blob: 462182f78236caed327c12e5080aefd5f73fe966 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -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 *
Jesper Juhl02c30a82005-05-05 16:16:16 -070010 * Authors: Ross Biro
Linus Torvalds1da177e2005-04-16 15:20:36 -070011 * 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
Linus Torvalds1da177e2005-04-16 15:20:36 -070025#include <linux/types.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070026#include <linux/ip.h>
27#include <linux/in.h>
Arnaldo Carvalho de Meloc9bdd4b2007-03-12 20:09:15 -030028#include <linux/skbuff.h>
Martin KaFai Lauf0b1e642017-12-01 12:52:30 -080029#include <linux/jhash.h>
Arnaldo Carvalho de Melo14c85022005-12-27 02:43:12 -020030
31#include <net/inet_sock.h>
Francesco Fuscoaa661582013-09-24 15:43:09 +020032#include <net/route.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070033#include <net/snmp.h>
KOVACS Krisztian86b08d82008-10-01 07:44:42 -070034#include <net/flow.h>
Jiri Pirko1bd758e2015-05-12 14:56:07 +020035#include <net/flow_dissector.h>
Martin KaFai Lauf0b1e642017-12-01 12:52:30 -080036#include <net/netns/hash.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070037
Gao Feng7ed14d92017-04-12 12:34:03 +080038#define IPV4_MAX_PMTU 65535U /* RFC 2675, Section 5.1 */
Eric Dumazetb5476022017-12-11 07:17:39 -080039#define IPV4_MIN_MTU 68 /* RFC 791 */
Gao Feng7ed14d92017-04-12 12:34:03 +080040
Linus Torvalds1da177e2005-04-16 15:20:36 -070041struct sock;
42
Eric Dumazetfd2c3ef2009-11-03 03:26:03 +000043struct inet_skb_parm {
David Ahern0b922b72016-05-10 11:19:51 -070044 int iif;
Linus Torvalds1da177e2005-04-16 15:20:36 -070045 struct ip_options opt; /* Compiled IP options */
David Aherna04a4802016-10-16 20:02:52 -070046 u16 flags;
Linus Torvalds1da177e2005-04-16 15:20:36 -070047
Hannes Frederic Sowadf4d9252015-01-23 12:01:26 +010048#define IPSKB_FORWARDED BIT(0)
49#define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
50#define IPSKB_XFRM_TRANSFORMED BIT(2)
51#define IPSKB_FRAG_COMPLETE BIT(3)
52#define IPSKB_REROUTED BIT(4)
53#define IPSKB_DOREDIRECT BIT(5)
Florian Westphald6b915e2015-05-22 16:32:51 +020054#define IPSKB_FRAG_PMTU BIT(6)
Lance Richardson9ee6c5d2016-11-02 16:36:17 -040055#define IPSKB_L3SLAVE BIT(7)
Patrick McHardy5f2d04f2012-08-26 19:13:55 +020056
57 u16 frag_max_size;
Linus Torvalds1da177e2005-04-16 15:20:36 -070058};
59
David Aherna04a4802016-10-16 20:02:52 -070060static inline bool ipv4_l3mdev_skb(u16 flags)
61{
62 return !!(flags & IPSKB_L3SLAVE);
63}
64
Arnaldo Carvalho de Meloc9bdd4b2007-03-12 20:09:15 -030065static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
66{
Arnaldo Carvalho de Meloeddc9ec2007-04-20 22:47:35 -070067 return ip_hdr(skb)->ihl * 4;
Arnaldo Carvalho de Meloc9bdd4b2007-03-12 20:09:15 -030068}
69
Eric Dumazetfd2c3ef2009-11-03 03:26:03 +000070struct ipcm_cookie {
Soheil Hassas Yeganeh24025c42016-04-02 23:08:10 -040071 struct sockcm_cookie sockc;
Al Viroc1d18f92006-09-27 18:28:28 -070072 __be32 addr;
Linus Torvalds1da177e2005-04-16 15:20:36 -070073 int oif;
Eric Dumazetf6d8bd02011-04-21 09:45:37 +000074 struct ip_options_rcu *opt;
Francesco Fuscof02db312013-09-24 15:43:08 +020075 __u8 ttl;
76 __s16 tos;
77 char priority;
Willem de Bruijnbec1f6f2018-04-26 13:42:17 -040078 __u16 gso_size;
Linus Torvalds1da177e2005-04-16 15:20:36 -070079};
80
Willem de Bruijn35178202018-07-06 10:12:54 -040081static inline void ipcm_init(struct ipcm_cookie *ipcm)
82{
83 *ipcm = (struct ipcm_cookie) { .tos = -1 };
84}
85
86static inline void ipcm_init_sk(struct ipcm_cookie *ipcm,
87 const struct inet_sock *inet)
88{
89 ipcm_init(ipcm);
90
91 ipcm->sockc.tsflags = inet->sk.sk_tsflags;
92 ipcm->oif = inet->sk.sk_bound_dev_if;
93 ipcm->addr = inet->inet_saddr;
94}
95
Linus Torvalds1da177e2005-04-16 15:20:36 -070096#define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
Hannes Frederic Sowa4b261c752014-01-20 03:43:08 +010097#define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
Linus Torvalds1da177e2005-04-16 15:20:36 -070098
David Ahernfb74c272017-08-07 08:44:16 -070099/* return enslaved device index if relevant */
100static inline int inet_sdif(struct sk_buff *skb)
101{
102#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
103 if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags))
104 return IPCB(skb)->iif;
105#endif
106 return 0;
107}
108
Kirill Tkhai5796ef72018-03-22 12:45:32 +0300109/* Special input handler for packets caught by router alert option.
110 They are selected only by protocol field, and then processed likely
111 local ones; but only if someone wants them! Otherwise, router
112 not running rsvpd will kill RSVP.
113
114 It is user level problem, what it will make with them.
115 I have no idea, how it will masquearde or NAT them (it is joke, joke :-)),
116 but receiver should be enough clever f.e. to forward mtrace requests,
117 sent to multicast group to reach destination designated router.
118 */
119
Eric Dumazetfd2c3ef2009-11-03 03:26:03 +0000120struct ip_ra_chain {
Eric Dumazet43a951e2010-10-25 03:32:44 +0000121 struct ip_ra_chain __rcu *next;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700122 struct sock *sk;
Eric Dumazet592fcb92010-06-09 16:21:07 +0000123 union {
124 void (*destructor)(struct sock *);
125 struct sock *saved_sk;
126 };
Eric Dumazet66018502010-06-07 03:12:08 +0000127 struct rcu_head rcu;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700128};
129
Linus Torvalds1da177e2005-04-16 15:20:36 -0700130/* IP flags. */
131#define IP_CE 0x8000 /* Flag: "Congestion" */
132#define IP_DF 0x4000 /* Flag: "Don't Fragment" */
133#define IP_MF 0x2000 /* Flag: "More Fragments" */
134#define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
135
136#define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
137
Arnaldo Carvalho de Melo14c85022005-12-27 02:43:12 -0200138struct msghdr;
139struct net_device;
140struct packet_type;
141struct rtable;
Arnaldo Carvalho de Melo14c85022005-12-27 02:43:12 -0200142struct sockaddr;
143
WANG Cong72c1d3b2014-01-10 16:09:45 -0800144int igmp_mc_init(void);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700145
146/*
147 * Functions provided by ip.c
148 */
149
Eric Dumazetcfe673b2015-09-25 07:39:16 -0700150int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700151 __be32 saddr, __be32 daddr,
152 struct ip_options_rcu *opt);
153int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
154 struct net_device *orig_dev);
Edward Cree17266ee2018-07-02 16:14:12 +0100155void ip_list_rcv(struct list_head *head, struct packet_type *pt,
156 struct net_device *orig_dev);
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700157int ip_local_deliver(struct sk_buff *skb);
158int ip_mr_input(struct sk_buff *skb);
Eric W. Biedermanede20592015-10-07 16:48:47 -0500159int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb);
160int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb);
Eric W. Biederman694869b2015-06-12 21:55:31 -0500161int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
162 int (*output)(struct net *, struct sock *, struct sk_buff *));
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700163void ip_send_check(struct iphdr *ip);
Eric W. Biedermancf91a992015-10-07 16:48:45 -0500164int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
Eric W. Biederman33224b12015-10-07 16:48:46 -0500165int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
Eric Dumazetaad88722014-04-15 13:47:15 -0400166
Xin Long69b9e1e2018-07-02 18:21:11 +0800167int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl,
168 __u8 tos);
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700169void ip_init(void);
170int ip_append_data(struct sock *sk, struct flowi4 *fl4,
171 int getfrag(void *from, char *to, int offset, int len,
172 int odd, struct sk_buff *skb),
173 void *from, int len, int protolen,
174 struct ipcm_cookie *ipc,
175 struct rtable **rt,
176 unsigned int flags);
177int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
178 struct sk_buff *skb);
179ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
180 int offset, size_t size, int flags);
181struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
182 struct sk_buff_head *queue,
183 struct inet_cork *cork);
184int ip_send_skb(struct net *net, struct sk_buff *skb);
185int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
186void ip_flush_pending_frames(struct sock *sk);
187struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
188 int getfrag(void *from, char *to, int offset,
189 int len, int odd, struct sk_buff *skb),
190 void *from, int length, int transhdrlen,
191 struct ipcm_cookie *ipc, struct rtable **rtp,
Willem de Bruijn1cd78842018-04-26 13:42:15 -0400192 struct inet_cork *cork, unsigned int flags);
Herbert Xu1c32c5a2011-03-01 02:36:47 +0000193
Xin Long69b9e1e2018-07-02 18:21:11 +0800194static inline int ip_queue_xmit(struct sock *sk, struct sk_buff *skb,
195 struct flowi *fl)
196{
197 return __ip_queue_xmit(sk, skb, fl, inet_sk(sk)->tos);
198}
199
David S. Miller77968b72011-05-08 17:12:19 -0700200static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
Herbert Xu1c32c5a2011-03-01 02:36:47 +0000201{
David S. Miller77968b72011-05-08 17:12:19 -0700202 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
Herbert Xu1c32c5a2011-03-01 02:36:47 +0000203}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700204
Francesco Fuscoaa661582013-09-24 15:43:09 +0200205static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
206{
207 return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
208}
209
210static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
211{
212 return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
213}
214
Linus Torvalds1da177e2005-04-16 15:20:36 -0700215/* datagram.c */
Eric Dumazet03645a12015-07-14 08:10:22 +0200216int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700217int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700218
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700219void ip4_datagram_release_cb(struct sock *sk);
Steffen Klassert8141ed92013-01-21 02:00:03 +0000220
Linus Torvalds1da177e2005-04-16 15:20:36 -0700221struct ip_reply_arg {
Stephen Hemminger82695b32018-02-27 15:48:21 -0800222 struct kvec iov[1];
KOVACS Krisztian88ef4a52008-10-01 07:41:00 -0700223 int flags;
Al Virod6f5493c2006-11-14 21:26:08 -0800224 __wsum csum;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700225 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
Stephen Hemminger82695b32018-02-27 15:48:21 -0800226 /* -1 if not needed */
Patrick McHardyf0e48db2007-06-04 21:32:46 -0700227 int bound_dev_if;
Eric Dumazet66b13d92011-10-24 03:06:21 -0400228 u8 tos;
Lorenzo Colittie2d118a2016-11-04 02:23:43 +0900229 kuid_t uid;
Stephen Hemminger82695b32018-02-27 15:48:21 -0800230};
Linus Torvalds1da177e2005-04-16 15:20:36 -0700231
KOVACS Krisztian88ef4a52008-10-01 07:41:00 -0700232#define IP_REPLY_ARG_NOSRCCHECK 1
233
KOVACS Krisztian86b08d82008-10-01 07:44:42 -0700234static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
235{
236 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
237}
238
Eric Dumazetbdbbb852015-01-29 21:35:05 -0800239void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
Eric Dumazet24a2d432014-09-27 09:50:55 -0700240 const struct ip_options *sopt,
241 __be32 daddr, __be32 saddr,
242 const struct ip_reply_arg *arg,
David S. Miller70e73412012-06-28 03:21:41 -0700243 unsigned int len);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700244
Eric Dumazet4ce3c182010-06-30 13:31:19 -0700245#define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
Eric Dumazet13415e42016-04-27 16:44:43 -0700246#define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field)
Eric Dumazet4ce3c182010-06-30 13:31:19 -0700247#define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
Eric Dumazet13415e42016-04-27 16:44:43 -0700248#define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
Eric Dumazet4ce3c182010-06-30 13:31:19 -0700249#define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
Eric Dumazet13415e42016-04-27 16:44:43 -0700250#define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
Pavel Emelyanov61a7e262008-07-18 04:03:08 -0700251#define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
Eric Dumazet13415e42016-04-27 16:44:43 -0700252#define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field)
David S. Millerf7324ac2014-03-06 15:03:17 -0500253#define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
Eric Dumazet13415e42016-04-27 16:44:43 -0700254#define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700255
Raghavendra K Tc4c6bc32015-08-30 11:29:41 +0530256u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offct);
WANG Cong698365f2014-05-05 15:55:55 -0700257unsigned long snmp_fold_field(void __percpu *mib, int offt);
Eric Dumazet4ce3c182010-06-30 13:31:19 -0700258#if BITS_PER_LONG==32
Raghavendra K Tc4c6bc32015-08-30 11:29:41 +0530259u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
260 size_t syncp_offset);
WANG Cong698365f2014-05-05 15:55:55 -0700261u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
Eric Dumazet4ce3c182010-06-30 13:31:19 -0700262#else
Raghavendra K Tc4c6bc32015-08-30 11:29:41 +0530263static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
264 size_t syncp_offset)
265{
266 return snmp_get_cpu_field(mib, cpu, offct);
267
268}
269
WANG Cong698365f2014-05-05 15:55:55 -0700270static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
Eric Dumazet4ce3c182010-06-30 13:31:19 -0700271{
272 return snmp_fold_field(mib, offt);
273}
274#endif
YOSHIFUJI Hideaki33490172007-04-20 15:57:15 -0700275
Jia He6348ef22016-09-30 11:28:58 +0800276#define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \
277{ \
278 int i, c; \
279 for_each_possible_cpu(c) { \
280 for (i = 0; stats_list[i].name; i++) \
281 buff64[i] += snmp_get_cpu_field64( \
282 mib_statistic, \
283 c, stats_list[i].entry, \
284 offset); \
285 } \
286}
287
288#define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
289{ \
290 int i, c; \
291 for_each_possible_cpu(c) { \
292 for (i = 0; stats_list[i].name; i++) \
293 buff[i] += snmp_get_cpu_field( \
294 mib_statistic, \
295 c, stats_list[i].entry); \
296 } \
297}
298
Eric W. Biederman0bbf87d2013-09-28 14:10:59 -0700299void inet_get_local_port_range(struct net *net, int *low, int *high);
Stephen Hemminger227b60f2007-10-10 17:30:46 -0700300
David S. Millerfcd77db2014-05-15 13:43:14 -0400301#ifdef CONFIG_SYSCTL
WANG Cong122ff242014-05-12 16:04:53 -0700302static inline int inet_is_local_reserved_port(struct net *net, int port)
Amerigo Wange3826f12010-05-05 00:27:06 +0000303{
WANG Cong122ff242014-05-12 16:04:53 -0700304 if (!net->ipv4.sysctl_local_reserved_ports)
305 return 0;
306 return test_bit(port, net->ipv4.sysctl_local_reserved_ports);
Amerigo Wange3826f12010-05-05 00:27:06 +0000307}
WANG Cong20e61da2014-07-25 15:25:08 -0700308
309static inline bool sysctl_dev_name_is_allowed(const char *name)
310{
311 return strcmp(name, "default") != 0 && strcmp(name, "all") != 0;
312}
313
Krister Johansen4548b682017-01-20 17:49:11 -0800314static inline int inet_prot_sock(struct net *net)
315{
316 return net->ipv4.sysctl_ip_prot_sock;
317}
318
WANG Cong122ff242014-05-12 16:04:53 -0700319#else
320static inline int inet_is_local_reserved_port(struct net *net, int port)
321{
322 return 0;
323}
Krister Johansen4548b682017-01-20 17:49:11 -0800324
325static inline int inet_prot_sock(struct net *net)
326{
327 return PROT_SOCK;
328}
WANG Cong122ff242014-05-12 16:04:53 -0700329#endif
Amerigo Wange3826f12010-05-05 00:27:06 +0000330
Deepa Dinamani822c8682016-02-27 00:32:15 -0800331__be32 inet_current_timestamp(void);
332
Arnaldo Carvalho de Melo20380732005-08-16 02:18:02 -0300333/* From inetpeer.c */
334extern int inet_peer_threshold;
335extern int inet_peer_minttl;
336extern int inet_peer_maxttl;
Arnaldo Carvalho de Melo20380732005-08-16 02:18:02 -0300337
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700338void ipfrag_init(void);
Arnaldo Carvalho de Melo20380732005-08-16 02:18:02 -0300339
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700340void ip_static_sysctl_init(void);
Al Virobd7b1532008-07-15 16:00:59 -0400341
Lorenzo Colittie1108612014-05-13 10:17:33 -0700342#define IP4_REPLY_MARK(net, mark) \
343 ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
344
David S. Millerd18cd552011-06-23 21:28:52 -0700345static inline bool ip_is_fragment(const struct iphdr *iph)
346{
347 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
348}
349
Linus Torvalds1da177e2005-04-16 15:20:36 -0700350#ifdef CONFIG_INET
Arnaldo Carvalho de Melo14c85022005-12-27 02:43:12 -0200351#include <net/dst.h>
352
Linus Torvalds1da177e2005-04-16 15:20:36 -0700353/* The function in 2.2 was invalid, producing wrong result for
354 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
355static inline
356int ip_decrease_ttl(struct iphdr *iph)
357{
Al Viro5c78f272006-11-14 21:42:26 -0800358 u32 check = (__force u32)iph->check;
359 check += (__force u32)htons(0x0100);
360 iph->check = (__force __sum16)(check + (check>=0xFFFF));
Linus Torvalds1da177e2005-04-16 15:20:36 -0700361 return --iph->ttl;
362}
363
Sabrina Dubrocad52e5a72018-03-14 10:21:14 +0100364static inline int ip_mtu_locked(const struct dst_entry *dst)
365{
366 const struct rtable *rt = (const struct rtable *)dst;
367
368 return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU);
369}
370
Linus Torvalds1da177e2005-04-16 15:20:36 -0700371static inline
Eric Dumazet4e3f5d72015-09-25 07:39:14 -0700372int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700373{
Eric Dumazet4e3f5d72015-09-25 07:39:14 -0700374 u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc);
375
376 return pmtudisc == IP_PMTUDISC_DO ||
377 (pmtudisc == IP_PMTUDISC_WANT &&
Sabrina Dubrocad52e5a72018-03-14 10:21:14 +0100378 !ip_mtu_locked(dst));
Linus Torvalds1da177e2005-04-16 15:20:36 -0700379}
380
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100381static inline bool ip_sk_accept_pmtu(const struct sock *sk)
382{
Hannes Frederic Sowa1b346572014-02-26 01:20:42 +0100383 return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
384 inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100385}
386
387static inline bool ip_sk_use_pmtu(const struct sock *sk)
388{
389 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
390}
391
WANG Cong60ff7462014-05-04 16:39:18 -0700392static inline bool ip_sk_ignore_df(const struct sock *sk)
Hannes Frederic Sowa1b346572014-02-26 01:20:42 +0100393{
394 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
395 inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
396}
397
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100398static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
399 bool forwarding)
400{
401 struct net *net = dev_net(dst->dev);
402
403 if (net->ipv4.sysctl_ip_fwd_use_pmtu ||
Sabrina Dubrocad52e5a72018-03-14 10:21:14 +0100404 ip_mtu_locked(dst) ||
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100405 !forwarding)
406 return dst_mtu(dst);
407
Eric Dumazetc780a042017-08-16 11:09:12 -0700408 return min(READ_ONCE(dst->dev->mtu), IP_MAX_MTU);
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100409}
410
Shmulik Ladkanifedbb6b42016-06-29 21:47:03 +0300411static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
412 const struct sk_buff *skb)
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100413{
Eric Dumazetcaf3f262015-10-04 21:08:08 -0700414 if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) {
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100415 bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
Eric Dumazetcaf3f262015-10-04 21:08:08 -0700416
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100417 return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100418 }
Eric Dumazetcaf3f262015-10-04 21:08:08 -0700419
Eric Dumazetc780a042017-08-16 11:09:12 -0700420 return min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
Hannes Frederic Sowaf87c10a2014-01-09 10:01:15 +0100421}
422
David Ahern767a2212018-10-04 20:07:51 -0700423struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
David Ahernd7e774f2018-11-06 12:51:15 -0800424 int fc_mx_len,
425 struct netlink_ext_ack *extack);
David Aherncc5f0eb2018-10-04 20:07:52 -0700426static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics)
427{
428 if (fib_metrics != &dst_default_metrics &&
429 refcount_dec_and_test(&fib_metrics->refcnt))
430 kfree(fib_metrics);
431}
David Aherna9195252018-04-17 17:33:07 -0700432
David Aherne1255ed2018-10-04 20:07:53 -0700433/* ipv4 and ipv6 both use refcounted metrics if it is not the default */
434static inline
435void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics)
436{
437 dst_init_metrics(dst, fib_metrics->metrics, true);
438
439 if (fib_metrics != &dst_default_metrics) {
440 dst->_metrics |= DST_METRICS_REFCOUNTED;
441 refcount_inc(&fib_metrics->refcnt);
442 }
443}
444
David Ahern1620a332018-10-04 20:07:54 -0700445static inline
446void ip_dst_metrics_put(struct dst_entry *dst)
447{
448 struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst);
449
450 if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt))
451 kfree(p);
452}
453
Eric Dumazet04ca6972014-07-26 08:58:10 +0200454u32 ip_idents_reserve(u32 hash, int segs);
Hannes Frederic Sowab6a77192015-03-25 17:07:44 +0100455void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
Eric Dumazet73f156a2014-06-02 05:26:03 -0700456
Hannes Frederic Sowab6a77192015-03-25 17:07:44 +0100457static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb,
458 struct sock *sk, int segs)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700459{
Ansis Atteka703133d2013-09-18 15:29:53 -0700460 struct iphdr *iph = ip_hdr(skb);
461
WANG Cong60ff7462014-05-04 16:39:18 -0700462 if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700463 /* This is only to work around buggy Windows95/2000
464 * VJ compression implementations. If the ID field
465 * does not change, they drop every other packet in
466 * a TCP stream using header compression.
467 */
Eric Dumazetc720c7e82009-10-15 06:30:45 +0000468 if (sk && inet_sk(sk)->inet_daddr) {
469 iph->id = htons(inet_sk(sk)->inet_id);
Eric Dumazet73f156a2014-06-02 05:26:03 -0700470 inet_sk(sk)->inet_id += segs;
471 } else {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700472 iph->id = 0;
Eric Dumazet73f156a2014-06-02 05:26:03 -0700473 }
474 } else {
Hannes Frederic Sowab6a77192015-03-25 17:07:44 +0100475 __ip_select_ident(net, iph, segs);
Eric Dumazet73f156a2014-06-02 05:26:03 -0700476 }
477}
478
Hannes Frederic Sowab6a77192015-03-25 17:07:44 +0100479static inline void ip_select_ident(struct net *net, struct sk_buff *skb,
480 struct sock *sk)
Eric Dumazet73f156a2014-06-02 05:26:03 -0700481{
Hannes Frederic Sowab6a77192015-03-25 17:07:44 +0100482 ip_select_ident_segs(net, skb, sk, 1);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700483}
484
Tom Herberted70fcf2014-05-02 16:29:38 -0700485static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto)
486{
487 return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr,
488 skb->len, proto, 0);
489}
490
Tom Herbertc3f83242015-06-04 09:16:40 -0700491/* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store
492 * Equivalent to : flow->v4addrs.src = iph->saddr;
493 * flow->v4addrs.dst = iph->daddr;
494 */
495static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow,
496 const struct iphdr *iph)
497{
498 BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) !=
499 offsetof(typeof(flow->addrs), v4addrs.src) +
500 sizeof(flow->addrs.v4addrs.src));
501 memcpy(&flow->addrs.v4addrs, &iph->saddr, sizeof(flow->addrs.v4addrs));
502 flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
503}
504
Tom Herbert1933a782014-08-22 13:34:04 -0700505static inline __wsum inet_gro_compute_pseudo(struct sk_buff *skb, int proto)
506{
507 const struct iphdr *iph = skb_gro_network_header(skb);
508
509 return csum_tcpudp_nofold(iph->saddr, iph->daddr,
510 skb_gro_len(skb), proto, 0);
511}
512
Linus Torvalds1da177e2005-04-16 15:20:36 -0700513/*
514 * Map a multicast IP onto multicast MAC for type ethernet.
515 */
516
Al Viro714e85b2006-11-14 20:51:49 -0800517static inline void ip_eth_mc_map(__be32 naddr, char *buf)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700518{
Al Viro714e85b2006-11-14 20:51:49 -0800519 __u32 addr=ntohl(naddr);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700520 buf[0]=0x01;
521 buf[1]=0x00;
522 buf[2]=0x5e;
523 buf[5]=addr&0xFF;
524 addr>>=8;
525 buf[4]=addr&0xFF;
526 addr>>=8;
527 buf[3]=addr&0x7F;
528}
529
530/*
531 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
532 * Leave P_Key as 0 to be filled in by driver.
533 */
534
Rolf Manderscheida9e527e2007-12-10 13:38:41 -0700535static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700536{
Al Viro714e85b2006-11-14 20:51:49 -0800537 __u32 addr;
Rolf Manderscheida9e527e2007-12-10 13:38:41 -0700538 unsigned char scope = broadcast[5] & 0xF;
539
Linus Torvalds1da177e2005-04-16 15:20:36 -0700540 buf[0] = 0; /* Reserved */
541 buf[1] = 0xff; /* Multicast QPN */
542 buf[2] = 0xff;
543 buf[3] = 0xff;
Al Viro714e85b2006-11-14 20:51:49 -0800544 addr = ntohl(naddr);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700545 buf[4] = 0xff;
Rolf Manderscheida9e527e2007-12-10 13:38:41 -0700546 buf[5] = 0x10 | scope; /* scope from broadcast address */
Linus Torvalds1da177e2005-04-16 15:20:36 -0700547 buf[6] = 0x40; /* IPv4 signature */
548 buf[7] = 0x1b;
Rolf Manderscheida9e527e2007-12-10 13:38:41 -0700549 buf[8] = broadcast[8]; /* P_Key */
550 buf[9] = broadcast[9];
Linus Torvalds1da177e2005-04-16 15:20:36 -0700551 buf[10] = 0;
552 buf[11] = 0;
553 buf[12] = 0;
554 buf[13] = 0;
555 buf[14] = 0;
556 buf[15] = 0;
557 buf[19] = addr & 0xff;
558 addr >>= 8;
559 buf[18] = addr & 0xff;
560 addr >>= 8;
561 buf[17] = addr & 0xff;
562 addr >>= 8;
563 buf[16] = addr & 0x0f;
564}
565
Timo Teräs93ca3bb2011-03-28 22:40:53 +0000566static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
567{
568 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
569 memcpy(buf, broadcast, 4);
570 else
571 memcpy(buf, &naddr, sizeof(naddr));
572}
573
Eric Dumazetdfd56b82011-12-10 09:48:31 +0000574#if IS_ENABLED(CONFIG_IPV6)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700575#include <linux/ipv6.h>
576#endif
577
578static __inline__ void inet_reset_saddr(struct sock *sk)
579{
Eric Dumazetc720c7e82009-10-15 06:30:45 +0000580 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
Eric Dumazetdfd56b82011-12-10 09:48:31 +0000581#if IS_ENABLED(CONFIG_IPV6)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700582 if (sk->sk_family == PF_INET6) {
583 struct ipv6_pinfo *np = inet6_sk(sk);
584
585 memset(&np->saddr, 0, sizeof(np->saddr));
Eric Dumazetefe42082013-10-03 15:42:29 -0700586 memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
Linus Torvalds1da177e2005-04-16 15:20:36 -0700587 }
588#endif
589}
590
591#endif
592
David Ahern72afa352015-08-27 16:06:59 -0700593static inline unsigned int ipv4_addr_hash(__be32 ip)
594{
595 return (__force unsigned int) ip;
596}
597
Martin KaFai Lauf0b1e642017-12-01 12:52:30 -0800598static inline u32 ipv4_portaddr_hash(const struct net *net,
599 __be32 saddr,
600 unsigned int port)
601{
602 return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
603}
604
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700605bool ip_call_ra_chain(struct sk_buff *skb);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700606
607/*
Rami Rosenb7982322008-01-01 21:13:09 -0800608 * Functions provided by ip_fragment.c
Linus Torvalds1da177e2005-04-16 15:20:36 -0700609 */
610
Eric Dumazetfd2c3ef2009-11-03 03:26:03 +0000611enum ip_defrag_users {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700612 IP_DEFRAG_LOCAL_DELIVER,
613 IP_DEFRAG_CALL_RA_CHAIN,
614 IP_DEFRAG_CONNTRACK_IN,
Alexey Dobriyan4be929b2010-05-24 14:33:03 -0700615 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700616 IP_DEFRAG_CONNTRACK_OUT,
Alexey Dobriyan4be929b2010-05-24 14:33:03 -0700617 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
Patrick McHardy8fa9ff62009-12-15 16:59:59 +0100618 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
Alexey Dobriyan4be929b2010-05-24 14:33:03 -0700619 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700620 IP_DEFRAG_VS_IN,
621 IP_DEFRAG_VS_OUT,
David S. Miller595fc712011-07-05 01:05:48 -0700622 IP_DEFRAG_VS_FWD,
623 IP_DEFRAG_AF_PACKET,
Eric Dumazetbc416d92011-10-06 10:28:31 +0000624 IP_DEFRAG_MACVLAN,
Linus Torvalds1da177e2005-04-16 15:20:36 -0700625};
626
Andy Zhou5cf42282015-05-15 14:15:35 -0700627/* Return true if the value of 'user' is between 'lower_bond'
628 * and 'upper_bond' inclusively.
629 */
630static inline bool ip_defrag_user_in_between(u32 user,
631 enum ip_defrag_users lower_bond,
632 enum ip_defrag_users upper_bond)
633{
634 return user >= lower_bond && user <= upper_bond;
635}
636
Eric W. Biederman19bcf9f2015-10-09 13:44:54 -0500637int ip_defrag(struct net *net, struct sk_buff *skb, u32 user);
Eric Dumazetbc416d92011-10-06 10:28:31 +0000638#ifdef CONFIG_INET
Eric W. Biederman19bcf9f2015-10-09 13:44:54 -0500639struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user);
Eric Dumazetbc416d92011-10-06 10:28:31 +0000640#else
Eric W. Biederman19bcf9f2015-10-09 13:44:54 -0500641static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user)
Eric Dumazetbc416d92011-10-06 10:28:31 +0000642{
643 return skb;
644}
645#endif
Linus Torvalds1da177e2005-04-16 15:20:36 -0700646
647/*
648 * Functions provided by ip_forward.c
649 */
Stephen Hemminger82695b32018-02-27 15:48:21 -0800650
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700651int ip_forward(struct sk_buff *skb);
Stephen Hemminger82695b32018-02-27 15:48:21 -0800652
Linus Torvalds1da177e2005-04-16 15:20:36 -0700653/*
654 * Functions provided by ip_options.c
655 */
Stephen Hemminger82695b32018-02-27 15:48:21 -0800656
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700657void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
658 __be32 daddr, struct rtable *rt, int is_frag);
Eric Dumazet24a2d432014-09-27 09:50:55 -0700659
Paolo Abeni91ed1e62017-08-03 18:07:06 +0200660int __ip_options_echo(struct net *net, struct ip_options *dopt,
661 struct sk_buff *skb, const struct ip_options *sopt);
662static inline int ip_options_echo(struct net *net, struct ip_options *dopt,
663 struct sk_buff *skb)
Eric Dumazet24a2d432014-09-27 09:50:55 -0700664{
Paolo Abeni91ed1e62017-08-03 18:07:06 +0200665 return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt);
Eric Dumazet24a2d432014-09-27 09:50:55 -0700666}
667
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700668void ip_options_fragment(struct sk_buff *skb);
669int ip_options_compile(struct net *net, struct ip_options *opt,
670 struct sk_buff *skb);
671int ip_options_get(struct net *net, struct ip_options_rcu **optp,
672 unsigned char *data, int optlen);
673int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
674 unsigned char __user *data, int optlen);
675void ip_options_undo(struct ip_options *opt);
676void ip_forward_options(struct sk_buff *skb);
677int ip_options_rcv_srr(struct sk_buff *skb);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700678
679/*
680 * Functions provided by ip_sockglue.c
681 */
682
Shawn Bohrerfbf88662013-10-07 11:01:40 -0500683void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
Paolo Abeniad959032016-11-04 11:28:58 +0100684void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk,
685 struct sk_buff *skb, int tlen, int offset);
Soheil Hassas Yeganeh24025c42016-04-02 23:08:10 -0400686int ip_cmsg_send(struct sock *sk, struct msghdr *msg,
Hannes Frederic Sowac8e6ad02014-02-18 21:38:08 +0100687 struct ipcm_cookie *ipc, bool allow_ipv6);
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700688int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
689 unsigned int optlen);
690int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
691 int __user *optlen);
692int compat_ip_setsockopt(struct sock *sk, int level, int optname,
693 char __user *optval, unsigned int optlen);
694int compat_ip_getsockopt(struct sock *sk, int level, int optname,
695 char __user *optval, int __user *optlen);
696int ip_ra_control(struct sock *sk, unsigned char on,
697 void (*destructor)(struct sock *));
Linus Torvalds1da177e2005-04-16 15:20:36 -0700698
Hannes Frederic Sowa85fbaa72013-11-23 00:46:12 +0100699int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700700void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
701 u32 info, u8 *payload);
702void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
703 u32 info);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700704
Tom Herbert5961de92015-01-05 13:56:16 -0800705static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
706{
Paolo Abeniad959032016-11-04 11:28:58 +0100707 ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0);
Tom Herbert5961de92015-01-05 13:56:16 -0800708}
709
Eric Dumazet4cdf5072014-09-19 07:38:40 -0700710bool icmp_global_allow(void);
711extern int sysctl_icmp_msgs_per_sec;
712extern int sysctl_icmp_msgs_burst;
713
Arnaldo Carvalho de Melo20380732005-08-16 02:18:02 -0300714#ifdef CONFIG_PROC_FS
Joe Perches5c3a0fd2013-09-21 10:22:42 -0700715int ip_misc_proc_init(void);
Arnaldo Carvalho de Melo20380732005-08-16 02:18:02 -0300716#endif
717
Roopa Prabhu404eb772018-05-22 14:03:27 -0700718int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto,
719 struct netlink_ext_ack *extack);
720
Linus Torvalds1da177e2005-04-16 15:20:36 -0700721#endif /* _IP_H */