Skip to contentSkip to conversation

Help: Marking UDP packets in a bridge

DISCUSSION Aravindhan Dhanasekaran 3 participants 4 messages Oct 10, 2014 – Oct 13, 2014

lore
Share & tools
lore
#0 lore raw

Hello,

I'm trying to mark UDP packets entering (or leaving) a bridge, destined to a particular UDP port on a machine on the other side of the bridge.

My simple topology looks like:
host1 [eth1] <-----> [s1-eth1] bridge [s1-eth2] <-----> [eth1] host2

I've added a rule to the FORWARD chain on the mangle table in the bridge to mark the packets that I require:

$ sudo iptables -t mangle -A FORWARD -p udp --dport 9917 -j MARK --set-mark 17
$ iptables -L FORWARD -t mangle -v
Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source       destination
    0     0 MARK       udp  --  any    any     anywhere     anywhere      udp
dpt:9917 MARK set 0x11

But, looks like none of packets are being marked (counters are all 0s in iptables output as shown above). I have traffic matching the above rule flowing through the bridge which I verified using tcpdump.

$ sudo tcpdump -i s1-eth1 udp dst port 9917
...
11:22:14.774417 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774597 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774731 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470^C

999 packets captured
1831 packets received by filter
795 packets dropped by kernel

I'm not sure what's wrong with my configuration. Any help would be appreciated.

Thanks,
/Aravind

#1 lore raw
Quoted 19 lines · Aravindhan Dhanasekaran

Hello,

I'm trying to mark UDP packets entering (or leaving) a bridge, destined to a particular UDP port on a machine on the other side of the bridge.

My simple topology looks like:
host1 [eth1] <-----> [s1-eth1] bridge [s1-eth2] <-----> [eth1] host2

I've added a rule to the FORWARD chain on the mangle table in the bridge to mark the packets that I require:

$ sudo iptables -t mangle -A FORWARD -p udp --dport 9917 -j MARK --set-mark 17
$ iptables -L FORWARD -t mangle -v
Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
  pkts bytes target     prot opt in     out     source       destination
     0     0 MARK       udp  --  any    any     anywhere     anywhere      udp
dpt:9917 MARK set 0x11
Quoted 8 lines

But, looks like none of packets are being marked (counters are all 0s in iptables output as shown above). I have traffic matching the above rule flowing through the bridge which I verified using tcpdump.

$ sudo tcpdump -i s1-eth1 udp dst port 9917
...
11:22:14.774417 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774597 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774731 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470^C

May be, we should do the packet marking using ebtables (instead of iptables) for Ethernet bridge. I am not sure..

vignesh

#3 lore raw
In reply to Aravindhan Dhanasekaran

Hello.

Check sysctl variable bridge-nf-call-iptables. It should be equals 1, otherwise bridged traffic don't pass iptables.

2014-10-10 19:28 GMT+04:00 Aravindhan Dhanasekaran <[email protected]>:

Quoted 41 lines

Hello,

I'm trying to mark UDP packets entering (or leaving) a bridge, destined to a particular UDP port on a machine on the other side of the bridge.

My simple topology looks like:
host1 [eth1] <-----> [s1-eth1] bridge [s1-eth2] <-----> [eth1] host2

I've added a rule to the FORWARD chain on the mangle table in the bridge to mark the packets that I require:

$ sudo iptables -t mangle -A FORWARD -p udp --dport 9917 -j MARK --set-mark 17
$ iptables -L FORWARD -t mangle -v
Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source       destination
    0     0 MARK       udp  --  any    any     anywhere     anywhere      udp
dpt:9917 MARK set 0x11

But, looks like none of packets are being marked (counters are all 0s in iptables output as shown above). I have traffic matching the above rule flowing through the bridge which I verified using tcpdump.

$ sudo tcpdump -i s1-eth1 udp dst port 9917
...
11:22:14.774417 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774597 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774731 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470^C

999 packets captured
1831 packets received by filter
795 packets dropped by kernel

I'm not sure what's wrong with my configuration. Any help would be appreciated.

Thanks,
/Aravind
--
To unsubscribe from this list: send the line "unsubscribe netfilter" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Signature · 1 line
Anton.
#2 lore raw
In reply to Vigneswaran R
Quoted 29 lines · Vigneswaran R
Quoted 19 lines · Aravindhan Dhanasekaran

Hello,

I'm trying to mark UDP packets entering (or leaving) a bridge, destined to a particular UDP port on a machine on the other side of the bridge.

My simple topology looks like:
host1 [eth1] <-----> [s1-eth1] bridge [s1-eth2] <-----> [eth1] host2

I've added a rule to the FORWARD chain on the mangle table in the bridge to mark the packets that I require:

$ sudo iptables -t mangle -A FORWARD -p udp --dport 9917 -j MARK --set-mark 17
$ iptables -L FORWARD -t mangle -v
Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
  pkts bytes target     prot opt in     out     source       destination
     0     0 MARK       udp  --  any    any     anywhere     anywhere      udp
dpt:9917 MARK set 0x11
Quoted 8 lines

But, looks like none of packets are being marked (counters are all 0s in iptables output as shown above). I have traffic matching the above rule flowing through the bridge which I verified using tcpdump.

$ sudo tcpdump -i s1-eth1 udp dst port 9917
...
11:22:14.774417 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774597 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470
11:22:14.774731 IP 10.0.0.2.49774 > 10.0.0.1.9917: UDP, length 1470^C
Quoted 2 lines

May be, we should do the packet marking using ebtables (instead of iptables) for Ethernet bridge. I am not sure..

When I added the above rule as a filter to my qdiscs, it worked as expected. So, I guess my packets were actually marked, just that the counters were 0.