Skip to contentSkip to conversation

netfilter fixes for 2.6.30-rc

PATCH Applied Pablo Neira Ayuso 2 participants 5 messages May 5, 2009

Subsystems (2)
lore
Share & tools
lore

Pablo Neira Ayuso

author maintainer
#0 lore raw

cover netfilter fixes for 2.6.30-rc

Hi Patrick,

These are a couple of fixes for 2.6.30-rc, the patch for the `cluster' is a resend (maybe you already have it in the tree). The other applies to ctnetlink, to fix a bug in the user-space reporting when a conntrack entry is updated.

---

Pablo Neira Ayuso (2):
netfilter: ctnetlink: fix wrong message type in user updates
netfilter: iptables: fix use of cluster match with 32 nodes

 include/linux/netfilter/xt_cluster.h |    2 +
 net/netfilter/nf_conntrack_netlink.c |   48 ++++++++++++++--------------------
 net/netfilter/xt_cluster.c           |    8 +++++-
 3 files changed, 29 insertions(+), 29 deletions(-)
author maintainer
#1 lore raw

PATCH 1/2 netfilter: iptables: fix use of cluster match with 32 nodes

This patch fixes a problem when you use 32 nodes in the cluster match:

% iptables -I PREROUTING -t mangle -i eth0 -m cluster \
--cluster-total-nodes 32 --cluster-local-node 32 \
--cluster-hash-seed 0xdeadbeef -j MARK --set-mark 0xffff
iptables: Invalid argument. Run `dmesg' for more information. % dmesg | tail -1
xt_cluster: this node mask cannot be higher than the total number of nodes

The problem is related to this checking:

if (info->node_mask >= (1 << info->total_nodes)) {
	printk(KERN_ERR "xt_cluster: this node mask cannot be "
			"higher than the total number of nodes\n");
	return false;
}

(1 << 32) is 1. Thus, the checking fails.

BTW, I said this before but I insist: I have only tested the cluster match with 2 nodes getting ~45% extra performance in an active-active setup. The maximum limit of 32 nodes is still completely arbitrary. I'd really appreciate if people that have more nodes in their setups let me know.

---

 include/linux/netfilter/xt_cluster.h |    2 ++
 net/netfilter/xt_cluster.c           |    8 +++++++-
 2 files changed, 9 insertions(+), 1 deletions(-)
include/linux/netfilter/xt_cluster.h +2 -0
index 5e0a0d0..8866826 100644
@@ -12,4 +12,6 @@ struct xt_cluster_match_info {
 	u_int32_t		flags;
 };
 
+#define XT_CLUSTER_NODES_MAX	32
+
 #endif /* _XT_CLUSTER_MATCH_H */
net/netfilter/xt_cluster.c +7 -1
index 6c48476..69a639f 100644
@@ -135,7 +135,13 @@ static bool xt_cluster_mt_checkentry(const struct xt_mtchk_param *par)
 {
 	struct xt_cluster_match_info *info = par->matchinfo;
 
-	if (info->node_mask >= (1 << info->total_nodes)) {
+	if (info->total_nodes > XT_CLUSTER_NODES_MAX) {
+		printk(KERN_ERR "xt_cluster: you have exceeded the maximum "
+				"number of cluster nodes (%u > %u)\n",
+				info->total_nodes, XT_CLUSTER_NODES_MAX);
+		return false;
+	}
+	if (info->node_mask >= (1ULL << info->total_nodes)) {
 		printk(KERN_ERR "xt_cluster: this node mask cannot be "
 				"higher than the total number of nodes\n");
 		return false;
#2 lore raw

Re: [PATCH 1/2] netfilter: iptables: fix use of cluster match with 32 nodes

Quoted 18 lines · Pablo Neira Ayuso

This patch fixes a problem when you use 32 nodes in the cluster match:

% iptables -I PREROUTING -t mangle -i eth0 -m cluster \
--cluster-total-nodes 32 --cluster-local-node 32 \
--cluster-hash-seed 0xdeadbeef -j MARK --set-mark 0xffff
iptables: Invalid argument. Run `dmesg' for more information. % dmesg | tail -1
xt_cluster: this node mask cannot be higher than the total number of nodes

The problem is related to this checking:

if (info->node_mask >= (1 << info->total_nodes)) {
	printk(KERN_ERR "xt_cluster: this node mask cannot be "
			"higher than the total number of nodes\n");
	return false;
}
Quoted 6 lines

(1 << 32) is 1. Thus, the checking fails.

BTW, I said this before but I insist: I have only tested the cluster match with 2 nodes getting ~45% extra performance in an active-active setup. The maximum limit of 32 nodes is still completely arbitrary. I'd really appreciate if people that have more nodes in their setups let me know.

Applied, thanks Pablo.

author maintainer
#3 lore raw

PATCH 2/2 netfilter: ctnetlink: fix wrong message type in user updates

In reply to Pablo Neira Ayuso

This patch fixes the wrong message type that are triggered by user updates, the following commands:

(term1)# conntrack -I -p tcp -s 1.1.1.1 -d 2.2.2.2 -t 10 --sport 10 --dport 20 --state LISTEN (term1)# conntrack -U -p tcp -s 1.1.1.1 -d 2.2.2.2 -t 10 --sport 10 --dport 20 --state SYN_SENT (term1)# conntrack -U -p tcp -s 1.1.1.1 -d 2.2.2.2 -t 10 --sport 10 --dport 20 --state SYN_RECV

only trigger event message of type NEW, when only the first is NEW while others should be UPDATE.

(term2)# conntrack -E

    [NEW] tcp      6 10 LISTEN src=1.1.1.1 dst=2.2.2.2 sport=10 dport=20 [UNREPLIED] src=2.2.2.2 dst=1.1.1.1 sport=20 dport=10 mark=0
    [NEW] tcp      6 10 SYN_SENT src=1.1.1.1 dst=2.2.2.2 sport=10 dport=20 [UNREPLIED] src=2.2.2.2 dst=1.1.1.1 sport=20 dport=10 mark=0
    [NEW] tcp      6 10 SYN_RECV src=1.1.1.1 dst=2.2.2.2 sport=10 dport=20 [UNREPLIED] src=2.2.2.2 dst=1.1.1.1 sport=20 dport=10 mark=0

This patch also removes IPCT_REFRESH from the bitmask since it is not of any use.

---

 net/netfilter/nf_conntrack_netlink.c |   48 ++++++++++++++--------------------
 1 files changed, 20 insertions(+), 28 deletions(-)
net/netfilter/nf_conntrack_netlink.c +20 -28
index f13fc57..c523f0b 100644
@@ -1186,28 +1186,6 @@ ctnetlink_change_conntrack(struct nf_conn *ct, struct nlattr *cda[])
 	return 0;
 }
 
-static inline void
-ctnetlink_event_report(struct nf_conn *ct, u32 pid, int report)
-{
-	unsigned int events = 0;
-
-	if (test_bit(IPS_EXPECTED_BIT, &ct->status))
-		events |= IPCT_RELATED;
-	else
-		events |= IPCT_NEW;
-
-	nf_conntrack_event_report(IPCT_STATUS |
-				  IPCT_HELPER |
-				  IPCT_REFRESH |
-				  IPCT_PROTOINFO |
-				  IPCT_NATSEQADJ |
-				  IPCT_MARK |
-				  events,
-				  ct,
-				  pid,
-				  report);
-}
-
 static struct nf_conn *
 ctnetlink_create_conntrack(struct nlattr *cda[],
 			   struct nf_conntrack_tuple *otuple,
@@ -1373,6 +1351,7 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb,
 		err = -ENOENT;
 		if (nlh->nlmsg_flags & NLM_F_CREATE) {
 			struct nf_conn *ct;
+			enum ip_conntrack_events events;
 
 			ct = ctnetlink_create_conntrack(cda, &otuple,
 							&rtuple, u3);
@@ -1383,9 +1362,18 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb,
 			err = 0;
 			nf_conntrack_get(&ct->ct_general);
 			spin_unlock_bh(&nf_conntrack_lock);
-			ctnetlink_event_report(ct,
-					       NETLINK_CB(skb).pid,
-					       nlmsg_report(nlh));
+			if (test_bit(IPS_EXPECTED_BIT, &ct->status))
+				events = IPCT_RELATED;
+			else
+				events = IPCT_NEW;
+
+			nf_conntrack_event_report(IPCT_STATUS |
+						  IPCT_HELPER |
+						  IPCT_PROTOINFO |
+						  IPCT_NATSEQADJ |
+						  IPCT_MARK | events,
+						  ct, NETLINK_CB(skb).pid,
+						  nlmsg_report(nlh));
 			nf_ct_put(ct);
 		} else
 			spin_unlock_bh(&nf_conntrack_lock);
@@ -1404,9 +1392,13 @@ ctnetlink_new_conntrack(struct sock *ctnl, struct sk_buff *skb,
 		if (err == 0) {
 			nf_conntrack_get(&ct->ct_general);
 			spin_unlock_bh(&nf_conntrack_lock);
-			ctnetlink_event_report(ct,
-					       NETLINK_CB(skb).pid,
-					       nlmsg_report(nlh));
+			nf_conntrack_event_report(IPCT_STATUS |
+						  IPCT_HELPER |
+						  IPCT_PROTOINFO |
+						  IPCT_NATSEQADJ |
+						  IPCT_MARK,
+						  ct, NETLINK_CB(skb).pid,
+						  nlmsg_report(nlh));
 			nf_ct_put(ct);
 		} else
 			spin_unlock_bh(&nf_conntrack_lock);
#4 lore raw

Re: [PATCH 2/2] netfilter: ctnetlink: fix wrong message type in user updates

Quoted 10 lines · Pablo Neira Ayuso

This patch fixes the wrong message type that are triggered by user updates, the following commands:

(term1)# conntrack -I -p tcp -s 1.1.1.1 -d 2.2.2.2 -t 10 --sport 10 --dport 20 --state LISTEN (term1)# conntrack -U -p tcp -s 1.1.1.1 -d 2.2.2.2 -t 10 --sport 10 --dport 20 --state SYN_SENT (term1)# conntrack -U -p tcp -s 1.1.1.1 -d 2.2.2.2 -t 10 --sport 10 --dport 20 --state SYN_RECV

only trigger event message of type NEW, when only the first is NEW while others should be UPDATE.

Quoted 7 lines

(term2)# conntrack -E

    [NEW] tcp      6 10 LISTEN src=1.1.1.1 dst=2.2.2.2 sport=10 dport=20 [UNREPLIED] src=2.2.2.2 dst=1.1.1.1 sport=20 dport=10 mark=0
    [NEW] tcp      6 10 SYN_SENT src=1.1.1.1 dst=2.2.2.2 sport=10 dport=20 [UNREPLIED] src=2.2.2.2 dst=1.1.1.1 sport=20 dport=10 mark=0
    [NEW] tcp      6 10 SYN_RECV src=1.1.1.1 dst=2.2.2.2 sport=10 dport=20 [UNREPLIED] src=2.2.2.2 dst=1.1.1.1 sport=20 dport=10 mark=0

This patch also removes IPCT_REFRESH from the bitmask since it is not of any use.

Also applied, thanks.