+  RHDN Forum Archive
|-+  Romhacking
| |-+  ROM Hacking Discussion
| | |-+  Asm guys, help me out
Pages: [1]
Author Topic: Asm guys, help me out  (Read 424 times)
D
Guest
« on: October 26, 2006, 12:09:39 pm »
Code:
004A6EA8                           L004A6EA8:
 004A6EA8  33D2                      xor edx,edx
 004A6EAA  8A1518344B00              mov dl,[L004B3418]
 004A6EB0  0315E4774B00              add edx,[L004B77E4]
 004A6EB6  33C9                      xor ecx,ecx
 004A6EB8  8A0DE1774B00              mov cl,[L004B77E1]
 004A6EBE  03D1                      add edx,ecx
 004A6EC0  33C9                      xor ecx,ecx
 004A6EC2  8A0E                      mov cl,[esi]
 004A6EC4  0FAFC8                    imul ecx,eax
 004A6EC7  8D0C49                    lea ecx,[ecx+ecx*2]
 004A6ECA  03D1                      add edx,ecx
 004A6ECC  33C9                      xor ecx,ecx
 004A6ECE  8A0DE0774B00              mov cl,[L004B77E0]
 004A6ED4  03D1                      add edx,ecx
 004A6ED6  8915E4774B00              mov [L004B77E4],edx
 004A6EDC  40                        inc eax
 004A6EDD  46                        inc esi
 004A6EDE  83F838                    cmp eax,00000038h
 004A6EE1  75C5                      jnz L004A6EA8
 004A6EE3  A1A4774B00                mov eax,[L004B77A4]
 004A6EE8  3B05E4774B00              cmp eax,[L004B77E4]
 004A6EEE  0F844D010000              jz L004A7041
 004A6EF4  6A10                      push 00000010h
 004A6EF6  6830764A00                push SSZ004A7630_Mystic_Balloon
 004A6EFB  68C4764A00                push SSZ004A76C4_Corrupt_game_record_found__Repla
 004A6F00  6A00                      push 00000000h
 004A6F02  E8C105F6FF                call jmp_user32.dll!MessageBoxA
 004A6F07  C60518344B0003            mov byte ptr [L004B3418],03h
 004A6F0E  B838000000                mov eax,00000038h
 004A6F13  BEA8774B00                mov esi,L004B77A8

How do I hack this to make it always validate instead of going down the Corrupt_game_record msgbox path?
D
Guest
« Reply #1 on: October 26, 2006, 12:18:25 pm »
Nevermind. I realized right after this post I needed to NOP the last CMP. It works now.
danke
Guest
« Reply #2 on: October 26, 2006, 12:48:51 pm »
03 unlocked : A800
00 cleared
      38 (default value)

Beating 1 level.
07 unlocked : C001
01 cleared   
      +3B
07 unlocked : FB01
02 cleared
      +3E
07 unlocked : 3902
03 cleared
      +41
07 unlocked : 7A02
04 cleared
      +44
07 unlocked : BE02
05 cleared
      +47
07 unlocked : 0503
06 cleared
      +4A
07 unlocked : 4F03
07 cleared
      +4D
07 unlocked : 9C03
08 cleared

...................

0B unlocked : 9E03
05 cleared   
      +47
0B unlocked : E503
06 cleared

Basically, the initial 38 is x, so x = x + 3, or x = (stages cleared * 3) + 38
The second byte of the checksum seems to be (levels unlocked)/2, rounded down to the nearest integer.
« Last Edit: October 26, 2006, 12:54:15 pm by danke »
byuu
Guest
« Reply #3 on: October 27, 2006, 06:14:28 pm »
Quote
Nevermind. I realized right after this post I needed to NOP the last CMP. It works now.

Just to be safe, you might want to leave the CMP alone (or NOP it, whatever, I usually modify the least amount of bytes possible in my hacks) and force the final branch to always occur. I'll admit to not being very well versed with which x86 opcodes set which flags, but one of the above opcodes might clear the zero flag causing the test to "randomly" fail on you if you only disable the CMP opcode. Or maybe not, but just in case:

Code:
004A6EEE  0F844D010000                jz  L004A7041

->

Code:
004A6EEE  E96D4D010000                jmp L004A7041

I *might* be wrong with 0f 84 -> e9 6d ... I used nasm and assembled similar code to get that, but it should be correct. Let me know if it isn't.

EDIT: actually, yeah. In this case, you're ok. The first loop will end with the zero flag set, and the mov opcode does not change the flags (at least not zero), so the next jz opcode will always pass.
At any rate, it's better to use code like the above than to disable CMPs, unless you want to examine the opcodes executed prior to make sure your conditional flag is always set the way you need it to be so you don't get random failures.
« Last Edit: October 27, 2006, 06:20:02 pm by byuu »
Pages: [1]  


Powered by SMF 1.1.4 | SMF © 2006-2007, Simple Machines LLC