notaz.gp2x.de / picodrive.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
improve menu autorepeat
[picodrive.git] / pico / draw_arm.s
CommitLineData
3d48f143 1@ vim:filetype=armasm\r
2\r
07abbab1 3@ ARM assembly versions of some funtions from draw.c\r
cc68a136 4@ this is highly specialized, be careful if changing related C code!\r
5\r
07abbab1 6@ (c) Copyright 2007-2008, Grazvydas "notaz" Ignotas\r
cc68a136 7@ All Rights Reserved\r
8\r
ea8c405f 9.include "port_config.s"\r
cc68a136 10\r
11.extern Pico\r
12.extern PicoOpt\r
13.extern HighCol\r
b6d7ac70 14.extern DrawScanline\r
cc68a136 15.extern HighSprZ\r
16.extern rendstatus\r
283fec1b 17.extern HighPreSpr\r
cc68a136 18.extern DrawLineDest\r
cc68a136 19.extern DrawStripInterlace\r
07abbab1 20.extern HighCacheS_ptr\r
cc68a136 21\r
283fec1b 22.equ PDRAW_SPRITES_MOVED, (1<<0)\r
23.equ PDRAW_WND_DIFF_PRIO, (1<<1)\r
24.equ PDRAW_ACC_SPRITES, (1<<2)\r
25.equ PDRAW_DIRTY_SPRITES, (1<<4)\r
26.equ PDRAW_PLANE_HI_PRIO, (1<<6)\r
27.equ PDRAW_SHHI_DONE, (1<<7)\r
cc68a136 28\r
29@ helper\r
30.macro TilePixel pat lsrr offs\r
31.if !\lsrr\r
32 ands r4, \pat, r2\r
33.else\r
34 ands r4, \pat, r2, lsr #\lsrr\r
35.endif\r
36 orrne r4, r3, r4\r
37 strneb r4, [r1,#\offs]\r
38.endm\r
39\r
40@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
41.macro TileNorm pat\r
42 TilePixel \pat, 12, 0 @ #0x0000f000\r
43 TilePixel \pat, 8, 1 @ #0x00000f00\r
44 TilePixel \pat, 4, 2 @ #0x000000f0\r
45 TilePixel \pat, 0, 3 @ #0x0000000f\r
46 TilePixel \pat, 28, 4 @ #0xf0000000\r
47 TilePixel \pat, 24, 5 @ #0x0f000000\r
48 TilePixel \pat, 20, 6 @ #0x00f00000\r
49 TilePixel \pat, 16, 7 @ #0x000f0000\r
50.endm\r
51\r
52@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
53.macro TileFlip pat\r
54 TilePixel \pat, 16, 0 @ #0x000f0000\r
55 TilePixel \pat, 20, 1 @ #0x00f00000\r
56 TilePixel \pat, 24, 2 @ #0x0f000000\r
57 TilePixel \pat, 28, 3 @ #0xf0000000\r
58 TilePixel \pat, 0, 4 @ #0x0000000f\r
59 TilePixel \pat, 4, 5 @ #0x000000f0\r
60 TilePixel \pat, 8, 6 @ #0x00000f00\r
61 TilePixel \pat, 12, 7 @ #0x0000f000\r
62.endm\r
63\r
64@ shadow/hilight mode\r
65\r
66@ this one is for hi priority layer\r
67.macro TilePixelShHP lsrr offs\r
68.if !\lsrr\r
69 ands r4, r12, r2\r
70.else\r
71 ands r4, r12, r2, lsr #\lsrr\r
72.endif\r
73 ldreqb r4, [r1,#\offs]\r
74 orrne r4, r3, r4\r
cc68a136 75 andeq r4, r4, #0x3f\r
07abbab1 76 strb r4, [r1,#\offs]\r
cc68a136 77.endm\r
78\r
7a7c6476 79@ TileNormShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
cc68a136 80.macro TileNormShHP\r
81 TilePixelShHP 12, 0 @ #0x0000f000\r
82 TilePixelShHP 8, 1 @ #0x00000f00\r
83 TilePixelShHP 4, 2 @ #0x000000f0\r
84 TilePixelShHP 0, 3 @ #0x0000000f\r
85 TilePixelShHP 28, 4 @ #0xf0000000\r
86 TilePixelShHP 24, 5 @ #0x0f000000\r
87 TilePixelShHP 20, 6 @ #0x00f00000\r
88 TilePixelShHP 16, 7 @ #0x000f0000\r
89.endm\r
90\r
7a7c6476 91@ TileFlipShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
cc68a136 92.macro TileFlipShHP\r
93 TilePixelShHP 16, 0 @ #0x000f0000\r
94 TilePixelShHP 20, 1 @ #0x00f00000\r
95 TilePixelShHP 24, 2 @ #0x0f000000\r
96 TilePixelShHP 28, 3 @ #0xf0000000\r
97 TilePixelShHP 0, 4 @ #0x0000000f\r
98 TilePixelShHP 4, 5 @ #0x000000f0\r
99 TilePixelShHP 8, 6 @ #0x00000f00\r
100 TilePixelShHP 12, 7 @ #0x0000f000\r
101.endm\r
102\r
103\r
104@ TileSingleSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx; r12: helper pattern 0xf\r
105.macro TileSingleSh\r
106 tst r0, #1 @ not aligned?\r
107 mov r7, #0x00c000\r
108 orr r7, r7, #0xc0\r
109 ldrneb r4, [r1]\r
110 ldreqh r4, [r1]\r
111 orr r4, r4, r7\r
112 strneb r4, [r1], #1\r
113 streqh r4, [r1], #2\r
114 ldrh r4, [r1]\r
115 orr r4, r4, r7\r
116 strh r4, [r1], #2\r
117 ldrh r4, [r1]\r
118 orr r4, r4, r7\r
119 strh r4, [r1], #2\r
120 ldrh r4, [r1]\r
121 orr r4, r4, r7\r
122 strh r4, [r1], #2\r
123 ldrneb r4, [r1]\r
124 orr r4, r4, r7\r
125 strneb r4, [r1], #1\r
126.endm\r
127\r
128@ TileSingleHi (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
129.macro TileSingleHi\r
130 tst r1, #1 @ not aligned?\r
131 mov r7, #0x008000\r
132 orr r7, r7, #0x80\r
133 ldrneb r4, [r1], #1\r
134 ldreqh r4, [r1], #2 @ 1ci\r
135 ldrh r12, [r1], #2\r
136 bic r4, r4, r7, lsr #1\r
137 orr r4, r4, r7\r
138 strneb r4, [r1, #-3]\r
139 streqh r4, [r1, #-4]\r
140 ldrh r4, [r1], #2\r
141 bic r12, r12, r7, lsr #1\r
142 orr r12, r12, r7\r
143 strh r12, [r1, #-4]\r
144 ldrh r12, [r1], #2\r
145 bic r4, r4, r7, lsr #1\r
146 orr r4, r4, r7\r
147 strh r4, [r1, #-4]\r
148 ldrneb r4, [r1]\r
149 bic r12, r12, r7, lsr #1\r
150 orr r12, r12, r7\r
151 strh r12, [r1, #-2]\r
152 bicne r4, r4, r7, lsr #1\r
153 orrne r4, r4, r7\r
154 strneb r4, [r1], #1\r
155 mov r12, #0xf\r
156.endm\r
157\r
158.macro TileDoShGenPixel shift ofs\r
159.if \shift\r
160 ands r4, r12, r2, lsr #\shift\r
161.else\r
162 ands r4, r12, r2\r
163.endif\r
07abbab1 164 beq 0f\r
cc68a136 165 cmp r4, #0xe\r
07abbab1 166 ldrgeb r4, [r1,#\ofs]\r
167 orrlt r4, r3, r4 @ normal\r
168\r
169 biceq r4, r4, #0xc0 @ hilight\r
170 orreq r4, r4, #0x80\r
171 orrgt r4, r4, #0xc0 @ shadow\r
172\r
cc68a136 173 strb r4, [r1,#\ofs]\r
07abbab1 1740:\r
cc68a136 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
07abbab1 201.macro TileDoShGenPixel_noop shift ofs\r
202.if \shift\r
203 and r4, r12, r2, lsr #\shift\r
204.else\r
205 and r4, r12, r2\r
206.endif\r
207 sub r7, r4, #1\r
208 cmp r7, #0xd\r
209 orrcc r4, r3, r4 @ 0-0xc (was 1-0xd)\r
210 strccb r4, [r1,#\ofs]\r
211.endm\r
212\r
213.macro TileFlipSh_noop\r
214 TileDoShGenPixel_noop 16, 0 @ #0x000f0000\r
215 TileDoShGenPixel_noop 20, 1 @ #0x00f00000\r
216 TileDoShGenPixel_noop 24, 2 @ #0x0f000000\r
217 TileDoShGenPixel_noop 28, 3 @ #0xf0000000\r
218 TileDoShGenPixel_noop 0, 4 @ #0x0000000f\r
219 TileDoShGenPixel_noop 4, 5 @ #0x000000f0\r
220 TileDoShGenPixel_noop 8, 6 @ #0x00000f00\r
221 TileDoShGenPixel_noop 12, 7 @ #0x0000f000\r
222.endm\r
223\r
224.macro TileNormSh_noop\r
225 TileDoShGenPixel_noop 12, 0 @ #0x0000f000\r
226 TileDoShGenPixel_noop 8, 1 @ #0x00000f00\r
227 TileDoShGenPixel_noop 4, 2 @ #0x000000f0\r
228 TileDoShGenPixel_noop 0, 3 @ #0x0000000f\r
229 TileDoShGenPixel_noop 28, 4 @ #0xf0000000\r
230 TileDoShGenPixel_noop 24, 5 @ #0x0f000000\r
231 TileDoShGenPixel_noop 20, 6 @ #0x00f00000\r
232 TileDoShGenPixel_noop 16, 7 @ #0x000f0000\r
233.endm\r
234\r
235.macro TileDoShGenPixel_onlyop_lp shift ofs\r
236.if \shift\r
237 ands r7, r12, r2, lsr #\shift\r
238.else\r
239 ands r7, r12, r2\r
240.endif\r
241 ldrneb r4, [r1,#\ofs]\r
242 tstne r4, #0x40\r
243 beq 0f\r
244\r
245 cmp r7, #0xe\r
246 biceq r4, r4, #0xc0 @ hilight\r
247 orreq r4, r4, #0x80\r
248 orrgt r4, r4, #0xc0 @ shadow\r
249 strgeb r4, [r1,#\ofs]\r
2500:\r
251.endm\r
252\r
253.macro TileFlipSh_onlyop_lp\r
254 TileDoShGenPixel_onlyop_lp 16, 0 @ #0x000f0000\r
255 TileDoShGenPixel_onlyop_lp 20, 1 @ #0x00f00000\r
256 TileDoShGenPixel_onlyop_lp 24, 2 @ #0x0f000000\r
257 TileDoShGenPixel_onlyop_lp 28, 3 @ #0xf0000000\r
258 TileDoShGenPixel_onlyop_lp 0, 4 @ #0x0000000f\r
259 TileDoShGenPixel_onlyop_lp 4, 5 @ #0x000000f0\r
260 TileDoShGenPixel_onlyop_lp 8, 6 @ #0x00000f00\r
261 TileDoShGenPixel_onlyop_lp 12, 7 @ #0x0000f000\r
262.endm\r
263\r
264.macro TileNormSh_onlyop_lp\r
265 TileDoShGenPixel_onlyop_lp 12, 0 @ #0x0000f000\r
266 TileDoShGenPixel_onlyop_lp 8, 1 @ #0x00000f00\r
267 TileDoShGenPixel_onlyop_lp 4, 2 @ #0x000000f0\r
268 TileDoShGenPixel_onlyop_lp 0, 3 @ #0x0000000f\r
269 TileDoShGenPixel_onlyop_lp 28, 4 @ #0xf0000000\r
270 TileDoShGenPixel_onlyop_lp 24, 5 @ #0x0f000000\r
271 TileDoShGenPixel_onlyop_lp 20, 6 @ #0x00f00000\r
272 TileDoShGenPixel_onlyop_lp 16, 7 @ #0x000f0000\r
273.endm\r
274\r
cc68a136 275\r
276@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
277\r
278@ struct TileStrip\r
279@ {\r
280@ int nametab; // 0x00\r
281@ int line; // 0x04\r
282@ int hscroll; // 0x08\r
283@ int xmask; // 0x0C\r
284@ int *hc; // 0x10 (pointer to cache buffer)\r
285@ int cells; // 0x14\r
286@ };\r
287\r
83c093a4 288@ void DrawLayer(int plane_sh, int *hcache, int cellskip, int maxcells);\r
cc68a136 289\r
83c093a4 290.global DrawLayer\r
cc68a136 291\r
292DrawLayer:\r
293 stmfd sp!, {r4-r11,lr}\r
294\r
295 ldr r11, =(Pico+0x22228) @ Pico.video\r
296 mov r8, #1\r
297\r
83c093a4 298 ldrb r7, [r11, #16] @ ??vv??hh\r
cc68a136 299\r
300 mov r6, r1 @ hcache\r
83c093a4 301 orr r9, r3, r0, lsl #30\r
302 orr r9, r9, r2, lsl #8 @ r9=sh[31]|cellskip[15:8]|maxcells[7:0] (tmp)\r
cc68a136 303\r
304 mov r1, r7, lsl #4\r
305 orr r1, r1, #0x00ff\r
306\r
307 and r10, r7, #3\r
308 cmp r10, #1\r
309 biclt r1, r1, #0xfc00\r
310 biceq r1, r1, #0xfe00\r
311 bicgt r1, r1, #0xff00 @ r1=ymask=(height<<8)|0xff; ...; // Y Mask in pixels\r
312\r
313 add r10, r10, #5\r
314 cmp r10, #7\r
315 subge r10, r10, #1 @ r10=shift[width] (5,6,6,7)\r
316\r
317 @ calculate xmask:\r
318 mov r5, r8, lsl r10\r
319 sub r5, r5, #1 @ r5=xmask\r
320\r
321 @ Find name table:\r
83c093a4 322 ands r0, r0, #1\r
cc68a136 323 ldreqb r12, [r11, #2]\r
324 ldrneb r12, [r11, #4]\r
325\r
b6d7ac70 326 ldr r2, =DrawScanline @ trying to make good use of pipeline here\r
cc68a136 327 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
328\r
329 moveq r12, r12, lsl #10\r
330 movne r12, r12, lsl #13\r
331 and r12, r12, #(7<<13) @ r12=(ts->nametab<<1) (halfword compliant)\r
332\r
333 ldrh r8, [r11, #12]\r
334 ldrb r7, [r11, #11]\r
335 ldr r2, [r2]\r
336\r
337 mov r4, r8, lsr #8 @ pvid->reg[13]\r
338 mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r
339 tst r7, #2\r
b6d7ac70 340 addne r4, r4, r2, lsl #2 @ htab+=DrawScanline<<1; // Offset by line\r
cc68a136 341 tst r7, #1\r
342 biceq r4, r4, #0x1f @ htab&=~0xf; // Offset by tile\r
343 add r4, r4, r0, lsl #1 @ htab+=plane\r
344 bic r4, r4, #0x00ff0000 @ just in case\r
345 ldrh r3, [lr, r4] @ r3=hscroll\r
346\r
347 tst r7, #4\r
348 bne .DrawStrip_vsscroll\r
349\r
350 @ Get vertical scroll value:\r
351 add r7, lr, #0x012000\r
352 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
353 ldr r7, [r7]\r
354\r
355 tst r8, #2\r
356 tstne r8, #4\r
357 bne .DrawStrip_interlace\r
358\r
359 tst r0, r0\r
360 movne r7, r7, lsr #16\r
361\r
362 @ Find the line in the name table\r
363 add r2, r2, r7\r
364 and r2, r2, r1\r
365 mov r4, r2, lsr #3\r
366 add r10, r10, #1 @ shift[width]++\r
367 add r12, r12, r4, lsl r10 @ nametab+=(ts.line>>3)<<shift[width];\r
368\r
369 @ ldmia r0, {r1,r2,r3,r5,r6,r9} @ r2=line, r3=ts->hscroll, r5=ts->xmask, r6=ts->hc, r9=ts->cells\r
cc68a136 370\r
371 and r10,r2, #7\r
372 mov r10,r10, lsl #1 @ r10=ty=(ts->line&7)<<1;\r
373 orr r10,r10, r9, lsl #24\r
374\r
375 rsb r8, r3, #0\r
376 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
377\r
378 sub r1, r3, #1\r
379 and r1, r1, #7\r
380 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
381\r
382 tst r9, #1<<31\r
383 mov r3, #0\r
740da8c6 384 orrne r10,r10, #1<<23 @ r10=(cells<<24|sh<<23|hi_not_empty<<22|had_output<<21|ty)\r
cc68a136 385 movne r3, #0x40 @ default to shadowed pal on sh mode\r
386\r
cc68a136 387 cmp r7, #8\r
388 addne r10,r10, #0x01000000 @ we will loop cells+1 times if there is scroll\r
389\r
83c093a4 390 and r9, r9, #0xff00\r
391 add r8, r8, r9, lsr #8 @ tilex+=cellskip\r
392 add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;\r
393 sub r10,r10,r9, lsl #16 @ cells-=cellskip\r
83c093a4 394\r
cc68a136 395 @ cache some stuff to avoid mem access\r
ea8c405f 396.if OVERRIDE_HIGHCOL\r
397 ldr r11,=HighCol\r
398 mov r0, #0xf\r
399 ldr r11,[r11]\r
400.else\r
cc68a136 401 ldr r11,=HighCol\r
402 mov r0, #0xf\r
ea8c405f 403.endif\r
404\r
405 mvn r9, #0 @ r9=prevcode=-1\r
cc68a136 406 add r1, r11, r7 @ r1=pdest\r
407\r
408\r
409 @ r4 & r7 are scratch in this loop\r
410.dsloop_subr1:\r
411 sub r1, r1, #8\r
412.dsloop: @ 40-41 times\r
413 subs r10,r10, #0x01000000\r
414 bmi .dsloop_exit\r
415\r
416.dsloop_enter:\r
417 and r7, r5, r8\r
418 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
419 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
420\r
421 add r1, r1, #8\r
422 add r8, r8, #1\r
423\r
424 tst r7, #0x8000\r
425 bne .DrawStrip_hiprio\r
426\r
427 cmp r7, r9\r
428 beq .DrawStrip_samecode @ we know stuff about this tile already\r
429\r
430 mov r9, r7 @ remember code\r
740da8c6 431 orr r10, r10, #1<<21 @ seen non hi-prio tile\r
cc68a136 432\r
433 movs r2, r9, lsl #20 @ if (code&0x1000)\r
434 mov r2, r2, lsl #1\r
435 add r2, r2, r10, lsl #17\r
436 mov r2, r2, lsr #17\r
437 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
438\r
439 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
440\r
441 bic r7, r3, #0x3f\r
442 and r3, r9, #0x6000\r
443 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
444\r
445.DrawStrip_samecode:\r
446 tst r2, r2\r
447 beq .dsloop @ tileline blank\r
448\r
449 cmp r2, r2, ror #4\r
450 beq .DrawStrip_SingleColor @ tileline singlecolor \r
451\r
452 tst r9, #0x0800\r
7a7c6476 453 bne .DrawStrip_TileFlip\r
cc68a136 454\r
455 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
cc68a136 456.DrawStrip_TileNorm:\r
457 TileNorm r0\r
458 b .dsloop\r
459\r
7a7c6476 460.DrawStrip_TileFlip:\r
461 TileFlip r0\r
462 b .dsloop\r
463\r
cc68a136 464.DrawStrip_SingleColor:\r
465 and r4, r2, #0xf\r
466 orr r4, r3, r4\r
467 orr r4, r4, r4, lsl #8\r
468 tst r1, #1 @ not aligned?\r
469 strneb r4, [r1], #1\r
470 streqh r4, [r1], #2\r
471 strh r4, [r1], #2\r
472 strh r4, [r1], #2\r
473 strh r4, [r1], #2\r
474 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
475 b .dsloop_subr1\r
476\r
cc68a136 477.DrawStrip_hiprio_maybempt:\r
478 cmp r7, r9\r
479 beq .dsloop @ must've been empty, otherwise we wouldn't get here\r
480 movs r2, r7, lsl #20 @ if (code&0x1000)\r
481 mov r2, r2, lsl #1\r
482 add r2, r2, r10, lsl #17\r
483 mov r2, r2, lsr #17\r
484 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
485 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
486 mov r9, r7 @ remember code\r
487 tst r2, r2\r
740da8c6 488 beq .dsloop\r
489 orr r10, r10, #1<<22\r
490\r
491.DrawStrip_hiprio:\r
492 tst r10, #0x00c00000\r
493 beq .DrawStrip_hiprio_maybempt\r
494 sub r0, r1, r11\r
495 orr r7, r7, r0, lsl #16\r
496 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
497 tst r7, #0x1000\r
498 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
499 str r7, [r6], #4 @ cache hi priority tile\r
500 mov r0, #0xf\r
cc68a136 501 b .dsloop\r
502\r
503.dsloop_exit:\r
740da8c6 504 tst r10, #1<<21 @ seen non hi-prio tile\r
505 ldreq r1, =rendstatus\r
cc68a136 506 mov r0, #0\r
740da8c6 507 ldreq r2, [r1]\r
cc68a136 508 str r0, [r6] @ terminate the cache list\r
283fec1b 509 orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
740da8c6 510 streq r2, [r1]\r
cc68a136 511\r
512 ldmfd sp!, {r4-r11,lr}\r
513 bx lr\r
514\r
6d7acf9e 515@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
cc68a136 516\r
517.DrawStrip_vsscroll:\r
6d7acf9e 518 rsb r8, r3, #0\r
519 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
7b802576 520 bic r8, r8, #0x3fc00000\r
521 orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|tilex[21:0])\r
cc68a136 522\r
b6d7ac70 523 ldr r4, =DrawScanline\r
6d7acf9e 524 orr r5, r1, r10, lsl #24\r
525 ldr r4, [r4]\r
526 sub r1, r3, #1\r
527 orr r5, r5, r4, lsl #16 @ r5=(shift_width[31:24]|scanline[23:16]|ymask[15:0])\r
528 and r1, r1, #7\r
529 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
cc68a136 530\r
6d7acf9e 531 mov r10,r9, lsl #16\r
83c093a4 532 tst r0, #1\r
6d7acf9e 533 orrne r10,r10, #0x8000\r
534 tst r9, #1<<31\r
535 mov r3, #0\r
536 orr r10,r10, #0xff000000 @ will be adjusted on entering loop\r
83c093a4 537 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
6d7acf9e 538 movne r3, #0x40 @ default to shadowed pal on sh mode\r
cc68a136 539\r
83c093a4 540 cmp r7, #8\r
541 subne r10,r10, #0x01000000 @ have hscroll, start with negative cell\r
542\r
543 and r9, r9, #0xff00\r
544 add r8, r8, r9, lsr #8 @ tilex+=cellskip\r
545 add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;\r
546 add r10,r10,r9, lsl #16 @ cell+=cellskip\r
cc68a136 547\r
6d7acf9e 548 @ cache some stuff to avoid mem access\r
ea8c405f 549.if OVERRIDE_HIGHCOL\r
6d7acf9e 550 ldr r11,=HighCol\r
551 mov r0, #0xf\r
ea8c405f 552 ldr r11,[r11]\r
553.else\r
554 ldr r11,=HighCol\r
555 mov r0, #0xf\r
556.endif\r
6d7acf9e 557\r
ea8c405f 558 mvn r9, #0 @ r9=prevcode=-1\r
559 add r1, r11, r7 @ r1=pdest\r
6d7acf9e 560\r
561 @ r4 & r7 are scratch in this loop\r
562.dsloop_vs_subr1:\r
563 sub r1, r1, #8\r
564.dsloop_vs: @ 40-41 times\r
565 add r10,r10, #0x01000000\r
566 and r4, r10, #0x003f0000\r
567 cmp r4, r10, asr #8\r
740da8c6 568 ble .dsloop_vs_exit\r
6d7acf9e 569\r
570 @ calc offset and read tileline code to r7, also calc ty\r
571 add r7, lr, #0x012000\r
572 add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r
573 add r7, r7, r10,asr #23 @ vsram + ((cell&~1)<<1)\r
574 bic r7, r7, #3\r
575 tst r10,#0x8000 @ plane1?\r
576 addne r7, r7, #2\r
577 ldrh r7, [r7] @ r7=vscroll\r
578\r
579 bic r10,r10,#0xff @ clear old ty\r
7b802576 580 and r4, r5, #0xff0000 @ scanline\r
581 add r4, r4, r7, lsl #16 @ ... += vscroll\r
582 and r4, r4, r5, lsl #16 @ ... &= ymask\r
6d7acf9e 583 and r7, r4, #0x70000\r
584 orr r10,r10,r7, lsr #15 @ new ty\r
585\r
586 mov r4, r4, lsr #19\r
587 mov r7, r5, lsr #24\r
588 mov r4, r4, lsl r7 @ nametabadd\r
589\r
590 and r7, r8, r8, lsr #25\r
591 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
592 add r7, r7, r4, lsl #1\r
593 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
594\r
595 add r1, r1, #8\r
596 add r8, r8, #1\r
597\r
598 tst r7, #0x8000\r
599 bne .DrawStrip_vs_hiprio\r
600\r
601 cmp r7, r9\r
602 beq .DrawStrip_vs_samecode @ we know stuff about this tile already\r
603\r
604 mov r9, r7 @ remember code\r
7b802576 605 orr r8, r8, #(1<<24)@ seen non hi-prio tile\r
6d7acf9e 606\r
607 movs r2, r9, lsl #20 @ if (code&0x1000)\r
608 mov r2, r2, lsl #1\r
609 add r2, r2, r10, lsl #17\r
610 mov r2, r2, lsr #17\r
611 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
612\r
613 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
614\r
615 bic r7, r3, #0x3f\r
616 and r3, r9, #0x6000\r
617 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
618\r
619.DrawStrip_vs_samecode:\r
620 tst r2, r2\r
621 beq .dsloop_vs @ tileline blank\r
622\r
623 cmp r2, r2, ror #4\r
624 beq .DrawStrip_vs_SingleColor @ tileline singlecolor \r
625\r
626 tst r9, #0x0800\r
7a7c6476 627 bne .DrawStrip_vs_TileFlip\r
6d7acf9e 628\r
629 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
6d7acf9e 630.DrawStrip_vs_TileNorm:\r
631 TileNorm r0\r
632 b .dsloop_vs\r
633\r
7a7c6476 634.DrawStrip_vs_TileFlip:\r
635 TileFlip r0\r
636 b .dsloop_vs\r
637\r
6d7acf9e 638.DrawStrip_vs_SingleColor:\r
639 and r4, r2, #0xf\r
640 orr r4, r3, r4\r
641 orr r4, r4, r4, lsl #8\r
642 tst r1, #1 @ not aligned?\r
643 strneb r4, [r1], #1\r
644 streqh r4, [r1], #2\r
645 strh r4, [r1], #2\r
646 strh r4, [r1], #2\r
647 strh r4, [r1], #2\r
648 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
649 b .dsloop_vs_subr1\r
650\r
651.DrawStrip_vs_hiprio:\r
652 tst r10, #0x00c00000\r
653 beq .DrawStrip_vs_hiprio_maybempt\r
654 sub r0, r1, r11\r
655 orr r7, r7, r0, lsl #16\r
656 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
657 tst r7, #0x1000\r
658 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
659 str r7, [r6], #4 @ cache hi priority tile\r
660 mov r0, #0xf\r
661 b .dsloop_vs\r
662\r
663.DrawStrip_vs_hiprio_maybempt:\r
664 cmp r7, r9\r
665 beq .dsloop_vs @ must've been empty, otherwise we wouldn't get here\r
666 movs r2, r7, lsl #20 @ if (code&0x1000)\r
667 mov r2, r2, lsl #1\r
668 add r2, r2, r10, lsl #17\r
669 mov r2, r2, lsr #17\r
670 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
671 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
672 mov r9, r7 @ remember code\r
673 tst r2, r2\r
674 orrne r10, r10, #1<<22\r
675 bne .DrawStrip_vs_hiprio\r
676 b .dsloop_vs\r
677\r
740da8c6 678.dsloop_vs_exit:\r
7b802576 679 tst r8, #(1<<24) @ seen non hi-prio tile\r
740da8c6 680 ldreq r1, =rendstatus\r
681 mov r0, #0\r
682 ldreq r2, [r1]\r
683 str r0, [r6] @ terminate the cache list\r
283fec1b 684 orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
740da8c6 685 streq r2, [r1]\r
686\r
687 ldmfd sp!, {r4-r11,lr}\r
688 bx lr\r
689\r
6d7acf9e 690\r
691@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
cc68a136 692\r
693@ interlace mode 2? Sonic 2?\r
694.DrawStrip_interlace:\r
695 tst r0, r0\r
696 moveq r7, r7, lsl #21\r
697 movne r7, r7, lsl #5\r
698\r
699 @ Find the line in the name table\r
b6d7ac70 700 add r2, r7, r2, lsl #22 @ r2=(vscroll+(DrawScanline<<1))<<21 (11 bits);\r
cc68a136 701 orr r1, r1, #0x80000000\r
702 and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
703 mov r2, r2, lsr #21\r
704 mov r4, r2, lsr #4\r
705 mov r12, r12, lsr #1 @ halfwords\r
706 add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];\r
707 and r9, r9, #0xff\r
708\r
709 sub sp, sp, #6*4\r
710 stmia sp, {r0,r2,r3,r5,r6,r9}\r
711\r
712 mov r0, sp\r
713 bl DrawStripInterlace @ struct TileStrip *ts\r
714\r
715 add sp, sp, #6*4\r
716 ldmfd sp!, {r4-r11,lr}\r
717 bx lr\r
718\r
719.pool\r
720\r
721@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
722\r
723\r
724.global BackFill @ int reg7, int sh\r
725\r
726BackFill:\r
727 stmfd sp!, {r4-r9,lr}\r
728\r
ea8c405f 729.if OVERRIDE_HIGHCOL\r
730 ldr lr, =HighCol\r
731 mov r0, r0, lsl #26\r
732 ldr lr, [lr]\r
733 mov r0, r0, lsr #26\r
734 add lr, lr, #8\r
735.else\r
cc68a136 736 ldr lr, =(HighCol+8)\r
cc68a136 737 mov r0, r0, lsl #26\r
738 mov r0, r0, lsr #26\r
ea8c405f 739.endif\r
740\r
cc68a136 741 orr r0, r0, r1, lsl #6\r
742 orr r0, r0, r0, lsl #8\r
743 orr r0, r0, r0, lsl #16\r
744\r
745 mov r1, r0\r
746 mov r2, r0\r
747 mov r3, r0\r
748 mov r4, r0\r
749 mov r5, r0\r
750 mov r6, r0\r
751 mov r7, r0\r
752\r
753 @ go go go!\r
754 stmia lr!, {r0-r7} @ 10*8*4\r
755 stmia lr!, {r0-r7}\r
756 stmia lr!, {r0-r7}\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\r
765 ldmfd sp!, {r4-r9,r12}\r
766 bx r12\r
767\r
768\r
769@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
770\r
771\r
7a7c6476 772.global DrawTilesFromCache @ int *hc, int sh, int rlim\r
cc68a136 773\r
774DrawTilesFromCache:\r
775 stmfd sp!, {r4-r8,r11,lr}\r
776\r
cc68a136 777 @ cache some stuff to avoid mem access\r
ea8c405f 778.if OVERRIDE_HIGHCOL\r
cc68a136 779 ldr r11,=HighCol\r
cc68a136 780 mov r12,#0xf\r
ea8c405f 781 ldr r11,[r11]\r
782.else\r
783 ldr r11,=HighCol\r
784 mov r12,#0xf\r
785.endif\r
786 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
cc68a136 787\r
740da8c6 788 mvn r5, #0 @ r5=prevcode=-1\r
7a7c6476 789 ands r8, r1, #1\r
790 orr r8, r8, r2, lsl #1\r
740da8c6 791 bne .dtfc_check_rendflags\r
792\r
cc68a136 793 @ scratch: r4, r7\r
794.dtfc_loop:\r
795 ldr r6, [r0], #4 @ read code\r
796 movs r1, r6, lsr #16 @ r1=dx;\r
797 ldmeqfd sp!, {r4-r8,r11,pc} @ dx is never zero, this must be a terminator, return\r
7a7c6476 798 bic r4, r1, #0xfe00\r
799 add r1, r11, r4 @ r1=pdest\r
cc68a136 800\r
801 mov r7, r6, lsl #16\r
802 cmp r5, r7, lsr #16\r
803 beq .dtfc_samecode @ if (code==prevcode)\r
804\r
805 mov r5, r7, lsr #16\r
806\r
807 mov r2, r5, lsl #21\r
808 mov r2, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4;\r
809 add r2, r2, r6, lsr #25 @ addr+=ty\r
810\r
811 and r3, r5, #0x6000\r
812 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
813\r
814 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
815\r
816.dtfc_samecode:\r
7a7c6476 817 rsbs r4, r4, r8, lsr #1\r
818 bmi .dtfc_cut_tile\r
819\r
820 tst r8, #1\r
cc68a136 821 bne .dtfc_shadow\r
822\r
823 tst r2, r2\r
824 beq .dtfc_loop\r
825\r
826 cmp r2, r2, ror #4\r
827 beq .dtfc_SingleColor @ tileline singlecolor \r
828\r
829 tst r5, #0x0800\r
7a7c6476 830 bne .dtfc_TileFlip\r
cc68a136 831\r
832 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 833.dtfc_TileNorm:\r
834 TileNorm r12\r
835 b .dtfc_loop\r
836\r
7a7c6476 837.dtfc_TileFlip:\r
838 TileFlip r12\r
839 b .dtfc_loop\r
840\r
cc68a136 841.dtfc_SingleColor:\r
842 and r4, r2, #0xf\r
843 orr r4, r3, r4\r
844 orr r4, r4, r4, lsl #8\r
845 tst r1, #1 @ not aligned?\r
846 strneb r4, [r1], #1\r
847 streqh r4, [r1], #2\r
848 strh r4, [r1], #2\r
849 strh r4, [r1], #2\r
850 strh r4, [r1], #2\r
851 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
852 b .dtfc_loop\r
853\r
854.dtfc_shadow:\r
855 tst r2, r2\r
856 beq .dtfc_shadow_blank\r
857\r
858 cmp r2, r2, ror #4\r
859 beq .dtfc_SingleColor @ tileline singlecolor \r
860\r
861 tst r5, #0x0800\r
7a7c6476 862 bne .dtfc_TileFlipShHP\r
cc68a136 863\r
864 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 865.dtfc_TileNormShHP:\r
866 TileNormShHP\r
867 b .dtfc_loop\r
868\r
7a7c6476 869.dtfc_TileFlipShHP:\r
870 TileFlipShHP\r
871 b .dtfc_loop\r
872\r
cc68a136 873.dtfc_shadow_blank:\r
07abbab1 874 tst r1, #1\r
875 ldrneb r4, [r1]\r
876 mov r6, #0x3f\r
877 and r4, r4, #0x3f\r
878 strneb r4, [r1], #1\r
879 ldrh r4, [r1]\r
880 orr r6, r6, r6, lsl #8\r
881 and r4, r4, r6\r
882 strh r4, [r1], #2\r
883 ldrh r4, [r1]\r
884 and r4, r4, r6\r
885 strh r4, [r1], #2\r
886 ldrh r4, [r1]\r
887 and r4, r4, r6\r
888 strh r4, [r1], #2\r
889 ldrh r4, [r1]\r
890 and r4, r4, r6\r
891 streqh r4, [r1]\r
892 strneb r4, [r1]\r
cc68a136 893 b .dtfc_loop\r
894\r
7a7c6476 895.dtfc_cut_tile:\r
896 add r4, r4, #7 @ 0-6\r
897 mov r4, r4, lsl #2\r
898 mov r12,#0xf<<28\r
899 mov r12,r12,asr r4\r
900 mov r2, r2, ror #16\r
901 tst r5, #0x0800 @ flipped?\r
902 mvnne r12,r12\r
903 and r2, r2, r12\r
904 mov r2, r2, ror #16\r
905 mov r12,#0xf\r
906 tst r8, #1\r
907 bne .dtfc_shadow\r
908 tst r2, r2\r
909 beq .dtfc_loop\r
910 tst r5, #0x0800\r
911 beq .dtfc_TileNorm\r
912 b .dtfc_TileFlip\r
913\r
740da8c6 914@ check if we have detected layer covered with hi-prio tiles:\r
915.dtfc_check_rendflags:\r
916 ldr r1, =rendstatus\r
917 ldr r2, [r1]\r
283fec1b 918 tst r2, #(PDRAW_PLANE_HI_PRIO|PDRAW_SHHI_DONE)\r
740da8c6 919 beq .dtfc_loop\r
7a7c6476 920 bic r8, r8, #1 @ sh/hi mode off\r
283fec1b 921 tst r2, #PDRAW_SHHI_DONE\r
740da8c6 922 bne .dtfc_loop @ already processed\r
283fec1b 923 orr r2, r2, #PDRAW_SHHI_DONE\r
740da8c6 924 str r2, [r1]\r
925\r
926 add r1, r11,#8\r
07abbab1 927 mov r3, #320/4/4\r
740da8c6 928 mov r6, #0x3f\r
929 orr r6, r6, r6, lsl #8\r
930 orr r6, r6, r6, lsl #16\r
931.dtfc_loop_shprep:\r
07abbab1 932 ldmia r1, {r2,r4,r5,r7}\r
740da8c6 933 subs r3, r3, #1\r
07abbab1 934 and r2, r2, r6\r
935 and r4, r4, r6\r
936 and r5, r5, r6\r
937 and r7, r7, r6\r
938 stmia r1!,{r2,r4,r5,r7}\r
939 bne .dtfc_loop_shprep\r
940\r
941 mvn r5, #0 @ r5=prevcode=-1\r
942 b .dtfc_loop\r
740da8c6 943\r
cc68a136 944.pool\r
945\r
946@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
947\r
948\r
fbc65db7 949.global DrawSpritesSHi @ unsigned char *sprited\r
e352c3af 950\r
951DrawSpritesSHi:\r
fbc65db7 952 ldr r3, [r0]\r
953 mov r12,#0xff\r
e352c3af 954 ands r3, r3, #0x7f\r
955 bxeq lr\r
cc68a136 956\r
cc68a136 957 stmfd sp!, {r4-r11,lr}\r
fbc65db7 958 strb r12,[r0,#2] @ set end marker\r
959 add r10,r0, #3 @ r10=HighLnSpr end\r
e352c3af 960 add r10,r10,r3 @ r10=HighLnSpr end\r
961\r
ea8c405f 962.if OVERRIDE_HIGHCOL\r
cc68a136 963 ldr r11,=HighCol\r
ea8c405f 964 mov r12,#0xf\r
965 ldr r11,[r11]\r
966.else\r
967 ldr r11,=HighCol\r
968 mov r12,#0xf\r
969.endif\r
cc68a136 970 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
cc68a136 971\r
cc68a136 972\r
e352c3af 973DrawSpriteSHi:\r
974 @ draw next sprite\r
975 ldrb r0, [r10,#-1]!\r
976 ldr r1, =HighPreSpr\r
fbc65db7 977@ ldr r8, [sp, #-4]\r
e352c3af 978 cmp r0, #0xff\r
979 ldmeqfd sp!, {r4-r11,pc} @ end of list\r
980 and r0, r0, #0x7f\r
981 add r0, r1, r0, lsl #3\r
982\r
983 ldr r9, [r0, #4] @ sprite[1]\r
984 mov r2, r9, asr #16 @ r2=sx\r
cc68a136 985\r
e352c3af 986 mov r9, r9, lsl #16\r
987 mov r3, r9, lsr #31 @ priority\r
988 mov r9, r9, lsr #16\r
fbc65db7 989@ orr r9, r9, r8, lsl #31 @ r9=code|sh[31] @@ sh is always on here now\r
e352c3af 990 and r4, r9, #0x6000\r
991 orr r9, r9, r4, lsl #16\r
fbc65db7 992 orr r9, r9, #0x90000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
e352c3af 993 cmp r12,r9, lsr #28 @ sh/hi with pal3?\r
fbc65db7 994 cmpne r3, #1 @ if not, is ir hi prio?\r
e352c3af 995 bne DrawSpriteSHi @ non-operator low sprite, already drawn\r
cc68a136 996\r
e352c3af 997 ldr r3, [r0] @ sprite[0]\r
998 ldr r7, =DrawScanline\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
cc68a136 1003\r
e352c3af 1004 mov r0, r3, lsl #16 @ r4=sy<<16 (tmp)\r
cc68a136 1005\r
e352c3af 1006 ldr r7, [r7]\r
1007 sub r7, r7, r0, asr #16 @ r7=row=DrawScanline-sy\r
cc68a136 1008\r
e352c3af 1009 tst r9, #0x1000\r
1010 movne r0, r5, lsl #3\r
1011 subne r0, r0, #1\r
1012 subne r7, r0, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
cc68a136 1013\r
e352c3af 1014 add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
1015 tst r9, #0x0800\r
1016 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1017 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
cc68a136 1018\r
e352c3af 1019 mov r8, r8, lsl #21\r
1020 mov r8, r8, lsr #17\r
1021 and r7, r7, #7\r
1022 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1023\r
1024 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
e352c3af 1025 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
1026\r
1027 add r6, r6, #1 @ inc now\r
1028 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1029 b .dsprShi_loop_enter\r
1030\r
1031.dsprShi_loop:\r
1032 subs r6, r6, #1 @ width--\r
1033 beq DrawSpriteSHi\r
1034 adds r0, r0, #8 @ sx+=8\r
1035 add r8, r8, r5 @ tile+=delta\r
1036\r
1037.dsprShi_loop_enter:\r
1038 ble .dsprShi_loop @ sx <= 0\r
cc68a136 1039 cmp r0, #328\r
e352c3af 1040 bge DrawSpriteSHi\r
cc68a136 1041\r
1042 mov r8, r8, lsl #17\r
e352c3af 1043 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
cc68a136 1044\r
e352c3af 1045 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1046 add r1, r11, r0 @ r1=pdest\r
cc68a136 1047 tst r2, r2\r
e352c3af 1048 beq .dsprShi_loop\r
cc68a136 1049\r
e352c3af 1050 cmp r12, r9, lsr #28\r
1051 beq .dsprShi_shadow\r
cc68a136 1052\r
1053 cmp r2, r2, ror #4\r
e352c3af 1054 beq .dsprShi_SingleColor @ tileline singlecolor \r
cc68a136 1055\r
e352c3af 1056 tst r9, #0x0800\r
1057 bne .dsprShi_TileFlip\r
cc68a136 1058\r
e352c3af 1059 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1060@ scratch: r4, r7\r
1061.dsprShi_TileNorm:\r
cc68a136 1062 TileNorm r12\r
e352c3af 1063 b .dsprShi_loop\r
cc68a136 1064\r
e352c3af 1065.dsprShi_TileFlip:\r
7a7c6476 1066 TileFlip r12\r
e352c3af 1067 b .dsprShi_loop\r
7a7c6476 1068\r
e352c3af 1069.dsprShi_SingleColor:\r
cc68a136 1070 and r4, r2, #0xf\r
1071 orr r4, r3, r4\r
1072 orr r4, r4, r4, lsl #8\r
e352c3af 1073 tst r0, #1 @ not aligned?\r
cc68a136 1074 strneb r4, [r1], #1\r
1075 streqh r4, [r1], #2\r
1076 strh r4, [r1], #2\r
1077 strh r4, [r1], #2\r
1078 strh r4, [r1], #2\r
1079 strneb r4, [r1], #1\r
e352c3af 1080 b .dsprShi_loop\r
cc68a136 1081\r
e352c3af 1082.dsprShi_shadow:\r
1083 tst r9, #0x8000\r
1084 beq .dsprShi_shadow_lowpri\r
07abbab1 1085\r
cc68a136 1086 cmp r2, r2, ror #4\r
e352c3af 1087 beq .dsprShi_singlec_sh\r
cc68a136 1088\r
e352c3af 1089 tst r9, #0x0800\r
1090 bne .dsprShi_TileFlip_sh\r
cc68a136 1091\r
1092 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
e352c3af 1093.dsprShi_TileNorm_sh:\r
cc68a136 1094 TileNormSh\r
e352c3af 1095 b .dsprShi_loop\r
cc68a136 1096\r
e352c3af 1097.dsprShi_TileFlip_sh:\r
7a7c6476 1098 TileFlipSh\r
e352c3af 1099 b .dsprShi_loop\r
7a7c6476 1100\r
e352c3af 1101.dsprShi_singlec_sh:\r
cc68a136 1102 cmp r2, #0xe0000000\r
e352c3af 1103 bcc .dsprShi_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
cc68a136 1104 tst r2, #0x10000000\r
e352c3af 1105 bne .dsprShi_sh_sh\r
cc68a136 1106 TileSingleHi\r
e352c3af 1107 b .dsprShi_loop\r
cc68a136 1108\r
e352c3af 1109.dsprShi_sh_sh:\r
cc68a136 1110 TileSingleSh\r
e352c3af 1111 b .dsprShi_loop\r
cc68a136 1112\r
e352c3af 1113.dsprShi_shadow_lowpri:\r
1114 tst r9, #0x800\r
1115 bne .dsprShi_TileFlip_sh_lp\r
07abbab1 1116\r
e352c3af 1117.dsprShi_TileNorm_sh_lp:\r
07abbab1 1118 TileNormSh_onlyop_lp\r
e352c3af 1119 b .dsprShi_loop\r
07abbab1 1120\r
e352c3af 1121.dsprShi_TileFlip_sh_lp:\r
07abbab1 1122 TileFlipSh_onlyop_lp\r
e352c3af 1123 b .dsprShi_loop\r
07abbab1 1124\r
cc68a136 1125.pool\r
1126\r
1127@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1128\r
fbc65db7 1129.global DrawAllSprites @ unsigned char *sprited, int prio, int sh\r
283fec1b 1130\r
1131DrawAllSprites:\r
1132 ldr r3, =rendstatus\r
fbc65db7 1133 orr r1, r2, r1, lsl #1\r
283fec1b 1134 ldr r12,[r3]\r
0fc0e241 1135 tst r12,#(PDRAW_DIRTY_SPRITES|PDRAW_SPRITES_MOVED)\r
283fec1b 1136 beq das_no_prep\r
fbc65db7 1137 stmfd sp!, {r0,r1,lr}\r
283fec1b 1138 and r0, r12,#PDRAW_DIRTY_SPRITES\r
0fc0e241 1139 bic r12,r12,#(PDRAW_DIRTY_SPRITES|PDRAW_SPRITES_MOVED)\r
283fec1b 1140 str r12,[r3]\r
1141 bl PrepareSprites\r
fbc65db7 1142 ldmfd sp!, {r0,r1,lr}\r
283fec1b 1143\r
1144das_no_prep:\r
fbc65db7 1145 ldr r3, [r0]\r
283fec1b 1146 ands r3, r3, #0x7f\r
283fec1b 1147 bxeq lr\r
1148\r
1149 @ time to do some real work\r
1150 stmfd sp!, {r4-r11,lr}\r
283fec1b 1151 mov r12,#0xff\r
fbc65db7 1152 strb r12,[r0,#2] @ set end marker\r
1153 add r10,r0, #3\r
283fec1b 1154 add r10,r10,r3 @ r10=HighLnSpr end\r
1155\r
97a7f774 1156 str r1, [sp, #-4] @ no calls after this point\r
283fec1b 1157\r
1158.if OVERRIDE_HIGHCOL\r
1159 ldr r11,=HighCol\r
1160 mov r12,#0xf\r
1161 ldr r11,[r11]\r
1162.else\r
1163 ldr r11,=HighCol\r
1164 mov r12,#0xf\r
1165.endif\r
1166 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
1167\r
1168@ + 0 : hhhhvvvv ----hhvv yyyyyyyy yyyyyyyy // v, h: horiz. size\r
cc68a136 1169@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
1170\r
283fec1b 1171DrawSprite: @ was: unsigned int *sprite, int sh, int acc_sprites\r
1172 @ draw next sprite\r
1173 ldrb r0, [r10,#-1]!\r
1174 ldr r1, =HighPreSpr\r
283fec1b 1175 ldr r8, [sp, #-4]\r
e352c3af 1176 mov r2, r0, lsr #7\r
1177 cmp r0, #0xff\r
1178 ldmeqfd sp!, {r4-r11,pc} @ end of list\r
97a7f774 1179 cmp r2, r8, lsr #1\r
e352c3af 1180 bne DrawSprite @ wrong priority\r
283fec1b 1181 and r0, r0, #0x7f\r
1182 add r0, r1, r0, lsl #3\r
cc68a136 1183\r
283fec1b 1184@ stmfd sp!, {r4-r9,r11,lr}\r
1185@ orr r8, r2, r1, lsl #4\r
cc68a136 1186\r
1187 ldr r3, [r0] @ sprite[0]\r
b6d7ac70 1188 ldr r7, =DrawScanline\r
cc68a136 1189 mov r6, r3, lsr #28\r
1190 sub r6, r6, #1 @ r6=width-1 (inc later)\r
1191 mov r5, r3, lsr #24\r
1192 and r5, r5, #7 @ r5=height\r
1193\r
1194 mov r4, r3, lsl #16 @ r4=sy<<16 (tmp)\r
1195\r
1196 ldr r7, [r7]\r
1197 ldr r9, [r0, #4]\r
b6d7ac70 1198 sub r7, r7, r4, asr #16 @ r7=row=DrawScanline-sy\r
cc68a136 1199\r
cc68a136 1200 mov r2, r9, asr #16 @ r2=sx\r
e5fa9817 1201 mov r9, r9, lsl #16\r
1202 mov r9, r9, lsr #16\r
97a7f774 1203 orr r9, r9, r8, lsl #31 @ r9=code|sh[31]\r
cc68a136 1204\r
1205 tst r9, #0x1000\r
1206 movne r4, r5, lsl #3\r
1207 subne r4, r4, #1\r
1208 subne r7, r4, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
1209\r
07abbab1 1210 add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
cc68a136 1211 tst r9, #0x0800\r
1212 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1213 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
1214\r
07abbab1 1215 mov r8, r8, lsl #21\r
1216 mov r8, r8, lsr #17\r
cc68a136 1217 and r7, r7, #7\r
1218 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1219\r
e5fa9817 1220.dspr_continue:\r
cc68a136 1221 @ cache some stuff to avoid mem access\r
cc68a136 1222 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
1223 and r4, r9, #0x6000\r
1224 orr r9, r9, r4, lsl #16\r
97a7f774 1225 orrs r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
cc68a136 1226\r
cc68a136 1227 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
97a7f774 1228 orrmi r3, r3, #0x40 @ for sh/hi\r
cc68a136 1229\r
1230 add r6, r6, #1 @ inc now\r
1231 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1232 b .dspr_loop_enter\r
1233\r
1234.dspr_loop:\r
1235 subs r6, r6, #1 @ width--\r
283fec1b 1236 beq DrawSprite\r
cc68a136 1237 adds r0, r0, #8 @ sx+=8\r
1238 add r8, r8, r5 @ tile+=delta\r
1239\r
1240.dspr_loop_enter:\r
1241 ble .dspr_loop @ sx <= 0\r
1242 cmp r0, #328\r
283fec1b 1243 bge DrawSprite\r
cc68a136 1244\r
1245 mov r8, r8, lsl #17\r
1246 mov r8, r8, lsr #17 @ tile&=0x7fff; // Clip tile address\r
1247\r
1248 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1249 add r1, r11, r0 @ r1=pdest\r
1250 tst r2, r2\r
1251 beq .dspr_loop\r
1252\r
1253 cmp r12, r9, lsr #28\r
1254 beq .dspr_shadow\r
1255\r
1256 cmp r2, r2, ror #4\r
1257 beq .dspr_SingleColor @ tileline singlecolor \r
1258\r
1259 tst r9, #0x0800\r
7a7c6476 1260 bne .dspr_TileFlip\r
cc68a136 1261\r
1262 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 1263@ scratch: r4, r7\r
1264.dspr_TileNorm:\r
1265 TileNorm r12\r
1266 b .dspr_loop\r
1267\r
7a7c6476 1268.dspr_TileFlip:\r
1269 TileFlip r12\r
1270 b .dspr_loop\r
1271\r
07abbab1 1272.dspr_singlec_sh:\r
1273 cmp r2, #0xe0000000\r
1274 bcs .dspr_loop @ operator tileline, ignore\r
1275\r
cc68a136 1276.dspr_SingleColor:\r
1277 and r4, r2, #0xf\r
1278 orr r4, r3, r4\r
1279 orr r4, r4, r4, lsl #8\r
1280 tst r0, #1 @ not aligned?\r
1281 strneb r4, [r1], #1\r
1282 streqh r4, [r1], #2\r
1283 strh r4, [r1], #2\r
1284 strh r4, [r1], #2\r
1285 strh r4, [r1], #2\r
1286 strneb r4, [r1], #1\r
1287 b .dspr_loop\r
1288\r
1289.dspr_shadow:\r
1290 cmp r2, r2, ror #4\r
1291 beq .dspr_singlec_sh\r
1292\r
1293 tst r9, #0x0800\r
7a7c6476 1294 bne .dspr_TileFlip_sh\r
cc68a136 1295\r
1296 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
cc68a136 1297.dspr_TileNorm_sh:\r
07abbab1 1298 TileNormSh_noop\r
cc68a136 1299 b .dspr_loop\r
1300\r
7a7c6476 1301.dspr_TileFlip_sh:\r
07abbab1 1302 TileFlipSh_noop\r
cc68a136 1303 b .dspr_loop\r
1304\r
1305\r
cc68a136 1306@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1307\r
1308.global DrawWindow @ int tstart, int tend, int prio, int sh // int *hcache\r
1309\r
1310DrawWindow:\r
1311 stmfd sp!, {r4-r11,lr}\r
1312\r
1313 ldr r11, =(Pico+0x22228) @ Pico.video\r
b6d7ac70 1314 ldr r10, =DrawScanline\r
cc68a136 1315 ldrb r12, [r11, #3] @ pvid->reg[3]\r
1316\r
1317 ldr r10, [r10]\r
1318 ldr r4, [r11, #12]\r
1319 mov r5, r10, lsr #3\r
1320 and r10, r10, #7\r
1321 mov r10, r10, lsl #1 @ r10=ty\r
1322\r
1323 mov r12, r12, lsl #10\r
1324\r
1325 tst r4, #1 @ 40 cell mode?\r
1326 andne r12, r12, #0xf000 @ 0x3c<<10\r
1327 andeq r12, r12, #0xf800\r
1328 addne r12, r12, r5, lsl #7\r
1329 addeq r12, r12, r5, lsl #6 @ nametab\r
1330 add r12, r12, r0, lsl #2 @ +starttile\r
1331\r
1332 ldr r6, =rendstatus\r
0fc0e241 1333 ldr lr, =(Pico+0x10000) @ lr=Pico.vram\r
07abbab1 1334 ldr r6, [r6]\r
cc68a136 1335\r
1336 @ fetch the first code now\r
1337 ldrh r7, [lr, r12]\r
1338\r
283fec1b 1339 ands r6, r6, #PDRAW_WND_DIFF_PRIO\r
cc68a136 1340 orr r6, r6, r2\r
cc68a136 1341\r
7292c709 1342 eoreq r8, r2, r7, lsr #15 @ do prio bits differ?\r
1343 cmpeq r8, #1\r
1344 ldmeqfd sp!, {r4-r11,pc} @ yes, assume that whole window uses same priority\r
cc68a136 1345\r
cc68a136 1346 orr r6, r6, r3, lsl #8 @ shadow mode\r
1347\r
1348 sub r8, r1, r0\r
cc68a136 1349\r
1350 @ cache some stuff to avoid mem access\r
ea8c405f 1351.if OVERRIDE_HIGHCOL\r
1352 ldr r11,=HighCol\r
1353 mov r8, r8, lsl #1 @ cells\r
1354 ldr r11,[r11]\r
1355 mvn r9, #0 @ r9=prevcode=-1\r
1356 add r11,r11,#8\r
1357.else\r
cc68a136 1358 ldr r11,=(HighCol+8)\r
ea8c405f 1359 mov r8, r8, lsl #1 @ cells\r
1360 mvn r9, #0 @ r9=prevcode=-1\r
1361.endif\r
07abbab1 1362 add r1, r11, r0, lsl #4 @ r1=pdest\r
cc68a136 1363 mov r0, #0xf\r
1364 b .dwloop_enter\r
1365\r
07abbab1 1366 @ r4,r5 are scratch in this loop\r
cc68a136 1367.dwloop:\r
1368 add r1, r1, #8\r
1369.dwloop_nor1:\r
1370 add r12, r12, #2 @ halfwords\r
1371 ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)\r
1372 subs r8, r8, #1\r
1373 beq .dwloop_end @ done\r
1374\r
1375 eor r5, r6, r7, lsr #15\r
1376 tst r5, #1\r
1377 orrne r6, r6, #2 @ wrong pri\r
1378 bne .dwloop\r
1379\r
1380 cmp r7, r9\r
1381 beq .dw_samecode @ we know stuff about this tile already\r
1382\r
1383.dwloop_enter:\r
1384 mov r9, r7 @ remember code\r
1385\r
1386 movs r2, r9, lsl #20 @ if (code&0x1000)\r
1387 mov r2, r2, lsl #1\r
1388 add r2, r10, r2, lsr #17 @ r2=addr=(code&0x7ff)<<4; addr+=ty\r
1389 eorcs r2, r2, #0x0e @ if (code&0x1000) addr^=0xe;\r
1390\r
1391 and r3, r9, #0x6000\r
1392 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
1393\r
1394 ldr r2, [lr, r2, lsl #1] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1395\r
1396.dw_samecode:\r
1397 tst r6, #0x100\r
1398 bne .dw_shadow\r
1399.dw_shadow_done:\r
1400 tst r2, r2\r
1401 beq .dwloop @ tileline blank\r
1402\r
1403 cmp r2, r2, ror #4\r
1404 beq .dw_SingleColor @ tileline singlecolor \r
1405\r
1406 tst r9, #0x0800\r
7a7c6476 1407 bne .dw_TileFlip\r
cc68a136 1408\r
1409 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
cc68a136 1410.dw_TileNorm:\r
1411 TileNorm r0\r
1412 b .dwloop\r
1413\r
7a7c6476 1414.dw_TileFlip:\r
1415 TileFlip r0\r
1416 b .dwloop\r
1417\r
cc68a136 1418.dw_SingleColor:\r
1419 and r4, r0, r2 @ #0x0000000f\r
1420 orr r4, r3, r4\r
1421 orr r4, r4, r4, lsl #8\r
1422 orr r4, r4, r4, lsl #16\r
1423 mov r5, r4\r
1424 stmia r1!, {r4,r5}\r
1425 b .dwloop_nor1 @ we incremeted r1 ourselves\r
1426\r
1427.dw_shadow:\r
1428 tst r6, #1 @ hi pri?\r
1429 orreq r3, r3, #0x40\r
1430 beq .dw_shadow_done\r
1431 ldr r4, [r1]\r
07abbab1 1432 mov r5, #0x3f\r
1433 orr r5, r5, r5, lsl #8\r
1434 orr r5, r5, r5, lsl #16\r
1435 and r4, r4, r5\r
cc68a136 1436 str r4, [r1]\r
1437 ldr r4, [r1,#4]\r
07abbab1 1438 and r4, r4, r5\r
cc68a136 1439 str r4, [r1,#4]\r
1440 b .dw_shadow_done\r
1441\r
1442.dwloop_end:\r
1443 ldr r0, =rendstatus\r
1444 ldr r1, [r0]\r
0fc0e241 1445 and r6, r6, #PDRAW_WND_DIFF_PRIO\r
cc68a136 1446 orr r1, r1, r6\r
1447 str r1, [r0]\r
1448\r
1449 ldmfd sp!, {r4-r11,r12}\r
1450 bx r12\r
1451\r
1452\r
1453@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1454\r
1455\r
1456@ hilights 2 pixels in RGB444/BGR444 format\r
1457.macro TileDoShHi2Pixels444 reg\r
1458 mov \reg, \reg, ror #12\r
1459 adds \reg, \reg, #0x40000000\r
1460 orrcs \reg, \reg, #0xf0000000\r
1461 mov \reg, \reg, ror #28\r
1462 adds \reg, \reg, #0x40000000\r
1463 orrcs \reg, \reg, #0xf0000000\r
1464 mov \reg, \reg, ror #28\r
1465 adds \reg, \reg, #0x40000000\r
1466 orrcs \reg, \reg, #0xf0000000\r
1467 mov \reg, \reg, ror #24\r
1468 adds \reg, \reg, #0x40000000\r
1469 orrcs \reg, \reg, #0xf0000000\r
1470 mov \reg, \reg, ror #28\r
1471 adds \reg, \reg, #0x40000000\r
1472 orrcs \reg, \reg, #0xf0000000\r
1473 mov \reg, \reg, ror #28\r
1474 adds \reg, \reg, #0x40000000\r
1475 orrcs \reg, \reg, #0xf0000000\r
1476 mov \reg, \reg, ror #12\r
1477.endm\r
1478\r
1479\r
1480.global FinalizeLineBGR444 @ int sh\r
1481\r
1482FinalizeLineBGR444:\r
1483 stmfd sp!, {r4-r6,lr}\r
1484 mov r6, r0\r
1485 ldr lr, =(Pico+0x22228) @ Pico.video\r
1486 ldr r0, =DrawLineDest\r
1487 ldrb r12, [lr, #12]\r
1488 ldr r0, [r0]\r
1489 sub r3, lr, #0x128 @ r3=Pico.cram\r
1490\r
1491 tst r12, #1\r
1492 movne r2, #320/4 @ len\r
1493 bne .fl_no32colBGR444\r
1494 ldr r4, =PicoOpt\r
1495 mov r2, #256/4\r
1496 ldr r4, [r4]\r
1497 tst r4, #0x100\r
1498 addeq r0, r0, #32*2\r
1499\r
1500.fl_no32colBGR444:\r
1501 tst r6, r6\r
1502 beq .fl_noshBGR444\r
1503\r
1504 ldr r4, =HighPal\r
1505\r
1506 ldrb r12, [lr, #-0x1a] @ 0x2220e ~ dirtyPal\r
1507 tst r12, r12\r
1508 moveq r3, r4\r
1509 beq .fl_noshBGR444\r
1510 mov r12, #0\r
1511 strb r12, [lr, #-0x1a]\r
1512\r
1513 mov lr, #0x40/8\r
1514 @ copy pal:\r
1515.fl_loopcpBGR444:\r
1516 ldmia r3!, {r1,r5,r6,r12}\r
1517 subs lr, lr, #1\r
1518 stmia r4!, {r1,r5,r6,r12}\r
1519 bne .fl_loopcpBGR444\r
1520\r
1521 @ shadowed pixels:\r
1522 mov r12, #0x0077\r
1523 orr r12,r12,#0x0700\r
1524 orr r12,r12,r12,lsl #16\r
1525 sub r3, r3, #0x40*2\r
1526 add r5, r4, #0x80*2\r
1527 mov lr, #0x40/4\r
1528.fl_loopcpBGR444_sh:\r
1529 ldmia r3!, {r1,r6}\r
1530 subs lr, lr, #1\r
1531 and r1, r12, r1, lsr #1\r
1532 and r6, r12, r6, lsr #1\r
1533 stmia r4!, {r1,r6}\r
1534 stmia r5!, {r1,r6}\r
1535 bne .fl_loopcpBGR444_sh\r
1536\r
1537 @ hilighted pixels:\r
1538 sub r3, r3, #0x40*2\r
1539 mov lr, #0x40/2\r
1540.fl_loopcpBGR444_hi:\r
1541 ldr r1, [r3], #4\r
1542 TileDoShHi2Pixels444 r1\r
1543 str r1, [r4], #4\r
1544 subs lr, lr, #1\r
1545 bne .fl_loopcpBGR444_hi\r
1546\r
1547 sub r3, r4, #0x40*3*2\r
e5fa9817 1548 mov r6, #1\r
cc68a136 1549\r
1550\r
1551.fl_noshBGR444:\r
e5fa9817 1552 ldr r12,=rendstatus\r
1553 eors r6, r6, #1 @ sh is 0\r
1554 ldr r12,[r12]\r
1555 mov lr, #0xff\r
283fec1b 1556 tstne r12,#PDRAW_ACC_SPRITES\r
e5fa9817 1557\r
ea8c405f 1558.if OVERRIDE_HIGHCOL\r
1559 ldr r1, =HighCol\r
e5fa9817 1560 movne lr, #0x3f\r
ea8c405f 1561 ldr r1, [r1]\r
1562 mov lr, lr, lsl #1\r
1563 add r1, r1, #8\r
1564.else\r
cc68a136 1565 ldr r1, =(HighCol+8)\r
e5fa9817 1566 movne lr, #0x3f\r
cc68a136 1567 mov lr, lr, lsl #1\r
ea8c405f 1568.endif\r
cc68a136 1569\r
1570.fl_loopBGR444:\r
cc68a136 1571 ldr r12, [r1], #4\r
1572 subs r2, r2, #1\r
1573\r
1574 and r4, lr, r12, lsl #1\r
1575 ldrh r4, [r3, r4]\r
1576 and r5, lr, r12, lsr #7\r
1577 ldrh r5, [r3, r5]\r
1578 and r6, lr, r12, lsr #15\r
1579 ldrh r6, [r3, r6]\r
e5fa9817 1580 and r12,lr, r12, lsr #23\r
1581 ldrh r12,[r3, r12] @ 1c.i.\r
cc68a136 1582 orr r4, r4, r5, lsl #16\r
e5fa9817 1583 orr r5, r6, r12,lsl #16\r
cc68a136 1584\r
1585 stmia r0!, {r4,r5}\r
1586 bne .fl_loopBGR444\r
1587\r
1588\r
1589 ldmfd sp!, {r4-r6,lr}\r
1590 bx lr\r
1591\r
1592\r
1593@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1594\r
1595\r
1596@ hilights 2 pixels in RGB555/BGR555 format\r
1597.macro TileDoShHi2Pixels555 reg\r
1598 adds \reg, \reg, #0x40000000\r
1599 orrcs \reg, \reg, #0xf0000000\r
1600 mov \reg, \reg, ror #27\r
1601 adds \reg, \reg, #0x40000000\r
1602 orrcs \reg, \reg, #0xf0000000\r
1603 mov \reg, \reg, ror #26\r
1604 adds \reg, \reg, #0x40000000\r
1605 orrcs \reg, \reg, #0xf0000000\r
1606 mov \reg, \reg, ror #27\r
1607 adds \reg, \reg, #0x40000000\r
1608 orrcs \reg, \reg, #0xf0000000\r
1609 mov \reg, \reg, ror #27\r
1610 adds \reg, \reg, #0x40000000\r
1611 orrcs \reg, \reg, #0xf0000000\r
1612 mov \reg, \reg, ror #26\r
1613 adds \reg, \reg, #0x40000000\r
1614 orrcs \reg, \reg, #0xf0000000\r
1615 mov \reg, \reg, ror #27\r
1616.endm\r
1617\r
1618\r
1619@ Convert 0000bbb0 ggg0rrr0\r
1620@ to rrrrrggg gggbbbbb\r
1621\r
1622@ r2,r3,r9 - scratch, lr = 0x001c001c, r8 = 0x00030003\r
1623.macro convRGB565 reg\r
1624 and r2, lr, \reg,lsl #1\r
1625 and r9, r8, \reg,lsr #2\r
1626 orr r2, r2, r9 @ r2=red\r
1627 and r3, lr, \reg,lsr #7\r
1628 and r9, r8, \reg,lsr #10\r
1629 orr r3, r3, r9 @ r3=blue\r
1630 and \reg, \reg, lr, lsl #3\r
1631 orr \reg, \reg, \reg,lsl #3 @ green\r
1632 orr \reg, \reg, r2, lsl #11 @ add red back\r
1633 orr \reg, \reg, r3 @ add blue back\r
1634.endm\r
1635\r
b2305d08 1636@ trashes: r2-r9,r12,lr; r0,r1 are advanced\r
1637.macro vidConvCpyRGB565_local\r
2ec14aec 1638 mov r12, r2, lsr #3 @ repeats\r
cc68a136 1639 mov lr, #0x001c0000\r
1640 orr lr, lr, #0x01c @ lr == pattern 0x001c001c\r
1641 mov r8, #0x00030000\r
2ec14aec 1642 orr r8, r8, #0x003\r
cc68a136 1643\r
b2305d08 16440:\r
cc68a136 1645 ldmia r1!, {r4-r7}\r
1646 subs r12, r12, #1\r
1647 convRGB565 r4\r
1648 str r4, [r0], #4\r
1649 convRGB565 r5\r
1650 str r5, [r0], #4\r
1651 convRGB565 r6\r
1652 str r6, [r0], #4\r
1653 convRGB565 r7\r
1654 str r7, [r0], #4\r
1655\r
b2305d08 1656 bgt 0b\r
1657.endm\r
1658\r
1659\r
1660.global vidConvCpyRGB565\r
cc68a136 1661\r
b2305d08 1662vidConvCpyRGB565: @ void *to, void *from, int pixels\r
1663 stmfd sp!, {r4-r9,lr}\r
1664 vidConvCpyRGB565_local\r
cc68a136 1665 ldmfd sp!, {r4-r9,lr}\r
1666 bx lr\r
1667\r
1668\r
b2305d08 1669.global PicoDoHighPal555 @ int sh\r
cc68a136 1670\r
b2305d08 1671PicoDoHighPal555:\r
1672 stmfd sp!, {r4-r9,lr}\r
1673 mov r1, #0\r
3d48f143 1674 ldr r8, =(Pico+0x22228) @ Pico.video\r
cc68a136 1675\r
b2305d08 1676PicoDoHighPal555_nopush:\r
1677 str r1, [sp, #-8] @ is called from FinalizeLineRGB555?\r
1678\r
1679 str r0, [sp, #-4]\r
1680 ldr r0, =HighPal\r
1681\r
cc68a136 1682 mov r1, #0\r
b2305d08 1683 strb r1, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
1684\r
3d48f143 1685 sub r1, r8, #0x128 @ r1=Pico.cram\r
cc68a136 1686 mov r2, #0x40\r
b2305d08 1687 vidConvCpyRGB565_local\r
cc68a136 1688\r
b2305d08 1689 ldr r0, [sp, #-4]\r
1690 tst r0, r0\r
1691 beq PicoDoHighPal555_end\r
1692\r
1693 ldr r3, =HighPal\r
cc68a136 1694\r
1695 @ shadowed pixels:\r
1696 mov r12, #0x008e\r
cc68a136 1697 add r4, r3, #0x40*2\r
b2305d08 1698 orr r12,r12,#0x7300\r
cc68a136 1699 add r5, r3, #0xc0*2\r
b2305d08 1700 orr r12,r12,r12,lsl #16\r
cc68a136 1701 mov lr, #0x40/4\r
1702.fl_loopcpRGB555_sh:\r
1703 ldmia r3!, {r1,r6}\r
1704 subs lr, lr, #1\r
1705 and r1, r12, r1, lsr #1\r
1706 and r6, r12, r6, lsr #1\r
1707 stmia r4!, {r1,r6}\r
1708 stmia r5!, {r1,r6}\r
1709 bne .fl_loopcpRGB555_sh\r
1710\r
1711 @ hilighted pixels:\r
1712 sub r3, r3, #0x40*2\r
1713 mov lr, #0x40/2\r
1714.fl_loopcpRGB555_hi:\r
1715 ldr r1, [r3], #4\r
1716 TileDoShHi2Pixels555 r1\r
1717 str r1, [r4], #4\r
1718 subs lr, lr, #1\r
1719 bne .fl_loopcpRGB555_hi\r
b2305d08 1720 mov r0, #1\r
cc68a136 1721\r
b2305d08 1722PicoDoHighPal555_end:\r
1723 ldr r1, [sp, #-8]\r
1724 tst r1, r1\r
1725 ldmeqfd sp!, {r4-r9,pc}\r
1726\r
1727 ldr r8, =(Pico+0x22228) @ Pico.video\r
1728 b FinalizeLineRGB555_pal_done\r
1729\r
1730\r
1731.global FinalizeLineRGB555 @ int sh\r
1732\r
1733FinalizeLineRGB555:\r
1734 stmfd sp!, {r4-r9,lr}\r
1735 ldr r8, =(Pico+0x22228) @ Pico.video\r
1736\r
1737 ldrb r2, [r8, #-0x1a] @ 0x2220e ~ dirtyPal\r
1738 mov r1, #1\r
1739 tst r2, r2\r
1740 bne PicoDoHighPal555_nopush\r
1741\r
1742FinalizeLineRGB555_pal_done:\r
1743 ldr r3, =HighPal\r
cc68a136 1744\r
e5fa9817 1745 ldr r12,=rendstatus\r
b2305d08 1746 eors r0, r0, #1 @ sh is 0\r
e5fa9817 1747 ldr r12,[r12]\r
1748 mov lr, #0xff\r
283fec1b 1749 tstne r12,#PDRAW_ACC_SPRITES\r
e5fa9817 1750 movne lr, #0x3f\r
1751\r
ea8c405f 1752.if OVERRIDE_HIGHCOL\r
1753 ldr r1, =HighCol\r
1754 ldr r0, =DrawLineDest\r
1755 ldr r1, [r1]\r
1756 ldr r0, [r0]\r
1757 add r1, r1, #8\r
1758.else\r
3d48f143 1759 ldr r0, =DrawLineDest\r
cc68a136 1760 ldr r1, =(HighCol+8)\r
3d48f143 1761 ldr r0, [r0]\r
ea8c405f 1762.endif\r
3d48f143 1763\r
1764 ldrb r12, [r8, #12]\r
cc68a136 1765 mov lr, lr, lsl #1\r
1766\r
3d48f143 1767 tst r12, #1\r
1768 movne r2, #320/8 @ len\r
1769 bne .fl_no32colRGB555\r
1770 ldr r4, =PicoOpt\r
1771 mov r2, #256/8\r
1772 ldr r4, [r4]\r
1773 tst r4, #0x4000\r
1774 bne .fl_32scale_RGB555\r
1775 tst r4, #0x0100\r
1776 addeq r0, r0, #32*2\r
1777\r
1778.fl_no32colRGB555:\r
cc68a136 1779\r
499a0be3 1780.if UNALIGNED_DRAWLINEDEST\r
1781 @ this is basically for Gizmondo, which has unaligned odd lines in the framebuffer\r
1782 tst r0, #2\r
1783 bne .fl_RGB555u\r
1784.endif\r
1785\r
1786.fl_loopRGB555:\r
cc68a136 1787 ldr r12, [r1], #4\r
1788 ldr r7, [r1], #4\r
1789\r
1790 and r4, lr, r12, lsl #1\r
1791 ldrh r4, [r3, r4]\r
1792 and r5, lr, r12, lsr #7\r
1793 ldrh r5, [r3, r5]\r
1794 and r6, lr, r12, lsr #15\r
1795 ldrh r6, [r3, r6]\r
1796 orr r4, r4, r5, lsl #16\r
1797\r
1798 and r5, lr, r12, lsr #23\r
1799 ldrh r5, [r3, r5]\r
1800 and r8, lr, r7, lsl #1\r
1801 ldrh r8, [r3, r8]\r
1802 orr r5, r6, r5, lsl #16\r
1803\r
1804 and r6, lr, r7, lsr #7\r
1805 ldrh r6, [r3, r6]\r
1806 and r12,lr, r7, lsr #15\r
1807 ldrh r12,[r3, r12]\r
499a0be3 1808 and r7, lr, r7, lsr #23\r
1809 ldrh r7, [r3, r7]\r
cc68a136 1810 orr r8, r8, r6, lsl #16\r
1811\r
cc68a136 1812 subs r2, r2, #1\r
499a0be3 1813 orr r12,r12, r7, lsl #16\r
cc68a136 1814\r
1815 stmia r0!, {r4,r5,r8,r12}\r
1816 bne .fl_loopRGB555\r
1817\r
b2305d08 1818 ldmfd sp!, {r4-r9,lr}\r
3d48f143 1819 bx lr\r
1820\r
1821\r
1822.fl_32scale_RGB555:\r
b2305d08 1823 stmfd sp!, {r10}\r
3d48f143 1824 mov r9, #0x3900 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1825 orr r9, r9, #0x00e7\r
1826\r
499a0be3 1827.if UNALIGNED_DRAWLINEDEST\r
1828 tst r0, #2\r
1829 bne .fl_32scale_RGB555u\r
1830.endif\r
1831\r
3d48f143 1832.fl_loop32scale_RGB555:\r
1833 ldr r12, [r1], #4\r
1834 ldr r7, [r1], #4\r
cc68a136 1835\r
3d48f143 1836 and r4, lr, r12,lsl #1\r
1837 ldrh r4, [r3, r4]\r
1838 and r5, lr, r12,lsr #7\r
1839 ldrh r5, [r3, r5]\r
1840 and r4, r4, r9, lsl #2\r
1841 orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0\r
1842 and r5, r5, r9, lsl #2\r
1843 sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1\r
1844 add r4, r4, r6, lsl #16 @ pix_d 0, 1\r
1845 and r6, lr, r12,lsr #15\r
1846 ldrh r6, [r3, r6]\r
1847 and r12,lr, r12,lsr #23\r
1848 ldrh r12,[r3, r12]\r
1849 and r6, r6, r9, lsl #2\r
1850 add r5, r5, r6\r
1851 mov r5, r5, lsr #1\r
1852 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1853 orr r5, r5, r6, lsl #16\r
1854\r
1855 and r6, lr, r7, lsl #1\r
1856 ldrh r6, [r3, r6]\r
1857 and r12,r12,r9, lsl #2\r
1858 add r5, r5, r12,lsl #14 @ pix_d 2, 3\r
1859 and r6, r6, r9, lsl #2\r
1860 orr r6, r12,r6, lsl #16 @ pix_d 4, 5\r
1861\r
1862 and r12,lr, r7, lsr #7\r
1863 ldrh r12,[r3, r12]\r
1864 and r10,lr, r7, lsr #15\r
1865 ldrh r10,[r3, r10]\r
1866 and r12,r12,r9, lsl #2\r
1867 sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1\r
1868 add r8, r8, r6, lsr #18\r
1869 and r7, lr, r7, lsr #23\r
1870 ldrh r7, [r3, r7]\r
1871 and r10,r10,r9, lsl #2\r
1872 orr r8, r8, r10,lsl #15\r
1873 add r8, r8, r12,lsl #15 @ pix_d 6, 7\r
1874 sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2\r
1875 and r7, r7, r9, lsl #2\r
1876 add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3\r
1877 orr r10,r10,r7, lsl #16 @ pix_d 8, 9\r
1878\r
1879 subs r2, r2, #1\r
1880\r
1881 stmia r0!, {r4,r5,r6,r8,r10}\r
1882 bne .fl_loop32scale_RGB555\r
1883\r
b2305d08 1884 ldmfd sp!, {r10}\r
1885 ldmfd sp!, {r4-r9,lr}\r
3d48f143 1886 bx lr\r
1887\r
499a0be3 1888.if UNALIGNED_DRAWLINEDEST\r
1889 @ unaligned versions of loops\r
9839d126 1890 @ warning: starts drawing 2bytes before dst\r
499a0be3 1891\r
1892.fl_RGB555u:\r
9839d126 1893 sub r0, r0, #2 @ initial adjustment\r
1894 mov r8, #0\r
499a0be3 1895\r
1896.fl_loopRGB555u:\r
1897 ldr r12, [r1], #4\r
1898 ldr r7, [r1], #4\r
1899\r
9839d126 1900 and r6, lr, r12,lsl #1\r
1901 ldrh r6, [r3, r6]\r
1902 and r5, lr, r12,lsr #7\r
499a0be3 1903 ldrh r5, [r3, r5]\r
9839d126 1904 orr r4, r8, r6, lsl #16\r
499a0be3 1905\r
9839d126 1906 and r6, lr, r12,lsr #15\r
499a0be3 1907 ldrh r6, [r3, r6]\r
9839d126 1908 and r8, lr, r12,lsr #23\r
499a0be3 1909 ldrh r8, [r3, r8]\r
9839d126 1910 orr r5, r5, r6, lsl #16\r
499a0be3 1911\r
9839d126 1912 and r6, lr, r7, lsl #1\r
499a0be3 1913 ldrh r6, [r3, r6]\r
9839d126 1914 and r12,lr, r7, lsr #7\r
499a0be3 1915 ldrh r12,[r3, r12]\r
9839d126 1916 orr r6, r8, r6, lsl #16\r
1917\r
1918 and r8, lr, r7, lsr #15\r
499a0be3 1919 ldrh r8, [r3, r8]\r
9839d126 1920 and r7, lr, r7, lsr #23\r
499a0be3 1921\r
1922 subs r2, r2, #1\r
9839d126 1923 orr r12,r12,r8, lsl #16\r
1924 ldrh r8, [r3, r7]\r
1925\r
1926 stmia r0!, {r4,r5,r6,r12}\r
499a0be3 1927 bne .fl_loopRGB555u\r
1928\r
1929 strh r8, [r0], #2\r
1930\r
b2305d08 1931 ldmfd sp!, {r4-r9,lr}\r
499a0be3 1932 bx lr\r
1933\r
1934\r
1935.fl_32scale_RGB555u:\r
9839d126 1936 sub r0, r0, #2 @ initial adjustment\r
1937 mov r4, #0\r
499a0be3 1938\r
1939 @ r9 f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1940.fl_loop32scale_RGB555u:\r
1941 ldr r12, [r1], #4\r
1942 ldr r7, [r1], #4\r
1943\r
9839d126 1944 and r6, lr, r12,lsl #1\r
1945 ldrh r6, [r3, r6]\r
499a0be3 1946 and r5, lr, r12,lsr #7\r
1947 ldrh r5, [r3, r5]\r
9839d126 1948 and r6, r6, r9, lsl #2\r
1949 orr r4, r4, r6, lsl #16 @ r4 = pix_d -1, 0\r
499a0be3 1950\r
1951 and r5, r5, r9, lsl #2\r
9839d126 1952 sub r8, r5, r5, lsr #2 @ r8 = 3/4 pix_s 1\r
1953 add r6, r8, r6, lsr #2 @ r6 = (1/4 pix_s 0) + (3/4 pix_s 1)\r
1954 orr r5, r6, r5, lsl #15\r
499a0be3 1955\r
1956 and r6, lr, r12,lsr #15\r
1957 ldrh r6, [r3, r6]\r
1958 and r12,lr, r12,lsr #23\r
1959 ldrh r12,[r3, r12]\r
1960 and r6, r6, r9, lsl #2\r
9839d126 1961 add r5, r5, r6, lsl #15 @ r5 = pix_d 1, 2\r
499a0be3 1962\r
9839d126 1963 and r8, lr, r7, lsl #1\r
1964 ldrh r8, [r3, r8]\r
1965 and r10,lr, r7, lsr #7\r
1966 ldrh r10,[r3, r10]\r
499a0be3 1967 and r12,r12,r9, lsl #2\r
9839d126 1968 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1969 add r6, r6, r12,lsr #2\r
1970 orr r6, r6, r12,lsl #16 @ r6 = pix_d 3, 4\r
499a0be3 1971\r
9839d126 1972 and r8, r8, r9, lsl #2\r
1973 and r10,r10,r9, lsl #2\r
1974 sub r12,r10,r10,lsr #2 @ r12 = 3/4 pix_s 5\r
1975 orr r8, r8, r8, lsl #14\r
1976 add r8, r8, r12,lsl #16 @ r8 = pix_d 5, 6\r
1977 and r12,lr, r7, lsr #15\r
499a0be3 1978 ldrh r12,[r3, r12]\r
499a0be3 1979 and r7, lr, r7, lsr #23\r
1980 ldrh r7, [r3, r7]\r
9839d126 1981 and r12,r12,r9, lsl #2\r
1982 add r10,r10,r12\r
1983 mov r10,r10, lsr #1\r
1984 sub r12,r12,r12,lsr #2 @ r12 = 3/4 pix_s 6\r
1985 orr r10,r10,r12,lsl #16\r
499a0be3 1986 and r7, r7, r9, lsl #2\r
9839d126 1987 add r10,r10,r7, lsl #14 @ r10 = pix_d 7, 8\r
499a0be3 1988\r
1989 subs r2, r2, #1\r
1990\r
1991 stmia r0!, {r4,r5,r6,r8,r10}\r
9839d126 1992 mov r4, r7\r
499a0be3 1993 bne .fl_loop32scale_RGB555u\r
1994\r
9839d126 1995 strh r4, [r0], #2\r
1996\r
b2305d08 1997 ldmfd sp!, {r10}\r
1998 ldmfd sp!, {r4-r9,lr}\r
499a0be3 1999 bx lr\r
2000\r
2001.endif @ UNALIGNED_DRAWLINEDEST\r
2002\r
cc68a136 2003\r
2004@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
2005\r
2006@ utility\r
2007.global blockcpy @ void *dst, void *src, size_t n\r
2008\r
2009blockcpy:\r
2010 stmfd sp!, {r4,r5}\r
2011 mov r2, r2, lsr #4\r
2012blockcpy_loop:\r
2013 ldmia r1!, {r3-r5,r12}\r
2014 subs r2, r2, #1\r
2015 stmia r0!, {r3-r5,r12}\r
2016 bne blockcpy_loop\r
2017 ldmfd sp!, {r4,r5}\r
2018 bx lr\r
2019\r
2020\r
2021.global blockcpy_or @ void *dst, void *src, size_t n, int pat\r
2022\r
2023blockcpy_or:\r
2024 stmfd sp!, {r4-r6}\r
2025 orr r3, r3, r3, lsl #8\r
2026 orr r3, r3, r3, lsl #16\r
2027 mov r2, r2, lsr #4\r
2028blockcpy_loop_or:\r
2029 ldmia r1!, {r4-r6,r12}\r
2030 subs r2, r2, #1\r
2031 orr r4, r4, r3\r
2032 orr r5, r5, r3\r
2033 orr r6, r6, r3\r
2034 orr r12,r12,r3\r
2035 stmia r0!, {r4-r6,r12}\r
2036 bne blockcpy_loop_or\r
2037 ldmfd sp!, {r4-r6}\r
2038 bx lr\r
2039\r
fast MegaDrive/MegaCD/32X emulator