ip.h 23 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797
  1. /* SPDX-License-Identifier: GPL-2.0-or-later */
  2. /*
  3. * INET An implementation of the TCP/IP protocol suite for the LINUX
  4. * operating system. INET is implemented using the BSD Socket
  5. * interface as the means of communication with the user level.
  6. *
  7. * Definitions for the IP module.
  8. *
  9. * Version: @(#)ip.h 1.0.2 05/07/93
  10. *
  11. * Authors: Ross Biro
  12. * Fred N. van Kempen, <[email protected]>
  13. * Alan Cox, <[email protected]>
  14. *
  15. * Changes:
  16. * Mike McLagan : Routing by source
  17. */
  18. #ifndef _IP_H
  19. #define _IP_H
  20. #include <linux/types.h>
  21. #include <linux/ip.h>
  22. #include <linux/in.h>
  23. #include <linux/skbuff.h>
  24. #include <linux/jhash.h>
  25. #include <linux/sockptr.h>
  26. #include <linux/static_key.h>
  27. #include <linux/android_kabi.h>
  28. #include <net/inet_sock.h>
  29. #include <net/route.h>
  30. #include <net/snmp.h>
  31. #include <net/flow.h>
  32. #include <net/flow_dissector.h>
  33. #include <net/netns/hash.h>
  34. #include <net/lwtunnel.h>
  35. #define IPV4_MAX_PMTU 65535U /* RFC 2675, Section 5.1 */
  36. #define IPV4_MIN_MTU 68 /* RFC 791 */
  37. extern unsigned int sysctl_fib_sync_mem;
  38. extern unsigned int sysctl_fib_sync_mem_min;
  39. extern unsigned int sysctl_fib_sync_mem_max;
  40. struct sock;
  41. struct inet_skb_parm {
  42. int iif;
  43. struct ip_options opt; /* Compiled IP options */
  44. u16 flags;
  45. #define IPSKB_FORWARDED BIT(0)
  46. #define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
  47. #define IPSKB_XFRM_TRANSFORMED BIT(2)
  48. #define IPSKB_FRAG_COMPLETE BIT(3)
  49. #define IPSKB_REROUTED BIT(4)
  50. #define IPSKB_DOREDIRECT BIT(5)
  51. #define IPSKB_FRAG_PMTU BIT(6)
  52. #define IPSKB_L3SLAVE BIT(7)
  53. #define IPSKB_NOPOLICY BIT(8)
  54. #define IPSKB_MULTIPATH BIT(9)
  55. u16 frag_max_size;
  56. };
  57. static inline bool ipv4_l3mdev_skb(u16 flags)
  58. {
  59. return !!(flags & IPSKB_L3SLAVE);
  60. }
  61. static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
  62. {
  63. return ip_hdr(skb)->ihl * 4;
  64. }
  65. struct ipcm_cookie {
  66. struct sockcm_cookie sockc;
  67. __be32 addr;
  68. int oif;
  69. struct ip_options_rcu *opt;
  70. __u8 protocol;
  71. __u8 ttl;
  72. __s16 tos;
  73. char priority;
  74. __u16 gso_size;
  75. ANDROID_KABI_RESERVE(1);
  76. };
  77. static inline void ipcm_init(struct ipcm_cookie *ipcm)
  78. {
  79. *ipcm = (struct ipcm_cookie) { .tos = -1 };
  80. }
  81. static inline void ipcm_init_sk(struct ipcm_cookie *ipcm,
  82. const struct inet_sock *inet)
  83. {
  84. ipcm_init(ipcm);
  85. ipcm->sockc.mark = READ_ONCE(inet->sk.sk_mark);
  86. ipcm->sockc.tsflags = inet->sk.sk_tsflags;
  87. ipcm->oif = READ_ONCE(inet->sk.sk_bound_dev_if);
  88. ipcm->addr = inet->inet_saddr;
  89. ipcm->protocol = inet->inet_num;
  90. }
  91. #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
  92. #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
  93. /* return enslaved device index if relevant */
  94. static inline int inet_sdif(const struct sk_buff *skb)
  95. {
  96. #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
  97. if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags))
  98. return IPCB(skb)->iif;
  99. #endif
  100. return 0;
  101. }
  102. /* Special input handler for packets caught by router alert option.
  103. They are selected only by protocol field, and then processed likely
  104. local ones; but only if someone wants them! Otherwise, router
  105. not running rsvpd will kill RSVP.
  106. It is user level problem, what it will make with them.
  107. I have no idea, how it will masquearde or NAT them (it is joke, joke :-)),
  108. but receiver should be enough clever f.e. to forward mtrace requests,
  109. sent to multicast group to reach destination designated router.
  110. */
  111. struct ip_ra_chain {
  112. struct ip_ra_chain __rcu *next;
  113. struct sock *sk;
  114. union {
  115. void (*destructor)(struct sock *);
  116. struct sock *saved_sk;
  117. };
  118. struct rcu_head rcu;
  119. };
  120. /* IP flags. */
  121. #define IP_CE 0x8000 /* Flag: "Congestion" */
  122. #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
  123. #define IP_MF 0x2000 /* Flag: "More Fragments" */
  124. #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
  125. #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
  126. struct msghdr;
  127. struct net_device;
  128. struct packet_type;
  129. struct rtable;
  130. struct sockaddr;
  131. int igmp_mc_init(void);
  132. /*
  133. * Functions provided by ip.c
  134. */
  135. int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
  136. __be32 saddr, __be32 daddr,
  137. struct ip_options_rcu *opt, u8 tos);
  138. int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
  139. struct net_device *orig_dev);
  140. void ip_list_rcv(struct list_head *head, struct packet_type *pt,
  141. struct net_device *orig_dev);
  142. int ip_local_deliver(struct sk_buff *skb);
  143. void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto);
  144. int ip_mr_input(struct sk_buff *skb);
  145. int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb);
  146. int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb);
  147. int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
  148. int (*output)(struct net *, struct sock *, struct sk_buff *));
  149. struct ip_fraglist_iter {
  150. struct sk_buff *frag;
  151. struct iphdr *iph;
  152. int offset;
  153. unsigned int hlen;
  154. };
  155. void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph,
  156. unsigned int hlen, struct ip_fraglist_iter *iter);
  157. void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter);
  158. static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter)
  159. {
  160. struct sk_buff *skb = iter->frag;
  161. iter->frag = skb->next;
  162. skb_mark_not_on_list(skb);
  163. return skb;
  164. }
  165. struct ip_frag_state {
  166. bool DF;
  167. unsigned int hlen;
  168. unsigned int ll_rs;
  169. unsigned int mtu;
  170. unsigned int left;
  171. int offset;
  172. int ptr;
  173. __be16 not_last_frag;
  174. };
  175. void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs,
  176. unsigned int mtu, bool DF, struct ip_frag_state *state);
  177. struct sk_buff *ip_frag_next(struct sk_buff *skb,
  178. struct ip_frag_state *state);
  179. void ip_send_check(struct iphdr *ip);
  180. int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
  181. int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
  182. int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl,
  183. __u8 tos);
  184. void ip_init(void);
  185. int ip_append_data(struct sock *sk, struct flowi4 *fl4,
  186. int getfrag(void *from, char *to, int offset, int len,
  187. int odd, struct sk_buff *skb),
  188. void *from, int len, int protolen,
  189. struct ipcm_cookie *ipc,
  190. struct rtable **rt,
  191. unsigned int flags);
  192. int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
  193. struct sk_buff *skb);
  194. ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
  195. int offset, size_t size, int flags);
  196. struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
  197. struct sk_buff_head *queue,
  198. struct inet_cork *cork);
  199. int ip_send_skb(struct net *net, struct sk_buff *skb);
  200. int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
  201. void ip_flush_pending_frames(struct sock *sk);
  202. struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
  203. int getfrag(void *from, char *to, int offset,
  204. int len, int odd, struct sk_buff *skb),
  205. void *from, int length, int transhdrlen,
  206. struct ipcm_cookie *ipc, struct rtable **rtp,
  207. struct inet_cork *cork, unsigned int flags);
  208. int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl);
  209. static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
  210. {
  211. return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
  212. }
  213. static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
  214. {
  215. return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
  216. }
  217. static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
  218. {
  219. return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
  220. }
  221. /* datagram.c */
  222. int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
  223. int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
  224. void ip4_datagram_release_cb(struct sock *sk);
  225. struct ip_reply_arg {
  226. struct kvec iov[1];
  227. int flags;
  228. __wsum csum;
  229. int csumoffset; /* u16 offset of csum in iov[0].iov_base */
  230. /* -1 if not needed */
  231. int bound_dev_if;
  232. u8 tos;
  233. kuid_t uid;
  234. };
  235. #define IP_REPLY_ARG_NOSRCCHECK 1
  236. static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
  237. {
  238. return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
  239. }
  240. void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
  241. const struct ip_options *sopt,
  242. __be32 daddr, __be32 saddr,
  243. const struct ip_reply_arg *arg,
  244. unsigned int len, u64 transmit_time, u32 txhash);
  245. #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
  246. #define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field)
  247. #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
  248. #define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
  249. #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
  250. #define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
  251. #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
  252. #define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field)
  253. #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
  254. #define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
  255. static inline u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offt)
  256. {
  257. return *(((unsigned long *)per_cpu_ptr(mib, cpu)) + offt);
  258. }
  259. unsigned long snmp_fold_field(void __percpu *mib, int offt);
  260. #if BITS_PER_LONG==32
  261. u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
  262. size_t syncp_offset);
  263. u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
  264. #else
  265. static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
  266. size_t syncp_offset)
  267. {
  268. return snmp_get_cpu_field(mib, cpu, offct);
  269. }
  270. static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
  271. {
  272. return snmp_fold_field(mib, offt);
  273. }
  274. #endif
  275. #define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \
  276. { \
  277. int i, c; \
  278. for_each_possible_cpu(c) { \
  279. for (i = 0; stats_list[i].name; i++) \
  280. buff64[i] += snmp_get_cpu_field64( \
  281. mib_statistic, \
  282. c, stats_list[i].entry, \
  283. offset); \
  284. } \
  285. }
  286. #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
  287. { \
  288. int i, c; \
  289. for_each_possible_cpu(c) { \
  290. for (i = 0; stats_list[i].name; i++) \
  291. buff[i] += snmp_get_cpu_field( \
  292. mib_statistic, \
  293. c, stats_list[i].entry); \
  294. } \
  295. }
  296. void inet_get_local_port_range(const struct net *net, int *low, int *high);
  297. void inet_sk_get_local_port_range(const struct sock *sk, int *low, int *high);
  298. #ifdef CONFIG_SYSCTL
  299. static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
  300. {
  301. if (!net->ipv4.sysctl_local_reserved_ports)
  302. return false;
  303. return test_bit(port, net->ipv4.sysctl_local_reserved_ports);
  304. }
  305. static inline bool sysctl_dev_name_is_allowed(const char *name)
  306. {
  307. return strcmp(name, "default") != 0 && strcmp(name, "all") != 0;
  308. }
  309. static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
  310. {
  311. return port < READ_ONCE(net->ipv4.sysctl_ip_prot_sock);
  312. }
  313. #else
  314. static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
  315. {
  316. return false;
  317. }
  318. static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
  319. {
  320. return port < PROT_SOCK;
  321. }
  322. #endif
  323. __be32 inet_current_timestamp(void);
  324. /* From inetpeer.c */
  325. extern int inet_peer_threshold;
  326. extern int inet_peer_minttl;
  327. extern int inet_peer_maxttl;
  328. void ipfrag_init(void);
  329. void ip_static_sysctl_init(void);
  330. #define IP4_REPLY_MARK(net, mark) \
  331. (READ_ONCE((net)->ipv4.sysctl_fwmark_reflect) ? (mark) : 0)
  332. static inline bool ip_is_fragment(const struct iphdr *iph)
  333. {
  334. return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
  335. }
  336. #ifdef CONFIG_INET
  337. #include <net/dst.h>
  338. /* The function in 2.2 was invalid, producing wrong result for
  339. * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
  340. static inline
  341. int ip_decrease_ttl(struct iphdr *iph)
  342. {
  343. u32 check = (__force u32)iph->check;
  344. check += (__force u32)htons(0x0100);
  345. iph->check = (__force __sum16)(check + (check>=0xFFFF));
  346. return --iph->ttl;
  347. }
  348. static inline int ip_mtu_locked(const struct dst_entry *dst)
  349. {
  350. const struct rtable *rt = (const struct rtable *)dst;
  351. return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU);
  352. }
  353. static inline
  354. int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst)
  355. {
  356. u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc);
  357. return pmtudisc == IP_PMTUDISC_DO ||
  358. (pmtudisc == IP_PMTUDISC_WANT &&
  359. !ip_mtu_locked(dst));
  360. }
  361. static inline bool ip_sk_accept_pmtu(const struct sock *sk)
  362. {
  363. return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
  364. inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
  365. }
  366. static inline bool ip_sk_use_pmtu(const struct sock *sk)
  367. {
  368. return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
  369. }
  370. static inline bool ip_sk_ignore_df(const struct sock *sk)
  371. {
  372. return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
  373. inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
  374. }
  375. static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
  376. bool forwarding)
  377. {
  378. const struct rtable *rt = container_of(dst, struct rtable, dst);
  379. struct net *net = dev_net(dst->dev);
  380. unsigned int mtu;
  381. if (READ_ONCE(net->ipv4.sysctl_ip_fwd_use_pmtu) ||
  382. ip_mtu_locked(dst) ||
  383. !forwarding) {
  384. mtu = rt->rt_pmtu;
  385. if (mtu && time_before(jiffies, rt->dst.expires))
  386. goto out;
  387. }
  388. /* 'forwarding = true' case should always honour route mtu */
  389. mtu = dst_metric_raw(dst, RTAX_MTU);
  390. if (mtu)
  391. goto out;
  392. mtu = READ_ONCE(dst->dev->mtu);
  393. if (unlikely(ip_mtu_locked(dst))) {
  394. if (rt->rt_uses_gateway && mtu > 576)
  395. mtu = 576;
  396. }
  397. out:
  398. mtu = min_t(unsigned int, mtu, IP_MAX_MTU);
  399. return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
  400. }
  401. static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
  402. const struct sk_buff *skb)
  403. {
  404. unsigned int mtu;
  405. if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) {
  406. bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
  407. return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
  408. }
  409. mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
  410. return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu);
  411. }
  412. struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
  413. int fc_mx_len,
  414. struct netlink_ext_ack *extack);
  415. static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics)
  416. {
  417. if (fib_metrics != &dst_default_metrics &&
  418. refcount_dec_and_test(&fib_metrics->refcnt))
  419. kfree(fib_metrics);
  420. }
  421. /* ipv4 and ipv6 both use refcounted metrics if it is not the default */
  422. static inline
  423. void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics)
  424. {
  425. dst_init_metrics(dst, fib_metrics->metrics, true);
  426. if (fib_metrics != &dst_default_metrics) {
  427. dst->_metrics |= DST_METRICS_REFCOUNTED;
  428. refcount_inc(&fib_metrics->refcnt);
  429. }
  430. }
  431. static inline
  432. void ip_dst_metrics_put(struct dst_entry *dst)
  433. {
  434. struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst);
  435. if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt))
  436. kfree(p);
  437. }
  438. void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
  439. static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb,
  440. struct sock *sk, int segs)
  441. {
  442. struct iphdr *iph = ip_hdr(skb);
  443. /* We had many attacks based on IPID, use the private
  444. * generator as much as we can.
  445. */
  446. if (sk && inet_sk(sk)->inet_daddr) {
  447. iph->id = htons(inet_sk(sk)->inet_id);
  448. inet_sk(sk)->inet_id += segs;
  449. return;
  450. }
  451. if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) {
  452. iph->id = 0;
  453. } else {
  454. /* Unfortunately we need the big hammer to get a suitable IPID */
  455. __ip_select_ident(net, iph, segs);
  456. }
  457. }
  458. static inline void ip_select_ident(struct net *net, struct sk_buff *skb,
  459. struct sock *sk)
  460. {
  461. ip_select_ident_segs(net, skb, sk, 1);
  462. }
  463. static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto)
  464. {
  465. return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr,
  466. skb->len, proto, 0);
  467. }
  468. /* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store
  469. * Equivalent to : flow->v4addrs.src = iph->saddr;
  470. * flow->v4addrs.dst = iph->daddr;
  471. */
  472. static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow,
  473. const struct iphdr *iph)
  474. {
  475. BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) !=
  476. offsetof(typeof(flow->addrs), v4addrs.src) +
  477. sizeof(flow->addrs.v4addrs.src));
  478. memcpy(&flow->addrs.v4addrs, &iph->addrs, sizeof(flow->addrs.v4addrs));
  479. flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
  480. }
  481. /*
  482. * Map a multicast IP onto multicast MAC for type ethernet.
  483. */
  484. static inline void ip_eth_mc_map(__be32 naddr, char *buf)
  485. {
  486. __u32 addr=ntohl(naddr);
  487. buf[0]=0x01;
  488. buf[1]=0x00;
  489. buf[2]=0x5e;
  490. buf[5]=addr&0xFF;
  491. addr>>=8;
  492. buf[4]=addr&0xFF;
  493. addr>>=8;
  494. buf[3]=addr&0x7F;
  495. }
  496. /*
  497. * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
  498. * Leave P_Key as 0 to be filled in by driver.
  499. */
  500. static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
  501. {
  502. __u32 addr;
  503. unsigned char scope = broadcast[5] & 0xF;
  504. buf[0] = 0; /* Reserved */
  505. buf[1] = 0xff; /* Multicast QPN */
  506. buf[2] = 0xff;
  507. buf[3] = 0xff;
  508. addr = ntohl(naddr);
  509. buf[4] = 0xff;
  510. buf[5] = 0x10 | scope; /* scope from broadcast address */
  511. buf[6] = 0x40; /* IPv4 signature */
  512. buf[7] = 0x1b;
  513. buf[8] = broadcast[8]; /* P_Key */
  514. buf[9] = broadcast[9];
  515. buf[10] = 0;
  516. buf[11] = 0;
  517. buf[12] = 0;
  518. buf[13] = 0;
  519. buf[14] = 0;
  520. buf[15] = 0;
  521. buf[19] = addr & 0xff;
  522. addr >>= 8;
  523. buf[18] = addr & 0xff;
  524. addr >>= 8;
  525. buf[17] = addr & 0xff;
  526. addr >>= 8;
  527. buf[16] = addr & 0x0f;
  528. }
  529. static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
  530. {
  531. if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
  532. memcpy(buf, broadcast, 4);
  533. else
  534. memcpy(buf, &naddr, sizeof(naddr));
  535. }
  536. #if IS_ENABLED(CONFIG_IPV6)
  537. #include <linux/ipv6.h>
  538. #endif
  539. static __inline__ void inet_reset_saddr(struct sock *sk)
  540. {
  541. inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
  542. #if IS_ENABLED(CONFIG_IPV6)
  543. if (sk->sk_family == PF_INET6) {
  544. struct ipv6_pinfo *np = inet6_sk(sk);
  545. memset(&np->saddr, 0, sizeof(np->saddr));
  546. memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
  547. }
  548. #endif
  549. }
  550. #endif
  551. static inline unsigned int ipv4_addr_hash(__be32 ip)
  552. {
  553. return (__force unsigned int) ip;
  554. }
  555. static inline u32 ipv4_portaddr_hash(const struct net *net,
  556. __be32 saddr,
  557. unsigned int port)
  558. {
  559. return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
  560. }
  561. bool ip_call_ra_chain(struct sk_buff *skb);
  562. /*
  563. * Functions provided by ip_fragment.c
  564. */
  565. enum ip_defrag_users {
  566. IP_DEFRAG_LOCAL_DELIVER,
  567. IP_DEFRAG_CALL_RA_CHAIN,
  568. IP_DEFRAG_CONNTRACK_IN,
  569. __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
  570. IP_DEFRAG_CONNTRACK_OUT,
  571. __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
  572. IP_DEFRAG_CONNTRACK_BRIDGE_IN,
  573. __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
  574. IP_DEFRAG_VS_IN,
  575. IP_DEFRAG_VS_OUT,
  576. IP_DEFRAG_VS_FWD,
  577. IP_DEFRAG_AF_PACKET,
  578. IP_DEFRAG_MACVLAN,
  579. };
  580. /* Return true if the value of 'user' is between 'lower_bond'
  581. * and 'upper_bond' inclusively.
  582. */
  583. static inline bool ip_defrag_user_in_between(u32 user,
  584. enum ip_defrag_users lower_bond,
  585. enum ip_defrag_users upper_bond)
  586. {
  587. return user >= lower_bond && user <= upper_bond;
  588. }
  589. int ip_defrag(struct net *net, struct sk_buff *skb, u32 user);
  590. #ifdef CONFIG_INET
  591. struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user);
  592. #else
  593. static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user)
  594. {
  595. return skb;
  596. }
  597. #endif
  598. /*
  599. * Functions provided by ip_forward.c
  600. */
  601. int ip_forward(struct sk_buff *skb);
  602. /*
  603. * Functions provided by ip_options.c
  604. */
  605. void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
  606. __be32 daddr, struct rtable *rt);
  607. int __ip_options_echo(struct net *net, struct ip_options *dopt,
  608. struct sk_buff *skb, const struct ip_options *sopt);
  609. static inline int ip_options_echo(struct net *net, struct ip_options *dopt,
  610. struct sk_buff *skb)
  611. {
  612. return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt);
  613. }
  614. void ip_options_fragment(struct sk_buff *skb);
  615. int __ip_options_compile(struct net *net, struct ip_options *opt,
  616. struct sk_buff *skb, __be32 *info);
  617. int ip_options_compile(struct net *net, struct ip_options *opt,
  618. struct sk_buff *skb);
  619. int ip_options_get(struct net *net, struct ip_options_rcu **optp,
  620. sockptr_t data, int optlen);
  621. void ip_options_undo(struct ip_options *opt);
  622. void ip_forward_options(struct sk_buff *skb);
  623. int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev);
  624. /*
  625. * Functions provided by ip_sockglue.c
  626. */
  627. void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
  628. void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk,
  629. struct sk_buff *skb, int tlen, int offset);
  630. int ip_cmsg_send(struct sock *sk, struct msghdr *msg,
  631. struct ipcm_cookie *ipc, bool allow_ipv6);
  632. DECLARE_STATIC_KEY_FALSE(ip4_min_ttl);
  633. int do_ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
  634. unsigned int optlen);
  635. int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
  636. unsigned int optlen);
  637. int do_ip_getsockopt(struct sock *sk, int level, int optname,
  638. sockptr_t optval, sockptr_t optlen);
  639. int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
  640. int __user *optlen);
  641. int ip_ra_control(struct sock *sk, unsigned char on,
  642. void (*destructor)(struct sock *));
  643. int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
  644. void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
  645. u32 info, u8 *payload);
  646. void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
  647. u32 info);
  648. static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
  649. {
  650. ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0);
  651. }
  652. bool icmp_global_allow(void);
  653. extern int sysctl_icmp_msgs_per_sec;
  654. extern int sysctl_icmp_msgs_burst;
  655. #ifdef CONFIG_PROC_FS
  656. int ip_misc_proc_init(void);
  657. #endif
  658. int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family,
  659. struct netlink_ext_ack *extack);
  660. static inline bool inetdev_valid_mtu(unsigned int mtu)
  661. {
  662. return likely(mtu >= IPV4_MIN_MTU);
  663. }
  664. void ip_sock_set_freebind(struct sock *sk);
  665. int ip_sock_set_mtu_discover(struct sock *sk, int val);
  666. void ip_sock_set_pktinfo(struct sock *sk);
  667. void ip_sock_set_recverr(struct sock *sk);
  668. void ip_sock_set_tos(struct sock *sk, int val);
  669. void __ip_sock_set_tos(struct sock *sk, int val);
  670. #endif /* _IP_H */