Arm Patching

[KabutHitam]

i need code fast like 

~A8 FMOV S0, #0x41980000 

but i want so fast

[maars]

Use the first site to get arm/8 opcode then the second to convert it into bytes code

• https://gcc.godbolt.org/
• https://armconverter.com/
  
  
  
  Make sure to configure gcc correctly by setting the language to c++ and the compiler to armv8-a or armv7-a

[XEKEX]

On 8/11/2023 at 4:49 AM, KabutHitam said:

i need code fast like 

~A8 FMOV S0, #0x41980000 

but i want so fast

what do u mean fast also is it 64 or 32 

 

[maars]

I guess he want it like this, dom might help xD

[KabutHitam]

2 hours ago, XEKEX said:

what do u mean fast also is it 64 or 32 

 

for 64, i using fmov for float 

1 hour ago, MAARS said:

I guess he want it like this, dom might help xD

dom can't help me ahhahaha

[maars]

33 minutes ago, KabutHitam said:

for 64, i using fmov for float 

2 hours ago, MAARS said:

Nice, but your request still lack context, what are you trying to do ?

[THETWINSOFFICIAL]

Have you tried it?

 

Like this ?

 

X1, x2

 

 X0, x1
 

[KabutHitam]

23 hours ago, MAARS said:

Nice, but your request still lack context, what are you trying to do ?

i just want opcode high value for speed

14 hours ago, AMIYA said:

Have you tried it?

 

Like this ?

 

X1, x2

 

 X0, x1
 

you get stuck when you use that, only fmov

[kiynox]

[ @KabutHitam ]
---

Quote

i just want opcode high value for speed

From what I understand, you're trying to set a large possible value. You can use this template:
---
[ Float 0-65535 ]

MOVW R0, #(1-65535)
VMOV SO, R0
VCVT.F32.S32 S0, S0
VMOV R0, S0
BX LR

---
[ Float 65535 - 131072 ]: R0 = R0 + R1

MOVW R0, #(0-65535)
MOVW R1, #(0-65535)
ADD R0, R0, R1
VMOV SO, R0
VCVT.F32.S32 SO, S0
VMOV RO, SO
BX LR

---
[ Float 131072 - 429503284]: R0 = (R0 x R1) + R1

MOVW R0, #(0-65535)
MOVW R1, #(0-65535) --multiply here
MUL R0, R0, R1
MOVW R1, #(0-65535) --addition here
ADD R0, R0, R1
VMOV S0, R0
VCVT.F32.S32 S0, S0
VMOV R0, S0
BX LR

---
*Above mnemonic are replicable for getting a really high value.
*Do note that most function has it's own Max Limit, so you need to change them first (example: MaxHP & MinHP)
*( credit: @BadCase )

[huanggua]

21 hours ago, Xaviesz said:

[ @KabutHitam ]
---

From what I understand, you're trying to set a large possible value. You can use this template:
---
[ Float 0-65535 ]

MOVW R0, #(1-65535)
VMOV SO, R0
VCVT.F32.S32 S0, S0
VMOV R0, S0
BX LR

---
[ Float 65535 - 131072 ]: R0 = R0 + R1

MOVW R0, #(0-65535)
MOVW R1, #(0-65535)
ADD R0, R0, R1
VMOV SO, R0
VCVT.F32.S32 SO, S0
VMOV RO, SO
BX LR

---
[ Float 131072 - 429503284]: R0 = (R0 x R1) + R1

MOVW R0, #(0-65535)
MOVW R1, #(0-65535) --multiply here
MUL R0, R0, R1
MOVW R1, #(0-65535) --addition here
ADD R0, R0, R1
VMOV S0, R0
VCVT.F32.S32 S0, S0
VMOV R0, S0
BX LR

---
*Above mnemonic are replicable for getting a really high value.
*Do note that most function has it's own Max Limit, so you need to change them first (example: MaxHP & MinHP)
*( credit: @BadCase )

How to use GameGuardian to modify so many assembly codes?

[kiynox]

[ @huanggua ]
---

Quote

How to use GameGuardian to modify so many assembly codes?

Make sure that you're know the difference between Arm and Thumb. The difference between both is that: Arm uses 4 hexes / 8 bytes, meaningwhile, Thumb usually uses only 2 hexes / 4 bytes. For example:

Quote

Assembly: NOP
Arm: 00 F0 20 E3
Thumb: 00 BF

---
It is important because you can only change one instruction with another one WITH the same hex length. You can replace the next instruction to fit your long-patches, but make sure that: if it's a function, you need to ends it with:

Arm32: BX LR
Arm64: RET

---

Edited August 17, 2023 by Xaviesz

[huanggua]

On 2023/8/18 at 2:05 AM, Xaviesz said:

[ @huanggua ]
---

Make sure that you're know the difference between Arm and Thumb. The difference between both is that: Arm uses 4 hexes / 8 bytes, meaningwhile, Thumb usually uses only 2 hexes / 4 bytes. For example:

---
It is important because you can only change one instruction with another one WITH the same hex length. You can replace the next instruction to fit your long-patches, but make sure that: if it's a function, you need to ends it with:

Arm32: BX LR
Arm64: RET

---

Do you mean that, for example, after I find the address of the MAXHP function, I can use GameGuardian to modify this value and some subsequent values to MOVW R0, #(1-65535)

VMOV SO, R0

VCVT.F32.S32 S0, S0

VMOV R0, S0

BX LR ?

Maybe I didn't express myself clearly, see if you can understand now.

[kiynox]

[ @huanggua ]
---

Quote

some subsequent values

That's right. If it's doesn't fit, just replace the next instruction.
---