Merge tag 'nf-23-10-18' of https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf

Florian Westphal says:

====================
netfilter: updates for net

First patch, from Phil Sutter, reduces number of audit notifications
when userspace requests to re-set stateful objects.
This change also comes with a selftest update.

Second patch, also from Phil, moves the nftables audit selftest
to its own netns to avoid interference with the init netns.

Third patch, from Pablo Neira, fixes an inconsistency with the "rbtree"
set backend: When set element X has expired, a request to delete element
X should fail (like with all other backends).

Finally, patch four, also from Pablo, reverts a recent attempt to speed
up abort of a large pending update with the "pipapo" set backend.

It could cause stray references to remain in the set, which then
results in a double-free.

* tag 'nf-23-10-18' of https://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nf_tables: revert do not remove elements if set backend implements .abort
  netfilter: nft_set_rbtree: .deactivate fails if element has expired
  selftests: netfilter: Run nft_audit.sh in its own netns
  netfilter: nf_tables: audit log object reset once per table
====================

Link: https://lore.kernel.org/r/20231018125605.27299-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Author: kuba-moo

net/netfilter/nf_tables_api.c

@@ -7607,6 +7607,16 @@ static int nf_tables_fill_obj_info(struct sk_buff *skb, struct net *net,
 	return -1;
 }
 
+static void audit_log_obj_reset(const struct nft_table *table,
+				unsigned int base_seq, unsigned int nentries)
+{
+	char *buf = kasprintf(GFP_ATOMIC, "%s:%u", table->name, base_seq);
+
+	audit_log_nfcfg(buf, table->family, nentries,
+			AUDIT_NFT_OP_OBJ_RESET, GFP_ATOMIC);
+	kfree(buf);
+}
+
 struct nft_obj_filter {
 	char		*table;
 	u32		type;
@@ -7621,8 +7631,10 @@ static int nf_tables_dump_obj(struct sk_buff *skb, struct netlink_callback *cb)
 	struct net *net = sock_net(skb->sk);
 	int family = nfmsg->nfgen_family;
 	struct nftables_pernet *nft_net;
+	unsigned int entries = 0;
 	struct nft_object *obj;
 	bool reset = false;
+	int rc = 0;
 
 	if (NFNL_MSG_TYPE(cb->nlh->nlmsg_type) == NFT_MSG_GETOBJ_RESET)
 		reset = true;
@@ -7635,6 +7647,7 @@ static int nf_tables_dump_obj(struct sk_buff *skb, struct netlink_callback *cb)
 		if (family != NFPROTO_UNSPEC && family != table->family)
 			continue;
 
+		entries = 0;
 		list_for_each_entry_rcu(obj, &table->objects, list) {
 			if (!nft_is_active(net, obj))
 				goto cont;
@@ -7650,34 +7663,27 @@ static int nf_tables_dump_obj(struct sk_buff *skb, struct netlink_callback *cb)
 			    filter->type != NFT_OBJECT_UNSPEC &&
 			    obj->ops->type->type != filter->type)
 				goto cont;
-			if (reset) {
-				char *buf = kasprintf(GFP_ATOMIC,
-						      "%s:%u",
-						      table->name,
-						      nft_net->base_seq);
-
-				audit_log_nfcfg(buf,
-						family,
-						obj->handle,
-						AUDIT_NFT_OP_OBJ_RESET,
-						GFP_ATOMIC);
-				kfree(buf);
-			}
 
-			if (nf_tables_fill_obj_info(skb, net, NETLINK_CB(cb->skb).portid,
-						    cb->nlh->nlmsg_seq,
-						    NFT_MSG_NEWOBJ,
-						    NLM_F_MULTI | NLM_F_APPEND,
-						    table->family, table,
-						    obj, reset) < 0)
-				goto done;
+			rc = nf_tables_fill_obj_info(skb, net,
+						     NETLINK_CB(cb->skb).portid,
+						     cb->nlh->nlmsg_seq,
+						     NFT_MSG_NEWOBJ,
+						     NLM_F_MULTI | NLM_F_APPEND,
+						     table->family, table,
+						     obj, reset);
+			if (rc < 0)
+				break;
 
+			entries++;
 			nl_dump_check_consistent(cb, nlmsg_hdr(skb));
 cont:
 			idx++;
 		}
+		if (reset && entries)
+			audit_log_obj_reset(table, nft_net->base_seq, entries);
+		if (rc < 0)
+			break;
 	}
-done:
 	rcu_read_unlock();
 
 	cb->args[0] = idx;
@@ -7782,7 +7788,7 @@ static int nf_tables_getobj(struct sk_buff *skb, const struct nfnl_info *info,
 
 		audit_log_nfcfg(buf,
 				family,
-				obj->handle,
+				1,
 				AUDIT_NFT_OP_OBJ_RESET,
 				GFP_ATOMIC);
 		kfree(buf);
@@ -10339,10 +10345,7 @@ static int __nf_tables_abort(struct net *net, enum nfnl_abort_action action)
 				break;
 			}
 			te = (struct nft_trans_elem *)trans->data;
-			if (!te->set->ops->abort ||
-			    nft_setelem_is_catchall(te->set, &te->elem))
-				nft_setelem_remove(net, te->set, &te->elem);
-
+			nft_setelem_remove(net, te->set, &te->elem);
 			if (!nft_setelem_is_catchall(te->set, &te->elem))
 				atomic_dec(&te->set->nelems);
 

net/netfilter/nft_set_rbtree.c

@@ -568,6 +568,8 @@ static void *nft_rbtree_deactivate(const struct net *net,
 				   nft_rbtree_interval_end(this)) {
 				parent = parent->rb_right;
 				continue;
+			} else if (nft_set_elem_expired(&rbe->ext)) {
+				break;
 			} else if (!nft_set_elem_active(&rbe->ext, genmask)) {
 				parent = parent->rb_left;
 				continue;

tools/testing/selftests/netfilter/nft_audit.sh

@@ -11,6 +11,12 @@ nft --version >/dev/null 2>&1 || {
 	exit $SKIP_RC
 }
 
+# Run everything in a separate network namespace
+[ "${1}" != "run" ] && { unshare -n "${0}" run; exit $?; }
+
+# give other scripts a chance to finish - audit_logread sees all activity
+sleep 1
+
 logfile=$(mktemp)
 rulefile=$(mktemp)
 echo "logging into $logfile"
@@ -93,6 +99,12 @@ do_test 'nft add counter t1 c1' \
 do_test 'nft add counter t2 c1; add counter t2 c2' \
 'table=t2 family=2 entries=2 op=nft_register_obj'
 
+for ((i = 3; i <= 500; i++)); do
+	echo "add counter t2 c$i"
+done >$rulefile
+do_test "nft -f $rulefile" \
+'table=t2 family=2 entries=498 op=nft_register_obj'
+
 # adding/updating quotas
 
 do_test 'nft add quota t1 q1 { 10 bytes }' \
@@ -101,6 +113,12 @@ do_test 'nft add quota t1 q1 { 10 bytes }' \
 do_test 'nft add quota t2 q1 { 10 bytes }; add quota t2 q2 { 10 bytes }' \
 'table=t2 family=2 entries=2 op=nft_register_obj'
 
+for ((i = 3; i <= 500; i++)); do
+	echo "add quota t2 q$i { 10 bytes }"
+done >$rulefile
+do_test "nft -f $rulefile" \
+'table=t2 family=2 entries=498 op=nft_register_obj'
+
 # changing the quota value triggers obj update path
 do_test 'nft add quota t1 q1 { 20 bytes }' \
 'table=t1 family=2 entries=1 op=nft_register_obj'
@@ -150,6 +168,40 @@ done
 do_test 'nft reset set t1 s' \
 'table=t1 family=2 entries=3 op=nft_reset_setelem'
 
+# resetting counters
+
+do_test 'nft reset counter t1 c1' \
+'table=t1 family=2 entries=1 op=nft_reset_obj'
+
+do_test 'nft reset counters t1' \
+'table=t1 family=2 entries=1 op=nft_reset_obj'
+
+do_test 'nft reset counters t2' \
+'table=t2 family=2 entries=342 op=nft_reset_obj
+table=t2 family=2 entries=158 op=nft_reset_obj'
+
+do_test 'nft reset counters' \
+'table=t1 family=2 entries=1 op=nft_reset_obj
+table=t2 family=2 entries=341 op=nft_reset_obj
+table=t2 family=2 entries=159 op=nft_reset_obj'
+
+# resetting quotas
+
+do_test 'nft reset quota t1 q1' \
+'table=t1 family=2 entries=1 op=nft_reset_obj'
+
+do_test 'nft reset quotas t1' \
+'table=t1 family=2 entries=1 op=nft_reset_obj'
+
+do_test 'nft reset quotas t2' \
+'table=t2 family=2 entries=315 op=nft_reset_obj
+table=t2 family=2 entries=185 op=nft_reset_obj'
+
+do_test 'nft reset quotas' \
+'table=t1 family=2 entries=1 op=nft_reset_obj
+table=t2 family=2 entries=314 op=nft_reset_obj
+table=t2 family=2 entries=186 op=nft_reset_obj'
+
 # deleting rules
 
 readarray -t handles < <(nft -a list chain t1 c1 | \