src/mame/video/rdptri.c
| r0 | r6716 | |
| 1 | static void triangle_ns_nt_nz(UINT32 w1, UINT32 w2) |
| 2 | { |
| 3 | int j; |
| 4 | int xleft, xright, xleft_inc, xright_inc; |
| 5 | int xstart, xend; |
| 6 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 7 | |
| 8 | INT32 yl, ym, yh; |
| 9 | INT32 xl, xm, xh; |
| 10 | INT32 dxldy, dxhdy, dxmdy; |
| 11 | UINT32 w3, w4, w5, w6, w7, w8; |
| 12 | |
| 13 | int k = 0; |
| 14 | |
| 15 | INT32 limcvg = 0; |
| 16 | INT32 startcvg = 0; |
| 17 | |
| 18 | int sign_dxldy = 0; |
| 19 | int sign_dxmdy = 0; |
| 20 | int samesign = 0; |
| 21 | |
| 22 | int sign_dxhdy = 0; |
| 23 | |
| 24 | int xfrac = 0; |
| 25 | |
| 26 | int m_inc; |
| 27 | UINT32 min=0, max=3; |
| 28 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 29 | |
| 30 | int spix = 0; // Current subpixel |
| 31 | int ycur; |
| 32 | int ylfar; |
| 33 | int ldflag; |
| 34 | int yhpix; |
| 35 | int ympix; |
| 36 | int ylpix; |
| 37 | |
| 38 | #define addleft(x) addleftcvg(x,k) |
| 39 | #define addright(x) addrightcvg(x,k) |
| 40 | |
| 41 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 42 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 43 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 44 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 45 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 46 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 47 | |
| 48 | yl = (w1 & 0x3fff); |
| 49 | ym = ((w2 >> 16) & 0x3fff); |
| 50 | yh = ((w2 >> 0) & 0x3fff); |
| 51 | xl = (INT32)(w3 & 0x3fffffff); |
| 52 | xh = (INT32)(w5 & 0x3fffffff); |
| 53 | xm = (INT32)(w7 & 0x3fffffff); |
| 54 | // Inverse slopes in 16.16 format |
| 55 | dxldy = (INT32)(w4); |
| 56 | dxhdy = (INT32)(w6); |
| 57 | dxmdy = (INT32)(w8); |
| 58 | |
| 59 | max_level = ((w1 >> 19) & 7); |
| 60 | |
| 61 | if (yl & 0x2000) yl |= 0xffffc000; |
| 62 | if (ym & 0x2000) ym |= 0xffffc000; |
| 63 | if (yh & 0x2000) yh |= 0xffffc000; |
| 64 | |
| 65 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 66 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 67 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 68 | |
| 69 | span[0].dymax = 0; |
| 70 | |
| 71 | xleft_inc = dxmdy >> 2; |
| 72 | xright_inc = dxhdy >> 2; |
| 73 | |
| 74 | xright = xh; |
| 75 | xleft = xm; |
| 76 | |
| 77 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 78 | if (limcvg < 0) |
| 79 | { |
| 80 | limcvg = 0; |
| 81 | } |
| 82 | |
| 83 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 84 | for (k = startcvg; k <= limcvg; k++) |
| 85 | { |
| 86 | memset((void*)&span[k].cvg[0],0,640); |
| 87 | } |
| 88 | |
| 89 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 90 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 91 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 92 | |
| 93 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 94 | |
| 95 | m_inc = flip ? 1 : -1; |
| 96 | |
| 97 | ycur = yh & ~3; |
| 98 | ylfar = yl | 3; |
| 99 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 100 | yhpix = yh >> 2; |
| 101 | ympix = ym >> 2; |
| 102 | ylpix = yl >> 2; |
| 103 | |
| 104 | for (k = ycur; k <= ylfar; k++) |
| 105 | { |
| 106 | if (k == ym) |
| 107 | { |
| 108 | xleft = xl; |
| 109 | xleft_inc = dxldy >> 2; |
| 110 | } |
| 111 | |
| 112 | xstart = xleft >> 16; |
| 113 | xend = xright >> 16; |
| 114 | j = k >> 2; |
| 115 | spix = k & 3; |
| 116 | |
| 117 | if (k >= 0 && k < 0x1000) |
| 118 | { |
| 119 | int m = 0; |
| 120 | int n = 0; |
| 121 | int length = 0; |
| 122 | min = 0; max = 3; |
| 123 | if (j == yhpix) |
| 124 | { |
| 125 | min = yh & 3; |
| 126 | } |
| 127 | if (j == ylpix) |
| 128 | { |
| 129 | max = yl & 3; |
| 130 | } |
| 131 | if (spix >= min && spix <= max) |
| 132 | { |
| 133 | if (spix == min) |
| 134 | { |
| 135 | minxmx = maxxmx = xstart; |
| 136 | minxhx = maxxhx = xend; |
| 137 | } |
| 138 | else |
| 139 | { |
| 140 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 141 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 142 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 143 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 144 | } |
| 145 | } |
| 146 | |
| 147 | if (spix == max) |
| 148 | { |
| 149 | if (flip) |
| 150 | { |
| 151 | span[j].lx = maxxmx; |
| 152 | span[j].rx = minxhx; |
| 153 | } |
| 154 | else |
| 155 | { |
| 156 | span[j].lx = minxmx; |
| 157 | span[j].rx = maxxhx; |
| 158 | } |
| 159 | } |
| 160 | |
| 161 | length = flip ? (xstart - xend) : (xend - xstart); |
| 162 | |
| 163 | if (spix == ldflag) |
| 164 | { |
| 165 | xfrac = ((xright >> 8) & 0xff); |
| 166 | } |
| 167 | |
| 168 | m = flip ? (xend+1) : (xend-1); |
| 169 | |
| 170 | if (k >= yh && length >= 0 && k <= yl) |
| 171 | { |
| 172 | if (xstart>=0 && xstart <1024) |
| 173 | { |
| 174 | if (!flip) |
| 175 | { |
| 176 | span[j].cvg[xstart] += addleft(xleft); |
| 177 | } |
| 178 | else |
| 179 | { |
| 180 | span[j].cvg[xstart] += addright(xleft); |
| 181 | } |
| 182 | } |
| 183 | if (xend>=0 && xend<1024) |
| 184 | { |
| 185 | if (xstart != xend) |
| 186 | { |
| 187 | if (!flip) |
| 188 | { |
| 189 | span[j].cvg[xend] += addright(xright); |
| 190 | } |
| 191 | else |
| 192 | { |
| 193 | span[j].cvg[xend] += addleft(xright); |
| 194 | } |
| 195 | } |
| 196 | else |
| 197 | { |
| 198 | if (!flip) |
| 199 | { |
| 200 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 201 | } |
| 202 | else |
| 203 | { |
| 204 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 205 | } |
| 206 | if (span[j].cvg[xend] > 200) |
| 207 | { |
| 208 | span[j].cvg[xend] = 0; |
| 209 | } |
| 210 | } |
| 211 | } |
| 212 | for (n = 0; n < (length - 1); n++) |
| 213 | { |
| 214 | if (m>=0 && m < 640) |
| 215 | { |
| 216 | span[j].cvg[m] += 2; |
| 217 | } |
| 218 | |
| 219 | m += m_inc; |
| 220 | } |
| 221 | } |
| 222 | } |
| 223 | |
| 224 | xleft += xleft_inc; |
| 225 | xright += xright_inc; |
| 226 | } |
| 227 | |
| 228 | switch (fb_size) // 8bpp needs to be implemented |
| 229 | { |
| 230 | case PIXEL_SIZE_16BIT: |
| 231 | if(flip) |
| 232 | { |
| 233 | render_spans_16_ns_nt_nz_f(yh>>2, yl>>2, &tile[0]); break; |
| 234 | } |
| 235 | else |
| 236 | { |
| 237 | render_spans_16_ns_nt_nz_nf(yh>>2, yl>>2, &tile[0]); break; |
| 238 | } |
| 239 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[0], 0, 0, 0, flip); break; |
| 240 | default: break; // V-Rally2 does this, fb_size=0 |
| 241 | } |
| 242 | } |
| 243 | |
| 244 | static void triangle_ns_nt_z(UINT32 w1, UINT32 w2) |
| 245 | { |
| 246 | int j; |
| 247 | int xleft, xright, xleft_inc, xright_inc; |
| 248 | int xstart, xend; |
| 249 | int z = 0; |
| 250 | int dzdx = 0; |
| 251 | int dzdy = 0; |
| 252 | int dzde = 0; |
| 253 | int tilenum; |
| 254 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 255 | |
| 256 | INT32 yl, ym, yh; |
| 257 | INT32 xl, xm, xh; |
| 258 | INT32 dxldy, dxhdy, dxmdy; |
| 259 | int dzdy_dz, dzdx_dz; |
| 260 | UINT32 w3, w4, w5, w6, w7, w8; |
| 261 | |
| 262 | int k = 0; |
| 263 | |
| 264 | INT32 limcvg = 0; |
| 265 | INT32 startcvg = 0; |
| 266 | |
| 267 | int sign_dxldy = 0; |
| 268 | int sign_dxmdy = 0; |
| 269 | int samesign = 0; |
| 270 | |
| 271 | int dzdiff = 0; |
| 272 | int sign_dxhdy = 0; |
| 273 | |
| 274 | int dzdeh = 0, dzdyh = 0; |
| 275 | int do_offset = 0; |
| 276 | |
| 277 | int xfrac = 0; |
| 278 | int dzeoff = 0; |
| 279 | |
| 280 | int dzdxh = 0; |
| 281 | |
| 282 | int m_inc; |
| 283 | UINT32 min=0, max=3; |
| 284 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 285 | |
| 286 | int spix = 0; // Current subpixel |
| 287 | int ycur; |
| 288 | int ylfar; |
| 289 | int ldflag; |
| 290 | int yhpix; |
| 291 | int ympix; |
| 292 | int ylpix; |
| 293 | |
| 294 | int zbuffer_base = rdp_cmd_cur + 8; |
| 295 | |
| 296 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 297 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 298 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 299 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 300 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 301 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 302 | |
| 303 | yl = (w1 & 0x3fff); |
| 304 | ym = ((w2 >> 16) & 0x3fff); |
| 305 | yh = ((w2 >> 0) & 0x3fff); |
| 306 | xl = (INT32)(w3 & 0x3fffffff); |
| 307 | xh = (INT32)(w5 & 0x3fffffff); |
| 308 | xm = (INT32)(w7 & 0x3fffffff); |
| 309 | // Inverse slopes in 16.16 format |
| 310 | dxldy = (INT32)(w4); |
| 311 | dxhdy = (INT32)(w6); |
| 312 | dxmdy = (INT32)(w8); |
| 313 | |
| 314 | max_level = ((w1 >> 19) & 7); |
| 315 | tilenum = (w1 >> 16) & 0x7; |
| 316 | |
| 317 | if (yl & 0x2000) yl |= 0xffffc000; |
| 318 | if (ym & 0x2000) ym |= 0xffffc000; |
| 319 | if (yh & 0x2000) yh |= 0xffffc000; |
| 320 | |
| 321 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 322 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 323 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 324 | |
| 325 | z = rdp_cmd_data[zbuffer_base+0]; |
| 326 | dzdx = rdp_cmd_data[zbuffer_base+1]; |
| 327 | dzde = rdp_cmd_data[zbuffer_base+2]; |
| 328 | dzdy = rdp_cmd_data[zbuffer_base+3]; |
| 329 | |
| 330 | span[0].dz.w = dzdx; |
| 331 | dzdy_dz = (dzdy >> 16) & 0xffff; |
| 332 | dzdx_dz = (dzdx >> 16) & 0xffff; |
| 333 | span[0].dzpix = ((dzdy_dz & 0x8000) ? ((~dzdy_dz) & 0x7fff) : dzdy_dz) + ((dzdx_dz & 0x8000) ? ((~dzdx_dz) & 0x7fff) : dzdx_dz); |
| 334 | span[0].dzpix = normalize_dzpix(span[0].dzpix); |
| 335 | |
| 336 | xleft_inc = dxmdy >> 2; |
| 337 | xright_inc = dxhdy >> 2; |
| 338 | |
| 339 | xright = xh; |
| 340 | xleft = xm; |
| 341 | |
| 342 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 343 | if (limcvg < 0) |
| 344 | { |
| 345 | limcvg = 0; |
| 346 | } |
| 347 | |
| 348 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 349 | for (k = startcvg; k <= limcvg; k++) |
| 350 | { |
| 351 | memset((void*)&span[k].cvg[0],0,640); |
| 352 | } |
| 353 | |
| 354 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 355 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 356 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 357 | |
| 358 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 359 | |
| 360 | do_offset = !(sign_dxhdy ^ (flip)); |
| 361 | |
| 362 | if (do_offset) |
| 363 | { |
| 364 | dzdeh = dzde >> 9; dzdyh = dzdy >> 9; |
| 365 | |
| 366 | dzdiff = (dzdeh*3 - dzdyh*3) << 7; |
| 367 | } |
| 368 | else |
| 369 | { |
| 370 | dzdiff = 0; |
| 371 | } |
| 372 | |
| 373 | if (do_offset) |
| 374 | { |
| 375 | dzeoff = (dzdeh*3) << 7; |
| 376 | } |
| 377 | else |
| 378 | { |
| 379 | dzeoff = 0; |
| 380 | } |
| 381 | |
| 382 | #define addleft(x) addleftcvg(x,k) |
| 383 | #define addright(x) addrightcvg(x,k) |
| 384 | #define setvalues() { \ |
| 385 | addvalues(); \ |
| 386 | adjust_attr(); \ |
| 387 | } |
| 388 | |
| 389 | dzdxh = dzdx >> 8; |
| 390 | |
| 391 | #define adjust_attr() \ |
| 392 | { \ |
| 393 | span[j].z.w = z + dzdiff - (xfrac * dzdxh); \ |
| 394 | } |
| 395 | |
| 396 | #define addvalues() { \ |
| 397 | z += dzde; \ |
| 398 | } |
| 399 | |
| 400 | m_inc = flip ? 1 : -1; |
| 401 | |
| 402 | ycur = yh & ~3; |
| 403 | ylfar = yl | 3; |
| 404 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 405 | yhpix = yh >> 2; |
| 406 | ympix = ym >> 2; |
| 407 | ylpix = yl >> 2; |
| 408 | |
| 409 | for (k = ycur; k <= ylfar; k++) |
| 410 | { |
| 411 | if (k == ym) |
| 412 | { |
| 413 | xleft = xl; |
| 414 | xleft_inc = dxldy >> 2; |
| 415 | } |
| 416 | |
| 417 | xstart = xleft >> 16; |
| 418 | xend = xright >> 16; |
| 419 | j = k >> 2; |
| 420 | spix = k & 3; |
| 421 | |
| 422 | if (k >= 0 && k < 0x1000) |
| 423 | { |
| 424 | int m = 0; |
| 425 | int n = 0; |
| 426 | int length = 0; |
| 427 | min = 0; max = 3; |
| 428 | if (j == yhpix) |
| 429 | { |
| 430 | min = yh & 3; |
| 431 | } |
| 432 | if (j == ylpix) |
| 433 | { |
| 434 | max = yl & 3; |
| 435 | } |
| 436 | if (spix >= min && spix <= max) |
| 437 | { |
| 438 | if (spix == min) |
| 439 | { |
| 440 | minxmx = maxxmx = xstart; |
| 441 | minxhx = maxxhx = xend; |
| 442 | } |
| 443 | else |
| 444 | { |
| 445 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 446 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 447 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 448 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 449 | } |
| 450 | } |
| 451 | |
| 452 | if (spix == max) |
| 453 | { |
| 454 | if (flip) |
| 455 | { |
| 456 | span[j].lx = maxxmx; |
| 457 | span[j].rx = minxhx; |
| 458 | } |
| 459 | else |
| 460 | { |
| 461 | span[j].lx = minxmx; |
| 462 | span[j].rx = maxxhx; |
| 463 | } |
| 464 | } |
| 465 | |
| 466 | length = flip ? (xstart - xend) : (xend - xstart); |
| 467 | |
| 468 | if (spix == ldflag) |
| 469 | { |
| 470 | xfrac = ((xright >> 8) & 0xff); |
| 471 | adjust_attr(); |
| 472 | } |
| 473 | |
| 474 | m = flip ? (xend+1) : (xend-1); |
| 475 | |
| 476 | if (k >= yh && length >= 0 && k <= yl) |
| 477 | { |
| 478 | if (xstart>=0 && xstart <1024) |
| 479 | { |
| 480 | if (!flip) |
| 481 | { |
| 482 | span[j].cvg[xstart] += addleft(xleft); |
| 483 | } |
| 484 | else |
| 485 | { |
| 486 | span[j].cvg[xstart] += addright(xleft); |
| 487 | } |
| 488 | } |
| 489 | if (xend>=0 && xend<1024) |
| 490 | { |
| 491 | if (xstart != xend) |
| 492 | { |
| 493 | if (!flip) |
| 494 | { |
| 495 | span[j].cvg[xend] += addright(xright); |
| 496 | } |
| 497 | else |
| 498 | { |
| 499 | span[j].cvg[xend] += addleft(xright); |
| 500 | } |
| 501 | } |
| 502 | else |
| 503 | { |
| 504 | if (!flip) |
| 505 | { |
| 506 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 507 | } |
| 508 | else |
| 509 | { |
| 510 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 511 | } |
| 512 | if (span[j].cvg[xend] > 200) |
| 513 | { |
| 514 | span[j].cvg[xend] = 0; |
| 515 | } |
| 516 | } |
| 517 | } |
| 518 | for (n = 0; n < (length - 1); n++) |
| 519 | { |
| 520 | if (m>=0 && m < 640) |
| 521 | { |
| 522 | span[j].cvg[m] += 2; |
| 523 | } |
| 524 | |
| 525 | m += m_inc; |
| 526 | } |
| 527 | } |
| 528 | } |
| 529 | |
| 530 | if (spix == 3) |
| 531 | { |
| 532 | addvalues(); |
| 533 | } |
| 534 | xleft += xleft_inc; |
| 535 | xright += xright_inc; |
| 536 | } |
| 537 | |
| 538 | switch (fb_size) // 8bpp needs to be implemented |
| 539 | { |
| 540 | case PIXEL_SIZE_16BIT: |
| 541 | if(flip) |
| 542 | { |
| 543 | render_spans_16_ns_nt_z_f(yh>>2, yl>>2, &tile[0]); break; |
| 544 | } |
| 545 | else |
| 546 | { |
| 547 | render_spans_16_ns_nt_z_nf(yh>>2, yl>>2, &tile[0]); break; |
| 548 | } |
| 549 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 0, 0, 1, flip); break; |
| 550 | default: break; // V-Rally2 does this, fb_size=0 |
| 551 | } |
| 552 | } |
| 553 | #undef addvalues |
| 554 | #undef adjust_attr |
| 555 | #undef addleft |
| 556 | #undef addright |
| 557 | #undef setvalues |
| 558 | |
| 559 | static void triangle_ns_t_nz(UINT32 w1, UINT32 w2) |
| 560 | { |
| 561 | int j; |
| 562 | int xleft, xright, xleft_inc, xright_inc; |
| 563 | int xstart, xend; |
| 564 | int s = 0, t = 0, w = 0; |
| 565 | int dsdx = 0, dtdx = 0, dwdx = 0; |
| 566 | int dsdy = 0, dtdy = 0, dwdy = 0; |
| 567 | int dsde = 0, dtde = 0, dwde = 0; |
| 568 | int tilenum; |
| 569 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 570 | |
| 571 | INT32 yl, ym, yh; |
| 572 | INT32 xl, xm, xh; |
| 573 | INT32 dxldy, dxhdy, dxmdy; |
| 574 | UINT32 w3, w4, w5, w6, w7, w8; |
| 575 | |
| 576 | int k = 0; |
| 577 | |
| 578 | INT32 limcvg = 0; |
| 579 | INT32 startcvg = 0; |
| 580 | |
| 581 | int sign_dxldy = 0; |
| 582 | int sign_dxmdy = 0; |
| 583 | int samesign = 0; |
| 584 | |
| 585 | int dsdiff = 0, dtdiff = 0, dwdiff = 0; |
| 586 | int sign_dxhdy = 0; |
| 587 | |
| 588 | int dsdeh = 0, dtdeh = 0, dwdeh = 0, dsdyh = 0, dtdyh = 0, dwdyh = 0; |
| 589 | int do_offset = 0; |
| 590 | |
| 591 | int xfrac = 0; |
| 592 | int dseoff = 0, dteoff = 0, dweoff = 0; |
| 593 | |
| 594 | int dsdxh = 0, dtdxh = 0, dwdxh = 0; |
| 595 | |
| 596 | int m_inc; |
| 597 | UINT32 min=0, max=3; |
| 598 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 599 | |
| 600 | int spix = 0; // Current subpixel |
| 601 | int ycur; |
| 602 | int ylfar; |
| 603 | int ldflag; |
| 604 | int yhpix; |
| 605 | int ympix; |
| 606 | int ylpix; |
| 607 | |
| 608 | int texture_base = rdp_cmd_cur + 8; |
| 609 | |
| 610 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 611 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 612 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 613 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 614 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 615 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 616 | |
| 617 | yl = (w1 & 0x3fff); |
| 618 | ym = ((w2 >> 16) & 0x3fff); |
| 619 | yh = ((w2 >> 0) & 0x3fff); |
| 620 | xl = (INT32)(w3 & 0x3fffffff); |
| 621 | xh = (INT32)(w5 & 0x3fffffff); |
| 622 | xm = (INT32)(w7 & 0x3fffffff); |
| 623 | // Inverse slopes in 16.16 format |
| 624 | dxldy = (INT32)(w4); |
| 625 | dxhdy = (INT32)(w6); |
| 626 | dxmdy = (INT32)(w8); |
| 627 | |
| 628 | max_level = ((w1 >> 19) & 7); |
| 629 | tilenum = (w1 >> 16) & 0x7; |
| 630 | |
| 631 | if (yl & 0x2000) yl |= 0xffffc000; |
| 632 | if (ym & 0x2000) ym |= 0xffffc000; |
| 633 | if (yh & 0x2000) yh |= 0xffffc000; |
| 634 | |
| 635 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 636 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 637 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 638 | |
| 639 | s = 0; t = 0; w = 0; |
| 640 | |
| 641 | s = (rdp_cmd_data[texture_base+0 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+4 ] >> 16) & 0x0000ffff); |
| 642 | t = ((rdp_cmd_data[texture_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+4 ] & 0x0000ffff); |
| 643 | w = (rdp_cmd_data[texture_base+1 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+5 ] >> 16) & 0x0000ffff); |
| 644 | dsdx = (rdp_cmd_data[texture_base+2 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+6 ] >> 16) & 0x0000ffff); |
| 645 | dtdx = ((rdp_cmd_data[texture_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+6 ] & 0x0000ffff); |
| 646 | dwdx = (rdp_cmd_data[texture_base+3 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+7 ] >> 16) & 0x0000ffff); |
| 647 | dsde = (rdp_cmd_data[texture_base+8 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+12] >> 16) & 0x0000ffff); |
| 648 | dtde = ((rdp_cmd_data[texture_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+12] & 0x0000ffff); |
| 649 | dwde = (rdp_cmd_data[texture_base+9 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+13] >> 16) & 0x0000ffff); |
| 650 | dsdy = (rdp_cmd_data[texture_base+10] & 0xffff0000) | ((rdp_cmd_data[texture_base+14] >> 16) & 0x0000ffff); |
| 651 | dtdy = ((rdp_cmd_data[texture_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+14] & 0x0000ffff); |
| 652 | dwdy = (rdp_cmd_data[texture_base+11] & 0xffff0000) | ((rdp_cmd_data[texture_base+15] >> 16) & 0x0000ffff); |
| 653 | |
| 654 | span[0].ds.w = dsdx; |
| 655 | span[0].dt.w = dtdx; |
| 656 | span[0].dw.w = dwdx; |
| 657 | |
| 658 | xleft_inc = dxmdy >> 2; |
| 659 | xright_inc = dxhdy >> 2; |
| 660 | |
| 661 | xright = xh; |
| 662 | xleft = xm; |
| 663 | |
| 664 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 665 | if (limcvg < 0) |
| 666 | { |
| 667 | limcvg = 0; |
| 668 | } |
| 669 | |
| 670 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 671 | for (k = startcvg; k <= limcvg; k++) |
| 672 | { |
| 673 | memset((void*)&span[k].cvg[0],0,640); |
| 674 | } |
| 675 | |
| 676 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 677 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 678 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 679 | |
| 680 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 681 | |
| 682 | do_offset = !(sign_dxhdy ^ (flip)); |
| 683 | |
| 684 | if (do_offset) |
| 685 | { |
| 686 | dsdeh = dsde >> 9; dsdyh = dsdy >> 9; |
| 687 | dtdeh = dtde >> 9; dtdyh = dtdy >> 9; |
| 688 | dwdeh = dwde >> 9; dwdyh = dwdy >> 9; |
| 689 | |
| 690 | dsdiff = (dsdeh*3 - dsdyh*3) << 7; |
| 691 | dtdiff = (dtdeh*3 - dtdyh*3) << 7; |
| 692 | dwdiff = (dwdeh*3 - dwdyh*3) << 7; |
| 693 | } |
| 694 | else |
| 695 | { |
| 696 | dsdiff = dtdiff = dwdiff = 0; |
| 697 | } |
| 698 | |
| 699 | if (do_offset) |
| 700 | { |
| 701 | dseoff = (dsdeh*3) << 7; |
| 702 | dteoff = (dtdeh*3) << 7; |
| 703 | dweoff = (dwdeh*3) << 7; |
| 704 | } |
| 705 | else |
| 706 | { |
| 707 | dseoff = dteoff = dweoff; |
| 708 | } |
| 709 | |
| 710 | #define addleft(x) addleftcvg(x,k) |
| 711 | #define addright(x) addrightcvg(x,k) |
| 712 | #define setvalues() { \ |
| 713 | addvalues(); \ |
| 714 | adjust_attr(); \ |
| 715 | } |
| 716 | |
| 717 | dsdxh = dsdx >> 8; |
| 718 | dtdxh = dtdx >> 8; |
| 719 | dwdxh = dwdx >> 8; |
| 720 | |
| 721 | #define adjust_attr() \ |
| 722 | { \ |
| 723 | span[j].s.w = (s + dsdiff - (xfrac * dsdxh)) & ~0x1f; \ |
| 724 | span[j].t.w = (t + dtdiff - (xfrac * dtdxh)) & ~0x1f; \ |
| 725 | span[j].w.w = (w + dwdiff - (xfrac * dwdxh)) & ~0x1f; \ |
| 726 | } |
| 727 | |
| 728 | #define addvalues() { \ |
| 729 | s += dsde; \ |
| 730 | t += dtde; \ |
| 731 | w += dwde; \ |
| 732 | } |
| 733 | |
| 734 | m_inc = flip ? 1 : -1; |
| 735 | |
| 736 | ycur = yh & ~3; |
| 737 | ylfar = yl | 3; |
| 738 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 739 | yhpix = yh >> 2; |
| 740 | ympix = ym >> 2; |
| 741 | ylpix = yl >> 2; |
| 742 | |
| 743 | for (k = ycur; k <= ylfar; k++) |
| 744 | { |
| 745 | if (k == ym) |
| 746 | { |
| 747 | xleft = xl; |
| 748 | xleft_inc = dxldy >> 2; |
| 749 | } |
| 750 | |
| 751 | xstart = xleft >> 16; |
| 752 | xend = xright >> 16; |
| 753 | j = k >> 2; |
| 754 | spix = k & 3; |
| 755 | |
| 756 | if (k >= 0 && k < 0x1000) |
| 757 | { |
| 758 | int m = 0; |
| 759 | int n = 0; |
| 760 | int length = 0; |
| 761 | min = 0; max = 3; |
| 762 | if (j == yhpix) |
| 763 | { |
| 764 | min = yh & 3; |
| 765 | } |
| 766 | if (j == ylpix) |
| 767 | { |
| 768 | max = yl & 3; |
| 769 | } |
| 770 | if (spix >= min && spix <= max) |
| 771 | { |
| 772 | if (spix == min) |
| 773 | { |
| 774 | minxmx = maxxmx = xstart; |
| 775 | minxhx = maxxhx = xend; |
| 776 | } |
| 777 | else |
| 778 | { |
| 779 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 780 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 781 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 782 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 783 | } |
| 784 | } |
| 785 | |
| 786 | if (spix == max) |
| 787 | { |
| 788 | if (flip) |
| 789 | { |
| 790 | span[j].lx = maxxmx; |
| 791 | span[j].rx = minxhx; |
| 792 | } |
| 793 | else |
| 794 | { |
| 795 | span[j].lx = minxmx; |
| 796 | span[j].rx = maxxhx; |
| 797 | } |
| 798 | } |
| 799 | |
| 800 | length = flip ? (xstart - xend) : (xend - xstart); |
| 801 | |
| 802 | if (spix == ldflag) |
| 803 | { |
| 804 | xfrac = ((xright >> 8) & 0xff); |
| 805 | adjust_attr(); |
| 806 | } |
| 807 | |
| 808 | m = flip ? (xend+1) : (xend-1); |
| 809 | |
| 810 | if (k >= yh && length >= 0 && k <= yl) |
| 811 | { |
| 812 | if (xstart>=0 && xstart <1024) |
| 813 | { |
| 814 | if (!flip) |
| 815 | { |
| 816 | span[j].cvg[xstart] += addleft(xleft); |
| 817 | } |
| 818 | else |
| 819 | { |
| 820 | span[j].cvg[xstart] += addright(xleft); |
| 821 | } |
| 822 | } |
| 823 | if (xend>=0 && xend<1024) |
| 824 | { |
| 825 | if (xstart != xend) |
| 826 | { |
| 827 | if (!flip) |
| 828 | { |
| 829 | span[j].cvg[xend] += addright(xright); |
| 830 | } |
| 831 | else |
| 832 | { |
| 833 | span[j].cvg[xend] += addleft(xright); |
| 834 | } |
| 835 | } |
| 836 | else |
| 837 | { |
| 838 | if (!flip) |
| 839 | { |
| 840 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 841 | } |
| 842 | else |
| 843 | { |
| 844 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 845 | } |
| 846 | if (span[j].cvg[xend] > 200) |
| 847 | { |
| 848 | span[j].cvg[xend] = 0; |
| 849 | } |
| 850 | } |
| 851 | } |
| 852 | for (n = 0; n < (length - 1); n++) |
| 853 | { |
| 854 | if (m>=0 && m < 640) |
| 855 | { |
| 856 | span[j].cvg[m] += 2; |
| 857 | } |
| 858 | |
| 859 | m += m_inc; |
| 860 | } |
| 861 | } |
| 862 | } |
| 863 | |
| 864 | if (spix == 3) |
| 865 | { |
| 866 | addvalues(); |
| 867 | } |
| 868 | xleft += xleft_inc; |
| 869 | xright += xright_inc; |
| 870 | } |
| 871 | |
| 872 | switch (fb_size) // 8bpp needs to be implemented |
| 873 | { |
| 874 | case PIXEL_SIZE_16BIT: |
| 875 | if(flip) |
| 876 | { |
| 877 | render_spans_16_ns_t_nz_f(yh>>2, yl>>2, &tile[0]); break; |
| 878 | } |
| 879 | else |
| 880 | { |
| 881 | render_spans_16_ns_t_nz_nf(yh>>2, yl>>2, &tile[0]); break; |
| 882 | } |
| 883 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 0, 1, 0, flip); break; |
| 884 | default: break; // V-Rally2 does this, fb_size=0 |
| 885 | } |
| 886 | } |
| 887 | #undef addvalues |
| 888 | #undef adjust_attr |
| 889 | #undef addleft |
| 890 | #undef addright |
| 891 | #undef setvalues |
| 892 | |
| 893 | static void triangle_ns_t_z(UINT32 w1, UINT32 w2) |
| 894 | { |
| 895 | int j; |
| 896 | int xleft, xright, xleft_inc, xright_inc; |
| 897 | int xstart, xend; |
| 898 | int z = 0, s = 0, t = 0, w = 0; |
| 899 | int dzdx = 0, dsdx = 0, dtdx = 0, dwdx = 0; |
| 900 | int dzdy = 0, dsdy = 0, dtdy = 0, dwdy = 0; |
| 901 | int dzde = 0, dsde = 0, dtde = 0, dwde = 0; |
| 902 | int tilenum; |
| 903 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 904 | |
| 905 | INT32 yl, ym, yh; |
| 906 | INT32 xl, xm, xh; |
| 907 | INT32 dxldy, dxhdy, dxmdy; |
| 908 | int dzdy_dz, dzdx_dz; |
| 909 | int dsdylod, dtdylod; |
| 910 | UINT32 w3, w4, w5, w6, w7, w8; |
| 911 | |
| 912 | int k = 0; |
| 913 | |
| 914 | INT32 limcvg = 0; |
| 915 | INT32 startcvg = 0; |
| 916 | |
| 917 | int sign_dxldy = 0; |
| 918 | int sign_dxmdy = 0; |
| 919 | int samesign = 0; |
| 920 | |
| 921 | int dsdiff = 0, dtdiff = 0, dwdiff = 0, dzdiff = 0; |
| 922 | int sign_dxhdy = 0; |
| 923 | |
| 924 | int dsdeh = 0, dtdeh = 0, dwdeh = 0, dzdeh = 0, dsdyh = 0, dtdyh = 0, dwdyh = 0, dzdyh = 0; |
| 925 | int do_offset = 0; |
| 926 | |
| 927 | int xfrac = 0; |
| 928 | int dseoff = 0, dteoff = 0, dweoff = 0, dzeoff = 0; |
| 929 | |
| 930 | int dsdxh = 0, dtdxh = 0, dwdxh = 0, dzdxh = 0; |
| 931 | |
| 932 | int m_inc; |
| 933 | UINT32 min=0, max=3; |
| 934 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 935 | |
| 936 | int spix = 0; // Current subpixel |
| 937 | int ycur; |
| 938 | int ylfar; |
| 939 | int ldflag; |
| 940 | int yhpix; |
| 941 | int ympix; |
| 942 | int ylpix; |
| 943 | |
| 944 | int texture_base = rdp_cmd_cur + 8; |
| 945 | int zbuffer_base = rdp_cmd_cur + 8; |
| 946 | |
| 947 | zbuffer_base += 16; |
| 948 | |
| 949 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 950 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 951 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 952 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 953 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 954 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 955 | |
| 956 | yl = (w1 & 0x3fff); |
| 957 | ym = ((w2 >> 16) & 0x3fff); |
| 958 | yh = ((w2 >> 0) & 0x3fff); |
| 959 | xl = (INT32)(w3 & 0x3fffffff); |
| 960 | xh = (INT32)(w5 & 0x3fffffff); |
| 961 | xm = (INT32)(w7 & 0x3fffffff); |
| 962 | // Inverse slopes in 16.16 format |
| 963 | dxldy = (INT32)(w4); |
| 964 | dxhdy = (INT32)(w6); |
| 965 | dxmdy = (INT32)(w8); |
| 966 | |
| 967 | max_level = ((w1 >> 19) & 7); |
| 968 | tilenum = (w1 >> 16) & 0x7; |
| 969 | |
| 970 | if (yl & 0x2000) yl |= 0xffffc000; |
| 971 | if (ym & 0x2000) ym |= 0xffffc000; |
| 972 | if (yh & 0x2000) yh |= 0xffffc000; |
| 973 | |
| 974 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 975 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 976 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 977 | |
| 978 | z = 0; |
| 979 | s = 0; t = 0; w = 0; |
| 980 | |
| 981 | s = (rdp_cmd_data[texture_base+0 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+4 ] >> 16) & 0x0000ffff); |
| 982 | t = ((rdp_cmd_data[texture_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+4 ] & 0x0000ffff); |
| 983 | w = (rdp_cmd_data[texture_base+1 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+5 ] >> 16) & 0x0000ffff); |
| 984 | dsdx = (rdp_cmd_data[texture_base+2 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+6 ] >> 16) & 0x0000ffff); |
| 985 | dtdx = ((rdp_cmd_data[texture_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+6 ] & 0x0000ffff); |
| 986 | dwdx = (rdp_cmd_data[texture_base+3 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+7 ] >> 16) & 0x0000ffff); |
| 987 | dsde = (rdp_cmd_data[texture_base+8 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+12] >> 16) & 0x0000ffff); |
| 988 | dtde = ((rdp_cmd_data[texture_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+12] & 0x0000ffff); |
| 989 | dwde = (rdp_cmd_data[texture_base+9 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+13] >> 16) & 0x0000ffff); |
| 990 | dsdy = (rdp_cmd_data[texture_base+10] & 0xffff0000) | ((rdp_cmd_data[texture_base+14] >> 16) & 0x0000ffff); |
| 991 | dtdy = ((rdp_cmd_data[texture_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+14] & 0x0000ffff); |
| 992 | dwdy = (rdp_cmd_data[texture_base+11] & 0xffff0000) | ((rdp_cmd_data[texture_base+15] >> 16) & 0x0000ffff); |
| 993 | z = rdp_cmd_data[zbuffer_base+0]; |
| 994 | dzdx = rdp_cmd_data[zbuffer_base+1]; |
| 995 | dzde = rdp_cmd_data[zbuffer_base+2]; |
| 996 | dzdy = rdp_cmd_data[zbuffer_base+3]; |
| 997 | |
| 998 | span[0].ds.w = dsdx; |
| 999 | span[0].dt.w = dtdx; |
| 1000 | span[0].dw.w = dwdx; |
| 1001 | span[0].dz.w = dzdx; |
| 1002 | dzdy_dz = (dzdy >> 16) & 0xffff; |
| 1003 | dzdx_dz = (dzdx >> 16) & 0xffff; |
| 1004 | span[0].dzpix = ((dzdy_dz & 0x8000) ? ((~dzdy_dz) & 0x7fff) : dzdy_dz) + ((dzdx_dz & 0x8000) ? ((~dzdx_dz) & 0x7fff) : dzdx_dz); |
| 1005 | span[0].dzpix = normalize_dzpix(span[0].dzpix); |
| 1006 | dsdylod = dsdy >> 16; |
| 1007 | dtdylod = dtdy >> 16; |
| 1008 | if (dsdylod & 0x20000) |
| 1009 | { |
| 1010 | dsdylod = ~dsdylod & 0x1ffff; |
| 1011 | } |
| 1012 | if (dtdylod & 0x20000) |
| 1013 | { |
| 1014 | dtdylod = ~dtdylod & 0x1ffff; |
| 1015 | } |
| 1016 | span[0].dymax = (dsdylod > dtdylod)? dsdylod : dtdylod; |
| 1017 | |
| 1018 | xleft_inc = dxmdy >> 2; |
| 1019 | xright_inc = dxhdy >> 2; |
| 1020 | |
| 1021 | xright = xh; |
| 1022 | xleft = xm; |
| 1023 | |
| 1024 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 1025 | if (limcvg < 0) |
| 1026 | { |
| 1027 | limcvg = 0; |
| 1028 | } |
| 1029 | |
| 1030 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 1031 | for (k = startcvg; k <= limcvg; k++) |
| 1032 | { |
| 1033 | memset((void*)&span[k].cvg[0],0,640); |
| 1034 | } |
| 1035 | |
| 1036 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 1037 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 1038 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 1039 | |
| 1040 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 1041 | |
| 1042 | do_offset = !(sign_dxhdy ^ (flip)); |
| 1043 | |
| 1044 | if (do_offset) |
| 1045 | { |
| 1046 | dsdeh = dsde >> 9; dsdyh = dsdy >> 9; |
| 1047 | dtdeh = dtde >> 9; dtdyh = dtdy >> 9; |
| 1048 | dwdeh = dwde >> 9; dwdyh = dwdy >> 9; |
| 1049 | dzdeh = dzde >> 9; dzdyh = dzdy >> 9; |
| 1050 | |
| 1051 | dsdiff = (dsdeh*3 - dsdyh*3) << 7; |
| 1052 | dtdiff = (dtdeh*3 - dtdyh*3) << 7; |
| 1053 | dwdiff = (dwdeh*3 - dwdyh*3) << 7; |
| 1054 | dzdiff = (dzdeh*3 - dzdyh*3) << 7; |
| 1055 | } |
| 1056 | else |
| 1057 | { |
| 1058 | dsdiff = dtdiff = dwdiff = dzdiff = 0; |
| 1059 | } |
| 1060 | |
| 1061 | if (do_offset) |
| 1062 | { |
| 1063 | dseoff = (dsdeh*3) << 7; |
| 1064 | dteoff = (dtdeh*3) << 7; |
| 1065 | dweoff = (dwdeh*3) << 7; |
| 1066 | dzeoff = (dzdeh*3) << 7; |
| 1067 | } |
| 1068 | else |
| 1069 | { |
| 1070 | dseoff = dteoff = dweoff = dzeoff = 0; |
| 1071 | } |
| 1072 | |
| 1073 | #define addleft(x) addleftcvg(x,k) |
| 1074 | #define addright(x) addrightcvg(x,k) |
| 1075 | #define setvalues() { \ |
| 1076 | addvalues(); \ |
| 1077 | adjust_attr(); \ |
| 1078 | } |
| 1079 | |
| 1080 | dsdxh = dsdx >> 8; |
| 1081 | dtdxh = dtdx >> 8; |
| 1082 | dwdxh = dwdx >> 8; |
| 1083 | dzdxh = dzdx >> 8; |
| 1084 | |
| 1085 | #define adjust_attr() \ |
| 1086 | { \ |
| 1087 | span[j].s.w = (s + dsdiff - (xfrac * dsdxh)) & ~0x1f; \ |
| 1088 | span[j].t.w = (t + dtdiff - (xfrac * dtdxh)) & ~0x1f; \ |
| 1089 | span[j].w.w = (w + dwdiff - (xfrac * dwdxh)) & ~0x1f; \ |
| 1090 | span[j].z.w = z + dzdiff - (xfrac * dzdxh); \ |
| 1091 | } |
| 1092 | |
| 1093 | #define addvalues() { \ |
| 1094 | s += dsde; \ |
| 1095 | t += dtde; \ |
| 1096 | w += dwde; \ |
| 1097 | z += dzde; \ |
| 1098 | } |
| 1099 | |
| 1100 | m_inc = flip ? 1 : -1; |
| 1101 | |
| 1102 | ycur = yh & ~3; |
| 1103 | ylfar = yl | 3; |
| 1104 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 1105 | yhpix = yh >> 2; |
| 1106 | ympix = ym >> 2; |
| 1107 | ylpix = yl >> 2; |
| 1108 | |
| 1109 | for (k = ycur; k <= ylfar; k++) |
| 1110 | { |
| 1111 | if (k == ym) |
| 1112 | { |
| 1113 | xleft = xl; |
| 1114 | xleft_inc = dxldy >> 2; |
| 1115 | } |
| 1116 | |
| 1117 | xstart = xleft >> 16; |
| 1118 | xend = xright >> 16; |
| 1119 | j = k >> 2; |
| 1120 | spix = k & 3; |
| 1121 | |
| 1122 | if (k >= 0 && k < 0x1000) |
| 1123 | { |
| 1124 | int m = 0; |
| 1125 | int n = 0; |
| 1126 | int length = 0; |
| 1127 | min = 0; max = 3; |
| 1128 | if (j == yhpix) |
| 1129 | { |
| 1130 | min = yh & 3; |
| 1131 | } |
| 1132 | if (j == ylpix) |
| 1133 | { |
| 1134 | max = yl & 3; |
| 1135 | } |
| 1136 | if (spix >= min && spix <= max) |
| 1137 | { |
| 1138 | if (spix == min) |
| 1139 | { |
| 1140 | minxmx = maxxmx = xstart; |
| 1141 | minxhx = maxxhx = xend; |
| 1142 | } |
| 1143 | else |
| 1144 | { |
| 1145 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 1146 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 1147 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 1148 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 1149 | } |
| 1150 | } |
| 1151 | |
| 1152 | if (spix == max) |
| 1153 | { |
| 1154 | if (flip) |
| 1155 | { |
| 1156 | span[j].lx = maxxmx; |
| 1157 | span[j].rx = minxhx; |
| 1158 | } |
| 1159 | else |
| 1160 | { |
| 1161 | span[j].lx = minxmx; |
| 1162 | span[j].rx = maxxhx; |
| 1163 | } |
| 1164 | } |
| 1165 | |
| 1166 | length = flip ? (xstart - xend) : (xend - xstart); |
| 1167 | |
| 1168 | if (spix == ldflag) |
| 1169 | { |
| 1170 | xfrac = ((xright >> 8) & 0xff); |
| 1171 | adjust_attr(); |
| 1172 | } |
| 1173 | |
| 1174 | m = flip ? (xend+1) : (xend-1); |
| 1175 | |
| 1176 | if (k >= yh && length >= 0 && k <= yl) |
| 1177 | { |
| 1178 | if (xstart>=0 && xstart <1024) |
| 1179 | { |
| 1180 | if (!flip) |
| 1181 | { |
| 1182 | span[j].cvg[xstart] += addleft(xleft); |
| 1183 | } |
| 1184 | else |
| 1185 | { |
| 1186 | span[j].cvg[xstart] += addright(xleft); |
| 1187 | } |
| 1188 | } |
| 1189 | if (xend>=0 && xend<1024) |
| 1190 | { |
| 1191 | if (xstart != xend) |
| 1192 | { |
| 1193 | if (!flip) |
| 1194 | { |
| 1195 | span[j].cvg[xend] += addright(xright); |
| 1196 | } |
| 1197 | else |
| 1198 | { |
| 1199 | span[j].cvg[xend] += addleft(xright); |
| 1200 | } |
| 1201 | } |
| 1202 | else |
| 1203 | { |
| 1204 | if (!flip) |
| 1205 | { |
| 1206 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 1207 | } |
| 1208 | else |
| 1209 | { |
| 1210 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 1211 | } |
| 1212 | if (span[j].cvg[xend] > 200) |
| 1213 | { |
| 1214 | span[j].cvg[xend] = 0; |
| 1215 | } |
| 1216 | } |
| 1217 | } |
| 1218 | for (n = 0; n < (length - 1); n++) |
| 1219 | { |
| 1220 | if (m>=0 && m < 640) |
| 1221 | { |
| 1222 | span[j].cvg[m] += 2; |
| 1223 | } |
| 1224 | |
| 1225 | m += m_inc; |
| 1226 | } |
| 1227 | } |
| 1228 | } |
| 1229 | |
| 1230 | if (spix == 3) |
| 1231 | { |
| 1232 | addvalues(); |
| 1233 | } |
| 1234 | xleft += xleft_inc; |
| 1235 | xright += xright_inc; |
| 1236 | } |
| 1237 | |
| 1238 | switch (fb_size) // 8bpp needs to be implemented |
| 1239 | { |
| 1240 | case PIXEL_SIZE_16BIT: |
| 1241 | if(flip) |
| 1242 | { |
| 1243 | render_spans_16_ns_t_z_f(yh>>2, yl>>2, &tile[0]); break; |
| 1244 | } |
| 1245 | else |
| 1246 | { |
| 1247 | render_spans_16_ns_t_z_nf(yh>>2, yl>>2, &tile[0]); break; |
| 1248 | } |
| 1249 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 0, 1, 1, flip); break; |
| 1250 | default: break; // V-Rally2 does this, fb_size=0 |
| 1251 | } |
| 1252 | } |
| 1253 | #undef addvalues |
| 1254 | #undef adjust_attr |
| 1255 | #undef addleft |
| 1256 | #undef addright |
| 1257 | #undef setvalues |
| 1258 | |
| 1259 | static void triangle_s_nt_nz(UINT32 w1, UINT32 w2) |
| 1260 | { |
| 1261 | int j; |
| 1262 | int xleft, xright, xleft_inc, xright_inc; |
| 1263 | int xstart, xend; |
| 1264 | int r = 0, g = 0, b = 0, a = 0; |
| 1265 | int dr, dg, db, da; |
| 1266 | int drdx = 0, dgdx = 0, dbdx = 0, dadx = 0; |
| 1267 | int drdy = 0, dgdy = 0, dbdy = 0, dady = 0; |
| 1268 | int drde = 0, dgde = 0, dbde = 0, dade = 0; |
| 1269 | int tilenum; |
| 1270 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 1271 | |
| 1272 | INT32 yl, ym, yh; |
| 1273 | INT32 xl, xm, xh; |
| 1274 | INT32 dxldy, dxhdy, dxmdy; |
| 1275 | UINT32 w3, w4, w5, w6, w7, w8; |
| 1276 | |
| 1277 | int k = 0; |
| 1278 | |
| 1279 | INT32 limcvg = 0; |
| 1280 | INT32 startcvg = 0; |
| 1281 | |
| 1282 | int sign_dxldy = 0; |
| 1283 | int sign_dxmdy = 0; |
| 1284 | int samesign = 0; |
| 1285 | |
| 1286 | int drdiff = 0, dgdiff = 0, dbdiff = 0, dadiff = 0; |
| 1287 | int sign_dxhdy = 0; |
| 1288 | |
| 1289 | int drdeh = 0, dgdeh = 0, dbdeh = 0, dadeh = 0, drdyh = 0, dgdyh = 0, dbdyh = 0, dadyh = 0; |
| 1290 | int do_offset = 0; |
| 1291 | |
| 1292 | int xfrac = 0; |
| 1293 | int dreoff = 0, dgeoff = 0, dbeoff = 0, daeoff = 0; |
| 1294 | |
| 1295 | int drdxh = 0, dgdxh = 0, dbdxh = 0, dadxh = 0; |
| 1296 | |
| 1297 | int m_inc; |
| 1298 | UINT32 min=0, max=3; |
| 1299 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 1300 | |
| 1301 | int spix = 0; // Current subpixel |
| 1302 | int ycur; |
| 1303 | int ylfar; |
| 1304 | int ldflag; |
| 1305 | int yhpix; |
| 1306 | int ympix; |
| 1307 | int ylpix; |
| 1308 | |
| 1309 | int shade_base = rdp_cmd_cur + 8; |
| 1310 | |
| 1311 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 1312 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 1313 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 1314 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 1315 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 1316 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 1317 | |
| 1318 | yl = (w1 & 0x3fff); |
| 1319 | ym = ((w2 >> 16) & 0x3fff); |
| 1320 | yh = ((w2 >> 0) & 0x3fff); |
| 1321 | xl = (INT32)(w3 & 0x3fffffff); |
| 1322 | xh = (INT32)(w5 & 0x3fffffff); |
| 1323 | xm = (INT32)(w7 & 0x3fffffff); |
| 1324 | // Inverse slopes in 16.16 format |
| 1325 | dxldy = (INT32)(w4); |
| 1326 | dxhdy = (INT32)(w6); |
| 1327 | dxmdy = (INT32)(w8); |
| 1328 | |
| 1329 | max_level = ((w1 >> 19) & 7); |
| 1330 | tilenum = (w1 >> 16) & 0x7; |
| 1331 | |
| 1332 | if (yl & 0x2000) yl |= 0xffffc000; |
| 1333 | if (ym & 0x2000) ym |= 0xffffc000; |
| 1334 | if (yh & 0x2000) yh |= 0xffffc000; |
| 1335 | |
| 1336 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 1337 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 1338 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 1339 | |
| 1340 | dr = 0; dg = 0; db = 0; da = 0; |
| 1341 | |
| 1342 | r = (rdp_cmd_data[shade_base+0 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+4 ] >> 16) & 0x0000ffff); |
| 1343 | g = ((rdp_cmd_data[shade_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+4 ] & 0x0000ffff); |
| 1344 | b = (rdp_cmd_data[shade_base+1 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+5 ] >> 16) & 0x0000ffff); |
| 1345 | a = ((rdp_cmd_data[shade_base+1 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+5 ] & 0x0000ffff); |
| 1346 | drdx = (rdp_cmd_data[shade_base+2 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+6 ] >> 16) & 0x0000ffff); |
| 1347 | dgdx = ((rdp_cmd_data[shade_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+6 ] & 0x0000ffff); |
| 1348 | dbdx = (rdp_cmd_data[shade_base+3 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+7 ] >> 16) & 0x0000ffff); |
| 1349 | dadx = ((rdp_cmd_data[shade_base+3 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+7 ] & 0x0000ffff); |
| 1350 | drde = (rdp_cmd_data[shade_base+8 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+12] >> 16) & 0x0000ffff); |
| 1351 | dgde = ((rdp_cmd_data[shade_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+12] & 0x0000ffff); |
| 1352 | dbde = (rdp_cmd_data[shade_base+9 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+13] >> 16) & 0x0000ffff); |
| 1353 | dade = ((rdp_cmd_data[shade_base+9 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+13] & 0x0000ffff); |
| 1354 | drdy = (rdp_cmd_data[shade_base+10] & 0xffff0000) | ((rdp_cmd_data[shade_base+14] >> 16) & 0x0000ffff); |
| 1355 | dgdy = ((rdp_cmd_data[shade_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+14] & 0x0000ffff); |
| 1356 | dbdy = (rdp_cmd_data[shade_base+11] & 0xffff0000) | ((rdp_cmd_data[shade_base+15] >> 16) & 0x0000ffff); |
| 1357 | dady = ((rdp_cmd_data[shade_base+11] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+15] & 0x0000ffff); |
| 1358 | |
| 1359 | span[0].dr.w = drdx & ~0x1f; |
| 1360 | span[0].dg.w = dgdx & ~0x1f; |
| 1361 | span[0].db.w = dbdx & ~0x1f; |
| 1362 | span[0].da.w = dadx & ~0x1f; |
| 1363 | |
| 1364 | xleft_inc = dxmdy >> 2; |
| 1365 | xright_inc = dxhdy >> 2; |
| 1366 | |
| 1367 | xright = xh; |
| 1368 | xleft = xm; |
| 1369 | |
| 1370 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 1371 | if (limcvg < 0) |
| 1372 | { |
| 1373 | limcvg = 0; |
| 1374 | } |
| 1375 | |
| 1376 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 1377 | for (k = startcvg; k <= limcvg; k++) |
| 1378 | { |
| 1379 | memset((void*)&span[k].cvg[0],0,640); |
| 1380 | } |
| 1381 | |
| 1382 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 1383 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 1384 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 1385 | |
| 1386 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 1387 | |
| 1388 | do_offset = !(sign_dxhdy ^ (flip)); |
| 1389 | |
| 1390 | if (do_offset) |
| 1391 | { |
| 1392 | drdeh = drde >> 9; drdyh = drdy >> 9; |
| 1393 | dgdeh = dgde >> 9; dgdyh = dgdy >> 9; |
| 1394 | dbdeh = dbde >> 9; dbdyh = dbdy >> 9; |
| 1395 | dadeh = dade >> 9; dadyh = dady >> 9; |
| 1396 | |
| 1397 | drdiff = (drdeh*3 - drdyh*3) << 7; |
| 1398 | dgdiff = (dgdeh*3 - dgdyh*3) << 7; |
| 1399 | dbdiff = (dbdeh*3 - dbdyh*3) << 7; |
| 1400 | dadiff = (dadeh*3 - dadyh*3) << 7; |
| 1401 | } |
| 1402 | else |
| 1403 | { |
| 1404 | drdiff = dgdiff = dbdiff = dadiff; |
| 1405 | } |
| 1406 | |
| 1407 | if (do_offset) |
| 1408 | { |
| 1409 | dreoff = (drdeh*3) << 7; |
| 1410 | dgeoff = (dgdeh*3) << 7; |
| 1411 | dbeoff = (dbdeh*3) << 7; |
| 1412 | daeoff = (dadeh*3) << 7; |
| 1413 | } |
| 1414 | else |
| 1415 | { |
| 1416 | dreoff = dgeoff = dbeoff = daeoff; |
| 1417 | } |
| 1418 | |
| 1419 | #define addleft(x) addleftcvg(x,k) |
| 1420 | #define addright(x) addrightcvg(x,k) |
| 1421 | #define setvalues() { \ |
| 1422 | addvalues(); \ |
| 1423 | adjust_attr(); \ |
| 1424 | } |
| 1425 | |
| 1426 | drdxh = drdx >> 8; |
| 1427 | dgdxh = dgdx >> 8; |
| 1428 | dbdxh = dbdx >> 8; |
| 1429 | dadxh = dadx >> 8; |
| 1430 | |
| 1431 | #define adjust_attr() \ |
| 1432 | { \ |
| 1433 | span[j].r.w = r + drdiff - (xfrac * drdxh); \ |
| 1434 | span[j].g.w = g + dgdiff - (xfrac * dgdxh); \ |
| 1435 | span[j].b.w = b + dbdiff - (xfrac * dbdxh); \ |
| 1436 | span[j].a.w = a + dadiff - (xfrac * dadxh); \ |
| 1437 | } |
| 1438 | |
| 1439 | #define addvalues() { \ |
| 1440 | r += drde; \ |
| 1441 | g += dgde; \ |
| 1442 | b += dbde; \ |
| 1443 | a += dade; \ |
| 1444 | } |
| 1445 | |
| 1446 | m_inc = flip ? 1 : -1; |
| 1447 | |
| 1448 | ycur = yh & ~3; |
| 1449 | ylfar = yl | 3; |
| 1450 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 1451 | yhpix = yh >> 2; |
| 1452 | ympix = ym >> 2; |
| 1453 | ylpix = yl >> 2; |
| 1454 | |
| 1455 | for (k = ycur; k <= ylfar; k++) |
| 1456 | { |
| 1457 | if (k == ym) |
| 1458 | { |
| 1459 | xleft = xl; |
| 1460 | xleft_inc = dxldy >> 2; |
| 1461 | } |
| 1462 | |
| 1463 | xstart = xleft >> 16; |
| 1464 | xend = xright >> 16; |
| 1465 | j = k >> 2; |
| 1466 | spix = k & 3; |
| 1467 | |
| 1468 | if (k >= 0 && k < 0x1000) |
| 1469 | { |
| 1470 | int m = 0; |
| 1471 | int n = 0; |
| 1472 | int length = 0; |
| 1473 | min = 0; max = 3; |
| 1474 | if (j == yhpix) |
| 1475 | { |
| 1476 | min = yh & 3; |
| 1477 | } |
| 1478 | if (j == ylpix) |
| 1479 | { |
| 1480 | max = yl & 3; |
| 1481 | } |
| 1482 | if (spix >= min && spix <= max) |
| 1483 | { |
| 1484 | if (spix == min) |
| 1485 | { |
| 1486 | minxmx = maxxmx = xstart; |
| 1487 | minxhx = maxxhx = xend; |
| 1488 | } |
| 1489 | else |
| 1490 | { |
| 1491 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 1492 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 1493 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 1494 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 1495 | } |
| 1496 | } |
| 1497 | |
| 1498 | if (spix == max) |
| 1499 | { |
| 1500 | if (flip) |
| 1501 | { |
| 1502 | span[j].lx = maxxmx; |
| 1503 | span[j].rx = minxhx; |
| 1504 | } |
| 1505 | else |
| 1506 | { |
| 1507 | span[j].lx = minxmx; |
| 1508 | span[j].rx = maxxhx; |
| 1509 | } |
| 1510 | } |
| 1511 | |
| 1512 | length = flip ? (xstart - xend) : (xend - xstart); |
| 1513 | |
| 1514 | if (spix == ldflag) |
| 1515 | { |
| 1516 | xfrac = ((xright >> 8) & 0xff); |
| 1517 | adjust_attr(); |
| 1518 | } |
| 1519 | |
| 1520 | m = flip ? (xend+1) : (xend-1); |
| 1521 | |
| 1522 | if (k >= yh && length >= 0 && k <= yl) |
| 1523 | { |
| 1524 | if (xstart>=0 && xstart <1024) |
| 1525 | { |
| 1526 | if (!flip) |
| 1527 | { |
| 1528 | span[j].cvg[xstart] += addleft(xleft); |
| 1529 | } |
| 1530 | else |
| 1531 | { |
| 1532 | span[j].cvg[xstart] += addright(xleft); |
| 1533 | } |
| 1534 | } |
| 1535 | if (xend>=0 && xend<1024) |
| 1536 | { |
| 1537 | if (xstart != xend) |
| 1538 | { |
| 1539 | if (!flip) |
| 1540 | { |
| 1541 | span[j].cvg[xend] += addright(xright); |
| 1542 | } |
| 1543 | else |
| 1544 | { |
| 1545 | span[j].cvg[xend] += addleft(xright); |
| 1546 | } |
| 1547 | } |
| 1548 | else |
| 1549 | { |
| 1550 | if (!flip) |
| 1551 | { |
| 1552 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 1553 | } |
| 1554 | else |
| 1555 | { |
| 1556 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 1557 | } |
| 1558 | if (span[j].cvg[xend] > 200) |
| 1559 | { |
| 1560 | span[j].cvg[xend] = 0; |
| 1561 | } |
| 1562 | } |
| 1563 | } |
| 1564 | for (n = 0; n < (length - 1); n++) |
| 1565 | { |
| 1566 | if (m>=0 && m < 640) |
| 1567 | { |
| 1568 | span[j].cvg[m] += 2; |
| 1569 | } |
| 1570 | |
| 1571 | m += m_inc; |
| 1572 | } |
| 1573 | } |
| 1574 | } |
| 1575 | |
| 1576 | if (spix == 3) |
| 1577 | { |
| 1578 | addvalues(); |
| 1579 | } |
| 1580 | xleft += xleft_inc; |
| 1581 | xright += xright_inc; |
| 1582 | } |
| 1583 | |
| 1584 | switch (fb_size) // 8bpp needs to be implemented |
| 1585 | { |
| 1586 | case PIXEL_SIZE_16BIT: |
| 1587 | if(flip) |
| 1588 | { |
| 1589 | render_spans_16_s_nt_nz_f(yh>>2, yl>>2, &tile[0]); break; |
| 1590 | } |
| 1591 | else |
| 1592 | { |
| 1593 | render_spans_16_s_nt_nz_nf(yh>>2, yl>>2, &tile[0]); break; |
| 1594 | } |
| 1595 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 1, 0, 0, flip); break; |
| 1596 | default: break; // V-Rally2 does this, fb_size=0 |
| 1597 | } |
| 1598 | } |
| 1599 | #undef addvalues |
| 1600 | #undef adjust_attr |
| 1601 | #undef addleft |
| 1602 | #undef addright |
| 1603 | #undef setvalues |
| 1604 | |
| 1605 | static void triangle_s_nt_z(UINT32 w1, UINT32 w2) |
| 1606 | { |
| 1607 | int j; |
| 1608 | int xleft, xright, xleft_inc, xright_inc; |
| 1609 | int xstart, xend; |
| 1610 | int r = 0, g = 0, b = 0, a = 0, z = 0; |
| 1611 | int dr, dg, db, da; |
| 1612 | int drdx = 0, dgdx = 0, dbdx = 0, dadx = 0, dzdx = 0; |
| 1613 | int drdy = 0, dgdy = 0, dbdy = 0, dady = 0, dzdy = 0; |
| 1614 | int drde = 0, dgde = 0, dbde = 0, dade = 0, dzde = 0; |
| 1615 | int tilenum; |
| 1616 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 1617 | |
| 1618 | INT32 yl, ym, yh; |
| 1619 | INT32 xl, xm, xh; |
| 1620 | INT32 dxldy, dxhdy, dxmdy; |
| 1621 | int dzdy_dz, dzdx_dz; |
| 1622 | UINT32 w3, w4, w5, w6, w7, w8; |
| 1623 | |
| 1624 | int k = 0; |
| 1625 | |
| 1626 | INT32 limcvg = 0; |
| 1627 | INT32 startcvg = 0; |
| 1628 | |
| 1629 | int sign_dxldy = 0; |
| 1630 | int sign_dxmdy = 0; |
| 1631 | int samesign = 0; |
| 1632 | |
| 1633 | int drdiff = 0, dgdiff = 0, dbdiff = 0, dadiff = 0, dzdiff = 0; |
| 1634 | int sign_dxhdy = 0; |
| 1635 | |
| 1636 | int drdeh = 0, dgdeh = 0, dbdeh = 0, dadeh = 0, dzdeh = 0, drdyh = 0, dgdyh = 0, dbdyh = 0, dadyh = 0, dzdyh = 0; |
| 1637 | int do_offset = 0; |
| 1638 | |
| 1639 | int xfrac = 0; |
| 1640 | int dreoff = 0, dgeoff = 0, dbeoff = 0, daeoff = 0, dzeoff = 0; |
| 1641 | |
| 1642 | int drdxh = 0, dgdxh = 0, dbdxh = 0, dadxh = 0, dzdxh = 0; |
| 1643 | |
| 1644 | int m_inc; |
| 1645 | UINT32 min=0, max=3; |
| 1646 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 1647 | |
| 1648 | int spix = 0; // Current subpixel |
| 1649 | int ycur; |
| 1650 | int ylfar; |
| 1651 | int ldflag; |
| 1652 | int yhpix; |
| 1653 | int ympix; |
| 1654 | int ylpix; |
| 1655 | |
| 1656 | int shade_base = rdp_cmd_cur + 8; |
| 1657 | int zbuffer_base = rdp_cmd_cur + 8; |
| 1658 | |
| 1659 | zbuffer_base += 16; |
| 1660 | |
| 1661 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 1662 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 1663 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 1664 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 1665 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 1666 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 1667 | |
| 1668 | yl = (w1 & 0x3fff); |
| 1669 | ym = ((w2 >> 16) & 0x3fff); |
| 1670 | yh = ((w2 >> 0) & 0x3fff); |
| 1671 | xl = (INT32)(w3 & 0x3fffffff); |
| 1672 | xh = (INT32)(w5 & 0x3fffffff); |
| 1673 | xm = (INT32)(w7 & 0x3fffffff); |
| 1674 | // Inverse slopes in 16.16 format |
| 1675 | dxldy = (INT32)(w4); |
| 1676 | dxhdy = (INT32)(w6); |
| 1677 | dxmdy = (INT32)(w8); |
| 1678 | |
| 1679 | max_level = ((w1 >> 19) & 7); |
| 1680 | tilenum = (w1 >> 16) & 0x7; |
| 1681 | |
| 1682 | if (yl & 0x2000) yl |= 0xffffc000; |
| 1683 | if (ym & 0x2000) ym |= 0xffffc000; |
| 1684 | if (yh & 0x2000) yh |= 0xffffc000; |
| 1685 | |
| 1686 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 1687 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 1688 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 1689 | |
| 1690 | z = 0; |
| 1691 | dr = 0; dg = 0; db = 0; da = 0; |
| 1692 | |
| 1693 | r = (rdp_cmd_data[shade_base+0 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+4 ] >> 16) & 0x0000ffff); |
| 1694 | g = ((rdp_cmd_data[shade_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+4 ] & 0x0000ffff); |
| 1695 | b = (rdp_cmd_data[shade_base+1 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+5 ] >> 16) & 0x0000ffff); |
| 1696 | a = ((rdp_cmd_data[shade_base+1 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+5 ] & 0x0000ffff); |
| 1697 | drdx = (rdp_cmd_data[shade_base+2 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+6 ] >> 16) & 0x0000ffff); |
| 1698 | dgdx = ((rdp_cmd_data[shade_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+6 ] & 0x0000ffff); |
| 1699 | dbdx = (rdp_cmd_data[shade_base+3 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+7 ] >> 16) & 0x0000ffff); |
| 1700 | dadx = ((rdp_cmd_data[shade_base+3 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+7 ] & 0x0000ffff); |
| 1701 | drde = (rdp_cmd_data[shade_base+8 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+12] >> 16) & 0x0000ffff); |
| 1702 | dgde = ((rdp_cmd_data[shade_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+12] & 0x0000ffff); |
| 1703 | dbde = (rdp_cmd_data[shade_base+9 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+13] >> 16) & 0x0000ffff); |
| 1704 | dade = ((rdp_cmd_data[shade_base+9 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+13] & 0x0000ffff); |
| 1705 | drdy = (rdp_cmd_data[shade_base+10] & 0xffff0000) | ((rdp_cmd_data[shade_base+14] >> 16) & 0x0000ffff); |
| 1706 | dgdy = ((rdp_cmd_data[shade_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+14] & 0x0000ffff); |
| 1707 | dbdy = (rdp_cmd_data[shade_base+11] & 0xffff0000) | ((rdp_cmd_data[shade_base+15] >> 16) & 0x0000ffff); |
| 1708 | dady = ((rdp_cmd_data[shade_base+11] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+15] & 0x0000ffff); |
| 1709 | z = rdp_cmd_data[zbuffer_base+0]; |
| 1710 | dzdx = rdp_cmd_data[zbuffer_base+1]; |
| 1711 | dzde = rdp_cmd_data[zbuffer_base+2]; |
| 1712 | dzdy = rdp_cmd_data[zbuffer_base+3]; |
| 1713 | |
| 1714 | span[0].dr.w = drdx & ~0x1f; |
| 1715 | span[0].dg.w = dgdx & ~0x1f; |
| 1716 | span[0].db.w = dbdx & ~0x1f; |
| 1717 | span[0].da.w = dadx & ~0x1f; |
| 1718 | span[0].dz.w = dzdx; |
| 1719 | dzdy_dz = (dzdy >> 16) & 0xffff; |
| 1720 | dzdx_dz = (dzdx >> 16) & 0xffff; |
| 1721 | span[0].dzpix = ((dzdy_dz & 0x8000) ? ((~dzdy_dz) & 0x7fff) : dzdy_dz) + ((dzdx_dz & 0x8000) ? ((~dzdx_dz) & 0x7fff) : dzdx_dz); |
| 1722 | span[0].dzpix = normalize_dzpix(span[0].dzpix); |
| 1723 | |
| 1724 | xleft_inc = dxmdy >> 2; |
| 1725 | xright_inc = dxhdy >> 2; |
| 1726 | |
| 1727 | xright = xh; |
| 1728 | xleft = xm; |
| 1729 | |
| 1730 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 1731 | if (limcvg < 0) |
| 1732 | { |
| 1733 | limcvg = 0; |
| 1734 | } |
| 1735 | |
| 1736 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 1737 | for (k = startcvg; k <= limcvg; k++) |
| 1738 | { |
| 1739 | memset((void*)&span[k].cvg[0],0,640); |
| 1740 | } |
| 1741 | |
| 1742 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 1743 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 1744 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 1745 | |
| 1746 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 1747 | |
| 1748 | do_offset = !(sign_dxhdy ^ (flip)); |
| 1749 | |
| 1750 | if (do_offset) |
| 1751 | { |
| 1752 | drdeh = drde >> 9; drdyh = drdy >> 9; |
| 1753 | dgdeh = dgde >> 9; dgdyh = dgdy >> 9; |
| 1754 | dbdeh = dbde >> 9; dbdyh = dbdy >> 9; |
| 1755 | dadeh = dade >> 9; dadyh = dady >> 9; |
| 1756 | dzdeh = dzde >> 9; dzdyh = dzdy >> 9; |
| 1757 | |
| 1758 | drdiff = (drdeh*3 - drdyh*3) << 7; |
| 1759 | dgdiff = (dgdeh*3 - dgdyh*3) << 7; |
| 1760 | dbdiff = (dbdeh*3 - dbdyh*3) << 7; |
| 1761 | dadiff = (dadeh*3 - dadyh*3) << 7; |
| 1762 | dzdiff = (dzdeh*3 - dzdyh*3) << 7; |
| 1763 | } |
| 1764 | else |
| 1765 | { |
| 1766 | drdiff = dgdiff = dbdiff = dadiff = dzdiff = 0; |
| 1767 | } |
| 1768 | |
| 1769 | if (do_offset) |
| 1770 | { |
| 1771 | dreoff = (drdeh*3) << 7; |
| 1772 | dgeoff = (dgdeh*3) << 7; |
| 1773 | dbeoff = (dbdeh*3) << 7; |
| 1774 | daeoff = (dadeh*3) << 7; |
| 1775 | dzeoff = (dzdeh*3) << 7; |
| 1776 | } |
| 1777 | else |
| 1778 | { |
| 1779 | dreoff = dgeoff = dbeoff = daeoff = dzeoff = 0; |
| 1780 | } |
| 1781 | |
| 1782 | #define addleft(x) addleftcvg(x,k) |
| 1783 | #define addright(x) addrightcvg(x,k) |
| 1784 | #define setvalues() { \ |
| 1785 | addvalues(); \ |
| 1786 | adjust_attr(); \ |
| 1787 | } |
| 1788 | |
| 1789 | drdxh = drdx >> 8; |
| 1790 | dgdxh = dgdx >> 8; |
| 1791 | dbdxh = dbdx >> 8; |
| 1792 | dadxh = dadx >> 8; |
| 1793 | dzdxh = dzdx >> 8; |
| 1794 | |
| 1795 | #define adjust_attr() \ |
| 1796 | { \ |
| 1797 | span[j].r.w = r + drdiff - (xfrac * drdxh); \ |
| 1798 | span[j].g.w = g + dgdiff - (xfrac * dgdxh); \ |
| 1799 | span[j].b.w = b + dbdiff - (xfrac * dbdxh); \ |
| 1800 | span[j].a.w = a + dadiff - (xfrac * dadxh); \ |
| 1801 | span[j].z.w = z + dzdiff - (xfrac * dzdxh); \ |
| 1802 | } |
| 1803 | |
| 1804 | #define addvalues() { \ |
| 1805 | r += drde; \ |
| 1806 | g += dgde; \ |
| 1807 | b += dbde; \ |
| 1808 | a += dade; \ |
| 1809 | z += dzde; \ |
| 1810 | } |
| 1811 | |
| 1812 | m_inc = flip ? 1 : -1; |
| 1813 | |
| 1814 | ycur = yh & ~3; |
| 1815 | ylfar = yl | 3; |
| 1816 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 1817 | yhpix = yh >> 2; |
| 1818 | ympix = ym >> 2; |
| 1819 | ylpix = yl >> 2; |
| 1820 | |
| 1821 | for (k = ycur; k <= ylfar; k++) |
| 1822 | { |
| 1823 | if (k == ym) |
| 1824 | { |
| 1825 | xleft = xl; |
| 1826 | xleft_inc = dxldy >> 2; |
| 1827 | } |
| 1828 | |
| 1829 | xstart = xleft >> 16; |
| 1830 | xend = xright >> 16; |
| 1831 | j = k >> 2; |
| 1832 | spix = k & 3; |
| 1833 | |
| 1834 | if (k >= 0 && k < 0x1000) |
| 1835 | { |
| 1836 | int m = 0; |
| 1837 | int n = 0; |
| 1838 | int length = 0; |
| 1839 | min = 0; max = 3; |
| 1840 | if (j == yhpix) |
| 1841 | { |
| 1842 | min = yh & 3; |
| 1843 | } |
| 1844 | if (j == ylpix) |
| 1845 | { |
| 1846 | max = yl & 3; |
| 1847 | } |
| 1848 | if (spix >= min && spix <= max) |
| 1849 | { |
| 1850 | if (spix == min) |
| 1851 | { |
| 1852 | minxmx = maxxmx = xstart; |
| 1853 | minxhx = maxxhx = xend; |
| 1854 | } |
| 1855 | else |
| 1856 | { |
| 1857 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 1858 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 1859 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 1860 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 1861 | } |
| 1862 | } |
| 1863 | |
| 1864 | if (spix == max) |
| 1865 | { |
| 1866 | if (flip) |
| 1867 | { |
| 1868 | span[j].lx = maxxmx; |
| 1869 | span[j].rx = minxhx; |
| 1870 | } |
| 1871 | else |
| 1872 | { |
| 1873 | span[j].lx = minxmx; |
| 1874 | span[j].rx = maxxhx; |
| 1875 | } |
| 1876 | } |
| 1877 | |
| 1878 | length = flip ? (xstart - xend) : (xend - xstart); |
| 1879 | |
| 1880 | if (spix == ldflag) |
| 1881 | { |
| 1882 | xfrac = ((xright >> 8) & 0xff); |
| 1883 | adjust_attr(); |
| 1884 | } |
| 1885 | |
| 1886 | m = flip ? (xend+1) : (xend-1); |
| 1887 | |
| 1888 | if (k >= yh && length >= 0 && k <= yl) |
| 1889 | { |
| 1890 | if (xstart>=0 && xstart <1024) |
| 1891 | { |
| 1892 | if (!flip) |
| 1893 | { |
| 1894 | span[j].cvg[xstart] += addleft(xleft); |
| 1895 | } |
| 1896 | else |
| 1897 | { |
| 1898 | span[j].cvg[xstart] += addright(xleft); |
| 1899 | } |
| 1900 | } |
| 1901 | if (xend>=0 && xend<1024) |
| 1902 | { |
| 1903 | if (xstart != xend) |
| 1904 | { |
| 1905 | if (!flip) |
| 1906 | { |
| 1907 | span[j].cvg[xend] += addright(xright); |
| 1908 | } |
| 1909 | else |
| 1910 | { |
| 1911 | span[j].cvg[xend] += addleft(xright); |
| 1912 | } |
| 1913 | } |
| 1914 | else |
| 1915 | { |
| 1916 | if (!flip) |
| 1917 | { |
| 1918 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 1919 | } |
| 1920 | else |
| 1921 | { |
| 1922 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 1923 | } |
| 1924 | if (span[j].cvg[xend] > 200) |
| 1925 | { |
| 1926 | span[j].cvg[xend] = 0; |
| 1927 | } |
| 1928 | } |
| 1929 | } |
| 1930 | for (n = 0; n < (length - 1); n++) |
| 1931 | { |
| 1932 | if (m>=0 && m < 640) |
| 1933 | { |
| 1934 | span[j].cvg[m] += 2; |
| 1935 | } |
| 1936 | |
| 1937 | m += m_inc; |
| 1938 | } |
| 1939 | } |
| 1940 | } |
| 1941 | |
| 1942 | if (spix == 3) |
| 1943 | { |
| 1944 | addvalues(); |
| 1945 | } |
| 1946 | xleft += xleft_inc; |
| 1947 | xright += xright_inc; |
| 1948 | } |
| 1949 | |
| 1950 | switch (fb_size) // 8bpp needs to be implemented |
| 1951 | { |
| 1952 | case PIXEL_SIZE_16BIT: |
| 1953 | if(flip) |
| 1954 | { |
| 1955 | render_spans_16_s_nt_z_f(yh>>2, yl>>2, &tile[0]); break; |
| 1956 | } |
| 1957 | else |
| 1958 | { |
| 1959 | render_spans_16_s_nt_z_nf(yh>>2, yl>>2, &tile[0]); break; |
| 1960 | } |
| 1961 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 1, 0, 1, flip); break; |
| 1962 | default: break; // V-Rally2 does this, fb_size=0 |
| 1963 | } |
| 1964 | } |
| 1965 | #undef addvalues |
| 1966 | #undef adjust_attr |
| 1967 | #undef addleft |
| 1968 | #undef addright |
| 1969 | #undef setvalues |
| 1970 | |
| 1971 | static void triangle_s_t_nz(UINT32 w1, UINT32 w2) |
| 1972 | { |
| 1973 | int j; |
| 1974 | int xleft, xright, xleft_inc, xright_inc; |
| 1975 | int xstart, xend; |
| 1976 | int r = 0, g = 0, b = 0, a = 0, s = 0, t = 0, w = 0; |
| 1977 | int dr, dg, db, da; |
| 1978 | int drdx = 0, dgdx = 0, dbdx = 0, dadx = 0, dsdx = 0, dtdx = 0, dwdx = 0; |
| 1979 | int drdy = 0, dgdy = 0, dbdy = 0, dady = 0, dsdy = 0, dtdy = 0, dwdy = 0; |
| 1980 | int drde = 0, dgde = 0, dbde = 0, dade = 0, dsde = 0, dtde = 0, dwde = 0; |
| 1981 | int tilenum; |
| 1982 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 1983 | |
| 1984 | INT32 yl, ym, yh; |
| 1985 | INT32 xl, xm, xh; |
| 1986 | INT32 dxldy, dxhdy, dxmdy; |
| 1987 | int dsdylod, dtdylod; |
| 1988 | UINT32 w3, w4, w5, w6, w7, w8; |
| 1989 | |
| 1990 | int k = 0; |
| 1991 | |
| 1992 | INT32 limcvg = 0; |
| 1993 | INT32 startcvg = 0; |
| 1994 | |
| 1995 | int sign_dxldy = 0; |
| 1996 | int sign_dxmdy = 0; |
| 1997 | int samesign = 0; |
| 1998 | |
| 1999 | int dsdiff = 0, dtdiff = 0, dwdiff = 0, drdiff = 0, dgdiff = 0, dbdiff = 0, dadiff = 0; |
| 2000 | int sign_dxhdy = 0; |
| 2001 | |
| 2002 | int dsdeh = 0, dtdeh = 0, dwdeh = 0, drdeh = 0, dgdeh = 0, dbdeh = 0, dadeh = 0, dsdyh = 0, dtdyh = 0, dwdyh = 0, drdyh = 0, dgdyh = 0, dbdyh = 0, dadyh = 0; |
| 2003 | int do_offset = 0; |
| 2004 | |
| 2005 | int xfrac = 0; |
| 2006 | int dseoff = 0, dteoff = 0, dweoff = 0, dreoff = 0, dgeoff = 0, dbeoff = 0, daeoff = 0; |
| 2007 | |
| 2008 | int dsdxh = 0, dtdxh = 0, dwdxh = 0, drdxh = 0, dgdxh = 0, dbdxh = 0, dadxh = 0; |
| 2009 | |
| 2010 | int m_inc; |
| 2011 | UINT32 min=0, max=3; |
| 2012 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 2013 | |
| 2014 | int spix = 0; // Current subpixel |
| 2015 | int ycur; |
| 2016 | int ylfar; |
| 2017 | int ldflag; |
| 2018 | int yhpix; |
| 2019 | int ympix; |
| 2020 | int ylpix; |
| 2021 | |
| 2022 | int shade_base = rdp_cmd_cur + 8; |
| 2023 | int texture_base = rdp_cmd_cur + 8; |
| 2024 | |
| 2025 | texture_base += 16; |
| 2026 | |
| 2027 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 2028 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 2029 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 2030 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 2031 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 2032 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 2033 | |
| 2034 | yl = (w1 & 0x3fff); |
| 2035 | ym = ((w2 >> 16) & 0x3fff); |
| 2036 | yh = ((w2 >> 0) & 0x3fff); |
| 2037 | xl = (INT32)(w3 & 0x3fffffff); |
| 2038 | xh = (INT32)(w5 & 0x3fffffff); |
| 2039 | xm = (INT32)(w7 & 0x3fffffff); |
| 2040 | // Inverse slopes in 16.16 format |
| 2041 | dxldy = (INT32)(w4); |
| 2042 | dxhdy = (INT32)(w6); |
| 2043 | dxmdy = (INT32)(w8); |
| 2044 | |
| 2045 | max_level = ((w1 >> 19) & 7); |
| 2046 | tilenum = (w1 >> 16) & 0x7; |
| 2047 | |
| 2048 | if (yl & 0x2000) yl |= 0xffffc000; |
| 2049 | if (ym & 0x2000) ym |= 0xffffc000; |
| 2050 | if (yh & 0x2000) yh |= 0xffffc000; |
| 2051 | |
| 2052 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 2053 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 2054 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 2055 | |
| 2056 | s = 0; t = 0; w = 0; |
| 2057 | dr = 0; dg = 0; db = 0; da = 0; |
| 2058 | |
| 2059 | r = (rdp_cmd_data[shade_base+0 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+4 ] >> 16) & 0x0000ffff); |
| 2060 | g = ((rdp_cmd_data[shade_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+4 ] & 0x0000ffff); |
| 2061 | b = (rdp_cmd_data[shade_base+1 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+5 ] >> 16) & 0x0000ffff); |
| 2062 | a = ((rdp_cmd_data[shade_base+1 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+5 ] & 0x0000ffff); |
| 2063 | drdx = (rdp_cmd_data[shade_base+2 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+6 ] >> 16) & 0x0000ffff); |
| 2064 | dgdx = ((rdp_cmd_data[shade_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+6 ] & 0x0000ffff); |
| 2065 | dbdx = (rdp_cmd_data[shade_base+3 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+7 ] >> 16) & 0x0000ffff); |
| 2066 | dadx = ((rdp_cmd_data[shade_base+3 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+7 ] & 0x0000ffff); |
| 2067 | drde = (rdp_cmd_data[shade_base+8 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+12] >> 16) & 0x0000ffff); |
| 2068 | dgde = ((rdp_cmd_data[shade_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+12] & 0x0000ffff); |
| 2069 | dbde = (rdp_cmd_data[shade_base+9 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+13] >> 16) & 0x0000ffff); |
| 2070 | dade = ((rdp_cmd_data[shade_base+9 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+13] & 0x0000ffff); |
| 2071 | drdy = (rdp_cmd_data[shade_base+10] & 0xffff0000) | ((rdp_cmd_data[shade_base+14] >> 16) & 0x0000ffff); |
| 2072 | dgdy = ((rdp_cmd_data[shade_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+14] & 0x0000ffff); |
| 2073 | dbdy = (rdp_cmd_data[shade_base+11] & 0xffff0000) | ((rdp_cmd_data[shade_base+15] >> 16) & 0x0000ffff); |
| 2074 | dady = ((rdp_cmd_data[shade_base+11] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+15] & 0x0000ffff); |
| 2075 | s = (rdp_cmd_data[texture_base+0 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+4 ] >> 16) & 0x0000ffff); |
| 2076 | t = ((rdp_cmd_data[texture_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+4 ] & 0x0000ffff); |
| 2077 | w = (rdp_cmd_data[texture_base+1 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+5 ] >> 16) & 0x0000ffff); |
| 2078 | dsdx = (rdp_cmd_data[texture_base+2 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+6 ] >> 16) & 0x0000ffff); |
| 2079 | dtdx = ((rdp_cmd_data[texture_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+6 ] & 0x0000ffff); |
| 2080 | dwdx = (rdp_cmd_data[texture_base+3 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+7 ] >> 16) & 0x0000ffff); |
| 2081 | dsde = (rdp_cmd_data[texture_base+8 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+12] >> 16) & 0x0000ffff); |
| 2082 | dtde = ((rdp_cmd_data[texture_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+12] & 0x0000ffff); |
| 2083 | dwde = (rdp_cmd_data[texture_base+9 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+13] >> 16) & 0x0000ffff); |
| 2084 | dsdy = (rdp_cmd_data[texture_base+10] & 0xffff0000) | ((rdp_cmd_data[texture_base+14] >> 16) & 0x0000ffff); |
| 2085 | dtdy = ((rdp_cmd_data[texture_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+14] & 0x0000ffff); |
| 2086 | dwdy = (rdp_cmd_data[texture_base+11] & 0xffff0000) | ((rdp_cmd_data[texture_base+15] >> 16) & 0x0000ffff); |
| 2087 | |
| 2088 | span[0].ds.w = dsdx; |
| 2089 | span[0].dt.w = dtdx; |
| 2090 | span[0].dw.w = dwdx; |
| 2091 | span[0].dr.w = drdx & ~0x1f; |
| 2092 | span[0].dg.w = dgdx & ~0x1f; |
| 2093 | span[0].db.w = dbdx & ~0x1f; |
| 2094 | span[0].da.w = dadx & ~0x1f; |
| 2095 | dsdylod = dsdy >> 16; |
| 2096 | dtdylod = dtdy >> 16; |
| 2097 | if (dsdylod & 0x20000) |
| 2098 | { |
| 2099 | dsdylod = ~dsdylod & 0x1ffff; |
| 2100 | } |
| 2101 | if (dtdylod & 0x20000) |
| 2102 | { |
| 2103 | dtdylod = ~dtdylod & 0x1ffff; |
| 2104 | } |
| 2105 | span[0].dymax = (dsdylod > dtdylod)? dsdylod : dtdylod; |
| 2106 | |
| 2107 | xleft_inc = dxmdy >> 2; |
| 2108 | xright_inc = dxhdy >> 2; |
| 2109 | |
| 2110 | xright = xh; |
| 2111 | xleft = xm; |
| 2112 | |
| 2113 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 2114 | if (limcvg < 0) |
| 2115 | { |
| 2116 | limcvg = 0; |
| 2117 | } |
| 2118 | |
| 2119 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 2120 | for (k = startcvg; k <= limcvg; k++) |
| 2121 | { |
| 2122 | memset((void*)&span[k].cvg[0],0,640); |
| 2123 | } |
| 2124 | |
| 2125 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 2126 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 2127 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 2128 | |
| 2129 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 2130 | |
| 2131 | do_offset = !(sign_dxhdy ^ (flip)); |
| 2132 | |
| 2133 | if (do_offset) |
| 2134 | { |
| 2135 | dsdeh = dsde >> 9; dsdyh = dsdy >> 9; |
| 2136 | dtdeh = dtde >> 9; dtdyh = dtdy >> 9; |
| 2137 | dwdeh = dwde >> 9; dwdyh = dwdy >> 9; |
| 2138 | drdeh = drde >> 9; drdyh = drdy >> 9; |
| 2139 | dgdeh = dgde >> 9; dgdyh = dgdy >> 9; |
| 2140 | dbdeh = dbde >> 9; dbdyh = dbdy >> 9; |
| 2141 | dadeh = dade >> 9; dadyh = dady >> 9; |
| 2142 | |
| 2143 | dsdiff = (dsdeh*3 - dsdyh*3) << 7; |
| 2144 | dtdiff = (dtdeh*3 - dtdyh*3) << 7; |
| 2145 | dwdiff = (dwdeh*3 - dwdyh*3) << 7; |
| 2146 | drdiff = (drdeh*3 - drdyh*3) << 7; |
| 2147 | dgdiff = (dgdeh*3 - dgdyh*3) << 7; |
| 2148 | dbdiff = (dbdeh*3 - dbdyh*3) << 7; |
| 2149 | dadiff = (dadeh*3 - dadyh*3) << 7; |
| 2150 | } |
| 2151 | else |
| 2152 | { |
| 2153 | dsdiff = dtdiff = dwdiff = drdiff = dgdiff = dbdiff = dadiff = 0; |
| 2154 | } |
| 2155 | |
| 2156 | if (do_offset) |
| 2157 | { |
| 2158 | dseoff = (dsdeh*3) << 7; |
| 2159 | dteoff = (dtdeh*3) << 7; |
| 2160 | dweoff = (dwdeh*3) << 7; |
| 2161 | dreoff = (drdeh*3) << 7; |
| 2162 | dgeoff = (dgdeh*3) << 7; |
| 2163 | dbeoff = (dbdeh*3) << 7; |
| 2164 | daeoff = (dadeh*3) << 7; |
| 2165 | } |
| 2166 | else |
| 2167 | { |
| 2168 | dseoff = dteoff = dweoff = dreoff = dgeoff = dbeoff = daeoff = 0; |
| 2169 | } |
| 2170 | |
| 2171 | #define addleft(x) addleftcvg(x,k) |
| 2172 | #define addright(x) addrightcvg(x,k) |
| 2173 | #define setvalues() { \ |
| 2174 | addvalues(); \ |
| 2175 | adjust_attr(); \ |
| 2176 | } |
| 2177 | |
| 2178 | dsdxh = dsdx >> 8; |
| 2179 | dtdxh = dtdx >> 8; |
| 2180 | dwdxh = dwdx >> 8; |
| 2181 | drdxh = drdx >> 8; |
| 2182 | dgdxh = dgdx >> 8; |
| 2183 | dbdxh = dbdx >> 8; |
| 2184 | dadxh = dadx >> 8; |
| 2185 | |
| 2186 | #define adjust_attr() \ |
| 2187 | { \ |
| 2188 | span[j].s.w = (s + dsdiff - (xfrac * dsdxh)) & ~0x1f; \ |
| 2189 | span[j].t.w = (t + dtdiff - (xfrac * dtdxh)) & ~0x1f; \ |
| 2190 | span[j].w.w = (w + dwdiff - (xfrac * dwdxh)) & ~0x1f; \ |
| 2191 | span[j].r.w = r + drdiff - (xfrac * drdxh); \ |
| 2192 | span[j].g.w = g + dgdiff - (xfrac * dgdxh); \ |
| 2193 | span[j].b.w = b + dbdiff - (xfrac * dbdxh); \ |
| 2194 | span[j].a.w = a + dadiff - (xfrac * dadxh); \ |
| 2195 | } |
| 2196 | |
| 2197 | #define addvalues() { \ |
| 2198 | s += dsde; \ |
| 2199 | t += dtde; \ |
| 2200 | w += dwde; \ |
| 2201 | r += drde; \ |
| 2202 | g += dgde; \ |
| 2203 | b += dbde; \ |
| 2204 | a += dade; \ |
| 2205 | } |
| 2206 | |
| 2207 | m_inc = flip ? 1 : -1; |
| 2208 | |
| 2209 | ycur = yh & ~3; |
| 2210 | ylfar = yl | 3; |
| 2211 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 2212 | yhpix = yh >> 2; |
| 2213 | ympix = ym >> 2; |
| 2214 | ylpix = yl >> 2; |
| 2215 | |
| 2216 | for (k = ycur; k <= ylfar; k++) |
| 2217 | { |
| 2218 | if (k == ym) |
| 2219 | { |
| 2220 | xleft = xl; |
| 2221 | xleft_inc = dxldy >> 2; |
| 2222 | } |
| 2223 | |
| 2224 | xstart = xleft >> 16; |
| 2225 | xend = xright >> 16; |
| 2226 | j = k >> 2; |
| 2227 | spix = k & 3; |
| 2228 | |
| 2229 | if (k >= 0 && k < 0x1000) |
| 2230 | { |
| 2231 | int m = 0; |
| 2232 | int n = 0; |
| 2233 | int length = 0; |
| 2234 | min = 0; max = 3; |
| 2235 | if (j == yhpix) |
| 2236 | { |
| 2237 | min = yh & 3; |
| 2238 | } |
| 2239 | if (j == ylpix) |
| 2240 | { |
| 2241 | max = yl & 3; |
| 2242 | } |
| 2243 | if (spix >= min && spix <= max) |
| 2244 | { |
| 2245 | if (spix == min) |
| 2246 | { |
| 2247 | minxmx = maxxmx = xstart; |
| 2248 | minxhx = maxxhx = xend; |
| 2249 | } |
| 2250 | else |
| 2251 | { |
| 2252 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 2253 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 2254 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 2255 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 2256 | } |
| 2257 | } |
| 2258 | |
| 2259 | if (spix == max) |
| 2260 | { |
| 2261 | if (flip) |
| 2262 | { |
| 2263 | span[j].lx = maxxmx; |
| 2264 | span[j].rx = minxhx; |
| 2265 | } |
| 2266 | else |
| 2267 | { |
| 2268 | span[j].lx = minxmx; |
| 2269 | span[j].rx = maxxhx; |
| 2270 | } |
| 2271 | } |
| 2272 | |
| 2273 | length = flip ? (xstart - xend) : (xend - xstart); |
| 2274 | |
| 2275 | if (spix == ldflag) |
| 2276 | { |
| 2277 | xfrac = ((xright >> 8) & 0xff); |
| 2278 | adjust_attr(); |
| 2279 | } |
| 2280 | |
| 2281 | m = flip ? (xend+1) : (xend-1); |
| 2282 | |
| 2283 | if (k >= yh && length >= 0 && k <= yl) |
| 2284 | { |
| 2285 | if (xstart>=0 && xstart <1024) |
| 2286 | { |
| 2287 | if (!flip) |
| 2288 | { |
| 2289 | span[j].cvg[xstart] += addleft(xleft); |
| 2290 | } |
| 2291 | else |
| 2292 | { |
| 2293 | span[j].cvg[xstart] += addright(xleft); |
| 2294 | } |
| 2295 | } |
| 2296 | if (xend>=0 && xend<1024) |
| 2297 | { |
| 2298 | if (xstart != xend) |
| 2299 | { |
| 2300 | if (!flip) |
| 2301 | { |
| 2302 | span[j].cvg[xend] += addright(xright); |
| 2303 | } |
| 2304 | else |
| 2305 | { |
| 2306 | span[j].cvg[xend] += addleft(xright); |
| 2307 | } |
| 2308 | } |
| 2309 | else |
| 2310 | { |
| 2311 | if (!flip) |
| 2312 | { |
| 2313 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 2314 | } |
| 2315 | else |
| 2316 | { |
| 2317 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 2318 | } |
| 2319 | if (span[j].cvg[xend] > 200) |
| 2320 | { |
| 2321 | span[j].cvg[xend] = 0; |
| 2322 | } |
| 2323 | } |
| 2324 | } |
| 2325 | for (n = 0; n < (length - 1); n++) |
| 2326 | { |
| 2327 | if (m>=0 && m < 640) |
| 2328 | { |
| 2329 | span[j].cvg[m] += 2; |
| 2330 | } |
| 2331 | |
| 2332 | m += m_inc; |
| 2333 | } |
| 2334 | } |
| 2335 | } |
| 2336 | |
| 2337 | if (spix == 3) |
| 2338 | { |
| 2339 | addvalues(); |
| 2340 | } |
| 2341 | xleft += xleft_inc; |
| 2342 | xright += xright_inc; |
| 2343 | } |
| 2344 | |
| 2345 | switch (fb_size) // 8bpp needs to be implemented |
| 2346 | { |
| 2347 | case PIXEL_SIZE_16BIT: |
| 2348 | if(flip) |
| 2349 | { |
| 2350 | render_spans_16_s_t_nz_f(yh>>2, yl>>2, &tile[0]); break; |
| 2351 | } |
| 2352 | else |
| 2353 | { |
| 2354 | render_spans_16_s_t_nz_nf(yh>>2, yl>>2, &tile[0]); break; |
| 2355 | } |
| 2356 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 1, 1, 0, flip); break; |
| 2357 | default: break; // V-Rally2 does this, fb_size=0 |
| 2358 | } |
| 2359 | } |
| 2360 | #undef addvalues |
| 2361 | #undef adjust_attr |
| 2362 | #undef addleft |
| 2363 | #undef addright |
| 2364 | #undef setvalues |
| 2365 | |
| 2366 | static void triangle_s_t_z(UINT32 w1, UINT32 w2) |
| 2367 | { |
| 2368 | int j; |
| 2369 | int xleft, xright, xleft_inc, xright_inc; |
| 2370 | int xstart, xend; |
| 2371 | int r = 0, g = 0, b = 0, a = 0, z = 0, s = 0, t = 0, w = 0; |
| 2372 | int dr, dg, db, da; |
| 2373 | int drdx = 0, dgdx = 0, dbdx = 0, dadx = 0, dzdx = 0, dsdx = 0, dtdx = 0, dwdx = 0; |
| 2374 | int drdy = 0, dgdy = 0, dbdy = 0, dady = 0, dzdy = 0, dsdy = 0, dtdy = 0, dwdy = 0; |
| 2375 | int drde = 0, dgde = 0, dbde = 0, dade = 0, dzde = 0, dsde = 0, dtde = 0, dwde = 0; |
| 2376 | int tilenum; |
| 2377 | int flip = (w1 & 0x800000) ? 1 : 0; |
| 2378 | |
| 2379 | INT32 yl, ym, yh; |
| 2380 | INT32 xl, xm, xh; |
| 2381 | INT32 dxldy, dxhdy, dxmdy; |
| 2382 | int dzdy_dz, dzdx_dz; |
| 2383 | int dsdylod, dtdylod; |
| 2384 | UINT32 w3, w4, w5, w6, w7, w8; |
| 2385 | |
| 2386 | int k = 0; |
| 2387 | |
| 2388 | INT32 limcvg = 0; |
| 2389 | INT32 startcvg = 0; |
| 2390 | |
| 2391 | int sign_dxldy = 0; |
| 2392 | int sign_dxmdy = 0; |
| 2393 | int samesign = 0; |
| 2394 | |
| 2395 | int dsdiff = 0, dtdiff = 0, dwdiff = 0, drdiff = 0, dgdiff = 0, dbdiff = 0, dadiff = 0, dzdiff = 0; |
| 2396 | int sign_dxhdy = 0; |
| 2397 | |
| 2398 | int dsdeh = 0, dtdeh = 0, dwdeh = 0, drdeh = 0, dgdeh = 0, dbdeh = 0, dadeh = 0, dzdeh = 0, dsdyh = 0, dtdyh = 0, dwdyh = 0, drdyh = 0, dgdyh = 0, dbdyh = 0, dadyh = 0, dzdyh = 0; |
| 2399 | int do_offset = 0; |
| 2400 | |
| 2401 | int xfrac = 0; |
| 2402 | int dseoff = 0, dteoff = 0, dweoff = 0, dreoff = 0, dgeoff = 0, dbeoff = 0, daeoff = 0, dzeoff = 0; |
| 2403 | |
| 2404 | int dsdxh = 0, dtdxh = 0, dwdxh = 0, drdxh = 0, dgdxh = 0, dbdxh = 0, dadxh = 0, dzdxh = 0; |
| 2405 | |
| 2406 | int m_inc; |
| 2407 | UINT32 min=0, max=3; |
| 2408 | INT32 maxxmx = 0, minxmx = 0, maxxhx = 0, minxhx = 0; |
| 2409 | |
| 2410 | int spix = 0; // Current subpixel |
| 2411 | int ycur; |
| 2412 | int ylfar; |
| 2413 | int ldflag; |
| 2414 | int yhpix; |
| 2415 | int ympix; |
| 2416 | int ylpix; |
| 2417 | |
| 2418 | int shade_base = rdp_cmd_cur + 8; |
| 2419 | int texture_base = rdp_cmd_cur + 24; |
| 2420 | int zbuffer_base = rdp_cmd_cur + 40; |
| 2421 | |
| 2422 | w3 = rdp_cmd_data[rdp_cmd_cur+2]; |
| 2423 | w4 = rdp_cmd_data[rdp_cmd_cur+3]; |
| 2424 | w5 = rdp_cmd_data[rdp_cmd_cur+4]; |
| 2425 | w6 = rdp_cmd_data[rdp_cmd_cur+5]; |
| 2426 | w7 = rdp_cmd_data[rdp_cmd_cur+6]; |
| 2427 | w8 = rdp_cmd_data[rdp_cmd_cur+7]; |
| 2428 | |
| 2429 | yl = (w1 & 0x3fff); |
| 2430 | ym = ((w2 >> 16) & 0x3fff); |
| 2431 | yh = ((w2 >> 0) & 0x3fff); |
| 2432 | xl = (INT32)(w3 & 0x3fffffff); |
| 2433 | xh = (INT32)(w5 & 0x3fffffff); |
| 2434 | xm = (INT32)(w7 & 0x3fffffff); |
| 2435 | // Inverse slopes in 16.16 format |
| 2436 | dxldy = (INT32)(w4); |
| 2437 | dxhdy = (INT32)(w6); |
| 2438 | dxmdy = (INT32)(w8); |
| 2439 | |
| 2440 | max_level = ((w1 >> 19) & 7); |
| 2441 | tilenum = (w1 >> 16) & 0x7; |
| 2442 | |
| 2443 | if (yl & 0x2000) yl |= 0xffffc000; |
| 2444 | if (ym & 0x2000) ym |= 0xffffc000; |
| 2445 | if (yh & 0x2000) yh |= 0xffffc000; |
| 2446 | |
| 2447 | if (xl & 0x20000000) xl |= 0xc0000000; |
| 2448 | if (xm & 0x20000000) xm |= 0xc0000000; |
| 2449 | if (xh & 0x20000000) xh |= 0xc0000000; |
| 2450 | |
| 2451 | z = 0; |
| 2452 | s = 0; t = 0; w = 0; |
| 2453 | dr = 0; dg = 0; db = 0; da = 0; |
| 2454 | |
| 2455 | r = (rdp_cmd_data[shade_base+0 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+4 ] >> 16) & 0x0000ffff); |
| 2456 | g = ((rdp_cmd_data[shade_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+4 ] & 0x0000ffff); |
| 2457 | b = (rdp_cmd_data[shade_base+1 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+5 ] >> 16) & 0x0000ffff); |
| 2458 | a = ((rdp_cmd_data[shade_base+1 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+5 ] & 0x0000ffff); |
| 2459 | drdx = (rdp_cmd_data[shade_base+2 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+6 ] >> 16) & 0x0000ffff); |
| 2460 | dgdx = ((rdp_cmd_data[shade_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+6 ] & 0x0000ffff); |
| 2461 | dbdx = (rdp_cmd_data[shade_base+3 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+7 ] >> 16) & 0x0000ffff); |
| 2462 | dadx = ((rdp_cmd_data[shade_base+3 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+7 ] & 0x0000ffff); |
| 2463 | drde = (rdp_cmd_data[shade_base+8 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+12] >> 16) & 0x0000ffff); |
| 2464 | dgde = ((rdp_cmd_data[shade_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+12] & 0x0000ffff); |
| 2465 | dbde = (rdp_cmd_data[shade_base+9 ] & 0xffff0000) | ((rdp_cmd_data[shade_base+13] >> 16) & 0x0000ffff); |
| 2466 | dade = ((rdp_cmd_data[shade_base+9 ] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+13] & 0x0000ffff); |
| 2467 | drdy = (rdp_cmd_data[shade_base+10] & 0xffff0000) | ((rdp_cmd_data[shade_base+14] >> 16) & 0x0000ffff); |
| 2468 | dgdy = ((rdp_cmd_data[shade_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+14] & 0x0000ffff); |
| 2469 | dbdy = (rdp_cmd_data[shade_base+11] & 0xffff0000) | ((rdp_cmd_data[shade_base+15] >> 16) & 0x0000ffff); |
| 2470 | dady = ((rdp_cmd_data[shade_base+11] << 16) & 0xffff0000) | (rdp_cmd_data[shade_base+15] & 0x0000ffff); |
| 2471 | s = (rdp_cmd_data[texture_base+0 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+4 ] >> 16) & 0x0000ffff); |
| 2472 | t = ((rdp_cmd_data[texture_base+0 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+4 ] & 0x0000ffff); |
| 2473 | w = (rdp_cmd_data[texture_base+1 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+5 ] >> 16) & 0x0000ffff); |
| 2474 | dsdx = (rdp_cmd_data[texture_base+2 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+6 ] >> 16) & 0x0000ffff); |
| 2475 | dtdx = ((rdp_cmd_data[texture_base+2 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+6 ] & 0x0000ffff); |
| 2476 | dwdx = (rdp_cmd_data[texture_base+3 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+7 ] >> 16) & 0x0000ffff); |
| 2477 | dsde = (rdp_cmd_data[texture_base+8 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+12] >> 16) & 0x0000ffff); |
| 2478 | dtde = ((rdp_cmd_data[texture_base+8 ] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+12] & 0x0000ffff); |
| 2479 | dwde = (rdp_cmd_data[texture_base+9 ] & 0xffff0000) | ((rdp_cmd_data[texture_base+13] >> 16) & 0x0000ffff); |
| 2480 | dsdy = (rdp_cmd_data[texture_base+10] & 0xffff0000) | ((rdp_cmd_data[texture_base+14] >> 16) & 0x0000ffff); |
| 2481 | dtdy = ((rdp_cmd_data[texture_base+10] << 16) & 0xffff0000) | (rdp_cmd_data[texture_base+14] & 0x0000ffff); |
| 2482 | dwdy = (rdp_cmd_data[texture_base+11] & 0xffff0000) | ((rdp_cmd_data[texture_base+15] >> 16) & 0x0000ffff); |
| 2483 | z = rdp_cmd_data[zbuffer_base+0]; |
| 2484 | dzdx = rdp_cmd_data[zbuffer_base+1]; |
| 2485 | dzde = rdp_cmd_data[zbuffer_base+2]; |
| 2486 | dzdy = rdp_cmd_data[zbuffer_base+3]; |
| 2487 | |
| 2488 | span[0].ds.w = dsdx; |
| 2489 | span[0].dt.w = dtdx; |
| 2490 | span[0].dw.w = dwdx; |
| 2491 | span[0].dr.w = drdx & ~0x1f; |
| 2492 | span[0].dg.w = dgdx & ~0x1f; |
| 2493 | span[0].db.w = dbdx & ~0x1f; |
| 2494 | span[0].da.w = dadx & ~0x1f; |
| 2495 | span[0].dz.w = dzdx; |
| 2496 | dzdy_dz = (dzdy >> 16) & 0xffff; |
| 2497 | dzdx_dz = (dzdx >> 16) & 0xffff; |
| 2498 | span[0].dzpix = ((dzdy_dz & 0x8000) ? ((~dzdy_dz) & 0x7fff) : dzdy_dz) + ((dzdx_dz & 0x8000) ? ((~dzdx_dz) & 0x7fff) : dzdx_dz); |
| 2499 | span[0].dzpix = normalize_dzpix(span[0].dzpix); |
| 2500 | dsdylod = dsdy >> 16; |
| 2501 | dtdylod = dtdy >> 16; |
| 2502 | if (dsdylod & 0x20000) |
| 2503 | { |
| 2504 | dsdylod = ~dsdylod & 0x1ffff; |
| 2505 | } |
| 2506 | if (dtdylod & 0x20000) |
| 2507 | { |
| 2508 | dtdylod = ~dtdylod & 0x1ffff; |
| 2509 | } |
| 2510 | span[0].dymax = (dsdylod > dtdylod)? dsdylod : dtdylod; |
| 2511 | |
| 2512 | xleft_inc = dxmdy >> 2; |
| 2513 | xright_inc = dxhdy >> 2; |
| 2514 | |
| 2515 | xright = xh; |
| 2516 | xleft = xm; |
| 2517 | |
| 2518 | limcvg = ((yl>>2) <= 1023) ? (yl>>2) : 1023; // Needed by 40 Winks |
| 2519 | if (limcvg < 0) |
| 2520 | { |
| 2521 | limcvg = 0; |
| 2522 | } |
| 2523 | |
| 2524 | startcvg = ((yh>>2)>=0) ? (yh>>2) : 0; |
| 2525 | for (k = startcvg; k <= limcvg; k++) |
| 2526 | { |
| 2527 | memset((void*)&span[k].cvg[0],0,640); |
| 2528 | } |
| 2529 | |
| 2530 | sign_dxldy = (dxldy & 0x80000000) ? 1 : 0; |
| 2531 | sign_dxmdy = (dxmdy & 0x80000000) ? 1 : 0; |
| 2532 | samesign = !(sign_dxldy ^ sign_dxmdy); |
| 2533 | |
| 2534 | sign_dxhdy = (dxhdy & 0x80000000) ? 1 : 0; |
| 2535 | |
| 2536 | do_offset = !(sign_dxhdy ^ (flip)); |
| 2537 | |
| 2538 | if (do_offset) |
| 2539 | { |
| 2540 | dsdeh = dsde >> 9; dsdyh = dsdy >> 9; |
| 2541 | dtdeh = dtde >> 9; dtdyh = dtdy >> 9; |
| 2542 | dwdeh = dwde >> 9; dwdyh = dwdy >> 9; |
| 2543 | drdeh = drde >> 9; drdyh = drdy >> 9; |
| 2544 | dgdeh = dgde >> 9; dgdyh = dgdy >> 9; |
| 2545 | dbdeh = dbde >> 9; dbdyh = dbdy >> 9; |
| 2546 | dadeh = dade >> 9; dadyh = dady >> 9; |
| 2547 | dzdeh = dzde >> 9; dzdyh = dzdy >> 9; |
| 2548 | |
| 2549 | dsdiff = (dsdeh*3 - dsdyh*3) << 7; |
| 2550 | dtdiff = (dtdeh*3 - dtdyh*3) << 7; |
| 2551 | dwdiff = (dwdeh*3 - dwdyh*3) << 7; |
| 2552 | drdiff = (drdeh*3 - drdyh*3) << 7; |
| 2553 | dgdiff = (dgdeh*3 - dgdyh*3) << 7; |
| 2554 | dbdiff = (dbdeh*3 - dbdyh*3) << 7; |
| 2555 | dadiff = (dadeh*3 - dadyh*3) << 7; |
| 2556 | dzdiff = (dzdeh*3 - dzdyh*3) << 7; |
| 2557 | } |
| 2558 | else |
| 2559 | { |
| 2560 | dsdiff = dtdiff = dwdiff = drdiff = dgdiff = dbdiff = dadiff = dzdiff = 0; |
| 2561 | } |
| 2562 | |
| 2563 | if (do_offset) |
| 2564 | { |
| 2565 | dseoff = (dsdeh*3) << 7; |
| 2566 | dteoff = (dtdeh*3) << 7; |
| 2567 | dweoff = (dwdeh*3) << 7; |
| 2568 | dreoff = (drdeh*3) << 7; |
| 2569 | dgeoff = (dgdeh*3) << 7; |
| 2570 | dbeoff = (dbdeh*3) << 7; |
| 2571 | daeoff = (dadeh*3) << 7; |
| 2572 | dzeoff = (dzdeh*3) << 7; |
| 2573 | } |
| 2574 | else |
| 2575 | { |
| 2576 | dseoff = dteoff = dweoff = dreoff = dgeoff = dbeoff = daeoff = dzeoff = 0; |
| 2577 | } |
| 2578 | |
| 2579 | #define addleft(x) addleftcvg(x,k) |
| 2580 | #define addright(x) addrightcvg(x,k) |
| 2581 | #define setvalues() { \ |
| 2582 | addvalues(); \ |
| 2583 | adjust_attr(); \ |
| 2584 | } |
| 2585 | |
| 2586 | dsdxh = dsdx >> 8; |
| 2587 | dtdxh = dtdx >> 8; |
| 2588 | dwdxh = dwdx >> 8; |
| 2589 | drdxh = drdx >> 8; |
| 2590 | dgdxh = dgdx >> 8; |
| 2591 | dbdxh = dbdx >> 8; |
| 2592 | dadxh = dadx >> 8; |
| 2593 | dzdxh = dzdx >> 8; |
| 2594 | |
| 2595 | #define adjust_attr() \ |
| 2596 | { \ |
| 2597 | span[j].s.w = (s + dsdiff - (xfrac * dsdxh)) & ~0x1f; \ |
| 2598 | span[j].t.w = (t + dtdiff - (xfrac * dtdxh)) & ~0x1f; \ |
| 2599 | span[j].w.w = (w + dwdiff - (xfrac * dwdxh)) & ~0x1f; \ |
| 2600 | span[j].r.w = r + drdiff - (xfrac * drdxh); \ |
| 2601 | span[j].g.w = g + dgdiff - (xfrac * dgdxh); \ |
| 2602 | span[j].b.w = b + dbdiff - (xfrac * dbdxh); \ |
| 2603 | span[j].a.w = a + dadiff - (xfrac * dadxh); \ |
| 2604 | span[j].z.w = z + dzdiff - (xfrac * dzdxh); \ |
| 2605 | } |
| 2606 | |
| 2607 | #define addvalues() { \ |
| 2608 | s += dsde; \ |
| 2609 | t += dtde; \ |
| 2610 | w += dwde; \ |
| 2611 | r += drde; \ |
| 2612 | g += dgde; \ |
| 2613 | b += dbde; \ |
| 2614 | a += dade; \ |
| 2615 | z += dzde; \ |
| 2616 | } |
| 2617 | |
| 2618 | m_inc = flip ? 1 : -1; |
| 2619 | |
| 2620 | ycur = yh & ~3; |
| 2621 | ylfar = yl | 3; |
| 2622 | ldflag = (sign_dxhdy ^ flip) ? 0 : 3; |
| 2623 | yhpix = yh >> 2; |
| 2624 | ympix = ym >> 2; |
| 2625 | ylpix = yl >> 2; |
| 2626 | |
| 2627 | for (k = ycur; k <= ylfar; k++) |
| 2628 | { |
| 2629 | if (k == ym) |
| 2630 | { |
| 2631 | xleft = xl; |
| 2632 | xleft_inc = dxldy >> 2; |
| 2633 | } |
| 2634 | |
| 2635 | xstart = xleft >> 16; |
| 2636 | xend = xright >> 16; |
| 2637 | j = k >> 2; |
| 2638 | spix = k & 3; |
| 2639 | |
| 2640 | if (k >= 0 && k < 0x1000) |
| 2641 | { |
| 2642 | int m = 0; |
| 2643 | int n = 0; |
| 2644 | int length = 0; |
| 2645 | min = 0; max = 3; |
| 2646 | if (j == yhpix) |
| 2647 | { |
| 2648 | min = yh & 3; |
| 2649 | } |
| 2650 | if (j == ylpix) |
| 2651 | { |
| 2652 | max = yl & 3; |
| 2653 | } |
| 2654 | if (spix >= min && spix <= max) |
| 2655 | { |
| 2656 | if (spix == min) |
| 2657 | { |
| 2658 | minxmx = maxxmx = xstart; |
| 2659 | minxhx = maxxhx = xend; |
| 2660 | } |
| 2661 | else |
| 2662 | { |
| 2663 | minxmx = (xstart < minxmx) ? xstart : minxmx; |
| 2664 | maxxmx = (xstart > maxxmx) ? xstart : maxxmx; |
| 2665 | minxhx = (xend < minxhx) ? xend : minxhx; |
| 2666 | maxxhx = (xend > maxxhx) ? xend : maxxhx; |
| 2667 | } |
| 2668 | } |
| 2669 | |
| 2670 | if (spix == max) |
| 2671 | { |
| 2672 | if (flip) |
| 2673 | { |
| 2674 | span[j].lx = maxxmx; |
| 2675 | span[j].rx = minxhx; |
| 2676 | } |
| 2677 | else |
| 2678 | { |
| 2679 | span[j].lx = minxmx; |
| 2680 | span[j].rx = maxxhx; |
| 2681 | } |
| 2682 | } |
| 2683 | |
| 2684 | length = flip ? (xstart - xend) : (xend - xstart); |
| 2685 | |
| 2686 | if (spix == ldflag) |
| 2687 | { |
| 2688 | xfrac = ((xright >> 8) & 0xff); |
| 2689 | adjust_attr(); |
| 2690 | } |
| 2691 | |
| 2692 | m = flip ? (xend+1) : (xend-1); |
| 2693 | |
| 2694 | if (k >= yh && length >= 0 && k <= yl) |
| 2695 | { |
| 2696 | if (xstart>=0 && xstart <1024) |
| 2697 | { |
| 2698 | if (!flip) |
| 2699 | { |
| 2700 | span[j].cvg[xstart] += addleft(xleft); |
| 2701 | } |
| 2702 | else |
| 2703 | { |
| 2704 | span[j].cvg[xstart] += addright(xleft); |
| 2705 | } |
| 2706 | } |
| 2707 | if (xend>=0 && xend<1024) |
| 2708 | { |
| 2709 | if (xstart != xend) |
| 2710 | { |
| 2711 | if (!flip) |
| 2712 | { |
| 2713 | span[j].cvg[xend] += addright(xright); |
| 2714 | } |
| 2715 | else |
| 2716 | { |
| 2717 | span[j].cvg[xend] += addleft(xright); |
| 2718 | } |
| 2719 | } |
| 2720 | else |
| 2721 | { |
| 2722 | if (!flip) |
| 2723 | { |
| 2724 | span[j].cvg[xend] -= (2 - addright(xright)); |
| 2725 | } |
| 2726 | else |
| 2727 | { |
| 2728 | span[j].cvg[xend] -= (2 - addleft(xright)); |
| 2729 | } |
| 2730 | if (span[j].cvg[xend] > 200) |
| 2731 | { |
| 2732 | span[j].cvg[xend] = 0; |
| 2733 | } |
| 2734 | } |
| 2735 | } |
| 2736 | for (n = 0; n < (length - 1); n++) |
| 2737 | { |
| 2738 | if (m>=0 && m < 640) |
| 2739 | { |
| 2740 | span[j].cvg[m] += 2; |
| 2741 | } |
| 2742 | |
| 2743 | m += m_inc; |
| 2744 | } |
| 2745 | } |
| 2746 | } |
| 2747 | |
| 2748 | if (spix == 3) |
| 2749 | { |
| 2750 | addvalues(); |
| 2751 | } |
| 2752 | xleft += xleft_inc; |
| 2753 | xright += xright_inc; |
| 2754 | } |
| 2755 | |
| 2756 | switch (fb_size) // 8bpp needs to be implemented |
| 2757 | { |
| 2758 | case PIXEL_SIZE_16BIT: |
| 2759 | if(flip) |
| 2760 | { |
| 2761 | render_spans_16_s_t_z_f(yh>>2, yl>>2, &tile[0]); break; |
| 2762 | } |
| 2763 | else |
| 2764 | { |
| 2765 | render_spans_16_s_t_z_nf(yh>>2, yl>>2, &tile[0]); break; |
| 2766 | } |
| 2767 | case PIXEL_SIZE_32BIT: render_spans_32(yh>>2, yl>>2, &tile[tilenum], 1, 1, 1, flip); break; |
| 2768 | default: break; // V-Rally2 does this, fb_size=0 |
| 2769 | } |
| 2770 | } |
| 2771 | #undef addvalues |
| 2772 | #undef adjust_attr |
| 2773 | #undef addleft |
| 2774 | #undef addright |
| 2775 | #undef setvalues |