tcp_ipv6.c 52.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	TCP over IPv6
3
 *	Linux INET6 implementation
L
Linus Torvalds 已提交
4 5
 *
 *	Authors:
6
 *	Pedro Roque		<roque@di.fc.ul.pt>
L
Linus Torvalds 已提交
7
 *
8
 *	Based on:
L
Linus Torvalds 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
 *	linux/net/ipv4/tcp.c
 *	linux/net/ipv4/tcp_input.c
 *	linux/net/ipv4/tcp_output.c
 *
 *	Fixes:
 *	Hideaki YOSHIFUJI	:	sin6_scope_id support
 *	YOSHIFUJI Hideaki @USAGI and:	Support IPV6_V6ONLY socket option, which
 *	Alexey Kuznetsov		allow both IPv4 and IPv6 sockets to bind
 *					a single port at the same time.
 *	YOSHIFUJI Hideaki @USAGI:	convert /proc/net/tcp6 to seq_file.
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

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

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

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

#include <asm/uaccess.h>

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

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

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

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

77 78
static struct inet_connection_sock_af_ops ipv6_mapped;
static struct inet_connection_sock_af_ops ipv6_specific;
79
#ifdef CONFIG_TCP_MD5SIG
80 81
static struct tcp_sock_af_ops tcp_sock_ipv6_specific;
static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific;
82 83 84 85 86 87
#else
static struct tcp_md5sig_key *tcp_v6_md5_do_lookup(struct sock *sk,
						   struct in6_addr *addr)
{
	return NULL;
}
88
#endif
L
Linus Torvalds 已提交
89 90 91 92

static void tcp_v6_hash(struct sock *sk)
{
	if (sk->sk_state != TCP_CLOSE) {
93
		if (inet_csk(sk)->icsk_af_ops == &ipv6_mapped) {
L
Linus Torvalds 已提交
94 95 96 97
			tcp_prot.hash(sk);
			return;
		}
		local_bh_disable();
98
		__inet6_hash(sk);
L
Linus Torvalds 已提交
99 100 101 102
		local_bh_enable();
	}
}

103
static __inline__ __sum16 tcp_v6_check(struct tcphdr *th, int len,
104 105
				   struct in6_addr *saddr,
				   struct in6_addr *daddr,
106
				   __wsum base)
L
Linus Torvalds 已提交
107 108 109 110
{
	return csum_ipv6_magic(saddr, daddr, len, IPPROTO_TCP, base);
}

111
static __u32 tcp_v6_init_sequence(struct sk_buff *skb)
L
Linus Torvalds 已提交
112
{
113 114
	return secure_tcpv6_sequence_number(ipv6_hdr(skb)->daddr.s6_addr32,
					    ipv6_hdr(skb)->saddr.s6_addr32,
115 116
					    tcp_hdr(skb)->dest,
					    tcp_hdr(skb)->source);
L
Linus Torvalds 已提交
117 118
}

119
static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
L
Linus Torvalds 已提交
120 121 122
			  int addr_len)
{
	struct sockaddr_in6 *usin = (struct sockaddr_in6 *) uaddr;
123
	struct inet_sock *inet = inet_sk(sk);
124
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
125 126 127 128 129 130 131 132
	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;

133
	if (addr_len < SIN6_LEN_RFC2133)
L
Linus Torvalds 已提交
134 135
		return -EINVAL;

136
	if (usin->sin6_family != AF_INET6)
L
Linus Torvalds 已提交
137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
		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);
		}
	}

	/*
155 156 157 158 159
	 *	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
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198

	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) {
199
		u32 exthdrlen = icsk->icsk_ext_hdr_len;
L
Linus Torvalds 已提交
200 201 202 203 204 205 206 207 208 209 210
		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];

211
		icsk->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
212
		sk->sk_backlog_rcv = tcp_v4_do_rcv;
213 214 215
#ifdef CONFIG_TCP_MD5SIG
		tp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif
L
Linus Torvalds 已提交
216 217 218 219

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

		if (err) {
220 221
			icsk->icsk_ext_hdr_len = exthdrlen;
			icsk->icsk_af_ops = &ipv6_specific;
L
Linus Torvalds 已提交
222
			sk->sk_backlog_rcv = tcp_v6_do_rcv;
223 224 225
#ifdef CONFIG_TCP_MD5SIG
			tp->af_specific = &tcp_sock_ipv6_specific;
#endif
L
Linus Torvalds 已提交
226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
			goto failure;
		} else {
			ipv6_addr_set(&np->saddr, 0, 0, htonl(0x0000FFFF),
				      inet->saddr);
			ipv6_addr_set(&np->rcv_saddr, 0, 0, htonl(0x0000FFFF),
				      inet->rcv_saddr);
		}

		return err;
	}

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

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

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

V
Venkat Yekkirala 已提交
255 256
	security_sk_classify_flow(sk, &fl);

L
Linus Torvalds 已提交
257 258 259 260 261 262
	err = ip6_dst_lookup(sk, &dst, &fl);
	if (err)
		goto failure;
	if (final_p)
		ipv6_addr_copy(&fl.fl6_dst, final_p);

263
	if ((err = __xfrm_lookup(&dst, &fl, sk, XFRM_LOOKUP_WAIT)) < 0) {
264 265 266 267 268
		if (err == -EREMOTE)
			err = ip6_dst_blackhole(sk, &dst, &fl);
		if (err < 0)
			goto failure;
	}
L
Linus Torvalds 已提交
269 270 271 272 273 274 275 276 277 278

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

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

H
Herbert Xu 已提交
279
	sk->sk_gso_type = SKB_GSO_TCPV6;
280
	__ip6_dst_store(sk, dst, NULL, NULL);
L
Linus Torvalds 已提交
281

282
	icsk->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
283
	if (np->opt)
284 285
		icsk->icsk_ext_hdr_len = (np->opt->opt_flen +
					  np->opt->opt_nflen);
L
Linus Torvalds 已提交
286 287 288 289 290 291

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

	inet->dport = usin->sin6_port;

	tcp_set_state(sk, TCP_SYN_SENT);
292
	err = inet6_hash_connect(&tcp_death_row, sk);
L
Linus Torvalds 已提交
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
	if (err)
		goto late_failure;

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

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

	return 0;

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

static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
318
		int type, int code, int offset, __be32 info)
L
Linus Torvalds 已提交
319 320
{
	struct ipv6hdr *hdr = (struct ipv6hdr*)skb->data;
321
	const struct tcphdr *th = (struct tcphdr *)(skb->data+offset);
L
Linus Torvalds 已提交
322 323 324
	struct ipv6_pinfo *np;
	struct sock *sk;
	int err;
325
	struct tcp_sock *tp;
L
Linus Torvalds 已提交
326
	__u32 seq;
327
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
328

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

	if (sk == NULL) {
333 334
		ICMP6_INC_STATS_BH(net, __in6_dev_get(skb->dev),
				   ICMP6_MIB_INERRORS);
L
Linus Torvalds 已提交
335 336 337 338
		return;
	}

	if (sk->sk_state == TCP_TIME_WAIT) {
339
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
340 341 342 343 344
		return;
	}

	bh_lock_sock(sk);
	if (sock_owned_by_user(sk))
345
		NET_INC_STATS_BH(net, LINUX_MIB_LOCKDROPPEDICMPS);
L
Linus Torvalds 已提交
346 347 348 349 350

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

	tp = tcp_sk(sk);
351
	seq = ntohl(th->seq);
L
Linus Torvalds 已提交
352 353
	if (sk->sk_state != TCP_LISTEN &&
	    !between(seq, tp->snd_una, tp->snd_nxt)) {
354
		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385
		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;
			fl.fl_ip_dport = inet->dport;
			fl.fl_ip_sport = inet->sport;
V
Venkat Yekkirala 已提交
386
			security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
387 388 389 390 391 392 393 394 395 396 397 398 399 400

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

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

		} else
			dst_hold(dst);

401
		if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
L
Linus Torvalds 已提交
402 403 404 405 406 407 408 409 410
			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);

411
	/* Might be for an request_sock */
L
Linus Torvalds 已提交
412
	switch (sk->sk_state) {
413
		struct request_sock *req, **prev;
L
Linus Torvalds 已提交
414 415 416 417
	case TCP_LISTEN:
		if (sock_owned_by_user(sk))
			goto out;

418 419
		req = inet6_csk_search_req(sk, &prev, th->dest, &hdr->daddr,
					   &hdr->saddr, inet6_iif(skb));
L
Linus Torvalds 已提交
420 421 422 423 424 425
		if (!req)
			goto out;

		/* ICMPs are not backlogged, hence we cannot get
		 * an established socket here.
		 */
426
		WARN_ON(req->sk != NULL);
L
Linus Torvalds 已提交
427

428
		if (seq != tcp_rsk(req)->snt_isn) {
429
			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
L
Linus Torvalds 已提交
430 431 432
			goto out;
		}

433
		inet_csk_reqsk_queue_drop(sk, req, prev);
L
Linus Torvalds 已提交
434 435 436 437
		goto out;

	case TCP_SYN_SENT:
	case TCP_SYN_RECV:  /* Cannot happen.
438
			       It can, it SYNs are crossed. --ANK */
L
Linus Torvalds 已提交
439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460
		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);
}


461
static int tcp_v6_send_synack(struct sock *sk, struct request_sock *req)
L
Linus Torvalds 已提交
462
{
463
	struct inet6_request_sock *treq = inet6_rsk(req);
L
Linus Torvalds 已提交
464 465 466 467 468
	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;
469
	struct dst_entry *dst;
L
Linus Torvalds 已提交
470 471 472 473
	int err = -1;

	memset(&fl, 0, sizeof(fl));
	fl.proto = IPPROTO_TCP;
474 475
	ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
	ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
476
	fl.fl6_flowlabel = 0;
477 478
	fl.oif = treq->iif;
	fl.fl_ip_dport = inet_rsk(req)->rmt_port;
L
Linus Torvalds 已提交
479
	fl.fl_ip_sport = inet_sk(sk)->sport;
480
	security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
481

482 483 484 485 486 487
	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
488 489
	}

490 491 492 493 494 495 496 497
	err = ip6_dst_lookup(sk, &dst, &fl);
	if (err)
		goto done;
	if (final_p)
		ipv6_addr_copy(&fl.fl6_dst, final_p);
	if ((err = xfrm_lookup(&dst, &fl, sk, 0)) < 0)
		goto done;

L
Linus Torvalds 已提交
498 499
	skb = tcp_make_synack(sk, dst, req);
	if (skb) {
500
		struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
501 502

		th->check = tcp_v6_check(th, skb->len,
503
					 &treq->loc_addr, &treq->rmt_addr,
L
Linus Torvalds 已提交
504 505
					 csum_partial((char *)th, skb->len, skb->csum));

506
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
507
		err = ip6_xmit(sk, skb, &fl, opt, 0);
508
		err = net_xmit_eval(err);
L
Linus Torvalds 已提交
509 510 511
	}

done:
512
	if (opt && opt != np->opt)
L
Linus Torvalds 已提交
513
		sock_kfree_s(sk, opt, opt->tot_len);
514
	dst_release(dst);
L
Linus Torvalds 已提交
515 516 517
	return err;
}

518 519 520 521 522 523 524 525 526 527 528 529 530 531
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));
}

532
static void tcp_v6_reqsk_destructor(struct request_sock *req)
L
Linus Torvalds 已提交
533
{
534 535
	if (inet6_rsk(req)->pktopts)
		kfree_skb(inet6_rsk(req)->pktopts);
L
Linus Torvalds 已提交
536 537
}

538 539 540 541 542 543 544 545 546 547 548 549 550
#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++) {
551
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, addr))
552
			return &tp->md5sig_info->keys6[i].base;
553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572
	}
	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 */
573
	struct tcp_md5sig_key *key;
574 575 576
	struct tcp_sock *tp = tcp_sk(sk);
	struct tcp6_md5sig_key *keys;

577
	key = tcp_v6_md5_do_lookup(sk, peer);
578 579
	if (key) {
		/* modify existing entry - just update that one */
580 581 582
		kfree(key->key);
		key->key = newkey;
		key->keylen = newkeylen;
583 584 585 586 587 588 589 590
	} 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;
			}
591
			sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
592
		}
593 594 595 596
		if (tcp_alloc_md5sig_pool() == NULL) {
			kfree(newkey);
			return -ENOMEM;
		}
597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618
		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);
619 620
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.key = newkey;
		tp->md5sig_info->keys6[tp->md5sig_info->entries6].base.keylen = newkeylen;
621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639

		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++) {
640
		if (ipv6_addr_equal(&tp->md5sig_info->keys6[i].addr, peer)) {
641
			/* Free the key */
642
			kfree(tp->md5sig_info->keys6[i].base.key);
643 644 645 646 647
			tp->md5sig_info->entries6--;

			if (tp->md5sig_info->entries6 == 0) {
				kfree(tp->md5sig_info->keys6);
				tp->md5sig_info->keys6 = NULL;
648
				tp->md5sig_info->alloced6 = 0;
649 650 651 652 653 654 655 656
			} 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]));
			}
657 658
			tcp_free_md5sig_pool();
			return 0;
659 660 661 662 663 664 665 666 667 668 669 670
		}
	}
	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++)
671
			kfree(tp->md5sig_info->keys6[i].base.key);
672 673 674 675 676 677 678 679 680 681
		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++)
682
			kfree(tp->md5sig_info->keys4[i].base.key);
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710
		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
711
		if (ipv6_addr_v4mapped(&sin6->sin6_addr))
712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727
			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;
728
		sk->sk_route_caps &= ~NETIF_F_GSO_MASK;
729 730
	}

A
Arnaldo Carvalho de Melo 已提交
731
	newkey = kmemdup(cmd.tcpm_key, cmd.tcpm_keylen, GFP_KERNEL);
732 733
	if (!newkey)
		return -ENOMEM;
B
Brian Haley 已提交
734
	if (ipv6_addr_v4mapped(&sin6->sin6_addr)) {
735 736 737 738 739 740
		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);
}

741 742 743
static int tcp_v6_md5_hash_pseudoheader(struct tcp_md5sig_pool *hp,
					struct in6_addr *daddr,
					struct in6_addr *saddr, int nbytes)
744 745
{
	struct tcp6_pseudohdr *bp;
746
	struct scatterlist sg;
747

748 749 750 751
	bp = &hp->md5_blk.ip6;
	/* 1. TCP pseudo-header (RFC2460) */
	ipv6_addr_copy(&bp->saddr, saddr);
	ipv6_addr_copy(&bp->daddr, daddr);
752
	bp->protocol = cpu_to_be32(IPPROTO_TCP);
A
Adam Langley 已提交
753
	bp->len = cpu_to_be32(nbytes);
754

755 756 757
	sg_init_one(&sg, bp, sizeof(*bp));
	return crypto_hash_update(&hp->md5_desc, &sg, sizeof(*bp));
}
758

759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779
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))
780 781 782 783
		goto clear_hash;

	tcp_put_md5sig_pool();
	return 0;
784

785 786 787 788
clear_hash:
	tcp_put_md5sig_pool();
clear_hash_noput:
	memset(md5_hash, 0, 16);
789
	return 1;
790 791
}

792 793 794
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)
795 796
{
	struct in6_addr *saddr, *daddr;
797 798 799
	struct tcp_md5sig_pool *hp;
	struct hash_desc *desc;
	struct tcphdr *th = tcp_hdr(skb);
800 801 802 803

	if (sk) {
		saddr = &inet6_sk(sk)->saddr;
		daddr = &inet6_sk(sk)->daddr;
804
	} else if (req) {
805 806
		saddr = &inet6_rsk(req)->loc_addr;
		daddr = &inet6_rsk(req)->rmt_addr;
807 808 809 810
	} else {
		struct ipv6hdr *ip6h = ipv6_hdr(skb);
		saddr = &ip6h->saddr;
		daddr = &ip6h->daddr;
811
	}
812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839

	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;
840 841 842 843 844 845
}

static int tcp_v6_inbound_md5_hash (struct sock *sk, struct sk_buff *skb)
{
	__u8 *hash_location = NULL;
	struct tcp_md5sig_key *hash_expected;
846
	struct ipv6hdr *ip6h = ipv6_hdr(skb);
847
	struct tcphdr *th = tcp_hdr(skb);
848 849 850 851
	int genhash;
	u8 newhash[16];

	hash_expected = tcp_v6_md5_do_lookup(sk, &ip6h->saddr);
852
	hash_location = tcp_parse_md5sig_option(th);
853

854 855 856 857 858 859
	/* 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);
860 861 862
		return 1;
	}

863 864
	if (!hash_expected && hash_location) {
		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
865 866 867 868
		return 1;
	}

	/* check the signature */
869 870 871 872
	genhash = tcp_v6_md5_hash_skb(newhash,
				      hash_expected,
				      NULL, NULL, skb);

873 874 875 876 877 878 879 880 881 882 883 884 885 886 887
	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
		if (net_ratelimit()) {
			printk(KERN_INFO "MD5 Hash %s for "
			       "(" NIP6_FMT ", %u)->"
			       "(" NIP6_FMT ", %u)\n",
			       genhash ? "failed" : "mismatch",
			       NIP6(ip6h->saddr), ntohs(th->source),
			       NIP6(ip6h->daddr), ntohs(th->dest));
		}
		return 1;
	}
	return 0;
}
#endif

888
struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
L
Linus Torvalds 已提交
889
	.family		=	AF_INET6,
890
	.obj_size	=	sizeof(struct tcp6_request_sock),
L
Linus Torvalds 已提交
891
	.rtx_syn_ack	=	tcp_v6_send_synack,
892 893
	.send_ack	=	tcp_v6_reqsk_send_ack,
	.destructor	=	tcp_v6_reqsk_destructor,
L
Linus Torvalds 已提交
894 895 896
	.send_reset	=	tcp_v6_send_reset
};

897
#ifdef CONFIG_TCP_MD5SIG
898
static struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
899 900
	.md5_lookup	=	tcp_v6_reqsk_md5_lookup,
};
901
#endif
902

903 904 905
static struct timewait_sock_ops tcp6_timewait_sock_ops = {
	.twsk_obj_size	= sizeof(struct tcp6_timewait_sock),
	.twsk_unique	= tcp_twsk_unique,
906
	.twsk_destructor= tcp_twsk_destructor,
907 908
};

909
static void tcp_v6_send_check(struct sock *sk, int len, struct sk_buff *skb)
L
Linus Torvalds 已提交
910 911
{
	struct ipv6_pinfo *np = inet6_sk(sk);
912
	struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
913

914
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
L
Linus Torvalds 已提交
915
		th->check = ~csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,  0);
916
		skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
917
		skb->csum_offset = offsetof(struct tcphdr, check);
L
Linus Torvalds 已提交
918
	} else {
919 920
		th->check = csum_ipv6_magic(&np->saddr, &np->daddr, len, IPPROTO_TCP,
					    csum_partial((char *)th, th->doff<<2,
L
Linus Torvalds 已提交
921 922 923 924
							 skb->csum));
	}
}

925 926 927 928 929 930 931 932
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;

933
	ipv6h = ipv6_hdr(skb);
934
	th = tcp_hdr(skb);
935 936 937 938

	th->check = 0;
	th->check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr, skb->len,
				     IPPROTO_TCP, 0);
939
	skb->csum_start = skb_transport_header(skb) - skb->head;
A
Al Viro 已提交
940
	skb->csum_offset = offsetof(struct tcphdr, check);
941
	skb->ip_summed = CHECKSUM_PARTIAL;
942 943
	return 0;
}
L
Linus Torvalds 已提交
944

945 946
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
947
{
948
	struct tcphdr *th = tcp_hdr(skb), *t1;
L
Linus Torvalds 已提交
949 950
	struct sk_buff *buff;
	struct flowi fl;
951
	struct net *net = dev_net(skb->dst->dev);
952
	struct sock *ctl_sk = net->ipv6.tcp_sk;
953
	unsigned int tot_len = sizeof(struct tcphdr);
954
	__be32 *topt;
L
Linus Torvalds 已提交
955

956 957
	if (ts)
		tot_len += TCPOLEN_TSTAMP_ALIGNED;
958 959 960 961 962 963
#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
964
			 GFP_ATOMIC);
965 966
	if (buff == NULL)
		return;
L
Linus Torvalds 已提交
967

968
	skb_reserve(buff, MAX_HEADER + sizeof(struct ipv6hdr) + tot_len);
L
Linus Torvalds 已提交
969

970
	t1 = (struct tcphdr *) skb_push(buff, tot_len);
L
Linus Torvalds 已提交
971 972 973 974 975

	/* Swap the send and the receive. */
	memset(t1, 0, sizeof(*t1));
	t1->dest = th->source;
	t1->source = th->dest;
976
	t1->doff = tot_len / 4;
977 978 979 980 981
	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
982

983 984
	topt = (__be32 *)(t1 + 1);

985 986 987 988 989 990 991
	if (ts) {
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
		*topt++ = htonl(tcp_time_stamp);
		*topt++ = htonl(ts);
	}

992 993
#ifdef CONFIG_TCP_MD5SIG
	if (key) {
994 995 996
		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
				(TCPOPT_MD5SIG << 8) | TCPOLEN_MD5SIG);
		tcp_v6_md5_hash_hdr((__u8 *)topt, key,
997 998
				    &ipv6_hdr(skb)->saddr,
				    &ipv6_hdr(skb)->daddr, t1);
999 1000 1001
	}
#endif

1002
	buff->csum = csum_partial((char *)t1, tot_len, 0);
L
Linus Torvalds 已提交
1003 1004

	memset(&fl, 0, sizeof(fl));
1005 1006
	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
1007 1008

	t1->check = csum_ipv6_magic(&fl.fl6_src, &fl.fl6_dst,
1009
				    tot_len, IPPROTO_TCP,
L
Linus Torvalds 已提交
1010 1011 1012
				    buff->csum);

	fl.proto = IPPROTO_TCP;
1013
	fl.oif = inet6_iif(skb);
L
Linus Torvalds 已提交
1014 1015
	fl.fl_ip_dport = t1->dest;
	fl.fl_ip_sport = t1->source;
V
Venkat Yekkirala 已提交
1016
	security_skb_classify_flow(skb, &fl);
L
Linus Torvalds 已提交
1017

1018 1019 1020 1021
	/* Pass a socket to ip6_dst_lookup either it is for RST
	 * Underlying function will use this to retrieve the network
	 * namespace
	 */
1022
	if (!ip6_dst_lookup(ctl_sk, &buff->dst, &fl)) {
A
Arnaldo Carvalho de Melo 已提交
1023
		if (xfrm_lookup(&buff->dst, &fl, NULL, 0) >= 0) {
1024
			ip6_xmit(ctl_sk, buff, &fl, NULL, 0);
1025
			TCP_INC_STATS_BH(net, TCP_MIB_OUTSEGS);
1026 1027
			if (rst)
				TCP_INC_STATS_BH(net, TCP_MIB_OUTRSTS);
L
Linus Torvalds 已提交
1028
			return;
A
Arnaldo Carvalho de Melo 已提交
1029
		}
L
Linus Torvalds 已提交
1030 1031 1032 1033 1034
	}

	kfree_skb(buff);
}

1035
static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
1036
{
1037 1038
	struct tcphdr *th = tcp_hdr(skb);
	u32 seq = 0, ack_seq = 0;
1039
	struct tcp_md5sig_key *key = NULL;
L
Linus Torvalds 已提交
1040

1041
	if (th->rst)
L
Linus Torvalds 已提交
1042 1043
		return;

1044 1045
	if (!ipv6_unicast_destination(skb))
		return;
L
Linus Torvalds 已提交
1046

1047
#ifdef CONFIG_TCP_MD5SIG
1048 1049
	if (sk)
		key = tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr);
1050 1051
#endif

1052 1053 1054 1055 1056
	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
1057

1058 1059
	tcp_v6_send_response(skb, seq, ack_seq, 0, 0, key, 1);
}
L
Linus Torvalds 已提交
1060

1061 1062 1063 1064
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
1065 1066 1067 1068
}

static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
{
1069
	struct inet_timewait_sock *tw = inet_twsk(sk);
1070
	struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
L
Linus Torvalds 已提交
1071

1072
	tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
1073
			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
1074
			tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw));
L
Linus Torvalds 已提交
1075

1076
	inet_twsk_put(tw);
L
Linus Torvalds 已提交
1077 1078
}

1079 1080
static void tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
				  struct request_sock *req)
L
Linus Torvalds 已提交
1081
{
1082
	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent,
1083
			tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr));
L
Linus Torvalds 已提交
1084 1085 1086 1087 1088
}


static struct sock *tcp_v6_hnd_req(struct sock *sk,struct sk_buff *skb)
{
1089
	struct request_sock *req, **prev;
1090
	const struct tcphdr *th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1091 1092 1093
	struct sock *nsk;

	/* Find possible connection requests. */
1094
	req = inet6_csk_search_req(sk, &prev, th->source,
1095 1096
				   &ipv6_hdr(skb)->saddr,
				   &ipv6_hdr(skb)->daddr, inet6_iif(skb));
L
Linus Torvalds 已提交
1097 1098 1099
	if (req)
		return tcp_check_req(sk, skb, req, prev);

1100
	nsk = __inet6_lookup_established(sock_net(sk), &tcp_hashinfo,
1101 1102
			&ipv6_hdr(skb)->saddr, th->source,
			&ipv6_hdr(skb)->daddr, ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1103 1104 1105 1106 1107 1108

	if (nsk) {
		if (nsk->sk_state != TCP_TIME_WAIT) {
			bh_lock_sock(nsk);
			return nsk;
		}
1109
		inet_twsk_put(inet_twsk(nsk));
L
Linus Torvalds 已提交
1110 1111 1112
		return NULL;
	}

1113
#ifdef CONFIG_SYN_COOKIES
L
Linus Torvalds 已提交
1114
	if (!th->rst && !th->syn && th->ack)
1115
		sk = cookie_v6_check(sk, skb);
L
Linus Torvalds 已提交
1116 1117 1118 1119 1120 1121 1122 1123 1124
#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)
{
1125
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1126 1127 1128
	struct ipv6_pinfo *np = inet6_sk(sk);
	struct tcp_options_received tmp_opt;
	struct tcp_sock *tp = tcp_sk(sk);
1129
	struct request_sock *req = NULL;
L
Linus Torvalds 已提交
1130
	__u32 isn = TCP_SKB_CB(skb)->when;
1131 1132 1133 1134 1135
#ifdef CONFIG_SYN_COOKIES
	int want_cookie = 0;
#else
#define want_cookie 0
#endif
L
Linus Torvalds 已提交
1136 1137 1138 1139 1140

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

	if (!ipv6_unicast_destination(skb))
1141
		goto drop;
L
Linus Torvalds 已提交
1142

1143
	if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
L
Linus Torvalds 已提交
1144
		if (net_ratelimit())
1145 1146 1147 1148 1149 1150
			syn_flood_warning(skb);
#ifdef CONFIG_SYN_COOKIES
		if (sysctl_tcp_syncookies)
			want_cookie = 1;
		else
#endif
1151
		goto drop;
L
Linus Torvalds 已提交
1152 1153
	}

1154
	if (sk_acceptq_is_full(sk) && inet_csk_reqsk_queue_young(sk) > 1)
L
Linus Torvalds 已提交
1155 1156
		goto drop;

1157
	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
L
Linus Torvalds 已提交
1158 1159 1160
	if (req == NULL)
		goto drop;

1161 1162 1163 1164
#ifdef CONFIG_TCP_MD5SIG
	tcp_rsk(req)->af_specific = &tcp_request_sock_ipv6_ops;
#endif

L
Linus Torvalds 已提交
1165 1166 1167 1168 1169 1170
	tcp_clear_options(&tmp_opt);
	tmp_opt.mss_clamp = IPV6_MIN_MTU - sizeof(struct tcphdr) - sizeof(struct ipv6hdr);
	tmp_opt.user_mss = tp->rx_opt.user_mss;

	tcp_parse_options(skb, &tmp_opt, 0);

1171
	if (want_cookie && !tmp_opt.saw_tstamp)
1172 1173
		tcp_clear_options(&tmp_opt);

L
Linus Torvalds 已提交
1174 1175 1176
	tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
	tcp_openreq_init(req, &tmp_opt, skb);

1177
	treq = inet6_rsk(req);
1178 1179
	ipv6_addr_copy(&treq->rmt_addr, &ipv6_hdr(skb)->saddr);
	ipv6_addr_copy(&treq->loc_addr, &ipv6_hdr(skb)->daddr);
1180 1181 1182 1183 1184
	if (!want_cookie)
		TCP_ECN_create_request(req, tcp_hdr(skb));

	if (want_cookie) {
		isn = cookie_v6_init_sequence(sk, skb, &req->mss);
1185
		req->cookie_ts = tmp_opt.tstamp_ok;
1186 1187 1188 1189 1190 1191 1192 1193
	} 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
1194

1195 1196 1197 1198
		/* 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
1199

1200
		isn = tcp_v6_init_sequence(skb);
1201
	}
L
Linus Torvalds 已提交
1202

1203
	tcp_rsk(req)->snt_isn = isn;
L
Linus Torvalds 已提交
1204

1205 1206
	security_inet_conn_request(sk, skb, req);

1207
	if (tcp_v6_send_synack(sk, req))
L
Linus Torvalds 已提交
1208 1209
		goto drop;

1210 1211 1212 1213
	if (!want_cookie) {
		inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
		return 0;
	}
L
Linus Torvalds 已提交
1214 1215 1216

drop:
	if (req)
1217
		reqsk_free(req);
L
Linus Torvalds 已提交
1218 1219 1220 1221 1222

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

static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
1223
					  struct request_sock *req,
L
Linus Torvalds 已提交
1224 1225
					  struct dst_entry *dst)
{
1226
	struct inet6_request_sock *treq;
L
Linus Torvalds 已提交
1227 1228 1229 1230 1231 1232
	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;
1233 1234 1235
#ifdef CONFIG_TCP_MD5SIG
	struct tcp_md5sig_key *key;
#endif
L
Linus Torvalds 已提交
1236 1237 1238 1239 1240 1241 1242 1243

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

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

1244
		if (newsk == NULL)
L
Linus Torvalds 已提交
1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
			return NULL;

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

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

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

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

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

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

1264
		inet_csk(newsk)->icsk_af_ops = &ipv6_mapped;
L
Linus Torvalds 已提交
1265
		newsk->sk_backlog_rcv = tcp_v4_do_rcv;
1266 1267 1268 1269
#ifdef CONFIG_TCP_MD5SIG
		newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
#endif

L
Linus Torvalds 已提交
1270 1271
		newnp->pktoptions  = NULL;
		newnp->opt	   = NULL;
1272
		newnp->mcast_oif   = inet6_iif(skb);
1273
		newnp->mcast_hops  = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1274

1275 1276 1277 1278
		/*
		 * 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
1279 1280 1281
		 */

		/* It is tricky place. Until this moment IPv4 tcp
1282
		   worked with IPv6 icsk.icsk_af_ops.
L
Linus Torvalds 已提交
1283 1284
		   Sync it now.
		 */
1285
		tcp_sync_mss(newsk, inet_csk(newsk)->icsk_pmtu_cookie);
L
Linus Torvalds 已提交
1286 1287 1288 1289

		return newsk;
	}

1290
	treq = inet6_rsk(req);
L
Linus Torvalds 已提交
1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
	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;
1302
		ipv6_addr_copy(&fl.fl6_dst, &treq->rmt_addr);
L
Linus Torvalds 已提交
1303 1304 1305 1306 1307 1308
		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;
		}
1309
		ipv6_addr_copy(&fl.fl6_src, &treq->loc_addr);
L
Linus Torvalds 已提交
1310
		fl.oif = sk->sk_bound_dev_if;
1311
		fl.fl_ip_dport = inet_rsk(req)->rmt_port;
L
Linus Torvalds 已提交
1312
		fl.fl_ip_sport = inet_sk(sk)->sport;
1313
		security_req_classify_flow(req, &fl);
L
Linus Torvalds 已提交
1314 1315 1316 1317 1318 1319 1320 1321 1322

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

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

		if ((xfrm_lookup(&dst, &fl, sk, 0)) < 0)
			goto out;
1323
	}
L
Linus Torvalds 已提交
1324 1325 1326 1327 1328

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

1329 1330 1331 1332 1333
	/*
	 * 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
1334

1335
	newsk->sk_gso_type = SKB_GSO_TCPV6;
1336
	__ip6_dst_store(newsk, dst, NULL, NULL);
L
Linus Torvalds 已提交
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346

	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));

1347 1348 1349 1350
	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
1351

1352
	/* Now IPv6 options...
L
Linus Torvalds 已提交
1353 1354 1355 1356

	   First: no IPv4 options.
	 */
	newinet->opt = NULL;
1357
	newnp->ipv6_fl_list = NULL;
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362 1363

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

	/* Clone pktoptions received with SYN */
	newnp->pktoptions = NULL;
1364 1365 1366 1367
	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
1368 1369 1370 1371
		if (newnp->pktoptions)
			skb_set_owner_r(newnp->pktoptions, newsk);
	}
	newnp->opt	  = NULL;
1372
	newnp->mcast_oif  = inet6_iif(skb);
1373
	newnp->mcast_hops = ipv6_hdr(skb)->hop_limit;
L
Linus Torvalds 已提交
1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386

	/* 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);
	}

1387
	inet_csk(newsk)->icsk_ext_hdr_len = 0;
L
Linus Torvalds 已提交
1388
	if (newnp->opt)
1389 1390
		inet_csk(newsk)->icsk_ext_hdr_len = (newnp->opt->opt_nflen +
						     newnp->opt->opt_flen);
L
Linus Torvalds 已提交
1391

J
John Heffner 已提交
1392
	tcp_mtup_init(newsk);
L
Linus Torvalds 已提交
1393 1394 1395 1396 1397 1398
	tcp_sync_mss(newsk, dst_mtu(dst));
	newtp->advmss = dst_metric(dst, RTAX_ADVMSS);
	tcp_initialize_rcv_mss(newsk);

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

1399 1400 1401 1402 1403 1404 1405 1406
#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
1407 1408
		char *newkey = kmemdup(key->key, key->keylen, GFP_ATOMIC);
		if (newkey != NULL)
1409 1410 1411 1412 1413
			tcp_v6_md5_do_add(newsk, &inet6_sk(sk)->daddr,
					  newkey, key->keylen);
	}
#endif

1414
	__inet6_hash(newsk);
1415
	__inet_inherit_port(sk, newsk);
L
Linus Torvalds 已提交
1416 1417 1418 1419

	return newsk;

out_overflow:
1420
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENOVERFLOWS);
L
Linus Torvalds 已提交
1421
out:
1422
	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428
	if (opt && opt != np->opt)
		sock_kfree_s(sk, opt, opt->tot_len);
	dst_release(dst);
	return NULL;
}

1429
static __sum16 tcp_v6_checksum_init(struct sk_buff *skb)
L
Linus Torvalds 已提交
1430
{
1431
	if (skb->ip_summed == CHECKSUM_COMPLETE) {
1432
		if (!tcp_v6_check(tcp_hdr(skb), skb->len, &ipv6_hdr(skb)->saddr,
1433
				  &ipv6_hdr(skb)->daddr, skb->csum)) {
1434
			skb->ip_summed = CHECKSUM_UNNECESSARY;
L
Linus Torvalds 已提交
1435
			return 0;
1436
		}
L
Linus Torvalds 已提交
1437
	}
1438

1439
	skb->csum = ~csum_unfold(tcp_v6_check(tcp_hdr(skb), skb->len,
1440 1441
					      &ipv6_hdr(skb)->saddr,
					      &ipv6_hdr(skb)->daddr, 0));
1442

L
Linus Torvalds 已提交
1443
	if (skb->len <= 76) {
1444
		return __skb_checksum_complete(skb);
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473
	}
	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);

1474 1475 1476 1477 1478
#ifdef CONFIG_TCP_MD5SIG
	if (tcp_v6_inbound_md5_hash (sk, skb))
		goto discard;
#endif

1479
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
		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.
1498
					       --ANK (980728)
L
Linus Torvalds 已提交
1499 1500 1501 1502 1503 1504
	 */
	if (np->rxopt.all)
		opt_skb = skb_clone(skb, GFP_ATOMIC);

	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
		TCP_CHECK_TIMER(sk);
1505
		if (tcp_rcv_established(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1506 1507 1508 1509 1510 1511 1512
			goto reset;
		TCP_CHECK_TIMER(sk);
		if (opt_skb)
			goto ipv6_pktoptions;
		return 0;
	}

1513
	if (skb->len < tcp_hdrlen(skb) || tcp_checksum_complete(skb))
L
Linus Torvalds 已提交
1514 1515
		goto csum_err;

1516
	if (sk->sk_state == TCP_LISTEN) {
L
Linus Torvalds 已提交
1517 1518 1519 1520 1521 1522 1523 1524 1525
		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..
		 */
1526
		if(nsk != sk) {
L
Linus Torvalds 已提交
1527 1528 1529 1530 1531 1532 1533 1534 1535
			if (tcp_child_process(sk, nsk, skb))
				goto reset;
			if (opt_skb)
				__kfree_skb(opt_skb);
			return 0;
		}
	}

	TCP_CHECK_TIMER(sk);
1536
	if (tcp_rcv_state_process(sk, skb, tcp_hdr(skb), skb->len))
L
Linus Torvalds 已提交
1537 1538 1539 1540 1541 1542 1543
		goto reset;
	TCP_CHECK_TIMER(sk);
	if (opt_skb)
		goto ipv6_pktoptions;
	return 0;

reset:
1544
	tcp_v6_send_reset(sk, skb);
L
Linus Torvalds 已提交
1545 1546 1547 1548 1549 1550
discard:
	if (opt_skb)
		__kfree_skb(opt_skb);
	kfree_skb(skb);
	return 0;
csum_err:
1551
	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
	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))) {
1566
		if (np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo)
1567
			np->mcast_oif = inet6_iif(opt_skb);
1568
		if (np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim)
1569
			np->mcast_hops = ipv6_hdr(opt_skb)->hop_limit;
L
Linus Torvalds 已提交
1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583
		if (ipv6_opt_accepted(sk, opt_skb)) {
			skb_set_owner_r(opt_skb, sk);
			opt_skb = xchg(&np->pktoptions, opt_skb);
		} else {
			__kfree_skb(opt_skb);
			opt_skb = xchg(&np->pktoptions, NULL);
		}
	}

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

1584
static int tcp_v6_rcv(struct sk_buff *skb)
L
Linus Torvalds 已提交
1585
{
1586
	struct tcphdr *th;
L
Linus Torvalds 已提交
1587 1588
	struct sock *sk;
	int ret;
1589
	struct net *net = dev_net(skb->dev);
L
Linus Torvalds 已提交
1590 1591 1592 1593 1594 1595 1596

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

	/*
	 *	Count it even if it's bad.
	 */
1597
	TCP_INC_STATS_BH(net, TCP_MIB_INSEGS);
L
Linus Torvalds 已提交
1598 1599 1600 1601

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

1602
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1603 1604 1605 1606 1607 1608

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

1609
	if (!skb_csum_unnecessary(skb) && tcp_v6_checksum_init(skb))
L
Linus Torvalds 已提交
1610 1611
		goto bad_packet;

1612
	th = tcp_hdr(skb);
L
Linus Torvalds 已提交
1613 1614 1615 1616 1617
	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;
1618
	TCP_SKB_CB(skb)->flags = ipv6_get_dsfield(ipv6_hdr(skb));
L
Linus Torvalds 已提交
1619 1620
	TCP_SKB_CB(skb)->sacked = 0;

1621
	sk = __inet6_lookup_skb(&tcp_hashinfo, skb, th->source, th->dest);
L
Linus Torvalds 已提交
1622 1623 1624 1625 1626 1627 1628 1629 1630 1631
	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;

1632
	if (sk_filter(sk, skb))
L
Linus Torvalds 已提交
1633 1634 1635 1636
		goto discard_and_relse;

	skb->dev = NULL;

1637
	bh_lock_sock_nested(sk);
L
Linus Torvalds 已提交
1638 1639
	ret = 0;
	if (!sock_owned_by_user(sk)) {
C
Chris Leech 已提交
1640
#ifdef CONFIG_NET_DMA
1641
		struct tcp_sock *tp = tcp_sk(sk);
1642 1643
		if (!tp->ucopy.dma_chan && tp->ucopy.pinned_list)
			tp->ucopy.dma_chan = get_softnet_dma();
1644 1645 1646
		if (tp->ucopy.dma_chan)
			ret = tcp_v6_do_rcv(sk, skb);
		else
C
Chris Leech 已提交
1647 1648 1649 1650 1651
#endif
		{
			if (!tcp_prequeue(sk, skb))
				ret = tcp_v6_do_rcv(sk, skb);
		}
L
Linus Torvalds 已提交
1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664
	} 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:
1665
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
L
Linus Torvalds 已提交
1666
	} else {
1667
		tcp_v6_send_reset(NULL, skb);
L
Linus Torvalds 已提交
1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
	}

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)) {
1685
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1686 1687 1688 1689
		goto discard_it;
	}

	if (skb->len < (th->doff<<2) || tcp_checksum_complete(skb)) {
1690
		TCP_INC_STATS_BH(net, TCP_MIB_INERRS);
1691
		inet_twsk_put(inet_twsk(sk));
L
Linus Torvalds 已提交
1692 1693 1694
		goto discard_it;
	}

1695
	switch (tcp_timewait_state_process(inet_twsk(sk), skb, th)) {
L
Linus Torvalds 已提交
1696 1697 1698 1699
	case TCP_TW_SYN:
	{
		struct sock *sk2;

1700
		sk2 = inet6_lookup_listener(dev_net(skb->dev), &tcp_hashinfo,
1701
					    &ipv6_hdr(skb)->daddr,
1702
					    ntohs(th->dest), inet6_iif(skb));
L
Linus Torvalds 已提交
1703
		if (sk2 != NULL) {
1704 1705 1706
			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
1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727
			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;
}

1728
static struct inet_connection_sock_af_ops ipv6_specific = {
1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739
	.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),
1740
	.bind_conflict	   = inet6_csk_bind_conflict,
1741
#ifdef CONFIG_COMPAT
1742 1743
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1744
#endif
L
Linus Torvalds 已提交
1745 1746
};

1747
#ifdef CONFIG_TCP_MD5SIG
1748
static struct tcp_sock_af_ops tcp_sock_ipv6_specific = {
1749
	.md5_lookup	=	tcp_v6_md5_lookup,
1750
	.calc_md5_hash	=	tcp_v6_md5_hash_skb,
1751 1752 1753
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1754
#endif
1755

L
Linus Torvalds 已提交
1756 1757 1758 1759
/*
 *	TCP over IPv4 via INET6 API
 */

1760
static struct inet_connection_sock_af_ops ipv6_mapped = {
1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771
	.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),
1772
	.bind_conflict	   = inet6_csk_bind_conflict,
1773
#ifdef CONFIG_COMPAT
1774 1775
	.compat_setsockopt = compat_ipv6_setsockopt,
	.compat_getsockopt = compat_ipv6_getsockopt,
1776
#endif
L
Linus Torvalds 已提交
1777 1778
};

1779
#ifdef CONFIG_TCP_MD5SIG
1780
static struct tcp_sock_af_ops tcp_sock_ipv6_mapped_specific = {
1781
	.md5_lookup	=	tcp_v4_md5_lookup,
1782
	.calc_md5_hash	=	tcp_v4_md5_hash_skb,
1783 1784 1785
	.md5_add	=	tcp_v6_md5_add_func,
	.md5_parse	=	tcp_v6_parse_md5_keys,
};
1786
#endif
1787

L
Linus Torvalds 已提交
1788 1789 1790 1791 1792
/* 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)
{
1793
	struct inet_connection_sock *icsk = inet_csk(sk);
L
Linus Torvalds 已提交
1794 1795 1796 1797 1798 1799
	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);

1800
	icsk->icsk_rto = TCP_TIMEOUT_INIT;
L
Linus Torvalds 已提交
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814
	tp->mdev = TCP_TIMEOUT_INIT;

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

	/* See draft-stevens-tcpca-spec-01 for discussion of the
	 * initialization of these values.
	 */
	tp->snd_ssthresh = 0x7fffffff;
	tp->snd_cwnd_clamp = ~0;
1815
	tp->mss_cache = 536;
L
Linus Torvalds 已提交
1816 1817 1818 1819 1820

	tp->reordering = sysctl_tcp_reordering;

	sk->sk_state = TCP_CLOSE;

1821
	icsk->icsk_af_ops = &ipv6_specific;
1822
	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
1823
	icsk->icsk_sync_mss = tcp_sync_mss;
L
Linus Torvalds 已提交
1824 1825 1826
	sk->sk_write_space = sk_stream_write_space;
	sock_set_flag(sk, SOCK_USE_WRITE_QUEUE);

1827 1828 1829 1830
#ifdef CONFIG_TCP_MD5SIG
	tp->af_specific = &tcp_sock_ipv6_specific;
#endif

L
Linus Torvalds 已提交
1831 1832 1833 1834 1835 1836 1837 1838
	sk->sk_sndbuf = sysctl_tcp_wmem[1];
	sk->sk_rcvbuf = sysctl_tcp_rmem[1];

	atomic_inc(&tcp_sockets_allocated);

	return 0;
}

1839
static void tcp_v6_destroy_sock(struct sock *sk)
L
Linus Torvalds 已提交
1840
{
1841 1842 1843 1844 1845
#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
1846
	tcp_v4_destroy_sock(sk);
1847
	inet6_destroy_sock(sk);
L
Linus Torvalds 已提交
1848 1849
}

1850
#ifdef CONFIG_PROC_FS
L
Linus Torvalds 已提交
1851
/* Proc filesystem TCPv6 sock list dumping. */
1852
static void get_openreq6(struct seq_file *seq,
1853
			 struct sock *sk, struct request_sock *req, int i, int uid)
L
Linus Torvalds 已提交
1854 1855
{
	int ttd = req->expires - jiffies;
1856 1857
	struct in6_addr *src = &inet6_rsk(req)->loc_addr;
	struct in6_addr *dest = &inet6_rsk(req)->rmt_addr;
L
Linus Torvalds 已提交
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870

	if (ttd < 0)
		ttd = 0;

	seq_printf(seq,
		   "%4d: %08X%08X%08X%08X:%04X %08X%08X%08X%08X:%04X "
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %d %d %p\n",
		   i,
		   src->s6_addr32[0], src->s6_addr32[1],
		   src->s6_addr32[2], src->s6_addr32[3],
		   ntohs(inet_sk(sk)->sport),
		   dest->s6_addr32[0], dest->s6_addr32[1],
		   dest->s6_addr32[2], dest->s6_addr32[3],
1871
		   ntohs(inet_rsk(req)->rmt_port),
L
Linus Torvalds 已提交
1872 1873
		   TCP_SYN_RECV,
		   0,0, /* could print option size, but that is af dependent. */
1874 1875
		   1,   /* timers active (only the expire timer) */
		   jiffies_to_clock_t(ttd),
L
Linus Torvalds 已提交
1876 1877
		   req->retrans,
		   uid,
1878
		   0,  /* non standard timer */
L
Linus Torvalds 已提交
1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890
		   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);
1891
	const struct inet_connection_sock *icsk = inet_csk(sp);
L
Linus Torvalds 已提交
1892 1893 1894 1895 1896 1897
	struct ipv6_pinfo *np = inet6_sk(sp);

	dest  = &np->daddr;
	src   = &np->rcv_saddr;
	destp = ntohs(inet->dport);
	srcp  = ntohs(inet->sport);
1898 1899

	if (icsk->icsk_pending == ICSK_TIME_RETRANS) {
L
Linus Torvalds 已提交
1900
		timer_active	= 1;
1901 1902
		timer_expires	= icsk->icsk_timeout;
	} else if (icsk->icsk_pending == ICSK_TIME_PROBE0) {
L
Linus Torvalds 已提交
1903
		timer_active	= 4;
1904
		timer_expires	= icsk->icsk_timeout;
L
Linus Torvalds 已提交
1905 1906 1907 1908 1909 1910 1911 1912 1913 1914
	} 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 "
1915
		   "%02X %08X:%08X %02X:%08lX %08X %5d %8d %lu %d %p %lu %lu %u %u %d\n",
L
Linus Torvalds 已提交
1916 1917 1918 1919 1920
		   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,
1921
		   sp->sk_state,
1922 1923
		   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
1924 1925
		   timer_active,
		   jiffies_to_clock_t(timer_expires - jiffies),
1926
		   icsk->icsk_retransmits,
L
Linus Torvalds 已提交
1927
		   sock_i_uid(sp),
1928
		   icsk->icsk_probes_out,
L
Linus Torvalds 已提交
1929 1930
		   sock_i_ino(sp),
		   atomic_read(&sp->sk_refcnt), sp,
1931 1932
		   jiffies_to_clock_t(icsk->icsk_rto),
		   jiffies_to_clock_t(icsk->icsk_ack.ato),
1933
		   (icsk->icsk_ack.quick << 1 ) | icsk->icsk_ack.pingpong,
L
Linus Torvalds 已提交
1934 1935 1936 1937
		   tp->snd_cwnd, tp->snd_ssthresh>=0xFFFF?-1:tp->snd_ssthresh
		   );
}

1938
static void get_timewait6_sock(struct seq_file *seq,
1939
			       struct inet_timewait_sock *tw, int i)
L
Linus Torvalds 已提交
1940 1941 1942
{
	struct in6_addr *dest, *src;
	__u16 destp, srcp;
1943
	struct inet6_timewait_sock *tw6 = inet6_twsk((struct sock *)tw);
L
Linus Torvalds 已提交
1944 1945 1946 1947 1948
	int ttd = tw->tw_ttd - jiffies;

	if (ttd < 0)
		ttd = 0;

1949 1950
	dest = &tw6->tw_v6_daddr;
	src  = &tw6->tw_v6_rcv_saddr;
L
Linus Torvalds 已提交
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
	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,
2001 2002 2003
	.seq_fops	= {
		.owner		= THIS_MODULE,
	},
2004 2005 2006
	.seq_ops	= {
		.show		= tcp6_seq_show,
	},
L
Linus Torvalds 已提交
2007 2008
};

2009
int tcp6_proc_init(struct net *net)
L
Linus Torvalds 已提交
2010
{
2011
	return tcp_proc_register(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2012 2013
}

2014
void tcp6_proc_exit(struct net *net)
L
Linus Torvalds 已提交
2015
{
2016
	tcp_proc_unregister(net, &tcp6_seq_afinfo);
L
Linus Torvalds 已提交
2017 2018 2019 2020 2021 2022 2023 2024 2025
}
#endif

struct proto tcpv6_prot = {
	.name			= "TCPv6",
	.owner			= THIS_MODULE,
	.close			= tcp_close,
	.connect		= tcp_v6_connect,
	.disconnect		= tcp_disconnect,
2026
	.accept			= inet_csk_accept,
L
Linus Torvalds 已提交
2027 2028 2029 2030 2031 2032 2033 2034 2035
	.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,
2036 2037
	.unhash			= inet_unhash,
	.get_port		= inet_csk_get_port,
L
Linus Torvalds 已提交
2038 2039 2040 2041
	.enter_memory_pressure	= tcp_enter_memory_pressure,
	.sockets_allocated	= &tcp_sockets_allocated,
	.memory_allocated	= &tcp_memory_allocated,
	.memory_pressure	= &tcp_memory_pressure,
2042
	.orphan_count		= &tcp_orphan_count,
L
Linus Torvalds 已提交
2043 2044 2045 2046 2047
	.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),
2048
	.twsk_prot		= &tcp6_timewait_sock_ops,
2049
	.rsk_prot		= &tcp6_request_sock_ops,
2050
	.h.hashinfo		= &tcp_hashinfo,
2051 2052 2053 2054
#ifdef CONFIG_COMPAT
	.compat_setsockopt	= compat_tcp_setsockopt,
	.compat_getsockopt	= compat_tcp_getsockopt,
#endif
L
Linus Torvalds 已提交
2055 2056 2057 2058 2059
};

static struct inet6_protocol tcpv6_protocol = {
	.handler	=	tcp_v6_rcv,
	.err_handler	=	tcp_v6_err,
2060
	.gso_send_check	=	tcp_v6_gso_send_check,
H
Herbert Xu 已提交
2061
	.gso_segment	=	tcp_tso_segment,
L
Linus Torvalds 已提交
2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
};

static struct inet_protosw tcpv6_protosw = {
	.type		=	SOCK_STREAM,
	.protocol	=	IPPROTO_TCP,
	.prot		=	&tcpv6_prot,
	.ops		=	&inet6_stream_ops,
	.capability	=	-1,
	.no_check	=	0,
2072 2073
	.flags		=	INET_PROTOSW_PERMANENT |
				INET_PROTOSW_ICSK,
L
Linus Torvalds 已提交
2074 2075
};

2076 2077
static int tcpv6_net_init(struct net *net)
{
2078 2079
	return inet_ctl_sock_create(&net->ipv6.tcp_sk, PF_INET6,
				    SOCK_RAW, IPPROTO_TCP, net);
2080 2081 2082 2083
}

static void tcpv6_net_exit(struct net *net)
{
2084
	inet_ctl_sock_destroy(net->ipv6.tcp_sk);
2085
	inet_twsk_purge(net, &tcp_hashinfo, &tcp_death_row, AF_INET6);
2086 2087 2088 2089 2090 2091 2092
}

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

2093
int __init tcpv6_init(void)
L
Linus Torvalds 已提交
2094
{
2095 2096 2097 2098 2099 2100
	int ret;

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

L
Linus Torvalds 已提交
2101
	/* register inet6 protocol */
2102 2103 2104 2105
	ret = inet6_register_protosw(&tcpv6_protosw);
	if (ret)
		goto out_tcpv6_protocol;

2106
	ret = register_pernet_subsys(&tcpv6_net_ops);
2107 2108 2109 2110
	if (ret)
		goto out_tcpv6_protosw;
out:
	return ret;
2111

2112 2113 2114 2115 2116 2117 2118
out_tcpv6_protocol:
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
out_tcpv6_protosw:
	inet6_unregister_protosw(&tcpv6_protosw);
	goto out;
}

2119
void tcpv6_exit(void)
2120
{
2121
	unregister_pernet_subsys(&tcpv6_net_ops);
2122 2123
	inet6_unregister_protosw(&tcpv6_protosw);
	inet6_del_protocol(&tcpv6_protocol, IPPROTO_TCP);
L
Linus Torvalds 已提交
2124
}