1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1980, 1986, 1993
5 * The Regents of the University of California. All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the University nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
18 *
19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE.
30 */
31
32#ifndef _NET_ROUTE_H_
33#define _NET_ROUTE_H_
34
35#include <net/vnet.h>
36
37/*
38 * Kernel resident routing tables.
39 *
40 * The routing tables are initialized when interface addresses
41 * are set by making entries for all directly connected interfaces.
42 */
43
44/*
45 * Struct route consiste of a destination address,
46 * a route entry pointer, link-layer prepend data pointer along
47 * with its length.
48 */
49struct route {
50 struct nhop_object *ro_nh;
51 struct llentry *ro_lle;
52 /*
53 * ro_prepend and ro_plen are only used for bpf to pass in a
54 * preformed header. They are not cacheable.
55 */
56 char *ro_prepend;
57 uint16_t ro_plen;
58 uint16_t ro_flags;
59 uint16_t ro_mtu; /* saved ro_rt mtu */
60 uint16_t spare;
61 struct sockaddr ro_dst;
62};
63
64#define RT_L2_ME_BIT 2 /* dst L2 addr is our address */
65#define RT_MAY_LOOP_BIT 3 /* dst may require loop copy */
66#define RT_HAS_HEADER_BIT 4 /* mbuf already have its header prepended */
67
68#define RT_L2_ME (1 << RT_L2_ME_BIT) /* 0x0004 */
69#define RT_MAY_LOOP (1 << RT_MAY_LOOP_BIT) /* 0x0008 */
70#define RT_HAS_HEADER (1 << RT_HAS_HEADER_BIT) /* 0x0010 */
71
72#define RT_REJECT 0x0020 /* Destination is reject */
73#define RT_BLACKHOLE 0x0040 /* Destination is blackhole */
74#define RT_HAS_GW 0x0080 /* Destination has GW */
75#define RT_LLE_CACHE 0x0100 /* Cache link layer */
76
77struct rt_metrics {
78 u_long rmx_locks; /* Kernel must leave these values alone */
79 u_long rmx_mtu; /* MTU for this path */
80 u_long rmx_hopcount; /* max hops expected */
81 u_long rmx_expire; /* lifetime for route, e.g. redirect */
82 u_long rmx_recvpipe; /* inbound delay-bandwidth product */
83 u_long rmx_sendpipe; /* outbound delay-bandwidth product */
84 u_long rmx_ssthresh; /* outbound gateway buffer limit */
85 u_long rmx_rtt; /* estimated round trip time */
86 u_long rmx_rttvar; /* estimated rtt variance */
87 u_long rmx_pksent; /* packets sent using this route */
88 u_long rmx_weight; /* route weight */
89 u_long rmx_nhidx; /* route nexhop index */
90 u_long rmx_filler[2]; /* will be used for T/TCP later */
91};
92
93/*
94 * rmx_rtt and rmx_rttvar are stored as microseconds;
95 */
96#define RTM_RTTUNIT 1000000 /* units for rtt, rttvar, as units per sec */
97
98/* lle state is exported in rmx_state rt_metrics field */
99#define rmx_state rmx_weight
100
101/* default route weight */
102#define RT_DEFAULT_WEIGHT 1
103#define RT_MAX_WEIGHT 16777215 /* 3 bytes */
104
105/*
106 * Keep a generation count of routing table, incremented on route addition,
107 * so we can invalidate caches. This is accessed without a lock, as precision
108 * is not required.
109 */
110typedef volatile u_int rt_gen_t; /* tree generation (for adds) */
111#define RT_GEN(fibnum, af) rt_tables_get_gen(fibnum, af)
112
113#define RT_DEFAULT_FIB 0 /* Explicitly mark fib=0 restricted cases */
114#define RT_ALL_FIBS -1 /* Announce event for every fib */
115#ifdef _KERNEL
116VNET_DECLARE(uint32_t, _rt_numfibs); /* number of existing route tables */
117#define V_rt_numfibs VNET(_rt_numfibs)
118/* temporary compat arg */
119#define rt_numfibs V_rt_numfibs
120VNET_DECLARE(u_int, rt_add_addr_allfibs); /* Announce interfaces to all fibs */
121#define V_rt_add_addr_allfibs VNET(rt_add_addr_allfibs)
122
123/* Calculate flowid for locally-originated packets */
124#define V_fib_hash_outbound VNET(fib_hash_outbound)
125VNET_DECLARE(u_int, fib_hash_outbound);
126
127/* Outbound flowid generation rules */
128#ifdef RSS
129
130#define fib4_calc_packet_hash xps_proto_software_hash_v4
131#define fib6_calc_packet_hash xps_proto_software_hash_v6
132#define CALC_FLOWID_OUTBOUND_SENDTO true
133
134#ifdef ROUTE_MPATH
135#define CALC_FLOWID_OUTBOUND V_fib_hash_outbound
136#else
137#define CALC_FLOWID_OUTBOUND false
138#endif
139
140#else /* !RSS */
141
142#define fib4_calc_packet_hash fib4_calc_software_hash
143#define fib6_calc_packet_hash fib6_calc_software_hash
144
145#ifdef ROUTE_MPATH
146#define CALC_FLOWID_OUTBOUND_SENDTO V_fib_hash_outbound
147#define CALC_FLOWID_OUTBOUND V_fib_hash_outbound
148#else
149#define CALC_FLOWID_OUTBOUND_SENDTO false
150#define CALC_FLOWID_OUTBOUND false
151#endif
152
153#endif /* RSS */
154
155
156#endif /* _KERNEL */
157
158/*
159 * We distinguish between routes to hosts and routes to networks,
160 * preferring the former if available. For each route we infer
161 * the interface to use from the gateway address supplied when
162 * the route was entered. Routes that forward packets through
163 * gateways are marked so that the output routines know to address the
164 * gateway rather than the ultimate destination.
165 */
166#define RTF_UP 0x1 /* route usable */
167#define RTF_GATEWAY 0x2 /* destination is a gateway */
168#define RTF_HOST 0x4 /* host entry (net otherwise) */
169#define RTF_REJECT 0x8 /* host or net unreachable */
170#define RTF_DYNAMIC 0x10 /* created dynamically (by redirect) */
171#define RTF_MODIFIED 0x20 /* modified dynamically (by redirect) */
172#define RTF_DONE 0x40 /* message confirmed */
173/* 0x80 unused, was RTF_DELCLONE */
174/* 0x100 unused, was RTF_CLONING */
175#define RTF_XRESOLVE 0x200 /* external daemon resolves name */
176#define RTF_LLINFO 0x400 /* DEPRECATED - exists ONLY for backward
177 compatibility */
178#define RTF_LLDATA 0x400 /* used by apps to add/del L2 entries */
179#define RTF_STATIC 0x800 /* manually added */
180#define RTF_BLACKHOLE 0x1000 /* just discard pkts (during updates) */
181#define RTF_PROTO2 0x4000 /* protocol specific routing flag */
182#define RTF_PROTO1 0x8000 /* protocol specific routing flag */
183/* 0x10000 unused, was RTF_PRCLONING */
184/* 0x20000 unused, was RTF_WASCLONED */
185#define RTF_PROTO3 0x40000 /* protocol specific routing flag */
186#define RTF_FIXEDMTU 0x80000 /* MTU was explicitly specified */
187#define RTF_PINNED 0x100000 /* route is immutable */
188#define RTF_LOCAL 0x200000 /* route represents a local address */
189#define RTF_BROADCAST 0x400000 /* route represents a bcast address */
190#define RTF_MULTICAST 0x800000 /* route represents a mcast address */
191 /* 0x8000000 and up unassigned */
192#define RTF_STICKY 0x10000000 /* always route dst->src */
193
194/* 0x40000000 unused, was RTF_RNH_LOCKED */
195
196#define RTF_GWFLAG_COMPAT 0x80000000 /* a compatibility bit for interacting
197 with existing routing apps */
198
199/* Mask of RTF flags that are allowed to be modified by RTM_CHANGE. */
200#define RTF_FMASK \
201 (RTF_PROTO1 | RTF_PROTO2 | RTF_PROTO3 | RTF_BLACKHOLE | \
202 RTF_REJECT | RTF_STATIC | RTF_STICKY)
203
204/*
205 * fib_ nexthop API flags.
206 */
207
208/* Consumer-visible nexthop info flags */
209#define NHF_MULTIPATH 0x0008 /* Nexhop is a nexthop group */
210#define NHF_REJECT 0x0010 /* RTF_REJECT */
211#define NHF_BLACKHOLE 0x0020 /* RTF_BLACKHOLE */
212#define NHF_REDIRECT 0x0040 /* RTF_DYNAMIC|RTF_MODIFIED */
213#define NHF_DEFAULT 0x0080 /* Default route */
214#define NHF_BROADCAST 0x0100 /* RTF_BROADCAST */
215#define NHF_GATEWAY 0x0200 /* RTF_GATEWAY */
216#define NHF_HOST 0x0400 /* RTF_HOST */
217
218/* Nexthop request flags */
219#define NHR_NONE 0x00 /* empty flags field */
220#define NHR_REF 0x01 /* reference nexhop */
221#define NHR_NODEFAULT 0x02 /* uRPF: do not consider default route */
222
223/* Control plane route request flags */
224#define NHR_COPY 0x100 /* Copy rte data */
225#define NHR_UNLOCKED 0x200 /* Do not lock table */
226
227/*
228 * Routing statistics.
229 */
230struct rtstat {
231 uint64_t rts_badredirect; /* bogus redirect calls */
232 uint64_t rts_dynamic; /* routes created by redirects */
233 uint64_t rts_newgateway; /* routes modified by redirects */
234 uint64_t rts_unreach; /* lookups which failed */
235 uint64_t rts_wildcard; /* lookups satisfied by a wildcard */
236 uint64_t rts_nh_idx_alloc_failure; /* nexthop index alloc failure*/
237 uint64_t rts_nh_alloc_failure; /* nexthop allocation failure*/
238 uint64_t rts_add_failure; /* # of route addition failures */
239 uint64_t rts_add_retry; /* # of route addition retries */
240 uint64_t rts_del_failure; /* # of route deletion failure */
241 uint64_t rts_del_retry; /* # of route deletion retries */
242 uint64_t rts_spare[5];
243};
244
245/*
246 * Structures for routing messages.
247 */
248struct rt_msghdr {
249 u_short rtm_msglen; /* to skip over non-understood messages */
250 u_char rtm_version; /* future binary compatibility */
251 u_char rtm_type; /* message type */
252 u_short rtm_index; /* index for associated ifp */
253 u_short _rtm_spare1;
254 int rtm_flags; /* flags, incl. kern & message, e.g. DONE */
255 int rtm_addrs; /* bitmask identifying sockaddrs in msg */
256 pid_t rtm_pid; /* identify sender */
257 int rtm_seq; /* for sender to identify action */
258 int rtm_errno; /* why failed */
259 int rtm_fmask; /* bitmask used in RTM_CHANGE message */
260 u_long rtm_inits; /* which metrics we are initializing */
261 struct rt_metrics rtm_rmx; /* metrics themselves */
262};
263
264#define RTM_VERSION 5 /* Up the ante and ignore older versions */
265
266#ifndef NETLINK_COMPAT
267/*
268 * Message types.
269 *
270 * The format for each message is annotated below using the following
271 * identifiers:
272 *
273 * (1) struct rt_msghdr
274 * (2) struct ifa_msghdr
275 * (3) struct if_msghdr
276 * (4) struct ifma_msghdr
277 * (5) struct if_announcemsghdr
278 *
279 */
280#define RTM_ADD 0x1 /* (1) Add Route */
281#define RTM_DELETE 0x2 /* (1) Delete Route */
282#define RTM_CHANGE 0x3 /* (1) Change Metrics or flags */
283#define RTM_GET 0x4 /* (1) Report Metrics */
284#define RTM_LOSING 0x5 /* (1) Kernel Suspects Partitioning */
285#define RTM_REDIRECT 0x6 /* (1) Told to use different route */
286#define RTM_MISS 0x7 /* (1) Lookup failed on this address */
287#define RTM_LOCK 0x8 /* (1) fix specified metrics */
288 /* 0x9 */
289 /* 0xa */
290#define RTM_RESOLVE 0xb /* (1) req to resolve dst to LL addr */
291#define RTM_NEWADDR 0xc /* (2) address being added to iface */
292#define RTM_DELADDR 0xd /* (2) address being removed from iface */
293#define RTM_IFINFO 0xe /* (3) iface going up/down etc. */
294#define RTM_NEWMADDR 0xf /* (4) mcast group membership being added to if */
295#define RTM_DELMADDR 0x10 /* (4) mcast group membership being deleted */
296#define RTM_IFANNOUNCE 0x11 /* (5) iface arrival/departure */
297#define RTM_IEEE80211 0x12 /* (5) IEEE80211 wireless event */
298#define RTM_IPFWLOG 0x13 /* (1) IPFW rule match log event */
299
300#endif /* NETLINK_COMPAT*/
301
302/*
303 * Bitmask values for rtm_inits and rmx_locks.
304 */
305#define RTV_MTU 0x1 /* init or lock _mtu */
306#define RTV_HOPCOUNT 0x2 /* init or lock _hopcount */
307#define RTV_EXPIRE 0x4 /* init or lock _expire */
308#define RTV_RPIPE 0x8 /* init or lock _recvpipe */
309#define RTV_SPIPE 0x10 /* init or lock _sendpipe */
310#define RTV_SSTHRESH 0x20 /* init or lock _ssthresh */
311#define RTV_RTT 0x40 /* init or lock _rtt */
312#define RTV_RTTVAR 0x80 /* init or lock _rttvar */
313#define RTV_WEIGHT 0x100 /* init or lock _weight */
314
315#ifndef NETLINK_COMPAT
316
317/*
318 * Bitmask values for rtm_addrs.
319 */
320#define RTA_DST 0x1 /* destination sockaddr present */
321#define RTA_GATEWAY 0x2 /* gateway sockaddr present */
322#define RTA_NETMASK 0x4 /* netmask sockaddr present */
323#define RTA_GENMASK 0x8 /* cloning mask sockaddr present */
324#define RTA_IFP 0x10 /* interface name sockaddr present */
325#define RTA_IFA 0x20 /* interface addr sockaddr present */
326#define RTA_AUTHOR 0x40 /* sockaddr for author of redirect */
327#define RTA_BRD 0x80 /* for NEWADDR, broadcast or p-p dest addr */
328
329#endif /* NETLINK_COMPAT*/
330
331/*
332 * Index offsets for sockaddr array for alternate internal encoding.
333 */
334#define RTAX_DST 0 /* destination sockaddr present */
335#define RTAX_GATEWAY 1 /* gateway sockaddr present */
336#define RTAX_NETMASK 2 /* netmask sockaddr present */
337#define RTAX_GENMASK 3 /* cloning mask sockaddr present */
338#define RTAX_IFP 4 /* interface name sockaddr present */
339#define RTAX_IFA 5 /* interface addr sockaddr present */
340#define RTAX_AUTHOR 6 /* sockaddr for author of redirect */
341#define RTAX_BRD 7 /* for NEWADDR, broadcast or p-p dest addr */
342#define RTAX_MAX 8 /* size of array to allocate */
343
344struct rtentry;
345struct nhop_object;
346typedef int rib_filter_f_t(const struct rtentry *, const struct nhop_object *,
347 void *);
348
349struct rt_addrinfo {
350 int rti_addrs; /* Route RTF_ flags */
351 int rti_flags; /* Route RTF_ flags */
352 struct sockaddr *rti_info[RTAX_MAX]; /* Sockaddr data */
353 struct ifaddr *rti_ifa; /* value of rt_ifa addr */
354 struct ifnet *rti_ifp; /* route interface */
355 rib_filter_f_t *rti_filter; /* filter function */
356 void *rti_filterdata; /* filter parameters */
357 u_long rti_mflags; /* metrics RTV_ flags */
358 u_long rti_spare; /* Will be used for fib */
359 struct rt_metrics *rti_rmx; /* Pointer to route metrics */
360};
361
362/*
363 * This macro returns the size of a struct sockaddr when passed
364 * through a routing socket. Basically we round up sa_len to
365 * a multiple of sizeof(long), with a minimum of sizeof(long).
366 * The case sa_len == 0 should only apply to empty structures.
367 */
368#define SA_SIZE(sa) \
369 ( (((struct sockaddr *)(sa))->sa_len == 0) ? \
370 sizeof(long) : \
371 1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
372
373#define sa_equal(a, b) ( \
374 (((const struct sockaddr *)(a))->sa_len == ((const struct sockaddr *)(b))->sa_len) && \
375 (bcmp((a), (b), ((const struct sockaddr *)(b))->sa_len) == 0))
376
377#ifdef _KERNEL
378
379#define RT_LINK_IS_UP(ifp) (!((ifp)->if_capabilities & IFCAP_LINKSTATE) \
380 || (ifp)->if_link_state == LINK_STATE_UP)
381
382#define RO_NHFREE(_ro) do { \
383 if ((_ro)->ro_nh) { \
384 NH_FREE((_ro)->ro_nh); \
385 (_ro)->ro_nh = NULL; \
386 } \
387} while (0)
388
389#define RO_INVALIDATE_CACHE(ro) do { \
390 if ((ro)->ro_lle != NULL) { \
391 LLE_FREE((ro)->ro_lle); \
392 (ro)->ro_lle = NULL; \
393 } \
394 if ((ro)->ro_nh != NULL) { \
395 NH_FREE((ro)->ro_nh); \
396 (ro)->ro_nh = NULL; \
397 } \
398 } while (0)
399
400#define RO_GET_FAMILY(ro, dst) ((ro) != NULL && \
401 (ro)->ro_flags & RT_HAS_GW \
402 ? (ro)->ro_dst.sa_family : (dst)->sa_family)
403
404/*
405 * Validate a cached route based on a supplied cookie. If there is an
406 * out-of-date cache, simply free it. Update the generation number
407 * for the new allocation
408 */
409#define NH_VALIDATE(ro, cookiep, fibnum) do { \
410 rt_gen_t cookie = RT_GEN(fibnum, (ro)->ro_dst.sa_family); \
411 if (*(cookiep) != cookie) { \
412 RO_INVALIDATE_CACHE(ro); \
413 *(cookiep) = cookie; \
414 } \
415} while (0)
416
417struct ifmultiaddr;
418struct rib_head;
419
420void rt_ieee80211msg(struct ifnet *, int, void *, size_t);
421void rt_ifmsg(struct ifnet *, int);
422void rt_missmsg(int, struct rt_addrinfo *, int, int);
423void rt_missmsg_fib(int, struct rt_addrinfo *, int, int, int);
424int rt_addrmsg(int, struct ifaddr *, int);
425int rt_routemsg(int, struct rtentry *, struct nhop_object *, int);
426int rt_routemsg_info(int, struct rt_addrinfo *, int);
427void rt_newmaddrmsg(int, struct ifmultiaddr *);
428void rt_maskedcopy(const struct sockaddr *, struct sockaddr *,
429 const struct sockaddr *);
430struct rib_head *rt_table_init(int, int, u_int);
431void rt_table_destroy(struct rib_head *);
432u_int rt_tables_get_gen(uint32_t table, sa_family_t family);
433
434struct sockaddr *rtsock_fix_netmask(const struct sockaddr *dst,
435 const struct sockaddr *smask, struct sockaddr_storage *dmask);
436
437void rt_updatemtu(struct ifnet *);
438
439void rt_flushifroutes(struct ifnet *ifp);
440
441/* XXX MRT NEW VERSIONS THAT USE FIBs
442 * For now the protocol indepedent versions are the same as the AF_INET ones
443 * but this will change..
444 */
445int rtioctl_fib(u_long, caddr_t, u_int);
446
447/* New API */
448void rib_flush_routes_family(int family);
449struct nhop_object *rib_lookup(uint32_t fibnum, const struct sockaddr *dst,
450 uint32_t flags, uint32_t flowid);
451const char *rib_print_family(int family);
452#endif
453
454#endif