Date   : Thu, 28 Jul 2005 00:27:49 +0100
From   : Rob <robert@...>
Subject: Re: Econet-Ethernet bridge

At 19:12 27/07/2005, Johan Heuseveldt wrote:

>There are two other hardware addresses - which don't access the 68B54
>- and allow/disallow the NMI from the 68B54 to the NMI line. They are
>called 'notINTOFF/notSTATID' and 'notINTON'.
>
>The NFS software also uses this to do some processing while servicing
>the NMI interrupt, and polling the device when it suits.

That's a curiosity I found when working on emulating the econet hardware - 
because the interrupt enable mechanism for the NMI line is basically a NAND 
gate (IC27) with the (inverted) notIRQ line from the ADLC and the latched 
enable signal, NFS actually delays NMI's !  The NFS disables the interrupts 
at certain points, something happens at the ADLC and it signals an 
interrupt, but it's being masked, so IRQ remains asserted but ignored.  A 
tad later, the NFS re-enables interrupts, because notIRQ is still asserted, 
it's let through to the CPU's NMI input, and an interrupt is immediately 
registered, and dealt with.  ("How to make an LDA cause an NMI"!)

Why it has to do it quite like this I don't know, but NFS3.60 needs it to 
work this way..

(For what it's worth, the hardware emulation all works now - I just got 
stalled on getting the data out into IP packets and across to another 
emulator on another machine!)

Rob