S?bastien Bourdeauducq wrote:
> Do you see errors on the bus?

This is a bit tricky because I can't monitor the bus for more than
a few frames. But let me think if I can set up things such that I
can detect errors after the fact and then raise a trigger.

> That's something messy, let's try to avoid it.

The problem is that there must be less than 1.3 us between the end
of the EOP of the DATAx packet and the SOP of the ACK. Is the
Navre fast enough for this ?

And that's at the device side. If we want a design that doesn't
need a lot more fixing when adding hubs, the delay should be even
shorter. The USB standard defines a turn-around time of only about
0.5 us for full-speed.

For auto-ACK, I was thinking of still managing all the state in
software. Software would tell the receiver beforehand what packet
type and maybe what maximum size to expect. Then the receiver
would need to maintain relatively little extra state.

- Werner
_______________________________________________
http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org
IRC: #milkymist@Freenode

Reply via email to