tcp_ipv6.c 55.8 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 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 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, NULL);
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
			goto failure;
		} else {
E
Eric Dumazet 已提交
229 230 231
			ipv6_addr_set_v4mapped(inet->inet_saddr, &np->saddr);
			ipv6_addr_set_v4mapped(inet->inet_rcv_saddr,
					       &np->rcv_saddr);
L
Linus Torvalds 已提交
232 233 234 235 236 237 238 239 240 241 242 243 244
		}

		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;
B
Brian Haley 已提交
245
	fl.mark = sk->sk_mark;
L
Linus Torvalds 已提交
246
	fl.fl_ip_dport = usin->sin6_port;
E
Eric Dumazet 已提交
247
	fl.fl_ip_sport = inet->inet_sport;
L
Linus Torvalds 已提交
248 249 250 251 252 253 254 255

	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

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

	/* set the source address */
	ipv6_addr_copy(&np->saddr, saddr);
E
Eric Dumazet 已提交
279
	inet->inet_rcv_saddr = LOOPBACK4_IPV6;
L
Linus Torvalds 已提交
280

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

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

E
Eric Dumazet 已提交
291
	inet->inet_dport = usin->sin6_port;
L
Linus Torvalds 已提交
292 293

	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
	if (err)
		goto late_failure;

	if (!tp->write_seq)
		tp->write_seq = secure_tcpv6_sequence_number(np->saddr.s6_addr32,
							     np->daddr.s6_addr32,
E
Eric Dumazet 已提交
301 302
							     inet->inet_sport,
							     inet->inet_dport);
L
Linus Torvalds 已提交
303 304 305 306 307 308 309 310 311 312 313

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

	return 0;

late_failure:
	tcp_set_state(sk, TCP_CLOSE);
	__sk_dst_reset(sk);
failure:
E
Eric Dumazet 已提交
314
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
315 316 317 318 319
	sk->sk_route_caps = 0;
	return err;
}

static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
320
		u8 type, u8 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
		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;
B
Brian Haley 已提交
386
			fl.mark = sk->sk_mark;
E
Eric Dumazet 已提交
387 388
			fl.fl_ip_dport = inet->inet_dport;
			fl.fl_ip_sport = inet->inet_sport;
V
Venkat Yekkirala 已提交
389
			security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
390 391 392 393 394 395

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

A
Alexey Dobriyan 已提交
396
			if ((err = xfrm_lookup(net, &dst, &fl, sk, 0)) < 0) {
L
Linus Torvalds 已提交
397 398 399 400 401 402 403
				sk->sk_err_soft = -err;
				goto out;
			}

		} else
			dst_hold(dst);

404
		if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
L
Linus Torvalds 已提交
405 406 407 408 409 410 411 412 413
			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);

414
	/* Might be for an request_sock */
L
Linus Torvalds 已提交
415
	switch (sk->sk_state) {
416
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
417 418 419 420
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

421 422
		req = inet6_csk_search_req(sk, &prev, th->dest, &hdr->daddr,
					   &hdr->saddr, inet6_iif(skb));
L
Linus Torvalds 已提交
423 424 425 426 427 428
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		 * an established socket here.
		 */
429
		WARN_ON(req->sk != NULL);
L
Linus Torvalds 已提交
430

431
		if (seq != tcp_rsk(req)->snt_isn) {
432
			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
433 434 435
			goto out;
		}

436
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
437 438 439 440
		goto out;

	case TCP_SYN_SENT:
	case TCP_SYN_RECV:  /* Cannot happen.
441
			       It can, it SYNs are crossed. --ANK */
L
Linus Torvalds 已提交
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
		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);
}


464 465
static int tcp_v6_send_synack(struct sock *sk, struct request_sock *req,
			      struct request_values *rvp)
L
Linus Torvalds 已提交
466
{
467
	struct inet6_request_sock *treq = inet6_rsk(req);
L
Linus Torvalds 已提交
468 469 470 471 472
	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;
473
	struct dst_entry *dst;
L
Linus Torvalds 已提交
474 475 476 477
	int err = -1;

	memset(&fl, 0, sizeof(fl));
	fl.proto = IPPROTO_TCP;
478 479
	ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
	ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
480
	fl.fl6_flowlabel = 0;
481
	fl.oif = treq->iif;
B
Brian Haley 已提交
482
	fl.mark = sk->sk_mark;
483
	fl.fl_ip_dport = inet_rsk(req)->rmt_port;
484
	fl.fl_ip_sport = inet_rsk(req)->loc_port;
485
	security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
486

487 488 489 490 491 492
	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
493 494
	}

495 496 497 498 499
	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
500
	if ((err = xfrm_lookup(sock_net(sk), &dst, &fl, sk, 0)) < 0)
501 502
		goto done;

503
	skb = tcp_make_synack(sk, dst, req, rvp);
L
Linus Torvalds 已提交
504
	if (skb) {
505
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
506

H
Herbert Xu 已提交
507
		th->check = tcp_v6_check(skb->len,
508
					 &treq->loc_addr, &treq->rmt_addr,
509
					 csum_partial(th, skb->len, skb->csum));
L
Linus Torvalds 已提交
510

511
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
512
		err = ip6_xmit(sk, skb, &fl, opt, 0);
513
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
514 515 516
	}

done:
517
	if (opt && opt != np->opt)
L
Linus Torvalds 已提交
518
		sock_kfree_s(sk, opt, opt->tot_len);
519
	dst_release(dst);
L
Linus Torvalds 已提交
520 521 522
	return err;
}

523 524 525 526 527 528 529
static int tcp_v6_rtx_synack(struct sock *sk, struct request_sock *req,
			     struct request_values *rvp)
{
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_RETRANSSEGS);
	return tcp_v6_send_synack(sk, req, rvp);
}

530 531 532 533 534 535 536 537 538 539 540 541 542 543
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));
}

544
static void tcp_v6_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
545
{
546
	kfree_skb(inet6_rsk(req)->pktopts);
L
Linus Torvalds 已提交
547 548
}

549 550 551 552 553 554 555 556 557 558 559 560 561
#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++) {
562
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, addr))
563
			return &tp->md5sig_info->keys6[i].base;
564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583
	}
	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 */
584
	struct tcp_md5sig_key *key;
585 586 587
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp6_md5sig_key *keys;

588
	key = tcp_v6_md5_do_lookup(sk, peer);
589 590
	if (key) {
		/* modify existing entry - just update that one */
591 592 593
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
594 595 596 597 598 599 600 601
	} 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;
			}
602
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
603
		}
604
		if (tcp_alloc_md5sig_pool(sk) == NULL) {
605 606 607
			kfree(newkey);
			return -ENOMEM;
		}
608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629
		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);
630 631
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.key = newkey;
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.keylen = newkeylen;
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650

		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++) {
651
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, peer)) {
652
			/* Free the key */
653
			kfree(tp->md5sig_info->keys6[i].base.key);
654 655 656 657 658
			tp->md5sig_info->entries6--;

			if (tp->md5sig_info->entries6 == 0) {
				kfree(tp->md5sig_info->keys6);
				tp->md5sig_info->keys6 = NULL;
659
				tp->md5sig_info->alloced6 = 0;
660 661 662 663 664 665 666 667
			} 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]));
			}
668 669
			tcp_free_md5sig_pool();
			return 0;
670 671 672 673 674 675 676 677 678 679 680 681
		}
	}
	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++)
682
			kfree(tp->md5sig_info->keys6[i].base.key);
683 684 685 686 687 688 689 690 691 692
		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++)
693
			kfree(tp->md5sig_info->keys4[i].base.key);
694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721
		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
722
		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738
			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;
739
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
740 741
	}

A
Arnaldo Carvalho de Melo 已提交
742
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
743 744
	if (!newkey)
		return -ENOMEM;
B
Brian Haley 已提交
745
	if (ipv6_addr_v4mapped(&sin6->sin6_addr)) {
746 747 748 749 750 751
		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);
}

752 753 754
static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp,
					struct in6_addr *daddr,
					struct in6_addr *saddr, int nbytes)
755 756
{
	struct tcp6_pseudohdr *bp;
757
	struct scatterlist sg;
758

759 760 761 762
	bp = &hp->md5_blk.ip6;
	/* 1. TCP pseudo-header (RFC2460) */
	ipv6_addr_copy(&bp->saddr, saddr);
	ipv6_addr_copy(&bp->daddr, daddr);
763
	bp->protocol = cpu_to_be32(IPPROTO_TCP);
A
Adam Langley 已提交
764
	bp->len = cpu_to_be32(nbytes);
765

766 767 768
	sg_init_one(&sg, bp, sizeof(*bp));
	return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp));
}
769

770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790
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))
791 792 793 794
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
795

796 797 798 799
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
800
	return 1;
801 802
}

803 804 805
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)
806 807
{
	struct in6_addr *saddr, *daddr;
808 809 810
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;
	struct tcphdr *th = tcp_hdr(skb);
811 812 813 814

	if (sk) {
		saddr = &inet6_sk(sk)->saddr;
		daddr = &inet6_sk(sk)->daddr;
815
	} else if (req) {
816 817
		saddr = &inet6_rsk(req)->loc_addr;
		daddr = &inet6_rsk(req)->rmt_addr;
818 819 820 821
	} else {
		struct ipv6hdr *ip6h = ipv6_hdr(skb);
		saddr = &ip6h->saddr;
		daddr = &ip6h->daddr;
822
	}
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850

	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;
851 852 853 854 855 856
}

static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb)
{
	__u8 *hash_location = NULL;
	struct tcp_md5sig_key *hash_expected;
857
	struct ipv6hdr *ip6h = ipv6_hdr(skb);
858
	struct tcphdr *th = tcp_hdr(skb);
859 860 861 862
	int genhash;
	u8 newhash[16];

	hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr);
863
	hash_location = tcp_parse_md5sig_option(th);
864

865 866 867 868 869 870
	/* 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);
871 872 873
		return 1;
	}

874 875
	if (!hash_expected && hash_location) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
876 877 878 879
		return 1;
	}

	/* check the signature */
880 881 882 883
	genhash = tcp_v6_md5_hash_skb(newhash,
				      hash_expected,
				      NULL, NULL, skb);

884 885
	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
886
			printk(KERN_INFO "MD5 Hash %s for [%pI6c]:%u->[%pI6c]:%u\n",
887
			       genhash ? "failed" : "mismatch",
888 889
			       &ip6h->saddr, ntohs(th->source),
			       &ip6h->daddr, ntohs(th->dest));
890 891 892 893 894 895 896
		}
		return 1;
	}
	return 0;
}
#endif

897
struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
898
	.family		=	AF_INET6,
899
	.obj_size	=	sizeof(struct tcp6_request_sock),
900
	.rtx_syn_ack	=	tcp_v6_rtx_synack,
901 902
	.send_ack	=	tcp_v6_reqsk_send_ack,
	.destructor	=	tcp_v6_reqsk_destructor,
903 904
	.send_reset	=	tcp_v6_send_reset,
	.syn_ack_timeout = 	tcp_syn_ack_timeout,
L
Linus Torvalds 已提交
905 906
};

907
#ifdef CONFIG_TCP_MD5SIG
908
static const struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
909
	.md5_lookup	=	tcp_v6_reqsk_md5_lookup,
910
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
911
};
912
#endif
913

914 915 916
static struct timewait_sock_ops tcp6_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp6_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
917
	.twsk_destructor= tcp_twsk_destructor,
918 919
};

920
static void tcp_v6_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
921 922
{
	struct ipv6_pinfo *np = inet6_sk(sk);
923
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
924

925
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
L
Linus Torvalds 已提交
926
		th->check = ~csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,  0);
927
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
928
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
929
	} else {
930
		th->check = csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,
931
					    csum_partial(th, th->doff<<2,
L
Linus Torvalds 已提交
932 933 934 935
							 skb->csum));
	}
}

936 937 938 939 940 941 942 943
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;

944
	ipv6h = ipv6_hdr(skb);
945
	th = tcp_hdr(skb);
946 947 948 949

	th->check = 0;
	th->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr, skb->len,
				     IPPROTO_TCP, 0);
950
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
951
	skb->csum_offset = offsetof(struct tcphdr, check);
952
	skb->ip_summed = CHECKSUM_PARTIAL;
953 954
	return 0;
}
L
Linus Torvalds 已提交
955

H
Herbert Xu 已提交
956 957
static struct sk_buff **tcp6_gro_receive(struct sk_buff **head,
					 struct sk_buff *skb)
H
Herbert Xu 已提交
958
{
H
Herbert Xu 已提交
959
	struct ipv6hdr *iph = skb_gro_network_header(skb);
H
Herbert Xu 已提交
960 961 962

	switch (skb->ip_summed) {
	case CHECKSUM_COMPLETE:
963
		if (!tcp_v6_check(skb_gro_len(skb), &iph->saddr, &iph->daddr,
H
Herbert Xu 已提交
964 965 966 967 968 969 970 971 972 973 974 975 976 977
				  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
978
static int tcp6_gro_complete(struct sk_buff *skb)
H
Herbert Xu 已提交
979 980 981 982 983 984 985 986 987 988 989
{
	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);
}

990 991
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
992
{
993
	struct tcphdr *th = tcp_hdr(skb), *t1;
L
Linus Torvalds 已提交
994 995
	struct sk_buff *buff;
	struct flowi fl;
E
Eric Dumazet 已提交
996
	struct net *net = dev_net(skb_dst(skb)->dev);
997
	struct sock *ctl_sk = net->ipv6.tcp_sk;
998
	unsigned int tot_len = sizeof(struct tcphdr);
E
Eric Dumazet 已提交
999
	struct dst_entry *dst;
1000
	__be32 *topt;
L
Linus Torvalds 已提交
1001

1002 1003
	if (ts)
		tot_len += TCPOLEN_TSTAMP_ALIGNED;
1004 1005 1006 1007 1008 1009
#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
1010
			 GFP_ATOMIC);
1011 1012
	if (buff == NULL)
		return;
L
Linus Torvalds 已提交
1013

1014
	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
L
Linus Torvalds 已提交
1015

1016
	t1 = (struct tcphdr *) skb_push(buff, tot_len);
1017
	skb_reset_transport_header(skb);
L
Linus Torvalds 已提交
1018 1019 1020 1021 1022

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
1023
	t1->doff = tot_len / 4;
1024 1025 1026 1027 1028
	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
1029

1030 1031
	topt = (__be32 *)(t1 + 1);

1032 1033 1034 1035 1036 1037 1038
	if (ts) {
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
		*topt++ = htonl(tcp_time_stamp);
		*topt++ = htonl(ts);
	}

1039 1040
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
1041 1042 1043
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG);
		tcp_v6_md5_hash_hdr((__u8 *)topt, key,
1044 1045
				    &ipv6_hdr(skb)->saddr,
				    &ipv6_hdr(skb)->daddr, t1);
1046 1047 1048
	}
#endif

1049
	buff->csum = csum_partial(t1, tot_len, 0);
L
Linus Torvalds 已提交
1050 1051

	memset(&fl, 0, sizeof(fl));
1052 1053
	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
1054 1055

	t1->check = csum_ipv6_magic(&fl.fl6_src, &fl.fl6_dst,
1056
				    tot_len, IPPROTO_TCP,
L
Linus Torvalds 已提交
1057 1058 1059
				    buff->csum);

	fl.proto = IPPROTO_TCP;
1060
	fl.oif = inet6_iif(skb);
L
Linus Torvalds 已提交
1061 1062
	fl.fl_ip_dport = t1->dest;
	fl.fl_ip_sport = t1->source;
V
Venkat Yekkirala 已提交
1063
	security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
1064

1065 1066 1067 1068
	/* 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
1069 1070 1071
	if (!ip6_dst_lookup(ctl_sk, &dst, &fl)) {
		if (xfrm_lookup(net, &dst, &fl, NULL, 0) >= 0) {
			skb_dst_set(buff, dst);
1072
			ip6_xmit(ctl_sk, buff, &fl, NULL, 0);
1073
			TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
1074 1075
			if (rst)
				TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
1076
			return;
A
Arnaldo Carvalho de Melo 已提交
1077
		}
L
Linus Torvalds 已提交
1078 1079 1080 1081 1082
	}

	kfree_skb(buff);
}

1083
static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
1084
{
1085 1086
	struct tcphdr *th = tcp_hdr(skb);
	u32 seq = 0, ack_seq = 0;
1087
	struct tcp_md5sig_key *key = NULL;
L
Linus Torvalds 已提交
1088

1089
	if (th->rst)
L
Linus Torvalds 已提交
1090 1091
		return;

1092 1093
	if (!ipv6_unicast_destination(skb))
		return;
L
Linus Torvalds 已提交
1094

1095
#ifdef CONFIG_TCP_MD5SIG
1096 1097
	if (sk)
		key = tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr);
1098 1099
#endif

1100 1101 1102 1103 1104
	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
1105

1106 1107
	tcp_v6_send_response(skb, seq, ack_seq, 0, 0, key, 1);
}
L
Linus Torvalds 已提交
1108

1109 1110 1111 1112
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
1113 1114 1115 1116
}

static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
1117
	struct inet_timewait_sock *tw = inet_twsk(sk);
1118
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
1119

1120
	tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
1121
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
1122
			tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw));
L
Linus Torvalds 已提交
1123

1124
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
1125 1126
}

1127 1128
static void tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
				  struct request_sock *req)
L
Linus Torvalds 已提交
1129
{
1130
	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent,
1131
			tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr));
L
Linus Torvalds 已提交
1132 1133 1134 1135 1136
}


static struct sock *tcp_v6_hnd_req(struct sock *sk,struct sk_buff *skb)
{
1137
	struct request_sock *req, **prev;
1138
	const struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1139 1140 1141
	struct sock *nsk;

	/* Find possible connection requests. */
1142
	req = inet6_csk_search_req(sk, &prev, th->source,
1143 1144
				   &ipv6_hdr(skb)->saddr,
				   &ipv6_hdr(skb)->daddr, inet6_iif(skb));
L
Linus Torvalds 已提交
1145 1146 1147
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1148
	nsk = __inet6_lookup_established(sock_net(sk), &tcp_hashinfo,
1149 1150
			&ipv6_hdr(skb)->saddr, th->source,
			&ipv6_hdr(skb)->daddr, ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1157
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1158 1159 1160
		return NULL;
	}

1161
#ifdef CONFIG_SYN_COOKIES
L
Linus Torvalds 已提交
1162
	if (!th->rst && !th->syn && th->ack)
1163
		sk = cookie_v6_check(sk, skb);
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169 1170 1171 1172
#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)
{
1173
	struct tcp_extend_values tmp_ext;
1174
	struct tcp_options_received tmp_opt;
1175
	u8 *hash_location;
1176
	struct request_sock *req;
1177
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1178 1179
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
1180
	__u32 isn = TCP_SKB_CB(skb)->when;
1181 1182 1183 1184 1185
#ifdef CONFIG_SYN_COOKIES
	int want_cookie = 0;
#else
#define want_cookie 0
#endif
L
Linus Torvalds 已提交
1186 1187 1188 1189 1190

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

	if (!ipv6_unicast_destination(skb))
1191
		goto drop;
L
Linus Torvalds 已提交
1192

1193
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1194
		if (net_ratelimit())
1195 1196 1197 1198 1199 1200
			syn_flood_warning(skb);
#ifdef CONFIG_SYN_COOKIES
		if (sysctl_tcp_syncookies)
			want_cookie = 1;
		else
#endif
1201
		goto drop;
L
Linus Torvalds 已提交
1202 1203
	}

1204
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1205 1206
		goto drop;

1207
	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
L
Linus Torvalds 已提交
1208 1209 1210
	if (req == NULL)
		goto drop;

1211 1212 1213 1214
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops;
#endif

L
Linus Torvalds 已提交
1215 1216 1217
	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;
1218
	tcp_parse_options(skb, &tmp_opt, &hash_location, 0);
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249

	if (tmp_opt.cookie_plus > 0 &&
	    tmp_opt.saw_tstamp &&
	    !tp->rx_opt.cookie_out_never &&
	    (sysctl_tcp_cookie_size > 0 ||
	     (tp->cookie_values != NULL &&
	      tp->cookie_values->cookie_desired > 0))) {
		u8 *c;
		u32 *d;
		u32 *mess = &tmp_ext.cookie_bakery[COOKIE_DIGEST_WORDS];
		int l = tmp_opt.cookie_plus - TCPOLEN_COOKIE_BASE;

		if (tcp_cookie_generator(&tmp_ext.cookie_bakery[0]) != 0)
			goto drop_and_free;

		/* Secret recipe starts with IP addresses */
		d = &ipv6_hdr(skb)->daddr.s6_addr32[0];
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		d = &ipv6_hdr(skb)->saddr.s6_addr32[0];
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		*mess++ ^= *d++;

		/* plus variable length Initiator Cookie */
		c = (u8 *)mess;
		while (l-- > 0)
			*c++ ^= *hash_location++;
L
Linus Torvalds 已提交
1250

1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
#ifdef CONFIG_SYN_COOKIES
		want_cookie = 0;	/* not our kind of cookie */
#endif
		tmp_ext.cookie_out_never = 0; /* false */
		tmp_ext.cookie_plus = tmp_opt.cookie_plus;
	} else if (!tp->rx_opt.cookie_in_always) {
		/* redundant indications, but ensure initialization. */
		tmp_ext.cookie_out_never = 1; /* true */
		tmp_ext.cookie_plus = 0;
	} else {
		goto drop_and_free;
	}
	tmp_ext.cookie_in_always = tp->rx_opt.cookie_in_always;
L
Linus Torvalds 已提交
1264

1265
	if (want_cookie && !tmp_opt.saw_tstamp)
1266 1267
		tcp_clear_options(&tmp_opt);

L
Linus Torvalds 已提交
1268 1269 1270
	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
	tcp_openreq_init(req, &tmp_opt, skb);

1271
	treq = inet6_rsk(req);
1272 1273
	ipv6_addr_copy(&treq->rmt_addr, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&treq->loc_addr, &ipv6_hdr(skb)->daddr);
1274 1275 1276 1277 1278
	if (!want_cookie)
		TCP_ECN_create_request(req, tcp_hdr(skb));

	if (want_cookie) {
		isn = cookie_v6_init_sequence(sk, skb, &req->mss);
1279
		req->cookie_ts = tmp_opt.tstamp_ok;
1280 1281 1282 1283 1284 1285 1286 1287
	} 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
1288

1289 1290 1291 1292
		/* 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
1293

1294
		isn = tcp_v6_init_sequence(skb);
1295
	}
1296
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1297

1298 1299
	security_inet_conn_request(sk, skb, req);

1300 1301 1302
	if (tcp_v6_send_synack(sk, req,
			       (struct request_values *)&tmp_ext) ||
	    want_cookie)
1303
		goto drop_and_free;
L
Linus Torvalds 已提交
1304

1305 1306
	inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
	return 0;
L
Linus Torvalds 已提交
1307

1308 1309
drop_and_free:
	reqsk_free(req);
L
Linus Torvalds 已提交
1310 1311 1312 1313 1314
drop:
	return 0; /* don't send reset */
}

static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
1315
					  struct request_sock *req,
L
Linus Torvalds 已提交
1316 1317
					  struct dst_entry *dst)
{
1318
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324
	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;
1325 1326 1327
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1328 1329 1330 1331 1332 1333 1334 1335

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

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

1336
		if (newsk == NULL)
L
Linus Torvalds 已提交
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
			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));

E
Eric Dumazet 已提交
1348
		ipv6_addr_set_v4mapped(newinet->inet_daddr, &newnp->daddr);
L
Linus Torvalds 已提交
1349

E
Eric Dumazet 已提交
1350
		ipv6_addr_set_v4mapped(newinet->inet_saddr, &newnp->saddr);
L
Linus Torvalds 已提交
1351 1352 1353

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

1354
		inet_csk(newsk)->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
1355
		newsk->sk_backlog_rcv = tcp_v4_do_rcv;
1356 1357 1358 1359
#ifdef CONFIG_TCP_MD5SIG
		newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif

L
Linus Torvalds 已提交
1360 1361
		newnp->pktoptions  = NULL;
		newnp->opt	   = NULL;
1362
		newnp->mcast_oif   = inet6_iif(skb);
1363
		newnp->mcast_hops  = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
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 1371
		 */

		/* It is tricky place. Until this moment IPv4 tcp
1372
		   worked with IPv6 icsk.icsk_af_ops.
L
Linus Torvalds 已提交
1373 1374
		   Sync it now.
		 */
1375
		tcp_sync_mss(newsk, inet_csk(newsk)->icsk_pmtu_cookie);
L
Linus Torvalds 已提交
1376 1377 1378 1379

		return newsk;
	}

1380
	treq = inet6_rsk(req);
L
Linus Torvalds 已提交
1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391
	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;
1392
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397 1398
		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;
		}
1399
		ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
1400
		fl.oif = sk->sk_bound_dev_if;
B
Brian Haley 已提交
1401
		fl.mark = sk->sk_mark;
1402
		fl.fl_ip_dport = inet_rsk(req)->rmt_port;
1403
		fl.fl_ip_sport = inet_rsk(req)->loc_port;
1404
		security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
1405 1406 1407 1408 1409 1410 1411

		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
1412
		if ((xfrm_lookup(sock_net(sk), &dst, &fl, sk, 0)) < 0)
L
Linus Torvalds 已提交
1413
			goto out;
1414
	}
L
Linus Torvalds 已提交
1415 1416 1417 1418 1419

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

1420 1421 1422 1423 1424
	/*
	 * 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
1425

1426
	newsk->sk_gso_type = SKB_GSO_TCPV6;
1427
	__ip6_dst_store(newsk, dst, NULL, NULL);
L
Linus Torvalds 已提交
1428 1429 1430 1431 1432 1433 1434 1435 1436 1437

	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));

1438 1439 1440 1441
	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
1442

1443
	/* Now IPv6 options...
L
Linus Torvalds 已提交
1444 1445 1446 1447

	   First: no IPv4 options.
	 */
	newinet->opt = NULL;
1448
	newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1449 1450 1451 1452 1453 1454

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

	/* Clone pktoptions received with SYN */
	newnp->pktoptions = NULL;
1455 1456 1457 1458
	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
1459 1460 1461 1462
		if (newnp->pktoptions)
			skb_set_owner_r(newnp->pktoptions, newsk);
	}
	newnp->opt	  = NULL;
1463
	newnp->mcast_oif  = inet6_iif(skb);
1464
	newnp->mcast_hops = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477

	/* 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);
	}

1478
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1479
	if (newnp->opt)
1480 1481
		inet_csk(newsk)->icsk_ext_hdr_len = (newnp->opt->opt_nflen +
						     newnp->opt->opt_flen);
L
Linus Torvalds 已提交
1482

J
John Heffner 已提交
1483
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1484 1485 1486 1487
	tcp_sync_mss(newsk, dst_mtu(dst));
	newtp->advmss = dst_metric(dst, RTAX_ADVMSS);
	tcp_initialize_rcv_mss(newsk);

E
Eric Dumazet 已提交
1488 1489
	newinet->inet_daddr = newinet->inet_saddr = LOOPBACK4_IPV6;
	newinet->inet_rcv_saddr = LOOPBACK4_IPV6;
L
Linus Torvalds 已提交
1490

1491 1492 1493 1494 1495 1496 1497 1498
#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
1499 1500
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1501
			tcp_v6_md5_do_add(newsk, &newnp->daddr,
1502 1503 1504 1505
					  newkey, key->keylen);
	}
#endif

1506
	__inet6_hash(newsk, NULL);
1507
	__inet_inherit_port(sk, newsk);
L
Linus Torvalds 已提交
1508 1509 1510 1511

	return newsk;

out_overflow:
1512
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
L
Linus Torvalds 已提交
1513
out:
1514
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520
	if (opt && opt != np->opt)
		sock_kfree_s(sk, opt, opt->tot_len);
	dst_release(dst);
	return NULL;
}

1521
static __sum16 tcp_v6_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1522
{
1523
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
H
Herbert Xu 已提交
1524
		if (!tcp_v6_check(skb->len, &ipv6_hdr(skb)->saddr,
1525
				  &ipv6_hdr(skb)->daddr, skb->csum)) {
1526
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1527
			return 0;
1528
		}
L
Linus Torvalds 已提交
1529
	}
1530

H
Herbert Xu 已提交
1531
	skb->csum = ~csum_unfold(tcp_v6_check(skb->len,
1532 1533
					      &ipv6_hdr(skb)->saddr,
					      &ipv6_hdr(skb)->daddr, 0));
1534

L
Linus Torvalds 已提交
1535
	if (skb->len <= 76) {
1536
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
	}
	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);

1566 1567 1568 1569 1570
#ifdef CONFIG_TCP_MD5SIG
	if (tcp_v6_inbound_md5_hash (sk, skb))
		goto discard;
#endif

1571
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589
		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.
1590
					       --ANK (980728)
L
Linus Torvalds 已提交
1591 1592 1593 1594 1595 1596
	 */
	if (np->rxopt.all)
		opt_skb = skb_clone(skb, GFP_ATOMIC);

	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1597
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1598 1599 1600 1601 1602 1603 1604
			goto reset;
		TCP_CHECK_TIMER(sk);
		if (opt_skb)
			goto ipv6_pktoptions;
		return 0;
	}

1605
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1606 1607
		goto csum_err;

1608
	if (sk->sk_state == TCP_LISTEN) {
L
Linus Torvalds 已提交
1609 1610 1611 1612 1613 1614 1615 1616 1617
		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..
		 */
1618
		if(nsk != sk) {
L
Linus Torvalds 已提交
1619 1620 1621 1622 1623 1624 1625 1626 1627
			if (tcp_child_process(sk, nsk, skb))
				goto reset;
			if (opt_skb)
				__kfree_skb(opt_skb);
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1628
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1629 1630 1631 1632 1633 1634 1635
		goto reset;
	TCP_CHECK_TIMER(sk);
	if (opt_skb)
		goto ipv6_pktoptions;
	return 0;

reset:
1636
	tcp_v6_send_reset(sk, skb);
L
Linus Torvalds 已提交
1637 1638 1639 1640 1641 1642
discard:
	if (opt_skb)
		__kfree_skb(opt_skb);
	kfree_skb(skb);
	return 0;
csum_err:
1643
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
	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))) {
1658
		if (np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo)
1659
			np->mcast_oif = inet6_iif(opt_skb);
1660
		if (np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim)
1661
			np->mcast_hops = ipv6_hdr(opt_skb)->hop_limit;
L
Linus Torvalds 已提交
1662 1663 1664 1665 1666 1667 1668 1669 1670
		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);
		}
	}

1671
	kfree_skb(opt_skb);
L
Linus Torvalds 已提交
1672 1673 1674
	return 0;
}

1675
static int tcp_v6_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
1676
{
1677
	struct tcphdr *th;
L
Linus Torvalds 已提交
1678 1679
	struct sock *sk;
	int ret;
1680
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1681 1682 1683 1684 1685 1686 1687

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

	/*
	 *	Count it even if it's bad.
	 */
1688
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1689 1690 1691 1692

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

1693
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1694 1695 1696 1697 1698 1699

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

1700
	if (!skb_csum_unnecessary(skb) && tcp_v6_checksum_init(skb))
L
Linus Torvalds 已提交
1701 1702
		goto bad_packet;

1703
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1704 1705 1706 1707 1708
	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;
1709
	TCP_SKB_CB(skb)->flags = ipv6_get_dsfield(ipv6_hdr(skb));
L
Linus Torvalds 已提交
1710 1711
	TCP_SKB_CB(skb)->sacked = 0;

1712
	sk = __inet6_lookup_skb(&tcp_hashinfo, skb, th->source, th->dest);
L
Linus Torvalds 已提交
1713 1714 1715 1716 1717 1718 1719 1720 1721 1722
	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;

1723
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1724 1725 1726 1727
		goto discard_and_relse;

	skb->dev = NULL;

1728
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1729 1730
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1731
#ifdef CONFIG_NET_DMA
1732
		struct tcp_sock *tp = tcp_sk(sk);
1733
		if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
1734
			tp->ucopy.dma_chan = dma_find_channel(DMA_MEMCPY);
1735 1736 1737
		if (tp->ucopy.dma_chan)
			ret = tcp_v6_do_rcv(sk, skb);
		else
C
Chris Leech 已提交
1738 1739 1740 1741 1742
#endif
		{
			if (!tcp_prequeue(sk, skb))
				ret = tcp_v6_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755
	} 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:
1756
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1757
	} else {
1758
		tcp_v6_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
	}

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)) {
1776
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1777 1778 1779 1780
		goto discard_it;
	}

	if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) {
1781
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1782
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1783 1784 1785
		goto discard_it;
	}

1786
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1787 1788 1789 1790
	case TCP_TW_SYN:
	{
		struct sock *sk2;

1791
		sk2 = inet6_lookup_listener(dev_net(skb->dev), &tcp_hashinfo,
1792
					    &ipv6_hdr(skb)->daddr,
1793
					    ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1794
		if (sk2 != NULL) {
1795 1796 1797
			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
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818
			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;
}

1819
static const struct inet_connection_sock_af_ops ipv6_specific = {
1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830
	.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),
1831
	.bind_conflict	   = inet6_csk_bind_conflict,
1832
#ifdef CONFIG_COMPAT
1833 1834
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1835
#endif
L
Linus Torvalds 已提交
1836 1837
};

1838
#ifdef CONFIG_TCP_MD5SIG
1839
static const struct tcp_sock_af_ops tcp_sock_ipv6_specific = {
1840
	.md5_lookup	=	tcp_v6_md5_lookup,
1841
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
1842 1843 1844
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1845
#endif
1846

L
Linus Torvalds 已提交
1847 1848 1849 1850
/*
 *	TCP over IPv4 via INET6 API
 */

1851
static const struct inet_connection_sock_af_ops ipv6_mapped = {
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862
	.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),
1863
	.bind_conflict	   = inet6_csk_bind_conflict,
1864
#ifdef CONFIG_COMPAT
1865 1866
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1867
#endif
L
Linus Torvalds 已提交
1868 1869
};

1870
#ifdef CONFIG_TCP_MD5SIG
1871
static const struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = {
1872
	.md5_lookup	=	tcp_v4_md5_lookup,
1873
	.calc_md5_hash	=	tcp_v4_md5_hash_skb,
1874 1875 1876
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1877
#endif
1878

L
Linus Torvalds 已提交
1879 1880 1881 1882 1883
/* 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)
{
1884
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1885 1886 1887 1888 1889 1890
	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);

1891
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903
	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.
	 */
I
Ilpo Järvinen 已提交
1904
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
L
Linus Torvalds 已提交
1905
	tp->snd_cwnd_clamp = ~0;
1906
	tp->mss_cache = TCP_MSS_DEFAULT;
L
Linus Torvalds 已提交
1907 1908 1909 1910 1911

	tp->reordering = sysctl_tcp_reordering;

	sk->sk_state = TCP_CLOSE;

1912
	icsk->icsk_af_ops = &ipv6_specific;
1913
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
1914
	icsk->icsk_sync_mss = tcp_sync_mss;
L
Linus Torvalds 已提交
1915 1916 1917
	sk->sk_write_space = sk_stream_write_space;
	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);

1918 1919 1920 1921
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv6_specific;
#endif

1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
	/* TCP Cookie Transactions */
	if (sysctl_tcp_cookie_size > 0) {
		/* Default, cookies without s_data_payload. */
		tp->cookie_values =
			kzalloc(sizeof(*tp->cookie_values),
				sk->sk_allocation);
		if (tp->cookie_values != NULL)
			kref_init(&tp->cookie_values->kref);
	}
	/* Presumed zeroed, in order of appearance:
	 *	cookie_in_always, cookie_out_never,
	 *	s_data_constant, s_data_in, s_data_out
	 */
L
Linus Torvalds 已提交
1935 1936 1937
	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

H
Herbert Xu 已提交
1938
	local_bh_disable();
1939
	percpu_counter_inc(&tcp_sockets_allocated);
H
Herbert Xu 已提交
1940
	local_bh_enable();
L
Linus Torvalds 已提交
1941 1942 1943 1944

	return 0;
}

1945
static void tcp_v6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1946
{
1947 1948 1949 1950 1951
#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
1952
	tcp_v4_destroy_sock(sk);
1953
	inet6_destroy_sock(sk);
L
Linus Torvalds 已提交
1954 1955
}

1956
#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1957
/* Proc filesystem TCPv6 sock list dumping. */
1958
static void get_openreq6(struct seq_file *seq,
1959
			 struct sock *sk, struct request_sock *req, int i, int uid)
L
Linus Torvalds 已提交
1960 1961
{
	int ttd = req->expires - jiffies;
1962 1963
	struct in6_addr *src = &inet6_rsk(req)->loc_addr;
	struct in6_addr *dest = &inet6_rsk(req)->rmt_addr;
L
Linus Torvalds 已提交
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973

	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],
1974
		   ntohs(inet_rsk(req)->loc_port),
L
Linus Torvalds 已提交
1975 1976
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3],
1977
		   ntohs(inet_rsk(req)->rmt_port),
L
Linus Torvalds 已提交
1978 1979
		   TCP_SYN_RECV,
		   0,0, /* could print option size, but that is af dependent. */
1980 1981
		   1,   /* timers active (only the expire timer) */
		   jiffies_to_clock_t(ttd),
L
Linus Torvalds 已提交
1982 1983
		   req->retrans,
		   uid,
1984
		   0,  /* non standard timer */
L
Linus Torvalds 已提交
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
		   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);
1997
	const struct inet_connection_sock *icsk = inet_csk(sp);
L
Linus Torvalds 已提交
1998 1999 2000 2001
	struct ipv6_pinfo *np = inet6_sk(sp);

	dest  = &np->daddr;
	src   = &np->rcv_saddr;
E
Eric Dumazet 已提交
2002 2003
	destp = ntohs(inet->inet_dport);
	srcp  = ntohs(inet->inet_sport);
2004 2005

	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
2006
		timer_active	= 1;
2007 2008
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
2009
		timer_active	= 4;
2010
		timer_expires	= icsk->icsk_timeout;
L
Linus Torvalds 已提交
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
	} 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 "
2021
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %lu %lu %u %u %d\n",
L
Linus Torvalds 已提交
2022 2023 2024 2025 2026
		   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,
2027
		   sp->sk_state,
2028 2029
		   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
2030 2031
		   timer_active,
		   jiffies_to_clock_t(timer_expires - jiffies),
2032
		   icsk->icsk_retransmits,
L
Linus Torvalds 已提交
2033
		   sock_i_uid(sp),
2034
		   icsk->icsk_probes_out,
L
Linus Torvalds 已提交
2035 2036
		   sock_i_ino(sp),
		   atomic_read(&sp->sk_refcnt), sp,
2037 2038
		   jiffies_to_clock_t(icsk->icsk_rto),
		   jiffies_to_clock_t(icsk->icsk_ack.ato),
2039
		   (icsk->icsk_ack.quick << 1 ) | icsk->icsk_ack.pingpong,
I
Ilpo Järvinen 已提交
2040 2041
		   tp->snd_cwnd,
		   tcp_in_initial_slowstart(tp) ? -1 : tp->snd_ssthresh
L
Linus Torvalds 已提交
2042 2043 2044
		   );
}

2045
static void get_timewait6_sock(struct seq_file *seq,
2046
			       struct inet_timewait_sock *tw, int i)
L
Linus Torvalds 已提交
2047 2048 2049
{
	struct in6_addr *dest, *src;
	__u16 destp, srcp;
2050
	struct inet6_timewait_sock *tw6 = inet6_twsk((struct sock *)tw);
L
Linus Torvalds 已提交
2051 2052 2053 2054 2055
	int ttd = tw->tw_ttd - jiffies;

	if (ttd < 0)
		ttd = 0;

2056 2057
	dest = &tw6->tw_v6_daddr;
	src  = &tw6->tw_v6_rcv_saddr;
L
Linus Torvalds 已提交
2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107
	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,
2108 2109 2110
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2111 2112 2113
	.seq_ops	= {
		.show		= tcp6_seq_show,
	},
L
Linus Torvalds 已提交
2114 2115
};

2116
int __net_init tcp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
2117
{
2118
	return tcp_proc_register(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2119 2120
}

2121
void tcp6_proc_exit(struct net *net)
L
Linus Torvalds 已提交
2122
{
2123
	tcp_proc_unregister(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2124 2125 2126 2127 2128 2129 2130 2131 2132
}
#endif

struct proto tcpv6_prot = {
	.name			= "TCPv6",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v6_connect,
	.disconnect		= tcp_disconnect,
2133
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2134 2135 2136 2137 2138 2139 2140 2141 2142
	.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,
2143 2144
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2145 2146 2147 2148
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
2149
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2150 2151 2152 2153 2154
	.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),
2155
	.slab_flags		= SLAB_DESTROY_BY_RCU,
2156
	.twsk_prot		= &tcp6_timewait_sock_ops,
2157
	.rsk_prot		= &tcp6_request_sock_ops,
2158
	.h.hashinfo		= &tcp_hashinfo,
2159 2160 2161 2162
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2163 2164
};

2165
static const struct inet6_protocol tcpv6_protocol = {
L
Linus Torvalds 已提交
2166 2167
	.handler	=	tcp_v6_rcv,
	.err_handler	=	tcp_v6_err,
2168
	.gso_send_check	=	tcp_v6_gso_send_check,
H
Herbert Xu 已提交
2169
	.gso_segment	=	tcp_tso_segment,
H
Herbert Xu 已提交
2170 2171
	.gro_receive	=	tcp6_gro_receive,
	.gro_complete	=	tcp6_gro_complete,
L
Linus Torvalds 已提交
2172 2173 2174 2175 2176 2177 2178 2179 2180
	.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,
	.no_check	=	0,
2181 2182
	.flags		=	INET_PROTOSW_PERMANENT |
				INET_PROTOSW_ICSK,
L
Linus Torvalds 已提交
2183 2184
};

2185
static int __net_init tcpv6_net_init(struct net *net)
2186
{
2187 2188
	return inet_ctl_sock_create(&net->ipv6.tcp_sk, PF_INET6,
				    SOCK_RAW, IPPROTO_TCP, net);
2189 2190
}

2191
static void __net_exit tcpv6_net_exit(struct net *net)
2192
{
2193
	inet_ctl_sock_destroy(net->ipv6.tcp_sk);
E
Eric W. Biederman 已提交
2194 2195
}

2196
static void __net_exit tcpv6_net_exit_batch(struct list_head *net_exit_list)
E
Eric W. Biederman 已提交
2197 2198
{
	inet_twsk_purge(&tcp_hashinfo, &tcp_death_row, AF_INET6);
2199 2200 2201
}

static struct pernet_operations tcpv6_net_ops = {
E
Eric W. Biederman 已提交
2202 2203 2204
	.init	    = tcpv6_net_init,
	.exit	    = tcpv6_net_exit,
	.exit_batch = tcpv6_net_exit_batch,
2205 2206
};

2207
int __init tcpv6_init(void)
L
Linus Torvalds 已提交
2208
{
2209 2210 2211 2212 2213 2214
	int ret;

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

L
Linus Torvalds 已提交
2215
	/* register inet6 protocol */
2216 2217 2218 2219
	ret = inet6_register_protosw(&tcpv6_protosw);
	if (ret)
		goto out_tcpv6_protocol;

2220
	ret = register_pernet_subsys(&tcpv6_net_ops);
2221 2222 2223 2224
	if (ret)
		goto out_tcpv6_protosw;
out:
	return ret;
2225

2226 2227 2228 2229 2230 2231 2232
out_tcpv6_protocol:
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
out_tcpv6_protosw:
	inet6_unregister_protosw(&tcpv6_protosw);
	goto out;
}

2233
void tcpv6_exit(void)
2234
{
2235
	unregister_pernet_subsys(&tcpv6_net_ops);
2236 2237
	inet6_unregister_protosw(&tcpv6_protosw);
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
L
Linus Torvalds 已提交
2238
}