 |
| Register | Login | |||||
 |
Main
| Memberlist
| Active users
| ACS
| Commons
| Calendar
| Online users Ranks | FAQ | Color Chart | Photo album | IRC Chat |
|
| | ||
|
| Register | Login | |||||
|
Main
| Memberlist
| Active users
| ACS
| Commons
| Calendar
| Online users Ranks | FAQ | Color Chart | Photo album | IRC Chat |
|
| | ||
| 0 user currently in Programming. | 3 guests |
| Acmlm's Board - I2 Archive - Programming - Help with some N64 code |
 |  |  |
| Add to favorites | "RSS" Feed | Next newer thread | Next older thread |
| User | Post | ||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4935/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
Alright, so this is a chunk of code I'm trying to hack into Zelda OoT (US V1.2) with a Gameshark:sw $a0,0000($sp) ;Save $a0 and $a1The code is placed at 80009300 in RAM and hooked with the instructions jal 9300 and nop at 803B04C0. What it's supposed to do is this: 1) Read a byte from 801C8D75 (low byte of the joypad input). 2) If bits 4 and 5 aren't set, exit. (Otherwise, L and R are pressed.) 3) Read a byte from 8011C089 (language select). 4) If it's 0, make it 1, and vice-versa. 5) Write it back to 8011C089 and exit. The problem, according to the game's useful on-board exception handler, is that somehow the wrong address is being accessed, because the first LUI instruction doesn't work. When it gets to lb $a0,0000($a1), $a1 should be 801C8D75, but it's FFFF8D75.  I've checked the code against the ASM and tried using $k1 in place of $a1, to no avail. (The exception handler doesn't show $k1, but it does show what appears to be the address that was being accessed, and it's FFFF8D75 just like with $a1.) I can't see any reason why this would be happening...Here's the actual Gameshark code. It's kinda messy because it's been hacked to use $k1 and thus the parts where $a1 is moved to and from the stack were NOPed out in case that was causing any problems, and other such things. 81009300 AFA4 81009302 0000 81009304 0000 81009306 0000 81009308 3C1B 8100930A 801C 8100930C 241B 8100930E 8D75 81009310 2404 81009312 0000 81009314 8364 81009316 0000 81009318 241B 8100931A 0030 8100931C 0364 8100931E 2024 81009320 149B 81009322 0008 81009324 0000 81009326 0000 81009328 3C1B 8100932A 8011 8100932C 241B 8100932E C089 81009330 2404 81009332 0000 81009334 8364 81009336 0000 81009338 2484 8100933A 0001 8100933C 3084 8100933E 0001 81009340 A364 81009342 0000 81009344 0000 81009346 0000 81009348 8FA4 8100934A 0000 8100934C 91CE 8100934E C089 81009350 03E0 81009352 0008 81009354 8519 81009356 4A56 813B04C0 0C00 813B04C2 24C0 813B04C4 2400 (edited by HyperHacker on 06-11-05 06:48 PM) |
|||
Ramsus Octoballoon Level: 19 Posts: 92/162 EXP: 34651 For next: 1126 Since: 01-24-05 From: United States Since last post: 39 days Last activity: 71 days |
| ||
(edited by Ramsus on 06-12-05 08:47 AM) |
|||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4949/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
| Hmm, that makes sense. Wouldn't I have to ensure the lower half of $a1 is zero before I do that though? | |||
|
Ramsus Octoballoon Level: 19 Posts: 95/162 EXP: 34651 For next: 1126 Since: 01-24-05 From: United States Since last post: 39 days Last activity: 71 days |
| ||
| I thought the same thing, but the description of lui I found says it zeros the bottom half. | |||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4950/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
Well that fixed the crashing, now I just need to figure out why it isn't doing anything.  |
|||
|
Ramsus Octoballoon Level: 19 Posts: 96/162 EXP: 34651 For next: 1126 Since: 01-24-05 From: United States Since last post: 39 days Last activity: 71 days |
| ||
| Now this makes me want to try some ROM hacking. And did you update the language address part too? That's about all I can suggest right now, since I don't know jack about N64 hardware or Zelda OoT. (edited by Ramsus on 06-12-05 11:56 AM) |
|||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4951/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
| Yeah, I think I just need to find another place to stick the code. It doesn't seem to be getting executed in some places. [edit] Well I found another place to put the code, tweaked it a bit ($s2 and $s3 aren't used so I don't have to use the stack anymore) and it's once again throwing logic out the window. lui $s2,801CIf L and R are pressed, it should be getting the language byte, modifying it, and storing it again. But it's not. If I put an invalid opcode in there, the CPU crashes when I hit the buttons just like it should, but somehow this code is just not doing anything. Also, I can't get the buttons to work right. The way it is now, it activates if L or R are pressed. Removing the 'and' that really doesn't need to be there anyway just disables it entirely.  (edited by HyperHacker on 06-12-05 03:34 PM) |
|||
|
Ramsus Octoballoon Level: 19 Posts: 97/162 EXP: 34651 For next: 1126 Since: 01-24-05 From: United States Since last post: 39 days Last activity: 71 days |
| ||
| How is the joypad status byte laid out? The L or R bug sounds kind of strange. And are you sure the language byte is used during the game? Maybe it's checked at the beginning and used to set a function pointer somewhere. Also, I'm only studying PowerPC and MMIX assembly, but try this: lui $s2,801C lbu $s3,8D75($s2) ;Get joypad status addiu $s2,$zero,0030 ;Value for L+R bne $s3,$s2,.end ;No nop lui $s2,8011 lb $s3,C089($s2) ;Language addr addiu $s3,$s3,0001 ;Next andi $s3,$s3,0001 ;But don't go past 1 sb $s3,C089($s2) ;Store it .end: [original op: addiu $s3,$zero,0004] j 0A2EBC [original op: or $s2,$zero,$zero] EDIT: I switched lb to lbu so it wouldn't set the upper 24 bits (or 56 bits if the registers are 64-bit) to whatever the sign bit of the loaded byte was. I think that might have been related to the L+R bug, but 0x30 in binary wouldn't have the sign bit anyway, so I doubt it. Let me know if it works, and if not, add the and instruction back in. (edited by Ramsus on 06-12-05 06:41 PM) (edited by Ramsus on 06-12-05 06:42 PM) (edited by Ramsus on 06-12-05 06:51 PM) |
|||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4958/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
| I was considering doing C089($s2) instead of setting $s2's lower half, but since it's above 7FFF I figured that might not work right (like 80110000 - 3F77 or something). Using an unsigned byte for the controller layout, though it seems right, doesn't seem to work. Also yes, for some weird reason the game constantly checks its language setting and switches the text and fonts accordingly. |
|||
|
Ramsus Octoballoon Level: 19 Posts: 98/162 EXP: 34651 For next: 1126 Since: 01-24-05 From: United States Since last post: 39 days Last activity: 71 days |
| ||
| They probably had something rigged for the translation team so they could switch between languages while testing the game. Say there's something missing in English, they could switch it to Japanese. I don't even know what the size of the instructions are, so follow what the programmers did with the game. Are there any good debugging tools for N64 emulators? That's about all I can think of trying at this point. |
|||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4959/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
| There's Nemu if by some miracle I can get it running, and the Gameshark which doesn't help a whole lot without step-by-step execution. All instructions are 4 bytes. | |||
neotransotaku Baby Mario 戻れたら、 誰も気が付く Level: 87 Posts: 3199/4016 EXP: 6220548 For next: 172226 Since: 03-15-04 From: Outside of Time/Space Since last post: 11 hours Last activity: 1 hour |
| ||
| how are you converting the ASM into Gameshark code? some things to consider: (1) have you tried lw? when you use lb/lbu, you have to consider endian issues. with lw, you don't have that problem unless the address you try to load from is not divisible by the size of a word (2) instead of using that and to do the comparison between $s2 and $s3, why don't you use beq? |
|||
|
HyperLamer <||bass> and this was the soloution i thought of that was guarinteed to piss off the greatest amount of people Sesshomaru Tamaranian Level: 118 Posts: 4979/8210 EXP: 18171887 For next: 211027 Since: 03-15-04 From: Canada, w00t! LOL FAD Since last post: 2 hours Last activity: 2 hours |
| ||
| The AND was so that if you pressed, say, L + R + A it would still activate. I figured there's no point doing that though, so I removed it. The conversion is a not-very-fun process of typing the ASM into Niew, copying the hex into Textpad, using regular expressions to format it into two-byte chunks and adding addresses. I can easily make simple changes to the code without having to do that though (like NOPing out an instruction or replacing it with an invalid one to see if it's even being executed). |
|||
|
neotransotaku Baby Mario 戻れたら、 誰も気が付く Level: 87 Posts: 3209/4016 EXP: 6220548 For next: 172226 Since: 03-15-04 From: Outside of Time/Space Since last post: 11 hours Last activity: 1 hour |
| ||
| maybe your branches don't work correctly when branching--other than monitoring registers, i'm not sure what else to help you with |
| Add to favorites | "RSS" Feed | Next newer thread | Next older thread |
| Acmlm's Board - I2 Archive - Programming - Help with some N64 code |
| | |
