Merge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf

Pablo Neira Ayuso says:

====================
Netfilter/IPVS fixes for net

The following patchset contains Netfilter/IPVS fixes for your net tree,
they are:

1) Fix handling of simultaneous open TCP connection in conntrack,
   from Jozsef Kadlecsik.

2) Insufficient sanitify check of xtables extension names, from
   Florian Westphal.

3) Skip unnecessary synchronize_rcu() call when transaction log
   is already empty, from Florian Westphal.

4) Incorrect destination mac validation in ebt_stp, from Stephen
   Hemminger.

5) xtables module reference counter leak in nft_compat, from
   Florian Westphal.

6) Incorrect connection reference counting logic in IPVS
   one-packet scheduler, from Julian Anastasov.

7) Wrong stats for 32-bits CPU in IPVS, also from Julian.

8) Calm down sparse error in netfilter core, also from Florian.

9) Use nla_strlcpy to fix compilation warning in nfnetlink_acct
   and nfnetlink_cthelper, again from Florian.

10) Missing module alias in icmp and icmp6 xtables extensions,
    from Florian Westphal.

11) Base chain statistics in nf_tables may be unset/null, from Florian.

12) Fix handling of large matchinfo size in nft_compat, this includes
    one preparation for before this fix. From Florian.

13) Fix bogus EBUSY error when deleting chains due to incorrect reference
    counting from the preparation phase of the two-phase commit protocol.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/netfilter/nf_tables.h

@@ -170,6 +170,7 @@ struct nft_data_desc {
 int nft_data_init(const struct nft_ctx *ctx,
 		  struct nft_data *data, unsigned int size,
 		  struct nft_data_desc *desc, const struct nlattr *nla);
+void nft_data_hold(const struct nft_data *data, enum nft_data_types type);
 void nft_data_release(const struct nft_data *data, enum nft_data_types type);
 int nft_data_dump(struct sk_buff *skb, int attr, const struct nft_data *data,
 		  enum nft_data_types type, unsigned int len);
@@ -736,6 +737,10 @@ struct nft_expr_ops {
 	int				(*init)(const struct nft_ctx *ctx,
 						const struct nft_expr *expr,
 						const struct nlattr * const tb[]);
+	void				(*activate)(const struct nft_ctx *ctx,
+						    const struct nft_expr *expr);
+	void				(*deactivate)(const struct nft_ctx *ctx,
+						      const struct nft_expr *expr);
 	void				(*destroy)(const struct nft_ctx *ctx,
 						   const struct nft_expr *expr);
 	int				(*dump)(struct sk_buff *skb,

include/uapi/linux/netfilter/nf_conntrack_tcp.h

@@ -46,6 +46,9 @@ enum tcp_conntrack {
 /* Marks possibility for expected RFC5961 challenge ACK */
 #define IP_CT_EXP_CHALLENGE_ACK 		0x40
 
+/* Simultaneous open initialized */
+#define IP_CT_TCP_SIMULTANEOUS_OPEN		0x80
+
 struct nf_ct_tcp_flags {
 	__u8 flags;
 	__u8 mask;

net/bridge/netfilter/ebt_stp.c

@@ -161,8 +161,8 @@ static int ebt_stp_mt_check(const struct xt_mtchk_param *par)
 	/* Make sure the match only receives stp frames */
 	if (!par->nft_compat &&
 	    (!ether_addr_equal(e->destmac, eth_stp_addr) ||
-	     !is_broadcast_ether_addr(e->destmsk) ||
-	     !(e->bitmask & EBT_DESTMAC)))
+	     !(e->bitmask & EBT_DESTMAC) ||
+	     !is_broadcast_ether_addr(e->destmsk)))
 		return -EINVAL;
 
 	return 0;

net/ipv4/netfilter/ip_tables.c

@@ -34,6 +34,7 @@
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Netfilter Core Team <coreteam@netfilter.org>");
 MODULE_DESCRIPTION("IPv4 packet filter");
+MODULE_ALIAS("ipt_icmp");
 
 void *ipt_alloc_initial_table(const struct xt_table *info)
 {

net/ipv6/netfilter/ip6_tables.c

@@ -38,6 +38,7 @@
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Netfilter Core Team <coreteam@netfilter.org>");
 MODULE_DESCRIPTION("IPv6 packet filter");
+MODULE_ALIAS("ip6t_icmp6");
 
 void *ip6t_alloc_initial_table(const struct xt_table *info)
 {

net/netfilter/core.c

@@ -585,7 +585,8 @@ void (*nf_nat_decode_session_hook)(struct sk_buff *, struct flowi *);
 EXPORT_SYMBOL(nf_nat_decode_session_hook);
 #endif
 
-static void __net_init __netfilter_net_init(struct nf_hook_entries **e, int max)
+static void __net_init
+__netfilter_net_init(struct nf_hook_entries __rcu **e, int max)
 {
 	int h;
 

net/netfilter/ipvs/ip_vs_conn.c

@@ -232,7 +232,10 @@ static inline int ip_vs_conn_unhash(struct ip_vs_conn *cp)
 static inline bool ip_vs_conn_unlink(struct ip_vs_conn *cp)
 {
 	unsigned int hash;
-	bool ret;
+	bool ret = false;
+
+	if (cp->flags & IP_VS_CONN_F_ONE_PACKET)
+		return refcount_dec_if_one(&cp->refcnt);
 
 	hash = ip_vs_conn_hashkey_conn(cp);
 
@@ -240,15 +243,13 @@ static inline bool ip_vs_conn_unlink(struct ip_vs_conn *cp)
 	spin_lock(&cp->lock);
 
 	if (cp->flags & IP_VS_CONN_F_HASHED) {
-		ret = false;
 		/* Decrease refcnt and unlink conn only if we are last user */
 		if (refcount_dec_if_one(&cp->refcnt)) {
 			hlist_del_rcu(&cp->c_list);
 			cp->flags &= ~IP_VS_CONN_F_HASHED;
 			ret = true;
 		}
-	} else
-		ret = refcount_read(&cp->refcnt) ? false : true;
+	}
 
 	spin_unlock(&cp->lock);
 	ct_write_unlock_bh(hash);
@@ -454,12 +455,6 @@ ip_vs_conn_out_get_proto(struct netns_ipvs *ipvs, int af,
 }
 EXPORT_SYMBOL_GPL(ip_vs_conn_out_get_proto);
 
-static void __ip_vs_conn_put_notimer(struct ip_vs_conn *cp)
-{
-	__ip_vs_conn_put(cp);
-	ip_vs_conn_expire(&cp->timer);
-}
-
 /*
  *      Put back the conn and restart its timer with its timeout
  */
@@ -478,7 +473,7 @@ void ip_vs_conn_put(struct ip_vs_conn *cp)
 	    (refcount_read(&cp->refcnt) == 1) &&
 	    !timer_pending(&cp->timer))
 		/* expire connection immediately */
-		__ip_vs_conn_put_notimer(cp);
+		ip_vs_conn_expire(&cp->timer);
 	else
 		__ip_vs_conn_put_timer(cp);
 }

net/netfilter/ipvs/ip_vs_core.c

@@ -119,6 +119,8 @@ ip_vs_in_stats(struct ip_vs_conn *cp, struct sk_buff *skb)
 		struct ip_vs_cpu_stats *s;
 		struct ip_vs_service *svc;
 
+		local_bh_disable();
+
 		s = this_cpu_ptr(dest->stats.cpustats);
 		u64_stats_update_begin(&s->syncp);
 		s->cnt.inpkts++;
@@ -137,6 +139,8 @@ ip_vs_in_stats(struct ip_vs_conn *cp, struct sk_buff *skb)
 		s->cnt.inpkts++;
 		s->cnt.inbytes += skb->len;
 		u64_stats_update_end(&s->syncp);
+
+		local_bh_enable();
 	}
 }
 
@@ -151,6 +155,8 @@ ip_vs_out_stats(struct ip_vs_conn *cp, struct sk_buff *skb)
 		struct ip_vs_cpu_stats *s;
 		struct ip_vs_service *svc;
 
+		local_bh_disable();
+
 		s = this_cpu_ptr(dest->stats.cpustats);
 		u64_stats_update_begin(&s->syncp);
 		s->cnt.outpkts++;
@@ -169,6 +175,8 @@ ip_vs_out_stats(struct ip_vs_conn *cp, struct sk_buff *skb)
 		s->cnt.outpkts++;
 		s->cnt.outbytes += skb->len;
 		u64_stats_update_end(&s->syncp);
+
+		local_bh_enable();
 	}
 }
 
@@ -179,6 +187,8 @@ ip_vs_conn_stats(struct ip_vs_conn *cp, struct ip_vs_service *svc)
 	struct netns_ipvs *ipvs = svc->ipvs;
 	struct ip_vs_cpu_stats *s;
 
+	local_bh_disable();
+
 	s = this_cpu_ptr(cp->dest->stats.cpustats);
 	u64_stats_update_begin(&s->syncp);
 	s->cnt.conns++;
@@ -193,6 +203,8 @@ ip_vs_conn_stats(struct ip_vs_conn *cp, struct ip_vs_service *svc)
 	u64_stats_update_begin(&s->syncp);
 	s->cnt.conns++;
 	u64_stats_update_end(&s->syncp);
+
+	local_bh_enable();
 }
 
 

net/netfilter/nf_conntrack_proto_tcp.c

@@ -981,6 +981,17 @@ static int tcp_packet(struct nf_conn *ct,
 			return NF_ACCEPT; /* Don't change state */
 		}
 		break;
+	case TCP_CONNTRACK_SYN_SENT2:
+		/* tcp_conntracks table is not smart enough to handle
+		 * simultaneous open.
+		 */
+		ct->proto.tcp.last_flags |= IP_CT_TCP_SIMULTANEOUS_OPEN;
+		break;
+	case TCP_CONNTRACK_SYN_RECV:
+		if (dir == IP_CT_DIR_REPLY && index == TCP_ACK_SET &&
+		    ct->proto.tcp.last_flags & IP_CT_TCP_SIMULTANEOUS_OPEN)
+			new_state = TCP_CONNTRACK_ESTABLISHED;
+		break;
 	case TCP_CONNTRACK_CLOSE:
 		if (index == TCP_RST_SET
 		    && (ct->proto.tcp.seen[!dir].flags & IP_CT_TCP_FLAG_MAXACK_SET)

net/netfilter/nf_tables_api.c

@@ -214,6 +214,34 @@ static int nft_delchain(struct nft_ctx *ctx)
 	return err;
 }
 
+static void nft_rule_expr_activate(const struct nft_ctx *ctx,
+				   struct nft_rule *rule)
+{
+	struct nft_expr *expr;
+
+	expr = nft_expr_first(rule);
+	while (expr != nft_expr_last(rule) && expr->ops) {
+		if (expr->ops->activate)
+			expr->ops->activate(ctx, expr);
+
+		expr = nft_expr_next(expr);
+	}
+}
+
+static void nft_rule_expr_deactivate(const struct nft_ctx *ctx,
+				     struct nft_rule *rule)
+{
+	struct nft_expr *expr;
+
+	expr = nft_expr_first(rule);
+	while (expr != nft_expr_last(rule) && expr->ops) {
+		if (expr->ops->deactivate)
+			expr->ops->deactivate(ctx, expr);
+
+		expr = nft_expr_next(expr);
+	}
+}
+
 static int
 nf_tables_delrule_deactivate(struct nft_ctx *ctx, struct nft_rule *rule)
 {
@@ -259,6 +287,7 @@ static int nft_delrule(struct nft_ctx *ctx, struct nft_rule *rule)
 		nft_trans_destroy(trans);
 		return err;
 	}
+	nft_rule_expr_deactivate(ctx, rule);
 
 	return 0;
 }
@@ -2238,6 +2267,13 @@ static void nf_tables_rule_destroy(const struct nft_ctx *ctx,
 	kfree(rule);
 }
 
+static void nf_tables_rule_release(const struct nft_ctx *ctx,
+				   struct nft_rule *rule)
+{
+	nft_rule_expr_deactivate(ctx, rule);
+	nf_tables_rule_destroy(ctx, rule);
+}
+
 #define NFT_RULE_MAXEXPRS	128
 
 static struct nft_expr_info *info;
@@ -2402,7 +2438,7 @@ static int nf_tables_newrule(struct net *net, struct sock *nlsk,
 	return 0;
 
 err2:
-	nf_tables_rule_destroy(&ctx, rule);
+	nf_tables_rule_release(&ctx, rule);
 err1:
 	for (i = 0; i < n; i++) {
 		if (info[i].ops != NULL)
@@ -4044,8 +4080,10 @@ static int nft_add_set_elem(struct nft_ctx *ctx, struct nft_set *set,
 			if (nft_set_ext_exists(ext, NFT_SET_EXT_DATA) ^
 			    nft_set_ext_exists(ext2, NFT_SET_EXT_DATA) ||
 			    nft_set_ext_exists(ext, NFT_SET_EXT_OBJREF) ^
-			    nft_set_ext_exists(ext2, NFT_SET_EXT_OBJREF))
-				return -EBUSY;
+			    nft_set_ext_exists(ext2, NFT_SET_EXT_OBJREF)) {
+				err = -EBUSY;
+				goto err5;
+			}
 			if ((nft_set_ext_exists(ext, NFT_SET_EXT_DATA) &&
 			     nft_set_ext_exists(ext2, NFT_SET_EXT_DATA) &&
 			     memcmp(nft_set_ext_data(ext),
@@ -4130,7 +4168,7 @@ static int nf_tables_newsetelem(struct net *net, struct sock *nlsk,
  *	NFT_GOTO verdicts. This function must be called on active data objects
  *	from the second phase of the commit protocol.
  */
-static void nft_data_hold(const struct nft_data *data, enum nft_data_types type)
+void nft_data_hold(const struct nft_data *data, enum nft_data_types type)
 {
 	if (type == NFT_DATA_VERDICT) {
 		switch (data->verdict.code) {
@@ -5761,7 +5799,7 @@ static void nft_chain_commit_update(struct nft_trans *trans)
 	}
 }
 
-static void nf_tables_commit_release(struct nft_trans *trans)
+static void nft_commit_release(struct nft_trans *trans)
 {
 	switch (trans->msg_type) {
 	case NFT_MSG_DELTABLE:
@@ -5790,6 +5828,21 @@ static void nf_tables_commit_release(struct nft_trans *trans)
 	kfree(trans);
 }
 
+static void nf_tables_commit_release(struct net *net)
+{
+	struct nft_trans *trans, *next;
+
+	if (list_empty(&net->nft.commit_list))
+		return;
+
+	synchronize_rcu();
+
+	list_for_each_entry_safe(trans, next, &net->nft.commit_list, list) {
+		list_del(&trans->list);
+		nft_commit_release(trans);
+	}
+}
+
 static int nf_tables_commit(struct net *net, struct sk_buff *skb)
 {
 	struct nft_trans *trans, *next;
@@ -5920,13 +5973,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb)
 		}
 	}
 
-	synchronize_rcu();
-
-	list_for_each_entry_safe(trans, next, &net->nft.commit_list, list) {
-		list_del(&trans->list);
-		nf_tables_commit_release(trans);
-	}
-
+	nf_tables_commit_release(net);
 	nf_tables_gen_notify(net, skb, NFT_MSG_NEWGEN);
 
 	return 0;
@@ -6006,10 +6053,12 @@ static int nf_tables_abort(struct net *net, struct sk_buff *skb)
 		case NFT_MSG_NEWRULE:
 			trans->ctx.chain->use--;
 			list_del_rcu(&nft_trans_rule(trans)->list);
+			nft_rule_expr_deactivate(&trans->ctx, nft_trans_rule(trans));
 			break;
 		case NFT_MSG_DELRULE:
 			trans->ctx.chain->use++;
 			nft_clear(trans->ctx.net, nft_trans_rule(trans));
+			nft_rule_expr_activate(&trans->ctx, nft_trans_rule(trans));
 			nft_trans_destroy(trans);
 			break;
 		case NFT_MSG_NEWSET:
@@ -6585,7 +6634,7 @@ int __nft_release_basechain(struct nft_ctx *ctx)
 	list_for_each_entry_safe(rule, nr, &ctx->chain->rules, list) {
 		list_del(&rule->list);
 		ctx->chain->use--;
-		nf_tables_rule_destroy(ctx, rule);
+		nf_tables_rule_release(ctx, rule);
 	}
 	list_del(&ctx->chain->list);
 	ctx->table->use--;
@@ -6623,7 +6672,7 @@ static void __nft_release_tables(struct net *net)
 			list_for_each_entry_safe(rule, nr, &chain->rules, list) {
 				list_del(&rule->list);
 				chain->use--;
-				nf_tables_rule_destroy(&ctx, rule);
+				nf_tables_rule_release(&ctx, rule);
 			}
 		}
 		list_for_each_entry_safe(flowtable, nf, &table->flowtables, list) {

net/netfilter/nf_tables_core.c

@@ -119,15 +119,22 @@ DEFINE_STATIC_KEY_FALSE(nft_counters_enabled);
 static noinline void nft_update_chain_stats(const struct nft_chain *chain,
 					    const struct nft_pktinfo *pkt)
 {
+	struct nft_base_chain *base_chain;
 	struct nft_stats *stats;
 
+	base_chain = nft_base_chain(chain);
+	if (!base_chain->stats)
+		return;
+
+	stats = this_cpu_ptr(rcu_dereference(base_chain->stats));
+	if (stats) {
 		local_bh_disable();
-	stats = this_cpu_ptr(rcu_dereference(nft_base_chain(chain)->stats));
 		u64_stats_update_begin(&stats->syncp);
 		stats->pkts++;
 		stats->bytes += pkt->skb->len;
 		u64_stats_update_end(&stats->syncp);
 		local_bh_enable();
+	}
 }
 
 struct nft_jumpstack {

net/netfilter/nfnetlink_acct.c

@@ -115,7 +115,7 @@ static int nfnl_acct_new(struct net *net, struct sock *nfnl,
 		nfacct->flags = flags;
 	}
 
-	strncpy(nfacct->name, nla_data(tb[NFACCT_NAME]), NFACCT_NAME_MAX);
+	nla_strlcpy(nfacct->name, nla_data(tb[NFACCT_NAME]), NFACCT_NAME_MAX);
 
 	if (tb[NFACCT_BYTES]) {
 		atomic64_set(&nfacct->bytes,

net/netfilter/nfnetlink_cthelper.c

@@ -149,7 +149,7 @@ nfnl_cthelper_expect_policy(struct nf_conntrack_expect_policy *expect_policy,
 	    !tb[NFCTH_POLICY_EXPECT_TIMEOUT])
 		return -EINVAL;
 
-	strncpy(expect_policy->name,
+	nla_strlcpy(expect_policy->name,
 		    nla_data(tb[NFCTH_POLICY_NAME]), NF_CT_HELPER_NAME_LEN);
 	expect_policy->max_expected =
 		ntohl(nla_get_be32(tb[NFCTH_POLICY_EXPECT_MAX]));
@@ -234,7 +234,8 @@ nfnl_cthelper_create(const struct nlattr * const tb[],
 	if (ret < 0)
 		goto err1;
 
-	strncpy(helper->name, nla_data(tb[NFCTH_NAME]), NF_CT_HELPER_NAME_LEN);
+	nla_strlcpy(helper->name,
+		    nla_data(tb[NFCTH_NAME]), NF_CT_HELPER_NAME_LEN);
 	size = ntohl(nla_get_be32(tb[NFCTH_PRIV_DATA_LEN]));
 	if (size > FIELD_SIZEOF(struct nf_conn_help, data)) {
 		ret = -ENOMEM;

net/netfilter/nft_compat.c

@@ -27,14 +27,31 @@ struct nft_xt {
 	struct list_head	head;
 	struct nft_expr_ops	ops;
 	unsigned int		refcnt;
+
+	/* Unlike other expressions, ops doesn't have static storage duration.
+	 * nft core assumes they do.  We use kfree_rcu so that nft core can
+	 * can check expr->ops->size even after nft_compat->destroy() frees
+	 * the nft_xt struct that holds the ops structure.
+	 */
+	struct rcu_head		rcu_head;
 };
 
-static void nft_xt_put(struct nft_xt *xt)
+/* Used for matches where *info is larger than X byte */
+#define NFT_MATCH_LARGE_THRESH	192
+
+struct nft_xt_match_priv {
+	void *info;
+};
+
+static bool nft_xt_put(struct nft_xt *xt)
 {
 	if (--xt->refcnt == 0) {
 		list_del(&xt->head);
-		kfree(xt);
+		kfree_rcu(xt, rcu_head);
+		return true;
 	}
+
+	return false;
 }
 
 static int nft_compat_chain_validate_dependency(const char *tablename,
@@ -226,6 +243,7 @@ nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	struct xt_target *target = expr->ops->data;
 	struct xt_tgchk_param par;
 	size_t size = XT_ALIGN(nla_len(tb[NFTA_TARGET_INFO]));
+	struct nft_xt *nft_xt;
 	u16 proto = 0;
 	bool inv = false;
 	union nft_entry e = {};
@@ -236,25 +254,22 @@ nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	if (ctx->nla[NFTA_RULE_COMPAT]) {
 		ret = nft_parse_compat(ctx->nla[NFTA_RULE_COMPAT], &proto, &inv);
 		if (ret < 0)
-			goto err;
+			return ret;
 	}
 
 	nft_target_set_tgchk_param(&par, ctx, target, info, &e, proto, inv);
 
 	ret = xt_check_target(&par, size, proto, inv);
 	if (ret < 0)
-		goto err;
+		return ret;
 
 	/* The standard target cannot be used */
-	if (target->target == NULL) {
-		ret = -EINVAL;
-		goto err;
-	}
+	if (!target->target)
+		return -EINVAL;
 
+	nft_xt = container_of(expr->ops, struct nft_xt, ops);
+	nft_xt->refcnt++;
 	return 0;
-err:
-	module_put(target->me);
-	return ret;
 }
 
 static void
@@ -271,7 +286,7 @@ nft_target_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
 	if (par.target->destroy != NULL)
 		par.target->destroy(&par);
 
-	nft_xt_put(container_of(expr->ops, struct nft_xt, ops));
+	if (nft_xt_put(container_of(expr->ops, struct nft_xt, ops)))
 		module_put(target->me);
 }
 
@@ -316,11 +331,11 @@ static int nft_target_validate(const struct nft_ctx *ctx,
 	return 0;
 }
 
-static void nft_match_eval(const struct nft_expr *expr,
+static void __nft_match_eval(const struct nft_expr *expr,
 			     struct nft_regs *regs,
-			   const struct nft_pktinfo *pkt)
+			     const struct nft_pktinfo *pkt,
+			     void *info)
 {
-	void *info = nft_expr_priv(expr);
 	struct xt_match *match = expr->ops->data;
 	struct sk_buff *skb = pkt->skb;
 	bool ret;
@@ -344,6 +359,22 @@ static void nft_match_eval(const struct nft_expr *expr,
 	}
 }
 
+static void nft_match_large_eval(const struct nft_expr *expr,
+				 struct nft_regs *regs,
+				 const struct nft_pktinfo *pkt)
+{
+	struct nft_xt_match_priv *priv = nft_expr_priv(expr);
+
+	__nft_match_eval(expr, regs, pkt, priv->info);
+}
+
+static void nft_match_eval(const struct nft_expr *expr,
+			   struct nft_regs *regs,
+			   const struct nft_pktinfo *pkt)
+{
+	__nft_match_eval(expr, regs, pkt, nft_expr_priv(expr));
+}
+
 static const struct nla_policy nft_match_policy[NFTA_MATCH_MAX + 1] = {
 	[NFTA_MATCH_NAME]	= { .type = NLA_NUL_STRING },
 	[NFTA_MATCH_REV]	= { .type = NLA_U32 },
@@ -404,13 +435,14 @@ static void match_compat_from_user(struct xt_match *m, void *in, void *out)
 }
 
 static int
-nft_match_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
-		const struct nlattr * const tb[])
+__nft_match_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
+		 const struct nlattr * const tb[],
+		 void *info)
 {
-	void *info = nft_expr_priv(expr);
 	struct xt_match *match = expr->ops->data;
 	struct xt_mtchk_param par;
 	size_t size = XT_ALIGN(nla_len(tb[NFTA_MATCH_INFO]));
+	struct nft_xt *nft_xt;
 	u16 proto = 0;
 	bool inv = false;
 	union nft_entry e = {};
@@ -421,26 +453,50 @@ nft_match_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
 	if (ctx->nla[NFTA_RULE_COMPAT]) {
 		ret = nft_parse_compat(ctx->nla[NFTA_RULE_COMPAT], &proto, &inv);
 		if (ret < 0)
-			goto err;
+			return ret;
 	}
 
 	nft_match_set_mtchk_param(&par, ctx, match, info, &e, proto, inv);
 
 	ret = xt_check_match(&par, size, proto, inv);
 	if (ret < 0)
-		goto err;
+		return ret;
 
+	nft_xt = container_of(expr->ops, struct nft_xt, ops);
+	nft_xt->refcnt++;
 	return 0;
-err:
-	module_put(match->me);
+}
+
+static int
+nft_match_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
+	       const struct nlattr * const tb[])
+{
+	return __nft_match_init(ctx, expr, tb, nft_expr_priv(expr));
+}
+
+static int
+nft_match_large_init(const struct nft_ctx *ctx, const struct nft_expr *expr,
+		     const struct nlattr * const tb[])
+{
+	struct nft_xt_match_priv *priv = nft_expr_priv(expr);
+	struct xt_match *m = expr->ops->data;
+	int ret;
+
+	priv->info = kmalloc(XT_ALIGN(m->matchsize), GFP_KERNEL);
+	if (!priv->info)
+		return -ENOMEM;
+
+	ret = __nft_match_init(ctx, expr, tb, priv->info);
+	if (ret)
+		kfree(priv->info);
 	return ret;
 }
 
 static void
-nft_match_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
+__nft_match_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr,
+		    void *info)
 {
 	struct xt_match *match = expr->ops->data;
-	void *info = nft_expr_priv(expr);
 	struct xt_mtdtor_param par;
 
 	par.net = ctx->net;
@@ -450,13 +506,28 @@ nft_match_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
 	if (par.match->destroy != NULL)
 		par.match->destroy(&par);
 
-	nft_xt_put(container_of(expr->ops, struct nft_xt, ops));
+	if (nft_xt_put(container_of(expr->ops, struct nft_xt, ops)))
 		module_put(match->me);
 }
 
-static int nft_match_dump(struct sk_buff *skb, const struct nft_expr *expr)
+static void
+nft_match_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
+{
+	__nft_match_destroy(ctx, expr, nft_expr_priv(expr));
+}
+
+static void
+nft_match_large_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr)
+{
+	struct nft_xt_match_priv *priv = nft_expr_priv(expr);
+
+	__nft_match_destroy(ctx, expr, priv->info);
+	kfree(priv->info);
+}
+
+static int __nft_match_dump(struct sk_buff *skb, const struct nft_expr *expr,
+			    void *info)
 {
-	void *info = nft_expr_priv(expr);
 	struct xt_match *match = expr->ops->data;
 
 	if (nla_put_string(skb, NFTA_MATCH_NAME, match->name) ||
@@ -470,6 +541,18 @@ static int nft_match_dump(struct sk_buff *skb, const struct nft_expr *expr)
 	return -1;
 }
 
+static int nft_match_dump(struct sk_buff *skb, const struct nft_expr *expr)
+{
+	return __nft_match_dump(skb, expr, nft_expr_priv(expr));
+}
+
+static int nft_match_large_dump(struct sk_buff *skb, const struct nft_expr *e)
+{
+	struct nft_xt_match_priv *priv = nft_expr_priv(e);
+
+	return __nft_match_dump(skb, e, priv->info);
+}
+
 static int nft_match_validate(const struct nft_ctx *ctx,
 			      const struct nft_expr *expr,
 			      const struct nft_data **data)
@@ -637,6 +720,7 @@ nft_match_select_ops(const struct nft_ctx *ctx,
 {
 	struct nft_xt *nft_match;
 	struct xt_match *match;
+	unsigned int matchsize;
 	char *mt_name;
 	u32 rev, family;
 	int err;
@@ -654,14 +738,9 @@ nft_match_select_ops(const struct nft_ctx *ctx,
 	list_for_each_entry(nft_match, &nft_match_list, head) {
 		struct xt_match *match = nft_match->ops.data;
 
-		if (nft_match_cmp(match, mt_name, rev, family)) {
-			if (!try_module_get(match->me))
-				return ERR_PTR(-ENOENT);
-
-			nft_match->refcnt++;
+		if (nft_match_cmp(match, mt_name, rev, family))
 			return &nft_match->ops;
 	}
-	}
 
 	match = xt_request_find_match(family, mt_name, rev);
 	if (IS_ERR(match))
@@ -679,9 +758,8 @@ nft_match_select_ops(const struct nft_ctx *ctx,
 		goto err;
 	}
 
-	nft_match->refcnt = 1;
+	nft_match->refcnt = 0;
 	nft_match->ops.type = &nft_match_type;
-	nft_match->ops.size = NFT_EXPR_SIZE(XT_ALIGN(match->matchsize));
 	nft_match->ops.eval = nft_match_eval;
 	nft_match->ops.init = nft_match_init;
 	nft_match->ops.destroy = nft_match_destroy;
@@ -689,6 +767,18 @@ nft_match_select_ops(const struct nft_ctx *ctx,
 	nft_match->ops.validate = nft_match_validate;
 	nft_match->ops.data = match;
 
+	matchsize = NFT_EXPR_SIZE(XT_ALIGN(match->matchsize));
+	if (matchsize > NFT_MATCH_LARGE_THRESH) {
+		matchsize = NFT_EXPR_SIZE(sizeof(struct nft_xt_match_priv));
+
+		nft_match->ops.eval = nft_match_large_eval;
+		nft_match->ops.init = nft_match_large_init;
+		nft_match->ops.destroy = nft_match_large_destroy;
+		nft_match->ops.dump = nft_match_large_dump;
+	}
+
+	nft_match->ops.size = matchsize;
+
 	list_add(&nft_match->head, &nft_match_list);
 
 	return &nft_match->ops;
@@ -739,14 +829,9 @@ nft_target_select_ops(const struct nft_ctx *ctx,
 	list_for_each_entry(nft_target, &nft_target_list, head) {
 		struct xt_target *target = nft_target->ops.data;
 
-		if (nft_target_cmp(target, tg_name, rev, family)) {
-			if (!try_module_get(target->me))
-				return ERR_PTR(-ENOENT);
-
-			nft_target->refcnt++;
+		if (nft_target_cmp(target, tg_name, rev, family))
 			return &nft_target->ops;
 	}
-	}
 
 	target = xt_request_find_target(family, tg_name, rev);
 	if (IS_ERR(target))
@@ -764,7 +849,7 @@ nft_target_select_ops(const struct nft_ctx *ctx,
 		goto err;
 	}
 
-	nft_target->refcnt = 1;
+	nft_target->refcnt = 0;
 	nft_target->ops.type = &nft_target_type;
 	nft_target->ops.size = NFT_EXPR_SIZE(XT_ALIGN(target->targetsize));
 	nft_target->ops.init = nft_target_init;
@@ -823,6 +908,32 @@ static int __init nft_compat_module_init(void)
 
 static void __exit nft_compat_module_exit(void)
 {
+	struct nft_xt *xt, *next;
+
+	/* list should be empty here, it can be non-empty only in case there
+	 * was an error that caused nft_xt expr to not be initialized fully
+	 * and noone else requested the same expression later.
+	 *
+	 * In this case, the lists contain 0-refcount entries that still
+	 * hold module reference.
+	 */
+	list_for_each_entry_safe(xt, next, &nft_target_list, head) {
+		struct xt_target *target = xt->ops.data;
+
+		if (WARN_ON_ONCE(xt->refcnt))
+			continue;
+		module_put(target->me);
+		kfree(xt);
+	}
+
+	list_for_each_entry_safe(xt, next, &nft_match_list, head) {
+		struct xt_match *match = xt->ops.data;
+
+		if (WARN_ON_ONCE(xt->refcnt))
+			continue;
+		module_put(match->me);
+		kfree(xt);
+	}
 	nfnetlink_subsys_unregister(&nfnl_compat_subsys);
 	nft_unregister_expr(&nft_target_type);
 	nft_unregister_expr(&nft_match_type);

net/netfilter/nft_immediate.c

@@ -69,7 +69,15 @@ static int nft_immediate_init(const struct nft_ctx *ctx,
 	return err;
 }
 
-static void nft_immediate_destroy(const struct nft_ctx *ctx,
+static void nft_immediate_activate(const struct nft_ctx *ctx,
+				   const struct nft_expr *expr)
+{
+	const struct nft_immediate_expr *priv = nft_expr_priv(expr);
+
+	return nft_data_hold(&priv->data, nft_dreg_to_type(priv->dreg));
+}
+
+static void nft_immediate_deactivate(const struct nft_ctx *ctx,
 				     const struct nft_expr *expr)
 {
 	const struct nft_immediate_expr *priv = nft_expr_priv(expr);
@@ -108,7 +116,8 @@ static const struct nft_expr_ops nft_imm_ops = {
 	.size		= NFT_EXPR_SIZE(sizeof(struct nft_immediate_expr)),
 	.eval		= nft_immediate_eval,
 	.init		= nft_immediate_init,
-	.destroy	= nft_immediate_destroy,
+	.activate	= nft_immediate_activate,
+	.deactivate	= nft_immediate_deactivate,
 	.dump		= nft_immediate_dump,
 	.validate	= nft_immediate_validate,
 };

net/netfilter/x_tables.c

@@ -183,6 +183,9 @@ struct xt_match *xt_find_match(u8 af, const char *name, u8 revision)
 	struct xt_match *m;
 	int err = -ENOENT;
 
+	if (strnlen(name, XT_EXTENSION_MAXNAMELEN) == XT_EXTENSION_MAXNAMELEN)
+		return ERR_PTR(-EINVAL);
+
 	mutex_lock(&xt[af].mutex);
 	list_for_each_entry(m, &xt[af].match, list) {
 		if (strcmp(m->name, name) == 0) {
@@ -229,6 +232,9 @@ struct xt_target *xt_find_target(u8 af, const char *name, u8 revision)
 	struct xt_target *t;
 	int err = -ENOENT;
 
+	if (strnlen(name, XT_EXTENSION_MAXNAMELEN) == XT_EXTENSION_MAXNAMELEN)
+		return ERR_PTR(-EINVAL);
+
 	mutex_lock(&xt[af].mutex);
 	list_for_each_entry(t, &xt[af].target, list) {
 		if (strcmp(t->name, name) == 0) {