> On Fri, 5 Feb 1999, Anders Magnusson wrote:
> 
> > The kernel does not rely on any floating-point, so that's no problem.
> > Does it use MTPR for console in/out? If it does, it will be simple to
> > get the kernel to say at least "hello". 
> 
> I'm going to have to guess no on that one.  I'm not sure how console I/O
> works on it, maybe like MicroVAX II? 
> 
MicroVAX II uses MTPR for console. Try:
>>> D/I 23 40
And see if something shows up on the screen.

> 
> > Part 2 of this is to make the uba routines to stop careing about map
> > registers, but that will also be rather simple I suppose.
> 
> Based on what Matt Thomas said about the effects on DMA due to the lack of
> these registers, will it really be easy to implement?  
> 
> Matt, I'm brushing up on my DMA I/O concepts with an old computer
> architecture textbook, but I may require a more detailed explanation of
> what you mean by doing DMA bounced or handling the scatter/gather
> ourselves?  
> 
> Without any mapping registers, how do the CPU and DMA I/O devices agree
> upon locations for transferring blocks of data?
> 
Actually, the principle for this is simple: All I/O devices are able to
access all of the computer internal memory. A 1:1 mapping (as in this case)
is much simpler to deal with than the mapping registers when you have
a smaller address space for the bus than the internal memory.

-- Ragge