EL-160 w / xoff-xon software flow control enabled

I noticed if a line tx buffer gets full(on the el-160), the el-160 issues a xoff to the hardware connected to the affected port. This is an expected event.

My question is, is this Xoff generated by the firmware in the etherlite hardware itself, when software flow control is enabled, or does it require the driver on the server to send the xoff. I would expect it to be localy generated, independant of the server/driver but want to verify that fact.

Thanks
JJ

This is configured/controlled from the driver/host.

I am aware that it is set/configurable from the host/driver, that is why I expect flow control characters.

But if the tx buffer for a line is full on the etherlite, and another character is inputed, isn’t it the etherlite el-160 that issues the actual Xoff, and not the driver? How would the driver know another charater was inputed, if the tx buffer is full? (not being emptied by the attached task).

It would seem that if I set flow control on the etherlite, The etherlite would issue the software flow control signals, based on the status of the ports rx and tx buffers.

Or is there some magic I’m missing that enables the driver/host to be aware of each ports tx&rx buffer status, and can bypass them to issue software flow control characters?

Thanks
JJ

The unit/firmware will send the Xoff.

However, as a sidenote, the driver will eventually beome aware of the extra data being received by the unit, because the RealPort protocol will send updates back the driver every so often.

Of course, these unit -> driver updates are no where near as fast as needed to make xon/xoff work well, which is why the unit/firmware does the actual enforcing of flow control.

Scottk

scottk,
That is the behavior I was hoping for. :^)

Thanks for the info…
JJ