]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame - include/net/ip.h
packet: Protect packet sk list with mutex (v2)
[mirror_ubuntu-artful-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
1da177e4
LT
45};
46
c9bdd4b5
ACM
47static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
48{
eddc9ec5 49 return ip_hdr(skb)->ihl * 4;
c9bdd4b5
ACM
50}
51
fd2c3ef7 52struct ipcm_cookie {
c1d18f9f 53 __be32 addr;
1da177e4 54 int oif;
f6d8bd05 55 struct ip_options_rcu *opt;
2244d07b 56 __u8 tx_flags;
1da177e4
LT
57};
58
59#define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
60
fd2c3ef7 61struct ip_ra_chain {
43a951e9 62 struct ip_ra_chain __rcu *next;
1da177e4 63 struct sock *sk;
592fcb9d
ED
64 union {
65 void (*destructor)(struct sock *);
66 struct sock *saved_sk;
67 };
66018506 68 struct rcu_head rcu;
1da177e4
LT
69};
70
43a951e9 71extern struct ip_ra_chain __rcu *ip_ra_chain;
1da177e4
LT
72
73/* IP flags. */
74#define IP_CE 0x8000 /* Flag: "Congestion" */
75#define IP_DF 0x4000 /* Flag: "Don't Fragment" */
76#define IP_MF 0x2000 /* Flag: "More Fragments" */
77#define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
78
79#define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
80
14c85021
ACM
81struct msghdr;
82struct net_device;
83struct packet_type;
84struct rtable;
14c85021
ACM
85struct sockaddr;
86
1da177e4
LT
87extern int igmp_mc_proc_init(void);
88
89/*
90 * Functions provided by ip.c
91 */
92
93extern int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
13d8eaa0 94 __be32 saddr, __be32 daddr,
f6d8bd05 95 struct ip_options_rcu *opt);
1da177e4 96extern int ip_rcv(struct sk_buff *skb, struct net_device *dev,
f2ccd8fa 97 struct packet_type *pt, struct net_device *orig_dev);
1da177e4
LT
98extern int ip_local_deliver(struct sk_buff *skb);
99extern int ip_mr_input(struct sk_buff *skb);
100extern int ip_output(struct sk_buff *skb);
101extern int ip_mc_output(struct sk_buff *skb);
2e2f7aef 102extern int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
1da177e4
LT
103extern int ip_do_nat(struct sk_buff *skb);
104extern void ip_send_check(struct iphdr *ip);
c439cb2e
HX
105extern int __ip_local_out(struct sk_buff *skb);
106extern int ip_local_out(struct sk_buff *skb);
d9d8da80 107extern int ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
1da177e4 108extern void ip_init(void);
f5fca608 109extern int ip_append_data(struct sock *sk, struct flowi4 *fl4,
1da177e4
LT
110 int getfrag(void *from, char *to, int offset, int len,
111 int odd, struct sk_buff *skb),
112 void *from, int len, int protolen,
113 struct ipcm_cookie *ipc,
2e77d89b 114 struct rtable **rt,
1da177e4
LT
115 unsigned int flags);
116extern int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb);
f5fca608 117extern ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
1da177e4 118 int offset, size_t size, int flags);
1c32c5ad 119extern struct sk_buff *__ip_make_skb(struct sock *sk,
77968b78 120 struct flowi4 *fl4,
1c32c5ad
HX
121 struct sk_buff_head *queue,
122 struct inet_cork *cork);
b5ec8eea 123extern int ip_send_skb(struct net *net, struct sk_buff *skb);
77968b78 124extern int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
1da177e4 125extern void ip_flush_pending_frames(struct sock *sk);
1c32c5ad 126extern struct sk_buff *ip_make_skb(struct sock *sk,
77968b78 127 struct flowi4 *fl4,
1c32c5ad
HX
128 int getfrag(void *from, char *to, int offset, int len,
129 int odd, struct sk_buff *skb),
130 void *from, int length, int transhdrlen,
131 struct ipcm_cookie *ipc,
132 struct rtable **rtp,
133 unsigned int flags);
134
77968b78 135static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
1c32c5ad 136{
77968b78 137 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
1c32c5ad 138}
1da177e4
LT
139
140/* datagram.c */
141extern int ip4_datagram_connect(struct sock *sk,
142 struct sockaddr *uaddr, int addr_len);
143
1da177e4
LT
144struct ip_reply_arg {
145 struct kvec iov[1];
88ef4a5a 146 int flags;
d6f5493c 147 __wsum csum;
1da177e4
LT
148 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
149 /* -1 if not needed */
f0e48dbf 150 int bound_dev_if;
66b13d99 151 u8 tos;
1da177e4
LT
152};
153
88ef4a5a
KK
154#define IP_REPLY_ARG_NOSRCCHECK 1
155
86b08d86
KK
156static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
157{
158 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
159}
160
be9f4a44 161void ip_send_unicast_reply(struct net *net, struct sk_buff *skb, __be32 daddr,
70e73416
DM
162 __be32 saddr, const struct ip_reply_arg *arg,
163 unsigned int len);
1da177e4 164
fd2c3ef7 165struct ipv4_config {
1da177e4 166 int log_martians;
1da177e4
LT
167 int no_pmtu_disc;
168};
169
170extern struct ipv4_config ipv4_config;
4ce3c183
ED
171#define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
172#define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
173#define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
174#define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
175#define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
176#define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
61a7e260
PE
177#define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
178#define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
179#define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
180#define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
181#define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
1da177e4 182
7d720c3e 183extern unsigned long snmp_fold_field(void __percpu *mib[], int offt);
4ce3c183
ED
184#if BITS_PER_LONG==32
185extern u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
186#else
187static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
188{
189 return snmp_fold_field(mib, offt);
190}
191#endif
1823e4c8 192extern int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
7d720c3e 193extern void snmp_mib_free(void __percpu *ptr[2]);
33490170 194
3c689b73
ED
195extern struct local_ports {
196 seqlock_t lock;
197 int range[2];
198} sysctl_local_ports;
227b60f5
SH
199extern void inet_get_local_port_range(int *low, int *high);
200
e3826f1e
AW
201extern unsigned long *sysctl_local_reserved_ports;
202static inline int inet_is_reserved_local_port(int port)
203{
204 return test_bit(port, sysctl_local_reserved_ports);
205}
206
cdac4e07 207extern int sysctl_ip_nonlocal_bind;
1da177e4 208
20380731
ACM
209/* From inetpeer.c */
210extern int inet_peer_threshold;
211extern int inet_peer_minttl;
212extern int inet_peer_maxttl;
20380731 213
6648bd7e
AD
214/* From ip_input.c */
215extern int sysctl_ip_early_demux;
216
20380731
ACM
217/* From ip_output.c */
218extern int sysctl_ip_dynaddr;
219
220extern void ipfrag_init(void);
221
bd7b1533
AV
222extern void ip_static_sysctl_init(void);
223
d18cd551
DM
224static inline bool ip_is_fragment(const struct iphdr *iph)
225{
226 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
227}
228
1da177e4 229#ifdef CONFIG_INET
14c85021
ACM
230#include <net/dst.h>
231
1da177e4
LT
232/* The function in 2.2 was invalid, producing wrong result for
233 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
234static inline
235int ip_decrease_ttl(struct iphdr *iph)
236{
5c78f275
AV
237 u32 check = (__force u32)iph->check;
238 check += (__force u32)htons(0x0100);
239 iph->check = (__force __sum16)(check + (check>=0xFFFF));
1da177e4
LT
240 return --iph->ttl;
241}
242
243static inline
244int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
245{
a02cec21 246 return inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
1da177e4 247 (inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
a02cec21 248 !(dst_metric_locked(dst, RTAX_MTU)));
1da177e4
LT
249}
250
251extern void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
252
253static inline void ip_select_ident(struct iphdr *iph, struct dst_entry *dst, struct sock *sk)
254{
255 if (iph->frag_off & htons(IP_DF)) {
256 /* This is only to work around buggy Windows95/2000
257 * VJ compression implementations. If the ID field
258 * does not change, they drop every other packet in
259 * a TCP stream using header compression.
260 */
c720c7e8
ED
261 iph->id = (sk && inet_sk(sk)->inet_daddr) ?
262 htons(inet_sk(sk)->inet_id++) : 0;
1da177e4
LT
263 } else
264 __ip_select_ident(iph, dst, 0);
265}
266
267static inline void ip_select_ident_more(struct iphdr *iph, struct dst_entry *dst, struct sock *sk, int more)
268{
269 if (iph->frag_off & htons(IP_DF)) {
c720c7e8
ED
270 if (sk && inet_sk(sk)->inet_daddr) {
271 iph->id = htons(inet_sk(sk)->inet_id);
272 inet_sk(sk)->inet_id += 1 + more;
1da177e4
LT
273 } else
274 iph->id = 0;
275 } else
276 __ip_select_ident(iph, dst, more);
277}
278
279/*
280 * Map a multicast IP onto multicast MAC for type ethernet.
281 */
282
714e85be 283static inline void ip_eth_mc_map(__be32 naddr, char *buf)
1da177e4 284{
714e85be 285 __u32 addr=ntohl(naddr);
1da177e4
LT
286 buf[0]=0x01;
287 buf[1]=0x00;
288 buf[2]=0x5e;
289 buf[5]=addr&0xFF;
290 addr>>=8;
291 buf[4]=addr&0xFF;
292 addr>>=8;
293 buf[3]=addr&0x7F;
294}
295
296/*
297 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
298 * Leave P_Key as 0 to be filled in by driver.
299 */
300
a9e527e3 301static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
1da177e4 302{
714e85be 303 __u32 addr;
a9e527e3
RM
304 unsigned char scope = broadcast[5] & 0xF;
305
1da177e4
LT
306 buf[0] = 0; /* Reserved */
307 buf[1] = 0xff; /* Multicast QPN */
308 buf[2] = 0xff;
309 buf[3] = 0xff;
714e85be 310 addr = ntohl(naddr);
1da177e4 311 buf[4] = 0xff;
a9e527e3 312 buf[5] = 0x10 | scope; /* scope from broadcast address */
1da177e4
LT
313 buf[6] = 0x40; /* IPv4 signature */
314 buf[7] = 0x1b;
a9e527e3
RM
315 buf[8] = broadcast[8]; /* P_Key */
316 buf[9] = broadcast[9];
1da177e4
LT
317 buf[10] = 0;
318 buf[11] = 0;
319 buf[12] = 0;
320 buf[13] = 0;
321 buf[14] = 0;
322 buf[15] = 0;
323 buf[19] = addr & 0xff;
324 addr >>= 8;
325 buf[18] = addr & 0xff;
326 addr >>= 8;
327 buf[17] = addr & 0xff;
328 addr >>= 8;
329 buf[16] = addr & 0x0f;
330}
331
93ca3bb5
TT
332static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
333{
334 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
335 memcpy(buf, broadcast, 4);
336 else
337 memcpy(buf, &naddr, sizeof(naddr));
338}
339
dfd56b8b 340#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
341#include <linux/ipv6.h>
342#endif
343
344static __inline__ void inet_reset_saddr(struct sock *sk)
345{
c720c7e8 346 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
dfd56b8b 347#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
348 if (sk->sk_family == PF_INET6) {
349 struct ipv6_pinfo *np = inet6_sk(sk);
350
351 memset(&np->saddr, 0, sizeof(np->saddr));
352 memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
353 }
354#endif
355}
356
357#endif
358
7ad6848c
OP
359static inline int sk_mc_loop(struct sock *sk)
360{
361 if (!sk)
362 return 1;
363 switch (sk->sk_family) {
364 case AF_INET:
365 return inet_sk(sk)->mc_loop;
dfd56b8b 366#if IS_ENABLED(CONFIG_IPV6)
7ad6848c
OP
367 case AF_INET6:
368 return inet6_sk(sk)->mc_loop;
369#endif
370 }
cd65c3c7 371 WARN_ON(1);
7ad6848c
OP
372 return 1;
373}
374
ba57b4db 375extern bool ip_call_ra_chain(struct sk_buff *skb);
1da177e4
LT
376
377/*
b798232f 378 * Functions provided by ip_fragment.c
1da177e4
LT
379 */
380
fd2c3ef7 381enum ip_defrag_users {
1da177e4
LT
382 IP_DEFRAG_LOCAL_DELIVER,
383 IP_DEFRAG_CALL_RA_CHAIN,
384 IP_DEFRAG_CONNTRACK_IN,
4be929be 385 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
1da177e4 386 IP_DEFRAG_CONNTRACK_OUT,
4be929be 387 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
8fa9ff68 388 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
4be929be 389 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
1da177e4
LT
390 IP_DEFRAG_VS_IN,
391 IP_DEFRAG_VS_OUT,
595fc71b
DM
392 IP_DEFRAG_VS_FWD,
393 IP_DEFRAG_AF_PACKET,
bc416d97 394 IP_DEFRAG_MACVLAN,
1da177e4
LT
395};
396
776c729e 397int ip_defrag(struct sk_buff *skb, u32 user);
bc416d97
ED
398#ifdef CONFIG_INET
399struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
400#else
401static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
402{
403 return skb;
404}
405#endif
6ddc0822 406int ip_frag_mem(struct net *net);
e5a2bb84 407int ip_frag_nqueues(struct net *net);
1da177e4
LT
408
409/*
410 * Functions provided by ip_forward.c
411 */
412
413extern int ip_forward(struct sk_buff *skb);
1da177e4
LT
414
415/*
416 * Functions provided by ip_options.c
417 */
418
f6d8bd05
ED
419extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
420 __be32 daddr, struct rtable *rt, int is_frag);
1da177e4
LT
421extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
422extern void ip_options_fragment(struct sk_buff *skb);
0e6bd4a1
DL
423extern int ip_options_compile(struct net *net,
424 struct ip_options *opt, struct sk_buff *skb);
f6d8bd05 425extern int ip_options_get(struct net *net, struct ip_options_rcu **optp,
4c6ea29d 426 unsigned char *data, int optlen);
f6d8bd05 427extern int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
4c6ea29d 428 unsigned char __user *data, int optlen);
1da177e4
LT
429extern void ip_options_undo(struct ip_options * opt);
430extern void ip_forward_options(struct sk_buff *skb);
431extern int ip_options_rcv_srr(struct sk_buff *skb);
432
433/*
434 * Functions provided by ip_sockglue.c
435 */
436
d826eb14 437extern void ipv4_pktinfo_prepare(struct sk_buff *skb);
1da177e4 438extern void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
7a6adb92
DL
439extern int ip_cmsg_send(struct net *net,
440 struct msghdr *msg, struct ipcm_cookie *ipc);
b7058842 441extern int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen);
1da177e4 442extern int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);
3fdadf7d 443extern int compat_ip_setsockopt(struct sock *sk, int level,
b7058842 444 int optname, char __user *optval, unsigned int optlen);
3fdadf7d
DM
445extern int compat_ip_getsockopt(struct sock *sk, int level,
446 int optname, char __user *optval, int __user *optlen);
1da177e4
LT
447extern int ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *));
448
449extern int ip_recv_error(struct sock *sk, struct msghdr *msg, int len);
450extern void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
35986b32 451 __be16 port, u32 info, u8 *payload);
0579016e 452extern void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
1da177e4
LT
453 u32 info);
454
20380731
ACM
455#ifdef CONFIG_PROC_FS
456extern int ip_misc_proc_init(void);
457#endif
458
1da177e4 459#endif /* _IP_H */