[Coco] [Color Computer] Create NitrOS9 L1 boot disk for Dragon with FD-502?

[Phill Harvey-Smith]

Robert Gault wrote:
> drag0n1969 wrote:
>> I need help creating a NitrOS9 Level 1 boot disk for my unique 
>> configuration: Tano Dragon with a Tandy FD-502 disk controller.
>>
>> I have replaced the FD-502's EPROM with a special version of Dragon 
>> SuperDOS, modified by Phill Harvey Smith to work with the Tandy 
>> controller hardware. This works perfectly under basic.
>>
>> Creating a NitrOS9 boot disk is a problem, since the Dragon Dos 
>> expects the boot sector on track 0 (from sector 3), and not track 34 
>> as on the CoCo (though both of them load the boot sector to same 
>> memory address at $2600)
>>
>> Am I correct that my boot disk should simply be a Dragon NitrOS9 boot 
>> disk with the modules BOOT and RB1773 from the CoCo (replacing the 
>> dragon's BOOT and DDISK)?
>>
> Phill should be the expert on this and if he does not see your message 
> here, you should e-mail him directly.

Yep getting the CoCo list mail ok, so continue posting here, that way 
any solutions etc will be in the list archives for others to see if they 
run across the same problems in future :)

> Far as I can tell, the Dragon uses the same disk controller as the Coco. 

In this case yes, it is a standard Tandy contoler with a replacement 
rom. The normal DragonDos controler however uses a WD2797 instead of a 
WD1793/1773. It also uses FIRQ/NMI instead of HALT/NMI.

> It is the disk structure that is changed. What you need to know is if 
> Dragon NitrOS-9 reserves T0 S1-3 for normal OS-9 use or normal Dragon 
> use. If Dragon NitrOS-9 places the boot track at T0 S3-18, then clearly 
> you will need a specialized boot module as well as several other modules 
> to handle the unique disk structure.

Yes the Dragon boot module takes care of this, and along with the, this 
only really causes a problem for DS80 disks as you have to increase the 
sectors/cluster to get it all to fit in the block map :( Toolshead 
however knows about this when asked to format a Dragon floppy, there is 
still however a problem with the native format util, it will however 
make SS or DS 40 disks properly, and also ss80.

> It seems to me that OS-9 on any system ought to use the same disk 
> structure. The Dragon should accommodate this with a special DOS command 
> expecting the boot track to be at T34. But that's just my opinion :)

Well to be honest as far as I can tell the convention of having the boot 
  track on track 35, was largely due to Tandy's choice of 35 track 
drives, it is in no way 'standard OS-9'. Also as far as I can tell other 
OS-9 systems would normally have what is on the boot track in rom, so 
not have this problem. (This is the way the Dragon Beta prototype was).

Also having the boot track on track 35, leads to all sorts of problems 
especially for 80 track disks where the boot is in the middle of the 
disk ! IMHO having it at the beginning where it will be in the same 
place irispective of the number of tracks was much more sensible :)

Having said that I did consider having a small boot program in the 
Dragon boot track that loaded the rest from T34, but ruled this out as I 
would still have to reserve the Dragon boot area and it would still mean 
a slightly different format.

>> If this is a correct, I guess the Coco boot module needs to be patched 
>> to read sector 0?
> That depends on the structure of the Dragon NitrOS-9 disk. Why don't you 
> use a raw disk editor to examine the structure of the Dragon NitrOS-9 
> disk to see exactly what is located in track 0 and track 34?

Track 0 contents (off the top of my memory....)

Purpose		Sector no	Size
LSN0		1		1
Block Map	2		1
Boot area	3		16

The Data area begins at LSN 18

>> I have also tried to "force" the dragon to boot a CoCo bootdisk, using 
>> a small program on the Dragon to read sector 34 from the CoCo boot 
>> disk, poke it into address 9728 ($2600)  onwards, and EXEC 9730. 
>> Doesn't seem to work either…
> That should work excluding other hardware changes between a Coco and a 
> Dragon. I don't think you can assume there are not significant hardware 
> changes which would prevent Coco NitrOS-9 from running on a Dragon.

That should work asuming that you are using the CoCo controler on the 
Dragon, the two machines are pretty similar.

Ok I guess it's time to get back to this, I have updated my copy of the 
Nitros source tree to support the new build architecture, I'll try and 
test it over the next couple of days, and then  do a commit. I can also 
check the Tano build as well.

Would it be helpfull if Mess supported this combo, Tano+RSdos cart + 
SuperDos ? If so I'll hack it back into the Dragon Driver, and see if I 
can get the Mess team to accept it...after all it's more than just me 
using it now :)

Phill.

-- 
Phill Harvey-Smith, Programmer, Hardware hacker, and general eccentric !

"You can twist perceptions, but reality won't budge" -- Rush.