notaz.gp2x.de / picodrive.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
removed some indirect jumps
[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
cc68a136 17.extern DrawStripInterlace\r
18\r
19\r
20@ helper\r
21.macro TilePixel pat lsrr offs\r
22.if !\lsrr\r
23 ands r4, \pat, r2\r
24.else\r
25 ands r4, \pat, r2, lsr #\lsrr\r
26.endif\r
27 orrne r4, r3, r4\r
28 strneb r4, [r1,#\offs]\r
29.endm\r
30\r
31@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
32.macro TileNorm pat\r
33 TilePixel \pat, 12, 0 @ #0x0000f000\r
34 TilePixel \pat, 8, 1 @ #0x00000f00\r
35 TilePixel \pat, 4, 2 @ #0x000000f0\r
36 TilePixel \pat, 0, 3 @ #0x0000000f\r
37 TilePixel \pat, 28, 4 @ #0xf0000000\r
38 TilePixel \pat, 24, 5 @ #0x0f000000\r
39 TilePixel \pat, 20, 6 @ #0x00f00000\r
40 TilePixel \pat, 16, 7 @ #0x000f0000\r
41.endm\r
42\r
43@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
44.macro TileFlip pat\r
45 TilePixel \pat, 16, 0 @ #0x000f0000\r
46 TilePixel \pat, 20, 1 @ #0x00f00000\r
47 TilePixel \pat, 24, 2 @ #0x0f000000\r
48 TilePixel \pat, 28, 3 @ #0xf0000000\r
49 TilePixel \pat, 0, 4 @ #0x0000000f\r
50 TilePixel \pat, 4, 5 @ #0x000000f0\r
51 TilePixel \pat, 8, 6 @ #0x00000f00\r
52 TilePixel \pat, 12, 7 @ #0x0000f000\r
53.endm\r
54\r
55@ shadow/hilight mode\r
56\r
57@ this one is for hi priority layer\r
58.macro TilePixelShHP lsrr offs\r
59.if !\lsrr\r
60 ands r4, r12, r2\r
61.else\r
62 ands r4, r12, r2, lsr #\lsrr\r
63.endif\r
64 ldreqb r4, [r1,#\offs]\r
65 orrne r4, r3, r4\r
66 strneb r4, [r1,#\offs]\r
67 tsteq r4, #0x80\r
68 andeq r4, r4, #0x3f\r
69 streqb r4, [r1,#\offs]\r
70.endm\r
71\r
72@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
73.macro TileNormShHP\r
74 TilePixelShHP 12, 0 @ #0x0000f000\r
75 TilePixelShHP 8, 1 @ #0x00000f00\r
76 TilePixelShHP 4, 2 @ #0x000000f0\r
77 TilePixelShHP 0, 3 @ #0x0000000f\r
78 TilePixelShHP 28, 4 @ #0xf0000000\r
79 TilePixelShHP 24, 5 @ #0x0f000000\r
80 TilePixelShHP 20, 6 @ #0x00f00000\r
81 TilePixelShHP 16, 7 @ #0x000f0000\r
82.endm\r
83\r
84@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
85.macro TileFlipShHP\r
86 TilePixelShHP 16, 0 @ #0x000f0000\r
87 TilePixelShHP 20, 1 @ #0x00f00000\r
88 TilePixelShHP 24, 2 @ #0x0f000000\r
89 TilePixelShHP 28, 3 @ #0xf0000000\r
90 TilePixelShHP 0, 4 @ #0x0000000f\r
91 TilePixelShHP 4, 5 @ #0x000000f0\r
92 TilePixelShHP 8, 6 @ #0x00000f00\r
93 TilePixelShHP 12, 7 @ #0x0000f000\r
94.endm\r
95\r
96\r
97@ TileSingleSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx; r12: helper pattern 0xf\r
98.macro TileSingleSh\r
99 tst r0, #1 @ not aligned?\r
100 mov r7, #0x00c000\r
101 orr r7, r7, #0xc0\r
102 ldrneb r4, [r1]\r
103 ldreqh r4, [r1]\r
104 orr r4, r4, r7\r
105 strneb r4, [r1], #1\r
106 streqh r4, [r1], #2\r
107 ldrh r4, [r1]\r
108 orr r4, r4, r7\r
109 strh r4, [r1], #2\r
110 ldrh r4, [r1]\r
111 orr r4, r4, r7\r
112 strh r4, [r1], #2\r
113 ldrh r4, [r1]\r
114 orr r4, r4, r7\r
115 strh r4, [r1], #2\r
116 ldrneb r4, [r1]\r
117 orr r4, r4, r7\r
118 strneb r4, [r1], #1\r
119.endm\r
120\r
121@ TileSingleHi (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
122.macro TileSingleHi\r
123 tst r1, #1 @ not aligned?\r
124 mov r7, #0x008000\r
125 orr r7, r7, #0x80\r
126 ldrneb r4, [r1], #1\r
127 ldreqh r4, [r1], #2 @ 1ci\r
128 ldrh r12, [r1], #2\r
129 bic r4, r4, r7, lsr #1\r
130 orr r4, r4, r7\r
131 strneb r4, [r1, #-3]\r
132 streqh r4, [r1, #-4]\r
133 ldrh r4, [r1], #2\r
134 bic r12, r12, r7, lsr #1\r
135 orr r12, r12, r7\r
136 strh r12, [r1, #-4]\r
137 ldrh r12, [r1], #2\r
138 bic r4, r4, r7, lsr #1\r
139 orr r4, r4, r7\r
140 strh r4, [r1, #-4]\r
141 ldrneb r4, [r1]\r
142 bic r12, r12, r7, lsr #1\r
143 orr r12, r12, r7\r
144 strh r12, [r1, #-2]\r
145 bicne r4, r4, r7, lsr #1\r
146 orrne r4, r4, r7\r
147 strneb r4, [r1], #1\r
148 mov r12, #0xf\r
149.endm\r
150\r
151.macro TileDoShGenPixel shift ofs\r
152.if \shift\r
153 ands r4, r12, r2, lsr #\shift\r
154.else\r
155 ands r4, r12, r2\r
156.endif\r
157 beq 3f\r
158 cmp r4, #0xe\r
159 beq 2f\r
160 bgt 1f\r
161 orr r4, r3, r4\r
162 strb r4, [r1,#\ofs]\r
163 b 3f\r
1641:\r
165 ldrb r4, [r1,#\ofs] @ 2ci\r
166 orr r4, r4, #0xc0\r
167 strb r4, [r1,#\ofs]\r
168 b 3f\r
1692:\r
170 ldrb r4, [r1,#\ofs] @ 2ci\r
171 bic r4, r4, #0xc0\r
172 orr r4, r4, #0x80\r
173 strb r4, [r1,#\ofs]\r
1743:\r
175.endm\r
176\r
177@ TileFlipSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
178.macro TileFlipSh\r
179 TileDoShGenPixel 16, 0 @ #0x000f0000\r
180 TileDoShGenPixel 20, 1 @ #0x00f00000\r
181 TileDoShGenPixel 24, 2 @ #0x0f000000\r
182 TileDoShGenPixel 28, 3 @ #0xf0000000\r
183 TileDoShGenPixel 0, 4 @ #0x0000000f\r
184 TileDoShGenPixel 4, 5 @ #0x000000f0\r
185 TileDoShGenPixel 8, 6 @ #0x00000f00\r
186 TileDoShGenPixel 12, 7 @ #0x0000f000\r
187.endm\r
188\r
189@ TileNormSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
190.macro TileNormSh\r
191 TileDoShGenPixel 12, 0 @ #0x0000f000\r
192 TileDoShGenPixel 8, 1 @ #0x00000f00\r
193 TileDoShGenPixel 4, 2 @ #0x000000f0\r
194 TileDoShGenPixel 0, 3 @ #0x0000000f\r
195 TileDoShGenPixel 28, 4 @ #0xf0000000\r
196 TileDoShGenPixel 24, 5 @ #0x0f000000\r
197 TileDoShGenPixel 20, 6 @ #0x00f00000\r
198 TileDoShGenPixel 16, 7 @ #0x000f0000\r
199.endm\r
200\r
201\r
202@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
203\r
204@ struct TileStrip\r
205@ {\r
206@ int nametab; // 0x00\r
207@ int line; // 0x04\r
208@ int hscroll; // 0x08\r
209@ int xmask; // 0x0C\r
210@ int *hc; // 0x10 (pointer to cache buffer)\r
211@ int cells; // 0x14\r
212@ };\r
213\r
214@ int DrawLayer(int plane, int *hcache, int maxcells, int sh)\r
215\r
216.global DrawLayer @ int plane, int *hcache, int maxcells, int sh\r
217\r
218DrawLayer:\r
219 stmfd sp!, {r4-r11,lr}\r
220\r
221 ldr r11, =(Pico+0x22228) @ Pico.video\r
222 mov r8, #1\r
223\r
224 ldrb r7, [r11, #16] @ ??hh??ww\r
225\r
226 mov r6, r1 @ hcache\r
227 orr r9, r2, r3, lsl #31 @ r9=maxcells|(sh<<31)\r
228\r
229 mov r1, r7, lsl #4\r
230 orr r1, r1, #0x00ff\r
231\r
232 and r10, r7, #3\r
233 cmp r10, #1\r
234 biclt r1, r1, #0xfc00\r
235 biceq r1, r1, #0xfe00\r
236 bicgt r1, r1, #0xff00 @ r1=ymask=(height<<8)|0xff; ...; // Y Mask in pixels\r
237\r
238 add r10, r10, #5\r
239 cmp r10, #7\r
240 subge r10, r10, #1 @ r10=shift[width] (5,6,6,7)\r
241\r
242 @ calculate xmask:\r
243 mov r5, r8, lsl r10\r
244 sub r5, r5, #1 @ r5=xmask\r
245\r
246 @ Find name table:\r
247 tst r0, r0\r
248 ldreqb r12, [r11, #2]\r
249 ldrneb r12, [r11, #4]\r
250\r
251 ldr r2, =Scanline @ trying to make good use of pipeline here\r
252 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
253\r
254 moveq r12, r12, lsl #10\r
255 movne r12, r12, lsl #13\r
256 and r12, r12, #(7<<13) @ r12=(ts->nametab<<1) (halfword compliant)\r
257\r
258 ldrh r8, [r11, #12]\r
259 ldrb r7, [r11, #11]\r
260 ldr r2, [r2]\r
261\r
262 mov r4, r8, lsr #8 @ pvid->reg[13]\r
263 mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r
264 tst r7, #2\r
265 addne r4, r4, r2, lsl #2 @ htab+=Scanline<<1; // Offset by line\r
266 tst r7, #1\r
267 biceq r4, r4, #0x1f @ htab&=~0xf; // Offset by tile\r
268 add r4, r4, r0, lsl #1 @ htab+=plane\r
269 bic r4, r4, #0x00ff0000 @ just in case\r
270 ldrh r3, [lr, r4] @ r3=hscroll\r
271\r
272 tst r7, #4\r
273 bne .DrawStrip_vsscroll\r
274\r
275 @ Get vertical scroll value:\r
276 add r7, lr, #0x012000\r
277 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
278 ldr r7, [r7]\r
279\r
280 tst r8, #2\r
281 tstne r8, #4\r
282 bne .DrawStrip_interlace\r
283\r
284 tst r0, r0\r
285 movne r7, r7, lsr #16\r
286\r
287 @ Find the line in the name table\r
288 add r2, r2, r7\r
289 and r2, r2, r1\r
290 mov r4, r2, lsr #3\r
291 add r10, r10, #1 @ shift[width]++\r
292 add r12, r12, r4, lsl r10 @ nametab+=(ts.line>>3)<<shift[width];\r
293\r
294 @ ldmia r0, {r1,r2,r3,r5,r6,r9} @ r2=line, r3=ts->hscroll, r5=ts->xmask, r6=ts->hc, r9=ts->cells\r
cc68a136 295\r
296 and r10,r2, #7\r
297 mov r10,r10, lsl #1 @ r10=ty=(ts->line&7)<<1;\r
298 orr r10,r10, r9, lsl #24\r
299\r
300 rsb r8, r3, #0\r
301 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
302\r
303 sub r1, r3, #1\r
304 and r1, r1, #7\r
305 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
306\r
307 tst r9, #1<<31\r
308 mov r3, #0\r
309 orrne r10,r10, #1<<23 @ r10=(cells<<24|sh<<23|hi_not_empty<<22|ty)\r
310 movne r3, #0x40 @ default to shadowed pal on sh mode\r
311\r
312 mvn r9, #0 @ r9=prevcode=-1\r
313\r
314 cmp r7, #8\r
315 addne r10,r10, #0x01000000 @ we will loop cells+1 times if there is scroll\r
316\r
317 @ cache some stuff to avoid mem access\r
318 ldr r11,=HighCol\r
319 mov r0, #0xf\r
320 add r1, r11, r7 @ r1=pdest\r
321\r
322\r
323 @ r4 & r7 are scratch in this loop\r
324.dsloop_subr1:\r
325 sub r1, r1, #8\r
326.dsloop: @ 40-41 times\r
327 subs r10,r10, #0x01000000\r
328 bmi .dsloop_exit\r
329\r
330.dsloop_enter:\r
331 and r7, r5, r8\r
332 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
333 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
334\r
335 add r1, r1, #8\r
336 add r8, r8, #1\r
337\r
338 tst r7, #0x8000\r
339 bne .DrawStrip_hiprio\r
340\r
341 cmp r7, r9\r
342 beq .DrawStrip_samecode @ we know stuff about this tile already\r
343\r
344 mov r9, r7 @ remember code\r
345\r
346 movs r2, r9, lsl #20 @ if (code&0x1000)\r
347 mov r2, r2, lsl #1\r
348 add r2, r2, r10, lsl #17\r
349 mov r2, r2, lsr #17\r
350 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
351\r
352 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
353\r
354 bic r7, r3, #0x3f\r
355 and r3, r9, #0x6000\r
356 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
357\r
358.DrawStrip_samecode:\r
359 tst r2, r2\r
360 beq .dsloop @ tileline blank\r
361\r
362 cmp r2, r2, ror #4\r
363 beq .DrawStrip_SingleColor @ tileline singlecolor \r
364\r
365 tst r9, #0x0800\r
366 beq .DrawStrip_TileNorm\r
367\r
368 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
369 TileFlip r0\r
370 b .dsloop\r
371\r
372.DrawStrip_TileNorm:\r
373 TileNorm r0\r
374 b .dsloop\r
375\r
376.DrawStrip_SingleColor:\r
377 and r4, r2, #0xf\r
378 orr r4, r3, r4\r
379 orr r4, r4, r4, lsl #8\r
380 tst r1, #1 @ not aligned?\r
381 strneb r4, [r1], #1\r
382 streqh r4, [r1], #2\r
383 strh r4, [r1], #2\r
384 strh r4, [r1], #2\r
385 strh r4, [r1], #2\r
386 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
387 b .dsloop_subr1\r
388\r
389.DrawStrip_hiprio:\r
390 tst r10, #0x00c00000\r
391 beq .DrawStrip_hiprio_maybempt\r
392 sub r0, r1, r11\r
393 orr r7, r7, r0, lsl #16\r
394 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
395 tst r7, #0x1000\r
396 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
397 str r7, [r6], #4 @ cache hi priority tile\r
398 mov r0, #0xf\r
399 b .dsloop\r
400\r
401.DrawStrip_hiprio_maybempt:\r
402 cmp r7, r9\r
403 beq .dsloop @ must've been empty, otherwise we wouldn't get here\r
404 movs r2, r7, lsl #20 @ if (code&0x1000)\r
405 mov r2, r2, lsl #1\r
406 add r2, r2, r10, lsl #17\r
407 mov r2, r2, lsr #17\r
408 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
409 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
410 mov r9, r7 @ remember code\r
411 tst r2, r2\r
412 orrne r10, r10, #1<<22\r
413 bne .DrawStrip_hiprio\r
414 b .dsloop\r
415\r
416.dsloop_exit:\r
417 mov r0, #0\r
418 str r0, [r6] @ terminate the cache list\r
419\r
420 ldmfd sp!, {r4-r11,lr}\r
421 bx lr\r
422\r
6d7acf9e 423@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
cc68a136 424\r
425.DrawStrip_vsscroll:\r
6d7acf9e 426 rsb r8, r3, #0\r
427 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
428 bic r8, r8, #0xff000000\r
429 orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|tilex[15:0])\r
cc68a136 430\r
6d7acf9e 431 ldr r4, =Scanline\r
432 orr r5, r1, r10, lsl #24\r
433 ldr r4, [r4]\r
434 sub r1, r3, #1\r
435 orr r5, r5, r4, lsl #16 @ r5=(shift_width[31:24]|scanline[23:16]|ymask[15:0])\r
436 and r1, r1, #7\r
437 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
cc68a136 438\r
6d7acf9e 439 mov r10,r9, lsl #16\r
440 tst r0, r0\r
441 orrne r10,r10, #0x8000\r
442 tst r9, #1<<31\r
443 mov r3, #0\r
444 orr r10,r10, #0xff000000 @ will be adjusted on entering loop\r
445 orrne r10,r10, #1<<23 @ r10=(cells[31:24]|sh[23]|hi_not_empty[22]|cells_max[21:16]|plane[15]|ty[14:0])\r
446 movne r3, #0x40 @ default to shadowed pal on sh mode\r
cc68a136 447\r
6d7acf9e 448 mvn r9, #0 @ r9=prevcode=-1\r
cc68a136 449\r
6d7acf9e 450 @ cache some stuff to avoid mem access\r
451 ldr r11,=HighCol\r
452 mov r0, #0xf\r
453 add r1, r11, r7 @ r1=pdest\r
454\r
455 cmp r7, #8\r
456 subne r10,r10, #0x01000000 @ have hscroll, start with negative cell\r
457\r
458\r
459 @ r4 & r7 are scratch in this loop\r
460.dsloop_vs_subr1:\r
461 sub r1, r1, #8\r
462.dsloop_vs: @ 40-41 times\r
463 add r10,r10, #0x01000000\r
464 and r4, r10, #0x003f0000\r
465 cmp r4, r10, asr #8\r
466 ble .dsloop_exit\r
467\r
468 @ calc offset and read tileline code to r7, also calc ty\r
469 add r7, lr, #0x012000\r
470 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
471 add r7, r7, r10,asr #23 @ vsram + ((cell&~1)<<1)\r
472 bic r7, r7, #3\r
473 tst r10,#0x8000 @ plane1?\r
474 addne r7, r7, #2\r
475 ldrh r7, [r7] @ r7=vscroll\r
476\r
477 bic r10,r10,#0xff @ clear old ty\r
478 and r4, r5, #0xff0000\r
479 add r4, r4, r7, lsl #16\r
480 and r4, r4, r5, lsl #16 @ r4=line<<16\r
481 and r7, r4, #0x70000\r
482 orr r10,r10,r7, lsr #15 @ new ty\r
483\r
484 mov r4, r4, lsr #19\r
485 mov r7, r5, lsr #24\r
486 mov r4, r4, lsl r7 @ nametabadd\r
487\r
488 and r7, r8, r8, lsr #25\r
489 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
490 add r7, r7, r4, lsl #1\r
491 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
492\r
493 add r1, r1, #8\r
494 add r8, r8, #1\r
495\r
496 tst r7, #0x8000\r
497 bne .DrawStrip_vs_hiprio\r
498\r
499 cmp r7, r9\r
500 beq .DrawStrip_vs_samecode @ we know stuff about this tile already\r
501\r
502 mov r9, r7 @ remember code\r
503\r
504 movs r2, r9, lsl #20 @ if (code&0x1000)\r
505 mov r2, r2, lsl #1\r
506 add r2, r2, r10, lsl #17\r
507 mov r2, r2, lsr #17\r
508 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
509\r
510 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
511\r
512 bic r7, r3, #0x3f\r
513 and r3, r9, #0x6000\r
514 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
515\r
516.DrawStrip_vs_samecode:\r
517 tst r2, r2\r
518 beq .dsloop_vs @ tileline blank\r
519\r
520 cmp r2, r2, ror #4\r
521 beq .DrawStrip_vs_SingleColor @ tileline singlecolor \r
522\r
523 tst r9, #0x0800\r
524 beq .DrawStrip_vs_TileNorm\r
525\r
526 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
527 TileFlip r0\r
528 b .dsloop_vs\r
529\r
530.DrawStrip_vs_TileNorm:\r
531 TileNorm r0\r
532 b .dsloop_vs\r
533\r
534.DrawStrip_vs_SingleColor:\r
535 and r4, r2, #0xf\r
536 orr r4, r3, r4\r
537 orr r4, r4, r4, lsl #8\r
538 tst r1, #1 @ not aligned?\r
539 strneb r4, [r1], #1\r
540 streqh r4, [r1], #2\r
541 strh r4, [r1], #2\r
542 strh r4, [r1], #2\r
543 strh r4, [r1], #2\r
544 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
545 b .dsloop_vs_subr1\r
546\r
547.DrawStrip_vs_hiprio:\r
548 tst r10, #0x00c00000\r
549 beq .DrawStrip_vs_hiprio_maybempt\r
550 sub r0, r1, r11\r
551 orr r7, r7, r0, lsl #16\r
552 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
553 tst r7, #0x1000\r
554 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
555 str r7, [r6], #4 @ cache hi priority tile\r
556 mov r0, #0xf\r
557 b .dsloop_vs\r
558\r
559.DrawStrip_vs_hiprio_maybempt:\r
560 cmp r7, r9\r
561 beq .dsloop_vs @ must've been empty, otherwise we wouldn't get here\r
562 movs r2, r7, lsl #20 @ if (code&0x1000)\r
563 mov r2, r2, lsl #1\r
564 add r2, r2, r10, lsl #17\r
565 mov r2, r2, lsr #17\r
566 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
567 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
568 mov r9, r7 @ remember code\r
569 tst r2, r2\r
570 orrne r10, r10, #1<<22\r
571 bne .DrawStrip_vs_hiprio\r
572 b .dsloop_vs\r
573\r
574\r
575@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
cc68a136 576\r
577@ interlace mode 2? Sonic 2?\r
578.DrawStrip_interlace:\r
579 tst r0, r0\r
580 moveq r7, r7, lsl #21\r
581 movne r7, r7, lsl #5\r
582\r
583 @ Find the line in the name table\r
584 add r2, r7, r2, lsl #22 @ r2=(vscroll+(Scanline<<1))<<21 (11 bits);\r
585 orr r1, r1, #0x80000000\r
586 and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
587 mov r2, r2, lsr #21\r
588 mov r4, r2, lsr #4\r
589 mov r12, r12, lsr #1 @ halfwords\r
590 add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];\r
591 and r9, r9, #0xff\r
592\r
593 sub sp, sp, #6*4\r
594 stmia sp, {r0,r2,r3,r5,r6,r9}\r
595\r
596 mov r0, sp\r
597 bl DrawStripInterlace @ struct TileStrip *ts\r
598\r
599 add sp, sp, #6*4\r
600 ldmfd sp!, {r4-r11,lr}\r
601 bx lr\r
602\r
603.pool\r
604\r
605@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
606\r
607\r
608.global BackFill @ int reg7, int sh\r
609\r
610BackFill:\r
611 stmfd sp!, {r4-r9,lr}\r
612\r
613 ldr lr, =(HighCol+8)\r
614\r
615 mov r0, r0, lsl #26\r
616 mov r0, r0, lsr #26\r
617 orr r0, r0, r1, lsl #6\r
618 orr r0, r0, r0, lsl #8\r
619 orr r0, r0, r0, lsl #16\r
620\r
621 mov r1, r0\r
622 mov r2, r0\r
623 mov r3, r0\r
624 mov r4, r0\r
625 mov r5, r0\r
626 mov r6, r0\r
627 mov r7, r0\r
628\r
629 @ go go go!\r
630 stmia lr!, {r0-r7} @ 10*8*4\r
631 stmia lr!, {r0-r7}\r
632 stmia lr!, {r0-r7}\r
633 stmia lr!, {r0-r7}\r
634 stmia lr!, {r0-r7}\r
635 stmia lr!, {r0-r7}\r
636 stmia lr!, {r0-r7}\r
637 stmia lr!, {r0-r7}\r
638 stmia lr!, {r0-r7}\r
639 stmia lr!, {r0-r7}\r
640\r
641 ldmfd sp!, {r4-r9,r12}\r
642 bx r12\r
643\r
644\r
645@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
646\r
647\r
648.global DrawTilesFromCache @ int *hc, int sh\r
649\r
650DrawTilesFromCache:\r
651 stmfd sp!, {r4-r8,r11,lr}\r
652\r
653 mvn r5, #0 @ r5=prevcode=-1\r
654 mov r8, r1\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 r12,#0xf\r
660\r
661 @ scratch: r4, r7\r
662.dtfc_loop:\r
663 ldr r6, [r0], #4 @ read code\r
664 movs r1, r6, lsr #16 @ r1=dx;\r
665 ldmeqfd sp!, {r4-r8,r11,pc} @ dx is never zero, this must be a terminator, return\r
666 bic r1, r1, #0xfe00\r
667 add r1, r11, r1 @ r1=pdest\r
668\r
669 mov r7, r6, lsl #16\r
670 cmp r5, r7, lsr #16\r
671 beq .dtfc_samecode @ if (code==prevcode)\r
672\r
673 mov r5, r7, lsr #16\r
674\r
675 mov r2, r5, lsl #21\r
676 mov r2, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4;\r
677 add r2, r2, r6, lsr #25 @ addr+=ty\r
678\r
679 and r3, r5, #0x6000\r
680 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
681\r
682 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
683\r
684.dtfc_samecode:\r
685 tst r8, r8\r
686 bne .dtfc_shadow\r
687\r
688 tst r2, r2\r
689 beq .dtfc_loop\r
690\r
691 cmp r2, r2, ror #4\r
692 beq .dtfc_SingleColor @ tileline singlecolor \r
693\r
694 tst r5, #0x0800\r
695 beq .dtfc_TileNorm\r
696\r
697 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
698 TileFlip r12\r
699 b .dtfc_loop\r
700\r
701.dtfc_TileNorm:\r
702 TileNorm r12\r
703 b .dtfc_loop\r
704\r
705.dtfc_SingleColor:\r
706 and r4, r2, #0xf\r
707 orr r4, r3, r4\r
708 orr r4, r4, r4, lsl #8\r
709 tst r1, #1 @ not aligned?\r
710 strneb r4, [r1], #1\r
711 streqh r4, [r1], #2\r
712 strh r4, [r1], #2\r
713 strh r4, [r1], #2\r
714 strh r4, [r1], #2\r
715 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
716 b .dtfc_loop\r
717\r
718.dtfc_shadow:\r
719 tst r2, r2\r
720 beq .dtfc_shadow_blank\r
721\r
722 cmp r2, r2, ror #4\r
723 beq .dtfc_SingleColor @ tileline singlecolor \r
724\r
725 tst r5, #0x0800\r
726 beq .dtfc_TileNormShHP\r
727\r
728 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
729 TileFlipShHP\r
730 b .dtfc_loop\r
731\r
732.dtfc_TileNormShHP:\r
733 TileNormShHP\r
734 b .dtfc_loop\r
735\r
736.dtfc_shadow_blank:\r
737 ldrb r4, [r1] @ 1ci\r
738 ldrb r12,[r1,#1]\r
739 tst r4, #0x80\r
740 andeq r4, r4,#0x3f\r
741 streqb r4, [r1]\r
742 tst r12,#0x80\r
743 ldrb r4, [r1,#2]\r
744 andeq r12,r12,#0x3f\r
745 streqb r12,[r1,#1]\r
746 tst r4, #0x80\r
747 ldrb r12,[r1,#3]\r
748 andeq r4, r4,#0x3f\r
749 streqb r4, [r1,#2]\r
750 tst r12,#0x80\r
751 ldrb r4, [r1,#4]\r
752 andeq r12,r12,#0x3f\r
753 streqb r12,[r1,#3]\r
754 tst r4, #0x80\r
755 ldrb r12,[r1,#5]\r
756 andeq r4, r4,#0x3f\r
757 streqb r4, [r1,#4]\r
758 tst r12,#0x80\r
759 ldrb r4, [r1,#6]\r
760 andeq r12,r12,#0x3f\r
761 streqb r12,[r1,#5]\r
762 tst r4, #0x80\r
763 ldrb r12,[r1,#7]\r
764 andeq r4, r4,#0x3f\r
765 streqb r4, [r1,#6]\r
766 tst r12,#0x80\r
767 andeq r12,r12,#0x3f\r
768 streqb r12,[r1,#7]\r
769 mov r12, #0xf\r
770 b .dtfc_loop\r
771\r
772.pool\r
773\r
774@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
775\r
776\r
777.global DrawSpritesFromCache @ int *hc, int sh\r
778\r
779DrawSpritesFromCache:\r
780 stmfd sp!, {r4-r11,lr}\r
781\r
782 @ cache some stuff to avoid mem access\r
783 ldr r11,=HighCol\r
784 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
785 mov r6, r1, lsl #31\r
786 orr r6, r6, #1<<30\r
787 mov r12,#0xf\r
788\r
789 mov r10, r0\r
790\r
791.dsfc_loop:\r
792 ldr r9, [r10], #4 @ read code\r
793 bic r6, r6, #7 @ using pipeline\r
794 tst r9, r9\r
795 ldmeqfd sp!, {r4-r11,pc}\r
796\r
797 mov r4, r9, lsl #28\r
798 orr r6, r6, r4, lsr #30\r
799 add r6, r6, #1 @ r6=s1cc???? ... ?????www (s=shadow/hilight, cc=pal, w=width)\r
800\r
801 and r5, r9, #3\r
802 add r5, r5, #1 @ r5=delta\r
803 tst r9, #0x10000\r
804 rsbne r5, r5, #0 @ Flip X\r
805 mov r5, r5, lsl #4\r
806\r
807 mov r2, r9, lsr #17\r
808 mov r8, r2, lsl #1 @ tile=((unsigned int)code>>17)<<1;\r
809\r
810 and r3, r9, #0x30 @ r3=pal=(code&0x30);\r
811\r
812 bic r6, r6, #3<<28\r
813 orr r6, r6, r3, lsl #24\r
814\r
815 mov r0, r9, lsl #16\r
816 mov r0, r0, asr #22 @ sx=(code<<16)>>22\r
817 adds r0, r0, #0 @ set ZV\r
818 b .dsfc_inloop_enter\r
819\r
820@ scratch: r4, r7\r
821.dsfc_inloop:\r
822 sub r6, r6, #1\r
823 tst r6, #7\r
824 beq .dsfc_loop\r
825 adds r0, r0, #8\r
826 add r8, r8, r5\r
827\r
828.dsfc_inloop_enter:\r
829 ble .dsfc_inloop\r
830 cmp r0, #328\r
831 bge .dsfc_loop\r
832\r
833 mov r8, r8, lsl #17\r
834 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
835\r
836 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+tile); // Get 8 pixels\r
837 add r1, r11, r0 @ r1=pdest\r
838 tst r2, r2\r
839 beq .dsfc_inloop\r
840\r
841 cmp r12, r6, lsr #28\r
842 beq .dsfc_shadow\r
843\r
844 cmp r2, r2, ror #4\r
845 beq .dsfc_SingleColor @ tileline singlecolor \r
846\r
847 tst r9, #0x10000\r
848 beq .dsfc_TileNorm\r
849\r
850 @ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
851 TileFlip r12\r
852 b .dsfc_inloop\r
853\r
854.dsfc_TileNorm:\r
855 TileNorm r12\r
856 b .dsfc_inloop\r
857\r
858.dsfc_SingleColor:\r
859 tst r0, #1 @ not aligned?\r
860 and r4, r2, #0xf\r
861 orr r4, r3, r4\r
862 orr r4, r4, r4, lsl #8\r
863 strneb r4, [r1], #1\r
864 streqh r4, [r1], #2\r
865 strh r4, [r1], #2\r
866 strh r4, [r1], #2\r
867 strh r4, [r1], #2\r
868 strneb r4, [r1], #1\r
869 b .dsfc_inloop\r
870\r
871.dsfc_shadow:\r
872 cmp r2, r2, ror #4\r
873 beq .dsfc_singlec_sh\r
874\r
875 tst r9, #0x10000\r
876 beq .dsfc_TileNorm_sh\r
877\r
878 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
879 TileFlipSh\r
880 b .dsfc_inloop\r
881\r
882.dsfc_TileNorm_sh:\r
883 TileNormSh\r
884 b .dsfc_inloop\r
885\r
886.dsfc_singlec_sh:\r
887 cmp r2, #0xe0000000\r
888 bcc .dsfc_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
889 tst r2, #0x10000000\r
890 bne .dsfc_sh_sh\r
891 TileSingleHi\r
892 b .dsfc_inloop\r
893\r
894.dsfc_sh_sh:\r
895 TileSingleSh\r
896 b .dsfc_inloop\r
897\r
898.pool\r
899\r
900@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
901\r
902@ + 0 : hhhhvvvv ab--hhvv yyyyyyyy yyyyyyyy // a: offscreen h, b: offs. v, h: horiz. size\r
903@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
904\r
905.global DrawSprite @ unsigned int *sprite, int **hc, int sh\r
906\r
907DrawSprite:\r
908 stmfd sp!, {r4-r9,r11,lr}\r
909\r
910 ldr r3, [r0] @ sprite[0]\r
911 ldr r7, =Scanline\r
912 mov r6, r3, lsr #28\r
913 sub r6, r6, #1 @ r6=width-1 (inc later)\r
914 mov r5, r3, lsr #24\r
915 and r5, r5, #7 @ r5=height\r
916\r
917 mov r4, r3, lsl #16 @ r4=sy<<16 (tmp)\r
918\r
919 ldr r7, [r7]\r
920 ldr r9, [r0, #4]\r
921 sub r7, r7, r4, asr #16 @ r7=row=Scanline-sy\r
922\r
923 tst r2, r2\r
924 mov r2, r9, asr #16 @ r2=sx\r
925 bic r9, r9, #0xfe000000\r
926 orrne r9, r9, #1<<31 @ r9=code|(sh<<31)\r
927\r
928 tst r9, #0x1000\r
929 movne r4, r5, lsl #3\r
930 subne r4, r4, #1\r
931 subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
932\r
933 mov r8, r9, lsl #21\r
934 mov r8, r8, lsr #21\r
935 add r8, r8, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
936 \r
937 tst r9, #0x0800\r
938 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
939 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
940\r
941 mov r8, r8, lsl #4\r
942 and r7, r7, #7\r
943 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
944\r
945 tst r9, #0x8000\r
946 bne .dspr_cache @ if(code&0x8000) // high priority - cache it\r
947\r
948 @ cache some stuff to avoid mem access\r
949 ldr r11,=HighCol\r
950 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
951 mov r12,#0xf\r
952\r
953 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
954 and r4, r9, #0x6000\r
955 orr r9, r9, r4, lsl #16\r
956 orr r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
957\r
958 tst r9, #1<<31\r
959 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
960 orrne r3, r3, #0x40 @ shadow by default\r
961\r
962 add r6, r6, #1 @ inc now\r
963 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
964 b .dspr_loop_enter\r
965\r
966.dspr_loop:\r
967 subs r6, r6, #1 @ width--\r
968 ldmeqfd sp!, {r4-r9,r11,pc}@ return\r
969 adds r0, r0, #8 @ sx+=8\r
970 add r8, r8, r5 @ tile+=delta\r
971\r
972.dspr_loop_enter:\r
973 ble .dspr_loop @ sx <= 0\r
974 cmp r0, #328\r
975 ldmgefd sp!, {r4-r9,r11,pc}@ return\r
976\r
977 mov r8, r8, lsl #17\r
978 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
979\r
980 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
981 add r1, r11, r0 @ r1=pdest\r
982 tst r2, r2\r
983 beq .dspr_loop\r
984\r
985 cmp r12, r9, lsr #28\r
986 beq .dspr_shadow\r
987\r
988 cmp r2, r2, ror #4\r
989 beq .dspr_SingleColor @ tileline singlecolor \r
990\r
991 tst r9, #0x0800\r
992 beq .dspr_TileNorm\r
993\r
994 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
995 TileFlip r12\r
996 b .dspr_loop\r
997\r
998@ scratch: r4, r7\r
999.dspr_TileNorm:\r
1000 TileNorm r12\r
1001 b .dspr_loop\r
1002\r
1003.dspr_SingleColor:\r
1004 and r4, r2, #0xf\r
1005 orr r4, r3, r4\r
1006 orr r4, r4, r4, lsl #8\r
1007 tst r0, #1 @ not aligned?\r
1008 strneb r4, [r1], #1\r
1009 streqh r4, [r1], #2\r
1010 strh r4, [r1], #2\r
1011 strh r4, [r1], #2\r
1012 strh r4, [r1], #2\r
1013 strneb r4, [r1], #1\r
1014 b .dspr_loop\r
1015\r
1016.dspr_shadow:\r
1017 cmp r2, r2, ror #4\r
1018 beq .dspr_singlec_sh\r
1019\r
1020 tst r9, #0x0800\r
1021 beq .dspr_TileNorm_sh\r
1022\r
1023 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1024 TileFlipSh\r
1025 b .dspr_loop\r
1026\r
1027.dspr_TileNorm_sh:\r
1028 TileNormSh\r
1029 b .dspr_loop\r
1030\r
1031.dspr_singlec_sh:\r
1032 cmp r2, #0xe0000000\r
1033 bcc .dspr_SingleColor @ normal tileline\r
1034 tst r2, #0x10000000\r
1035 bne .dspr_sh_sh\r
1036 TileSingleHi\r
1037 b .dspr_loop\r
1038\r
1039.dspr_sh_sh:\r
1040 TileSingleSh\r
1041 b .dspr_loop\r
1042\r
1043\r
1044.dspr_cache:\r
1045 @ *(*hc)++ = (tile<<16)|((code&0x0800)<<5)|((sx<<6)&0x0000ffc0)|((code>>9)&0x30)|((sprite[0]>>24)&0xf);\r
1046 mov r4, r8, lsl #16 @ tile\r
1047 tst r9, #0x0800\r
1048 orrne r4, r4, #0x10000 @ code&0x0800\r
1049 mov r2, r2, lsl #22\r
1050 orr r4, r4, r2, lsr #16 @ (sx<<6)&0x0000ffc0\r
1051 and r2, r9, #0x6000\r
1052 orr r4, r4, r2, lsr #9 @ (code>>9)&0x30\r
1053 mov r3, r3, lsl #12\r
1054 ldr r2, [r1]\r
1055 orr r4, r4, r3, lsr #28 @ (sprite[0]>>24)&0xf\r
1056\r
1057 str r4, [r2], #4\r
1058 str r2, [r1]\r
1059\r
1060 ldmfd sp!, {r4-r9,r11,lr}\r
1061 bx lr\r
1062\r
1063@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1064\r
1065.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache\r
1066\r
1067DrawWindow:\r
1068 stmfd sp!, {r4-r11,lr}\r
1069\r
1070 ldr r11, =(Pico+0x22228) @ Pico.video\r
1071 ldr r10, =Scanline\r
1072 ldrb r12, [r11, #3] @ pvid->reg[3]\r
1073\r
1074 ldr r10, [r10]\r
1075 ldr r4, [r11, #12]\r
1076 mov r5, r10, lsr #3\r
1077 and r10, r10, #7\r
1078 mov r10, r10, lsl #1 @ r10=ty\r
1079\r
1080 mov r12, r12, lsl #10\r
1081\r
1082 tst r4, #1 @ 40 cell mode?\r
1083 andne r12, r12, #0xf000 @ 0x3c<<10\r
1084 andeq r12, r12, #0xf800\r
1085 addne r12, r12, r5, lsl #7\r
1086 addeq r12, r12, r5, lsl #6 @ nametab\r
1087 add r12, r12, r0, lsl #2 @ +starttile\r
1088\r
1089 ldr r6, =rendstatus\r
1090 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
1091 ldrb r6, [r6]\r
1092\r
1093 @ fetch the first code now\r
1094 ldrh r7, [lr, r12]\r
1095\r
1096 ands r6, r6, #2 @ we care about bit 1 only\r
1097 orr r6, r6, r2\r
1098 bne .dw_no_sameprio\r
1099\r
1100 cmp r2, r7, lsr #15\r
1101 ldmnefd sp!, {r4-r11,pc} @ assume that whole window uses same priority\r
1102\r
1103.dw_no_sameprio:\r
1104 orr r6, r6, r3, lsl #8 @ shadow mode\r
1105\r
1106 sub r8, r1, r0\r
1107 mov r8, r8, lsl #1 @ cells\r
1108\r
1109 mvn r9, #0 @ r9=prevcode=-1\r
1110\r
1111 @ cache some stuff to avoid mem access\r
1112 ldr r11,=(HighCol+8)\r
1113 add r1, r11, r0, lsl #4 @ r1=pdest\r
1114 mov r0, #0xf\r
1115 b .dwloop_enter\r
1116\r
1117 @ r4,r5 & r7 are scratch in this loop\r
1118.dwloop:\r
1119 add r1, r1, #8\r
1120.dwloop_nor1:\r
1121 add r12, r12, #2 @ halfwords\r
1122 ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)\r
1123 subs r8, r8, #1\r
1124 beq .dwloop_end @ done\r
1125\r
1126 eor r5, r6, r7, lsr #15\r
1127 tst r5, #1\r
1128 orrne r6, r6, #2 @ wrong pri\r
1129 bne .dwloop\r
1130\r
1131 cmp r7, r9\r
1132 beq .dw_samecode @ we know stuff about this tile already\r
1133\r
1134.dwloop_enter:\r
1135 mov r9, r7 @ remember code\r
1136\r
1137 movs r2, r9, lsl #20 @ if (code&0x1000)\r
1138 mov r2, r2, lsl #1\r
1139 add r2, r10, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4; addr+=ty\r
1140 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
1141\r
1142 and r3, r9, #0x6000\r
1143 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
1144\r
1145 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1146\r
1147.dw_samecode:\r
1148 tst r6, #0x100\r
1149 bne .dw_shadow\r
1150.dw_shadow_done:\r
1151 tst r2, r2\r
1152 beq .dwloop @ tileline blank\r
1153\r
1154 cmp r2, r2, ror #4\r
1155 beq .dw_SingleColor @ tileline singlecolor \r
1156\r
1157 tst r9, #0x0800\r
1158 beq .dw_TileNorm\r
1159\r
1160 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
1161 TileFlip r0\r
1162 b .dwloop\r
1163\r
1164.dw_TileNorm:\r
1165 TileNorm r0\r
1166 b .dwloop\r
1167\r
1168.dw_SingleColor:\r
1169 and r4, r0, r2 @ #0x0000000f\r
1170 orr r4, r3, r4\r
1171 orr r4, r4, r4, lsl #8\r
1172 orr r4, r4, r4, lsl #16\r
1173 mov r5, r4\r
1174 stmia r1!, {r4,r5}\r
1175 b .dwloop_nor1 @ we incremeted r1 ourselves\r
1176\r
1177.dw_shadow:\r
1178 tst r6, #1 @ hi pri?\r
1179 orreq r3, r3, #0x40\r
1180 beq .dw_shadow_done\r
1181 ldr r4, [r1]\r
1182 tst r4, #0x00000080\r
1183 biceq r4, r4, #0x000000c0\r
1184 tst r4, #0x00008000\r
1185 biceq r4, r4, #0x0000c000\r
1186 tst r4, #0x00800000\r
1187 biceq r4, r4, #0x00c00000\r
1188 tst r4, #0x80000000\r
1189 biceq r4, r4, #0xc0000000\r
1190 str r4, [r1]\r
1191 ldr r4, [r1,#4]\r
1192 tst r4, #0x00000080\r
1193 biceq r4, r4, #0x000000c0\r
1194 tst r4, #0x00008000\r
1195 biceq r4, r4, #0x0000c000\r
1196 tst r4, #0x00800000\r
1197 biceq r4, r4, #0x00c00000\r
1198 tst r4, #0x80000000\r
1199 biceq r4, r4, #0xc0000000\r
1200 str r4, [r1,#4]\r
1201 b .dw_shadow_done\r
1202\r
1203.dwloop_end:\r
1204 ldr r0, =rendstatus\r
1205 ldr r1, [r0]\r
1206 and r6, r6, #2\r
1207 orr r1, r1, r6\r
1208 str r1, [r0]\r
1209\r
1210 ldmfd sp!, {r4-r11,r12}\r
1211 bx r12\r
1212\r
1213\r
1214@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1215\r
1216\r
1217@ hilights 2 pixels in RGB444/BGR444 format\r
1218.macro TileDoShHi2Pixels444 reg\r
1219 mov \reg, \reg, ror #12\r
1220 adds \reg, \reg, #0x40000000\r
1221 orrcs \reg, \reg, #0xf0000000\r
1222 mov \reg, \reg, ror #28\r
1223 adds \reg, \reg, #0x40000000\r
1224 orrcs \reg, \reg, #0xf0000000\r
1225 mov \reg, \reg, ror #28\r
1226 adds \reg, \reg, #0x40000000\r
1227 orrcs \reg, \reg, #0xf0000000\r
1228 mov \reg, \reg, ror #24\r
1229 adds \reg, \reg, #0x40000000\r
1230 orrcs \reg, \reg, #0xf0000000\r
1231 mov \reg, \reg, ror #28\r
1232 adds \reg, \reg, #0x40000000\r
1233 orrcs \reg, \reg, #0xf0000000\r
1234 mov \reg, \reg, ror #28\r
1235 adds \reg, \reg, #0x40000000\r
1236 orrcs \reg, \reg, #0xf0000000\r
1237 mov \reg, \reg, ror #12\r
1238.endm\r
1239\r
1240\r
1241.global FinalizeLineBGR444 @ int sh\r
1242\r
1243FinalizeLineBGR444:\r
1244 stmfd sp!, {r4-r6,lr}\r
1245 mov r6, r0\r
1246 ldr lr, =(Pico+0x22228) @ Pico.video\r
1247 ldr r0, =DrawLineDest\r
1248 ldrb r12, [lr, #12]\r
1249 ldr r0, [r0]\r
1250 sub r3, lr, #0x128 @ r3=Pico.cram\r
1251\r
1252 tst r12, #1\r
1253 movne r2, #320/4 @ len\r
1254 bne .fl_no32colBGR444\r
1255 ldr r4, =PicoOpt\r
1256 mov r2, #256/4\r
1257 ldr r4, [r4]\r
1258 tst r4, #0x100\r
1259 addeq r0, r0, #32*2\r
1260\r
1261.fl_no32colBGR444:\r
1262 tst r6, r6\r
1263 beq .fl_noshBGR444\r
1264\r
1265 ldr r4, =HighPal\r
1266\r
1267 ldrb r12, [lr, #-0x1a] @ 0x2220e ~ dirtyPal\r
1268 tst r12, r12\r
1269 moveq r3, r4\r
1270 beq .fl_noshBGR444\r
1271 mov r12, #0\r
1272 strb r12, [lr, #-0x1a]\r
1273\r
1274 mov lr, #0x40/8\r
1275 @ copy pal:\r
1276.fl_loopcpBGR444:\r
1277 ldmia r3!, {r1,r5,r6,r12}\r
1278 subs lr, lr, #1\r
1279 stmia r4!, {r1,r5,r6,r12}\r
1280 bne .fl_loopcpBGR444\r
1281\r
1282 @ shadowed pixels:\r
1283 mov r12, #0x0077\r
1284 orr r12,r12,#0x0700\r
1285 orr r12,r12,r12,lsl #16\r
1286 sub r3, r3, #0x40*2\r
1287 add r5, r4, #0x80*2\r
1288 mov lr, #0x40/4\r
1289.fl_loopcpBGR444_sh:\r
1290 ldmia r3!, {r1,r6}\r
1291 subs lr, lr, #1\r
1292 and r1, r12, r1, lsr #1\r
1293 and r6, r12, r6, lsr #1\r
1294 stmia r4!, {r1,r6}\r
1295 stmia r5!, {r1,r6}\r
1296 bne .fl_loopcpBGR444_sh\r
1297\r
1298 @ hilighted pixels:\r
1299 sub r3, r3, #0x40*2\r
1300 mov lr, #0x40/2\r
1301.fl_loopcpBGR444_hi:\r
1302 ldr r1, [r3], #4\r
1303 TileDoShHi2Pixels444 r1\r
1304 str r1, [r4], #4\r
1305 subs lr, lr, #1\r
1306 bne .fl_loopcpBGR444_hi\r
1307\r
1308 sub r3, r4, #0x40*3*2\r
1309\r
1310\r
1311.fl_noshBGR444:\r
1312 ldr r1, =(HighCol+8)\r
1313 mov lr, #0xff\r
1314 mov lr, lr, lsl #1\r
1315\r
1316.fl_loopBGR444:\r
1317\r
1318 ldr r12, [r1], #4\r
1319 subs r2, r2, #1\r
1320\r
1321 and r4, lr, r12, lsl #1\r
1322 ldrh r4, [r3, r4]\r
1323 and r5, lr, r12, lsr #7\r
1324 ldrh r5, [r3, r5]\r
1325 and r6, lr, r12, lsr #15\r
1326 ldrh r6, [r3, r6]\r
1327 orr r4, r4, r5, lsl #16\r
1328\r
1329 and r5, lr, r12, lsr #23\r
1330 ldrh r5, [r3, r5] @ 2c.i.\r
1331 orr r5, r6, r5, lsl #16\r
1332\r
1333 stmia r0!, {r4,r5}\r
1334 bne .fl_loopBGR444\r
1335\r
1336\r
1337 ldmfd sp!, {r4-r6,lr}\r
1338 bx lr\r
1339\r
1340\r
1341@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1342\r
1343\r
1344@ hilights 2 pixels in RGB555/BGR555 format\r
1345.macro TileDoShHi2Pixels555 reg\r
1346 adds \reg, \reg, #0x40000000\r
1347 orrcs \reg, \reg, #0xf0000000\r
1348 mov \reg, \reg, ror #27\r
1349 adds \reg, \reg, #0x40000000\r
1350 orrcs \reg, \reg, #0xf0000000\r
1351 mov \reg, \reg, ror #26\r
1352 adds \reg, \reg, #0x40000000\r
1353 orrcs \reg, \reg, #0xf0000000\r
1354 mov \reg, \reg, ror #27\r
1355 adds \reg, \reg, #0x40000000\r
1356 orrcs \reg, \reg, #0xf0000000\r
1357 mov \reg, \reg, ror #27\r
1358 adds \reg, \reg, #0x40000000\r
1359 orrcs \reg, \reg, #0xf0000000\r
1360 mov \reg, \reg, ror #26\r
1361 adds \reg, \reg, #0x40000000\r
1362 orrcs \reg, \reg, #0xf0000000\r
1363 mov \reg, \reg, ror #27\r
1364.endm\r
1365\r
1366\r
1367@ Convert 0000bbb0 ggg0rrr0\r
1368@ to rrrrrggg gggbbbbb\r
1369\r
1370@ r2,r3,r9 - scratch, lr = 0x001c001c, r8 = 0x00030003\r
1371.macro convRGB565 reg\r
1372 and r2, lr, \reg,lsl #1\r
1373 and r9, r8, \reg,lsr #2\r
1374 orr r2, r2, r9 @ r2=red\r
1375 and r3, lr, \reg,lsr #7\r
1376 and r9, r8, \reg,lsr #10\r
1377 orr r3, r3, r9 @ r3=blue\r
1378 and \reg, \reg, lr, lsl #3\r
1379 orr \reg, \reg, \reg,lsl #3 @ green\r
1380 orr \reg, \reg, r2, lsl #11 @ add red back\r
1381 orr \reg, \reg, r3 @ add blue back\r
1382.endm\r
1383\r
1384vidConvCpyRGB565: @ void *to, void *from, int pixels\r
1385 stmfd sp!, {r4-r9,lr}\r
1386\r
1387 mov r12, r2, lsr #3 @ repeats\r
1388 mov lr, #0x001c0000\r
1389 orr lr, lr, #0x01c @ lr == pattern 0x001c001c\r
1390 mov r8, #0x00030000\r
1391 orr r8, r8, #0x003 @ lr == pattern 0x001c001c\r
1392\r
1393.loopRGB565:\r
1394 ldmia r1!, {r4-r7}\r
1395 subs r12, r12, #1\r
1396 convRGB565 r4\r
1397 str r4, [r0], #4\r
1398 convRGB565 r5\r
1399 str r5, [r0], #4\r
1400 convRGB565 r6\r
1401 str r6, [r0], #4\r
1402 convRGB565 r7\r
1403 str r7, [r0], #4\r
1404\r
1405 bgt .loopRGB565\r
1406\r
1407 ldmfd sp!, {r4-r9,lr}\r
1408 bx lr\r
1409\r
1410\r
1411\r
1412.global FinalizeLineRGB555 @ int sh\r
1413\r
1414FinalizeLineRGB555:\r
1415 stmfd sp!, {r4-r8,lr}\r
3d48f143 1416 ldr r8, =(Pico+0x22228) @ Pico.video\r
cc68a136 1417 ldr r4, =HighPal\r
1418\r
3d48f143 1419 ldrb r7, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
cc68a136 1420 mov r6, r0\r
1421 mov r1, #0\r
1422 tst r7, r7\r
1423 beq .fl_noconvRGB555\r
3d48f143 1424 strb r1, [r8, #-0x1a]\r
1425 sub r1, r8, #0x128 @ r1=Pico.cram\r
cc68a136 1426 mov r0, r4\r
1427 mov r2, #0x40\r
1428 bl vidConvCpyRGB565\r
1429\r
1430.fl_noconvRGB555:\r
cc68a136 1431 mov r3, r4\r
1432 tst r6, r6\r
1433 beq .fl_noshRGB555\r
1434 tst r7, r7\r
1435 beq .fl_noshRGB555\r
1436\r
1437 @ shadowed pixels:\r
1438 mov r12, #0x008e\r
1439 orr r12,r12,#0x7300\r
1440 orr r12,r12,r12,lsl #16\r
1441 add r4, r3, #0x40*2\r
1442 add r5, r3, #0xc0*2\r
1443 mov lr, #0x40/4\r
1444.fl_loopcpRGB555_sh:\r
1445 ldmia r3!, {r1,r6}\r
1446 subs lr, lr, #1\r
1447 and r1, r12, r1, lsr #1\r
1448 and r6, r12, r6, lsr #1\r
1449 stmia r4!, {r1,r6}\r
1450 stmia r5!, {r1,r6}\r
1451 bne .fl_loopcpRGB555_sh\r
1452\r
1453 @ hilighted pixels:\r
1454 sub r3, r3, #0x40*2\r
1455 mov lr, #0x40/2\r
1456.fl_loopcpRGB555_hi:\r
1457 ldr r1, [r3], #4\r
1458 TileDoShHi2Pixels555 r1\r
1459 str r1, [r4], #4\r
1460 subs lr, lr, #1\r
1461 bne .fl_loopcpRGB555_hi\r
1462\r
1463 sub r3, r3, #0x40*2\r
1464\r
cc68a136 1465.fl_noshRGB555:\r
3d48f143 1466 ldr r0, =DrawLineDest\r
cc68a136 1467 ldr r1, =(HighCol+8)\r
3d48f143 1468 ldr r0, [r0]\r
1469\r
1470 ldrb r12, [r8, #12]\r
cc68a136 1471 mov lr, #0xff\r
1472 mov lr, lr, lsl #1\r
1473\r
3d48f143 1474 tst r12, #1\r
1475 movne r2, #320/8 @ len\r
1476 bne .fl_no32colRGB555\r
1477 ldr r4, =PicoOpt\r
1478 mov r2, #256/8\r
1479 ldr r4, [r4]\r
1480 tst r4, #0x4000\r
1481 bne .fl_32scale_RGB555\r
1482 tst r4, #0x0100\r
1483 addeq r0, r0, #32*2\r
1484\r
1485.fl_no32colRGB555:\r
cc68a136 1486.fl_loopRGB555:\r
1487\r
1488 ldr r12, [r1], #4\r
1489 ldr r7, [r1], #4\r
1490\r
1491 and r4, lr, r12, lsl #1\r
1492 ldrh r4, [r3, r4]\r
1493 and r5, lr, r12, lsr #7\r
1494 ldrh r5, [r3, r5]\r
1495 and r6, lr, r12, lsr #15\r
1496 ldrh r6, [r3, r6]\r
1497 orr r4, r4, r5, lsl #16\r
1498\r
1499 and r5, lr, r12, lsr #23\r
1500 ldrh r5, [r3, r5]\r
1501 and r8, lr, r7, lsl #1\r
1502 ldrh r8, [r3, r8]\r
1503 orr r5, r6, r5, lsl #16\r
1504\r
1505 and r6, lr, r7, lsr #7\r
1506 ldrh r6, [r3, r6]\r
1507 and r12,lr, r7, lsr #15\r
1508 ldrh r12,[r3, r12]\r
1509 orr r8, r8, r6, lsl #16\r
1510\r
1511 and r6, lr, r7, lsr #23\r
1512 ldrh r6, [r3, r6] @ 1 cycle interlock here (r6)\r
1513 subs r2, r2, #1\r
1514 orr r12,r12, r6, lsl #16\r
1515\r
1516 stmia r0!, {r4,r5,r8,r12}\r
1517 bne .fl_loopRGB555\r
1518\r
3d48f143 1519 ldmfd sp!, {r4-r8,lr}\r
1520 bx lr\r
1521\r
1522\r
1523.fl_32scale_RGB555:\r
1524 stmfd sp!, {r9,r10}\r
1525 mov r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1526 orr r9, r9, #0x00e7\r
1527\r
1528.fl_loop32scale_RGB555:\r
1529 ldr r12, [r1], #4\r
1530 ldr r7, [r1], #4\r
cc68a136 1531\r
3d48f143 1532 and r4, lr, r12,lsl #1\r
1533 ldrh r4, [r3, r4]\r
1534 and r5, lr, r12,lsr #7\r
1535 ldrh r5, [r3, r5]\r
1536 and r4, r4, r9, lsl #2\r
1537 orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0\r
1538 and r5, r5, r9, lsl #2\r
1539 sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1\r
1540 add r4, r4, r6, lsl #16 @ pix_d 0, 1\r
1541 and r6, lr, r12,lsr #15\r
1542 ldrh r6, [r3, r6]\r
1543 and r12,lr, r12,lsr #23\r
1544 ldrh r12,[r3, r12]\r
1545 and r6, r6, r9, lsl #2\r
1546 add r5, r5, r6\r
1547 mov r5, r5, lsr #1\r
1548 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1549 orr r5, r5, r6, lsl #16\r
1550\r
1551 and r6, lr, r7, lsl #1\r
1552 ldrh r6, [r3, r6]\r
1553 and r12,r12,r9, lsl #2\r
1554 add r5, r5, r12,lsl #14 @ pix_d 2, 3\r
1555 and r6, r6, r9, lsl #2\r
1556 orr r6, r12,r6, lsl #16 @ pix_d 4, 5\r
1557\r
1558 and r12,lr, r7, lsr #7\r
1559 ldrh r12,[r3, r12]\r
1560 and r10,lr, r7, lsr #15\r
1561 ldrh r10,[r3, r10]\r
1562 and r12,r12,r9, lsl #2\r
1563 sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1\r
1564 add r8, r8, r6, lsr #18\r
1565 and r7, lr, r7, lsr #23\r
1566 ldrh r7, [r3, r7]\r
1567 and r10,r10,r9, lsl #2\r
1568 orr r8, r8, r10,lsl #15\r
1569 add r8, r8, r12,lsl #15 @ pix_d 6, 7\r
1570 sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2\r
1571 and r7, r7, r9, lsl #2\r
1572 add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3\r
1573 orr r10,r10,r7, lsl #16 @ pix_d 8, 9\r
1574\r
1575 subs r2, r2, #1\r
1576\r
1577 stmia r0!, {r4,r5,r6,r8,r10}\r
1578 bne .fl_loop32scale_RGB555\r
1579\r
1580 ldmfd sp!, {r9,r10}\r
cc68a136 1581 ldmfd sp!, {r4-r8,lr}\r
3d48f143 1582 bx lr\r
1583\r
cc68a136 1584\r
1585@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1586\r
1587@ utility\r
1588.global blockcpy @ void *dst, void *src, size_t n\r
1589\r
1590blockcpy:\r
1591 stmfd sp!, {r4,r5}\r
1592 mov r2, r2, lsr #4\r
1593blockcpy_loop:\r
1594 ldmia r1!, {r3-r5,r12}\r
1595 subs r2, r2, #1\r
1596 stmia r0!, {r3-r5,r12}\r
1597 bne blockcpy_loop\r
1598 ldmfd sp!, {r4,r5}\r
1599 bx lr\r
1600\r
1601\r
1602.global blockcpy_or @ void *dst, void *src, size_t n, int pat\r
1603\r
1604blockcpy_or:\r
1605 stmfd sp!, {r4-r6}\r
1606 orr r3, r3, r3, lsl #8\r
1607 orr r3, r3, r3, lsl #16\r
1608 mov r2, r2, lsr #4\r
1609blockcpy_loop_or:\r
1610 ldmia r1!, {r4-r6,r12}\r
1611 subs r2, r2, #1\r
1612 orr r4, r4, r3\r
1613 orr r5, r5, r3\r
1614 orr r6, r6, r3\r
1615 orr r12,r12,r3\r
1616 stmia r0!, {r4-r6,r12}\r
1617 bne blockcpy_loop_or\r
1618 ldmfd sp!, {r4-r6}\r
1619 bx lr\r
1620\r
fast MegaDrive/MegaCD/32X emulator