[Open-graphics] bootstrapping the video mode

[Timothy Miller]

On 8/3/06, Nicolas Boulay <nicolas.boulay at gmail.com> wrote:
> Such a cpu is tiny compare to the reste of the GPU in terme of space.
> It offer great flexibility and provide means to correct some hardware
> bug and adding some feature that could be very complexe and slow in
> pure hardware.
>

I'm thinking about returning to the nanocontroller idea for VGA (and
saying so-what to the massive latency for VGA reads).  Some VGA stuff
could possibly require specialized assist hardware, but the majority
of it could be done in software.  This controller could run
double-duty for a number of other purposes.

We would give this processor access to graphics memory and some user
I/O signals, at least.  VGA accesses would cause interrupts.  We could
also give it access to the GPU and the DMA contoller.