Merge tag 'for-5.19/block-2022-06-02' of git://git.kernel.dk/linux-block
[linux-block.git] / include / net / ip.h
CommitLineData
2874c5fd 1/* SPDX-License-Identifier: GPL-2.0-or-later */
1da177e4
LT
2/*
3 * INET An implementation of the TCP/IP protocol suite for the LINUX
4 * operating system. INET is implemented using the BSD Socket
5 * interface as the means of communication with the user level.
6 *
7 * Definitions for the IP module.
8 *
9 * Version: @(#)ip.h 1.0.2 05/07/93
10 *
02c30a84 11 * Authors: Ross Biro
1da177e4
LT
12 * Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
13 * Alan Cox, <gw4pts@gw4pts.ampr.org>
14 *
15 * Changes:
16 * Mike McLagan : Routing by source
1da177e4
LT
17 */
18#ifndef _IP_H
19#define _IP_H
20
1da177e4 21#include <linux/types.h>
1da177e4
LT
22#include <linux/ip.h>
23#include <linux/in.h>
c9bdd4b5 24#include <linux/skbuff.h>
f0b1e64c 25#include <linux/jhash.h>
de40a3e8 26#include <linux/sockptr.h>
020e71a3 27#include <linux/static_key.h>
14c85021
ACM
28
29#include <net/inet_sock.h>
aa661581 30#include <net/route.h>
1da177e4 31#include <net/snmp.h>
86b08d86 32#include <net/flow.h>
1bd758eb 33#include <net/flow_dissector.h>
f0b1e64c 34#include <net/netns/hash.h>
fade5641 35#include <net/lwtunnel.h>
1da177e4 36
7ed14d97 37#define IPV4_MAX_PMTU 65535U /* RFC 2675, Section 5.1 */
b5476022 38#define IPV4_MIN_MTU 68 /* RFC 791 */
7ed14d97 39
9ab948a9
DA
40extern unsigned int sysctl_fib_sync_mem;
41extern unsigned int sysctl_fib_sync_mem_min;
42extern unsigned int sysctl_fib_sync_mem_max;
43
1da177e4
LT
44struct sock;
45
fd2c3ef7 46struct inet_skb_parm {
0b922b7a 47 int iif;
1da177e4 48 struct ip_options opt; /* Compiled IP options */
a04a480d 49 u16 flags;
1da177e4 50
df4d9254
HFS
51#define IPSKB_FORWARDED BIT(0)
52#define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
53#define IPSKB_XFRM_TRANSFORMED BIT(2)
54#define IPSKB_FRAG_COMPLETE BIT(3)
55#define IPSKB_REROUTED BIT(4)
56#define IPSKB_DOREDIRECT BIT(5)
d6b915e2 57#define IPSKB_FRAG_PMTU BIT(6)
9ee6c5dc 58#define IPSKB_L3SLAVE BIT(7)
e6175a2e 59#define IPSKB_NOPOLICY BIT(8)
5f2d04f1
PM
60
61 u16 frag_max_size;
1da177e4
LT
62};
63
a04a480d
DA
64static inline bool ipv4_l3mdev_skb(u16 flags)
65{
66 return !!(flags & IPSKB_L3SLAVE);
67}
68
c9bdd4b5
ACM
69static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
70{
eddc9ec5 71 return ip_hdr(skb)->ihl * 4;
c9bdd4b5
ACM
72}
73
fd2c3ef7 74struct ipcm_cookie {
24025c46 75 struct sockcm_cookie sockc;
c1d18f9f 76 __be32 addr;
1da177e4 77 int oif;
f6d8bd05 78 struct ip_options_rcu *opt;
f02db315
FF
79 __u8 ttl;
80 __s16 tos;
81 char priority;
bec1f6f6 82 __u16 gso_size;
1da177e4
LT
83};
84
35178206
WB
85static inline void ipcm_init(struct ipcm_cookie *ipcm)
86{
87 *ipcm = (struct ipcm_cookie) { .tos = -1 };
88}
89
90static inline void ipcm_init_sk(struct ipcm_cookie *ipcm,
91 const struct inet_sock *inet)
92{
93 ipcm_init(ipcm);
94
c6af0c22 95 ipcm->sockc.mark = inet->sk.sk_mark;
35178206 96 ipcm->sockc.tsflags = inet->sk.sk_tsflags;
4c971d2f 97 ipcm->oif = READ_ONCE(inet->sk.sk_bound_dev_if);
35178206
WB
98 ipcm->addr = inet->inet_saddr;
99}
100
1da177e4 101#define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
4b261c75 102#define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
1da177e4 103
fb74c277 104/* return enslaved device index if relevant */
d6bb2d1e 105static inline int inet_sdif(const struct sk_buff *skb)
fb74c277
DA
106{
107#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
108 if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags))
109 return IPCB(skb)->iif;
110#endif
111 return 0;
112}
113
5796ef75
KT
114/* Special input handler for packets caught by router alert option.
115 They are selected only by protocol field, and then processed likely
116 local ones; but only if someone wants them! Otherwise, router
117 not running rsvpd will kill RSVP.
118
119 It is user level problem, what it will make with them.
120 I have no idea, how it will masquearde or NAT them (it is joke, joke :-)),
121 but receiver should be enough clever f.e. to forward mtrace requests,
122 sent to multicast group to reach destination designated router.
123 */
124
fd2c3ef7 125struct ip_ra_chain {
43a951e9 126 struct ip_ra_chain __rcu *next;
1da177e4 127 struct sock *sk;
592fcb9d
ED
128 union {
129 void (*destructor)(struct sock *);
130 struct sock *saved_sk;
131 };
66018506 132 struct rcu_head rcu;
1da177e4
LT
133};
134
1da177e4
LT
135/* IP flags. */
136#define IP_CE 0x8000 /* Flag: "Congestion" */
137#define IP_DF 0x4000 /* Flag: "Don't Fragment" */
138#define IP_MF 0x2000 /* Flag: "More Fragments" */
139#define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
140
141#define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
142
14c85021
ACM
143struct msghdr;
144struct net_device;
145struct packet_type;
146struct rtable;
14c85021
ACM
147struct sockaddr;
148
72c1d3bd 149int igmp_mc_init(void);
1da177e4
LT
150
151/*
152 * Functions provided by ip.c
153 */
154
cfe673b0 155int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
5c3a0fd7 156 __be32 saddr, __be32 daddr,
de033b7d 157 struct ip_options_rcu *opt, u8 tos);
5c3a0fd7
JP
158int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
159 struct net_device *orig_dev);
17266ee9
EC
160void ip_list_rcv(struct list_head *head, struct packet_type *pt,
161 struct net_device *orig_dev);
5c3a0fd7 162int ip_local_deliver(struct sk_buff *skb);
68cb7d53 163void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto);
5c3a0fd7 164int ip_mr_input(struct sk_buff *skb);
ede2059d
EB
165int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb);
166int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb);
694869b3
EB
167int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
168 int (*output)(struct net *, struct sock *, struct sk_buff *));
c8b17be0
PNA
169
170struct ip_fraglist_iter {
c8b17be0
PNA
171 struct sk_buff *frag;
172 struct iphdr *iph;
173 int offset;
174 unsigned int hlen;
175};
176
177void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph,
178 unsigned int hlen, struct ip_fraglist_iter *iter);
179void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter);
180
181static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter)
182{
183 struct sk_buff *skb = iter->frag;
184
185 iter->frag = skb->next;
186 skb_mark_not_on_list(skb);
187
188 return skb;
189}
190
065ff79f 191struct ip_frag_state {
e7a409c3 192 bool DF;
065ff79f
PNA
193 unsigned int hlen;
194 unsigned int ll_rs;
195 unsigned int mtu;
196 unsigned int left;
197 int offset;
198 int ptr;
199 __be16 not_last_frag;
200};
201
202void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs,
e7a409c3 203 unsigned int mtu, bool DF, struct ip_frag_state *state);
065ff79f
PNA
204struct sk_buff *ip_frag_next(struct sk_buff *skb,
205 struct ip_frag_state *state);
206
5c3a0fd7 207void ip_send_check(struct iphdr *ip);
cf91a99d 208int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
33224b16 209int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
aad88724 210
69b9e1e0
XL
211int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl,
212 __u8 tos);
5c3a0fd7
JP
213void ip_init(void);
214int ip_append_data(struct sock *sk, struct flowi4 *fl4,
215 int getfrag(void *from, char *to, int offset, int len,
216 int odd, struct sk_buff *skb),
217 void *from, int len, int protolen,
218 struct ipcm_cookie *ipc,
219 struct rtable **rt,
220 unsigned int flags);
221int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
222 struct sk_buff *skb);
223ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
224 int offset, size_t size, int flags);
225struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
226 struct sk_buff_head *queue,
227 struct inet_cork *cork);
228int ip_send_skb(struct net *net, struct sk_buff *skb);
229int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
230void ip_flush_pending_frames(struct sock *sk);
231struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
232 int getfrag(void *from, char *to, int offset,
233 int len, int odd, struct sk_buff *skb),
234 void *from, int length, int transhdrlen,
235 struct ipcm_cookie *ipc, struct rtable **rtp,
1cd7884d 236 struct inet_cork *cork, unsigned int flags);
1c32c5ad 237
05e22e83 238int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl);
69b9e1e0 239
77968b78 240static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
1c32c5ad 241{
77968b78 242 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
1c32c5ad 243}
1da177e4 244
aa661581
FF
245static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
246{
247 return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
248}
249
250static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
251{
252 return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
253}
254
1da177e4 255/* datagram.c */
03645a11 256int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
5c3a0fd7 257int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
1da177e4 258
5c3a0fd7 259void ip4_datagram_release_cb(struct sock *sk);
8141ed9f 260
1da177e4 261struct ip_reply_arg {
82695b30 262 struct kvec iov[1];
88ef4a5a 263 int flags;
d6f5493c 264 __wsum csum;
1da177e4 265 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
82695b30 266 /* -1 if not needed */
f0e48dbf 267 int bound_dev_if;
66b13d99 268 u8 tos;
e2d118a1 269 kuid_t uid;
82695b30 270};
1da177e4 271
88ef4a5a
KK
272#define IP_REPLY_ARG_NOSRCCHECK 1
273
86b08d86
KK
274static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
275{
276 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
277}
278
bdbbb852 279void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
24a2d43d
ED
280 const struct ip_options *sopt,
281 __be32 daddr, __be32 saddr,
282 const struct ip_reply_arg *arg,
d6fb396c 283 unsigned int len, u64 transmit_time);
1da177e4 284
4ce3c183 285#define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
13415e46 286#define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field)
4ce3c183 287#define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
13415e46 288#define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
4ce3c183 289#define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
13415e46 290#define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
61a7e260 291#define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
13415e46 292#define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field)
f7324acd 293#define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
13415e46 294#define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
1da177e4 295
59f09ae8
ED
296static inline u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offt)
297{
298 return *(((unsigned long *)per_cpu_ptr(mib, cpu)) + offt);
299}
300
698365fa 301unsigned long snmp_fold_field(void __percpu *mib, int offt);
4ce3c183 302#if BITS_PER_LONG==32
c4c6bc31
R
303u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
304 size_t syncp_offset);
698365fa 305u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
4ce3c183 306#else
c4c6bc31
R
307static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
308 size_t syncp_offset)
309{
310 return snmp_get_cpu_field(mib, cpu, offct);
311
312}
313
698365fa 314static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
4ce3c183
ED
315{
316 return snmp_fold_field(mib, offt);
317}
318#endif
33490170 319
6348ef2d
JH
320#define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \
321{ \
322 int i, c; \
323 for_each_possible_cpu(c) { \
324 for (i = 0; stats_list[i].name; i++) \
325 buff64[i] += snmp_get_cpu_field64( \
326 mib_statistic, \
327 c, stats_list[i].entry, \
328 offset); \
329 } \
330}
331
332#define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
333{ \
334 int i, c; \
335 for_each_possible_cpu(c) { \
336 for (i = 0; stats_list[i].name; i++) \
337 buff[i] += snmp_get_cpu_field( \
338 mib_statistic, \
339 c, stats_list[i].entry); \
340 } \
341}
342
0bbf87d8 343void inet_get_local_port_range(struct net *net, int *low, int *high);
227b60f5 344
fcd77db0 345#ifdef CONFIG_SYSCTL
66e2f5f7 346static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
e3826f1e 347{
122ff243 348 if (!net->ipv4.sysctl_local_reserved_ports)
30429fba 349 return false;
122ff243 350 return test_bit(port, net->ipv4.sysctl_local_reserved_ports);
e3826f1e 351}
20e61da7
WC
352
353static inline bool sysctl_dev_name_is_allowed(const char *name)
354{
355 return strcmp(name, "default") != 0 && strcmp(name, "all") != 0;
356}
357
82f31ebf 358static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
4548b683 359{
82f31ebf 360 return port < net->ipv4.sysctl_ip_prot_sock;
4548b683
KJ
361}
362
122ff243 363#else
66e2f5f7 364static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
122ff243 365{
30429fba 366 return false;
122ff243 367}
4548b683 368
82f31ebf 369static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
4548b683 370{
82f31ebf 371 return port < PROT_SOCK;
4548b683 372}
122ff243 373#endif
e3826f1e 374
822c8685
DD
375__be32 inet_current_timestamp(void);
376
20380731
ACM
377/* From inetpeer.c */
378extern int inet_peer_threshold;
379extern int inet_peer_minttl;
380extern int inet_peer_maxttl;
20380731 381
5c3a0fd7 382void ipfrag_init(void);
20380731 383
5c3a0fd7 384void ip_static_sysctl_init(void);
bd7b1533 385
e110861f
LC
386#define IP4_REPLY_MARK(net, mark) \
387 ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
388
d18cd551
DM
389static inline bool ip_is_fragment(const struct iphdr *iph)
390{
391 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
392}
393
1da177e4 394#ifdef CONFIG_INET
14c85021
ACM
395#include <net/dst.h>
396
1da177e4
LT
397/* The function in 2.2 was invalid, producing wrong result for
398 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
399static inline
400int ip_decrease_ttl(struct iphdr *iph)
401{
5c78f275
AV
402 u32 check = (__force u32)iph->check;
403 check += (__force u32)htons(0x0100);
404 iph->check = (__force __sum16)(check + (check>=0xFFFF));
1da177e4
LT
405 return --iph->ttl;
406}
407
d52e5a7e
SD
408static inline int ip_mtu_locked(const struct dst_entry *dst)
409{
410 const struct rtable *rt = (const struct rtable *)dst;
411
412 return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU);
413}
414
1da177e4 415static inline
4e3f5d72 416int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst)
1da177e4 417{
4e3f5d72
ED
418 u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc);
419
420 return pmtudisc == IP_PMTUDISC_DO ||
421 (pmtudisc == IP_PMTUDISC_WANT &&
d52e5a7e 422 !ip_mtu_locked(dst));
1da177e4
LT
423}
424
f87c10a8
HFS
425static inline bool ip_sk_accept_pmtu(const struct sock *sk)
426{
1b346576
HFS
427 return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
428 inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
f87c10a8
HFS
429}
430
431static inline bool ip_sk_use_pmtu(const struct sock *sk)
432{
433 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
434}
435
60ff7467 436static inline bool ip_sk_ignore_df(const struct sock *sk)
1b346576
HFS
437{
438 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
439 inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
440}
441
f87c10a8
HFS
442static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
443 bool forwarding)
444{
ac6627a2 445 const struct rtable *rt = container_of(dst, struct rtable, dst);
f87c10a8 446 struct net *net = dev_net(dst->dev);
02a1b175 447 unsigned int mtu;
f87c10a8
HFS
448
449 if (net->ipv4.sysctl_ip_fwd_use_pmtu ||
d52e5a7e 450 ip_mtu_locked(dst) ||
ac6627a2
VF
451 !forwarding) {
452 mtu = rt->rt_pmtu;
453 if (mtu && time_before(jiffies, rt->dst.expires))
454 goto out;
455 }
f87c10a8 456
02a1b175
MÅ»
457 /* 'forwarding = true' case should always honour route mtu */
458 mtu = dst_metric_raw(dst, RTAX_MTU);
ac6627a2
VF
459 if (mtu)
460 goto out;
461
462 mtu = READ_ONCE(dst->dev->mtu);
463
464 if (unlikely(ip_mtu_locked(dst))) {
465 if (rt->rt_uses_gateway && mtu > 576)
466 mtu = 576;
467 }
468
469out:
470 mtu = min_t(unsigned int, mtu, IP_MAX_MTU);
02a1b175 471
fade5641 472 return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
f87c10a8
HFS
473}
474
fedbb6b4
SL
475static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
476 const struct sk_buff *skb)
f87c10a8 477{
fade5641
VF
478 unsigned int mtu;
479
caf3f267 480 if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) {
f87c10a8 481 bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
caf3f267 482
f87c10a8 483 return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
f87c10a8 484 }
caf3f267 485
fade5641
VF
486 mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
487 return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu);
f87c10a8
HFS
488}
489
767a2217 490struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
d7e774f3
DA
491 int fc_mx_len,
492 struct netlink_ext_ack *extack);
cc5f0eb2
DA
493static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics)
494{
495 if (fib_metrics != &dst_default_metrics &&
496 refcount_dec_and_test(&fib_metrics->refcnt))
497 kfree(fib_metrics);
498}
a919525a 499
e1255ed4
DA
500/* ipv4 and ipv6 both use refcounted metrics if it is not the default */
501static inline
502void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics)
503{
504 dst_init_metrics(dst, fib_metrics->metrics, true);
505
506 if (fib_metrics != &dst_default_metrics) {
507 dst->_metrics |= DST_METRICS_REFCOUNTED;
508 refcount_inc(&fib_metrics->refcnt);
509 }
510}
511
1620a336
DA
512static inline
513void ip_dst_metrics_put(struct dst_entry *dst)
514{
515 struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst);
516
517 if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt))
518 kfree(p);
519}
520
b6a7719a 521void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
73f156a6 522
b6a7719a
HFS
523static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb,
524 struct sock *sk, int segs)
1da177e4 525{
703133de
AA
526 struct iphdr *iph = ip_hdr(skb);
527
23f57406
ED
528 /* We had many attacks based on IPID, use the private
529 * generator as much as we can.
530 */
531 if (sk && inet_sk(sk)->inet_daddr) {
532 iph->id = htons(inet_sk(sk)->inet_id);
533 inet_sk(sk)->inet_id += segs;
534 return;
535 }
60ff7467 536 if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) {
23f57406 537 iph->id = 0;
73f156a6 538 } else {
23f57406 539 /* Unfortunately we need the big hammer to get a suitable IPID */
b6a7719a 540 __ip_select_ident(net, iph, segs);
73f156a6
ED
541 }
542}
543
b6a7719a
HFS
544static inline void ip_select_ident(struct net *net, struct sk_buff *skb,
545 struct sock *sk)
73f156a6 546{
b6a7719a 547 ip_select_ident_segs(net, skb, sk, 1);
1da177e4
LT
548}
549
ed70fcfc
TH
550static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto)
551{
552 return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr,
553 skb->len, proto, 0);
554}
555
c3f83241
TH
556/* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store
557 * Equivalent to : flow->v4addrs.src = iph->saddr;
558 * flow->v4addrs.dst = iph->daddr;
559 */
560static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow,
561 const struct iphdr *iph)
562{
563 BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) !=
564 offsetof(typeof(flow->addrs), v4addrs.src) +
565 sizeof(flow->addrs.v4addrs.src));
566 memcpy(&flow->addrs.v4addrs, &iph->saddr, sizeof(flow->addrs.v4addrs));
567 flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
568}
569
1da177e4
LT
570/*
571 * Map a multicast IP onto multicast MAC for type ethernet.
572 */
573
714e85be 574static inline void ip_eth_mc_map(__be32 naddr, char *buf)
1da177e4 575{
714e85be 576 __u32 addr=ntohl(naddr);
1da177e4
LT
577 buf[0]=0x01;
578 buf[1]=0x00;
579 buf[2]=0x5e;
580 buf[5]=addr&0xFF;
581 addr>>=8;
582 buf[4]=addr&0xFF;
583 addr>>=8;
584 buf[3]=addr&0x7F;
585}
586
587/*
588 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
589 * Leave P_Key as 0 to be filled in by driver.
590 */
591
a9e527e3 592static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
1da177e4 593{
714e85be 594 __u32 addr;
a9e527e3
RM
595 unsigned char scope = broadcast[5] & 0xF;
596
1da177e4
LT
597 buf[0] = 0; /* Reserved */
598 buf[1] = 0xff; /* Multicast QPN */
599 buf[2] = 0xff;
600 buf[3] = 0xff;
714e85be 601 addr = ntohl(naddr);
1da177e4 602 buf[4] = 0xff;
a9e527e3 603 buf[5] = 0x10 | scope; /* scope from broadcast address */
1da177e4
LT
604 buf[6] = 0x40; /* IPv4 signature */
605 buf[7] = 0x1b;
a9e527e3
RM
606 buf[8] = broadcast[8]; /* P_Key */
607 buf[9] = broadcast[9];
1da177e4
LT
608 buf[10] = 0;
609 buf[11] = 0;
610 buf[12] = 0;
611 buf[13] = 0;
612 buf[14] = 0;
613 buf[15] = 0;
614 buf[19] = addr & 0xff;
615 addr >>= 8;
616 buf[18] = addr & 0xff;
617 addr >>= 8;
618 buf[17] = addr & 0xff;
619 addr >>= 8;
620 buf[16] = addr & 0x0f;
621}
622
93ca3bb5
TT
623static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
624{
625 if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
626 memcpy(buf, broadcast, 4);
627 else
628 memcpy(buf, &naddr, sizeof(naddr));
629}
630
dfd56b8b 631#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
632#include <linux/ipv6.h>
633#endif
634
635static __inline__ void inet_reset_saddr(struct sock *sk)
636{
c720c7e8 637 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
dfd56b8b 638#if IS_ENABLED(CONFIG_IPV6)
1da177e4
LT
639 if (sk->sk_family == PF_INET6) {
640 struct ipv6_pinfo *np = inet6_sk(sk);
641
642 memset(&np->saddr, 0, sizeof(np->saddr));
efe4208f 643 memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
1da177e4
LT
644 }
645#endif
646}
647
648#endif
649
72afa352
DA
650static inline unsigned int ipv4_addr_hash(__be32 ip)
651{
652 return (__force unsigned int) ip;
653}
654
f0b1e64c
MKL
655static inline u32 ipv4_portaddr_hash(const struct net *net,
656 __be32 saddr,
657 unsigned int port)
658{
659 return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port;
660}
661
5c3a0fd7 662bool ip_call_ra_chain(struct sk_buff *skb);
1da177e4
LT
663
664/*
b798232f 665 * Functions provided by ip_fragment.c
1da177e4
LT
666 */
667
fd2c3ef7 668enum ip_defrag_users {
1da177e4
LT
669 IP_DEFRAG_LOCAL_DELIVER,
670 IP_DEFRAG_CALL_RA_CHAIN,
671 IP_DEFRAG_CONNTRACK_IN,
4be929be 672 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
1da177e4 673 IP_DEFRAG_CONNTRACK_OUT,
4be929be 674 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
8fa9ff68 675 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
4be929be 676 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
1da177e4
LT
677 IP_DEFRAG_VS_IN,
678 IP_DEFRAG_VS_OUT,
595fc71b
DM
679 IP_DEFRAG_VS_FWD,
680 IP_DEFRAG_AF_PACKET,
bc416d97 681 IP_DEFRAG_MACVLAN,
1da177e4
LT
682};
683
5cf42280
AZ
684/* Return true if the value of 'user' is between 'lower_bond'
685 * and 'upper_bond' inclusively.
686 */
687static inline bool ip_defrag_user_in_between(u32 user,
688 enum ip_defrag_users lower_bond,
689 enum ip_defrag_users upper_bond)
690{
691 return user >= lower_bond && user <= upper_bond;
692}
693
19bcf9f2 694int ip_defrag(struct net *net, struct sk_buff *skb, u32 user);
bc416d97 695#ifdef CONFIG_INET
19bcf9f2 696struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user);
bc416d97 697#else
19bcf9f2 698static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user)
bc416d97
ED
699{
700 return skb;
701}
702#endif
1da177e4
LT
703
704/*
705 * Functions provided by ip_forward.c
706 */
82695b30 707
5c3a0fd7 708int ip_forward(struct sk_buff *skb);
82695b30 709
1da177e4
LT
710/*
711 * Functions provided by ip_options.c
712 */
82695b30 713
5c3a0fd7 714void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
4f0e3040 715 __be32 daddr, struct rtable *rt);
24a2d43d 716
91ed1e66
PA
717int __ip_options_echo(struct net *net, struct ip_options *dopt,
718 struct sk_buff *skb, const struct ip_options *sopt);
719static inline int ip_options_echo(struct net *net, struct ip_options *dopt,
720 struct sk_buff *skb)
24a2d43d 721{
91ed1e66 722 return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt);
24a2d43d
ED
723}
724
5c3a0fd7 725void ip_options_fragment(struct sk_buff *skb);
3da1ed7a
NS
726int __ip_options_compile(struct net *net, struct ip_options *opt,
727 struct sk_buff *skb, __be32 *info);
5c3a0fd7
JP
728int ip_options_compile(struct net *net, struct ip_options *opt,
729 struct sk_buff *skb);
730int ip_options_get(struct net *net, struct ip_options_rcu **optp,
de40a3e8 731 sockptr_t data, int optlen);
5c3a0fd7
JP
732void ip_options_undo(struct ip_options *opt);
733void ip_forward_options(struct sk_buff *skb);
8c83f2df 734int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev);
1da177e4
LT
735
736/*
737 * Functions provided by ip_sockglue.c
738 */
739
fbf8866d 740void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
ad959036
PA
741void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk,
742 struct sk_buff *skb, int tlen, int offset);
24025c46 743int ip_cmsg_send(struct sock *sk, struct msghdr *msg,
c8e6ad08 744 struct ipcm_cookie *ipc, bool allow_ipv6);
020e71a3 745DECLARE_STATIC_KEY_FALSE(ip4_min_ttl);
a7b75c5a 746int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
5c3a0fd7
JP
747 unsigned int optlen);
748int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
749 int __user *optlen);
5c3a0fd7
JP
750int ip_ra_control(struct sock *sk, unsigned char on,
751 void (*destructor)(struct sock *));
752
85fbaa75 753int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
5c3a0fd7
JP
754void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
755 u32 info, u8 *payload);
756void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
757 u32 info);
1da177e4 758
5961de9f
TH
759static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
760{
ad959036 761 ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0);
5961de9f
TH
762}
763
4cdf507d
ED
764bool icmp_global_allow(void);
765extern int sysctl_icmp_msgs_per_sec;
766extern int sysctl_icmp_msgs_burst;
767
20380731 768#ifdef CONFIG_PROC_FS
5c3a0fd7 769int ip_misc_proc_init(void);
20380731
ACM
770#endif
771
5e1a99ea 772int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family,
404eb77e
RP
773 struct netlink_ext_ack *extack);
774
501a90c9
ED
775static inline bool inetdev_valid_mtu(unsigned int mtu)
776{
777 return likely(mtu >= IPV4_MIN_MTU);
778}
779
c4e446bf 780void ip_sock_set_freebind(struct sock *sk);
2de569bd 781int ip_sock_set_mtu_discover(struct sock *sk, int val);
c1f9ec57 782void ip_sock_set_pktinfo(struct sock *sk);
db45c0ef 783void ip_sock_set_recverr(struct sock *sk);
6ebf71ba 784void ip_sock_set_tos(struct sock *sk, int val);
4f47d5d5 785void __ip_sock_set_tos(struct sock *sk, int val);
6ebf71ba 786
1da177e4 787#endif /* _IP_H */