On 8/15/06, Jack Carroll <[EMAIL PROTECTED]> wrote:

        That's why I suggested putting some bare-bones minimal logic in the
TRV10 to allow it to use the SPI port to probe for an external configuration
tool, from which it can read in the raw binary image of the video mode
registers.  This would get the display up and running at boot time in text
mode, with zero CPU intervention.  That makes the console available to begin
debugging the very first low-level code on the target system.  The parts
cost for a minimal initialization tool implemented with DIP switches and
shift registers came to around $25 in moderate quantities, as I recall.
There might not be all that many initialization tools sold, but it could
mean a big difference in chips sold into the embedded market.

I can see the benefit of having this.  But almost nothing in any
computer can boot up and do anything useful before the CPU does
something to initialize it.  Why should this be any different, from,
say, configuring kernel page tables so that the CPU can execute the
rest of the boot ROM?  I don't know how much logic would be required
to do what you're asking for, but it goes against the general
philosophy of pushing off as much work as is practical onto the host
CPU.

Here are the things that need to be initialized:

- Program memory controller with appropriate timings
- Load video program into video controller
- Load VGA program into VGA/DMA controller

Admittedly, this COULD mostly be just a bunch of copying.  I just
loath the idea of adding yet another controller to this thing.  I
suppose it's on the list of "if it can fit" items.

You know, I was thinking... one solution we may want to consider here
is selling a VT102 terminal as a product.  We could implement
everything on an FPGA, so in a small box, you see power, PS/2
keyboard, monitor, and serial ports.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to