On Mon, 14 Jan 2002, gabriel rosenkoetter wrote:

> On Mon, Jan 14, 2002 at 02:04:50PM -0800, David A. Gatwood wrote:
> > Well, the difference between mac68k and macppc is kind of obvious, but the
> > differences between a nubus ppc and a PCI ppc are pretty small by
> > comparison.  Other than a little bit of tweaking at the kernel entry
> > point, the only huge issue is the lack of OF.
> 
> Hrm.
> 
> I'm not sure I agree. I think a PDM (which stands for Piltdown Man,
> for those following along at home; as in, the missing link... they
> were thinking between 68k macs and Taligent's Tesserect, but it's
> ironically appropriate now) is closer to the 68k line than to the
> later PCI PowerPC line in many respects. Am I mistaken?

It's pretty similar to PCI.  The only thing it had in common with the
nubus machines, at least from a programming perspective, was the existence
of nubus itself, which is supported through a BART bridge chip so that it
is relatively isolated from the CPU bus.

Interrupt handling is like PCI, in that all interrupts except for those
from actual nubus cards can be acked in the code for the interrupt
controller itself (nubus cards being constrained by compatibility with
existing cards).  All the motherboard chips are roughly the same as the
7x00 PCI machines (Cuda, MACE, 53C94 SCSI, AWACS).


> > That's an awful lot of work for two systems as similar as the nubus and
> > pci PowerMacs....
> 
> Are they really so similar? The *processor* is, sure, but isn't most
> of the motherboard hardware fairly different? (I'm not talking about
> SCSI or even the swim3 here, but RAM, expansion ports, video
> hardware... yes, I know that we don't care, you don't think we
> should deal with it yet, and we can hijack it from Mac OS, so those
> aren't the best examples, but...)

Video hardware is different, yes, but video hardware is different from one
PCI machine to the next as well.  I count ten different types of video
that have been used on various powermac models -- and that's only if you
fold all the different ATIs into a single entry.  :-)  RAM is
non-contiguous, but that should be relatively easy to deal with, given
that the VM system already has to support that for mac68k.



> > 7.  601 support
> 
> Well, I lack any PDM (I typically type x100, but this works just
> as well) PowerMacs, but I do seem to have a 601 with a PCI bus
> handy. So maybe I'll play with that, eh? :^>

Yeah, that'll be a really good step in the right direction.


> > #7-8 can be ignored as long as you have a Newer accelerator card
> > and are using Apple-supplied video (i.e. not a nubus card).
> 
> Sure, but #7 buys us 7200 (and 7400? Is that the right model
> number?) support and gives support for PDMs another leg to stand on.

7500.  And yes, it definitely makes sense to work on it for that reason.
It just isn't a "show-stopper".  :-)


> > The old 601 processor reference from Motorola, 'cept that it's out of
> > print.  You can probably find a copy used somewhere or borrow one from
> > someone.  If I had a copy, I'd loan it to you.
> 
> You mean like this:
> 
>   http://www.mot.com/pub/SPS/PowerPC/library/user_man/601UM.pdf

Maybe.  I'm getting a 10 byte file, but that may be an IE issue.


> [Lose old PPCs or lose old OF?]
> > My guess is... both?  ;-)
> 
> My guess is force-feed marketing of newer systems, actually. Apple's
> making more money from hardware than software, last I heard.

Well, there's a pretty heavy cost in maintaining support for a bunch of
old systems, even just in terms of all the testing and verification needed
to make sure that they work.  Not to mention that it would be a
second^h^h^h^h^h^hthird booting mechanism to support....


David

---------------------------------------------------------------------
                    Check out my weekly web comic:
                     http://www.techmagazine.org