netfilter: move route indirection to struct nf_ipv6_ops

We cannot make a direct call to nf_ip6_route() because that would result
in autoloading the 'ipv6' module because of symbol dependencies.
Therefore, define route indirection in nf_ipv6_ops where this really
belongs to.

For IPv4, we can indeed make a direct function call, which is faster,
given IPv4 is built-in in the networking code by default. Still,
CONFIG_INET=n and CONFIG_NETFILTER=y is possible, so define empty inline
stub for IPv4 in such case.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

include/linux/netfilter.h

@@ -311,8 +311,6 @@ struct nf_queue_entry;
 
 struct nf_afinfo {
 	unsigned short	family;
-	int		(*route)(struct net *net, struct dst_entry **dst,
-				 struct flowi *fl, bool strict);
 	int		(*reroute)(struct net *net, struct sk_buff *skb,
 				   const struct nf_queue_entry *entry);
 	int		route_key_size;
@@ -331,6 +329,8 @@ __sum16 nf_checksum(struct sk_buff *skb, unsigned int hook,
 __sum16 nf_checksum_partial(struct sk_buff *skb, unsigned int hook,
 			    unsigned int dataoff, unsigned int len,
 			    u_int8_t protocol, unsigned short family);
+int nf_route(struct net *net, struct dst_entry **dst, struct flowi *fl,
+	     bool strict, unsigned short family);
 
 int nf_register_afinfo(const struct nf_afinfo *afinfo);
 void nf_unregister_afinfo(const struct nf_afinfo *afinfo);

include/linux/netfilter_ipv4.h

@@ -24,6 +24,8 @@ __sum16 nf_ip_checksum(struct sk_buff *skb, unsigned int hook,
 __sum16 nf_ip_checksum_partial(struct sk_buff *skb, unsigned int hook,
 			       unsigned int dataoff, unsigned int len,
 			       u_int8_t protocol);
+int nf_ip_route(struct net *net, struct dst_entry **dst, struct flowi *fl,
+		bool strict);
 #else
 static inline __sum16 nf_ip_checksum(struct sk_buff *skb, unsigned int hook,
 				     unsigned int dataoff, u_int8_t protocol)
@@ -38,6 +40,11 @@ static inline __sum16 nf_ip_checksum_partial(struct sk_buff *skb,
 {
 	return 0;
 }
+static inline int nf_ip_route(struct net *net, struct dst_entry **dst,
+			      struct flowi *fl, bool strict)
+{
+	return -EOPNOTSUPP;
+}
 #endif /* CONFIG_INET */
 
 #endif /*__LINUX_IP_NETFILTER_H*/

include/linux/netfilter_ipv6.h

@@ -33,6 +33,8 @@ struct nf_ipv6_ops {
 	__sum16 (*checksum_partial)(struct sk_buff *skb, unsigned int hook,
 				    unsigned int dataoff, unsigned int len,
 				    u_int8_t protocol);
+	int (*route)(struct net *net, struct dst_entry **dst, struct flowi *fl,
+		     bool strict);
 };
 
 #ifdef CONFIG_NETFILTER

net/bridge/netfilter/nf_tables_bridge.c

@@ -101,15 +101,8 @@ static int nf_br_reroute(struct net *net, struct sk_buff *skb,
 	return 0;
 }
 
-static int nf_br_route(struct net *net, struct dst_entry **dst,
-		       struct flowi *fl, bool strict __always_unused)
-{
-	return 0;
-}
-
 static const struct nf_afinfo nf_br_afinfo = {
 	.family                 = AF_BRIDGE,
-	.route                  = nf_br_route,
 	.reroute                = nf_br_reroute,
 	.route_key_size         = 0,
 };

net/ipv4/netfilter.c

@@ -150,8 +150,8 @@ __sum16 nf_ip_checksum_partial(struct sk_buff *skb, unsigned int hook,
 }
 EXPORT_SYMBOL_GPL(nf_ip_checksum_partial);
 
-static int nf_ip_route(struct net *net, struct dst_entry **dst,
+int nf_ip_route(struct net *net, struct dst_entry **dst, struct flowi *fl,
-		       struct flowi *fl, bool strict __always_unused)
+		bool strict __always_unused)
 {
 	struct rtable *rt = ip_route_output_key(net, &fl->u.ip4);
 	if (IS_ERR(rt))
@@ -159,10 +159,10 @@ static int nf_ip_route(struct net *net, struct dst_entry **dst,
 	*dst = &rt->dst;
 	return 0;
 }
+EXPORT_SYMBOL_GPL(nf_ip_route);
 
 static const struct nf_afinfo nf_ip_afinfo = {
 	.family			= AF_INET,
-	.route			= nf_ip_route,
 	.reroute		= nf_ip_reroute,
 	.route_key_size		= sizeof(struct ip_rt_info),
 };

net/ipv6/netfilter.c

@@ -170,11 +170,11 @@ static const struct nf_ipv6_ops ipv6ops = {
 	.fragment		= ip6_fragment,
 	.checksum		= nf_ip6_checksum,
 	.checksum_partial	= nf_ip6_checksum_partial,
+	.route			= nf_ip6_route,
 };
 
 static const struct nf_afinfo nf_ip6_afinfo = {
 	.family			= AF_INET6,
-	.route			= nf_ip6_route,
 	.reroute		= nf_ip6_reroute,
 	.route_key_size		= sizeof(struct ip6_rt_info),
 };

net/ipv6/netfilter/nft_fib_ipv6.c

@@ -60,7 +60,6 @@ static u32 __nft_fib6_eval_type(const struct nft_fib *priv,
 {
 	const struct net_device *dev = NULL;
 	const struct nf_ipv6_ops *v6ops;
-	const struct nf_afinfo *afinfo;
 	int route_err, addrtype;
 	struct rt6_info *rt;
 	struct flowi6 fl6 = {
@@ -69,8 +68,8 @@ static u32 __nft_fib6_eval_type(const struct nft_fib *priv,
 	};
 	u32 ret = 0;
 
-	afinfo = nf_get_afinfo(NFPROTO_IPV6);
+	v6ops = nf_get_ipv6_ops();
-	if (!afinfo)
+	if (!v6ops)
 		return RTN_UNREACHABLE;
 
 	if (priv->flags & NFTA_FIB_F_IIF)
@@ -80,12 +79,11 @@ static u32 __nft_fib6_eval_type(const struct nft_fib *priv,
 
 	nft_fib6_flowi_init(&fl6, priv, pkt, dev, iph);
 
-	v6ops = nf_get_ipv6_ops();
+	if (dev && v6ops->chk_addr(nft_net(pkt), &fl6.daddr, dev, true))
-	if (dev && v6ops && v6ops->chk_addr(nft_net(pkt), &fl6.daddr, dev, true))
 		ret = RTN_LOCAL;
 
-	route_err = afinfo->route(nft_net(pkt), (struct dst_entry **)&rt,
+	route_err = v6ops->route(nft_net(pkt), (struct dst_entry **)&rt,
-				  flowi6_to_flowi(&fl6), false);
+				 flowi6_to_flowi(&fl6), false);
 	if (route_err)
 		goto err;
 

net/netfilter/nf_conntrack_h323_main.c

@@ -24,6 +24,7 @@
 #include <linux/skbuff.h>
 #include <net/route.h>
 #include <net/ip6_route.h>
+#include <linux/netfilter_ipv6.h>
 
 #include <net/netfilter/nf_conntrack.h>
 #include <net/netfilter/nf_conntrack_core.h>
@@ -721,14 +722,8 @@ static int callforward_do_filter(struct net *net,
 				 const union nf_inet_addr *dst,
 				 u_int8_t family)
 {
-	const struct nf_afinfo *afinfo;
 	int ret = 0;
 
-	/* rcu_read_lock()ed by nf_hook_thresh */
-	afinfo = nf_get_afinfo(family);
-	if (!afinfo)
-		return 0;
-
 	switch (family) {
 	case AF_INET: {
 		struct flowi4 fl1, fl2;
@@ -739,10 +734,10 @@ static int callforward_do_filter(struct net *net,
 
 		memset(&fl2, 0, sizeof(fl2));
 		fl2.daddr = dst->ip;
-		if (!afinfo->route(net, (struct dst_entry **)&rt1,
+		if (!nf_ip_route(net, (struct dst_entry **)&rt1,
-				   flowi4_to_flowi(&fl1), false)) {
+				 flowi4_to_flowi(&fl1), false)) {
-			if (!afinfo->route(net, (struct dst_entry **)&rt2,
+			if (!nf_ip_route(net, (struct dst_entry **)&rt2,
-					   flowi4_to_flowi(&fl2), false)) {
+					 flowi4_to_flowi(&fl2), false)) {
 				if (rt_nexthop(rt1, fl1.daddr) ==
 				    rt_nexthop(rt2, fl2.daddr) &&
 				    rt1->dst.dev  == rt2->dst.dev)
@@ -755,18 +750,23 @@ static int callforward_do_filter(struct net *net,
 	}
 #if IS_ENABLED(CONFIG_NF_CONNTRACK_IPV6)
 	case AF_INET6: {
-		struct flowi6 fl1, fl2;
+		const struct nf_ipv6_ops *v6ops;
 		struct rt6_info *rt1, *rt2;
+		struct flowi6 fl1, fl2;
+
+		v6ops = nf_get_ipv6_ops();
+		if (!v6ops)
+			return 0;
 
 		memset(&fl1, 0, sizeof(fl1));
 		fl1.daddr = src->in6;
 
 		memset(&fl2, 0, sizeof(fl2));
 		fl2.daddr = dst->in6;
-		if (!afinfo->route(net, (struct dst_entry **)&rt1,
+		if (!v6ops->route(net, (struct dst_entry **)&rt1,
-				   flowi6_to_flowi(&fl1), false)) {
+				  flowi6_to_flowi(&fl1), false)) {
-			if (!afinfo->route(net, (struct dst_entry **)&rt2,
+			if (!v6ops->route(net, (struct dst_entry **)&rt2,
-					   flowi6_to_flowi(&fl2), false)) {
+					  flowi6_to_flowi(&fl2), false)) {
 				if (ipv6_addr_equal(rt6_nexthop(rt1, &fl1.daddr),
 						    rt6_nexthop(rt2, &fl2.daddr)) &&
 				    rt1->dst.dev == rt2->dst.dev)

net/netfilter/nft_rt.c

@@ -27,7 +27,7 @@ static u16 get_tcpmss(const struct nft_pktinfo *pkt, const struct dst_entry *skb
 {
 	u32 minlen = sizeof(struct ipv6hdr), mtu = dst_mtu(skbdst);
 	const struct sk_buff *skb = pkt->skb;
-	const struct nf_afinfo *ai;
+	struct dst_entry *dst = NULL;
 	struct flowi fl;
 
 	memset(&fl, 0, sizeof(fl));
@@ -43,15 +43,10 @@ static u16 get_tcpmss(const struct nft_pktinfo *pkt, const struct dst_entry *skb
 		break;
 	}
 
-	ai = nf_get_afinfo(nft_pf(pkt));
+	nf_route(nft_net(pkt), &dst, &fl, false, nft_pf(pkt));
-	if (ai) {
+	if (dst) {
-		struct dst_entry *dst = NULL;
+		mtu = min(mtu, dst_mtu(dst));
-
+		dst_release(dst);
-		ai->route(nft_net(pkt), &dst, &fl, false);
-		if (dst) {
-			mtu = min(mtu, dst_mtu(dst));
-			dst_release(dst);
-		}
 	}
 
 	if (mtu <= minlen || mtu > 0xffff)

net/netfilter/utils.c

@@ -48,3 +48,24 @@ __sum16 nf_checksum_partial(struct sk_buff *skb, unsigned int hook,
 	return csum;
 }
 EXPORT_SYMBOL_GPL(nf_checksum_partial);
+
+int nf_route(struct net *net, struct dst_entry **dst, struct flowi *fl,
+	     bool strict, unsigned short family)
+{
+	const struct nf_ipv6_ops *v6ops;
+	int ret = 0;
+
+	switch (family) {
+	case AF_INET:
+		ret = nf_ip_route(net, dst, fl, strict);
+		break;
+	case AF_INET6:
+		v6ops = rcu_dereference(nf_ipv6_ops);
+		if (v6ops)
+			ret = v6ops->route(net, dst, fl, strict);
+		break;
+	}
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(nf_route);

net/netfilter/xt_TCPMSS.c

@@ -48,7 +48,6 @@ static u_int32_t tcpmss_reverse_mtu(struct net *net,
 				    unsigned int family)
 {
 	struct flowi fl;
-	const struct nf_afinfo *ai;
 	struct rtable *rt = NULL;
 	u_int32_t mtu     = ~0U;
 
@@ -62,10 +61,8 @@ static u_int32_t tcpmss_reverse_mtu(struct net *net,
 		memset(fl6, 0, sizeof(*fl6));
 		fl6->daddr = ipv6_hdr(skb)->saddr;
 	}
-	ai = nf_get_afinfo(family);
-	if (ai != NULL)
-		ai->route(net, (struct dst_entry **)&rt, &fl, false);
 
+	nf_route(net, (struct dst_entry **)&rt, &fl, false, family);
 	if (rt != NULL) {
 		mtu = dst_mtu(&rt->dst);
 		dst_release(&rt->dst);

net/netfilter/xt_addrtype.c

@@ -36,7 +36,7 @@ MODULE_ALIAS("ip6t_addrtype");
 static u32 match_lookup_rt6(struct net *net, const struct net_device *dev,
 			    const struct in6_addr *addr, u16 mask)
 {
-	const struct nf_afinfo *afinfo;
+	const struct nf_ipv6_ops *v6ops;
 	struct flowi6 flow;
 	struct rt6_info *rt;
 	u32 ret = 0;
@@ -47,17 +47,14 @@ static u32 match_lookup_rt6(struct net *net, const struct net_device *dev,
 	if (dev)
 		flow.flowi6_oif = dev->ifindex;
 
-	afinfo = nf_get_afinfo(NFPROTO_IPV6);
+	v6ops = nf_get_ipv6_ops();
-	if (afinfo != NULL) {
+	if (v6ops) {
-		const struct nf_ipv6_ops *v6ops;
-
 		if (dev && (mask & XT_ADDRTYPE_LOCAL)) {
-			v6ops = nf_get_ipv6_ops();
+			if (v6ops->chk_addr(net, addr, dev, true))
-			if (v6ops && v6ops->chk_addr(net, addr, dev, true))
 				ret = XT_ADDRTYPE_LOCAL;
 		}
-		route_err = afinfo->route(net, (struct dst_entry **)&rt,
+		route_err = v6ops->route(net, (struct dst_entry **)&rt,
-					  flowi6_to_flowi(&flow), false);
+					 flowi6_to_flowi(&flow), false);
 	} else {
 		route_err = 1;
 	}