On Thursday, January 18, 2007 9:45 PM [GMT+1],
Rich Mellor <rich@xxxxxxxxxxxxxxxxxx> wrote:
Ah but the comments need to be in English as well :-)
We are the ones developing the divIDE Plus which has been referred to.
wow!
I developed a clone with a totally different mass storage interface
(sd/mmc card). Garry Lancaster (ResiDOS and +3e rom author) made (is
making) a custom release of both, where the sd card is used in place of
compactflash or ide disk.
In order to support ResiDOS, he asked me to implement some additional
features on the hardware such as more ram, nonvolatile option, the
ability to disable my built-in NMI context switch, a nonstandard
registers lock bit and a 'fast paging' register that let the operating
system map any 16K bank out of 32 (512KBytes chip) on address space 0 -
$3FFF, as long as any ROM bank (same size).
We finally agreed to this scheme for the fastpage register:
I/O $54DF, rd/wr
D7 = RAM WR enable (1 = enabled, stand alone)
D6 = fastpaging enable (1 = selected bank IS paged over 0-3FFF)
D5 = ram/rom select: 1 = ram is paged in; 0 = rom is paged in
D4:D0 = 5 bit bank select (32 banks)
The reason for having a separate and standalone bit for WR control is
that, in this way, the program could enable the ram in WR only over the
0-3FFF space, while the previously active ROM is still readable. As a
result, a plain LDIR can be used to copy rom content to the selected ram
bank before activating it in RD mode too.
When Mario said that a lot of operating systems are available for
DivIDE, I was wondering about the hardware resources of this interface
to see if this memory scheme could support them one day.
Ciao!
