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 <net/if_inet6.h>
19 #include <net/ndisc.h>
23 #define SIN6_LEN_RFC2133 24
25 #define IPV6_MAXPLEN 65535
28 * NextHeader field of IPv6 header
31 #define NEXTHDR_HOP 0 /* Hop-by-hop option header. */
32 #define NEXTHDR_TCP 6 /* TCP segment. */
33 #define NEXTHDR_UDP 17 /* UDP message. */
34 #define NEXTHDR_IPV6 41 /* IPv6 in IPv6 */
35 #define NEXTHDR_ROUTING 43 /* Routing header. */
36 #define NEXTHDR_FRAGMENT 44 /* Fragmentation/reassembly header. */
37 #define NEXTHDR_ESP 50 /* Encapsulating security payload. */
38 #define NEXTHDR_AUTH 51 /* Authentication header. */
39 #define NEXTHDR_ICMP 58 /* ICMP for IPv6. */
40 #define NEXTHDR_NONE 59 /* No next header */
41 #define NEXTHDR_DEST 60 /* Destination options header. */
42 #define NEXTHDR_MOBILITY 135 /* Mobility header. */
44 #define NEXTHDR_MAX 255
48 #define IPV6_DEFAULT_HOPLIMIT 64
49 #define IPV6_DEFAULT_MCASTHOPS 1
54 * type - unicast | multicast
55 * scope - local | site | global
62 #define IPV6_ADDR_ANY 0x0000U
64 #define IPV6_ADDR_UNICAST 0x0001U
65 #define IPV6_ADDR_MULTICAST 0x0002U
67 #define IPV6_ADDR_LOOPBACK 0x0010U
68 #define IPV6_ADDR_LINKLOCAL 0x0020U
69 #define IPV6_ADDR_SITELOCAL 0x0040U
71 #define IPV6_ADDR_COMPATv4 0x0080U
73 #define IPV6_ADDR_SCOPE_MASK 0x00f0U
75 #define IPV6_ADDR_MAPPED 0x1000U
80 #define IPV6_ADDR_MC_SCOPE(a) \
81 ((a)->s6_addr[1] & 0x0f) /* nonstandard */
82 #define __IPV6_ADDR_SCOPE_INVALID -1
83 #define IPV6_ADDR_SCOPE_NODELOCAL 0x01
84 #define IPV6_ADDR_SCOPE_LINKLOCAL 0x02
85 #define IPV6_ADDR_SCOPE_SITELOCAL 0x05
86 #define IPV6_ADDR_SCOPE_ORGLOCAL 0x08
87 #define IPV6_ADDR_SCOPE_GLOBAL 0x0e
92 #define IPV6_ADDR_MC_FLAG_TRANSIENT(a) \
93 ((a)->s6_addr[1] & 0x10)
94 #define IPV6_ADDR_MC_FLAG_PREFIX(a) \
95 ((a)->s6_addr[1] & 0x20)
96 #define IPV6_ADDR_MC_FLAG_RENDEZVOUS(a) \
97 ((a)->s6_addr[1] & 0x40)
100 * fragmentation header
107 __be32 identification
;
110 #define IP6_MF 0x0001
112 #include <net/sock.h>
115 extern int sysctl_mld_max_msf
;
117 #define _DEVINC(net, statname, modifier, idev, field) \
119 struct inet6_dev *_idev = (idev); \
120 if (likely(_idev != NULL)) \
121 SNMP_INC_STATS##modifier((_idev)->stats.statname, (field)); \
122 SNMP_INC_STATS##modifier((net)->mib.statname##_statistics, (field));\
125 /* per device counters are atomic_long_t */
126 #define _DEVINCATOMIC(net, statname, modifier, idev, field) \
128 struct inet6_dev *_idev = (idev); \
129 if (likely(_idev != NULL)) \
130 SNMP_INC_STATS_ATOMIC_LONG((_idev)->stats.statname##dev, (field)); \
131 SNMP_INC_STATS##modifier((net)->mib.statname##_statistics, (field));\
134 /* per device and per net counters are atomic_long_t */
135 #define _DEVINC_ATOMIC_ATOMIC(net, statname, idev, field) \
137 struct inet6_dev *_idev = (idev); \
138 if (likely(_idev != NULL)) \
139 SNMP_INC_STATS_ATOMIC_LONG((_idev)->stats.statname##dev, (field)); \
140 SNMP_INC_STATS_ATOMIC_LONG((net)->mib.statname##_statistics, (field));\
143 #define _DEVADD(net, statname, modifier, idev, field, val) \
145 struct inet6_dev *_idev = (idev); \
146 if (likely(_idev != NULL)) \
147 SNMP_ADD_STATS##modifier((_idev)->stats.statname, (field), (val)); \
148 SNMP_ADD_STATS##modifier((net)->mib.statname##_statistics, (field), (val));\
151 #define _DEVUPD(net, statname, modifier, idev, field, val) \
153 struct inet6_dev *_idev = (idev); \
154 if (likely(_idev != NULL)) \
155 SNMP_UPD_PO_STATS##modifier((_idev)->stats.statname, field, (val)); \
156 SNMP_UPD_PO_STATS##modifier((net)->mib.statname##_statistics, field, (val));\
161 #define IP6_INC_STATS(net, idev,field) \
162 _DEVINC(net, ipv6, 64, idev, field)
163 #define IP6_INC_STATS_BH(net, idev,field) \
164 _DEVINC(net, ipv6, 64_BH, idev, field)
165 #define IP6_ADD_STATS(net, idev,field,val) \
166 _DEVADD(net, ipv6, 64, idev, field, val)
167 #define IP6_ADD_STATS_BH(net, idev,field,val) \
168 _DEVADD(net, ipv6, 64_BH, idev, field, val)
169 #define IP6_UPD_PO_STATS(net, idev,field,val) \
170 _DEVUPD(net, ipv6, 64, idev, field, val)
171 #define IP6_UPD_PO_STATS_BH(net, idev,field,val) \
172 _DEVUPD(net, ipv6, 64_BH, idev, field, val)
173 #define ICMP6_INC_STATS(net, idev, field) \
174 _DEVINCATOMIC(net, icmpv6, , idev, field)
175 #define ICMP6_INC_STATS_BH(net, idev, field) \
176 _DEVINCATOMIC(net, icmpv6, _BH, idev, field)
178 #define ICMP6MSGOUT_INC_STATS(net, idev, field) \
179 _DEVINC_ATOMIC_ATOMIC(net, icmpv6msg, idev, field +256)
180 #define ICMP6MSGOUT_INC_STATS_BH(net, idev, field) \
181 _DEVINC_ATOMIC_ATOMIC(net, icmpv6msg, idev, field +256)
182 #define ICMP6MSGIN_INC_STATS_BH(net, idev, field) \
183 _DEVINC_ATOMIC_ATOMIC(net, icmpv6msg, idev, field)
185 struct ip6_ra_chain
{
186 struct ip6_ra_chain
*next
;
189 void (*destructor
)(struct sock
*);
192 extern struct ip6_ra_chain
*ip6_ra_chain
;
193 extern rwlock_t ip6_ra_lock
;
196 This structure is prepared by protocol, when parsing
197 ancillary data and passed to IPv6.
200 struct ipv6_txoptions
{
201 /* Length of this structure */
204 /* length of extension headers */
206 __u16 opt_flen
; /* after fragment hdr */
207 __u16 opt_nflen
; /* before fragment hdr */
209 struct ipv6_opt_hdr
*hopopt
;
210 struct ipv6_opt_hdr
*dst0opt
;
211 struct ipv6_rt_hdr
*srcrt
; /* Routing Header */
212 struct ipv6_opt_hdr
*dst1opt
;
214 /* Option buffer, as read by IPV6_PKTOPTIONS, starts here. */
217 struct ip6_flowlabel
{
218 struct ip6_flowlabel
*next
;
222 struct ipv6_txoptions
*opt
;
223 unsigned long linger
;
226 unsigned long lastuse
;
227 unsigned long expires
;
231 #define IPV6_FLOWINFO_MASK cpu_to_be32(0x0FFFFFFF)
232 #define IPV6_FLOWLABEL_MASK cpu_to_be32(0x000FFFFF)
234 struct ipv6_fl_socklist
{
235 struct ipv6_fl_socklist
*next
;
236 struct ip6_flowlabel
*fl
;
239 extern struct ip6_flowlabel
*fl6_sock_lookup(struct sock
*sk
, __be32 label
);
240 extern struct ipv6_txoptions
*fl6_merge_options(struct ipv6_txoptions
* opt_space
,
241 struct ip6_flowlabel
* fl
,
242 struct ipv6_txoptions
* fopt
);
243 extern void fl6_free_socklist(struct sock
*sk
);
244 extern int ipv6_flowlabel_opt(struct sock
*sk
, char __user
*optval
, int optlen
);
245 extern int ip6_flowlabel_init(void);
246 extern void ip6_flowlabel_cleanup(void);
248 static inline void fl6_sock_release(struct ip6_flowlabel
*fl
)
251 atomic_dec(&fl
->users
);
254 extern int ip6_ra_control(struct sock
*sk
, int sel
);
256 extern int ipv6_parse_hopopts(struct sk_buff
*skb
);
258 extern struct ipv6_txoptions
* ipv6_dup_options(struct sock
*sk
, struct ipv6_txoptions
*opt
);
259 extern struct ipv6_txoptions
* ipv6_renew_options(struct sock
*sk
, struct ipv6_txoptions
*opt
,
261 struct ipv6_opt_hdr __user
*newopt
,
263 struct ipv6_txoptions
*ipv6_fixup_options(struct ipv6_txoptions
*opt_space
,
264 struct ipv6_txoptions
*opt
);
266 extern int ipv6_opt_accepted(struct sock
*sk
, struct sk_buff
*skb
);
268 int ip6_frag_nqueues(struct net
*net
);
269 int ip6_frag_mem(struct net
*net
);
271 #define IPV6_FRAG_HIGH_THRESH (256 * 1024) /* 262144 */
272 #define IPV6_FRAG_LOW_THRESH (192 * 1024) /* 196608 */
273 #define IPV6_FRAG_TIMEOUT (60 * HZ) /* 60 seconds */
275 extern int __ipv6_addr_type(const struct in6_addr
*addr
);
276 static inline int ipv6_addr_type(const struct in6_addr
*addr
)
278 return __ipv6_addr_type(addr
) & 0xffff;
281 static inline int ipv6_addr_scope(const struct in6_addr
*addr
)
283 return __ipv6_addr_type(addr
) & IPV6_ADDR_SCOPE_MASK
;
286 static inline int __ipv6_addr_src_scope(int type
)
288 return (type
== IPV6_ADDR_ANY
) ? __IPV6_ADDR_SCOPE_INVALID
: (type
>> 16);
291 static inline int ipv6_addr_src_scope(const struct in6_addr
*addr
)
293 return __ipv6_addr_src_scope(__ipv6_addr_type(addr
));
296 static inline int ipv6_addr_cmp(const struct in6_addr
*a1
, const struct in6_addr
*a2
)
298 return memcmp(a1
, a2
, sizeof(struct in6_addr
));
302 ipv6_masked_addr_cmp(const struct in6_addr
*a1
, const struct in6_addr
*m
,
303 const struct in6_addr
*a2
)
305 return !!(((a1
->s6_addr32
[0] ^ a2
->s6_addr32
[0]) & m
->s6_addr32
[0]) |
306 ((a1
->s6_addr32
[1] ^ a2
->s6_addr32
[1]) & m
->s6_addr32
[1]) |
307 ((a1
->s6_addr32
[2] ^ a2
->s6_addr32
[2]) & m
->s6_addr32
[2]) |
308 ((a1
->s6_addr32
[3] ^ a2
->s6_addr32
[3]) & m
->s6_addr32
[3]));
311 static inline void ipv6_addr_prefix(struct in6_addr
*pfx
,
312 const struct in6_addr
*addr
,
315 /* caller must guarantee 0 <= plen <= 128 */
319 memset(pfx
->s6_addr
, 0, sizeof(pfx
->s6_addr
));
320 memcpy(pfx
->s6_addr
, addr
, o
);
322 pfx
->s6_addr
[o
] = addr
->s6_addr
[o
] & (0xff00 >> b
);
325 static inline void ipv6_addr_set(struct in6_addr
*addr
,
326 __be32 w1
, __be32 w2
,
327 __be32 w3
, __be32 w4
)
329 addr
->s6_addr32
[0] = w1
;
330 addr
->s6_addr32
[1] = w2
;
331 addr
->s6_addr32
[2] = w3
;
332 addr
->s6_addr32
[3] = w4
;
335 static inline int ipv6_addr_equal(const struct in6_addr
*a1
,
336 const struct in6_addr
*a2
)
338 return ((a1
->s6_addr32
[0] ^ a2
->s6_addr32
[0]) |
339 (a1
->s6_addr32
[1] ^ a2
->s6_addr32
[1]) |
340 (a1
->s6_addr32
[2] ^ a2
->s6_addr32
[2]) |
341 (a1
->s6_addr32
[3] ^ a2
->s6_addr32
[3])) == 0;
344 static inline int __ipv6_prefix_equal(const __be32
*a1
, const __be32
*a2
,
345 unsigned int prefixlen
)
347 unsigned int pdw
, pbi
;
349 /* check complete u32 in prefix */
350 pdw
= prefixlen
>> 5;
351 if (pdw
&& memcmp(a1
, a2
, pdw
<< 2))
354 /* check incomplete u32 in prefix */
355 pbi
= prefixlen
& 0x1f;
356 if (pbi
&& ((a1
[pdw
] ^ a2
[pdw
]) & htonl((0xffffffff) << (32 - pbi
))))
362 static inline int ipv6_prefix_equal(const struct in6_addr
*a1
,
363 const struct in6_addr
*a2
,
364 unsigned int prefixlen
)
366 return __ipv6_prefix_equal(a1
->s6_addr32
, a2
->s6_addr32
,
370 struct inet_frag_queue
;
372 enum ip6_defrag_users
{
373 IP6_DEFRAG_LOCAL_DELIVER
,
374 IP6_DEFRAG_CONNTRACK_IN
,
375 __IP6_DEFRAG_CONNTRACK_IN
= IP6_DEFRAG_CONNTRACK_IN
+ USHRT_MAX
,
376 IP6_DEFRAG_CONNTRACK_OUT
,
377 __IP6_DEFRAG_CONNTRACK_OUT
= IP6_DEFRAG_CONNTRACK_OUT
+ USHRT_MAX
,
378 IP6_DEFRAG_CONNTRACK_BRIDGE_IN
,
379 __IP6_DEFRAG_CONNTRACK_BRIDGE_IN
= IP6_DEFRAG_CONNTRACK_BRIDGE_IN
+ USHRT_MAX
,
382 struct ip6_create_arg
{
385 const struct in6_addr
*src
;
386 const struct in6_addr
*dst
;
389 void ip6_frag_init(struct inet_frag_queue
*q
, void *a
);
390 int ip6_frag_match(struct inet_frag_queue
*q
, void *a
);
392 static inline int ipv6_addr_any(const struct in6_addr
*a
)
394 return (a
->s6_addr32
[0] | a
->s6_addr32
[1] |
395 a
->s6_addr32
[2] | a
->s6_addr32
[3]) == 0;
398 static inline int ipv6_addr_loopback(const struct in6_addr
*a
)
400 return (a
->s6_addr32
[0] | a
->s6_addr32
[1] |
401 a
->s6_addr32
[2] | (a
->s6_addr32
[3] ^ htonl(1))) == 0;
404 static inline int ipv6_addr_v4mapped(const struct in6_addr
*a
)
406 return (a
->s6_addr32
[0] | a
->s6_addr32
[1] |
407 (a
->s6_addr32
[2] ^ htonl(0x0000ffff))) == 0;
411 * Check for a RFC 4843 ORCHID address
412 * (Overlay Routable Cryptographic Hash Identifiers)
414 static inline int ipv6_addr_orchid(const struct in6_addr
*a
)
416 return (a
->s6_addr32
[0] & htonl(0xfffffff0)) == htonl(0x20010010);
419 static inline void ipv6_addr_set_v4mapped(const __be32 addr
,
420 struct in6_addr
*v4mapped
)
422 ipv6_addr_set(v4mapped
,
429 * find the first different bit between two addresses
430 * length of address must be a multiple of 32bits
432 static inline int __ipv6_addr_diff(const void *token1
, const void *token2
, int addrlen
)
434 const __be32
*a1
= token1
, *a2
= token2
;
439 for (i
= 0; i
< addrlen
; i
++) {
440 __be32 xb
= a1
[i
] ^ a2
[i
];
442 return i
* 32 + 31 - __fls(ntohl(xb
));
446 * we should *never* get to this point since that
447 * would mean the addrs are equal
449 * However, we do get to it 8) And exacly, when
450 * addresses are equal 8)
452 * ip route add 1111::/128 via ...
453 * ip route add 1111::/64 via ...
456 * Ideally, this function should stop comparison
457 * at prefix length. It does not, but it is still OK,
458 * if returned value is greater than prefix length.
464 static inline int ipv6_addr_diff(const struct in6_addr
*a1
, const struct in6_addr
*a2
)
466 return __ipv6_addr_diff(a1
, a2
, sizeof(struct in6_addr
));
469 extern void ipv6_select_ident(struct frag_hdr
*fhdr
, struct rt6_info
*rt
);
472 * Prototypes exported by ipv6
476 * rcv function (called from netdevice level)
479 extern int ipv6_rcv(struct sk_buff
*skb
,
480 struct net_device
*dev
,
481 struct packet_type
*pt
,
482 struct net_device
*orig_dev
);
484 extern int ip6_rcv_finish(struct sk_buff
*skb
);
487 * upper-layer output functions
489 extern int ip6_xmit(struct sock
*sk
,
492 struct ipv6_txoptions
*opt
,
495 extern int ip6_nd_hdr(struct sock
*sk
,
497 struct net_device
*dev
,
498 const struct in6_addr
*saddr
,
499 const struct in6_addr
*daddr
,
502 extern int ip6_find_1stfragopt(struct sk_buff
*skb
, u8
**nexthdr
);
504 extern int ip6_append_data(struct sock
*sk
,
505 int getfrag(void *from
, char *to
, int offset
, int len
, int odd
, struct sk_buff
*skb
),
511 struct ipv6_txoptions
*opt
,
517 extern int ip6_push_pending_frames(struct sock
*sk
);
519 extern void ip6_flush_pending_frames(struct sock
*sk
);
521 extern int ip6_dst_lookup(struct sock
*sk
,
522 struct dst_entry
**dst
,
524 extern struct dst_entry
* ip6_dst_lookup_flow(struct sock
*sk
,
526 const struct in6_addr
*final_dst
,
528 extern struct dst_entry
* ip6_sk_dst_lookup_flow(struct sock
*sk
,
530 const struct in6_addr
*final_dst
,
532 extern struct dst_entry
* ip6_blackhole_route(struct net
*net
,
533 struct dst_entry
*orig_dst
);
536 * skb processing functions
539 extern int ip6_output(struct sk_buff
*skb
);
540 extern int ip6_forward(struct sk_buff
*skb
);
541 extern int ip6_input(struct sk_buff
*skb
);
542 extern int ip6_mc_input(struct sk_buff
*skb
);
544 extern int __ip6_local_out(struct sk_buff
*skb
);
545 extern int ip6_local_out(struct sk_buff
*skb
);
548 * Extension header (options) processing
551 extern void ipv6_push_nfrag_opts(struct sk_buff
*skb
,
552 struct ipv6_txoptions
*opt
,
554 struct in6_addr
**daddr_p
);
555 extern void ipv6_push_frag_opts(struct sk_buff
*skb
,
556 struct ipv6_txoptions
*opt
,
559 extern int ipv6_skip_exthdr(const struct sk_buff
*, int start
,
560 u8
*nexthdrp
, __be16
*frag_offp
);
562 extern int ipv6_ext_hdr(u8 nexthdr
);
564 extern int ipv6_find_tlv(struct sk_buff
*skb
, int offset
, int type
);
566 extern struct in6_addr
*fl6_update_dst(struct flowi6
*fl6
,
567 const struct ipv6_txoptions
*opt
,
568 struct in6_addr
*orig
);
571 * socket options (ipv6_sockglue.c)
574 extern int ipv6_setsockopt(struct sock
*sk
, int level
,
577 unsigned int optlen
);
578 extern int ipv6_getsockopt(struct sock
*sk
, int level
,
582 extern int compat_ipv6_setsockopt(struct sock
*sk
,
586 unsigned int optlen
);
587 extern int compat_ipv6_getsockopt(struct sock
*sk
,
593 extern int ip6_datagram_connect(struct sock
*sk
,
594 struct sockaddr
*addr
, int addr_len
);
596 extern int ipv6_recv_error(struct sock
*sk
, struct msghdr
*msg
, int len
);
597 extern int ipv6_recv_rxpmtu(struct sock
*sk
, struct msghdr
*msg
, int len
);
598 extern void ipv6_icmp_error(struct sock
*sk
, struct sk_buff
*skb
, int err
, __be16 port
,
599 u32 info
, u8
*payload
);
600 extern void ipv6_local_error(struct sock
*sk
, int err
, struct flowi6
*fl6
, u32 info
);
601 extern void ipv6_local_rxpmtu(struct sock
*sk
, struct flowi6
*fl6
, u32 mtu
);
603 extern int inet6_release(struct socket
*sock
);
604 extern int inet6_bind(struct socket
*sock
, struct sockaddr
*uaddr
,
606 extern int inet6_getname(struct socket
*sock
, struct sockaddr
*uaddr
,
607 int *uaddr_len
, int peer
);
608 extern int inet6_ioctl(struct socket
*sock
, unsigned int cmd
,
611 extern int inet6_hash_connect(struct inet_timewait_death_row
*death_row
,
617 extern const struct proto_ops inet6_stream_ops
;
618 extern const struct proto_ops inet6_dgram_ops
;
620 struct group_source_req
;
623 extern int ip6_mc_source(int add
, int omode
, struct sock
*sk
,
624 struct group_source_req
*pgsr
);
625 extern int ip6_mc_msfilter(struct sock
*sk
, struct group_filter
*gsf
);
626 extern int ip6_mc_msfget(struct sock
*sk
, struct group_filter
*gsf
,
627 struct group_filter __user
*optval
,
629 extern unsigned int inet6_hash_frag(__be32 id
, const struct in6_addr
*saddr
,
630 const struct in6_addr
*daddr
, u32 rnd
);
632 #ifdef CONFIG_PROC_FS
633 extern int ac6_proc_init(struct net
*net
);
634 extern void ac6_proc_exit(struct net
*net
);
635 extern int raw6_proc_init(void);
636 extern void raw6_proc_exit(void);
637 extern int tcp6_proc_init(struct net
*net
);
638 extern void tcp6_proc_exit(struct net
*net
);
639 extern int udp6_proc_init(struct net
*net
);
640 extern void udp6_proc_exit(struct net
*net
);
641 extern int udplite6_proc_init(void);
642 extern void udplite6_proc_exit(void);
643 extern int ipv6_misc_proc_init(void);
644 extern void ipv6_misc_proc_exit(void);
645 extern int snmp6_register_dev(struct inet6_dev
*idev
);
646 extern int snmp6_unregister_dev(struct inet6_dev
*idev
);
649 static inline int ac6_proc_init(struct net
*net
) { return 0; }
650 static inline void ac6_proc_exit(struct net
*net
) { }
651 static inline int snmp6_register_dev(struct inet6_dev
*idev
) { return 0; }
652 static inline int snmp6_unregister_dev(struct inet6_dev
*idev
) { return 0; }
656 extern ctl_table ipv6_route_table_template
[];
657 extern ctl_table ipv6_icmp_table_template
[];
659 extern struct ctl_table
*ipv6_icmp_sysctl_init(struct net
*net
);
660 extern struct ctl_table
*ipv6_route_sysctl_init(struct net
*net
);
661 extern int ipv6_sysctl_register(void);
662 extern void ipv6_sysctl_unregister(void);
665 #endif /* _NET_IPV6_H */