Dissidia: Final Fantasy - ULUS10437

[Karen]

the game is working pretty fine in the new revision, the only problem is the same that happens in all the others the msg that "blablabla compiler has been disabled" but nothing happens and the game keeps working fine....dunno what it means

[awesomeness]

Ok, I used r1964 and booted the game, used the save from the old (v0.5 r1697) revision and then started quick match.

What's the problem again? I don't quite understand.

The problem is that after the loading screen (after picking a battle) nothing will show up, it just freezes, although the music still plays.

[Itaru]

Starting with r1955, I'm getting graphical corruption (screen capture included in attachment) at the beginning of an EX Burst move in the tutorial battle. Also, when the guy says, "Intervene and be slain!" his mouth doesn't move. In r1954, his mouth moves during that line and also there's no corruption at the beginning of the EX Burst move.

The regression is due to the changes made in jpcsp.Allegrex.compiler.nativeCode.NativeCodeSequence and/or jpcsp.Allegrex.compiler.nativeCode.NativeCodeManager. From r1955, if I revert just those 2 files to r1954 and also revert the method call in jpcsp.Allegrex.compiler.CompilerContext, the game runs fine without the missing mouth movements and graphical corruption.

The regression still occurs in r1964 BTW.

UPDATE
Okay, I've found the problem. In r1964, class jpcsp.Allegrex.compiler.nativeCode.NativeCodeManager is missing a check condition in line 298. I think it should be:

if (opcode == 0 || !nativeCodeSequence.isMatching(i, opcode)) {

This fixes the missing mouth movements and graphical corruption.

[jm0069]

It can't run encrypted games.. WHY T______T

[serio]

r1965 can. i managed to run all of mine anyway, though some games might be still problematic.

[Kai3213]

It can't run encrypted games.. WHY T______T

Because they have encrypted eboots thats why. You need to decrypt them with your psp before you can run it on jpcsp.

[serio]

not anymore, since r1965 jpcsp decrypts the eboot files by itself (well 1964, but it had a prx decryption bug that prevented most of the games from loading).

[Itaru]

Encrypted games have tags in their headers which tells the PSP which encryption key and method are used so that the PSP is able to decrypt them using the correct key and method based on the tag. The problem is that this game as well as a couple of others that use the tag "C0CB167C" cannot be decrypted by Jpcsp since the method of encryption/mangling for that particular tag is not known yet AFAIK. The only way to decrypt such games with that particular tag is by using a real PSP since its firmware has the necessary function to do the proper decryption. Without knowing the algorithm used to encrypt/mangle those games with the "C0CB167C" tag, there's no way to emulate the proper decryption in Jpcsp even if all the keys are known.

I hope I'm wrong about this though, and that the dev team actually already knows the algorithm to decrypt games with the "C0CB167C" tag but they just haven't gotten around to implementing it in Jpcsp yet.

[gid15]

Starting with r1955, I'm getting graphical corruption (screen capture included in attachment) at the beginning of an EX Burst move in the tutorial battle. Also, when the guy says, "Intervene and be slain!" his mouth doesn't move. In r1954, his mouth moves during that line and also there's no corruption at the beginning of the EX Burst move.

The regression is due to the changes made in jpcsp.Allegrex.compiler.nativeCode.NativeCodeSequence and/or jpcsp.Allegrex.compiler.nativeCode.NativeCodeManager. From r1955, if I revert just those 2 files to r1954 and also revert the method call in jpcsp.Allegrex.compiler.CompilerContext, the game runs fine without the missing mouth movements and graphical corruption.

The regression still occurs in r1964 BTW.

UPDATE
Okay, I've found the problem. In r1964, class jpcsp.Allegrex.compiler.nativeCode.NativeCodeManager is missing a check condition in line 298. I think it should be:

if (opcode == 0 || !nativeCodeSequence.isMatching(i, opcode)) {

This fixes the missing mouth movements and graphical corruption.

Thanks!
The real problem has been fixed in r1968 (was located in jpcsp.Allegrex.compiler.nativeCode.graphics.sceGu.sceGuCopyImage)

[str8mjed]

not anymore, since r1965 jpcsp decrypts the eboot files by itself (well 1964, but it had a prx decryption bug that prevented most of the games from loading).

hmm.. the JPCSP r1969 doesn't load my Encrypted Dissidia....