xref: /xnu-10002.61.3/bsd/net/route_private.h (revision 0f4c859e951fba394238ab619495c4e1d54d0f34)
1 /*
2  * Copyright (c) 2000-2022 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_PRIVATE_H_
65 #define _NET_ROUTE_PRIVATE_H_
66 #include <net/route.h>
67 #include <stdint.h>
68 #include <sys/types.h>
69 #include <sys/socket.h>
70 #include <uuid/uuid.h>
71 
72 struct route_old {
73 	void            *ro_rt;
74 	uint32_t        ro_flags;
75 	struct sockaddr ro_dst;
76 };
77 
78 #ifdef BSD_KERNEL_PRIVATE
79 #include <kern/locks.h>
80 #include <net/radix.h>
81 #include <sys/eventhandler.h>
82 #include <net/if_dl.h>
83 #include <netinet/in_private.h>
84 
85 extern boolean_t trigger_v6_defrtr_select;
86 /*
87  * Kernel resident routing tables.
88  *
89  * The routing tables are initialized when interface addresses
90  * are set by making entries for all directly connected interfaces.
91  */
92 
93 /* forward declarations */
94 struct ifnet_llreach_info;
95 struct rt_reach_info;
96 
97 /*
98  * IP route structure
99  *
100  * A route consists of a destination address and a reference
101  * to a routing entry.  These are often held by protocols
102  * in their control blocks, e.g. inpcb.
103  */
104 struct route {
105 	/*
106 	 * N.B: struct route must begin with ro_{rt, lle, srcia, flags}
107 	 * because the code does some casts of a 'struct route_in6 *'
108 	 * to a 'struct route *'.
109 	 */
110 	struct rtentry        *ro_rt;
111 	struct ifaddr         *ro_srcia;
112 	uint32_t              ro_flags;       /* route flags (see below) */
113 #if __has_ptrcheck
114 	struct sockaddr_in    ro_dst;
115 #else
116 	struct sockaddr       ro_dst;
117 #endif
118 };
119 
120 #define ROF_SRCIF_SELECTED      0x0001  /* source interface was selected */
121 
122 #define ROUTE_UNUSABLE(_ro)                                             \
123 	((_ro)->ro_rt == NULL ||                                        \
124 	((_ro)->ro_rt->rt_flags & (RTF_UP|RTF_CONDEMNED)) != RTF_UP ||  \
125 	RT_GENID_OUTOFSYNC((_ro)->ro_rt))
126 
127 #define _ROUTE_RELEASE_COMMON(_ro, _rnh_locked) do {                    \
128 	if ((_ro)->ro_rt != NULL) {                                     \
129 	        RT_LOCK_ASSERT_NOTHELD((_ro)->ro_rt);                   \
130 	        if (_rnh_locked)                                        \
131 	                rtfree_locked((_ro)->ro_rt);                    \
132 	        else                                                    \
133 	                rtfree((_ro)->ro_rt);                           \
134 	        (_ro)->ro_rt = NULL;                                    \
135 	}                                                               \
136 	if ((_ro)->ro_srcia != NULL) {                                  \
137 	        IFA_REMREF((_ro)->ro_srcia);                            \
138 	        (_ro)->ro_srcia = NULL;                                 \
139 	        (_ro)->ro_flags &= ~ROF_SRCIF_SELECTED;                 \
140 	}                                                               \
141 } while (0)
142 
143 #define ROUTE_RELEASE_LOCKED(_ro)       _ROUTE_RELEASE_COMMON(_ro, TRUE)
144 #define ROUTE_RELEASE(_ro)              _ROUTE_RELEASE_COMMON(_ro, FALSE)
145 
146 /*
147  * We distinguish between routes to hosts and routes to networks,
148  * preferring the former if available.  For each route we infer
149  * the interface to use from the gateway address supplied when
150  * the route was entered.  Routes that forward packets through
151  * gateways are marked so that the output routines know to address the
152  * gateway rather than the ultimate destination.
153  */
154 
155 #define NRTT_HIST       10
156 /*
157  * Kernel routing entry structure.
158  */
159 struct rtentry {
160 	struct  radix_node rt_nodes[2]; /* tree glue, and other values */
161 #define rt_key(r)       (SA((r)->rt_nodes->rn_key))
162 #define rt_mask(r)      (SA((r)->rt_nodes->rn_mask))
163 	/*
164 	 * See bsd/net/route.c for synchronization notes.
165 	 */
166 	decl_lck_mtx_data(, rt_lock);   /* lock for routing entry */
167 	uint32_t rt_refcnt;             /* # held references */
168 	uint32_t rt_flags;              /* up/down?, host/net */
169 	uint32_t rt_genid;              /* route generation id */
170 	struct sockaddr *rt_gateway;    /* value */
171 	struct ifnet *rt_ifp;           /* the answer: interface to use */
172 	struct ifaddr *rt_ifa;          /* the answer: interface addr to use */
173 	struct sockaddr *rt_genmask;    /* for generation of cloned routes */
174 	void *rt_llinfo;                /* pointer to link level info cache */
175 	void (*rt_llinfo_get_ri)        /* llinfo get reachability info fn */
176 	(struct rtentry *, struct rt_reach_info *);
177 	void (*rt_llinfo_get_iflri)     /* ifnet llinfo get reach. info fn */
178 	(struct rtentry *, struct ifnet_llreach_info *);
179 	void (*rt_llinfo_purge)(struct rtentry *); /* llinfo purge fn */
180 	void (*rt_llinfo_free)(void *); /* link level info free function */
181 	void (*rt_llinfo_refresh) (struct rtentry *); /* expedite llinfo refresh */
182 	struct rt_metrics rt_rmx;       /* metrics used by rx'ing protocols */
183 #define rt_use rt_rmx.rmx_pksent
184 	struct rtentry *rt_gwroute;     /* implied entry for gatewayed routes */
185 	struct rtentry *rt_parent;      /* cloning parent of this route */
186 	struct nstat_counts *rt_stats;  /* route stats */
187 	void (*rt_if_ref_fn)(struct ifnet *, int); /* interface ref func */
188 
189 	uint32_t *rt_tree_genid;        /* ptr to per-tree route_genid */
190 	uint64_t rt_expire;             /* expiration time in uptime seconds */
191 	uint64_t base_calendartime;     /* calendar time upon entry creation */
192 	uint64_t base_uptime;           /* uptime upon entry creation */
193 	u_int32_t rtt_hist[NRTT_HIST];  /* RTT history sample by TCP connections */
194 	u_int32_t rtt_min;              /* minimum RTT computed from history */
195 	u_int32_t rtt_expire_ts;        /* RTT history expire timestamp */
196 	u_int8_t rtt_index;             /* Index into RTT history */
197 	/* Event handler context for the rtentrt */
198 	struct eventhandler_lists_ctxt rt_evhdlr_ctxt;
199 };
200 
201 #define rt_key_free(r) ({ \
202 	void *__r = rt_key(r); \
203 	kheap_free_addr(KHEAP_DATA_BUFFERS, __r); \
204 })
205 
206 enum {
207 	ROUTE_STATUS_UPDATE = 1,
208 	ROUTE_ENTRY_REFRESH,
209 	ROUTE_ENTRY_DELETED,
210 	ROUTE_LLENTRY_RESOLVED,
211 	ROUTE_LLENTRY_UNREACH,
212 	ROUTE_LLENTRY_CHANGED,
213 	ROUTE_LLENTRY_STALE,
214 	ROUTE_LLENTRY_TIMEDOUT,
215 	ROUTE_LLENTRY_DELETED,
216 	ROUTE_LLENTRY_EXPIRED,
217 	ROUTE_LLENTRY_PROBED,
218 	ROUTE_EVHDLR_DEREGISTER,
219 };
220 
221 extern const char * route_event2str(int route_event);
222 
223 typedef void (*route_event_fn) (struct eventhandler_entry_arg,
224     struct sockaddr *, int, struct sockaddr *, int);
225 EVENTHANDLER_DECLARE(route_event, route_event_fn);
226 
227 /*
228  * Synchronize route entry's generation ID with the tree's.
229  */
230 #define RT_GENID_SYNC(_rt) do {                                         \
231 	if ((_rt)->rt_tree_genid != NULL)                               \
232 	        (_rt)->rt_genid = *(_rt)->rt_tree_genid;                \
233 } while (0)
234 
235 /*
236  * Indicates whether or not the route entry's generation ID is stale.
237  */
238 #define RT_GENID_OUTOFSYNC(_rt)                                         \
239 	((_rt)->rt_tree_genid != NULL &&                                \
240 	*(_rt)->rt_tree_genid != (_rt)->rt_genid)
241 
242 enum {
243 	ROUTE_OP_READ,
244 	ROUTE_OP_WRITE,
245 };
246 
247 extern int route_op_entitlement_check(struct socket *, kauth_cred_t, int, boolean_t);
248 #endif /* BSD_KERNEL_PRIVATE */
249 
250 struct kev_netevent_apnfallbk_data {
251 	pid_t           epid;           /* effective PID */
252 	uuid_t          euuid;          /* effective UUID */
253 };
254 
255 /*
256  * Route reachability info.
257  */
258 struct rt_reach_info {
259 	u_int32_t       ri_refcnt;      /* reference count */
260 	u_int32_t       ri_probes;      /* total # of probes */
261 	u_int64_t       ri_snd_expire;  /* tx expiration (calendar) time */
262 	u_int64_t       ri_rcv_expire;  /* rx expiration (calendar) time */
263 	int32_t         ri_rssi;        /* received signal strength */
264 	int32_t         ri_lqm;         /* link quality metric */
265 	int32_t         ri_npm;         /* node proximity metric */
266 };
267 
268 /*
269  * Extended routing message header (private).
270  */
271 struct rt_msghdr_ext {
272 	u_short rtm_msglen;     /* to skip over non-understood messages */
273 	u_char  rtm_version;    /* future binary compatibility */
274 	u_char  rtm_type;       /* message type */
275 	u_int32_t rtm_index;    /* index for associated ifp */
276 	u_int32_t rtm_flags;    /* flags, incl. kern & message, e.g. DONE */
277 	u_int32_t rtm_reserved; /* for future use */
278 	u_int32_t rtm_addrs;    /* bitmask identifying sockaddrs in msg */
279 	pid_t   rtm_pid;        /* identify sender */
280 	int     rtm_seq;        /* for sender to identify action */
281 	int     rtm_errno;      /* why failed */
282 	u_int32_t rtm_use;      /* from rtentry */
283 	u_int32_t rtm_inits;    /* which metrics we are initializing */
284 	struct rt_metrics rtm_rmx;      /* metrics themselves */
285 	struct rt_reach_info rtm_ri;    /* route reachability info */
286 };
287 
288 /*
289  * Message types.
290  */
291 #define RTM_GET_SILENT  0x11
292 #define RTM_GET_EXT     0x15
293 
294 /*
295  * Bitmask values for rtm_inits and rmx_locks.
296  */
297 #define RTV_REFRESH_HOST        0x100   /* init host route to expedite refresh */
298 
299 /*
300  * For scoped routing; a zero interface scope value means nil/no scope.
301  */
302 #define IFSCOPE_NONE    0
303 #define IFSCOPE_UNKNOWN IFSCOPE_NONE
304 
305 #ifdef BSD_KERNEL_PRIVATE
306 /*
307  * Generic call trace used by some subsystems (e.g. route, ifaddr)
308  */
309 #define CTRACE_STACK_SIZE       8               /* depth of stack trace */
310 #define CTRACE_HIST_SIZE        4               /* refcnt history size */
311 typedef struct ctrace {
312 	void    *th;                            /* thread ptr */
313 	void    *pc[CTRACE_STACK_SIZE];         /* PC stack trace */
314 } ctrace_t;
315 
316 extern void ctrace_record(ctrace_t *);
317 
318 #define RT_LOCK_ASSERT_HELD(_rt)                                        \
319 	LCK_MTX_ASSERT(&(_rt)->rt_lock, LCK_MTX_ASSERT_OWNED)
320 
321 #define RT_LOCK_ASSERT_NOTHELD(_rt)                                     \
322 	LCK_MTX_ASSERT(&(_rt)->rt_lock, LCK_MTX_ASSERT_NOTOWNED)
323 
324 #define RT_LOCK(_rt) do {                                               \
325 	rt_lock(_rt, FALSE);                                            \
326 } while (0)
327 
328 #define RT_LOCK_SPIN(_rt) do {                                          \
329 	rt_lock(_rt, TRUE);                                             \
330 } while (0)
331 
332 #define RT_CONVERT_LOCK(_rt) do {                                       \
333 	RT_LOCK_ASSERT_HELD(_rt);                                       \
334 	lck_mtx_convert_spin(&(_rt)->rt_lock);                          \
335 } while (0)
336 
337 #define RT_UNLOCK(_rt) do {                                             \
338 	rt_unlock(_rt);                                                 \
339 } while (0)
340 
341 #define RT_ADDREF_LOCKED(_rt) do {                                      \
342 	rtref(_rt);                                                     \
343 } while (0)
344 
345 /*
346  * Spin variant mutex is used here; caller is responsible for
347  * converting any previously-held similar lock to full mutex.
348  */
349 #define RT_ADDREF(_rt) do {                                             \
350 	RT_LOCK_SPIN(_rt);                                              \
351 	RT_ADDREF_LOCKED(_rt);                                          \
352 	RT_UNLOCK(_rt);                                                 \
353 } while (0)
354 
355 #define RT_REMREF_LOCKED(_rt) do {                                      \
356 	(void) rtunref(_rt);                                            \
357 } while (0)
358 
359 /*
360  * Spin variant mutex is used here; caller is responsible for
361  * converting any previously-held similar lock to full mutex.
362  */
363 #define RT_REMREF(_rt) do {                                             \
364 	RT_LOCK_SPIN(_rt);                                              \
365 	RT_REMREF_LOCKED(_rt);                                          \
366 	RT_UNLOCK(_rt);                                                 \
367 } while (0)
368 
369 /*
370  * This macro calculates skew in wall clock, just in case the user changes the
371  * system time. This skew adjustment is required because we now keep the
372  * expiration times in uptime terms in the kernel, but the userland still
373  * expects expiration times in terms of calendar times.  This is used when
374  * reporting rt_expire, ln_expire, etc. values to user space.
375  */
376 #define NET_CALCULATE_CLOCKSKEW(cc, ic, cu, iu)                         \
377 	((cc.tv_sec - ic) - (cu - iu))
378 
379 extern unsigned int rt_verbose;
380 extern struct radix_node_head *rt_tables[AF_MAX + 1];
381 extern lck_mtx_t rnh_lock_data;
382 #define rnh_lock (&rnh_lock_data)
383 extern uint32_t route_genid_inet;       /* INET route generation count */
384 extern uint32_t route_genid_inet6;      /* INET6 route generation count */
385 extern int rttrash;
386 extern unsigned int rte_debug;
387 
388 struct ifmultiaddr;
389 struct proc;
390 
391 extern void route_init(void);
392 extern void routegenid_update(void);
393 extern void routegenid_inet_update(void);
394 extern void routegenid_inet6_update(void);
395 extern void rt_ifmsg(struct ifnet *);
396 extern void rt_missmsg(u_char, struct rt_addrinfo *, int, int);
397 extern void rt_newaddrmsg(u_char, struct ifaddr *, int, struct rtentry *);
398 extern void rt_newmaddrmsg(u_char, struct ifmultiaddr *);
399 extern int rt_setgate(struct rtentry *, struct sockaddr *, struct sockaddr *);
400 extern void set_primary_ifscope(int, unsigned int);
401 extern unsigned int get_primary_ifscope(int);
402 extern boolean_t rt_primary_default(struct rtentry *, struct sockaddr *);
403 extern struct rtentry *rt_lookup(boolean_t, struct sockaddr *,
404     struct sockaddr *, struct radix_node_head *, unsigned int);
405 extern struct rtentry *rt_lookup_coarse(boolean_t, struct sockaddr *,
406     struct sockaddr *, struct radix_node_head *);
407 extern void rtalloc(struct route *);
408 extern void rtalloc_scoped(struct route *, unsigned int);
409 extern void rtalloc_ign(struct route *, uint32_t);
410 extern void rtalloc_scoped_ign(struct route *, uint32_t, unsigned int);
411 extern struct rtentry *rtalloc1(struct sockaddr *, int, uint32_t);
412 extern struct rtentry *rtalloc1_scoped(struct sockaddr *, int, uint32_t,
413     unsigned int);
414 extern struct rtentry *rtalloc1_scoped_locked(struct sockaddr *, int,
415     uint32_t, unsigned int);
416 extern void rtfree_locked(struct rtentry *);
417 extern void rtfree(struct rtentry *);
418 extern void rtref(struct rtentry *);
419 /*
420  * rtunref will decrement the refcount, rtfree will decrement and free if
421  * the refcount has reached zero and the route is not up.
422  * Unless you have good reason to do otherwise, use rtfree.
423  */
424 extern int rtunref(struct rtentry *);
425 extern void rtsetifa(struct rtentry *, struct ifaddr *);
426 extern int rtinit(struct ifaddr *, uint8_t, int);
427 extern int rtinit_locked(struct ifaddr *, uint8_t, int);
428 extern int rtioctl(unsigned long, caddr_t, struct proc *);
429 extern void rtredirect(struct ifnet *, struct sockaddr *, struct sockaddr *,
430     struct sockaddr *, int, struct sockaddr *, struct rtentry **);
431 extern int rtrequest(int, struct sockaddr *,
432     struct sockaddr *, struct sockaddr *, int, struct rtentry **);
433 extern int rtrequest_scoped(int, struct sockaddr *, struct sockaddr *,
434     struct sockaddr *, int, struct rtentry **, unsigned int);
435 extern int rtrequest_locked(int, struct sockaddr *,
436     struct sockaddr *, struct sockaddr *, int, struct rtentry **);
437 extern int rtrequest_scoped_locked(int, struct sockaddr *, struct sockaddr *,
438     struct sockaddr *, int, struct rtentry **, unsigned int);
439 extern void sin_set_ifscope(struct sockaddr *, unsigned int);
440 extern unsigned int sin_get_ifscope(struct sockaddr *);
441 extern unsigned int sin6_get_ifscope(struct sockaddr *);
442 extern void rt_lock(struct rtentry *, boolean_t);
443 extern void rt_unlock(struct rtentry *);
444 extern struct sockaddr *rtm_scrub(int, int, struct sockaddr *,
445     struct sockaddr *, void *, uint32_t, kauth_cred_t *);
446 extern boolean_t rt_validate(struct rtentry *);
447 extern void rt_set_proxy(struct rtentry *, boolean_t);
448 extern void rt_set_gwroute(struct rtentry *, struct sockaddr *,
449     struct rtentry *);
450 extern void rt_revalidate_gwroute(struct rtentry *, struct rtentry *);
451 extern errno_t route_to_gwroute(const struct sockaddr *, struct rtentry *,
452     struct rtentry **);
453 extern void rt_setexpire(struct rtentry *, uint64_t);
454 extern void rt_str(struct rtentry *, char *, uint32_t, char *, uint32_t);
455 extern const char *rtm2str(int);
456 extern void route_clear(struct route *);
457 extern void route_copyin(struct route *, struct route *, size_t);
458 extern void route_copyout(struct route *, const struct route *, size_t);
459 extern boolean_t rt_ifa_is_dst(struct sockaddr *, struct ifaddr *);
460 extern struct sockaddr *sa_copy(struct sockaddr *, struct sockaddr_storage *,
461     unsigned int *);
462 
463 /*
464  * The following is used to enqueue work items for route events
465  * and also used to pass route event while walking the tree
466  */
467 struct route_event {
468 	struct rtentry *rt;
469 	/*
470 	 * There's no reference taken on gwrt.
471 	 * We only use it to check whether we should
472 	 * point to rt_gateway or the embedded rt_addr
473 	 * structure.
474 	 */
475 	struct rtentry *gwrt;
476 	union {
477 		union sockaddr_in_4_6 _rtev_ipaddr;
478 		char _rtev_addr_bytes[DLIL_SDLMAXLEN];
479 	} rt_addr;
480 	uint32_t route_event_code;
481 	eventhandler_tag evtag;
482 };
483 
484 #define rtev_ipaddr     rt_addr._rtev_ipaddr
485 #define rtev_addr_bytes rt_addr._rtev_addr_bytes
486 
487 extern void route_event_init(struct route_event *p_route_ev, struct rtentry *rt,
488     struct rtentry *gwrt, int route_ev_code);
489 extern int route_event_walktree(struct radix_node *rn, void *arg);
490 extern void route_event_enqueue_nwk_wq_entry(struct rtentry *, struct rtentry *,
491     uint32_t, eventhandler_tag, boolean_t);
492 #endif /* BSD_KERNEL_PRIVATE */
493 #endif /* _NET_ROUTE_PRIVATE_H_ */
494