Panasonic PanaXSeries MN1030 Series, Руководство пользователя

Страница: 30 / 340

Chapter 2 Program Development Flow
18 Programming with Assembler
Programming style
It is important to use a consistent style for program coding from start to finish. When several people are
to create a program, they should meet in advance to decide on a common style.
You should consider the following points regarding the fixed style of the Cross Assembler.
• Header files
Constants and variables used in all files and define identifiers used in common should be
gathered into a single header file. As a result, changes can be made at just one location in the
header file.
• Library files
Subroutine programs frequently used by different files should be gathered by function as library
files to make programs easier to use.
• Declaration position global directives
Use one position for global directive declarations. The global directive can be declared
anywhere within a program, but confusion will result if the declaration positions differ across
source files.
• Unify radix and notation directives
Choose a common default radix for coding numbers, constant values, strings, etc.
• Comment statements
Comments reveal program algorithms and processing details within a program. Choose a
common format for coding comment statements.
Optimization
This Series' optimizations apply to unconditional branches, data transfer instructions, arithmetic
instructions, logical instructions, bit manipulation instructions, and user-defined instructions.
• Unconditional branches that undergo optimization
• Data transfer, arithmetic, logical, bit manipulation, and user-defined instructions that undergo
optimization
Coding is not a simple task if the programmer must always select the optimal instruction from the
above instructions. In particular, it is nearly impossible to select the optimal instructions when coding a
program divided between files in section format.
The optimization functions provide a solution to these problems. The assembler and linker use them to
produce the optimal code no matter what the source code.
The assembler evaluates the source statement notation. It evaluates the immediate data, memory
specifications, and displacement data appearing as operands to a data transfer, arithmetic, logical, bit
manipulation, and user-defined instructions and selects the shortest version of the instruction.
The assembler also examines unconditional branches, choosing the shortest versions for the CALL,
CALLS, JMP, and JSR instructions.
The linker evaluates instructions that were the object of optimization, and selects the optimal codes.
As a result, the programmer must be aware that the generated code will differ from the source
statements coded in the list file.