Fast graphics path screen fill working

Makefile

@@ -11,7 +11,7 @@
 
 CL65=cl65
 AC=AppleCommander.jar
-ADDR=6000
+ADDR=800
 
 PGM=gscats
 

equates.s

@@ -1,10 +1,12 @@
-TEXTCOLOR		= $c022
+TEXTCOLOR		= $e0c022
-NEWVIDEO		= $c029
+NEWVIDEO		= $e0c029
-BORDERCOLOR		= $c034
+BORDERCOLOR		= $e0c034
-CLOCKCTL		= $c034
+CLOCKCTL		= $e0c034
+SHADOW			= $e0c035
+STACKCTL		= $e0c068
+KBD				= $e0c000
+KBDSTROBE		= $e0c010
 COUT			= $fded
-KBD				= $c000
-KBDSTROBE		= $c010
 
 
 ; Zero page locations we use (unused by Monitor, Applesoft, or ProDOS)

graphics.s

@@ -1,19 +1,31 @@
-;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 ; colorFill
-; Fills the screen with a color (or two). Not very fast.
+; Fills the screen with a color (or two). Pretty fast, but not fastest possible
 ; A 4:4:4:4 = Palette entries
+;
+; Trashes Y
 
 colorFill:
-	phx
+	FASTGRAPHICS
-	ldx #32000
+
+	lda #$9d00-1	; Point stack to end of VRAM
+	tcs
+
+	ldy #200
 
 colorFillLoop:
-	sta $e11ffe,x
+	; 80 PHXs, for 1 line
-	dex
+	; We could do the entire screen with PHXs, but this is a
-	dex
+	; balance between speed and super-verbose code
-	bne colorFillLoop
+	.byte $da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da
-	plx
+	.byte $da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da
+	.byte $da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da
+	.byte $da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da,$da
 
+	dey
+	bne colorFillLoop
+
+	SLOWGRAPHICS
 	rts
 
 
@@ -30,7 +42,7 @@ initSCBs:
 initSCBsLoop:
 	dex
 	dex
-	sta $e19cfe,x
+	sta $e19d00,x
 	bne initSCBsLoop
 	rts
 
@@ -87,8 +99,19 @@ setPaletteLoop:
 	iny
 	inx
 	inx
-	cmp #16
+	cpx #32
 	bne setPaletteLoop
 
 	RESTORE_XY
 	rts
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+; caches
+shadowRegister:
+	.byte 0
+stackRegister:
+	.byte 0
+stackPtr:
+	.word 0
+

gscats.s

@@ -10,13 +10,49 @@
 .include "equates.s"
 
 
-.org $6000
+.org $800
 
 main:
 	NATIVE
 
+mainCopyStart:
+	ldx #0
+	lda #mainBank2
+	sta mainCopyDest+1
+
+mainCopyLoop:
+	lda mainBank2,x
+
+mainCopyDest:
+	sta $020800,x
+	inx
+	cpx #endMainBank2-mainBank2
+	bne mainCopyLoop
+
+	lda #returnToProDOS
+	sta proDOSLongJump
+	lda #mainBank2
+	sta mainLongJump
+	jml (mainLongJump)
+
+returnToProDOS:
+	SYNCDBR
+	EMULATION
+	rts
+
+mainLongJump:
+	.byte 00,08,02
+proDOSLongJump:
+	.byte 00,00,00
+
+
+mainBank2:
+	SYNCDBR
+
+	BITS8
 	lda #$f0
 	sta TEXTCOLOR
+	BITS16
 
 	jsr initSCBs
 	SHRVIDEO
@@ -26,14 +62,13 @@ main:
 	lda #0
 	jsr setPalette
 	
-	lda #$3333
+	ldx #$1111
 	jsr colorFill
 
 	jsr kbdWait
 	CLASSICVIDEO
 
-	EMULATION
+	jml (proDOSLongJump)
-	rts
 
 
 
@@ -62,6 +97,8 @@ basePalette:
 .include "graphics.s"
 
 
+endMainBank2:
+
 
 ; Suppress some linker warnings - Must be the last thing in the file
 ; This is because Quinn doesn't really know how to use ca65 properly

macros.s

@@ -1,44 +1,114 @@
+.macro OP8
+	.i8
+	.a8
+.endmacro
+
+
+.macro OP16
+	.i16
+	.a16
+.endmacro
+
+
+.macro BITS8
+	sep #%00110000
+	OP8
+.endmacro
+
+
+.macro BITS16
+	rep #%00110000
+	OP16
+.endmacro
+
+
+.macro DBR bankNum
+	BITS8
+	lda bankNum
+	pha
+	plb
+	BITS16
+.endmacro
+
+
+.macro SYNCDBR
+	phk
+	plb
+.endmacro
+
+
 .macro EMULATION
 	sec		; Enable 8-bit mode
 	xce
-	.i8
+	OP8
-	.a8
 .endmacro
 
 
 .macro NATIVE
 	clc				; Enable 16-bit mode
 	xce
-	rep	#$30
+	BITS16
-	.i16
-	.a16
 .endmacro
 
+
 .macro SHRVIDEO
-	sep #$30
+	BITS8
-	.i8
-	.a8
 	lda NEWVIDEO
 	ora	#%11000001
 	sta NEWVIDEO
-	rep	#$30
+	BITS16
-	.i16
-	.a16
 .endmacro
 
+
 .macro CLASSICVIDEO
-	sep #$30
+	BITS8
-	.i8
-	.a8
 	lda NEWVIDEO
 	and #%00111111
 	sta NEWVIDEO
-	rep	#$30
+	BITS16
-	.i16
-	.a16
 .endmacro
 
 
+.macro FASTGRAPHICS
+	sei
+	sep #%00100000	; 16-bit A only, to preserve X/Y
+	.a8
+
+	lda SHADOW
+	sta shadowRegister
+	lda #0
+	sta SHADOW
+
+	lda STACKCTL
+	sta stackRegister
+	ora #$30
+	sta STACKCTL
+
+	rep #%00100000
+	.a16
+	tsc
+	sta stackPtr
+.endmacro
+
+
+.macro SLOWGRAPHICS
+	sep #%00100000	; 16-bit A only, to preserve X/Y
+	.a8
+
+	lda shadowRegister
+	sta SHADOW
+
+	lda stackRegister
+	sta STACKCTL
+
+	rep #%00100000
+	.a16
+	lda stackPtr
+	tcs
+
+	cli
+.endmacro
+
 
 
 ;;;;;;;;;;