Post by Michael SchwingenPost by Didier MéquignonI know nothing about the Milan, but the TOS404 works on the M5484LITE
Coldfire board so nothing is impossible, if the flash size is 512 KB I
can create a 3rd target.
The Milan has 512KB Flash - part of that is used by bootblock and modular
drivers (eg. for PCI SCSI cards), but since the TOS is compressed, there is
plenty of space left.
I use also an LZ77 compression for the PCI part (above 0xEA0000), I
think also use space between 0xFC0000 and 0xFF0000 for get free space
for USB drivers.
Post by Michael SchwingenPost by Didier MéquignonThe PCI bridge isn't a problem the PLX 9054 is compatible with the PLX
9080, but I need some sources for insert some patchs like serial port,
keyboard and mouse, disks, there are an IKBD on the Milan ? Else it's
impossible without lost time, without hardware to create another target.
Maybe another problem the PCI drivers and TOS patches are GPL.
GPL would definitely be a problem - I doubt that you can legally incorporate
GPL code into TOS on any machine, because TOS is not GPL and not
GPL-compatible. Distributing patches only moves that problem to the user.
However, even without the licensing issues, it would not be that easy: the
Milan uses a heavily modified TOS which integrates code for the Milan
hardware, PCI BIOS, hardware emulation etc. - I doubt any patches against a
binary TOS 4.04 would apply easily.
With sources of specific hardware, it's easy, else without sources and
hardware for test it, it's just impossible. I have worked on M5484LITE
board patches, on month before get the card, and 2 months for get a TOS
working with the Radeon.
Post by Michael SchwingenHowever, i do not see why TOS patches should be needed at all: the graphics
driver can be loaded from the auto folder (or as a pre-auto flash resident
module on the Milan). That only requires basic standard VGA init code in the
bootblock - does ATI provide documentation on that part that is usually
handled by the X86 BIOS ROM?
I don't know, actually the X86EMU used works fine for start the VGA BIOS
(91KB), but I have tried INT 10 without success (black screen), it's not
important for me because the init of the card is finished and the radeon
driver works fine (256 / 65K / 16M), the driver start before the
Setscreen 1st call. The VDI rebuild in assembly language in ROM is also
faster than fVDI because direct calls are used for replace generic fVDI
calls, something than 10 times faster on blitting and text functions.
And the BIOS part is not inside the FVDI.SYS drivers, so it's impossible
to use new Setscreen modes, and other functions added like VIDIX
(actually fulscreen with the scaler for Aniplayer with a YUV or RVB
source), or TV port on the Radeon AIW (this lastest part isn't
finished). Sure it's always possible to work on a TSR program started
before fVDI.
Regards,
Didier.
--
Didier MEQUIGNON Aniplayer download: http://aniplay.atari.org
CT60 package download: http://ct60conf.atari.org
CTPCI: http://ctpci.atari.org