Commit | Line | Data |
---|---|---|
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> |
14c85021 ACM |
27 | |
28 | #include <net/inet_sock.h> | |
aa661581 | 29 | #include <net/route.h> |
1da177e4 | 30 | #include <net/snmp.h> |
86b08d86 | 31 | #include <net/flow.h> |
1bd758eb | 32 | #include <net/flow_dissector.h> |
f0b1e64c | 33 | #include <net/netns/hash.h> |
1da177e4 | 34 | |
7ed14d97 | 35 | #define IPV4_MAX_PMTU 65535U /* RFC 2675, Section 5.1 */ |
b5476022 | 36 | #define IPV4_MIN_MTU 68 /* RFC 791 */ |
7ed14d97 | 37 | |
9ab948a9 DA |
38 | extern unsigned int sysctl_fib_sync_mem; |
39 | extern unsigned int sysctl_fib_sync_mem_min; | |
40 | extern unsigned int sysctl_fib_sync_mem_max; | |
41 | ||
1da177e4 LT |
42 | struct sock; |
43 | ||
fd2c3ef7 | 44 | struct inet_skb_parm { |
0b922b7a | 45 | int iif; |
1da177e4 | 46 | struct ip_options opt; /* Compiled IP options */ |
a04a480d | 47 | u16 flags; |
1da177e4 | 48 | |
df4d9254 HFS |
49 | #define IPSKB_FORWARDED BIT(0) |
50 | #define IPSKB_XFRM_TUNNEL_SIZE BIT(1) | |
51 | #define IPSKB_XFRM_TRANSFORMED BIT(2) | |
52 | #define IPSKB_FRAG_COMPLETE BIT(3) | |
53 | #define IPSKB_REROUTED BIT(4) | |
54 | #define IPSKB_DOREDIRECT BIT(5) | |
d6b915e2 | 55 | #define IPSKB_FRAG_PMTU BIT(6) |
9ee6c5dc | 56 | #define IPSKB_L3SLAVE BIT(7) |
5f2d04f1 PM |
57 | |
58 | u16 frag_max_size; | |
1da177e4 LT |
59 | }; |
60 | ||
a04a480d DA |
61 | static inline bool ipv4_l3mdev_skb(u16 flags) |
62 | { | |
63 | return !!(flags & IPSKB_L3SLAVE); | |
64 | } | |
65 | ||
c9bdd4b5 ACM |
66 | static inline unsigned int ip_hdrlen(const struct sk_buff *skb) |
67 | { | |
eddc9ec5 | 68 | return ip_hdr(skb)->ihl * 4; |
c9bdd4b5 ACM |
69 | } |
70 | ||
fd2c3ef7 | 71 | struct ipcm_cookie { |
24025c46 | 72 | struct sockcm_cookie sockc; |
c1d18f9f | 73 | __be32 addr; |
1da177e4 | 74 | int oif; |
f6d8bd05 | 75 | struct ip_options_rcu *opt; |
f02db315 FF |
76 | __u8 ttl; |
77 | __s16 tos; | |
78 | char priority; | |
bec1f6f6 | 79 | __u16 gso_size; |
1da177e4 LT |
80 | }; |
81 | ||
35178206 WB |
82 | static inline void ipcm_init(struct ipcm_cookie *ipcm) |
83 | { | |
84 | *ipcm = (struct ipcm_cookie) { .tos = -1 }; | |
85 | } | |
86 | ||
87 | static inline void ipcm_init_sk(struct ipcm_cookie *ipcm, | |
88 | const struct inet_sock *inet) | |
89 | { | |
90 | ipcm_init(ipcm); | |
91 | ||
c6af0c22 | 92 | ipcm->sockc.mark = inet->sk.sk_mark; |
35178206 WB |
93 | ipcm->sockc.tsflags = inet->sk.sk_tsflags; |
94 | ipcm->oif = inet->sk.sk_bound_dev_if; | |
95 | ipcm->addr = inet->inet_saddr; | |
96 | } | |
97 | ||
1da177e4 | 98 | #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb)) |
4b261c75 | 99 | #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb)) |
1da177e4 | 100 | |
fb74c277 DA |
101 | /* return enslaved device index if relevant */ |
102 | static inline int inet_sdif(struct sk_buff *skb) | |
103 | { | |
104 | #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV) | |
105 | if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags)) | |
106 | return IPCB(skb)->iif; | |
107 | #endif | |
108 | return 0; | |
109 | } | |
110 | ||
5796ef75 KT |
111 | /* Special input handler for packets caught by router alert option. |
112 | They are selected only by protocol field, and then processed likely | |
113 | local ones; but only if someone wants them! Otherwise, router | |
114 | not running rsvpd will kill RSVP. | |
115 | ||
116 | It is user level problem, what it will make with them. | |
117 | I have no idea, how it will masquearde or NAT them (it is joke, joke :-)), | |
118 | but receiver should be enough clever f.e. to forward mtrace requests, | |
119 | sent to multicast group to reach destination designated router. | |
120 | */ | |
121 | ||
fd2c3ef7 | 122 | struct ip_ra_chain { |
43a951e9 | 123 | struct ip_ra_chain __rcu *next; |
1da177e4 | 124 | struct sock *sk; |
592fcb9d ED |
125 | union { |
126 | void (*destructor)(struct sock *); | |
127 | struct sock *saved_sk; | |
128 | }; | |
66018506 | 129 | struct rcu_head rcu; |
1da177e4 LT |
130 | }; |
131 | ||
1da177e4 LT |
132 | /* IP flags. */ |
133 | #define IP_CE 0x8000 /* Flag: "Congestion" */ | |
134 | #define IP_DF 0x4000 /* Flag: "Don't Fragment" */ | |
135 | #define IP_MF 0x2000 /* Flag: "More Fragments" */ | |
136 | #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */ | |
137 | ||
138 | #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */ | |
139 | ||
14c85021 ACM |
140 | struct msghdr; |
141 | struct net_device; | |
142 | struct packet_type; | |
143 | struct rtable; | |
14c85021 ACM |
144 | struct sockaddr; |
145 | ||
72c1d3bd | 146 | int igmp_mc_init(void); |
1da177e4 LT |
147 | |
148 | /* | |
149 | * Functions provided by ip.c | |
150 | */ | |
151 | ||
cfe673b0 | 152 | int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk, |
5c3a0fd7 | 153 | __be32 saddr, __be32 daddr, |
de033b7d | 154 | struct ip_options_rcu *opt, u8 tos); |
5c3a0fd7 JP |
155 | int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, |
156 | struct net_device *orig_dev); | |
17266ee9 EC |
157 | void ip_list_rcv(struct list_head *head, struct packet_type *pt, |
158 | struct net_device *orig_dev); | |
5c3a0fd7 | 159 | int ip_local_deliver(struct sk_buff *skb); |
68cb7d53 | 160 | void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto); |
5c3a0fd7 | 161 | int ip_mr_input(struct sk_buff *skb); |
ede2059d EB |
162 | int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb); |
163 | int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb); | |
694869b3 EB |
164 | int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb, |
165 | int (*output)(struct net *, struct sock *, struct sk_buff *)); | |
c8b17be0 PNA |
166 | |
167 | struct ip_fraglist_iter { | |
c8b17be0 PNA |
168 | struct sk_buff *frag; |
169 | struct iphdr *iph; | |
170 | int offset; | |
171 | unsigned int hlen; | |
172 | }; | |
173 | ||
174 | void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph, | |
175 | unsigned int hlen, struct ip_fraglist_iter *iter); | |
176 | void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter); | |
177 | ||
178 | static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter) | |
179 | { | |
180 | struct sk_buff *skb = iter->frag; | |
181 | ||
182 | iter->frag = skb->next; | |
183 | skb_mark_not_on_list(skb); | |
184 | ||
185 | return skb; | |
186 | } | |
187 | ||
065ff79f | 188 | struct ip_frag_state { |
e7a409c3 | 189 | bool DF; |
065ff79f PNA |
190 | unsigned int hlen; |
191 | unsigned int ll_rs; | |
192 | unsigned int mtu; | |
193 | unsigned int left; | |
194 | int offset; | |
195 | int ptr; | |
196 | __be16 not_last_frag; | |
197 | }; | |
198 | ||
199 | void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs, | |
e7a409c3 | 200 | unsigned int mtu, bool DF, struct ip_frag_state *state); |
065ff79f PNA |
201 | struct sk_buff *ip_frag_next(struct sk_buff *skb, |
202 | struct ip_frag_state *state); | |
203 | ||
5c3a0fd7 | 204 | void ip_send_check(struct iphdr *ip); |
cf91a99d | 205 | int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb); |
33224b16 | 206 | int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb); |
aad88724 | 207 | |
69b9e1e0 XL |
208 | int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl, |
209 | __u8 tos); | |
5c3a0fd7 JP |
210 | void ip_init(void); |
211 | int ip_append_data(struct sock *sk, struct flowi4 *fl4, | |
212 | int getfrag(void *from, char *to, int offset, int len, | |
213 | int odd, struct sk_buff *skb), | |
214 | void *from, int len, int protolen, | |
215 | struct ipcm_cookie *ipc, | |
216 | struct rtable **rt, | |
217 | unsigned int flags); | |
218 | int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, | |
219 | struct sk_buff *skb); | |
220 | ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page, | |
221 | int offset, size_t size, int flags); | |
222 | struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4, | |
223 | struct sk_buff_head *queue, | |
224 | struct inet_cork *cork); | |
225 | int ip_send_skb(struct net *net, struct sk_buff *skb); | |
226 | int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4); | |
227 | void ip_flush_pending_frames(struct sock *sk); | |
228 | struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4, | |
229 | int getfrag(void *from, char *to, int offset, | |
230 | int len, int odd, struct sk_buff *skb), | |
231 | void *from, int length, int transhdrlen, | |
232 | struct ipcm_cookie *ipc, struct rtable **rtp, | |
1cd7884d | 233 | struct inet_cork *cork, unsigned int flags); |
1c32c5ad | 234 | |
05e22e83 | 235 | int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl); |
69b9e1e0 | 236 | |
77968b78 | 237 | static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4) |
1c32c5ad | 238 | { |
77968b78 | 239 | return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base); |
1c32c5ad | 240 | } |
1da177e4 | 241 | |
aa661581 FF |
242 | static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet) |
243 | { | |
244 | return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos); | |
245 | } | |
246 | ||
247 | static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk) | |
248 | { | |
249 | return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk); | |
250 | } | |
251 | ||
1da177e4 | 252 | /* datagram.c */ |
03645a11 | 253 | int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); |
5c3a0fd7 | 254 | int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); |
1da177e4 | 255 | |
5c3a0fd7 | 256 | void ip4_datagram_release_cb(struct sock *sk); |
8141ed9f | 257 | |
1da177e4 | 258 | struct ip_reply_arg { |
82695b30 | 259 | struct kvec iov[1]; |
88ef4a5a | 260 | int flags; |
d6f5493c | 261 | __wsum csum; |
1da177e4 | 262 | int csumoffset; /* u16 offset of csum in iov[0].iov_base */ |
82695b30 | 263 | /* -1 if not needed */ |
f0e48dbf | 264 | int bound_dev_if; |
66b13d99 | 265 | u8 tos; |
e2d118a1 | 266 | kuid_t uid; |
82695b30 | 267 | }; |
1da177e4 | 268 | |
88ef4a5a KK |
269 | #define IP_REPLY_ARG_NOSRCCHECK 1 |
270 | ||
86b08d86 KK |
271 | static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg) |
272 | { | |
273 | return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0; | |
274 | } | |
275 | ||
bdbbb852 | 276 | void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb, |
24a2d43d ED |
277 | const struct ip_options *sopt, |
278 | __be32 daddr, __be32 saddr, | |
279 | const struct ip_reply_arg *arg, | |
d6fb396c | 280 | unsigned int len, u64 transmit_time); |
1da177e4 | 281 | |
4ce3c183 | 282 | #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field) |
13415e46 | 283 | #define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field) |
4ce3c183 | 284 | #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) |
13415e46 | 285 | #define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) |
4ce3c183 | 286 | #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) |
13415e46 | 287 | #define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) |
61a7e260 | 288 | #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field) |
13415e46 | 289 | #define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field) |
f7324acd | 290 | #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd) |
13415e46 | 291 | #define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd) |
1da177e4 | 292 | |
c4c6bc31 | 293 | u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offct); |
698365fa | 294 | unsigned long snmp_fold_field(void __percpu *mib, int offt); |
4ce3c183 | 295 | #if BITS_PER_LONG==32 |
c4c6bc31 R |
296 | u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct, |
297 | size_t syncp_offset); | |
698365fa | 298 | u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off); |
4ce3c183 | 299 | #else |
c4c6bc31 R |
300 | static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct, |
301 | size_t syncp_offset) | |
302 | { | |
303 | return snmp_get_cpu_field(mib, cpu, offct); | |
304 | ||
305 | } | |
306 | ||
698365fa | 307 | static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off) |
4ce3c183 ED |
308 | { |
309 | return snmp_fold_field(mib, offt); | |
310 | } | |
311 | #endif | |
33490170 | 312 | |
6348ef2d JH |
313 | #define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \ |
314 | { \ | |
315 | int i, c; \ | |
316 | for_each_possible_cpu(c) { \ | |
317 | for (i = 0; stats_list[i].name; i++) \ | |
318 | buff64[i] += snmp_get_cpu_field64( \ | |
319 | mib_statistic, \ | |
320 | c, stats_list[i].entry, \ | |
321 | offset); \ | |
322 | } \ | |
323 | } | |
324 | ||
325 | #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \ | |
326 | { \ | |
327 | int i, c; \ | |
328 | for_each_possible_cpu(c) { \ | |
329 | for (i = 0; stats_list[i].name; i++) \ | |
330 | buff[i] += snmp_get_cpu_field( \ | |
331 | mib_statistic, \ | |
332 | c, stats_list[i].entry); \ | |
333 | } \ | |
334 | } | |
335 | ||
0bbf87d8 | 336 | void inet_get_local_port_range(struct net *net, int *low, int *high); |
227b60f5 | 337 | |
fcd77db0 | 338 | #ifdef CONFIG_SYSCTL |
66e2f5f7 | 339 | static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) |
e3826f1e | 340 | { |
122ff243 | 341 | if (!net->ipv4.sysctl_local_reserved_ports) |
30429fba | 342 | return false; |
122ff243 | 343 | return test_bit(port, net->ipv4.sysctl_local_reserved_ports); |
e3826f1e | 344 | } |
20e61da7 WC |
345 | |
346 | static inline bool sysctl_dev_name_is_allowed(const char *name) | |
347 | { | |
348 | return strcmp(name, "default") != 0 && strcmp(name, "all") != 0; | |
349 | } | |
350 | ||
82f31ebf | 351 | static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port) |
4548b683 | 352 | { |
82f31ebf | 353 | return port < net->ipv4.sysctl_ip_prot_sock; |
4548b683 KJ |
354 | } |
355 | ||
122ff243 | 356 | #else |
66e2f5f7 | 357 | static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) |
122ff243 | 358 | { |
30429fba | 359 | return false; |
122ff243 | 360 | } |
4548b683 | 361 | |
82f31ebf | 362 | static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port) |
4548b683 | 363 | { |
82f31ebf | 364 | return port < PROT_SOCK; |
4548b683 | 365 | } |
122ff243 | 366 | #endif |
e3826f1e | 367 | |
822c8685 DD |
368 | __be32 inet_current_timestamp(void); |
369 | ||
20380731 ACM |
370 | /* From inetpeer.c */ |
371 | extern int inet_peer_threshold; | |
372 | extern int inet_peer_minttl; | |
373 | extern int inet_peer_maxttl; | |
20380731 | 374 | |
5c3a0fd7 | 375 | void ipfrag_init(void); |
20380731 | 376 | |
5c3a0fd7 | 377 | void ip_static_sysctl_init(void); |
bd7b1533 | 378 | |
e110861f LC |
379 | #define IP4_REPLY_MARK(net, mark) \ |
380 | ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0) | |
381 | ||
d18cd551 DM |
382 | static inline bool ip_is_fragment(const struct iphdr *iph) |
383 | { | |
384 | return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0; | |
385 | } | |
386 | ||
1da177e4 | 387 | #ifdef CONFIG_INET |
14c85021 ACM |
388 | #include <net/dst.h> |
389 | ||
1da177e4 LT |
390 | /* The function in 2.2 was invalid, producing wrong result for |
391 | * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */ | |
392 | static inline | |
393 | int ip_decrease_ttl(struct iphdr *iph) | |
394 | { | |
5c78f275 AV |
395 | u32 check = (__force u32)iph->check; |
396 | check += (__force u32)htons(0x0100); | |
397 | iph->check = (__force __sum16)(check + (check>=0xFFFF)); | |
1da177e4 LT |
398 | return --iph->ttl; |
399 | } | |
400 | ||
d52e5a7e SD |
401 | static inline int ip_mtu_locked(const struct dst_entry *dst) |
402 | { | |
403 | const struct rtable *rt = (const struct rtable *)dst; | |
404 | ||
405 | return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU); | |
406 | } | |
407 | ||
1da177e4 | 408 | static inline |
4e3f5d72 | 409 | int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst) |
1da177e4 | 410 | { |
4e3f5d72 ED |
411 | u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc); |
412 | ||
413 | return pmtudisc == IP_PMTUDISC_DO || | |
414 | (pmtudisc == IP_PMTUDISC_WANT && | |
d52e5a7e | 415 | !ip_mtu_locked(dst)); |
1da177e4 LT |
416 | } |
417 | ||
f87c10a8 HFS |
418 | static inline bool ip_sk_accept_pmtu(const struct sock *sk) |
419 | { | |
1b346576 HFS |
420 | return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE && |
421 | inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT; | |
f87c10a8 HFS |
422 | } |
423 | ||
424 | static inline bool ip_sk_use_pmtu(const struct sock *sk) | |
425 | { | |
426 | return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE; | |
427 | } | |
428 | ||
60ff7467 | 429 | static inline bool ip_sk_ignore_df(const struct sock *sk) |
1b346576 HFS |
430 | { |
431 | return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO || | |
432 | inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT; | |
433 | } | |
434 | ||
f87c10a8 HFS |
435 | static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst, |
436 | bool forwarding) | |
437 | { | |
438 | struct net *net = dev_net(dst->dev); | |
02a1b175 | 439 | unsigned int mtu; |
f87c10a8 HFS |
440 | |
441 | if (net->ipv4.sysctl_ip_fwd_use_pmtu || | |
d52e5a7e | 442 | ip_mtu_locked(dst) || |
f87c10a8 HFS |
443 | !forwarding) |
444 | return dst_mtu(dst); | |
445 | ||
02a1b175 MÅ» |
446 | /* 'forwarding = true' case should always honour route mtu */ |
447 | mtu = dst_metric_raw(dst, RTAX_MTU); | |
448 | if (mtu) | |
449 | return mtu; | |
450 | ||
c780a049 | 451 | return min(READ_ONCE(dst->dev->mtu), IP_MAX_MTU); |
f87c10a8 HFS |
452 | } |
453 | ||
fedbb6b4 SL |
454 | static inline unsigned int ip_skb_dst_mtu(struct sock *sk, |
455 | const struct sk_buff *skb) | |
f87c10a8 | 456 | { |
caf3f267 | 457 | if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) { |
f87c10a8 | 458 | bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED; |
caf3f267 | 459 | |
f87c10a8 | 460 | return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding); |
f87c10a8 | 461 | } |
caf3f267 | 462 | |
c780a049 | 463 | return min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU); |
f87c10a8 HFS |
464 | } |
465 | ||
767a2217 | 466 | struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx, |
d7e774f3 DA |
467 | int fc_mx_len, |
468 | struct netlink_ext_ack *extack); | |
cc5f0eb2 DA |
469 | static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics) |
470 | { | |
471 | if (fib_metrics != &dst_default_metrics && | |
472 | refcount_dec_and_test(&fib_metrics->refcnt)) | |
473 | kfree(fib_metrics); | |
474 | } | |
a919525a | 475 | |
e1255ed4 DA |
476 | /* ipv4 and ipv6 both use refcounted metrics if it is not the default */ |
477 | static inline | |
478 | void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics) | |
479 | { | |
480 | dst_init_metrics(dst, fib_metrics->metrics, true); | |
481 | ||
482 | if (fib_metrics != &dst_default_metrics) { | |
483 | dst->_metrics |= DST_METRICS_REFCOUNTED; | |
484 | refcount_inc(&fib_metrics->refcnt); | |
485 | } | |
486 | } | |
487 | ||
1620a336 DA |
488 | static inline |
489 | void ip_dst_metrics_put(struct dst_entry *dst) | |
490 | { | |
491 | struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst); | |
492 | ||
493 | if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt)) | |
494 | kfree(p); | |
495 | } | |
496 | ||
04ca6973 | 497 | u32 ip_idents_reserve(u32 hash, int segs); |
b6a7719a | 498 | void __ip_select_ident(struct net *net, struct iphdr *iph, int segs); |
73f156a6 | 499 | |
b6a7719a HFS |
500 | static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb, |
501 | struct sock *sk, int segs) | |
1da177e4 | 502 | { |
703133de AA |
503 | struct iphdr *iph = ip_hdr(skb); |
504 | ||
60ff7467 | 505 | if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) { |
1da177e4 LT |
506 | /* This is only to work around buggy Windows95/2000 |
507 | * VJ compression implementations. If the ID field | |
508 | * does not change, they drop every other packet in | |
509 | * a TCP stream using header compression. | |
510 | */ | |
c720c7e8 ED |
511 | if (sk && inet_sk(sk)->inet_daddr) { |
512 | iph->id = htons(inet_sk(sk)->inet_id); | |
73f156a6 ED |
513 | inet_sk(sk)->inet_id += segs; |
514 | } else { | |
1da177e4 | 515 | iph->id = 0; |
73f156a6 ED |
516 | } |
517 | } else { | |
b6a7719a | 518 | __ip_select_ident(net, iph, segs); |
73f156a6 ED |
519 | } |
520 | } | |
521 | ||
b6a7719a HFS |
522 | static inline void ip_select_ident(struct net *net, struct sk_buff *skb, |
523 | struct sock *sk) | |
73f156a6 | 524 | { |
b6a7719a | 525 | ip_select_ident_segs(net, skb, sk, 1); |
1da177e4 LT |
526 | } |
527 | ||
ed70fcfc TH |
528 | static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto) |
529 | { | |
530 | return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr, | |
531 | skb->len, proto, 0); | |
532 | } | |
533 | ||
c3f83241 TH |
534 | /* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store |
535 | * Equivalent to : flow->v4addrs.src = iph->saddr; | |
536 | * flow->v4addrs.dst = iph->daddr; | |
537 | */ | |
538 | static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow, | |
539 | const struct iphdr *iph) | |
540 | { | |
541 | BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) != | |
542 | offsetof(typeof(flow->addrs), v4addrs.src) + | |
543 | sizeof(flow->addrs.v4addrs.src)); | |
544 | memcpy(&flow->addrs.v4addrs, &iph->saddr, sizeof(flow->addrs.v4addrs)); | |
545 | flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS; | |
546 | } | |
547 | ||
1933a785 TH |
548 | static inline __wsum inet_gro_compute_pseudo(struct sk_buff *skb, int proto) |
549 | { | |
550 | const struct iphdr *iph = skb_gro_network_header(skb); | |
551 | ||
552 | return csum_tcpudp_nofold(iph->saddr, iph->daddr, | |
553 | skb_gro_len(skb), proto, 0); | |
554 | } | |
555 | ||
1da177e4 LT |
556 | /* |
557 | * Map a multicast IP onto multicast MAC for type ethernet. | |
558 | */ | |
559 | ||
714e85be | 560 | static inline void ip_eth_mc_map(__be32 naddr, char *buf) |
1da177e4 | 561 | { |
714e85be | 562 | __u32 addr=ntohl(naddr); |
1da177e4 LT |
563 | buf[0]=0x01; |
564 | buf[1]=0x00; | |
565 | buf[2]=0x5e; | |
566 | buf[5]=addr&0xFF; | |
567 | addr>>=8; | |
568 | buf[4]=addr&0xFF; | |
569 | addr>>=8; | |
570 | buf[3]=addr&0x7F; | |
571 | } | |
572 | ||
573 | /* | |
574 | * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand. | |
575 | * Leave P_Key as 0 to be filled in by driver. | |
576 | */ | |
577 | ||
a9e527e3 | 578 | static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) |
1da177e4 | 579 | { |
714e85be | 580 | __u32 addr; |
a9e527e3 RM |
581 | unsigned char scope = broadcast[5] & 0xF; |
582 | ||
1da177e4 LT |
583 | buf[0] = 0; /* Reserved */ |
584 | buf[1] = 0xff; /* Multicast QPN */ | |
585 | buf[2] = 0xff; | |
586 | buf[3] = 0xff; | |
714e85be | 587 | addr = ntohl(naddr); |
1da177e4 | 588 | buf[4] = 0xff; |
a9e527e3 | 589 | buf[5] = 0x10 | scope; /* scope from broadcast address */ |
1da177e4 LT |
590 | buf[6] = 0x40; /* IPv4 signature */ |
591 | buf[7] = 0x1b; | |
a9e527e3 RM |
592 | buf[8] = broadcast[8]; /* P_Key */ |
593 | buf[9] = broadcast[9]; | |
1da177e4 LT |
594 | buf[10] = 0; |
595 | buf[11] = 0; | |
596 | buf[12] = 0; | |
597 | buf[13] = 0; | |
598 | buf[14] = 0; | |
599 | buf[15] = 0; | |
600 | buf[19] = addr & 0xff; | |
601 | addr >>= 8; | |
602 | buf[18] = addr & 0xff; | |
603 | addr >>= 8; | |
604 | buf[17] = addr & 0xff; | |
605 | addr >>= 8; | |
606 | buf[16] = addr & 0x0f; | |
607 | } | |
608 | ||
93ca3bb5 TT |
609 | static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) |
610 | { | |
611 | if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0) | |
612 | memcpy(buf, broadcast, 4); | |
613 | else | |
614 | memcpy(buf, &naddr, sizeof(naddr)); | |
615 | } | |
616 | ||
dfd56b8b | 617 | #if IS_ENABLED(CONFIG_IPV6) |
1da177e4 LT |
618 | #include <linux/ipv6.h> |
619 | #endif | |
620 | ||
621 | static __inline__ void inet_reset_saddr(struct sock *sk) | |
622 | { | |
c720c7e8 | 623 | inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0; |
dfd56b8b | 624 | #if IS_ENABLED(CONFIG_IPV6) |
1da177e4 LT |
625 | if (sk->sk_family == PF_INET6) { |
626 | struct ipv6_pinfo *np = inet6_sk(sk); | |
627 | ||
628 | memset(&np->saddr, 0, sizeof(np->saddr)); | |
efe4208f | 629 | memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr)); |
1da177e4 LT |
630 | } |
631 | #endif | |
632 | } | |
633 | ||
634 | #endif | |
635 | ||
72afa352 DA |
636 | static inline unsigned int ipv4_addr_hash(__be32 ip) |
637 | { | |
638 | return (__force unsigned int) ip; | |
639 | } | |
640 | ||
f0b1e64c MKL |
641 | static inline u32 ipv4_portaddr_hash(const struct net *net, |
642 | __be32 saddr, | |
643 | unsigned int port) | |
644 | { | |
645 | return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port; | |
646 | } | |
647 | ||
5c3a0fd7 | 648 | bool ip_call_ra_chain(struct sk_buff *skb); |
1da177e4 LT |
649 | |
650 | /* | |
b798232f | 651 | * Functions provided by ip_fragment.c |
1da177e4 LT |
652 | */ |
653 | ||
fd2c3ef7 | 654 | enum ip_defrag_users { |
1da177e4 LT |
655 | IP_DEFRAG_LOCAL_DELIVER, |
656 | IP_DEFRAG_CALL_RA_CHAIN, | |
657 | IP_DEFRAG_CONNTRACK_IN, | |
4be929be | 658 | __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX, |
1da177e4 | 659 | IP_DEFRAG_CONNTRACK_OUT, |
4be929be | 660 | __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX, |
8fa9ff68 | 661 | IP_DEFRAG_CONNTRACK_BRIDGE_IN, |
4be929be | 662 | __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX, |
1da177e4 LT |
663 | IP_DEFRAG_VS_IN, |
664 | IP_DEFRAG_VS_OUT, | |
595fc71b DM |
665 | IP_DEFRAG_VS_FWD, |
666 | IP_DEFRAG_AF_PACKET, | |
bc416d97 | 667 | IP_DEFRAG_MACVLAN, |
1da177e4 LT |
668 | }; |
669 | ||
5cf42280 AZ |
670 | /* Return true if the value of 'user' is between 'lower_bond' |
671 | * and 'upper_bond' inclusively. | |
672 | */ | |
673 | static inline bool ip_defrag_user_in_between(u32 user, | |
674 | enum ip_defrag_users lower_bond, | |
675 | enum ip_defrag_users upper_bond) | |
676 | { | |
677 | return user >= lower_bond && user <= upper_bond; | |
678 | } | |
679 | ||
19bcf9f2 | 680 | int ip_defrag(struct net *net, struct sk_buff *skb, u32 user); |
bc416d97 | 681 | #ifdef CONFIG_INET |
19bcf9f2 | 682 | struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user); |
bc416d97 | 683 | #else |
19bcf9f2 | 684 | static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user) |
bc416d97 ED |
685 | { |
686 | return skb; | |
687 | } | |
688 | #endif | |
1da177e4 LT |
689 | |
690 | /* | |
691 | * Functions provided by ip_forward.c | |
692 | */ | |
82695b30 | 693 | |
5c3a0fd7 | 694 | int ip_forward(struct sk_buff *skb); |
82695b30 | 695 | |
1da177e4 LT |
696 | /* |
697 | * Functions provided by ip_options.c | |
698 | */ | |
82695b30 | 699 | |
5c3a0fd7 JP |
700 | void ip_options_build(struct sk_buff *skb, struct ip_options *opt, |
701 | __be32 daddr, struct rtable *rt, int is_frag); | |
24a2d43d | 702 | |
91ed1e66 PA |
703 | int __ip_options_echo(struct net *net, struct ip_options *dopt, |
704 | struct sk_buff *skb, const struct ip_options *sopt); | |
705 | static inline int ip_options_echo(struct net *net, struct ip_options *dopt, | |
706 | struct sk_buff *skb) | |
24a2d43d | 707 | { |
91ed1e66 | 708 | return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt); |
24a2d43d ED |
709 | } |
710 | ||
5c3a0fd7 | 711 | void ip_options_fragment(struct sk_buff *skb); |
3da1ed7a NS |
712 | int __ip_options_compile(struct net *net, struct ip_options *opt, |
713 | struct sk_buff *skb, __be32 *info); | |
5c3a0fd7 JP |
714 | int ip_options_compile(struct net *net, struct ip_options *opt, |
715 | struct sk_buff *skb); | |
716 | int ip_options_get(struct net *net, struct ip_options_rcu **optp, | |
de40a3e8 | 717 | sockptr_t data, int optlen); |
5c3a0fd7 JP |
718 | void ip_options_undo(struct ip_options *opt); |
719 | void ip_forward_options(struct sk_buff *skb); | |
8c83f2df | 720 | int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev); |
1da177e4 LT |
721 | |
722 | /* | |
723 | * Functions provided by ip_sockglue.c | |
724 | */ | |
725 | ||
fbf8866d | 726 | void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb); |
ad959036 PA |
727 | void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk, |
728 | struct sk_buff *skb, int tlen, int offset); | |
24025c46 | 729 | int ip_cmsg_send(struct sock *sk, struct msghdr *msg, |
c8e6ad08 | 730 | struct ipcm_cookie *ipc, bool allow_ipv6); |
a7b75c5a | 731 | int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval, |
5c3a0fd7 JP |
732 | unsigned int optlen); |
733 | int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, | |
734 | int __user *optlen); | |
5c3a0fd7 JP |
735 | int ip_ra_control(struct sock *sk, unsigned char on, |
736 | void (*destructor)(struct sock *)); | |
737 | ||
85fbaa75 | 738 | int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len); |
5c3a0fd7 JP |
739 | void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port, |
740 | u32 info, u8 *payload); | |
741 | void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport, | |
742 | u32 info); | |
1da177e4 | 743 | |
5961de9f TH |
744 | static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb) |
745 | { | |
ad959036 | 746 | ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0); |
5961de9f TH |
747 | } |
748 | ||
4cdf507d ED |
749 | bool icmp_global_allow(void); |
750 | extern int sysctl_icmp_msgs_per_sec; | |
751 | extern int sysctl_icmp_msgs_burst; | |
752 | ||
20380731 | 753 | #ifdef CONFIG_PROC_FS |
5c3a0fd7 | 754 | int ip_misc_proc_init(void); |
20380731 ACM |
755 | #endif |
756 | ||
5e1a99ea | 757 | int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family, |
404eb77e RP |
758 | struct netlink_ext_ack *extack); |
759 | ||
501a90c9 ED |
760 | static inline bool inetdev_valid_mtu(unsigned int mtu) |
761 | { | |
762 | return likely(mtu >= IPV4_MIN_MTU); | |
763 | } | |
764 | ||
c4e446bf | 765 | void ip_sock_set_freebind(struct sock *sk); |
2de569bd | 766 | int ip_sock_set_mtu_discover(struct sock *sk, int val); |
c1f9ec57 | 767 | void ip_sock_set_pktinfo(struct sock *sk); |
db45c0ef | 768 | void ip_sock_set_recverr(struct sock *sk); |
6ebf71ba CH |
769 | void ip_sock_set_tos(struct sock *sk, int val); |
770 | ||
1da177e4 | 771 | #endif /* _IP_H */ |