]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/blame - include/net/ip.h
ethernet: use likely() for common Ethernet encap
[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>
aa661581 31#include <net/route.h>
1da177e4 32#include <net/snmp.h>
86b08d86 33#include <net/flow.h>
1da177e4
LT
34
35struct sock;
36
fd2c3ef7 37struct inet_skb_parm {
1da177e4
LT
38 struct ip_options opt; /* Compiled IP options */
39 unsigned char flags;
40
8cdfab8a
PM
41#define IPSKB_FORWARDED 1
42#define IPSKB_XFRM_TUNNEL_SIZE 2
3e3850e9
PM
43#define IPSKB_XFRM_TRANSFORMED 4
44#define IPSKB_FRAG_COMPLETE 8
48d5cad8 45#define IPSKB_REROUTED 16
5f2d04f1
PM
46
47 u16 frag_max_size;
1da177e4
LT
48};
49
c9bdd4b5
ACM
50static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
51{
eddc9ec5 52 return ip_hdr(skb)->ihl * 4;
c9bdd4b5
ACM
53}
54
fd2c3ef7 55struct ipcm_cookie {
c1d18f9f 56 __be32 addr;
1da177e4 57 int oif;
f6d8bd05 58 struct ip_options_rcu *opt;
2244d07b 59 __u8 tx_flags;
f02db315
FF
60 __u8 ttl;
61 __s16 tos;
62 char priority;
1da177e4
LT
63};
64
65#define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
66
fd2c3ef7 67struct ip_ra_chain {
43a951e9 68 struct ip_ra_chain __rcu *next;
1da177e4 69 struct sock *sk;
592fcb9d
ED
70 union {
71 void (*destructor)(struct sock *);
72 struct sock *saved_sk;
73 };
66018506 74 struct rcu_head rcu;
1da177e4
LT
75};
76
43a951e9 77extern struct ip_ra_chain __rcu *ip_ra_chain;
1da177e4
LT
78
79/* IP flags. */
80#define IP_CE 0x8000 /* Flag: "Congestion" */
81#define IP_DF 0x4000 /* Flag: "Don't Fragment" */
82#define IP_MF 0x2000 /* Flag: "More Fragments" */
83#define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
84
85#define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
86
14c85021
ACM
87struct msghdr;
88struct net_device;
89struct packet_type;
90struct rtable;
14c85021
ACM
91struct sockaddr;
92
5c3a0fd7 93int igmp_mc_proc_init(void);
1da177e4
LT
94
95/*
96 * Functions provided by ip.c
97 */
98
5c3a0fd7
JP
99int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
100 __be32 saddr, __be32 daddr,
101 struct ip_options_rcu *opt);
102int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
103 struct net_device *orig_dev);
104int ip_local_deliver(struct sk_buff *skb);
105int ip_mr_input(struct sk_buff *skb);
106int ip_output(struct sk_buff *skb);
107int ip_mc_output(struct sk_buff *skb);
108int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
109int ip_do_nat(struct sk_buff *skb);
110void ip_send_check(struct iphdr *ip);
111int __ip_local_out(struct sk_buff *skb);
112int ip_local_out(struct sk_buff *skb);
113int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
114void ip_init(void);
115int ip_append_data(struct sock *sk, struct flowi4 *fl4,
116 int getfrag(void *from, char *to, int offset, int len,
117 int odd, struct sk_buff *skb),
118 void *from, int len, int protolen,
119 struct ipcm_cookie *ipc,
120 struct rtable **rt,
121 unsigned int flags);
122int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
123 struct sk_buff *skb);
124ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
125 int offset, size_t size, int flags);
126struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
127 struct sk_buff_head *queue,
128 struct inet_cork *cork);
129int ip_send_skb(struct net *net, struct sk_buff *skb);
130int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
131void ip_flush_pending_frames(struct sock *sk);
132struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
133 int getfrag(void *from, char *to, int offset,
134 int len, int odd, struct sk_buff *skb),
135 void *from, int length, int transhdrlen,
136 struct ipcm_cookie *ipc, struct rtable **rtp,
137 unsigned int flags);
1c32c5ad 138
77968b78 139static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
1c32c5ad 140{
77968b78 141 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
1c32c5ad 142}
1da177e4 143
aa661581
FF
144static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
145{
146 return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
147}
148
149static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
150{
151 return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
152}
153
1da177e4 154/* datagram.c */
5c3a0fd7 155int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
1da177e4 156
5c3a0fd7 157void ip4_datagram_release_cb(struct sock *sk);
8141ed9f 158
1da177e4
LT
159struct ip_reply_arg {
160 struct kvec iov[1];
88ef4a5a 161 int flags;
d6f5493c 162 __wsum csum;
1da177e4
LT
163 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
164 /* -1 if not needed */
f0e48dbf 165 int bound_dev_if;
66b13d99 166 u8 tos;
1da177e4
LT
167};
168
88ef4a5a
KK
169#define IP_REPLY_ARG_NOSRCCHECK 1
170
86b08d86
KK
171static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
172{
173 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
174}
175
be9f4a44 176void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr,
70e73416
DM
177 __be32 saddr, const struct ip_reply_arg *arg,
178 unsigned int len);
1da177e4 179
fd2c3ef7 180struct ipv4_config {
1da177e4 181 int log_martians;
1da177e4
LT
182 int no_pmtu_disc;
183};
184
185extern struct ipv4_config ipv4_config;
4ce3c183
ED
186#define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
187#define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
188#define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
189#define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
190#define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
191#define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
61a7e260
PE
192#define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
193#define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
194#define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
195#define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
196#define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
1da177e4 197
5c3a0fd7 198unsigned long snmp_fold_field(void __percpu *mib[], int offt);
4ce3c183 199#if BITS_PER_LONG==32
5c3a0fd7 200u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
4ce3c183
ED
201#else
202static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
203{
204 return snmp_fold_field(mib, offt);
205}
206#endif
5c3a0fd7 207int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
5a17a390
CW
208
209static inline void snmp_mib_free(void __percpu *ptr[SNMP_ARRAY_SZ])
210{
211 int i;
212
213 BUG_ON(ptr == NULL);
214 for (i = 0; i < SNMP_ARRAY_SZ; i++) {
215 free_percpu(ptr[i]);
216 ptr[i] = NULL;
217 }
218}
33490170 219
3c689b73
ED
220extern struct local_ports {
221 seqlock_t lock;
222 int range[2];
223} sysctl_local_ports;
5c3a0fd7 224void inet_get_local_port_range(int *low, int *high);
227b60f5 225
e3826f1e
AW
226extern unsigned long *sysctl_local_reserved_ports;
227static inline int inet_is_reserved_local_port(int port)
228{
229 return test_bit(port, sysctl_local_reserved_ports);
230}
231
cdac4e07 232extern int sysctl_ip_nonlocal_bind;
1da177e4 233
20380731
ACM
234/* From inetpeer.c */
235extern int inet_peer_threshold;
236extern int inet_peer_minttl;
237extern int inet_peer_maxttl;
20380731 238
6648bd7e
AD
239/* From ip_input.c */
240extern int sysctl_ip_early_demux;
241
20380731
ACM
242/* From ip_output.c */
243extern int sysctl_ip_dynaddr;
244
5c3a0fd7 245void ipfrag_init(void);
20380731 246
5c3a0fd7 247void ip_static_sysctl_init(void);
bd7b1533 248
d18cd551
DM
249static inline bool ip_is_fragment(const struct iphdr *iph)
250{
251 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
252}
253
1da177e4 254#ifdef CONFIG_INET
14c85021
ACM
255#include <net/dst.h>
256
1da177e4
LT
257/* The function in 2.2 was invalid, producing wrong result for
258 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
259static inline
260int ip_decrease_ttl(struct iphdr *iph)
261{
5c78f275
AV
262 u32 check = (__force u32)iph->check;
263 check += (__force u32)htons(0x0100);
264 iph->check = (__force __sum16)(check + (check>=0xFFFF));
1da177e4
LT
265 return --iph->ttl;
266}
267
268static inline
269int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
270{
a02cec21 271 return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
1da177e4 272 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
a02cec21 273 !(dst_metric_locked(dst, RTAX_MTU)));
1da177e4
LT
274}
275
5c3a0fd7 276void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
1da177e4 277
703133de 278static inline void ip_select_ident(struct sk_buff *skb, struct dst_entry *dst, struct sock *sk)
1da177e4 279{
703133de
AA
280 struct iphdr *iph = ip_hdr(skb);
281
282 if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
1da177e4
LT
283 /* This is only to work around buggy Windows95/2000
284 * VJ compression implementations. If the ID field
285 * does not change, they drop every other packet in
286 * a TCP stream using header compression.
287 */
c720c7e8
ED
288 iph->id = (sk && inet_sk(sk)->inet_daddr) ?
289 htons(inet_sk(sk)->inet_id++) : 0;
1da177e4
LT
290 } else
291 __ip_select_ident(iph, dst, 0);
292}
293
703133de 294static inline void ip_select_ident_more(struct sk_buff *skb, struct dst_entry *dst, struct sock *sk, int more)
1da177e4 295{
703133de
AA
296 struct iphdr *iph = ip_hdr(skb);
297
298 if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
c720c7e8
ED
299 if (sk && inet_sk(sk)->inet_daddr) {
300 iph->id = htons(inet_sk(sk)->inet_id);
301 inet_sk(sk)->inet_id += 1 + more;
1da177e4
LT
302 } else
303 iph->id = 0;
304 } else
305 __ip_select_ident(iph, dst, more);
306}
307
308/*
309 * Map a multicast IP onto multicast MAC for type ethernet.
310 */
311
714e85be 312static inline void ip_eth_mc_map(__be32 naddr, char *buf)
1da177e4 313{
714e85be 314 __u32 addr=ntohl(naddr);
1da177e4
LT
315 buf[0]=0x01;
316 buf[1]=0x00;
317 buf[2]=0x5e;
318 buf[5]=addr&0xFF;
319 addr>>=8;
320 buf[4]=addr&0xFF;
321 addr>>=8;
322 buf[3]=addr&0x7F;
323}
324
325/*
326 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
327 * Leave P_Key as 0 to be filled in by driver.
328 */
329
a9e527e3 330static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
1da177e4 331{
714e85be 332 __u32 addr;
a9e527e3
RM
333 unsigned char scope = broadcast[5] & 0xF;
334
1da177e4
LT
335 buf[0] = 0; /* Reserved */
336 buf[1] = 0xff; /* Multicast QPN */
337 buf[2] = 0xff;
338 buf[3] = 0xff;
714e85be 339 addr = ntohl(naddr);
1da177e4 340 buf[4] = 0xff;
a9e527e3 341 buf[5] = 0x10 | scope; /* scope from broadcast address */
1da177e4
LT
342 buf[6] = 0x40; /* IPv4 signature */
343 buf[7] = 0x1b;
a9e527e3
RM
344 buf[8] = broadcast[8]; /* P_Key */
345 buf[9] = broadcast[9];
1da177e4
LT
346 buf[10] = 0;
347 buf[11] = 0;
348 buf[12] = 0;
349 buf[13] = 0;
350 buf[14] = 0;
351 buf[15] = 0;
352 buf[19] = addr & 0xff;
353 addr >>= 8;
354 buf[18] = addr & 0xff;
355 addr >>= 8;
356 buf[17] = addr & 0xff;
357 addr >>= 8;
358 buf[16] = addr & 0x0f;
359}
360
93ca3bb5
TT
361static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
362{
363 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
364 memcpy(buf, broadcast, 4);
365 else
366 memcpy(buf, &naddr, sizeof(naddr));
367}
368
dfd56b8b 369#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
370#include <linux/ipv6.h>
371#endif
372
373static __inline__ void inet_reset_saddr(struct sock *sk)
374{
c720c7e8 375 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
dfd56b8b 376#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
377 if (sk->sk_family == PF_INET6) {
378 struct ipv6_pinfo *np = inet6_sk(sk);
379
380 memset(&np->saddr, 0, sizeof(np->saddr));
381 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
382 }
383#endif
384}
385
386#endif
387
7ad6848c
OP
388static inline int sk_mc_loop(struct sock *sk)
389{
390 if (!sk)
391 return 1;
392 switch (sk->sk_family) {
393 case AF_INET:
394 return inet_sk(sk)->mc_loop;
dfd56b8b 395#if IS_ENABLED(CONFIG_IPV6)
7ad6848c
OP
396 case AF_INET6:
397 return inet6_sk(sk)->mc_loop;
398#endif
399 }
cd65c3c7 400 WARN_ON(1);
7ad6848c
OP
401 return 1;
402}
403
5c3a0fd7 404bool ip_call_ra_chain(struct sk_buff *skb);
1da177e4
LT
405
406/*
b798232f 407 * Functions provided by ip_fragment.c
1da177e4
LT
408 */
409
fd2c3ef7 410enum ip_defrag_users {
1da177e4
LT
411 IP_DEFRAG_LOCAL_DELIVER,
412 IP_DEFRAG_CALL_RA_CHAIN,
413 IP_DEFRAG_CONNTRACK_IN,
4be929be 414 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
1da177e4 415 IP_DEFRAG_CONNTRACK_OUT,
4be929be 416 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
8fa9ff68 417 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
4be929be 418 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
1da177e4
LT
419 IP_DEFRAG_VS_IN,
420 IP_DEFRAG_VS_OUT,
595fc71b
DM
421 IP_DEFRAG_VS_FWD,
422 IP_DEFRAG_AF_PACKET,
bc416d97 423 IP_DEFRAG_MACVLAN,
1da177e4
LT
424};
425
776c729e 426int ip_defrag(struct sk_buff *skb, u32 user);
bc416d97
ED
427#ifdef CONFIG_INET
428struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
429#else
430static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
431{
432 return skb;
433}
434#endif
6ddc0822 435int ip_frag_mem(struct net *net);
e5a2bb84 436int ip_frag_nqueues(struct net *net);
1da177e4
LT
437
438/*
439 * Functions provided by ip_forward.c
440 */
441
5c3a0fd7 442int ip_forward(struct sk_buff *skb);
1da177e4
LT
443
444/*
445 * Functions provided by ip_options.c
446 */
447
5c3a0fd7
JP
448void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
449 __be32 daddr, struct rtable *rt, int is_frag);
450int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
451void ip_options_fragment(struct sk_buff *skb);
452int ip_options_compile(struct net *net, struct ip_options *opt,
453 struct sk_buff *skb);
454int ip_options_get(struct net *net, struct ip_options_rcu **optp,
455 unsigned char *data, int optlen);
456int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
457 unsigned char __user *data, int optlen);
458void ip_options_undo(struct ip_options *opt);
459void ip_forward_options(struct sk_buff *skb);
460int ip_options_rcv_srr(struct sk_buff *skb);
1da177e4
LT
461
462/*
463 * Functions provided by ip_sockglue.c
464 */
465
5c3a0fd7
JP
466void ipv4_pktinfo_prepare(struct sk_buff *skb);
467void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
468int ip_cmsg_send(struct net *net, struct msghdr *msg, struct ipcm_cookie *ipc);
469int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
470 unsigned int optlen);
471int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
472 int __user *optlen);
473int compat_ip_setsockopt(struct sock *sk, int level, int optname,
474 char __user *optval, unsigned int optlen);
475int compat_ip_getsockopt(struct sock *sk, int level, int optname,
476 char __user *optval, int __user *optlen);
477int ip_ra_control(struct sock *sk, unsigned char on,
478 void (*destructor)(struct sock *));
479
480int ip_recv_error(struct sock *sk, struct msghdr *msg, int len);
481void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
482 u32 info, u8 *payload);
483void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
484 u32 info);
1da177e4 485
20380731 486#ifdef CONFIG_PROC_FS
5c3a0fd7 487int ip_misc_proc_init(void);
20380731
ACM
488#endif
489
1da177e4 490#endif /* _IP_H */