Geoff Body wrote:
Steve,
I have been talking with Tony who was kind enough to send a network capture which shows the packets from the Ethernet card in the IIgs being incorrectly formed with the length field being 14 bytes longer than it should. The 14 bytes being the source and destination mac address 6 bytes each + 2 byte length field, the first 14 bytes of the packet which should be excluded from the length. Linux was not responding to these packets, so it would appear to be discarding these packets as faulty.
What's odd is that Tony tried the IIgs through a Farallon etherwave bridge and had similar (bad) results. I believe he mentioned trying a Gatorbox, too.
But, it seems you may be on to something.
The main change was that Phase 2 allowed a range of network addresses instead of a single network address to be allocated to a network allowing more than 253 to be on the network at once. Phase 2 also saw the introduction of Tokentalk. Zones existed prior to Phase 2.
That's what I thought. The claim on an old usenet post about Linux DDP lacking support for Phase 1 is, I believe, incorrect.
I don't think bridges are required to have zones as they make the device appear as if it is on the other side of the bridge. and there might not be a router on the other side of the bridge to provide the zone or network address information. I haven't played with bridges and that is one of the reasons I was asking if a network capture could be provided. It may help me to modify Linux appletalk/netatalk to support these types of devices.
So, to be sure my terminology is correct: My Gatorbox, and its predecessor Kinetics FastPath 4, are routers rather than bridges? Or, can they operate in a bridged mode as well?
Steve