1 /* $NetBSD: ip_flow.c,v 1.33 2006/06/07 22:34:01 kardel Exp $ */ 2 3 /*- 4 * Copyright (c) 1998 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by the 3am Software Foundry ("3am"). It was developed by Matt Thomas. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. All advertising materials mentioning features or use of this software 19 * must display the following acknowledgement: 20 * This product includes software developed by the NetBSD 21 * Foundation, Inc. and its contributors. 22 * 4. Neither the name of The NetBSD Foundation nor the names of its 23 * contributors may be used to endorse or promote products derived 24 * from this software without specific prior written permission. 25 * 26 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 27 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 28 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 29 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 30 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 31 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 32 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 33 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 34 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 35 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 36 * POSSIBILITY OF SUCH DAMAGE. 37 */ 38 39 #include <sys/cdefs.h> 40 __KERNEL_RCSID(0, "$NetBSD: ip_flow.c,v 1.33 2006/06/07 22:34:01 kardel Exp $"); 41 42 #include <sys/param.h> 43 #include <sys/systm.h> 44 #include <sys/malloc.h> 45 #include <sys/mbuf.h> 46 #include <sys/domain.h> 47 #include <sys/protosw.h> 48 #include <sys/socket.h> 49 #include <sys/socketvar.h> 50 #include <sys/errno.h> 51 #include <sys/time.h> 52 #include <sys/kernel.h> 53 #include <sys/pool.h> 54 #include <sys/sysctl.h> 55 56 #include <net/if.h> 57 #include <net/if_dl.h> 58 #include <net/route.h> 59 #include <net/pfil.h> 60 61 #include <netinet/in.h> 62 #include <netinet/in_systm.h> 63 #include <netinet/ip.h> 64 #include <netinet/in_pcb.h> 65 #include <netinet/in_var.h> 66 #include <netinet/ip_var.h> 67 68 POOL_INIT(ipflow_pool, sizeof(struct ipflow), 0, 0, 0, "ipflowpl", NULL); 69 70 LIST_HEAD(ipflowhead, ipflow); 71 72 #define IPFLOW_TIMER (5 * PR_SLOWHZ) 73 #define IPFLOW_HASHSIZE (1 << IPFLOW_HASHBITS) 74 75 static struct ipflowhead ipflowtable[IPFLOW_HASHSIZE]; 76 static struct ipflowhead ipflowlist; 77 static int ipflow_inuse; 78 79 #define IPFLOW_INSERT(bucket, ipf) \ 80 do { \ 81 LIST_INSERT_HEAD((bucket), (ipf), ipf_hash); \ 82 LIST_INSERT_HEAD(&ipflowlist, (ipf), ipf_list); \ 83 } while (/*CONSTCOND*/ 0) 84 85 #define IPFLOW_REMOVE(ipf) \ 86 do { \ 87 LIST_REMOVE((ipf), ipf_hash); \ 88 LIST_REMOVE((ipf), ipf_list); \ 89 } while (/*CONSTCOND*/ 0) 90 91 #ifndef IPFLOW_MAX 92 #define IPFLOW_MAX 256 93 #endif 94 int ip_maxflows = IPFLOW_MAX; 95 96 static unsigned 97 ipflow_hash(struct in_addr dst, struct in_addr src, unsigned tos) 98 { 99 unsigned hash = tos; 100 int idx; 101 for (idx = 0; idx < 32; idx += IPFLOW_HASHBITS) 102 hash += (dst.s_addr >> (32 - idx)) + (src.s_addr >> idx); 103 return hash & (IPFLOW_HASHSIZE-1); 104 } 105 106 static struct ipflow * 107 ipflow_lookup(const struct ip *ip) 108 { 109 unsigned hash; 110 struct ipflow *ipf; 111 112 hash = ipflow_hash(ip->ip_dst, ip->ip_src, ip->ip_tos); 113 114 LIST_FOREACH(ipf, &ipflowtable[hash], ipf_hash) { 115 if (ip->ip_dst.s_addr == ipf->ipf_dst.s_addr 116 && ip->ip_src.s_addr == ipf->ipf_src.s_addr 117 && ip->ip_tos == ipf->ipf_tos) 118 break; 119 } 120 return ipf; 121 } 122 123 void 124 ipflow_init(void) 125 { 126 int i; 127 128 LIST_INIT(&ipflowlist); 129 for (i = 0; i < IPFLOW_HASHSIZE; i++) 130 LIST_INIT(&ipflowtable[i]); 131 } 132 133 int 134 ipflow_fastforward(struct mbuf *m) 135 { 136 struct ip *ip, ip_store; 137 struct ipflow *ipf; 138 struct rtentry *rt; 139 struct sockaddr *dst; 140 int error; 141 int iplen; 142 143 /* 144 * Are we forwarding packets? Big enough for an IP packet? 145 */ 146 if (!ipforwarding || ipflow_inuse == 0 || m->m_len < sizeof(struct ip)) 147 return 0; 148 149 /* 150 * Was packet received as a link-level multicast or broadcast? 151 * If so, don't try to fast forward.. 152 */ 153 if ((m->m_flags & (M_BCAST|M_MCAST)) != 0) 154 return 0; 155 156 /* 157 * IP header with no option and valid version and length 158 */ 159 if (IP_HDR_ALIGNED_P(mtod(m, caddr_t))) 160 ip = mtod(m, struct ip *); 161 else { 162 memcpy(&ip_store, mtod(m, caddr_t), sizeof(ip_store)); 163 ip = &ip_store; 164 } 165 iplen = ntohs(ip->ip_len); 166 if (ip->ip_v != IPVERSION || ip->ip_hl != (sizeof(struct ip) >> 2) || 167 iplen < sizeof(struct ip) || iplen > m->m_pkthdr.len) 168 return 0; 169 /* 170 * Find a flow. 171 */ 172 if ((ipf = ipflow_lookup(ip)) == NULL) 173 return 0; 174 175 /* 176 * Verify the IP header checksum. 177 */ 178 switch (m->m_pkthdr.csum_flags & 179 ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) | 180 M_CSUM_IPv4_BAD)) { 181 case M_CSUM_IPv4|M_CSUM_IPv4_BAD: 182 return (0); 183 184 case M_CSUM_IPv4: 185 /* Checksum was okay. */ 186 break; 187 188 default: 189 /* Must compute it ourselves. */ 190 if (in_cksum(m, sizeof(struct ip)) != 0) 191 return (0); 192 break; 193 } 194 195 /* 196 * Route and interface still up? 197 */ 198 rt = ipf->ipf_ro.ro_rt; 199 if ((rt->rt_flags & RTF_UP) == 0 || 200 (rt->rt_ifp->if_flags & IFF_UP) == 0) 201 return 0; 202 203 /* 204 * Packet size OK? TTL? 205 */ 206 if (m->m_pkthdr.len > rt->rt_ifp->if_mtu || ip->ip_ttl <= IPTTLDEC) 207 return 0; 208 209 /* 210 * Clear any in-bound checksum flags for this packet. 211 */ 212 m->m_pkthdr.csum_flags = 0; 213 214 /* 215 * Everything checks out and so we can forward this packet. 216 * Modify the TTL and incrementally change the checksum. 217 * 218 * This method of adding the checksum works on either endian CPU. 219 * If htons() is inlined, all the arithmetic is folded; otherwise 220 * the htons()s are combined by CSE due to the const attribute. 221 * 222 * Don't bother using HW checksumming here -- the incremental 223 * update is pretty fast. 224 */ 225 ip->ip_ttl -= IPTTLDEC; 226 if (ip->ip_sum >= (u_int16_t) ~htons(IPTTLDEC << 8)) 227 ip->ip_sum -= ~htons(IPTTLDEC << 8); 228 else 229 ip->ip_sum += htons(IPTTLDEC << 8); 230 231 /* 232 * Done modifying the header; copy it back, if necessary. 233 */ 234 if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) 235 memcpy(mtod(m, caddr_t), &ip_store, sizeof(ip_store)); 236 237 /* 238 * Trim the packet in case it's too long.. 239 */ 240 if (m->m_pkthdr.len > iplen) { 241 if (m->m_len == m->m_pkthdr.len) { 242 m->m_len = iplen; 243 m->m_pkthdr.len = iplen; 244 } else 245 m_adj(m, iplen - m->m_pkthdr.len); 246 } 247 248 /* 249 * Send the packet on it's way. All we can get back is ENOBUFS 250 */ 251 ipf->ipf_uses++; 252 PRT_SLOW_ARM(ipf->ipf_timer, IPFLOW_TIMER); 253 254 if (rt->rt_flags & RTF_GATEWAY) 255 dst = rt->rt_gateway; 256 else 257 dst = &ipf->ipf_ro.ro_dst; 258 259 if ((error = (*rt->rt_ifp->if_output)(rt->rt_ifp, m, dst, rt)) != 0) { 260 if (error == ENOBUFS) 261 ipf->ipf_dropped++; 262 else 263 ipf->ipf_errors++; 264 } 265 return 1; 266 } 267 268 static void 269 ipflow_addstats(struct ipflow *ipf) 270 { 271 ipf->ipf_ro.ro_rt->rt_use += ipf->ipf_uses; 272 ipstat.ips_cantforward += ipf->ipf_errors + ipf->ipf_dropped; 273 ipstat.ips_forward += ipf->ipf_uses; 274 ipstat.ips_fastforward += ipf->ipf_uses; 275 } 276 277 static void 278 ipflow_free(struct ipflow *ipf) 279 { 280 int s; 281 /* 282 * Remove the flow from the hash table (at elevated IPL). 283 * Once it's off the list, we can deal with it at normal 284 * network IPL. 285 */ 286 s = splnet(); 287 IPFLOW_REMOVE(ipf); 288 splx(s); 289 ipflow_addstats(ipf); 290 RTFREE(ipf->ipf_ro.ro_rt); 291 ipflow_inuse--; 292 pool_put(&ipflow_pool, ipf); 293 } 294 295 struct ipflow * 296 ipflow_reap(int just_one) 297 { 298 while (just_one || ipflow_inuse > ip_maxflows) { 299 struct ipflow *ipf, *maybe_ipf = NULL; 300 int s; 301 302 ipf = LIST_FIRST(&ipflowlist); 303 while (ipf != NULL) { 304 /* 305 * If this no longer points to a valid route 306 * reclaim it. 307 */ 308 if ((ipf->ipf_ro.ro_rt->rt_flags & RTF_UP) == 0) 309 goto done; 310 /* 311 * choose the one that's been least recently 312 * used or has had the least uses in the 313 * last 1.5 intervals. 314 */ 315 if (maybe_ipf == NULL || 316 ipf->ipf_timer < maybe_ipf->ipf_timer || 317 (ipf->ipf_timer == maybe_ipf->ipf_timer && 318 ipf->ipf_last_uses + ipf->ipf_uses < 319 maybe_ipf->ipf_last_uses + 320 maybe_ipf->ipf_uses)) 321 maybe_ipf = ipf; 322 ipf = LIST_NEXT(ipf, ipf_list); 323 } 324 ipf = maybe_ipf; 325 done: 326 /* 327 * Remove the entry from the flow table. 328 */ 329 s = splnet(); 330 IPFLOW_REMOVE(ipf); 331 splx(s); 332 ipflow_addstats(ipf); 333 RTFREE(ipf->ipf_ro.ro_rt); 334 if (just_one) 335 return ipf; 336 pool_put(&ipflow_pool, ipf); 337 ipflow_inuse--; 338 } 339 return NULL; 340 } 341 342 void 343 ipflow_slowtimo(void) 344 { 345 struct ipflow *ipf, *next_ipf; 346 347 for (ipf = LIST_FIRST(&ipflowlist); ipf != NULL; ipf = next_ipf) { 348 next_ipf = LIST_NEXT(ipf, ipf_list); 349 if (PRT_SLOW_ISEXPIRED(ipf->ipf_timer)) { 350 ipflow_free(ipf); 351 } else { 352 ipf->ipf_last_uses = ipf->ipf_uses; 353 ipf->ipf_ro.ro_rt->rt_use += ipf->ipf_uses; 354 ipstat.ips_forward += ipf->ipf_uses; 355 ipstat.ips_fastforward += ipf->ipf_uses; 356 ipf->ipf_uses = 0; 357 } 358 } 359 } 360 361 void 362 ipflow_create(const struct route *ro, struct mbuf *m) 363 { 364 const struct ip *const ip = mtod(m, struct ip *); 365 struct ipflow *ipf; 366 unsigned hash; 367 int s; 368 369 /* 370 * Don't create cache entries for ICMP messages. 371 */ 372 if (ip_maxflows == 0 || ip->ip_p == IPPROTO_ICMP) 373 return; 374 /* 375 * See if an existing flow struct exists. If so remove it from it's 376 * list and free the old route. If not, try to malloc a new one 377 * (if we aren't at our limit). 378 */ 379 ipf = ipflow_lookup(ip); 380 if (ipf == NULL) { 381 if (ipflow_inuse >= ip_maxflows) { 382 ipf = ipflow_reap(1); 383 } else { 384 ipf = pool_get(&ipflow_pool, PR_NOWAIT); 385 if (ipf == NULL) 386 return; 387 ipflow_inuse++; 388 } 389 bzero((caddr_t) ipf, sizeof(*ipf)); 390 } else { 391 s = splnet(); 392 IPFLOW_REMOVE(ipf); 393 splx(s); 394 ipflow_addstats(ipf); 395 RTFREE(ipf->ipf_ro.ro_rt); 396 ipf->ipf_uses = ipf->ipf_last_uses = 0; 397 ipf->ipf_errors = ipf->ipf_dropped = 0; 398 } 399 400 /* 401 * Fill in the updated information. 402 */ 403 ipf->ipf_ro = *ro; 404 ro->ro_rt->rt_refcnt++; 405 ipf->ipf_dst = ip->ip_dst; 406 ipf->ipf_src = ip->ip_src; 407 ipf->ipf_tos = ip->ip_tos; 408 PRT_SLOW_ARM(ipf->ipf_timer, IPFLOW_TIMER); 409 ipf->ipf_start = time_uptime; 410 /* 411 * Insert into the approriate bucket of the flow table. 412 */ 413 hash = ipflow_hash(ip->ip_dst, ip->ip_src, ip->ip_tos); 414 s = splnet(); 415 IPFLOW_INSERT(&ipflowtable[hash], ipf); 416 splx(s); 417 } 418 419 void 420 ipflow_invalidate_all(void) 421 { 422 struct ipflow *ipf, *next_ipf; 423 int s; 424 425 s = splnet(); 426 for (ipf = LIST_FIRST(&ipflowlist); ipf != NULL; ipf = next_ipf) { 427 next_ipf = LIST_NEXT(ipf, ipf_list); 428 ipflow_free(ipf); 429 } 430 splx(s); 431 } 432