Computer Architecture and Assembly Language Assignment 2

Programming environment
For this assignment you must ensure your work executes correctly on Arduino
boards in ECS 249. If you have installed AVR Studio on your own computer then you
are welcome to do much of the programming work on your machine. If this is the
case, however, then you must allow enough time to ensure your solutions work on
the lab machines. If your submission fails to work on a lab machine, the fault is very
rarely that the lab workstations. “It worked on my machine!” will be treated as the
equivalent of “The dog ate my homework!”
Individual work
This assignment is to be completed by each individual student (i.e., no group work).
Naturally you will want to discuss aspects of the problem with fellow students, and
such discussion is encouraged. However, sharing of code fragments is strictly
forbidden without the express written permission of the course instructor
(Zastre). If you are still unsure regarding what is permitted or have other questions
about what constitutes appropriate collaboration, please contact me as soon as
possible. (Code-similarity analysis tools will be used to examine submitted work.)
The URLs of significant code fragments you have found and used in your solution
must be cited in comments just before where such code has been used.
Objectives of this assignment
• Write and use functions.
• Implement parameter passing using registers and stack.
• Implement return values using registers.
• Use stack frames where needed for parameter passing.
• Use the Arduino mega2560 board in the course lab to implement a Morse
code display.
Page 2 of 7
Morse code
Morse code is a method for transmitting messages over a distance by using a
combination of long and short signals (i.e., long or short light flashes, long or short
electrical pulses, etc.) Until recently, the ability to send and receive Morse code
messages was an important skill for anyone involved in communication over radio
(such as airplane pilots, military personnel, amateur-radio operators, etc.). For a bit
more about Morse code you can read the Wikipedia article at:
https://en.wikipedia.org/wiki/Morse_code
Our assignment will use International Morse Code to display a message using the 26
letters of the English alphabet. A demonstration can be found at:
https://youtu.be/PwtoJwlKhdw
The starter file is named a2_morse.asm. There is only one file for this assignment.
Your work will be in five parts, ordered from easy to more difficult:
a) Write the functions leds_on and leds_off
b) Write the function morse_flash
c) Write the function flash_message
d) Write the function letter_to_code
e) Write the function encode_message
Part (a): Write functions leds_on and leds_off
leds_on:
parameters: one in r16, pass-by-value
return value: none
leds_off:
parameters: none
return value: none
The parameter to leds_on determines how many of the Ardujno board’s six LEDs to
turn on (i.e., the ones which you would have used during some lab exercises).
Therefore if 6 is stored in r16, all six LEDs will be turned on. If 2 is stored in r16,
then only two LEDs will be turned on (and you can choose which two those are).
leds_off turns off all LEDs. We could eliminate leds_off by saying that storing
zero into r16 and calling leds_on turns off all LEDs. However, this assumes leds_on
is correctly implemented. To reduce your frustration, please implement leds_off.
Page 3 of 7
Part (b): Write the function morse_flash
morse_flash:
parameters: one in r16, pass-by-value
return value: none
Later in part (d) of this assignment, you will write code to convert dots and dashes
for a letter (e.g., “. . .” for ‘S’, and “- - -” for ‘O’) into a one-byte equivalent. For now,
however, you are to write a function that takes a one-byte equivalent (pass into the
function within a register) and flashes the LEDs appropriate to the contents of that
byte. The function doesn’t need to know the letter to be flashed; all it needs is the
byte.
Below are examples of dot-dash sequences beside their one-byte equivalents:
Each byte consists of a high nybble and a low nybble:
• The high nybble (left-most four bits) encodes the length of the sequence.
• The low nybble (right-most four bits) encodes the dot-dash sequence itself
(where 0 is a “dot” and 1 is a “dash”).
In the first example above (the Morse for ‘F’), the high nybble encodes the number 4
(i.e., the length of the sequence), and the low nybble contains that sequence (0, 0, 1,
0). In the last example (the Morse for ‘T’), the high nybble encodes the number 1
(i.e., the length of the sequence), and the low nybble contains that sequence (first
three 0s are ignored, while last 1 is the dash). Notice that the low-nybble bits for
sequences of length three, two and one will contain leading zeros; these leading
zeros must be ignored (i.e., they are not dots).
The one-byte equivalent is to be turned into a series of calls to leds_on and
leds_off, with a delay between calls visually distinguishing dots from dashes:
. . — . 0b01000010
— — — 0b00110111
. — . 0b00110010
— 0b00010001
dot dot
off
on
dash
delay_long
time
delay_short
Page 4 of 7
The previous diagram shows the timing pattern for “. . – .” The functions
delay_long and delay_short are already provided for you in a2_morse.asm1. So a
dot would be similar (in a pseudo-code manner) to:
leds_on
delay_short
leds_off
delay_long
while a dash would be:
leds_on
delay_long
leds_off
delay_long
Notice at the end of the timing diagram there appears to be an additional
delay_long with the LEDs still off; this visually separates Morse code sequences.
There is a special one-byte value: 0xff. It represents a space between words. For
this, morse_flash must keep the LEDs off for three calls to delay_long.
Tip: When writing your implementation of morse_flash you may find the swap
opcode helpful.
Part (c): Write the function flash_message
flash_message:
parameters: encoded-message address on stack (two bytes);
in effect pass-by-reference.
return value: none
Encoded messages are stored in SRAM and consists of a sequence of one-byte
equivalents. A sequence is terminated by the 0 value (i.e., null, just as in ending a
string). For example, the one-byte equivalent sequence for “SOS” consists of the
following four bytes:
0x30 0x37 0x30 0x00
The code provided for you in a2_morse.asm stores this sequence at the
TESTADDRESS memory location.
1 The implementations of delay_long and delay_short are rather unsatisfying as
they are just wrappers for repeated calls to delay and delay_busywait. In later
assignments we will see how using timers and interrupts yield much more elegant
implementations of such delays. For this assignment, however, your code must use
only delay_long and delay_short.
Page 5 of 7
This function must use the address passed on the stack as the starting location of the
sequence, and then loop through that sequence, calling morse_flash for each byte,
until encountering the end of the sequence. Once the sequence ends, the function
may return. Given the form of parameter passing used, you will implement a stackframe-like caller and callee sequence of instructions.
Part (d): Write the function letter_to_code
letter_to_code:
parameters: letter on stack for conversion to one-byte equivalent; pass-by-value
return value: one-byte equivalent stored in r0
This may seem the most complicated of the assignment functions. It must:
• Obtain the letter to be converted from the stack.
• Locate in the table of codes the dot-dash sequence for the letter (given to you
in a2_morse.asm, starting at ITU_MORSE in program memory).
• Convert the dots (‘.’) and dashes (‘–‘) into the one-byte equivalent for the
letter.
However, a couple of features of characters will help us here. For example, our
letters are actually bytes (that is, ASCII characters) which can be directly compared
with other bytes. Consider for example:
ldi r16, 65
ldi r17, 'A' ; must use single quotes!!!
cp r16, r17
The comparison will set the Z flag. Why? Because the ASCII code for ‘A’ is 65. Note,
however, that the ASCII code for ‘a’ is 97. You may assume this function will be
given messages with only upper-case letters (and spaces) for encoding.
Another assist for you is that the table is aligned on an eight-byte boundary.
Therefore even though some letters have longer Morse code sequences than others,
each table entry has the same length. To indicate the end of a letter’s sequence we
use our good friend, zero. For example, here is a snippet from the table:
...
.db "N", "-.", 0, 0, 0, 0, 0
.db "O", "---", 0, 0, 0, 0
.db "P", ".--.", 0, 0, 0
....
The sequence for “N” consists of a dash and a dot; that for “O” is three dashes; that
for “P” a dot, dash, dash, and dot. All, however, end with one or many zeros such that
each line consists of exactly eight bytes of data.
Page 6 of 7
The roughest of pseudocode solutions is shown below to suggest an implementation
strategy. The operation mem[Z] refers to memory access (either loading or storing,
either SRAM or program memory) through using the Z register as the address.
Z = ITU_MORSE
while (mem[Z] != 0)
if mem[Z] equals letter-to-be-converted:
Z = Z + 1
while mem[Z] != 0:
do something if mem[Z] is a dot or
do something else if mem[Z] is a dash
Z = Z + 1
finished (i.e., break out of outermost loop)
Z = Z + 8
// At this point the encoding of letter is complete
Tip: Remember that ld, lds, and ldd are used for SRAM, while lpm is used for
program memory (i.e., where ITU_MORSE is stored). Also remember that the AVR
assembler treats program memory addresses as word addresses, so we must leftshift by one to convert such address by byte addresses.
Advanced tip: Strictly speaking the outermost loop is not needed, but only because of
the byte-alignment used for the table. However, an acceptable and correct solution
for this part (d) may include and outermost loop.
Part (e): Write the function encode_message
encode_message:
parameters: address of message (i.e., sequence of upper-case letters plus spaces)
to be encoded into one-byte-equivalent sequene; address of buffer in SRAM in
which one-byte equivalent sequence is to be stored;
both addresses pushed onto the stack;
both parameters, in effect, are passed-by-reference
return value: none
This function is quite straightforward. The only tricky bit will be the correct use of
the stack frame, and any bugs in letter_to_code – especially the way registers are
saved and restored – can cause problems in encode_message. That said, all this
function needs to do is (1) read each character in the original message, and for each
character (2) convert it into one-byte equivalent by calling letter_to_morse and
storing its result into the buffer. (A message is terminated with a zero.)
Once encode_message is completed, displaying the Morse code for a text message M
made up of uppercase letters and spaces requires two steps:
1. encode the message M into some buffer B
Page 7 of 7
2. call flash_message with the buffer B passed as a parameter
Given the form of parameter passing used, you will implement a stack-frame-like
caller and callee sequence of instructions.
What you must submit
• Your completed work in the single source code file (a2_morse.asm); do not
change the name of this file!
• Your work must use the provided skeleton a2_morse.asm. Any other kinds of
solutions will not be accepted.
Evaluation
• 1 marks: Solution for leds_on and leds_off
• 4 mark: Solution for morse_flash
• 4 marks: Solution for flash_message
• 7 marks: Solution for letter_to_morse
• 4 mark: Solution for encode_message
Therefore the total mark for this assignment is 20.
Some of the evaluation above will also take into account whether or not submitted
code is properly formatted (i.e., indenting and commenting are suitably used), and
the file correctly named.