developing drivers for custom boards

This is just awesome.

Now I suppose things like RESET thread will have to go away. And I assume it stands to reason my step speed will be the same if not slower using the port this way. but overall I think I will be much happier with the setup.

Great idea. So emc will send the step high, then my card will reset it low after some finite time based on my driver needs. I like that alot!!

See, this is why I need to get these things ironed out before I write the code for my cards. At least the how it will work part.

Not sure about the protocol, but its no problem if I know what it wants to be. It will already have a very fundamental protocol in the fact that 4 bytes will be sent- and they will be read as such. So for instance R0 = steps, R1 = dir, R2 = outputs, R3 = spindle speed. For now I will assume all 4 reach the card as they should. I am thinking for error checking, maybe set or clear the MSB's for the 4 bytes read back. 1 is good transmission and 0 is bad. This will be based soley on receiving 4 bytes. if less or more are received then it is an error. Maybe that is not needed? The reason for using all 4 msb's is just in case something goes wrong reading back the info. If emc gets at least 1 byte it will know there is a problem.

On the other hand, the step pins will always be the same pins - not configurable. So unless I am missing something (good chance of that), seems like I could just always drive the steps low x-time after receiving from the port.

How about one address write to reset your state machine on the EPP remote
so it cannot get out of sync with the host, that is write the step signals with an address write
and the other signals with data writes/reads which increment your data register pointer

Oh so many great things I would never have thought of. so on the first outb instead of outb (R0,base+4) use base +3. That is just too easy. Thanks PCW.
For whatever reason, I was taking it at face value that if I did an address write, I needed to send an address. Doh! that will just use the address write line from the EPP port to signal to my card to reset the pointer no matter what.

Still plugging away at this. Been BUSY at work so no real progress just yet. But I have another question. Assuming I reset the steps in hardware, won't I need to tell EMC/HAL not to worry with it? I get that not setting a reset will tend to that part, but how do I tell it to always generate the start but never worry about the reset? Or if I need to keep the reset thread in there, how to not actually send the reset to the port?

Or maybe I am half asleep (again), it is Monday morning afterall.