Page 1 of 2

Bugs ?

Posted: Wed Nov 29, 2006 12:06 pm
by KanedaFr
Be free to add here the bug you got with KMod

Please follow this template


Tool & version : ex GensKMod 0.5
Windows version : ex XP SP 2 fr (include language)
Country settings : ex Japan
CD Bios used for this country : ex Mega CD BIOS V2.00 (E).bin or jp_mcd1_9111.bin (if needed)
Others config : ex Spy activated
Description : ex bad colors

I'll try to resolve them ;)
Please don't ask for new features, it's only for BUG

Kaneda

Posted: Tue Dec 12, 2006 10:35 am
by TmEE co.(TM)
Gens KMod version : 0.5
Windows version : Win98SE 4.10.2222a, with unofficial SP 10, english
Country settings : J,U,E
Description : instant crash if Alien3 is loaded (This game does DAC and 6th FM ch at the same time, maybe that's why).

Posted: Sat Jan 27, 2007 10:57 pm
by GrayLight
Tool & version : GensKMod 0.6
Windows version : XP SP 2 ru
Country settings : J,U,E
Description : Debug->Layers "Show Layer B" checkbox switches ALL layers (even Sprites), uncheck makes all invisible or, sometimes, leaves some "garbage"..
p.s. Sorry, if my english is not well... )

Posted: Tue Jan 30, 2007 8:52 am
by KanedaFr
GrayLight wrote:Tool & version : GensKMod 0.6
Windows version : XP SP 2 ru
Country settings : J,U,E
Description : Debug->Layers "Show Layer B" checkbox switches ALL layers (even Sprites), uncheck makes all invisible or, sometimes, leaves some "garbage"..
p.s. Sorry, if my english is not well... )
someone can confirm this ???
do you have this problem on a special game or all the games ?



ps: no problem for the english, I use the same ;)

Posted: Tue Jan 30, 2007 5:29 pm
by GrayLight
I've used this roms:
Phantasy_Star_II_(UE)_(REV_02)_[!]
Landstalker_-_The_Treasures_of_King_Nole_(U)_[!]

..With KMod_05 layers debug works fine..

Posted: Sun Feb 04, 2007 9:32 pm
by KanedaFr
it seems the way I handled shadow broke some stuff :(
sorry for that, I'll fix it soon

Posted: Mon Feb 12, 2007 6:29 pm
by TmEE co.(TM)
KMod 0.6 VDP tile scroller shows incorrect address

Posted: Mon Feb 12, 2007 7:36 pm
by Stef
Tool & version : GensKMod 0.6 beta (with DAC update rate feature)
Windows version : XP SP 2 fr
Country settings : US
Description : keys respond problems -->
When i open some debug window (as YM2612 one), it looks that after sometime GensKMod doesn't reply to key events and seems somewhat freezed (but it isn't).

By the way, your "DAC update rate" is very usefull when you code some Z80 driver for DAC playing, i encourage you to keep this feature in future version ;)

Posted: Tue Feb 13, 2007 8:41 am
by KanedaFr
TmEE co.(TM) wrote:KMod 0.6 VDP tile scroller shows incorrect address
are you sure ?
I fixed it because it was already wrong so I don't want to come back to my first version...

Posted: Tue Feb 13, 2007 8:42 am
by KanedaFr
Stef wrote:Tool & version : GensKMod 0.6 beta (with DAC update rate feature)
Windows version : XP SP 2 fr
Country settings : US
Description : keys respond problems -->
When i open some debug window (as YM2612 one), it looks that after sometime GensKMod doesn't reply to key events and seems somewhat freezed (but it isn't).
when you activate a window, the key events aren't send to Gens main window...
are you sure it's not that ? I ask because I often made the mistake ;)
By the way, your "DAC update rate" is very usefull when you code some Z80 driver for DAC playing, i encourage you to keep this feature in future version ;)
it's still here ;)

Posted: Tue Feb 13, 2007 1:28 pm
by TmEE co.(TM)
KanedaFr wrote:
TmEE co.(TM) wrote:KMod 0.6 VDP tile scroller shows incorrect address
are you sure ?
I fixed it because it was already wrong so I don't want to come back to my first version...
I scrolled down to the last row and the address was 5D0

Posted: Tue Feb 13, 2007 2:59 pm
by Stef
KanedaFr wrote:
Stef wrote:Tool & version : GensKMod 0.6 beta (with DAC update rate feature)
Windows version : XP SP 2 fr
Country settings : US
Description : keys respond problems -->
When i open some debug window (as YM2612 one), it looks that after sometime GensKMod doesn't reply to key events and seems somewhat freezed (but it isn't).
when you activate a window, the key events aren't send to Gens main window...
are you sure it's not that ? I ask because I often made the mistake ;)
Yeah i know but even wit focus on the Gens main window, keys aren't always handled...

By the way, your "DAC update rate" is very usefull when you code some Z80 driver for DAC playing, i encourage you to keep this feature in future version ;)
it's still here ;)
good new :D

Posted: Thu Feb 22, 2007 2:56 pm
by KanedaFr
Stef wrote: Yeah i know but even wit focus on the Gens main window, keys aren't always handled...
I'm unable to make this bug...
perhaps it's the latency....I got key respond problem when some debug windows are opened : I need to keep my keypress ;)
sorry for that but as I said since the first version, KMod isn'"t suitable for games play :)
I scrolled down to the last row and the address was 5D0
VRAM idx is now the tile number not the VRAM adress...
it was requested so I changed it
I'll give it another look....

Posted: Thu Feb 22, 2007 3:06 pm
by Stef
KanedaFr wrote:
Stef wrote: Yeah i know but even wit focus on the Gens main window, keys aren't always handled...
I'm unable to make this bug...
perhaps it's the latency....I got key respond problem when some debug windows are opened : I need to keep my keypress ;)
sorry for that but as I said since the first version, KMod isn'"t suitable for games play :)
Thanks for taking time to examine it. I was using key just for testing purpose so that's not so problematic ;)

Posted: Fri Dec 28, 2007 7:00 am
by HardWareMan
Emulator: see below.
Windows: does not matter.
Here is the problem:
1: GENS's debugger wrong decodes ORI.W and some other opcodes.
2: I found a bug in emulation M68K in MOVE.L (A0)+,D1. First time it's work fine, after some calculations it's read garbage in D1, but A0 is still correct. MOVE.B (A0)+,D1 works perfectly. Here source, that not working:

Code: Select all

           move.l      #$FF0000,d2     *
           move.l      d2,a1           *
           subq.l      #1,d0           *
BuildLn:   clr.l       d1
           move.l      (a0)+,d1        * << Here is buggy command
           move.w      #39,d2          *
BuildLp:   move.b      (a0),d3         *
           cmp.b       #0,d3           *
           beq         BuildSkp        *
           move.b      (a0),(a1)+      *
           addq.l      #1,a0           *
           bra         BuildFin        *
BuildSkp:  move.b      #$20,(a1)+      *
BuildFin:  dbra        d2,BuildLp      *
BuildZer:  move.b      (a0),d3         *
           cmp.b       #0,d3           *
           beq         BuildNxt        *
           addq.l      #1,a0           *
           bra         BuildZer        *
BuildNxt:  addq        #1,a0           *
           move.l      d1,(a1)         *
           add.l       #24,a1          *
           dbra        d0,BuildLn      *
If I replace that command this:

Code: Select all

           move.b      (a0)+,d1        *
           rol.l       #8,d1           *
           move.b      (a0)+,d1        *
           rol.l       #8,d1           *
           move.b      (a0)+,d1        *
           rol.l       #8,d1           *
           move.b      (a0)+,d1        *
All works perfectly. I'd made some tests and here is result:
Test passed emulators:
DGen v1.21
Fusion v3.51
Megasis v0.06a
Test not passed emulators:
Gens32 Surreal v1.7
Gens32 (Gens v2.10)
Gens32 V2.26
Gens v2.12a KMod 0.7b
Gens v2.11
Gens+ v0.0.3.38
GensD (Gens v2.10)
So, it seems this is "Gens" only problem in M68K emulation code.
Test ROM you can get here.
PS. That command loads melody pointer. If it buggy, then melody can not be played correctly.
PPS. Can you do something with that terrible 16bit screen mode? That screen flashes is killing me... (Gens32 use 32bit screen mode) :(