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.
6 * Definitions for the IP module.
8 * Version: @(#)ip.h 1.0.2 05/07/93
15 * Mike McLagan : Routing by source
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.
25 #include <linux/types.h>
28 #include <linux/skbuff.h>
30 #include <net/inet_sock.h>
31 #include <net/route.h>
34 #include <net/flow_keys.h>
38 struct inet_skb_parm {
39 struct ip_options opt; /* Compiled IP options */
42 #define IPSKB_FORWARDED 1
43 #define IPSKB_XFRM_TUNNEL_SIZE 2
44 #define IPSKB_XFRM_TRANSFORMED 4
45 #define IPSKB_FRAG_COMPLETE 8
46 #define IPSKB_REROUTED 16
51 static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
53 return ip_hdr(skb)->ihl * 4;
59 struct ip_options_rcu *opt;
66 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
67 #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
70 struct ip_ra_chain __rcu *next;
73 void (*destructor)(struct sock *);
74 struct sock *saved_sk;
79 extern struct ip_ra_chain __rcu *ip_ra_chain;
82 #define IP_CE 0x8000 /* Flag: "Congestion" */
83 #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
84 #define IP_MF 0x2000 /* Flag: "More Fragments" */
85 #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
87 #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
95 int igmp_mc_init(void);
98 * Functions provided by ip.c
101 int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
102 __be32 saddr, __be32 daddr,
103 struct ip_options_rcu *opt);
104 int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
105 struct net_device *orig_dev);
106 int ip_local_deliver(struct sk_buff *skb);
107 int ip_mr_input(struct sk_buff *skb);
108 int ip_output(struct sock *sk, struct sk_buff *skb);
109 int ip_mc_output(struct sock *sk, struct sk_buff *skb);
110 int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
111 int ip_do_nat(struct sk_buff *skb);
112 void ip_send_check(struct iphdr *ip);
113 int __ip_local_out(struct sk_buff *skb);
114 int ip_local_out_sk(struct sock *sk, struct sk_buff *skb);
115 static inline int ip_local_out(struct sk_buff *skb)
117 return ip_local_out_sk(skb->sk, skb);
120 int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl);
122 int ip_append_data(struct sock *sk, struct flowi4 *fl4,
123 int getfrag(void *from, char *to, int offset, int len,
124 int odd, struct sk_buff *skb),
125 void *from, int len, int protolen,
126 struct ipcm_cookie *ipc,
129 int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
130 struct sk_buff *skb);
131 ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
132 int offset, size_t size, int flags);
133 struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
134 struct sk_buff_head *queue,
135 struct inet_cork *cork);
136 int ip_send_skb(struct net *net, struct sk_buff *skb);
137 int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
138 void ip_flush_pending_frames(struct sock *sk);
139 struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
140 int getfrag(void *from, char *to, int offset,
141 int len, int odd, struct sk_buff *skb),
142 void *from, int length, int transhdrlen,
143 struct ipcm_cookie *ipc, struct rtable **rtp,
146 static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
148 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
151 static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
153 return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
156 static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
158 return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
162 int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
164 void ip4_datagram_release_cb(struct sock *sk);
166 struct ip_reply_arg {
170 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
171 /* -1 if not needed */
176 #define IP_REPLY_ARG_NOSRCCHECK 1
178 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
180 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
183 void ip_send_unicast_reply(struct net *net, struct sk_buff *skb,
184 const struct ip_options *sopt,
185 __be32 daddr, __be32 saddr,
186 const struct ip_reply_arg *arg,
189 #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
190 #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
191 #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
192 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
193 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
194 #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
195 #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
196 #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
197 #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
198 #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
199 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
200 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
202 unsigned long snmp_fold_field(void __percpu *mib, int offt);
203 #if BITS_PER_LONG==32
204 u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
206 static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
208 return snmp_fold_field(mib, offt);
212 void inet_get_local_port_range(struct net *net, int *low, int *high);
215 static inline int inet_is_local_reserved_port(struct net *net, int port)
217 if (!net->ipv4.sysctl_local_reserved_ports)
219 return test_bit(port, net->ipv4.sysctl_local_reserved_ports);
222 static inline bool sysctl_dev_name_is_allowed(const char *name)
224 return strcmp(name, "default") != 0 && strcmp(name, "all") != 0;
228 static inline int inet_is_local_reserved_port(struct net *net, int port)
234 /* From inetpeer.c */
235 extern int inet_peer_threshold;
236 extern int inet_peer_minttl;
237 extern int inet_peer_maxttl;
239 /* From ip_input.c */
240 extern int sysctl_ip_early_demux;
242 /* From ip_output.c */
243 extern int sysctl_ip_dynaddr;
245 void ipfrag_init(void);
247 void ip_static_sysctl_init(void);
249 #define IP4_REPLY_MARK(net, mark) \
250 ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
252 static inline bool ip_is_fragment(const struct iphdr *iph)
254 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
260 /* The function in 2.2 was invalid, producing wrong result for
261 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
263 int ip_decrease_ttl(struct iphdr *iph)
265 u32 check = (__force u32)iph->check;
266 check += (__force u32)htons(0x0100);
267 iph->check = (__force __sum16)(check + (check>=0xFFFF));
272 int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
274 return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
275 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
276 !(dst_metric_locked(dst, RTAX_MTU)));
279 static inline bool ip_sk_accept_pmtu(const struct sock *sk)
281 return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
282 inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
285 static inline bool ip_sk_use_pmtu(const struct sock *sk)
287 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
290 static inline bool ip_sk_ignore_df(const struct sock *sk)
292 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
293 inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
296 static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
299 struct net *net = dev_net(dst->dev);
301 if (net->ipv4.sysctl_ip_fwd_use_pmtu ||
302 dst_metric_locked(dst, RTAX_MTU) ||
306 return min(dst->dev->mtu, IP_MAX_MTU);
309 static inline unsigned int ip_skb_dst_mtu(const struct sk_buff *skb)
311 if (!skb->sk || ip_sk_use_pmtu(skb->sk)) {
312 bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
313 return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
315 return min(skb_dst(skb)->dev->mtu, IP_MAX_MTU);
319 u32 ip_idents_reserve(u32 hash, int segs);
320 void __ip_select_ident(struct iphdr *iph, int segs);
322 static inline void ip_select_ident_segs(struct sk_buff *skb, struct sock *sk, int segs)
324 struct iphdr *iph = ip_hdr(skb);
326 if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) {
327 /* This is only to work around buggy Windows95/2000
328 * VJ compression implementations. If the ID field
329 * does not change, they drop every other packet in
330 * a TCP stream using header compression.
332 if (sk && inet_sk(sk)->inet_daddr) {
333 iph->id = htons(inet_sk(sk)->inet_id);
334 inet_sk(sk)->inet_id += segs;
339 __ip_select_ident(iph, segs);
343 static inline void ip_select_ident(struct sk_buff *skb, struct sock *sk)
345 ip_select_ident_segs(skb, sk, 1);
348 static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto)
350 return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr,
354 static inline void inet_set_txhash(struct sock *sk)
356 struct inet_sock *inet = inet_sk(sk);
357 struct flow_keys keys;
359 keys.src = inet->inet_saddr;
360 keys.dst = inet->inet_daddr;
361 keys.port16[0] = inet->inet_sport;
362 keys.port16[1] = inet->inet_dport;
364 sk->sk_txhash = flow_hash_from_keys(&keys);
367 static inline __wsum inet_gro_compute_pseudo(struct sk_buff *skb, int proto)
369 const struct iphdr *iph = skb_gro_network_header(skb);
371 return csum_tcpudp_nofold(iph->saddr, iph->daddr,
372 skb_gro_len(skb), proto, 0);
376 * Map a multicast IP onto multicast MAC for type ethernet.
379 static inline void ip_eth_mc_map(__be32 naddr, char *buf)
381 __u32 addr=ntohl(naddr);
393 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
394 * Leave P_Key as 0 to be filled in by driver.
397 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
400 unsigned char scope = broadcast[5] & 0xF;
402 buf[0] = 0; /* Reserved */
403 buf[1] = 0xff; /* Multicast QPN */
408 buf[5] = 0x10 | scope; /* scope from broadcast address */
409 buf[6] = 0x40; /* IPv4 signature */
411 buf[8] = broadcast[8]; /* P_Key */
412 buf[9] = broadcast[9];
419 buf[19] = addr & 0xff;
421 buf[18] = addr & 0xff;
423 buf[17] = addr & 0xff;
425 buf[16] = addr & 0x0f;
428 static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
430 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
431 memcpy(buf, broadcast, 4);
433 memcpy(buf, &naddr, sizeof(naddr));
436 #if IS_ENABLED(CONFIG_IPV6)
437 #include <linux/ipv6.h>
440 static __inline__ void inet_reset_saddr(struct sock *sk)
442 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
443 #if IS_ENABLED(CONFIG_IPV6)
444 if (sk->sk_family == PF_INET6) {
445 struct ipv6_pinfo *np = inet6_sk(sk);
447 memset(&np->saddr, 0, sizeof(np->saddr));
448 memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
455 static inline int sk_mc_loop(struct sock *sk)
459 switch (sk->sk_family) {
461 return inet_sk(sk)->mc_loop;
462 #if IS_ENABLED(CONFIG_IPV6)
464 return inet6_sk(sk)->mc_loop;
471 bool ip_call_ra_chain(struct sk_buff *skb);
474 * Functions provided by ip_fragment.c
477 enum ip_defrag_users {
478 IP_DEFRAG_LOCAL_DELIVER,
479 IP_DEFRAG_CALL_RA_CHAIN,
480 IP_DEFRAG_CONNTRACK_IN,
481 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
482 IP_DEFRAG_CONNTRACK_OUT,
483 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
484 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
485 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
493 int ip_defrag(struct sk_buff *skb, u32 user);
495 struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
497 static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
502 int ip_frag_mem(struct net *net);
505 * Functions provided by ip_forward.c
508 int ip_forward(struct sk_buff *skb);
511 * Functions provided by ip_options.c
514 void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
515 __be32 daddr, struct rtable *rt, int is_frag);
517 int __ip_options_echo(struct ip_options *dopt, struct sk_buff *skb,
518 const struct ip_options *sopt);
519 static inline int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb)
521 return __ip_options_echo(dopt, skb, &IPCB(skb)->opt);
524 void ip_options_fragment(struct sk_buff *skb);
525 int ip_options_compile(struct net *net, struct ip_options *opt,
526 struct sk_buff *skb);
527 int ip_options_get(struct net *net, struct ip_options_rcu **optp,
528 unsigned char *data, int optlen);
529 int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
530 unsigned char __user *data, int optlen);
531 void ip_options_undo(struct ip_options *opt);
532 void ip_forward_options(struct sk_buff *skb);
533 int ip_options_rcv_srr(struct sk_buff *skb);
536 * Functions provided by ip_sockglue.c
539 void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
540 void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
541 int ip_cmsg_send(struct net *net, struct msghdr *msg,
542 struct ipcm_cookie *ipc, bool allow_ipv6);
543 int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
544 unsigned int optlen);
545 int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
547 int compat_ip_setsockopt(struct sock *sk, int level, int optname,
548 char __user *optval, unsigned int optlen);
549 int compat_ip_getsockopt(struct sock *sk, int level, int optname,
550 char __user *optval, int __user *optlen);
551 int ip_ra_control(struct sock *sk, unsigned char on,
552 void (*destructor)(struct sock *));
554 int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
555 void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
556 u32 info, u8 *payload);
557 void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
560 bool icmp_global_allow(void);
561 extern int sysctl_icmp_msgs_per_sec;
562 extern int sysctl_icmp_msgs_burst;
564 #ifdef CONFIG_PROC_FS
565 int ip_misc_proc_init(void);