route.h 11 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369
  1. /*
  2. * INET An implementation of the TCP/IP protocol suite for the LINUX
  3. * operating system. INET is implemented using the BSD Socket
  4. * interface as the means of communication with the user level.
  5. *
  6. * Definitions for the IP router.
  7. *
  8. * Version: @(#)route.h 1.0.4 05/27/93
  9. *
  10. * Authors: Ross Biro
  11. * Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
  12. * Fixes:
  13. * Alan Cox : Reformatted. Added ip_rt_local()
  14. * Alan Cox : Support for TCP parameters.
  15. * Alexey Kuznetsov: Major changes for new routing code.
  16. * Mike McLagan : Routing by source
  17. * Robert Olsson : Added rt_cache statistics
  18. *
  19. * This program is free software; you can redistribute it and/or
  20. * modify it under the terms of the GNU General Public License
  21. * as published by the Free Software Foundation; either version
  22. * 2 of the License, or (at your option) any later version.
  23. */
  24. #ifndef _ROUTE_H
  25. #define _ROUTE_H
  26. #include <net/dst.h>
  27. #include <net/inetpeer.h>
  28. #include <net/flow.h>
  29. #include <net/inet_sock.h>
  30. #include <linux/in_route.h>
  31. #include <linux/rtnetlink.h>
  32. #include <linux/route.h>
  33. #include <linux/ip.h>
  34. #include <linux/cache.h>
  35. #include <linux/security.h>
  36. #define RTO_ONLINK 0x01
  37. #define RT_CONN_FLAGS(sk) (RT_TOS(inet_sk(sk)->tos) | sock_flag(sk, SOCK_LOCALROUTE))
  38. struct fib_nh;
  39. struct inet_peer;
  40. struct fib_info;
  41. struct rtable {
  42. struct dst_entry dst;
  43. /* Lookup key. */
  44. __be32 rt_key_dst;
  45. __be32 rt_key_src;
  46. int rt_genid;
  47. unsigned int rt_flags;
  48. __u16 rt_type;
  49. __u8 rt_key_tos;
  50. __be32 rt_dst; /* Path destination */
  51. __be32 rt_src; /* Path source */
  52. int rt_route_iif;
  53. int rt_iif;
  54. int rt_oif;
  55. __u32 rt_mark;
  56. /* Info on neighbour */
  57. __be32 rt_gateway;
  58. /* Miscellaneous cached information */
  59. __be32 rt_spec_dst; /* RFC1122 specific destination */
  60. u32 rt_peer_genid;
  61. unsigned long _peer; /* long-living peer info */
  62. struct fib_info *fi; /* for client ref to shared metrics */
  63. };
  64. static inline struct inet_peer *rt_peer_ptr(struct rtable *rt)
  65. {
  66. return inetpeer_ptr(rt->_peer);
  67. }
  68. static inline bool rt_has_peer(struct rtable *rt)
  69. {
  70. return inetpeer_ptr_is_peer(rt->_peer);
  71. }
  72. static inline void __rt_set_peer(struct rtable *rt, struct inet_peer *peer)
  73. {
  74. __inetpeer_ptr_set_peer(&rt->_peer, peer);
  75. }
  76. static inline bool rt_set_peer(struct rtable *rt, struct inet_peer *peer)
  77. {
  78. return inetpeer_ptr_set_peer(&rt->_peer, peer);
  79. }
  80. static inline void rt_init_peer(struct rtable *rt, struct inet_peer_base *base)
  81. {
  82. inetpeer_init_ptr(&rt->_peer, base);
  83. }
  84. static inline void rt_transfer_peer(struct rtable *rt, struct rtable *ort)
  85. {
  86. rt->_peer = ort->_peer;
  87. if (rt_has_peer(ort)) {
  88. struct inet_peer *peer = rt_peer_ptr(ort);
  89. atomic_inc(&peer->refcnt);
  90. }
  91. }
  92. static inline bool rt_is_input_route(const struct rtable *rt)
  93. {
  94. return rt->rt_route_iif != 0;
  95. }
  96. static inline bool rt_is_output_route(const struct rtable *rt)
  97. {
  98. return rt->rt_route_iif == 0;
  99. }
  100. struct ip_rt_acct {
  101. __u32 o_bytes;
  102. __u32 o_packets;
  103. __u32 i_bytes;
  104. __u32 i_packets;
  105. };
  106. struct rt_cache_stat {
  107. unsigned int in_hit;
  108. unsigned int in_slow_tot;
  109. unsigned int in_slow_mc;
  110. unsigned int in_no_route;
  111. unsigned int in_brd;
  112. unsigned int in_martian_dst;
  113. unsigned int in_martian_src;
  114. unsigned int out_hit;
  115. unsigned int out_slow_tot;
  116. unsigned int out_slow_mc;
  117. unsigned int gc_total;
  118. unsigned int gc_ignored;
  119. unsigned int gc_goal_miss;
  120. unsigned int gc_dst_overflow;
  121. unsigned int in_hlist_search;
  122. unsigned int out_hlist_search;
  123. };
  124. extern struct ip_rt_acct __percpu *ip_rt_acct;
  125. struct in_device;
  126. extern int ip_rt_init(void);
  127. extern void ip_rt_redirect(__be32 old_gw, __be32 dst, __be32 new_gw,
  128. __be32 src, struct net_device *dev);
  129. extern void rt_cache_flush(struct net *net, int how);
  130. extern void rt_cache_flush_batch(struct net *net);
  131. extern struct rtable *__ip_route_output_key(struct net *, struct flowi4 *flp);
  132. extern struct rtable *ip_route_output_flow(struct net *, struct flowi4 *flp,
  133. struct sock *sk);
  134. extern struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_orig);
  135. static inline struct rtable *ip_route_output_key(struct net *net, struct flowi4 *flp)
  136. {
  137. return ip_route_output_flow(net, flp, NULL);
  138. }
  139. static inline struct rtable *ip_route_output(struct net *net, __be32 daddr,
  140. __be32 saddr, u8 tos, int oif)
  141. {
  142. struct flowi4 fl4 = {
  143. .flowi4_oif = oif,
  144. .flowi4_tos = tos,
  145. .daddr = daddr,
  146. .saddr = saddr,
  147. };
  148. return ip_route_output_key(net, &fl4);
  149. }
  150. static inline struct rtable *ip_route_output_ports(struct net *net, struct flowi4 *fl4,
  151. struct sock *sk,
  152. __be32 daddr, __be32 saddr,
  153. __be16 dport, __be16 sport,
  154. __u8 proto, __u8 tos, int oif)
  155. {
  156. flowi4_init_output(fl4, oif, sk ? sk->sk_mark : 0, tos,
  157. RT_SCOPE_UNIVERSE, proto,
  158. sk ? inet_sk_flowi_flags(sk) : 0,
  159. daddr, saddr, dport, sport);
  160. if (sk)
  161. security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
  162. return ip_route_output_flow(net, fl4, sk);
  163. }
  164. static inline struct rtable *ip_route_output_gre(struct net *net, struct flowi4 *fl4,
  165. __be32 daddr, __be32 saddr,
  166. __be32 gre_key, __u8 tos, int oif)
  167. {
  168. memset(fl4, 0, sizeof(*fl4));
  169. fl4->flowi4_oif = oif;
  170. fl4->daddr = daddr;
  171. fl4->saddr = saddr;
  172. fl4->flowi4_tos = tos;
  173. fl4->flowi4_proto = IPPROTO_GRE;
  174. fl4->fl4_gre_key = gre_key;
  175. return ip_route_output_key(net, fl4);
  176. }
  177. extern int ip_route_input_common(struct sk_buff *skb, __be32 dst, __be32 src,
  178. u8 tos, struct net_device *devin, bool noref, bool nocache);
  179. static inline int ip_route_input(struct sk_buff *skb, __be32 dst, __be32 src,
  180. u8 tos, struct net_device *devin)
  181. {
  182. return ip_route_input_common(skb, dst, src, tos, devin, false, false);
  183. }
  184. static inline int ip_route_input_noref(struct sk_buff *skb, __be32 dst, __be32 src,
  185. u8 tos, struct net_device *devin, bool nocache)
  186. {
  187. return ip_route_input_common(skb, dst, src, tos, devin, true, nocache);
  188. }
  189. extern void ipv4_update_pmtu(struct sk_buff *skb, struct net *net, u32 mtu,
  190. int oif, u32 mark, u8 protocol, int flow_flags);
  191. extern void ipv4_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, u32 mtu);
  192. extern void ip_rt_send_redirect(struct sk_buff *skb);
  193. extern unsigned int inet_addr_type(struct net *net, __be32 addr);
  194. extern unsigned int inet_dev_addr_type(struct net *net, const struct net_device *dev, __be32 addr);
  195. extern void ip_rt_multicast_event(struct in_device *);
  196. extern int ip_rt_ioctl(struct net *, unsigned int cmd, void __user *arg);
  197. extern void ip_rt_get_source(u8 *src, struct sk_buff *skb, struct rtable *rt);
  198. extern int ip_rt_dump(struct sk_buff *skb, struct netlink_callback *cb);
  199. struct in_ifaddr;
  200. extern void fib_add_ifaddr(struct in_ifaddr *);
  201. extern void fib_del_ifaddr(struct in_ifaddr *, struct in_ifaddr *);
  202. static inline void ip_rt_put(struct rtable * rt)
  203. {
  204. if (rt)
  205. dst_release(&rt->dst);
  206. }
  207. #define IPTOS_RT_MASK (IPTOS_TOS_MASK & ~3)
  208. extern const __u8 ip_tos2prio[16];
  209. static inline char rt_tos2priority(u8 tos)
  210. {
  211. return ip_tos2prio[IPTOS_TOS(tos)>>1];
  212. }
  213. /* ip_route_connect() and ip_route_newports() work in tandem whilst
  214. * binding a socket for a new outgoing connection.
  215. *
  216. * In order to use IPSEC properly, we must, in the end, have a
  217. * route that was looked up using all available keys including source
  218. * and destination ports.
  219. *
  220. * However, if a source port needs to be allocated (the user specified
  221. * a wildcard source port) we need to obtain addressing information
  222. * in order to perform that allocation.
  223. *
  224. * So ip_route_connect() looks up a route using wildcarded source and
  225. * destination ports in the key, simply so that we can get a pair of
  226. * addresses to use for port allocation.
  227. *
  228. * Later, once the ports are allocated, ip_route_newports() will make
  229. * another route lookup if needed to make sure we catch any IPSEC
  230. * rules keyed on the port information.
  231. *
  232. * The callers allocate the flow key on their stack, and must pass in
  233. * the same flowi4 object to both the ip_route_connect() and the
  234. * ip_route_newports() calls.
  235. */
  236. static inline void ip_route_connect_init(struct flowi4 *fl4, __be32 dst, __be32 src,
  237. u32 tos, int oif, u8 protocol,
  238. __be16 sport, __be16 dport,
  239. struct sock *sk, bool can_sleep)
  240. {
  241. __u8 flow_flags = 0;
  242. if (inet_sk(sk)->transparent)
  243. flow_flags |= FLOWI_FLAG_ANYSRC;
  244. if (protocol == IPPROTO_TCP)
  245. flow_flags |= FLOWI_FLAG_PRECOW_METRICS;
  246. if (can_sleep)
  247. flow_flags |= FLOWI_FLAG_CAN_SLEEP;
  248. flowi4_init_output(fl4, oif, sk->sk_mark, tos, RT_SCOPE_UNIVERSE,
  249. protocol, flow_flags, dst, src, dport, sport);
  250. }
  251. static inline struct rtable *ip_route_connect(struct flowi4 *fl4,
  252. __be32 dst, __be32 src, u32 tos,
  253. int oif, u8 protocol,
  254. __be16 sport, __be16 dport,
  255. struct sock *sk, bool can_sleep)
  256. {
  257. struct net *net = sock_net(sk);
  258. struct rtable *rt;
  259. ip_route_connect_init(fl4, dst, src, tos, oif, protocol,
  260. sport, dport, sk, can_sleep);
  261. if (!dst || !src) {
  262. rt = __ip_route_output_key(net, fl4);
  263. if (IS_ERR(rt))
  264. return rt;
  265. ip_rt_put(rt);
  266. flowi4_update_output(fl4, oif, tos, fl4->daddr, fl4->saddr);
  267. }
  268. security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
  269. return ip_route_output_flow(net, fl4, sk);
  270. }
  271. static inline struct rtable *ip_route_newports(struct flowi4 *fl4, struct rtable *rt,
  272. __be16 orig_sport, __be16 orig_dport,
  273. __be16 sport, __be16 dport,
  274. struct sock *sk)
  275. {
  276. if (sport != orig_sport || dport != orig_dport) {
  277. fl4->fl4_dport = dport;
  278. fl4->fl4_sport = sport;
  279. ip_rt_put(rt);
  280. flowi4_update_output(fl4, sk->sk_bound_dev_if,
  281. RT_CONN_FLAGS(sk), fl4->daddr,
  282. fl4->saddr);
  283. security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
  284. return ip_route_output_flow(sock_net(sk), fl4, sk);
  285. }
  286. return rt;
  287. }
  288. extern void rt_bind_peer(struct rtable *rt, __be32 daddr, int create);
  289. static inline struct inet_peer *__rt_get_peer(struct rtable *rt, __be32 daddr, int create)
  290. {
  291. if (rt_has_peer(rt))
  292. return rt_peer_ptr(rt);
  293. rt_bind_peer(rt, daddr, create);
  294. return (rt_has_peer(rt) ? rt_peer_ptr(rt) : NULL);
  295. }
  296. static inline struct inet_peer *rt_get_peer(struct rtable *rt, __be32 daddr)
  297. {
  298. return __rt_get_peer(rt, daddr, 0);
  299. }
  300. static inline struct inet_peer *rt_get_peer_create(struct rtable *rt, __be32 daddr)
  301. {
  302. return __rt_get_peer(rt, daddr, 1);
  303. }
  304. static inline int inet_iif(const struct sk_buff *skb)
  305. {
  306. return skb_rtable(skb)->rt_iif;
  307. }
  308. extern int sysctl_ip_default_ttl;
  309. static inline int ip4_dst_hoplimit(const struct dst_entry *dst)
  310. {
  311. int hoplimit = dst_metric_raw(dst, RTAX_HOPLIMIT);
  312. if (hoplimit == 0)
  313. hoplimit = sysctl_ip_default_ttl;
  314. return hoplimit;
  315. }
  316. #endif /* _ROUTE_H */