FWSI Analogic CD32 Diskdrive issue

[Trigger58]

Originally Posted by alexh

And you've tried

loadmodule dh1:/fwsi/low_level_fwsi.library

But it says the two modules are identical?

No. I didn't tried that. I thought the names have to be identical to work. If that doesn't matter then i will give it another try this evening. No time to test it yet.

[alexh]

Originally Posted by Trigger58

I thought the names have to be identical to work.

I assume the library name is part of the binary and the filename is irrelevant.

[Keir]

Originally Posted by alexh

I assume the library name is part of the binary and the filename is irrelevant.

Yes, there's a name string embedded in the library.

[Trigger58]

Hello,
I haven't tested the new version with the higher version number and LoadModule in the startup sequence yet. But I have done a few tests with the new lowlevel.library.

The following equipment was used for the tests:
- Amiga CD32
- HoneyBee Joypad
- Sum 4000/CD32 keyboard adapter with wired Acer USB keyboard
- wireless mouse with DB9 wireless adapter (from the Kickstarter campaign)
- fWSI Wallstreet module with modified Gotek drive (converted from Amiga to PC drive) and Sandisk USB stick
- TF330 turbo card with 64MB RAM (MMU active, caches active)
- 32GB Sandisk Extreme Compact Flash card with CF adapter connected directly to the TF330 without a cable
- CF divided into 4 partitions: DH0-DH3
DH0 = 4 GB with game boot menu
DH1 = 305 MB with OS3.1
DH2 = 305 MB as data
DH3 = approx. 3300 MB for separate WHDLoad games

Main boot partition is DH0.

Here is the startup sequence of the boot partition:

assign root: ""
assign env: ram: >nil:
assign t: ram: >nil:
assign nvsaves: nv_saves >nil:
assign envarc: prefs/env-archive >nil:

Assign >NIL: RAD: EXISTS
IF not warn

	if exists rad:tempboot
		execute rad:tempboot >nil:	
		quit
	endif

endif

setpatch quiet
c:FastRAD RAD:

if not exists devs/Kickstarts/kick34005.a500
	echo kick34005.a500 not found in devs\Kickstarts
	wait 1
endif
if not exists devs/Kickstarts/kick40068.a1200
	echo kick40068.a1200 not found in devs\Kickstarts
	wait 1
endif

vol1:System/setmap d

menu.exe  > env:game
cd "games/$game"


if exists runme
	execute runme
else
	whdload "$game.slave" preload
endif

reboot

I used 2 games for tests: Flink & Cecconoid

Beginning with the game FLINK:

Test scenario 1:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot.
Loading game from cd.

Result:
Games works perfectly with correct buttons.


Test scenario 2:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot. Loading game from CF card, WHDLoad installation.

Result:
Game don't work correctly. Buttons still all wrong.


Test scenario 3:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot.
First, all files from the CD were copied into a games folder on a CF card. These were then loaded into a RAD disk and started from there.

Result:
Game works perfect. Joypad buttons all ok.


Test scenario 4:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot.
WHDload version was loaded into the RAD disk including all necessary files like kickrom files. Then the game was started from RAD disk.

Result:
The same problem as starting directly from the CF card. The controls don't work. The joypad buttons are wrong.



Next test game is Cecconoid (there you can control the shooting direction with all 4 buttons). I tested the function of the CD32 joypad in advance in WinUAE. There, shooting works perfectly using all 4 buttons and different directions.

Test scenario 1:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot. Loading game from CF card, WHDLoad installation.

Result:
Game don't work correctly. Buttons still all wrong.


Test scenario 2:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot.
WHDload version was loaded into the RAD disk. Then the game was started from RAD disk.

Result:
The same problem as starting directly from the CF card. The controls don't work. The joypad buttons are wrong.


Test scenario 3:
Boot the lowlevel.library from diskette via the Gotek. The library is loaded with LoadModule and then a reboot occurs. The diskette is removed after the reboot.
Booting the game from Gotek as ADF File (game is a NDOS disk).

Result:
The controls don't work. The joypad buttons are wrong.



My conclusion so far:

Basically, the patch seems to work with the new lowlevel.library. At least with the original CD32 titles.

The patch didn't work for games that were loaded via WHDLoad.

If everything goes well, I'll receive a new patched ROM this week. I'll then repeat all the tests and see if anything has changed in the behavior of the games.

[Trigger58]

Today i got the new EPROM with the lowlevel.library patch in it.

@Keir
Sorry to say but the patch works only for games on CD or games which were loaded in RAD disk and aro NO whdload games.

ALL games loaded by WHDload, JST or other tools the patch don't work. It doesn't matter if i load them into RAD or not. All games loaded from a hard drive don't work

I tested this with Flink, Cecconoid and Chuck Rock.

I think WHDLoad is the problem here. It does something different as a normal loaded game from CD.

[alexh]

WHDload and the games running inside probably do not use the OS and so don't use lowlevel.library but instead hit the CIA HW directly and so of course won't work even with a new ROM.

Each slave may need patching to use the alternate CIA slave addresses.

There is a slim chance that games that DO use lowlevel.library in WHDLoad but are provided with an emulated Kickstart (KickEmu) can be patched with a new WHDload release.

[Trigger58]

Originally Posted by alexh

WHDload and the games running inside probably do not use the OS and so don't use lowlevel.library but instead hit the CIA HW directly and so of course won't work even with a new ROM.

Each slave may need patching to use the alternate CIA slave addresses.

There is a slim chance that games that DO use lowlevel.library in WHDLoad but are provided with an emulated Kickstart (KickEmu) can be patched with a new WHDload release.

Yeah, that's what i thought. I tested many games in the last 2 hours and ALL games using WHDLoad don't work with the patch.



It doesn't matter whether you load the games from CD, RAD or hard drive. As soon as WHDLoad is in the game, the CD32 joypad no longer works properly. That's really rubbish.

[alexh]

Originally Posted by Trigger58

As soon as WHDLoad is in the game, the CD32 joypad no longer works properly. That's really rubbish.

Blame Mick Tinker, the guy who designed and made the FWSI and it's fake CIAs. Nothing you could do now other than reverse engineer the FWSI CPLD and reprogram it.

If you talk to WEPL there is a slim chance he can make you a custom build of WHDload for any OS legal games... otherwise you'll have to ask individual slave authors if they will make you a custom slave build.

[Trigger58]

Originally Posted by alexh

Blame Mick Tinker, the guy who designed and made the FWSI and it's fake CIAs. Nothing you could do now other than reverse engineer the FWSI CPLD and reprogram it.

If you talk to WEPL there is a slim chance he can make you a custom build of WHDload for any OS legal games... otherwise you'll have to ask individual slave authors if they will make you a custom slave build.

I have no skills in programming or reverse engineering


Thank you for your help. I wrote Wepl a pn. I hope he can help us

[Wepl]

Originally Posted by alexh

WHDload and the games running inside probably do not use the OS and so don't use lowlevel.library but instead hit the CIA HW directly and so of course won't work even with a new ROM.

Each slave may need patching to use the alternate CIA slave addresses.

There is a slim chance that games that DO use lowlevel.library in WHDLoad but are provided with an emulated Kickstart (KickEmu) can be patched with a new WHDload release.

Yes, nothing outside (OS, Kick) is used for programs running under WHDLoad.
Some installs use a file base lowlevel.library. These can be fixed manually by replacing the lowlevel.library contained in the install data directory with a modified one.
For all other installs the code to check the pad is contained in the slave or the game binary data. These installs need to be fixed individually all. Recently added is a function inside WHDLoad v19 to read the controllers. This one could be modified. But currently there are no installs using it.

[Trigger58]

Originally Posted by Wepl

Yes, nothing outside (OS, Kick) is used for programs running under WHDLoad.
Some installs use a file base lowlevel.library. These can be fixed manually by replacing the lowlevel.library contained in the install data directory with a modified one.
For all other installs the code to check the pad is contained in the slave or the game binary data. These installs need to be fixed individually all. Recently added is a function inside WHDLoad v19 to read the controllers. This one could be modified. But currently there are no installs using it.

Thanks for the explanation.

I was afraid that it wouldn't be so easy to do. The only thing left now is reverse engineering the chip on the fWSI.

I would be willing to donate €100 for it.

I'm not the only one with such a module. Maybe we can get a few more together. Then the fWSI module would finally be fully compatible. The brilliant thing about the module is the second connection that is normally used to plug in a real-time clock module. But a TF330 also works with it. This would give you a cost-effective alternative to an SX1 or an SX32. That's why I bought it.

[alexh]

Originally Posted by Trigger58

The only thing left now is reverse engineering the chip on the fWSI. I would be willing to donate €100 for it.

There are 3 chips on the fWSI but I imagine examining the schematic / photos someone could work out which is which... maybe.

Originally Posted by Trigger58

I'm not the only one with such a module.

I've owned one for 20+ years.

Originally Posted by Trigger58

The brilliant thing about the module is the second connection that is normally used to plug in a real-time clock module. But a TF330 also works with it.

Hmmm I don't think that is true. I don't think the TF328, TF330 or TF360 works with the fWSI as the "riser". If I remember correctly the floppy disk drive logic drives the signal /PUNT low instead of high impedance when not in use which breaks anything in 2nd slot that would use it. A simple RAM card might work. But not an accelerator/IDE controller.

Originally Posted by Trigger58

This would give you a cost-effective alternative to an SX1 or an SX32. That's why I bought it.

That is why I bought mine too but I was able to bag an SX32 MK1 a few weeks later and it's sat in the box next to my desk ever since.

[Trigger58]

Originally Posted by alexh

There are 3 chips on the fWSI but I imagine examining the schematic / photos someone could work out which is which... maybe.


I've owned one for 20+ years.


Hmmm I don't think that is true. I don't think the TF328, TF330 or TF360 works with the fWSI as the "riser". If I remember correctly the floppy disk drive logic drives the signal /PUNT low instead of high impedance when not in use which breaks anything in 2nd slot that would use it. A simple RAM card might work. But not an accelerator/IDE controller.


That is why I bought mine too but I was able to bag an SX32 MK1 a few weeks later and it's sat in the box next to my desk ever since.

I can say a TF330 works fine with it

I use this combo by myself.

[Trigger58]

Originally Posted by alexh

There are 3 chips on the fWSI but I imagine examining the schematic / photos someone could work out which is which... maybe.

I will make a photo and post it here. A manual doesn't exist or better i never found one. I asked in different forums and searched the whole net. No manual, no service manual, really no documentations to find.

[alexh]

Realistically not enough people care so it's extremely unlikely to get done unless you do a lot of the easy but time consuming work.