xref: /freebsd/sys/netinet/in_pcb.h (revision 884a2a699669ec61e2366e3e358342dbc94be24a)
1 /*-
2  * Copyright (c) 1982, 1986, 1990, 1993
3  *	The Regents of the University of California.
4  * Copyright (c) 2010-2011 Juniper Networks, Inc.
5  * All rights reserved.
6  *
7  * Portions of this software were developed by Robert N. M. Watson under
8  * contract to Juniper Networks, Inc.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 4. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  *
34  *	@(#)in_pcb.h	8.1 (Berkeley) 6/10/93
35  * $FreeBSD$
36  */
37 
38 #ifndef _NETINET_IN_PCB_H_
39 #define _NETINET_IN_PCB_H_
40 
41 #include <sys/queue.h>
42 #include <sys/_lock.h>
43 #include <sys/_mutex.h>
44 #include <sys/_rwlock.h>
45 
46 #ifdef _KERNEL
47 #include <sys/rwlock.h>
48 #include <net/vnet.h>
49 #include <vm/uma.h>
50 #endif
51 
52 #define	in6pcb		inpcb	/* for KAME src sync over BSD*'s */
53 #define	in6p_sp		inp_sp	/* for KAME src sync over BSD*'s */
54 struct inpcbpolicy;
55 
56 /*
57  * struct inpcb is the common protocol control block structure used in most
58  * IP transport protocols.
59  *
60  * Pointers to local and foreign host table entries, local and foreign socket
61  * numbers, and pointers up (to a socket structure) and down (to a
62  * protocol-specific control block) are stored here.
63  */
64 LIST_HEAD(inpcbhead, inpcb);
65 LIST_HEAD(inpcbporthead, inpcbport);
66 typedef	u_quad_t	inp_gen_t;
67 
68 /*
69  * PCB with AF_INET6 null bind'ed laddr can receive AF_INET input packet.
70  * So, AF_INET6 null laddr is also used as AF_INET null laddr, by utilizing
71  * the following structure.
72  */
73 struct in_addr_4in6 {
74 	u_int32_t	ia46_pad32[3];
75 	struct	in_addr	ia46_addr4;
76 };
77 
78 /*
79  * NOTE: ipv6 addrs should be 64-bit aligned, per RFC 2553.  in_conninfo has
80  * some extra padding to accomplish this.
81  */
82 struct in_endpoints {
83 	u_int16_t	ie_fport;		/* foreign port */
84 	u_int16_t	ie_lport;		/* local port */
85 	/* protocol dependent part, local and foreign addr */
86 	union {
87 		/* foreign host table entry */
88 		struct	in_addr_4in6 ie46_foreign;
89 		struct	in6_addr ie6_foreign;
90 	} ie_dependfaddr;
91 	union {
92 		/* local host table entry */
93 		struct	in_addr_4in6 ie46_local;
94 		struct	in6_addr ie6_local;
95 	} ie_dependladdr;
96 };
97 #define	ie_faddr	ie_dependfaddr.ie46_foreign.ia46_addr4
98 #define	ie_laddr	ie_dependladdr.ie46_local.ia46_addr4
99 #define	ie6_faddr	ie_dependfaddr.ie6_foreign
100 #define	ie6_laddr	ie_dependladdr.ie6_local
101 
102 /*
103  * XXX The defines for inc_* are hacks and should be changed to direct
104  * references.
105  */
106 struct in_conninfo {
107 	u_int8_t	inc_flags;
108 	u_int8_t	inc_len;
109 	u_int16_t	inc_fibnum;	/* XXX was pad, 16 bits is plenty */
110 	/* protocol dependent part */
111 	struct	in_endpoints inc_ie;
112 };
113 
114 /*
115  * Flags for inc_flags.
116  */
117 #define	INC_ISIPV6	0x01
118 
119 #define inc_isipv6	inc_flags	/* temp compatability */
120 #define	inc_fport	inc_ie.ie_fport
121 #define	inc_lport	inc_ie.ie_lport
122 #define	inc_faddr	inc_ie.ie_faddr
123 #define	inc_laddr	inc_ie.ie_laddr
124 #define	inc6_faddr	inc_ie.ie6_faddr
125 #define	inc6_laddr	inc_ie.ie6_laddr
126 
127 struct	icmp6_filter;
128 
129 /*-
130  * struct inpcb captures the network layer state for TCP, UDP, and raw IPv4
131  * and IPv6 sockets.  In the case of TCP, further per-connection state is
132  * hung off of inp_ppcb most of the time.  Almost all fields of struct inpcb
133  * are static after creation or protected by a per-inpcb rwlock, inp_lock.  A
134  * few fields also require the global pcbinfo lock for the inpcb to be held,
135  * when modified, such as the global connection lists and hashes, as well as
136  * binding information (which affects which hash a connection is on).  This
137  * model means that connections can be looked up without holding the
138  * per-connection lock, which is important for performance when attempting to
139  * find the connection for a packet given its IP and port tuple.  Writing to
140  * these fields that write locks be held on both the inpcb and global locks.
141  *
142  * Key:
143  * (c) - Constant after initialization
144  * (i) - Protected by the inpcb lock
145  * (p) - Protected by the pcbinfo lock for the inpcb
146  * (s) - Protected by another subsystem's locks
147  * (x) - Undefined locking
148  *
149  * A few other notes:
150  *
151  * When a read lock is held, stability of the field is guaranteed; to write
152  * to a field, a write lock must generally be held.
153  *
154  * netinet/netinet6-layer code should not assume that the inp_socket pointer
155  * is safe to dereference without inp_lock being held, even for protocols
156  * other than TCP (where the inpcb persists during TIMEWAIT even after the
157  * socket has been freed), or there may be close(2)-related races.
158  *
159  * The inp_vflag field is overloaded, and would otherwise ideally be (c).
160  */
161 struct inpcb {
162 	LIST_ENTRY(inpcb) inp_hash;	/* (i/p) hash list */
163 	LIST_ENTRY(inpcb) inp_list;	/* (i/p) list for all PCBs for proto */
164 	void	*inp_ppcb;		/* (i) pointer to per-protocol pcb */
165 	struct	inpcbinfo *inp_pcbinfo;	/* (c) PCB list info */
166 	struct	socket *inp_socket;	/* (i) back pointer to socket */
167 	struct	ucred	*inp_cred;	/* (c) cache of socket cred */
168 	u_int32_t inp_flow;		/* (i) IPv6 flow information */
169 	int	inp_flags;		/* (i) generic IP/datagram flags */
170 	int	inp_flags2;		/* (i) generic IP/datagram flags #2*/
171 	u_char	inp_vflag;		/* (i) IP version flag (v4/v6) */
172 	u_char	inp_ip_ttl;		/* (i) time to live proto */
173 	u_char	inp_ip_p;		/* (c) protocol proto */
174 	u_char	inp_ip_minttl;		/* (i) minimum TTL or drop */
175 	uint32_t inp_flowid;		/* (x) flow id / queue id */
176 	u_int	inp_refcount;		/* (i) refcount */
177 	void	*inp_pspare[4];		/* (x) rtentry / general use */
178 	u_int	inp_ispare[4];		/* general use */
179 
180 	/* Local and foreign ports, local and foreign addr. */
181 	struct	in_conninfo inp_inc;	/* (i/p) list for PCB's local port */
182 
183 	/* MAC and IPSEC policy information. */
184 	struct	label *inp_label;	/* (i) MAC label */
185 	struct	inpcbpolicy *inp_sp;    /* (s) for IPSEC */
186 
187 	/* Protocol-dependent part; options. */
188 	struct {
189 		u_char	inp4_ip_tos;		/* (i) type of service proto */
190 		struct	mbuf *inp4_options;	/* (i) IP options */
191 		struct	ip_moptions *inp4_moptions; /* (i) IP mcast options */
192 	} inp_depend4;
193 	struct {
194 		/* (i) IP options */
195 		struct	mbuf *inp6_options;
196 		/* (i) IP6 options for outgoing packets */
197 		struct	ip6_pktopts *inp6_outputopts;
198 		/* (i) IP multicast options */
199 		struct	ip6_moptions *inp6_moptions;
200 		/* (i) ICMPv6 code type filter */
201 		struct	icmp6_filter *inp6_icmp6filt;
202 		/* (i) IPV6_CHECKSUM setsockopt */
203 		int	inp6_cksum;
204 		short	inp6_hops;
205 	} inp_depend6;
206 	LIST_ENTRY(inpcb) inp_portlist;	/* (i/p) */
207 	struct	inpcbport *inp_phd;	/* (i/p) head of this list */
208 #define inp_zero_size offsetof(struct inpcb, inp_gencnt)
209 	inp_gen_t	inp_gencnt;	/* (c) generation count */
210 	struct llentry	*inp_lle;	/* cached L2 information */
211 	struct rtentry	*inp_rt;	/* cached L3 information */
212 	struct rwlock	inp_lock;
213 };
214 #define	inp_fport	inp_inc.inc_fport
215 #define	inp_lport	inp_inc.inc_lport
216 #define	inp_faddr	inp_inc.inc_faddr
217 #define	inp_laddr	inp_inc.inc_laddr
218 #define	inp_ip_tos	inp_depend4.inp4_ip_tos
219 #define	inp_options	inp_depend4.inp4_options
220 #define	inp_moptions	inp_depend4.inp4_moptions
221 
222 #define	in6p_faddr	inp_inc.inc6_faddr
223 #define	in6p_laddr	inp_inc.inc6_laddr
224 #define	in6p_hops	inp_depend6.inp6_hops	/* default hop limit */
225 #define	in6p_flowinfo	inp_flow
226 #define	in6p_options	inp_depend6.inp6_options
227 #define	in6p_outputopts	inp_depend6.inp6_outputopts
228 #define	in6p_moptions	inp_depend6.inp6_moptions
229 #define	in6p_icmp6filt	inp_depend6.inp6_icmp6filt
230 #define	in6p_cksum	inp_depend6.inp6_cksum
231 
232 #define	inp_vnet	inp_pcbinfo->ipi_vnet
233 
234 /*
235  * The range of the generation count, as used in this implementation, is 9e19.
236  * We would have to create 300 billion connections per second for this number
237  * to roll over in a year.  This seems sufficiently unlikely that we simply
238  * don't concern ourselves with that possibility.
239  */
240 
241 /*
242  * Interface exported to userland by various protocols which use inpcbs.  Hack
243  * alert -- only define if struct xsocket is in scope.
244  */
245 #ifdef _SYS_SOCKETVAR_H_
246 struct	xinpcb {
247 	size_t	xi_len;		/* length of this structure */
248 	struct	inpcb xi_inp;
249 	struct	xsocket xi_socket;
250 	u_quad_t	xi_alignment_hack;
251 };
252 
253 struct	xinpgen {
254 	size_t	xig_len;	/* length of this structure */
255 	u_int	xig_count;	/* number of PCBs at this time */
256 	inp_gen_t xig_gen;	/* generation count at this time */
257 	so_gen_t xig_sogen;	/* socket generation count at this time */
258 };
259 #endif /* _SYS_SOCKETVAR_H_ */
260 
261 struct inpcbport {
262 	LIST_ENTRY(inpcbport) phd_hash;
263 	struct inpcbhead phd_pcblist;
264 	u_short phd_port;
265 };
266 
267 /*-
268  * Global data structure for each high-level protocol (UDP, TCP, ...) in both
269  * IPv4 and IPv6.  Holds inpcb lists and information for managing them.
270  *
271  * Each pcbinfo is protected by two locks: ipi_lock and ipi_hash_lock,
272  * the former covering mutable global fields (such as the global pcb list),
273  * and the latter covering the hashed lookup tables.  The lock order is:
274  *
275  *    ipi_lock (before) inpcb locks (before) ipi_hash_lock
276  *
277  * Locking key:
278  *
279  * (c) Constant or nearly constant after initialisation
280  * (g) Locked by ipi_lock
281  * (h) Read using either ipi_hash_lock or inpcb lock; write requires both.
282  * (x) Synchronisation properties poorly defined
283  */
284 struct inpcbinfo {
285 	/*
286 	 * Global lock protecting global inpcb list, inpcb count, etc.
287 	 */
288 	struct rwlock		 ipi_lock;
289 
290 	/*
291 	 * Global list of inpcbs on the protocol.
292 	 */
293 	struct inpcbhead	*ipi_listhead;		/* (g) */
294 	u_int			 ipi_count;		/* (g) */
295 
296 	/*
297 	 * Generation count -- incremented each time a connection is allocated
298 	 * or freed.
299 	 */
300 	u_quad_t		 ipi_gencnt;		/* (g) */
301 
302 	/*
303 	 * Fields associated with port lookup and allocation.
304 	 */
305 	u_short			 ipi_lastport;		/* (x) */
306 	u_short			 ipi_lastlow;		/* (x) */
307 	u_short			 ipi_lasthi;		/* (x) */
308 
309 	/*
310 	 * UMA zone from which inpcbs are allocated for this protocol.
311 	 */
312 	struct	uma_zone	*ipi_zone;		/* (c) */
313 
314 	/*
315 	 * Global lock protecting hash lookup tables.
316 	 */
317 	struct rwlock		 ipi_hash_lock;
318 
319 	/*
320 	 * Global hash of inpcbs, hashed by local and foreign addresses and
321 	 * port numbers.
322 	 */
323 	struct inpcbhead	*ipi_hashbase;		/* (h) */
324 	u_long			 ipi_hashmask;		/* (h) */
325 
326 	/*
327 	 * Global hash of inpcbs, hashed by only local port number.
328 	 */
329 	struct inpcbporthead	*ipi_porthashbase;	/* (h) */
330 	u_long			 ipi_porthashmask;	/* (h) */
331 
332 	/*
333 	 * Pointer to network stack instance
334 	 */
335 	struct vnet		*ipi_vnet;		/* (c) */
336 
337 	/*
338 	 * general use 2
339 	 */
340 	void 			*ipi_pspare[2];
341 };
342 
343 #define INP_LOCK_INIT(inp, d, t) \
344 	rw_init_flags(&(inp)->inp_lock, (t), RW_RECURSE |  RW_DUPOK)
345 #define INP_LOCK_DESTROY(inp)	rw_destroy(&(inp)->inp_lock)
346 #define INP_RLOCK(inp)		rw_rlock(&(inp)->inp_lock)
347 #define INP_WLOCK(inp)		rw_wlock(&(inp)->inp_lock)
348 #define INP_TRY_RLOCK(inp)	rw_try_rlock(&(inp)->inp_lock)
349 #define INP_TRY_WLOCK(inp)	rw_try_wlock(&(inp)->inp_lock)
350 #define INP_RUNLOCK(inp)	rw_runlock(&(inp)->inp_lock)
351 #define INP_WUNLOCK(inp)	rw_wunlock(&(inp)->inp_lock)
352 #define	INP_TRY_UPGRADE(inp)	rw_try_upgrade(&(inp)->inp_lock)
353 #define	INP_DOWNGRADE(inp)	rw_downgrade(&(inp)->inp_lock)
354 #define	INP_WLOCKED(inp)	rw_wowned(&(inp)->inp_lock)
355 #define	INP_LOCK_ASSERT(inp)	rw_assert(&(inp)->inp_lock, RA_LOCKED)
356 #define	INP_RLOCK_ASSERT(inp)	rw_assert(&(inp)->inp_lock, RA_RLOCKED)
357 #define	INP_WLOCK_ASSERT(inp)	rw_assert(&(inp)->inp_lock, RA_WLOCKED)
358 #define	INP_UNLOCK_ASSERT(inp)	rw_assert(&(inp)->inp_lock, RA_UNLOCKED)
359 
360 #ifdef _KERNEL
361 /*
362  * These locking functions are for inpcb consumers outside of sys/netinet,
363  * more specifically, they were added for the benefit of TOE drivers. The
364  * macros are reserved for use by the stack.
365  */
366 void inp_wlock(struct inpcb *);
367 void inp_wunlock(struct inpcb *);
368 void inp_rlock(struct inpcb *);
369 void inp_runlock(struct inpcb *);
370 
371 #ifdef INVARIANTS
372 void inp_lock_assert(struct inpcb *);
373 void inp_unlock_assert(struct inpcb *);
374 #else
375 static __inline void
376 inp_lock_assert(struct inpcb *inp __unused)
377 {
378 }
379 
380 static __inline void
381 inp_unlock_assert(struct inpcb *inp __unused)
382 {
383 }
384 
385 #endif
386 
387 void	inp_apply_all(void (*func)(struct inpcb *, void *), void *arg);
388 int 	inp_ip_tos_get(const struct inpcb *inp);
389 void 	inp_ip_tos_set(struct inpcb *inp, int val);
390 struct socket *
391 	inp_inpcbtosocket(struct inpcb *inp);
392 struct tcpcb *
393 	inp_inpcbtotcpcb(struct inpcb *inp);
394 void 	inp_4tuple_get(struct inpcb *inp, uint32_t *laddr, uint16_t *lp,
395 		uint32_t *faddr, uint16_t *fp);
396 
397 #endif /* _KERNEL */
398 
399 #define INP_INFO_LOCK_INIT(ipi, d) \
400 	rw_init_flags(&(ipi)->ipi_lock, (d), RW_RECURSE)
401 #define INP_INFO_LOCK_DESTROY(ipi)  rw_destroy(&(ipi)->ipi_lock)
402 #define INP_INFO_RLOCK(ipi)	rw_rlock(&(ipi)->ipi_lock)
403 #define INP_INFO_WLOCK(ipi)	rw_wlock(&(ipi)->ipi_lock)
404 #define INP_INFO_TRY_RLOCK(ipi)	rw_try_rlock(&(ipi)->ipi_lock)
405 #define INP_INFO_TRY_WLOCK(ipi)	rw_try_wlock(&(ipi)->ipi_lock)
406 #define INP_INFO_TRY_UPGRADE(ipi)	rw_try_upgrade(&(ipi)->ipi_lock)
407 #define INP_INFO_RUNLOCK(ipi)	rw_runlock(&(ipi)->ipi_lock)
408 #define INP_INFO_WUNLOCK(ipi)	rw_wunlock(&(ipi)->ipi_lock)
409 #define	INP_INFO_LOCK_ASSERT(ipi)	rw_assert(&(ipi)->ipi_lock, RA_LOCKED)
410 #define INP_INFO_RLOCK_ASSERT(ipi)	rw_assert(&(ipi)->ipi_lock, RA_RLOCKED)
411 #define INP_INFO_WLOCK_ASSERT(ipi)	rw_assert(&(ipi)->ipi_lock, RA_WLOCKED)
412 #define INP_INFO_UNLOCK_ASSERT(ipi)	rw_assert(&(ipi)->ipi_lock, RA_UNLOCKED)
413 
414 #define	INP_HASH_LOCK_INIT(ipi, d) \
415 	rw_init_flags(&(ipi)->ipi_hash_lock, (d), 0)
416 #define	INP_HASH_LOCK_DESTROY(ipi)	rw_destroy(&(ipi)->ipi_hash_lock)
417 #define	INP_HASH_RLOCK(ipi)		rw_rlock(&(ipi)->ipi_hash_lock)
418 #define	INP_HASH_WLOCK(ipi)		rw_wlock(&(ipi)->ipi_hash_lock)
419 #define	INP_HASH_RUNLOCK(ipi)		rw_runlock(&(ipi)->ipi_hash_lock)
420 #define	INP_HASH_WUNLOCK(ipi)		rw_wunlock(&(ipi)->ipi_hash_lock)
421 #define	INP_HASH_LOCK_ASSERT(ipi)	rw_assert(&(ipi)->ipi_hash_lock, \
422 					    RA_LOCKED)
423 #define	INP_HASH_WLOCK_ASSERT(ipi)	rw_assert(&(ipi)->ipi_hash_lock, \
424 					    RA_WLOCKED)
425 
426 #define INP_PCBHASH(faddr, lport, fport, mask) \
427 	(((faddr) ^ ((faddr) >> 16) ^ ntohs((lport) ^ (fport))) & (mask))
428 #define INP_PCBPORTHASH(lport, mask) \
429 	(ntohs((lport)) & (mask))
430 
431 /*
432  * Flags for inp_vflags -- historically version flags only
433  */
434 #define	INP_IPV4	0x1
435 #define	INP_IPV6	0x2
436 #define	INP_IPV6PROTO	0x4		/* opened under IPv6 protocol */
437 
438 /*
439  * Flags for inp_flags.
440  */
441 #define	INP_RECVOPTS		0x00000001 /* receive incoming IP options */
442 #define	INP_RECVRETOPTS		0x00000002 /* receive IP options for reply */
443 #define	INP_RECVDSTADDR		0x00000004 /* receive IP dst address */
444 #define	INP_HDRINCL		0x00000008 /* user supplies entire IP header */
445 #define	INP_HIGHPORT		0x00000010 /* user wants "high" port binding */
446 #define	INP_LOWPORT		0x00000020 /* user wants "low" port binding */
447 #define	INP_ANONPORT		0x00000040 /* port chosen for user */
448 #define	INP_RECVIF		0x00000080 /* receive incoming interface */
449 #define	INP_MTUDISC		0x00000100 /* user can do MTU discovery */
450 #define	INP_FAITH		0x00000200 /* accept FAITH'ed connections */
451 #define	INP_RECVTTL		0x00000400 /* receive incoming IP TTL */
452 #define	INP_DONTFRAG		0x00000800 /* don't fragment packet */
453 #define	INP_BINDANY		0x00001000 /* allow bind to any address */
454 #define	INP_INHASHLIST		0x00002000 /* in_pcbinshash() has been called */
455 #define	IN6P_IPV6_V6ONLY	0x00008000 /* restrict AF_INET6 socket for v6 */
456 #define	IN6P_PKTINFO		0x00010000 /* receive IP6 dst and I/F */
457 #define	IN6P_HOPLIMIT		0x00020000 /* receive hoplimit */
458 #define	IN6P_HOPOPTS		0x00040000 /* receive hop-by-hop options */
459 #define	IN6P_DSTOPTS		0x00080000 /* receive dst options after rthdr */
460 #define	IN6P_RTHDR		0x00100000 /* receive routing header */
461 #define	IN6P_RTHDRDSTOPTS	0x00200000 /* receive dstoptions before rthdr */
462 #define	IN6P_TCLASS		0x00400000 /* receive traffic class value */
463 #define	IN6P_AUTOFLOWLABEL	0x00800000 /* attach flowlabel automatically */
464 #define	INP_TIMEWAIT		0x01000000 /* in TIMEWAIT, ppcb is tcptw */
465 #define	INP_ONESBCAST		0x02000000 /* send all-ones broadcast */
466 #define	INP_DROPPED		0x04000000 /* protocol drop flag */
467 #define	INP_SOCKREF		0x08000000 /* strong socket reference */
468 #define	INP_SW_FLOWID           0x10000000 /* software generated flow id */
469 #define	INP_HW_FLOWID           0x20000000 /* hardware generated flow id */
470 #define	IN6P_RFC2292		0x40000000 /* used RFC2292 API on the socket */
471 #define	IN6P_MTU		0x80000000 /* receive path MTU */
472 
473 #define	INP_CONTROLOPTS		(INP_RECVOPTS|INP_RECVRETOPTS|INP_RECVDSTADDR|\
474 				 INP_RECVIF|INP_RECVTTL|\
475 				 IN6P_PKTINFO|IN6P_HOPLIMIT|IN6P_HOPOPTS|\
476 				 IN6P_DSTOPTS|IN6P_RTHDR|IN6P_RTHDRDSTOPTS|\
477 				 IN6P_TCLASS|IN6P_AUTOFLOWLABEL|IN6P_RFC2292|\
478 				 IN6P_MTU)
479 
480 /*
481  * Flags for inp_flags2.
482  */
483 #define	INP_LLE_VALID		0x00000001 /* cached lle is valid */
484 #define	INP_RT_VALID		0x00000002 /* cached rtentry is valid */
485 
486 /*
487  * Flags passed to in_pcblookup*() functions.
488  */
489 #define	INPLOOKUP_WILDCARD	0x00000001	/* Allow wildcard sockets. */
490 #define	INPLOOKUP_RLOCKPCB	0x00000002	/* Return inpcb read-locked. */
491 #define	INPLOOKUP_WLOCKPCB	0x00000004	/* Return inpcb write-locked. */
492 
493 #define	INPLOOKUP_MASK	(INPLOOKUP_WILDCARD | INPLOOKUP_RLOCKPCB | \
494 			    INPLOOKUP_WLOCKPCB)
495 
496 #define	sotoinpcb(so)	((struct inpcb *)(so)->so_pcb)
497 #define	sotoin6pcb(so)	sotoinpcb(so) /* for KAME src sync over BSD*'s */
498 
499 #define	INP_SOCKAF(so) so->so_proto->pr_domain->dom_family
500 
501 #define	INP_CHECK_SOCKAF(so, af)	(INP_SOCKAF(so) == af)
502 
503 #ifdef _KERNEL
504 VNET_DECLARE(int, ipport_reservedhigh);
505 VNET_DECLARE(int, ipport_reservedlow);
506 VNET_DECLARE(int, ipport_lowfirstauto);
507 VNET_DECLARE(int, ipport_lowlastauto);
508 VNET_DECLARE(int, ipport_firstauto);
509 VNET_DECLARE(int, ipport_lastauto);
510 VNET_DECLARE(int, ipport_hifirstauto);
511 VNET_DECLARE(int, ipport_hilastauto);
512 VNET_DECLARE(int, ipport_randomized);
513 VNET_DECLARE(int, ipport_randomcps);
514 VNET_DECLARE(int, ipport_randomtime);
515 VNET_DECLARE(int, ipport_stoprandom);
516 VNET_DECLARE(int, ipport_tcpallocs);
517 
518 #define	V_ipport_reservedhigh	VNET(ipport_reservedhigh)
519 #define	V_ipport_reservedlow	VNET(ipport_reservedlow)
520 #define	V_ipport_lowfirstauto	VNET(ipport_lowfirstauto)
521 #define	V_ipport_lowlastauto	VNET(ipport_lowlastauto)
522 #define	V_ipport_firstauto	VNET(ipport_firstauto)
523 #define	V_ipport_lastauto	VNET(ipport_lastauto)
524 #define	V_ipport_hifirstauto	VNET(ipport_hifirstauto)
525 #define	V_ipport_hilastauto	VNET(ipport_hilastauto)
526 #define	V_ipport_randomized	VNET(ipport_randomized)
527 #define	V_ipport_randomcps	VNET(ipport_randomcps)
528 #define	V_ipport_randomtime	VNET(ipport_randomtime)
529 #define	V_ipport_stoprandom	VNET(ipport_stoprandom)
530 #define	V_ipport_tcpallocs	VNET(ipport_tcpallocs)
531 
532 void	in_pcbinfo_destroy(struct inpcbinfo *);
533 void	in_pcbinfo_init(struct inpcbinfo *, const char *, struct inpcbhead *,
534 	    int, int, char *, uma_init, uma_fini, uint32_t);
535 
536 void	in_pcbpurgeif0(struct inpcbinfo *, struct ifnet *);
537 int	in_pcballoc(struct socket *, struct inpcbinfo *);
538 int	in_pcbbind(struct inpcb *, struct sockaddr *, struct ucred *);
539 int	in_pcb_lport(struct inpcb *, struct in_addr *, u_short *,
540 	    struct ucred *, int);
541 int	in_pcbbind_setup(struct inpcb *, struct sockaddr *, in_addr_t *,
542 	    u_short *, struct ucred *);
543 int	in_pcbconnect(struct inpcb *, struct sockaddr *, struct ucred *);
544 int	in_pcbconnect_mbuf(struct inpcb *, struct sockaddr *, struct ucred *,
545 	    struct mbuf *);
546 int	in_pcbconnect_setup(struct inpcb *, struct sockaddr *, in_addr_t *,
547 	    u_short *, in_addr_t *, u_short *, struct inpcb **,
548 	    struct ucred *);
549 void	in_pcbdetach(struct inpcb *);
550 void	in_pcbdisconnect(struct inpcb *);
551 void	in_pcbdrop(struct inpcb *);
552 void	in_pcbfree(struct inpcb *);
553 int	in_pcbinshash(struct inpcb *);
554 struct inpcb *
555 	in_pcblookup_local(struct inpcbinfo *,
556 	    struct in_addr, u_short, int, struct ucred *);
557 struct inpcb *
558 	in_pcblookup(struct inpcbinfo *, struct in_addr, u_int,
559 	    struct in_addr, u_int, int, struct ifnet *);
560 struct inpcb *
561 	in_pcblookup_mbuf(struct inpcbinfo *, struct in_addr, u_int,
562 	    struct in_addr, u_int, int, struct ifnet *, struct mbuf *);
563 void	in_pcbnotifyall(struct inpcbinfo *pcbinfo, struct in_addr,
564 	    int, struct inpcb *(*)(struct inpcb *, int));
565 void	in_pcbref(struct inpcb *);
566 void	in_pcbrehash(struct inpcb *);
567 void	in_pcbrehash_mbuf(struct inpcb *, struct mbuf *);
568 int	in_pcbrele(struct inpcb *);
569 int	in_pcbrele_rlocked(struct inpcb *);
570 int	in_pcbrele_wlocked(struct inpcb *);
571 void	in_pcbsetsolabel(struct socket *so);
572 int	in_getpeeraddr(struct socket *so, struct sockaddr **nam);
573 int	in_getsockaddr(struct socket *so, struct sockaddr **nam);
574 struct sockaddr *
575 	in_sockaddr(in_port_t port, struct in_addr *addr);
576 void	in_pcbsosetlabel(struct socket *so);
577 #endif /* _KERNEL */
578 
579 #endif /* !_NETINET_IN_PCB_H_ */
580