Implicit Behaviors of Storage Model Pseudoinstructions

<span>.MODEL</span> pseudoinstructions not only define the storage model of the program but also automatically execute several important low-level settings. These implicit actions greatly simplify the writing of assembly language programs.

Core Implicit Actions

1. **Automatic Creation of Segment Group (DGROUP)**：

• Combines <span>_DATA</span>, <span>CONST</span>, <span>_BSS</span>, and <span>STACK</span> into a segment group named DGROUP
• In the SMALL model, all data segments share the same physical segment

ASSUME CS:code_segment_name, DS:DGROUP, SS:DGROUP

• The default name for the code segment in the SMALL model is <span>_TEXT</span>
• The name of the code segment in medium/large models varies according to the model

Code Examples and Analysis

Example 1: Implicit Behavior in the SMALL Model

; Program Name: IMPLICIT.ASM
; Function: Demonstrate implicit actions of .MODEL

.MODEL SMALL    ; Implicit: DGROUP GROUP _DATA, CONST, _BSS, STACK
                ; Implicit: ASSUME CS:_TEXT, DS:DGROUP, SS:DGROUP

.STACK 100H     ; Belongs to DGROUP

.DATA           ; Belongs to DGROUP (actual segment name _DATA)
    MSG DB 'DGROUP Demonstration', '$'

.CODE           ; Segment name _TEXT (SMALL model)
MAIN PROC
    ; No need to explicitly set ASSUME
    MOV AX, @DATA   ; @DATA = address of DGROUP
    MOV DS, AX      ; Initialize DS to point to DGROUP

    MOV AH, 09H
    LEA DX, MSG     ; Access data in DGROUP
    INT 21H

    MOV AH, 4CH
    INT 21H
MAIN ENDP

END MAIN

Key Point Explanation:

1. <span>@DATA</span> is defined by the assembler as the address of DGROUP
2. All data segments (_DATA, .DATA?, .CONST) share the same segment register
3. The code segment defaults to NEAR calls

Changes in Segment Names Across Different Models

Example 2: Segment Names in the MEDIUM Model

.MODEL MEDIUM

.STACK 200H     ; Still belongs to DGROUP

.DATA           ; _DATA
    BUFFER DB 100 DUP(?)

.CODE           ; Segment name varies with the model
MAIN PROC FAR   ; Requires FAR calls
    MOV AX, @DATA
    MOV DS, AX

    ; Cross-segment subroutine call
    CALL OTHER_SEGMENT

    RET
MAIN ENDP

.CODE OTHER     ; Additional code segment
OTHER_SEGMENT PROC FAR
    RET
OTHER_SEGMENT ENDP

END MAIN

Explicit ASSUME Usage Scenarios

Although .MODEL provides default ASSUME settings, explicit declarations are still necessary in certain cases:

Example 3: Situations Requiring Explicit ASSUME

.MODEL SMALL

EXTRN FARDATA:BYTE  ; External far data

.DATA
    LOCAL_VAR DB ?

.CODE
MAIN PROC
    MOV AX, @DATA
    MOV DS, AX      ; DS points to DGROUP

    ; Situation 1: Accessing far data requires ES
    ASSUME ES:NOTHING
    MOV AX, SEG FARDATA
    MOV ES, AX
    ASSUME ES:SEG FARDATA

    MOV AL, ES:FARDATA

    ; Situation 2: Switching DS usage
    ASSUME DS:NOTHING
    MOV AX, CS      ; Make DS point to code segment
    MOV DS, AX
    ASSUME DS:CODE

    ; Restore DS
    MOV AX, @DATA
    MOV DS, AX
    ASSUME DS:DGROUP

    MOV AH, 4CH
    INT 21H
MAIN ENDP

END MAIN

Advanced Applications: Managing Multiple Data Segments

Example 4: Explicit Control of Multiple Data Segments

.MODEL COMPACT

FAR_DATA SEGMENT PARA 'FAR_DATA'
    BIG_ARRAY DB 65536 DUP(?)
FAR_DATA ENDS

.DATA
    SMALL_VAR DW ?

.CODE
MAIN PROC
    ; Default DS settings
    MOV AX, @DATA
    MOV DS, AX
    ASSUME DS:DGROUP

    ; Access far data segment
    MOV AX, SEG FAR_DATA
    MOV ES, AX
    ASSUME ES:SEG FAR_DATA

    ; Initialize large array
    MOV CX, 32768
    LEA DI, ES:BIG_ARRAY
    MOV AL, 0
    REP STOSW

    ; Switch DS to far data segment
    PUSH DS
    MOV AX, ES
    MOV DS, AX
    ASSUME DS:SEG FAR_DATA

    ; Directly access far data
    MOV BIG_ARRAY[0], 1

    ; Restore DS
    POP DS
    ASSUME DS:DGROUP

    MOV AH, 4CH
    INT 21H
MAIN ENDP

END MAIN

Practical Development Recommendations

1. Understand Default Behaviors：

• Be clear about the implicit actions of the .MODEL pseudoinstruction under specific models
• Use <span>/Sg</span> option to view the segment definitions generated by the assembler

• In the small model, rely on default settings
• Large programs require explicit management of additional segment registers like ES/FS/GS

• Check the actual values of segment registers in the debugger
• Pay attention to the matching of ASSUME declarations with actual settings

• Maintain consistent segment conventions when interacting with high-level languages
• Be aware of segment settings during function calls

Performance Optimization Considerations

1. Advantages of the Small Model：

• Reduces segment register load instructions
• All transfers are of NEAR type

• Frequently accessed data should be placed within DGROUP
• Large data should use a separate segment and be accessed via ES

• Keep critical path code within the same code segment
• Separate cold code into other segments

By deeply understanding the implicit actions of the .MODEL pseudoinstruction, one can write concise and efficient assembly programs. Especially in modern development environments, these automated processes can significantly enhance development efficiency.