On Wed, Apr 10, 2002 at 09:56:51AM -0400, gabriel rosenkoetter wrote:
> I share other people's concerns about taking anything from next68k.
> It's far from mature. But if you think it works, great. :^>

Well, the bus_dma code is from the alpha, modified somewhat to be
applied on the NeXT platform.  I understand that the x68k came from
i386 originally, but they're quite similar, in any case.  The
bus_dmamap_sync() is the main place where the next68k and x68k
implementations were different.  The next68k implementation looked
cleaner to me, but that may not mean anything...

> (Granted, neither my turbocube nor my 030 cube gets far enough along
> for me to have any concrete "here's why SCSI's bad on NeXTs"
> examples, but...)

This is actually more a problem with the DMA controller, I think,
since the NeXT DMA controller had to be reverse engineered with fewer
docs that we usually have and with tools much more primitive than
MacsBug...

Of course, we do want a working bus_dma, so we'll do what we need to
get there.  :-)

-allen

-- 
 Allen Briggs                     briggs@wasabisystems.com
 http://www.wasabisystems.com/    Quality NetBSD CDs, Sales, Support, Service
NetBSD development for Alpha, ARM, M68K, MIPS, PowerPC, SuperH, XScale, etc...