On Mon, Jun 14, 2010 at 10:20 PM, Patrick Coleman <[email protected]> wrote: > On Tue, Jun 15, 2010 at 1:03 PM, LeviaComm Networks NOC > <[email protected]> wrote: >> It would be best if you had a working switch to test with, the switch may > be >> forwarding packets to the OpenBSD box because its MAC table is broken. The >> switch may be the cause, please confirm that it isn't before making noise. >> I am sure that no one wants to waste time casing down a bug and then >> finding out that it was the switch all along. > > Sure, I acknowledge there may be something broken there. But tcpdump > on the OpenBSD box indicates the MAC addresses of the traffic received > do not match any MAC address on the OpenBSD box. In this case OpenBSD > should be simply discarding the packets, not transmitting spoofed RSTs > for TCP conversations it is not involved in. > > The situation is basically the same as if OpenBSD was connected to a > hub, not a switch. In that case, it would be receiving every packet > traversing the local subnet. > > I'm not denying I might have configured OpenBSD wrong somehow - if so, > any ideas as to where would be greatly appreciated.
This thread is of some interest to me. I have experienced a somewhat similar condition. I don't wish to add noise to this thread, as I don't have much data to contribute, but I thought I'd share the following info/story. It may not be of much use I'm afraid. I started using trunk(4) interface with my ibook (at work) in failover mode with ethernet port and wifi ports; master port being the wired ethernet port (gem0). What I noticed was that whenever I would unplug the ibook, the adjacent PC would have all its TCP sessions dropped. The PC and ibook in my office shared the same netgear switch (FS105). First few times I thought it was just a fluke, but soon I correlated it with I unplugging the ethernet cable from the ibook. IIRC, pflog showed 'block in' rule on the ibook acting on traffic not indented for the ibook. Unfortunately, I didn't have much chance to debug this further as our work wifi network was soon redone by $newcompany. The new wifi routers were pretty much useless; $newcompanyIT blamed the building's structure for reliability issues. An initial assumption was an incorrect/invalid set-up on my part -- it was my first attempt to play with trunk(4) -- also I couldn't find anyone else reporting such problem on m...@. Before playing with trunk(4), when I would disconnect the ibook and hop on the wifi network I would manually `ifconfig gem0 delete; route delete default; ifconfig bwi0 nwid $nwid ... up ; route add default $gw`. This procedure never caused above described issue. I'm going to lurk now, --patrick
