2 * Linux INET6 implementation
5 * Pedro Roque <roque@di.fc.ul.pt>
7 * This program is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation; either version
10 * 2 of the License, or (at your option) any later version.
16 #include <linux/ipv6.h>
17 #include <linux/hardirq.h>
18 #include <linux/jhash.h>
19 #include <net/if_inet6.h>
20 #include <net/ndisc.h>
24 #define SIN6_LEN_RFC2133 24
26 #define IPV6_MAXPLEN 65535
29 * NextHeader field of IPv6 header
32 #define NEXTHDR_HOP 0 /* Hop-by-hop option header. */
33 #define NEXTHDR_TCP 6 /* TCP segment. */
34 #define NEXTHDR_UDP 17 /* UDP message. */
35 #define NEXTHDR_IPV6 41 /* IPv6 in IPv6 */
36 #define NEXTHDR_ROUTING 43 /* Routing header. */
37 #define NEXTHDR_FRAGMENT 44 /* Fragmentation/reassembly header. */
38 #define NEXTHDR_ESP 50 /* Encapsulating security payload. */
39 #define NEXTHDR_AUTH 51 /* Authentication header. */
40 #define NEXTHDR_ICMP 58 /* ICMP for IPv6. */
41 #define NEXTHDR_NONE 59 /* No next header */
42 #define NEXTHDR_DEST 60 /* Destination options header. */
43 #define NEXTHDR_MOBILITY 135 /* Mobility header. */
45 #define NEXTHDR_MAX 255
49 #define IPV6_DEFAULT_HOPLIMIT 64
50 #define IPV6_DEFAULT_MCASTHOPS 1
55 * type - unicast | multicast
56 * scope - local | site | global
63 #define IPV6_ADDR_ANY 0x0000U
65 #define IPV6_ADDR_UNICAST 0x0001U
66 #define IPV6_ADDR_MULTICAST 0x0002U
68 #define IPV6_ADDR_LOOPBACK 0x0010U
69 #define IPV6_ADDR_LINKLOCAL 0x0020U
70 #define IPV6_ADDR_SITELOCAL 0x0040U
72 #define IPV6_ADDR_COMPATv4 0x0080U
74 #define IPV6_ADDR_SCOPE_MASK 0x00f0U
76 #define IPV6_ADDR_MAPPED 0x1000U
81 #define IPV6_ADDR_MC_SCOPE(a) \
82 ((a)->s6_addr[1] & 0x0f) /* nonstandard */
83 #define __IPV6_ADDR_SCOPE_INVALID -1
84 #define IPV6_ADDR_SCOPE_NODELOCAL 0x01
85 #define IPV6_ADDR_SCOPE_LINKLOCAL 0x02
86 #define IPV6_ADDR_SCOPE_SITELOCAL 0x05
87 #define IPV6_ADDR_SCOPE_ORGLOCAL 0x08
88 #define IPV6_ADDR_SCOPE_GLOBAL 0x0e
93 #define IPV6_ADDR_MC_FLAG_TRANSIENT(a) \
94 ((a)->s6_addr[1] & 0x10)
95 #define IPV6_ADDR_MC_FLAG_PREFIX(a) \
96 ((a)->s6_addr[1] & 0x20)
97 #define IPV6_ADDR_MC_FLAG_RENDEZVOUS(a) \
98 ((a)->s6_addr[1] & 0x40)
101 * fragmentation header
108 __be32 identification;
111 #define IP6_MF 0x0001
113 #include <net/sock.h>
116 extern int sysctl_mld_max_msf;
117 extern struct ctl_path net_ipv6_ctl_path[];
119 #define _DEVINC(net, statname, modifier, idev, field) \
121 struct inet6_dev *_idev = (idev); \
122 if (likely(_idev != NULL)) \
123 SNMP_INC_STATS##modifier((_idev)->stats.statname, (field)); \
124 SNMP_INC_STATS##modifier((net)->mib.statname##_statistics, (field));\
127 /* per device counters are atomic_long_t */
128 #define _DEVINCATOMIC(net, statname, modifier, idev, field) \
130 struct inet6_dev *_idev = (idev); \
131 if (likely(_idev != NULL)) \
132 SNMP_INC_STATS_ATOMIC_LONG((_idev)->stats.statname##dev, (field)); \
133 SNMP_INC_STATS##modifier((net)->mib.statname##_statistics, (field));\
136 #define _DEVADD(net, statname, modifier, idev, field, val) \
138 struct inet6_dev *_idev = (idev); \
139 if (likely(_idev != NULL)) \
140 SNMP_ADD_STATS##modifier((_idev)->stats.statname, (field), (val)); \
141 SNMP_ADD_STATS##modifier((net)->mib.statname##_statistics, (field), (val));\
144 #define _DEVUPD(net, statname, modifier, idev, field, val) \
146 struct inet6_dev *_idev = (idev); \
147 if (likely(_idev != NULL)) \
148 SNMP_UPD_PO_STATS##modifier((_idev)->stats.statname, field, (val)); \
149 SNMP_UPD_PO_STATS##modifier((net)->mib.statname##_statistics, field, (val));\
154 #define IP6_INC_STATS(net, idev,field) \
155 _DEVINC(net, ipv6, 64, idev, field)
156 #define IP6_INC_STATS_BH(net, idev,field) \
157 _DEVINC(net, ipv6, 64_BH, idev, field)
158 #define IP6_ADD_STATS(net, idev,field,val) \
159 _DEVADD(net, ipv6, 64, idev, field, val)
160 #define IP6_ADD_STATS_BH(net, idev,field,val) \
161 _DEVADD(net, ipv6, 64_BH, idev, field, val)
162 #define IP6_UPD_PO_STATS(net, idev,field,val) \
163 _DEVUPD(net, ipv6, 64, idev, field, val)
164 #define IP6_UPD_PO_STATS_BH(net, idev,field,val) \
165 _DEVUPD(net, ipv6, 64_BH, idev, field, val)
166 #define ICMP6_INC_STATS(net, idev, field) \
167 _DEVINCATOMIC(net, icmpv6, , idev, field)
168 #define ICMP6_INC_STATS_BH(net, idev, field) \
169 _DEVINCATOMIC(net, icmpv6, _BH, idev, field)
171 #define ICMP6MSGOUT_INC_STATS(net, idev, field) \
172 _DEVINCATOMIC(net, icmpv6msg, , idev, field +256)
173 #define ICMP6MSGOUT_INC_STATS_BH(net, idev, field) \
174 _DEVINCATOMIC(net, icmpv6msg, _BH, idev, field +256)
175 #define ICMP6MSGIN_INC_STATS_BH(net, idev, field) \
176 _DEVINCATOMIC(net, icmpv6msg, _BH, idev, field)
178 struct ip6_ra_chain {
179 struct ip6_ra_chain *next;
182 void (*destructor)(struct sock *);
185 extern struct ip6_ra_chain *ip6_ra_chain;
186 extern rwlock_t ip6_ra_lock;
189 This structure is prepared by protocol, when parsing
190 ancillary data and passed to IPv6.
193 struct ipv6_txoptions {
195 /* Length of this structure */
198 /* length of extension headers */
200 __u16 opt_flen; /* after fragment hdr */
201 __u16 opt_nflen; /* before fragment hdr */
203 struct ipv6_opt_hdr *hopopt;
204 struct ipv6_opt_hdr *dst0opt;
205 struct ipv6_rt_hdr *srcrt; /* Routing Header */
206 struct ipv6_opt_hdr *dst1opt;
208 /* Option buffer, as read by IPV6_PKTOPTIONS, starts here. */
211 struct ip6_flowlabel {
212 struct ip6_flowlabel *next;
216 struct ipv6_txoptions *opt;
217 unsigned long linger;
220 unsigned long lastuse;
221 unsigned long expires;
225 #define IPV6_FLOWINFO_MASK cpu_to_be32(0x0FFFFFFF)
226 #define IPV6_FLOWLABEL_MASK cpu_to_be32(0x000FFFFF)
228 struct ipv6_fl_socklist {
229 struct ipv6_fl_socklist *next;
230 struct ip6_flowlabel *fl;
233 static inline struct ipv6_txoptions *txopt_get(const struct ipv6_pinfo *np)
235 struct ipv6_txoptions *opt;
238 opt = rcu_dereference(np->opt);
239 if (opt && !atomic_inc_not_zero(&opt->refcnt))
245 static inline void txopt_put(struct ipv6_txoptions *opt)
247 if (opt && atomic_dec_and_test(&opt->refcnt))
251 extern struct ip6_flowlabel *fl6_sock_lookup(struct sock *sk, __be32 label);
252 extern struct ipv6_txoptions *fl6_merge_options(struct ipv6_txoptions * opt_space,
253 struct ip6_flowlabel * fl,
254 struct ipv6_txoptions * fopt);
255 extern void fl6_free_socklist(struct sock *sk);
256 extern int ipv6_flowlabel_opt(struct sock *sk, char __user *optval, int optlen);
257 extern int ip6_flowlabel_init(void);
258 extern void ip6_flowlabel_cleanup(void);
260 static inline void fl6_sock_release(struct ip6_flowlabel *fl)
263 atomic_dec(&fl->users);
266 extern int ip6_ra_control(struct sock *sk, int sel);
268 extern int ipv6_parse_hopopts(struct sk_buff *skb);
270 extern struct ipv6_txoptions * ipv6_dup_options(struct sock *sk, struct ipv6_txoptions *opt);
271 extern struct ipv6_txoptions * ipv6_renew_options(struct sock *sk, struct ipv6_txoptions *opt,
273 struct ipv6_opt_hdr __user *newopt,
275 struct ipv6_txoptions *ipv6_fixup_options(struct ipv6_txoptions *opt_space,
276 struct ipv6_txoptions *opt);
278 extern int ipv6_opt_accepted(struct sock *sk, struct sk_buff *skb);
280 int ip6_frag_nqueues(struct net *net);
281 int ip6_frag_mem(struct net *net);
283 #define IPV6_FRAG_HIGH_THRESH (256 * 1024) /* 262144 */
284 #define IPV6_FRAG_LOW_THRESH (192 * 1024) /* 196608 */
285 #define IPV6_FRAG_TIMEOUT (60 * HZ) /* 60 seconds */
287 extern int __ipv6_addr_type(const struct in6_addr *addr);
288 static inline int ipv6_addr_type(const struct in6_addr *addr)
290 return __ipv6_addr_type(addr) & 0xffff;
293 static inline int ipv6_addr_scope(const struct in6_addr *addr)
295 return __ipv6_addr_type(addr) & IPV6_ADDR_SCOPE_MASK;
298 static inline int __ipv6_addr_src_scope(int type)
300 return (type == IPV6_ADDR_ANY) ? __IPV6_ADDR_SCOPE_INVALID : (type >> 16);
303 static inline int ipv6_addr_src_scope(const struct in6_addr *addr)
305 return __ipv6_addr_src_scope(__ipv6_addr_type(addr));
308 static inline int ipv6_addr_cmp(const struct in6_addr *a1, const struct in6_addr *a2)
310 return memcmp(a1, a2, sizeof(struct in6_addr));
314 ipv6_masked_addr_cmp(const struct in6_addr *a1, const struct in6_addr *m,
315 const struct in6_addr *a2)
317 return !!(((a1->s6_addr32[0] ^ a2->s6_addr32[0]) & m->s6_addr32[0]) |
318 ((a1->s6_addr32[1] ^ a2->s6_addr32[1]) & m->s6_addr32[1]) |
319 ((a1->s6_addr32[2] ^ a2->s6_addr32[2]) & m->s6_addr32[2]) |
320 ((a1->s6_addr32[3] ^ a2->s6_addr32[3]) & m->s6_addr32[3]));
323 static inline void ipv6_addr_copy(struct in6_addr *a1, const struct in6_addr *a2)
325 memcpy(a1, a2, sizeof(struct in6_addr));
328 static inline void ipv6_addr_prefix(struct in6_addr *pfx,
329 const struct in6_addr *addr,
332 /* caller must guarantee 0 <= plen <= 128 */
336 memset(pfx->s6_addr, 0, sizeof(pfx->s6_addr));
337 memcpy(pfx->s6_addr, addr, o);
339 pfx->s6_addr[o] = addr->s6_addr[o] & (0xff00 >> b);
342 static inline void ipv6_addr_set(struct in6_addr *addr,
343 __be32 w1, __be32 w2,
344 __be32 w3, __be32 w4)
346 addr->s6_addr32[0] = w1;
347 addr->s6_addr32[1] = w2;
348 addr->s6_addr32[2] = w3;
349 addr->s6_addr32[3] = w4;
352 static inline int ipv6_addr_equal(const struct in6_addr *a1,
353 const struct in6_addr *a2)
355 return ((a1->s6_addr32[0] ^ a2->s6_addr32[0]) |
356 (a1->s6_addr32[1] ^ a2->s6_addr32[1]) |
357 (a1->s6_addr32[2] ^ a2->s6_addr32[2]) |
358 (a1->s6_addr32[3] ^ a2->s6_addr32[3])) == 0;
361 static inline int __ipv6_prefix_equal(const __be32 *a1, const __be32 *a2,
362 unsigned int prefixlen)
366 /* check complete u32 in prefix */
367 pdw = prefixlen >> 5;
368 if (pdw && memcmp(a1, a2, pdw << 2))
371 /* check incomplete u32 in prefix */
372 pbi = prefixlen & 0x1f;
373 if (pbi && ((a1[pdw] ^ a2[pdw]) & htonl((0xffffffff) << (32 - pbi))))
379 static inline int ipv6_prefix_equal(const struct in6_addr *a1,
380 const struct in6_addr *a2,
381 unsigned int prefixlen)
383 return __ipv6_prefix_equal(a1->s6_addr32, a2->s6_addr32,
387 struct inet_frag_queue;
389 enum ip6_defrag_users {
390 IP6_DEFRAG_LOCAL_DELIVER,
391 IP6_DEFRAG_CONNTRACK_IN,
392 __IP6_DEFRAG_CONNTRACK_IN = IP6_DEFRAG_CONNTRACK_IN + USHRT_MAX,
393 IP6_DEFRAG_CONNTRACK_OUT,
394 __IP6_DEFRAG_CONNTRACK_OUT = IP6_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
395 IP6_DEFRAG_CONNTRACK_BRIDGE_IN,
396 __IP6_DEFRAG_CONNTRACK_BRIDGE_IN = IP6_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
399 struct ip6_create_arg {
402 const struct in6_addr *src;
403 const struct in6_addr *dst;
407 void ip6_frag_init(struct inet_frag_queue *q, void *a);
408 int ip6_frag_match(struct inet_frag_queue *q, void *a);
410 /* more secured version of ipv6_addr_hash() */
411 static inline u32 __ipv6_addr_jhash(const struct in6_addr *a, const u32 initval)
413 u32 v = (__force u32)a->s6_addr32[0] ^ (__force u32)a->s6_addr32[1];
415 return jhash_3words(v,
416 (__force u32)a->s6_addr32[2],
417 (__force u32)a->s6_addr32[3],
421 static inline u32 ipv6_addr_jhash(const struct in6_addr *a)
423 return __ipv6_addr_jhash(a, ipv6_hash_secret);
426 static inline int ipv6_addr_any(const struct in6_addr *a)
428 return (a->s6_addr32[0] | a->s6_addr32[1] |
429 a->s6_addr32[2] | a->s6_addr32[3]) == 0;
432 static inline int ipv6_addr_loopback(const struct in6_addr *a)
434 return (a->s6_addr32[0] | a->s6_addr32[1] |
435 a->s6_addr32[2] | (a->s6_addr32[3] ^ htonl(1))) == 0;
438 static inline int ipv6_addr_v4mapped(const struct in6_addr *a)
440 return (a->s6_addr32[0] | a->s6_addr32[1] |
441 (a->s6_addr32[2] ^ htonl(0x0000ffff))) == 0;
445 * Check for a RFC 4843 ORCHID address
446 * (Overlay Routable Cryptographic Hash Identifiers)
448 static inline int ipv6_addr_orchid(const struct in6_addr *a)
450 return (a->s6_addr32[0] & htonl(0xfffffff0)) == htonl(0x20010010);
453 static inline void ipv6_addr_set_v4mapped(const __be32 addr,
454 struct in6_addr *v4mapped)
456 ipv6_addr_set(v4mapped,
463 * find the first different bit between two addresses
464 * length of address must be a multiple of 32bits
466 static inline int __ipv6_addr_diff(const void *token1, const void *token2, int addrlen)
468 const __be32 *a1 = token1, *a2 = token2;
473 for (i = 0; i < addrlen; i++) {
474 __be32 xb = a1[i] ^ a2[i];
476 return i * 32 + 31 - __fls(ntohl(xb));
480 * we should *never* get to this point since that
481 * would mean the addrs are equal
483 * However, we do get to it 8) And exacly, when
484 * addresses are equal 8)
486 * ip route add 1111::/128 via ...
487 * ip route add 1111::/64 via ...
490 * Ideally, this function should stop comparison
491 * at prefix length. It does not, but it is still OK,
492 * if returned value is greater than prefix length.
498 static inline int ipv6_addr_diff(const struct in6_addr *a1, const struct in6_addr *a2)
500 return __ipv6_addr_diff(a1, a2, sizeof(struct in6_addr));
503 extern void ipv6_select_ident(struct frag_hdr *fhdr, struct rt6_info *rt);
504 void ipv6_proxy_select_ident(struct sk_buff *skb);
507 * Prototypes exported by ipv6
511 * rcv function (called from netdevice level)
514 extern int ipv6_rcv(struct sk_buff *skb,
515 struct net_device *dev,
516 struct packet_type *pt,
517 struct net_device *orig_dev);
519 extern int ip6_rcv_finish(struct sk_buff *skb);
522 * upper-layer output functions
524 extern int ip6_xmit(struct sock *sk,
527 struct ipv6_txoptions *opt,
530 extern int ip6_nd_hdr(struct sock *sk,
532 struct net_device *dev,
533 const struct in6_addr *saddr,
534 const struct in6_addr *daddr,
537 extern int ip6_find_1stfragopt(struct sk_buff *skb, u8 **nexthdr);
539 extern int ip6_append_data(struct sock *sk,
540 int getfrag(void *from, char *to, int offset, int len, int odd, struct sk_buff *skb),
546 struct ipv6_txoptions *opt,
552 extern int ip6_push_pending_frames(struct sock *sk);
554 extern void ip6_flush_pending_frames(struct sock *sk);
556 extern int ip6_dst_lookup(struct sock *sk,
557 struct dst_entry **dst,
559 extern struct dst_entry * ip6_dst_lookup_flow(struct sock *sk,
561 const struct in6_addr *final_dst,
563 extern struct dst_entry * ip6_sk_dst_lookup_flow(struct sock *sk,
565 const struct in6_addr *final_dst,
567 extern struct dst_entry * ip6_blackhole_route(struct net *net,
568 struct dst_entry *orig_dst);
571 * skb processing functions
574 extern int ip6_output(struct sk_buff *skb);
575 extern int ip6_forward(struct sk_buff *skb);
576 extern int ip6_input(struct sk_buff *skb);
577 extern int ip6_mc_input(struct sk_buff *skb);
579 extern int __ip6_local_out(struct sk_buff *skb);
580 extern int ip6_local_out(struct sk_buff *skb);
583 * Extension header (options) processing
586 extern void ipv6_push_nfrag_opts(struct sk_buff *skb,
587 struct ipv6_txoptions *opt,
589 struct in6_addr **daddr_p);
590 extern void ipv6_push_frag_opts(struct sk_buff *skb,
591 struct ipv6_txoptions *opt,
594 extern int ipv6_skip_exthdr(const struct sk_buff *, int start,
597 extern int ipv6_ext_hdr(u8 nexthdr);
599 extern int ipv6_find_tlv(struct sk_buff *skb, int offset, int type);
601 extern struct in6_addr *fl6_update_dst(struct flowi6 *fl6,
602 const struct ipv6_txoptions *opt,
603 struct in6_addr *orig);
606 * socket options (ipv6_sockglue.c)
609 extern int ipv6_setsockopt(struct sock *sk, int level,
612 unsigned int optlen);
613 extern int ipv6_getsockopt(struct sock *sk, int level,
617 extern int compat_ipv6_setsockopt(struct sock *sk,
621 unsigned int optlen);
622 extern int compat_ipv6_getsockopt(struct sock *sk,
628 extern int ip6_datagram_connect(struct sock *sk,
629 struct sockaddr *addr, int addr_len);
631 extern int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len,
633 extern int ipv6_recv_rxpmtu(struct sock *sk, struct msghdr *msg, int len,
635 extern void ipv6_icmp_error(struct sock *sk, struct sk_buff *skb, int err, __be16 port,
636 u32 info, u8 *payload);
637 extern void ipv6_local_error(struct sock *sk, int err, struct flowi6 *fl6, u32 info);
638 extern void ipv6_local_rxpmtu(struct sock *sk, struct flowi6 *fl6, u32 mtu);
640 extern int inet6_release(struct socket *sock);
641 extern int inet6_bind(struct socket *sock, struct sockaddr *uaddr,
643 extern int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
644 int *uaddr_len, int peer);
645 extern int inet6_ioctl(struct socket *sock, unsigned int cmd,
648 extern int inet6_hash_connect(struct inet_timewait_death_row *death_row,
654 extern const struct proto_ops inet6_stream_ops;
655 extern const struct proto_ops inet6_dgram_ops;
657 struct group_source_req;
660 extern int ip6_mc_source(int add, int omode, struct sock *sk,
661 struct group_source_req *pgsr);
662 extern int ip6_mc_msfilter(struct sock *sk, struct group_filter *gsf);
663 extern int ip6_mc_msfget(struct sock *sk, struct group_filter *gsf,
664 struct group_filter __user *optval,
666 extern unsigned int inet6_hash_frag(__be32 id, const struct in6_addr *saddr,
667 const struct in6_addr *daddr, u32 rnd);
669 #ifdef CONFIG_PROC_FS
670 extern int ac6_proc_init(struct net *net);
671 extern void ac6_proc_exit(struct net *net);
672 extern int raw6_proc_init(void);
673 extern void raw6_proc_exit(void);
674 extern int tcp6_proc_init(struct net *net);
675 extern void tcp6_proc_exit(struct net *net);
676 extern int udp6_proc_init(struct net *net);
677 extern void udp6_proc_exit(struct net *net);
678 extern int udplite6_proc_init(void);
679 extern void udplite6_proc_exit(void);
680 extern int ipv6_misc_proc_init(void);
681 extern void ipv6_misc_proc_exit(void);
682 extern int snmp6_register_dev(struct inet6_dev *idev);
683 extern int snmp6_unregister_dev(struct inet6_dev *idev);
686 static inline int ac6_proc_init(struct net *net) { return 0; }
687 static inline void ac6_proc_exit(struct net *net) { }
688 static inline int snmp6_register_dev(struct inet6_dev *idev) { return 0; }
689 static inline int snmp6_unregister_dev(struct inet6_dev *idev) { return 0; }
693 extern ctl_table ipv6_route_table_template[];
694 extern ctl_table ipv6_icmp_table_template[];
696 extern struct ctl_table *ipv6_icmp_sysctl_init(struct net *net);
697 extern struct ctl_table *ipv6_route_sysctl_init(struct net *net);
698 extern int ipv6_sysctl_register(void);
699 extern void ipv6_sysctl_unregister(void);
700 extern int ipv6_static_sysctl_register(void);
701 extern void ipv6_static_sysctl_unregister(void);
704 #endif /* _NET_IPV6_H */