On 3 May 2013 16:28, Florian Obser <[email protected]> wrote: > Chris Ivancic & Colin Ligertwood reported in January that the > "SolarWinds NetFlow Traffic Analyzer" doesn't like it if the > ingress/egress interface index is always 0 (v5) or not present at all > (v9/v10). This keeps track on which interface a packet for a state was > first observed / first left and exports it in pflow(4). > > OK? >
OK, but I think all pads should to be gone from struct pf_state since it's not packed. Whatever was the intention, it doesn't seem to be relevant anymore.
