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 25 #include <linux/types.h> 26 #include <linux/ip.h> 27 #include <linux/in.h> 28 #include <linux/skbuff.h> 29 30 #include <net/inet_sock.h> 31 #include <net/snmp.h> 32 #include <net/flow.h> 33 34 struct sock; 35 36 struct inet_skb_parm { 37 struct ip_options opt; /* Compiled IP options */ 38 unsigned char flags; 39 40 #define IPSKB_FORWARDED 1 41 #define IPSKB_XFRM_TUNNEL_SIZE 2 42 #define IPSKB_XFRM_TRANSFORMED 4 43 #define IPSKB_FRAG_COMPLETE 8 44 #define IPSKB_REROUTED 16 45 46 u16 frag_max_size; 47 }; 48 49 static inline unsigned int ip_hdrlen(const struct sk_buff *skb) 50 { 51 return ip_hdr(skb)->ihl * 4; 52 } 53 54 struct ipcm_cookie { 55 __be32 addr; 56 int oif; 57 struct ip_options_rcu *opt; 58 __u8 tx_flags; 59 }; 60 61 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb)) 62 63 struct ip_ra_chain { 64 struct ip_ra_chain __rcu *next; 65 struct sock *sk; 66 union { 67 void (*destructor)(struct sock *); 68 struct sock *saved_sk; 69 }; 70 struct rcu_head rcu; 71 }; 72 73 extern struct ip_ra_chain __rcu *ip_ra_chain; 74 75 /* IP flags. */ 76 #define IP_CE 0x8000 /* Flag: "Congestion" */ 77 #define IP_DF 0x4000 /* Flag: "Don't Fragment" */ 78 #define IP_MF 0x2000 /* Flag: "More Fragments" */ 79 #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */ 80 81 #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */ 82 83 struct msghdr; 84 struct net_device; 85 struct packet_type; 86 struct rtable; 87 struct sockaddr; 88 89 extern int igmp_mc_proc_init(void); 90 91 /* 92 * Functions provided by ip.c 93 */ 94 95 extern int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk, 96 __be32 saddr, __be32 daddr, 97 struct ip_options_rcu *opt); 98 extern int ip_rcv(struct sk_buff *skb, struct net_device *dev, 99 struct packet_type *pt, struct net_device *orig_dev); 100 extern int ip_local_deliver(struct sk_buff *skb); 101 extern int ip_mr_input(struct sk_buff *skb); 102 extern int ip_output(struct sk_buff *skb); 103 extern int ip_mc_output(struct sk_buff *skb); 104 extern int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *)); 105 extern int ip_do_nat(struct sk_buff *skb); 106 extern void ip_send_check(struct iphdr *ip); 107 extern int __ip_local_out(struct sk_buff *skb); 108 extern int ip_local_out(struct sk_buff *skb); 109 extern int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl); 110 extern void ip_init(void); 111 extern int ip_append_data(struct sock *sk, struct flowi4 *fl4, 112 int getfrag(void *from, char *to, int offset, int len, 113 int odd, struct sk_buff *skb), 114 void *from, int len, int protolen, 115 struct ipcm_cookie *ipc, 116 struct rtable **rt, 117 unsigned int flags); 118 extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb); 119 extern ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page, 120 int offset, size_t size, int flags); 121 extern struct sk_buff *__ip_make_skb(struct sock *sk, 122 struct flowi4 *fl4, 123 struct sk_buff_head *queue, 124 struct inet_cork *cork); 125 extern int ip_send_skb(struct net *net, struct sk_buff *skb); 126 extern int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4); 127 extern void ip_flush_pending_frames(struct sock *sk); 128 extern struct sk_buff *ip_make_skb(struct sock *sk, 129 struct flowi4 *fl4, 130 int getfrag(void *from, char *to, int offset, int len, 131 int odd, struct sk_buff *skb), 132 void *from, int length, int transhdrlen, 133 struct ipcm_cookie *ipc, 134 struct rtable **rtp, 135 unsigned int flags); 136 137 static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4) 138 { 139 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base); 140 } 141 142 /* datagram.c */ 143 extern int ip4_datagram_connect(struct sock *sk, 144 struct sockaddr *uaddr, int addr_len); 145 146 struct ip_reply_arg { 147 struct kvec iov[1]; 148 int flags; 149 __wsum csum; 150 int csumoffset; /* u16 offset of csum in iov[0].iov_base */ 151 /* -1 if not needed */ 152 int bound_dev_if; 153 u8 tos; 154 }; 155 156 #define IP_REPLY_ARG_NOSRCCHECK 1 157 158 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg) 159 { 160 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0; 161 } 162 163 void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr, 164 __be32 saddr, const struct ip_reply_arg *arg, 165 unsigned int len); 166 167 struct ipv4_config { 168 int log_martians; 169 int no_pmtu_disc; 170 }; 171 172 extern struct ipv4_config ipv4_config; 173 #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field) 174 #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field) 175 #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) 176 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val) 177 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) 178 #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val) 179 #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field) 180 #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field) 181 #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field) 182 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd) 183 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd) 184 185 extern unsigned long snmp_fold_field(void __percpu *mib[], int offt); 186 #if BITS_PER_LONG==32 187 extern u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off); 188 #else 189 static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off) 190 { 191 return snmp_fold_field(mib, offt); 192 } 193 #endif 194 extern int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align); 195 extern void snmp_mib_free(void __percpu *ptr[2]); 196 197 extern struct local_ports { 198 seqlock_t lock; 199 int range[2]; 200 } sysctl_local_ports; 201 extern void inet_get_local_port_range(int *low, int *high); 202 203 extern unsigned long *sysctl_local_reserved_ports; 204 static inline int inet_is_reserved_local_port(int port) 205 { 206 return test_bit(port, sysctl_local_reserved_ports); 207 } 208 209 extern int sysctl_ip_nonlocal_bind; 210 211 /* From inetpeer.c */ 212 extern int inet_peer_threshold; 213 extern int inet_peer_minttl; 214 extern int inet_peer_maxttl; 215 216 /* From ip_input.c */ 217 extern int sysctl_ip_early_demux; 218 219 /* From ip_output.c */ 220 extern int sysctl_ip_dynaddr; 221 222 extern void ipfrag_init(void); 223 224 extern void ip_static_sysctl_init(void); 225 226 static inline bool ip_is_fragment(const struct iphdr *iph) 227 { 228 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0; 229 } 230 231 #ifdef CONFIG_INET 232 #include <net/dst.h> 233 234 /* The function in 2.2 was invalid, producing wrong result for 235 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */ 236 static inline 237 int ip_decrease_ttl(struct iphdr *iph) 238 { 239 u32 check = (__force u32)iph->check; 240 check += (__force u32)htons(0x0100); 241 iph->check = (__force __sum16)(check + (check>=0xFFFF)); 242 return --iph->ttl; 243 } 244 245 static inline 246 int ip_dont_fragment(struct sock *sk, struct dst_entry *dst) 247 { 248 return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO || 249 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT && 250 !(dst_metric_locked(dst, RTAX_MTU))); 251 } 252 253 extern void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more); 254 255 static inline void ip_select_ident(struct iphdr *iph, struct dst_entry *dst, struct sock *sk) 256 { 257 if (iph->frag_off & htons(IP_DF)) { 258 /* This is only to work around buggy Windows95/2000 259 * VJ compression implementations. If the ID field 260 * does not change, they drop every other packet in 261 * a TCP stream using header compression. 262 */ 263 iph->id = (sk && inet_sk(sk)->inet_daddr) ? 264 htons(inet_sk(sk)->inet_id++) : 0; 265 } else 266 __ip_select_ident(iph, dst, 0); 267 } 268 269 static inline void ip_select_ident_more(struct iphdr *iph, struct dst_entry *dst, struct sock *sk, int more) 270 { 271 if (iph->frag_off & htons(IP_DF)) { 272 if (sk && inet_sk(sk)->inet_daddr) { 273 iph->id = htons(inet_sk(sk)->inet_id); 274 inet_sk(sk)->inet_id += 1 + more; 275 } else 276 iph->id = 0; 277 } else 278 __ip_select_ident(iph, dst, more); 279 } 280 281 /* 282 * Map a multicast IP onto multicast MAC for type ethernet. 283 */ 284 285 static inline void ip_eth_mc_map(__be32 naddr, char *buf) 286 { 287 __u32 addr=ntohl(naddr); 288 buf[0]=0x01; 289 buf[1]=0x00; 290 buf[2]=0x5e; 291 buf[5]=addr&0xFF; 292 addr>>=8; 293 buf[4]=addr&0xFF; 294 addr>>=8; 295 buf[3]=addr&0x7F; 296 } 297 298 /* 299 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand. 300 * Leave P_Key as 0 to be filled in by driver. 301 */ 302 303 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) 304 { 305 __u32 addr; 306 unsigned char scope = broadcast[5] & 0xF; 307 308 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 } 333 334 static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) 335 { 336 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0) 337 memcpy(buf, broadcast, 4); 338 else 339 memcpy(buf, &naddr, sizeof(naddr)); 340 } 341 342 #if IS_ENABLED(CONFIG_IPV6) 343 #include <linux/ipv6.h> 344 #endif 345 346 static __inline__ void inet_reset_saddr(struct sock *sk) 347 { 348 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0; 349 #if IS_ENABLED(CONFIG_IPV6) 350 if (sk->sk_family == PF_INET6) { 351 struct ipv6_pinfo *np = inet6_sk(sk); 352 353 memset(&np->saddr, 0, sizeof(np->saddr)); 354 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr)); 355 } 356 #endif 357 } 358 359 #endif 360 361 static inline int sk_mc_loop(struct sock *sk) 362 { 363 if (!sk) 364 return 1; 365 switch (sk->sk_family) { 366 case AF_INET: 367 return inet_sk(sk)->mc_loop; 368 #if IS_ENABLED(CONFIG_IPV6) 369 case AF_INET6: 370 return inet6_sk(sk)->mc_loop; 371 #endif 372 } 373 WARN_ON(1); 374 return 1; 375 } 376 377 extern bool ip_call_ra_chain(struct sk_buff *skb); 378 379 /* 380 * Functions provided by ip_fragment.c 381 */ 382 383 enum ip_defrag_users { 384 IP_DEFRAG_LOCAL_DELIVER, 385 IP_DEFRAG_CALL_RA_CHAIN, 386 IP_DEFRAG_CONNTRACK_IN, 387 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX, 388 IP_DEFRAG_CONNTRACK_OUT, 389 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX, 390 IP_DEFRAG_CONNTRACK_BRIDGE_IN, 391 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX, 392 IP_DEFRAG_VS_IN, 393 IP_DEFRAG_VS_OUT, 394 IP_DEFRAG_VS_FWD, 395 IP_DEFRAG_AF_PACKET, 396 IP_DEFRAG_MACVLAN, 397 }; 398 399 int ip_defrag(struct sk_buff *skb, u32 user); 400 #ifdef CONFIG_INET 401 struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user); 402 #else 403 static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user) 404 { 405 return skb; 406 } 407 #endif 408 int ip_frag_mem(struct net *net); 409 int ip_frag_nqueues(struct net *net); 410 411 /* 412 * Functions provided by ip_forward.c 413 */ 414 415 extern int ip_forward(struct sk_buff *skb); 416 417 /* 418 * Functions provided by ip_options.c 419 */ 420 421 extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt, 422 __be32 daddr, struct rtable *rt, int is_frag); 423 extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb); 424 extern void ip_options_fragment(struct sk_buff *skb); 425 extern int ip_options_compile(struct net *net, 426 struct ip_options *opt, struct sk_buff *skb); 427 extern int ip_options_get(struct net *net, struct ip_options_rcu **optp, 428 unsigned char *data, int optlen); 429 extern int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp, 430 unsigned char __user *data, int optlen); 431 extern void ip_options_undo(struct ip_options * opt); 432 extern void ip_forward_options(struct sk_buff *skb); 433 extern int ip_options_rcv_srr(struct sk_buff *skb); 434 435 /* 436 * Functions provided by ip_sockglue.c 437 */ 438 439 extern void ipv4_pktinfo_prepare(struct sk_buff *skb); 440 extern void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb); 441 extern int ip_cmsg_send(struct net *net, 442 struct msghdr *msg, struct ipcm_cookie *ipc); 443 extern int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen); 444 extern int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen); 445 extern int compat_ip_setsockopt(struct sock *sk, int level, 446 int optname, char __user *optval, unsigned int optlen); 447 extern int compat_ip_getsockopt(struct sock *sk, int level, 448 int optname, char __user *optval, int __user *optlen); 449 extern int ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *)); 450 451 extern int ip_recv_error(struct sock *sk, struct msghdr *msg, int len); 452 extern void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, 453 __be16 port, u32 info, u8 *payload); 454 extern void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport, 455 u32 info); 456 457 #ifdef CONFIG_PROC_FS 458 extern int ip_misc_proc_init(void); 459 #endif 460 461 #endif /* _IP_H */ 462