1 /*
2  * Copyright (c) 2000-2008 Apple Inc. All rights reserved.
3  *
4  * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
5  *
6  * This file contains Original Code and/or Modifications of Original Code
7  * as defined in and that are subject to the Apple Public Source License
8  * Version 2.0 (the 'License'). You may not use this file except in
9  * compliance with the License. The rights granted to you under the License
10  * may not be used to create, or enable the creation or redistribution of,
11  * unlawful or unlicensed copies of an Apple operating system, or to
12  * circumvent, violate, or enable the circumvention or violation of, any
13  * terms of an Apple operating system software license agreement.
14  *
15  * Please obtain a copy of the License at
16  * http://www.opensource.apple.com/apsl/ and read it before using this file.
17  *
18  * The Original Code and all software distributed under the License are
19  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
20  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
21  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
22  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
23  * Please see the License for the specific language governing rights and
24  * limitations under the License.
25  *
26  * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
27  */
28 /*
29  * Copyright (c) 1980, 1986, 1993
30  *	The Regents of the University of California.  All rights reserved.
31  *
32  * Redistribution and use in source and binary forms, with or without
33  * modification, are permitted provided that the following conditions
34  * are met:
35  * 1. Redistributions of source code must retain the above copyright
36  *    notice, this list of conditions and the following disclaimer.
37  * 2. Redistributions in binary form must reproduce the above copyright
38  *    notice, this list of conditions and the following disclaimer in the
39  *    documentation and/or other materials provided with the distribution.
40  * 3. All advertising materials mentioning features or use of this software
41  *    must display the following acknowledgement:
42  *	This product includes software developed by the University of
43  *	California, Berkeley and its contributors.
44  * 4. Neither the name of the University nor the names of its contributors
45  *    may be used to endorse or promote products derived from this software
46  *    without specific prior written permission.
47  *
48  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58  * SUCH DAMAGE.
59  *
60  *	@(#)route.h	8.3 (Berkeley) 4/19/94
61  * $FreeBSD: src/sys/net/route.h,v 1.36.2.1 2000/08/16 06:14:23 jayanth Exp $
62  */
63 
64 #ifndef _NET_ROUTE_H_
65 #define _NET_ROUTE_H_
66 #include <sys/appleapiopts.h>
67 #include <stdint.h>
68 #include <sys/types.h>
69 #include <sys/socket.h>
70 
71 /*
72  * Kernel resident routing tables.
73  *
74  * The routing tables are initialized when interface addresses
75  * are set by making entries for all directly connected interfaces.
76  */
77 
78 /*
79  * A route consists of a destination address and a reference
80  * to a routing entry.  These are often held by protocols
81  * in their control blocks, e.g. inpcb.
82  */
83 #ifdef PRIVATE
84 struct  rtentry;
85 struct route {
86 	/*
87 	 * N.B: struct route must begin with ro_rt and ro_flags
88 	 * because the code does some casts of a 'struct route_in6 *'
89 	 * to a 'struct route *'.
90 	 */
91 	struct rtentry	*ro_rt;
92 	uint32_t	ro_flags;	/* route flags (see below) */
93 	struct sockaddr	ro_dst;
94 };
95 
96 #define	ROF_SRCIF_SELECTED	0x1 /* source interface was selected */
97 
98 #else
99 struct route;
100 #endif /* PRIVATE */
101 
102 /*
103  * These numbers are used by reliable protocols for determining
104  * retransmission behavior and are included in the routing structure.
105  */
106 struct rt_metrics {
107 	u_int32_t	rmx_locks;	/* Kernel must leave these values alone */
108 	u_int32_t	rmx_mtu;	/* MTU for this path */
109 	u_int32_t	rmx_hopcount;	/* max hops expected */
110 	int32_t		rmx_expire;	/* lifetime for route, e.g. redirect */
111 	u_int32_t	rmx_recvpipe;	/* inbound delay-bandwidth product */
112 	u_int32_t	rmx_sendpipe;	/* outbound delay-bandwidth product */
113 	u_int32_t	rmx_ssthresh;	/* outbound gateway buffer limit */
114 	u_int32_t	rmx_rtt;	/* estimated round trip time */
115 	u_int32_t	rmx_rttvar;	/* estimated rtt variance */
116 	u_int32_t	rmx_pksent;	/* packets sent using this route */
117 	u_int32_t	rmx_filler[4];	/* will be used for T/TCP later */
118 };
119 
120 /*
121  * rmx_rtt and rmx_rttvar are stored as microseconds;
122  */
123 #define	RTM_RTTUNIT	1000000	/* units for rtt, rttvar, as units per sec */
124 
125 /*
126  * We distinguish between routes to hosts and routes to networks,
127  * preferring the former if available.  For each route we infer
128  * the interface to use from the gateway address supplied when
129  * the route was entered.  Routes that forward packets through
130  * gateways are marked so that the output routines know to address the
131  * gateway rather than the ultimate destination.
132  */
133 #ifdef KERNEL_PRIVATE
134 #include <kern/locks.h>
135 #ifndef RNF_NORMAL
136 #include <net/radix.h>
137 #endif
138 /*
139  * Kernel routing entry structure (private).
140  */
141 struct rtentry {
142 	struct	radix_node rt_nodes[2];	/* tree glue, and other values */
143 #define	rt_key(r)	((struct sockaddr *)((r)->rt_nodes->rn_key))
144 #define	rt_mask(r)	((struct sockaddr *)((r)->rt_nodes->rn_mask))
145 	struct	sockaddr *rt_gateway;	/* value */
146 	int32_t	rt_refcnt;		/* # held references */
147 	uint32_t rt_flags;		/* up/down?, host/net */
148 	struct	ifnet *rt_ifp;		/* the answer: interface to use */
149 	struct	ifaddr *rt_ifa;		/* the answer: interface addr to use */
150 	struct	sockaddr *rt_genmask;	/* for generation of cloned routes */
151 	void	*rt_llinfo;		/* pointer to link level info cache */
152 	void	(*rt_llinfo_free)(void *); /* link level info free function */
153 	struct	rt_metrics rt_rmx;	/* metrics used by rx'ing protocols */
154 	struct	rtentry *rt_gwroute;	/* implied entry for gatewayed routes */
155 	struct	rtentry *rt_parent;	/* cloning parent of this route */
156 	uint32_t generation_id;		/* route generation id */
157 	/*
158 	 * See bsd/net/route.c for synchronization notes.
159 	 */
160 	decl_lck_mtx_data(, rt_lock);	/* lock for routing entry */
161 };
162 #endif /* KERNEL_PRIVATE */
163 
164 #ifdef KERNEL_PRIVATE
165 #define rt_use rt_rmx.rmx_pksent
166 #endif /* KERNEL_PRIVATE */
167 
168 #define	RTF_UP		0x1		/* route usable */
169 #define	RTF_GATEWAY	0x2		/* destination is a gateway */
170 #define	RTF_HOST	0x4		/* host entry (net otherwise) */
171 #define	RTF_REJECT	0x8		/* host or net unreachable */
172 #define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
173 #define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
174 #define RTF_DONE	0x40		/* message confirmed */
175 #define RTF_DELCLONE	0x80		/* delete cloned route */
176 #define RTF_CLONING	0x100		/* generate new routes on use */
177 #define RTF_XRESOLVE	0x200		/* external daemon resolves name */
178 #define RTF_LLINFO	0x400		/* generated by link layer (e.g. ARP) */
179 #define RTF_STATIC	0x800		/* manually added */
180 #define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
181 #define RTF_PROTO2	0x4000		/* protocol specific routing flag */
182 #define RTF_PROTO1	0x8000		/* protocol specific routing flag */
183 
184 #define RTF_PRCLONING	0x10000		/* protocol requires cloning */
185 #define RTF_WASCLONED	0x20000		/* route generated through cloning */
186 #define RTF_PROTO3	0x40000		/* protocol specific routing flag */
187 					/* 0x80000 unused */
188 #define RTF_PINNED	0x100000	/* future use */
189 #define	RTF_LOCAL	0x200000	/* route represents a local address */
190 #define	RTF_BROADCAST	0x400000	/* route represents a bcast address */
191 #define	RTF_MULTICAST	0x800000	/* route represents a mcast address */
192 #define RTF_IFSCOPE	0x1000000	/* has valid interface scope */
193 #define RTF_CONDEMNED	0x2000000	/* defunct; no longer modifiable */
194 					/* 0x4000000 and up unassigned */
195 
196 /*
197  * Routing statistics.
198  */
199 struct	rtstat {
200 	short	rts_badredirect;	/* bogus redirect calls */
201 	short	rts_dynamic;		/* routes created by redirects */
202 	short	rts_newgateway;		/* routes modified by redirects */
203 	short	rts_unreach;		/* lookups which failed */
204 	short	rts_wildcard;		/* lookups satisfied by a wildcard */
205 };
206 
207 /*
208  * Structures for routing messages.
209  */
210 struct rt_msghdr {
211 	u_short	rtm_msglen;		/* to skip over non-understood messages */
212 	u_char	rtm_version;		/* future binary compatibility */
213 	u_char	rtm_type;		/* message type */
214 	u_short	rtm_index;		/* index for associated ifp */
215 	int	rtm_flags;		/* flags, incl. kern & message, e.g. DONE */
216 	int	rtm_addrs;		/* bitmask identifying sockaddrs in msg */
217 	pid_t	rtm_pid;		/* identify sender */
218 	int	rtm_seq;		/* for sender to identify action */
219 	int	rtm_errno;		/* why failed */
220 	int	rtm_use;		/* from rtentry */
221 	u_int32_t rtm_inits;		/* which metrics we are initializing */
222 	struct rt_metrics rtm_rmx;	/* metrics themselves */
223 };
224 
225 struct rt_msghdr2 {
226 	u_short	rtm_msglen;		/* to skip over non-understood messages */
227 	u_char	rtm_version;		/* future binary compatibility */
228 	u_char	rtm_type;		/* message type */
229 	u_short	rtm_index;		/* index for associated ifp */
230 	int	rtm_flags;		/* flags, incl. kern & message, e.g. DONE */
231 	int	rtm_addrs;		/* bitmask identifying sockaddrs in msg */
232 	int32_t	rtm_refcnt;		/* reference count */
233 	int	rtm_parentflags;	/* flags of the parent route */
234 	int	rtm_reserved;		/* reserved field set to 0 */
235 	int	rtm_use;		/* from rtentry */
236 	u_int32_t rtm_inits;		/* which metrics we are initializing */
237 	struct rt_metrics rtm_rmx;	/* metrics themselves */
238 };
239 
240 
241 #define RTM_VERSION	5	/* Up the ante and ignore older versions */
242 
243 /*
244  * Message types.
245  */
246 #define RTM_ADD		0x1	/* Add Route */
247 #define RTM_DELETE	0x2	/* Delete Route */
248 #define RTM_CHANGE	0x3	/* Change Metrics or flags */
249 #define RTM_GET		0x4	/* Report Metrics */
250 #define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
251 #define RTM_REDIRECT	0x6	/* Told to use different route */
252 #define RTM_MISS	0x7	/* Lookup failed on this address */
253 #define RTM_LOCK	0x8	/* fix specified metrics */
254 #define RTM_OLDADD	0x9	/* caused by SIOCADDRT */
255 #define RTM_OLDDEL	0xa	/* caused by SIOCDELRT */
256 #define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
257 #define RTM_NEWADDR	0xc	/* address being added to iface */
258 #define RTM_DELADDR	0xd	/* address being removed from iface */
259 #define RTM_IFINFO	0xe	/* iface going up/down etc. */
260 #define	RTM_NEWMADDR	0xf	/* mcast group membership being added to if */
261 #define	RTM_DELMADDR	0x10	/* mcast group membership being deleted */
262 #ifdef PRIVATE
263 #define RTM_GET_SILENT	0x11
264 #endif /* PRIVATE */
265 #define RTM_IFINFO2	0x12	/* */
266 #define RTM_NEWMADDR2	0x13	/* */
267 #define RTM_GET2	0x14	/* */
268 
269 /*
270  * Bitmask values for rtm_inits and rmx_locks.
271  */
272 #define RTV_MTU		0x1	/* init or lock _mtu */
273 #define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
274 #define RTV_EXPIRE	0x4	/* init or lock _expire */
275 #define RTV_RPIPE	0x8	/* init or lock _recvpipe */
276 #define RTV_SPIPE	0x10	/* init or lock _sendpipe */
277 #define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
278 #define RTV_RTT		0x40	/* init or lock _rtt */
279 #define RTV_RTTVAR	0x80	/* init or lock _rttvar */
280 
281 /*
282  * Bitmask values for rtm_addrs.
283  */
284 #define RTA_DST		0x1	/* destination sockaddr present */
285 #define RTA_GATEWAY	0x2	/* gateway sockaddr present */
286 #define RTA_NETMASK	0x4	/* netmask sockaddr present */
287 #define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
288 #define RTA_IFP		0x10	/* interface name sockaddr present */
289 #define RTA_IFA		0x20	/* interface addr sockaddr present */
290 #define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
291 #define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
292 
293 /*
294  * Index offsets for sockaddr array for alternate internal encoding.
295  */
296 #define RTAX_DST	0	/* destination sockaddr present */
297 #define RTAX_GATEWAY	1	/* gateway sockaddr present */
298 #define RTAX_NETMASK	2	/* netmask sockaddr present */
299 #define RTAX_GENMASK	3	/* cloning mask sockaddr present */
300 #define RTAX_IFP	4	/* interface name sockaddr present */
301 #define RTAX_IFA	5	/* interface addr sockaddr present */
302 #define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
303 #define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
304 #define RTAX_MAX	8	/* size of array to allocate */
305 
306 struct rt_addrinfo {
307 	int	rti_addrs;
308 	struct	sockaddr *rti_info[RTAX_MAX];
309 };
310 
311 struct route_cb {
312 	int	ip_count;
313 	int	ip6_count;
314 	int	ipx_count;
315 	int	ns_count;
316 	int	iso_count;
317 	int	any_count;
318 };
319 
320 #ifdef PRIVATE
321 /*
322  * For scoped routing; a zero interface scope value means nil/no scope.
323  */
324 #define	IFSCOPE_NONE	0
325 #endif /* PRIVATE */
326 
327 #ifdef KERNEL_PRIVATE
328 /*
329  * Generic call trace used by some subsystems (e.g. route, ifaddr)
330  */
331 #define	CTRACE_STACK_SIZE	8		/* depth of stack trace */
332 #define	CTRACE_HIST_SIZE	4		/* refcnt history size */
333 typedef struct ctrace {
334 	void	*th;				/* thread ptr */
335 	void	*pc[CTRACE_STACK_SIZE];		/* PC stack trace */
336 } ctrace_t;
337 
338 extern void ctrace_record(ctrace_t *);
339 
340 #define	RT_LOCK_ASSERT_HELD(_rt)					\
341 	lck_mtx_assert(&(_rt)->rt_lock, LCK_MTX_ASSERT_OWNED)
342 
343 #define	RT_LOCK_ASSERT_NOTHELD(_rt)					\
344 	lck_mtx_assert(&(_rt)->rt_lock, LCK_MTX_ASSERT_NOTOWNED)
345 
346 #define	RT_LOCK(_rt) do {						\
347 	if (!rte_debug)							\
348 		lck_mtx_lock(&(_rt)->rt_lock);				\
349 	else								\
350 		rt_lock(_rt, FALSE);					\
351 } while (0)
352 
353 #define	RT_LOCK_SPIN(_rt) do {						\
354 	if (!rte_debug)							\
355 		lck_mtx_lock_spin(&(_rt)->rt_lock);			\
356 	else								\
357 		rt_lock(_rt, TRUE);					\
358 } while (0)
359 
360 #define	RT_CONVERT_LOCK(_rt) do {					\
361 	RT_LOCK_ASSERT_HELD(_rt);					\
362 	lck_mtx_convert_spin(&(_rt)->rt_lock);				\
363 } while (0)
364 
365 #define	RT_UNLOCK(_rt) do {						\
366 	if (!rte_debug)							\
367 		lck_mtx_unlock(&(_rt)->rt_lock);			\
368 	else								\
369 		rt_unlock(_rt);						\
370 } while (0)
371 
372 #define	RT_ADDREF_LOCKED(_rt) do {					\
373 	if (!rte_debug) {						\
374 		RT_LOCK_ASSERT_HELD(_rt);				\
375 		if (++(_rt)->rt_refcnt == 0)				\
376 			panic("RT_ADDREF(%p) bad refcnt\n", _rt);	\
377 	} else {							\
378 		rtref(_rt);						\
379 	}								\
380 } while (0)
381 
382 /*
383  * Spin variant mutex is used here; caller is responsible for
384  * converting any previously-held similar lock to full mutex.
385  */
386 #define	RT_ADDREF(_rt) do {						\
387 	RT_LOCK_SPIN(_rt);						\
388 	RT_ADDREF_LOCKED(_rt);						\
389 	RT_UNLOCK(_rt);							\
390 } while (0)
391 
392 #define	RT_REMREF_LOCKED(_rt) do {					\
393 	if (!rte_debug) {						\
394 		RT_LOCK_ASSERT_HELD(_rt);				\
395 		if ((_rt)->rt_refcnt == 0)				\
396 			panic("RT_REMREF(%p) bad refcnt\n", _rt);	\
397 		--(_rt)->rt_refcnt;					\
398 	} else {							\
399 		(void) rtunref(_rt);					\
400 	}								\
401 } while (0)
402 
403 /*
404  * Spin variant mutex is used here; caller is responsible for
405  * converting any previously-held similar lock to full mutex.
406  */
407 #define	RT_REMREF(_rt) do {						\
408 	RT_LOCK_SPIN(_rt);						\
409 	RT_REMREF_LOCKED(_rt);						\
410 	RT_UNLOCK(_rt);							\
411 } while (0)
412 
413 #define RTFREE(_rt)		rtfree(_rt)
414 #define RTFREE_LOCKED(_rt)	rtfree_locked(_rt)
415 
416 extern struct route_cb route_cb;
417 extern struct radix_node_head *rt_tables[AF_MAX+1];
418 __private_extern__ lck_mtx_t *rnh_lock;
419 __private_extern__ int use_routegenid;
420 __private_extern__ uint32_t route_generation;
421 __private_extern__ int rttrash;
422 __private_extern__ unsigned int rte_debug;
423 
424 struct ifmultiaddr;
425 struct proc;
426 
427 extern void route_init(void) __attribute__((section("__TEXT, initcode")));
428 extern void routegenid_update(void);
429 extern void rt_ifmsg(struct ifnet *);
430 extern void rt_missmsg(int, struct rt_addrinfo *, int, int);
431 extern void rt_newaddrmsg(int, struct ifaddr *, int, struct rtentry *);
432 extern void rt_newmaddrmsg(int, struct ifmultiaddr *);
433 extern int rt_setgate(struct rtentry *, struct sockaddr *, struct sockaddr *);
434 extern void set_primary_ifscope(unsigned int);
435 extern unsigned int get_primary_ifscope(void);
436 extern boolean_t rt_inet_default(struct rtentry *, struct sockaddr *);
437 extern struct rtentry *rt_lookup(boolean_t, struct sockaddr *,
438     struct sockaddr *, struct radix_node_head *, unsigned int);
439 extern void rtalloc(struct route *);
440 extern void rtalloc_ign(struct route *, uint32_t);
441 extern void rtalloc_ign_locked(struct route *, uint32_t);
442 extern void rtalloc_scoped_ign(struct route *, uint32_t, unsigned int);
443 extern void rtalloc_scoped_ign_locked(struct route *, uint32_t, unsigned int);
444 extern struct rtentry *rtalloc1(struct sockaddr *, int, uint32_t);
445 extern struct rtentry *rtalloc1_locked(struct sockaddr *, int, uint32_t);
446 extern struct rtentry *rtalloc1_scoped(struct sockaddr *, int, uint32_t,
447     unsigned int);
448 extern struct rtentry *rtalloc1_scoped_locked(struct sockaddr *, int,
449     uint32_t, unsigned int);
450 extern void rtfree(struct rtentry *);
451 extern void rtfree_locked(struct rtentry *);
452 extern void rtref(struct rtentry *);
453 /*
454  * rtunref will decrement the refcount, rtfree will decrement and free if
455  * the refcount has reached zero and the route is not up.
456  * Unless you have good reason to do otherwise, use rtfree.
457  */
458 extern int rtunref(struct rtentry *);
459 extern void rtsetifa(struct rtentry *, struct ifaddr *);
460 extern int rtinit(struct ifaddr *, int, int);
461 extern int rtinit_locked(struct ifaddr *, int, int);
462 extern int rtioctl(unsigned long, caddr_t, struct proc *);
463 extern void rtredirect(struct ifnet *, struct sockaddr *, struct sockaddr *,
464     struct sockaddr *, int, struct sockaddr *, struct rtentry **);
465 extern int rtrequest(int, struct sockaddr *,
466     struct sockaddr *, struct sockaddr *, int, struct rtentry **);
467 extern int rtrequest_locked(int, struct sockaddr *,
468     struct sockaddr *, struct sockaddr *, int, struct rtentry **);
469 extern int rtrequest_scoped_locked(int, struct sockaddr *, struct sockaddr *,
470     struct sockaddr *, int, struct rtentry **, unsigned int);
471 extern unsigned int sa_get_ifscope(struct sockaddr *);
472 extern void rt_lock(struct rtentry *, boolean_t);
473 extern void rt_unlock(struct rtentry *);
474 extern struct sockaddr *rtm_scrub_ifscope(int, struct sockaddr *,
475     struct sockaddr *, struct sockaddr_storage *);
476 #endif /* KERNEL_PRIVATE */
477 
478 #endif
479