tcp_ipv6.c 53.9 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	TCP over IPv6
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7
 *
8
 *	Based on:
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
 *	linux/net/ipv4/tcp.c
 *	linux/net/ipv4/tcp_input.c
 *	linux/net/ipv4/tcp_output.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *	YOSHIFUJI Hideaki @USAGI:	convert /proc/net/tcp6 to seq_file.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/module.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/jiffies.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/init.h>
#include <linux/jhash.h>
#include <linux/ipsec.h>
#include <linux/times.h>

#include <linux/ipv6.h>
#include <linux/icmpv6.h>
#include <linux/random.h>

#include <net/tcp.h>
#include <net/ndisc.h>
47
#include <net/inet6_hashtables.h>
48
#include <net/inet6_connection_sock.h>
L
Linus Torvalds 已提交
49 50 51 52 53 54 55 56 57 58
#include <net/ipv6.h>
#include <net/transp_v6.h>
#include <net/addrconf.h>
#include <net/ip6_route.h>
#include <net/ip6_checksum.h>
#include <net/inet_ecn.h>
#include <net/protocol.h>
#include <net/xfrm.h>
#include <net/snmp.h>
#include <net/dsfield.h>
59
#include <net/timewait_sock.h>
60
#include <net/netdma.h>
61
#include <net/inet_common.h>
L
Linus Torvalds 已提交
62 63 64 65 66 67

#include <asm/uaccess.h>

#include <linux/proc_fs.h>
#include <linux/seq_file.h>

68 69 70 71
#include <linux/crypto.h>
#include <linux/scatterlist.h>

static void	tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb);
72
static void	tcp_v6_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req);
L
Linus Torvalds 已提交
73 74 75

static int	tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb);

76 77
static struct inet_connection_sock_af_ops ipv6_mapped;
static struct inet_connection_sock_af_ops ipv6_specific;
78
#ifdef CONFIG_TCP_MD5SIG
79 80
static struct tcp_sock_af_ops tcp_sock_ipv6_specific;
static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific;
81 82 83 84 85 86
#else
static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk,
						   struct in6_addr *addr)
{
	return NULL;
}
87
#endif
L
Linus Torvalds 已提交
88 89 90 91

static void tcp_v6_hash(struct sock *sk)
{
	if (sk->sk_state != TCP_CLOSE) {
92
		if (inet_csk(sk)->icsk_af_ops == &ipv6_mapped) {
L
Linus Torvalds 已提交
93 94 95 96
			tcp_prot.hash(sk);
			return;
		}
		local_bh_disable();
97
		__inet6_hash(sk);
L
Linus Torvalds 已提交
98 99 100 101
		local_bh_enable();
	}
}

102
static __inline__ __sum16 tcp_v6_check(struct tcphdr *th, int len,
103 104
				   struct in6_addr *saddr,
				   struct in6_addr *daddr,
105
				   __wsum base)
L
Linus Torvalds 已提交
106 107 108 109
{
	return csum_ipv6_magic(saddr, daddr, len, IPPROTO_TCP, base);
}

110
static __u32 tcp_v6_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
111
{
112 113
	return secure_tcpv6_sequence_number(ipv6_hdr(skb)->daddr.s6_addr32,
					    ipv6_hdr(skb)->saddr.s6_addr32,
114 115
					    tcp_hdr(skb)->dest,
					    tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
116 117
}

118
static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
L
Linus Torvalds 已提交
119 120 121
			  int addr_len)
{
	struct sockaddr_in6 *usin = (struct sockaddr_in6 *) uaddr;
122
	struct inet_sock *inet = inet_sk(sk);
123
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
124 125 126 127 128 129 130 131
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
	struct in6_addr *saddr = NULL, *final_p = NULL, final;
	struct flowi fl;
	struct dst_entry *dst;
	int addr_type;
	int err;

132
	if (addr_len < SIN6_LEN_RFC2133)
L
Linus Torvalds 已提交
133 134
		return -EINVAL;

135
	if (usin->sin6_family != AF_INET6)
L
Linus Torvalds 已提交
136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
		return(-EAFNOSUPPORT);

	memset(&fl, 0, sizeof(fl));

	if (np->sndflow) {
		fl.fl6_flowlabel = usin->sin6_flowinfo&IPV6_FLOWINFO_MASK;
		IP6_ECN_flow_init(fl.fl6_flowlabel);
		if (fl.fl6_flowlabel&IPV6_FLOWLABEL_MASK) {
			struct ip6_flowlabel *flowlabel;
			flowlabel = fl6_sock_lookup(sk, fl.fl6_flowlabel);
			if (flowlabel == NULL)
				return -EINVAL;
			ipv6_addr_copy(&usin->sin6_addr, &flowlabel->dst);
			fl6_sock_release(flowlabel);
		}
	}

	/*
154 155 156 157 158
	 *	connect() to INADDR_ANY means loopback (BSD'ism).
	 */

	if(ipv6_addr_any(&usin->sin6_addr))
		usin->sin6_addr.s6_addr[15] = 0x1;
L
Linus Torvalds 已提交
159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197

	addr_type = ipv6_addr_type(&usin->sin6_addr);

	if(addr_type & IPV6_ADDR_MULTICAST)
		return -ENETUNREACH;

	if (addr_type&IPV6_ADDR_LINKLOCAL) {
		if (addr_len >= sizeof(struct sockaddr_in6) &&
		    usin->sin6_scope_id) {
			/* If interface is set while binding, indices
			 * must coincide.
			 */
			if (sk->sk_bound_dev_if &&
			    sk->sk_bound_dev_if != usin->sin6_scope_id)
				return -EINVAL;

			sk->sk_bound_dev_if = usin->sin6_scope_id;
		}

		/* Connect to link-local address requires an interface */
		if (!sk->sk_bound_dev_if)
			return -EINVAL;
	}

	if (tp->rx_opt.ts_recent_stamp &&
	    !ipv6_addr_equal(&np->daddr, &usin->sin6_addr)) {
		tp->rx_opt.ts_recent = 0;
		tp->rx_opt.ts_recent_stamp = 0;
		tp->write_seq = 0;
	}

	ipv6_addr_copy(&np->daddr, &usin->sin6_addr);
	np->flow_label = fl.fl6_flowlabel;

	/*
	 *	TCP over IPv4
	 */

	if (addr_type == IPV6_ADDR_MAPPED) {
198
		u32 exthdrlen = icsk->icsk_ext_hdr_len;
L
Linus Torvalds 已提交
199 200 201 202 203 204 205 206 207 208 209
		struct sockaddr_in sin;

		SOCK_DEBUG(sk, "connect: ipv4 mapped\n");

		if (__ipv6_only_sock(sk))
			return -ENETUNREACH;

		sin.sin_family = AF_INET;
		sin.sin_port = usin->sin6_port;
		sin.sin_addr.s_addr = usin->sin6_addr.s6_addr32[3];

210
		icsk->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
211
		sk->sk_backlog_rcv = tcp_v4_do_rcv;
212 213 214
#ifdef CONFIG_TCP_MD5SIG
		tp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif
L
Linus Torvalds 已提交
215 216 217 218

		err = tcp_v4_connect(sk, (struct sockaddr *)&sin, sizeof(sin));

		if (err) {
219 220
			icsk->icsk_ext_hdr_len = exthdrlen;
			icsk->icsk_af_ops = &ipv6_specific;
L
Linus Torvalds 已提交
221
			sk->sk_backlog_rcv = tcp_v6_do_rcv;
222 223 224
#ifdef CONFIG_TCP_MD5SIG
			tp->af_specific = &tcp_sock_ipv6_specific;
#endif
L
Linus Torvalds 已提交
225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253
			goto failure;
		} else {
			ipv6_addr_set(&np->saddr, 0, 0, htonl(0x0000FFFF),
				      inet->saddr);
			ipv6_addr_set(&np->rcv_saddr, 0, 0, htonl(0x0000FFFF),
				      inet->rcv_saddr);
		}

		return err;
	}

	if (!ipv6_addr_any(&np->rcv_saddr))
		saddr = &np->rcv_saddr;

	fl.proto = IPPROTO_TCP;
	ipv6_addr_copy(&fl.fl6_dst, &np->daddr);
	ipv6_addr_copy(&fl.fl6_src,
		       (saddr ? saddr : &np->saddr));
	fl.oif = sk->sk_bound_dev_if;
	fl.fl_ip_dport = usin->sin6_port;
	fl.fl_ip_sport = inet->sport;

	if (np->opt && np->opt->srcrt) {
		struct rt0_hdr *rt0 = (struct rt0_hdr *)np->opt->srcrt;
		ipv6_addr_copy(&final, &fl.fl6_dst);
		ipv6_addr_copy(&fl.fl6_dst, rt0->addr);
		final_p = &final;
	}

V
Venkat Yekkirala 已提交
254 255
	security_sk_classify_flow(sk, &fl);

L
Linus Torvalds 已提交
256 257 258 259 260 261
	err = ip6_dst_lookup(sk, &dst, &fl);
	if (err)
		goto failure;
	if (final_p)
		ipv6_addr_copy(&fl.fl6_dst, final_p);

262
	if ((err = __xfrm_lookup(&dst, &fl, sk, XFRM_LOOKUP_WAIT)) < 0) {
263 264 265 266 267
		if (err == -EREMOTE)
			err = ip6_dst_blackhole(sk, &dst, &fl);
		if (err < 0)
			goto failure;
	}
L
Linus Torvalds 已提交
268 269 270 271 272 273 274 275 276 277

	if (saddr == NULL) {
		saddr = &fl.fl6_src;
		ipv6_addr_copy(&np->rcv_saddr, saddr);
	}

	/* set the source address */
	ipv6_addr_copy(&np->saddr, saddr);
	inet->rcv_saddr = LOOPBACK4_IPV6;

H
Herbert Xu 已提交
278
	sk->sk_gso_type = SKB_GSO_TCPV6;
279
	__ip6_dst_store(sk, dst, NULL, NULL);
L
Linus Torvalds 已提交
280

281
	icsk->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
282
	if (np->opt)
283 284
		icsk->icsk_ext_hdr_len = (np->opt->opt_flen +
					  np->opt->opt_nflen);
L
Linus Torvalds 已提交
285 286 287 288 289 290

	tp->rx_opt.mss_clamp = IPV6_MIN_MTU - sizeof(struct tcphdr) - sizeof(struct ipv6hdr);

	inet->dport = usin->sin6_port;

	tcp_set_state(sk, TCP_SYN_SENT);
291
	err = inet6_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
	if (err)
		goto late_failure;

	if (!tp->write_seq)
		tp->write_seq = secure_tcpv6_sequence_number(np->saddr.s6_addr32,
							     np->daddr.s6_addr32,
							     inet->sport,
							     inet->dport);

	err = tcp_connect(sk);
	if (err)
		goto late_failure;

	return 0;

late_failure:
	tcp_set_state(sk, TCP_CLOSE);
	__sk_dst_reset(sk);
failure:
	inet->dport = 0;
	sk->sk_route_caps = 0;
	return err;
}

static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
317
		int type, int code, int offset, __be32 info)
L
Linus Torvalds 已提交
318 319
{
	struct ipv6hdr *hdr = (struct ipv6hdr*)skb->data;
320
	const struct tcphdr *th = (struct tcphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
321 322 323
	struct ipv6_pinfo *np;
	struct sock *sk;
	int err;
324
	struct tcp_sock *tp;
L
Linus Torvalds 已提交
325
	__u32 seq;
326
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
327

328
	sk = inet6_lookup(net, &tcp_hashinfo, &hdr->daddr,
329
			th->dest, &hdr->saddr, th->source, skb->dev->ifindex);
L
Linus Torvalds 已提交
330 331 332 333 334 335 336

	if (sk == NULL) {
		ICMP6_INC_STATS_BH(__in6_dev_get(skb->dev), ICMP6_MIB_INERRORS);
		return;
	}

	if (sk->sk_state == TCP_TIME_WAIT) {
337
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
338 339 340 341 342
		return;
	}

	bh_lock_sock(sk);
	if (sock_owned_by_user(sk))
343
		NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS);
L
Linus Torvalds 已提交
344 345 346 347 348

	if (sk->sk_state == TCP_CLOSE)
		goto out;

	tp = tcp_sk(sk);
349
	seq = ntohl(th->seq);
L
Linus Torvalds 已提交
350 351
	if (sk->sk_state != TCP_LISTEN &&
	    !between(seq, tp->snd_una, tp->snd_nxt)) {
352
		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
		goto out;
	}

	np = inet6_sk(sk);

	if (type == ICMPV6_PKT_TOOBIG) {
		struct dst_entry *dst = NULL;

		if (sock_owned_by_user(sk))
			goto out;
		if ((1 << sk->sk_state) & (TCPF_LISTEN | TCPF_CLOSE))
			goto out;

		/* icmp should have updated the destination cache entry */
		dst = __sk_dst_check(sk, np->dst_cookie);

		if (dst == NULL) {
			struct inet_sock *inet = inet_sk(sk);
			struct flowi fl;

			/* BUGGG_FUTURE: Again, it is not clear how
			   to handle rthdr case. Ignore this complexity
			   for now.
			 */
			memset(&fl, 0, sizeof(fl));
			fl.proto = IPPROTO_TCP;
			ipv6_addr_copy(&fl.fl6_dst, &np->daddr);
			ipv6_addr_copy(&fl.fl6_src, &np->saddr);
			fl.oif = sk->sk_bound_dev_if;
			fl.fl_ip_dport = inet->dport;
			fl.fl_ip_sport = inet->sport;
V
Venkat Yekkirala 已提交
384
			security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
385 386 387 388 389 390 391 392 393 394 395 396 397 398

			if ((err = ip6_dst_lookup(sk, &dst, &fl))) {
				sk->sk_err_soft = -err;
				goto out;
			}

			if ((err = xfrm_lookup(&dst, &fl, sk, 0)) < 0) {
				sk->sk_err_soft = -err;
				goto out;
			}

		} else
			dst_hold(dst);

399
		if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
L
Linus Torvalds 已提交
400 401 402 403 404 405 406 407 408
			tcp_sync_mss(sk, dst_mtu(dst));
			tcp_simple_retransmit(sk);
		} /* else let the usual retransmit timer handle it */
		dst_release(dst);
		goto out;
	}

	icmpv6_err_convert(type, code, &err);

409
	/* Might be for an request_sock */
L
Linus Torvalds 已提交
410
	switch (sk->sk_state) {
411
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
412 413 414 415
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

416 417
		req = inet6_csk_search_req(sk, &prev, th->dest, &hdr->daddr,
					   &hdr->saddr, inet6_iif(skb));
L
Linus Torvalds 已提交
418 419 420 421 422 423
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		 * an established socket here.
		 */
424
		WARN_ON(req->sk != NULL);
L
Linus Torvalds 已提交
425

426
		if (seq != tcp_rsk(req)->snt_isn) {
427
			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
428 429 430
			goto out;
		}

431
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
432 433 434 435
		goto out;

	case TCP_SYN_SENT:
	case TCP_SYN_RECV:  /* Cannot happen.
436
			       It can, it SYNs are crossed. --ANK */
L
Linus Torvalds 已提交
437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
		if (!sock_owned_by_user(sk)) {
			sk->sk_err = err;
			sk->sk_error_report(sk);		/* Wake people up to see the error (see connect in sock.c) */

			tcp_done(sk);
		} else
			sk->sk_err_soft = err;
		goto out;
	}

	if (!sock_owned_by_user(sk) && np->recverr) {
		sk->sk_err = err;
		sk->sk_error_report(sk);
	} else
		sk->sk_err_soft = err;

out:
	bh_unlock_sock(sk);
	sock_put(sk);
}


459
static int tcp_v6_send_synack(struct sock *sk, struct request_sock *req)
L
Linus Torvalds 已提交
460
{
461
	struct inet6_request_sock *treq = inet6_rsk(req);
L
Linus Torvalds 已提交
462 463 464 465 466
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct sk_buff * skb;
	struct ipv6_txoptions *opt = NULL;
	struct in6_addr * final_p = NULL, final;
	struct flowi fl;
467
	struct dst_entry *dst;
L
Linus Torvalds 已提交
468 469 470 471
	int err = -1;

	memset(&fl, 0, sizeof(fl));
	fl.proto = IPPROTO_TCP;
472 473
	ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
	ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
474
	fl.fl6_flowlabel = 0;
475 476
	fl.oif = treq->iif;
	fl.fl_ip_dport = inet_rsk(req)->rmt_port;
L
Linus Torvalds 已提交
477
	fl.fl_ip_sport = inet_sk(sk)->sport;
478
	security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
479

480 481 482 483 484 485
	opt = np->opt;
	if (opt && opt->srcrt) {
		struct rt0_hdr *rt0 = (struct rt0_hdr *) opt->srcrt;
		ipv6_addr_copy(&final, &fl.fl6_dst);
		ipv6_addr_copy(&fl.fl6_dst, rt0->addr);
		final_p = &final;
L
Linus Torvalds 已提交
486 487
	}

488 489 490 491 492 493 494 495
	err = ip6_dst_lookup(sk, &dst, &fl);
	if (err)
		goto done;
	if (final_p)
		ipv6_addr_copy(&fl.fl6_dst, final_p);
	if ((err = xfrm_lookup(&dst, &fl, sk, 0)) < 0)
		goto done;

L
Linus Torvalds 已提交
496 497
	skb = tcp_make_synack(sk, dst, req);
	if (skb) {
498
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
499 500

		th->check = tcp_v6_check(th, skb->len,
501
					 &treq->loc_addr, &treq->rmt_addr,
L
Linus Torvalds 已提交
502 503
					 csum_partial((char *)th, skb->len, skb->csum));

504
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
505
		err = ip6_xmit(sk, skb, &fl, opt, 0);
506
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
507 508 509
	}

done:
510
	if (opt && opt != np->opt)
L
Linus Torvalds 已提交
511
		sock_kfree_s(sk, opt, opt->tot_len);
512
	dst_release(dst);
L
Linus Torvalds 已提交
513 514 515
	return err;
}

516 517 518 519 520 521 522 523 524 525 526 527 528 529
static inline void syn_flood_warning(struct sk_buff *skb)
{
#ifdef CONFIG_SYN_COOKIES
	if (sysctl_tcp_syncookies)
		printk(KERN_INFO
		       "TCPv6: Possible SYN flooding on port %d. "
		       "Sending cookies.\n", ntohs(tcp_hdr(skb)->dest));
	else
#endif
		printk(KERN_INFO
		       "TCPv6: Possible SYN flooding on port %d. "
		       "Dropping request.\n", ntohs(tcp_hdr(skb)->dest));
}

530
static void tcp_v6_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
531
{
532 533
	if (inet6_rsk(req)->pktopts)
		kfree_skb(inet6_rsk(req)->pktopts);
L
Linus Torvalds 已提交
534 535
}

536 537 538 539 540 541 542 543 544 545 546 547 548
#ifdef CONFIG_TCP_MD5SIG
static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk,
						   struct in6_addr *addr)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int i;

	BUG_ON(tp == NULL);

	if (!tp->md5sig_info || !tp->md5sig_info->entries6)
		return NULL;

	for (i = 0; i < tp->md5sig_info->entries6; i++) {
549
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, addr))
550
			return &tp->md5sig_info->keys6[i].base;
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
	}
	return NULL;
}

static struct tcp_md5sig_key *tcp_v6_md5_lookup(struct sock *sk,
						struct sock *addr_sk)
{
	return tcp_v6_md5_do_lookup(sk, &inet6_sk(addr_sk)->daddr);
}

static struct tcp_md5sig_key *tcp_v6_reqsk_md5_lookup(struct sock *sk,
						      struct request_sock *req)
{
	return tcp_v6_md5_do_lookup(sk, &inet6_rsk(req)->rmt_addr);
}

static int tcp_v6_md5_do_add(struct sock *sk, struct in6_addr *peer,
			     char *newkey, u8 newkeylen)
{
	/* Add key to the list */
571
	struct tcp_md5sig_key *key;
572 573 574
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp6_md5sig_key *keys;

575
	key = tcp_v6_md5_do_lookup(sk, peer);
576 577
	if (key) {
		/* modify existing entry - just update that one */
578 579 580
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
581 582 583 584 585 586 587 588
	} else {
		/* reallocate new list if current one is full. */
		if (!tp->md5sig_info) {
			tp->md5sig_info = kzalloc(sizeof(*tp->md5sig_info), GFP_ATOMIC);
			if (!tp->md5sig_info) {
				kfree(newkey);
				return -ENOMEM;
			}
589
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
590
		}
591 592 593 594
		if (tcp_alloc_md5sig_pool() == NULL) {
			kfree(newkey);
			return -ENOMEM;
		}
595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616
		if (tp->md5sig_info->alloced6 == tp->md5sig_info->entries6) {
			keys = kmalloc((sizeof (tp->md5sig_info->keys6[0]) *
				       (tp->md5sig_info->entries6 + 1)), GFP_ATOMIC);

			if (!keys) {
				tcp_free_md5sig_pool();
				kfree(newkey);
				return -ENOMEM;
			}

			if (tp->md5sig_info->entries6)
				memmove(keys, tp->md5sig_info->keys6,
					(sizeof (tp->md5sig_info->keys6[0]) *
					 tp->md5sig_info->entries6));

			kfree(tp->md5sig_info->keys6);
			tp->md5sig_info->keys6 = keys;
			tp->md5sig_info->alloced6++;
		}

		ipv6_addr_copy(&tp->md5sig_info->keys6[tp->md5sig_info->entries6].addr,
			       peer);
617 618
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.key = newkey;
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.keylen = newkeylen;
619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637

		tp->md5sig_info->entries6++;
	}
	return 0;
}

static int tcp_v6_md5_add_func(struct sock *sk, struct sock *addr_sk,
			       u8 *newkey, __u8 newkeylen)
{
	return tcp_v6_md5_do_add(sk, &inet6_sk(addr_sk)->daddr,
				 newkey, newkeylen);
}

static int tcp_v6_md5_do_del(struct sock *sk, struct in6_addr *peer)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int i;

	for (i = 0; i < tp->md5sig_info->entries6; i++) {
638
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, peer)) {
639
			/* Free the key */
640
			kfree(tp->md5sig_info->keys6[i].base.key);
641 642 643 644 645
			tp->md5sig_info->entries6--;

			if (tp->md5sig_info->entries6 == 0) {
				kfree(tp->md5sig_info->keys6);
				tp->md5sig_info->keys6 = NULL;
646
				tp->md5sig_info->alloced6 = 0;
647 648 649 650 651 652 653 654
			} else {
				/* shrink the database */
				if (tp->md5sig_info->entries6 != i)
					memmove(&tp->md5sig_info->keys6[i],
						&tp->md5sig_info->keys6[i+1],
						(tp->md5sig_info->entries6 - i)
						* sizeof (tp->md5sig_info->keys6[0]));
			}
655 656
			tcp_free_md5sig_pool();
			return 0;
657 658 659 660 661 662 663 664 665 666 667 668
		}
	}
	return -ENOENT;
}

static void tcp_v6_clear_md5_list (struct sock *sk)
{
	struct tcp_sock *tp = tcp_sk(sk);
	int i;

	if (tp->md5sig_info->entries6) {
		for (i = 0; i < tp->md5sig_info->entries6; i++)
669
			kfree(tp->md5sig_info->keys6[i].base.key);
670 671 672 673 674 675 676 677 678 679
		tp->md5sig_info->entries6 = 0;
		tcp_free_md5sig_pool();
	}

	kfree(tp->md5sig_info->keys6);
	tp->md5sig_info->keys6 = NULL;
	tp->md5sig_info->alloced6 = 0;

	if (tp->md5sig_info->entries4) {
		for (i = 0; i < tp->md5sig_info->entries4; i++)
680
			kfree(tp->md5sig_info->keys4[i].base.key);
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708
		tp->md5sig_info->entries4 = 0;
		tcp_free_md5sig_pool();
	}

	kfree(tp->md5sig_info->keys4);
	tp->md5sig_info->keys4 = NULL;
	tp->md5sig_info->alloced4 = 0;
}

static int tcp_v6_parse_md5_keys (struct sock *sk, char __user *optval,
				  int optlen)
{
	struct tcp_md5sig cmd;
	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)&cmd.tcpm_addr;
	u8 *newkey;

	if (optlen < sizeof(cmd))
		return -EINVAL;

	if (copy_from_user(&cmd, optval, sizeof(cmd)))
		return -EFAULT;

	if (sin6->sin6_family != AF_INET6)
		return -EINVAL;

	if (!cmd.tcpm_keylen) {
		if (!tcp_sk(sk)->md5sig_info)
			return -ENOENT;
B
Brian Haley 已提交
709
		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725
			return tcp_v4_md5_do_del(sk, sin6->sin6_addr.s6_addr32[3]);
		return tcp_v6_md5_do_del(sk, &sin6->sin6_addr);
	}

	if (cmd.tcpm_keylen > TCP_MD5SIG_MAXKEYLEN)
		return -EINVAL;

	if (!tcp_sk(sk)->md5sig_info) {
		struct tcp_sock *tp = tcp_sk(sk);
		struct tcp_md5sig_info *p;

		p = kzalloc(sizeof(struct tcp_md5sig_info), GFP_KERNEL);
		if (!p)
			return -ENOMEM;

		tp->md5sig_info = p;
726
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
727 728
	}

A
Arnaldo Carvalho de Melo 已提交
729
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
730 731
	if (!newkey)
		return -ENOMEM;
B
Brian Haley 已提交
732
	if (ipv6_addr_v4mapped(&sin6->sin6_addr)) {
733 734 735 736 737 738
		return tcp_v4_md5_do_add(sk, sin6->sin6_addr.s6_addr32[3],
					 newkey, cmd.tcpm_keylen);
	}
	return tcp_v6_md5_do_add(sk, &sin6->sin6_addr, newkey, cmd.tcpm_keylen);
}

739 740 741
static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp,
					struct in6_addr *daddr,
					struct in6_addr *saddr, int nbytes)
742 743
{
	struct tcp6_pseudohdr *bp;
744
	struct scatterlist sg;
745

746 747 748 749
	bp = &hp->md5_blk.ip6;
	/* 1. TCP pseudo-header (RFC2460) */
	ipv6_addr_copy(&bp->saddr, saddr);
	ipv6_addr_copy(&bp->daddr, daddr);
750 751
	bp->protocol = cpu_to_be32(IPPROTO_TCP);
	bp->len = cpu_to_be16(nbytes);
752

753 754 755
	sg_init_one(&sg, bp, sizeof(*bp));
	return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp));
}
756

757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
static int tcp_v6_md5_hash_hdr(char *md5_hash, struct tcp_md5sig_key *key,
			       struct in6_addr *daddr, struct in6_addr *saddr,
			       struct tcphdr *th)
{
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;
	desc = &hp->md5_desc;

	if (crypto_hash_init(desc))
		goto clear_hash;
	if (tcp_v6_md5_hash_pseudoheader(hp, daddr, saddr, th->doff << 2))
		goto clear_hash;
	if (tcp_md5_hash_header(hp, th))
		goto clear_hash;
	if (tcp_md5_hash_key(hp, key))
		goto clear_hash;
	if (crypto_hash_final(desc, md5_hash))
778 779 780 781
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
782

783 784 785 786
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
787
	return 1;
788 789
}

790 791 792
static int tcp_v6_md5_hash_skb(char *md5_hash, struct tcp_md5sig_key *key,
			       struct sock *sk, struct request_sock *req,
			       struct sk_buff *skb)
793 794
{
	struct in6_addr *saddr, *daddr;
795 796 797
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;
	struct tcphdr *th = tcp_hdr(skb);
798 799 800 801

	if (sk) {
		saddr = &inet6_sk(sk)->saddr;
		daddr = &inet6_sk(sk)->daddr;
802
	} else if (req) {
803 804
		saddr = &inet6_rsk(req)->loc_addr;
		daddr = &inet6_rsk(req)->rmt_addr;
805 806 807 808
	} else {
		struct ipv6hdr *ip6h = ipv6_hdr(skb);
		saddr = &ip6h->saddr;
		daddr = &ip6h->daddr;
809
	}
810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;
	desc = &hp->md5_desc;

	if (crypto_hash_init(desc))
		goto clear_hash;

	if (tcp_v6_md5_hash_pseudoheader(hp, daddr, saddr, skb->len))
		goto clear_hash;
	if (tcp_md5_hash_header(hp, th))
		goto clear_hash;
	if (tcp_md5_hash_skb_data(hp, skb, th->doff << 2))
		goto clear_hash;
	if (tcp_md5_hash_key(hp, key))
		goto clear_hash;
	if (crypto_hash_final(desc, md5_hash))
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;

clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
	return 1;
838 839 840 841 842 843
}

static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb)
{
	__u8 *hash_location = NULL;
	struct tcp_md5sig_key *hash_expected;
844
	struct ipv6hdr *ip6h = ipv6_hdr(skb);
845
	struct tcphdr *th = tcp_hdr(skb);
846 847 848 849
	int genhash;
	u8 newhash[16];

	hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr);
850
	hash_location = tcp_parse_md5sig_option(th);
851

852 853 854 855 856 857
	/* We've parsed the options - do we have a hash? */
	if (!hash_expected && !hash_location)
		return 0;

	if (hash_expected && !hash_location) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND);
858 859 860
		return 1;
	}

861 862
	if (!hash_expected && hash_location) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
863 864 865 866
		return 1;
	}

	/* check the signature */
867 868 869 870
	genhash = tcp_v6_md5_hash_skb(newhash,
				      hash_expected,
				      NULL, NULL, skb);

871 872 873 874 875 876 877 878 879 880 881 882 883 884 885
	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
			printk(KERN_INFO "MD5 Hash %s for "
			       "(" NIP6_FMT ", %u)->"
			       "(" NIP6_FMT ", %u)\n",
			       genhash ? "failed" : "mismatch",
			       NIP6(ip6h->saddr), ntohs(th->source),
			       NIP6(ip6h->daddr), ntohs(th->dest));
		}
		return 1;
	}
	return 0;
}
#endif

886
struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
887
	.family		=	AF_INET6,
888
	.obj_size	=	sizeof(struct tcp6_request_sock),
L
Linus Torvalds 已提交
889
	.rtx_syn_ack	=	tcp_v6_send_synack,
890 891
	.send_ack	=	tcp_v6_reqsk_send_ack,
	.destructor	=	tcp_v6_reqsk_destructor,
L
Linus Torvalds 已提交
892 893 894
	.send_reset	=	tcp_v6_send_reset
};

895
#ifdef CONFIG_TCP_MD5SIG
896
static struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
897 898
	.md5_lookup	=	tcp_v6_reqsk_md5_lookup,
};
899
#endif
900

901 902 903
static struct timewait_sock_ops tcp6_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp6_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
904
	.twsk_destructor= tcp_twsk_destructor,
905 906
};

907
static void tcp_v6_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
908 909
{
	struct ipv6_pinfo *np = inet6_sk(sk);
910
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
911

912
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
L
Linus Torvalds 已提交
913
		th->check = ~csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,  0);
914
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
915
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
916
	} else {
917 918
		th->check = csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,
					    csum_partial((char *)th, th->doff<<2,
L
Linus Torvalds 已提交
919 920 921 922
							 skb->csum));
	}
}

923 924 925 926 927 928 929 930
static int tcp_v6_gso_send_check(struct sk_buff *skb)
{
	struct ipv6hdr *ipv6h;
	struct tcphdr *th;

	if (!pskb_may_pull(skb, sizeof(*th)))
		return -EINVAL;

931
	ipv6h = ipv6_hdr(skb);
932
	th = tcp_hdr(skb);
933 934 935 936

	th->check = 0;
	th->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr, skb->len,
				     IPPROTO_TCP, 0);
937
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
938
	skb->csum_offset = offsetof(struct tcphdr, check);
939
	skb->ip_summed = CHECKSUM_PARTIAL;
940 941
	return 0;
}
L
Linus Torvalds 已提交
942

943
static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
944
{
945
	struct tcphdr *th = tcp_hdr(skb), *t1;
L
Linus Torvalds 已提交
946 947
	struct sk_buff *buff;
	struct flowi fl;
948
	struct net *net = dev_net(skb->dst->dev);
949
	struct sock *ctl_sk = net->ipv6.tcp_sk;
950
	unsigned int tot_len = sizeof(*th);
951 952 953
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
954 955 956 957 958

	if (th->rst)
		return;

	if (!ipv6_unicast_destination(skb))
959
		return;
L
Linus Torvalds 已提交
960

961 962
#ifdef CONFIG_TCP_MD5SIG
	if (sk)
963
		key = tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr);
964 965 966 967 968 969 970
	else
		key = NULL;

	if (key)
		tot_len += TCPOLEN_MD5SIG_ALIGNED;
#endif

L
Linus Torvalds 已提交
971 972 973 974 975
	/*
	 * We need to grab some memory, and put together an RST,
	 * and then put it into the queue to be sent.
	 */

976
	buff = alloc_skb(MAX_HEADER + sizeof(struct ipv6hdr) + tot_len,
L
Linus Torvalds 已提交
977
			 GFP_ATOMIC);
978 979
	if (buff == NULL)
		return;
L
Linus Torvalds 已提交
980

981
	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
L
Linus Torvalds 已提交
982

983
	t1 = (struct tcphdr *) skb_push(buff, tot_len);
L
Linus Torvalds 已提交
984 985 986 987 988

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
989
	t1->doff = tot_len / 4;
L
Linus Torvalds 已提交
990
	t1->rst = 1;
991

L
Linus Torvalds 已提交
992
	if(th->ack) {
993
		t1->seq = th->ack_seq;
L
Linus Torvalds 已提交
994 995 996 997 998 999
	} else {
		t1->ack = 1;
		t1->ack_seq = htonl(ntohl(th->seq) + th->syn + th->fin
				    + skb->len - (th->doff<<2));
	}

1000 1001
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
1002
		__be32 *opt = (__be32*)(t1 + 1);
1003 1004 1005 1006
		opt[0] = htonl((TCPOPT_NOP << 24) |
			       (TCPOPT_NOP << 16) |
			       (TCPOPT_MD5SIG << 8) |
			       TCPOLEN_MD5SIG);
1007 1008 1009
		tcp_v6_md5_hash_hdr((__u8 *)&opt[1], key,
				    &ipv6_hdr(skb)->daddr,
				    &ipv6_hdr(skb)->saddr, t1);
1010 1011 1012
	}
#endif

L
Linus Torvalds 已提交
1013 1014 1015
	buff->csum = csum_partial((char *)t1, sizeof(*t1), 0);

	memset(&fl, 0, sizeof(fl));
1016 1017
	ipv6_addr_copy(&fl.fl6_dst, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&fl.fl6_src, &ipv6_hdr(skb)->daddr);
L
Linus Torvalds 已提交
1018 1019 1020 1021 1022 1023

	t1->check = csum_ipv6_magic(&fl.fl6_src, &fl.fl6_dst,
				    sizeof(*t1), IPPROTO_TCP,
				    buff->csum);

	fl.proto = IPPROTO_TCP;
1024
	fl.oif = inet6_iif(skb);
L
Linus Torvalds 已提交
1025 1026
	fl.fl_ip_dport = t1->dest;
	fl.fl_ip_sport = t1->source;
V
Venkat Yekkirala 已提交
1027
	security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
1028

1029 1030 1031 1032
	/* Pass a socket to ip6_dst_lookup either it is for RST
	 * Underlying function will use this to retrieve the network
	 * namespace
	 */
1033
	if (!ip6_dst_lookup(ctl_sk, &buff->dst, &fl)) {
L
Linus Torvalds 已提交
1034

A
Arnaldo Carvalho de Melo 已提交
1035
		if (xfrm_lookup(&buff->dst, &fl, NULL, 0) >= 0) {
1036
			ip6_xmit(ctl_sk, buff, &fl, NULL, 0);
1037 1038
			TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
			TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
1039
			return;
A
Arnaldo Carvalho de Melo 已提交
1040
		}
L
Linus Torvalds 已提交
1041 1042 1043 1044 1045
	}

	kfree_skb(buff);
}

1046 1047
static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts,
			    struct tcp_md5sig_key *key)
L
Linus Torvalds 已提交
1048
{
1049
	struct tcphdr *th = tcp_hdr(skb), *t1;
L
Linus Torvalds 已提交
1050 1051
	struct sk_buff *buff;
	struct flowi fl;
1052
	struct net *net = dev_net(skb->dev);
1053
	struct sock *ctl_sk = net->ipv6.tcp_sk;
1054
	unsigned int tot_len = sizeof(struct tcphdr);
A
Al Viro 已提交
1055
	__be32 *topt;
L
Linus Torvalds 已提交
1056 1057

	if (ts)
1058
		tot_len += TCPOLEN_TSTAMP_ALIGNED;
1059 1060 1061 1062
#ifdef CONFIG_TCP_MD5SIG
	if (key)
		tot_len += TCPOLEN_MD5SIG_ALIGNED;
#endif
L
Linus Torvalds 已提交
1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081

	buff = alloc_skb(MAX_HEADER + sizeof(struct ipv6hdr) + tot_len,
			 GFP_ATOMIC);
	if (buff == NULL)
		return;

	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);

	t1 = (struct tcphdr *) skb_push(buff,tot_len);

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
	t1->doff = tot_len/4;
	t1->seq = htonl(seq);
	t1->ack_seq = htonl(ack);
	t1->ack = 1;
	t1->window = htons(win);
1082

A
Al Viro 已提交
1083
	topt = (__be32 *)(t1 + 1);
1084

L
Linus Torvalds 已提交
1085
	if (ts) {
1086 1087 1088 1089
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
		*topt++ = htonl(tcp_time_stamp);
		*topt = htonl(ts);
L
Linus Torvalds 已提交
1090 1091
	}

1092 1093 1094 1095
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG);
1096 1097 1098
		tcp_v6_md5_hash_hdr((__u8 *)topt, key,
				    &ipv6_hdr(skb)->daddr,
				    &ipv6_hdr(skb)->saddr, t1);
1099 1100 1101
	}
#endif

L
Linus Torvalds 已提交
1102 1103 1104
	buff->csum = csum_partial((char *)t1, tot_len, 0);

	memset(&fl, 0, sizeof(fl));
1105 1106
	ipv6_addr_copy(&fl.fl6_dst, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&fl.fl6_src, &ipv6_hdr(skb)->daddr);
L
Linus Torvalds 已提交
1107 1108 1109 1110 1111 1112

	t1->check = csum_ipv6_magic(&fl.fl6_src, &fl.fl6_dst,
				    tot_len, IPPROTO_TCP,
				    buff->csum);

	fl.proto = IPPROTO_TCP;
1113
	fl.oif = inet6_iif(skb);
L
Linus Torvalds 已提交
1114 1115
	fl.fl_ip_dport = t1->dest;
	fl.fl_ip_sport = t1->source;
V
Venkat Yekkirala 已提交
1116
	security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
1117

1118
	if (!ip6_dst_lookup(ctl_sk, &buff->dst, &fl)) {
A
Arnaldo Carvalho de Melo 已提交
1119
		if (xfrm_lookup(&buff->dst, &fl, NULL, 0) >= 0) {
1120
			ip6_xmit(ctl_sk, buff, &fl, NULL, 0);
1121
			TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
L
Linus Torvalds 已提交
1122
			return;
A
Arnaldo Carvalho de Melo 已提交
1123
		}
L
Linus Torvalds 已提交
1124 1125 1126 1127 1128 1129 1130
	}

	kfree_skb(buff);
}

static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
1131
	struct inet_timewait_sock *tw = inet_twsk(sk);
1132
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
1133

1134
	tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
1135
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
1136
			tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw));
L
Linus Torvalds 已提交
1137

1138
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
1139 1140
}

1141
static void tcp_v6_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req)
L
Linus Torvalds 已提交
1142
{
1143 1144
	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent,
			tcp_v6_md5_do_lookup(skb->sk, &ipv6_hdr(skb)->daddr));
L
Linus Torvalds 已提交
1145 1146 1147 1148 1149
}


static struct sock *tcp_v6_hnd_req(struct sock *sk,struct sk_buff *skb)
{
1150
	struct request_sock *req, **prev;
1151
	const struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1152 1153 1154
	struct sock *nsk;

	/* Find possible connection requests. */
1155
	req = inet6_csk_search_req(sk, &prev, th->source,
1156 1157
				   &ipv6_hdr(skb)->saddr,
				   &ipv6_hdr(skb)->daddr, inet6_iif(skb));
L
Linus Torvalds 已提交
1158 1159 1160
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1161
	nsk = __inet6_lookup_established(sock_net(sk), &tcp_hashinfo,
1162 1163
			&ipv6_hdr(skb)->saddr, th->source,
			&ipv6_hdr(skb)->daddr, ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1170
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1171 1172 1173
		return NULL;
	}

1174
#ifdef CONFIG_SYN_COOKIES
L
Linus Torvalds 已提交
1175
	if (!th->rst && !th->syn && th->ack)
1176
		sk = cookie_v6_check(sk, skb);
L
Linus Torvalds 已提交
1177 1178 1179 1180 1181 1182 1183 1184 1185
#endif
	return sk;
}

/* FIXME: this is substantially similar to the ipv4 code.
 * Can some kind of merge be done? -- erics
 */
static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
{
1186
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1187 1188 1189
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_options_received tmp_opt;
	struct tcp_sock *tp = tcp_sk(sk);
1190
	struct request_sock *req = NULL;
L
Linus Torvalds 已提交
1191
	__u32 isn = TCP_SKB_CB(skb)->when;
1192 1193 1194 1195 1196
#ifdef CONFIG_SYN_COOKIES
	int want_cookie = 0;
#else
#define want_cookie 0
#endif
L
Linus Torvalds 已提交
1197 1198 1199 1200 1201

	if (skb->protocol == htons(ETH_P_IP))
		return tcp_v4_conn_request(sk, skb);

	if (!ipv6_unicast_destination(skb))
1202
		goto drop;
L
Linus Torvalds 已提交
1203

1204
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1205
		if (net_ratelimit())
1206 1207 1208 1209 1210 1211
			syn_flood_warning(skb);
#ifdef CONFIG_SYN_COOKIES
		if (sysctl_tcp_syncookies)
			want_cookie = 1;
		else
#endif
1212
		goto drop;
L
Linus Torvalds 已提交
1213 1214
	}

1215
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1216 1217
		goto drop;

1218
	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
L
Linus Torvalds 已提交
1219 1220 1221
	if (req == NULL)
		goto drop;

1222 1223 1224 1225
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops;
#endif

L
Linus Torvalds 已提交
1226 1227 1228 1229 1230 1231
	tcp_clear_options(&tmp_opt);
	tmp_opt.mss_clamp = IPV6_MIN_MTU - sizeof(struct tcphdr) - sizeof(struct ipv6hdr);
	tmp_opt.user_mss = tp->rx_opt.user_mss;

	tcp_parse_options(skb, &tmp_opt, 0);

1232
	if (want_cookie && !tmp_opt.saw_tstamp)
1233 1234
		tcp_clear_options(&tmp_opt);

L
Linus Torvalds 已提交
1235 1236 1237
	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
	tcp_openreq_init(req, &tmp_opt, skb);

1238
	treq = inet6_rsk(req);
1239 1240
	ipv6_addr_copy(&treq->rmt_addr, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&treq->loc_addr, &ipv6_hdr(skb)->daddr);
1241 1242 1243 1244 1245
	if (!want_cookie)
		TCP_ECN_create_request(req, tcp_hdr(skb));

	if (want_cookie) {
		isn = cookie_v6_init_sequence(sk, skb, &req->mss);
1246
		req->cookie_ts = tmp_opt.tstamp_ok;
1247 1248 1249 1250 1251 1252 1253 1254
	} else if (!isn) {
		if (ipv6_opt_accepted(sk, skb) ||
		    np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo ||
		    np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim) {
			atomic_inc(&skb->users);
			treq->pktopts = skb;
		}
		treq->iif = sk->sk_bound_dev_if;
L
Linus Torvalds 已提交
1255

1256 1257 1258 1259
		/* So that link locals have meaning */
		if (!sk->sk_bound_dev_if &&
		    ipv6_addr_type(&treq->rmt_addr) & IPV6_ADDR_LINKLOCAL)
			treq->iif = inet6_iif(skb);
L
Linus Torvalds 已提交
1260

1261
		isn = tcp_v6_init_sequence(skb);
1262
	}
L
Linus Torvalds 已提交
1263

1264
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1265

1266 1267
	security_inet_conn_request(sk, skb, req);

1268
	if (tcp_v6_send_synack(sk, req))
L
Linus Torvalds 已提交
1269 1270
		goto drop;

1271 1272 1273 1274
	if (!want_cookie) {
		inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
		return 0;
	}
L
Linus Torvalds 已提交
1275 1276 1277

drop:
	if (req)
1278
		reqsk_free(req);
L
Linus Torvalds 已提交
1279 1280 1281 1282 1283

	return 0; /* don't send reset */
}

static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
1284
					  struct request_sock *req,
L
Linus Torvalds 已提交
1285 1286
					  struct dst_entry *dst)
{
1287
	struct inet6_request_sock *treq = inet6_rsk(req);
L
Linus Torvalds 已提交
1288 1289 1290 1291 1292 1293
	struct ipv6_pinfo *newnp, *np = inet6_sk(sk);
	struct tcp6_sock *newtcp6sk;
	struct inet_sock *newinet;
	struct tcp_sock *newtp;
	struct sock *newsk;
	struct ipv6_txoptions *opt;
1294 1295 1296
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1297 1298 1299 1300 1301 1302 1303 1304

	if (skb->protocol == htons(ETH_P_IP)) {
		/*
		 *	v6 mapped
		 */

		newsk = tcp_v4_syn_recv_sock(sk, skb, req, dst);

1305
		if (newsk == NULL)
L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324
			return NULL;

		newtcp6sk = (struct tcp6_sock *)newsk;
		inet_sk(newsk)->pinet6 = &newtcp6sk->inet6;

		newinet = inet_sk(newsk);
		newnp = inet6_sk(newsk);
		newtp = tcp_sk(newsk);

		memcpy(newnp, np, sizeof(struct ipv6_pinfo));

		ipv6_addr_set(&newnp->daddr, 0, 0, htonl(0x0000FFFF),
			      newinet->daddr);

		ipv6_addr_set(&newnp->saddr, 0, 0, htonl(0x0000FFFF),
			      newinet->saddr);

		ipv6_addr_copy(&newnp->rcv_saddr, &newnp->saddr);

1325
		inet_csk(newsk)->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
1326
		newsk->sk_backlog_rcv = tcp_v4_do_rcv;
1327 1328 1329 1330
#ifdef CONFIG_TCP_MD5SIG
		newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif

L
Linus Torvalds 已提交
1331 1332
		newnp->pktoptions  = NULL;
		newnp->opt	   = NULL;
1333
		newnp->mcast_oif   = inet6_iif(skb);
1334
		newnp->mcast_hops  = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1335

1336 1337 1338 1339
		/*
		 * No need to charge this sock to the relevant IPv6 refcnt debug socks count
		 * here, tcp_create_openreq_child now does this for us, see the comment in
		 * that function for the gory details. -acme
L
Linus Torvalds 已提交
1340 1341 1342
		 */

		/* It is tricky place. Until this moment IPv4 tcp
1343
		   worked with IPv6 icsk.icsk_af_ops.
L
Linus Torvalds 已提交
1344 1345
		   Sync it now.
		 */
1346
		tcp_sync_mss(newsk, inet_csk(newsk)->icsk_pmtu_cookie);
L
Linus Torvalds 已提交
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361

		return newsk;
	}

	opt = np->opt;

	if (sk_acceptq_is_full(sk))
		goto out_overflow;

	if (dst == NULL) {
		struct in6_addr *final_p = NULL, final;
		struct flowi fl;

		memset(&fl, 0, sizeof(fl));
		fl.proto = IPPROTO_TCP;
1362
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
1363 1364 1365 1366 1367 1368
		if (opt && opt->srcrt) {
			struct rt0_hdr *rt0 = (struct rt0_hdr *) opt->srcrt;
			ipv6_addr_copy(&final, &fl.fl6_dst);
			ipv6_addr_copy(&fl.fl6_dst, rt0->addr);
			final_p = &final;
		}
1369
		ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
1370
		fl.oif = sk->sk_bound_dev_if;
1371
		fl.fl_ip_dport = inet_rsk(req)->rmt_port;
L
Linus Torvalds 已提交
1372
		fl.fl_ip_sport = inet_sk(sk)->sport;
1373
		security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
1374 1375 1376 1377 1378 1379 1380 1381 1382

		if (ip6_dst_lookup(sk, &dst, &fl))
			goto out;

		if (final_p)
			ipv6_addr_copy(&fl.fl6_dst, final_p);

		if ((xfrm_lookup(&dst, &fl, sk, 0)) < 0)
			goto out;
1383
	}
L
Linus Torvalds 已提交
1384 1385 1386 1387 1388

	newsk = tcp_create_openreq_child(sk, req, skb);
	if (newsk == NULL)
		goto out;

1389 1390 1391 1392 1393
	/*
	 * No need to charge this sock to the relevant IPv6 refcnt debug socks
	 * count here, tcp_create_openreq_child now does this for us, see the
	 * comment in that function for the gory details. -acme
	 */
L
Linus Torvalds 已提交
1394

1395
	newsk->sk_gso_type = SKB_GSO_TCPV6;
1396
	__ip6_dst_store(newsk, dst, NULL, NULL);
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406

	newtcp6sk = (struct tcp6_sock *)newsk;
	inet_sk(newsk)->pinet6 = &newtcp6sk->inet6;

	newtp = tcp_sk(newsk);
	newinet = inet_sk(newsk);
	newnp = inet6_sk(newsk);

	memcpy(newnp, np, sizeof(struct ipv6_pinfo));

1407 1408 1409 1410
	ipv6_addr_copy(&newnp->daddr, &treq->rmt_addr);
	ipv6_addr_copy(&newnp->saddr, &treq->loc_addr);
	ipv6_addr_copy(&newnp->rcv_saddr, &treq->loc_addr);
	newsk->sk_bound_dev_if = treq->iif;
L
Linus Torvalds 已提交
1411

1412
	/* Now IPv6 options...
L
Linus Torvalds 已提交
1413 1414 1415 1416

	   First: no IPv4 options.
	 */
	newinet->opt = NULL;
1417
	newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1418 1419 1420 1421 1422 1423

	/* Clone RX bits */
	newnp->rxopt.all = np->rxopt.all;

	/* Clone pktoptions received with SYN */
	newnp->pktoptions = NULL;
1424 1425 1426 1427
	if (treq->pktopts != NULL) {
		newnp->pktoptions = skb_clone(treq->pktopts, GFP_ATOMIC);
		kfree_skb(treq->pktopts);
		treq->pktopts = NULL;
L
Linus Torvalds 已提交
1428 1429 1430 1431
		if (newnp->pktoptions)
			skb_set_owner_r(newnp->pktoptions, newsk);
	}
	newnp->opt	  = NULL;
1432
	newnp->mcast_oif  = inet6_iif(skb);
1433
	newnp->mcast_hops = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446

	/* Clone native IPv6 options from listening socket (if any)

	   Yes, keeping reference count would be much more clever,
	   but we make one more one thing there: reattach optmem
	   to newsk.
	 */
	if (opt) {
		newnp->opt = ipv6_dup_options(newsk, opt);
		if (opt != np->opt)
			sock_kfree_s(sk, opt, opt->tot_len);
	}

1447
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1448
	if (newnp->opt)
1449 1450
		inet_csk(newsk)->icsk_ext_hdr_len = (newnp->opt->opt_nflen +
						     newnp->opt->opt_flen);
L
Linus Torvalds 已提交
1451

J
John Heffner 已提交
1452
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1453 1454 1455 1456 1457 1458
	tcp_sync_mss(newsk, dst_mtu(dst));
	newtp->advmss = dst_metric(dst, RTAX_ADVMSS);
	tcp_initialize_rcv_mss(newsk);

	newinet->daddr = newinet->saddr = newinet->rcv_saddr = LOOPBACK4_IPV6;

1459 1460 1461 1462 1463 1464 1465 1466
#ifdef CONFIG_TCP_MD5SIG
	/* Copy over the MD5 key from the original socket */
	if ((key = tcp_v6_md5_do_lookup(sk, &newnp->daddr)) != NULL) {
		/* We're using one, so create a matching key
		 * on the newsk structure. If we fail to get
		 * memory, then we end up not copying the key
		 * across. Shucks.
		 */
A
Arnaldo Carvalho de Melo 已提交
1467 1468
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1469 1470 1471 1472 1473
			tcp_v6_md5_do_add(newsk, &inet6_sk(sk)->daddr,
					  newkey, key->keylen);
	}
#endif

1474
	__inet6_hash(newsk);
1475
	__inet_inherit_port(sk, newsk);
L
Linus Torvalds 已提交
1476 1477 1478 1479

	return newsk;

out_overflow:
1480
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
L
Linus Torvalds 已提交
1481
out:
1482
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
L
Linus Torvalds 已提交
1483 1484 1485 1486 1487 1488
	if (opt && opt != np->opt)
		sock_kfree_s(sk, opt, opt->tot_len);
	dst_release(dst);
	return NULL;
}

1489
static __sum16 tcp_v6_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1490
{
1491
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
1492
		if (!tcp_v6_check(tcp_hdr(skb), skb->len, &ipv6_hdr(skb)->saddr,
1493
				  &ipv6_hdr(skb)->daddr, skb->csum)) {
1494
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1495
			return 0;
1496
		}
L
Linus Torvalds 已提交
1497
	}
1498

1499
	skb->csum = ~csum_unfold(tcp_v6_check(tcp_hdr(skb), skb->len,
1500 1501
					      &ipv6_hdr(skb)->saddr,
					      &ipv6_hdr(skb)->daddr, 0));
1502

L
Linus Torvalds 已提交
1503
	if (skb->len <= 76) {
1504
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533
	}
	return 0;
}

/* The socket must have it's spinlock held when we get
 * here.
 *
 * We have a potential double-lock case here, so even when
 * doing backlog processing we use the BH locking scheme.
 * This is because we cannot sleep with the original spinlock
 * held.
 */
static int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_sock *tp;
	struct sk_buff *opt_skb = NULL;

	/* Imagine: socket is IPv6. IPv4 packet arrives,
	   goes to IPv4 receive handler and backlogged.
	   From backlog it always goes here. Kerboom...
	   Fortunately, tcp_rcv_established and rcv_established
	   handle them correctly, but it is not case with
	   tcp_v6_hnd_req and tcp_v6_send_reset().   --ANK
	 */

	if (skb->protocol == htons(ETH_P_IP))
		return tcp_v4_do_rcv(sk, skb);

1534 1535 1536 1537 1538
#ifdef CONFIG_TCP_MD5SIG
	if (tcp_v6_inbound_md5_hash (sk, skb))
		goto discard;
#endif

1539
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
		goto discard;

	/*
	 *	socket locking is here for SMP purposes as backlog rcv
	 *	is currently called with bh processing disabled.
	 */

	/* Do Stevens' IPV6_PKTOPTIONS.

	   Yes, guys, it is the only place in our code, where we
	   may make it not affecting IPv4.
	   The rest of code is protocol independent,
	   and I do not like idea to uglify IPv4.

	   Actually, all the idea behind IPV6_PKTOPTIONS
	   looks not very well thought. For now we latch
	   options, received in the last packet, enqueued
	   by tcp. Feel free to propose better solution.
1558
					       --ANK (980728)
L
Linus Torvalds 已提交
1559 1560 1561 1562 1563 1564
	 */
	if (np->rxopt.all)
		opt_skb = skb_clone(skb, GFP_ATOMIC);

	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1565
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1566 1567 1568 1569 1570 1571 1572
			goto reset;
		TCP_CHECK_TIMER(sk);
		if (opt_skb)
			goto ipv6_pktoptions;
		return 0;
	}

1573
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1574 1575
		goto csum_err;

1576
	if (sk->sk_state == TCP_LISTEN) {
L
Linus Torvalds 已提交
1577 1578 1579 1580 1581 1582 1583 1584 1585
		struct sock *nsk = tcp_v6_hnd_req(sk, skb);
		if (!nsk)
			goto discard;

		/*
		 * Queue it on the new socket if the new socket is active,
		 * otherwise we just shortcircuit this and continue with
		 * the new socket..
		 */
1586
		if(nsk != sk) {
L
Linus Torvalds 已提交
1587 1588 1589 1590 1591 1592 1593 1594 1595
			if (tcp_child_process(sk, nsk, skb))
				goto reset;
			if (opt_skb)
				__kfree_skb(opt_skb);
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1596
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1597 1598 1599 1600 1601 1602 1603
		goto reset;
	TCP_CHECK_TIMER(sk);
	if (opt_skb)
		goto ipv6_pktoptions;
	return 0;

reset:
1604
	tcp_v6_send_reset(sk, skb);
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609 1610
discard:
	if (opt_skb)
		__kfree_skb(opt_skb);
	kfree_skb(skb);
	return 0;
csum_err:
1611
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
	goto discard;


ipv6_pktoptions:
	/* Do you ask, what is it?

	   1. skb was enqueued by tcp.
	   2. skb is added to tail of read queue, rather than out of order.
	   3. socket is not in passive state.
	   4. Finally, it really contains options, which user wants to receive.
	 */
	tp = tcp_sk(sk);
	if (TCP_SKB_CB(opt_skb)->end_seq == tp->rcv_nxt &&
	    !((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN))) {
1626
		if (np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo)
1627
			np->mcast_oif = inet6_iif(opt_skb);
1628
		if (np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim)
1629
			np->mcast_hops = ipv6_hdr(opt_skb)->hop_limit;
L
Linus Torvalds 已提交
1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643
		if (ipv6_opt_accepted(sk, opt_skb)) {
			skb_set_owner_r(opt_skb, sk);
			opt_skb = xchg(&np->pktoptions, opt_skb);
		} else {
			__kfree_skb(opt_skb);
			opt_skb = xchg(&np->pktoptions, NULL);
		}
	}

	if (opt_skb)
		kfree_skb(opt_skb);
	return 0;
}

1644
static int tcp_v6_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
1645
{
1646
	struct tcphdr *th;
L
Linus Torvalds 已提交
1647 1648
	struct sock *sk;
	int ret;
1649
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1650 1651 1652 1653 1654 1655 1656

	if (skb->pkt_type != PACKET_HOST)
		goto discard_it;

	/*
	 *	Count it even if it's bad.
	 */
1657
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1658 1659 1660 1661

	if (!pskb_may_pull(skb, sizeof(struct tcphdr)))
		goto discard_it;

1662
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1663 1664 1665 1666 1667 1668

	if (th->doff < sizeof(struct tcphdr)/4)
		goto bad_packet;
	if (!pskb_may_pull(skb, th->doff*4))
		goto discard_it;

1669
	if (!skb_csum_unnecessary(skb) && tcp_v6_checksum_init(skb))
L
Linus Torvalds 已提交
1670 1671
		goto bad_packet;

1672
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1673 1674 1675 1676 1677
	TCP_SKB_CB(skb)->seq = ntohl(th->seq);
	TCP_SKB_CB(skb)->end_seq = (TCP_SKB_CB(skb)->seq + th->syn + th->fin +
				    skb->len - th->doff*4);
	TCP_SKB_CB(skb)->ack_seq = ntohl(th->ack_seq);
	TCP_SKB_CB(skb)->when = 0;
1678
	TCP_SKB_CB(skb)->flags = ipv6_get_dsfield(ipv6_hdr(skb));
L
Linus Torvalds 已提交
1679 1680
	TCP_SKB_CB(skb)->sacked = 0;

1681
	sk = __inet6_lookup(net, &tcp_hashinfo,
1682 1683 1684
			&ipv6_hdr(skb)->saddr, th->source,
			&ipv6_hdr(skb)->daddr, ntohs(th->dest),
			inet6_iif(skb));
L
Linus Torvalds 已提交
1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695

	if (!sk)
		goto no_tcp_socket;

process:
	if (sk->sk_state == TCP_TIME_WAIT)
		goto do_time_wait;

	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto discard_and_relse;

1696
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1697 1698 1699 1700
		goto discard_and_relse;

	skb->dev = NULL;

1701
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1702 1703
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1704
#ifdef CONFIG_NET_DMA
1705
		struct tcp_sock *tp = tcp_sk(sk);
1706 1707
		if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
			tp->ucopy.dma_chan = get_softnet_dma();
1708 1709 1710
		if (tp->ucopy.dma_chan)
			ret = tcp_v6_do_rcv(sk, skb);
		else
C
Chris Leech 已提交
1711 1712 1713 1714 1715
#endif
		{
			if (!tcp_prequeue(sk, skb))
				ret = tcp_v6_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728
	} else
		sk_add_backlog(sk, skb);
	bh_unlock_sock(sk);

	sock_put(sk);
	return ret ? -1 : 0;

no_tcp_socket:
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
		goto discard_it;

	if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) {
bad_packet:
1729
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1730
	} else {
1731
		tcp_v6_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748
	}

discard_it:

	/*
	 *	Discard frame
	 */

	kfree_skb(skb);
	return 0;

discard_and_relse:
	sock_put(sk);
	goto discard_it;

do_time_wait:
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
1749
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1750 1751 1752 1753
		goto discard_it;
	}

	if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) {
1754
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1755
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1756 1757 1758
		goto discard_it;
	}

1759
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1760 1761 1762 1763
	case TCP_TW_SYN:
	{
		struct sock *sk2;

1764
		sk2 = inet6_lookup_listener(dev_net(skb->dev), &tcp_hashinfo,
1765
					    &ipv6_hdr(skb)->daddr,
1766
					    ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1767
		if (sk2 != NULL) {
1768 1769 1770
			struct inet_timewait_sock *tw = inet_twsk(sk);
			inet_twsk_deschedule(tw, &tcp_death_row);
			inet_twsk_put(tw);
L
Linus Torvalds 已提交
1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
			sk = sk2;
			goto process;
		}
		/* Fall through to ACK */
	}
	case TCP_TW_ACK:
		tcp_v6_timewait_ack(sk, skb);
		break;
	case TCP_TW_RST:
		goto no_tcp_socket;
	case TCP_TW_SUCCESS:;
	}
	goto discard_it;
}

static int tcp_v6_remember_stamp(struct sock *sk)
{
	/* Alas, not yet... */
	return 0;
}

1792
static struct inet_connection_sock_af_ops ipv6_specific = {
1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
	.queue_xmit	   = inet6_csk_xmit,
	.send_check	   = tcp_v6_send_check,
	.rebuild_header	   = inet6_sk_rebuild_header,
	.conn_request	   = tcp_v6_conn_request,
	.syn_recv_sock	   = tcp_v6_syn_recv_sock,
	.remember_stamp	   = tcp_v6_remember_stamp,
	.net_header_len	   = sizeof(struct ipv6hdr),
	.setsockopt	   = ipv6_setsockopt,
	.getsockopt	   = ipv6_getsockopt,
	.addr2sockaddr	   = inet6_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in6),
1804
	.bind_conflict	   = inet6_csk_bind_conflict,
1805
#ifdef CONFIG_COMPAT
1806 1807
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1808
#endif
L
Linus Torvalds 已提交
1809 1810
};

1811
#ifdef CONFIG_TCP_MD5SIG
1812
static struct tcp_sock_af_ops tcp_sock_ipv6_specific = {
1813
	.md5_lookup	=	tcp_v6_md5_lookup,
1814
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
1815 1816 1817
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1818
#endif
1819

L
Linus Torvalds 已提交
1820 1821 1822 1823
/*
 *	TCP over IPv4 via INET6 API
 */

1824
static struct inet_connection_sock_af_ops ipv6_mapped = {
1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835
	.queue_xmit	   = ip_queue_xmit,
	.send_check	   = tcp_v4_send_check,
	.rebuild_header	   = inet_sk_rebuild_header,
	.conn_request	   = tcp_v6_conn_request,
	.syn_recv_sock	   = tcp_v6_syn_recv_sock,
	.remember_stamp	   = tcp_v4_remember_stamp,
	.net_header_len	   = sizeof(struct iphdr),
	.setsockopt	   = ipv6_setsockopt,
	.getsockopt	   = ipv6_getsockopt,
	.addr2sockaddr	   = inet6_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in6),
1836
	.bind_conflict	   = inet6_csk_bind_conflict,
1837
#ifdef CONFIG_COMPAT
1838 1839
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1840
#endif
L
Linus Torvalds 已提交
1841 1842
};

1843
#ifdef CONFIG_TCP_MD5SIG
1844
static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = {
1845
	.md5_lookup	=	tcp_v4_md5_lookup,
1846
	.calc_md5_hash	=	tcp_v4_md5_hash_skb,
1847 1848 1849
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1850
#endif
1851

L
Linus Torvalds 已提交
1852 1853 1854 1855 1856
/* NOTE: A lot of things set to zero explicitly by call to
 *       sk_alloc() so need not be done here.
 */
static int tcp_v6_init_sock(struct sock *sk)
{
1857
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1858 1859 1860 1861 1862 1863
	struct tcp_sock *tp = tcp_sk(sk);

	skb_queue_head_init(&tp->out_of_order_queue);
	tcp_init_xmit_timers(sk);
	tcp_prequeue_init(tp);

1864
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878
	tp->mdev = TCP_TIMEOUT_INIT;

	/* So many TCP implementations out there (incorrectly) count the
	 * initial SYN frame in their delayed-ACK and congestion control
	 * algorithms that we must have the following bandaid to talk
	 * efficiently to them.  -DaveM
	 */
	tp->snd_cwnd = 2;

	/* See draft-stevens-tcpca-spec-01 for discussion of the
	 * initialization of these values.
	 */
	tp->snd_ssthresh = 0x7fffffff;
	tp->snd_cwnd_clamp = ~0;
1879
	tp->mss_cache = 536;
L
Linus Torvalds 已提交
1880 1881 1882 1883 1884

	tp->reordering = sysctl_tcp_reordering;

	sk->sk_state = TCP_CLOSE;

1885
	icsk->icsk_af_ops = &ipv6_specific;
1886
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
1887
	icsk->icsk_sync_mss = tcp_sync_mss;
L
Linus Torvalds 已提交
1888 1889 1890
	sk->sk_write_space = sk_stream_write_space;
	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);

1891 1892 1893 1894
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv6_specific;
#endif

L
Linus Torvalds 已提交
1895 1896 1897 1898 1899 1900 1901 1902
	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

	atomic_inc(&tcp_sockets_allocated);

	return 0;
}

1903
static void tcp_v6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1904
{
1905 1906 1907 1908 1909
#ifdef CONFIG_TCP_MD5SIG
	/* Clean up the MD5 key list */
	if (tcp_sk(sk)->md5sig_info)
		tcp_v6_clear_md5_list(sk);
#endif
L
Linus Torvalds 已提交
1910
	tcp_v4_destroy_sock(sk);
1911
	inet6_destroy_sock(sk);
L
Linus Torvalds 已提交
1912 1913
}

1914
#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1915
/* Proc filesystem TCPv6 sock list dumping. */
1916
static void get_openreq6(struct seq_file *seq,
1917
			 struct sock *sk, struct request_sock *req, int i, int uid)
L
Linus Torvalds 已提交
1918 1919
{
	int ttd = req->expires - jiffies;
1920 1921
	struct in6_addr *src = &inet6_rsk(req)->loc_addr;
	struct in6_addr *dest = &inet6_rsk(req)->rmt_addr;
L
Linus Torvalds 已提交
1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934

	if (ttd < 0)
		ttd = 0;

	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p\n",
		   i,
		   src->s6_addr32[0], src->s6_addr32[1],
		   src->s6_addr32[2], src->s6_addr32[3],
		   ntohs(inet_sk(sk)->sport),
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3],
1935
		   ntohs(inet_rsk(req)->rmt_port),
L
Linus Torvalds 已提交
1936 1937
		   TCP_SYN_RECV,
		   0,0, /* could print option size, but that is af dependent. */
1938 1939
		   1,   /* timers active (only the expire timer) */
		   jiffies_to_clock_t(ttd),
L
Linus Torvalds 已提交
1940 1941
		   req->retrans,
		   uid,
1942
		   0,  /* non standard timer */
L
Linus Torvalds 已提交
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954
		   0, /* open_requests have no inode */
		   0, req);
}

static void get_tcp6_sock(struct seq_file *seq, struct sock *sp, int i)
{
	struct in6_addr *dest, *src;
	__u16 destp, srcp;
	int timer_active;
	unsigned long timer_expires;
	struct inet_sock *inet = inet_sk(sp);
	struct tcp_sock *tp = tcp_sk(sp);
1955
	const struct inet_connection_sock *icsk = inet_csk(sp);
L
Linus Torvalds 已提交
1956 1957 1958 1959 1960 1961
	struct ipv6_pinfo *np = inet6_sk(sp);

	dest  = &np->daddr;
	src   = &np->rcv_saddr;
	destp = ntohs(inet->dport);
	srcp  = ntohs(inet->sport);
1962 1963

	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
1964
		timer_active	= 1;
1965 1966
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
1967
		timer_active	= 4;
1968
		timer_expires	= icsk->icsk_timeout;
L
Linus Torvalds 已提交
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978
	} else if (timer_pending(&sp->sk_timer)) {
		timer_active	= 2;
		timer_expires	= sp->sk_timer.expires;
	} else {
		timer_active	= 0;
		timer_expires = jiffies;
	}

	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
1979
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %lu %lu %u %u %d\n",
L
Linus Torvalds 已提交
1980 1981 1982 1983 1984
		   i,
		   src->s6_addr32[0], src->s6_addr32[1],
		   src->s6_addr32[2], src->s6_addr32[3], srcp,
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3], destp,
1985
		   sp->sk_state,
1986 1987
		   tp->write_seq-tp->snd_una,
		   (sp->sk_state == TCP_LISTEN) ? sp->sk_ack_backlog : (tp->rcv_nxt - tp->copied_seq),
L
Linus Torvalds 已提交
1988 1989
		   timer_active,
		   jiffies_to_clock_t(timer_expires - jiffies),
1990
		   icsk->icsk_retransmits,
L
Linus Torvalds 已提交
1991
		   sock_i_uid(sp),
1992
		   icsk->icsk_probes_out,
L
Linus Torvalds 已提交
1993 1994
		   sock_i_ino(sp),
		   atomic_read(&sp->sk_refcnt), sp,
1995 1996
		   jiffies_to_clock_t(icsk->icsk_rto),
		   jiffies_to_clock_t(icsk->icsk_ack.ato),
1997
		   (icsk->icsk_ack.quick << 1 ) | icsk->icsk_ack.pingpong,
L
Linus Torvalds 已提交
1998 1999 2000 2001
		   tp->snd_cwnd, tp->snd_ssthresh>=0xFFFF?-1:tp->snd_ssthresh
		   );
}

2002
static void get_timewait6_sock(struct seq_file *seq,
2003
			       struct inet_timewait_sock *tw, int i)
L
Linus Torvalds 已提交
2004 2005 2006
{
	struct in6_addr *dest, *src;
	__u16 destp, srcp;
2007
	struct inet6_timewait_sock *tw6 = inet6_twsk((struct sock *)tw);
L
Linus Torvalds 已提交
2008 2009 2010 2011 2012
	int ttd = tw->tw_ttd - jiffies;

	if (ttd < 0)
		ttd = 0;

2013 2014
	dest = &tw6->tw_v6_daddr;
	src  = &tw6->tw_v6_rcv_saddr;
L
Linus Torvalds 已提交
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064
	destp = ntohs(tw->tw_dport);
	srcp  = ntohs(tw->tw_sport);

	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p\n",
		   i,
		   src->s6_addr32[0], src->s6_addr32[1],
		   src->s6_addr32[2], src->s6_addr32[3], srcp,
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3], destp,
		   tw->tw_substate, 0, 0,
		   3, jiffies_to_clock_t(ttd), 0, 0, 0, 0,
		   atomic_read(&tw->tw_refcnt), tw);
}

static int tcp6_seq_show(struct seq_file *seq, void *v)
{
	struct tcp_iter_state *st;

	if (v == SEQ_START_TOKEN) {
		seq_puts(seq,
			 "  sl  "
			 "local_address                         "
			 "remote_address                        "
			 "st tx_queue rx_queue tr tm->when retrnsmt"
			 "   uid  timeout inode\n");
		goto out;
	}
	st = seq->private;

	switch (st->state) {
	case TCP_SEQ_STATE_LISTENING:
	case TCP_SEQ_STATE_ESTABLISHED:
		get_tcp6_sock(seq, v, st->num);
		break;
	case TCP_SEQ_STATE_OPENREQ:
		get_openreq6(seq, st->syn_wait_sk, v, st->num, st->uid);
		break;
	case TCP_SEQ_STATE_TIME_WAIT:
		get_timewait6_sock(seq, v, st->num);
		break;
	}
out:
	return 0;
}

static struct tcp_seq_afinfo tcp6_seq_afinfo = {
	.name		= "tcp6",
	.family		= AF_INET6,
2065 2066 2067
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2068 2069 2070
	.seq_ops	= {
		.show		= tcp6_seq_show,
	},
L
Linus Torvalds 已提交
2071 2072
};

2073
int tcp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
2074
{
2075
	return tcp_proc_register(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2076 2077
}

2078
void tcp6_proc_exit(struct net *net)
L
Linus Torvalds 已提交
2079
{
2080
	tcp_proc_unregister(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2081 2082 2083 2084 2085 2086 2087 2088 2089
}
#endif

struct proto tcpv6_prot = {
	.name			= "TCPv6",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v6_connect,
	.disconnect		= tcp_disconnect,
2090
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2091 2092 2093 2094 2095 2096 2097 2098 2099
	.ioctl			= tcp_ioctl,
	.init			= tcp_v6_init_sock,
	.destroy		= tcp_v6_destroy_sock,
	.shutdown		= tcp_shutdown,
	.setsockopt		= tcp_setsockopt,
	.getsockopt		= tcp_getsockopt,
	.recvmsg		= tcp_recvmsg,
	.backlog_rcv		= tcp_v6_do_rcv,
	.hash			= tcp_v6_hash,
2100 2101
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2102 2103 2104 2105
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
2106
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2107 2108 2109 2110 2111
	.sysctl_mem		= sysctl_tcp_mem,
	.sysctl_wmem		= sysctl_tcp_wmem,
	.sysctl_rmem		= sysctl_tcp_rmem,
	.max_header		= MAX_TCP_HEADER,
	.obj_size		= sizeof(struct tcp6_sock),
2112
	.twsk_prot		= &tcp6_timewait_sock_ops,
2113
	.rsk_prot		= &tcp6_request_sock_ops,
2114
	.h.hashinfo		= &tcp_hashinfo,
2115 2116 2117 2118
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2119 2120 2121 2122 2123
};

static struct inet6_protocol tcpv6_protocol = {
	.handler	=	tcp_v6_rcv,
	.err_handler	=	tcp_v6_err,
2124
	.gso_send_check	=	tcp_v6_gso_send_check,
H
Herbert Xu 已提交
2125
	.gso_segment	=	tcp_tso_segment,
L
Linus Torvalds 已提交
2126 2127 2128 2129 2130 2131 2132 2133 2134 2135
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

static struct inet_protosw tcpv6_protosw = {
	.type		=	SOCK_STREAM,
	.protocol	=	IPPROTO_TCP,
	.prot		=	&tcpv6_prot,
	.ops		=	&inet6_stream_ops,
	.capability	=	-1,
	.no_check	=	0,
2136 2137
	.flags		=	INET_PROTOSW_PERMANENT |
				INET_PROTOSW_ICSK,
L
Linus Torvalds 已提交
2138 2139
};

2140 2141
static int tcpv6_net_init(struct net *net)
{
2142 2143
	return inet_ctl_sock_create(&net->ipv6.tcp_sk, PF_INET6,
				    SOCK_RAW, IPPROTO_TCP, net);
2144 2145 2146 2147
}

static void tcpv6_net_exit(struct net *net)
{
2148
	inet_ctl_sock_destroy(net->ipv6.tcp_sk);
2149 2150 2151 2152 2153 2154 2155
}

static struct pernet_operations tcpv6_net_ops = {
	.init = tcpv6_net_init,
	.exit = tcpv6_net_exit,
};

2156
int __init tcpv6_init(void)
L
Linus Torvalds 已提交
2157
{
2158 2159 2160 2161 2162 2163
	int ret;

	ret = inet6_add_protocol(&tcpv6_protocol, IPPROTO_TCP);
	if (ret)
		goto out;

L
Linus Torvalds 已提交
2164
	/* register inet6 protocol */
2165 2166 2167 2168
	ret = inet6_register_protosw(&tcpv6_protosw);
	if (ret)
		goto out_tcpv6_protocol;

2169
	ret = register_pernet_subsys(&tcpv6_net_ops);
2170 2171 2172 2173
	if (ret)
		goto out_tcpv6_protosw;
out:
	return ret;
2174

2175 2176 2177 2178 2179 2180 2181
out_tcpv6_protocol:
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
out_tcpv6_protosw:
	inet6_unregister_protosw(&tcpv6_protosw);
	goto out;
}

2182
void tcpv6_exit(void)
2183
{
2184
	unregister_pernet_subsys(&tcpv6_net_ops);
2185 2186
	inet6_unregister_protosw(&tcpv6_protosw);
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
L
Linus Torvalds 已提交
2187
}