ip6_output.c 43.9 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
#include <linux/bpf-cgroup.h>
L
Linus Torvalds 已提交
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
#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>
58
#include <linux/mroute6.h>
D
David Ahern 已提交
59
#include <net/l3mdev.h>
60
#include <net/lwtunnel.h>
L
Linus Torvalds 已提交
61

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

70
	if (ipv6_addr_is_multicast(&ipv6_hdr(skb)->daddr)) {
E
Eric Dumazet 已提交
71
		struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb));
L
Linus Torvalds 已提交
72

73
		if (!(dev->flags & IFF_LOOPBACK) && sk_mc_loop(sk) &&
74
		    ((mroute6_socket(net, skb) &&
75
		     !(IP6CB(skb)->flags & IP6SKB_FORWARDED)) ||
76 77
		     ipv6_chk_mcast_addr(dev, &ipv6_hdr(skb)->daddr,
					 &ipv6_hdr(skb)->saddr))) {
L
Linus Torvalds 已提交
78 79 80 81 82 83
			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)
84
				NF_HOOK(NFPROTO_IPV6, NF_INET_POST_ROUTING,
85
					net, sk, newskb, NULL, newskb->dev,
86
					dev_loopback_xmit);
L
Linus Torvalds 已提交
87

88
			if (ipv6_hdr(skb)->hop_limit == 0) {
89
				IP6_INC_STATS(net, idev,
90
					      IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
91 92 93 94 95
				kfree_skb(skb);
				return 0;
			}
		}

96
		IP6_UPD_PO_STATS(net, idev, IPSTATS_MIB_OUTMCAST, skb->len);
97 98 99 100 101 102 103

		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
104 105
	}

106 107 108 109 110 111 112
	if (lwtunnel_xmit_redirect(dst->lwtstate)) {
		int res = lwtunnel_xmit(skb);

		if (res < 0 || res == LWTUNNEL_XMIT_DONE)
			return res;
	}

113
	rcu_read_lock_bh();
114
	nexthop = rt6_nexthop((struct rt6_info *)dst, &ipv6_hdr(skb)->daddr);
115 116 117 118
	neigh = __ipv6_neigh_lookup_noref(dst->dev, nexthop);
	if (unlikely(!neigh))
		neigh = __neigh_create(&nd_tbl, nexthop, dst->dev, false);
	if (!IS_ERR(neigh)) {
119
		sock_confirm_neigh(skb, neigh);
120
		ret = neigh_output(neigh, skb);
121 122 123 124
		rcu_read_unlock_bh();
		return ret;
	}
	rcu_read_unlock_bh();
125

126
	IP6_INC_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTNOROUTES);
127 128
	kfree_skb(skb);
	return -EINVAL;
L
Linus Torvalds 已提交
129 130
}

131
static int ip6_finish_output(struct net *net, struct sock *sk, struct sk_buff *skb)
132
{
133 134 135 136 137 138 139 140
	int ret;

	ret = BPF_CGROUP_RUN_PROG_INET_EGRESS(sk, skb);
	if (ret) {
		kfree_skb(skb);
		return ret;
	}

141
	if ((skb->len > ip6_skb_dst_mtu(skb) && !skb_is_gso(skb)) ||
142 143
	    dst_allfrag(skb_dst(skb)) ||
	    (IP6CB(skb)->frag_max_size && skb->len > IP6CB(skb)->frag_max_size))
144
		return ip6_fragment(net, sk, skb, ip6_finish_output2);
145
	else
146
		return ip6_finish_output2(net, sk, skb);
147 148
}

E
Eric W. Biederman 已提交
149
int ip6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
L
Linus Torvalds 已提交
150
{
151
	struct net_device *dev = skb_dst(skb)->dev;
E
Eric Dumazet 已提交
152
	struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb));
153

154 155 156
	skb->protocol = htons(ETH_P_IPV6);
	skb->dev = dev;

157
	if (unlikely(idev->cnf.disable_ipv6)) {
158
		IP6_INC_STATS(net, idev, IPSTATS_MIB_OUTDISCARDS);
159 160 161 162
		kfree_skb(skb);
		return 0;
	}

163 164
	return NF_HOOK_COND(NFPROTO_IPV6, NF_INET_POST_ROUTING,
			    net, sk, skb, NULL, dev,
165 166
			    ip6_finish_output,
			    !(IP6CB(skb)->flags & IP6SKB_REROUTED));
L
Linus Torvalds 已提交
167 168 169
}

/*
170 171 172 173
 * 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
174
 */
175
int ip6_xmit(const struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6,
176
	     __u32 mark, struct ipv6_txoptions *opt, int tclass)
L
Linus Torvalds 已提交
177
{
178
	struct net *net = sock_net(sk);
179
	const struct ipv6_pinfo *np = inet6_sk(sk);
180
	struct in6_addr *first_hop = &fl6->daddr;
E
Eric Dumazet 已提交
181
	struct dst_entry *dst = skb_dst(skb);
L
Linus Torvalds 已提交
182
	struct ipv6hdr *hdr;
183
	u8  proto = fl6->flowi6_proto;
L
Linus Torvalds 已提交
184
	int seg_len = skb->len;
G
Gerrit Renker 已提交
185
	int hlimit = -1;
L
Linus Torvalds 已提交
186 187 188
	u32 mtu;

	if (opt) {
189
		unsigned int head_room;
L
Linus Torvalds 已提交
190 191 192 193 194 195 196 197 198 199

		/* 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);
200
			if (!skb2) {
E
Eric Dumazet 已提交
201
				IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
202 203
					      IPSTATS_MIB_OUTDISCARDS);
				kfree_skb(skb);
L
Linus Torvalds 已提交
204 205
				return -ENOBUFS;
			}
206
			consume_skb(skb);
207
			skb = skb2;
208 209 210 211
			/* 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
212 213 214 215
		}
		if (opt->opt_flen)
			ipv6_push_frag_opts(skb, opt, &proto);
		if (opt->opt_nflen)
216 217
			ipv6_push_nfrag_opts(skb, opt, &proto, &first_hop,
					     &fl6->saddr);
L
Linus Torvalds 已提交
218 219
	}

220 221
	skb_push(skb, sizeof(struct ipv6hdr));
	skb_reset_network_header(skb);
222
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
223 224 225 226

	/*
	 *	Fill in the IPv6 header
	 */
227
	if (np)
L
Linus Torvalds 已提交
228 229
		hlimit = np->hop_limit;
	if (hlimit < 0)
230
		hlimit = ip6_dst_hoplimit(dst);
L
Linus Torvalds 已提交
231

232
	ip6_flow_hdr(hdr, tclass, ip6_make_flowlabel(net, skb, fl6->flowlabel,
233
						     np->autoflowlabel, fl6));
234

L
Linus Torvalds 已提交
235 236 237 238
	hdr->payload_len = htons(seg_len);
	hdr->nexthdr = proto;
	hdr->hop_limit = hlimit;

A
Alexey Dobriyan 已提交
239 240
	hdr->saddr = fl6->saddr;
	hdr->daddr = *first_hop;
L
Linus Torvalds 已提交
241

242
	skb->protocol = htons(ETH_P_IPV6);
243
	skb->priority = sk->sk_priority;
244
	skb->mark = mark;
245

L
Linus Torvalds 已提交
246
	mtu = dst_mtu(dst);
W
WANG Cong 已提交
247
	if ((skb->len <= mtu) || skb->ignore_df || skb_is_gso(skb)) {
E
Eric Dumazet 已提交
248
		IP6_UPD_PO_STATS(net, ip6_dst_idev(skb_dst(skb)),
249
			      IPSTATS_MIB_OUT, skb->len);
250 251 252 253 254 255 256 257

		/* if egress device is enslaved to an L3 master device pass the
		 * skb to its handler for processing
		 */
		skb = l3mdev_ip6_out((struct sock *)sk, skb);
		if (unlikely(!skb))
			return 0;

258 259 260
		/* hooks should never assume socket lock is held.
		 * we promote our socket to non const
		 */
261
		return NF_HOOK(NFPROTO_IPV6, NF_INET_LOCAL_OUT,
262
			       net, (struct sock *)sk, skb, NULL, dst->dev,
263
			       dst_output);
L
Linus Torvalds 已提交
264 265 266
	}

	skb->dev = dst->dev;
267 268 269 270 271
	/* 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
272
	IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)), IPSTATS_MIB_FRAGFAILS);
L
Linus Torvalds 已提交
273 274 275
	kfree_skb(skb);
	return -EMSGSIZE;
}
276 277
EXPORT_SYMBOL(ip6_xmit);

L
Linus Torvalds 已提交
278 279 280 281 282 283 284 285
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;
286 287 288
		if (sk && ra->sel == sel &&
		    (!sk->sk_bound_dev_if ||
		     sk->sk_bound_dev_if == skb->dev->ifindex)) {
L
Linus Torvalds 已提交
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
			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;
}

307 308
static int ip6_forward_proxy_check(struct sk_buff *skb)
{
309
	struct ipv6hdr *hdr = ipv6_hdr(skb);
310
	u8 nexthdr = hdr->nexthdr;
311
	__be16 frag_off;
312 313 314
	int offset;

	if (ipv6_ext_hdr(nexthdr)) {
315
		offset = ipv6_skip_exthdr(skb, sizeof(*hdr), &nexthdr, &frag_off);
316 317 318 319 320 321 322 323
		if (offset < 0)
			return 0;
	} else
		offset = sizeof(struct ipv6hdr);

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

324 325
		if (!pskb_may_pull(skb, (skb_network_header(skb) +
					 offset + 1 - skb->data)))
326 327
			return 0;

328
		icmp6 = (struct icmp6hdr *)(skb_network_header(skb) + offset);
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345

		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;
		}
	}

346 347 348 349 350 351 352 353 354 355
	/*
	 * 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;
	}

356 357 358
	return 0;
}

359 360
static inline int ip6_forward_finish(struct net *net, struct sock *sk,
				     struct sk_buff *skb)
L
Linus Torvalds 已提交
361
{
362
	return dst_output(net, sk, skb);
L
Linus Torvalds 已提交
363 364
}

365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385
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;
}

386 387
static bool ip6_pkt_too_big(const struct sk_buff *skb, unsigned int mtu)
{
388
	if (skb->len <= mtu)
389 390
		return false;

W
WANG Cong 已提交
391
	/* ipv6 conntrack defrag sets max_frag_size + ignore_df */
392 393 394
	if (IP6CB(skb)->frag_max_size && IP6CB(skb)->frag_max_size > mtu)
		return true;

W
WANG Cong 已提交
395
	if (skb->ignore_df)
396 397
		return false;

398
	if (skb_is_gso(skb) && skb_gso_validate_mtu(skb, mtu))
399 400 401 402 403
		return false;

	return true;
}

L
Linus Torvalds 已提交
404 405
int ip6_forward(struct sk_buff *skb)
{
E
Eric Dumazet 已提交
406
	struct dst_entry *dst = skb_dst(skb);
407
	struct ipv6hdr *hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
408
	struct inet6_skb_parm *opt = IP6CB(skb);
409
	struct net *net = dev_net(dst->dev);
410
	u32 mtu;
411

412
	if (net->ipv6.devconf_all->forwarding == 0)
L
Linus Torvalds 已提交
413 414
		goto error;

415 416 417
	if (skb->pkt_type != PACKET_HOST)
		goto drop;

418 419 420
	if (unlikely(skb->sk))
		goto drop;

421 422 423
	if (skb_warn_if_lro(skb))
		goto drop;

L
Linus Torvalds 已提交
424
	if (!xfrm6_policy_check(NULL, XFRM_POLICY_FWD, skb)) {
E
Eric Dumazet 已提交
425 426
		__IP6_INC_STATS(net, ip6_dst_idev(dst),
				IPSTATS_MIB_INDISCARDS);
L
Linus Torvalds 已提交
427 428 429
		goto drop;
	}

430
	skb_forward_csum(skb);
L
Linus Torvalds 已提交
431 432 433 434 435 436 437 438 439 440 441 442 443 444

	/*
	 *	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
	 */
445 446
	if (unlikely(opt->flags & IP6SKB_ROUTERALERT)) {
		if (ip6_call_ra_chain(skb, ntohs(opt->ra)))
L
Linus Torvalds 已提交
447 448 449 450 451 452 453 454 455
			return 0;
	}

	/*
	 *	check and decrement ttl
	 */
	if (hdr->hop_limit <= 1) {
		/* Force OUTPUT device used as source address */
		skb->dev = dst->dev;
456
		icmpv6_send(skb, ICMPV6_TIME_EXCEED, ICMPV6_EXC_HOPLIMIT, 0);
E
Eric Dumazet 已提交
457 458
		__IP6_INC_STATS(net, ip6_dst_idev(dst),
				IPSTATS_MIB_INHDRERRORS);
L
Linus Torvalds 已提交
459 460 461 462 463

		kfree_skb(skb);
		return -ETIMEDOUT;
	}

464
	/* XXX: idev->cnf.proxy_ndp? */
465
	if (net->ipv6.devconf_all->proxy_ndp &&
466
	    pneigh_lookup(&nd_tbl, net, &hdr->daddr, skb->dev, 0)) {
467 468
		int proxied = ip6_forward_proxy_check(skb);
		if (proxied > 0)
469
			return ip6_input(skb);
470
		else if (proxied < 0) {
E
Eric Dumazet 已提交
471 472
			__IP6_INC_STATS(net, ip6_dst_idev(dst),
					IPSTATS_MIB_INDISCARDS);
473 474
			goto drop;
		}
475 476
	}

L
Linus Torvalds 已提交
477
	if (!xfrm6_route_forward(skb)) {
E
Eric Dumazet 已提交
478 479
		__IP6_INC_STATS(net, ip6_dst_idev(dst),
				IPSTATS_MIB_INDISCARDS);
L
Linus Torvalds 已提交
480 481
		goto drop;
	}
E
Eric Dumazet 已提交
482
	dst = skb_dst(skb);
L
Linus Torvalds 已提交
483 484 485

	/* IPv6 specs say nothing about it, but it is clear that we cannot
	   send redirects to source routed frames.
486
	   We don't send redirects to frames decapsulated from IPsec.
L
Linus Torvalds 已提交
487
	 */
488
	if (skb->dev == dst->dev && opt->srcrt == 0 && !skb_sec_path(skb)) {
L
Linus Torvalds 已提交
489
		struct in6_addr *target = NULL;
490
		struct inet_peer *peer;
L
Linus Torvalds 已提交
491 492 493 494 495 496 497 498
		struct rt6_info *rt;

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

		rt = (struct rt6_info *) dst;
499 500
		if (rt->rt6i_flags & RTF_GATEWAY)
			target = &rt->rt6i_gateway;
L
Linus Torvalds 已提交
501 502 503
		else
			target = &hdr->daddr;

504
		peer = inet_getpeer_v6(net->ipv6.peers, &hdr->daddr, 1);
505

L
Linus Torvalds 已提交
506 507 508
		/* Limit redirects both by destination (here)
		   and by source (inside ndisc_send_redirect)
		 */
509
		if (inet_peer_xrlim_allow(peer, 1*HZ))
510
			ndisc_send_redirect(skb, target);
511 512
		if (peer)
			inet_putpeer(peer);
513 514 515
	} else {
		int addrtype = ipv6_addr_type(&hdr->saddr);

L
Linus Torvalds 已提交
516
		/* This check is security critical. */
517 518
		if (addrtype == IPV6_ADDR_ANY ||
		    addrtype & (IPV6_ADDR_MULTICAST | IPV6_ADDR_LOOPBACK))
519 520 521
			goto error;
		if (addrtype & IPV6_ADDR_LINKLOCAL) {
			icmpv6_send(skb, ICMPV6_DEST_UNREACH,
522
				    ICMPV6_NOT_NEIGHBOUR, 0);
523 524
			goto error;
		}
L
Linus Torvalds 已提交
525 526
	}

527
	mtu = ip6_dst_mtu_forward(dst);
528 529 530
	if (mtu < IPV6_MIN_MTU)
		mtu = IPV6_MIN_MTU;

531
	if (ip6_pkt_too_big(skb, mtu)) {
L
Linus Torvalds 已提交
532 533
		/* Again, force OUTPUT device used as source address */
		skb->dev = dst->dev;
534
		icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
E
Eric Dumazet 已提交
535 536 537 538
		__IP6_INC_STATS(net, ip6_dst_idev(dst),
				IPSTATS_MIB_INTOOBIGERRORS);
		__IP6_INC_STATS(net, ip6_dst_idev(dst),
				IPSTATS_MIB_FRAGFAILS);
L
Linus Torvalds 已提交
539 540 541 542 543
		kfree_skb(skb);
		return -EMSGSIZE;
	}

	if (skb_cow(skb, dst->dev->hard_header_len)) {
E
Eric Dumazet 已提交
544 545
		__IP6_INC_STATS(net, ip6_dst_idev(dst),
				IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
546 547 548
		goto drop;
	}

549
	hdr = ipv6_hdr(skb);
L
Linus Torvalds 已提交
550 551

	/* Mangling hops number delayed to point after skb COW */
552

L
Linus Torvalds 已提交
553 554
	hdr->hop_limit--;

E
Eric Dumazet 已提交
555 556
	__IP6_INC_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTFORWDATAGRAMS);
	__IP6_ADD_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTOCTETS, skb->len);
557 558
	return NF_HOOK(NFPROTO_IPV6, NF_INET_FORWARD,
		       net, NULL, skb, skb->dev, dst->dev,
559
		       ip6_forward_finish);
L
Linus Torvalds 已提交
560 561

error:
E
Eric Dumazet 已提交
562
	__IP6_INC_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_INADDRERRORS);
L
Linus Torvalds 已提交
563 564 565 566 567 568 569 570 571 572
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
573 574
	skb_dst_drop(to);
	skb_dst_set(to, dst_clone(skb_dst(from)));
L
Linus Torvalds 已提交
575
	to->dev = from->dev;
T
Thomas Graf 已提交
576
	to->mark = from->mark;
L
Linus Torvalds 已提交
577 578 579 580

#ifdef CONFIG_NET_SCHED
	to->tc_index = from->tc_index;
#endif
581
	nf_copy(to, from);
582
	skb_copy_secmark(to, from);
L
Linus Torvalds 已提交
583 584
}

585 586
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
587 588
{
	struct sk_buff *frag;
589
	struct rt6_info *rt = (struct rt6_info *)skb_dst(skb);
590 591
	struct ipv6_pinfo *np = skb->sk && !dev_recursion_level() ?
				inet6_sk(skb->sk) : NULL;
L
Linus Torvalds 已提交
592 593 594
	struct ipv6hdr *tmp_hdr;
	struct frag_hdr *fh;
	unsigned int mtu, hlen, left, len;
595
	int hroom, troom;
596
	__be32 frag_id;
597
	int ptr, offset = 0, err = 0;
L
Linus Torvalds 已提交
598 599
	u8 *prevhdr, nexthdr = 0;

600 601
	err = ip6_find_1stfragopt(skb, &prevhdr);
	if (err < 0)
602
		goto fail;
603
	hlen = err;
L
Linus Torvalds 已提交
604 605
	nexthdr = *prevhdr;

J
John Heffner 已提交
606
	mtu = ip6_skb_dst_mtu(skb);
607 608

	/* We must not fragment if the socket is set to force MTU discovery
609
	 * or if the skb it not generated by a local socket.
610
	 */
611 612
	if (unlikely(!skb->ignore_df && skb->len > mtu))
		goto fail_toobig;
613

614 615 616 617 618 619 620 621
	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;
622 623
	}

624 625 626 627
	if (np && np->frag_size < mtu) {
		if (np->frag_size)
			mtu = np->frag_size;
	}
628
	if (mtu < hlen + sizeof(struct frag_hdr) + 8)
629
		goto fail_toobig;
630
	mtu -= hlen + sizeof(struct frag_hdr);
L
Linus Torvalds 已提交
631

632 633
	frag_id = ipv6_select_ident(net, &ipv6_hdr(skb)->daddr,
				    &ipv6_hdr(skb)->saddr);
634

635 636 637 638
	if (skb->ip_summed == CHECKSUM_PARTIAL &&
	    (err = skb_checksum_help(skb)))
		goto fail;

639
	hroom = LL_RESERVED_SPACE(rt->dst.dev);
640
	if (skb_has_frag_list(skb)) {
641
		unsigned int first_len = skb_pagelen(skb);
642
		struct sk_buff *frag2;
L
Linus Torvalds 已提交
643 644 645

		if (first_len - hlen > mtu ||
		    ((first_len - hlen) & 7) ||
646 647
		    skb_cloned(skb) ||
		    skb_headroom(skb) < (hroom + sizeof(struct frag_hdr)))
L
Linus Torvalds 已提交
648 649
			goto slow_path;

650
		skb_walk_frags(skb, frag) {
L
Linus Torvalds 已提交
651 652 653
			/* Correct geometry. */
			if (frag->len > mtu ||
			    ((frag->len & 7) && frag->next) ||
654
			    skb_headroom(frag) < (hlen + hroom + sizeof(struct frag_hdr)))
655
				goto slow_path_clean;
L
Linus Torvalds 已提交
656 657 658

			/* Partially cloned skb? */
			if (skb_shared(frag))
659
				goto slow_path_clean;
660 661 662 663 664 665

			BUG_ON(frag->sk);
			if (skb->sk) {
				frag->sk = skb->sk;
				frag->destructor = sock_wfree;
			}
666
			skb->truesize -= frag->truesize;
L
Linus Torvalds 已提交
667 668 669 670 671 672
		}

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

Y
YOSHIFUJI Hideaki 已提交
673
		*prevhdr = NEXTHDR_FRAGMENT;
674
		tmp_hdr = kmemdup(skb_network_header(skb), hlen, GFP_ATOMIC);
L
Linus Torvalds 已提交
675
		if (!tmp_hdr) {
676 677
			err = -ENOMEM;
			goto fail;
L
Linus Torvalds 已提交
678
		}
679 680
		frag = skb_shinfo(skb)->frag_list;
		skb_frag_list_init(skb);
L
Linus Torvalds 已提交
681 682

		__skb_pull(skb, hlen);
683
		fh = __skb_push(skb, sizeof(struct frag_hdr));
684 685
		__skb_push(skb, hlen);
		skb_reset_network_header(skb);
686
		memcpy(skb_network_header(skb), tmp_hdr, hlen);
L
Linus Torvalds 已提交
687 688 689 690

		fh->nexthdr = nexthdr;
		fh->reserved = 0;
		fh->frag_off = htons(IP6_MF);
691
		fh->identification = frag_id;
L
Linus Torvalds 已提交
692 693 694 695

		first_len = skb_pagelen(skb);
		skb->data_len = first_len - skb_headlen(skb);
		skb->len = first_len;
696 697
		ipv6_hdr(skb)->payload_len = htons(first_len -
						   sizeof(struct ipv6hdr));
698

L
Linus Torvalds 已提交
699 700 701 702 703
		for (;;) {
			/* Prepare header of the next frame,
			 * before previous one went down. */
			if (frag) {
				frag->ip_summed = CHECKSUM_NONE;
704
				skb_reset_transport_header(frag);
705
				fh = __skb_push(frag, sizeof(struct frag_hdr));
706 707
				__skb_push(frag, hlen);
				skb_reset_network_header(frag);
708 709
				memcpy(skb_network_header(frag), tmp_hdr,
				       hlen);
L
Linus Torvalds 已提交
710 711 712 713
				offset += skb->len - hlen - sizeof(struct frag_hdr);
				fh->nexthdr = nexthdr;
				fh->reserved = 0;
				fh->frag_off = htons(offset);
714
				if (frag->next)
L
Linus Torvalds 已提交
715 716
					fh->frag_off |= htons(IP6_MF);
				fh->identification = frag_id;
717 718 719
				ipv6_hdr(frag)->payload_len =
						htons(frag->len -
						      sizeof(struct ipv6hdr));
L
Linus Torvalds 已提交
720 721
				ip6_copy_metadata(frag, skb);
			}
722

723
			err = output(net, sk, skb);
724
			if (!err)
725
				IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
726
					      IPSTATS_MIB_FRAGCREATES);
727

L
Linus Torvalds 已提交
728 729 730 731 732 733 734 735
			if (err || !frag)
				break;

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

J
Jesper Juhl 已提交
736
		kfree(tmp_hdr);
L
Linus Torvalds 已提交
737 738

		if (err == 0) {
739
			IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
740
				      IPSTATS_MIB_FRAGOKS);
L
Linus Torvalds 已提交
741 742 743
			return 0;
		}

744
		kfree_skb_list(frag);
L
Linus Torvalds 已提交
745

746
		IP6_INC_STATS(net, ip6_dst_idev(&rt->dst),
747
			      IPSTATS_MIB_FRAGFAILS);
L
Linus Torvalds 已提交
748
		return err;
749 750 751 752 753 754 755 756 757

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
758 759 760 761 762 763 764 765 766 767
	}

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

	/*
	 *	Fragment the datagram.
	 */

768
	troom = rt->dst.dev->needed_tailroom;
L
Linus Torvalds 已提交
769 770 771 772

	/*
	 *	Keep copying data until we run out.
	 */
773
	while (left > 0)	{
774 775
		u8 *fragnexthdr_offset;

L
Linus Torvalds 已提交
776 777 778 779
		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
780
		/* IF: we are not sending up to and including the packet end
L
Linus Torvalds 已提交
781 782 783 784 785
		   then align the next start on an eight byte boundary */
		if (len < left)	{
			len &= ~7;
		}

786 787 788 789
		/* Allocate buffer */
		frag = alloc_skb(len + hlen + sizeof(struct frag_hdr) +
				 hroom + troom, GFP_ATOMIC);
		if (!frag) {
L
Linus Torvalds 已提交
790 791 792 793 794 795 796 797 798
			err = -ENOMEM;
			goto fail;
		}

		/*
		 *	Set up data on packet
		 */

		ip6_copy_metadata(frag, skb);
799
		skb_reserve(frag, hroom);
L
Linus Torvalds 已提交
800
		skb_put(frag, len + hlen + sizeof(struct frag_hdr));
801
		skb_reset_network_header(frag);
802
		fh = (struct frag_hdr *)(skb_network_header(frag) + hlen);
803 804
		frag->transport_header = (frag->network_header + hlen +
					  sizeof(struct frag_hdr));
L
Linus Torvalds 已提交
805 806 807 808 809 810 811 812 813 814 815

		/*
		 *	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.
		 */
816
		skb_copy_from_linear_data(skb, skb_network_header(frag), hlen);
L
Linus Torvalds 已提交
817

818 819 820 821
		fragnexthdr_offset = skb_network_header(frag);
		fragnexthdr_offset += prevhdr - skb_network_header(skb);
		*fragnexthdr_offset = NEXTHDR_FRAGMENT;

L
Linus Torvalds 已提交
822 823 824 825 826
		/*
		 *	Build fragment header.
		 */
		fh->nexthdr = nexthdr;
		fh->reserved = 0;
827
		fh->identification = frag_id;
L
Linus Torvalds 已提交
828 829 830 831

		/*
		 *	Copy a block of the IP datagram.
		 */
H
Himangi Saraogi 已提交
832 833
		BUG_ON(skb_copy_bits(skb, ptr, skb_transport_header(frag),
				     len));
L
Linus Torvalds 已提交
834 835 836 837 838
		left -= len;

		fh->frag_off = htons(offset);
		if (left > 0)
			fh->frag_off |= htons(IP6_MF);
839 840
		ipv6_hdr(frag)->payload_len = htons(frag->len -
						    sizeof(struct ipv6hdr));
L
Linus Torvalds 已提交
841 842 843 844 845 846 847

		ptr += len;
		offset += len;

		/*
		 *	Put this fragment into the sending queue.
		 */
848
		err = output(net, sk, frag);
L
Linus Torvalds 已提交
849 850
		if (err)
			goto fail;
851

E
Eric Dumazet 已提交
852
		IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
853
			      IPSTATS_MIB_FRAGCREATES);
L
Linus Torvalds 已提交
854
	}
E
Eric Dumazet 已提交
855
	IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
856
		      IPSTATS_MIB_FRAGOKS);
857
	consume_skb(skb);
L
Linus Torvalds 已提交
858 859
	return err;

860 861 862 863 864 865 866
fail_toobig:
	if (skb->sk && dst_allfrag(skb_dst(skb)))
		sk_nocaps_add(skb->sk, NETIF_F_GSO_MASK);

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

L
Linus Torvalds 已提交
867
fail:
E
Eric Dumazet 已提交
868
	IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),
869
		      IPSTATS_MIB_FRAGFAILS);
870
	kfree_skb(skb);
L
Linus Torvalds 已提交
871 872 873
	return err;
}

874 875 876
static inline int ip6_rt_check(const struct rt6key *rt_key,
			       const struct in6_addr *fl_addr,
			       const struct in6_addr *addr_cache)
877
{
E
Eric Dumazet 已提交
878
	return (rt_key->plen != 128 || !ipv6_addr_equal(fl_addr, &rt_key->addr)) &&
879
		(!addr_cache || !ipv6_addr_equal(fl_addr, addr_cache));
880 881
}

882 883
static struct dst_entry *ip6_sk_dst_check(struct sock *sk,
					  struct dst_entry *dst,
884
					  const struct flowi6 *fl6)
L
Linus Torvalds 已提交
885
{
886
	struct ipv6_pinfo *np = inet6_sk(sk);
887
	struct rt6_info *rt;
L
Linus Torvalds 已提交
888

889 890 891
	if (!dst)
		goto out;

892 893 894 895 896 897
	if (dst->ops->family != AF_INET6) {
		dst_release(dst);
		return NULL;
	}

	rt = (struct rt6_info *)dst;
898 899 900
	/* Yes, checking route validity in not connected
	 * case is not very simple. Take into account,
	 * that we do not support routing by source, TOS,
901
	 * and MSG_DONTROUTE		--ANK (980726)
902
	 *
903 904
	 * 1. ip6_rt_check(): If route was host route,
	 *    check that cached destination is current.
905 906 907 908 909 910 911 912 913 914
	 *    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.
	 */
915
	if (ip6_rt_check(&rt->rt6i_dst, &fl6->daddr, np->daddr_cache) ||
916
#ifdef CONFIG_IPV6_SUBTREES
917
	    ip6_rt_check(&rt->rt6i_src, &fl6->saddr, np->saddr_cache) ||
918
#endif
D
David Ahern 已提交
919 920
	   (!(fl6->flowi6_flags & FLOWI_FLAG_SKIP_NH_OIF) &&
	      (fl6->flowi6_oif && fl6->flowi6_oif != dst->dev->ifindex))) {
921 922
		dst_release(dst);
		dst = NULL;
L
Linus Torvalds 已提交
923 924
	}

925 926 927 928
out:
	return dst;
}

929
static int ip6_dst_lookup_tail(struct net *net, const struct sock *sk,
930
			       struct dst_entry **dst, struct flowi6 *fl6)
931
{
932 933
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
	struct neighbour *n;
934
	struct rt6_info *rt;
935 936
#endif
	int err;
937
	int flags = 0;
938

939 940 941 942 943 944 945 946 947 948 949 950
	/* 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
951

952 953 954
		if (!had_dst)
			*dst = ip6_route_output(net, sk, fl6);
		rt = (*dst)->error ? NULL : (struct rt6_info *)*dst;
955 956 957
		err = ip6_route_get_saddr(net, rt, &fl6->daddr,
					  sk ? inet6_sk(sk)->srcprefs : 0,
					  &fl6->saddr);
958
		if (err)
L
Linus Torvalds 已提交
959
			goto out_err_release;
960 961 962 963 964 965 966 967 968

		/* 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;
		}
969 970 971

		if (fl6->flowi6_oif)
			flags |= RT6_LOOKUP_F_IFACE;
L
Linus Torvalds 已提交
972 973
	}

974
	if (!*dst)
975
		*dst = ip6_route_output_flags(net, sk, fl6, flags);
976 977 978 979 980

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

981
#ifdef CONFIG_IPV6_OPTIMISTIC_DAD
982 983 984 985 986 987 988 989
	/*
	 * 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
	 */
990
	rt = (struct rt6_info *) *dst;
991
	rcu_read_lock_bh();
992 993
	n = __ipv6_neigh_lookup_noref(rt->dst.dev,
				      rt6_nexthop(rt, &fl6->daddr));
994 995 996 997
	err = n && !(n->nud_state & NUD_VALID) ? -EINVAL : 0;
	rcu_read_unlock_bh();

	if (err) {
998
		struct inet6_ifaddr *ifp;
999
		struct flowi6 fl_gw6;
1000 1001
		int redirect;

1002
		ifp = ipv6_get_ifaddr(net, &fl6->saddr,
1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
				      (*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);
1015 1016 1017
			memcpy(&fl_gw6, fl6, sizeof(struct flowi6));
			memset(&fl_gw6.daddr, 0, sizeof(struct in6_addr));
			*dst = ip6_route_output(net, sk, &fl_gw6);
1018 1019
			err = (*dst)->error;
			if (err)
1020
				goto out_err_release;
1021
		}
1022
	}
1023
#endif
1024
	if (ipv6_addr_v4mapped(&fl6->saddr) &&
1025 1026 1027 1028
	    !(ipv6_addr_v4mapped(&fl6->daddr) || ipv6_addr_any(&fl6->daddr))) {
		err = -EAFNOSUPPORT;
		goto out_err_release;
	}
1029

L
Linus Torvalds 已提交
1030 1031 1032 1033 1034
	return 0;

out_err_release:
	dst_release(*dst);
	*dst = NULL;
1035

1036 1037
	if (err == -ENETUNREACH)
		IP6_INC_STATS(net, NULL, IPSTATS_MIB_OUTNOROUTES);
L
Linus Torvalds 已提交
1038 1039
	return err;
}
A
Adrian Bunk 已提交
1040

1041 1042 1043 1044
/**
 *	ip6_dst_lookup - perform route lookup on flow
 *	@sk: socket which provides route info
 *	@dst: pointer to dst_entry * for result
1045
 *	@fl6: flow to lookup
1046 1047 1048 1049 1050
 *
 *	This function performs a route lookup on the given flow.
 *
 *	It returns zero on success, or a standard errno code on error.
 */
1051 1052
int ip6_dst_lookup(struct net *net, struct sock *sk, struct dst_entry **dst,
		   struct flowi6 *fl6)
1053 1054
{
	*dst = NULL;
1055
	return ip6_dst_lookup_tail(net, sk, dst, fl6);
1056
}
1057 1058
EXPORT_SYMBOL_GPL(ip6_dst_lookup);

1059
/**
1060 1061
 *	ip6_dst_lookup_flow - perform route lookup on flow with ipsec
 *	@sk: socket which provides route info
1062
 *	@fl6: flow to lookup
1063 1064 1065 1066 1067 1068 1069
 *	@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.
 */
1070
struct dst_entry *ip6_dst_lookup_flow(const struct sock *sk, struct flowi6 *fl6,
1071
				      const struct in6_addr *final_dst)
1072 1073 1074 1075
{
	struct dst_entry *dst = NULL;
	int err;

1076
	err = ip6_dst_lookup_tail(sock_net(sk), sk, &dst, fl6);
1077 1078 1079
	if (err)
		return ERR_PTR(err);
	if (final_dst)
A
Alexey Dobriyan 已提交
1080
		fl6->daddr = *final_dst;
1081

1082
	return xfrm_lookup_route(sock_net(sk), dst, flowi6_to_flowi(fl6), sk, 0);
1083 1084 1085 1086 1087
}
EXPORT_SYMBOL_GPL(ip6_dst_lookup_flow);

/**
 *	ip6_sk_dst_lookup_flow - perform socket cached route lookup on flow
1088
 *	@sk: socket which provides the dst cache and route info
1089
 *	@fl6: flow to lookup
1090
 *	@final_dst: final destination address for ipsec lookup
1091 1092 1093 1094 1095 1096
 *
 *	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.
 *
1097 1098
 *	It returns a valid dst pointer on success, or a pointer encoded
 *	error code.
1099
 */
1100
struct dst_entry *ip6_sk_dst_lookup_flow(struct sock *sk, struct flowi6 *fl6,
1101
					 const struct in6_addr *final_dst)
1102
{
1103
	struct dst_entry *dst = sk_dst_check(sk, inet6_sk(sk)->dst_cookie);
1104

1105
	dst = ip6_sk_dst_check(sk, dst, fl6);
1106 1107
	if (!dst)
		dst = ip6_dst_lookup_flow(sk, fl6, final_dst);
1108

1109
	return dst;
1110
}
1111
EXPORT_SYMBOL_GPL(ip6_sk_dst_lookup_flow);
1112

1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
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;
}

1125
static void ip6_append_data_mtu(unsigned int *mtu,
G
Gao feng 已提交
1126 1127 1128
				int *maxfraglen,
				unsigned int fragheaderlen,
				struct sk_buff *skb,
1129
				struct rt6_info *rt,
1130
				unsigned int orig_mtu)
G
Gao feng 已提交
1131 1132
{
	if (!(rt->dst.flags & DST_XFRM_TUNNEL)) {
1133
		if (!skb) {
G
Gao feng 已提交
1134
			/* first fragment, reserve header_len */
1135
			*mtu = orig_mtu - rt->dst.header_len;
G
Gao feng 已提交
1136 1137 1138 1139 1140 1141

		} else {
			/*
			 * this fragment is not first, the headers
			 * space is regarded as data space.
			 */
1142
			*mtu = orig_mtu;
G
Gao feng 已提交
1143 1144 1145 1146 1147 1148
		}
		*maxfraglen = ((*mtu - fragheaderlen) & ~7)
			      + fragheaderlen - sizeof(struct frag_hdr);
	}
}

1149
static int ip6_setup_cork(struct sock *sk, struct inet_cork_full *cork,
W
Wei Wang 已提交
1150
			  struct inet6_cork *v6_cork, struct ipcm6_cookie *ipc6,
1151 1152 1153 1154
			  struct rt6_info *rt, struct flowi6 *fl6)
{
	struct ipv6_pinfo *np = inet6_sk(sk);
	unsigned int mtu;
W
Wei Wang 已提交
1155
	struct ipv6_txoptions *opt = ipc6->opt;
1156 1157 1158 1159 1160 1161 1162 1163

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

1164
		v6_cork->opt = kzalloc(sizeof(*opt), sk->sk_allocation);
1165
		if (unlikely(!v6_cork->opt))
1166 1167
			return -ENOBUFS;

1168
		v6_cork->opt->tot_len = sizeof(*opt);
1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
		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;
W
Wei Wang 已提交
1197 1198
	v6_cork->hop_limit = ipc6->hlimit;
	v6_cork->tclass = ipc6->tclass;
1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
	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;
}

1217 1218 1219 1220 1221 1222 1223 1224 1225
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,
W
Wei Wang 已提交
1226
			     unsigned int flags, struct ipcm6_cookie *ipc6,
1227
			     const struct sockcm_cookie *sockc)
L
Linus Torvalds 已提交
1228
{
G
Gao feng 已提交
1229
	struct sk_buff *skb, *skb_prev = NULL;
1230
	unsigned int maxfraglen, fragheaderlen, mtu, orig_mtu;
1231 1232
	int exthdrlen = 0;
	int dst_exthdrlen = 0;
L
Linus Torvalds 已提交
1233 1234 1235 1236
	int hh_len;
	int copy;
	int err;
	int offset = 0;
1237
	__u8 tx_flags = 0;
1238
	u32 tskey = 0;
1239 1240
	struct rt6_info *rt = (struct rt6_info *)cork->dst;
	struct ipv6_txoptions *opt = v6_cork->opt;
1241
	int csummode = CHECKSUM_NONE;
1242
	unsigned int maxnonfragsize, headersize;
L
Linus Torvalds 已提交
1243

1244 1245 1246
	skb = skb_peek_tail(queue);
	if (!skb) {
		exthdrlen = opt ? opt->opt_flen : 0;
1247
		dst_exthdrlen = rt->dst.header_len - rt->rt6i_nfheader_len;
L
Linus Torvalds 已提交
1248
	}
1249

1250
	mtu = cork->fragsize;
1251
	orig_mtu = mtu;
L
Linus Torvalds 已提交
1252

1253
	hh_len = LL_RESERVED_SPACE(rt->dst.dev);
L
Linus Torvalds 已提交
1254

1255
	fragheaderlen = sizeof(struct ipv6hdr) + rt->rt6i_nfheader_len +
1256
			(opt ? opt->opt_nflen : 0);
1257 1258
	maxfraglen = ((mtu - fragheaderlen) & ~7) + fragheaderlen -
		     sizeof(struct frag_hdr);
L
Linus Torvalds 已提交
1259

1260 1261 1262 1263 1264 1265
	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;

W
Wei Wang 已提交
1266
	if (cork->length + length > mtu - headersize && ipc6->dontfrag &&
1267 1268 1269 1270 1271 1272
	    (sk->sk_protocol == IPPROTO_UDP ||
	     sk->sk_protocol == IPPROTO_RAW)) {
		ipv6_local_rxpmtu(sk, fl6, mtu - headersize +
				sizeof(struct ipv6hdr));
		goto emsgsize;
	}
1273

1274 1275 1276 1277
	if (ip6_sk_ignore_df(sk))
		maxnonfragsize = sizeof(struct ipv6hdr) + IPV6_MAXPLEN;
	else
		maxnonfragsize = mtu;
1278

1279
	if (cork->length + length > maxnonfragsize - headersize) {
1280
emsgsize:
1281 1282 1283 1284
		ipv6_local_error(sk, EMSGSIZE, fl6,
				 mtu - headersize +
				 sizeof(struct ipv6hdr));
		return -EMSGSIZE;
L
Linus Torvalds 已提交
1285 1286
	}

1287 1288 1289 1290 1291
	/* 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) &&
1292
	    length <= mtu - headersize &&
1293
	    !(flags & MSG_MORE) &&
1294
	    rt->dst.dev->features & (NETIF_F_IPV6_CSUM | NETIF_F_HW_CSUM))
1295 1296
		csummode = CHECKSUM_PARTIAL;

1297
	if (sk->sk_type == SOCK_DGRAM || sk->sk_type == SOCK_RAW) {
1298
		sock_tx_timestamp(sk, sockc->tsflags, &tx_flags);
1299 1300 1301 1302
		if (tx_flags & SKBTX_ANY_SW_TSTAMP &&
		    sk->sk_tsflags & SOF_TIMESTAMPING_OPT_ID)
			tskey = sk->sk_tskey++;
	}
1303

L
Linus Torvalds 已提交
1304 1305 1306 1307 1308 1309 1310
	/*
	 * 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
1311
	 * of the buffer to the new fragment when we split
L
Linus Torvalds 已提交
1312 1313
	 * the message.
	 *
1314
	 * FIXME: It may be fragmented into multiple chunks
L
Linus Torvalds 已提交
1315 1316
	 *        at once if non-fragmentable extension headers
	 *        are too large.
1317
	 * --yoshfuji
L
Linus Torvalds 已提交
1318 1319
	 */

1320 1321
	cork->length += length;
	if (!skb)
L
Linus Torvalds 已提交
1322 1323 1324 1325
		goto alloc_new_skb;

	while (length > 0) {
		/* Check if the remaining data fits into current packet. */
1326
		copy = (cork->length <= mtu && !(cork->flags & IPCORK_ALLFRAG) ? mtu : maxfraglen) - skb->len;
L
Linus Torvalds 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
		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
1338 1339
			if (skb)
				fraggap = skb->len - maxfraglen;
L
Linus Torvalds 已提交
1340 1341
			else
				fraggap = 0;
G
Gao feng 已提交
1342
			/* update mtu and maxfraglen if necessary */
1343
			if (!skb || !skb_prev)
G
Gao feng 已提交
1344
				ip6_append_data_mtu(&mtu, &maxfraglen,
1345
						    fragheaderlen, skb, rt,
1346
						    orig_mtu);
G
Gao feng 已提交
1347 1348

			skb_prev = skb;
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354 1355

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

G
Gao feng 已提交
1356 1357
			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
1358
			if ((flags & MSG_MORE) &&
1359
			    !(rt->dst.dev->features&NETIF_F_SG))
L
Linus Torvalds 已提交
1360 1361 1362 1363
				alloclen = mtu;
			else
				alloclen = datalen + fragheaderlen;

1364 1365
			alloclen += dst_exthdrlen;

G
Gao feng 已提交
1366 1367 1368 1369 1370 1371 1372 1373 1374 1375
			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
1376 1377 1378

			/*
			 * We just reserve space for fragment header.
1379
			 * Note: this may be overallocation if the message
L
Linus Torvalds 已提交
1380 1381 1382 1383
			 * (without MSG_MORE) fits into the MTU.
			 */
			alloclen += sizeof(struct frag_hdr);

1384 1385 1386 1387 1388
			copy = datalen - transhdrlen - fraggap;
			if (copy < 0) {
				err = -EINVAL;
				goto error;
			}
L
Linus Torvalds 已提交
1389 1390 1391 1392 1393 1394
			if (transhdrlen) {
				skb = sock_alloc_send_skb(sk,
						alloclen + hh_len,
						(flags & MSG_DONTWAIT), &err);
			} else {
				skb = NULL;
1395
				if (refcount_read(&sk->sk_wmem_alloc) <=
L
Linus Torvalds 已提交
1396 1397 1398 1399
				    2 * sk->sk_sndbuf)
					skb = sock_wmalloc(sk,
							   alloclen + hh_len, 1,
							   sk->sk_allocation);
1400
				if (unlikely(!skb))
L
Linus Torvalds 已提交
1401 1402
					err = -ENOBUFS;
			}
1403
			if (!skb)
L
Linus Torvalds 已提交
1404 1405 1406 1407
				goto error;
			/*
			 *	Fill in the control structures
			 */
1408
			skb->protocol = htons(ETH_P_IPV6);
1409
			skb->ip_summed = csummode;
L
Linus Torvalds 已提交
1410
			skb->csum = 0;
1411 1412 1413
			/* reserve for fragmentation and ipsec header */
			skb_reserve(skb, hh_len + sizeof(struct frag_hdr) +
				    dst_exthdrlen);
L
Linus Torvalds 已提交
1414

1415 1416 1417
			/* Only the initial fragment is time stamped */
			skb_shinfo(skb)->tx_flags = tx_flags;
			tx_flags = 0;
1418 1419
			skb_shinfo(skb)->tskey = tskey;
			tskey = 0;
1420

L
Linus Torvalds 已提交
1421 1422 1423
			/*
			 *	Find where to start putting bytes
			 */
1424 1425 1426
			data = skb_put(skb, fraglen);
			skb_set_network_header(skb, exthdrlen);
			data += fragheaderlen;
1427 1428
			skb->transport_header = (skb->network_header +
						 fragheaderlen);
L
Linus Torvalds 已提交
1429 1430 1431 1432 1433 1434 1435
			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;
1436
				pskb_trim_unique(skb_prev, maxfraglen);
L
Linus Torvalds 已提交
1437
			}
1438 1439 1440
			if (copy > 0 &&
			    getfrag(from, data + transhdrlen, offset,
				    copy, fraggap, skb) < 0) {
L
Linus Torvalds 已提交
1441 1442 1443 1444 1445 1446 1447 1448 1449
				err = -EFAULT;
				kfree_skb(skb);
				goto error;
			}

			offset += copy;
			length -= datalen - fraggap;
			transhdrlen = 0;
			exthdrlen = 0;
1450
			dst_exthdrlen = 0;
L
Linus Torvalds 已提交
1451

1452 1453 1454
			if ((flags & MSG_CONFIRM) && !skb_prev)
				skb_set_dst_pending_confirm(skb, 1);

L
Linus Torvalds 已提交
1455 1456 1457
			/*
			 * Put the packet on the pending queue
			 */
1458
			__skb_queue_tail(queue, skb);
L
Linus Torvalds 已提交
1459 1460 1461 1462 1463 1464
			continue;
		}

		if (copy > length)
			copy = length;

1465
		if (!(rt->dst.dev->features&NETIF_F_SG)) {
L
Linus Torvalds 已提交
1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477
			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;

1478 1479
			err = -ENOMEM;
			if (!sk_page_frag_refill(sk, pfrag))
L
Linus Torvalds 已提交
1480
				goto error;
1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491

			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
1492
			}
1493
			copy = min_t(int, copy, pfrag->size - pfrag->offset);
E
Eric Dumazet 已提交
1494
			if (getfrag(from,
1495 1496 1497 1498 1499 1500
				    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
1501 1502
			skb->len += copy;
			skb->data_len += copy;
1503
			skb->truesize += copy;
1504
			refcount_add(copy, &sk->sk_wmem_alloc);
L
Linus Torvalds 已提交
1505 1506 1507 1508
		}
		offset += copy;
		length -= copy;
	}
1509

L
Linus Torvalds 已提交
1510
	return 0;
1511 1512 1513

error_efault:
	err = -EFAULT;
L
Linus Torvalds 已提交
1514
error:
1515
	cork->length -= length;
1516
	IP6_INC_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
1517 1518
	return err;
}
1519 1520 1521 1522

int ip6_append_data(struct sock *sk,
		    int getfrag(void *from, char *to, int offset, int len,
				int odd, struct sk_buff *skb),
W
Wei Wang 已提交
1523 1524 1525
		    void *from, int length, int transhdrlen,
		    struct ipcm6_cookie *ipc6, struct flowi6 *fl6,
		    struct rt6_info *rt, unsigned int flags,
1526
		    const struct sockcm_cookie *sockc)
1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
{
	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
		 */
W
Wei Wang 已提交
1539 1540
		err = ip6_setup_cork(sk, &inet->cork, &np->cork,
				     ipc6, rt, fl6);
1541 1542 1543
		if (err)
			return err;

W
Wei Wang 已提交
1544
		exthdrlen = (ipc6->opt ? ipc6->opt->opt_flen : 0);
1545 1546 1547 1548 1549 1550 1551 1552 1553
		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,
W
Wei Wang 已提交
1554
				 from, length, transhdrlen, flags, ipc6, sockc);
1555
}
1556
EXPORT_SYMBOL_GPL(ip6_append_data);
L
Linus Torvalds 已提交
1557

1558 1559
static void ip6_cork_release(struct inet_cork_full *cork,
			     struct inet6_cork *v6_cork)
1560
{
1561 1562 1563 1564 1565 1566 1567
	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;
1568 1569
	}

1570 1571 1572 1573
	if (cork->base.dst) {
		dst_release(cork->base.dst);
		cork->base.dst = NULL;
		cork->base.flags &= ~IPCORK_ALLFRAG;
1574
	}
1575
	memset(&cork->fl, 0, sizeof(cork->fl));
1576 1577
}

V
Vlad Yasevich 已提交
1578 1579 1580 1581
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
1582 1583 1584 1585 1586
{
	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);
1587
	struct net *net = sock_net(sk);
L
Linus Torvalds 已提交
1588
	struct ipv6hdr *hdr;
V
Vlad Yasevich 已提交
1589 1590 1591
	struct ipv6_txoptions *opt = v6_cork->opt;
	struct rt6_info *rt = (struct rt6_info *)cork->base.dst;
	struct flowi6 *fl6 = &cork->fl.u.ip6;
1592
	unsigned char proto = fl6->flowi6_proto;
L
Linus Torvalds 已提交
1593

V
Vlad Yasevich 已提交
1594
	skb = __skb_dequeue(queue);
1595
	if (!skb)
L
Linus Torvalds 已提交
1596 1597 1598 1599
		goto out;
	tail_skb = &(skb_shinfo(skb)->frag_list);

	/* move skb->data to ip header from ext header */
1600
	if (skb->data < skb_network_header(skb))
1601
		__skb_pull(skb, skb_network_offset(skb));
V
Vlad Yasevich 已提交
1602
	while ((tmp_skb = __skb_dequeue(queue)) != NULL) {
1603
		__skb_pull(tmp_skb, skb_network_header_len(skb));
L
Linus Torvalds 已提交
1604 1605 1606 1607 1608 1609 1610 1611 1612
		*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;
	}

1613
	/* Allow local fragmentation. */
W
WANG Cong 已提交
1614
	skb->ignore_df = ip6_sk_ignore_df(sk);
1615

A
Alexey Dobriyan 已提交
1616
	*final_dst = fl6->daddr;
1617
	__skb_pull(skb, skb_network_header_len(skb));
L
Linus Torvalds 已提交
1618 1619 1620
	if (opt && opt->opt_flen)
		ipv6_push_frag_opts(skb, opt, &proto);
	if (opt && opt->opt_nflen)
1621
		ipv6_push_nfrag_opts(skb, opt, &proto, &final_dst, &fl6->saddr);
L
Linus Torvalds 已提交
1622

1623 1624
	skb_push(skb, sizeof(struct ipv6hdr));
	skb_reset_network_header(skb);
1625
	hdr = ipv6_hdr(skb);
1626

V
Vlad Yasevich 已提交
1627
	ip6_flow_hdr(hdr, v6_cork->tclass,
1628
		     ip6_make_flowlabel(net, skb, fl6->flowlabel,
1629
					np->autoflowlabel, fl6));
V
Vlad Yasevich 已提交
1630
	hdr->hop_limit = v6_cork->hop_limit;
L
Linus Torvalds 已提交
1631
	hdr->nexthdr = proto;
A
Alexey Dobriyan 已提交
1632 1633
	hdr->saddr = fl6->saddr;
	hdr->daddr = *final_dst;
L
Linus Torvalds 已提交
1634

1635
	skb->priority = sk->sk_priority;
1636
	skb->mark = sk->sk_mark;
1637

1638
	skb_dst_set(skb, dst_clone(&rt->dst));
1639
	IP6_UPD_PO_STATS(net, rt->rt6i_idev, IPSTATS_MIB_OUT, skb->len);
1640
	if (proto == IPPROTO_ICMPV6) {
E
Eric Dumazet 已提交
1641
		struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb));
1642

1643 1644
		ICMP6MSGOUT_INC_STATS(net, idev, icmp6_hdr(skb)->icmp6_type);
		ICMP6_INC_STATS(net, idev, ICMP6_MIB_OUTMSGS);
1645 1646
	}

V
Vlad Yasevich 已提交
1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
	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;

1658
	err = ip6_local_out(net, skb->sk, skb);
L
Linus Torvalds 已提交
1659 1660
	if (err) {
		if (err > 0)
E
Eric Dumazet 已提交
1661
			err = net_xmit_errno(err);
L
Linus Torvalds 已提交
1662
		if (err)
V
Vlad Yasevich 已提交
1663 1664
			IP6_INC_STATS(net, rt->rt6i_idev,
				      IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
1665 1666 1667
	}

	return err;
V
Vlad Yasevich 已提交
1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678
}

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
1679
}
1680
EXPORT_SYMBOL_GPL(ip6_push_pending_frames);
L
Linus Torvalds 已提交
1681

1682
static void __ip6_flush_pending_frames(struct sock *sk,
V
Vlad Yasevich 已提交
1683 1684 1685
				       struct sk_buff_head *queue,
				       struct inet_cork_full *cork,
				       struct inet6_cork *v6_cork)
L
Linus Torvalds 已提交
1686 1687 1688
{
	struct sk_buff *skb;

1689
	while ((skb = __skb_dequeue_tail(queue)) != NULL) {
E
Eric Dumazet 已提交
1690 1691
		if (skb_dst(skb))
			IP6_INC_STATS(sock_net(sk), ip6_dst_idev(skb_dst(skb)),
1692
				      IPSTATS_MIB_OUTDISCARDS);
L
Linus Torvalds 已提交
1693 1694 1695
		kfree_skb(skb);
	}

V
Vlad Yasevich 已提交
1696
	ip6_cork_release(cork, v6_cork);
L
Linus Torvalds 已提交
1697
}
1698 1699 1700

void ip6_flush_pending_frames(struct sock *sk)
{
V
Vlad Yasevich 已提交
1701 1702
	__ip6_flush_pending_frames(sk, &sk->sk_write_queue,
				   &inet_sk(sk)->cork, &inet6_sk(sk)->cork);
1703
}
1704
EXPORT_SYMBOL_GPL(ip6_flush_pending_frames);
V
Vlad Yasevich 已提交
1705 1706 1707 1708 1709

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,
W
Wei Wang 已提交
1710
			     struct ipcm6_cookie *ipc6, struct flowi6 *fl6,
V
Vlad Yasevich 已提交
1711
			     struct rt6_info *rt, unsigned int flags,
W
Wei Wang 已提交
1712
			     const struct sockcm_cookie *sockc)
V
Vlad Yasevich 已提交
1713 1714 1715 1716
{
	struct inet_cork_full cork;
	struct inet6_cork v6_cork;
	struct sk_buff_head queue;
W
Wei Wang 已提交
1717
	int exthdrlen = (ipc6->opt ? ipc6->opt->opt_flen : 0);
V
Vlad Yasevich 已提交
1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728
	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;
W
Wei Wang 已提交
1729
	err = ip6_setup_cork(sk, &cork, &v6_cork, ipc6, rt, fl6);
V
Vlad Yasevich 已提交
1730 1731 1732
	if (err)
		return ERR_PTR(err);

W
Wei Wang 已提交
1733 1734
	if (ipc6->dontfrag < 0)
		ipc6->dontfrag = inet6_sk(sk)->dontfrag;
V
Vlad Yasevich 已提交
1735 1736 1737 1738

	err = __ip6_append_data(sk, fl6, &queue, &cork.base, &v6_cork,
				&current->task_frag, getfrag, from,
				length + exthdrlen, transhdrlen + exthdrlen,
W
Wei Wang 已提交
1739
				flags, ipc6, sockc);
V
Vlad Yasevich 已提交
1740 1741 1742 1743 1744 1745 1746
	if (err) {
		__ip6_flush_pending_frames(sk, &queue, &cork, &v6_cork);
		return ERR_PTR(err);
	}

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