Revert "netfilter: move nat hlist_head to nf_conn"

This reverts commit 7c96643 as it is
not working properly.  Please move to 4.9 to get the full fix.

Reported-by: Pablo Neira Ayuso <pablo@netfilter.org>
Cc: Florian Westphal <fw@strlen.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

include/net/netfilter/nf_conntrack.h

@@ -117,9 +117,6 @@ struct nf_conn {
 	/* Extensions */
 	struct nf_ct_ext *ext;
 
-#if IS_ENABLED(CONFIG_NF_NAT)
-	struct hlist_node	nat_bysource;
-#endif
 	/* Storage reserved for other modules, must be the last member */
 	union nf_conntrack_proto proto;
 };

include/net/netfilter/nf_conntrack_extend.h

@@ -99,6 +99,9 @@ void *__nf_ct_ext_add_length(struct nf_conn *ct, enum nf_ct_ext_id id,
 struct nf_ct_ext_type {
 	/* Destroys relationships (can be NULL). */
 	void (*destroy)(struct nf_conn *ct);
+	/* Called when realloacted (can be NULL).
+	   Contents has already been moved. */
+	void (*move)(void *new, void *old);
 
 	enum nf_ct_ext_id id;
 

include/net/netfilter/nf_nat.h

@@ -29,6 +29,8 @@ struct nf_conn;
 
 /* The structure embedded in the conntrack structure. */
 struct nf_conn_nat {
+	struct hlist_node bysource;
+	struct nf_conn *ct;
 	union nf_conntrack_nat_help help;
 #if IS_ENABLED(CONFIG_NF_NAT_MASQUERADE_IPV4) || \
     IS_ENABLED(CONFIG_NF_NAT_MASQUERADE_IPV6)

net/netfilter/nf_conntrack_extend.c

@@ -73,7 +73,7 @@ void *__nf_ct_ext_add_length(struct nf_conn *ct, enum nf_ct_ext_id id,
 			     size_t var_alloc_len, gfp_t gfp)
 {
 	struct nf_ct_ext *old, *new;
-	int newlen, newoff;
+	int i, newlen, newoff;
 	struct nf_ct_ext_type *t;
 
 	/* Conntrack must not be confirmed to avoid races on reallocation. */
@@ -99,8 +99,19 @@ void *__nf_ct_ext_add_length(struct nf_conn *ct, enum nf_ct_ext_id id,
 		return NULL;
 
 	if (new != old) {
+		for (i = 0; i < NF_CT_EXT_NUM; i++) {
+			if (!__nf_ct_ext_exist(old, i))
+				continue;
+
+			rcu_read_lock();
+			t = rcu_dereference(nf_ct_ext_types[i]);
+			if (t && t->move)
+				t->move((void *)new + new->offset[i],
+					(void *)old + old->offset[i]);
+			rcu_read_unlock();
+		}
 		kfree_rcu(old, rcu);
-		rcu_assign_pointer(ct->ext, new);
+		ct->ext = new;
 	}
 
 	new->offset[id] = newoff;

net/netfilter/nf_nat_core.c

@@ -198,9 +198,11 @@ find_appropriate_src(struct net *net,
 		     const struct nf_nat_range *range)
 {
 	unsigned int h = hash_by_src(net, tuple);
+	const struct nf_conn_nat *nat;
 	const struct nf_conn *ct;
 
-	hlist_for_each_entry_rcu(ct, &nf_nat_bysource[h], nat_bysource) {
+	hlist_for_each_entry_rcu(nat, &nf_nat_bysource[h], bysource) {
+		ct = nat->ct;
 		if (same_src(ct, tuple) &&
 		    net_eq(net, nf_ct_net(ct)) &&
 		    nf_ct_zone_equal(ct, zone, IP_CT_DIR_ORIGINAL)) {
@@ -434,7 +436,8 @@ nf_nat_setup_info(struct nf_conn *ct,
 		spin_lock_bh(&nf_nat_lock);
 		/* nf_conntrack_alter_reply might re-allocate extension aera */
 		nat = nfct_nat(ct);
-		hlist_add_head_rcu(&ct->nat_bysource,
+		nat->ct = ct;
+		hlist_add_head_rcu(&nat->bysource,
 				   &nf_nat_bysource[srchash]);
 		spin_unlock_bh(&nf_nat_lock);
 	}
@@ -541,7 +544,7 @@ static int nf_nat_proto_clean(struct nf_conn *ct, void *data)
 	if (nf_nat_proto_remove(ct, data))
 		return 1;
 
-	if (!nat)
+	if (!nat || !nat->ct)
 		return 0;
 
 	/* This netns is being destroyed, and conntrack has nat null binding.
@@ -554,8 +557,9 @@ static int nf_nat_proto_clean(struct nf_conn *ct, void *data)
 		return 1;
 
 	spin_lock_bh(&nf_nat_lock);
-	hlist_del_rcu(&ct->nat_bysource);
+	hlist_del_rcu(&nat->bysource);
 	ct->status &= ~IPS_NAT_DONE_MASK;
+	nat->ct = NULL;
 	spin_unlock_bh(&nf_nat_lock);
 
 	add_timer(&ct->timeout);
@@ -685,20 +689,35 @@ static void nf_nat_cleanup_conntrack(struct nf_conn *ct)
 {
 	struct nf_conn_nat *nat = nf_ct_ext_find(ct, NF_CT_EXT_NAT);
 
-	if (!nat)
+	if (nat == NULL || nat->ct == NULL)
 		return;
 
-	NF_CT_ASSERT(ct->status & IPS_SRC_NAT_DONE);
+	NF_CT_ASSERT(nat->ct->status & IPS_SRC_NAT_DONE);
+
+	spin_lock_bh(&nf_nat_lock);
+	hlist_del_rcu(&nat->bysource);
+	spin_unlock_bh(&nf_nat_lock);
+}
+
+static void nf_nat_move_storage(void *new, void *old)
+{
+	struct nf_conn_nat *new_nat = new;
+	struct nf_conn_nat *old_nat = old;
+	struct nf_conn *ct = old_nat->ct;
+
+	if (!ct || !(ct->status & IPS_SRC_NAT_DONE))
+		return;
 
 	spin_lock_bh(&nf_nat_lock);
-	hlist_del_rcu(&ct->nat_bysource);
+	hlist_replace_rcu(&old_nat->bysource, &new_nat->bysource);
 	spin_unlock_bh(&nf_nat_lock);
 }
 
 static struct nf_ct_ext_type nat_extend __read_mostly = {
 	.len		= sizeof(struct nf_conn_nat),
 	.align		= __alignof__(struct nf_conn_nat),
 	.destroy	= nf_nat_cleanup_conntrack,
+	.move		= nf_nat_move_storage,
 	.id		= NF_CT_EXT_NAT,
 	.flags		= NF_CT_EXT_F_PREALLOC,
 };