VisualBoyAdvance-M

That's not true. On the 22nd October, 2008 I first discovered that issue.

 

Yesterday, I started the game from scratch (without any save states or battery saves) because the save state here was corrupted, and I noticed that freeze on exactly the same position and under the same conditions, so it is definitely not related to a single save state.

Then something is wrong with your VBA-M configuration.

 

have you audio synching enabled?

of course

 

and vsync

 

 

 

I just found out, that when the game freezes while the boss created a sound effect by one of his attacks, that sound effect will also repeat besides the BGM.

 

 

Oh, and now the freeze appeared even earlier.

 

I was recording this now as a video, so I'll upload that on YouTube for ya.

http://www.youtube.com/watch?v=XYZdIMEep0E

I think its still definitely something on your end dude [img]<fileStore.core_Emoticons>/emoticons/huh.png[/img]/emoticons/huh@2x.png 2x" width="20" height="20" />

I can't reproduce these issues what so ever.

 

Post up your ini file.

Illegal Reads throughout most of the level.

Illegal Writes during the explosion of the final boss.

 

Biggest issue however, would be the Audio buffers are not filling fast enough. This could lead to a freeze.....

I attached my ini.

 

But you are right with the illegal reads.

 

I get endless of these in the intro level:

Illegal halfword read: 0400004c at 080214a6
Illegal halfword read: 0400004c at 08020572

Illegal halfword read: 0400004c at 080212aa

 

GBATEK says "400004Ch - MOSAIC - Mosaic Size (W)"

So it can be written only, not read.

Xaudio core has more buffer refill issues then openal... but still no freezes.

 

I really think you have a bad dump spacy :\.

 

try my ini.

hm, CRC in ROM information is correct, though. Not sure if is just for the ROM header though.

 

 

Looking at the disassembler, it looks like the game is jumping to all kinds of funny code areas. One of these is around 2dd0360c. "0000 => LSL r0, r0, 0x00". This looks like a NOP to me [img]<fileStore.core_Emoticons>/emoticons/laugh.png[/img]/emoticons/laugh@2x.png 2x" width="20" height="20" />

 

Oh, and when I look up that adress, it's already out of any memory area.

 

 

Edit: Your ini does not help either

 

 

Edit2:

I start to understand the issue.

 

The game locks up

BECAUSE it is emulating an endless amount of NOPs

BECAUSE the program counter is out of any valid memory area

BECAUSE it was sent there by the opcode at the adress in R14 (0801E091)

BECAUSE there is something wrong with an opcode being at an uneven adress.

 

More to follow [img]<fileStore.core_Emoticons>/emoticons/laugh.png[/img]/emoticons/laugh@2x.png 2x" width="20" height="20" />

 

Now I need breakpoints to stop at the right time... hm

 

 

Edit3:

Btw, my ROM's CRC is 57930C6A

 

 

Edit4:

On a second ROM, the program jumped to another bad place somewhere at the beginning of the memory map.

 

Edit5:

This bug also occurs on the US version, BIOS enabled or disabled does not matter. http://spacy51.sp.funpic.de/VBA-M/crash_saves/Iridion%20II%20(USA)2.7z . press start to return from pause and then press A button to activate bug. save state needs BIOS and has cheats enabled, but that does not matter, because the other save state does not have both.

 

It is also quite unprobable that two ROMs from the no-intro dat-set are bad dumps.

 

Edit6:

Oh, now I played it on No$GBA and it also tells me "The rom-image has crashed"

 

Edit7:

Found someone else with the same problem on almost REAL HARDWARE:

http://gbatemp.net/index.php?showtopic=76956

Well, it seems to be related to wrong dumping

 

Edit8:

The bug was already noticed two years ago:

http://forums.ngemu.com/vba-m-discussion/76942-developers-wanted-6.html

Well it seems i have a US dump, rather then E

 

CRC: 1A992D7C

It is also quite unprobable that two ROMs from the no-intro dat-set are bad dumps.

 

no, its actually not improbable. and since im not getting any crashes, its actually very probable.