[net-next-2.6.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, <waltje@uWalt.NL.Mugnet.ORG>
 *		Donald Becker, <becker@super.org>
 *		Alan Cox, <alan@lxorguk.ukuu.org.uk>
 *		Richard Underwood
 *		Stefan Becker, <stefanb@yello.ping.de>
 *		Jorge Cwik, <jorge@laser.satlink.net>
 *		Arnt Gulbrandsen, <agulbra@nvg.unit.no>
 *
 *
 * 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 Mark.Bush@prg.ox.ac.uk)
 *		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.
 */

#include <asm/system.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.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 <linux/netfilter_ipv4.h>
#include <net/xfrm.h>
#include <linux/mroute.h>
#include <linux/netlink.h>

/*
 *	Process Router Attention IP option
 */
int 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;

	read_lock(&ip_ra_lock);
	for (ra = ip_ra_chain; ra; ra = 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)->num == protocol &&
		    (!sk->sk_bound_dev_if ||
		     sk->sk_bound_dev_if == dev->ifindex) &&
		    sock_net(sk) == dev_net(dev)) {
			if (ip_hdr(skb)->frag_off & htons(IP_MF | IP_OFFSET)) {
				if (ip_defrag(skb, IP_DEFRAG_CALL_RA_CHAIN)) {
					read_unlock(&ip_ra_lock);
					return 1;
				}
			}
			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);
		read_unlock(&ip_ra_lock);
		return 1;
	}
	read_unlock(&ip_ra_lock);
	return 0;
}

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

	__skb_pull(skb, ip_hdrlen(skb));

	/* Point into the IP datagram, just past the header. */
	skb_reset_transport_header(skb);

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

	resubmit:
		raw = raw_local_deliver(skb, protocol);

		hash = protocol & (MAX_INET_PROTOS - 1);
		ipprot = rcu_dereference(inet_protos[hash]);
		if (ipprot != NULL) {
			int ret;

			if (!net_eq(net, &init_net) && !ipprot->netns_ok) {
				if (net_ratelimit())
					printk("%s: proto %d isn't netns-ready\n",
						__func__, protocol);
				kfree_skb(skb);
				goto out;
			}

			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);
				}
			} else
				IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
			kfree_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_hdr(skb)->frag_off & htons(IP_MF | IP_OFFSET)) {
		if (ip_defrag(skb, IP_DEFRAG_LOCAL_DELIVER))
			return 0;
	}

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

static inline int ip_rcv_options(struct sk_buff *skb)
{
	struct ip_options *opt;
	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(dev);
		if (in_dev) {
			if (!IN_DEV_SOURCE_ROUTE(in_dev)) {
				if (IN_DEV_LOG_MARTIANS(in_dev) &&
				    net_ratelimit())
					printk(KERN_INFO "source route option %pI4 -> %pI4\n",
					       &iph->saddr, &iph->daddr);
				in_dev_put(in_dev);
				goto drop;
			}

			in_dev_put(in_dev);
		}

		if (ip_options_rcv_srr(skb))
			goto drop;
	}

	return 0;
drop:
	return -1;
}

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

	/*
	 *	Initialise the virtual path cache for the packet. It describes
	 *	how the packet travels inside Linux networking.
	 */
	if (skb_dst(skb) == NULL) {
		int err = ip_route_input(skb, iph->daddr, iph->saddr, iph->tos,
					 skb->dev);
		if (unlikely(err)) {
			if (err == -EHOSTUNREACH)
				IP_INC_STATS_BH(dev_net(skb->dev),
						IPSTATS_MIB_INADDRERRORS);
			else if (err == -ENETUNREACH)
				IP_INC_STATS_BH(dev_net(skb->dev),
						IPSTATS_MIB_INNOROUTES);
			goto drop;
		}
	}

#ifdef CONFIG_NET_CLS_ROUTE
	if (unlikely(skb_dst(skb)->tclassid)) {
		struct ip_rt_acct *st = per_cpu_ptr(ip_rt_acct, smp_processor_id());
		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->u.dst.dev), IPSTATS_MIB_INMCAST,
				skb->len);
	} else if (rt->rt_type == RTN_BROADCAST)
		IP_UPD_PO_STATS_BH(dev_net(rt->u.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)
{
	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;

	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 inhdr_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;
	}

	/* 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(PF_INET, NF_INET_PRE_ROUTING, skb, dev, NULL,
		       ip_rcv_finish);

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, <waltje@uWalt.NL.Mugnet.ORG>
1da177e4 LT	10	* Donald Becker, <becker@super.org>
113aa838	11	* Alan Cox, <alan@lxorguk.ukuu.org.uk>
1da177e4 LT	12	* Richard Underwood
	13	* Stefan Becker, <stefanb@yello.ping.de>
	14	* Jorge Cwik, <jorge@laser.satlink.net>
	15	* Arnt Gulbrandsen, <agulbra@nvg.unit.no>
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 Mark.Bush@prg.ox.ac.uk)
	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
	116	#include <asm/system.h>
	117	#include <linux/module.h>
	118	#include <linux/types.h>
	119	#include <linux/kernel.h>
	120	#include <linux/string.h>
	121	#include <linux/errno.h>
1da177e4 LT	122
	123	#include <linux/net.h>
	124	#include <linux/socket.h>
	125	#include <linux/sockios.h>
	126	#include <linux/in.h>
	127	#include <linux/inet.h>
14c85021	128	#include <linux/inetdevice.h>
1da177e4 LT	129	#include <linux/netdevice.h>
	130	#include <linux/etherdevice.h>
	131
	132	#include <net/snmp.h>
	133	#include <net/ip.h>
	134	#include <net/protocol.h>
	135	#include <net/route.h>
	136	#include <linux/skbuff.h>
	137	#include <net/sock.h>
	138	#include <net/arp.h>
	139	#include <net/icmp.h>
	140	#include <net/raw.h>
	141	#include <net/checksum.h>
	142	#include <linux/netfilter_ipv4.h>
	143	#include <net/xfrm.h>
	144	#include <linux/mroute.h>
	145	#include <linux/netlink.h>
	146
1da177e4 LT	147	/*
1da177e4 LT	148	* Process Router Attention IP option
e905a9ed	149	*/
1da177e4 LT	150	int ip_call_ra_chain(struct sk_buff *skb)
	151	{
	152	struct ip_ra_chain *ra;
eddc9ec5	153	u8 protocol = ip_hdr(skb)->protocol;
1da177e4	154	struct sock *last = NULL;
cb84663e	155	struct net_device *dev = skb->dev;
1da177e4 LT	156
	157	read_lock(&ip_ra_lock);
	158	for (ra = ip_ra_chain; ra; ra = ra->next) {
	159	struct sock *sk = ra->sk;
	160
	161	/* If socket is bound to an interface, only report
	162	* the packet if it came from that interface.
	163	*/
	164	if (sk && inet_sk(sk)->num == protocol &&
	165	(!sk->sk_bound_dev_if \|\|
cb84663e	166	sk->sk_bound_dev_if == dev->ifindex) &&
3b1e0a65	167	sock_net(sk) == dev_net(dev)) {
eddc9ec5	168	if (ip_hdr(skb)->frag_off & htons(IP_MF \| IP_OFFSET)) {
776c729e	169	if (ip_defrag(skb, IP_DEFRAG_CALL_RA_CHAIN)) {
1da177e4 LT	170	read_unlock(&ip_ra_lock);
	171	return 1;
	172	}
	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);
	185	read_unlock(&ip_ra_lock);
	186	return 1;
	187	}
	188	read_unlock(&ip_ra_lock);
	189	return 0;
	190	}
	191
861d0486	192	static int ip_local_deliver_finish(struct sk_buff *skb)
1da177e4	193	{
c346dca1	194	struct net *net = dev_net(skb->dev);
f145049a	195
c9bdd4b5	196	__skb_pull(skb, ip_hdrlen(skb));
1da177e4	197
e905a9ed	198	/* Point into the IP datagram, just past the header. */
badff6d0	199	skb_reset_transport_header(skb);
1da177e4 LT	200
	201	rcu_read_lock();
	202	{
eddc9ec5	203	int protocol = ip_hdr(skb)->protocol;
7bc54c90	204	int hash, raw;
1da177e4 LT	205	struct net_protocol *ipprot;
	206
	207	resubmit:
7bc54c90	208	raw = raw_local_deliver(skb, protocol);
1da177e4	209
7bc54c90	210	hash = protocol & (MAX_INET_PROTOS - 1);
f145049a	211	ipprot = rcu_dereference(inet_protos[hash]);
9c0188ac	212	if (ipprot != NULL) {
1da177e4 LT	213	int ret;
1da177e4 LT	214
9c0188ac AD	215	if (!net_eq(net, &init_net) && !ipprot->netns_ok) {
	216	if (net_ratelimit())
	217	printk("%s: proto %d isn't netns-ready\n",
	218	__func__, protocol);
	219	kfree_skb(skb);
	220	goto out;
	221	}
	222
b59c2701 PM	223	if (!ipprot->no_policy) {
	224	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
	225	kfree_skb(skb);
	226	goto out;
	227	}
	228	nf_reset(skb);
1da177e4 LT	229	}
	230	ret = ipprot->handler(skb);
	231	if (ret < 0) {
	232	protocol = -ret;
	233	goto resubmit;
	234	}
7c73a6fa	235	IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
1da177e4	236	} else {
7bc54c90	237	if (!raw) {
1da177e4	238	if (xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
7c73a6fa	239	IP_INC_STATS_BH(net, IPSTATS_MIB_INUNKNOWNPROTOS);
1da177e4 LT	240	icmp_send(skb, ICMP_DEST_UNREACH,
	241	ICMP_PROT_UNREACH, 0);
	242	}
	243	} else
7c73a6fa	244	IP_INC_STATS_BH(net, IPSTATS_MIB_INDELIVERS);
1da177e4 LT	245	kfree_skb(skb);
	246	}
	247	}
	248	out:
	249	rcu_read_unlock();
	250
	251	return 0;
	252	}
	253
	254	/*
	255	* Deliver IP Packets to the higher protocol layers.
e905a9ed	256	*/
1da177e4 LT	257	int ip_local_deliver(struct sk_buff *skb)
	258	{
	259	/*
	260	* Reassemble IP fragments.
	261	*/
	262
eddc9ec5	263	if (ip_hdr(skb)->frag_off & htons(IP_MF \| IP_OFFSET)) {
776c729e	264	if (ip_defrag(skb, IP_DEFRAG_LOCAL_DELIVER))
1da177e4 LT	265	return 0;
	266	}
	267
6e23ae2a	268	return NF_HOOK(PF_INET, NF_INET_LOCAL_IN, skb, skb->dev, NULL,
1da177e4 LT	269	ip_local_deliver_finish);
	270	}
	271
d245407e TG	272	static inline int ip_rcv_options(struct sk_buff *skb)
	273	{
	274	struct ip_options *opt;
	275	struct iphdr *iph;
	276	struct net_device *dev = skb->dev;
	277
	278	/* It looks as overkill, because not all
	279	IP options require packet mangling.
	280	But it is the easiest for now, especially taking
	281	into account that combination of IP options
	282	and running sniffer is extremely rare condition.
	283	--ANK (980813)
	284	*/
	285	if (skb_cow(skb, skb_headroom(skb))) {
7c73a6fa	286	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
d245407e TG	287	goto drop;
	288	}
	289
eddc9ec5	290	iph = ip_hdr(skb);
22aba383 DL	291	opt = &(IPCB(skb)->opt);
22aba383 DL	292	opt->optlen = iph->ihl*4 - sizeof(struct iphdr);
d245407e	293
c346dca1	294	if (ip_options_compile(dev_net(dev), opt, skb)) {
7c73a6fa	295	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
d245407e TG	296	goto drop;
	297	}
	298
d245407e TG	299	if (unlikely(opt->srr)) {
	300	struct in_device *in_dev = in_dev_get(dev);
	301	if (in_dev) {
	302	if (!IN_DEV_SOURCE_ROUTE(in_dev)) {
	303	if (IN_DEV_LOG_MARTIANS(in_dev) &&
	304	net_ratelimit())
673d57e7 HH	305	printk(KERN_INFO "source route option %pI4 -> %pI4\n",
673d57e7 HH	306	&iph->saddr, &iph->daddr);
d245407e TG	307	in_dev_put(in_dev);
	308	goto drop;
	309	}
	310
	311	in_dev_put(in_dev);
	312	}
	313
	314	if (ip_options_rcv_srr(skb))
	315	goto drop;
	316	}
	317
	318	return 0;
	319	drop:
	320	return -1;
	321	}
	322
861d0486	323	static int ip_rcv_finish(struct sk_buff *skb)
1da177e4	324	{
eddc9ec5	325	const struct iphdr *iph = ip_hdr(skb);
5506b54b	326	struct rtable *rt;
1da177e4 LT	327
	328	/*
	329	* Initialise the virtual path cache for the packet. It describes
	330	* how the packet travels inside Linux networking.
e905a9ed	331	*/
adf30907	332	if (skb_dst(skb) == NULL) {
3e192bea TG	333	int err = ip_route_input(skb, iph->daddr, iph->saddr, iph->tos,
	334	skb->dev);
	335	if (unlikely(err)) {
2c2910a4	336	if (err == -EHOSTUNREACH)
7c73a6fa PE	337	IP_INC_STATS_BH(dev_net(skb->dev),
7c73a6fa PE	338	IPSTATS_MIB_INADDRERRORS);
e91a47eb	339	else if (err == -ENETUNREACH)
7c73a6fa PE	340	IP_INC_STATS_BH(dev_net(skb->dev),
7c73a6fa PE	341	IPSTATS_MIB_INNOROUTES);
e905a9ed	342	goto drop;
2c2910a4	343	}
1da177e4 LT	344	}
	345
	346	#ifdef CONFIG_NET_CLS_ROUTE
adf30907	347	if (unlikely(skb_dst(skb)->tclassid)) {
8dbde28d	348	struct ip_rt_acct *st = per_cpu_ptr(ip_rt_acct, smp_processor_id());
adf30907	349	u32 idx = skb_dst(skb)->tclassid;
1da177e4	350	st[idx&0xFF].o_packets++;
fd3f8c4c	351	st[idx&0xFF].o_bytes += skb->len;
1da177e4	352	st[(idx>>16)&0xFF].i_packets++;
fd3f8c4c	353	st[(idx>>16)&0xFF].i_bytes += skb->len;
1da177e4 LT	354	}
	355	#endif
	356
d245407e TG	357	if (iph->ihl > 5 && ip_rcv_options(skb))
d245407e TG	358	goto drop;
1da177e4	359
511c3f92	360	rt = skb_rtable(skb);
edf391ff NH	361	if (rt->rt_type == RTN_MULTICAST) {
	362	IP_UPD_PO_STATS_BH(dev_net(rt->u.dst.dev), IPSTATS_MIB_INMCAST,
	363	skb->len);
	364	} else if (rt->rt_type == RTN_BROADCAST)
	365	IP_UPD_PO_STATS_BH(dev_net(rt->u.dst.dev), IPSTATS_MIB_INBCAST,
	366	skb->len);
5506b54b	367
1da177e4 LT	368	return dst_input(skb);
1da177e4 LT	369
1da177e4	370	drop:
e905a9ed YH	371	kfree_skb(skb);
e905a9ed YH	372	return NET_RX_DROP;
1da177e4 LT	373	}
	374
	375	/*
	376	* Main IP Receive routine.
e905a9ed	377	*/
f2ccd8fa	378	int ip_rcv(struct sk_buff skb, struct net_device dev, struct packet_type pt, struct net_device orig_dev)
1da177e4 LT	379	{
1da177e4 LT	380	struct iphdr *iph;
58615242	381	u32 len;
1da177e4 LT	382
	383	/* When the interface is in promisc. mode, drop all the crap
	384	* that it receives, do not try to analyse it.
	385	*/
	386	if (skb->pkt_type == PACKET_OTHERHOST)
	387	goto drop;
	388
edf391ff NH	389
edf391ff NH	390	IP_UPD_PO_STATS_BH(dev_net(dev), IPSTATS_MIB_IN, skb->len);
1da177e4 LT	391
1da177e4 LT	392	if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) {
7c73a6fa	393	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
1da177e4 LT	394	goto out;
	395	}
	396
	397	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
	398	goto inhdr_error;
	399
eddc9ec5	400	iph = ip_hdr(skb);
1da177e4 LT	401
1da177e4 LT	402	/*
c67fa027	403	* RFC1122: 3.2.1.2 MUST silently discard any IP frame that fails the checksum.
1da177e4 LT	404	*
	405	* Is the datagram acceptable?
	406	*
	407	* 1. Length at least the size of an ip header
	408	* 2. Version of 4
	409	* 3. Checksums correctly. [Speed optimisation for later, skip loopback checksums]
	410	* 4. Doesn't have a bogus length
	411	*/
	412
	413	if (iph->ihl < 5 \|\| iph->version != 4)
58615242	414	goto inhdr_error;
1da177e4 LT	415
	416	if (!pskb_may_pull(skb, iph->ihl*4))
	417	goto inhdr_error;
	418
eddc9ec5	419	iph = ip_hdr(skb);
1da177e4	420
e9c60422	421	if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl)))
58615242	422	goto inhdr_error;
1da177e4	423
58615242	424	len = ntohs(iph->tot_len);
704aed53	425	if (skb->len < len) {
7c73a6fa	426	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INTRUNCATEDPKTS);
704aed53 MC	427	goto drop;
704aed53 MC	428	} else if (len < (iph->ihl*4))
58615242	429	goto inhdr_error;
1da177e4	430
58615242 TG	431	/* Our transport medium may have padded the buffer out. Now we know it
	432	* is IP we can trim to the true length of the frame.
	433	* Note this now means skb->len holds ntohs(iph->tot_len).
	434	*/
	435	if (pskb_trim_rcsum(skb, len)) {
7c73a6fa	436	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
58615242	437	goto drop;
1da177e4 LT	438	}
1da177e4 LT	439
53602f92	440	/* Remove any debris in the socket control block */
d569f1d7	441	memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
53602f92	442
71f9dacd HX	443	/* Must drop socket now because of tproxy. */
	444	skb_orphan(skb);
	445
6e23ae2a	446	return NF_HOOK(PF_INET, NF_INET_PRE_ROUTING, skb, dev, NULL,
1da177e4 LT	447	ip_rcv_finish);
	448
	449	inhdr_error:
7c73a6fa	450	IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
1da177e4	451	drop:
e905a9ed	452	kfree_skb(skb);
1da177e4	453	out:
e905a9ed	454	return NET_RX_DROP;
1da177e4	455	}