npf_inet.c revision 1.33 1 1.33 mlelstv /* $NetBSD: npf_inet.c,v 1.33 2015/12/17 12:17:13 mlelstv Exp $ */
2 1.1 rmind
3 1.1 rmind /*-
4 1.29 rmind * Copyright (c) 2009-2014 The NetBSD Foundation, Inc.
5 1.1 rmind * All rights reserved.
6 1.1 rmind *
7 1.1 rmind * This material is based upon work partially supported by The
8 1.1 rmind * NetBSD Foundation under a contract with Mindaugas Rasiukevicius.
9 1.1 rmind *
10 1.1 rmind * Redistribution and use in source and binary forms, with or without
11 1.1 rmind * modification, are permitted provided that the following conditions
12 1.1 rmind * are met:
13 1.1 rmind * 1. Redistributions of source code must retain the above copyright
14 1.1 rmind * notice, this list of conditions and the following disclaimer.
15 1.1 rmind * 2. Redistributions in binary form must reproduce the above copyright
16 1.1 rmind * notice, this list of conditions and the following disclaimer in the
17 1.1 rmind * documentation and/or other materials provided with the distribution.
18 1.1 rmind *
19 1.1 rmind * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
20 1.1 rmind * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
21 1.1 rmind * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22 1.1 rmind * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
23 1.1 rmind * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24 1.1 rmind * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25 1.1 rmind * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26 1.1 rmind * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27 1.1 rmind * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28 1.1 rmind * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29 1.1 rmind * POSSIBILITY OF SUCH DAMAGE.
30 1.1 rmind */
31 1.1 rmind
32 1.1 rmind /*
33 1.22 rmind * Various protocol related helper routines.
34 1.12 rmind *
35 1.12 rmind * This layer manipulates npf_cache_t structure i.e. caches requested headers
36 1.12 rmind * and stores which information was cached in the information bit field.
37 1.12 rmind * It is also responsibility of this layer to update or invalidate the cache
38 1.12 rmind * on rewrites (e.g. by translation routines).
39 1.1 rmind */
40 1.1 rmind
41 1.1 rmind #include <sys/cdefs.h>
42 1.33 mlelstv __KERNEL_RCSID(0, "$NetBSD: npf_inet.c,v 1.33 2015/12/17 12:17:13 mlelstv Exp $");
43 1.1 rmind
44 1.1 rmind #include <sys/param.h>
45 1.11 rmind #include <sys/types.h>
46 1.1 rmind
47 1.4 rmind #include <net/pfil.h>
48 1.4 rmind #include <net/if.h>
49 1.4 rmind #include <net/ethertypes.h>
50 1.4 rmind #include <net/if_ether.h>
51 1.4 rmind
52 1.1 rmind #include <netinet/in_systm.h>
53 1.1 rmind #include <netinet/in.h>
54 1.33 mlelstv #include <netinet6/in6_var.h>
55 1.1 rmind #include <netinet/ip.h>
56 1.4 rmind #include <netinet/ip6.h>
57 1.1 rmind #include <netinet/tcp.h>
58 1.1 rmind #include <netinet/udp.h>
59 1.1 rmind #include <netinet/ip_icmp.h>
60 1.1 rmind
61 1.1 rmind #include "npf_impl.h"
62 1.1 rmind
63 1.1 rmind /*
64 1.27 rmind * npf_fixup{16,32}_cksum: incremental update of the Internet checksum.
65 1.1 rmind */
66 1.1 rmind
67 1.1 rmind uint16_t
68 1.1 rmind npf_fixup16_cksum(uint16_t cksum, uint16_t odatum, uint16_t ndatum)
69 1.1 rmind {
70 1.1 rmind uint32_t sum;
71 1.1 rmind
72 1.1 rmind /*
73 1.1 rmind * RFC 1624:
74 1.1 rmind * HC' = ~(~HC + ~m + m')
75 1.27 rmind *
76 1.27 rmind * Note: 1's complement sum is endian-independent (RFC 1071, page 2).
77 1.1 rmind */
78 1.27 rmind sum = ~cksum & 0xffff;
79 1.27 rmind sum += (~odatum & 0xffff) + ndatum;
80 1.1 rmind sum = (sum >> 16) + (sum & 0xffff);
81 1.1 rmind sum += (sum >> 16);
82 1.1 rmind
83 1.27 rmind return ~sum & 0xffff;
84 1.1 rmind }
85 1.1 rmind
86 1.1 rmind uint16_t
87 1.1 rmind npf_fixup32_cksum(uint16_t cksum, uint32_t odatum, uint32_t ndatum)
88 1.1 rmind {
89 1.27 rmind uint32_t sum;
90 1.27 rmind
91 1.27 rmind /*
92 1.27 rmind * Checksum 32-bit datum as as two 16-bit. Note, the first
93 1.27 rmind * 32->16 bit reduction is not necessary.
94 1.27 rmind */
95 1.27 rmind sum = ~cksum & 0xffff;
96 1.27 rmind sum += (~odatum & 0xffff) + (ndatum & 0xffff);
97 1.1 rmind
98 1.27 rmind sum += (~odatum >> 16) + (ndatum >> 16);
99 1.27 rmind sum = (sum >> 16) + (sum & 0xffff);
100 1.27 rmind sum += (sum >> 16);
101 1.27 rmind return ~sum & 0xffff;
102 1.1 rmind }
103 1.1 rmind
104 1.1 rmind /*
105 1.4 rmind * npf_addr_cksum: calculate checksum of the address, either IPv4 or IPv6.
106 1.4 rmind */
107 1.4 rmind uint16_t
108 1.19 rmind npf_addr_cksum(uint16_t cksum, int sz, const npf_addr_t *oaddr,
109 1.19 rmind const npf_addr_t *naddr)
110 1.4 rmind {
111 1.19 rmind const uint32_t *oip32 = (const uint32_t *)oaddr;
112 1.19 rmind const uint32_t *nip32 = (const uint32_t *)naddr;
113 1.4 rmind
114 1.4 rmind KASSERT(sz % sizeof(uint32_t) == 0);
115 1.4 rmind do {
116 1.4 rmind cksum = npf_fixup32_cksum(cksum, *oip32++, *nip32++);
117 1.4 rmind sz -= sizeof(uint32_t);
118 1.4 rmind } while (sz);
119 1.4 rmind
120 1.4 rmind return cksum;
121 1.4 rmind }
122 1.4 rmind
123 1.4 rmind /*
124 1.26 rmind * npf_addr_sum: provide IP addresses as a XORed 32-bit integer.
125 1.4 rmind * Note: used for hash function.
126 1.1 rmind */
127 1.4 rmind uint32_t
128 1.26 rmind npf_addr_mix(const int sz, const npf_addr_t *a1, const npf_addr_t *a2)
129 1.1 rmind {
130 1.4 rmind uint32_t mix = 0;
131 1.1 rmind
132 1.5 rmind KASSERT(sz > 0 && a1 != NULL && a2 != NULL);
133 1.5 rmind
134 1.26 rmind for (int i = 0; i < (sz >> 2); i++) {
135 1.26 rmind mix ^= a1->s6_addr32[i];
136 1.26 rmind mix ^= a2->s6_addr32[i];
137 1.4 rmind }
138 1.4 rmind return mix;
139 1.4 rmind }
140 1.1 rmind
141 1.13 rmind /*
142 1.13 rmind * npf_addr_mask: apply the mask to a given address and store the result.
143 1.13 rmind */
144 1.13 rmind void
145 1.13 rmind npf_addr_mask(const npf_addr_t *addr, const npf_netmask_t mask,
146 1.13 rmind const int alen, npf_addr_t *out)
147 1.12 rmind {
148 1.13 rmind const int nwords = alen >> 2;
149 1.12 rmind uint_fast8_t length = mask;
150 1.12 rmind
151 1.12 rmind /* Note: maximum length is 32 for IPv4 and 128 for IPv6. */
152 1.12 rmind KASSERT(length <= NPF_MAX_NETMASK);
153 1.12 rmind
154 1.13 rmind for (int i = 0; i < nwords; i++) {
155 1.13 rmind uint32_t wordmask;
156 1.13 rmind
157 1.12 rmind if (length >= 32) {
158 1.13 rmind wordmask = htonl(0xffffffff);
159 1.12 rmind length -= 32;
160 1.13 rmind } else if (length) {
161 1.13 rmind wordmask = htonl(0xffffffff << (32 - length));
162 1.13 rmind length = 0;
163 1.12 rmind } else {
164 1.13 rmind wordmask = 0;
165 1.12 rmind }
166 1.13 rmind out->s6_addr32[i] = addr->s6_addr32[i] & wordmask;
167 1.12 rmind }
168 1.12 rmind }
169 1.12 rmind
170 1.12 rmind /*
171 1.12 rmind * npf_addr_cmp: compare two addresses, either IPv4 or IPv6.
172 1.12 rmind *
173 1.13 rmind * => Return 0 if equal and negative/positive if less/greater accordingly.
174 1.12 rmind * => Ignore the mask, if NPF_NO_NETMASK is specified.
175 1.12 rmind */
176 1.12 rmind int
177 1.12 rmind npf_addr_cmp(const npf_addr_t *addr1, const npf_netmask_t mask1,
178 1.13 rmind const npf_addr_t *addr2, const npf_netmask_t mask2, const int alen)
179 1.12 rmind {
180 1.13 rmind npf_addr_t realaddr1, realaddr2;
181 1.12 rmind
182 1.12 rmind if (mask1 != NPF_NO_NETMASK) {
183 1.13 rmind npf_addr_mask(addr1, mask1, alen, &realaddr1);
184 1.13 rmind addr1 = &realaddr1;
185 1.12 rmind }
186 1.12 rmind if (mask2 != NPF_NO_NETMASK) {
187 1.13 rmind npf_addr_mask(addr2, mask2, alen, &realaddr2);
188 1.13 rmind addr2 = &realaddr2;
189 1.12 rmind }
190 1.13 rmind return memcmp(addr1, addr2, alen);
191 1.12 rmind }
192 1.12 rmind
193 1.4 rmind /*
194 1.4 rmind * npf_tcpsaw: helper to fetch SEQ, ACK, WIN and return TCP data length.
195 1.12 rmind *
196 1.12 rmind * => Returns all values in host byte-order.
197 1.4 rmind */
198 1.4 rmind int
199 1.12 rmind npf_tcpsaw(const npf_cache_t *npc, tcp_seq *seq, tcp_seq *ack, uint32_t *win)
200 1.4 rmind {
201 1.19 rmind const struct tcphdr *th = npc->npc_l4.tcp;
202 1.8 rmind u_int thlen;
203 1.1 rmind
204 1.7 zoltan KASSERT(npf_iscached(npc, NPC_TCP));
205 1.1 rmind
206 1.4 rmind *seq = ntohl(th->th_seq);
207 1.4 rmind *ack = ntohl(th->th_ack);
208 1.4 rmind *win = (uint32_t)ntohs(th->th_win);
209 1.8 rmind thlen = th->th_off << 2;
210 1.1 rmind
211 1.7 zoltan if (npf_iscached(npc, NPC_IP4)) {
212 1.19 rmind const struct ip *ip = npc->npc_ip.v4;
213 1.21 rmind return ntohs(ip->ip_len) - npc->npc_hlen - thlen;
214 1.12 rmind } else if (npf_iscached(npc, NPC_IP6)) {
215 1.19 rmind const struct ip6_hdr *ip6 = npc->npc_ip.v6;
216 1.8 rmind return ntohs(ip6->ip6_plen) - thlen;
217 1.7 zoltan }
218 1.7 zoltan return 0;
219 1.1 rmind }
220 1.1 rmind
221 1.1 rmind /*
222 1.4 rmind * npf_fetch_tcpopts: parse and return TCP options.
223 1.1 rmind */
224 1.1 rmind bool
225 1.32 rmind npf_fetch_tcpopts(npf_cache_t *npc, uint16_t *mss, int *wscale)
226 1.1 rmind {
227 1.32 rmind nbuf_t *nbuf = npc->npc_nbuf;
228 1.19 rmind const struct tcphdr *th = npc->npc_l4.tcp;
229 1.4 rmind int topts_len, step;
230 1.19 rmind void *nptr;
231 1.4 rmind uint8_t val;
232 1.19 rmind bool ok;
233 1.4 rmind
234 1.7 zoltan KASSERT(npf_iscached(npc, NPC_IP46));
235 1.7 zoltan KASSERT(npf_iscached(npc, NPC_TCP));
236 1.10 rmind
237 1.4 rmind /* Determine if there are any TCP options, get their length. */
238 1.4 rmind topts_len = (th->th_off << 2) - sizeof(struct tcphdr);
239 1.4 rmind if (topts_len <= 0) {
240 1.4 rmind /* No options. */
241 1.1 rmind return false;
242 1.4 rmind }
243 1.4 rmind KASSERT(topts_len <= MAX_TCPOPTLEN);
244 1.1 rmind
245 1.4 rmind /* First step: IP and TCP header up to options. */
246 1.21 rmind step = npc->npc_hlen + sizeof(struct tcphdr);
247 1.19 rmind nbuf_reset(nbuf);
248 1.4 rmind next:
249 1.19 rmind if ((nptr = nbuf_advance(nbuf, step, 1)) == NULL) {
250 1.19 rmind ok = false;
251 1.19 rmind goto done;
252 1.4 rmind }
253 1.19 rmind val = *(uint8_t *)nptr;
254 1.12 rmind
255 1.4 rmind switch (val) {
256 1.4 rmind case TCPOPT_EOL:
257 1.4 rmind /* Done. */
258 1.19 rmind ok = true;
259 1.19 rmind goto done;
260 1.4 rmind case TCPOPT_NOP:
261 1.4 rmind topts_len--;
262 1.4 rmind step = 1;
263 1.4 rmind break;
264 1.4 rmind case TCPOPT_MAXSEG:
265 1.19 rmind if ((nptr = nbuf_advance(nbuf, 2, 2)) == NULL) {
266 1.19 rmind ok = false;
267 1.19 rmind goto done;
268 1.4 rmind }
269 1.4 rmind if (mss) {
270 1.19 rmind if (*mss) {
271 1.19 rmind memcpy(nptr, mss, sizeof(uint16_t));
272 1.19 rmind } else {
273 1.19 rmind memcpy(mss, nptr, sizeof(uint16_t));
274 1.19 rmind }
275 1.4 rmind }
276 1.4 rmind topts_len -= TCPOLEN_MAXSEG;
277 1.19 rmind step = 2;
278 1.4 rmind break;
279 1.4 rmind case TCPOPT_WINDOW:
280 1.10 rmind /* TCP Window Scaling (RFC 1323). */
281 1.19 rmind if ((nptr = nbuf_advance(nbuf, 2, 1)) == NULL) {
282 1.19 rmind ok = false;
283 1.19 rmind goto done;
284 1.4 rmind }
285 1.19 rmind val = *(uint8_t *)nptr;
286 1.4 rmind *wscale = (val > TCP_MAX_WINSHIFT) ? TCP_MAX_WINSHIFT : val;
287 1.4 rmind topts_len -= TCPOLEN_WINDOW;
288 1.19 rmind step = 1;
289 1.4 rmind break;
290 1.4 rmind default:
291 1.19 rmind if ((nptr = nbuf_advance(nbuf, 1, 1)) == NULL) {
292 1.19 rmind ok = false;
293 1.19 rmind goto done;
294 1.4 rmind }
295 1.19 rmind val = *(uint8_t *)nptr;
296 1.16 rmind if (val < 2 || val > topts_len) {
297 1.19 rmind ok = false;
298 1.19 rmind goto done;
299 1.4 rmind }
300 1.4 rmind topts_len -= val;
301 1.4 rmind step = val - 1;
302 1.4 rmind }
303 1.12 rmind
304 1.6 rmind /* Any options left? */
305 1.4 rmind if (__predict_true(topts_len > 0)) {
306 1.4 rmind goto next;
307 1.4 rmind }
308 1.19 rmind ok = true;
309 1.19 rmind done:
310 1.19 rmind if (nbuf_flag_p(nbuf, NBUF_DATAREF_RESET)) {
311 1.32 rmind npf_recache(npc);
312 1.19 rmind }
313 1.19 rmind return ok;
314 1.1 rmind }
315 1.1 rmind
316 1.19 rmind static int
317 1.19 rmind npf_cache_ip(npf_cache_t *npc, nbuf_t *nbuf)
318 1.1 rmind {
319 1.19 rmind const void *nptr = nbuf_dataptr(nbuf);
320 1.19 rmind const uint8_t ver = *(const uint8_t *)nptr;
321 1.19 rmind int flags = 0;
322 1.12 rmind
323 1.4 rmind switch (ver >> 4) {
324 1.12 rmind case IPVERSION: {
325 1.19 rmind struct ip *ip;
326 1.12 rmind
327 1.19 rmind ip = nbuf_ensure_contig(nbuf, sizeof(struct ip));
328 1.19 rmind if (ip == NULL) {
329 1.19 rmind return 0;
330 1.4 rmind }
331 1.12 rmind
332 1.4 rmind /* Check header length and fragment offset. */
333 1.10 rmind if ((u_int)(ip->ip_hl << 2) < sizeof(struct ip)) {
334 1.19 rmind return 0;
335 1.4 rmind }
336 1.4 rmind if (ip->ip_off & ~htons(IP_DF | IP_RF)) {
337 1.4 rmind /* Note fragmentation. */
338 1.19 rmind flags |= NPC_IPFRAG;
339 1.4 rmind }
340 1.12 rmind
341 1.4 rmind /* Cache: layer 3 - IPv4. */
342 1.14 rmind npc->npc_alen = sizeof(struct in_addr);
343 1.28 rmind npc->npc_ips[NPF_SRC] = (npf_addr_t *)&ip->ip_src;
344 1.28 rmind npc->npc_ips[NPF_DST] = (npf_addr_t *)&ip->ip_dst;
345 1.7 zoltan npc->npc_hlen = ip->ip_hl << 2;
346 1.19 rmind npc->npc_proto = ip->ip_p;
347 1.19 rmind
348 1.19 rmind npc->npc_ip.v4 = ip;
349 1.19 rmind flags |= NPC_IP4;
350 1.4 rmind break;
351 1.12 rmind }
352 1.4 rmind
353 1.12 rmind case (IPV6_VERSION >> 4): {
354 1.19 rmind struct ip6_hdr *ip6;
355 1.19 rmind struct ip6_ext *ip6e;
356 1.19 rmind size_t off, hlen;
357 1.19 rmind
358 1.19 rmind ip6 = nbuf_ensure_contig(nbuf, sizeof(struct ip6_hdr));
359 1.19 rmind if (ip6 == NULL) {
360 1.19 rmind return 0;
361 1.7 zoltan }
362 1.19 rmind
363 1.19 rmind /* Set initial next-protocol value. */
364 1.19 rmind hlen = sizeof(struct ip6_hdr);
365 1.19 rmind npc->npc_proto = ip6->ip6_nxt;
366 1.13 rmind npc->npc_hlen = hlen;
367 1.7 zoltan
368 1.12 rmind /*
369 1.19 rmind * Advance by the length of the current header.
370 1.12 rmind */
371 1.19 rmind off = nbuf_offset(nbuf);
372 1.19 rmind while (nbuf_advance(nbuf, hlen, 0) != NULL) {
373 1.19 rmind ip6e = nbuf_ensure_contig(nbuf, sizeof(*ip6e));
374 1.19 rmind if (ip6e == NULL) {
375 1.19 rmind return 0;
376 1.19 rmind }
377 1.19 rmind
378 1.13 rmind /*
379 1.13 rmind * Determine whether we are going to continue.
380 1.13 rmind */
381 1.19 rmind switch (npc->npc_proto) {
382 1.13 rmind case IPPROTO_HOPOPTS:
383 1.7 zoltan case IPPROTO_DSTOPTS:
384 1.7 zoltan case IPPROTO_ROUTING:
385 1.19 rmind hlen = (ip6e->ip6e_len + 1) << 3;
386 1.7 zoltan break;
387 1.7 zoltan case IPPROTO_FRAGMENT:
388 1.13 rmind hlen = sizeof(struct ip6_frag);
389 1.19 rmind flags |= NPC_IPFRAG;
390 1.7 zoltan break;
391 1.7 zoltan case IPPROTO_AH:
392 1.19 rmind hlen = (ip6e->ip6e_len + 2) << 2;
393 1.7 zoltan break;
394 1.7 zoltan default:
395 1.13 rmind hlen = 0;
396 1.13 rmind break;
397 1.13 rmind }
398 1.13 rmind
399 1.13 rmind if (!hlen) {
400 1.7 zoltan break;
401 1.7 zoltan }
402 1.19 rmind npc->npc_proto = ip6e->ip6e_nxt;
403 1.13 rmind npc->npc_hlen += hlen;
404 1.13 rmind }
405 1.7 zoltan
406 1.23 rmind /*
407 1.23 rmind * Re-fetch the header pointers (nbufs might have been
408 1.23 rmind * reallocated). Restore the original offset (if any).
409 1.23 rmind */
410 1.19 rmind nbuf_reset(nbuf);
411 1.23 rmind ip6 = nbuf_dataptr(nbuf);
412 1.19 rmind if (off) {
413 1.19 rmind nbuf_advance(nbuf, off, 0);
414 1.19 rmind }
415 1.19 rmind
416 1.12 rmind /* Cache: layer 3 - IPv6. */
417 1.14 rmind npc->npc_alen = sizeof(struct in6_addr);
418 1.28 rmind npc->npc_ips[NPF_SRC] = (npf_addr_t *)&ip6->ip6_src;
419 1.28 rmind npc->npc_ips[NPF_DST]= (npf_addr_t *)&ip6->ip6_dst;
420 1.19 rmind
421 1.19 rmind npc->npc_ip.v6 = ip6;
422 1.19 rmind flags |= NPC_IP6;
423 1.7 zoltan break;
424 1.12 rmind }
425 1.4 rmind default:
426 1.19 rmind break;
427 1.4 rmind }
428 1.19 rmind return flags;
429 1.1 rmind }
430 1.1 rmind
431 1.1 rmind /*
432 1.4 rmind * npf_cache_all: general routine to cache all relevant IP (v4 or v6)
433 1.12 rmind * and TCP, UDP or ICMP headers.
434 1.19 rmind *
435 1.19 rmind * => nbuf offset shall be set accordingly.
436 1.1 rmind */
437 1.10 rmind int
438 1.32 rmind npf_cache_all(npf_cache_t *npc)
439 1.1 rmind {
440 1.32 rmind nbuf_t *nbuf = npc->npc_nbuf;
441 1.19 rmind int flags, l4flags;
442 1.19 rmind u_int hlen;
443 1.19 rmind
444 1.19 rmind /*
445 1.19 rmind * This routine is a main point where the references are cached,
446 1.19 rmind * therefore clear the flag as we reset.
447 1.19 rmind */
448 1.19 rmind again:
449 1.19 rmind nbuf_unset_flag(nbuf, NBUF_DATAREF_RESET);
450 1.1 rmind
451 1.19 rmind /*
452 1.19 rmind * First, cache the L3 header (IPv4 or IPv6). If IP packet is
453 1.19 rmind * fragmented, then we cannot look into L4.
454 1.19 rmind */
455 1.19 rmind flags = npf_cache_ip(npc, nbuf);
456 1.19 rmind if ((flags & NPC_IP46) == 0 || (flags & NPC_IPFRAG) != 0) {
457 1.23 rmind nbuf_unset_flag(nbuf, NBUF_DATAREF_RESET);
458 1.19 rmind npc->npc_info |= flags;
459 1.19 rmind return flags;
460 1.1 rmind }
461 1.19 rmind hlen = npc->npc_hlen;
462 1.19 rmind
463 1.19 rmind switch (npc->npc_proto) {
464 1.1 rmind case IPPROTO_TCP:
465 1.19 rmind /* Cache: layer 4 - TCP. */
466 1.19 rmind npc->npc_l4.tcp = nbuf_advance(nbuf, hlen,
467 1.19 rmind sizeof(struct tcphdr));
468 1.19 rmind l4flags = NPC_LAYER4 | NPC_TCP;
469 1.10 rmind break;
470 1.1 rmind case IPPROTO_UDP:
471 1.19 rmind /* Cache: layer 4 - UDP. */
472 1.19 rmind npc->npc_l4.udp = nbuf_advance(nbuf, hlen,
473 1.19 rmind sizeof(struct udphdr));
474 1.19 rmind l4flags = NPC_LAYER4 | NPC_UDP;
475 1.10 rmind break;
476 1.1 rmind case IPPROTO_ICMP:
477 1.19 rmind /* Cache: layer 4 - ICMPv4. */
478 1.19 rmind npc->npc_l4.icmp = nbuf_advance(nbuf, hlen,
479 1.19 rmind offsetof(struct icmp, icmp_void));
480 1.19 rmind l4flags = NPC_LAYER4 | NPC_ICMP;
481 1.19 rmind break;
482 1.15 spz case IPPROTO_ICMPV6:
483 1.19 rmind /* Cache: layer 4 - ICMPv6. */
484 1.19 rmind npc->npc_l4.icmp6 = nbuf_advance(nbuf, hlen,
485 1.19 rmind offsetof(struct icmp6_hdr, icmp6_data32));
486 1.19 rmind l4flags = NPC_LAYER4 | NPC_ICMP;
487 1.19 rmind break;
488 1.19 rmind default:
489 1.19 rmind l4flags = 0;
490 1.10 rmind break;
491 1.1 rmind }
492 1.19 rmind
493 1.19 rmind if (nbuf_flag_p(nbuf, NBUF_DATAREF_RESET)) {
494 1.19 rmind goto again;
495 1.19 rmind }
496 1.19 rmind
497 1.19 rmind /* Add the L4 flags if nbuf_advance() succeeded. */
498 1.19 rmind if (l4flags && npc->npc_l4.hdr) {
499 1.19 rmind flags |= l4flags;
500 1.19 rmind }
501 1.19 rmind npc->npc_info |= flags;
502 1.19 rmind return flags;
503 1.19 rmind }
504 1.19 rmind
505 1.19 rmind void
506 1.32 rmind npf_recache(npf_cache_t *npc)
507 1.19 rmind {
508 1.32 rmind nbuf_t *nbuf = npc->npc_nbuf;
509 1.24 martin const int mflags __diagused = npc->npc_info & (NPC_IP46 | NPC_LAYER4);
510 1.25 mrg int flags __diagused;
511 1.19 rmind
512 1.19 rmind nbuf_reset(nbuf);
513 1.19 rmind npc->npc_info = 0;
514 1.32 rmind flags = npf_cache_all(npc);
515 1.32 rmind
516 1.19 rmind KASSERT((flags & mflags) == mflags);
517 1.19 rmind KASSERT(nbuf_flag_p(nbuf, NBUF_DATAREF_RESET) == 0);
518 1.1 rmind }
519 1.1 rmind
520 1.1 rmind /*
521 1.19 rmind * npf_rwrip: rewrite required IP address.
522 1.4 rmind */
523 1.4 rmind bool
524 1.28 rmind npf_rwrip(const npf_cache_t *npc, u_int which, const npf_addr_t *addr)
525 1.4 rmind {
526 1.4 rmind KASSERT(npf_iscached(npc, NPC_IP46));
527 1.28 rmind KASSERT(which == NPF_SRC || which == NPF_DST);
528 1.4 rmind
529 1.28 rmind memcpy(npc->npc_ips[which], addr, npc->npc_alen);
530 1.4 rmind return true;
531 1.4 rmind }
532 1.4 rmind
533 1.4 rmind /*
534 1.19 rmind * npf_rwrport: rewrite required TCP/UDP port.
535 1.1 rmind */
536 1.1 rmind bool
537 1.28 rmind npf_rwrport(const npf_cache_t *npc, u_int which, const in_port_t port)
538 1.1 rmind {
539 1.21 rmind const int proto = npc->npc_proto;
540 1.4 rmind in_port_t *oport;
541 1.1 rmind
542 1.4 rmind KASSERT(npf_iscached(npc, NPC_TCP) || npf_iscached(npc, NPC_UDP));
543 1.1 rmind KASSERT(proto == IPPROTO_TCP || proto == IPPROTO_UDP);
544 1.28 rmind KASSERT(which == NPF_SRC || which == NPF_DST);
545 1.1 rmind
546 1.19 rmind /* Get the offset and store the port in it. */
547 1.4 rmind if (proto == IPPROTO_TCP) {
548 1.19 rmind struct tcphdr *th = npc->npc_l4.tcp;
549 1.28 rmind oport = (which == NPF_SRC) ? &th->th_sport : &th->th_dport;
550 1.1 rmind } else {
551 1.19 rmind struct udphdr *uh = npc->npc_l4.udp;
552 1.28 rmind oport = (which == NPF_SRC) ? &uh->uh_sport : &uh->uh_dport;
553 1.1 rmind }
554 1.19 rmind memcpy(oport, &port, sizeof(in_port_t));
555 1.1 rmind return true;
556 1.1 rmind }
557 1.1 rmind
558 1.1 rmind /*
559 1.19 rmind * npf_rwrcksum: rewrite IPv4 and/or TCP/UDP checksum.
560 1.1 rmind */
561 1.1 rmind bool
562 1.28 rmind npf_rwrcksum(const npf_cache_t *npc, u_int which,
563 1.19 rmind const npf_addr_t *addr, const in_port_t port)
564 1.1 rmind {
565 1.28 rmind const npf_addr_t *oaddr = npc->npc_ips[which];
566 1.21 rmind const int proto = npc->npc_proto;
567 1.19 rmind const int alen = npc->npc_alen;
568 1.18 rmind uint16_t *ocksum;
569 1.18 rmind in_port_t oport;
570 1.18 rmind
571 1.19 rmind KASSERT(npf_iscached(npc, NPC_LAYER4));
572 1.28 rmind KASSERT(which == NPF_SRC || which == NPF_DST);
573 1.18 rmind
574 1.4 rmind if (npf_iscached(npc, NPC_IP4)) {
575 1.19 rmind struct ip *ip = npc->npc_ip.v4;
576 1.19 rmind uint16_t ipsum = ip->ip_sum;
577 1.4 rmind
578 1.19 rmind /* Recalculate IPv4 checksum and rewrite. */
579 1.19 rmind ip->ip_sum = npf_addr_cksum(ipsum, alen, oaddr, addr);
580 1.4 rmind } else {
581 1.4 rmind /* No checksum for IPv6. */
582 1.4 rmind KASSERT(npf_iscached(npc, NPC_IP6));
583 1.4 rmind }
584 1.4 rmind
585 1.18 rmind /* Nothing else to do for ICMP. */
586 1.30 rmind if (proto == IPPROTO_ICMP || proto == IPPROTO_ICMPV6) {
587 1.4 rmind return true;
588 1.4 rmind }
589 1.7 zoltan KASSERT(npf_iscached(npc, NPC_TCP) || npf_iscached(npc, NPC_UDP));
590 1.4 rmind
591 1.18 rmind /*
592 1.18 rmind * Calculate TCP/UDP checksum:
593 1.18 rmind * - Skip if UDP and the current checksum is zero.
594 1.18 rmind * - Fixup the IP address change.
595 1.18 rmind * - Fixup the port change, if required (non-zero).
596 1.18 rmind */
597 1.4 rmind if (proto == IPPROTO_TCP) {
598 1.19 rmind struct tcphdr *th = npc->npc_l4.tcp;
599 1.4 rmind
600 1.18 rmind ocksum = &th->th_sum;
601 1.28 rmind oport = (which == NPF_SRC) ? th->th_sport : th->th_dport;
602 1.4 rmind } else {
603 1.19 rmind struct udphdr *uh = npc->npc_l4.udp;
604 1.4 rmind
605 1.4 rmind KASSERT(proto == IPPROTO_UDP);
606 1.18 rmind ocksum = &uh->uh_sum;
607 1.18 rmind if (*ocksum == 0) {
608 1.4 rmind /* No need to update. */
609 1.4 rmind return true;
610 1.4 rmind }
611 1.28 rmind oport = (which == NPF_SRC) ? uh->uh_sport : uh->uh_dport;
612 1.18 rmind }
613 1.18 rmind
614 1.19 rmind uint16_t cksum = npf_addr_cksum(*ocksum, alen, oaddr, addr);
615 1.18 rmind if (port) {
616 1.18 rmind cksum = npf_fixup16_cksum(cksum, oport, port);
617 1.4 rmind }
618 1.1 rmind
619 1.19 rmind /* Rewrite TCP/UDP checksum. */
620 1.19 rmind memcpy(ocksum, &cksum, sizeof(uint16_t));
621 1.4 rmind return true;
622 1.4 rmind }
623 1.4 rmind
624 1.29 rmind /*
625 1.30 rmind * npf_napt_rwr: perform address and/or port translation.
626 1.30 rmind */
627 1.30 rmind int
628 1.30 rmind npf_napt_rwr(const npf_cache_t *npc, u_int which,
629 1.30 rmind const npf_addr_t *addr, const in_addr_t port)
630 1.30 rmind {
631 1.30 rmind const unsigned proto = npc->npc_proto;
632 1.30 rmind
633 1.30 rmind /*
634 1.30 rmind * Rewrite IP and/or TCP/UDP checksums first, since we need the
635 1.30 rmind * current (old) address/port for the calculations. Then perform
636 1.30 rmind * the address translation i.e. rewrite source or destination.
637 1.30 rmind */
638 1.30 rmind if (!npf_rwrcksum(npc, which, addr, port)) {
639 1.30 rmind return EINVAL;
640 1.30 rmind }
641 1.30 rmind if (!npf_rwrip(npc, which, addr)) {
642 1.30 rmind return EINVAL;
643 1.30 rmind }
644 1.30 rmind if (port == 0) {
645 1.30 rmind /* Done. */
646 1.30 rmind return 0;
647 1.30 rmind }
648 1.30 rmind
649 1.30 rmind switch (proto) {
650 1.30 rmind case IPPROTO_TCP:
651 1.30 rmind case IPPROTO_UDP:
652 1.30 rmind /* Rewrite source/destination port. */
653 1.30 rmind if (!npf_rwrport(npc, which, port)) {
654 1.30 rmind return EINVAL;
655 1.30 rmind }
656 1.30 rmind break;
657 1.30 rmind case IPPROTO_ICMP:
658 1.30 rmind case IPPROTO_ICMPV6:
659 1.30 rmind KASSERT(npf_iscached(npc, NPC_ICMP));
660 1.30 rmind /* Nothing. */
661 1.30 rmind break;
662 1.30 rmind default:
663 1.30 rmind return ENOTSUP;
664 1.30 rmind }
665 1.30 rmind return 0;
666 1.30 rmind }
667 1.30 rmind
668 1.30 rmind /*
669 1.29 rmind * IPv6-to-IPv6 Network Prefix Translation (NPTv6), as per RFC 6296.
670 1.29 rmind */
671 1.29 rmind
672 1.29 rmind int
673 1.29 rmind npf_npt66_rwr(const npf_cache_t *npc, u_int which, const npf_addr_t *pref,
674 1.29 rmind npf_netmask_t len, uint16_t adj)
675 1.29 rmind {
676 1.29 rmind npf_addr_t *addr = npc->npc_ips[which];
677 1.29 rmind unsigned remnant, word, preflen = len >> 4;
678 1.29 rmind uint32_t sum;
679 1.29 rmind
680 1.29 rmind KASSERT(which == NPF_SRC || which == NPF_DST);
681 1.29 rmind
682 1.29 rmind if (!npf_iscached(npc, NPC_IP6)) {
683 1.29 rmind return EINVAL;
684 1.29 rmind }
685 1.29 rmind if (len <= 48) {
686 1.29 rmind /*
687 1.29 rmind * The word to adjust. Cannot translate the 0xffff
688 1.29 rmind * subnet if /48 or shorter.
689 1.29 rmind */
690 1.29 rmind word = 3;
691 1.29 rmind if (addr->s6_addr16[word] == 0xffff) {
692 1.29 rmind return EINVAL;
693 1.29 rmind }
694 1.29 rmind } else {
695 1.29 rmind /*
696 1.29 rmind * Also, all 0s or 1s in the host part are disallowed for
697 1.29 rmind * longer than /48 prefixes.
698 1.29 rmind */
699 1.29 rmind if ((addr->s6_addr32[2] == 0 && addr->s6_addr32[3] == 0) ||
700 1.29 rmind (addr->s6_addr32[2] == ~0U && addr->s6_addr32[3] == ~0U))
701 1.29 rmind return EINVAL;
702 1.29 rmind
703 1.29 rmind /* Determine the 16-bit word to adjust. */
704 1.29 rmind for (word = 4; word < 8; word++)
705 1.29 rmind if (addr->s6_addr16[word] != 0xffff)
706 1.29 rmind break;
707 1.29 rmind }
708 1.29 rmind
709 1.29 rmind /* Rewrite the prefix. */
710 1.29 rmind for (unsigned i = 0; i < preflen; i++) {
711 1.29 rmind addr->s6_addr16[i] = pref->s6_addr16[i];
712 1.29 rmind }
713 1.29 rmind
714 1.29 rmind /*
715 1.29 rmind * If prefix length is within a 16-bit word (not dividable by 16),
716 1.29 rmind * then prepare a mask, determine the word and adjust it.
717 1.29 rmind */
718 1.29 rmind if ((remnant = len - (preflen << 4)) != 0) {
719 1.29 rmind const uint16_t wordmask = (1U << remnant) - 1;
720 1.29 rmind const unsigned i = preflen;
721 1.29 rmind
722 1.29 rmind addr->s6_addr16[i] = (pref->s6_addr16[i] & wordmask) |
723 1.29 rmind (addr->s6_addr16[i] & ~wordmask);
724 1.29 rmind }
725 1.29 rmind
726 1.29 rmind /*
727 1.29 rmind * Performing 1's complement sum/difference.
728 1.29 rmind */
729 1.29 rmind sum = addr->s6_addr16[word] + adj;
730 1.29 rmind while (sum >> 16) {
731 1.29 rmind sum = (sum >> 16) + (sum & 0xffff);
732 1.29 rmind }
733 1.29 rmind if (sum == 0xffff) {
734 1.29 rmind /* RFC 1071. */
735 1.29 rmind sum = 0x0000;
736 1.29 rmind }
737 1.29 rmind addr->s6_addr16[word] = sum;
738 1.29 rmind return 0;
739 1.29 rmind }
740 1.29 rmind
741 1.13 rmind #if defined(DDB) || defined(_NPF_TESTING)
742 1.13 rmind
743 1.31 rmind const char *
744 1.31 rmind npf_addr_dump(const npf_addr_t *addr, int alen)
745 1.13 rmind {
746 1.31 rmind if (alen == sizeof(struct in_addr)) {
747 1.31 rmind struct in_addr ip;
748 1.31 rmind memcpy(&ip, addr, alen);
749 1.31 rmind return inet_ntoa(ip);
750 1.31 rmind }
751 1.33 mlelstv return ip6_sprintf(addr);
752 1.13 rmind }
753 1.13 rmind
754 1.13 rmind #endif
755