Merge tag 'nf-24-04-11' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf

netfilter pull request 24-04-11

Pablo Neira Ayuso says:

====================
Netfilter fixes for net

The following patchset contains Netfilter fixes for net:

Patches #1 and #2 add missing rcu read side lock when iterating over
expression and object type list which could race with module removal.

Patch #3 prevents promisc packet from visiting the bridge/input hook
	 to amend a recent fix to address conntrack confirmation race
	 in br_netfilter and nf_conntrack_bridge.

Patch #4 adds and uses iterate decorator type to fetch the current
	 pipapo set backend datastructure view when netlink dumps the
	 set elements.

Patch #5 fixes removal of duplicate elements in the pipapo set backend.

Patch #6 flowtable validates pppoe header before accessing it.

Patch #7 fixes flowtable datapath for pppoe packets, otherwise lookup
         fails and pppoe packets follow classic path.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/netfilter/nf_flow_table.h

@@ -336,7 +336,7 @@ int nf_flow_rule_route_ipv6(struct net *net, struct flow_offload *flow,
 int nf_flow_table_offload_init(void);
 void nf_flow_table_offload_exit(void);
 
-static inline __be16 nf_flow_pppoe_proto(const struct sk_buff *skb)
+static inline __be16 __nf_flow_pppoe_proto(const struct sk_buff *skb)
 {
 	__be16 proto;
 
@@ -352,6 +352,16 @@ static inline __be16 nf_flow_pppoe_proto(const struct sk_buff *skb)
 	return 0;
 }
 
+static inline bool nf_flow_pppoe_proto(struct sk_buff *skb, __be16 *inner_proto)
+{
+	if (!pskb_may_pull(skb, PPPOE_SES_HLEN))
+		return false;
+
+	*inner_proto = __nf_flow_pppoe_proto(skb);
+
+	return true;
+}
+
 #define NF_FLOW_TABLE_STAT_INC(net, count) __this_cpu_inc((net)->ft.stat->count)
 #define NF_FLOW_TABLE_STAT_DEC(net, count) __this_cpu_dec((net)->ft.stat->count)
 #define NF_FLOW_TABLE_STAT_INC_ATOMIC(net, count)	\

include/net/netfilter/nf_tables.h

@@ -307,9 +307,23 @@ static inline void *nft_elem_priv_cast(const struct nft_elem_priv *priv)
 	return (void *)priv;
 }
 
+
+/**
+ * enum nft_iter_type - nftables set iterator type
+ *
+ * @NFT_ITER_READ: read-only iteration over set elements
+ * @NFT_ITER_UPDATE: iteration under mutex to update set element state
+ */
+enum nft_iter_type {
+	NFT_ITER_UNSPEC,
+	NFT_ITER_READ,
+	NFT_ITER_UPDATE,
+};
+
 struct nft_set;
 struct nft_set_iter {
 	u8		genmask;
+	enum nft_iter_type type:8;
 	unsigned int	count;
 	unsigned int	skip;
 	int		err;

net/bridge/br_input.c

@@ -30,7 +30,7 @@ br_netif_receive_skb(struct net *net, struct sock *sk, struct sk_buff *skb)
 	return netif_receive_skb(skb);
 }
 
-static int br_pass_frame_up(struct sk_buff *skb)
+static int br_pass_frame_up(struct sk_buff *skb, bool promisc)
 {
 	struct net_device *indev, *brdev = BR_INPUT_SKB_CB(skb)->brdev;
 	struct net_bridge *br = netdev_priv(brdev);
@@ -65,6 +65,8 @@ static int br_pass_frame_up(struct sk_buff *skb)
 	br_multicast_count(br, NULL, skb, br_multicast_igmp_type(skb),
 			   BR_MCAST_DIR_TX);
 
+	BR_INPUT_SKB_CB(skb)->promisc = promisc;
+
 	return NF_HOOK(NFPROTO_BRIDGE, NF_BR_LOCAL_IN,
 		       dev_net(indev), NULL, skb, indev, NULL,
 		       br_netif_receive_skb);
@@ -82,6 +84,7 @@ int br_handle_frame_finish(struct net *net, struct sock *sk, struct sk_buff *skb
 	struct net_bridge_mcast *brmctx;
 	struct net_bridge_vlan *vlan;
 	struct net_bridge *br;
+	bool promisc;
 	u16 vid = 0;
 	u8 state;
 
@@ -137,7 +140,9 @@ int br_handle_frame_finish(struct net *net, struct sock *sk, struct sk_buff *skb
 	if (p->flags & BR_LEARNING)
 		br_fdb_update(br, p, eth_hdr(skb)->h_source, vid, 0);
 
-	local_rcv = !!(br->dev->flags & IFF_PROMISC);
+	promisc = !!(br->dev->flags & IFF_PROMISC);
+	local_rcv = promisc;
+
 	if (is_multicast_ether_addr(eth_hdr(skb)->h_dest)) {
 		/* by definition the broadcast is also a multicast address */
 		if (is_broadcast_ether_addr(eth_hdr(skb)->h_dest)) {
@@ -200,7 +205,7 @@ int br_handle_frame_finish(struct net *net, struct sock *sk, struct sk_buff *skb
 		unsigned long now = jiffies;
 
 		if (test_bit(BR_FDB_LOCAL, &dst->flags))
-			return br_pass_frame_up(skb);
+			return br_pass_frame_up(skb, false);
 
 		if (now != dst->used)
 			dst->used = now;
@@ -213,7 +218,7 @@ int br_handle_frame_finish(struct net *net, struct sock *sk, struct sk_buff *skb
 	}
 
 	if (local_rcv)
-		return br_pass_frame_up(skb);
+		return br_pass_frame_up(skb, promisc);
 
 out:
 	return 0;
@@ -386,6 +391,8 @@ static rx_handler_result_t br_handle_frame(struct sk_buff **pskb)
 				goto forward;
 		}
 
+		BR_INPUT_SKB_CB(skb)->promisc = false;
+
 		/* The else clause should be hit when nf_hook():
 		 *   - returns < 0 (drop/error)
 		 *   - returns = 0 (stolen/nf_queue)

net/bridge/br_netfilter_hooks.c

@@ -600,11 +600,17 @@ static unsigned int br_nf_local_in(void *priv,
 				   struct sk_buff *skb,
 				   const struct nf_hook_state *state)
 {
+	bool promisc = BR_INPUT_SKB_CB(skb)->promisc;
 	struct nf_conntrack *nfct = skb_nfct(skb);
 	const struct nf_ct_hook *ct_hook;
 	struct nf_conn *ct;
 	int ret;
 
+	if (promisc) {
+		nf_reset_ct(skb);
+		return NF_ACCEPT;
+	}
+
 	if (!nfct || skb->pkt_type == PACKET_HOST)
 		return NF_ACCEPT;
 

net/bridge/br_private.h

@@ -589,6 +589,7 @@ struct br_input_skb_cb {
 #endif
 	u8 proxyarp_replied:1;
 	u8 src_port_isolated:1;
+	u8 promisc:1;
 #ifdef CONFIG_BRIDGE_VLAN_FILTERING
 	u8 vlan_filtered:1;
 #endif

net/bridge/netfilter/nf_conntrack_bridge.c

@@ -294,18 +294,24 @@ static unsigned int nf_ct_bridge_pre(void *priv, struct sk_buff *skb,
 static unsigned int nf_ct_bridge_in(void *priv, struct sk_buff *skb,
 				    const struct nf_hook_state *state)
 {
-	enum ip_conntrack_info ctinfo;
+	bool promisc = BR_INPUT_SKB_CB(skb)->promisc;
+	struct nf_conntrack *nfct = skb_nfct(skb);
 	struct nf_conn *ct;
 
-	if (skb->pkt_type == PACKET_HOST)
+	if (promisc) {
+		nf_reset_ct(skb);
+		return NF_ACCEPT;
+	}
+
+	if (!nfct || skb->pkt_type == PACKET_HOST)
 		return NF_ACCEPT;
 
 	/* nf_conntrack_confirm() cannot handle concurrent clones,
 	 * this happens for broad/multicast frames with e.g. macvlan on top
 	 * of the bridge device.
 	 */
-	ct = nf_ct_get(skb, &ctinfo);
-	if (!ct || nf_ct_is_confirmed(ct) || nf_ct_is_template(ct))
+	ct = container_of(nfct, struct nf_conn, ct_general);
+	if (nf_ct_is_confirmed(ct) || nf_ct_is_template(ct))
 		return NF_ACCEPT;
 
 	/* let inet prerouting call conntrack again */

net/netfilter/nf_flow_table_inet.c

@@ -21,7 +21,8 @@ nf_flow_offload_inet_hook(void *priv, struct sk_buff *skb,
 		proto = veth->h_vlan_encapsulated_proto;
 		break;
 	case htons(ETH_P_PPP_SES):
-		proto = nf_flow_pppoe_proto(skb);
+		if (!nf_flow_pppoe_proto(skb, &proto))
+			return NF_ACCEPT;
 		break;
 	default:
 		proto = skb->protocol;

net/netfilter/nf_flow_table_ip.c

@@ -157,7 +157,7 @@ static void nf_flow_tuple_encap(struct sk_buff *skb,
 		tuple->encap[i].proto = skb->protocol;
 		break;
 	case htons(ETH_P_PPP_SES):
-		phdr = (struct pppoe_hdr *)skb_mac_header(skb);
+		phdr = (struct pppoe_hdr *)skb_network_header(skb);
 		tuple->encap[i].id = ntohs(phdr->sid);
 		tuple->encap[i].proto = skb->protocol;
 		break;
@@ -273,10 +273,11 @@ static unsigned int nf_flow_xmit_xfrm(struct sk_buff *skb,
 	return NF_STOLEN;
 }
 
-static bool nf_flow_skb_encap_protocol(const struct sk_buff *skb, __be16 proto,
+static bool nf_flow_skb_encap_protocol(struct sk_buff *skb, __be16 proto,
 				       u32 *offset)
 {
 	struct vlan_ethhdr *veth;
+	__be16 inner_proto;
 
 	switch (skb->protocol) {
 	case htons(ETH_P_8021Q):
@@ -287,7 +288,8 @@ static bool nf_flow_skb_encap_protocol(const struct sk_buff *skb, __be16 proto,
 		}
 		break;
 	case htons(ETH_P_PPP_SES):
-		if (nf_flow_pppoe_proto(skb) == proto) {
+		if (nf_flow_pppoe_proto(skb, &inner_proto) &&
+		    inner_proto == proto) {
 			*offset += PPPOE_SES_HLEN;
 			return true;
 		}
@@ -316,7 +318,7 @@ static void nf_flow_encap_pop(struct sk_buff *skb,
 			skb_reset_network_header(skb);
 			break;
 		case htons(ETH_P_PPP_SES):
-			skb->protocol = nf_flow_pppoe_proto(skb);
+			skb->protocol = __nf_flow_pppoe_proto(skb);
 			skb_pull(skb, PPPOE_SES_HLEN);
 			skb_reset_network_header(skb);
 			break;

net/netfilter/nf_tables_api.c

@@ -626,6 +626,7 @@ static void nft_map_deactivate(const struct nft_ctx *ctx, struct nft_set *set)
 {
 	struct nft_set_iter iter = {
 		.genmask	= nft_genmask_next(ctx->net),
+		.type		= NFT_ITER_UPDATE,
 		.fn		= nft_mapelem_deactivate,
 	};
 
@@ -3060,7 +3061,7 @@ static const struct nft_expr_type *__nft_expr_type_get(u8 family,
 {
 	const struct nft_expr_type *type, *candidate = NULL;
 
-	list_for_each_entry(type, &nf_tables_expressions, list) {
+	list_for_each_entry_rcu(type, &nf_tables_expressions, list) {
 		if (!nla_strcmp(nla, type->name)) {
 			if (!type->family && !candidate)
 				candidate = type;
@@ -3092,9 +3093,13 @@ static const struct nft_expr_type *nft_expr_type_get(struct net *net,
 	if (nla == NULL)
 		return ERR_PTR(-EINVAL);
 
+	rcu_read_lock();
 	type = __nft_expr_type_get(family, nla);
-	if (type != NULL && try_module_get(type->owner))
+	if (type != NULL && try_module_get(type->owner)) {
+		rcu_read_unlock();
 		return type;
+	}
+	rcu_read_unlock();
 
 	lockdep_nfnl_nft_mutex_not_held();
 #ifdef CONFIG_MODULES
@@ -5441,6 +5446,7 @@ int nf_tables_bind_set(const struct nft_ctx *ctx, struct nft_set *set,
 		}
 
 		iter.genmask	= nft_genmask_next(ctx->net);
+		iter.type	= NFT_ITER_UPDATE;
 		iter.skip 	= 0;
 		iter.count	= 0;
 		iter.err	= 0;
@@ -5514,6 +5520,7 @@ static void nft_map_activate(const struct nft_ctx *ctx, struct nft_set *set)
 {
 	struct nft_set_iter iter = {
 		.genmask	= nft_genmask_next(ctx->net),
+		.type		= NFT_ITER_UPDATE,
 		.fn		= nft_mapelem_activate,
 	};
 
@@ -5888,6 +5895,7 @@ static int nf_tables_dump_set(struct sk_buff *skb, struct netlink_callback *cb)
 	args.skb		= skb;
 	args.reset		= dump_ctx->reset;
 	args.iter.genmask	= nft_genmask_cur(net);
+	args.iter.type		= NFT_ITER_READ;
 	args.iter.skip		= cb->args[0];
 	args.iter.count		= 0;
 	args.iter.err		= 0;
@@ -7372,6 +7380,7 @@ static int nft_set_flush(struct nft_ctx *ctx, struct nft_set *set, u8 genmask)
 {
 	struct nft_set_iter iter = {
 		.genmask	= genmask,
+		.type		= NFT_ITER_UPDATE,
 		.fn		= nft_setelem_flush,
 	};
 
@@ -7607,7 +7616,7 @@ static const struct nft_object_type *__nft_obj_type_get(u32 objtype, u8 family)
 {
 	const struct nft_object_type *type;
 
-	list_for_each_entry(type, &nf_tables_objects, list) {
+	list_for_each_entry_rcu(type, &nf_tables_objects, list) {
 		if (type->family != NFPROTO_UNSPEC &&
 		    type->family != family)
 			continue;
@@ -7623,9 +7632,13 @@ nft_obj_type_get(struct net *net, u32 objtype, u8 family)
 {
 	const struct nft_object_type *type;
 
+	rcu_read_lock();
 	type = __nft_obj_type_get(objtype, family);
-	if (type != NULL && try_module_get(type->owner))
+	if (type != NULL && try_module_get(type->owner)) {
+		rcu_read_unlock();
 		return type;
+	}
+	rcu_read_unlock();
 
 	lockdep_nfnl_nft_mutex_not_held();
 #ifdef CONFIG_MODULES
@@ -10871,6 +10884,7 @@ static int nf_tables_check_loops(const struct nft_ctx *ctx,
 				continue;
 
 			iter.genmask	= nft_genmask_next(ctx->net);
+			iter.type	= NFT_ITER_UPDATE;
 			iter.skip 	= 0;
 			iter.count	= 0;
 			iter.err	= 0;

net/netfilter/nft_set_pipapo.c

@@ -2077,6 +2077,8 @@ static void nft_pipapo_remove(const struct net *net, const struct nft_set *set,
 		rules_fx = rules_f0;
 
 		nft_pipapo_for_each_field(f, i, m) {
+			bool last = i == m->field_count - 1;
+
 			if (!pipapo_match_field(f, start, rules_fx,
 						match_start, match_end))
 				break;
@@ -2089,16 +2091,18 @@ static void nft_pipapo_remove(const struct net *net, const struct nft_set *set,
 
 			match_start += NFT_PIPAPO_GROUPS_PADDED_SIZE(f);
 			match_end += NFT_PIPAPO_GROUPS_PADDED_SIZE(f);
-		}
 
-		if (i == m->field_count) {
+			if (last && f->mt[rulemap[i].to].e == e) {
 				priv->dirty = true;
 				pipapo_drop(m, rulemap);
 				return;
 			}
+		}
 
 		first_rule += rules_f0;
 	}
+
+	WARN_ON_ONCE(1); /* elem_priv not found */
 }
 
 /**
@@ -2115,13 +2119,14 @@ static void nft_pipapo_walk(const struct nft_ctx *ctx, struct nft_set *set,
 			    struct nft_set_iter *iter)
 {
 	struct nft_pipapo *priv = nft_set_priv(set);
-	struct net *net = read_pnet(&set->net);
 	const struct nft_pipapo_match *m;
 	const struct nft_pipapo_field *f;
 	unsigned int i, r;
 
+	WARN_ON_ONCE(iter->type == NFT_ITER_UNSPEC);
+
 	rcu_read_lock();
-	if (iter->genmask == nft_genmask_cur(net))
+	if (iter->type == NFT_ITER_READ)
 		m = rcu_dereference(priv->match);
 	else
 		m = priv->clone;