On Mon, 2 Aug 1999, Colin Wood wrote:

> well, it'll just involve a little bit of digging into pre-existing kernel
> code :-)  of course, i'm looking at the pas driver now, and it might be a
> little more difficult than i originally thought.  it's pretty damn
> i386-specific at the moment (yuck!).

...

> well, given that the card appears to more or less be a soundblaster clone,
> you'd need to familiarize yourself with the pas driver, the soundblaster
> driver (sbdsp.c) and the MI audio driver.  you'd also need to learn the
> mac68k nubus driver.  once you've got all that straight, you basically
> have to figure out where the card keeps its registers in nubus space, map
> them in, and then operate on them in ways similar to how the sb driver
> does.  of course, the ISA version is DMA-based, and i don't know exactly
> how the nubus version would handle this...
> 
> so, you can probably begin to see the size of the task if you want to pick
> something like this up...

Hmmm.  I don't mind checking out the existing code, but since most of the
time I spend on my NetBSD machine is remotely connected from work, I don't
imagine that I'd have much time to play around finding hardware addresses
and stuff, since I'd probably need to be at the console.  Also, figuring
out how to do the ISA DMA stuff on NuBus could be daunting.  Another
concern is that some features of the card never worked well anyway, even
with multiple MacOS driver revisions.  I'd heard that it was a hardware
bug.  That might make it even harder to work with. 

I was kinda hoping for something a bit simpler - like an advanced "Hello,
world!" equivalent, maybe  :-)

/*======================================================================
"Don't die wondering..."                http://www.cldc.howard.edu/~cpg
                                              email: cpg@scs.howard.edu
chris out-              Christopher P. Gill
  peace.        C.L.D.C. Senior System Operator (Ret.)
======================================================================*/