[mirror_ubuntu-jammy-kernel.git] / net / netfilter / nf_flow_table_ip.c

// SPDX-License-Identifier: GPL-2.0-only
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/netfilter.h>
#include <linux/rhashtable.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/netdevice.h>
#include <net/ip.h>
#include <net/ipv6.h>
#include <net/ip6_route.h>
#include <net/neighbour.h>
#include <net/netfilter/nf_flow_table.h>
#include <net/netfilter/nf_conntrack_acct.h>
/* For layer 4 checksum field offset. */
#include <linux/tcp.h>
#include <linux/udp.h>

static int nf_flow_state_check(struct flow_offload *flow, int proto,
			       struct sk_buff *skb, unsigned int thoff)
{
	struct tcphdr *tcph;

	if (proto != IPPROTO_TCP)
		return 0;

	tcph = (void *)(skb_network_header(skb) + thoff);
	if (unlikely(tcph->fin || tcph->rst)) {
		flow_offload_teardown(flow);
		return -1;
	}

	return 0;
}

static void nf_flow_nat_ip_tcp(struct sk_buff *skb, unsigned int thoff,
			       __be32 addr, __be32 new_addr)
{
	struct tcphdr *tcph;

	tcph = (void *)(skb_network_header(skb) + thoff);
	inet_proto_csum_replace4(&tcph->check, skb, addr, new_addr, true);
}

static void nf_flow_nat_ip_udp(struct sk_buff *skb, unsigned int thoff,
			       __be32 addr, __be32 new_addr)
{
	struct udphdr *udph;

	udph = (void *)(skb_network_header(skb) + thoff);
	if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
		inet_proto_csum_replace4(&udph->check, skb, addr,
					 new_addr, true);
		if (!udph->check)
			udph->check = CSUM_MANGLED_0;
	}
}

static void nf_flow_nat_ip_l4proto(struct sk_buff *skb, struct iphdr *iph,
				   unsigned int thoff, __be32 addr,
				   __be32 new_addr)
{
	switch (iph->protocol) {
	case IPPROTO_TCP:
		nf_flow_nat_ip_tcp(skb, thoff, addr, new_addr);
		break;
	case IPPROTO_UDP:
		nf_flow_nat_ip_udp(skb, thoff, addr, new_addr);
		break;
	}
}

static void nf_flow_snat_ip(const struct flow_offload *flow,
			    struct sk_buff *skb, struct iphdr *iph,
			    unsigned int thoff, enum flow_offload_tuple_dir dir)
{
	__be32 addr, new_addr;

	switch (dir) {
	case FLOW_OFFLOAD_DIR_ORIGINAL:
		addr = iph->saddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_v4.s_addr;
		iph->saddr = new_addr;
		break;
	case FLOW_OFFLOAD_DIR_REPLY:
		addr = iph->daddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v4.s_addr;
		iph->daddr = new_addr;
		break;
	}
	csum_replace4(&iph->check, addr, new_addr);

	nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
}

static void nf_flow_dnat_ip(const struct flow_offload *flow,
			    struct sk_buff *skb, struct iphdr *iph,
			    unsigned int thoff, enum flow_offload_tuple_dir dir)
{
	__be32 addr, new_addr;

	switch (dir) {
	case FLOW_OFFLOAD_DIR_ORIGINAL:
		addr = iph->daddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_v4.s_addr;
		iph->daddr = new_addr;
		break;
	case FLOW_OFFLOAD_DIR_REPLY:
		addr = iph->saddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v4.s_addr;
		iph->saddr = new_addr;
		break;
	}
	csum_replace4(&iph->check, addr, new_addr);

	nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
}

static void nf_flow_nat_ip(const struct flow_offload *flow, struct sk_buff *skb,
			  unsigned int thoff, enum flow_offload_tuple_dir dir,
			  struct iphdr *iph)
{
	if (test_bit(NF_FLOW_SNAT, &flow->flags)) {
		nf_flow_snat_port(flow, skb, thoff, iph->protocol, dir);
		nf_flow_snat_ip(flow, skb, iph, thoff, dir);
	}
	if (test_bit(NF_FLOW_DNAT, &flow->flags)) {
		nf_flow_dnat_port(flow, skb, thoff, iph->protocol, dir);
		nf_flow_dnat_ip(flow, skb, iph, thoff, dir);
	}
}

static bool ip_has_options(unsigned int thoff)
{
	return thoff != sizeof(struct iphdr);
}

static int nf_flow_tuple_ip(struct sk_buff *skb, const struct net_device *dev,
			    struct flow_offload_tuple *tuple, u32 *hdrsize)
{
	struct flow_ports *ports;
	unsigned int thoff;
	struct iphdr *iph;

	if (!pskb_may_pull(skb, sizeof(*iph)))
		return -1;

	iph = ip_hdr(skb);
	thoff = iph->ihl * 4;

	if (ip_is_fragment(iph) ||
	    unlikely(ip_has_options(thoff)))
		return -1;

	switch (iph->protocol) {
	case IPPROTO_TCP:
		*hdrsize = sizeof(struct tcphdr);
		break;
	case IPPROTO_UDP:
		*hdrsize = sizeof(struct udphdr);
		break;
	default:
		return -1;
	}

	if (iph->ttl <= 1)
		return -1;

	thoff = iph->ihl * 4;
	if (!pskb_may_pull(skb, thoff + *hdrsize))
		return -1;

	iph = ip_hdr(skb);
	ports = (struct flow_ports *)(skb_network_header(skb) + thoff);

	tuple->src_v4.s_addr	= iph->saddr;
	tuple->dst_v4.s_addr	= iph->daddr;
	tuple->src_port		= ports->source;
	tuple->dst_port		= ports->dest;
	tuple->l3proto		= AF_INET;
	tuple->l4proto		= iph->protocol;
	tuple->iifidx		= dev->ifindex;

	return 0;
}

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

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

	return true;
}

static unsigned int nf_flow_xmit_xfrm(struct sk_buff *skb,
				      const struct nf_hook_state *state,
				      struct dst_entry *dst)
{
	skb_orphan(skb);
	skb_dst_set_noref(skb, dst);
	dst_output(state->net, state->sk, skb);
	return NF_STOLEN;
}

unsigned int
nf_flow_offload_ip_hook(void *priv, struct sk_buff *skb,
			const struct nf_hook_state *state)
{
	struct flow_offload_tuple_rhash *tuplehash;
	struct nf_flowtable *flow_table = priv;
	struct flow_offload_tuple tuple = {};
	enum flow_offload_tuple_dir dir;
	struct flow_offload *flow;
	struct net_device *outdev;
	struct rtable *rt;
	unsigned int thoff;
	struct iphdr *iph;
	__be32 nexthop;
	u32 hdrsize;

	if (skb->protocol != htons(ETH_P_IP))
		return NF_ACCEPT;

	if (nf_flow_tuple_ip(skb, state->in, &tuple, &hdrsize) < 0)
		return NF_ACCEPT;

	tuplehash = flow_offload_lookup(flow_table, &tuple);
	if (tuplehash == NULL)
		return NF_ACCEPT;

	dir = tuplehash->tuple.dir;
	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	rt = (struct rtable *)flow->tuplehash[dir].tuple.dst_cache;
	outdev = rt->dst.dev;

	if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
		return NF_ACCEPT;

	iph = ip_hdr(skb);
	thoff = iph->ihl * 4;
	if (nf_flow_state_check(flow, iph->protocol, skb, thoff))
		return NF_ACCEPT;

	if (!dst_check(&rt->dst, 0)) {
		flow_offload_teardown(flow);
		return NF_ACCEPT;
	}

	if (skb_try_make_writable(skb, thoff + hdrsize))
		return NF_DROP;

	flow_offload_refresh(flow_table, flow);

	iph = ip_hdr(skb);
	nf_flow_nat_ip(flow, skb, thoff, dir, iph);

	ip_decrease_ttl(iph);
	skb->tstamp = 0;

	if (flow_table->flags & NF_FLOWTABLE_COUNTER)
		nf_ct_acct_update(flow->ct, tuplehash->tuple.dir, skb->len);

	if (unlikely(dst_xfrm(&rt->dst))) {
		memset(skb->cb, 0, sizeof(struct inet_skb_parm));
		IPCB(skb)->iif = skb->dev->ifindex;
		IPCB(skb)->flags = IPSKB_FORWARDED;
		return nf_flow_xmit_xfrm(skb, state, &rt->dst);
	}

	skb->dev = outdev;
	nexthop = rt_nexthop(rt, flow->tuplehash[!dir].tuple.src_v4.s_addr);
	skb_dst_set_noref(skb, &rt->dst);
	neigh_xmit(NEIGH_ARP_TABLE, outdev, &nexthop, skb);

	return NF_STOLEN;
}
EXPORT_SYMBOL_GPL(nf_flow_offload_ip_hook);

static void nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
				 struct in6_addr *addr,
				 struct in6_addr *new_addr,
				 struct ipv6hdr *ip6h)
{
	struct tcphdr *tcph;

	tcph = (void *)(skb_network_header(skb) + thoff);
	inet_proto_csum_replace16(&tcph->check, skb, addr->s6_addr32,
				  new_addr->s6_addr32, true);
}

static void nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
				 struct in6_addr *addr,
				 struct in6_addr *new_addr)
{
	struct udphdr *udph;

	udph = (void *)(skb_network_header(skb) + thoff);
	if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
		inet_proto_csum_replace16(&udph->check, skb, addr->s6_addr32,
					  new_addr->s6_addr32, true);
		if (!udph->check)
			udph->check = CSUM_MANGLED_0;
	}
}

static void nf_flow_nat_ipv6_l4proto(struct sk_buff *skb, struct ipv6hdr *ip6h,
				     unsigned int thoff, struct in6_addr *addr,
				     struct in6_addr *new_addr)
{
	switch (ip6h->nexthdr) {
	case IPPROTO_TCP:
		nf_flow_nat_ipv6_tcp(skb, thoff, addr, new_addr, ip6h);
		break;
	case IPPROTO_UDP:
		nf_flow_nat_ipv6_udp(skb, thoff, addr, new_addr);
		break;
	}
}

static void nf_flow_snat_ipv6(const struct flow_offload *flow,
			      struct sk_buff *skb, struct ipv6hdr *ip6h,
			      unsigned int thoff,
			      enum flow_offload_tuple_dir dir)
{
	struct in6_addr addr, new_addr;

	switch (dir) {
	case FLOW_OFFLOAD_DIR_ORIGINAL:
		addr = ip6h->saddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_v6;
		ip6h->saddr = new_addr;
		break;
	case FLOW_OFFLOAD_DIR_REPLY:
		addr = ip6h->daddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v6;
		ip6h->daddr = new_addr;
		break;
	}

	nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
}

static void nf_flow_dnat_ipv6(const struct flow_offload *flow,
			      struct sk_buff *skb, struct ipv6hdr *ip6h,
			      unsigned int thoff,
			      enum flow_offload_tuple_dir dir)
{
	struct in6_addr addr, new_addr;

	switch (dir) {
	case FLOW_OFFLOAD_DIR_ORIGINAL:
		addr = ip6h->daddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_v6;
		ip6h->daddr = new_addr;
		break;
	case FLOW_OFFLOAD_DIR_REPLY:
		addr = ip6h->saddr;
		new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v6;
		ip6h->saddr = new_addr;
		break;
	}

	nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
}

static void nf_flow_nat_ipv6(const struct flow_offload *flow,
			     struct sk_buff *skb,
			     enum flow_offload_tuple_dir dir,
			     struct ipv6hdr *ip6h)
{
	unsigned int thoff = sizeof(*ip6h);

	if (test_bit(NF_FLOW_SNAT, &flow->flags)) {
		nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir);
		nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir);
	}
	if (test_bit(NF_FLOW_DNAT, &flow->flags)) {
		nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir);
		nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir);
	}
}

static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev,
			      struct flow_offload_tuple *tuple, u32 *hdrsize)
{
	struct flow_ports *ports;
	struct ipv6hdr *ip6h;
	unsigned int thoff;

	if (!pskb_may_pull(skb, sizeof(*ip6h)))
		return -1;

	ip6h = ipv6_hdr(skb);

	switch (ip6h->nexthdr) {
	case IPPROTO_TCP:
		*hdrsize = sizeof(struct tcphdr);
		break;
	case IPPROTO_UDP:
		*hdrsize = sizeof(struct udphdr);
		break;
	default:
		return -1;
	}

	if (ip6h->hop_limit <= 1)
		return -1;

	thoff = sizeof(*ip6h);
	if (!pskb_may_pull(skb, thoff + *hdrsize))
		return -1;

	ip6h = ipv6_hdr(skb);
	ports = (struct flow_ports *)(skb_network_header(skb) + thoff);

	tuple->src_v6		= ip6h->saddr;
	tuple->dst_v6		= ip6h->daddr;
	tuple->src_port		= ports->source;
	tuple->dst_port		= ports->dest;
	tuple->l3proto		= AF_INET6;
	tuple->l4proto		= ip6h->nexthdr;
	tuple->iifidx		= dev->ifindex;

	return 0;
}

unsigned int
nf_flow_offload_ipv6_hook(void *priv, struct sk_buff *skb,
			  const struct nf_hook_state *state)
{
	struct flow_offload_tuple_rhash *tuplehash;
	struct nf_flowtable *flow_table = priv;
	struct flow_offload_tuple tuple = {};
	enum flow_offload_tuple_dir dir;
	const struct in6_addr *nexthop;
	struct flow_offload *flow;
	struct net_device *outdev;
	struct ipv6hdr *ip6h;
	struct rt6_info *rt;
	u32 hdrsize;

	if (skb->protocol != htons(ETH_P_IPV6))
		return NF_ACCEPT;

	if (nf_flow_tuple_ipv6(skb, state->in, &tuple, &hdrsize) < 0)
		return NF_ACCEPT;

	tuplehash = flow_offload_lookup(flow_table, &tuple);
	if (tuplehash == NULL)
		return NF_ACCEPT;

	dir = tuplehash->tuple.dir;
	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	rt = (struct rt6_info *)flow->tuplehash[dir].tuple.dst_cache;
	outdev = rt->dst.dev;

	if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
		return NF_ACCEPT;

	if (nf_flow_state_check(flow, ipv6_hdr(skb)->nexthdr, skb,
				sizeof(*ip6h)))
		return NF_ACCEPT;

	if (!dst_check(&rt->dst, 0)) {
		flow_offload_teardown(flow);
		return NF_ACCEPT;
	}

	if (skb_try_make_writable(skb, sizeof(*ip6h) + hdrsize))
		return NF_DROP;

	flow_offload_refresh(flow_table, flow);

	ip6h = ipv6_hdr(skb);
	nf_flow_nat_ipv6(flow, skb, dir, ip6h);

	ip6h->hop_limit--;
	skb->tstamp = 0;

	if (flow_table->flags & NF_FLOWTABLE_COUNTER)
		nf_ct_acct_update(flow->ct, tuplehash->tuple.dir, skb->len);

	if (unlikely(dst_xfrm(&rt->dst))) {
		memset(skb->cb, 0, sizeof(struct inet6_skb_parm));
		IP6CB(skb)->iif = skb->dev->ifindex;
		IP6CB(skb)->flags = IP6SKB_FORWARDED;
		return nf_flow_xmit_xfrm(skb, state, &rt->dst);
	}

	skb->dev = outdev;
	nexthop = rt6_nexthop(rt, &flow->tuplehash[!dir].tuple.src_v6);
	skb_dst_set_noref(skb, &rt->dst);
	neigh_xmit(NEIGH_ND_TABLE, outdev, nexthop, skb);

	return NF_STOLEN;
}
EXPORT_SYMBOL_GPL(nf_flow_offload_ipv6_hook);
Commit	Line	Data
457c8996	1	// SPDX-License-Identifier: GPL-2.0-only
7d208687 FF	2	#include <linux/kernel.h>
	3	#include <linux/init.h>
	4	#include <linux/module.h>
	5	#include <linux/netfilter.h>
	6	#include <linux/rhashtable.h>
	7	#include <linux/ip.h>
a908fdec	8	#include <linux/ipv6.h>
7d208687 FF	9	#include <linux/netdevice.h>
7d208687 FF	10	#include <net/ip.h>
a908fdec FF	11	#include <net/ipv6.h>
a908fdec FF	12	#include <net/ip6_route.h>
7d208687 FF	13	#include <net/neighbour.h>
7d208687 FF	14	#include <net/netfilter/nf_flow_table.h>
53c2b289	15	#include <net/netfilter/nf_conntrack_acct.h>
7d208687 FF	16	/* For layer 4 checksum field offset. */
	17	#include <linux/tcp.h>
	18	#include <linux/udp.h>
	19
33894c36 FF	20	static int nf_flow_state_check(struct flow_offload *flow, int proto,
33894c36 FF	21	struct sk_buff *skb, unsigned int thoff)
b6f27d32 FF	22	{
	23	struct tcphdr *tcph;
	24
33894c36 FF	25	if (proto != IPPROTO_TCP)
	26	return 0;
	27
b6f27d32 FF	28	tcph = (void *)(skb_network_header(skb) + thoff);
	29	if (unlikely(tcph->fin \|\| tcph->rst)) {
	30	flow_offload_teardown(flow);
	31	return -1;
	32	}
	33
	34	return 0;
	35	}
	36
f4401262 PNA	37	static void nf_flow_nat_ip_tcp(struct sk_buff *skb, unsigned int thoff,
f4401262 PNA	38	__be32 addr, __be32 new_addr)
7d208687 FF	39	{
	40	struct tcphdr *tcph;
	41
7d208687 FF	42	tcph = (void *)(skb_network_header(skb) + thoff);
7d208687 FF	43	inet_proto_csum_replace4(&tcph->check, skb, addr, new_addr, true);
7d208687 FF	44	}
7d208687 FF	45
f4401262 PNA	46	static void nf_flow_nat_ip_udp(struct sk_buff *skb, unsigned int thoff,
f4401262 PNA	47	__be32 addr, __be32 new_addr)
7d208687 FF	48	{
	49	struct udphdr *udph;
	50
7d208687 FF	51	udph = (void *)(skb_network_header(skb) + thoff);
	52	if (udph->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	53	inet_proto_csum_replace4(&udph->check, skb, addr,
	54	new_addr, true);
	55	if (!udph->check)
	56	udph->check = CSUM_MANGLED_0;
	57	}
7d208687 FF	58	}
7d208687 FF	59
f4401262 PNA	60	static void nf_flow_nat_ip_l4proto(struct sk_buff skb, struct iphdr iph,
	61	unsigned int thoff, __be32 addr,
	62	__be32 new_addr)
7d208687 FF	63	{
	64	switch (iph->protocol) {
	65	case IPPROTO_TCP:
f4401262	66	nf_flow_nat_ip_tcp(skb, thoff, addr, new_addr);
7d208687 FF	67	break;
7d208687 FF	68	case IPPROTO_UDP:
f4401262	69	nf_flow_nat_ip_udp(skb, thoff, addr, new_addr);
7d208687 FF	70	break;
7d208687 FF	71	}
7d208687 FF	72	}
7d208687 FF	73
f4401262 PNA	74	static void nf_flow_snat_ip(const struct flow_offload *flow,
	75	struct sk_buff skb, struct iphdr iph,
	76	unsigned int thoff, enum flow_offload_tuple_dir dir)
7d208687 FF	77	{
	78	__be32 addr, new_addr;
	79
	80	switch (dir) {
	81	case FLOW_OFFLOAD_DIR_ORIGINAL:
	82	addr = iph->saddr;
	83	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_v4.s_addr;
	84	iph->saddr = new_addr;
	85	break;
	86	case FLOW_OFFLOAD_DIR_REPLY:
	87	addr = iph->daddr;
	88	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v4.s_addr;
	89	iph->daddr = new_addr;
	90	break;
7d208687 FF	91	}
	92	csum_replace4(&iph->check, addr, new_addr);
	93
f4401262	94	nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
7d208687 FF	95	}
7d208687 FF	96
f4401262 PNA	97	static void nf_flow_dnat_ip(const struct flow_offload *flow,
	98	struct sk_buff skb, struct iphdr iph,
	99	unsigned int thoff, enum flow_offload_tuple_dir dir)
7d208687 FF	100	{
	101	__be32 addr, new_addr;
	102
	103	switch (dir) {
	104	case FLOW_OFFLOAD_DIR_ORIGINAL:
	105	addr = iph->daddr;
	106	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_v4.s_addr;
	107	iph->daddr = new_addr;
	108	break;
	109	case FLOW_OFFLOAD_DIR_REPLY:
	110	addr = iph->saddr;
	111	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v4.s_addr;
	112	iph->saddr = new_addr;
	113	break;
7d208687 FF	114	}
	115	csum_replace4(&iph->check, addr, new_addr);
	116
f4401262	117	nf_flow_nat_ip_l4proto(skb, iph, thoff, addr, new_addr);
7d208687 FF	118	}
7d208687 FF	119
f4401262	120	static void nf_flow_nat_ip(const struct flow_offload flow, struct sk_buff skb,
2fc11745 PNA	121	unsigned int thoff, enum flow_offload_tuple_dir dir,
2fc11745 PNA	122	struct iphdr *iph)
7d208687	123	{
f4401262 PNA	124	if (test_bit(NF_FLOW_SNAT, &flow->flags)) {
	125	nf_flow_snat_port(flow, skb, thoff, iph->protocol, dir);
	126	nf_flow_snat_ip(flow, skb, iph, thoff, dir);
	127	}
	128	if (test_bit(NF_FLOW_DNAT, &flow->flags)) {
	129	nf_flow_dnat_port(flow, skb, thoff, iph->protocol, dir);
	130	nf_flow_dnat_ip(flow, skb, iph, thoff, dir);
	131	}
7d208687 FF	132	}
	133
	134	static bool ip_has_options(unsigned int thoff)
	135	{
	136	return thoff != sizeof(struct iphdr);
	137	}
	138
	139	static int nf_flow_tuple_ip(struct sk_buff skb, const struct net_device dev,
2fc11745	140	struct flow_offload_tuple tuple, u32 hdrsize)
7d208687 FF	141	{
7d208687 FF	142	struct flow_ports *ports;
2fc11745	143	unsigned int thoff;
7d208687 FF	144	struct iphdr *iph;
	145
	146	if (!pskb_may_pull(skb, sizeof(*iph)))
	147	return -1;
	148
	149	iph = ip_hdr(skb);
	150	thoff = iph->ihl * 4;
	151
	152	if (ip_is_fragment(iph) \|\|
	153	unlikely(ip_has_options(thoff)))
	154	return -1;
	155
793d5d61 PNA	156	switch (iph->protocol) {
793d5d61 PNA	157	case IPPROTO_TCP:
2fc11745	158	*hdrsize = sizeof(struct tcphdr);
793d5d61 PNA	159	break;
793d5d61 PNA	160	case IPPROTO_UDP:
2fc11745	161	*hdrsize = sizeof(struct udphdr);
793d5d61 PNA	162	break;
793d5d61 PNA	163	default:
7d208687	164	return -1;
793d5d61	165	}
7d208687	166
33cc3c0c TY	167	if (iph->ttl <= 1)
	168	return -1;
	169
7d208687	170	thoff = iph->ihl * 4;
2fc11745	171	if (!pskb_may_pull(skb, thoff + *hdrsize))
7d208687 FF	172	return -1;
7d208687 FF	173
41e9ec5a	174	iph = ip_hdr(skb);
7d208687 FF	175	ports = (struct flow_ports *)(skb_network_header(skb) + thoff);
	176
	177	tuple->src_v4.s_addr = iph->saddr;
	178	tuple->dst_v4.s_addr = iph->daddr;
	179	tuple->src_port = ports->source;
	180	tuple->dst_port = ports->dest;
	181	tuple->l3proto = AF_INET;
	182	tuple->l4proto = iph->protocol;
	183	tuple->iifidx = dev->ifindex;
	184
	185	return 0;
	186	}
	187
	188	/* Based on ip_exceeds_mtu(). */
	189	static bool nf_flow_exceeds_mtu(const struct sk_buff *skb, unsigned int mtu)
	190	{
	191	if (skb->len <= mtu)
	192	return false;
	193
7d208687 FF	194	if (skb_is_gso(skb) && skb_gso_validate_network_len(skb, mtu))
	195	return false;
	196
	197	return true;
	198	}
	199
589b474a FW	200	static unsigned int nf_flow_xmit_xfrm(struct sk_buff *skb,
	201	const struct nf_hook_state *state,
	202	struct dst_entry *dst)
	203	{
	204	skb_orphan(skb);
	205	skb_dst_set_noref(skb, dst);
589b474a FW	206	dst_output(state->net, state->sk, skb);
	207	return NF_STOLEN;
	208	}
	209
7d208687 FF	210	unsigned int
	211	nf_flow_offload_ip_hook(void priv, struct sk_buff skb,
	212	const struct nf_hook_state *state)
	213	{
	214	struct flow_offload_tuple_rhash *tuplehash;
	215	struct nf_flowtable *flow_table = priv;
	216	struct flow_offload_tuple tuple = {};
	217	enum flow_offload_tuple_dir dir;
	218	struct flow_offload *flow;
	219	struct net_device *outdev;
2a79fd39	220	struct rtable *rt;
b6f27d32	221	unsigned int thoff;
7d208687 FF	222	struct iphdr *iph;
7d208687 FF	223	__be32 nexthop;
2fc11745	224	u32 hdrsize;
7d208687 FF	225
	226	if (skb->protocol != htons(ETH_P_IP))
	227	return NF_ACCEPT;
	228
2fc11745	229	if (nf_flow_tuple_ip(skb, state->in, &tuple, &hdrsize) < 0)
7d208687 FF	230	return NF_ACCEPT;
	231
	232	tuplehash = flow_offload_lookup(flow_table, &tuple);
	233	if (tuplehash == NULL)
	234	return NF_ACCEPT;
	235
7d208687 FF	236	dir = tuplehash->tuple.dir;
7d208687 FF	237	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
2a79fd39	238	rt = (struct rtable *)flow->tuplehash[dir].tuple.dst_cache;
227e1e4d	239	outdev = rt->dst.dev;
7d208687	240
e75b3e1c	241	if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
7d208687 FF	242	return NF_ACCEPT;
7d208687 FF	243
2fc11745 PNA	244	iph = ip_hdr(skb);
2fc11745 PNA	245	thoff = iph->ihl * 4;
2fc11745	246	if (nf_flow_state_check(flow, iph->protocol, skb, thoff))
b6f27d32 FF	247	return NF_ACCEPT;
b6f27d32 FF	248
e5075c0b	249	if (!dst_check(&rt->dst, 0)) {
589b474a FW	250	flow_offload_teardown(flow);
	251	return NF_ACCEPT;
	252	}
	253
2babb46c PNA	254	if (skb_try_make_writable(skb, thoff + hdrsize))
	255	return NF_DROP;
	256
1b9cd769 PNA	257	flow_offload_refresh(flow_table, flow);
1b9cd769 PNA	258
2babb46c	259	iph = ip_hdr(skb);
f4401262	260	nf_flow_nat_ip(flow, skb, thoff, dir, iph);
7d208687	261
7d208687	262	ip_decrease_ttl(iph);
de20900f	263	skb->tstamp = 0;
7d208687	264
53c2b289 PNA	265	if (flow_table->flags & NF_FLOWTABLE_COUNTER)
	266	nf_ct_acct_update(flow->ct, tuplehash->tuple.dir, skb->len);
	267
589b474a FW	268	if (unlikely(dst_xfrm(&rt->dst))) {
	269	memset(skb->cb, 0, sizeof(struct inet_skb_parm));
	270	IPCB(skb)->iif = skb->dev->ifindex;
	271	IPCB(skb)->flags = IPSKB_FORWARDED;
	272	return nf_flow_xmit_xfrm(skb, state, &rt->dst);
	273	}
	274
7d208687 FF	275	skb->dev = outdev;
7d208687 FF	276	nexthop = rt_nexthop(rt, flow->tuplehash[!dir].tuple.src_v4.s_addr);
2a79fd39	277	skb_dst_set_noref(skb, &rt->dst);
7d208687 FF	278	neigh_xmit(NEIGH_ARP_TABLE, outdev, &nexthop, skb);
	279
	280	return NF_STOLEN;
	281	}
	282	EXPORT_SYMBOL_GPL(nf_flow_offload_ip_hook);
a908fdec	283
f4401262 PNA	284	static void nf_flow_nat_ipv6_tcp(struct sk_buff *skb, unsigned int thoff,
	285	struct in6_addr *addr,
	286	struct in6_addr *new_addr,
	287	struct ipv6hdr *ip6h)
a908fdec FF	288	{
	289	struct tcphdr *tcph;
	290
a908fdec FF	291	tcph = (void *)(skb_network_header(skb) + thoff);
	292	inet_proto_csum_replace16(&tcph->check, skb, addr->s6_addr32,
	293	new_addr->s6_addr32, true);
a908fdec FF	294	}
a908fdec FF	295
f4401262 PNA	296	static void nf_flow_nat_ipv6_udp(struct sk_buff *skb, unsigned int thoff,
	297	struct in6_addr *addr,
	298	struct in6_addr *new_addr)
a908fdec FF	299	{
	300	struct udphdr *udph;
	301
a908fdec FF	302	udph = (void *)(skb_network_header(skb) + thoff);
	303	if (udph->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	304	inet_proto_csum_replace16(&udph->check, skb, addr->s6_addr32,
	305	new_addr->s6_addr32, true);
	306	if (!udph->check)
	307	udph->check = CSUM_MANGLED_0;
	308	}
a908fdec FF	309	}
a908fdec FF	310
f4401262 PNA	311	static void nf_flow_nat_ipv6_l4proto(struct sk_buff skb, struct ipv6hdr ip6h,
	312	unsigned int thoff, struct in6_addr *addr,
	313	struct in6_addr *new_addr)
a908fdec FF	314	{
	315	switch (ip6h->nexthdr) {
	316	case IPPROTO_TCP:
f4401262	317	nf_flow_nat_ipv6_tcp(skb, thoff, addr, new_addr, ip6h);
a908fdec FF	318	break;
a908fdec FF	319	case IPPROTO_UDP:
f4401262	320	nf_flow_nat_ipv6_udp(skb, thoff, addr, new_addr);
a908fdec FF	321	break;
a908fdec FF	322	}
a908fdec FF	323	}
a908fdec FF	324
f4401262 PNA	325	static void nf_flow_snat_ipv6(const struct flow_offload *flow,
	326	struct sk_buff skb, struct ipv6hdr ip6h,
	327	unsigned int thoff,
	328	enum flow_offload_tuple_dir dir)
a908fdec FF	329	{
	330	struct in6_addr addr, new_addr;
	331
	332	switch (dir) {
	333	case FLOW_OFFLOAD_DIR_ORIGINAL:
	334	addr = ip6h->saddr;
	335	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_v6;
	336	ip6h->saddr = new_addr;
	337	break;
	338	case FLOW_OFFLOAD_DIR_REPLY:
	339	addr = ip6h->daddr;
	340	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_v6;
	341	ip6h->daddr = new_addr;
	342	break;
a908fdec FF	343	}
a908fdec FF	344
f4401262	345	nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
a908fdec FF	346	}
a908fdec FF	347
f4401262 PNA	348	static void nf_flow_dnat_ipv6(const struct flow_offload *flow,
	349	struct sk_buff skb, struct ipv6hdr ip6h,
	350	unsigned int thoff,
	351	enum flow_offload_tuple_dir dir)
a908fdec FF	352	{
	353	struct in6_addr addr, new_addr;
	354
	355	switch (dir) {
	356	case FLOW_OFFLOAD_DIR_ORIGINAL:
	357	addr = ip6h->daddr;
	358	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_v6;
	359	ip6h->daddr = new_addr;
	360	break;
	361	case FLOW_OFFLOAD_DIR_REPLY:
	362	addr = ip6h->saddr;
	363	new_addr = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_v6;
	364	ip6h->saddr = new_addr;
	365	break;
a908fdec FF	366	}
a908fdec FF	367
f4401262	368	nf_flow_nat_ipv6_l4proto(skb, ip6h, thoff, &addr, &new_addr);
a908fdec FF	369	}
a908fdec FF	370
f4401262 PNA	371	static void nf_flow_nat_ipv6(const struct flow_offload *flow,
	372	struct sk_buff *skb,
	373	enum flow_offload_tuple_dir dir,
	374	struct ipv6hdr *ip6h)
a908fdec	375	{
a908fdec FF	376	unsigned int thoff = sizeof(*ip6h);
a908fdec FF	377
f4401262 PNA	378	if (test_bit(NF_FLOW_SNAT, &flow->flags)) {
	379	nf_flow_snat_port(flow, skb, thoff, ip6h->nexthdr, dir);
	380	nf_flow_snat_ipv6(flow, skb, ip6h, thoff, dir);
	381	}
	382	if (test_bit(NF_FLOW_DNAT, &flow->flags)) {
	383	nf_flow_dnat_port(flow, skb, thoff, ip6h->nexthdr, dir);
	384	nf_flow_dnat_ipv6(flow, skb, ip6h, thoff, dir);
	385	}
a908fdec FF	386	}
	387
	388	static int nf_flow_tuple_ipv6(struct sk_buff skb, const struct net_device dev,
2fc11745	389	struct flow_offload_tuple tuple, u32 hdrsize)
a908fdec FF	390	{
	391	struct flow_ports *ports;
	392	struct ipv6hdr *ip6h;
2fc11745	393	unsigned int thoff;
a908fdec FF	394
	395	if (!pskb_may_pull(skb, sizeof(*ip6h)))
	396	return -1;
	397
	398	ip6h = ipv6_hdr(skb);
	399
793d5d61 PNA	400	switch (ip6h->nexthdr) {
793d5d61 PNA	401	case IPPROTO_TCP:
2fc11745	402	*hdrsize = sizeof(struct tcphdr);
793d5d61 PNA	403	break;
793d5d61 PNA	404	case IPPROTO_UDP:
2fc11745	405	*hdrsize = sizeof(struct udphdr);
793d5d61 PNA	406	break;
793d5d61 PNA	407	default:
a908fdec	408	return -1;
793d5d61	409	}
a908fdec	410
33cc3c0c TY	411	if (ip6h->hop_limit <= 1)
	412	return -1;
	413
a908fdec	414	thoff = sizeof(*ip6h);
2fc11745	415	if (!pskb_may_pull(skb, thoff + *hdrsize))
a908fdec FF	416	return -1;
a908fdec FF	417
41e9ec5a	418	ip6h = ipv6_hdr(skb);
a908fdec FF	419	ports = (struct flow_ports *)(skb_network_header(skb) + thoff);
	420
	421	tuple->src_v6 = ip6h->saddr;
	422	tuple->dst_v6 = ip6h->daddr;
	423	tuple->src_port = ports->source;
	424	tuple->dst_port = ports->dest;
	425	tuple->l3proto = AF_INET6;
	426	tuple->l4proto = ip6h->nexthdr;
	427	tuple->iifidx = dev->ifindex;
	428
	429	return 0;
	430	}
	431
	432	unsigned int
	433	nf_flow_offload_ipv6_hook(void priv, struct sk_buff skb,
	434	const struct nf_hook_state *state)
	435	{
	436	struct flow_offload_tuple_rhash *tuplehash;
	437	struct nf_flowtable *flow_table = priv;
	438	struct flow_offload_tuple tuple = {};
	439	enum flow_offload_tuple_dir dir;
9b1c1ef1	440	const struct in6_addr *nexthop;
a908fdec FF	441	struct flow_offload *flow;
a908fdec FF	442	struct net_device *outdev;
a908fdec FF	443	struct ipv6hdr *ip6h;
a908fdec FF	444	struct rt6_info *rt;
2fc11745	445	u32 hdrsize;
a908fdec FF	446
	447	if (skb->protocol != htons(ETH_P_IPV6))
	448	return NF_ACCEPT;
	449
2fc11745	450	if (nf_flow_tuple_ipv6(skb, state->in, &tuple, &hdrsize) < 0)
a908fdec FF	451	return NF_ACCEPT;
	452
	453	tuplehash = flow_offload_lookup(flow_table, &tuple);
	454	if (tuplehash == NULL)
	455	return NF_ACCEPT;
	456
a908fdec FF	457	dir = tuplehash->tuple.dir;
	458	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	459	rt = (struct rt6_info *)flow->tuplehash[dir].tuple.dst_cache;
227e1e4d	460	outdev = rt->dst.dev;
a908fdec FF	461
	462	if (unlikely(nf_flow_exceeds_mtu(skb, flow->tuplehash[dir].tuple.mtu)))
	463	return NF_ACCEPT;
	464
33894c36 FF	465	if (nf_flow_state_check(flow, ipv6_hdr(skb)->nexthdr, skb,
33894c36 FF	466	sizeof(*ip6h)))
b6f27d32 FF	467	return NF_ACCEPT;
b6f27d32 FF	468
e5075c0b	469	if (!dst_check(&rt->dst, 0)) {
589b474a FW	470	flow_offload_teardown(flow);
	471	return NF_ACCEPT;
	472	}
	473
2fc11745	474	if (skb_try_make_writable(skb, sizeof(*ip6h) + hdrsize))
a908fdec FF	475	return NF_DROP;
a908fdec FF	476
1b9cd769 PNA	477	flow_offload_refresh(flow_table, flow);
1b9cd769 PNA	478
2fc11745	479	ip6h = ipv6_hdr(skb);
f4401262	480	nf_flow_nat_ipv6(flow, skb, dir, ip6h);
a908fdec	481
a908fdec	482	ip6h->hop_limit--;
de20900f	483	skb->tstamp = 0;
a908fdec	484
53c2b289 PNA	485	if (flow_table->flags & NF_FLOWTABLE_COUNTER)
	486	nf_ct_acct_update(flow->ct, tuplehash->tuple.dir, skb->len);
	487
589b474a FW	488	if (unlikely(dst_xfrm(&rt->dst))) {
	489	memset(skb->cb, 0, sizeof(struct inet6_skb_parm));
	490	IP6CB(skb)->iif = skb->dev->ifindex;
	491	IP6CB(skb)->flags = IP6SKB_FORWARDED;
	492	return nf_flow_xmit_xfrm(skb, state, &rt->dst);
	493	}
	494
a908fdec FF	495	skb->dev = outdev;
a908fdec FF	496	nexthop = rt6_nexthop(rt, &flow->tuplehash[!dir].tuple.src_v6);
2a79fd39	497	skb_dst_set_noref(skb, &rt->dst);
a908fdec FF	498	neigh_xmit(NEIGH_ND_TABLE, outdev, nexthop, skb);
	499
	500	return NF_STOLEN;
	501	}
	502	EXPORT_SYMBOL_GPL(nf_flow_offload_ipv6_hook);