Well, I suppose if someone decides to send me a TCP segment bigger than 1500 bytes, it would result in several fragmented IP datagrams. Fragmented datagrams are discarded by the stack. They don't frighten me at all.

Concerning the counters, all counters that HACK-RTOS uses are signed 8-bit ones. That is, I only have to test for 255 system ticks or seconds depending on the counter type. Lucky me!

What I want to test now is a communication over a router, where eZ80 is in one subnet and the client is in another. The immediate problem I'm facing is that I don't have a router. Will have to find a way to make a PC to do the job.

I have done what I wanted today. I have made a router out of my PC, and have got two subnets, plugged eZ80 in one of them, started several TCP clients in both subnets plus a few pings, and guess what? It just started to work! Truthfully, I expected troubles because I never much tested that "default router" thing in the stack, but no. After 5 hours of testing no problems whatsoever.

John, if I may ask, when you need to update the firmware of the unit in the field what do you do? Does RZK 1.34 provide means to do it remotely over Ethernet, or do you need to climb towers and ride helicopters?

See if you can find yourself a used Cisco rack-mount router with their store and forward, that might be a better test, and then than feeding a hubbed net - we catch more problems that way. You want the subjects under test to see lots of traffic, in lots of combinations, with lots of different setups. You'll also find out that different MAC / PHY chips behave differently with different Ciscos...and I'll leave that for you to discover;)

Also find out the differences between the Acclaim! MCU and the Acclaim!Plus. They aren't the same, and they are not drop-in replacements for each other. You'll find that out also.

Yes, we use a remote update system over Ethernet, but because the equipment has to stay operating, it will fail over to the backup controller during a firmware upgrade - nothing is ever, ever powered down or reset. The system is vulnerable though because it absolutely can't get hung during the firmware upgrade - that might require a physical presence.

We really, really try to not release anything until its tested and working -we send a minimal amount of software upgrades out (one per year at most unless there is a major problem). We don't believe in making the end customers be beta testers, as much as we can.

Today, while testing a new thing for HACK-RTOS - the remote update capability, I have run into some troubles with my TCP stack. Apparently, the new thing uses it in a way different than that of my old TCP test client. There is still work for me to do.

You know, John, maybe you should remind Zilog of your original request. They might believe I have successfully put it out of your mind with my HACK-RTOS.