[deliverable/linux.git] / net / ipv6 / netfilter / nf_nat_l3proto_ipv6.c

/*
 * Copyright (c) 2011 Patrick McHardy <kaber@trash.net>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * Development of IPv6 NAT funded by Astaro.
 */
#include <linux/types.h>
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/ipv6.h>
#include <linux/netfilter.h>
#include <linux/netfilter_ipv6.h>
#include <net/secure_seq.h>
#include <net/checksum.h>
#include <net/ip6_checksum.h>
#include <net/ip6_route.h>
#include <net/ipv6.h>

#include <net/netfilter/nf_conntrack_core.h>
#include <net/netfilter/nf_conntrack.h>
#include <net/netfilter/nf_nat_core.h>
#include <net/netfilter/nf_nat_l3proto.h>
#include <net/netfilter/nf_nat_l4proto.h>

static const struct nf_nat_l3proto nf_nat_l3proto_ipv6;

#ifdef CONFIG_XFRM
static void nf_nat_ipv6_decode_session(struct sk_buff *skb,
				       const struct nf_conn *ct,
				       enum ip_conntrack_dir dir,
				       unsigned long statusbit,
				       struct flowi *fl)
{
	const struct nf_conntrack_tuple *t = &ct->tuplehash[dir].tuple;
	struct flowi6 *fl6 = &fl->u.ip6;

	if (ct->status & statusbit) {
		fl6->daddr = t->dst.u3.in6;
		if (t->dst.protonum == IPPROTO_TCP ||
		    t->dst.protonum == IPPROTO_UDP ||
		    t->dst.protonum == IPPROTO_UDPLITE ||
		    t->dst.protonum == IPPROTO_DCCP ||
		    t->dst.protonum == IPPROTO_SCTP)
			fl6->fl6_dport = t->dst.u.all;
	}

	statusbit ^= IPS_NAT_MASK;

	if (ct->status & statusbit) {
		fl6->saddr = t->src.u3.in6;
		if (t->dst.protonum == IPPROTO_TCP ||
		    t->dst.protonum == IPPROTO_UDP ||
		    t->dst.protonum == IPPROTO_UDPLITE ||
		    t->dst.protonum == IPPROTO_DCCP ||
		    t->dst.protonum == IPPROTO_SCTP)
			fl6->fl6_sport = t->src.u.all;
	}
}
#endif

static bool nf_nat_ipv6_in_range(const struct nf_conntrack_tuple *t,
				 const struct nf_nat_range *range)
{
	return ipv6_addr_cmp(&t->src.u3.in6, &range->min_addr.in6) >= 0 &&
	       ipv6_addr_cmp(&t->src.u3.in6, &range->max_addr.in6) <= 0;
}

static u32 nf_nat_ipv6_secure_port(const struct nf_conntrack_tuple *t,
				   __be16 dport)
{
	return secure_ipv6_port_ephemeral(t->src.u3.ip6, t->dst.u3.ip6, dport);
}

static bool nf_nat_ipv6_manip_pkt(struct sk_buff *skb,
				  unsigned int iphdroff,
				  const struct nf_nat_l4proto *l4proto,
				  const struct nf_conntrack_tuple *target,
				  enum nf_nat_manip_type maniptype)
{
	struct ipv6hdr *ipv6h;
	__be16 frag_off;
	int hdroff;
	u8 nexthdr;

	if (!skb_make_writable(skb, iphdroff + sizeof(*ipv6h)))
		return false;

	ipv6h = (void *)skb->data + iphdroff;
	nexthdr = ipv6h->nexthdr;
	hdroff = ipv6_skip_exthdr(skb, iphdroff + sizeof(*ipv6h),
				  &nexthdr, &frag_off);
	if (hdroff < 0)
		goto manip_addr;

	if ((frag_off & htons(~0x7)) == 0 &&
	    !l4proto->manip_pkt(skb, &nf_nat_l3proto_ipv6, iphdroff, hdroff,
				target, maniptype))
		return false;
manip_addr:
	if (maniptype == NF_NAT_MANIP_SRC)
		ipv6h->saddr = target->src.u3.in6;
	else
		ipv6h->daddr = target->dst.u3.in6;

	return true;
}

static void nf_nat_ipv6_csum_update(struct sk_buff *skb,
				    unsigned int iphdroff, __sum16 *check,
				    const struct nf_conntrack_tuple *t,
				    enum nf_nat_manip_type maniptype)
{
	const struct ipv6hdr *ipv6h = (struct ipv6hdr *)(skb->data + iphdroff);
	const struct in6_addr *oldip, *newip;

	if (maniptype == NF_NAT_MANIP_SRC) {
		oldip = &ipv6h->saddr;
		newip = &t->src.u3.in6;
	} else {
		oldip = &ipv6h->daddr;
		newip = &t->dst.u3.in6;
	}
	inet_proto_csum_replace16(check, skb, oldip->s6_addr32,
				  newip->s6_addr32, true);
}

static void nf_nat_ipv6_csum_recalc(struct sk_buff *skb,
				    u8 proto, void *data, __sum16 *check,
				    int datalen, int oldlen)
{
	const struct ipv6hdr *ipv6h = ipv6_hdr(skb);
	struct rt6_info *rt = (struct rt6_info *)skb_dst(skb);

	if (skb->ip_summed != CHECKSUM_PARTIAL) {
		if (!(rt->rt6i_flags & RTF_LOCAL) &&
		    (!skb->dev || skb->dev->features & NETIF_F_V6_CSUM)) {
			skb->ip_summed = CHECKSUM_PARTIAL;
			skb->csum_start = skb_headroom(skb) +
					  skb_network_offset(skb) +
					  (data - (void *)skb->data);
			skb->csum_offset = (void *)check - data;
			*check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
						  datalen, proto, 0);
		} else {
			*check = 0;
			*check = csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
						 datalen, proto,
						 csum_partial(data, datalen,
							      0));
			if (proto == IPPROTO_UDP && !*check)
				*check = CSUM_MANGLED_0;
		}
	} else
		inet_proto_csum_replace2(check, skb,
					 htons(oldlen), htons(datalen), true);
}

#if IS_ENABLED(CONFIG_NF_CT_NETLINK)
static int nf_nat_ipv6_nlattr_to_range(struct nlattr *tb[],
				       struct nf_nat_range *range)
{
	if (tb[CTA_NAT_V6_MINIP]) {
		nla_memcpy(&range->min_addr.ip6, tb[CTA_NAT_V6_MINIP],
			   sizeof(struct in6_addr));
		range->flags |= NF_NAT_RANGE_MAP_IPS;
	}

	if (tb[CTA_NAT_V6_MAXIP])
		nla_memcpy(&range->max_addr.ip6, tb[CTA_NAT_V6_MAXIP],
			   sizeof(struct in6_addr));
	else
		range->max_addr = range->min_addr;

	return 0;
}
#endif

static const struct nf_nat_l3proto nf_nat_l3proto_ipv6 = {
	.l3proto		= NFPROTO_IPV6,
	.secure_port		= nf_nat_ipv6_secure_port,
	.in_range		= nf_nat_ipv6_in_range,
	.manip_pkt		= nf_nat_ipv6_manip_pkt,
	.csum_update		= nf_nat_ipv6_csum_update,
	.csum_recalc		= nf_nat_ipv6_csum_recalc,
#if IS_ENABLED(CONFIG_NF_CT_NETLINK)
	.nlattr_to_range	= nf_nat_ipv6_nlattr_to_range,
#endif
#ifdef CONFIG_XFRM
	.decode_session	= nf_nat_ipv6_decode_session,
#endif
};

int nf_nat_icmpv6_reply_translation(struct sk_buff *skb,
				    struct nf_conn *ct,
				    enum ip_conntrack_info ctinfo,
				    unsigned int hooknum,
				    unsigned int hdrlen)
{
	struct {
		struct icmp6hdr	icmp6;
		struct ipv6hdr	ip6;
	} *inside;
	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
	enum nf_nat_manip_type manip = HOOK2MANIP(hooknum);
	const struct nf_nat_l4proto *l4proto;
	struct nf_conntrack_tuple target;
	unsigned long statusbit;

	NF_CT_ASSERT(ctinfo == IP_CT_RELATED || ctinfo == IP_CT_RELATED_REPLY);

	if (!skb_make_writable(skb, hdrlen + sizeof(*inside)))
		return 0;
	if (nf_ip6_checksum(skb, hooknum, hdrlen, IPPROTO_ICMPV6))
		return 0;

	inside = (void *)skb->data + hdrlen;
	if (inside->icmp6.icmp6_type == NDISC_REDIRECT) {
		if ((ct->status & IPS_NAT_DONE_MASK) != IPS_NAT_DONE_MASK)
			return 0;
		if (ct->status & IPS_NAT_MASK)
			return 0;
	}

	if (manip == NF_NAT_MANIP_SRC)
		statusbit = IPS_SRC_NAT;
	else
		statusbit = IPS_DST_NAT;

	/* Invert if this is reply direction */
	if (dir == IP_CT_DIR_REPLY)
		statusbit ^= IPS_NAT_MASK;

	if (!(ct->status & statusbit))
		return 1;

	l4proto = __nf_nat_l4proto_find(NFPROTO_IPV6, inside->ip6.nexthdr);
	if (!nf_nat_ipv6_manip_pkt(skb, hdrlen + sizeof(inside->icmp6),
				   l4proto, &ct->tuplehash[!dir].tuple, !manip))
		return 0;

	if (skb->ip_summed != CHECKSUM_PARTIAL) {
		struct ipv6hdr *ipv6h = ipv6_hdr(skb);
		inside = (void *)skb->data + hdrlen;
		inside->icmp6.icmp6_cksum = 0;
		inside->icmp6.icmp6_cksum =
			csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
					skb->len - hdrlen, IPPROTO_ICMPV6,
					csum_partial(&inside->icmp6,
						     skb->len - hdrlen, 0));
	}

	nf_ct_invert_tuplepr(&target, &ct->tuplehash[!dir].tuple);
	l4proto = __nf_nat_l4proto_find(NFPROTO_IPV6, IPPROTO_ICMPV6);
	if (!nf_nat_ipv6_manip_pkt(skb, 0, l4proto, &target, manip))
		return 0;

	return 1;
}
EXPORT_SYMBOL_GPL(nf_nat_icmpv6_reply_translation);

unsigned int
nf_nat_ipv6_fn(void *priv, struct sk_buff *skb,
	       const struct nf_hook_state *state,
	       unsigned int (*do_chain)(void *priv,
					struct sk_buff *skb,
					const struct nf_hook_state *state,
					struct nf_conn *ct))
{
	struct nf_conn *ct;
	enum ip_conntrack_info ctinfo;
	struct nf_conn_nat *nat;
	enum nf_nat_manip_type maniptype = HOOK2MANIP(state->hook);
	__be16 frag_off;
	int hdrlen;
	u8 nexthdr;

	ct = nf_ct_get(skb, &ctinfo);
	/* Can't track?  It's not due to stress, or conntrack would
	 * have dropped it.  Hence it's the user's responsibilty to
	 * packet filter it out, or implement conntrack/NAT for that
	 * protocol. 8) --RR
	 */
	if (!ct)
		return NF_ACCEPT;

	/* Don't try to NAT if this packet is not conntracked */
	if (nf_ct_is_untracked(ct))
		return NF_ACCEPT;

	nat = nf_ct_nat_ext_add(ct);
	if (nat == NULL)
		return NF_ACCEPT;

	switch (ctinfo) {
	case IP_CT_RELATED:
	case IP_CT_RELATED_REPLY:
		nexthdr = ipv6_hdr(skb)->nexthdr;
		hdrlen = ipv6_skip_exthdr(skb, sizeof(struct ipv6hdr),
					  &nexthdr, &frag_off);

		if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
			if (!nf_nat_icmpv6_reply_translation(skb, ct, ctinfo,
							     state->hook,
							     hdrlen))
				return NF_DROP;
			else
				return NF_ACCEPT;
		}
		/* Fall thru... (Only ICMPs can be IP_CT_IS_REPLY) */
	case IP_CT_NEW:
		/* Seen it before?  This can happen for loopback, retrans,
		 * or local packets.
		 */
		if (!nf_nat_initialized(ct, maniptype)) {
			unsigned int ret;

			ret = do_chain(priv, skb, state, ct);
			if (ret != NF_ACCEPT)
				return ret;

			if (nf_nat_initialized(ct, HOOK2MANIP(state->hook)))
				break;

			ret = nf_nat_alloc_null_binding(ct, state->hook);
			if (ret != NF_ACCEPT)
				return ret;
		} else {
			pr_debug("Already setup manip %s for ct %p\n",
				 maniptype == NF_NAT_MANIP_SRC ? "SRC" : "DST",
				 ct);
			if (nf_nat_oif_changed(state->hook, ctinfo, nat, state->out))
				goto oif_changed;
		}
		break;

	default:
		/* ESTABLISHED */
		NF_CT_ASSERT(ctinfo == IP_CT_ESTABLISHED ||
			     ctinfo == IP_CT_ESTABLISHED_REPLY);
		if (nf_nat_oif_changed(state->hook, ctinfo, nat, state->out))
			goto oif_changed;
	}

	return nf_nat_packet(ct, ctinfo, state->hook, skb);

oif_changed:
	nf_ct_kill_acct(ct, ctinfo, skb);
	return NF_DROP;
}
EXPORT_SYMBOL_GPL(nf_nat_ipv6_fn);

unsigned int
nf_nat_ipv6_in(void *priv, struct sk_buff *skb,
	       const struct nf_hook_state *state,
	       unsigned int (*do_chain)(void *priv,
					struct sk_buff *skb,
					const struct nf_hook_state *state,
					struct nf_conn *ct))
{
	unsigned int ret;
	struct in6_addr daddr = ipv6_hdr(skb)->daddr;

	ret = nf_nat_ipv6_fn(priv, skb, state, do_chain);
	if (ret != NF_DROP && ret != NF_STOLEN &&
	    ipv6_addr_cmp(&daddr, &ipv6_hdr(skb)->daddr))
		skb_dst_drop(skb);

	return ret;
}
EXPORT_SYMBOL_GPL(nf_nat_ipv6_in);

unsigned int
nf_nat_ipv6_out(void *priv, struct sk_buff *skb,
		const struct nf_hook_state *state,
		unsigned int (*do_chain)(void *priv,
					 struct sk_buff *skb,
					 const struct nf_hook_state *state,
					 struct nf_conn *ct))
{
#ifdef CONFIG_XFRM
	const struct nf_conn *ct;
	enum ip_conntrack_info ctinfo;
	int err;
#endif
	unsigned int ret;

	/* root is playing with raw sockets. */
	if (skb->len < sizeof(struct ipv6hdr))
		return NF_ACCEPT;

	ret = nf_nat_ipv6_fn(priv, skb, state, do_chain);
#ifdef CONFIG_XFRM
	if (ret != NF_DROP && ret != NF_STOLEN &&
	    !(IP6CB(skb)->flags & IP6SKB_XFRM_TRANSFORMED) &&
	    (ct = nf_ct_get(skb, &ctinfo)) != NULL) {
		enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);

		if (!nf_inet_addr_cmp(&ct->tuplehash[dir].tuple.src.u3,
				      &ct->tuplehash[!dir].tuple.dst.u3) ||
		    (ct->tuplehash[dir].tuple.dst.protonum != IPPROTO_ICMPV6 &&
		     ct->tuplehash[dir].tuple.src.u.all !=
		     ct->tuplehash[!dir].tuple.dst.u.all)) {
			err = nf_xfrm_me_harder(state->net, skb, AF_INET6);
			if (err < 0)
				ret = NF_DROP_ERR(err);
		}
	}
#endif
	return ret;
}
EXPORT_SYMBOL_GPL(nf_nat_ipv6_out);

unsigned int
nf_nat_ipv6_local_fn(void *priv, struct sk_buff *skb,
		     const struct nf_hook_state *state,
		     unsigned int (*do_chain)(void *priv,
					      struct sk_buff *skb,
					      const struct nf_hook_state *state,
					      struct nf_conn *ct))
{
	const struct nf_conn *ct;
	enum ip_conntrack_info ctinfo;
	unsigned int ret;
	int err;

	/* root is playing with raw sockets. */
	if (skb->len < sizeof(struct ipv6hdr))
		return NF_ACCEPT;

	ret = nf_nat_ipv6_fn(priv, skb, state, do_chain);
	if (ret != NF_DROP && ret != NF_STOLEN &&
	    (ct = nf_ct_get(skb, &ctinfo)) != NULL) {
		enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);

		if (!nf_inet_addr_cmp(&ct->tuplehash[dir].tuple.dst.u3,
				      &ct->tuplehash[!dir].tuple.src.u3)) {
			err = ip6_route_me_harder(state->net, skb);
			if (err < 0)
				ret = NF_DROP_ERR(err);
		}
#ifdef CONFIG_XFRM
		else if (!(IP6CB(skb)->flags & IP6SKB_XFRM_TRANSFORMED) &&
			 ct->tuplehash[dir].tuple.dst.protonum != IPPROTO_ICMPV6 &&
			 ct->tuplehash[dir].tuple.dst.u.all !=
			 ct->tuplehash[!dir].tuple.src.u.all) {
			err = nf_xfrm_me_harder(state->net, skb, AF_INET6);
			if (err < 0)
				ret = NF_DROP_ERR(err);
		}
#endif
	}
	return ret;
}
EXPORT_SYMBOL_GPL(nf_nat_ipv6_local_fn);

static int __init nf_nat_l3proto_ipv6_init(void)
{
	int err;

	err = nf_nat_l4proto_register(NFPROTO_IPV6, &nf_nat_l4proto_icmpv6);
	if (err < 0)
		goto err1;
	err = nf_nat_l3proto_register(&nf_nat_l3proto_ipv6);
	if (err < 0)
		goto err2;
	return err;

err2:
	nf_nat_l4proto_unregister(NFPROTO_IPV6, &nf_nat_l4proto_icmpv6);
err1:
	return err;
}

static void __exit nf_nat_l3proto_ipv6_exit(void)
{
	nf_nat_l3proto_unregister(&nf_nat_l3proto_ipv6);
	nf_nat_l4proto_unregister(NFPROTO_IPV6, &nf_nat_l4proto_icmpv6);
}

MODULE_LICENSE("GPL");
MODULE_ALIAS("nf-nat-" __stringify(AF_INET6));

module_init(nf_nat_l3proto_ipv6_init);
module_exit(nf_nat_l3proto_ipv6_exit);
Commit	Line	Data
58a317f1 PM	1	/*
	2	* Copyright (c) 2011 Patrick McHardy <kaber@trash.net>
	3	*
	4	* This program is free software; you can redistribute it and/or modify
	5	* it under the terms of the GNU General Public License version 2 as
	6	* published by the Free Software Foundation.
	7	*
	8	* Development of IPv6 NAT funded by Astaro.
	9	*/
	10	#include <linux/types.h>
	11	#include <linux/module.h>
	12	#include <linux/skbuff.h>
	13	#include <linux/ipv6.h>
	14	#include <linux/netfilter.h>
	15	#include <linux/netfilter_ipv6.h>
	16	#include <net/secure_seq.h>
	17	#include <net/checksum.h>
2c969322	18	#include <net/ip6_checksum.h>
58a317f1 PM	19	#include <net/ip6_route.h>
	20	#include <net/ipv6.h>
	21
	22	#include <net/netfilter/nf_conntrack_core.h>
	23	#include <net/netfilter/nf_conntrack.h>
	24	#include <net/netfilter/nf_nat_core.h>
	25	#include <net/netfilter/nf_nat_l3proto.h>
	26	#include <net/netfilter/nf_nat_l4proto.h>
	27
	28	static const struct nf_nat_l3proto nf_nat_l3proto_ipv6;
	29
	30	#ifdef CONFIG_XFRM
	31	static void nf_nat_ipv6_decode_session(struct sk_buff *skb,
	32	const struct nf_conn *ct,
	33	enum ip_conntrack_dir dir,
	34	unsigned long statusbit,
	35	struct flowi *fl)
	36	{
	37	const struct nf_conntrack_tuple *t = &ct->tuplehash[dir].tuple;
	38	struct flowi6 *fl6 = &fl->u.ip6;
	39
	40	if (ct->status & statusbit) {
	41	fl6->daddr = t->dst.u3.in6;
	42	if (t->dst.protonum == IPPROTO_TCP \|\|
	43	t->dst.protonum == IPPROTO_UDP \|\|
	44	t->dst.protonum == IPPROTO_UDPLITE \|\|
	45	t->dst.protonum == IPPROTO_DCCP \|\|
	46	t->dst.protonum == IPPROTO_SCTP)
	47	fl6->fl6_dport = t->dst.u.all;
	48	}
	49
	50	statusbit ^= IPS_NAT_MASK;
	51
	52	if (ct->status & statusbit) {
	53	fl6->saddr = t->src.u3.in6;
	54	if (t->dst.protonum == IPPROTO_TCP \|\|
	55	t->dst.protonum == IPPROTO_UDP \|\|
	56	t->dst.protonum == IPPROTO_UDPLITE \|\|
	57	t->dst.protonum == IPPROTO_DCCP \|\|
	58	t->dst.protonum == IPPROTO_SCTP)
	59	fl6->fl6_sport = t->src.u.all;
	60	}
	61	}
	62	#endif
	63
	64	static bool nf_nat_ipv6_in_range(const struct nf_conntrack_tuple *t,
	65	const struct nf_nat_range *range)
	66	{
	67	return ipv6_addr_cmp(&t->src.u3.in6, &range->min_addr.in6) >= 0 &&
	68	ipv6_addr_cmp(&t->src.u3.in6, &range->max_addr.in6) <= 0;
	69	}
	70
	71	static u32 nf_nat_ipv6_secure_port(const struct nf_conntrack_tuple *t,
	72	__be16 dport)
	73	{
	74	return secure_ipv6_port_ephemeral(t->src.u3.ip6, t->dst.u3.ip6, dport);
	75	}
	76
	77	static bool nf_nat_ipv6_manip_pkt(struct sk_buff *skb,
	78	unsigned int iphdroff,
	79	const struct nf_nat_l4proto *l4proto,
	80	const struct nf_conntrack_tuple *target,
	81	enum nf_nat_manip_type maniptype)
	82	{
83	struct ipv6hdr *ipv6h;
84	__be16 frag_off;
85	int hdroff;
86	u8 nexthdr;
87
88	if (!skb_make_writable(skb, iphdroff + sizeof(*ipv6h)))
89	return false;
90
91	ipv6h = (void *)skb->data + iphdroff;
92	nexthdr = ipv6h->nexthdr;
93	hdroff = ipv6_skip_exthdr(skb, iphdroff + sizeof(*ipv6h),
94	&nexthdr, &frag_off);
95	if (hdroff < 0)
96	goto manip_addr;
97
98	if ((frag_off & htons(~0x7)) == 0 &&
99	!l4proto->manip_pkt(skb, &nf_nat_l3proto_ipv6, iphdroff, hdroff,
100	target, maniptype))
101	return false;
102	manip_addr:
103	if (maniptype == NF_NAT_MANIP_SRC)
104	ipv6h->saddr = target->src.u3.in6;
105	else
106	ipv6h->daddr = target->dst.u3.in6;
107
108	return true;
109	}
110
111	static void nf_nat_ipv6_csum_update(struct sk_buff *skb,
112	unsigned int iphdroff, __sum16 *check,
113	const struct nf_conntrack_tuple *t,
114	enum nf_nat_manip_type maniptype)
115	{
116	const struct ipv6hdr ipv6h = (struct ipv6hdr )(skb->data + iphdroff);
117	const struct in6_addr oldip, newip;
118
119	if (maniptype == NF_NAT_MANIP_SRC) {
120	oldip = &ipv6h->saddr;
121	newip = &t->src.u3.in6;
122	} else {
123	oldip = &ipv6h->daddr;
124	newip = &t->dst.u3.in6;
125	}
126	inet_proto_csum_replace16(check, skb, oldip->s6_addr32,
4b048d6d	127	newip->s6_addr32, true);
58a317f1 PM	128	}
	129
	130	static void nf_nat_ipv6_csum_recalc(struct sk_buff *skb,
	131	u8 proto, void data, __sum16 check,
	132	int datalen, int oldlen)
	133	{
	134	const struct ipv6hdr *ipv6h = ipv6_hdr(skb);
	135	struct rt6_info rt = (struct rt6_info )skb_dst(skb);
	136
	137	if (skb->ip_summed != CHECKSUM_PARTIAL) {
	138	if (!(rt->rt6i_flags & RTF_LOCAL) &&
	139	(!skb->dev \|\| skb->dev->features & NETIF_F_V6_CSUM)) {
	140	skb->ip_summed = CHECKSUM_PARTIAL;
	141	skb->csum_start = skb_headroom(skb) +
	142	skb_network_offset(skb) +
	143	(data - (void *)skb->data);
	144	skb->csum_offset = (void *)check - data;
	145	*check = ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
	146	datalen, proto, 0);
	147	} else {
	148	*check = 0;
	149	*check = csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
	150	datalen, proto,
	151	csum_partial(data, datalen,
	152	0));
	153	if (proto == IPPROTO_UDP && !*check)
	154	*check = CSUM_MANGLED_0;
	155	}
	156	} else
	157	inet_proto_csum_replace2(check, skb,
4b048d6d	158	htons(oldlen), htons(datalen), true);
58a317f1 PM	159	}
58a317f1 PM	160
24de3d37	161	#if IS_ENABLED(CONFIG_NF_CT_NETLINK)
58a317f1 PM	162	static int nf_nat_ipv6_nlattr_to_range(struct nlattr *tb[],
	163	struct nf_nat_range *range)
	164	{
	165	if (tb[CTA_NAT_V6_MINIP]) {
	166	nla_memcpy(&range->min_addr.ip6, tb[CTA_NAT_V6_MINIP],
	167	sizeof(struct in6_addr));
	168	range->flags \|= NF_NAT_RANGE_MAP_IPS;
	169	}
	170
	171	if (tb[CTA_NAT_V6_MAXIP])
	172	nla_memcpy(&range->max_addr.ip6, tb[CTA_NAT_V6_MAXIP],
	173	sizeof(struct in6_addr));
	174	else
	175	range->max_addr = range->min_addr;
	176
	177	return 0;
	178	}
24de3d37	179	#endif
58a317f1 PM	180
	181	static const struct nf_nat_l3proto nf_nat_l3proto_ipv6 = {
	182	.l3proto = NFPROTO_IPV6,
	183	.secure_port = nf_nat_ipv6_secure_port,
	184	.in_range = nf_nat_ipv6_in_range,
	185	.manip_pkt = nf_nat_ipv6_manip_pkt,
	186	.csum_update = nf_nat_ipv6_csum_update,
	187	.csum_recalc = nf_nat_ipv6_csum_recalc,
24de3d37	188	#if IS_ENABLED(CONFIG_NF_CT_NETLINK)
58a317f1	189	.nlattr_to_range = nf_nat_ipv6_nlattr_to_range,
24de3d37	190	#endif
58a317f1 PM	191	#ifdef CONFIG_XFRM
	192	.decode_session = nf_nat_ipv6_decode_session,
	193	#endif
	194	};
	195
	196	int nf_nat_icmpv6_reply_translation(struct sk_buff *skb,
	197	struct nf_conn *ct,
	198	enum ip_conntrack_info ctinfo,
	199	unsigned int hooknum,
	200	unsigned int hdrlen)
	201	{
	202	struct {
	203	struct icmp6hdr icmp6;
	204	struct ipv6hdr ip6;
	205	} *inside;
	206	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
	207	enum nf_nat_manip_type manip = HOOK2MANIP(hooknum);
	208	const struct nf_nat_l4proto *l4proto;
	209	struct nf_conntrack_tuple target;
	210	unsigned long statusbit;
	211
	212	NF_CT_ASSERT(ctinfo == IP_CT_RELATED \|\| ctinfo == IP_CT_RELATED_REPLY);
	213
	214	if (!skb_make_writable(skb, hdrlen + sizeof(*inside)))
	215	return 0;
	216	if (nf_ip6_checksum(skb, hooknum, hdrlen, IPPROTO_ICMPV6))
	217	return 0;
	218
	219	inside = (void *)skb->data + hdrlen;
	220	if (inside->icmp6.icmp6_type == NDISC_REDIRECT) {
	221	if ((ct->status & IPS_NAT_DONE_MASK) != IPS_NAT_DONE_MASK)
	222	return 0;
	223	if (ct->status & IPS_NAT_MASK)
	224	return 0;
	225	}
	226
	227	if (manip == NF_NAT_MANIP_SRC)
	228	statusbit = IPS_SRC_NAT;
	229	else
	230	statusbit = IPS_DST_NAT;
	231
	232	/* Invert if this is reply direction */
	233	if (dir == IP_CT_DIR_REPLY)
	234	statusbit ^= IPS_NAT_MASK;
	235
	236	if (!(ct->status & statusbit))
	237	return 1;
	238
	239	l4proto = __nf_nat_l4proto_find(NFPROTO_IPV6, inside->ip6.nexthdr);
	240	if (!nf_nat_ipv6_manip_pkt(skb, hdrlen + sizeof(inside->icmp6),
	241	l4proto, &ct->tuplehash[!dir].tuple, !manip))
	242	return 0;
	243
	244	if (skb->ip_summed != CHECKSUM_PARTIAL) {
	245	struct ipv6hdr *ipv6h = ipv6_hdr(skb);
	246	inside = (void *)skb->data + hdrlen;
	247	inside->icmp6.icmp6_cksum = 0;
	248	inside->icmp6.icmp6_cksum =
	249	csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr,
	250	skb->len - hdrlen, IPPROTO_ICMPV6,
	251	csum_partial(&inside->icmp6,
	252	skb->len - hdrlen, 0));
	253	}
	254
255	nf_ct_invert_tuplepr(&target, &ct->tuplehash[!dir].tuple);
256	l4proto = __nf_nat_l4proto_find(NFPROTO_IPV6, IPPROTO_ICMPV6);
257	if (!nf_nat_ipv6_manip_pkt(skb, 0, l4proto, &target, manip))
258	return 0;
259
260	return 1;
261	}
262	EXPORT_SYMBOL_GPL(nf_nat_icmpv6_reply_translation);
263
2a5538e9	264	unsigned int
06198b34	265	nf_nat_ipv6_fn(void priv, struct sk_buff skb,
8fe22382	266	const struct nf_hook_state *state,
06198b34	267	unsigned int (do_chain)(void priv,
2a5538e9	268	struct sk_buff *skb,
8fe22382	269	const struct nf_hook_state *state,
2a5538e9 PNA	270	struct nf_conn *ct))
	271	{
	272	struct nf_conn *ct;
	273	enum ip_conntrack_info ctinfo;
	274	struct nf_conn_nat *nat;
082a758f	275	enum nf_nat_manip_type maniptype = HOOK2MANIP(state->hook);
2a5538e9 PNA	276	__be16 frag_off;
	277	int hdrlen;
	278	u8 nexthdr;
	279
	280	ct = nf_ct_get(skb, &ctinfo);
	281	/* Can't track? It's not due to stress, or conntrack would
	282	* have dropped it. Hence it's the user's responsibilty to
	283	* packet filter it out, or implement conntrack/NAT for that
	284	* protocol. 8) --RR
	285	*/
	286	if (!ct)
	287	return NF_ACCEPT;
	288
	289	/* Don't try to NAT if this packet is not conntracked */
	290	if (nf_ct_is_untracked(ct))
	291	return NF_ACCEPT;
	292
	293	nat = nf_ct_nat_ext_add(ct);
	294	if (nat == NULL)
	295	return NF_ACCEPT;
	296
	297	switch (ctinfo) {
	298	case IP_CT_RELATED:
	299	case IP_CT_RELATED_REPLY:
	300	nexthdr = ipv6_hdr(skb)->nexthdr;
	301	hdrlen = ipv6_skip_exthdr(skb, sizeof(struct ipv6hdr),
	302	&nexthdr, &frag_off);
	303
	304	if (hdrlen >= 0 && nexthdr == IPPROTO_ICMPV6) {
	305	if (!nf_nat_icmpv6_reply_translation(skb, ct, ctinfo,
082a758f	306	state->hook,
2a5538e9 PNA	307	hdrlen))
	308	return NF_DROP;
	309	else
	310	return NF_ACCEPT;
	311	}
	312	/* Fall thru... (Only ICMPs can be IP_CT_IS_REPLY) */
	313	case IP_CT_NEW:
	314	/* Seen it before? This can happen for loopback, retrans,
	315	* or local packets.
	316	*/
	317	if (!nf_nat_initialized(ct, maniptype)) {
	318	unsigned int ret;
	319
06198b34	320	ret = do_chain(priv, skb, state, ct);
2a5538e9 PNA	321	if (ret != NF_ACCEPT)
	322	return ret;
	323
082a758f	324	if (nf_nat_initialized(ct, HOOK2MANIP(state->hook)))
2a5538e9 PNA	325	break;
2a5538e9 PNA	326
082a758f	327	ret = nf_nat_alloc_null_binding(ct, state->hook);
2a5538e9 PNA	328	if (ret != NF_ACCEPT)
	329	return ret;
	330	} else {
	331	pr_debug("Already setup manip %s for ct %p\n",
	332	maniptype == NF_NAT_MANIP_SRC ? "SRC" : "DST",
	333	ct);
082a758f	334	if (nf_nat_oif_changed(state->hook, ctinfo, nat, state->out))
2a5538e9 PNA	335	goto oif_changed;
	336	}
	337	break;
	338
	339	default:
	340	/* ESTABLISHED */
	341	NF_CT_ASSERT(ctinfo == IP_CT_ESTABLISHED \|\|
	342	ctinfo == IP_CT_ESTABLISHED_REPLY);
082a758f	343	if (nf_nat_oif_changed(state->hook, ctinfo, nat, state->out))
2a5538e9 PNA	344	goto oif_changed;
	345	}
	346
082a758f	347	return nf_nat_packet(ct, ctinfo, state->hook, skb);
2a5538e9 PNA	348
	349	oif_changed:
	350	nf_ct_kill_acct(ct, ctinfo, skb);
	351	return NF_DROP;
	352	}
	353	EXPORT_SYMBOL_GPL(nf_nat_ipv6_fn);
	354
	355	unsigned int
06198b34	356	nf_nat_ipv6_in(void priv, struct sk_buff skb,
8fe22382	357	const struct nf_hook_state *state,
06198b34	358	unsigned int (do_chain)(void priv,
2a5538e9	359	struct sk_buff *skb,
8fe22382	360	const struct nf_hook_state *state,
2a5538e9 PNA	361	struct nf_conn *ct))
	362	{
	363	unsigned int ret;
	364	struct in6_addr daddr = ipv6_hdr(skb)->daddr;
	365
06198b34	366	ret = nf_nat_ipv6_fn(priv, skb, state, do_chain);
2a5538e9 PNA	367	if (ret != NF_DROP && ret != NF_STOLEN &&
	368	ipv6_addr_cmp(&daddr, &ipv6_hdr(skb)->daddr))
	369	skb_dst_drop(skb);
	370
	371	return ret;
	372	}
	373	EXPORT_SYMBOL_GPL(nf_nat_ipv6_in);
	374
	375	unsigned int
06198b34	376	nf_nat_ipv6_out(void priv, struct sk_buff skb,
8fe22382	377	const struct nf_hook_state *state,
06198b34	378	unsigned int (do_chain)(void priv,
2a5538e9	379	struct sk_buff *skb,
8fe22382	380	const struct nf_hook_state *state,
2a5538e9 PNA	381	struct nf_conn *ct))
	382	{
	383	#ifdef CONFIG_XFRM
	384	const struct nf_conn *ct;
	385	enum ip_conntrack_info ctinfo;
	386	int err;
	387	#endif
	388	unsigned int ret;
	389
	390	/* root is playing with raw sockets. */
	391	if (skb->len < sizeof(struct ipv6hdr))
	392	return NF_ACCEPT;
	393
06198b34	394	ret = nf_nat_ipv6_fn(priv, skb, state, do_chain);
2a5538e9 PNA	395	#ifdef CONFIG_XFRM
	396	if (ret != NF_DROP && ret != NF_STOLEN &&
	397	!(IP6CB(skb)->flags & IP6SKB_XFRM_TRANSFORMED) &&
	398	(ct = nf_ct_get(skb, &ctinfo)) != NULL) {
	399	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
	400
	401	if (!nf_inet_addr_cmp(&ct->tuplehash[dir].tuple.src.u3,
	402	&ct->tuplehash[!dir].tuple.dst.u3) \|\|
	403	(ct->tuplehash[dir].tuple.dst.protonum != IPPROTO_ICMPV6 &&
	404	ct->tuplehash[dir].tuple.src.u.all !=
	405	ct->tuplehash[!dir].tuple.dst.u.all)) {
c7af6483	406	err = nf_xfrm_me_harder(state->net, skb, AF_INET6);
2a5538e9 PNA	407	if (err < 0)
	408	ret = NF_DROP_ERR(err);
	409	}
	410	}
	411	#endif
	412	return ret;
	413	}
	414	EXPORT_SYMBOL_GPL(nf_nat_ipv6_out);
	415
	416	unsigned int
06198b34	417	nf_nat_ipv6_local_fn(void priv, struct sk_buff skb,
8fe22382	418	const struct nf_hook_state *state,
06198b34	419	unsigned int (do_chain)(void priv,
2a5538e9	420	struct sk_buff *skb,
8fe22382	421	const struct nf_hook_state *state,
2a5538e9 PNA	422	struct nf_conn *ct))
	423	{
	424	const struct nf_conn *ct;
	425	enum ip_conntrack_info ctinfo;
	426	unsigned int ret;
	427	int err;
	428
	429	/* root is playing with raw sockets. */
	430	if (skb->len < sizeof(struct ipv6hdr))
	431	return NF_ACCEPT;
	432
06198b34	433	ret = nf_nat_ipv6_fn(priv, skb, state, do_chain);
2a5538e9 PNA	434	if (ret != NF_DROP && ret != NF_STOLEN &&
	435	(ct = nf_ct_get(skb, &ctinfo)) != NULL) {
	436	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
	437
	438	if (!nf_inet_addr_cmp(&ct->tuplehash[dir].tuple.dst.u3,
	439	&ct->tuplehash[!dir].tuple.src.u3)) {
5f5d74d7	440	err = ip6_route_me_harder(state->net, skb);
2a5538e9 PNA	441	if (err < 0)
	442	ret = NF_DROP_ERR(err);
	443	}
	444	#ifdef CONFIG_XFRM
	445	else if (!(IP6CB(skb)->flags & IP6SKB_XFRM_TRANSFORMED) &&
	446	ct->tuplehash[dir].tuple.dst.protonum != IPPROTO_ICMPV6 &&
	447	ct->tuplehash[dir].tuple.dst.u.all !=
	448	ct->tuplehash[!dir].tuple.src.u.all) {
c7af6483	449	err = nf_xfrm_me_harder(state->net, skb, AF_INET6);
2a5538e9 PNA	450	if (err < 0)
	451	ret = NF_DROP_ERR(err);
	452	}
	453	#endif
	454	}
	455	return ret;
	456	}
	457	EXPORT_SYMBOL_GPL(nf_nat_ipv6_local_fn);
	458
58a317f1 PM	459	static int __init nf_nat_l3proto_ipv6_init(void)
	460	{
	461	int err;
	462
	463	err = nf_nat_l4proto_register(NFPROTO_IPV6, &nf_nat_l4proto_icmpv6);
	464	if (err < 0)
	465	goto err1;
	466	err = nf_nat_l3proto_register(&nf_nat_l3proto_ipv6);
	467	if (err < 0)
	468	goto err2;
	469	return err;
	470
	471	err2:
	472	nf_nat_l4proto_unregister(NFPROTO_IPV6, &nf_nat_l4proto_icmpv6);
	473	err1:
	474	return err;
	475	}
	476
	477	static void __exit nf_nat_l3proto_ipv6_exit(void)
	478	{
	479	nf_nat_l3proto_unregister(&nf_nat_l3proto_ipv6);
	480	nf_nat_l4proto_unregister(NFPROTO_IPV6, &nf_nat_l4proto_icmpv6);
	481	}
	482
	483	MODULE_LICENSE("GPL");
	484	MODULE_ALIAS("nf-nat-" __stringify(AF_INET6));
	485
	486	module_init(nf_nat_l3proto_ipv6_init);
	487	module_exit(nf_nat_l3proto_ipv6_exit);