frr/lib/routemap.c

3266 lines
85 KiB
C
Raw Normal View History

2002-12-13 21:15:29 +01:00
/* Route map function.
* Copyright (C) 1998, 1999 Kunihiro Ishiguro
*
* This file is part of GNU Zebra.
*
* GNU Zebra is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2, or (at your option) any
* later version.
*
* GNU Zebra is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with this program; see the file COPYING; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
2002-12-13 21:15:29 +01:00
#include <zebra.h>
#include "linklist.h"
#include "memory.h"
#include "vector.h"
#include "prefix.h"
#include "vty.h"
2002-12-13 21:15:29 +01:00
#include "routemap.h"
#include "command.h"
2004-09-13 Jose Luis Rubio <jrubio@dit.upm.es> (at Technical University of Madrid as part of Euro6ix Project) Enhanced Route Server functionality and Route-Maps: * bgpd/bgpd.h: Modified 'struct peer' and 'struct bgp_filter' to support rs-clients. A 'struct bgp_table *rib' has been added to the first (to mantain a separated RIB for each rs-client) and two new route-maps have been added to the last (for import/export policies). Added the following #defines: RMAP_{IN|OUT|IMPORT|EXPORT|MAX}, PEER_RMAP_TYPE_{IMPORT|EXPORT} and BGP_CLEAR_SOFT_RSCLIENT. * bgpd/bgpd.c: Modified the functions that create/delete/etc peers in order to consider the new fields included in 'struct peer' for supporting rs-clients, i.e. the import/export route-maps and the 'struct bgp_table'. * bgpd/bgp_route.{ch}: Modified several functions related with receiving/sending announces in order to support the new Route Server capabilities. Function 'bgp_process' has been reorganized, creating an auxiliar function for best path selection ('bgp_best_selection'). Modified 'bgp_show' and 'bgp_show_route' for displaying information about any RIB (and not only the main bgp RIB). Added commands for displaying information about RS-clients RIBs: 'show bgp rsclient (A.B.C.D|X:X::X:X)', 'show bgp rsclient (A.B.C.D|X:X::X:X) X:X::X:X/M', etc * bgpd/bgp_table.{ch}: The structure 'struct bgp_table' now has two new fields: type (which can take the values BGP_TABLE_{MAIN|RSCLIENT}) and 'void *owner' which points to 'struct bgp' or 'struct peer' which owns the table. When creating a new bgp_table by default 'type=BGP_TABLE_MAIN' is set. * bgpd/bgp_vty.c: The commands 'neighbor ... route-server-client' and 'no neighbor ... route-server-client' now not only set/unset the flag PEER_FLAG_RSERVER_CLIENT, but they create/destroy the 'struct bgp_table' of the peer. Special actions are taken for peer_groups. Command 'neighbor ... route-map WORD (in|out)' now also supports two new kinds of route-map: 'import' and 'export'. Added commands 'clear bgp * rsclient', etc. These commands allow a new kind of soft_reconfig which affects only the RIB of the specified RS-client. Added commands 'show bgp rsclient summary', etc which display a summary of the rs-clients configured for the corresponding address family. * bgpd/bgp_routemap.c: A new match statement is available, 'match peer (A.B.C.D|X:X::X:X)'. This statement can only be used in import/export route-maps, and it matches when the peer who announces (when used in an import route-map) or is going to receive (when used in an export route-map) the route is the same than the one specified in the statement. For peer-groups the statement matches if the specified peer is member of the peer-group. A special version of the command, 'match peer local', matches with routes originated by the Route Server (defined with 'network ...', redistributed routes and default-originate). * lib/routemap.{ch}: Added a new clause 'call NAME' for use in route-maps. It jumps into the specified route-map and when it returns the first route-map ends if the called RM returns DENY_MATCH, or continues in other case.
2004-09-13 07:12:46 +02:00
#include "log.h"
#include "hash.h"
#include "libfrr.h"
#include "lib_errors.h"
#include "table.h"
DEFINE_MTYPE_STATIC(LIB, ROUTE_MAP, "Route map")
DEFINE_MTYPE(LIB, ROUTE_MAP_NAME, "Route map name")
DEFINE_MTYPE_STATIC(LIB, ROUTE_MAP_INDEX, "Route map index")
DEFINE_MTYPE(LIB, ROUTE_MAP_RULE, "Route map rule")
DEFINE_MTYPE_STATIC(LIB, ROUTE_MAP_RULE_STR, "Route map rule str")
DEFINE_MTYPE(LIB, ROUTE_MAP_COMPILED, "Route map compiled")
DEFINE_MTYPE_STATIC(LIB, ROUTE_MAP_DEP, "Route map dependency")
DEFINE_MTYPE_STATIC(LIB, ROUTE_MAP_DEP_DATA, "Route map dependency data")
DEFINE_QOBJ_TYPE(route_map_index)
DEFINE_QOBJ_TYPE(route_map)
#define IPv4_PREFIX_LIST "ip address prefix-list"
#define IPv6_PREFIX_LIST "ipv6 address prefix-list"
#define IPv4_MATCH_RULE "ip "
#define IPv6_MATCH_RULE "ipv6 "
#define IS_RULE_IPv4_PREFIX_LIST(S) \
(strncmp(S, IPv4_PREFIX_LIST, strlen(IPv4_PREFIX_LIST)) == 0)
#define IS_RULE_IPv6_PREFIX_LIST(S) \
(strncmp(S, IPv6_PREFIX_LIST, strlen(IPv6_PREFIX_LIST)) == 0)
#define IS_IPv4_RULE(S) \
(strncmp(S, IPv4_MATCH_RULE, strlen(IPv4_MATCH_RULE)) == 0)
#define IS_IPv6_RULE(S) \
(strncmp(S, IPv6_MATCH_RULE, strlen(IPv6_MATCH_RULE)) == 0)
struct route_map_pentry_dep {
struct prefix_list_entry *pentry;
const char *plist_name;
route_map_event_t event;
};
2002-12-13 21:15:29 +01:00
/* Vector for route match rules. */
static vector route_match_vec;
/* Vector for route set rules. */
static vector route_set_vec;
static void route_map_pfx_tbl_update(route_map_event_t event,
struct route_map_index *index, afi_t afi,
const char *plist_name);
static void route_map_pfx_table_add_default(afi_t afi,
struct route_map_index *index);
static void route_map_pfx_table_del_default(afi_t afi,
struct route_map_index *index);
static void route_map_add_plist_entries(afi_t afi,
struct route_map_index *index,
const char *plist_name,
struct prefix_list_entry *entry);
static void route_map_del_plist_entries(afi_t afi,
struct route_map_index *index,
const char *plist_name,
struct prefix_list_entry *entry);
static bool route_map_is_ip_rule_present(struct route_map_index *index);
static bool route_map_is_ipv6_rule_present(struct route_map_index *index);
static struct hash *route_map_get_dep_hash(route_map_event_t event);
struct route_map_match_set_hooks rmap_match_set_hook;
/* match interface */
void route_map_match_interface_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_interface = func;
}
/* no match interface */
void route_map_no_match_interface_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_interface = func;
}
/* match ip address */
void route_map_match_ip_address_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ip_address = func;
}
/* no match ip address */
void route_map_no_match_ip_address_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ip_address = func;
}
/* match ip address prefix list */
void route_map_match_ip_address_prefix_list_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ip_address_prefix_list = func;
}
/* no match ip address prefix list */
void route_map_no_match_ip_address_prefix_list_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ip_address_prefix_list = func;
}
/* match ip next hop */
void route_map_match_ip_next_hop_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ip_next_hop = func;
}
/* no match ip next hop */
void route_map_no_match_ip_next_hop_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ip_next_hop = func;
}
/* match ip next hop prefix list */
void route_map_match_ip_next_hop_prefix_list_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ip_next_hop_prefix_list = func;
}
/* no match ip next hop prefix list */
void route_map_no_match_ip_next_hop_prefix_list_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ip_next_hop_prefix_list = func;
}
/* match ip next-hop type */
void route_map_match_ip_next_hop_type_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ip_next_hop_type = func;
}
/* no match ip next-hop type */
void route_map_no_match_ip_next_hop_type_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ip_next_hop_type = func;
}
/* match ipv6 address */
void route_map_match_ipv6_address_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ipv6_address = func;
}
/* no match ipv6 address */
void route_map_no_match_ipv6_address_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ipv6_address = func;
}
/* match ipv6 address prefix list */
void route_map_match_ipv6_address_prefix_list_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ipv6_address_prefix_list = func;
}
/* no match ipv6 address prefix list */
void route_map_no_match_ipv6_address_prefix_list_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ipv6_address_prefix_list = func;
}
/* match ipv6 next-hop type */
void route_map_match_ipv6_next_hop_type_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_ipv6_next_hop_type = func;
}
/* no match ipv6 next-hop type */
void route_map_no_match_ipv6_next_hop_type_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_ipv6_next_hop_type = func;
}
/* match metric */
void route_map_match_metric_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.match_metric = func;
}
/* no match metric */
void route_map_no_match_metric_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_metric = func;
}
/* match tag */
void route_map_match_tag_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command, const char *arg,
route_map_event_t type))
{
rmap_match_set_hook.match_tag = func;
}
/* no match tag */
void route_map_no_match_tag_hook(int (*func)(
struct vty *vty, struct route_map_index *index, const char *command,
const char *arg, route_map_event_t type))
{
rmap_match_set_hook.no_match_tag = func;
}
/* set ip nexthop */
void route_map_set_ip_nexthop_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command,
const char *arg))
{
rmap_match_set_hook.set_ip_nexthop = func;
}
/* no set ip nexthop */
void route_map_no_set_ip_nexthop_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command,
const char *arg))
{
rmap_match_set_hook.no_set_ip_nexthop = func;
}
/* set ipv6 nexthop local */
void route_map_set_ipv6_nexthop_local_hook(
int (*func)(struct vty *vty, struct route_map_index *index,
const char *command, const char *arg))
{
rmap_match_set_hook.set_ipv6_nexthop_local = func;
}
/* no set ipv6 nexthop local */
void route_map_no_set_ipv6_nexthop_local_hook(
int (*func)(struct vty *vty, struct route_map_index *index,
const char *command, const char *arg))
{
rmap_match_set_hook.no_set_ipv6_nexthop_local = func;
}
/* set metric */
void route_map_set_metric_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command,
const char *arg))
{
rmap_match_set_hook.set_metric = func;
}
/* no set metric */
void route_map_no_set_metric_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command,
const char *arg))
{
rmap_match_set_hook.no_set_metric = func;
}
/* set tag */
void route_map_set_tag_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command, const char *arg))
{
rmap_match_set_hook.set_tag = func;
}
/* no set tag */
void route_map_no_set_tag_hook(int (*func)(struct vty *vty,
struct route_map_index *index,
const char *command,
const char *arg))
{
rmap_match_set_hook.no_set_tag = func;
}
int generic_match_add(struct vty *vty, struct route_map_index *index,
const char *command, const char *arg,
route_map_event_t type)
{
enum rmap_compile_rets ret;
ret = route_map_add_match(index, command, arg, type);
switch (ret) {
case RMAP_RULE_MISSING:
if (vty)
vty_out(vty, "%% [%s] Can't find rule.\n",
frr_protonameinst);
else
zlog_warn("Can't find rule: %s", command);
return CMD_WARNING_CONFIG_FAILED;
case RMAP_COMPILE_ERROR:
if (vty)
vty_out(vty,
"%% [%s] Argument form is unsupported or malformed.\n",
frr_protonameinst);
else
zlog_warn("Argument form is unsupported or malformed: "
"%s %s", command, arg);
return CMD_WARNING_CONFIG_FAILED;
case RMAP_COMPILE_SUCCESS:
/*
* Nothing to do here move along
*/
break;
}
return CMD_SUCCESS;
}
int generic_match_delete(struct vty *vty, struct route_map_index *index,
const char *command, const char *arg,
route_map_event_t type)
{
enum rmap_compile_rets ret;
int retval = CMD_SUCCESS;
char *dep_name = NULL;
const char *tmpstr;
char *rmap_name = NULL;
if (type != RMAP_EVENT_MATCH_DELETED) {
/* ignore the mundane, the types without any dependency */
if (arg == NULL) {
if ((tmpstr = route_map_get_match_arg(index, command))
!= NULL)
dep_name =
XSTRDUP(MTYPE_ROUTE_MAP_RULE, tmpstr);
} else {
dep_name = XSTRDUP(MTYPE_ROUTE_MAP_RULE, arg);
}
rmap_name = XSTRDUP(MTYPE_ROUTE_MAP_NAME, index->map->name);
}
ret = route_map_delete_match(index, command, dep_name, type);
switch (ret) {
case RMAP_RULE_MISSING:
if (vty)
vty_out(vty, "%% [%s] Can't find rule.\n",
frr_protonameinst);
else
zlog_warn("Can't find rule: %s", command);
retval = CMD_WARNING_CONFIG_FAILED;
break;
case RMAP_COMPILE_ERROR:
if (vty)
vty_out(vty,
"%% [%s] Argument form is unsupported or malformed.\n",
frr_protonameinst);
else
zlog_warn("Argument form is unsupported or malformed: "
"%s %s", command, arg);
retval = CMD_WARNING_CONFIG_FAILED;
break;
case RMAP_COMPILE_SUCCESS:
/*
* Nothing to do here
*/
break;
}
XFREE(MTYPE_ROUTE_MAP_RULE, dep_name);
XFREE(MTYPE_ROUTE_MAP_NAME, rmap_name);
return retval;
}
int generic_set_add(struct vty *vty, struct route_map_index *index,
const char *command, const char *arg)
2002-12-13 21:15:29 +01:00
{
enum rmap_compile_rets ret;
ret = route_map_add_set(index, command, arg);
switch (ret) {
case RMAP_RULE_MISSING:
if (vty)
vty_out(vty, "%% [%s] Can't find rule.\n", frr_protonameinst);
else
zlog_warn("Can't find rule: %s", command);
return CMD_WARNING_CONFIG_FAILED;
case RMAP_COMPILE_ERROR:
if (vty)
vty_out(vty,
"%% [%s] Argument form is unsupported or malformed.\n",
frr_protonameinst);
else
zlog_warn("Argument form is unsupported or malformed: "
"%s %s", command, arg);
return CMD_WARNING_CONFIG_FAILED;
case RMAP_COMPILE_SUCCESS:
break;
}
return CMD_SUCCESS;
}
int generic_set_delete(struct vty *vty, struct route_map_index *index,
const char *command, const char *arg)
{
enum rmap_compile_rets ret;
ret = route_map_delete_set(index, command, arg);
switch (ret) {
case RMAP_RULE_MISSING:
if (vty)
vty_out(vty, "%% [%s] Can't find rule.\n", frr_protonameinst);
else
zlog_warn("Can't find rule: %s", command);
return CMD_WARNING_CONFIG_FAILED;
case RMAP_COMPILE_ERROR:
if (vty)
vty_out(vty,
"%% [%s] Argument form is unsupported or malformed.\n",
frr_protonameinst);
else
zlog_warn("Argument form is unsupported or malformed: "
"%s %s", command, arg);
return CMD_WARNING_CONFIG_FAILED;
case RMAP_COMPILE_SUCCESS:
break;
}
return CMD_SUCCESS;
}
2002-12-13 21:15:29 +01:00
/* Master list of route map. */
struct route_map_list route_map_master = {NULL, NULL, NULL, NULL, NULL};
struct hash *route_map_master_hash = NULL;
static unsigned int route_map_hash_key_make(const void *p)
{
const struct route_map *map = p;
return string_hash_make(map->name);
}
static bool route_map_hash_cmp(const void *p1, const void *p2)
{
const struct route_map *map1 = p1;
const struct route_map *map2 = p2;
if (map1->deleted == map2->deleted) {
if (map1->name && map2->name) {
if (!strcmp(map1->name, map2->name)) {
return true;
}
} else if (!map1->name && !map2->name) {
return true;
}
}
return false;
}
enum route_map_upd8_type {
ROUTE_MAP_ADD = 1,
ROUTE_MAP_DEL,
};
/* all possible route-map dependency types */
enum route_map_dep_type {
ROUTE_MAP_DEP_RMAP = 1,
ROUTE_MAP_DEP_CLIST,
ROUTE_MAP_DEP_ECLIST,
ROUTE_MAP_DEP_LCLIST,
ROUTE_MAP_DEP_PLIST,
ROUTE_MAP_DEP_ASPATH,
ROUTE_MAP_DEP_FILTER,
ROUTE_MAP_DEP_MAX,
};
struct route_map_dep {
char *dep_name;
struct hash *dep_rmap_hash;
struct hash *this_hash; /* ptr to the hash structure this is part of */
};
2002-12-13 21:15:29 +01:00
struct route_map_dep_data {
/* Route-map name.
*/
char *rname;
/* Count of number of sequences of this
* route-map that depend on the same entity.
*/
uint16_t refcnt;
};
/* Hashes maintaining dependency between various sublists used by route maps */
static struct hash *route_map_dep_hash[ROUTE_MAP_DEP_MAX];
static unsigned int route_map_dep_hash_make_key(const void *p);
static void route_map_clear_all_references(char *rmap_name);
static void route_map_rule_delete(struct route_map_rule_list *,
struct route_map_rule *);
static bool rmap_debug;
2002-12-13 21:15:29 +01:00
/* New route map allocation. Please note route map's name must be
specified. */
static struct route_map *route_map_new(const char *name)
2002-12-13 21:15:29 +01:00
{
struct route_map *new;
2002-12-13 21:15:29 +01:00
new = XCALLOC(MTYPE_ROUTE_MAP, sizeof(struct route_map));
new->name = XSTRDUP(MTYPE_ROUTE_MAP_NAME, name);
QOBJ_REG(new, route_map);
return new;
2002-12-13 21:15:29 +01:00
}
/* Add new name to route_map. */
static struct route_map *route_map_add(const char *name)
{
struct route_map *map;
struct route_map_list *list;
map = route_map_new(name);
list = &route_map_master;
/* Add map to the hash */
hash_get(route_map_master_hash, map, hash_alloc_intern);
/* Add new entry to the head of the list to match how it is added in the
* hash table. This is to ensure that if the same route-map has been
* created more than once and then marked for deletion (which can happen
* if prior deletions haven't completed as BGP hasn't yet done the
* route-map processing), the order of the entities is the same in both
* the list and the hash table. Otherwise, since there is nothing to
* distinguish between the two entries, the wrong entry could get freed.
* TODO: This needs to be re-examined to handle it better - e.g., revive
* a deleted entry if the route-map is created again.
*/
map->prev = NULL;
map->next = list->head;
if (list->head)
list->head->prev = map;
list->head = map;
if (!list->tail)
list->tail = map;
/* Execute hook. */
if (route_map_master.add_hook) {
(*route_map_master.add_hook)(name);
route_map_notify_dependencies(name, RMAP_EVENT_CALL_ADDED);
}
if (!map->ipv4_prefix_table)
map->ipv4_prefix_table = route_table_init();
if (!map->ipv6_prefix_table)
map->ipv6_prefix_table = route_table_init();
if (rmap_debug)
zlog_debug("Add route-map %s", name);
return map;
2002-12-13 21:15:29 +01:00
}
/* this is supposed to be called post processing by
* the delete hook function. Don't invoke delete_hook
* again in this routine.
*/
static void route_map_free_map(struct route_map *map)
2002-12-13 21:15:29 +01:00
{
struct route_map_list *list;
struct route_map_index *index;
if (map == NULL)
return;
while ((index = map->head) != NULL)
route_map_index_delete(index, 0);
2002-12-13 21:15:29 +01:00
if (rmap_debug)
zlog_debug("Deleting route-map %s", map->name);
list = &route_map_master;
2002-12-13 21:15:29 +01:00
QOBJ_UNREG(map);
if (map->next)
map->next->prev = map->prev;
else
list->tail = map->prev;
2002-12-13 21:15:29 +01:00
if (map->prev)
map->prev->next = map->next;
else
list->head = map->next;
2002-12-13 21:15:29 +01:00
hash_release(route_map_master_hash, map);
XFREE(MTYPE_ROUTE_MAP_NAME, map->name);
XFREE(MTYPE_ROUTE_MAP, map);
}
2002-12-13 21:15:29 +01:00
/* Route map delete from list. */
void route_map_delete(struct route_map *map)
{
struct route_map_index *index;
char *name;
while ((index = map->head) != NULL)
route_map_index_delete(index, 0);
name = map->name;
map->head = NULL;
/* Clear all dependencies */
route_map_clear_all_references(name);
map->deleted = true;
/* Execute deletion hook. */
if (route_map_master.delete_hook) {
(*route_map_master.delete_hook)(name);
route_map_notify_dependencies(name, RMAP_EVENT_CALL_DELETED);
}
2002-12-13 21:15:29 +01:00
if (!map->to_be_processed) {
route_map_free_map(map);
}
2002-12-13 21:15:29 +01:00
}
/* Lookup route map by route map name string. */
struct route_map *route_map_lookup_by_name(const char *name)
2002-12-13 21:15:29 +01:00
{
struct route_map *map;
struct route_map tmp_map;
2002-12-13 21:15:29 +01:00
if (!name)
return NULL;
// map.deleted is 0 via memset
memset(&tmp_map, 0, sizeof(struct route_map));
tmp_map.name = XSTRDUP(MTYPE_ROUTE_MAP_NAME, name);
map = hash_lookup(route_map_master_hash, &tmp_map);
XFREE(MTYPE_ROUTE_MAP_NAME, tmp_map.name);
return map;
2002-12-13 21:15:29 +01:00
}
/* Simple helper to warn if route-map does not exist. */
struct route_map *route_map_lookup_warn_noexist(struct vty *vty, const char *name)
{
struct route_map *route_map = route_map_lookup_by_name(name);
if (!route_map)
if (vty_shell_serv(vty))
vty_out(vty, "The route-map '%s' does not exist.\n", name);
return route_map;
}
int route_map_mark_updated(const char *name)
{
struct route_map *map;
int ret = -1;
struct route_map tmp_map;
if (!name)
return (ret);
map = route_map_lookup_by_name(name);
/* If we did not find the routemap with deleted=false try again
* with deleted=true
*/
if (!map) {
memset(&tmp_map, 0, sizeof(struct route_map));
tmp_map.name = XSTRDUP(MTYPE_ROUTE_MAP_NAME, name);
tmp_map.deleted = true;
map = hash_lookup(route_map_master_hash, &tmp_map);
XFREE(MTYPE_ROUTE_MAP_NAME, tmp_map.name);
}
if (map) {
map->to_be_processed = true;
ret = 0;
}
return (ret);
}
static int route_map_clear_updated(struct route_map *map)
{
int ret = -1;
if (map) {
map->to_be_processed = false;
if (map->deleted)
route_map_free_map(map);
}
return (ret);
}
2002-12-13 21:15:29 +01:00
/* Lookup route map. If there isn't route map create one and return
it. */
struct route_map *route_map_get(const char *name)
2002-12-13 21:15:29 +01:00
{
struct route_map *map;
2002-12-13 21:15:29 +01:00
map = route_map_lookup_by_name(name);
if (map == NULL)
map = route_map_add(name);
return map;
2002-12-13 21:15:29 +01:00
}
void route_map_walk_update_list(void (*route_map_update_fn)(char *name))
{
struct route_map *node;
struct route_map *nnode = NULL;
for (node = route_map_master.head; node; node = nnode) {
if (node->to_be_processed) {
/* DD: Should we add any thread yield code here */
route_map_update_fn(node->name);
nnode = node->next;
route_map_clear_updated(node);
} else
nnode = node->next;
}
}
2002-12-13 21:15:29 +01:00
/* Return route map's type string. */
static const char *route_map_type_str(enum route_map_type type)
{
switch (type) {
case RMAP_PERMIT:
return "permit";
case RMAP_DENY:
return "deny";
case RMAP_ANY:
return "";
}
return "";
2002-12-13 21:15:29 +01:00
}
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
static const char *route_map_cmd_result_str(enum route_map_cmd_result_t res)
{
switch (res) {
case RMAP_MATCH:
return "match";
case RMAP_NOMATCH:
return "no match";
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
case RMAP_NOOP:
return "noop";
case RMAP_ERROR:
return "error";
case RMAP_OKAY:
return "okay";
}
return "invalid";
}
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
static const char *route_map_result_str(route_map_result_t res)
{
switch (res) {
case RMAP_DENYMATCH:
return "deny";
case RMAP_PERMITMATCH:
return "permit";
}
return "invalid";
}
/* show route-map */
static void vty_show_route_map_entry(struct vty *vty, struct route_map *map)
{
struct route_map_index *index;
struct route_map_rule *rule;
vty_out(vty, "route-map: %s Invoked: %" PRIu64 " Optimization: %s\n",
map->name, map->applied - map->applied_clear,
map->optimization_disabled ? "disabled" : "enabled");
for (index = map->head; index; index = index->next) {
vty_out(vty, " %s, sequence %d Invoked %" PRIu64 "\n",
route_map_type_str(index->type), index->pref,
index->applied - index->applied_clear);
/* Description */
if (index->description)
vty_out(vty, " Description:\n %s\n",
index->description);
/* Match clauses */
vty_out(vty, " Match clauses:\n");
for (rule = index->match_list.head; rule; rule = rule->next)
vty_out(vty, " %s %s\n", rule->cmd->str,
rule->rule_str);
vty_out(vty, " Set clauses:\n");
for (rule = index->set_list.head; rule; rule = rule->next)
vty_out(vty, " %s %s\n", rule->cmd->str,
rule->rule_str);
/* Call clause */
vty_out(vty, " Call clause:\n");
if (index->nextrm)
vty_out(vty, " Call %s\n", index->nextrm);
/* Exit Policy */
vty_out(vty, " Action:\n");
if (index->exitpolicy == RMAP_GOTO)
vty_out(vty, " Goto %d\n", index->nextpref);
else if (index->exitpolicy == RMAP_NEXT)
vty_out(vty, " Continue to next entry\n");
else if (index->exitpolicy == RMAP_EXIT)
vty_out(vty, " Exit routemap\n");
}
}
static int sort_route_map(const void **map1, const void **map2)
{
const struct route_map *m1 = *map1;
const struct route_map *m2 = *map2;
return strcmp(m1->name, m2->name);
}
static int vty_show_route_map(struct vty *vty, const char *name)
{
struct route_map *map;
vty_out(vty, "%s:\n", frr_protonameinst);
if (name) {
map = route_map_lookup_by_name(name);
if (map) {
vty_show_route_map_entry(vty, map);
return CMD_SUCCESS;
} else {
vty_out(vty, "%s: 'route-map %s' not found\n",
frr_protonameinst, name);
return CMD_SUCCESS;
}
} else {
struct list *maplist = list_new();
struct listnode *ln;
for (map = route_map_master.head; map; map = map->next)
listnode_add(maplist, map);
list_sort(maplist, sort_route_map);
for (ALL_LIST_ELEMENTS_RO(maplist, ln, map))
vty_show_route_map_entry(vty, map);
list_delete(&maplist);
}
return CMD_SUCCESS;
}
/* Unused route map details */
static int vty_show_unused_route_map(struct vty *vty)
{
struct list *maplist = list_new();
struct listnode *ln;
struct route_map *map;
for (map = route_map_master.head; map; map = map->next) {
/* If use_count is zero, No protocol is using this routemap.
* so adding to the list.
*/
if (!map->use_count)
listnode_add(maplist, map);
}
if (maplist->count > 0) {
vty_out(vty, "\n%s:\n", frr_protonameinst);
list_sort(maplist, sort_route_map);
for (ALL_LIST_ELEMENTS_RO(maplist, ln, map))
vty_show_route_map_entry(vty, map);
} else {
vty_out(vty, "\n%s: None\n", frr_protonameinst);
}
list_delete(&maplist);
return CMD_SUCCESS;
}
2002-12-13 21:15:29 +01:00
/* New route map allocation. Please note route map's name must be
specified. */
static struct route_map_index *route_map_index_new(void)
2002-12-13 21:15:29 +01:00
{
struct route_map_index *new;
2002-12-13 21:15:29 +01:00
new = XCALLOC(MTYPE_ROUTE_MAP_INDEX, sizeof(struct route_map_index));
new->exitpolicy = RMAP_EXIT; /* Default to Cisco-style */
TAILQ_INIT(&new->rhclist);
QOBJ_REG(new, route_map_index);
return new;
2002-12-13 21:15:29 +01:00
}
/* Free route map index. */
void route_map_index_delete(struct route_map_index *index, int notify)
2002-12-13 21:15:29 +01:00
{
struct routemap_hook_context *rhc;
struct route_map_rule *rule;
2002-12-13 21:15:29 +01:00
QOBJ_UNREG(index);
if (rmap_debug)
zlog_debug("Deleting route-map %s sequence %d",
index->map->name, index->pref);
/* Free route map northbound hook contexts. */
while ((rhc = TAILQ_FIRST(&index->rhclist)) != NULL)
routemap_hook_context_free(rhc);
/* Free route match. */
while ((rule = index->match_list.head) != NULL) {
if (IS_RULE_IPv4_PREFIX_LIST(rule->cmd->str))
route_map_pfx_tbl_update(RMAP_EVENT_PLIST_DELETED,
index, AFI_IP, rule->rule_str);
else if (IS_RULE_IPv6_PREFIX_LIST(rule->cmd->str))
route_map_pfx_tbl_update(RMAP_EVENT_PLIST_DELETED,
index, AFI_IP6,
rule->rule_str);
route_map_rule_delete(&index->match_list, rule);
}
2002-12-13 21:15:29 +01:00
/* Free route set. */
while ((rule = index->set_list.head) != NULL)
route_map_rule_delete(&index->set_list, rule);
2002-12-13 21:15:29 +01:00
/* Remove index from route map list. */
if (index->next)
index->next->prev = index->prev;
else
index->map->tail = index->prev;
2002-12-13 21:15:29 +01:00
if (index->prev)
index->prev->next = index->next;
else
index->map->head = index->next;
2002-12-13 21:15:29 +01:00
/* Free 'char *nextrm' if not NULL */
XFREE(MTYPE_ROUTE_MAP_NAME, index->nextrm);
2004-09-13 Jose Luis Rubio <jrubio@dit.upm.es> (at Technical University of Madrid as part of Euro6ix Project) Enhanced Route Server functionality and Route-Maps: * bgpd/bgpd.h: Modified 'struct peer' and 'struct bgp_filter' to support rs-clients. A 'struct bgp_table *rib' has been added to the first (to mantain a separated RIB for each rs-client) and two new route-maps have been added to the last (for import/export policies). Added the following #defines: RMAP_{IN|OUT|IMPORT|EXPORT|MAX}, PEER_RMAP_TYPE_{IMPORT|EXPORT} and BGP_CLEAR_SOFT_RSCLIENT. * bgpd/bgpd.c: Modified the functions that create/delete/etc peers in order to consider the new fields included in 'struct peer' for supporting rs-clients, i.e. the import/export route-maps and the 'struct bgp_table'. * bgpd/bgp_route.{ch}: Modified several functions related with receiving/sending announces in order to support the new Route Server capabilities. Function 'bgp_process' has been reorganized, creating an auxiliar function for best path selection ('bgp_best_selection'). Modified 'bgp_show' and 'bgp_show_route' for displaying information about any RIB (and not only the main bgp RIB). Added commands for displaying information about RS-clients RIBs: 'show bgp rsclient (A.B.C.D|X:X::X:X)', 'show bgp rsclient (A.B.C.D|X:X::X:X) X:X::X:X/M', etc * bgpd/bgp_table.{ch}: The structure 'struct bgp_table' now has two new fields: type (which can take the values BGP_TABLE_{MAIN|RSCLIENT}) and 'void *owner' which points to 'struct bgp' or 'struct peer' which owns the table. When creating a new bgp_table by default 'type=BGP_TABLE_MAIN' is set. * bgpd/bgp_vty.c: The commands 'neighbor ... route-server-client' and 'no neighbor ... route-server-client' now not only set/unset the flag PEER_FLAG_RSERVER_CLIENT, but they create/destroy the 'struct bgp_table' of the peer. Special actions are taken for peer_groups. Command 'neighbor ... route-map WORD (in|out)' now also supports two new kinds of route-map: 'import' and 'export'. Added commands 'clear bgp * rsclient', etc. These commands allow a new kind of soft_reconfig which affects only the RIB of the specified RS-client. Added commands 'show bgp rsclient summary', etc which display a summary of the rs-clients configured for the corresponding address family. * bgpd/bgp_routemap.c: A new match statement is available, 'match peer (A.B.C.D|X:X::X:X)'. This statement can only be used in import/export route-maps, and it matches when the peer who announces (when used in an import route-map) or is going to receive (when used in an export route-map) the route is the same than the one specified in the statement. For peer-groups the statement matches if the specified peer is member of the peer-group. A special version of the command, 'match peer local', matches with routes originated by the Route Server (defined with 'network ...', redistributed routes and default-originate). * lib/routemap.{ch}: Added a new clause 'call NAME' for use in route-maps. It jumps into the specified route-map and when it returns the first route-map ends if the called RM returns DENY_MATCH, or continues in other case.
2004-09-13 07:12:46 +02:00
route_map_pfx_tbl_update(RMAP_EVENT_INDEX_DELETED, index, 0, NULL);
/* Execute event hook. */
if (route_map_master.event_hook && notify) {
(*route_map_master.event_hook)(index->map->name);
route_map_notify_dependencies(index->map->name,
RMAP_EVENT_CALL_ADDED);
}
XFREE(MTYPE_ROUTE_MAP_INDEX, index);
2002-12-13 21:15:29 +01:00
}
/* Lookup index from route map. */
static struct route_map_index *route_map_index_lookup(struct route_map *map,
enum route_map_type type,
int pref)
2002-12-13 21:15:29 +01:00
{
struct route_map_index *index;
2002-12-13 21:15:29 +01:00
for (index = map->head; index; index = index->next)
if ((index->type == type || type == RMAP_ANY)
&& index->pref == pref)
return index;
return NULL;
2002-12-13 21:15:29 +01:00
}
/* Add new index to route map. */
static struct route_map_index *
route_map_index_add(struct route_map *map, enum route_map_type type, int pref)
{
struct route_map_index *index;
struct route_map_index *point;
/* Allocate new route map inex. */
index = route_map_index_new();
index->map = map;
index->type = type;
index->pref = pref;
/* Compare preference. */
for (point = map->head; point; point = point->next)
if (point->pref >= pref)
break;
if (map->head == NULL) {
map->head = map->tail = index;
} else if (point == NULL) {
index->prev = map->tail;
map->tail->next = index;
map->tail = index;
} else if (point == map->head) {
index->next = map->head;
map->head->prev = index;
map->head = index;
} else {
index->next = point;
index->prev = point->prev;
if (point->prev)
point->prev->next = index;
point->prev = index;
}
route_map_pfx_tbl_update(RMAP_EVENT_INDEX_ADDED, index, 0, NULL);
/* Execute event hook. */
if (route_map_master.event_hook) {
(*route_map_master.event_hook)(map->name);
route_map_notify_dependencies(map->name, RMAP_EVENT_CALL_ADDED);
}
if (rmap_debug)
zlog_debug("Route-map %s add sequence %d, type: %s",
map->name, pref, route_map_type_str(type));
return index;
2002-12-13 21:15:29 +01:00
}
/* Get route map index. */
struct route_map_index *
route_map_index_get(struct route_map *map, enum route_map_type type, int pref)
2002-12-13 21:15:29 +01:00
{
struct route_map_index *index;
2002-12-13 21:15:29 +01:00
index = route_map_index_lookup(map, RMAP_ANY, pref);
if (index && index->type != type) {
/* Delete index from route map. */
route_map_index_delete(index, 1);
index = NULL;
}
if (index == NULL)
index = route_map_index_add(map, type, pref);
return index;
2002-12-13 21:15:29 +01:00
}
/* New route map rule */
static struct route_map_rule *route_map_rule_new(void)
2002-12-13 21:15:29 +01:00
{
struct route_map_rule *new;
2002-12-13 21:15:29 +01:00
new = XCALLOC(MTYPE_ROUTE_MAP_RULE, sizeof(struct route_map_rule));
return new;
2002-12-13 21:15:29 +01:00
}
2002-12-13 21:15:29 +01:00
/* Install rule command to the match list. */
void route_map_install_match(const struct route_map_rule_cmd *cmd)
2002-12-13 21:15:29 +01:00
{
vector_set(route_match_vec, (void *)cmd);
2002-12-13 21:15:29 +01:00
}
/* Install rule command to the set list. */
void route_map_install_set(const struct route_map_rule_cmd *cmd)
2002-12-13 21:15:29 +01:00
{
vector_set(route_set_vec, (void *)cmd);
2002-12-13 21:15:29 +01:00
}
/* Lookup rule command from match list. */
static const struct route_map_rule_cmd *route_map_lookup_match(const char *name)
2002-12-13 21:15:29 +01:00
{
unsigned int i;
const struct route_map_rule_cmd *rule;
2002-12-13 21:15:29 +01:00
for (i = 0; i < vector_active(route_match_vec); i++)
if ((rule = vector_slot(route_match_vec, i)) != NULL)
if (strcmp(rule->str, name) == 0)
return rule;
return NULL;
2002-12-13 21:15:29 +01:00
}
/* Lookup rule command from set list. */
static const struct route_map_rule_cmd *route_map_lookup_set(const char *name)
2002-12-13 21:15:29 +01:00
{
unsigned int i;
const struct route_map_rule_cmd *rule;
2002-12-13 21:15:29 +01:00
for (i = 0; i < vector_active(route_set_vec); i++)
if ((rule = vector_slot(route_set_vec, i)) != NULL)
if (strcmp(rule->str, name) == 0)
return rule;
return NULL;
2002-12-13 21:15:29 +01:00
}
/* Add match and set rule to rule list. */
static void route_map_rule_add(struct route_map_rule_list *list,
struct route_map_rule *rule)
2002-12-13 21:15:29 +01:00
{
rule->next = NULL;
rule->prev = list->tail;
if (list->tail)
list->tail->next = rule;
else
list->head = rule;
list->tail = rule;
2002-12-13 21:15:29 +01:00
}
/* Delete rule from rule list. */
static void route_map_rule_delete(struct route_map_rule_list *list,
struct route_map_rule *rule)
2002-12-13 21:15:29 +01:00
{
if (rule->cmd->func_free)
(*rule->cmd->func_free)(rule->value);
2002-12-13 21:15:29 +01:00
XFREE(MTYPE_ROUTE_MAP_RULE_STR, rule->rule_str);
2002-12-13 21:15:29 +01:00
if (rule->next)
rule->next->prev = rule->prev;
else
list->tail = rule->prev;
if (rule->prev)
rule->prev->next = rule->next;
else
list->head = rule->next;
2002-12-13 21:15:29 +01:00
XFREE(MTYPE_ROUTE_MAP_RULE, rule);
2002-12-13 21:15:29 +01:00
}
/* strcmp wrapper function which don't crush even argument is NULL. */
static int rulecmp(const char *dst, const char *src)
{
if (dst == NULL) {
if (src == NULL)
return 0;
else
return 1;
} else {
if (src == NULL)
return 1;
else
return strcmp(dst, src);
}
2002-12-13 21:15:29 +01:00
return 1;
}
/* Use this to return the already specified argument for this match. This is
* useful to get the specified argument with a route map match rule when the
* rule is being deleted and the argument is not provided.
*/
const char *route_map_get_match_arg(struct route_map_index *index,
const char *match_name)
{
struct route_map_rule *rule;
const struct route_map_rule_cmd *cmd;
/* First lookup rule for add match statement. */
cmd = route_map_lookup_match(match_name);
if (cmd == NULL)
return NULL;
for (rule = index->match_list.head; rule; rule = rule->next)
if (rule->cmd == cmd && rule->rule_str != NULL)
return (rule->rule_str);
return NULL;
}
static route_map_event_t get_route_map_delete_event(route_map_event_t type)
{
switch (type) {
case RMAP_EVENT_CALL_ADDED:
return RMAP_EVENT_CALL_DELETED;
case RMAP_EVENT_PLIST_ADDED:
return RMAP_EVENT_PLIST_DELETED;
case RMAP_EVENT_CLIST_ADDED:
return RMAP_EVENT_CLIST_DELETED;
case RMAP_EVENT_ECLIST_ADDED:
return RMAP_EVENT_ECLIST_DELETED;
case RMAP_EVENT_LLIST_ADDED:
return RMAP_EVENT_LLIST_DELETED;
case RMAP_EVENT_ASLIST_ADDED:
return RMAP_EVENT_ASLIST_DELETED;
case RMAP_EVENT_FILTER_ADDED:
return RMAP_EVENT_FILTER_DELETED;
case RMAP_EVENT_SET_ADDED:
case RMAP_EVENT_SET_DELETED:
case RMAP_EVENT_SET_REPLACED:
case RMAP_EVENT_MATCH_ADDED:
case RMAP_EVENT_MATCH_DELETED:
case RMAP_EVENT_MATCH_REPLACED:
case RMAP_EVENT_INDEX_ADDED:
case RMAP_EVENT_INDEX_DELETED:
case RMAP_EVENT_CALL_DELETED:
case RMAP_EVENT_PLIST_DELETED:
case RMAP_EVENT_CLIST_DELETED:
case RMAP_EVENT_ECLIST_DELETED:
case RMAP_EVENT_LLIST_DELETED:
case RMAP_EVENT_ASLIST_DELETED:
case RMAP_EVENT_FILTER_DELETED:
/* This function returns the appropriate 'deleted' event type
* for every 'added' event type passed to this function.
* This is done only for named entities used in the
* route-map match commands.
* This function is not to be invoked for any of the other event
* types.
*/
assert(0);
}
assert(0);
/*
* Return to make c happy but if we get here something has gone
* terribly terribly wrong, so yes this return makes no sense.
*/
return RMAP_EVENT_CALL_ADDED;
}
2002-12-13 21:15:29 +01:00
/* Add match statement to route map. */
enum rmap_compile_rets route_map_add_match(struct route_map_index *index,
const char *match_name,
const char *match_arg,
route_map_event_t type)
{
struct route_map_rule *rule;
struct route_map_rule *next;
const struct route_map_rule_cmd *cmd;
void *compile;
int8_t delete_rmap_event_type = 0;
const char *rule_key;
/* First lookup rule for add match statement. */
cmd = route_map_lookup_match(match_name);
if (cmd == NULL)
return RMAP_RULE_MISSING;
/* Next call compile function for this match statement. */
if (cmd->func_compile) {
compile = (*cmd->func_compile)(match_arg);
if (compile == NULL)
return RMAP_COMPILE_ERROR;
} else
compile = NULL;
/* use the compiled results if applicable */
if (compile && cmd->func_get_rmap_rule_key)
rule_key = (*cmd->func_get_rmap_rule_key)
(compile);
else
rule_key = match_arg;
/* If argument is completely same ignore it. */
for (rule = index->match_list.head; rule; rule = next) {
next = rule->next;
if (rule->cmd == cmd) {
/* If the configured route-map match rule is exactly
* the same as the existing configuration then,
* ignore the duplicate configuration.
*/
if (strcmp(match_arg, rule->rule_str) == 0) {
if (cmd->func_free)
(*cmd->func_free)(compile);
return RMAP_COMPILE_SUCCESS;
}
/* If IPv4 or IPv6 prefix-list match criteria
* has been delete to the route-map index, update
* the route-map's prefix table.
*/
if (IS_RULE_IPv4_PREFIX_LIST(match_name))
route_map_pfx_tbl_update(
RMAP_EVENT_PLIST_DELETED, index, AFI_IP,
rule->rule_str);
else if (IS_RULE_IPv6_PREFIX_LIST(match_name))
route_map_pfx_tbl_update(
RMAP_EVENT_PLIST_DELETED, index,
AFI_IP6, rule->rule_str);
/* Remove the dependency of the route-map on the rule
* that is being replaced.
*/
if (type >= RMAP_EVENT_CALL_ADDED) {
delete_rmap_event_type =
get_route_map_delete_event(type);
route_map_upd8_dependency(
delete_rmap_event_type,
rule_key,
index->map->name);
}
route_map_rule_delete(&index->match_list, rule);
}
2002-12-13 21:15:29 +01:00
}
/* Add new route map match rule. */
rule = route_map_rule_new();
rule->cmd = cmd;
rule->value = compile;
if (match_arg)
rule->rule_str = XSTRDUP(MTYPE_ROUTE_MAP_RULE_STR, match_arg);
else
rule->rule_str = NULL;
2002-12-13 21:15:29 +01:00
/* Add new route match rule to linked list. */
route_map_rule_add(&index->match_list, rule);
2002-12-13 21:15:29 +01:00
/* If IPv4 or IPv6 prefix-list match criteria
* has been added to the route-map index, update
* the route-map's prefix table.
*/
if (IS_RULE_IPv4_PREFIX_LIST(match_name)) {
route_map_pfx_tbl_update(RMAP_EVENT_PLIST_ADDED, index, AFI_IP,
match_arg);
} else if (IS_RULE_IPv6_PREFIX_LIST(match_name)) {
route_map_pfx_tbl_update(RMAP_EVENT_PLIST_ADDED, index, AFI_IP6,
match_arg);
} else {
/* If IPv4 match criteria has been added to the route-map
* index, check for IPv6 prefix-list match rule presence and
* remove this index from the trie node created for each of the
* prefix-entry within the prefix-list. If no IPv6 prefix-list
* match rule is present, remove this index from the IPv6
* default route's trie node.
*/
if (IS_IPv4_RULE(match_name))
route_map_del_plist_entries(AFI_IP6, index, NULL, NULL);
/* If IPv6 match criteria has been added to the route-map
* index, check for IPv4 prefix-list match rule presence and
* remove this index from the trie node created for each of the
* prefix-entry within the prefix-list. If no IPv4 prefix-list
* match rule is present, remove this index from the IPv4
* default route's trie node.
*/
else if (IS_IPv6_RULE(match_name))
route_map_del_plist_entries(AFI_IP, index, NULL, NULL);
}
/* Execute event hook. */
if (route_map_master.event_hook) {
(*route_map_master.event_hook)(index->map->name);
route_map_notify_dependencies(index->map->name,
RMAP_EVENT_CALL_ADDED);
}
if (type != RMAP_EVENT_MATCH_ADDED)
route_map_upd8_dependency(type, rule_key, index->map->name);
2002-12-13 21:15:29 +01:00
return RMAP_COMPILE_SUCCESS;
2002-12-13 21:15:29 +01:00
}
/* Delete specified route match rule. */
enum rmap_compile_rets route_map_delete_match(struct route_map_index *index,
const char *match_name,
const char *match_arg,
route_map_event_t type)
{
struct route_map_rule *rule;
const struct route_map_rule_cmd *cmd;
const char *rule_key;
cmd = route_map_lookup_match(match_name);
if (cmd == NULL)
return RMAP_RULE_MISSING;
for (rule = index->match_list.head; rule; rule = rule->next)
if (rule->cmd == cmd && (rulecmp(rule->rule_str, match_arg) == 0
|| match_arg == NULL)) {
/* Execute event hook. */
if (route_map_master.event_hook) {
(*route_map_master.event_hook)(index->map->name);
route_map_notify_dependencies(
index->map->name,
RMAP_EVENT_CALL_ADDED);
}
if (cmd->func_get_rmap_rule_key)
rule_key = (*cmd->func_get_rmap_rule_key)
(rule->value);
else
rule_key = match_arg;
if (type != RMAP_EVENT_MATCH_DELETED && rule_key)
route_map_upd8_dependency(type, rule_key,
index->map->name);
route_map_rule_delete(&index->match_list, rule);
/* If IPv4 or IPv6 prefix-list match criteria
* has been delete to the route-map index, update
* the route-map's prefix table.
*/
if (IS_RULE_IPv4_PREFIX_LIST(match_name)) {
route_map_pfx_tbl_update(
RMAP_EVENT_PLIST_DELETED, index, AFI_IP,
match_arg);
} else if (IS_RULE_IPv6_PREFIX_LIST(match_name)) {
route_map_pfx_tbl_update(
RMAP_EVENT_PLIST_DELETED, index,
AFI_IP6, match_arg);
} else {
/* If no more IPv4 match rules are present in
* this index, check for IPv6 prefix-list match
* rule presence and add this index to trie node
* created for each of the prefix-entry within
* the prefix-list. If no IPv6 prefix-list match
* rule is present, add this index to the IPv6
* default route's trie node.
*/
if (!route_map_is_ip_rule_present(index))
route_map_add_plist_entries(
AFI_IP6, index, NULL, NULL);
/* If no more IPv6 match rules are present in
* this index, check for IPv4 prefix-list match
* rule presence and add this index to trie node
* created for each of the prefix-entry within
* the prefix-list. If no IPv6 prefix-list match
* rule is present, add this index to the IPv4
* default route's trie node.
*/
if (!route_map_is_ipv6_rule_present(index))
route_map_add_plist_entries(
AFI_IP, index, NULL, NULL);
}
return RMAP_COMPILE_SUCCESS;
}
/* Can't find matched rule. */
return RMAP_RULE_MISSING;
2002-12-13 21:15:29 +01:00
}
/* Add route-map set statement to the route map. */
enum rmap_compile_rets route_map_add_set(struct route_map_index *index,
const char *set_name,
const char *set_arg)
{
struct route_map_rule *rule;
struct route_map_rule *next;
const struct route_map_rule_cmd *cmd;
void *compile;
cmd = route_map_lookup_set(set_name);
if (cmd == NULL)
return RMAP_RULE_MISSING;
/* Next call compile function for this match statement. */
if (cmd->func_compile) {
compile = (*cmd->func_compile)(set_arg);
if (compile == NULL)
return RMAP_COMPILE_ERROR;
} else
compile = NULL;
/* Add by WJL. if old set command of same kind exist, delete it first
to ensure only one set command of same kind exist under a
route_map_index. */
for (rule = index->set_list.head; rule; rule = next) {
next = rule->next;
if (rule->cmd == cmd)
route_map_rule_delete(&index->set_list, rule);
}
/* Add new route map match rule. */
rule = route_map_rule_new();
rule->cmd = cmd;
rule->value = compile;
if (set_arg)
rule->rule_str = XSTRDUP(MTYPE_ROUTE_MAP_RULE_STR, set_arg);
else
rule->rule_str = NULL;
/* Add new route match rule to linked list. */
route_map_rule_add(&index->set_list, rule);
/* Execute event hook. */
if (route_map_master.event_hook) {
(*route_map_master.event_hook)(index->map->name);
route_map_notify_dependencies(index->map->name,
RMAP_EVENT_CALL_ADDED);
2002-12-13 21:15:29 +01:00
}
return RMAP_COMPILE_SUCCESS;
2002-12-13 21:15:29 +01:00
}
/* Delete route map set rule. */
enum rmap_compile_rets route_map_delete_set(struct route_map_index *index,
const char *set_name,
const char *set_arg)
{
struct route_map_rule *rule;
const struct route_map_rule_cmd *cmd;
cmd = route_map_lookup_set(set_name);
if (cmd == NULL)
return RMAP_RULE_MISSING;
for (rule = index->set_list.head; rule; rule = rule->next)
if ((rule->cmd == cmd) && (rulecmp(rule->rule_str, set_arg) == 0
|| set_arg == NULL)) {
route_map_rule_delete(&index->set_list, rule);
/* Execute event hook. */
if (route_map_master.event_hook) {
(*route_map_master.event_hook)(index->map->name);
route_map_notify_dependencies(
index->map->name,
RMAP_EVENT_CALL_ADDED);
}
return RMAP_COMPILE_SUCCESS;
}
/* Can't find matched rule. */
return RMAP_RULE_MISSING;
2002-12-13 21:15:29 +01:00
}
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
static enum route_map_cmd_result_t
route_map_apply_match(struct route_map_rule_list *match_list,
const struct prefix *prefix, route_map_object_t type,
void *object)
{
enum route_map_cmd_result_t ret = RMAP_NOMATCH;
struct route_map_rule *match;
bool is_matched = false;
/* Check all match rule and if there is no match rule, go to the
set statement. */
if (!match_list->head)
ret = RMAP_MATCH;
else {
for (match = match_list->head; match; match = match->next) {
/*
* Try each match statement. If any match does not
* return RMAP_MATCH or RMAP_NOOP, return.
* Otherwise continue on to next match statement.
* All match statements must MATCH for
* end-result to be a match.
* (Exception:If match stmts result in a mix of
* MATCH/NOOP, then also end-result is a match)
* If all result in NOOP, end-result is NOOP.
*/
ret = (*match->cmd->func_apply)(match->value, prefix,
type, object);
/*
* If the consolidated result of func_apply is:
* -----------------------------------------------
* | MATCH | NOMATCH | NOOP | Final Result |
* ------------------------------------------------
* | yes | yes | yes | NOMATCH |
* | no | no | yes | NOOP |
* | yes | no | yes | MATCH |
* | no | yes | yes | NOMATCH |
* |-----------------------------------------------
*
* Traditionally, all rules within route-map
* should match for it to MATCH.
* If there are noops within the route-map rules,
* it follows the above matrix.
*
* Eg: route-map rm1 permit 10
* match rule1
* match rule2
* match rule3
* ....
* route-map rm1 permit 20
* match ruleX
* match ruleY
* ...
*/
switch (ret) {
case RMAP_MATCH:
is_matched = true;
break;
case RMAP_NOMATCH:
return ret;
case RMAP_NOOP:
if (is_matched)
ret = RMAP_MATCH;
break;
default:
break;
}
}
}
return ret;
}
static struct list *route_map_get_index_list(struct route_node **rn,
const struct prefix *prefix,
struct route_table *table)
{
struct route_node *tmp_rn = NULL;
if (!(*rn)) {
*rn = route_node_match(table, prefix);
if (!(*rn))
return NULL;
if ((*rn)->info)
return (struct list *)((*rn)->info);
/* If rn->info is NULL, get the parent.
* Store the rn in tmp_rn and unlock it later.
*/
tmp_rn = *rn;
}
do {
*rn = (*rn)->parent;
if (tmp_rn)
route_unlock_node(tmp_rn);
if (!(*rn))
break;
if ((*rn)->info) {
route_lock_node(*rn);
return (struct list *)((*rn)->info);
}
} while (!(*rn)->info);
return NULL;
}
/*
* This function returns the route-map index that best matches the prefix.
*/
static struct route_map_index *
route_map_get_index(struct route_map *map, const struct prefix *prefix,
route_map_object_t type, void *object, uint8_t *match_ret)
{
int ret = 0;
struct list *candidate_rmap_list = NULL;
struct route_node *rn = NULL;
struct listnode *ln = NULL, *nn = NULL;
struct route_map_index *index = NULL, *best_index = NULL;
struct route_map_index *head_index = NULL;
struct route_table *table = NULL;
unsigned char family = prefix->family;
if (family == AF_INET)
table = map->ipv4_prefix_table;
else
table = map->ipv6_prefix_table;
if (!table)
return NULL;
do {
candidate_rmap_list =
route_map_get_index_list(&rn, prefix, table);
if (!rn)
break;
/* If the index at the head of the list is of seq higher
* than that in best_index, ignore the list and get the
* parent node's list.
*/
head_index = (struct route_map_index *)(listgetdata(
listhead(candidate_rmap_list)));
if (best_index && head_index
&& (best_index->pref < head_index->pref)) {
route_unlock_node(rn);
continue;
}
for (ALL_LIST_ELEMENTS(candidate_rmap_list, ln, nn, index)) {
/* If the index is of seq higher than that in
* best_index, ignore the list and get the parent
* node's list.
*/
if (best_index && (best_index->pref < index->pref))
break;
ret = route_map_apply_match(&index->match_list, prefix,
type, object);
if (ret == RMAP_MATCH) {
*match_ret = ret;
best_index = index;
break;
} else if (ret == RMAP_NOOP) {
/*
* If match_ret is denymatch, even if we see
* more noops, we retain this return value and
* return this eventually if there are no
* matches.
*/
if (*match_ret != RMAP_NOMATCH)
*match_ret = ret;
} else {
/*
* ret is RMAP_NOMATCH.
*/
*match_ret = ret;
}
}
route_unlock_node(rn);
} while (rn);
return best_index;
}
static int route_map_candidate_list_cmp(struct route_map_index *idx1,
struct route_map_index *idx2)
{
if (!idx1)
return -1;
if (!idx2)
return 1;
return (idx1->pref - idx2->pref);
}
/*
* This function adds the route-map index into the default route's
* route-node in the route-map's IPv4/IPv6 prefix-table.
*/
static void route_map_pfx_table_add_default(afi_t afi,
struct route_map_index *index)
{
struct route_node *rn = NULL;
struct list *rmap_candidate_list = NULL;
struct prefix p;
bool updated_rn = false;
struct route_table *table = NULL;
memset(&p, 0, sizeof(p));
p.family = afi2family(afi);
p.prefixlen = 0;
if (p.family == AF_INET) {
table = index->map->ipv4_prefix_table;
if (!table)
index->map->ipv4_prefix_table = route_table_init();
table = index->map->ipv4_prefix_table;
} else {
table = index->map->ipv6_prefix_table;
if (!table)
index->map->ipv6_prefix_table = route_table_init();
table = index->map->ipv6_prefix_table;
}
/* Add default route to table */
rn = route_node_get(table, &p);
if (!rn)
return;
if (!rn->info) {
rmap_candidate_list = list_new();
rmap_candidate_list->cmp =
(int (*)(void *, void *))route_map_candidate_list_cmp;
rn->info = rmap_candidate_list;
} else {
rmap_candidate_list = (struct list *)rn->info;
updated_rn = true;
}
listnode_add_sort_nodup(rmap_candidate_list, index);
if (updated_rn)
route_unlock_node(rn);
}
/*
* This function removes the route-map index from the default route's
* route-node in the route-map's IPv4/IPv6 prefix-table.
*/
static void route_map_pfx_table_del_default(afi_t afi,
struct route_map_index *index)
{
struct route_node *rn = NULL;
struct list *rmap_candidate_list = NULL;
struct prefix p;
struct route_table *table = NULL;
memset(&p, 0, sizeof(p));
p.family = afi2family(afi);
p.prefixlen = 0;
if (p.family == AF_INET)
table = index->map->ipv4_prefix_table;
else
table = index->map->ipv6_prefix_table;
/* Remove RMAP index from default route in table */
rn = route_node_lookup(table, &p);
if (!rn || !rn->info)
return;
rmap_candidate_list = (struct list *)rn->info;
listnode_delete(rmap_candidate_list, index);
if (listcount(rmap_candidate_list) == 0) {
list_delete(&rmap_candidate_list);
rn->info = NULL;
route_unlock_node(rn);
}
route_unlock_node(rn);
}
/*
* This function adds the route-map index to the route-node for
* the prefix-entry in the route-map's IPv4/IPv6 prefix-table.
*/
static void route_map_pfx_table_add(struct route_table *table,
struct route_map_index *index,
struct prefix_list_entry *pentry)
{
struct route_node *rn = NULL;
struct list *rmap_candidate_list = NULL;
bool updated_rn = false;
rn = route_node_get(table, &pentry->prefix);
if (!rn)
return;
if (!rn->info) {
rmap_candidate_list = list_new();
rmap_candidate_list->cmp =
(int (*)(void *, void *))route_map_candidate_list_cmp;
rn->info = rmap_candidate_list;
} else {
rmap_candidate_list = (struct list *)rn->info;
updated_rn = true;
}
listnode_add_sort_nodup(rmap_candidate_list, index);
if (updated_rn)
route_unlock_node(rn);
}
/*
* This function removes the route-map index from the route-node for
* the prefix-entry in the route-map's IPv4/IPv6 prefix-table.
*/
static void route_map_pfx_table_del(struct route_table *table,
struct route_map_index *index,
struct prefix_list_entry *pentry)
{
struct route_node *rn = NULL;
struct list *rmap_candidate_list = NULL;
rn = route_node_lookup(table, &pentry->prefix);
if (!rn || !rn->info)
return;
rmap_candidate_list = (struct list *)rn->info;
listnode_delete(rmap_candidate_list, index);
if (listcount(rmap_candidate_list) == 0) {
list_delete(&rmap_candidate_list);
rn->info = NULL;
route_unlock_node(rn);
}
route_unlock_node(rn);
}
/* This function checks for the presence of an IPv4 match rule
* in the given route-map index.
*/
static bool route_map_is_ip_rule_present(struct route_map_index *index)
{
struct route_map_rule_list *match_list = NULL;
struct route_map_rule *rule = NULL;
match_list = &index->match_list;
for (rule = match_list->head; rule; rule = rule->next)
if (IS_IPv4_RULE(rule->cmd->str))
return true;
return false;
}
/* This function checks for the presence of an IPv6 match rule
* in the given route-map index.
*/
static bool route_map_is_ipv6_rule_present(struct route_map_index *index)
{
struct route_map_rule_list *match_list = NULL;
struct route_map_rule *rule = NULL;
match_list = &index->match_list;
for (rule = match_list->head; rule; rule = rule->next)
if (IS_IPv6_RULE(rule->cmd->str))
return true;
return false;
}
/* This function does the following:
* 1) If plist_name is not present, search for a IPv4 or IPv6 prefix-list
* match clause (based on the afi passed to this foo) and get the
* prefix-list name.
* 2) Look up the prefix-list using the name.
* 3) If the prefix-list is not found then, add the index to the IPv4/IPv6
* default-route's node in the trie (based on the afi passed to this foo).
* 4) If the prefix-list is found then, remove the index from the IPv4/IPv6
* default-route's node in the trie (based on the afi passed to this foo).
* 5) If a prefix-entry is passed then, create a route-node for this entry and
* add this index to the route-node.
* 6) If prefix-entry is not passed then, for every prefix-entry in the
* prefix-list, create a route-node for this entry and
* add this index to the route-node.
*/
static void route_map_add_plist_entries(afi_t afi,
struct route_map_index *index,
const char *plist_name,
struct prefix_list_entry *entry)
{
struct route_map_rule_list *match_list = NULL;
struct route_map_rule *match = NULL;
struct prefix_list *plist = NULL;
struct prefix_list_entry *pentry = NULL;
bool plist_rule_is_present = false;
if (!plist_name) {
match_list = &index->match_list;
for (match = match_list->head; match; match = match->next) {
if (afi == AFI_IP) {
if (IS_RULE_IPv4_PREFIX_LIST(match->cmd->str)) {
plist_rule_is_present = true;
break;
}
} else {
if (IS_RULE_IPv6_PREFIX_LIST(match->cmd->str)) {
plist_rule_is_present = true;
break;
}
}
}
if (plist_rule_is_present)
plist = prefix_list_lookup(afi, match->rule_str);
} else {
plist = prefix_list_lookup(afi, plist_name);
}
if (!plist) {
route_map_pfx_table_add_default(afi, index);
return;
}
route_map_pfx_table_del_default(afi, index);
if (entry) {
if (afi == AFI_IP) {
route_map_pfx_table_add(index->map->ipv4_prefix_table,
index, entry);
} else {
route_map_pfx_table_add(index->map->ipv6_prefix_table,
index, entry);
}
} else {
for (pentry = plist->head; pentry; pentry = pentry->next) {
if (afi == AFI_IP) {
route_map_pfx_table_add(
index->map->ipv4_prefix_table, index,
pentry);
} else {
route_map_pfx_table_add(
index->map->ipv6_prefix_table, index,
pentry);
}
}
}
}
/* This function does the following:
* 1) If plist_name is not present, search for a IPv4 or IPv6 prefix-list
* match clause (based on the afi passed to this foo) and get the
* prefix-list name.
* 2) Look up the prefix-list using the name.
* 3) If the prefix-list is not found then, delete the index from the IPv4/IPv6
* default-route's node in the trie (based on the afi passed to this foo).
* 4) If a prefix-entry is passed then, remove this index from the route-node
* for the prefix in this prefix-entry.
* 5) If prefix-entry is not passed then, for every prefix-entry in the
* prefix-list, remove this index from the route-node
* for the prefix in this prefix-entry.
*/
static void route_map_del_plist_entries(afi_t afi,
struct route_map_index *index,
const char *plist_name,
struct prefix_list_entry *entry)
{
struct route_map_rule_list *match_list = NULL;
struct route_map_rule *match = NULL;
struct prefix_list *plist = NULL;
struct prefix_list_entry *pentry = NULL;
bool plist_rule_is_present = false;
if (!plist_name) {
match_list = &index->match_list;
for (match = match_list->head; match; match = match->next) {
if (afi == AFI_IP) {
if (IS_RULE_IPv4_PREFIX_LIST(match->cmd->str)) {
plist_rule_is_present = true;
break;
}
} else {
if (IS_RULE_IPv6_PREFIX_LIST(match->cmd->str)) {
plist_rule_is_present = true;
break;
}
}
}
if (plist_rule_is_present)
plist = prefix_list_lookup(afi, match->rule_str);
} else {
plist = prefix_list_lookup(afi, plist_name);
}
if (!plist) {
route_map_pfx_table_del_default(afi, index);
return;
}
if (entry) {
if (afi == AFI_IP) {
route_map_pfx_table_del(index->map->ipv4_prefix_table,
index, entry);
} else {
route_map_pfx_table_del(index->map->ipv6_prefix_table,
index, entry);
}
} else {
for (pentry = plist->head; pentry; pentry = pentry->next) {
if (afi == AFI_IP) {
route_map_pfx_table_del(
index->map->ipv4_prefix_table, index,
pentry);
} else {
route_map_pfx_table_del(
index->map->ipv6_prefix_table, index,
pentry);
}
}
}
}
/*
* This function handles the cases where a prefix-list is added/removed
* as a match command from a particular route-map index.
* It updates the prefix-table of the route-map accordingly.
*/
static void route_map_trie_update(afi_t afi, route_map_event_t event,
struct route_map_index *index,
const char *plist_name)
{
if (event == RMAP_EVENT_PLIST_ADDED) {
if (afi == AFI_IP) {
if (!route_map_is_ipv6_rule_present(index)) {
route_map_pfx_table_del_default(AFI_IP6, index);
route_map_add_plist_entries(afi, index,
plist_name, NULL);
} else {
route_map_del_plist_entries(AFI_IP6, index,
NULL, NULL);
}
} else {
if (!route_map_is_ip_rule_present(index)) {
route_map_pfx_table_del_default(AFI_IP, index);
route_map_add_plist_entries(afi, index,
plist_name, NULL);
} else {
route_map_del_plist_entries(AFI_IP, index, NULL,
NULL);
}
}
} else if (event == RMAP_EVENT_PLIST_DELETED) {
if (afi == AFI_IP) {
route_map_del_plist_entries(afi, index, plist_name,
NULL);
if (!route_map_is_ipv6_rule_present(index))
route_map_pfx_table_add_default(afi, index);
if (!route_map_is_ip_rule_present(index))
route_map_add_plist_entries(AFI_IP6, index,
NULL, NULL);
} else {
route_map_del_plist_entries(afi, index, plist_name,
NULL);
if (!route_map_is_ip_rule_present(index))
route_map_pfx_table_add_default(afi, index);
if (!route_map_is_ipv6_rule_present(index))
route_map_add_plist_entries(AFI_IP, index, NULL,
NULL);
}
}
}
/*
* This function handles the cases where a route-map index and
* prefix-list is added/removed.
* It updates the prefix-table of the route-map accordingly.
*/
static void route_map_pfx_tbl_update(route_map_event_t event,
struct route_map_index *index, afi_t afi,
const char *plist_name)
{
struct route_map *rmap = NULL;
if (!index)
return;
if (event == RMAP_EVENT_INDEX_ADDED) {
route_map_pfx_table_add_default(AFI_IP, index);
route_map_pfx_table_add_default(AFI_IP6, index);
return;
}
if (event == RMAP_EVENT_INDEX_DELETED) {
route_map_pfx_table_del_default(AFI_IP, index);
route_map_pfx_table_del_default(AFI_IP6, index);
if ((index->map->head == NULL) && (index->map->tail == NULL)) {
rmap = index->map;
if (rmap->ipv4_prefix_table) {
route_table_finish(rmap->ipv4_prefix_table);
rmap->ipv4_prefix_table = NULL;
}
if (rmap->ipv6_prefix_table) {
route_table_finish(rmap->ipv6_prefix_table);
rmap->ipv6_prefix_table = NULL;
}
}
return;
}
/* Handle prefix-list match rule addition/deletion.
*/
route_map_trie_update(afi, event, index, plist_name);
}
/*
* This function handles the cases where a new prefix-entry is added to
* a prefix-list or, an existing prefix-entry is removed from the prefix-list.
* It updates the prefix-table of the route-map accordingly.
*/
static void route_map_pentry_update(route_map_event_t event,
const char *plist_name,
struct route_map_index *index,
struct prefix_list_entry *pentry)
{
struct prefix_list *plist = NULL;
afi_t afi;
unsigned char family = pentry->prefix.family;
if (family == AF_INET) {
afi = AFI_IP;
plist = prefix_list_lookup(AFI_IP, plist_name);
} else {
afi = AFI_IP6;
plist = prefix_list_lookup(AFI_IP6, plist_name);
}
if (event == RMAP_EVENT_PLIST_ADDED) {
if (plist->count == 1) {
if (afi == AFI_IP) {
if (!route_map_is_ipv6_rule_present(index))
route_map_add_plist_entries(
afi, index, plist_name, pentry);
} else {
if (!route_map_is_ip_rule_present(index))
route_map_add_plist_entries(
afi, index, plist_name, pentry);
}
} else {
route_map_add_plist_entries(afi, index, plist_name,
pentry);
}
} else if (event == RMAP_EVENT_PLIST_DELETED) {
route_map_del_plist_entries(afi, index, plist_name, pentry);
if (plist->count == 1) {
if (afi == AFI_IP) {
if (!route_map_is_ipv6_rule_present(index))
route_map_pfx_table_add_default(afi,
index);
} else {
if (!route_map_is_ip_rule_present(index))
route_map_pfx_table_add_default(afi,
index);
}
}
}
}
static void route_map_pentry_process_dependency(struct hash_bucket *backet,
void *data)
{
char *rmap_name = NULL;
struct route_map *rmap = NULL;
struct route_map_index *index = NULL;
struct route_map_rule_list *match_list = NULL;
struct route_map_rule *match = NULL;
struct route_map_dep_data *dep_data = NULL;
struct route_map_pentry_dep *pentry_dep =
(struct route_map_pentry_dep *)data;
unsigned char family = pentry_dep->pentry->prefix.family;
dep_data = (struct route_map_dep_data *)backet->data;
if (!dep_data)
return;
rmap_name = dep_data->rname;
rmap = route_map_lookup_by_name(rmap_name);
if (!rmap || !rmap->head)
return;
for (index = rmap->head; index; index = index->next) {
match_list = &index->match_list;
if (!match_list)
continue;
for (match = match_list->head; match; match = match->next) {
if (strcmp(match->rule_str, pentry_dep->plist_name)
== 0) {
if (IS_RULE_IPv4_PREFIX_LIST(match->cmd->str)
&& family == AF_INET) {
route_map_pentry_update(
pentry_dep->event,
pentry_dep->plist_name, index,
pentry_dep->pentry);
} else if (IS_RULE_IPv6_PREFIX_LIST(
match->cmd->str)
&& family == AF_INET6) {
route_map_pentry_update(
pentry_dep->event,
pentry_dep->plist_name, index,
pentry_dep->pentry);
}
}
}
}
}
void route_map_notify_pentry_dependencies(const char *affected_name,
struct prefix_list_entry *pentry,
route_map_event_t event)
{
struct route_map_dep *dep = NULL;
struct hash *upd8_hash = NULL;
struct route_map_pentry_dep pentry_dep;
if (!affected_name || !pentry)
return;
upd8_hash = route_map_get_dep_hash(event);
if (!upd8_hash)
return;
dep = (struct route_map_dep *)hash_get(upd8_hash, (void *)affected_name,
NULL);
if (dep) {
if (!dep->this_hash)
dep->this_hash = upd8_hash;
memset(&pentry_dep, 0, sizeof(struct route_map_pentry_dep));
pentry_dep.pentry = pentry;
pentry_dep.plist_name = affected_name;
pentry_dep.event = event;
hash_iterate(dep->dep_rmap_hash,
route_map_pentry_process_dependency,
(void *)&pentry_dep);
}
}
/* Apply route map's each index to the object.
The matrix for a route-map looks like this:
(note, this includes the description for the "NEXT"
and "GOTO" frobs now
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
| Match | No Match | No op
|-----------|--------------|-------
permit | action | cont | cont.
| | default:deny | default:permit
-------------------+-----------------------
| deny | cont | cont.
deny | | default:deny | default:permit
|-----------|--------------|--------
2004-09-13 Jose Luis Rubio <jrubio@dit.upm.es> (at Technical University of Madrid as part of Euro6ix Project) Enhanced Route Server functionality and Route-Maps: * bgpd/bgpd.h: Modified 'struct peer' and 'struct bgp_filter' to support rs-clients. A 'struct bgp_table *rib' has been added to the first (to mantain a separated RIB for each rs-client) and two new route-maps have been added to the last (for import/export policies). Added the following #defines: RMAP_{IN|OUT|IMPORT|EXPORT|MAX}, PEER_RMAP_TYPE_{IMPORT|EXPORT} and BGP_CLEAR_SOFT_RSCLIENT. * bgpd/bgpd.c: Modified the functions that create/delete/etc peers in order to consider the new fields included in 'struct peer' for supporting rs-clients, i.e. the import/export route-maps and the 'struct bgp_table'. * bgpd/bgp_route.{ch}: Modified several functions related with receiving/sending announces in order to support the new Route Server capabilities. Function 'bgp_process' has been reorganized, creating an auxiliar function for best path selection ('bgp_best_selection'). Modified 'bgp_show' and 'bgp_show_route' for displaying information about any RIB (and not only the main bgp RIB). Added commands for displaying information about RS-clients RIBs: 'show bgp rsclient (A.B.C.D|X:X::X:X)', 'show bgp rsclient (A.B.C.D|X:X::X:X) X:X::X:X/M', etc * bgpd/bgp_table.{ch}: The structure 'struct bgp_table' now has two new fields: type (which can take the values BGP_TABLE_{MAIN|RSCLIENT}) and 'void *owner' which points to 'struct bgp' or 'struct peer' which owns the table. When creating a new bgp_table by default 'type=BGP_TABLE_MAIN' is set. * bgpd/bgp_vty.c: The commands 'neighbor ... route-server-client' and 'no neighbor ... route-server-client' now not only set/unset the flag PEER_FLAG_RSERVER_CLIENT, but they create/destroy the 'struct bgp_table' of the peer. Special actions are taken for peer_groups. Command 'neighbor ... route-map WORD (in|out)' now also supports two new kinds of route-map: 'import' and 'export'. Added commands 'clear bgp * rsclient', etc. These commands allow a new kind of soft_reconfig which affects only the RIB of the specified RS-client. Added commands 'show bgp rsclient summary', etc which display a summary of the rs-clients configured for the corresponding address family. * bgpd/bgp_routemap.c: A new match statement is available, 'match peer (A.B.C.D|X:X::X:X)'. This statement can only be used in import/export route-maps, and it matches when the peer who announces (when used in an import route-map) or is going to receive (when used in an export route-map) the route is the same than the one specified in the statement. For peer-groups the statement matches if the specified peer is member of the peer-group. A special version of the command, 'match peer local', matches with routes originated by the Route Server (defined with 'network ...', redistributed routes and default-originate). * lib/routemap.{ch}: Added a new clause 'call NAME' for use in route-maps. It jumps into the specified route-map and when it returns the first route-map ends if the called RM returns DENY_MATCH, or continues in other case.
2004-09-13 07:12:46 +02:00
action)
-Apply Set statements, accept route
-If Call statement is present jump to the specified route-map, if it
denies the route we finish.
2004-09-13 Jose Luis Rubio <jrubio@dit.upm.es> (at Technical University of Madrid as part of Euro6ix Project) Enhanced Route Server functionality and Route-Maps: * bgpd/bgpd.h: Modified 'struct peer' and 'struct bgp_filter' to support rs-clients. A 'struct bgp_table *rib' has been added to the first (to mantain a separated RIB for each rs-client) and two new route-maps have been added to the last (for import/export policies). Added the following #defines: RMAP_{IN|OUT|IMPORT|EXPORT|MAX}, PEER_RMAP_TYPE_{IMPORT|EXPORT} and BGP_CLEAR_SOFT_RSCLIENT. * bgpd/bgpd.c: Modified the functions that create/delete/etc peers in order to consider the new fields included in 'struct peer' for supporting rs-clients, i.e. the import/export route-maps and the 'struct bgp_table'. * bgpd/bgp_route.{ch}: Modified several functions related with receiving/sending announces in order to support the new Route Server capabilities. Function 'bgp_process' has been reorganized, creating an auxiliar function for best path selection ('bgp_best_selection'). Modified 'bgp_show' and 'bgp_show_route' for displaying information about any RIB (and not only the main bgp RIB). Added commands for displaying information about RS-clients RIBs: 'show bgp rsclient (A.B.C.D|X:X::X:X)', 'show bgp rsclient (A.B.C.D|X:X::X:X) X:X::X:X/M', etc * bgpd/bgp_table.{ch}: The structure 'struct bgp_table' now has two new fields: type (which can take the values BGP_TABLE_{MAIN|RSCLIENT}) and 'void *owner' which points to 'struct bgp' or 'struct peer' which owns the table. When creating a new bgp_table by default 'type=BGP_TABLE_MAIN' is set. * bgpd/bgp_vty.c: The commands 'neighbor ... route-server-client' and 'no neighbor ... route-server-client' now not only set/unset the flag PEER_FLAG_RSERVER_CLIENT, but they create/destroy the 'struct bgp_table' of the peer. Special actions are taken for peer_groups. Command 'neighbor ... route-map WORD (in|out)' now also supports two new kinds of route-map: 'import' and 'export'. Added commands 'clear bgp * rsclient', etc. These commands allow a new kind of soft_reconfig which affects only the RIB of the specified RS-client. Added commands 'show bgp rsclient summary', etc which display a summary of the rs-clients configured for the corresponding address family. * bgpd/bgp_routemap.c: A new match statement is available, 'match peer (A.B.C.D|X:X::X:X)'. This statement can only be used in import/export route-maps, and it matches when the peer who announces (when used in an import route-map) or is going to receive (when used in an export route-map) the route is the same than the one specified in the statement. For peer-groups the statement matches if the specified peer is member of the peer-group. A special version of the command, 'match peer local', matches with routes originated by the Route Server (defined with 'network ...', redistributed routes and default-originate). * lib/routemap.{ch}: Added a new clause 'call NAME' for use in route-maps. It jumps into the specified route-map and when it returns the first route-map ends if the called RM returns DENY_MATCH, or continues in other case.
2004-09-13 07:12:46 +02:00
-If NEXT is specified, goto NEXT statement
-If GOTO is specified, goto the first clause where pref > nextpref
-If nothing is specified, do as Cisco and finish
deny)
-Route is denied by route-map.
cont)
-Goto Next index
If we get no matches after we've processed all updates, then the route
is dropped too.
2004-09-13 Jose Luis Rubio <jrubio@dit.upm.es> (at Technical University of Madrid as part of Euro6ix Project) Enhanced Route Server functionality and Route-Maps: * bgpd/bgpd.h: Modified 'struct peer' and 'struct bgp_filter' to support rs-clients. A 'struct bgp_table *rib' has been added to the first (to mantain a separated RIB for each rs-client) and two new route-maps have been added to the last (for import/export policies). Added the following #defines: RMAP_{IN|OUT|IMPORT|EXPORT|MAX}, PEER_RMAP_TYPE_{IMPORT|EXPORT} and BGP_CLEAR_SOFT_RSCLIENT. * bgpd/bgpd.c: Modified the functions that create/delete/etc peers in order to consider the new fields included in 'struct peer' for supporting rs-clients, i.e. the import/export route-maps and the 'struct bgp_table'. * bgpd/bgp_route.{ch}: Modified several functions related with receiving/sending announces in order to support the new Route Server capabilities. Function 'bgp_process' has been reorganized, creating an auxiliar function for best path selection ('bgp_best_selection'). Modified 'bgp_show' and 'bgp_show_route' for displaying information about any RIB (and not only the main bgp RIB). Added commands for displaying information about RS-clients RIBs: 'show bgp rsclient (A.B.C.D|X:X::X:X)', 'show bgp rsclient (A.B.C.D|X:X::X:X) X:X::X:X/M', etc * bgpd/bgp_table.{ch}: The structure 'struct bgp_table' now has two new fields: type (which can take the values BGP_TABLE_{MAIN|RSCLIENT}) and 'void *owner' which points to 'struct bgp' or 'struct peer' which owns the table. When creating a new bgp_table by default 'type=BGP_TABLE_MAIN' is set. * bgpd/bgp_vty.c: The commands 'neighbor ... route-server-client' and 'no neighbor ... route-server-client' now not only set/unset the flag PEER_FLAG_RSERVER_CLIENT, but they create/destroy the 'struct bgp_table' of the peer. Special actions are taken for peer_groups. Command 'neighbor ... route-map WORD (in|out)' now also supports two new kinds of route-map: 'import' and 'export'. Added commands 'clear bgp * rsclient', etc. These commands allow a new kind of soft_reconfig which affects only the RIB of the specified RS-client. Added commands 'show bgp rsclient summary', etc which display a summary of the rs-clients configured for the corresponding address family. * bgpd/bgp_routemap.c: A new match statement is available, 'match peer (A.B.C.D|X:X::X:X)'. This statement can only be used in import/export route-maps, and it matches when the peer who announces (when used in an import route-map) or is going to receive (when used in an export route-map) the route is the same than the one specified in the statement. For peer-groups the statement matches if the specified peer is member of the peer-group. A special version of the command, 'match peer local', matches with routes originated by the Route Server (defined with 'network ...', redistributed routes and default-originate). * lib/routemap.{ch}: Added a new clause 'call NAME' for use in route-maps. It jumps into the specified route-map and when it returns the first route-map ends if the called RM returns DENY_MATCH, or continues in other case.
2004-09-13 07:12:46 +02:00
Some notes on the new "CALL", "NEXT" and "GOTO"
call WORD - If this clause is matched, then the set statements
are executed and then we jump to route-map 'WORD'. If
this route-map denies the route, we finish, in other
case we
do whatever the exit policy (EXIT, NEXT or GOTO) tells.
on-match next - If this clause is matched, then the set statements
are executed and then we drop through to the next clause
on-match goto n - If this clause is matched, then the set statments
are executed and then we goto the nth clause, or the
first clause greater than this. In order to ensure
route-maps *always* exit, you cannot jump backwards.
Sorry ;)
We need to make sure our route-map processing matches the above
2002-12-13 21:15:29 +01:00
*/
route_map_result_t route_map_apply(struct route_map *map,
const struct prefix *prefix,
route_map_object_t type, void *object)
{
static int recursion = 0;
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
enum route_map_cmd_result_t match_ret = RMAP_NOMATCH;
route_map_result_t ret = RMAP_PERMITMATCH;
struct route_map_index *index = NULL;
struct route_map_rule *set = NULL;
char buf[PREFIX_STRLEN];
bool skip_match_clause = false;
if (recursion > RMAP_RECURSION_LIMIT) {
flog_warn(
EC_LIB_RMAP_RECURSION_LIMIT,
"route-map recursion limit (%d) reached, discarding route",
RMAP_RECURSION_LIMIT);
recursion = 0;
return RMAP_DENYMATCH;
}
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
if (map == NULL || map->head == NULL) {
ret = RMAP_DENYMATCH;
goto route_map_apply_end;
}
map->applied++;
if ((!map->optimization_disabled)
&& (map->ipv4_prefix_table || map->ipv6_prefix_table)) {
index = route_map_get_index(map, prefix, type, object,
(uint8_t *)&match_ret);
if (index) {
if (rmap_debug)
zlog_debug(
"Best match route-map: %s, sequence: %d for pfx: %s, result: %s",
map->name, index->pref,
prefix2str(prefix, buf, sizeof(buf)),
route_map_cmd_result_str(match_ret));
} else {
if (rmap_debug)
zlog_debug(
"No best match sequence for pfx: %s in route-map: %s, result: %s",
prefix2str(prefix, buf, sizeof(buf)),
map->name,
route_map_cmd_result_str(match_ret));
/*
* No index matches this prefix. Return deny unless,
* match_ret = RMAP_NOOP.
*/
if (match_ret == RMAP_NOOP)
ret = RMAP_PERMITMATCH;
else
ret = RMAP_DENYMATCH;
goto route_map_apply_end;
}
skip_match_clause = true;
} else {
index = map->head;
}
for (; index; index = index->next) {
if (skip_match_clause == false) {
/* Apply this index. */
match_ret = route_map_apply_match(&index->match_list,
prefix, type, object);
if (rmap_debug) {
zlog_debug(
"Route-map: %s, sequence: %d, prefix: %s, result: %s",
map->name, index->pref,
prefix2str(prefix, buf, sizeof(buf)),
route_map_cmd_result_str(match_ret));
}
} else
skip_match_clause = false;
/* Now we apply the matrix from above */
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
if (match_ret == RMAP_NOOP)
/*
* Do not change the return value. Retain the previous
* return value. Previous values can be:
* 1)permitmatch (if a nomatch was never
* seen before in this route-map.)
* 2)denymatch (if a nomatch was seen earlier in one
* of the previous sequences)
*/
/*
* 'cont' from matrix - continue to next route-map
* sequence
*/
continue;
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
else if (match_ret == RMAP_NOMATCH) {
/*
* The return value is now changed to denymatch.
* So from here on out, even if we see more noops,
* we retain this return value and return this
* eventually if there are no matches.
*/
ret = RMAP_DENYMATCH;
/*
* 'cont' from matrix - continue to next route-map
* sequence
*/
continue;
} else if (match_ret == RMAP_MATCH) {
if (index->type == RMAP_PERMIT)
/* 'action' */
{
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
/* Match succeeded, rmap is of type permit */
ret = RMAP_PERMITMATCH;
/* permit+match must execute sets */
for (set = index->set_list.head; set;
set = set->next)
lib: Introducing a 3rd state for route-map match cmd: RMAP_NOOP Introducing a 3rd state for route_map_apply library function: RMAP_NOOP Traditionally route map MATCH rule apis were designed to return a binary response, consisting of either RMAP_MATCH or RMAP_NOMATCH. (Route-map SET rule apis return RMAP_OKAY or RMAP_ERROR). Depending on this response, the following statemachine decided the course of action: State1: If match cmd returns RMAP_MATCH then, keep existing behaviour. If routemap type is PERMIT, execute set cmds or call cmds if applicable, otherwise PERMIT! Else If routemap type is DENY, we DENYMATCH right away State2: If match cmd returns RMAP_NOMATCH, continue on to next route-map. If there are no other rules or if all the rules return RMAP_NOMATCH, return DENYMATCH We require a 3rd state because of the following situation: The issue - what if, the rule api needs to abort or ignore a rule?: "match evpn vni xx" route-map filter can be applied to incoming routes regardless of whether the tunnel type is vxlan or mpls. This rule should be N/A for mpls based evpn route, but applicable to only vxlan based evpn route. Also, this rule should be applicable for routes with VNI label only, and not for routes without labels. For example, type 3 and type 4 EVPN routes do not have labels, so, this match cmd should let them through. Today, the filter produces either a match or nomatch response regardless of whether it is mpls/vxlan, resulting in either permitting or denying the route.. So an mpls evpn route may get filtered out incorrectly. Eg: "route-map RM1 permit 10 ; match evpn vni 20" or "route-map RM2 deny 20 ; match vni 20" With the introduction of the 3rd state, we can abort this rule check safely. How? The rules api can now return RMAP_NOOP to indicate that it encountered an invalid check, and needs to abort just that rule, but continue with other rules. As a result we have a 3rd state: State3: If match cmd returned RMAP_NOOP Then, proceed to other route-map, otherwise if there are no more rules or if all the rules return RMAP_NOOP, then, return RMAP_PERMITMATCH. Signed-off-by: Lakshman Krishnamoorthy <lkrishnamoor@vmware.com>
2019-06-19 23:04:36 +02:00
/*
* set cmds return RMAP_OKAY or
* RMAP_ERROR. We do not care if
* set succeeded or not. So, ignore
* return code.
*/
(void) (*set->cmd->func_apply)(
set->value, prefix, type,
object);
/* Call another route-map if available */
if (index->nextrm) {
struct route_map *nextrm =
route_map_lookup_by_name(
index->nextrm);
if (nextrm) /* Target route-map found,
jump to it */
{
recursion++;
ret = route_map_apply(
nextrm, prefix, type,
object);
recursion--;
}
/* If nextrm returned 'deny', finish. */
if (ret == RMAP_DENYMATCH)
goto route_map_apply_end;
}
switch (index->exitpolicy) {
case RMAP_EXIT:
goto route_map_apply_end;
case RMAP_NEXT:
continue;
case RMAP_GOTO: {
/* Find the next clause to jump to */
struct route_map_index *next =
index->next;
int nextpref = index->nextpref;
while (next && next->pref < nextpref) {
index = next;
next = next->next;
}
if (next == NULL) {
/* No clauses match! */
goto route_map_apply_end;
}
}
}
} else if (index->type == RMAP_DENY)
/* 'deny' */
{
ret = RMAP_DENYMATCH;
goto route_map_apply_end;
}
}
}
route_map_apply_end:
if (rmap_debug) {
zlog_debug("Route-map: %s, prefix: %s, result: %s",
(map ? map->name : "null"),
prefix2str(prefix, buf, sizeof(buf)),
route_map_result_str(ret));
}
return (ret);
}
void route_map_add_hook(void (*func)(const char *))
{
route_map_master.add_hook = func;
}
void route_map_delete_hook(void (*func)(const char *))
{
route_map_master.delete_hook = func;
}
void route_map_event_hook(void (*func)(const char *name))
{
route_map_master.event_hook = func;
2002-12-13 21:15:29 +01:00
}
/* Routines for route map dependency lists and dependency processing */
static bool route_map_rmap_hash_cmp(const void *p1, const void *p2)
{
return strcmp(((const struct route_map_dep_data *)p1)->rname,
((const struct route_map_dep_data *)p2)->rname)
== 0;
}
static bool route_map_dep_hash_cmp(const void *p1, const void *p2)
{
return (strcmp(((const struct route_map_dep *)p1)->dep_name,
(const char *)p2)
== 0);
}
static void route_map_clear_reference(struct hash_bucket *bucket, void *arg)
{
struct route_map_dep *dep = bucket->data;
struct route_map_dep_data *dep_data = NULL, tmp_dep_data;
if (arg) {
memset(&tmp_dep_data, 0, sizeof(struct route_map_dep_data));
tmp_dep_data.rname = arg;
dep_data = hash_release(dep->dep_rmap_hash,
&tmp_dep_data);
if (dep_data) {
XFREE(MTYPE_ROUTE_MAP_NAME, dep_data->rname);
XFREE(MTYPE_ROUTE_MAP_DEP_DATA, dep_data);
}
if (!dep->dep_rmap_hash->count) {
dep = hash_release(dep->this_hash,
(void *)dep->dep_name);
hash_free(dep->dep_rmap_hash);
XFREE(MTYPE_ROUTE_MAP_NAME, dep->dep_name);
XFREE(MTYPE_ROUTE_MAP_DEP, dep);
}
}
}
static void route_map_clear_all_references(char *rmap_name)
{
int i;
for (i = 1; i < ROUTE_MAP_DEP_MAX; i++) {
hash_iterate(route_map_dep_hash[i], route_map_clear_reference,
(void *)rmap_name);
}
}
static unsigned int route_map_dep_data_hash_make_key(const void *p)
{
const struct route_map_dep_data *dep_data = p;
return string_hash_make(dep_data->rname);
}
static void *route_map_dep_hash_alloc(void *p)
{
char *dep_name = (char *)p;
struct route_map_dep *dep_entry;
dep_entry = XCALLOC(MTYPE_ROUTE_MAP_DEP, sizeof(struct route_map_dep));
dep_entry->dep_name = XSTRDUP(MTYPE_ROUTE_MAP_NAME, dep_name);
dep_entry->dep_rmap_hash =
hash_create_size(8, route_map_dep_data_hash_make_key,
route_map_rmap_hash_cmp, "Route Map Dep Hash");
dep_entry->this_hash = NULL;
return dep_entry;
}
static void *route_map_name_hash_alloc(void *p)
{
struct route_map_dep_data *dep_data = NULL, *tmp_dep_data = NULL;
dep_data = XCALLOC(MTYPE_ROUTE_MAP_DEP_DATA,
sizeof(struct route_map_dep_data));
tmp_dep_data = p;
dep_data->rname = XSTRDUP(MTYPE_ROUTE_MAP_NAME, tmp_dep_data->rname);
return dep_data;
}
static unsigned int route_map_dep_hash_make_key(const void *p)
{
return (string_hash_make((char *)p));
}
static void route_map_print_dependency(struct hash_bucket *bucket, void *data)
{
struct route_map_dep_data *dep_data = bucket->data;
char *rmap_name = dep_data->rname;
char *dep_name = data;
zlog_debug("%s: Dependency for %s: %s", __FUNCTION__, dep_name,
rmap_name);
}
static int route_map_dep_update(struct hash *dephash, const char *dep_name,
const char *rmap_name, route_map_event_t type)
{
struct route_map_dep *dep = NULL;
char *dname, *rname;
int ret = 0;
struct route_map_dep_data *dep_data = NULL, *ret_dep_data = NULL;
struct route_map_dep_data tmp_dep_data;
dname = XSTRDUP(MTYPE_ROUTE_MAP_NAME, dep_name);
rname = XSTRDUP(MTYPE_ROUTE_MAP_NAME, rmap_name);
switch (type) {
case RMAP_EVENT_PLIST_ADDED:
case RMAP_EVENT_CLIST_ADDED:
case RMAP_EVENT_ECLIST_ADDED:
case RMAP_EVENT_ASLIST_ADDED:
case RMAP_EVENT_LLIST_ADDED:
case RMAP_EVENT_CALL_ADDED:
case RMAP_EVENT_FILTER_ADDED:
if (rmap_debug)
zlog_debug("Adding dependency for filter %s in route-map %s",
dep_name, rmap_name);
dep = (struct route_map_dep *)hash_get(
dephash, dname, route_map_dep_hash_alloc);
if (!dep) {
ret = -1;
goto out;
}
if (!dep->this_hash)
dep->this_hash = dephash;
memset(&tmp_dep_data, 0, sizeof(struct route_map_dep_data));
tmp_dep_data.rname = rname;
dep_data = hash_lookup(dep->dep_rmap_hash, &tmp_dep_data);
if (!dep_data)
dep_data = hash_get(dep->dep_rmap_hash, &tmp_dep_data,
route_map_name_hash_alloc);
dep_data->refcnt++;
break;
case RMAP_EVENT_PLIST_DELETED:
case RMAP_EVENT_CLIST_DELETED:
case RMAP_EVENT_ECLIST_DELETED:
case RMAP_EVENT_ASLIST_DELETED:
case RMAP_EVENT_LLIST_DELETED:
case RMAP_EVENT_CALL_DELETED:
case RMAP_EVENT_FILTER_DELETED:
if (rmap_debug)
zlog_debug("Deleting dependency for filter %s in route-map %s",
dep_name, rmap_name);
dep = (struct route_map_dep *)hash_get(dephash, dname, NULL);
if (!dep) {
goto out;
}
memset(&tmp_dep_data, 0, sizeof(struct route_map_dep_data));
tmp_dep_data.rname = rname;
dep_data = hash_lookup(dep->dep_rmap_hash, &tmp_dep_data);
dep_data->refcnt--;
if (!dep_data->refcnt) {
ret_dep_data = hash_release(dep->dep_rmap_hash,
&tmp_dep_data);
if (ret_dep_data) {
XFREE(MTYPE_ROUTE_MAP_NAME,
ret_dep_data->rname);
XFREE(MTYPE_ROUTE_MAP_DEP_DATA, ret_dep_data);
}
}
if (!dep->dep_rmap_hash->count) {
dep = hash_release(dephash, dname);
hash_free(dep->dep_rmap_hash);
XFREE(MTYPE_ROUTE_MAP_NAME, dep->dep_name);
XFREE(MTYPE_ROUTE_MAP_DEP, dep);
}
break;
case RMAP_EVENT_SET_ADDED:
case RMAP_EVENT_SET_DELETED:
case RMAP_EVENT_SET_REPLACED:
case RMAP_EVENT_MATCH_ADDED:
case RMAP_EVENT_MATCH_DELETED:
case RMAP_EVENT_MATCH_REPLACED:
case RMAP_EVENT_INDEX_ADDED:
case RMAP_EVENT_INDEX_DELETED:
break;
}
if (dep) {
if (rmap_debug)
hash_iterate(dep->dep_rmap_hash,
route_map_print_dependency, dname);
}
out:
XFREE(MTYPE_ROUTE_MAP_NAME, rname);
XFREE(MTYPE_ROUTE_MAP_NAME, dname);
return ret;
}
static struct hash *route_map_get_dep_hash(route_map_event_t event)
{
struct hash *upd8_hash = NULL;
switch (event) {
case RMAP_EVENT_PLIST_ADDED:
case RMAP_EVENT_PLIST_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_PLIST];
break;
case RMAP_EVENT_CLIST_ADDED:
case RMAP_EVENT_CLIST_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_CLIST];
break;
case RMAP_EVENT_ECLIST_ADDED:
case RMAP_EVENT_ECLIST_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_ECLIST];
break;
case RMAP_EVENT_ASLIST_ADDED:
case RMAP_EVENT_ASLIST_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_ASPATH];
break;
case RMAP_EVENT_LLIST_ADDED:
case RMAP_EVENT_LLIST_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_LCLIST];
break;
case RMAP_EVENT_CALL_ADDED:
case RMAP_EVENT_CALL_DELETED:
case RMAP_EVENT_MATCH_ADDED:
case RMAP_EVENT_MATCH_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_RMAP];
break;
case RMAP_EVENT_FILTER_ADDED:
case RMAP_EVENT_FILTER_DELETED:
upd8_hash = route_map_dep_hash[ROUTE_MAP_DEP_FILTER];
break;
/*
* Should we actually be ignoring these?
* I am not sure but at this point in time, let
* us get them into this switch and we can peel
* them into the appropriate place in the future
*/
case RMAP_EVENT_SET_ADDED:
case RMAP_EVENT_SET_DELETED:
case RMAP_EVENT_SET_REPLACED:
case RMAP_EVENT_MATCH_REPLACED:
case RMAP_EVENT_INDEX_ADDED:
case RMAP_EVENT_INDEX_DELETED:
upd8_hash = NULL;
break;
}
return (upd8_hash);
}
static void route_map_process_dependency(struct hash_bucket *bucket, void *data)
{
struct route_map_dep_data *dep_data = NULL;
char *rmap_name = NULL;
dep_data = bucket->data;
rmap_name = dep_data->rname;
if (rmap_debug)
zlog_debug("Notifying %s of dependency", rmap_name);
if (route_map_master.event_hook)
(*route_map_master.event_hook)(rmap_name);
}
void route_map_upd8_dependency(route_map_event_t type, const char *arg,
const char *rmap_name)
{
struct hash *upd8_hash = NULL;
if ((upd8_hash = route_map_get_dep_hash(type))) {
route_map_dep_update(upd8_hash, arg, rmap_name, type);
if (type == RMAP_EVENT_CALL_ADDED) {
/* Execute hook. */
if (route_map_master.add_hook)
(*route_map_master.add_hook)(rmap_name);
} else if (type == RMAP_EVENT_CALL_DELETED) {
/* Execute hook. */
if (route_map_master.delete_hook)
(*route_map_master.delete_hook)(rmap_name);
}
}
}
void route_map_notify_dependencies(const char *affected_name,
route_map_event_t event)
{
struct route_map_dep *dep;
struct hash *upd8_hash;
char *name;
if (!affected_name)
return;
name = XSTRDUP(MTYPE_ROUTE_MAP_NAME, affected_name);
if ((upd8_hash = route_map_get_dep_hash(event)) == NULL) {
XFREE(MTYPE_ROUTE_MAP_NAME, name);
return;
}
dep = (struct route_map_dep *)hash_get(upd8_hash, name, NULL);
if (dep) {
if (!dep->this_hash)
dep->this_hash = upd8_hash;
if (rmap_debug)
zlog_debug("Filter %s updated", dep->dep_name);
hash_iterate(dep->dep_rmap_hash, route_map_process_dependency,
(void *)event);
}
XFREE(MTYPE_ROUTE_MAP_NAME, name);
}
2002-12-13 21:15:29 +01:00
/* VTY related functions. */
DEFUN(no_routemap_optimization, no_routemap_optimization_cmd,
"no route-map optimization",
NO_STR
"route-map\n"
"optimization\n")
{
VTY_DECLVAR_CONTEXT(route_map_index, index);
index->map->optimization_disabled = true;
return CMD_SUCCESS;
}
DEFUN(routemap_optimization, routemap_optimization_cmd,
"route-map optimization",
"route-map\n"
"optimization\n")
{
VTY_DECLVAR_CONTEXT(route_map_index, index);
index->map->optimization_disabled = false;
return CMD_SUCCESS;
}
static void clear_route_map_helper(struct route_map *map)
{
struct route_map_index *index;
map->applied_clear = map->applied;
for (index = map->head; index; index = index->next)
index->applied_clear = index->applied;
}
DEFUN (rmap_clear_counters,
rmap_clear_counters_cmd,
"clear route-map counters [WORD]",
CLEAR_STR
"route-map information\n"
"counters associated with the specified route-map\n"
"route-map name\n")
{
int idx_word = 2;
struct route_map *map;
const char *name = (argc == 3 ) ? argv[idx_word]->arg : NULL;
if (name) {
map = route_map_lookup_by_name(name);
if (map)
clear_route_map_helper(map);
else {
vty_out(vty, "%s: 'route-map %s' not found\n",
frr_protonameinst, name);
return CMD_SUCCESS;
}
} else {
for (map = route_map_master.head; map; map = map->next)
clear_route_map_helper(map);
}
return CMD_SUCCESS;
}
DEFUN (rmap_show_name,
rmap_show_name_cmd,
[zebra] Routemap support on received routes, with 'set src' command (linux) 2007-05-01 David L Stevens <dlstevens@us.ibm.com> * (general) These changes collectively add route-map and prefix-list support to zebra and fix a bug in "show route-map" (with no argument). * doc/main.texi: added route-map, prefix-list, ip protocol and set src documentation * lib/command.h: added PROTOCOL_NODE type * lib/log.c: (proto_name2num) new function, protocol name to number translation. * lib/routemap.c: (vty_show_route_map) fixed "show route-map" without route-map name * lib/routemap.h: added RMAP_ZEBRA type * lib/zebra.h: added proto_name2num() prototype * vtysh/extract.pl.in: added VTYSH_ZEBRA flag for route-map and plist * vtysh/Makefile.am: added zebra_routemap.c * vtysh/vtysh.h: added VTYSH_ZEBRA flag to VTYSH_RMAP * zebra/connected.c: (connected_up_ipv4) added src preference argument to rib_add_ipv4() * zebra/kernel_socket.c: (rtm_read) ditto * zebra/main.c: added prefix list initialization * zebra/Makefile.am: added zebra_routemap.c source file * zebra/rib.h: added generic address union "g_addr" and use in existing places that had an explicit union. Added "src" to struct nexthop. Added preferred src arg to nexthop_ipv4_add and rib_add_ipv4. * zebra/rt_netlink.c: (netlink_routing_table) set preferred source on netlink messages. (netlink_route_change) ditto (netlink_route_multipath) ditto. * zebra/rtread_getmsg.c: (handle_route_entry) added (NULL) src to rib_add_ipv4() call. * zebra/rtread_proc.c: (proc_route_read) ditto * zebra/zebra_rib.c: (nexthop_ipv4_add) add src argument. (nexthop_ipv4_ifindex_add) ditto (rib_add_ipv4) ditto (nexthop_active_check) Add route-map processing. * zebra/zebra_routemap.c: new file for zebra route-map commands. * zebra/zebra_vty.c: (ip_protocol_cmd) Apply route-map to protocol (vty_show_ip_route_detail) added "src" printing (vty_show_ip_route) ditto (show_ip_protocol_cmd) new command, list routemaps. (config_write_protocol) write out routemap protocl config. (zebra_vty_init) Install the new routemap protocol commands. * zebra/zserv.c: (zread_ipv4_add) added (NULL) src arg (zebra_init) init zebra route-maps. * zebra/zserv.h: add zebra_route_map_init
2007-05-02 18:05:35 +02:00
"show route-map [WORD]",
SHOW_STR
"route-map information\n"
"route-map name\n")
{
int idx_word = 2;
const char *name = (argc == 3) ? argv[idx_word]->arg : NULL;
return vty_show_route_map(vty, name);
}
DEFUN (rmap_show_unused,
rmap_show_unused_cmd,
"show route-map-unused",
SHOW_STR
"unused route-map information\n")
{
return vty_show_unused_route_map(vty);
}
DEFUN (debug_rmap,
debug_rmap_cmd,
"debug route-map",
DEBUG_STR
"Debug option set for route-maps\n")
{
rmap_debug = true;
return CMD_SUCCESS;
}
DEFUN (no_debug_rmap,
no_debug_rmap_cmd,
"no debug route-map",
NO_STR
DEBUG_STR
"Debug option set for route-maps\n")
{
rmap_debug = false;
return CMD_SUCCESS;
}
/* Debug node. */
static struct cmd_node rmap_debug_node = {RMAP_DEBUG_NODE, "", 1};
2002-12-13 21:15:29 +01:00
/* Configuration write function. */
static int rmap_config_write_debug(struct vty *vty)
{
int write = 0;
if (rmap_debug) {
vty_out(vty, "debug route-map\n");
write++;
}
return write;
}
/* Common route map rules */
void *route_map_rule_tag_compile(const char *arg)
{
unsigned long int tmp;
char *endptr;
route_tag_t *tag;
errno = 0;
tmp = strtoul(arg, &endptr, 0);
if (arg[0] == '\0' || *endptr != '\0' || errno || tmp > ROUTE_TAG_MAX)
return NULL;
tag = XMALLOC(MTYPE_ROUTE_MAP_COMPILED, sizeof(*tag));
*tag = tmp;
return tag;
}
void route_map_rule_tag_free(void *rule)
{
XFREE(MTYPE_ROUTE_MAP_COMPILED, rule);
}
void route_map_finish(void)
{
int i;
vector_free(route_match_vec);
route_match_vec = NULL;
vector_free(route_set_vec);
route_set_vec = NULL;
/*
* All protocols are setting these to NULL
* by default on shutdown( route_map_finish )
* Why are we making them do this work?
*/
route_map_master.add_hook = NULL;
route_map_master.delete_hook = NULL;
route_map_master.event_hook = NULL;
/* cleanup route_map */
while (route_map_master.head) {
struct route_map *map = route_map_master.head;
map->to_be_processed = false;
route_map_delete(map);
}
for (i = 1; i < ROUTE_MAP_DEP_MAX; i++) {
hash_free(route_map_dep_hash[i]);
route_map_dep_hash[i] = NULL;
}
hash_free(route_map_master_hash);
route_map_master_hash = NULL;
}
/* Increment the use_count counter while attaching the route map */
void route_map_counter_increment(struct route_map *map)
{
if (map)
map->use_count++;
}
/* Decrement the use_count counter while detaching the route map. */
void route_map_counter_decrement(struct route_map *map)
{
if (map) {
if (map->use_count <= 0)
return;
map->use_count--;
}
}
DEFUN_HIDDEN(show_route_map_pfx_tbl, show_route_map_pfx_tbl_cmd,
"show route-map WORD prefix-table",
SHOW_STR
"route-map\n"
"route-map name\n"
"internal prefix-table\n")
{
const char *rmap_name = argv[2]->arg;
struct route_map *rmap = NULL;
struct route_table *rm_pfx_tbl4 = NULL;
struct route_table *rm_pfx_tbl6 = NULL;
struct route_node *rn = NULL, *prn = NULL;
struct list *rmap_index_list = NULL;
struct listnode *ln = NULL, *nln = NULL;
struct route_map_index *index = NULL;
struct prefix *p = NULL, *pp = NULL;
char buf[SU_ADDRSTRLEN], pbuf[SU_ADDRSTRLEN];
uint8_t len = 54;
vty_out(vty, "%s:\n", frr_protonameinst);
rmap = route_map_lookup_by_name(rmap_name);
if (rmap) {
rm_pfx_tbl4 = rmap->ipv4_prefix_table;
if (rm_pfx_tbl4) {
vty_out(vty, "\n%s%43s%s\n", "IPv4 Prefix", "",
"Route-map Index List");
vty_out(vty, "%s%39s%s\n", "_______________", "",
"____________________");
for (rn = route_top(rm_pfx_tbl4); rn;
rn = route_next(rn)) {
p = &rn->p;
vty_out(vty, " %s/%d (%d)\n",
inet_ntop(p->family, &p->u.prefix, buf,
SU_ADDRSTRLEN),
p->prefixlen, rn->lock);
vty_out(vty, "(P) ");
prn = rn->parent;
if (prn) {
pp = &prn->p;
vty_out(vty, "%s/%d\n",
inet_ntop(pp->family,
&pp->u.prefix, pbuf,
SU_ADDRSTRLEN),
pp->prefixlen);
}
vty_out(vty, "\n");
rmap_index_list = (struct list *)rn->info;
if (!rmap_index_list
|| !listcount(rmap_index_list))
vty_out(vty, "%*s%s\n", len, "", "-");
else
for (ALL_LIST_ELEMENTS(rmap_index_list,
ln, nln,
index)) {
vty_out(vty, "%*s%s seq %d\n",
len, "",
index->map->name,
index->pref);
}
vty_out(vty, "\n");
}
}
rm_pfx_tbl6 = rmap->ipv6_prefix_table;
if (rm_pfx_tbl6) {
vty_out(vty, "\n%s%43s%s\n", "IPv6 Prefix", "",
"Route-map Index List");
vty_out(vty, "%s%39s%s\n", "_______________", "",
"____________________");
for (rn = route_top(rm_pfx_tbl6); rn;
rn = route_next(rn)) {
p = &rn->p;
vty_out(vty, " %s/%d (%d)\n",
inet_ntop(p->family, &p->u.prefix, buf,
SU_ADDRSTRLEN),
p->prefixlen, rn->lock);
vty_out(vty, "(P) ");
prn = rn->parent;
if (prn) {
pp = &prn->p;
vty_out(vty, "%s/%d\n",
inet_ntop(pp->family,
&pp->u.prefix, pbuf,
SU_ADDRSTRLEN),
pp->prefixlen);
}
vty_out(vty, "\n");
rmap_index_list = (struct list *)rn->info;
if (!rmap_index_list
|| !listcount(rmap_index_list))
vty_out(vty, "%*s%s\n", len, "", "-");
else
for (ALL_LIST_ELEMENTS(rmap_index_list,
ln, nln,
index)) {
vty_out(vty, "%*s%s seq %d\n",
len, "",
index->map->name,
index->pref);
}
vty_out(vty, "\n");
}
}
}
vty_out(vty, "\n");
return CMD_SUCCESS;
}
2002-12-13 21:15:29 +01:00
/* Initialization of route map vector. */
void route_map_init(void)
2002-12-13 21:15:29 +01:00
{
int i;
/* Make vector for match and set. */
route_match_vec = vector_init(1);
route_set_vec = vector_init(1);
route_map_master_hash =
hash_create_size(8, route_map_hash_key_make, route_map_hash_cmp,
"Route Map Master Hash");
for (i = 1; i < ROUTE_MAP_DEP_MAX; i++)
route_map_dep_hash[i] = hash_create_size(
8, route_map_dep_hash_make_key, route_map_dep_hash_cmp,
"Route Map Dep Hash");
rmap_debug = false;
route_map_cli_init();
/* Install route map top node. */
install_node(&rmap_debug_node, rmap_config_write_debug);
/* Install route map commands. */
install_element(CONFIG_NODE, &debug_rmap_cmd);
install_element(CONFIG_NODE, &no_debug_rmap_cmd);
/* Install show command */
install_element(ENABLE_NODE, &rmap_clear_counters_cmd);
install_element(ENABLE_NODE, &rmap_show_name_cmd);
install_element(ENABLE_NODE, &rmap_show_unused_cmd);
install_element(ENABLE_NODE, &debug_rmap_cmd);
install_element(ENABLE_NODE, &no_debug_rmap_cmd);
install_element(RMAP_NODE, &routemap_optimization_cmd);
install_element(RMAP_NODE, &no_routemap_optimization_cmd);
install_element(ENABLE_NODE, &show_route_map_pfx_tbl_cmd);
2002-12-13 21:15:29 +01:00
}