include/net/ip.h

   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 extern void ip4_datagram_release_cb(struct sock *sk);
 147
 148 struct ip_reply_arg {
 149         struct kvec iov[1];
 150         int         flags;
 151         __wsum      csum;
 152         int         csumoffset; /* u16 offset of csum in iov[0].iov_base */
 153                                 /* -1 if not needed */
 154         int         bound_dev_if;
 155         u8          tos;
 156 };
 157
 158 #define IP_REPLY_ARG_NOSRCCHECK 1
 159
 160 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
 161 {
 162         return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
 163 }
 164
 165 void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr,
 166                            __be32 saddr, const struct ip_reply_arg *arg,
 167                            unsigned int len);
 168
 169 struct ipv4_config {
 170         int     log_martians;
 171         int     no_pmtu_disc;
 172 };
 173
 174 extern struct ipv4_config ipv4_config;
 175 #define IP_INC_STATS(net, field)        SNMP_INC_STATS64((net)->mib.ip_statistics, field)
 176 #define IP_INC_STATS_BH(net, field)     SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
 177 #define IP_ADD_STATS(net, field, val)   SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
 178 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
 179 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
 180 #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
 181 #define NET_INC_STATS(net, field)       SNMP_INC_STATS((net)->mib.net_statistics, field)
 182 #define NET_INC_STATS_BH(net, field)    SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
 183 #define NET_INC_STATS_USER(net, field)  SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
 184 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
 185 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
 186
 187 extern unsigned long snmp_fold_field(void __percpu *mib[], int offt);
 188 #if BITS_PER_LONG==32
 189 extern u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
 190 #else
 191 static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
 192 {
 193         return snmp_fold_field(mib, offt);
 194 }
 195 #endif
 196 extern int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
 197 extern void snmp_mib_free(void __percpu *ptr[2]);
 198
 199 extern struct local_ports {
 200         seqlock_t       lock;
 201         int             range[2];
 202 } sysctl_local_ports;
 203 extern void inet_get_local_port_range(int *low, int *high);
 204
 205 extern unsigned long *sysctl_local_reserved_ports;
 206 static inline int inet_is_reserved_local_port(int port)
 207 {
 208         return test_bit(port, sysctl_local_reserved_ports);
 209 }
 210
 211 extern int sysctl_ip_nonlocal_bind;
 212
 213 /* From inetpeer.c */
 214 extern int inet_peer_threshold;
 215 extern int inet_peer_minttl;
 216 extern int inet_peer_maxttl;
 217
 218 /* From ip_input.c */
 219 extern int sysctl_ip_early_demux;
 220
 221 /* From ip_output.c */
 222 extern int sysctl_ip_dynaddr;
 223
 224 extern void ipfrag_init(void);
 225
 226 extern void ip_static_sysctl_init(void);
 227
 228 static inline bool ip_is_fragment(const struct iphdr *iph)
 229 {
 230         return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
 231 }
 232
 233 #ifdef CONFIG_INET
 234 #include <net/dst.h>
 235
 236 /* The function in 2.2 was invalid, producing wrong result for
 237  * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
 238 static inline
 239 int ip_decrease_ttl(struct iphdr *iph)
 240 {
 241         u32 check = (__force u32)iph->check;
 242         check += (__force u32)htons(0x0100);
 243         iph->check = (__force __sum16)(check + (check>=0xFFFF));
 244         return --iph->ttl;
 245 }
 246
 247 static inline
 248 int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
 249 {
 250         return  inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
 251                 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
 252                  !(dst_metric_locked(dst, RTAX_MTU)));
 253 }
 254
 255 extern void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
 256
 257 static inline void ip_select_ident(struct iphdr *iph, struct dst_entry *dst, struct sock *sk)
 258 {
 259         if (iph->frag_off & htons(IP_DF)) {
 260                 /* This is only to work around buggy Windows95/2000
 261                  * VJ compression implementations.  If the ID field
 262                  * does not change, they drop every other packet in
 263                  * a TCP stream using header compression.
 264                  */
 265                 iph->id = (sk && inet_sk(sk)->inet_daddr) ?
 266                                         htons(inet_sk(sk)->inet_id++) : 0;
 267         } else
 268                 __ip_select_ident(iph, dst, 0);
 269 }
 270
 271 static inline void ip_select_ident_more(struct iphdr *iph, struct dst_entry *dst, struct sock *sk, int more)
 272 {
 273         if (iph->frag_off & htons(IP_DF)) {
 274                 if (sk && inet_sk(sk)->inet_daddr) {
 275                         iph->id = htons(inet_sk(sk)->inet_id);
 276                         inet_sk(sk)->inet_id += 1 + more;
 277                 } else
 278                         iph->id = 0;
 279         } else
 280                 __ip_select_ident(iph, dst, more);
 281 }
 282
 283 /*
 284  *      Map a multicast IP onto multicast MAC for type ethernet.
 285  */
 286
 287 static inline void ip_eth_mc_map(__be32 naddr, char *buf)
 288 {
 289         __u32 addr=ntohl(naddr);
 290         buf[0]=0x01;
 291         buf[1]=0x00;
 292         buf[2]=0x5e;
 293         buf[5]=addr&0xFF;
 294         addr>>=8;
 295         buf[4]=addr&0xFF;
 296         addr>>=8;
 297         buf[3]=addr&0x7F;
 298 }
 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
 305 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
 306 {
 307         __u32 addr;
 308         unsigned char scope = broadcast[5] & 0xF;
 309
 310         buf[0]  = 0;            /* Reserved */
 311         buf[1]  = 0xff;         /* Multicast QPN */
 312         buf[2]  = 0xff;
 313         buf[3]  = 0xff;
 314         addr    = ntohl(naddr);
 315         buf[4]  = 0xff;
 316         buf[5]  = 0x10 | scope; /* scope from broadcast address */
 317         buf[6]  = 0x40;         /* IPv4 signature */
 318         buf[7]  = 0x1b;
 319         buf[8]  = broadcast[8];         /* P_Key */
 320         buf[9]  = broadcast[9];
 321         buf[10] = 0;
 322         buf[11] = 0;
 323         buf[12] = 0;
 324         buf[13] = 0;
 325         buf[14] = 0;
 326         buf[15] = 0;
 327         buf[19] = addr & 0xff;
 328         addr  >>= 8;
 329         buf[18] = addr & 0xff;
 330         addr  >>= 8;
 331         buf[17] = addr & 0xff;
 332         addr  >>= 8;
 333         buf[16] = addr & 0x0f;
 334 }
 335
 336 static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
 337 {
 338         if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
 339                 memcpy(buf, broadcast, 4);
 340         else
 341                 memcpy(buf, &naddr, sizeof(naddr));
 342 }
 343
 344 #if IS_ENABLED(CONFIG_IPV6)
 345 #include <linux/ipv6.h>
 346 #endif
 347
 348 static __inline__ void inet_reset_saddr(struct sock *sk)
 349 {
 350         inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
 351 #if IS_ENABLED(CONFIG_IPV6)
 352         if (sk->sk_family == PF_INET6) {
 353                 struct ipv6_pinfo *np = inet6_sk(sk);
 354
 355                 memset(&np->saddr, 0, sizeof(np->saddr));
 356                 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
 357         }
 358 #endif
 359 }
 360
 361 #endif
 362
 363 static inline int sk_mc_loop(struct sock *sk)
 364 {
 365         if (!sk)
 366                 return 1;
 367         switch (sk->sk_family) {
 368         case AF_INET:
 369                 return inet_sk(sk)->mc_loop;
 370 #if IS_ENABLED(CONFIG_IPV6)
 371         case AF_INET6:
 372                 return inet6_sk(sk)->mc_loop;
 373 #endif
 374         }
 375         WARN_ON(1);
 376         return 1;
 377 }
 378
 379 extern bool ip_call_ra_chain(struct sk_buff *skb);
 380
 381 /*
 382  *      Functions provided by ip_fragment.c
 383  */
 384
 385 enum ip_defrag_users {
 386         IP_DEFRAG_LOCAL_DELIVER,
 387         IP_DEFRAG_CALL_RA_CHAIN,
 388         IP_DEFRAG_CONNTRACK_IN,
 389         __IP_DEFRAG_CONNTRACK_IN_END    = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
 390         IP_DEFRAG_CONNTRACK_OUT,
 391         __IP_DEFRAG_CONNTRACK_OUT_END   = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
 392         IP_DEFRAG_CONNTRACK_BRIDGE_IN,
 393         __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
 394         IP_DEFRAG_VS_IN,
 395         IP_DEFRAG_VS_OUT,
 396         IP_DEFRAG_VS_FWD,
 397         IP_DEFRAG_AF_PACKET,
 398         IP_DEFRAG_MACVLAN,
 399 };
 400
 401 int ip_defrag(struct sk_buff *skb, u32 user);
 402 #ifdef CONFIG_INET
 403 struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
 404 #else
 405 static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
 406 {
 407         return skb;
 408 }
 409 #endif
 410 int ip_frag_mem(struct net *net);
 411 int ip_frag_nqueues(struct net *net);
 412
 413 /*
 414  *      Functions provided by ip_forward.c
 415  */
 416
 417 extern int ip_forward(struct sk_buff *skb);
 418
 419 /*
 420  *      Functions provided by ip_options.c
 421  */
 422
 423 extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
 424                              __be32 daddr, struct rtable *rt, int is_frag);
 425 extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
 426 extern void ip_options_fragment(struct sk_buff *skb);
 427 extern int ip_options_compile(struct net *net,
 428                               struct ip_options *opt, struct sk_buff *skb);
 429 extern int ip_options_get(struct net *net, struct ip_options_rcu **optp,
 430                           unsigned char *data, int optlen);
 431 extern int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
 432                                     unsigned char __user *data, int optlen);
 433 extern void ip_options_undo(struct ip_options * opt);
 434 extern void ip_forward_options(struct sk_buff *skb);
 435 extern int ip_options_rcv_srr(struct sk_buff *skb);
 436
 437 /*
 438  *      Functions provided by ip_sockglue.c
 439  */
 440
 441 extern void     ipv4_pktinfo_prepare(struct sk_buff *skb);
 442 extern void     ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
 443 extern int      ip_cmsg_send(struct net *net,
 444                              struct msghdr *msg, struct ipcm_cookie *ipc);
 445 extern int      ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen);
 446 extern int      ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);
 447 extern int      compat_ip_setsockopt(struct sock *sk, int level,
 448                         int optname, char __user *optval, unsigned int optlen);
 449 extern int      compat_ip_getsockopt(struct sock *sk, int level,
 450                         int optname, char __user *optval, int __user *optlen);
 451 extern int      ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *));
 452
 453 extern int      ip_recv_error(struct sock *sk, struct msghdr *msg, int len);
 454 extern void     ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
 455                               __be16 port, u32 info, u8 *payload);
 456 extern void     ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
 457                                u32 info);
 458
 459 #ifdef CONFIG_PROC_FS
 460 extern int ip_misc_proc_init(void);
 461 #endif
 462
 463 #endif  /* _IP_H */