tcp_ipv6.c 53.4 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 41 42 43 44 45 46 47
#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>
48
#include <net/inet6_hashtables.h>
49
#include <net/inet6_connection_sock.h>
L
Linus Torvalds 已提交
50 51 52 53 54 55 56 57 58 59
#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>
60
#include <net/timewait_sock.h>
61
#include <net/netdma.h>
62
#include <net/inet_common.h>
L
Linus Torvalds 已提交
63 64 65 66 67 68

#include <asm/uaccess.h>

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

69 70 71 72
#include <linux/crypto.h>
#include <linux/scatterlist.h>

static void	tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb);
73 74
static void	tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
				      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 struct inet_connection_sock_af_ops ipv6_mapped;
static struct inet_connection_sock_af_ops ipv6_specific;
80
#ifdef CONFIG_TCP_MD5SIG
81 82
static struct tcp_sock_af_ops tcp_sock_ipv6_specific;
static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific;
83 84 85 86 87 88
#else
static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk,
						   struct in6_addr *addr)
{
	return NULL;
}
89
#endif
L
Linus Torvalds 已提交
90 91 92 93

static void tcp_v6_hash(struct sock *sk)
{
	if (sk->sk_state != TCP_CLOSE) {
94
		if (inet_csk(sk)->icsk_af_ops == &ipv6_mapped) {
L
Linus Torvalds 已提交
95 96 97 98
			tcp_prot.hash(sk);
			return;
		}
		local_bh_disable();
99
		__inet6_hash(sk);
L
Linus Torvalds 已提交
100 101 102 103
		local_bh_enable();
	}
}

H
Herbert Xu 已提交
104
static __inline__ __sum16 tcp_v6_check(int len,
105 106
				   struct in6_addr *saddr,
				   struct in6_addr *daddr,
107
				   __wsum base)
L
Linus Torvalds 已提交
108 109 110 111
{
	return csum_ipv6_magic(saddr, daddr, len, IPPROTO_TCP, base);
}

112
static __u32 tcp_v6_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
113
{
114 115
	return secure_tcpv6_sequence_number(ipv6_hdr(skb)->daddr.s6_addr32,
					    ipv6_hdr(skb)->saddr.s6_addr32,
116 117
					    tcp_hdr(skb)->dest,
					    tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
118 119
}

120
static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
L
Linus Torvalds 已提交
121 122 123
			  int addr_len)
{
	struct sockaddr_in6 *usin = (struct sockaddr_in6 *) uaddr;
124
	struct inet_sock *inet = inet_sk(sk);
125
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
126 127 128 129 130 131 132 133
	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;

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

137
	if (usin->sin6_family != AF_INET6)
L
Linus Torvalds 已提交
138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
		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);
		}
	}

	/*
156 157 158 159 160
	 *	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
161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199

	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) {
200
		u32 exthdrlen = icsk->icsk_ext_hdr_len;
L
Linus Torvalds 已提交
201 202 203 204 205 206 207 208 209 210 211
		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];

212
		icsk->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
213
		sk->sk_backlog_rcv = tcp_v4_do_rcv;
214 215 216
#ifdef CONFIG_TCP_MD5SIG
		tp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif
L
Linus Torvalds 已提交
217 218 219 220

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

		if (err) {
221 222
			icsk->icsk_ext_hdr_len = exthdrlen;
			icsk->icsk_af_ops = &ipv6_specific;
L
Linus Torvalds 已提交
223
			sk->sk_backlog_rcv = tcp_v6_do_rcv;
224 225 226
#ifdef CONFIG_TCP_MD5SIG
			tp->af_specific = &tcp_sock_ipv6_specific;
#endif
L
Linus Torvalds 已提交
227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255
			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
256 257
	security_sk_classify_flow(sk, &fl);

L
Linus Torvalds 已提交
258 259 260 261 262 263
	err = ip6_dst_lookup(sk, &dst, &fl);
	if (err)
		goto failure;
	if (final_p)
		ipv6_addr_copy(&fl.fl6_dst, final_p);

A
Alexey Dobriyan 已提交
264 265
	err = __xfrm_lookup(sock_net(sk), &dst, &fl, sk, XFRM_LOOKUP_WAIT);
	if (err < 0) {
266 267 268 269 270
		if (err == -EREMOTE)
			err = ip6_dst_blackhole(sk, &dst, &fl);
		if (err < 0)
			goto failure;
	}
L
Linus Torvalds 已提交
271 272 273 274 275 276 277 278 279 280

	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
281
	sk->sk_gso_type = SKB_GSO_TCPV6;
282
	__ip6_dst_store(sk, dst, NULL, NULL);
L
Linus Torvalds 已提交
283

284
	icsk->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
285
	if (np->opt)
286 287
		icsk->icsk_ext_hdr_len = (np->opt->opt_flen +
					  np->opt->opt_nflen);
L
Linus Torvalds 已提交
288 289 290 291 292 293

	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);
294
	err = inet6_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
	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,
320
		int type, int code, int offset, __be32 info)
L
Linus Torvalds 已提交
321 322
{
	struct ipv6hdr *hdr = (struct ipv6hdr*)skb->data;
323
	const struct tcphdr *th = (struct tcphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
324 325 326
	struct ipv6_pinfo *np;
	struct sock *sk;
	int err;
327
	struct tcp_sock *tp;
L
Linus Torvalds 已提交
328
	__u32 seq;
329
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
330

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

	if (sk == NULL) {
335 336
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
337 338 339 340
		return;
	}

	if (sk->sk_state == TCP_TIME_WAIT) {
341
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
342 343 344 345 346
		return;
	}

	bh_lock_sock(sk);
	if (sock_owned_by_user(sk))
347
		NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS);
L
Linus Torvalds 已提交
348 349 350 351 352

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

	tp = tcp_sk(sk);
353
	seq = ntohl(th->seq);
L
Linus Torvalds 已提交
354 355
	if (sk->sk_state != TCP_LISTEN &&
	    !between(seq, tp->snd_una, tp->snd_nxt)) {
356
		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
		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
388
			security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
389 390 391 392 393 394

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

A
Alexey Dobriyan 已提交
395
			if ((err = xfrm_lookup(net, &dst, &fl, sk, 0)) < 0) {
L
Linus Torvalds 已提交
396 397 398 399 400 401 402
				sk->sk_err_soft = -err;
				goto out;
			}

		} else
			dst_hold(dst);

403
		if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
L
Linus Torvalds 已提交
404 405 406 407 408 409 410 411 412
			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);

413
	/* Might be for an request_sock */
L
Linus Torvalds 已提交
414
	switch (sk->sk_state) {
415
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
416 417 418 419
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

420 421
		req = inet6_csk_search_req(sk, &prev, th->dest, &hdr->daddr,
					   &hdr->saddr, inet6_iif(skb));
L
Linus Torvalds 已提交
422 423 424 425 426 427
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		 * an established socket here.
		 */
428
		WARN_ON(req->sk != NULL);
L
Linus Torvalds 已提交
429

430
		if (seq != tcp_rsk(req)->snt_isn) {
431
			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
432 433 434
			goto out;
		}

435
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
436 437 438 439
		goto out;

	case TCP_SYN_SENT:
	case TCP_SYN_RECV:  /* Cannot happen.
440
			       It can, it SYNs are crossed. --ANK */
L
Linus Torvalds 已提交
441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462
		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);
}


463
static int tcp_v6_send_synack(struct sock *sk, struct request_sock *req)
L
Linus Torvalds 已提交
464
{
465
	struct inet6_request_sock *treq = inet6_rsk(req);
L
Linus Torvalds 已提交
466 467 468 469 470
	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;
471
	struct dst_entry *dst;
L
Linus Torvalds 已提交
472 473 474 475
	int err = -1;

	memset(&fl, 0, sizeof(fl));
	fl.proto = IPPROTO_TCP;
476 477
	ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
	ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
478
	fl.fl6_flowlabel = 0;
479 480
	fl.oif = treq->iif;
	fl.fl_ip_dport = inet_rsk(req)->rmt_port;
481
	fl.fl_ip_sport = inet_rsk(req)->loc_port;
482
	security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
483

484 485 486 487 488 489
	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
490 491
	}

492 493 494 495 496
	err = ip6_dst_lookup(sk, &dst, &fl);
	if (err)
		goto done;
	if (final_p)
		ipv6_addr_copy(&fl.fl6_dst, final_p);
A
Alexey Dobriyan 已提交
497
	if ((err = xfrm_lookup(sock_net(sk), &dst, &fl, sk, 0)) < 0)
498 499
		goto done;

L
Linus Torvalds 已提交
500 501
	skb = tcp_make_synack(sk, dst, req);
	if (skb) {
502
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
503

H
Herbert Xu 已提交
504
		th->check = tcp_v6_check(skb->len,
505
					 &treq->loc_addr, &treq->rmt_addr,
506
					 csum_partial(th, skb->len, skb->csum));
L
Linus Torvalds 已提交
507

508
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
509
		err = ip6_xmit(sk, skb, &fl, opt, 0);
510
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
511 512 513
	}

done:
514
	if (opt && opt != np->opt)
L
Linus Torvalds 已提交
515
		sock_kfree_s(sk, opt, opt->tot_len);
516
	dst_release(dst);
L
Linus Torvalds 已提交
517 518 519
	return err;
}

520 521 522 523 524 525 526 527 528 529 530 531 532 533
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));
}

534
static void tcp_v6_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
535
{
536
	kfree_skb(inet6_rsk(req)->pktopts);
L
Linus Torvalds 已提交
537 538
}

539 540 541 542 543 544 545 546 547 548 549 550 551
#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++) {
552
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, addr))
553
			return &tp->md5sig_info->keys6[i].base;
554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573
	}
	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 */
574
	struct tcp_md5sig_key *key;
575 576 577
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp6_md5sig_key *keys;

578
	key = tcp_v6_md5_do_lookup(sk, peer);
579 580
	if (key) {
		/* modify existing entry - just update that one */
581 582 583
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
584 585 586 587 588 589 590 591
	} 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;
			}
592
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
593
		}
594 595 596 597
		if (tcp_alloc_md5sig_pool() == NULL) {
			kfree(newkey);
			return -ENOMEM;
		}
598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619
		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);
620 621
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.key = newkey;
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.keylen = newkeylen;
622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640

		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++) {
641
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, peer)) {
642
			/* Free the key */
643
			kfree(tp->md5sig_info->keys6[i].base.key);
644 645 646 647 648
			tp->md5sig_info->entries6--;

			if (tp->md5sig_info->entries6 == 0) {
				kfree(tp->md5sig_info->keys6);
				tp->md5sig_info->keys6 = NULL;
649
				tp->md5sig_info->alloced6 = 0;
650 651 652 653 654 655 656 657
			} 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]));
			}
658 659
			tcp_free_md5sig_pool();
			return 0;
660 661 662 663 664 665 666 667 668 669 670 671
		}
	}
	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++)
672
			kfree(tp->md5sig_info->keys6[i].base.key);
673 674 675 676 677 678 679 680 681 682
		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++)
683
			kfree(tp->md5sig_info->keys4[i].base.key);
684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711
		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
712
		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
			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;
729
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
730 731
	}

A
Arnaldo Carvalho de Melo 已提交
732
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
733 734
	if (!newkey)
		return -ENOMEM;
B
Brian Haley 已提交
735
	if (ipv6_addr_v4mapped(&sin6->sin6_addr)) {
736 737 738 739 740 741
		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);
}

742 743 744
static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp,
					struct in6_addr *daddr,
					struct in6_addr *saddr, int nbytes)
745 746
{
	struct tcp6_pseudohdr *bp;
747
	struct scatterlist sg;
748

749 750 751 752
	bp = &hp->md5_blk.ip6;
	/* 1. TCP pseudo-header (RFC2460) */
	ipv6_addr_copy(&bp->saddr, saddr);
	ipv6_addr_copy(&bp->daddr, daddr);
753
	bp->protocol = cpu_to_be32(IPPROTO_TCP);
A
Adam Langley 已提交
754
	bp->len = cpu_to_be32(nbytes);
755

756 757 758
	sg_init_one(&sg, bp, sizeof(*bp));
	return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp));
}
759

760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
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))
781 782 783 784
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
785

786 787 788 789
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
790
	return 1;
791 792
}

793 794 795
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)
796 797
{
	struct in6_addr *saddr, *daddr;
798 799 800
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;
	struct tcphdr *th = tcp_hdr(skb);
801 802 803 804

	if (sk) {
		saddr = &inet6_sk(sk)->saddr;
		daddr = &inet6_sk(sk)->daddr;
805
	} else if (req) {
806 807
		saddr = &inet6_rsk(req)->loc_addr;
		daddr = &inet6_rsk(req)->rmt_addr;
808 809 810 811
	} else {
		struct ipv6hdr *ip6h = ipv6_hdr(skb);
		saddr = &ip6h->saddr;
		daddr = &ip6h->daddr;
812
	}
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840

	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;
841 842 843 844 845 846
}

static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb)
{
	__u8 *hash_location = NULL;
	struct tcp_md5sig_key *hash_expected;
847
	struct ipv6hdr *ip6h = ipv6_hdr(skb);
848
	struct tcphdr *th = tcp_hdr(skb);
849 850 851 852
	int genhash;
	u8 newhash[16];

	hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr);
853
	hash_location = tcp_parse_md5sig_option(th);
854

855 856 857 858 859 860
	/* 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);
861 862 863
		return 1;
	}

864 865
	if (!hash_expected && hash_location) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
866 867 868 869
		return 1;
	}

	/* check the signature */
870 871 872 873
	genhash = tcp_v6_md5_hash_skb(newhash,
				      hash_expected,
				      NULL, NULL, skb);

874 875
	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
H
Harvey Harrison 已提交
876
			printk(KERN_INFO "MD5 Hash %s for (%pI6, %u)->(%pI6, %u)\n",
877
			       genhash ? "failed" : "mismatch",
878 879
			       &ip6h->saddr, ntohs(th->source),
			       &ip6h->daddr, ntohs(th->dest));
880 881 882 883 884 885 886
		}
		return 1;
	}
	return 0;
}
#endif

887
struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
888
	.family		=	AF_INET6,
889
	.obj_size	=	sizeof(struct tcp6_request_sock),
L
Linus Torvalds 已提交
890
	.rtx_syn_ack	=	tcp_v6_send_synack,
891 892
	.send_ack	=	tcp_v6_reqsk_send_ack,
	.destructor	=	tcp_v6_reqsk_destructor,
L
Linus Torvalds 已提交
893 894 895
	.send_reset	=	tcp_v6_send_reset
};

896
#ifdef CONFIG_TCP_MD5SIG
897
static struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
898 899
	.md5_lookup	=	tcp_v6_reqsk_md5_lookup,
};
900
#endif
901

902 903 904
static struct timewait_sock_ops tcp6_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp6_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
905
	.twsk_destructor= tcp_twsk_destructor,
906 907
};

908
static void tcp_v6_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
909 910
{
	struct ipv6_pinfo *np = inet6_sk(sk);
911
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
912

913
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
L
Linus Torvalds 已提交
914
		th->check = ~csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,  0);
915
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
916
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
917
	} else {
918
		th->check = csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,
919
					    csum_partial(th, th->doff<<2,
L
Linus Torvalds 已提交
920 921 922 923
							 skb->csum));
	}
}

924 925 926 927 928 929 930 931
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;

932
	ipv6h = ipv6_hdr(skb);
933
	th = tcp_hdr(skb);
934 935 936 937

	th->check = 0;
	th->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr, skb->len,
				     IPPROTO_TCP, 0);
938
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
939
	skb->csum_offset = offsetof(struct tcphdr, check);
940
	skb->ip_summed = CHECKSUM_PARTIAL;
941 942
	return 0;
}
L
Linus Torvalds 已提交
943

H
Herbert Xu 已提交
944 945
static struct sk_buff **tcp6_gro_receive(struct sk_buff **head,
					 struct sk_buff *skb)
H
Herbert Xu 已提交
946
{
H
Herbert Xu 已提交
947
	struct ipv6hdr *iph = skb_gro_network_header(skb);
H
Herbert Xu 已提交
948 949 950

	switch (skb->ip_summed) {
	case CHECKSUM_COMPLETE:
951
		if (!tcp_v6_check(skb_gro_len(skb), &iph->saddr, &iph->daddr,
H
Herbert Xu 已提交
952 953 954 955 956 957 958 959 960 961 962 963 964 965
				  skb->csum)) {
			skb->ip_summed = CHECKSUM_UNNECESSARY;
			break;
		}

		/* fall through */
	case CHECKSUM_NONE:
		NAPI_GRO_CB(skb)->flush = 1;
		return NULL;
	}

	return tcp_gro_receive(head, skb);
}

H
Herbert Xu 已提交
966
static int tcp6_gro_complete(struct sk_buff *skb)
H
Herbert Xu 已提交
967 968 969 970 971 972 973 974 975 976 977
{
	struct ipv6hdr *iph = ipv6_hdr(skb);
	struct tcphdr *th = tcp_hdr(skb);

	th->check = ~tcp_v6_check(skb->len - skb_transport_offset(skb),
				  &iph->saddr, &iph->daddr, 0);
	skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;

	return tcp_gro_complete(skb);
}

978 979
static void tcp_v6_send_response(struct sk_buff *skb, u32 seq, u32 ack, u32 win,
				 u32 ts, struct tcp_md5sig_key *key, int rst)
L
Linus Torvalds 已提交
980
{
981
	struct tcphdr *th = tcp_hdr(skb), *t1;
L
Linus Torvalds 已提交
982 983
	struct sk_buff *buff;
	struct flowi fl;
E
Eric Dumazet 已提交
984
	struct net *net = dev_net(skb_dst(skb)->dev);
985
	struct sock *ctl_sk = net->ipv6.tcp_sk;
986
	unsigned int tot_len = sizeof(struct tcphdr);
E
Eric Dumazet 已提交
987
	struct dst_entry *dst;
988
	__be32 *topt;
L
Linus Torvalds 已提交
989

990 991
	if (ts)
		tot_len += TCPOLEN_TSTAMP_ALIGNED;
992 993 994 995 996 997
#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
998
			 GFP_ATOMIC);
999 1000
	if (buff == NULL)
		return;
L
Linus Torvalds 已提交
1001

1002
	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
L
Linus Torvalds 已提交
1003

1004
	t1 = (struct tcphdr *) skb_push(buff, tot_len);
L
Linus Torvalds 已提交
1005 1006 1007 1008 1009

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
1010
	t1->doff = tot_len / 4;
1011 1012 1013 1014 1015
	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
1016

1017 1018
	topt = (__be32 *)(t1 + 1);

1019 1020 1021 1022 1023 1024 1025
	if (ts) {
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
		*topt++ = htonl(tcp_time_stamp);
		*topt++ = htonl(ts);
	}

1026 1027
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
1028 1029 1030
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG);
		tcp_v6_md5_hash_hdr((__u8 *)topt, key,
1031 1032
				    &ipv6_hdr(skb)->saddr,
				    &ipv6_hdr(skb)->daddr, t1);
1033 1034 1035
	}
#endif

1036
	buff->csum = csum_partial(t1, tot_len, 0);
L
Linus Torvalds 已提交
1037 1038

	memset(&fl, 0, sizeof(fl));
1039 1040
	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
1041 1042

	t1->check = csum_ipv6_magic(&fl.fl6_src, &fl.fl6_dst,
1043
				    tot_len, IPPROTO_TCP,
L
Linus Torvalds 已提交
1044 1045 1046
				    buff->csum);

	fl.proto = IPPROTO_TCP;
1047
	fl.oif = inet6_iif(skb);
L
Linus Torvalds 已提交
1048 1049
	fl.fl_ip_dport = t1->dest;
	fl.fl_ip_sport = t1->source;
V
Venkat Yekkirala 已提交
1050
	security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
1051

1052 1053 1054 1055
	/* Pass a socket to ip6_dst_lookup either it is for RST
	 * Underlying function will use this to retrieve the network
	 * namespace
	 */
E
Eric Dumazet 已提交
1056 1057 1058
	if (!ip6_dst_lookup(ctl_sk, &dst, &fl)) {
		if (xfrm_lookup(net, &dst, &fl, NULL, 0) >= 0) {
			skb_dst_set(buff, dst);
1059
			ip6_xmit(ctl_sk, buff, &fl, NULL, 0);
1060
			TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
1061 1062
			if (rst)
				TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
1063
			return;
A
Arnaldo Carvalho de Melo 已提交
1064
		}
L
Linus Torvalds 已提交
1065 1066 1067 1068 1069
	}

	kfree_skb(buff);
}

1070
static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
1071
{
1072 1073
	struct tcphdr *th = tcp_hdr(skb);
	u32 seq = 0, ack_seq = 0;
1074
	struct tcp_md5sig_key *key = NULL;
L
Linus Torvalds 已提交
1075

1076
	if (th->rst)
L
Linus Torvalds 已提交
1077 1078
		return;

1079 1080
	if (!ipv6_unicast_destination(skb))
		return;
L
Linus Torvalds 已提交
1081

1082
#ifdef CONFIG_TCP_MD5SIG
1083 1084
	if (sk)
		key = tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr);
1085 1086
#endif

1087 1088 1089 1090 1091
	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
1092

1093 1094
	tcp_v6_send_response(skb, seq, ack_seq, 0, 0, key, 1);
}
L
Linus Torvalds 已提交
1095

1096 1097 1098 1099
static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts,
			    struct tcp_md5sig_key *key)
{
	tcp_v6_send_response(skb, seq, ack, win, ts, key, 0);
L
Linus Torvalds 已提交
1100 1101 1102 1103
}

static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
1104
	struct inet_timewait_sock *tw = inet_twsk(sk);
1105
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
1106

1107
	tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
1108
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
1109
			tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw));
L
Linus Torvalds 已提交
1110

1111
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
1112 1113
}

1114 1115
static void tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
				  struct request_sock *req)
L
Linus Torvalds 已提交
1116
{
1117
	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent,
1118
			tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr));
L
Linus Torvalds 已提交
1119 1120 1121 1122 1123
}


static struct sock *tcp_v6_hnd_req(struct sock *sk,struct sk_buff *skb)
{
1124
	struct request_sock *req, **prev;
1125
	const struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1126 1127 1128
	struct sock *nsk;

	/* Find possible connection requests. */
1129
	req = inet6_csk_search_req(sk, &prev, th->source,
1130 1131
				   &ipv6_hdr(skb)->saddr,
				   &ipv6_hdr(skb)->daddr, inet6_iif(skb));
L
Linus Torvalds 已提交
1132 1133 1134
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1135
	nsk = __inet6_lookup_established(sock_net(sk), &tcp_hashinfo,
1136 1137
			&ipv6_hdr(skb)->saddr, th->source,
			&ipv6_hdr(skb)->daddr, ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1138 1139 1140 1141 1142 1143

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1144
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1145 1146 1147
		return NULL;
	}

1148
#ifdef CONFIG_SYN_COOKIES
L
Linus Torvalds 已提交
1149
	if (!th->rst && !th->syn && th->ack)
1150
		sk = cookie_v6_check(sk, skb);
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159
#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)
{
1160
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1161 1162 1163
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_options_received tmp_opt;
	struct tcp_sock *tp = tcp_sk(sk);
1164
	struct request_sock *req = NULL;
L
Linus Torvalds 已提交
1165
	__u32 isn = TCP_SKB_CB(skb)->when;
1166 1167 1168 1169 1170
#ifdef CONFIG_SYN_COOKIES
	int want_cookie = 0;
#else
#define want_cookie 0
#endif
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175

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

	if (!ipv6_unicast_destination(skb))
1176
		goto drop;
L
Linus Torvalds 已提交
1177

1178
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1179
		if (net_ratelimit())
1180 1181 1182 1183 1184 1185
			syn_flood_warning(skb);
#ifdef CONFIG_SYN_COOKIES
		if (sysctl_tcp_syncookies)
			want_cookie = 1;
		else
#endif
1186
		goto drop;
L
Linus Torvalds 已提交
1187 1188
	}

1189
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1190 1191
		goto drop;

1192
	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
L
Linus Torvalds 已提交
1193 1194 1195
	if (req == NULL)
		goto drop;

1196 1197 1198 1199
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops;
#endif

L
Linus Torvalds 已提交
1200 1201 1202 1203 1204 1205
	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);

1206
	if (want_cookie && !tmp_opt.saw_tstamp)
1207 1208
		tcp_clear_options(&tmp_opt);

L
Linus Torvalds 已提交
1209 1210 1211
	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
	tcp_openreq_init(req, &tmp_opt, skb);

1212
	treq = inet6_rsk(req);
1213 1214
	ipv6_addr_copy(&treq->rmt_addr, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&treq->loc_addr, &ipv6_hdr(skb)->daddr);
1215 1216 1217 1218 1219
	if (!want_cookie)
		TCP_ECN_create_request(req, tcp_hdr(skb));

	if (want_cookie) {
		isn = cookie_v6_init_sequence(sk, skb, &req->mss);
1220
		req->cookie_ts = tmp_opt.tstamp_ok;
1221 1222 1223 1224 1225 1226 1227 1228
	} 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
1229

1230 1231 1232 1233
		/* 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
1234

1235
		isn = tcp_v6_init_sequence(skb);
1236
	}
L
Linus Torvalds 已提交
1237

1238
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1239

1240 1241
	security_inet_conn_request(sk, skb, req);

1242
	if (tcp_v6_send_synack(sk, req))
L
Linus Torvalds 已提交
1243 1244
		goto drop;

1245 1246 1247 1248
	if (!want_cookie) {
		inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
		return 0;
	}
L
Linus Torvalds 已提交
1249 1250 1251

drop:
	if (req)
1252
		reqsk_free(req);
L
Linus Torvalds 已提交
1253 1254 1255 1256 1257

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

static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
1258
					  struct request_sock *req,
L
Linus Torvalds 已提交
1259 1260
					  struct dst_entry *dst)
{
1261
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1262 1263 1264 1265 1266 1267
	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;
1268 1269 1270
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1271 1272 1273 1274 1275 1276 1277 1278

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

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

1279
		if (newsk == NULL)
L
Linus Torvalds 已提交
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
			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);

1299
		inet_csk(newsk)->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
1300
		newsk->sk_backlog_rcv = tcp_v4_do_rcv;
1301 1302 1303 1304
#ifdef CONFIG_TCP_MD5SIG
		newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif

L
Linus Torvalds 已提交
1305 1306
		newnp->pktoptions  = NULL;
		newnp->opt	   = NULL;
1307
		newnp->mcast_oif   = inet6_iif(skb);
1308
		newnp->mcast_hops  = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1309

1310 1311 1312 1313
		/*
		 * 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
1314 1315 1316
		 */

		/* It is tricky place. Until this moment IPv4 tcp
1317
		   worked with IPv6 icsk.icsk_af_ops.
L
Linus Torvalds 已提交
1318 1319
		   Sync it now.
		 */
1320
		tcp_sync_mss(newsk, inet_csk(newsk)->icsk_pmtu_cookie);
L
Linus Torvalds 已提交
1321 1322 1323 1324

		return newsk;
	}

1325
	treq = inet6_rsk(req);
L
Linus Torvalds 已提交
1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
	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;
1337
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
1338 1339 1340 1341 1342 1343
		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;
		}
1344
		ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
1345
		fl.oif = sk->sk_bound_dev_if;
1346
		fl.fl_ip_dport = inet_rsk(req)->rmt_port;
1347
		fl.fl_ip_sport = inet_rsk(req)->loc_port;
1348
		security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355

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

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

A
Alexey Dobriyan 已提交
1356
		if ((xfrm_lookup(sock_net(sk), &dst, &fl, sk, 0)) < 0)
L
Linus Torvalds 已提交
1357
			goto out;
1358
	}
L
Linus Torvalds 已提交
1359 1360 1361 1362 1363

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

1364 1365 1366 1367 1368
	/*
	 * 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
1369

1370
	newsk->sk_gso_type = SKB_GSO_TCPV6;
1371
	__ip6_dst_store(newsk, dst, NULL, NULL);
L
Linus Torvalds 已提交
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381

	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));

1382 1383 1384 1385
	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
1386

1387
	/* Now IPv6 options...
L
Linus Torvalds 已提交
1388 1389 1390 1391

	   First: no IPv4 options.
	 */
	newinet->opt = NULL;
1392
	newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397 1398

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

	/* Clone pktoptions received with SYN */
	newnp->pktoptions = NULL;
1399 1400 1401 1402
	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
1403 1404 1405 1406
		if (newnp->pktoptions)
			skb_set_owner_r(newnp->pktoptions, newsk);
	}
	newnp->opt	  = NULL;
1407
	newnp->mcast_oif  = inet6_iif(skb);
1408
	newnp->mcast_hops = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421

	/* 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);
	}

1422
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1423
	if (newnp->opt)
1424 1425
		inet_csk(newsk)->icsk_ext_hdr_len = (newnp->opt->opt_nflen +
						     newnp->opt->opt_flen);
L
Linus Torvalds 已提交
1426

J
John Heffner 已提交
1427
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1428 1429 1430 1431 1432 1433
	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;

1434 1435 1436 1437 1438 1439 1440 1441
#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
1442 1443
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1444 1445 1446 1447 1448
			tcp_v6_md5_do_add(newsk, &inet6_sk(sk)->daddr,
					  newkey, key->keylen);
	}
#endif

1449
	__inet6_hash(newsk);
1450
	__inet_inherit_port(sk, newsk);
L
Linus Torvalds 已提交
1451 1452 1453 1454

	return newsk;

out_overflow:
1455
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
L
Linus Torvalds 已提交
1456
out:
1457
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
L
Linus Torvalds 已提交
1458 1459 1460 1461 1462 1463
	if (opt && opt != np->opt)
		sock_kfree_s(sk, opt, opt->tot_len);
	dst_release(dst);
	return NULL;
}

1464
static __sum16 tcp_v6_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1465
{
1466
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
H
Herbert Xu 已提交
1467
		if (!tcp_v6_check(skb->len, &ipv6_hdr(skb)->saddr,
1468
				  &ipv6_hdr(skb)->daddr, skb->csum)) {
1469
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1470
			return 0;
1471
		}
L
Linus Torvalds 已提交
1472
	}
1473

H
Herbert Xu 已提交
1474
	skb->csum = ~csum_unfold(tcp_v6_check(skb->len,
1475 1476
					      &ipv6_hdr(skb)->saddr,
					      &ipv6_hdr(skb)->daddr, 0));
1477

L
Linus Torvalds 已提交
1478
	if (skb->len <= 76) {
1479
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
	}
	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);

1509 1510 1511 1512 1513
#ifdef CONFIG_TCP_MD5SIG
	if (tcp_v6_inbound_md5_hash (sk, skb))
		goto discard;
#endif

1514
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532
		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.
1533
					       --ANK (980728)
L
Linus Torvalds 已提交
1534 1535 1536 1537 1538 1539
	 */
	if (np->rxopt.all)
		opt_skb = skb_clone(skb, GFP_ATOMIC);

	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1540
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1541 1542 1543 1544 1545 1546 1547
			goto reset;
		TCP_CHECK_TIMER(sk);
		if (opt_skb)
			goto ipv6_pktoptions;
		return 0;
	}

1548
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1549 1550
		goto csum_err;

1551
	if (sk->sk_state == TCP_LISTEN) {
L
Linus Torvalds 已提交
1552 1553 1554 1555 1556 1557 1558 1559 1560
		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..
		 */
1561
		if(nsk != sk) {
L
Linus Torvalds 已提交
1562 1563 1564 1565 1566 1567 1568 1569 1570
			if (tcp_child_process(sk, nsk, skb))
				goto reset;
			if (opt_skb)
				__kfree_skb(opt_skb);
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1571
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576 1577 1578
		goto reset;
	TCP_CHECK_TIMER(sk);
	if (opt_skb)
		goto ipv6_pktoptions;
	return 0;

reset:
1579
	tcp_v6_send_reset(sk, skb);
L
Linus Torvalds 已提交
1580 1581 1582 1583 1584 1585
discard:
	if (opt_skb)
		__kfree_skb(opt_skb);
	kfree_skb(skb);
	return 0;
csum_err:
1586
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600
	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))) {
1601
		if (np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo)
1602
			np->mcast_oif = inet6_iif(opt_skb);
1603
		if (np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim)
1604
			np->mcast_hops = ipv6_hdr(opt_skb)->hop_limit;
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609 1610 1611 1612 1613
		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);
		}
	}

1614
	kfree_skb(opt_skb);
L
Linus Torvalds 已提交
1615 1616 1617
	return 0;
}

1618
static int tcp_v6_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
1619
{
1620
	struct tcphdr *th;
L
Linus Torvalds 已提交
1621 1622
	struct sock *sk;
	int ret;
1623
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1624 1625 1626 1627 1628 1629 1630

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

	/*
	 *	Count it even if it's bad.
	 */
1631
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1632 1633 1634 1635

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

1636
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1637 1638 1639 1640 1641 1642

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

1643
	if (!skb_csum_unnecessary(skb) && tcp_v6_checksum_init(skb))
L
Linus Torvalds 已提交
1644 1645
		goto bad_packet;

1646
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1647 1648 1649 1650 1651
	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;
1652
	TCP_SKB_CB(skb)->flags = ipv6_get_dsfield(ipv6_hdr(skb));
L
Linus Torvalds 已提交
1653 1654
	TCP_SKB_CB(skb)->sacked = 0;

1655
	sk = __inet6_lookup_skb(&tcp_hashinfo, skb, th->source, th->dest);
L
Linus Torvalds 已提交
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665
	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;

1666
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1667 1668 1669 1670
		goto discard_and_relse;

	skb->dev = NULL;

1671
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1672 1673
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1674
#ifdef CONFIG_NET_DMA
1675
		struct tcp_sock *tp = tcp_sk(sk);
1676
		if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
1677
			tp->ucopy.dma_chan = dma_find_channel(DMA_MEMCPY);
1678 1679 1680
		if (tp->ucopy.dma_chan)
			ret = tcp_v6_do_rcv(sk, skb);
		else
C
Chris Leech 已提交
1681 1682 1683 1684 1685
#endif
		{
			if (!tcp_prequeue(sk, skb))
				ret = tcp_v6_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698
	} 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:
1699
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1700
	} else {
1701
		tcp_v6_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
	}

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)) {
1719
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1720 1721 1722 1723
		goto discard_it;
	}

	if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) {
1724
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1725
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1726 1727 1728
		goto discard_it;
	}

1729
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1730 1731 1732 1733
	case TCP_TW_SYN:
	{
		struct sock *sk2;

1734
		sk2 = inet6_lookup_listener(dev_net(skb->dev), &tcp_hashinfo,
1735
					    &ipv6_hdr(skb)->daddr,
1736
					    ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1737
		if (sk2 != NULL) {
1738 1739 1740
			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
1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
			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;
}

1762
static struct inet_connection_sock_af_ops ipv6_specific = {
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773
	.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),
1774
	.bind_conflict	   = inet6_csk_bind_conflict,
1775
#ifdef CONFIG_COMPAT
1776 1777
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1778
#endif
L
Linus Torvalds 已提交
1779 1780
};

1781
#ifdef CONFIG_TCP_MD5SIG
1782
static struct tcp_sock_af_ops tcp_sock_ipv6_specific = {
1783
	.md5_lookup	=	tcp_v6_md5_lookup,
1784
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
1785 1786 1787
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1788
#endif
1789

L
Linus Torvalds 已提交
1790 1791 1792 1793
/*
 *	TCP over IPv4 via INET6 API
 */

1794
static struct inet_connection_sock_af_ops ipv6_mapped = {
1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805
	.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),
1806
	.bind_conflict	   = inet6_csk_bind_conflict,
1807
#ifdef CONFIG_COMPAT
1808 1809
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1810
#endif
L
Linus Torvalds 已提交
1811 1812
};

1813
#ifdef CONFIG_TCP_MD5SIG
1814
static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = {
1815
	.md5_lookup	=	tcp_v4_md5_lookup,
1816
	.calc_md5_hash	=	tcp_v4_md5_hash_skb,
1817 1818 1819
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1820
#endif
1821

L
Linus Torvalds 已提交
1822 1823 1824 1825 1826
/* 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)
{
1827
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1828 1829 1830 1831 1832 1833
	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);

1834
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848
	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;
1849
	tp->mss_cache = 536;
L
Linus Torvalds 已提交
1850 1851 1852 1853 1854

	tp->reordering = sysctl_tcp_reordering;

	sk->sk_state = TCP_CLOSE;

1855
	icsk->icsk_af_ops = &ipv6_specific;
1856
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
1857
	icsk->icsk_sync_mss = tcp_sync_mss;
L
Linus Torvalds 已提交
1858 1859 1860
	sk->sk_write_space = sk_stream_write_space;
	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);

1861 1862 1863 1864
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv6_specific;
#endif

L
Linus Torvalds 已提交
1865 1866 1867
	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

H
Herbert Xu 已提交
1868
	local_bh_disable();
1869
	percpu_counter_inc(&tcp_sockets_allocated);
H
Herbert Xu 已提交
1870
	local_bh_enable();
L
Linus Torvalds 已提交
1871 1872 1873 1874

	return 0;
}

1875
static void tcp_v6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1876
{
1877 1878 1879 1880 1881
#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
1882
	tcp_v4_destroy_sock(sk);
1883
	inet6_destroy_sock(sk);
L
Linus Torvalds 已提交
1884 1885
}

1886
#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1887
/* Proc filesystem TCPv6 sock list dumping. */
1888
static void get_openreq6(struct seq_file *seq,
1889
			 struct sock *sk, struct request_sock *req, int i, int uid)
L
Linus Torvalds 已提交
1890 1891
{
	int ttd = req->expires - jiffies;
1892 1893
	struct in6_addr *src = &inet6_rsk(req)->loc_addr;
	struct in6_addr *dest = &inet6_rsk(req)->rmt_addr;
L
Linus Torvalds 已提交
1894 1895 1896 1897 1898 1899 1900 1901 1902 1903

	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],
1904
		   ntohs(inet_rsk(req)->loc_port),
L
Linus Torvalds 已提交
1905 1906
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3],
1907
		   ntohs(inet_rsk(req)->rmt_port),
L
Linus Torvalds 已提交
1908 1909
		   TCP_SYN_RECV,
		   0,0, /* could print option size, but that is af dependent. */
1910 1911
		   1,   /* timers active (only the expire timer) */
		   jiffies_to_clock_t(ttd),
L
Linus Torvalds 已提交
1912 1913
		   req->retrans,
		   uid,
1914
		   0,  /* non standard timer */
L
Linus Torvalds 已提交
1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926
		   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);
1927
	const struct inet_connection_sock *icsk = inet_csk(sp);
L
Linus Torvalds 已提交
1928 1929 1930 1931 1932 1933
	struct ipv6_pinfo *np = inet6_sk(sp);

	dest  = &np->daddr;
	src   = &np->rcv_saddr;
	destp = ntohs(inet->dport);
	srcp  = ntohs(inet->sport);
1934 1935

	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
1936
		timer_active	= 1;
1937 1938
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
1939
		timer_active	= 4;
1940
		timer_expires	= icsk->icsk_timeout;
L
Linus Torvalds 已提交
1941 1942 1943 1944 1945 1946 1947 1948 1949 1950
	} 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 "
1951
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %lu %lu %u %u %d\n",
L
Linus Torvalds 已提交
1952 1953 1954 1955 1956
		   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,
1957
		   sp->sk_state,
1958 1959
		   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
1960 1961
		   timer_active,
		   jiffies_to_clock_t(timer_expires - jiffies),
1962
		   icsk->icsk_retransmits,
L
Linus Torvalds 已提交
1963
		   sock_i_uid(sp),
1964
		   icsk->icsk_probes_out,
L
Linus Torvalds 已提交
1965 1966
		   sock_i_ino(sp),
		   atomic_read(&sp->sk_refcnt), sp,
1967 1968
		   jiffies_to_clock_t(icsk->icsk_rto),
		   jiffies_to_clock_t(icsk->icsk_ack.ato),
1969
		   (icsk->icsk_ack.quick << 1 ) | icsk->icsk_ack.pingpong,
L
Linus Torvalds 已提交
1970 1971 1972 1973
		   tp->snd_cwnd, tp->snd_ssthresh>=0xFFFF?-1:tp->snd_ssthresh
		   );
}

1974
static void get_timewait6_sock(struct seq_file *seq,
1975
			       struct inet_timewait_sock *tw, int i)
L
Linus Torvalds 已提交
1976 1977 1978
{
	struct in6_addr *dest, *src;
	__u16 destp, srcp;
1979
	struct inet6_timewait_sock *tw6 = inet6_twsk((struct sock *)tw);
L
Linus Torvalds 已提交
1980 1981 1982 1983 1984
	int ttd = tw->tw_ttd - jiffies;

	if (ttd < 0)
		ttd = 0;

1985 1986
	dest = &tw6->tw_v6_daddr;
	src  = &tw6->tw_v6_rcv_saddr;
L
Linus Torvalds 已提交
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036
	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,
2037 2038 2039
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2040 2041 2042
	.seq_ops	= {
		.show		= tcp6_seq_show,
	},
L
Linus Torvalds 已提交
2043 2044
};

2045
int tcp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
2046
{
2047
	return tcp_proc_register(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2048 2049
}

2050
void tcp6_proc_exit(struct net *net)
L
Linus Torvalds 已提交
2051
{
2052
	tcp_proc_unregister(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2053 2054 2055 2056 2057 2058 2059 2060 2061
}
#endif

struct proto tcpv6_prot = {
	.name			= "TCPv6",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v6_connect,
	.disconnect		= tcp_disconnect,
2062
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2063 2064 2065 2066 2067 2068 2069 2070 2071
	.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,
2072 2073
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2074 2075 2076 2077
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
2078
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2079 2080 2081 2082 2083
	.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),
2084
	.slab_flags		= SLAB_DESTROY_BY_RCU,
2085
	.twsk_prot		= &tcp6_timewait_sock_ops,
2086
	.rsk_prot		= &tcp6_request_sock_ops,
2087
	.h.hashinfo		= &tcp_hashinfo,
2088 2089 2090 2091
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2092 2093 2094 2095 2096
};

static struct inet6_protocol tcpv6_protocol = {
	.handler	=	tcp_v6_rcv,
	.err_handler	=	tcp_v6_err,
2097
	.gso_send_check	=	tcp_v6_gso_send_check,
H
Herbert Xu 已提交
2098
	.gso_segment	=	tcp_tso_segment,
H
Herbert Xu 已提交
2099 2100
	.gro_receive	=	tcp6_gro_receive,
	.gro_complete	=	tcp6_gro_complete,
L
Linus Torvalds 已提交
2101 2102 2103 2104 2105 2106 2107 2108 2109 2110
	.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,
2111 2112
	.flags		=	INET_PROTOSW_PERMANENT |
				INET_PROTOSW_ICSK,
L
Linus Torvalds 已提交
2113 2114
};

2115 2116
static int tcpv6_net_init(struct net *net)
{
2117 2118
	return inet_ctl_sock_create(&net->ipv6.tcp_sk, PF_INET6,
				    SOCK_RAW, IPPROTO_TCP, net);
2119 2120 2121 2122
}

static void tcpv6_net_exit(struct net *net)
{
2123
	inet_ctl_sock_destroy(net->ipv6.tcp_sk);
2124
	inet_twsk_purge(net, &tcp_hashinfo, &tcp_death_row, AF_INET6);
2125 2126 2127 2128 2129 2130 2131
}

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

2132
int __init tcpv6_init(void)
L
Linus Torvalds 已提交
2133
{
2134 2135 2136 2137 2138 2139
	int ret;

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

L
Linus Torvalds 已提交
2140
	/* register inet6 protocol */
2141 2142 2143 2144
	ret = inet6_register_protosw(&tcpv6_protosw);
	if (ret)
		goto out_tcpv6_protocol;

2145
	ret = register_pernet_subsys(&tcpv6_net_ops);
2146 2147 2148 2149
	if (ret)
		goto out_tcpv6_protosw;
out:
	return ret;
2150

2151 2152 2153 2154 2155 2156 2157
out_tcpv6_protocol:
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
out_tcpv6_protosw:
	inet6_unregister_protosw(&tcpv6_protosw);
	goto out;
}

2158
void tcpv6_exit(void)
2159
{
2160
	unregister_pernet_subsys(&tcpv6_net_ops);
2161 2162
	inet6_unregister_protosw(&tcpv6_protosw);
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
L
Linus Torvalds 已提交
2163
}