notaz.gp2x.de / picodrive.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
timing hack removed, seems to be no longer needed
[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
7a7c6476 72@ TileNormShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
cc68a136 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
7a7c6476 84@ TileFlipShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
cc68a136 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
740da8c6 309 orrne r10,r10, #1<<23 @ r10=(cells<<24|sh<<23|hi_not_empty<<22|had_output<<21|ty)\r
cc68a136 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
740da8c6 345 orr r10, r10, #1<<21 @ seen non hi-prio tile\r
cc68a136 346\r
347 movs r2, r9, lsl #20 @ if (code&0x1000)\r
348 mov r2, r2, lsl #1\r
349 add r2, r2, r10, lsl #17\r
350 mov r2, r2, lsr #17\r
351 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
352\r
353 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
354\r
355 bic r7, r3, #0x3f\r
356 and r3, r9, #0x6000\r
357 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
358\r
359.DrawStrip_samecode:\r
360 tst r2, r2\r
361 beq .dsloop @ tileline blank\r
362\r
363 cmp r2, r2, ror #4\r
364 beq .DrawStrip_SingleColor @ tileline singlecolor \r
365\r
366 tst r9, #0x0800\r
7a7c6476 367 bne .DrawStrip_TileFlip\r
cc68a136 368\r
369 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
cc68a136 370.DrawStrip_TileNorm:\r
371 TileNorm r0\r
372 b .dsloop\r
373\r
7a7c6476 374.DrawStrip_TileFlip:\r
375 TileFlip r0\r
376 b .dsloop\r
377\r
cc68a136 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
cc68a136 391.DrawStrip_hiprio_maybempt:\r
392 cmp r7, r9\r
393 beq .dsloop @ must've been empty, otherwise we wouldn't get here\r
394 movs r2, r7, lsl #20 @ if (code&0x1000)\r
395 mov r2, r2, lsl #1\r
396 add r2, r2, r10, lsl #17\r
397 mov r2, r2, lsr #17\r
398 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
399 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
400 mov r9, r7 @ remember code\r
401 tst r2, r2\r
740da8c6 402 beq .dsloop\r
403 orr r10, r10, #1<<22\r
404\r
405.DrawStrip_hiprio:\r
406 tst r10, #0x00c00000\r
407 beq .DrawStrip_hiprio_maybempt\r
408 sub r0, r1, r11\r
409 orr r7, r7, r0, lsl #16\r
410 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
411 tst r7, #0x1000\r
412 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
413 str r7, [r6], #4 @ cache hi priority tile\r
414 mov r0, #0xf\r
cc68a136 415 b .dsloop\r
416\r
417.dsloop_exit:\r
740da8c6 418 tst r10, #1<<21 @ seen non hi-prio tile\r
419 ldreq r1, =rendstatus\r
cc68a136 420 mov r0, #0\r
740da8c6 421 ldreq r2, [r1]\r
cc68a136 422 str r0, [r6] @ terminate the cache list\r
740da8c6 423 orreq r2, r2, #0x40 @ had a layer with all hi-prio tiles\r
424 streq r2, [r1]\r
cc68a136 425\r
426 ldmfd sp!, {r4-r11,lr}\r
427 bx lr\r
428\r
6d7acf9e 429@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
cc68a136 430\r
431.DrawStrip_vsscroll:\r
6d7acf9e 432 rsb r8, r3, #0\r
433 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
434 bic r8, r8, #0xff000000\r
740da8c6 435 orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|tilex[15:0])\r
cc68a136 436\r
6d7acf9e 437 ldr r4, =Scanline\r
438 orr r5, r1, r10, lsl #24\r
439 ldr r4, [r4]\r
440 sub r1, r3, #1\r
441 orr r5, r5, r4, lsl #16 @ r5=(shift_width[31:24]|scanline[23:16]|ymask[15:0])\r
442 and r1, r1, #7\r
443 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
cc68a136 444\r
6d7acf9e 445 mov r10,r9, lsl #16\r
446 tst r0, r0\r
447 orrne r10,r10, #0x8000\r
448 tst r9, #1<<31\r
449 mov r3, #0\r
450 orr r10,r10, #0xff000000 @ will be adjusted on entering loop\r
451 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
452 movne r3, #0x40 @ default to shadowed pal on sh mode\r
cc68a136 453\r
6d7acf9e 454 mvn r9, #0 @ r9=prevcode=-1\r
cc68a136 455\r
6d7acf9e 456 @ cache some stuff to avoid mem access\r
457 ldr r11,=HighCol\r
458 mov r0, #0xf\r
459 add r1, r11, r7 @ r1=pdest\r
460\r
461 cmp r7, #8\r
462 subne r10,r10, #0x01000000 @ have hscroll, start with negative cell\r
463\r
464\r
465 @ r4 & r7 are scratch in this loop\r
466.dsloop_vs_subr1:\r
467 sub r1, r1, #8\r
468.dsloop_vs: @ 40-41 times\r
469 add r10,r10, #0x01000000\r
470 and r4, r10, #0x003f0000\r
471 cmp r4, r10, asr #8\r
740da8c6 472 ble .dsloop_vs_exit\r
6d7acf9e 473\r
474 @ calc offset and read tileline code to r7, also calc ty\r
475 add r7, lr, #0x012000\r
476 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
477 add r7, r7, r10,asr #23 @ vsram + ((cell&~1)<<1)\r
478 bic r7, r7, #3\r
479 tst r10,#0x8000 @ plane1?\r
480 addne r7, r7, #2\r
481 ldrh r7, [r7] @ r7=vscroll\r
482\r
483 bic r10,r10,#0xff @ clear old ty\r
484 and r4, r5, #0xff0000\r
485 add r4, r4, r7, lsl #16\r
486 and r4, r4, r5, lsl #16 @ r4=line<<16\r
487 and r7, r4, #0x70000\r
488 orr r10,r10,r7, lsr #15 @ new ty\r
489\r
490 mov r4, r4, lsr #19\r
491 mov r7, r5, lsr #24\r
492 mov r4, r4, lsl r7 @ nametabadd\r
493\r
494 and r7, r8, r8, lsr #25\r
495 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
496 add r7, r7, r4, lsl #1\r
497 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
498\r
499 add r1, r1, #8\r
500 add r8, r8, #1\r
501\r
502 tst r7, #0x8000\r
503 bne .DrawStrip_vs_hiprio\r
504\r
505 cmp r7, r9\r
506 beq .DrawStrip_vs_samecode @ we know stuff about this tile already\r
507\r
508 mov r9, r7 @ remember code\r
740da8c6 509 orr r8, r8, #1<<24 @ seen non hi-prio tile\r
6d7acf9e 510\r
511 movs r2, r9, lsl #20 @ if (code&0x1000)\r
512 mov r2, r2, lsl #1\r
513 add r2, r2, r10, lsl #17\r
514 mov r2, r2, lsr #17\r
515 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
516\r
517 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
518\r
519 bic r7, r3, #0x3f\r
520 and r3, r9, #0x6000\r
521 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
522\r
523.DrawStrip_vs_samecode:\r
524 tst r2, r2\r
525 beq .dsloop_vs @ tileline blank\r
526\r
527 cmp r2, r2, ror #4\r
528 beq .DrawStrip_vs_SingleColor @ tileline singlecolor \r
529\r
530 tst r9, #0x0800\r
7a7c6476 531 bne .DrawStrip_vs_TileFlip\r
6d7acf9e 532\r
533 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
6d7acf9e 534.DrawStrip_vs_TileNorm:\r
535 TileNorm r0\r
536 b .dsloop_vs\r
537\r
7a7c6476 538.DrawStrip_vs_TileFlip:\r
539 TileFlip r0\r
540 b .dsloop_vs\r
541\r
6d7acf9e 542.DrawStrip_vs_SingleColor:\r
543 and r4, r2, #0xf\r
544 orr r4, r3, r4\r
545 orr r4, r4, r4, lsl #8\r
546 tst r1, #1 @ not aligned?\r
547 strneb r4, [r1], #1\r
548 streqh r4, [r1], #2\r
549 strh r4, [r1], #2\r
550 strh r4, [r1], #2\r
551 strh r4, [r1], #2\r
552 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
553 b .dsloop_vs_subr1\r
554\r
555.DrawStrip_vs_hiprio:\r
556 tst r10, #0x00c00000\r
557 beq .DrawStrip_vs_hiprio_maybempt\r
558 sub r0, r1, r11\r
559 orr r7, r7, r0, lsl #16\r
560 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
561 tst r7, #0x1000\r
562 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
563 str r7, [r6], #4 @ cache hi priority tile\r
564 mov r0, #0xf\r
565 b .dsloop_vs\r
566\r
567.DrawStrip_vs_hiprio_maybempt:\r
568 cmp r7, r9\r
569 beq .dsloop_vs @ must've been empty, otherwise we wouldn't get here\r
570 movs r2, r7, lsl #20 @ if (code&0x1000)\r
571 mov r2, r2, lsl #1\r
572 add r2, r2, r10, lsl #17\r
573 mov r2, r2, lsr #17\r
574 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
575 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
576 mov r9, r7 @ remember code\r
577 tst r2, r2\r
578 orrne r10, r10, #1<<22\r
579 bne .DrawStrip_vs_hiprio\r
580 b .dsloop_vs\r
581\r
740da8c6 582.dsloop_vs_exit:\r
583 tst r8, #1<<24 @ seen non hi-prio tile\r
584 ldreq r1, =rendstatus\r
585 mov r0, #0\r
586 ldreq r2, [r1]\r
587 str r0, [r6] @ terminate the cache list\r
588 orreq r2, r2, #0x40 @ had a layer with all hi-prio tiles\r
589 streq r2, [r1]\r
590\r
591 ldmfd sp!, {r4-r11,lr}\r
592 bx lr\r
593\r
6d7acf9e 594\r
595@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
cc68a136 596\r
597@ interlace mode 2? Sonic 2?\r
598.DrawStrip_interlace:\r
599 tst r0, r0\r
600 moveq r7, r7, lsl #21\r
601 movne r7, r7, lsl #5\r
602\r
603 @ Find the line in the name table\r
604 add r2, r7, r2, lsl #22 @ r2=(vscroll+(Scanline<<1))<<21 (11 bits);\r
605 orr r1, r1, #0x80000000\r
606 and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
607 mov r2, r2, lsr #21\r
608 mov r4, r2, lsr #4\r
609 mov r12, r12, lsr #1 @ halfwords\r
610 add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];\r
611 and r9, r9, #0xff\r
612\r
613 sub sp, sp, #6*4\r
614 stmia sp, {r0,r2,r3,r5,r6,r9}\r
615\r
616 mov r0, sp\r
617 bl DrawStripInterlace @ struct TileStrip *ts\r
618\r
619 add sp, sp, #6*4\r
620 ldmfd sp!, {r4-r11,lr}\r
621 bx lr\r
622\r
623.pool\r
624\r
625@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
626\r
627\r
628.global BackFill @ int reg7, int sh\r
629\r
630BackFill:\r
631 stmfd sp!, {r4-r9,lr}\r
632\r
633 ldr lr, =(HighCol+8)\r
634\r
635 mov r0, r0, lsl #26\r
636 mov r0, r0, lsr #26\r
637 orr r0, r0, r1, lsl #6\r
638 orr r0, r0, r0, lsl #8\r
639 orr r0, r0, r0, lsl #16\r
640\r
641 mov r1, r0\r
642 mov r2, r0\r
643 mov r3, r0\r
644 mov r4, r0\r
645 mov r5, r0\r
646 mov r6, r0\r
647 mov r7, r0\r
648\r
649 @ go go go!\r
650 stmia lr!, {r0-r7} @ 10*8*4\r
651 stmia lr!, {r0-r7}\r
652 stmia lr!, {r0-r7}\r
653 stmia lr!, {r0-r7}\r
654 stmia lr!, {r0-r7}\r
655 stmia lr!, {r0-r7}\r
656 stmia lr!, {r0-r7}\r
657 stmia lr!, {r0-r7}\r
658 stmia lr!, {r0-r7}\r
659 stmia lr!, {r0-r7}\r
660\r
661 ldmfd sp!, {r4-r9,r12}\r
662 bx r12\r
663\r
664\r
665@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
666\r
667\r
7a7c6476 668.global DrawTilesFromCache @ int *hc, int sh, int rlim\r
cc68a136 669\r
670DrawTilesFromCache:\r
671 stmfd sp!, {r4-r8,r11,lr}\r
672\r
cc68a136 673 @ cache some stuff to avoid mem access\r
674 ldr r11,=HighCol\r
675 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
676 mov r12,#0xf\r
677\r
740da8c6 678 mvn r5, #0 @ r5=prevcode=-1\r
7a7c6476 679 ands r8, r1, #1\r
680 orr r8, r8, r2, lsl #1\r
740da8c6 681 bne .dtfc_check_rendflags\r
682\r
cc68a136 683 @ scratch: r4, r7\r
684.dtfc_loop:\r
685 ldr r6, [r0], #4 @ read code\r
686 movs r1, r6, lsr #16 @ r1=dx;\r
687 ldmeqfd sp!, {r4-r8,r11,pc} @ dx is never zero, this must be a terminator, return\r
7a7c6476 688 bic r4, r1, #0xfe00\r
689 add r1, r11, r4 @ r1=pdest\r
cc68a136 690\r
691 mov r7, r6, lsl #16\r
692 cmp r5, r7, lsr #16\r
693 beq .dtfc_samecode @ if (code==prevcode)\r
694\r
695 mov r5, r7, lsr #16\r
696\r
697 mov r2, r5, lsl #21\r
698 mov r2, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4;\r
699 add r2, r2, r6, lsr #25 @ addr+=ty\r
700\r
701 and r3, r5, #0x6000\r
702 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
703\r
704 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
705\r
706.dtfc_samecode:\r
7a7c6476 707 rsbs r4, r4, r8, lsr #1\r
708 bmi .dtfc_cut_tile\r
709\r
710 tst r8, #1\r
cc68a136 711 bne .dtfc_shadow\r
712\r
713 tst r2, r2\r
714 beq .dtfc_loop\r
715\r
716 cmp r2, r2, ror #4\r
717 beq .dtfc_SingleColor @ tileline singlecolor \r
718\r
719 tst r5, #0x0800\r
7a7c6476 720 bne .dtfc_TileFlip\r
cc68a136 721\r
722 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 723.dtfc_TileNorm:\r
724 TileNorm r12\r
725 b .dtfc_loop\r
726\r
7a7c6476 727.dtfc_TileFlip:\r
728 TileFlip r12\r
729 b .dtfc_loop\r
730\r
cc68a136 731.dtfc_SingleColor:\r
732 and r4, r2, #0xf\r
733 orr r4, r3, r4\r
734 orr r4, r4, r4, lsl #8\r
735 tst r1, #1 @ not aligned?\r
736 strneb r4, [r1], #1\r
737 streqh r4, [r1], #2\r
738 strh r4, [r1], #2\r
739 strh r4, [r1], #2\r
740 strh r4, [r1], #2\r
741 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
742 b .dtfc_loop\r
743\r
744.dtfc_shadow:\r
745 tst r2, r2\r
746 beq .dtfc_shadow_blank\r
747\r
748 cmp r2, r2, ror #4\r
749 beq .dtfc_SingleColor @ tileline singlecolor \r
750\r
751 tst r5, #0x0800\r
7a7c6476 752 bne .dtfc_TileFlipShHP\r
cc68a136 753\r
754 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 755.dtfc_TileNormShHP:\r
756 TileNormShHP\r
757 b .dtfc_loop\r
758\r
7a7c6476 759.dtfc_TileFlipShHP:\r
760 TileFlipShHP\r
761 b .dtfc_loop\r
762\r
cc68a136 763.dtfc_shadow_blank:\r
764 ldrb r4, [r1] @ 1ci\r
765 ldrb r12,[r1,#1]\r
766 tst r4, #0x80\r
767 andeq r4, r4,#0x3f\r
768 streqb r4, [r1]\r
769 tst r12,#0x80\r
770 ldrb r4, [r1,#2]\r
771 andeq r12,r12,#0x3f\r
772 streqb r12,[r1,#1]\r
773 tst r4, #0x80\r
774 ldrb r12,[r1,#3]\r
775 andeq r4, r4,#0x3f\r
776 streqb r4, [r1,#2]\r
777 tst r12,#0x80\r
778 ldrb r4, [r1,#4]\r
779 andeq r12,r12,#0x3f\r
780 streqb r12,[r1,#3]\r
781 tst r4, #0x80\r
782 ldrb r12,[r1,#5]\r
783 andeq r4, r4,#0x3f\r
784 streqb r4, [r1,#4]\r
785 tst r12,#0x80\r
786 ldrb r4, [r1,#6]\r
787 andeq r12,r12,#0x3f\r
788 streqb r12,[r1,#5]\r
789 tst r4, #0x80\r
790 ldrb r12,[r1,#7]\r
791 andeq r4, r4,#0x3f\r
792 streqb r4, [r1,#6]\r
793 tst r12,#0x80\r
794 andeq r12,r12,#0x3f\r
795 streqb r12,[r1,#7]\r
796 mov r12, #0xf\r
797 b .dtfc_loop\r
798\r
7a7c6476 799.dtfc_cut_tile:\r
800 add r4, r4, #7 @ 0-6\r
801 mov r4, r4, lsl #2\r
802 mov r12,#0xf<<28\r
803 mov r12,r12,asr r4\r
804 mov r2, r2, ror #16\r
805 tst r5, #0x0800 @ flipped?\r
806 mvnne r12,r12\r
807 and r2, r2, r12\r
808 mov r2, r2, ror #16\r
809 mov r12,#0xf\r
810 tst r8, #1\r
811 bne .dtfc_shadow\r
812 tst r2, r2\r
813 beq .dtfc_loop\r
814 tst r5, #0x0800\r
815 beq .dtfc_TileNorm\r
816 b .dtfc_TileFlip\r
817\r
740da8c6 818@ check if we have detected layer covered with hi-prio tiles:\r
819.dtfc_check_rendflags:\r
820 ldr r1, =rendstatus\r
821 ldr r2, [r1]\r
822 tst r2, #0xc0\r
823 beq .dtfc_loop\r
7a7c6476 824 bic r8, r8, #1 @ sh/hi mode off\r
740da8c6 825 tst r2, #0x80\r
826 bne .dtfc_loop @ already processed\r
827 orr r2, r2, #0x80\r
828 str r2, [r1]\r
829\r
830 add r1, r11,#8\r
831 mov r3, #320/4\r
832 mov r7, #0x80\r
833 orr r7, r7, r7, lsl #8\r
834 orr r7, r7, r7, lsl #16\r
835 mov r6, #0x3f\r
836 orr r6, r6, r6, lsl #8\r
837 orr r6, r6, r6, lsl #16\r
838.dtfc_loop_shprep:\r
839 subs r3, r3, #1\r
840 bmi .dtfc_loop @ done\r
841 ldr r2, [r1]\r
842 tst r2, r7\r
843 andeq r2, r2, r6\r
844 streq r2, [r1], #4\r
845 beq .dtfc_loop_shprep\r
846 tst r2, #0x80000000\r
847 biceq r2, r2, #0xc0000000\r
848 tst r2, #0x00800000\r
849 biceq r2, r2, #0x00c00000\r
850 tst r2, #0x00008000\r
851 biceq r2, r2, #0x0000c000\r
852 tst r2, #0x00000080\r
853 biceq r2, r2, #0x000000c0\r
854 str r2, [r1], #4\r
855 b .dtfc_loop_shprep\r
856\r
cc68a136 857.pool\r
858\r
859@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
860\r
861\r
862.global DrawSpritesFromCache @ int *hc, int sh\r
863\r
864DrawSpritesFromCache:\r
865 stmfd sp!, {r4-r11,lr}\r
866\r
867 @ cache some stuff to avoid mem access\r
868 ldr r11,=HighCol\r
869 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
870 mov r6, r1, lsl #31\r
871 orr r6, r6, #1<<30\r
872 mov r12,#0xf\r
873\r
874 mov r10, r0\r
875\r
876.dsfc_loop:\r
877 ldr r9, [r10], #4 @ read code\r
878 bic r6, r6, #7 @ using pipeline\r
879 tst r9, r9\r
880 ldmeqfd sp!, {r4-r11,pc}\r
881\r
882 mov r4, r9, lsl #28\r
883 orr r6, r6, r4, lsr #30\r
884 add r6, r6, #1 @ r6=s1cc???? ... ?????www (s=shadow/hilight, cc=pal, w=width)\r
885\r
886 and r5, r9, #3\r
887 add r5, r5, #1 @ r5=delta\r
888 tst r9, #0x10000\r
889 rsbne r5, r5, #0 @ Flip X\r
890 mov r5, r5, lsl #4\r
891\r
892 mov r2, r9, lsr #17\r
893 mov r8, r2, lsl #1 @ tile=((unsigned int)code>>17)<<1;\r
894\r
895 and r3, r9, #0x30 @ r3=pal=(code&0x30);\r
896\r
897 bic r6, r6, #3<<28\r
898 orr r6, r6, r3, lsl #24\r
899\r
900 mov r0, r9, lsl #16\r
901 mov r0, r0, asr #22 @ sx=(code<<16)>>22\r
902 adds r0, r0, #0 @ set ZV\r
903 b .dsfc_inloop_enter\r
904\r
905@ scratch: r4, r7\r
906.dsfc_inloop:\r
907 sub r6, r6, #1\r
908 tst r6, #7\r
909 beq .dsfc_loop\r
910 adds r0, r0, #8\r
911 add r8, r8, r5\r
912\r
913.dsfc_inloop_enter:\r
914 ble .dsfc_inloop\r
915 cmp r0, #328\r
916 bge .dsfc_loop\r
917\r
918 mov r8, r8, lsl #17\r
919 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
920\r
921 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+tile); // Get 8 pixels\r
922 add r1, r11, r0 @ r1=pdest\r
923 tst r2, r2\r
924 beq .dsfc_inloop\r
925\r
926 cmp r12, r6, lsr #28\r
927 beq .dsfc_shadow\r
928\r
929 cmp r2, r2, ror #4\r
930 beq .dsfc_SingleColor @ tileline singlecolor \r
931\r
932 tst r9, #0x10000\r
7a7c6476 933 bne .dsfc_TileFlip\r
cc68a136 934\r
935 @ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 936.dsfc_TileNorm:\r
937 TileNorm r12\r
938 b .dsfc_inloop\r
939\r
7a7c6476 940.dsfc_TileFlip:\r
941 TileFlip r12\r
942 b .dsfc_inloop\r
943\r
cc68a136 944.dsfc_SingleColor:\r
945 tst r0, #1 @ not aligned?\r
946 and r4, r2, #0xf\r
947 orr r4, r3, r4\r
948 orr r4, r4, r4, lsl #8\r
949 strneb r4, [r1], #1\r
950 streqh r4, [r1], #2\r
951 strh r4, [r1], #2\r
952 strh r4, [r1], #2\r
953 strh r4, [r1], #2\r
954 strneb r4, [r1], #1\r
955 b .dsfc_inloop\r
956\r
957.dsfc_shadow:\r
958 cmp r2, r2, ror #4\r
959 beq .dsfc_singlec_sh\r
960\r
961 tst r9, #0x10000\r
7a7c6476 962 bne .dsfc_TileFlip_sh\r
cc68a136 963\r
964 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 965.dsfc_TileNorm_sh:\r
966 TileNormSh\r
967 b .dsfc_inloop\r
968\r
7a7c6476 969.dsfc_TileFlip_sh:\r
970 TileFlipSh\r
971 b .dsfc_inloop\r
972\r
cc68a136 973.dsfc_singlec_sh:\r
974 cmp r2, #0xe0000000\r
975 bcc .dsfc_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
976 tst r2, #0x10000000\r
977 bne .dsfc_sh_sh\r
978 TileSingleHi\r
979 b .dsfc_inloop\r
980\r
981.dsfc_sh_sh:\r
982 TileSingleSh\r
983 b .dsfc_inloop\r
984\r
985.pool\r
986\r
987@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
988\r
989@ + 0 : hhhhvvvv ab--hhvv yyyyyyyy yyyyyyyy // a: offscreen h, b: offs. v, h: horiz. size\r
990@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
991\r
992.global DrawSprite @ unsigned int *sprite, int **hc, int sh\r
993\r
994DrawSprite:\r
995 stmfd sp!, {r4-r9,r11,lr}\r
996\r
997 ldr r3, [r0] @ sprite[0]\r
998 ldr r7, =Scanline\r
999 mov r6, r3, lsr #28\r
1000 sub r6, r6, #1 @ r6=width-1 (inc later)\r
1001 mov r5, r3, lsr #24\r
1002 and r5, r5, #7 @ r5=height\r
1003\r
1004 mov r4, r3, lsl #16 @ r4=sy<<16 (tmp)\r
1005\r
1006 ldr r7, [r7]\r
1007 ldr r9, [r0, #4]\r
1008 sub r7, r7, r4, asr #16 @ r7=row=Scanline-sy\r
1009\r
1010 tst r2, r2\r
1011 mov r2, r9, asr #16 @ r2=sx\r
1012 bic r9, r9, #0xfe000000\r
1013 orrne r9, r9, #1<<31 @ r9=code|(sh<<31)\r
1014\r
1015 tst r9, #0x1000\r
1016 movne r4, r5, lsl #3\r
1017 subne r4, r4, #1\r
1018 subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
1019\r
1020 mov r8, r9, lsl #21\r
1021 mov r8, r8, lsr #21\r
1022 add r8, r8, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
1023 \r
1024 tst r9, #0x0800\r
1025 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1026 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
1027\r
1028 mov r8, r8, lsl #4\r
1029 and r7, r7, #7\r
1030 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1031\r
1032 tst r9, #0x8000\r
1033 bne .dspr_cache @ if(code&0x8000) // high priority - cache it\r
1034\r
1035 @ cache some stuff to avoid mem access\r
1036 ldr r11,=HighCol\r
1037 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
1038 mov r12,#0xf\r
1039\r
1040 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
1041 and r4, r9, #0x6000\r
1042 orr r9, r9, r4, lsl #16\r
1043 orr r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
1044\r
1045 tst r9, #1<<31\r
1046 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
1047 orrne r3, r3, #0x40 @ shadow by default\r
1048\r
1049 add r6, r6, #1 @ inc now\r
1050 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1051 b .dspr_loop_enter\r
1052\r
1053.dspr_loop:\r
1054 subs r6, r6, #1 @ width--\r
1055 ldmeqfd sp!, {r4-r9,r11,pc}@ return\r
1056 adds r0, r0, #8 @ sx+=8\r
1057 add r8, r8, r5 @ tile+=delta\r
1058\r
1059.dspr_loop_enter:\r
1060 ble .dspr_loop @ sx <= 0\r
1061 cmp r0, #328\r
1062 ldmgefd sp!, {r4-r9,r11,pc}@ return\r
1063\r
1064 mov r8, r8, lsl #17\r
1065 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
1066\r
1067 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1068 add r1, r11, r0 @ r1=pdest\r
1069 tst r2, r2\r
1070 beq .dspr_loop\r
1071\r
1072 cmp r12, r9, lsr #28\r
1073 beq .dspr_shadow\r
1074\r
1075 cmp r2, r2, ror #4\r
1076 beq .dspr_SingleColor @ tileline singlecolor \r
1077\r
1078 tst r9, #0x0800\r
7a7c6476 1079 bne .dspr_TileFlip\r
cc68a136 1080\r
1081 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 1082@ scratch: r4, r7\r
1083.dspr_TileNorm:\r
1084 TileNorm r12\r
1085 b .dspr_loop\r
1086\r
7a7c6476 1087.dspr_TileFlip:\r
1088 TileFlip r12\r
1089 b .dspr_loop\r
1090\r
cc68a136 1091.dspr_SingleColor:\r
1092 and r4, r2, #0xf\r
1093 orr r4, r3, r4\r
1094 orr r4, r4, r4, lsl #8\r
1095 tst r0, #1 @ not aligned?\r
1096 strneb r4, [r1], #1\r
1097 streqh r4, [r1], #2\r
1098 strh r4, [r1], #2\r
1099 strh r4, [r1], #2\r
1100 strh r4, [r1], #2\r
1101 strneb r4, [r1], #1\r
1102 b .dspr_loop\r
1103\r
1104.dspr_shadow:\r
1105 cmp r2, r2, ror #4\r
1106 beq .dspr_singlec_sh\r
1107\r
1108 tst r9, #0x0800\r
7a7c6476 1109 bne .dspr_TileFlip_sh\r
cc68a136 1110\r
1111 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 1112.dspr_TileNorm_sh:\r
1113 TileNormSh\r
1114 b .dspr_loop\r
1115\r
7a7c6476 1116.dspr_TileFlip_sh:\r
1117 TileFlipSh\r
1118 b .dspr_loop\r
1119\r
cc68a136 1120.dspr_singlec_sh:\r
1121 cmp r2, #0xe0000000\r
1122 bcc .dspr_SingleColor @ normal tileline\r
1123 tst r2, #0x10000000\r
1124 bne .dspr_sh_sh\r
1125 TileSingleHi\r
1126 b .dspr_loop\r
1127\r
1128.dspr_sh_sh:\r
1129 TileSingleSh\r
1130 b .dspr_loop\r
1131\r
1132\r
1133.dspr_cache:\r
1134 @ *(*hc)++ = (tile<<16)|((code&0x0800)<<5)|((sx<<6)&0x0000ffc0)|((code>>9)&0x30)|((sprite[0]>>24)&0xf);\r
1135 mov r4, r8, lsl #16 @ tile\r
1136 tst r9, #0x0800\r
1137 orrne r4, r4, #0x10000 @ code&0x0800\r
1138 mov r2, r2, lsl #22\r
1139 orr r4, r4, r2, lsr #16 @ (sx<<6)&0x0000ffc0\r
1140 and r2, r9, #0x6000\r
1141 orr r4, r4, r2, lsr #9 @ (code>>9)&0x30\r
1142 mov r3, r3, lsl #12\r
1143 ldr r2, [r1]\r
1144 orr r4, r4, r3, lsr #28 @ (sprite[0]>>24)&0xf\r
1145\r
1146 str r4, [r2], #4\r
1147 str r2, [r1]\r
1148\r
1149 ldmfd sp!, {r4-r9,r11,lr}\r
1150 bx lr\r
1151\r
1152@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1153\r
1154.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache\r
1155\r
1156DrawWindow:\r
1157 stmfd sp!, {r4-r11,lr}\r
1158\r
1159 ldr r11, =(Pico+0x22228) @ Pico.video\r
1160 ldr r10, =Scanline\r
1161 ldrb r12, [r11, #3] @ pvid->reg[3]\r
1162\r
1163 ldr r10, [r10]\r
1164 ldr r4, [r11, #12]\r
1165 mov r5, r10, lsr #3\r
1166 and r10, r10, #7\r
1167 mov r10, r10, lsl #1 @ r10=ty\r
1168\r
1169 mov r12, r12, lsl #10\r
1170\r
1171 tst r4, #1 @ 40 cell mode?\r
1172 andne r12, r12, #0xf000 @ 0x3c<<10\r
1173 andeq r12, r12, #0xf800\r
1174 addne r12, r12, r5, lsl #7\r
1175 addeq r12, r12, r5, lsl #6 @ nametab\r
1176 add r12, r12, r0, lsl #2 @ +starttile\r
1177\r
1178 ldr r6, =rendstatus\r
1179 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
1180 ldrb r6, [r6]\r
1181\r
1182 @ fetch the first code now\r
1183 ldrh r7, [lr, r12]\r
1184\r
1185 ands r6, r6, #2 @ we care about bit 1 only\r
1186 orr r6, r6, r2\r
1187 bne .dw_no_sameprio\r
1188\r
1189 cmp r2, r7, lsr #15\r
1190 ldmnefd sp!, {r4-r11,pc} @ assume that whole window uses same priority\r
1191\r
1192.dw_no_sameprio:\r
1193 orr r6, r6, r3, lsl #8 @ shadow mode\r
1194\r
1195 sub r8, r1, r0\r
1196 mov r8, r8, lsl #1 @ cells\r
1197\r
1198 mvn r9, #0 @ r9=prevcode=-1\r
1199\r
1200 @ cache some stuff to avoid mem access\r
1201 ldr r11,=(HighCol+8)\r
1202 add r1, r11, r0, lsl #4 @ r1=pdest\r
1203 mov r0, #0xf\r
1204 b .dwloop_enter\r
1205\r
1206 @ r4,r5 & r7 are scratch in this loop\r
1207.dwloop:\r
1208 add r1, r1, #8\r
1209.dwloop_nor1:\r
1210 add r12, r12, #2 @ halfwords\r
1211 ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)\r
1212 subs r8, r8, #1\r
1213 beq .dwloop_end @ done\r
1214\r
1215 eor r5, r6, r7, lsr #15\r
1216 tst r5, #1\r
1217 orrne r6, r6, #2 @ wrong pri\r
1218 bne .dwloop\r
1219\r
1220 cmp r7, r9\r
1221 beq .dw_samecode @ we know stuff about this tile already\r
1222\r
1223.dwloop_enter:\r
1224 mov r9, r7 @ remember code\r
1225\r
1226 movs r2, r9, lsl #20 @ if (code&0x1000)\r
1227 mov r2, r2, lsl #1\r
1228 add r2, r10, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4; addr+=ty\r
1229 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
1230\r
1231 and r3, r9, #0x6000\r
1232 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
1233\r
1234 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1235\r
1236.dw_samecode:\r
1237 tst r6, #0x100\r
1238 bne .dw_shadow\r
1239.dw_shadow_done:\r
1240 tst r2, r2\r
1241 beq .dwloop @ tileline blank\r
1242\r
1243 cmp r2, r2, ror #4\r
1244 beq .dw_SingleColor @ tileline singlecolor \r
1245\r
1246 tst r9, #0x0800\r
7a7c6476 1247 bne .dw_TileFlip\r
cc68a136 1248\r
1249 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
cc68a136 1250.dw_TileNorm:\r
1251 TileNorm r0\r
1252 b .dwloop\r
1253\r
7a7c6476 1254.dw_TileFlip:\r
1255 TileFlip r0\r
1256 b .dwloop\r
1257\r
cc68a136 1258.dw_SingleColor:\r
1259 and r4, r0, r2 @ #0x0000000f\r
1260 orr r4, r3, r4\r
1261 orr r4, r4, r4, lsl #8\r
1262 orr r4, r4, r4, lsl #16\r
1263 mov r5, r4\r
1264 stmia r1!, {r4,r5}\r
1265 b .dwloop_nor1 @ we incremeted r1 ourselves\r
1266\r
1267.dw_shadow:\r
1268 tst r6, #1 @ hi pri?\r
1269 orreq r3, r3, #0x40\r
1270 beq .dw_shadow_done\r
1271 ldr r4, [r1]\r
1272 tst r4, #0x00000080\r
1273 biceq r4, r4, #0x000000c0\r
1274 tst r4, #0x00008000\r
1275 biceq r4, r4, #0x0000c000\r
1276 tst r4, #0x00800000\r
1277 biceq r4, r4, #0x00c00000\r
1278 tst r4, #0x80000000\r
1279 biceq r4, r4, #0xc0000000\r
1280 str r4, [r1]\r
1281 ldr r4, [r1,#4]\r
1282 tst r4, #0x00000080\r
1283 biceq r4, r4, #0x000000c0\r
1284 tst r4, #0x00008000\r
1285 biceq r4, r4, #0x0000c000\r
1286 tst r4, #0x00800000\r
1287 biceq r4, r4, #0x00c00000\r
1288 tst r4, #0x80000000\r
1289 biceq r4, r4, #0xc0000000\r
1290 str r4, [r1,#4]\r
1291 b .dw_shadow_done\r
1292\r
1293.dwloop_end:\r
1294 ldr r0, =rendstatus\r
1295 ldr r1, [r0]\r
1296 and r6, r6, #2\r
1297 orr r1, r1, r6\r
1298 str r1, [r0]\r
1299\r
1300 ldmfd sp!, {r4-r11,r12}\r
1301 bx r12\r
1302\r
1303\r
1304@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1305\r
1306\r
1307@ hilights 2 pixels in RGB444/BGR444 format\r
1308.macro TileDoShHi2Pixels444 reg\r
1309 mov \reg, \reg, ror #12\r
1310 adds \reg, \reg, #0x40000000\r
1311 orrcs \reg, \reg, #0xf0000000\r
1312 mov \reg, \reg, ror #28\r
1313 adds \reg, \reg, #0x40000000\r
1314 orrcs \reg, \reg, #0xf0000000\r
1315 mov \reg, \reg, ror #28\r
1316 adds \reg, \reg, #0x40000000\r
1317 orrcs \reg, \reg, #0xf0000000\r
1318 mov \reg, \reg, ror #24\r
1319 adds \reg, \reg, #0x40000000\r
1320 orrcs \reg, \reg, #0xf0000000\r
1321 mov \reg, \reg, ror #28\r
1322 adds \reg, \reg, #0x40000000\r
1323 orrcs \reg, \reg, #0xf0000000\r
1324 mov \reg, \reg, ror #28\r
1325 adds \reg, \reg, #0x40000000\r
1326 orrcs \reg, \reg, #0xf0000000\r
1327 mov \reg, \reg, ror #12\r
1328.endm\r
1329\r
1330\r
1331.global FinalizeLineBGR444 @ int sh\r
1332\r
1333FinalizeLineBGR444:\r
1334 stmfd sp!, {r4-r6,lr}\r
1335 mov r6, r0\r
1336 ldr lr, =(Pico+0x22228) @ Pico.video\r
1337 ldr r0, =DrawLineDest\r
1338 ldrb r12, [lr, #12]\r
1339 ldr r0, [r0]\r
1340 sub r3, lr, #0x128 @ r3=Pico.cram\r
1341\r
1342 tst r12, #1\r
1343 movne r2, #320/4 @ len\r
1344 bne .fl_no32colBGR444\r
1345 ldr r4, =PicoOpt\r
1346 mov r2, #256/4\r
1347 ldr r4, [r4]\r
1348 tst r4, #0x100\r
1349 addeq r0, r0, #32*2\r
1350\r
1351.fl_no32colBGR444:\r
1352 tst r6, r6\r
1353 beq .fl_noshBGR444\r
1354\r
1355 ldr r4, =HighPal\r
1356\r
1357 ldrb r12, [lr, #-0x1a] @ 0x2220e ~ dirtyPal\r
1358 tst r12, r12\r
1359 moveq r3, r4\r
1360 beq .fl_noshBGR444\r
1361 mov r12, #0\r
1362 strb r12, [lr, #-0x1a]\r
1363\r
1364 mov lr, #0x40/8\r
1365 @ copy pal:\r
1366.fl_loopcpBGR444:\r
1367 ldmia r3!, {r1,r5,r6,r12}\r
1368 subs lr, lr, #1\r
1369 stmia r4!, {r1,r5,r6,r12}\r
1370 bne .fl_loopcpBGR444\r
1371\r
1372 @ shadowed pixels:\r
1373 mov r12, #0x0077\r
1374 orr r12,r12,#0x0700\r
1375 orr r12,r12,r12,lsl #16\r
1376 sub r3, r3, #0x40*2\r
1377 add r5, r4, #0x80*2\r
1378 mov lr, #0x40/4\r
1379.fl_loopcpBGR444_sh:\r
1380 ldmia r3!, {r1,r6}\r
1381 subs lr, lr, #1\r
1382 and r1, r12, r1, lsr #1\r
1383 and r6, r12, r6, lsr #1\r
1384 stmia r4!, {r1,r6}\r
1385 stmia r5!, {r1,r6}\r
1386 bne .fl_loopcpBGR444_sh\r
1387\r
1388 @ hilighted pixels:\r
1389 sub r3, r3, #0x40*2\r
1390 mov lr, #0x40/2\r
1391.fl_loopcpBGR444_hi:\r
1392 ldr r1, [r3], #4\r
1393 TileDoShHi2Pixels444 r1\r
1394 str r1, [r4], #4\r
1395 subs lr, lr, #1\r
1396 bne .fl_loopcpBGR444_hi\r
1397\r
1398 sub r3, r4, #0x40*3*2\r
1399\r
1400\r
1401.fl_noshBGR444:\r
1402 ldr r1, =(HighCol+8)\r
1403 mov lr, #0xff\r
1404 mov lr, lr, lsl #1\r
1405\r
1406.fl_loopBGR444:\r
1407\r
1408 ldr r12, [r1], #4\r
1409 subs r2, r2, #1\r
1410\r
1411 and r4, lr, r12, lsl #1\r
1412 ldrh r4, [r3, r4]\r
1413 and r5, lr, r12, lsr #7\r
1414 ldrh r5, [r3, r5]\r
1415 and r6, lr, r12, lsr #15\r
1416 ldrh r6, [r3, r6]\r
1417 orr r4, r4, r5, lsl #16\r
1418\r
1419 and r5, lr, r12, lsr #23\r
1420 ldrh r5, [r3, r5] @ 2c.i.\r
1421 orr r5, r6, r5, lsl #16\r
1422\r
1423 stmia r0!, {r4,r5}\r
1424 bne .fl_loopBGR444\r
1425\r
1426\r
1427 ldmfd sp!, {r4-r6,lr}\r
1428 bx lr\r
1429\r
1430\r
1431@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1432\r
1433\r
1434@ hilights 2 pixels in RGB555/BGR555 format\r
1435.macro TileDoShHi2Pixels555 reg\r
1436 adds \reg, \reg, #0x40000000\r
1437 orrcs \reg, \reg, #0xf0000000\r
1438 mov \reg, \reg, ror #27\r
1439 adds \reg, \reg, #0x40000000\r
1440 orrcs \reg, \reg, #0xf0000000\r
1441 mov \reg, \reg, ror #26\r
1442 adds \reg, \reg, #0x40000000\r
1443 orrcs \reg, \reg, #0xf0000000\r
1444 mov \reg, \reg, ror #27\r
1445 adds \reg, \reg, #0x40000000\r
1446 orrcs \reg, \reg, #0xf0000000\r
1447 mov \reg, \reg, ror #27\r
1448 adds \reg, \reg, #0x40000000\r
1449 orrcs \reg, \reg, #0xf0000000\r
1450 mov \reg, \reg, ror #26\r
1451 adds \reg, \reg, #0x40000000\r
1452 orrcs \reg, \reg, #0xf0000000\r
1453 mov \reg, \reg, ror #27\r
1454.endm\r
1455\r
1456\r
1457@ Convert 0000bbb0 ggg0rrr0\r
1458@ to rrrrrggg gggbbbbb\r
1459\r
1460@ r2,r3,r9 - scratch, lr = 0x001c001c, r8 = 0x00030003\r
1461.macro convRGB565 reg\r
1462 and r2, lr, \reg,lsl #1\r
1463 and r9, r8, \reg,lsr #2\r
1464 orr r2, r2, r9 @ r2=red\r
1465 and r3, lr, \reg,lsr #7\r
1466 and r9, r8, \reg,lsr #10\r
1467 orr r3, r3, r9 @ r3=blue\r
1468 and \reg, \reg, lr, lsl #3\r
1469 orr \reg, \reg, \reg,lsl #3 @ green\r
1470 orr \reg, \reg, r2, lsl #11 @ add red back\r
1471 orr \reg, \reg, r3 @ add blue back\r
1472.endm\r
1473\r
1474vidConvCpyRGB565: @ void *to, void *from, int pixels\r
1475 stmfd sp!, {r4-r9,lr}\r
1476\r
1477 mov r12, r2, lsr #3 @ repeats\r
1478 mov lr, #0x001c0000\r
1479 orr lr, lr, #0x01c @ lr == pattern 0x001c001c\r
1480 mov r8, #0x00030000\r
1481 orr r8, r8, #0x003 @ lr == pattern 0x001c001c\r
1482\r
1483.loopRGB565:\r
1484 ldmia r1!, {r4-r7}\r
1485 subs r12, r12, #1\r
1486 convRGB565 r4\r
1487 str r4, [r0], #4\r
1488 convRGB565 r5\r
1489 str r5, [r0], #4\r
1490 convRGB565 r6\r
1491 str r6, [r0], #4\r
1492 convRGB565 r7\r
1493 str r7, [r0], #4\r
1494\r
1495 bgt .loopRGB565\r
1496\r
1497 ldmfd sp!, {r4-r9,lr}\r
1498 bx lr\r
1499\r
1500\r
1501\r
1502.global FinalizeLineRGB555 @ int sh\r
1503\r
1504FinalizeLineRGB555:\r
1505 stmfd sp!, {r4-r8,lr}\r
3d48f143 1506 ldr r8, =(Pico+0x22228) @ Pico.video\r
cc68a136 1507 ldr r4, =HighPal\r
1508\r
3d48f143 1509 ldrb r7, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
cc68a136 1510 mov r6, r0\r
1511 mov r1, #0\r
1512 tst r7, r7\r
1513 beq .fl_noconvRGB555\r
3d48f143 1514 strb r1, [r8, #-0x1a]\r
1515 sub r1, r8, #0x128 @ r1=Pico.cram\r
cc68a136 1516 mov r0, r4\r
1517 mov r2, #0x40\r
1518 bl vidConvCpyRGB565\r
1519\r
1520.fl_noconvRGB555:\r
cc68a136 1521 mov r3, r4\r
1522 tst r6, r6\r
1523 beq .fl_noshRGB555\r
1524 tst r7, r7\r
1525 beq .fl_noshRGB555\r
1526\r
1527 @ shadowed pixels:\r
1528 mov r12, #0x008e\r
1529 orr r12,r12,#0x7300\r
1530 orr r12,r12,r12,lsl #16\r
1531 add r4, r3, #0x40*2\r
1532 add r5, r3, #0xc0*2\r
1533 mov lr, #0x40/4\r
1534.fl_loopcpRGB555_sh:\r
1535 ldmia r3!, {r1,r6}\r
1536 subs lr, lr, #1\r
1537 and r1, r12, r1, lsr #1\r
1538 and r6, r12, r6, lsr #1\r
1539 stmia r4!, {r1,r6}\r
1540 stmia r5!, {r1,r6}\r
1541 bne .fl_loopcpRGB555_sh\r
1542\r
1543 @ hilighted pixels:\r
1544 sub r3, r3, #0x40*2\r
1545 mov lr, #0x40/2\r
1546.fl_loopcpRGB555_hi:\r
1547 ldr r1, [r3], #4\r
1548 TileDoShHi2Pixels555 r1\r
1549 str r1, [r4], #4\r
1550 subs lr, lr, #1\r
1551 bne .fl_loopcpRGB555_hi\r
1552\r
1553 sub r3, r3, #0x40*2\r
1554\r
cc68a136 1555.fl_noshRGB555:\r
3d48f143 1556 ldr r0, =DrawLineDest\r
cc68a136 1557 ldr r1, =(HighCol+8)\r
3d48f143 1558 ldr r0, [r0]\r
1559\r
1560 ldrb r12, [r8, #12]\r
cc68a136 1561 mov lr, #0xff\r
1562 mov lr, lr, lsl #1\r
1563\r
3d48f143 1564 tst r12, #1\r
1565 movne r2, #320/8 @ len\r
1566 bne .fl_no32colRGB555\r
1567 ldr r4, =PicoOpt\r
1568 mov r2, #256/8\r
1569 ldr r4, [r4]\r
1570 tst r4, #0x4000\r
1571 bne .fl_32scale_RGB555\r
1572 tst r4, #0x0100\r
1573 addeq r0, r0, #32*2\r
1574\r
1575.fl_no32colRGB555:\r
cc68a136 1576.fl_loopRGB555:\r
1577\r
1578 ldr r12, [r1], #4\r
1579 ldr r7, [r1], #4\r
1580\r
1581 and r4, lr, r12, lsl #1\r
1582 ldrh r4, [r3, r4]\r
1583 and r5, lr, r12, lsr #7\r
1584 ldrh r5, [r3, r5]\r
1585 and r6, lr, r12, lsr #15\r
1586 ldrh r6, [r3, r6]\r
1587 orr r4, r4, r5, lsl #16\r
1588\r
1589 and r5, lr, r12, lsr #23\r
1590 ldrh r5, [r3, r5]\r
1591 and r8, lr, r7, lsl #1\r
1592 ldrh r8, [r3, r8]\r
1593 orr r5, r6, r5, lsl #16\r
1594\r
1595 and r6, lr, r7, lsr #7\r
1596 ldrh r6, [r3, r6]\r
1597 and r12,lr, r7, lsr #15\r
1598 ldrh r12,[r3, r12]\r
1599 orr r8, r8, r6, lsl #16\r
1600\r
1601 and r6, lr, r7, lsr #23\r
1602 ldrh r6, [r3, r6] @ 1 cycle interlock here (r6)\r
1603 subs r2, r2, #1\r
1604 orr r12,r12, r6, lsl #16\r
1605\r
1606 stmia r0!, {r4,r5,r8,r12}\r
1607 bne .fl_loopRGB555\r
1608\r
3d48f143 1609 ldmfd sp!, {r4-r8,lr}\r
1610 bx lr\r
1611\r
1612\r
1613.fl_32scale_RGB555:\r
1614 stmfd sp!, {r9,r10}\r
1615 mov r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1616 orr r9, r9, #0x00e7\r
1617\r
1618.fl_loop32scale_RGB555:\r
1619 ldr r12, [r1], #4\r
1620 ldr r7, [r1], #4\r
cc68a136 1621\r
3d48f143 1622 and r4, lr, r12,lsl #1\r
1623 ldrh r4, [r3, r4]\r
1624 and r5, lr, r12,lsr #7\r
1625 ldrh r5, [r3, r5]\r
1626 and r4, r4, r9, lsl #2\r
1627 orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0\r
1628 and r5, r5, r9, lsl #2\r
1629 sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1\r
1630 add r4, r4, r6, lsl #16 @ pix_d 0, 1\r
1631 and r6, lr, r12,lsr #15\r
1632 ldrh r6, [r3, r6]\r
1633 and r12,lr, r12,lsr #23\r
1634 ldrh r12,[r3, r12]\r
1635 and r6, r6, r9, lsl #2\r
1636 add r5, r5, r6\r
1637 mov r5, r5, lsr #1\r
1638 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1639 orr r5, r5, r6, lsl #16\r
1640\r
1641 and r6, lr, r7, lsl #1\r
1642 ldrh r6, [r3, r6]\r
1643 and r12,r12,r9, lsl #2\r
1644 add r5, r5, r12,lsl #14 @ pix_d 2, 3\r
1645 and r6, r6, r9, lsl #2\r
1646 orr r6, r12,r6, lsl #16 @ pix_d 4, 5\r
1647\r
1648 and r12,lr, r7, lsr #7\r
1649 ldrh r12,[r3, r12]\r
1650 and r10,lr, r7, lsr #15\r
1651 ldrh r10,[r3, r10]\r
1652 and r12,r12,r9, lsl #2\r
1653 sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1\r
1654 add r8, r8, r6, lsr #18\r
1655 and r7, lr, r7, lsr #23\r
1656 ldrh r7, [r3, r7]\r
1657 and r10,r10,r9, lsl #2\r
1658 orr r8, r8, r10,lsl #15\r
1659 add r8, r8, r12,lsl #15 @ pix_d 6, 7\r
1660 sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2\r
1661 and r7, r7, r9, lsl #2\r
1662 add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3\r
1663 orr r10,r10,r7, lsl #16 @ pix_d 8, 9\r
1664\r
1665 subs r2, r2, #1\r
1666\r
1667 stmia r0!, {r4,r5,r6,r8,r10}\r
1668 bne .fl_loop32scale_RGB555\r
1669\r
1670 ldmfd sp!, {r9,r10}\r
cc68a136 1671 ldmfd sp!, {r4-r8,lr}\r
3d48f143 1672 bx lr\r
1673\r
cc68a136 1674\r
1675@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1676\r
1677@ utility\r
1678.global blockcpy @ void *dst, void *src, size_t n\r
1679\r
1680blockcpy:\r
1681 stmfd sp!, {r4,r5}\r
1682 mov r2, r2, lsr #4\r
1683blockcpy_loop:\r
1684 ldmia r1!, {r3-r5,r12}\r
1685 subs r2, r2, #1\r
1686 stmia r0!, {r3-r5,r12}\r
1687 bne blockcpy_loop\r
1688 ldmfd sp!, {r4,r5}\r
1689 bx lr\r
1690\r
1691\r
1692.global blockcpy_or @ void *dst, void *src, size_t n, int pat\r
1693\r
1694blockcpy_or:\r
1695 stmfd sp!, {r4-r6}\r
1696 orr r3, r3, r3, lsl #8\r
1697 orr r3, r3, r3, lsl #16\r
1698 mov r2, r2, lsr #4\r
1699blockcpy_loop_or:\r
1700 ldmia r1!, {r4-r6,r12}\r
1701 subs r2, r2, #1\r
1702 orr r4, r4, r3\r
1703 orr r5, r5, r3\r
1704 orr r6, r6, r3\r
1705 orr r12,r12,r3\r
1706 stmia r0!, {r4-r6,r12}\r
1707 bne blockcpy_loop_or\r
1708 ldmfd sp!, {r4-r6}\r
1709 bx lr\r
1710\r
fast MegaDrive/MegaCD/32X emulator