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 2006 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 #pragma ident "%Z%%M% %I% %E% SMI" 31 32 #ifdef __cplusplus 33 extern "C" { 34 #endif 35 36 #define IPV6_LL_PREFIXLEN 10 /* Number of bits in link-local pref */ 37 38 #define IP_CACHE_TABLE_SIZE 256 39 #define IP_MRTUN_TABLE_SIZE 256 /* Mobile IP reverse tunnel table */ 40 /* size. Only used by mipagent */ 41 #define IP_SRCIF_TABLE_SIZE 256 /* Per interface routing table size */ 42 #define IP_MASK_TABLE_SIZE (IP_ABITS + 1) /* 33 ptrs */ 43 44 #define IP6_FTABLE_HASH_SIZE 32 /* size of each hash table in ptrs */ 45 #define IP6_CACHE_TABLE_SIZE 256 46 #define IP6_MASK_TABLE_SIZE (IPV6_ABITS + 1) /* 129 ptrs */ 47 48 /* 49 * We use the common modulo hash function. In ip_ire_init(), we make 50 * sure that the cache table size is always a power of 2. That's why 51 * we can use & instead of %. Also note that we try hard to make sure 52 * the lower bits of an address capture most info from the whole address. 53 * The reason being that since our hash table is probably a lot smaller 54 * than 2^32 buckets so the lower bits are the most important. 55 */ 56 #define IRE_ADDR_HASH(addr, table_size) \ 57 (((addr) ^ ((addr) >> 8) ^ ((addr) >> 16) ^ ((addr) >> 24)) & \ 58 ((table_size) - 1)) 59 60 /* 61 * Exclusive-or those bytes that are likely to contain the MAC 62 * address. Assumes EUI-64 format for good hashing. 63 */ 64 #define IRE_ADDR_HASH_V6(addr, table_size) \ 65 (((addr).s6_addr32[3] ^ \ 66 (((addr).s6_addr32[3] ^ (addr).s6_addr32[2]) >> 12)) & \ 67 ((table_size) - 1)) 68 /* This assumes that the ftable size is a power of 2. */ 69 #define IRE_ADDR_MASK_HASH_V6(addr, mask, table_size) \ 70 ((((addr).s6_addr8[8] & (mask).s6_addr8[8]) ^ \ 71 ((addr).s6_addr8[9] & (mask).s6_addr8[9]) ^ \ 72 ((addr).s6_addr8[10] & (mask).s6_addr8[10]) ^ \ 73 ((addr).s6_addr8[13] & (mask).s6_addr8[13]) ^ \ 74 ((addr).s6_addr8[14] & (mask).s6_addr8[14]) ^ \ 75 ((addr).s6_addr8[15] & (mask).s6_addr8[15])) & ((table_size) - 1)) 76 77 /* 78 * match parameter definitions for 79 * IRE lookup routines. 80 */ 81 #define MATCH_IRE_DSTONLY 0x0000 /* Match just the address */ 82 #define MATCH_IRE_TYPE 0x0001 /* Match IRE type */ 83 #define MATCH_IRE_SRC 0x0002 /* Match IRE source address */ 84 #define MATCH_IRE_MASK 0x0004 /* Match IRE mask */ 85 #define MATCH_IRE_WQ 0x0008 /* Match IRE Write Q */ 86 #define MATCH_IRE_GW 0x0010 /* Match IRE gateway */ 87 #define MATCH_IRE_IPIF 0x0020 /* Match IRE ipif */ 88 #define MATCH_IRE_RECURSIVE 0x0040 /* Do recursive lookup if necessary */ 89 #define MATCH_IRE_DEFAULT 0x0080 /* Return default route if no route */ 90 /* found. */ 91 #define MATCH_IRE_RJ_BHOLE 0x0100 /* During lookup if we hit an ire */ 92 /* with RTF_REJECT or RTF_BLACKHOLE, */ 93 /* return the ire. No recursive */ 94 /* lookup should be done. */ 95 #define MATCH_IRE_IHANDLE 0x0200 /* Match IRE on ihandle */ 96 #define MATCH_IRE_MARK_HIDDEN 0x0400 /* Match IRE ire_marks with */ 97 /* IRE_MARK_HIDDEN. */ 98 /* 99 * MATCH_IRE_ILL is used whenever we want to specifically match an IRE 100 * whose ire_ipif->ipif_ill or (ill_t *)ire_stq->q_ptr matches a given 101 * ill. When MATCH_IRE_ILL is used to locate an IRE_CACHE, it implies 102 * that the packet will not be load balanced. This is normally used 103 * by in.mpathd to send out failure detection probes. 104 * 105 * MATCH_IRE_ILL_GROUP is used whenever we are not specific about which 106 * interface (ill) the packet should be sent out. This implies that the 107 * packets will be subjected to load balancing and it might go out on 108 * any interface in the group. When there is only interface in the group, 109 * MATCH_IRE_ILL_GROUP becomes MATCH_IRE_ILL. Most of the code uses 110 * MATCH_IRE_ILL_GROUP and MATCH_IRE_ILL is used in very few cases where 111 * we want to disable load balancing. 112 * 113 * MATCH_IRE_PARENT is used whenever we unconditionally want to get the 114 * parent IRE (sire) while recursively searching IREs for an offsubnet 115 * destination. With this flag, even if no IRE_CACHETABLE or IRE_INTERFACE 116 * is found to help resolving IRE_OFFSUBNET in lookup routines, the 117 * IRE_OFFSUBNET sire, if any, is returned to the caller. 118 */ 119 #define MATCH_IRE_ILL_GROUP 0x0800 /* Match IRE on ill or the ill_group. */ 120 #define MATCH_IRE_ILL 0x1000 /* Match IRE on the ill only */ 121 122 #define MATCH_IRE_PARENT 0x2000 /* Match parent ire, if any, */ 123 /* even if ire is not matched. */ 124 #define MATCH_IRE_ZONEONLY 0x4000 /* Match IREs in specified zone, ie */ 125 /* don't match IRE_LOCALs from other */ 126 /* zones or shared IREs */ 127 #define MATCH_IRE_MARK_PRIVATE_ADDR 0x8000 /* Match IRE ire_marks with */ 128 /* IRE_MARK_PRIVATE_ADDR. */ 129 #define MATCH_IRE_SECATTR 0x10000 /* Match gateway security attributes */ 130 #define MATCH_IRE_COMPLETE 0x20000 /* ire_ftable_lookup() can return */ 131 /* IRE_CACHE entry only if it is */ 132 /* ND_REACHABLE */ 133 134 /* 135 * Any ire to nce association is long term, and 136 * the refhold and refrele may be done by different 137 * threads. So all cases of making or breaking ire to 138 * nce association should all effectively use the NOTR variants. 139 * To understand the *effectively* part read on. 140 * 141 * ndp_lookup() and ndp_add() implicitly does NCE_REFHOLD. So wherever we 142 * make ire to nce association after calling these functions, 143 * we effectively want to end up with NCE_REFHOLD_NOTR, 144 * We call this macro to achieve this effect. This macro changes 145 * a NCE_REFHOLD to a NCE_REFHOLD_NOTR. The macro's NCE_REFRELE 146 * cancels off ndp_lookup[ndp_add]'s implicit NCE_REFHOLD, and what 147 * you are left with is a NCE_REFHOLD_NOTR 148 */ 149 #define NCE_REFHOLD_TO_REFHOLD_NOTR(nce) { \ 150 NCE_REFHOLD_NOTR(nce); \ 151 NCE_REFRELE(nce); \ 152 } 153 154 /* Structure for ire_cache_count() */ 155 typedef struct { 156 int icc_total; /* Total number of IRE_CACHE */ 157 int icc_unused; /* # off/no PMTU unused since last reclaim */ 158 int icc_offlink; /* # offlink without PMTU information */ 159 int icc_pmtu; /* # offlink with PMTU information */ 160 int icc_onlink; /* # onlink */ 161 } ire_cache_count_t; 162 163 /* 164 * Structure for ire_cache_reclaim(). Each field is a fraction i.e. 1 meaning 165 * reclaim all, N meaning reclaim 1/Nth of all entries, 0 meaning reclaim none. 166 */ 167 typedef struct { 168 int icr_unused; /* Fraction for unused since last reclaim */ 169 int icr_offlink; /* Fraction for offlink without PMTU info */ 170 int icr_pmtu; /* Fraction for offlink with PMTU info */ 171 int icr_onlink; /* Fraction for onlink */ 172 } ire_cache_reclaim_t; 173 174 typedef struct { 175 uint64_t ire_stats_alloced; /* # of ires alloced */ 176 uint64_t ire_stats_freed; /* # of ires freed */ 177 uint64_t ire_stats_inserted; /* # of ires inserted in the bucket */ 178 uint64_t ire_stats_deleted; /* # of ires deleted from the bucket */ 179 } ire_stats_t; 180 181 extern ire_stats_t ire_stats_v4; 182 extern uint32_t ip_cache_table_size; 183 extern uint32_t ip6_cache_table_size; 184 extern irb_t *ip_cache_table; 185 extern uint32_t ip6_ftable_hash_size; 186 187 /* 188 * We use atomics so that we get an accurate accounting on the ires. 189 * Otherwise we can't determine leaks correctly. 190 */ 191 #define BUMP_IRE_STATS(ire_stats, x) atomic_add_64(&(ire_stats).x, 1) 192 193 extern irb_t *ip_forwarding_table_v6[]; 194 extern irb_t *ip_cache_table_v6; 195 extern irb_t *ip_mrtun_table; 196 extern irb_t *ip_srcif_table; 197 extern kmutex_t ire_ft_init_lock; 198 extern kmutex_t ire_mrtun_lock; 199 extern kmutex_t ire_srcif_table_lock; 200 extern ire_stats_t ire_stats_v6; 201 extern uint_t ire_mrtun_count; 202 extern uint_t ire_srcif_table_count; 203 204 #ifdef _KERNEL 205 struct ts_label_s; 206 207 extern ipaddr_t ip_plen_to_mask(uint_t); 208 extern in6_addr_t *ip_plen_to_mask_v6(uint_t, in6_addr_t *); 209 210 extern int ip_ire_advise(queue_t *, mblk_t *, cred_t *); 211 extern int ip_ire_delete(queue_t *, mblk_t *, cred_t *); 212 extern boolean_t ip_ire_clookup_and_delete(ipaddr_t, ipif_t *); 213 extern void ip_ire_clookup_and_delete_v6(const in6_addr_t *); 214 215 extern int ip_ire_report(queue_t *, mblk_t *, caddr_t, cred_t *); 216 extern int ip_ire_report_mrtun(queue_t *, mblk_t *, caddr_t, cred_t *); 217 extern int ip_ire_report_srcif(queue_t *, mblk_t *, caddr_t, cred_t *); 218 extern int ip_ire_report_v6(queue_t *, mblk_t *, caddr_t, cred_t *); 219 extern void ire_report_ftable(ire_t *, char *); 220 221 extern void ip_ire_req(queue_t *, mblk_t *); 222 223 extern int ip_mask_to_plen(ipaddr_t); 224 extern int ip_mask_to_plen_v6(const in6_addr_t *); 225 226 extern ire_t *ipif_to_ire(const ipif_t *); 227 extern ire_t *ipif_to_ire_v6(const ipif_t *); 228 229 extern int ire_add(ire_t **, queue_t *, mblk_t *, ipsq_func_t, boolean_t); 230 extern int ire_add_mrtun(ire_t **, queue_t *, mblk_t *, ipsq_func_t); 231 extern void ire_add_then_send(queue_t *, ire_t *, mblk_t *); 232 extern int ire_add_v6(ire_t **, queue_t *, mblk_t *, ipsq_func_t); 233 extern int ire_atomic_start(irb_t *irb_ptr, ire_t *ire, queue_t *q, 234 mblk_t *mp, ipsq_func_t func); 235 extern void ire_atomic_end(irb_t *irb_ptr, ire_t *ire); 236 237 extern void ire_cache_count(ire_t *, char *); 238 extern ire_t *ire_cache_lookup(ipaddr_t, zoneid_t, 239 const struct ts_label_s *); 240 extern ire_t *ire_cache_lookup_v6(const in6_addr_t *, zoneid_t, 241 const struct ts_label_s *); 242 extern void ire_cache_reclaim(ire_t *, char *); 243 244 extern void ire_check_bcast_present(ipif_t *, ipaddr_t, int, boolean_t *, 245 boolean_t *); 246 247 extern ire_t *ire_create_mp(uchar_t *, uchar_t *, uchar_t *, uchar_t *, 248 uchar_t *, uint_t, mblk_t *, queue_t *, queue_t *, ushort_t, mblk_t *, 249 ipif_t *, ill_t *, ipaddr_t, uint32_t, uint32_t, uint32_t, const iulp_t *, 250 tsol_gc_t *, tsol_gcgrp_t *); 251 252 extern ire_t *ire_create(uchar_t *, uchar_t *, uchar_t *, uchar_t *, 253 uchar_t *, uint_t *, mblk_t *, queue_t *, queue_t *, ushort_t, mblk_t *, 254 ipif_t *, ill_t *, ipaddr_t, uint32_t, uint32_t, uint32_t, const iulp_t *, 255 tsol_gc_t *, tsol_gcgrp_t *); 256 257 extern ire_t **ire_check_and_create_bcast(ipif_t *, ipaddr_t, 258 ire_t **, int); 259 extern ire_t **ire_create_bcast(ipif_t *, ipaddr_t, ire_t **); 260 extern ire_t *ire_init(ire_t *, uchar_t *, uchar_t *, uchar_t *, 261 uchar_t *, uchar_t *, uint_t *, mblk_t *, queue_t *, queue_t *, ushort_t, 262 mblk_t *, ipif_t *, ill_t *, ipaddr_t, uint32_t, uint32_t, uint32_t, 263 const iulp_t *, tsol_gc_t *, tsol_gcgrp_t *); 264 265 extern boolean_t ire_init_common(ire_t *, uint_t *, mblk_t *, queue_t *, 266 queue_t *, ushort_t, mblk_t *, ipif_t *, ill_t *, uint32_t, 267 uint32_t, uint32_t, uchar_t, const iulp_t *, tsol_gc_t *, tsol_gcgrp_t *); 268 269 extern ire_t *ire_create_v6(const in6_addr_t *, const in6_addr_t *, 270 const in6_addr_t *, const in6_addr_t *, uint_t *, mblk_t *, queue_t *, 271 queue_t *, ushort_t, mblk_t *, ipif_t *, 272 const in6_addr_t *, uint32_t, uint32_t, uint_t, const iulp_t *, 273 tsol_gc_t *, tsol_gcgrp_t *); 274 275 extern ire_t *ire_create_mp_v6(const in6_addr_t *, const in6_addr_t *, 276 const in6_addr_t *, const in6_addr_t *, mblk_t *, queue_t *, 277 queue_t *, ushort_t, mblk_t *, ipif_t *, 278 const in6_addr_t *, uint32_t, uint32_t, uint_t, const iulp_t *, 279 tsol_gc_t *, tsol_gcgrp_t *); 280 281 extern ire_t *ire_init_v6(ire_t *, const in6_addr_t *, const in6_addr_t *, 282 const in6_addr_t *, const in6_addr_t *, uint_t *, mblk_t *, queue_t *, 283 queue_t *, ushort_t, mblk_t *, ipif_t *, 284 const in6_addr_t *, uint32_t, uint32_t, uint_t, const iulp_t *, 285 tsol_gc_t *, tsol_gcgrp_t *); 286 287 extern void ire_clookup_delete_cache_gw(ipaddr_t, zoneid_t); 288 extern void ire_clookup_delete_cache_gw_v6(const in6_addr_t *, zoneid_t); 289 290 extern ire_t *ire_ctable_lookup(ipaddr_t, ipaddr_t, int, const ipif_t *, 291 zoneid_t, const struct ts_label_s *, int); 292 293 extern ire_t *ire_ctable_lookup_v6(const in6_addr_t *, const in6_addr_t *, 294 int, const ipif_t *, zoneid_t, const struct ts_label_s *, int); 295 296 extern void ire_delete(ire_t *); 297 extern void ire_delete_cache_gw(ire_t *, char *); 298 extern void ire_delete_cache_gw_v6(ire_t *, char *); 299 extern void ire_delete_cache_v6(ire_t *, char *); 300 extern void ire_delete_srcif(ire_t *); 301 extern void ire_delete_v6(ire_t *); 302 303 extern void ire_expire(ire_t *, char *); 304 extern void ire_fastpath_flush(ire_t *, void *); 305 extern boolean_t ire_fastpath_update(ire_t *, void *); 306 307 extern void ire_flush_cache_v4(ire_t *, int); 308 extern void ire_flush_cache_v6(ire_t *, int); 309 310 extern ire_t *ire_ftable_lookup_v6(const in6_addr_t *, const in6_addr_t *, 311 const in6_addr_t *, int, const ipif_t *, ire_t **, zoneid_t, 312 uint32_t, const struct ts_label_s *, int); 313 314 extern ire_t *ire_ihandle_lookup_onlink(ire_t *); 315 extern ire_t *ire_ihandle_lookup_offlink(ire_t *, ire_t *); 316 extern ire_t *ire_ihandle_lookup_offlink_v6(ire_t *, ire_t *); 317 318 extern ire_t *ire_lookup_local(zoneid_t); 319 extern ire_t *ire_lookup_local_v6(zoneid_t); 320 321 extern ire_t *ire_lookup_multi(ipaddr_t, zoneid_t); 322 extern ire_t *ire_lookup_multi_v6(const in6_addr_t *, zoneid_t); 323 324 extern ire_t *ire_mrtun_lookup(ipaddr_t, ill_t *); 325 326 extern void ire_refrele(ire_t *); 327 extern void ire_refrele_notr(ire_t *); 328 extern ire_t *ire_route_lookup(ipaddr_t, ipaddr_t, ipaddr_t, int, 329 const ipif_t *, ire_t **, zoneid_t, const struct ts_label_s *, int); 330 331 extern ire_t *ire_route_lookup_v6(const in6_addr_t *, const in6_addr_t *, 332 const in6_addr_t *, int, const ipif_t *, ire_t **, zoneid_t, 333 const struct ts_label_s *, int); 334 335 extern ire_t *ire_srcif_table_lookup(ipaddr_t, int, ipif_t *, ill_t *, int); 336 extern ill_t *ire_to_ill(const ire_t *); 337 338 extern void ire_walk(pfv_t, void *); 339 extern void ire_walk_ill(uint_t, uint_t, pfv_t, void *, ill_t *); 340 extern void ire_walk_ill_mrtun(uint_t, uint_t, pfv_t, void *, ill_t *); 341 extern void ire_walk_ill_v4(uint_t, uint_t, pfv_t, void *, ill_t *); 342 extern void ire_walk_ill_v6(uint_t, uint_t, pfv_t, void *, ill_t *); 343 extern void ire_walk_v4(pfv_t, void *, zoneid_t); 344 extern void ire_walk_ill_tables(uint_t match_flags, uint_t ire_type, 345 pfv_t func, void *arg, size_t ftbl_sz, size_t htbl_sz, 346 irb_t **ipftbl, size_t ctbl_sz, irb_t *ipctbl, ill_t *ill, 347 zoneid_t zoneid); 348 extern void ire_walk_srcif_table_v4(pfv_t, void *); 349 extern void ire_walk_v6(pfv_t, void *, zoneid_t); 350 351 extern boolean_t ire_multirt_lookup(ire_t **, ire_t **, uint32_t, 352 const struct ts_label_s *); 353 extern boolean_t ire_multirt_need_resolve(ipaddr_t, 354 const struct ts_label_s *); 355 extern boolean_t ire_multirt_lookup_v6(ire_t **, ire_t **, uint32_t, 356 const struct ts_label_s *); 357 extern boolean_t ire_multirt_need_resolve_v6(const in6_addr_t *, 358 const struct ts_label_s *); 359 360 extern ire_t *ipif_lookup_multi_ire(ipif_t *, ipaddr_t); 361 extern ire_t *ipif_lookup_multi_ire_v6(ipif_t *, const in6_addr_t *); 362 363 extern void ire_fastpath_list_dispatch(ill_t *, 364 boolean_t (*)(ire_t *, void *), void *); 365 extern void ire_fastpath_list_delete(ill_t *, ire_t *); 366 367 extern mblk_t *ip_nexthop_route(const struct sockaddr *, char *); 368 extern mblk_t *ip_nexthop(const struct sockaddr *, const char *); 369 370 extern ire_t *ire_get_next_bcast_ire(ire_t *, ire_t *); 371 extern ire_t *ire_get_next_default_ire(ire_t *, ire_t *); 372 373 extern void ire_arpresolve(ire_t *, ill_t *); 374 extern void ire_freemblk(ire_t *); 375 extern void ire_fastpath(ire_t *); 376 extern boolean_t ire_match_args(ire_t *, ipaddr_t, ipaddr_t, ipaddr_t, 377 int, const ipif_t *, zoneid_t, uint32_t, const struct ts_label_s *, int); 378 extern int ire_nce_init(ire_t *, mblk_t *, mblk_t *); 379 extern boolean_t ire_walk_ill_match(uint_t, uint_t, ire_t *, ill_t *, 380 zoneid_t); 381 382 #endif /* _KERNEL */ 383 384 #ifdef __cplusplus 385 } 386 #endif 387 388 #endif /* _INET_IP_IRE_H */ 389