Kodewerx

Nope.
First we assume that the offset is cleared before your codes (offset=0).
Then let's say that at 0x021C4D28 you have 0x02000000, and that at 0x02000004 you have 0x02234568

Now B21C4D28 00000000 loads the value at [offset+0x021C4D28] into the offset. That means the offset is now 0x02000000.

B0000004 00000000 loads the value at [offset+0x00000004] into the offset. As the offset was 0x02000000, the offset is now [0x02000004] = 0x02234568.

After that, your code
0000AA3C EDB88320
writes 0xEDB88320 at [offset+0x0000AA3C], ie at 0x02234568+0x0000AA3C = 0x223EFA4.

Sounds like No$GBA may be a case of pointer-to-pointer-to-data, where the data and second pointer move around. It may even be intentionally trying to make this procedure difficult. If nothing else, you can locate the RAM and edit emulators.cfg every time you run the emulator.

_________________
143
HyperNova Software is now live (but may take a few tries to load) currently down; check out my PSP/DS/Game Boy/Windows/Linux homebrew, ROM hacks, and Gameshark codes!

i just use the +10000h pointer, and all the codes that i use there i just add those 10000h, i hope viper fixes that later.

@kenobi
Then this is like a addition to the offset? like your hack.

_________________

DO NOT send me code requests/conversions via pm. I DON'T GIVE PERMISSION, READ THE FUCKING RULES.
I Hate You So Much!!! Fuck & Peace!

Everyone has AIDS!
My grandma and my dog 'ol blue
My father
My sister
My uncle and my cousin and her best friend
The pope has got it and so do you.

My SITE

My Official Code Breaker DS Codes

Does the pointer always point to 0x10000 before the beginning of RAM? Maybe all that's needed then is the ability to add an offset to the pointer.

Actually NDS RAM could be a problem, since it has remappable VRAM and ITCM. Who knows how a given emulator handles that...

_________________
143
HyperNova Software is now live (but may take a few tries to load) currently down; check out my PSP/DS/Game Boy/Windows/Linux homebrew, ROM hacks, and Gameshark codes!

yes, there are only fixed pointers with +10000h of distance or more. i checked in a debugger, it's coded in the emulator, so this pointer is fixed.

i was talking with Viper in mirc, if he can use the artmoney emulator configuration file as base, since it works with me and believe with everyone lol

NO$GBA 2.4b;NO$GBA.EXE
;NDSRAM 4mb;02000000;P0046D998,10000;400000

;text;Base address;pointer address;pointer offset add or sub;ram size

_________________

DO NOT send me code requests/conversions via pm. I DON'T GIVE PERMISSION, READ THE FUCKING RULES.
I Hate You So Much!!! Fuck & Peace!

Everyone has AIDS!
My grandma and my dog 'ol blue
My father
My sister
My uncle and my cousin and her best friend
The pope has got it and so do you.

My SITE

My Official Code Breaker DS Codes

I don't really get how to combine the buttons.

BITWISE AND!

I think this will make him more confused

_________________

DO NOT send me code requests/conversions via pm. I DON'T GIVE PERMISSION, READ THE FUCKING RULES.
I Hate You So Much!!! Fuck & Peace!

Everyone has AIDS!
My grandma and my dog 'ol blue
My father
My sister
My uncle and my cousin and her best friend
The pope has got it and so do you.

My SITE

My Official Code Breaker DS Codes

Windows calculator has an AND button so have at it.

_________________

That is, if he can figure out how to put it in Scientific mode. Assuming he has a brain, it shouldn't be difficult.

_________________
143
HyperNova Software is now live (but may take a few tries to load) currently down; check out my PSP/DS/Game Boy/Windows/Linux homebrew, ROM hacks, and Gameshark codes!

Just an early heads up, new Action Replay DS firmware is due out later this week or next week.

Only thing new are the code types present in the Trainer Toolkit that the AR itself didn't have.

Just upgraded to 1.54.

They seem to have made some "big" changes in their code handler's code type handling (they use a table instead of fixed jumps, so some 'old' ar hack would not work anymore). I'll study it and report what's new/changed.

Edit :
NDS AR HACKs #4 and #5 should still be working correctly (ie. asm execution hack, and memory copy hack). However the registers used changed, so if your hack relied on the AR's Code handler register, it'll not work properly.

Codes type C4, C5 and C6 have been added (check the TT's code list to know what they do).

Code type DB's bug is corrected.

The TT's code handler seems to be the same than the one that came with the AR 1.52 (its versions are : body v0.08, hook v0.04, patcher v0.03, code engine v1.22...).

And if an "IF" code type's address is zeroed, the offset is used.


Edit2 :
Small warning about the C5 code type. It doesn't work the same way on the AR v1.54 and the Trainer Toolkit :/
On the AR, the counter is increased each time a C5 code type is encountered (even if the execution status is false), compared on the TT where is it 'timer based'. That is bad, because if two different person make codes using the C5 code type, the result will be wrong (the counter will be increased by 2 everytime the AR code handler is executed). Also, the address at which the counter is stored is weird (it's 023200B8 - that could be/is in the middle of the game data?).
So I advice not using the C5 code type until it has been fixed. I'll make some hack(s) later (at least to change the address).

is it possible for it to be use on an emulator?

No$GBA can run the ROM, but there's no way to load a game.

_________________
143
HyperNova Software is now live (but may take a few tries to load) currently down; check out my PSP/DS/Game Boy/Windows/Linux homebrew, ROM hacks, and Gameshark codes!

Small hack for some AR v1.54 codes type (they'll only work for the AR v1.54, using the auto hook system (ie. when no custom (m) code are used)):

This hack will make the C5's code type's counter be inside the code handler (at 0x023FE578) :
523FE324 E5923000
023FE324 E59F324C
023FE32C E58F3244
D2000000 00000000
Could prevent some crashes/bugs when using the C5 code type.


This other 'hack' will make the C5's code type work 'properly', ie. the counter will only be increased once each time the code handler is executed (and not each time a C5 code is executed), and the counter will be inside the code handler (at 0x023FE578) :
523FE324 E5923000
023FE324 E59F324C
023FE328 E1A00000
023FE32C E58F3244
D2000000 00000000
D9000000 023FE578
D4000000 00000001
D6000000 023FE578
D2000000 00000000


Finally, this hack will make the AR If... codes use address+offset if the address of the code is not 0ed (instead of the address alone). Not quite sure if it'll be useful or not... :
Enable address+offset on If... codes :
023FE128 11A04005

Disable address+offset on If... codes :
023FE128 11A04000

I might update the docs in the first post later on.

With firmware v1.54 and the asm hack, what's the current status of the register usage?

I already learned the hard way that I must preserve r6 now. Can I still use r8 and r12 freely? (I assume there's no problem with r0-r3).

Can I still calculate and put the offset into r10 in order to be used outside the asm afterwards?

Why don't you just push everything before your routine and pop everything afterwords?

But the best part about AR hack #4 is not having to find hooks OR push registers! :/

_________________

Well, why spend cycles pushpopping when you don't need to? According to no$gba, stmfd [r13]!, {r0-r12, r14} takes a whopping 15 cycles and the corresponding ldmfd even 16 cycles on the ARM7TDMI, and those are probably only machine cycles not taking into account memory access.

Anyways, sometimes one wants to make use of the registers that have useful functions, and thus one might want to know what the exact functions are ...

I'll look at the new registers soon, and update the informations.

Bleh that previous post was unneeded, now I have to double post to let you know I added the data in the from page.

Here is what I added :

Update : ARv1.54 Registers :
These ones can be used freely :
r0 = holds the address of the E code type plus the offset (0XXXXXXX + offset)
r1 = holds the starting address of the E code type data (what I used for the bx)
r2 = Data (holds the number of bytes of data you entered in the E code type)
r3 = Code type (0x0000000E)
r4 = Address (without code type), or offset if address=0.
r5 = holds the address of the E code type plus the offset (0XXXXXXX + offset)
r8 = holds what the next 'execution status' data would look when set to true
r11 = holds what the next 'execution status' data would look when set to false
r12 = Current execution status


These ones shouldn't be changed, unless you know what you are doing :
r6 = (holds the number of bytes of data you entered in the E code type)
r7 = Holds the offset of the E code type data
r9 = Offset
r10 = (holds the 'execution status' data)
r13 = (SP)
r14 = (LR)

That's excellent, it helps me a lot Thank you so much.

I corrected the r7 register information (used to be 'address of the next code to be executed').

FYI, the address of the next code to be executed is somewhat computed like that :
x = r7+(((r6+7)>>3)<<3)

Does it seriously do two shifts when a bit clear does the same thing? XD

_________________

Erm sorry. I guess I was a bit tired (stupid excuse, heh?).
r7 is not the address, but an offset (to add to the starting address of the code list).

So my phrase should be changed to :

"FYI, the offset of the next code to be executed is computed like that :
x = r7+(((r6+7)>>3)<<1)"

Then the AR adds that offset<<2 to the starting address of the code list.

I might also add that :
[r13] = # of times to repeat when a D1/2 code is encountered (=data of the C code type)
[r13+4h] = offset of the code to go to when a repeat is executed.
[r13+8h] = execution status poped when a repeat is executed.
[r13+Ch] = Dx Data register
[r13+10h] = # of codes in the code list
[r13+14h] = address in ram of the start of the code list

Updated ASM to ARDS to version 1.01. Added support for the Trainer Toolkit and Nitro Hax.
Check the bottom of the first post of this thread to find the tool.

I'm looking at this code type:



and I'm wondering - is there a way to do the reverse? I mean, instead of add the value, subtract it? I used to do some code hacking on the PS1 Gameshark/AR, and it had code types for both incrementing and decrementing, which is why I'm surprised I don't see an obvious way to do this on DS.

A substract code is 'useless'. To substract Y, just add the 32-bits value not(Y)+1...

Ie. if you want to substract 1, add 0xFFFFFFFF. If you want to subtract 0x123, add 0xFFFFFEDD...

hello,i am trying to understand and use ar codes and asm with it.i have trying using this code on my ds,(i use TTDS card,i have mario ds original cartidge and this for homebrew,and because it has ar chat ,i use also for experiment with cheats.
After several attempts ,and i thought maybe this card only support normal ar cheats and doent this kind of chats (executing asm) ,i was able to make it run.
The version is EU .

And this is the code:


023FE074 E3520003
E2098930 00000004 //02098930 is memory region for lives in eu cartidge
00000001 00000000
023FE074 012FFF11
94000130 FFF30000 //it does not run
E0000000 00000014
E59F0008 E3A02063
E5802000 E12FFF1E
02098930 00000000 //02098930 is memory region for lives in eu cartidge
D2000000 00000000
0209F2F4 00000063


I dont understand why you put 023FE074 012FFF31 if according your info is 11,altough it works also

and can you explain what do the two last lines,if i dont use ,it doesnt run,but it is suppoused the code ends when you freeze the hack and finish with the data used by the first instruction


The patch key also doesnt work,i use also the r4cce editor and add it with it ,but the code runs without press the key

My question would be why you have 0209F2F4 00000063 after ending the code with D2... if you wanted to end the if statement it should be D0.

_________________