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 1
42 #define IPSKB_XFRM_TUNNEL_SIZE 2
43 #define IPSKB_XFRM_TRANSFORMED 4
44 #define IPSKB_FRAG_COMPLETE 8
45 #define IPSKB_REROUTED 16
50 static inline unsigned int ip_hdrlen(const struct sk_buff
*skb
)
52 return ip_hdr(skb
)->ihl
* 4;
58 struct ip_options_rcu
*opt
;
65 #define IPCB(skb) ((struct inet_skb_parm*)((skb)->cb))
66 #define PKTINFO_SKB_CB(skb) ((struct in_pktinfo *)((skb)->cb))
69 struct ip_ra_chain __rcu
*next
;
72 void (*destructor
)(struct sock
*);
73 struct sock
*saved_sk
;
78 extern struct ip_ra_chain __rcu
*ip_ra_chain
;
81 #define IP_CE 0x8000 /* Flag: "Congestion" */
82 #define IP_DF 0x4000 /* Flag: "Don't Fragment" */
83 #define IP_MF 0x2000 /* Flag: "More Fragments" */
84 #define IP_OFFSET 0x1FFF /* "Fragment Offset" part */
86 #define IP_FRAG_TIME (30 * HZ) /* fragment lifetime */
94 int igmp_mc_init(void);
97 * Functions provided by ip.c
100 int ip_build_and_send_pkt(struct sk_buff
*skb
, struct sock
*sk
,
101 __be32 saddr
, __be32 daddr
,
102 struct ip_options_rcu
*opt
);
103 int ip_rcv(struct sk_buff
*skb
, struct net_device
*dev
, struct packet_type
*pt
,
104 struct net_device
*orig_dev
);
105 int ip_local_deliver(struct sk_buff
*skb
);
106 int ip_mr_input(struct sk_buff
*skb
);
107 int ip_output(struct sock
*sk
, struct sk_buff
*skb
);
108 int ip_mc_output(struct sock
*sk
, struct sk_buff
*skb
);
109 int ip_fragment(struct sk_buff
*skb
, int (*output
)(struct sk_buff
*));
110 int ip_do_nat(struct sk_buff
*skb
);
111 void ip_send_check(struct iphdr
*ip
);
112 int __ip_local_out(struct sk_buff
*skb
);
113 int ip_local_out_sk(struct sock
*sk
, struct sk_buff
*skb
);
114 static inline int ip_local_out(struct sk_buff
*skb
)
116 return ip_local_out_sk(skb
->sk
, skb
);
119 int ip_queue_xmit(struct sock
*sk
, struct sk_buff
*skb
, struct flowi
*fl
);
121 int ip_append_data(struct sock
*sk
, struct flowi4
*fl4
,
122 int getfrag(void *from
, char *to
, int offset
, int len
,
123 int odd
, struct sk_buff
*skb
),
124 void *from
, int len
, int protolen
,
125 struct ipcm_cookie
*ipc
,
128 int ip_generic_getfrag(void *from
, char *to
, int offset
, int len
, int odd
,
129 struct sk_buff
*skb
);
130 ssize_t
ip_append_page(struct sock
*sk
, struct flowi4
*fl4
, struct page
*page
,
131 int offset
, size_t size
, int flags
);
132 struct sk_buff
*__ip_make_skb(struct sock
*sk
, struct flowi4
*fl4
,
133 struct sk_buff_head
*queue
,
134 struct inet_cork
*cork
);
135 int ip_send_skb(struct net
*net
, struct sk_buff
*skb
);
136 int ip_push_pending_frames(struct sock
*sk
, struct flowi4
*fl4
);
137 void ip_flush_pending_frames(struct sock
*sk
);
138 struct sk_buff
*ip_make_skb(struct sock
*sk
, struct flowi4
*fl4
,
139 int getfrag(void *from
, char *to
, int offset
,
140 int len
, int odd
, struct sk_buff
*skb
),
141 void *from
, int length
, int transhdrlen
,
142 struct ipcm_cookie
*ipc
, struct rtable
**rtp
,
145 static inline struct sk_buff
*ip_finish_skb(struct sock
*sk
, struct flowi4
*fl4
)
147 return __ip_make_skb(sk
, fl4
, &sk
->sk_write_queue
, &inet_sk(sk
)->cork
.base
);
150 static inline __u8
get_rttos(struct ipcm_cookie
* ipc
, struct inet_sock
*inet
)
152 return (ipc
->tos
!= -1) ? RT_TOS(ipc
->tos
) : RT_TOS(inet
->tos
);
155 static inline __u8
get_rtconn_flags(struct ipcm_cookie
* ipc
, struct sock
* sk
)
157 return (ipc
->tos
!= -1) ? RT_CONN_FLAGS_TOS(sk
, ipc
->tos
) : RT_CONN_FLAGS(sk
);
161 int ip4_datagram_connect(struct sock
*sk
, struct sockaddr
*uaddr
, int addr_len
);
163 void ip4_datagram_release_cb(struct sock
*sk
);
165 struct ip_reply_arg
{
169 int csumoffset
; /* u16 offset of csum in iov[0].iov_base */
170 /* -1 if not needed */
175 #define IP_REPLY_ARG_NOSRCCHECK 1
177 static inline __u8
ip_reply_arg_flowi_flags(const struct ip_reply_arg
*arg
)
179 return (arg
->flags
& IP_REPLY_ARG_NOSRCCHECK
) ? FLOWI_FLAG_ANYSRC
: 0;
182 void ip_send_unicast_reply(struct net
*net
, struct sk_buff
*skb
, __be32 daddr
,
183 __be32 saddr
, const struct ip_reply_arg
*arg
,
186 #define IP_INC_STATS(net, field) SNMP_INC_STATS64((net)->mib.ip_statistics, field)
187 #define IP_INC_STATS_BH(net, field) SNMP_INC_STATS64_BH((net)->mib.ip_statistics, field)
188 #define IP_ADD_STATS(net, field, val) SNMP_ADD_STATS64((net)->mib.ip_statistics, field, val)
189 #define IP_ADD_STATS_BH(net, field, val) SNMP_ADD_STATS64_BH((net)->mib.ip_statistics, field, val)
190 #define IP_UPD_PO_STATS(net, field, val) SNMP_UPD_PO_STATS64((net)->mib.ip_statistics, field, val)
191 #define IP_UPD_PO_STATS_BH(net, field, val) SNMP_UPD_PO_STATS64_BH((net)->mib.ip_statistics, field, val)
192 #define NET_INC_STATS(net, field) SNMP_INC_STATS((net)->mib.net_statistics, field)
193 #define NET_INC_STATS_BH(net, field) SNMP_INC_STATS_BH((net)->mib.net_statistics, field)
194 #define NET_INC_STATS_USER(net, field) SNMP_INC_STATS_USER((net)->mib.net_statistics, field)
195 #define NET_ADD_STATS(net, field, adnd) SNMP_ADD_STATS((net)->mib.net_statistics, field, adnd)
196 #define NET_ADD_STATS_BH(net, field, adnd) SNMP_ADD_STATS_BH((net)->mib.net_statistics, field, adnd)
197 #define NET_ADD_STATS_USER(net, field, adnd) SNMP_ADD_STATS_USER((net)->mib.net_statistics, field, adnd)
199 unsigned long snmp_fold_field(void __percpu
*mib
[], int offt
);
200 #if BITS_PER_LONG==32
201 u64
snmp_fold_field64(void __percpu
*mib
[], int offt
, size_t sync_off
);
203 static inline u64
snmp_fold_field64(void __percpu
*mib
[], int offt
, size_t syncp_off
)
205 return snmp_fold_field(mib
, offt
);
208 int snmp_mib_init(void __percpu
*ptr
[2], size_t mibsize
, size_t align
);
210 static inline void snmp_mib_free(void __percpu
*ptr
[SNMP_ARRAY_SZ
])
215 for (i
= 0; i
< SNMP_ARRAY_SZ
; i
++) {
221 void inet_get_local_port_range(struct net
*net
, int *low
, int *high
);
223 extern unsigned long *sysctl_local_reserved_ports
;
224 static inline int inet_is_reserved_local_port(int port
)
226 return test_bit(port
, sysctl_local_reserved_ports
);
229 extern int sysctl_ip_nonlocal_bind
;
231 /* From inetpeer.c */
232 extern int inet_peer_threshold
;
233 extern int inet_peer_minttl
;
234 extern int inet_peer_maxttl
;
236 /* From ip_input.c */
237 extern int sysctl_ip_early_demux
;
239 /* From ip_output.c */
240 extern int sysctl_ip_dynaddr
;
242 void ipfrag_init(void);
244 void ip_static_sysctl_init(void);
246 static inline bool ip_is_fragment(const struct iphdr
*iph
)
248 return (iph
->frag_off
& htons(IP_MF
| IP_OFFSET
)) != 0;
254 /* The function in 2.2 was invalid, producing wrong result for
255 * check=0xFEFF. It was noticed by Arthur Skawina _year_ ago. --ANK(000625) */
257 int ip_decrease_ttl(struct iphdr
*iph
)
259 u32 check
= (__force u32
)iph
->check
;
260 check
+= (__force u32
)htons(0x0100);
261 iph
->check
= (__force __sum16
)(check
+ (check
>=0xFFFF));
266 int ip_dont_fragment(struct sock
*sk
, struct dst_entry
*dst
)
268 return inet_sk(sk
)->pmtudisc
== IP_PMTUDISC_DO
||
269 (inet_sk(sk
)->pmtudisc
== IP_PMTUDISC_WANT
&&
270 !(dst_metric_locked(dst
, RTAX_MTU
)));
273 static inline bool ip_sk_accept_pmtu(const struct sock
*sk
)
275 return inet_sk(sk
)->pmtudisc
!= IP_PMTUDISC_INTERFACE
&&
276 inet_sk(sk
)->pmtudisc
!= IP_PMTUDISC_OMIT
;
279 static inline bool ip_sk_use_pmtu(const struct sock
*sk
)
281 return inet_sk(sk
)->pmtudisc
< IP_PMTUDISC_PROBE
;
284 static inline bool ip_sk_local_df(const struct sock
*sk
)
286 return inet_sk(sk
)->pmtudisc
< IP_PMTUDISC_DO
||
287 inet_sk(sk
)->pmtudisc
== IP_PMTUDISC_OMIT
;
290 static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry
*dst
,
293 struct net
*net
= dev_net(dst
->dev
);
295 if (net
->ipv4
.sysctl_ip_fwd_use_pmtu
||
296 dst_metric_locked(dst
, RTAX_MTU
) ||
300 return min(dst
->dev
->mtu
, IP_MAX_MTU
);
303 static inline unsigned int ip_skb_dst_mtu(const struct sk_buff
*skb
)
305 if (!skb
->sk
|| ip_sk_use_pmtu(skb
->sk
)) {
306 bool forwarding
= IPCB(skb
)->flags
& IPSKB_FORWARDED
;
307 return ip_dst_mtu_maybe_forward(skb_dst(skb
), forwarding
);
309 return min(skb_dst(skb
)->dev
->mtu
, IP_MAX_MTU
);
313 void __ip_select_ident(struct iphdr
*iph
, struct dst_entry
*dst
, int more
);
315 static inline void ip_select_ident(struct sk_buff
*skb
, struct dst_entry
*dst
, struct sock
*sk
)
317 struct iphdr
*iph
= ip_hdr(skb
);
319 if ((iph
->frag_off
& htons(IP_DF
)) && !skb
->local_df
) {
320 /* This is only to work around buggy Windows95/2000
321 * VJ compression implementations. If the ID field
322 * does not change, they drop every other packet in
323 * a TCP stream using header compression.
325 iph
->id
= (sk
&& inet_sk(sk
)->inet_daddr
) ?
326 htons(inet_sk(sk
)->inet_id
++) : 0;
328 __ip_select_ident(iph
, dst
, 0);
331 static inline void ip_select_ident_more(struct sk_buff
*skb
, struct dst_entry
*dst
, struct sock
*sk
, int more
)
333 struct iphdr
*iph
= ip_hdr(skb
);
335 if ((iph
->frag_off
& htons(IP_DF
)) && !skb
->local_df
) {
336 if (sk
&& inet_sk(sk
)->inet_daddr
) {
337 iph
->id
= htons(inet_sk(sk
)->inet_id
);
338 inet_sk(sk
)->inet_id
+= 1 + more
;
342 __ip_select_ident(iph
, dst
, more
);
345 static inline __wsum
inet_compute_pseudo(struct sk_buff
*skb
, int proto
)
347 return csum_tcpudp_nofold(ip_hdr(skb
)->saddr
, ip_hdr(skb
)->daddr
,
352 * Map a multicast IP onto multicast MAC for type ethernet.
355 static inline void ip_eth_mc_map(__be32 naddr
, char *buf
)
357 __u32 addr
=ntohl(naddr
);
369 * Map a multicast IP onto multicast MAC for type IP-over-InfiniBand.
370 * Leave P_Key as 0 to be filled in by driver.
373 static inline void ip_ib_mc_map(__be32 naddr
, const unsigned char *broadcast
, char *buf
)
376 unsigned char scope
= broadcast
[5] & 0xF;
378 buf
[0] = 0; /* Reserved */
379 buf
[1] = 0xff; /* Multicast QPN */
384 buf
[5] = 0x10 | scope
; /* scope from broadcast address */
385 buf
[6] = 0x40; /* IPv4 signature */
387 buf
[8] = broadcast
[8]; /* P_Key */
388 buf
[9] = broadcast
[9];
395 buf
[19] = addr
& 0xff;
397 buf
[18] = addr
& 0xff;
399 buf
[17] = addr
& 0xff;
401 buf
[16] = addr
& 0x0f;
404 static inline void ip_ipgre_mc_map(__be32 naddr
, const unsigned char *broadcast
, char *buf
)
406 if ((broadcast
[0] | broadcast
[1] | broadcast
[2] | broadcast
[3]) != 0)
407 memcpy(buf
, broadcast
, 4);
409 memcpy(buf
, &naddr
, sizeof(naddr
));
412 #if IS_ENABLED(CONFIG_IPV6)
413 #include <linux/ipv6.h>
416 static __inline__
void inet_reset_saddr(struct sock
*sk
)
418 inet_sk(sk
)->inet_rcv_saddr
= inet_sk(sk
)->inet_saddr
= 0;
419 #if IS_ENABLED(CONFIG_IPV6)
420 if (sk
->sk_family
== PF_INET6
) {
421 struct ipv6_pinfo
*np
= inet6_sk(sk
);
423 memset(&np
->saddr
, 0, sizeof(np
->saddr
));
424 memset(&sk
->sk_v6_rcv_saddr
, 0, sizeof(sk
->sk_v6_rcv_saddr
));
431 static inline int sk_mc_loop(struct sock
*sk
)
435 switch (sk
->sk_family
) {
437 return inet_sk(sk
)->mc_loop
;
438 #if IS_ENABLED(CONFIG_IPV6)
440 return inet6_sk(sk
)->mc_loop
;
447 bool ip_call_ra_chain(struct sk_buff
*skb
);
450 * Functions provided by ip_fragment.c
453 enum ip_defrag_users
{
454 IP_DEFRAG_LOCAL_DELIVER
,
455 IP_DEFRAG_CALL_RA_CHAIN
,
456 IP_DEFRAG_CONNTRACK_IN
,
457 __IP_DEFRAG_CONNTRACK_IN_END
= IP_DEFRAG_CONNTRACK_IN
+ USHRT_MAX
,
458 IP_DEFRAG_CONNTRACK_OUT
,
459 __IP_DEFRAG_CONNTRACK_OUT_END
= IP_DEFRAG_CONNTRACK_OUT
+ USHRT_MAX
,
460 IP_DEFRAG_CONNTRACK_BRIDGE_IN
,
461 __IP_DEFRAG_CONNTRACK_BRIDGE_IN
= IP_DEFRAG_CONNTRACK_BRIDGE_IN
+ USHRT_MAX
,
469 int ip_defrag(struct sk_buff
*skb
, u32 user
);
471 struct sk_buff
*ip_check_defrag(struct sk_buff
*skb
, u32 user
);
473 static inline struct sk_buff
*ip_check_defrag(struct sk_buff
*skb
, u32 user
)
478 int ip_frag_mem(struct net
*net
);
479 int ip_frag_nqueues(struct net
*net
);
482 * Functions provided by ip_forward.c
485 int ip_forward(struct sk_buff
*skb
);
488 * Functions provided by ip_options.c
491 void ip_options_build(struct sk_buff
*skb
, struct ip_options
*opt
,
492 __be32 daddr
, struct rtable
*rt
, int is_frag
);
493 int ip_options_echo(struct ip_options
*dopt
, struct sk_buff
*skb
);
494 void ip_options_fragment(struct sk_buff
*skb
);
495 int ip_options_compile(struct net
*net
, struct ip_options
*opt
,
496 struct sk_buff
*skb
);
497 int ip_options_get(struct net
*net
, struct ip_options_rcu
**optp
,
498 unsigned char *data
, int optlen
);
499 int ip_options_get_from_user(struct net
*net
, struct ip_options_rcu
**optp
,
500 unsigned char __user
*data
, int optlen
);
501 void ip_options_undo(struct ip_options
*opt
);
502 void ip_forward_options(struct sk_buff
*skb
);
503 int ip_options_rcv_srr(struct sk_buff
*skb
);
506 * Functions provided by ip_sockglue.c
509 void ipv4_pktinfo_prepare(const struct sock
*sk
, struct sk_buff
*skb
);
510 void ip_cmsg_recv(struct msghdr
*msg
, struct sk_buff
*skb
);
511 int ip_cmsg_send(struct net
*net
, struct msghdr
*msg
,
512 struct ipcm_cookie
*ipc
, bool allow_ipv6
);
513 int ip_setsockopt(struct sock
*sk
, int level
, int optname
, char __user
*optval
,
514 unsigned int optlen
);
515 int ip_getsockopt(struct sock
*sk
, int level
, int optname
, char __user
*optval
,
517 int compat_ip_setsockopt(struct sock
*sk
, int level
, int optname
,
518 char __user
*optval
, unsigned int optlen
);
519 int compat_ip_getsockopt(struct sock
*sk
, int level
, int optname
,
520 char __user
*optval
, int __user
*optlen
);
521 int ip_ra_control(struct sock
*sk
, unsigned char on
,
522 void (*destructor
)(struct sock
*));
524 int ip_recv_error(struct sock
*sk
, struct msghdr
*msg
, int len
, int *addr_len
);
525 void ip_icmp_error(struct sock
*sk
, struct sk_buff
*skb
, int err
, __be16 port
,
526 u32 info
, u8
*payload
);
527 void ip_local_error(struct sock
*sk
, int err
, __be32 daddr
, __be16 dport
,
530 #ifdef CONFIG_PROC_FS
531 int ip_misc_proc_init(void);