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.
6 * Definitions for the IP module.
8 * Version: @(#)ip.h 1.0.2 05/07/93
11 * Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
12 * Alan Cox, <gw4pts@gw4pts.ampr.org>
15 * Mike McLagan : Routing by source
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.
25 #include <linux/types.h>
28 #include <linux/skbuff.h>
30 #include <net/inet_sock.h>
31 #include <net/route.h>
37 struct inet_skb_parm
{
38 struct ip_options opt
; /* Compiled IP options */
41 #define IPSKB_FORWARDED BIT(0)
42 #define IPSKB_XFRM_TUNNEL_SIZE BIT(1)
43 #define IPSKB_XFRM_TRANSFORMED BIT(2)
44 #define IPSKB_FRAG_COMPLETE BIT(3)
45 #define IPSKB_REROUTED BIT(4)
46 #define IPSKB_DOREDIRECT BIT(5)
51 static inline unsigned int ip_hdrlen(const struct sk_buff
*skb
)
53 return ip_hdr(skb
)->ihl
* 4;
59 struct ip_options_rcu
*opt
;
66 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
67 #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
70 struct ip_ra_chain __rcu
*next
;
73 void (*destructor
)(struct sock
*);
74 struct sock
*saved_sk
;
79 extern struct ip_ra_chain __rcu
*ip_ra_chain
;
82 #define IP_CE 0x8000 /* Flag: "Congestion" */
83 #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
84 #define IP_MF 0x2000 /* Flag: "More Fragments" */
85 #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
87 #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
95 int igmp_mc_init(void);
98 * Functions provided by ip.c
101 int ip_build_and_send_pkt(struct sk_buff
*skb
, struct sock
*sk
,
102 __be32 saddr
, __be32 daddr
,
103 struct ip_options_rcu
*opt
);
104 int ip_rcv(struct sk_buff
*skb
, struct net_device
*dev
, struct packet_type
*pt
,
105 struct net_device
*orig_dev
);
106 int ip_local_deliver(struct sk_buff
*skb
);
107 int ip_mr_input(struct sk_buff
*skb
);
108 int ip_output(struct sock
*sk
, struct sk_buff
*skb
);
109 int ip_mc_output(struct sock
*sk
, struct sk_buff
*skb
);
110 int ip_fragment(struct sk_buff
*skb
, int (*output
)(struct sk_buff
*));
111 int ip_do_nat(struct sk_buff
*skb
);
112 void ip_send_check(struct iphdr
*ip
);
113 int __ip_local_out(struct sk_buff
*skb
);
114 int ip_local_out_sk(struct sock
*sk
, struct sk_buff
*skb
);
115 static inline int ip_local_out(struct sk_buff
*skb
)
117 return ip_local_out_sk(skb
->sk
, skb
);
120 int ip_queue_xmit(struct sock
*sk
, struct sk_buff
*skb
, struct flowi
*fl
);
122 int ip_append_data(struct sock
*sk
, struct flowi4
*fl4
,
123 int getfrag(void *from
, char *to
, int offset
, int len
,
124 int odd
, struct sk_buff
*skb
),
125 void *from
, int len
, int protolen
,
126 struct ipcm_cookie
*ipc
,
129 int ip_generic_getfrag(void *from
, char *to
, int offset
, int len
, int odd
,
130 struct sk_buff
*skb
);
131 ssize_t
ip_append_page(struct sock
*sk
, struct flowi4
*fl4
, struct page
*page
,
132 int offset
, size_t size
, int flags
);
133 struct sk_buff
*__ip_make_skb(struct sock
*sk
, struct flowi4
*fl4
,
134 struct sk_buff_head
*queue
,
135 struct inet_cork
*cork
);
136 int ip_send_skb(struct net
*net
, struct sk_buff
*skb
);
137 int ip_push_pending_frames(struct sock
*sk
, struct flowi4
*fl4
);
138 void ip_flush_pending_frames(struct sock
*sk
);
139 struct sk_buff
*ip_make_skb(struct sock
*sk
, struct flowi4
*fl4
,
140 int getfrag(void *from
, char *to
, int offset
,
141 int len
, int odd
, struct sk_buff
*skb
),
142 void *from
, int length
, int transhdrlen
,
143 struct ipcm_cookie
*ipc
, struct rtable
**rtp
,
146 static inline struct sk_buff
*ip_finish_skb(struct sock
*sk
, struct flowi4
*fl4
)
148 return __ip_make_skb(sk
, fl4
, &sk
->sk_write_queue
, &inet_sk(sk
)->cork
.base
);
151 static inline __u8
get_rttos(struct ipcm_cookie
* ipc
, struct inet_sock
*inet
)
153 return (ipc
->tos
!= -1) ? RT_TOS(ipc
->tos
) : RT_TOS(inet
->tos
);
156 static inline __u8
get_rtconn_flags(struct ipcm_cookie
* ipc
, struct sock
* sk
)
158 return (ipc
->tos
!= -1) ? RT_CONN_FLAGS_TOS(sk
, ipc
->tos
) : RT_CONN_FLAGS(sk
);
162 int __ip4_datagram_connect(struct sock
*sk
, struct sockaddr
*uaddr
, int addr_len
);
163 int ip4_datagram_connect(struct sock
*sk
, struct sockaddr
*uaddr
, int addr_len
);
165 void ip4_datagram_release_cb(struct sock
*sk
);
167 struct ip_reply_arg
{
171 int csumoffset
; /* u16 offset of csum in iov[0].iov_base */
172 /* -1 if not needed */
177 #define IP_REPLY_ARG_NOSRCCHECK 1
179 static inline __u8
ip_reply_arg_flowi_flags(const struct ip_reply_arg
*arg
)
181 return (arg
->flags
& IP_REPLY_ARG_NOSRCCHECK
) ? FLOWI_FLAG_ANYSRC
: 0;
184 void ip_send_unicast_reply(struct sock
*sk
, struct sk_buff
*skb
, __be32 daddr
,
185 __be32 saddr
, const struct ip_reply_arg
*arg
,
188 #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
189 #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
190 #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
191 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
192 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
193 #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
194 #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
195 #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
196 #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
197 #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
198 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
199 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
201 unsigned long snmp_fold_field(void __percpu
*mib
, int offt
);
202 #if BITS_PER_LONG==32
203 u64
snmp_fold_field64(void __percpu
*mib
, int offt
, size_t sync_off
);
205 static inline u64
snmp_fold_field64(void __percpu
*mib
, int offt
, size_t syncp_off
)
207 return snmp_fold_field(mib
, offt
);
211 void inet_get_local_port_range(struct net
*net
, int *low
, int *high
);
214 static inline int inet_is_local_reserved_port(struct net
*net
, int port
)
216 if (!net
->ipv4
.sysctl_local_reserved_ports
)
218 return test_bit(port
, net
->ipv4
.sysctl_local_reserved_ports
);
221 static inline int inet_is_local_reserved_port(struct net
*net
, int port
)
227 extern int sysctl_ip_nonlocal_bind
;
229 /* From inetpeer.c */
230 extern int inet_peer_threshold
;
231 extern int inet_peer_minttl
;
232 extern int inet_peer_maxttl
;
234 /* From ip_input.c */
235 extern int sysctl_ip_early_demux
;
237 /* From ip_output.c */
238 extern int sysctl_ip_dynaddr
;
240 void ipfrag_init(void);
242 void ip_static_sysctl_init(void);
244 #define IP4_REPLY_MARK(net, mark) \
245 ((net)->ipv4.sysctl_fwmark_reflect ? (mark) : 0)
247 static inline bool ip_is_fragment(const struct iphdr
*iph
)
249 return (iph
->frag_off
& htons(IP_MF
| IP_OFFSET
)) != 0;
255 /* The function in 2.2 was invalid, producing wrong result for
256 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
258 int ip_decrease_ttl(struct iphdr
*iph
)
260 u32 check
= (__force u32
)iph
->check
;
261 check
+= (__force u32
)htons(0x0100);
262 iph
->check
= (__force __sum16
)(check
+ (check
>=0xFFFF));
267 int ip_dont_fragment(struct sock
*sk
, struct dst_entry
*dst
)
269 return inet_sk(sk
)->pmtudisc
== IP_PMTUDISC_DO
||
270 (inet_sk(sk
)->pmtudisc
== IP_PMTUDISC_WANT
&&
271 !(dst_metric_locked(dst
, RTAX_MTU
)));
274 static inline bool ip_sk_accept_pmtu(const struct sock
*sk
)
276 return inet_sk(sk
)->pmtudisc
!= IP_PMTUDISC_INTERFACE
&&
277 inet_sk(sk
)->pmtudisc
!= IP_PMTUDISC_OMIT
;
280 static inline bool ip_sk_use_pmtu(const struct sock
*sk
)
282 return inet_sk(sk
)->pmtudisc
< IP_PMTUDISC_PROBE
;
285 static inline bool ip_sk_ignore_df(const struct sock
*sk
)
287 return inet_sk(sk
)->pmtudisc
< IP_PMTUDISC_DO
||
288 inet_sk(sk
)->pmtudisc
== IP_PMTUDISC_OMIT
;
291 static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry
*dst
,
294 struct net
*net
= dev_net(dst
->dev
);
296 if (net
->ipv4
.sysctl_ip_fwd_use_pmtu
||
297 dst_metric_locked(dst
, RTAX_MTU
) ||
301 return min(dst
->dev
->mtu
, IP_MAX_MTU
);
304 static inline unsigned int ip_skb_dst_mtu(const struct sk_buff
*skb
)
306 if (!skb
->sk
|| ip_sk_use_pmtu(skb
->sk
)) {
307 bool forwarding
= IPCB(skb
)->flags
& IPSKB_FORWARDED
;
308 return ip_dst_mtu_maybe_forward(skb_dst(skb
), forwarding
);
310 return min(skb_dst(skb
)->dev
->mtu
, IP_MAX_MTU
);
314 u32
ip_idents_reserve(u32 hash
, int segs
);
315 void __ip_select_ident(struct iphdr
*iph
, int segs
);
317 static inline void ip_select_ident_segs(struct sk_buff
*skb
, struct sock
*sk
, int segs
)
319 struct iphdr
*iph
= ip_hdr(skb
);
321 if ((iph
->frag_off
& htons(IP_DF
)) && !skb
->ignore_df
) {
322 /* This is only to work around buggy Windows95/2000
323 * VJ compression implementations. If the ID field
324 * does not change, they drop every other packet in
325 * a TCP stream using header compression.
327 if (sk
&& inet_sk(sk
)->inet_daddr
) {
328 iph
->id
= htons(inet_sk(sk
)->inet_id
);
329 inet_sk(sk
)->inet_id
+= segs
;
334 __ip_select_ident(iph
, segs
);
338 static inline void ip_select_ident(struct sk_buff
*skb
, struct sock
*sk
)
340 ip_select_ident_segs(skb
, sk
, 1);
343 static inline __wsum
inet_compute_pseudo(struct sk_buff
*skb
, int proto
)
345 return csum_tcpudp_nofold(ip_hdr(skb
)->saddr
, ip_hdr(skb
)->daddr
,
350 * Map a multicast IP onto multicast MAC for type ethernet.
353 static inline void ip_eth_mc_map(__be32 naddr
, char *buf
)
355 __u32 addr
=ntohl(naddr
);
367 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
368 * Leave P_Key as 0 to be filled in by driver.
371 static inline void ip_ib_mc_map(__be32 naddr
, const unsigned char *broadcast
, char *buf
)
374 unsigned char scope
= broadcast
[5] & 0xF;
376 buf
[0] = 0; /* Reserved */
377 buf
[1] = 0xff; /* Multicast QPN */
382 buf
[5] = 0x10 | scope
; /* scope from broadcast address */
383 buf
[6] = 0x40; /* IPv4 signature */
385 buf
[8] = broadcast
[8]; /* P_Key */
386 buf
[9] = broadcast
[9];
393 buf
[19] = addr
& 0xff;
395 buf
[18] = addr
& 0xff;
397 buf
[17] = addr
& 0xff;
399 buf
[16] = addr
& 0x0f;
402 static inline void ip_ipgre_mc_map(__be32 naddr
, const unsigned char *broadcast
, char *buf
)
404 if ((broadcast
[0] | broadcast
[1] | broadcast
[2] | broadcast
[3]) != 0)
405 memcpy(buf
, broadcast
, 4);
407 memcpy(buf
, &naddr
, sizeof(naddr
));
410 #if IS_ENABLED(CONFIG_IPV6)
411 #include <linux/ipv6.h>
414 static __inline__
void inet_reset_saddr(struct sock
*sk
)
416 inet_sk(sk
)->inet_rcv_saddr
= inet_sk(sk
)->inet_saddr
= 0;
417 #if IS_ENABLED(CONFIG_IPV6)
418 if (sk
->sk_family
== PF_INET6
) {
419 struct ipv6_pinfo
*np
= inet6_sk(sk
);
421 memset(&np
->saddr
, 0, sizeof(np
->saddr
));
422 memset(&sk
->sk_v6_rcv_saddr
, 0, sizeof(sk
->sk_v6_rcv_saddr
));
429 bool ip_call_ra_chain(struct sk_buff
*skb
);
432 * Functions provided by ip_fragment.c
435 enum ip_defrag_users
{
436 IP_DEFRAG_LOCAL_DELIVER
,
437 IP_DEFRAG_CALL_RA_CHAIN
,
438 IP_DEFRAG_CONNTRACK_IN
,
439 __IP_DEFRAG_CONNTRACK_IN_END
= IP_DEFRAG_CONNTRACK_IN
+ USHRT_MAX
,
440 IP_DEFRAG_CONNTRACK_OUT
,
441 __IP_DEFRAG_CONNTRACK_OUT_END
= IP_DEFRAG_CONNTRACK_OUT
+ USHRT_MAX
,
442 IP_DEFRAG_CONNTRACK_BRIDGE_IN
,
443 __IP_DEFRAG_CONNTRACK_BRIDGE_IN
= IP_DEFRAG_CONNTRACK_BRIDGE_IN
+ USHRT_MAX
,
451 int ip_defrag(struct sk_buff
*skb
, u32 user
);
453 struct sk_buff
*ip_check_defrag(struct sk_buff
*skb
, u32 user
);
455 static inline struct sk_buff
*ip_check_defrag(struct sk_buff
*skb
, u32 user
)
460 int ip_frag_mem(struct net
*net
);
461 int ip_frag_nqueues(struct net
*net
);
464 * Functions provided by ip_forward.c
467 int ip_forward(struct sk_buff
*skb
);
470 * Functions provided by ip_options.c
473 void ip_options_build(struct sk_buff
*skb
, struct ip_options
*opt
,
474 __be32 daddr
, struct rtable
*rt
, int is_frag
);
475 int ip_options_echo(struct ip_options
*dopt
, struct sk_buff
*skb
);
476 void ip_options_fragment(struct sk_buff
*skb
);
477 int ip_options_compile(struct net
*net
, struct ip_options
*opt
,
478 struct sk_buff
*skb
);
479 int ip_options_get(struct net
*net
, struct ip_options_rcu
**optp
,
480 unsigned char *data
, int optlen
);
481 int ip_options_get_from_user(struct net
*net
, struct ip_options_rcu
**optp
,
482 unsigned char __user
*data
, int optlen
);
483 void ip_options_undo(struct ip_options
*opt
);
484 void ip_forward_options(struct sk_buff
*skb
);
485 int ip_options_rcv_srr(struct sk_buff
*skb
);
488 * Functions provided by ip_sockglue.c
491 void ipv4_pktinfo_prepare(const struct sock
*sk
, struct sk_buff
*skb
);
492 void ip_cmsg_recv(struct msghdr
*msg
, struct sk_buff
*skb
);
493 int ip_cmsg_send(struct net
*net
, struct msghdr
*msg
,
494 struct ipcm_cookie
*ipc
, bool allow_ipv6
);
495 int ip_setsockopt(struct sock
*sk
, int level
, int optname
, char __user
*optval
,
496 unsigned int optlen
);
497 int ip_getsockopt(struct sock
*sk
, int level
, int optname
, char __user
*optval
,
499 int compat_ip_setsockopt(struct sock
*sk
, int level
, int optname
,
500 char __user
*optval
, unsigned int optlen
);
501 int compat_ip_getsockopt(struct sock
*sk
, int level
, int optname
,
502 char __user
*optval
, int __user
*optlen
);
503 int ip_ra_control(struct sock
*sk
, unsigned char on
,
504 void (*destructor
)(struct sock
*));
506 int ip_recv_error(struct sock
*sk
, struct msghdr
*msg
, int len
, int *addr_len
);
507 void ip_icmp_error(struct sock
*sk
, struct sk_buff
*skb
, int err
, __be16 port
,
508 u32 info
, u8
*payload
);
509 void ip_local_error(struct sock
*sk
, int err
, __be32 daddr
, __be16 dport
,
512 #ifdef CONFIG_PROC_FS
513 int ip_misc_proc_init(void);