[Request] Real machine test

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 4:16 am

Can someone test my ROM in a real gen/md?

here!

If you have a male plug on the ext port please test it (with a controller)

Thanks

HardWareMan · Post by **HardWareMan** » Sat Jan 24, 2009 1:14 pm

Don't work on my MD1 (board IC BD M5 PAL VA4). Seems like you don't do proper init of security system. Or something like that.

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 1:22 pm

Bah! Or maybe it's hanging on the Z80 loading...

HardWareMan · Post by **HardWareMan** » Sat Jan 24, 2009 1:24 pm

Maybe, but even on emulator sometimes it generate an exception.

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 1:29 pm

Ops I missed that! I just tested it in fusion and gens.

Ok seems that the joypad read has a problem, it gives different results in gens/regen/fusion

(accidental emulator detection)

BTW how did you trigger the exception?

AamirM · Post by **AamirM** » Sat Jan 24, 2009 2:25 pm

I pressed 'A' then 'Start' and got to that screen HardWareMan posted.

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 2:28 pm

Reposted... Please test (again

)

Well option A has a yet unknown problem.

When A is pressed:

In Regen:
An exception occurs (probably address error), the screen is grey. But when start is pressed the program is resetted and something goes wrong.

In Fusion:
The screen deforms (plane dimensions changed) but the mappings are the same.

In Gens:
A gradient from black to red (should be blue) appears, but scrolling it makes it weird green and flicker (raster bugs)

TmEE co.(TM) · Post by **TmEE co.(TM)** » Sat Jan 24, 2009 6:09 pm

my TMSS bypassed MD2 gave white screen which means the VDP is not initialized properly or you trigger an DMA during reg init and it crashes the system.

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 6:38 pm

Well the code is started with a move #$2700 to sr, so that accidental dma is not started.

I think that "Clr.w VDPdataport" is not allowed, yes?
And "Clr.xx AnyRamLocation" ?

Now the Address error issue (HBL buggy code) in regen is fixed, and the 3 emus show the same gradient.

BIN reuploaded

TmEE co.(TM) · Post by **TmEE co.(TM)** » Sat Jan 24, 2009 8:23 pm

CLR, BTST and some other instructions cannot be used with VDP

EDIT: Latest greatest still shows white screen, and $2700 in SR will not prevent DMA's from happening, put a $80 into last DMA reg, if you don't have it there.

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 9:05 pm

Well I though you were referring to Vblank DMAs, before the code inits properly. That's why the move to sr is right at the start.

But now I'm doing a "MOVE.w #$8100, VDPControlport" after deactivating HBL/VBL. This should end with the accidental DMAs...

Re-Reuploaded

I'm getting out of ideas

TmEE co.(TM) · Post by **TmEE co.(TM)** » Sat Jan 24, 2009 9:08 pm

Now it gives a black screen instead of white

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 9:12 pm

At least it DMA filled Vram with 0s...

Going to check if more clears/tests are done on the vdp ports...

OPS! I am...

Done! Removed every clear to the vdp ports. there weren't any tsts/btsts.

Re-Re-Reuploaded

Graz · Post by **Graz** » Sat Jan 24, 2009 10:04 pm

Jorge Nuno wrote:Well I though you were referring to Vblank DMAs, before the code inits properly. That's why the move to sr is right at the start.

But now I'm doing a "MOVE.w #$8100, VDPControlport" after deactivating HBL/VBL. This should end with the accidental DMAs...

I think you may be confusing DMA and interrupts. Both of the measures you've described disable interrupts and have nothing to do with DMA.

Jorge Nuno · Post by **Jorge Nuno** » Sat Jan 24, 2009 10:10 pm

I'm not

It's that I only do DMA copies in VBlank. And clearing register $81 should disable all kinds of DMA, unless there is some obscure bug.