[linux.git] / net / ipv4 / ip_input.c

/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		The Internet Protocol (IP) module.
 *
 * Authors:	Ross Biro
 *		Fred N. van Kempen, <[email protected]>
 *		Donald Becker, <[email protected]>
 *		Alan Cox, <[email protected]>
 *		Richard Underwood
 *		Stefan Becker, <[email protected]>
 *		Jorge Cwik, <[email protected]>
 *		Arnt Gulbrandsen, <[email protected]>
 *
 *
 * Fixes:
 *		Alan Cox	:	Commented a couple of minor bits of surplus code
 *		Alan Cox	:	Undefining IP_FORWARD doesn't include the code
 *					(just stops a compiler warning).
 *		Alan Cox	:	Frames with >=MAX_ROUTE record routes, strict routes or loose routes
 *					are junked rather than corrupting things.
 *		Alan Cox	:	Frames to bad broadcast subnets are dumped
 *					We used to process them non broadcast and
 *					boy could that cause havoc.
 *		Alan Cox	:	ip_forward sets the free flag on the
 *					new frame it queues. Still crap because
 *					it copies the frame but at least it
 *					doesn't eat memory too.
 *		Alan Cox	:	Generic queue code and memory fixes.
 *		Fred Van Kempen :	IP fragment support (borrowed from NET2E)
 *		Gerhard Koerting:	Forward fragmented frames correctly.
 *		Gerhard Koerting: 	Fixes to my fix of the above 8-).
 *		Gerhard Koerting:	IP interface addressing fix.
 *		Linus Torvalds	:	More robustness checks
 *		Alan Cox	:	Even more checks: Still not as robust as it ought to be
 *		Alan Cox	:	Save IP header pointer for later
 *		Alan Cox	:	ip option setting
 *		Alan Cox	:	Use ip_tos/ip_ttl settings
 *		Alan Cox	:	Fragmentation bogosity removed
 *					(Thanks to [email protected])
 *		Dmitry Gorodchanin :	Send of a raw packet crash fix.
 *		Alan Cox	:	Silly ip bug when an overlength
 *					fragment turns up. Now frees the
 *					queue.
 *		Linus Torvalds/ :	Memory leakage on fragmentation
 *		Alan Cox	:	handling.
 *		Gerhard Koerting:	Forwarding uses IP priority hints
 *		Teemu Rantanen	:	Fragment problems.
 *		Alan Cox	:	General cleanup, comments and reformat
 *		Alan Cox	:	SNMP statistics
 *		Alan Cox	:	BSD address rule semantics. Also see
 *					UDP as there is a nasty checksum issue
 *					if you do things the wrong way.
 *		Alan Cox	:	Always defrag, moved IP_FORWARD to the config.in file
 *		Alan Cox	: 	IP options adjust sk->priority.
 *		Pedro Roque	:	Fix mtu/length error in ip_forward.
 *		Alan Cox	:	Avoid ip_chk_addr when possible.
 *	Richard Underwood	:	IP multicasting.
 *		Alan Cox	:	Cleaned up multicast handlers.
 *		Alan Cox	:	RAW sockets demultiplex in the BSD style.
 *		Gunther Mayer	:	Fix the SNMP reporting typo
 *		Alan Cox	:	Always in group 224.0.0.1
 *	Pauline Middelink	:	Fast ip_checksum update when forwarding
 *					Masquerading support.
 *		Alan Cox	:	Multicast loopback error for 224.0.0.1
 *		Alan Cox	:	IP_MULTICAST_LOOP option.
 *		Alan Cox	:	Use notifiers.
 *		Bjorn Ekwall	:	Removed ip_csum (from slhc.c too)
 *		Bjorn Ekwall	:	Moved ip_fast_csum to ip.h (inline!)
 *		Stefan Becker   :       Send out ICMP HOST REDIRECT
 *	Arnt Gulbrandsen	:	ip_build_xmit
 *		Alan Cox	:	Per socket routing cache
 *		Alan Cox	:	Fixed routing cache, added header cache.
 *		Alan Cox	:	Loopback didn't work right in original ip_build_xmit - fixed it.
 *		Alan Cox	:	Only send ICMP_REDIRECT if src/dest are the same net.
 *		Alan Cox	:	Incoming IP option handling.
 *		Alan Cox	:	Set saddr on raw output frames as per BSD.
 *		Alan Cox	:	Stopped broadcast source route explosions.
 *		Alan Cox	:	Can disable source routing
 *		Takeshi Sone    :	Masquerading didn't work.
 *	Dave Bonn,Alan Cox	:	Faster IP forwarding whenever possible.
 *		Alan Cox	:	Memory leaks, tramples, misc debugging.
 *		Alan Cox	:	Fixed multicast (by popular demand 8))
 *		Alan Cox	:	Fixed forwarding (by even more popular demand 8))
 *		Alan Cox	:	Fixed SNMP statistics [I think]
 *	Gerhard Koerting	:	IP fragmentation forwarding fix
 *		Alan Cox	:	Device lock against page fault.
 *		Alan Cox	:	IP_HDRINCL facility.
 *	Werner Almesberger	:	Zero fragment bug
 *		Alan Cox	:	RAW IP frame length bug
 *		Alan Cox	:	Outgoing firewall on build_xmit
 *		A.N.Kuznetsov	:	IP_OPTIONS support throughout the kernel
 *		Alan Cox	:	Multicast routing hooks
 *		Jos Vos		:	Do accounting *before* call_in_firewall
 *	Willy Konynenberg	:	Transparent proxying support
 *
 *
 *
 * To Fix:
 *		IP fragmentation wants rewriting cleanly. The RFC815 algorithm is much more efficient
 *		and could be made very efficient with the addition of some virtual memory hacks to permit
 *		the allocation of a buffer that can then be 'grown' by twiddling page tables.
 *		Output fragmentation wants updating along with the buffer management to use a single
 *		interleaved copy algorithm so that fragmenting has a one copy overhead. Actual packet
 *		output should probably do its own fragmentation at the UDP/RAW layer. TCP shouldn't cause
 *		fragmentation anyway.
 *
 *		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.
 */

#define pr_fmt(fmt) "IPv4: " fmt

#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/slab.h>

#include <linux/net.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/inet.h>
#include <linux/inetdevice.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>

#include <net/snmp.h>
#include <net/ip.h>
#include <net/protocol.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/arp.h>
#include <net/icmp.h>
#include <net/raw.h>
#include <net/checksum.h>
#include <net/inet_ecn.h>
#include <linux/netfilter_ipv4.h>
#include <net/xfrm.h>
#include <linux/mroute.h>
#include <linux/netlink.h>

/*
 *	Process Router Attention IP option (RFC 2113)
 */
bool ip_call_ra_chain(struct sk_buff *skb)
{
	struct ip_ra_chain *ra;
	u8 protocol = ip_hdr(skb)->protocol;
	struct sock *last = NULL;
	struct net_device *dev = skb->dev;

	for (ra = rcu_dereference(ip_ra_chain); ra; ra = rcu_dereference(ra->next)) {
		struct sock *sk = ra->sk;

		/* If socket is bound to an interface, only report
		 * the packet if it came  from that interface.
		 */
		if (sk && inet_sk(sk)->inet_num == protocol &&
		    (!sk->sk_bound_dev_if ||
		     sk->sk_bound_dev_if == dev->ifindex) &&
		    net_eq(sock_net(sk), dev_net(dev))) {
			if (ip_is_fragment(ip_hdr(skb))) {
				if (ip_defrag(skb, IP_DEFRAG_CALL_RA_CHAIN))
					return true;
			}
			if (last) {
				struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
				if (skb2)
					raw_rcv(last, skb2);
			}
			last = sk;
		}
	}

	if (last) {
		raw_rcv(last, skb);
		return true;
	}
	return false;
}

static int ip_local_deliver_finish(struct sk_buff *skb)
{
	struct net *net = dev_net(skb->dev);

	__skb_pull(skb, skb_network_header_len(skb));

	rcu_read_lock();
	{
		int protocol = ip_hdr(skb)->protocol;
		const struct net_protocol *ipprot;
		int raw;

	resubmit:
		raw = raw_local_deliver(skb, protocol);

		ipprot = rcu_dereference(inet_protos[protocol]);
		if (ipprot != NULL) {
			int ret;

			if (!ipprot->no_policy) {
				if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
					kfree_skb(skb);
					goto out;
				}
				nf_reset(skb);
			}
			ret = ipprot->handler(skb);
			if (ret < 0) {
				protocol = -ret;
				goto resubmit;
			}
			IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
		} else {
			if (!raw) {
				if (xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
					IP_INC_STATS_BH(net, IPSTATS_MIB_INUNKNOWNPROTOS);
					icmp_send(skb, ICMP_DEST_UNREACH,
						  ICMP_PROT_UNREACH, 0);
				}
				kfree_skb(skb);
			} else {
				IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
				consume_skb(skb);
			}
		}
	}
 out:
	rcu_read_unlock();

	return 0;
}

/*
 * 	Deliver IP Packets to the higher protocol layers.
 */
int ip_local_deliver(struct sk_buff *skb)
{
	/*
	 *	Reassemble IP fragments.
	 */

	if (ip_is_fragment(ip_hdr(skb))) {
		if (ip_defrag(skb, IP_DEFRAG_LOCAL_DELIVER))
			return 0;
	}

	return NF_HOOK(NFPROTO_IPV4, NF_INET_LOCAL_IN, skb, skb->dev, NULL,
		       ip_local_deliver_finish);
}

static inline bool ip_rcv_options(struct sk_buff *skb)
{
	struct ip_options *opt;
	const struct iphdr *iph;
	struct net_device *dev = skb->dev;

	/* It looks as overkill, because not all
	   IP options require packet mangling.
	   But it is the easiest for now, especially taking
	   into account that combination of IP options
	   and running sniffer is extremely rare condition.
					      --ANK (980813)
	*/
	if (skb_cow(skb, skb_headroom(skb))) {
		IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
		goto drop;
	}

	iph = ip_hdr(skb);
	opt = &(IPCB(skb)->opt);
	opt->optlen = iph->ihl*4 - sizeof(struct iphdr);

	if (ip_options_compile(dev_net(dev), opt, skb)) {
		IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
		goto drop;
	}

	if (unlikely(opt->srr)) {
		struct in_device *in_dev = __in_dev_get_rcu(dev);

		if (in_dev) {
			if (!IN_DEV_SOURCE_ROUTE(in_dev)) {
				if (IN_DEV_LOG_MARTIANS(in_dev))
					net_info_ratelimited("source route option %pI4 -> %pI4\n",
							     &iph->saddr,
							     &iph->daddr);
				goto drop;
			}
		}

		if (ip_options_rcv_srr(skb))
			goto drop;
	}

	return false;
drop:
	return true;
}

int sysctl_ip_early_demux __read_mostly = 1;
EXPORT_SYMBOL(sysctl_ip_early_demux);

static int ip_rcv_finish(struct sk_buff *skb)
{
	const struct iphdr *iph = ip_hdr(skb);
	struct rtable *rt;

	if (sysctl_ip_early_demux && !skb_dst(skb) && skb->sk == NULL) {
		const struct net_protocol *ipprot;
		int protocol = iph->protocol;

		ipprot = rcu_dereference(inet_protos[protocol]);
		if (ipprot && ipprot->early_demux) {
			ipprot->early_demux(skb);
			/* must reload iph, skb->head might have changed */
			iph = ip_hdr(skb);
		}
	}

	/*
	 *	Initialise the virtual path cache for the packet. It describes
	 *	how the packet travels inside Linux networking.
	 */
	if (!skb_dst(skb)) {
		int err = ip_route_input_noref(skb, iph->daddr, iph->saddr,
					       iph->tos, skb->dev);
		if (unlikely(err)) {
			if (err == -EXDEV)
				NET_INC_STATS_BH(dev_net(skb->dev),
						 LINUX_MIB_IPRPFILTER);
			goto drop;
		}
	}

#ifdef CONFIG_IP_ROUTE_CLASSID
	if (unlikely(skb_dst(skb)->tclassid)) {
		struct ip_rt_acct *st = this_cpu_ptr(ip_rt_acct);
		u32 idx = skb_dst(skb)->tclassid;
		st[idx&0xFF].o_packets++;
		st[idx&0xFF].o_bytes += skb->len;
		st[(idx>>16)&0xFF].i_packets++;
		st[(idx>>16)&0xFF].i_bytes += skb->len;
	}
#endif

	if (iph->ihl > 5 && ip_rcv_options(skb))
		goto drop;

	rt = skb_rtable(skb);
	if (rt->rt_type == RTN_MULTICAST) {
		IP_UPD_PO_STATS_BH(dev_net(rt->dst.dev), IPSTATS_MIB_INMCAST,
				skb->len);
	} else if (rt->rt_type == RTN_BROADCAST)
		IP_UPD_PO_STATS_BH(dev_net(rt->dst.dev), IPSTATS_MIB_INBCAST,
				skb->len);

	return dst_input(skb);

drop:
	kfree_skb(skb);
	return NET_RX_DROP;
}

/*
 * 	Main IP Receive routine.
 */
int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev)
{
	const struct iphdr *iph;
	u32 len;

	/* When the interface is in promisc. mode, drop all the crap
	 * that it receives, do not try to analyse it.
	 */
	if (skb->pkt_type == PACKET_OTHERHOST)
		goto drop;


	IP_UPD_PO_STATS_BH(dev_net(dev), IPSTATS_MIB_IN, skb->len);

	if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) {
		IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
		goto out;
	}

	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
		goto inhdr_error;

	iph = ip_hdr(skb);

	/*
	 *	RFC1122: 3.2.1.2 MUST silently discard any IP frame that fails the checksum.
	 *
	 *	Is the datagram acceptable?
	 *
	 *	1.	Length at least the size of an ip header
	 *	2.	Version of 4
	 *	3.	Checksums correctly. [Speed optimisation for later, skip loopback checksums]
	 *	4.	Doesn't have a bogus length
	 */

	if (iph->ihl < 5 || iph->version != 4)
		goto inhdr_error;

	BUILD_BUG_ON(IPSTATS_MIB_ECT1PKTS != IPSTATS_MIB_NOECTPKTS + INET_ECN_ECT_1);
	BUILD_BUG_ON(IPSTATS_MIB_ECT0PKTS != IPSTATS_MIB_NOECTPKTS + INET_ECN_ECT_0);
	BUILD_BUG_ON(IPSTATS_MIB_CEPKTS != IPSTATS_MIB_NOECTPKTS + INET_ECN_CE);
	IP_ADD_STATS_BH(dev_net(dev),
			IPSTATS_MIB_NOECTPKTS + (iph->tos & INET_ECN_MASK),
			max_t(unsigned short, 1, skb_shinfo(skb)->gso_segs));

	if (!pskb_may_pull(skb, iph->ihl*4))
		goto inhdr_error;

	iph = ip_hdr(skb);

	if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl)))
		goto csum_error;

	len = ntohs(iph->tot_len);
	if (skb->len < len) {
		IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INTRUNCATEDPKTS);
		goto drop;
	} else if (len < (iph->ihl*4))
		goto inhdr_error;

	/* Our transport medium may have padded the buffer out. Now we know it
	 * is IP we can trim to the true length of the frame.
	 * Note this now means skb->len holds ntohs(iph->tot_len).
	 */
	if (pskb_trim_rcsum(skb, len)) {
		IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
		goto drop;
	}

	skb->transport_header = skb->network_header + iph->ihl*4;

	/* Remove any debris in the socket control block */
	memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));

	/* Must drop socket now because of tproxy. */
	skb_orphan(skb);

	return NF_HOOK(NFPROTO_IPV4, NF_INET_PRE_ROUTING, skb, dev, NULL,
		       ip_rcv_finish);

csum_error:
	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_CSUMERRORS);
inhdr_error:
	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
drop:
	kfree_skb(skb);
out:
	return NET_RX_DROP;
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* INET An implementation of the TCP/IP protocol suite for the LINUX
	3	* operating system. INET is implemented using the BSD Socket
	4	* interface as the means of communication with the user level.
	5	*
	6	* The Internet Protocol (IP) module.
	7	*
02c30a84	8	* Authors: Ross Biro
1da177e4 LT	9	* Fred N. van Kempen, <[email protected]>
1da177e4 LT	10	* Donald Becker, <[email protected]>
113aa838	11	* Alan Cox, <[email protected]>
1da177e4 LT	12	* Richard Underwood
	13	* Stefan Becker, <[email protected]>
	14	* Jorge Cwik, <[email protected]>
	15	* Arnt Gulbrandsen, <[email protected]>
e905a9ed	16	*
1da177e4 LT	17	*
	18	* Fixes:
	19	* Alan Cox : Commented a couple of minor bits of surplus code
	20	* Alan Cox : Undefining IP_FORWARD doesn't include the code
	21	* (just stops a compiler warning).
	22	* Alan Cox : Frames with >=MAX_ROUTE record routes, strict routes or loose routes
	23	* are junked rather than corrupting things.
	24	* Alan Cox : Frames to bad broadcast subnets are dumped
	25	* We used to process them non broadcast and
	26	* boy could that cause havoc.
	27	* Alan Cox : ip_forward sets the free flag on the
	28	* new frame it queues. Still crap because
	29	* it copies the frame but at least it
	30	* doesn't eat memory too.
	31	* Alan Cox : Generic queue code and memory fixes.
	32	* Fred Van Kempen : IP fragment support (borrowed from NET2E)
	33	* Gerhard Koerting: Forward fragmented frames correctly.
	34	* Gerhard Koerting: Fixes to my fix of the above 8-).
	35	* Gerhard Koerting: IP interface addressing fix.
	36	* Linus Torvalds : More robustness checks
	37	* Alan Cox : Even more checks: Still not as robust as it ought to be
	38	* Alan Cox : Save IP header pointer for later
	39	* Alan Cox : ip option setting
	40	* Alan Cox : Use ip_tos/ip_ttl settings
	41	* Alan Cox : Fragmentation bogosity removed
	42	* (Thanks to [email protected])
	43	* Dmitry Gorodchanin : Send of a raw packet crash fix.
	44	* Alan Cox : Silly ip bug when an overlength
	45	* fragment turns up. Now frees the
	46	* queue.
	47	* Linus Torvalds/ : Memory leakage on fragmentation
	48	* Alan Cox : handling.
	49	* Gerhard Koerting: Forwarding uses IP priority hints
	50	* Teemu Rantanen : Fragment problems.
	51	* Alan Cox : General cleanup, comments and reformat
	52	* Alan Cox : SNMP statistics
	53	* Alan Cox : BSD address rule semantics. Also see
	54	* UDP as there is a nasty checksum issue
	55	* if you do things the wrong way.
	56	* Alan Cox : Always defrag, moved IP_FORWARD to the config.in file
	57	* Alan Cox : IP options adjust sk->priority.
	58	* Pedro Roque : Fix mtu/length error in ip_forward.
	59	* Alan Cox : Avoid ip_chk_addr when possible.
	60	* Richard Underwood : IP multicasting.
	61	* Alan Cox : Cleaned up multicast handlers.
	62	* Alan Cox : RAW sockets demultiplex in the BSD style.
	63	* Gunther Mayer : Fix the SNMP reporting typo
	64	* Alan Cox : Always in group 224.0.0.1
	65	* Pauline Middelink : Fast ip_checksum update when forwarding
	66	* Masquerading support.
	67	* Alan Cox : Multicast loopback error for 224.0.0.1
	68	* Alan Cox : IP_MULTICAST_LOOP option.
	69	* Alan Cox : Use notifiers.
	70	* Bjorn Ekwall : Removed ip_csum (from slhc.c too)
	71	* Bjorn Ekwall : Moved ip_fast_csum to ip.h (inline!)
	72	* Stefan Becker : Send out ICMP HOST REDIRECT
	73	* Arnt Gulbrandsen : ip_build_xmit
	74	* Alan Cox : Per socket routing cache
	75	* Alan Cox : Fixed routing cache, added header cache.
	76	* Alan Cox : Loopback didn't work right in original ip_build_xmit - fixed it.
	77	* Alan Cox : Only send ICMP_REDIRECT if src/dest are the same net.
	78	* Alan Cox : Incoming IP option handling.
	79	* Alan Cox : Set saddr on raw output frames as per BSD.
	80	* Alan Cox : Stopped broadcast source route explosions.
81	* Alan Cox : Can disable source routing
82	* Takeshi Sone : Masquerading didn't work.
83	* Dave Bonn,Alan Cox : Faster IP forwarding whenever possible.
84	* Alan Cox : Memory leaks, tramples, misc debugging.
85	* Alan Cox : Fixed multicast (by popular demand 8))
86	* Alan Cox : Fixed forwarding (by even more popular demand 8))
87	* Alan Cox : Fixed SNMP statistics [I think]
88	* Gerhard Koerting : IP fragmentation forwarding fix
89	* Alan Cox : Device lock against page fault.
90	* Alan Cox : IP_HDRINCL facility.
91	* Werner Almesberger : Zero fragment bug
92	* Alan Cox : RAW IP frame length bug
93	* Alan Cox : Outgoing firewall on build_xmit
94	* A.N.Kuznetsov : IP_OPTIONS support throughout the kernel
95	* Alan Cox : Multicast routing hooks
96	* Jos Vos : Do accounting before call_in_firewall
97	* Willy Konynenberg : Transparent proxying support
98	*
e905a9ed	99	*
1da177e4 LT	100	*
	101	* To Fix:
	102	* IP fragmentation wants rewriting cleanly. The RFC815 algorithm is much more efficient
	103	* and could be made very efficient with the addition of some virtual memory hacks to permit
	104	* the allocation of a buffer that can then be 'grown' by twiddling page tables.
e905a9ed	105	* Output fragmentation wants updating along with the buffer management to use a single
1da177e4 LT	106	* interleaved copy algorithm so that fragmenting has a one copy overhead. Actual packet
	107	* output should probably do its own fragmentation at the UDP/RAW layer. TCP shouldn't cause
	108	* fragmentation anyway.
	109	*
	110	* This program is free software; you can redistribute it and/or
	111	* modify it under the terms of the GNU General Public License
	112	* as published by the Free Software Foundation; either version
	113	* 2 of the License, or (at your option) any later version.
	114	*/
	115
afd46503 JP	116	#define pr_fmt(fmt) "IPv4: " fmt
afd46503 JP	117
1da177e4 LT	118	#include <linux/module.h>
	119	#include <linux/types.h>
	120	#include <linux/kernel.h>
	121	#include <linux/string.h>
	122	#include <linux/errno.h>
5a0e3ad6	123	#include <linux/slab.h>
1da177e4 LT	124
	125	#include <linux/net.h>
	126	#include <linux/socket.h>
	127	#include <linux/sockios.h>
	128	#include <linux/in.h>
	129	#include <linux/inet.h>
14c85021	130	#include <linux/inetdevice.h>
1da177e4 LT	131	#include <linux/netdevice.h>
	132	#include <linux/etherdevice.h>
	133
	134	#include <net/snmp.h>
	135	#include <net/ip.h>
	136	#include <net/protocol.h>
	137	#include <net/route.h>
	138	#include <linux/skbuff.h>
	139	#include <net/sock.h>
	140	#include <net/arp.h>
	141	#include <net/icmp.h>
	142	#include <net/raw.h>
	143	#include <net/checksum.h>
1f07d03e	144	#include <net/inet_ecn.h>
1da177e4 LT	145	#include <linux/netfilter_ipv4.h>
	146	#include <net/xfrm.h>
	147	#include <linux/mroute.h>
	148	#include <linux/netlink.h>
	149
1da177e4	150	/*
66018506	151	* Process Router Attention IP option (RFC 2113)
e905a9ed	152	*/
ba57b4db	153	bool ip_call_ra_chain(struct sk_buff *skb)
1da177e4 LT	154	{
1da177e4 LT	155	struct ip_ra_chain *ra;
eddc9ec5	156	u8 protocol = ip_hdr(skb)->protocol;
1da177e4	157	struct sock *last = NULL;
cb84663e	158	struct net_device *dev = skb->dev;
1da177e4	159
66018506	160	for (ra = rcu_dereference(ip_ra_chain); ra; ra = rcu_dereference(ra->next)) {
1da177e4 LT	161	struct sock *sk = ra->sk;
	162
	163	/* If socket is bound to an interface, only report
	164	* the packet if it came from that interface.
	165	*/
c720c7e8	166	if (sk && inet_sk(sk)->inet_num == protocol &&
1da177e4	167	(!sk->sk_bound_dev_if \|\|
cb84663e	168	sk->sk_bound_dev_if == dev->ifindex) &&
09ad9bc7	169	net_eq(sock_net(sk), dev_net(dev))) {
56f8a75c	170	if (ip_is_fragment(ip_hdr(skb))) {
66018506	171	if (ip_defrag(skb, IP_DEFRAG_CALL_RA_CHAIN))
ba57b4db	172	return true;
1da177e4 LT	173	}
	174	if (last) {
	175	struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
	176	if (skb2)
	177	raw_rcv(last, skb2);
	178	}
	179	last = sk;
	180	}
	181	}
	182
	183	if (last) {
	184	raw_rcv(last, skb);
ba57b4db	185	return true;
1da177e4	186	}
ba57b4db	187	return false;
1da177e4 LT	188	}
1da177e4 LT	189
861d0486	190	static int ip_local_deliver_finish(struct sk_buff *skb)
1da177e4	191	{
c346dca1	192	struct net *net = dev_net(skb->dev);
f145049a	193
21d1196a	194	__skb_pull(skb, skb_network_header_len(skb));
1da177e4 LT	195
	196	rcu_read_lock();
	197	{
eddc9ec5	198	int protocol = ip_hdr(skb)->protocol;
32613090	199	const struct net_protocol *ipprot;
f9242b6b	200	int raw;
1da177e4 LT	201
1da177e4 LT	202	resubmit:
7bc54c90	203	raw = raw_local_deliver(skb, protocol);
1da177e4	204
f9242b6b	205	ipprot = rcu_dereference(inet_protos[protocol]);
9c0188ac	206	if (ipprot != NULL) {
1da177e4 LT	207	int ret;
1da177e4 LT	208
b59c2701 PM	209	if (!ipprot->no_policy) {
	210	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
	211	kfree_skb(skb);
	212	goto out;
	213	}
	214	nf_reset(skb);
1da177e4 LT	215	}
	216	ret = ipprot->handler(skb);
	217	if (ret < 0) {
	218	protocol = -ret;
	219	goto resubmit;
	220	}
7c73a6fa	221	IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
1da177e4	222	} else {
7bc54c90	223	if (!raw) {
1da177e4	224	if (xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
7c73a6fa	225	IP_INC_STATS_BH(net, IPSTATS_MIB_INUNKNOWNPROTOS);
1da177e4 LT	226	icmp_send(skb, ICMP_DEST_UNREACH,
	227	ICMP_PROT_UNREACH, 0);
	228	}
d8c6f4b9 NH	229	kfree_skb(skb);
d8c6f4b9 NH	230	} else {
7c73a6fa	231	IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
d8c6f4b9 NH	232	consume_skb(skb);
d8c6f4b9 NH	233	}
1da177e4 LT	234	}
	235	}
	236	out:
	237	rcu_read_unlock();
	238
	239	return 0;
	240	}
	241
	242	/*
	243	* Deliver IP Packets to the higher protocol layers.
e905a9ed	244	*/
1da177e4 LT	245	int ip_local_deliver(struct sk_buff *skb)
	246	{
	247	/*
	248	* Reassemble IP fragments.
	249	*/
	250
56f8a75c	251	if (ip_is_fragment(ip_hdr(skb))) {
776c729e	252	if (ip_defrag(skb, IP_DEFRAG_LOCAL_DELIVER))
1da177e4 LT	253	return 0;
	254	}
	255
9bbc768a	256	return NF_HOOK(NFPROTO_IPV4, NF_INET_LOCAL_IN, skb, skb->dev, NULL,
1da177e4 LT	257	ip_local_deliver_finish);
	258	}
	259
6a91395f	260	static inline bool ip_rcv_options(struct sk_buff *skb)
d245407e TG	261	{
d245407e TG	262	struct ip_options *opt;
b71d1d42	263	const struct iphdr *iph;
d245407e TG	264	struct net_device *dev = skb->dev;
	265
	266	/* It looks as overkill, because not all
	267	IP options require packet mangling.
	268	But it is the easiest for now, especially taking
	269	into account that combination of IP options
	270	and running sniffer is extremely rare condition.
	271	--ANK (980813)
	272	*/
	273	if (skb_cow(skb, skb_headroom(skb))) {
7c73a6fa	274	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
d245407e TG	275	goto drop;
	276	}
	277
eddc9ec5	278	iph = ip_hdr(skb);
22aba383 DL	279	opt = &(IPCB(skb)->opt);
22aba383 DL	280	opt->optlen = iph->ihl*4 - sizeof(struct iphdr);
d245407e	281
c346dca1	282	if (ip_options_compile(dev_net(dev), opt, skb)) {
7c73a6fa	283	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
d245407e TG	284	goto drop;
	285	}
	286
d245407e	287	if (unlikely(opt->srr)) {
6e8b11b4 ED	288	struct in_device *in_dev = __in_dev_get_rcu(dev);
6e8b11b4 ED	289
d245407e TG	290	if (in_dev) {
d245407e TG	291	if (!IN_DEV_SOURCE_ROUTE(in_dev)) {
e87cc472 JP	292	if (IN_DEV_LOG_MARTIANS(in_dev))
	293	net_info_ratelimited("source route option %pI4 -> %pI4\n",
	294	&iph->saddr,
	295	&iph->daddr);
d245407e TG	296	goto drop;
d245407e TG	297	}
d245407e TG	298	}
	299
	300	if (ip_options_rcv_srr(skb))
	301	goto drop;
	302	}
	303
6a91395f	304	return false;
d245407e	305	drop:
6a91395f	306	return true;
d245407e TG	307	}
d245407e TG	308
6648bd7e	309	int sysctl_ip_early_demux __read_mostly = 1;
c7109986	310	EXPORT_SYMBOL(sysctl_ip_early_demux);
6648bd7e	311
861d0486	312	static int ip_rcv_finish(struct sk_buff *skb)
1da177e4	313	{
eddc9ec5	314	const struct iphdr *iph = ip_hdr(skb);
5506b54b	315	struct rtable *rt;
1da177e4	316
a452ce34	317	if (sysctl_ip_early_demux && !skb_dst(skb) && skb->sk == NULL) {
160eb5a6 DM	318	const struct net_protocol *ipprot;
	319	int protocol = iph->protocol;
	320
160eb5a6	321	ipprot = rcu_dereference(inet_protos[protocol]);
9cb429d6	322	if (ipprot && ipprot->early_demux) {
160eb5a6	323	ipprot->early_demux(skb);
9cb429d6 ED	324	/* must reload iph, skb->head might have changed */
	325	iph = ip_hdr(skb);
	326	}
160eb5a6 DM	327	}
160eb5a6 DM	328
1da177e4 LT	329	/*
	330	* Initialise the virtual path cache for the packet. It describes
	331	* how the packet travels inside Linux networking.
e905a9ed	332	*/
160eb5a6	333	if (!skb_dst(skb)) {
c6cffba4 DM	334	int err = ip_route_input_noref(skb, iph->daddr, iph->saddr,
c6cffba4 DM	335	iph->tos, skb->dev);
160eb5a6 DM	336	if (unlikely(err)) {
	337	if (err == -EXDEV)
	338	NET_INC_STATS_BH(dev_net(skb->dev),
	339	LINUX_MIB_IPRPFILTER);
	340	goto drop;
2c2910a4	341	}
1da177e4 LT	342	}
1da177e4 LT	343
c7066f70	344	#ifdef CONFIG_IP_ROUTE_CLASSID
adf30907	345	if (unlikely(skb_dst(skb)->tclassid)) {
7a9b2d59	346	struct ip_rt_acct *st = this_cpu_ptr(ip_rt_acct);
adf30907	347	u32 idx = skb_dst(skb)->tclassid;
1da177e4	348	st[idx&0xFF].o_packets++;
fd3f8c4c	349	st[idx&0xFF].o_bytes += skb->len;
1da177e4	350	st[(idx>>16)&0xFF].i_packets++;
fd3f8c4c	351	st[(idx>>16)&0xFF].i_bytes += skb->len;
1da177e4 LT	352	}
	353	#endif
	354
d245407e TG	355	if (iph->ihl > 5 && ip_rcv_options(skb))
d245407e TG	356	goto drop;
1da177e4	357
511c3f92	358	rt = skb_rtable(skb);
edf391ff	359	if (rt->rt_type == RTN_MULTICAST) {
d8d1f30b	360	IP_UPD_PO_STATS_BH(dev_net(rt->dst.dev), IPSTATS_MIB_INMCAST,
edf391ff NH	361	skb->len);
edf391ff NH	362	} else if (rt->rt_type == RTN_BROADCAST)
d8d1f30b	363	IP_UPD_PO_STATS_BH(dev_net(rt->dst.dev), IPSTATS_MIB_INBCAST,
edf391ff	364	skb->len);
5506b54b	365
1da177e4 LT	366	return dst_input(skb);
1da177e4 LT	367
1da177e4	368	drop:
e905a9ed YH	369	kfree_skb(skb);
e905a9ed YH	370	return NET_RX_DROP;
1da177e4 LT	371	}
	372
	373	/*
	374	* Main IP Receive routine.
e905a9ed	375	*/
f2ccd8fa	376	int ip_rcv(struct sk_buff skb, struct net_device dev, struct packet_type pt, struct net_device orig_dev)
1da177e4	377	{
b71d1d42	378	const struct iphdr *iph;
58615242	379	u32 len;
1da177e4 LT	380
	381	/* When the interface is in promisc. mode, drop all the crap
	382	* that it receives, do not try to analyse it.
	383	*/
	384	if (skb->pkt_type == PACKET_OTHERHOST)
	385	goto drop;
	386
edf391ff NH	387
edf391ff NH	388	IP_UPD_PO_STATS_BH(dev_net(dev), IPSTATS_MIB_IN, skb->len);
1da177e4 LT	389
1da177e4 LT	390	if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) {
7c73a6fa	391	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
1da177e4 LT	392	goto out;
	393	}
	394
	395	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
	396	goto inhdr_error;
	397
eddc9ec5	398	iph = ip_hdr(skb);
1da177e4 LT	399
1da177e4 LT	400	/*
c67fa027	401	* RFC1122: 3.2.1.2 MUST silently discard any IP frame that fails the checksum.
1da177e4 LT	402	*
	403	* Is the datagram acceptable?
	404	*
	405	* 1. Length at least the size of an ip header
	406	* 2. Version of 4
	407	* 3. Checksums correctly. [Speed optimisation for later, skip loopback checksums]
	408	* 4. Doesn't have a bogus length
	409	*/
	410
	411	if (iph->ihl < 5 \|\| iph->version != 4)
58615242	412	goto inhdr_error;
1da177e4	413
1f07d03e ED	414	BUILD_BUG_ON(IPSTATS_MIB_ECT1PKTS != IPSTATS_MIB_NOECTPKTS + INET_ECN_ECT_1);
	415	BUILD_BUG_ON(IPSTATS_MIB_ECT0PKTS != IPSTATS_MIB_NOECTPKTS + INET_ECN_ECT_0);
	416	BUILD_BUG_ON(IPSTATS_MIB_CEPKTS != IPSTATS_MIB_NOECTPKTS + INET_ECN_CE);
	417	IP_ADD_STATS_BH(dev_net(dev),
	418	IPSTATS_MIB_NOECTPKTS + (iph->tos & INET_ECN_MASK),
	419	max_t(unsigned short, 1, skb_shinfo(skb)->gso_segs));
	420
1da177e4 LT	421	if (!pskb_may_pull(skb, iph->ihl*4))
	422	goto inhdr_error;
	423
eddc9ec5	424	iph = ip_hdr(skb);
1da177e4	425
e9c60422	426	if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl)))
6a5dc9e5	427	goto csum_error;
1da177e4	428
58615242	429	len = ntohs(iph->tot_len);
704aed53	430	if (skb->len < len) {
7c73a6fa	431	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INTRUNCATEDPKTS);
704aed53 MC	432	goto drop;
704aed53 MC	433	} else if (len < (iph->ihl*4))
58615242	434	goto inhdr_error;
1da177e4	435
58615242 TG	436	/* Our transport medium may have padded the buffer out. Now we know it
	437	* is IP we can trim to the true length of the frame.
	438	* Note this now means skb->len holds ntohs(iph->tot_len).
	439	*/
	440	if (pskb_trim_rcsum(skb, len)) {
7c73a6fa	441	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
58615242	442	goto drop;
1da177e4 LT	443	}
1da177e4 LT	444
21d1196a ED	445	skb->transport_header = skb->network_header + iph->ihl*4;
21d1196a ED	446
53602f92	447	/* Remove any debris in the socket control block */
d569f1d7	448	memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
53602f92	449
71f9dacd HX	450	/* Must drop socket now because of tproxy. */
	451	skb_orphan(skb);
	452
9bbc768a	453	return NF_HOOK(NFPROTO_IPV4, NF_INET_PRE_ROUTING, skb, dev, NULL,
1da177e4 LT	454	ip_rcv_finish);
1da177e4 LT	455
6a5dc9e5 ED	456	csum_error:
6a5dc9e5 ED	457	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_CSUMERRORS);
1da177e4	458	inhdr_error:
7c73a6fa	459	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
1da177e4	460	drop:
e905a9ed	461	kfree_skb(skb);
1da177e4	462	out:
e905a9ed	463	return NET_RX_DROP;
1da177e4	464	}