]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/blame - include/net/ip.h
ipv4: IP_TOS and IP_TTL can be specified as ancillary data
[mirror_ubuntu-eoan-kernel.git] / include / net / ip.h
CommitLineData
1da177e4
LT
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 *
02c30a84 10 * Authors: Ross Biro
1da177e4
LT
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
1da177e4 25#include <linux/types.h>
1da177e4
LT
26#include <linux/ip.h>
27#include <linux/in.h>
c9bdd4b5 28#include <linux/skbuff.h>
14c85021
ACM
29
30#include <net/inet_sock.h>
1da177e4 31#include <net/snmp.h>
86b08d86 32#include <net/flow.h>
1da177e4
LT
33
34struct sock;
35
fd2c3ef7 36struct inet_skb_parm {
1da177e4
LT
37 struct ip_options opt; /* Compiled IP options */
38 unsigned char flags;
39
8cdfab8a
PM
40#define IPSKB_FORWARDED 1
41#define IPSKB_XFRM_TUNNEL_SIZE 2
3e3850e9
PM
42#define IPSKB_XFRM_TRANSFORMED 4
43#define IPSKB_FRAG_COMPLETE 8
48d5cad8 44#define IPSKB_REROUTED 16
5f2d04f1
PM
45
46 u16 frag_max_size;
1da177e4
LT
47};
48
c9bdd4b5
ACM
49static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
50{
eddc9ec5 51 return ip_hdr(skb)->ihl * 4;
c9bdd4b5
ACM
52}
53
fd2c3ef7 54struct ipcm_cookie {
c1d18f9f 55 __be32 addr;
1da177e4 56 int oif;
f6d8bd05 57 struct ip_options_rcu *opt;
2244d07b 58 __u8 tx_flags;
f02db315
FF
59 __u8 ttl;
60 __s16 tos;
61 char priority;
1da177e4
LT
62};
63
64#define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
65
fd2c3ef7 66struct ip_ra_chain {
43a951e9 67 struct ip_ra_chain __rcu *next;
1da177e4 68 struct sock *sk;
592fcb9d
ED
69 union {
70 void (*destructor)(struct sock *);
71 struct sock *saved_sk;
72 };
66018506 73 struct rcu_head rcu;
1da177e4
LT
74};
75
43a951e9 76extern struct ip_ra_chain __rcu *ip_ra_chain;
1da177e4
LT
77
78/* IP flags. */
79#define IP_CE 0x8000 /* Flag: "Congestion" */
80#define IP_DF 0x4000 /* Flag: "Don't Fragment" */
81#define IP_MF 0x2000 /* Flag: "More Fragments" */
82#define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
83
84#define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
85
14c85021
ACM
86struct msghdr;
87struct net_device;
88struct packet_type;
89struct rtable;
14c85021
ACM
90struct sockaddr;
91
5c3a0fd7 92int igmp_mc_proc_init(void);
1da177e4
LT
93
94/*
95 * Functions provided by ip.c
96 */
97
5c3a0fd7
JP
98int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
99 __be32 saddr, __be32 daddr,
100 struct ip_options_rcu *opt);
101int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
102 struct net_device *orig_dev);
103int ip_local_deliver(struct sk_buff *skb);
104int ip_mr_input(struct sk_buff *skb);
105int ip_output(struct sk_buff *skb);
106int ip_mc_output(struct sk_buff *skb);
107int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
108int ip_do_nat(struct sk_buff *skb);
109void ip_send_check(struct iphdr *ip);
110int __ip_local_out(struct sk_buff *skb);
111int ip_local_out(struct sk_buff *skb);
112int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
113void ip_init(void);
114int ip_append_data(struct sock *sk, struct flowi4 *fl4,
115 int getfrag(void *from, char *to, int offset, int len,
116 int odd, struct sk_buff *skb),
117 void *from, int len, int protolen,
118 struct ipcm_cookie *ipc,
119 struct rtable **rt,
120 unsigned int flags);
121int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
122 struct sk_buff *skb);
123ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
124 int offset, size_t size, int flags);
125struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
126 struct sk_buff_head *queue,
127 struct inet_cork *cork);
128int ip_send_skb(struct net *net, struct sk_buff *skb);
129int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
130void ip_flush_pending_frames(struct sock *sk);
131struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
132 int getfrag(void *from, char *to, int offset,
133 int len, int odd, struct sk_buff *skb),
134 void *from, int length, int transhdrlen,
135 struct ipcm_cookie *ipc, struct rtable **rtp,
136 unsigned int flags);
1c32c5ad 137
77968b78 138static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
1c32c5ad 139{
77968b78 140 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
1c32c5ad 141}
1da177e4
LT
142
143/* datagram.c */
5c3a0fd7 144int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
1da177e4 145
5c3a0fd7 146void ip4_datagram_release_cb(struct sock *sk);
8141ed9f 147
1da177e4
LT
148struct ip_reply_arg {
149 struct kvec iov[1];
88ef4a5a 150 int flags;
d6f5493c 151 __wsum csum;
1da177e4
LT
152 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
153 /* -1 if not needed */
f0e48dbf 154 int bound_dev_if;
66b13d99 155 u8 tos;
1da177e4
LT
156};
157
88ef4a5a
KK
158#define IP_REPLY_ARG_NOSRCCHECK 1
159
86b08d86
KK
160static 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
be9f4a44 165void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr,
70e73416
DM
166 __be32 saddr, const struct ip_reply_arg *arg,
167 unsigned int len);
1da177e4 168
fd2c3ef7 169struct ipv4_config {
1da177e4 170 int log_martians;
1da177e4
LT
171 int no_pmtu_disc;
172};
173
174extern struct ipv4_config ipv4_config;
4ce3c183
ED
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)
61a7e260
PE
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)
1da177e4 186
5c3a0fd7 187unsigned long snmp_fold_field(void __percpu *mib[], int offt);
4ce3c183 188#if BITS_PER_LONG==32
5c3a0fd7 189u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
4ce3c183
ED
190#else
191static 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
5c3a0fd7 196int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
5a17a390
CW
197
198static inline void snmp_mib_free(void __percpu *ptr[SNMP_ARRAY_SZ])
199{
200 int i;
201
202 BUG_ON(ptr == NULL);
203 for (i = 0; i < SNMP_ARRAY_SZ; i++) {
204 free_percpu(ptr[i]);
205 ptr[i] = NULL;
206 }
207}
33490170 208
3c689b73
ED
209extern struct local_ports {
210 seqlock_t lock;
211 int range[2];
212} sysctl_local_ports;
5c3a0fd7 213void inet_get_local_port_range(int *low, int *high);
227b60f5 214
e3826f1e
AW
215extern unsigned long *sysctl_local_reserved_ports;
216static inline int inet_is_reserved_local_port(int port)
217{
218 return test_bit(port, sysctl_local_reserved_ports);
219}
220
cdac4e07 221extern int sysctl_ip_nonlocal_bind;
1da177e4 222
20380731
ACM
223/* From inetpeer.c */
224extern int inet_peer_threshold;
225extern int inet_peer_minttl;
226extern int inet_peer_maxttl;
20380731 227
6648bd7e
AD
228/* From ip_input.c */
229extern int sysctl_ip_early_demux;
230
20380731
ACM
231/* From ip_output.c */
232extern int sysctl_ip_dynaddr;
233
5c3a0fd7 234void ipfrag_init(void);
20380731 235
5c3a0fd7 236void ip_static_sysctl_init(void);
bd7b1533 237
d18cd551
DM
238static inline bool ip_is_fragment(const struct iphdr *iph)
239{
240 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
241}
242
1da177e4 243#ifdef CONFIG_INET
14c85021
ACM
244#include <net/dst.h>
245
1da177e4
LT
246/* The function in 2.2 was invalid, producing wrong result for
247 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
248static inline
249int ip_decrease_ttl(struct iphdr *iph)
250{
5c78f275
AV
251 u32 check = (__force u32)iph->check;
252 check += (__force u32)htons(0x0100);
253 iph->check = (__force __sum16)(check + (check>=0xFFFF));
1da177e4
LT
254 return --iph->ttl;
255}
256
257static inline
258int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
259{
a02cec21 260 return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
1da177e4 261 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
a02cec21 262 !(dst_metric_locked(dst, RTAX_MTU)));
1da177e4
LT
263}
264
5c3a0fd7 265void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
1da177e4 266
703133de 267static inline void ip_select_ident(struct sk_buff *skb, struct dst_entry *dst, struct sock *sk)
1da177e4 268{
703133de
AA
269 struct iphdr *iph = ip_hdr(skb);
270
271 if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
1da177e4
LT
272 /* This is only to work around buggy Windows95/2000
273 * VJ compression implementations. If the ID field
274 * does not change, they drop every other packet in
275 * a TCP stream using header compression.
276 */
c720c7e8
ED
277 iph->id = (sk && inet_sk(sk)->inet_daddr) ?
278 htons(inet_sk(sk)->inet_id++) : 0;
1da177e4
LT
279 } else
280 __ip_select_ident(iph, dst, 0);
281}
282
703133de 283static inline void ip_select_ident_more(struct sk_buff *skb, struct dst_entry *dst, struct sock *sk, int more)
1da177e4 284{
703133de
AA
285 struct iphdr *iph = ip_hdr(skb);
286
287 if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
c720c7e8
ED
288 if (sk && inet_sk(sk)->inet_daddr) {
289 iph->id = htons(inet_sk(sk)->inet_id);
290 inet_sk(sk)->inet_id += 1 + more;
1da177e4
LT
291 } else
292 iph->id = 0;
293 } else
294 __ip_select_ident(iph, dst, more);
295}
296
297/*
298 * Map a multicast IP onto multicast MAC for type ethernet.
299 */
300
714e85be 301static inline void ip_eth_mc_map(__be32 naddr, char *buf)
1da177e4 302{
714e85be 303 __u32 addr=ntohl(naddr);
1da177e4
LT
304 buf[0]=0x01;
305 buf[1]=0x00;
306 buf[2]=0x5e;
307 buf[5]=addr&0xFF;
308 addr>>=8;
309 buf[4]=addr&0xFF;
310 addr>>=8;
311 buf[3]=addr&0x7F;
312}
313
314/*
315 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
316 * Leave P_Key as 0 to be filled in by driver.
317 */
318
a9e527e3 319static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
1da177e4 320{
714e85be 321 __u32 addr;
a9e527e3
RM
322 unsigned char scope = broadcast[5] & 0xF;
323
1da177e4
LT
324 buf[0] = 0; /* Reserved */
325 buf[1] = 0xff; /* Multicast QPN */
326 buf[2] = 0xff;
327 buf[3] = 0xff;
714e85be 328 addr = ntohl(naddr);
1da177e4 329 buf[4] = 0xff;
a9e527e3 330 buf[5] = 0x10 | scope; /* scope from broadcast address */
1da177e4
LT
331 buf[6] = 0x40; /* IPv4 signature */
332 buf[7] = 0x1b;
a9e527e3
RM
333 buf[8] = broadcast[8]; /* P_Key */
334 buf[9] = broadcast[9];
1da177e4
LT
335 buf[10] = 0;
336 buf[11] = 0;
337 buf[12] = 0;
338 buf[13] = 0;
339 buf[14] = 0;
340 buf[15] = 0;
341 buf[19] = addr & 0xff;
342 addr >>= 8;
343 buf[18] = addr & 0xff;
344 addr >>= 8;
345 buf[17] = addr & 0xff;
346 addr >>= 8;
347 buf[16] = addr & 0x0f;
348}
349
93ca3bb5
TT
350static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
351{
352 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
353 memcpy(buf, broadcast, 4);
354 else
355 memcpy(buf, &naddr, sizeof(naddr));
356}
357
dfd56b8b 358#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
359#include <linux/ipv6.h>
360#endif
361
362static __inline__ void inet_reset_saddr(struct sock *sk)
363{
c720c7e8 364 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
dfd56b8b 365#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
366 if (sk->sk_family == PF_INET6) {
367 struct ipv6_pinfo *np = inet6_sk(sk);
368
369 memset(&np->saddr, 0, sizeof(np->saddr));
370 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
371 }
372#endif
373}
374
375#endif
376
7ad6848c
OP
377static inline int sk_mc_loop(struct sock *sk)
378{
379 if (!sk)
380 return 1;
381 switch (sk->sk_family) {
382 case AF_INET:
383 return inet_sk(sk)->mc_loop;
dfd56b8b 384#if IS_ENABLED(CONFIG_IPV6)
7ad6848c
OP
385 case AF_INET6:
386 return inet6_sk(sk)->mc_loop;
387#endif
388 }
cd65c3c7 389 WARN_ON(1);
7ad6848c
OP
390 return 1;
391}
392
5c3a0fd7 393bool ip_call_ra_chain(struct sk_buff *skb);
1da177e4
LT
394
395/*
b798232f 396 * Functions provided by ip_fragment.c
1da177e4
LT
397 */
398
fd2c3ef7 399enum ip_defrag_users {
1da177e4
LT
400 IP_DEFRAG_LOCAL_DELIVER,
401 IP_DEFRAG_CALL_RA_CHAIN,
402 IP_DEFRAG_CONNTRACK_IN,
4be929be 403 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
1da177e4 404 IP_DEFRAG_CONNTRACK_OUT,
4be929be 405 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
8fa9ff68 406 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
4be929be 407 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
1da177e4
LT
408 IP_DEFRAG_VS_IN,
409 IP_DEFRAG_VS_OUT,
595fc71b
DM
410 IP_DEFRAG_VS_FWD,
411 IP_DEFRAG_AF_PACKET,
bc416d97 412 IP_DEFRAG_MACVLAN,
1da177e4
LT
413};
414
776c729e 415int ip_defrag(struct sk_buff *skb, u32 user);
bc416d97
ED
416#ifdef CONFIG_INET
417struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
418#else
419static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
420{
421 return skb;
422}
423#endif
6ddc0822 424int ip_frag_mem(struct net *net);
e5a2bb84 425int ip_frag_nqueues(struct net *net);
1da177e4
LT
426
427/*
428 * Functions provided by ip_forward.c
429 */
430
5c3a0fd7 431int ip_forward(struct sk_buff *skb);
1da177e4
LT
432
433/*
434 * Functions provided by ip_options.c
435 */
436
5c3a0fd7
JP
437void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
438 __be32 daddr, struct rtable *rt, int is_frag);
439int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
440void ip_options_fragment(struct sk_buff *skb);
441int ip_options_compile(struct net *net, struct ip_options *opt,
442 struct sk_buff *skb);
443int ip_options_get(struct net *net, struct ip_options_rcu **optp,
444 unsigned char *data, int optlen);
445int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
446 unsigned char __user *data, int optlen);
447void ip_options_undo(struct ip_options *opt);
448void ip_forward_options(struct sk_buff *skb);
449int ip_options_rcv_srr(struct sk_buff *skb);
1da177e4
LT
450
451/*
452 * Functions provided by ip_sockglue.c
453 */
454
5c3a0fd7
JP
455void ipv4_pktinfo_prepare(struct sk_buff *skb);
456void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
457int ip_cmsg_send(struct net *net, struct msghdr *msg, struct ipcm_cookie *ipc);
458int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
459 unsigned int optlen);
460int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
461 int __user *optlen);
462int compat_ip_setsockopt(struct sock *sk, int level, int optname,
463 char __user *optval, unsigned int optlen);
464int compat_ip_getsockopt(struct sock *sk, int level, int optname,
465 char __user *optval, int __user *optlen);
466int ip_ra_control(struct sock *sk, unsigned char on,
467 void (*destructor)(struct sock *));
468
469int ip_recv_error(struct sock *sk, struct msghdr *msg, int len);
470void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
471 u32 info, u8 *payload);
472void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
473 u32 info);
1da177e4 474
20380731 475#ifdef CONFIG_PROC_FS
5c3a0fd7 476int ip_misc_proc_init(void);
20380731
ACM
477#endif
478
1da177e4 479#endif /* _IP_H */