Re: ZIP Chip Heartbreak

[toddpw@frappe.ugcs.caltech.edu (Todd P. Whitesel)]

gtolar@blkbox.com (Glynne Tolar) writes:

>supported by all //e accelerators) couldn't the CPU when it detects a DMA 
>signal (that I know makes it to the CPU!) just flush its cache?  I know 
>you'ld have some speed loss.  Just a thought.

If you could get the DMA signal to the CPU, sure.

However the Zip8 would need a test clip for that, and that's more trouble
during installation than many customers want to deal with.

On the 8 bit Apple CPU socket, the only way to detect DMA is to watch for
a suspiciously long delay in the main input clock, and this is not easy
to do reliably without some tricky analog circuitry. I think Zip did the
right thing to not tackle it.

Todd Whitesel
toddpw @ ugcs.caltech.edu