> On Thu, 4 Mar 1999, Martin Mares wrote:
>
> > > And how do you use more than 4 gigabytes of RAM on these boring Intel
> > > boxes _without_ segmentation? ;)
> > > [... and don't say "ram drives" - unless you have a new meaning for that ;)]
> >
> > The answer is "You don't want to do this". Using >4G of RAM on a 32-bit
> > CPU is brain-damaged.
>
> Still I think this brain-damage has already been implemented somewhere...
> maybe even at the brain-damage center at redmond ;)
> ... and it definitely was implemented for 32-bit memory addressing on
> 16-bit cpus ;)
Sure, and everyone thought that was brain damaged too. It made for some
very complicated user space code, and I dread to think how complex many
286 UNIX variants became. Moving to a flat address space was the "thing
to do" when writing 386 DOS applications.
Just for contrast, it is possible to get a 1Mb ram pack for the C64, an
8-bit machine with technically only 64kb address space. It was achieved
using bank switching, and it was a stupid idea then too. It was simpler
to just get a 16-bit machine, and often faster/cheaper as well.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/