tcp_ipv6.c 51.3 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
 *	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.
 */

H
Herbert Xu 已提交
26
#include <linux/bottom_half.h>
L
Linus Torvalds 已提交
27 28 29 30 31 32 33 34 35 36 37 38 39 40
#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>
41
#include <linux/slab.h>
W
Wang Yufen 已提交
42
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
43 44 45 46 47 48
#include <linux/ipv6.h>
#include <linux/icmpv6.h>
#include <linux/random.h>

#include <net/tcp.h>
#include <net/ndisc.h>
49
#include <net/inet6_hashtables.h>
50
#include <net/inet6_connection_sock.h>
L
Linus Torvalds 已提交
51 52 53 54 55 56 57 58 59 60
#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>
61
#include <net/timewait_sock.h>
62
#include <net/inet_common.h>
63
#include <net/secure_seq.h>
64
#include <net/busy_poll.h>
L
Linus Torvalds 已提交
65 66 67 68

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

H
Herbert Xu 已提交
69
#include <crypto/hash.h>
70 71
#include <linux/scatterlist.h>

72 73
static void	tcp_v6_send_reset(const struct sock *sk, struct sk_buff *skb);
static void	tcp_v6_reqsk_send_ack(const struct sock *sk, struct sk_buff *skb,
74
				      struct request_sock *req);
L
Linus Torvalds 已提交
75 76 77

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

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

91 92 93 94
static void inet6_sk_rx_dst_set(struct sock *sk, const struct sk_buff *skb)
{
	struct dst_entry *dst = skb_dst(skb);

E
Eric Dumazet 已提交
95
	if (dst && dst_hold_safe(dst)) {
96 97 98 99
		const struct rt6_info *rt = (const struct rt6_info *)dst;

		sk->sk_rx_dst = dst;
		inet_sk(sk)->rx_dst_ifindex = skb->skb_iif;
100
		inet6_sk(sk)->rx_dst_cookie = rt6_get_cookie(rt);
101
	}
102 103
}

104
static u32 tcp_v6_init_seq(const struct sk_buff *skb)
L
Linus Torvalds 已提交
105
{
106 107 108 109 110 111
	return secure_tcpv6_seq(ipv6_hdr(skb)->daddr.s6_addr32,
				ipv6_hdr(skb)->saddr.s6_addr32,
				tcp_hdr(skb)->dest,
				tcp_hdr(skb)->source);
}

112
static u32 tcp_v6_init_ts_off(const struct net *net, const struct sk_buff *skb)
113
{
114
	return secure_tcpv6_ts_off(net, ipv6_hdr(skb)->daddr.s6_addr32,
115
				   ipv6_hdr(skb)->saddr.s6_addr32);
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
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
126
	struct in6_addr *saddr = NULL, *final_p, final;
127
	struct ipv6_txoptions *opt;
128
	struct flowi6 fl6;
L
Linus Torvalds 已提交
129 130 131
	struct dst_entry *dst;
	int addr_type;
	int err;
132
	struct inet_timewait_death_row *tcp_death_row = &sock_net(sk)->ipv4.tcp_death_row;
L
Linus Torvalds 已提交
133

134
	if (addr_len < SIN6_LEN_RFC2133)
L
Linus Torvalds 已提交
135 136
		return -EINVAL;

137
	if (usin->sin6_family != AF_INET6)
E
Eric Dumazet 已提交
138
		return -EAFNOSUPPORT;
L
Linus Torvalds 已提交
139

140
	memset(&fl6, 0, sizeof(fl6));
L
Linus Torvalds 已提交
141 142

	if (np->sndflow) {
143 144 145
		fl6.flowlabel = usin->sin6_flowinfo&IPV6_FLOWINFO_MASK;
		IP6_ECN_flow_init(fl6.flowlabel);
		if (fl6.flowlabel&IPV6_FLOWLABEL_MASK) {
L
Linus Torvalds 已提交
146
			struct ip6_flowlabel *flowlabel;
147
			flowlabel = fl6_sock_lookup(sk, fl6.flowlabel);
148
			if (!flowlabel)
L
Linus Torvalds 已提交
149 150 151 152 153 154
				return -EINVAL;
			fl6_sock_release(flowlabel);
		}
	}

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

158 159 160 161 162 163 164
	if (ipv6_addr_any(&usin->sin6_addr)) {
		if (ipv6_addr_v4mapped(&sk->sk_v6_rcv_saddr))
			ipv6_addr_set_v4mapped(htonl(INADDR_LOOPBACK),
					       &usin->sin6_addr);
		else
			usin->sin6_addr = in6addr_loopback;
	}
L
Linus Torvalds 已提交
165 166 167

	addr_type = ipv6_addr_type(&usin->sin6_addr);

W
Weilong Chen 已提交
168
	if (addr_type & IPV6_ADDR_MULTICAST)
L
Linus Torvalds 已提交
169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
		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 &&
190
	    !ipv6_addr_equal(&sk->sk_v6_daddr, &usin->sin6_addr)) {
L
Linus Torvalds 已提交
191 192 193 194 195
		tp->rx_opt.ts_recent = 0;
		tp->rx_opt.ts_recent_stamp = 0;
		tp->write_seq = 0;
	}

196
	sk->sk_v6_daddr = usin->sin6_addr;
197
	np->flow_label = fl6.flowlabel;
L
Linus Torvalds 已提交
198 199 200 201 202

	/*
	 *	TCP over IPv4
	 */

203
	if (addr_type & IPV6_ADDR_MAPPED) {
204
		u32 exthdrlen = icsk->icsk_ext_hdr_len;
L
Linus Torvalds 已提交
205 206 207 208 209 210 211 212 213 214 215
		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];

216
		icsk->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
217
		sk->sk_backlog_rcv = tcp_v4_do_rcv;
218 219 220
#ifdef CONFIG_TCP_MD5SIG
		tp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif
L
Linus Torvalds 已提交
221 222 223 224

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

		if (err) {
225 226
			icsk->icsk_ext_hdr_len = exthdrlen;
			icsk->icsk_af_ops = &ipv6_specific;
L
Linus Torvalds 已提交
227
			sk->sk_backlog_rcv = tcp_v6_do_rcv;
228 229 230
#ifdef CONFIG_TCP_MD5SIG
			tp->af_specific = &tcp_sock_ipv6_specific;
#endif
L
Linus Torvalds 已提交
231 232
			goto failure;
		}
233
		np->saddr = sk->sk_v6_rcv_saddr;
L
Linus Torvalds 已提交
234 235 236 237

		return err;
	}

238 239
	if (!ipv6_addr_any(&sk->sk_v6_rcv_saddr))
		saddr = &sk->sk_v6_rcv_saddr;
L
Linus Torvalds 已提交
240

241
	fl6.flowi6_proto = IPPROTO_TCP;
242
	fl6.daddr = sk->sk_v6_daddr;
A
Alexey Dobriyan 已提交
243
	fl6.saddr = saddr ? *saddr : np->saddr;
244 245
	fl6.flowi6_oif = sk->sk_bound_dev_if;
	fl6.flowi6_mark = sk->sk_mark;
246 247
	fl6.fl6_dport = usin->sin6_port;
	fl6.fl6_sport = inet->inet_sport;
248
	fl6.flowi6_uid = sk->sk_uid;
L
Linus Torvalds 已提交
249

250
	opt = rcu_dereference_protected(np->opt, lockdep_sock_is_held(sk));
251
	final_p = fl6_update_dst(&fl6, opt, &final);
L
Linus Torvalds 已提交
252

253
	security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
V
Venkat Yekkirala 已提交
254

255
	dst = ip6_dst_lookup_flow(sk, &fl6, final_p);
256 257
	if (IS_ERR(dst)) {
		err = PTR_ERR(dst);
L
Linus Torvalds 已提交
258
		goto failure;
259
	}
L
Linus Torvalds 已提交
260

261
	if (!saddr) {
262
		saddr = &fl6.saddr;
263
		sk->sk_v6_rcv_saddr = *saddr;
L
Linus Torvalds 已提交
264 265 266
	}

	/* set the source address */
A
Alexey Dobriyan 已提交
267
	np->saddr = *saddr;
E
Eric Dumazet 已提交
268
	inet->inet_rcv_saddr = LOOPBACK4_IPV6;
L
Linus Torvalds 已提交
269

H
Herbert Xu 已提交
270
	sk->sk_gso_type = SKB_GSO_TCPV6;
E
Eric Dumazet 已提交
271
	ip6_dst_store(sk, dst, NULL, NULL);
L
Linus Torvalds 已提交
272

273
	icsk->icsk_ext_hdr_len = 0;
274 275 276
	if (opt)
		icsk->icsk_ext_hdr_len = opt->opt_flen +
					 opt->opt_nflen;
L
Linus Torvalds 已提交
277 278 279

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

E
Eric Dumazet 已提交
280
	inet->inet_dport = usin->sin6_port;
L
Linus Torvalds 已提交
281 282

	tcp_set_state(sk, TCP_SYN_SENT);
283
	err = inet6_hash_connect(tcp_death_row, sk);
L
Linus Torvalds 已提交
284 285 286
	if (err)
		goto late_failure;

287
	sk_set_txhash(sk);
288

289 290
	if (likely(!tp->repair)) {
		if (!tp->write_seq)
291 292 293 294
			tp->write_seq = secure_tcpv6_seq(np->saddr.s6_addr32,
							 sk->sk_v6_daddr.s6_addr32,
							 inet->inet_sport,
							 inet->inet_dport);
295 296
		tp->tsoffset = secure_tcpv6_ts_off(sock_net(sk),
						   np->saddr.s6_addr32,
297
						   sk->sk_v6_daddr.s6_addr32);
298
	}
L
Linus Torvalds 已提交
299

W
Wei Wang 已提交
300 301 302 303 304
	if (tcp_fastopen_defer_connect(sk, &err))
		return err;
	if (err)
		goto late_failure;

L
Linus Torvalds 已提交
305 306 307 308 309 310 311 312 313
	err = tcp_connect(sk);
	if (err)
		goto late_failure;

	return 0;

late_failure:
	tcp_set_state(sk, TCP_CLOSE);
failure:
E
Eric Dumazet 已提交
314
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
315 316 317 318
	sk->sk_route_caps = 0;
	return err;
}

319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335
static void tcp_v6_mtu_reduced(struct sock *sk)
{
	struct dst_entry *dst;

	if ((1 << sk->sk_state) & (TCPF_LISTEN | TCPF_CLOSE))
		return;

	dst = inet6_csk_update_pmtu(sk, tcp_sk(sk)->mtu_info);
	if (!dst)
		return;

	if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
		tcp_sync_mss(sk, dst_mtu(dst));
		tcp_simple_retransmit(sk);
	}
}

L
Linus Torvalds 已提交
336
static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
337
		u8 type, u8 code, int offset, __be32 info)
L
Linus Torvalds 已提交
338
{
W
Weilong Chen 已提交
339
	const struct ipv6hdr *hdr = (const struct ipv6hdr *)skb->data;
340
	const struct tcphdr *th = (struct tcphdr *)(skb->data+offset);
341 342
	struct net *net = dev_net(skb->dev);
	struct request_sock *fastopen;
L
Linus Torvalds 已提交
343
	struct ipv6_pinfo *np;
344
	struct tcp_sock *tp;
345
	__u32 seq, snd_una;
346
	struct sock *sk;
347
	bool fatal;
348
	int err;
L
Linus Torvalds 已提交
349

350 351 352 353
	sk = __inet6_lookup_established(net, &tcp_hashinfo,
					&hdr->daddr, th->dest,
					&hdr->saddr, ntohs(th->source),
					skb->dev->ifindex);
L
Linus Torvalds 已提交
354

355
	if (!sk) {
E
Eric Dumazet 已提交
356 357
		__ICMP6_INC_STATS(net, __in6_dev_get(skb->dev),
				  ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
358 359 360 361
		return;
	}

	if (sk->sk_state == TCP_TIME_WAIT) {
362
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
363 364
		return;
	}
365
	seq = ntohl(th->seq);
366
	fatal = icmpv6_err_convert(type, code, &err);
367
	if (sk->sk_state == TCP_NEW_SYN_RECV)
368
		return tcp_req_err(sk, seq, fatal);
L
Linus Torvalds 已提交
369 370

	bh_lock_sock(sk);
371
	if (sock_owned_by_user(sk) && type != ICMPV6_PKT_TOOBIG)
372
		__NET_INC_STATS(net, LINUX_MIB_LOCKDROPPEDICMPS);
L
Linus Torvalds 已提交
373 374 375 376

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

377
	if (ipv6_hdr(skb)->hop_limit < inet6_sk(sk)->min_hopcount) {
378
		__NET_INC_STATS(net, LINUX_MIB_TCPMINTTLDROP);
379 380 381
		goto out;
	}

L
Linus Torvalds 已提交
382
	tp = tcp_sk(sk);
383 384 385
	/* XXX (TFO) - tp->snd_una should be ISN (tcp_create_openreq_child() */
	fastopen = tp->fastopen_rsk;
	snd_una = fastopen ? tcp_rsk(fastopen)->snt_isn : tp->snd_una;
L
Linus Torvalds 已提交
386
	if (sk->sk_state != TCP_LISTEN &&
387
	    !between(seq, snd_una, tp->snd_nxt)) {
388
		__NET_INC_STATS(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
389 390 391 392 393
		goto out;
	}

	np = inet6_sk(sk);

394
	if (type == NDISC_REDIRECT) {
395 396
		if (!sock_owned_by_user(sk)) {
			struct dst_entry *dst = __sk_dst_check(sk, np->dst_cookie);
397

398 399 400
			if (dst)
				dst->ops->redirect(dst, sk, skb);
		}
401
		goto out;
402 403
	}

L
Linus Torvalds 已提交
404
	if (type == ICMPV6_PKT_TOOBIG) {
405 406 407 408 409 410 411
		/* We are not interested in TCP_LISTEN and open_requests
		 * (SYN-ACKs send out by Linux are always <576bytes so
		 * they should go through unfragmented).
		 */
		if (sk->sk_state == TCP_LISTEN)
			goto out;

412 413 414
		if (!ip6_sk_accept_pmtu(sk))
			goto out;

415 416 417
		tp->mtu_info = ntohl(info);
		if (!sock_owned_by_user(sk))
			tcp_v6_mtu_reduced(sk);
418
		else if (!test_and_set_bit(TCP_MTU_REDUCED_DEFERRED,
419
					   &sk->sk_tsq_flags))
420
			sock_hold(sk);
L
Linus Torvalds 已提交
421 422 423 424
		goto out;
	}


425
	/* Might be for an request_sock */
L
Linus Torvalds 已提交
426 427
	switch (sk->sk_state) {
	case TCP_SYN_SENT:
428 429 430 431
	case TCP_SYN_RECV:
		/* Only in fast or simultaneous open. If a fast open socket is
		 * is already accepted it is treated as a connected one below.
		 */
432
		if (fastopen && !fastopen->sk)
433 434
			break;

L
Linus Torvalds 已提交
435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456
		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);
}


457
static int tcp_v6_send_synack(const struct sock *sk, struct dst_entry *dst,
458
			      struct flowi *fl,
459
			      struct request_sock *req,
460
			      struct tcp_fastopen_cookie *foc,
461
			      enum tcp_synack_type synack_type)
L
Linus Torvalds 已提交
462
{
463
	struct inet_request_sock *ireq = inet_rsk(req);
L
Linus Torvalds 已提交
464
	struct ipv6_pinfo *np = inet6_sk(sk);
465
	struct ipv6_txoptions *opt;
466
	struct flowi6 *fl6 = &fl->u.ip6;
W
Weilong Chen 已提交
467
	struct sk_buff *skb;
468
	int err = -ENOMEM;
L
Linus Torvalds 已提交
469

470
	/* First, grab a route. */
471 472
	if (!dst && (dst = inet6_csk_route_req(sk, fl6, req,
					       IPPROTO_TCP)) == NULL)
473
		goto done;
474

475
	skb = tcp_make_synack(sk, dst, req, foc, synack_type);
476

L
Linus Torvalds 已提交
477
	if (skb) {
478 479
		__tcp_v6_send_check(skb, &ireq->ir_v6_loc_addr,
				    &ireq->ir_v6_rmt_addr);
L
Linus Torvalds 已提交
480

481
		fl6->daddr = ireq->ir_v6_rmt_addr;
482
		if (np->repflow && ireq->pktopts)
483 484
			fl6->flowlabel = ip6_flowlabel(ipv6_hdr(ireq->pktopts));

485
		rcu_read_lock();
486 487 488
		opt = ireq->ipv6_opt;
		if (!opt)
			opt = rcu_dereference(np->opt);
489
		err = ip6_xmit(sk, skb, fl6, sk->sk_mark, opt, np->tclass);
490
		rcu_read_unlock();
491
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
492 493 494 495 496 497
	}

done:
	return err;
}

498

499
static void tcp_v6_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
500
{
501
	kfree(inet_rsk(req)->ipv6_opt);
502
	kfree_skb(inet_rsk(req)->pktopts);
L
Linus Torvalds 已提交
503 504
}

505
#ifdef CONFIG_TCP_MD5SIG
506
static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(const struct sock *sk,
507
						   const struct in6_addr *addr)
508
{
E
Eric Dumazet 已提交
509
	return tcp_md5_do_lookup(sk, (union tcp_md5_addr *)addr, AF_INET6);
510 511
}

512
static struct tcp_md5sig_key *tcp_v6_md5_lookup(const struct sock *sk,
513
						const struct sock *addr_sk)
514
{
515
	return tcp_v6_md5_do_lookup(sk, &addr_sk->sk_v6_daddr);
516 517
}

518 519
static int tcp_v6_parse_md5_keys(struct sock *sk, int optname,
				 char __user *optval, int optlen)
520 521 522
{
	struct tcp_md5sig cmd;
	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)&cmd.tcpm_addr;
523
	u8 prefixlen;
524 525 526 527 528 529 530 531 532 533

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

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

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

534 535 536 537 538 539 540 541 542 543
	if (optname == TCP_MD5SIG_EXT &&
	    cmd.tcpm_flags & TCP_MD5SIG_FLAG_PREFIX) {
		prefixlen = cmd.tcpm_prefixlen;
		if (prefixlen > 128 || (ipv6_addr_v4mapped(&sin6->sin6_addr) &&
					prefixlen > 32))
			return -EINVAL;
	} else {
		prefixlen = ipv6_addr_v4mapped(&sin6->sin6_addr) ? 32 : 128;
	}

544
	if (!cmd.tcpm_keylen) {
B
Brian Haley 已提交
545
		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
E
Eric Dumazet 已提交
546
			return tcp_md5_do_del(sk, (union tcp_md5_addr *)&sin6->sin6_addr.s6_addr32[3],
547
					      AF_INET, prefixlen);
E
Eric Dumazet 已提交
548
		return tcp_md5_do_del(sk, (union tcp_md5_addr *)&sin6->sin6_addr,
549
				      AF_INET6, prefixlen);
550 551 552 553 554
	}

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

E
Eric Dumazet 已提交
555 556
	if (ipv6_addr_v4mapped(&sin6->sin6_addr))
		return tcp_md5_do_add(sk, (union tcp_md5_addr *)&sin6->sin6_addr.s6_addr32[3],
557
				      AF_INET, prefixlen, cmd.tcpm_key,
558
				      cmd.tcpm_keylen, GFP_KERNEL);
559

E
Eric Dumazet 已提交
560
	return tcp_md5_do_add(sk, (union tcp_md5_addr *)&sin6->sin6_addr,
561 562
			      AF_INET6, prefixlen, cmd.tcpm_key,
			      cmd.tcpm_keylen, GFP_KERNEL);
563 564
}

565 566 567 568
static int tcp_v6_md5_hash_headers(struct tcp_md5sig_pool *hp,
				   const struct in6_addr *daddr,
				   const struct in6_addr *saddr,
				   const struct tcphdr *th, int nbytes)
569 570
{
	struct tcp6_pseudohdr *bp;
571
	struct scatterlist sg;
572
	struct tcphdr *_th;
573

574
	bp = hp->scratch;
575
	/* 1. TCP pseudo-header (RFC2460) */
A
Alexey Dobriyan 已提交
576 577
	bp->saddr = *saddr;
	bp->daddr = *daddr;
578
	bp->protocol = cpu_to_be32(IPPROTO_TCP);
A
Adam Langley 已提交
579
	bp->len = cpu_to_be32(nbytes);
580

581 582 583 584 585 586 587
	_th = (struct tcphdr *)(bp + 1);
	memcpy(_th, th, sizeof(*th));
	_th->check = 0;

	sg_init_one(&sg, bp, sizeof(*bp) + sizeof(*th));
	ahash_request_set_crypt(hp->md5_req, &sg, NULL,
				sizeof(*bp) + sizeof(*th));
H
Herbert Xu 已提交
588
	return crypto_ahash_update(hp->md5_req);
589
}
590

591
static int tcp_v6_md5_hash_hdr(char *md5_hash, const struct tcp_md5sig_key *key,
592
			       const struct in6_addr *daddr, struct in6_addr *saddr,
E
Eric Dumazet 已提交
593
			       const struct tcphdr *th)
594 595
{
	struct tcp_md5sig_pool *hp;
H
Herbert Xu 已提交
596
	struct ahash_request *req;
597 598 599 600

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;
H
Herbert Xu 已提交
601
	req = hp->md5_req;
602

H
Herbert Xu 已提交
603
	if (crypto_ahash_init(req))
604
		goto clear_hash;
605
	if (tcp_v6_md5_hash_headers(hp, daddr, saddr, th, th->doff << 2))
606 607 608
		goto clear_hash;
	if (tcp_md5_hash_key(hp, key))
		goto clear_hash;
H
Herbert Xu 已提交
609 610
	ahash_request_set_crypt(req, NULL, md5_hash, 0);
	if (crypto_ahash_final(req))
611 612 613 614
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
615

616 617 618 619
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
620
	return 1;
621 622
}

623 624
static int tcp_v6_md5_hash_skb(char *md5_hash,
			       const struct tcp_md5sig_key *key,
E
Eric Dumazet 已提交
625 626
			       const struct sock *sk,
			       const struct sk_buff *skb)
627
{
628
	const struct in6_addr *saddr, *daddr;
629
	struct tcp_md5sig_pool *hp;
H
Herbert Xu 已提交
630
	struct ahash_request *req;
E
Eric Dumazet 已提交
631
	const struct tcphdr *th = tcp_hdr(skb);
632

633 634
	if (sk) { /* valid for establish/request sockets */
		saddr = &sk->sk_v6_rcv_saddr;
635
		daddr = &sk->sk_v6_daddr;
636
	} else {
637
		const struct ipv6hdr *ip6h = ipv6_hdr(skb);
638 639
		saddr = &ip6h->saddr;
		daddr = &ip6h->daddr;
640
	}
641 642 643 644

	hp = tcp_get_md5sig_pool();
	if (!hp)
		goto clear_hash_noput;
H
Herbert Xu 已提交
645
	req = hp->md5_req;
646

H
Herbert Xu 已提交
647
	if (crypto_ahash_init(req))
648 649
		goto clear_hash;

650
	if (tcp_v6_md5_hash_headers(hp, daddr, saddr, th, skb->len))
651 652 653 654 655
		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;
H
Herbert Xu 已提交
656 657
	ahash_request_set_crypt(req, NULL, md5_hash, 0);
	if (crypto_ahash_final(req))
658 659 660 661 662 663 664 665 666 667
		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;
668 669
}

670 671 672 673
#endif

static bool tcp_v6_inbound_md5_hash(const struct sock *sk,
				    const struct sk_buff *skb)
674
{
675
#ifdef CONFIG_TCP_MD5SIG
676
	const __u8 *hash_location = NULL;
677
	struct tcp_md5sig_key *hash_expected;
678
	const struct ipv6hdr *ip6h = ipv6_hdr(skb);
E
Eric Dumazet 已提交
679
	const struct tcphdr *th = tcp_hdr(skb);
680 681 682 683
	int genhash;
	u8 newhash[16];

	hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr);
684
	hash_location = tcp_parse_md5sig_option(th);
685

686 687
	/* We've parsed the options - do we have a hash? */
	if (!hash_expected && !hash_location)
688
		return false;
689 690

	if (hash_expected && !hash_location) {
691
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND);
692
		return true;
693 694
	}

695
	if (!hash_expected && hash_location) {
696
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
697
		return true;
698 699 700
	}

	/* check the signature */
701 702
	genhash = tcp_v6_md5_hash_skb(newhash,
				      hash_expected,
703
				      NULL, skb);
704

705
	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
706
		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5FAILURE);
707 708 709 710
		net_info_ratelimited("MD5 Hash %s for [%pI6c]:%u->[%pI6c]:%u\n",
				     genhash ? "failed" : "mismatch",
				     &ip6h->saddr, ntohs(th->source),
				     &ip6h->daddr, ntohs(th->dest));
711
		return true;
712
	}
713
#endif
714
	return false;
715 716
}

717 718
static void tcp_v6_init_req(struct request_sock *req,
			    const struct sock *sk_listener,
719 720 721
			    struct sk_buff *skb)
{
	struct inet_request_sock *ireq = inet_rsk(req);
722
	const struct ipv6_pinfo *np = inet6_sk(sk_listener);
723 724 725 726 727

	ireq->ir_v6_rmt_addr = ipv6_hdr(skb)->saddr;
	ireq->ir_v6_loc_addr = ipv6_hdr(skb)->daddr;

	/* So that link locals have meaning */
728
	if (!sk_listener->sk_bound_dev_if &&
729
	    ipv6_addr_type(&ireq->ir_v6_rmt_addr) & IPV6_ADDR_LINKLOCAL)
E
Eric Dumazet 已提交
730
		ireq->ir_iif = tcp_v6_iif(skb);
731

732
	if (!TCP_SKB_CB(skb)->tcp_tw_isn &&
733
	    (ipv6_opt_accepted(sk_listener, skb, &TCP_SKB_CB(skb)->header.h6) ||
734
	     np->rxopt.bits.rxinfo ||
735 736 737 738 739 740 741
	     np->rxopt.bits.rxoinfo || np->rxopt.bits.rxhlim ||
	     np->rxopt.bits.rxohlim || np->repflow)) {
		atomic_inc(&skb->users);
		ireq->pktopts = skb;
	}
}

742 743
static struct dst_entry *tcp_v6_route_req(const struct sock *sk,
					  struct flowi *fl,
744
					  const struct request_sock *req)
745
{
746
	return inet6_csk_route_req(sk, &fl->u.ip6, req, IPPROTO_TCP);
747 748
}

749
struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
750
	.family		=	AF_INET6,
751
	.obj_size	=	sizeof(struct tcp6_request_sock),
752
	.rtx_syn_ack	=	tcp_rtx_synack,
753 754
	.send_ack	=	tcp_v6_reqsk_send_ack,
	.destructor	=	tcp_v6_reqsk_destructor,
755
	.send_reset	=	tcp_v6_send_reset,
W
Wang Yufen 已提交
756
	.syn_ack_timeout =	tcp_syn_ack_timeout,
L
Linus Torvalds 已提交
757 758
};

759
static const struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
760 761
	.mss_clamp	=	IPV6_MIN_MTU - sizeof(struct tcphdr) -
				sizeof(struct ipv6hdr),
762
#ifdef CONFIG_TCP_MD5SIG
763
	.req_md5_lookup	=	tcp_v6_md5_lookup,
764
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
765
#endif
766
	.init_req	=	tcp_v6_init_req,
767 768 769
#ifdef CONFIG_SYN_COOKIES
	.cookie_init_seq =	cookie_v6_init_sequence,
#endif
770
	.route_req	=	tcp_v6_route_req,
771 772
	.init_seq	=	tcp_v6_init_seq,
	.init_ts_off	=	tcp_v6_init_ts_off,
773
	.send_synack	=	tcp_v6_send_synack,
774
};
775

776
static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32 seq,
777 778
				 u32 ack, u32 win, u32 tsval, u32 tsecr,
				 int oif, struct tcp_md5sig_key *key, int rst,
779
				 u8 tclass, __be32 label)
L
Linus Torvalds 已提交
780
{
781 782
	const struct tcphdr *th = tcp_hdr(skb);
	struct tcphdr *t1;
L
Linus Torvalds 已提交
783
	struct sk_buff *buff;
784
	struct flowi6 fl6;
785
	struct net *net = sk ? sock_net(sk) : dev_net(skb_dst(skb)->dev);
786
	struct sock *ctl_sk = net->ipv6.tcp_sk;
787
	unsigned int tot_len = sizeof(struct tcphdr);
E
Eric Dumazet 已提交
788
	struct dst_entry *dst;
789
	__be32 *topt;
L
Linus Torvalds 已提交
790

791
	if (tsecr)
792
		tot_len += TCPOLEN_TSTAMP_ALIGNED;
793 794 795 796 797 798
#ifdef CONFIG_TCP_MD5SIG
	if (key)
		tot_len += TCPOLEN_MD5SIG_ALIGNED;
#endif

	buff = alloc_skb(MAX_HEADER + sizeof(struct ipv6hdr) + tot_len,
L
Linus Torvalds 已提交
799
			 GFP_ATOMIC);
800
	if (!buff)
801
		return;
L
Linus Torvalds 已提交
802

803
	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
L
Linus Torvalds 已提交
804

805
	t1 = skb_push(buff, tot_len);
806
	skb_reset_transport_header(buff);
L
Linus Torvalds 已提交
807 808 809 810 811

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
812
	t1->doff = tot_len / 4;
813 814 815 816 817
	t1->seq = htonl(seq);
	t1->ack_seq = htonl(ack);
	t1->ack = !rst || !th->ack;
	t1->rst = rst;
	t1->window = htons(win);
L
Linus Torvalds 已提交
818

819 820
	topt = (__be32 *)(t1 + 1);

821
	if (tsecr) {
822 823
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
824 825
		*topt++ = htonl(tsval);
		*topt++ = htonl(tsecr);
826 827
	}

828 829
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
830 831 832
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG);
		tcp_v6_md5_hash_hdr((__u8 *)topt, key,
833 834
				    &ipv6_hdr(skb)->saddr,
				    &ipv6_hdr(skb)->daddr, t1);
835 836 837
	}
#endif

838
	memset(&fl6, 0, sizeof(fl6));
A
Alexey Dobriyan 已提交
839 840
	fl6.daddr = ipv6_hdr(skb)->saddr;
	fl6.saddr = ipv6_hdr(skb)->daddr;
841
	fl6.flowlabel = label;
L
Linus Torvalds 已提交
842

843 844 845
	buff->ip_summed = CHECKSUM_PARTIAL;
	buff->csum = 0;

846
	__tcp_v6_send_check(buff, &fl6.saddr, &fl6.daddr);
L
Linus Torvalds 已提交
847

848
	fl6.flowi6_proto = IPPROTO_TCP;
849
	if (rt6_need_strict(&fl6.daddr) && !oif)
E
Eric Dumazet 已提交
850
		fl6.flowi6_oif = tcp_v6_iif(skb);
851 852 853 854 855 856
	else {
		if (!oif && netif_index_is_l3_master(net, skb->skb_iif))
			oif = skb->skb_iif;

		fl6.flowi6_oif = oif;
	}
857

858
	fl6.flowi6_mark = IP6_REPLY_MARK(net, skb->mark);
859 860
	fl6.fl6_dport = t1->dest;
	fl6.fl6_sport = t1->source;
861
	fl6.flowi6_uid = sock_net_uid(net, sk && sk_fullsock(sk) ? sk : NULL);
862
	security_skb_classify_flow(skb, flowi6_to_flowi(&fl6));
L
Linus Torvalds 已提交
863

864 865 866 867
	/* Pass a socket to ip6_dst_lookup either it is for RST
	 * Underlying function will use this to retrieve the network
	 * namespace
	 */
868
	dst = ip6_dst_lookup_flow(ctl_sk, &fl6, NULL);
869 870
	if (!IS_ERR(dst)) {
		skb_dst_set(buff, dst);
871
		ip6_xmit(ctl_sk, buff, &fl6, fl6.flowi6_mark, NULL, tclass);
872
		TCP_INC_STATS(net, TCP_MIB_OUTSEGS);
873
		if (rst)
874
			TCP_INC_STATS(net, TCP_MIB_OUTRSTS);
875
		return;
L
Linus Torvalds 已提交
876 877 878 879 880
	}

	kfree_skb(buff);
}

881
static void tcp_v6_send_reset(const struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
882
{
883
	const struct tcphdr *th = tcp_hdr(skb);
884
	u32 seq = 0, ack_seq = 0;
885
	struct tcp_md5sig_key *key = NULL;
886 887 888 889 890 891 892
#ifdef CONFIG_TCP_MD5SIG
	const __u8 *hash_location = NULL;
	struct ipv6hdr *ipv6h = ipv6_hdr(skb);
	unsigned char newhash[16];
	int genhash;
	struct sock *sk1 = NULL;
#endif
W
Wang Yufen 已提交
893
	int oif;
L
Linus Torvalds 已提交
894

895
	if (th->rst)
L
Linus Torvalds 已提交
896 897
		return;

898 899 900 901
	/* If sk not NULL, it means we did a successful lookup and incoming
	 * route had to be correct. prequeue might have dropped our dst.
	 */
	if (!sk && !ipv6_unicast_destination(skb))
902
		return;
L
Linus Torvalds 已提交
903

904
#ifdef CONFIG_TCP_MD5SIG
905
	rcu_read_lock();
906
	hash_location = tcp_parse_md5sig_option(th);
907
	if (sk && sk_fullsock(sk)) {
908 909
		key = tcp_v6_md5_do_lookup(sk, &ipv6h->saddr);
	} else if (hash_location) {
910 911 912 913 914 915 916 917
		/*
		 * active side is lost. Try to find listening socket through
		 * source port, and then find md5 key through listening socket.
		 * we are not loose security here:
		 * Incoming packet is checked with md5 hash with finding key,
		 * no RST generated if md5 hash doesn't match.
		 */
		sk1 = inet6_lookup_listener(dev_net(skb_dst(skb)->dev),
918 919
					   &tcp_hashinfo, NULL, 0,
					   &ipv6h->saddr,
920
					   th->source, &ipv6h->daddr,
E
Eric Dumazet 已提交
921
					   ntohs(th->source), tcp_v6_iif(skb));
922
		if (!sk1)
923
			goto out;
924 925 926

		key = tcp_v6_md5_do_lookup(sk1, &ipv6h->saddr);
		if (!key)
927
			goto out;
928

929
		genhash = tcp_v6_md5_hash_skb(newhash, key, NULL, skb);
930
		if (genhash || memcmp(hash_location, newhash, 16) != 0)
931
			goto out;
932
	}
933 934
#endif

935 936 937 938 939
	if (th->ack)
		seq = ntohl(th->ack_seq);
	else
		ack_seq = ntohl(th->seq) + th->syn + th->fin + skb->len -
			  (th->doff << 2);
L
Linus Torvalds 已提交
940

W
Wang Yufen 已提交
941
	oif = sk ? sk->sk_bound_dev_if : 0;
942
	tcp_v6_send_response(sk, skb, seq, ack_seq, 0, 0, 0, oif, key, 1, 0, 0);
943 944

#ifdef CONFIG_TCP_MD5SIG
945 946
out:
	rcu_read_unlock();
947
#endif
948
}
L
Linus Torvalds 已提交
949

950
static void tcp_v6_send_ack(const struct sock *sk, struct sk_buff *skb, u32 seq,
951
			    u32 ack, u32 win, u32 tsval, u32 tsecr, int oif,
952
			    struct tcp_md5sig_key *key, u8 tclass,
953
			    __be32 label)
954
{
955 956
	tcp_v6_send_response(sk, skb, seq, ack, win, tsval, tsecr, oif, key, 0,
			     tclass, label);
L
Linus Torvalds 已提交
957 958 959 960
}

static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
961
	struct inet_timewait_sock *tw = inet_twsk(sk);
962
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
963

964
	tcp_v6_send_ack(sk, skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
965
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
966
			tcp_time_stamp_raw() + tcptw->tw_ts_offset,
W
Wang Yufen 已提交
967
			tcptw->tw_ts_recent, tw->tw_bound_dev_if, tcp_twsk_md5_key(tcptw),
968
			tw->tw_tclass, cpu_to_be32(tw->tw_flowlabel));
L
Linus Torvalds 已提交
969

970
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
971 972
}

973
static void tcp_v6_reqsk_send_ack(const struct sock *sk, struct sk_buff *skb,
974
				  struct request_sock *req)
L
Linus Torvalds 已提交
975
{
976 977 978
	/* sk->sk_state == TCP_LISTEN -> for regular TCP_SYN_RECV
	 * sk->sk_state == TCP_SYN_RECV -> for Fast Open.
	 */
979 980 981 982 983
	/* RFC 7323 2.3
	 * The window field (SEG.WND) of every outgoing segment, with the
	 * exception of <SYN> segments, MUST be right-shifted by
	 * Rcv.Wind.Shift bits:
	 */
984
	tcp_v6_send_ack(sk, skb, (sk->sk_state == TCP_LISTEN) ?
985
			tcp_rsk(req)->snt_isn + 1 : tcp_sk(sk)->snd_nxt,
986 987
			tcp_rsk(req)->rcv_nxt,
			req->rsk_rcv_wnd >> inet_rsk(req)->rcv_wscale,
988
			tcp_time_stamp_raw() + tcp_rsk(req)->ts_off,
989
			req->ts_recent, sk->sk_bound_dev_if,
990 991
			tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr),
			0, 0);
L
Linus Torvalds 已提交
992 993 994
}


995
static struct sock *tcp_v6_cookie_check(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
996
{
997
#ifdef CONFIG_SYN_COOKIES
998
	const struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
999

1000
	if (!th->syn)
1001
		sk = cookie_v6_check(sk, skb);
L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011
#endif
	return sk;
}

static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
{
	if (skb->protocol == htons(ETH_P_IP))
		return tcp_v4_conn_request(sk, skb);

	if (!ipv6_unicast_destination(skb))
1012
		goto drop;
L
Linus Torvalds 已提交
1013

O
Octavian Purdila 已提交
1014 1015
	return tcp_conn_request(&tcp6_request_sock_ops,
				&tcp_request_sock_ipv6_ops, sk, skb);
L
Linus Torvalds 已提交
1016 1017

drop:
1018
	tcp_listendrop(sk);
L
Linus Torvalds 已提交
1019 1020 1021
	return 0; /* don't send reset */
}

1022 1023 1024 1025 1026 1027 1028 1029 1030 1031
static void tcp_v6_restore_cb(struct sk_buff *skb)
{
	/* We need to move header back to the beginning if xfrm6_policy_check()
	 * and tcp_v6_fill_cb() are going to be called again.
	 * ip6_datagram_recv_specific_ctl() also expects IP6CB to be there.
	 */
	memmove(IP6CB(skb), &TCP_SKB_CB(skb)->header.h6,
		sizeof(struct inet6_skb_parm));
}

1032
static struct sock *tcp_v6_syn_recv_sock(const struct sock *sk, struct sk_buff *skb,
W
Weilong Chen 已提交
1033
					 struct request_sock *req,
1034 1035 1036
					 struct dst_entry *dst,
					 struct request_sock *req_unhash,
					 bool *own_req)
L
Linus Torvalds 已提交
1037
{
1038
	struct inet_request_sock *ireq;
1039 1040
	struct ipv6_pinfo *newnp;
	const struct ipv6_pinfo *np = inet6_sk(sk);
1041
	struct ipv6_txoptions *opt;
L
Linus Torvalds 已提交
1042 1043 1044 1045
	struct tcp6_sock *newtcp6sk;
	struct inet_sock *newinet;
	struct tcp_sock *newtp;
	struct sock *newsk;
1046 1047 1048
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
1049
	struct flowi6 fl6;
L
Linus Torvalds 已提交
1050 1051 1052 1053 1054 1055

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

1056 1057
		newsk = tcp_v4_syn_recv_sock(sk, skb, req, dst,
					     req_unhash, own_req);
L
Linus Torvalds 已提交
1058

1059
		if (!newsk)
L
Linus Torvalds 已提交
1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
			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));

1071
		newnp->saddr = newsk->sk_v6_rcv_saddr;
L
Linus Torvalds 已提交
1072

1073
		inet_csk(newsk)->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
1074
		newsk->sk_backlog_rcv = tcp_v4_do_rcv;
1075 1076 1077 1078
#ifdef CONFIG_TCP_MD5SIG
		newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif

1079
		newnp->ipv6_mc_list = NULL;
1080 1081
		newnp->ipv6_ac_list = NULL;
		newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1082 1083
		newnp->pktoptions  = NULL;
		newnp->opt	   = NULL;
E
Eric Dumazet 已提交
1084
		newnp->mcast_oif   = tcp_v6_iif(skb);
1085
		newnp->mcast_hops  = ipv6_hdr(skb)->hop_limit;
1086
		newnp->rcv_flowinfo = ip6_flowinfo(ipv6_hdr(skb));
1087 1088
		if (np->repflow)
			newnp->flow_label = ip6_flowlabel(ipv6_hdr(skb));
L
Linus Torvalds 已提交
1089

1090 1091 1092 1093
		/*
		 * 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
1094 1095 1096
		 */

		/* It is tricky place. Until this moment IPv4 tcp
1097
		   worked with IPv6 icsk.icsk_af_ops.
L
Linus Torvalds 已提交
1098 1099
		   Sync it now.
		 */
1100
		tcp_sync_mss(newsk, inet_csk(newsk)->icsk_pmtu_cookie);
L
Linus Torvalds 已提交
1101 1102 1103 1104

		return newsk;
	}

1105
	ireq = inet_rsk(req);
L
Linus Torvalds 已提交
1106 1107 1108 1109

	if (sk_acceptq_is_full(sk))
		goto out_overflow;

1110
	if (!dst) {
1111
		dst = inet6_csk_route_req(sk, &fl6, req, IPPROTO_TCP);
1112
		if (!dst)
L
Linus Torvalds 已提交
1113
			goto out;
1114
	}
L
Linus Torvalds 已提交
1115 1116

	newsk = tcp_create_openreq_child(sk, req, skb);
1117
	if (!newsk)
1118
		goto out_nonewsk;
L
Linus Torvalds 已提交
1119

1120 1121 1122 1123 1124
	/*
	 * 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
1125

1126
	newsk->sk_gso_type = SKB_GSO_TCPV6;
E
Eric Dumazet 已提交
1127
	ip6_dst_store(newsk, dst, NULL, NULL);
1128
	inet6_sk_rx_dst_set(newsk, skb);
L
Linus Torvalds 已提交
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138

	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));

1139 1140 1141 1142
	newsk->sk_v6_daddr = ireq->ir_v6_rmt_addr;
	newnp->saddr = ireq->ir_v6_loc_addr;
	newsk->sk_v6_rcv_saddr = ireq->ir_v6_loc_addr;
	newsk->sk_bound_dev_if = ireq->ir_iif;
L
Linus Torvalds 已提交
1143

1144
	/* Now IPv6 options...
L
Linus Torvalds 已提交
1145 1146 1147

	   First: no IPv4 options.
	 */
1148
	newinet->inet_opt = NULL;
1149
	newnp->ipv6_mc_list = NULL;
1150
	newnp->ipv6_ac_list = NULL;
1151
	newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1152 1153 1154 1155 1156 1157

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

	newnp->pktoptions = NULL;
	newnp->opt	  = NULL;
E
Eric Dumazet 已提交
1158
	newnp->mcast_oif  = tcp_v6_iif(skb);
1159
	newnp->mcast_hops = ipv6_hdr(skb)->hop_limit;
1160
	newnp->rcv_flowinfo = ip6_flowinfo(ipv6_hdr(skb));
1161 1162
	if (np->repflow)
		newnp->flow_label = ip6_flowlabel(ipv6_hdr(skb));
L
Linus Torvalds 已提交
1163 1164 1165 1166 1167 1168 1169

	/* 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.
	 */
1170 1171 1172
	opt = ireq->ipv6_opt;
	if (!opt)
		opt = rcu_dereference(np->opt);
1173 1174 1175 1176
	if (opt) {
		opt = ipv6_dup_options(newsk, opt);
		RCU_INIT_POINTER(newnp->opt, opt);
	}
1177
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
1178 1179 1180
	if (opt)
		inet_csk(newsk)->icsk_ext_hdr_len = opt->opt_nflen +
						    opt->opt_flen;
L
Linus Torvalds 已提交
1181

1182 1183
	tcp_ca_openreq_child(newsk, dst);

L
Linus Torvalds 已提交
1184
	tcp_sync_mss(newsk, dst_mtu(dst));
E
Eric Dumazet 已提交
1185
	newtp->advmss = tcp_mss_clamp(tcp_sk(sk), dst_metric_advmss(dst));
1186

L
Linus Torvalds 已提交
1187 1188
	tcp_initialize_rcv_mss(newsk);

E
Eric Dumazet 已提交
1189 1190
	newinet->inet_daddr = newinet->inet_saddr = LOOPBACK4_IPV6;
	newinet->inet_rcv_saddr = LOOPBACK4_IPV6;
L
Linus Torvalds 已提交
1191

1192 1193
#ifdef CONFIG_TCP_MD5SIG
	/* Copy over the MD5 key from the original socket */
W
Wang Yufen 已提交
1194
	key = tcp_v6_md5_do_lookup(sk, &newsk->sk_v6_daddr);
1195
	if (key) {
1196 1197 1198 1199 1200
		/* 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.
		 */
1201
		tcp_md5_do_add(newsk, (union tcp_md5_addr *)&newsk->sk_v6_daddr,
1202
			       AF_INET6, 128, key->key, key->keylen,
1203
			       sk_gfp_mask(sk, GFP_ATOMIC));
1204 1205 1206
	}
#endif

1207
	if (__inet_inherit_port(sk, newsk) < 0) {
1208 1209
		inet_csk_prepare_forced_close(newsk);
		tcp_done(newsk);
1210 1211
		goto out;
	}
1212
	*own_req = inet_ehash_nolisten(newsk, req_to_sk(req_unhash));
E
Eric Dumazet 已提交
1213
	if (*own_req) {
1214
		tcp_move_syn(newtp, req);
E
Eric Dumazet 已提交
1215 1216 1217 1218

		/* Clone pktoptions received with SYN, if we own the req */
		if (ireq->pktopts) {
			newnp->pktoptions = skb_clone(ireq->pktopts,
1219
						      sk_gfp_mask(sk, GFP_ATOMIC));
E
Eric Dumazet 已提交
1220 1221
			consume_skb(ireq->pktopts);
			ireq->pktopts = NULL;
1222 1223
			if (newnp->pktoptions) {
				tcp_v6_restore_cb(newnp->pktoptions);
E
Eric Dumazet 已提交
1224
				skb_set_owner_r(newnp->pktoptions, newsk);
1225
			}
E
Eric Dumazet 已提交
1226
		}
E
Eric Dumazet 已提交
1227
	}
L
Linus Torvalds 已提交
1228 1229 1230 1231

	return newsk;

out_overflow:
1232
	__NET_INC_STATS(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
1233
out_nonewsk:
L
Linus Torvalds 已提交
1234
	dst_release(dst);
1235
out:
1236
	tcp_listendrop(sk);
L
Linus Torvalds 已提交
1237 1238 1239 1240
	return NULL;
}

/* The socket must have it's spinlock held when we get
1241
 * here, unless it is a TCP_LISTEN socket.
L
Linus Torvalds 已提交
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
 *
 * 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);

	/*
	 *	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.
1281
					       --ANK (980728)
L
Linus Torvalds 已提交
1282 1283
	 */
	if (np->rxopt.all)
1284
		opt_skb = skb_clone(skb, sk_gfp_mask(sk, GFP_ATOMIC));
L
Linus Torvalds 已提交
1285 1286

	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
E
Eric Dumazet 已提交
1287 1288
		struct dst_entry *dst = sk->sk_rx_dst;

1289
		sock_rps_save_rxhash(sk, skb);
1290
		sk_mark_napi_id(sk, skb);
E
Eric Dumazet 已提交
1291 1292 1293 1294 1295 1296 1297 1298
		if (dst) {
			if (inet_sk(sk)->rx_dst_ifindex != skb->skb_iif ||
			    dst->ops->check(dst, np->rx_dst_cookie) == NULL) {
				dst_release(dst);
				sk->sk_rx_dst = NULL;
			}
		}

1299
		tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len);
L
Linus Torvalds 已提交
1300 1301 1302 1303 1304
		if (opt_skb)
			goto ipv6_pktoptions;
		return 0;
	}

E
Eric Dumazet 已提交
1305
	if (tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1306 1307
		goto csum_err;

1308
	if (sk->sk_state == TCP_LISTEN) {
1309 1310
		struct sock *nsk = tcp_v6_cookie_check(sk, skb);

L
Linus Torvalds 已提交
1311 1312 1313
		if (!nsk)
			goto discard;

W
Weilong Chen 已提交
1314
		if (nsk != sk) {
L
Linus Torvalds 已提交
1315 1316 1317 1318 1319 1320
			if (tcp_child_process(sk, nsk, skb))
				goto reset;
			if (opt_skb)
				__kfree_skb(opt_skb);
			return 0;
		}
1321
	} else
1322
		sock_rps_save_rxhash(sk, skb);
L
Linus Torvalds 已提交
1323

1324
	if (tcp_rcv_state_process(sk, skb))
L
Linus Torvalds 已提交
1325 1326 1327 1328 1329 1330
		goto reset;
	if (opt_skb)
		goto ipv6_pktoptions;
	return 0;

reset:
1331
	tcp_v6_send_reset(sk, skb);
L
Linus Torvalds 已提交
1332 1333 1334 1335 1336 1337
discard:
	if (opt_skb)
		__kfree_skb(opt_skb);
	kfree_skb(skb);
	return 0;
csum_err:
1338 1339
	TCP_INC_STATS(sock_net(sk), TCP_MIB_CSUMERRORS);
	TCP_INC_STATS(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353
	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))) {
1354
		if (np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo)
E
Eric Dumazet 已提交
1355
			np->mcast_oif = tcp_v6_iif(opt_skb);
1356
		if (np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim)
1357
			np->mcast_hops = ipv6_hdr(opt_skb)->hop_limit;
1358
		if (np->rxopt.bits.rxflow || np->rxopt.bits.rxtclass)
1359
			np->rcv_flowinfo = ip6_flowinfo(ipv6_hdr(opt_skb));
1360 1361
		if (np->repflow)
			np->flow_label = ip6_flowlabel(ipv6_hdr(opt_skb));
1362
		if (ipv6_opt_accepted(sk, opt_skb, &TCP_SKB_CB(opt_skb)->header.h6)) {
L
Linus Torvalds 已提交
1363
			skb_set_owner_r(opt_skb, sk);
1364
			tcp_v6_restore_cb(opt_skb);
L
Linus Torvalds 已提交
1365 1366 1367 1368 1369 1370 1371
			opt_skb = xchg(&np->pktoptions, opt_skb);
		} else {
			__kfree_skb(opt_skb);
			opt_skb = xchg(&np->pktoptions, NULL);
		}
	}

1372
	kfree_skb(opt_skb);
L
Linus Torvalds 已提交
1373 1374 1375
	return 0;
}

1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
static void tcp_v6_fill_cb(struct sk_buff *skb, const struct ipv6hdr *hdr,
			   const struct tcphdr *th)
{
	/* This is tricky: we move IP6CB at its correct location into
	 * TCP_SKB_CB(). It must be done after xfrm6_policy_check(), because
	 * _decode_session6() uses IP6CB().
	 * barrier() makes sure compiler won't play aliasing games.
	 */
	memmove(&TCP_SKB_CB(skb)->header.h6, IP6CB(skb),
		sizeof(struct inet6_skb_parm));
	barrier();

	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)->tcp_flags = tcp_flag_byte(th);
	TCP_SKB_CB(skb)->tcp_tw_isn = 0;
	TCP_SKB_CB(skb)->ip_dsfield = ipv6_get_dsfield(hdr);
	TCP_SKB_CB(skb)->sacked = 0;
}

1398
static int tcp_v6_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
1399
{
1400
	const struct tcphdr *th;
1401
	const struct ipv6hdr *hdr;
1402
	bool refcounted;
L
Linus Torvalds 已提交
1403 1404
	struct sock *sk;
	int ret;
1405
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1406 1407 1408 1409 1410 1411 1412

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

	/*
	 *	Count it even if it's bad.
	 */
E
Eric Dumazet 已提交
1413
	__TCP_INC_STATS(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1414 1415 1416 1417

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

1418
	th = (const struct tcphdr *)skb->data;
L
Linus Torvalds 已提交
1419

1420
	if (unlikely(th->doff < sizeof(struct tcphdr)/4))
L
Linus Torvalds 已提交
1421 1422 1423 1424
		goto bad_packet;
	if (!pskb_may_pull(skb, th->doff*4))
		goto discard_it;

1425
	if (skb_checksum_init(skb, IPPROTO_TCP, ip6_compute_pseudo))
1426
		goto csum_error;
L
Linus Torvalds 已提交
1427

1428
	th = (const struct tcphdr *)skb->data;
1429
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
1430

1431
lookup:
1432
	sk = __inet6_lookup_skb(&tcp_hashinfo, skb, __tcp_hdrlen(th),
1433 1434
				th->source, th->dest, inet6_iif(skb),
				&refcounted);
L
Linus Torvalds 已提交
1435 1436 1437 1438 1439 1440 1441
	if (!sk)
		goto no_tcp_socket;

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

1442 1443
	if (sk->sk_state == TCP_NEW_SYN_RECV) {
		struct request_sock *req = inet_reqsk(sk);
1444
		struct sock *nsk;
1445 1446 1447 1448

		sk = req->rsk_listener;
		tcp_v6_fill_cb(skb, hdr, th);
		if (tcp_v6_inbound_md5_hash(sk, skb)) {
1449
			sk_drops_add(sk, skb);
1450 1451 1452
			reqsk_put(req);
			goto discard_it;
		}
1453
		if (unlikely(sk->sk_state != TCP_LISTEN)) {
1454
			inet_csk_reqsk_queue_drop_and_put(sk, req);
1455 1456
			goto lookup;
		}
1457
		sock_hold(sk);
1458
		refcounted = true;
1459
		nsk = tcp_check_req(sk, skb, req, false);
1460 1461
		if (!nsk) {
			reqsk_put(req);
1462
			goto discard_and_relse;
1463 1464 1465 1466 1467 1468
		}
		if (nsk == sk) {
			reqsk_put(req);
			tcp_v6_restore_cb(skb);
		} else if (tcp_child_process(sk, nsk, skb)) {
			tcp_v6_send_reset(nsk, skb);
1469
			goto discard_and_relse;
1470
		} else {
1471
			sock_put(sk);
1472 1473 1474
			return 0;
		}
	}
1475
	if (hdr->hop_limit < inet6_sk(sk)->min_hopcount) {
1476
		__NET_INC_STATS(net, LINUX_MIB_TCPMINTTLDROP);
1477 1478 1479
		goto discard_and_relse;
	}

L
Linus Torvalds 已提交
1480 1481 1482
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto discard_and_relse;

1483 1484
	tcp_v6_fill_cb(skb, hdr, th);

1485 1486 1487
	if (tcp_v6_inbound_md5_hash(sk, skb))
		goto discard_and_relse;

1488
	if (tcp_filter(sk, skb))
L
Linus Torvalds 已提交
1489
		goto discard_and_relse;
1490 1491
	th = (const struct tcphdr *)skb->data;
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
1492 1493 1494

	skb->dev = NULL;

1495 1496 1497 1498 1499 1500 1501
	if (sk->sk_state == TCP_LISTEN) {
		ret = tcp_v6_do_rcv(sk, skb);
		goto put_and_return;
	}

	sk_incoming_cpu_update(sk);

1502
	bh_lock_sock_nested(sk);
1503
	tcp_segs_in(tcp_sk(sk), skb);
L
Linus Torvalds 已提交
1504 1505
	ret = 0;
	if (!sock_owned_by_user(sk)) {
D
Dan Williams 已提交
1506
		if (!tcp_prequeue(sk, skb))
1507
			ret = tcp_v6_do_rcv(sk, skb);
E
Eric Dumazet 已提交
1508
	} else if (tcp_add_backlog(sk, skb)) {
Z
Zhu Yi 已提交
1509 1510
		goto discard_and_relse;
	}
L
Linus Torvalds 已提交
1511 1512
	bh_unlock_sock(sk);

1513
put_and_return:
1514 1515
	if (refcounted)
		sock_put(sk);
L
Linus Torvalds 已提交
1516 1517 1518 1519 1520 1521
	return ret ? -1 : 0;

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

1522 1523
	tcp_v6_fill_cb(skb, hdr, th);

E
Eric Dumazet 已提交
1524
	if (tcp_checksum_complete(skb)) {
1525
csum_error:
E
Eric Dumazet 已提交
1526
		__TCP_INC_STATS(net, TCP_MIB_CSUMERRORS);
L
Linus Torvalds 已提交
1527
bad_packet:
E
Eric Dumazet 已提交
1528
		__TCP_INC_STATS(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1529
	} else {
1530
		tcp_v6_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1531 1532 1533 1534 1535 1536 1537
	}

discard_it:
	kfree_skb(skb);
	return 0;

discard_and_relse:
1538
	sk_drops_add(sk, skb);
1539 1540
	if (refcounted)
		sock_put(sk);
L
Linus Torvalds 已提交
1541 1542 1543 1544
	goto discard_it;

do_time_wait:
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
1545
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1546 1547 1548
		goto discard_it;
	}

1549 1550
	tcp_v6_fill_cb(skb, hdr, th);

1551 1552 1553
	if (tcp_checksum_complete(skb)) {
		inet_twsk_put(inet_twsk(sk));
		goto csum_error;
L
Linus Torvalds 已提交
1554 1555
	}

1556
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1557 1558 1559 1560
	case TCP_TW_SYN:
	{
		struct sock *sk2;

1561
		sk2 = inet6_lookup_listener(dev_net(skb->dev), &tcp_hashinfo,
1562
					    skb, __tcp_hdrlen(th),
1563
					    &ipv6_hdr(skb)->saddr, th->source,
1564
					    &ipv6_hdr(skb)->daddr,
E
Eric Dumazet 已提交
1565
					    ntohs(th->dest), tcp_v6_iif(skb));
1566
		if (sk2) {
1567
			struct inet_timewait_sock *tw = inet_twsk(sk);
1568
			inet_twsk_deschedule_put(tw);
L
Linus Torvalds 已提交
1569
			sk = sk2;
1570
			tcp_v6_restore_cb(skb);
1571
			refcounted = false;
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576 1577 1578 1579
			goto process;
		}
		/* Fall through to ACK */
	}
	case TCP_TW_ACK:
		tcp_v6_timewait_ack(sk, skb);
		break;
	case TCP_TW_RST:
1580
		tcp_v6_restore_cb(skb);
1581 1582 1583
		tcp_v6_send_reset(sk, skb);
		inet_twsk_deschedule_put(inet_twsk(sk));
		goto discard_it;
W
Wang Yufen 已提交
1584 1585
	case TCP_TW_SUCCESS:
		;
L
Linus Torvalds 已提交
1586 1587 1588 1589
	}
	goto discard_it;
}

E
Eric Dumazet 已提交
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607
static void tcp_v6_early_demux(struct sk_buff *skb)
{
	const struct ipv6hdr *hdr;
	const struct tcphdr *th;
	struct sock *sk;

	if (skb->pkt_type != PACKET_HOST)
		return;

	if (!pskb_may_pull(skb, skb_transport_offset(skb) + sizeof(struct tcphdr)))
		return;

	hdr = ipv6_hdr(skb);
	th = tcp_hdr(skb);

	if (th->doff < sizeof(struct tcphdr) / 4)
		return;

E
Eric Dumazet 已提交
1608
	/* Note : We use inet6_iif() here, not tcp_v6_iif() */
E
Eric Dumazet 已提交
1609 1610 1611 1612 1613 1614 1615
	sk = __inet6_lookup_established(dev_net(skb->dev), &tcp_hashinfo,
					&hdr->saddr, th->source,
					&hdr->daddr, ntohs(th->dest),
					inet6_iif(skb));
	if (sk) {
		skb->sk = sk;
		skb->destructor = sock_edemux;
1616
		if (sk_fullsock(sk)) {
1617
			struct dst_entry *dst = READ_ONCE(sk->sk_rx_dst);
1618

E
Eric Dumazet 已提交
1619
			if (dst)
E
Eric Dumazet 已提交
1620
				dst = dst_check(dst, inet6_sk(sk)->rx_dst_cookie);
E
Eric Dumazet 已提交
1621
			if (dst &&
1622
			    inet_sk(sk)->rx_dst_ifindex == skb->skb_iif)
E
Eric Dumazet 已提交
1623 1624 1625 1626 1627
				skb_dst_set_noref(skb, dst);
		}
	}
}

1628 1629 1630
static struct timewait_sock_ops tcp6_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp6_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
W
Wang Yufen 已提交
1631
	.twsk_destructor = tcp_twsk_destructor,
1632 1633
};

1634
static const struct inet_connection_sock_af_ops ipv6_specific = {
1635 1636 1637
	.queue_xmit	   = inet6_csk_xmit,
	.send_check	   = tcp_v6_send_check,
	.rebuild_header	   = inet6_sk_rebuild_header,
E
Eric Dumazet 已提交
1638
	.sk_rx_dst_set	   = inet6_sk_rx_dst_set,
1639 1640 1641
	.conn_request	   = tcp_v6_conn_request,
	.syn_recv_sock	   = tcp_v6_syn_recv_sock,
	.net_header_len	   = sizeof(struct ipv6hdr),
1642
	.net_frag_header_len = sizeof(struct frag_hdr),
1643 1644 1645 1646
	.setsockopt	   = ipv6_setsockopt,
	.getsockopt	   = ipv6_getsockopt,
	.addr2sockaddr	   = inet6_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in6),
1647
#ifdef CONFIG_COMPAT
1648 1649
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1650
#endif
1651
	.mtu_reduced	   = tcp_v6_mtu_reduced,
L
Linus Torvalds 已提交
1652 1653
};

1654
#ifdef CONFIG_TCP_MD5SIG
1655
static const struct tcp_sock_af_ops tcp_sock_ipv6_specific = {
1656
	.md5_lookup	=	tcp_v6_md5_lookup,
1657
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
1658 1659
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1660
#endif
1661

L
Linus Torvalds 已提交
1662 1663 1664
/*
 *	TCP over IPv4 via INET6 API
 */
1665
static const struct inet_connection_sock_af_ops ipv6_mapped = {
1666 1667 1668
	.queue_xmit	   = ip_queue_xmit,
	.send_check	   = tcp_v4_send_check,
	.rebuild_header	   = inet_sk_rebuild_header,
1669
	.sk_rx_dst_set	   = inet_sk_rx_dst_set,
1670 1671 1672 1673 1674 1675 1676
	.conn_request	   = tcp_v6_conn_request,
	.syn_recv_sock	   = tcp_v6_syn_recv_sock,
	.net_header_len	   = sizeof(struct iphdr),
	.setsockopt	   = ipv6_setsockopt,
	.getsockopt	   = ipv6_getsockopt,
	.addr2sockaddr	   = inet6_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in6),
1677
#ifdef CONFIG_COMPAT
1678 1679
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1680
#endif
1681
	.mtu_reduced	   = tcp_v4_mtu_reduced,
L
Linus Torvalds 已提交
1682 1683
};

1684
#ifdef CONFIG_TCP_MD5SIG
1685
static const struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = {
1686
	.md5_lookup	=	tcp_v4_md5_lookup,
1687
	.calc_md5_hash	=	tcp_v4_md5_hash_skb,
1688 1689
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1690
#endif
1691

L
Linus Torvalds 已提交
1692 1693 1694 1695 1696
/* 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)
{
1697
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1698

1699
	tcp_init_sock(sk);
L
Linus Torvalds 已提交
1700

1701
	icsk->icsk_af_ops = &ipv6_specific;
L
Linus Torvalds 已提交
1702

1703
#ifdef CONFIG_TCP_MD5SIG
1704
	tcp_sk(sk)->af_specific = &tcp_sock_ipv6_specific;
1705 1706
#endif

L
Linus Torvalds 已提交
1707 1708 1709
	return 0;
}

1710
static void tcp_v6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1711 1712
{
	tcp_v4_destroy_sock(sk);
1713
	inet6_destroy_sock(sk);
L
Linus Torvalds 已提交
1714 1715
}

1716
#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1717
/* Proc filesystem TCPv6 sock list dumping. */
1718
static void get_openreq6(struct seq_file *seq,
E
Eric Dumazet 已提交
1719
			 const struct request_sock *req, int i)
L
Linus Torvalds 已提交
1720
{
1721
	long ttd = req->rsk_timer.expires - jiffies;
1722 1723
	const struct in6_addr *src = &inet_rsk(req)->ir_v6_loc_addr;
	const struct in6_addr *dest = &inet_rsk(req)->ir_v6_rmt_addr;
L
Linus Torvalds 已提交
1724 1725 1726 1727 1728 1729

	if (ttd < 0)
		ttd = 0;

	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
1730
		   "%02X %08X:%08X %02X:%08lX %08X %5u %8d %d %d %pK\n",
L
Linus Torvalds 已提交
1731 1732 1733
		   i,
		   src->s6_addr32[0], src->s6_addr32[1],
		   src->s6_addr32[2], src->s6_addr32[3],
E
Eric Dumazet 已提交
1734
		   inet_rsk(req)->ir_num,
L
Linus Torvalds 已提交
1735 1736
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3],
1737
		   ntohs(inet_rsk(req)->ir_rmt_port),
L
Linus Torvalds 已提交
1738
		   TCP_SYN_RECV,
W
Weilong Chen 已提交
1739
		   0, 0, /* could print option size, but that is af dependent. */
1740 1741
		   1,   /* timers active (only the expire timer) */
		   jiffies_to_clock_t(ttd),
1742
		   req->num_timeout,
E
Eric Dumazet 已提交
1743 1744
		   from_kuid_munged(seq_user_ns(seq),
				    sock_i_uid(req->rsk_listener)),
1745
		   0,  /* non standard timer */
L
Linus Torvalds 已提交
1746 1747 1748 1749 1750 1751
		   0, /* open_requests have no inode */
		   0, req);
}

static void get_tcp6_sock(struct seq_file *seq, struct sock *sp, int i)
{
1752
	const struct in6_addr *dest, *src;
L
Linus Torvalds 已提交
1753 1754 1755
	__u16 destp, srcp;
	int timer_active;
	unsigned long timer_expires;
1756 1757
	const struct inet_sock *inet = inet_sk(sp);
	const struct tcp_sock *tp = tcp_sk(sp);
1758
	const struct inet_connection_sock *icsk = inet_csk(sp);
1759
	const struct fastopen_queue *fastopenq = &icsk->icsk_accept_queue.fastopenq;
1760 1761
	int rx_queue;
	int state;
L
Linus Torvalds 已提交
1762

1763 1764
	dest  = &sp->sk_v6_daddr;
	src   = &sp->sk_v6_rcv_saddr;
E
Eric Dumazet 已提交
1765 1766
	destp = ntohs(inet->inet_dport);
	srcp  = ntohs(inet->inet_sport);
1767

1768
	if (icsk->icsk_pending == ICSK_TIME_RETRANS ||
1769
	    icsk->icsk_pending == ICSK_TIME_REO_TIMEOUT ||
1770
	    icsk->icsk_pending == ICSK_TIME_LOSS_PROBE) {
L
Linus Torvalds 已提交
1771
		timer_active	= 1;
1772 1773
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
1774
		timer_active	= 4;
1775
		timer_expires	= icsk->icsk_timeout;
L
Linus Torvalds 已提交
1776 1777 1778 1779 1780 1781 1782 1783
	} else if (timer_pending(&sp->sk_timer)) {
		timer_active	= 2;
		timer_expires	= sp->sk_timer.expires;
	} else {
		timer_active	= 0;
		timer_expires = jiffies;
	}

1784 1785 1786 1787 1788 1789 1790 1791 1792
	state = sk_state_load(sp);
	if (state == TCP_LISTEN)
		rx_queue = sp->sk_ack_backlog;
	else
		/* Because we don't lock the socket,
		 * we might find a transient negative value.
		 */
		rx_queue = max_t(int, tp->rcv_nxt - tp->copied_seq, 0);

L
Linus Torvalds 已提交
1793 1794
	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
1795
		   "%02X %08X:%08X %02X:%08lX %08X %5u %8d %lu %d %pK %lu %lu %u %u %d\n",
L
Linus Torvalds 已提交
1796 1797 1798 1799 1800
		   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,
1801 1802 1803
		   state,
		   tp->write_seq - tp->snd_una,
		   rx_queue,
L
Linus Torvalds 已提交
1804
		   timer_active,
1805
		   jiffies_delta_to_clock_t(timer_expires - jiffies),
1806
		   icsk->icsk_retransmits,
1807
		   from_kuid_munged(seq_user_ns(seq), sock_i_uid(sp)),
1808
		   icsk->icsk_probes_out,
L
Linus Torvalds 已提交
1809 1810
		   sock_i_ino(sp),
		   atomic_read(&sp->sk_refcnt), sp,
1811 1812
		   jiffies_to_clock_t(icsk->icsk_rto),
		   jiffies_to_clock_t(icsk->icsk_ack.ato),
W
Weilong Chen 已提交
1813
		   (icsk->icsk_ack.quick << 1) | icsk->icsk_ack.pingpong,
I
Ilpo Järvinen 已提交
1814
		   tp->snd_cwnd,
1815
		   state == TCP_LISTEN ?
1816
			fastopenq->max_qlen :
1817
			(tcp_in_initial_slowstart(tp) ? -1 : tp->snd_ssthresh)
L
Linus Torvalds 已提交
1818 1819 1820
		   );
}

1821
static void get_timewait6_sock(struct seq_file *seq,
1822
			       struct inet_timewait_sock *tw, int i)
L
Linus Torvalds 已提交
1823
{
1824
	long delta = tw->tw_timer.expires - jiffies;
1825
	const struct in6_addr *dest, *src;
L
Linus Torvalds 已提交
1826 1827
	__u16 destp, srcp;

1828 1829
	dest = &tw->tw_v6_daddr;
	src  = &tw->tw_v6_rcv_saddr;
L
Linus Torvalds 已提交
1830 1831 1832 1833 1834
	destp = ntohs(tw->tw_dport);
	srcp  = ntohs(tw->tw_sport);

	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
D
Dan Rosenberg 已提交
1835
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %pK\n",
L
Linus Torvalds 已提交
1836 1837 1838 1839 1840 1841
		   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,
1842
		   3, jiffies_delta_to_clock_t(delta), 0, 0, 0, 0,
L
Linus Torvalds 已提交
1843 1844 1845 1846 1847 1848
		   atomic_read(&tw->tw_refcnt), tw);
}

static int tcp6_seq_show(struct seq_file *seq, void *v)
{
	struct tcp_iter_state *st;
E
Eric Dumazet 已提交
1849
	struct sock *sk = v;
L
Linus Torvalds 已提交
1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861

	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;

1862 1863 1864
	if (sk->sk_state == TCP_TIME_WAIT)
		get_timewait6_sock(seq, v, st->num);
	else if (sk->sk_state == TCP_NEW_SYN_RECV)
E
Eric Dumazet 已提交
1865
		get_openreq6(seq, v, st->num);
1866 1867
	else
		get_tcp6_sock(seq, v, st->num);
L
Linus Torvalds 已提交
1868 1869 1870 1871
out:
	return 0;
}

1872 1873 1874 1875 1876 1877 1878 1879
static const struct file_operations tcp6_afinfo_seq_fops = {
	.owner   = THIS_MODULE,
	.open    = tcp_seq_open,
	.read    = seq_read,
	.llseek  = seq_lseek,
	.release = seq_release_net
};

L
Linus Torvalds 已提交
1880 1881 1882
static struct tcp_seq_afinfo tcp6_seq_afinfo = {
	.name		= "tcp6",
	.family		= AF_INET6,
1883
	.seq_fops	= &tcp6_afinfo_seq_fops,
1884 1885 1886
	.seq_ops	= {
		.show		= tcp6_seq_show,
	},
L
Linus Torvalds 已提交
1887 1888
};

1889
int __net_init tcp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
1890
{
1891
	return tcp_proc_register(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
1892 1893
}

1894
void tcp6_proc_exit(struct net *net)
L
Linus Torvalds 已提交
1895
{
1896
	tcp_proc_unregister(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
1897 1898 1899 1900 1901 1902 1903 1904 1905
}
#endif

struct proto tcpv6_prot = {
	.name			= "TCPv6",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v6_connect,
	.disconnect		= tcp_disconnect,
1906
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
1907 1908 1909 1910 1911 1912
	.ioctl			= tcp_ioctl,
	.init			= tcp_v6_init_sock,
	.destroy		= tcp_v6_destroy_sock,
	.shutdown		= tcp_shutdown,
	.setsockopt		= tcp_setsockopt,
	.getsockopt		= tcp_getsockopt,
1913
	.keepalive		= tcp_set_keepalive,
L
Linus Torvalds 已提交
1914
	.recvmsg		= tcp_recvmsg,
1915 1916
	.sendmsg		= tcp_sendmsg,
	.sendpage		= tcp_sendpage,
L
Linus Torvalds 已提交
1917
	.backlog_rcv		= tcp_v6_do_rcv,
E
Eric Dumazet 已提交
1918
	.release_cb		= tcp_release_cb,
1919
	.hash			= inet6_hash,
1920 1921
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
1922
	.enter_memory_pressure	= tcp_enter_memory_pressure,
1923
	.leave_memory_pressure	= tcp_leave_memory_pressure,
1924
	.stream_memory_free	= tcp_stream_memory_free,
L
Linus Torvalds 已提交
1925 1926 1927
	.sockets_allocated	= &tcp_sockets_allocated,
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
1928
	.orphan_count		= &tcp_orphan_count,
1929
	.sysctl_mem		= sysctl_tcp_mem,
L
Linus Torvalds 已提交
1930 1931 1932 1933
	.sysctl_wmem		= sysctl_tcp_wmem,
	.sysctl_rmem		= sysctl_tcp_rmem,
	.max_header		= MAX_TCP_HEADER,
	.obj_size		= sizeof(struct tcp6_sock),
1934
	.slab_flags		= SLAB_TYPESAFE_BY_RCU,
1935
	.twsk_prot		= &tcp6_timewait_sock_ops,
1936
	.rsk_prot		= &tcp6_request_sock_ops,
1937
	.h.hashinfo		= &tcp_hashinfo,
1938
	.no_autobind		= true,
1939 1940 1941
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
G
Glauber Costa 已提交
1942
#endif
1943
	.diag_destroy		= tcp_abort,
L
Linus Torvalds 已提交
1944 1945
};

1946
static struct inet6_protocol tcpv6_protocol = {
E
Eric Dumazet 已提交
1947
	.early_demux	=	tcp_v6_early_demux,
1948
	.early_demux_handler =  tcp_v6_early_demux,
L
Linus Torvalds 已提交
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958
	.handler	=	tcp_v6_rcv,
	.err_handler	=	tcp_v6_err,
	.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,
1959 1960
	.flags		=	INET_PROTOSW_PERMANENT |
				INET_PROTOSW_ICSK,
L
Linus Torvalds 已提交
1961 1962
};

1963
static int __net_init tcpv6_net_init(struct net *net)
1964
{
1965 1966
	return inet_ctl_sock_create(&net->ipv6.tcp_sk, PF_INET6,
				    SOCK_RAW, IPPROTO_TCP, net);
1967 1968
}

1969
static void __net_exit tcpv6_net_exit(struct net *net)
1970
{
1971
	inet_ctl_sock_destroy(net->ipv6.tcp_sk);
E
Eric W. Biederman 已提交
1972 1973
}

1974
static void __net_exit tcpv6_net_exit_batch(struct list_head *net_exit_list)
E
Eric W. Biederman 已提交
1975
{
1976
	inet_twsk_purge(&tcp_hashinfo, AF_INET6);
1977 1978 1979
}

static struct pernet_operations tcpv6_net_ops = {
E
Eric W. Biederman 已提交
1980 1981 1982
	.init	    = tcpv6_net_init,
	.exit	    = tcpv6_net_exit,
	.exit_batch = tcpv6_net_exit_batch,
1983 1984
};

1985
int __init tcpv6_init(void)
L
Linus Torvalds 已提交
1986
{
1987 1988
	int ret;

1989 1990
	ret = inet6_add_protocol(&tcpv6_protocol, IPPROTO_TCP);
	if (ret)
1991
		goto out;
1992

L
Linus Torvalds 已提交
1993
	/* register inet6 protocol */
1994 1995 1996 1997
	ret = inet6_register_protosw(&tcpv6_protosw);
	if (ret)
		goto out_tcpv6_protocol;

1998
	ret = register_pernet_subsys(&tcpv6_net_ops);
1999 2000 2001 2002
	if (ret)
		goto out_tcpv6_protosw;
out:
	return ret;
2003

2004 2005
out_tcpv6_protosw:
	inet6_unregister_protosw(&tcpv6_protosw);
2006 2007
out_tcpv6_protocol:
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
2008 2009 2010
	goto out;
}

2011
void tcpv6_exit(void)
2012
{
2013
	unregister_pernet_subsys(&tcpv6_net_ops);
2014 2015
	inet6_unregister_protosw(&tcpv6_protosw);
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
L
Linus Torvalds 已提交
2016
}