ECE Experiment 6: Stack Operations, Subroutines, and Memory Solution

Instructional Objectives:

To understand the way that the stack is implemented to save register values. To use the stack to pass information to and from a subroutine.

To understand the way that banked memory and paging is implemented in the HC(S)12 microcontroller

Introduction:

Previous lab procedures have required that students use the HC(S)12 instruction set to write code to interface with the lab I/O board. However, most programming is not done in assembly because of its lack of portability from one processor to another. One drawback of using higher level languages is not being able to observe the several uses of the stack. The purpose of the following experiment is to give students a better appreciation for the tasks the stack fulfills in a program, as well as a better understanding of the way that memory is structured in the HC(S)12 used in lab.

Stack Manipulation:

The stack is a Last In First Out data structure that programmers use to temporarily store and retrieve data. In previous experiments, the stack has been implemented in short delay subroutines, but it is important to understand how to manipulate the stack, as well as what it’s functions are in a programming environment.




Note: Remember that the way to properly initialize the stack pointer for the s12e128 used in lab is to use the constant __SEG_END_SSTACK that is referenced at the top of a new program. This is a value that is generated by the linker file. Its value is whatever the starting address of RAM is plus the value of STACKSIZE. On default, this value is 0x0500, but this value can be changed to increment or decrement the size of the stack by changing the value of STACKSIZE in the linker file.



On the next page, the several different uses of the stack will be explored in depth. The stack has five primary functions in the programming environment:

Saving the return address for properly returning from a subroutine or a function. Temporarily saving registers for a subroutine or a function.

Saving local variables used in a subroutine or function. Passing function parameters to a subroutine or function. Returning data from a subroutine or function

The next couple of pages have some examples of how the stack is implemented.



[1]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


Preserving the return address

Consider the following code to call the delay subroutine and the accompanying stack frame:



SP




SP+2


Stack Address
Byte Data


$4FD



$4FE
$C0


$4FF
$06




Above is a sample portion of code from the keypad from experiment 5.2. Suppose that the stack pointer currently holds the value $500. When the instruction JSR delay is executed, the program pushes the contents of the program counter onto the stack, and the stack pointer is automatically decremented by two so that it now holds the value $4FE. Later at the end of the subroutine, when the RTS instruction is executed (see below), the program pulls the data that the stack pointer is pointing to from the stack, and the stack pointer is automatically incremented by two.



Temporarily saving register values

Continuing with the example above, consider the following subroutine and accompanying stack frame:

SP
Stack Address
Byte Data







$4FC
XH



SP+2
$4FD
XL







$4FE
$C0



SP+4
$4FF
$06








The above example is the delay loop that was implemented in lab 4 for the stepper motor. Since index register X is being used in the subroutine, at the beginning of the subroutine, the contents of index register X is pushed on the stack, and at the end of the subroutine, the contents of index register X is pulled from the stack. Since the HC(S)12 only has a limited number of registers for use, it is important to preserve the contents of any registers that are used in a subroutine on the stack.



[2]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


Passing arguments to a subroutine

The following is an excerpt of the code from lab 6.2. Please note the accompanying assembly code and stack frame.





























The above function call is compiled down to six assembly instructions. This function accepts three arguments. The first is the two-byte effective address of the character array Data. The second argument is the two-byte integer value numberChars which holds the number of characters in the array. The third argument is a dummy argument whose sole purpose is to suppress a compiler warning. The first two arguments are passed to the subroutine through the stack. The third argument remains in accumulator D. For any function call using the default compiler settings in CodeWarrior, all of the arguments except for the last argument are passed through the stack. The program then uses the CALL instruction to preserve the return address and page number and jump to the subroutine.
















[3]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


Extended Memory:

One new thing to take away from the above example is the use of the instruction CALL instead of JSR to call the function. It’s important to understand the memory model of the HC(S)12 and distinction between banked and non-banked memory. Remember that the number of addressable bytes with a 16-bit data bus is 216 = 65,536 bytes. However, the s12e128 used in the lab has 128k of flash memory available for use. To access these addition locations, a technique called “paging” is implemented.


































The above figure shows the local memory map for an HC(S)12 device. There are two common characteristics for devices in the HC(S)12 family. The first is that the space boundaries of register space, EEPROM, and RAM may differ between devices. The other is that the lower 48kB hosts the flash memory. On the HC(S)12, flash memory is partitioned into 16kB sections. The middle 16kB region is called the flash page window.

Certain local addresses do not always point to the same physical address location. These special address ranges are known as page windows. Local addresses in these ranges are addresses of 16 bits and do not have enough information for the controller to determine what the physical location of this memory space is. This addition information is stored to the PPAGE register, which is used to select the part of physical memory that the page window points to.




[4]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17

















The above was taken from the linker.prm file in a default CodeWarrior project. Under paged flash, the first byte is the information stored in the PPAGE register and the second and third byte is the local address. The HC(S)12 family chooses pages in sequential order in such a way that the last page of memory is given the number 0x3F. Although the concept of paging and page numbers are only useful when accessing banked locations, it is important to understand that the division of numbered pages is done for all of memory. For more information on the memory scheme in the S12 architecture, visit the following resource from NXP: http://www.nxp.com/assets/documents/data/en/application-notes/AN3784.pdf

The reason that the memory scheme is brought up is because most c projects in CodeWarrior, including project 6.2 and the final project, implement this page window. In order to preserve the physical page when jumping to a subroutine, the CALL instruction is used in place of the JSR instruction. CALL preserves the page in the PAGE register as well as the return address. To return from CALL, the RTC (Return from call) instruction should be used instead of the RTS instruction, or else the program will not pull the correct value from the PPAGE register and thus will not return properly.




























[5]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


Experimental Procedure:

Laboratory 6.1: Stack Manipulation

Make a new program. At the beginning of the program, initialize the stack pointer using the __SEG_END_SSTACK variable generated by the linker file. Load the registers with the following values:

Accumulator A = $11 Accumulator B = $22

Index Register X = $3333 Index Register Y = $4444

After the registers have been loaded with the values above, call a subroutine. In this subroutine, push the registers onto the stack in the order that they are loaded above. Clear all the registers. Pull the values back from the stack in the opposite order they were pulled.

When running the program in the debugger, go to the location of the stack pointer in the memory window and single step through the program. Fill out the corresponding stack frame below.




Stack Address



Byte Data


































[6]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


Laboratory 6.2: Passing Parameters

The objective of this exercise is to write an assembly language subroutine to interface with a C language main program and a C language subroutine. The subroutine to be written is:

void SendMessageInfo(char* Data, int numberChars, int dummy)

The above function call passes three 16-bit arguments to the subroutine SendMessageInfo. The first argument is the starting effective address of a character string. The second argument is the number of characters in the string. The third argument is a dummy argument to suppress a compiler warning that can be ignored.

The first and second arguments are passed to the function through the stack. The third argument remains in accumulator D. Since this program uses banked memory, both the return address and page number are saved on the stack via the CALL instruction. Structure of stack after the CALL instruction is shown as below:




















Download and extract the file Lab6-2 from the lab section of the course site. Create a new assembly file and add it to the project. Call the subroutine SendMessageInfo. This subroutine should do the following:

Load the effective address of array named Data from the stack into one of the index registers. Load the counter(number of characters) into one of the remaining 16-bit registers.
Load the 8-bit values from Data one at a time.

Push the argument through the stack to the function void SendsChr(char OutValue, int Dummy). This function accepts two arguments. The first is the 8-bit value from the array. The second is a dummy argument passed through accumulator D and can be ignored.

Use the CALL instruction to call the SendsChr function.

Adjust the stack pointer so that the program returns from SendMessageInfo properly. Decrement and check the value of numberChars. If the count has not reached zero,

branch back and load the next element of Data. Else, use the RTC instruction to properly return from the array.

[7]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


    • Remember that SendMessageInfo and SendsChr both must be defined and referenced using the appropriate XDEF and XREF assembler directives.

EXTRA INFO: In assembly, SendMessageInfo function call is compiled down into the following instructions:










Laboratory 6.3: LCD and Potentiometer

The object of this lab is to create an assembly program that reads the values from the potentiometer on the I/O board and display them on the LCD. Extract and open the file Lab6-3 from the lab section of the course site.

To display a string on the LCD, take the following steps (the first two are already completed in the file found in the course site):

Create a string that contains 33 characters in the variables section. To indicate the end of the string, use a null character (0) for the terminator.

Initialize the characters of the string one at a time in the code section using the MOVB instruction.

Initialize the LCD by calling the init_LCD function. This is a function of type void that takes no arguments. Since this project does not implement banked memory, use the JSR instruction to call this function. This should only be done once at the start of the program. Doing this multiple times will cause the LCD to flicker.

To display a string on the LCD, load the effective address of the string into accumulator D, then call the display_string function to display the string on the LCD. This is a function of type void and takes the effective address of a char array as it’s only argument. Reaching a null character will terminate this function. Repeat this step any time that a change is made to the string.




Note: The LCD utilizes the ASCII table to display characters. To change an individual character in a string, put the new character in the corresponding memory address of that string. The value of the new character must be in ASCII format.









[8]    © J. Lee, N. Wheeler
ECE 362 – Experiment 6    Rev – 2/17


To read a value from the potentiometer, call the Read_Pot function. This is a function of type void that takes no arguments. After this function is called, the value of the potentiometer is stored into the two byte variable pot_value. The value of the potentiometer is only one-byte, and is in the lower byte of pot_value. This value can take the range from 0 to 255, but on the lab I/O board it generally only goes up to around 127.


To complete this exercise, take the following steps (note that steps 1 and 3 are already done in the Lab6-3 file):

Include the following files in the project:

o    funct.h    (Header file for the LCD and Potentiometer)

o    lcddisp.c    (C code for init_LCD and display_string)

        o potentiometer.c(C code for the read_pot and pot_value)

Reference the functions and variables that are going to be used in the program. Create the string that will be displayed to the LCD. The string should be:

o  “The value of the pot is:
“,0

Initialize the LCD using the init_LCD function.

Get the value from the potentiometer using the read_pot function.

Store the value of the potentiometer in the string. The value of the potentiometer is from

0 to 255. To do so, follow the following three steps:
o  Separate the digits. For example, if the value in the potentiometer is 147:
        ◦ 147/100: Quotient = 1, Remainder = 47. Store the quotient into memory.

        ◦ 47/10: Quotient = 4, Remainder = 7 (47 is the remainder from 147/100). Store the quotient and remainder into memory.
    o Add #$30 to each digit to convert to its ASCII equivalent.

        o Store the hundreds, tens, and ones digit in order into the string after the ‘:’ character.

Display the string to the LCD using the display_string function. Repeat the previous three steps.






















[9]    © J. Lee, N. Wheeler