• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * INET		An implementation of the TCP/IP protocol suite for the LINUX
3  *		operating system.  INET is implemented using the  BSD Socket
4  *		interface as the means of communication with the user level.
5  *
6  *		Definitions for the IP module.
7  *
8  * Version:	@(#)ip.h	1.0.2	05/07/93
9  *
10  * Authors:	Ross Biro
11  *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
12  *		Alan Cox, <gw4pts@gw4pts.ampr.org>
13  *
14  * Changes:
15  *		Mike McLagan    :       Routing by source
16  *
17  *		This program is free software; you can redistribute it and/or
18  *		modify it under the terms of the GNU General Public License
19  *		as published by the Free Software Foundation; either version
20  *		2 of the License, or (at your option) any later version.
21  */
22 #ifndef _IP_H
23 #define _IP_H
24 
25 #include <linux/types.h>
26 #include <linux/ip.h>
27 #include <linux/in.h>
28 #include <linux/skbuff.h>
29 
30 #include <net/inet_sock.h>
31 #include <net/snmp.h>
32 #include <net/flow.h>
33 
34 struct sock;
35 
36 struct inet_skb_parm {
37 	struct ip_options	opt;		/* Compiled IP options		*/
38 	unsigned char		flags;
39 
40 #define IPSKB_FORWARDED		BIT(0)
41 #define IPSKB_XFRM_TUNNEL_SIZE	BIT(1)
42 #define IPSKB_XFRM_TRANSFORMED	BIT(2)
43 #define IPSKB_FRAG_COMPLETE	BIT(3)
44 #define IPSKB_REROUTED		BIT(4)
45 #define IPSKB_DOREDIRECT	BIT(5)
46 
47 	u16			frag_max_size;
48 };
49 
ip_hdrlen(const struct sk_buff * skb)50 static inline unsigned int ip_hdrlen(const struct sk_buff *skb)
51 {
52 	return ip_hdr(skb)->ihl * 4;
53 }
54 
55 struct ipcm_cookie {
56 	__be32			addr;
57 	int			oif;
58 	struct ip_options_rcu	*opt;
59 	__u8			tx_flags;
60 };
61 
62 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
63 
64 struct ip_ra_chain {
65 	struct ip_ra_chain __rcu *next;
66 	struct sock		*sk;
67 	union {
68 		void			(*destructor)(struct sock *);
69 		struct sock		*saved_sk;
70 	};
71 	struct rcu_head		rcu;
72 };
73 
74 extern struct ip_ra_chain __rcu *ip_ra_chain;
75 
76 /* IP flags. */
77 #define IP_CE		0x8000		/* Flag: "Congestion"		*/
78 #define IP_DF		0x4000		/* Flag: "Don't Fragment"	*/
79 #define IP_MF		0x2000		/* Flag: "More Fragments"	*/
80 #define IP_OFFSET	0x1FFF		/* "Fragment Offset" part	*/
81 
82 #define IP_FRAG_TIME	(30 * HZ)		/* fragment lifetime	*/
83 
84 struct msghdr;
85 struct net_device;
86 struct packet_type;
87 struct rtable;
88 struct sockaddr;
89 
90 extern int		igmp_mc_proc_init(void);
91 
92 /*
93  *	Functions provided by ip.c
94  */
95 
96 extern int		ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
97 					      __be32 saddr, __be32 daddr,
98 					      struct ip_options_rcu *opt);
99 extern int		ip_rcv(struct sk_buff *skb, struct net_device *dev,
100 			       struct packet_type *pt, struct net_device *orig_dev);
101 extern int		ip_local_deliver(struct sk_buff *skb);
102 extern int		ip_mr_input(struct sk_buff *skb);
103 extern int		ip_output(struct sk_buff *skb);
104 extern int		ip_mc_output(struct sk_buff *skb);
105 extern int		ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *));
106 extern int		ip_do_nat(struct sk_buff *skb);
107 extern void		ip_send_check(struct iphdr *ip);
108 extern int		__ip_local_out(struct sk_buff *skb);
109 extern int		ip_local_out(struct sk_buff *skb);
110 extern int		ip_queue_xmit(struct sk_buff *skb, struct flowi *fl);
111 extern void		ip_init(void);
112 extern int		ip_append_data(struct sock *sk, struct flowi4 *fl4,
113 				       int getfrag(void *from, char *to, int offset, int len,
114 						   int odd, struct sk_buff *skb),
115 				void *from, int len, int protolen,
116 				struct ipcm_cookie *ipc,
117 				struct rtable **rt,
118 				unsigned int flags);
119 extern int		ip_generic_getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb);
120 extern ssize_t		ip_append_page(struct sock *sk, struct flowi4 *fl4, struct page *page,
121 				int offset, size_t size, int flags);
122 extern struct sk_buff  *__ip_make_skb(struct sock *sk,
123 				      struct flowi4 *fl4,
124 				      struct sk_buff_head *queue,
125 				      struct inet_cork *cork);
126 extern int		ip_send_skb(struct net *net, struct sk_buff *skb);
127 extern int		ip_push_pending_frames(struct sock *sk, struct flowi4 *fl4);
128 extern void		ip_flush_pending_frames(struct sock *sk);
129 extern struct sk_buff  *ip_make_skb(struct sock *sk,
130 				    struct flowi4 *fl4,
131 				    int getfrag(void *from, char *to, int offset, int len,
132 						int odd, struct sk_buff *skb),
133 				    void *from, int length, int transhdrlen,
134 				    struct ipcm_cookie *ipc,
135 				    struct rtable **rtp,
136 				    unsigned int flags);
137 
ip_finish_skb(struct sock * sk,struct flowi4 * fl4)138 static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4)
139 {
140 	return __ip_make_skb(sk, fl4, &sk->sk_write_queue, &inet_sk(sk)->cork.base);
141 }
142 
143 /* datagram.c */
144 extern int		ip4_datagram_connect(struct sock *sk,
145 					     struct sockaddr *uaddr, int addr_len);
146 
147 extern void ip4_datagram_release_cb(struct sock *sk);
148 
149 struct ip_reply_arg {
150 	struct kvec iov[1];
151 	int	    flags;
152 	__wsum 	    csum;
153 	int	    csumoffset; /* u16 offset of csum in iov[0].iov_base */
154 				/* -1 if not needed */
155 	int	    bound_dev_if;
156 	u8  	    tos;
157 	kuid_t	    uid;
158 };
159 
160 #define IP_REPLY_ARG_NOSRCCHECK 1
161 
ip_reply_arg_flowi_flags(const struct ip_reply_arg * arg)162 static inline __u8 ip_reply_arg_flowi_flags(const struct ip_reply_arg *arg)
163 {
164 	return (arg->flags & IP_REPLY_ARG_NOSRCCHECK) ? FLOWI_FLAG_ANYSRC : 0;
165 }
166 
167 void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb, __be32 daddr,
168 			   __be32 saddr, const struct ip_reply_arg *arg,
169 			   unsigned int len);
170 
171 struct ipv4_config {
172 	int	log_martians;
173 	int	no_pmtu_disc;
174 };
175 
176 extern struct ipv4_config ipv4_config;
177 #define IP_INC_STATS(net, field)	SNMP_INC_STATS64((net)->mib.ip_statistics, field)
178 #define IP_INC_STATS_BH(net, field)	SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
179 #define IP_ADD_STATS(net, field, val)	SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
180 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
181 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
182 #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
183 #define NET_INC_STATS(net, field)	SNMP_INC_STATS((net)->mib.net_statistics, field)
184 #define NET_INC_STATS_BH(net, field)	SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
185 #define NET_INC_STATS_USER(net, field) 	SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
186 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
187 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
188 
189 extern unsigned long snmp_fold_field(void __percpu *mib[], int offt);
190 #if BITS_PER_LONG==32
191 extern u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t sync_off);
192 #else
snmp_fold_field64(void __percpu * mib[],int offt,size_t syncp_off)193 static inline u64 snmp_fold_field64(void __percpu *mib[], int offt, size_t syncp_off)
194 {
195 	return snmp_fold_field(mib, offt);
196 }
197 #endif
198 extern int snmp_mib_init(void __percpu *ptr[2], size_t mibsize, size_t align);
199 extern void snmp_mib_free(void __percpu *ptr[2]);
200 
201 extern struct local_ports {
202 	seqlock_t	lock;
203 	int		range[2];
204 } sysctl_local_ports;
205 extern void inet_get_local_port_range(int *low, int *high);
206 
207 extern unsigned long *sysctl_local_reserved_ports;
inet_is_reserved_local_port(int port)208 static inline int inet_is_reserved_local_port(int port)
209 {
210 	return test_bit(port, sysctl_local_reserved_ports);
211 }
212 
213 extern int sysctl_ip_nonlocal_bind;
214 
215 /* From inetpeer.c */
216 extern int inet_peer_threshold;
217 extern int inet_peer_minttl;
218 extern int inet_peer_maxttl;
219 
220 /* From ip_input.c */
221 extern int sysctl_ip_early_demux;
222 
223 /* From ip_output.c */
224 extern int sysctl_ip_dynaddr;
225 
226 extern void ipfrag_init(void);
227 
228 extern void ip_static_sysctl_init(void);
229 
230 #define IP4_REPLY_MARK(net, mark) \
231 	((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
232 
ip_is_fragment(const struct iphdr * iph)233 static inline bool ip_is_fragment(const struct iphdr *iph)
234 {
235 	return (iph->frag_off & htons(IP_MF | IP_OFFSET)) != 0;
236 }
237 
238 #ifdef CONFIG_INET
239 #include <net/dst.h>
240 
241 /* The function in 2.2 was invalid, producing wrong result for
242  * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
243 static inline
ip_decrease_ttl(struct iphdr * iph)244 int ip_decrease_ttl(struct iphdr *iph)
245 {
246 	u32 check = (__force u32)iph->check;
247 	check += (__force u32)htons(0x0100);
248 	iph->check = (__force __sum16)(check + (check>=0xFFFF));
249 	return --iph->ttl;
250 }
251 
252 static inline
ip_dont_fragment(struct sock * sk,struct dst_entry * dst)253 int ip_dont_fragment(struct sock *sk, struct dst_entry *dst)
254 {
255 	return  inet_sk(sk)->pmtudisc == IP_PMTUDISC_DO ||
256 		(inet_sk(sk)->pmtudisc == IP_PMTUDISC_WANT &&
257 		 !(dst_metric_locked(dst, RTAX_MTU)));
258 }
259 
260 extern void __ip_select_ident(struct iphdr *iph, struct dst_entry *dst, int more);
261 
ip_select_ident(struct iphdr * iph,struct dst_entry * dst,struct sock * sk)262 static inline void ip_select_ident(struct iphdr *iph, struct dst_entry *dst, struct sock *sk)
263 {
264 	if (iph->frag_off & htons(IP_DF)) {
265 		/* This is only to work around buggy Windows95/2000
266 		 * VJ compression implementations.  If the ID field
267 		 * does not change, they drop every other packet in
268 		 * a TCP stream using header compression.
269 		 */
270 		iph->id = (sk && inet_sk(sk)->inet_daddr) ?
271 					htons(inet_sk(sk)->inet_id++) : 0;
272 	} else
273 		__ip_select_ident(iph, dst, 0);
274 }
275 
ip_select_ident_more(struct iphdr * iph,struct dst_entry * dst,struct sock * sk,int more)276 static inline void ip_select_ident_more(struct iphdr *iph, struct dst_entry *dst, struct sock *sk, int more)
277 {
278 	if (iph->frag_off & htons(IP_DF)) {
279 		if (sk && inet_sk(sk)->inet_daddr) {
280 			iph->id = htons(inet_sk(sk)->inet_id);
281 			inet_sk(sk)->inet_id += 1 + more;
282 		} else
283 			iph->id = 0;
284 	} else
285 		__ip_select_ident(iph, dst, more);
286 }
287 
288 /*
289  *	Map a multicast IP onto multicast MAC for type ethernet.
290  */
291 
ip_eth_mc_map(__be32 naddr,char * buf)292 static inline void ip_eth_mc_map(__be32 naddr, char *buf)
293 {
294 	__u32 addr=ntohl(naddr);
295 	buf[0]=0x01;
296 	buf[1]=0x00;
297 	buf[2]=0x5e;
298 	buf[5]=addr&0xFF;
299 	addr>>=8;
300 	buf[4]=addr&0xFF;
301 	addr>>=8;
302 	buf[3]=addr&0x7F;
303 }
304 
305 /*
306  *	Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
307  *	Leave P_Key as 0 to be filled in by driver.
308  */
309 
ip_ib_mc_map(__be32 naddr,const unsigned char * broadcast,char * buf)310 static inline void ip_ib_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
311 {
312 	__u32 addr;
313 	unsigned char scope = broadcast[5] & 0xF;
314 
315 	buf[0]  = 0;		/* Reserved */
316 	buf[1]  = 0xff;		/* Multicast QPN */
317 	buf[2]  = 0xff;
318 	buf[3]  = 0xff;
319 	addr    = ntohl(naddr);
320 	buf[4]  = 0xff;
321 	buf[5]  = 0x10 | scope;	/* scope from broadcast address */
322 	buf[6]  = 0x40;		/* IPv4 signature */
323 	buf[7]  = 0x1b;
324 	buf[8]  = broadcast[8];		/* P_Key */
325 	buf[9]  = broadcast[9];
326 	buf[10] = 0;
327 	buf[11] = 0;
328 	buf[12] = 0;
329 	buf[13] = 0;
330 	buf[14] = 0;
331 	buf[15] = 0;
332 	buf[19] = addr & 0xff;
333 	addr  >>= 8;
334 	buf[18] = addr & 0xff;
335 	addr  >>= 8;
336 	buf[17] = addr & 0xff;
337 	addr  >>= 8;
338 	buf[16] = addr & 0x0f;
339 }
340 
ip_ipgre_mc_map(__be32 naddr,const unsigned char * broadcast,char * buf)341 static inline void ip_ipgre_mc_map(__be32 naddr, const unsigned char *broadcast, char *buf)
342 {
343 	if ((broadcast[0] | broadcast[1] | broadcast[2] | broadcast[3]) != 0)
344 		memcpy(buf, broadcast, 4);
345 	else
346 		memcpy(buf, &naddr, sizeof(naddr));
347 }
348 
349 #if IS_ENABLED(CONFIG_IPV6)
350 #include <linux/ipv6.h>
351 #endif
352 
inet_reset_saddr(struct sock * sk)353 static __inline__ void inet_reset_saddr(struct sock *sk)
354 {
355 	inet_sk(sk)->inet_rcv_saddr = inet_sk(sk)->inet_saddr = 0;
356 #if IS_ENABLED(CONFIG_IPV6)
357 	if (sk->sk_family == PF_INET6) {
358 		struct ipv6_pinfo *np = inet6_sk(sk);
359 
360 		memset(&np->saddr, 0, sizeof(np->saddr));
361 		memset(&np->rcv_saddr, 0, sizeof(np->rcv_saddr));
362 	}
363 #endif
364 }
365 
366 #endif
367 
sk_mc_loop(struct sock * sk)368 static inline int sk_mc_loop(struct sock *sk)
369 {
370 	if (!sk)
371 		return 1;
372 	switch (sk->sk_family) {
373 	case AF_INET:
374 		return inet_sk(sk)->mc_loop;
375 #if IS_ENABLED(CONFIG_IPV6)
376 	case AF_INET6:
377 		return inet6_sk(sk)->mc_loop;
378 #endif
379 	}
380 	WARN_ON(1);
381 	return 1;
382 }
383 
384 extern bool ip_call_ra_chain(struct sk_buff *skb);
385 
386 /*
387  *	Functions provided by ip_fragment.c
388  */
389 
390 enum ip_defrag_users {
391 	IP_DEFRAG_LOCAL_DELIVER,
392 	IP_DEFRAG_CALL_RA_CHAIN,
393 	IP_DEFRAG_CONNTRACK_IN,
394 	__IP_DEFRAG_CONNTRACK_IN_END	= IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
395 	IP_DEFRAG_CONNTRACK_OUT,
396 	__IP_DEFRAG_CONNTRACK_OUT_END	= IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
397 	IP_DEFRAG_CONNTRACK_BRIDGE_IN,
398 	__IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
399 	IP_DEFRAG_VS_IN,
400 	IP_DEFRAG_VS_OUT,
401 	IP_DEFRAG_VS_FWD,
402 	IP_DEFRAG_AF_PACKET,
403 	IP_DEFRAG_MACVLAN,
404 };
405 
406 int ip_defrag(struct sk_buff *skb, u32 user);
407 #ifdef CONFIG_INET
408 struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user);
409 #else
ip_check_defrag(struct sk_buff * skb,u32 user)410 static inline struct sk_buff *ip_check_defrag(struct sk_buff *skb, u32 user)
411 {
412 	return skb;
413 }
414 #endif
415 int ip_frag_mem(struct net *net);
416 int ip_frag_nqueues(struct net *net);
417 
418 /*
419  *	Functions provided by ip_forward.c
420  */
421 
422 extern int ip_forward(struct sk_buff *skb);
423 
424 /*
425  *	Functions provided by ip_options.c
426  */
427 
428 extern void ip_options_build(struct sk_buff *skb, struct ip_options *opt,
429 			     __be32 daddr, struct rtable *rt, int is_frag);
430 extern int ip_options_echo(struct ip_options *dopt, struct sk_buff *skb);
431 extern void ip_options_fragment(struct sk_buff *skb);
432 extern int ip_options_compile(struct net *net,
433 			      struct ip_options *opt, struct sk_buff *skb);
434 extern int ip_options_get(struct net *net, struct ip_options_rcu **optp,
435 			  unsigned char *data, int optlen);
436 extern int ip_options_get_from_user(struct net *net, struct ip_options_rcu **optp,
437 				    unsigned char __user *data, int optlen);
438 extern void ip_options_undo(struct ip_options * opt);
439 extern void ip_forward_options(struct sk_buff *skb);
440 extern int ip_options_rcv_srr(struct sk_buff *skb);
441 
442 /*
443  *	Functions provided by ip_sockglue.c
444  */
445 
446 extern void	ipv4_pktinfo_prepare(struct sk_buff *skb);
447 extern void	ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb);
448 extern int	ip_cmsg_send(struct net *net,
449 			     struct msghdr *msg, struct ipcm_cookie *ipc);
450 extern int	ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval, unsigned int optlen);
451 extern int	ip_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen);
452 extern int	compat_ip_setsockopt(struct sock *sk, int level,
453 			int optname, char __user *optval, unsigned int optlen);
454 extern int	compat_ip_getsockopt(struct sock *sk, int level,
455 			int optname, char __user *optval, int __user *optlen);
456 extern int	ip_ra_control(struct sock *sk, unsigned char on, void (*destructor)(struct sock *));
457 
458 extern int 	ip_recv_error(struct sock *sk, struct msghdr *msg, int len, int *addr_len);
459 extern void	ip_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
460 			      __be16 port, u32 info, u8 *payload);
461 extern void	ip_local_error(struct sock *sk, int err, __be32 daddr, __be16 dport,
462 			       u32 info);
463 
464 #ifdef CONFIG_PROC_FS
465 extern int ip_misc_proc_init(void);
466 #endif
467 
468 #endif	/* _IP_H */
469