[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.
 *
 * Version:	$Id: ip_input.c,v 1.55 2002/01/12 07:39:45 davem Exp $
 *
 * Authors:	Ross Biro
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *		Donald Becker, <becker@super.org>
 *		Alan Cox, <Alan.Cox@linux.org>
 *		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>

/*
 *	SNMP management statistics
 */

DEFINE_SNMP_STAT(struct ipstats_mib, ip_statistics) __read_mostly;

/*
 *	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) &&
		    sk->sk_net == dev->nd_net) {
			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)
{
	__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);
		if ((ipprot = rcu_dereference(inet_protos[hash])) != 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(IPSTATS_MIB_INDELIVERS);
		} else {
			if (!raw) {
				if (xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
					IP_INC_STATS_BH(IPSTATS_MIB_INUNKNOWNPROTOS);
					icmp_send(skb, ICMP_DEST_UNREACH,
						  ICMP_PROT_UNREACH, 0);
				}
			} else
				IP_INC_STATS_BH(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(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->nd_net, opt, skb)) {
		IP_INC_STATS_BH(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 "
					       "%u.%u.%u.%u -> %u.%u.%u.%u\n",
					       NIPQUAD(iph->saddr),
					       NIPQUAD(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 == 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(IPSTATS_MIB_INADDRERRORS);
			else if (err == -ENETUNREACH)
				IP_INC_STATS_BH(IPSTATS_MIB_INNOROUTES);
			goto drop;
		}
	}

#ifdef CONFIG_NET_CLS_ROUTE
	if (unlikely(skb->dst->tclassid)) {
		struct ip_rt_acct *st = per_cpu_ptr(ip_rt_acct, smp_processor_id());
		u32 idx = skb->dst->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;
	if (rt->rt_type == RTN_MULTICAST)
		IP_INC_STATS_BH(IPSTATS_MIB_INMCASTPKTS);
	else if (rt->rt_type == RTN_BROADCAST)
		IP_INC_STATS_BH(IPSTATS_MIB_INBCASTPKTS);

	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;

	if (dev->nd_net != &init_net)
		goto drop;

	/* 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_INC_STATS_BH(IPSTATS_MIB_INRECEIVES);

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

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

	iph = ip_hdr(skb);

	/*
	 *	RFC1122: 3.1.2.2 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(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(IPSTATS_MIB_INDISCARDS);
		goto drop;
	}

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

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

inhdr_error:
	IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS);
drop:
	kfree_skb(skb);
out:
	return NET_RX_DROP;
}

EXPORT_SYMBOL(ip_statistics);
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	*
	8	* Version: $Id: ip_input.c,v 1.55 2002/01/12 07:39:45 davem Exp $
	9	*
02c30a84	10	* Authors: Ross Biro
1da177e4 LT	11	* Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
	12	* Donald Becker, <becker@super.org>
	13	* Alan Cox, <Alan.Cox@linux.org>
	14	* Richard Underwood
	15	* Stefan Becker, <stefanb@yello.ping.de>
	16	* Jorge Cwik, <jorge@laser.satlink.net>
	17	* Arnt Gulbrandsen, <agulbra@nvg.unit.no>
e905a9ed	18	*
1da177e4 LT	19	*
	20	* Fixes:
	21	* Alan Cox : Commented a couple of minor bits of surplus code
	22	* Alan Cox : Undefining IP_FORWARD doesn't include the code
	23	* (just stops a compiler warning).
	24	* Alan Cox : Frames with >=MAX_ROUTE record routes, strict routes or loose routes
	25	* are junked rather than corrupting things.
	26	* Alan Cox : Frames to bad broadcast subnets are dumped
	27	* We used to process them non broadcast and
	28	* boy could that cause havoc.
	29	* Alan Cox : ip_forward sets the free flag on the
	30	* new frame it queues. Still crap because
	31	* it copies the frame but at least it
	32	* doesn't eat memory too.
	33	* Alan Cox : Generic queue code and memory fixes.
	34	* Fred Van Kempen : IP fragment support (borrowed from NET2E)
	35	* Gerhard Koerting: Forward fragmented frames correctly.
	36	* Gerhard Koerting: Fixes to my fix of the above 8-).
	37	* Gerhard Koerting: IP interface addressing fix.
	38	* Linus Torvalds : More robustness checks
	39	* Alan Cox : Even more checks: Still not as robust as it ought to be
	40	* Alan Cox : Save IP header pointer for later
	41	* Alan Cox : ip option setting
	42	* Alan Cox : Use ip_tos/ip_ttl settings
	43	* Alan Cox : Fragmentation bogosity removed
	44	* (Thanks to Mark.Bush@prg.ox.ac.uk)
	45	* Dmitry Gorodchanin : Send of a raw packet crash fix.
	46	* Alan Cox : Silly ip bug when an overlength
	47	* fragment turns up. Now frees the
	48	* queue.
	49	* Linus Torvalds/ : Memory leakage on fragmentation
	50	* Alan Cox : handling.
	51	* Gerhard Koerting: Forwarding uses IP priority hints
	52	* Teemu Rantanen : Fragment problems.
	53	* Alan Cox : General cleanup, comments and reformat
	54	* Alan Cox : SNMP statistics
	55	* Alan Cox : BSD address rule semantics. Also see
	56	* UDP as there is a nasty checksum issue
	57	* if you do things the wrong way.
	58	* Alan Cox : Always defrag, moved IP_FORWARD to the config.in file
	59	* Alan Cox : IP options adjust sk->priority.
	60	* Pedro Roque : Fix mtu/length error in ip_forward.
	61	* Alan Cox : Avoid ip_chk_addr when possible.
	62	* Richard Underwood : IP multicasting.
	63	* Alan Cox : Cleaned up multicast handlers.
	64	* Alan Cox : RAW sockets demultiplex in the BSD style.
	65	* Gunther Mayer : Fix the SNMP reporting typo
	66	* Alan Cox : Always in group 224.0.0.1
	67	* Pauline Middelink : Fast ip_checksum update when forwarding
	68	* Masquerading support.
	69	* Alan Cox : Multicast loopback error for 224.0.0.1
	70	* Alan Cox : IP_MULTICAST_LOOP option.
	71	* Alan Cox : Use notifiers.
	72	* Bjorn Ekwall : Removed ip_csum (from slhc.c too)
	73	* Bjorn Ekwall : Moved ip_fast_csum to ip.h (inline!)
	74	* Stefan Becker : Send out ICMP HOST REDIRECT
	75	* Arnt Gulbrandsen : ip_build_xmit
	76	* Alan Cox : Per socket routing cache
	77	* Alan Cox : Fixed routing cache, added header cache.
	78	* Alan Cox : Loopback didn't work right in original ip_build_xmit - fixed it.
	79	* Alan Cox : Only send ICMP_REDIRECT if src/dest are the same net.
	80	* Alan Cox : Incoming IP option handling.
	81	* Alan Cox : Set saddr on raw output frames as per BSD.
	82	* Alan Cox : Stopped broadcast source route explosions.
83	* Alan Cox : Can disable source routing
84	* Takeshi Sone : Masquerading didn't work.
85	* Dave Bonn,Alan Cox : Faster IP forwarding whenever possible.
86	* Alan Cox : Memory leaks, tramples, misc debugging.
87	* Alan Cox : Fixed multicast (by popular demand 8))
88	* Alan Cox : Fixed forwarding (by even more popular demand 8))
89	* Alan Cox : Fixed SNMP statistics [I think]
90	* Gerhard Koerting : IP fragmentation forwarding fix
91	* Alan Cox : Device lock against page fault.
92	* Alan Cox : IP_HDRINCL facility.
93	* Werner Almesberger : Zero fragment bug
94	* Alan Cox : RAW IP frame length bug
95	* Alan Cox : Outgoing firewall on build_xmit
96	* A.N.Kuznetsov : IP_OPTIONS support throughout the kernel
97	* Alan Cox : Multicast routing hooks
98	* Jos Vos : Do accounting before call_in_firewall
99	* Willy Konynenberg : Transparent proxying support
100	*
e905a9ed	101	*
1da177e4 LT	102	*
	103	* To Fix:
	104	* IP fragmentation wants rewriting cleanly. The RFC815 algorithm is much more efficient
	105	* and could be made very efficient with the addition of some virtual memory hacks to permit
	106	* the allocation of a buffer that can then be 'grown' by twiddling page tables.
e905a9ed	107	* Output fragmentation wants updating along with the buffer management to use a single
1da177e4 LT	108	* interleaved copy algorithm so that fragmenting has a one copy overhead. Actual packet
	109	* output should probably do its own fragmentation at the UDP/RAW layer. TCP shouldn't cause
	110	* fragmentation anyway.
	111	*
	112	* This program is free software; you can redistribute it and/or
	113	* modify it under the terms of the GNU General Public License
	114	* as published by the Free Software Foundation; either version
	115	* 2 of the License, or (at your option) any later version.
	116	*/
	117
	118	#include <asm/system.h>
	119	#include <linux/module.h>
	120	#include <linux/types.h>
	121	#include <linux/kernel.h>
	122	#include <linux/string.h>
	123	#include <linux/errno.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>
	144	#include <linux/netfilter_ipv4.h>
	145	#include <net/xfrm.h>
	146	#include <linux/mroute.h>
	147	#include <linux/netlink.h>
	148
	149	/*
	150	* SNMP management statistics
	151	*/
	152
ba89966c	153	DEFINE_SNMP_STAT(struct ipstats_mib, ip_statistics) __read_mostly;
1da177e4 LT	154
	155	/*
	156	* Process Router Attention IP option
e905a9ed	157	*/
1da177e4 LT	158	int ip_call_ra_chain(struct sk_buff *skb)
	159	{
	160	struct ip_ra_chain *ra;
eddc9ec5	161	u8 protocol = ip_hdr(skb)->protocol;
1da177e4	162	struct sock *last = NULL;
cb84663e	163	struct net_device *dev = skb->dev;
1da177e4 LT	164
	165	read_lock(&ip_ra_lock);
	166	for (ra = ip_ra_chain; ra; ra = ra->next) {
	167	struct sock *sk = ra->sk;
	168
	169	/* If socket is bound to an interface, only report
	170	* the packet if it came from that interface.
	171	*/
	172	if (sk && inet_sk(sk)->num == protocol &&
	173	(!sk->sk_bound_dev_if \|\|
cb84663e DL	174	sk->sk_bound_dev_if == dev->ifindex) &&
cb84663e DL	175	sk->sk_net == dev->nd_net) {
eddc9ec5	176	if (ip_hdr(skb)->frag_off & htons(IP_MF \| IP_OFFSET)) {
776c729e	177	if (ip_defrag(skb, IP_DEFRAG_CALL_RA_CHAIN)) {
1da177e4 LT	178	read_unlock(&ip_ra_lock);
	179	return 1;
	180	}
	181	}
	182	if (last) {
	183	struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
	184	if (skb2)
	185	raw_rcv(last, skb2);
	186	}
	187	last = sk;
	188	}
	189	}
	190
	191	if (last) {
	192	raw_rcv(last, skb);
	193	read_unlock(&ip_ra_lock);
	194	return 1;
	195	}
	196	read_unlock(&ip_ra_lock);
	197	return 0;
	198	}
	199
861d0486	200	static int ip_local_deliver_finish(struct sk_buff *skb)
1da177e4	201	{
c9bdd4b5	202	__skb_pull(skb, ip_hdrlen(skb));
1da177e4	203
e905a9ed	204	/* Point into the IP datagram, just past the header. */
badff6d0	205	skb_reset_transport_header(skb);
1da177e4 LT	206
	207	rcu_read_lock();
	208	{
eddc9ec5	209	int protocol = ip_hdr(skb)->protocol;
7bc54c90	210	int hash, raw;
1da177e4 LT	211	struct net_protocol *ipprot;
	212
	213	resubmit:
7bc54c90	214	raw = raw_local_deliver(skb, protocol);
1da177e4	215
7bc54c90	216	hash = protocol & (MAX_INET_PROTOS - 1);
1da177e4 LT	217	if ((ipprot = rcu_dereference(inet_protos[hash])) != NULL) {
	218	int ret;
	219
b59c2701 PM	220	if (!ipprot->no_policy) {
	221	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
	222	kfree_skb(skb);
	223	goto out;
	224	}
	225	nf_reset(skb);
1da177e4 LT	226	}
	227	ret = ipprot->handler(skb);
	228	if (ret < 0) {
	229	protocol = -ret;
	230	goto resubmit;
	231	}
	232	IP_INC_STATS_BH(IPSTATS_MIB_INDELIVERS);
	233	} else {
7bc54c90	234	if (!raw) {
1da177e4 LT	235	if (xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
	236	IP_INC_STATS_BH(IPSTATS_MIB_INUNKNOWNPROTOS);
	237	icmp_send(skb, ICMP_DEST_UNREACH,
	238	ICMP_PROT_UNREACH, 0);
	239	}
	240	} else
	241	IP_INC_STATS_BH(IPSTATS_MIB_INDELIVERS);
	242	kfree_skb(skb);
	243	}
	244	}
	245	out:
	246	rcu_read_unlock();
	247
	248	return 0;
	249	}
	250
	251	/*
	252	* Deliver IP Packets to the higher protocol layers.
e905a9ed	253	*/
1da177e4 LT	254	int ip_local_deliver(struct sk_buff *skb)
	255	{
	256	/*
	257	* Reassemble IP fragments.
	258	*/
	259
eddc9ec5	260	if (ip_hdr(skb)->frag_off & htons(IP_MF \| IP_OFFSET)) {
776c729e	261	if (ip_defrag(skb, IP_DEFRAG_LOCAL_DELIVER))
1da177e4 LT	262	return 0;
	263	}
	264
6e23ae2a	265	return NF_HOOK(PF_INET, NF_INET_LOCAL_IN, skb, skb->dev, NULL,
1da177e4 LT	266	ip_local_deliver_finish);
	267	}
	268
d245407e TG	269	static inline int ip_rcv_options(struct sk_buff *skb)
	270	{
	271	struct ip_options *opt;
	272	struct iphdr *iph;
	273	struct net_device *dev = skb->dev;
	274
	275	/* It looks as overkill, because not all
	276	IP options require packet mangling.
	277	But it is the easiest for now, especially taking
	278	into account that combination of IP options
	279	and running sniffer is extremely rare condition.
	280	--ANK (980813)
	281	*/
	282	if (skb_cow(skb, skb_headroom(skb))) {
	283	IP_INC_STATS_BH(IPSTATS_MIB_INDISCARDS);
	284	goto drop;
	285	}
	286
eddc9ec5	287	iph = ip_hdr(skb);
22aba383 DL	288	opt = &(IPCB(skb)->opt);
22aba383 DL	289	opt->optlen = iph->ihl*4 - sizeof(struct iphdr);
d245407e	290
cb84663e	291	if (ip_options_compile(dev->nd_net, opt, skb)) {
d245407e TG	292	IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS);
	293	goto drop;
	294	}
	295
d245407e TG	296	if (unlikely(opt->srr)) {
	297	struct in_device *in_dev = in_dev_get(dev);
	298	if (in_dev) {
	299	if (!IN_DEV_SOURCE_ROUTE(in_dev)) {
	300	if (IN_DEV_LOG_MARTIANS(in_dev) &&
	301	net_ratelimit())
	302	printk(KERN_INFO "source route option "
	303	"%u.%u.%u.%u -> %u.%u.%u.%u\n",
	304	NIPQUAD(iph->saddr),
	305	NIPQUAD(iph->daddr));
	306	in_dev_put(in_dev);
	307	goto drop;
	308	}
	309
	310	in_dev_put(in_dev);
	311	}
	312
	313	if (ip_options_rcv_srr(skb))
	314	goto drop;
	315	}
	316
	317	return 0;
	318	drop:
	319	return -1;
	320	}
	321
861d0486	322	static int ip_rcv_finish(struct sk_buff *skb)
1da177e4	323	{
eddc9ec5	324	const struct iphdr *iph = ip_hdr(skb);
5506b54b	325	struct rtable *rt;
1da177e4 LT	326
	327	/*
	328	* Initialise the virtual path cache for the packet. It describes
	329	* how the packet travels inside Linux networking.
e905a9ed	330	*/
0182bd2b	331	if (skb->dst == NULL) {
3e192bea TG	332	int err = ip_route_input(skb, iph->daddr, iph->saddr, iph->tos,
	333	skb->dev);
	334	if (unlikely(err)) {
2c2910a4 DE	335	if (err == -EHOSTUNREACH)
2c2910a4 DE	336	IP_INC_STATS_BH(IPSTATS_MIB_INADDRERRORS);
e91a47eb MC	337	else if (err == -ENETUNREACH)
e91a47eb MC	338	IP_INC_STATS_BH(IPSTATS_MIB_INNOROUTES);
e905a9ed	339	goto drop;
2c2910a4	340	}
1da177e4 LT	341	}
	342
	343	#ifdef CONFIG_NET_CLS_ROUTE
3e192bea	344	if (unlikely(skb->dst->tclassid)) {
8dbde28d	345	struct ip_rt_acct *st = per_cpu_ptr(ip_rt_acct, smp_processor_id());
1da177e4 LT	346	u32 idx = skb->dst->tclassid;
	347	st[idx&0xFF].o_packets++;
	348	st[idx&0xFF].o_bytes+=skb->len;
	349	st[(idx>>16)&0xFF].i_packets++;
	350	st[(idx>>16)&0xFF].i_bytes+=skb->len;
	351	}
	352	#endif
	353
d245407e TG	354	if (iph->ihl > 5 && ip_rcv_options(skb))
d245407e TG	355	goto drop;
1da177e4	356
ee6b9673	357	rt = skb->rtable;
5506b54b MC	358	if (rt->rt_type == RTN_MULTICAST)
	359	IP_INC_STATS_BH(IPSTATS_MIB_INMCASTPKTS);
	360	else if (rt->rt_type == RTN_BROADCAST)
	361	IP_INC_STATS_BH(IPSTATS_MIB_INBCASTPKTS);
	362
1da177e4 LT	363	return dst_input(skb);
1da177e4 LT	364
1da177e4	365	drop:
e905a9ed YH	366	kfree_skb(skb);
e905a9ed YH	367	return NET_RX_DROP;
1da177e4 LT	368	}
	369
	370	/*
	371	* Main IP Receive routine.
e905a9ed	372	*/
f2ccd8fa	373	int ip_rcv(struct sk_buff skb, struct net_device dev, struct packet_type pt, struct net_device orig_dev)
1da177e4 LT	374	{
1da177e4 LT	375	struct iphdr *iph;
58615242	376	u32 len;
1da177e4	377
e730c155 EB	378	if (dev->nd_net != &init_net)
	379	goto drop;
	380
1da177e4 LT	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
	387	IP_INC_STATS_BH(IPSTATS_MIB_INRECEIVES);
	388
	389	if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) {
	390	IP_INC_STATS_BH(IPSTATS_MIB_INDISCARDS);
	391	goto out;
	392	}
	393
	394	if (!pskb_may_pull(skb, sizeof(struct iphdr)))
	395	goto inhdr_error;
	396
eddc9ec5	397	iph = ip_hdr(skb);
1da177e4 LT	398
	399	/*
	400	* RFC1122: 3.1.2.2 MUST silently discard any IP frame that fails the checksum.
	401	*
	402	* Is the datagram acceptable?
	403	*
	404	* 1. Length at least the size of an ip header
	405	* 2. Version of 4
	406	* 3. Checksums correctly. [Speed optimisation for later, skip loopback checksums]
	407	* 4. Doesn't have a bogus length
	408	*/
	409
	410	if (iph->ihl < 5 \|\| iph->version != 4)
58615242	411	goto inhdr_error;
1da177e4 LT	412
	413	if (!pskb_may_pull(skb, iph->ihl*4))
	414	goto inhdr_error;
	415
eddc9ec5	416	iph = ip_hdr(skb);
1da177e4	417
e9c60422	418	if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl)))
58615242	419	goto inhdr_error;
1da177e4	420
58615242	421	len = ntohs(iph->tot_len);
704aed53 MC	422	if (skb->len < len) {
	423	IP_INC_STATS_BH(IPSTATS_MIB_INTRUNCATEDPKTS);
	424	goto drop;
	425	} else if (len < (iph->ihl*4))
58615242	426	goto inhdr_error;
1da177e4	427
58615242 TG	428	/* Our transport medium may have padded the buffer out. Now we know it
	429	* is IP we can trim to the true length of the frame.
	430	* Note this now means skb->len holds ntohs(iph->tot_len).
	431	*/
	432	if (pskb_trim_rcsum(skb, len)) {
	433	IP_INC_STATS_BH(IPSTATS_MIB_INDISCARDS);
	434	goto drop;
1da177e4 LT	435	}
1da177e4 LT	436
53602f92	437	/* Remove any debris in the socket control block */
d569f1d7	438	memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
53602f92	439
6e23ae2a	440	return NF_HOOK(PF_INET, NF_INET_PRE_ROUTING, skb, dev, NULL,
1da177e4 LT	441	ip_rcv_finish);
	442
	443	inhdr_error:
	444	IP_INC_STATS_BH(IPSTATS_MIB_INHDRERRORS);
	445	drop:
e905a9ed	446	kfree_skb(skb);
1da177e4	447	out:
e905a9ed	448	return NET_RX_DROP;
1da177e4 LT	449	}
1da177e4 LT	450
1da177e4	451	EXPORT_SYMBOL(ip_statistics);