in_var.h revision 222143
12Srgrimes/*-
22Srgrimes * Copyright (c) 1985, 1986, 1993
32Srgrimes *	The Regents of the University of California.  All rights reserved.
42Srgrimes *
52Srgrimes * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 * 4. Neither the name of the University nor the names of its contributors
14 *    may be used to endorse or promote products derived from this software
15 *    without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 *
29 *	@(#)in_var.h	8.2 (Berkeley) 1/9/95
30 * $FreeBSD: head/sys/netinet/in_var.h 222143 2011-05-20 19:12:20Z qingli $
31 */
32
33#ifndef _NETINET_IN_VAR_H_
34#define _NETINET_IN_VAR_H_
35
36#include <sys/queue.h>
37#include <sys/fnv_hash.h>
38#include <sys/tree.h>
39
40struct igmp_ifinfo;
41struct in_multi;
42struct lltable;
43
44/*
45 * IPv4 per-interface state.
46 */
47struct in_ifinfo {
48	struct lltable		*ii_llt;	/* ARP state */
49	struct igmp_ifinfo	*ii_igmp;	/* IGMP state */
50	struct in_multi		*ii_allhosts;	/* 224.0.0.1 membership */
51};
52
53/*
54 * Interface address, Internet version.  One of these structures
55 * is allocated for each Internet address on an interface.
56 * The ifaddr structure contains the protocol-independent part
57 * of the structure and is assumed to be first.
58 */
59struct in_ifaddr {
60	struct	ifaddr ia_ifa;		/* protocol-independent info */
61#define	ia_ifp		ia_ifa.ifa_ifp
62#define ia_flags	ia_ifa.ifa_flags
63					/* ia_{,sub}net{,mask} in host order */
64	u_long	ia_net;			/* network number of interface */
65	u_long	ia_netmask;		/* mask of net part */
66	u_long	ia_subnet;		/* subnet number, including net */
67	u_long	ia_subnetmask;		/* mask of subnet part */
68	struct	in_addr ia_netbroadcast; /* to recognize net broadcasts */
69	LIST_ENTRY(in_ifaddr) ia_hash;	/* entry in bucket of inet addresses */
70	TAILQ_ENTRY(in_ifaddr) ia_link;	/* list of internet addresses */
71	struct	sockaddr_in ia_addr;	/* reserve space for interface name */
72	struct	sockaddr_in ia_dstaddr; /* reserve space for broadcast addr */
73#define	ia_broadaddr	ia_dstaddr
74	struct	sockaddr_in ia_sockmask; /* reserve space for general netmask */
75};
76
77struct	in_aliasreq {
78	char	ifra_name[IFNAMSIZ];		/* if name, e.g. "en0" */
79	struct	sockaddr_in ifra_addr;
80	struct	sockaddr_in ifra_broadaddr;
81#define ifra_dstaddr ifra_broadaddr
82	struct	sockaddr_in ifra_mask;
83};
84/*
85 * Given a pointer to an in_ifaddr (ifaddr),
86 * return a pointer to the addr as a sockaddr_in.
87 */
88#define IA_SIN(ia)    (&(((struct in_ifaddr *)(ia))->ia_addr))
89#define IA_DSTSIN(ia) (&(((struct in_ifaddr *)(ia))->ia_dstaddr))
90
91#define IN_LNAOF(in, ifa) \
92	((ntohl((in).s_addr) & ~((struct in_ifaddr *)(ifa)->ia_subnetmask))
93
94
95#ifdef	_KERNEL
96extern	u_char	inetctlerrmap[];
97
98#define LLTABLE(ifp)	\
99	((struct in_ifinfo *)(ifp)->if_afdata[AF_INET])->ii_llt
100/*
101 * Hash table for IP addresses.
102 */
103TAILQ_HEAD(in_ifaddrhead, in_ifaddr);
104LIST_HEAD(in_ifaddrhashhead, in_ifaddr);
105
106VNET_DECLARE(struct in_ifaddrhashhead *, in_ifaddrhashtbl);
107VNET_DECLARE(struct in_ifaddrhead, in_ifaddrhead);
108VNET_DECLARE(u_long, in_ifaddrhmask);		/* mask for hash table */
109
110#define	V_in_ifaddrhashtbl	VNET(in_ifaddrhashtbl)
111#define	V_in_ifaddrhead		VNET(in_ifaddrhead)
112#define	V_in_ifaddrhmask	VNET(in_ifaddrhmask)
113
114#define INADDR_NHASH_LOG2       9
115#define INADDR_NHASH		(1 << INADDR_NHASH_LOG2)
116#define INADDR_HASHVAL(x)	fnv_32_buf((&(x)), sizeof(x), FNV1_32_INIT)
117#define INADDR_HASH(x) \
118	(&V_in_ifaddrhashtbl[INADDR_HASHVAL(x) & V_in_ifaddrhmask])
119
120extern	struct rwlock in_ifaddr_lock;
121
122#define	IN_IFADDR_LOCK_ASSERT()	rw_assert(&in_ifaddr_lock, RA_LOCKED)
123#define	IN_IFADDR_RLOCK()	rw_rlock(&in_ifaddr_lock)
124#define	IN_IFADDR_RLOCK_ASSERT()	rw_assert(&in_ifaddr_lock, RA_RLOCKED)
125#define	IN_IFADDR_RUNLOCK()	rw_runlock(&in_ifaddr_lock)
126#define	IN_IFADDR_WLOCK()	rw_wlock(&in_ifaddr_lock)
127#define	IN_IFADDR_WLOCK_ASSERT()	rw_assert(&in_ifaddr_lock, RA_WLOCKED)
128#define	IN_IFADDR_WUNLOCK()	rw_wunlock(&in_ifaddr_lock)
129
130/*
131 * Macro for finding the internet address structure (in_ifaddr)
132 * corresponding to one of our IP addresses (in_addr).
133 */
134#define INADDR_TO_IFADDR(addr, ia) \
135	/* struct in_addr addr; */ \
136	/* struct in_ifaddr *ia; */ \
137do { \
138\
139	LIST_FOREACH(ia, INADDR_HASH((addr).s_addr), ia_hash) \
140		if (IA_SIN(ia)->sin_addr.s_addr == (addr).s_addr) \
141			break; \
142} while (0)
143
144/*
145 * Macro for finding the interface (ifnet structure) corresponding to one
146 * of our IP addresses.
147 */
148#define INADDR_TO_IFP(addr, ifp) \
149	/* struct in_addr addr; */ \
150	/* struct ifnet *ifp; */ \
151{ \
152	struct in_ifaddr *ia; \
153\
154	INADDR_TO_IFADDR(addr, ia); \
155	(ifp) = (ia == NULL) ? NULL : ia->ia_ifp; \
156}
157
158/*
159 * Macro for finding the internet address structure (in_ifaddr) corresponding
160 * to a given interface (ifnet structure).
161 */
162#define IFP_TO_IA(ifp, ia)						\
163	/* struct ifnet *ifp; */					\
164	/* struct in_ifaddr *ia; */					\
165{									\
166	for ((ia) = TAILQ_FIRST(&V_in_ifaddrhead);			\
167	    (ia) != NULL && (ia)->ia_ifp != (ifp);			\
168	    (ia) = TAILQ_NEXT((ia), ia_link))				\
169		continue;						\
170	if ((ia) != NULL)						\
171		ifa_ref(&(ia)->ia_ifa);					\
172}
173#endif
174
175/*
176 * IP datagram reassembly.
177 */
178#define	IPREASS_NHASH_LOG2	6
179#define	IPREASS_NHASH		(1 << IPREASS_NHASH_LOG2)
180#define	IPREASS_HMASK		(IPREASS_NHASH - 1)
181#define	IPREASS_HASH(x,y) \
182	(((((x) & 0xF) | ((((x) >> 8) & 0xF) << 4)) ^ (y)) & IPREASS_HMASK)
183
184/*
185 * Legacy IPv4 IGMP per-link structure.
186 */
187struct router_info {
188	struct ifnet *rti_ifp;
189	int    rti_type; /* type of router which is querier on this interface */
190	int    rti_time; /* # of slow timeouts since last old query */
191	SLIST_ENTRY(router_info) rti_list;
192};
193
194/*
195 * Per-interface IGMP router version information.
196 */
197struct igmp_ifinfo {
198	LIST_ENTRY(igmp_ifinfo) igi_link;
199	struct ifnet *igi_ifp;	/* interface this instance belongs to */
200	uint32_t igi_version;	/* IGMPv3 Host Compatibility Mode */
201	uint32_t igi_v1_timer;	/* IGMPv1 Querier Present timer (s) */
202	uint32_t igi_v2_timer;	/* IGMPv2 Querier Present timer (s) */
203	uint32_t igi_v3_timer;	/* IGMPv3 General Query (interface) timer (s)*/
204	uint32_t igi_flags;	/* IGMP per-interface flags */
205	uint32_t igi_rv;	/* IGMPv3 Robustness Variable */
206	uint32_t igi_qi;	/* IGMPv3 Query Interval (s) */
207	uint32_t igi_qri;	/* IGMPv3 Query Response Interval (s) */
208	uint32_t igi_uri;	/* IGMPv3 Unsolicited Report Interval (s) */
209	SLIST_HEAD(,in_multi)	igi_relinmhead; /* released groups */
210	struct ifqueue	 igi_gq;	/* queue of general query responses */
211};
212
213#define IGIF_SILENT	0x00000001	/* Do not use IGMP on this ifp */
214#define IGIF_LOOPBACK	0x00000002	/* Send IGMP reports to loopback */
215
216/*
217 * IPv4 multicast IGMP-layer source entry.
218 */
219struct ip_msource {
220	RB_ENTRY(ip_msource)	ims_link;	/* RB tree links */
221	in_addr_t		ims_haddr;	/* host byte order */
222	struct ims_st {
223		uint16_t	ex;		/* # of exclusive members */
224		uint16_t	in;		/* # of inclusive members */
225	}			ims_st[2];	/* state at t0, t1 */
226	uint8_t			ims_stp;	/* pending query */
227};
228
229/*
230 * IPv4 multicast PCB-layer source entry.
231 */
232struct in_msource {
233	RB_ENTRY(ip_msource)	ims_link;	/* RB tree links */
234	in_addr_t		ims_haddr;	/* host byte order */
235	uint8_t			imsl_st[2];	/* state before/at commit */
236};
237
238RB_HEAD(ip_msource_tree, ip_msource);	/* define struct ip_msource_tree */
239
240static __inline int
241ip_msource_cmp(const struct ip_msource *a, const struct ip_msource *b)
242{
243
244	if (a->ims_haddr < b->ims_haddr)
245		return (-1);
246	if (a->ims_haddr == b->ims_haddr)
247		return (0);
248	return (1);
249}
250RB_PROTOTYPE(ip_msource_tree, ip_msource, ims_link, ip_msource_cmp);
251
252/*
253 * IPv4 multicast PCB-layer group filter descriptor.
254 */
255struct in_mfilter {
256	struct ip_msource_tree	imf_sources; /* source list for (S,G) */
257	u_long			imf_nsrc;    /* # of source entries */
258	uint8_t			imf_st[2];   /* state before/at commit */
259};
260
261/*
262 * IPv4 group descriptor.
263 *
264 * For every entry on an ifnet's if_multiaddrs list which represents
265 * an IP multicast group, there is one of these structures.
266 *
267 * If any source filters are present, then a node will exist in the RB-tree
268 * to permit fast lookup by source whenever an operation takes place.
269 * This permits pre-order traversal when we issue reports.
270 * Source filter trees are kept separately from the socket layer to
271 * greatly simplify locking.
272 *
273 * When IGMPv3 is active, inm_timer is the response to group query timer.
274 * The state-change timer inm_sctimer is separate; whenever state changes
275 * for the group the state change record is generated and transmitted,
276 * and kept if retransmissions are necessary.
277 *
278 * FUTURE: inm_link is now only used when groups are being purged
279 * on a detaching ifnet. It could be demoted to a SLIST_ENTRY, but
280 * because it is at the very start of the struct, we can't do this
281 * w/o breaking the ABI for ifmcstat.
282 */
283struct in_multi {
284	LIST_ENTRY(in_multi) inm_link;	/* to-be-released by in_ifdetach */
285	struct	in_addr inm_addr;	/* IP multicast address, convenience */
286	struct	ifnet *inm_ifp;		/* back pointer to ifnet */
287	struct	ifmultiaddr *inm_ifma;	/* back pointer to ifmultiaddr */
288	u_int	inm_timer;		/* IGMPv1/v2 group / v3 query timer */
289	u_int	inm_state;		/* state of the membership */
290	void	*inm_rti;		/* unused, legacy field */
291	u_int	inm_refcount;		/* reference count */
292
293	/* New fields for IGMPv3 follow. */
294	struct igmp_ifinfo	*inm_igi;	/* IGMP info */
295	SLIST_ENTRY(in_multi)	 inm_nrele;	/* to-be-released by IGMP */
296	struct ip_msource_tree	 inm_srcs;	/* tree of sources */
297	u_long			 inm_nsrc;	/* # of tree entries */
298
299	struct ifqueue		 inm_scq;	/* queue of pending
300						 * state-change packets */
301	struct timeval		 inm_lastgsrtv;	/* Time of last G-S-R query */
302	uint16_t		 inm_sctimer;	/* state-change timer */
303	uint16_t		 inm_scrv;	/* state-change rexmit count */
304
305	/*
306	 * SSM state counters which track state at T0 (the time the last
307	 * state-change report's RV timer went to zero) and T1
308	 * (time of pending report, i.e. now).
309	 * Used for computing IGMPv3 state-change reports. Several refcounts
310	 * are maintained here to optimize for common use-cases.
311	 */
312	struct inm_st {
313		uint16_t	iss_fmode;	/* IGMP filter mode */
314		uint16_t	iss_asm;	/* # of ASM listeners */
315		uint16_t	iss_ex;		/* # of exclusive members */
316		uint16_t	iss_in;		/* # of inclusive members */
317		uint16_t	iss_rec;	/* # of recorded sources */
318	}			inm_st[2];	/* state at t0, t1 */
319};
320
321/*
322 * Helper function to derive the filter mode on a source entry
323 * from its internal counters. Predicates are:
324 *  A source is only excluded if all listeners exclude it.
325 *  A source is only included if no listeners exclude it,
326 *  and at least one listener includes it.
327 * May be used by ifmcstat(8).
328 */
329static __inline uint8_t
330ims_get_mode(const struct in_multi *inm, const struct ip_msource *ims,
331    uint8_t t)
332{
333
334	t = !!t;
335	if (inm->inm_st[t].iss_ex > 0 &&
336	    inm->inm_st[t].iss_ex == ims->ims_st[t].ex)
337		return (MCAST_EXCLUDE);
338	else if (ims->ims_st[t].in > 0 && ims->ims_st[t].ex == 0)
339		return (MCAST_INCLUDE);
340	return (MCAST_UNDEFINED);
341}
342
343#ifdef _KERNEL
344
345#ifdef SYSCTL_DECL
346SYSCTL_DECL(_net_inet);
347SYSCTL_DECL(_net_inet_ip);
348SYSCTL_DECL(_net_inet_raw);
349#endif
350
351/*
352 * Lock macros for IPv4 layer multicast address lists.  IPv4 lock goes
353 * before link layer multicast locks in the lock order.  In most cases,
354 * consumers of IN_*_MULTI() macros should acquire the locks before
355 * calling them; users of the in_{add,del}multi() functions should not.
356 */
357extern struct mtx in_multi_mtx;
358#define	IN_MULTI_LOCK()		mtx_lock(&in_multi_mtx)
359#define	IN_MULTI_UNLOCK()	mtx_unlock(&in_multi_mtx)
360#define	IN_MULTI_LOCK_ASSERT()	mtx_assert(&in_multi_mtx, MA_OWNED)
361#define	IN_MULTI_UNLOCK_ASSERT() mtx_assert(&in_multi_mtx, MA_NOTOWNED)
362
363/*
364 * Function for looking up an in_multi record for an IPv4 multicast address
365 * on a given interface. ifp must be valid. If no record found, return NULL.
366 * The IN_MULTI_LOCK and IF_ADDR_LOCK on ifp must be held.
367 */
368static __inline struct in_multi *
369inm_lookup_locked(struct ifnet *ifp, const struct in_addr ina)
370{
371	struct ifmultiaddr *ifma;
372	struct in_multi *inm;
373
374	IN_MULTI_LOCK_ASSERT();
375	IF_ADDR_LOCK_ASSERT(ifp);
376
377	inm = NULL;
378	TAILQ_FOREACH(ifma, &((ifp)->if_multiaddrs), ifma_link) {
379		if (ifma->ifma_addr->sa_family == AF_INET) {
380			inm = (struct in_multi *)ifma->ifma_protospec;
381			if (inm->inm_addr.s_addr == ina.s_addr)
382				break;
383			inm = NULL;
384		}
385	}
386	return (inm);
387}
388
389/*
390 * Wrapper for inm_lookup_locked().
391 * The IF_ADDR_LOCK will be taken on ifp and released on return.
392 */
393static __inline struct in_multi *
394inm_lookup(struct ifnet *ifp, const struct in_addr ina)
395{
396	struct in_multi *inm;
397
398	IN_MULTI_LOCK_ASSERT();
399	IF_ADDR_LOCK(ifp);
400	inm = inm_lookup_locked(ifp, ina);
401	IF_ADDR_UNLOCK(ifp);
402
403	return (inm);
404}
405
406/* Acquire an in_multi record. */
407static __inline void
408inm_acquire_locked(struct in_multi *inm)
409{
410
411	IN_MULTI_LOCK_ASSERT();
412	++inm->inm_refcount;
413}
414
415/*
416 * Return values for imo_multi_filter().
417 */
418#define MCAST_PASS		0	/* Pass */
419#define MCAST_NOTGMEMBER	1	/* This host not a member of group */
420#define MCAST_NOTSMEMBER	2	/* This host excluded source */
421#define MCAST_MUTED		3	/* [deprecated] */
422
423struct	rtentry;
424struct	route;
425struct	ip_moptions;
426
427int	imo_multi_filter(const struct ip_moptions *, const struct ifnet *,
428	    const struct sockaddr *, const struct sockaddr *);
429void	inm_commit(struct in_multi *);
430void	inm_clear_recorded(struct in_multi *);
431void	inm_print(const struct in_multi *);
432int	inm_record_source(struct in_multi *inm, const in_addr_t);
433void	inm_release(struct in_multi *);
434void	inm_release_locked(struct in_multi *);
435struct	in_multi *
436	in_addmulti(struct in_addr *, struct ifnet *);
437void	in_delmulti(struct in_multi *);
438int	in_joingroup(struct ifnet *, const struct in_addr *,
439	    /*const*/ struct in_mfilter *, struct in_multi **);
440int	in_joingroup_locked(struct ifnet *, const struct in_addr *,
441	    /*const*/ struct in_mfilter *, struct in_multi **);
442int	in_leavegroup(struct in_multi *, /*const*/ struct in_mfilter *);
443int	in_leavegroup_locked(struct in_multi *,
444	    /*const*/ struct in_mfilter *);
445int	in_control(struct socket *, u_long, caddr_t, struct ifnet *,
446	    struct thread *);
447void	in_rtqdrain(void);
448void	ip_input(struct mbuf *);
449int	in_ifadown(struct ifaddr *ifa, int);
450void	in_ifscrub(struct ifnet *, struct in_ifaddr *, u_int);
451struct	mbuf	*ip_fastforward(struct mbuf *);
452void	*in_domifattach(struct ifnet *);
453void	in_domifdetach(struct ifnet *, void *);
454
455
456/* XXX */
457void	 in_rtalloc_ign(struct route *ro, u_long ignflags, u_int fibnum);
458void	 in_rtalloc(struct route *ro, u_int fibnum);
459struct rtentry *in_rtalloc1(struct sockaddr *, int, u_long, u_int);
460void	 in_rtredirect(struct sockaddr *, struct sockaddr *,
461	    struct sockaddr *, int, struct sockaddr *, u_int);
462int	 in_rtrequest(int, struct sockaddr *,
463	    struct sockaddr *, struct sockaddr *, int, struct rtentry **, u_int);
464
465#if 0
466int	 in_rt_getifa(struct rt_addrinfo *, u_int fibnum);
467int	 in_rtioctl(u_long, caddr_t, u_int);
468int	 in_rtrequest1(int, struct rt_addrinfo *, struct rtentry **, u_int);
469#endif
470#endif /* _KERNEL */
471
472/* INET6 stuff */
473#include <netinet6/in6_var.h>
474
475#endif /* _NETINET_IN_VAR_H_ */
476