Motorola Axiom CMD11E1 Guía de usuario descargar pdf (Pagina 33)

3.4. WRITING YOUR FIRST FUNCTION 33

next and then the program section, though it is easier to follow the code if the

program section precedes the data section as in the previous example.

The program section should start with the main code, i.e. the code you want

to execute. The main code should be followed by various functions. The order is

not important.

What is a function? A function, also known as a subroutine is a self contained

code that implements a well deﬁned functionality. What do I mean by self con-

tained? You should b e able to draw a line above and below your code for the

function, and make sure that

1. Only way to branch out of the two lines is with JSR or BSR or RTS instruc-

tions. If you ﬁnd any other branching instruction such as BRA, BEQ etc. then

your code is most likely incorrect

2. Only way to branch into an instruction between the two lines from an

instruction outside the two lines is with a JSR or BSR instruction. If you

ﬁnd any other branching instruction such as BRA, BEQ etc. then your code

is most likely incorrect.

The function should terminate with a return from subroutine RTS instruction. It

is a good programming practice not to have more than one RTS statement in any

function.

Once you have written your function, it is there for you to use as many times

as you need. To use the function, you should know where its ﬁrst instruction is

located in memory (technically known as the entry point). If you use the assembler

to create the function, you can place a label before the very ﬁrst instruction. The

assembler will automatically EQUate the label with the ﬁrst instruction. If the

function needs any additional information, they will have to be supplied by the

user prior to using the function (technically known as binding). The function you

will be writing will use one of the registers for binding. Also, many functions will

return some useful value to the caller. In this case, it is a good idea to return the

value in one of the registers.

Unless you write functions that do absolutely nothing (technically known as

stubs), the function will use and modify one or more registers. If the caller hap-

pens to keep valuable data in one of these registers, then you have a potential

problem. There are two possible solutions: The caller could save the values in

the registers before calling your function, and then restore it after your function

returns. Alternatively, your function can save the values in the registers it uses

and then restore the values before it returns. The ﬁrst approach more eﬃcient

but the second approach will result in fewer bugs. I strongly recommend that you

get into the habit of writing functions that clean up after themselves and restore

1 2 ... 28 29 30 31 32 33 34 35 36 37 38 ... 100 101

Sin comentarios

Motorola Axiom CMD11E1 Guía de usuario Pagina 33