נושא: בוחן באסמבלי

< http-equiv="Content-" content="text/; charset=utf-8">< name="ProgId" content="Word.">< name="Generator" content="Microsoft Word 12">< name="Originator" content="Microsoft Word 12"><link rel="File-List" href="file:///C:%5CDOCUME%7E1%5C2DBD%7E1%5CLOCALS%7E1%5CTemp%5Cmsohtmlclip1%5C01%5Cclip_filelist.xml" target="_blank"><link rel="themeData" href="file:///C:%5CDOCUME%7E1%5C2DBD%7E1%5CLOCALS%7E1%5CTemp%5Cmsohtmlclip1%5C01%5Cclip_themedata.thmx" target="_blank"><link rel="colorSchemeMapping" href="file:///C:%5CDOCUME%7E1%5C2DBD%7E1%5CLOCALS%7E1%5CTemp%5Cmsohtmlclip1%5C01%5Cclip_colorschememapping.xml" target="_blank">< style>

1. Consider the following 80X86 code segment.

2.          section .data

3.   

4.           mov    eax, 0

5.        l: shr    ebx, 1

6.           adc    eax, 0

7.            add    ebx, 0

8.            jnz    l

We wish to write this code without using any form of LOOP, and with no conditional jumps (we can use other forms of JMP, however). Write a code segment that provides the same results for eax (but may use or destroy other registers). You may NOT "unfold" the above loop!

2.      Write 80X86 instructions to get a value of 1 in the ax register on an 80X86:

1. List 5 different ways to do this with ONE instruction.
2. List 1 way to do this with 2 instructions, each different from all the instructions used in part 1.
2. List the SHORTEST POSSIBLE CODE for adding 5 to y, subtracting 2 from x, and adding 1 to z, which are defined as follows (consecutive locations):

4.      x    db     1

5.      z    dw     330

6.      y    db     7

If we know that subtracting 2 from x and adding 1 to z cause no overflow, is there a shorter way to do that? How? Do this: a) For an 80X86 machine, 2) For Motorola 68000, which is a big-endian machine (assume that it has an instruction: ADDI.S #num, var that adds a number num to a location in memory at address var. S is the operand size, one of: B (byte), W (16 bit word), L (32 bit longword).

4.      Consider the following 80X86 code segment. Assume that when the code is run, CX is non-zero.

5.       MOV    AX, 0

6.    L: ADD    AX, BX

7.       ADC    DX, 0

8.       LOOP   L

1. What is the result of running the code?
2. Re-write the above code segment for better efficiency.

 

 

 

9. Consider the following code for Motorola 68000 (comments state what each instruction does, according to the 68000 instruction manual (no, we did not teach you that machine, you should learn the relevant details on your own from the exercise). Note that the 68000 has 32-bit registers D0-D7 and A0-A7, where A7 is also the STACK POINTER.

10. F: MOVE.L   D0, -(A7)     ; D0 to memory - predecrement mode

11.    SUBQ.L   #1, D0        ; Subtract immediate - long (32 bit) operand

12.    BMI      N             ; Branch (jump) if result was negative

13.    JSR      F             ; CALL subroutine F (push PC then jump to F)

14.    SUB.L   (A7)+, D2      ; Signed subtract memory (postincrement)

15.                             ; with D2, result in D2

16.    RTS                    ; Return from procedure/subroutine

17. N: MOVE.L   (A7)+, D0     ; Move memory to D0 - postincrement mode

18.    MOVEQ.L  #0, D2        ; Move immediate to D2

19.    RTS

What happens if we execute an instruction JSR F, with D0=0? D0=1? Other values of D0 (call that value k)?