How is the stack pointer changed in this program with call and ret

Question

My questions pertain to the actions that seem to happen between the line when context is changed especially concerning RSP and RBP.

Given this very simple program:

Reading symbols from ./function_call...done.
(gdb) disass main
Dump of assembler code for function main:
   0x00000000004004d6 <+0>: push   rbp
   0x00000000004004d7 <+1>: mov    rbp,rsp
   0x00000000004004da <+4>: mov    esi,0x2
   0x00000000004004df <+9>: mov    edi,0x1
   0x00000000004004e4 <+14>:    call   0x4004b6 <add_and_7>
   0x00000000004004e9 <+19>:    mov    eax,0x0
   0x00000000004004ee <+24>:    pop    rbp
   0x00000000004004ef <+25>:    ret    
End of assembler dump.
(gdb) disass add_and_7
Dump of assembler code for function add_and_7:
   0x00000000004004b6 <+0>: push   rbp
   0x00000000004004b7 <+1>: mov    rbp,rsp
   0x00000000004004ba <+4>: mov    DWORD PTR [rbp-0x14],edi
   0x00000000004004bd <+7>: mov    DWORD PTR [rbp-0x18],esi
   0x00000000004004c0 <+10>:    mov    DWORD PTR [rbp-0x4],0x7
   0x00000000004004c7 <+17>:    mov    edx,DWORD PTR [rbp-0x14]
   0x00000000004004ca <+20>:    mov    eax,DWORD PTR [rbp-0x18]
   0x00000000004004cd <+23>:    add    edx,eax
   0x00000000004004cf <+25>:    mov    eax,DWORD PTR [rbp-0x4]
   0x00000000004004d2 <+28>:    add    eax,edx
   0x00000000004004d4 <+30>:    pop    rbp
   0x00000000004004d5 <+31>:    ret    
End of assembler dump.
(gdb) list
1   int add_and_7( int num1, int num2 ) {
2       int seven = 7;
3       return num1 + num2 + seven;
4   }
5   
6   int main() {
7       add_and_7( 1, 2 );
8       return 0;
9   }

All functions start off with push rbp which I as I understand it is preserving the parent context onto the stack. How does the parent function know how to rebuild itself? Are the necessary steps built into call and ret?

Then the rsp is always moved to rbp. As I have read this sets the new stack base to be in the context of the current function. What I can't seem to figure out is when or how stack pointer was set to that point in the first place. My best guess is the assembly function call does this, is that whats happening?

Lastly when a method returns it seems like eax is the register that is used for the parent function to utilize the return of its child function. Is eax explicitly used for this or is this just a convention with my compiler and architecture?

Answer 1

How does the parent function know how to rebuild itself ? Are the necessary steps built into call and ret?

Before calling a function, current status of registers are saved, as well as the return address. call instruction jumps to particular address, where the called function begins. The return address is pushed onto stack. When called function returns, ret instruction pops previously pushed return address and goes to that location.

Then the rsp is always moved to rbp

rbp is previously pushed onto stack to be able to restore rbp's value from caller's function. Then, rsp is moved to rbp to create a new stack frame for callee function. The new base pointer has been set up. So currently, rbp and rsp points to the same addresses. If there are other push instructions, esp is automatically adjusted. When function is done, the pop ebp instruction restores previously pushed stack base pointer address.

Answer 2

Push and Pop modify the stack pointer - SP.

Call pushes FLAGS - status register as well as the RA - return address. Ret pops the FLAGS pops and jumps to the return address.

As rkhb said, the need to keep certain registers as they are comes from the calling conventions.