xref: /titanic_50/usr/src/uts/common/inet/ip_ire.h (revision 3c112a2b34403220c06c3e2fcac403358cfba168)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License (the "License").
6  * You may not use this file except in compliance with the License.
7  *
8  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9  * or http://www.opensolaris.org/os/licensing.
10  * See the License for the specific language governing permissions
11  * and limitations under the License.
12  *
13  * When distributing Covered Code, include this CDDL HEADER in each
14  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15  * If applicable, add the following below this CDDL HEADER, with the
16  * fields enclosed by brackets "[]" replaced with your own identifying
17  * information: Portions Copyright [yyyy] [name of copyright owner]
18  *
19  * CDDL HEADER END
20  */
21 /*
22  * Copyright 2010 Sun Microsystems, Inc.  All rights reserved.
23  * Use is subject to license terms.
24  */
25 /* Copyright (c) 1990 Mentat Inc. */
26 
27 #ifndef	_INET_IP_IRE_H
28 #define	_INET_IP_IRE_H
29 
30 #ifdef	__cplusplus
31 extern "C" {
32 #endif
33 
34 #define	IPV6_LL_PREFIXLEN	10	/* Number of bits in link-local pref */
35 
36 #define	IP_CACHE_TABLE_SIZE	256
37 #define	IP_MASK_TABLE_SIZE	(IP_ABITS + 1)		/* 33 ptrs */
38 
39 #define	IP6_FTABLE_HASH_SIZE	32	/* size of each hash table in ptrs */
40 #define	IP6_CACHE_TABLE_SIZE	256
41 #define	IP6_MASK_TABLE_SIZE	(IPV6_ABITS + 1)	/* 129 ptrs */
42 
43 /*
44  * We use the common modulo hash function.  In ip_ire_init(), we make
45  * sure that the cache table size is always a power of 2.  That's why
46  * we can use & instead of %.  Also note that we try hard to make sure
47  * the lower bits of an address capture most info from the whole address.
48  * The reason being that since our hash table is probably a lot smaller
49  * than 2^32 buckets so the lower bits are the most important.
50  */
51 #define	IRE_ADDR_HASH(addr, table_size) \
52 	(((addr) ^ ((addr) >> 8) ^ ((addr) >> 16) ^ ((addr) >> 24)) &	\
53 	((table_size) - 1))
54 
55 /*
56  * To make a byte-order neutral hash for IPv6, just take all the
57  * bytes in the bottom 32 bits into account.
58  */
59 #define	IRE_ADDR_HASH_V6(addr, table_size) 				\
60 	IRE_ADDR_HASH((addr).s6_addr32[3], table_size)
61 
62 /*
63  * This assumes that the ftable size is a power of 2.
64  * We include some high-order bytes to avoid all IRE_LOCALs in the same
65  * bucket for performance reasons.
66  */
67 #define	IRE_ADDR_MASK_HASH_V6(addr, mask, table_size) 			\
68 	((((addr).s6_addr8[0] & (mask).s6_addr8[0]) ^ 			\
69 	((addr).s6_addr8[1] & (mask).s6_addr8[1]) ^ 			\
70 	((addr).s6_addr8[6] & (mask).s6_addr8[6]) ^ 			\
71 	((addr).s6_addr8[7] & (mask).s6_addr8[7]) ^ 			\
72 	((addr).s6_addr8[8] & (mask).s6_addr8[8]) ^ 			\
73 	((addr).s6_addr8[9] & (mask).s6_addr8[9]) ^			\
74 	((addr).s6_addr8[10] & (mask).s6_addr8[10]) ^ 			\
75 	((addr).s6_addr8[13] & (mask).s6_addr8[13]) ^ 			\
76 	((addr).s6_addr8[14] & (mask).s6_addr8[14]) ^ 			\
77 	((addr).s6_addr8[15] & (mask).s6_addr8[15])) & ((table_size) - 1))
78 
79 #define	IRE_HIDDEN_TYPE(ire_type) ((ire_type) &			\
80 	(IRE_HOST | IRE_PREFIX | IRE_DEFAULT | IRE_IF_ALL | IRE_BROADCAST))
81 
82 /*
83  * match parameter definitions for IRE lookup routines.
84  */
85 #define	MATCH_IRE_DSTONLY	0x0000	/* Match just the address */
86 #define	MATCH_IRE_TYPE		0x0001	/* Match IRE type */
87 #define	MATCH_IRE_MASK		0x0002	/* Match IRE mask */
88 #define	MATCH_IRE_SHORTERMASK	0x0004	/* A mask shorter than the argument */
89 #define	MATCH_IRE_GW		0x0008	/* Match IRE gateway */
90 #define	MATCH_IRE_ILL		0x0010	/* Match IRE on the ill */
91 #define	MATCH_IRE_ZONEONLY	0x0020	/* Match IREs in specified zone, ie */
92 					/* don't match IRE_LOCALs from other */
93 					/* zones or shared IREs */
94 #define	MATCH_IRE_SECATTR	0x0040	/* Match gateway security attributes */
95 #define	MATCH_IRE_TESTHIDDEN 	0x0080	/* Match ire_testhidden IREs */
96 #define	MATCH_IRE_SRC_ILL	0x0100	/* ire_ill uses a src address on ill */
97 #define	MATCH_IRE_DIRECT	0x0200	/* Don't match indirect routes */
98 
99 #define	MAX_IRE_RECURSION	4	/* Max IREs in ire_route_recursive */
100 
101 
102 /*
103  * We use atomics so that we get an accurate accounting on the ires.
104  * Otherwise we can't determine leaks correctly.
105  */
106 #define	BUMP_IRE_STATS(ire_stats, x) atomic_add_64(&(ire_stats).x, 1)
107 
108 #ifdef _KERNEL
109 struct ts_label_s;
110 struct nce_s;
111 /*
112  * structure for passing args between ire_ftable_lookup and ire_find_best_route
113  */
114 typedef struct ire_ftable_args_s {
115 	in6_addr_t		ift_addr_v6;
116 	in6_addr_t		ift_mask_v6;
117 	in6_addr_t		ift_gateway_v6;
118 #define	ift_addr		V4_PART_OF_V6(ift_addr_v6)
119 #define	ift_mask		V4_PART_OF_V6(ift_mask_v6)
120 #define	ift_gateway		V4_PART_OF_V6(ift_gateway_v6)
121 	int			ift_type;
122 	const ill_t		*ift_ill;
123 	zoneid_t		ift_zoneid;
124 	const ts_label_t	*ift_tsl;
125 	int			ift_flags;
126 	ire_t			*ift_best_ire;
127 } ire_ftable_args_t;
128 
129 extern	ipaddr_t	ip_plen_to_mask(uint_t);
130 extern	in6_addr_t	*ip_plen_to_mask_v6(uint_t, in6_addr_t *);
131 
132 extern	int	ip_ire_advise(queue_t *, mblk_t *, cred_t *);
133 extern	int	ip_ire_delete(queue_t *, mblk_t *, cred_t *);
134 extern	void	ip_ire_reclaim(void *);
135 
136 extern	int	ip_mask_to_plen(ipaddr_t);
137 extern	int	ip_mask_to_plen_v6(const in6_addr_t *);
138 
139 extern	ire_t	*ire_add(ire_t *);
140 extern	ire_t	*ire_add_v6(ire_t *);
141 extern	int	ire_atomic_start(irb_t *irb_ptr, ire_t *ire);
142 extern	void	ire_atomic_end(irb_t *irb_ptr, ire_t *ire);
143 
144 extern	ire_t	*ire_create(uchar_t *, uchar_t *, uchar_t *,
145     ushort_t, ill_t *, zoneid_t, uint_t, tsol_gc_t *, ip_stack_t *);
146 
147 extern	ire_t	**ire_create_bcast(ill_t *, ipaddr_t, zoneid_t, ire_t **);
148 extern	ire_t	*ire_create_if_clone(ire_t *, const in6_addr_t *, uint_t *);
149 extern	ire_t	*ire_lookup_bcast(ill_t *, ipaddr_t, zoneid_t);
150 extern	int	ire_init_v4(ire_t *, uchar_t *, uchar_t *, uchar_t *,
151     ushort_t, ill_t *, zoneid_t, uint_t, tsol_gc_t *, ip_stack_t *);
152 extern	int	ire_init_v6(ire_t *, const in6_addr_t *, const in6_addr_t *,
153     const in6_addr_t *, ushort_t, ill_t *, zoneid_t, uint_t, tsol_gc_t *,
154     ip_stack_t *);
155 
156 extern	int	ire_init_common(ire_t *, ushort_t, ill_t *, zoneid_t, uint_t,
157     uchar_t, tsol_gc_t *, ip_stack_t *);
158 
159 extern	ire_t	*ire_create_v6(const in6_addr_t *, const in6_addr_t *,
160     const in6_addr_t *, ushort_t, ill_t *, zoneid_t, uint_t,
161     tsol_gc_t *, ip_stack_t *);
162 
163 extern	void	ire_delete(ire_t *);
164 extern	void	ire_delete_v6(ire_t *);
165 
166 extern	boolean_t ire_dep_build(ire_t *[], uint_t [], uint_t);
167 extern	void	ire_dep_delete_if_clone(ire_t *);
168 extern	void	ire_dep_incr_generation(ire_t *);
169 extern	void	ire_dep_remove(ire_t *);
170 extern	void	ire_dep_unbuild(ire_t *[], uint_t);
171 extern	uint_t	ire_dep_validate_generations(ire_t *);
172 extern	void	ire_dep_invalidate_generations(ire_t *);
173 extern	boolean_t ire_determine_nce_capable(ire_t *);
174 
175 extern	void	ire_flush_cache_v4(ire_t *, int);
176 extern	void	ire_flush_cache_v6(ire_t *, int);
177 
178 extern	ire_t	*ire_ftable_lookup_v4(ipaddr_t, ipaddr_t, ipaddr_t, int,
179     const ill_t *, zoneid_t, const struct ts_label_s *, int, uint32_t,
180     ip_stack_t *, uint_t *);
181 extern	ire_t	*ire_ftable_lookup_v6(const in6_addr_t *, const in6_addr_t *,
182     const in6_addr_t *, int, const ill_t *, zoneid_t,
183     const struct ts_label_s *, int, uint32_t, ip_stack_t *, uint_t *);
184 
185 extern	ire_t	*ire_ftable_lookup_simple_v4(ipaddr_t, uint32_t, ip_stack_t *,
186     uint_t *);
187 extern	ire_t	*ire_ftable_lookup_simple_v6(const in6_addr_t *, uint32_t,
188     ip_stack_t *, uint_t *);
189 
190 extern boolean_t ire_gateway_ok_zone_v4(ipaddr_t, zoneid_t, ill_t *,
191     const ts_label_t *, ip_stack_t *, boolean_t);
192 extern boolean_t ire_gateway_ok_zone_v6(const in6_addr_t *, zoneid_t, ill_t *,
193     const ts_label_t *, ip_stack_t *, boolean_t);
194 
195 extern ire_t	*ire_alt_local(ire_t *, zoneid_t, const ts_label_t *,
196     const ill_t *, uint_t *);
197 
198 extern  ill_t	*ire_lookup_multi_ill_v4(ipaddr_t, zoneid_t, ip_stack_t *,
199     boolean_t *, ipaddr_t *);
200 extern  ill_t	*ire_lookup_multi_ill_v6(const in6_addr_t *, zoneid_t,
201     ip_stack_t *, boolean_t *, in6_addr_t *);
202 
203 extern	ire_t	*ire_nexthop(ire_t *);
204 extern	ill_t	*ire_nexthop_ill(ire_t *);
205 extern	ill_t	*ire_nce_ill(ire_t *);
206 
207 extern	ire_t	*ire_reject(ip_stack_t *, boolean_t);
208 extern	ire_t	*ire_blackhole(ip_stack_t *, boolean_t);
209 extern	ire_t	*ire_multicast(ill_t *);
210 
211 /* The different ire_recvfn functions */
212 extern void	ire_recv_forward_v4(ire_t *, mblk_t *, void *,
213     ip_recv_attr_t *);
214 extern void	ire_recv_noroute_v4(ire_t *, mblk_t *, void *,
215     ip_recv_attr_t *);
216 extern void	ire_recv_broadcast_v4(ire_t *, mblk_t *, void *,
217     ip_recv_attr_t *);
218 extern void	ire_recv_multicast_v4(ire_t *, mblk_t *, void *,
219     ip_recv_attr_t *);
220 extern void	ire_recv_multirt_v4(ire_t *, mblk_t *, void *,
221     ip_recv_attr_t *);
222 extern void	ire_recv_loopback_v4(ire_t *, mblk_t *, void *,
223     ip_recv_attr_t *);
224 extern void	ire_recv_local_v4(ire_t *, mblk_t *, void *,
225     ip_recv_attr_t *);
226 extern void	ire_recv_noaccept_v4(ire_t *, mblk_t *, void *,
227     ip_recv_attr_t *);
228 
229 extern void	ire_recv_forward_v6(ire_t *, mblk_t *, void *,
230     ip_recv_attr_t *);
231 extern void	ire_recv_noroute_v6(ire_t *, mblk_t *, void *,
232     ip_recv_attr_t *);
233 extern void	ire_recv_multicast_v6(ire_t *, mblk_t *, void *,
234     ip_recv_attr_t *);
235 extern void	ire_recv_multirt_v6(ire_t *, mblk_t *, void *,
236     ip_recv_attr_t *);
237 extern void	ire_recv_loopback_v6(ire_t *, mblk_t *, void *,
238     ip_recv_attr_t *);
239 extern void	ire_recv_local_v6(ire_t *, mblk_t *, void *, ip_recv_attr_t *);
240 extern void	ire_recv_noaccept_v6(ire_t *, mblk_t *, void *,
241     ip_recv_attr_t *);
242 
243 extern	void	irb_refhold(irb_t *);
244 extern	void	irb_refhold_locked(irb_t *);
245 extern	void	irb_refrele(irb_t *);
246 extern  void	irb_increment_generation(irb_t *);
247 
248 extern	void	ire_refhold(ire_t *);
249 extern	void	ire_refhold_notr(ire_t *);
250 extern	void	ire_refhold_locked(ire_t *);
251 extern	void	ire_refrele(ire_t *);
252 extern	void	ire_refrele_notr(ire_t *);
253 extern	void	ire_make_condemned(ire_t *);
254 extern	boolean_t ire_no_good(ire_t *);
255 extern	nce_t	*ire_handle_condemned_nce(nce_t *, ire_t *, ipha_t *, ip6_t *,
256     boolean_t);
257 
258 extern ire_t   	*ire_round_robin(irb_t *, ire_ftable_args_t *, uint_t,
259     ire_t *, ip_stack_t *);
260 
261 extern ire_t	*ire_route_recursive_v4(ipaddr_t, uint_t, const ill_t *,
262     zoneid_t, const ts_label_t *, uint_t, uint_t, uint32_t, ip_stack_t *,
263     ipaddr_t *, tsol_ire_gw_secattr_t **, uint_t *);
264 extern ire_t	*ire_route_recursive_v6(const in6_addr_t *, uint_t,
265     const ill_t *, zoneid_t, const ts_label_t *, uint_t, uint_t, uint32_t,
266     ip_stack_t *, in6_addr_t *, tsol_ire_gw_secattr_t **, uint_t *);
267 extern ire_t	*ire_route_recursive_dstonly_v4(ipaddr_t, uint_t,
268     uint32_t, ip_stack_t *);
269 extern ire_t	*ire_route_recursive_dstonly_v6(const in6_addr_t *, uint_t,
270     uint32_t, ip_stack_t *);
271 extern ire_t	*ire_route_recursive_impl_v4(ire_t *ire, ipaddr_t, uint_t,
272     const ill_t *, zoneid_t, const ts_label_t *, uint_t, uint_t, uint32_t,
273     ip_stack_t *, ipaddr_t *, tsol_ire_gw_secattr_t **, uint_t *);
274 extern ire_t	*ire_route_recursive_impl_v6(ire_t *ire, const in6_addr_t *,
275     uint_t, const ill_t *, zoneid_t, const ts_label_t *, uint_t, uint_t,
276     uint32_t, ip_stack_t *, in6_addr_t *, tsol_ire_gw_secattr_t **, uint_t *);
277 
278 /* The different ire_sendfn functions */
279 extern int	ire_send_local_v4(ire_t *, mblk_t *, void *,
280     ip_xmit_attr_t *, uint32_t *);
281 extern int	ire_send_multirt_v4(ire_t *, mblk_t *, void *,
282     ip_xmit_attr_t *, uint32_t *);
283 extern int	ire_send_noroute_v4(ire_t *, mblk_t *, void *,
284     ip_xmit_attr_t *, uint32_t *);
285 extern int	ire_send_multicast_v4(ire_t *, mblk_t *, void *,
286     ip_xmit_attr_t *, uint32_t *);
287 extern int	ire_send_broadcast_v4(ire_t *, mblk_t *, void *,
288     ip_xmit_attr_t *, uint32_t *);
289 extern int	ire_send_wire_v4(ire_t *, mblk_t *, void *,
290     ip_xmit_attr_t *, uint32_t *);
291 extern int	ire_send_local_v6(ire_t *, mblk_t *, void *,
292     ip_xmit_attr_t *, uint32_t *);
293 extern int	ire_send_multirt_v6(ire_t *, mblk_t *, void *,
294     ip_xmit_attr_t *, uint32_t *);
295 extern int	ire_send_noroute_v6(ire_t *, mblk_t *, void *,
296     ip_xmit_attr_t *, uint32_t *);
297 extern int	ire_send_multicast_v6(ire_t *, mblk_t *, void *,
298     ip_xmit_attr_t *, uint32_t *);
299 extern int	ire_send_wire_v6(ire_t *, mblk_t *, void *,
300     ip_xmit_attr_t *, uint32_t *);
301 
302 extern nce_t	*ire_to_nce_pkt(ire_t *, mblk_t *);
303 extern nce_t	*ire_to_nce(ire_t *, ipaddr_t, const in6_addr_t *);
304 
305 /* Different ire_postfragfn functions */
306 extern int	ip_xmit(mblk_t *, struct nce_s *,
307     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
308 extern int	ip_postfrag_loopcheck(mblk_t *, struct nce_s *,
309     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
310 extern int	ip_postfrag_multirt_v4(mblk_t *, struct nce_s *,
311     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
312 extern int	ip_postfrag_multirt_v6(mblk_t *, struct nce_s *,
313     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
314 
315 extern void	ip_postfrag_loopback(mblk_t *, struct nce_s *,
316     iaflags_t, uint_t, zoneid_t);
317 extern int	ire_revalidate_nce(ire_t *);
318 
319 extern ire_t	*ip_select_route_pkt(mblk_t *, ip_xmit_attr_t *,
320     uint_t *, int *, boolean_t *);
321 extern ire_t	*ip_select_route(const in6_addr_t *, const in6_addr_t,
322     ip_xmit_attr_t *, uint_t *, in6_addr_t *, int *, boolean_t *);
323 extern ire_t	*ip_select_route_v4(ipaddr_t, ipaddr_t, ip_xmit_attr_t *,
324     uint_t *, ipaddr_t *, int *, boolean_t *);
325 extern ire_t	*ip_select_route_v6(const in6_addr_t *, const in6_addr_t,
326     ip_xmit_attr_t *, uint_t *, in6_addr_t *, int *, boolean_t *);
327 
328 extern	void	ire_walk(pfv_t, void *, ip_stack_t *);
329 extern	void	ire_walk_ill(uint_t, uint_t, pfv_t, void *, ill_t *);
330 extern	void	ire_walk_v4(pfv_t, void *, zoneid_t, ip_stack_t *);
331 extern  void	ire_walk_ill_tables(uint_t match_flags, uint_t ire_type,
332     pfv_t func, void *arg, size_t ftbl_sz, size_t htbl_sz,
333     irb_t **ipftbl, ill_t *ill,
334     zoneid_t zoneid, ip_stack_t *);
335 extern	void	ire_walk_v6(pfv_t, void *, zoneid_t, ip_stack_t *);
336 
337 extern boolean_t	ire_match_args(ire_t *, ipaddr_t, ipaddr_t, ipaddr_t,
338     int, const ill_t *, zoneid_t, const struct ts_label_s *, int);
339 extern boolean_t	ire_match_args_v6(ire_t *, const in6_addr_t *,
340     const in6_addr_t *, const in6_addr_t *, int, const ill_t *, zoneid_t,
341     const ts_label_t *, int);
342 
343 extern  struct nce_s	*arp_nce_init(ill_t *, in_addr_t, int);
344 extern  boolean_t	ire_walk_ill_match(uint_t, uint_t, ire_t *, ill_t *,
345     zoneid_t, ip_stack_t *);
346 extern  void ire_increment_generation(ire_t *);
347 extern  void ire_increment_multicast_generation(ip_stack_t *, boolean_t);
348 extern	void ire_rebind(ire_t *);
349 
350 #endif /* _KERNEL */
351 
352 #ifdef	__cplusplus
353 }
354 #endif
355 
356 #endif	/* _INET_IP_IRE_H */
357