The call instruction, behind the scenes, puts the address of RIP (the
instruction pointer) onto the stack, then jumps to the specified label.
When we use ret, the process is the opposite: the address of the caller is
popped from the stack, then the program jumps to it.
When we want to call functions from other programs, there are some calling conventions, which specify things like:
-
where should we put the parameters (in order): registers, stack, etc.
-
which registers the called function should preserve for the caller
-
which registers the caller should save before calling
-
to which number of bytes the stack should be aligned before the call
-
etc.
For now, what is important to know is that in Linux x64 the parameters are passed via registers, in this order: RDI, RSI, RDX, RCX, R8, R9.