Final Exam of System Programming

NCNU System Programming Final Exam

Date: January 19th, 2007

Time: 09:10-

(10%) Consider the assembly code with macro invocations in the following program. After the macros are expanded, can we use a 1-pass assembler to handle this program? Why?

XMAS	START	1000
PRINT	MACRO	&CH
$LOOP	TD	OUTDEV
	JEQ	$LOOP
	LDCH	#&CH
	WD	OUTDEV
	MEND
OUTDEV	BYTE	x'06'
HEIGHT	WORD	4
MARGIN	WORD	5
I	RESW	1
J	RESW	1
EXPR1	RESW	1
FIRST	LDA	#1
L1B	STA	I
	COMP	HEIGHT
	JGT	L1E
	LDA	I
	ADD	#2
	STA	EXPR1
	LDA	I
L2B	STA	J
	COMP	EXPR1
	JGT	L2E
	LDA	MARGIN
	ADD	#1
	SUB	J
	RMO	A,S
	LDX	#1
L3B	COMPR	X,S
	JGT	L3E
	PRINT	32
	TIXR	X
	J	L3B
L3E	LDA	J
	MUL	#2
	SUB	#1
	RMO	A,S
	LDX	#1
L4B	COMPR	X,S
	JGT	L4E
	PRINT	42
	TIXR	X
	J	L4B
L4E	PRINT	00
	LDA	J
	ADD	#1
	J	L2B
L2E	LDA	I
	ADD	#1
	J	L1B
L1E	LDA	#1
L5B	STA	I
	COMP	HEIGHT
	JGT	L5E
	LDA	MARGIN
	SUB	#1
	RMO	A,S
	LDX	#1
L6B	COMPR	X,S
	JGT	L6E
	PRINT	32
	TIXR	X
	J	L6B
L6E	LDX	#1
	LDS	#3
L7B	COMPR	X,S
	JGT	L7E
	PRINT	42
	TIXR	X
	J	L7B
L7E	PRINT	00
	LDA	I
	ADD	#1
	J	L5B
L5E	RSUB
	END	FIRST

(10%) If we use a two-pass assembler to assemble it to an object file, what is the value of the program size in the Header record?
(10%) Show the output (at device 06) that we shall obtain when we run the above SIC/XE assembly program.

(10%) Translate (by hand) the following assembly program to SIC/XE object code.

REF	START	1000
. Subroutine EXCH
EXCH	LDA	@P1
	STA	TEMP
	LDA	@P2
	STA	@P1
	LDA	TEMP
	STA	@P2
	RSUB
P1	RESW	1
P2	RESW	1
TEMP	RESW	1
. Main program
MAIN	LDA	#1
	STA	I
	LDA	#3
	STA	J
. Call a subroutine
	LDA	#I
	STA	P1
	LDA	#J
	STA	P2
	JSUB	EXCH
I	RESW	1
J	RESW	1
	END	MAIN

(10%) By specifying the memory addresses, explain how the "indirect addressing" is utilized in the above subroutine to modify the values of variables in the main program.

(10%) Suppose we made the following changes to the program in Figure 2.9:

Delete the LTORG statement on line 93.
Change the statement on line 45 to +LDA... .
Change the operands on lines 135 and 145 to use literals (and delete line 185).

(10%) In the program of Figure 2.11, suppose we used only two program blocks: the default block and CBLKS. Assume that the data items in CDATA are to be included in the default block. What changes in the source program would accomplish this? Show the length and address of these program blocks as shown in Page 80.

(10%) Consider an extended version of SIC/XE that has a new register R. The contents of R cannot be accessed or changed by the user program. When a program is loaded, however, the loader sets register R so that it contains the starting address of the program. For simplicity, assume that this version of SIC has no program-counter or base relative addressing - thus, all instructions that refer to memory must use Format 4.
Each time the program refers to an address in memory, the contents of register R are automatically added into the target address calculation. Suppose, for example, that an assembled instruction specifies an address of 800 (hexadecimal). If R contains 5000, executing this instruction would actually refer to memory address 5800. If R contains 8000, executing the same instruction would actually refer to memory address 8800.
Consider the control sections shown in Figure 3.8. Assume that these control sections are being loaded and linked at the addresses shown in Figure 3.10; thus the loader will set register R to the value 4000.

What value should appear in the External Symbol Table of the loader for the symbol LISTB?
What should the instruction labeled REF2 in the control section PROGC look like after all loading and linking operations have been performed?

(10%) Suppose we have the macro definition of ABSDIFF as

#define ABSDIFF(X,Y) X > Y ? X - Y : Y - X
#define DISPLAY(EXPR) printf(#EXPR "= %d\n", EXPR)
1. Expand the macro invocation
  DISPLAY( ABSDIFF(3-1, 9+3) );
2. If we execute the C program containing this statement, what output will be produced?

(10%) Refer to the definition of RDBUFF that appears in Figure 4.8(a). Each of the following macro invocation statements contains an error. Which of these errors would be detected by the macro processor, and which would be detected by the assembler?

RDBUFF F3, BUF, RECL, ZZ

{ illegal value specified for &EOR }

RDBUFF F3, BUF, RECL, 04, 2048, 01
- { too many arguments }
RDBUFF F3, ,RECL, 04
- { no value specified for &BUFADR }
RDBUFF F3, RECL, BUF
- { arguments specified in wrong order }