cff531af |
1 | /*\r |
2 | * assembly optimized versions of most funtions from draw2.c\r |
3 | * (C) notaz, 2006-2008\r |
4 | *\r |
5 | * This work is licensed under the terms of MAME license.\r |
6 | * See COPYING file in the top-level directory.\r |
7 | *\r |
8 | * this is highly specialized, be careful if changing related C code!\r |
9 | */\r |
cc68a136 |
10 | \r |
11 | .extern Pico\r |
e5f426aa |
12 | .extern PicoDraw2FB\r |
cc68a136 |
13 | \r |
14 | @ define these constants in your include file:\r |
15 | @ .equiv START_ROW, 1\r |
16 | @ .equiv END_ROW, 27\r |
17 | @ one row means 8 pixels. If above example was used, (27-1)*8=208 lines would be rendered.\r |
18 | .include "port_config.s"\r |
19 | \r |
20 | \r |
21 | .global BackFillFull @ int reg7\r |
22 | \r |
23 | BackFillFull:\r |
24 | stmfd sp!, {r4-r9,lr}\r |
25 | \r |
e5f426aa |
26 | ldr lr, =PicoDraw2FB @ lr=PicoDraw2FB\r |
ea8c405f |
27 | mov r0, r0, lsl #26\r |
cc68a136 |
28 | ldr lr, [lr]\r |
ea8c405f |
29 | mov r0, r0, lsr #26\r |
cc68a136 |
30 | add lr, lr, #328*8\r |
31 | \r |
cc68a136 |
32 | orr r0, r0, r0, lsl #8\r |
33 | orr r0, r0, r0, lsl #16\r |
34 | \r |
35 | mov r1, r0 @ 25 opcodes wasted?\r |
36 | mov r2, r0\r |
37 | mov r3, r0\r |
38 | mov r4, r0\r |
39 | mov r5, r0\r |
40 | mov r6, r0\r |
41 | mov r7, r0\r |
42 | mov r8, r0\r |
43 | mov r9, r0\r |
44 | \r |
45 | mov r12, #(END_ROW-START_ROW)*8\r |
46 | \r |
47 | @ go go go!\r |
48 | .bff_loop:\r |
49 | add lr, lr, #8\r |
50 | subs r12, r12, #1\r |
51 | \r |
52 | stmia lr!, {r0-r9} @ 10*4*8\r |
53 | stmia lr!, {r0-r9}\r |
54 | stmia lr!, {r0-r9}\r |
55 | stmia lr!, {r0-r9}\r |
56 | stmia lr!, {r0-r9}\r |
57 | stmia lr!, {r0-r9}\r |
58 | stmia lr!, {r0-r9}\r |
59 | stmia lr!, {r0-r9}\r |
60 | \r |
61 | bne .bff_loop\r |
62 | \r |
63 | ldmfd sp!, {r4-r9,r12}\r |
64 | bx r12\r |
65 | \r |
66 | .pool\r |
67 | \r |
68 | @ -------- some macros --------\r |
69 | \r |
70 | \r |
71 | @ helper\r |
72 | @ TileLineSinglecol (r1=pdest, r2=pixels8, r3=pal) r4: scratch, r0: pixels8_old\r |
73 | .macro TileLineSinglecol notsinglecol=0\r |
74 | and r2, r2, #0xf @ #0x0000000f\r |
75 | .if !\notsinglecol\r |
76 | cmp r2, r0, lsr #28 @ if these don't match,\r |
77 | bicne r9, r9, #2 @ it is a sign that whole tile is not singlecolor (only it's lines may be)\r |
78 | .endif\r |
79 | orr r4, r3, r2\r |
80 | orr r4, r4, r4, lsl #8\r |
81 | \r |
82 | tst r1, #1 @ not aligned?\r |
83 | strneb r4, [r1], #1\r |
84 | streqh r4, [r1], #2\r |
85 | strh r4, [r1], #2\r |
86 | strh r4, [r1], #2\r |
87 | strh r4, [r1], #2\r |
88 | strneb r4, [r1], #1 @ have a remaining unaligned pixel?\r |
89 | sub r1, r1, #8\r |
90 | .if !\notsinglecol\r |
91 | mov r0, #0xf\r |
92 | orr r0, r0, r2, lsl #28 @ we will need the old palindex later\r |
93 | .endif\r |
94 | .endm\r |
95 | \r |
96 | @ TileNorm (r1=pdest, r2=pixels8, r3=pal) r0,r4: scratch\r |
97 | .macro TileLineNorm\r |
98 | ands r4, r0, r2, lsr #12 @ #0x0000f000\r |
99 | orrne r4, r3, r4\r |
100 | strneb r4, [r1]\r |
101 | ands r4, r0, r2, lsr #8 @ #0x00000f00\r |
102 | orrne r4, r3, r4\r |
103 | strneb r4, [r1,#1]\r |
104 | ands r4, r0, r2, lsr #4 @ #0x000000f0\r |
105 | orrne r4, r3, r4\r |
106 | strneb r4, [r1,#2]\r |
107 | ands r4, r0, r2 @ #0x0000000f\r |
108 | orrne r4, r3, r4\r |
109 | strneb r4, [r1,#3]\r |
110 | ands r4, r0, r2, lsr #28 @ #0xf0000000\r |
111 | orrne r4, r3, r4\r |
112 | strneb r4, [r1,#4]\r |
113 | ands r4, r0, r2, lsr #24 @ #0x0f000000\r |
114 | orrne r4, r3, r4\r |
115 | strneb r4, [r1,#5]\r |
116 | ands r4, r0, r2, lsr #20 @ #0x00f00000\r |
117 | orrne r4, r3, r4\r |
118 | strneb r4, [r1,#6]\r |
119 | ands r4, r0, r2, lsr #16 @ #0x000f0000\r |
120 | orrne r4, r3, r4\r |
121 | strneb r4, [r1,#7]\r |
122 | .endm\r |
123 | \r |
124 | @ TileFlip (r1=pdest, r2=pixels8, r3=pal) r0,r4: scratch\r |
125 | .macro TileLineFlip\r |
126 | ands r4, r0, r2, lsr #16 @ #0x000f0000\r |
127 | orrne r4, r3, r4\r |
128 | strneb r4, [r1]\r |
129 | ands r4, r0, r2, lsr #20 @ #0x00f00000\r |
130 | orrne r4, r3, r4\r |
131 | strneb r4, [r1,#1]\r |
132 | ands r4, r0, r2, lsr #24 @ #0x0f000000\r |
133 | orrne r4, r3, r4\r |
134 | strneb r4, [r1,#2]\r |
135 | ands r4, r0, r2, lsr #28 @ #0xf0000000\r |
136 | orrne r4, r3, r4\r |
137 | strneb r4, [r1,#3]\r |
138 | ands r4, r0, r2 @ #0x0000000f\r |
139 | orrne r4, r3, r4\r |
140 | strneb r4, [r1,#4]\r |
141 | ands r4, r0, r2, lsr #4 @ #0x000000f0\r |
142 | orrne r4, r3, r4\r |
143 | strneb r4, [r1,#5]\r |
144 | ands r4, r0, r2, lsr #8 @ #0x00000f00\r |
145 | orrne r4, r3, r4\r |
146 | strneb r4, [r1,#6]\r |
147 | ands r4, r0, r2, lsr #12 @ #0x0000f000\r |
148 | orrne r4, r3, r4\r |
149 | strneb r4, [r1,#7]\r |
150 | .endm\r |
151 | \r |
152 | @ Tile (r1=pdest, r3=pal, r9=prevcode, r10=Pico.vram) r2,r4,r7: scratch, r0=0xf\r |
153 | .macro Tile hflip vflip\r |
154 | mov r7, r9, lsl #13 @ r9=code<<8; addr=(code&0x7ff)<<4;\r |
155 | add r7, r10, r7, lsr #16\r |
156 | orr r9, r9, #3 @ emptytile=singlecolor=1, r9 must be <code_16> 00000xxx\r |
157 | .if \vflip\r |
158 | @ we read tilecodes in reverse order if we have vflip\r |
159 | add r7, r7, #8*4\r |
160 | .endif\r |
161 | @ loop through 8 lines\r |
162 | orr r9, r9, #(7<<24)\r |
163 | b 1f @ loop_enter\r |
164 | \r |
165 | 0: @ singlecol_loop\r |
166 | subs r9, r9, #(1<<24)\r |
167 | add r1, r1, #328 @ set pointer to next line\r |
168 | bmi 8f @ loop_exit with r0 restore\r |
169 | 1:\r |
170 | .if \vflip\r |
171 | ldr r2, [r7, #-4]! @ pack=*(unsigned int *)(Pico.vram+addr); // Get 8 pixels\r |
172 | .else\r |
173 | ldr r2, [r7], #4\r |
174 | .endif\r |
175 | tst r2, r2\r |
176 | beq 2f @ empty line\r |
177 | bic r9, r9, #1\r |
178 | cmp r2, r2, ror #4\r |
179 | bne 3f @ not singlecolor\r |
180 | TileLineSinglecol\r |
181 | b 0b\r |
182 | \r |
183 | 2:\r |
184 | bic r9, r9, #2\r |
185 | 2: @ empty_loop\r |
186 | subs r9, r9, #(1<<24)\r |
187 | add r1, r1, #328 @ set pointer to next line\r |
188 | bmi 8f @ loop_exit with r0 restore\r |
189 | .if \vflip\r |
190 | ldr r2, [r7, #-4]! @ next pack\r |
191 | .else\r |
192 | ldr r2, [r7], #4\r |
193 | .endif\r |
194 | mov r0, #0xf @ singlecol_loop might have messed r0\r |
195 | tst r2, r2\r |
196 | beq 2b\r |
197 | \r |
198 | bic r9, r9, #3 @ if we are here, it means we have empty and not empty line\r |
199 | b 5f\r |
200 | \r |
201 | 3: @ not empty, not singlecol\r |
202 | mov r0, #0xf\r |
203 | bic r9, r9, #3\r |
204 | b 6f\r |
205 | \r |
206 | 4: @ not empty, not singlecol loop\r |
207 | subs r9, r9, #(1<<24)\r |
208 | add r1, r1, #328 @ set pointer to next line\r |
209 | bmi 9f @ loop_exit\r |
210 | .if \vflip\r |
211 | ldr r2, [r7, #-4]! @ next pack\r |
212 | .else\r |
213 | ldr r2, [r7], #4\r |
214 | .endif\r |
215 | tst r2, r2\r |
216 | beq 4b @ empty line\r |
217 | 5:\r |
218 | cmp r2, r2, ror #4\r |
219 | beq 7f @ singlecolor line\r |
220 | 6:\r |
221 | .if \hflip\r |
222 | TileLineFlip\r |
223 | .else\r |
224 | TileLineNorm\r |
225 | .endif\r |
226 | b 4b\r |
227 | 7:\r |
228 | TileLineSinglecol 1\r |
229 | b 4b\r |
230 | \r |
231 | 8:\r |
232 | mov r0, #0xf\r |
233 | 9: @ loop_exit\r |
234 | add r9, r9, #(1<<24) @ fix r9\r |
235 | sub r1, r1, #328*8 @ restore pdest pointer\r |
236 | .endm\r |
237 | \r |
238 | \r |
239 | @ TileLineSinglecolAl (r1=pdest, r4,r7=color)\r |
240 | .macro TileLineSinglecolAl0\r |
241 | stmia r1!, {r4,r7}\r |
242 | add r1, r1, #320\r |
243 | .endm\r |
244 | \r |
245 | .macro TileLineSinglecolAl1\r |
246 | strb r4, [r1], #1\r |
247 | strh r4, [r1], #2\r |
248 | str r4, [r1], #4\r |
249 | strb r4, [r1], #1+320\r |
250 | @ add r1, r1, #320\r |
251 | .endm\r |
252 | \r |
253 | .macro TileLineSinglecolAl2\r |
254 | strh r4, [r1], #2\r |
255 | str r4, [r1], #4\r |
256 | strh r4, [r1], #2\r |
257 | add r1, r1, #320\r |
258 | .endm\r |
259 | \r |
260 | .macro TileLineSinglecolAl3\r |
261 | strb r4, [r1], #1\r |
262 | str r4, [r1], #4\r |
263 | strh r4, [r1], #2\r |
264 | strb r4, [r1], #1+320\r |
265 | @ add r1, r1, #320\r |
266 | .endm\r |
267 | \r |
268 | @ TileSinglecol (r1=pdest, r2=pixels8, r3=pal) r4,r7: scratch, r0=0xf\r |
269 | @ kaligned==1, if dest is always aligned\r |
270 | .macro TileSinglecol kaligned=0\r |
271 | and r4, r2, #0xf @ we assume we have good r2 from previous time\r |
272 | orr r4, r4, r3\r |
273 | orr r4, r4, r4, lsl #8\r |
274 | orr r4, r4, r4, lsl #16\r |
275 | mov r7, r4\r |
276 | \r |
277 | .if !\kaligned\r |
278 | tst r1, #2 @ not aligned?\r |
279 | bne 2f\r |
280 | tst r1, #1\r |
281 | bne 1f\r |
282 | .endif\r |
283 | \r |
284 | TileLineSinglecolAl0\r |
285 | TileLineSinglecolAl0\r |
286 | TileLineSinglecolAl0\r |
287 | TileLineSinglecolAl0\r |
288 | TileLineSinglecolAl0\r |
289 | TileLineSinglecolAl0\r |
290 | TileLineSinglecolAl0\r |
291 | TileLineSinglecolAl0\r |
292 | \r |
293 | .if !\kaligned\r |
294 | b 4f\r |
295 | 1:\r |
296 | TileLineSinglecolAl1\r |
297 | TileLineSinglecolAl1\r |
298 | TileLineSinglecolAl1\r |
299 | TileLineSinglecolAl1\r |
300 | TileLineSinglecolAl1\r |
301 | TileLineSinglecolAl1\r |
302 | TileLineSinglecolAl1\r |
303 | TileLineSinglecolAl1\r |
304 | b 4f\r |
305 | \r |
306 | 2:\r |
307 | tst r1, #1\r |
308 | bne 3f\r |
309 | \r |
310 | TileLineSinglecolAl2\r |
311 | TileLineSinglecolAl2\r |
312 | TileLineSinglecolAl2\r |
313 | TileLineSinglecolAl2\r |
314 | TileLineSinglecolAl2\r |
315 | TileLineSinglecolAl2\r |
316 | TileLineSinglecolAl2\r |
317 | TileLineSinglecolAl2\r |
318 | b 4f\r |
319 | \r |
320 | 3:\r |
321 | TileLineSinglecolAl3\r |
322 | TileLineSinglecolAl3\r |
323 | TileLineSinglecolAl3\r |
324 | TileLineSinglecolAl3\r |
325 | TileLineSinglecolAl3\r |
326 | TileLineSinglecolAl3\r |
327 | TileLineSinglecolAl3\r |
328 | TileLineSinglecolAl3\r |
329 | \r |
330 | 4:\r |
331 | .endif\r |
332 | sub r1, r1, #328*8 @ restore pdest pointer\r |
333 | .endm\r |
334 | \r |
335 | \r |
336 | \r |
337 | @ DrawLayerTiles(*hcache, *scrpos, (cells<<24)|(nametab<<9)|(vscroll&0x3ff)<<11|(shift[width]<<8)|planeend, (ymask<<24)|(planestart<<16)|[htab||hscroll]\r |
338 | \r |
339 | @static void DrawLayerFull(int plane, int *hcache, int planestart, int planeend)\r |
340 | \r |
341 | .global DrawLayerFull\r |
342 | \r |
343 | DrawLayerFull:\r |
344 | stmfd sp!, {r4-r11,lr}\r |
345 | \r |
346 | mov r6, r1 @ hcache\r |
347 | \r |
ea8c405f |
348 | ldr r11, =(Pico+0x22228) @ Pico.video\r |
349 | ldr r10, =(Pico+0x10000) @ r10=Pico.vram\r |
350 | ldrb r5, [r11, #13] @ pvid->reg[13]\r |
351 | ldrb r7, [r11, #11]\r |
352 | \r |
cc68a136 |
353 | sub lr, r3, r2\r |
354 | and lr, lr, #0x00ff0000 @ lr=cells\r |
355 | \r |
cc68a136 |
356 | mov r5, r5, lsl #10 @ htab=pvid->reg[13]<<9; (halfwords)\r |
357 | add r5, r5, r0, lsl #1 @ htab+=plane\r |
358 | bic r5, r5, #0x00ff0000 @ just in case\r |
359 | \r |
cc68a136 |
360 | tst r7, #3 @ full screen scroll? (if ==0)\r |
ea8c405f |
361 | ldrb r7, [r11, #16] @ ??hh??ww\r |
cc68a136 |
362 | ldreqh r5, [r10, r5]\r |
363 | biceq r5, r5, #0x0000fc00 @ r5=hscroll (0-0x3ff)\r |
364 | movne r5, r5, lsr #1\r |
365 | orrne r5, r5, #0x8000 @ this marks that we have htab pointer, not hscroll here\r |
366 | \r |
cc68a136 |
367 | and r8, r7, #3\r |
368 | \r |
369 | orr r5, r5, r7, lsl #1+24\r |
370 | orr r5, r5, #0x1f000000\r |
371 | cmp r8, #1\r |
372 | biclt r5, r5, #0x80000000\r |
373 | biceq r5, r5, #0xc0000000\r |
374 | bicgt r5, r5, #0xe0000000\r |
375 | \r |
376 | mov r9, r2, lsl #24\r |
377 | orr r5, r5, r9, lsr #8 @ r5=(ymask<<24)|(trow<<16)|[htab||hscroll]\r |
378 | \r |
379 | add r4, r8, #5\r |
380 | cmp r4, #7\r |
381 | subge r4, r4, #1 @ r4=shift[width] (5,6,6,7)\r |
382 | \r |
383 | orr lr, lr, r4 \r |
384 | orr lr, lr, r3, lsl #24 @ lr=(planeend<<24)|(cells<<16)|shift[width]\r |
385 | \r |
386 | @ calculate xmask:\r |
387 | mov r8, r8, lsl #24+5\r |
388 | orr r8, r8, #0x1f000000\r |
389 | \r |
ea8c405f |
390 | @ Find name table:\r |
cc68a136 |
391 | tst r0, r0\r |
392 | ldreqb r4, [r11, #2]\r |
393 | moveq r4, r4, lsr #3\r |
394 | ldrneb r4, [r11, #4]\r |
395 | and r4, r4, #7\r |
396 | orr lr, lr, r4, lsl #13 @ lr|=nametab_bits{3}<<13\r |
397 | \r |
e5f426aa |
398 | ldr r11, =PicoDraw2FB @ r11=PicoDraw2FB\r |
cc68a136 |
399 | sub r4, r9, #(START_ROW<<24)\r |
ea8c405f |
400 | ldr r11, [r11]\r |
cc68a136 |
401 | mov r4, r4, asr #24\r |
402 | mov r7, #328*8\r |
403 | mla r11, r4, r7, r11 @ scrpos+=8*328*(planestart-START_ROW);\r |
404 | \r |
ea8c405f |
405 | @ Get vertical scroll value:\r |
cc68a136 |
406 | add r7, r10, #0x012000\r |
407 | add r7, r7, #0x000180 @ r7=Pico.vsram (Pico+0x22180)\r |
408 | ldr r7, [r7]\r |
409 | tst r0, r0\r |
410 | moveq r7, r7, lsl #22\r |
411 | movne r7, r7, lsl #6\r |
412 | mov r7, r7, lsr #22 @ r7=vscroll (10 bits)\r |
413 | \r |
414 | orr lr, lr, r7, lsl #3\r |
415 | mov lr, lr, ror #24 @ packed: cccccccc nnnvvvvv vvvvvsss pppppppp: cells, nametab, vscroll, shift[width], planeend\r |
416 | \r |
417 | ands r7, r7, #7\r |
418 | addne lr, lr, #1 @ we have vertically clipped tiles due to vscroll, so we need 1 more row\r |
419 | \r |
420 | rsb r7, r7, #8\r |
421 | str r7, [r6], #4 @ push y-offset to tilecache\r |
422 | mov r4, #328\r |
423 | mla r11, r4, r7, r11 @ scrpos+=(8-(vscroll&7))*328;\r |
424 | \r |
425 | mov r9, #0xff000000 @ r9=(prevcode<<8)|flags: 1~tile empty, 2~tile singlecolor\r |
426 | \r |
427 | .rtrloop_outer:\r |
428 | mov r4, lr, lsl #11\r |
429 | mov r4, r4, lsr #25 @ r4=vscroll>>3 (7 bits)\r |
430 | add r4, r4, r5, lsr #16 @ +trow\r |
431 | and r4, r4, r5, lsr #24 @ &=ymask\r |
432 | mov r7, lr, lsr #8\r |
433 | and r7, r7, #7 @ shift[width]\r |
434 | mov r0, lr, lsr #9\r |
435 | and r0, r0, #0x7000 @ nametab\r |
436 | add r12,r0, r4, lsl r7 @ nametab_row = nametab + (((trow+(vscroll>>3))&ymask)<<shift[width]); \r |
437 | \r |
438 | mov r4, lr, lsr #24\r |
439 | orr r12,r12,r4, lsl #23\r |
440 | mov r12,r12,lsl #1 @ (nametab_row|(cells<<24)) (halfword compliant)\r |
441 | \r |
442 | @ htab?\r |
443 | tst r5, #0x8000\r |
444 | moveq r7, r5, lsl #22 @ hscroll (0-3FFh)\r |
445 | moveq r7, r7, lsr #22\r |
446 | beq .rtr_hscroll_done\r |
447 | \r |
448 | @ get hscroll from htab\r |
449 | mov r7, r5, lsl #17\r |
450 | ands r4, r5, #0x00ff0000\r |
451 | add r7, r7, r4, lsl #5 @ +=trow<<4\r |
452 | andne r4, lr, #0x3800\r |
453 | subne r7, r7, r4, lsl #7 @ if(trow) htaddr-=(vscroll&7)<<1;\r |
454 | mov r7, r7, lsr #16 @ halfwords\r |
455 | ldrh r7, [r10, r7]\r |
456 | \r |
457 | .rtr_hscroll_done:\r |
ea8c405f |
458 | and r8, r8, #0xff000000\r |
cc68a136 |
459 | rsb r4, r7, #0 @ r4=tilex=(-ts->hscroll)>>3\r |
460 | mov r4, r4, asr #3\r |
461 | and r4, r4, #0xff\r |
cc68a136 |
462 | orr r8, r8, r4 @ r8=(xmask<<24)|tilex\r |
463 | \r |
464 | sub r7, r7, #1\r |
465 | and r7, r7, #7\r |
466 | add r7, r7, #1 @ r7=dx=((ts->hscroll-1)&7)+1\r |
467 | \r |
468 | cmp r7, #8\r |
469 | subeq r12,r12, #0x01000000 @ we will loop cells+1 times, so loop less when there is no hscroll\r |
470 | \r |
471 | add r1, r11, r7 @ r1=pdest\r |
472 | mov r0, #0xf\r |
473 | b .rtrloop_enter\r |
474 | \r |
475 | @ r4 & r7 are scratch in this loop\r |
476 | .rtrloop: @ 40-41 times\r |
477 | add r1, r1, #8\r |
478 | subs r12,r12, #0x01000000\r |
479 | add r8, r8, #1\r |
480 | bmi .rtrloop_exit\r |
481 | \r |
482 | .rtrloop_enter:\r |
483 | and r7, r8, r8, lsr #24\r |
484 | add r7, r10, r7, lsl #1\r |
485 | bic r4, r12, #0xff000000 @ Pico.vram[nametab_row+(tilex&xmask)];\r |
486 | ldrh r7, [r7, r4] @ r7=code (int, but from unsigned, no sign extend)\r |
487 | \r |
488 | tst r7, #0x8000\r |
489 | bne .rtr_hiprio\r |
490 | \r |
491 | cmp r7, r9, lsr #8\r |
492 | bne .rtr_notsamecode\r |
493 | @ we know stuff about this tile already\r |
494 | tst r9, #1\r |
495 | bne .rtrloop @ empty tile\r |
496 | tst r9, #2\r |
497 | bne .rtr_singlecolor @ singlecolor tile\r |
498 | b .rtr_samecode\r |
499 | \r |
500 | .rtr_notsamecode:\r |
501 | and r4, r9, #0x600000\r |
502 | mov r9, r7, lsl #8 @ remember new code\r |
503 | \r |
504 | @ update cram\r |
505 | and r7, r7, #0x6000\r |
506 | mov r3, r7, asr #9 @ r3=pal=((code&0x6000)>>9);\r |
507 | \r |
508 | .rtr_samecode:\r |
509 | tst r9, #0x100000 @ vflip?\r |
510 | bne .rtr_vflip\r |
511 | \r |
512 | tst r9, #0x080000 @ hflip?\r |
513 | bne .rtr_hflip\r |
514 | \r |
515 | @ Tile (r1=pdest, r3=pal, r9=prevcode, r10=Pico.vram) r2,r4,r7: scratch, r0=0xf\r |
516 | Tile 0, 0\r |
517 | b .rtrloop\r |
518 | \r |
519 | .rtr_hflip:\r |
520 | Tile 1, 0\r |
521 | b .rtrloop\r |
522 | \r |
523 | .rtr_vflip:\r |
524 | tst r9, #0x080000 @ hflip?\r |
525 | bne .rtr_vflip_hflip\r |
526 | \r |
527 | Tile 0, 1\r |
528 | b .rtrloop\r |
529 | \r |
530 | .rtr_vflip_hflip:\r |
531 | Tile 1, 1\r |
532 | b .rtrloop\r |
533 | \r |
534 | .rtr_singlecolor:\r |
535 | TileSinglecol\r |
536 | b .rtrloop\r |
537 | \r |
538 | .rtr_hiprio:\r |
539 | @ *(*hcache)++ = code|(dx<<16)|(trow<<27);\r |
540 | sub r4, r1, r11\r |
541 | orr r7, r7, r4, lsl #16\r |
542 | and r4, r5, #0x00ff0000\r |
543 | orr r7, r7, r4, lsl #11 @ (trow<<27)\r |
544 | str r7, [r6], #4 @ cache hi priority tile\r |
545 | b .rtrloop\r |
546 | \r |
547 | .rtrloop_exit:\r |
548 | add r5, r5, #0x00010000\r |
549 | mov r4, r5, lsl #8\r |
550 | cmp r4, lr, lsl #24\r |
551 | bge .rtrloop_outer_exit\r |
552 | add r11, r11, #328*8\r |
553 | b .rtrloop_outer\r |
554 | \r |
555 | .rtrloop_outer_exit:\r |
556 | \r |
557 | @ terminate cache list\r |
558 | mov r0, #0\r |
559 | str r0, [r6] @ save cache pointer\r |
560 | \r |
561 | ldmfd sp!, {r4-r11,lr}\r |
562 | bx lr\r |
563 | \r |
564 | .pool\r |
565 | \r |
566 | \r |
567 | \r |
568 | .global DrawTilesFromCacheF @ int *hc\r |
569 | \r |
570 | DrawTilesFromCacheF:\r |
571 | stmfd sp!, {r4-r10,lr}\r |
572 | \r |
573 | mov r9, #0xff000000 @ r9=prevcode=-1\r |
574 | mvn r6, #0 @ r6=prevy=-1\r |
575 | \r |
e5f426aa |
576 | ldr r4, =PicoDraw2FB @ r4=PicoDraw2FB\r |
cc68a136 |
577 | ldr r1, [r0], #4 @ read y offset\r |
ea8c405f |
578 | ldr r4, [r4]\r |
cc68a136 |
579 | mov r7, #328\r |
580 | mla r1, r7, r1, r4\r |
581 | sub r12, r1, #(328*8*START_ROW) @ r12=scrpos\r |
582 | \r |
583 | ldr r10, =(Pico+0x10000) @ r10=Pico.vram\r |
584 | mov r8, r0 @ hc\r |
585 | mov r0, #0xf\r |
586 | \r |
587 | @ scratch: r4, r7\r |
588 | @ *hcache++ = code|(dx<<16)|(trow<<27); // cache it\r |
589 | \r |
590 | .dtfcf_loop:\r |
591 | ldr r7, [r8], #4 @ read code\r |
592 | movs r1, r7, lsr #16 @ r1=dx;\r |
593 | ldmeqfd sp!, {r4-r10,pc} @ dx is never zero, this must be a terminator, return\r |
594 | \r |
ea8c405f |
595 | @ row changed?\r |
cc68a136 |
596 | cmp r6, r7, lsr #27\r |
597 | movne r6, r7, lsr #27\r |
598 | movne r4, #328*8\r |
599 | mlane r5, r4, r6, r12 @ r5=pd = scrpos + prevy*328*8\r |
600 | \r |
601 | bic r1, r1, #0xf800\r |
602 | add r1, r5, r1 @ r1=pdest (halfwords)\r |
603 | \r |
604 | mov r7, r7, lsl #16\r |
605 | mov r7, r7, lsr #16\r |
606 | \r |
607 | cmp r7, r9, lsr #8\r |
608 | bne .dtfcf_notsamecode\r |
609 | @ we know stuff about this tile already\r |
610 | tst r9, #1\r |
611 | bne .dtfcf_loop @ empty tile\r |
612 | tst r9, #2\r |
613 | bne .dtfcf_singlecolor @ singlecolor tile\r |
614 | b .dtfcf_samecode\r |
615 | \r |
616 | .dtfcf_notsamecode:\r |
617 | and r4, r9, #0x600000\r |
618 | mov r9, r7, lsl #8 @ remember new code\r |
619 | \r |
620 | @ update cram val\r |
621 | and r7, r7, #0x6000\r |
622 | mov r3, r7, asr #9 @ r3=pal=((code&0x6000)>>9);\r |
623 | \r |
624 | \r |
625 | .dtfcf_samecode:\r |
626 | \r |
627 | tst r9, #0x100000 @ vflip?\r |
628 | bne .dtfcf_vflip\r |
629 | \r |
630 | tst r9, #0x080000 @ hflip?\r |
631 | bne .dtfcf_hflip\r |
632 | \r |
633 | @ Tile (r1=pdest, r3=pal, r9=prevcode, r10=Pico.vram) r2,r4,r7: scratch, r0=0xf\r |
634 | Tile 0, 0\r |
635 | b .dtfcf_loop\r |
636 | \r |
637 | .dtfcf_hflip:\r |
638 | Tile 1, 0\r |
639 | b .dtfcf_loop\r |
640 | \r |
641 | .dtfcf_vflip:\r |
642 | tst r9, #0x080000 @ hflip?\r |
643 | bne .dtfcf_vflip_hflip\r |
644 | \r |
645 | Tile 0, 1\r |
646 | b .dtfcf_loop\r |
647 | \r |
648 | .dtfcf_vflip_hflip:\r |
649 | Tile 1, 1\r |
650 | b .dtfcf_loop\r |
651 | \r |
652 | .dtfcf_singlecolor:\r |
653 | TileSinglecol\r |
654 | b .dtfcf_loop\r |
655 | \r |
656 | .pool\r |
657 | \r |
658 | \r |
659 | @ @@@@@@@@@@@@@@@\r |
660 | \r |
661 | @ (tile_start<<16)|row_start\r |
662 | .global DrawWindowFull @ int tstart, int tend, int prio\r |
663 | \r |
664 | DrawWindowFull:\r |
665 | stmfd sp!, {r4-r11,lr}\r |
666 | \r |
667 | ldr r11, =(Pico+0x22228) @ Pico.video\r |
668 | ldrb r12, [r11, #3] @ pvid->reg[3]\r |
669 | mov r12, r12, lsl #10\r |
670 | \r |
671 | ldr r4, [r11, #12]\r |
672 | mov r5, #1 @ nametab_step\r |
673 | tst r4, #1 @ 40 cell mode?\r |
674 | andne r12, r12, #0xf000 @ 0x3c<<10\r |
675 | andeq r12, r12, #0xf800\r |
676 | movne r5, r5, lsl #7\r |
677 | moveq r5, r5, lsl #6 @ nametab_step\r |
678 | \r |
679 | and r4, r0, #0xff\r |
680 | mla r12, r5, r4, r12 @ nametab += nametab_step*start;\r |
681 | \r |
682 | mov r4, r0, lsr #16 @ r4=start_cell_h\r |
683 | add r7, r12, r4, lsl #1\r |
684 | \r |
685 | @ fetch the first code now\r |
686 | ldr r10, =(Pico+0x10000) @ lr=Pico.vram\r |
687 | ldrh r7, [r10, r7]\r |
688 | cmp r2, r7, lsr #15\r |
689 | ldmnefd sp!, {r4-r11,pc} @ hack: simply assume that whole window uses same priority\r |
690 | \r |
691 | rsb r8, r4, r1, lsr #16 @ cells (h)\r |
692 | orr r8, r8, r4, lsl #8\r |
693 | mov r4, r1, lsl #24\r |
694 | sub r4, r4, r0, lsl #24\r |
695 | orr r8, r8, r4, lsr #8 @ r8=cells_h|(start_cell_h<<8)|(cells_v<<16)\r |
696 | sub r8, r8, #0x010000 @ adjust for algo\r |
697 | \r |
698 | mov r9, #0xff000000 @ r9=prevcode=-1\r |
699 | \r |
e5f426aa |
700 | ldr r11, =PicoDraw2FB @ r11=scrpos\r |
ea8c405f |
701 | and r4, r0, #0xff\r |
cc68a136 |
702 | ldr r11, [r11]\r |
ea8c405f |
703 | sub r4, r4, #START_ROW\r |
cc68a136 |
704 | add r11, r11, #328*8\r |
705 | add r11, r11, #8\r |
706 | \r |
cc68a136 |
707 | mov r7, #328*8\r |
708 | mla r11, r7, r4, r11 @ scrpos+=8*328*(start-START_ROW);\r |
709 | mov r0, #0xf\r |
710 | \r |
711 | .dwfloop_outer:\r |
712 | and r6, r8, #0xff00 @ r6=tilex\r |
713 | add r1, r11, r6, lsr #5 @ r1=pdest\r |
714 | add r6, r12, r6, lsr #7\r |
715 | add r6, r10, r6 @ r6=Pico.vram+nametab+tilex\r |
716 | orr r8, r8, r8, lsl #24\r |
717 | sub r8, r8, #0x01000000 @ cell loop counter\r |
718 | b .dwfloop_enter\r |
719 | \r |
720 | @ r4 & r7 are scratch in this loop\r |
721 | .dwfloop:\r |
722 | add r1, r1, #8\r |
723 | subs r8, r8, #0x01000000\r |
724 | bmi .dwfloop_exit\r |
725 | \r |
726 | .dwfloop_enter:\r |
727 | ldrh r7, [r6], #2 @ r7=code\r |
728 | \r |
729 | cmp r7, r9, lsr #8\r |
730 | bne .dwf_notsamecode\r |
731 | @ we know stuff about this tile already\r |
732 | tst r9, #1\r |
733 | bne .dwfloop @ empty tile\r |
734 | tst r9, #2\r |
735 | bne .dwf_singlecolor @ singlecolor tile\r |
736 | b .dwf_samecode\r |
737 | \r |
738 | .dwf_notsamecode:\r |
739 | and r4, r9, #0x600000\r |
740 | mov r9, r7, lsl #8 @ remember new code\r |
741 | \r |
742 | @ update cram val\r |
743 | and r7, r7, #0x6000\r |
744 | mov r3, r7, asr #9 @ r3=pal=((code&0x6000)>>9);\r |
745 | \r |
746 | .dwf_samecode:\r |
747 | \r |
748 | tst r9, #0x100000 @ vflip?\r |
749 | bne .dwf_vflip\r |
750 | \r |
751 | tst r9, #0x080000 @ hflip?\r |
752 | bne .dwf_hflip\r |
753 | \r |
754 | @ Tile (r1=pdest, r3=pal, r9=prevcode, r10=Pico.vram) r2,r4,r7: scratch, r0=0xf\r |
755 | Tile 0, 0\r |
756 | b .dwfloop\r |
757 | \r |
758 | .dwf_hflip:\r |
759 | Tile 1, 0\r |
760 | b .dwfloop\r |
761 | \r |
762 | .dwf_vflip:\r |
763 | tst r9, #0x080000 @ hflip?\r |
764 | bne .dwf_vflip_hflip\r |
765 | \r |
766 | Tile 0, 1\r |
767 | b .dwfloop\r |
768 | \r |
769 | .dwf_vflip_hflip:\r |
770 | Tile 1, 1\r |
771 | b .dwfloop\r |
772 | \r |
773 | .dwf_singlecolor:\r |
774 | TileSinglecol 1\r |
775 | b .dwfloop\r |
776 | \r |
777 | .dwfloop_exit:\r |
778 | bic r8, r8, #0xff000000 @ fix r8\r |
779 | subs r8, r8, #0x010000\r |
780 | ldmmifd sp!, {r4-r11,pc}\r |
781 | add r11, r11, #328*8\r |
782 | add r12, r12, r5 @ nametab+=nametab_step\r |
783 | b .dwfloop_outer\r |
784 | \r |
785 | .pool\r |
786 | \r |
787 | \r |
788 | @ ---------------- sprites ---------------\r |
789 | \r |
790 | .macro SpriteLoop hflip vflip\r |
791 | .if \vflip\r |
792 | mov r1, r5, lsr #24 @ height\r |
793 | mov r0, #328*8\r |
794 | mla r11, r1, r0, r11 @ scrpos+=height*328*8;\r |
795 | add r12, r12, r1, lsl #3 @ sy+=height*8\r |
796 | .endif\r |
797 | mov r0, #0xf\r |
798 | .if \hflip\r |
799 | and r1, r5, #0xff\r |
800 | add r8, r8, r1, lsl #3 @ sx+=width*8\r |
801 | 58:\r |
802 | cmp r8, #336\r |
803 | blt 51f\r |
804 | add r9, r9, r5, lsr #16\r |
805 | sub r5, r5, #1 @ sub width\r |
806 | sub r8, r8, #8\r |
807 | b 58b\r |
808 | .else\r |
809 | cmp r8, #0 @ skip tiles hidden on the left of screen\r |
810 | bgt 51f\r |
811 | 58:\r |
812 | add r9, r9, r5, lsr #16\r |
813 | sub r5, r5, #1\r |
814 | adds r8, r8, #8\r |
815 | ble 58b\r |
816 | b 51f\r |
817 | .endif\r |
818 | \r |
819 | 50: @ outer\r |
820 | .if !\hflip\r |
821 | add r8, r8, #8 @ sx+=8\r |
822 | .endif\r |
823 | bic r5, r5, #0xff000000 @ fix height\r |
824 | orr r5, r5, r5, lsl #16\r |
825 | \r |
826 | 51: @ outer_enter\r |
827 | sub r5, r5, #1 @ width--\r |
828 | movs r1, r5, lsl #24\r |
829 | ldmmifd sp!, {r4-r11,pc} @ end of tile\r |
830 | .if \hflip\r |
831 | subs r8, r8, #8 @ sx-=8\r |
832 | ldmlefd sp!, {r4-r11,pc} @ tile offscreen\r |
833 | .else\r |
834 | cmp r8, #328\r |
835 | ldmgefd sp!, {r4-r11,pc} @ tile offscreen\r |
836 | .endif\r |
837 | mov r6, r12 @ r6=sy\r |
838 | add r1, r11, r8 @ pdest=scrpos+sx\r |
839 | b 53f\r |
840 | \r |
841 | 52: @ inner\r |
842 | add r9, r9, #1<<8 @ tile++\r |
843 | .if !\vflip\r |
844 | add r6, r6, #8 @ sy+=8\r |
845 | add r1, r1, #328*8\r |
846 | .endif\r |
847 | \r |
848 | 53: @ inner_enter\r |
849 | @ end of sprite?\r |
850 | subs r5, r5, #0x01000000\r |
851 | bmi 50b @ ->outer\r |
852 | .if \vflip\r |
853 | sub r6, r6, #8 @ sy-=8\r |
854 | sub r1, r1, #328*8\r |
855 | .endif\r |
856 | \r |
857 | @ offscreen?\r |
858 | cmp r6, #(START_ROW*8)\r |
859 | ble 52b\r |
860 | \r |
861 | cmp r6, #(END_ROW*8+8)\r |
862 | bge 52b\r |
863 | \r |
864 | @ Tile (r1=pdest, r3=pal, r9=prevcode, r10=Pico.vram) r2,r4,r7: scratch, r0=0xf\r |
865 | Tile \hflip, \vflip\r |
866 | b 52b\r |
867 | .endm\r |
868 | \r |
869 | \r |
870 | .global DrawSpriteFull @ unsigned int *sprite\r |
871 | \r |
872 | DrawSpriteFull:\r |
873 | stmfd sp!, {r4-r11,lr}\r |
874 | \r |
875 | ldr r3, [r0] @ sprite[0]\r |
876 | mov r5, r3, lsl #4\r |
877 | mov r6, r5, lsr #30\r |
878 | add r6, r6, #1 @ r6=width\r |
879 | mov r5, r5, lsl #2\r |
880 | mov r5, r5, lsr #30\r |
881 | add r5, r5, #1 @ r5=height\r |
882 | \r |
883 | mov r12, r3, lsl #23\r |
884 | mov r12, r12, lsr #23\r |
cc68a136 |
885 | \r |
886 | ldr lr, [r0, #4] @ lr=code\r |
ea8c405f |
887 | sub r12, r12, #0x78 @ r12=sy\r |
cc68a136 |
888 | mov r8, lr, lsl #7\r |
889 | mov r8, r8, lsr #23\r |
890 | sub r8, r8, #0x78 @ r8=sx\r |
891 | \r |
892 | mov r9, lr, lsl #21\r |
893 | mov r9, r9, lsr #13 @ r9=tile<<8\r |
894 | \r |
895 | and r3, lr, #0x6000\r |
896 | mov r3, r3, lsr #9 @ r3=pal=((code>>9)&0x30);\r |
897 | \r |
e5f426aa |
898 | ldr r11, =PicoDraw2FB @ r11=scrpos\r |
ea8c405f |
899 | ldr r10, =(Pico+0x10000) @ r10=Pico.vram\r |
cc68a136 |
900 | ldr r11, [r11]\r |
901 | sub r1, r12, #(START_ROW*8)\r |
902 | mov r0, #328\r |
903 | mla r11, r1, r0, r11 @ scrpos+=(sy-START_ROW*8)*328;\r |
904 | \r |
905 | orr r5, r5, r5, lsl #16 @\r |
906 | orr r5, r6, r5, lsl #8 @ r5=width|(height<<8)|(height<<24)\r |
907 | \r |
908 | tst lr, #0x1000 @ vflip?\r |
909 | bne .dsf_vflip\r |
910 | \r |
911 | tst lr, #0x0800 @ hflip?\r |
912 | bne .dsf_hflip\r |
913 | \r |
914 | SpriteLoop 0, 0\r |
915 | \r |
916 | .dsf_hflip:\r |
917 | SpriteLoop 1, 0\r |
918 | \r |
919 | .dsf_vflip:\r |
920 | tst lr, #0x0800 @ hflip?\r |
921 | bne .dsf_vflip_hflip\r |
922 | \r |
923 | SpriteLoop 0, 1\r |
924 | \r |
925 | .dsf_vflip_hflip:\r |
926 | SpriteLoop 1, 1\r |
927 | \r |
928 | .pool\r |
929 | \r |
cff531af |
930 | @ vim:filetype=armasm\r |