa, adj->snpa, ETH_ALEN);
if (comp > 0) {
excess = node;
continue;
}
}
2003-12-23 09:09:43 +01:00
list_delete_node(adjs, excess);
return;
}
const char *vtype2string(enum vertextype vtype)
2003-12-23 09:09:43 +01:00
{
switch (vtype) {
case VTYPE_PSEUDO_IS:
return "pseudo_IS";
case VTYPE_PSEUDO_TE_IS:
return "pseudo_TE-IS";
case VTYPE_NONPSEUDO_IS:
return "IS";
case VTYPE_NONPSEUDO_TE_IS:
return "TE-IS";
case VTYPE_ES:
return "ES";
case VTYPE_IPREACH_INTERNAL:
return "IP internal";
case VTYPE_IPREACH_EXTERNAL:
return "IP external";
case VTYPE_IPREACH_TE:
return "IP TE";
case VTYPE_IP6REACH_INTERNAL:
return "IP6 internal";
case VTYPE_IP6REACH_EXTERNAL:
return "IP6 external";
default:
return "UNKNOWN";
}
return NULL; /* Not reached */
2003-12-23 09:09:43 +01:00
}
const char *vid2string(const struct isis_vertex *vertex, char *buff, int size)
2003-12-23 09:09:43 +01:00
{
if (VTYPE_IS(vertex->type) || VTYPE_ES(vertex->type)) {
const char *hostname = print_sys_hostname(vertex->N.id);
strlcpy(buff, hostname, size);
return buff;
}
if (VTYPE_IP(vertex->type)) {
srcdest2str(&vertex->N.ip.p.dest, &vertex->N.ip.p.src, buff,
size);
return buff;
}
return "UNKNOWN";
2003-12-23 09:09:43 +01:00
}
static bool prefix_sid_cmp(const void *value1, const void *value2)
{
const struct isis_vertex *c1 = value1;
const struct isis_vertex *c2 = value2;
if (CHECK_FLAG(c1->N.ip.sr.sid.flags,
ISIS_PREFIX_SID_VALUE | ISIS_PREFIX_SID_LOCAL)
!= CHECK_FLAG(c2->N.ip.sr.sid.flags,
ISIS_PREFIX_SID_VALUE | ISIS_PREFIX_SID_LOCAL))
return false;
return c1->N.ip.sr.sid.value == c2->N.ip.sr.sid.value;
}
static unsigned int prefix_sid_key_make(const void *value)
{
const struct isis_vertex *vertex = value;
return jhash_1word(vertex->N.ip.sr.sid.value, 0);
}
struct isis_vertex *isis_spf_prefix_sid_lookup(struct isis_spftree *spftree,
struct isis_prefix_sid *psid)
{
struct isis_vertex lookup = {};
lookup.N.ip.sr.sid = *psid;
return hash_lookup(spftree->prefix_sids, &lookup);
}
void isis_vertex_adj_free(void *arg)
{
struct isis_vertex_adj *vadj = arg;
XFREE(MTYPE_ISIS_VERTEX_ADJ, vadj);
}
static struct isis_vertex *isis_vertex_new(struct isis_spftree *spftree,
void *id,
enum vertextype vtype)
{
struct isis_vertex *vertex;
vertex = XCALLOC(MTYPE_ISIS_VERTEX, sizeof(struct isis_vertex));
isis_vertex_id_init(vertex, id, vtype);
2012-03-24 16:35:20 +01:00
vertex->Adj_N = list_new();
vertex->Adj_N->del = isis_vertex_adj_free;
2012-03-24 16:35:20 +01:00
vertex->parents = list_new();
if (CHECK_FLAG(spftree->flags, F_SPFTREE_HOPCOUNT_METRIC)) {
vertex->firsthops = hash_create(isis_vertex_queue_hash_key,
isis_vertex_queue_hash_cmp,
NULL);
}
2012-03-24 16:35:20 +01:00
return vertex;
2003-12-23 09:09:43 +01:00
}
void isis_vertex_del(struct isis_vertex *vertex)
{
list_delete(&vertex->Adj_N);
list_delete(&vertex->parents);
if (vertex->firsthops) {
hash_clean(vertex->firsthops, NULL);
hash_free(vertex->firsthops);
vertex->firsthops = NULL;
}
memset(vertex, 0, sizeof(struct isis_vertex));
XFREE(MTYPE_ISIS_VERTEX, vertex);
}
struct isis_vertex_adj *
isis_vertex_adj_add(struct isis_spftree *spftree, struct isis_vertex *vertex,
struct list *vadj_list, struct isis_spf_adj *sadj,
struct isis_prefix_sid *psid, bool last_hop)
{
struct isis_vertex_adj *vadj;
vadj = XCALLOC(MTYPE_ISIS_VERTEX_ADJ, sizeof(*vadj));
vadj->sadj = sadj;
if (spftree->area->srdb.enabled && psid) {
if (vertex->N.ip.sr.present
&& vertex->N.ip.sr.sid.value != psid->value)
zlog_warn(
"ISIS-SPF: ignoring different Prefix-SID for route %pFX",
&vertex->N.ip.p.dest);
else {
vadj->sr.sid = *psid;
vadj->sr.label = sr_prefix_out_label(
spftree->lspdb, vertex->N.ip.p.dest.family,
psid, sadj->id, last_hop);
if (vadj->sr.label != MPLS_INVALID_LABEL)
vadj->sr.present = true;
}
}
listnode_add(vadj_list, vadj);
return vadj;
}
2012-03-24 16:35:20 +01:00
static void isis_vertex_adj_del(struct isis_vertex *vertex,
struct isis_adjacency *adj)
{
struct isis_vertex_adj *vadj;
2012-03-24 16:35:20 +01:00
struct listnode *node, *nextnode;
2012-03-24 16:35:20 +01:00
if (!vertex)
return;
for (ALL_LIST_ELEMENTS(vertex->Adj_N, node, nextnode, vadj)) {
if (vadj->sadj->adj == adj) {
listnode_delete(vertex->Adj_N, vadj);
isis_vertex_adj_free(vadj);
}
2012-03-24 16:35:20 +01:00
}
return;
}
bool isis_vertex_adj_exists(const struct isis_spftree *spftree,
const struct isis_vertex *vertex,
const struct isis_spf_adj *sadj)
{
struct isis_vertex_adj *tmp;
struct listnode *node;
for (ALL_LIST_ELEMENTS_RO(vertex->Adj_N, node, tmp)) {
if (CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ADJACENCIES)) {
if (memcmp(sadj->id, tmp->sadj->id, sizeof(sadj->id))
== 0)
return true;
} else {
if (sadj->adj == tmp->sadj->adj)
return true;
}
}
return false;
}
static void isis_spf_adj_free(void *arg)
{
struct isis_spf_adj *sadj = arg;
XFREE(MTYPE_ISIS_SPF_ADJ, sadj);
}
struct isis_spftree *isis_spftree_new(struct isis_area *area,
struct lspdb_head *lspdb,
const uint8_t *sysid, int level,
enum spf_tree_id tree_id,
enum spf_type type, uint8_t flags)
2012-03-24 16:35:20 +01:00
{
struct isis_spftree *tree;
2012-03-24 16:35:20 +01:00
tree = XCALLOC(MTYPE_ISIS_SPFTREE, sizeof(struct isis_spftree));
isis_vertex_queue_init(&tree->tents, "IS-IS SPF tents", true);
isis_vertex_queue_init(&tree->paths, "IS-IS SPF paths", false);
tree->route_table = srcdest_table_init();
tree->route_table->cleanup = isis_route_node_cleanup;
tree->route_table_backup = srcdest_table_init();
tree->route_table_backup->cleanup = isis_route_node_cleanup;
2012-03-24 16:35:20 +01:00
tree->area = area;
tree->lspdb = lspdb;
tree->prefix_sids = hash_create(prefix_sid_key_make, prefix_sid_cmp,
"SR Prefix-SID Entries");
tree->sadj_list = list_new();
tree->sadj_list->del = isis_spf_adj_free;
2012-03-28 08:48:05 +02:00
tree->last_run_timestamp = 0;
tree->last_run_monotime = 0;
2012-03-28 08:48:05 +02:00
tree->last_run_duration = 0;
2012-03-24 16:35:20 +01:00
tree->runcount = 0;
tree->type = type;
memcpy(tree->sysid, sysid, ISIS_SYS_ID_LEN);
tree->level = level;
tree->tree_id = tree_id;
tree->family = (tree->tree_id == SPFTREE_IPV4) ? AF_INET : AF_INET6;
tree->flags = flags;
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
isis_rlfa_list_init(tree);
tree->lfa.remote.pc_spftrees = list_new();
tree->lfa.remote.pc_spftrees->del = (void (*)(void *))isis_spftree_del;
if (tree->type == SPF_TYPE_RLFA || tree->type == SPF_TYPE_TI_LFA) {
isis_spf_node_list_init(&tree->lfa.p_space);
isis_spf_node_list_init(&tree->lfa.q_space);
}
2012-03-24 16:35:20 +01:00
return tree;
}
2003-12-23 09:09:43 +01:00
void isis_spftree_del(struct isis_spftree *spftree)
{
hash_clean(spftree->prefix_sids, NULL);
hash_free(spftree->prefix_sids);
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
isis_zebra_rlfa_unregister_all(spftree);
isis_rlfa_list_clear(spftree);
list_delete(&spftree->lfa.remote.pc_spftrees);
if (spftree->type == SPF_TYPE_RLFA
|| spftree->type == SPF_TYPE_TI_LFA) {
isis_spf_node_list_clear(&spftree->lfa.q_space);
isis_spf_node_list_clear(&spftree->lfa.p_space);
}
isis_spf_node_list_clear(&spftree->adj_nodes);
list_delete(&spftree->sadj_list);
isis_vertex_queue_free(&spftree->tents);
isis_vertex_queue_free(&spftree->paths);
route_table_finish(spftree->route_table);
route_table_finish(spftree->route_table_backup);
spftree->route_table = NULL;
2003-12-23 09:09:43 +01:00
XFREE(MTYPE_ISIS_SPFTREE, spftree);
2003-12-23 09:09:43 +01:00
return;
}
2012-03-24 16:35:20 +01:00
static void isis_spftree_adj_del(struct isis_spftree *spftree,
struct isis_adjacency *adj)
{
struct listnode *node;
struct isis_vertex *v;
2012-03-24 16:35:20 +01:00
if (!adj)
return;
assert(!isis_vertex_queue_count(&spftree->tents));
for (ALL_QUEUE_ELEMENTS_RO(&spftree->paths, node, v))
isis_vertex_adj_del(v, adj);
2012-03-24 16:35:20 +01:00
return;
}
2003-12-23 09:09:43 +01:00
void spftree_area_init(struct isis_area *area)
{
for (int tree = SPFTREE_IPV4; tree < SPFTREE_COUNT; tree++) {
for (int level = ISIS_LEVEL1; level <= ISIS_LEVEL2; level++) {
if (!(area->is_type & level))
continue;
if (area->spftree[tree][level - 1])
continue;
area->spftree[tree][level - 1] =
isis_spftree_new(area, &area->lspdb[level - 1],
area->isis->sysid, level, tree,
SPF_TYPE_FORWARD, 0);
}
2012-03-24 16:35:20 +01:00
}
2003-12-23 09:09:43 +01:00
}
2012-03-24 16:35:20 +01:00
void spftree_area_del(struct isis_area *area)
2003-12-23 09:09:43 +01:00
{
for (int tree = SPFTREE_IPV4; tree < SPFTREE_COUNT; tree++) {
for (int level = ISIS_LEVEL1; level <= ISIS_LEVEL2; level++) {
if (!(area->is_type & level))
continue;
if (!area->spftree[tree][level - 1])
continue;
isis_spftree_del(area->spftree[tree][level - 1]);
}
}
2012-03-24 16:35:20 +01:00
}
static int spf_adj_state_change(struct isis_adjacency *adj)
2012-03-24 16:35:20 +01:00
{
struct isis_area *area = adj->circuit->area;
if (adj->adj_state == ISIS_ADJ_UP)
return 0;
/* Remove adjacency from all SPF trees. */
for (int tree = SPFTREE_IPV4; tree < SPFTREE_COUNT; tree++) {
for (int level = ISIS_LEVEL1; level <= ISIS_LEVEL2; level++) {
if (!(area->is_type & level))
continue;
if (!area->spftree[tree][level - 1])
continue;
isis_spftree_adj_del(area->spftree[tree][level - 1],
adj);
}
2012-03-24 16:35:20 +01:00
}
if (fabricd_spftree(area) != NULL)
isis_spftree_adj_del(fabricd_spftree(area), adj);
return 0;
2012-03-24 16:35:20 +01:00
}
/*
* Find the system LSP: returns the LSP in our LSP database
* associated with the given system ID.
*/
struct isis_lsp *isis_root_system_lsp(struct lspdb_head *lspdb,
const uint8_t *sysid)
2012-03-24 16:35:20 +01:00
{
2012-03-28 08:48:05 +02:00
struct isis_lsp *lsp;
uint8_t lspid[ISIS_SYS_ID_LEN + 2];
2012-03-24 16:35:20 +01:00
memcpy(lspid, sysid, ISIS_SYS_ID_LEN);
LSP_PSEUDO_ID(lspid) = 0;
LSP_FRAGMENT(lspid) = 0;
lsp = lsp_search(lspdb, lspid);
if (lsp && lsp->hdr.rem_lifetime != 0)
2012-03-28 08:48:05 +02:00
return lsp;
return NULL;
2003-12-23 09:09:43 +01:00
}
/*
* Add this IS to the root of SPT
*/
static struct isis_vertex *isis_spf_add_root(struct isis_spftree *spftree)
2003-12-23 09:09:43 +01:00
{
struct isis_vertex *vertex;
#ifdef EXTREME_DEBUG
char buff[VID2STR_BUFFER];
2003-12-23 09:09:43 +01:00
#endif /* EXTREME_DEBUG */
vertex = isis_vertex_new(spftree, spftree->sysid,
spftree->area->oldmetric
? VTYPE_NONPSEUDO_IS
: VTYPE_NONPSEUDO_TE_IS);
isis_vertex_queue_append(&spftree->paths, vertex);
2003-12-23 09:09:43 +01:00
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: added this IS %s %s depth %d dist %d to PATHS",
vtype2string(vertex->type),
vid2string(vertex, buff, sizeof(buff)), vertex->depth,
vertex->d_N);
2003-12-23 09:09:43 +01:00
#endif /* EXTREME_DEBUG */
2012-03-24 16:35:20 +01:00
return vertex;
2003-12-23 09:09:43 +01:00
}
static void vertex_add_parent_firsthop(struct hash_bucket *bucket, void *arg)
{
struct isis_vertex *vertex = arg;
struct isis_vertex *hop = bucket->data;
(void)hash_get(vertex->firsthops, hop, hash_alloc_intern);
}
static void vertex_update_firsthops(struct isis_vertex *vertex,
struct isis_vertex *parent)
{
if (vertex->d_N <= 2)
(void)hash_get(vertex->firsthops, vertex, hash_alloc_intern);
if (vertex->d_N < 2 || !parent)
return;
hash_iterate(parent->firsthops, vertex_add_parent_firsthop, vertex);
}
2003-12-23 09:09:43 +01:00
/*
* Add a vertex to TENT sorted by cost and by vertextype on tie break situation
*/
static struct isis_vertex *
isis_spf_add2tent(struct isis_spftree *spftree, enum vertextype vtype, void *id,
uint32_t cost, int depth, struct isis_spf_adj *sadj,
struct isis_prefix_sid *psid, struct isis_vertex *parent)
2003-12-23 09:09:43 +01:00
{
struct isis_vertex *vertex;
2003-12-23 09:09:43 +01:00
struct listnode *node;
bool last_hop;
char buff[VID2STR_BUFFER];
2003-12-23 09:09:43 +01:00
vertex = isis_find_vertex(&spftree->paths, id, vtype);
if (vertex != NULL) {
zlog_err(
"%s: vertex %s of type %s already in PATH; check for sysId collisions with established neighbors",
__func__, vid2string(vertex, buff, sizeof(buff)),
vtype2string(vertex->type));
return NULL;
}
vertex = isis_find_vertex(&spftree->tents, id, vtype);
if (vertex != NULL) {
zlog_err(
"%s: vertex %s of type %s already in TENT; check for sysId collisions with established neighbors",
__func__, vid2string(vertex, buff, sizeof(buff)),
vtype2string(vertex->type));
return NULL;
}
vertex = isis_vertex_new(spftree, id, vtype);
2003-12-23 09:09:43 +01:00
vertex->d_N = cost;
vertex->depth = depth;
if (VTYPE_IP(vtype) && spftree->area->srdb.enabled && psid) {
struct isis_area *area = spftree->area;
struct isis_vertex *vertex_psid;
/*
* Check if the Prefix-SID is already in use by another prefix.
*/
vertex_psid = isis_spf_prefix_sid_lookup(spftree, psid);
if (vertex_psid
&& !prefix_same(&vertex_psid->N.ip.p.dest,
&vertex->N.ip.p.dest)) {
flog_warn(
EC_ISIS_SID_COLLISION,
"ISIS-Sr (%s): collision detected, prefixes %pFX and %pFX share the same SID %s (%u)",
area->area_tag, &vertex->N.ip.p.dest,
&vertex_psid->N.ip.p.dest,
CHECK_FLAG(psid->flags, ISIS_PREFIX_SID_VALUE)
? "label"
: "index",
psid->value);
psid = NULL;
} else {
bool local;
local = (vertex->depth == 1);
vertex->N.ip.sr.sid = *psid;
vertex->N.ip.sr.label =
sr_prefix_in_label(area, psid, local);
if (vertex->N.ip.sr.label != MPLS_INVALID_LABEL)
vertex->N.ip.sr.present = true;
(void)hash_get(spftree->prefix_sids, vertex,
hash_alloc_intern);
}
}
2012-03-24 16:35:20 +01:00
if (parent) {
listnode_add(vertex->parents, parent);
}
if (CHECK_FLAG(spftree->flags, F_SPFTREE_HOPCOUNT_METRIC))
vertex_update_firsthops(vertex, parent);
last_hop = (vertex->depth == 2);
2012-03-24 16:35:20 +01:00
if (parent && parent->Adj_N && listcount(parent->Adj_N) > 0) {
struct isis_vertex_adj *parent_vadj;
for (ALL_LIST_ELEMENTS_RO(parent->Adj_N, node, parent_vadj))
isis_vertex_adj_add(spftree, vertex, vertex->Adj_N,
parent_vadj->sadj, psid, last_hop);
} else if (sadj) {
isis_vertex_adj_add(spftree, vertex, vertex->Adj_N, sadj, psid,
last_hop);
2012-03-24 16:35:20 +01:00
}
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: add to TENT %s %s %s depth %d dist %d adjcount %d",
print_sys_hostname(vertex->N.id),
vtype2string(vertex->type),
vid2string(vertex, buff, sizeof(buff)), vertex->depth,
vertex->d_N, listcount(vertex->Adj_N));
2003-12-23 09:09:43 +01:00
#endif /* EXTREME_DEBUG */
2012-03-24 16:35:20 +01:00
isis_vertex_queue_insert(&spftree->tents, vertex);
return vertex;
}
static void isis_spf_add_local(struct isis_spftree *spftree,
enum vertextype vtype, void *id,
struct isis_spf_adj *sadj, uint32_t cost,
struct isis_prefix_sid *psid,
2012-03-24 16:35:20 +01:00
struct isis_vertex *parent)
{
2003-12-23 09:09:43 +01:00
struct isis_vertex *vertex;
vertex = isis_find_vertex(&spftree->tents, id, vtype);
if (vertex) {
/* C.2.5 c) */
if (vertex->d_N == cost) {
if (sadj) {
bool last_hop = (vertex->depth == 2);
isis_vertex_adj_add(spftree, vertex,
vertex->Adj_N, sadj, psid,
last_hop);
}
/* d) */
if (!CHECK_FLAG(spftree->flags,
F_SPFTREE_NO_ADJACENCIES)
&& listcount(vertex->Adj_N) > ISIS_MAX_PATH_SPLITS)
remove_excess_adjs(vertex->Adj_N);
if (parent && (listnode_lookup(vertex->parents, parent)
== NULL))
2012-03-24 16:35:20 +01:00
listnode_add(vertex->parents, parent);
return;
} else if (vertex->d_N < cost) {
/* e) do nothing */
return;
} else { /* vertex->d_N > cost */
/* f) */
isis_vertex_queue_delete(&spftree->tents, vertex);
2012-03-24 16:35:20 +01:00
isis_vertex_del(vertex);
}
}
isis_spf_add2tent(spftree, vtype, id, cost, 1, sadj, psid, parent);
2012-03-24 16:35:20 +01:00
return;
2003-12-23 09:09:43 +01:00
}
static void process_N(struct isis_spftree *spftree, enum vertextype vtype,
void *id, uint32_t dist, uint16_t depth,
struct isis_prefix_sid *psid, struct isis_vertex *parent)
2003-12-23 09:09:43 +01:00
{
struct isis_vertex *vertex;
#ifdef EXTREME_DEBUG
char buff[VID2STR_BUFFER];
2003-12-23 09:09:43 +01:00
#endif
2012-03-24 16:35:20 +01:00
assert(spftree && parent);
if (CHECK_FLAG(spftree->flags, F_SPFTREE_HOPCOUNT_METRIC)
&& !VTYPE_IS(vtype))
return;
struct prefix_pair p;
if (vtype >= VTYPE_IPREACH_INTERNAL) {
memcpy(&p, id, sizeof(p));
apply_mask(&p.dest);
apply_mask(&p.src);
id = &p;
}
2012-03-24 16:35:20 +01:00
/* RFC3787 section 5.1 */
if (spftree->area->newmetric == 1) {
if (dist > MAX_WIDE_PATH_METRIC)
return;
}
2003-12-23 09:09:43 +01:00
/* C.2.6 b) */
2012-03-24 16:35:20 +01:00
else if (spftree->area->oldmetric == 1) {
if (dist > MAX_NARROW_PATH_METRIC)
return;
}
2003-12-23 09:09:43 +01:00
/* c) */
vertex = isis_find_vertex(&spftree->paths, id, vtype);
if (vertex) {
2003-12-23 09:09:43 +01:00
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: process_N %s %s %s dist %d already found from PATH",
print_sys_hostname(vertex->N.id),
vtype2string(vtype),
vid2string(vertex, buff, sizeof(buff)), dist);
2003-12-23 09:09:43 +01:00
#endif /* EXTREME_DEBUG */
assert(dist >= vertex->d_N);
return;
}
vertex = isis_find_vertex(&spftree->tents, id, vtype);
/* d) */
if (vertex) {
/* 1) */
2003-12-23 09:09:43 +01:00
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: process_N %s %s %s dist %d parent %s adjcount %d",
print_sys_hostname(vertex->N.id),
vtype2string(vtype),
vid2string(vertex, buff, sizeof(buff)), dist,
(parent ? print_sys_hostname(parent->N.id)
: "null"),
(parent ? listcount(parent->Adj_N) : 0));
2003-12-23 09:09:43 +01:00
#endif /* EXTREME_DEBUG */
if (vertex->d_N == dist) {
2012-03-24 16:35:20 +01:00
struct listnode *node;
struct isis_vertex_adj *parent_vadj;
2012-03-24 16:35:20 +01:00
for (ALL_LIST_ELEMENTS_RO(parent->Adj_N, node,
parent_vadj))
if (!isis_vertex_adj_exists(
spftree, vertex,
parent_vadj->sadj)) {
bool last_hop = (vertex->depth == 2);
isis_vertex_adj_add(spftree, vertex,
vertex->Adj_N,
parent_vadj->sadj,
psid, last_hop);
}
if (CHECK_FLAG(spftree->flags,
F_SPFTREE_HOPCOUNT_METRIC))
vertex_update_firsthops(vertex, parent);
/* 2) */
if (!CHECK_FLAG(spftree->flags,
F_SPFTREE_NO_ADJACENCIES)
&& listcount(vertex->Adj_N) > ISIS_MAX_PATH_SPLITS)
remove_excess_adjs(vertex->Adj_N);
2012-03-24 16:35:20 +01:00
if (listnode_lookup(vertex->parents, parent) == NULL)
listnode_add(vertex->parents, parent);
return;
} else if (vertex->d_N < dist) {
return;
/* 4) */
} else {
isis_vertex_queue_delete(&spftree->tents, vertex);
2012-03-24 16:35:20 +01:00
isis_vertex_del(vertex);
}
}
2012-03-24 16:35:20 +01:00
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: process_N add2tent %s %s dist %d parent %s",
print_sys_hostname(id), vtype2string(vtype), dist,
(parent ? print_sys_hostname(parent->N.id) : "null"));
2012-03-24 16:35:20 +01:00
#endif /* EXTREME_DEBUG */
isis_spf_add2tent(spftree, vtype, id, dist, depth, NULL, psid, parent);
2003-12-23 09:09:43 +01:00
return;
}
/*
* C.2.6 Step 1
*/
static int isis_spf_process_lsp(struct isis_spftree *spftree,
struct isis_lsp *lsp, uint32_t cost,
uint16_t depth, uint8_t *root_sysid,
2012-03-24 16:35:20 +01:00
struct isis_vertex *parent)
2003-12-23 09:09:43 +01:00
{
bool pseudo_lsp = LSP_PSEUDO_ID(lsp->hdr.lsp_id);
struct listnode *fragnode = NULL;
2012-03-24 16:35:20 +01:00
uint32_t dist;
2003-12-23 09:09:43 +01:00
enum vertextype vtype;
static const uint8_t null_sysid[ISIS_SYS_ID_LEN];
struct isis_mt_router_info *mt_router_info = NULL;
struct prefix_pair ip_info;
bool has_valid_psid;
if (isis_lfa_excise_node_check(spftree, lsp->hdr.lsp_id)) {
if (IS_DEBUG_LFA)
zlog_debug("ISIS-LFA: excising node %s",
print_sys_hostname(lsp->hdr.lsp_id));
return ISIS_OK;
}
if (!lsp->tlvs)
return ISIS_OK;
if (spftree->mtid != ISIS_MT_IPV4_UNICAST)
mt_router_info = isis_tlvs_lookup_mt_router_info(lsp->tlvs,
spftree->mtid);
if (!pseudo_lsp && (spftree->mtid == ISIS_MT_IPV4_UNICAST
&& !speaks(lsp->tlvs->protocols_supported.protocols,
lsp->tlvs->protocols_supported.count,
spftree->family))
&& !mt_router_info)
2003-12-23 09:09:43 +01:00
return ISIS_OK;
/* RFC3787 section 4 SHOULD ignore overload bit in pseudo LSPs */
bool no_overload = (pseudo_lsp
|| (spftree->mtid == ISIS_MT_IPV4_UNICAST
&& !ISIS_MASK_LSP_OL_BIT(lsp->hdr.lsp_bits))
|| (mt_router_info && !mt_router_info->overload));
lspfragloop:
if (lsp->hdr.seqno == 0) {
zlog_warn("%s: lsp with 0 seq_num - ignore", __func__);
return ISIS_WARNING;
2003-12-23 09:09:43 +01:00
}
2012-03-24 16:35:20 +01:00
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug("ISIS-SPF: process_lsp %s",
print_sys_hostname(lsp->hdr.lsp_id));
2012-03-24 16:35:20 +01:00
#endif /* EXTREME_DEBUG */
if (no_overload) {
if ((pseudo_lsp || spftree->mtid == ISIS_MT_IPV4_UNICAST)
&& spftree->area->oldmetric) {
struct isis_oldstyle_reach *r;
for (r = (struct isis_oldstyle_reach *)
lsp->tlvs->oldstyle_reach.head;
r; r = r->next) {
if (fabricd)
continue;
2012-03-24 16:35:20 +01:00
/* C.2.6 a) */
/* Two way connectivity */
if (!LSP_PSEUDO_ID(r->id)
&& !memcmp(r->id, root_sysid,
ISIS_SYS_ID_LEN))
continue;
if (!pseudo_lsp
&& !memcmp(r->id, null_sysid,
ISIS_SYS_ID_LEN))
2012-03-24 16:35:20 +01:00
continue;
dist = cost + r->metric;
process_N(spftree,
LSP_PSEUDO_ID(r->id)
? VTYPE_PSEUDO_IS
: VTYPE_NONPSEUDO_IS,
(void *)r->id, dist, depth + 1, NULL,
parent);
}
}
if (spftree->area->newmetric) {
struct isis_item_list *te_neighs = NULL;
if (pseudo_lsp || spftree->mtid == ISIS_MT_IPV4_UNICAST)
te_neighs = &lsp->tlvs->extended_reach;
else
te_neighs = isis_lookup_mt_items(
&lsp->tlvs->mt_reach, spftree->mtid);
struct isis_extended_reach *er;
for (er = te_neighs ? (struct isis_extended_reach *)
te_neighs->head
: NULL;
er; er = er->next) {
/* C.2.6 a) */
/* Two way connectivity */
if (!LSP_PSEUDO_ID(er->id)
&& !memcmp(er->id, root_sysid,
ISIS_SYS_ID_LEN))
continue;
if (!pseudo_lsp
&& !memcmp(er->id, null_sysid,
ISIS_SYS_ID_LEN))
continue;
dist = cost
+ (CHECK_FLAG(spftree->flags,
F_SPFTREE_HOPCOUNT_METRIC)
? 1
: er->metric);
process_N(spftree,
LSP_PSEUDO_ID(er->id)
? VTYPE_PSEUDO_TE_IS
: VTYPE_NONPSEUDO_TE_IS,
(void *)er->id, dist, depth + 1, NULL,
parent);
}
}
}
if (!fabricd && !pseudo_lsp && spftree->family == AF_INET
&& spftree->mtid == ISIS_MT_IPV4_UNICAST
&& spftree->area->oldmetric) {
struct isis_item_list *reachs[] = {
&lsp->tlvs->oldstyle_ip_reach,
&lsp->tlvs->oldstyle_ip_reach_ext};
for (unsigned int i = 0; i < array_size(reachs); i++) {
vtype = i ? VTYPE_IPREACH_EXTERNAL
: VTYPE_IPREACH_INTERNAL;
memset(&ip_info, 0, sizeof(ip_info));
ip_info.dest.family = AF_INET;
struct isis_oldstyle_ip_reach *r;
for (r = (struct isis_oldstyle_ip_reach *)reachs[i]
->head;
r; r = r->next) {
dist = cost + r->metric;
ip_info.dest.u.prefix4 = r->prefix.prefix;
ip_info.dest.prefixlen = r->prefix.prefixlen;
process_N(spftree, vtype, &ip_info,
dist, depth + 1, NULL, parent);
}
}
}
/* we can skip all the rest if we're using metric style narrow */
if (!spftree->area->newmetric)
goto end;
if (!pseudo_lsp && spftree->family == AF_INET) {
struct isis_item_list *ipv4_reachs;
if (spftree->mtid == ISIS_MT_IPV4_UNICAST)
ipv4_reachs = &lsp->tlvs->extended_ip_reach;
else
ipv4_reachs = isis_lookup_mt_items(
&lsp->tlvs->mt_ip_reach, spftree->mtid);
memset(&ip_info, 0, sizeof(ip_info));
ip_info.dest.family = AF_INET;
struct isis_extended_ip_reach *r;
for (r = ipv4_reachs
? (struct isis_extended_ip_reach *)
ipv4_reachs->head
: NULL;
r; r = r->next) {
dist = cost + r->metric;
ip_info.dest.u.prefix4 = r->prefix.prefix;
ip_info.dest.prefixlen = r->prefix.prefixlen;
/* Parse list of Prefix-SID subTLVs if SR is enabled */
has_valid_psid = false;
if (spftree->area->srdb.enabled && r->subtlvs) {
for (struct isis_item *i =
r->subtlvs->prefix_sids.head;
i; i = i->next) {
struct isis_prefix_sid *psid =
(struct isis_prefix_sid *)i;
if (psid->algorithm != SR_ALGORITHM_SPF)
continue;
has_valid_psid = true;
process_N(spftree, VTYPE_IPREACH_TE,
&ip_info, dist, depth + 1,
psid, parent);
/*
* Stop the Prefix-SID iteration since
* we only support the SPF algorithm for
* now.
*/
break;
}
}
if (!has_valid_psid)
process_N(spftree, VTYPE_IPREACH_TE, &ip_info,
dist, depth + 1, NULL, parent);
}
}
if (!pseudo_lsp && spftree->family == AF_INET6) {
struct isis_item_list *ipv6_reachs;
if (spftree->mtid == ISIS_MT_IPV4_UNICAST)
ipv6_reachs = &lsp->tlvs->ipv6_reach;
else
ipv6_reachs = isis_lookup_mt_items(
&lsp->tlvs->mt_ipv6_reach, spftree->mtid);
struct isis_ipv6_reach *r;
for (r = ipv6_reachs
? (struct isis_ipv6_reach *)ipv6_reachs->head
: NULL;
r; r = r->next) {
dist = cost + r->metric;
vtype = r->external ? VTYPE_IP6REACH_EXTERNAL
: VTYPE_IP6REACH_INTERNAL;
memset(&ip_info, 0, sizeof(ip_info));
ip_info.dest.family = AF_INET6;
ip_info.dest.u.prefix6 = r->prefix.prefix;
ip_info.dest.prefixlen = r->prefix.prefixlen;
if (spftree->area->srdb.enabled && r->subtlvs &&
r->subtlvs->source_prefix &&
r->subtlvs->source_prefix->prefixlen) {
if (spftree->tree_id != SPFTREE_DSTSRC) {
char buff[VID2STR_BUFFER];
zlog_warn("Ignoring dest-src route %s in non dest-src topology",
srcdest2str(
&ip_info.dest,
r->subtlvs->source_prefix,
buff, sizeof(buff)
)
);
continue;
}
ip_info.src = *r->subtlvs->source_prefix;
}
/* Parse list of Prefix-SID subTLVs */
has_valid_psid = false;
if (r->subtlvs) {
for (struct isis_item *i =
r->subtlvs->prefix_sids.head;
i; i = i->next) {
struct isis_prefix_sid *psid =
(struct isis_prefix_sid *)i;
if (psid->algorithm != SR_ALGORITHM_SPF)
continue;
has_valid_psid = true;
process_N(spftree, vtype, &ip_info,
dist, depth + 1, psid,
parent);
/*
* Stop the Prefix-SID iteration since
* we only support the SPF algorithm for
* now.
*/
break;
}
}
if (!has_valid_psid)
process_N(spftree, vtype, &ip_info, dist,
depth + 1, NULL, parent);
}
}
end:
/* if attach bit set in LSP, attached-bit receive ignore is
* not configured, we are a level-1 area and we have no other
* level-2 | level1-2 areas then add a default route toward
* this neighbor
*/
if ((lsp->hdr.lsp_bits & LSPBIT_ATT) == LSPBIT_ATT
&& !spftree->area->attached_bit_rcv_ignore
&& (spftree->area->is_type & IS_LEVEL_1)
&& !isis_level2_adj_up(spftree->area)) {
struct prefix_pair ip_info = { {0} };
if (IS_DEBUG_RTE_EVENTS)
zlog_debug("ISIS-Spf (%s): add default %s route",
rawlspid_print(lsp->hdr.lsp_id),
spftree->family == AF_INET ? "ipv4"
: "ipv6");
if (spftree->family == AF_INET) {
ip_info.dest.family = AF_INET;
vtype = VTYPE_IPREACH_INTERNAL;
} else {
ip_info.dest.family = AF_INET6;
vtype = VTYPE_IP6REACH_INTERNAL;
}
process_N(spftree, vtype, &ip_info, cost, depth + 1, NULL,
parent);
}
2003-12-23 09:09:43 +01:00
if (fragnode == NULL)
fragnode = listhead(lsp->lspu.frags);
else
fragnode = listnextnode(fragnode);
if (fragnode) {
lsp = listgetdata(fragnode);
goto lspfragloop;
}
2003-12-23 09:09:43 +01:00
return ISIS_OK;
}
static struct isis_adjacency *adj_find(struct list *adj_list, const uint8_t *id,
int level, uint16_t mtid, int family)
{
2003-12-23 09:09:43 +01:00
struct isis_adjacency *adj;
struct listnode *node;
for (ALL_LIST_ELEMENTS_RO(adj_list, node, adj)) {
if (!(adj->level & level))
continue;
if (memcmp(adj->sysid, id, ISIS_SYS_ID_LEN) != 0)
continue;
if (adj->adj_state != ISIS_ADJ_UP)
continue;
if (!adj_has_mt(adj, mtid))
continue;
if (mtid == ISIS_MT_IPV4_UNICAST
&& !speaks(adj->nlpids.nlpids, adj->nlpids.count, family))
continue;
return adj;
}
return NULL;
}
struct spf_preload_tent_ip_reach_args {
struct isis_spftree *spftree;
struct isis_vertex *parent;
};
static int isis_spf_preload_tent_ip_reach_cb(const struct prefix *prefix,
uint32_t metric, bool external,
struct isis_subtlvs *subtlvs,
void *arg)
{
struct spf_preload_tent_ip_reach_args *args = arg;
struct isis_spftree *spftree = args->spftree;
struct isis_vertex *parent = args->parent;
struct prefix_pair ip_info;
enum vertextype vtype;
bool has_valid_psid = false;
if (external)
return LSP_ITER_CONTINUE;
assert(spftree->family == prefix->family);
memset(&ip_info, 0, sizeof(ip_info));
prefix_copy(&ip_info.dest, prefix);
apply_mask(&ip_info.dest);
if (prefix->family == AF_INET)
vtype = VTYPE_IPREACH_INTERNAL;
else
vtype = VTYPE_IP6REACH_INTERNAL;
/* Parse list of Prefix-SID subTLVs if SR is enabled */
if (spftree->area->srdb.enabled && subtlvs) {
for (struct isis_item *i = subtlvs->prefix_sids.head; i;
i = i->next) {
struct isis_prefix_sid *psid =
(struct isis_prefix_sid *)i;
if (psid->algorithm != SR_ALGORITHM_SPF)
continue;
has_valid_psid = true;
isis_spf_add_local(spftree, vtype, &ip_info, NULL, 0,
psid, parent);
/*
* Stop the Prefix-SID iteration since we only support
* the SPF algorithm for now.
*/
break;
}
}
if (!has_valid_psid)
isis_spf_add_local(spftree, vtype, &ip_info, NULL, 0, NULL,
parent);
return LSP_ITER_CONTINUE;
}
static void isis_spf_preload_tent(struct isis_spftree *spftree,
uint8_t *root_sysid,
struct isis_lsp *root_lsp,
struct isis_vertex *parent)
{
struct spf_preload_tent_ip_reach_args ip_reach_args;
struct isis_spf_adj *sadj;
struct listnode *node;
if (!CHECK_FLAG(spftree->flags, F_SPFTREE_HOPCOUNT_METRIC)) {
ip_reach_args.spftree = spftree;
ip_reach_args.parent = parent;
isis_lsp_iterate_ip_reach(
root_lsp, spftree->family, spftree->mtid,
isis_spf_preload_tent_ip_reach_cb, &ip_reach_args);
}
/* Iterate over adjacencies. */
for (ALL_LIST_ELEMENTS_RO(spftree->sadj_list, node, sadj)) {
const uint8_t *adj_id;
uint32_t metric;
if (CHECK_FLAG(sadj->flags, F_ISIS_SPF_ADJ_BROADCAST))
adj_id = sadj->lan.desig_is_id;
else
adj_id = sadj->id;
if (isis_lfa_excise_adj_check(spftree, adj_id)) {
if (IS_DEBUG_LFA)
zlog_debug("ISIS-SPF: excising adjacency %s",
isis_format_id(sadj->id,
ISIS_SYS_ID_LEN + 1));
continue;
}
metric = CHECK_FLAG(spftree->flags, F_SPFTREE_HOPCOUNT_METRIC)
? 1
: sadj->metric;
if (!LSP_PSEUDO_ID(sadj->id)) {
isis_spf_add_local(spftree,
CHECK_FLAG(sadj->flags,
F_ISIS_SPF_ADJ_OLDMETRIC)
? VTYPE_NONPSEUDO_IS
: VTYPE_NONPSEUDO_TE_IS,
sadj->id, sadj, metric, NULL,
parent);
} else if (sadj->lsp) {
isis_spf_process_lsp(spftree, sadj->lsp, metric, 0,
spftree->sysid, parent);
}
}
}
struct spf_adj_find_reverse_metric_args {
const uint8_t *id_self;
uint32_t reverse_metric;
};
static int spf_adj_find_reverse_metric_cb(const uint8_t *id, uint32_t metric,
bool oldmetric,
struct isis_ext_subtlvs *subtlvs,
void *arg)
{
struct spf_adj_find_reverse_metric_args *args = arg;
if (memcmp(id, args->id_self, ISIS_SYS_ID_LEN))
return LSP_ITER_CONTINUE;
args->reverse_metric = metric;
return LSP_ITER_STOP;
}
/*
* Change all SPF adjacencies to use the link cost in the direction from the
* next hop back towards root in place of the link cost in the direction away
* from root towards the next hop.
*/
static void spf_adj_get_reverse_metrics(struct isis_spftree *spftree)
{
struct isis_spf_adj *sadj;
struct listnode *node, *nnode;
for (ALL_LIST_ELEMENTS(spftree->sadj_list, node, nnode, sadj)) {
uint8_t lspid[ISIS_SYS_ID_LEN + 2];
struct isis_lsp *lsp_adj;
const uint8_t *id_self;
struct spf_adj_find_reverse_metric_args args;
/* Skip pseudonodes. */
if (LSP_PSEUDO_ID(sadj->id))
continue;
/* Find LSP of the corresponding adjacency. */
memcpy(lspid, sadj->id, ISIS_SYS_ID_LEN);
LSP_PSEUDO_ID(lspid) = 0;
LSP_FRAGMENT(lspid) = 0;
lsp_adj = lsp_search(spftree->lspdb, lspid);
if (lsp_adj == NULL || lsp_adj->hdr.rem_lifetime == 0) {
/* Delete one-way adjacency. */
listnode_delete(spftree->sadj_list, sadj);
isis_spf_adj_free(sadj);
continue;
}
/* Find root node in the LSP of the adjacent router. */
if (CHECK_FLAG(sadj->flags, F_ISIS_SPF_ADJ_BROADCAST))
id_self = sadj->lan.desig_is_id;
else
id_self = spftree->sysid;
args.id_self = id_self;
args.reverse_metric = UINT32_MAX;
isis_lsp_iterate_is_reach(lsp_adj, spftree->mtid,
spf_adj_find_reverse_metric_cb,
&args);
if (args.reverse_metric == UINT32_MAX) {
/* Delete one-way adjacency. */
listnode_delete(spftree->sadj_list, sadj);
isis_spf_adj_free(sadj);
continue;
}
sadj->metric = args.reverse_metric;
}
}
static void spf_adj_list_parse_tlv(struct isis_spftree *spftree,
struct list *adj_list, const uint8_t *id,
const uint8_t *desig_is_id,
uint32_t pseudo_metric, uint32_t metric,
bool oldmetric,
struct isis_ext_subtlvs *subtlvs)
{
struct isis_spf_adj *sadj;
uint8_t lspid[ISIS_SYS_ID_LEN + 2];
struct isis_lsp *lsp;
uint8_t flags = 0;
/* Skip self in the pseudonode. */
if (desig_is_id && !memcmp(id, spftree->sysid, ISIS_SYS_ID_LEN))
return;
/* Find LSP from the adjacency. */
memcpy(lspid, id, ISIS_SYS_ID_LEN + 1);
LSP_FRAGMENT(lspid) = 0;
lsp = lsp_search(spftree->lspdb, lspid);
if (lsp == NULL || lsp->hdr.rem_lifetime == 0) {
zlog_warn("ISIS-SPF: No LSP found from root to L%d %s",
spftree->level, rawlspid_print(lspid));
return;
}
sadj = XCALLOC(MTYPE_ISIS_SPF_ADJ, sizeof(*sadj));
memcpy(sadj->id, id, sizeof(sadj->id));
if (desig_is_id) {
memcpy(sadj->lan.desig_is_id, desig_is_id,
sizeof(sadj->lan.desig_is_id));
SET_FLAG(flags, F_ISIS_SPF_ADJ_BROADCAST);
sadj->metric = pseudo_metric;
} else
sadj->metric = metric;
if (oldmetric)
SET_FLAG(flags, F_ISIS_SPF_ADJ_OLDMETRIC);
sadj->lsp = lsp;
sadj->subtlvs = subtlvs;
sadj->flags = flags;
if ((oldmetric && metric == ISIS_NARROW_METRIC_INFINITY)
|| (!oldmetric && metric == ISIS_WIDE_METRIC_INFINITY))
SET_FLAG(flags, F_ISIS_SPF_ADJ_METRIC_INFINITY);
/* Set real adjacency. */
if (!CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ADJACENCIES)
&& !LSP_PSEUDO_ID(id)) {
struct isis_adjacency *adj;
adj = adj_find(adj_list, id, spftree->level, spftree->mtid,
spftree->family);
if (!adj) {
XFREE(MTYPE_ISIS_SPF_ADJ, sadj);
return;
}
listnode_delete(adj_list, adj);
sadj->adj = adj;
}
/* Add adjacency to the list. */
listnode_add(spftree->sadj_list, sadj);
if (!LSP_PSEUDO_ID(id)) {
struct isis_spf_node *node;
node = isis_spf_node_find(&spftree->adj_nodes, id);
if (!node)
node = isis_spf_node_new(&spftree->adj_nodes, id);
if (node->best_metric == 0 || sadj->metric < node->best_metric)
node->best_metric = sadj->metric;
listnode_add(node->adjacencies, sadj);
}
/* Parse pseudonode LSP too. */
if (LSP_PSEUDO_ID(id))
spf_adj_list_parse_lsp(spftree, adj_list, lsp, id, metric);
}
static void spf_adj_list_parse_lsp(struct isis_spftree *spftree,
struct list *adj_list, struct isis_lsp *lsp,
const uint8_t *pseudo_nodeid,
uint32_t pseudo_metric)
{
bool pseudo_lsp = LSP_PSEUDO_ID(lsp->hdr.lsp_id);
struct isis_lsp *frag;
struct listnode *node;
struct isis_item *head;
struct isis_item_list *te_neighs;
if (lsp->hdr.seqno == 0 || lsp->hdr.rem_lifetime == 0)
return;
isisd: fix infinite loop when parsing LSPs Fixing the crash: > #0 0x0000560aa80f8e30 in lspdb_const_find (h=<error reading variable: Cannot access memory at address 0x7fff5e95efe8>, item=<error reading variable: Cannot access memory at address 0x7fff5e95efe0>) at ./isisd/isis_lsp.h:64 > #1 0x0000560aa80f8e9d in lspdb_find (h=0x560aaa1ed3b8, item=0x7fff5e95f050) at ./isisd/isis_lsp.h:64 > #2 0x0000560aa80f92f9 in lsp_search (head=0x560aaa1ed3b8, id=0x7fff5e95f200 "") at isisd/isis_lsp.c:100 > #3 0x0000560aa8113d69 in spf_adj_list_parse_tlv (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, id=0x560aad331a78 "", desig_is_id=0x0, pseudo_metric=0, metric=3, oldmetric=false, subtlvs=0x0) at isisd/isis_spf.c:1330 > #4 0x0000560aa811419d in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1429 > #5 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1ff8e0, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #6 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > (...) > #65507 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1ff8e0, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65508 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65509 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1ff8e0, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65510 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65511 0x0000560aa8114313 in isis_spf_build_adj_list (spftree=0x560aaa1f09d0, lsp=0x560aaa1f4e50) at isisd/isis_spf.c:1455 > #65512 0x0000560aa8114f09 in isis_run_spf (spftree=0x560aaa1f09d0) at isisd/isis_spf.c:1775 > #65513 0x0000560aa8115057 in isis_run_spf_with_protection (area=0x560aaa1ed3b0, spftree=0x560aaa1f09d0) at isisd/isis_spf.c:1801 > #65514 0x0000560aa8115311 in isis_run_spf_cb (thread=0x7fff5f15e5a0) at isisd/isis_spf.c:1859 > #65515 0x00007f90bac66dcc in thread_call (thread=0x7fff5f15e5a0) at lib/thread.c:2002 > #65516 0x00007f90bac013ee in frr_run (master=0x560aa9f5cb40) at lib/libfrr.c:1196 > #65517 0x0000560aa80e7da2 in main (argc=2, argv=0x7fff5f15e7b8, envp=0x7fff5f15e7d0) at isisd/isis_main.c:273 The fix is similar to the crash fix included in d9884a758c ("isisd: Prepare IS-IS for Link State support"). The fix was: > diff --git a/isisd/isis_lsp.c b/isisd/isis_lsp.c > index 94353a5bc8..92d329f035 100644 > --- a/isisd/isis_lsp.c > +++ b/isisd/isis_lsp.c > @@ -2166,7 +2178,7 @@ int isis_lsp_iterate_ip_reach(struct isis_lsp *lsp, int family, uint16_t mtid, > if (lsp->hdr.seqno == 0 || lsp->hdr.rem_lifetime == 0) > return LSP_ITER_CONTINUE; > > - /* Parse main LSP. */ > + /* Parse LSP */ > if (lsp->tlvs) { > if (!fabricd && !pseudo_lsp && family == AF_INET > && mtid == ISIS_MT_IPV4_UNICAST) { > @@ -2236,13 +2248,17 @@ int isis_lsp_iterate_ip_reach(struct isis_lsp *lsp, int family, uint16_t mtid, > } > } > > - /* Parse LSP fragments. */ > - for (ALL_LIST_ELEMENTS_RO(lsp->lspu.frags, node, frag)) { > - if (!frag->tlvs) > - continue; > + /* Parse LSP fragments if it is not a fragment itself */ > + if (!LSP_FRAGMENT(lsp->hdr.lsp_id)) > + for (ALL_LIST_ELEMENTS_RO(lsp->lspu.frags, node, frag)) { > + if (!frag->tlvs) > + continue; > > - isis_lsp_iterate_ip_reach(frag, family, mtid, cb, arg); > - } > + if (isis_lsp_iterate_ip_reach(frag, family, mtid, cb, > + arg) > + == LSP_ITER_STOP) > + return LSP_ITER_STOP; > + } > > return LSP_ITER_CONTINUE; > } Fixes: 7b36d36e0e ("isisd: make the SPF code more modular") Fixes: 5e56a50559 ("isisd: fix infinite loop when parsing LSPs") Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
2022-05-27 10:42:53 +02:00
/* Parse LSP. */
if (lsp->tlvs) {
if (pseudo_lsp || spftree->mtid == ISIS_MT_IPV4_UNICAST) {
head = lsp->tlvs->oldstyle_reach.head;
for (struct isis_oldstyle_reach *reach =
(struct isis_oldstyle_reach *)head;
reach; reach = reach->next) {
spf_adj_list_parse_tlv(
spftree, adj_list, reach->id,
pseudo_nodeid, pseudo_metric,
reach->metric, true, NULL);
}
}
if (pseudo_lsp || spftree->mtid == ISIS_MT_IPV4_UNICAST)
te_neighs = &lsp->tlvs->extended_reach;
else
te_neighs = isis_get_mt_items(&lsp->tlvs->mt_reach,
spftree->mtid);
if (te_neighs) {
head = te_neighs->head;
for (struct isis_extended_reach *reach =
(struct isis_extended_reach *)head;
reach; reach = reach->next) {
spf_adj_list_parse_tlv(
spftree, adj_list, reach->id,
pseudo_nodeid, pseudo_metric,
reach->metric, false, reach->subtlvs);
}
}
}
2003-12-23 09:09:43 +01:00
isisd: fix infinite loop when parsing LSPs Fixing the crash: > #0 0x0000560aa80f8e30 in lspdb_const_find (h=<error reading variable: Cannot access memory at address 0x7fff5e95efe8>, item=<error reading variable: Cannot access memory at address 0x7fff5e95efe0>) at ./isisd/isis_lsp.h:64 > #1 0x0000560aa80f8e9d in lspdb_find (h=0x560aaa1ed3b8, item=0x7fff5e95f050) at ./isisd/isis_lsp.h:64 > #2 0x0000560aa80f92f9 in lsp_search (head=0x560aaa1ed3b8, id=0x7fff5e95f200 "") at isisd/isis_lsp.c:100 > #3 0x0000560aa8113d69 in spf_adj_list_parse_tlv (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, id=0x560aad331a78 "", desig_is_id=0x0, pseudo_metric=0, metric=3, oldmetric=false, subtlvs=0x0) at isisd/isis_spf.c:1330 > #4 0x0000560aa811419d in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1429 > #5 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1ff8e0, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #6 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > (...) > #65507 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1ff8e0, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65508 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65509 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1ff8e0, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65510 0x0000560aa81141fe in spf_adj_list_parse_lsp (spftree=0x560aaa1f09d0, adj_list=0x560aaa214480, lsp=0x560aaa1f4e50, pseudo_nodeid=0x0, pseudo_metric=0) at isisd/isis_spf.c:1442 > #65511 0x0000560aa8114313 in isis_spf_build_adj_list (spftree=0x560aaa1f09d0, lsp=0x560aaa1f4e50) at isisd/isis_spf.c:1455 > #65512 0x0000560aa8114f09 in isis_run_spf (spftree=0x560aaa1f09d0) at isisd/isis_spf.c:1775 > #65513 0x0000560aa8115057 in isis_run_spf_with_protection (area=0x560aaa1ed3b0, spftree=0x560aaa1f09d0) at isisd/isis_spf.c:1801 > #65514 0x0000560aa8115311 in isis_run_spf_cb (thread=0x7fff5f15e5a0) at isisd/isis_spf.c:1859 > #65515 0x00007f90bac66dcc in thread_call (thread=0x7fff5f15e5a0) at lib/thread.c:2002 > #65516 0x00007f90bac013ee in frr_run (master=0x560aa9f5cb40) at lib/libfrr.c:1196 > #65517 0x0000560aa80e7da2 in main (argc=2, argv=0x7fff5f15e7b8, envp=0x7fff5f15e7d0) at isisd/isis_main.c:273 The fix is similar to the crash fix included in d9884a758c ("isisd: Prepare IS-IS for Link State support"). The fix was: > diff --git a/isisd/isis_lsp.c b/isisd/isis_lsp.c > index 94353a5bc8..92d329f035 100644 > --- a/isisd/isis_lsp.c > +++ b/isisd/isis_lsp.c > @@ -2166,7 +2178,7 @@ int isis_lsp_iterate_ip_reach(struct isis_lsp *lsp, int family, uint16_t mtid, > if (lsp->hdr.seqno == 0 || lsp->hdr.rem_lifetime == 0) > return LSP_ITER_CONTINUE; > > - /* Parse main LSP. */ > + /* Parse LSP */ > if (lsp->tlvs) { > if (!fabricd && !pseudo_lsp && family == AF_INET > && mtid == ISIS_MT_IPV4_UNICAST) { > @@ -2236,13 +2248,17 @@ int isis_lsp_iterate_ip_reach(struct isis_lsp *lsp, int family, uint16_t mtid, > } > } > > - /* Parse LSP fragments. */ > - for (ALL_LIST_ELEMENTS_RO(lsp->lspu.frags, node, frag)) { > - if (!frag->tlvs) > - continue; > + /* Parse LSP fragments if it is not a fragment itself */ > + if (!LSP_FRAGMENT(lsp->hdr.lsp_id)) > + for (ALL_LIST_ELEMENTS_RO(lsp->lspu.frags, node, frag)) { > + if (!frag->tlvs) > + continue; > > - isis_lsp_iterate_ip_reach(frag, family, mtid, cb, arg); > - } > + if (isis_lsp_iterate_ip_reach(frag, family, mtid, cb, > + arg) > + == LSP_ITER_STOP) > + return LSP_ITER_STOP; > + } > > return LSP_ITER_CONTINUE; > } Fixes: 7b36d36e0e ("isisd: make the SPF code more modular") Fixes: 5e56a50559 ("isisd: fix infinite loop when parsing LSPs") Signed-off-by: Louis Scalbert <louis.scalbert@6wind.com>
2022-05-27 10:42:53 +02:00
if (LSP_FRAGMENT(lsp->hdr.lsp_id))
return;
/* Parse LSP fragments. */
for (ALL_LIST_ELEMENTS_RO(lsp->lspu.frags, node, frag)) {
if (!frag->tlvs)
continue;
spf_adj_list_parse_lsp(spftree, adj_list, frag, pseudo_nodeid,
pseudo_metric);
}
}
static void isis_spf_build_adj_list(struct isis_spftree *spftree,
struct isis_lsp *lsp)
{
struct list *adj_list = NULL;
if (!CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ADJACENCIES))
adj_list = list_dup(spftree->area->adjacency_list);
spf_adj_list_parse_lsp(spftree, adj_list, lsp, NULL, 0);
if (!CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ADJACENCIES))
list_delete(&adj_list);
if (spftree->type == SPF_TYPE_REVERSE)
spf_adj_get_reverse_metrics(spftree);
2003-12-23 09:09:43 +01:00
}
/*
* The parent(s) for vertex is set when added to TENT list
* now we just put the child pointer(s) in place
*/
static void add_to_paths(struct isis_spftree *spftree,
struct isis_vertex *vertex)
2003-12-23 09:09:43 +01:00
{
#ifdef EXTREME_DEBUG
char buff[VID2STR_BUFFER];
#endif /* EXTREME_DEBUG */
2012-03-24 16:35:20 +01:00
if (isis_find_vertex(&spftree->paths, &vertex->N, vertex->type))
2012-03-24 16:35:20 +01:00
return;
isis_vertex_queue_append(&spftree->paths, vertex);
2003-12-23 09:09:43 +01:00
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug("ISIS-SPF: added %s %s %s depth %d dist %d to PATHS",
print_sys_hostname(vertex->N.id),
vtype2string(vertex->type),
vid2string(vertex, buff, sizeof(buff)),
vertex->depth, vertex->d_N);
#endif /* EXTREME_DEBUG */
}
static void init_spt(struct isis_spftree *spftree, int mtid)
{
/* Clear data from previous run. */
hash_clean(spftree->prefix_sids, NULL);
isis_spf_node_list_clear(&spftree->adj_nodes);
list_delete_all_node(spftree->sadj_list);
isis_vertex_queue_clear(&spftree->tents);
isis_vertex_queue_clear(&spftree->paths);
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
isis_zebra_rlfa_unregister_all(spftree);
isis_rlfa_list_clear(spftree);
list_delete_all_node(spftree->lfa.remote.pc_spftrees);
memset(&spftree->lfa.protection_counters, 0,
sizeof(spftree->lfa.protection_counters));
spftree->mtid = mtid;
}
static enum spf_prefix_priority
spf_prefix_priority(struct isis_spftree *spftree, struct isis_vertex *vertex)
{
struct isis_area *area = spftree->area;
struct prefix *prefix = &vertex->N.ip.p.dest;
for (int priority = SPF_PREFIX_PRIO_CRITICAL;
priority <= SPF_PREFIX_PRIO_MEDIUM; priority++) {
struct spf_prefix_priority_acl *ppa;
enum filter_type ret = FILTER_PERMIT;
ppa = &area->spf_prefix_priorities[priority];
switch (spftree->family) {
case AF_INET:
ret = access_list_apply(ppa->list_v4, prefix);
break;
case AF_INET6:
ret = access_list_apply(ppa->list_v6, prefix);
break;
default:
break;
}
if (ret == FILTER_PERMIT)
return priority;
}
/* Assign medium priority to loopback prefixes by default. */
if (is_host_route(prefix))
return SPF_PREFIX_PRIO_MEDIUM;
return SPF_PREFIX_PRIO_LOW;
}
static void spf_path_process(struct isis_spftree *spftree,
struct isis_vertex *vertex)
{
struct isis_area *area = spftree->area;
int level = spftree->level;
char buff[VID2STR_BUFFER];
2012-03-24 16:35:20 +01:00
if (spftree->type == SPF_TYPE_TI_LFA && VTYPE_IS(vertex->type)
&& !CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ADJACENCIES)) {
if (listcount(vertex->Adj_N) > 0) {
struct isis_adjacency *adj;
if (isis_tilfa_check(spftree, vertex) != 0)
return;
adj = isis_adj_find(area, level, vertex->N.id);
if (adj)
sr_adj_sid_add_single(adj, spftree->family,
true, vertex->Adj_N);
} else if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: no adjacencies, do not install backup Adj-SID for %s depth %d dist %d",
vid2string(vertex, buff, sizeof(buff)),
vertex->depth, vertex->d_N);
}
if (VTYPE_IP(vertex->type)
&& !CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ROUTES)) {
enum spf_prefix_priority priority;
priority = spf_prefix_priority(spftree, vertex);
vertex->N.ip.priority = priority;
if (vertex->depth == 1 || listcount(vertex->Adj_N) > 0) {
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
struct isis_spftree *pre_spftree;
struct route_table *route_table = NULL;
bool allow_ecmp = false;
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
switch (spftree->type) {
case SPF_TYPE_RLFA:
case SPF_TYPE_TI_LFA:
if (priority
> area->lfa_priority_limit[level - 1]) {
if (IS_DEBUG_LFA)
zlog_debug(
"ISIS-LFA: skipping %s %s (low prefix priority)",
vtype2string(
vertex->type),
vid2string(
vertex, buff,
sizeof(buff)));
return;
}
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
break;
case SPF_TYPE_FORWARD:
case SPF_TYPE_REVERSE:
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
break;
}
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
switch (spftree->type) {
case SPF_TYPE_RLFA:
isis_rlfa_check(spftree, vertex);
return;
case SPF_TYPE_TI_LFA:
if (isis_tilfa_check(spftree, vertex) != 0)
return;
pre_spftree = spftree->lfa.old.spftree;
route_table = pre_spftree->route_table_backup;
allow_ecmp = area->lfa_load_sharing[level - 1];
pre_spftree->lfa.protection_counters
.tilfa[vertex->N.ip.priority] += 1;
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
break;
case SPF_TYPE_FORWARD:
case SPF_TYPE_REVERSE:
route_table = spftree->route_table;
allow_ecmp = true;
/*
* Update LFA protection counters (ignore local
* routes).
*/
if (vertex->depth > 1) {
spftree->lfa.protection_counters
.total[priority] += 1;
if (listcount(vertex->Adj_N) > 1)
spftree->lfa.protection_counters
.ecmp[priority] += 1;
}
isisd: implement Remote LFA Remote LFA (RFC 7490) is an extension to the base LFA mechanism that uses dynamically determined tunnels to extend the IP-FRR protection coverage. RLFA is similar to TI-LFA in that it computes a post-convergence SPT (with the protected interface pruned from the network topology) and the P/Q spaces based on that SPT. There are a few differences however: * RLFAs can push at most one label, so the P/Q spaces need to intersect otherwise the destination can't be protected (the protection coverage is topology dependent). * isisd needs to interface with ldpd to obtain the labels it needs to create a tunnel to the PQ node. That interaction needs to be done asynchronously to prevent blocking the daemon for too long. With TI-LFA all required labels are already available in the LSPDB. RLFA and TI-LFA have more similarities than differences though, and thanks to that both features share a lot of code. Limitations: * Only RLFA link protection is implemented. The algorithm used to find node-protecting RLFAs (RFC 8102) is too CPU intensive and doesn't always work. Most vendors implement RLFA link protection only. * RFC 7490 says it should be a local matter whether the repair path selection policy favors LFA repairs over RLFA repairs. It might be desirable, for instance, to prefer RLFAs that satisfy the downstream condition over LFAs that don't. In this implementation, however, RLFAs are only computed for destinations that can't be protected by local LFAs. Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
2020-11-26 03:39:09 +01:00
break;
}
isis_route_create(
&vertex->N.ip.p.dest, &vertex->N.ip.p.src,
vertex->d_N, vertex->depth, &vertex->N.ip.sr,
vertex->Adj_N, allow_ecmp, area, route_table);
} else if (IS_DEBUG_SPF_EVENTS)
2012-03-24 16:35:20 +01:00
zlog_debug(
"ISIS-SPF: no adjacencies, do not install route for %s depth %d dist %d",
vid2string(vertex, buff, sizeof(buff)),
2012-03-24 16:35:20 +01:00
vertex->depth, vertex->d_N);
}
}
static void isis_spf_loop(struct isis_spftree *spftree,
uint8_t *root_sysid)
{
struct isis_vertex *vertex;
struct isis_lsp *lsp;
struct listnode *node;
while (isis_vertex_queue_count(&spftree->tents)) {
vertex = isis_vertex_queue_pop(&spftree->tents);
#ifdef EXTREME_DEBUG
if (IS_DEBUG_SPF_EVENTS)
zlog_debug(
"ISIS-SPF: get TENT node %s %s depth %d dist %d to PATHS",
print_sys_hostname(vertex->N.id),
vtype2string(vertex->type), vertex->depth,
vertex->d_N);
#endif /* EXTREME_DEBUG */
add_to_paths(spftree, vertex);
if (!VTYPE_IS(vertex->type))
continue;
lsp = lsp_for_vertex(spftree, vertex);
if (!lsp) {
zlog_warn("ISIS-SPF: No LSP found for %s",
isis_format_id(vertex->N.id,
sizeof(vertex->N.id)));
continue;
}
isis_spf_process_lsp(spftree, lsp, vertex->d_N, vertex->depth,
root_sysid, vertex);
}
/* Generate routes once the SPT is formed. */
for (ALL_QUEUE_ELEMENTS_RO(&spftree->paths, node, vertex)) {
/* New-style TLVs take precedence over the old-style TLVs. */
switch (vertex->type) {
case VTYPE_IPREACH_INTERNAL:
case VTYPE_IPREACH_EXTERNAL:
if (isis_find_vertex(&spftree->paths, &vertex->N,
VTYPE_IPREACH_TE))
continue;
break;
case VTYPE_PSEUDO_IS:
case VTYPE_PSEUDO_TE_IS:
case VTYPE_NONPSEUDO_IS:
case VTYPE_NONPSEUDO_TE_IS:
case VTYPE_ES:
case VTYPE_IPREACH_TE:
case VTYPE_IP6REACH_INTERNAL:
case VTYPE_IP6REACH_EXTERNAL:
break;
}
spf_path_process(spftree, vertex);
}
}
struct isis_spftree *isis_run_hopcount_spf(struct isis_area *area,
uint8_t *sysid,
struct isis_spftree *spftree)
{
if (!spftree)
spftree = isis_spftree_new(area, &area->lspdb[IS_LEVEL_2 - 1],
sysid, ISIS_LEVEL2, SPFTREE_IPV4,
SPF_TYPE_FORWARD,
F_SPFTREE_HOPCOUNT_METRIC);
init_spt(spftree, ISIS_MT_IPV4_UNICAST);
if (!memcmp(sysid, area->isis->sysid, ISIS_SYS_ID_LEN)) {
struct isis_lsp *root_lsp;
struct isis_vertex *root_vertex;
root_lsp = isis_root_system_lsp(spftree->lspdb, spftree->sysid);
if (root_lsp) {
/*
* If we are running locally, initialize with
* information from adjacencies
*/
root_vertex = isis_spf_add_root(spftree);
isis_spf_preload_tent(spftree, sysid, root_lsp,
root_vertex);
}
} else {
isis_vertex_queue_insert(
&spftree->tents,
isis_vertex_new(spftree, sysid, VTYPE_NONPSEUDO_TE_IS));
}
isis_spf_loop(spftree, sysid);
return spftree;
2003-12-23 09:09:43 +01:00
}
void isis_run_spf(struct isis_spftree *spftree)
2003-12-23 09:09:43 +01:00
{
struct isis_lsp *root_lsp;
2012-03-24 16:35:20 +01:00
struct isis_vertex *root_vertex;
struct timeval time_start;
struct timeval time_end;
struct isis_mt_router_info *mt_router_info;
uint16_t mtid = 0;
2012-03-28 08:48:05 +02:00
/* Get time that can't roll backwards. */
monotime(&time_start);
root_lsp = isis_root_system_lsp(spftree->lspdb, spftree->sysid);
if (root_lsp == NULL) {
zlog_err("ISIS-SPF: could not find own l%d LSP!",
spftree->level);
return;
}
/* Get Multi-Topology ID. */
switch (spftree->tree_id) {
case SPFTREE_IPV4:
mtid = ISIS_MT_IPV4_UNICAST;
break;
case SPFTREE_IPV6:
mt_router_info = isis_tlvs_lookup_mt_router_info(
root_lsp->tlvs, ISIS_MT_IPV6_UNICAST);
if (mt_router_info)
mtid = ISIS_MT_IPV6_UNICAST;
else
mtid = ISIS_MT_IPV4_UNICAST;
break;
case SPFTREE_DSTSRC:
mtid = ISIS_MT_IPV6_DSTSRC;
break;
case SPFTREE_COUNT:
zlog_err(
"%s should never be called with SPFTREE_COUNT as argument!",
__func__);
exit(1);
}
2003-12-23 09:09:43 +01:00
/*
* C.2.5 Step 0
*/
init_spt(spftree, mtid);
2003-12-23 09:09:43 +01:00
/* a) */
root_vertex = isis_spf_add_root(spftree);
2003-12-23 09:09:43 +01:00
/* b) */
isis_spf_build_adj_list(spftree, root_lsp);
isis_spf_preload_tent(spftree, spftree->sysid, root_lsp, root_vertex);
2003-12-23 09:09:43 +01:00
/*
* C.2.7 Step 2
*/
if (!isis_vertex_queue_count(&spftree->tents)
&& (IS_DEBUG_SPF_EVENTS)) {
zlog_warn("ISIS-SPF: TENT is empty SPF-root:%s",
print_sys_hostname(spftree->sysid));
}
isis_spf_loop(spftree, spftree->sysid);
2012-03-28 08:48:05 +02:00
spftree->runcount++;
spftree->last_run_timestamp = time(NULL);
spftree->last_run_monotime = monotime(&time_end);
spftree->last_run_duration =
((time_end.tv_sec - time_start.tv_sec) * 1000000)
+ (time_end.tv_usec - time_start.tv_usec);
2003-12-23 09:09:43 +01:00
}
static void isis_run_spf_with_protection(struct isis_area *area,
struct isis_spftree *spftree)
{
/* Run forward SPF locally. */
memcpy(spftree->sysid, area->isis->sysid, ISIS_SYS_ID_LEN);
isis_run_spf(spftree);
/* Run LFA protection if configured. */
if (area->lfa_protected_links[spftree->level - 1] > 0
|| area->tilfa_protected_links[spftree->level - 1] > 0)
isis_spf_run_lfa(area, spftree);
}
void isis_spf_verify_routes(struct isis_area *area, struct isis_spftree **trees)
{
if (area->is_type == IS_LEVEL_1) {
isis_route_verify_table(area, trees[0]->route_table,
trees[0]->route_table_backup);
} else if (area->is_type == IS_LEVEL_2) {
isis_route_verify_table(area, trees[1]->route_table,
trees[1]->route_table_backup);
} else {
isis_route_verify_merge(area, trees[0]->route_table,
trees[0]->route_table_backup,
trees[1]->route_table,
trees[1]->route_table_backup);
}
}
void isis_spf_invalidate_routes(struct isis_spftree *tree)
{
isis_route_invalidate_table(tree->area, tree->route_table);
/* Delete backup routes. */
route_table_finish(tree->route_table_backup);
tree->route_table_backup = srcdest_table_init();
tree->route_table_backup->cleanup = isis_route_node_cleanup;
}
void isis_spf_switchover_routes(struct isis_area *area,
struct isis_spftree **trees, int family,
union g_addr *nexthop_ip, ifindex_t ifindex,
int level)
{
isis_route_switchover_nexthop(area, trees[level - 1]->route_table,
family, nexthop_ip, ifindex);
}
static void isis_run_spf_cb(struct thread *thread)
2003-12-23 09:09:43 +01:00
{
struct isis_spf_run *run = THREAD_ARG(thread);
struct isis_area *area = run->area;
int level = run->level;
int have_run = 0;
XFREE(MTYPE_ISIS_SPF_RUN, run);
if (!(area->is_type & level)) {
if (IS_DEBUG_SPF_EVENTS)
zlog_warn("ISIS-SPF (%s) area does not share level",
area->area_tag);
return;
2003-12-23 09:09:43 +01:00
}
isis_area_delete_backup_adj_sids(area, level);
isis_area_invalidate_routes(area, level);
if (IS_DEBUG_SPF_EVENTS)
zlog_debug("ISIS-SPF (%s) L%d SPF needed, periodic SPF",
area->area_tag, level);
if (area->ip_circuits) {
isis_run_spf_with_protection(
area, area->spftree[SPFTREE_IPV4][level - 1]);
have_run = 1;
}
if (area->ipv6_circuits) {
isis_run_spf_with_protection(
area, area->spftree[SPFTREE_IPV6][level - 1]);
have_run = 1;
}
if (area->ipv6_circuits && isis_area_ipv6_dstsrc_enabled(area)) {
isis_run_spf_with_protection(
area, area->spftree[SPFTREE_DSTSRC][level - 1]);
have_run = 1;
}
if (have_run)
area->spf_run_count[level]++;
2003-12-23 09:09:43 +01:00
isis_area_verify_routes(area);
/* walk all circuits and reset any spf specific flags */
struct listnode *node;
struct isis_circuit *circuit;
for (ALL_LIST_ELEMENTS_RO(area->circuit_list, node, circuit))
UNSET_FLAG(circuit->flags, ISIS_CIRCUIT_FLAPPED_AFTER_SPF);
fabricd_run_spf(area);
2003-12-23 09:09:43 +01:00
}
static struct isis_spf_run *isis_run_spf_arg(struct isis_area *area, int level)
{
struct isis_spf_run *run = XMALLOC(MTYPE_ISIS_SPF_RUN, sizeof(*run));
run->area = area;
run->level = level;
return run;
2003-12-23 09:09:43 +01:00
}
void isis_spf_timer_free(void *run)
{
XFREE(MTYPE_ISIS_SPF_RUN, run);
}
int _isis_spf_schedule(struct isis_area *area, int level,
const char *func, const char *file, int line)
2003-12-23 09:09:43 +01:00
{
struct isis_spftree *spftree;
time_t now;
long tree_diff, diff;
int tree;
now = monotime(NULL);
diff = 0;
for (tree = SPFTREE_IPV4; tree < SPFTREE_COUNT; tree++) {
spftree = area->spftree[tree][level - 1];
tree_diff = difftime(now - spftree->last_run_monotime, 0);
if (tree_diff != now && (diff == 0 || tree_diff < diff))
diff = tree_diff;
}
if (CHECK_FLAG(im->options, F_ISIS_UNIT_TEST))
return 0;
assert(diff >= 0);
assert(area->is_type & level);
if (IS_DEBUG_SPF_EVENTS) {
zlog_debug(
"ISIS-SPF (%s) L%d SPF schedule called, lastrun %ld sec ago Caller: %s %s:%d",
area->area_tag, level, diff, func, file, line);
}
THREAD_OFF(area->t_rlfa_rib_update);
2012-03-24 16:35:20 +01:00
if (area->spf_delay_ietf[level - 1]) {
/* Need to call schedule function also if spf delay is running
* to
2012-03-24 16:35:20 +01:00
* restart holdoff timer - compare
* draft-ietf-rtgwg-backoff-algo-04 */
long delay =
spf_backoff_schedule(area->spf_delay_ietf[level - 1]);
2012-03-24 16:35:20 +01:00
if (area->spf_timer[level - 1])
return ISIS_OK;
2012-03-24 16:35:20 +01:00
thread_add_timer_msec(master, isis_run_spf_cb,
isis_run_spf_arg(area, level), delay,
&area->spf_timer[level - 1]);
return ISIS_OK;
}
2012-03-24 16:35:20 +01:00
if (area->spf_timer[level - 1])
return ISIS_OK;
2012-03-24 16:35:20 +01:00
/* wait configured min_spf_interval before doing the SPF */
long timer;
isisd: Fast RIB recovery from BFD recognized link failures Unfortunately as the topotests show a fast recovery after failure detection due to BFD is currently not possible because of the following issue: There are multiple scheduling mechanisms within isisd to prevent overload situations. Regarding our problem these two are important: * scheduler for regenerating ISIS Link State PDUs scheduler for managing * consecutive SPF calculations In fact both schedulers are coupled, the first one triggers the second one, which again is triggered by isis_adj_state_change (which again is triggered by a BFD 'down' message). The re-calculation of SPF paths finally triggers updates in zebra for the RIB. Both schedulers work as a throttle, e.g. they allow the regeneration of Link State PDUs or a re-calculation for SPF paths only once within a certain time interval which is configurable (and by default different!). This means that a request can go through the first scheduler but might still be 'stuck' at the second one for a while. Or a request can be 'stuck' at the first scheduler even though the second one is ready. This also explains the 'random' behaviour one can observe testing since a 'fast' recovery is only possible if both schedulers are ready to process this request. Note that the solution in this commit is 'thread safe' in the sense that both schedulers use the same thread master such that the introduced flags are only used exactly one time (and one after another) for a 'fast' execution. Further there are some irritating comments and logs which I partially removed. They seems to be not valid anymore due to changes in thread management (or they were never valid in the first place). Signed-off-by: GalaxyGorilla <sascha@netdef.org>
2020-05-19 13:52:04 +02:00
if (diff >= area->min_spf_interval[level - 1]
|| area->bfd_force_spf_refresh) {
/*
* Last run is more than min interval ago or BFD signalled a
* 'down' message, schedule immediate run
*/
timer = 0;
isisd: Fast RIB recovery from BFD recognized link failures Unfortunately as the topotests show a fast recovery after failure detection due to BFD is currently not possible because of the following issue: There are multiple scheduling mechanisms within isisd to prevent overload situations. Regarding our problem these two are important: * scheduler for regenerating ISIS Link State PDUs scheduler for managing * consecutive SPF calculations In fact both schedulers are coupled, the first one triggers the second one, which again is triggered by isis_adj_state_change (which again is triggered by a BFD 'down' message). The re-calculation of SPF paths finally triggers updates in zebra for the RIB. Both schedulers work as a throttle, e.g. they allow the regeneration of Link State PDUs or a re-calculation for SPF paths only once within a certain time interval which is configurable (and by default different!). This means that a request can go through the first scheduler but might still be 'stuck' at the second one for a while. Or a request can be 'stuck' at the first scheduler even though the second one is ready. This also explains the 'random' behaviour one can observe testing since a 'fast' recovery is only possible if both schedulers are ready to process this request. Note that the solution in this commit is 'thread safe' in the sense that both schedulers use the same thread master such that the introduced flags are only used exactly one time (and one after another) for a 'fast' execution. Further there are some irritating comments and logs which I partially removed. They seems to be not valid anymore due to changes in thread management (or they were never valid in the first place). Signed-off-by: GalaxyGorilla <sascha@netdef.org>
2020-05-19 13:52:04 +02:00
if (area->bfd_force_spf_refresh) {
zlog_debug(
"ISIS-SPF (%s) L%d SPF scheduled immediately due to BFD 'down' message",
isisd: Fast RIB recovery from BFD recognized link failures Unfortunately as the topotests show a fast recovery after failure detection due to BFD is currently not possible because of the following issue: There are multiple scheduling mechanisms within isisd to prevent overload situations. Regarding our problem these two are important: * scheduler for regenerating ISIS Link State PDUs scheduler for managing * consecutive SPF calculations In fact both schedulers are coupled, the first one triggers the second one, which again is triggered by isis_adj_state_change (which again is triggered by a BFD 'down' message). The re-calculation of SPF paths finally triggers updates in zebra for the RIB. Both schedulers work as a throttle, e.g. they allow the regeneration of Link State PDUs or a re-calculation for SPF paths only once within a certain time interval which is configurable (and by default different!). This means that a request can go through the first scheduler but might still be 'stuck' at the second one for a while. Or a request can be 'stuck' at the first scheduler even though the second one is ready. This also explains the 'random' behaviour one can observe testing since a 'fast' recovery is only possible if both schedulers are ready to process this request. Note that the solution in this commit is 'thread safe' in the sense that both schedulers use the same thread master such that the introduced flags are only used exactly one time (and one after another) for a 'fast' execution. Further there are some irritating comments and logs which I partially removed. They seems to be not valid anymore due to changes in thread management (or they were never valid in the first place). Signed-off-by: GalaxyGorilla <sascha@netdef.org>
2020-05-19 13:52:04 +02:00
area->area_tag, level);
area->bfd_force_spf_refresh = false;
}
} else {
timer = area->min_spf_interval[level - 1] - diff;
}
2012-03-24 16:35:20 +01:00
thread_add_timer(master, isis_run_spf_cb, isis_run_spf_arg(area, level),
timer, &area->spf_timer[level - 1]);
if (IS_DEBUG_SPF_EVENTS)
zlog_debug("ISIS-SPF (%s) L%d SPF scheduled %ld sec from now",
area->area_tag, level, timer);
2012-03-24 16:35:20 +01:00
return ISIS_OK;
}
static void isis_print_paths(struct vty *vty, struct isis_vertex_queue *queue,
uint8_t *root_sysid)
{
struct listnode *node;
2003-12-23 09:09:43 +01:00
struct isis_vertex *vertex;
char buff[VID2STR_BUFFER];
vty_out(vty,
"Vertex Type Metric Next-Hop Interface Parent\n");
for (ALL_QUEUE_ELEMENTS_RO(queue, node, vertex)) {
if (VTYPE_IS(vertex->type)
&& memcmp(vertex->N.id, root_sysid, ISIS_SYS_ID_LEN) == 0) {
2012-03-24 16:35:20 +01:00
vty_out(vty, "%-20s %-12s %-6s",
print_sys_hostname(root_sysid), "", "");
vty_out(vty, "%-30s\n", "");
continue;
}
int rows = 0;
struct listnode *anode = listhead(vertex->Adj_N);
struct listnode *pnode = listhead(vertex->parents);
struct isis_vertex_adj *vadj;
struct isis_vertex *pvertex;
vty_out(vty, "%-20s %-12s %-6u ",
vid2string(vertex, buff, sizeof(buff)),
vtype2string(vertex->type), vertex->d_N);
for (unsigned int i = 0;
i < MAX(vertex->Adj_N ? listcount(vertex->Adj_N) : 0,
vertex->parents ? listcount(vertex->parents) : 0);
i++) {
if (anode) {
vadj = listgetdata(anode);
anode = anode->next;
} else {
vadj = NULL;
}
if (pnode) {
pvertex = listgetdata(pnode);
pnode = pnode->next;
} else {
pvertex = NULL;
}
if (rows) {
vty_out(vty, "\n");
vty_out(vty, "%-20s %-12s %-6s ", "", "", "");
}
if (vadj) {
struct isis_spf_adj *sadj = vadj->sadj;
vty_out(vty, "%-20s %-9s ",
print_sys_hostname(sadj->id),
sadj->adj ? sadj->adj->circuit
->interface->name
: "-");
}
if (pvertex) {
if (!vadj)
vty_out(vty, "%-20s %-9s ", "", "");
2012-03-24 16:35:20 +01:00
vty_out(vty, "%s(%d)",
vid2string(pvertex, buff, sizeof(buff)),
2012-03-24 16:35:20 +01:00
pvertex->type);
}
++rows;
}
vty_out(vty, "\n");
}
2003-12-23 09:09:43 +01:00
}
void isis_print_spftree(struct vty *vty, struct isis_spftree *spftree)
{
const char *tree_id_text = NULL;
if (!spftree || !isis_vertex_queue_count(&spftree->paths))
return;
switch (spftree->tree_id) {
case SPFTREE_IPV4:
tree_id_text = "that speak IP";
break;
case SPFTREE_IPV6:
tree_id_text = "that speak IPv6";
break;
case SPFTREE_DSTSRC:
tree_id_text = "that support IPv6 dst-src routing";
break;
case SPFTREE_COUNT:
assert(!"isis_print_spftree shouldn't be called with SPFTREE_COUNT as type");
return;
}
vty_out(vty, "IS-IS paths to level-%d routers %s\n", spftree->level,
tree_id_text);
isis_print_paths(vty, &spftree->paths, spftree->sysid);
vty_out(vty, "\n");
}
static void show_isis_topology_common(struct vty *vty, int levels,
struct isis *isis)
2003-12-23 09:09:43 +01:00
{
struct listnode *node;
struct isis_area *area;
2003-12-23 09:09:43 +01:00
if (!isis->area_list || isis->area_list->count == 0)
return;
for (ALL_LIST_ELEMENTS_RO(isis->area_list, node, area)) {
vty_out(vty, "Area %s:\n",
area->area_tag ? area->area_tag : "null");
for (int level = ISIS_LEVEL1; level <= ISIS_LEVELS; level++) {
if ((level & levels) == 0)
continue;
if (area->ip_circuits > 0) {
isis_print_spftree(
vty,
area->spftree[SPFTREE_IPV4][level - 1]);
}
if (area->ipv6_circuits > 0) {
isis_print_spftree(
vty,
area->spftree[SPFTREE_IPV6][level - 1]);
}
if (isis_area_ipv6_dstsrc_enabled(area)) {
isis_print_spftree(vty,
area->spftree[SPFTREE_DSTSRC]
[level - 1]);
}
}
if (fabricd_spftree(area)) {
vty_out(vty,
"IS-IS paths to level-2 routers with hop-by-hop metric\n");
isis_print_paths(vty, &fabricd_spftree(area)->paths, isis->sysid);
vty_out(vty, "\n");
}
vty_out(vty, "\n");
}
}
DEFUN(show_isis_topology, show_isis_topology_cmd,
"show " PROTO_NAME
" [vrf <NAME|all>] topology"
#ifndef FABRICD
" [<level-1|level-2>]"
#endif
,
SHOW_STR PROTO_HELP VRF_CMD_HELP_STR
"All VRFs\n"
"IS-IS paths to Intermediate Systems\n"
#ifndef FABRICD
"Paths to all level-1 routers in the area\n"
"Paths to all level-2 routers in the domain\n"
#endif
)
{
int levels = ISIS_LEVELS;
struct listnode *node;
struct isis *isis = NULL;
int idx = 0;
const char *vrf_name = VRF_DEFAULT_NAME;
bool all_vrf = false;
int idx_vrf = 0;
if (argv_find(argv, argc, "topology", &idx)) {
if (argc < idx + 2)
levels = ISIS_LEVEL1 | ISIS_LEVEL2;
else if (strmatch(argv[idx + 1]->arg, "level-1"))
levels = ISIS_LEVEL1;
else
levels = ISIS_LEVEL2;
}
if (!im) {
vty_out(vty, "IS-IS Routing Process not enabled\n");
return CMD_SUCCESS;
}
ISIS_FIND_VRF_ARGS(argv, argc, idx_vrf, vrf_name, all_vrf);
if (vrf_name) {
if (all_vrf) {
for (ALL_LIST_ELEMENTS_RO(im->isis, node, isis))
show_isis_topology_common(vty, levels, isis);
return CMD_SUCCESS;
}
isis = isis_lookup_by_vrfname(vrf_name);
if (isis != NULL)
show_isis_topology_common(vty, levels, isis);
}
2012-03-24 16:35:20 +01:00
2003-12-23 09:09:43 +01:00
return CMD_SUCCESS;
}
2003-12-23 09:09:43 +01:00
static void isis_print_route(struct ttable *tt, const struct prefix *prefix,
struct isis_route_info *rinfo, bool prefix_sid,
bool no_adjacencies)
{
struct isis_nexthop *nexthop;
struct listnode *node;
bool first = true;
char buf_prefix[BUFSIZ];
(void)prefix2str(prefix, buf_prefix, sizeof(buf_prefix));
for (ALL_LIST_ELEMENTS_RO(rinfo->nexthops, node, nexthop)) {
struct interface *ifp;
char buf_iface[BUFSIZ];
char buf_nhop[BUFSIZ];
if (!no_adjacencies) {
inet_ntop(nexthop->family, &nexthop->ip, buf_nhop,
sizeof(buf_nhop));
ifp = if_lookup_by_index(nexthop->ifindex, VRF_DEFAULT);
if (ifp)
strlcpy(buf_iface, ifp->name,
sizeof(buf_iface));
else
snprintf(buf_iface, sizeof(buf_iface),
"ifindex %u", nexthop->ifindex);
} else {
strlcpy(buf_nhop, print_sys_hostname(nexthop->sysid),
sizeof(buf_nhop));
strlcpy(buf_iface, "-", sizeof(buf_iface));
}
if (prefix_sid) {
char buf_sid[BUFSIZ] = {};
char buf_lblop[BUFSIZ] = {};
if (nexthop->sr.present) {
snprintf(buf_sid, sizeof(buf_sid), "%u",
nexthop->sr.sid.value);
sr_op2str(buf_lblop, sizeof(buf_lblop),
rinfo->sr.label, nexthop->sr.label);
} else {
strlcpy(buf_sid, "-", sizeof(buf_sid));
strlcpy(buf_lblop, "-", sizeof(buf_lblop));
}
if (first) {
ttable_add_row(tt, "%s|%u|%s|%s|%s|%s",
buf_prefix, rinfo->cost,
buf_iface, buf_nhop, buf_sid,
buf_lblop);
first = false;
} else
ttable_add_row(tt, "||%s|%s|%s|%s", buf_iface,
buf_nhop, buf_sid, buf_lblop);
} else {
char buf_labels[BUFSIZ] = {};
if (nexthop->label_stack) {
for (int i = 0;
i < nexthop->label_stack->num_labels;
i++) {
char buf_label[BUFSIZ];
label2str(
nexthop->label_stack->label[i],
buf_label, sizeof(buf_label));
if (i != 0)
strlcat(buf_labels, "/",
sizeof(buf_labels));
strlcat(buf_labels, buf_label,
sizeof(buf_labels));
}
} else if (nexthop->sr.present)
label2str(nexthop->sr.label, buf_labels,
sizeof(buf_labels));
else
strlcpy(buf_labels, "-", sizeof(buf_labels));
if (first) {
ttable_add_row(tt, "%s|%u|%s|%s|%s", buf_prefix,
rinfo->cost, buf_iface, buf_nhop,
buf_labels);
first = false;
} else
ttable_add_row(tt, "||%s|%s|%s", buf_iface,
buf_nhop, buf_labels);
}
}
if (list_isempty(rinfo->nexthops)) {
if (prefix_sid) {
char buf_sid[BUFSIZ] = {};
char buf_lblop[BUFSIZ] = {};
if (rinfo->sr.present) {
snprintf(buf_sid, sizeof(buf_sid), "%u",
rinfo->sr.sid.value);
sr_op2str(buf_lblop, sizeof(buf_lblop),
rinfo->sr.label,
MPLS_LABEL_IMPLICIT_NULL);
} else {
strlcpy(buf_sid, "-", sizeof(buf_sid));
strlcpy(buf_lblop, "-", sizeof(buf_lblop));
}
ttable_add_row(tt, "%s|%u|%s|%s|%s|%s", buf_prefix,
rinfo->cost, "-", "-", buf_sid,
buf_lblop);
} else
ttable_add_row(tt, "%s|%u|%s|%s|%s", buf_prefix,
rinfo->cost, "-", "-", "-");
}
}
void isis_print_routes(struct vty *vty, struct isis_spftree *spftree,
bool prefix_sid, bool backup)
{
struct route_table *route_table;
struct ttable *tt;
struct route_node *rn;
bool no_adjacencies = false;
const char *tree_id_text = NULL;
if (!spftree)
return;
switch (spftree->tree_id) {
case SPFTREE_IPV4:
tree_id_text = "IPv4";
break;
case SPFTREE_IPV6:
tree_id_text = "IPv6";
break;
case SPFTREE_DSTSRC:
tree_id_text = "IPv6 (dst-src routing)";
break;
case SPFTREE_COUNT:
assert(!"isis_print_routes shouldn't be called with SPFTREE_COUNT as type");
return;
}
vty_out(vty, "IS-IS %s %s routing table:\n\n",
circuit_t2string(spftree->level), tree_id_text);
/* Prepare table. */
tt = ttable_new(&ttable_styles[TTSTYLE_BLANK]);
if (prefix_sid)
ttable_add_row(tt, "Prefix|Metric|Interface|Nexthop|SID|Label Op.");
else
ttable_add_row(tt, "Prefix|Metric|Interface|Nexthop|Label(s)");
tt->style.cell.rpad = 2;
tt->style.corner = '+';
ttable_restyle(tt);
ttable_rowseps(tt, 0, BOTTOM, true, '-');
if (CHECK_FLAG(spftree->flags, F_SPFTREE_NO_ADJACENCIES))
no_adjacencies = true;
route_table =
(backup) ? spftree->route_table_backup : spftree->route_table;
for (rn = route_top(route_table); rn; rn = route_next(rn)) {
struct isis_route_info *rinfo;
rinfo = rn->info;
if (!rinfo)
continue;
isis_print_route(tt, &rn->p, rinfo, prefix_sid, no_adjacencies);
}
/* Dump the generated table. */
if (tt->nrows > 1) {
char *table;
table = ttable_dump(tt, "\n");
vty_out(vty, "%s\n", table);
XFREE(MTYPE_TMP, table);
}
ttable_del(tt);
}
static void show_isis_route_common(struct vty *vty, int levels,
struct isis *isis, bool prefix_sid,
bool backup)
{
struct listnode *node;
struct isis_area *area;
if (!isis->area_list || isis->area_list->count == 0)
return;
for (ALL_LIST_ELEMENTS_RO(isis->area_list, node, area)) {
vty_out(vty, "Area %s:\n",
area->area_tag ? area->area_tag : "null");
for (int level = ISIS_LEVEL1; level <= ISIS_LEVELS; level++) {
if ((level & levels) == 0)
continue;
if (area->ip_circuits > 0) {
isis_print_routes(
vty,
area->spftree[SPFTREE_IPV4][level - 1],
prefix_sid, backup);
}
if (area->ipv6_circuits > 0) {
isis_print_routes(
vty,
area->spftree[SPFTREE_IPV6][level - 1],
prefix_sid, backup);
}
if (isis_area_ipv6_dstsrc_enabled(area)) {
isis_print_routes(vty,
area->spftree[SPFTREE_DSTSRC]
[level - 1],
prefix_sid, backup);
}
}
}
}
DEFUN(show_isis_route, show_isis_route_cmd,
"show " PROTO_NAME
" [vrf <NAME|all>] route"
#ifndef FABRICD
" [<level-1|level-2>]"
#endif
" [<prefix-sid|backup>]",
SHOW_STR PROTO_HELP VRF_FULL_CMD_HELP_STR
"IS-IS routing table\n"
#ifndef FABRICD
"level-1 routes\n"
"level-2 routes\n"
#endif
"Show Prefix-SID information\n"
"Show backup routes\n")
{
int levels;
struct isis *isis;
struct listnode *node;
const char *vrf_name = VRF_DEFAULT_NAME;
bool all_vrf = false;
bool prefix_sid = false;
bool backup = false;
int idx = 0;
if (argv_find(argv, argc, "level-1", &idx))
levels = ISIS_LEVEL1;
else if (argv_find(argv, argc, "level-2", &idx))
levels = ISIS_LEVEL2;
else
levels = ISIS_LEVEL1 | ISIS_LEVEL2;
if (!im) {
vty_out(vty, "IS-IS Routing Process not enabled\n");
return CMD_SUCCESS;
}
ISIS_FIND_VRF_ARGS(argv, argc, idx, vrf_name, all_vrf);
if (argv_find(argv, argc, "prefix-sid", &idx))
prefix_sid = true;
if (argv_find(argv, argc, "backup", &idx))
backup = true;
if (vrf_name) {
if (all_vrf) {
for (ALL_LIST_ELEMENTS_RO(im->isis, node, isis))
show_isis_route_common(vty, levels, isis,
prefix_sid, backup);
return CMD_SUCCESS;
}
isis = isis_lookup_by_vrfname(vrf_name);
if (isis != NULL)
show_isis_route_common(vty, levels, isis, prefix_sid,
backup);
}
return CMD_SUCCESS;
}
static void isis_print_frr_summary_line(struct ttable *tt,
const char *protection,
uint32_t counters[SPF_PREFIX_PRIO_MAX])
{
uint32_t critical, high, medium, low, total;
critical = counters[SPF_PREFIX_PRIO_CRITICAL];
high = counters[SPF_PREFIX_PRIO_HIGH];
medium = counters[SPF_PREFIX_PRIO_MEDIUM];
low = counters[SPF_PREFIX_PRIO_LOW];
total = critical + high + medium + low;
ttable_add_row(tt, "%s|%u|%u|%u|%u|%u", protection, critical, high,
medium, low, total);
}
static void
isis_print_frr_summary_line_coverage(struct ttable *tt, const char *protection,
double counters[SPF_PREFIX_PRIO_MAX],
double total)
{
double critical, high, medium, low;
critical = counters[SPF_PREFIX_PRIO_CRITICAL] * 100;
high = counters[SPF_PREFIX_PRIO_HIGH] * 100;
medium = counters[SPF_PREFIX_PRIO_MEDIUM] * 100;
low = counters[SPF_PREFIX_PRIO_LOW] * 100;
total *= 100;
ttable_add_row(tt, "%s|%.2f%%|%.2f%%|%.2f%%|%.2f%%|%.2f%%", protection,
critical, high, medium, low, total);
}
static void isis_print_frr_summary(struct vty *vty,
struct isis_spftree *spftree)
{
struct ttable *tt;
char *table;
const char *tree_id_text = NULL;
uint32_t protectd[SPF_PREFIX_PRIO_MAX] = {0};
uint32_t unprotected[SPF_PREFIX_PRIO_MAX] = {0};
double coverage[SPF_PREFIX_PRIO_MAX] = {0};
uint32_t protected_total = 0, grand_total = 0;
double coverage_total;
if (!spftree)
return;
switch (spftree->tree_id) {
case SPFTREE_IPV4:
tree_id_text = "IPv4";
break;
case SPFTREE_IPV6:
tree_id_text = "IPv6";
break;
case SPFTREE_DSTSRC:
tree_id_text = "IPv6 (dst-src routing)";
break;
case SPFTREE_COUNT:
assert(!"isis_print_frr_summary shouldn't be called with SPFTREE_COUNT as type");
return;
}
vty_out(vty, " IS-IS %s %s Fast ReRoute summary:\n\n",
circuit_t2string(spftree->level), tree_id_text);
/* Prepare table. */
tt = ttable_new(&ttable_styles[TTSTYLE_BLANK]);
ttable_add_row(
tt,
"Protection \\ Priority|Critical|High |Medium |Low |Total");
tt->style.cell.rpad = 2;
tt->style.corner = '+';
ttable_restyle(tt);
ttable_rowseps(tt, 0, BOTTOM, true, '-');
/* Compute unprotected and coverage totals. */
for (int priority = SPF_PREFIX_PRIO_CRITICAL;
priority < SPF_PREFIX_PRIO_MAX; priority++) {
uint32_t *lfa = spftree->lfa.protection_counters.lfa;
uint32_t *rlfa = spftree->lfa.protection_counters.rlfa;
uint32_t *tilfa = spftree->lfa.protection_counters.tilfa;
uint32_t *ecmp = spftree->lfa.protection_counters.ecmp;
uint32_t *total = spftree->lfa.protection_counters.total;
protectd[priority] = lfa[priority] + rlfa[priority]
+ tilfa[priority] + ecmp[priority];
/* Safeguard to protect against possible inconsistencies. */
if (protectd[priority] > total[priority])
protectd[priority] = total[priority];
unprotected[priority] = total[priority] - protectd[priority];
protected_total += protectd[priority];
grand_total += total[priority];
if (!total[priority])
coverage[priority] = 0;
else
coverage[priority] =
protectd[priority] / (double)total[priority];
}
if (!grand_total)
coverage_total = 0;
else
coverage_total = protected_total / (double)grand_total;
/* Add rows. */
isis_print_frr_summary_line(tt, "Classic LFA",
spftree->lfa.protection_counters.lfa);
isis_print_frr_summary_line(tt, "Remote LFA",
spftree->lfa.protection_counters.rlfa);
isis_print_frr_summary_line(tt, "Topology Independent LFA",
spftree->lfa.protection_counters.tilfa);
isis_print_frr_summary_line(tt, "ECMP",
spftree->lfa.protection_counters.ecmp);
isis_print_frr_summary_line(tt, "Unprotected", unprotected);
isis_print_frr_summary_line_coverage(tt, "Protection coverage",
coverage, coverage_total);
/* Dump the generated table. */
table = ttable_dump(tt, "\n");
vty_out(vty, "%s\n", table);
XFREE(MTYPE_TMP, table);
ttable_del(tt);
}
static void show_isis_frr_summary_common(struct vty *vty, int levels,
struct isis *isis)
{
struct listnode *node;
struct isis_area *area;
if (!isis->area_list || isis->area_list->count == 0)
return;
for (ALL_LIST_ELEMENTS_RO(isis->area_list, node, area)) {
vty_out(vty, "Area %s:\n",
area->area_tag ? area->area_tag : "null");
for (int level = ISIS_LEVEL1; level <= ISIS_LEVELS; level++) {
if ((level & levels) == 0)
continue;
if (area->ip_circuits > 0) {
isis_print_frr_summary(
vty,
area->spftree[SPFTREE_IPV4][level - 1]);
}
if (area->ipv6_circuits > 0) {
isis_print_frr_summary(
vty,
area->spftree[SPFTREE_IPV6][level - 1]);
}
if (isis_area_ipv6_dstsrc_enabled(area)) {
isis_print_frr_summary(
vty, area->spftree[SPFTREE_DSTSRC]
[level - 1]);
}
}
}
}
DEFUN(show_isis_frr_summary, show_isis_frr_summary_cmd,
"show " PROTO_NAME
" [vrf <NAME|all>] fast-reroute summary"
#ifndef FABRICD
" [<level-1|level-2>]"
#endif
,
SHOW_STR PROTO_HELP VRF_FULL_CMD_HELP_STR
"IS-IS FRR information\n"
"FRR summary\n"
#ifndef FABRICD
"level-1 routes\n"
"level-2 routes\n"
#endif
)
{
int levels;
struct isis *isis;
struct listnode *node;
const char *vrf_name = VRF_DEFAULT_NAME;
bool all_vrf = false;
int idx = 0;
if (argv_find(argv, argc, "level-1", &idx))
levels = ISIS_LEVEL1;
else if (argv_find(argv, argc, "level-2", &idx))
levels = ISIS_LEVEL2;
else
levels = ISIS_LEVEL1 | ISIS_LEVEL2;
if (!im) {
vty_out(vty, "IS-IS Routing Process not enabled\n");
return CMD_SUCCESS;
}
ISIS_FIND_VRF_ARGS(argv, argc, idx, vrf_name, all_vrf);
if (vrf_name) {
if (all_vrf) {
for (ALL_LIST_ELEMENTS_RO(im->isis, node, isis))
show_isis_frr_summary_common(vty, levels, isis);
return CMD_SUCCESS;
}
isis = isis_lookup_by_vrfname(vrf_name);
if (isis != NULL)
show_isis_frr_summary_common(vty, levels, isis);
}
return CMD_SUCCESS;
}
void isis_spf_init(void)
2003-12-23 09:09:43 +01:00
{
install_element(VIEW_NODE, &show_isis_topology_cmd);
install_element(VIEW_NODE, &show_isis_route_cmd);
install_element(VIEW_NODE, &show_isis_frr_summary_cmd);
/* Register hook(s). */
hook_register(isis_adj_state_change_hook, spf_adj_state_change);
2003-12-23 09:09:43 +01:00
}
void isis_spf_print(struct isis_spftree *spftree, struct vty *vty)
{
uint64_t last_run_duration = spftree->last_run_duration;
vty_out(vty, " last run elapsed : ");
vty_out_timestr(vty, spftree->last_run_timestamp);
vty_out(vty, "\n");
vty_out(vty, " last run duration : %" PRIu64 " usec\n",
last_run_duration);
vty_out(vty, " run count : %u\n", spftree->runcount);
}
void isis_spf_print_json(struct isis_spftree *spftree, struct json_object *json)
{
char uptime[MONOTIME_STRLEN];
time_t cur;
cur = time(NULL);
cur -= spftree->last_run_timestamp;
frrtime_to_interval(cur, uptime, sizeof(uptime));
json_object_string_add(json, "last-run-elapsed", uptime);
json_object_int_add(json, "last-run-duration-usec",
spftree->last_run_duration);
json_object_int_add(json, "last-run-count", spftree->runcount);
}