ip6_output.c 45.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2
/*
 *	IPv6 output functions
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 9 10 11 12 13 14 15 16 17 18 19 20 21 22
 *
 *	Based on linux/net/ipv4/ip_output.c
 *
 *	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.
 *
 *	Changes:
 *	A.N.Kuznetsov	:	airthmetics in fragmentation.
 *				extension headers are implemented.
 *				route changes now work.
 *				ip6_forward does not confuse sniffers.
 *				etc.
 *
 *      H. von Brand    :       Added missing #include <linux/string.h>
23
 *	Imran Patel	:	frag id should be in NBO
L
Linus Torvalds 已提交
24 25 26 27 28 29
 *      Kazunori MIYAZAWA @USAGI
 *			:       add ip6_append_data and related functions
 *				for datagram xmit
 */

#include <linux/errno.h>
H
Herbert Xu 已提交
30
#include <linux/kernel.h>
L
Linus Torvalds 已提交
31 32 33 34 35 36 37 38
#include <linux/string.h>
#include <linux/socket.h>
#include <linux/net.h>
#include <linux/netdevice.h>
#include <linux/if_arp.h>
#include <linux/in6.h>
#include <linux/tcp.h>
#include <linux/route.h>
39
#include <linux/module.h>
40
#include <linux/slab.h>
L
Linus Torvalds 已提交
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

#include <linux/netfilter.h>
#include <linux/netfilter_ipv6.h>

#include <net/sock.h>
#include <net/snmp.h>

#include <net/ipv6.h>
#include <net/ndisc.h>
#include <net/protocol.h>
#include <net/ip6_route.h>
#include <net/addrconf.h>
#include <net/rawv6.h>
#include <net/icmp.h>
#include <net/xfrm.h>
#include <net/checksum.h>
57
#include <linux/mroute6.h>
D
David Ahern 已提交
58
#include <net/l3mdev.h>
L
Linus Torvalds 已提交
59

60
static int ip6_finish_output2(struct net *net, struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
61
{
E
Eric Dumazet 已提交
62
	struct dst_entry *dst = skb_dst(skb);
L
Linus Torvalds 已提交
63
	struct net_device *dev = dst->dev;
64
	struct neighbour *neigh;
65 66
	struct in6_addr *nexthop;
	int ret;
L
Linus Torvalds 已提交
67 68 69 70

	skb->protocol = htons(ETH_P_IPV6);
	skb->dev = dev;

71
	if (ipv6_addr_is_multicast(&ipv6_hdr(skb)->daddr)) {
E
Eric Dumazet 已提交
72
		struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb));
L
Linus Torvalds 已提交
73

74
		if (!(dev->flags & IFF_LOOPBACK) && sk_mc_loop(sk) &&
75
		    ((mroute6_socket(net, skb) &&
76
		     !(IP6CB(skb)->flags & IP6SKB_FORWARDED)) ||
77 78
		     ipv6_chk_mcast_addr(dev, &ipv6_hdr(skb)->daddr,
					 &ipv6_hdr(skb)->saddr))) {
L
Linus Torvalds 已提交
79 80 81 82 83 84
			struct sk_buff *newskb = skb_clone(skb, GFP_ATOMIC);

			/* Do not check for IFF_ALLMULTI; multicast routing
			   is not supported in any case.
			 */
			if (newskb)
85
				NF_HOOK(NFPROTO_IPV6, NF_INET_POST_ROUTING,
86
					net, sk, newskb, NULL, newskb->dev,
87
					dev_loopback_xmit);
L
Linus Torvalds 已提交
88

89
			if (ipv6_hdr(skb)->hop_limit == 0) {
90
				IP6_INC_STATS(net, idev,
91
					      IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
92 93 94 95 96
				kfree_skb(skb);
				return 0;
			}
		}

97
		IP6_UPD_PO_STATS(net, idev, IPSTATS_MIB_OUTMCAST, skb->len);
98 99 100 101 102 103 104

		if (IPV6_ADDR_MC_SCOPE(&ipv6_hdr(skb)->daddr) <=
		    IPV6_ADDR_SCOPE_NODELOCAL &&
		    !(dev->flags & IFF_LOOPBACK)) {
			kfree_skb(skb);
			return 0;
		}
L
Linus Torvalds 已提交
105 106
	}

107
	rcu_read_lock_bh();
108
	nexthop = rt6_nexthop((struct rt6_info *)dst, &ipv6_hdr(skb)->daddr);
109 110 111 112 113 114 115 116 117
	neigh = __ipv6_neigh_lookup_noref(dst->dev, nexthop);
	if (unlikely(!neigh))
		neigh = __neigh_create(&nd_tbl, nexthop, dst->dev, false);
	if (!IS_ERR(neigh)) {
		ret = dst_neigh_output(dst, neigh, skb);
		rcu_read_unlock_bh();
		return ret;
	}
	rcu_read_unlock_bh();
118

119
	IP6_INC_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTNOROUTES);
120 121
	kfree_skb(skb);
	return -EINVAL;
L
Linus Torvalds 已提交
122 123
}

124
static int ip6_finish_output(struct net *net, struct sock *sk, struct sk_buff *skb)
125 126
{
	if ((skb->len > ip6_skb_dst_mtu(skb) && !skb_is_gso(skb)) ||
127 128
	    dst_allfrag(skb_dst(skb)) ||
	    (IP6CB(skb)->frag_max_size && skb->len > IP6CB(skb)->frag_max_size))
129
		return ip6_fragment(net, sk, skb, ip6_finish_output2);
130
	else
131
		return ip6_finish_output2(net, sk, skb);
132 133
}

E
Eric W. Biederman 已提交
134
int ip6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
135
{
136
	struct net_device *dev = skb_dst(skb)->dev;
E
Eric Dumazet 已提交
137
	struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb));
138

139
	if (unlikely(idev->cnf.disable_ipv6)) {
140
		IP6_INC_STATS(net, idev, IPSTATS_MIB_OUTDISCARDS);
141 142 143 144
		kfree_skb(skb);
		return 0;
	}

145 146
	return NF_HOOK_COND(NFPROTO_IPV6, NF_INET_POST_ROUTING,
			    net, sk, skb, NULL, dev,
147 148
			    ip6_finish_output,
			    !(IP6CB(skb)->flags & IP6SKB_REROUTED));
L
Linus Torvalds 已提交
149 150 151
}

/*
152 153 154 155
 * xmit an sk_buff (used by TCP, SCTP and DCCP)
 * Note : socket lock is not held for SYNACK packets, but might be modified
 * by calls to skb_set_owner_w() and ipv6_local_error(),
 * which are using proper atomic operations or spinlocks.
L
Linus Torvalds 已提交
156
 */
157
int ip6_xmit(const struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6,
158
	     struct ipv6_txoptions *opt, int tclass)
L
Linus Torvalds 已提交
159
{
160
	struct net *net = sock_net(sk);
161
	const struct ipv6_pinfo *np = inet6_sk(sk);
162
	struct in6_addr *first_hop = &fl6->daddr;
E
Eric Dumazet 已提交
163
	struct dst_entry *dst = skb_dst(skb);
L
Linus Torvalds 已提交
164
	struct ipv6hdr *hdr;
165
	u8  proto = fl6->flowi6_proto;
L
Linus Torvalds 已提交
166
	int seg_len = skb->len;
G
Gerrit Renker 已提交
167
	int hlimit = -1;
L
Linus Torvalds 已提交
168 169 170
	u32 mtu;

	if (opt) {
171
		unsigned int head_room;
L
Linus Torvalds 已提交
172 173 174 175 176 177 178 179 180 181

		/* First: exthdrs may take lots of space (~8K for now)
		   MAX_HEADER is not enough.
		 */
		head_room = opt->opt_nflen + opt->opt_flen;
		seg_len += head_room;
		head_room += sizeof(struct ipv6hdr) + LL_RESERVED_SPACE(dst->dev);

		if (skb_headroom(skb) < head_room) {
			struct sk_buff *skb2 = skb_realloc_headroom(skb, head_room);
182
			if (!skb2) {
E
Eric Dumazet 已提交
183
				IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
184 185
					      IPSTATS_MIB_OUTDISCARDS);
				kfree_skb(skb);
L
Linus Torvalds 已提交
186 187
				return -ENOBUFS;
			}
188
			consume_skb(skb);
189
			skb = skb2;
190 191 192 193
			/* skb_set_owner_w() changes sk->sk_wmem_alloc atomically,
			 * it is safe to call in our context (socket lock not held)
			 */
			skb_set_owner_w(skb, (struct sock *)sk);
L
Linus Torvalds 已提交
194 195 196 197 198 199 200
		}
		if (opt->opt_flen)
			ipv6_push_frag_opts(skb, opt, &proto);
		if (opt->opt_nflen)
			ipv6_push_nfrag_opts(skb, opt, &proto, &first_hop);
	}

201 202
	skb_push(skb, sizeof(struct ipv6hdr));
	skb_reset_network_header(skb);
203
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
204 205 206 207

	/*
	 *	Fill in the IPv6 header
	 */
208
	if (np)
L
Linus Torvalds 已提交
209 210
		hlimit = np->hop_limit;
	if (hlimit < 0)
211
		hlimit = ip6_dst_hoplimit(dst);
L
Linus Torvalds 已提交
212

213
	ip6_flow_hdr(hdr, tclass, ip6_make_flowlabel(net, skb, fl6->flowlabel,
214
						     np->autoflowlabel, fl6));
215

L
Linus Torvalds 已提交
216 217 218 219
	hdr->payload_len = htons(seg_len);
	hdr->nexthdr = proto;
	hdr->hop_limit = hlimit;

A
Alexey Dobriyan 已提交
220 221
	hdr->saddr = fl6->saddr;
	hdr->daddr = *first_hop;
L
Linus Torvalds 已提交
222

223
	skb->protocol = htons(ETH_P_IPV6);
224
	skb->priority = sk->sk_priority;
225
	skb->mark = sk->sk_mark;
226

L
Linus Torvalds 已提交
227
	mtu = dst_mtu(dst);
W
WANG Cong 已提交
228
	if ((skb->len <= mtu) || skb->ignore_df || skb_is_gso(skb)) {
E
Eric Dumazet 已提交
229
		IP6_UPD_PO_STATS(net, ip6_dst_idev(skb_dst(skb)),
230
			      IPSTATS_MIB_OUT, skb->len);
231 232 233
		/* hooks should never assume socket lock is held.
		 * we promote our socket to non const
		 */
234
		return NF_HOOK(NFPROTO_IPV6, NF_INET_LOCAL_OUT,
235
			       net, (struct sock *)sk, skb, NULL, dst->dev,
236
			       dst_output);
L
Linus Torvalds 已提交
237 238 239
	}

	skb->dev = dst->dev;
240 241 242 243 244
	/* ipv6_local_error() does not require socket lock,
	 * we promote our socket to non const
	 */
	ipv6_local_error((struct sock *)sk, EMSGSIZE, fl6, mtu);

E
Eric Dumazet 已提交
245
	IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)), IPSTATS_MIB_FRAGFAILS);
L
Linus Torvalds 已提交
246 247 248
	kfree_skb(skb);
	return -EMSGSIZE;
}
249 250
EXPORT_SYMBOL(ip6_xmit);

L
Linus Torvalds 已提交
251 252 253 254 255 256 257 258
static int ip6_call_ra_chain(struct sk_buff *skb, int sel)
{
	struct ip6_ra_chain *ra;
	struct sock *last = NULL;

	read_lock(&ip6_ra_lock);
	for (ra = ip6_ra_chain; ra; ra = ra->next) {
		struct sock *sk = ra->sk;
259 260 261
		if (sk && ra->sel == sel &&
		    (!sk->sk_bound_dev_if ||
		     sk->sk_bound_dev_if == skb->dev->ifindex)) {
L
Linus Torvalds 已提交
262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279
			if (last) {
				struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
				if (skb2)
					rawv6_rcv(last, skb2);
			}
			last = sk;
		}
	}

	if (last) {
		rawv6_rcv(last, skb);
		read_unlock(&ip6_ra_lock);
		return 1;
	}
	read_unlock(&ip6_ra_lock);
	return 0;
}

280 281
static int ip6_forward_proxy_check(struct sk_buff *skb)
{
282
	struct ipv6hdr *hdr = ipv6_hdr(skb);
283
	u8 nexthdr = hdr->nexthdr;
284
	__be16 frag_off;
285 286 287
	int offset;

	if (ipv6_ext_hdr(nexthdr)) {
288
		offset = ipv6_skip_exthdr(skb, sizeof(*hdr), &nexthdr, &frag_off);
289 290 291 292 293 294 295 296
		if (offset < 0)
			return 0;
	} else
		offset = sizeof(struct ipv6hdr);

	if (nexthdr == IPPROTO_ICMPV6) {
		struct icmp6hdr *icmp6;

297 298
		if (!pskb_may_pull(skb, (skb_network_header(skb) +
					 offset + 1 - skb->data)))
299 300
			return 0;

301
		icmp6 = (struct icmp6hdr *)(skb_network_header(skb) + offset);
302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318

		switch (icmp6->icmp6_type) {
		case NDISC_ROUTER_SOLICITATION:
		case NDISC_ROUTER_ADVERTISEMENT:
		case NDISC_NEIGHBOUR_SOLICITATION:
		case NDISC_NEIGHBOUR_ADVERTISEMENT:
		case NDISC_REDIRECT:
			/* For reaction involving unicast neighbor discovery
			 * message destined to the proxied address, pass it to
			 * input function.
			 */
			return 1;
		default:
			break;
		}
	}

319 320 321 322 323 324 325 326 327 328
	/*
	 * The proxying router can't forward traffic sent to a link-local
	 * address, so signal the sender and discard the packet. This
	 * behavior is clarified by the MIPv6 specification.
	 */
	if (ipv6_addr_type(&hdr->daddr) & IPV6_ADDR_LINKLOCAL) {
		dst_link_failure(skb);
		return -1;
	}

329 330 331
	return 0;
}

332 333
static inline int ip6_forward_finish(struct net *net, struct sock *sk,
				     struct sk_buff *skb)
L
Linus Torvalds 已提交
334
{
335
	skb_sender_cpu_clear(skb);
336
	return dst_output(net, sk, skb);
L
Linus Torvalds 已提交
337 338
}

339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
static unsigned int ip6_dst_mtu_forward(const struct dst_entry *dst)
{
	unsigned int mtu;
	struct inet6_dev *idev;

	if (dst_metric_locked(dst, RTAX_MTU)) {
		mtu = dst_metric_raw(dst, RTAX_MTU);
		if (mtu)
			return mtu;
	}

	mtu = IPV6_MIN_MTU;
	rcu_read_lock();
	idev = __in6_dev_get(dst->dev);
	if (idev)
		mtu = idev->cnf.mtu6;
	rcu_read_unlock();

	return mtu;
}

360 361
static bool ip6_pkt_too_big(const struct sk_buff *skb, unsigned int mtu)
{
362
	if (skb->len <= mtu)
363 364
		return false;

W
WANG Cong 已提交
365
	/* ipv6 conntrack defrag sets max_frag_size + ignore_df */
366 367 368
	if (IP6CB(skb)->frag_max_size && IP6CB(skb)->frag_max_size > mtu)
		return true;

W
WANG Cong 已提交
369
	if (skb->ignore_df)
370 371
		return false;

372 373 374 375 376 377
	if (skb_is_gso(skb) && skb_gso_network_seglen(skb) <= mtu)
		return false;

	return true;
}

L
Linus Torvalds 已提交
378 379
int ip6_forward(struct sk_buff *skb)
{
E
Eric Dumazet 已提交
380
	struct dst_entry *dst = skb_dst(skb);
381
	struct ipv6hdr *hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
382
	struct inet6_skb_parm *opt = IP6CB(skb);
383
	struct net *net = dev_net(dst->dev);
384
	u32 mtu;
385

386
	if (net->ipv6.devconf_all->forwarding == 0)
L
Linus Torvalds 已提交
387 388
		goto error;

389 390 391
	if (skb->pkt_type != PACKET_HOST)
		goto drop;

392 393 394
	if (unlikely(skb->sk))
		goto drop;

395 396 397
	if (skb_warn_if_lro(skb))
		goto drop;

L
Linus Torvalds 已提交
398
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_FWD, skb)) {
399 400
		IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
				 IPSTATS_MIB_INDISCARDS);
L
Linus Torvalds 已提交
401 402 403
		goto drop;
	}

404
	skb_forward_csum(skb);
L
Linus Torvalds 已提交
405 406 407 408 409 410 411 412 413 414 415 416 417 418

	/*
	 *	We DO NOT make any processing on
	 *	RA packets, pushing them to user level AS IS
	 *	without ane WARRANTY that application will be able
	 *	to interpret them. The reason is that we
	 *	cannot make anything clever here.
	 *
	 *	We are not end-node, so that if packet contains
	 *	AH/ESP, we cannot make anything.
	 *	Defragmentation also would be mistake, RA packets
	 *	cannot be fragmented, because there is no warranty
	 *	that different fragments will go along one path. --ANK
	 */
419 420
	if (unlikely(opt->flags & IP6SKB_ROUTERALERT)) {
		if (ip6_call_ra_chain(skb, ntohs(opt->ra)))
L
Linus Torvalds 已提交
421 422 423 424 425 426 427 428 429
			return 0;
	}

	/*
	 *	check and decrement ttl
	 */
	if (hdr->hop_limit <= 1) {
		/* Force OUTPUT device used as source address */
		skb->dev = dst->dev;
430
		icmpv6_send(skb, ICMPV6_TIME_EXCEED, ICMPV6_EXC_HOPLIMIT, 0);
431 432
		IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
				 IPSTATS_MIB_INHDRERRORS);
L
Linus Torvalds 已提交
433 434 435 436 437

		kfree_skb(skb);
		return -ETIMEDOUT;
	}

438
	/* XXX: idev->cnf.proxy_ndp? */
439
	if (net->ipv6.devconf_all->proxy_ndp &&
440
	    pneigh_lookup(&nd_tbl, net, &hdr->daddr, skb->dev, 0)) {
441 442
		int proxied = ip6_forward_proxy_check(skb);
		if (proxied > 0)
443
			return ip6_input(skb);
444
		else if (proxied < 0) {
445 446
			IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
					 IPSTATS_MIB_INDISCARDS);
447 448
			goto drop;
		}
449 450
	}

L
Linus Torvalds 已提交
451
	if (!xfrm6_route_forward(skb)) {
452 453
		IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
				 IPSTATS_MIB_INDISCARDS);
L
Linus Torvalds 已提交
454 455
		goto drop;
	}
E
Eric Dumazet 已提交
456
	dst = skb_dst(skb);
L
Linus Torvalds 已提交
457 458 459

	/* IPv6 specs say nothing about it, but it is clear that we cannot
	   send redirects to source routed frames.
460
	   We don't send redirects to frames decapsulated from IPsec.
L
Linus Torvalds 已提交
461
	 */
462
	if (skb->dev == dst->dev && opt->srcrt == 0 && !skb_sec_path(skb)) {
L
Linus Torvalds 已提交
463
		struct in6_addr *target = NULL;
464
		struct inet_peer *peer;
L
Linus Torvalds 已提交
465 466 467 468 469 470 471 472
		struct rt6_info *rt;

		/*
		 *	incoming and outgoing devices are the same
		 *	send a redirect.
		 */

		rt = (struct rt6_info *) dst;
473 474
		if (rt->rt6i_flags & RTF_GATEWAY)
			target = &rt->rt6i_gateway;
L
Linus Torvalds 已提交
475 476 477
		else
			target = &hdr->daddr;

478
		peer = inet_getpeer_v6(net->ipv6.peers, &hdr->daddr, 1);
479

L
Linus Torvalds 已提交
480 481 482
		/* Limit redirects both by destination (here)
		   and by source (inside ndisc_send_redirect)
		 */
483
		if (inet_peer_xrlim_allow(peer, 1*HZ))
484
			ndisc_send_redirect(skb, target);
485 486
		if (peer)
			inet_putpeer(peer);
487 488 489
	} else {
		int addrtype = ipv6_addr_type(&hdr->saddr);

L
Linus Torvalds 已提交
490
		/* This check is security critical. */
491 492
		if (addrtype == IPV6_ADDR_ANY ||
		    addrtype & (IPV6_ADDR_MULTICAST | IPV6_ADDR_LOOPBACK))
493 494 495
			goto error;
		if (addrtype & IPV6_ADDR_LINKLOCAL) {
			icmpv6_send(skb, ICMPV6_DEST_UNREACH,
496
				    ICMPV6_NOT_NEIGHBOUR, 0);
497 498
			goto error;
		}
L
Linus Torvalds 已提交
499 500
	}

501
	mtu = ip6_dst_mtu_forward(dst);
502 503 504
	if (mtu < IPV6_MIN_MTU)
		mtu = IPV6_MIN_MTU;

505
	if (ip6_pkt_too_big(skb, mtu)) {
L
Linus Torvalds 已提交
506 507
		/* Again, force OUTPUT device used as source address */
		skb->dev = dst->dev;
508
		icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
509 510 511 512
		IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
				 IPSTATS_MIB_INTOOBIGERRORS);
		IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
				 IPSTATS_MIB_FRAGFAILS);
L
Linus Torvalds 已提交
513 514 515 516 517
		kfree_skb(skb);
		return -EMSGSIZE;
	}

	if (skb_cow(skb, dst->dev->hard_header_len)) {
518 519
		IP6_INC_STATS_BH(net, ip6_dst_idev(dst),
				 IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
520 521 522
		goto drop;
	}

523
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
524 525

	/* Mangling hops number delayed to point after skb COW */
526

L
Linus Torvalds 已提交
527 528
	hdr->hop_limit--;

529
	IP6_INC_STATS_BH(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTFORWDATAGRAMS);
530
	IP6_ADD_STATS_BH(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTOCTETS, skb->len);
531 532
	return NF_HOOK(NFPROTO_IPV6, NF_INET_FORWARD,
		       net, NULL, skb, skb->dev, dst->dev,
533
		       ip6_forward_finish);
L
Linus Torvalds 已提交
534 535

error:
536
	IP6_INC_STATS_BH(net, ip6_dst_idev(dst), IPSTATS_MIB_INADDRERRORS);
L
Linus Torvalds 已提交
537 538 539 540 541 542 543 544 545 546
drop:
	kfree_skb(skb);
	return -EINVAL;
}

static void ip6_copy_metadata(struct sk_buff *to, struct sk_buff *from)
{
	to->pkt_type = from->pkt_type;
	to->priority = from->priority;
	to->protocol = from->protocol;
E
Eric Dumazet 已提交
547 548
	skb_dst_drop(to);
	skb_dst_set(to, dst_clone(skb_dst(from)));
L
Linus Torvalds 已提交
549
	to->dev = from->dev;
T
Thomas Graf 已提交
550
	to->mark = from->mark;
L
Linus Torvalds 已提交
551 552 553 554

#ifdef CONFIG_NET_SCHED
	to->tc_index = from->tc_index;
#endif
555
	nf_copy(to, from);
556
	skb_copy_secmark(to, from);
L
Linus Torvalds 已提交
557 558
}

559 560
int ip6_fragment(struct net *net, struct sock *sk, struct sk_buff *skb,
		 int (*output)(struct net *, struct sock *, struct sk_buff *))
L
Linus Torvalds 已提交
561 562
{
	struct sk_buff *frag;
563
	struct rt6_info *rt = (struct rt6_info *)skb_dst(skb);
564 565
	struct ipv6_pinfo *np = skb->sk && !dev_recursion_level() ?
				inet6_sk(skb->sk) : NULL;
L
Linus Torvalds 已提交
566 567 568
	struct ipv6hdr *tmp_hdr;
	struct frag_hdr *fh;
	unsigned int mtu, hlen, left, len;
569
	int hroom, troom;
570
	__be32 frag_id;
571
	int ptr, offset = 0, err = 0;
L
Linus Torvalds 已提交
572 573 574 575 576
	u8 *prevhdr, nexthdr = 0;

	hlen = ip6_find_1stfragopt(skb, &prevhdr);
	nexthdr = *prevhdr;

J
John Heffner 已提交
577
	mtu = ip6_skb_dst_mtu(skb);
578 579

	/* We must not fragment if the socket is set to force MTU discovery
580
	 * or if the skb it not generated by a local socket.
581
	 */
582 583
	if (unlikely(!skb->ignore_df && skb->len > mtu))
		goto fail_toobig;
584

585 586 587 588 589 590 591 592
	if (IP6CB(skb)->frag_max_size) {
		if (IP6CB(skb)->frag_max_size > mtu)
			goto fail_toobig;

		/* don't send fragments larger than what we received */
		mtu = IP6CB(skb)->frag_max_size;
		if (mtu < IPV6_MIN_MTU)
			mtu = IPV6_MIN_MTU;
593 594
	}

595 596 597 598
	if (np && np->frag_size < mtu) {
		if (np->frag_size)
			mtu = np->frag_size;
	}
599
	if (mtu < hlen + sizeof(struct frag_hdr) + 8)
600
		goto fail_toobig;
601
	mtu -= hlen + sizeof(struct frag_hdr);
L
Linus Torvalds 已提交
602

603 604
	frag_id = ipv6_select_ident(net, &ipv6_hdr(skb)->daddr,
				    &ipv6_hdr(skb)->saddr);
605

606 607 608 609
	if (skb->ip_summed == CHECKSUM_PARTIAL &&
	    (err = skb_checksum_help(skb)))
		goto fail;

610
	hroom = LL_RESERVED_SPACE(rt->dst.dev);
611
	if (skb_has_frag_list(skb)) {
L
Linus Torvalds 已提交
612
		int first_len = skb_pagelen(skb);
613
		struct sk_buff *frag2;
L
Linus Torvalds 已提交
614 615 616

		if (first_len - hlen > mtu ||
		    ((first_len - hlen) & 7) ||
617 618
		    skb_cloned(skb) ||
		    skb_headroom(skb) < (hroom + sizeof(struct frag_hdr)))
L
Linus Torvalds 已提交
619 620
			goto slow_path;

621
		skb_walk_frags(skb, frag) {
L
Linus Torvalds 已提交
622 623 624
			/* Correct geometry. */
			if (frag->len > mtu ||
			    ((frag->len & 7) && frag->next) ||
625
			    skb_headroom(frag) < (hlen + hroom + sizeof(struct frag_hdr)))
626
				goto slow_path_clean;
L
Linus Torvalds 已提交
627 628 629

			/* Partially cloned skb? */
			if (skb_shared(frag))
630
				goto slow_path_clean;
631 632 633 634 635 636

			BUG_ON(frag->sk);
			if (skb->sk) {
				frag->sk = skb->sk;
				frag->destructor = sock_wfree;
			}
637
			skb->truesize -= frag->truesize;
L
Linus Torvalds 已提交
638 639 640 641 642 643
		}

		err = 0;
		offset = 0;
		/* BUILD HEADER */

Y
YOSHIFUJI Hideaki 已提交
644
		*prevhdr = NEXTHDR_FRAGMENT;
645
		tmp_hdr = kmemdup(skb_network_header(skb), hlen, GFP_ATOMIC);
L
Linus Torvalds 已提交
646
		if (!tmp_hdr) {
E
Eric Dumazet 已提交
647
			IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
648
				      IPSTATS_MIB_FRAGFAILS);
649 650
			err = -ENOMEM;
			goto fail;
L
Linus Torvalds 已提交
651
		}
652 653
		frag = skb_shinfo(skb)->frag_list;
		skb_frag_list_init(skb);
L
Linus Torvalds 已提交
654 655

		__skb_pull(skb, hlen);
656
		fh = (struct frag_hdr *)__skb_push(skb, sizeof(struct frag_hdr));
657 658
		__skb_push(skb, hlen);
		skb_reset_network_header(skb);
659
		memcpy(skb_network_header(skb), tmp_hdr, hlen);
L
Linus Torvalds 已提交
660 661 662 663

		fh->nexthdr = nexthdr;
		fh->reserved = 0;
		fh->frag_off = htons(IP6_MF);
664
		fh->identification = frag_id;
L
Linus Torvalds 已提交
665 666 667 668

		first_len = skb_pagelen(skb);
		skb->data_len = first_len - skb_headlen(skb);
		skb->len = first_len;
669 670
		ipv6_hdr(skb)->payload_len = htons(first_len -
						   sizeof(struct ipv6hdr));
671

672
		dst_hold(&rt->dst);
L
Linus Torvalds 已提交
673 674 675 676 677 678

		for (;;) {
			/* Prepare header of the next frame,
			 * before previous one went down. */
			if (frag) {
				frag->ip_summed = CHECKSUM_NONE;
679
				skb_reset_transport_header(frag);
680
				fh = (struct frag_hdr *)__skb_push(frag, sizeof(struct frag_hdr));
681 682
				__skb_push(frag, hlen);
				skb_reset_network_header(frag);
683 684
				memcpy(skb_network_header(frag), tmp_hdr,
				       hlen);
L
Linus Torvalds 已提交
685 686 687 688
				offset += skb->len - hlen - sizeof(struct frag_hdr);
				fh->nexthdr = nexthdr;
				fh->reserved = 0;
				fh->frag_off = htons(offset);
689
				if (frag->next)
L
Linus Torvalds 已提交
690 691
					fh->frag_off |= htons(IP6_MF);
				fh->identification = frag_id;
692 693 694
				ipv6_hdr(frag)->payload_len =
						htons(frag->len -
						      sizeof(struct ipv6hdr));
L
Linus Torvalds 已提交
695 696
				ip6_copy_metadata(frag, skb);
			}
697

698
			err = output(net, sk, skb);
699
			if (!err)
700
				IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
701
					      IPSTATS_MIB_FRAGCREATES);
702

L
Linus Torvalds 已提交
703 704 705 706 707 708 709 710
			if (err || !frag)
				break;

			skb = frag;
			frag = skb->next;
			skb->next = NULL;
		}

J
Jesper Juhl 已提交
711
		kfree(tmp_hdr);
L
Linus Torvalds 已提交
712 713

		if (err == 0) {
714
			IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
715
				      IPSTATS_MIB_FRAGOKS);
A
Amerigo Wang 已提交
716
			ip6_rt_put(rt);
L
Linus Torvalds 已提交
717 718 719
			return 0;
		}

720
		kfree_skb_list(frag);
L
Linus Torvalds 已提交
721

722
		IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
723
			      IPSTATS_MIB_FRAGFAILS);
A
Amerigo Wang 已提交
724
		ip6_rt_put(rt);
L
Linus Torvalds 已提交
725
		return err;
726 727 728 729 730 731 732 733 734

slow_path_clean:
		skb_walk_frags(skb, frag2) {
			if (frag2 == frag)
				break;
			frag2->sk = NULL;
			frag2->destructor = NULL;
			skb->truesize += frag2->truesize;
		}
L
Linus Torvalds 已提交
735 736 737 738 739 740 741 742 743 744 745
	}

slow_path:
	left = skb->len - hlen;		/* Space per frame */
	ptr = hlen;			/* Where to start from */

	/*
	 *	Fragment the datagram.
	 */

	*prevhdr = NEXTHDR_FRAGMENT;
746
	troom = rt->dst.dev->needed_tailroom;
L
Linus Torvalds 已提交
747 748 749 750

	/*
	 *	Keep copying data until we run out.
	 */
751
	while (left > 0)	{
L
Linus Torvalds 已提交
752 753 754 755
		len = left;
		/* IF: it doesn't fit, use 'mtu' - the data space left */
		if (len > mtu)
			len = mtu;
L
Lucas De Marchi 已提交
756
		/* IF: we are not sending up to and including the packet end
L
Linus Torvalds 已提交
757 758 759 760 761
		   then align the next start on an eight byte boundary */
		if (len < left)	{
			len &= ~7;
		}

762 763 764 765
		/* Allocate buffer */
		frag = alloc_skb(len + hlen + sizeof(struct frag_hdr) +
				 hroom + troom, GFP_ATOMIC);
		if (!frag) {
E
Eric Dumazet 已提交
766
			IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
767
				      IPSTATS_MIB_FRAGFAILS);
L
Linus Torvalds 已提交
768 769 770 771 772 773 774 775 776
			err = -ENOMEM;
			goto fail;
		}

		/*
		 *	Set up data on packet
		 */

		ip6_copy_metadata(frag, skb);
777
		skb_reserve(frag, hroom);
L
Linus Torvalds 已提交
778
		skb_put(frag, len + hlen + sizeof(struct frag_hdr));
779
		skb_reset_network_header(frag);
780
		fh = (struct frag_hdr *)(skb_network_header(frag) + hlen);
781 782
		frag->transport_header = (frag->network_header + hlen +
					  sizeof(struct frag_hdr));
L
Linus Torvalds 已提交
783 784 785 786 787 788 789 790 791 792 793

		/*
		 *	Charge the memory for the fragment to any owner
		 *	it might possess
		 */
		if (skb->sk)
			skb_set_owner_w(frag, skb->sk);

		/*
		 *	Copy the packet header into the new buffer.
		 */
794
		skb_copy_from_linear_data(skb, skb_network_header(frag), hlen);
L
Linus Torvalds 已提交
795 796 797 798 799 800

		/*
		 *	Build fragment header.
		 */
		fh->nexthdr = nexthdr;
		fh->reserved = 0;
801
		fh->identification = frag_id;
L
Linus Torvalds 已提交
802 803 804 805

		/*
		 *	Copy a block of the IP datagram.
		 */
H
Himangi Saraogi 已提交
806 807
		BUG_ON(skb_copy_bits(skb, ptr, skb_transport_header(frag),
				     len));
L
Linus Torvalds 已提交
808 809 810 811 812
		left -= len;

		fh->frag_off = htons(offset);
		if (left > 0)
			fh->frag_off |= htons(IP6_MF);
813 814
		ipv6_hdr(frag)->payload_len = htons(frag->len -
						    sizeof(struct ipv6hdr));
L
Linus Torvalds 已提交
815 816 817 818 819 820 821

		ptr += len;
		offset += len;

		/*
		 *	Put this fragment into the sending queue.
		 */
822
		err = output(net, sk, frag);
L
Linus Torvalds 已提交
823 824
		if (err)
			goto fail;
825

E
Eric Dumazet 已提交
826
		IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
827
			      IPSTATS_MIB_FRAGCREATES);
L
Linus Torvalds 已提交
828
	}
E
Eric Dumazet 已提交
829
	IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
830
		      IPSTATS_MIB_FRAGOKS);
831
	consume_skb(skb);
L
Linus Torvalds 已提交
832 833
	return err;

834 835 836 837 838 839 840 841
fail_toobig:
	if (skb->sk && dst_allfrag(skb_dst(skb)))
		sk_nocaps_add(skb->sk, NETIF_F_GSO_MASK);

	skb->dev = skb_dst(skb)->dev;
	icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
	err = -EMSGSIZE;

L
Linus Torvalds 已提交
842
fail:
E
Eric Dumazet 已提交
843
	IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
844
		      IPSTATS_MIB_FRAGFAILS);
845
	kfree_skb(skb);
L
Linus Torvalds 已提交
846 847 848
	return err;
}

849 850 851
static inline int ip6_rt_check(const struct rt6key *rt_key,
			       const struct in6_addr *fl_addr,
			       const struct in6_addr *addr_cache)
852
{
E
Eric Dumazet 已提交
853
	return (rt_key->plen != 128 || !ipv6_addr_equal(fl_addr, &rt_key->addr)) &&
854
		(!addr_cache || !ipv6_addr_equal(fl_addr, addr_cache));
855 856
}

857 858
static struct dst_entry *ip6_sk_dst_check(struct sock *sk,
					  struct dst_entry *dst,
859
					  const struct flowi6 *fl6)
L
Linus Torvalds 已提交
860
{
861
	struct ipv6_pinfo *np = inet6_sk(sk);
862
	struct rt6_info *rt;
L
Linus Torvalds 已提交
863

864 865 866
	if (!dst)
		goto out;

867 868 869 870 871 872
	if (dst->ops->family != AF_INET6) {
		dst_release(dst);
		return NULL;
	}

	rt = (struct rt6_info *)dst;
873 874 875
	/* Yes, checking route validity in not connected
	 * case is not very simple. Take into account,
	 * that we do not support routing by source, TOS,
876
	 * and MSG_DONTROUTE		--ANK (980726)
877
	 *
878 879
	 * 1. ip6_rt_check(): If route was host route,
	 *    check that cached destination is current.
880 881 882 883 884 885 886 887 888 889
	 *    If it is network route, we still may
	 *    check its validity using saved pointer
	 *    to the last used address: daddr_cache.
	 *    We do not want to save whole address now,
	 *    (because main consumer of this service
	 *    is tcp, which has not this problem),
	 *    so that the last trick works only on connected
	 *    sockets.
	 * 2. oif also should be the same.
	 */
890
	if (ip6_rt_check(&rt->rt6i_dst, &fl6->daddr, np->daddr_cache) ||
891
#ifdef CONFIG_IPV6_SUBTREES
892
	    ip6_rt_check(&rt->rt6i_src, &fl6->saddr, np->saddr_cache) ||
893
#endif
D
David Ahern 已提交
894 895
	   (!(fl6->flowi6_flags & FLOWI_FLAG_SKIP_NH_OIF) &&
	      (fl6->flowi6_oif && fl6->flowi6_oif != dst->dev->ifindex))) {
896 897
		dst_release(dst);
		dst = NULL;
L
Linus Torvalds 已提交
898 899
	}

900 901 902 903
out:
	return dst;
}

904
static int ip6_dst_lookup_tail(struct net *net, const struct sock *sk,
905
			       struct dst_entry **dst, struct flowi6 *fl6)
906
{
907 908
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	struct neighbour *n;
909
	struct rt6_info *rt;
910 911
#endif
	int err;
912
	int flags = 0;
913

914 915 916 917 918 919 920 921 922 923 924 925
	/* The correct way to handle this would be to do
	 * ip6_route_get_saddr, and then ip6_route_output; however,
	 * the route-specific preferred source forces the
	 * ip6_route_output call _before_ ip6_route_get_saddr.
	 *
	 * In source specific routing (no src=any default route),
	 * ip6_route_output will fail given src=any saddr, though, so
	 * that's why we try it again later.
	 */
	if (ipv6_addr_any(&fl6->saddr) && (!*dst || !(*dst)->error)) {
		struct rt6_info *rt;
		bool had_dst = *dst != NULL;
L
Linus Torvalds 已提交
926

927 928 929
		if (!had_dst)
			*dst = ip6_route_output(net, sk, fl6);
		rt = (*dst)->error ? NULL : (struct rt6_info *)*dst;
930 931 932
		err = ip6_route_get_saddr(net, rt, &fl6->daddr,
					  sk ? inet6_sk(sk)->srcprefs : 0,
					  &fl6->saddr);
933
		if (err)
L
Linus Torvalds 已提交
934
			goto out_err_release;
935 936 937 938 939 940 941 942 943

		/* If we had an erroneous initial result, pretend it
		 * never existed and let the SA-enabled version take
		 * over.
		 */
		if (!had_dst && (*dst)->error) {
			dst_release(*dst);
			*dst = NULL;
		}
944 945 946

		if (fl6->flowi6_oif)
			flags |= RT6_LOOKUP_F_IFACE;
L
Linus Torvalds 已提交
947 948
	}

949
	if (!*dst)
950
		*dst = ip6_route_output_flags(net, sk, fl6, flags);
951 952 953 954 955

	err = (*dst)->error;
	if (err)
		goto out_err_release;

956
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
957 958 959 960 961 962 963 964
	/*
	 * Here if the dst entry we've looked up
	 * has a neighbour entry that is in the INCOMPLETE
	 * state and the src address from the flow is
	 * marked as OPTIMISTIC, we release the found
	 * dst entry and replace it instead with the
	 * dst entry of the nexthop router
	 */
965
	rt = (struct rt6_info *) *dst;
966
	rcu_read_lock_bh();
967 968
	n = __ipv6_neigh_lookup_noref(rt->dst.dev,
				      rt6_nexthop(rt, &fl6->daddr));
969 970 971 972
	err = n && !(n->nud_state & NUD_VALID) ? -EINVAL : 0;
	rcu_read_unlock_bh();

	if (err) {
973
		struct inet6_ifaddr *ifp;
974
		struct flowi6 fl_gw6;
975 976
		int redirect;

977
		ifp = ipv6_get_ifaddr(net, &fl6->saddr,
978 979 980 981 982 983 984 985 986 987 988 989
				      (*dst)->dev, 1);

		redirect = (ifp && ifp->flags & IFA_F_OPTIMISTIC);
		if (ifp)
			in6_ifa_put(ifp);

		if (redirect) {
			/*
			 * We need to get the dst entry for the
			 * default router instead
			 */
			dst_release(*dst);
990 991 992
			memcpy(&fl_gw6, fl6, sizeof(struct flowi6));
			memset(&fl_gw6.daddr, 0, sizeof(struct in6_addr));
			*dst = ip6_route_output(net, sk, &fl_gw6);
993 994
			err = (*dst)->error;
			if (err)
995
				goto out_err_release;
996
		}
997
	}
998 999
#endif

L
Linus Torvalds 已提交
1000 1001 1002
	return 0;

out_err_release:
1003
	if (err == -ENETUNREACH)
1004
		IP6_INC_STATS(net, NULL, IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
1005 1006 1007 1008
	dst_release(*dst);
	*dst = NULL;
	return err;
}
A
Adrian Bunk 已提交
1009

1010 1011 1012 1013
/**
 *	ip6_dst_lookup - perform route lookup on flow
 *	@sk: socket which provides route info
 *	@dst: pointer to dst_entry * for result
1014
 *	@fl6: flow to lookup
1015 1016 1017 1018 1019
 *
 *	This function performs a route lookup on the given flow.
 *
 *	It returns zero on success, or a standard errno code on error.
 */
1020 1021
int ip6_dst_lookup(struct net *net, struct sock *sk, struct dst_entry **dst,
		   struct flowi6 *fl6)
1022 1023
{
	*dst = NULL;
1024
	return ip6_dst_lookup_tail(net, sk, dst, fl6);
1025
}
1026 1027
EXPORT_SYMBOL_GPL(ip6_dst_lookup);

1028
/**
1029 1030
 *	ip6_dst_lookup_flow - perform route lookup on flow with ipsec
 *	@sk: socket which provides route info
1031
 *	@fl6: flow to lookup
1032 1033 1034 1035 1036 1037 1038
 *	@final_dst: final destination address for ipsec lookup
 *
 *	This function performs a route lookup on the given flow.
 *
 *	It returns a valid dst pointer on success, or a pointer encoded
 *	error code.
 */
1039
struct dst_entry *ip6_dst_lookup_flow(const struct sock *sk, struct flowi6 *fl6,
1040
				      const struct in6_addr *final_dst)
1041 1042 1043 1044
{
	struct dst_entry *dst = NULL;
	int err;

1045
	err = ip6_dst_lookup_tail(sock_net(sk), sk, &dst, fl6);
1046 1047 1048
	if (err)
		return ERR_PTR(err);
	if (final_dst)
A
Alexey Dobriyan 已提交
1049
		fl6->daddr = *final_dst;
1050
	if (!fl6->flowi6_oif)
D
David Ahern 已提交
1051
		fl6->flowi6_oif = l3mdev_fib_oif(dst->dev);
1052

1053
	return xfrm_lookup_route(sock_net(sk), dst, flowi6_to_flowi(fl6), sk, 0);
1054 1055 1056 1057 1058
}
EXPORT_SYMBOL_GPL(ip6_dst_lookup_flow);

/**
 *	ip6_sk_dst_lookup_flow - perform socket cached route lookup on flow
1059
 *	@sk: socket which provides the dst cache and route info
1060
 *	@fl6: flow to lookup
1061
 *	@final_dst: final destination address for ipsec lookup
1062 1063 1064 1065 1066 1067
 *
 *	This function performs a route lookup on the given flow with the
 *	possibility of using the cached route in the socket if it is valid.
 *	It will take the socket dst lock when operating on the dst cache.
 *	As a result, this function can only be used in process context.
 *
1068 1069
 *	It returns a valid dst pointer on success, or a pointer encoded
 *	error code.
1070
 */
1071
struct dst_entry *ip6_sk_dst_lookup_flow(struct sock *sk, struct flowi6 *fl6,
1072
					 const struct in6_addr *final_dst)
1073
{
1074 1075
	struct dst_entry *dst = sk_dst_check(sk, inet6_sk(sk)->dst_cookie);
	int err;
1076

1077
	dst = ip6_sk_dst_check(sk, dst, fl6);
1078

1079
	err = ip6_dst_lookup_tail(sock_net(sk), sk, &dst, fl6);
1080 1081 1082
	if (err)
		return ERR_PTR(err);
	if (final_dst)
A
Alexey Dobriyan 已提交
1083
		fl6->daddr = *final_dst;
1084

1085
	return xfrm_lookup_route(sock_net(sk), dst, flowi6_to_flowi(fl6), sk, 0);
1086
}
1087
EXPORT_SYMBOL_GPL(ip6_sk_dst_lookup_flow);
1088

A
Adrian Bunk 已提交
1089
static inline int ip6_ufo_append_data(struct sock *sk,
1090
			struct sk_buff_head *queue,
1091 1092 1093
			int getfrag(void *from, char *to, int offset, int len,
			int odd, struct sk_buff *skb),
			void *from, int length, int hh_len, int fragheaderlen,
1094
			int transhdrlen, int mtu, unsigned int flags,
1095
			const struct flowi6 *fl6)
1096 1097 1098 1099 1100 1101 1102 1103 1104

{
	struct sk_buff *skb;
	int err;

	/* There is support for UDP large send offload by network
	 * device, so create one single skb packet containing complete
	 * udp datagram
	 */
1105
	skb = skb_peek_tail(queue);
1106
	if (!skb) {
1107 1108 1109
		skb = sock_alloc_send_skb(sk,
			hh_len + fragheaderlen + transhdrlen + 20,
			(flags & MSG_DONTWAIT), &err);
1110
		if (!skb)
1111
			return err;
1112 1113 1114 1115 1116

		/* reserve space for Hardware header */
		skb_reserve(skb, hh_len);

		/* create space for UDP/IP header */
1117
		skb_put(skb, fragheaderlen + transhdrlen);
1118 1119

		/* initialize network header pointer */
1120
		skb_reset_network_header(skb);
1121 1122

		/* initialize protocol header pointer */
1123
		skb->transport_header = skb->network_header + fragheaderlen;
1124

1125
		skb->protocol = htons(ETH_P_IPV6);
1126 1127
		skb->csum = 0;

1128
		__skb_queue_tail(queue, skb);
1129 1130
	} else if (skb_is_gso(skb)) {
		goto append;
1131 1132
	}

1133 1134 1135 1136 1137 1138 1139
	skb->ip_summed = CHECKSUM_PARTIAL;
	/* Specify the length of each IPv6 datagram fragment.
	 * It has to be a multiple of 8.
	 */
	skb_shinfo(skb)->gso_size = (mtu - fragheaderlen -
				     sizeof(struct frag_hdr)) & ~7;
	skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
1140 1141 1142
	skb_shinfo(skb)->ip6_frag_id = ipv6_select_ident(sock_net(sk),
							 &fl6->daddr,
							 &fl6->saddr);
1143 1144

append:
1145 1146
	return skb_append_datato_frags(sk, skb, getfrag, from,
				       (length - transhdrlen));
1147
}
L
Linus Torvalds 已提交
1148

1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
static inline struct ipv6_opt_hdr *ip6_opt_dup(struct ipv6_opt_hdr *src,
					       gfp_t gfp)
{
	return src ? kmemdup(src, (src->hdrlen + 1) * 8, gfp) : NULL;
}

static inline struct ipv6_rt_hdr *ip6_rthdr_dup(struct ipv6_rt_hdr *src,
						gfp_t gfp)
{
	return src ? kmemdup(src, (src->hdrlen + 1) * 8, gfp) : NULL;
}

1161
static void ip6_append_data_mtu(unsigned int *mtu,
G
Gao feng 已提交
1162 1163 1164
				int *maxfraglen,
				unsigned int fragheaderlen,
				struct sk_buff *skb,
1165
				struct rt6_info *rt,
1166
				unsigned int orig_mtu)
G
Gao feng 已提交
1167 1168
{
	if (!(rt->dst.flags & DST_XFRM_TUNNEL)) {
1169
		if (!skb) {
G
Gao feng 已提交
1170
			/* first fragment, reserve header_len */
1171
			*mtu = orig_mtu - rt->dst.header_len;
G
Gao feng 已提交
1172 1173 1174 1175 1176 1177

		} else {
			/*
			 * this fragment is not first, the headers
			 * space is regarded as data space.
			 */
1178
			*mtu = orig_mtu;
G
Gao feng 已提交
1179 1180 1181 1182 1183 1184
		}
		*maxfraglen = ((*mtu - fragheaderlen) & ~7)
			      + fragheaderlen - sizeof(struct frag_hdr);
	}
}

1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200
static int ip6_setup_cork(struct sock *sk, struct inet_cork_full *cork,
			  struct inet6_cork *v6_cork,
			  int hlimit, int tclass, struct ipv6_txoptions *opt,
			  struct rt6_info *rt, struct flowi6 *fl6)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	unsigned int mtu;

	/*
	 * setup for corking
	 */
	if (opt) {
		if (WARN_ON(v6_cork->opt))
			return -EINVAL;

		v6_cork->opt = kzalloc(opt->tot_len, sk->sk_allocation);
1201
		if (unlikely(!v6_cork->opt))
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
			return -ENOBUFS;

		v6_cork->opt->tot_len = opt->tot_len;
		v6_cork->opt->opt_flen = opt->opt_flen;
		v6_cork->opt->opt_nflen = opt->opt_nflen;

		v6_cork->opt->dst0opt = ip6_opt_dup(opt->dst0opt,
						    sk->sk_allocation);
		if (opt->dst0opt && !v6_cork->opt->dst0opt)
			return -ENOBUFS;

		v6_cork->opt->dst1opt = ip6_opt_dup(opt->dst1opt,
						    sk->sk_allocation);
		if (opt->dst1opt && !v6_cork->opt->dst1opt)
			return -ENOBUFS;

		v6_cork->opt->hopopt = ip6_opt_dup(opt->hopopt,
						   sk->sk_allocation);
		if (opt->hopopt && !v6_cork->opt->hopopt)
			return -ENOBUFS;

		v6_cork->opt->srcrt = ip6_rthdr_dup(opt->srcrt,
						    sk->sk_allocation);
		if (opt->srcrt && !v6_cork->opt->srcrt)
			return -ENOBUFS;

		/* need source address above miyazawa*/
	}
	dst_hold(&rt->dst);
	cork->base.dst = &rt->dst;
	cork->fl.u.ip6 = *fl6;
	v6_cork->hop_limit = hlimit;
	v6_cork->tclass = tclass;
	if (rt->dst.flags & DST_XFRM_TUNNEL)
		mtu = np->pmtudisc >= IPV6_PMTUDISC_PROBE ?
		      rt->dst.dev->mtu : dst_mtu(&rt->dst);
	else
		mtu = np->pmtudisc >= IPV6_PMTUDISC_PROBE ?
		      rt->dst.dev->mtu : dst_mtu(rt->dst.path);
	if (np->frag_size < mtu) {
		if (np->frag_size)
			mtu = np->frag_size;
	}
	cork->base.fragsize = mtu;
	if (dst_allfrag(rt->dst.path))
		cork->base.flags |= IPCORK_ALLFRAG;
	cork->base.length = 0;

	return 0;
}

1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
static int __ip6_append_data(struct sock *sk,
			     struct flowi6 *fl6,
			     struct sk_buff_head *queue,
			     struct inet_cork *cork,
			     struct inet6_cork *v6_cork,
			     struct page_frag *pfrag,
			     int getfrag(void *from, char *to, int offset,
					 int len, int odd, struct sk_buff *skb),
			     void *from, int length, int transhdrlen,
			     unsigned int flags, int dontfrag)
L
Linus Torvalds 已提交
1263
{
G
Gao feng 已提交
1264
	struct sk_buff *skb, *skb_prev = NULL;
1265
	unsigned int maxfraglen, fragheaderlen, mtu, orig_mtu;
1266 1267
	int exthdrlen = 0;
	int dst_exthdrlen = 0;
L
Linus Torvalds 已提交
1268 1269 1270 1271
	int hh_len;
	int copy;
	int err;
	int offset = 0;
1272
	__u8 tx_flags = 0;
1273
	u32 tskey = 0;
1274 1275
	struct rt6_info *rt = (struct rt6_info *)cork->dst;
	struct ipv6_txoptions *opt = v6_cork->opt;
1276
	int csummode = CHECKSUM_NONE;
1277
	unsigned int maxnonfragsize, headersize;
L
Linus Torvalds 已提交
1278

1279 1280 1281
	skb = skb_peek_tail(queue);
	if (!skb) {
		exthdrlen = opt ? opt->opt_flen : 0;
1282
		dst_exthdrlen = rt->dst.header_len - rt->rt6i_nfheader_len;
L
Linus Torvalds 已提交
1283
	}
1284

1285
	mtu = cork->fragsize;
1286
	orig_mtu = mtu;
L
Linus Torvalds 已提交
1287

1288
	hh_len = LL_RESERVED_SPACE(rt->dst.dev);
L
Linus Torvalds 已提交
1289

1290
	fragheaderlen = sizeof(struct ipv6hdr) + rt->rt6i_nfheader_len +
1291
			(opt ? opt->opt_nflen : 0);
1292 1293
	maxfraglen = ((mtu - fragheaderlen) & ~7) + fragheaderlen -
		     sizeof(struct frag_hdr);
L
Linus Torvalds 已提交
1294

1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307
	headersize = sizeof(struct ipv6hdr) +
		     (opt ? opt->opt_flen + opt->opt_nflen : 0) +
		     (dst_allfrag(&rt->dst) ?
		      sizeof(struct frag_hdr) : 0) +
		     rt->rt6i_nfheader_len;

	if (cork->length + length > mtu - headersize && dontfrag &&
	    (sk->sk_protocol == IPPROTO_UDP ||
	     sk->sk_protocol == IPPROTO_RAW)) {
		ipv6_local_rxpmtu(sk, fl6, mtu - headersize +
				sizeof(struct ipv6hdr));
		goto emsgsize;
	}
1308

1309 1310 1311 1312
	if (ip6_sk_ignore_df(sk))
		maxnonfragsize = sizeof(struct ipv6hdr) + IPV6_MAXPLEN;
	else
		maxnonfragsize = mtu;
1313

1314
	if (cork->length + length > maxnonfragsize - headersize) {
1315
emsgsize:
1316 1317 1318 1319
		ipv6_local_error(sk, EMSGSIZE, fl6,
				 mtu - headersize +
				 sizeof(struct ipv6hdr));
		return -EMSGSIZE;
L
Linus Torvalds 已提交
1320 1321
	}

1322 1323 1324 1325 1326 1327 1328
	/* CHECKSUM_PARTIAL only with no extension headers and when
	 * we are not going to fragment
	 */
	if (transhdrlen && sk->sk_protocol == IPPROTO_UDP &&
	    headersize == sizeof(struct ipv6hdr) &&
	    length < mtu - headersize &&
	    !(flags & MSG_MORE) &&
1329
	    rt->dst.dev->features & (NETIF_F_IPV6_CSUM | NETIF_F_HW_CSUM))
1330 1331
		csummode = CHECKSUM_PARTIAL;

1332
	if (sk->sk_type == SOCK_DGRAM || sk->sk_type == SOCK_RAW) {
1333
		sock_tx_timestamp(sk, &tx_flags);
1334 1335 1336 1337
		if (tx_flags & SKBTX_ANY_SW_TSTAMP &&
		    sk->sk_tsflags & SOF_TIMESTAMPING_OPT_ID)
			tskey = sk->sk_tskey++;
	}
1338

L
Linus Torvalds 已提交
1339 1340 1341 1342 1343 1344 1345
	/*
	 * Let's try using as much space as possible.
	 * Use MTU if total length of the message fits into the MTU.
	 * Otherwise, we need to reserve fragment header and
	 * fragment alignment (= 8-15 octects, in total).
	 *
	 * Note that we may need to "move" the data from the tail of
1346
	 * of the buffer to the new fragment when we split
L
Linus Torvalds 已提交
1347 1348
	 * the message.
	 *
1349
	 * FIXME: It may be fragmented into multiple chunks
L
Linus Torvalds 已提交
1350 1351
	 *        at once if non-fragmentable extension headers
	 *        are too large.
1352
	 * --yoshfuji
L
Linus Torvalds 已提交
1353 1354
	 */

1355 1356 1357 1358
	cork->length += length;
	if (((length > mtu) ||
	     (skb && skb_is_gso(skb))) &&
	    (sk->sk_protocol == IPPROTO_UDP) &&
1359
	    (rt->dst.dev->features & NETIF_F_UFO) &&
1360
	    (sk->sk_type == SOCK_DGRAM) && !udp_get_no_check6_tx(sk)) {
1361
		err = ip6_ufo_append_data(sk, queue, getfrag, from, length,
1362
					  hh_len, fragheaderlen,
1363
					  transhdrlen, mtu, flags, fl6);
1364 1365 1366
		if (err)
			goto error;
		return 0;
1367
	}
L
Linus Torvalds 已提交
1368

1369
	if (!skb)
L
Linus Torvalds 已提交
1370 1371 1372 1373
		goto alloc_new_skb;

	while (length > 0) {
		/* Check if the remaining data fits into current packet. */
1374
		copy = (cork->length <= mtu && !(cork->flags & IPCORK_ALLFRAG) ? mtu : maxfraglen) - skb->len;
L
Linus Torvalds 已提交
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385
		if (copy < length)
			copy = maxfraglen - skb->len;

		if (copy <= 0) {
			char *data;
			unsigned int datalen;
			unsigned int fraglen;
			unsigned int fraggap;
			unsigned int alloclen;
alloc_new_skb:
			/* There's no room in the current skb */
G
Gao feng 已提交
1386 1387
			if (skb)
				fraggap = skb->len - maxfraglen;
L
Linus Torvalds 已提交
1388 1389
			else
				fraggap = 0;
G
Gao feng 已提交
1390
			/* update mtu and maxfraglen if necessary */
1391
			if (!skb || !skb_prev)
G
Gao feng 已提交
1392
				ip6_append_data_mtu(&mtu, &maxfraglen,
1393
						    fragheaderlen, skb, rt,
1394
						    orig_mtu);
G
Gao feng 已提交
1395 1396

			skb_prev = skb;
L
Linus Torvalds 已提交
1397 1398 1399 1400 1401 1402 1403

			/*
			 * If remaining data exceeds the mtu,
			 * we know we need more fragment(s).
			 */
			datalen = length + fraggap;

G
Gao feng 已提交
1404 1405
			if (datalen > (cork->length <= mtu && !(cork->flags & IPCORK_ALLFRAG) ? mtu : maxfraglen) - fragheaderlen)
				datalen = maxfraglen - fragheaderlen - rt->dst.trailer_len;
L
Linus Torvalds 已提交
1406
			if ((flags & MSG_MORE) &&
1407
			    !(rt->dst.dev->features&NETIF_F_SG))
L
Linus Torvalds 已提交
1408 1409 1410 1411
				alloclen = mtu;
			else
				alloclen = datalen + fragheaderlen;

1412 1413
			alloclen += dst_exthdrlen;

G
Gao feng 已提交
1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
			if (datalen != length + fraggap) {
				/*
				 * this is not the last fragment, the trailer
				 * space is regarded as data space.
				 */
				datalen += rt->dst.trailer_len;
			}

			alloclen += rt->dst.trailer_len;
			fraglen = datalen + fragheaderlen;
L
Linus Torvalds 已提交
1424 1425 1426

			/*
			 * We just reserve space for fragment header.
1427
			 * Note: this may be overallocation if the message
L
Linus Torvalds 已提交
1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442
			 * (without MSG_MORE) fits into the MTU.
			 */
			alloclen += sizeof(struct frag_hdr);

			if (transhdrlen) {
				skb = sock_alloc_send_skb(sk,
						alloclen + hh_len,
						(flags & MSG_DONTWAIT), &err);
			} else {
				skb = NULL;
				if (atomic_read(&sk->sk_wmem_alloc) <=
				    2 * sk->sk_sndbuf)
					skb = sock_wmalloc(sk,
							   alloclen + hh_len, 1,
							   sk->sk_allocation);
1443
				if (unlikely(!skb))
L
Linus Torvalds 已提交
1444 1445
					err = -ENOBUFS;
			}
1446
			if (!skb)
L
Linus Torvalds 已提交
1447 1448 1449 1450
				goto error;
			/*
			 *	Fill in the control structures
			 */
1451
			skb->protocol = htons(ETH_P_IPV6);
1452
			skb->ip_summed = csummode;
L
Linus Torvalds 已提交
1453
			skb->csum = 0;
1454 1455 1456
			/* reserve for fragmentation and ipsec header */
			skb_reserve(skb, hh_len + sizeof(struct frag_hdr) +
				    dst_exthdrlen);
L
Linus Torvalds 已提交
1457

1458 1459 1460
			/* Only the initial fragment is time stamped */
			skb_shinfo(skb)->tx_flags = tx_flags;
			tx_flags = 0;
1461 1462
			skb_shinfo(skb)->tskey = tskey;
			tskey = 0;
1463

L
Linus Torvalds 已提交
1464 1465 1466
			/*
			 *	Find where to start putting bytes
			 */
1467 1468 1469
			data = skb_put(skb, fraglen);
			skb_set_network_header(skb, exthdrlen);
			data += fragheaderlen;
1470 1471
			skb->transport_header = (skb->network_header +
						 fragheaderlen);
L
Linus Torvalds 已提交
1472 1473 1474 1475 1476 1477 1478
			if (fraggap) {
				skb->csum = skb_copy_and_csum_bits(
					skb_prev, maxfraglen,
					data + transhdrlen, fraggap, 0);
				skb_prev->csum = csum_sub(skb_prev->csum,
							  skb->csum);
				data += fraggap;
1479
				pskb_trim_unique(skb_prev, maxfraglen);
L
Linus Torvalds 已提交
1480 1481
			}
			copy = datalen - transhdrlen - fraggap;
1482

L
Linus Torvalds 已提交
1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496
			if (copy < 0) {
				err = -EINVAL;
				kfree_skb(skb);
				goto error;
			} else if (copy > 0 && getfrag(from, data + transhdrlen, offset, copy, fraggap, skb) < 0) {
				err = -EFAULT;
				kfree_skb(skb);
				goto error;
			}

			offset += copy;
			length -= datalen - fraggap;
			transhdrlen = 0;
			exthdrlen = 0;
1497
			dst_exthdrlen = 0;
L
Linus Torvalds 已提交
1498 1499 1500 1501

			/*
			 * Put the packet on the pending queue
			 */
1502
			__skb_queue_tail(queue, skb);
L
Linus Torvalds 已提交
1503 1504 1505 1506 1507 1508
			continue;
		}

		if (copy > length)
			copy = length;

1509
		if (!(rt->dst.dev->features&NETIF_F_SG)) {
L
Linus Torvalds 已提交
1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521
			unsigned int off;

			off = skb->len;
			if (getfrag(from, skb_put(skb, copy),
						offset, copy, off, skb) < 0) {
				__skb_trim(skb, off);
				err = -EFAULT;
				goto error;
			}
		} else {
			int i = skb_shinfo(skb)->nr_frags;

1522 1523
			err = -ENOMEM;
			if (!sk_page_frag_refill(sk, pfrag))
L
Linus Torvalds 已提交
1524
				goto error;
1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535

			if (!skb_can_coalesce(skb, i, pfrag->page,
					      pfrag->offset)) {
				err = -EMSGSIZE;
				if (i == MAX_SKB_FRAGS)
					goto error;

				__skb_fill_page_desc(skb, i, pfrag->page,
						     pfrag->offset, 0);
				skb_shinfo(skb)->nr_frags = ++i;
				get_page(pfrag->page);
L
Linus Torvalds 已提交
1536
			}
1537
			copy = min_t(int, copy, pfrag->size - pfrag->offset);
E
Eric Dumazet 已提交
1538
			if (getfrag(from,
1539 1540 1541 1542 1543 1544
				    page_address(pfrag->page) + pfrag->offset,
				    offset, copy, skb->len, skb) < 0)
				goto error_efault;

			pfrag->offset += copy;
			skb_frag_size_add(&skb_shinfo(skb)->frags[i - 1], copy);
L
Linus Torvalds 已提交
1545 1546
			skb->len += copy;
			skb->data_len += copy;
1547 1548
			skb->truesize += copy;
			atomic_add(copy, &sk->sk_wmem_alloc);
L
Linus Torvalds 已提交
1549 1550 1551 1552
		}
		offset += copy;
		length -= copy;
	}
1553

L
Linus Torvalds 已提交
1554
	return 0;
1555 1556 1557

error_efault:
	err = -EFAULT;
L
Linus Torvalds 已提交
1558
error:
1559
	cork->length -= length;
1560
	IP6_INC_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
1561 1562
	return err;
}
1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598

int ip6_append_data(struct sock *sk,
		    int getfrag(void *from, char *to, int offset, int len,
				int odd, struct sk_buff *skb),
		    void *from, int length, int transhdrlen, int hlimit,
		    int tclass, struct ipv6_txoptions *opt, struct flowi6 *fl6,
		    struct rt6_info *rt, unsigned int flags, int dontfrag)
{
	struct inet_sock *inet = inet_sk(sk);
	struct ipv6_pinfo *np = inet6_sk(sk);
	int exthdrlen;
	int err;

	if (flags&MSG_PROBE)
		return 0;
	if (skb_queue_empty(&sk->sk_write_queue)) {
		/*
		 * setup for corking
		 */
		err = ip6_setup_cork(sk, &inet->cork, &np->cork, hlimit,
				     tclass, opt, rt, fl6);
		if (err)
			return err;

		exthdrlen = (opt ? opt->opt_flen : 0);
		length += exthdrlen;
		transhdrlen += exthdrlen;
	} else {
		fl6 = &inet->cork.fl.u.ip6;
		transhdrlen = 0;
	}

	return __ip6_append_data(sk, fl6, &sk->sk_write_queue, &inet->cork.base,
				 &np->cork, sk_page_frag(sk), getfrag,
				 from, length, transhdrlen, flags, dontfrag);
}
1599
EXPORT_SYMBOL_GPL(ip6_append_data);
L
Linus Torvalds 已提交
1600

1601 1602
static void ip6_cork_release(struct inet_cork_full *cork,
			     struct inet6_cork *v6_cork)
1603
{
1604 1605 1606 1607 1608 1609 1610
	if (v6_cork->opt) {
		kfree(v6_cork->opt->dst0opt);
		kfree(v6_cork->opt->dst1opt);
		kfree(v6_cork->opt->hopopt);
		kfree(v6_cork->opt->srcrt);
		kfree(v6_cork->opt);
		v6_cork->opt = NULL;
1611 1612
	}

1613 1614 1615 1616
	if (cork->base.dst) {
		dst_release(cork->base.dst);
		cork->base.dst = NULL;
		cork->base.flags &= ~IPCORK_ALLFRAG;
1617
	}
1618
	memset(&cork->fl, 0, sizeof(cork->fl));
1619 1620
}

V
Vlad Yasevich 已提交
1621 1622 1623 1624
struct sk_buff *__ip6_make_skb(struct sock *sk,
			       struct sk_buff_head *queue,
			       struct inet_cork_full *cork,
			       struct inet6_cork *v6_cork)
L
Linus Torvalds 已提交
1625 1626 1627 1628 1629
{
	struct sk_buff *skb, *tmp_skb;
	struct sk_buff **tail_skb;
	struct in6_addr final_dst_buf, *final_dst = &final_dst_buf;
	struct ipv6_pinfo *np = inet6_sk(sk);
1630
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1631
	struct ipv6hdr *hdr;
V
Vlad Yasevich 已提交
1632 1633 1634
	struct ipv6_txoptions *opt = v6_cork->opt;
	struct rt6_info *rt = (struct rt6_info *)cork->base.dst;
	struct flowi6 *fl6 = &cork->fl.u.ip6;
1635
	unsigned char proto = fl6->flowi6_proto;
L
Linus Torvalds 已提交
1636

V
Vlad Yasevich 已提交
1637
	skb = __skb_dequeue(queue);
1638
	if (!skb)
L
Linus Torvalds 已提交
1639 1640 1641 1642
		goto out;
	tail_skb = &(skb_shinfo(skb)->frag_list);

	/* move skb->data to ip header from ext header */
1643
	if (skb->data < skb_network_header(skb))
1644
		__skb_pull(skb, skb_network_offset(skb));
V
Vlad Yasevich 已提交
1645
	while ((tmp_skb = __skb_dequeue(queue)) != NULL) {
1646
		__skb_pull(tmp_skb, skb_network_header_len(skb));
L
Linus Torvalds 已提交
1647 1648 1649 1650 1651 1652 1653 1654 1655
		*tail_skb = tmp_skb;
		tail_skb = &(tmp_skb->next);
		skb->len += tmp_skb->len;
		skb->data_len += tmp_skb->len;
		skb->truesize += tmp_skb->truesize;
		tmp_skb->destructor = NULL;
		tmp_skb->sk = NULL;
	}

1656
	/* Allow local fragmentation. */
W
WANG Cong 已提交
1657
	skb->ignore_df = ip6_sk_ignore_df(sk);
1658

A
Alexey Dobriyan 已提交
1659
	*final_dst = fl6->daddr;
1660
	__skb_pull(skb, skb_network_header_len(skb));
L
Linus Torvalds 已提交
1661 1662 1663 1664 1665
	if (opt && opt->opt_flen)
		ipv6_push_frag_opts(skb, opt, &proto);
	if (opt && opt->opt_nflen)
		ipv6_push_nfrag_opts(skb, opt, &proto, &final_dst);

1666 1667
	skb_push(skb, sizeof(struct ipv6hdr));
	skb_reset_network_header(skb);
1668
	hdr = ipv6_hdr(skb);
1669

V
Vlad Yasevich 已提交
1670
	ip6_flow_hdr(hdr, v6_cork->tclass,
1671
		     ip6_make_flowlabel(net, skb, fl6->flowlabel,
1672
					np->autoflowlabel, fl6));
V
Vlad Yasevich 已提交
1673
	hdr->hop_limit = v6_cork->hop_limit;
L
Linus Torvalds 已提交
1674
	hdr->nexthdr = proto;
A
Alexey Dobriyan 已提交
1675 1676
	hdr->saddr = fl6->saddr;
	hdr->daddr = *final_dst;
L
Linus Torvalds 已提交
1677

1678
	skb->priority = sk->sk_priority;
1679
	skb->mark = sk->sk_mark;
1680

1681
	skb_dst_set(skb, dst_clone(&rt->dst));
1682
	IP6_UPD_PO_STATS(net, rt->rt6i_idev, IPSTATS_MIB_OUT, skb->len);
1683
	if (proto == IPPROTO_ICMPV6) {
E
Eric Dumazet 已提交
1684
		struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb));
1685

1686 1687
		ICMP6MSGOUT_INC_STATS(net, idev, icmp6_hdr(skb)->icmp6_type);
		ICMP6_INC_STATS(net, idev, ICMP6_MIB_OUTMSGS);
1688 1689
	}

V
Vlad Yasevich 已提交
1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700
	ip6_cork_release(cork, v6_cork);
out:
	return skb;
}

int ip6_send_skb(struct sk_buff *skb)
{
	struct net *net = sock_net(skb->sk);
	struct rt6_info *rt = (struct rt6_info *)skb_dst(skb);
	int err;

1701
	err = ip6_local_out(net, skb->sk, skb);
L
Linus Torvalds 已提交
1702 1703
	if (err) {
		if (err > 0)
E
Eric Dumazet 已提交
1704
			err = net_xmit_errno(err);
L
Linus Torvalds 已提交
1705
		if (err)
V
Vlad Yasevich 已提交
1706 1707
			IP6_INC_STATS(net, rt->rt6i_idev,
				      IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
1708 1709 1710
	}

	return err;
V
Vlad Yasevich 已提交
1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721
}

int ip6_push_pending_frames(struct sock *sk)
{
	struct sk_buff *skb;

	skb = ip6_finish_skb(sk);
	if (!skb)
		return 0;

	return ip6_send_skb(skb);
L
Linus Torvalds 已提交
1722
}
1723
EXPORT_SYMBOL_GPL(ip6_push_pending_frames);
L
Linus Torvalds 已提交
1724

1725
static void __ip6_flush_pending_frames(struct sock *sk,
V
Vlad Yasevich 已提交
1726 1727 1728
				       struct sk_buff_head *queue,
				       struct inet_cork_full *cork,
				       struct inet6_cork *v6_cork)
L
Linus Torvalds 已提交
1729 1730 1731
{
	struct sk_buff *skb;

1732
	while ((skb = __skb_dequeue_tail(queue)) != NULL) {
E
Eric Dumazet 已提交
1733 1734
		if (skb_dst(skb))
			IP6_INC_STATS(sock_net(sk), ip6_dst_idev(skb_dst(skb)),
1735
				      IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
1736 1737 1738
		kfree_skb(skb);
	}

V
Vlad Yasevich 已提交
1739
	ip6_cork_release(cork, v6_cork);
L
Linus Torvalds 已提交
1740
}
1741 1742 1743

void ip6_flush_pending_frames(struct sock *sk)
{
V
Vlad Yasevich 已提交
1744 1745
	__ip6_flush_pending_frames(sk, &sk->sk_write_queue,
				   &inet_sk(sk)->cork, &inet6_sk(sk)->cork);
1746
}
1747
EXPORT_SYMBOL_GPL(ip6_flush_pending_frames);
V
Vlad Yasevich 已提交
1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790

struct sk_buff *ip6_make_skb(struct sock *sk,
			     int getfrag(void *from, char *to, int offset,
					 int len, int odd, struct sk_buff *skb),
			     void *from, int length, int transhdrlen,
			     int hlimit, int tclass,
			     struct ipv6_txoptions *opt, struct flowi6 *fl6,
			     struct rt6_info *rt, unsigned int flags,
			     int dontfrag)
{
	struct inet_cork_full cork;
	struct inet6_cork v6_cork;
	struct sk_buff_head queue;
	int exthdrlen = (opt ? opt->opt_flen : 0);
	int err;

	if (flags & MSG_PROBE)
		return NULL;

	__skb_queue_head_init(&queue);

	cork.base.flags = 0;
	cork.base.addr = 0;
	cork.base.opt = NULL;
	v6_cork.opt = NULL;
	err = ip6_setup_cork(sk, &cork, &v6_cork, hlimit, tclass, opt, rt, fl6);
	if (err)
		return ERR_PTR(err);

	if (dontfrag < 0)
		dontfrag = inet6_sk(sk)->dontfrag;

	err = __ip6_append_data(sk, fl6, &queue, &cork.base, &v6_cork,
				&current->task_frag, getfrag, from,
				length + exthdrlen, transhdrlen + exthdrlen,
				flags, dontfrag);
	if (err) {
		__ip6_flush_pending_frames(sk, &queue, &cork, &v6_cork);
		return ERR_PTR(err);
	}

	return __ip6_make_skb(sk, &queue, &cork, &v6_cork);
}