Difference between revisions of "SPO600 64-bit Assembly Language Lab"

From CDOT Wiki
Jump to: navigation, search
 
(46 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
[[Category:SPO600 Labs]][[Category:Assembly Language]]
 
[[Category:SPO600 Labs]][[Category:Assembly Language]]
 
{{Admon/lab|Purpose of this Lab|In this lab, you will experiment with assembler on the x86_64 and aarch64 platforms.}}
 
{{Admon/lab|Purpose of this Lab|In this lab, you will experiment with assembler on the x86_64 and aarch64 platforms.}}
{{Admon/tip|SPO600 Servers|Perform this lab on [[SPO600 Servers]] (you may use your own x86_64 system if desired, along with the AArch64 server).}}
+
{{Admon/tip|SPO600 Servers|Perform this lab on [[SPO600 Servers]] (you may use your own systems if they are of the right architecture and appropriately configured).}}
  
== Lab 3 ==
+
== Lab 4 ==
<!--
 
  
### THIS COMMENTED-OUT SECTION DESCRIBES THE
 
### CONFIGURATION USED FOR THE WINTER 2014
 
### OFFERING OF THE SPO600 COURSE, WHERE THE
 
### AARCH64 WORK WAS DONE IN EMULATION ALONGSIDE
 
### THE X86_64 WORK ON THE INTEL HOST "IRELAND".
 
### IN FALL 2014, AARCH64 HARDWARE WAS AVAILABLE,
 
### AND IRELAND HAD FAILED, SO WE SWITCHED TO
 
### THOSE HOSTS.
 
 
=== Ireland - Configuration ===
 
 
The host ''Ireland'' (ireland.proximity.on.ca) has been set up so that you can use it normally as an x86_64 host, or [[SPO600 aarch64 QEMU on Ireland|use an emulation environment to build and run aarch64 binaries]].
 
 
The directory <code>~/arm64/spo600/examples</code>, which is also accessible as <code>~/spo600-examples</code>, contains these files:
 
 
── hello                        # 'hello world' example programs
 
    ├── assembler
 
    │  ├── aarch64              # aarch64 assembler version
 
    │  │  ├── hello.s
 
    │  │  └── Makefile
 
    │  └── x86_64                # x86_64 assembler versions
 
    │      ├── hello-gas.s      # 64-bit instructions with AT&T/gnu assembler syntax (called 'gas', /usr/bin/as)
 
    │      ├── hello-nasm.s      # 32-bit instructions with Intel/nasm assembler syntax (/usr/bin/nasm)
 
    │      └── Makefile
 
    └── c
 
        ├── hello2.c              # C version using the write() syscall wrapper
 
        ├── hello.c              # C version using printf()
 
        └── Makefile
 
 
Throughout this lab, take advantage of ''[[make and Makefiles|make]]'' whenever possible.
 
 
-->
 
 
=== Code Examples ===
 
=== Code Examples ===
  
The code examples for this lab are available at either of these links:
+
The code examples for this lab are available in the file <code>/public/spo600-assembler-lab-examples.tgz</code> on each of the [[SPO600 Servers]].
* Outside Seneca: http://england.cdot.systems/spo600/spo600-assembler-lab-examples.tgz
 
* Inside Seneca: http://england.internal.cdot.systems/spo600/spo600-assembler-lab-examples.tgz
 
 
 
Please download this archive to your accounts on the x86_64 and AArch64 systems (<code>wget</code> is a good way to do this), and unpack the archive on both systems.
 
  
 
Unpacking the archive in your home directory will produce the following directory structure:
 
Unpacking the archive in your home directory will produce the following directory structure:
 
 
  spo600
 
  spo600
  `-- examples
+
  └── examples
     `-- hello                    # "hello world" example programs
+
     └── hello                    # "hello world" example programs
         |-- assembler
+
         ├── assembler
         |  |-- aarch64          # aarch64 gas assembly language version
+
         │   ├── aarch64          # aarch64 gas assembly language version
         |  |  |-- hello.s
+
         │   │   ├── hello.s
         |  |  `-- Makefile
+
         │   │   └── Makefile
         |  `-- x86_64            # x86_64 assembly language versions
+
        │   ├── Makefile
         |      |-- hello-gas.s  # ... gas syntax
+
         │   └── x86_64            # x86_64 assembly language versions
         |      |-- hello-nasm.s  # ... nasm syntax
+
         │      ├── hello-gas.s  # ... gas syntax
         |      `-- Makefile
+
         │      ├── hello-nasm.s  # ... nasm syntax
         `-- c                    # Portable C versions
+
         │      └── Makefile
             |-- hello2.c          # syscall wrapper version
+
         └── c                    # Portable C versions
             |-- hello.c          # printf version
+
             ├── hello2.c          # ... using write()
             `-- Makefile
+
            ├── hello3.c          # ... using syscall()
 +
             ├── hello.c          # ... using printf()
 +
             └── Makefile
  
 
Throughout this lab, take advantage of ''[[make and Makefiles|make]]'' whenever possible.
 
Throughout this lab, take advantage of ''[[make and Makefiles|make]]'' whenever possible.
  
=== References ===
+
=== Resources ===
* [[Assembler Basics]]
+
* [[Assembler Basics]] (includes instructions on how to use the GNU Assembler)
 +
* [[Syscalls]]
 
* [[x86_64 Register and Instruction Quick Start]]
 
* [[x86_64 Register and Instruction Quick Start]]
 
* [[aarch64 Register and Instruction Quick Start]]
 
* [[aarch64 Register and Instruction Quick Start]]
  
=== Group Lab Tasks ===
+
=== Optional Investigation ===
 +
 
 +
1. Build and run the three C versions of the program for x86_64 and aarch64, using <code>make</code>. Take a look at the differences in the code.
 +
 
 +
2. Use the <code>objdump -d</code> command to dump (print) the object code (machine code) and disassemble it into assembler for each of the binaries. Find the <code><nowiki><main></nowiki></code> section and take a look at the code. Also notice the total amount of code.
  
1. Build and run the two C versions of the program for x86_64. Take a look at the differences in the code.
+
3. Review, build, and run the x86_64 assembly language programs using <code>make</code>, taking note of the commands that are executed to assemble and link the code. Take a look at the code using <code>objdump -d '''objectfile'''</code> and compare it to the source code. Notice the absence of other code (compared to the C binary, which had a lot of extra code).
  
2. Use the <code>objdump -d</code> command to dump the object code (machine code) for the two binaries. Find the <code><nowiki><main></nowiki></code> section and make sure you understand the code.
+
4. Build and run the assembly language version of the program for aarch64 using <code>make</code>, taking note of the commands that are executed to assemble and link the code. Verify that you can disassemble the object code in the ELF binary using <code>objdump -d ''objectfile''</code> and take a look at the code.
  
3. Review, build, and run the x86_64 assembly language programs. Make sure you understand the code.
+
=== Lab Tasks ===
  
4. Build and run the two C versions of the program for aarch64. Verify that you can disassemble the object code in the ELF binary using <code>objdump -d</code>
+
<!-- {{Admon/tip|Answers in the Video!|The answers to the first three steps below are contained in the associated [https://web.microsoftstream.com/video/8c3c1353-5729-4217-b1ba-371410f14ad4 lecture video.]}} -->
  
5. Review, build, and run the aarch64 assembly language programs. Make sure you understand the code.
+
1. Review, build, and run the aarch64 assembly language programs. Take a look at the code using <code>objdump -d '''objectfile'''</code> and compare it to the source code.
  
6. Here is a basic loop in x86_64 assembler - this loops from 0 to 9, using r15 as the index (loop control) counter:
+
2. Here is a basic loop in AArch64 assembler - this loops from 0 to 9, using r19 as the index (loop control) counter:
  
 
  .text
 
  .text
  .globl   _start
+
  .globl _start
 
   
 
   
  start = 0                       /* starting value for the loop index; note that this is a symbol (constant), not a variable */
+
  min = 0                         /* starting value for the loop index; '''note that this is a symbol (constant)''', not a variable */
  max = 10                       /* loop exits when the index hits this number (loop condition is i<max) */
+
  max = 10                         /* loop exits when the index hits this number (loop condition is i<max) */
 
   
 
   
 
  _start:
 
  _start:
     mov    $start,%r15        /* loop index */
+
 +
     mov    x19, min
 
   
 
   
 
  loop:
 
  loop:
    /* ... body of the loop ... do something useful here ... */
 
 
   
 
   
     inc    %r15                /* increment index */
+
     /* '''... body of the loop ... do something useful here ...''' */
     cmp    $max,%r15          /* see if we're done */
+
     jne    loop                /* loop if we're not */
+
    add    x19, x19, 1
 +
     cmp    x19, max
 +
     b.ne    loop
 
   
 
   
     mov    $0,%rdi            /* exit status */
+
     mov    x0, 0           /* status -> 0 */
     mov    $60,%rax            /* syscall sys_exit */
+
     mov    x8, 93          /* exit is syscall #93 */
     syscall
+
     svc    0              /* invoke syscall */
 +
 
 +
This code doesn't actually do anything while looping, because the body of the loop is empty. On an AArch64 machine, combine this code with code from the "Hello World" assembley-language example, so that it prints a word each time it loops:
 +
 
 +
Loop
 +
Loop
 +
Loop
 +
Loop
 +
Loop
 +
Loop
 +
Loop
 +
Loop
 +
Loop
 +
Loop
  
Extend this code, combining it with code from the "Hello World" example, so that it prints each digit from 0 to 9 like this:
+
Then modify the message so that it includes the loop index values, showing each digit from 0 to 9 like this:
  
 
  Loop: 0
 
  Loop: 0
Line 118: Line 102:
 
  Loop: 9
 
  Loop: 9
  
{{Admon/tip|Character conversion|In order to print the loop index value, you will need to convert from an integer to digit character. In ASCII/ISO-9959-1/Unicode UTF-8, the digit characters are in the range 48-57 (0x30-0x39). You will also need to assemble the message to be printed for each line - you can do this by writing the digit into the message buffer before outputting it to stdout, which is probably the best approach, or you can perform a sequence of writes for the thee parts of the message ('Loop: ', number, '\n'). You may want to refer to the manpage for <code>ascii</code>.}}
+
{{Admon/tip|Character conversion|In order to print the loop index value, you will need to convert from an integer to digit character. In ASCII/ISO-8859-1/Unicode UTF-8, the digit characters are in the range 48-57 (0x30-0x39). You will also need to assemble the message to be printed for each line - you can do this by writing the digit into the message buffer before outputting it to stdout, which is probably the best approach, or you can perform a sequence of writes for the thee parts of the message ('Loop: ', number, '\n'). You may want to refer to the manpage for <code>ascii</code>.}}
 +
 
 +
{{Admon/tip|6502 Implementation|For reference, here is a [[6502 Counting Loop Example|6502 implementation of this loop]].}}
 +
 
 +
3. Repeat the previous step for x86_64.
 +
 
 +
For reference, here is the loop code in x86_64 assembler:
 +
 
 +
.text
 +
.globl    _start
 +
 +
min = 0                        /* starting value for the loop index; '''note that this is a symbol (constant)''', not a variable */
 +
max = 10                        /* loop exits when the index hits this number (loop condition is i<max) */
 +
 +
_start:
 +
    mov    $min,%r15          /* loop index */
 +
 +
loop:
 +
    /* '''... body of the loop ... do something useful here ...''' */
 +
 +
    inc    %r15                /* increment index */
 +
    cmp    $max,%r15          /* see if we're done */
 +
    jne    loop                /* loop if we're not */
 +
 +
    mov    $0,%rdi            /* exit status */
 +
    mov    $60,%rax            /* syscall sys_exit */
 +
    syscall
  
7. Repeat step 6 for aarch64.
+
4. Extend the AArch64 code to loop from 00-30, printing each value as a 2-digit decimal number.
  
8. Extend the code to loop from 00-30, printing each value as a 2-digit decimal number.
+
{{Admon/tip|2-Digit Conversion|You will need to take the loop index and convert it to a 2-digit decimal number by dividing by 10. Read the description of the division instruction carefully. On x86_64, you need to set up specific registers before performing a division. On AArch64, you will need to use a second instruction to find the remainder after a division.}}
  
{{Admon/tip|2-Digit Conversion|You will need to take the loop index and convert it to a 2-digit decimal number by dividing by 10. To do this, use the <code>div</code> instruction, which takes the dividend from rax and the divisor from register supplied as an argument. The quotient will be placed in rax and the remainder will be placed in rdx.}}
+
5. Change the code as needed to suppress the leading zero (printing 0-30 instead of 00-30).
  
9. Repeat step 8 for aarch64.
+
5. Repeat the previous two steps for x86_64.
  
 
=== Deliverables ===
 
=== Deliverables ===
  
1. Complete the group lab section, above.
+
1. Complete the lab section, above.
 
 
2. Extend the assembler programs (both x86_64 and aarch64) to suppress the high digit when it is 0. In other words, the printed values should progress from 0-30 instead of from 00-30. It is OK to output a space in place of the suppressed digit (this will cause the numbers to be aligned vertically in the output).
 
  
3. Blog about the programs you've written. Describe the experience of writing and debugging in assembler, as compared to writing in other languages. Contrast x86_64 and aarch64 assembler, your experience with each, and your opinions of each. Include links to the source code for both of your assembler programs.
+
2. Blog about the programs you've written. Describe the experience of writing and debugging in assembler, as compared to writing in other languages. Contrast x86_64 and aarch64 assembler, your experience with each, and your opinions of each. Include links to the source code for each of your assembler programs.
  
 
=== Optional Challenge ===
 
=== Optional Challenge ===

Latest revision as of 21:56, 5 October 2022

Lab icon.png
Purpose of this Lab
In this lab, you will experiment with assembler on the x86_64 and aarch64 platforms.
Idea.png
SPO600 Servers
Perform this lab on SPO600 Servers (you may use your own systems if they are of the right architecture and appropriately configured).

Lab 4

Code Examples

The code examples for this lab are available in the file /public/spo600-assembler-lab-examples.tgz on each of the SPO600 Servers.

Unpacking the archive in your home directory will produce the following directory structure:

spo600
└── examples
    └── hello                     # "hello world" example programs
        ├── assembler
        │   ├── aarch64           # aarch64 gas assembly language version
        │   │   ├── hello.s
        │   │   └── Makefile
        │   ├── Makefile
        │   └── x86_64            # x86_64 assembly language versions
        │       ├── hello-gas.s   # ... gas syntax
        │       ├── hello-nasm.s  # ... nasm syntax
        │       └── Makefile
        └── c                     # Portable C versions
            ├── hello2.c          # ... using write()
            ├── hello3.c          # ... using syscall()
            ├── hello.c           # ... using printf()
            └── Makefile

Throughout this lab, take advantage of make whenever possible.

Resources

Optional Investigation

1. Build and run the three C versions of the program for x86_64 and aarch64, using make. Take a look at the differences in the code.

2. Use the objdump -d command to dump (print) the object code (machine code) and disassemble it into assembler for each of the binaries. Find the <main> section and take a look at the code. Also notice the total amount of code.

3. Review, build, and run the x86_64 assembly language programs using make, taking note of the commands that are executed to assemble and link the code. Take a look at the code using objdump -d objectfile and compare it to the source code. Notice the absence of other code (compared to the C binary, which had a lot of extra code).

4. Build and run the assembly language version of the program for aarch64 using make, taking note of the commands that are executed to assemble and link the code. Verify that you can disassemble the object code in the ELF binary using objdump -d objectfile and take a look at the code.

Lab Tasks

1. Review, build, and run the aarch64 assembly language programs. Take a look at the code using objdump -d objectfile and compare it to the source code.

2. Here is a basic loop in AArch64 assembler - this loops from 0 to 9, using r19 as the index (loop control) counter:

.text
.globl _start

min = 0                          /* starting value for the loop index; note that this is a symbol (constant), not a variable */
max = 10                         /* loop exits when the index hits this number (loop condition is i<max) */

_start:

    mov     x19, min

loop:

    /* ... body of the loop ... do something useful here ... */

    add     x19, x19, 1
    cmp     x19, max
    b.ne    loop

    mov     x0, 0           /* status -> 0 */
    mov     x8, 93          /* exit is syscall #93 */
    svc     0               /* invoke syscall */

This code doesn't actually do anything while looping, because the body of the loop is empty. On an AArch64 machine, combine this code with code from the "Hello World" assembley-language example, so that it prints a word each time it loops:

Loop
Loop
Loop
Loop
Loop
Loop
Loop
Loop
Loop
Loop

Then modify the message so that it includes the loop index values, showing each digit from 0 to 9 like this:

Loop: 0
Loop: 1
Loop: 2
Loop: 3
Loop: 4
Loop: 5
Loop: 6
Loop: 7
Loop: 8
Loop: 9
Idea.png
Character conversion
In order to print the loop index value, you will need to convert from an integer to digit character. In ASCII/ISO-8859-1/Unicode UTF-8, the digit characters are in the range 48-57 (0x30-0x39). You will also need to assemble the message to be printed for each line - you can do this by writing the digit into the message buffer before outputting it to stdout, which is probably the best approach, or you can perform a sequence of writes for the thee parts of the message ('Loop: ', number, '\n'). You may want to refer to the manpage for ascii.
Idea.png
6502 Implementation
For reference, here is a 6502 implementation of this loop.

3. Repeat the previous step for x86_64.

For reference, here is the loop code in x86_64 assembler:

.text
.globl    _start

min = 0                         /* starting value for the loop index; note that this is a symbol (constant), not a variable */
max = 10                        /* loop exits when the index hits this number (loop condition is i<max) */

_start:
    mov     $min,%r15           /* loop index */

loop:
    /* ... body of the loop ... do something useful here ... */

    inc     %r15                /* increment index */
    cmp     $max,%r15           /* see if we're done */
    jne     loop                /* loop if we're not */

    mov     $0,%rdi             /* exit status */
    mov     $60,%rax            /* syscall sys_exit */
    syscall

4. Extend the AArch64 code to loop from 00-30, printing each value as a 2-digit decimal number.

Idea.png
2-Digit Conversion
You will need to take the loop index and convert it to a 2-digit decimal number by dividing by 10. Read the description of the division instruction carefully. On x86_64, you need to set up specific registers before performing a division. On AArch64, you will need to use a second instruction to find the remainder after a division.

5. Change the code as needed to suppress the leading zero (printing 0-30 instead of 00-30).

5. Repeat the previous two steps for x86_64.

Deliverables

1. Complete the lab section, above.

2. Blog about the programs you've written. Describe the experience of writing and debugging in assembler, as compared to writing in other languages. Contrast x86_64 and aarch64 assembler, your experience with each, and your opinions of each. Include links to the source code for each of your assembler programs.

Optional Challenge

Write a program in aarch64 assembly language to print the times tables from 1-12 ("1 x 1 = 1" through "12 x 12 = 144"). Add a spacer between each table, and use a function/subroutine to format the numbers with leading-zero suppression.

The output could look something like this:

 1 x  1 =   1
 2 x  1 =   2
 3 x  1 =   3
 4 x  1 =   4
 5 x  1 =   5
 6 x  1 =   6
 7 x  1 =   7
 8 x  1 =   8
 9 x  1 =   9
10 x  1 =  10
11 x  1 =  11
12 x  1 =  12
-------------
 1 x  2 =   2
 2 x  2 =   4
 3 x  2 =   6
 4 x  2 =   8
 5 x  2 =  10

  ...lines snipped for space... 

11 x 12 = 132
-------------
 1 x 12 =  12
 2 x 12 =  24
 3 x 12 =  36
 4 x 12 =  48
 5 x 12 =  60
 6 x 12 =  72
 7 x 12 =  84
 8 x 12 =  96
 9 x 12 = 108
10 x 12 = 120
11 x 12 = 132
12 x 12 = 144