|
Last change
on this file since 464 was 287, checked in by aitotat@…, 13 years ago |
|
Changes to Assembly Library:
- Hopefully fixed a problem of wrong sized MDA cursor.
- Some minor improvements.
|
|
File size:
1.7 KB
|
| Line | |
|---|
| 1 | ; Project name : Assembly Library
|
|---|
| 2 | ; Description : Delay macros.
|
|---|
| 3 | %ifndef DELAY_INC
|
|---|
| 4 | %define DELAY_INC
|
|---|
| 5 |
|
|---|
| 6 | ;--------------------------------------------------------------------
|
|---|
| 7 | ; Clears prefetch queue by jumping to next instruction.
|
|---|
| 8 | ; This delays much more than nop instruction of fast systems.
|
|---|
| 9 | ;
|
|---|
| 10 | ; JMP_DELAY
|
|---|
| 11 | ; Parameters
|
|---|
| 12 | ; Nothing
|
|---|
| 13 | ; Returns:
|
|---|
| 14 | ; Nothing
|
|---|
| 15 | ; Corrupts registers:
|
|---|
| 16 | ; Nothing
|
|---|
| 17 | ;--------------------------------------------------------------------
|
|---|
| 18 | %macro JMP_DELAY 0
|
|---|
| 19 | jmp SHORT %%NextInstruction
|
|---|
| 20 | %%NextInstruction:
|
|---|
| 21 | %endmacro
|
|---|
| 22 |
|
|---|
| 23 |
|
|---|
| 24 | ;--------------------------------------------------------------------
|
|---|
| 25 | ; Mimimun delays (without fetching) with some CPU architectures:
|
|---|
| 26 | ; 8088/8086: 17 cycles for jump + 5 cycles for last comparison
|
|---|
| 27 | ; 286: 10 cycles for jump + 4 cycles for last comparison
|
|---|
| 28 | ; 386: 13 cycles for jump + ? cycles for last comparison
|
|---|
| 29 | ; 486: 7 cycles for jump + 6 cycles for last comparison
|
|---|
| 30 | ;
|
|---|
| 31 | ; LOOP instruction uses two bytes so aligned fetching will require:
|
|---|
| 32 | ; 8088: 8 cycles (two BYTE reads)
|
|---|
| 33 | ; 8086: 4 cycles (one WORD read)
|
|---|
| 34 | ; 286: 2 cycles + wait states (usually 1)
|
|---|
| 35 | ; 386: ?
|
|---|
| 36 | ; 486: Fetched only once to internal cache
|
|---|
| 37 | ;
|
|---|
| 38 | ; DELAY_WITH_LOOP_INSTRUCTION_NA ; No JUMP_ALIGN
|
|---|
| 39 | ; DELAY_WITH_LOOP_INSTRUCTION
|
|---|
| 40 | ; Parameters
|
|---|
| 41 | ; CX: Loop iterations (0 is maximum delay with 65536 iterations)
|
|---|
| 42 | ; Returns:
|
|---|
| 43 | ; CX: Zero
|
|---|
| 44 | ; Corrupts registers:
|
|---|
| 45 | ; Nothing
|
|---|
| 46 | ;--------------------------------------------------------------------
|
|---|
| 47 | %macro DELAY_WITH_LOOP_INSTRUCTION_NA 0
|
|---|
| 48 | %%StartOfLoop:
|
|---|
| 49 | loop %%StartOfLoop
|
|---|
| 50 | %endmacro
|
|---|
| 51 |
|
|---|
| 52 | %macro DELAY_WITH_LOOP_INSTRUCTION 0
|
|---|
| 53 | ALIGN JUMP_ALIGN
|
|---|
| 54 | %%StartOfLoop:
|
|---|
| 55 | loop %%StartOfLoop
|
|---|
| 56 | %endmacro
|
|---|
| 57 |
|
|---|
| 58 |
|
|---|
| 59 | %endif ; DELAY_INC
|
|---|
Note:
See
TracBrowser
for help on using the repository browser.
