Skip to contentSkip to conversation

net: fix duplicate logs of iptables TRACE target

PATCH In review kai zhang 2 participants 2 messages Jan 24, 2022

Subsystems (3)
lore
Share & tools
lore

kai zhang

author
#0 lore raw

PATCH 1/1 net: fix duplicate logs of iptables TRACE target

Below configuration, mangle,filter and security tables have no rule:

There are 5 logs for incoming ssh packet:

kernel: [ 7018.727278] TRACE: raw:PREROUTING:policy:2 IN=enp9s0 ...
kernel: [ 7018.727304] TRACE: mangle:PREROUTING:policy:1 IN=enp9s0 ...
kernel: [ 7018.727327] TRACE: mangle:INPUT:policy:1 IN=enp9s0 ...
kernel: [ 7018.727343] TRACE: filter:INPUT:policy:1 IN=enp9s0 ...
kernel: [ 7018.727359] TRACE: security:INPUT:policy:1 IN=enp9s0 ...

---
 net/ipv4/netfilter/ip_tables.c  | 4 +++-
 net/ipv6/netfilter/ip6_tables.c | 4 +++-
 2 files changed, 6 insertions(+), 2 deletions(-)
net/ipv4/netfilter/ip_tables.c +3 -1
index 2ed7c58b4..5f0e6096e 100644
@@ -304,9 +304,11 @@ ipt_do_table(void *priv,
 
 #if IS_ENABLED(CONFIG_NETFILTER_XT_TARGET_TRACE)
 		/* The packet is traced: log it */
-		if (unlikely(skb->nf_trace))
+		if (unlikely(skb->nf_trace)) {
 			trace_packet(state->net, skb, hook, state->in,
 				     state->out, table->name, private, e);
+			nf_reset_trace(skb);
+		}
 #endif
 		/* Standard target? */
 		if (!t->u.kernel.target->target) {
net/ipv6/netfilter/ip6_tables.c +3 -1
index 2d816277f..ae842a835 100644
@@ -327,9 +327,11 @@ ip6t_do_table(void *priv, struct sk_buff *skb,
 
 #if IS_ENABLED(CONFIG_NETFILTER_XT_TARGET_TRACE)
 		/* The packet is traced: log it */
-		if (unlikely(skb->nf_trace))
+		if (unlikely(skb->nf_trace)) {
 			trace_packet(state->net, skb, hook, state->in,
 				     state->out, table->name, private, e);
+			nf_reset_trace(skb);
+		}
 #endif
 		/* Standard target? */
 		if (!t->u.kernel.target->target) {
Signature · 1 line
2.30.2
maintainer
#1 lore raw
Quoted 9 lines · kai zhang

Below configuration, mangle,filter and security tables have no rule:

There are 5 logs for incoming ssh packet:

kernel: [ 7018.727278] TRACE: raw:PREROUTING:policy:2 IN=enp9s0 ...
kernel: [ 7018.727304] TRACE: mangle:PREROUTING:policy:1 IN=enp9s0 ...
kernel: [ 7018.727327] TRACE: mangle:INPUT:policy:1 IN=enp9s0 ...
kernel: [ 7018.727343] TRACE: filter:INPUT:policy:1 IN=enp9s0 ...
kernel: [ 7018.727359] TRACE: security:INPUT:policy:1 IN=enp9s0 ...

Thats correct and exactly whats supposed to happen.

net/ipv4/netfilter/ip_tables.c ipt_do_table(void *priv, · line 311

Quoted 8 lines

#if IS_ENABLED(CONFIG_NETFILTER_XT_TARGET_TRACE)

net/ipv4/netfilter/ip_tables.c +3 -1
 		/* The packet is traced: log it */
-		if (unlikely(skb->nf_trace))
+		if (unlikely(skb->nf_trace)) {
 			trace_packet(state->net, skb, hook, state->in,
 				     state->out, table->name, private, e);
+			nf_reset_trace(skb);
+		}

This breaks the long established behavior of TRACE,
we don't want users to have to TRACE tables individually which may also be hard when nat is involved.