DesMuMe mods !!!

[shashClp]

It's, afaik, a timing problem, related to the IRQs. Check the Zelda Video Demo (it isn't running), and you'll see it's the same problem. I've worked on it a few hours, with mid to good results, but it's far for finished.

[NHervé]

Thanks. I can tell you RaymanDS is not working because of a strange problem in the ARM9. The ARM9 freezes on this opcode :
MCR CP15, 0, R0, CR7, CR0, 4
which means "Wait for IRQ". IE=0x00040019 and IF=0x000C0019, so IRQ should be triggered, but it is not.

[shashClp]

Originally Posted by NHervé

Thanks. I can tell you RaymanDS is not working because of a strange problem in the ARM9. The ARM9 freezes on this opcode :
MCR CP15, 0, R0, CR7, CR0, 4
which means "Wait for IRQ". IE=0x00040019 and IF=0x000C0019, so IRQ should be triggered, but it is not.

Same problem as Mario Kart, Pokemon, Zelda Video Demo and probably others :P

[NHervé]

This is VERY strange. IE&IF=0x00040019, so IRQs should be triggered. DeSmuME does'nt trigger these IRQs, I don't know why. Very strange.

PS : Why do you don't commit your changes to CVS ?

[shashClp]

Originally Posted by NHervé

(...)

PS : Why do you don't commit your changes to CVS ?

Why should I? That's the interesting question. Anyway, to make a long story short, I'm not very interested on seeing games running on desmume, but making it a good development tool. As of now, the included debugger tools are far from good, so if it's not going to be useful for homebrew development, there isn't much point for me.

Of course I make an incompatible game run once and then, just because the challenge is fun.

[NHervé]

Okay. I think I'll have to debug my incompatible games myself (Worms2, MarioKart, RaymanDS). I think the RaymanDS bug is very strange.

[shashClp]

Oh, and I forgot that I won't be committing minor fixes on official Desmume for a long while (as I did before, basically supporting homebrewers requests), you can find the reason somewhere in the forums.

[NHervé]

I found what's wrong with RaymanDS : The ARM9 freezes on this MCR opcode because the I flag is set, which means "Disable IRQ".

[shashClp]

Originally Posted by NHervé

I found what's wrong with RaymanDS : The ARM9 freezes on this MCR opcode because the I flag is set, which means "Disable IRQ".

Hacking it to one makes the game run ? :P

[NHervé]

PS : I hacked this by forcing the I flag to be 0, I got over 2 intros further but the game still doesn't work. After reading the ARM datasheet, I found that this MCR opcode switches the proc in low-power mode, and any IRQ can wake it up even if IRQs are disabled (I flag set to 1), which DeSmuME doesn't support. I'll try to implement this feature and I'll tell you.

[shashClp]

Originally Posted by NHervé

PS : I hacked this by forcing the I flag to be 0, I got over 2 intros further but the game still doesn't work. After reading the ARM datasheet, I found that this MCR opcode switches the proc in low-power mode, and any IRQ can wake it up even if IRQs are disabled (I flag set to 1), which DeSmuME doesn't support. I'll try to implement this feature and I'll tell you.

Didn't know that it was meant to ignore the I register, good luck implementing it.

[NHervé]

I'm gathering more and more infos : before calling the MCR opcode, the game stores the 8 first bits of R2 into the control bits of the CPSR. Maybe R2 is having wrong value. I think any NDS program shouldn't set I flag to 1 before calling a "wait for IRQ" opcode.

[refraction]

Just thought id put my 2 cents in with my lack of knowledge.

from reading this

FIQ (Fast Interrupt Request)
This interrupt is generated by a LOW level on the nFIQ input. It is supposed to process timing critical interrupts at a high priority, as fast as possible.
Additionally to the common banked registers (R13_fiq,R14_fiq), five extra banked registers (R8_fiq-R12_fiq) are available in FIQ mode. The exception handler may freely access these registers without modifying the main programs R8-R12 registers (and without having to save that registers on stack).
In privileged (non-user) modes, FIQs may be also manually disabled by setting the F Bit in CPSR.

IRQ (Normal Interrupt Request)
This interrupt is generated by a LOW level on the nIRQ input. Unlike FIQ, the IRQ mode is not having its own banked R8-R12 registers.
IRQ is having lower priority than FIQ, and IRQs are automatically disabled when a FIQ exception becomes executed. In privileged (non-user) modes, IRQs may be also manually disabled by setting the I Bit in CPSR.
To return from IRQ Mode (continuing at following opcode):
SUBS PC,R14,4 ;both PC=R14_irq-4, and CPSR=SPSR_irq

It sounds like (what you are describing) could be a Fast IRQ, which looking at that description can only be disabled by the F bit in the CPSR reg, totally ignoring the I bit. So it might be worth checking some of the other register settings.

Im probably completely wrong, but if it gets the cogs ticking, thats all that matters

[NHervé]

Maybe it's that, or maybe the game is disabling IRQs for a special purpose and it uses that MCR opcode that ignores the I flag.

I tried by adding a "lowPower" variable in the armcpu_t struct, setting it to 1 when the MCR opcode is called and to 0 when a IRQ occurs, but it is still not correct. It makes the game freezes in a strange location (0x0FFF0000).

[refraction]

do you know when the IRQ disable flag is being set? or is it being done manually and not being reset? i dunno cos ive not looked at the source

[NHervé]

Yes. The game calls a subroutine in order to set the CPSR to 0x1F (I=0), then calls the MCR that succeds, calls then another subroutine that sets the CPSR to 0x9F (I=1), then calls the same MCR. This time the MCR lamentably freezes the ARM9 due to I=1.

[refraction]

okay, any idea what causes the interrupt?, maybe it expects a response by then.

Meh, ill leave you guys to it, i really dont know the DS system

[NHervé]

I don't know. When I hack it I get 3 intros instead of 1, but it's still not correct. It seems that what I described in my previous post is done by the game in order to "test" something. Maybe it's the init sequence of the SDK that's been used to develop this game.

[Normmatt]

NHervé, can you post a disassembly of the last few opcodes before it freezes, maybe it will tell us what its doing and expecting from the hardware.

[NHervé]

Well, I'll do it.
Sorry to be late, but the forums recently had problems due to moving servers.