[linux-2.6-block.git] / net / ipv4 / ip_forward.c

// SPDX-License-Identifier: GPL-2.0
/*
 * 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 IP forwarding functionality.
 *
 * Authors:	see ip.c
 *
 * Fixes:
 *		Many		:	Split from ip.c , see ip_input.c for
 *					history.
 *		Dave Gregorich	:	NULL ip_rt_put fix for multicast
 *					routing.
 *		Jos Vos		:	Add call_out_firewall before sending,
 *					use output device for accounting.
 *		Jos Vos		:	Call forward firewall after routing
 *					(always use output device).
 *		Mike McLagan	:	Routing by source
 */

#include <linux/types.h>
#include <linux/mm.h>
#include <linux/skbuff.h>
#include <linux/ip.h>
#include <linux/icmp.h>
#include <linux/netdevice.h>
#include <linux/slab.h>
#include <net/sock.h>
#include <net/ip.h>
#include <net/tcp.h>
#include <net/udp.h>
#include <net/icmp.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/netfilter_ipv4.h>
#include <net/checksum.h>
#include <linux/route.h>
#include <net/route.h>
#include <net/xfrm.h>

static bool ip_exceeds_mtu(const struct sk_buff *skb, unsigned int mtu)
{
	if (skb->len <= mtu)
		return false;

	if (unlikely((ip_hdr(skb)->frag_off & htons(IP_DF)) == 0))
		return false;

	/* original fragment exceeds mtu and DF is set */
	if (unlikely(IPCB(skb)->frag_max_size > mtu))
		return true;

	if (skb->ignore_df)
		return false;

	if (skb_is_gso(skb) && skb_gso_validate_network_len(skb, mtu))
		return false;

	return true;
}


static int ip_forward_finish(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	struct ip_options *opt	= &(IPCB(skb)->opt);

#ifdef CONFIG_NET_SWITCHDEV
	if (skb->offload_l3_fwd_mark) {
		consume_skb(skb);
		return 0;
	}
#endif

	if (unlikely(opt->optlen))
		ip_forward_options(skb);

	skb_clear_tstamp(skb);
	return dst_output(net, sk, skb);
}

int ip_forward(struct sk_buff *skb)
{
	u32 mtu;
	struct iphdr *iph;	/* Our header */
	struct rtable *rt;	/* Route we use */
	struct ip_options *opt	= &(IPCB(skb)->opt);
	struct net *net;
	SKB_DR(reason);

	/* that should never happen */
	if (skb->pkt_type != PACKET_HOST)
		goto drop;

	if (unlikely(skb->sk))
		goto drop;

	if (skb_warn_if_lro(skb))
		goto drop;

	if (!xfrm4_policy_check(NULL, XFRM_POLICY_FWD, skb)) {
		SKB_DR_SET(reason, XFRM_POLICY);
		goto drop;
	}

	if (IPCB(skb)->opt.router_alert && ip_call_ra_chain(skb))
		return NET_RX_SUCCESS;

	skb_forward_csum(skb);
	net = dev_net(skb->dev);

	/*
	 *	According to the RFC, we must first decrease the TTL field. If
	 *	that reaches zero, we must reply an ICMP control message telling
	 *	that the packet's lifetime expired.
	 */
	if (ip_hdr(skb)->ttl <= 1)
		goto too_many_hops;

	if (!xfrm4_route_forward(skb)) {
		SKB_DR_SET(reason, XFRM_POLICY);
		goto drop;
	}

	rt = skb_rtable(skb);

	if (opt->is_strictroute && rt->rt_uses_gateway)
		goto sr_failed;

	__IP_INC_STATS(net, IPSTATS_MIB_OUTFORWDATAGRAMS);

	IPCB(skb)->flags |= IPSKB_FORWARDED;
	mtu = ip_dst_mtu_maybe_forward(&rt->dst, true);
	if (ip_exceeds_mtu(skb, mtu)) {
		IP_INC_STATS(net, IPSTATS_MIB_FRAGFAILS);
		icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED,
			  htonl(mtu));
		SKB_DR_SET(reason, PKT_TOO_BIG);
		goto drop;
	}

	/* We are about to mangle packet. Copy it! */
	if (skb_cow(skb, LL_RESERVED_SPACE(rt->dst.dev)+rt->dst.header_len))
		goto drop;
	iph = ip_hdr(skb);

	/* Decrease ttl after skb cow done */
	ip_decrease_ttl(iph);

	/*
	 *	We now generate an ICMP HOST REDIRECT giving the route
	 *	we calculated.
	 */
	if (IPCB(skb)->flags & IPSKB_DOREDIRECT && !opt->srr &&
	    !skb_sec_path(skb))
		ip_rt_send_redirect(skb);

	if (READ_ONCE(net->ipv4.sysctl_ip_fwd_update_priority))
		skb->priority = rt_tos2priority(iph->tos);

	return NF_HOOK(NFPROTO_IPV4, NF_INET_FORWARD,
		       net, NULL, skb, skb->dev, rt->dst.dev,
		       ip_forward_finish);

sr_failed:
	/*
	 *	Strict routing permits no gatewaying
	 */
	 icmp_send(skb, ICMP_DEST_UNREACH, ICMP_SR_FAILED, 0);
	 goto drop;

too_many_hops:
	/* Tell the sender its packet died... */
	__IP_INC_STATS(net, IPSTATS_MIB_INHDRERRORS);
	icmp_send(skb, ICMP_TIME_EXCEEDED, ICMP_EXC_TTL, 0);
	SKB_DR_SET(reason, IP_INHDR);
drop:
	kfree_skb_reason(skb, reason);
	return NET_RX_DROP;
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* INET An implementation of the TCP/IP protocol suite for the LINUX
	4	* operating system. INET is implemented using the BSD Socket
	5	* interface as the means of communication with the user level.
	6	*
	7	* The IP forwarding functionality.
e905a9ed	8	*
1da177e4 LT	9	* Authors: see ip.c
	10	*
	11	* Fixes:
e905a9ed	12	* Many : Split from ip.c , see ip_input.c for
1da177e4	13	* history.
e905a9ed	14	* Dave Gregorich : NULL ip_rt_put fix for multicast
1da177e4 LT	15	* routing.
	16	* Jos Vos : Add call_out_firewall before sending,
	17	* use output device for accounting.
	18	* Jos Vos : Call forward firewall after routing
	19	* (always use output device).
	20	* Mike McLagan : Routing by source
	21	*/
	22
1da177e4 LT	23	#include <linux/types.h>
1da177e4 LT	24	#include <linux/mm.h>
1da177e4 LT	25	#include <linux/skbuff.h>
	26	#include <linux/ip.h>
	27	#include <linux/icmp.h>
	28	#include <linux/netdevice.h>
5a0e3ad6	29	#include <linux/slab.h>
1da177e4 LT	30	#include <net/sock.h>
	31	#include <net/ip.h>
	32	#include <net/tcp.h>
	33	#include <net/udp.h>
	34	#include <net/icmp.h>
	35	#include <linux/tcp.h>
	36	#include <linux/udp.h>
	37	#include <linux/netfilter_ipv4.h>
	38	#include <net/checksum.h>
	39	#include <linux/route.h>
	40	#include <net/route.h>
	41	#include <net/xfrm.h>
	42
fe6cc55f FW	43	static bool ip_exceeds_mtu(const struct sk_buff *skb, unsigned int mtu)
fe6cc55f FW	44	{
ca6c5d4a	45	if (skb->len <= mtu)
fe6cc55f FW	46	return false;
fe6cc55f FW	47
cf826244 FW	48	if (unlikely((ip_hdr(skb)->frag_off & htons(IP_DF)) == 0))
	49	return false;
	50
	51	/* original fragment exceeds mtu and DF is set */
	52	if (unlikely(IPCB(skb)->frag_max_size > mtu))
	53	return true;
	54
	55	if (skb->ignore_df)
	56	return false;
	57
779b7931	58	if (skb_is_gso(skb) && skb_gso_validate_network_len(skb, mtu))
fe6cc55f FW	59	return false;
	60
	61	return true;
	62	}
	63
fe6cc55f	64
0c4b51f0	65	static int ip_forward_finish(struct net net, struct sock sk, struct sk_buff *skb)
1da177e4	66	{
5e73ea1a	67	struct ip_options *opt = &(IPCB(skb)->opt);
1da177e4	68
f839a6c9 IS	69	#ifdef CONFIG_NET_SWITCHDEV
	70	if (skb->offload_l3_fwd_mark) {
	71	consume_skb(skb);
	72	return 0;
	73	}
	74	#endif
	75
1da177e4 LT	76	if (unlikely(opt->optlen))
	77	ip_forward_options(skb);
	78
de799101	79	skb_clear_tstamp(skb);
13206b6b	80	return dst_output(net, sk, skb);
1da177e4 LT	81	}
	82
	83	int ip_forward(struct sk_buff *skb)
	84	{
f87c10a8	85	u32 mtu;
1da177e4 LT	86	struct iphdr iph; / Our header */
1da177e4 LT	87	struct rtable rt; / Route we use */
5e73ea1a	88	struct ip_options *opt = &(IPCB(skb)->opt);
fcad0ac2	89	struct net *net;
2edc1a38	90	SKB_DR(reason);
1da177e4	91
d4f2fa6a DK	92	/* that should never happen */
	93	if (skb->pkt_type != PACKET_HOST)
	94	goto drop;
	95
2ab95749 SP	96	if (unlikely(skb->sk))
	97	goto drop;
	98
4497b076 BH	99	if (skb_warn_if_lro(skb))
	100	goto drop;
	101
2edc1a38 MD	102	if (!xfrm4_policy_check(NULL, XFRM_POLICY_FWD, skb)) {
2edc1a38 MD	103	SKB_DR_SET(reason, XFRM_POLICY);
1da177e4	104	goto drop;
2edc1a38	105	}
1da177e4 LT	106
	107	if (IPCB(skb)->opt.router_alert && ip_call_ra_chain(skb))
	108	return NET_RX_SUCCESS;
	109
35fc92a9	110	skb_forward_csum(skb);
fcad0ac2	111	net = dev_net(skb->dev);
e905a9ed	112
1da177e4 LT	113	/*
	114	* According to the RFC, we must first decrease the TTL field. If
	115	* that reaches zero, we must reply an ICMP control message telling
	116	* that the packet's lifetime expired.
	117	*/
eddc9ec5	118	if (ip_hdr(skb)->ttl <= 1)
e905a9ed	119	goto too_many_hops;
1da177e4	120
2edc1a38 MD	121	if (!xfrm4_route_forward(skb)) {
2edc1a38 MD	122	SKB_DR_SET(reason, XFRM_POLICY);
1da177e4	123	goto drop;
2edc1a38	124	}
1da177e4	125
511c3f92	126	rt = skb_rtable(skb);
1da177e4	127
77d5bc7e	128	if (opt->is_strictroute && rt->rt_uses_gateway)
1da177e4 LT	129	goto sr_failed;
1da177e4 LT	130
cf8b49fb HG	131	__IP_INC_STATS(net, IPSTATS_MIB_OUTFORWDATAGRAMS);
cf8b49fb HG	132
9ee6c5dc	133	IPCB(skb)->flags \|= IPSKB_FORWARDED;
f87c10a8	134	mtu = ip_dst_mtu_maybe_forward(&rt->dst, true);
cf826244	135	if (ip_exceeds_mtu(skb, mtu)) {
fcad0ac2	136	IP_INC_STATS(net, IPSTATS_MIB_FRAGFAILS);
9af3912e	137	icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED,
f87c10a8	138	htonl(mtu));
2edc1a38	139	SKB_DR_SET(reason, PKT_TOO_BIG);
9af3912e JH	140	goto drop;
	141	}
	142
1da177e4	143	/* We are about to mangle packet. Copy it! */
d8d1f30b	144	if (skb_cow(skb, LL_RESERVED_SPACE(rt->dst.dev)+rt->dst.header_len))
1da177e4	145	goto drop;
eddc9ec5	146	iph = ip_hdr(skb);
1da177e4 LT	147
	148	/* Decrease ttl after skb cow done */
	149	ip_decrease_ttl(iph);
	150
	151	/*
	152	* We now generate an ICMP HOST REDIRECT giving the route
	153	* we calculated.
	154	*/
df4d9254 HFS	155	if (IPCB(skb)->flags & IPSKB_DOREDIRECT && !opt->srr &&
df4d9254 HFS	156	!skb_sec_path(skb))
1da177e4 LT	157	ip_rt_send_redirect(skb);
1da177e4 LT	158
7bf9e18d	159	if (READ_ONCE(net->ipv4.sysctl_ip_fwd_update_priority))
432e05d3	160	skb->priority = rt_tos2priority(iph->tos);
1da177e4	161
29a26a56 EB	162	return NF_HOOK(NFPROTO_IPV4, NF_INET_FORWARD,
	163	net, NULL, skb, skb->dev, rt->dst.dev,
	164	ip_forward_finish);
1da177e4 LT	165
1da177e4 LT	166	sr_failed:
e905a9ed	167	/*
1da177e4 LT	168	* Strict routing permits no gatewaying
1da177e4 LT	169	*/
e905a9ed YH	170	icmp_send(skb, ICMP_DEST_UNREACH, ICMP_SR_FAILED, 0);
e905a9ed YH	171	goto drop;
1da177e4 LT	172
1da177e4 LT	173	too_many_hops:
e905a9ed	174	/* Tell the sender its packet died... */
b45386ef	175	__IP_INC_STATS(net, IPSTATS_MIB_INHDRERRORS);
e905a9ed	176	icmp_send(skb, ICMP_TIME_EXCEEDED, ICMP_EXC_TTL, 0);
2edc1a38	177	SKB_DR_SET(reason, IP_INHDR);
1da177e4	178	drop:
2edc1a38	179	kfree_skb_reason(skb, reason);
1da177e4 LT	180	return NET_RX_DROP;
1da177e4 LT	181	}