RET - x86-64 instruction reference

retNear return, default size 
Pops the 8-byte return address from the stack into rip and transfers control to it.
8086 and later
Intel's reference: RET 
Used in The x86-64 instruction setx86-64 assembly
call, ret and the System V conventionx86-64 assembly
SSE2 scalar and packed arithmeticx86-64 assembly
Interrupts, exceptions and signalsx86-64 assembly
A subroutine with its arguments in registersx86-64 assembly
Stack arguments and a stack framex86-64 assembly
Recursion: factorial and Fibonaccix86-64 assembly
A jump tablex86-64 assembly
Binary searchx86-64 assembly
Print an unsigned number in bases 2 through 36x86-64 assembly
 
CFPFAFZFSFDFOF –––––––
Changes no flag
Forms  ret  
 ret   
Operands imm16 An unsigned 16-bit byte count, 0 to 65535, added to the stack pointer after the return address is popped. 
In 64-bit mode it pops an 8-byte return address from [rsp] into rip and advances rsp by 8. ret and retn are synonyms; the q suffix explicitly requests the 64-bit form.
The optional sole operand is an unsigned 16-bit count of additional bytes to remove from the stack after popping the return address. For example, ret 16 advances rsp by 24 in total. A negative spelling does not subtract stack space: encoded 0xffff means cleanup of 65535 bytes.
 global _start

section .text
_start:
        mov rax, 12
        call double
        jmp done

double:
        add rax, rax
        ret                 ; back to the line after the call

done:

        mov rax, 60         ; syscall 60: exit
        xor rdi, rdi
        syscall