ip.h 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523
  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 module.
  7. *
  8. * Version: @(#)ip.h 1.0.2 05/07/93
  9. *
  10. * Authors: Ross Biro
  11. * Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
  12. * Alan Cox, <gw4pts@gw4pts.ampr.org>
  13. *
  14. * Changes:
  15. * Mike McLagan : Routing by source
  16. *
  17. * This program is free software; you can redistribute it and/or
  18. * modify it under the terms of the GNU General Public License
  19. * as published by the Free Software Foundation; either version
  20. * 2 of the License, or (at your option) any later version.
  21. */
  22. #ifndef _IP_H
  23. #define _IP_H
  24. #include <linux/types.h>
  25. #include <linux/ip.h>
  26. #include <linux/in.h>
  27. #include <linux/skbuff.h>
  28. #include <net/inet_sock.h>
  29. #include <net/route.h>
  30. #include <net/snmp.h>
  31. #include <net/flow.h>
  32. struct sock;
  33. struct inet_skb_parm {
  34. struct ip_options opt; /* Compiled IP options */
  35. unsigned char flags;
  36. #define IPSKB_FORWARDED 1
  37. #define IPSKB_XFRM_TUNNEL_SIZE 2
  38. #define IPSKB_XFRM_TRANSFORMED 4
  39. #define IPSKB_FRAG_COMPLETE 8
  40. #define IPSKB_REROUTED 16
  41. u16 frag_max_size;
  42. };
  43. static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
  44. {
  45. return ip_hdr(skb)->ihl * 4;
  46. }
  47. struct ipcm_cookie {
  48. __be32 addr;
  49. int oif;
  50. struct ip_options_rcu *opt;
  51. __u8 tx_flags;
  52. __u8 ttl;
  53. __s16 tos;
  54. char priority;
  55. };
  56. #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
  57. #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
  58. struct ip_ra_chain {
  59. struct ip_ra_chain __rcu *next;
  60. struct sock *sk;
  61. union {
  62. void (*destructor)(struct sock *);
  63. struct sock *saved_sk;
  64. };
  65. struct rcu_head rcu;
  66. };
  67. extern struct ip_ra_chain __rcu *ip_ra_chain;
  68. /* IP flags. */
  69. #define IP_CE 0x8000 /* Flag: "Congestion" */
  70. #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
  71. #define IP_MF 0x2000 /* Flag: "More Fragments" */
  72. #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
  73. #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
  74. struct msghdr;
  75. struct net_device;
  76. struct packet_type;
  77. struct rtable;
  78. struct sockaddr;
  79. int igmp_mc_init(void);
  80. /*
  81. * Functions provided by ip.c
  82. */
  83. int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
  84. __be32 saddr, __be32 daddr,
  85. struct ip_options_rcu *opt);
  86. int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
  87. struct net_device *orig_dev);
  88. int ip_local_deliver(struct sk_buff *skb);
  89. int ip_mr_input(struct sk_buff *skb);
  90. int ip_output(struct sk_buff *skb);
  91. int ip_mc_output(struct sk_buff *skb);
  92. int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
  93. int ip_do_nat(struct sk_buff *skb);
  94. void ip_send_check(struct iphdr *ip);
  95. int __ip_local_out(struct sk_buff *skb);
  96. int ip_local_out(struct sk_buff *skb);
  97. int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
  98. void ip_init(void);
  99. int ip_append_data(struct sock *sk, struct flowi4 *fl4,
  100. int getfrag(void *from, char *to, int offset, int len,
  101. int odd, struct sk_buff *skb),
  102. void *from, int len, int protolen,
  103. struct ipcm_cookie *ipc,
  104. struct rtable **rt,
  105. unsigned int flags);
  106. int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
  107. struct sk_buff *skb);
  108. ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
  109. int offset, size_t size, int flags);
  110. struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
  111. struct sk_buff_head *queue,
  112. struct inet_cork *cork);
  113. int ip_send_skb(struct net *net, struct sk_buff *skb);
  114. int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
  115. void ip_flush_pending_frames(struct sock *sk);
  116. struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
  117. int getfrag(void *from, char *to, int offset,
  118. int len, int odd, struct sk_buff *skb),
  119. void *from, int length, int transhdrlen,
  120. struct ipcm_cookie *ipc, struct rtable **rtp,
  121. unsigned int flags);
  122. static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
  123. {
  124. return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
  125. }
  126. static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
  127. {
  128. return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
  129. }
  130. static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
  131. {
  132. return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
  133. }
  134. /* datagram.c */
  135. int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
  136. void ip4_datagram_release_cb(struct sock *sk);
  137. struct ip_reply_arg {
  138. struct kvec iov[1];
  139. int flags;
  140. __wsum csum;
  141. int csumoffset; /* u16 offset of csum in iov[0].iov_base */
  142. /* -1 if not needed */
  143. int bound_dev_if;
  144. u8 tos;
  145. };
  146. #define IP_REPLY_ARG_NOSRCCHECK 1
  147. static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
  148. {
  149. return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
  150. }
  151. void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr,
  152. __be32 saddr, const struct ip_reply_arg *arg,
  153. unsigned int len);
  154. #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
  155. #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
  156. #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
  157. #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
  158. #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
  159. #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
  160. #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
  161. #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
  162. #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
  163. #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
  164. #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
  165. #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
  166. unsigned long snmp_fold_field(void __percpu *mib[], int offt);
  167. #if BITS_PER_LONG==32
  168. u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
  169. #else
  170. static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
  171. {
  172. return snmp_fold_field(mib, offt);
  173. }
  174. #endif
  175. int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
  176. static inline void snmp_mib_free(void __percpu *ptr[SNMP_ARRAY_SZ])
  177. {
  178. int i;
  179. BUG_ON(ptr == NULL);
  180. for (i = 0; i < SNMP_ARRAY_SZ; i++) {
  181. free_percpu(ptr[i]);
  182. ptr[i] = NULL;
  183. }
  184. }
  185. void inet_get_local_port_range(struct net *net, int *low, int *high);
  186. extern unsigned long *sysctl_local_reserved_ports;
  187. static inline int inet_is_reserved_local_port(int port)
  188. {
  189. return test_bit(port, sysctl_local_reserved_ports);
  190. }
  191. extern int sysctl_ip_nonlocal_bind;
  192. /* From inetpeer.c */
  193. extern int inet_peer_threshold;
  194. extern int inet_peer_minttl;
  195. extern int inet_peer_maxttl;
  196. /* From ip_input.c */
  197. extern int sysctl_ip_early_demux;
  198. /* From ip_output.c */
  199. extern int sysctl_ip_dynaddr;
  200. void ipfrag_init(void);
  201. void ip_static_sysctl_init(void);
  202. static inline bool ip_is_fragment(const struct iphdr *iph)
  203. {
  204. return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
  205. }
  206. #ifdef CONFIG_INET
  207. #include <net/dst.h>
  208. /* The function in 2.2 was invalid, producing wrong result for
  209. * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
  210. static inline
  211. int ip_decrease_ttl(struct iphdr *iph)
  212. {
  213. u32 check = (__force u32)iph->check;
  214. check += (__force u32)htons(0x0100);
  215. iph->check = (__force __sum16)(check + (check>=0xFFFF));
  216. return --iph->ttl;
  217. }
  218. static inline
  219. int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
  220. {
  221. return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
  222. (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
  223. !(dst_metric_locked(dst, RTAX_MTU)));
  224. }
  225. static inline bool ip_sk_accept_pmtu(const struct sock *sk)
  226. {
  227. return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
  228. inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
  229. }
  230. static inline bool ip_sk_use_pmtu(const struct sock *sk)
  231. {
  232. return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
  233. }
  234. static inline bool ip_sk_local_df(const struct sock *sk)
  235. {
  236. return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
  237. inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
  238. }
  239. static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
  240. bool forwarding)
  241. {
  242. struct net *net = dev_net(dst->dev);
  243. if (net->ipv4.sysctl_ip_fwd_use_pmtu ||
  244. dst_metric_locked(dst, RTAX_MTU) ||
  245. !forwarding)
  246. return dst_mtu(dst);
  247. return min(dst->dev->mtu, IP_MAX_MTU);
  248. }
  249. static inline unsigned int ip_skb_dst_mtu(const struct sk_buff *skb)
  250. {
  251. if (!skb->sk || ip_sk_use_pmtu(skb->sk)) {
  252. bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
  253. return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
  254. } else {
  255. return min(skb_dst(skb)->dev->mtu, IP_MAX_MTU);
  256. }
  257. }
  258. void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
  259. static inline void ip_select_ident(struct sk_buff *skb, struct dst_entry *dst, struct sock *sk)
  260. {
  261. struct iphdr *iph = ip_hdr(skb);
  262. if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
  263. /* This is only to work around buggy Windows95/2000
  264. * VJ compression implementations. If the ID field
  265. * does not change, they drop every other packet in
  266. * a TCP stream using header compression.
  267. */
  268. iph->id = (sk && inet_sk(sk)->inet_daddr) ?
  269. htons(inet_sk(sk)->inet_id++) : 0;
  270. } else
  271. __ip_select_ident(iph, dst, 0);
  272. }
  273. static inline void ip_select_ident_more(struct sk_buff *skb, struct dst_entry *dst, struct sock *sk, int more)
  274. {
  275. struct iphdr *iph = ip_hdr(skb);
  276. if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
  277. if (sk && inet_sk(sk)->inet_daddr) {
  278. iph->id = htons(inet_sk(sk)->inet_id);
  279. inet_sk(sk)->inet_id += 1 + more;
  280. } else
  281. iph->id = 0;
  282. } else
  283. __ip_select_ident(iph, dst, more);
  284. }
  285. /*
  286. * Map a multicast IP onto multicast MAC for type ethernet.
  287. */
  288. static inline void ip_eth_mc_map(__be32 naddr, char *buf)
  289. {
  290. __u32 addr=ntohl(naddr);
  291. buf[0]=0x01;
  292. buf[1]=0x00;
  293. buf[2]=0x5e;
  294. buf[5]=addr&0xFF;
  295. addr>>=8;
  296. buf[4]=addr&0xFF;
  297. addr>>=8;
  298. buf[3]=addr&0x7F;
  299. }
  300. /*
  301. * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
  302. * Leave P_Key as 0 to be filled in by driver.
  303. */
  304. static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
  305. {
  306. __u32 addr;
  307. unsigned char scope = broadcast[5] & 0xF;
  308. buf[0] = 0; /* Reserved */
  309. buf[1] = 0xff; /* Multicast QPN */
  310. buf[2] = 0xff;
  311. buf[3] = 0xff;
  312. addr = ntohl(naddr);
  313. buf[4] = 0xff;
  314. buf[5] = 0x10 | scope; /* scope from broadcast address */
  315. buf[6] = 0x40; /* IPv4 signature */
  316. buf[7] = 0x1b;
  317. buf[8] = broadcast[8]; /* P_Key */
  318. buf[9] = broadcast[9];
  319. buf[10] = 0;
  320. buf[11] = 0;
  321. buf[12] = 0;
  322. buf[13] = 0;
  323. buf[14] = 0;
  324. buf[15] = 0;
  325. buf[19] = addr & 0xff;
  326. addr >>= 8;
  327. buf[18] = addr & 0xff;
  328. addr >>= 8;
  329. buf[17] = addr & 0xff;
  330. addr >>= 8;
  331. buf[16] = addr & 0x0f;
  332. }
  333. static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
  334. {
  335. if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
  336. memcpy(buf, broadcast, 4);
  337. else
  338. memcpy(buf, &naddr, sizeof(naddr));
  339. }
  340. #if IS_ENABLED(CONFIG_IPV6)
  341. #include <linux/ipv6.h>
  342. #endif
  343. static __inline__ void inet_reset_saddr(struct sock *sk)
  344. {
  345. inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
  346. #if IS_ENABLED(CONFIG_IPV6)
  347. if (sk->sk_family == PF_INET6) {
  348. struct ipv6_pinfo *np = inet6_sk(sk);
  349. memset(&np->saddr, 0, sizeof(np->saddr));
  350. memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
  351. }
  352. #endif
  353. }
  354. #endif
  355. static inline int sk_mc_loop(struct sock *sk)
  356. {
  357. if (!sk)
  358. return 1;
  359. switch (sk->sk_family) {
  360. case AF_INET:
  361. return inet_sk(sk)->mc_loop;
  362. #if IS_ENABLED(CONFIG_IPV6)
  363. case AF_INET6:
  364. return inet6_sk(sk)->mc_loop;
  365. #endif
  366. }
  367. WARN_ON(1);
  368. return 1;
  369. }
  370. bool ip_call_ra_chain(struct sk_buff *skb);
  371. /*
  372. * Functions provided by ip_fragment.c
  373. */
  374. enum ip_defrag_users {
  375. IP_DEFRAG_LOCAL_DELIVER,
  376. IP_DEFRAG_CALL_RA_CHAIN,
  377. IP_DEFRAG_CONNTRACK_IN,
  378. __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
  379. IP_DEFRAG_CONNTRACK_OUT,
  380. __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
  381. IP_DEFRAG_CONNTRACK_BRIDGE_IN,
  382. __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
  383. IP_DEFRAG_VS_IN,
  384. IP_DEFRAG_VS_OUT,
  385. IP_DEFRAG_VS_FWD,
  386. IP_DEFRAG_AF_PACKET,
  387. IP_DEFRAG_MACVLAN,
  388. };
  389. int ip_defrag(struct sk_buff *skb, u32 user);
  390. #ifdef CONFIG_INET
  391. struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
  392. #else
  393. static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
  394. {
  395. return skb;
  396. }
  397. #endif
  398. int ip_frag_mem(struct net *net);
  399. int ip_frag_nqueues(struct net *net);
  400. /*
  401. * Functions provided by ip_forward.c
  402. */
  403. int ip_forward(struct sk_buff *skb);
  404. /*
  405. * Functions provided by ip_options.c
  406. */
  407. void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
  408. __be32 daddr, struct rtable *rt, int is_frag);
  409. int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
  410. void ip_options_fragment(struct sk_buff *skb);
  411. int ip_options_compile(struct net *net, struct ip_options *opt,
  412. struct sk_buff *skb);
  413. int ip_options_get(struct net *net, struct ip_options_rcu **optp,
  414. unsigned char *data, int optlen);
  415. int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
  416. unsigned char __user *data, int optlen);
  417. void ip_options_undo(struct ip_options *opt);
  418. void ip_forward_options(struct sk_buff *skb);
  419. int ip_options_rcv_srr(struct sk_buff *skb);
  420. /*
  421. * Functions provided by ip_sockglue.c
  422. */
  423. void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
  424. void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
  425. int ip_cmsg_send(struct net *net, struct msghdr *msg,
  426. struct ipcm_cookie *ipc, bool allow_ipv6);
  427. int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
  428. unsigned int optlen);
  429. int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
  430. int __user *optlen);
  431. int compat_ip_setsockopt(struct sock *sk, int level, int optname,
  432. char __user *optval, unsigned int optlen);
  433. int compat_ip_getsockopt(struct sock *sk, int level, int optname,
  434. char __user *optval, int __user *optlen);
  435. int ip_ra_control(struct sock *sk, unsigned char on,
  436. void (*destructor)(struct sock *));
  437. int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
  438. void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
  439. u32 info, u8 *payload);
  440. void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
  441. u32 info);
  442. #ifdef CONFIG_PROC_FS
  443. int ip_misc_proc_init(void);
  444. #endif
  445. #endif /* _IP_H */