x86 (32-bit) VS x64 (64-bit)

Assembly - Nightmareguyinatuxedo.github.io

1. General comparisons

	x86	x64
How are arguments passed to functions in an ELF (Linux) architecture?	Passed on the stack	Passed via registers
Which register is used to return a value from a C function?	eax register	rax register
Maximum register size	4 byte (32 bit)	8 byte (64 bit)
...

2. Registers

As mentioned before, the x64 architecture has 8 byte registers, while x86 can only access up to the 4 byte registers. There also exists 2 byte and a single byte register, as defined in the table below:

Taken directly from source

Taking the 8 byte register rax as an example, we can see that the lower 4, 2 and single byte(s) register will be the eax, ax and al respectively.

• eax: Lower 4 bytes of the rax register

• ax: Lower 2 bytes of the rax register

• al: Last byte of the rax register

3. Passing of arguments to functions

As discussed before, we know that x86 architecture passes arguments to functions via the stack, while x64 does it via registers (rdi, rsi, rdx, etc.). In this section, we will discuss examples of how this can be achieved

x86 (via stack)

Most compilers have the convention of storing the argument value in the eax, ebx,ecx_, etc. registers (depending on the number of arguments) temporarily, right before the values are pushed to the stack

• In the example below, the function_to_call function receives two arguments: 2 and 22:

mov eax, 2       ; first argument
push eax         ; push first argument onto stack
mov ebx, 22      ; second argument
push ebx         ; push second argument onto stack
call function_to_call

x64 (via registers)

Function arguments are passed via the following registers: rdi, rsi, rdx, rcx, r8, r9 (1st to 6th argument)

• In the example below, the function_to_call function receives two arguments: 2 and 22:

mov rdi, 2      ; first argument
mov rsi, 22     ; second argument
call function_to_call