1 /* SPDX-License-Identifier: GPL-2.0-or-later */
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 *
11 * Authors: Ross Biro
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
17 */
18 #ifndef _IP_H
19 #define _IP_H
20
21 #include <linux/types.h>
22 #include <linux/ip.h>
23 #include <linux/in.h>
24 #include <linux/skbuff.h>
25 #include <linux/jhash.h>
26 #include <linux/sockptr.h>
27
28 #include <net/inet_sock.h>
29 #include <net/route.h>
30 #include <net/snmp.h>
31 #include <net/flow.h>
32 #include <net/flow_dissector.h>
33 #include <net/netns/hash.h>
34 #ifndef __GENKSYMS__
35 #include <net/lwtunnel.h>
36 #endif
37
38 #define IPV4_MAX_PMTU 65535U /* RFC 2675, Section 5.1 */
39 #define IPV4_MIN_MTU 68 /* RFC 791 */
40
41 extern unsigned int sysctl_fib_sync_mem;
42 extern unsigned int sysctl_fib_sync_mem_min;
43 extern unsigned int sysctl_fib_sync_mem_max;
44
45 struct sock;
46
47 struct inet_skb_parm {
48 int iif;
49 struct ip_options opt; /* Compiled IP options */
50 u16 flags;
51
52 #define IPSKB_FORWARDED BIT(0)
53 #define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
54 #define IPSKB_XFRM_TRANSFORMED BIT(2)
55 #define IPSKB_FRAG_COMPLETE BIT(3)
56 #define IPSKB_REROUTED BIT(4)
57 #define IPSKB_DOREDIRECT BIT(5)
58 #define IPSKB_FRAG_PMTU BIT(6)
59 #define IPSKB_L3SLAVE BIT(7)
60
61 u16 frag_max_size;
62 };
63
ipv4_l3mdev_skb(u16 flags)64 static inline bool ipv4_l3mdev_skb(u16 flags)
65 {
66 return !!(flags & IPSKB_L3SLAVE);
67 }
68
ip_hdrlen(const struct sk_buff * skb)69 static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
70 {
71 return ip_hdr(skb)->ihl * 4;
72 }
73
74 struct ipcm_cookie {
75 struct sockcm_cookie sockc;
76 __be32 addr;
77 int oif;
78 struct ip_options_rcu *opt;
79 __u8 ttl;
80 __s16 tos;
81 char priority;
82 __u16 gso_size;
83 };
84
ipcm_init(struct ipcm_cookie * ipcm)85 static inline void ipcm_init(struct ipcm_cookie *ipcm)
86 {
87 *ipcm = (struct ipcm_cookie) { .tos = -1 };
88 }
89
ipcm_init_sk(struct ipcm_cookie * ipcm,const struct inet_sock * inet)90 static inline void ipcm_init_sk(struct ipcm_cookie *ipcm,
91 const struct inet_sock *inet)
92 {
93 ipcm_init(ipcm);
94
95 ipcm->sockc.mark = inet->sk.sk_mark;
96 ipcm->sockc.tsflags = inet->sk.sk_tsflags;
97 ipcm->oif = inet->sk.sk_bound_dev_if;
98 ipcm->addr = inet->inet_saddr;
99 }
100
101 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
102 #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
103
104 /* return enslaved device index if relevant */
inet_sdif(struct sk_buff * skb)105 static inline int inet_sdif(struct sk_buff *skb)
106 {
107 #if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
108 if (skb && ipv4_l3mdev_skb(IPCB(skb)->flags))
109 return IPCB(skb)->iif;
110 #endif
111 return 0;
112 }
113
114 /* Special input handler for packets caught by router alert option.
115 They are selected only by protocol field, and then processed likely
116 local ones; but only if someone wants them! Otherwise, router
117 not running rsvpd will kill RSVP.
118
119 It is user level problem, what it will make with them.
120 I have no idea, how it will masquearde or NAT them (it is joke, joke :-)),
121 but receiver should be enough clever f.e. to forward mtrace requests,
122 sent to multicast group to reach destination designated router.
123 */
124
125 struct ip_ra_chain {
126 struct ip_ra_chain __rcu *next;
127 struct sock *sk;
128 union {
129 void (*destructor)(struct sock *);
130 struct sock *saved_sk;
131 };
132 struct rcu_head rcu;
133 };
134
135 /* IP flags. */
136 #define IP_CE 0x8000 /* Flag: "Congestion" */
137 #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
138 #define IP_MF 0x2000 /* Flag: "More Fragments" */
139 #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
140
141 #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
142
143 struct msghdr;
144 struct net_device;
145 struct packet_type;
146 struct rtable;
147 struct sockaddr;
148
149 int igmp_mc_init(void);
150
151 /*
152 * Functions provided by ip.c
153 */
154
155 int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
156 __be32 saddr, __be32 daddr,
157 struct ip_options_rcu *opt, u8 tos);
158 int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt,
159 struct net_device *orig_dev);
160 void ip_list_rcv(struct list_head *head, struct packet_type *pt,
161 struct net_device *orig_dev);
162 int ip_local_deliver(struct sk_buff *skb);
163 void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int proto);
164 int ip_mr_input(struct sk_buff *skb);
165 int ip_output(struct net *net, struct sock *sk, struct sk_buff *skb);
166 int ip_mc_output(struct net *net, struct sock *sk, struct sk_buff *skb);
167 int ip_do_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
168 int (*output)(struct net *, struct sock *, struct sk_buff *));
169
170 struct ip_fraglist_iter {
171 struct sk_buff *frag;
172 struct iphdr *iph;
173 int offset;
174 unsigned int hlen;
175 };
176
177 void ip_fraglist_init(struct sk_buff *skb, struct iphdr *iph,
178 unsigned int hlen, struct ip_fraglist_iter *iter);
179 void ip_fraglist_prepare(struct sk_buff *skb, struct ip_fraglist_iter *iter);
180
ip_fraglist_next(struct ip_fraglist_iter * iter)181 static inline struct sk_buff *ip_fraglist_next(struct ip_fraglist_iter *iter)
182 {
183 struct sk_buff *skb = iter->frag;
184
185 iter->frag = skb->next;
186 skb_mark_not_on_list(skb);
187
188 return skb;
189 }
190
191 struct ip_frag_state {
192 bool DF;
193 unsigned int hlen;
194 unsigned int ll_rs;
195 unsigned int mtu;
196 unsigned int left;
197 int offset;
198 int ptr;
199 __be16 not_last_frag;
200 };
201
202 void ip_frag_init(struct sk_buff *skb, unsigned int hlen, unsigned int ll_rs,
203 unsigned int mtu, bool DF, struct ip_frag_state *state);
204 struct sk_buff *ip_frag_next(struct sk_buff *skb,
205 struct ip_frag_state *state);
206
207 void ip_send_check(struct iphdr *ip);
208 int __ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
209 int ip_local_out(struct net *net, struct sock *sk, struct sk_buff *skb);
210
211 int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl,
212 __u8 tos);
213 void ip_init(void);
214 int ip_append_data(struct sock *sk, struct flowi4 *fl4,
215 int getfrag(void *from, char *to, int offset, int len,
216 int odd, struct sk_buff *skb),
217 void *from, int len, int protolen,
218 struct ipcm_cookie *ipc,
219 struct rtable **rt,
220 unsigned int flags);
221 int ip_generic_getfrag(void *from, char *to, int offset, int len, int odd,
222 struct sk_buff *skb);
223 ssize_t ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
224 int offset, size_t size, int flags);
225 struct sk_buff *__ip_make_skb(struct sock *sk, struct flowi4 *fl4,
226 struct sk_buff_head *queue,
227 struct inet_cork *cork);
228 int ip_send_skb(struct net *net, struct sk_buff *skb);
229 int ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
230 void ip_flush_pending_frames(struct sock *sk);
231 struct sk_buff *ip_make_skb(struct sock *sk, struct flowi4 *fl4,
232 int getfrag(void *from, char *to, int offset,
233 int len, int odd, struct sk_buff *skb),
234 void *from, int length, int transhdrlen,
235 struct ipcm_cookie *ipc, struct rtable **rtp,
236 struct inet_cork *cork, unsigned int flags);
237
238 int ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl);
239
ip_finish_skb(struct sock * sk,struct flowi4 * fl4)240 static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
241 {
242 return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
243 }
244
get_rttos(struct ipcm_cookie * ipc,struct inet_sock * inet)245 static inline __u8 get_rttos(struct ipcm_cookie* ipc, struct inet_sock *inet)
246 {
247 return (ipc->tos != -1) ? RT_TOS(ipc->tos) : RT_TOS(inet->tos);
248 }
249
get_rtconn_flags(struct ipcm_cookie * ipc,struct sock * sk)250 static inline __u8 get_rtconn_flags(struct ipcm_cookie* ipc, struct sock* sk)
251 {
252 return (ipc->tos != -1) ? RT_CONN_FLAGS_TOS(sk, ipc->tos) : RT_CONN_FLAGS(sk);
253 }
254
255 /* datagram.c */
256 int __ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
257 int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
258
259 void ip4_datagram_release_cb(struct sock *sk);
260
261 struct ip_reply_arg {
262 struct kvec iov[1];
263 int flags;
264 __wsum csum;
265 int csumoffset; /* u16 offset of csum in iov[0].iov_base */
266 /* -1 if not needed */
267 int bound_dev_if;
268 u8 tos;
269 kuid_t uid;
270 };
271
272 #define IP_REPLY_ARG_NOSRCCHECK 1
273
ip_reply_arg_flowi_flags(const struct ip_reply_arg * arg)274 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
275 {
276 return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
277 }
278
279 void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
280 const struct ip_options *sopt,
281 __be32 daddr, __be32 saddr,
282 const struct ip_reply_arg *arg,
283 unsigned int len, u64 transmit_time);
284
285 #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
286 #define __IP_INC_STATS(net, field) __SNMP_INC_STATS64((net)->mib.ip_statistics, field)
287 #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
288 #define __IP_ADD_STATS(net, field, val) __SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
289 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
290 #define __IP_UPD_PO_STATS(net, field, val) __SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
291 #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
292 #define __NET_INC_STATS(net, field) __SNMP_INC_STATS((net)->mib.net_statistics, field)
293 #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
294 #define __NET_ADD_STATS(net, field, adnd) __SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
295
296 u64 snmp_get_cpu_field(void __percpu *mib, int cpu, int offct);
297 unsigned long snmp_fold_field(void __percpu *mib, int offt);
298 #if BITS_PER_LONG==32
299 u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
300 size_t syncp_offset);
301 u64 snmp_fold_field64(void __percpu *mib, int offt, size_t sync_off);
302 #else
snmp_get_cpu_field64(void __percpu * mib,int cpu,int offct,size_t syncp_offset)303 static inline u64 snmp_get_cpu_field64(void __percpu *mib, int cpu, int offct,
304 size_t syncp_offset)
305 {
306 return snmp_get_cpu_field(mib, cpu, offct);
307
308 }
309
snmp_fold_field64(void __percpu * mib,int offt,size_t syncp_off)310 static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_off)
311 {
312 return snmp_fold_field(mib, offt);
313 }
314 #endif
315
316 #define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \
317 { \
318 int i, c; \
319 for_each_possible_cpu(c) { \
320 for (i = 0; stats_list[i].name; i++) \
321 buff64[i] += snmp_get_cpu_field64( \
322 mib_statistic, \
323 c, stats_list[i].entry, \
324 offset); \
325 } \
326 }
327
328 #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
329 { \
330 int i, c; \
331 for_each_possible_cpu(c) { \
332 for (i = 0; stats_list[i].name; i++) \
333 buff[i] += snmp_get_cpu_field( \
334 mib_statistic, \
335 c, stats_list[i].entry); \
336 } \
337 }
338
339 void inet_get_local_port_range(struct net *net, int *low, int *high);
340
341 #ifdef CONFIG_SYSCTL
inet_is_local_reserved_port(struct net * net,unsigned short port)342 static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
343 {
344 if (!net->ipv4.sysctl_local_reserved_ports)
345 return false;
346 return test_bit(port, net->ipv4.sysctl_local_reserved_ports);
347 }
348
inet_is_local_unbindable_port(struct net * net,unsigned short port)349 static inline bool inet_is_local_unbindable_port(struct net *net, unsigned short port)
350 {
351 if (!net->ipv4.sysctl_local_unbindable_ports)
352 return false;
353 return test_bit(port, net->ipv4.sysctl_local_unbindable_ports);
354 }
355
sysctl_dev_name_is_allowed(const char * name)356 static inline bool sysctl_dev_name_is_allowed(const char *name)
357 {
358 return strcmp(name, "default") != 0 && strcmp(name, "all") != 0;
359 }
360
inet_port_requires_bind_service(struct net * net,unsigned short port)361 static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
362 {
363 return port < READ_ONCE(net->ipv4.sysctl_ip_prot_sock);
364 }
365
366 #else
inet_is_local_reserved_port(struct net * net,unsigned short port)367 static inline bool inet_is_local_reserved_port(struct net *net, unsigned short port)
368 {
369 return false;
370 }
371
inet_is_local_unbindable_port(struct net * net,unsigned short port)372 static inline bool inet_is_local_unbindable_port(struct net *net, unsigned short port)
373 {
374 return false;
375 }
376
inet_port_requires_bind_service(struct net * net,unsigned short port)377 static inline bool inet_port_requires_bind_service(struct net *net, unsigned short port)
378 {
379 return port < PROT_SOCK;
380 }
381 #endif
382
383 __be32 inet_current_timestamp(void);
384
385 /* From inetpeer.c */
386 extern int inet_peer_threshold;
387 extern int inet_peer_minttl;
388 extern int inet_peer_maxttl;
389
390 void ipfrag_init(void);
391
392 void ip_static_sysctl_init(void);
393
394 #define IP4_REPLY_MARK(net, mark) \
395 (READ_ONCE((net)->ipv4.sysctl_fwmark_reflect) ? (mark) : 0)
396
ip_is_fragment(const struct iphdr * iph)397 static inline bool ip_is_fragment(const struct iphdr *iph)
398 {
399 return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
400 }
401
402 #ifdef CONFIG_INET
403 #include <net/dst.h>
404
405 /* The function in 2.2 was invalid, producing wrong result for
406 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
407 static inline
ip_decrease_ttl(struct iphdr * iph)408 int ip_decrease_ttl(struct iphdr *iph)
409 {
410 u32 check = (__force u32)iph->check;
411 check += (__force u32)htons(0x0100);
412 iph->check = (__force __sum16)(check + (check>=0xFFFF));
413 return --iph->ttl;
414 }
415
ip_mtu_locked(const struct dst_entry * dst)416 static inline int ip_mtu_locked(const struct dst_entry *dst)
417 {
418 const struct rtable *rt = (const struct rtable *)dst;
419
420 return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU);
421 }
422
423 static inline
ip_dont_fragment(const struct sock * sk,const struct dst_entry * dst)424 int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst)
425 {
426 u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc);
427
428 return pmtudisc == IP_PMTUDISC_DO ||
429 (pmtudisc == IP_PMTUDISC_WANT &&
430 !ip_mtu_locked(dst));
431 }
432
ip_sk_accept_pmtu(const struct sock * sk)433 static inline bool ip_sk_accept_pmtu(const struct sock *sk)
434 {
435 return inet_sk(sk)->pmtudisc != IP_PMTUDISC_INTERFACE &&
436 inet_sk(sk)->pmtudisc != IP_PMTUDISC_OMIT;
437 }
438
ip_sk_use_pmtu(const struct sock * sk)439 static inline bool ip_sk_use_pmtu(const struct sock *sk)
440 {
441 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_PROBE;
442 }
443
ip_sk_ignore_df(const struct sock * sk)444 static inline bool ip_sk_ignore_df(const struct sock *sk)
445 {
446 return inet_sk(sk)->pmtudisc < IP_PMTUDISC_DO ||
447 inet_sk(sk)->pmtudisc == IP_PMTUDISC_OMIT;
448 }
449
ip_dst_mtu_maybe_forward(const struct dst_entry * dst,bool forwarding)450 static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
451 bool forwarding)
452 {
453 struct net *net = dev_net(dst->dev);
454 unsigned int mtu;
455
456 if (READ_ONCE(net->ipv4.sysctl_ip_fwd_use_pmtu) ||
457 ip_mtu_locked(dst) ||
458 !forwarding)
459 return dst_mtu(dst);
460
461 /* 'forwarding = true' case should always honour route mtu */
462 mtu = dst_metric_raw(dst, RTAX_MTU);
463 if (!mtu)
464 mtu = min(READ_ONCE(dst->dev->mtu), IP_MAX_MTU);
465
466 return mtu - lwtunnel_headroom(dst->lwtstate, mtu);
467 }
468
ip_skb_dst_mtu(struct sock * sk,const struct sk_buff * skb)469 static inline unsigned int ip_skb_dst_mtu(struct sock *sk,
470 const struct sk_buff *skb)
471 {
472 unsigned int mtu;
473
474 if (!sk || !sk_fullsock(sk) || ip_sk_use_pmtu(sk)) {
475 bool forwarding = IPCB(skb)->flags & IPSKB_FORWARDED;
476
477 return ip_dst_mtu_maybe_forward(skb_dst(skb), forwarding);
478 }
479
480 mtu = min(READ_ONCE(skb_dst(skb)->dev->mtu), IP_MAX_MTU);
481 return mtu - lwtunnel_headroom(skb_dst(skb)->lwtstate, mtu);
482 }
483
484 struct dst_metrics *ip_fib_metrics_init(struct net *net, struct nlattr *fc_mx,
485 int fc_mx_len,
486 struct netlink_ext_ack *extack);
ip_fib_metrics_put(struct dst_metrics * fib_metrics)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 }
493
494 /* ipv4 and ipv6 both use refcounted metrics if it is not the default */
495 static inline
ip_dst_init_metrics(struct dst_entry * dst,struct dst_metrics * fib_metrics)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
506 static inline
ip_dst_metrics_put(struct dst_entry * dst)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
515 u32 ip_idents_reserve(u32 hash, int segs);
516 void __ip_select_ident(struct net *net, struct iphdr *iph, int segs);
517
ip_select_ident_segs(struct net * net,struct sk_buff * skb,struct sock * sk,int segs)518 static inline void ip_select_ident_segs(struct net *net, struct sk_buff *skb,
519 struct sock *sk, int segs)
520 {
521 struct iphdr *iph = ip_hdr(skb);
522
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 }
531 if ((iph->frag_off & htons(IP_DF)) && !skb->ignore_df) {
532 iph->id = 0;
533 } else {
534 /* Unfortunately we need the big hammer to get a suitable IPID */
535 __ip_select_ident(net, iph, segs);
536 }
537 }
538
ip_select_ident(struct net * net,struct sk_buff * skb,struct sock * sk)539 static inline void ip_select_ident(struct net *net, struct sk_buff *skb,
540 struct sock *sk)
541 {
542 ip_select_ident_segs(net, skb, sk, 1);
543 }
544
inet_compute_pseudo(struct sk_buff * skb,int proto)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
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 */
iph_to_flow_copy_v4addrs(struct flow_keys * flow,const struct iphdr * iph)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->addrs, sizeof(flow->addrs.v4addrs));
562 flow->control.addr_type = FLOW_DISSECTOR_KEY_IPV4_ADDRS;
563 }
564
inet_gro_compute_pseudo(struct sk_buff * skb,int proto)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
573 /*
574 * Map a multicast IP onto multicast MAC for type ethernet.
575 */
576
ip_eth_mc_map(__be32 naddr,char * buf)577 static inline void ip_eth_mc_map(__be32 naddr, char *buf)
578 {
579 __u32 addr=ntohl(naddr);
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
ip_ib_mc_map(__be32 naddr,const unsigned char * broadcast,char * buf)595 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
596 {
597 __u32 addr;
598 unsigned char scope = broadcast[5] & 0xF;
599
600 buf[0] = 0; /* Reserved */
601 buf[1] = 0xff; /* Multicast QPN */
602 buf[2] = 0xff;
603 buf[3] = 0xff;
604 addr = ntohl(naddr);
605 buf[4] = 0xff;
606 buf[5] = 0x10 | scope; /* scope from broadcast address */
607 buf[6] = 0x40; /* IPv4 signature */
608 buf[7] = 0x1b;
609 buf[8] = broadcast[8]; /* P_Key */
610 buf[9] = broadcast[9];
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
ip_ipgre_mc_map(__be32 naddr,const unsigned char * broadcast,char * buf)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
634 #if IS_ENABLED(CONFIG_IPV6)
635 #include <linux/ipv6.h>
636 #endif
637
inet_reset_saddr(struct sock * sk)638 static __inline__ void inet_reset_saddr(struct sock *sk)
639 {
640 inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
641 #if IS_ENABLED(CONFIG_IPV6)
642 if (sk->sk_family == PF_INET6) {
643 struct ipv6_pinfo *np = inet6_sk(sk);
644
645 memset(&np->saddr, 0, sizeof(np->saddr));
646 memset(&sk->sk_v6_rcv_saddr, 0, sizeof(sk->sk_v6_rcv_saddr));
647 }
648 #endif
649 }
650
651 #endif
652
ipv4_addr_hash(__be32 ip)653 static inline unsigned int ipv4_addr_hash(__be32 ip)
654 {
655 return (__force unsigned int) ip;
656 }
657
ipv4_portaddr_hash(const struct net * net,__be32 saddr,unsigned int port)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
665 bool ip_call_ra_chain(struct sk_buff *skb);
666
667 /*
668 * Functions provided by ip_fragment.c
669 */
670
671 enum ip_defrag_users {
672 IP_DEFRAG_LOCAL_DELIVER,
673 IP_DEFRAG_CALL_RA_CHAIN,
674 IP_DEFRAG_CONNTRACK_IN,
675 __IP_DEFRAG_CONNTRACK_IN_END = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
676 IP_DEFRAG_CONNTRACK_OUT,
677 __IP_DEFRAG_CONNTRACK_OUT_END = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
678 IP_DEFRAG_CONNTRACK_BRIDGE_IN,
679 __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
680 IP_DEFRAG_VS_IN,
681 IP_DEFRAG_VS_OUT,
682 IP_DEFRAG_VS_FWD,
683 IP_DEFRAG_AF_PACKET,
684 IP_DEFRAG_MACVLAN,
685 };
686
687 /* Return true if the value of 'user' is between 'lower_bond'
688 * and 'upper_bond' inclusively.
689 */
ip_defrag_user_in_between(u32 user,enum ip_defrag_users lower_bond,enum ip_defrag_users upper_bond)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
697 int ip_defrag(struct net *net, struct sk_buff *skb, u32 user);
698 #ifdef CONFIG_INET
699 struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user);
700 #else
ip_check_defrag(struct net * net,struct sk_buff * skb,u32 user)701 static inline struct sk_buff *ip_check_defrag(struct net *net, struct sk_buff *skb, u32 user)
702 {
703 return skb;
704 }
705 #endif
706
707 /*
708 * Functions provided by ip_forward.c
709 */
710
711 int ip_forward(struct sk_buff *skb);
712
713 /*
714 * Functions provided by ip_options.c
715 */
716
717 void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
718 __be32 daddr, struct rtable *rt, int is_frag);
719
720 int __ip_options_echo(struct net *net, struct ip_options *dopt,
721 struct sk_buff *skb, const struct ip_options *sopt);
ip_options_echo(struct net * net,struct ip_options * dopt,struct sk_buff * skb)722 static inline int ip_options_echo(struct net *net, struct ip_options *dopt,
723 struct sk_buff *skb)
724 {
725 return __ip_options_echo(net, dopt, skb, &IPCB(skb)->opt);
726 }
727
728 void ip_options_fragment(struct sk_buff *skb);
729 int __ip_options_compile(struct net *net, struct ip_options *opt,
730 struct sk_buff *skb, __be32 *info);
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,
734 sockptr_t data, int optlen);
735 void ip_options_undo(struct ip_options *opt);
736 void ip_forward_options(struct sk_buff *skb);
737 int ip_options_rcv_srr(struct sk_buff *skb, struct net_device *dev);
738
739 /*
740 * Functions provided by ip_sockglue.c
741 */
742
743 void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb);
744 void ip_cmsg_recv_offset(struct msghdr *msg, struct sock *sk,
745 struct sk_buff *skb, int tlen, int offset);
746 int ip_cmsg_send(struct sock *sk, struct msghdr *msg,
747 struct ipcm_cookie *ipc, bool allow_ipv6);
748 int ip_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
749 unsigned int optlen);
750 int ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
751 int __user *optlen);
752 int ip_ra_control(struct sock *sk, unsigned char on,
753 void (*destructor)(struct sock *));
754
755 int ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
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);
760
ip_cmsg_recv(struct msghdr * msg,struct sk_buff * skb)761 static inline void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
762 {
763 ip_cmsg_recv_offset(msg, skb->sk, skb, 0, 0);
764 }
765
766 bool icmp_global_allow(void);
767 extern int sysctl_icmp_msgs_per_sec;
768 extern int sysctl_icmp_msgs_burst;
769
770 #ifdef CONFIG_PROC_FS
771 int ip_misc_proc_init(void);
772 #endif
773
774 int rtm_getroute_parse_ip_proto(struct nlattr *attr, u8 *ip_proto, u8 family,
775 struct netlink_ext_ack *extack);
776
inetdev_valid_mtu(unsigned int mtu)777 static inline bool inetdev_valid_mtu(unsigned int mtu)
778 {
779 return likely(mtu >= IPV4_MIN_MTU);
780 }
781
782 void ip_sock_set_freebind(struct sock *sk);
783 int ip_sock_set_mtu_discover(struct sock *sk, int val);
784 void ip_sock_set_pktinfo(struct sock *sk);
785 void ip_sock_set_recverr(struct sock *sk);
786 void ip_sock_set_tos(struct sock *sk, int val);
787
788 #endif /* _IP_H */
789