The binary of opcode in assembly

问题:

I have the following code (after producing the listing file, written for intel 80x86):

 1                                    global _start
 2                                  
 3                                  section .data
 4 00000000 03000000                x:  dd  3
 5                                  
 6                                  ;section .text
 7                                  
 8                                  _start:
 9 00000004 8B0D[00000000]              mov ecx, [x]
10 0000000A 000D[16000000]          r:  add byte [l+6], cl
11 00000010 C605[00000000]30        l:  mov byte [x], 48
12 00000017 51                          push    ecx
13 00000018 B804000000                  mov eax, 4      ; For "Write" system call
14 0000001D BB01000000                  mov ebx, 1      ; to standard output
15 00000022 B9[00000000]                mov ecx, x      ; "buffer"
16 00000027 BA01000000                  mov edx, 1      ; byte counter
17 0000002C CD80                        int 0x80
18 0000002E 59                          pop ecx
19 0000002F E2D9                        loop    r, ecx
20                                      
21 00000031 BB00000000                  mov ebx, 0
22 00000036 B801000000                  mov eax, 1      ; For "exit" system call
23 0000003B CD80                        int 0x80

I'm concentrating now on row 19, and I don't completely understand it. I understand that the binary of the opcode 'loop' is E2.

But from where the D9 byte? how it was calculated?

回答1:

19 0000002F E2D9 loop r, ecx

Where does the second opcode (D9) come from?

The second opcode (0xD9 in this case) is the relative destination address in two's complement - since you are jumping backwards, it is negative in this case:

  0x00000031   (The address following the loop instruction)
+ 0xFFFFFFD9   (Signed-extended representation of 0xD9 - actually a negative number, -39 decimal)
============
  0x0000000A   (The address of the r label)

Note that the destination address is calculated based on the address after the loop instruction.

See also http://www.mathemainzel.info/files/x86asmref.html#loop