notaz.gp2x.de / picodrive.git / blame_incremental
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (non-incremental) | history | HEAD
remove some dead code
[picodrive.git] / pico / draw_arm.S
... / ...
CommitLineData
1/*\r
2 * assembly optimized versions of most funtions from draw.c\r
3 * (C) notaz, 2006-2010\r
4 *\r
5 * This work is licensed under the terms of MAME license.\r
6 * See COPYING file in the top-level directory.\r
7 *\r
8 * this is highly specialized, be careful if changing related C code!\r
9 */\r
10\r
11.extern Pico\r
12.extern PicoOpt\r
13.extern HighCol\r
14.extern DrawScanline\r
15.extern HighSprZ\r
16.extern rendstatus\r
17.extern HighPreSpr\r
18.extern DrawLineDest\r
19.extern DrawStripInterlace\r
20.extern HighCacheS_ptr\r
21\r
22.equiv OVERRIDE_HIGHCOL, 1\r
23\r
24.equ PDRAW_SPRITES_MOVED, (1<<0)\r
25.equ PDRAW_WND_DIFF_PRIO, (1<<1)\r
26.equ PDRAW_ACC_SPRITES, (1<<2)\r
27.equ PDRAW_DIRTY_SPRITES, (1<<4)\r
28.equ PDRAW_PLANE_HI_PRIO, (1<<6)\r
29.equ PDRAW_SHHI_DONE, (1<<7)\r
30\r
31@ helper\r
32.macro TilePixel pat lsrr offs\r
33.if !\lsrr\r
34 ands r4, \pat, r2\r
35.else\r
36 ands r4, \pat, r2, lsr #\lsrr\r
37.endif\r
38 orrne r4, r3, r4\r
39 strneb r4, [r1,#\offs]\r
40.endm\r
41\r
42@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
43.macro TileNorm pat\r
44 TilePixel \pat, 12, 0 @ #0x0000f000\r
45 TilePixel \pat, 8, 1 @ #0x00000f00\r
46 TilePixel \pat, 4, 2 @ #0x000000f0\r
47 TilePixel \pat, 0, 3 @ #0x0000000f\r
48 TilePixel \pat, 28, 4 @ #0xf0000000\r
49 TilePixel \pat, 24, 5 @ #0x0f000000\r
50 TilePixel \pat, 20, 6 @ #0x00f00000\r
51 TilePixel \pat, 16, 7 @ #0x000f0000\r
52.endm\r
53\r
54@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
55.macro TileFlip pat\r
56 TilePixel \pat, 16, 0 @ #0x000f0000\r
57 TilePixel \pat, 20, 1 @ #0x00f00000\r
58 TilePixel \pat, 24, 2 @ #0x0f000000\r
59 TilePixel \pat, 28, 3 @ #0xf0000000\r
60 TilePixel \pat, 0, 4 @ #0x0000000f\r
61 TilePixel \pat, 4, 5 @ #0x000000f0\r
62 TilePixel \pat, 8, 6 @ #0x00000f00\r
63 TilePixel \pat, 12, 7 @ #0x0000f000\r
64.endm\r
65\r
66@ shadow/hilight mode\r
67\r
68@ this one is for hi priority layer\r
69.macro TilePixelShHP lsrr offs\r
70.if !\lsrr\r
71 ands r4, r12, r2\r
72.else\r
73 ands r4, r12, r2, lsr #\lsrr\r
74.endif\r
75 ldreqb r4, [r1,#\offs]\r
76 orrne r4, r3, r4\r
77 andeq r4, r4, #0xbf\r
78 strb r4, [r1,#\offs]\r
79.endm\r
80\r
81@ TileNormShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
82.macro TileNormShHP\r
83 TilePixelShHP 12, 0 @ #0x0000f000\r
84 TilePixelShHP 8, 1 @ #0x00000f00\r
85 TilePixelShHP 4, 2 @ #0x000000f0\r
86 TilePixelShHP 0, 3 @ #0x0000000f\r
87 TilePixelShHP 28, 4 @ #0xf0000000\r
88 TilePixelShHP 24, 5 @ #0x0f000000\r
89 TilePixelShHP 20, 6 @ #0x00f00000\r
90 TilePixelShHP 16, 7 @ #0x000f0000\r
91.endm\r
92\r
93@ TileFlipShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
94.macro TileFlipShHP\r
95 TilePixelShHP 16, 0 @ #0x000f0000\r
96 TilePixelShHP 20, 1 @ #0x00f00000\r
97 TilePixelShHP 24, 2 @ #0x0f000000\r
98 TilePixelShHP 28, 3 @ #0xf0000000\r
99 TilePixelShHP 0, 4 @ #0x0000000f\r
100 TilePixelShHP 4, 5 @ #0x000000f0\r
101 TilePixelShHP 8, 6 @ #0x00000f00\r
102 TilePixelShHP 12, 7 @ #0x0000f000\r
103.endm\r
104\r
105\r
106@ TileSingleSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx; r12: helper pattern 0xf\r
107.macro TileSingleSh\r
108 tst r0, #1 @ not aligned?\r
109 mov r7, #0x00c000\r
110 orr r7, r7, #0xc0\r
111 ldrneb r4, [r1]\r
112 ldreqh r4, [r1]\r
113 orr r4, r4, r7\r
114 strneb r4, [r1], #1\r
115 streqh r4, [r1], #2\r
116 ldrh r4, [r1]\r
117 orr r4, r4, r7\r
118 strh r4, [r1], #2\r
119 ldrh r4, [r1]\r
120 orr r4, r4, r7\r
121 strh r4, [r1], #2\r
122 ldrh r4, [r1]\r
123 orr r4, r4, r7\r
124 strh r4, [r1], #2\r
125 ldrneb r4, [r1]\r
126 orr r4, r4, r7\r
127 strneb r4, [r1], #1\r
128.endm\r
129\r
130@ TileSingleHi (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
131.macro TileSingleHi\r
132 tst r1, #1 @ not aligned?\r
133 mov r7, #0x008000\r
134 orr r7, r7, #0x80\r
135 ldrneb r4, [r1], #1\r
136 ldreqh r4, [r1], #2 @ 1ci\r
137 ldrh r12, [r1], #2\r
138 bic r4, r4, r7, lsr #1\r
139 orr r4, r4, r7\r
140 strneb r4, [r1, #-3]\r
141 streqh r4, [r1, #-4]\r
142 ldrh r4, [r1], #2\r
143 bic r12, r12, r7, lsr #1\r
144 orr r12, r12, r7\r
145 strh r12, [r1, #-4]\r
146 ldrh r12, [r1], #2\r
147 bic r4, r4, r7, lsr #1\r
148 orr r4, r4, r7\r
149 strh r4, [r1, #-4]\r
150 ldrneb r4, [r1]\r
151 bic r12, r12, r7, lsr #1\r
152 orr r12, r12, r7\r
153 strh r12, [r1, #-2]\r
154 bicne r4, r4, r7, lsr #1\r
155 orrne r4, r4, r7\r
156 strneb r4, [r1], #1\r
157 mov r12, #0xf\r
158.endm\r
159\r
160.macro TileDoShGenPixel shift ofs\r
161.if \shift\r
162 ands r4, r12, r2, lsr #\shift\r
163.else\r
164 ands r4, r12, r2\r
165.endif\r
166 beq 0f\r
167 cmp r4, #0xe\r
168 ldrgeb r7, [r1,#\ofs]\r
169 orrlt r7, r3, r4 @ normal\r
170\r
171 bicge r7, r7, #0xc0\r
172 orrge r7, r7, r4, lsl #6\r
173 strb r7, [r1,#\ofs]\r
1740:\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.macro TileDoShGenPixel_markop shift ofs\r
202.if \shift\r
203 ands r4, r12, r2, lsr #\shift\r
204.else\r
205 ands r4, r12, r2\r
206.endif\r
207 beq 0f\r
208 cmp r4, #0xe\r
209 ldrgeb r4, [r1,#\ofs]\r
210 orrlt r4, r3, r4\r
211 orrge r4, r4, #0x80\r
212 strb r4, [r1,#\ofs]\r
2130:\r
214.endm\r
215\r
216.macro TileFlipSh_markop\r
217 TileDoShGenPixel_markop 16, 0 @ #0x000f0000\r
218 TileDoShGenPixel_markop 20, 1 @ #0x00f00000\r
219 TileDoShGenPixel_markop 24, 2 @ #0x0f000000\r
220 TileDoShGenPixel_markop 28, 3 @ #0xf0000000\r
221 TileDoShGenPixel_markop 0, 4 @ #0x0000000f\r
222 TileDoShGenPixel_markop 4, 5 @ #0x000000f0\r
223 TileDoShGenPixel_markop 8, 6 @ #0x00000f00\r
224 TileDoShGenPixel_markop 12, 7 @ #0x0000f000\r
225.endm\r
226\r
227.macro TileNormSh_markop\r
228 TileDoShGenPixel_markop 12, 0 @ #0x0000f000\r
229 TileDoShGenPixel_markop 8, 1 @ #0x00000f00\r
230 TileDoShGenPixel_markop 4, 2 @ #0x000000f0\r
231 TileDoShGenPixel_markop 0, 3 @ #0x0000000f\r
232 TileDoShGenPixel_markop 28, 4 @ #0xf0000000\r
233 TileDoShGenPixel_markop 24, 5 @ #0x0f000000\r
234 TileDoShGenPixel_markop 20, 6 @ #0x00f00000\r
235 TileDoShGenPixel_markop 16, 7 @ #0x000f0000\r
236.endm\r
237\r
238.macro TileDoShGenPixel_onlyop_lp shift ofs\r
239.if \shift\r
240 ands r7, r12, r2, lsr #\shift\r
241.else\r
242 ands r7, r12, r2\r
243.endif\r
244 ldrneb r4, [r1,#\ofs]\r
245 cmp r7, #0xe\r
246 blt 0f\r
247\r
248 tst r4, #0xc0\r
249 bicne r4, r4, #0xc0\r
250 orrne r4, r4, r7, lsl #6\r
251 strneb r4, [r1,#\ofs]\r
2520:\r
253.endm\r
254\r
255.macro TileFlipSh_onlyop_lp\r
256 TileDoShGenPixel_onlyop_lp 16, 0 @ #0x000f0000\r
257 TileDoShGenPixel_onlyop_lp 20, 1 @ #0x00f00000\r
258 TileDoShGenPixel_onlyop_lp 24, 2 @ #0x0f000000\r
259 TileDoShGenPixel_onlyop_lp 28, 3 @ #0xf0000000\r
260 TileDoShGenPixel_onlyop_lp 0, 4 @ #0x0000000f\r
261 TileDoShGenPixel_onlyop_lp 4, 5 @ #0x000000f0\r
262 TileDoShGenPixel_onlyop_lp 8, 6 @ #0x00000f00\r
263 TileDoShGenPixel_onlyop_lp 12, 7 @ #0x0000f000\r
264.endm\r
265\r
266.macro TileNormSh_onlyop_lp\r
267 TileDoShGenPixel_onlyop_lp 12, 0 @ #0x0000f000\r
268 TileDoShGenPixel_onlyop_lp 8, 1 @ #0x00000f00\r
269 TileDoShGenPixel_onlyop_lp 4, 2 @ #0x000000f0\r
270 TileDoShGenPixel_onlyop_lp 0, 3 @ #0x0000000f\r
271 TileDoShGenPixel_onlyop_lp 28, 4 @ #0xf0000000\r
272 TileDoShGenPixel_onlyop_lp 24, 5 @ #0x0f000000\r
273 TileDoShGenPixel_onlyop_lp 20, 6 @ #0x00f00000\r
274 TileDoShGenPixel_onlyop_lp 16, 7 @ #0x000f0000\r
275.endm\r
276\r
277\r
278@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
279\r
280@ struct TileStrip\r
281@ {\r
282@ int nametab; // 0x00\r
283@ int line; // 0x04\r
284@ int hscroll; // 0x08\r
285@ int xmask; // 0x0C\r
286@ int *hc; // 0x10 (pointer to cache buffer)\r
287@ int cells; // 0x14\r
288@ };\r
289\r
290@ void DrawLayer(int plane_sh, int *hcache, int cellskip, int maxcells);\r
291\r
292.global DrawLayer\r
293\r
294DrawLayer:\r
295 stmfd sp!, {r4-r11,lr}\r
296\r
297 ldr r11, =(Pico+0x22228) @ Pico.video\r
298 mov r8, #1\r
299\r
300 ldrb r7, [r11, #16] @ ??vv??hh\r
301\r
302 mov r6, r1 @ hcache\r
303 orr r9, r3, r0, lsl #30\r
304 orr r9, r9, r2, lsl #8 @ r9=sh[31]|cellskip[15:8]|maxcells[7:0] (tmp)\r
305\r
306 mov r1, r7, lsl #4\r
307 orr r1, r1, #0x00ff\r
308\r
309 and r10, r7, #3\r
310 cmp r10, #1\r
311 biclt r1, r1, #0xfc00\r
312 biceq r1, r1, #0xfe00\r
313 bicgt r1, r1, #0xff00 @ r1=ymask=(height<<8)|0xff; ...; // Y Mask in pixels\r
314\r
315 add r10, r10, #5\r
316 cmp r10, #7\r
317 subge r10, r10, #1 @ r10=shift[width] (5,6,6,7)\r
318\r
319 @ calculate xmask:\r
320 mov r5, r8, lsl r10\r
321 sub r5, r5, #1 @ r5=xmask\r
322\r
323 @ Find name table:\r
324 ands r0, r0, #1\r
325 ldreqb r12, [r11, #2]\r
326 ldrneb r12, [r11, #4]\r
327\r
328 ldr r2, =DrawScanline @ trying to make good use of pipeline here\r
329 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
330\r
331 moveq r12, r12, lsl #10\r
332 movne r12, r12, lsl #13\r
333 and r12, r12, #(7<<13) @ r12=(ts->nametab<<1) (halfword compliant)\r
334\r
335 ldrh r8, [r11, #12]\r
336 ldrb r7, [r11, #11]\r
337 ldr r2, [r2]\r
338\r
339 mov r4, r8, lsr #8 @ pvid->reg[13]\r
340 mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r
341 tst r7, #2\r
342 addne r4, r4, r2, lsl #2 @ htab+=DrawScanline<<1; // Offset by line\r
343 tst r7, #1\r
344 biceq r4, r4, #0x1f @ htab&=~0xf; // Offset by tile\r
345 add r4, r4, r0, lsl #1 @ htab+=plane\r
346 bic r4, r4, #0x00ff0000 @ just in case\r
347 ldrh r3, [lr, r4] @ r3=hscroll\r
348\r
349 tst r7, #4\r
350 bne .DrawStrip_vsscroll\r
351\r
352 @ Get vertical scroll value:\r
353 add r7, lr, #0x012000\r
354 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
355 ldr r7, [r7]\r
356\r
357 tst r8, #2\r
358 tstne r8, #4\r
359 bne .DrawStrip_interlace\r
360\r
361 tst r0, r0\r
362 movne r7, r7, lsr #16\r
363\r
364 @ Find the line in the name table\r
365 add r2, r2, r7\r
366 and r2, r2, r1\r
367 mov r4, r2, lsr #3\r
368 add r10, r10, #1 @ shift[width]++\r
369 add r12, r12, r4, lsl r10 @ nametab+=(ts.line>>3)<<shift[width];\r
370\r
371 @ ldmia r0, {r1,r2,r3,r5,r6,r9} @ r2=line, r3=ts->hscroll, r5=ts->xmask, r6=ts->hc, r9=ts->cells\r
372\r
373 and r10,r2, #7\r
374 mov r10,r10, lsl #1 @ r10=ty=(ts->line&7)<<1;\r
375 orr r10,r10, r9, lsl #24\r
376\r
377 rsb r8, r3, #0\r
378 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
379\r
380 sub r1, r3, #1\r
381 and r1, r1, #7\r
382 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
383\r
384 tst r9, #1<<31\r
385 mov r3, #0\r
386 orrne r10,r10, #1<<23 @ r10=(cells<<24|sh<<23|hi_not_empty<<22|had_output<<21|ty)\r
387 movne r3, #0x40 @ default to shadowed pal on sh mode\r
388\r
389 cmp r7, #8\r
390 addne r10,r10, #0x01000000 @ we will loop cells+1 times if there is scroll\r
391\r
392 and r9, r9, #0xff00\r
393 add r8, r8, r9, lsr #8 @ tilex+=cellskip\r
394 add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;\r
395 sub r10,r10,r9, lsl #16 @ cells-=cellskip\r
396\r
397 @ cache some stuff to avoid mem access\r
398.if OVERRIDE_HIGHCOL\r
399 ldr r11,=HighCol\r
400 mov r0, #0xf\r
401 ldr r11,[r11]\r
402.else\r
403 ldr r11,=HighCol\r
404 mov r0, #0xf\r
405.endif\r
406\r
407 mvn r9, #0 @ r9=prevcode=-1\r
408 add r1, r11, r7 @ r1=pdest\r
409\r
410\r
411 @ r4 & r7 are scratch in this loop\r
412.dsloop_subr1:\r
413 sub r1, r1, #8\r
414.dsloop: @ 40-41 times\r
415 subs r10,r10, #0x01000000\r
416 bmi .dsloop_exit\r
417\r
418.dsloop_enter:\r
419 and r7, r5, r8\r
420 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
421 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
422\r
423 add r1, r1, #8\r
424 add r8, r8, #1\r
425\r
426 tst r7, #0x8000\r
427 bne .DrawStrip_hiprio\r
428\r
429 cmp r7, r9\r
430 beq .DrawStrip_samecode @ we know stuff about this tile already\r
431\r
432 mov r9, r7 @ remember code\r
433 orr r10, r10, #1<<21 @ seen non hi-prio tile\r
434\r
435 movs r2, r9, lsl #20 @ if (code&0x1000)\r
436 mov r2, r2, lsl #1\r
437 add r2, r2, r10, lsl #17\r
438 mov r2, r2, lsr #17\r
439 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
440\r
441 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
442\r
443 bic r7, r3, #0x3f\r
444 and r3, r9, #0x6000\r
445 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
446\r
447.DrawStrip_samecode:\r
448 tst r2, r2\r
449 beq .dsloop @ tileline blank\r
450\r
451 cmp r2, r2, ror #4\r
452 beq .DrawStrip_SingleColor @ tileline singlecolor \r
453\r
454 tst r9, #0x0800\r
455 bne .DrawStrip_TileFlip\r
456\r
457 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
458.DrawStrip_TileNorm:\r
459 TileNorm r0\r
460 b .dsloop\r
461\r
462.DrawStrip_TileFlip:\r
463 TileFlip r0\r
464 b .dsloop\r
465\r
466.DrawStrip_SingleColor:\r
467 and r4, r2, #0xf\r
468 orr r4, r3, r4\r
469 orr r4, r4, r4, lsl #8\r
470 tst r1, #1 @ not aligned?\r
471 strneb r4, [r1], #1\r
472 streqh r4, [r1], #2\r
473 strh r4, [r1], #2\r
474 strh r4, [r1], #2\r
475 strh r4, [r1], #2\r
476 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
477 b .dsloop_subr1\r
478\r
479.DrawStrip_hiprio_maybempt:\r
480 cmp r7, r9\r
481 beq .dsloop @ must've been empty, otherwise we wouldn't get here\r
482 movs r2, r7, lsl #20 @ if (code&0x1000)\r
483 mov r2, r2, lsl #1\r
484 add r2, r2, r10, lsl #17\r
485 mov r2, r2, lsr #17\r
486 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
487 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
488 mov r9, r7 @ remember code\r
489 tst r2, r2\r
490 beq .dsloop\r
491 orr r10, r10, #1<<22\r
492\r
493.DrawStrip_hiprio:\r
494 tst r10, #0x00c00000\r
495 beq .DrawStrip_hiprio_maybempt\r
496 sub r0, r1, r11\r
497 orr r7, r7, r0, lsl #16\r
498 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
499 tst r7, #0x1000\r
500 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
501 str r7, [r6], #4 @ cache hi priority tile\r
502 mov r0, #0xf\r
503 b .dsloop\r
504\r
505.dsloop_exit:\r
506 tst r10, #1<<21 @ seen non hi-prio tile\r
507 ldreq r1, =rendstatus\r
508 mov r0, #0\r
509 ldreq r2, [r1]\r
510 str r0, [r6] @ terminate the cache list\r
511 orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
512 streq r2, [r1]\r
513\r
514 ldmfd sp!, {r4-r11,lr}\r
515 bx lr\r
516\r
517@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
518\r
519.DrawStrip_vsscroll:\r
520 rsb r8, r3, #0\r
521 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
522 bic r8, r8, #0x3fc00000\r
523 orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|tilex[21:0])\r
524\r
525 ldr r4, =DrawScanline\r
526 orr r5, r1, r10, lsl #24\r
527 ldr r4, [r4]\r
528 sub r1, r3, #1\r
529 orr r5, r5, r4, lsl #16 @ r5=(shift_width[31:24]|scanline[23:16]|ymask[15:0])\r
530 and r1, r1, #7\r
531 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
532\r
533 mov r10,r9, lsl #16\r
534 tst r0, #1\r
535 orrne r10,r10, #0x8000\r
536 tst r9, #1<<31\r
537 mov r3, #0\r
538 orr r10,r10, #0xff000000 @ will be adjusted on entering loop\r
539 orrne r10,r10, #1<<23 @ r10=(cell[31:24]|sh[23]|hi_not_empty[22]|cells_max[21:16]|plane[15]|ty[14:0])\r
540 movne r3, #0x40 @ default to shadowed pal on sh mode\r
541\r
542 cmp r7, #8\r
543 subne r10,r10, #0x01000000 @ have hscroll, start with negative cell\r
544\r
545 and r9, r9, #0xff00\r
546 add r8, r8, r9, lsr #8 @ tilex+=cellskip\r
547 add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;\r
548 add r10,r10,r9, lsl #16 @ cell+=cellskip\r
549\r
550 @ cache some stuff to avoid mem access\r
551.if OVERRIDE_HIGHCOL\r
552 ldr r11,=HighCol\r
553 mov r0, #0xf\r
554 ldr r11,[r11]\r
555.else\r
556 ldr r11,=HighCol\r
557 mov r0, #0xf\r
558.endif\r
559\r
560 mvn r9, #0 @ r9=prevcode=-1\r
561 add r1, r11, r7 @ r1=pdest\r
562\r
563 @ r4 & r7 are scratch in this loop\r
564.dsloop_vs_subr1:\r
565 sub r1, r1, #8\r
566.dsloop_vs: @ 40-41 times\r
567 add r10,r10, #0x01000000\r
568 and r4, r10, #0x003f0000\r
569 cmp r4, r10, asr #8\r
570 ble .dsloop_vs_exit\r
571\r
572 @ calc offset and read tileline code to r7, also calc ty\r
573 add r7, lr, #0x012000\r
574 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
575 add r7, r7, r10,asr #23 @ vsram + ((cell&~1)<<1)\r
576 bic r7, r7, #3\r
577 tst r10,#0x8000 @ plane1?\r
578 addne r7, r7, #2\r
579 ldrh r7, [r7] @ r7=vscroll\r
580\r
581 bic r10,r10,#0xff @ clear old ty\r
582 and r4, r5, #0xff0000 @ scanline\r
583 add r4, r4, r7, lsl #16 @ ... += vscroll\r
584 and r4, r4, r5, lsl #16 @ ... &= ymask\r
585 and r7, r4, #0x70000\r
586 orr r10,r10,r7, lsr #15 @ new ty\r
587\r
588 mov r4, r4, lsr #19\r
589 mov r7, r5, lsr #24\r
590 mov r4, r4, lsl r7 @ nametabadd\r
591\r
592 and r7, r8, r8, lsr #25\r
593 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
594 add r7, r7, r4, lsl #1\r
595 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
596\r
597 add r1, r1, #8\r
598 add r8, r8, #1\r
599\r
600 tst r7, #0x8000\r
601 bne .DrawStrip_vs_hiprio\r
602\r
603 cmp r7, r9\r
604 beq .DrawStrip_vs_samecode @ we know stuff about this tile already\r
605\r
606 mov r9, r7 @ remember code\r
607 orr r8, r8, #(1<<24)@ seen non hi-prio tile\r
608\r
609 movs r2, r9, lsl #20 @ if (code&0x1000)\r
610 mov r2, r2, lsl #1\r
611 add r2, r2, r10, lsl #17\r
612 mov r2, r2, lsr #17\r
613 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
614\r
615 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
616\r
617 bic r7, r3, #0x3f\r
618 and r3, r9, #0x6000\r
619 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
620\r
621.DrawStrip_vs_samecode:\r
622 tst r2, r2\r
623 beq .dsloop_vs @ tileline blank\r
624\r
625 cmp r2, r2, ror #4\r
626 beq .DrawStrip_vs_SingleColor @ tileline singlecolor \r
627\r
628 tst r9, #0x0800\r
629 bne .DrawStrip_vs_TileFlip\r
630\r
631 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
632.DrawStrip_vs_TileNorm:\r
633 TileNorm r0\r
634 b .dsloop_vs\r
635\r
636.DrawStrip_vs_TileFlip:\r
637 TileFlip r0\r
638 b .dsloop_vs\r
639\r
640.DrawStrip_vs_SingleColor:\r
641 and r4, r2, #0xf\r
642 orr r4, r3, r4\r
643 orr r4, r4, r4, lsl #8\r
644 tst r1, #1 @ not aligned?\r
645 strneb r4, [r1], #1\r
646 streqh r4, [r1], #2\r
647 strh r4, [r1], #2\r
648 strh r4, [r1], #2\r
649 strh r4, [r1], #2\r
650 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
651 b .dsloop_vs_subr1\r
652\r
653.DrawStrip_vs_hiprio:\r
654 tst r10, #0x00c00000\r
655 beq .DrawStrip_vs_hiprio_maybempt\r
656 sub r0, r1, r11\r
657 orr r7, r7, r0, lsl #16\r
658 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
659 tst r7, #0x1000\r
660 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
661 str r7, [r6], #4 @ cache hi priority tile\r
662 mov r0, #0xf\r
663 b .dsloop_vs\r
664\r
665.DrawStrip_vs_hiprio_maybempt:\r
666 cmp r7, r9\r
667 beq .dsloop_vs @ must've been empty, otherwise we wouldn't get here\r
668 movs r2, r7, lsl #20 @ if (code&0x1000)\r
669 mov r2, r2, lsl #1\r
670 add r2, r2, r10, lsl #17\r
671 mov r2, r2, lsr #17\r
672 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
673 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
674 mov r9, r7 @ remember code\r
675 tst r2, r2\r
676 orrne r10, r10, #1<<22\r
677 bne .DrawStrip_vs_hiprio\r
678 b .dsloop_vs\r
679\r
680.dsloop_vs_exit:\r
681 tst r8, #(1<<24) @ seen non hi-prio tile\r
682 ldreq r1, =rendstatus\r
683 mov r0, #0\r
684 ldreq r2, [r1]\r
685 str r0, [r6] @ terminate the cache list\r
686 orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
687 streq r2, [r1]\r
688\r
689 ldmfd sp!, {r4-r11,lr}\r
690 bx lr\r
691\r
692\r
693@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
694\r
695@ interlace mode 2? Sonic 2?\r
696.DrawStrip_interlace:\r
697 tst r0, r0\r
698 moveq r7, r7, lsl #21\r
699 movne r7, r7, lsl #5\r
700\r
701 @ Find the line in the name table\r
702 add r2, r7, r2, lsl #22 @ r2=(vscroll+(DrawScanline<<1))<<21 (11 bits);\r
703 orr r1, r1, #0x80000000\r
704 and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
705 mov r2, r2, lsr #21\r
706 mov r4, r2, lsr #4\r
707 mov r12, r12, lsr #1 @ halfwords\r
708 add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];\r
709 and r9, r9, #0xff\r
710\r
711 sub sp, sp, #6*4\r
712 stmia sp, {r0,r2,r3,r5,r6,r9}\r
713\r
714 mov r0, sp\r
715 bl DrawStripInterlace @ struct TileStrip *ts\r
716\r
717 add sp, sp, #6*4\r
718 ldmfd sp!, {r4-r11,lr}\r
719 bx lr\r
720\r
721.pool\r
722\r
723@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
724\r
725\r
726.global BackFill @ int reg7, int sh\r
727\r
728BackFill:\r
729 stmfd sp!, {r4-r9,lr}\r
730\r
731.if OVERRIDE_HIGHCOL\r
732 ldr lr, =HighCol\r
733 mov r0, r0, lsl #26\r
734 ldr lr, [lr]\r
735 mov r0, r0, lsr #26\r
736 add lr, lr, #8\r
737.else\r
738 ldr lr, =(HighCol+8)\r
739 mov r0, r0, lsl #26\r
740 mov r0, r0, lsr #26\r
741.endif\r
742\r
743 orr r0, r0, r1, lsl #6\r
744 orr r0, r0, r0, lsl #8\r
745 orr r0, r0, r0, lsl #16\r
746\r
747 mov r1, r0\r
748 mov r2, r0\r
749 mov r3, r0\r
750 mov r4, r0\r
751 mov r5, r0\r
752 mov r6, r0\r
753 mov r7, r0\r
754\r
755 @ go go go!\r
756 stmia lr!, {r0-r7} @ 10*8*4\r
757 stmia lr!, {r0-r7}\r
758 stmia lr!, {r0-r7}\r
759 stmia lr!, {r0-r7}\r
760 stmia lr!, {r0-r7}\r
761 stmia lr!, {r0-r7}\r
762 stmia lr!, {r0-r7}\r
763 stmia lr!, {r0-r7}\r
764 stmia lr!, {r0-r7}\r
765 stmia lr!, {r0-r7}\r
766\r
767 ldmfd sp!, {r4-r9,r12}\r
768 bx r12\r
769\r
770\r
771@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
772\r
773\r
774.global DrawTilesFromCache @ int *hc, int sh, int rlim\r
775\r
776DrawTilesFromCache:\r
777 stmfd sp!, {r4-r8,r11,lr}\r
778\r
779 @ cache some stuff to avoid mem access\r
780.if OVERRIDE_HIGHCOL\r
781 ldr r11,=HighCol\r
782 mov r12,#0xf\r
783 ldr r11,[r11]\r
784.else\r
785 ldr r11,=HighCol\r
786 mov r12,#0xf\r
787.endif\r
788 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
789\r
790 mvn r5, #0 @ r5=prevcode=-1\r
791 ands r8, r1, #1\r
792 orr r8, r8, r2, lsl #1\r
793 bne .dtfc_check_rendflags\r
794\r
795 @ scratch: r4, r7\r
796.dtfc_loop:\r
797 ldr r6, [r0], #4 @ read code\r
798 movs r1, r6, lsr #16 @ r1=dx;\r
799 ldmeqfd sp!, {r4-r8,r11,pc} @ dx is never zero, this must be a terminator, return\r
800 bic r4, r1, #0xfe00\r
801 add r1, r11, r4 @ r1=pdest\r
802\r
803 mov r7, r6, lsl #16\r
804 cmp r5, r7, lsr #16\r
805 beq .dtfc_samecode @ if (code==prevcode)\r
806\r
807 mov r5, r7, lsr #16\r
808\r
809 mov r2, r5, lsl #21\r
810 mov r2, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4;\r
811 add r2, r2, r6, lsr #25 @ addr+=ty\r
812\r
813 and r3, r5, #0x6000\r
814 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
815\r
816 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
817\r
818.dtfc_samecode:\r
819 rsbs r4, r4, r8, lsr #1\r
820 bmi .dtfc_cut_tile\r
821\r
822 tst r8, #1\r
823 bne .dtfc_shadow\r
824\r
825 tst r2, r2\r
826 beq .dtfc_loop\r
827\r
828 cmp r2, r2, ror #4\r
829 beq .dtfc_SingleColor @ tileline singlecolor \r
830\r
831 tst r5, #0x0800\r
832 bne .dtfc_TileFlip\r
833\r
834 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
835.dtfc_TileNorm:\r
836 TileNorm r12\r
837 b .dtfc_loop\r
838\r
839.dtfc_TileFlip:\r
840 TileFlip r12\r
841 b .dtfc_loop\r
842\r
843.dtfc_SingleColor:\r
844 and r4, r2, #0xf\r
845 orr r4, r3, r4\r
846 orr r4, r4, r4, lsl #8\r
847 tst r1, #1 @ not aligned?\r
848 strneb r4, [r1], #1\r
849 streqh r4, [r1], #2\r
850 strh r4, [r1], #2\r
851 strh r4, [r1], #2\r
852 strh r4, [r1], #2\r
853 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
854 b .dtfc_loop\r
855\r
856.dtfc_shadow:\r
857 tst r2, r2\r
858 beq .dtfc_shadow_blank\r
859\r
860 cmp r2, r2, ror #4\r
861 beq .dtfc_SingleColor @ tileline singlecolor \r
862\r
863 tst r5, #0x0800\r
864 bne .dtfc_TileFlipShHP\r
865\r
866 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
867.dtfc_TileNormShHP:\r
868 TileNormShHP\r
869 b .dtfc_loop\r
870\r
871.dtfc_TileFlipShHP:\r
872 TileFlipShHP\r
873 b .dtfc_loop\r
874\r
875.dtfc_shadow_blank:\r
876 tst r1, #1\r
877 ldrneb r4, [r1]\r
878 mov r6, #0xbf\r
879 and r4, r4, #0xbf\r
880 strneb r4, [r1], #1\r
881 ldrh r4, [r1]\r
882 orr r6, r6, r6, lsl #8\r
883 and r4, r4, r6\r
884 strh r4, [r1], #2\r
885 ldrh r4, [r1]\r
886 and r4, r4, r6\r
887 strh r4, [r1], #2\r
888 ldrh r4, [r1]\r
889 and r4, r4, r6\r
890 strh r4, [r1], #2\r
891 ldrh r4, [r1]\r
892 and r4, r4, r6\r
893 streqh r4, [r1]\r
894 strneb r4, [r1]\r
895 b .dtfc_loop\r
896\r
897.dtfc_cut_tile:\r
898 add r4, r4, #7 @ 0-6\r
899 mov r4, r4, lsl #2\r
900 mov r12,#0xf<<28\r
901 mov r12,r12,asr r4\r
902 mov r2, r2, ror #16\r
903 tst r5, #0x0800 @ flipped?\r
904 mvnne r12,r12\r
905 and r2, r2, r12\r
906 mov r2, r2, ror #16\r
907 mov r12,#0xf\r
908 tst r8, #1\r
909 bne .dtfc_shadow\r
910 tst r2, r2\r
911 beq .dtfc_loop\r
912 tst r5, #0x0800\r
913 beq .dtfc_TileNorm\r
914 b .dtfc_TileFlip\r
915\r
916@ check if we have detected layer covered with hi-prio tiles:\r
917.dtfc_check_rendflags:\r
918 ldr r1, =rendstatus\r
919 ldr r2, [r1]\r
920 tst r2, #(PDRAW_PLANE_HI_PRIO|PDRAW_SHHI_DONE)\r
921 beq .dtfc_loop\r
922 bic r8, r8, #1 @ sh/hi mode off\r
923 tst r2, #PDRAW_SHHI_DONE\r
924 bne .dtfc_loop @ already processed\r
925 orr r2, r2, #PDRAW_SHHI_DONE\r
926 str r2, [r1]\r
927\r
928 add r1, r11,#8\r
929 mov r3, #320/4/4\r
930 mov r6, #0xbf\r
931 orr r6, r6, r6, lsl #8\r
932 orr r6, r6, r6, lsl #16\r
933.dtfc_loop_shprep:\r
934 ldmia r1, {r2,r4,r5,r7}\r
935 subs r3, r3, #1\r
936 and r2, r2, r6\r
937 and r4, r4, r6\r
938 and r5, r5, r6\r
939 and r7, r7, r6\r
940 stmia r1!,{r2,r4,r5,r7}\r
941 bne .dtfc_loop_shprep\r
942\r
943 mvn r5, #0 @ r5=prevcode=-1\r
944 b .dtfc_loop\r
945\r
946.pool\r
947\r
948@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
949\r
950\r
951.global DrawSpritesSHi @ unsigned char *sprited\r
952\r
953DrawSpritesSHi:\r
954 ldr r3, [r0]\r
955 mov r12,#0xff\r
956 ands r3, r3, #0x7f\r
957 bxeq lr\r
958\r
959 stmfd sp!, {r4-r11,lr}\r
960 strb r12,[r0,#2] @ set end marker\r
961 add r10,r0, #3 @ r10=HighLnSpr end\r
962 add r10,r10,r3 @ r10=HighLnSpr end\r
963\r
964.if OVERRIDE_HIGHCOL\r
965 ldr r11,=HighCol\r
966 mov r12,#0xf\r
967 ldr r11,[r11]\r
968.else\r
969 ldr r11,=HighCol\r
970 mov r12,#0xf\r
971.endif\r
972 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
973\r
974\r
975DrawSpriteSHi:\r
976 @ draw next sprite\r
977 ldrb r0, [r10,#-1]!\r
978 ldr r1, =HighPreSpr\r
979@ ldr r8, [sp, #-4]\r
980 cmp r0, #0xff\r
981 ldmeqfd sp!, {r4-r11,pc} @ end of list\r
982 and r0, r0, #0x7f\r
983 add r0, r1, r0, lsl #3\r
984\r
985 ldr r9, [r0, #4] @ sprite[1]\r
986 mov r2, r9, asr #16 @ r2=sx\r
987\r
988 mov r9, r9, lsl #16\r
989 mov r3, r9, lsr #31 @ priority\r
990 mov r9, r9, lsr #16\r
991@ orr r9, r9, r8, lsl #31 @ r9=code|sh[31] @@ sh is always on here now\r
992 and r4, r9, #0x6000\r
993 orr r9, r9, r4, lsl #16\r
994 orr r9, r9, #0x90000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
995 cmp r12,r9, lsr #28 @ sh/hi with pal3?\r
996 cmpne r3, #1 @ if not, is it hi prio?\r
997 bne DrawSpriteSHi @ non-operator low sprite, already drawn\r
998\r
999 ldr r3, [r0] @ sprite[0]\r
1000 ldr r7, =DrawScanline\r
1001 mov r6, r3, lsr #28\r
1002 sub r6, r6, #1 @ r6=width-1 (inc later)\r
1003 mov r5, r3, lsr #24\r
1004 and r5, r5, #7 @ r5=height\r
1005\r
1006 mov r0, r3, lsl #16 @ r4=sy<<16 (tmp)\r
1007\r
1008 ldr r7, [r7]\r
1009 sub r7, r7, r0, asr #16 @ r7=row=DrawScanline-sy\r
1010\r
1011 tst r9, #0x1000\r
1012 movne r0, r5, lsl #3\r
1013 subne r0, r0, #1\r
1014 subne r7, r0, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
1015\r
1016 add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
1017 tst r9, #0x0800\r
1018 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1019 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
1020\r
1021 mov r8, r8, lsl #21\r
1022 mov r8, r8, lsr #17\r
1023 and r7, r7, #7\r
1024 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1025\r
1026 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
1027 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
1028\r
1029 add r6, r6, #1 @ inc now\r
1030 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1031 b .dsprShi_loop_enter\r
1032\r
1033.dsprShi_loop:\r
1034 subs r6, r6, #1 @ width--\r
1035 beq DrawSpriteSHi\r
1036 adds r0, r0, #8 @ sx+=8\r
1037 add r8, r8, r5 @ tile+=delta\r
1038\r
1039.dsprShi_loop_enter:\r
1040 ble .dsprShi_loop @ sx <= 0\r
1041 cmp r0, #328\r
1042 bge DrawSpriteSHi\r
1043\r
1044 mov r8, r8, lsl #17\r
1045 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
1046\r
1047 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1048 add r1, r11, r0 @ r1=pdest\r
1049 tst r2, r2\r
1050 beq .dsprShi_loop\r
1051\r
1052 cmp r12, r9, lsr #28\r
1053 beq .dsprShi_shadow\r
1054\r
1055 cmp r2, r2, ror #4\r
1056 beq .dsprShi_SingleColor @ tileline singlecolor \r
1057\r
1058 tst r9, #0x0800\r
1059 bne .dsprShi_TileFlip\r
1060\r
1061 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1062@ scratch: r4, r7\r
1063.dsprShi_TileNorm:\r
1064 TileNorm r12\r
1065 b .dsprShi_loop\r
1066\r
1067.dsprShi_TileFlip:\r
1068 TileFlip r12\r
1069 b .dsprShi_loop\r
1070\r
1071.dsprShi_SingleColor:\r
1072 and r4, r2, #0xf\r
1073 orr r4, r3, r4\r
1074 orr r4, r4, r4, lsl #8\r
1075 tst r0, #1 @ not aligned?\r
1076 strneb r4, [r1], #1\r
1077 streqh r4, [r1], #2\r
1078 strh r4, [r1], #2\r
1079 strh r4, [r1], #2\r
1080 strh r4, [r1], #2\r
1081 strneb r4, [r1], #1\r
1082 b .dsprShi_loop\r
1083\r
1084.dsprShi_shadow:\r
1085 tst r9, #0x8000\r
1086 beq .dsprShi_shadow_lowpri\r
1087\r
1088 cmp r2, r2, ror #4\r
1089 beq .dsprShi_singlec_sh\r
1090\r
1091 tst r9, #0x0800\r
1092 bne .dsprShi_TileFlip_sh\r
1093\r
1094 @ (r1=pdest, r2=pixels8, r3=pal) r4, r7: scratch, r12: helper pattern\r
1095.dsprShi_TileNorm_sh:\r
1096 TileNormSh\r
1097 b .dsprShi_loop\r
1098\r
1099.dsprShi_TileFlip_sh:\r
1100 TileFlipSh\r
1101 b .dsprShi_loop\r
1102\r
1103.dsprShi_singlec_sh:\r
1104 cmp r2, #0xe0000000\r
1105 bcc .dsprShi_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
1106 tst r2, #0x10000000\r
1107 bne .dsprShi_sh_sh\r
1108 TileSingleHi\r
1109 b .dsprShi_loop\r
1110\r
1111.dsprShi_sh_sh:\r
1112 TileSingleSh\r
1113 b .dsprShi_loop\r
1114\r
1115.dsprShi_shadow_lowpri:\r
1116 tst r9, #0x800\r
1117 bne .dsprShi_TileFlip_sh_lp\r
1118\r
1119.dsprShi_TileNorm_sh_lp:\r
1120 TileNormSh_onlyop_lp\r
1121 b .dsprShi_loop\r
1122\r
1123.dsprShi_TileFlip_sh_lp:\r
1124 TileFlipSh_onlyop_lp\r
1125 b .dsprShi_loop\r
1126\r
1127.pool\r
1128\r
1129@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1130\r
1131.global DrawAllSprites @ unsigned char *sprited, int prio, int sh\r
1132\r
1133DrawAllSprites:\r
1134 ldr r3, =rendstatus\r
1135 orr r1, r2, r1, lsl #1\r
1136 ldr r12,[r3]\r
1137 tst r12,#(PDRAW_DIRTY_SPRITES|PDRAW_SPRITES_MOVED)\r
1138 beq das_no_prep\r
1139 stmfd sp!, {r0,r1,lr}\r
1140 and r0, r12,#PDRAW_DIRTY_SPRITES\r
1141 bic r12,r12,#(PDRAW_DIRTY_SPRITES|PDRAW_SPRITES_MOVED)\r
1142 str r12,[r3]\r
1143 bl PrepareSprites\r
1144 ldmfd sp!, {r0,r1,lr}\r
1145\r
1146das_no_prep:\r
1147 ldr r3, [r0]\r
1148 ands r3, r3, #0x7f\r
1149 bxeq lr\r
1150\r
1151 @ time to do some real work\r
1152 stmfd sp!, {r4-r11,lr}\r
1153 mov r12,#0xff\r
1154 strb r12,[r0,#2] @ set end marker\r
1155 add r10,r0, #3\r
1156 add r10,r10,r3 @ r10=HighLnSpr end\r
1157\r
1158 str r1, [sp, #-4] @ no calls after this point\r
1159\r
1160.if OVERRIDE_HIGHCOL\r
1161 ldr r11,=HighCol\r
1162 mov r12,#0xf\r
1163 ldr r11,[r11]\r
1164.else\r
1165 ldr r11,=HighCol\r
1166 mov r12,#0xf\r
1167.endif\r
1168 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
1169\r
1170@ + 0 : hhhhvvvv ----hhvv yyyyyyyy yyyyyyyy // v, h: horiz. size\r
1171@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
1172\r
1173DrawSprite: @ was: unsigned int *sprite, int sh, int acc_sprites\r
1174 @ draw next sprite\r
1175 ldrb r0, [r10,#-1]!\r
1176 ldr r1, =HighPreSpr\r
1177 ldr r8, [sp, #-4]\r
1178 mov r2, r0, lsr #7\r
1179 cmp r0, #0xff\r
1180 ldmeqfd sp!, {r4-r11,pc} @ end of list\r
1181 cmp r2, r8, lsr #1\r
1182 bne DrawSprite @ wrong priority\r
1183 and r0, r0, #0x7f\r
1184 add r0, r1, r0, lsl #3\r
1185\r
1186@ stmfd sp!, {r4-r9,r11,lr}\r
1187@ orr r8, r2, r1, lsl #4\r
1188\r
1189 ldr r3, [r0] @ sprite[0]\r
1190 ldr r7, =DrawScanline\r
1191 mov r6, r3, lsr #28\r
1192 sub r6, r6, #1 @ r6=width-1 (inc later)\r
1193 mov r5, r3, lsr #24\r
1194 and r5, r5, #7 @ r5=height\r
1195\r
1196 mov r4, r3, lsl #16 @ r4=sy<<16 (tmp)\r
1197\r
1198 ldr r7, [r7]\r
1199 ldr r9, [r0, #4]\r
1200 sub r7, r7, r4, asr #16 @ r7=row=DrawScanline-sy\r
1201\r
1202 mov r2, r9, asr #16 @ r2=sx\r
1203 mov r9, r9, lsl #16\r
1204 mov r9, r9, lsr #16\r
1205 orr r9, r9, r8, lsl #31 @ r9=code|sh[31]\r
1206\r
1207 tst r9, #0x1000\r
1208 movne r4, r5, lsl #3\r
1209 subne r4, r4, #1\r
1210 subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
1211\r
1212 add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
1213 tst r9, #0x0800\r
1214 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1215 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
1216\r
1217 mov r8, r8, lsl #21\r
1218 mov r8, r8, lsr #17\r
1219 and r7, r7, #7\r
1220 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1221\r
1222.dspr_continue:\r
1223 @ cache some stuff to avoid mem access\r
1224 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
1225 and r4, r9, #0x6000\r
1226 orr r9, r9, r4, lsl #16\r
1227 orrs r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
1228\r
1229 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
1230 orrmi r3, r3, #0x40 @ for sh/hi\r
1231\r
1232 add r6, r6, #1 @ inc now\r
1233 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1234 b .dspr_loop_enter\r
1235\r
1236.dspr_loop:\r
1237 subs r6, r6, #1 @ width--\r
1238 beq DrawSprite\r
1239 adds r0, r0, #8 @ sx+=8\r
1240 add r8, r8, r5 @ tile+=delta\r
1241\r
1242.dspr_loop_enter:\r
1243 ble .dspr_loop @ sx <= 0\r
1244 cmp r0, #328\r
1245 bge DrawSprite\r
1246\r
1247 mov r8, r8, lsl #17\r
1248 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
1249\r
1250 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1251 add r1, r11, r0 @ r1=pdest\r
1252 tst r2, r2\r
1253 beq .dspr_loop\r
1254\r
1255 cmp r12, r9, lsr #28\r
1256 beq .dspr_shadow\r
1257\r
1258 cmp r2, r2, ror #4\r
1259 beq .dspr_SingleColor @ tileline singlecolor \r
1260\r
1261 tst r9, #0x0800\r
1262 bne .dspr_TileFlip\r
1263\r
1264 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1265@ scratch: r4, r7\r
1266.dspr_TileNorm:\r
1267 TileNorm r12\r
1268 b .dspr_loop\r
1269\r
1270.dspr_TileFlip:\r
1271 TileFlip r12\r
1272 b .dspr_loop\r
1273\r
1274.dspr_singlec_sh:\r
1275 cmp r2, #0xe0000000\r
1276 bcs .dspr_TileNorm_sh @ op. tileline, markop. XXX: maybe add a spec. handler?\r
1277\r
1278.dspr_SingleColor:\r
1279 and r4, r2, #0xf\r
1280 orr r4, r3, r4\r
1281 orr r4, r4, r4, lsl #8\r
1282 tst r0, #1 @ not aligned?\r
1283 strneb r4, [r1], #1\r
1284 streqh r4, [r1], #2\r
1285 strh r4, [r1], #2\r
1286 strh r4, [r1], #2\r
1287 strh r4, [r1], #2\r
1288 strneb r4, [r1], #1\r
1289 b .dspr_loop\r
1290\r
1291.dspr_shadow:\r
1292 cmp r2, r2, ror #4\r
1293 beq .dspr_singlec_sh\r
1294\r
1295 tst r9, #0x0800\r
1296 bne .dspr_TileFlip_sh\r
1297\r
1298 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1299.dspr_TileNorm_sh:\r
1300 TileNormSh_markop\r
1301 b .dspr_loop\r
1302\r
1303.dspr_TileFlip_sh:\r
1304 TileFlipSh_markop\r
1305 b .dspr_loop\r
1306\r
1307\r
1308@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1309\r
1310.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache\r
1311\r
1312DrawWindow:\r
1313 stmfd sp!, {r4-r11,lr}\r
1314\r
1315 ldr r11, =(Pico+0x22228) @ Pico.video\r
1316 ldr r10, =DrawScanline\r
1317 ldrb r12, [r11, #3] @ pvid->reg[3]\r
1318\r
1319 ldr r10, [r10]\r
1320 ldr r4, [r11, #12]\r
1321 mov r5, r10, lsr #3\r
1322 and r10, r10, #7\r
1323 mov r10, r10, lsl #1 @ r10=ty\r
1324\r
1325 mov r12, r12, lsl #10\r
1326\r
1327 tst r4, #1 @ 40 cell mode?\r
1328 andne r12, r12, #0xf000 @ 0x3c<<10\r
1329 andeq r12, r12, #0xf800\r
1330 addne r12, r12, r5, lsl #7\r
1331 addeq r12, r12, r5, lsl #6 @ nametab\r
1332 add r12, r12, r0, lsl #2 @ +starttile\r
1333\r
1334 ldr r6, =rendstatus\r
1335 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
1336 ldr r6, [r6]\r
1337\r
1338 @ fetch the first code now\r
1339 ldrh r7, [lr, r12]\r
1340\r
1341 ands r6, r6, #PDRAW_WND_DIFF_PRIO\r
1342 orr r6, r6, r2\r
1343\r
1344 eoreq r8, r2, r7, lsr #15 @ do prio bits differ?\r
1345 cmpeq r8, #1\r
1346 ldmeqfd sp!, {r4-r11,pc} @ yes, assume that whole window uses same priority\r
1347\r
1348 orr r6, r6, r3, lsl #8 @ shadow mode\r
1349\r
1350 sub r8, r1, r0\r
1351\r
1352 @ cache some stuff to avoid mem access\r
1353.if OVERRIDE_HIGHCOL\r
1354 ldr r11,=HighCol\r
1355 mov r8, r8, lsl #1 @ cells\r
1356 ldr r11,[r11]\r
1357 mvn r9, #0 @ r9=prevcode=-1\r
1358 add r11,r11,#8\r
1359.else\r
1360 ldr r11,=(HighCol+8)\r
1361 mov r8, r8, lsl #1 @ cells\r
1362 mvn r9, #0 @ r9=prevcode=-1\r
1363.endif\r
1364 add r1, r11, r0, lsl #4 @ r1=pdest\r
1365 mov r0, #0xf\r
1366 b .dwloop_enter\r
1367\r
1368 @ r4,r5 are scratch in this loop\r
1369.dwloop:\r
1370 add r1, r1, #8\r
1371.dwloop_nor1:\r
1372 add r12, r12, #2 @ halfwords\r
1373 ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)\r
1374 subs r8, r8, #1\r
1375 beq .dwloop_end @ done\r
1376\r
1377 eor r5, r6, r7, lsr #15\r
1378 tst r5, #1\r
1379 orrne r6, r6, #2 @ wrong pri\r
1380 bne .dwloop\r
1381\r
1382 cmp r7, r9\r
1383 beq .dw_samecode @ we know stuff about this tile already\r
1384\r
1385.dwloop_enter:\r
1386 mov r9, r7 @ remember code\r
1387\r
1388 movs r2, r9, lsl #20 @ if (code&0x1000)\r
1389 mov r2, r2, lsl #1\r
1390 add r2, r10, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4; addr+=ty\r
1391 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
1392\r
1393 and r3, r9, #0x6000\r
1394 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
1395\r
1396 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1397\r
1398.dw_samecode:\r
1399 tst r6, #0x100\r
1400 bne .dw_shadow\r
1401.dw_shadow_done:\r
1402 tst r2, r2\r
1403 beq .dwloop @ tileline blank\r
1404\r
1405 cmp r2, r2, ror #4\r
1406 beq .dw_SingleColor @ tileline singlecolor \r
1407\r
1408 tst r9, #0x0800\r
1409 bne .dw_TileFlip\r
1410\r
1411 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
1412.dw_TileNorm:\r
1413 TileNorm r0\r
1414 b .dwloop\r
1415\r
1416.dw_TileFlip:\r
1417 TileFlip r0\r
1418 b .dwloop\r
1419\r
1420.dw_SingleColor:\r
1421 and r4, r0, r2 @ #0x0000000f\r
1422 orr r4, r3, r4\r
1423 orr r4, r4, r4, lsl #8\r
1424 orr r4, r4, r4, lsl #16\r
1425 mov r5, r4\r
1426 stmia r1!, {r4,r5}\r
1427 b .dwloop_nor1 @ we incremeted r1 ourselves\r
1428\r
1429.dw_shadow:\r
1430 tst r6, #1 @ hi pri?\r
1431 orreq r3, r3, #0x40\r
1432 beq .dw_shadow_done\r
1433 ldr r4, [r1]\r
1434 mov r5, #0x3f\r
1435 orr r5, r5, r5, lsl #8\r
1436 orr r5, r5, r5, lsl #16\r
1437 and r4, r4, r5\r
1438 str r4, [r1]\r
1439 ldr r4, [r1,#4]\r
1440 and r4, r4, r5\r
1441 str r4, [r1,#4]\r
1442 b .dw_shadow_done\r
1443\r
1444.dwloop_end:\r
1445 ldr r0, =rendstatus\r
1446 ldr r1, [r0]\r
1447 and r6, r6, #PDRAW_WND_DIFF_PRIO\r
1448 orr r1, r1, r6\r
1449 str r1, [r0]\r
1450\r
1451 ldmfd sp!, {r4-r11,r12}\r
1452 bx r12\r
1453\r
1454\r
1455@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1456\r
1457\r
1458@ hilights 2 pixels in RGB444/BGR444 format\r
1459.macro TileDoShHi2Pixels444 reg\r
1460 mov \reg, \reg, ror #12\r
1461 adds \reg, \reg, #0x40000000\r
1462 orrcs \reg, \reg, #0xf0000000\r
1463 mov \reg, \reg, ror #28\r
1464 adds \reg, \reg, #0x40000000\r
1465 orrcs \reg, \reg, #0xf0000000\r
1466 mov \reg, \reg, ror #28\r
1467 adds \reg, \reg, #0x40000000\r
1468 orrcs \reg, \reg, #0xf0000000\r
1469 mov \reg, \reg, ror #24\r
1470 adds \reg, \reg, #0x40000000\r
1471 orrcs \reg, \reg, #0xf0000000\r
1472 mov \reg, \reg, ror #28\r
1473 adds \reg, \reg, #0x40000000\r
1474 orrcs \reg, \reg, #0xf0000000\r
1475 mov \reg, \reg, ror #28\r
1476 adds \reg, \reg, #0x40000000\r
1477 orrcs \reg, \reg, #0xf0000000\r
1478 mov \reg, \reg, ror #12\r
1479.endm\r
1480\r
1481\r
1482@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1483\r
1484\r
1485@ Convert 0000bbb0 ggg0rrr0\r
1486@ to rrrrrggg gggbbbbb\r
1487\r
1488@ r2,r3 - scratch, lr = 0x001c001c, r8 = 0x08610861\r
1489.macro convRGB565 reg\r
1490 and r2, lr, \reg,lsr #7 @ b\r
1491 and r3, lr, \reg,lsr #3 @ g\r
1492 and \reg, lr, \reg,lsl #1 @ r\r
1493 orr r2, r2, r3, lsl #6\r
1494 orr \reg, r2, \reg,lsl #11\r
1495\r
1496 and r2, r8, \reg,lsr #4\r
1497 orr \reg, \reg, r2\r
1498.endm\r
1499\r
1500@ trashes: r2-r8,r12,lr; r8 = 0x08610861; r0,r1 are advanced\r
1501.macro vidConvCpyRGB565_local\r
1502 mov r12, r2, lsr #3 @ repeats\r
1503 mov lr, #0x001c0000\r
1504 orr lr, lr, #0x01c @ lr == pattern 0x001c001c\r
1505\r
15060:\r
1507 ldmia r1!, {r4-r7}\r
1508 subs r12, r12, #1\r
1509 convRGB565 r4\r
1510 str r4, [r0], #4\r
1511 convRGB565 r5\r
1512 str r5, [r0], #4\r
1513 convRGB565 r6\r
1514 str r6, [r0], #4\r
1515 convRGB565 r7\r
1516 str r7, [r0], #4\r
1517\r
1518 bgt 0b\r
1519.endm\r
1520\r
1521\r
1522.global vidConvCpyRGB565\r
1523\r
1524vidConvCpyRGB565: @ void *to, void *from, int pixels\r
1525 stmfd sp!, {r4-r9,lr}\r
1526 mov r8, #0x0061\r
1527 orr r8, r8, #0x0800\r
1528 orr r8, r8, r8, lsl #16\r
1529 vidConvCpyRGB565_local\r
1530 ldmfd sp!, {r4-r9,lr}\r
1531 bx lr\r
1532\r
1533\r
1534.global PicoDoHighPal555 @ int sh\r
1535\r
1536PicoDoHighPal555:\r
1537 stmfd sp!, {r4-r9,lr}\r
1538 mov r1, #0\r
1539 ldr r8, =(Pico+0x22228) @ Pico.video\r
1540\r
1541PicoDoHighPal555_nopush:\r
1542 orr r9, r1, r0, lsl #31 @ 0:called from FinalizeLine555, 31: s/h\r
1543\r
1544 ldr r0, =HighPal\r
1545\r
1546 mov r1, #0\r
1547 strb r1, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
1548\r
1549 sub r1, r8, #0x128 @ r1=Pico.cram\r
1550 mov r2, #0x40\r
1551 mov r8, #0x0061\r
1552 orr r8, r8, #0x0800\r
1553 orr r8, r8, r8, lsl #16\r
1554\r
1555 vidConvCpyRGB565_local\r
1556\r
1557 tst r9, #(1<<31)\r
1558 beq PicoDoHighPal555_end\r
1559\r
1560 ldr r3, =HighPal\r
1561\r
1562 @ shadowed pixels:\r
1563 mov r12, #0x008e\r
1564 add r4, r3, #0x40*2\r
1565 orr r12,r12,#0x7300\r
1566 add r5, r3, #0xc0*2\r
1567 orr r12,r12,r12,lsl #16\r
1568 mov lr, #0x40/4\r
1569.fl_loopcpRGB555_sh:\r
1570 ldmia r3!, {r1,r6}\r
1571 subs lr, lr, #1\r
1572 and r1, r12, r1, lsr #1\r
1573 and r6, r12, r6, lsr #1\r
1574 stmia r4!, {r1,r6}\r
1575 stmia r5!, {r1,r6}\r
1576 bne .fl_loopcpRGB555_sh\r
1577\r
1578 @ hilighted pixels:\r
1579 @ t = ((dpal[i] >> 1) & 0x738e738e) + 0x738e738e;\r
1580 @ t |= (t >> 4) & 0x08610861;\r
1581 @ r8=0x08610861\r
1582 sub r3, r3, #0x40*2\r
1583 mov lr, #0x40/4\r
1584.fl_loopcpRGB555_hi:\r
1585 ldmia r3!, {r1,r6}\r
1586 and r1, r12, r1, lsr #1\r
1587 and r6, r12, r6, lsr #1\r
1588 add r1, r12, r1\r
1589 add r6, r12, r6\r
1590 and r5, r8, r1, lsr #4\r
1591 and r7, r8, r6, lsr #4\r
1592 orr r1, r1, r5\r
1593 orr r6, r6, r7\r
1594 stmia r4!, {r1,r6}\r
1595 subs lr, lr, #1\r
1596 bne .fl_loopcpRGB555_hi\r
1597 mov r0, #1\r
1598\r
1599PicoDoHighPal555_end:\r
1600 tst r9, #1\r
1601 ldmeqfd sp!, {r4-r9,pc}\r
1602\r
1603 ldr r8, =(Pico+0x22228) @ Pico.video\r
1604 b FinalizeLineRGB555_pal_done\r
1605\r
1606\r
1607.global FinalizeLine555 @ int sh\r
1608\r
1609FinalizeLine555:\r
1610 stmfd sp!, {r4-r9,lr}\r
1611 ldr r8, =(Pico+0x22228) @ Pico.video\r
1612\r
1613 ldrb r2, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
1614 mov r1, #1\r
1615 tst r2, r2\r
1616 bne PicoDoHighPal555_nopush\r
1617\r
1618FinalizeLineRGB555_pal_done:\r
1619 ldr r3, =HighPal\r
1620\r
1621 ldr r12,=rendstatus\r
1622 eors r0, r0, #1 @ sh is 0\r
1623 ldr r12,[r12]\r
1624 mov lr, #0xff\r
1625 tstne r12,#PDRAW_ACC_SPRITES\r
1626 movne lr, #0x3f\r
1627\r
1628.if OVERRIDE_HIGHCOL\r
1629 ldr r1, =HighCol\r
1630 ldr r0, =DrawLineDest\r
1631 ldr r1, [r1]\r
1632 ldr r0, [r0]\r
1633 add r1, r1, #8\r
1634.else\r
1635 ldr r0, =DrawLineDest\r
1636 ldr r1, =(HighCol+8)\r
1637 ldr r0, [r0]\r
1638.endif\r
1639\r
1640 ldrb r12, [r8, #12]\r
1641 mov lr, lr, lsl #1\r
1642\r
1643 tst r12, #1\r
1644 movne r2, #320/8 @ len\r
1645 bne .fl_no32colRGB555\r
1646 ldr r4, =PicoOpt\r
1647 mov r2, #256/8\r
1648 ldr r4, [r4]\r
1649 tst r4, #0x4000\r
1650 bne .fl_32scale_RGB555\r
1651 tst r4, #0x0100\r
1652 addeq r0, r0, #32*2\r
1653\r
1654.fl_no32colRGB555:\r
1655\r
1656#ifdef UNALIGNED_DRAWLINEDEST\r
1657 @ this is basically for Gizmondo, which has unaligned odd lines in the framebuffer\r
1658 tst r0, #2\r
1659 bne .fl_RGB555u\r
1660#endif\r
1661\r
1662.fl_loopRGB555:\r
1663 ldr r12, [r1], #4\r
1664 ldr r7, [r1], #4\r
1665\r
1666 and r4, lr, r12, lsl #1\r
1667 ldrh r4, [r3, r4]\r
1668 and r5, lr, r12, lsr #7\r
1669 ldrh r5, [r3, r5]\r
1670 and r6, lr, r12, lsr #15\r
1671 ldrh r6, [r3, r6]\r
1672 orr r4, r4, r5, lsl #16\r
1673\r
1674 and r5, lr, r12, lsr #23\r
1675 ldrh r5, [r3, r5]\r
1676 and r8, lr, r7, lsl #1\r
1677 ldrh r8, [r3, r8]\r
1678 orr r5, r6, r5, lsl #16\r
1679\r
1680 and r6, lr, r7, lsr #7\r
1681 ldrh r6, [r3, r6]\r
1682 and r12,lr, r7, lsr #15\r
1683 ldrh r12,[r3, r12]\r
1684 and r7, lr, r7, lsr #23\r
1685 ldrh r7, [r3, r7]\r
1686 orr r8, r8, r6, lsl #16\r
1687\r
1688 subs r2, r2, #1\r
1689 orr r12,r12, r7, lsl #16\r
1690\r
1691 stmia r0!, {r4,r5,r8,r12}\r
1692 bne .fl_loopRGB555\r
1693\r
1694 ldmfd sp!, {r4-r9,lr}\r
1695 bx lr\r
1696\r
1697\r
1698.fl_32scale_RGB555:\r
1699 stmfd sp!, {r10}\r
1700 mov r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1701 orr r9, r9, #0x00e7\r
1702\r
1703#ifdef UNALIGNED_DRAWLINEDEST\r
1704 tst r0, #2\r
1705 bne .fl_32scale_RGB555u\r
1706#endif\r
1707\r
1708.fl_loop32scale_RGB555:\r
1709 ldr r12, [r1], #4\r
1710 ldr r7, [r1], #4\r
1711\r
1712 and r4, lr, r12,lsl #1\r
1713 ldrh r4, [r3, r4]\r
1714 and r5, lr, r12,lsr #7\r
1715 ldrh r5, [r3, r5]\r
1716 and r4, r4, r9, lsl #2\r
1717 orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0\r
1718 and r5, r5, r9, lsl #2\r
1719 sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1\r
1720 add r4, r4, r6, lsl #16 @ pix_d 0, 1\r
1721 and r6, lr, r12,lsr #15\r
1722 ldrh r6, [r3, r6]\r
1723 and r12,lr, r12,lsr #23\r
1724 ldrh r12,[r3, r12]\r
1725 and r6, r6, r9, lsl #2\r
1726 add r5, r5, r6\r
1727 mov r5, r5, lsr #1\r
1728 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1729 orr r5, r5, r6, lsl #16\r
1730\r
1731 and r6, lr, r7, lsl #1\r
1732 ldrh r6, [r3, r6]\r
1733 and r12,r12,r9, lsl #2\r
1734 add r5, r5, r12,lsl #14 @ pix_d 2, 3\r
1735 and r6, r6, r9, lsl #2\r
1736 orr r6, r12,r6, lsl #16 @ pix_d 4, 5\r
1737\r
1738 and r12,lr, r7, lsr #7\r
1739 ldrh r12,[r3, r12]\r
1740 and r10,lr, r7, lsr #15\r
1741 ldrh r10,[r3, r10]\r
1742 and r12,r12,r9, lsl #2\r
1743 sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1\r
1744 add r8, r8, r6, lsr #18\r
1745 and r7, lr, r7, lsr #23\r
1746 ldrh r7, [r3, r7]\r
1747 and r10,r10,r9, lsl #2\r
1748 orr r8, r8, r10,lsl #15\r
1749 add r8, r8, r12,lsl #15 @ pix_d 6, 7\r
1750 sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2\r
1751 and r7, r7, r9, lsl #2\r
1752 add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3\r
1753 orr r10,r10,r7, lsl #16 @ pix_d 8, 9\r
1754\r
1755 subs r2, r2, #1\r
1756\r
1757 stmia r0!, {r4,r5,r6,r8,r10}\r
1758 bne .fl_loop32scale_RGB555\r
1759\r
1760 ldmfd sp!, {r10}\r
1761 ldmfd sp!, {r4-r9,lr}\r
1762 bx lr\r
1763\r
1764#ifdef UNALIGNED_DRAWLINEDEST\r
1765 @ unaligned versions of loops\r
1766 @ warning: starts drawing 2bytes before dst\r
1767\r
1768.fl_RGB555u:\r
1769 sub r0, r0, #2 @ initial adjustment\r
1770 mov r8, #0\r
1771\r
1772.fl_loopRGB555u:\r
1773 ldr r12, [r1], #4\r
1774 ldr r7, [r1], #4\r
1775\r
1776 and r6, lr, r12,lsl #1\r
1777 ldrh r6, [r3, r6]\r
1778 and r5, lr, r12,lsr #7\r
1779 ldrh r5, [r3, r5]\r
1780 orr r4, r8, r6, lsl #16\r
1781\r
1782 and r6, lr, r12,lsr #15\r
1783 ldrh r6, [r3, r6]\r
1784 and r8, lr, r12,lsr #23\r
1785 ldrh r8, [r3, r8]\r
1786 orr r5, r5, r6, lsl #16\r
1787\r
1788 and r6, lr, r7, lsl #1\r
1789 ldrh r6, [r3, r6]\r
1790 and r12,lr, r7, lsr #7\r
1791 ldrh r12,[r3, r12]\r
1792 orr r6, r8, r6, lsl #16\r
1793\r
1794 and r8, lr, r7, lsr #15\r
1795 ldrh r8, [r3, r8]\r
1796 and r7, lr, r7, lsr #23\r
1797\r
1798 subs r2, r2, #1\r
1799 orr r12,r12,r8, lsl #16\r
1800 ldrh r8, [r3, r7]\r
1801\r
1802 stmia r0!, {r4,r5,r6,r12}\r
1803 bne .fl_loopRGB555u\r
1804\r
1805 strh r8, [r0], #2\r
1806\r
1807 ldmfd sp!, {r4-r9,lr}\r
1808 bx lr\r
1809\r
1810\r
1811.fl_32scale_RGB555u:\r
1812 sub r0, r0, #2 @ initial adjustment\r
1813 mov r4, #0\r
1814\r
1815 @ r9 f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1816.fl_loop32scale_RGB555u:\r
1817 ldr r12, [r1], #4\r
1818 ldr r7, [r1], #4\r
1819\r
1820 and r6, lr, r12,lsl #1\r
1821 ldrh r6, [r3, r6]\r
1822 and r5, lr, r12,lsr #7\r
1823 ldrh r5, [r3, r5]\r
1824 and r6, r6, r9, lsl #2\r
1825 orr r4, r4, r6, lsl #16 @ r4 = pix_d -1, 0\r
1826\r
1827 and r5, r5, r9, lsl #2\r
1828 sub r8, r5, r5, lsr #2 @ r8 = 3/4 pix_s 1\r
1829 add r6, r8, r6, lsr #2 @ r6 = (1/4 pix_s 0) + (3/4 pix_s 1)\r
1830 orr r5, r6, r5, lsl #15\r
1831\r
1832 and r6, lr, r12,lsr #15\r
1833 ldrh r6, [r3, r6]\r
1834 and r12,lr, r12,lsr #23\r
1835 ldrh r12,[r3, r12]\r
1836 and r6, r6, r9, lsl #2\r
1837 add r5, r5, r6, lsl #15 @ r5 = pix_d 1, 2\r
1838\r
1839 and r8, lr, r7, lsl #1\r
1840 ldrh r8, [r3, r8]\r
1841 and r10,lr, r7, lsr #7\r
1842 ldrh r10,[r3, r10]\r
1843 and r12,r12,r9, lsl #2\r
1844 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1845 add r6, r6, r12,lsr #2\r
1846 orr r6, r6, r12,lsl #16 @ r6 = pix_d 3, 4\r
1847\r
1848 and r8, r8, r9, lsl #2\r
1849 and r10,r10,r9, lsl #2\r
1850 sub r12,r10,r10,lsr #2 @ r12 = 3/4 pix_s 5\r
1851 orr r8, r8, r8, lsl #14\r
1852 add r8, r8, r12,lsl #16 @ r8 = pix_d 5, 6\r
1853 and r12,lr, r7, lsr #15\r
1854 ldrh r12,[r3, r12]\r
1855 and r7, lr, r7, lsr #23\r
1856 ldrh r7, [r3, r7]\r
1857 and r12,r12,r9, lsl #2\r
1858 add r10,r10,r12\r
1859 mov r10,r10, lsr #1\r
1860 sub r12,r12,r12,lsr #2 @ r12 = 3/4 pix_s 6\r
1861 orr r10,r10,r12,lsl #16\r
1862 and r7, r7, r9, lsl #2\r
1863 add r10,r10,r7, lsl #14 @ r10 = pix_d 7, 8\r
1864\r
1865 subs r2, r2, #1\r
1866\r
1867 stmia r0!, {r4,r5,r6,r8,r10}\r
1868 mov r4, r7\r
1869 bne .fl_loop32scale_RGB555u\r
1870\r
1871 strh r4, [r0], #2\r
1872\r
1873 ldmfd sp!, {r10}\r
1874 ldmfd sp!, {r4-r9,lr}\r
1875 bx lr\r
1876\r
1877#endif /* UNALIGNED_DRAWLINEDEST */\r
1878\r
1879\r
1880@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1881\r
1882@ utility\r
1883.global blockcpy @ void *dst, void *src, size_t n\r
1884\r
1885blockcpy:\r
1886 stmfd sp!, {r4,r5}\r
1887 mov r2, r2, lsr #4\r
1888blockcpy_loop:\r
1889 ldmia r1!, {r3-r5,r12}\r
1890 subs r2, r2, #1\r
1891 stmia r0!, {r3-r5,r12}\r
1892 bne blockcpy_loop\r
1893 ldmfd sp!, {r4,r5}\r
1894 bx lr\r
1895\r
1896\r
1897.global blockcpy_or @ void *dst, void *src, size_t n, int pat\r
1898\r
1899blockcpy_or:\r
1900 stmfd sp!, {r4-r6}\r
1901 orr r3, r3, r3, lsl #8\r
1902 orr r3, r3, r3, lsl #16\r
1903 mov r2, r2, lsr #4\r
1904blockcpy_loop_or:\r
1905 ldmia r1!, {r4-r6,r12}\r
1906 subs r2, r2, #1\r
1907 orr r4, r4, r3\r
1908 orr r5, r5, r3\r
1909 orr r6, r6, r3\r
1910 orr r12,r12,r3\r
1911 stmia r0!, {r4-r6,r12}\r
1912 bne blockcpy_loop_or\r
1913 ldmfd sp!, {r4-r6}\r
1914 bx lr\r
1915\r
1916@ vim:filetype=armasm\r
fast MegaDrive/MegaCD/32X emulator