![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to d_polysa.s CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [Quake] / quake1|
Return to d_polysa.s CVS log | Up to [Quake] / quake1 |
1.1 ! root 1: // ! 2: // d_polysa.s ! 3: // x86 assembly-language polygon model drawing code ! 4: // ! 5: ! 6: #include "asm_i386.h" ! 7: #include "quakeasm.h" ! 8: #include "asm_draw.h" ! 9: #include "d_ifacea.h" ! 10: ! 11: #if id386 ! 12: ! 13: // !!! if this is changed, it must be changed in d_polyse.c too !!! ! 14: #define DPS_MAXSPANS MAXHEIGHT+1 ! 15: // 1 extra for spanpackage that marks end ! 16: ! 17: #define SPAN_SIZE (((DPS_MAXSPANS + 1 + ((CACHE_SIZE - 1) / spanpackage_t_size)) + 1) * spanpackage_t_size) ! 18: ! 19: ! 20: .data ! 21: ! 22: .align 4 ! 23: p10_minus_p20: .single 0 ! 24: p01_minus_p21: .single 0 ! 25: temp0: .single 0 ! 26: temp1: .single 0 ! 27: Ltemp: .single 0 ! 28: ! 29: aff8entryvec_table: .long LDraw8, LDraw7, LDraw6, LDraw5 ! 30: .long LDraw4, LDraw3, LDraw2, LDraw1 ! 31: ! 32: lzistepx: .long 0 ! 33: ! 34: ! 35: .text ! 36: ! 37: #ifndef NeXT ! 38: .extern C(D_PolysetSetEdgeTable) ! 39: .extern C(D_RasterizeAliasPolySmooth) ! 40: #endif ! 41: ! 42: //---------------------------------------------------------------------- ! 43: // affine triangle gradient calculation code ! 44: //---------------------------------------------------------------------- ! 45: ! 46: #define skinwidth 4+0 ! 47: ! 48: .globl C(D_PolysetCalcGradients) ! 49: C(D_PolysetCalcGradients): ! 50: ! 51: // p00_minus_p20 = r_p0[0] - r_p2[0]; ! 52: // p01_minus_p21 = r_p0[1] - r_p2[1]; ! 53: // p10_minus_p20 = r_p1[0] - r_p2[0]; ! 54: // p11_minus_p21 = r_p1[1] - r_p2[1]; ! 55: // ! 56: // xstepdenominv = 1.0 / (p10_minus_p20 * p01_minus_p21 - ! 57: // p00_minus_p20 * p11_minus_p21); ! 58: // ! 59: // ystepdenominv = -xstepdenominv; ! 60: ! 61: fildl C(r_p0)+0 // r_p0[0] ! 62: fildl C(r_p2)+0 // r_p2[0] | r_p0[0] ! 63: fildl C(r_p0)+4 // r_p0[1] | r_p2[0] | r_p0[0] ! 64: fildl C(r_p2)+4 // r_p2[1] | r_p0[1] | r_p2[0] | r_p0[0] ! 65: fildl C(r_p1)+0 // r_p1[0] | r_p2[1] | r_p0[1] | r_p2[0] | r_p0[0] ! 66: fildl C(r_p1)+4 // r_p1[1] | r_p1[0] | r_p2[1] | r_p0[1] | ! 67: // r_p2[0] | r_p0[0] ! 68: fxch %st(3) // r_p0[1] | r_p1[0] | r_p2[1] | r_p1[1] | ! 69: // r_p2[0] | r_p0[0] ! 70: fsub %st(2),%st(0) // p01_minus_p21 | r_p1[0] | r_p2[1] | r_p1[1] | ! 71: // r_p2[0] | r_p0[0] ! 72: fxch %st(1) // r_p1[0] | p01_minus_p21 | r_p2[1] | r_p1[1] | ! 73: // r_p2[0] | r_p0[0] ! 74: fsub %st(4),%st(0) // p10_minus_p20 | p01_minus_p21 | r_p2[1] | ! 75: // r_p1[1] | r_p2[0] | r_p0[0] ! 76: fxch %st(5) // r_p0[0] | p01_minus_p21 | r_p2[1] | ! 77: // r_p1[1] | r_p2[0] | p10_minus_p20 ! 78: fsubp %st(0),%st(4) // p01_minus_p21 | r_p2[1] | r_p1[1] | ! 79: // p00_minus_p20 | p10_minus_p20 ! 80: fxch %st(2) // r_p1[1] | r_p2[1] | p01_minus_p21 | ! 81: // p00_minus_p20 | p10_minus_p20 ! 82: fsubp %st(0),%st(1) // p11_minus_p21 | p01_minus_p21 | ! 83: // p00_minus_p20 | p10_minus_p20 ! 84: fxch %st(1) // p01_minus_p21 | p11_minus_p21 | ! 85: // p00_minus_p20 | p10_minus_p20 ! 86: flds C(d_xdenom) // d_xdenom | p01_minus_p21 | p11_minus_p21 | ! 87: // p00_minus_p20 | p10_minus_p20 ! 88: fxch %st(4) // p10_minus_p20 | p01_minus_p21 | p11_minus_p21 | ! 89: // p00_minus_p20 | d_xdenom ! 90: fstps p10_minus_p20 // p01_minus_p21 | p11_minus_p21 | ! 91: // p00_minus_p20 | d_xdenom ! 92: fstps p01_minus_p21 // p11_minus_p21 | p00_minus_p20 | xstepdenominv ! 93: fxch %st(2) // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 94: ! 95: //// ceil () for light so positive steps are exaggerated, negative steps ! 96: //// diminished, pushing us away from underflow toward overflow. Underflow is ! 97: //// very visible, overflow is very unlikely, because of ambient lighting ! 98: // t0 = r_p0[4] - r_p2[4]; ! 99: // t1 = r_p1[4] - r_p2[4]; ! 100: ! 101: fildl C(r_p2)+16 // r_p2[4] | xstepdenominv | p00_minus_p20 | ! 102: // p11_minus_p21 ! 103: fildl C(r_p0)+16 // r_p0[4] | r_p2[4] | xstepdenominv | ! 104: // p00_minus_p20 | p11_minus_p21 ! 105: fildl C(r_p1)+16 // r_p1[4] | r_p0[4] | r_p2[4] | xstepdenominv | ! 106: // p00_minus_p20 | p11_minus_p21 ! 107: fxch %st(2) // r_p2[4] | r_p0[4] | r_p1[4] | xstepdenominv | ! 108: // p00_minus_p20 | p11_minus_p21 ! 109: fld %st(0) // r_p2[4] | r_p2[4] | r_p0[4] | r_p1[4] | ! 110: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 111: fsubrp %st(0),%st(2) // r_p2[4] | t0 | r_p1[4] | xstepdenominv | ! 112: // p00_minus_p20 | p11_minus_p21 ! 113: fsubrp %st(0),%st(2) // t0 | t1 | xstepdenominv | p00_minus_p20 | ! 114: // p11_minus_p21 ! 115: ! 116: // r_lstepx = (int) ! 117: // ceil((t1 * p01_minus_p21 - t0 * p11_minus_p21) * xstepdenominv); ! 118: // r_lstepy = (int) ! 119: // ceil((t1 * p00_minus_p20 - t0 * p10_minus_p20) * ystepdenominv); ! 120: ! 121: fld %st(0) // t0 | t0 | t1 | xstepdenominv | p00_minus_p20 | ! 122: // p11_minus_p21 ! 123: fmul %st(5),%st(0) // t0*p11_minus_p21 | t0 | t1 | xstepdenominv | ! 124: // p00_minus_p20 | p11_minus_p21 ! 125: fxch %st(2) // t1 | t0 | t0*p11_minus_p21 | xstepdenominv | ! 126: // p00_minus_p20 | p11_minus_p21 ! 127: fld %st(0) // t1 | t1 | t0 | t0*p11_minus_p21 | ! 128: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 129: fmuls p01_minus_p21 // t1*p01_minus_p21 | t1 | t0 | t0*p11_minus_p21 | ! 130: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 131: fxch %st(2) // t0 | t1 | t1*p01_minus_p21 | t0*p11_minus_p21 | ! 132: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 133: fmuls p10_minus_p20 // t0*p10_minus_p20 | t1 | t1*p01_minus_p21 | ! 134: // t0*p11_minus_p21 | xstepdenominv | ! 135: // p00_minus_p20 | p11_minus_p21 ! 136: fxch %st(1) // t1 | t0*p10_minus_p20 | t1*p01_minus_p21 | ! 137: // t0*p11_minus_p21 | xstepdenominv | ! 138: // p00_minus_p20 | p11_minus_p21 ! 139: fmul %st(5),%st(0) // t1*p00_minus_p20 | t0*p10_minus_p20 | ! 140: // t1*p01_minus_p21 | t0*p11_minus_p21 | ! 141: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 142: fxch %st(2) // t1*p01_minus_p21 | t0*p10_minus_p20 | ! 143: // t1*p00_minus_p20 | t0*p11_minus_p21 | ! 144: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 145: fsubp %st(0),%st(3) // t0*p10_minus_p20 | t1*p00_minus_p20 | ! 146: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 147: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 148: fsubrp %st(0),%st(1) // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 149: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 150: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 151: fld %st(2) // xstepdenominv | ! 152: // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 153: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 154: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 155: fmuls float_minus_1 // ystepdenominv | ! 156: // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 157: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 158: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 159: fxch %st(2) // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 160: // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 161: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 162: // p11_minus_p21 ! 163: fmul %st(3),%st(0) // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 164: // xstepdenominv | ! 165: // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 166: // | ystepdenominv | xstepdenominv | ! 167: // p00_minus_p20 | p11_minus_p21 ! 168: fxch %st(1) // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 169: // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 170: // xstepdenominv | ystepdenominv | ! 171: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 172: fmul %st(2),%st(0) // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 173: // ystepdenominv | ! 174: // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 175: // xstepdenominv | ystepdenominv | ! 176: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 177: fldcw ceil_cw ! 178: fistpl C(r_lstepy) // r_lstepx | ystepdenominv | xstepdenominv | ! 179: // p00_minus_p20 | p11_minus_p21 ! 180: fistpl C(r_lstepx) // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 181: // p11_minus_p21 ! 182: fldcw single_cw ! 183: ! 184: // t0 = r_p0[2] - r_p2[2]; ! 185: // t1 = r_p1[2] - r_p2[2]; ! 186: ! 187: fildl C(r_p2)+8 // r_p2[2] | ystepdenominv | xstepdenominv | ! 188: // p00_minus_p20 | p11_minus_p21 ! 189: fildl C(r_p0)+8 // r_p0[2] | r_p2[2] | ystepdenominv | ! 190: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 191: fildl C(r_p1)+8 // r_p1[2] | r_p0[2] | r_p2[2] | ystepdenominv | ! 192: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 193: fxch %st(2) // r_p2[2] | r_p0[2] | r_p1[2] | ystepdenominv | ! 194: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 195: fld %st(0) // r_p2[2] | r_p2[2] | r_p0[2] | r_p1[2] | ! 196: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 197: // p11_minus_p21 ! 198: fsubrp %st(0),%st(2) // r_p2[2] | t0 | r_p1[2] | ystepdenominv | ! 199: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 200: fsubrp %st(0),%st(2) // t0 | t1 | ystepdenominv | xstepdenominv | ! 201: // p00_minus_p20 | p11_minus_p21 ! 202: ! 203: // r_sstepx = (int)((t1 * p01_minus_p21 - t0 * p11_minus_p21) * ! 204: // xstepdenominv); ! 205: // r_sstepy = (int)((t1 * p00_minus_p20 - t0 * p10_minus_p20) * ! 206: // ystepdenominv); ! 207: ! 208: fld %st(0) // t0 | t0 | t1 | ystepdenominv | xstepdenominv ! 209: fmul %st(6),%st(0) // t0*p11_minus_p21 | t0 | t1 | ystepdenominv | ! 210: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 211: fxch %st(2) // t1 | t0 | t0*p11_minus_p21 | ystepdenominv | ! 212: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 213: fld %st(0) // t1 | t1 | t0 | t0*p11_minus_p21 | ! 214: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 215: // p11_minus_p21 ! 216: fmuls p01_minus_p21 // t1*p01_minus_p21 | t1 | t0 | t0*p11_minus_p21 | ! 217: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 218: // p11_minus_p21 ! 219: fxch %st(2) // t0 | t1 | t1*p01_minus_p21 | t0*p11_minus_p21 | ! 220: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 221: // p11_minus_p21 ! 222: fmuls p10_minus_p20 // t0*p10_minus_p20 | t1 | t1*p01_minus_p21 | ! 223: // t0*p11_minus_p21 | ystepdenominv | ! 224: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 225: fxch %st(1) // t1 | t0*p10_minus_p20 | t1*p01_minus_p21 | ! 226: // t0*p11_minus_p21 | ystepdenominv | ! 227: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 228: fmul %st(6),%st(0) // t1*p00_minus_p20 | t0*p10_minus_p20 | ! 229: // t1*p01_minus_p21 | t0*p11_minus_p21 | ! 230: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 231: // p11_minus_p21 ! 232: fxch %st(2) // t1*p01_minus_p21 | t0*p10_minus_p20 | ! 233: // t1*p00_minus_p20 | t0*p11_minus_p21 | ! 234: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 235: // p11_minus_p21 ! 236: fsubp %st(0),%st(3) // t0*p10_minus_p20 | t1*p00_minus_p20 | ! 237: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 238: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 239: // p11_minus_p21 ! 240: fsubrp %st(0),%st(1) // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 241: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 242: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 243: // p11_minus_p21 ! 244: fmul %st(2),%st(0) // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 245: // ystepdenominv | ! 246: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 247: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 248: // p11_minus_p21 ! 249: fxch %st(1) // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 250: // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 251: // ystepdenominv | ystepdenominv | ! 252: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 253: fmul %st(3),%st(0) // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 254: // xstepdenominv | ! 255: // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 256: // ystepdenominv | ystepdenominv | ! 257: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 258: fxch %st(1) // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 259: // ystepdenominv | ! 260: // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 261: // xstepdenominv | ystepdenominv | ! 262: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 263: fistpl C(r_sstepy) // r_sstepx | ystepdenominv | xstepdenominv | ! 264: // p00_minus_p20 | p11_minus_p21 ! 265: fistpl C(r_sstepx) // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 266: // p11_minus_p21 ! 267: ! 268: // t0 = r_p0[3] - r_p2[3]; ! 269: // t1 = r_p1[3] - r_p2[3]; ! 270: ! 271: fildl C(r_p2)+12 // r_p2[3] | ystepdenominv | xstepdenominv | ! 272: // p00_minus_p20 | p11_minus_p21 ! 273: fildl C(r_p0)+12 // r_p0[3] | r_p2[3] | ystepdenominv | ! 274: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 275: fildl C(r_p1)+12 // r_p1[3] | r_p0[3] | r_p2[3] | ystepdenominv | ! 276: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 277: fxch %st(2) // r_p2[3] | r_p0[3] | r_p1[3] | ystepdenominv | ! 278: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 279: fld %st(0) // r_p2[3] | r_p2[3] | r_p0[3] | r_p1[3] | ! 280: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 281: // p11_minus_p21 ! 282: fsubrp %st(0),%st(2) // r_p2[3] | t0 | r_p1[3] | ystepdenominv | ! 283: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 284: fsubrp %st(0),%st(2) // t0 | t1 | ystepdenominv | xstepdenominv | ! 285: // p00_minus_p20 | p11_minus_p21 ! 286: ! 287: // r_tstepx = (int)((t1 * p01_minus_p21 - t0 * p11_minus_p21) * ! 288: // xstepdenominv); ! 289: // r_tstepy = (int)((t1 * p00_minus_p20 - t0 * p10_minus_p20) * ! 290: // ystepdenominv); ! 291: ! 292: fld %st(0) // t0 | t0 | t1 | ystepdenominv | xstepdenominv | ! 293: // p00_minus_p20 | p11_minus_p21 ! 294: fmul %st(6),%st(0) // t0*p11_minus_p21 | t0 | t1 | ystepdenominv | ! 295: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 296: fxch %st(2) // t1 | t0 | t0*p11_minus_p21 | ystepdenominv | ! 297: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 298: fld %st(0) // t1 | t1 | t0 | t0*p11_minus_p21 | ! 299: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 300: // p11_minus_p21 ! 301: fmuls p01_minus_p21 // t1*p01_minus_p21 | t1 | t0 | t0*p11_minus_p21 | ! 302: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 303: // p11_minus_p21 ! 304: fxch %st(2) // t0 | t1 | t1*p01_minus_p21 | t0*p11_minus_p21 | ! 305: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 306: // p11_minus_p21 ! 307: fmuls p10_minus_p20 // t0*p10_minus_p20 | t1 | t1*p01_minus_p21 | ! 308: // t0*p11_minus_p21 | ystepdenominv | ! 309: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 310: fxch %st(1) // t1 | t0*p10_minus_p20 | t1*p01_minus_p21 | ! 311: // t0*p11_minus_p21 | ystepdenominv | ! 312: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 313: fmul %st(6),%st(0) // t1*p00_minus_p20 | t0*p10_minus_p20 | ! 314: // t1*p01_minus_p21 | t0*p11_minus_p21 | ! 315: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 316: // p11_minus_p21 ! 317: fxch %st(2) // t1*p01_minus_p21 | t0*p10_minus_p20 | ! 318: // t1*p00_minus_p20 | t0*p11_minus_p21 | ! 319: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 320: // p11_minus_p21 ! 321: fsubp %st(0),%st(3) // t0*p10_minus_p20 | t1*p00_minus_p20 | ! 322: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 323: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 324: // p11_minus_p21 ! 325: fsubrp %st(0),%st(1) // t1*p00_minus_p20 - t0*p10_minus_p20 | ! 326: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 327: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 328: // p11_minus_p21 ! 329: fmul %st(2),%st(0) // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 330: // ystepdenominv | ! 331: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 332: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 333: // p11_minus_p21 ! 334: fxch %st(1) // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 335: // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 336: // ystepdenominv | ystepdenominv | ! 337: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 338: fmul %st(3),%st(0) // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 339: // xstepdenominv | ! 340: // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 341: // ystepdenominv | ystepdenominv | ! 342: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 343: fxch %st(1) // (t1*p00_minus_p20 - t0*p10_minus_p20)* ! 344: // ystepdenominv | ! 345: // (t1*p01_minus_p21 - t0*p11_minus_p21)* ! 346: // xstepdenominv | ystepdenominv | ! 347: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 348: fistpl C(r_tstepy) // r_tstepx | ystepdenominv | xstepdenominv | ! 349: // p00_minus_p20 | p11_minus_p21 ! 350: fistpl C(r_tstepx) // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 351: // p11_minus_p21 ! 352: ! 353: // t0 = r_p0[5] - r_p2[5]; ! 354: // t1 = r_p1[5] - r_p2[5]; ! 355: ! 356: fildl C(r_p2)+20 // r_p2[5] | ystepdenominv | xstepdenominv | ! 357: // p00_minus_p20 | p11_minus_p21 ! 358: fildl C(r_p0)+20 // r_p0[5] | r_p2[5] | ystepdenominv | ! 359: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 360: fildl C(r_p1)+20 // r_p1[5] | r_p0[5] | r_p2[5] | ystepdenominv | ! 361: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 362: fxch %st(2) // r_p2[5] | r_p0[5] | r_p1[5] | ystepdenominv | ! 363: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 364: fld %st(0) // r_p2[5] | r_p2[5] | r_p0[5] | r_p1[5] | ! 365: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 366: // p11_minus_p21 ! 367: fsubrp %st(0),%st(2) // r_p2[5] | t0 | r_p1[5] | ystepdenominv | ! 368: // xstepdenominv | p00_minus_p20 | p11_minus_p21 ! 369: fsubrp %st(0),%st(2) // t0 | t1 | ystepdenominv | xstepdenominv | ! 370: // p00_minus_p20 | p11_minus_p21 ! 371: ! 372: // r_zistepx = (int)((t1 * p01_minus_p21 - t0 * p11_minus_p21) * ! 373: // xstepdenominv); ! 374: // r_zistepy = (int)((t1 * p00_minus_p20 - t0 * p10_minus_p20) * ! 375: // ystepdenominv); ! 376: ! 377: fld %st(0) // t0 | t0 | t1 | ystepdenominv | xstepdenominv | ! 378: // p00_minus_p20 | p11_minus_p21 ! 379: fmulp %st(0),%st(6) // t0 | t1 | ystepdenominv | xstepdenominv | ! 380: // p00_minus_p20 | t0*p11_minus_p21 ! 381: fxch %st(1) // t1 | t0 | ystepdenominv | xstepdenominv | ! 382: // p00_minus_p20 | t0*p11_minus_p21 ! 383: fld %st(0) // t1 | t1 | t0 | ystepdenominv | xstepdenominv | ! 384: // p00_minus_p20 | t0*p11_minus_p21 ! 385: fmuls p01_minus_p21 // t1*p01_minus_p21 | t1 | t0 | ystepdenominv | ! 386: // xstepdenominv | p00_minus_p20 | ! 387: // t0*p11_minus_p21 ! 388: fxch %st(2) // t0 | t1 | t1*p01_minus_p21 | ystepdenominv | ! 389: // xstepdenominv | p00_minus_p20 | ! 390: // t0*p11_minus_p21 ! 391: fmuls p10_minus_p20 // t0*p10_minus_p20 | t1 | t1*p01_minus_p21 | ! 392: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 393: // t0*p11_minus_p21 ! 394: fxch %st(1) // t1 | t0*p10_minus_p20 | t1*p01_minus_p21 | ! 395: // ystepdenominv | xstepdenominv | p00_minus_p20 | ! 396: // t0*p11_minus_p21 ! 397: fmulp %st(0),%st(5) // t0*p10_minus_p20 | t1*p01_minus_p21 | ! 398: // ystepdenominv | xstepdenominv | ! 399: // t1*p00_minus_p20 | t0*p11_minus_p21 ! 400: fxch %st(5) // t0*p11_minus_p21 | t1*p01_minus_p21 | ! 401: // ystepdenominv | xstepdenominv | ! 402: // t1*p00_minus_p20 | t0*p10_minus_p20 ! 403: fsubrp %st(0),%st(1) // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 404: // ystepdenominv | xstepdenominv | ! 405: // t1*p00_minus_p20 | t0*p10_minus_p20 ! 406: fxch %st(3) // t1*p00_minus_p20 | ystepdenominv | ! 407: // xstepdenominv | ! 408: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 409: // t0*p10_minus_p20 ! 410: fsubp %st(0),%st(4) // ystepdenominv | xstepdenominv | ! 411: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 412: // t1*p00_minus_p20 - t0*p10_minus_p20 ! 413: fxch %st(1) // xstepdenominv | ystepdenominv | ! 414: // t1*p01_minus_p21 - t0*p11_minus_p21 | ! 415: // t1*p00_minus_p20 - t0*p10_minus_p20 ! 416: fmulp %st(0),%st(2) // ystepdenominv | ! 417: // (t1*p01_minus_p21 - t0*p11_minus_p21) * ! 418: // xstepdenominv | ! 419: // t1*p00_minus_p20 - t0*p10_minus_p20 ! 420: fmulp %st(0),%st(2) // (t1*p01_minus_p21 - t0*p11_minus_p21) * ! 421: // xstepdenominv | ! 422: // (t1*p00_minus_p20 - t0*p10_minus_p20) * ! 423: // ystepdenominv ! 424: fistpl C(r_zistepx) // (t1*p00_minus_p20 - t0*p10_minus_p20) * ! 425: // ystepdenominv ! 426: fistpl C(r_zistepy) ! 427: ! 428: // a_sstepxfrac = r_sstepx << 16; ! 429: // a_tstepxfrac = r_tstepx << 16; ! 430: // ! 431: // a_ststepxwhole = r_affinetridesc.skinwidth * (r_tstepx >> 16) + ! 432: // (r_sstepx >> 16); ! 433: ! 434: movl C(r_sstepx),%eax ! 435: movl C(r_tstepx),%edx ! 436: shll $16,%eax ! 437: shll $16,%edx ! 438: movl %eax,C(a_sstepxfrac) ! 439: movl %edx,C(a_tstepxfrac) ! 440: ! 441: movl C(r_sstepx),%ecx ! 442: movl C(r_tstepx),%eax ! 443: sarl $16,%ecx ! 444: sarl $16,%eax ! 445: imull skinwidth(%esp) ! 446: addl %ecx,%eax ! 447: movl %eax,C(a_ststepxwhole) ! 448: ! 449: ret ! 450: ! 451: ! 452: //---------------------------------------------------------------------- ! 453: // recursive subdivision affine triangle drawing code ! 454: // ! 455: // not C-callable because of stdcall return ! 456: //---------------------------------------------------------------------- ! 457: ! 458: #define lp1 4+16 ! 459: #define lp2 8+16 ! 460: #define lp3 12+16 ! 461: ! 462: .globl C(D_PolysetRecursiveTriangle) ! 463: C(D_PolysetRecursiveTriangle): ! 464: pushl %ebp // preserve caller stack frame pointer ! 465: pushl %esi // preserve register variables ! 466: pushl %edi ! 467: pushl %ebx ! 468: ! 469: // int *temp; ! 470: // int d; ! 471: // int new[6]; ! 472: // int i; ! 473: // int z; ! 474: // short *zbuf; ! 475: movl lp2(%esp),%esi ! 476: movl lp1(%esp),%ebx ! 477: movl lp3(%esp),%edi ! 478: ! 479: // d = lp2[0] - lp1[0]; ! 480: // if (d < -1 || d > 1) ! 481: // goto split; ! 482: movl 0(%esi),%eax ! 483: ! 484: movl 0(%ebx),%edx ! 485: movl 4(%esi),%ebp ! 486: ! 487: subl %edx,%eax ! 488: movl 4(%ebx),%ecx ! 489: ! 490: subl %ecx,%ebp ! 491: incl %eax ! 492: ! 493: cmpl $2,%eax ! 494: ja LSplit ! 495: ! 496: // d = lp2[1] - lp1[1]; ! 497: // if (d < -1 || d > 1) ! 498: // goto split; ! 499: movl 0(%edi),%eax ! 500: incl %ebp ! 501: ! 502: cmpl $2,%ebp ! 503: ja LSplit ! 504: ! 505: // d = lp3[0] - lp2[0]; ! 506: // if (d < -1 || d > 1) ! 507: // goto split2; ! 508: movl 0(%esi),%edx ! 509: movl 4(%edi),%ebp ! 510: ! 511: subl %edx,%eax ! 512: movl 4(%esi),%ecx ! 513: ! 514: subl %ecx,%ebp ! 515: incl %eax ! 516: ! 517: cmpl $2,%eax ! 518: ja LSplit2 ! 519: ! 520: // d = lp3[1] - lp2[1]; ! 521: // if (d < -1 || d > 1) ! 522: // goto split2; ! 523: movl 0(%ebx),%eax ! 524: incl %ebp ! 525: ! 526: cmpl $2,%ebp ! 527: ja LSplit2 ! 528: ! 529: // d = lp1[0] - lp3[0]; ! 530: // if (d < -1 || d > 1) ! 531: // goto split3; ! 532: movl 0(%edi),%edx ! 533: movl 4(%ebx),%ebp ! 534: ! 535: subl %edx,%eax ! 536: movl 4(%edi),%ecx ! 537: ! 538: subl %ecx,%ebp ! 539: incl %eax ! 540: ! 541: incl %ebp ! 542: movl %ebx,%edx ! 543: ! 544: cmpl $2,%eax ! 545: ja LSplit3 ! 546: ! 547: // d = lp1[1] - lp3[1]; ! 548: // if (d < -1 || d > 1) ! 549: // { ! 550: //split3: ! 551: // temp = lp1; ! 552: // lp3 = lp2; ! 553: // lp1 = lp3; ! 554: // lp2 = temp; ! 555: // goto split; ! 556: // } ! 557: // ! 558: // return; // entire tri is filled ! 559: // ! 560: cmpl $2,%ebp ! 561: jna LDone ! 562: ! 563: LSplit3: ! 564: movl %edi,%ebx ! 565: movl %esi,%edi ! 566: movl %edx,%esi ! 567: jmp LSplit ! 568: ! 569: //split2: ! 570: LSplit2: ! 571: ! 572: // temp = lp1; ! 573: // lp1 = lp2; ! 574: // lp2 = lp3; ! 575: // lp3 = temp; ! 576: movl %ebx,%eax ! 577: movl %esi,%ebx ! 578: movl %edi,%esi ! 579: movl %eax,%edi ! 580: ! 581: //split: ! 582: LSplit: ! 583: ! 584: subl $24,%esp // allocate space for a new vertex ! 585: ! 586: //// split this edge ! 587: // new[0] = (lp1[0] + lp2[0]) >> 1; ! 588: // new[1] = (lp1[1] + lp2[1]) >> 1; ! 589: // new[2] = (lp1[2] + lp2[2]) >> 1; ! 590: // new[3] = (lp1[3] + lp2[3]) >> 1; ! 591: // new[5] = (lp1[5] + lp2[5]) >> 1; ! 592: movl 8(%ebx),%eax ! 593: ! 594: movl 8(%esi),%edx ! 595: movl 12(%ebx),%ecx ! 596: ! 597: addl %edx,%eax ! 598: movl 12(%esi),%edx ! 599: ! 600: sarl $1,%eax ! 601: addl %edx,%ecx ! 602: ! 603: movl %eax,8(%esp) ! 604: movl 20(%ebx),%eax ! 605: ! 606: sarl $1,%ecx ! 607: movl 20(%esi),%edx ! 608: ! 609: movl %ecx,12(%esp) ! 610: addl %edx,%eax ! 611: ! 612: movl 0(%ebx),%ecx ! 613: movl 0(%esi),%edx ! 614: ! 615: sarl $1,%eax ! 616: addl %ecx,%edx ! 617: ! 618: movl %eax,20(%esp) ! 619: movl 4(%ebx),%eax ! 620: ! 621: sarl $1,%edx ! 622: movl 4(%esi),%ebp ! 623: ! 624: movl %edx,0(%esp) ! 625: addl %eax,%ebp ! 626: ! 627: sarl $1,%ebp ! 628: movl %ebp,4(%esp) ! 629: ! 630: //// draw the point if splitting a leading edge ! 631: // if (lp2[1] > lp1[1]) ! 632: // goto nodraw; ! 633: cmpl %eax,4(%esi) ! 634: jg LNoDraw ! 635: ! 636: // if ((lp2[1] == lp1[1]) && (lp2[0] < lp1[0])) ! 637: // goto nodraw; ! 638: movl 0(%esi),%edx ! 639: jnz LDraw ! 640: ! 641: cmpl %ecx,%edx ! 642: jl LNoDraw ! 643: ! 644: LDraw: ! 645: ! 646: // z = new[5] >> 16; ! 647: movl 20(%esp),%edx ! 648: movl 4(%esp),%ecx ! 649: ! 650: sarl $16,%edx ! 651: movl 0(%esp),%ebp ! 652: ! 653: // zbuf = zspantable[new[1]] + new[0]; ! 654: movl C(zspantable)(,%ecx,4),%eax ! 655: ! 656: // if (z >= *zbuf) ! 657: // { ! 658: cmpw (%eax,%ebp,2),%dx ! 659: jnge LNoDraw ! 660: ! 661: // int pix; ! 662: // ! 663: // *zbuf = z; ! 664: movw %dx,(%eax,%ebp,2) ! 665: ! 666: // pix = d_pcolormap[skintable[new[3]>>16][new[2]>>16]]; ! 667: movl 12(%esp),%eax ! 668: ! 669: sarl $16,%eax ! 670: movl 8(%esp),%edx ! 671: ! 672: sarl $16,%edx ! 673: subl %ecx,%ecx ! 674: ! 675: movl C(skintable)(,%eax,4),%eax ! 676: movl 4(%esp),%ebp ! 677: ! 678: movb (%eax,%edx,),%cl ! 679: movl C(d_pcolormap),%edx ! 680: ! 681: movb (%edx,%ecx,),%dl ! 682: movl 0(%esp),%ecx ! 683: ! 684: // d_viewbuffer[d_scantable[new[1]] + new[0]] = pix; ! 685: movl C(d_scantable)(,%ebp,4),%eax ! 686: addl %eax,%ecx ! 687: movl C(d_viewbuffer),%eax ! 688: movb %dl,(%eax,%ecx,1) ! 689: ! 690: // } ! 691: // ! 692: //nodraw: ! 693: LNoDraw: ! 694: ! 695: //// recursively continue ! 696: // D_PolysetRecursiveTriangle (lp3, lp1, new); ! 697: pushl %esp ! 698: pushl %ebx ! 699: pushl %edi ! 700: call C(D_PolysetRecursiveTriangle) ! 701: ! 702: // D_PolysetRecursiveTriangle (lp3, new, lp2); ! 703: movl %esp,%ebx ! 704: pushl %esi ! 705: pushl %ebx ! 706: pushl %edi ! 707: call C(D_PolysetRecursiveTriangle) ! 708: addl $24,%esp ! 709: ! 710: LDone: ! 711: popl %ebx // restore register variables ! 712: popl %edi ! 713: popl %esi ! 714: popl %ebp // restore caller stack frame pointer ! 715: ret $12 ! 716: ! 717: ! 718: //---------------------------------------------------------------------- ! 719: // 8-bpp horizontal span drawing code for affine polygons, with smooth ! 720: // shading and no transparency ! 721: //---------------------------------------------------------------------- ! 722: ! 723: #define pspans 4+8 ! 724: ! 725: .globl C(D_PolysetAff8Start) ! 726: C(D_PolysetAff8Start): ! 727: ! 728: .globl C(D_PolysetDrawSpans8) ! 729: C(D_PolysetDrawSpans8): ! 730: pushl %esi // preserve register variables ! 731: pushl %ebx ! 732: ! 733: movl pspans(%esp),%esi // point to the first span descriptor ! 734: movl C(r_zistepx),%ecx ! 735: ! 736: pushl %ebp // preserve caller's stack frame ! 737: pushl %edi ! 738: ! 739: rorl $16,%ecx // put high 16 bits of 1/z step in low word ! 740: movl spanpackage_t_count(%esi),%edx ! 741: ! 742: movl %ecx,lzistepx ! 743: ! 744: LSpanLoop: ! 745: ! 746: // lcount = d_aspancount - pspanpackage->count; ! 747: // ! 748: // errorterm += erroradjustup; ! 749: // if (errorterm >= 0) ! 750: // { ! 751: // d_aspancount += d_countextrastep; ! 752: // errorterm -= erroradjustdown; ! 753: // } ! 754: // else ! 755: // { ! 756: // d_aspancount += ubasestep; ! 757: // } ! 758: movl C(d_aspancount),%eax ! 759: subl %edx,%eax ! 760: ! 761: movl C(erroradjustup),%edx ! 762: movl C(errorterm),%ebx ! 763: addl %edx,%ebx ! 764: js LNoTurnover ! 765: ! 766: movl C(erroradjustdown),%edx ! 767: movl C(d_countextrastep),%edi ! 768: subl %edx,%ebx ! 769: movl C(d_aspancount),%ebp ! 770: movl %ebx,C(errorterm) ! 771: addl %edi,%ebp ! 772: movl %ebp,C(d_aspancount) ! 773: jmp LRightEdgeStepped ! 774: ! 775: LNoTurnover: ! 776: movl C(d_aspancount),%edi ! 777: movl C(ubasestep),%edx ! 778: movl %ebx,C(errorterm) ! 779: addl %edx,%edi ! 780: movl %edi,C(d_aspancount) ! 781: ! 782: LRightEdgeStepped: ! 783: cmpl $1,%eax ! 784: ! 785: jl LNextSpan ! 786: jz LExactlyOneLong ! 787: ! 788: // ! 789: // set up advancetable ! 790: // ! 791: movl C(a_ststepxwhole),%ecx ! 792: movl C(r_affinetridesc)+atd_skinwidth,%edx ! 793: ! 794: movl %ecx,advancetable+4 // advance base in t ! 795: addl %edx,%ecx ! 796: ! 797: movl %ecx,advancetable // advance extra in t ! 798: movl C(a_tstepxfrac),%ecx ! 799: ! 800: movw C(r_lstepx),%cx ! 801: movl %eax,%edx // count ! 802: ! 803: movl %ecx,tstep ! 804: addl $7,%edx ! 805: ! 806: shrl $3,%edx // count of full and partial loops ! 807: movl spanpackage_t_sfrac(%esi),%ebx ! 808: ! 809: movw %dx,%bx ! 810: movl spanpackage_t_pz(%esi),%ecx ! 811: ! 812: negl %eax ! 813: ! 814: movl spanpackage_t_pdest(%esi),%edi ! 815: andl $7,%eax // 0->0, 1->7, 2->6, ... , 7->1 ! 816: ! 817: subl %eax,%edi // compensate for hardwired offsets ! 818: subl %eax,%ecx ! 819: ! 820: subl %eax,%ecx ! 821: movl spanpackage_t_tfrac(%esi),%edx ! 822: ! 823: movw spanpackage_t_light(%esi),%dx ! 824: movl spanpackage_t_zi(%esi),%ebp ! 825: ! 826: rorl $16,%ebp // put high 16 bits of 1/z in low word ! 827: pushl %esi ! 828: ! 829: movl spanpackage_t_ptex(%esi),%esi ! 830: jmp aff8entryvec_table(,%eax,4) ! 831: ! 832: // %bx = count of full and partial loops ! 833: // %ebx high word = sfrac ! 834: // %ecx = pz ! 835: // %dx = light ! 836: // %edx high word = tfrac ! 837: // %esi = ptex ! 838: // %edi = pdest ! 839: // %ebp = 1/z ! 840: // tstep low word = C(r_lstepx) ! 841: // tstep high word = C(a_tstepxfrac) ! 842: // C(a_sstepxfrac) low word = 0 ! 843: // C(a_sstepxfrac) high word = C(a_sstepxfrac) ! 844: ! 845: LDrawLoop: ! 846: ! 847: // FIXME: do we need to clamp light? We may need at least a buffer bit to ! 848: // keep it from poking into tfrac and causing problems ! 849: ! 850: LDraw8: ! 851: cmpw (%ecx),%bp ! 852: jl Lp1 ! 853: xorl %eax,%eax ! 854: movb %dh,%ah ! 855: movb (%esi),%al ! 856: movw %bp,(%ecx) ! 857: movb 0x12345678(%eax),%al ! 858: LPatch8: ! 859: movb %al,(%edi) ! 860: Lp1: ! 861: addl tstep,%edx ! 862: sbbl %eax,%eax ! 863: addl lzistepx,%ebp ! 864: adcl $0,%ebp ! 865: addl C(a_sstepxfrac),%ebx ! 866: adcl advancetable+4(,%eax,4),%esi ! 867: ! 868: LDraw7: ! 869: cmpw 2(%ecx),%bp ! 870: jl Lp2 ! 871: xorl %eax,%eax ! 872: movb %dh,%ah ! 873: movb (%esi),%al ! 874: movw %bp,2(%ecx) ! 875: movb 0x12345678(%eax),%al ! 876: LPatch7: ! 877: movb %al,1(%edi) ! 878: Lp2: ! 879: addl tstep,%edx ! 880: sbbl %eax,%eax ! 881: addl lzistepx,%ebp ! 882: adcl $0,%ebp ! 883: addl C(a_sstepxfrac),%ebx ! 884: adcl advancetable+4(,%eax,4),%esi ! 885: ! 886: LDraw6: ! 887: cmpw 4(%ecx),%bp ! 888: jl Lp3 ! 889: xorl %eax,%eax ! 890: movb %dh,%ah ! 891: movb (%esi),%al ! 892: movw %bp,4(%ecx) ! 893: movb 0x12345678(%eax),%al ! 894: LPatch6: ! 895: movb %al,2(%edi) ! 896: Lp3: ! 897: addl tstep,%edx ! 898: sbbl %eax,%eax ! 899: addl lzistepx,%ebp ! 900: adcl $0,%ebp ! 901: addl C(a_sstepxfrac),%ebx ! 902: adcl advancetable+4(,%eax,4),%esi ! 903: ! 904: LDraw5: ! 905: cmpw 6(%ecx),%bp ! 906: jl Lp4 ! 907: xorl %eax,%eax ! 908: movb %dh,%ah ! 909: movb (%esi),%al ! 910: movw %bp,6(%ecx) ! 911: movb 0x12345678(%eax),%al ! 912: LPatch5: ! 913: movb %al,3(%edi) ! 914: Lp4: ! 915: addl tstep,%edx ! 916: sbbl %eax,%eax ! 917: addl lzistepx,%ebp ! 918: adcl $0,%ebp ! 919: addl C(a_sstepxfrac),%ebx ! 920: adcl advancetable+4(,%eax,4),%esi ! 921: ! 922: LDraw4: ! 923: cmpw 8(%ecx),%bp ! 924: jl Lp5 ! 925: xorl %eax,%eax ! 926: movb %dh,%ah ! 927: movb (%esi),%al ! 928: movw %bp,8(%ecx) ! 929: movb 0x12345678(%eax),%al ! 930: LPatch4: ! 931: movb %al,4(%edi) ! 932: Lp5: ! 933: addl tstep,%edx ! 934: sbbl %eax,%eax ! 935: addl lzistepx,%ebp ! 936: adcl $0,%ebp ! 937: addl C(a_sstepxfrac),%ebx ! 938: adcl advancetable+4(,%eax,4),%esi ! 939: ! 940: LDraw3: ! 941: cmpw 10(%ecx),%bp ! 942: jl Lp6 ! 943: xorl %eax,%eax ! 944: movb %dh,%ah ! 945: movb (%esi),%al ! 946: movw %bp,10(%ecx) ! 947: movb 0x12345678(%eax),%al ! 948: LPatch3: ! 949: movb %al,5(%edi) ! 950: Lp6: ! 951: addl tstep,%edx ! 952: sbbl %eax,%eax ! 953: addl lzistepx,%ebp ! 954: adcl $0,%ebp ! 955: addl C(a_sstepxfrac),%ebx ! 956: adcl advancetable+4(,%eax,4),%esi ! 957: ! 958: LDraw2: ! 959: cmpw 12(%ecx),%bp ! 960: jl Lp7 ! 961: xorl %eax,%eax ! 962: movb %dh,%ah ! 963: movb (%esi),%al ! 964: movw %bp,12(%ecx) ! 965: movb 0x12345678(%eax),%al ! 966: LPatch2: ! 967: movb %al,6(%edi) ! 968: Lp7: ! 969: addl tstep,%edx ! 970: sbbl %eax,%eax ! 971: addl lzistepx,%ebp ! 972: adcl $0,%ebp ! 973: addl C(a_sstepxfrac),%ebx ! 974: adcl advancetable+4(,%eax,4),%esi ! 975: ! 976: LDraw1: ! 977: cmpw 14(%ecx),%bp ! 978: jl Lp8 ! 979: xorl %eax,%eax ! 980: movb %dh,%ah ! 981: movb (%esi),%al ! 982: movw %bp,14(%ecx) ! 983: movb 0x12345678(%eax),%al ! 984: LPatch1: ! 985: movb %al,7(%edi) ! 986: Lp8: ! 987: addl tstep,%edx ! 988: sbbl %eax,%eax ! 989: addl lzistepx,%ebp ! 990: adcl $0,%ebp ! 991: addl C(a_sstepxfrac),%ebx ! 992: adcl advancetable+4(,%eax,4),%esi ! 993: ! 994: addl $8,%edi ! 995: addl $16,%ecx ! 996: ! 997: decw %bx ! 998: jnz LDrawLoop ! 999: ! 1000: popl %esi // restore spans pointer ! 1001: LNextSpan: ! 1002: addl $(spanpackage_t_size),%esi // point to next span ! 1003: LNextSpanESISet: ! 1004: movl spanpackage_t_count(%esi),%edx ! 1005: cmpl $-999999,%edx // any more spans? ! 1006: jnz LSpanLoop // yes ! 1007: ! 1008: popl %edi ! 1009: popl %ebp // restore the caller's stack frame ! 1010: popl %ebx // restore register variables ! 1011: popl %esi ! 1012: ret ! 1013: ! 1014: ! 1015: // draw a one-long span ! 1016: ! 1017: LExactlyOneLong: ! 1018: ! 1019: movl spanpackage_t_pz(%esi),%ecx ! 1020: movl spanpackage_t_zi(%esi),%ebp ! 1021: ! 1022: rorl $16,%ebp // put high 16 bits of 1/z in low word ! 1023: movl spanpackage_t_ptex(%esi),%ebx ! 1024: ! 1025: cmpw (%ecx),%bp ! 1026: jl LNextSpan ! 1027: xorl %eax,%eax ! 1028: movl spanpackage_t_pdest(%esi),%edi ! 1029: movb spanpackage_t_light+1(%esi),%ah ! 1030: addl $(spanpackage_t_size),%esi // point to next span ! 1031: movb (%ebx),%al ! 1032: movw %bp,(%ecx) ! 1033: movb 0x12345678(%eax),%al ! 1034: LPatch9: ! 1035: movb %al,(%edi) ! 1036: ! 1037: jmp LNextSpanESISet ! 1038: ! 1039: .globl C(D_PolysetAff8End) ! 1040: C(D_PolysetAff8End): ! 1041: ! 1042: ! 1043: #define pcolormap 4 ! 1044: ! 1045: .globl C(D_Aff8Patch) ! 1046: C(D_Aff8Patch): ! 1047: movl pcolormap(%esp),%eax ! 1048: movl %eax,LPatch1-4 ! 1049: movl %eax,LPatch2-4 ! 1050: movl %eax,LPatch3-4 ! 1051: movl %eax,LPatch4-4 ! 1052: movl %eax,LPatch5-4 ! 1053: movl %eax,LPatch6-4 ! 1054: movl %eax,LPatch7-4 ! 1055: movl %eax,LPatch8-4 ! 1056: movl %eax,LPatch9-4 ! 1057: ! 1058: ret ! 1059: ! 1060: ! 1061: //---------------------------------------------------------------------- ! 1062: // Alias model polygon dispatching code, combined with subdivided affine ! 1063: // triangle drawing code ! 1064: //---------------------------------------------------------------------- ! 1065: ! 1066: .globl C(D_PolysetDraw) ! 1067: C(D_PolysetDraw): ! 1068: ! 1069: // spanpackage_t spans[DPS_MAXSPANS + 1 + ! 1070: // ((CACHE_SIZE - 1) / sizeof(spanpackage_t)) + 1]; ! 1071: // // one extra because of cache line pretouching ! 1072: // ! 1073: // a_spans = (spanpackage_t *) ! 1074: // (((long)&spans[0] + CACHE_SIZE - 1) & ~(CACHE_SIZE - 1)); ! 1075: subl $(SPAN_SIZE),%esp ! 1076: movl %esp,%eax ! 1077: addl $(CACHE_SIZE - 1),%eax ! 1078: andl $(~(CACHE_SIZE - 1)),%eax ! 1079: movl %eax,C(a_spans) ! 1080: ! 1081: // if (r_affinetridesc.drawtype) ! 1082: // D_DrawSubdiv (); ! 1083: // else ! 1084: // D_DrawNonSubdiv (); ! 1085: movl C(r_affinetridesc)+atd_drawtype,%eax ! 1086: testl %eax,%eax ! 1087: jz C(D_DrawNonSubdiv) ! 1088: ! 1089: pushl %ebp // preserve caller stack frame pointer ! 1090: ! 1091: // lnumtriangles = r_affinetridesc.numtriangles; ! 1092: movl C(r_affinetridesc)+atd_numtriangles,%ebp ! 1093: ! 1094: pushl %esi // preserve register variables ! 1095: shll $4,%ebp ! 1096: ! 1097: pushl %ebx ! 1098: // ptri = r_affinetridesc.ptriangles; ! 1099: movl C(r_affinetridesc)+atd_ptriangles,%ebx ! 1100: ! 1101: pushl %edi ! 1102: ! 1103: // mtriangle_t *ptri; ! 1104: // finalvert_t *pfv, *index0, *index1, *index2; ! 1105: // int i; ! 1106: // int lnumtriangles; ! 1107: // int s0, s1, s2; ! 1108: ! 1109: // pfv = r_affinetridesc.pfinalverts; ! 1110: movl C(r_affinetridesc)+atd_pfinalverts,%edi ! 1111: ! 1112: // for (i=0 ; i<lnumtriangles ; i++) ! 1113: // { ! 1114: ! 1115: Llooptop: ! 1116: ! 1117: // index0 = pfv + ptri[i].vertindex[0]; ! 1118: // index1 = pfv + ptri[i].vertindex[1]; ! 1119: // index2 = pfv + ptri[i].vertindex[2]; ! 1120: movl mtri_vertindex-16+0(%ebx,%ebp,),%ecx ! 1121: movl mtri_vertindex-16+4(%ebx,%ebp,),%esi ! 1122: ! 1123: shll $(fv_shift),%ecx ! 1124: movl mtri_vertindex-16+8(%ebx,%ebp,),%edx ! 1125: ! 1126: shll $(fv_shift),%esi ! 1127: addl %edi,%ecx ! 1128: ! 1129: shll $(fv_shift),%edx ! 1130: addl %edi,%esi ! 1131: ! 1132: addl %edi,%edx ! 1133: ! 1134: // if (((index0->v[1]-index1->v[1]) * ! 1135: // (index0->v[0]-index2->v[0]) - ! 1136: // (index0->v[0]-index1->v[0])*(index0->v[1]-index2->v[1])) >= 0) ! 1137: // { ! 1138: // continue; ! 1139: // } ! 1140: // ! 1141: // d_pcolormap = &((byte *)acolormap)[index0->v[4] & 0xFF00]; ! 1142: fildl fv_v+4(%ecx) // i0v1 ! 1143: fildl fv_v+4(%esi) // i1v1 | i0v1 ! 1144: fildl fv_v+0(%ecx) // i0v0 | i1v1 | i0v1 ! 1145: fildl fv_v+0(%edx) // i2v0 | i0v0 | i1v1 | i0v1 ! 1146: fxch %st(2) // i1v1 | i0v0 | i2v0 | i0v1 ! 1147: fsubr %st(3),%st(0) // i0v1-i1v1 | i0v0 | i2v0 | i0v1 ! 1148: fildl fv_v+0(%esi) // i1v0 | i0v1-i1v1 | i0v0 | i2v0 | i0v1 ! 1149: fxch %st(2) // i0v0 | i0v1-i1v1 | i1v0 | i2v0 | i0v1 ! 1150: fsub %st(0),%st(3) // i0v0 | i0v1-i1v1 | i1v0 | i0v0-i2v0 | i0v1 ! 1151: fildl fv_v+4(%edx) // i2v1 | i0v0 | i0v1-i1v1 | i1v0 | i0v0-i2v0| i0v1 ! 1152: fxch %st(1) // i0v0 | i2v1 | i0v1-i1v1 | i1v0 | i0v0-i2v0| i0v1 ! 1153: fsubp %st(0),%st(3) // i2v1 | i0v1-i1v1 | i0v0-i1v0 | i0v0-i2v0 | i0v1 ! 1154: fxch %st(1) // i0v1-i1v1 | i2v1 | i0v0-i1v0 | i0v0-i2v0 | i0v1 ! 1155: fmulp %st(0),%st(3) // i2v1 | i0v0-i1v0 | i0v1-i1v1*i0v0-i2v0 | i0v1 ! 1156: fsubrp %st(0),%st(3) // i0v0-i1v0 | i0v1-i1v1*i0v0-i2v0 | i0v1-i2v1 ! 1157: movl fv_v+16(%ecx),%eax ! 1158: andl $0xFF00,%eax ! 1159: fmulp %st(0),%st(2) // i0v1-i1v1*i0v0-i2v0 | i0v0-i1v0*i0v1-i2v1 ! 1160: addl C(acolormap),%eax ! 1161: fsubp %st(0),%st(1) // (i0v1-i1v1)*(i0v0-i2v0)-(i0v0-i1v0)*(i0v1-i2v1) ! 1162: movl %eax,C(d_pcolormap) ! 1163: fstps Ltemp ! 1164: movl Ltemp,%eax ! 1165: subl $0x80000001,%eax ! 1166: jc Lskip ! 1167: ! 1168: // if (ptri[i].facesfront) ! 1169: // { ! 1170: // D_PolysetRecursiveTriangle(index0->v, index1->v, index2->v); ! 1171: movl mtri_facesfront-16(%ebx,%ebp,),%eax ! 1172: testl %eax,%eax ! 1173: jz Lfacesback ! 1174: ! 1175: pushl %edx ! 1176: pushl %esi ! 1177: pushl %ecx ! 1178: call C(D_PolysetRecursiveTriangle) ! 1179: ! 1180: subl $16,%ebp ! 1181: jnz Llooptop ! 1182: jmp Ldone2 ! 1183: ! 1184: // } ! 1185: // else ! 1186: // { ! 1187: Lfacesback: ! 1188: ! 1189: // s0 = index0->v[2]; ! 1190: // s1 = index1->v[2]; ! 1191: // s2 = index2->v[2]; ! 1192: movl fv_v+8(%ecx),%eax ! 1193: pushl %eax ! 1194: movl fv_v+8(%esi),%eax ! 1195: pushl %eax ! 1196: movl fv_v+8(%edx),%eax ! 1197: pushl %eax ! 1198: pushl %ecx ! 1199: pushl %edx ! 1200: ! 1201: // if (index0->flags & ALIAS_ONSEAM) ! 1202: // index0->v[2] += r_affinetridesc.seamfixupX16; ! 1203: movl C(r_affinetridesc)+atd_seamfixupX16,%eax ! 1204: testl $(ALIAS_ONSEAM),fv_flags(%ecx) ! 1205: jz Lp11 ! 1206: addl %eax,fv_v+8(%ecx) ! 1207: Lp11: ! 1208: ! 1209: // if (index1->flags & ALIAS_ONSEAM) ! 1210: // index1->v[2] += r_affinetridesc.seamfixupX16; ! 1211: testl $(ALIAS_ONSEAM),fv_flags(%esi) ! 1212: jz Lp12 ! 1213: addl %eax,fv_v+8(%esi) ! 1214: Lp12: ! 1215: ! 1216: // if (index2->flags & ALIAS_ONSEAM) ! 1217: // index2->v[2] += r_affinetridesc.seamfixupX16; ! 1218: testl $(ALIAS_ONSEAM),fv_flags(%edx) ! 1219: jz Lp13 ! 1220: addl %eax,fv_v+8(%edx) ! 1221: Lp13: ! 1222: ! 1223: // D_PolysetRecursiveTriangle(index0->v, index1->v, index2->v); ! 1224: pushl %edx ! 1225: pushl %esi ! 1226: pushl %ecx ! 1227: call C(D_PolysetRecursiveTriangle) ! 1228: ! 1229: // index0->v[2] = s0; ! 1230: // index1->v[2] = s1; ! 1231: // index2->v[2] = s2; ! 1232: popl %edx ! 1233: popl %ecx ! 1234: popl %eax ! 1235: movl %eax,fv_v+8(%edx) ! 1236: popl %eax ! 1237: movl %eax,fv_v+8(%esi) ! 1238: popl %eax ! 1239: movl %eax,fv_v+8(%ecx) ! 1240: ! 1241: // } ! 1242: // } ! 1243: Lskip: ! 1244: subl $16,%ebp ! 1245: jnz Llooptop ! 1246: ! 1247: Ldone2: ! 1248: popl %edi // restore the caller's stack frame ! 1249: popl %ebx ! 1250: popl %esi // restore register variables ! 1251: popl %ebp ! 1252: ! 1253: addl $(SPAN_SIZE),%esp ! 1254: ! 1255: ret ! 1256: ! 1257: ! 1258: //---------------------------------------------------------------------- ! 1259: // Alias model triangle left-edge scanning code ! 1260: //---------------------------------------------------------------------- ! 1261: ! 1262: #define height 4+16 ! 1263: ! 1264: .globl C(D_PolysetScanLeftEdge) ! 1265: C(D_PolysetScanLeftEdge): ! 1266: pushl %ebp // preserve caller stack frame pointer ! 1267: pushl %esi // preserve register variables ! 1268: pushl %edi ! 1269: pushl %ebx ! 1270: ! 1271: movl height(%esp),%eax ! 1272: movl C(d_sfrac),%ecx ! 1273: andl $0xFFFF,%eax ! 1274: movl C(d_ptex),%ebx ! 1275: orl %eax,%ecx ! 1276: movl C(d_pedgespanpackage),%esi ! 1277: movl C(d_tfrac),%edx ! 1278: movl C(d_light),%edi ! 1279: movl C(d_zi),%ebp ! 1280: ! 1281: // %eax: scratch ! 1282: // %ebx: d_ptex ! 1283: // %ecx: d_sfrac in high word, count in low word ! 1284: // %edx: d_tfrac ! 1285: // %esi: d_pedgespanpackage, errorterm, scratch alternately ! 1286: // %edi: d_light ! 1287: // %ebp: d_zi ! 1288: ! 1289: // do ! 1290: // { ! 1291: ! 1292: LScanLoop: ! 1293: ! 1294: // d_pedgespanpackage->ptex = ptex; ! 1295: // d_pedgespanpackage->pdest = d_pdest; ! 1296: // d_pedgespanpackage->pz = d_pz; ! 1297: // d_pedgespanpackage->count = d_aspancount; ! 1298: // d_pedgespanpackage->light = d_light; ! 1299: // d_pedgespanpackage->zi = d_zi; ! 1300: // d_pedgespanpackage->sfrac = d_sfrac << 16; ! 1301: // d_pedgespanpackage->tfrac = d_tfrac << 16; ! 1302: movl %ebx,spanpackage_t_ptex(%esi) ! 1303: movl C(d_pdest),%eax ! 1304: movl %eax,spanpackage_t_pdest(%esi) ! 1305: movl C(d_pz),%eax ! 1306: movl %eax,spanpackage_t_pz(%esi) ! 1307: movl C(d_aspancount),%eax ! 1308: movl %eax,spanpackage_t_count(%esi) ! 1309: movl %edi,spanpackage_t_light(%esi) ! 1310: movl %ebp,spanpackage_t_zi(%esi) ! 1311: movl %ecx,spanpackage_t_sfrac(%esi) ! 1312: movl %edx,spanpackage_t_tfrac(%esi) ! 1313: ! 1314: // pretouch the next cache line ! 1315: movb spanpackage_t_size(%esi),%al ! 1316: ! 1317: // d_pedgespanpackage++; ! 1318: addl $(spanpackage_t_size),%esi ! 1319: movl C(erroradjustup),%eax ! 1320: movl %esi,C(d_pedgespanpackage) ! 1321: ! 1322: // errorterm += erroradjustup; ! 1323: movl C(errorterm),%esi ! 1324: addl %eax,%esi ! 1325: movl C(d_pdest),%eax ! 1326: ! 1327: // if (errorterm >= 0) ! 1328: // { ! 1329: js LNoLeftEdgeTurnover ! 1330: ! 1331: // errorterm -= erroradjustdown; ! 1332: // d_pdest += d_pdestextrastep; ! 1333: subl C(erroradjustdown),%esi ! 1334: addl C(d_pdestextrastep),%eax ! 1335: movl %esi,C(errorterm) ! 1336: movl %eax,C(d_pdest) ! 1337: ! 1338: // d_pz += d_pzextrastep; ! 1339: // d_aspancount += d_countextrastep; ! 1340: // d_ptex += d_ptexextrastep; ! 1341: // d_sfrac += d_sfracextrastep; ! 1342: // d_ptex += d_sfrac >> 16; ! 1343: // d_sfrac &= 0xFFFF; ! 1344: // d_tfrac += d_tfracextrastep; ! 1345: movl C(d_pz),%eax ! 1346: movl C(d_aspancount),%esi ! 1347: addl C(d_pzextrastep),%eax ! 1348: addl C(d_sfracextrastep),%ecx ! 1349: adcl C(d_ptexextrastep),%ebx ! 1350: addl C(d_countextrastep),%esi ! 1351: movl %eax,C(d_pz) ! 1352: movl C(d_tfracextrastep),%eax ! 1353: movl %esi,C(d_aspancount) ! 1354: addl %eax,%edx ! 1355: ! 1356: // if (d_tfrac & 0x10000) ! 1357: // { ! 1358: jnc LSkip1 ! 1359: ! 1360: // d_ptex += r_affinetridesc.skinwidth; ! 1361: // d_tfrac &= 0xFFFF; ! 1362: addl C(r_affinetridesc)+atd_skinwidth,%ebx ! 1363: ! 1364: // } ! 1365: ! 1366: LSkip1: ! 1367: ! 1368: // d_light += d_lightextrastep; ! 1369: // d_zi += d_ziextrastep; ! 1370: addl C(d_lightextrastep),%edi ! 1371: addl C(d_ziextrastep),%ebp ! 1372: ! 1373: // } ! 1374: movl C(d_pedgespanpackage),%esi ! 1375: decl %ecx ! 1376: testl $0xFFFF,%ecx ! 1377: jnz LScanLoop ! 1378: ! 1379: popl %ebx ! 1380: popl %edi ! 1381: popl %esi ! 1382: popl %ebp ! 1383: ret ! 1384: ! 1385: // else ! 1386: // { ! 1387: ! 1388: LNoLeftEdgeTurnover: ! 1389: movl %esi,C(errorterm) ! 1390: ! 1391: // d_pdest += d_pdestbasestep; ! 1392: addl C(d_pdestbasestep),%eax ! 1393: movl %eax,C(d_pdest) ! 1394: ! 1395: // d_pz += d_pzbasestep; ! 1396: // d_aspancount += ubasestep; ! 1397: // d_ptex += d_ptexbasestep; ! 1398: // d_sfrac += d_sfracbasestep; ! 1399: // d_ptex += d_sfrac >> 16; ! 1400: // d_sfrac &= 0xFFFF; ! 1401: movl C(d_pz),%eax ! 1402: movl C(d_aspancount),%esi ! 1403: addl C(d_pzbasestep),%eax ! 1404: addl C(d_sfracbasestep),%ecx ! 1405: adcl C(d_ptexbasestep),%ebx ! 1406: addl C(ubasestep),%esi ! 1407: movl %eax,C(d_pz) ! 1408: movl %esi,C(d_aspancount) ! 1409: ! 1410: // d_tfrac += d_tfracbasestep; ! 1411: movl C(d_tfracbasestep),%esi ! 1412: addl %esi,%edx ! 1413: ! 1414: // if (d_tfrac & 0x10000) ! 1415: // { ! 1416: jnc LSkip2 ! 1417: ! 1418: // d_ptex += r_affinetridesc.skinwidth; ! 1419: // d_tfrac &= 0xFFFF; ! 1420: addl C(r_affinetridesc)+atd_skinwidth,%ebx ! 1421: ! 1422: // } ! 1423: ! 1424: LSkip2: ! 1425: ! 1426: // d_light += d_lightbasestep; ! 1427: // d_zi += d_zibasestep; ! 1428: addl C(d_lightbasestep),%edi ! 1429: addl C(d_zibasestep),%ebp ! 1430: ! 1431: // } ! 1432: // } while (--height); ! 1433: movl C(d_pedgespanpackage),%esi ! 1434: decl %ecx ! 1435: testl $0xFFFF,%ecx ! 1436: jnz LScanLoop ! 1437: ! 1438: popl %ebx ! 1439: popl %edi ! 1440: popl %esi ! 1441: popl %ebp ! 1442: ret ! 1443: ! 1444: ! 1445: //---------------------------------------------------------------------- ! 1446: // Alias model vertex drawing code ! 1447: //---------------------------------------------------------------------- ! 1448: ! 1449: #define fv 4+8 ! 1450: #define numverts 8+8 ! 1451: ! 1452: .globl C(D_PolysetDrawFinalVerts) ! 1453: C(D_PolysetDrawFinalVerts): ! 1454: pushl %ebp // preserve caller stack frame pointer ! 1455: pushl %ebx ! 1456: ! 1457: // int i, z; ! 1458: // short *zbuf; ! 1459: ! 1460: movl numverts(%esp),%ecx ! 1461: movl fv(%esp),%ebx ! 1462: ! 1463: pushl %esi // preserve register variables ! 1464: pushl %edi ! 1465: ! 1466: LFVLoop: ! 1467: ! 1468: // for (i=0 ; i<numverts ; i++, fv++) ! 1469: // { ! 1470: // // valid triangle coordinates for filling can include the bottom and ! 1471: // // right clip edges, due to the fill rule; these shouldn't be drawn ! 1472: // if ((fv->v[0] < r_refdef.vrectright) && ! 1473: // (fv->v[1] < r_refdef.vrectbottom)) ! 1474: // { ! 1475: movl fv_v+0(%ebx),%eax ! 1476: movl C(r_refdef)+rd_vrectright,%edx ! 1477: cmpl %edx,%eax ! 1478: jge LNextVert ! 1479: movl fv_v+4(%ebx),%esi ! 1480: movl C(r_refdef)+rd_vrectbottom,%edx ! 1481: cmpl %edx,%esi ! 1482: jge LNextVert ! 1483: ! 1484: // zbuf = zspantable[fv->v[1]] + fv->v[0]; ! 1485: movl C(zspantable)(,%esi,4),%edi ! 1486: ! 1487: // z = fv->v[5]>>16; ! 1488: movl fv_v+20(%ebx),%edx ! 1489: shrl $16,%edx ! 1490: ! 1491: // if (z >= *zbuf) ! 1492: // { ! 1493: // int pix; ! 1494: cmpw (%edi,%eax,2),%dx ! 1495: jl LNextVert ! 1496: ! 1497: // *zbuf = z; ! 1498: movw %dx,(%edi,%eax,2) ! 1499: ! 1500: // pix = skintable[fv->v[3]>>16][fv->v[2]>>16]; ! 1501: movl fv_v+12(%ebx),%edi ! 1502: shrl $16,%edi ! 1503: movl C(skintable)(,%edi,4),%edi ! 1504: movl fv_v+8(%ebx),%edx ! 1505: shrl $16,%edx ! 1506: movb (%edi,%edx),%dl ! 1507: ! 1508: // pix = ((byte *)acolormap)[pix + (fv->v[4] & 0xFF00)]; ! 1509: movl fv_v+16(%ebx),%edi ! 1510: andl $0xFF00,%edi ! 1511: andl $0x00FF,%edx ! 1512: addl %edx,%edi ! 1513: movl C(acolormap),%edx ! 1514: movb (%edx,%edi,1),%dl ! 1515: ! 1516: // d_viewbuffer[d_scantable[fv->v[1]] + fv->v[0]] = pix; ! 1517: movl C(d_scantable)(,%esi,4),%edi ! 1518: movl C(d_viewbuffer),%esi ! 1519: addl %eax,%edi ! 1520: movb %dl,(%esi,%edi) ! 1521: ! 1522: // } ! 1523: // } ! 1524: // } ! 1525: LNextVert: ! 1526: addl $(fv_size),%ebx ! 1527: decl %ecx ! 1528: jnz LFVLoop ! 1529: ! 1530: popl %edi ! 1531: popl %esi ! 1532: popl %ebx ! 1533: popl %ebp ! 1534: ret ! 1535: ! 1536: ! 1537: //---------------------------------------------------------------------- ! 1538: // Alias model non-subdivided polygon dispatching code ! 1539: // ! 1540: // not C-callable because of stack buffer cleanup ! 1541: //---------------------------------------------------------------------- ! 1542: ! 1543: .globl C(D_DrawNonSubdiv) ! 1544: C(D_DrawNonSubdiv): ! 1545: pushl %ebp // preserve caller stack frame pointer ! 1546: movl C(r_affinetridesc)+atd_numtriangles,%ebp ! 1547: pushl %ebx ! 1548: shll $(mtri_shift),%ebp ! 1549: pushl %esi // preserve register variables ! 1550: movl C(r_affinetridesc)+atd_ptriangles,%esi ! 1551: pushl %edi ! 1552: ! 1553: // mtriangle_t *ptri; ! 1554: // finalvert_t *pfv, *index0, *index1, *index2; ! 1555: // int i; ! 1556: // int lnumtriangles; ! 1557: ! 1558: // pfv = r_affinetridesc.pfinalverts; ! 1559: // ptri = r_affinetridesc.ptriangles; ! 1560: // lnumtriangles = r_affinetridesc.numtriangles; ! 1561: ! 1562: LNDLoop: ! 1563: ! 1564: // for (i=0 ; i<lnumtriangles ; i++, ptri++) ! 1565: // { ! 1566: // index0 = pfv + ptri->vertindex[0]; ! 1567: // index1 = pfv + ptri->vertindex[1]; ! 1568: // index2 = pfv + ptri->vertindex[2]; ! 1569: movl C(r_affinetridesc)+atd_pfinalverts,%edi ! 1570: movl mtri_vertindex+0-mtri_size(%esi,%ebp,1),%ecx ! 1571: shll $(fv_shift),%ecx ! 1572: movl mtri_vertindex+4-mtri_size(%esi,%ebp,1),%edx ! 1573: shll $(fv_shift),%edx ! 1574: movl mtri_vertindex+8-mtri_size(%esi,%ebp,1),%ebx ! 1575: shll $(fv_shift),%ebx ! 1576: addl %edi,%ecx ! 1577: addl %edi,%edx ! 1578: addl %edi,%ebx ! 1579: ! 1580: // d_xdenom = (index0->v[1]-index1->v[1]) * ! 1581: // (index0->v[0]-index2->v[0]) - ! 1582: // (index0->v[0]-index1->v[0])*(index0->v[1]-index2->v[1]); ! 1583: movl fv_v+4(%ecx),%eax ! 1584: movl fv_v+0(%ecx),%esi ! 1585: subl fv_v+4(%edx),%eax ! 1586: subl fv_v+0(%ebx),%esi ! 1587: imull %esi,%eax ! 1588: movl fv_v+0(%ecx),%esi ! 1589: movl fv_v+4(%ecx),%edi ! 1590: subl fv_v+0(%edx),%esi ! 1591: subl fv_v+4(%ebx),%edi ! 1592: imull %esi,%edi ! 1593: subl %edi,%eax ! 1594: ! 1595: // if (d_xdenom >= 0) ! 1596: // { ! 1597: // continue; ! 1598: jns LNextTri ! 1599: ! 1600: // } ! 1601: ! 1602: movl %eax,C(d_xdenom) ! 1603: fildl C(d_xdenom) ! 1604: ! 1605: // r_p0[0] = index0->v[0]; // u ! 1606: // r_p0[1] = index0->v[1]; // v ! 1607: // r_p0[2] = index0->v[2]; // s ! 1608: // r_p0[3] = index0->v[3]; // t ! 1609: // r_p0[4] = index0->v[4]; // light ! 1610: // r_p0[5] = index0->v[5]; // iz ! 1611: movl fv_v+0(%ecx),%eax ! 1612: movl fv_v+4(%ecx),%esi ! 1613: movl %eax,C(r_p0)+0 ! 1614: movl %esi,C(r_p0)+4 ! 1615: movl fv_v+8(%ecx),%eax ! 1616: movl fv_v+12(%ecx),%esi ! 1617: movl %eax,C(r_p0)+8 ! 1618: movl %esi,C(r_p0)+12 ! 1619: movl fv_v+16(%ecx),%eax ! 1620: movl fv_v+20(%ecx),%esi ! 1621: movl %eax,C(r_p0)+16 ! 1622: movl %esi,C(r_p0)+20 ! 1623: ! 1624: fdivrs float_1 ! 1625: ! 1626: // r_p1[0] = index1->v[0]; ! 1627: // r_p1[1] = index1->v[1]; ! 1628: // r_p1[2] = index1->v[2]; ! 1629: // r_p1[3] = index1->v[3]; ! 1630: // r_p1[4] = index1->v[4]; ! 1631: // r_p1[5] = index1->v[5]; ! 1632: movl fv_v+0(%edx),%eax ! 1633: movl fv_v+4(%edx),%esi ! 1634: movl %eax,C(r_p1)+0 ! 1635: movl %esi,C(r_p1)+4 ! 1636: movl fv_v+8(%edx),%eax ! 1637: movl fv_v+12(%edx),%esi ! 1638: movl %eax,C(r_p1)+8 ! 1639: movl %esi,C(r_p1)+12 ! 1640: movl fv_v+16(%edx),%eax ! 1641: movl fv_v+20(%edx),%esi ! 1642: movl %eax,C(r_p1)+16 ! 1643: movl %esi,C(r_p1)+20 ! 1644: ! 1645: // r_p2[0] = index2->v[0]; ! 1646: // r_p2[1] = index2->v[1]; ! 1647: // r_p2[2] = index2->v[2]; ! 1648: // r_p2[3] = index2->v[3]; ! 1649: // r_p2[4] = index2->v[4]; ! 1650: // r_p2[5] = index2->v[5]; ! 1651: movl fv_v+0(%ebx),%eax ! 1652: movl fv_v+4(%ebx),%esi ! 1653: movl %eax,C(r_p2)+0 ! 1654: movl %esi,C(r_p2)+4 ! 1655: movl fv_v+8(%ebx),%eax ! 1656: movl fv_v+12(%ebx),%esi ! 1657: movl %eax,C(r_p2)+8 ! 1658: movl %esi,C(r_p2)+12 ! 1659: movl fv_v+16(%ebx),%eax ! 1660: movl fv_v+20(%ebx),%esi ! 1661: movl %eax,C(r_p2)+16 ! 1662: movl C(r_affinetridesc)+atd_ptriangles,%edi ! 1663: movl %esi,C(r_p2)+20 ! 1664: movl mtri_facesfront-mtri_size(%edi,%ebp,1),%eax ! 1665: ! 1666: // if (!ptri->facesfront) ! 1667: // { ! 1668: testl %eax,%eax ! 1669: jnz LFacesFront ! 1670: ! 1671: // if (index0->flags & ALIAS_ONSEAM) ! 1672: // r_p0[2] += r_affinetridesc.seamfixupX16; ! 1673: movl fv_flags(%ecx),%eax ! 1674: movl fv_flags(%edx),%esi ! 1675: movl fv_flags(%ebx),%edi ! 1676: testl $(ALIAS_ONSEAM),%eax ! 1677: movl C(r_affinetridesc)+atd_seamfixupX16,%eax ! 1678: jz LOnseamDone0 ! 1679: addl %eax,C(r_p0)+8 ! 1680: LOnseamDone0: ! 1681: ! 1682: // if (index1->flags & ALIAS_ONSEAM) ! 1683: // r_p1[2] += r_affinetridesc.seamfixupX16; ! 1684: testl $(ALIAS_ONSEAM),%esi ! 1685: jz LOnseamDone1 ! 1686: addl %eax,C(r_p1)+8 ! 1687: LOnseamDone1: ! 1688: ! 1689: // if (index2->flags & ALIAS_ONSEAM) ! 1690: // r_p2[2] += r_affinetridesc.seamfixupX16; ! 1691: testl $(ALIAS_ONSEAM),%edi ! 1692: jz LOnseamDone2 ! 1693: addl %eax,C(r_p2)+8 ! 1694: LOnseamDone2: ! 1695: ! 1696: // } ! 1697: ! 1698: LFacesFront: ! 1699: ! 1700: fstps C(d_xdenom) ! 1701: ! 1702: // D_PolysetSetEdgeTable (); ! 1703: // D_RasterizeAliasPolySmooth (); ! 1704: call C(D_PolysetSetEdgeTable) ! 1705: call C(D_RasterizeAliasPolySmooth) ! 1706: ! 1707: LNextTri: ! 1708: movl C(r_affinetridesc)+atd_ptriangles,%esi ! 1709: subl $16,%ebp ! 1710: jnz LNDLoop ! 1711: // } ! 1712: ! 1713: popl %edi ! 1714: popl %esi ! 1715: popl %ebx ! 1716: popl %ebp ! 1717: ! 1718: addl $(SPAN_SIZE),%esp ! 1719: ! 1720: ret ! 1721: ! 1722: ! 1723: #endif // id386 ! 1724:
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.