Removed unnecessary zero page usage.

LDY <absolute>,X without page-crossing (sort of surprisingly) requires only 4 cycles, so there's after all no need at all to resort to LDY <zeropage>,X to stay inside the 4 cycle budget.
This commit is contained in:
Oliver Schmidt 2020-05-12 23:01:21 +02:00
parent c445802cfa
commit 406a0628bd

View File

@ -5,17 +5,18 @@
.macpack apple2
.pc02
src := $00
dst := $02
rd_count := $04
wr_count := $05
scroll_row := $06
scroll_col := $07
scroll_opc := $08
; free zero page locations according to ProDOS TechRefMan A.4
src := $CE
dst := $D6
rd_count := $EB
wr_count := $EC
scroll_row := $ED
scroll_col := $EE
scroll_opc := $EF
program := $100B ; adressed via jmp_table !
buffer := $9000
zp_save := $9100
physical := $9100
px = $8F ; force page crossing
silence = $FF
@ -118,22 +119,15 @@ chr:lda a:$0000,x ; patched
stream:
; Save zero page
ldx #$00
: lda $00,x
sta zp_save,x
inx
bne :-
; Crate unrolled loop program
jsr unroll
; Create socket 0 physical read memory lookup table in zero page
; Create socket 0 physical read memory lookup table
ldx #$00
: txa
and #>$1FFF
ora #>$6000
sta $00,x
sta physical,x
inx
bne :-
@ -152,13 +146,6 @@ stream:
; Run unrolled loop program
jsr program
; Restore zero page
ldx #$00
: lda zp_save,x
sta $00,x
inx
bne :-
; Switch to text screen
bit $C051
@ -473,17 +460,6 @@ scroll_patch:
:.endmacro
; 14 bytes, 9 cycles
.macro spkr_long
lsr a ; 2
bcs :+ ; 2 3
nop ; 2
bra :++ ; 3
.res 5
: bit $C030 ; 4
:.endmacro
; 13 bytes, 9 cycles
.macro spkr_jsr addr
lsr a ; 2
@ -550,8 +526,8 @@ recv_prolog:
spkr ; bit 2
ldx $C0F7 ; high byte
spkr ; bit 3
ldy $00,x ; high byte -> physical
spkr_long ; bit 4
ldy physical,x ; high byte -> physical
spkr ; bit 4
ldx $C0F7 ; low byte
spkr ; bit 5
sty $C0F5 ; read addr high