notaz.gp2x.de / picodrive.git / blame_incremental
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (non-incremental) | history | HEAD
pandora: fix readme and pxml version
[picodrive.git] / pico / draw_arm.S
... / ...
CommitLineData
1/*\r
2 * assembly optimized versions of most funtions from draw.c\r
3 * (C) notaz, 2006-2010,2017\r
4 * (C) irixxxx, 2020-2024\r
5 *\r
6 * This work is licensed under the terms of MAME license.\r
7 * See COPYING file in the top-level directory.\r
8 *\r
9 * this is highly specialized, be careful if changing related C code!\r
10 *\r
11 * NB only does RGB565 output, BGR isn't supported\r
12 */\r
13\r
14#include "pico_int_offs.h"\r
15\r
16.extern DrawStripInterlace\r
17\r
18.equ PDRAW_WND_DIFF_PRIO, (1<<1)\r
19.equ PDRAW_PLANE_HI_PRIO, (1<<6)\r
20.equ PDRAW_SHHI_DONE, (1<<7)\r
21.equ PDRAW_BORDER_32, (1<<9)\r
22.equ PDRAW_32X_SCALE, (1<<12)\r
23.equ PDRAW_BGC_DMA, (1<<14)\r
24.equ PDRAW_SOFTSCALE, (1<<15)\r
25\r
26@ helpers\r
27.macro add_c24 d s c\r
28 add \d, \s, #(\c & 0x00ff00)\r
29.if \c & 0x0000ff\r
30 add \d, \d, #(\c & 0x0000ff)\r
31.endif\r
32.if \c & 0xff0000\r
33 add \d, \d, #(\c & 0xff0000)\r
34.endif\r
35.endm\r
36\r
37.macro TilePixel pat lsrr offs\r
38.if !\lsrr\r
39 ands r4, \pat, r2\r
40.else\r
41 ands r4, \pat, r2, lsr #\lsrr\r
42.endif\r
43 orrne r4, r3, r4\r
44 strneb r4, [r1,#\offs]\r
45.endm\r
46\r
47@ TileNorm (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
48.macro TileNorm pat\r
49 TilePixel \pat, 12, 0 @ #0x0000f000\r
50 TilePixel \pat, 8, 1 @ #0x00000f00\r
51 TilePixel \pat, 4, 2 @ #0x000000f0\r
52 TilePixel \pat, 0, 3 @ #0x0000000f\r
53 TilePixel \pat, 28, 4 @ #0xf0000000\r
54 TilePixel \pat, 24, 5 @ #0x0f000000\r
55 TilePixel \pat, 20, 6 @ #0x00f00000\r
56 TilePixel \pat, 16, 7 @ #0x000f0000\r
57.endm\r
58\r
59@ TileFlip (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
60.macro TileFlip pat\r
61 TilePixel \pat, 16, 0 @ #0x000f0000\r
62 TilePixel \pat, 20, 1 @ #0x00f00000\r
63 TilePixel \pat, 24, 2 @ #0x0f000000\r
64 TilePixel \pat, 28, 3 @ #0xf0000000\r
65 TilePixel \pat, 0, 4 @ #0x0000000f\r
66 TilePixel \pat, 4, 5 @ #0x000000f0\r
67 TilePixel \pat, 8, 6 @ #0x00000f00\r
68 TilePixel \pat, 12, 7 @ #0x0000f000\r
69.endm\r
70\r
71@ shadow/hilight mode\r
72\r
73@ \r
74.macro TilePixelNonSH pat lsrr offs\r
75.if !\lsrr\r
76 ands r4, \pat, r2\r
77.else\r
78 ands r4, \pat, r2, lsr #\lsrr\r
79.endif\r
80 beq 0f\r
81 cmp r4, #0xe\r
82 orr r4, r3, r4\r
83 biceq r4, r4, #0x80\r
84 strb r4, [r1,#\offs]\r
850:\r
86.endm\r
87\r
88@ TileNormNonSH (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
89.macro TileNormNonSH pat\r
90 TilePixelNonSH \pat, 12, 0 @ #0x0000f000\r
91 TilePixelNonSH \pat, 8, 1 @ #0x00000f00\r
92 TilePixelNonSH \pat, 4, 2 @ #0x000000f0\r
93 TilePixelNonSH \pat, 0, 3 @ #0x0000000f\r
94 TilePixelNonSH \pat, 28, 4 @ #0xf0000000\r
95 TilePixelNonSH \pat, 24, 5 @ #0x0f000000\r
96 TilePixelNonSH \pat, 20, 6 @ #0x00f00000\r
97 TilePixelNonSH \pat, 16, 7 @ #0x000f0000\r
98.endm\r
99\r
100@ TileFlipNonSH (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
101.macro TileFlipNonSH pat\r
102 TilePixelNonSH \pat, 16, 0 @ #0x000f0000\r
103 TilePixelNonSH \pat, 20, 1 @ #0x00f00000\r
104 TilePixelNonSH \pat, 24, 2 @ #0x0f000000\r
105 TilePixelNonSH \pat, 28, 3 @ #0xf0000000\r
106 TilePixelNonSH \pat, 0, 4 @ #0x0000000f\r
107 TilePixelNonSH \pat, 4, 5 @ #0x000000f0\r
108 TilePixelNonSH \pat, 8, 6 @ #0x00000f00\r
109 TilePixelNonSH \pat, 12, 7 @ #0x0000f000\r
110.endm\r
111\r
112@ this one is for hi priority layer\r
113.macro TilePixelShHP lsrr offs\r
114.if !\lsrr\r
115 ands r4, r12, r2\r
116.else\r
117 ands r4, r12, r2, lsr #\lsrr\r
118.endif\r
119 ldreqb r4, [r1,#\offs]\r
120 orrne r4, r3, r4\r
121 andeq r4, r4, #0x7f\r
122 strb r4, [r1,#\offs]\r
123.endm\r
124\r
125@ TileNormShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: register with helper pattern 0xf, touches r3 high bits\r
126.macro TileNormShHP\r
127 TilePixelShHP 12, 0 @ #0x0000f000\r
128 TilePixelShHP 8, 1 @ #0x00000f00\r
129 TilePixelShHP 4, 2 @ #0x000000f0\r
130 TilePixelShHP 0, 3 @ #0x0000000f\r
131 TilePixelShHP 28, 4 @ #0xf0000000\r
132 TilePixelShHP 24, 5 @ #0x0f000000\r
133 TilePixelShHP 20, 6 @ #0x00f00000\r
134 TilePixelShHP 16, 7 @ #0x000f0000\r
135.endm\r
136\r
137@ TileFlipShHP (r1=pdest, r2=pixels8, r3=pal) r4: scratch, pat: register with helper pattern 0xf\r
138.macro TileFlipShHP\r
139 TilePixelShHP 16, 0 @ #0x000f0000\r
140 TilePixelShHP 20, 1 @ #0x00f00000\r
141 TilePixelShHP 24, 2 @ #0x0f000000\r
142 TilePixelShHP 28, 3 @ #0xf0000000\r
143 TilePixelShHP 0, 4 @ #0x0000000f\r
144 TilePixelShHP 4, 5 @ #0x000000f0\r
145 TilePixelShHP 8, 6 @ #0x00000f00\r
146 TilePixelShHP 12, 7 @ #0x0000f000\r
147.endm\r
148\r
149\r
150@ TileSingleSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx; r12: helper pattern 0xf\r
151.macro TileSingleSh\r
152 tst r0, #1 @ not aligned?\r
153 mov r7, #0x008000\r
154 orr r7, r7, #0x80\r
155 ldrneb r4, [r1], #1\r
156 ldreqh r4, [r1], #2 @ 1ci\r
157 ldrh r12, [r1], #2\r
158 orr r4, r4, r7\r
159 strneb r4, [r1, #-3]\r
160 streqh r4, [r1, #-4]\r
161 ldrh r4, [r1], #2\r
162 orr r12, r12, r7\r
163 strh r12, [r1, #-4]\r
164 ldrh r12, [r1], #2\r
165 orr r4, r4, r7\r
166 strh r4, [r1, #-4]\r
167 ldrneb r4, [r1]\r
168 orr r12, r12, r7\r
169 strh r12, [r1, #-2]\r
170 orrne r4, r4, r7\r
171 strneb r4, [r1], #1\r
172 mov r12, #0xf\r
173.endm\r
174\r
175@ TileSingleHi (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
176.macro TileSingleHi\r
177 tst r1, #1 @ not aligned?\r
178 mov r7, #0x004000\r
179 orr r7, r7, #0x40\r
180 ldrneb r4, [r1], #1\r
181 ldreqh r4, [r1], #2 @ 1ci\r
182 ldrh r12, [r1], #2\r
183 orr r4, r4, r7\r
184 strneb r4, [r1, #-3]\r
185 streqh r4, [r1, #-4]\r
186 ldrh r4, [r1], #2\r
187 orr r12, r12, r7\r
188 strh r12, [r1, #-4]\r
189 ldrh r12, [r1], #2\r
190 orr r4, r4, r7\r
191 strh r4, [r1, #-4]\r
192 ldrneb r4, [r1]\r
193 orr r12, r12, r7\r
194 strh r12, [r1, #-2]\r
195 orrne r4, r4, r7\r
196 strneb r4, [r1], #1\r
197 mov r12, #0xf\r
198.endm\r
199\r
200.macro TileDoShGenPixel shift ofs\r
201.if \shift\r
202 ands r4, r12, r2, lsr #\shift\r
203.else\r
204 ands r4, r12, r2\r
205.endif\r
206 beq 0f\r
207 cmp r4, #0xe\r
208 ldrgeb r7, [r1,#\ofs]\r
209 orrlt r7, r3, r4 @ normal\r
210\r
211 subge r4, r4, #1\r
212 orrge r7, r7, r4, lsl #6\r
213 strb r7, [r1,#\ofs]\r
2140:\r
215.endm\r
216\r
217@ TileFlipSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
218.macro TileFlipSh\r
219 TileDoShGenPixel 16, 0 @ #0x000f0000\r
220 TileDoShGenPixel 20, 1 @ #0x00f00000\r
221 TileDoShGenPixel 24, 2 @ #0x0f000000\r
222 TileDoShGenPixel 28, 3 @ #0xf0000000\r
223 TileDoShGenPixel 0, 4 @ #0x0000000f\r
224 TileDoShGenPixel 4, 5 @ #0x000000f0\r
225 TileDoShGenPixel 8, 6 @ #0x00000f00\r
226 TileDoShGenPixel 12, 7 @ #0x0000f000\r
227.endm\r
228\r
229@ TileNormSh (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=sx, r12: register with helper pattern 0xf\r
230.macro TileNormSh\r
231 TileDoShGenPixel 12, 0 @ #0x0000f000\r
232 TileDoShGenPixel 8, 1 @ #0x00000f00\r
233 TileDoShGenPixel 4, 2 @ #0x000000f0\r
234 TileDoShGenPixel 0, 3 @ #0x0000000f\r
235 TileDoShGenPixel 28, 4 @ #0xf0000000\r
236 TileDoShGenPixel 24, 5 @ #0x0f000000\r
237 TileDoShGenPixel 20, 6 @ #0x00f00000\r
238 TileDoShGenPixel 16, 7 @ #0x000f0000\r
239.endm\r
240\r
241.macro TileDoShGenPixel_markop shift ofs\r
242.if \shift\r
243 ands r4, r12, r2, lsr #\shift\r
244.else\r
245 ands r4, r12, r2\r
246.endif\r
247 beq 0f\r
248 cmp r4, #0xe\r
249 ldrgeb r4, [r1,#\ofs]\r
250 orrlt r4, r3, r4\r
251 orrge r4, r4, #0x40\r
252 strb r4, [r1,#\ofs]\r
2530:\r
254.endm\r
255\r
256.macro TileFlipSh_markop\r
257 TileDoShGenPixel_markop 16, 0 @ #0x000f0000\r
258 TileDoShGenPixel_markop 20, 1 @ #0x00f00000\r
259 TileDoShGenPixel_markop 24, 2 @ #0x0f000000\r
260 TileDoShGenPixel_markop 28, 3 @ #0xf0000000\r
261 TileDoShGenPixel_markop 0, 4 @ #0x0000000f\r
262 TileDoShGenPixel_markop 4, 5 @ #0x000000f0\r
263 TileDoShGenPixel_markop 8, 6 @ #0x00000f00\r
264 TileDoShGenPixel_markop 12, 7 @ #0x0000f000\r
265.endm\r
266\r
267.macro TileNormSh_markop\r
268 TileDoShGenPixel_markop 12, 0 @ #0x0000f000\r
269 TileDoShGenPixel_markop 8, 1 @ #0x00000f00\r
270 TileDoShGenPixel_markop 4, 2 @ #0x000000f0\r
271 TileDoShGenPixel_markop 0, 3 @ #0x0000000f\r
272 TileDoShGenPixel_markop 28, 4 @ #0xf0000000\r
273 TileDoShGenPixel_markop 24, 5 @ #0x0f000000\r
274 TileDoShGenPixel_markop 20, 6 @ #0x00f00000\r
275 TileDoShGenPixel_markop 16, 7 @ #0x000f0000\r
276.endm\r
277\r
278.macro TileDoShGenPixel_onlyop_lp shift ofs\r
279.if \shift\r
280 ands r7, r12, r2, lsr #\shift\r
281.else\r
282 ands r7, r12, r2\r
283.endif\r
284 ldrneb r4, [r1,#\ofs]\r
285 cmp r7, #0xe\r
286 blt 0f\r
287\r
288 tst r4, #0x40\r
289 bicne r4, r4, #0x40\r
290 subne r7, r7, #1\r
291 orrne r4, r4, r7, lsl #6\r
292 strneb r4, [r1,#\ofs]\r
2930:\r
294.endm\r
295\r
296.macro TileFlipSh_onlyop_lp\r
297 TileDoShGenPixel_onlyop_lp 16, 0 @ #0x000f0000\r
298 TileDoShGenPixel_onlyop_lp 20, 1 @ #0x00f00000\r
299 TileDoShGenPixel_onlyop_lp 24, 2 @ #0x0f000000\r
300 TileDoShGenPixel_onlyop_lp 28, 3 @ #0xf0000000\r
301 TileDoShGenPixel_onlyop_lp 0, 4 @ #0x0000000f\r
302 TileDoShGenPixel_onlyop_lp 4, 5 @ #0x000000f0\r
303 TileDoShGenPixel_onlyop_lp 8, 6 @ #0x00000f00\r
304 TileDoShGenPixel_onlyop_lp 12, 7 @ #0x0000f000\r
305.endm\r
306\r
307.macro TileNormSh_onlyop_lp\r
308 TileDoShGenPixel_onlyop_lp 12, 0 @ #0x0000f000\r
309 TileDoShGenPixel_onlyop_lp 8, 1 @ #0x00000f00\r
310 TileDoShGenPixel_onlyop_lp 4, 2 @ #0x000000f0\r
311 TileDoShGenPixel_onlyop_lp 0, 3 @ #0x0000000f\r
312 TileDoShGenPixel_onlyop_lp 28, 4 @ #0xf0000000\r
313 TileDoShGenPixel_onlyop_lp 24, 5 @ #0x0f000000\r
314 TileDoShGenPixel_onlyop_lp 20, 6 @ #0x00f00000\r
315 TileDoShGenPixel_onlyop_lp 16, 7 @ #0x000f0000\r
316.endm\r
317\r
318\r
319@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
320\r
321@ struct TileStrip\r
322@ {\r
323@ int nametab; // 0x00\r
324@ int line; // 0x04\r
325@ int hscroll; // 0x08\r
326@ int xmask; // 0x0C\r
327@ int *hc; // 0x10 (pointer to cache buffer)\r
328@ int cells; // 0x14\r
329@ };\r
330\r
331@ void DrawLayer(int lflags, int *hcache, int cellskip, int maxcells,\r
332@ struct PicoEState *est)\r
333\r
334.global DrawLayer\r
335\r
336DrawLayer:\r
337 ldr r12, [sp] @ est\r
338 stmfd sp!, {r4-r11,lr}\r
339\r
340 ldr r11, [r12, #OFS_EST_Pico]\r
341 mov r8, #1\r
342\r
343 ldrb r7, [r11, #OFS_Pico_video_reg+16] @ ??vv??hh\r
344\r
345 mov r6, r1 @ hcache\r
346 orr r9, r3, r0, lsl #29 @ r9=force[31]|sh[30]|plane[29]\r
347 orr r9, r9, r2, lsl #8 @ |cellskip[15:8]|maxcells[7:0] (tmp)\r
348\r
349 mov r1, r7, lsl #4\r
350 orr r1, r1, #0x00ff\r
351\r
352 and r10, r7, #3\r
353 cmp r10, #1\r
354 biclt r1, r1, #0xfc00\r
355 biceq r1, r1, #0xfe00\r
356 cmp r10, #2\r
357 moveq r1, #0x0007\r
358 movgt r1, #0x00ff @ r1=ymask=(height<<8)|0xff; ...; // Y Mask in pixels\r
359\r
360 cmp r10, #2\r
361 addlt r10, r10, #5\r
362 moveq r10, #5\r
363 movgt r10, #7 @ r10=shift[width] (5,6,5,7)\r
364\r
365 ldr r2, [r12, #OFS_EST_DrawScanline]\r
366 ldr lr, [r12, #OFS_EST_PicoMem_vram]\r
367\r
368 @ Find name table:\r
369 ands r0, r0, #1\r
370 ldreqb r12, [r11, #OFS_Pico_video_reg+2]\r
371 ldrneb r12, [r11, #OFS_Pico_video_reg+4]\r
372\r
373 @ calculate xmask:\r
374 mov r5, r8, lsl r10\r
375 sub r5, r5, #1 @ r5=xmask\r
376\r
377 moveq r12, r12, lsl #10\r
378 movne r12, r12, lsl #13\r
379 and r12, r12, #(7<<13) @ r12=(ts->nametab<<1) (halfword compliant)\r
380\r
381 ldrh r8, [r11, #OFS_Pico_video_reg+12]\r
382 ldrb r7, [r11, #OFS_Pico_video_reg+11]\r
383\r
384 mov r4, r8, lsr #8 @ pvid->reg[13]\r
385 mov r4, r4, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r
386\r
387 ands r3, r7, #0x03\r
388 beq 0f\r
389 cmp r3, #2\r
390 mov r3, r2, lsl #2 @ htab+=DrawScanline<<1; // Offset by line\r
391 biceq r3, r3, #0x1f @ htab&=~0xf; // Offset by tile\r
392 andlt r3, r3, #0x1f\r
393 add r4, r4, r3\r
3940: add r4, r4, r0, lsl #1 @ htab+=plane\r
395 bic r4, r4, #0x00ff0000 @ just in case\r
396 ldrh r3, [lr, r4] @ r3=hscroll\r
397\r
398 tst r7, #4\r
399 bne .DrawStrip_vsscroll\r
400\r
401 @ Get vertical scroll value:\r
402 add_c24 r7, lr, (OFS_PMEM_vsram-OFS_PMEM_vram)\r
403 ldr r7, [r7]\r
404\r
405 tst r8, #2\r
406 tstne r8, #4\r
407 bne .DrawStrip_interlace\r
408\r
409 tst r0, r0\r
410 moveq r7, r7, lsl #16\r
411 mov r7, r7, lsr #16\r
412\r
413 @ Find the line in the name table\r
414 add r2, r2, r7\r
415 and r2, r2, r1\r
416 mov r4, r2, lsr #3\r
417 add r10, r10, #1 @ shift[width]++\r
418 add r12, r12, r4, lsl r10 @ nametab+=(ts.line>>3)<<shift[width];\r
419\r
420 @ ldmia r0, {r1,r2,r3,r5,r6,r9} @ r2=line, r3=ts->hscroll, r5=ts->xmask, r6=ts->hc, r9=ts->cells\r
421\r
422 and r10,r2, #7\r
423 mov r10,r10, lsl #1 @ r10=ty=(ts->line&7)<<1;\r
424 orr r10,r10, r9, lsl #24\r
425\r
426 rsb r8, r3, #0\r
427 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
428\r
429 sub r1, r3, #1\r
430 and r1, r1, #7\r
431 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
432\r
433 movs r3, r9, lsl #1 @ (force[31]|sh[30]) << 1\r
434 mov r3, #0\r
435 orrmi r10,r10, #1<<23 @ r10=cells[31:24]|sh[23]|hi_not_empty[22]\r
436@ orrcc r10,r10, #1<<20 @ |had_output[21]|!force[20]|hscroll[18:16]|ty[15:0]\r
437 movmi r3, #0x80 @ default to shadowed pal on sh mode\r
438\r
439 and r4, r7, #7\r
440 orr r10,r10, r4, lsl #16 @ we will process cells+1 if there is scroll\r
441\r
442 and r9, r9, #0xff00\r
443 add r8, r8, r9, lsr #8 @ tilex+=cellskip\r
444 add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;\r
445 sub r10,r10,r9, lsl #16 @ cells-=cellskip\r
446\r
447 @ cache some stuff to avoid mem access\r
448 ldr r11,[sp, #9*4] @ est\r
449 mov r0, #0xf\r
450 ldr r11,[r11, #OFS_EST_HighCol]\r
451\r
452 mvn r9, #0 @ r9=prevcode=-1\r
453 add r1, r11, r7 @ r1=pdest\r
454\r
455 @ r10=cells[31:24]|sh[23]|hi_not_empty[22]|had_output[21]|!force[20]|hscroll[18:16]|ty[15:0]\r
456 @ r1=pd+dx r2=pack r3=pal r5=xmask r6=hc r8=tilex r9=prevcode r11=HighCol r12=nametab lr=vram\r
457 @ r4 & r7 are scratch in this loop\r
458\r
459 ands r4, r10, #7<<16 @ hscroll?\r
460 beq .dsloop_subr1\r
461 subs r10,r10, #0x01000000\r
462 bmi .dsloop_exit\r
463\r
464 and r7, r5, r8 @ do first cut tile\r
465 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
466 ldrh r9, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
467\r
468 add r8, r8, #1\r
469\r
470 tst r9, #0x1000 @ if (code&0x1000)\r
471 mov r2, r9, lsl #21\r
472 add r2, r2, r10, lsl #17\r
473 eorne r2, r2, #0xe<<17 @ if (code&0x1000) addr^=0xe;\r
474\r
475 ldr r2, [lr, r2, lsr #16] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
476\r
477 mvn r7, #0\r
478 mov r4, r4, lsr #16-2 @ (dx&7)*4\r
479 tst r9, #0x0800\r
480 moveq r7, r7, lsl r4 @ mask = ~0 [shift] (dx&7)*4\r
481 movne r7, r7, lsr r4\r
482 bic r2, r2, r7, ror #16 @ pack&~mask\r
483\r
484 orr r9, r9, #0x80000000 @ invalidate oldcode since pack is masked\r
485 b .DrawStrip_samecode\r
486\r
487.dsloop_subr1:\r
488 sub r1, r1, #8\r
489.dsloop: @ 40-41 times\r
490 subs r10,r10, #0x01000000\r
491 bmi .dsloop_exit\r
492\r
493 and r7, r5, r8\r
494 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
495 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
496\r
497 add r1, r1, #8\r
498 add r8, r8, #1\r
499\r
500 cmp r7, r9\r
501 beq .DrawStrip_samecode @ we know stuff about this tile already\r
502\r
503 mov r9, r7 @ remember code\r
504\r
505 tst r9, #0x1000 @ if (code&0x1000)\r
506 mov r2, r9, lsl #21\r
507 add r2, r2, r10, lsl #17\r
508 eorne r2, r2, #0x0e<<17 @ if (code&0x1000) addr^=0xe;\r
509\r
510 ldr r2, [lr, r2, lsr #16] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
511.DrawStrip_samecode:\r
512 tst r9, #0x8000\r
513@ tstne r10, #1<<20 @ !force[20]\r
514 bne .DrawStrip_hiprio\r
515\r
516 orr r10, r10, #1<<21 @ seen non hi-prio tile\r
517 tst r2, r2\r
518 beq .dsloop @ tileline blank\r
519\r
520 bic r7, r3, #0x7f\r
521 and r3, r9, #0x6000\r
522 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
523\r
524 cmp r2, r2, ror #4\r
525 beq .DrawStrip_SingleColor @ tileline singlecolor \r
526\r
527 tst r9, #0x0800\r
528 bne .DrawStrip_TileFlip\r
529\r
530 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
531.DrawStrip_TileNorm:\r
532 TileNorm r0\r
533 b .dsloop\r
534\r
535.DrawStrip_TileFlip:\r
536 TileFlip r0\r
537 b .dsloop\r
538\r
539.DrawStrip_SingleColor:\r
540 and r4, r2, #0xf\r
541 orr r4, r3, r4\r
542 orr r4, r4, r4, lsl #8\r
543 tst r1, #1 @ not aligned?\r
544 strneb r4, [r1], #1\r
545 streqh r4, [r1], #2\r
546 strh r4, [r1], #2\r
547 strh r4, [r1], #2\r
548 strh r4, [r1], #2\r
549 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
550 b .dsloop_subr1\r
551\r
552.DrawStrip_hiprio:\r
553 tst r10, #(1<<23) @ sh[23]\r
554 tsteq r2, r2 @ if (!sh[23] && code==blank) continue\r
555 beq .dsloop\r
556\r
557@ orr r10, r10, #1<<22 @ hi_not_empty[22]\r
558 sub r7, r1, r11\r
559 orr r7, r9, r7, lsl #16\r
560 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
561 tst r9, #0x1000\r
562 eorne r7, r7, #0xe<<25 @ if(code&0x1000) cval^=0xe<<25;\r
563 str r7, [r6], #4 @ cache hi priority tile code\r
564 str r2, [r6], #4 @ cache hi priority tile data\r
565 b .dsloop\r
566\r
567.dsloop_exit:\r
568 ands r4,r10, #7<<16 @ hscroll?\r
569 beq .DrawStrip_noscroll\r
570\r
571 and r7, r5, r8 @ do one more cut tile\r
572 add r7, lr, r7, lsl #1 @ Pico.vram+((tilex&ts->xmask) as halfwords)\r
573 ldrh r9, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
574\r
575 add r1, r1, #8\r
576\r
577 tst r9, #0x1000 @ if (code&0x1000)\r
578 mov r2, r9, lsl #21\r
579 add r2, r2, r10, lsl #17\r
580 eorne r2, r2, #0x0e<<17 @ if (code&0x1000) addr^=0xe;\r
581\r
582 ldr r2, [lr, r2, lsr #16] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
583\r
584 mvn r7, #0\r
585 mov r4, r4, lsr #16-2 @ (dx&7)*4\r
586 tst r9, #0x0800\r
587 moveq r7, r7, lsl r4 @ mask = ~0 [shift] (dx&7)*4\r
588 movne r7, r7, lsr r4\r
589 and r2, r2, r7, ror #16 @ pack&mask\r
590\r
591 bic r10,r10, #7<<16\r
592 b .DrawStrip_samecode @ one last time, with last tile now masked\r
593\r
594.DrawStrip_noscroll:\r
595 tst r10, #1<<21 @ seen non hi-prio tile\r
596 ldr r1, [sp, #9*4] @ est\r
597 mov r0, #0\r
598 ldreq r2, [r1, #OFS_EST_rendstatus]\r
599 str r0, [r6] @ terminate the cache list\r
600 orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
601 streq r2, [r1, #OFS_EST_rendstatus]\r
602\r
603 ldmfd sp!, {r4-r11,pc}\r
604\r
605@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
606\r
607.DrawStrip_vsscroll:\r
608 tst r8, #1 @ if h40: lflags |= 0x10000\r
609 orrne r0, r0, #0x10000\r
610\r
611 rsb r8, r3, #0\r
612 mov r8, r8, lsr #3 @ r8=tilex=(-ts->hscroll)>>3\r
613 bic r8, r8, #0x3fc00000\r
614 orr r8, r8, r5, lsl #25 @ r8=(xmask[31:25]|had_output[24]|!force[23]|tilex[21:0])\r
615\r
616 ldr r11, [sp, #9*4] @ est\r
617 orr r5, r1, r10, lsl #24\r
618 ldr r4, [r11, #OFS_EST_DrawScanline]\r
619 sub r1, r3, #1\r
620 orr r5, r5, r4, lsl #16 @ r5=(shift_width[31:24]|scanline[23:16]|ymask[15:0])\r
621 and r1, r1, #7\r
622 add r7, r1, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r
623\r
624 mov r10,r9, lsl #16\r
625 orr r10,r10, #0xff000000 @ will be adjusted on entering loop\r
626 tst r0, #1\r
627 orrne r10,r10, #0x8000\r
628 tst r3, #0x0f @ hscroll & 0x0f?\r
629 beq 0f\r
630 eor r3, r3, r7\r
631 sub r10,r10, #1<<24 @ cell-- // start from negative for hscroll\r
632 tst r3, #0x08\r
633 subne r10,r10, #1<<16 @ cells--\r
634 subne r10,r10, #1<<24 @ cell-- // even more negative\r
635\r
636 add_c24 r1, lr, (OFS_PMEM_vsram-OFS_PMEM_vram)\r
637 tst r0, #0x10000 @ h40?\r
638 ldrne r3, [r1, #0x00] @ r3=vsram[0x00..0x01]\r
639 ldreq r3, [r1, #0x40] @ r3=vsram[0x20..0x21]\r
640 str r3, [r1, #0x7c] @ vsram[0x3e..0x3f]=r3\r
6410:\r
642 tst r9, #1<<30\r
643 mov r3, #0\r
644 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
645 movne r3, #0x80 @ default to shadowed pal on sh mode\r
646@ tst r9, #1<<31\r
647@ orreq r8, r8, #1<<23\r
648\r
649 and r9, r9, #0xff00\r
650 add r8, r8, r9, lsr #8 @ tilex+=cellskip\r
651 add r7, r7, r9, lsr #5 @ dx+=cellskip<<3;\r
652 add r10,r10,r9, lsl #16 @ cell+=cellskip\r
653\r
654 @ cache some stuff to avoid mem access\r
655 ldr r11,[sp, #9*4] @ est\r
656 mov r0, #0xf\r
657 ldr r11,[r11, #OFS_EST_HighCol]\r
658\r
659 mvn r9, #0 @ r9=prevcode=-1\r
660 add r1, r11, r7 @ r1=pdest\r
661\r
662 @ r10=cells[31:24]|sh[23]|hi_not_empty[22]|cells_max[21:16]|plane[15]|ty[14:0]\r
663 @ r8=xmask[31:25]|had_output[24]|!force[23]|tilex[21:0]\r
664 @ r5=shift_width[31:24]|scanline[23:16]|ymask[15:0]\r
665 @ r3=nametabadd[31:16]|must_be_0[15:8]|pal[7:0]\r
666 @ r1=pd+dx r2=pack r6=hc r9=prevcode r11=HighCol r12=nametab lr=vram\r
667 @ r4 & r7 are scratch in this loop\r
668\r
669 @ need to calc new ty?\r
670 movs r7, r10, lsl #7 @ (cell&1)?\r
671 bmi .dsloop_vs_subr1\r
672\r
673 @ calc offset and read tileline code to r7, also calc ty\r
674 add_c24 r7, lr, (OFS_PMEM_vsram-OFS_PMEM_vram)\r
675 and r4, r10, #0x3e000000\r
676 add r7, r7, r4, asr #23 @ vsram + ((cell&0x3e)<<1)\r
677 tst r10,#0x8000 @ plane1?\r
678 addne r7, r7, #2\r
679 ldrh r7, [r7] @ r7=vscroll\r
680\r
681 bic r10,r10,#0xff @ clear old ty\r
682 and r4, r5, #0xff0000 @ scanline\r
683 add r4, r4, r7, lsl #16 @ ... += vscroll\r
684 and r4, r4, r5, lsl #16 @ ... &= ymask\r
685 and r7, r4, #0x70000\r
686 orr r10,r10,r7, lsr #15 @ new ty\r
687\r
688 mov r4, r4, lsr #19\r
689 mov r7, r5, lsr #24\r
690 mov r4, r4, lsl r7 @ nametabadd\r
691 and r3, r3, #0xff\r
692 orr r3, r3, r4, lsl #16 @ r3=(nametabadd[31:16],pal[15:0])\r
693\r
694.dsloop_vs_subr1:\r
695 sub r1, r1, #8\r
696.dsloop_vs: @ 40-41 times\r
697 add r10,r10, #0x01000000\r
698 and r4, r10, #0x003f0000\r
699 cmp r4, r10, asr #8\r
700 ble .dsloop_vs_exit\r
701\r
702 @ need to calc new ty?\r
703 movs r7, r10, lsl #7 @ (cell&1)?\r
704 bmi 0f\r
705\r
706 @ calc offset and read tileline code to r7, also calc ty\r
707 add_c24 r7, lr, (OFS_PMEM_vsram-OFS_PMEM_vram)\r
708 and r4, r10, #0x3e000000\r
709 add r7, r7, r4, asr #23 @ vsram + ((cell&0x3e)<<1)\r
710 tst r10,#0x8000 @ plane1?\r
711 addne r7, r7, #2\r
712 ldrh r7, [r7] @ r7=vscroll\r
713\r
714 bic r10,r10,#0xff @ clear old ty\r
715 and r4, r5, #0xff0000 @ scanline\r
716 add r4, r4, r7, lsl #16 @ ... += vscroll\r
717 and r4, r4, r5, lsl #16 @ ... &= ymask\r
718 and r7, r4, #0x70000\r
719 orr r10,r10,r7, lsr #15 @ new ty\r
720\r
721 mov r4, r4, lsr #19\r
722 mov r7, r5, lsr #24\r
723 mov r4, r4, lsl r7 @ nametabadd\r
724 and r3, r3, #0xff\r
725 orr r3, r3, r4, lsl #16 @ r3=(nametabadd[31:16],pal[15:0])\r
7260:\r
727 and r7, r8, r8, lsr #25\r
728 add r7, lr, r7, lsl #1 @ PicoMem.vram+((tilex&ts->xmask) as halfwords)\r
729 add r7, r7, r3, lsr #15\r
730 ldrh r7, [r7, r12] @ r7=code (int, but from unsigned, no sign extend)\r
731\r
732 add r1, r1, #8\r
733 add r8, r8, #1\r
734\r
735 orr r7, r7, r10, lsl #24 @ code | (ty << 24)\r
736 cmp r7, r9\r
737 beq .DrawStrip_vs_samecode @ we know stuff about this tile already\r
738\r
739 mov r9, r7 @ remember code\r
740\r
741 tst r9, #0x1000 @ if (code&0x1000)\r
742 mov r2, r9, lsl #21\r
743 add r2, r2, r10, lsl #17\r
744 eorne r2, r2, #0x0e<<17 @ if (code&0x1000) addr^=0xe;\r
745\r
746 ldr r2, [lr, r2, lsr #16] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
747\r
748.DrawStrip_vs_samecode:\r
749 tst r9, #0x8000\r
750@ tstne r8, #1<<23 @ !force[23]\r
751 bne .DrawStrip_vs_hiprio\r
752\r
753 orr r8, r8, #(1<<24)@ seen non hi-prio tile\r
754 tst r2, r2\r
755 beq .dsloop_vs @ tileline blank\r
756\r
757 bic r7, r3, #0x7f\r
758 and r3, r9, #0x6000\r
759 add r3, r7, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
760\r
761 cmp r2, r2, ror #4\r
762 beq .DrawStrip_vs_SingleColor @ tileline singlecolor \r
763\r
764 tst r9, #0x0800\r
765 bne .DrawStrip_vs_TileFlip\r
766\r
767 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
768.DrawStrip_vs_TileNorm:\r
769 TileNorm r0\r
770 b .dsloop_vs\r
771\r
772.DrawStrip_vs_TileFlip:\r
773 TileFlip r0\r
774 b .dsloop_vs\r
775\r
776.DrawStrip_vs_SingleColor:\r
777 and r4, r2, #0xf\r
778 orr r4, r3, r4\r
779 orr r4, r4, r4, lsl #8\r
780 tst r1, #1 @ not aligned?\r
781 strneb r4, [r1], #1\r
782 streqh r4, [r1], #2\r
783 strh r4, [r1], #2\r
784 strh r4, [r1], #2\r
785 strh r4, [r1], #2\r
786 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
787 b .dsloop_vs_subr1\r
788\r
789.DrawStrip_vs_hiprio:\r
790 tst r10, #(1<<23) @ sh[23]\r
791 tsteq r2, r2 @ if (!sh[23] && code==blank) continue\r
792 beq .dsloop_vs\r
793\r
794@ orr r10, r10, #1<<22 @ hi_not_empty[22]\r
795 sub r7, r1, r11\r
796 orr r7, r9, r7, lsl #16\r
797 orr r7, r7, r10, lsl #25 @ (ty<<25)\r
798 tst r9, #0x1000\r
799 eorne r7, r7, #7<<26 @ if(code&0x1000) cval^=7<<26;\r
800 str r7, [r6], #4 @ cache hi priority tile code\r
801 str r2, [r6], #4 @ cache hi priority tile data\r
802 b .dsloop_vs\r
803\r
804.dsloop_vs_exit:\r
805 tst r8, #(1<<24) @ seen non hi-prio tile\r
806 ldr r1, [sp, #9*4] @ est\r
807 mov r0, #0\r
808 ldreq r2, [r1, #OFS_EST_rendstatus]\r
809 str r0, [r6] @ terminate the cache list\r
810 orreq r2, r2, #PDRAW_PLANE_HI_PRIO @ had a layer with all hi-prio tiles\r
811 streq r2, [r1, #OFS_EST_rendstatus]\r
812\r
813 ldmfd sp!, {r4-r11,pc}\r
814\r
815\r
816@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
817\r
818@ interlace mode 2? Sonic 2?\r
819.DrawStrip_interlace:\r
820 tst r0, r0\r
821 movne r7, r7, lsr #16\r
822 mov r7, r7, lsl #21\r
823\r
824 @ Find the line in the name table\r
825 add r2, r7, r2, lsl #22 @ r2=(vscroll+(DrawScanline<<1))<<21 (11 bits);\r
826 orr r1, r1, #0x80000000\r
827 and r2, r2, r1, ror #10 @ &((ymask<<1)|1)<<21;\r
828 mov r2, r2, lsr #21\r
829 mov r4, r2, lsr #4\r
830 mov r12, r12, lsr #1 @ halfwords\r
831 add r0, r12, r4, lsl r10 @ nametab+=(ts.line>>4)<<shift[width];\r
832 and r9, r9, #0xff\r
833\r
834 sub sp, sp, #6*4\r
835 stmia sp, {r0,r2,r3,r5,r6,r9}\r
836\r
837 mov r0, sp\r
838 mov r1, r9, lsr #29\r
839 mov r2, r9, lsr #8\r
840 and r2, r2, #0xff\r
841 bl DrawStripInterlace @ struct TileStrip *ts, int plane_sh, int cellskip\r
842\r
843 add sp, sp, #6*4\r
844 ldmfd sp!, {r4-r11,pc}\r
845\r
846.pool\r
847\r
848@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
849\r
850@ void BackFill(int reg7, int sh, struct PicoEState *est)\r
851\r
852.global BackFill\r
853\r
854BackFill:\r
855 stmfd sp!, {r4-r9,lr}\r
856\r
857 ldr lr, [r2, #OFS_EST_HighCol]\r
858 orr r0, r0, r1, lsl #7\r
859 orr r0, r0, r0, lsl #8\r
860 orr r0, r0, r0, lsl #16\r
861\r
862 mov r1, r0\r
863 mov r2, r0\r
864 mov r3, r0\r
865 mov r4, r0\r
866 mov r5, r0\r
867 mov r6, r0\r
868 mov r7, r0\r
869\r
870 @ go go go!\r
871 add lr, lr, #8\r
872 stmia lr!, {r0-r7} @ 10*8*4\r
873 stmia lr!, {r0-r7}\r
874 stmia lr!, {r0-r7}\r
875 stmia lr!, {r0-r7}\r
876 stmia lr!, {r0-r7}\r
877 stmia lr!, {r0-r7}\r
878 stmia lr!, {r0-r7}\r
879 stmia lr!, {r0-r7}\r
880 stmia lr!, {r0-r7}\r
881 stmia lr!, {r0-r7}\r
882\r
883 ldmfd sp!, {r4-r9,pc}\r
884\r
885\r
886@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
887\r
888@ void DrawTilesFromCache(int *hc, int sh, int rlim, struct PicoEState *est)\r
889\r
890.global DrawTilesFromCache\r
891\r
892DrawTilesFromCache:\r
893 stmfd sp!, {r4-r9,r11,lr}\r
894\r
895 @ cache some stuff to avoid mem access\r
896 ldr r11,[r3, #OFS_EST_HighCol]\r
897 mov r12,#0xf\r
898 ldr lr, [r3, #OFS_EST_PicoMem_vram]\r
899 mov r9, r3 @ est\r
900\r
901 ands r8, r1, #1\r
902 orr r8, r8, r2, lsl #1\r
903 bne .dtfc_check_rendflags\r
904\r
905 @ scratch: r4, r7\r
906.dtfc_loop:\r
907 ldr r6, [r0], #4 @ read code\r
908 movs r1, r6, lsr #16 @ r1=dx;\r
909 ldmeqfd sp!, {r4-r9,r11,pc} @ dx is never zero, this must be a terminator, return\r
910 bic r4, r1, #0xfe00\r
911 add r1, r11, r4 @ r1=pdest\r
912\r
913 ldr r2, [r0], #4 @ read pixel data\r
914\r
915 and r3, r6, #0x6000\r
916 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
917\r
918 rsbs r4, r4, r8, lsr #1\r
919 bmi .dtfc_cut_tile\r
920\r
921 tst r8, #1\r
922 bne .dtfc_shadow\r
923\r
924 tst r2, r2\r
925 beq .dtfc_loop\r
926\r
927 cmp r2, r2, ror #4\r
928 beq .dtfc_SingleColor @ tileline singlecolor \r
929\r
930 tst r6, #0x0800\r
931 bne .dtfc_TileFlip\r
932\r
933 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
934.dtfc_TileNorm:\r
935 TileNorm r12\r
936 b .dtfc_loop\r
937\r
938.dtfc_TileFlip:\r
939 TileFlip r12\r
940 b .dtfc_loop\r
941\r
942.dtfc_SingleColor:\r
943 and r4, r2, #0xf\r
944 orr r4, r3, r4\r
945 orr r4, r4, r4, lsl #8\r
946 tst r1, #1 @ not aligned?\r
947 strneb r4, [r1], #1\r
948 streqh r4, [r1], #2\r
949 strh r4, [r1], #2\r
950 strh r4, [r1], #2\r
951 strh r4, [r1], #2\r
952 strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r
953 b .dtfc_loop\r
954\r
955.dtfc_shadow:\r
956 tst r2, r2\r
957 beq .dtfc_shadow_blank\r
958\r
959 cmp r2, r2, ror #4\r
960 beq .dtfc_SingleColor @ tileline singlecolor \r
961\r
962 tst r6, #0x0800\r
963 bne .dtfc_TileFlipShHP\r
964\r
965 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
966.dtfc_TileNormShHP:\r
967 TileNormShHP\r
968 b .dtfc_loop\r
969\r
970.dtfc_TileFlipShHP:\r
971 TileFlipShHP\r
972 b .dtfc_loop\r
973\r
974.dtfc_shadow_blank:\r
975 tst r1, #1\r
976 ldrneb r4, [r1]\r
977 mov r6, #0x7f\r
978 and r4, r4, r6\r
979 strneb r4, [r1], #1\r
980 ldrh r4, [r1]\r
981 orr r6, r6, r6, lsl #8\r
982 and r4, r4, r6\r
983 strh r4, [r1], #2\r
984 ldrh r4, [r1]\r
985 and r4, r4, r6\r
986 strh r4, [r1], #2\r
987 ldrh r4, [r1]\r
988 and r4, r4, r6\r
989 strh r4, [r1], #2\r
990 ldrh r4, [r1]\r
991 and r4, r4, r6\r
992 streqh r4, [r1]\r
993 strneb r4, [r1]\r
994 b .dtfc_loop\r
995\r
996.dtfc_cut_tile:\r
997 cmn r4, #8\r
998 ble .dtfc_loop @ off limits\r
999\r
1000 rsb r4, r4, #0 @ 1-7\r
1001 mov r4, r4, lsl #2\r
1002 mvn r12,#0\r
1003 tst r6, #0x0800 @ flipped?\r
1004 moveq r12,r12, lsl r4\r
1005 movne r12,r12, lsr r4\r
1006 and r2, r2, r12, ror #16\r
1007 mov r12,#0xf\r
1008 tst r8, #1\r
1009 bne .dtfc_shadow\r
1010 tst r2, r2\r
1011 beq .dtfc_loop\r
1012 tst r6, #0x0800\r
1013 beq .dtfc_TileNorm\r
1014 b .dtfc_TileFlip\r
1015\r
1016@ check if we have detected layer covered with hi-prio tiles:\r
1017.dtfc_check_rendflags:\r
1018 ldr r2, [r9, #OFS_EST_rendstatus]\r
1019 tst r2, #(PDRAW_PLANE_HI_PRIO|PDRAW_SHHI_DONE)\r
1020 beq .dtfc_loop\r
1021 bic r8, r8, #1 @ sh/hi mode off\r
1022 tst r2, #PDRAW_SHHI_DONE\r
1023 bne .dtfc_loop @ already processed\r
1024 orr r2, r2, #PDRAW_SHHI_DONE\r
1025 str r2, [r9, #OFS_EST_rendstatus]\r
1026\r
1027 add r1, r11,#8\r
1028 mov r3, #320/4/4\r
1029 mov r6, #0x7f\r
1030 orr r6, r6, r6, lsl #8\r
1031 orr r6, r6, r6, lsl #16\r
1032.dtfc_loop_shprep:\r
1033 ldmia r1, {r2,r4,r5,r7}\r
1034 subs r3, r3, #1\r
1035 and r2, r2, r6\r
1036 and r4, r4, r6\r
1037 and r5, r5, r6\r
1038 and r7, r7, r6\r
1039 stmia r1!,{r2,r4,r5,r7}\r
1040 bne .dtfc_loop_shprep\r
1041\r
1042 b .dtfc_loop\r
1043\r
1044.pool\r
1045\r
1046@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1047\r
1048\r
1049@ void DrawSpritesSHi(unsigned char *sprited, struct PicoEState *est)\r
1050\r
1051.global DrawSpritesSHi\r
1052\r
1053DrawSpritesSHi:\r
1054 ldrb r3, [r0]\r
1055 mov r12,#0xff\r
1056 ands r3, r3, #0x7f\r
1057 bxeq lr\r
1058\r
1059 stmfd sp!, {r1,r3-r11,lr} @ +est\r
1060 strb r12,[r0,#3] @ set end marker\r
1061 ldrb r12,[r0,#1]\r
1062 add r10,r0, #4 @ r10=HighLnSpr end\r
1063 mvn r12,r12\r
1064 tst r12,#0x6 @ masking in slot 1 and tile ovfl?\r
1065 ldmeqfd sp!, {r1,r3-r11,pc}\r
1066 add r10,r10,r3 @ r10=HighLnSpr end\r
1067\r
1068 ldrb r12,[r10,#0] @ width of last sprite\r
1069 ldr r11,[r1, #OFS_EST_HighCol]\r
1070 str r12,[sp, #4]\r
1071 mov r12,#0xf\r
1072 ldr lr, [r1, #OFS_EST_PicoMem_vram]\r
1073\r
1074\r
1075DrawSpriteSHi:\r
1076 @ draw next sprite\r
1077 ldr r7, [sp] @ est\r
1078 ldrb r0, [r10,#-1]!\r
1079 ldr r1, [r7, #OFS_EST_HighPreSpr]\r
1080 cmp r0, #0xff\r
1081 ldmeqfd sp!, {r1,r3-r11,pc} @ end of list\r
1082 and r0, r0, #0x7f\r
1083 add r0, r1, r0, lsl #3\r
1084\r
1085 ldr r9, [r0, #4] @ sprite[1]\r
1086 mov r2, r9, asr #16 @ r2=sx\r
1087\r
1088 mov r9, r9, lsl #16\r
1089 mov r3, r9, lsr #31 @ priority\r
1090 mov r9, r9, lsr #16\r
1091@ orr r9, r9, r8, lsl #31 @ r9=code|sh[31] @@ sh is always on here now\r
1092 and r4, r9, #0x6000\r
1093 orr r9, r9, r4, lsl #16\r
1094 orr r9, r9, #0x90000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
1095 cmp r12,r9, lsr #28 @ sh/hi with pal3?\r
1096 cmpne r3, #1 @ if not, is it hi prio?\r
1097 strne r3, [sp, #4] @ reset last sprite width\r
1098 bne DrawSpriteSHi @ non-operator low sprite, already drawn\r
1099\r
1100 ldr r3, [r0] @ sprite[0]\r
1101 mov r6, r3, lsr #28\r
1102 sub r6, r6, #1 @ r6=width-1 (inc later)\r
1103 mov r5, r3, lsr #24\r
1104 and r5, r5, #7 @ r5=height\r
1105\r
1106 ldr r7, [r7, #OFS_EST_DrawScanline]\r
1107 mov r0, r3, lsl #16 @ r4=sy<<16 (tmp)\r
1108\r
1109 sub r7, r7, r0, asr #16 @ r7=row=DrawScanline-sy\r
1110\r
1111 tst r9, #0x1000\r
1112 movne r0, r5, lsl #3\r
1113 subne r0, r0, #1\r
1114 subne r7, r0, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
1115\r
1116 add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
1117 tst r9, #0x0800\r
1118 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1119 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
1120\r
1121 mov r8, r8, lsl #21\r
1122 mov r8, r8, lsr #17\r
1123 and r7, r7, #7\r
1124 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1125\r
1126 ldr r0, [sp, #4]\r
1127 add r6, r6, #1 @ inc now\r
1128 cmp r0, #0 @ check width of last sprite\r
1129 movne r6, r0\r
1130 movne r0, #0\r
1131 strne r0, [sp, #4]\r
1132\r
1133 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
1134 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
1135\r
1136 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1137 b .dsprShi_loop_enter\r
1138\r
1139.dsprShi_loop:\r
1140 subs r6, r6, #1 @ width--\r
1141 beq DrawSpriteSHi\r
1142 adds r0, r0, #8 @ sx+=8\r
1143 add r8, r8, r5 @ tile+=delta\r
1144\r
1145.dsprShi_loop_enter:\r
1146 ble .dsprShi_loop @ sx <= 0\r
1147 cmp r0, #328\r
1148 bge DrawSpriteSHi\r
1149\r
1150 bic r8, r8, #0xf8000 @ tile&=0x7fff; // Clip tile address\r
1151 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(PicoMem.vram+addr); // Get 8 pixels\r
1152 add r1, r11, r0 @ r1=pdest\r
1153 tst r2, r2\r
1154 beq .dsprShi_loop\r
1155\r
1156 cmp r12, r9, lsr #28\r
1157 beq .dsprShi_shadow\r
1158\r
1159 cmp r2, r2, ror #4\r
1160 beq .dsprShi_SingleColor @ tileline singlecolor \r
1161\r
1162 tst r9, #0x0800\r
1163 bne .dsprShi_TileFlip\r
1164\r
1165 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1166@ scratch: r4, r7\r
1167.dsprShi_TileNorm:\r
1168 TileNorm r12\r
1169 b .dsprShi_loop\r
1170\r
1171.dsprShi_TileFlip:\r
1172 TileFlip r12\r
1173 b .dsprShi_loop\r
1174\r
1175.dsprShi_SingleColor:\r
1176 and r4, r2, #0xf\r
1177 orr r4, r3, r4\r
1178 orr r4, r4, r4, lsl #8\r
1179 tst r0, #1 @ not aligned?\r
1180 strneb r4, [r1], #1\r
1181 streqh r4, [r1], #2\r
1182 strh r4, [r1], #2\r
1183 strh r4, [r1], #2\r
1184 strh r4, [r1], #2\r
1185 strneb r4, [r1], #1\r
1186 b .dsprShi_loop\r
1187\r
1188.dsprShi_shadow:\r
1189 tst r9, #0x8000\r
1190 beq .dsprShi_shadow_lowpri\r
1191\r
1192 cmp r2, r2, ror #4\r
1193 beq .dsprShi_singlec_sh\r
1194\r
1195 tst r9, #0x0800\r
1196 bne .dsprShi_TileFlip_sh\r
1197\r
1198 @ (r1=pdest, r2=pixels8, r3=pal) r4, r7: scratch, r12: helper pattern\r
1199.dsprShi_TileNorm_sh:\r
1200 TileNormSh\r
1201 b .dsprShi_loop\r
1202\r
1203.dsprShi_TileFlip_sh:\r
1204 TileFlipSh\r
1205 b .dsprShi_loop\r
1206\r
1207.dsprShi_singlec_sh:\r
1208 cmp r2, #0xe0000000\r
1209 bcc .dsprShi_SingleColor @ normal singlecolor tileline (carry inverted in ARM)\r
1210 tst r2, #0x10000000\r
1211 bne .dsprShi_sh_sh\r
1212 TileSingleHi\r
1213 b .dsprShi_loop\r
1214\r
1215.dsprShi_sh_sh:\r
1216 TileSingleSh\r
1217 b .dsprShi_loop\r
1218\r
1219.dsprShi_shadow_lowpri:\r
1220 tst r9, #0x800\r
1221 bne .dsprShi_TileFlip_sh_lp\r
1222\r
1223.dsprShi_TileNorm_sh_lp:\r
1224 TileNormSh_onlyop_lp\r
1225 b .dsprShi_loop\r
1226\r
1227.dsprShi_TileFlip_sh_lp:\r
1228 TileFlipSh_onlyop_lp\r
1229 b .dsprShi_loop\r
1230\r
1231.pool\r
1232\r
1233@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1234\r
1235@ void DrawAllSprites(unsigned char *sprited, int prio, int sh,\r
1236@ struct PicoEState *est)\r
1237\r
1238.global DrawAllSprites\r
1239\r
1240DrawAllSprites:\r
1241 orr r1, r2, r1, lsl #1\r
1242 ldr r2, [r0]\r
1243 ands r2, r2, #0x7f\r
1244 bxeq lr\r
1245\r
1246 @ time to do some real work\r
1247 stmfd sp!, {r1,r3-r11,lr} @ +sh|prio<<1 +est\r
1248 mov r12,#0xff\r
1249 strb r12,[r0,#3] @ set end marker\r
1250 ldrb r12,[r0,#1]\r
1251 add r10,r0 ,#4\r
1252 mvn r12,r12\r
1253 tst r12,#0x6 @ masking in slot 1 and tile ovfl?\r
1254 ldmeqfd sp!, {r1,r3-r11,pc}\r
1255 add r10,r10,r2 @ r10=HighLnSpr end\r
1256\r
1257 ldrb r12,[r10,#0] @ width of last sprite\r
1258 ldr r11,[r3, #OFS_EST_HighCol]\r
1259 orr r1 ,r1 ,r12,lsl #24\r
1260 str r1, [sp]\r
1261 mov r12,#0xf\r
1262 ldr lr, [r3, #OFS_EST_PicoMem_vram]\r
1263\r
1264@ + 0 : hhhhvvvv ----hhvv yyyyyyyy yyyyyyyy // v, h: horiz. size\r
1265@ + 4 : xxxxxxxx xxxxxxxx pccvhnnn nnnnnnnn // x: x coord + 8\r
1266\r
1267DrawSprite:\r
1268 @ draw next sprite\r
1269 ldrb r0, [r10,#-1]!\r
1270 ldr r4, [sp] @ sh|prio<<1|lastw<<24\r
1271 ldr r7, [sp, #4] @ est\r
1272 mov r2, r0, lsl #24\r
1273 cmp r0, #0xff\r
1274 ldmeqfd sp!, {r1,r3-r11,pc} @ end of list\r
1275 eors r2, r2, r4, lsl #30\r
1276 bic r2, r4, #0xff000000\r
1277 str r2, [sp]\r
1278 bmi DrawSprite @ wrong priority\r
1279 ldr r1, [r7, #OFS_EST_HighPreSpr]\r
1280 and r0, r0, #0x7f\r
1281 add r0, r1, r0, lsl #3\r
1282\r
1283 ldr r3, [r0] @ sprite[0]\r
1284 ldr r7, [r7, #OFS_EST_DrawScanline]\r
1285 mov r6, r3, lsr #28\r
1286 sub r6, r6, #1 @ r6=width-1 (inc later)\r
1287 mov r5, r3, lsr #24\r
1288 and r5, r5, #7 @ r5=height\r
1289\r
1290 mov r8, r3, lsl #16 @ r8=sy<<16 (tmp)\r
1291\r
1292 ldr r9, [r0, #4]\r
1293 sub r7, r7, r8, asr #16 @ r7=row=DrawScanline-sy\r
1294\r
1295 mov r2, r9, asr #16 @ r2=sx\r
1296 mov r9, r9, lsl #16\r
1297 mov r9, r9, lsr #16\r
1298 orr r9, r9, r4, lsl #31 @ r9=code|sh[31]\r
1299\r
1300 tst r9, #0x1000\r
1301 movne r8, r5, lsl #3\r
1302 subne r8, r8, #1\r
1303 subne r7, r8, r7 @ if (code&0x1000) row=(height<<3)-1-row; // Flip Y\r
1304\r
1305 add r8, r9, r7, lsr #3 @ tile+=row>>3; // Tile number increases going down\r
1306 tst r9, #0x0800\r
1307 mlane r8, r5, r6, r8 @ if (code&0x0800) { tile+=delta*(width-1);\r
1308 rsbne r5, r5, #0 @ delta=-delta; } // r5=delta now\r
1309\r
1310 mov r8, r8, lsl #21\r
1311 mov r8, r8, lsr #17\r
1312 and r7, r7, #7\r
1313 add r8, r8, r7, lsl #1 @ tile+=(row&7)<<1; // Tile address\r
1314\r
1315 add r6, r6, #1 @ inc now\r
1316 cmp r4, #0x1000000 @ check width of last sprite\r
1317 movhs r6, r4, lsr #24\r
1318\r
1319 @ cache some stuff to avoid mem access\r
1320 mov r5, r5, lsl #4 @ delta<<=4; // Delta of address\r
1321 and r4, r9, #0x6000\r
1322 orr r9, r9, r4, lsl #16\r
1323 orrs r9, r9, #0x10000000 @ r9=scc1 ???? ... <code> (s=shadow/hilight, cc=pal)\r
1324\r
1325 mov r3, r4, lsr #9 @ r3=pal=((code>>9)&0x30);\r
1326 orrmi r3, r3, #0x80 @ for sh/hi\r
1327\r
1328 adds r0, r2, #0 @ mov sx to r0 and set ZV flags\r
1329 b .dspr_loop_enter\r
1330\r
1331.dspr_loop:\r
1332 subs r6, r6, #1 @ width--\r
1333 beq DrawSprite\r
1334 adds r0, r0, #8 @ sx+=8\r
1335 add r8, r8, r5 @ tile+=delta\r
1336\r
1337.dspr_loop_enter:\r
1338 ble .dspr_loop @ sx <= 0\r
1339 cmp r0, #328\r
1340 bge DrawSprite\r
1341\r
1342 bic r8, r8, #0xf8000 @ tile&=0x7fff; // Clip tile address\r
1343 ldr r2, [lr, r8, lsl #1] @ pack=*(unsigned int *)(PicoMem.vram+addr); // Get 8 pixels\r
1344 add r1, r11, r0 @ r1=pdest\r
1345 tst r2, r2\r
1346 beq .dspr_loop\r
1347\r
1348 cmp r12, r9, lsr #28\r
1349 beq .dspr_shadow\r
1350\r
1351 tst r9, #0x80000000\r
1352 bne .dspr_shnonsh\r
1353\r
1354 cmp r2, r2, ror #4\r
1355 beq .dspr_SingleColor @ tileline singlecolor \r
1356\r
1357 tst r9, #0x0800\r
1358 bne .dspr_TileFlip\r
1359\r
1360 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1361@ scratch: r4, r7\r
1362.dspr_TileNorm:\r
1363 TileNorm r12\r
1364 b .dspr_loop\r
1365\r
1366.dspr_TileFlip:\r
1367 TileFlip r12\r
1368 b .dspr_loop\r
1369\r
1370.dspr_singlec_sh:\r
1371 cmp r2, #0xe0000000\r
1372 bcs .dspr_TileNorm_sh @ op. tileline, markop. XXX: maybe add a spec. handler?\r
1373\r
1374.dspr_SingleColor:\r
1375 and r4, r2, #0xf\r
1376 orr r4, r3, r4\r
1377 orr r4, r4, r4, lsl #8\r
1378 tst r0, #1 @ not aligned?\r
1379 strneb r4, [r1], #1\r
1380 streqh r4, [r1], #2\r
1381 strh r4, [r1], #2\r
1382 strh r4, [r1], #2\r
1383 strh r4, [r1], #2\r
1384 strneb r4, [r1], #1\r
1385 b .dspr_loop\r
1386\r
1387.dspr_shnonsh:\r
1388 tst r9, #0x0800\r
1389 bne .dspr_TileFlipNonSH\r
1390\r
1391 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1392@ scratch: r4, r7\r
1393.dspr_TileNormNonSH:\r
1394 TileNormNonSH r12\r
1395 b .dspr_loop\r
1396\r
1397.dspr_TileFlipNonSH:\r
1398 TileFlipNonSH r12\r
1399 b .dspr_loop\r
1400\r
1401.dspr_shadow:\r
1402 cmp r2, r2, ror #4\r
1403 beq .dspr_singlec_sh\r
1404\r
1405 tst r9, #0x0800\r
1406 bne .dspr_TileFlip_sh\r
1407\r
1408 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r12: helper pattern\r
1409.dspr_TileNorm_sh:\r
1410 TileNormSh_markop\r
1411 b .dspr_loop\r
1412\r
1413.dspr_TileFlip_sh:\r
1414 TileFlipSh_markop\r
1415 b .dspr_loop\r
1416\r
1417\r
1418@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1419\r
1420@ void DrawWindow(int tstart, int tend, int prio, int sh\r
1421@ struct PicoEState *est)\r
1422\r
1423.global DrawWindow\r
1424\r
1425DrawWindow:\r
1426 ldr r12, [sp] @ est\r
1427 stmfd sp!, {r4-r11,lr}\r
1428\r
1429 ldr r6, [r12, #OFS_EST_Pico]\r
1430 ldr r10, [r12, #OFS_EST_DrawScanline]\r
1431 mov r11, r12 @ est\r
1432 ldrb r12, [r6, #OFS_Pico_video_reg+3] @ pvid->reg[3]\r
1433\r
1434 ldr r4, [r6, #OFS_Pico_video_reg+12]\r
1435 mov r5, r10, lsr #3\r
1436 and r10, r10, #7\r
1437 mov r10, r10, lsl #1 @ r10=ty\r
1438\r
1439 ldr r6, [r11, #OFS_EST_rendstatus]\r
1440 ldr lr, [r11, #OFS_EST_PicoMem_vram]\r
1441\r
1442 mov r12, r12, lsl #10\r
1443\r
1444 tst r4, #1 @ 40 cell mode?\r
1445 andne r12, r12, #0xf000 @ 0x3c<<10\r
1446 andeq r12, r12, #0xf800\r
1447 addne r12, r12, r5, lsl #7\r
1448 addeq r12, r12, r5, lsl #6 @ nametab\r
1449 add r12, r12, r0, lsl #2 @ +starttile\r
1450\r
1451 ands r6, r6, #PDRAW_WND_DIFF_PRIO\r
1452 cmpeq r2, #1 @ prio && !(rendstatus & WND_DIFF_PRIO)?\r
1453 ldmeqfd sp!, {r4-r11,pc} @ yes, assume that whole window uses same priority\r
1454\r
1455 orr r6, r6, r2\r
1456 orr r6, r6, r3, lsl #8 @ shadow mode\r
1457\r
1458 sub r8, r1, r0\r
1459\r
1460 @ cache some stuff to avoid mem access\r
1461 ldr r11, [r11, #OFS_EST_HighCol]\r
1462 mov r8, r8, lsl #1 @ cells\r
1463 mvn r9, #0 @ r9=prevcode=-1\r
1464 add r1, r11, r0, lsl #4 @ r1=pdest=HighCol+starttile (+8 added in loop)\r
1465 mov r0, #0xf\r
1466\r
1467 @ r4,r5 are scratch in this loop\r
1468.dwloop:\r
1469 add r1, r1, #8\r
1470.dwloop_nor1:\r
1471 ldrh r7, [lr, r12] @ r7=code (int, but from unsigned, no sign extend)\r
1472 add r12, r12, #2 @ halfwords\r
1473 subs r8, r8, #1\r
1474 bmi .dwloop_end @ done\r
1475\r
1476 eor r5, r6, r7, lsr #15\r
1477 tst r5, #1\r
1478 orrne r6, r6, #PDRAW_WND_DIFF_PRIO @ wrong pri\r
1479 bne .dwloop\r
1480\r
1481 cmp r7, r9\r
1482 beq .dw_samecode @ we know stuff about this tile already\r
1483\r
1484 mov r9, r7 @ remember code\r
1485\r
1486 tst r9, #0x1000 @ if (code&0x1000)\r
1487 mov r2, r9, lsl #21\r
1488 add r2, r2, r10, lsl #17\r
1489 eorne r2, r2, #0xe<<17 @ if (code&0x1000) addr^=0xe;\r
1490\r
1491 ldr r2, [lr, r2, lsr #16] @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r
1492\r
1493 and r3, r9, #0x6000\r
1494 mov r3, r3, lsr #9 @ r3=pal=((code&0x6000)>>9);\r
1495\r
1496.dw_samecode:\r
1497 tst r6, #0x100\r
1498 bne .dw_shadow\r
1499.dw_shadow_done:\r
1500 tst r2, r2\r
1501 beq .dwloop @ tileline blank\r
1502\r
1503 cmp r2, r2, ror #4\r
1504 beq .dw_SingleColor @ tileline singlecolor \r
1505\r
1506 tst r9, #0x0800\r
1507 bne .dw_TileFlip\r
1508\r
1509 @ (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: helper pattern\r
1510.dw_TileNorm:\r
1511 TileNorm r0\r
1512 b .dwloop\r
1513\r
1514.dw_TileFlip:\r
1515 TileFlip r0\r
1516 b .dwloop\r
1517\r
1518.dw_SingleColor:\r
1519 and r4, r0, r2 @ #0x0000000f\r
1520 orr r4, r3, r4\r
1521 orr r4, r4, r4, lsl #8\r
1522 orr r4, r4, r4, lsl #16\r
1523 mov r5, r4\r
1524 stmia r1!, {r4,r5}\r
1525 b .dwloop_nor1 @ we incremented r1 ourselves\r
1526\r
1527.dw_shadow:\r
1528 tst r6, #1 @ hi pri?\r
1529 orreq r3, r3, #0x80\r
1530 beq .dw_shadow_done\r
1531 ldr r4, [r1]\r
1532 mov r5, #0x7f\r
1533 orr r5, r5, r5, lsl #8\r
1534 orr r5, r5, r5, lsl #16\r
1535 and r4, r4, r5\r
1536 str r4, [r1]\r
1537 ldr r4, [r1,#4]\r
1538 and r4, r4, r5\r
1539 str r4, [r1,#4]\r
1540 b .dw_shadow_done\r
1541\r
1542.dwloop_end:\r
1543 and r2, r6, #PDRAW_WND_DIFF_PRIO\r
1544 ldmfd sp!, {r4-r11,lr}\r
1545 ldr r0, [sp]\r
1546 ldr r1, [r0, #OFS_EST_rendstatus]\r
1547 orr r1, r1, r2\r
1548 str r1, [r0, #OFS_EST_rendstatus]\r
1549\r
1550 bx lr\r
1551\r
1552\r
1553@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
1554\r
1555\r
1556@ Convert 0000bbb0 ggg0rrr0\r
1557@ to rrrrrggg gggbbbbb\r
1558\r
1559@ r2,r3 - scratch, lr = 0x001c001c, r8 = 0x08610861\r
1560.macro convRGB565 reg\r
1561 and r2, lr, \reg,lsr #7 @ b\r
1562 and r3, lr, \reg,lsr #3 @ g\r
1563 and \reg, lr, \reg,lsl #1 @ r\r
1564 orr r2, r2, r3, lsl #6\r
1565 orr \reg, r2, \reg,lsl #11\r
1566\r
1567 and r2, r8, \reg,lsr #4\r
1568 orr \reg, \reg, r2\r
1569\r
1570 orr r3, \reg, \reg,lsr #1\r
1571 and r3, r3, r8,lsl #2\r
1572 bic r3, r3, r2,lsl #2\r
1573 orr \reg, \reg, r3,lsr #1\r
1574.endm\r
1575\r
1576@ trashes: r2-r8,r12,lr; r8 = 0x08610861; r0,r1 are advanced\r
1577.macro vidConvCpyRGB565_local\r
1578 mov r12, r2, lsr #3 @ repeats\r
1579 mov lr, #0x001c0000\r
1580 orr lr, lr, #0x01c @ lr == pattern 0x001c001c\r
1581\r
15820:\r
1583 ldmia r1!, {r4-r7}\r
1584 subs r12, r12, #1\r
1585 convRGB565 r4\r
1586 str r4, [r0], #4\r
1587 convRGB565 r5\r
1588 str r5, [r0], #4\r
1589 convRGB565 r6\r
1590 str r6, [r0], #4\r
1591 convRGB565 r7\r
1592 str r7, [r0], #4\r
1593\r
1594 bgt 0b\r
1595.endm\r
1596\r
1597\r
1598.global vidConvCpyRGB565\r
1599\r
1600vidConvCpyRGB565: @ void *to, void *from, int pixels\r
1601 stmfd sp!, {r4-r9,lr}\r
1602 mov r8, #0x0061\r
1603 orr r8, r8, #0x0800\r
1604 orr r8, r8, r8, lsl #16\r
1605 vidConvCpyRGB565_local\r
1606 ldmfd sp!, {r4-r9,pc}\r
1607\r
1608\r
1609@ void PicoDoHighPal555(int sh, int line, struct PicoEState *est)\r
1610\r
1611.global PicoDoHighPal555\r
1612\r
1613PicoDoHighPal555:\r
1614 stmfd sp!, {r4-r10,lr}\r
1615 mov r10,r2 @ est\r
1616 ldr r8, [r10, #OFS_EST_Pico]\r
1617\r
1618 mov r9, r0\r
1619\r
1620 add r0, r10, #OFS_EST_HighPal\r
1621\r
1622 mov r1, #0\r
1623 strb r1, [r8, #OFS_Pico_m_dirtyPal]\r
1624\r
1625 ldr r1, [r10, #OFS_EST_PicoMem_cram]\r
1626 mov r2, #0x40\r
1627 mov r8, #0x0061\r
1628 orr r8, r8, #0x0800\r
1629 orr r8, r8, r8, lsl #16\r
1630\r
1631 vidConvCpyRGB565_local\r
1632\r
1633 cmp r9, #0\r
1634 beq PicoDoHighPal555_end\r
1635\r
1636 add r3, r10, #OFS_EST_HighPal\r
1637 add r4, r3, #0x40*2\r
1638\r
1639 @ hilighted pixels (0x40-0x7f):\r
1640 @ t = ((dpal[i] >> 1) & 0x738e738e) + 0x738e738e;\r
1641 @ t += 0x08610861;\r
1642 @ r8=0x08610861\r
1643 mov r12, #0x008e\r
1644 orr r12,r12,#0x7300\r
1645 orr r12,r12,r12,lsl #16\r
1646 add r10,r12,r8\r
1647 mov lr, #0x40/4\r
1648.fl_loopcpRGB555_hi:\r
1649 ldmia r3!, {r1,r6}\r
1650 and r1, r12, r1, lsr #1\r
1651 and r6, r12, r6, lsr #1\r
1652 add r1, r10, r1\r
1653 add r6, r10, r6\r
1654 stmia r4!, {r1,r6}\r
1655 subs lr, lr, #1\r
1656 bne .fl_loopcpRGB555_hi\r
1657\r
1658 sub r3, r3, #0x40*2\r
1659 @ shadowed (0x80-0xbf), shadow|hilight (aka normal, 0xc0-0xff) pixels:\r
1660 @ t = (dpal[i] >> 1) & 0x738e738e;\r
1661 @ t += ((t>>2|t>>1|t>>0) & (0x08610861<<1));\r
1662 add r5, r3, #0xc0*2\r
1663 mov lr, #0x40/4\r
1664.fl_loopcpRGB555_sh:\r
1665 ldmia r3!, {r1,r6}\r
1666 subs lr, lr, #1\r
1667 stmia r5!, {r1,r6} @ 0xc0, normal\r
1668 and r1, r12, r1, lsr #1\r
1669 orr r0, r1, r1, lsr #1\r
1670 orr r0, r0, r1, lsr #2\r
1671 and r0, r0, r8, lsl #1\r
1672 add r1, r1, r0\r
1673 and r6, r12, r6, lsr #1\r
1674 orr r0, r6, r6, lsr #1\r
1675 orr r0, r0, r6, lsr #2\r
1676 and r0, r0, r8, lsl #1\r
1677 add r6, r6, r0\r
1678 stmia r4!, {r1,r6}\r
1679 bne .fl_loopcpRGB555_sh\r
1680\r
1681 mov r0, #1\r
1682PicoDoHighPal555_end:\r
1683 ldmfd sp!, {r4-r10,pc}\r
1684\r
1685\r
1686@ void FinalizeLine555(int sh, int line, struct PicoEState *est)\r
1687\r
1688.global FinalizeLine555\r
1689\r
1690FinalizeLine555:\r
1691 ldr r3, [r2, #OFS_EST_rendstatus]\r
1692 mov r0, r2\r
1693 tst r3, #PDRAW_BGC_DMA\r
1694 bne BgcDMA\r
1695\r
1696 stmfd sp!, {r4-r11,lr}\r
1697 mov r11,r2 @ est\r
1698 mov r4, r3\r
1699\r
1700 bl PicoDrawUpdateHighPal\r
1701\r
1702 ldr r8, [r11, #OFS_EST_Pico]\r
1703 add r3, r11, #OFS_EST_HighPal\r
1704\r
1705 mov lr, #0xff\r
1706 mov lr, lr, lsl #1\r
1707\r
1708 ldr r5, [r11, #OFS_EST_PicoOpt]\r
1709 ldr r1, [r11, #OFS_EST_HighCol]\r
1710 ldr r0, [r11, #OFS_EST_DrawLineDest]\r
1711 ldr r7, [r5, #OFS_PicoIn_AHW-OFS_PicoIn_opt]\r
1712 ldrb r12,[r8, #OFS_Pico_video_reg+12]\r
1713 ldrb r6, [r8, #OFS_Pico_video_reg+0]\r
1714 ldr r2, [r8, #OFS_Pico_m_hardware]\r
1715 add r1, r1, #8\r
1716\r
1717 tst r7, #0x20 @ GG ?\r
1718 tstne r2, #0x2 @ LCD ?\r
1719 bne .fl_gg20col\r
1720\r
1721 tst r7, #0x10 @ SMS ?\r
1722 beq .fl_noSMS\r
1723\r
1724 tst r6, #0x20\r
1725 movne r2, #248/8 @ len = 248\r
1726 addne r1, r1, #8 @ ps += 8\r
1727 moveq r2, #256/8 @ len = 256\r
1728 b .fl_check32scaling\r
1729\r
1730.fl_gg20col:\r
1731 mov r2, #160/8 @ len = 160\r
1732 tst r4, #PDRAW_SOFTSCALE\r
1733 bne .fl_20scale_RGB555 @ scale 160->320\r
1734 b .fl_checkborder\r
1735\r
1736.fl_noSMS:\r
1737 tst r12, #1 @ h32?\r
1738 movne r2, #320/8 @ len = 320\r
1739 bne .fl_40colRGB555\r
1740 mov r2, #256/8 @ len = 256\r
1741\r
1742.fl_check32scaling:\r
1743 tst r4, #PDRAW_SOFTSCALE\r
1744 rsbne r7, r2, #256/8\r
1745 addne r0, r0, r7, lsl #3 @ pd += (256-len)>>1\r
1746 bne .fl_32scale_RGB555 @ scale 256->320\r
1747\r
1748.fl_checkborder:\r
1749 tst r4, #PDRAW_BORDER_32\r
1750 rsbne r7, r2, #320/8 @ pd += (320-len)/2\r
1751 addne r0, r0, r7, lsl #3\r
1752\r
1753.fl_40colRGB555:\r
1754#ifdef UNALIGNED_DRAWLINEDEST\r
1755 @ this is basically for Gizmondo, which has unaligned odd lines in the framebuffer\r
1756 tst r0, #2\r
1757 bne .fl_RGB555u\r
1758#endif\r
1759\r
1760.fl_loopRGB555:\r
1761 ldr r12, [r1], #4\r
1762 ldr r7, [r1], #4\r
1763\r
1764 and r4, lr, r12, lsl #1\r
1765 ldrh r4, [r3, r4]\r
1766 and r5, lr, r12, lsr #7\r
1767 ldrh r5, [r3, r5]\r
1768 and r6, lr, r12, lsr #15\r
1769 ldrh r6, [r3, r6]\r
1770 orr r4, r4, r5, lsl #16\r
1771\r
1772 and r5, lr, r12, lsr #23\r
1773 ldrh r5, [r3, r5]\r
1774 and r8, lr, r7, lsl #1\r
1775 ldrh r8, [r3, r8]\r
1776 orr r5, r6, r5, lsl #16\r
1777\r
1778 and r6, lr, r7, lsr #7\r
1779 ldrh r6, [r3, r6]\r
1780 and r12,lr, r7, lsr #15\r
1781 ldrh r12,[r3, r12]\r
1782 and r7, lr, r7, lsr #23\r
1783 ldrh r7, [r3, r7]\r
1784 orr r8, r8, r6, lsl #16\r
1785\r
1786 subs r2, r2, #1\r
1787 orr r12,r12, r7, lsl #16\r
1788\r
1789 stmia r0!, {r4,r5,r8,r12}\r
1790 bne .fl_loopRGB555\r
1791\r
1792 ldmfd sp!, {r4-r11,pc}\r
1793\r
1794\r
1795.fl_32scale_RGB555:\r
1796 ldr r5, [r5, #OFS_PicoIn_filter-OFS_PicoIn_opt]\r
1797\r
1798 mov r9, #0xf700 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
1799 orr r9, r9, #0x00de\r
1800\r
1801#ifdef UNALIGNED_DRAWLINEDEST\r
1802 tst r0, #2\r
1803 bne .fl_32scale_RGB555u\r
1804#endif\r
1805\r
1806 and r5, r5, #0x3\r
1807 add pc, pc, r5, lsl #2\r
1808 nop\r
1809 b .fl_32scale_nn\r
1810 b .fl_32scale_snn\r
1811 b .fl_32scale_bl2\r
1812 b .fl_32scale_bl4\r
1813\r
1814.fl_32scale_nn:\r
1815 ldr r12, [r1], #4\r
1816 ldr r7, [r1], #4\r
1817\r
1818 and r4, lr, r12, lsl #1\r
1819 ldrh r4, [r3, r4]\r
1820 and r5, lr, r12, lsr #7\r
1821 ldrh r5, [r3, r5]\r
1822 and r6, lr, r12, lsr #15\r
1823 ldrh r6, [r3, r6]\r
1824 and r10,lr, r12, lsr #23\r
1825 ldrh r10,[r3, r10]\r
1826\r
1827 orr r4, r4, r5, lsl #16\r
1828 orr r5, r6, r6, lsl #16\r
1829\r
1830 and r6, lr, r7, lsl #1\r
1831 ldrh r6, [r3, r6]\r
1832 and r8, lr, r7, lsr #7\r
1833 ldrh r8, [r3, r8]\r
1834 and r12,lr, r7, lsr #15\r
1835 ldrh r12,[r3, r12]\r
1836 and r7, lr, r7, lsr #23\r
1837 ldrh r7, [r3, r7]\r
1838\r
1839 orr r6, r10,r6, lsl #16\r
1840 orr r8, r8,r12, lsl #16\r
1841\r
1842 subs r2, r2, #1\r
1843\r
1844 orr r10,r12,r7, lsl #16\r
1845\r
1846 stmia r0!, {r4,r5,r6,r8,r10}\r
1847 bne .fl_32scale_nn\r
1848\r
1849 b .fl_32scale_8bit\r
1850\r
1851.fl_32scale_snn:\r
1852 ldr r12, [r1], #4\r
1853 ldr r7, [r1], #4\r
1854\r
1855 and r4, lr, r12, lsl #1\r
1856 ldrh r4, [r3, r4]\r
1857 and r5, lr, r12, lsr #7\r
1858 ldrh r5, [r3, r5]\r
1859 and r6, lr, r12, lsr #15\r
1860 ldrh r6, [r3, r6]\r
1861 and r10,lr, r12, lsr #23\r
1862 ldrh r10,[r3, r10]\r
1863\r
1864 and r4, r4, r9\r
1865 and r5, r5, r9\r
1866 orr r4, r4, r5, lsl #16\r
1867 and r6, r6, r9\r
1868 add r5, r5, r6\r
1869 mov r5, r5, lsr #1\r
1870 orr r5, r5, r6, lsl #16\r
1871\r
1872 and r6, lr, r7, lsl #1\r
1873 ldrh r6, [r3, r6]\r
1874 and r8, lr, r7, lsr #7\r
1875 ldrh r8, [r3, r8]\r
1876 and r12,lr, r7, lsr #15\r
1877 ldrh r12,[r3, r12]\r
1878 and r7, lr, r7, lsr #23\r
1879 ldrh r7, [r3, r7]\r
1880\r
1881 and r6, r6, r9\r
1882 and r10,r10,r9\r
1883 orr r6, r10,r6, lsl #16\r
1884 and r12,r12,r9\r
1885 and r7, r7, r9\r
1886 orr r10,r12,r7, lsl #16\r
1887\r
1888 and r8, r8, r9\r
1889 add r12,r12,r8\r
1890 mov r12,r12,lsr #1\r
1891 orr r8, r8,r12, lsl #16\r
1892\r
1893 subs r2, r2, #1\r
1894\r
1895 stmia r0!, {r4,r5,r6,r8,r10}\r
1896 bne .fl_32scale_snn\r
1897\r
1898 b .fl_32scale_8bit\r
1899\r
1900.fl_32scale_bl2:\r
1901 ldr r12, [r1], #4\r
1902 ldr r7, [r1], #4\r
1903\r
1904 and r4, lr, r12, lsl #1\r
1905 ldrh r4, [r3, r4]\r
1906 and r5, lr, r12, lsr #7\r
1907 ldrh r5, [r3, r5]\r
1908 and r6, lr, r12, lsr #15\r
1909 ldrh r6, [r3, r6]\r
1910\r
1911 and r4, r4, r9\r
1912 and r5, r5, r9\r
1913 add r10,r4, r5\r
1914 mov r10,r10,lsr #1\r
1915 orr r4, r4, r10,lsl #16 @ px0 | (px0+px1)/2\r
1916\r
1917 and r6, r6, r9\r
1918 add r5, r5, r6\r
1919 mov r5, r5, lsr #1\r
1920 orr r5, r5, r6, lsl #16 @ (px1+px2)/2 | px2\r
1921\r
1922 and r10,lr, r12, lsr #23\r
1923 ldrh r10,[r3, r10]\r
1924 and r8, lr, r7, lsl #1\r
1925 ldrh r8, [r3, r8]\r
1926\r
1927 and r10,r10,r9\r
1928 and r8, r8, r9\r
1929 orr r6, r10,r8, lsl #16 @ px3 | px4\r
1930\r
1931 and r12,lr, r7, lsr #15\r
1932 ldrh r12,[r3, r12]\r
1933 and r10, lr, r7, lsr #23\r
1934 ldrh r10, [r3, r10]\r
1935 and r7, lr, r7, lsr #7\r
1936 ldrh r7, [r3, r7]\r
1937\r
1938 and r12,r12,r9\r
1939 and r10,r10,r9\r
1940 orr r10,r12,r10, lsl #16 @ px6 | px7\r
1941\r
1942 and r7, r7, r9\r
1943 add r12,r12,r7\r
1944 add r8, r8, r7\r
1945 mov r8, r8, lsr #1\r
1946 mov r12,r12,lsr #1\r
1947 orr r8, r8,r12, lsl #16 @ (px4+px5)/2 | (px5+px6)/2\r
1948\r
1949 subs r2, r2, #1\r
1950\r
1951 stmia r0!, {r4,r5,r6,r8,r10}\r
1952 bne .fl_32scale_bl2\r
1953\r
1954 b .fl_32scale_8bit\r
1955\r
1956.fl_32scale_bl4:\r
1957 // TODO this should reflect the bl4 C algorithm, but it doesn't, it's bln.\r
1958 and r9, r9, r9, lsl #1 @ nuke 2 LSBs to avoid spilling for n/4\r
1959.fl_32loop_bl4:\r
1960 ldr r12, [r1], #4\r
1961 ldr r7, [r1], #4\r
1962\r
1963 and r4, lr, r12,lsl #1\r
1964 ldrh r4, [r3, r4]\r
1965 and r5, lr, r12,lsr #7\r
1966 ldrh r5, [r3, r5]\r
1967\r
1968 @ r4 = 1/4px0+3/4px1 : px0\r
1969 and r4, r4, r9\r
1970 orr r4, r4, r4, lsl #14 @ r4[31:16] = 1/4 pix_s 0\r
1971 and r5, r5, r9\r
1972 sub r6, r5, r5, lsr #2 @ r6 = 3/4 pix_s 1\r
1973 add r4, r4, r6, lsl #16 @ pix_d 0, 1\r
1974\r
1975 and r6, lr, r12,lsr #15\r
1976 ldrh r6, [r3, r6]\r
1977 and r12,lr, r12,lsr #23\r
1978 ldrh r12,[r3, r12]\r
1979\r
1980 @ r5 = 3/4px2+1/4px3 : (px1+px2)/2\r
1981 and r6, r6, r9\r
1982 add r5, r5, r6\r
1983 mov r5, r5, lsr #1\r
1984 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
1985 orr r5, r5, r6, lsl #16\r
1986\r
1987 and r6, lr, r7, lsl #1\r
1988 ldrh r6, [r3, r6]\r
1989 and r12,r12,r9\r
1990 add r5, r5, r12,lsl #14 @ pix_d 2, 3\r
1991\r
1992 @ r6 = px4 : px3\r
1993 and r6, r6, r9\r
1994 orr r6, r12,r6, lsl #16 @ pix_d 4, 5\r
1995\r
1996 @ r8 = (px5+px6)/2 : 1/4px4+3/4px5\r
1997 and r12,lr, r7, lsr #7\r
1998 ldrh r12,[r3, r12]\r
1999 and r10,lr, r7, lsr #15\r
2000 ldrh r10,[r3, r10]\r
2001 and r12,r12,r9\r
2002 sub r8, r12,r12,lsr #2 @ r8 = 3/4 pix_s 1\r
2003 add r8, r8, r6, lsr #18\r
2004\r
2005 and r7, lr, r7, lsr #23\r
2006 ldrh r7, [r3, r7]\r
2007 and r10,r10,r9\r
2008 orr r8, r8, r10,lsl #15\r
2009 add r8, r8, r12,lsl #15 @ pix_d 6, 7\r
2010\r
2011 @ r10 = px7 : 3/4px6+1/4px7\r
2012 sub r10,r10,r10,lsr #2 @ r10= 3/4 pix_s 2\r
2013 and r7, r7, r9\r
2014 add r10,r10,r7, lsr #2 @ += 1/4 pix_s 3\r
2015 orr r10,r10,r7, lsl #16 @ pix_d 8, 9\r
2016\r
2017 subs r2, r2, #1\r
2018\r
2019 stmia r0!, {r4,r5,r6,r8,r10}\r
2020 bne .fl_32loop_bl4\r
2021\r
2022.fl_32scale_8bit:\r
2023 ldr r4, [r11, #OFS_EST_rendstatus]\r
2024 add r0, r1, #320-256\r
2025 mov r2, #256/8\r
2026 tst r4, #PDRAW_32X_SCALE\r
2027 ldmeqfd sp!, {r4-r11,pc}\r
2028 mov lr, #0xff\r
2029\r
2030.fl_32scale_8bit_nn:\r
2031 ldr r7, [r1, #-4]!\r
2032 ldr r12, [r1, #-4]!\r
2033\r
2034 and r4, lr, r12, lsl #0\r
2035 and r5, lr, r12, lsr #8\r
2036 and r6, lr, r12, lsr #16\r
2037 and r10,lr, r12, lsr #24\r
2038 \r
2039 orr r4, r4, r5, lsl #8\r
2040 orr r5, r6, r6, lsl #8\r
2041\r
2042 and r6, lr, r7, lsl #0\r
2043 and r8, lr, r7, lsr #8\r
2044 and r12,lr, r7, lsr #16\r
2045 and r7, lr, r7, lsr #24\r
2046\r
2047 orr r6, r10,r6, lsl #8\r
2048 orr r8, r8,r12, lsl #8\r
2049\r
2050 subs r2, r2, #1\r
2051\r
2052 orr r10,r12,r7, lsl #8\r
2053\r
2054 strh r10, [r0, #-2]!\r
2055 strh r8, [r0, #-2]!\r
2056 strh r6, [r0, #-2]!\r
2057 strh r5, [r0, #-2]!\r
2058 strh r4, [r0, #-2]!\r
2059\r
2060 bne .fl_32scale_8bit_nn\r
2061\r
2062 ldmfd sp!, {r4-r11,pc}\r
2063\r
2064\r
2065.fl_20scale_RGB555:\r
2066 ldr r5, [r5, #OFS_PicoIn_filter-OFS_PicoIn_opt]\r
2067\r
2068 mov r9, #0xf700 @ f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
2069 orr r9, r9, #0x00de\r
2070\r
2071#ifdef UNALIGNED_DRAWLINEDEST\r
2072 tst r0, #2\r
2073 bne .fl_20scale_RGB555u\r
2074#endif\r
2075\r
2076 and r5, r5, #0x2\r
2077 add pc, pc, r5, lsl #1\r
2078 nop\r
2079 b .fl_20scale_nn\r
2080 b .fl_20scale_bl2\r
2081\r
2082.fl_20scale_nn:\r
2083 ldr r12, [r1], #4\r
2084 ldr r7, [r1], #4\r
2085\r
2086 and r4, lr, r12, lsl #1\r
2087 ldrh r4, [r3, r4]\r
2088 and r5, lr, r12, lsr #7\r
2089 ldrh r5, [r3, r5]\r
2090 and r6, lr, r12, lsr #15\r
2091 ldrh r6, [r3, r6]\r
2092 and r8 ,lr, r12, lsr #23\r
2093 ldrh r8 ,[r3, r8]\r
2094\r
2095 orr r4, r4, r4, lsl #16\r
2096 orr r5, r5, r5, lsl #16\r
2097 orr r6, r6, r6, lsl #16\r
2098 orr r8, r8, r8, lsl #16\r
2099 stmia r0!, {r4,r5,r6,r8}\r
2100\r
2101 and r4, lr, r7, lsl #1\r
2102 ldrh r4, [r3, r4]\r
2103 and r5, lr, r7, lsr #7\r
2104 ldrh r5, [r3, r5]\r
2105 and r6 ,lr, r7, lsr #15\r
2106 ldrh r6 ,[r3, r6]\r
2107 and r8, lr, r7, lsr #23\r
2108 ldrh r8, [r3, r8]\r
2109\r
2110 orr r4, r4, r4, lsl #16\r
2111 orr r5, r5, r5, lsl #16\r
2112 orr r6, r6, r6, lsl #16\r
2113 orr r8, r8, r8, lsl #16\r
2114 stmia r0!, {r4,r5,r6,r8}\r
2115\r
2116 subs r2, r2, #1\r
2117 bne .fl_20scale_nn\r
2118\r
2119 ldmfd sp!, {r4-r11,pc}\r
2120\r
2121\r
2122.fl_20scale_bl2:\r
2123 ldr r8, [r1]\r
2124 and r8, lr, r8, lsl #1\r
2125 ldrh r8, [r3, r8]\r
2126 and r8, r8, r9\r
2127 mov r8, r8, lsl #16\r
2128\r
2129.fl_20loop_bl2:\r
2130 ldr r12, [r1], #4\r
2131 ldr r7, [r1], #4\r
2132\r
2133 and r4, lr, r12, lsl #1\r
2134 ldrh r4, [r3, r4]\r
2135 and r5, lr, r12, lsr #7\r
2136 ldrh r5, [r3, r5]\r
2137 and r6, lr, r12, lsr #15\r
2138 ldrh r6, [r3, r6]\r
2139\r
2140 and r4, r4, r9\r
2141 add r10,r4, r8, lsr #16\r
2142 mov r10,r10,lsr #1\r
2143 orr r4, r10,r4, lsl #16 @ (px-1+px0)/2 | px0\r
2144\r
2145 and r8 ,lr, r12, lsr #23\r
2146 ldrh r8 ,[r3, r8]\r
2147\r
2148 and r5, r5, r9\r
2149 add r10,r5, r4, lsr #16\r
2150 mov r10,r10,lsr #1\r
2151 orr r5, r10,r5, lsl #16 @ (px0 +px1)/2 | px1\r
2152\r
2153 and r6, r6, r9\r
2154 add r10,r6, r5, lsr #16\r
2155 mov r10,r10,lsr #1\r
2156 orr r6, r10,r6, lsl #16 @ (px1 +px2)/2 | px2\r
2157\r
2158 and r8, r8, r9\r
2159 add r10,r8, r6, lsr #16\r
2160 mov r10,r10,lsr #1\r
2161 orr r8, r10,r8, lsl #16 @ (px2 +px3)/2 | px3\r
2162\r
2163 stmia r0!, {r4,r5,r6,r8}\r
2164\r
2165 and r4, lr, r7, lsl #1\r
2166 ldrh r4, [r3, r4]\r
2167 and r5, lr, r7, lsr #7\r
2168 ldrh r5, [r3, r5]\r
2169 and r6, lr, r7, lsr #15\r
2170 ldrh r6, [r3, r6]\r
2171\r
2172 and r4, r4, r9\r
2173 add r10,r4, r8, lsr #16\r
2174 mov r10,r10,lsr #1\r
2175 orr r4, r10,r4, lsl #16 @ (px-1+px0)/2 | px0\r
2176\r
2177 and r8 ,lr, r7, lsr #23\r
2178 ldrh r8 ,[r3, r8]\r
2179\r
2180 and r5, r5, r9\r
2181 add r10,r5, r4, lsr #16\r
2182 mov r10,r10,lsr #1\r
2183 orr r5, r10,r5, lsl #16 @ (px0 +px1)/2 | px1\r
2184\r
2185 and r6, r6, r9\r
2186 add r10,r6, r5, lsr #16\r
2187 mov r10,r10,lsr #1\r
2188 orr r6, r10,r6, lsl #16 @ (px1 +px2)/2 | px2\r
2189\r
2190 and r8, r8, r9\r
2191 add r10,r8, r6, lsr #16\r
2192 mov r10,r10,lsr #1\r
2193 orr r8, r10,r8, lsl #16 @ (px2 +px3)/2 | px3\r
2194\r
2195 subs r2, r2, #1\r
2196 stmia r0!, {r4,r5,r6,r8}\r
2197 bne .fl_20loop_bl2\r
2198\r
2199 ldmfd sp!, {r4-r11,pc}\r
2200\r
2201\r
2202#ifdef UNALIGNED_DRAWLINEDEST\r
2203 @ unaligned versions of loops\r
2204 @ warning: starts drawing 2bytes before dst\r
2205\r
2206.fl_RGB555u:\r
2207 sub r0, r0, #2 @ initial adjustment\r
2208 mov r8, #0\r
2209\r
2210.fl_loopRGB555u:\r
2211 ldr r12, [r1], #4\r
2212 ldr r7, [r1], #4\r
2213\r
2214 and r6, lr, r12,lsl #1\r
2215 ldrh r6, [r3, r6]\r
2216 and r5, lr, r12,lsr #7\r
2217 ldrh r5, [r3, r5]\r
2218 orr r4, r8, r6, lsl #16\r
2219\r
2220 and r6, lr, r12,lsr #15\r
2221 ldrh r6, [r3, r6]\r
2222 and r8, lr, r12,lsr #23\r
2223 ldrh r8, [r3, r8]\r
2224 orr r5, r5, r6, lsl #16\r
2225\r
2226 and r6, lr, r7, lsl #1\r
2227 ldrh r6, [r3, r6]\r
2228 and r12,lr, r7, lsr #7\r
2229 ldrh r12,[r3, r12]\r
2230 orr r6, r8, r6, lsl #16\r
2231\r
2232 and r8, lr, r7, lsr #15\r
2233 ldrh r8, [r3, r8]\r
2234 and r7, lr, r7, lsr #23\r
2235\r
2236 subs r2, r2, #1\r
2237 orr r12,r12,r8, lsl #16\r
2238 ldrh r8, [r3, r7]\r
2239\r
2240 stmia r0!, {r4,r5,r6,r12}\r
2241 bne .fl_loopRGB555u\r
2242\r
2243 strh r8, [r0], #2\r
2244\r
2245 ldmfd sp!, {r4-r11,pc}\r
2246\r
2247\r
2248.fl_32scale_RGB555u:\r
2249 sub r0, r0, #2 @ initial adjustment\r
2250 mov r4, #0\r
2251\r
2252 @ r9 f800 07e0 001f | e000 0780 001c | 3800 01e0 0007\r
2253.fl_loop32scale_RGB555u:\r
2254 ldr r12, [r1], #4\r
2255 ldr r7, [r1], #4\r
2256\r
2257 and r6, lr, r12,lsl #1\r
2258 ldrh r6, [r3, r6]\r
2259 and r5, lr, r12,lsr #7\r
2260 ldrh r5, [r3, r5]\r
2261 and r6, r6, r9\r
2262 orr r4, r4, r6, lsl #16 @ r4 = pix_d -1, 0\r
2263\r
2264 and r5, r5, r9\r
2265 sub r8, r5, r5, lsr #2 @ r8 = 3/4 pix_s 1\r
2266 add r6, r8, r6, lsr #2 @ r6 = (1/4 pix_s 0) + (3/4 pix_s 1)\r
2267 orr r5, r6, r5, lsl #15\r
2268\r
2269 and r6, lr, r12,lsr #15\r
2270 ldrh r6, [r3, r6]\r
2271 and r12,lr, r12,lsr #23\r
2272 ldrh r12,[r3, r12]\r
2273 and r6, r6, r9\r
2274 add r5, r5, r6, lsl #15 @ r5 = pix_d 1, 2\r
2275\r
2276 and r8, lr, r7, lsl #1\r
2277 ldrh r8, [r3, r8]\r
2278 and r10,lr, r7, lsr #7\r
2279 ldrh r10,[r3, r10]\r
2280 and r12,r12,r9\r
2281 sub r6, r6, r6, lsr #2 @ r6 = 3/4 pix_s 2\r
2282 add r6, r6, r12,lsr #2\r
2283 orr r6, r6, r12,lsl #16 @ r6 = pix_d 3, 4\r
2284\r
2285 and r8, r8, r9\r
2286 and r10,r10,r9\r
2287 sub r12,r10,r10,lsr #2 @ r12 = 3/4 pix_s 5\r
2288 orr r8, r8, r8, lsl #14\r
2289 add r8, r8, r12,lsl #16 @ r8 = pix_d 5, 6\r
2290 and r12,lr, r7, lsr #15\r
2291 ldrh r12,[r3, r12]\r
2292 and r7, lr, r7, lsr #23\r
2293 ldrh r7, [r3, r7]\r
2294 and r12,r12,r9\r
2295 add r10,r10,r12\r
2296 mov r10,r10, lsr #1\r
2297 sub r12,r12,r12,lsr #2 @ r12 = 3/4 pix_s 6\r
2298 orr r10,r10,r12,lsl #16\r
2299 and r7, r7, r9\r
2300 add r10,r10,r7, lsl #14 @ r10 = pix_d 7, 8\r
2301\r
2302 subs r2, r2, #1\r
2303\r
2304 stmia r0!, {r4,r5,r6,r8,r10}\r
2305 mov r4, r7\r
2306 bne .fl_loop32scale_RGB555u\r
2307\r
2308 strh r4, [r0], #2\r
2309\r
2310 ldmfd sp!, {r4-r11,pc}\r
2311\r
2312#endif /* UNALIGNED_DRAWLINEDEST */\r
2313\r
2314\r
2315@ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\r
2316\r
2317@ utility\r
2318.global blockcpy @ void *dst, void *src, size_t n\r
2319\r
2320blockcpy:\r
2321 stmfd sp!, {r4,r5}\r
2322 cmp r0, r1\r
2323 bhs blockcpyhi\r
2324\r
2325 subs r2, r2, #16\r
2326 blt blockcpy2\r
2327blockcpy_loop:\r
2328 ldmia r1!, {r3-r5,r12}\r
2329 subs r2, r2, #16\r
2330 stmia r0!, {r3-r5,r12}\r
2331 bge blockcpy_loop\r
2332\r
2333blockcpy2:\r
2334 adds r2, r2, #16-4\r
2335 ldmltfd sp!, {r4,r5}\r
2336 bxlt lr\r
2337\r
2338blockcpy_loop2:\r
2339 ldr r3, [r1], #4\r
2340 subs r2, r2, #4\r
2341 str r3, [r0], #4\r
2342 bge blockcpy_loop2\r
2343\r
2344 ldmfd sp!, {r4,r5}\r
2345 bx lr\r
2346\r
2347blockcpyhi:\r
2348 add r0, r0, r2\r
2349 add r1, r1, r2\r
2350\r
2351 subs r2, r2, #16\r
2352 blt blockcpyhi2\r
2353blockcpyhi_loop:\r
2354 ldmdb r1!, {r3-r5,r12}\r
2355 subs r2, r2, #16\r
2356 stmdb r0!, {r3-r5,r12}\r
2357 bge blockcpyhi_loop\r
2358\r
2359blockcpyhi2:\r
2360 adds r2, r2, #16-4\r
2361 ldmltfd sp!, {r4,r5}\r
2362 bxlt lr\r
2363\r
2364blockcpyhi_loop2:\r
2365 ldr r3, [r1, #-4]!\r
2366 subs r2, r2, #4\r
2367 str r3, [r0, #-4]!\r
2368 bge blockcpyhi_loop2\r
2369\r
2370 ldmfd sp!, {r4,r5}\r
2371 bx lr\r
2372\r
2373\r
2374.global blockcpy_or @ void *dst, void *src, size_t n, int pat\r
2375\r
2376blockcpy_or:\r
2377 stmfd sp!, {r4-r6}\r
2378 orr r3, r3, r3, lsl #8\r
2379 orr r3, r3, r3, lsl #16\r
2380 cmp r0, r1\r
2381 bhs blockcpyhi_or\r
2382\r
2383 subs r2, r2, #16\r
2384 blt blockcpy_or2\r
2385blockcpy_loop_or:\r
2386 ldmia r1!, {r4-r6,r12}\r
2387 subs r2, r2, #16\r
2388 orr r4, r4, r3\r
2389 orr r5, r5, r3\r
2390 orr r6, r6, r3\r
2391 orr r12,r12,r3\r
2392 stmia r0!, {r4-r6,r12}\r
2393 bge blockcpy_loop_or\r
2394\r
2395blockcpy_or2:\r
2396 adds r2, r2, #16-4\r
2397 ldmltfd sp!, {r4-r6}\r
2398 bxlt lr\r
2399\r
2400blockcpy_loop_or2:\r
2401 ldr r4, [r1], #4\r
2402 subs r2, r2, #4\r
2403 orr r4, r4, r3\r
2404 str r4, [r0], #4\r
2405 bge blockcpy_loop_or2\r
2406\r
2407 ldmfd sp!, {r4-r6}\r
2408 bx lr\r
2409\r
2410blockcpyhi_or:\r
2411 add r0, r0, r2\r
2412 add r1, r1, r2\r
2413\r
2414 subs r2, r2, #16\r
2415 blt blockcpyhi_or2\r
2416blockcpyhi_loop_or:\r
2417 ldmdb r1!, {r4-r6,r12}\r
2418 subs r2, r2, #16\r
2419 orr r4, r4, r3\r
2420 orr r5, r5, r3\r
2421 orr r6, r6, r3\r
2422 orr r12,r12,r3\r
2423 stmdb r0!, {r4-r6,r12}\r
2424 bge blockcpyhi_loop_or\r
2425\r
2426blockcpyhi_or2:\r
2427 adds r2, r2, #16-4\r
2428 ldmltfd sp!, {r4-r6}\r
2429 bxlt lr\r
2430\r
2431blockcpyhi_loop_or2:\r
2432 ldr r4, [r1, #-4]!\r
2433 subs r2, r2, #4\r
2434 orr r4, r4, r3\r
2435 str r4, [r0, #-4]!\r
2436 bge blockcpyhi_loop_or2\r
2437\r
2438 ldmfd sp!, {r4-r6}\r
2439 bx lr\r
2440\r
2441@ vim:filetype=armasm\r
fast MegaDrive/MegaCD/32X emulator