Understanding ARM Assembly and ARM GNU Assembly

1. Two Ways of ARM Assembly Development

ARM assembly development refers to the use of ARM-provided assembly instructions for ARM program development.

There are two ways of ARM assembly development: one is using ARM assembly, and the other is using ARM GNU assembly. The assembly instructions used in both development methods are exactly the same; the difference lies in the macro instructions, pseudo-instructions, and pseudo-operations. The real distinction between the two methods is the compilation tools used.

For ARM assembly, the compiler developed by ARM is used, while ARM GNU assembly uses the GNU compiler developed for the ARM instruction set, specifically arm-gcc.

2. ARM Compilation Development Environment

Two commonly used ARM compilation development environments:

• DS5: An integrated development software provided by ARM. It uses the toolchain provided by ARM for program compilation.

• GNU Development Environment: Composed of the GNU assembler as, cross-compiler gcc, and linker ld, etc.

3. Pseudo-operations, Macro Instructions, and Pseudo-instructions

Pseudo-operations: Some special instruction mnemonics in ARM assembly language programs, mainly used to prepare for various tasks in the assembly program, processed by the assembler during the assembly of the source program, rather than executed by the machine during runtime. For example, defining a program segment is a pseudo-operation.

Macro instructions: A segment of independent program code that can be inserted into the source program, defined through pseudo-operations.

Pseudo-instructions: Some special instruction mnemonics in ARM assembly language programs that are not executed during processor runtime; instead, they are replaced by appropriate ARM machine instructions during assembly, thus achieving actual instruction operations.

4. ARM Assembly Pseudo-operations

Pseudo-operation	Syntax Format	Function
GBLA	GBLA Variable	Declare a global arithmetic variable and initialize it to 0
GBLL	GBLL Variable	Declare a global logical variable and initialize it to {FALSE}
GBLS	GBLS Variable	Declare a global string variable and initialize it to an empty string
LCLA	LCLA Variable	Declare a local arithmetic variable and initialize it to 0
LCLL	LCLL Variable	Declare a local logical variable and initialize it to {FALSE}
LCLS	LCLS Variable	Declare a local string variable and initialize it to an empty string
SETA	SETA Variable expr	Assign a value to a global or local arithmetic variable
SETL	SETL Variable expr	Assign a value to a global or local logical variable
SETS	SETS Variable expr	Assign a value to a global or local string variable
RLIST	name LIST {list of registers}	Define a name for a list of general-purpose registers
CN	name CN expr	Define a name for a coprocessor register
CP	name CP expr	Define a name for a coprocessor
DN/SN	name DN/SN expr	Define a name for a double/single precision VFP register
FN	name FN expr	Define a name for an FPA floating-point register
LTORG	LTORG	Declare the start of a data buffer pool (literal pool)
MAP	MAP expr {, base-register}	Define the starting address of a structured memory table (storage map)
FIELD	{label} FIELD expr	Define a data field in a structured memory table
SPACE	{label} SPACE expr	Allocate a block of contiguous memory units and initialize with 0
DCB	{label} DCB expr {,expr}..	Allocate a block of byte memory units and initialize with expr
DCD/ DCDU	{label} DCD/DCDU expr {,expr}…	Allocate a block of word memory units and initialize with expr
DCDO	{label} DCDO expr {,expr}…	Allocate a block of word-aligned memory units and initialize with expr
DCFD/DCFDU	{label} DCFD{U} fpliteral ,{,fpliteral}…	Allocate word-aligned memory units for double precision floating-point numbers
DCFS/DCFSU	{label} DCFS{U} fpliteral ,{,fpliteral}…	Allocate word-aligned memory units for single precision floating-point numbers
DCI	{label} DCI expr, {expr}…	ARM code allocates a segment of word-aligned memory units, filling expr (binary instruction code); in THUMB code, allocates a segment of half-word aligned half-word memory units.
DCQ/ DCQU	{label} DCQ{U} {-} literal, {, {-} literal}…	Allocate a segment of memory in double words (8 bytes)
DCW/DCWU	{label} DCW{U} {-} literal, {, {-} literal}…	DCW is used to allocate a segment of half-word aligned half-word memory units

1. AREA

Create a new program code or data area.

Format: AREA name, {,attr,} …

Where name is the segment name, attr is the segment name attribute.

For attributes, there are the following:

• CODE: Used to define a code segment, default is READONLY.

• DATA: Used to define a data segment, default is READWRITE.

• READONLY: Specifies that the contents of this segment are read-only.

• READWRITE: Specifies that the contents of this segment are readable and writable.

• ALIGN: Specifies alignment to a power of 2.

• COMMON: Defines a common segment. It does not contain any user code and data. COMMON segments with the same name in different source files share the same storage unit.

2. ALIGN

Specifies alignment.

ALIGN 4 indicates 4-byte address alignment.

ALIGN 8 indicates 8-byte address alignment.

Note: The difference between using ALIGN in AREA and using ALIGN alone lies in the format and alignment calculation.

3. ENTRY

Specifies the entry point of the assembly program.

A program must have at least one entry point, and it can have multiple entry points, but there can be at most one ENTRY in a source file. When multiple source files have ENTRY, the linker specifies the actual entry point of the program.

4. END

Indicates the end of the source program.

Therefore, assembly language source files must end with END; when the assembler encounters END, it will terminate compilation.

5. EXPORT

Format: EXPORT label [,WEAK]

Declare a global label that can be used by other source files. WEAK indicates that if there are other labels with the same name, the other labels take priority.

6. IMPORT

Format: IMPORT label [,WEAK]

Indicates that the referenced label is in other source files and only needs to be referenced in the current file. WEAK indicates that if the label cannot be found, no error will be reported; generally, the label is set to 0. If this label is used in B or BL instructions, the instruction will be set to nop.

This label will be added to the symbol table of the current source file.

7. EXTERN

Similar to IMPORT, except that if the current file does not reference this label, the label will not be added to the symbol table of the current source file.

8. GET (or INCLUDE)

Includes a source file into the current source file.

9. EQU

Assigns a constant label a value.

Format: name EQU expression

Where name is the symbol name, and expression is a fixed value related to the register or program.

For example:

num EQU 2; assigns the number 2 to the symbol.

EQU is equivalent to #define in C language for defining a constant.

10. SPACE

Used to allocate a block of contiguous memory units and initialize with 0. SPACE can be replaced with %.

Format: {label} SPACE expr

label: is an optional label.

expr: number of bytes allocated.

For example:

stack SPACE 100; allocates 100 bytes of memory units and initializes with 0. The label stack is the starting address of this space.

11. DCB

Used to allocate segment byte memory units and initialize with expr in the pseudo-operation.

Format: {label} DCB expr {,expr}

label: is an optional label.

expr: can be a value from -128 to 255 or a string.

For example:

string DCB “HELLO”; allocates space for the string HELLO, where string is the starting address of this space.

12. DCD and DCDU

Used to allocate segment word memory units (the allocated memory is word-aligned, DCDU is not strictly word-aligned) and initialize with expr in the pseudo-operation. DCD can be replaced with &.

Format: {label} DCD expr, {,expr}

label: is an optional label, indicating the starting address of the memory unit.

expr: numeric expression or label in the program.

For example:

data DCD 1,2,3,4; allocates word-aligned word unit space, initialized to 1, 2, 3, 4.

5. ARM Assembly Pseudo-instructions

ARM pseudo-instructions include: ADR, ADRL, LDR, NOP.

THUMB pseudo-instructions include: ADR, LDR, NOP.

Pseudo-instruction	Syntax Format	Function
ADR	ADR{cond} register, expr	Reads a PC-relative or register-relative address value into a register. Small range address read.
ADRL	ADRL{cond} register, expr	Reads a PC-relative or register-relative address value into a register. Medium range address read.
LDR	LDR {cond} register, =[expr|label]	Reads a 32-bit immediate value or an address value into a register. Large range address read.
NOP	NOP	Replaced by a no-operation during assembly.

6. ARM GNU Compilation Environment

Pseudo-operation	Syntax Format	Function
.byte	.byte expr {,expr}…	Allocates a segment of byte memory units and initializes with expr.
.hword/.short	.hword expr {,expr}…	Allocates a segment of half-word memory units and initializes with expr.
.ascii	.ascii expr {,expr}…	Defines string expr.
.asciz/.string	.asciz expr {,expr}…	Defines string expr (adds /0 as a terminator).
.float/.single	.float expr {,expr}…	Defines 32-bit IEEE floating-point number expr.
.double	.double expr {,expr}…	Defines 64-bit IEEE floating-point number expr.
.word/.long/.int	.word expr {,expr}…	Allocates a segment of word memory units and initializes with expr.
.fill	.fill repeat {,size} {,value}	Allocates a segment of byte memory units, filling repeat times with value of size length.
.zero	.zero size	Allocates a segment of byte memory units and fills the memory with 0.
.space/.skip	.space size, {,value}	Allocates a segment of memory units and initializes the memory with value.
.section	.section expr	Defines a section.
.text	.text {subsection}	Code segment.
.data	.data{subsection}	Data segment.
.bss	.bss{subsection}	bss segment.
.cond 16/.thumb	.code 16/.thumb	Indicates that subsequent assembly instructions use the THUMB instruction set.
.code 32/.arm	.code 32/.arm	Indicates that subsequent assembly instructions use the ARM instruction set.
.end	.end	Marks the end of the assembly file.
.include	.include “filename”	Includes a source file into the current source file.
.align/.balign	.align {alignment} {,fill},{max}	Fills bytes to satisfy certain alignment format.

7. Differences Between the Two Development Environments

The assembly code under the two development environments has many differences, mainly in symbols and pseudo-operations.

Pseudo-operators of ARM Assembly	Pseudo-operators of GNU Assembly
INCLUDE	.include
NUM EQU 25	.equ NUM, 25
EXPORT	.global
IMPORT	.extern
DCD	.long
IF: DEF:	.ifdef
ELSE	.else
ENDIF	.endif
OR	|
SHL	<<
RN	.req
GBLA	.global
NUM SETA 16	.equ NUM , 16
MACRO	.macro
MEND	.endm
END	.end
AREA WORD, CODE, READONLY	.text
AREA BLOCK, DATA, READWRITE	.data
CODE32	.arm
CODE16	.thumb
LTORG	.ltorg
%	.fill
ENTRY	ENTRY:
ldr x0,=0xff	ldr x0,=0xff

Recruitment Information

Training Information

You can also directly click the URL to visit

http://gk.chinaaet.com