blob: 8175e3ecd741e75920ce56e342770a1ed1a36e01 [file] [log] [blame]
lh9ed821d2023-04-07 01:36:19 -07001/*
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
34struct sock;
35
36struct inet_skb_parm {
37 struct ip_options opt; /* Compiled IP options */
38 unsigned char flags;
39
40#define IPSKB_FORWARDED BIT(0)
41#define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
42#define IPSKB_XFRM_TRANSFORMED BIT(2)
43#define IPSKB_FRAG_COMPLETE BIT(3)
44#define IPSKB_REROUTED BIT(4)
45#define IPSKB_DOREDIRECT BIT(5)
46
47};
48
49static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
50{
51 return ip_hdr(skb)->ihl * 4;
52}
53
54struct 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
63struct 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
73extern 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
83struct msghdr;
84struct net_device;
85struct packet_type;
86struct rtable;
87struct sockaddr;
88
89extern int igmp_mc_proc_init(void);
90
91/*
92 * Functions provided by ip.c
93 */
94
95extern int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
96 __be32 saddr, __be32 daddr,
97 struct ip_options_rcu *opt);
98extern int ip_rcv(struct sk_buff *skb, struct net_device *dev,
99 struct packet_type *pt, struct net_device *orig_dev);
100extern int ip_local_deliver(struct sk_buff *skb);
101extern int ip_mr_input(struct sk_buff *skb);
102extern int ip_output(struct sk_buff *skb);
103extern int ip_mc_output(struct sk_buff *skb);
104extern int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
105extern int ip_do_nat(struct sk_buff *skb);
106extern void ip_send_check(struct iphdr *ip);
107extern int __ip_local_out(struct sk_buff *skb);
108extern int ip_local_out(struct sk_buff *skb);
109extern int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
110extern void ip_init(void);
111extern 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);
118extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb);
119extern ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
120 int offset, size_t size, int flags);
121extern struct sk_buff *__ip_make_skb(struct sock *sk,
122 struct flowi4 *fl4,
123 struct sk_buff_head *queue,
124 struct inet_cork *cork);
125extern int ip_send_skb(struct sk_buff *skb);
126extern int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
127extern void ip_flush_pending_frames(struct sock *sk);
128extern 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
137static 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 */
143extern int ip4_datagram_connect(struct sock *sk,
144 struct sockaddr *uaddr, int addr_len);
145
146/*
147 * Map a multicast IP onto multicast MAC for type Token Ring.
148 * This conforms to RFC1469 Option 2 Multicasting i.e.
149 * using a functional address to transmit / receive
150 * multicast packets.
151 */
152
153static inline void ip_tr_mc_map(__be32 addr, char *buf)
154{
155 buf[0]=0xC0;
156 buf[1]=0x00;
157 buf[2]=0x00;
158 buf[3]=0x04;
159 buf[4]=0x00;
160 buf[5]=0x00;
161}
162
163struct ip_reply_arg {
164 struct kvec iov[1];
165 int flags;
166 __wsum csum;
167 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
168 /* -1 if not needed */
169 int bound_dev_if;
170 u8 tos;
171};
172
173#define IP_REPLY_ARG_NOSRCCHECK 1
174
175static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
176{
177 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
178}
179
180void ip_send_reply(struct sock *sk, struct sk_buff *skb, __be32 daddr,
181 const struct ip_reply_arg *arg, unsigned int len);
182
183struct ipv4_config {
184 int log_martians;
185 int no_pmtu_disc;
186};
187
188extern struct ipv4_config ipv4_config;
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_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
199#define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
200
201extern unsigned long snmp_fold_field(void __percpu *mib[], int offt);
202#if BITS_PER_LONG==32
203extern u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
204#else
205static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
206{
207 return snmp_fold_field(mib, offt);
208}
209#endif
210extern int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
211extern void snmp_mib_free(void __percpu *ptr[2]);
212
213extern struct local_ports {
214 seqlock_t lock;
215 int range[2];
216} sysctl_local_ports;
217extern void inet_get_local_port_range(int *low, int *high);
218
219extern unsigned long *sysctl_local_reserved_ports;
220static inline int inet_is_reserved_local_port(int port)
221{
222 return test_bit(port, sysctl_local_reserved_ports);
223}
224
225extern int sysctl_ip_nonlocal_bind;
226
227extern struct ctl_path net_core_path[];
228extern struct ctl_path net_ipv4_ctl_path[];
229
230/* From inetpeer.c */
231extern int inet_peer_threshold;
232extern int inet_peer_minttl;
233extern int inet_peer_maxttl;
234
235/* From ip_output.c */
236extern int sysctl_ip_dynaddr;
237
238extern void ipfrag_init(void);
239
240extern void ip_static_sysctl_init(void);
241
242static inline bool ip_is_fragment(const struct iphdr *iph)
243{
244 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
245}
246
247#ifdef CONFIG_INET
248#include <net/dst.h>
249
250/* The function in 2.2 was invalid, producing wrong result for
251 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
252static inline
253int ip_decrease_ttl(struct iphdr *iph)
254{
255 u32 check = (__force u32)iph->check;
256 check += (__force u32)htons(0x0100);
257 iph->check = (__force __sum16)(check + (check>=0xFFFF));
258 return --iph->ttl;
259}
260
261static inline
262int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
263{
264 return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
265 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
266 !(dst_metric_locked(dst, RTAX_MTU)));
267}
268
269u32 ip_idents_reserve(u32 hash, int segs);
270void __ip_select_ident(struct iphdr *iph, int segs);
271
272static inline void ip_select_ident_segs(struct sk_buff *skb, struct sock *sk, int segs)
273{
274 struct iphdr *iph = ip_hdr(skb);
275
276 if ((iph->frag_off & htons(IP_DF)) && !skb->local_df) {
277 /* This is only to work around buggy Windows95/2000
278 * VJ compression implementations. If the ID field
279 * does not change, they drop every other packet in
280 * a TCP stream using header compression.
281 */
282 if (sk && inet_sk(sk)->inet_daddr) {
283 iph->id = htons(inet_sk(sk)->inet_id);
284 inet_sk(sk)->inet_id += segs;
285 } else {
286 iph->id = 0;
287 }
288 } else {
289 __ip_select_ident(iph, segs);
290 }
291}
292
293static inline void ip_select_ident(struct sk_buff *skb, struct sock *sk)
294{
295 ip_select_ident_segs(skb, sk, 1);
296}
297
298/*
299 * Map a multicast IP onto multicast MAC for type ethernet.
300 */
301
302static inline void ip_eth_mc_map(__be32 naddr, char *buf)
303{
304 __u32 addr=ntohl(naddr);
305 buf[0]=0x01;
306 buf[1]=0x00;
307 buf[2]=0x5e;
308 buf[5]=addr&0xFF;
309 addr>>=8;
310 buf[4]=addr&0xFF;
311 addr>>=8;
312 buf[3]=addr&0x7F;
313}
314
315/*
316 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
317 * Leave P_Key as 0 to be filled in by driver.
318 */
319
320static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
321{
322 __u32 addr;
323 unsigned char scope = broadcast[5] & 0xF;
324
325 buf[0] = 0; /* Reserved */
326 buf[1] = 0xff; /* Multicast QPN */
327 buf[2] = 0xff;
328 buf[3] = 0xff;
329 addr = ntohl(naddr);
330 buf[4] = 0xff;
331 buf[5] = 0x10 | scope; /* scope from broadcast address */
332 buf[6] = 0x40; /* IPv4 signature */
333 buf[7] = 0x1b;
334 buf[8] = broadcast[8]; /* P_Key */
335 buf[9] = broadcast[9];
336 buf[10] = 0;
337 buf[11] = 0;
338 buf[12] = 0;
339 buf[13] = 0;
340 buf[14] = 0;
341 buf[15] = 0;
342 buf[19] = addr & 0xff;
343 addr >>= 8;
344 buf[18] = addr & 0xff;
345 addr >>= 8;
346 buf[17] = addr & 0xff;
347 addr >>= 8;
348 buf[16] = addr & 0x0f;
349}
350
351static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
352{
353 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
354 memcpy(buf, broadcast, 4);
355 else
356 memcpy(buf, &naddr, sizeof(naddr));
357}
358
359#if IS_ENABLED(CONFIG_IPV6)
360#include <linux/ipv6.h>
361#endif
362
363static __inline__ void inet_reset_saddr(struct sock *sk)
364{
365 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
366#if IS_ENABLED(CONFIG_IPV6)
367 if (sk->sk_family == PF_INET6) {
368 struct ipv6_pinfo *np = inet6_sk(sk);
369
370 memset(&np->saddr, 0, sizeof(np->saddr));
371 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
372 }
373#endif
374}
375
376#endif
377
378static inline int sk_mc_loop(struct sock *sk)
379{
380 if (!sk)
381 return 1;
382 switch (sk->sk_family) {
383 case AF_INET:
384 return inet_sk(sk)->mc_loop;
385#if IS_ENABLED(CONFIG_IPV6)
386 case AF_INET6:
387 return inet6_sk(sk)->mc_loop;
388#endif
389 }
390 WARN_ON(1);
391 return 1;
392}
393
394extern bool ip_call_ra_chain(struct sk_buff *skb);
395
396/*
397 * Functions provided by ip_fragment.c
398 */
399
400enum ip_defrag_users {
401 IP_DEFRAG_LOCAL_DELIVER,
402 IP_DEFRAG_CALL_RA_CHAIN,
403 IP_DEFRAG_CONNTRACK_IN,
404 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
405 IP_DEFRAG_CONNTRACK_OUT,
406 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
407 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
408 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
409 IP_DEFRAG_VS_IN,
410 IP_DEFRAG_VS_OUT,
411 IP_DEFRAG_VS_FWD,
412 IP_DEFRAG_AF_PACKET,
413 IP_DEFRAG_MACVLAN,
414};
415
416int ip_defrag(struct sk_buff *skb, u32 user);
417#ifdef CONFIG_INET
418struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
419#else
420static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
421{
422 return skb;
423}
424#endif
425int ip_frag_mem(struct net *net);
426int ip_frag_nqueues(struct net *net);
427
428/*
429 * Functions provided by ip_forward.c
430 */
431
432extern int ip_forward(struct sk_buff *skb);
433
434/*
435 * Functions provided by ip_options.c
436 */
437
438extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
439 __be32 daddr, struct rtable *rt, int is_frag);
440extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
441extern void ip_options_fragment(struct sk_buff *skb);
442extern int ip_options_compile(struct net *net,
443 struct ip_options *opt, struct sk_buff *skb);
444extern int ip_options_get(struct net *net, struct ip_options_rcu **optp,
445 unsigned char *data, int optlen);
446extern int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
447 unsigned char __user *data, int optlen);
448extern void ip_options_undo(struct ip_options * opt);
449extern void ip_forward_options(struct sk_buff *skb);
450extern int ip_options_rcv_srr(struct sk_buff *skb);
451
452/*
453 * Functions provided by ip_sockglue.c
454 */
455
456extern void ipv4_pktinfo_prepare(struct sk_buff *skb);
457extern void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
458extern int ip_cmsg_send(struct net *net,
459 struct msghdr *msg, struct ipcm_cookie *ipc);
460extern int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen);
461extern int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);
462extern int compat_ip_setsockopt(struct sock *sk, int level,
463 int optname, char __user *optval, unsigned int optlen);
464extern int compat_ip_getsockopt(struct sock *sk, int level,
465 int optname, char __user *optval, int __user *optlen);
466extern int ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *));
467
468extern int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
469extern void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
470 __be16 port, u32 info, u8 *payload);
471extern void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
472 u32 info);
473
474#ifdef CONFIG_PROC_FS
475extern int ip_misc_proc_init(void);
476#endif
477
478#endif /* _IP_H */