xref: /titanic_41/usr/src/uts/common/inet/ip_ire.h (revision 35a5a3587fd94b666239c157d3722745250ccbd7)
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 2009 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 
97 #define	MAX_IRE_RECURSION	4	/* Max IREs in ire_route_recursive */
98 
99 
100 /*
101  * We use atomics so that we get an accurate accounting on the ires.
102  * Otherwise we can't determine leaks correctly.
103  */
104 #define	BUMP_IRE_STATS(ire_stats, x) atomic_add_64(&(ire_stats).x, 1)
105 
106 #ifdef _KERNEL
107 struct ts_label_s;
108 struct nce_s;
109 /*
110  * structure for passing args between ire_ftable_lookup and ire_find_best_route
111  */
112 typedef struct ire_ftable_args_s {
113 	in6_addr_t		ift_addr_v6;
114 	in6_addr_t		ift_mask_v6;
115 	in6_addr_t		ift_gateway_v6;
116 #define	ift_addr		V4_PART_OF_V6(ift_addr_v6)
117 #define	ift_mask		V4_PART_OF_V6(ift_mask_v6)
118 #define	ift_gateway		V4_PART_OF_V6(ift_gateway_v6)
119 	int			ift_type;
120 	const ill_t		*ift_ill;
121 	zoneid_t		ift_zoneid;
122 	const ts_label_t	*ift_tsl;
123 	int			ift_flags;
124 	ire_t			*ift_best_ire;
125 } ire_ftable_args_t;
126 
127 extern	ipaddr_t	ip_plen_to_mask(uint_t);
128 extern	in6_addr_t	*ip_plen_to_mask_v6(uint_t, in6_addr_t *);
129 
130 extern	int	ip_ire_advise(queue_t *, mblk_t *, cred_t *);
131 extern	int	ip_ire_delete(queue_t *, mblk_t *, cred_t *);
132 extern	void	ip_ire_reclaim(void *);
133 
134 extern	int	ip_mask_to_plen(ipaddr_t);
135 extern	int	ip_mask_to_plen_v6(const in6_addr_t *);
136 
137 extern	ire_t	*ire_add(ire_t *);
138 extern	ire_t	*ire_add_v6(ire_t *);
139 extern	int	ire_atomic_start(irb_t *irb_ptr, ire_t *ire);
140 extern	void	ire_atomic_end(irb_t *irb_ptr, ire_t *ire);
141 
142 extern	ire_t	*ire_create(uchar_t *, uchar_t *, uchar_t *,
143     ushort_t, ill_t *, zoneid_t, uint_t, tsol_gc_t *, ip_stack_t *);
144 
145 extern	ire_t	**ire_create_bcast(ill_t *, ipaddr_t, zoneid_t, ire_t **);
146 extern	ire_t	*ire_create_if_clone(ire_t *, const in6_addr_t *, uint_t *);
147 extern	ire_t	*ire_lookup_bcast(ill_t *, ipaddr_t, zoneid_t);
148 extern	int	ire_init_v4(ire_t *, uchar_t *, uchar_t *, uchar_t *,
149     ushort_t, ill_t *, zoneid_t, uint_t, tsol_gc_t *, ip_stack_t *);
150 extern	int	ire_init_v6(ire_t *, const in6_addr_t *, const in6_addr_t *,
151     const in6_addr_t *, ushort_t, ill_t *, zoneid_t, uint_t, tsol_gc_t *,
152     ip_stack_t *);
153 
154 extern	int	ire_init_common(ire_t *, ushort_t, ill_t *, zoneid_t, uint_t,
155     uchar_t, tsol_gc_t *, ip_stack_t *);
156 
157 extern	ire_t	*ire_create_v6(const in6_addr_t *, const in6_addr_t *,
158     const in6_addr_t *, ushort_t, ill_t *, zoneid_t, uint_t,
159     tsol_gc_t *, ip_stack_t *);
160 
161 extern	void	ire_delete(ire_t *);
162 extern	void	ire_delete_v6(ire_t *);
163 
164 /*
165  * ire_pref used to make sure we don't set up routing loops in the ire_dep
166  * chain.
167  */
168 extern	int	ire_pref(ire_t *);
169 extern	boolean_t ire_dep_build(ire_t *[], uint_t [], uint_t);
170 extern	void	ire_dep_delete_if_clone(ire_t *);
171 extern	void	ire_dep_incr_generation(ire_t *);
172 extern	void	ire_dep_remove(ire_t *);
173 extern	void	ire_dep_unbuild(ire_t *[], uint_t);
174 extern	uint_t	ire_dep_validate_generations(ire_t *);
175 extern	void	ire_dep_invalidate_generations(ire_t *);
176 extern	boolean_t ire_determine_nce_capable(ire_t *);
177 
178 extern	void	ire_flush_cache_v4(ire_t *, int);
179 extern	void	ire_flush_cache_v6(ire_t *, int);
180 
181 extern	ire_t	*ire_ftable_lookup_v4(ipaddr_t, ipaddr_t, ipaddr_t, int,
182     const ill_t *, zoneid_t, const struct ts_label_s *, int, uint32_t,
183     ip_stack_t *, uint_t *);
184 extern	ire_t	*ire_ftable_lookup_v6(const in6_addr_t *, const in6_addr_t *,
185     const in6_addr_t *, int, const ill_t *, zoneid_t,
186     const struct ts_label_s *, int, uint32_t, ip_stack_t *, uint_t *);
187 
188 extern	ire_t	*ire_ftable_lookup_simple_v4(ipaddr_t, uint32_t, ip_stack_t *,
189     uint_t *);
190 extern	ire_t	*ire_ftable_lookup_simple_v6(const in6_addr_t *, uint32_t,
191     ip_stack_t *, uint_t *);
192 
193 extern boolean_t ire_gateway_ok_zone_v4(ipaddr_t, zoneid_t, ill_t *,
194     const ts_label_t *, ip_stack_t *, boolean_t);
195 extern boolean_t ire_gateway_ok_zone_v6(const in6_addr_t *, zoneid_t, ill_t *,
196     const ts_label_t *, ip_stack_t *, boolean_t);
197 
198 extern ire_t	*ire_alt_local(ire_t *, zoneid_t, const ts_label_t *,
199     const ill_t *, uint_t *);
200 
201 extern  ill_t	*ire_lookup_multi_ill_v4(ipaddr_t, zoneid_t, ip_stack_t *,
202     boolean_t *, ipaddr_t *);
203 extern  ill_t	*ire_lookup_multi_ill_v6(const in6_addr_t *, zoneid_t,
204     ip_stack_t *, boolean_t *, in6_addr_t *);
205 
206 extern	ire_t	*ire_nexthop(ire_t *);
207 extern	ill_t	*ire_nexthop_ill(ire_t *);
208 extern	ill_t	*ire_nce_ill(ire_t *);
209 
210 extern	ire_t	*ire_reject(ip_stack_t *, boolean_t);
211 extern	ire_t	*ire_blackhole(ip_stack_t *, boolean_t);
212 extern	ire_t	*ire_multicast(ill_t *);
213 
214 /* The different ire_recvfn functions */
215 extern void	ire_recv_forward_v4(ire_t *, mblk_t *, void *,
216     ip_recv_attr_t *);
217 extern void	ire_recv_noroute_v4(ire_t *, mblk_t *, void *,
218     ip_recv_attr_t *);
219 extern void	ire_recv_broadcast_v4(ire_t *, mblk_t *, void *,
220     ip_recv_attr_t *);
221 extern void	ire_recv_multicast_v4(ire_t *, mblk_t *, void *,
222     ip_recv_attr_t *);
223 extern void	ire_recv_multirt_v4(ire_t *, mblk_t *, void *,
224     ip_recv_attr_t *);
225 extern void	ire_recv_loopback_v4(ire_t *, mblk_t *, void *,
226     ip_recv_attr_t *);
227 extern void	ire_recv_local_v4(ire_t *, mblk_t *, void *,
228     ip_recv_attr_t *);
229 extern void	ire_recv_noaccept_v4(ire_t *, mblk_t *, void *,
230     ip_recv_attr_t *);
231 
232 extern void	ire_recv_forward_v6(ire_t *, mblk_t *, void *,
233     ip_recv_attr_t *);
234 extern void	ire_recv_noroute_v6(ire_t *, mblk_t *, void *,
235     ip_recv_attr_t *);
236 extern void	ire_recv_multicast_v6(ire_t *, mblk_t *, void *,
237     ip_recv_attr_t *);
238 extern void	ire_recv_multirt_v6(ire_t *, mblk_t *, void *,
239     ip_recv_attr_t *);
240 extern void	ire_recv_loopback_v6(ire_t *, mblk_t *, void *,
241     ip_recv_attr_t *);
242 extern void	ire_recv_local_v6(ire_t *, mblk_t *, void *, ip_recv_attr_t *);
243 extern void	ire_recv_noaccept_v6(ire_t *, mblk_t *, void *,
244     ip_recv_attr_t *);
245 
246 extern	void	irb_refhold(irb_t *);
247 extern	void	irb_refhold_locked(irb_t *);
248 extern	void	irb_refrele(irb_t *);
249 extern  void	irb_increment_generation(irb_t *);
250 
251 extern	void	ire_refhold(ire_t *);
252 extern	void	ire_refhold_notr(ire_t *);
253 extern	void	ire_refhold_locked(ire_t *);
254 extern	void	ire_refrele(ire_t *);
255 extern	void	ire_refrele_notr(ire_t *);
256 extern	void	ire_make_condemned(ire_t *);
257 extern	boolean_t ire_no_good(ire_t *);
258 extern	nce_t	*ire_handle_condemned_nce(nce_t *, ire_t *, ipha_t *, ip6_t *,
259     boolean_t);
260 
261 extern ire_t   	*ire_round_robin(irb_t *, ire_ftable_args_t *, uint_t,
262     ire_t *, ip_stack_t *);
263 
264 extern ire_t	*ire_route_recursive_v4(ipaddr_t, uint_t, const ill_t *,
265     zoneid_t, const ts_label_t *, uint_t, boolean_t, uint32_t, ip_stack_t *,
266     ipaddr_t *, tsol_ire_gw_secattr_t **, uint_t *);
267 extern ire_t	*ire_route_recursive_v6(const in6_addr_t *, uint_t,
268     const ill_t *, zoneid_t, const ts_label_t *, uint_t, boolean_t, uint32_t,
269     ip_stack_t *, in6_addr_t *, tsol_ire_gw_secattr_t **, uint_t *);
270 extern ire_t	*ire_route_recursive_dstonly_v4(ipaddr_t, boolean_t,
271     uint32_t, ip_stack_t *);
272 extern ire_t	*ire_route_recursive_dstonly_v6(const in6_addr_t *, boolean_t,
273     uint32_t, ip_stack_t *);
274 extern ire_t	*ire_route_recursive_impl_v4(ire_t *ire, ipaddr_t, uint_t,
275     const ill_t *, zoneid_t, const ts_label_t *, uint_t, boolean_t, uint32_t,
276     ip_stack_t *, ipaddr_t *, tsol_ire_gw_secattr_t **, uint_t *);
277 extern ire_t	*ire_route_recursive_impl_v6(ire_t *ire, const in6_addr_t *,
278     uint_t, const ill_t *, zoneid_t, const ts_label_t *, uint_t, boolean_t,
279     uint32_t, ip_stack_t *, in6_addr_t *, tsol_ire_gw_secattr_t **, uint_t *);
280 
281 /* The different ire_sendfn functions */
282 extern int	ire_send_local_v4(ire_t *, mblk_t *, void *,
283     ip_xmit_attr_t *, uint32_t *);
284 extern int	ire_send_multirt_v4(ire_t *, mblk_t *, void *,
285     ip_xmit_attr_t *, uint32_t *);
286 extern int	ire_send_noroute_v4(ire_t *, mblk_t *, void *,
287     ip_xmit_attr_t *, uint32_t *);
288 extern int	ire_send_multicast_v4(ire_t *, mblk_t *, void *,
289     ip_xmit_attr_t *, uint32_t *);
290 extern int	ire_send_broadcast_v4(ire_t *, mblk_t *, void *,
291     ip_xmit_attr_t *, uint32_t *);
292 extern int	ire_send_wire_v4(ire_t *, mblk_t *, void *,
293     ip_xmit_attr_t *, uint32_t *);
294 extern int	ire_send_local_v6(ire_t *, mblk_t *, void *,
295     ip_xmit_attr_t *, uint32_t *);
296 extern int	ire_send_multirt_v6(ire_t *, mblk_t *, void *,
297     ip_xmit_attr_t *, uint32_t *);
298 extern int	ire_send_noroute_v6(ire_t *, mblk_t *, void *,
299     ip_xmit_attr_t *, uint32_t *);
300 extern int	ire_send_multicast_v6(ire_t *, mblk_t *, void *,
301     ip_xmit_attr_t *, uint32_t *);
302 extern int	ire_send_wire_v6(ire_t *, mblk_t *, void *,
303     ip_xmit_attr_t *, uint32_t *);
304 
305 extern nce_t	*ire_to_nce_pkt(ire_t *, mblk_t *);
306 extern nce_t	*ire_to_nce(ire_t *, ipaddr_t, const in6_addr_t *);
307 
308 /* Different ire_postfragfn functions */
309 extern int	ip_xmit(mblk_t *, struct nce_s *,
310     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
311 extern int	ip_postfrag_loopcheck(mblk_t *, struct nce_s *,
312     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
313 extern int	ip_postfrag_multirt_v4(mblk_t *, struct nce_s *,
314     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
315 extern int	ip_postfrag_multirt_v6(mblk_t *, struct nce_s *,
316     iaflags_t, uint_t, uint32_t, zoneid_t, zoneid_t, uintptr_t *);
317 
318 extern void	ip_postfrag_loopback(mblk_t *, struct nce_s *,
319     iaflags_t, uint_t, zoneid_t);
320 extern int	ire_revalidate_nce(ire_t *);
321 
322 extern ire_t	*ip_select_route_pkt(mblk_t *, ip_xmit_attr_t *,
323     uint_t *, int *, boolean_t *);
324 extern ire_t	*ip_select_route(const in6_addr_t *, ip_xmit_attr_t *,
325     uint_t *, in6_addr_t *, int *, boolean_t *);
326 extern ire_t	*ip_select_route_v4(ipaddr_t, ip_xmit_attr_t *,
327     uint_t *, ipaddr_t *, int *, boolean_t *);
328 extern ire_t	*ip_select_route_v6(const in6_addr_t *, ip_xmit_attr_t *,
329     uint_t *, in6_addr_t *, int *, boolean_t *);
330 
331 extern	void	ire_walk(pfv_t, void *, ip_stack_t *);
332 extern	void	ire_walk_ill(uint_t, uint_t, pfv_t, void *, ill_t *);
333 extern	void	ire_walk_v4(pfv_t, void *, zoneid_t, ip_stack_t *);
334 extern  void	ire_walk_ill_tables(uint_t match_flags, uint_t ire_type,
335     pfv_t func, void *arg, size_t ftbl_sz, size_t htbl_sz,
336     irb_t **ipftbl, ill_t *ill,
337     zoneid_t zoneid, ip_stack_t *);
338 extern	void	ire_walk_v6(pfv_t, void *, zoneid_t, ip_stack_t *);
339 
340 extern boolean_t	ire_match_args(ire_t *, ipaddr_t, ipaddr_t, ipaddr_t,
341     int, const ill_t *, zoneid_t, const struct ts_label_s *, int);
342 extern boolean_t	ire_match_args_v6(ire_t *, const in6_addr_t *,
343     const in6_addr_t *, const in6_addr_t *, int, const ill_t *, zoneid_t,
344     const ts_label_t *, int);
345 
346 extern  struct nce_s	*arp_nce_init(ill_t *, in_addr_t, int);
347 extern  boolean_t	ire_walk_ill_match(uint_t, uint_t, ire_t *, ill_t *,
348     zoneid_t, ip_stack_t *);
349 extern  void ire_increment_generation(ire_t *);
350 extern  void ire_increment_multicast_generation(ip_stack_t *, boolean_t);
351 
352 #endif /* _KERNEL */
353 
354 #ifdef	__cplusplus
355 }
356 #endif
357 
358 #endif	/* _INET_IP_IRE_H */
359