Home | History | Annotate | Line # | Download | only in npf
npf_inet.c revision 1.50 
      1 /*	$NetBSD: npf_inet.c,v 1.50 2018/04/08 05:51:45 maxv Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 2009-2014 The NetBSD Foundation, Inc.
      5  * All rights reserved.
      6  *
      7  * This material is based upon work partially supported by The
      8  * NetBSD Foundation under a contract with Mindaugas Rasiukevicius.
      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  *
     19  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     20  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     21  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     22  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     23  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     24  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     25  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     26  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     27  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     28  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     29  * POSSIBILITY OF SUCH DAMAGE.
     30  */
     31 
     32 /*
     33  * Various protocol related helper routines.
     34  *
     35  * This layer manipulates npf_cache_t structure i.e. caches requested headers
     36  * and stores which information was cached in the information bit field.
     37  * It is also responsibility of this layer to update or invalidate the cache
     38  * on rewrites (e.g. by translation routines).
     39  */
     40 
     41 #ifdef _KERNEL
     42 #include <sys/cdefs.h>
     43 __KERNEL_RCSID(0, "$NetBSD: npf_inet.c,v 1.50 2018/04/08 05:51:45 maxv Exp $");
     44 
     45 #include <sys/param.h>
     46 #include <sys/types.h>
     47 
     48 #include <net/pfil.h>
     49 #include <net/if.h>
     50 #include <net/ethertypes.h>
     51 #include <net/if_ether.h>
     52 
     53 #include <netinet/in_systm.h>
     54 #include <netinet/in.h>
     55 #include <netinet6/in6_var.h>
     56 #include <netinet/ip.h>
     57 #include <netinet/ip6.h>
     58 #include <netinet/tcp.h>
     59 #include <netinet/udp.h>
     60 #include <netinet/ip_icmp.h>
     61 #endif
     62 
     63 #include "npf_impl.h"
     64 
     65 /*
     66  * npf_fixup{16,32}_cksum: incremental update of the Internet checksum.
     67  */
     68 
     69 uint16_t
     70 npf_fixup16_cksum(uint16_t cksum, uint16_t odatum, uint16_t ndatum)
     71 {
     72 	uint32_t sum;
     73 
     74 	/*
     75 	 * RFC 1624:
     76 	 *	HC' = ~(~HC + ~m + m')
     77 	 *
     78 	 * Note: 1's complement sum is endian-independent (RFC 1071, page 2).
     79 	 */
     80 	sum = ~cksum & 0xffff;
     81 	sum += (~odatum & 0xffff) + ndatum;
     82 	sum = (sum >> 16) + (sum & 0xffff);
     83 	sum += (sum >> 16);
     84 
     85 	return ~sum & 0xffff;
     86 }
     87 
     88 uint16_t
     89 npf_fixup32_cksum(uint16_t cksum, uint32_t odatum, uint32_t ndatum)
     90 {
     91 	uint32_t sum;
     92 
     93 	/*
     94 	 * Checksum 32-bit datum as as two 16-bit.  Note, the first
     95 	 * 32->16 bit reduction is not necessary.
     96 	 */
     97 	sum = ~cksum & 0xffff;
     98 	sum += (~odatum & 0xffff) + (ndatum & 0xffff);
     99 
    100 	sum += (~odatum >> 16) + (ndatum >> 16);
    101 	sum = (sum >> 16) + (sum & 0xffff);
    102 	sum += (sum >> 16);
    103 	return ~sum & 0xffff;
    104 }
    105 
    106 /*
    107  * npf_addr_cksum: calculate checksum of the address, either IPv4 or IPv6.
    108  */
    109 uint16_t
    110 npf_addr_cksum(uint16_t cksum, int sz, const npf_addr_t *oaddr,
    111     const npf_addr_t *naddr)
    112 {
    113 	const uint32_t *oip32 = (const uint32_t *)oaddr;
    114 	const uint32_t *nip32 = (const uint32_t *)naddr;
    115 
    116 	KASSERT(sz % sizeof(uint32_t) == 0);
    117 	do {
    118 		cksum = npf_fixup32_cksum(cksum, *oip32++, *nip32++);
    119 		sz -= sizeof(uint32_t);
    120 	} while (sz);
    121 
    122 	return cksum;
    123 }
    124 
    125 /*
    126  * npf_addr_sum: provide IP addresses as a XORed 32-bit integer.
    127  * Note: used for hash function.
    128  */
    129 uint32_t
    130 npf_addr_mix(const int sz, const npf_addr_t *a1, const npf_addr_t *a2)
    131 {
    132 	uint32_t mix = 0;
    133 
    134 	KASSERT(sz > 0 && a1 != NULL && a2 != NULL);
    135 
    136 	for (int i = 0; i < (sz >> 2); i++) {
    137 		mix ^= a1->word32[i];
    138 		mix ^= a2->word32[i];
    139 	}
    140 	return mix;
    141 }
    142 
    143 /*
    144  * npf_addr_mask: apply the mask to a given address and store the result.
    145  */
    146 void
    147 npf_addr_mask(const npf_addr_t *addr, const npf_netmask_t mask,
    148     const int alen, npf_addr_t *out)
    149 {
    150 	const int nwords = alen >> 2;
    151 	uint_fast8_t length = mask;
    152 
    153 	/* Note: maximum length is 32 for IPv4 and 128 for IPv6. */
    154 	KASSERT(length <= NPF_MAX_NETMASK);
    155 
    156 	for (int i = 0; i < nwords; i++) {
    157 		uint32_t wordmask;
    158 
    159 		if (length >= 32) {
    160 			wordmask = htonl(0xffffffff);
    161 			length -= 32;
    162 		} else if (length) {
    163 			wordmask = htonl(0xffffffff << (32 - length));
    164 			length = 0;
    165 		} else {
    166 			wordmask = 0;
    167 		}
    168 		out->word32[i] = addr->word32[i] & wordmask;
    169 	}
    170 }
    171 
    172 /*
    173  * npf_addr_cmp: compare two addresses, either IPv4 or IPv6.
    174  *
    175  * => Return 0 if equal and negative/positive if less/greater accordingly.
    176  * => Ignore the mask, if NPF_NO_NETMASK is specified.
    177  */
    178 int
    179 npf_addr_cmp(const npf_addr_t *addr1, const npf_netmask_t mask1,
    180     const npf_addr_t *addr2, const npf_netmask_t mask2, const int alen)
    181 {
    182 	npf_addr_t realaddr1, realaddr2;
    183 
    184 	if (mask1 != NPF_NO_NETMASK) {
    185 		npf_addr_mask(addr1, mask1, alen, &realaddr1);
    186 		addr1 = &realaddr1;
    187 	}
    188 	if (mask2 != NPF_NO_NETMASK) {
    189 		npf_addr_mask(addr2, mask2, alen, &realaddr2);
    190 		addr2 = &realaddr2;
    191 	}
    192 	return memcmp(addr1, addr2, alen);
    193 }
    194 
    195 /*
    196  * npf_tcpsaw: helper to fetch SEQ, ACK, WIN and return TCP data length.
    197  *
    198  * => Returns all values in host byte-order.
    199  */
    200 int
    201 npf_tcpsaw(const npf_cache_t *npc, tcp_seq *seq, tcp_seq *ack, uint32_t *win)
    202 {
    203 	const struct tcphdr *th = npc->npc_l4.tcp;
    204 	u_int thlen;
    205 
    206 	KASSERT(npf_iscached(npc, NPC_TCP));
    207 
    208 	*seq = ntohl(th->th_seq);
    209 	*ack = ntohl(th->th_ack);
    210 	*win = (uint32_t)ntohs(th->th_win);
    211 	thlen = th->th_off << 2;
    212 
    213 	if (npf_iscached(npc, NPC_IP4)) {
    214 		const struct ip *ip = npc->npc_ip.v4;
    215 		return ntohs(ip->ip_len) - npc->npc_hlen - thlen;
    216 	} else if (npf_iscached(npc, NPC_IP6)) {
    217 		const struct ip6_hdr *ip6 = npc->npc_ip.v6;
    218 		return ntohs(ip6->ip6_plen) -
    219 		    (npc->npc_hlen - sizeof(*ip6)) - thlen;
    220 	}
    221 	return 0;
    222 }
    223 
    224 /*
    225  * npf_fetch_tcpopts: parse and return TCP options.
    226  */
    227 bool
    228 npf_fetch_tcpopts(npf_cache_t *npc, uint16_t *mss, int *wscale)
    229 {
    230 	nbuf_t *nbuf = npc->npc_nbuf;
    231 	const struct tcphdr *th = npc->npc_l4.tcp;
    232 	int cnt, optlen = 0;
    233 	bool setmss = false;
    234 	uint8_t *cp, opt;
    235 	uint8_t val;
    236 	bool ok;
    237 
    238 	KASSERT(npf_iscached(npc, NPC_IP46));
    239 	KASSERT(npf_iscached(npc, NPC_TCP));
    240 
    241 	/* Determine if there are any TCP options, get their length. */
    242 	cnt = (th->th_off << 2) - sizeof(struct tcphdr);
    243 	if (cnt <= 0) {
    244 		/* No options. */
    245 		return false;
    246 	}
    247 	KASSERT(cnt <= MAX_TCPOPTLEN);
    248 
    249 	/* Determine if we want to set or get the mss. */
    250 	if (mss) {
    251 		setmss = (*mss != 0);
    252 	}
    253 
    254 	/* Fetch all the options at once. */
    255 	nbuf_reset(nbuf);
    256 	const int step = npc->npc_hlen + sizeof(struct tcphdr);
    257 	if ((cp = nbuf_advance(nbuf, step, cnt)) == NULL) {
    258 		ok = false;
    259 		goto done;
    260 	}
    261 
    262 	/* Scan the options. */
    263 	for (; cnt > 0; cnt -= optlen, cp += optlen) {
    264 		opt = cp[0];
    265 		if (opt == TCPOPT_EOL)
    266 			break;
    267 		if (opt == TCPOPT_NOP)
    268 			optlen = 1;
    269 		else {
    270 			if (cnt < 2)
    271 				break;
    272 			optlen = cp[1];
    273 			if (optlen < 2 || optlen > cnt)
    274 				break;
    275 		}
    276 
    277 		switch (opt) {
    278 		case TCPOPT_MAXSEG:
    279 			if (optlen != TCPOLEN_MAXSEG)
    280 				continue;
    281 			if (mss) {
    282 				if (setmss) {
    283 					memcpy(cp + 2, mss, sizeof(uint16_t));
    284 				} else {
    285 					memcpy(mss, cp + 2, sizeof(uint16_t));
    286 				}
    287 			}
    288 			break;
    289 		case TCPOPT_WINDOW:
    290 			if (optlen != TCPOLEN_WINDOW)
    291 				continue;
    292 			val = *(cp + 2);
    293 			*wscale = (val > TCP_MAX_WINSHIFT) ? TCP_MAX_WINSHIFT : val;
    294 			break;
    295 		default:
    296 			break;
    297 		}
    298 	}
    299 
    300 	ok = true;
    301 done:
    302 	if (nbuf_flag_p(nbuf, NBUF_DATAREF_RESET)) {
    303 		npf_recache(npc);
    304 	}
    305 	return ok;
    306 }
    307 
    308 static int
    309 npf_cache_ip(npf_cache_t *npc, nbuf_t *nbuf)
    310 {
    311 	const void *nptr = nbuf_dataptr(nbuf);
    312 	const uint8_t ver = *(const uint8_t *)nptr;
    313 	int flags = 0;
    314 
    315 	/*
    316 	 * We intentionally don't read the L4 payload after IPPROTO_AH.
    317 	 */
    318 
    319 	switch (ver >> 4) {
    320 	case IPVERSION: {
    321 		struct ip *ip;
    322 
    323 		ip = nbuf_ensure_contig(nbuf, sizeof(struct ip));
    324 		if (ip == NULL) {
    325 			return NPC_FMTERR;
    326 		}
    327 
    328 		/* Retrieve the complete header. */
    329 		if ((u_int)(ip->ip_hl << 2) < sizeof(struct ip)) {
    330 			return NPC_FMTERR;
    331 		}
    332 		ip = nbuf_ensure_contig(nbuf, (u_int)(ip->ip_hl << 2));
    333 		if (ip == NULL) {
    334 			return NPC_FMTERR;
    335 		}
    336 
    337 		if (ip->ip_off & ~htons(IP_DF | IP_RF)) {
    338 			/* Note fragmentation. */
    339 			flags |= NPC_IPFRAG;
    340 		}
    341 
    342 		/* Cache: layer 3 - IPv4. */
    343 		npc->npc_alen = sizeof(struct in_addr);
    344 		npc->npc_ips[NPF_SRC] = (npf_addr_t *)&ip->ip_src;
    345 		npc->npc_ips[NPF_DST] = (npf_addr_t *)&ip->ip_dst;
    346 		npc->npc_hlen = ip->ip_hl << 2;
    347 		npc->npc_proto = ip->ip_p;
    348 
    349 		npc->npc_ip.v4 = ip;
    350 		flags |= NPC_IP4;
    351 		break;
    352 	}
    353 
    354 	case (IPV6_VERSION >> 4): {
    355 		struct ip6_hdr *ip6;
    356 		struct ip6_ext *ip6e;
    357 		struct ip6_frag *ip6f;
    358 		size_t off, hlen;
    359 		int frag_present;
    360 
    361 		ip6 = nbuf_ensure_contig(nbuf, sizeof(struct ip6_hdr));
    362 		if (ip6 == NULL) {
    363 			return NPC_FMTERR;
    364 		}
    365 
    366 		/*
    367 		 * XXX: We don't handle IPv6 Jumbograms.
    368 		 */
    369 
    370 		/* Set initial next-protocol value. */
    371 		hlen = sizeof(struct ip6_hdr);
    372 		npc->npc_proto = ip6->ip6_nxt;
    373 		npc->npc_hlen = hlen;
    374 
    375 		frag_present = 0;
    376 
    377 		/*
    378 		 * Advance by the length of the current header.
    379 		 */
    380 		off = nbuf_offset(nbuf);
    381 		while ((ip6e = nbuf_advance(nbuf, hlen, sizeof(*ip6e))) != NULL) {
    382 			/*
    383 			 * Determine whether we are going to continue.
    384 			 */
    385 			switch (npc->npc_proto) {
    386 			case IPPROTO_HOPOPTS:
    387 			case IPPROTO_DSTOPTS:
    388 			case IPPROTO_ROUTING:
    389 				hlen = (ip6e->ip6e_len + 1) << 3;
    390 				break;
    391 			case IPPROTO_FRAGMENT:
    392 				if (frag_present++)
    393 					return NPC_FMTERR;
    394 				ip6f = nbuf_ensure_contig(nbuf, sizeof(*ip6f));
    395 				if (ip6f == NULL)
    396 					return NPC_FMTERR;
    397 
    398 				/* RFC6946: Skip dummy fragments. */
    399 				if (!ntohs(ip6f->ip6f_offlg & IP6F_OFF_MASK) &&
    400 				    !(ip6f->ip6f_offlg & IP6F_MORE_FRAG)) {
    401 					hlen = sizeof(struct ip6_frag);
    402 					break;
    403 				}
    404 
    405 				hlen = 0;
    406 				flags |= NPC_IPFRAG;
    407 
    408 				break;
    409 			default:
    410 				hlen = 0;
    411 				break;
    412 			}
    413 
    414 			if (!hlen) {
    415 				break;
    416 			}
    417 			npc->npc_proto = ip6e->ip6e_nxt;
    418 			npc->npc_hlen += hlen;
    419 		}
    420 
    421 		if (ip6e == NULL) {
    422 			return NPC_FMTERR;
    423 		}
    424 
    425 		/*
    426 		 * Re-fetch the header pointers (nbufs might have been
    427 		 * reallocated).  Restore the original offset (if any).
    428 		 */
    429 		nbuf_reset(nbuf);
    430 		ip6 = nbuf_dataptr(nbuf);
    431 		if (off) {
    432 			nbuf_advance(nbuf, off, 0);
    433 		}
    434 
    435 		/* Cache: layer 3 - IPv6. */
    436 		npc->npc_alen = sizeof(struct in6_addr);
    437 		npc->npc_ips[NPF_SRC] = (npf_addr_t *)&ip6->ip6_src;
    438 		npc->npc_ips[NPF_DST] = (npf_addr_t *)&ip6->ip6_dst;
    439 
    440 		npc->npc_ip.v6 = ip6;
    441 		flags |= NPC_IP6;
    442 		break;
    443 	}
    444 	default:
    445 		break;
    446 	}
    447 	return flags;
    448 }
    449 
    450 /*
    451  * npf_cache_all: general routine to cache all relevant IP (v4 or v6)
    452  * and TCP, UDP or ICMP headers.
    453  *
    454  * => nbuf offset shall be set accordingly.
    455  */
    456 int
    457 npf_cache_all(npf_cache_t *npc)
    458 {
    459 	nbuf_t *nbuf = npc->npc_nbuf;
    460 	int flags, l4flags;
    461 	u_int hlen;
    462 
    463 	/*
    464 	 * This routine is a main point where the references are cached,
    465 	 * therefore clear the flag as we reset.
    466 	 */
    467 again:
    468 	nbuf_unset_flag(nbuf, NBUF_DATAREF_RESET);
    469 
    470 	/*
    471 	 * First, cache the L3 header (IPv4 or IPv6).  If IP packet is
    472 	 * fragmented, then we cannot look into L4.
    473 	 */
    474 	flags = npf_cache_ip(npc, nbuf);
    475 	if ((flags & NPC_IP46) == 0 || (flags & NPC_IPFRAG) != 0 ||
    476 	    (flags & NPC_FMTERR) != 0) {
    477 		goto out;
    478 	}
    479 	hlen = npc->npc_hlen;
    480 
    481 	/*
    482 	 * Note: we guarantee that the potential "Query Id" field of the
    483 	 * ICMPv4/ICMPv6 packets is in the nbuf. This field is used in the
    484 	 * ICMP ALG.
    485 	 */
    486 	switch (npc->npc_proto) {
    487 	case IPPROTO_TCP:
    488 		/* Cache: layer 4 - TCP. */
    489 		npc->npc_l4.tcp = nbuf_advance(nbuf, hlen,
    490 		    sizeof(struct tcphdr));
    491 		l4flags = NPC_LAYER4 | NPC_TCP;
    492 		break;
    493 	case IPPROTO_UDP:
    494 		/* Cache: layer 4 - UDP. */
    495 		npc->npc_l4.udp = nbuf_advance(nbuf, hlen,
    496 		    sizeof(struct udphdr));
    497 		l4flags = NPC_LAYER4 | NPC_UDP;
    498 		break;
    499 	case IPPROTO_ICMP:
    500 		/* Cache: layer 4 - ICMPv4. */
    501 		npc->npc_l4.icmp = nbuf_advance(nbuf, hlen,
    502 		    ICMP_MINLEN);
    503 		l4flags = NPC_LAYER4 | NPC_ICMP;
    504 		break;
    505 	case IPPROTO_ICMPV6:
    506 		/* Cache: layer 4 - ICMPv6. */
    507 		npc->npc_l4.icmp6 = nbuf_advance(nbuf, hlen,
    508 		    sizeof(struct icmp6_hdr));
    509 		l4flags = NPC_LAYER4 | NPC_ICMP;
    510 		break;
    511 	default:
    512 		l4flags = 0;
    513 		break;
    514 	}
    515 
    516 	/* Error out if nbuf_advance failed. */
    517 	if (l4flags && npc->npc_l4.hdr == NULL) {
    518 		goto err;
    519 	}
    520 
    521 	if (nbuf_flag_p(nbuf, NBUF_DATAREF_RESET)) {
    522 		goto again;
    523 	}
    524 
    525 	flags |= l4flags;
    526 	npc->npc_info |= flags;
    527 	return flags;
    528 
    529 err:
    530 	flags = NPC_FMTERR;
    531 out:
    532 	nbuf_unset_flag(nbuf, NBUF_DATAREF_RESET);
    533 	npc->npc_info |= flags;
    534 	return flags;
    535 }
    536 
    537 void
    538 npf_recache(npf_cache_t *npc)
    539 {
    540 	nbuf_t *nbuf = npc->npc_nbuf;
    541 	const int mflags __diagused = npc->npc_info & (NPC_IP46 | NPC_LAYER4);
    542 	int flags __diagused;
    543 
    544 	nbuf_reset(nbuf);
    545 	npc->npc_info = 0;
    546 	flags = npf_cache_all(npc);
    547 
    548 	KASSERT((flags & mflags) == mflags);
    549 	KASSERT(nbuf_flag_p(nbuf, NBUF_DATAREF_RESET) == 0);
    550 }
    551 
    552 /*
    553  * npf_rwrip: rewrite required IP address.
    554  */
    555 bool
    556 npf_rwrip(const npf_cache_t *npc, u_int which, const npf_addr_t *addr)
    557 {
    558 	KASSERT(npf_iscached(npc, NPC_IP46));
    559 	KASSERT(which == NPF_SRC || which == NPF_DST);
    560 
    561 	memcpy(npc->npc_ips[which], addr, npc->npc_alen);
    562 	return true;
    563 }
    564 
    565 /*
    566  * npf_rwrport: rewrite required TCP/UDP port.
    567  */
    568 bool
    569 npf_rwrport(const npf_cache_t *npc, u_int which, const in_port_t port)
    570 {
    571 	const int proto = npc->npc_proto;
    572 	in_port_t *oport;
    573 
    574 	KASSERT(npf_iscached(npc, NPC_TCP) || npf_iscached(npc, NPC_UDP));
    575 	KASSERT(proto == IPPROTO_TCP || proto == IPPROTO_UDP);
    576 	KASSERT(which == NPF_SRC || which == NPF_DST);
    577 
    578 	/* Get the offset and store the port in it. */
    579 	if (proto == IPPROTO_TCP) {
    580 		struct tcphdr *th = npc->npc_l4.tcp;
    581 		oport = (which == NPF_SRC) ? &th->th_sport : &th->th_dport;
    582 	} else {
    583 		struct udphdr *uh = npc->npc_l4.udp;
    584 		oport = (which == NPF_SRC) ? &uh->uh_sport : &uh->uh_dport;
    585 	}
    586 	memcpy(oport, &port, sizeof(in_port_t));
    587 	return true;
    588 }
    589 
    590 /*
    591  * npf_rwrcksum: rewrite IPv4 and/or TCP/UDP checksum.
    592  */
    593 bool
    594 npf_rwrcksum(const npf_cache_t *npc, u_int which,
    595     const npf_addr_t *addr, const in_port_t port)
    596 {
    597 	const npf_addr_t *oaddr = npc->npc_ips[which];
    598 	const int proto = npc->npc_proto;
    599 	const int alen = npc->npc_alen;
    600 	uint16_t *ocksum;
    601 	in_port_t oport;
    602 
    603 	KASSERT(npf_iscached(npc, NPC_LAYER4));
    604 	KASSERT(which == NPF_SRC || which == NPF_DST);
    605 
    606 	if (npf_iscached(npc, NPC_IP4)) {
    607 		struct ip *ip = npc->npc_ip.v4;
    608 		uint16_t ipsum = ip->ip_sum;
    609 
    610 		/* Recalculate IPv4 checksum and rewrite. */
    611 		ip->ip_sum = npf_addr_cksum(ipsum, alen, oaddr, addr);
    612 	} else {
    613 		/* No checksum for IPv6. */
    614 		KASSERT(npf_iscached(npc, NPC_IP6));
    615 	}
    616 
    617 	/* Nothing else to do for ICMP. */
    618 	if (proto == IPPROTO_ICMP || proto == IPPROTO_ICMPV6) {
    619 		return true;
    620 	}
    621 	KASSERT(npf_iscached(npc, NPC_TCP) || npf_iscached(npc, NPC_UDP));
    622 
    623 	/*
    624 	 * Calculate TCP/UDP checksum:
    625 	 * - Skip if UDP and the current checksum is zero.
    626 	 * - Fixup the IP address change.
    627 	 * - Fixup the port change, if required (non-zero).
    628 	 */
    629 	if (proto == IPPROTO_TCP) {
    630 		struct tcphdr *th = npc->npc_l4.tcp;
    631 
    632 		ocksum = &th->th_sum;
    633 		oport = (which == NPF_SRC) ? th->th_sport : th->th_dport;
    634 	} else {
    635 		struct udphdr *uh = npc->npc_l4.udp;
    636 
    637 		KASSERT(proto == IPPROTO_UDP);
    638 		ocksum = &uh->uh_sum;
    639 		if (*ocksum == 0) {
    640 			/* No need to update. */
    641 			return true;
    642 		}
    643 		oport = (which == NPF_SRC) ? uh->uh_sport : uh->uh_dport;
    644 	}
    645 
    646 	uint16_t cksum = npf_addr_cksum(*ocksum, alen, oaddr, addr);
    647 	if (port) {
    648 		cksum = npf_fixup16_cksum(cksum, oport, port);
    649 	}
    650 
    651 	/* Rewrite TCP/UDP checksum. */
    652 	memcpy(ocksum, &cksum, sizeof(uint16_t));
    653 	return true;
    654 }
    655 
    656 /*
    657  * npf_napt_rwr: perform address and/or port translation.
    658  */
    659 int
    660 npf_napt_rwr(const npf_cache_t *npc, u_int which,
    661     const npf_addr_t *addr, const in_addr_t port)
    662 {
    663 	const unsigned proto = npc->npc_proto;
    664 
    665 	/*
    666 	 * Rewrite IP and/or TCP/UDP checksums first, since we need the
    667 	 * current (old) address/port for the calculations.  Then perform
    668 	 * the address translation i.e. rewrite source or destination.
    669 	 */
    670 	if (!npf_rwrcksum(npc, which, addr, port)) {
    671 		return EINVAL;
    672 	}
    673 	if (!npf_rwrip(npc, which, addr)) {
    674 		return EINVAL;
    675 	}
    676 	if (port == 0) {
    677 		/* Done. */
    678 		return 0;
    679 	}
    680 
    681 	switch (proto) {
    682 	case IPPROTO_TCP:
    683 	case IPPROTO_UDP:
    684 		/* Rewrite source/destination port. */
    685 		if (!npf_rwrport(npc, which, port)) {
    686 			return EINVAL;
    687 		}
    688 		break;
    689 	case IPPROTO_ICMP:
    690 	case IPPROTO_ICMPV6:
    691 		KASSERT(npf_iscached(npc, NPC_ICMP));
    692 		/* Nothing. */
    693 		break;
    694 	default:
    695 		return ENOTSUP;
    696 	}
    697 	return 0;
    698 }
    699 
    700 /*
    701  * IPv6-to-IPv6 Network Prefix Translation (NPTv6), as per RFC 6296.
    702  */
    703 
    704 int
    705 npf_npt66_rwr(const npf_cache_t *npc, u_int which, const npf_addr_t *pref,
    706     npf_netmask_t len, uint16_t adj)
    707 {
    708 	npf_addr_t *addr = npc->npc_ips[which];
    709 	unsigned remnant, word, preflen = len >> 4;
    710 	uint32_t sum;
    711 
    712 	KASSERT(which == NPF_SRC || which == NPF_DST);
    713 
    714 	if (!npf_iscached(npc, NPC_IP6)) {
    715 		return EINVAL;
    716 	}
    717 	if (len <= 48) {
    718 		/*
    719 		 * The word to adjust.  Cannot translate the 0xffff
    720 		 * subnet if /48 or shorter.
    721 		 */
    722 		word = 3;
    723 		if (addr->word16[word] == 0xffff) {
    724 			return EINVAL;
    725 		}
    726 	} else {
    727 		/*
    728 		 * Also, all 0s or 1s in the host part are disallowed for
    729 		 * longer than /48 prefixes.
    730 		 */
    731 		if ((addr->word32[2] == 0 && addr->word32[3] == 0) ||
    732 		    (addr->word32[2] == ~0U && addr->word32[3] == ~0U))
    733 			return EINVAL;
    734 
    735 		/* Determine the 16-bit word to adjust. */
    736 		for (word = 4; word < 8; word++)
    737 			if (addr->word16[word] != 0xffff)
    738 				break;
    739 	}
    740 
    741 	/* Rewrite the prefix. */
    742 	for (unsigned i = 0; i < preflen; i++) {
    743 		addr->word16[i] = pref->word16[i];
    744 	}
    745 
    746 	/*
    747 	 * If prefix length is within a 16-bit word (not dividable by 16),
    748 	 * then prepare a mask, determine the word and adjust it.
    749 	 */
    750 	if ((remnant = len - (preflen << 4)) != 0) {
    751 		const uint16_t wordmask = (1U << remnant) - 1;
    752 		const unsigned i = preflen;
    753 
    754 		addr->word16[i] = (pref->word16[i] & wordmask) |
    755 		    (addr->word16[i] & ~wordmask);
    756 	}
    757 
    758 	/*
    759 	 * Performing 1's complement sum/difference.
    760 	 */
    761 	sum = addr->word16[word] + adj;
    762 	while (sum >> 16) {
    763 		sum = (sum >> 16) + (sum & 0xffff);
    764 	}
    765 	if (sum == 0xffff) {
    766 		/* RFC 1071. */
    767 		sum = 0x0000;
    768 	}
    769 	addr->word16[word] = sum;
    770 	return 0;
    771 }
    772 
    773 #if defined(DDB) || defined(_NPF_TESTING)
    774 
    775 const char *
    776 npf_addr_dump(const npf_addr_t *addr, int alen)
    777 {
    778 	if (alen == sizeof(struct in_addr)) {
    779 		struct in_addr ip;
    780 		memcpy(&ip, addr, alen);
    781 		return inet_ntoa(ip);
    782 	}
    783 	return "[IPv6]";
    784 }
    785 
    786 #endif
    787