ip.h 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486
  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/snmp.h>
  30. #include <net/flow.h>
  31. struct sock;
  32. struct inet_skb_parm {
  33. struct ip_options opt; /* Compiled IP options */
  34. unsigned char flags;
  35. #define IPSKB_FORWARDED BIT(0)
  36. #define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
  37. #define IPSKB_XFRM_TRANSFORMED BIT(2)
  38. #define IPSKB_FRAG_COMPLETE BIT(3)
  39. #define IPSKB_REROUTED BIT(4)
  40. #define IPSKB_DOREDIRECT BIT(5)
  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. };
  53. #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
  54. struct ip_ra_chain {
  55. struct ip_ra_chain __rcu *next;
  56. struct sock *sk;
  57. union {
  58. void (*destructor)(struct sock *);
  59. struct sock *saved_sk;
  60. };
  61. struct rcu_head rcu;
  62. };
  63. extern struct ip_ra_chain __rcu *ip_ra_chain;
  64. /* IP flags. */
  65. #define IP_CE 0x8000 /* Flag: "Congestion" */
  66. #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
  67. #define IP_MF 0x2000 /* Flag: "More Fragments" */
  68. #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
  69. #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
  70. struct msghdr;
  71. struct net_device;
  72. struct packet_type;
  73. struct rtable;
  74. struct sockaddr;
  75. extern int igmp_mc_proc_init(void);
  76. /*
  77. * Functions provided by ip.c
  78. */
  79. extern int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
  80. __be32 saddr, __be32 daddr,
  81. struct ip_options_rcu *opt);
  82. extern int ip_rcv(struct sk_buff *skb, struct net_device *dev,
  83. struct packet_type *pt, struct net_device *orig_dev);
  84. extern int ip_local_deliver(struct sk_buff *skb);
  85. extern int ip_mr_input(struct sk_buff *skb);
  86. extern int ip_output(struct sk_buff *skb);
  87. extern int ip_mc_output(struct sk_buff *skb);
  88. extern int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
  89. extern int ip_do_nat(struct sk_buff *skb);
  90. extern void ip_send_check(struct iphdr *ip);
  91. extern int __ip_local_out(struct sk_buff *skb);
  92. extern int ip_local_out(struct sk_buff *skb);
  93. extern int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
  94. extern void ip_init(void);
  95. extern int ip_append_data(struct sock *sk, struct flowi4 *fl4,
  96. int getfrag(void *from, char *to, int offset, int len,
  97. int odd, struct sk_buff *skb),
  98. void *from, int len, int protolen,
  99. struct ipcm_cookie *ipc,
  100. struct rtable **rt,
  101. unsigned int flags);
  102. extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb);
  103. extern ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
  104. int offset, size_t size, int flags);
  105. extern struct sk_buff *__ip_make_skb(struct sock *sk,
  106. struct flowi4 *fl4,
  107. struct sk_buff_head *queue,
  108. struct inet_cork *cork);
  109. extern int ip_send_skb(struct sk_buff *skb);
  110. extern int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
  111. extern void ip_flush_pending_frames(struct sock *sk);
  112. extern struct sk_buff *ip_make_skb(struct sock *sk,
  113. struct flowi4 *fl4,
  114. int getfrag(void *from, char *to, int offset, int len,
  115. int odd, struct sk_buff *skb),
  116. void *from, int length, int transhdrlen,
  117. struct ipcm_cookie *ipc,
  118. struct rtable **rtp,
  119. unsigned int flags);
  120. static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
  121. {
  122. return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
  123. }
  124. /* datagram.c */
  125. extern int ip4_datagram_connect(struct sock *sk,
  126. struct sockaddr *uaddr, int addr_len);
  127. /*
  128. * Map a multicast IP onto multicast MAC for type Token Ring.
  129. * This conforms to RFC1469 Option 2 Multicasting i.e.
  130. * using a functional address to transmit / receive
  131. * multicast packets.
  132. */
  133. static inline void ip_tr_mc_map(__be32 addr, char *buf)
  134. {
  135. buf[0]=0xC0;
  136. buf[1]=0x00;
  137. buf[2]=0x00;
  138. buf[3]=0x04;
  139. buf[4]=0x00;
  140. buf[5]=0x00;
  141. }
  142. struct ip_reply_arg {
  143. struct kvec iov[1];
  144. int flags;
  145. __wsum csum;
  146. int csumoffset; /* u16 offset of csum in iov[0].iov_base */
  147. /* -1 if not needed */
  148. int bound_dev_if;
  149. u8 tos;
  150. kuid_t uid;
  151. };
  152. #define IP_REPLY_ARG_NOSRCCHECK 1
  153. static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
  154. {
  155. return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
  156. }
  157. void ip_send_reply(struct sock *sk, struct sk_buff *skb, __be32 daddr,
  158. const struct ip_reply_arg *arg, unsigned int len);
  159. struct ipv4_config {
  160. int log_martians;
  161. int no_pmtu_disc;
  162. };
  163. extern struct ipv4_config ipv4_config;
  164. #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
  165. #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
  166. #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
  167. #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
  168. #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
  169. #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
  170. #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
  171. #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
  172. #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
  173. #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
  174. #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
  175. extern unsigned long snmp_fold_field(void __percpu *mib[], int offt);
  176. #if BITS_PER_LONG==32
  177. extern u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
  178. #else
  179. static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
  180. {
  181. return snmp_fold_field(mib, offt);
  182. }
  183. #endif
  184. extern int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
  185. extern void snmp_mib_free(void __percpu *ptr[2]);
  186. extern struct local_ports {
  187. seqlock_t lock;
  188. int range[2];
  189. } sysctl_local_ports;
  190. extern void inet_get_local_port_range(int *low, int *high);
  191. extern unsigned long *sysctl_local_reserved_ports;
  192. static inline int inet_is_reserved_local_port(int port)
  193. {
  194. return test_bit(port, sysctl_local_reserved_ports);
  195. }
  196. extern int sysctl_ip_nonlocal_bind;
  197. extern struct ctl_path net_core_path[];
  198. extern struct ctl_path net_ipv4_ctl_path[];
  199. /* From inetpeer.c */
  200. extern int inet_peer_threshold;
  201. extern int inet_peer_minttl;
  202. extern int inet_peer_maxttl;
  203. /* From ip_output.c */
  204. extern int sysctl_ip_dynaddr;
  205. extern void ipfrag_init(void);
  206. extern void ip_static_sysctl_init(void);
  207. #define IP4_REPLY_MARK(net, mark) \
  208. ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
  209. static inline bool ip_is_fragment(const struct iphdr *iph)
  210. {
  211. return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
  212. }
  213. #ifdef CONFIG_INET
  214. #include <net/dst.h>
  215. /* The function in 2.2 was invalid, producing wrong result for
  216. * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
  217. static inline
  218. int ip_decrease_ttl(struct iphdr *iph)
  219. {
  220. u32 check = (__force u32)iph->check;
  221. check += (__force u32)htons(0x0100);
  222. iph->check = (__force __sum16)(check + (check>=0xFFFF));
  223. return --iph->ttl;
  224. }
  225. static inline
  226. int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
  227. {
  228. return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
  229. (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
  230. !(dst_metric_locked(dst, RTAX_MTU)));
  231. }
  232. u32 ip_idents_reserve(u32 hash, int segs);
  233. void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
  234. static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb,
  235. struct sock *sk, int segs)
  236. {
  237. struct iphdr *iph = ip_hdr(skb);
  238. if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
  239. /* This is only to work around buggy Windows95/2000
  240. * VJ compression implementations. If the ID field
  241. * does not change, they drop every other packet in
  242. * a TCP stream using header compression.
  243. */
  244. if (sk && inet_sk(sk)->inet_daddr) {
  245. iph->id = htons(inet_sk(sk)->inet_id);
  246. inet_sk(sk)->inet_id += segs;
  247. } else {
  248. iph->id = 0;
  249. }
  250. } else {
  251. __ip_select_ident(net, iph, segs);
  252. }
  253. }
  254. static inline void ip_select_ident(struct net *net, struct sk_buff *skb,
  255. struct sock *sk)
  256. {
  257. ip_select_ident_segs(net, skb, sk, 1);
  258. }
  259. /*
  260. * Map a multicast IP onto multicast MAC for type ethernet.
  261. */
  262. static inline void ip_eth_mc_map(__be32 naddr, char *buf)
  263. {
  264. __u32 addr=ntohl(naddr);
  265. buf[0]=0x01;
  266. buf[1]=0x00;
  267. buf[2]=0x5e;
  268. buf[5]=addr&0xFF;
  269. addr>>=8;
  270. buf[4]=addr&0xFF;
  271. addr>>=8;
  272. buf[3]=addr&0x7F;
  273. }
  274. /*
  275. * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
  276. * Leave P_Key as 0 to be filled in by driver.
  277. */
  278. static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
  279. {
  280. __u32 addr;
  281. unsigned char scope = broadcast[5] & 0xF;
  282. buf[0] = 0; /* Reserved */
  283. buf[1] = 0xff; /* Multicast QPN */
  284. buf[2] = 0xff;
  285. buf[3] = 0xff;
  286. addr = ntohl(naddr);
  287. buf[4] = 0xff;
  288. buf[5] = 0x10 | scope; /* scope from broadcast address */
  289. buf[6] = 0x40; /* IPv4 signature */
  290. buf[7] = 0x1b;
  291. buf[8] = broadcast[8]; /* P_Key */
  292. buf[9] = broadcast[9];
  293. buf[10] = 0;
  294. buf[11] = 0;
  295. buf[12] = 0;
  296. buf[13] = 0;
  297. buf[14] = 0;
  298. buf[15] = 0;
  299. buf[19] = addr & 0xff;
  300. addr >>= 8;
  301. buf[18] = addr & 0xff;
  302. addr >>= 8;
  303. buf[17] = addr & 0xff;
  304. addr >>= 8;
  305. buf[16] = addr & 0x0f;
  306. }
  307. static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
  308. {
  309. if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
  310. memcpy(buf, broadcast, 4);
  311. else
  312. memcpy(buf, &naddr, sizeof(naddr));
  313. }
  314. #if IS_ENABLED(CONFIG_IPV6)
  315. #include <linux/ipv6.h>
  316. #endif
  317. static __inline__ void inet_reset_saddr(struct sock *sk)
  318. {
  319. inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
  320. #if IS_ENABLED(CONFIG_IPV6)
  321. if (sk->sk_family == PF_INET6) {
  322. struct ipv6_pinfo *np = inet6_sk(sk);
  323. memset(&np->saddr, 0, sizeof(np->saddr));
  324. memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
  325. }
  326. #endif
  327. }
  328. #endif
  329. static inline int sk_mc_loop(struct sock *sk)
  330. {
  331. if (!sk)
  332. return 1;
  333. switch (sk->sk_family) {
  334. case AF_INET:
  335. return inet_sk(sk)->mc_loop;
  336. #if IS_ENABLED(CONFIG_IPV6)
  337. case AF_INET6:
  338. return inet6_sk(sk)->mc_loop;
  339. #endif
  340. }
  341. WARN_ON(1);
  342. return 1;
  343. }
  344. extern bool ip_call_ra_chain(struct sk_buff *skb);
  345. /*
  346. * Functions provided by ip_fragment.c
  347. */
  348. enum ip_defrag_users {
  349. IP_DEFRAG_LOCAL_DELIVER,
  350. IP_DEFRAG_CALL_RA_CHAIN,
  351. IP_DEFRAG_CONNTRACK_IN,
  352. __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
  353. IP_DEFRAG_CONNTRACK_OUT,
  354. __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
  355. IP_DEFRAG_CONNTRACK_BRIDGE_IN,
  356. __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
  357. IP_DEFRAG_VS_IN,
  358. IP_DEFRAG_VS_OUT,
  359. IP_DEFRAG_VS_FWD,
  360. IP_DEFRAG_AF_PACKET,
  361. IP_DEFRAG_MACVLAN,
  362. };
  363. int ip_defrag(struct sk_buff *skb, u32 user);
  364. #ifdef CONFIG_INET
  365. struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
  366. #else
  367. static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
  368. {
  369. return skb;
  370. }
  371. #endif
  372. int ip_frag_mem(struct net *net);
  373. int ip_frag_nqueues(struct net *net);
  374. /*
  375. * Functions provided by ip_forward.c
  376. */
  377. extern int ip_forward(struct sk_buff *skb);
  378. /*
  379. * Functions provided by ip_options.c
  380. */
  381. extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
  382. __be32 daddr, struct rtable *rt, int is_frag);
  383. extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
  384. extern void ip_options_fragment(struct sk_buff *skb);
  385. extern int ip_options_compile(struct net *net,
  386. struct ip_options *opt, struct sk_buff *skb);
  387. extern int ip_options_get(struct net *net, struct ip_options_rcu **optp,
  388. unsigned char *data, int optlen);
  389. extern int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
  390. unsigned char __user *data, int optlen);
  391. extern void ip_options_undo(struct ip_options * opt);
  392. extern void ip_forward_options(struct sk_buff *skb);
  393. extern int ip_options_rcv_srr(struct sk_buff *skb);
  394. /*
  395. * Functions provided by ip_sockglue.c
  396. */
  397. extern void ipv4_pktinfo_prepare(struct sk_buff *skb);
  398. extern void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
  399. extern int ip_cmsg_send(struct net *net,
  400. struct msghdr *msg, struct ipcm_cookie *ipc);
  401. extern int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen);
  402. extern int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);
  403. extern int compat_ip_setsockopt(struct sock *sk, int level,
  404. int optname, char __user *optval, unsigned int optlen);
  405. extern int compat_ip_getsockopt(struct sock *sk, int level,
  406. int optname, char __user *optval, int __user *optlen);
  407. extern int ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *));
  408. extern int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
  409. extern void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
  410. __be16 port, u32 info, u8 *payload);
  411. extern void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
  412. u32 info);
  413. #ifdef CONFIG_PROC_FS
  414. extern int ip_misc_proc_init(void);
  415. #endif
  416. #endif /* _IP_H */