netgraph issue

[sheeda]

hello everyone,

i am new to netgraph and have been experimenting with few very basic configurations and facing some issues.

i have got 2 machines, A & C, connected via third machine B with dual nics, with everything in same subnet.

A <-----> B <-----> C

i have used the bridge script in /usr/share/examples/netgraph and i am successfully able to ping between A and C.

i did some experiments and getting weird results:

1. i cleaned up the tee node and compiled a pipe node by removing the lefttoright and righttoleft hooks. once i load the ng_pipe.ko module and place it between the two ether nodes on lower hooks, i can't ping between A and C. i even switched the 2 interfaces to promisc mode and disabled the mac auto source thing through the msg command.

(lnc0)-----(pipe)-----(lnc1)

the ping strangely starts working if i start and stop the bridge script and then connect my pipe node, even after unloading the ng_bridge module.

2. i suspected some issue with my pipe netgraph node, therefore connected the two ether nodes directly via lower hooks and had the same issue. the A & C starts communicating again once i start and stop the bridge script and then connect the lower hooks and set the promisious modes.

(lnc0)----------------(lnc1)
            lower

all my 3 machines are freebsd release 3 in vmware. i have even tried setting up the static arps on A and C but it makes no difference.

another interesting thing i found was that if i shutdown my pipe node, the two ether nodes connect directly on lower hooks and the ping continues to work.

please let me know if this is all the default behaviour or if i am missing something.

thanks to everyone for their help in advance.

regards,

sheeda