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
11 * Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
12 * Alan Cox, <gw4pts@gw4pts.ampr.org>
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>
38 struct ip_options opt; /* Compiled IP options */
41 #define IPSKB_FORWARDED 1
42 #define IPSKB_XFRM_TUNNEL_SIZE 2
43 #define IPSKB_XFRM_TRANSFORMED 4
44 #define IPSKB_FRAG_COMPLETE 8
45 #define IPSKB_REROUTED 16
48 static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
50 return ip_hdr(skb)->ihl * 4;
57 struct ip_options *opt;
60 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
64 struct ip_ra_chain *next;
66 void (*destructor)(struct sock *);
69 extern struct ip_ra_chain *ip_ra_chain;
70 extern rwlock_t ip_ra_lock;
73 #define IP_CE 0x8000 /* Flag: "Congestion" */
74 #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
75 #define IP_MF 0x2000 /* Flag: "More Fragments" */
76 #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
78 #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
86 extern int igmp_mc_proc_init(void);
89 * Functions provided by ip.c
92 extern int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
93 __be32 saddr, __be32 daddr,
94 struct ip_options *opt);
95 extern int ip_rcv(struct sk_buff *skb, struct net_device *dev,
96 struct packet_type *pt, struct net_device *orig_dev);
97 extern int ip_local_deliver(struct sk_buff *skb);
98 extern int ip_mr_input(struct sk_buff *skb);
99 extern int ip_output(struct sk_buff *skb);
100 extern int ip_mc_output(struct sk_buff *skb);
101 extern int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
102 extern int ip_do_nat(struct sk_buff *skb);
103 extern void ip_send_check(struct iphdr *ip);
104 extern int __ip_local_out(struct sk_buff *skb);
105 extern int ip_local_out(struct sk_buff *skb);
106 extern int ip_queue_xmit(struct sk_buff *skb, int ipfragok);
107 extern void ip_init(void);
108 extern int ip_append_data(struct sock *sk,
109 int getfrag(void *from, char *to, int offset, int len,
110 int odd, struct sk_buff *skb),
111 void *from, int len, int protolen,
112 struct ipcm_cookie *ipc,
115 extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb);
116 extern ssize_t ip_append_page(struct sock *sk, struct page *page,
117 int offset, size_t size, int flags);
118 extern int ip_push_pending_frames(struct sock *sk);
119 extern void ip_flush_pending_frames(struct sock *sk);
122 extern int ip4_datagram_connect(struct sock *sk,
123 struct sockaddr *uaddr, int addr_len);
126 * Map a multicast IP onto multicast MAC for type Token Ring.
127 * This conforms to RFC1469 Option 2 Multicasting i.e.
128 * using a functional address to transmit / receive
132 static inline void ip_tr_mc_map(__be32 addr, char *buf)
142 struct ip_reply_arg {
146 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
147 /* -1 if not needed */
151 #define IP_REPLY_ARG_NOSRCCHECK 1
153 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
155 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
158 void ip_send_reply(struct sock *sk, struct sk_buff *skb, struct ip_reply_arg *arg,
167 extern struct ipv4_config ipv4_config;
168 #define IP_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.ip_statistics, field)
169 #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.ip_statistics, field)
170 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS_BH((net)->mib.ip_statistics, field, val)
171 #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
172 #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
173 #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
174 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
175 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
177 extern unsigned long snmp_fold_field(void *mib[], int offt);
178 extern int snmp_mib_init(void *ptr[2], size_t mibsize);
179 extern void snmp_mib_free(void *ptr[2]);
181 extern struct local_ports {
184 } sysctl_local_ports;
185 extern void inet_get_local_port_range(int *low, int *high);
187 extern int sysctl_ip_default_ttl;
188 extern int sysctl_ip_nonlocal_bind;
190 extern struct ctl_path net_core_path[];
191 extern struct ctl_path net_ipv4_ctl_path[];
193 /* From inetpeer.c */
194 extern int inet_peer_threshold;
195 extern int inet_peer_minttl;
196 extern int inet_peer_maxttl;
197 extern int inet_peer_gc_mintime;
198 extern int inet_peer_gc_maxtime;
200 /* From ip_output.c */
201 extern int sysctl_ip_dynaddr;
203 extern void ipfrag_init(void);
205 extern void ip_static_sysctl_init(void);
210 /* The function in 2.2 was invalid, producing wrong result for
211 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
213 int ip_decrease_ttl(struct iphdr *iph)
215 u32 check = (__force u32)iph->check;
216 check += (__force u32)htons(0x0100);
217 iph->check = (__force __sum16)(check + (check>=0xFFFF));
222 int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
224 return (inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
225 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
226 !(dst_metric_locked(dst, RTAX_MTU))));
229 extern void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
231 static inline void ip_select_ident(struct iphdr *iph, struct dst_entry *dst, struct sock *sk)
233 if (iph->frag_off & htons(IP_DF)) {
234 /* This is only to work around buggy Windows95/2000
235 * VJ compression implementations. If the ID field
236 * does not change, they drop every other packet in
237 * a TCP stream using header compression.
239 iph->id = (sk && inet_sk(sk)->daddr) ?
240 htons(inet_sk(sk)->id++) : 0;
242 __ip_select_ident(iph, dst, 0);
245 static inline void ip_select_ident_more(struct iphdr *iph, struct dst_entry *dst, struct sock *sk, int more)
247 if (iph->frag_off & htons(IP_DF)) {
248 if (sk && inet_sk(sk)->daddr) {
249 iph->id = htons(inet_sk(sk)->id);
250 inet_sk(sk)->id += 1 + more;
254 __ip_select_ident(iph, dst, more);
258 * Map a multicast IP onto multicast MAC for type ethernet.
261 static inline void ip_eth_mc_map(__be32 naddr, char *buf)
263 __u32 addr=ntohl(naddr);
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.
279 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
282 unsigned char scope = broadcast[5] & 0xF;
284 buf[0] = 0; /* Reserved */
285 buf[1] = 0xff; /* Multicast QPN */
290 buf[5] = 0x10 | scope; /* scope from broadcast address */
291 buf[6] = 0x40; /* IPv4 signature */
293 buf[8] = broadcast[8]; /* P_Key */
294 buf[9] = broadcast[9];
301 buf[19] = addr & 0xff;
303 buf[18] = addr & 0xff;
305 buf[17] = addr & 0xff;
307 buf[16] = addr & 0x0f;
310 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
311 #include <linux/ipv6.h>
314 static __inline__ void inet_reset_saddr(struct sock *sk)
316 inet_sk(sk)->rcv_saddr = inet_sk(sk)->saddr = 0;
317 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
318 if (sk->sk_family == PF_INET6) {
319 struct ipv6_pinfo *np = inet6_sk(sk);
321 memset(&np->saddr, 0, sizeof(np->saddr));
322 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
329 extern int ip_call_ra_chain(struct sk_buff *skb);
332 * Functions provided by ip_fragment.c
337 IP_DEFRAG_LOCAL_DELIVER,
338 IP_DEFRAG_CALL_RA_CHAIN,
339 IP_DEFRAG_CONNTRACK_IN,
340 IP_DEFRAG_CONNTRACK_OUT,
346 int ip_defrag(struct sk_buff *skb, u32 user);
347 int ip_frag_mem(struct net *net);
348 int ip_frag_nqueues(struct net *net);
351 * Functions provided by ip_forward.c
354 extern int ip_forward(struct sk_buff *skb);
357 * Functions provided by ip_options.c
360 extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt, __be32 daddr, struct rtable *rt, int is_frag);
361 extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
362 extern void ip_options_fragment(struct sk_buff *skb);
363 extern int ip_options_compile(struct net *net,
364 struct ip_options *opt, struct sk_buff *skb);
365 extern int ip_options_get(struct net *net, struct ip_options **optp,
366 unsigned char *data, int optlen);
367 extern int ip_options_get_from_user(struct net *net, struct ip_options **optp,
368 unsigned char __user *data, int optlen);
369 extern void ip_options_undo(struct ip_options * opt);
370 extern void ip_forward_options(struct sk_buff *skb);
371 extern int ip_options_rcv_srr(struct sk_buff *skb);
374 * Functions provided by ip_sockglue.c
377 extern void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
378 extern int ip_cmsg_send(struct net *net,
379 struct msghdr *msg, struct ipcm_cookie *ipc);
380 extern int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, int optlen);
381 extern int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);
382 extern int compat_ip_setsockopt(struct sock *sk, int level,
383 int optname, char __user *optval, int optlen);
384 extern int compat_ip_getsockopt(struct sock *sk, int level,
385 int optname, char __user *optval, int __user *optlen);
386 extern int ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *));
388 extern int ip_recv_error(struct sock *sk, struct msghdr *msg, int len);
389 extern void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
390 __be16 port, u32 info, u8 *payload);
391 extern void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
394 /* sysctl helpers - any sysctl which holds a value that ends up being
395 * fed into the routing cache should use these handlers.
397 int ipv4_doint_and_flush(ctl_table *ctl, int write,
398 struct file* filp, void __user *buffer,
399 size_t *lenp, loff_t *ppos);
400 int ipv4_doint_and_flush_strategy(ctl_table *table,
401 void __user *oldval, size_t __user *oldlenp,
402 void __user *newval, size_t newlen);
403 #ifdef CONFIG_PROC_FS
404 extern int ip_misc_proc_init(void);