mirror of
https://github.com/RaySollium99/picodrive.git
synced 2025-09-05 15:27:46 -04:00
9839d126cc
git-svn-id: file:///home/notaz/opt/svn/PicoDrive@261 be3aeb3a-fb24-0410-a615-afba39da0efa
1910 lines
51 KiB
ArmAsm
1910 lines
51 KiB
ArmAsm
@ vim:filetype=armasm
|
|
|
|
@ assembly "optimized" version of some funtions from draw.c
|
|
@ this is highly specialized, be careful if changing related C code!
|
|
|
|
@ (c) Copyright 2007, Grazvydas "notaz" Ignotas
|
|
@ All Rights Reserved
|
|
|
|
.include "port_config.s"
|
|
|
|
.extern Pico
|
|
.extern PicoOpt
|
|
.extern HighCol
|
|
.extern Scanline
|
|
.extern HighSprZ
|
|
.extern rendstatus
|
|
.extern DrawLineDest
|
|
.extern DrawStripInterlace
|
|
|
|
|
|
@ helper
|
|
.macro TilePixel pat lsrr offs
|
|
.if !\lsrr
|
|
ands r4, \pat, r2
|
|
.else
|
|
ands r4, \pat, r2, lsr #\lsrr
|
|
.endif
|
|
orrne r4, r3, r4
|
|
strneb r4, [r1,#\offs]
|
|
.endm
|
|
|
|
@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf
|
|
.macro TileNorm pat
|
|
TilePixel \pat, 12, 0 @ #0x0000f000
|
|
TilePixel \pat, 8, 1 @ #0x00000f00
|
|
TilePixel \pat, 4, 2 @ #0x000000f0
|
|
TilePixel \pat, 0, 3 @ #0x0000000f
|
|
TilePixel \pat, 28, 4 @ #0xf0000000
|
|
TilePixel \pat, 24, 5 @ #0x0f000000
|
|
TilePixel \pat, 20, 6 @ #0x00f00000
|
|
TilePixel \pat, 16, 7 @ #0x000f0000
|
|
.endm
|
|
|
|
@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf
|
|
.macro TileFlip pat
|
|
TilePixel \pat, 16, 0 @ #0x000f0000
|
|
TilePixel \pat, 20, 1 @ #0x00f00000
|
|
TilePixel \pat, 24, 2 @ #0x0f000000
|
|
TilePixel \pat, 28, 3 @ #0xf0000000
|
|
TilePixel \pat, 0, 4 @ #0x0000000f
|
|
TilePixel \pat, 4, 5 @ #0x000000f0
|
|
TilePixel \pat, 8, 6 @ #0x00000f00
|
|
TilePixel \pat, 12, 7 @ #0x0000f000
|
|
.endm
|
|
|
|
@ shadow/hilight mode
|
|
|
|
@ this one is for hi priority layer
|
|
.macro TilePixelShHP lsrr offs
|
|
.if !\lsrr
|
|
ands r4, r12, r2
|
|
.else
|
|
ands r4, r12, r2, lsr #\lsrr
|
|
.endif
|
|
ldreqb r4, [r1,#\offs]
|
|
orrne r4, r3, r4
|
|
strneb r4, [r1,#\offs]
|
|
tsteq r4, #0x80
|
|
andeq r4, r4, #0x3f
|
|
streqb r4, [r1,#\offs]
|
|
.endm
|
|
|
|
@ TileNormShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits
|
|
.macro TileNormShHP
|
|
TilePixelShHP 12, 0 @ #0x0000f000
|
|
TilePixelShHP 8, 1 @ #0x00000f00
|
|
TilePixelShHP 4, 2 @ #0x000000f0
|
|
TilePixelShHP 0, 3 @ #0x0000000f
|
|
TilePixelShHP 28, 4 @ #0xf0000000
|
|
TilePixelShHP 24, 5 @ #0x0f000000
|
|
TilePixelShHP 20, 6 @ #0x00f00000
|
|
TilePixelShHP 16, 7 @ #0x000f0000
|
|
.endm
|
|
|
|
@ TileFlipShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf
|
|
.macro TileFlipShHP
|
|
TilePixelShHP 16, 0 @ #0x000f0000
|
|
TilePixelShHP 20, 1 @ #0x00f00000
|
|
TilePixelShHP 24, 2 @ #0x0f000000
|
|
TilePixelShHP 28, 3 @ #0xf0000000
|
|
TilePixelShHP 0, 4 @ #0x0000000f
|
|
TilePixelShHP 4, 5 @ #0x000000f0
|
|
TilePixelShHP 8, 6 @ #0x00000f00
|
|
TilePixelShHP 12, 7 @ #0x0000f000
|
|
.endm
|
|
|
|
|
|
@ TileSingleSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx; r12: helper pattern 0xf
|
|
.macro TileSingleSh
|
|
tst r0, #1 @ not aligned?
|
|
mov r7, #0x00c000
|
|
orr r7, r7, #0xc0
|
|
ldrneb r4, [r1]
|
|
ldreqh r4, [r1]
|
|
orr r4, r4, r7
|
|
strneb r4, [r1], #1
|
|
streqh r4, [r1], #2
|
|
ldrh r4, [r1]
|
|
orr r4, r4, r7
|
|
strh r4, [r1], #2
|
|
ldrh r4, [r1]
|
|
orr r4, r4, r7
|
|
strh r4, [r1], #2
|
|
ldrh r4, [r1]
|
|
orr r4, r4, r7
|
|
strh r4, [r1], #2
|
|
ldrneb r4, [r1]
|
|
orr r4, r4, r7
|
|
strneb r4, [r1], #1
|
|
.endm
|
|
|
|
@ TileSingleHi (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf
|
|
.macro TileSingleHi
|
|
tst r1, #1 @ not aligned?
|
|
mov r7, #0x008000
|
|
orr r7, r7, #0x80
|
|
ldrneb r4, [r1], #1
|
|
ldreqh r4, [r1], #2 @ 1ci
|
|
ldrh r12, [r1], #2
|
|
bic r4, r4, r7, lsr #1
|
|
orr r4, r4, r7
|
|
strneb r4, [r1, #-3]
|
|
streqh r4, [r1, #-4]
|
|
ldrh r4, [r1], #2
|
|
bic r12, r12, r7, lsr #1
|
|
orr r12, r12, r7
|
|
strh r12, [r1, #-4]
|
|
ldrh r12, [r1], #2
|
|
bic r4, r4, r7, lsr #1
|
|
orr r4, r4, r7
|
|
strh r4, [r1, #-4]
|
|
ldrneb r4, [r1]
|
|
bic r12, r12, r7, lsr #1
|
|
orr r12, r12, r7
|
|
strh r12, [r1, #-2]
|
|
bicne r4, r4, r7, lsr #1
|
|
orrne r4, r4, r7
|
|
strneb r4, [r1], #1
|
|
mov r12, #0xf
|
|
.endm
|
|
|
|
.macro TileDoShGenPixel shift ofs
|
|
.if \shift
|
|
ands r4, r12, r2, lsr #\shift
|
|
.else
|
|
ands r4, r12, r2
|
|
.endif
|
|
beq 3f
|
|
cmp r4, #0xe
|
|
beq 2f
|
|
bgt 1f
|
|
orr r4, r3, r4
|
|
strb r4, [r1,#\ofs]
|
|
b 3f
|
|
1:
|
|
ldrb r4, [r1,#\ofs] @ 2ci
|
|
orr r4, r4, #0xc0
|
|
strb r4, [r1,#\ofs]
|
|
b 3f
|
|
2:
|
|
ldrb r4, [r1,#\ofs] @ 2ci
|
|
bic r4, r4, #0xc0
|
|
orr r4, r4, #0x80
|
|
strb r4, [r1,#\ofs]
|
|
3:
|
|
.endm
|
|
|
|
@ TileFlipSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf
|
|
.macro TileFlipSh
|
|
TileDoShGenPixel 16, 0 @ #0x000f0000
|
|
TileDoShGenPixel 20, 1 @ #0x00f00000
|
|
TileDoShGenPixel 24, 2 @ #0x0f000000
|
|
TileDoShGenPixel 28, 3 @ #0xf0000000
|
|
TileDoShGenPixel 0, 4 @ #0x0000000f
|
|
TileDoShGenPixel 4, 5 @ #0x000000f0
|
|
TileDoShGenPixel 8, 6 @ #0x00000f00
|
|
TileDoShGenPixel 12, 7 @ #0x0000f000
|
|
.endm
|
|
|
|
@ TileNormSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf
|
|
.macro TileNormSh
|
|
TileDoShGenPixel 12, 0 @ #0x0000f000
|
|
TileDoShGenPixel 8, 1 @ #0x00000f00
|
|
TileDoShGenPixel 4, 2 @ #0x000000f0
|
|
TileDoShGenPixel 0, 3 @ #0x0000000f
|
|
TileDoShGenPixel 28, 4 @ #0xf0000000
|
|
TileDoShGenPixel 24, 5 @ #0x0f000000
|
|
TileDoShGenPixel 20, 6 @ #0x00f00000
|
|
TileDoShGenPixel 16, 7 @ #0x000f0000
|
|
.endm
|
|
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
@ struct TileStrip
|
|
@ {
|
|
@ int nametab; // 0x00
|
|
@ int line; // 0x04
|
|
@ int hscroll; // 0x08
|
|
@ int xmask; // 0x0C
|
|
@ int *hc; // 0x10 (pointer to cache buffer)
|
|
@ int cells; // 0x14
|
|
@ };
|
|
|
|
@ void DrawLayer(int plane_sh, int *hcache, int cellskip, int maxcells);
|
|
|
|
.global DrawLayer
|
|
|
|
DrawLayer:
|
|
stmfd sp!, {r4-r11,lr}
|
|
|
|
ldr r11, =(Pico+0x22228) @ Pico.video
|
|
mov r8, #1
|
|
|
|
ldrb r7, [r11, #16] @ ??vv??hh
|
|
|
|
mov r6, r1 @ hcache
|
|
orr r9, r3, r0, lsl #30
|
|
orr r9, r9, r2, lsl #8 @ r9=sh[31]|cellskip[15:8]|maxcells[7:0] (tmp)
|
|
|
|
mov r1, r7, lsl #4
|
|
orr r1, r1, #0x00ff
|
|
|
|
and r10, r7, #3
|
|
cmp r10, #1
|
|
biclt r1, r1, #0xfc00
|
|
biceq r1, r1, #0xfe00
|
|
bicgt r1, r1, #0xff00 @ r1=ymask=(height<<8)|0xff; ...; // Y Mask in pixels
|
|
|
|
add r10, r10, #5
|
|
cmp r10, #7
|
|
subge r10, r10, #1 @ r10=shift[width] (5,6,6,7)
|
|
|
|
@ calculate xmask:
|
|
mov r5, r8, lsl r10
|
|
sub r5, r5, #1 @ r5=xmask
|
|
|
|
@ Find name table:
|
|
ands r0, r0, #1
|
|
ldreqb r12, [r11, #2]
|
|
ldrneb r12, [r11, #4]
|
|
|
|
ldr r2, =Scanline @ trying to make good use of pipeline here
|
|
ldr lr, =(Pico+0x10000) @ lr=Pico.vram
|
|
|
|
moveq r12, r12, lsl #10
|
|
movne r12, r12, lsl #13
|
|
and r12, r12, #(7<<13) @ r12=(ts->nametab<<1) (halfword compliant)
|
|
|
|
ldrh r8, [r11, #12]
|
|
ldrb r7, [r11, #11]
|
|
ldr r2, [r2]
|
|
|
|
mov r4, r8, lsr #8 @ pvid->reg[13]
|
|
mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)
|
|
tst r7, #2
|
|
addne r4, r4, r2, lsl #2 @ htab+=Scanline<<1; // Offset by line
|
|
tst r7, #1
|
|
biceq r4, r4, #0x1f @ htab&=~0xf; // Offset by tile
|
|
add r4, r4, r0, lsl #1 @ htab+=plane
|
|
bic r4, r4, #0x00ff0000 @ just in case
|
|
ldrh r3, [lr, r4] @ r3=hscroll
|
|
|
|
tst r7, #4
|
|
bne .DrawStrip_vsscroll
|
|
|
|
@ Get vertical scroll value:
|
|
add r7, lr, #0x012000
|
|
add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)
|
|
ldr r7, [r7]
|
|
|
|
tst r8, #2
|
|
tstne r8, #4
|
|
bne .DrawStrip_interlace
|
|
|
|
tst r0, r0
|
|
movne r7, r7, lsr #16
|
|
|
|
@ Find the line in the name table
|
|
add r2, r2, r7
|
|
and r2, r2, r1
|
|
mov r4, r2, lsr #3
|
|
add r10, r10, #1 @ shift[width]++
|
|
add r12, r12, r4, lsl r10 @ nametab+=(ts.line>>3)<<shift[width];
|
|
|
|
@ ldmia r0, {r1,r2,r3,r5,r6,r9} @ r2=line, r3=ts->hscroll, r5=ts->xmask, r6=ts->hc, r9=ts->cells
|
|
|
|
and r10,r2, #7
|
|
mov r10,r10, lsl #1 @ r10=ty=(ts->line&7)<<1;
|
|
orr r10,r10, r9, lsl #24
|
|
|
|
rsb r8, r3, #0
|
|
mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3
|
|
|
|
sub r1, r3, #1
|
|
and r1, r1, #7
|
|
add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1
|
|
|
|
tst r9, #1<<31
|
|
mov r3, #0
|
|
orrne r10,r10, #1<<23 @ r10=(cells<<24|sh<<23|hi_not_empty<<22|had_output<<21|ty)
|
|
movne r3, #0x40 @ default to shadowed pal on sh mode
|
|
|
|
cmp r7, #8
|
|
addne r10,r10, #0x01000000 @ we will loop cells+1 times if there is scroll
|
|
|
|
and r9, r9, #0xff00
|
|
add r8, r8, r9, lsr #8 @ tilex+=cellskip
|
|
add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;
|
|
sub r10,r10,r9, lsl #16 @ cells-=cellskip
|
|
|
|
@ cache some stuff to avoid mem access
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r11,=HighCol
|
|
mov r0, #0xf
|
|
ldr r11,[r11]
|
|
.else
|
|
ldr r11,=HighCol
|
|
mov r0, #0xf
|
|
.endif
|
|
|
|
mvn r9, #0 @ r9=prevcode=-1
|
|
add r1, r11, r7 @ r1=pdest
|
|
|
|
|
|
@ r4 & r7 are scratch in this loop
|
|
.dsloop_subr1:
|
|
sub r1, r1, #8
|
|
.dsloop: @ 40-41 times
|
|
subs r10,r10, #0x01000000
|
|
bmi .dsloop_exit
|
|
|
|
.dsloop_enter:
|
|
and r7, r5, r8
|
|
add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)
|
|
ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)
|
|
|
|
add r1, r1, #8
|
|
add r8, r8, #1
|
|
|
|
tst r7, #0x8000
|
|
bne .DrawStrip_hiprio
|
|
|
|
cmp r7, r9
|
|
beq .DrawStrip_samecode @ we know stuff about this tile already
|
|
|
|
mov r9, r7 @ remember code
|
|
orr r10, r10, #1<<21 @ seen non hi-prio tile
|
|
|
|
movs r2, r9, lsl #20 @ if (code&0x1000)
|
|
mov r2, r2, lsl #1
|
|
add r2, r2, r10, lsl #17
|
|
mov r2, r2, lsr #17
|
|
eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;
|
|
|
|
ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
|
|
bic r7, r3, #0x3f
|
|
and r3, r9, #0x6000
|
|
add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);
|
|
|
|
.DrawStrip_samecode:
|
|
tst r2, r2
|
|
beq .dsloop @ tileline blank
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .DrawStrip_SingleColor @ tileline singlecolor
|
|
|
|
tst r9, #0x0800
|
|
bne .DrawStrip_TileFlip
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern
|
|
.DrawStrip_TileNorm:
|
|
TileNorm r0
|
|
b .dsloop
|
|
|
|
.DrawStrip_TileFlip:
|
|
TileFlip r0
|
|
b .dsloop
|
|
|
|
.DrawStrip_SingleColor:
|
|
and r4, r2, #0xf
|
|
orr r4, r3, r4
|
|
orr r4, r4, r4, lsl #8
|
|
tst r1, #1 @ not aligned?
|
|
strneb r4, [r1], #1
|
|
streqh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strneb r4, [r1], #1 @ have a remaining unaligned pixel?
|
|
b .dsloop_subr1
|
|
|
|
.DrawStrip_hiprio_maybempt:
|
|
cmp r7, r9
|
|
beq .dsloop @ must've been empty, otherwise we wouldn't get here
|
|
movs r2, r7, lsl #20 @ if (code&0x1000)
|
|
mov r2, r2, lsl #1
|
|
add r2, r2, r10, lsl #17
|
|
mov r2, r2, lsr #17
|
|
eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;
|
|
ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
mov r9, r7 @ remember code
|
|
tst r2, r2
|
|
beq .dsloop
|
|
orr r10, r10, #1<<22
|
|
|
|
.DrawStrip_hiprio:
|
|
tst r10, #0x00c00000
|
|
beq .DrawStrip_hiprio_maybempt
|
|
sub r0, r1, r11
|
|
orr r7, r7, r0, lsl #16
|
|
orr r7, r7, r10, lsl #25 @ (ty<<25)
|
|
tst r7, #0x1000
|
|
eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;
|
|
str r7, [r6], #4 @ cache hi priority tile
|
|
mov r0, #0xf
|
|
b .dsloop
|
|
|
|
.dsloop_exit:
|
|
tst r10, #1<<21 @ seen non hi-prio tile
|
|
ldreq r1, =rendstatus
|
|
mov r0, #0
|
|
ldreq r2, [r1]
|
|
str r0, [r6] @ terminate the cache list
|
|
orreq r2, r2, #0x40 @ had a layer with all hi-prio tiles
|
|
streq r2, [r1]
|
|
|
|
ldmfd sp!, {r4-r11,lr}
|
|
bx lr
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
.DrawStrip_vsscroll:
|
|
rsb r8, r3, #0
|
|
mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3
|
|
bic r8, r8, #0xff000000
|
|
orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|tilex[23:0])
|
|
|
|
ldr r4, =Scanline
|
|
orr r5, r1, r10, lsl #24
|
|
ldr r4, [r4]
|
|
sub r1, r3, #1
|
|
orr r5, r5, r4, lsl #16 @ r5=(shift_width[31:24]|scanline[23:16]|ymask[15:0])
|
|
and r1, r1, #7
|
|
add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1
|
|
|
|
mov r10,r9, lsl #16
|
|
tst r0, #1
|
|
orrne r10,r10, #0x8000
|
|
tst r9, #1<<31
|
|
mov r3, #0
|
|
orr r10,r10, #0xff000000 @ will be adjusted on entering loop
|
|
orrne r10,r10, #1<<23 @ r10=(cell[31:24]|sh[23]|hi_not_empty[22]|cells_max[21:16]|plane[15]|ty[14:0])
|
|
movne r3, #0x40 @ default to shadowed pal on sh mode
|
|
|
|
cmp r7, #8
|
|
subne r10,r10, #0x01000000 @ have hscroll, start with negative cell
|
|
|
|
and r9, r9, #0xff00
|
|
add r8, r8, r9, lsr #8 @ tilex+=cellskip
|
|
add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;
|
|
add r10,r10,r9, lsl #16 @ cell+=cellskip
|
|
|
|
@ cache some stuff to avoid mem access
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r11,=HighCol
|
|
mov r0, #0xf
|
|
ldr r11,[r11]
|
|
.else
|
|
ldr r11,=HighCol
|
|
mov r0, #0xf
|
|
.endif
|
|
|
|
mvn r9, #0 @ r9=prevcode=-1
|
|
add r1, r11, r7 @ r1=pdest
|
|
|
|
@ r4 & r7 are scratch in this loop
|
|
.dsloop_vs_subr1:
|
|
sub r1, r1, #8
|
|
.dsloop_vs: @ 40-41 times
|
|
add r10,r10, #0x01000000
|
|
and r4, r10, #0x003f0000
|
|
cmp r4, r10, asr #8
|
|
ble .dsloop_vs_exit
|
|
|
|
@ calc offset and read tileline code to r7, also calc ty
|
|
add r7, lr, #0x012000
|
|
add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)
|
|
add r7, r7, r10,asr #23 @ vsram + ((cell&~1)<<1)
|
|
bic r7, r7, #3
|
|
tst r10,#0x8000 @ plane1?
|
|
addne r7, r7, #2
|
|
ldrh r7, [r7] @ r7=vscroll
|
|
|
|
bic r10,r10,#0xff @ clear old ty
|
|
and r4, r5, #0xff0000
|
|
add r4, r4, r7, lsl #16
|
|
and r4, r4, r5, lsl #16 @ r4=line<<16
|
|
and r7, r4, #0x70000
|
|
orr r10,r10,r7, lsr #15 @ new ty
|
|
|
|
mov r4, r4, lsr #19
|
|
mov r7, r5, lsr #24
|
|
mov r4, r4, lsl r7 @ nametabadd
|
|
|
|
and r7, r8, r8, lsr #25
|
|
add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)
|
|
add r7, r7, r4, lsl #1
|
|
ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)
|
|
|
|
add r1, r1, #8
|
|
add r8, r8, #1
|
|
|
|
tst r7, #0x8000
|
|
bne .DrawStrip_vs_hiprio
|
|
|
|
cmp r7, r9
|
|
beq .DrawStrip_vs_samecode @ we know stuff about this tile already
|
|
|
|
mov r9, r7 @ remember code
|
|
orr r8, r8, #1<<24 @ seen non hi-prio tile
|
|
|
|
movs r2, r9, lsl #20 @ if (code&0x1000)
|
|
mov r2, r2, lsl #1
|
|
add r2, r2, r10, lsl #17
|
|
mov r2, r2, lsr #17
|
|
eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;
|
|
|
|
ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
|
|
bic r7, r3, #0x3f
|
|
and r3, r9, #0x6000
|
|
add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);
|
|
|
|
.DrawStrip_vs_samecode:
|
|
tst r2, r2
|
|
beq .dsloop_vs @ tileline blank
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .DrawStrip_vs_SingleColor @ tileline singlecolor
|
|
|
|
tst r9, #0x0800
|
|
bne .DrawStrip_vs_TileFlip
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern
|
|
.DrawStrip_vs_TileNorm:
|
|
TileNorm r0
|
|
b .dsloop_vs
|
|
|
|
.DrawStrip_vs_TileFlip:
|
|
TileFlip r0
|
|
b .dsloop_vs
|
|
|
|
.DrawStrip_vs_SingleColor:
|
|
and r4, r2, #0xf
|
|
orr r4, r3, r4
|
|
orr r4, r4, r4, lsl #8
|
|
tst r1, #1 @ not aligned?
|
|
strneb r4, [r1], #1
|
|
streqh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strneb r4, [r1], #1 @ have a remaining unaligned pixel?
|
|
b .dsloop_vs_subr1
|
|
|
|
.DrawStrip_vs_hiprio:
|
|
tst r10, #0x00c00000
|
|
beq .DrawStrip_vs_hiprio_maybempt
|
|
sub r0, r1, r11
|
|
orr r7, r7, r0, lsl #16
|
|
orr r7, r7, r10, lsl #25 @ (ty<<25)
|
|
tst r7, #0x1000
|
|
eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;
|
|
str r7, [r6], #4 @ cache hi priority tile
|
|
mov r0, #0xf
|
|
b .dsloop_vs
|
|
|
|
.DrawStrip_vs_hiprio_maybempt:
|
|
cmp r7, r9
|
|
beq .dsloop_vs @ must've been empty, otherwise we wouldn't get here
|
|
movs r2, r7, lsl #20 @ if (code&0x1000)
|
|
mov r2, r2, lsl #1
|
|
add r2, r2, r10, lsl #17
|
|
mov r2, r2, lsr #17
|
|
eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;
|
|
ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
mov r9, r7 @ remember code
|
|
tst r2, r2
|
|
orrne r10, r10, #1<<22
|
|
bne .DrawStrip_vs_hiprio
|
|
b .dsloop_vs
|
|
|
|
.dsloop_vs_exit:
|
|
tst r8, #1<<24 @ seen non hi-prio tile
|
|
ldreq r1, =rendstatus
|
|
mov r0, #0
|
|
ldreq r2, [r1]
|
|
str r0, [r6] @ terminate the cache list
|
|
orreq r2, r2, #0x40 @ had a layer with all hi-prio tiles
|
|
streq r2, [r1]
|
|
|
|
ldmfd sp!, {r4-r11,lr}
|
|
bx lr
|
|
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
@ interlace mode 2? Sonic 2?
|
|
.DrawStrip_interlace:
|
|
tst r0, r0
|
|
moveq r7, r7, lsl #21
|
|
movne r7, r7, lsl #5
|
|
|
|
@ Find the line in the name table
|
|
add r2, r7, r2, lsl #22 @ r2=(vscroll+(Scanline<<1))<<21 (11 bits);
|
|
orr r1, r1, #0x80000000
|
|
and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;
|
|
mov r2, r2, lsr #21
|
|
mov r4, r2, lsr #4
|
|
mov r12, r12, lsr #1 @ halfwords
|
|
add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];
|
|
and r9, r9, #0xff
|
|
|
|
sub sp, sp, #6*4
|
|
stmia sp, {r0,r2,r3,r5,r6,r9}
|
|
|
|
mov r0, sp
|
|
bl DrawStripInterlace @ struct TileStrip *ts
|
|
|
|
add sp, sp, #6*4
|
|
ldmfd sp!, {r4-r11,lr}
|
|
bx lr
|
|
|
|
.pool
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
|
|
.global BackFill @ int reg7, int sh
|
|
|
|
BackFill:
|
|
stmfd sp!, {r4-r9,lr}
|
|
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr lr, =HighCol
|
|
mov r0, r0, lsl #26
|
|
ldr lr, [lr]
|
|
mov r0, r0, lsr #26
|
|
add lr, lr, #8
|
|
.else
|
|
ldr lr, =(HighCol+8)
|
|
mov r0, r0, lsl #26
|
|
mov r0, r0, lsr #26
|
|
.endif
|
|
|
|
orr r0, r0, r1, lsl #6
|
|
orr r0, r0, r0, lsl #8
|
|
orr r0, r0, r0, lsl #16
|
|
|
|
mov r1, r0
|
|
mov r2, r0
|
|
mov r3, r0
|
|
mov r4, r0
|
|
mov r5, r0
|
|
mov r6, r0
|
|
mov r7, r0
|
|
|
|
@ go go go!
|
|
stmia lr!, {r0-r7} @ 10*8*4
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
stmia lr!, {r0-r7}
|
|
|
|
ldmfd sp!, {r4-r9,r12}
|
|
bx r12
|
|
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
|
|
.global DrawTilesFromCache @ int *hc, int sh, int rlim
|
|
|
|
DrawTilesFromCache:
|
|
stmfd sp!, {r4-r8,r11,lr}
|
|
|
|
@ cache some stuff to avoid mem access
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r11,=HighCol
|
|
mov r12,#0xf
|
|
ldr r11,[r11]
|
|
.else
|
|
ldr r11,=HighCol
|
|
mov r12,#0xf
|
|
.endif
|
|
ldr lr, =(Pico+0x10000) @ lr=Pico.vram
|
|
|
|
mvn r5, #0 @ r5=prevcode=-1
|
|
ands r8, r1, #1
|
|
orr r8, r8, r2, lsl #1
|
|
bne .dtfc_check_rendflags
|
|
|
|
@ scratch: r4, r7
|
|
.dtfc_loop:
|
|
ldr r6, [r0], #4 @ read code
|
|
movs r1, r6, lsr #16 @ r1=dx;
|
|
ldmeqfd sp!, {r4-r8,r11,pc} @ dx is never zero, this must be a terminator, return
|
|
bic r4, r1, #0xfe00
|
|
add r1, r11, r4 @ r1=pdest
|
|
|
|
mov r7, r6, lsl #16
|
|
cmp r5, r7, lsr #16
|
|
beq .dtfc_samecode @ if (code==prevcode)
|
|
|
|
mov r5, r7, lsr #16
|
|
|
|
mov r2, r5, lsl #21
|
|
mov r2, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4;
|
|
add r2, r2, r6, lsr #25 @ addr+=ty
|
|
|
|
and r3, r5, #0x6000
|
|
mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);
|
|
|
|
ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
|
|
.dtfc_samecode:
|
|
rsbs r4, r4, r8, lsr #1
|
|
bmi .dtfc_cut_tile
|
|
|
|
tst r8, #1
|
|
bne .dtfc_shadow
|
|
|
|
tst r2, r2
|
|
beq .dtfc_loop
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .dtfc_SingleColor @ tileline singlecolor
|
|
|
|
tst r5, #0x0800
|
|
bne .dtfc_TileFlip
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern
|
|
.dtfc_TileNorm:
|
|
TileNorm r12
|
|
b .dtfc_loop
|
|
|
|
.dtfc_TileFlip:
|
|
TileFlip r12
|
|
b .dtfc_loop
|
|
|
|
.dtfc_SingleColor:
|
|
and r4, r2, #0xf
|
|
orr r4, r3, r4
|
|
orr r4, r4, r4, lsl #8
|
|
tst r1, #1 @ not aligned?
|
|
strneb r4, [r1], #1
|
|
streqh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strneb r4, [r1], #1 @ have a remaining unaligned pixel?
|
|
b .dtfc_loop
|
|
|
|
.dtfc_shadow:
|
|
tst r2, r2
|
|
beq .dtfc_shadow_blank
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .dtfc_SingleColor @ tileline singlecolor
|
|
|
|
tst r5, #0x0800
|
|
bne .dtfc_TileFlipShHP
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern
|
|
.dtfc_TileNormShHP:
|
|
TileNormShHP
|
|
b .dtfc_loop
|
|
|
|
.dtfc_TileFlipShHP:
|
|
TileFlipShHP
|
|
b .dtfc_loop
|
|
|
|
.dtfc_shadow_blank:
|
|
ldrb r4, [r1] @ 1ci
|
|
ldrb r12,[r1,#1]
|
|
tst r4, #0x80
|
|
andeq r4, r4,#0x3f
|
|
streqb r4, [r1]
|
|
tst r12,#0x80
|
|
ldrb r4, [r1,#2]
|
|
andeq r12,r12,#0x3f
|
|
streqb r12,[r1,#1]
|
|
tst r4, #0x80
|
|
ldrb r12,[r1,#3]
|
|
andeq r4, r4,#0x3f
|
|
streqb r4, [r1,#2]
|
|
tst r12,#0x80
|
|
ldrb r4, [r1,#4]
|
|
andeq r12,r12,#0x3f
|
|
streqb r12,[r1,#3]
|
|
tst r4, #0x80
|
|
ldrb r12,[r1,#5]
|
|
andeq r4, r4,#0x3f
|
|
streqb r4, [r1,#4]
|
|
tst r12,#0x80
|
|
ldrb r4, [r1,#6]
|
|
andeq r12,r12,#0x3f
|
|
streqb r12,[r1,#5]
|
|
tst r4, #0x80
|
|
ldrb r12,[r1,#7]
|
|
andeq r4, r4,#0x3f
|
|
streqb r4, [r1,#6]
|
|
tst r12,#0x80
|
|
andeq r12,r12,#0x3f
|
|
streqb r12,[r1,#7]
|
|
mov r12, #0xf
|
|
b .dtfc_loop
|
|
|
|
.dtfc_cut_tile:
|
|
add r4, r4, #7 @ 0-6
|
|
mov r4, r4, lsl #2
|
|
mov r12,#0xf<<28
|
|
mov r12,r12,asr r4
|
|
mov r2, r2, ror #16
|
|
tst r5, #0x0800 @ flipped?
|
|
mvnne r12,r12
|
|
and r2, r2, r12
|
|
mov r2, r2, ror #16
|
|
mov r12,#0xf
|
|
tst r8, #1
|
|
bne .dtfc_shadow
|
|
tst r2, r2
|
|
beq .dtfc_loop
|
|
tst r5, #0x0800
|
|
beq .dtfc_TileNorm
|
|
b .dtfc_TileFlip
|
|
|
|
@ check if we have detected layer covered with hi-prio tiles:
|
|
.dtfc_check_rendflags:
|
|
ldr r1, =rendstatus
|
|
ldr r2, [r1]
|
|
tst r2, #0xc0
|
|
beq .dtfc_loop
|
|
bic r8, r8, #1 @ sh/hi mode off
|
|
tst r2, #0x80
|
|
bne .dtfc_loop @ already processed
|
|
orr r2, r2, #0x80
|
|
str r2, [r1]
|
|
|
|
add r1, r11,#8
|
|
mov r3, #320/4
|
|
mov r7, #0x80
|
|
orr r7, r7, r7, lsl #8
|
|
orr r7, r7, r7, lsl #16
|
|
mov r6, #0x3f
|
|
orr r6, r6, r6, lsl #8
|
|
orr r6, r6, r6, lsl #16
|
|
.dtfc_loop_shprep:
|
|
subs r3, r3, #1
|
|
bmi .dtfc_loop @ done
|
|
ldr r2, [r1]
|
|
tst r2, r7
|
|
andeq r2, r2, r6
|
|
streq r2, [r1], #4
|
|
beq .dtfc_loop_shprep
|
|
tst r2, #0x80000000
|
|
biceq r2, r2, #0xc0000000
|
|
tst r2, #0x00800000
|
|
biceq r2, r2, #0x00c00000
|
|
tst r2, #0x00008000
|
|
biceq r2, r2, #0x0000c000
|
|
tst r2, #0x00000080
|
|
biceq r2, r2, #0x000000c0
|
|
str r2, [r1], #4
|
|
b .dtfc_loop_shprep
|
|
|
|
.pool
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
|
|
.global DrawSpritesFromCache @ int *hc, int sh
|
|
|
|
DrawSpritesFromCache:
|
|
stmfd sp!, {r4-r11,lr}
|
|
|
|
@ cache some stuff to avoid mem access
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r11,=HighCol
|
|
mov r12,#0xf
|
|
ldr r11,[r11]
|
|
.else
|
|
ldr r11,=HighCol
|
|
mov r12,#0xf
|
|
.endif
|
|
ldr lr, =(Pico+0x10000) @ lr=Pico.vram
|
|
mov r6, r1, lsl #31
|
|
orr r6, r6, #1<<30
|
|
|
|
mov r10, r0
|
|
|
|
.dsfc_loop:
|
|
ldr r9, [r10], #4 @ read code
|
|
bic r6, r6, #7 @ using pipeline
|
|
tst r9, r9
|
|
ldmeqfd sp!, {r4-r11,pc}
|
|
|
|
mov r4, r9, lsl #28
|
|
orr r6, r6, r4, lsr #30
|
|
add r6, r6, #1 @ r6=s1cc???? ... ?????www (s=shadow/hilight, cc=pal, w=width)
|
|
|
|
and r5, r9, #3
|
|
add r5, r5, #1 @ r5=delta
|
|
tst r9, #0x10000
|
|
rsbne r5, r5, #0 @ Flip X
|
|
mov r5, r5, lsl #4
|
|
|
|
mov r2, r9, lsr #17
|
|
mov r8, r2, lsl #1 @ tile=((unsigned int)code>>17)<<1;
|
|
|
|
and r3, r9, #0x30 @ r3=pal=(code&0x30);
|
|
|
|
bic r6, r6, #3<<28
|
|
orr r6, r6, r3, lsl #24
|
|
|
|
mov r0, r9, lsl #16
|
|
mov r0, r0, asr #22 @ sx=(code<<16)>>22
|
|
adds r0, r0, #0 @ set ZV
|
|
b .dsfc_inloop_enter
|
|
|
|
@ scratch: r4, r7
|
|
.dsfc_inloop:
|
|
sub r6, r6, #1
|
|
tst r6, #7
|
|
beq .dsfc_loop
|
|
adds r0, r0, #8
|
|
add r8, r8, r5
|
|
|
|
.dsfc_inloop_enter:
|
|
ble .dsfc_inloop
|
|
cmp r0, #328
|
|
bge .dsfc_loop
|
|
|
|
mov r8, r8, lsl #17
|
|
mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address
|
|
|
|
ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+tile); // Get 8 pixels
|
|
add r1, r11, r0 @ r1=pdest
|
|
tst r2, r2
|
|
beq .dsfc_inloop
|
|
|
|
cmp r12, r6, lsr #28
|
|
beq .dsfc_shadow
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .dsfc_SingleColor @ tileline singlecolor
|
|
|
|
tst r9, #0x10000
|
|
bne .dsfc_TileFlip
|
|
|
|
@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern
|
|
.dsfc_TileNorm:
|
|
TileNorm r12
|
|
b .dsfc_inloop
|
|
|
|
.dsfc_TileFlip:
|
|
TileFlip r12
|
|
b .dsfc_inloop
|
|
|
|
.dsfc_SingleColor:
|
|
tst r0, #1 @ not aligned?
|
|
and r4, r2, #0xf
|
|
orr r4, r3, r4
|
|
orr r4, r4, r4, lsl #8
|
|
strneb r4, [r1], #1
|
|
streqh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strneb r4, [r1], #1
|
|
b .dsfc_inloop
|
|
|
|
.dsfc_shadow:
|
|
cmp r2, r2, ror #4
|
|
beq .dsfc_singlec_sh
|
|
|
|
tst r9, #0x10000
|
|
bne .dsfc_TileFlip_sh
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern
|
|
.dsfc_TileNorm_sh:
|
|
TileNormSh
|
|
b .dsfc_inloop
|
|
|
|
.dsfc_TileFlip_sh:
|
|
TileFlipSh
|
|
b .dsfc_inloop
|
|
|
|
.dsfc_singlec_sh:
|
|
cmp r2, #0xe0000000
|
|
bcc .dsfc_SingleColor @ normal singlecolor tileline (carry inverted in ARM)
|
|
tst r2, #0x10000000
|
|
bne .dsfc_sh_sh
|
|
TileSingleHi
|
|
b .dsfc_inloop
|
|
|
|
.dsfc_sh_sh:
|
|
TileSingleSh
|
|
b .dsfc_inloop
|
|
|
|
.pool
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
@ + 0 : hhhhvvvv ab--hhvv yyyyyyyy yyyyyyyy // a: offscreen h, b: offs. v, h: horiz. size
|
|
@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8
|
|
|
|
.global DrawSprite @ unsigned int *sprite, int **hc, int sh
|
|
|
|
DrawSprite:
|
|
stmfd sp!, {r4-r9,r11,lr}
|
|
|
|
ldr r3, [r0] @ sprite[0]
|
|
ldr r7, =Scanline
|
|
mov r6, r3, lsr #28
|
|
sub r6, r6, #1 @ r6=width-1 (inc later)
|
|
mov r5, r3, lsr #24
|
|
and r5, r5, #7 @ r5=height
|
|
|
|
mov r4, r3, lsl #16 @ r4=sy<<16 (tmp)
|
|
|
|
ldr r7, [r7]
|
|
ldr r9, [r0, #4]
|
|
sub r7, r7, r4, asr #16 @ r7=row=Scanline-sy
|
|
|
|
tst r2, r2
|
|
mov r2, r9, asr #16 @ r2=sx
|
|
bic r9, r9, #0xfe000000
|
|
orrne r9, r9, #1<<31 @ r9=code|(sh<<31)
|
|
|
|
tst r9, #0x1000
|
|
movne r4, r5, lsl #3
|
|
subne r4, r4, #1
|
|
subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y
|
|
|
|
mov r8, r9, lsl #21
|
|
mov r8, r8, lsr #21
|
|
add r8, r8, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down
|
|
|
|
tst r9, #0x0800
|
|
mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);
|
|
rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now
|
|
|
|
mov r8, r8, lsl #4
|
|
and r7, r7, #7
|
|
add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address
|
|
|
|
tst r9, #0x8000
|
|
bne .dspr_cache @ if(code&0x8000) // high priority - cache it
|
|
|
|
@ cache some stuff to avoid mem access
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r11,=HighCol
|
|
mov r12,#0xf
|
|
ldr r11,[r11]
|
|
.else
|
|
ldr r11,=HighCol
|
|
mov r12,#0xf
|
|
.endif
|
|
ldr lr, =(Pico+0x10000) @ lr=Pico.vram
|
|
|
|
mov r5, r5, lsl #4 @ delta<<=4; // Delta of address
|
|
and r4, r9, #0x6000
|
|
orr r9, r9, r4, lsl #16
|
|
orr r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)
|
|
|
|
tst r9, #1<<31
|
|
mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);
|
|
orrne r3, r3, #0x40 @ shadow by default
|
|
|
|
add r6, r6, #1 @ inc now
|
|
adds r0, r2, #0 @ mov sx to r0 and set ZV flags
|
|
b .dspr_loop_enter
|
|
|
|
.dspr_loop:
|
|
subs r6, r6, #1 @ width--
|
|
ldmeqfd sp!, {r4-r9,r11,pc}@ return
|
|
adds r0, r0, #8 @ sx+=8
|
|
add r8, r8, r5 @ tile+=delta
|
|
|
|
.dspr_loop_enter:
|
|
ble .dspr_loop @ sx <= 0
|
|
cmp r0, #328
|
|
ldmgefd sp!, {r4-r9,r11,pc}@ return
|
|
|
|
mov r8, r8, lsl #17
|
|
mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address
|
|
|
|
ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
add r1, r11, r0 @ r1=pdest
|
|
tst r2, r2
|
|
beq .dspr_loop
|
|
|
|
cmp r12, r9, lsr #28
|
|
beq .dspr_shadow
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .dspr_SingleColor @ tileline singlecolor
|
|
|
|
tst r9, #0x0800
|
|
bne .dspr_TileFlip
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern
|
|
@ scratch: r4, r7
|
|
.dspr_TileNorm:
|
|
TileNorm r12
|
|
b .dspr_loop
|
|
|
|
.dspr_TileFlip:
|
|
TileFlip r12
|
|
b .dspr_loop
|
|
|
|
.dspr_SingleColor:
|
|
and r4, r2, #0xf
|
|
orr r4, r3, r4
|
|
orr r4, r4, r4, lsl #8
|
|
tst r0, #1 @ not aligned?
|
|
strneb r4, [r1], #1
|
|
streqh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strh r4, [r1], #2
|
|
strneb r4, [r1], #1
|
|
b .dspr_loop
|
|
|
|
.dspr_shadow:
|
|
cmp r2, r2, ror #4
|
|
beq .dspr_singlec_sh
|
|
|
|
tst r9, #0x0800
|
|
bne .dspr_TileFlip_sh
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern
|
|
.dspr_TileNorm_sh:
|
|
TileNormSh
|
|
b .dspr_loop
|
|
|
|
.dspr_TileFlip_sh:
|
|
TileFlipSh
|
|
b .dspr_loop
|
|
|
|
.dspr_singlec_sh:
|
|
cmp r2, #0xe0000000
|
|
bcc .dspr_SingleColor @ normal tileline
|
|
tst r2, #0x10000000
|
|
bne .dspr_sh_sh
|
|
TileSingleHi
|
|
b .dspr_loop
|
|
|
|
.dspr_sh_sh:
|
|
TileSingleSh
|
|
b .dspr_loop
|
|
|
|
|
|
.dspr_cache:
|
|
@ *(*hc)++ = (tile<<16)|((code&0x0800)<<5)|((sx<<6)&0x0000ffc0)|((code>>9)&0x30)|((sprite[0]>>24)&0xf);
|
|
mov r4, r8, lsl #16 @ tile
|
|
tst r9, #0x0800
|
|
orrne r4, r4, #0x10000 @ code&0x0800
|
|
mov r2, r2, lsl #22
|
|
orr r4, r4, r2, lsr #16 @ (sx<<6)&0x0000ffc0
|
|
and r2, r9, #0x6000
|
|
orr r4, r4, r2, lsr #9 @ (code>>9)&0x30
|
|
mov r3, r3, lsl #12
|
|
ldr r2, [r1]
|
|
orr r4, r4, r3, lsr #28 @ (sprite[0]>>24)&0xf
|
|
|
|
str r4, [r2], #4
|
|
str r2, [r1]
|
|
|
|
ldmfd sp!, {r4-r9,r11,lr}
|
|
bx lr
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache
|
|
|
|
DrawWindow:
|
|
stmfd sp!, {r4-r11,lr}
|
|
|
|
ldr r11, =(Pico+0x22228) @ Pico.video
|
|
ldr r10, =Scanline
|
|
ldrb r12, [r11, #3] @ pvid->reg[3]
|
|
|
|
ldr r10, [r10]
|
|
ldr r4, [r11, #12]
|
|
mov r5, r10, lsr #3
|
|
and r10, r10, #7
|
|
mov r10, r10, lsl #1 @ r10=ty
|
|
|
|
mov r12, r12, lsl #10
|
|
|
|
tst r4, #1 @ 40 cell mode?
|
|
andne r12, r12, #0xf000 @ 0x3c<<10
|
|
andeq r12, r12, #0xf800
|
|
addne r12, r12, r5, lsl #7
|
|
addeq r12, r12, r5, lsl #6 @ nametab
|
|
add r12, r12, r0, lsl #2 @ +starttile
|
|
|
|
ldr r6, =rendstatus
|
|
ldr lr, =(Pico+0x10000) @ lr=Pico.vram
|
|
ldrb r6, [r6]
|
|
|
|
@ fetch the first code now
|
|
ldrh r7, [lr, r12]
|
|
|
|
ands r6, r6, #2 @ we care about bit 1 only
|
|
orr r6, r6, r2
|
|
bne .dw_no_sameprio
|
|
|
|
cmp r2, r7, lsr #15
|
|
ldmnefd sp!, {r4-r11,pc} @ assume that whole window uses same priority
|
|
|
|
.dw_no_sameprio:
|
|
orr r6, r6, r3, lsl #8 @ shadow mode
|
|
|
|
sub r8, r1, r0
|
|
|
|
@ cache some stuff to avoid mem access
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r11,=HighCol
|
|
mov r8, r8, lsl #1 @ cells
|
|
ldr r11,[r11]
|
|
mvn r9, #0 @ r9=prevcode=-1
|
|
add r11,r11,#8
|
|
.else
|
|
ldr r11,=(HighCol+8)
|
|
mov r8, r8, lsl #1 @ cells
|
|
mvn r9, #0 @ r9=prevcode=-1
|
|
.endif
|
|
add r1, r11, r0, lsl #4 @ r1=pdest
|
|
mov r0, #0xf
|
|
b .dwloop_enter
|
|
|
|
@ r4,r5 & r7 are scratch in this loop
|
|
.dwloop:
|
|
add r1, r1, #8
|
|
.dwloop_nor1:
|
|
add r12, r12, #2 @ halfwords
|
|
ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)
|
|
subs r8, r8, #1
|
|
beq .dwloop_end @ done
|
|
|
|
eor r5, r6, r7, lsr #15
|
|
tst r5, #1
|
|
orrne r6, r6, #2 @ wrong pri
|
|
bne .dwloop
|
|
|
|
cmp r7, r9
|
|
beq .dw_samecode @ we know stuff about this tile already
|
|
|
|
.dwloop_enter:
|
|
mov r9, r7 @ remember code
|
|
|
|
movs r2, r9, lsl #20 @ if (code&0x1000)
|
|
mov r2, r2, lsl #1
|
|
add r2, r10, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4; addr+=ty
|
|
eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;
|
|
|
|
and r3, r9, #0x6000
|
|
mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);
|
|
|
|
ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels
|
|
|
|
.dw_samecode:
|
|
tst r6, #0x100
|
|
bne .dw_shadow
|
|
.dw_shadow_done:
|
|
tst r2, r2
|
|
beq .dwloop @ tileline blank
|
|
|
|
cmp r2, r2, ror #4
|
|
beq .dw_SingleColor @ tileline singlecolor
|
|
|
|
tst r9, #0x0800
|
|
bne .dw_TileFlip
|
|
|
|
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern
|
|
.dw_TileNorm:
|
|
TileNorm r0
|
|
b .dwloop
|
|
|
|
.dw_TileFlip:
|
|
TileFlip r0
|
|
b .dwloop
|
|
|
|
.dw_SingleColor:
|
|
and r4, r0, r2 @ #0x0000000f
|
|
orr r4, r3, r4
|
|
orr r4, r4, r4, lsl #8
|
|
orr r4, r4, r4, lsl #16
|
|
mov r5, r4
|
|
stmia r1!, {r4,r5}
|
|
b .dwloop_nor1 @ we incremeted r1 ourselves
|
|
|
|
.dw_shadow:
|
|
tst r6, #1 @ hi pri?
|
|
orreq r3, r3, #0x40
|
|
beq .dw_shadow_done
|
|
ldr r4, [r1]
|
|
tst r4, #0x00000080
|
|
biceq r4, r4, #0x000000c0
|
|
tst r4, #0x00008000
|
|
biceq r4, r4, #0x0000c000
|
|
tst r4, #0x00800000
|
|
biceq r4, r4, #0x00c00000
|
|
tst r4, #0x80000000
|
|
biceq r4, r4, #0xc0000000
|
|
str r4, [r1]
|
|
ldr r4, [r1,#4]
|
|
tst r4, #0x00000080
|
|
biceq r4, r4, #0x000000c0
|
|
tst r4, #0x00008000
|
|
biceq r4, r4, #0x0000c000
|
|
tst r4, #0x00800000
|
|
biceq r4, r4, #0x00c00000
|
|
tst r4, #0x80000000
|
|
biceq r4, r4, #0xc0000000
|
|
str r4, [r1,#4]
|
|
b .dw_shadow_done
|
|
|
|
.dwloop_end:
|
|
ldr r0, =rendstatus
|
|
ldr r1, [r0]
|
|
and r6, r6, #2
|
|
orr r1, r1, r6
|
|
str r1, [r0]
|
|
|
|
ldmfd sp!, {r4-r11,r12}
|
|
bx r12
|
|
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
|
|
@ hilights 2 pixels in RGB444/BGR444 format
|
|
.macro TileDoShHi2Pixels444 reg
|
|
mov \reg, \reg, ror #12
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #28
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #28
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #24
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #28
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #28
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #12
|
|
.endm
|
|
|
|
|
|
.global FinalizeLineBGR444 @ int sh
|
|
|
|
FinalizeLineBGR444:
|
|
stmfd sp!, {r4-r6,lr}
|
|
mov r6, r0
|
|
ldr lr, =(Pico+0x22228) @ Pico.video
|
|
ldr r0, =DrawLineDest
|
|
ldrb r12, [lr, #12]
|
|
ldr r0, [r0]
|
|
sub r3, lr, #0x128 @ r3=Pico.cram
|
|
|
|
tst r12, #1
|
|
movne r2, #320/4 @ len
|
|
bne .fl_no32colBGR444
|
|
ldr r4, =PicoOpt
|
|
mov r2, #256/4
|
|
ldr r4, [r4]
|
|
tst r4, #0x100
|
|
addeq r0, r0, #32*2
|
|
|
|
.fl_no32colBGR444:
|
|
tst r6, r6
|
|
beq .fl_noshBGR444
|
|
|
|
ldr r4, =HighPal
|
|
|
|
ldrb r12, [lr, #-0x1a] @ 0x2220e ~ dirtyPal
|
|
tst r12, r12
|
|
moveq r3, r4
|
|
beq .fl_noshBGR444
|
|
mov r12, #0
|
|
strb r12, [lr, #-0x1a]
|
|
|
|
mov lr, #0x40/8
|
|
@ copy pal:
|
|
.fl_loopcpBGR444:
|
|
ldmia r3!, {r1,r5,r6,r12}
|
|
subs lr, lr, #1
|
|
stmia r4!, {r1,r5,r6,r12}
|
|
bne .fl_loopcpBGR444
|
|
|
|
@ shadowed pixels:
|
|
mov r12, #0x0077
|
|
orr r12,r12,#0x0700
|
|
orr r12,r12,r12,lsl #16
|
|
sub r3, r3, #0x40*2
|
|
add r5, r4, #0x80*2
|
|
mov lr, #0x40/4
|
|
.fl_loopcpBGR444_sh:
|
|
ldmia r3!, {r1,r6}
|
|
subs lr, lr, #1
|
|
and r1, r12, r1, lsr #1
|
|
and r6, r12, r6, lsr #1
|
|
stmia r4!, {r1,r6}
|
|
stmia r5!, {r1,r6}
|
|
bne .fl_loopcpBGR444_sh
|
|
|
|
@ hilighted pixels:
|
|
sub r3, r3, #0x40*2
|
|
mov lr, #0x40/2
|
|
.fl_loopcpBGR444_hi:
|
|
ldr r1, [r3], #4
|
|
TileDoShHi2Pixels444 r1
|
|
str r1, [r4], #4
|
|
subs lr, lr, #1
|
|
bne .fl_loopcpBGR444_hi
|
|
|
|
sub r3, r4, #0x40*3*2
|
|
|
|
|
|
.fl_noshBGR444:
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r1, =HighCol
|
|
mov lr, #0xff
|
|
ldr r1, [r1]
|
|
mov lr, lr, lsl #1
|
|
add r1, r1, #8
|
|
.else
|
|
ldr r1, =(HighCol+8)
|
|
mov lr, #0xff
|
|
mov lr, lr, lsl #1
|
|
.endif
|
|
|
|
.fl_loopBGR444:
|
|
|
|
ldr r12, [r1], #4
|
|
subs r2, r2, #1
|
|
|
|
and r4, lr, r12, lsl #1
|
|
ldrh r4, [r3, r4]
|
|
and r5, lr, r12, lsr #7
|
|
ldrh r5, [r3, r5]
|
|
and r6, lr, r12, lsr #15
|
|
ldrh r6, [r3, r6]
|
|
orr r4, r4, r5, lsl #16
|
|
|
|
and r5, lr, r12, lsr #23
|
|
ldrh r5, [r3, r5] @ 2c.i.
|
|
orr r5, r6, r5, lsl #16
|
|
|
|
stmia r0!, {r4,r5}
|
|
bne .fl_loopBGR444
|
|
|
|
|
|
ldmfd sp!, {r4-r6,lr}
|
|
bx lr
|
|
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
|
|
@ hilights 2 pixels in RGB555/BGR555 format
|
|
.macro TileDoShHi2Pixels555 reg
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #27
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #26
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #27
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #27
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #26
|
|
adds \reg, \reg, #0x40000000
|
|
orrcs \reg, \reg, #0xf0000000
|
|
mov \reg, \reg, ror #27
|
|
.endm
|
|
|
|
|
|
@ Convert 0000bbb0 ggg0rrr0
|
|
@ to rrrrrggg gggbbbbb
|
|
|
|
@ r2,r3,r9 - scratch, lr = 0x001c001c, r8 = 0x00030003
|
|
.macro convRGB565 reg
|
|
and r2, lr, \reg,lsl #1
|
|
and r9, r8, \reg,lsr #2
|
|
orr r2, r2, r9 @ r2=red
|
|
and r3, lr, \reg,lsr #7
|
|
and r9, r8, \reg,lsr #10
|
|
orr r3, r3, r9 @ r3=blue
|
|
and \reg, \reg, lr, lsl #3
|
|
orr \reg, \reg, \reg,lsl #3 @ green
|
|
orr \reg, \reg, r2, lsl #11 @ add red back
|
|
orr \reg, \reg, r3 @ add blue back
|
|
.endm
|
|
|
|
.global vidConvCpyRGB565
|
|
|
|
vidConvCpyRGB565: @ void *to, void *from, int pixels
|
|
stmfd sp!, {r4-r9,lr}
|
|
|
|
mov r12, r2, lsr #3 @ repeats
|
|
mov lr, #0x001c0000
|
|
orr lr, lr, #0x01c @ lr == pattern 0x001c001c
|
|
mov r8, #0x00030000
|
|
orr r8, r8, #0x003
|
|
|
|
.loopRGB565:
|
|
ldmia r1!, {r4-r7}
|
|
subs r12, r12, #1
|
|
convRGB565 r4
|
|
str r4, [r0], #4
|
|
convRGB565 r5
|
|
str r5, [r0], #4
|
|
convRGB565 r6
|
|
str r6, [r0], #4
|
|
convRGB565 r7
|
|
str r7, [r0], #4
|
|
|
|
bgt .loopRGB565
|
|
|
|
ldmfd sp!, {r4-r9,lr}
|
|
bx lr
|
|
|
|
|
|
|
|
.global FinalizeLineRGB555 @ int sh
|
|
|
|
FinalizeLineRGB555:
|
|
stmfd sp!, {r4-r8,lr}
|
|
ldr r8, =(Pico+0x22228) @ Pico.video
|
|
ldr r4, =HighPal
|
|
|
|
ldrb r7, [r8, #-0x1a] @ 0x2220e ~ dirtyPal
|
|
mov r6, r0
|
|
mov r1, #0
|
|
tst r7, r7
|
|
beq .fl_noconvRGB555
|
|
strb r1, [r8, #-0x1a]
|
|
sub r1, r8, #0x128 @ r1=Pico.cram
|
|
mov r0, r4
|
|
mov r2, #0x40
|
|
bl vidConvCpyRGB565
|
|
|
|
.fl_noconvRGB555:
|
|
mov r3, r4
|
|
tst r6, r6
|
|
beq .fl_noshRGB555
|
|
tst r7, r7
|
|
beq .fl_noshRGB555
|
|
|
|
@ shadowed pixels:
|
|
mov r12, #0x008e
|
|
orr r12,r12,#0x7300
|
|
orr r12,r12,r12,lsl #16
|
|
add r4, r3, #0x40*2
|
|
add r5, r3, #0xc0*2
|
|
mov lr, #0x40/4
|
|
.fl_loopcpRGB555_sh:
|
|
ldmia r3!, {r1,r6}
|
|
subs lr, lr, #1
|
|
and r1, r12, r1, lsr #1
|
|
and r6, r12, r6, lsr #1
|
|
stmia r4!, {r1,r6}
|
|
stmia r5!, {r1,r6}
|
|
bne .fl_loopcpRGB555_sh
|
|
|
|
@ hilighted pixels:
|
|
sub r3, r3, #0x40*2
|
|
mov lr, #0x40/2
|
|
.fl_loopcpRGB555_hi:
|
|
ldr r1, [r3], #4
|
|
TileDoShHi2Pixels555 r1
|
|
str r1, [r4], #4
|
|
subs lr, lr, #1
|
|
bne .fl_loopcpRGB555_hi
|
|
|
|
sub r3, r3, #0x40*2
|
|
|
|
.fl_noshRGB555:
|
|
.if OVERRIDE_HIGHCOL
|
|
ldr r1, =HighCol
|
|
ldr r0, =DrawLineDest
|
|
ldr r1, [r1]
|
|
ldr r0, [r0]
|
|
add r1, r1, #8
|
|
.else
|
|
ldr r0, =DrawLineDest
|
|
ldr r1, =(HighCol+8)
|
|
ldr r0, [r0]
|
|
.endif
|
|
|
|
ldrb r12, [r8, #12]
|
|
mov lr, #0xff
|
|
mov lr, lr, lsl #1
|
|
|
|
tst r12, #1
|
|
movne r2, #320/8 @ len
|
|
bne .fl_no32colRGB555
|
|
ldr r4, =PicoOpt
|
|
mov r2, #256/8
|
|
ldr r4, [r4]
|
|
tst r4, #0x4000
|
|
bne .fl_32scale_RGB555
|
|
tst r4, #0x0100
|
|
addeq r0, r0, #32*2
|
|
|
|
.fl_no32colRGB555:
|
|
|
|
.if UNALIGNED_DRAWLINEDEST
|
|
@ this is basically for Gizmondo, which has unaligned odd lines in the framebuffer
|
|
tst r0, #2
|
|
bne .fl_RGB555u
|
|
.endif
|
|
|
|
.fl_loopRGB555:
|
|
ldr r12, [r1], #4
|
|
ldr r7, [r1], #4
|
|
|
|
and r4, lr, r12, lsl #1
|
|
ldrh r4, [r3, r4]
|
|
and r5, lr, r12, lsr #7
|
|
ldrh r5, [r3, r5]
|
|
and r6, lr, r12, lsr #15
|
|
ldrh r6, [r3, r6]
|
|
orr r4, r4, r5, lsl #16
|
|
|
|
and r5, lr, r12, lsr #23
|
|
ldrh r5, [r3, r5]
|
|
and r8, lr, r7, lsl #1
|
|
ldrh r8, [r3, r8]
|
|
orr r5, r6, r5, lsl #16
|
|
|
|
and r6, lr, r7, lsr #7
|
|
ldrh r6, [r3, r6]
|
|
and r12,lr, r7, lsr #15
|
|
ldrh r12,[r3, r12]
|
|
and r7, lr, r7, lsr #23
|
|
ldrh r7, [r3, r7]
|
|
orr r8, r8, r6, lsl #16
|
|
|
|
subs r2, r2, #1
|
|
orr r12,r12, r7, lsl #16
|
|
|
|
stmia r0!, {r4,r5,r8,r12}
|
|
bne .fl_loopRGB555
|
|
|
|
ldmfd sp!, {r4-r8,lr}
|
|
bx lr
|
|
|
|
|
|
.fl_32scale_RGB555:
|
|
stmfd sp!, {r9,r10}
|
|
mov r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007
|
|
orr r9, r9, #0x00e7
|
|
|
|
.if UNALIGNED_DRAWLINEDEST
|
|
tst r0, #2
|
|
bne .fl_32scale_RGB555u
|
|
.endif
|
|
|
|
.fl_loop32scale_RGB555:
|
|
ldr r12, [r1], #4
|
|
ldr r7, [r1], #4
|
|
|
|
and r4, lr, r12,lsl #1
|
|
ldrh r4, [r3, r4]
|
|
and r5, lr, r12,lsr #7
|
|
ldrh r5, [r3, r5]
|
|
and r4, r4, r9, lsl #2
|
|
orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0
|
|
and r5, r5, r9, lsl #2
|
|
sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1
|
|
add r4, r4, r6, lsl #16 @ pix_d 0, 1
|
|
and r6, lr, r12,lsr #15
|
|
ldrh r6, [r3, r6]
|
|
and r12,lr, r12,lsr #23
|
|
ldrh r12,[r3, r12]
|
|
and r6, r6, r9, lsl #2
|
|
add r5, r5, r6
|
|
mov r5, r5, lsr #1
|
|
sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2
|
|
orr r5, r5, r6, lsl #16
|
|
|
|
and r6, lr, r7, lsl #1
|
|
ldrh r6, [r3, r6]
|
|
and r12,r12,r9, lsl #2
|
|
add r5, r5, r12,lsl #14 @ pix_d 2, 3
|
|
and r6, r6, r9, lsl #2
|
|
orr r6, r12,r6, lsl #16 @ pix_d 4, 5
|
|
|
|
and r12,lr, r7, lsr #7
|
|
ldrh r12,[r3, r12]
|
|
and r10,lr, r7, lsr #15
|
|
ldrh r10,[r3, r10]
|
|
and r12,r12,r9, lsl #2
|
|
sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1
|
|
add r8, r8, r6, lsr #18
|
|
and r7, lr, r7, lsr #23
|
|
ldrh r7, [r3, r7]
|
|
and r10,r10,r9, lsl #2
|
|
orr r8, r8, r10,lsl #15
|
|
add r8, r8, r12,lsl #15 @ pix_d 6, 7
|
|
sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2
|
|
and r7, r7, r9, lsl #2
|
|
add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3
|
|
orr r10,r10,r7, lsl #16 @ pix_d 8, 9
|
|
|
|
subs r2, r2, #1
|
|
|
|
stmia r0!, {r4,r5,r6,r8,r10}
|
|
bne .fl_loop32scale_RGB555
|
|
|
|
ldmfd sp!, {r9,r10}
|
|
ldmfd sp!, {r4-r8,lr}
|
|
bx lr
|
|
|
|
.if UNALIGNED_DRAWLINEDEST
|
|
@ unaligned versions of loops
|
|
@ warning: starts drawing 2bytes before dst
|
|
|
|
.fl_RGB555u:
|
|
sub r0, r0, #2 @ initial adjustment
|
|
mov r8, #0
|
|
|
|
.fl_loopRGB555u:
|
|
ldr r12, [r1], #4
|
|
ldr r7, [r1], #4
|
|
|
|
and r6, lr, r12,lsl #1
|
|
ldrh r6, [r3, r6]
|
|
and r5, lr, r12,lsr #7
|
|
ldrh r5, [r3, r5]
|
|
orr r4, r8, r6, lsl #16
|
|
|
|
and r6, lr, r12,lsr #15
|
|
ldrh r6, [r3, r6]
|
|
and r8, lr, r12,lsr #23
|
|
ldrh r8, [r3, r8]
|
|
orr r5, r5, r6, lsl #16
|
|
|
|
and r6, lr, r7, lsl #1
|
|
ldrh r6, [r3, r6]
|
|
and r12,lr, r7, lsr #7
|
|
ldrh r12,[r3, r12]
|
|
orr r6, r8, r6, lsl #16
|
|
|
|
and r8, lr, r7, lsr #15
|
|
ldrh r8, [r3, r8]
|
|
and r7, lr, r7, lsr #23
|
|
|
|
subs r2, r2, #1
|
|
orr r12,r12,r8, lsl #16
|
|
ldrh r8, [r3, r7]
|
|
|
|
stmia r0!, {r4,r5,r6,r12}
|
|
bne .fl_loopRGB555u
|
|
|
|
strh r8, [r0], #2
|
|
|
|
ldmfd sp!, {r4-r8,lr}
|
|
bx lr
|
|
|
|
|
|
.fl_32scale_RGB555u:
|
|
sub r0, r0, #2 @ initial adjustment
|
|
mov r4, #0
|
|
|
|
@ r9 f800 07e0 001f | e000 0780 001c | 3800 01e0 0007
|
|
.fl_loop32scale_RGB555u:
|
|
ldr r12, [r1], #4
|
|
ldr r7, [r1], #4
|
|
|
|
and r6, lr, r12,lsl #1
|
|
ldrh r6, [r3, r6]
|
|
and r5, lr, r12,lsr #7
|
|
ldrh r5, [r3, r5]
|
|
and r6, r6, r9, lsl #2
|
|
orr r4, r4, r6, lsl #16 @ r4 = pix_d -1, 0
|
|
|
|
and r5, r5, r9, lsl #2
|
|
sub r8, r5, r5, lsr #2 @ r8 = 3/4 pix_s 1
|
|
add r6, r8, r6, lsr #2 @ r6 = (1/4 pix_s 0) + (3/4 pix_s 1)
|
|
orr r5, r6, r5, lsl #15
|
|
|
|
and r6, lr, r12,lsr #15
|
|
ldrh r6, [r3, r6]
|
|
and r12,lr, r12,lsr #23
|
|
ldrh r12,[r3, r12]
|
|
and r6, r6, r9, lsl #2
|
|
add r5, r5, r6, lsl #15 @ r5 = pix_d 1, 2
|
|
|
|
and r8, lr, r7, lsl #1
|
|
ldrh r8, [r3, r8]
|
|
and r10,lr, r7, lsr #7
|
|
ldrh r10,[r3, r10]
|
|
and r12,r12,r9, lsl #2
|
|
sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2
|
|
add r6, r6, r12,lsr #2
|
|
orr r6, r6, r12,lsl #16 @ r6 = pix_d 3, 4
|
|
|
|
and r8, r8, r9, lsl #2
|
|
and r10,r10,r9, lsl #2
|
|
sub r12,r10,r10,lsr #2 @ r12 = 3/4 pix_s 5
|
|
orr r8, r8, r8, lsl #14
|
|
add r8, r8, r12,lsl #16 @ r8 = pix_d 5, 6
|
|
and r12,lr, r7, lsr #15
|
|
ldrh r12,[r3, r12]
|
|
and r7, lr, r7, lsr #23
|
|
ldrh r7, [r3, r7]
|
|
and r12,r12,r9, lsl #2
|
|
add r10,r10,r12
|
|
mov r10,r10, lsr #1
|
|
sub r12,r12,r12,lsr #2 @ r12 = 3/4 pix_s 6
|
|
orr r10,r10,r12,lsl #16
|
|
and r7, r7, r9, lsl #2
|
|
add r10,r10,r7, lsl #14 @ r10 = pix_d 7, 8
|
|
|
|
subs r2, r2, #1
|
|
|
|
stmia r0!, {r4,r5,r6,r8,r10}
|
|
mov r4, r7
|
|
bne .fl_loop32scale_RGB555u
|
|
|
|
strh r4, [r0], #2
|
|
|
|
ldmfd sp!, {r9,r10}
|
|
ldmfd sp!, {r4-r8,lr}
|
|
bx lr
|
|
|
|
.endif @ UNALIGNED_DRAWLINEDEST
|
|
|
|
|
|
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
|
|
|
|
@ utility
|
|
.global blockcpy @ void *dst, void *src, size_t n
|
|
|
|
blockcpy:
|
|
stmfd sp!, {r4,r5}
|
|
mov r2, r2, lsr #4
|
|
blockcpy_loop:
|
|
ldmia r1!, {r3-r5,r12}
|
|
subs r2, r2, #1
|
|
stmia r0!, {r3-r5,r12}
|
|
bne blockcpy_loop
|
|
ldmfd sp!, {r4,r5}
|
|
bx lr
|
|
|
|
|
|
.global blockcpy_or @ void *dst, void *src, size_t n, int pat
|
|
|
|
blockcpy_or:
|
|
stmfd sp!, {r4-r6}
|
|
orr r3, r3, r3, lsl #8
|
|
orr r3, r3, r3, lsl #16
|
|
mov r2, r2, lsr #4
|
|
blockcpy_loop_or:
|
|
ldmia r1!, {r4-r6,r12}
|
|
subs r2, r2, #1
|
|
orr r4, r4, r3
|
|
orr r5, r5, r3
|
|
orr r6, r6, r3
|
|
orr r12,r12,r3
|
|
stmia r0!, {r4-r6,r12}
|
|
bne blockcpy_loop_or
|
|
ldmfd sp!, {r4-r6}
|
|
bx lr
|
|
|