|
From: Paul K. <pk...@eq...> - 2002-06-17 20:44:20
|
>>>>> "Gregg" == Gregg C Levine <dr...@wo...> writes: Gregg> ... Can you post a recipe behind how you did Gregg> your testing? As I mentioned, I used ethertap, which is a configurable feature in Linux 2.4 (and 2.2 as well, I think). The biggest issue with ethertap (apart from inadequate documentation) is that out of the box it only simulates an ethernet port, not a LAN. In other words, if you have several of them, sending to one doesn't cause the data in question to appear on the other ports. So I wrote a little daemon which makes that happen, by posting reads on each of the tap<n> units and propagating any received packets to the other tap<n> units. I sent that to the author of the ethertap facility, but I don't know what has happened to it since. If there's interest, I can post it here; it's small enough. So then I configured tap0, tap1, and tap2. I hook them up with ethertapd, tell decnet/linux to use tap0, tell ersatz-11 to simulate a UNA or QNA on tap1, and give tap2 to tcpdump. Prestochango -- a LAN in a laptop, not bothering anyone, with full packet trace and two separate decnet implementations talking to each other. This is how I did the rmterm support for decnet/e... paul |
