Hi, i discovered the same behavior. It's seem to be related to the usage of pcap. If i use the following example, it works (no loop). FromDevice(eth0, METHOD LINUX, PROMISC true)->Queue->ToDevice(eth0, METHOD LINUX );
The next script causes the loop: FromDevice(eth0, METHOD PCAP, PROMISC true)->Queue->ToDevice(eth0); You didn't use the param METHOD, but in the git PCAP is the default capture method. Can you please check, whether you script causes a loop, if you use "METHOD LINUX" ? Best regards, Robert On Samstag, 9. Juli 2011, Antonie Henning wrote: > Hi, > > Something I discovered with the latest git pull code: > > Using the same interface in a fromdevice and todevice element, seems to > cause a loop e.g. > > Eth0 receives 1 pps but transmits e.g. 80k pps @ 1500bytes or line rate > 1gbps. This can be replicated with the config below: > > FromDevice(eth0, PROMISC true)->Queue->ToDevice(eth0); > > To test this the following example can be used at the source of the > packets: > > // Where ethX is connected to eth0: > RatedSource(DATA \<DATA>, RATE 1, LIMIT 10, ACTIVE true, STOP true) -> > Queue -> ToDevice(ethX); FromDevice(ethX, PROMISC true) > -> c::Counter -> Queue -> Discard; > > > Script( > wait 1, > print c.rate, > loop > ); > > Changing the interface in the ToDevice to a different interface e.g. eth1 > solves the problem... > > When adding another fromdevice and todevice pair even with different > interfaces (as long as a fromdevice interface and todevice interface match > somewhere in the config) the problem returns e.g FromDevice(eth0, PROMISC > true)->Queue->ToDevice(eth1); > FromDevice(eth1, PROMISC true)->Queue->ToDevice(eth0); > > I have checked whether this is because of promiscuous mode but found the > same result when removing promisc and using EtherEncap. > > A > _______________________________________________ > click mailing list > [email protected] > https://amsterdam.lcs.mit.edu/mailman/listinfo/click _______________________________________________ click mailing list [email protected] https://amsterdam.lcs.mit.edu/mailman/listinfo/click
