tcp_ipv6.c 58.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	TCP over IPv6
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7
 *
8
 *	Based on:
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
 *	linux/net/ipv4/tcp.c
 *	linux/net/ipv4/tcp_input.c
 *	linux/net/ipv4/tcp_output.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *	YOSHIFUJI Hideaki @USAGI:	convert /proc/net/tcp6 to seq_file.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

H
Herbert Xu 已提交
26
#include <linux/bottom_half.h>
L
Linus Torvalds 已提交
27 28 29 30 31 32 33 34 35 36 37 38 39 40
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/jiffies.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <linux/netdevice.h>
#include <linux/init.h>
#include <linux/jhash.h>
#include <linux/ipsec.h>
#include <linux/times.h>
41
#include <linux/slab.h>
L
Linus Torvalds 已提交
42 43 44 45 46 47 48

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

#include <net/tcp.h>
#include <net/ndisc.h>
49
#include <net/inet6_hashtables.h>
50
#include <net/inet6_connection_sock.h>
L
Linus Torvalds 已提交
51 52 53 54 55 56 57 58 59 60
#include <net/ipv6.h>
#include <net/transp_v6.h>
#include <net/addrconf.h>
#include <net/ip6_route.h>
#include <net/ip6_checksum.h>
#include <net/inet_ecn.h>
#include <net/protocol.h>
#include <net/xfrm.h>
#include <net/snmp.h>
#include <net/dsfield.h>
61
#include <net/timewait_sock.h>
62
#include <net/netdma.h>
63
#include <net/inet_common.h>
L
Linus Torvalds 已提交
64 65 66 67 68 69

#include <asm/uaccess.h>

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

70 71 72 73
#include <linux/crypto.h>
#include <linux/scatterlist.h>

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

static int	tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb);
78 79 80
static void	__tcp_v6_send_check(struct sk_buff *skb,
				    struct in6_addr *saddr,
				    struct in6_addr *daddr);
L
Linus Torvalds 已提交
81

82 83
static const struct inet_connection_sock_af_ops ipv6_mapped;
static const struct inet_connection_sock_af_ops ipv6_specific;
84
#ifdef CONFIG_TCP_MD5SIG
85 86
static const struct tcp_sock_af_ops tcp_sock_ipv6_specific;
static const struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific;
87 88 89 90 91 92
#else
static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk,
						   struct in6_addr *addr)
{
	return NULL;
}
93
#endif
L
Linus Torvalds 已提交
94 95 96 97

static void tcp_v6_hash(struct sock *sk)
{
	if (sk->sk_state != TCP_CLOSE) {
98
		if (inet_csk(sk)->icsk_af_ops == &ipv6_mapped) {
L
Linus Torvalds 已提交
99 100 101 102
			tcp_prot.hash(sk);
			return;
		}
		local_bh_disable();
103
		__inet6_hash(sk, NULL);
L
Linus Torvalds 已提交
104 105 106 107
		local_bh_enable();
	}
}

H
Herbert Xu 已提交
108
static __inline__ __sum16 tcp_v6_check(int len,
109 110
				   struct in6_addr *saddr,
				   struct in6_addr *daddr,
111
				   __wsum base)
L
Linus Torvalds 已提交
112 113 114 115
{
	return csum_ipv6_magic(saddr, daddr, len, IPPROTO_TCP, base);
}

116
static __u32 tcp_v6_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
117
{
118 119
	return secure_tcpv6_sequence_number(ipv6_hdr(skb)->daddr.s6_addr32,
					    ipv6_hdr(skb)->saddr.s6_addr32,
120 121
					    tcp_hdr(skb)->dest,
					    tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
122 123
}

124
static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
L
Linus Torvalds 已提交
125 126 127
			  int addr_len)
{
	struct sockaddr_in6 *usin = (struct sockaddr_in6 *) uaddr;
128
	struct inet_sock *inet = inet_sk(sk);
129
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
130 131
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
132
	struct in6_addr *saddr = NULL, *final_p, final;
133
	struct rt6_info *rt;
L
Linus Torvalds 已提交
134 135 136 137 138
	struct flowi fl;
	struct dst_entry *dst;
	int addr_type;
	int err;

139
	if (addr_len < SIN6_LEN_RFC2133)
L
Linus Torvalds 已提交
140 141
		return -EINVAL;

142
	if (usin->sin6_family != AF_INET6)
E
Eric Dumazet 已提交
143
		return -EAFNOSUPPORT;
L
Linus Torvalds 已提交
144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160

	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);
		}
	}

	/*
161 162 163 164 165
	 *	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
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204

	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) {
205
		u32 exthdrlen = icsk->icsk_ext_hdr_len;
L
Linus Torvalds 已提交
206 207 208 209 210 211 212 213 214 215 216
		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];

217
		icsk->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
218
		sk->sk_backlog_rcv = tcp_v4_do_rcv;
219 220 221
#ifdef CONFIG_TCP_MD5SIG
		tp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif
L
Linus Torvalds 已提交
222 223 224 225

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

		if (err) {
226 227
			icsk->icsk_ext_hdr_len = exthdrlen;
			icsk->icsk_af_ops = &ipv6_specific;
L
Linus Torvalds 已提交
228
			sk->sk_backlog_rcv = tcp_v6_do_rcv;
229 230 231
#ifdef CONFIG_TCP_MD5SIG
			tp->af_specific = &tcp_sock_ipv6_specific;
#endif
L
Linus Torvalds 已提交
232 233
			goto failure;
		} else {
E
Eric Dumazet 已提交
234 235 236
			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
237 238 239 240 241 242 243 244 245 246 247 248 249
		}

		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
250
	fl.mark = sk->sk_mark;
L
Linus Torvalds 已提交
251
	fl.fl_ip_dport = usin->sin6_port;
E
Eric Dumazet 已提交
252
	fl.fl_ip_sport = inet->inet_sport;
L
Linus Torvalds 已提交
253

254
	final_p = fl6_update_dst(&fl, np->opt, &final);
L
Linus Torvalds 已提交
255

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

284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
	rt = (struct rt6_info *) dst;
	if (tcp_death_row.sysctl_tw_recycle &&
	    !tp->rx_opt.ts_recent_stamp &&
	    ipv6_addr_equal(&rt->rt6i_dst.addr, &np->daddr)) {
		struct inet_peer *peer = rt6_get_peer(rt);
		/*
		 * VJ's idea. We save last timestamp seen from
		 * the destination in peer table, when entering state
		 * TIME-WAIT * and initialize rx_opt.ts_recent from it,
		 * when trying new connection.
		 */
		if (peer) {
			inet_peer_refcheck(peer);
			if ((u32)get_seconds() - peer->tcp_ts_stamp <= TCP_PAWS_MSL) {
				tp->rx_opt.ts_recent_stamp = peer->tcp_ts_stamp;
				tp->rx_opt.ts_recent = peer->tcp_ts;
			}
		}
	}

304
	icsk->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
305
	if (np->opt)
306 307
		icsk->icsk_ext_hdr_len = (np->opt->opt_flen +
					  np->opt->opt_nflen);
L
Linus Torvalds 已提交
308 309 310

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

E
Eric Dumazet 已提交
311
	inet->inet_dport = usin->sin6_port;
L
Linus Torvalds 已提交
312 313

	tcp_set_state(sk, TCP_SYN_SENT);
314
	err = inet6_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
315 316 317 318 319 320
	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
321 322
							     inet->inet_sport,
							     inet->inet_dport);
L
Linus Torvalds 已提交
323 324 325 326 327 328 329 330 331 332 333

	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
334
	inet->inet_dport = 0;
L
Linus Torvalds 已提交
335 336 337 338 339
	sk->sk_route_caps = 0;
	return err;
}

static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
340
		u8 type, u8 code, int offset, __be32 info)
L
Linus Torvalds 已提交
341 342
{
	struct ipv6hdr *hdr = (struct ipv6hdr*)skb->data;
343
	const struct tcphdr *th = (struct tcphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
344 345 346
	struct ipv6_pinfo *np;
	struct sock *sk;
	int err;
347
	struct tcp_sock *tp;
L
Linus Torvalds 已提交
348
	__u32 seq;
349
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
350

351
	sk = inet6_lookup(net, &tcp_hashinfo, &hdr->daddr,
352
			th->dest, &hdr->saddr, th->source, skb->dev->ifindex);
L
Linus Torvalds 已提交
353 354

	if (sk == NULL) {
355 356
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
357 358 359 360
		return;
	}

	if (sk->sk_state == TCP_TIME_WAIT) {
361
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
362 363 364 365 366
		return;
	}

	bh_lock_sock(sk);
	if (sock_owned_by_user(sk))
367
		NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS);
L
Linus Torvalds 已提交
368 369 370 371

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

372 373 374 375 376
	if (ipv6_hdr(skb)->hop_limit < inet6_sk(sk)->min_hopcount) {
		NET_INC_STATS_BH(net, LINUX_MIB_TCPMINTTLDROP);
		goto out;
	}

L
Linus Torvalds 已提交
377
	tp = tcp_sk(sk);
378
	seq = ntohl(th->seq);
L
Linus Torvalds 已提交
379 380
	if (sk->sk_state != TCP_LISTEN &&
	    !between(seq, tp->snd_una, tp->snd_nxt)) {
381
		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410
		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
411
			fl.mark = sk->sk_mark;
E
Eric Dumazet 已提交
412 413
			fl.fl_ip_dport = inet->inet_dport;
			fl.fl_ip_sport = inet->inet_sport;
V
Venkat Yekkirala 已提交
414
			security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
415 416 417 418 419 420

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

A
Alexey Dobriyan 已提交
421
			if ((err = xfrm_lookup(net, &dst, &fl, sk, 0)) < 0) {
L
Linus Torvalds 已提交
422 423 424 425 426 427 428
				sk->sk_err_soft = -err;
				goto out;
			}

		} else
			dst_hold(dst);

429
		if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
L
Linus Torvalds 已提交
430 431 432 433 434 435 436 437 438
			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);

439
	/* Might be for an request_sock */
L
Linus Torvalds 已提交
440
	switch (sk->sk_state) {
441
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
442 443 444 445
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

446 447
		req = inet6_csk_search_req(sk, &prev, th->dest, &hdr->daddr,
					   &hdr->saddr, inet6_iif(skb));
L
Linus Torvalds 已提交
448 449 450 451 452 453
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		 * an established socket here.
		 */
454
		WARN_ON(req->sk != NULL);
L
Linus Torvalds 已提交
455

456
		if (seq != tcp_rsk(req)->snt_isn) {
457
			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
458 459 460
			goto out;
		}

461
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
462 463 464 465
		goto out;

	case TCP_SYN_SENT:
	case TCP_SYN_RECV:  /* Cannot happen.
466
			       It can, it SYNs are crossed. --ANK */
L
Linus Torvalds 已提交
467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488
		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);
}


489 490
static int tcp_v6_send_synack(struct sock *sk, struct request_sock *req,
			      struct request_values *rvp)
L
Linus Torvalds 已提交
491
{
492
	struct inet6_request_sock *treq = inet6_rsk(req);
L
Linus Torvalds 已提交
493 494 495
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct sk_buff * skb;
	struct ipv6_txoptions *opt = NULL;
496
	struct in6_addr * final_p, final;
L
Linus Torvalds 已提交
497
	struct flowi fl;
498
	struct dst_entry *dst;
L
Linus Torvalds 已提交
499 500 501 502
	int err = -1;

	memset(&fl, 0, sizeof(fl));
	fl.proto = IPPROTO_TCP;
503 504
	ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
	ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
505
	fl.fl6_flowlabel = 0;
506
	fl.oif = treq->iif;
B
Brian Haley 已提交
507
	fl.mark = sk->sk_mark;
508
	fl.fl_ip_dport = inet_rsk(req)->rmt_port;
509
	fl.fl_ip_sport = inet_rsk(req)->loc_port;
510
	security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
511

512
	opt = np->opt;
513
	final_p = fl6_update_dst(&fl, opt, &final);
L
Linus Torvalds 已提交
514

515 516 517 518 519
	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
520
	if ((err = xfrm_lookup(sock_net(sk), &dst, &fl, sk, 0)) < 0)
521 522
		goto done;

523
	skb = tcp_make_synack(sk, dst, req, rvp);
L
Linus Torvalds 已提交
524
	if (skb) {
525
		__tcp_v6_send_check(skb, &treq->loc_addr, &treq->rmt_addr);
L
Linus Torvalds 已提交
526

527
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
528
		err = ip6_xmit(sk, skb, &fl, opt);
529
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
530 531 532
	}

done:
533
	if (opt && opt != np->opt)
L
Linus Torvalds 已提交
534
		sock_kfree_s(sk, opt, opt->tot_len);
535
	dst_release(dst);
L
Linus Torvalds 已提交
536 537 538
	return err;
}

539 540 541 542 543 544 545
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);
}

546 547 548 549 550 551 552 553 554 555 556 557 558 559
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));
}

560
static void tcp_v6_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
561
{
562
	kfree_skb(inet6_rsk(req)->pktopts);
L
Linus Torvalds 已提交
563 564
}

565 566 567 568 569 570 571 572 573 574 575 576 577
#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++) {
578
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, addr))
579
			return &tp->md5sig_info->keys6[i].base;
580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599
	}
	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 */
600
	struct tcp_md5sig_key *key;
601 602 603
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp6_md5sig_key *keys;

604
	key = tcp_v6_md5_do_lookup(sk, peer);
605 606
	if (key) {
		/* modify existing entry - just update that one */
607 608 609
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
610 611 612 613 614 615 616 617
	} 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;
			}
E
Eric Dumazet 已提交
618
			sk_nocaps_add(sk, NETIF_F_GSO_MASK);
619
		}
620
		if (tcp_alloc_md5sig_pool(sk) == NULL) {
621 622 623
			kfree(newkey);
			return -ENOMEM;
		}
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645
		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);
646 647
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.key = newkey;
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.keylen = newkeylen;
648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666

		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++) {
667
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, peer)) {
668
			/* Free the key */
669
			kfree(tp->md5sig_info->keys6[i].base.key);
670 671 672 673 674
			tp->md5sig_info->entries6--;

			if (tp->md5sig_info->entries6 == 0) {
				kfree(tp->md5sig_info->keys6);
				tp->md5sig_info->keys6 = NULL;
675
				tp->md5sig_info->alloced6 = 0;
676 677 678 679 680 681 682 683
			} 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]));
			}
684 685
			tcp_free_md5sig_pool();
			return 0;
686 687 688 689 690 691 692 693 694 695 696 697
		}
	}
	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++)
698
			kfree(tp->md5sig_info->keys6[i].base.key);
699 700 701 702 703 704 705 706 707 708
		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++)
709
			kfree(tp->md5sig_info->keys4[i].base.key);
710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
		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
738
		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
			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;
E
Eric Dumazet 已提交
755
		sk_nocaps_add(sk, NETIF_F_GSO_MASK);
756 757
	}

A
Arnaldo Carvalho de Melo 已提交
758
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
759 760
	if (!newkey)
		return -ENOMEM;
B
Brian Haley 已提交
761
	if (ipv6_addr_v4mapped(&sin6->sin6_addr)) {
762 763 764 765 766 767
		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);
}

768 769 770
static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp,
					struct in6_addr *daddr,
					struct in6_addr *saddr, int nbytes)
771 772
{
	struct tcp6_pseudohdr *bp;
773
	struct scatterlist sg;
774

775 776 777 778
	bp = &hp->md5_blk.ip6;
	/* 1. TCP pseudo-header (RFC2460) */
	ipv6_addr_copy(&bp->saddr, saddr);
	ipv6_addr_copy(&bp->daddr, daddr);
779
	bp->protocol = cpu_to_be32(IPPROTO_TCP);
A
Adam Langley 已提交
780
	bp->len = cpu_to_be32(nbytes);
781

782 783 784
	sg_init_one(&sg, bp, sizeof(*bp));
	return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp));
}
785

786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806
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))
807 808 809 810
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
811

812 813 814 815
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
816
	return 1;
817 818
}

819 820 821
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)
822 823
{
	struct in6_addr *saddr, *daddr;
824 825 826
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;
	struct tcphdr *th = tcp_hdr(skb);
827 828 829 830

	if (sk) {
		saddr = &inet6_sk(sk)->saddr;
		daddr = &inet6_sk(sk)->daddr;
831
	} else if (req) {
832 833
		saddr = &inet6_rsk(req)->loc_addr;
		daddr = &inet6_rsk(req)->rmt_addr;
834 835 836 837
	} else {
		struct ipv6hdr *ip6h = ipv6_hdr(skb);
		saddr = &ip6h->saddr;
		daddr = &ip6h->daddr;
838
	}
839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866

	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;
867 868 869 870 871 872
}

static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb)
{
	__u8 *hash_location = NULL;
	struct tcp_md5sig_key *hash_expected;
873
	struct ipv6hdr *ip6h = ipv6_hdr(skb);
874
	struct tcphdr *th = tcp_hdr(skb);
875 876 877 878
	int genhash;
	u8 newhash[16];

	hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr);
879
	hash_location = tcp_parse_md5sig_option(th);
880

881 882 883 884 885 886
	/* 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);
887 888 889
		return 1;
	}

890 891
	if (!hash_expected && hash_location) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
892 893 894 895
		return 1;
	}

	/* check the signature */
896 897 898 899
	genhash = tcp_v6_md5_hash_skb(newhash,
				      hash_expected,
				      NULL, NULL, skb);

900 901
	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
902
			printk(KERN_INFO "MD5 Hash %s for [%pI6c]:%u->[%pI6c]:%u\n",
903
			       genhash ? "failed" : "mismatch",
904 905
			       &ip6h->saddr, ntohs(th->source),
			       &ip6h->daddr, ntohs(th->dest));
906 907 908 909 910 911 912
		}
		return 1;
	}
	return 0;
}
#endif

913
struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
914
	.family		=	AF_INET6,
915
	.obj_size	=	sizeof(struct tcp6_request_sock),
916
	.rtx_syn_ack	=	tcp_v6_rtx_synack,
917 918
	.send_ack	=	tcp_v6_reqsk_send_ack,
	.destructor	=	tcp_v6_reqsk_destructor,
919 920
	.send_reset	=	tcp_v6_send_reset,
	.syn_ack_timeout = 	tcp_syn_ack_timeout,
L
Linus Torvalds 已提交
921 922
};

923
#ifdef CONFIG_TCP_MD5SIG
924
static const struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
925
	.md5_lookup	=	tcp_v6_reqsk_md5_lookup,
926
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
927
};
928
#endif
929

930 931
static void __tcp_v6_send_check(struct sk_buff *skb,
				struct in6_addr *saddr, struct in6_addr *daddr)
L
Linus Torvalds 已提交
932
{
933
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
934

935
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
936
		th->check = ~tcp_v6_check(skb->len, saddr, daddr, 0);
937
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
938
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
939
	} else {
940 941 942
		th->check = tcp_v6_check(skb->len, saddr, daddr,
					 csum_partial(th, th->doff << 2,
						      skb->csum));
L
Linus Torvalds 已提交
943 944 945
	}
}

946
static void tcp_v6_send_check(struct sock *sk, struct sk_buff *skb)
947 948 949 950 951 952
{
	struct ipv6_pinfo *np = inet6_sk(sk);

	__tcp_v6_send_check(skb, &np->saddr, &np->daddr);
}

953 954 955 956 957 958 959 960
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;

961
	ipv6h = ipv6_hdr(skb);
962
	th = tcp_hdr(skb);
963 964

	th->check = 0;
965
	skb->ip_summed = CHECKSUM_PARTIAL;
966
	__tcp_v6_send_check(skb, &ipv6h->saddr, &ipv6h->daddr);
967 968
	return 0;
}
L
Linus Torvalds 已提交
969

H
Herbert Xu 已提交
970 971
static struct sk_buff **tcp6_gro_receive(struct sk_buff **head,
					 struct sk_buff *skb)
H
Herbert Xu 已提交
972
{
H
Herbert Xu 已提交
973
	struct ipv6hdr *iph = skb_gro_network_header(skb);
H
Herbert Xu 已提交
974 975 976

	switch (skb->ip_summed) {
	case CHECKSUM_COMPLETE:
977
		if (!tcp_v6_check(skb_gro_len(skb), &iph->saddr, &iph->daddr,
H
Herbert Xu 已提交
978 979 980 981 982 983 984 985 986 987 988 989 990 991
				  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
992
static int tcp6_gro_complete(struct sk_buff *skb)
H
Herbert Xu 已提交
993 994 995 996 997 998 999 1000 1001 1002 1003
{
	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);
}

1004 1005
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
1006
{
1007
	struct tcphdr *th = tcp_hdr(skb), *t1;
L
Linus Torvalds 已提交
1008 1009
	struct sk_buff *buff;
	struct flowi fl;
E
Eric Dumazet 已提交
1010
	struct net *net = dev_net(skb_dst(skb)->dev);
1011
	struct sock *ctl_sk = net->ipv6.tcp_sk;
1012
	unsigned int tot_len = sizeof(struct tcphdr);
E
Eric Dumazet 已提交
1013
	struct dst_entry *dst;
1014
	__be32 *topt;
L
Linus Torvalds 已提交
1015

1016 1017
	if (ts)
		tot_len += TCPOLEN_TSTAMP_ALIGNED;
1018 1019 1020 1021 1022 1023
#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
1024
			 GFP_ATOMIC);
1025 1026
	if (buff == NULL)
		return;
L
Linus Torvalds 已提交
1027

1028
	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
L
Linus Torvalds 已提交
1029

1030
	t1 = (struct tcphdr *) skb_push(buff, tot_len);
1031
	skb_reset_transport_header(buff);
L
Linus Torvalds 已提交
1032 1033 1034 1035 1036

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
1037
	t1->doff = tot_len / 4;
1038 1039 1040 1041 1042
	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
1043

1044 1045
	topt = (__be32 *)(t1 + 1);

1046 1047 1048 1049 1050 1051 1052
	if (ts) {
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
		*topt++ = htonl(tcp_time_stamp);
		*topt++ = htonl(ts);
	}

1053 1054
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
1055 1056 1057
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG);
		tcp_v6_md5_hash_hdr((__u8 *)topt, key,
1058 1059
				    &ipv6_hdr(skb)->saddr,
				    &ipv6_hdr(skb)->daddr, t1);
1060 1061 1062
	}
#endif

L
Linus Torvalds 已提交
1063
	memset(&fl, 0, sizeof(fl));
1064 1065
	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
1066

1067 1068 1069
	buff->ip_summed = CHECKSUM_PARTIAL;
	buff->csum = 0;

1070
	__tcp_v6_send_check(buff, &fl.fl6_src, &fl.fl6_dst);
L
Linus Torvalds 已提交
1071 1072

	fl.proto = IPPROTO_TCP;
1073
	fl.oif = inet6_iif(skb);
L
Linus Torvalds 已提交
1074 1075
	fl.fl_ip_dport = t1->dest;
	fl.fl_ip_sport = t1->source;
V
Venkat Yekkirala 已提交
1076
	security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
1077

1078 1079 1080 1081
	/* 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
1082 1083 1084
	if (!ip6_dst_lookup(ctl_sk, &dst, &fl)) {
		if (xfrm_lookup(net, &dst, &fl, NULL, 0) >= 0) {
			skb_dst_set(buff, dst);
1085
			ip6_xmit(ctl_sk, buff, &fl, NULL);
1086
			TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
1087 1088
			if (rst)
				TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
1089
			return;
A
Arnaldo Carvalho de Melo 已提交
1090
		}
L
Linus Torvalds 已提交
1091 1092 1093 1094 1095
	}

	kfree_skb(buff);
}

1096
static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
1097
{
1098 1099
	struct tcphdr *th = tcp_hdr(skb);
	u32 seq = 0, ack_seq = 0;
1100
	struct tcp_md5sig_key *key = NULL;
L
Linus Torvalds 已提交
1101

1102
	if (th->rst)
L
Linus Torvalds 已提交
1103 1104
		return;

1105 1106
	if (!ipv6_unicast_destination(skb))
		return;
L
Linus Torvalds 已提交
1107

1108
#ifdef CONFIG_TCP_MD5SIG
1109 1110
	if (sk)
		key = tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr);
1111 1112
#endif

1113 1114 1115 1116 1117
	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
1118

1119 1120
	tcp_v6_send_response(skb, seq, ack_seq, 0, 0, key, 1);
}
L
Linus Torvalds 已提交
1121

1122 1123 1124 1125
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
1126 1127 1128 1129
}

static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
1130
	struct inet_timewait_sock *tw = inet_twsk(sk);
1131
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
1132

1133
	tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
1134
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
1135
			tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw));
L
Linus Torvalds 已提交
1136

1137
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
1138 1139
}

1140 1141
static void tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
				  struct request_sock *req)
L
Linus Torvalds 已提交
1142
{
1143
	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent,
1144
			tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr));
L
Linus Torvalds 已提交
1145 1146 1147 1148 1149
}


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

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

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

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

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

/* FIXME: this is substantially similar to the ipv4 code.
 * Can some kind of merge be done? -- erics
 */
static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
{
1186
	struct tcp_extend_values tmp_ext;
1187
	struct tcp_options_received tmp_opt;
1188
	u8 *hash_location;
1189
	struct request_sock *req;
1190
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1191 1192
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_sock *tp = tcp_sk(sk);
1193
	__u32 isn = TCP_SKB_CB(skb)->when;
1194
	struct dst_entry *dst = NULL;
1195 1196 1197 1198 1199
#ifdef CONFIG_SYN_COOKIES
	int want_cookie = 0;
#else
#define want_cookie 0
#endif
L
Linus Torvalds 已提交
1200 1201 1202 1203 1204

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

	if (!ipv6_unicast_destination(skb))
1205
		goto drop;
L
Linus Torvalds 已提交
1206

1207
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1208
		if (net_ratelimit())
1209 1210 1211 1212 1213 1214
			syn_flood_warning(skb);
#ifdef CONFIG_SYN_COOKIES
		if (sysctl_tcp_syncookies)
			want_cookie = 1;
		else
#endif
1215
		goto drop;
L
Linus Torvalds 已提交
1216 1217
	}

1218
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1219 1220
		goto drop;

1221
	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
L
Linus Torvalds 已提交
1222 1223 1224
	if (req == NULL)
		goto drop;

1225 1226 1227 1228
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops;
#endif

L
Linus Torvalds 已提交
1229 1230 1231
	tcp_clear_options(&tmp_opt);
	tmp_opt.mss_clamp = IPV6_MIN_MTU - sizeof(struct tcphdr) - sizeof(struct ipv6hdr);
	tmp_opt.user_mss = tp->rx_opt.user_mss;
1232
	tcp_parse_options(skb, &tmp_opt, &hash_location, 0);
1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248

	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 */
1249
		d = (__force u32 *)&ipv6_hdr(skb)->daddr.s6_addr32[0];
1250 1251 1252 1253
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		*mess++ ^= *d++;
		*mess++ ^= *d++;
1254
		d = (__force u32 *)&ipv6_hdr(skb)->saddr.s6_addr32[0];
1255 1256 1257 1258 1259 1260 1261 1262 1263
		*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
1264

1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277
#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
1278

1279
	if (want_cookie && !tmp_opt.saw_tstamp)
1280 1281
		tcp_clear_options(&tmp_opt);

L
Linus Torvalds 已提交
1282 1283 1284
	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
	tcp_openreq_init(req, &tmp_opt, skb);

1285
	treq = inet6_rsk(req);
1286 1287
	ipv6_addr_copy(&treq->rmt_addr, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&treq->loc_addr, &ipv6_hdr(skb)->daddr);
1288
	if (!want_cookie || tmp_opt.tstamp_ok)
1289 1290
		TCP_ECN_create_request(req, tcp_hdr(skb));

1291
	if (!isn) {
1292 1293
		struct inet_peer *peer = NULL;

1294 1295 1296 1297 1298 1299 1300
		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
1301

1302 1303 1304 1305
		/* 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);
1306 1307

		if (want_cookie) {
1308 1309
			isn = cookie_v6_init_sequence(sk, skb, &req->mss);
			req->cookie_ts = tmp_opt.tstamp_ok;
1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
			goto have_isn;
		}

		/* VJ's idea. We save last timestamp seen
		 * from the destination in peer table, when entering
		 * state TIME-WAIT, and check against it before
		 * accepting new connection request.
		 *
		 * If "isn" is not zero, this request hit alive
		 * timewait bucket, so that all the necessary checks
		 * are made in the function processing timewait state.
		 */
		if (tmp_opt.saw_tstamp &&
		    tcp_death_row.sysctl_tw_recycle &&
		    (dst = inet6_csk_route_req(sk, req)) != NULL &&
		    (peer = rt6_get_peer((struct rt6_info *)dst)) != NULL &&
1326
		    ipv6_addr_equal((struct in6_addr *)peer->daddr.addr.a6,
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
				    &treq->rmt_addr)) {
			inet_peer_refcheck(peer);
			if ((u32)get_seconds() - peer->tcp_ts_stamp < TCP_PAWS_MSL &&
			    (s32)(peer->tcp_ts - req->ts_recent) >
							TCP_PAWS_WINDOW) {
				NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_PAWSPASSIVEREJECTED);
				goto drop_and_release;
			}
		}
		/* Kill the following clause, if you dislike this way. */
		else if (!sysctl_tcp_syncookies &&
			 (sysctl_max_syn_backlog - inet_csk_reqsk_queue_len(sk) <
			  (sysctl_max_syn_backlog >> 2)) &&
			 (!peer || !peer->tcp_ts_stamp) &&
			 (!dst || !dst_metric(dst, RTAX_RTT))) {
			/* Without syncookies last quarter of
			 * backlog is filled with destinations,
			 * proven to be alive.
			 * It means that we continue to communicate
			 * to destinations, already remembered
			 * to the moment of synflood.
			 */
			LIMIT_NETDEBUG(KERN_DEBUG "TCP: drop open request from %pI6/%u\n",
				       &treq->rmt_addr, ntohs(tcp_hdr(skb)->source));
			goto drop_and_release;
1352
		}
1353 1354

		isn = tcp_v6_init_sequence(skb);
1355
	}
1356
have_isn:
1357
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1358

1359 1360
	security_inet_conn_request(sk, skb, req);

1361 1362 1363
	if (tcp_v6_send_synack(sk, req,
			       (struct request_values *)&tmp_ext) ||
	    want_cookie)
1364
		goto drop_and_free;
L
Linus Torvalds 已提交
1365

1366 1367
	inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
	return 0;
L
Linus Torvalds 已提交
1368

1369 1370
drop_and_release:
	dst_release(dst);
1371 1372
drop_and_free:
	reqsk_free(req);
L
Linus Torvalds 已提交
1373 1374 1375 1376 1377
drop:
	return 0; /* don't send reset */
}

static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
1378
					  struct request_sock *req,
L
Linus Torvalds 已提交
1379 1380
					  struct dst_entry *dst)
{
1381
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1382 1383 1384 1385 1386 1387
	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;
1388 1389 1390
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1391 1392 1393 1394 1395 1396 1397 1398

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

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

1399
		if (newsk == NULL)
L
Linus Torvalds 已提交
1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410
			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
1411
		ipv6_addr_set_v4mapped(newinet->inet_daddr, &newnp->daddr);
L
Linus Torvalds 已提交
1412

E
Eric Dumazet 已提交
1413
		ipv6_addr_set_v4mapped(newinet->inet_saddr, &newnp->saddr);
L
Linus Torvalds 已提交
1414 1415 1416

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

1417
		inet_csk(newsk)->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
1418
		newsk->sk_backlog_rcv = tcp_v4_do_rcv;
1419 1420 1421 1422
#ifdef CONFIG_TCP_MD5SIG
		newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif

L
Linus Torvalds 已提交
1423 1424
		newnp->pktoptions  = NULL;
		newnp->opt	   = NULL;
1425
		newnp->mcast_oif   = inet6_iif(skb);
1426
		newnp->mcast_hops  = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1427

1428 1429 1430 1431
		/*
		 * 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
1432 1433 1434
		 */

		/* It is tricky place. Until this moment IPv4 tcp
1435
		   worked with IPv6 icsk.icsk_af_ops.
L
Linus Torvalds 已提交
1436 1437
		   Sync it now.
		 */
1438
		tcp_sync_mss(newsk, inet_csk(newsk)->icsk_pmtu_cookie);
L
Linus Torvalds 已提交
1439 1440 1441 1442

		return newsk;
	}

1443
	treq = inet6_rsk(req);
L
Linus Torvalds 已提交
1444 1445 1446 1447 1448
	opt = np->opt;

	if (sk_acceptq_is_full(sk))
		goto out_overflow;

1449 1450 1451
	if (!dst) {
		dst = inet6_csk_route_req(sk, req);
		if (!dst)
L
Linus Torvalds 已提交
1452
			goto out;
1453
	}
L
Linus Torvalds 已提交
1454 1455 1456

	newsk = tcp_create_openreq_child(sk, req, skb);
	if (newsk == NULL)
1457
		goto out_nonewsk;
L
Linus Torvalds 已提交
1458

1459 1460 1461 1462 1463
	/*
	 * 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
1464

1465
	newsk->sk_gso_type = SKB_GSO_TCPV6;
1466
	__ip6_dst_store(newsk, dst, NULL, NULL);
L
Linus Torvalds 已提交
1467 1468 1469 1470 1471 1472 1473 1474 1475 1476

	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));

1477 1478 1479 1480
	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
1481

1482
	/* Now IPv6 options...
L
Linus Torvalds 已提交
1483 1484 1485 1486

	   First: no IPv4 options.
	 */
	newinet->opt = NULL;
1487
	newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1488 1489 1490 1491 1492 1493

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

	/* Clone pktoptions received with SYN */
	newnp->pktoptions = NULL;
1494 1495 1496 1497
	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
1498 1499 1500 1501
		if (newnp->pktoptions)
			skb_set_owner_r(newnp->pktoptions, newsk);
	}
	newnp->opt	  = NULL;
1502
	newnp->mcast_oif  = inet6_iif(skb);
1503
	newnp->mcast_hops = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516

	/* 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);
	}

1517
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1518
	if (newnp->opt)
1519 1520
		inet_csk(newsk)->icsk_ext_hdr_len = (newnp->opt->opt_nflen +
						     newnp->opt->opt_flen);
L
Linus Torvalds 已提交
1521

J
John Heffner 已提交
1522
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1523
	tcp_sync_mss(newsk, dst_mtu(dst));
1524
	newtp->advmss = dst_metric_advmss(dst);
L
Linus Torvalds 已提交
1525 1526
	tcp_initialize_rcv_mss(newsk);

E
Eric Dumazet 已提交
1527 1528
	newinet->inet_daddr = newinet->inet_saddr = LOOPBACK4_IPV6;
	newinet->inet_rcv_saddr = LOOPBACK4_IPV6;
L
Linus Torvalds 已提交
1529

1530 1531 1532 1533 1534 1535 1536 1537
#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
1538 1539
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1540
			tcp_v6_md5_do_add(newsk, &newnp->daddr,
1541 1542 1543 1544
					  newkey, key->keylen);
	}
#endif

1545 1546 1547 1548
	if (__inet_inherit_port(sk, newsk) < 0) {
		sock_put(newsk);
		goto out;
	}
1549
	__inet6_hash(newsk, NULL);
L
Linus Torvalds 已提交
1550 1551 1552 1553

	return newsk;

out_overflow:
1554
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
1555
out_nonewsk:
L
Linus Torvalds 已提交
1556 1557 1558
	if (opt && opt != np->opt)
		sock_kfree_s(sk, opt, opt->tot_len);
	dst_release(dst);
1559 1560
out:
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
L
Linus Torvalds 已提交
1561 1562 1563
	return NULL;
}

1564
static __sum16 tcp_v6_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1565
{
1566
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
H
Herbert Xu 已提交
1567
		if (!tcp_v6_check(skb->len, &ipv6_hdr(skb)->saddr,
1568
				  &ipv6_hdr(skb)->daddr, skb->csum)) {
1569
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1570
			return 0;
1571
		}
L
Linus Torvalds 已提交
1572
	}
1573

H
Herbert Xu 已提交
1574
	skb->csum = ~csum_unfold(tcp_v6_check(skb->len,
1575 1576
					      &ipv6_hdr(skb)->saddr,
					      &ipv6_hdr(skb)->daddr, 0));
1577

L
Linus Torvalds 已提交
1578
	if (skb->len <= 76) {
1579
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608
	}
	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);

1609 1610 1611 1612 1613
#ifdef CONFIG_TCP_MD5SIG
	if (tcp_v6_inbound_md5_hash (sk, skb))
		goto discard;
#endif

1614
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632
		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.
1633
					       --ANK (980728)
L
Linus Torvalds 已提交
1634 1635 1636 1637 1638 1639
	 */
	if (np->rxopt.all)
		opt_skb = skb_clone(skb, GFP_ATOMIC);

	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1640
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1641 1642 1643 1644 1645 1646 1647
			goto reset;
		TCP_CHECK_TIMER(sk);
		if (opt_skb)
			goto ipv6_pktoptions;
		return 0;
	}

1648
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1649 1650
		goto csum_err;

1651
	if (sk->sk_state == TCP_LISTEN) {
L
Linus Torvalds 已提交
1652 1653 1654 1655 1656 1657 1658 1659 1660
		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..
		 */
1661
		if(nsk != sk) {
L
Linus Torvalds 已提交
1662 1663 1664 1665 1666 1667 1668 1669 1670
			if (tcp_child_process(sk, nsk, skb))
				goto reset;
			if (opt_skb)
				__kfree_skb(opt_skb);
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1671
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1672 1673 1674 1675 1676 1677 1678
		goto reset;
	TCP_CHECK_TIMER(sk);
	if (opt_skb)
		goto ipv6_pktoptions;
	return 0;

reset:
1679
	tcp_v6_send_reset(sk, skb);
L
Linus Torvalds 已提交
1680 1681 1682 1683 1684 1685
discard:
	if (opt_skb)
		__kfree_skb(opt_skb);
	kfree_skb(skb);
	return 0;
csum_err:
1686
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700
	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))) {
1701
		if (np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo)
1702
			np->mcast_oif = inet6_iif(opt_skb);
1703
		if (np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim)
1704
			np->mcast_hops = ipv6_hdr(opt_skb)->hop_limit;
L
Linus Torvalds 已提交
1705 1706 1707 1708 1709 1710 1711 1712 1713
		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);
		}
	}

1714
	kfree_skb(opt_skb);
L
Linus Torvalds 已提交
1715 1716 1717
	return 0;
}

1718
static int tcp_v6_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
1719
{
1720
	struct tcphdr *th;
1721
	struct ipv6hdr *hdr;
L
Linus Torvalds 已提交
1722 1723
	struct sock *sk;
	int ret;
1724
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1725 1726 1727 1728 1729 1730 1731

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

	/*
	 *	Count it even if it's bad.
	 */
1732
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1733 1734 1735 1736

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

1737
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1738 1739 1740 1741 1742 1743

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

1744
	if (!skb_csum_unnecessary(skb) && tcp_v6_checksum_init(skb))
L
Linus Torvalds 已提交
1745 1746
		goto bad_packet;

1747
	th = tcp_hdr(skb);
1748
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
1749 1750 1751 1752 1753
	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;
1754
	TCP_SKB_CB(skb)->flags = ipv6_get_dsfield(hdr);
L
Linus Torvalds 已提交
1755 1756
	TCP_SKB_CB(skb)->sacked = 0;

1757
	sk = __inet6_lookup_skb(&tcp_hashinfo, skb, th->source, th->dest);
L
Linus Torvalds 已提交
1758 1759 1760 1761 1762 1763 1764
	if (!sk)
		goto no_tcp_socket;

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

1765 1766 1767 1768 1769
	if (hdr->hop_limit < inet6_sk(sk)->min_hopcount) {
		NET_INC_STATS_BH(net, LINUX_MIB_TCPMINTTLDROP);
		goto discard_and_relse;
	}

L
Linus Torvalds 已提交
1770 1771 1772
	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
		goto discard_and_relse;

1773
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1774 1775 1776 1777
		goto discard_and_relse;

	skb->dev = NULL;

1778
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1779 1780
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1781
#ifdef CONFIG_NET_DMA
1782
		struct tcp_sock *tp = tcp_sk(sk);
1783
		if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
1784
			tp->ucopy.dma_chan = dma_find_channel(DMA_MEMCPY);
1785 1786 1787
		if (tp->ucopy.dma_chan)
			ret = tcp_v6_do_rcv(sk, skb);
		else
C
Chris Leech 已提交
1788 1789 1790 1791 1792
#endif
		{
			if (!tcp_prequeue(sk, skb))
				ret = tcp_v6_do_rcv(sk, skb);
		}
1793
	} else if (unlikely(sk_add_backlog(sk, skb))) {
Z
Zhu Yi 已提交
1794
		bh_unlock_sock(sk);
1795
		NET_INC_STATS_BH(net, LINUX_MIB_TCPBACKLOGDROP);
Z
Zhu Yi 已提交
1796 1797
		goto discard_and_relse;
	}
L
Linus Torvalds 已提交
1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808
	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:
1809
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1810
	} else {
1811
		tcp_v6_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828
	}

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)) {
1829
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1830 1831 1832 1833
		goto discard_it;
	}

	if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) {
1834
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1835
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1836 1837 1838
		goto discard_it;
	}

1839
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1840 1841 1842 1843
	case TCP_TW_SYN:
	{
		struct sock *sk2;

1844
		sk2 = inet6_lookup_listener(dev_net(skb->dev), &tcp_hashinfo,
1845
					    &ipv6_hdr(skb)->daddr,
1846
					    ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1847
		if (sk2 != NULL) {
1848 1849 1850
			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
1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865
			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;
}

1866 1867
static struct inet_peer *tcp_v6_get_peer(struct sock *sk, bool *release_it)
{
1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879
	struct rt6_info *rt = (struct rt6_info *) __sk_dst_get(sk);
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct inet_peer *peer;

	if (!rt ||
	    !ipv6_addr_equal(&np->daddr, &rt->rt6i_dst.addr)) {
		peer = inet_getpeer_v6(&np->daddr, 1);
		*release_it = true;
	} else {
		if (!rt->rt6i_peer)
			rt6_bind_peer(rt, 1);
		peer = rt->rt6i_peer;
1880
		*release_it = false;
1881 1882 1883
	}

	return peer;
1884 1885 1886
}

static void *tcp_v6_tw_get_peer(struct sock *sk)
L
Linus Torvalds 已提交
1887
{
1888
	struct inet6_timewait_sock *tw6 = inet6_twsk(sk);
1889 1890 1891 1892 1893
	struct inet_timewait_sock *tw = inet_twsk(sk);

	if (tw->tw_family == AF_INET)
		return tcp_v4_tw_get_peer(sk);

1894
	return inet_getpeer_v6(&tw6->tw_v6_daddr, 1);
L
Linus Torvalds 已提交
1895 1896
}

1897 1898 1899 1900 1901 1902 1903
static struct timewait_sock_ops tcp6_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp6_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
	.twsk_destructor= tcp_twsk_destructor,
	.twsk_getpeer	= tcp_v6_tw_get_peer,
};

1904
static const struct inet_connection_sock_af_ops ipv6_specific = {
1905 1906 1907 1908 1909
	.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,
1910
	.get_peer	   = tcp_v6_get_peer,
1911 1912 1913 1914 1915
	.net_header_len	   = sizeof(struct ipv6hdr),
	.setsockopt	   = ipv6_setsockopt,
	.getsockopt	   = ipv6_getsockopt,
	.addr2sockaddr	   = inet6_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in6),
1916
	.bind_conflict	   = inet6_csk_bind_conflict,
1917
#ifdef CONFIG_COMPAT
1918 1919
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1920
#endif
L
Linus Torvalds 已提交
1921 1922
};

1923
#ifdef CONFIG_TCP_MD5SIG
1924
static const struct tcp_sock_af_ops tcp_sock_ipv6_specific = {
1925
	.md5_lookup	=	tcp_v6_md5_lookup,
1926
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
1927 1928 1929
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1930
#endif
1931

L
Linus Torvalds 已提交
1932 1933 1934 1935
/*
 *	TCP over IPv4 via INET6 API
 */

1936
static const struct inet_connection_sock_af_ops ipv6_mapped = {
1937 1938 1939 1940 1941
	.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,
1942
	.get_peer	   = tcp_v4_get_peer,
1943 1944 1945 1946 1947
	.net_header_len	   = sizeof(struct iphdr),
	.setsockopt	   = ipv6_setsockopt,
	.getsockopt	   = ipv6_getsockopt,
	.addr2sockaddr	   = inet6_csk_addr2sockaddr,
	.sockaddr_len	   = sizeof(struct sockaddr_in6),
1948
	.bind_conflict	   = inet6_csk_bind_conflict,
1949
#ifdef CONFIG_COMPAT
1950 1951
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1952
#endif
L
Linus Torvalds 已提交
1953 1954
};

1955
#ifdef CONFIG_TCP_MD5SIG
1956
static const struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = {
1957
	.md5_lookup	=	tcp_v4_md5_lookup,
1958
	.calc_md5_hash	=	tcp_v4_md5_hash_skb,
1959 1960 1961
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1962
#endif
1963

L
Linus Torvalds 已提交
1964 1965 1966 1967 1968
/* 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)
{
1969
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1970 1971 1972 1973 1974 1975
	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);

1976
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988
	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
1989
	tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
L
Linus Torvalds 已提交
1990
	tp->snd_cwnd_clamp = ~0;
1991
	tp->mss_cache = TCP_MSS_DEFAULT;
L
Linus Torvalds 已提交
1992 1993 1994 1995 1996

	tp->reordering = sysctl_tcp_reordering;

	sk->sk_state = TCP_CLOSE;

1997
	icsk->icsk_af_ops = &ipv6_specific;
1998
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
1999
	icsk->icsk_sync_mss = tcp_sync_mss;
L
Linus Torvalds 已提交
2000 2001 2002
	sk->sk_write_space = sk_stream_write_space;
	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);

2003 2004 2005 2006
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv6_specific;
#endif

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
	/* 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
2020 2021 2022
	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

H
Herbert Xu 已提交
2023
	local_bh_disable();
2024
	percpu_counter_inc(&tcp_sockets_allocated);
H
Herbert Xu 已提交
2025
	local_bh_enable();
L
Linus Torvalds 已提交
2026 2027 2028 2029

	return 0;
}

2030
static void tcp_v6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
2031
{
2032 2033 2034 2035 2036
#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
2037
	tcp_v4_destroy_sock(sk);
2038
	inet6_destroy_sock(sk);
L
Linus Torvalds 已提交
2039 2040
}

2041
#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
2042
/* Proc filesystem TCPv6 sock list dumping. */
2043
static void get_openreq6(struct seq_file *seq,
2044
			 struct sock *sk, struct request_sock *req, int i, int uid)
L
Linus Torvalds 已提交
2045 2046
{
	int ttd = req->expires - jiffies;
2047 2048
	struct in6_addr *src = &inet6_rsk(req)->loc_addr;
	struct in6_addr *dest = &inet6_rsk(req)->rmt_addr;
L
Linus Torvalds 已提交
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058

	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],
2059
		   ntohs(inet_rsk(req)->loc_port),
L
Linus Torvalds 已提交
2060 2061
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3],
2062
		   ntohs(inet_rsk(req)->rmt_port),
L
Linus Torvalds 已提交
2063 2064
		   TCP_SYN_RECV,
		   0,0, /* could print option size, but that is af dependent. */
2065 2066
		   1,   /* timers active (only the expire timer) */
		   jiffies_to_clock_t(ttd),
L
Linus Torvalds 已提交
2067 2068
		   req->retrans,
		   uid,
2069
		   0,  /* non standard timer */
L
Linus Torvalds 已提交
2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081
		   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);
2082
	const struct inet_connection_sock *icsk = inet_csk(sp);
L
Linus Torvalds 已提交
2083 2084 2085 2086
	struct ipv6_pinfo *np = inet6_sk(sp);

	dest  = &np->daddr;
	src   = &np->rcv_saddr;
E
Eric Dumazet 已提交
2087 2088
	destp = ntohs(inet->inet_dport);
	srcp  = ntohs(inet->inet_sport);
2089 2090

	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
2091
		timer_active	= 1;
2092 2093
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
2094
		timer_active	= 4;
2095
		timer_expires	= icsk->icsk_timeout;
L
Linus Torvalds 已提交
2096 2097 2098 2099 2100 2101 2102 2103 2104 2105
	} 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 "
2106
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %lu %lu %u %u %d\n",
L
Linus Torvalds 已提交
2107 2108 2109 2110 2111
		   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,
2112
		   sp->sk_state,
2113 2114
		   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
2115 2116
		   timer_active,
		   jiffies_to_clock_t(timer_expires - jiffies),
2117
		   icsk->icsk_retransmits,
L
Linus Torvalds 已提交
2118
		   sock_i_uid(sp),
2119
		   icsk->icsk_probes_out,
L
Linus Torvalds 已提交
2120 2121
		   sock_i_ino(sp),
		   atomic_read(&sp->sk_refcnt), sp,
2122 2123
		   jiffies_to_clock_t(icsk->icsk_rto),
		   jiffies_to_clock_t(icsk->icsk_ack.ato),
2124
		   (icsk->icsk_ack.quick << 1 ) | icsk->icsk_ack.pingpong,
I
Ilpo Järvinen 已提交
2125 2126
		   tp->snd_cwnd,
		   tcp_in_initial_slowstart(tp) ? -1 : tp->snd_ssthresh
L
Linus Torvalds 已提交
2127 2128 2129
		   );
}

2130
static void get_timewait6_sock(struct seq_file *seq,
2131
			       struct inet_timewait_sock *tw, int i)
L
Linus Torvalds 已提交
2132 2133 2134
{
	struct in6_addr *dest, *src;
	__u16 destp, srcp;
2135
	struct inet6_timewait_sock *tw6 = inet6_twsk((struct sock *)tw);
L
Linus Torvalds 已提交
2136 2137 2138 2139 2140
	int ttd = tw->tw_ttd - jiffies;

	if (ttd < 0)
		ttd = 0;

2141 2142
	dest = &tw6->tw_v6_daddr;
	src  = &tw6->tw_v6_rcv_saddr;
L
Linus Torvalds 已提交
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192
	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,
2193 2194 2195
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2196 2197 2198
	.seq_ops	= {
		.show		= tcp6_seq_show,
	},
L
Linus Torvalds 已提交
2199 2200
};

2201
int __net_init tcp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
2202
{
2203
	return tcp_proc_register(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2204 2205
}

2206
void tcp6_proc_exit(struct net *net)
L
Linus Torvalds 已提交
2207
{
2208
	tcp_proc_unregister(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2209 2210 2211 2212 2213 2214 2215 2216 2217
}
#endif

struct proto tcpv6_prot = {
	.name			= "TCPv6",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v6_connect,
	.disconnect		= tcp_disconnect,
2218
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2219 2220 2221 2222 2223 2224 2225
	.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,
2226 2227
	.sendmsg		= tcp_sendmsg,
	.sendpage		= tcp_sendpage,
L
Linus Torvalds 已提交
2228 2229
	.backlog_rcv		= tcp_v6_do_rcv,
	.hash			= tcp_v6_hash,
2230 2231
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2232 2233 2234 2235
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
2236
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2237 2238 2239 2240 2241
	.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),
2242
	.slab_flags		= SLAB_DESTROY_BY_RCU,
2243
	.twsk_prot		= &tcp6_timewait_sock_ops,
2244
	.rsk_prot		= &tcp6_request_sock_ops,
2245
	.h.hashinfo		= &tcp_hashinfo,
2246
	.no_autobind		= true,
2247 2248 2249 2250
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2251 2252
};

2253
static const struct inet6_protocol tcpv6_protocol = {
L
Linus Torvalds 已提交
2254 2255
	.handler	=	tcp_v6_rcv,
	.err_handler	=	tcp_v6_err,
2256
	.gso_send_check	=	tcp_v6_gso_send_check,
H
Herbert Xu 已提交
2257
	.gso_segment	=	tcp_tso_segment,
H
Herbert Xu 已提交
2258 2259
	.gro_receive	=	tcp6_gro_receive,
	.gro_complete	=	tcp6_gro_complete,
L
Linus Torvalds 已提交
2260 2261 2262 2263 2264 2265 2266 2267 2268
	.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,
2269 2270
	.flags		=	INET_PROTOSW_PERMANENT |
				INET_PROTOSW_ICSK,
L
Linus Torvalds 已提交
2271 2272
};

2273
static int __net_init tcpv6_net_init(struct net *net)
2274
{
2275 2276
	return inet_ctl_sock_create(&net->ipv6.tcp_sk, PF_INET6,
				    SOCK_RAW, IPPROTO_TCP, net);
2277 2278
}

2279
static void __net_exit tcpv6_net_exit(struct net *net)
2280
{
2281
	inet_ctl_sock_destroy(net->ipv6.tcp_sk);
E
Eric W. Biederman 已提交
2282 2283
}

2284
static void __net_exit tcpv6_net_exit_batch(struct list_head *net_exit_list)
E
Eric W. Biederman 已提交
2285 2286
{
	inet_twsk_purge(&tcp_hashinfo, &tcp_death_row, AF_INET6);
2287 2288 2289
}

static struct pernet_operations tcpv6_net_ops = {
E
Eric W. Biederman 已提交
2290 2291 2292
	.init	    = tcpv6_net_init,
	.exit	    = tcpv6_net_exit,
	.exit_batch = tcpv6_net_exit_batch,
2293 2294
};

2295
int __init tcpv6_init(void)
L
Linus Torvalds 已提交
2296
{
2297 2298 2299 2300 2301 2302
	int ret;

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

L
Linus Torvalds 已提交
2303
	/* register inet6 protocol */
2304 2305 2306 2307
	ret = inet6_register_protosw(&tcpv6_protosw);
	if (ret)
		goto out_tcpv6_protocol;

2308
	ret = register_pernet_subsys(&tcpv6_net_ops);
2309 2310 2311 2312
	if (ret)
		goto out_tcpv6_protosw;
out:
	return ret;
2313

2314 2315 2316 2317 2318 2319 2320
out_tcpv6_protocol:
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
out_tcpv6_protosw:
	inet6_unregister_protosw(&tcpv6_protosw);
	goto out;
}

2321
void tcpv6_exit(void)
2322
{
2323
	unregister_pernet_subsys(&tcpv6_net_ops);
2324 2325
	inet6_unregister_protosw(&tcpv6_protosw);
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
L
Linus Torvalds 已提交
2326
}