]>
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> |
fade5641 | 34 | #include <net/lwtunnel.h> |
1da177e4 | 35 | |
7ed14d97 | 36 | #define IPV4_MAX_PMTU 65535U /* RFC 2675, Section 5.1 */ |
b5476022 | 37 | #define IPV4_MIN_MTU 68 /* RFC 791 */ |
7ed14d97 | 38 | |
9ab948a9 DA |
39 | extern unsigned int sysctl_fib_sync_mem; |
40 | extern unsigned int sysctl_fib_sync_mem_min; | |
41 | extern unsigned int sysctl_fib_sync_mem_max; | |
42 | ||
1da177e4 LT |
43 | struct sock; |
44 | ||
fd2c3ef7 | 45 | struct inet_skb_parm { |
0b922b7a | 46 | int iif; |
1da177e4 | 47 | struct ip_options opt; /* Compiled IP options */ |
a04a480d | 48 | u16 flags; |
1da177e4 | 49 | |
df4d9254 HFS |
50 | #define IPSKB_FORWARDED BIT(0) |
51 | #define IPSKB_XFRM_TUNNEL_SIZE BIT(1) | |
52 | #define IPSKB_XFRM_TRANSFORMED BIT(2) | |
53 | #define IPSKB_FRAG_COMPLETE BIT(3) | |
54 | #define IPSKB_REROUTED BIT(4) | |
55 | #define IPSKB_DOREDIRECT BIT(5) | |
d6b915e2 | 56 | #define IPSKB_FRAG_PMTU BIT(6) |
9ee6c5dc | 57 | #define IPSKB_L3SLAVE BIT(7) |
5f2d04f1 PM |
58 | |
59 | u16 frag_max_size; | |
1da177e4 LT |
60 | }; |
61 | ||
a04a480d DA |
62 | static inline bool ipv4_l3mdev_skb(u16 flags) |
63 | { | |
64 | return !!(flags & IPSKB_L3SLAVE); | |
65 | } | |
66 | ||
c9bdd4b5 ACM |
67 | static inline unsigned int ip_hdrlen(const struct sk_buff *skb) |
68 | { | |
eddc9ec5 | 69 | return ip_hdr(skb)->ihl * 4; |
c9bdd4b5 ACM |
70 | } |
71 | ||
fd2c3ef7 | 72 | struct ipcm_cookie { |
24025c46 | 73 | struct sockcm_cookie sockc; |
c1d18f9f | 74 | __be32 addr; |
1da177e4 | 75 | int oif; |
f6d8bd05 | 76 | struct ip_options_rcu *opt; |
f02db315 FF |
77 | __u8 ttl; |
78 | __s16 tos; | |
79 | char priority; | |
bec1f6f6 | 80 | __u16 gso_size; |
1da177e4 LT |
81 | }; |
82 | ||
35178206 WB |
83 | static inline void ipcm_init(struct ipcm_cookie *ipcm) |
84 | { | |
85 | *ipcm = (struct ipcm_cookie) { .tos = -1 }; | |
86 | } | |
87 | ||
88 | static inline void ipcm_init_sk(struct ipcm_cookie *ipcm, | |
89 | const struct inet_sock *inet) | |
90 | { | |
91 | ipcm_init(ipcm); | |
92 | ||
c6af0c22 | 93 | ipcm->sockc.mark = inet->sk.sk_mark; |
35178206 WB |
94 | ipcm->sockc.tsflags = inet->sk.sk_tsflags; |
95 | ipcm->oif = inet->sk.sk_bound_dev_if; | |
96 | ipcm->addr = inet->inet_saddr; | |
97 | } | |
98 | ||
1da177e4 | 99 | #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb)) |
4b261c75 | 100 | #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb)) |
1da177e4 | 101 | |
fb74c277 | 102 | /* return enslaved device index if relevant */ |
d6bb2d1e | 103 | static inline int inet_sdif(const struct sk_buff *skb) |
fb74c277 DA |
104 | { |
105 | #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV) | |
106 | if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags)) | |
107 | return IPCB(skb)->iif; | |
108 | #endif | |
109 | return 0; | |
110 | } | |
111 | ||
5796ef75 KT |
112 | /* Special input handler for packets caught by router alert option. |
113 | They are selected only by protocol field, and then processed likely | |
114 | local ones; but only if someone wants them! Otherwise, router | |
115 | not running rsvpd will kill RSVP. | |
116 | ||
117 | It is user level problem, what it will make with them. | |
118 | I have no idea, how it will masquearde or NAT them (it is joke, joke :-)), | |
119 | but receiver should be enough clever f.e. to forward mtrace requests, | |
120 | sent to multicast group to reach destination designated router. | |
121 | */ | |
122 | ||
fd2c3ef7 | 123 | struct ip_ra_chain { |
43a951e9 | 124 | struct ip_ra_chain __rcu *next; |
1da177e4 | 125 | struct sock *sk; |
592fcb9d ED |
126 | union { |
127 | void (*destructor)(struct sock *); | |
128 | struct sock *saved_sk; | |
129 | }; | |
66018506 | 130 | struct rcu_head rcu; |
1da177e4 LT |
131 | }; |
132 | ||
1da177e4 LT |
133 | /* IP flags. */ |
134 | #define IP_CE 0x8000 /* Flag: "Congestion" */ | |
135 | #define IP_DF 0x4000 /* Flag: "Don't Fragment" */ | |
136 | #define IP_MF 0x2000 /* Flag: "More Fragments" */ | |
137 | #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */ | |
138 | ||
139 | #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */ | |
140 | ||
14c85021 ACM |
141 | struct msghdr; |
142 | struct net_device; | |
143 | struct packet_type; | |
144 | struct rtable; | |
14c85021 ACM |
145 | struct sockaddr; |
146 | ||
72c1d3bd | 147 | int igmp_mc_init(void); |
1da177e4 LT |
148 | |
149 | /* | |
150 | * Functions provided by ip.c | |
151 | */ | |
152 | ||
cfe673b0 | 153 | int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk, |
5c3a0fd7 | 154 | __be32 saddr, __be32 daddr, |
de033b7d | 155 | struct ip_options_rcu *opt, u8 tos); |
5c3a0fd7 JP |
156 | int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, |
157 | struct net_device *orig_dev); | |
17266ee9 EC |
158 | void ip_list_rcv(struct list_head *head, struct packet_type *pt, |
159 | struct net_device *orig_dev); | |
5c3a0fd7 | 160 | int ip_local_deliver(struct sk_buff *skb); |
68cb7d53 | 161 | void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto); |
5c3a0fd7 | 162 | int ip_mr_input(struct sk_buff *skb); |
ede2059d EB |
163 | int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb); |
164 | int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb); | |
694869b3 EB |
165 | int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb, |
166 | int (*output)(struct net *, struct sock *, struct sk_buff *)); | |
c8b17be0 PNA |
167 | |
168 | struct ip_fraglist_iter { | |
c8b17be0 PNA |
169 | struct sk_buff *frag; |
170 | struct iphdr *iph; | |
171 | int offset; | |
172 | unsigned int hlen; | |
173 | }; | |
174 | ||
175 | void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph, | |
176 | unsigned int hlen, struct ip_fraglist_iter *iter); | |
177 | void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter); | |
178 | ||
179 | static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter) | |
180 | { | |
181 | struct sk_buff *skb = iter->frag; | |
182 | ||
183 | iter->frag = skb->next; | |
184 | skb_mark_not_on_list(skb); | |
185 | ||
186 | return skb; | |
187 | } | |
188 | ||
065ff79f | 189 | struct ip_frag_state { |
e7a409c3 | 190 | bool DF; |
065ff79f PNA |
191 | unsigned int hlen; |
192 | unsigned int ll_rs; | |
193 | unsigned int mtu; | |
194 | unsigned int left; | |
195 | int offset; | |
196 | int ptr; | |
197 | __be16 not_last_frag; | |
198 | }; | |
199 | ||
200 | void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs, | |
e7a409c3 | 201 | unsigned int mtu, bool DF, struct ip_frag_state *state); |
065ff79f PNA |
202 | struct sk_buff *ip_frag_next(struct sk_buff *skb, |
203 | struct ip_frag_state *state); | |
204 | ||
5c3a0fd7 | 205 | void ip_send_check(struct iphdr *ip); |
cf91a99d | 206 | int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb); |
33224b16 | 207 | int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb); |
aad88724 | 208 | |
69b9e1e0 XL |
209 | int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl, |
210 | __u8 tos); | |
5c3a0fd7 JP |
211 | void ip_init(void); |
212 | int ip_append_data(struct sock *sk, struct flowi4 *fl4, | |
213 | int getfrag(void *from, char *to, int offset, int len, | |
214 | int odd, struct sk_buff *skb), | |
215 | void *from, int len, int protolen, | |
216 | struct ipcm_cookie *ipc, | |
217 | struct rtable **rt, | |
218 | unsigned int flags); | |
219 | int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, | |
220 | struct sk_buff *skb); | |
221 | ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page, | |
222 | int offset, size_t size, int flags); | |
223 | struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4, | |
224 | struct sk_buff_head *queue, | |
225 | struct inet_cork *cork); | |
226 | int ip_send_skb(struct net *net, struct sk_buff *skb); | |
227 | int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4); | |
228 | void ip_flush_pending_frames(struct sock *sk); | |
229 | struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4, | |
230 | int getfrag(void *from, char *to, int offset, | |
231 | int len, int odd, struct sk_buff *skb), | |
232 | void *from, int length, int transhdrlen, | |
233 | struct ipcm_cookie *ipc, struct rtable **rtp, | |
1cd7884d | 234 | struct inet_cork *cork, unsigned int flags); |
1c32c5ad | 235 | |
05e22e83 | 236 | int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl); |
69b9e1e0 | 237 | |
77968b78 | 238 | static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4) |
1c32c5ad | 239 | { |
77968b78 | 240 | return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base); |
1c32c5ad | 241 | } |
1da177e4 | 242 | |
aa661581 FF |
243 | static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet) |
244 | { | |
245 | return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos); | |
246 | } | |
247 | ||
248 | static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk) | |
249 | { | |
250 | return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk); | |
251 | } | |
252 | ||
1da177e4 | 253 | /* datagram.c */ |
03645a11 | 254 | int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); |
5c3a0fd7 | 255 | int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len); |
1da177e4 | 256 | |
5c3a0fd7 | 257 | void ip4_datagram_release_cb(struct sock *sk); |
8141ed9f | 258 | |
1da177e4 | 259 | struct ip_reply_arg { |
82695b30 | 260 | struct kvec iov[1]; |
88ef4a5a | 261 | int flags; |
d6f5493c | 262 | __wsum csum; |
1da177e4 | 263 | int csumoffset; /* u16 offset of csum in iov[0].iov_base */ |
82695b30 | 264 | /* -1 if not needed */ |
f0e48dbf | 265 | int bound_dev_if; |
66b13d99 | 266 | u8 tos; |
e2d118a1 | 267 | kuid_t uid; |
82695b30 | 268 | }; |
1da177e4 | 269 | |
88ef4a5a KK |
270 | #define IP_REPLY_ARG_NOSRCCHECK 1 |
271 | ||
86b08d86 KK |
272 | static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg) |
273 | { | |
274 | return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0; | |
275 | } | |
276 | ||
bdbbb852 | 277 | void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb, |
24a2d43d ED |
278 | const struct ip_options *sopt, |
279 | __be32 daddr, __be32 saddr, | |
280 | const struct ip_reply_arg *arg, | |
d6fb396c | 281 | unsigned int len, u64 transmit_time); |
1da177e4 | 282 | |
4ce3c183 | 283 | #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field) |
13415e46 | 284 | #define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field) |
4ce3c183 | 285 | #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) |
13415e46 | 286 | #define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val) |
4ce3c183 | 287 | #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) |
13415e46 | 288 | #define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val) |
61a7e260 | 289 | #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field) |
13415e46 | 290 | #define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field) |
f7324acd | 291 | #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd) |
13415e46 | 292 | #define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd) |
1da177e4 | 293 | |
c4c6bc31 | 294 | u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offct); |
698365fa | 295 | unsigned long snmp_fold_field(void __percpu *mib, int offt); |
4ce3c183 | 296 | #if BITS_PER_LONG==32 |
c4c6bc31 R |
297 | u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct, |
298 | size_t syncp_offset); | |
698365fa | 299 | u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off); |
4ce3c183 | 300 | #else |
c4c6bc31 R |
301 | static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct, |
302 | size_t syncp_offset) | |
303 | { | |
304 | return snmp_get_cpu_field(mib, cpu, offct); | |
305 | ||
306 | } | |
307 | ||
698365fa | 308 | static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off) |
4ce3c183 ED |
309 | { |
310 | return snmp_fold_field(mib, offt); | |
311 | } | |
312 | #endif | |
33490170 | 313 | |
6348ef2d JH |
314 | #define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \ |
315 | { \ | |
316 | int i, c; \ | |
317 | for_each_possible_cpu(c) { \ | |
318 | for (i = 0; stats_list[i].name; i++) \ | |
319 | buff64[i] += snmp_get_cpu_field64( \ | |
320 | mib_statistic, \ | |
321 | c, stats_list[i].entry, \ | |
322 | offset); \ | |
323 | } \ | |
324 | } | |
325 | ||
326 | #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \ | |
327 | { \ | |
328 | int i, c; \ | |
329 | for_each_possible_cpu(c) { \ | |
330 | for (i = 0; stats_list[i].name; i++) \ | |
331 | buff[i] += snmp_get_cpu_field( \ | |
332 | mib_statistic, \ | |
333 | c, stats_list[i].entry); \ | |
334 | } \ | |
335 | } | |
336 | ||
0bbf87d8 | 337 | void inet_get_local_port_range(struct net *net, int *low, int *high); |
227b60f5 | 338 | |
fcd77db0 | 339 | #ifdef CONFIG_SYSCTL |
66e2f5f7 | 340 | static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) |
e3826f1e | 341 | { |
122ff243 | 342 | if (!net->ipv4.sysctl_local_reserved_ports) |
30429fba | 343 | return false; |
122ff243 | 344 | return test_bit(port, net->ipv4.sysctl_local_reserved_ports); |
e3826f1e | 345 | } |
20e61da7 WC |
346 | |
347 | static inline bool sysctl_dev_name_is_allowed(const char *name) | |
348 | { | |
349 | return strcmp(name, "default") != 0 && strcmp(name, "all") != 0; | |
350 | } | |
351 | ||
82f31ebf | 352 | static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port) |
4548b683 | 353 | { |
82f31ebf | 354 | return port < net->ipv4.sysctl_ip_prot_sock; |
4548b683 KJ |
355 | } |
356 | ||
122ff243 | 357 | #else |
66e2f5f7 | 358 | static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port) |
122ff243 | 359 | { |
30429fba | 360 | return false; |
122ff243 | 361 | } |
4548b683 | 362 | |
82f31ebf | 363 | static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port) |
4548b683 | 364 | { |
82f31ebf | 365 | return port < PROT_SOCK; |
4548b683 | 366 | } |
122ff243 | 367 | #endif |
e3826f1e | 368 | |
822c8685 DD |
369 | __be32 inet_current_timestamp(void); |
370 | ||
20380731 ACM |
371 | /* From inetpeer.c */ |
372 | extern int inet_peer_threshold; | |
373 | extern int inet_peer_minttl; | |
374 | extern int inet_peer_maxttl; | |
20380731 | 375 | |
5c3a0fd7 | 376 | void ipfrag_init(void); |
20380731 | 377 | |
5c3a0fd7 | 378 | void ip_static_sysctl_init(void); |
bd7b1533 | 379 | |
e110861f LC |
380 | #define IP4_REPLY_MARK(net, mark) \ |
381 | ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0) | |
382 | ||
d18cd551 DM |
383 | static inline bool ip_is_fragment(const struct iphdr *iph) |
384 | { | |
385 | return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0; | |
386 | } | |
387 | ||
1da177e4 | 388 | #ifdef CONFIG_INET |
14c85021 ACM |
389 | #include <net/dst.h> |
390 | ||
1da177e4 LT |
391 | /* The function in 2.2 was invalid, producing wrong result for |
392 | * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */ | |
393 | static inline | |
394 | int ip_decrease_ttl(struct iphdr *iph) | |
395 | { | |
5c78f275 AV |
396 | u32 check = (__force u32)iph->check; |
397 | check += (__force u32)htons(0x0100); | |
398 | iph->check = (__force __sum16)(check + (check>=0xFFFF)); | |
1da177e4 LT |
399 | return --iph->ttl; |
400 | } | |
401 | ||
d52e5a7e SD |
402 | static inline int ip_mtu_locked(const struct dst_entry *dst) |
403 | { | |
404 | const struct rtable *rt = (const struct rtable *)dst; | |
405 | ||
406 | return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU); | |
407 | } | |
408 | ||
1da177e4 | 409 | static inline |
4e3f5d72 | 410 | int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst) |
1da177e4 | 411 | { |
4e3f5d72 ED |
412 | u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc); |
413 | ||
414 | return pmtudisc == IP_PMTUDISC_DO || | |
415 | (pmtudisc == IP_PMTUDISC_WANT && | |
d52e5a7e | 416 | !ip_mtu_locked(dst)); |
1da177e4 LT |
417 | } |
418 | ||
f87c10a8 HFS |
419 | static inline bool ip_sk_accept_pmtu(const struct sock *sk) |
420 | { | |
1b346576 HFS |
421 | return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE && |
422 | inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT; | |
f87c10a8 HFS |
423 | } |
424 | ||
425 | static inline bool ip_sk_use_pmtu(const struct sock *sk) | |
426 | { | |
427 | return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE; | |
428 | } | |
429 | ||
60ff7467 | 430 | static inline bool ip_sk_ignore_df(const struct sock *sk) |
1b346576 HFS |
431 | { |
432 | return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO || | |
433 | inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT; | |
434 | } | |
435 | ||
f87c10a8 HFS |
436 | static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst, |
437 | bool forwarding) | |
438 | { | |
ac6627a2 | 439 | const struct rtable *rt = container_of(dst, struct rtable, dst); |
f87c10a8 | 440 | struct net *net = dev_net(dst->dev); |
02a1b175 | 441 | unsigned int mtu; |
f87c10a8 HFS |
442 | |
443 | if (net->ipv4.sysctl_ip_fwd_use_pmtu || | |
d52e5a7e | 444 | ip_mtu_locked(dst) || |
ac6627a2 VF |
445 | !forwarding) { |
446 | mtu = rt->rt_pmtu; | |
447 | if (mtu && time_before(jiffies, rt->dst.expires)) | |
448 | goto out; | |
449 | } | |
f87c10a8 | 450 | |
02a1b175 MÅ» |
451 | /* 'forwarding = true' case should always honour route mtu */ |
452 | mtu = dst_metric_raw(dst, RTAX_MTU); | |
ac6627a2 VF |
453 | if (mtu) |
454 | goto out; | |
455 | ||
456 | mtu = READ_ONCE(dst->dev->mtu); | |
457 | ||
458 | if (unlikely(ip_mtu_locked(dst))) { | |
459 | if (rt->rt_uses_gateway && mtu > 576) | |
460 | mtu = 576; | |
461 | } | |
462 | ||
463 | out: | |
464 | mtu = min_t(unsigned int, mtu, IP_MAX_MTU); | |
02a1b175 | 465 | |
fade5641 | 466 | return mtu - lwtunnel_headroom(dst->lwtstate, mtu); |
f87c10a8 HFS |
467 | } |
468 | ||
fedbb6b4 SL |
469 | static inline unsigned int ip_skb_dst_mtu(struct sock *sk, |
470 | const struct sk_buff *skb) | |
f87c10a8 | 471 | { |
fade5641 VF |
472 | unsigned int mtu; |
473 | ||
caf3f267 | 474 | if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) { |
f87c10a8 | 475 | bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED; |
caf3f267 | 476 | |
f87c10a8 | 477 | return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding); |
f87c10a8 | 478 | } |
caf3f267 | 479 | |
fade5641 VF |
480 | mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU); |
481 | return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu); | |
f87c10a8 HFS |
482 | } |
483 | ||
767a2217 | 484 | struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx, |
d7e774f3 DA |
485 | int fc_mx_len, |
486 | struct netlink_ext_ack *extack); | |
cc5f0eb2 DA |
487 | static inline void ip_fib_metrics_put(struct dst_metrics *fib_metrics) |
488 | { | |
489 | if (fib_metrics != &dst_default_metrics && | |
490 | refcount_dec_and_test(&fib_metrics->refcnt)) | |
491 | kfree(fib_metrics); | |
492 | } | |
a919525a | 493 | |
e1255ed4 DA |
494 | /* ipv4 and ipv6 both use refcounted metrics if it is not the default */ |
495 | static inline | |
496 | void ip_dst_init_metrics(struct dst_entry *dst, struct dst_metrics *fib_metrics) | |
497 | { | |
498 | dst_init_metrics(dst, fib_metrics->metrics, true); | |
499 | ||
500 | if (fib_metrics != &dst_default_metrics) { | |
501 | dst->_metrics |= DST_METRICS_REFCOUNTED; | |
502 | refcount_inc(&fib_metrics->refcnt); | |
503 | } | |
504 | } | |
505 | ||
1620a336 DA |
506 | static inline |
507 | void ip_dst_metrics_put(struct dst_entry *dst) | |
508 | { | |
509 | struct dst_metrics *p = (struct dst_metrics *)DST_METRICS_PTR(dst); | |
510 | ||
511 | if (p != &dst_default_metrics && refcount_dec_and_test(&p->refcnt)) | |
512 | kfree(p); | |
513 | } | |
514 | ||
04ca6973 | 515 | u32 ip_idents_reserve(u32 hash, int segs); |
b6a7719a | 516 | void __ip_select_ident(struct net *net, struct iphdr *iph, int segs); |
73f156a6 | 517 | |
b6a7719a HFS |
518 | static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb, |
519 | struct sock *sk, int segs) | |
1da177e4 | 520 | { |
703133de AA |
521 | struct iphdr *iph = ip_hdr(skb); |
522 | ||
2a0239eb ED |
523 | /* We had many attacks based on IPID, use the private |
524 | * generator as much as we can. | |
525 | */ | |
526 | if (sk && inet_sk(sk)->inet_daddr) { | |
527 | iph->id = htons(inet_sk(sk)->inet_id); | |
528 | inet_sk(sk)->inet_id += segs; | |
529 | return; | |
530 | } | |
60ff7467 | 531 | if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) { |
2a0239eb | 532 | iph->id = 0; |
73f156a6 | 533 | } else { |
2a0239eb | 534 | /* Unfortunately we need the big hammer to get a suitable IPID */ |
b6a7719a | 535 | __ip_select_ident(net, iph, segs); |
73f156a6 ED |
536 | } |
537 | } | |
538 | ||
b6a7719a HFS |
539 | static inline void ip_select_ident(struct net *net, struct sk_buff *skb, |
540 | struct sock *sk) | |
73f156a6 | 541 | { |
b6a7719a | 542 | ip_select_ident_segs(net, skb, sk, 1); |
1da177e4 LT |
543 | } |
544 | ||
ed70fcfc TH |
545 | static inline __wsum inet_compute_pseudo(struct sk_buff *skb, int proto) |
546 | { | |
547 | return csum_tcpudp_nofold(ip_hdr(skb)->saddr, ip_hdr(skb)->daddr, | |
548 | skb->len, proto, 0); | |
549 | } | |
550 | ||
c3f83241 TH |
551 | /* copy IPv4 saddr & daddr to flow_keys, possibly using 64bit load/store |
552 | * Equivalent to : flow->v4addrs.src = iph->saddr; | |
553 | * flow->v4addrs.dst = iph->daddr; | |
554 | */ | |
555 | static inline void iph_to_flow_copy_v4addrs(struct flow_keys *flow, | |
556 | const struct iphdr *iph) | |
557 | { | |
558 | BUILD_BUG_ON(offsetof(typeof(flow->addrs), v4addrs.dst) != | |
559 | offsetof(typeof(flow->addrs), v4addrs.src) + | |
560 | sizeof(flow->addrs.v4addrs.src)); | |
561 | memcpy(&flow->addrs.v4addrs, &iph->saddr, sizeof(flow->addrs.v4addrs)); | |
562 | flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS; | |
563 | } | |
564 | ||
1933a785 TH |
565 | static inline __wsum inet_gro_compute_pseudo(struct sk_buff *skb, int proto) |
566 | { | |
567 | const struct iphdr *iph = skb_gro_network_header(skb); | |
568 | ||
569 | return csum_tcpudp_nofold(iph->saddr, iph->daddr, | |
570 | skb_gro_len(skb), proto, 0); | |
571 | } | |
572 | ||
1da177e4 LT |
573 | /* |
574 | * Map a multicast IP onto multicast MAC for type ethernet. | |
575 | */ | |
576 | ||
714e85be | 577 | static inline void ip_eth_mc_map(__be32 naddr, char *buf) |
1da177e4 | 578 | { |
714e85be | 579 | __u32 addr=ntohl(naddr); |
1da177e4 LT |
580 | buf[0]=0x01; |
581 | buf[1]=0x00; | |
582 | buf[2]=0x5e; | |
583 | buf[5]=addr&0xFF; | |
584 | addr>>=8; | |
585 | buf[4]=addr&0xFF; | |
586 | addr>>=8; | |
587 | buf[3]=addr&0x7F; | |
588 | } | |
589 | ||
590 | /* | |
591 | * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand. | |
592 | * Leave P_Key as 0 to be filled in by driver. | |
593 | */ | |
594 | ||
a9e527e3 | 595 | static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) |
1da177e4 | 596 | { |
714e85be | 597 | __u32 addr; |
a9e527e3 RM |
598 | unsigned char scope = broadcast[5] & 0xF; |
599 | ||
1da177e4 LT |
600 | buf[0] = 0; /* Reserved */ |
601 | buf[1] = 0xff; /* Multicast QPN */ | |
602 | buf[2] = 0xff; | |
603 | buf[3] = 0xff; | |
714e85be | 604 | addr = ntohl(naddr); |
1da177e4 | 605 | buf[4] = 0xff; |
a9e527e3 | 606 | buf[5] = 0x10 | scope; /* scope from broadcast address */ |
1da177e4 LT |
607 | buf[6] = 0x40; /* IPv4 signature */ |
608 | buf[7] = 0x1b; | |
a9e527e3 RM |
609 | buf[8] = broadcast[8]; /* P_Key */ |
610 | buf[9] = broadcast[9]; | |
1da177e4 LT |
611 | buf[10] = 0; |
612 | buf[11] = 0; | |
613 | buf[12] = 0; | |
614 | buf[13] = 0; | |
615 | buf[14] = 0; | |
616 | buf[15] = 0; | |
617 | buf[19] = addr & 0xff; | |
618 | addr >>= 8; | |
619 | buf[18] = addr & 0xff; | |
620 | addr >>= 8; | |
621 | buf[17] = addr & 0xff; | |
622 | addr >>= 8; | |
623 | buf[16] = addr & 0x0f; | |
624 | } | |
625 | ||
93ca3bb5 TT |
626 | static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf) |
627 | { | |
628 | if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0) | |
629 | memcpy(buf, broadcast, 4); | |
630 | else | |
631 | memcpy(buf, &naddr, sizeof(naddr)); | |
632 | } | |
633 | ||
dfd56b8b | 634 | #if IS_ENABLED(CONFIG_IPV6) |
1da177e4 LT |
635 | #include <linux/ipv6.h> |
636 | #endif | |
637 | ||
638 | static __inline__ void inet_reset_saddr(struct sock *sk) | |
639 | { | |
c720c7e8 | 640 | inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0; |
dfd56b8b | 641 | #if IS_ENABLED(CONFIG_IPV6) |
1da177e4 LT |
642 | if (sk->sk_family == PF_INET6) { |
643 | struct ipv6_pinfo *np = inet6_sk(sk); | |
644 | ||
645 | memset(&np->saddr, 0, sizeof(np->saddr)); | |
efe4208f | 646 | memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr)); |
1da177e4 LT |
647 | } |
648 | #endif | |
649 | } | |
650 | ||
651 | #endif | |
652 | ||
72afa352 DA |
653 | static inline unsigned int ipv4_addr_hash(__be32 ip) |
654 | { | |
655 | return (__force unsigned int) ip; | |
656 | } | |
657 | ||
f0b1e64c MKL |
658 | static inline u32 ipv4_portaddr_hash(const struct net *net, |
659 | __be32 saddr, | |
660 | unsigned int port) | |
661 | { | |
662 | return jhash_1word((__force u32)saddr, net_hash_mix(net)) ^ port; | |
663 | } | |
664 | ||
5c3a0fd7 | 665 | bool ip_call_ra_chain(struct sk_buff *skb); |
1da177e4 LT |
666 | |
667 | /* | |
b798232f | 668 | * Functions provided by ip_fragment.c |
1da177e4 LT |
669 | */ |
670 | ||
fd2c3ef7 | 671 | enum ip_defrag_users { |
1da177e4 LT |
672 | IP_DEFRAG_LOCAL_DELIVER, |
673 | IP_DEFRAG_CALL_RA_CHAIN, | |
674 | IP_DEFRAG_CONNTRACK_IN, | |
4be929be | 675 | __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX, |
1da177e4 | 676 | IP_DEFRAG_CONNTRACK_OUT, |
4be929be | 677 | __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX, |
8fa9ff68 | 678 | IP_DEFRAG_CONNTRACK_BRIDGE_IN, |
4be929be | 679 | __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX, |
1da177e4 LT |
680 | IP_DEFRAG_VS_IN, |
681 | IP_DEFRAG_VS_OUT, | |
595fc71b DM |
682 | IP_DEFRAG_VS_FWD, |
683 | IP_DEFRAG_AF_PACKET, | |
bc416d97 | 684 | IP_DEFRAG_MACVLAN, |
1da177e4 LT |
685 | }; |
686 | ||
5cf42280 AZ |
687 | /* Return true if the value of 'user' is between 'lower_bond' |
688 | * and 'upper_bond' inclusively. | |
689 | */ | |
690 | static inline bool ip_defrag_user_in_between(u32 user, | |
691 | enum ip_defrag_users lower_bond, | |
692 | enum ip_defrag_users upper_bond) | |
693 | { | |
694 | return user >= lower_bond && user <= upper_bond; | |
695 | } | |
696 | ||
19bcf9f2 | 697 | int ip_defrag(struct net *net, struct sk_buff *skb, u32 user); |
bc416d97 | 698 | #ifdef CONFIG_INET |
19bcf9f2 | 699 | struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user); |
bc416d97 | 700 | #else |
19bcf9f2 | 701 | static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user) |
bc416d97 ED |
702 | { |
703 | return skb; | |
704 | } | |
705 | #endif | |
1da177e4 LT |
706 | |
707 | /* | |
708 | * Functions provided by ip_forward.c | |
709 | */ | |
82695b30 | 710 | |
5c3a0fd7 | 711 | int ip_forward(struct sk_buff *skb); |
82695b30 | 712 | |
1da177e4 LT |
713 | /* |
714 | * Functions provided by ip_options.c | |
715 | */ | |
82695b30 | 716 | |
5c3a0fd7 JP |
717 | void ip_options_build(struct sk_buff *skb, struct ip_options *opt, |
718 | __be32 daddr, struct rtable *rt, int is_frag); | |
24a2d43d | 719 | |
91ed1e66 PA |
720 | int __ip_options_echo(struct net *net, struct ip_options *dopt, |
721 | struct sk_buff *skb, const struct ip_options *sopt); | |
722 | static inline int ip_options_echo(struct net *net, struct ip_options *dopt, | |
723 | struct sk_buff *skb) | |
24a2d43d | 724 | { |
91ed1e66 | 725 | return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt); |
24a2d43d ED |
726 | } |
727 | ||
5c3a0fd7 | 728 | void ip_options_fragment(struct sk_buff *skb); |
3da1ed7a NS |
729 | int __ip_options_compile(struct net *net, struct ip_options *opt, |
730 | struct sk_buff *skb, __be32 *info); | |
5c3a0fd7 JP |
731 | int ip_options_compile(struct net *net, struct ip_options *opt, |
732 | struct sk_buff *skb); | |
733 | int ip_options_get(struct net *net, struct ip_options_rcu **optp, | |
de40a3e8 | 734 | sockptr_t data, int optlen); |
5c3a0fd7 JP |
735 | void ip_options_undo(struct ip_options *opt); |
736 | void ip_forward_options(struct sk_buff *skb); | |
8c83f2df | 737 | int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev); |
1da177e4 LT |
738 | |
739 | /* | |
740 | * Functions provided by ip_sockglue.c | |
741 | */ | |
742 | ||
fbf8866d | 743 | void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb); |
ad959036 PA |
744 | void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk, |
745 | struct sk_buff *skb, int tlen, int offset); | |
24025c46 | 746 | int ip_cmsg_send(struct sock *sk, struct msghdr *msg, |
c8e6ad08 | 747 | struct ipcm_cookie *ipc, bool allow_ipv6); |
a7b75c5a | 748 | int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval, |
5c3a0fd7 JP |
749 | unsigned int optlen); |
750 | int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, | |
751 | int __user *optlen); | |
5c3a0fd7 JP |
752 | int ip_ra_control(struct sock *sk, unsigned char on, |
753 | void (*destructor)(struct sock *)); | |
754 | ||
85fbaa75 | 755 | int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len); |
5c3a0fd7 JP |
756 | void ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port, |
757 | u32 info, u8 *payload); | |
758 | void ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport, | |
759 | u32 info); | |
1da177e4 | 760 | |
5961de9f TH |
761 | static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb) |
762 | { | |
ad959036 | 763 | ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0); |
5961de9f TH |
764 | } |
765 | ||
4cdf507d ED |
766 | bool icmp_global_allow(void); |
767 | extern int sysctl_icmp_msgs_per_sec; | |
768 | extern int sysctl_icmp_msgs_burst; | |
769 | ||
20380731 | 770 | #ifdef CONFIG_PROC_FS |
5c3a0fd7 | 771 | int ip_misc_proc_init(void); |
20380731 ACM |
772 | #endif |
773 | ||
5e1a99ea | 774 | int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family, |
404eb77e RP |
775 | struct netlink_ext_ack *extack); |
776 | ||
501a90c9 ED |
777 | static inline bool inetdev_valid_mtu(unsigned int mtu) |
778 | { | |
779 | return likely(mtu >= IPV4_MIN_MTU); | |
780 | } | |
781 | ||
c4e446bf | 782 | void ip_sock_set_freebind(struct sock *sk); |
2de569bd | 783 | int ip_sock_set_mtu_discover(struct sock *sk, int val); |
c1f9ec57 | 784 | void ip_sock_set_pktinfo(struct sock *sk); |
db45c0ef | 785 | void ip_sock_set_recverr(struct sock *sk); |
6ebf71ba CH |
786 | void ip_sock_set_tos(struct sock *sk, int val); |
787 | ||
1da177e4 | 788 | #endif /* _IP_H */ |