@ vim:filetype=armasm\r
\r
-@ assembly "optimized" version of some funtions from draw.c\r
+@ ARM assembly versions of some funtions from draw.c\r
@ this is highly specialized, be careful if changing related C code!\r
\r
-@ (c) Copyright 2007, Grazvydas "notaz" Ignotas\r
+@ (c) Copyright 2007-2008, Grazvydas "notaz" Ignotas\r
@ All Rights Reserved\r
\r
.include "port_config.s"\r
.extern Pico\r
.extern PicoOpt\r
.extern HighCol\r
-.extern Scanline\r
+.extern DrawScanline\r
.extern HighSprZ\r
.extern rendstatus\r
+.extern HighPreSpr\r
.extern DrawLineDest\r
.extern DrawStripInterlace\r
+.extern HighCacheS_ptr\r
\r
+.equ PDRAW_SPRITES_MOVED, (1<<0)\r
+.equ PDRAW_WND_DIFF_PRIO, (1<<1)\r
+.equ PDRAW_ACC_SPRITES, (1<<2)\r
+.equ PDRAW_DIRTY_SPRITES, (1<<4)\r
+.equ PDRAW_PLANE_HI_PRIO, (1<<6)\r
+.equ PDRAW_SHHI_DONE, (1<<7)\r
+.equ MAX_LINE_SPRITES, 30\r
\r
@ helper\r
.macro TilePixel pat lsrr offs\r
.endif\r
ldreqb r4, [r1,#\offs]\r
orrne r4, r3, r4\r
- strneb r4, [r1,#\offs]\r
- tsteq r4, #0x80\r
andeq r4, r4, #0x3f\r
- streqb r4, [r1,#\offs]\r
+ strb r4, [r1,#\offs]\r
.endm\r
\r
@ TileNormShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
.else\r
ands r4, r12, r2\r
.endif\r
- beq 3f\r
+ beq 0f\r
cmp r4, #0xe\r
- beq 2f\r
- bgt 1f\r
- orr r4, r3, r4\r
- strb r4, [r1,#\ofs]\r
- b 3f\r
-1:\r
- ldrb r4, [r1,#\ofs] @ 2ci\r
- orr r4, r4, #0xc0\r
- strb r4, [r1,#\ofs]\r
- b 3f\r
-2:\r
- ldrb r4, [r1,#\ofs] @ 2ci\r
- bic r4, r4, #0xc0\r
- orr r4, r4, #0x80\r
+ ldrgeb r4, [r1,#\ofs]\r
+ orrlt r4, r3, r4 @ normal\r
+\r
+ biceq r4, r4, #0xc0 @ hilight\r
+ orreq r4, r4, #0x80\r
+ orrgt r4, r4, #0xc0 @ shadow\r
+\r
strb r4, [r1,#\ofs]\r
-3:\r
+0:\r
.endm\r
\r
@ TileFlipSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
TileDoShGenPixel 16, 7 @ #0x000f0000\r
.endm\r
\r
+.macro TileDoShGenPixel_noop shift ofs\r
+.if \shift\r
+ and r4, r12, r2, lsr #\shift\r
+.else\r
+ and r4, r12, r2\r
+.endif\r
+ sub r7, r4, #1\r
+ cmp r7, #0xd\r
+ orrcc r4, r3, r4 @ 0-0xc (was 1-0xd)\r
+ strccb r4, [r1,#\ofs]\r
+.endm\r
+\r
+.macro TileFlipSh_noop\r
+ TileDoShGenPixel_noop 16, 0 @ #0x000f0000\r
+ TileDoShGenPixel_noop 20, 1 @ #0x00f00000\r
+ TileDoShGenPixel_noop 24, 2 @ #0x0f000000\r
+ TileDoShGenPixel_noop 28, 3 @ #0xf0000000\r
+ TileDoShGenPixel_noop 0, 4 @ #0x0000000f\r
+ TileDoShGenPixel_noop 4, 5 @ #0x000000f0\r
+ TileDoShGenPixel_noop 8, 6 @ #0x00000f00\r
+ TileDoShGenPixel_noop 12, 7 @ #0x0000f000\r
+.endm\r
+\r
+.macro TileNormSh_noop\r
+ TileDoShGenPixel_noop 12, 0 @ #0x0000f000\r
+ TileDoShGenPixel_noop 8, 1 @ #0x00000f00\r
+ TileDoShGenPixel_noop 4, 2 @ #0x000000f0\r
+ TileDoShGenPixel_noop 0, 3 @ #0x0000000f\r
+ TileDoShGenPixel_noop 28, 4 @ #0xf0000000\r
+ TileDoShGenPixel_noop 24, 5 @ #0x0f000000\r
+ TileDoShGenPixel_noop 20, 6 @ #0x00f00000\r
+ TileDoShGenPixel_noop 16, 7 @ #0x000f0000\r
+.endm\r
+\r
+.macro TileDoShGenPixel_onlyop_lp shift ofs\r
+.if \shift\r
+ ands r7, r12, r2, lsr #\shift\r
+.else\r
+ ands r7, r12, r2\r
+.endif\r
+ ldrneb r4, [r1,#\ofs]\r
+ tstne r4, #0x40\r
+ beq 0f\r
+\r
+ cmp r7, #0xe\r
+ biceq r4, r4, #0xc0 @ hilight\r
+ orreq r4, r4, #0x80\r
+ orrgt r4, r4, #0xc0 @ shadow\r
+ strgeb r4, [r1,#\ofs]\r
+0:\r
+.endm\r
+\r
+.macro TileFlipSh_onlyop_lp\r
+ TileDoShGenPixel_onlyop_lp 16, 0 @ #0x000f0000\r
+ TileDoShGenPixel_onlyop_lp 20, 1 @ #0x00f00000\r
+ TileDoShGenPixel_onlyop_lp 24, 2 @ #0x0f000000\r
+ TileDoShGenPixel_onlyop_lp 28, 3 @ #0xf0000000\r
+ TileDoShGenPixel_onlyop_lp 0, 4 @ #0x0000000f\r
+ TileDoShGenPixel_onlyop_lp 4, 5 @ #0x000000f0\r
+ TileDoShGenPixel_onlyop_lp 8, 6 @ #0x00000f00\r
+ TileDoShGenPixel_onlyop_lp 12, 7 @ #0x0000f000\r
+.endm\r
+\r
+.macro TileNormSh_onlyop_lp\r
+ TileDoShGenPixel_onlyop_lp 12, 0 @ #0x0000f000\r
+ TileDoShGenPixel_onlyop_lp 8, 1 @ #0x00000f00\r
+ TileDoShGenPixel_onlyop_lp 4, 2 @ #0x000000f0\r
+ TileDoShGenPixel_onlyop_lp 0, 3 @ #0x0000000f\r
+ TileDoShGenPixel_onlyop_lp 28, 4 @ #0xf0000000\r
+ TileDoShGenPixel_onlyop_lp 24, 5 @ #0x0f000000\r
+ TileDoShGenPixel_onlyop_lp 20, 6 @ #0x00f00000\r
+ TileDoShGenPixel_onlyop_lp 16, 7 @ #0x000f0000\r
+.endm\r
+\r
\r
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
\r
ldreqb r12, [r11, #2]\r
ldrneb r12, [r11, #4]\r
\r
- ldr r2, =Scanline @ trying to make good use of pipeline here\r
+ ldr r2, =DrawScanline @ trying to make good use of pipeline here\r
ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
\r
moveq r12, r12, lsl #10\r
mov r4, r8, lsr #8 @ pvid->reg[13]\r
mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r
tst r7, #2\r
- addne r4, r4, r2, lsl #2 @ htab+=Scanline<<1; // Offset by line\r
+ addne r4, r4, r2, lsl #2 @ htab+=DrawScanline<<1; // Offset by line\r
tst r7, #1\r
biceq r4, r4, #0x1f @ htab&=~0xf; // Offset by tile\r
add r4, r4, r0, lsl #1 @ htab+=plane\r
mov r0, #0\r
ldreq r2, [r1]\r
str r0, [r6] @ terminate the cache list\r
- orreq r2, r2, #0x40 @ had a layer with all hi-prio tiles\r
+ orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
streq r2, [r1]\r
\r
ldmfd sp!, {r4-r11,lr}\r
.DrawStrip_vsscroll:\r
rsb r8, r3, #0\r
mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
- bic r8, r8, #0xff000000\r
- orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|tilex[23:0])\r
+ bic r8, r8, #0x3fc00000\r
+ orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|tilex[21:0])\r
\r
- ldr r4, =Scanline\r
+ ldr r4, =DrawScanline\r
orr r5, r1, r10, lsl #24\r
ldr r4, [r4]\r
sub r1, r3, #1\r
ldrh r7, [r7] @ r7=vscroll\r
\r
bic r10,r10,#0xff @ clear old ty\r
- and r4, r5, #0xff0000\r
- add r4, r4, r7, lsl #16\r
- and r4, r4, r5, lsl #16 @ r4=line<<16\r
+ and r4, r5, #0xff0000 @ scanline\r
+ add r4, r4, r7, lsl #16 @ ... += vscroll\r
+ and r4, r4, r5, lsl #16 @ ... &= ymask\r
and r7, r4, #0x70000\r
orr r10,r10,r7, lsr #15 @ new ty\r
\r
beq .DrawStrip_vs_samecode @ we know stuff about this tile already\r
\r
mov r9, r7 @ remember code\r
- orr r8, r8, #1<<24 @ seen non hi-prio tile\r
+ orr r8, r8, #(1<<24)@ seen non hi-prio tile\r
\r
movs r2, r9, lsl #20 @ if (code&0x1000)\r
mov r2, r2, lsl #1\r
b .dsloop_vs\r
\r
.dsloop_vs_exit:\r
- tst r8, #1<<24 @ seen non hi-prio tile\r
+ tst r8, #(1<<24) @ seen non hi-prio tile\r
ldreq r1, =rendstatus\r
mov r0, #0\r
ldreq r2, [r1]\r
str r0, [r6] @ terminate the cache list\r
- orreq r2, r2, #0x40 @ had a layer with all hi-prio tiles\r
+ orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
streq r2, [r1]\r
\r
ldmfd sp!, {r4-r11,lr}\r
movne r7, r7, lsl #5\r
\r
@ Find the line in the name table\r
- add r2, r7, r2, lsl #22 @ r2=(vscroll+(Scanline<<1))<<21 (11 bits);\r
+ add r2, r7, r2, lsl #22 @ r2=(vscroll+(DrawScanline<<1))<<21 (11 bits);\r
orr r1, r1, #0x80000000\r
and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
mov r2, r2, lsr #21\r
b .dtfc_loop\r
\r
.dtfc_shadow_blank:\r
- ldrb r4, [r1] @ 1ci\r
- ldrb r12,[r1,#1]\r
- tst r4, #0x80\r
- andeq r4, r4,#0x3f\r
- streqb r4, [r1]\r
- tst r12,#0x80\r
- ldrb r4, [r1,#2]\r
- andeq r12,r12,#0x3f\r
- streqb r12,[r1,#1]\r
- tst r4, #0x80\r
- ldrb r12,[r1,#3]\r
- andeq r4, r4,#0x3f\r
- streqb r4, [r1,#2]\r
- tst r12,#0x80\r
- ldrb r4, [r1,#4]\r
- andeq r12,r12,#0x3f\r
- streqb r12,[r1,#3]\r
- tst r4, #0x80\r
- ldrb r12,[r1,#5]\r
- andeq r4, r4,#0x3f\r
- streqb r4, [r1,#4]\r
- tst r12,#0x80\r
- ldrb r4, [r1,#6]\r
- andeq r12,r12,#0x3f\r
- streqb r12,[r1,#5]\r
- tst r4, #0x80\r
- ldrb r12,[r1,#7]\r
- andeq r4, r4,#0x3f\r
- streqb r4, [r1,#6]\r
- tst r12,#0x80\r
- andeq r12,r12,#0x3f\r
- streqb r12,[r1,#7]\r
- mov r12, #0xf\r
+ tst r1, #1\r
+ ldrneb r4, [r1]\r
+ mov r6, #0x3f\r
+ and r4, r4, #0x3f\r
+ strneb r4, [r1], #1\r
+ ldrh r4, [r1]\r
+ orr r6, r6, r6, lsl #8\r
+ and r4, r4, r6\r
+ strh r4, [r1], #2\r
+ ldrh r4, [r1]\r
+ and r4, r4, r6\r
+ strh r4, [r1], #2\r
+ ldrh r4, [r1]\r
+ and r4, r4, r6\r
+ strh r4, [r1], #2\r
+ ldrh r4, [r1]\r
+ and r4, r4, r6\r
+ streqh r4, [r1]\r
+ strneb r4, [r1]\r
b .dtfc_loop\r
\r
.dtfc_cut_tile:\r
.dtfc_check_rendflags:\r
ldr r1, =rendstatus\r
ldr r2, [r1]\r
- tst r2, #0xc0\r
+ tst r2, #(PDRAW_PLANE_HI_PRIO|PDRAW_SHHI_DONE)\r
beq .dtfc_loop\r
bic r8, r8, #1 @ sh/hi mode off\r
- tst r2, #0x80\r
+ tst r2, #PDRAW_SHHI_DONE\r
bne .dtfc_loop @ already processed\r
- orr r2, r2, #0x80\r
+ orr r2, r2, #PDRAW_SHHI_DONE\r
str r2, [r1]\r
\r
add r1, r11,#8\r
- mov r3, #320/4\r
- mov r7, #0x80\r
- orr r7, r7, r7, lsl #8\r
- orr r7, r7, r7, lsl #16\r
+ mov r3, #320/4/4\r
mov r6, #0x3f\r
orr r6, r6, r6, lsl #8\r
orr r6, r6, r6, lsl #16\r
.dtfc_loop_shprep:\r
+ ldmia r1, {r2,r4,r5,r7}\r
subs r3, r3, #1\r
- bmi .dtfc_loop @ done\r
- ldr r2, [r1]\r
- tst r2, r7\r
- andeq r2, r2, r6\r
- streq r2, [r1], #4\r
- beq .dtfc_loop_shprep\r
- tst r2, #0x80000000\r
- biceq r2, r2, #0xc0000000\r
- tst r2, #0x00800000\r
- biceq r2, r2, #0x00c00000\r
- tst r2, #0x00008000\r
- biceq r2, r2, #0x0000c000\r
- tst r2, #0x00000080\r
- biceq r2, r2, #0x000000c0\r
- str r2, [r1], #4\r
- b .dtfc_loop_shprep\r
+ and r2, r2, r6\r
+ and r4, r4, r6\r
+ and r5, r5, r6\r
+ and r7, r7, r6\r
+ stmia r1!,{r2,r4,r5,r7}\r
+ bne .dtfc_loop_shprep\r
+\r
+ mvn r5, #0 @ r5=prevcode=-1\r
+ b .dtfc_loop\r
\r
.pool\r
\r
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
\r
\r
-.global DrawSpritesFromCache @ int *hc, int sh\r
+.global DrawSpritesSHi @ int prio_unused, int sh\r
+\r
+DrawSpritesSHi:\r
+ ldr r3, =DrawScanline\r
+ ldr r2, =HighLnSpr\r
+ ldr r12,[r3]\r
+ mov r3, #(MAX_LINE_SPRITES+2)\r
+ mla r2, r12, r3, r2\r
+ ldr r3, [r2]\r
+ ands r3, r3, #0x7f\r
+ bxeq lr\r
\r
-DrawSpritesFromCache:\r
stmfd sp!, {r4-r11,lr}\r
+ mov r12,#0xff\r
+ strb r12,[r2,#1] @ set end marker\r
+ add r10,r2, #2\r
+ add r10,r10,r3 @ r10=HighLnSpr end\r
+\r
+ str r1, [sp, #-4] @ no calls after this point\r
\r
- @ cache some stuff to avoid mem access\r
.if OVERRIDE_HIGHCOL\r
ldr r11,=HighCol\r
mov r12,#0xf\r
mov r12,#0xf\r
.endif\r
ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
- mov r6, r1, lsl #31\r
- orr r6, r6, #1<<30\r
\r
- mov r10, r0\r
\r
-.dsfc_loop:\r
- ldr r9, [r10], #4 @ read code\r
- bic r6, r6, #7 @ using pipeline\r
- tst r9, r9\r
- ldmeqfd sp!, {r4-r11,pc}\r
+DrawSpriteSHi:\r
+ @ draw next sprite\r
+ ldrb r0, [r10,#-1]!\r
+ ldr r1, =HighPreSpr\r
+ ldr r8, [sp, #-4]\r
+ cmp r0, #0xff\r
+ ldmeqfd sp!, {r4-r11,pc} @ end of list\r
+ and r0, r0, #0x7f\r
+ add r0, r1, r0, lsl #3\r
\r
- mov r4, r9, lsl #28\r
- orr r6, r6, r4, lsr #30\r
- add r6, r6, #1 @ r6=s1cc???? ... ?????www (s=shadow/hilight, cc=pal, w=width)\r
+ ldr r9, [r0, #4] @ sprite[1]\r
+ mov r2, r9, asr #16 @ r2=sx\r
\r
- and r5, r9, #3\r
- add r5, r5, #1 @ r5=delta\r
- tst r9, #0x10000\r
- rsbne r5, r5, #0 @ Flip X\r
- mov r5, r5, lsl #4\r
+ mov r9, r9, lsl #16\r
+ mov r3, r9, lsr #31 @ priority\r
+ mov r9, r9, lsr #16\r
+ orr r9, r9, r8, lsl #31 @ r9=code|sh[31]\r
+ and r4, r9, #0x6000\r
+ orr r9, r9, r4, lsl #16\r
+ orr r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
+ cmp r12,r9, lsr #28 @ sh/hi with pal3?\r
+ cmpne r3, #1 @ if not, is hi prio\r
+ bne DrawSpriteSHi @ non-operator low sprite, already drawn\r
\r
- mov r2, r9, lsr #17\r
- mov r8, r2, lsl #1 @ tile=((unsigned int)code>>17)<<1;\r
+ ldr r3, [r0] @ sprite[0]\r
+ ldr r7, =DrawScanline\r
+ mov r6, r3, lsr #28\r
+ sub r6, r6, #1 @ r6=width-1 (inc later)\r
+ mov r5, r3, lsr #24\r
+ and r5, r5, #7 @ r5=height\r
\r
- and r3, r9, #0x30 @ r3=pal=(code&0x30);\r
+ mov r0, r3, lsl #16 @ r4=sy<<16 (tmp)\r
\r
- bic r6, r6, #3<<28\r
- orr r6, r6, r3, lsl #24\r
+ ldr r7, [r7]\r
+ sub r7, r7, r0, asr #16 @ r7=row=DrawScanline-sy\r
\r
- mov r0, r9, lsl #16\r
- mov r0, r0, asr #22 @ sx=(code<<16)>>22\r
- adds r0, r0, #0 @ set ZV\r
- b .dsfc_inloop_enter\r
+ tst r9, #0x1000\r
+ movne r0, r5, lsl #3\r
+ subne r0, r0, #1\r
+ subne r7, r0, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
\r
-@ scratch: r4, r7\r
-.dsfc_inloop:\r
- sub r6, r6, #1\r
- tst r6, #7\r
- beq .dsfc_loop\r
- adds r0, r0, #8\r
- add r8, r8, r5\r
-\r
-.dsfc_inloop_enter:\r
- ble .dsfc_inloop\r
+ add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
+ tst r9, #0x0800\r
+ mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
+ rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
+\r
+ mov r8, r8, lsl #21\r
+ mov r8, r8, lsr #17\r
+ and r7, r7, #7\r
+ add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
+\r
+ mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
+\r
+ orrs r3, r9, r9, lsl #4\r
+ mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
+\r
+ add r6, r6, #1 @ inc now\r
+ adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
+ b .dsprShi_loop_enter\r
+\r
+.dsprShi_loop:\r
+ subs r6, r6, #1 @ width--\r
+ beq DrawSpriteSHi\r
+ adds r0, r0, #8 @ sx+=8\r
+ add r8, r8, r5 @ tile+=delta\r
+\r
+.dsprShi_loop_enter:\r
+ ble .dsprShi_loop @ sx <= 0\r
cmp r0, #328\r
- bge .dsfc_loop\r
+ bge DrawSpriteSHi\r
\r
mov r8, r8, lsl #17\r
- mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
+ mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
\r
- ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+tile); // Get 8 pixels\r
- add r1, r11, r0 @ r1=pdest\r
+ ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
+ add r1, r11, r0 @ r1=pdest\r
tst r2, r2\r
- beq .dsfc_inloop\r
+ beq .dsprShi_loop\r
\r
- cmp r12, r6, lsr #28\r
- beq .dsfc_shadow\r
+ cmp r12, r9, lsr #28\r
+ beq .dsprShi_shadow\r
\r
cmp r2, r2, ror #4\r
- beq .dsfc_SingleColor @ tileline singlecolor \r
+ beq .dsprShi_SingleColor @ tileline singlecolor \r
\r
- tst r9, #0x10000\r
- bne .dsfc_TileFlip\r
+ tst r9, #0x0800\r
+ bne .dsprShi_TileFlip\r
\r
- @ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
-.dsfc_TileNorm:\r
+ @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
+@ scratch: r4, r7\r
+.dsprShi_TileNorm:\r
TileNorm r12\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
\r
-.dsfc_TileFlip:\r
+.dsprShi_TileFlip:\r
TileFlip r12\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
\r
-.dsfc_SingleColor:\r
- tst r0, #1 @ not aligned?\r
+.dsprShi_SingleColor:\r
and r4, r2, #0xf\r
orr r4, r3, r4\r
orr r4, r4, r4, lsl #8\r
+ tst r0, #1 @ not aligned?\r
strneb r4, [r1], #1\r
streqh r4, [r1], #2\r
strh r4, [r1], #2\r
strh r4, [r1], #2\r
strh r4, [r1], #2\r
strneb r4, [r1], #1\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
+\r
+.dsprShi_shadow:\r
+ tst r9, #0x8000\r
+ beq .dsprShi_shadow_lowpri\r
\r
-.dsfc_shadow:\r
cmp r2, r2, ror #4\r
- beq .dsfc_singlec_sh\r
+ beq .dsprShi_singlec_sh\r
\r
- tst r9, #0x10000\r
- bne .dsfc_TileFlip_sh\r
+ tst r9, #0x0800\r
+ bne .dsprShi_TileFlip_sh\r
\r
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
-.dsfc_TileNorm_sh:\r
+.dsprShi_TileNorm_sh:\r
TileNormSh\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
\r
-.dsfc_TileFlip_sh:\r
+.dsprShi_TileFlip_sh:\r
TileFlipSh\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
\r
-.dsfc_singlec_sh:\r
+.dsprShi_singlec_sh:\r
cmp r2, #0xe0000000\r
- bcc .dsfc_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
+ bcc .dsprShi_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
tst r2, #0x10000000\r
- bne .dsfc_sh_sh\r
+ bne .dsprShi_sh_sh\r
TileSingleHi\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
\r
-.dsfc_sh_sh:\r
+.dsprShi_sh_sh:\r
TileSingleSh\r
- b .dsfc_inloop\r
+ b .dsprShi_loop\r
+\r
+.dsprShi_shadow_lowpri:\r
+ tst r9, #0x800\r
+ bne .dsprShi_TileFlip_sh_lp\r
+\r
+.dsprShi_TileNorm_sh_lp:\r
+ TileNormSh_onlyop_lp\r
+ b .dsprShi_loop\r
+\r
+.dsprShi_TileFlip_sh_lp:\r
+ TileFlipSh_onlyop_lp\r
+ b .dsprShi_loop\r
\r
.pool\r
\r
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
\r
-@ + 0 : hhhhvvvv ab--hhvv yyyyyyyy yyyyyyyy // a: offscreen h, b: offs. v, h: horiz. size\r
-@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
+.global DrawAllSprites @ int prio, int sh\r
+\r
+DrawAllSprites:\r
+ ldr r3, =rendstatus\r
+ orr r1, r1, r0, lsl #1\r
+ ldr r12,[r3]\r
+ tst r12,#(PDRAW_ACC_SPRITES|PDRAW_SPRITES_MOVED)\r
+ beq das_no_prep\r
+ stmfd sp!, {r1,lr}\r
+ and r0, r12,#PDRAW_DIRTY_SPRITES\r
+ bic r12,r12,#(PDRAW_ACC_SPRITES|PDRAW_SPRITES_MOVED)\r
+ str r12,[r3]\r
+ bl PrepareSprites\r
+ ldmfd sp!, {r1,lr}\r
+\r
+das_no_prep:\r
+ ldr r3, =DrawScanline\r
+ ldr r2, =HighLnSpr\r
+ ldr r12,[r3]\r
+ mov r3, #(MAX_LINE_SPRITES+2)\r
+ mla r2, r12, r3, r2\r
+ ldr r3, [r2]\r
+ ands r3, r3, #0x7f\r
+ bxeq lr\r
+\r
+ @ time to do some real work\r
+ stmfd sp!, {r4-r11,lr}\r
+ mov r12,#0xff\r
+ strb r12,[r2,#1] @ set end marker\r
+ add r10,r2, #2\r
+ add r10,r10,r3 @ r10=HighLnSpr end\r
\r
-.global DrawSprite @ unsigned int *sprite, int **hc, int sh\r
+ str r1, [sp, #-4] @ no calls after this point\r
\r
-DrawSprite:\r
- stmfd sp!, {r4-r9,r11,lr}\r
+.if OVERRIDE_HIGHCOL\r
+ ldr r11,=HighCol\r
+ mov r12,#0xf\r
+ ldr r11,[r11]\r
+.else\r
+ ldr r11,=HighCol\r
+ mov r12,#0xf\r
+.endif\r
+ ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
+\r
+@ + 0 : hhhhvvvv ----hhvv yyyyyyyy yyyyyyyy // v, h: horiz. size\r
+@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
+\r
+DrawSprite: @ was: unsigned int *sprite, int sh, int acc_sprites\r
+ @ draw next sprite\r
+ ldrb r0, [r10,#-1]!\r
+ ldr r1, =HighPreSpr\r
+ ldr r8, [sp, #-4]\r
+ mov r2, r0, lsr #7\r
+ cmp r0, #0xff\r
+ ldmeqfd sp!, {r4-r11,pc} @ end of list\r
+ cmp r2, r8, lsr #1\r
+ bne DrawSprite @ wrong priority\r
+ and r0, r0, #0x7f\r
+ add r0, r1, r0, lsl #3\r
+\r
+@ stmfd sp!, {r4-r9,r11,lr}\r
+@ orr r8, r2, r1, lsl #4\r
\r
ldr r3, [r0] @ sprite[0]\r
- ldr r7, =Scanline\r
+ ldr r7, =DrawScanline\r
mov r6, r3, lsr #28\r
sub r6, r6, #1 @ r6=width-1 (inc later)\r
mov r5, r3, lsr #24\r
\r
ldr r7, [r7]\r
ldr r9, [r0, #4]\r
- sub r7, r7, r4, asr #16 @ r7=row=Scanline-sy\r
+ sub r7, r7, r4, asr #16 @ r7=row=DrawScanline-sy\r
\r
- tst r2, r2\r
mov r2, r9, asr #16 @ r2=sx\r
- bic r9, r9, #0xfe000000\r
- orrne r9, r9, #1<<31 @ r9=code|(sh<<31)\r
+ mov r9, r9, lsl #16\r
+ mov r9, r9, lsr #16\r
+ orr r9, r9, r8, lsl #31 @ r9=code|sh[31]\r
\r
tst r9, #0x1000\r
movne r4, r5, lsl #3\r
subne r4, r4, #1\r
subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
\r
- mov r8, r9, lsl #21\r
- mov r8, r8, lsr #21\r
- add r8, r8, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
- \r
+ add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
tst r9, #0x0800\r
mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
\r
- mov r8, r8, lsl #4\r
+ mov r8, r8, lsl #21\r
+ mov r8, r8, lsr #17\r
and r7, r7, #7\r
add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
\r
- tst r9, #0x8000\r
- bne .dspr_cache @ if(code&0x8000) // high priority - cache it\r
-\r
+.dspr_continue:\r
@ cache some stuff to avoid mem access\r
-.if OVERRIDE_HIGHCOL\r
- ldr r11,=HighCol\r
- mov r12,#0xf\r
- ldr r11,[r11]\r
-.else\r
- ldr r11,=HighCol\r
- mov r12,#0xf\r
-.endif\r
- ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
-\r
mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
and r4, r9, #0x6000\r
orr r9, r9, r4, lsl #16\r
- orr r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
+ orrs r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
\r
- tst r9, #1<<31\r
mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
- orrne r3, r3, #0x40 @ shadow by default\r
+ orrmi r3, r3, #0x40 @ for sh/hi\r
\r
add r6, r6, #1 @ inc now\r
adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
\r
.dspr_loop:\r
subs r6, r6, #1 @ width--\r
- ldmeqfd sp!, {r4-r9,r11,pc}@ return\r
+ beq DrawSprite\r
adds r0, r0, #8 @ sx+=8\r
add r8, r8, r5 @ tile+=delta\r
\r
.dspr_loop_enter:\r
ble .dspr_loop @ sx <= 0\r
cmp r0, #328\r
- ldmgefd sp!, {r4-r9,r11,pc}@ return\r
+ bge DrawSprite\r
\r
mov r8, r8, lsl #17\r
mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
TileFlip r12\r
b .dspr_loop\r
\r
+.dspr_singlec_sh:\r
+ cmp r2, #0xe0000000\r
+ bcs .dspr_loop @ operator tileline, ignore\r
+\r
.dspr_SingleColor:\r
and r4, r2, #0xf\r
orr r4, r3, r4\r
\r
@ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
.dspr_TileNorm_sh:\r
- TileNormSh\r
+ TileNormSh_noop\r
b .dspr_loop\r
\r
.dspr_TileFlip_sh:\r
- TileFlipSh\r
- b .dspr_loop\r
-\r
-.dspr_singlec_sh:\r
- cmp r2, #0xe0000000\r
- bcc .dspr_SingleColor @ normal tileline\r
- tst r2, #0x10000000\r
- bne .dspr_sh_sh\r
- TileSingleHi\r
- b .dspr_loop\r
-\r
-.dspr_sh_sh:\r
- TileSingleSh\r
+ TileFlipSh_noop\r
b .dspr_loop\r
\r
\r
-.dspr_cache:\r
- @ *(*hc)++ = (tile<<16)|((code&0x0800)<<5)|((sx<<6)&0x0000ffc0)|((code>>9)&0x30)|((sprite[0]>>24)&0xf);\r
- mov r4, r8, lsl #16 @ tile\r
- tst r9, #0x0800\r
- orrne r4, r4, #0x10000 @ code&0x0800\r
- mov r2, r2, lsl #22\r
- orr r4, r4, r2, lsr #16 @ (sx<<6)&0x0000ffc0\r
- and r2, r9, #0x6000\r
- orr r4, r4, r2, lsr #9 @ (code>>9)&0x30\r
- mov r3, r3, lsl #12\r
- ldr r2, [r1]\r
- orr r4, r4, r3, lsr #28 @ (sprite[0]>>24)&0xf\r
-\r
- str r4, [r2], #4\r
- str r2, [r1]\r
-\r
- ldmfd sp!, {r4-r9,r11,lr}\r
- bx lr\r
-\r
@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
\r
.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache\r
stmfd sp!, {r4-r11,lr}\r
\r
ldr r11, =(Pico+0x22228) @ Pico.video\r
- ldr r10, =Scanline\r
+ ldr r10, =DrawScanline\r
ldrb r12, [r11, #3] @ pvid->reg[3]\r
\r
ldr r10, [r10]\r
\r
ldr r6, =rendstatus\r
ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
- ldrb r6, [r6]\r
+ ldr r6, [r6]\r
\r
@ fetch the first code now\r
ldrh r7, [lr, r12]\r
\r
- ands r6, r6, #2 @ we care about bit 1 only\r
+ ands r6, r6, #PDRAW_WND_DIFF_PRIO\r
orr r6, r6, r2\r
- bne .dw_no_sameprio\r
\r
- cmp r2, r7, lsr #15\r
- ldmnefd sp!, {r4-r11,pc} @ assume that whole window uses same priority\r
+ eoreq r8, r2, r7, lsr #15 @ do prio bits differ?\r
+ cmpeq r8, #1\r
+ ldmeqfd sp!, {r4-r11,pc} @ yes, assume that whole window uses same priority\r
\r
-.dw_no_sameprio:\r
orr r6, r6, r3, lsl #8 @ shadow mode\r
\r
sub r8, r1, r0\r
mov r8, r8, lsl #1 @ cells\r
mvn r9, #0 @ r9=prevcode=-1\r
.endif\r
- add r1, r11, r0, lsl #4 @ r1=pdest\r
+ add r1, r11, r0, lsl #4 @ r1=pdest\r
mov r0, #0xf\r
b .dwloop_enter\r
\r
- @ r4,r5 & r7 are scratch in this loop\r
+ @ r4,r5 are scratch in this loop\r
.dwloop:\r
add r1, r1, #8\r
.dwloop_nor1:\r
orreq r3, r3, #0x40\r
beq .dw_shadow_done\r
ldr r4, [r1]\r
- tst r4, #0x00000080\r
- biceq r4, r4, #0x000000c0\r
- tst r4, #0x00008000\r
- biceq r4, r4, #0x0000c000\r
- tst r4, #0x00800000\r
- biceq r4, r4, #0x00c00000\r
- tst r4, #0x80000000\r
- biceq r4, r4, #0xc0000000\r
+ mov r5, #0x3f\r
+ orr r5, r5, r5, lsl #8\r
+ orr r5, r5, r5, lsl #16\r
+ and r4, r4, r5\r
str r4, [r1]\r
ldr r4, [r1,#4]\r
- tst r4, #0x00000080\r
- biceq r4, r4, #0x000000c0\r
- tst r4, #0x00008000\r
- biceq r4, r4, #0x0000c000\r
- tst r4, #0x00800000\r
- biceq r4, r4, #0x00c00000\r
- tst r4, #0x80000000\r
- biceq r4, r4, #0xc0000000\r
+ and r4, r4, r5\r
str r4, [r1,#4]\r
b .dw_shadow_done\r
\r
bne .fl_loopcpBGR444_hi\r
\r
sub r3, r4, #0x40*3*2\r
+ mov r6, #1\r
\r
\r
.fl_noshBGR444:\r
+ ldr r12,=rendstatus\r
+ eors r6, r6, #1 @ sh is 0\r
+ ldr r12,[r12]\r
+ mov lr, #0xff\r
+ tstne r12,#PDRAW_ACC_SPRITES\r
+\r
.if OVERRIDE_HIGHCOL\r
ldr r1, =HighCol\r
- mov lr, #0xff\r
+ movne lr, #0x3f\r
ldr r1, [r1]\r
mov lr, lr, lsl #1\r
add r1, r1, #8\r
.else\r
ldr r1, =(HighCol+8)\r
- mov lr, #0xff\r
+ movne lr, #0x3f\r
mov lr, lr, lsl #1\r
.endif\r
\r
.fl_loopBGR444:\r
-\r
ldr r12, [r1], #4\r
subs r2, r2, #1\r
\r
ldrh r5, [r3, r5]\r
and r6, lr, r12, lsr #15\r
ldrh r6, [r3, r6]\r
+ and r12,lr, r12, lsr #23\r
+ ldrh r12,[r3, r12] @ 1c.i.\r
orr r4, r4, r5, lsl #16\r
-\r
- and r5, lr, r12, lsr #23\r
- ldrh r5, [r3, r5] @ 2c.i.\r
- orr r5, r6, r5, lsl #16\r
+ orr r5, r6, r12,lsl #16\r
\r
stmia r0!, {r4,r5}\r
bne .fl_loopBGR444\r
bne .fl_loopcpRGB555_hi\r
\r
sub r3, r3, #0x40*2\r
+ mov r6, #1\r
\r
.fl_noshRGB555:\r
+ ldr r12,=rendstatus\r
+ eors r6, r6, #1 @ sh is 0\r
+ ldr r12,[r12]\r
+ mov lr, #0xff\r
+ tstne r12,#PDRAW_ACC_SPRITES\r
+ movne lr, #0x3f\r
+\r
.if OVERRIDE_HIGHCOL\r
ldr r1, =HighCol\r
ldr r0, =DrawLineDest\r
.endif\r
\r
ldrb r12, [r8, #12]\r
- mov lr, #0xff\r
mov lr, lr, lsl #1\r
\r
tst r12, #1\r
notaz.gp2x.de / picodrive.git / blobdiff