notaz.gp2x.de / picodrive.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
better mmuhack usage
[picodrive.git] / Pico / Draw.s
CommitLineData
3d48f143 1@ vim:filetype=armasm\r
2\r
cc68a136 3@ assembly "optimized" version of some funtions from draw.c\r
4@ this is highly specialized, be careful if changing related C code!\r
5\r
6cadc2da 6@ (c) Copyright 2007, Grazvydas "notaz" Ignotas\r
cc68a136 7@ All Rights Reserved\r
8\r
9\r
10.extern Pico\r
11.extern PicoOpt\r
12.extern HighCol\r
13.extern Scanline\r
14.extern HighSprZ\r
15.extern rendstatus\r
16.extern DrawLineDest\r
17.extern DrawStripVSRam\r
18.extern DrawStripInterlace\r
19\r
20\r
21@ helper\r
22.macro TilePixel pat lsrr offs\r
23.if !\lsrr\r
24 ands r4, \pat, r2\r
25.else\r
26 ands r4, \pat, r2, lsr #\lsrr\r
27.endif\r
28 orrne r4, r3, r4\r
29 strneb r4, [r1,#\offs]\r
30.endm\r
31\r
32@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
33.macro TileNorm pat\r
34 TilePixel \pat, 12, 0 @ #0x0000f000\r
35 TilePixel \pat, 8, 1 @ #0x00000f00\r
36 TilePixel \pat, 4, 2 @ #0x000000f0\r
37 TilePixel \pat, 0, 3 @ #0x0000000f\r
38 TilePixel \pat, 28, 4 @ #0xf0000000\r
39 TilePixel \pat, 24, 5 @ #0x0f000000\r
40 TilePixel \pat, 20, 6 @ #0x00f00000\r
41 TilePixel \pat, 16, 7 @ #0x000f0000\r
42.endm\r
43\r
44@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
45.macro TileFlip pat\r
46 TilePixel \pat, 16, 0 @ #0x000f0000\r
47 TilePixel \pat, 20, 1 @ #0x00f00000\r
48 TilePixel \pat, 24, 2 @ #0x0f000000\r
49 TilePixel \pat, 28, 3 @ #0xf0000000\r
50 TilePixel \pat, 0, 4 @ #0x0000000f\r
51 TilePixel \pat, 4, 5 @ #0x000000f0\r
52 TilePixel \pat, 8, 6 @ #0x00000f00\r
53 TilePixel \pat, 12, 7 @ #0x0000f000\r
54.endm\r
55\r
56@ shadow/hilight mode\r
57\r
58@ this one is for hi priority layer\r
59.macro TilePixelShHP lsrr offs\r
60.if !\lsrr\r
61 ands r4, r12, r2\r
62.else\r
63 ands r4, r12, r2, lsr #\lsrr\r
64.endif\r
65 ldreqb r4, [r1,#\offs]\r
66 orrne r4, r3, r4\r
67 strneb r4, [r1,#\offs]\r
68 tsteq r4, #0x80\r
69 andeq r4, r4, #0x3f\r
70 streqb r4, [r1,#\offs]\r
71.endm\r
72\r
73@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
74.macro TileNormShHP\r
75 TilePixelShHP 12, 0 @ #0x0000f000\r
76 TilePixelShHP 8, 1 @ #0x00000f00\r
77 TilePixelShHP 4, 2 @ #0x000000f0\r
78 TilePixelShHP 0, 3 @ #0x0000000f\r
79 TilePixelShHP 28, 4 @ #0xf0000000\r
80 TilePixelShHP 24, 5 @ #0x0f000000\r
81 TilePixelShHP 20, 6 @ #0x00f00000\r
82 TilePixelShHP 16, 7 @ #0x000f0000\r
83.endm\r
84\r
85@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
86.macro TileFlipShHP\r
87 TilePixelShHP 16, 0 @ #0x000f0000\r
88 TilePixelShHP 20, 1 @ #0x00f00000\r
89 TilePixelShHP 24, 2 @ #0x0f000000\r
90 TilePixelShHP 28, 3 @ #0xf0000000\r
91 TilePixelShHP 0, 4 @ #0x0000000f\r
92 TilePixelShHP 4, 5 @ #0x000000f0\r
93 TilePixelShHP 8, 6 @ #0x00000f00\r
94 TilePixelShHP 12, 7 @ #0x0000f000\r
95.endm\r
96\r
97\r
98@ TileSingleSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx; r12: helper pattern 0xf\r
99.macro TileSingleSh\r
100 tst r0, #1 @ not aligned?\r
101 mov r7, #0x00c000\r
102 orr r7, r7, #0xc0\r
103 ldrneb r4, [r1]\r
104 ldreqh r4, [r1]\r
105 orr r4, r4, r7\r
106 strneb r4, [r1], #1\r
107 streqh r4, [r1], #2\r
108 ldrh r4, [r1]\r
109 orr r4, r4, r7\r
110 strh r4, [r1], #2\r
111 ldrh r4, [r1]\r
112 orr r4, r4, r7\r
113 strh r4, [r1], #2\r
114 ldrh r4, [r1]\r
115 orr r4, r4, r7\r
116 strh r4, [r1], #2\r
117 ldrneb r4, [r1]\r
118 orr r4, r4, r7\r
119 strneb r4, [r1], #1\r
120.endm\r
121\r
122@ TileSingleHi (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
123.macro TileSingleHi\r
124 tst r1, #1 @ not aligned?\r
125 mov r7, #0x008000\r
126 orr r7, r7, #0x80\r
127 ldrneb r4, [r1], #1\r
128 ldreqh r4, [r1], #2 @ 1ci\r
129 ldrh r12, [r1], #2\r
130 bic r4, r4, r7, lsr #1\r
131 orr r4, r4, r7\r
132 strneb r4, [r1, #-3]\r
133 streqh r4, [r1, #-4]\r
134 ldrh r4, [r1], #2\r
135 bic r12, r12, r7, lsr #1\r
136 orr r12, r12, r7\r
137 strh r12, [r1, #-4]\r
138 ldrh r12, [r1], #2\r
139 bic r4, r4, r7, lsr #1\r
140 orr r4, r4, r7\r
141 strh r4, [r1, #-4]\r
142 ldrneb r4, [r1]\r
143 bic r12, r12, r7, lsr #1\r
144 orr r12, r12, r7\r
145 strh r12, [r1, #-2]\r
146 bicne r4, r4, r7, lsr #1\r
147 orrne r4, r4, r7\r
148 strneb r4, [r1], #1\r
149 mov r12, #0xf\r
150.endm\r
151\r
152.macro TileDoShGenPixel shift ofs\r
153.if \shift\r
154 ands r4, r12, r2, lsr #\shift\r
155.else\r
156 ands r4, r12, r2\r
157.endif\r
158 beq 3f\r
159 cmp r4, #0xe\r
160 beq 2f\r
161 bgt 1f\r
162 orr r4, r3, r4\r
163 strb r4, [r1,#\ofs]\r
164 b 3f\r
1651:\r
166 ldrb r4, [r1,#\ofs] @ 2ci\r
167 orr r4, r4, #0xc0\r
168 strb r4, [r1,#\ofs]\r
169 b 3f\r
1702:\r
171 ldrb r4, [r1,#\ofs] @ 2ci\r
172 bic r4, r4, #0xc0\r
173 orr r4, r4, #0x80\r
174 strb r4, [r1,#\ofs]\r
1753:\r
176.endm\r
177\r
178@ TileFlipSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
179.macro TileFlipSh\r
180 TileDoShGenPixel 16, 0 @ #0x000f0000\r
181 TileDoShGenPixel 20, 1 @ #0x00f00000\r
182 TileDoShGenPixel 24, 2 @ #0x0f000000\r
183 TileDoShGenPixel 28, 3 @ #0xf0000000\r
184 TileDoShGenPixel 0, 4 @ #0x0000000f\r
185 TileDoShGenPixel 4, 5 @ #0x000000f0\r
186 TileDoShGenPixel 8, 6 @ #0x00000f00\r
187 TileDoShGenPixel 12, 7 @ #0x0000f000\r
188.endm\r
189\r
190@ TileNormSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
191.macro TileNormSh\r
192 TileDoShGenPixel 12, 0 @ #0x0000f000\r
193 TileDoShGenPixel 8, 1 @ #0x00000f00\r
194 TileDoShGenPixel 4, 2 @ #0x000000f0\r
195 TileDoShGenPixel 0, 3 @ #0x0000000f\r
196 TileDoShGenPixel 28, 4 @ #0xf0000000\r
197 TileDoShGenPixel 24, 5 @ #0x0f000000\r
198 TileDoShGenPixel 20, 6 @ #0x00f00000\r
199 TileDoShGenPixel 16, 7 @ #0x000f0000\r
200.endm\r
201\r
202\r
203@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
204\r
205@ struct TileStrip\r
206@ {\r
207@ int nametab; // 0x00\r
208@ int line; // 0x04\r
209@ int hscroll; // 0x08\r
210@ int xmask; // 0x0C\r
211@ int *hc; // 0x10 (pointer to cache buffer)\r
212@ int cells; // 0x14\r
213@ };\r
214\r
215@ int DrawLayer(int plane, int *hcache, int maxcells, int sh)\r
216\r
217.global DrawLayer @ int plane, int *hcache, int maxcells, int sh\r
218\r
219DrawLayer:\r
220 stmfd sp!, {r4-r11,lr}\r
221\r
222 ldr r11, =(Pico+0x22228) @ Pico.video\r
223 mov r8, #1\r
224\r
225 ldrb r7, [r11, #16] @ ??hh??ww\r
226\r
227 mov r6, r1 @ hcache\r
228 orr r9, r2, r3, lsl #31 @ r9=maxcells|(sh<<31)\r
229\r
230 mov r1, r7, lsl #4\r
231 orr r1, r1, #0x00ff\r
232\r
233 and r10, r7, #3\r
234 cmp r10, #1\r
235 biclt r1, r1, #0xfc00\r
236 biceq r1, r1, #0xfe00\r
237 bicgt r1, r1, #0xff00 @ r1=ymask=(height<<8)|0xff; ...; // Y Mask in pixels\r
238\r
239 add r10, r10, #5\r
240 cmp r10, #7\r
241 subge r10, r10, #1 @ r10=shift[width] (5,6,6,7)\r
242\r
243 @ calculate xmask:\r
244 mov r5, r8, lsl r10\r
245 sub r5, r5, #1 @ r5=xmask\r
246\r
247 @ Find name table:\r
248 tst r0, r0\r
249 ldreqb r12, [r11, #2]\r
250 ldrneb r12, [r11, #4]\r
251\r
252 ldr r2, =Scanline @ trying to make good use of pipeline here\r
253 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
254\r
255 moveq r12, r12, lsl #10\r
256 movne r12, r12, lsl #13\r
257 and r12, r12, #(7<<13) @ r12=(ts->nametab<<1) (halfword compliant)\r
258\r
259 ldrh r8, [r11, #12]\r
260 ldrb r7, [r11, #11]\r
261 ldr r2, [r2]\r
262\r
263 mov r4, r8, lsr #8 @ pvid->reg[13]\r
264 mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r
265 tst r7, #2\r
266 addne r4, r4, r2, lsl #2 @ htab+=Scanline<<1; // Offset by line\r
267 tst r7, #1\r
268 biceq r4, r4, #0x1f @ htab&=~0xf; // Offset by tile\r
269 add r4, r4, r0, lsl #1 @ htab+=plane\r
270 bic r4, r4, #0x00ff0000 @ just in case\r
271 ldrh r3, [lr, r4] @ r3=hscroll\r
272\r
273 tst r7, #4\r
274 bne .DrawStrip_vsscroll\r
275\r
276 @ Get vertical scroll value:\r
277 add r7, lr, #0x012000\r
278 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
279 ldr r7, [r7]\r
280\r
281 tst r8, #2\r
282 tstne r8, #4\r
283 bne .DrawStrip_interlace\r
284\r
285 tst r0, r0\r
286 movne r7, r7, lsr #16\r
287\r
288 @ Find the line in the name table\r
289 add r2, r2, r7\r
290 and r2, r2, r1\r
291 mov r4, r2, lsr #3\r
292 add r10, r10, #1 @ shift[width]++\r
293 add r12, r12, r4, lsl r10 @ nametab+=(ts.line>>3)<<shift[width];\r
294\r
295 @ ldmia r0, {r1,r2,r3,r5,r6,r9} @ r2=line, r3=ts->hscroll, r5=ts->xmask, r6=ts->hc, r9=ts->cells\r
296@ mov r12,r1, lsl #1 @ r12=(ts->nametab<<1) (halfword compliant)\r
297\r
298 and r10,r2, #7\r
299 mov r10,r10, lsl #1 @ r10=ty=(ts->line&7)<<1;\r
300 orr r10,r10, r9, lsl #24\r
301\r
302 rsb r8, r3, #0\r
303 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
304\r
305 sub r1, r3, #1\r
306 and r1, r1, #7\r
307 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
308\r
309 tst r9, #1<<31\r
310 mov r3, #0\r
311 orrne r10,r10, #1<<23 @ r10=(cells<<24|sh<<23|hi_not_empty<<22|ty)\r
312 movne r3, #0x40 @ default to shadowed pal on sh mode\r
313\r
314 mvn r9, #0 @ r9=prevcode=-1\r
315\r
316 cmp r7, #8\r
317 addne r10,r10, #0x01000000 @ we will loop cells+1 times if there is scroll\r
318\r
319 @ cache some stuff to avoid mem access\r
320 ldr r11,=HighCol\r
321 mov r0, #0xf\r
322 add r1, r11, r7 @ r1=pdest\r
323\r
324\r
325 @ r4 & r7 are scratch in this loop\r
326.dsloop_subr1:\r
327 sub r1, r1, #8\r
328.dsloop: @ 40-41 times\r
329 subs r10,r10, #0x01000000\r
330 bmi .dsloop_exit\r
331\r
332.dsloop_enter:\r
333 and r7, r5, r8\r
334 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
335 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
336\r
337 add r1, r1, #8\r
338 add r8, r8, #1\r
339\r
340 tst r7, #0x8000\r
341 bne .DrawStrip_hiprio\r
342\r
343 cmp r7, r9\r
344 beq .DrawStrip_samecode @ we know stuff about this tile already\r
345\r
346 mov r9, r7 @ remember code\r
347\r
348 movs r2, r9, lsl #20 @ if (code&0x1000)\r
349 mov r2, r2, lsl #1\r
350 add r2, r2, r10, lsl #17\r
351 mov r2, r2, lsr #17\r
352 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
353\r
354 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
355\r
356 bic r7, r3, #0x3f\r
357 and r3, r9, #0x6000\r
358 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
359\r
360.DrawStrip_samecode:\r
361 tst r2, r2\r
362 beq .dsloop @ tileline blank\r
363\r
364 cmp r2, r2, ror #4\r
365 beq .DrawStrip_SingleColor @ tileline singlecolor \r
366\r
367 tst r9, #0x0800\r
368 beq .DrawStrip_TileNorm\r
369\r
370 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
371 TileFlip r0\r
372 b .dsloop\r
373\r
374.DrawStrip_TileNorm:\r
375 TileNorm r0\r
376 b .dsloop\r
377\r
378.DrawStrip_SingleColor:\r
379 and r4, r2, #0xf\r
380 orr r4, r3, r4\r
381 orr r4, r4, r4, lsl #8\r
382 tst r1, #1 @ not aligned?\r
383 strneb r4, [r1], #1\r
384 streqh r4, [r1], #2\r
385 strh r4, [r1], #2\r
386 strh r4, [r1], #2\r
387 strh r4, [r1], #2\r
388 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
389 b .dsloop_subr1\r
390\r
391.DrawStrip_hiprio:\r
392 tst r10, #0x00c00000\r
393 beq .DrawStrip_hiprio_maybempt\r
394 sub r0, r1, r11\r
395 orr r7, r7, r0, lsl #16\r
396 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
397 tst r7, #0x1000\r
398 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
399 str r7, [r6], #4 @ cache hi priority tile\r
400 mov r0, #0xf\r
401 b .dsloop\r
402\r
403.DrawStrip_hiprio_maybempt:\r
404 cmp r7, r9\r
405 beq .dsloop @ must've been empty, otherwise we wouldn't get here\r
406 movs r2, r7, lsl #20 @ if (code&0x1000)\r
407 mov r2, r2, lsl #1\r
408 add r2, r2, r10, lsl #17\r
409 mov r2, r2, lsr #17\r
410 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
411 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
412 mov r9, r7 @ remember code\r
413 tst r2, r2\r
414 orrne r10, r10, #1<<22\r
415 bne .DrawStrip_hiprio\r
416 b .dsloop\r
417\r
418.dsloop_exit:\r
419 mov r0, #0\r
420 str r0, [r6] @ terminate the cache list\r
421\r
422 ldmfd sp!, {r4-r11,lr}\r
423 bx lr\r
424\r
425\r
426.DrawStrip_vsscroll:\r
427 @ shit, we have 2-cell column based vscroll\r
428 @ let the c code handle this (for now)\r
429\r
430 @ int nametab; // 0x00\r
431 @ int line; // 0x04\r
432 @ int hscroll; // 0x08\r
433 @ int xmask; // 0x0C\r
434 @ int *hc; // 0x10 (pointer to cache buffer)\r
435 @ int cells; // 0x14\r
436\r
437 sub sp, sp, #6*4\r
438 orr r2, r1, r10, lsl #24 @ ts.line=ymask|(shift[width]<<24); // save some stuff instead of line\r
439 mov r1, r0 @ plane\r
440 mov r0, r12, lsr #1 @ halfwords\r
441 and r9, r9, #0xff\r
442 stmia sp, {r0,r2,r3,r5,r6,r9}\r
443\r
444 mov r0, sp\r
445 bl DrawStripVSRam @ struct TileStrip *ts, int plane\r
446\r
447 add sp, sp, #6*4\r
448 ldmfd sp!, {r4-r11,lr}\r
449 bx lr\r
450\r
451@ interlace mode 2? Sonic 2?\r
452.DrawStrip_interlace:\r
453 tst r0, r0\r
454 moveq r7, r7, lsl #21\r
455 movne r7, r7, lsl #5\r
456\r
457 @ Find the line in the name table\r
458 add r2, r7, r2, lsl #22 @ r2=(vscroll+(Scanline<<1))<<21 (11 bits);\r
459 orr r1, r1, #0x80000000\r
460 and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
461 mov r2, r2, lsr #21\r
462 mov r4, r2, lsr #4\r
463 mov r12, r12, lsr #1 @ halfwords\r
464 add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];\r
465 and r9, r9, #0xff\r
466\r
467 sub sp, sp, #6*4\r
468 stmia sp, {r0,r2,r3,r5,r6,r9}\r
469\r
470 mov r0, sp\r
471 bl DrawStripInterlace @ struct TileStrip *ts\r
472\r
473 add sp, sp, #6*4\r
474 ldmfd sp!, {r4-r11,lr}\r
475 bx lr\r
476\r
477.pool\r
478\r
479@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
480\r
481\r
482.global BackFill @ int reg7, int sh\r
483\r
484BackFill:\r
485 stmfd sp!, {r4-r9,lr}\r
486\r
487 ldr lr, =(HighCol+8)\r
488\r
489 mov r0, r0, lsl #26\r
490 mov r0, r0, lsr #26\r
491 orr r0, r0, r1, lsl #6\r
492 orr r0, r0, r0, lsl #8\r
493 orr r0, r0, r0, lsl #16\r
494\r
495 mov r1, r0\r
496 mov r2, r0\r
497 mov r3, r0\r
498 mov r4, r0\r
499 mov r5, r0\r
500 mov r6, r0\r
501 mov r7, r0\r
502\r
503 @ go go go!\r
504 stmia lr!, {r0-r7} @ 10*8*4\r
505 stmia lr!, {r0-r7}\r
506 stmia lr!, {r0-r7}\r
507 stmia lr!, {r0-r7}\r
508 stmia lr!, {r0-r7}\r
509 stmia lr!, {r0-r7}\r
510 stmia lr!, {r0-r7}\r
511 stmia lr!, {r0-r7}\r
512 stmia lr!, {r0-r7}\r
513 stmia lr!, {r0-r7}\r
514\r
515 ldmfd sp!, {r4-r9,r12}\r
516 bx r12\r
517\r
518\r
519@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
520\r
521\r
522.global DrawTilesFromCache @ int *hc, int sh\r
523\r
524DrawTilesFromCache:\r
525 stmfd sp!, {r4-r8,r11,lr}\r
526\r
527 mvn r5, #0 @ r5=prevcode=-1\r
528 mov r8, r1\r
529\r
530 @ cache some stuff to avoid mem access\r
531 ldr r11,=HighCol\r
532 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
533 mov r12,#0xf\r
534\r
535 @ scratch: r4, r7\r
536.dtfc_loop:\r
537 ldr r6, [r0], #4 @ read code\r
538 movs r1, r6, lsr #16 @ r1=dx;\r
539 ldmeqfd sp!, {r4-r8,r11,pc} @ dx is never zero, this must be a terminator, return\r
540 bic r1, r1, #0xfe00\r
541 add r1, r11, r1 @ r1=pdest\r
542\r
543 mov r7, r6, lsl #16\r
544 cmp r5, r7, lsr #16\r
545 beq .dtfc_samecode @ if (code==prevcode)\r
546\r
547 mov r5, r7, lsr #16\r
548\r
549 mov r2, r5, lsl #21\r
550 mov r2, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4;\r
551 add r2, r2, r6, lsr #25 @ addr+=ty\r
552\r
553 and r3, r5, #0x6000\r
554 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
555\r
556 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
557\r
558.dtfc_samecode:\r
559 tst r8, r8\r
560 bne .dtfc_shadow\r
561\r
562 tst r2, r2\r
563 beq .dtfc_loop\r
564\r
565 cmp r2, r2, ror #4\r
566 beq .dtfc_SingleColor @ tileline singlecolor \r
567\r
568 tst r5, #0x0800\r
569 beq .dtfc_TileNorm\r
570\r
571 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
572 TileFlip r12\r
573 b .dtfc_loop\r
574\r
575.dtfc_TileNorm:\r
576 TileNorm r12\r
577 b .dtfc_loop\r
578\r
579.dtfc_SingleColor:\r
580 and r4, r2, #0xf\r
581 orr r4, r3, r4\r
582 orr r4, r4, r4, lsl #8\r
583 tst r1, #1 @ not aligned?\r
584 strneb r4, [r1], #1\r
585 streqh r4, [r1], #2\r
586 strh r4, [r1], #2\r
587 strh r4, [r1], #2\r
588 strh r4, [r1], #2\r
589 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
590 b .dtfc_loop\r
591\r
592.dtfc_shadow:\r
593 tst r2, r2\r
594 beq .dtfc_shadow_blank\r
595\r
596 cmp r2, r2, ror #4\r
597 beq .dtfc_SingleColor @ tileline singlecolor \r
598\r
599 tst r5, #0x0800\r
600 beq .dtfc_TileNormShHP\r
601\r
602 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
603 TileFlipShHP\r
604 b .dtfc_loop\r
605\r
606.dtfc_TileNormShHP:\r
607 TileNormShHP\r
608 b .dtfc_loop\r
609\r
610.dtfc_shadow_blank:\r
611 ldrb r4, [r1] @ 1ci\r
612 ldrb r12,[r1,#1]\r
613 tst r4, #0x80\r
614 andeq r4, r4,#0x3f\r
615 streqb r4, [r1]\r
616 tst r12,#0x80\r
617 ldrb r4, [r1,#2]\r
618 andeq r12,r12,#0x3f\r
619 streqb r12,[r1,#1]\r
620 tst r4, #0x80\r
621 ldrb r12,[r1,#3]\r
622 andeq r4, r4,#0x3f\r
623 streqb r4, [r1,#2]\r
624 tst r12,#0x80\r
625 ldrb r4, [r1,#4]\r
626 andeq r12,r12,#0x3f\r
627 streqb r12,[r1,#3]\r
628 tst r4, #0x80\r
629 ldrb r12,[r1,#5]\r
630 andeq r4, r4,#0x3f\r
631 streqb r4, [r1,#4]\r
632 tst r12,#0x80\r
633 ldrb r4, [r1,#6]\r
634 andeq r12,r12,#0x3f\r
635 streqb r12,[r1,#5]\r
636 tst r4, #0x80\r
637 ldrb r12,[r1,#7]\r
638 andeq r4, r4,#0x3f\r
639 streqb r4, [r1,#6]\r
640 tst r12,#0x80\r
641 andeq r12,r12,#0x3f\r
642 streqb r12,[r1,#7]\r
643 mov r12, #0xf\r
644 b .dtfc_loop\r
645\r
646.pool\r
647\r
648@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
649\r
650\r
651.global DrawSpritesFromCache @ int *hc, int sh\r
652\r
653DrawSpritesFromCache:\r
654 stmfd sp!, {r4-r11,lr}\r
655\r
656 @ cache some stuff to avoid mem access\r
657 ldr r11,=HighCol\r
658 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
659 mov r6, r1, lsl #31\r
660 orr r6, r6, #1<<30\r
661 mov r12,#0xf\r
662\r
663 mov r10, r0\r
664\r
665.dsfc_loop:\r
666 ldr r9, [r10], #4 @ read code\r
667 bic r6, r6, #7 @ using pipeline\r
668 tst r9, r9\r
669 ldmeqfd sp!, {r4-r11,pc}\r
670\r
671 mov r4, r9, lsl #28\r
672 orr r6, r6, r4, lsr #30\r
673 add r6, r6, #1 @ r6=s1cc???? ... ?????www (s=shadow/hilight, cc=pal, w=width)\r
674\r
675 and r5, r9, #3\r
676 add r5, r5, #1 @ r5=delta\r
677 tst r9, #0x10000\r
678 rsbne r5, r5, #0 @ Flip X\r
679 mov r5, r5, lsl #4\r
680\r
681 mov r2, r9, lsr #17\r
682 mov r8, r2, lsl #1 @ tile=((unsigned int)code>>17)<<1;\r
683\r
684 and r3, r9, #0x30 @ r3=pal=(code&0x30);\r
685\r
686 bic r6, r6, #3<<28\r
687 orr r6, r6, r3, lsl #24\r
688\r
689 mov r0, r9, lsl #16\r
690 mov r0, r0, asr #22 @ sx=(code<<16)>>22\r
691 adds r0, r0, #0 @ set ZV\r
692 b .dsfc_inloop_enter\r
693\r
694@ scratch: r4, r7\r
695.dsfc_inloop:\r
696 sub r6, r6, #1\r
697 tst r6, #7\r
698 beq .dsfc_loop\r
699 adds r0, r0, #8\r
700 add r8, r8, r5\r
701\r
702.dsfc_inloop_enter:\r
703 ble .dsfc_inloop\r
704 cmp r0, #328\r
705 bge .dsfc_loop\r
706\r
707 mov r8, r8, lsl #17\r
708 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
709\r
710 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+tile); // Get 8 pixels\r
711 add r1, r11, r0 @ r1=pdest\r
712 tst r2, r2\r
713 beq .dsfc_inloop\r
714\r
715 cmp r12, r6, lsr #28\r
716 beq .dsfc_shadow\r
717\r
718 cmp r2, r2, ror #4\r
719 beq .dsfc_SingleColor @ tileline singlecolor \r
720\r
721 tst r9, #0x10000\r
722 beq .dsfc_TileNorm\r
723\r
724 @ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
725 TileFlip r12\r
726 b .dsfc_inloop\r
727\r
728.dsfc_TileNorm:\r
729 TileNorm r12\r
730 b .dsfc_inloop\r
731\r
732.dsfc_SingleColor:\r
733 tst r0, #1 @ not aligned?\r
734 and r4, r2, #0xf\r
735 orr r4, r3, r4\r
736 orr r4, r4, r4, lsl #8\r
737 strneb r4, [r1], #1\r
738 streqh r4, [r1], #2\r
739 strh r4, [r1], #2\r
740 strh r4, [r1], #2\r
741 strh r4, [r1], #2\r
742 strneb r4, [r1], #1\r
743 b .dsfc_inloop\r
744\r
745.dsfc_shadow:\r
746 cmp r2, r2, ror #4\r
747 beq .dsfc_singlec_sh\r
748\r
749 tst r9, #0x10000\r
750 beq .dsfc_TileNorm_sh\r
751\r
752 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
753 TileFlipSh\r
754 b .dsfc_inloop\r
755\r
756.dsfc_TileNorm_sh:\r
757 TileNormSh\r
758 b .dsfc_inloop\r
759\r
760.dsfc_singlec_sh:\r
761 cmp r2, #0xe0000000\r
762 bcc .dsfc_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
763 tst r2, #0x10000000\r
764 bne .dsfc_sh_sh\r
765 TileSingleHi\r
766 b .dsfc_inloop\r
767\r
768.dsfc_sh_sh:\r
769 TileSingleSh\r
770 b .dsfc_inloop\r
771\r
772.pool\r
773\r
774@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
775\r
776@ + 0 : hhhhvvvv ab--hhvv yyyyyyyy yyyyyyyy // a: offscreen h, b: offs. v, h: horiz. size\r
777@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
778\r
779.global DrawSprite @ unsigned int *sprite, int **hc, int sh\r
780\r
781DrawSprite:\r
782 stmfd sp!, {r4-r9,r11,lr}\r
783\r
784 ldr r3, [r0] @ sprite[0]\r
785 ldr r7, =Scanline\r
786 mov r6, r3, lsr #28\r
787 sub r6, r6, #1 @ r6=width-1 (inc later)\r
788 mov r5, r3, lsr #24\r
789 and r5, r5, #7 @ r5=height\r
790\r
791 mov r4, r3, lsl #16 @ r4=sy<<16 (tmp)\r
792\r
793 ldr r7, [r7]\r
794 ldr r9, [r0, #4]\r
795 sub r7, r7, r4, asr #16 @ r7=row=Scanline-sy\r
796\r
797 tst r2, r2\r
798 mov r2, r9, asr #16 @ r2=sx\r
799 bic r9, r9, #0xfe000000\r
800 orrne r9, r9, #1<<31 @ r9=code|(sh<<31)\r
801\r
802 tst r9, #0x1000\r
803 movne r4, r5, lsl #3\r
804 subne r4, r4, #1\r
805 subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
806\r
807 mov r8, r9, lsl #21\r
808 mov r8, r8, lsr #21\r
809 add r8, r8, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
810 \r
811 tst r9, #0x0800\r
812 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
813 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
814\r
815 mov r8, r8, lsl #4\r
816 and r7, r7, #7\r
817 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
818\r
819 tst r9, #0x8000\r
820 bne .dspr_cache @ if(code&0x8000) // high priority - cache it\r
821\r
822 @ cache some stuff to avoid mem access\r
823 ldr r11,=HighCol\r
824 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
825 mov r12,#0xf\r
826\r
827 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
828 and r4, r9, #0x6000\r
829 orr r9, r9, r4, lsl #16\r
830 orr r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
831\r
832 tst r9, #1<<31\r
833 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
834 orrne r3, r3, #0x40 @ shadow by default\r
835\r
836 add r6, r6, #1 @ inc now\r
837 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
838 b .dspr_loop_enter\r
839\r
840.dspr_loop:\r
841 subs r6, r6, #1 @ width--\r
842 ldmeqfd sp!, {r4-r9,r11,pc}@ return\r
843 adds r0, r0, #8 @ sx+=8\r
844 add r8, r8, r5 @ tile+=delta\r
845\r
846.dspr_loop_enter:\r
847 ble .dspr_loop @ sx <= 0\r
848 cmp r0, #328\r
849 ldmgefd sp!, {r4-r9,r11,pc}@ return\r
850\r
851 mov r8, r8, lsl #17\r
852 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
853\r
854 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
855 add r1, r11, r0 @ r1=pdest\r
856 tst r2, r2\r
857 beq .dspr_loop\r
858\r
859 cmp r12, r9, lsr #28\r
860 beq .dspr_shadow\r
861\r
862 cmp r2, r2, ror #4\r
863 beq .dspr_SingleColor @ tileline singlecolor \r
864\r
865 tst r9, #0x0800\r
866 beq .dspr_TileNorm\r
867\r
868 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
869 TileFlip r12\r
870 b .dspr_loop\r
871\r
872@ scratch: r4, r7\r
873.dspr_TileNorm:\r
874 TileNorm r12\r
875 b .dspr_loop\r
876\r
877.dspr_SingleColor:\r
878 and r4, r2, #0xf\r
879 orr r4, r3, r4\r
880 orr r4, r4, r4, lsl #8\r
881 tst r0, #1 @ not aligned?\r
882 strneb r4, [r1], #1\r
883 streqh r4, [r1], #2\r
884 strh r4, [r1], #2\r
885 strh r4, [r1], #2\r
886 strh r4, [r1], #2\r
887 strneb r4, [r1], #1\r
888 b .dspr_loop\r
889\r
890.dspr_shadow:\r
891 cmp r2, r2, ror #4\r
892 beq .dspr_singlec_sh\r
893\r
894 tst r9, #0x0800\r
895 beq .dspr_TileNorm_sh\r
896\r
897 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
898 TileFlipSh\r
899 b .dspr_loop\r
900\r
901.dspr_TileNorm_sh:\r
902 TileNormSh\r
903 b .dspr_loop\r
904\r
905.dspr_singlec_sh:\r
906 cmp r2, #0xe0000000\r
907 bcc .dspr_SingleColor @ normal tileline\r
908 tst r2, #0x10000000\r
909 bne .dspr_sh_sh\r
910 TileSingleHi\r
911 b .dspr_loop\r
912\r
913.dspr_sh_sh:\r
914 TileSingleSh\r
915 b .dspr_loop\r
916\r
917\r
918.dspr_cache:\r
919 @ *(*hc)++ = (tile<<16)|((code&0x0800)<<5)|((sx<<6)&0x0000ffc0)|((code>>9)&0x30)|((sprite[0]>>24)&0xf);\r
920 mov r4, r8, lsl #16 @ tile\r
921 tst r9, #0x0800\r
922 orrne r4, r4, #0x10000 @ code&0x0800\r
923 mov r2, r2, lsl #22\r
924 orr r4, r4, r2, lsr #16 @ (sx<<6)&0x0000ffc0\r
925 and r2, r9, #0x6000\r
926 orr r4, r4, r2, lsr #9 @ (code>>9)&0x30\r
927 mov r3, r3, lsl #12\r
928 ldr r2, [r1]\r
929 orr r4, r4, r3, lsr #28 @ (sprite[0]>>24)&0xf\r
930\r
931 str r4, [r2], #4\r
932 str r2, [r1]\r
933\r
934 ldmfd sp!, {r4-r9,r11,lr}\r
935 bx lr\r
936\r
937@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
938\r
939.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache\r
940\r
941DrawWindow:\r
942 stmfd sp!, {r4-r11,lr}\r
943\r
944 ldr r11, =(Pico+0x22228) @ Pico.video\r
945 ldr r10, =Scanline\r
946 ldrb r12, [r11, #3] @ pvid->reg[3]\r
947\r
948 ldr r10, [r10]\r
949 ldr r4, [r11, #12]\r
950 mov r5, r10, lsr #3\r
951 and r10, r10, #7\r
952 mov r10, r10, lsl #1 @ r10=ty\r
953\r
954 mov r12, r12, lsl #10\r
955\r
956 tst r4, #1 @ 40 cell mode?\r
957 andne r12, r12, #0xf000 @ 0x3c<<10\r
958 andeq r12, r12, #0xf800\r
959 addne r12, r12, r5, lsl #7\r
960 addeq r12, r12, r5, lsl #6 @ nametab\r
961 add r12, r12, r0, lsl #2 @ +starttile\r
962\r
963 ldr r6, =rendstatus\r
964 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
965 ldrb r6, [r6]\r
966\r
967 @ fetch the first code now\r
968 ldrh r7, [lr, r12]\r
969\r
970 ands r6, r6, #2 @ we care about bit 1 only\r
971 orr r6, r6, r2\r
972 bne .dw_no_sameprio\r
973\r
974 cmp r2, r7, lsr #15\r
975 ldmnefd sp!, {r4-r11,pc} @ assume that whole window uses same priority\r
976\r
977.dw_no_sameprio:\r
978 orr r6, r6, r3, lsl #8 @ shadow mode\r
979\r
980 sub r8, r1, r0\r
981 mov r8, r8, lsl #1 @ cells\r
982\r
983 mvn r9, #0 @ r9=prevcode=-1\r
984\r
985 @ cache some stuff to avoid mem access\r
986 ldr r11,=(HighCol+8)\r
987 add r1, r11, r0, lsl #4 @ r1=pdest\r
988 mov r0, #0xf\r
989 b .dwloop_enter\r
990\r
991 @ r4,r5 & r7 are scratch in this loop\r
992.dwloop:\r
993 add r1, r1, #8\r
994.dwloop_nor1:\r
995 add r12, r12, #2 @ halfwords\r
996 ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)\r
997 subs r8, r8, #1\r
998 beq .dwloop_end @ done\r
999\r
1000 eor r5, r6, r7, lsr #15\r
1001 tst r5, #1\r
1002 orrne r6, r6, #2 @ wrong pri\r
1003 bne .dwloop\r
1004\r
1005 cmp r7, r9\r
1006 beq .dw_samecode @ we know stuff about this tile already\r
1007\r
1008.dwloop_enter:\r
1009 mov r9, r7 @ remember code\r
1010\r
1011 movs r2, r9, lsl #20 @ if (code&0x1000)\r
1012 mov r2, r2, lsl #1\r
1013 add r2, r10, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4; addr+=ty\r
1014 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
1015\r
1016 and r3, r9, #0x6000\r
1017 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
1018\r
1019 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1020\r
1021.dw_samecode:\r
1022 tst r6, #0x100\r
1023 bne .dw_shadow\r
1024.dw_shadow_done:\r
1025 tst r2, r2\r
1026 beq .dwloop @ tileline blank\r
1027\r
1028 cmp r2, r2, ror #4\r
1029 beq .dw_SingleColor @ tileline singlecolor \r
1030\r
1031 tst r9, #0x0800\r
1032 beq .dw_TileNorm\r
1033\r
1034 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
1035 TileFlip r0\r
1036 b .dwloop\r
1037\r
1038.dw_TileNorm:\r
1039 TileNorm r0\r
1040 b .dwloop\r
1041\r
1042.dw_SingleColor:\r
1043 and r4, r0, r2 @ #0x0000000f\r
1044 orr r4, r3, r4\r
1045 orr r4, r4, r4, lsl #8\r
1046 orr r4, r4, r4, lsl #16\r
1047 mov r5, r4\r
1048 stmia r1!, {r4,r5}\r
1049 b .dwloop_nor1 @ we incremeted r1 ourselves\r
1050\r
1051.dw_shadow:\r
1052 tst r6, #1 @ hi pri?\r
1053 orreq r3, r3, #0x40\r
1054 beq .dw_shadow_done\r
1055 ldr r4, [r1]\r
1056 tst r4, #0x00000080\r
1057 biceq r4, r4, #0x000000c0\r
1058 tst r4, #0x00008000\r
1059 biceq r4, r4, #0x0000c000\r
1060 tst r4, #0x00800000\r
1061 biceq r4, r4, #0x00c00000\r
1062 tst r4, #0x80000000\r
1063 biceq r4, r4, #0xc0000000\r
1064 str r4, [r1]\r
1065 ldr r4, [r1,#4]\r
1066 tst r4, #0x00000080\r
1067 biceq r4, r4, #0x000000c0\r
1068 tst r4, #0x00008000\r
1069 biceq r4, r4, #0x0000c000\r
1070 tst r4, #0x00800000\r
1071 biceq r4, r4, #0x00c00000\r
1072 tst r4, #0x80000000\r
1073 biceq r4, r4, #0xc0000000\r
1074 str r4, [r1,#4]\r
1075 b .dw_shadow_done\r
1076\r
1077.dwloop_end:\r
1078 ldr r0, =rendstatus\r
1079 ldr r1, [r0]\r
1080 and r6, r6, #2\r
1081 orr r1, r1, r6\r
1082 str r1, [r0]\r
1083\r
1084 ldmfd sp!, {r4-r11,r12}\r
1085 bx r12\r
1086\r
1087\r
1088@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1089\r
1090\r
1091@ hilights 2 pixels in RGB444/BGR444 format\r
1092.macro TileDoShHi2Pixels444 reg\r
1093 mov \reg, \reg, ror #12\r
1094 adds \reg, \reg, #0x40000000\r
1095 orrcs \reg, \reg, #0xf0000000\r
1096 mov \reg, \reg, ror #28\r
1097 adds \reg, \reg, #0x40000000\r
1098 orrcs \reg, \reg, #0xf0000000\r
1099 mov \reg, \reg, ror #28\r
1100 adds \reg, \reg, #0x40000000\r
1101 orrcs \reg, \reg, #0xf0000000\r
1102 mov \reg, \reg, ror #24\r
1103 adds \reg, \reg, #0x40000000\r
1104 orrcs \reg, \reg, #0xf0000000\r
1105 mov \reg, \reg, ror #28\r
1106 adds \reg, \reg, #0x40000000\r
1107 orrcs \reg, \reg, #0xf0000000\r
1108 mov \reg, \reg, ror #28\r
1109 adds \reg, \reg, #0x40000000\r
1110 orrcs \reg, \reg, #0xf0000000\r
1111 mov \reg, \reg, ror #12\r
1112.endm\r
1113\r
1114\r
1115.global FinalizeLineBGR444 @ int sh\r
1116\r
1117FinalizeLineBGR444:\r
1118 stmfd sp!, {r4-r6,lr}\r
1119 mov r6, r0\r
1120 ldr lr, =(Pico+0x22228) @ Pico.video\r
1121 ldr r0, =DrawLineDest\r
1122 ldrb r12, [lr, #12]\r
1123 ldr r0, [r0]\r
1124 sub r3, lr, #0x128 @ r3=Pico.cram\r
1125\r
1126 tst r12, #1\r
1127 movne r2, #320/4 @ len\r
1128 bne .fl_no32colBGR444\r
1129 ldr r4, =PicoOpt\r
1130 mov r2, #256/4\r
1131 ldr r4, [r4]\r
1132 tst r4, #0x100\r
1133 addeq r0, r0, #32*2\r
1134\r
1135.fl_no32colBGR444:\r
1136 tst r6, r6\r
1137 beq .fl_noshBGR444\r
1138\r
1139 ldr r4, =HighPal\r
1140\r
1141 ldrb r12, [lr, #-0x1a] @ 0x2220e ~ dirtyPal\r
1142 tst r12, r12\r
1143 moveq r3, r4\r
1144 beq .fl_noshBGR444\r
1145 mov r12, #0\r
1146 strb r12, [lr, #-0x1a]\r
1147\r
1148 mov lr, #0x40/8\r
1149 @ copy pal:\r
1150.fl_loopcpBGR444:\r
1151 ldmia r3!, {r1,r5,r6,r12}\r
1152 subs lr, lr, #1\r
1153 stmia r4!, {r1,r5,r6,r12}\r
1154 bne .fl_loopcpBGR444\r
1155\r
1156 @ shadowed pixels:\r
1157 mov r12, #0x0077\r
1158 orr r12,r12,#0x0700\r
1159 orr r12,r12,r12,lsl #16\r
1160 sub r3, r3, #0x40*2\r
1161 add r5, r4, #0x80*2\r
1162 mov lr, #0x40/4\r
1163.fl_loopcpBGR444_sh:\r
1164 ldmia r3!, {r1,r6}\r
1165 subs lr, lr, #1\r
1166 and r1, r12, r1, lsr #1\r
1167 and r6, r12, r6, lsr #1\r
1168 stmia r4!, {r1,r6}\r
1169 stmia r5!, {r1,r6}\r
1170 bne .fl_loopcpBGR444_sh\r
1171\r
1172 @ hilighted pixels:\r
1173 sub r3, r3, #0x40*2\r
1174 mov lr, #0x40/2\r
1175.fl_loopcpBGR444_hi:\r
1176 ldr r1, [r3], #4\r
1177 TileDoShHi2Pixels444 r1\r
1178 str r1, [r4], #4\r
1179 subs lr, lr, #1\r
1180 bne .fl_loopcpBGR444_hi\r
1181\r
1182 sub r3, r4, #0x40*3*2\r
1183\r
1184\r
1185.fl_noshBGR444:\r
1186 ldr r1, =(HighCol+8)\r
1187 mov lr, #0xff\r
1188 mov lr, lr, lsl #1\r
1189\r
1190.fl_loopBGR444:\r
1191\r
1192 ldr r12, [r1], #4\r
1193 subs r2, r2, #1\r
1194\r
1195 and r4, lr, r12, lsl #1\r
1196 ldrh r4, [r3, r4]\r
1197 and r5, lr, r12, lsr #7\r
1198 ldrh r5, [r3, r5]\r
1199 and r6, lr, r12, lsr #15\r
1200 ldrh r6, [r3, r6]\r
1201 orr r4, r4, r5, lsl #16\r
1202\r
1203 and r5, lr, r12, lsr #23\r
1204 ldrh r5, [r3, r5] @ 2c.i.\r
1205 orr r5, r6, r5, lsl #16\r
1206\r
1207 stmia r0!, {r4,r5}\r
1208 bne .fl_loopBGR444\r
1209\r
1210\r
1211 ldmfd sp!, {r4-r6,lr}\r
1212 bx lr\r
1213\r
1214\r
1215@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1216\r
1217\r
1218@ hilights 2 pixels in RGB555/BGR555 format\r
1219.macro TileDoShHi2Pixels555 reg\r
1220 adds \reg, \reg, #0x40000000\r
1221 orrcs \reg, \reg, #0xf0000000\r
1222 mov \reg, \reg, ror #27\r
1223 adds \reg, \reg, #0x40000000\r
1224 orrcs \reg, \reg, #0xf0000000\r
1225 mov \reg, \reg, ror #26\r
1226 adds \reg, \reg, #0x40000000\r
1227 orrcs \reg, \reg, #0xf0000000\r
1228 mov \reg, \reg, ror #27\r
1229 adds \reg, \reg, #0x40000000\r
1230 orrcs \reg, \reg, #0xf0000000\r
1231 mov \reg, \reg, ror #27\r
1232 adds \reg, \reg, #0x40000000\r
1233 orrcs \reg, \reg, #0xf0000000\r
1234 mov \reg, \reg, ror #26\r
1235 adds \reg, \reg, #0x40000000\r
1236 orrcs \reg, \reg, #0xf0000000\r
1237 mov \reg, \reg, ror #27\r
1238.endm\r
1239\r
1240\r
1241@ Convert 0000bbb0 ggg0rrr0\r
1242@ to rrrrrggg gggbbbbb\r
1243\r
1244@ r2,r3,r9 - scratch, lr = 0x001c001c, r8 = 0x00030003\r
1245.macro convRGB565 reg\r
1246 and r2, lr, \reg,lsl #1\r
1247 and r9, r8, \reg,lsr #2\r
1248 orr r2, r2, r9 @ r2=red\r
1249 and r3, lr, \reg,lsr #7\r
1250 and r9, r8, \reg,lsr #10\r
1251 orr r3, r3, r9 @ r3=blue\r
1252 and \reg, \reg, lr, lsl #3\r
1253 orr \reg, \reg, \reg,lsl #3 @ green\r
1254 orr \reg, \reg, r2, lsl #11 @ add red back\r
1255 orr \reg, \reg, r3 @ add blue back\r
1256.endm\r
1257\r
1258vidConvCpyRGB565: @ void *to, void *from, int pixels\r
1259 stmfd sp!, {r4-r9,lr}\r
1260\r
1261 mov r12, r2, lsr #3 @ repeats\r
1262 mov lr, #0x001c0000\r
1263 orr lr, lr, #0x01c @ lr == pattern 0x001c001c\r
1264 mov r8, #0x00030000\r
1265 orr r8, r8, #0x003 @ lr == pattern 0x001c001c\r
1266\r
1267.loopRGB565:\r
1268 ldmia r1!, {r4-r7}\r
1269 subs r12, r12, #1\r
1270 convRGB565 r4\r
1271 str r4, [r0], #4\r
1272 convRGB565 r5\r
1273 str r5, [r0], #4\r
1274 convRGB565 r6\r
1275 str r6, [r0], #4\r
1276 convRGB565 r7\r
1277 str r7, [r0], #4\r
1278\r
1279 bgt .loopRGB565\r
1280\r
1281 ldmfd sp!, {r4-r9,lr}\r
1282 bx lr\r
1283\r
1284\r
1285\r
1286.global FinalizeLineRGB555 @ int sh\r
1287\r
1288FinalizeLineRGB555:\r
1289 stmfd sp!, {r4-r8,lr}\r
3d48f143 1290 ldr r8, =(Pico+0x22228) @ Pico.video\r
cc68a136 1291 ldr r4, =HighPal\r
1292\r
3d48f143 1293 ldrb r7, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
cc68a136 1294 mov r6, r0\r
1295 mov r1, #0\r
1296 tst r7, r7\r
1297 beq .fl_noconvRGB555\r
3d48f143 1298 strb r1, [r8, #-0x1a]\r
1299 sub r1, r8, #0x128 @ r1=Pico.cram\r
cc68a136 1300 mov r0, r4\r
1301 mov r2, #0x40\r
1302 bl vidConvCpyRGB565\r
1303\r
1304.fl_noconvRGB555:\r
cc68a136 1305 mov r3, r4\r
1306 tst r6, r6\r
1307 beq .fl_noshRGB555\r
1308 tst r7, r7\r
1309 beq .fl_noshRGB555\r
1310\r
1311 @ shadowed pixels:\r
1312 mov r12, #0x008e\r
1313 orr r12,r12,#0x7300\r
1314 orr r12,r12,r12,lsl #16\r
1315 add r4, r3, #0x40*2\r
1316 add r5, r3, #0xc0*2\r
1317 mov lr, #0x40/4\r
1318.fl_loopcpRGB555_sh:\r
1319 ldmia r3!, {r1,r6}\r
1320 subs lr, lr, #1\r
1321 and r1, r12, r1, lsr #1\r
1322 and r6, r12, r6, lsr #1\r
1323 stmia r4!, {r1,r6}\r
1324 stmia r5!, {r1,r6}\r
1325 bne .fl_loopcpRGB555_sh\r
1326\r
1327 @ hilighted pixels:\r
1328 sub r3, r3, #0x40*2\r
1329 mov lr, #0x40/2\r
1330.fl_loopcpRGB555_hi:\r
1331 ldr r1, [r3], #4\r
1332 TileDoShHi2Pixels555 r1\r
1333 str r1, [r4], #4\r
1334 subs lr, lr, #1\r
1335 bne .fl_loopcpRGB555_hi\r
1336\r
1337 sub r3, r3, #0x40*2\r
1338\r
cc68a136 1339.fl_noshRGB555:\r
3d48f143 1340 ldr r0, =DrawLineDest\r
cc68a136 1341 ldr r1, =(HighCol+8)\r
3d48f143 1342 ldr r0, [r0]\r
1343\r
1344 ldrb r12, [r8, #12]\r
cc68a136 1345 mov lr, #0xff\r
1346 mov lr, lr, lsl #1\r
1347\r
3d48f143 1348 tst r12, #1\r
1349 movne r2, #320/8 @ len\r
1350 bne .fl_no32colRGB555\r
1351 ldr r4, =PicoOpt\r
1352 mov r2, #256/8\r
1353 ldr r4, [r4]\r
1354 tst r4, #0x4000\r
1355 bne .fl_32scale_RGB555\r
1356 tst r4, #0x0100\r
1357 addeq r0, r0, #32*2\r
1358\r
1359.fl_no32colRGB555:\r
cc68a136 1360.fl_loopRGB555:\r
1361\r
1362 ldr r12, [r1], #4\r
1363 ldr r7, [r1], #4\r
1364\r
1365 and r4, lr, r12, lsl #1\r
1366 ldrh r4, [r3, r4]\r
1367 and r5, lr, r12, lsr #7\r
1368 ldrh r5, [r3, r5]\r
1369 and r6, lr, r12, lsr #15\r
1370 ldrh r6, [r3, r6]\r
1371 orr r4, r4, r5, lsl #16\r
1372\r
1373 and r5, lr, r12, lsr #23\r
1374 ldrh r5, [r3, r5]\r
1375 and r8, lr, r7, lsl #1\r
1376 ldrh r8, [r3, r8]\r
1377 orr r5, r6, r5, lsl #16\r
1378\r
1379 and r6, lr, r7, lsr #7\r
1380 ldrh r6, [r3, r6]\r
1381 and r12,lr, r7, lsr #15\r
1382 ldrh r12,[r3, r12]\r
1383 orr r8, r8, r6, lsl #16\r
1384\r
1385 and r6, lr, r7, lsr #23\r
1386 ldrh r6, [r3, r6] @ 1 cycle interlock here (r6)\r
1387 subs r2, r2, #1\r
1388 orr r12,r12, r6, lsl #16\r
1389\r
1390 stmia r0!, {r4,r5,r8,r12}\r
1391 bne .fl_loopRGB555\r
1392\r
3d48f143 1393 ldmfd sp!, {r4-r8,lr}\r
1394 bx lr\r
1395\r
1396\r
1397.fl_32scale_RGB555:\r
1398 stmfd sp!, {r9,r10}\r
1399 mov r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1400 orr r9, r9, #0x00e7\r
1401\r
1402.fl_loop32scale_RGB555:\r
1403 ldr r12, [r1], #4\r
1404 ldr r7, [r1], #4\r
cc68a136 1405\r
3d48f143 1406 and r4, lr, r12,lsl #1\r
1407 ldrh r4, [r3, r4]\r
1408 and r5, lr, r12,lsr #7\r
1409 ldrh r5, [r3, r5]\r
1410 and r4, r4, r9, lsl #2\r
1411 orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0\r
1412 and r5, r5, r9, lsl #2\r
1413 sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1\r
1414 add r4, r4, r6, lsl #16 @ pix_d 0, 1\r
1415 and r6, lr, r12,lsr #15\r
1416 ldrh r6, [r3, r6]\r
1417 and r12,lr, r12,lsr #23\r
1418 ldrh r12,[r3, r12]\r
1419 and r6, r6, r9, lsl #2\r
1420 add r5, r5, r6\r
1421 mov r5, r5, lsr #1\r
1422 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1423 orr r5, r5, r6, lsl #16\r
1424\r
1425 and r6, lr, r7, lsl #1\r
1426 ldrh r6, [r3, r6]\r
1427 and r12,r12,r9, lsl #2\r
1428 add r5, r5, r12,lsl #14 @ pix_d 2, 3\r
1429 and r6, r6, r9, lsl #2\r
1430 orr r6, r12,r6, lsl #16 @ pix_d 4, 5\r
1431\r
1432 and r12,lr, r7, lsr #7\r
1433 ldrh r12,[r3, r12]\r
1434 and r10,lr, r7, lsr #15\r
1435 ldrh r10,[r3, r10]\r
1436 and r12,r12,r9, lsl #2\r
1437 sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1\r
1438 add r8, r8, r6, lsr #18\r
1439 and r7, lr, r7, lsr #23\r
1440 ldrh r7, [r3, r7]\r
1441 and r10,r10,r9, lsl #2\r
1442 orr r8, r8, r10,lsl #15\r
1443 add r8, r8, r12,lsl #15 @ pix_d 6, 7\r
1444 sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2\r
1445 and r7, r7, r9, lsl #2\r
1446 add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3\r
1447 orr r10,r10,r7, lsl #16 @ pix_d 8, 9\r
1448\r
1449 subs r2, r2, #1\r
1450\r
1451 stmia r0!, {r4,r5,r6,r8,r10}\r
1452 bne .fl_loop32scale_RGB555\r
1453\r
1454 ldmfd sp!, {r9,r10}\r
cc68a136 1455 ldmfd sp!, {r4-r8,lr}\r
3d48f143 1456 bx lr\r
1457\r
cc68a136 1458\r
1459@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1460\r
1461@ utility\r
1462.global blockcpy @ void *dst, void *src, size_t n\r
1463\r
1464blockcpy:\r
1465 stmfd sp!, {r4,r5}\r
1466 mov r2, r2, lsr #4\r
1467blockcpy_loop:\r
1468 ldmia r1!, {r3-r5,r12}\r
1469 subs r2, r2, #1\r
1470 stmia r0!, {r3-r5,r12}\r
1471 bne blockcpy_loop\r
1472 ldmfd sp!, {r4,r5}\r
1473 bx lr\r
1474\r
1475\r
1476.global blockcpy_or @ void *dst, void *src, size_t n, int pat\r
1477\r
1478blockcpy_or:\r
1479 stmfd sp!, {r4-r6}\r
1480 orr r3, r3, r3, lsl #8\r
1481 orr r3, r3, r3, lsl #16\r
1482 mov r2, r2, lsr #4\r
1483blockcpy_loop_or:\r
1484 ldmia r1!, {r4-r6,r12}\r
1485 subs r2, r2, #1\r
1486 orr r4, r4, r3\r
1487 orr r5, r5, r3\r
1488 orr r6, r6, r3\r
1489 orr r12,r12,r3\r
1490 stmia r0!, {r4-r6,r12}\r
1491 bne blockcpy_loop_or\r
1492 ldmfd sp!, {r4-r6}\r
1493 bx lr\r
1494\r
fast MegaDrive/MegaCD/32X emulator