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> |
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 |
40 | extern unsigned int sysctl_fib_sync_mem; |
41 | extern unsigned int sysctl_fib_sync_mem_min; | |
42 | extern unsigned int sysctl_fib_sync_mem_max; | |
43 | ||
1da177e4 LT |
44 | struct sock; |
45 | ||
fd2c3ef7 | 46 | struct 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) |
5f2d04f1 PM |
59 | |
60 | u16 frag_max_size; | |
1da177e4 LT |
61 | }; |
62 | ||
a04a480d DA |
63 | static inline bool ipv4_l3mdev_skb(u16 flags) |
64 | { | |
65 | return !!(flags & IPSKB_L3SLAVE); | |
66 | } | |
67 | ||
c9bdd4b5 ACM |
68 | static inline unsigned int ip_hdrlen(const struct sk_buff *skb) |
69 | { | |
eddc9ec5 | 70 | return ip_hdr(skb)->ihl * 4; |
c9bdd4b5 ACM |
71 | } |
72 | ||
fd2c3ef7 | 73 | struct ipcm_cookie { |
24025c46 | 74 | struct sockcm_cookie sockc; |
c1d18f9f | 75 | __be32 addr; |
1da177e4 | 76 | int oif; |
f6d8bd05 | 77 | struct ip_options_rcu *opt; |
f02db315 FF |
78 | __u8 ttl; |
79 | __s16 tos; | |
80 | char priority; | |
bec1f6f6 | 81 | __u16 gso_size; |
1da177e4 LT |
82 | }; |
83 | ||
35178206 WB |
84 | static inline void ipcm_init(struct ipcm_cookie *ipcm) |
85 | { | |
86 | *ipcm = (struct ipcm_cookie) { .tos = -1 }; | |
87 | } | |
88 | ||
89 | static inline void ipcm_init_sk(struct ipcm_cookie *ipcm, | |
90 | const struct inet_sock *inet) | |
91 | { | |
92 | ipcm_init(ipcm); | |
93 | ||
c6af0c22 | 94 | ipcm->sockc.mark = inet->sk.sk_mark; |
35178206 WB |
95 | ipcm->sockc.tsflags = inet->sk.sk_tsflags; |
96 | ipcm->oif = inet->sk.sk_bound_dev_if; | |
97 | ipcm->addr = inet->inet_saddr; | |
98 | } | |
99 | ||
1da177e4 | 100 | #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb)) |
4b261c75 | 101 | #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb)) |
1da177e4 | 102 | |
fb74c277 | 103 | /* return enslaved device index if relevant */ |
d6bb2d1e | 104 | static inline int inet_sdif(const struct sk_buff *skb) |
fb74c277 DA |
105 | { |
106 | #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV) | |
107 | if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags)) | |
108 | return IPCB(skb)->iif; | |
109 | #endif | |
110 | return 0; | |
111 | } | |
112 | ||
5796ef75 KT |
113 | /* Special input handler for packets caught by router alert option. |
114 | They are selected only by protocol field, and then processed likely | |
115 | local ones; but only if someone wants them! Otherwise, router | |
116 | not running rsvpd will kill RSVP. | |
117 | ||
118 | It is user level problem, what it will make with them. | |
119 | I have no idea, how it will masquearde or NAT them (it is joke, joke :-)), | |
120 | but receiver should be enough clever f.e. to forward mtrace requests, | |
121 | sent to multicast group to reach destination designated router. | |
122 | */ | |
123 | ||
fd2c3ef7 | 124 | struct ip_ra_chain { |
43a951e9 | 125 | struct ip_ra_chain __rcu *next; |
1da177e4 | 126 | struct sock *sk; |
592fcb9d ED |
127 | union { |
128 | void (*destructor)(struct sock *); | |
129 | struct sock *saved_sk; | |
130 | }; | |
66018506 | 131 | struct rcu_head rcu; |
1da177e4 LT |
132 | }; |
133 | ||
1da177e4 LT |
134 | /* IP flags. */ |
135 | #define IP_CE 0x8000 /* Flag: "Congestion" */ | |
136 | #define IP_DF 0x4000 /* Flag: "Don't Fragment" */ | |
137 | #define IP_MF 0x2000 /* Flag: "More Fragments" */ | |
138 | #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */ | |
139 | ||
140 | #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */ | |
141 | ||
14c85021 ACM |
142 | struct msghdr; |
143 | struct net_device; | |
144 | struct packet_type; | |
145 | struct rtable; | |
14c85021 ACM |
146 | struct sockaddr; |
147 | ||
72c1d3bd | 148 | int igmp_mc_init(void); |
1da177e4 LT |
149 | |
150 | /* | |
151 | * Functions provided by ip.c | |
152 | */ | |
153 | ||
cfe673b0 | 154 | int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk, |
5c3a0fd7 | 155 | __be32 saddr, __be32 daddr, |
de033b7d | 156 | struct ip_options_rcu *opt, u8 tos); |
5c3a0fd7 JP |
157 | int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, |
158 | struct net_device *orig_dev); | |
17266ee9 EC |
159 | void ip_list_rcv(struct list_head *head, struct packet_type *pt, |
160 | struct net_device *orig_dev); | |
5c3a0fd7 | 161 | int ip_local_deliver(struct sk_buff *skb); |
68cb7d53 | 162 | void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto); |
5c3a0fd7 | 163 | int ip_mr_input(struct sk_buff *skb); |
ede2059d EB |
164 | int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb); |
165 | int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb); | |
694869b3 EB |
166 | int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb, |
167 | int (*output)(struct net *, struct sock *, struct sk_buff *)); | |
c8b17be0 PNA |
168 | |
169 | struct ip_fraglist_iter { | |
c8b17be0 PNA |
170 | struct sk_buff *frag; |
171 | struct iphdr *iph; | |
172 | int offset; | |
173 | unsigned int hlen; | |
174 | }; | |
175 | ||
176 | void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph, | |
177 | unsigned int hlen, struct ip_fraglist_iter *iter); | |
178 | void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter); | |
179 | ||
180 | static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter) | |
181 | { | |
182 | struct sk_buff *skb = iter->frag; | |
183 | ||
184 | iter->frag = skb->next; | |
185 | skb_mark_not_on_list(skb); | |
186 | ||
187 | return skb; | |
188 | } | |
189 | ||
065ff79f | 190 | struct ip_frag_state { |
e7a409c3 | 191 | bool DF; |
065ff79f PNA |
192 | unsigned int hlen; |
193 | unsigned int ll_rs; | |
194 | unsigned int mtu; | |
195 | unsigned int left; | |
196 | int offset; | |
197 | int ptr; | |
198 | __be16 not_last_frag; | |
199 | }; | |
200 | ||
201 | void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs, | |
e7a409c3 | 202 | unsigned int mtu, bool DF, struct ip_frag_state *state); |
065ff79f PNA |
203 | struct sk_buff *ip_frag_next(struct sk_buff *skb, |
204 | struct ip_frag_state *state); | |
205 | ||
5c3a0fd7 | 206 | void ip_send_check(struct iphdr *ip); |
cf91a99d | 207 | int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb); |
33224b16 | 208 | int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb); |
aad88724 | 209 | |
69b9e1e0 XL |
210 | int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl, |
211 | __u8 tos); | |
5c3a0fd7 JP |
212 | void ip_init(void); |
213 | int ip_append_data(struct sock *sk, struct flowi4 *fl4, | |
214 | int getfrag(void *from, char *to, int offset, int len, | |
215 | int odd, struct sk_buff *skb), | |
216 | void *from, int len, int protolen, | |
217 | struct ipcm_cookie *ipc, | |
218 | struct rtable **rt, | |
219 | unsigned int flags); | |
220 | int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, | |
221 | struct sk_buff *skb); | |
222 | ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page, | |
223 | int offset, size_t size, int flags); | |
224 | struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4, | |
225 | struct sk_buff_head *queue, | |
226 | struct inet_cork *cork); | |
227 | int ip_send_skb(struct net *net, struct sk_buff *skb); | |
228 | int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4); | |
229 | void ip_flush_pending_frames(struct sock *sk); | |
230 | struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4, | |
231 | int getfrag(void *from, char *to, int offset, | |
232 | int len, int odd, struct sk_buff *skb), | |
233 | void *from, int length, int transhdrlen, | |
234 | struct ipcm_cookie *ipc, struct rtable **rtp, | |
1cd7884d | 235 | struct inet_cork *cork, unsigned int flags); |
1c32c5ad | 236 | |
05e22e83 | 237 | int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl); |
69b9e1e0 | 238 | |
77968b78 | 239 | static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4) |
1c32c5ad | 240 | { |
77968b78 | 241 | return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base); |
1c32c5ad | 242 | } |
1da177e4 | 243 | |
aa661581 FF |
244 | static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet) |
245 | { | |
246 | return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos); | |
247 | } | |
248 | ||
249 | static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk) | |
250 | { | |
251 | return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk); | |
252 | } | |
253 | ||
1da177e4 | 254 | /* datagram.c */ |
03645a11 | 255 | int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); |
5c3a0fd7 | 256 | int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); |
1da177e4 | 257 | |
5c3a0fd7 | 258 | void ip4_datagram_release_cb(struct sock *sk); |
8141ed9f | 259 | |
1da177e4 | 260 | struct ip_reply_arg { |
82695b30 | 261 | struct kvec iov[1]; |
88ef4a5a | 262 | int flags; |
d6f5493c | 263 | __wsum csum; |
1da177e4 | 264 | int csumoffset; /* u16 offset of csum in iov[0].iov_base */ |
82695b30 | 265 | /* -1 if not needed */ |
f0e48dbf | 266 | int bound_dev_if; |
66b13d99 | 267 | u8 tos; |
e2d118a1 | 268 | kuid_t uid; |
82695b30 | 269 | }; |
1da177e4 | 270 | |
88ef4a5a KK |
271 | #define IP_REPLY_ARG_NOSRCCHECK 1 |
272 | ||
86b08d86 KK |
273 | static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg) |
274 | { | |
275 | return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0; | |
276 | } | |
277 | ||
bdbbb852 | 278 | void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb, |
24a2d43d ED |
279 | const struct ip_options *sopt, |
280 | __be32 daddr, __be32 saddr, | |
281 | const struct ip_reply_arg *arg, | |
d6fb396c | 282 | unsigned int len, u64 transmit_time); |
1da177e4 | 283 | |
4ce3c183 | 284 | #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field) |
13415e46 | 285 | #define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field) |
4ce3c183 | 286 | #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) |
13415e46 | 287 | #define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) |
4ce3c183 | 288 | #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) |
13415e46 | 289 | #define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) |
61a7e260 | 290 | #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field) |
13415e46 | 291 | #define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field) |
f7324acd | 292 | #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd) |
13415e46 | 293 | #define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd) |
1da177e4 | 294 | |
59f09ae8 ED |
295 | static inline u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offt) |
296 | { | |
297 | return *(((unsigned long *)per_cpu_ptr(mib, cpu)) + offt); | |
298 | } | |
299 | ||
698365fa | 300 | unsigned long snmp_fold_field(void __percpu *mib, int offt); |
4ce3c183 | 301 | #if BITS_PER_LONG==32 |
c4c6bc31 R |
302 | u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct, |
303 | size_t syncp_offset); | |
698365fa | 304 | u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off); |
4ce3c183 | 305 | #else |
c4c6bc31 R |
306 | static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct, |
307 | size_t syncp_offset) | |
308 | { | |
309 | return snmp_get_cpu_field(mib, cpu, offct); | |
310 | ||
311 | } | |
312 | ||
698365fa | 313 | static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off) |
4ce3c183 ED |
314 | { |
315 | return snmp_fold_field(mib, offt); | |
316 | } | |
317 | #endif | |
33490170 | 318 | |
6348ef2d JH |
319 | #define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \ |
320 | { \ | |
321 | int i, c; \ | |
322 | for_each_possible_cpu(c) { \ | |
323 | for (i = 0; stats_list[i].name; i++) \ | |
324 | buff64[i] += snmp_get_cpu_field64( \ | |
325 | mib_statistic, \ | |
326 | c, stats_list[i].entry, \ | |
327 | offset); \ | |
328 | } \ | |
329 | } | |
330 | ||
331 | #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \ | |
332 | { \ | |
333 | int i, c; \ | |
334 | for_each_possible_cpu(c) { \ | |
335 | for (i = 0; stats_list[i].name; i++) \ | |
336 | buff[i] += snmp_get_cpu_field( \ | |
337 | mib_statistic, \ | |
338 | c, stats_list[i].entry); \ | |
339 | } \ | |
340 | } | |
341 | ||
0bbf87d8 | 342 | void inet_get_local_port_range(struct net *net, int *low, int *high); |
227b60f5 | 343 | |
fcd77db0 | 344 | #ifdef CONFIG_SYSCTL |
66e2f5f7 | 345 | static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) |
e3826f1e | 346 | { |
122ff243 | 347 | if (!net->ipv4.sysctl_local_reserved_ports) |
30429fba | 348 | return false; |
122ff243 | 349 | return test_bit(port, net->ipv4.sysctl_local_reserved_ports); |
e3826f1e | 350 | } |
20e61da7 WC |
351 | |
352 | static inline bool sysctl_dev_name_is_allowed(const char *name) | |
353 | { | |
354 | return strcmp(name, "default") != 0 && strcmp(name, "all") != 0; | |
355 | } | |
356 | ||
82f31ebf | 357 | static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port) |
4548b683 | 358 | { |
82f31ebf | 359 | return port < net->ipv4.sysctl_ip_prot_sock; |
4548b683 KJ |
360 | } |
361 | ||
122ff243 | 362 | #else |
66e2f5f7 | 363 | static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) |
122ff243 | 364 | { |
30429fba | 365 | return false; |
122ff243 | 366 | } |
4548b683 | 367 | |
82f31ebf | 368 | static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port) |
4548b683 | 369 | { |
82f31ebf | 370 | return port < PROT_SOCK; |
4548b683 | 371 | } |
122ff243 | 372 | #endif |
e3826f1e | 373 | |
822c8685 DD |
374 | __be32 inet_current_timestamp(void); |
375 | ||
20380731 ACM |
376 | /* From inetpeer.c */ |
377 | extern int inet_peer_threshold; | |
378 | extern int inet_peer_minttl; | |
379 | extern int inet_peer_maxttl; | |
20380731 | 380 | |
5c3a0fd7 | 381 | void ipfrag_init(void); |
20380731 | 382 | |
5c3a0fd7 | 383 | void ip_static_sysctl_init(void); |
bd7b1533 | 384 | |
e110861f LC |
385 | #define IP4_REPLY_MARK(net, mark) \ |
386 | ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0) | |
387 | ||
d18cd551 DM |
388 | static inline bool ip_is_fragment(const struct iphdr *iph) |
389 | { | |
390 | return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0; | |
391 | } | |
392 | ||
1da177e4 | 393 | #ifdef CONFIG_INET |
14c85021 ACM |
394 | #include <net/dst.h> |
395 | ||
1da177e4 LT |
396 | /* The function in 2.2 was invalid, producing wrong result for |
397 | * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */ | |
398 | static inline | |
399 | int ip_decrease_ttl(struct iphdr *iph) | |
400 | { | |
5c78f275 AV |
401 | u32 check = (__force u32)iph->check; |
402 | check += (__force u32)htons(0x0100); | |
403 | iph->check = (__force __sum16)(check + (check>=0xFFFF)); | |
1da177e4 LT |
404 | return --iph->ttl; |
405 | } | |
406 | ||
d52e5a7e SD |
407 | static inline int ip_mtu_locked(const struct dst_entry *dst) |
408 | { | |
409 | const struct rtable *rt = (const struct rtable *)dst; | |
410 | ||
411 | return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU); | |
412 | } | |
413 | ||
1da177e4 | 414 | static inline |
4e3f5d72 | 415 | int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst) |
1da177e4 | 416 | { |
4e3f5d72 ED |
417 | u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc); |
418 | ||
419 | return pmtudisc == IP_PMTUDISC_DO || | |
420 | (pmtudisc == IP_PMTUDISC_WANT && | |
d52e5a7e | 421 | !ip_mtu_locked(dst)); |
1da177e4 LT |
422 | } |
423 | ||
f87c10a8 HFS |
424 | static inline bool ip_sk_accept_pmtu(const struct sock *sk) |
425 | { | |
1b346576 HFS |
426 | return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE && |
427 | inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT; | |
f87c10a8 HFS |
428 | } |
429 | ||
430 | static inline bool ip_sk_use_pmtu(const struct sock *sk) | |
431 | { | |
432 | return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE; | |
433 | } | |
434 | ||
60ff7467 | 435 | static inline bool ip_sk_ignore_df(const struct sock *sk) |
1b346576 HFS |
436 | { |
437 | return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO || | |
438 | inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT; | |
439 | } | |
440 | ||
f87c10a8 HFS |
441 | static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst, |
442 | bool forwarding) | |
443 | { | |
ac6627a2 | 444 | const struct rtable *rt = container_of(dst, struct rtable, dst); |
f87c10a8 | 445 | struct net *net = dev_net(dst->dev); |
02a1b175 | 446 | unsigned int mtu; |
f87c10a8 HFS |
447 | |
448 | if (net->ipv4.sysctl_ip_fwd_use_pmtu || | |
d52e5a7e | 449 | ip_mtu_locked(dst) || |
ac6627a2 VF |
450 | !forwarding) { |
451 | mtu = rt->rt_pmtu; | |
452 | if (mtu && time_before(jiffies, rt->dst.expires)) | |
453 | goto out; | |
454 | } | |
f87c10a8 | 455 | |
02a1b175 MÅ» |
456 | /* 'forwarding = true' case should always honour route mtu */ |
457 | mtu = dst_metric_raw(dst, RTAX_MTU); | |
ac6627a2 VF |
458 | if (mtu) |
459 | goto out; | |
460 | ||
461 | mtu = READ_ONCE(dst->dev->mtu); | |
462 | ||
463 | if (unlikely(ip_mtu_locked(dst))) { | |
464 | if (rt->rt_uses_gateway && mtu > 576) | |
465 | mtu = 576; | |
466 | } | |
467 | ||
468 | out: | |
469 | mtu = min_t(unsigned int, mtu, IP_MAX_MTU); | |
02a1b175 | 470 | |
fade5641 | 471 | return mtu - lwtunnel_headroom(dst->lwtstate, mtu); |
f87c10a8 HFS |
472 | } |
473 | ||
fedbb6b4 SL |
474 | static inline unsigned int ip_skb_dst_mtu(struct sock *sk, |
475 | const struct sk_buff *skb) | |
f87c10a8 | 476 | { |
fade5641 VF |
477 | unsigned int mtu; |
478 | ||
caf3f267 | 479 | if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) { |
f87c10a8 | 480 | bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED; |
caf3f267 | 481 | |
f87c10a8 | 482 | return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding); |
f87c10a8 | 483 | } |
caf3f267 | 484 | |
fade5641 VF |
485 | mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU); |
486 | return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu); | |
f87c10a8 HFS |
487 | } |
488 | ||
767a2217 | 489 | struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx, |
d7e774f3 DA |
490 | int fc_mx_len, |
491 | struct netlink_ext_ack *extack); | |
cc5f0eb2 DA |
492 | static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics) |
493 | { | |
494 | if (fib_metrics != &dst_default_metrics && | |
495 | refcount_dec_and_test(&fib_metrics->refcnt)) | |
496 | kfree(fib_metrics); | |
497 | } | |
a919525a | 498 | |
e1255ed4 DA |
499 | /* ipv4 and ipv6 both use refcounted metrics if it is not the default */ |
500 | static inline | |
501 | void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics) | |
502 | { | |
503 | dst_init_metrics(dst, fib_metrics->metrics, true); | |
504 | ||
505 | if (fib_metrics != &dst_default_metrics) { | |
506 | dst->_metrics |= DST_METRICS_REFCOUNTED; | |
507 | refcount_inc(&fib_metrics->refcnt); | |
508 | } | |
509 | } | |
510 | ||
1620a336 DA |
511 | static inline |
512 | void ip_dst_metrics_put(struct dst_entry *dst) | |
513 | { | |
514 | struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst); | |
515 | ||
516 | if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt)) | |
517 | kfree(p); | |
518 | } | |
519 | ||
04ca6973 | 520 | u32 ip_idents_reserve(u32 hash, int segs); |
b6a7719a | 521 | void __ip_select_ident(struct net *net, struct iphdr *iph, int segs); |
73f156a6 | 522 | |
b6a7719a HFS |
523 | static 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 | ||
60ff7467 | 528 | if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) { |
1da177e4 LT |
529 | /* This is only to work around buggy Windows95/2000 |
530 | * VJ compression implementations. If the ID field | |
531 | * does not change, they drop every other packet in | |
532 | * a TCP stream using header compression. | |
533 | */ | |
c720c7e8 ED |
534 | if (sk && inet_sk(sk)->inet_daddr) { |
535 | iph->id = htons(inet_sk(sk)->inet_id); | |
73f156a6 ED |
536 | inet_sk(sk)->inet_id += segs; |
537 | } else { | |
1da177e4 | 538 | iph->id = 0; |
73f156a6 ED |
539 | } |
540 | } else { | |
b6a7719a | 541 | __ip_select_ident(net, iph, segs); |
73f156a6 ED |
542 | } |
543 | } | |
544 | ||
b6a7719a HFS |
545 | static inline void ip_select_ident(struct net *net, struct sk_buff *skb, |
546 | struct sock *sk) | |
73f156a6 | 547 | { |
b6a7719a | 548 | ip_select_ident_segs(net, skb, sk, 1); |
1da177e4 LT |
549 | } |
550 | ||
ed70fcfc TH |
551 | static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto) |
552 | { | |
553 | return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr, | |
554 | skb->len, proto, 0); | |
555 | } | |
556 | ||
c3f83241 TH |
557 | /* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store |
558 | * Equivalent to : flow->v4addrs.src = iph->saddr; | |
559 | * flow->v4addrs.dst = iph->daddr; | |
560 | */ | |
561 | static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow, | |
562 | const struct iphdr *iph) | |
563 | { | |
564 | BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) != | |
565 | offsetof(typeof(flow->addrs), v4addrs.src) + | |
566 | sizeof(flow->addrs.v4addrs.src)); | |
567 | memcpy(&flow->addrs.v4addrs, &iph->saddr, sizeof(flow->addrs.v4addrs)); | |
568 | flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS; | |
569 | } | |
570 | ||
1933a785 TH |
571 | static inline __wsum inet_gro_compute_pseudo(struct sk_buff *skb, int proto) |
572 | { | |
573 | const struct iphdr *iph = skb_gro_network_header(skb); | |
574 | ||
575 | return csum_tcpudp_nofold(iph->saddr, iph->daddr, | |
576 | skb_gro_len(skb), proto, 0); | |
577 | } | |
578 | ||
1da177e4 LT |
579 | /* |
580 | * Map a multicast IP onto multicast MAC for type ethernet. | |
581 | */ | |
582 | ||
714e85be | 583 | static inline void ip_eth_mc_map(__be32 naddr, char *buf) |
1da177e4 | 584 | { |
714e85be | 585 | __u32 addr=ntohl(naddr); |
1da177e4 LT |
586 | buf[0]=0x01; |
587 | buf[1]=0x00; | |
588 | buf[2]=0x5e; | |
589 | buf[5]=addr&0xFF; | |
590 | addr>>=8; | |
591 | buf[4]=addr&0xFF; | |
592 | addr>>=8; | |
593 | buf[3]=addr&0x7F; | |
594 | } | |
595 | ||
596 | /* | |
597 | * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand. | |
598 | * Leave P_Key as 0 to be filled in by driver. | |
599 | */ | |
600 | ||
a9e527e3 | 601 | static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) |
1da177e4 | 602 | { |
714e85be | 603 | __u32 addr; |
a9e527e3 RM |
604 | unsigned char scope = broadcast[5] & 0xF; |
605 | ||
1da177e4 LT |
606 | buf[0] = 0; /* Reserved */ |
607 | buf[1] = 0xff; /* Multicast QPN */ | |
608 | buf[2] = 0xff; | |
609 | buf[3] = 0xff; | |
714e85be | 610 | addr = ntohl(naddr); |
1da177e4 | 611 | buf[4] = 0xff; |
a9e527e3 | 612 | buf[5] = 0x10 | scope; /* scope from broadcast address */ |
1da177e4 LT |
613 | buf[6] = 0x40; /* IPv4 signature */ |
614 | buf[7] = 0x1b; | |
a9e527e3 RM |
615 | buf[8] = broadcast[8]; /* P_Key */ |
616 | buf[9] = broadcast[9]; | |
1da177e4 LT |
617 | buf[10] = 0; |
618 | buf[11] = 0; | |
619 | buf[12] = 0; | |
620 | buf[13] = 0; | |
621 | buf[14] = 0; | |
622 | buf[15] = 0; | |
623 | buf[19] = addr & 0xff; | |
624 | addr >>= 8; | |
625 | buf[18] = addr & 0xff; | |
626 | addr >>= 8; | |
627 | buf[17] = addr & 0xff; | |
628 | addr >>= 8; | |
629 | buf[16] = addr & 0x0f; | |
630 | } | |
631 | ||
93ca3bb5 TT |
632 | static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) |
633 | { | |
634 | if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0) | |
635 | memcpy(buf, broadcast, 4); | |
636 | else | |
637 | memcpy(buf, &naddr, sizeof(naddr)); | |
638 | } | |
639 | ||
dfd56b8b | 640 | #if IS_ENABLED(CONFIG_IPV6) |
1da177e4 LT |
641 | #include <linux/ipv6.h> |
642 | #endif | |
643 | ||
644 | static __inline__ void inet_reset_saddr(struct sock *sk) | |
645 | { | |
c720c7e8 | 646 | inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0; |
dfd56b8b | 647 | #if IS_ENABLED(CONFIG_IPV6) |
1da177e4 LT |
648 | if (sk->sk_family == PF_INET6) { |
649 | struct ipv6_pinfo *np = inet6_sk(sk); | |
650 | ||
651 | memset(&np->saddr, 0, sizeof(np->saddr)); | |
efe4208f | 652 | memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr)); |
1da177e4 LT |
653 | } |
654 | #endif | |
655 | } | |
656 | ||
657 | #endif | |
658 | ||
72afa352 DA |
659 | static inline unsigned int ipv4_addr_hash(__be32 ip) |
660 | { | |
661 | return (__force unsigned int) ip; | |
662 | } | |
663 | ||
f0b1e64c MKL |
664 | static inline u32 ipv4_portaddr_hash(const struct net *net, |
665 | __be32 saddr, | |
666 | unsigned int port) | |
667 | { | |
668 | return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port; | |
669 | } | |
670 | ||
5c3a0fd7 | 671 | bool ip_call_ra_chain(struct sk_buff *skb); |
1da177e4 LT |
672 | |
673 | /* | |
b798232f | 674 | * Functions provided by ip_fragment.c |
1da177e4 LT |
675 | */ |
676 | ||
fd2c3ef7 | 677 | enum ip_defrag_users { |
1da177e4 LT |
678 | IP_DEFRAG_LOCAL_DELIVER, |
679 | IP_DEFRAG_CALL_RA_CHAIN, | |
680 | IP_DEFRAG_CONNTRACK_IN, | |
4be929be | 681 | __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX, |
1da177e4 | 682 | IP_DEFRAG_CONNTRACK_OUT, |
4be929be | 683 | __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX, |
8fa9ff68 | 684 | IP_DEFRAG_CONNTRACK_BRIDGE_IN, |
4be929be | 685 | __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX, |
1da177e4 LT |
686 | IP_DEFRAG_VS_IN, |
687 | IP_DEFRAG_VS_OUT, | |
595fc71b DM |
688 | IP_DEFRAG_VS_FWD, |
689 | IP_DEFRAG_AF_PACKET, | |
bc416d97 | 690 | IP_DEFRAG_MACVLAN, |
1da177e4 LT |
691 | }; |
692 | ||
5cf42280 AZ |
693 | /* Return true if the value of 'user' is between 'lower_bond' |
694 | * and 'upper_bond' inclusively. | |
695 | */ | |
696 | static inline bool ip_defrag_user_in_between(u32 user, | |
697 | enum ip_defrag_users lower_bond, | |
698 | enum ip_defrag_users upper_bond) | |
699 | { | |
700 | return user >= lower_bond && user <= upper_bond; | |
701 | } | |
702 | ||
19bcf9f2 | 703 | int ip_defrag(struct net *net, struct sk_buff *skb, u32 user); |
bc416d97 | 704 | #ifdef CONFIG_INET |
19bcf9f2 | 705 | struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user); |
bc416d97 | 706 | #else |
19bcf9f2 | 707 | static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user) |
bc416d97 ED |
708 | { |
709 | return skb; | |
710 | } | |
711 | #endif | |
1da177e4 LT |
712 | |
713 | /* | |
714 | * Functions provided by ip_forward.c | |
715 | */ | |
82695b30 | 716 | |
5c3a0fd7 | 717 | int ip_forward(struct sk_buff *skb); |
82695b30 | 718 | |
1da177e4 LT |
719 | /* |
720 | * Functions provided by ip_options.c | |
721 | */ | |
82695b30 | 722 | |
5c3a0fd7 JP |
723 | void ip_options_build(struct sk_buff *skb, struct ip_options *opt, |
724 | __be32 daddr, struct rtable *rt, int is_frag); | |
24a2d43d | 725 | |
91ed1e66 PA |
726 | int __ip_options_echo(struct net *net, struct ip_options *dopt, |
727 | struct sk_buff *skb, const struct ip_options *sopt); | |
728 | static inline int ip_options_echo(struct net *net, struct ip_options *dopt, | |
729 | struct sk_buff *skb) | |
24a2d43d | 730 | { |
91ed1e66 | 731 | return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt); |
24a2d43d ED |
732 | } |
733 | ||
5c3a0fd7 | 734 | void ip_options_fragment(struct sk_buff *skb); |
3da1ed7a NS |
735 | int __ip_options_compile(struct net *net, struct ip_options *opt, |
736 | struct sk_buff *skb, __be32 *info); | |
5c3a0fd7 JP |
737 | int ip_options_compile(struct net *net, struct ip_options *opt, |
738 | struct sk_buff *skb); | |
739 | int ip_options_get(struct net *net, struct ip_options_rcu **optp, | |
de40a3e8 | 740 | sockptr_t data, int optlen); |
5c3a0fd7 JP |
741 | void ip_options_undo(struct ip_options *opt); |
742 | void ip_forward_options(struct sk_buff *skb); | |
8c83f2df | 743 | int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev); |
1da177e4 LT |
744 | |
745 | /* | |
746 | * Functions provided by ip_sockglue.c | |
747 | */ | |
748 | ||
fbf8866d | 749 | void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb); |
ad959036 PA |
750 | void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk, |
751 | struct sk_buff *skb, int tlen, int offset); | |
24025c46 | 752 | int ip_cmsg_send(struct sock *sk, struct msghdr *msg, |
c8e6ad08 | 753 | struct ipcm_cookie *ipc, bool allow_ipv6); |
020e71a3 | 754 | DECLARE_STATIC_KEY_FALSE(ip4_min_ttl); |
a7b75c5a | 755 | int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval, |
5c3a0fd7 JP |
756 | unsigned int optlen); |
757 | int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, | |
758 | int __user *optlen); | |
5c3a0fd7 JP |
759 | int ip_ra_control(struct sock *sk, unsigned char on, |
760 | void (*destructor)(struct sock *)); | |
761 | ||
85fbaa75 | 762 | int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len); |
5c3a0fd7 JP |
763 | void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port, |
764 | u32 info, u8 *payload); | |
765 | void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport, | |
766 | u32 info); | |
1da177e4 | 767 | |
5961de9f TH |
768 | static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb) |
769 | { | |
ad959036 | 770 | ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0); |
5961de9f TH |
771 | } |
772 | ||
4cdf507d ED |
773 | bool icmp_global_allow(void); |
774 | extern int sysctl_icmp_msgs_per_sec; | |
775 | extern int sysctl_icmp_msgs_burst; | |
776 | ||
20380731 | 777 | #ifdef CONFIG_PROC_FS |
5c3a0fd7 | 778 | int ip_misc_proc_init(void); |
20380731 ACM |
779 | #endif |
780 | ||
5e1a99ea | 781 | int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family, |
404eb77e RP |
782 | struct netlink_ext_ack *extack); |
783 | ||
501a90c9 ED |
784 | static inline bool inetdev_valid_mtu(unsigned int mtu) |
785 | { | |
786 | return likely(mtu >= IPV4_MIN_MTU); | |
787 | } | |
788 | ||
c4e446bf | 789 | void ip_sock_set_freebind(struct sock *sk); |
2de569bd | 790 | int ip_sock_set_mtu_discover(struct sock *sk, int val); |
c1f9ec57 | 791 | void ip_sock_set_pktinfo(struct sock *sk); |
db45c0ef | 792 | void ip_sock_set_recverr(struct sock *sk); |
6ebf71ba CH |
793 | void ip_sock_set_tos(struct sock *sk, int val); |
794 | ||
1da177e4 | 795 | #endif /* _IP_H */ |