Lab 3 -- for loops with an incrementing or decrementing index

By now, you all are familiar with logging into SNOWBALL account, compiling a C program to get an assembly language program. In the next exercise we will be looking at how a for loop would look in assembly language.

First, log in to your account on SNOWBALL. You will use the command "ssh your_account@snowball.cs.gsu.edu" 


cascade:Desktop> ssh snowball.cs.gsu.edu
mweeks@snowball.cs.gsu.edu's password:
Last login: Mon Dec 4 12:25:41 2023 from 131.96.221.170
+
| GSU Computer Science
| Instructional Server
| SNOWBALL.cs.gsu.edu
+

[mweeks@gsuad.gsu.edu@snowball ~]$ vi loop.c
Next, get the "loop.c" file. You could copy and paste it from this assignment, or you could even type it yourself. We are looking at how a "for" loop works here. In the "for" loop, we have initialization statement, test expression and update statement. Initialization statement is executed only once. Then, test expression is evaluated. If the test expression is evaluated to false then "for" loop is terminated. However if the test expression is evaluated to true then, statements inside the "for" loop body are executed and the update expression is updated. Again the test expression is evaluated to continue for the next iteration. This process goes on until the test expression is false. When the test expression is false, the loop terminates.

If you have not started the log, start it now. 


[mweeks@gsuad.gsu.edu@snowball ~]$ script lab3.log
Script started, file is lab3.log
Next, we use the "cat" command to display the contents of the "loop.c" file. 

[mweeks@gsuad.gsu.edu@snowball ~]$ cat loop.c
// Print numbers from 0 to 5
#include <stdio.h>

int main() {
  int i;

  for (i = 0; i < 6; i++) {
    printf("%d ", i);
  }
  printf("\n");
  return 0;
}
In the next step, we compile the "loop.c" program. "gcc" is the GNU C compiler, and "-o loop" specifies that we want the output to be a new file called "loop". Be careful with this step; if you were to type "-o loop.c", it would over-write the C file, and you would have to start over. 

[mweeks@gsuad.gsu.edu@snowball ~]$ gcc loop.c -o loop
The file "loop" is executable, and so we run it in the next command. The "./" specifies where to find the file, and "." is short-hand for the current directory. 

[mweeks@gsuad.gsu.edu@snowball ~]$ ./loop
0 1 2 3 4 5
The program ran as expected, and gave us the line of output. Here, in the program "i" is the initialization variable assigned with 0. As the test expression is i<6, the "for" loop executed from 0 to 5, incrementing i value at every iteration. We gave the update statement as i++, which means i=i+1. At i=6, i<6 expression failed as 6 is not less than 6, so "for" loop terminated. As we see, the output is 0 1 2 3 4 5. So far, we have obtained a C program, showed its contents, compiled it, and ran it.

Next, we will compile it again. Notice the "-S" (and yes, that is a capital S). This generates a new file, "loop.s". We then "cat" its contents, like before. 


[mweeks@gsuad.gsu.edu@snowball ~]$ gcc loop.c -S
[mweeks@gsuad.gsu.edu@snowball ~]$ cat loop.s
	.file	"loop.c"
	.section	.rodata
.LC0:
	.string	"%d "
	.text
	.globl	main
	.type	main, @function
main:
.LFB0:
	.cfi_startproc
	pushq	%rbp
	.cfi_def_cfa_offset 16
	.cfi_offset 6, -16
	movq	%rsp, %rbp
	.cfi_def_cfa_register 6
	subq	$16, %rsp
	movl	$0, -4(%rbp)
	jmp	.L2
.L3:
	movl	-4(%rbp), %eax
	movl	%eax, %esi
	movl	$.LC0, %edi
	movl	$0, %eax
	call	printf
	addl	$1, -4(%rbp)
.L2:
	cmpl	$5, -4(%rbp)
	jle	.L3
	movl	$10, %edi
	call	putchar
	movl	$0, %eax
	leave
	.cfi_def_cfa 7, 8
	ret
	.cfi_endproc
.LFE0:
	.size	main, .-main
	.ident	"GCC: (GNU) 4.8.5 20150623 (Red Hat 4.8.5-44)"
	.section	.note.GNU-stack,"",@progbits
At this point, we have an assembly-language program for the "for" loop.

You may wonder what this code does: 


        movl    -4(%rbp), %eax
        movl    %eax, %esi
        movl    $.LC0, %edi
        movl    $0, %eax
        call    printf
Effectively, it prints out the number stored in a local variable.
"movl -4(%rbp), %eax" copies the local variable's value to EAX.
"movl %eax, %esi" copies the value from EAX to ESI. Why not copy it directly? It looks like you can.
"movl $.LC0, %edi" copies the label's memory location .LC0 to EDI.
"movl $0, %eax" sets EAX to 0.
"call printf" then calls the function printf to output the value that ESI has, with the format specified by EDI.

As an exercise, we can try changing the test expression to see decrementing expression which is i--, which is evaluated as i=i-1. Create a loop_decrement.c file. You can add program contents as below: 


[mweeks@gsuad.gsu.edu@snowball ~]$ vi loop_decrement.c
[mweeks@gsuad.gsu.edu@snowball ~]$ cat loop_decrement.c
// Print numbers from 5 to 1
#include <stdio.h>

int main() {
  int i;

  for (i = 5; i > 0; i--) {
    printf("%d ", i);
  }
  printf("\n");
  return 0;
}
As you can see, here we are initializing i with 5 and then decrementing its value with every iteration. Here, the test expression is i>0 so the loop will continue from 5 to 1 until i=1. As the test expression is i>0 (means 1>0) after the final iteration, the test expression is false so the loop terminates. Thus, the output is 5 4 3 2 1.

Next, we can follow the similar steps as above to compile using "gcc" and executing it. 


[mweeks@gsuad.gsu.edu@snowball ~]$ gcc loop_decrement.c -o loop_decrement
[mweeks@gsuad.gsu.edu@snowball ~]$ ./loop_decrement
5 4 3 2 1
[mweeks@gsuad.gsu.edu@snowball ~]$ gcc loop_decrement.c -S
[mweeks@gsuad.gsu.edu@snowball ~]$ cat loop_decrement.s
	.file	"loop_decrement.c"
	.section	.rodata
.LC0:
	.string	"%d "
	.text
	.globl	main
	.type	main, @function
main:
.LFB0:
	.cfi_startproc
	pushq	%rbp
	.cfi_def_cfa_offset 16
	.cfi_offset 6, -16
	movq	%rsp, %rbp
	.cfi_def_cfa_register 6
	subq	$16, %rsp
	movl	$5, -4(%rbp)
	jmp	.L2
.L3:
	movl	-4(%rbp), %eax
	movl	%eax, %esi
	movl	$.LC0, %edi
	movl	$0, %eax
	call	printf
	subl	$1, -4(%rbp)
.L2:
	cmpl	$0, -4(%rbp)
	jg	.L3
	movl	$10, %edi
	call	putchar
	movl	$0, %eax
	leave
	.cfi_def_cfa 7, 8
	ret
	.cfi_endproc
.LFE0:
	.size	main, .-main
	.ident	"GCC: (GNU) 4.8.5 20150623 (Red Hat 4.8.5-44)"
	.section	.note.GNU-stack,"",@progbits
We have assembly language program of "for" loop with decrementing test expression. Notice the difference in assembly language code for incrementing and decrementing test expressions. You can see this plainly with the "diff" command, as below. 

[mweeks@gsuad.gsu.edu@snowball ~]$ diff loop.s loop_decrement.s
1c1
< 	.file	"loop.c"
---
> 	.file	"loop_decrement.c"
17c17
< 	movl	$0, -4(%rbp)
---
> 	movl	$5, -4(%rbp)
25c25
< 	addl	$1, -4(%rbp)
---
> 	subl	$1, -4(%rbp)
27,28c27,28
< 	cmpl	$5, -4(%rbp)
< 	jle	.L3
---
> 	cmpl	$0, -4(%rbp)
> 	jg	.L3
Notice what the "diff" command outputs. The line "1c1" means to change line 1. The line after has < meaning that it is how it looks in the first file. Then "---" separates the entries. The line after that starts with > meaning that it is how it looks in the second file. Taking the first 4 lines together, we can say that the text in file 1 is "loop" while the text in file 2 has "loop_decrement" instead, for that line.

Looking at the differences, explain the others. You can write your explanations as a text file. You should recall that "movl" is a "long" move, that moves a 32 bit value to a 64 bit register, with zeros for the high bits. Similarly, "addl", "subl", and "cmpl" are the "long" versions of "add", "sub" (subtract) and "cmp" (compare). The dollar sign ("$") is used to indicate an immediate value, e.g. $7 simply means the number 7. ".L3" is a label, specifiing a point in the code to go to. "jle" means "jump less than or equal to" while "jg" is for "jump if greater".

Next, run the original "loop" program. Then run the "loop_decrement" program. 


[mweeks@gsuad.gsu.edu@snowball ~]$ ./loop
1 2 3 4 5
[mweeks@gsuad.gsu.edu@snowball ~]$ ./loop_decrement
5 4 3 2 1
As an exercise, create assembly language code for the tasks below. You are welcome to copy and change the .s file directly, especially for the first one. For the second and third, you may want to try changing the C file to get the assembly language code. This requires changing the logic in the "for" loop body to print only even/odd numbers. To check if a number is even we can use expression <number>/2. If expression evaluates to 0 then it is even number else, odd number.

Use in-line comments within the programs (starting with the # character) to say what lines you changed. Show the contents of the .s programs from above (use "cat"). Also, compile them, and run them to show that they work.

Finally, we exit from the script command. This informs the "script" utility that we are done. 


[mweeks@gsuad.gsu.edu@snowball ~]$ exit
exit
Script done, file is lab3.log
You might want to next use "cat" on the "lab3.log" file, copy the output, and paste it into a text file called "lab3.txt". This should get rid of any extraneous (control) characters, though you should look it over to make sure. There are other ways to do this: one would be to make your own filtering program, that only allows characters with ASCII values 32 to 126, along with the newline (linefeed) character (10). Tab, character 9, might be useful too, maybe mapped to 4 spaces.

Then we exit (log out) from the server. You can use "sftp" to get your log file from the server. 


[mweeks@gsuad.gsu.edu@snowball ~]$ exit
logout
Connection to snowball.cs.gsu.edu closed.
In this lab, we have seen :