Yu-Gi-Oh GX Tag Force 3 and 5

[Ekaseo]

The game works almost perfect now, but there is one big problem, you cant travel between instances, it shows a black screen and music still runs, but nothing loads:
http://www.fastimages.eu/images/ygogxtf31.jpg
http://www.fastimages.eu/images/ygogxtf32.jpg
http://www.fastimages.eu/images/ygogxtf33.jpg
http://www.fastimages.eu/images/ygogxtf34.jpg

[hrydgard]

Got logs when that happen?

[Ekaseo]

im not sure if i done it right but here it is:
https://docs.google.com/file/d/0B1sguG1C1mNAS1VMQm0wRWtmdkU/edit?usp=sharing
(uploaded it 3rd time... this time ther should not be unnecesary stuff)

[Ekaseo]

what is wrong with it? it loads normaly to the map, but then it wont go on

[kahaya]

http://pastebin.com/QNKEZPbv

Here's another one, should I open a new issue for Tag Force 5?
Because 5 doesn't even go into the Story Mode, in this game everything works but the changing between places.

[sliferdragon]

Issue still here after 9.2

[unknownbrackets]

Among the log reports, I noticed a lot of "bad font" errors:
http://report.ppsspp.org/logs/game/ULJM05940_1.01

Does anyone get these in the console while playing these games? From the screenshots, it seems like at least some fonts work.

-[Unknown]

[thedax]

So I took a peek in the debugger for this game, and here's what I see as far as threads go, when the game hangs after selecting a location to travel to:

Does anything look out of the ordinary (serious question)?

[unknownbrackets]

Hard to tell. The user_main one is the only one that can effectively do anything there (aside from interrupts.)

What happens if you right click and force wakeup on the ones after user_main (try one at a time)? Also, if you double click user_main, and set a breakpoint after the wait, what HLE does it call - anything that might wake those threads up?

-[Unknown]

[hrydgard]

Hm, I looked at logs from yu gi oh tag force 5D a while ago which hangs the same way, saw async I/O stuff in there right before the hang... anything like that going on?

[thedax]

No, it doesn't seem to do any async stuff. I turned the log level up to verbose and unchecked everything except sceIO.

Resuming the threads one by one doesn't seem to help either. They end up back in the same state they're in as depicted in the screenshot.

[hrydgard]

Here's a piece of strongly filtered Yu Gi Oh 5 log at the hang:

24:13:621 idle0        D[CPU]: HLE\sceGe.cpp:111 Entering interrupt handler 08807bd0
24:13:621 user_main    D[CTRL]: HLE\sceCtrl.cpp:482 1=sceCtrlPeekBufferPositive(0bfff638, 1)
24:13:636 ehscript     D[IO]: HLE\sceIo.cpp:653 sceIoGetstat(ms0:/PSP/SAVEDATA/ULES01474DAT/FILE467.DAT, 0bfe8d88) : FILE NOT FOUND
24:13:638 ehscript     D[IO]: HLE\sceIo.cpp:653 sceIoGetstat(ms0:/PSP/SAVEDATA/ULES01474DAT/FILE467.DAT, 0bfe7948) : FILE NOT FOUND
24:13:639 ehscript     D[IO]: HLE\sceUmd.cpp:302 0x32=sceUmdGetDriveStat()
24:13:639 ehscript     D[FileSys]: FileSystems\ISOFileSystem.cpp:395 Got a raw sector open: /sce_lbn0x7fca0_size0x403f, sector 0007fca0, size 0000403f

24:13:640 ehscript     D[IO]: HLE\sceIo.cpp:1162 4=sceIoOpen(disc0:/sce_lbn0x7fca0_size0x403f, 00000001, 00000000)
24:13:640 ehscript     D[IO]: HLE\sceUmd.cpp:302 0x32=sceUmdGetDriveStat()
24:13:640 ehscript     D[IO]: HLE\sceIo.cpp:783 sceIoRead(4, 09ed1080, 403f): deferring result
24:13:640 ehscript     D[IO]: HLE\sceIo.cpp:1170 sceIoClose(4)
24:13:663 idle0        D[CPU]: HLE\sceGe.cpp:111 Entering interrupt handler 08807bd0
24:13:663 user_main    D[CTRL]: HLE\sceCtrl.cpp:482 1=sceCtrlPeekBufferPositive(0bfff638, 1)

deferring result and then a close? Hm right, that's just i/o threading.

[thedax]

Also, if you double click user_main, and set a breakpoint after the wait, what HLE does it call - anything that might wake those threads up?

Sorry, I missed this previously: Tag Force 3's user_main seems to pause on a sceDisplayGetAccumulatedHcount count, and then the next HLE it calls after a while is sceDisplaySetFramebuf.

[thedax]

Since we don't seem to have a full debug log of tag force 3 posted, here's one. It expands to 15 mbytes.

[unknownbrackets]

Maybe sceAtracGetLoopStatus(), or sceSasGetEndFlag()/sceSasGetEnvelopeHeight() (if it's not supposed to have ended yet or something), or maybe sceDisplayGetAccumulatedHcount() although I think that mostly works now...

-[Unknown]

[thedax]

Any suggestions on tweaks? I'm open tonight for testing things.

[unknownbrackets]

Sorry, not really sure. I mean, I guess you could try hardcoding return values for any of those, but I'm not really sure what it could be looking for.

-[Unknown]

[thedax]

I guess the logical place to start would be with a JPCSPTrace of those 3 funcs to see if anything is out of the ordinary. I'll grab a trace soon.

Edit: Okay, here's the JPCSPTrace log.

[MohamedElayadi]

I Have the same problem with all yu-gi-oh games on psp

[thedax]

Something I just noticed: Is it normal for a thread to be created, started, and returned from so quickly?

17:30:040              I[MODULE]: HLE\sceKernelModule.cpp:1459 334=sceKernelLoadModule(name=disc0:/sce_lbn0x72f50_size0x3dc74,flag=00000000,(...))
17:30:040              I[MODULE]: HLE\sceKernelModule.cpp:1504 sceKernelStartModule(334,asize=00000004,aptr=0bfe91ac,retptr=089248bc,00000000)
17:30:040              I[KERNEL]: HLE\sceKernelThread.cpp:2110 335=sceKernelCreateThread(name=modfield, entry=090be300, prio=20, stacksize=262144)
17:30:040              I[KERNEL]: HLE\sceKernelThread.cpp:2161 sceKernelStartThread(thread=335, argSize=4, argPtr=0bfe91ac)
17:30:040 modfield     I[KERNEL]: HLE\sceKernelThread.cpp:2241 __KernelReturnFromThread: 0

This is an excerpt from Tag Force 3, when exiting to the world map.

[unknownbrackets]

Well, probably not, but it could be. Can't say without knowing what the thread does. Following a breakpoint at 090be300 might help.

-[Unknown]

[thedax]

I'll check soon.

Edit: It doesn't seem to do anything interesting, so I don't think that thread is the problem.

[thedax]

PPSSPP seems to open some LBNs that JPCSP doesn't via sceIoOpen..

[dbz400]

For yu gi oh tag force 5D , it stopped after this screen

and this one should be the next one to display (from JPCSP)

[thedax]

Confirmed, traveling in GX 3 works in story mode now. If the entire game is playable remains to be seen, however.

[Zakamaru]

Hi Thedax, i have the same problem of the black screen. It seems that u find a solution ? How can i do for solve this problem please, thank you very much

[thedax]

Just use 0.9.7 or later and it should work fine.

[Zakamaru]

Well i try to use 0.9.7 but it show me that i miss some MSVCP120.dll or MSVCR120.dll. Then i have searched theses documents on internet then i found how can i fix this problem. But then when i want to launch the programm, they say "error 0xc00007b".. how can i fix that?

[thedax]

Delete the bad DLLs and install the proper runtimes. 64 bit if you're on x64 Windows, or 32 bit if you're on x86 (or on x64 using 32-bit PPSSPP).