search

xoreaxeaxeax / movfuscator

The single instruction C compiler
Other
9.39k stars 396 forks source link

I want to build a corresponding movfuscator on the ARM architecture. How should I pay attention to it or how to carry out this project #48

Open dnffabs opened 6 months ago

dnffabs commented 6 months ago

I want to apply this project to the ARM architecture. What should I pay attention to, how should I plan the entire process, and can the author provide some feedback

edcdecl commented 5 months ago

you can't. the entire reason this is possible on x86 is because the mov instruction can do math in its operands.

gogo2464 commented 1 month ago

I do not think it could be mathematically impossible. We can alway use arm code and temp stack variable. I still think it will be hard to implement.

I would like to vcreate a study group for : https://www.reddit.com/r/REMath/ do you wanna join us? :)

edcdecl commented 1 month ago

I do not think it could be mathematically impossible. We can alway use arm code and temp stack variable. I still think it will be hard to implement.

I would like to vcreate a study group for : https://www.reddit.com/r/REMath/ do you wanna join us? :)

it's not impossible but it wouldn't be the same. the whole reason this works is because it abuses x86 by doing everything through lookup tables. you could definitely make a compiler that is only allowed to generate a couple of instructions that make lookup tables possible, and the resulting binary would most likely still look really obfuscated.