Every process, as we saw earlier, has an area of memory called the stack. It grows towards lower addresses, as can be seen in that picture.
The stack is controlled by one register called the stack pointer, or RSP. If we try to do the following:
mov rax, [rsp]
the top element of the stack is copied to RAX, assuming that there is a 64-bit element on the top of the stack. If we know the sizes of other elements that come previously on the stack, we can access them by using offsets (in bytes). For example, assuming again that there is a 64-bit element on top, and a 16-bit element before it, we could access this second element by:
mov ax, [rsp + 8]
We passed the offset of 8, which is the size of the top element. We can also allocate space on the stack (usually done for local variables) by subtracting from RSP:
sub rsp, 8
mov [rsp], 50
Above, we allocated 8 bytes of space on the stack and put a value of 50
there. See sp.asm.
Of course, the most common way of using the stack is by the push and pop
instructions (see push_pop.asm):
push rax ; pushes the value of RAX onto the stack
pop rax ; pops the value that is in the top of the stack off to RAX
Only 16-bit and 64-bit registers can be pushed and popped. The same hold for
immediates and values from memory, but of course you can use the size
keywords word and qword to "cast" these values into 16 and 64 bits,
respectively.