IP address woes

[Chris Johns]

Angelo Fraietta wrote:
> 
> You are right in that it is not in the arp cache - I tested with arp - a 
> , and the RTEMS machine was definately not there on the list - even 
> though it responded back directly. Is there a way to set it  to load it 
> to the cache when it receives a message?
> 

I assume the 'arp -a' command was run on MacOS. If so it would seem 
interesting to have this happen as the Steven's book is based on a BSD 
stack and so is MacOS X and RTEMS. I assume this is MacOS X.

There are ways to directly play with the ARP cache but I would not 
recommend this approach.

> Also, when the RTEMS is unable to respond to the broadcast from the mac, 
> it is able to respond to a broadcast from the Windows machine.

What do mean by cannot respond ?

> Why is 
> that the case? When this occurs, I unplug and replug my mac back in, and 
> RTEMS can now find it - but the mac still can't find the RTEMS.

Not sure. Maybe the Mac's ARP cache is being flushed when disconnected.

I would check the ARP cache entries against the tcpdump data.

> How does a ping work? The machine pinging does not yet have the ethernet 
> address before the ping?

Ping is an ICMP echo command. ICMP sits next to IP and uses the same ARP 
cache in the same manner as IP.

What are the IP network configurations of each machine involved ?

> I don't think that this is the case.

You can prove this with all machines connected and tcpdump. It may be a 
good to check this out.

Also how are these machine connected, ie hub, switch etc ?

-- 
  Chris Johns