| | 1 | /* |
| | 2 | * Copyright (C) 2011 Gilead Kutnick "Exophase" <exophase@gmail.com> |
| | 3 | * |
| | 4 | * This program is free software; you can redistribute it and/or |
| | 5 | * modify it under the terms of the GNU General Public License as |
| | 6 | * published by the Free Software Foundation; either version 2 of |
| | 7 | * the License, or (at your option) any later version. |
| | 8 | * |
| | 9 | * This program is distributed in the hope that it will be useful, |
| | 10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| | 11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| | 12 | * General Public License for more details. |
| | 13 | */ |
| | 14 | |
| | 15 | #define MAX_SPANS 512 |
| | 16 | #define MAX_BLOCKS 64 |
| | 17 | #define MAX_BLOCKS_PER_ROW 128 |
| | 18 | |
| | 19 | #include "psx_gpu_offsets.h" |
| | 20 | |
| | 21 | #define psx_gpu_b_dx_offset (psx_gpu_b_block_span_offset + 4) |
| | 22 | |
| | 23 | #define edge_data_left_x_offset 0 |
| | 24 | #define edge_data_num_blocks_offset 2 |
| | 25 | #define edge_data_right_mask_offset 4 |
| | 26 | #define edge_data_y_offset 6 |
| | 27 | |
| | 28 | |
| | 29 | #define psx_gpu r0 |
| | 30 | #define v_a r1 |
| | 31 | #define v_b r2 |
| | 32 | #define v_c r3 |
| | 33 | |
| | 34 | #define x0 r4 |
| | 35 | #define x1 r5 |
| | 36 | #define x2 r6 |
| | 37 | #define x0_x1 r5 |
| | 38 | #define x1_x2 r6 |
| | 39 | #define y0 r7 |
| | 40 | #define y1 r8 |
| | 41 | #define y2 r9 |
| | 42 | #define y0_y1 r7 |
| | 43 | #define y1_y2 r8 |
| | 44 | #define b0 r9 |
| | 45 | #define b1 r10 |
| | 46 | #define b2 r11 |
| | 47 | #define b0_b1 r10 |
| | 48 | #define b1_b2 r11 |
| | 49 | |
| | 50 | |
| | 51 | #define area_r_s r5 |
| | 52 | |
| | 53 | #define g_bx0 r2 |
| | 54 | #define g_bx r3 |
| | 55 | #define g_bx2 r4 |
| | 56 | #define g_bx3 r5 |
| | 57 | #define b_base r6 |
| | 58 | #define g_by r8 |
| | 59 | |
| | 60 | #define gs_bx r7 |
| | 61 | #define gs_by r10 |
| | 62 | |
| | 63 | #define ga_bx g_bx |
| | 64 | #define ga_by g_by |
| | 65 | |
| | 66 | #define gw_bx_h g_bx |
| | 67 | #define gw_by_h g_by |
| | 68 | |
| | 69 | #define gw_bx_l r11 |
| | 70 | #define gw_by_l gw_bx_l |
| | 71 | |
| | 72 | #define store_a r0 |
| | 73 | #define store_b r1 |
| | 74 | #define store_inc r5 |
| | 75 | |
| | 76 | |
| | 77 | #define v0 q0 |
| | 78 | #define uvrgb0 d0 |
| | 79 | #define x0_y0 d1 |
| | 80 | |
| | 81 | #define v1 q1 |
| | 82 | #define uvrgb1 d2 |
| | 83 | #define x1_y1 d3 |
| | 84 | |
| | 85 | #define v2 q2 |
| | 86 | #define uvrgb2 d4 |
| | 87 | #define x2_y2 d5 |
| | 88 | |
| | 89 | #define x0_ab q3 |
| | 90 | #define uvrg_xxxx0 q3 |
| | 91 | #define uvrg0 d6 |
| | 92 | #define xxxx0 d7 |
| | 93 | |
| | 94 | #define x1_ab q4 |
| | 95 | #define uvrg_xxxx1 q4 |
| | 96 | #define uvrg1 d8 |
| | 97 | #define xxxx1 d9 |
| | 98 | |
| | 99 | #define x2_ab q5 |
| | 100 | #define uvrg_xxxx2 q5 |
| | 101 | #define uvrg2 d10 |
| | 102 | #define xxxx2 d11 |
| | 103 | |
| | 104 | #define y0_ab q6 |
| | 105 | #define yyyy_uvrg0 q6 |
| | 106 | #define yyyy0 d12 |
| | 107 | #define uvrg0b d13 |
| | 108 | |
| | 109 | #define y1_ab q7 |
| | 110 | #define yyyy_uvrg1 q7 |
| | 111 | #define yyyy1 d14 |
| | 112 | #define uvrg1b d15 |
| | 113 | |
| | 114 | #define y2_ab q8 |
| | 115 | #define yyyy_uvrg2 q8 |
| | 116 | #define yyyy2 d16 |
| | 117 | #define uvrg2b d17 |
| | 118 | |
| | 119 | #define d0_ab q9 |
| | 120 | #define d0_a d18 |
| | 121 | #define d0_b d19 |
| | 122 | |
| | 123 | #define d1_ab q10 |
| | 124 | #define d1_a d20 |
| | 125 | #define d1_b d21 |
| | 126 | |
| | 127 | #define d2_ab q11 |
| | 128 | #define d2_a d22 |
| | 129 | #define d2_b d23 |
| | 130 | |
| | 131 | #define d3_ab q12 |
| | 132 | #define d3_a d24 |
| | 133 | #define d3_b d25 |
| | 134 | |
| | 135 | #define ga_uvrg_x q1 |
| | 136 | #define ga_uvrg_y q4 |
| | 137 | |
| | 138 | #define dx x0_x1 |
| | 139 | #define dy y0_y1 |
| | 140 | #define db b0_b1 |
| | 141 | |
| | 142 | #define uvrg_base q11 |
| | 143 | |
| | 144 | #define gs_uvrg_x q5 |
| | 145 | #define gs_uvrg_y q6 |
| | 146 | |
| | 147 | #define g_uvrg_x q1 |
| | 148 | #define ga_uv_x d2 |
| | 149 | #define g_uv_x d2 |
| | 150 | #define ga_rg_x d3 |
| | 151 | #define g_rg_x d3 |
| | 152 | |
| | 153 | #define g_uvrg_y q4 |
| | 154 | #define ga_uv_y d8 |
| | 155 | #define g_uv_y d8 |
| | 156 | #define ga_rg_y d9 |
| | 157 | #define g_rg_y d9 |
| | 158 | |
| | 159 | #define gw_uv_x q1 |
| | 160 | #define gw_rg_x q2 |
| | 161 | #define gw_uv_y q4 |
| | 162 | #define gw_rg_y q3 |
| | 163 | |
| | 164 | #define w_mask q9 |
| | 165 | #define w_mask_l d18 |
| | 166 | |
| | 167 | #define r_shift q10 |
| | 168 | |
| | 169 | #define uvrg_dx0 q0 |
| | 170 | #define uvrg_dx0l d0 |
| | 171 | #define uvrg_dx0h d1 |
| | 172 | |
| | 173 | #define uvrg_dx1 q1 |
| | 174 | #define uvrg_dx1l d2 |
| | 175 | #define uvrg_dx1h d3 |
| | 176 | |
| | 177 | #define uvrg_dx2 q2 |
| | 178 | #define uvrg_dx2l d4 |
| | 179 | #define uvrg_dx2h d5 |
| | 180 | |
| | 181 | #define uvrg_dx3 q3 |
| | 182 | #define uvrg_dx3l d6 |
| | 183 | #define uvrg_dx3h d7 |
| | 184 | |
| | 185 | |
| | 186 | .align 4 |
| | 187 | |
| | 188 | /* FIXME: users of this should be in psx_gpu instead */ |
| | 189 | #ifndef __PIC__ |
| | 190 | #define load_pointer(register, pointer) \ |
| | 191 | movw register, :lower16:pointer; \ |
| | 192 | movt register, :upper16:pointer; \ |
| | 193 | |
| | 194 | #else |
| | 195 | #define load_pointer(register, pointer) \ |
| | 196 | ldr register, =pointer \ |
| | 197 | |
| | 198 | #endif |
| | 199 | |
| | 200 | #define function(name) \ |
| | 201 | .global name; \ |
| | 202 | name: \ |
| | 203 | |
| | 204 | @ r0: psx_gpu |
| | 205 | @ r1: v_a |
| | 206 | @ r2: v_b |
| | 207 | @ r3: v_c |
| | 208 | |
| | 209 | function(compute_all_gradients) |
| | 210 | // First compute the triangle area reciprocal and shift. The division will |
| | 211 | // happen concurrently with much of the work which follows. |
| | 212 | @ r12 = psx_gpu->triangle_area |
| | 213 | ldr r12, [ psx_gpu, #psx_gpu_triangle_area_offset ] |
| | 214 | stmdb sp!, { r4 - r11, lr } |
| | 215 | |
| | 216 | @ load exponent of 62 into upper half of double |
| | 217 | movw r4, #0 |
| | 218 | clz r14, r12 @ r14 = shift |
| | 219 | |
| | 220 | movt r4, #((62 + 1023) << 4) |
| | 221 | mov r12, r12, lsl r14 @ r12 = triangle_area_normalized |
| | 222 | |
| | 223 | @ load area normalized into lower half of double |
| | 224 | mov r5, r12, lsr #10 |
| | 225 | vmov.f64 d30, r5, r4 @ d30 = (1 << 62) + ta_n |
| | 226 | |
| | 227 | movt r4, #((1022 + 31) << 4) |
| | 228 | mov r5, r12, lsl #20 |
| | 229 | |
| | 230 | add r4, r4, r12, lsr #11 |
| | 231 | vmov.f64 d31, r5, r4 |
| | 232 | |
| | 233 | vdiv.f64 d30, d30, d31 @ d30 = ((1 << 62) + ta_n) / ta_n |
| | 234 | |
| | 235 | // ((x1 - x0) * (y2 - y1)) - ((x2 - x1) * (y1 - y0)) = |
| | 236 | // ( d0 * d1 ) - ( d2 * d3 ) = |
| | 237 | // ( m0 ) - ( m1 ) = gradient |
| | 238 | |
| | 239 | // This is split to do 12 elements at a time over three sets: a, b, and c. |
| | 240 | // Technically we only need to do 10 elements (uvrgb_x and uvrgb_y), so |
| | 241 | // two of the slots are unused. |
| | 242 | |
| | 243 | // Inputs are all 16-bit signed. The m0/m1 results are 32-bit signed, as |
| | 244 | // is g. |
| | 245 | |
| | 246 | // First type is: uvrg bxxx xxxx |
| | 247 | // Second type is: yyyy ybyy uvrg |
| | 248 | // Since x_a and y_c are the same the same variable is used for both. |
| | 249 | |
| | 250 | vld1.u32 { v0 }, [ v_a, : 128 ] @ v0 = { uvrg0, b0, x0, y0 } |
| | 251 | ldrsh x0, [ v_a, #8 ] @ load x0 |
| | 252 | |
| | 253 | vld1.u32 { v1 }, [ v_b, : 128 ] @ v1 = { uvrg1, b1, x1, y1} |
| | 254 | ldrh x1, [ v_b, #8 ] @ load x1 |
| | 255 | |
| | 256 | vld1.u32 { v2 }, [ v_c, : 128 ] @ v2 = { uvrg2, b2, x2, y2 } |
| | 257 | ldrh x2, [ v_c, #8 ] @ load x2 |
| | 258 | |
| | 259 | vmovl.u8 uvrg_xxxx0, uvrgb0 @ uvrg_xxxx0 = { uv0, rg0, b0-, -- } |
| | 260 | ldrh y0, [ v_a, #10 ] @ load y0 |
| | 261 | |
| | 262 | vmovl.u8 uvrg_xxxx1, uvrgb1 @ uvrg_xxxx1 = { uv1, rg1, b1-, -- } |
| | 263 | ldrh y1, [ v_b, #10 ] @ load y1 |
| | 264 | |
| | 265 | vmovl.u8 uvrg_xxxx2, uvrgb2 @ uvrg_xxxx2 = { uv2, rg2, b2-, -- } |
| | 266 | ldrh y2, [ v_c, #10 ] @ load y2 |
| | 267 | |
| | 268 | vmov.u8 uvrg0b, uvrg0 @ uvrg0b = { uv0, rg0 } |
| | 269 | vdup.u16 xxxx0, x0_y0[0] @ xxxx0 = { xx0, xx0 } |
| | 270 | |
| | 271 | orr x1_x2, x1, x2, lsl #16 @ x1_x2 = { x1, x2 } |
| | 272 | pkhbt x0_x1, x0, x1, lsl #16 @ x0_x1 = { x0, x1 } |
| | 273 | |
| | 274 | vmov.u8 uvrg1b, uvrg1 @ uvrg1b = { uv1, rg1 } |
| | 275 | vdup.u16 xxxx1, x1_y1[0] @ xxxx1 = { xx1, xx1 } |
| | 276 | |
| | 277 | vmov.u8 uvrg2b, uvrg2 @ uvrg2b = { uv2, rg2 } |
| | 278 | vdup.u16 xxxx2, x2_y2[0] @ xxxx2 = { xx2, xx2 } |
| | 279 | |
| | 280 | ldrb b2, [ v_c, #4 ] @ load b2 |
| | 281 | orr y0_y1, y0, y1, lsl #16 @ y0_y1 = { y0, y1 } |
| | 282 | |
| | 283 | ldrb b1, [ v_b, #4 ] @ load b1 |
| | 284 | orr y1_y2, y1, y2, lsl #16 @ y1_y2 = { y1, y2 } |
| | 285 | |
| | 286 | vdup.u16 yyyy0, x0_y0[1] @ yyyy0 = { yy0, yy0 } |
| | 287 | vsub.s16 d0_ab, x1_ab, x0_ab |
| | 288 | |
| | 289 | ldrb b0, [ v_a, #4 ] @ load b0 |
| | 290 | orr b1_b2, b1, b2, lsl #16 @ b1_b2 = { b1, b2 } |
| | 291 | |
| | 292 | vdup.u16 yyyy1, x1_y1[1] @ yyyy1 = { yy1, yy1 } |
| | 293 | vsub.s16 d2_ab, x2_ab, x1_ab |
| | 294 | |
| | 295 | vdup.u16 yyyy2, x2_y2[1] @ yyyy2 = { yy2, yy2 } |
| | 296 | vsub.s16 d1_ab, y2_ab, y1_ab |
| | 297 | |
| | 298 | orr b0_b1, b0, b1, lsl #16 @ b1_b2 = { b1, b2 } |
| | 299 | ssub16 dx, x1_x2, x0_x1 @ dx = { x1 - x0, x2 - x1 } |
| | 300 | |
| | 301 | ssub16 dy, y1_y2, y0_y1 @ dy = { y1 - y0, y2 - y1 } |
| | 302 | ssub16 db, b1_b2, b0_b1 @ db = { b1 - b0, b2 - b1 } |
| | 303 | |
| | 304 | vsub.s16 d3_ab, y1_ab, y0_ab |
| | 305 | smusdx ga_by, dx, db @ ga_by = ((x1 - x0) * (b2 - b1)) - |
| | 306 | @ ((x2 - X1) * (b1 - b0)) |
| | 307 | vmull.s16 ga_uvrg_x, d0_a, d1_a |
| | 308 | smusdx ga_bx, db, dy @ ga_bx = ((b1 - b0) * (y2 - y1)) - |
| | 309 | @ ((b2 - b1) * (y1 - y0)) |
| | 310 | vmlsl.s16 ga_uvrg_x, d2_a, d3_a |
| | 311 | movs gs_bx, ga_bx, asr #31 |
| | 312 | |
| | 313 | vmull.s16 ga_uvrg_y, d0_b, d1_b |
| | 314 | rsbmi ga_bx, ga_bx, #0 |
| | 315 | |
| | 316 | vmlsl.s16 ga_uvrg_y, d2_b, d3_b |
| | 317 | movs gs_by, ga_by, asr #31 |
| | 318 | |
| | 319 | vshr.u64 d0, d30, #22 |
| | 320 | mov b_base, b0, lsl #16 |
| | 321 | |
| | 322 | rsbmi ga_by, ga_by, #0 |
| | 323 | vclt.s32 gs_uvrg_x, ga_uvrg_x, #0 @ gs_uvrg_x = ga_uvrg_x < 0 |
| | 324 | |
| | 325 | @ r12 = psx_gpu->triangle_winding_offset |
| | 326 | ldrb r12, [ psx_gpu, #psx_gpu_triangle_winding_offset ] |
| | 327 | vclt.s32 gs_uvrg_y, ga_uvrg_y, #0 @ gs_uvrg_y = ga_uvrg_y < 0 |
| | 328 | |
| | 329 | add b_base, b_base, #0x8000 |
| | 330 | rsb r12, r12, #0 @ r12 = -(triangle->winding) |
| | 331 | |
| | 332 | vdup.u32 w_mask, r12 @ w_mask = { -w, -w, -w, -w } |
| | 333 | sub r14, r14, #(62 - 12) @ r14 = shift - (62 - FIXED_BITS) |
| | 334 | |
| | 335 | vshll.u16 uvrg_base, uvrg0, #16 @ uvrg_base = uvrg0 << 16 |
| | 336 | vdup.u32 r_shift, r14 @ r_shift = { shift, shift, shift, shift } |
| | 337 | |
| | 338 | vorr.u32 uvrg_base, #0x8000 |
| | 339 | vabs.s32 ga_uvrg_x, ga_uvrg_x @ ga_uvrg_x = abs(ga_uvrg_x) |
| | 340 | |
| | 341 | vmov area_r_s, s0 @ area_r_s = triangle_reciprocal |
| | 342 | vabs.s32 ga_uvrg_y, ga_uvrg_y @ ga_uvrg_y = abs(ga_uvrg_y) |
| | 343 | |
| | 344 | vmull.u32 gw_rg_x, ga_rg_x, d0[0] |
| | 345 | vmull.u32 gw_uv_x, ga_uv_x, d0[0] |
| | 346 | vmull.u32 gw_rg_y, ga_rg_y, d0[0] |
| | 347 | vmull.u32 gw_uv_y, ga_uv_y, d0[0] |
| | 348 | |
| | 349 | vshl.u64 gw_rg_x, gw_rg_x, r_shift |
| | 350 | vshl.u64 gw_uv_x, gw_uv_x, r_shift |
| | 351 | vshl.u64 gw_rg_y, gw_rg_y, r_shift |
| | 352 | vshl.u64 gw_uv_y, gw_uv_y, r_shift |
| | 353 | |
| | 354 | veor.u32 gs_uvrg_x, gs_uvrg_x, w_mask |
| | 355 | vmovn.u64 g_uv_x, gw_uv_x |
| | 356 | |
| | 357 | veor.u32 gs_uvrg_y, gs_uvrg_y, w_mask |
| | 358 | vmovn.u64 g_rg_x, gw_rg_x |
| | 359 | |
| | 360 | veor.u32 g_uvrg_x, g_uvrg_x, gs_uvrg_x |
| | 361 | vmovn.u64 g_uv_y, gw_uv_y |
| | 362 | |
| | 363 | vsub.u32 g_uvrg_x, g_uvrg_x, gs_uvrg_x |
| | 364 | vmovn.u64 g_rg_y, gw_rg_y |
| | 365 | |
| | 366 | veor.u32 g_uvrg_y, g_uvrg_y, gs_uvrg_y |
| | 367 | mov ga_bx, ga_bx, lsl #13 |
| | 368 | |
| | 369 | vsub.u32 g_uvrg_y, g_uvrg_y, gs_uvrg_y |
| | 370 | mov ga_by, ga_by, lsl #13 |
| | 371 | |
| | 372 | vdup.u32 x0_y0, x0 |
| | 373 | umull gw_bx_l, gw_bx_h, ga_bx, area_r_s |
| | 374 | |
| | 375 | vshl.u32 g_uvrg_x, g_uvrg_x, #4 |
| | 376 | vshl.u32 g_uvrg_y, g_uvrg_y, #4 |
| | 377 | |
| | 378 | umull gw_by_l, gw_by_h, ga_by, area_r_s |
| | 379 | vmls.s32 uvrg_base, ga_uvrg_x, x0_y0[0] |
| | 380 | |
| | 381 | eor gs_bx, gs_bx, r12 |
| | 382 | vadd.u32 uvrg_dx2, uvrg_dx1, uvrg_dx1 |
| | 383 | |
| | 384 | veor.u32 uvrg_dx0, uvrg_dx0, uvrg_dx0 |
| | 385 | eor gs_by, gs_by, r12 |
| | 386 | |
| | 387 | rsb r11, r14, #0 @ r11 = negative shift for scalar lsr |
| | 388 | add store_a, psx_gpu, #psx_gpu_uvrg_offset |
| | 389 | |
| | 390 | sub r11, r11, #(32 - 13) |
| | 391 | |
| | 392 | add store_b, store_a, #16 |
| | 393 | mov store_inc, #32 |
| | 394 | |
| | 395 | vadd.u32 uvrg_dx3, uvrg_dx2, uvrg_dx1 |
| | 396 | vst1.u32 { uvrg_base }, [ store_a, : 128 ], store_inc |
| | 397 | |
| | 398 | vst1.u32 { uvrg_dx1 }, [ store_b, : 128 ], store_inc |
| | 399 | mov g_bx, gw_bx_h, lsr r11 |
| | 400 | |
| | 401 | vst1.u32 { g_uvrg_y }, [ store_a, : 128 ], store_inc |
| | 402 | mov g_by, gw_by_h, lsr r11 |
| | 403 | |
| | 404 | vst4.u32 { uvrg_dx0l, uvrg_dx1l, uvrg_dx2l, uvrg_dx3l }, \ |
| | 405 | [ store_b, : 128 ], store_inc |
| | 406 | eor g_bx, g_bx, gs_bx |
| | 407 | |
| | 408 | vst4.u32 { uvrg_dx0h, uvrg_dx1h, uvrg_dx2h, uvrg_dx3h }, \ |
| | 409 | [ store_b, : 128 ], store_inc |
| | 410 | sub g_bx, g_bx, gs_bx |
| | 411 | |
| | 412 | lsl g_bx, g_bx, #4 |
| | 413 | eor g_by, g_by, gs_by |
| | 414 | |
| | 415 | mls b_base, g_bx, x0, b_base |
| | 416 | sub g_by, g_by, gs_by |
| | 417 | |
| | 418 | lsl g_by, g_by, #4 |
| | 419 | mov g_bx0, #0 |
| | 420 | |
| | 421 | add g_bx2, g_bx, g_bx |
| | 422 | add g_bx3, g_bx, g_bx2 |
| | 423 | |
| | 424 | stmia store_b, { g_bx0, g_bx, g_bx2, g_bx3, b_base, g_by } |
| | 425 | |
| | 426 | ldmia sp!, { r4 - r11, pc } |
| | 427 | |
| | 428 | |
| | 429 | #define psx_gpu r0 |
| | 430 | #define v_a r1 |
| | 431 | #define v_b r2 |
| | 432 | #define v_c r3 |
| | 433 | |
| | 434 | #define temp r14 |
| | 435 | |
| | 436 | #define x_a r4 |
| | 437 | #define x_b r5 |
| | 438 | #define x_c r6 |
| | 439 | #define y_a r1 |
| | 440 | #define y_b r2 |
| | 441 | #define y_c r3 |
| | 442 | |
| | 443 | #define height_minor_a r7 |
| | 444 | #define height_minor_b r8 |
| | 445 | #define height_major r9 |
| | 446 | #define height r9 |
| | 447 | |
| | 448 | #define reciprocal_table_ptr r10 |
| | 449 | |
| | 450 | #define edge_alt_low r4 |
| | 451 | #define edge_alt_high r5 |
| | 452 | #define edge_dx_dy_alt r6 |
| | 453 | #define edge_shift_alt r10 |
| | 454 | |
| | 455 | #define edge_dx_dy_alt_low r4 |
| | 456 | #define edge_dx_dy_alt_high r5 |
| | 457 | |
| | 458 | #define span_edge_data r4 |
| | 459 | #define span_uvrg_offset r5 |
| | 460 | #define span_b_offset r6 |
| | 461 | |
| | 462 | #define clip r14 |
| | 463 | |
| | 464 | #define b r11 |
| | 465 | #define b_dy r12 |
| | 466 | |
| | 467 | |
| | 468 | #define alternate_x q0 |
| | 469 | #define alternate_dx_dy q1 |
| | 470 | #define alternate_x_32 q2 |
| | 471 | |
| | 472 | #define alternate_x_low d0 |
| | 473 | #define alternate_x_high d1 |
| | 474 | #define alternate_dx_dy_low d2 |
| | 475 | #define alternate_dx_dy_high d3 |
| | 476 | #define alternate_x_32_low d4 |
| | 477 | #define alternate_x_32_high d5 |
| | 478 | |
| | 479 | #define left_x q3 |
| | 480 | #define right_x q4 |
| | 481 | #define left_dx_dy q5 |
| | 482 | #define right_dx_dy q6 |
| | 483 | #define left_edge q7 |
| | 484 | #define right_edge q8 |
| | 485 | |
| | 486 | #define left_x_low d6 |
| | 487 | #define left_x_high d7 |
| | 488 | #define right_x_low d8 |
| | 489 | #define right_x_high d9 |
| | 490 | #define left_dx_dy_low d10 |
| | 491 | #define left_dx_dy_high d11 |
| | 492 | #define right_dx_dy_low d12 |
| | 493 | #define right_dx_dy_high d13 |
| | 494 | #define left_edge_low d14 |
| | 495 | #define left_edge_high d15 |
| | 496 | #define right_edge_low d16 |
| | 497 | #define right_edge_high d17 |
| | 498 | |
| | 499 | #define y_mid_point d18 |
| | 500 | #define c_0x0004 d19 |
| | 501 | |
| | 502 | #define left_right_x_16 q11 |
| | 503 | #define span_shifts_y q12 |
| | 504 | #define c_0x0001 q13 |
| | 505 | |
| | 506 | #define span_shifts d24 |
| | 507 | #define y_x4 d25 |
| | 508 | #define c_0xFFFE d26 |
| | 509 | #define c_0x0007 d27 |
| | 510 | |
| | 511 | #define left_right_x_16_low d22 |
| | 512 | #define left_right_x_16_high d23 |
| | 513 | |
| | 514 | #define uvrg q14 |
| | 515 | #define uvrg_dy q15 |
| | 516 | |
| | 517 | #define alternate_x_16 d4 |
| | 518 | |
| | 519 | #define v_clip q3 |
| | 520 | #define v_clip_low d6 |
| | 521 | |
| | 522 | #define right_x_32 q10 |
| | 523 | #define left_x_32 q11 |
| | 524 | #define alternate_select d24 |
| | 525 | |
| | 526 | #define right_x_32_low d20 |
| | 527 | #define right_x_32_high d21 |
| | 528 | #define left_x_32_low d22 |
| | 529 | #define left_x_32_high d23 |
| | 530 | |
| | 531 | #define edges_xy q0 |
| | 532 | #define edges_dx_dy d2 |
| | 533 | #define edge_shifts d3 |
| | 534 | #define edge_shifts_64 q2 |
| | 535 | |
| | 536 | #define edges_xy_left d0 |
| | 537 | #define edges_xy_right d1 |
| | 538 | |
| | 539 | #define height_reciprocals d6 |
| | 540 | #define heights d7 |
| | 541 | |
| | 542 | #define widths d8 |
| | 543 | #define c_0x01 d9 |
| | 544 | #define x_starts d10 |
| | 545 | #define x_ends d11 |
| | 546 | |
| | 547 | #define heights_b d12 |
| | 548 | #define edges_dx_dy_64 q10 |
| | 549 | |
| | 550 | #define edges_dx_dy_64_left d20 |
| | 551 | #define edges_dx_dy_64_right d21 |
| | 552 | |
| | 553 | |
| | 554 | #define setup_spans_prologue() \ |
| | 555 | stmdb sp!, { r4 - r11, lr }; \ |
| | 556 | \ |
| | 557 | ldrsh x_a, [ v_a, #8 ]; \ |
| | 558 | ldrsh x_b, [ v_b, #8 ]; \ |
| | 559 | ldrsh x_c, [ v_c, #8 ]; \ |
| | 560 | ldrsh y_a, [ v_a, #10 ]; \ |
| | 561 | ldrsh y_b, [ v_b, #10 ]; \ |
| | 562 | ldrsh y_c, [ v_c, #10 ]; \ |
| | 563 | \ |
| | 564 | add temp, psx_gpu, #psx_gpu_uvrg_offset; \ |
| | 565 | vld1.32 { uvrg }, [ temp ]; \ |
| | 566 | add temp, psx_gpu, #psx_gpu_uvrg_dy_offset; \ |
| | 567 | vld1.32 { uvrg_dy }, [ temp ]; \ |
| | 568 | load_pointer(reciprocal_table_ptr, reciprocal_table); \ |
| | 569 | \ |
| | 570 | vmov.u32 c_0x01, #0x01 \ |
| | 571 | |
| | 572 | #define setup_spans_load_b() \ |
| | 573 | ldr b, [ psx_gpu, #psx_gpu_b_offset ]; \ |
| | 574 | ldr b_dy, [ psx_gpu, #psx_gpu_b_dy_offset ] \ |
| | 575 | |
| | 576 | #define setup_spans_prologue_b() \ |
| | 577 | add span_uvrg_offset, psx_gpu, #psx_gpu_span_uvrg_offset_offset; \ |
| | 578 | add temp, psx_gpu, #psx_gpu_viewport_start_x_offset; \ |
| | 579 | \ |
| | 580 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset; \ |
| | 581 | vmov.u16 c_0x0004, #0x0004; \ |
| | 582 | \ |
| | 583 | add span_b_offset, psx_gpu, #psx_gpu_span_b_offset_offset; \ |
| | 584 | vmov.u16 c_0x0001, #0x0001; \ |
| | 585 | \ |
| | 586 | vld1.u16 { left_edge_low[], left_edge_high[] }, [ temp ]; \ |
| | 587 | add temp, psx_gpu, #psx_gpu_viewport_end_x_offset; \ |
| | 588 | \ |
| | 589 | vld1.u16 { right_edge_low[], right_edge_high[] }, [ temp ]; \ |
| | 590 | vadd.u16 right_edge, right_edge, c_0x0001; \ |
| | 591 | \ |
| | 592 | vmov.u16 c_0x0007, #0x0007; \ |
| | 593 | vmvn.u16 c_0xFFFE, #0x0001 \ |
| | 594 | |
| | 595 | |
| | 596 | #define compute_edge_delta_x2() \ |
| | 597 | ldr temp, [ reciprocal_table_ptr, height, lsl #2 ]; \ |
| | 598 | \ |
| | 599 | vdup.u32 heights, height; \ |
| | 600 | vsub.u32 widths, x_ends, x_starts; \ |
| | 601 | \ |
| | 602 | vdup.u32 edge_shifts, temp; \ |
| | 603 | vsub.u32 heights_b, heights, c_0x01; \ |
| | 604 | vshr.u32 height_reciprocals, edge_shifts, #10; \ |
| | 605 | \ |
| | 606 | vmla.s32 heights_b, x_starts, heights; \ |
| | 607 | vbic.u16 edge_shifts, #0xE0; \ |
| | 608 | vmul.s32 edges_dx_dy, widths, height_reciprocals; \ |
| | 609 | vmull.s32 edges_xy, heights_b, height_reciprocals \ |
| | 610 | |
| | 611 | #define width_alt r6 |
| | 612 | #define height_reciprocal_alt r11 |
| | 613 | #define height_b_alt r12 |
| | 614 | |
| | 615 | #define compute_edge_delta_x3(start_c, height_a, height_b) \ |
| | 616 | vmov.u32 heights, height_a, height_b; \ |
| | 617 | ldr temp, [ reciprocal_table_ptr, height_a, lsl #2 ]; \ |
| | 618 | vmov.u32 edge_shifts[0], temp; \ |
| | 619 | ldr temp, [ reciprocal_table_ptr, height_b, lsl #2 ]; \ |
| | 620 | vmov.u32 edge_shifts[1], temp; \ |
| | 621 | ldr edge_shift_alt, [ reciprocal_table_ptr, height_minor_b, lsl #2 ]; \ |
| | 622 | \ |
| | 623 | vsub.u32 widths, x_ends, x_starts; \ |
| | 624 | sub width_alt, x_c, start_c; \ |
| | 625 | \ |
| | 626 | vsub.u32 heights_b, heights, c_0x01; \ |
| | 627 | sub height_b_alt, height_minor_b, #1; \ |
| | 628 | \ |
| | 629 | vshr.u32 height_reciprocals, edge_shifts, #10; \ |
| | 630 | lsr height_reciprocal_alt, edge_shift_alt, #10; \ |
| | 631 | \ |
| | 632 | vmla.s32 heights_b, x_starts, heights; \ |
| | 633 | mla height_b_alt, height_minor_b, start_c, height_b_alt; \ |
| | 634 | \ |
| | 635 | vbic.u16 edge_shifts, #0xE0; \ |
| | 636 | and edge_shift_alt, edge_shift_alt, #0x1F; \ |
| | 637 | \ |
| | 638 | vmul.s32 edges_dx_dy, widths, height_reciprocals; \ |
| | 639 | mul edge_dx_dy_alt, width_alt, height_reciprocal_alt; \ |
| | 640 | \ |
| | 641 | vmull.s32 edges_xy, heights_b, height_reciprocals; \ |
| | 642 | smull edge_alt_low, edge_alt_high, height_b_alt, height_reciprocal_alt \ |
| | 643 | |
| | 644 | |
| | 645 | #define setup_spans_adjust_y_up() \ |
| | 646 | vsub.u32 y_x4, y_x4, c_0x0004 \ |
| | 647 | |
| | 648 | #define setup_spans_adjust_y_down() \ |
| | 649 | vadd.u32 y_x4, y_x4, c_0x0004 \ |
| | 650 | |
| | 651 | #define setup_spans_adjust_interpolants_up() \ |
| | 652 | vsub.u32 uvrg, uvrg, uvrg_dy; \ |
| | 653 | sub b, b, b_dy \ |
| | 654 | |
| | 655 | #define setup_spans_adjust_interpolants_down() \ |
| | 656 | vadd.u32 uvrg, uvrg, uvrg_dy; \ |
| | 657 | add b, b, b_dy \ |
| | 658 | |
| | 659 | |
| | 660 | #define setup_spans_clip_interpolants_increment() \ |
| | 661 | mla b, b_dy, clip, b; \ |
| | 662 | vmla.s32 uvrg, uvrg_dy, v_clip \ |
| | 663 | |
| | 664 | #define setup_spans_clip_interpolants_decrement() \ |
| | 665 | mls b, b_dy, clip, b; \ |
| | 666 | vmls.s32 uvrg, uvrg_dy, v_clip \ |
| | 667 | |
| | 668 | #define setup_spans_clip_alternate_yes() \ |
| | 669 | smlal edge_alt_low, edge_alt_high, edge_dx_dy_alt, clip \ |
| | 670 | |
| | 671 | #define setup_spans_clip_alternate_no() \ |
| | 672 | |
| | 673 | #define setup_spans_clip(direction, alternate_active) \ |
| | 674 | vdup.u32 v_clip, clip; \ |
| | 675 | setup_spans_clip_alternate_##alternate_active(); \ |
| | 676 | setup_spans_clip_interpolants_##direction(); \ |
| | 677 | vmlal.s32 edges_xy, edges_dx_dy, v_clip_low \ |
| | 678 | |
| | 679 | |
| | 680 | #define setup_spans_adjust_edges_alternate_no(left_index, right_index) \ |
| | 681 | vmovl.s32 edge_shifts_64, edge_shifts; \ |
| | 682 | vmovl.s32 edges_dx_dy_64, edges_dx_dy; \ |
| | 683 | \ |
| | 684 | vshl.s64 edges_xy, edges_xy, edge_shifts_64; \ |
| | 685 | vshl.s64 edges_dx_dy_64, edges_dx_dy_64, edge_shifts_64; \ |
| | 686 | \ |
| | 687 | vmov left_x_low, edges_xy_##left_index; \ |
| | 688 | vmov right_x_low, edges_xy_##right_index; \ |
| | 689 | \ |
| | 690 | vmov left_dx_dy_low, edges_dx_dy_64_##left_index; \ |
| | 691 | vmov left_dx_dy_high, edges_dx_dy_64_##left_index; \ |
| | 692 | vmov right_dx_dy_low, edges_dx_dy_64_##right_index; \ |
| | 693 | vmov right_dx_dy_high, edges_dx_dy_64_##right_index; \ |
| | 694 | \ |
| | 695 | vadd.u64 left_x_high, left_x_low, left_dx_dy_low; \ |
| | 696 | vadd.u64 right_x_high, right_x_low, right_dx_dy_low; \ |
| | 697 | \ |
| | 698 | vadd.u64 left_dx_dy, left_dx_dy, left_dx_dy; \ |
| | 699 | vadd.u64 right_dx_dy, right_dx_dy, right_dx_dy \ |
| | 700 | |
| | 701 | |
| | 702 | #define setup_spans_adjust_edges_alternate_yes(left_index, right_index) \ |
| | 703 | setup_spans_adjust_edges_alternate_no(left_index, right_index); \ |
| | 704 | \ |
| | 705 | vdup.u16 y_mid_point, y_b; \ |
| | 706 | rsb temp, edge_shift_alt, #32; \ |
| | 707 | \ |
| | 708 | lsl edge_alt_high, edge_alt_high, edge_shift_alt; \ |
| | 709 | orr edge_alt_high, edge_alt_high, edge_alt_low, lsr temp; \ |
| | 710 | lsl edge_alt_low, edge_alt_low, edge_shift_alt; \ |
| | 711 | vmov alternate_x_low, edge_alt_low, edge_alt_high; \ |
| | 712 | \ |
| | 713 | asr edge_dx_dy_alt_high, edge_dx_dy_alt, temp; \ |
| | 714 | lsl edge_dx_dy_alt_low, edge_dx_dy_alt, edge_shift_alt; \ |
| | 715 | vmov alternate_dx_dy_low, edge_dx_dy_alt_low, edge_dx_dy_alt_high; \ |
| | 716 | vmov alternate_dx_dy_high, alternate_dx_dy_low; \ |
| | 717 | \ |
| | 718 | vadd.u64 alternate_x_high, alternate_x_low, alternate_dx_dy_low; \ |
| | 719 | vadd.u64 alternate_dx_dy, alternate_dx_dy, alternate_dx_dy \ |
| | 720 | |
| | 721 | |
| | 722 | #define setup_spans_y_select_up() \ |
| | 723 | vclt.s16 alternate_select, y_x4, y_mid_point \ |
| | 724 | |
| | 725 | #define setup_spans_y_select_down() \ |
| | 726 | vcgt.s16 alternate_select, y_x4, y_mid_point \ |
| | 727 | |
| | 728 | |
| | 729 | #define setup_spans_alternate_select_left() \ |
| | 730 | vbit.u16 left_right_x_16_low, alternate_x_16, alternate_select \ |
| | 731 | |
| | 732 | #define setup_spans_alternate_select_right() \ |
| | 733 | vbit.u16 left_right_x_16_high, alternate_x_16, alternate_select \ |
| | 734 | |
| | 735 | |
| | 736 | #define setup_spans_set_x4_alternate_yes(alternate, direction) \ |
| | 737 | vshrn.s64 alternate_x_32_low, alternate_x, #32; \ |
| | 738 | vshrn.s64 left_x_32_low, left_x, #32; \ |
| | 739 | vshrn.s64 right_x_32_low, right_x, #32; \ |
| | 740 | \ |
| | 741 | vadd.u64 alternate_x, alternate_x, alternate_dx_dy; \ |
| | 742 | vadd.u64 left_x, left_x, left_dx_dy; \ |
| | 743 | vadd.u64 right_x, right_x, right_dx_dy; \ |
| | 744 | \ |
| | 745 | vshrn.s64 alternate_x_32_high, alternate_x, #32; \ |
| | 746 | vshrn.s64 left_x_32_high, left_x, #32; \ |
| | 747 | vshrn.s64 right_x_32_high, right_x, #32; \ |
| | 748 | \ |
| | 749 | vadd.u64 alternate_x, alternate_x, alternate_dx_dy; \ |
| | 750 | vadd.u64 left_x, left_x, left_dx_dy; \ |
| | 751 | vadd.u64 right_x, right_x, right_dx_dy; \ |
| | 752 | \ |
| | 753 | vmovn.u32 alternate_x_16, alternate_x_32; \ |
| | 754 | setup_spans_y_select_##direction(); \ |
| | 755 | vmovn.u32 left_right_x_16_low, left_x_32; \ |
| | 756 | \ |
| | 757 | vmovn.u32 left_right_x_16_high, right_x_32; \ |
| | 758 | setup_spans_alternate_select_##alternate(); \ |
| | 759 | \ |
| | 760 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 761 | str b, [ span_b_offset ], #4; \ |
| | 762 | setup_spans_adjust_interpolants_##direction(); \ |
| | 763 | \ |
| | 764 | vmax.s16 left_right_x_16, left_right_x_16, left_edge; \ |
| | 765 | \ |
| | 766 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 767 | str b, [ span_b_offset ], #4; \ |
| | 768 | setup_spans_adjust_interpolants_##direction(); \ |
| | 769 | \ |
| | 770 | vmin.s16 left_right_x_16, left_right_x_16, right_edge; \ |
| | 771 | \ |
| | 772 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 773 | str b, [ span_b_offset ], #4; \ |
| | 774 | setup_spans_adjust_interpolants_##direction(); \ |
| | 775 | \ |
| | 776 | vsub.u16 left_right_x_16_high, left_right_x_16_high, left_right_x_16_low; \ |
| | 777 | vadd.u16 left_right_x_16_high, left_right_x_16_high, c_0x0007; \ |
| | 778 | vand.u16 span_shifts, left_right_x_16_high, c_0x0007; \ |
| | 779 | \ |
| | 780 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 781 | str b, [ span_b_offset ], #4; \ |
| | 782 | setup_spans_adjust_interpolants_##direction(); \ |
| | 783 | \ |
| | 784 | vshr.u16 left_right_x_16_high, left_right_x_16_high, #3; \ |
| | 785 | vshl.u16 span_shifts, c_0xFFFE, span_shifts; \ |
| | 786 | \ |
| | 787 | vst4.u16 { left_right_x_16, span_shifts_y }, [ span_edge_data ]!; \ |
| | 788 | \ |
| | 789 | setup_spans_adjust_y_##direction() \ |
| | 790 | |
| | 791 | |
| | 792 | #define setup_spans_set_x4_alternate_no(alternate, direction) \ |
| | 793 | vshrn.s64 left_x_32_low, left_x, #32; \ |
| | 794 | vshrn.s64 right_x_32_low, right_x, #32; \ |
| | 795 | \ |
| | 796 | vadd.u64 left_x, left_x, left_dx_dy; \ |
| | 797 | vadd.u64 right_x, right_x, right_dx_dy; \ |
| | 798 | \ |
| | 799 | vshrn.s64 left_x_32_high, left_x, #32; \ |
| | 800 | vshrn.s64 right_x_32_high, right_x, #32; \ |
| | 801 | \ |
| | 802 | vadd.u64 left_x, left_x, left_dx_dy; \ |
| | 803 | vadd.u64 right_x, right_x, right_dx_dy; \ |
| | 804 | \ |
| | 805 | vmovn.u32 left_right_x_16_low, left_x_32; \ |
| | 806 | vmovn.u32 left_right_x_16_high, right_x_32; \ |
| | 807 | \ |
| | 808 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 809 | str b, [ span_b_offset ], #4; \ |
| | 810 | setup_spans_adjust_interpolants_##direction(); \ |
| | 811 | \ |
| | 812 | vmax.s16 left_right_x_16, left_right_x_16, left_edge; \ |
| | 813 | \ |
| | 814 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 815 | str b, [ span_b_offset ], #4; \ |
| | 816 | setup_spans_adjust_interpolants_##direction(); \ |
| | 817 | \ |
| | 818 | vmin.s16 left_right_x_16, left_right_x_16, right_edge; \ |
| | 819 | \ |
| | 820 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 821 | str b, [ span_b_offset ], #4; \ |
| | 822 | setup_spans_adjust_interpolants_##direction(); \ |
| | 823 | \ |
| | 824 | vsub.u16 left_right_x_16_high, left_right_x_16_high, left_right_x_16_low; \ |
| | 825 | vadd.u16 left_right_x_16_high, left_right_x_16_high, c_0x0007; \ |
| | 826 | vand.u16 span_shifts, left_right_x_16_high, c_0x0007; \ |
| | 827 | \ |
| | 828 | vst1.u32 { uvrg }, [ span_uvrg_offset, :128 ]!; \ |
| | 829 | str b, [ span_b_offset ], #4; \ |
| | 830 | setup_spans_adjust_interpolants_##direction(); \ |
| | 831 | \ |
| | 832 | vshl.u16 span_shifts, c_0xFFFE, span_shifts; \ |
| | 833 | vshr.u16 left_right_x_16_high, left_right_x_16_high, #3; \ |
| | 834 | \ |
| | 835 | vst4.u16 { left_right_x_16, span_shifts_y }, [ span_edge_data ]!; \ |
| | 836 | \ |
| | 837 | setup_spans_adjust_y_##direction() \ |
| | 838 | |
| | 839 | |
| | 840 | #define edge_adjust_low r11 |
| | 841 | #define edge_adjust_high r12 |
| | 842 | |
| | 843 | #define setup_spans_alternate_adjust_yes() \ |
| | 844 | smull edge_adjust_low, edge_adjust_high, edge_dx_dy_alt, height_minor_a; \ |
| | 845 | subs edge_alt_low, edge_alt_low, edge_adjust_low; \ |
| | 846 | sbc edge_alt_high, edge_alt_high, edge_adjust_high \ |
| | 847 | |
| | 848 | #define setup_spans_alternate_adjust_no() \ |
| | 849 | |
| | 850 | |
| | 851 | #define setup_spans_down(left_index, right_index, alternate, alternate_active) \ |
| | 852 | setup_spans_alternate_adjust_##alternate_active(); \ |
| | 853 | setup_spans_load_b(); \ |
| | 854 | \ |
| | 855 | ldrsh temp, [ psx_gpu, #psx_gpu_viewport_end_y_offset ]; \ |
| | 856 | subs y_c, y_c, temp; \ |
| | 857 | subgt height, height, y_c; \ |
| | 858 | addgt height, height, #1; \ |
| | 859 | \ |
| | 860 | ldrsh temp, [ psx_gpu, #psx_gpu_viewport_start_y_offset ]; \ |
| | 861 | subs clip, temp, y_a; \ |
| | 862 | ble 0f; \ |
| | 863 | \ |
| | 864 | sub height, height, clip; \ |
| | 865 | add y_a, y_a, clip; \ |
| | 866 | setup_spans_clip(increment, alternate_active); \ |
| | 867 | \ |
| | 868 | 0: \ |
| | 869 | cmp height, #0; \ |
| | 870 | ble 1f; \ |
| | 871 | \ |
| | 872 | orr temp, y_a, y_a, lsl #16; \ |
| | 873 | add temp, temp, #(1 << 16); \ |
| | 874 | add y_a, temp, #2; \ |
| | 875 | add y_a, y_a, #(2 << 16); \ |
| | 876 | vmov.u32 y_x4, temp, y_a; \ |
| | 877 | \ |
| | 878 | setup_spans_adjust_edges_alternate_##alternate_active(left_index, \ |
| | 879 | right_index); \ |
| | 880 | setup_spans_prologue_b(); \ |
| | 881 | \ |
| | 882 | strh height, [ psx_gpu, #psx_gpu_num_spans_offset ]; \ |
| | 883 | \ |
| | 884 | 2: \ |
| | 885 | setup_spans_set_x4_alternate_##alternate_active(alternate, down); \ |
| | 886 | subs height, height, #4; \ |
| | 887 | bhi 2b; \ |
| | 888 | \ |
| | 889 | 1: \ |
| | 890 | |
| | 891 | |
| | 892 | #define setup_spans_alternate_pre_increment_yes() \ |
| | 893 | adds edge_alt_low, edge_alt_low, edge_dx_dy_alt; \ |
| | 894 | adc edge_alt_high, edge_alt_high, edge_dx_dy_alt, asr #31 \ |
| | 895 | |
| | 896 | #define setup_spans_alternate_pre_increment_no() \ |
| | 897 | |
| | 898 | |
| | 899 | #define setup_spans_up_decrement_yes() \ |
| | 900 | suble height, height, #1 \ |
| | 901 | |
| | 902 | #define setup_spans_up_decrement_no() \ |
| | 903 | |
| | 904 | |
| | 905 | #define setup_spans_up(left_index, right_index, alternate, alternate_active) \ |
| | 906 | setup_spans_alternate_adjust_##alternate_active(); \ |
| | 907 | setup_spans_load_b(); \ |
| | 908 | sub y_a, y_a, #1; \ |
| | 909 | \ |
| | 910 | ldrh temp, [ psx_gpu, #psx_gpu_viewport_start_y_offset ]; \ |
| | 911 | subs temp, temp, y_c; \ |
| | 912 | subgt height, height, temp; \ |
| | 913 | setup_spans_up_decrement_##alternate_active(); \ |
| | 914 | \ |
| | 915 | ldrh temp, [ psx_gpu, #psx_gpu_viewport_end_y_offset ]; \ |
| | 916 | subs clip, y_a, temp; \ |
| | 917 | ble 0f; \ |
| | 918 | \ |
| | 919 | sub height, height, clip; \ |
| | 920 | sub y_a, y_a, clip; \ |
| | 921 | setup_spans_clip(decrement, alternate_active); \ |
| | 922 | \ |
| | 923 | 0: \ |
| | 924 | cmp height, #0; \ |
| | 925 | ble 1f; \ |
| | 926 | \ |
| | 927 | orr temp, y_a, y_a, lsl #16; \ |
| | 928 | sub temp, temp, #(1 << 16); \ |
| | 929 | sub y_a, temp, #2; \ |
| | 930 | sub y_a, y_a, #(2 << 16); \ |
| | 931 | vmov.u32 y_x4, temp, y_a; \ |
| | 932 | \ |
| | 933 | vaddw.s32 edges_xy, edges_xy, edges_dx_dy; \ |
| | 934 | \ |
| | 935 | setup_spans_alternate_pre_increment_##alternate_active(); \ |
| | 936 | setup_spans_adjust_edges_alternate_##alternate_active(left_index, \ |
| | 937 | right_index); \ |
| | 938 | setup_spans_adjust_interpolants_up(); \ |
| | 939 | setup_spans_prologue_b(); \ |
| | 940 | \ |
| | 941 | strh height, [ psx_gpu, #psx_gpu_num_spans_offset ]; \ |
| | 942 | \ |
| | 943 | 2: \ |
| | 944 | setup_spans_set_x4_alternate_##alternate_active(alternate, up); \ |
| | 945 | subs height, height, #4; \ |
| | 946 | bhi 2b; \ |
| | 947 | \ |
| | 948 | 1: \ |
| | 949 | |
| | 950 | |
| | 951 | #define setup_spans_epilogue() \ |
| | 952 | ldmia sp!, { r4 - r11, pc } \ |
| | 953 | |
| | 954 | |
| | 955 | #define setup_spans_up_up(minor, major) \ |
| | 956 | setup_spans_prologue(); \ |
| | 957 | sub height_minor_a, y_a, y_b; \ |
| | 958 | sub height_minor_b, y_b, y_c; \ |
| | 959 | sub height, y_a, y_c; \ |
| | 960 | \ |
| | 961 | vdup.u32 x_starts, x_a; \ |
| | 962 | vmov.u32 x_ends, x_c, x_b; \ |
| | 963 | \ |
| | 964 | compute_edge_delta_x3(x_b, height_major, height_minor_a); \ |
| | 965 | setup_spans_up(major, minor, minor, yes); \ |
| | 966 | setup_spans_epilogue() \ |
| | 967 | |
| | 968 | function(setup_spans_up_left) |
| | 969 | setup_spans_up_up(left, right) |
| | 970 | |
| | 971 | function(setup_spans_up_right) |
| | 972 | setup_spans_up_up(right, left) |
| | 973 | |
| | 974 | .pool |
| | 975 | |
| | 976 | #define setup_spans_down_down(minor, major) \ |
| | 977 | setup_spans_prologue(); \ |
| | 978 | sub height_minor_a, y_b, y_a; \ |
| | 979 | sub height_minor_b, y_c, y_b; \ |
| | 980 | sub height, y_c, y_a; \ |
| | 981 | \ |
| | 982 | vdup.u32 x_starts, x_a; \ |
| | 983 | vmov.u32 x_ends, x_c, x_b; \ |
| | 984 | \ |
| | 985 | compute_edge_delta_x3(x_b, height_major, height_minor_a); \ |
| | 986 | setup_spans_down(major, minor, minor, yes); \ |
| | 987 | setup_spans_epilogue() \ |
| | 988 | |
| | 989 | function(setup_spans_down_left) |
| | 990 | setup_spans_down_down(left, right) |
| | 991 | |
| | 992 | function(setup_spans_down_right) |
| | 993 | setup_spans_down_down(right, left) |
| | 994 | |
| | 995 | |
| | 996 | #define setup_spans_up_flat() \ |
| | 997 | sub height, y_a, y_c; \ |
| | 998 | \ |
| | 999 | compute_edge_delta_x2(); \ |
| | 1000 | setup_spans_up(left, right, none, no); \ |
| | 1001 | setup_spans_epilogue() \ |
| | 1002 | |
| | 1003 | function(setup_spans_up_a) |
| | 1004 | setup_spans_prologue() |
| | 1005 | |
| | 1006 | vmov.u32 x_starts, x_a, x_b |
| | 1007 | vdup.u32 x_ends, x_c |
| | 1008 | |
| | 1009 | setup_spans_up_flat() |
| | 1010 | |
| | 1011 | function(setup_spans_up_b) |
| | 1012 | setup_spans_prologue() |
| | 1013 | |
| | 1014 | vdup.u32 x_starts, x_a |
| | 1015 | vmov.u32 x_ends, x_b, x_c |
| | 1016 | |
| | 1017 | setup_spans_up_flat() |
| | 1018 | |
| | 1019 | #define setup_spans_down_flat() \ |
| | 1020 | sub height, y_c, y_a; \ |
| | 1021 | \ |
| | 1022 | compute_edge_delta_x2(); \ |
| | 1023 | setup_spans_down(left, right, none, no); \ |
| | 1024 | setup_spans_epilogue() \ |
| | 1025 | |
| | 1026 | function(setup_spans_down_a) |
| | 1027 | setup_spans_prologue() |
| | 1028 | |
| | 1029 | vmov.u32 x_starts, x_a, x_b |
| | 1030 | vdup.u32 x_ends, x_c |
| | 1031 | |
| | 1032 | setup_spans_down_flat() |
| | 1033 | |
| | 1034 | function(setup_spans_down_b) |
| | 1035 | setup_spans_prologue() |
| | 1036 | |
| | 1037 | vdup.u32 x_starts, x_a |
| | 1038 | vmov.u32 x_ends, x_b, x_c |
| | 1039 | |
| | 1040 | setup_spans_down_flat() |
| | 1041 | |
| | 1042 | |
| | 1043 | #define middle_y r9 |
| | 1044 | |
| | 1045 | #define edges_xy_b q11 |
| | 1046 | #define edges_dx_dy_b d26 |
| | 1047 | #define edge_shifts_b d27 |
| | 1048 | #define edges_dx_dy_and_shifts_b q13 |
| | 1049 | #define height_increment d20 |
| | 1050 | |
| | 1051 | #define edges_dx_dy_and_shifts q1 |
| | 1052 | |
| | 1053 | #define edges_xy_b_left d22 |
| | 1054 | #define edges_xy_b_right d23 |
| | 1055 | |
| | 1056 | #define setup_spans_up_down_load_edge_set_b() \ |
| | 1057 | vmov edges_xy, edges_xy_b; \ |
| | 1058 | vmov edges_dx_dy_and_shifts, edges_dx_dy_and_shifts_b \ |
| | 1059 | |
| | 1060 | |
| | 1061 | function(setup_spans_up_down) |
| | 1062 | setup_spans_prologue() |
| | 1063 | |
| | 1064 | // s32 middle_y = y_a; |
| | 1065 | sub height_minor_a, y_a, y_b |
| | 1066 | sub height_minor_b, y_c, y_a |
| | 1067 | sub height_major, y_c, y_b |
| | 1068 | |
| | 1069 | vmov.u32 x_starts, x_a, x_c |
| | 1070 | vdup.u32 x_ends, x_b |
| | 1071 | |
| | 1072 | compute_edge_delta_x3(x_a, height_minor_a, height_major) |
| | 1073 | |
| | 1074 | mov temp, #0 |
| | 1075 | vmov.u32 height_increment, temp, height_minor_b |
| | 1076 | vmlal.s32 edges_xy, edges_dx_dy, height_increment |
| | 1077 | |
| | 1078 | vmov edges_xy_b_left, edge_alt_low, edge_alt_high |
| | 1079 | vmov edges_xy_b_right, edges_xy_right |
| | 1080 | |
| | 1081 | vmov edge_shifts_b, edge_shifts |
| | 1082 | vmov.u32 edge_shifts_b[0], edge_shift_alt |
| | 1083 | |
| | 1084 | vneg.s32 edges_dx_dy_b, edges_dx_dy |
| | 1085 | vmov.u32 edges_dx_dy_b[0], edge_dx_dy_alt |
| | 1086 | |
| | 1087 | mov middle_y, y_a |
| | 1088 | |
| | 1089 | setup_spans_load_b() |
| | 1090 | sub y_a, y_a, #1 |
| | 1091 | |
| | 1092 | ldrh temp, [ psx_gpu, #psx_gpu_viewport_start_y_offset ] |
| | 1093 | subs temp, temp, y_b |
| | 1094 | subgt height_minor_a, height_minor_a, temp |
| | 1095 | |
| | 1096 | ldrh temp, [ psx_gpu, #psx_gpu_viewport_end_y_offset ] |
| | 1097 | subs clip, y_a, temp |
| | 1098 | ble 0f |
| | 1099 | |
| | 1100 | sub height_minor_a, height_minor_a, clip |
| | 1101 | sub y_a, y_a, clip |
| | 1102 | setup_spans_clip(decrement, no) |
| | 1103 | |
| | 1104 | 0: |
| | 1105 | cmp height_minor_a, #0 |
| | 1106 | ble 3f |
| | 1107 | |
| | 1108 | orr temp, y_a, y_a, lsl #16 |
| | 1109 | sub temp, temp, #(1 << 16) |
| | 1110 | sub y_a, temp, #2 |
| | 1111 | sub y_a, y_a, #(2 << 16) |
| | 1112 | vmov.u32 y_x4, temp, y_a |
| | 1113 | |
| | 1114 | vaddw.s32 edges_xy, edges_xy, edges_dx_dy |
| | 1115 | |
| | 1116 | strh height_minor_a, [ psx_gpu, #psx_gpu_num_spans_offset ] |
| | 1117 | |
| | 1118 | setup_spans_adjust_edges_alternate_no(left, right); |
| | 1119 | setup_spans_adjust_interpolants_up() |
| | 1120 | setup_spans_up_down_load_edge_set_b() |
| | 1121 | |
| | 1122 | setup_spans_prologue_b() |
| | 1123 | |
| | 1124 | |
| | 1125 | 2: |
| | 1126 | setup_spans_set_x4_alternate_no(none, up) |
| | 1127 | subs height_minor_a, height_minor_a, #4 |
| | 1128 | bhi 2b |
| | 1129 | |
| | 1130 | add span_edge_data, span_edge_data, height_minor_a, lsl #3 |
| | 1131 | add span_uvrg_offset, span_uvrg_offset, height_minor_a, lsl #4 |
| | 1132 | add span_b_offset, span_b_offset, height_minor_a, lsl #2 |
| | 1133 | |
| | 1134 | 4: |
| | 1135 | add temp, psx_gpu, #psx_gpu_uvrg_offset |
| | 1136 | vld1.32 { uvrg }, [ temp ] |
| | 1137 | mov y_a, middle_y |
| | 1138 | |
| | 1139 | setup_spans_load_b() |
| | 1140 | |
| | 1141 | ldrh temp, [ psx_gpu, #psx_gpu_viewport_end_y_offset ] |
| | 1142 | subs y_c, y_c, temp |
| | 1143 | subgt height_minor_b, height_minor_b, y_c |
| | 1144 | addgt height_minor_b, height_minor_b, #1 |
| | 1145 | |
| | 1146 | ldrh temp, [ psx_gpu, #psx_gpu_viewport_start_y_offset ] |
| | 1147 | subs clip, temp, y_a |
| | 1148 | ble 0f |
| | 1149 | |
| | 1150 | sub height_minor_b, height_minor_b, clip |
| | 1151 | add y_a, y_a, clip |
| | 1152 | setup_spans_clip(increment, no) |
| | 1153 | |
| | 1154 | 0: |
| | 1155 | cmp height_minor_b, #0 |
| | 1156 | ble 1f |
| | 1157 | |
| | 1158 | orr temp, y_a, y_a, lsl #16 |
| | 1159 | add temp, temp, #(1 << 16) |
| | 1160 | add y_a, temp, #2 |
| | 1161 | add y_a, y_a, #(2 << 16) |
| | 1162 | vmov.u32 y_x4, temp, y_a |
| | 1163 | |
| | 1164 | setup_spans_adjust_edges_alternate_no(left, right) |
| | 1165 | |
| | 1166 | ldrh temp, [ psx_gpu, #psx_gpu_num_spans_offset ] |
| | 1167 | add temp, temp, height_minor_b |
| | 1168 | strh temp, [ psx_gpu, #psx_gpu_num_spans_offset ] |
| | 1169 | |
| | 1170 | 2: |
| | 1171 | setup_spans_set_x4_alternate_no(none, down) |
| | 1172 | subs height_minor_b, height_minor_b, #4 |
| | 1173 | bhi 2b |
| | 1174 | |
| | 1175 | 1: |
| | 1176 | setup_spans_epilogue() |
| | 1177 | |
| | 1178 | 3: |
| | 1179 | setup_spans_up_down_load_edge_set_b() |
| | 1180 | setup_spans_prologue_b() |
| | 1181 | bal 4b |
| | 1182 | |
| | 1183 | .pool |
| | 1184 | |
| | 1185 | #undef span_uvrg_offset |
| | 1186 | #undef span_edge_data |
| | 1187 | #undef span_b_offset |
| | 1188 | #undef left_x |
| | 1189 | #undef b |
| | 1190 | |
| | 1191 | #define psx_gpu r0 |
| | 1192 | #define num_spans r1 |
| | 1193 | #define span_uvrg_offset r2 |
| | 1194 | #define span_edge_data r3 |
| | 1195 | #define span_b_offset r4 |
| | 1196 | #define b_dx r5 |
| | 1197 | #define span_num_blocks r6 |
| | 1198 | #define y r7 |
| | 1199 | #define left_x r8 |
| | 1200 | #define b r9 |
| | 1201 | #define dither_offset_ptr r10 |
| | 1202 | #define block_ptr_a r11 |
| | 1203 | #define fb_ptr r12 |
| | 1204 | #define num_blocks r14 |
| | 1205 | |
| | 1206 | #define uvrg_dx_ptr r2 |
| | 1207 | #define texture_mask_ptr r3 |
| | 1208 | #define dither_shift r8 |
| | 1209 | #define dither_row r10 |
| | 1210 | |
| | 1211 | #define c_32 r7 |
| | 1212 | #define b_dx4 r8 |
| | 1213 | #define b_dx8 r9 |
| | 1214 | #define block_ptr_b r10 |
| | 1215 | |
| | 1216 | #define block_span_ptr r10 |
| | 1217 | #define right_mask r8 |
| | 1218 | |
| | 1219 | #define color r2 |
| | 1220 | #define color_r r3 |
| | 1221 | #define color_g r4 |
| | 1222 | #define color_b r5 |
| | 1223 | |
| | 1224 | #undef uvrg |
| | 1225 | |
| | 1226 | #define u_block q0 |
| | 1227 | #define v_block q1 |
| | 1228 | #define r_block q2 |
| | 1229 | #define g_block q3 |
| | 1230 | #define b_block q4 |
| | 1231 | |
| | 1232 | #define uv_dx4 d10 |
| | 1233 | #define rg_dx4 d11 |
| | 1234 | #define uv_dx8 d12 |
| | 1235 | #define rg_dx8 d13 |
| | 1236 | #define b_whole_8 d14 |
| | 1237 | #define fb_mask_ptrs d15 |
| | 1238 | |
| | 1239 | #define uvrg_dx4 q5 |
| | 1240 | #define uvrg_dx8 q6 |
| | 1241 | #define uv_dx8 d12 |
| | 1242 | #define rg_dx8 d13 |
| | 1243 | |
| | 1244 | #define u_whole q8 |
| | 1245 | #define v_whole q9 |
| | 1246 | #define r_whole q10 |
| | 1247 | #define g_whole q11 |
| | 1248 | #define b_whole q12 |
| | 1249 | |
| | 1250 | #define u_whole_low d16 |
| | 1251 | #define u_whole_high d17 |
| | 1252 | #define v_whole_low d18 |
| | 1253 | #define v_whole_high d19 |
| | 1254 | #define r_whole_low d20 |
| | 1255 | #define r_whole_high d21 |
| | 1256 | #define g_whole_low d22 |
| | 1257 | #define g_whole_high d23 |
| | 1258 | #define b_whole_low d24 |
| | 1259 | #define b_whole_high d25 |
| | 1260 | |
| | 1261 | #define dx4 q13 |
| | 1262 | #define dx8 q13 |
| | 1263 | |
| | 1264 | #define u_whole_8 d26 |
| | 1265 | #define v_whole_8 d27 |
| | 1266 | #define u_whole_8b d24 |
| | 1267 | #define r_whole_8 d24 |
| | 1268 | #define g_whole_8 d25 |
| | 1269 | |
| | 1270 | #define uv_whole_8 q13 |
| | 1271 | #define uv_whole_8b q14 |
| | 1272 | |
| | 1273 | #define dither_offsets q14 |
| | 1274 | #define texture_mask q15 |
| | 1275 | #define texture_mask_u d30 |
| | 1276 | #define texture_mask_v d31 |
| | 1277 | |
| | 1278 | #define dither_offsets_short d28 |
| | 1279 | |
| | 1280 | #define v_left_x q8 |
| | 1281 | #define uvrg q9 |
| | 1282 | #define block_span q10 |
| | 1283 | |
| | 1284 | #define uv d18 |
| | 1285 | #define rg d19 |
| | 1286 | |
| | 1287 | #define draw_mask q1 |
| | 1288 | #define draw_mask_edge q13 |
| | 1289 | #define test_mask q0 |
| | 1290 | |
| | 1291 | #define uvrg_dx q3 |
| | 1292 | |
| | 1293 | #define colors q2 |
| | 1294 | |
| | 1295 | #define setup_blocks_texture_swizzled() \ |
| | 1296 | vand.u8 u_whole_8b, u_whole_8, texture_mask_u; \ |
| | 1297 | vsli.u8 u_whole_8, v_whole_8, #4; \ |
| | 1298 | vsri.u8 v_whole_8, u_whole_8b, #4 \ |
| | 1299 | |
| | 1300 | #define setup_blocks_texture_unswizzled() \ |
| | 1301 | |
| | 1302 | |
| | 1303 | #define setup_blocks_shaded_textured_builder(swizzling) \ |
| | 1304 | .align 3; \ |
| | 1305 | \ |
| | 1306 | function(setup_blocks_shaded_textured_dithered_##swizzling##_indirect) \ |
| | 1307 | ldrh num_spans, [ psx_gpu, #psx_gpu_num_spans_offset ]; \ |
| | 1308 | add uvrg_dx_ptr, psx_gpu, #psx_gpu_uvrg_dx_offset; \ |
| | 1309 | \ |
| | 1310 | vld1.u32 { uvrg_dx }, [ uvrg_dx_ptr, :128 ]; \ |
| | 1311 | add texture_mask_ptr, psx_gpu, #psx_gpu_texture_mask_width_offset; \ |
| | 1312 | \ |
| | 1313 | cmp num_spans, #0; \ |
| | 1314 | bxeq lr; \ |
| | 1315 | \ |
| | 1316 | stmdb sp!, { r4 - r11, r14 }; \ |
| | 1317 | vshl.u32 uvrg_dx4, uvrg_dx, #2; \ |
| | 1318 | \ |
| | 1319 | ldr b_dx, [ psx_gpu, #psx_gpu_b_dx_offset ]; \ |
| | 1320 | vshl.u32 uvrg_dx8, uvrg_dx, #3; \ |
| | 1321 | \ |
| | 1322 | vld2.u8 { texture_mask_u[], texture_mask_v[] }, [ texture_mask_ptr, :16 ]; \ |
| | 1323 | add span_uvrg_offset, psx_gpu, #psx_gpu_span_uvrg_offset_offset; \ |
| | 1324 | \ |
| | 1325 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 1326 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset; \ |
| | 1327 | \ |
| | 1328 | add span_b_offset, psx_gpu, #psx_gpu_span_b_offset_offset; \ |
| | 1329 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 1330 | \ |
| | 1331 | add block_ptr_a, block_ptr_a, num_blocks, lsl #6; \ |
| | 1332 | \ |
| | 1333 | 0: \ |
| | 1334 | vmov.u8 fb_mask_ptrs, #0; \ |
| | 1335 | \ |
| | 1336 | ldrh span_num_blocks, [ span_edge_data, #edge_data_num_blocks_offset ]; \ |
| | 1337 | add dither_offset_ptr, psx_gpu, #psx_gpu_dither_table_offset; \ |
| | 1338 | \ |
| | 1339 | ldrh y, [ span_edge_data, #edge_data_y_offset ]; \ |
| | 1340 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ]; \ |
| | 1341 | \ |
| | 1342 | cmp span_num_blocks, #0; \ |
| | 1343 | beq 1f; \ |
| | 1344 | \ |
| | 1345 | ldrh left_x, [ span_edge_data, #edge_data_left_x_offset ]; \ |
| | 1346 | add num_blocks, span_num_blocks, num_blocks; \ |
| | 1347 | \ |
| | 1348 | cmp num_blocks, #MAX_BLOCKS; \ |
| | 1349 | bgt 2f; \ |
| | 1350 | \ |
| | 1351 | 3: \ |
| | 1352 | ldr b, [ span_b_offset ]; \ |
| | 1353 | add fb_ptr, fb_ptr, y, lsl #11; \ |
| | 1354 | \ |
| | 1355 | vdup.u32 v_left_x, left_x; \ |
| | 1356 | and y, y, #0x3; \ |
| | 1357 | \ |
| | 1358 | ldr dither_row, [ dither_offset_ptr, y, lsl #2 ]; \ |
| | 1359 | add fb_ptr, fb_ptr, left_x, lsl #1; \ |
| | 1360 | \ |
| | 1361 | mla b, b_dx, left_x, b; \ |
| | 1362 | and dither_shift, left_x, #0x03; \ |
| | 1363 | \ |
| | 1364 | vld1.u32 { uvrg }, [ span_uvrg_offset, :128 ]; \ |
| | 1365 | vshr.u32 uvrg_dx, uvrg_dx4, #2; \ |
| | 1366 | \ |
| | 1367 | mov dither_shift, dither_shift, lsl #3; \ |
| | 1368 | vmla.u32 uvrg, uvrg_dx, v_left_x; \ |
| | 1369 | \ |
| | 1370 | mov c_32, #32; \ |
| | 1371 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 1372 | \ |
| | 1373 | mov dither_row, dither_row, ror dither_shift; \ |
| | 1374 | mov b_dx4, b_dx, lsl #2; \ |
| | 1375 | \ |
| | 1376 | vdup.u32 dither_offsets_short, dither_row; \ |
| | 1377 | add block_span_ptr, psx_gpu, #psx_gpu_u_block_span_offset; \ |
| | 1378 | \ |
| | 1379 | vdup.u32 b_block, b; \ |
| | 1380 | vshll.s8 dither_offsets, dither_offsets_short, #4; \ |
| | 1381 | \ |
| | 1382 | vdup.u32 u_block, uv[0]; \ |
| | 1383 | mov b_dx8, b_dx, lsl #3; \ |
| | 1384 | \ |
| | 1385 | vdup.u32 v_block, uv[1]; \ |
| | 1386 | vdup.u32 r_block, rg[0]; \ |
| | 1387 | vdup.u32 g_block, rg[1]; \ |
| | 1388 | \ |
| | 1389 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 1390 | \ |
| | 1391 | vadd.u32 u_block, u_block, block_span; \ |
| | 1392 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 1393 | \ |
| | 1394 | vadd.u32 v_block, v_block, block_span; \ |
| | 1395 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 1396 | \ |
| | 1397 | vadd.u32 r_block, r_block, block_span; \ |
| | 1398 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 1399 | \ |
| | 1400 | vadd.u32 g_block, g_block, block_span; \ |
| | 1401 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]; \ |
| | 1402 | \ |
| | 1403 | vadd.u32 b_block, b_block, block_span; \ |
| | 1404 | add block_ptr_b, block_ptr_a, #16; \ |
| | 1405 | \ |
| | 1406 | vshrn.u32 u_whole_low, u_block, #16; \ |
| | 1407 | vshrn.u32 v_whole_low, v_block, #16; \ |
| | 1408 | vshrn.u32 r_whole_low, r_block, #16; \ |
| | 1409 | vshrn.u32 g_whole_low, g_block, #16; \ |
| | 1410 | \ |
| | 1411 | vdup.u32 dx4, uv_dx4[0]; \ |
| | 1412 | vshrn.u32 b_whole_low, b_block, #16; \ |
| | 1413 | \ |
| | 1414 | vaddhn.u32 u_whole_high, u_block, dx4; \ |
| | 1415 | vdup.u32 dx4, uv_dx4[1]; \ |
| | 1416 | \ |
| | 1417 | vaddhn.u32 v_whole_high, v_block, dx4; \ |
| | 1418 | vdup.u32 dx4, rg_dx4[0]; \ |
| | 1419 | \ |
| | 1420 | vaddhn.u32 r_whole_high, r_block, dx4; \ |
| | 1421 | vdup.u32 dx4, rg_dx4[1]; \ |
| | 1422 | \ |
| | 1423 | vaddhn.u32 g_whole_high, g_block, dx4; \ |
| | 1424 | vdup.u32 dx4, b_dx4; \ |
| | 1425 | \ |
| | 1426 | vaddhn.u32 b_whole_high, b_block, dx4; \ |
| | 1427 | vdup.u32 dx8, uv_dx8[0]; \ |
| | 1428 | \ |
| | 1429 | vadd.u32 u_block, u_block, dx8; \ |
| | 1430 | vdup.u32 dx8, uv_dx8[1]; \ |
| | 1431 | \ |
| | 1432 | vadd.u32 v_block, v_block, dx8; \ |
| | 1433 | vdup.u32 dx8, rg_dx8[0]; \ |
| | 1434 | \ |
| | 1435 | vadd.u32 r_block, r_block, dx8; \ |
| | 1436 | vdup.u32 dx8, rg_dx8[1]; \ |
| | 1437 | \ |
| | 1438 | vadd.u32 g_block, g_block, dx8; \ |
| | 1439 | vdup.u32 dx8, b_dx8; \ |
| | 1440 | \ |
| | 1441 | vadd.u32 b_block, b_block, dx8; \ |
| | 1442 | vmovn.u16 u_whole_8, u_whole; \ |
| | 1443 | \ |
| | 1444 | vmovn.u16 v_whole_8, v_whole; \ |
| | 1445 | \ |
| | 1446 | vmovn.u16 b_whole_8, b_whole; \ |
| | 1447 | pld [ fb_ptr ]; \ |
| | 1448 | vmov.u32 fb_mask_ptrs[1], fb_ptr; \ |
| | 1449 | \ |
| | 1450 | vand.u8 uv_whole_8, uv_whole_8, texture_mask; \ |
| | 1451 | setup_blocks_texture_##swizzling(); \ |
| | 1452 | \ |
| | 1453 | vmovn.u16 r_whole_8, r_whole; \ |
| | 1454 | beq 5f; \ |
| | 1455 | \ |
| | 1456 | 4: \ |
| | 1457 | vmovn.u16 g_whole_8, g_whole; \ |
| | 1458 | vshrn.u32 u_whole_low, u_block, #16; \ |
| | 1459 | \ |
| | 1460 | vst2.u8 { u_whole_8, v_whole_8 }, [ block_ptr_a, :128 ], c_32; \ |
| | 1461 | vshrn.u32 v_whole_low, v_block, #16; \ |
| | 1462 | \ |
| | 1463 | vst1.u32 { r_whole_8, g_whole_8 }, [ block_ptr_b, :128 ], c_32; \ |
| | 1464 | vshrn.u32 r_whole_low, r_block, #16; \ |
| | 1465 | \ |
| | 1466 | vst1.u32 { b_whole_8, fb_mask_ptrs }, [ block_ptr_a, :128 ], c_32; \ |
| | 1467 | vshrn.u32 g_whole_low, g_block, #16; \ |
| | 1468 | \ |
| | 1469 | vdup.u32 dx4, uv_dx4[0]; \ |
| | 1470 | vshrn.u32 b_whole_low, b_block, #16; \ |
| | 1471 | \ |
| | 1472 | vaddhn.u32 u_whole_high, u_block, dx4; \ |
| | 1473 | vdup.u32 dx4, uv_dx4[1]; \ |
| | 1474 | \ |
| | 1475 | vaddhn.u32 v_whole_high, v_block, dx4; \ |
| | 1476 | vdup.u32 dx4, rg_dx4[0]; \ |
| | 1477 | \ |
| | 1478 | vaddhn.u32 r_whole_high, r_block, dx4; \ |
| | 1479 | vdup.u32 dx4, rg_dx4[1]; \ |
| | 1480 | \ |
| | 1481 | vaddhn.u32 g_whole_high, g_block, dx4; \ |
| | 1482 | vdup.u32 dx4, b_dx4; \ |
| | 1483 | \ |
| | 1484 | vaddhn.u32 b_whole_high, b_block, dx4; \ |
| | 1485 | vdup.u32 dx8, uv_dx8[0]; \ |
| | 1486 | \ |
| | 1487 | vadd.u32 u_block, u_block, dx8; \ |
| | 1488 | vdup.u32 dx8, uv_dx8[1]; \ |
| | 1489 | \ |
| | 1490 | vadd.u32 v_block, v_block, dx8; \ |
| | 1491 | vdup.u32 dx8, rg_dx8[0]; \ |
| | 1492 | \ |
| | 1493 | vadd.u32 r_block, r_block, dx8; \ |
| | 1494 | vdup.u32 dx8, rg_dx8[1]; \ |
| | 1495 | \ |
| | 1496 | vadd.u32 g_block, g_block, dx8; \ |
| | 1497 | vdup.u32 dx8, b_dx8; \ |
| | 1498 | \ |
| | 1499 | vadd.u32 b_block, b_block, dx8; \ |
| | 1500 | vmovn.u16 u_whole_8, u_whole; \ |
| | 1501 | \ |
| | 1502 | add fb_ptr, fb_ptr, #16; \ |
| | 1503 | vmovn.u16 v_whole_8, v_whole; \ |
| | 1504 | \ |
| | 1505 | vst1.u32 { dither_offsets }, [ block_ptr_b, :128 ], c_32; \ |
| | 1506 | vmovn.u16 b_whole_8, b_whole; \ |
| | 1507 | \ |
| | 1508 | pld [ fb_ptr ]; \ |
| | 1509 | \ |
| | 1510 | vmov.u32 fb_mask_ptrs[1], fb_ptr; \ |
| | 1511 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 1512 | \ |
| | 1513 | vand.u8 uv_whole_8, uv_whole_8, texture_mask; \ |
| | 1514 | setup_blocks_texture_##swizzling(); \ |
| | 1515 | \ |
| | 1516 | vmovn.u16 r_whole_8, r_whole; \ |
| | 1517 | bne 4b; \ |
| | 1518 | \ |
| | 1519 | 5: \ |
| | 1520 | vmovn.u16 g_whole_8, g_whole; \ |
| | 1521 | ldrh right_mask, [ span_edge_data, #edge_data_right_mask_offset ]; \ |
| | 1522 | \ |
| | 1523 | vld1.u32 { test_mask }, [ psx_gpu, :128 ]; \ |
| | 1524 | vdup.u8 draw_mask, right_mask; \ |
| | 1525 | \ |
| | 1526 | vmov.u32 fb_mask_ptrs[0], right_mask; \ |
| | 1527 | vtst.u16 draw_mask, draw_mask, test_mask; \ |
| | 1528 | vzip.u8 u_whole_8, v_whole_8; \ |
| | 1529 | \ |
| | 1530 | vbic.u16 uv_whole_8, uv_whole_8, draw_mask; \ |
| | 1531 | vst1.u32 { r_whole_8, g_whole_8 }, [ block_ptr_b, :128 ], c_32; \ |
| | 1532 | vst1.u32 { uv_whole_8 }, [ block_ptr_a, :128 ], c_32; \ |
| | 1533 | vst1.u32 { dither_offsets }, [ block_ptr_b, :128 ], c_32; \ |
| | 1534 | vst1.u32 { b_whole_8, fb_mask_ptrs }, [ block_ptr_a, :128 ], c_32; \ |
| | 1535 | \ |
| | 1536 | 1: \ |
| | 1537 | add span_uvrg_offset, span_uvrg_offset, #16; \ |
| | 1538 | add span_b_offset, span_b_offset, #4; \ |
| | 1539 | \ |
| | 1540 | add span_edge_data, span_edge_data, #8; \ |
| | 1541 | subs num_spans, num_spans, #1; \ |
| | 1542 | \ |
| | 1543 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 1544 | bne 0b; \ |
| | 1545 | \ |
| | 1546 | ldmia sp!, { r4 - r11, pc }; \ |
| | 1547 | \ |
| | 1548 | 2: \ |
| | 1549 | /* TODO: Load from psx_gpu instead of saving/restoring these */\ |
| | 1550 | vpush { texture_mask }; \ |
| | 1551 | vpush { uvrg_dx4 }; \ |
| | 1552 | \ |
| | 1553 | stmdb sp!, { r0 - r3, r12, r14 }; \ |
| | 1554 | bl flush_render_block_buffer; \ |
| | 1555 | ldmia sp!, { r0 - r3, r12, r14 }; \ |
| | 1556 | \ |
| | 1557 | vpop { uvrg_dx4 }; \ |
| | 1558 | vpop { texture_mask }; \ |
| | 1559 | \ |
| | 1560 | vadd.u32 uvrg_dx8, uvrg_dx4, uvrg_dx4; \ |
| | 1561 | vmov.u8 fb_mask_ptrs, #0; \ |
| | 1562 | \ |
| | 1563 | mov num_blocks, span_num_blocks; \ |
| | 1564 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 1565 | bal 3b \ |
| | 1566 | |
| | 1567 | |
| | 1568 | setup_blocks_shaded_textured_builder(swizzled) |
| | 1569 | setup_blocks_shaded_textured_builder(unswizzled) |
| | 1570 | |
| | 1571 | |
| | 1572 | #define setup_blocks_unshaded_textured_builder(swizzling) \ |
| | 1573 | .align 3; \ |
| | 1574 | \ |
| | 1575 | function(setup_blocks_unshaded_textured_dithered_##swizzling##_indirect) \ |
| | 1576 | ldrh num_spans, [ psx_gpu, #psx_gpu_num_spans_offset ]; \ |
| | 1577 | add uvrg_dx_ptr, psx_gpu, #psx_gpu_uvrg_dx_offset; \ |
| | 1578 | \ |
| | 1579 | vld1.u32 { uvrg_dx }, [ uvrg_dx_ptr, :128 ]; \ |
| | 1580 | add texture_mask_ptr, psx_gpu, #psx_gpu_texture_mask_width_offset; \ |
| | 1581 | \ |
| | 1582 | cmp num_spans, #0; \ |
| | 1583 | bxeq lr; \ |
| | 1584 | \ |
| | 1585 | stmdb sp!, { r4 - r11, r14 }; \ |
| | 1586 | vshl.u32 uvrg_dx4, uvrg_dx, #2; \ |
| | 1587 | \ |
| | 1588 | vshl.u32 uvrg_dx8, uvrg_dx, #3; \ |
| | 1589 | \ |
| | 1590 | vld2.u8 { texture_mask_u[], texture_mask_v[] }, [ texture_mask_ptr, :16 ]; \ |
| | 1591 | add span_uvrg_offset, psx_gpu, #psx_gpu_span_uvrg_offset_offset; \ |
| | 1592 | \ |
| | 1593 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 1594 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset; \ |
| | 1595 | \ |
| | 1596 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 1597 | \ |
| | 1598 | add block_ptr_a, block_ptr_a, num_blocks, lsl #6; \ |
| | 1599 | \ |
| | 1600 | 0: \ |
| | 1601 | vmov.u8 fb_mask_ptrs, #0; \ |
| | 1602 | \ |
| | 1603 | ldrh span_num_blocks, [ span_edge_data, #edge_data_num_blocks_offset ]; \ |
| | 1604 | add dither_offset_ptr, psx_gpu, #psx_gpu_dither_table_offset; \ |
| | 1605 | \ |
| | 1606 | ldrh y, [ span_edge_data, #edge_data_y_offset ]; \ |
| | 1607 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ]; \ |
| | 1608 | \ |
| | 1609 | cmp span_num_blocks, #0; \ |
| | 1610 | beq 1f; \ |
| | 1611 | \ |
| | 1612 | ldrh left_x, [ span_edge_data, #edge_data_left_x_offset ]; \ |
| | 1613 | add num_blocks, span_num_blocks, num_blocks; \ |
| | 1614 | \ |
| | 1615 | cmp num_blocks, #MAX_BLOCKS; \ |
| | 1616 | bgt 2f; \ |
| | 1617 | \ |
| | 1618 | 3: \ |
| | 1619 | add fb_ptr, fb_ptr, y, lsl #11; \ |
| | 1620 | \ |
| | 1621 | vdup.u32 v_left_x, left_x; \ |
| | 1622 | and y, y, #0x3; \ |
| | 1623 | \ |
| | 1624 | ldr dither_row, [ dither_offset_ptr, y, lsl #2 ]; \ |
| | 1625 | add fb_ptr, fb_ptr, left_x, lsl #1; \ |
| | 1626 | \ |
| | 1627 | and dither_shift, left_x, #0x03; \ |
| | 1628 | \ |
| | 1629 | vld1.u32 { uvrg }, [ span_uvrg_offset, :128 ]; \ |
| | 1630 | vshr.u32 uvrg_dx, uvrg_dx4, #2; \ |
| | 1631 | \ |
| | 1632 | mov dither_shift, dither_shift, lsl #3; \ |
| | 1633 | vmla.u32 uvrg, uvrg_dx, v_left_x; \ |
| | 1634 | \ |
| | 1635 | mov c_32, #32; \ |
| | 1636 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 1637 | \ |
| | 1638 | mov dither_row, dither_row, ror dither_shift; \ |
| | 1639 | \ |
| | 1640 | vdup.u32 dither_offsets_short, dither_row; \ |
| | 1641 | add block_span_ptr, psx_gpu, #psx_gpu_u_block_span_offset; \ |
| | 1642 | \ |
| | 1643 | vshll.s8 dither_offsets, dither_offsets_short, #4; \ |
| | 1644 | \ |
| | 1645 | vdup.u32 u_block, uv[0]; \ |
| | 1646 | \ |
| | 1647 | vdup.u32 v_block, uv[1]; \ |
| | 1648 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 1649 | \ |
| | 1650 | vadd.u32 u_block, u_block, block_span; \ |
| | 1651 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 1652 | \ |
| | 1653 | vadd.u32 v_block, v_block, block_span; \ |
| | 1654 | add block_ptr_b, block_ptr_a, #16; \ |
| | 1655 | \ |
| | 1656 | vshrn.u32 u_whole_low, u_block, #16; \ |
| | 1657 | vshrn.u32 v_whole_low, v_block, #16; \ |
| | 1658 | \ |
| | 1659 | vdup.u32 dx4, uv_dx4[0]; \ |
| | 1660 | \ |
| | 1661 | vaddhn.u32 u_whole_high, u_block, dx4; \ |
| | 1662 | vdup.u32 dx4, uv_dx4[1]; \ |
| | 1663 | \ |
| | 1664 | vaddhn.u32 v_whole_high, v_block, dx4; \ |
| | 1665 | vdup.u32 dx8, uv_dx8[0]; \ |
| | 1666 | \ |
| | 1667 | vadd.u32 u_block, u_block, dx8; \ |
| | 1668 | vdup.u32 dx8, uv_dx8[1]; \ |
| | 1669 | \ |
| | 1670 | vadd.u32 v_block, v_block, dx8; \ |
| | 1671 | vmovn.u16 u_whole_8, u_whole; \ |
| | 1672 | \ |
| | 1673 | vmovn.u16 v_whole_8, v_whole; \ |
| | 1674 | \ |
| | 1675 | pld [ fb_ptr ]; \ |
| | 1676 | vmov.u32 fb_mask_ptrs[1], fb_ptr; \ |
| | 1677 | \ |
| | 1678 | vand.u8 uv_whole_8, uv_whole_8, texture_mask; \ |
| | 1679 | setup_blocks_texture_##swizzling(); \ |
| | 1680 | \ |
| | 1681 | beq 5f; \ |
| | 1682 | \ |
| | 1683 | 4: \ |
| | 1684 | vshrn.u32 u_whole_low, u_block, #16; \ |
| | 1685 | \ |
| | 1686 | vst2.u8 { u_whole_8, v_whole_8 }, [ block_ptr_a, :128 ], c_32; \ |
| | 1687 | vshrn.u32 v_whole_low, v_block, #16; \ |
| | 1688 | \ |
| | 1689 | add block_ptr_b, block_ptr_b, #32; \ |
| | 1690 | vst1.u32 { b_whole_8, fb_mask_ptrs }, [ block_ptr_a, :128 ], c_32; \ |
| | 1691 | \ |
| | 1692 | vdup.u32 dx4, uv_dx4[0]; \ |
| | 1693 | vaddhn.u32 u_whole_high, u_block, dx4; \ |
| | 1694 | vdup.u32 dx4, uv_dx4[1]; \ |
| | 1695 | \ |
| | 1696 | vaddhn.u32 v_whole_high, v_block, dx4; \ |
| | 1697 | vdup.u32 dx8, uv_dx8[0]; \ |
| | 1698 | \ |
| | 1699 | vadd.u32 u_block, u_block, dx8; \ |
| | 1700 | vdup.u32 dx8, uv_dx8[1]; \ |
| | 1701 | \ |
| | 1702 | vadd.u32 v_block, v_block, dx8; \ |
| | 1703 | vmovn.u16 u_whole_8, u_whole; \ |
| | 1704 | \ |
| | 1705 | add fb_ptr, fb_ptr, #16; \ |
| | 1706 | vmovn.u16 v_whole_8, v_whole; \ |
| | 1707 | \ |
| | 1708 | vst1.u32 { dither_offsets }, [ block_ptr_b, :128 ], c_32; \ |
| | 1709 | pld [ fb_ptr ]; \ |
| | 1710 | \ |
| | 1711 | vmov.u32 fb_mask_ptrs[1], fb_ptr; \ |
| | 1712 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 1713 | \ |
| | 1714 | vand.u8 uv_whole_8, uv_whole_8, texture_mask; \ |
| | 1715 | setup_blocks_texture_##swizzling(); \ |
| | 1716 | \ |
| | 1717 | bne 4b; \ |
| | 1718 | \ |
| | 1719 | 5: \ |
| | 1720 | ldrh right_mask, [ span_edge_data, #edge_data_right_mask_offset ]; \ |
| | 1721 | \ |
| | 1722 | vld1.u32 { test_mask }, [ psx_gpu, :128 ]; \ |
| | 1723 | vdup.u8 draw_mask, right_mask; \ |
| | 1724 | \ |
| | 1725 | vmov.u32 fb_mask_ptrs[0], right_mask; \ |
| | 1726 | vtst.u16 draw_mask, draw_mask, test_mask; \ |
| | 1727 | vzip.u8 u_whole_8, v_whole_8; \ |
| | 1728 | \ |
| | 1729 | vbic.u16 uv_whole_8, uv_whole_8, draw_mask; \ |
| | 1730 | add block_ptr_b, block_ptr_b, #32; \ |
| | 1731 | vst1.u32 { uv_whole_8 }, [ block_ptr_a, :128 ], c_32; \ |
| | 1732 | vst1.u32 { dither_offsets }, [ block_ptr_b, :128 ], c_32; \ |
| | 1733 | vst1.u32 { b_whole_8, fb_mask_ptrs }, [ block_ptr_a, :128 ], c_32; \ |
| | 1734 | \ |
| | 1735 | 1: \ |
| | 1736 | add span_uvrg_offset, span_uvrg_offset, #16; \ |
| | 1737 | add span_edge_data, span_edge_data, #8; \ |
| | 1738 | subs num_spans, num_spans, #1; \ |
| | 1739 | \ |
| | 1740 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 1741 | bne 0b; \ |
| | 1742 | \ |
| | 1743 | ldmia sp!, { r4 - r11, pc }; \ |
| | 1744 | \ |
| | 1745 | 2: \ |
| | 1746 | /* TODO: Load from psx_gpu instead of saving/restoring these */\ |
| | 1747 | vpush { texture_mask }; \ |
| | 1748 | vpush { uvrg_dx4 }; \ |
| | 1749 | \ |
| | 1750 | stmdb sp!, { r0 - r3, r12, r14 }; \ |
| | 1751 | bl flush_render_block_buffer; \ |
| | 1752 | ldmia sp!, { r0 - r3, r12, r14 }; \ |
| | 1753 | \ |
| | 1754 | vpop { uvrg_dx4 }; \ |
| | 1755 | vpop { texture_mask }; \ |
| | 1756 | \ |
| | 1757 | vadd.u32 uvrg_dx8, uvrg_dx4, uvrg_dx4; \ |
| | 1758 | vmov.u8 fb_mask_ptrs, #0; \ |
| | 1759 | \ |
| | 1760 | mov num_blocks, span_num_blocks; \ |
| | 1761 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 1762 | bal 3b \ |
| | 1763 | |
| | 1764 | |
| | 1765 | setup_blocks_unshaded_textured_builder(swizzled) |
| | 1766 | setup_blocks_unshaded_textured_builder(unswizzled) |
| | 1767 | |
| | 1768 | |
| | 1769 | .align 3 |
| | 1770 | |
| | 1771 | function(setup_blocks_unshaded_untextured_undithered_unswizzled_indirect) |
| | 1772 | ldrh num_spans, [ psx_gpu, #psx_gpu_num_spans_offset ] |
| | 1773 | veor.u32 draw_mask, draw_mask, draw_mask |
| | 1774 | |
| | 1775 | cmp num_spans, #0 |
| | 1776 | bxeq lr |
| | 1777 | |
| | 1778 | stmdb sp!, { r4 - r11, r14 } |
| | 1779 | vld1.u32 { test_mask }, [ psx_gpu, :128 ] |
| | 1780 | |
| | 1781 | ldr color, [ psx_gpu, #psx_gpu_triangle_color_offset ] |
| | 1782 | |
| | 1783 | ubfx color_r, color, #3, #5 |
| | 1784 | ubfx color_g, color, #11, #5 |
| | 1785 | ubfx color_b, color, #19, #5 |
| | 1786 | |
| | 1787 | orr color, color_r, color_b, lsl #10 |
| | 1788 | orr color, color, color_g, lsl #5 |
| | 1789 | |
| | 1790 | vdup.u16 colors, color |
| | 1791 | |
| | 1792 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 1793 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset |
| | 1794 | |
| | 1795 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset |
| | 1796 | add block_ptr_a, block_ptr_a, num_blocks, lsl #6 |
| | 1797 | |
| | 1798 | 0: |
| | 1799 | ldrh span_num_blocks, [ span_edge_data, #edge_data_num_blocks_offset ] |
| | 1800 | ldrh y, [ span_edge_data, #edge_data_y_offset ] |
| | 1801 | |
| | 1802 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ] |
| | 1803 | |
| | 1804 | cmp span_num_blocks, #0 |
| | 1805 | beq 1f |
| | 1806 | |
| | 1807 | ldrh left_x, [ span_edge_data, #edge_data_left_x_offset ] |
| | 1808 | add num_blocks, span_num_blocks, num_blocks |
| | 1809 | |
| | 1810 | cmp num_blocks, #MAX_BLOCKS |
| | 1811 | bgt 2f |
| | 1812 | |
| | 1813 | 3: |
| | 1814 | add fb_ptr, fb_ptr, y, lsl #11 |
| | 1815 | and y, y, #0x3 |
| | 1816 | |
| | 1817 | add fb_ptr, fb_ptr, left_x, lsl #1 |
| | 1818 | mov c_32, #32 |
| | 1819 | |
| | 1820 | subs span_num_blocks, span_num_blocks, #1 |
| | 1821 | |
| | 1822 | add block_ptr_b, block_ptr_a, #16 |
| | 1823 | pld [ fb_ptr ] |
| | 1824 | |
| | 1825 | vmov.u32 fb_mask_ptrs[1], fb_ptr |
| | 1826 | beq 5f |
| | 1827 | |
| | 1828 | 4: |
| | 1829 | vst1.u32 { draw_mask }, [ block_ptr_a, :128 ], c_32 |
| | 1830 | vst1.u32 { colors }, [ block_ptr_b, :128 ], c_32 |
| | 1831 | vst1.u32 { b_whole_8, fb_mask_ptrs }, [ block_ptr_a, :128 ], c_32 |
| | 1832 | |
| | 1833 | add fb_ptr, fb_ptr, #16 |
| | 1834 | add block_ptr_b, block_ptr_b, #32 |
| | 1835 | |
| | 1836 | pld [ fb_ptr ] |
| | 1837 | |
| | 1838 | vmov.u32 fb_mask_ptrs[1], fb_ptr |
| | 1839 | subs span_num_blocks, span_num_blocks, #1 |
| | 1840 | |
| | 1841 | bne 4b |
| | 1842 | |
| | 1843 | 5: |
| | 1844 | ldrh right_mask, [ span_edge_data, #edge_data_right_mask_offset ] |
| | 1845 | |
| | 1846 | vdup.u8 draw_mask_edge, right_mask |
| | 1847 | vtst.u16 draw_mask_edge, draw_mask_edge, test_mask |
| | 1848 | |
| | 1849 | vst1.u32 { colors }, [ block_ptr_b, :128 ], c_32 |
| | 1850 | vst1.u32 { draw_mask_edge }, [ block_ptr_a, :128 ], c_32 |
| | 1851 | add block_ptr_b, block_ptr_b, #32 |
| | 1852 | vst1.u32 { b_whole_8, fb_mask_ptrs }, [ block_ptr_a, :128 ], c_32 |
| | 1853 | |
| | 1854 | 1: |
| | 1855 | add span_edge_data, span_edge_data, #8 |
| | 1856 | subs num_spans, num_spans, #1 |
| | 1857 | |
| | 1858 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 1859 | bne 0b |
| | 1860 | |
| | 1861 | ldmia sp!, { r4 - r11, pc } |
| | 1862 | |
| | 1863 | 2: |
| | 1864 | vpush { colors } |
| | 1865 | |
| | 1866 | stmdb sp!, { r0 - r3, r12, r14 } |
| | 1867 | bl flush_render_block_buffer |
| | 1868 | ldmia sp!, { r0 - r3, r12, r14 } |
| | 1869 | |
| | 1870 | vpop { colors } |
| | 1871 | |
| | 1872 | vld1.u32 { test_mask }, [ psx_gpu, :128 ] |
| | 1873 | veor.u32 draw_mask, draw_mask, draw_mask |
| | 1874 | |
| | 1875 | mov num_blocks, span_num_blocks |
| | 1876 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset |
| | 1877 | bal 3b |
| | 1878 | |
| | 1879 | |
| | 1880 | #define mask_msb_scalar r14 |
| | 1881 | |
| | 1882 | #define msb_mask q15 |
| | 1883 | |
| | 1884 | #define pixels_low d16 |
| | 1885 | |
| | 1886 | #define msb_mask_low d30 |
| | 1887 | #define msb_mask_high d31 |
| | 1888 | |
| | 1889 | |
| | 1890 | .align 3 |
| | 1891 | |
| | 1892 | function(setup_blocks_unshaded_untextured_undithered_unswizzled_direct) |
| | 1893 | ldrh num_spans, [ psx_gpu, #psx_gpu_num_spans_offset ] |
| | 1894 | |
| | 1895 | cmp num_spans, #0 |
| | 1896 | bxeq lr |
| | 1897 | |
| | 1898 | stmdb sp!, { r4 - r11, r14 } |
| | 1899 | |
| | 1900 | ldr color, [ psx_gpu, #psx_gpu_triangle_color_offset ] |
| | 1901 | |
| | 1902 | ubfx color_r, color, #3, #5 |
| | 1903 | ubfx color_g, color, #11, #5 |
| | 1904 | |
| | 1905 | ldrh mask_msb_scalar, [ psx_gpu, #psx_gpu_mask_msb_offset ] |
| | 1906 | ubfx color_b, color, #19, #5 |
| | 1907 | |
| | 1908 | orr color, color_r, color_b, lsl #10 |
| | 1909 | orr color, color, color_g, lsl #5 |
| | 1910 | orr color, color, mask_msb_scalar |
| | 1911 | |
| | 1912 | vdup.u16 colors, color |
| | 1913 | |
| | 1914 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset |
| | 1915 | orr color, color, lsl #16 |
| | 1916 | |
| | 1917 | |
| | 1918 | 0: |
| | 1919 | ldrh span_num_blocks, [ span_edge_data, #edge_data_num_blocks_offset ] |
| | 1920 | ldrh y, [ span_edge_data, #edge_data_y_offset ] |
| | 1921 | |
| | 1922 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ] |
| | 1923 | |
| | 1924 | cmp span_num_blocks, #0 |
| | 1925 | beq 1f |
| | 1926 | |
| | 1927 | ldrh left_x, [ span_edge_data, #edge_data_left_x_offset ] |
| | 1928 | |
| | 1929 | add fb_ptr, fb_ptr, y, lsl #11 |
| | 1930 | subs span_num_blocks, span_num_blocks, #1 |
| | 1931 | |
| | 1932 | add fb_ptr, fb_ptr, left_x, lsl #1 |
| | 1933 | beq 3f |
| | 1934 | |
| | 1935 | 2: |
| | 1936 | vst1.u32 { colors }, [ fb_ptr ]! |
| | 1937 | subs span_num_blocks, span_num_blocks, #1 |
| | 1938 | |
| | 1939 | bne 2b |
| | 1940 | |
| | 1941 | 3: |
| | 1942 | ldrb right_mask, [ span_edge_data, #edge_data_right_mask_offset ] |
| | 1943 | |
| | 1944 | cmp right_mask, #0x0 |
| | 1945 | beq 5f |
| | 1946 | |
| | 1947 | tst right_mask, #0xF |
| | 1948 | streq color, [ fb_ptr ], #4 |
| | 1949 | moveq right_mask, right_mask, lsr #4 |
| | 1950 | streq color, [ fb_ptr ], #4 |
| | 1951 | |
| | 1952 | tst right_mask, #0x3 |
| | 1953 | streq color, [ fb_ptr ], #4 |
| | 1954 | moveq right_mask, right_mask, lsr #2 |
| | 1955 | |
| | 1956 | tst right_mask, #0x1 |
| | 1957 | streqh color, [ fb_ptr ] |
| | 1958 | |
| | 1959 | 1: |
| | 1960 | add span_edge_data, span_edge_data, #8 |
| | 1961 | subs num_spans, num_spans, #1 |
| | 1962 | bne 0b |
| | 1963 | |
| | 1964 | ldmia sp!, { r4 - r11, pc } |
| | 1965 | |
| | 1966 | 5: |
| | 1967 | vst1.u32 { colors }, [ fb_ptr ] |
| | 1968 | bal 1b |
| | 1969 | |
| | 1970 | |
| | 1971 | #undef c_64 |
| | 1972 | |
| | 1973 | #define c_64 r7 |
| | 1974 | #define rg_dx_ptr r2 |
| | 1975 | |
| | 1976 | |
| | 1977 | #undef r_block |
| | 1978 | #undef g_block |
| | 1979 | #undef b_block |
| | 1980 | #undef r_whole |
| | 1981 | #undef g_whole |
| | 1982 | #undef b_whole |
| | 1983 | #undef r_whole_low |
| | 1984 | #undef r_whole_high |
| | 1985 | #undef g_whole_low |
| | 1986 | #undef g_whole_high |
| | 1987 | #undef b_whole_low |
| | 1988 | #undef b_whole_high |
| | 1989 | #undef r_whole_8 |
| | 1990 | #undef g_whole_8 |
| | 1991 | #undef b_whole_8 |
| | 1992 | #undef dither_offsets |
| | 1993 | #undef rg_dx4 |
| | 1994 | #undef rg_dx8 |
| | 1995 | #undef dx4 |
| | 1996 | #undef dx8 |
| | 1997 | #undef v_left_x |
| | 1998 | #undef uvrg |
| | 1999 | #undef block_span |
| | 2000 | #undef rg |
| | 2001 | #undef draw_mask |
| | 2002 | #undef test_mask |
| | 2003 | |
| | 2004 | #define r_block q0 |
| | 2005 | #define g_block q1 |
| | 2006 | #define b_block q2 |
| | 2007 | |
| | 2008 | #define r_whole q3 |
| | 2009 | #define g_whole q4 |
| | 2010 | #define b_whole q5 |
| | 2011 | |
| | 2012 | #define r_whole_low d6 |
| | 2013 | #define r_whole_high d7 |
| | 2014 | #define g_whole_low d8 |
| | 2015 | #define g_whole_high d9 |
| | 2016 | #define b_whole_low d10 |
| | 2017 | #define b_whole_high d11 |
| | 2018 | |
| | 2019 | #define gb_whole_8 q6 |
| | 2020 | |
| | 2021 | #define g_whole_8 d12 |
| | 2022 | #define b_whole_8 d13 |
| | 2023 | |
| | 2024 | #define r_whole_8 d14 |
| | 2025 | |
| | 2026 | #define pixels q8 |
| | 2027 | |
| | 2028 | #define rg_dx4 d18 |
| | 2029 | #define rg_dx8 d19 |
| | 2030 | |
| | 2031 | #define dx4 q10 |
| | 2032 | #define dx8 q10 |
| | 2033 | |
| | 2034 | #define v_left_x d6 |
| | 2035 | #define uvrg q4 |
| | 2036 | #define block_span q5 |
| | 2037 | |
| | 2038 | #define rg d9 |
| | 2039 | |
| | 2040 | #define d64_1 d22 |
| | 2041 | #define d64_128 d23 |
| | 2042 | |
| | 2043 | #define d128_4 q12 |
| | 2044 | #define d128_0x7 q13 |
| | 2045 | |
| | 2046 | #define d64_4 d24 |
| | 2047 | |
| | 2048 | #define dither_offsets q14 |
| | 2049 | #define draw_mask q15 |
| | 2050 | |
| | 2051 | #define dither_offsets_low d28 |
| | 2052 | |
| | 2053 | #define rg_dx d0 |
| | 2054 | #define test_mask q10 |
| | 2055 | |
| | 2056 | |
| | 2057 | #define setup_blocks_shaded_untextured_dither_a_dithered() \ |
| | 2058 | vqadd.u8 r_whole_8, r_whole_8, dither_offsets_low; \ |
| | 2059 | vqadd.u8 gb_whole_8, gb_whole_8, dither_offsets; \ |
| | 2060 | |
| | 2061 | #define setup_blocks_shaded_untextured_dither_b_dithered() \ |
| | 2062 | vqsub.u8 r_whole_8, r_whole_8, d64_4; \ |
| | 2063 | vqsub.u8 gb_whole_8, gb_whole_8, d128_4 \ |
| | 2064 | |
| | 2065 | #define setup_blocks_shaded_untextured_dither_a_undithered() \ |
| | 2066 | |
| | 2067 | #define setup_blocks_shaded_untextured_dither_b_undithered() \ |
| | 2068 | |
| | 2069 | |
| | 2070 | #define setup_blocks_shaded_untextured_indirect_builder(dithering) \ |
| | 2071 | .align 3; \ |
| | 2072 | \ |
| | 2073 | function(setup_blocks_shaded_untextured_##dithering##_unswizzled_indirect) \ |
| | 2074 | ldrh num_spans, [ psx_gpu, #psx_gpu_num_spans_offset ]; \ |
| | 2075 | add rg_dx_ptr, psx_gpu, #(psx_gpu_uvrg_dx_offset + 8); \ |
| | 2076 | \ |
| | 2077 | vld1.u32 { rg_dx }, [ rg_dx_ptr, :64 ]; \ |
| | 2078 | \ |
| | 2079 | cmp num_spans, #0; \ |
| | 2080 | bxeq lr; \ |
| | 2081 | \ |
| | 2082 | stmdb sp!, { r4 - r11, r14 }; \ |
| | 2083 | vshl.u32 rg_dx4, rg_dx, #2; \ |
| | 2084 | \ |
| | 2085 | ldr b_dx, [ psx_gpu, #psx_gpu_b_dx_offset ]; \ |
| | 2086 | vshl.u32 rg_dx8, rg_dx, #3; \ |
| | 2087 | \ |
| | 2088 | add span_uvrg_offset, psx_gpu, #psx_gpu_span_uvrg_offset_offset; \ |
| | 2089 | \ |
| | 2090 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 2091 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset; \ |
| | 2092 | \ |
| | 2093 | add span_b_offset, psx_gpu, #psx_gpu_span_b_offset_offset; \ |
| | 2094 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 2095 | \ |
| | 2096 | add block_ptr_a, block_ptr_a, num_blocks, lsl #6; \ |
| | 2097 | vmov.u8 d64_1, #1; \ |
| | 2098 | \ |
| | 2099 | vmov.u8 d128_4, #4; \ |
| | 2100 | vmov.u8 d64_128, #128; \ |
| | 2101 | \ |
| | 2102 | vmov.u8 d128_0x7, #0x7; \ |
| | 2103 | \ |
| | 2104 | 0: \ |
| | 2105 | ldrh span_num_blocks, [ span_edge_data, #edge_data_num_blocks_offset ]; \ |
| | 2106 | add dither_offset_ptr, psx_gpu, #psx_gpu_dither_table_offset; \ |
| | 2107 | \ |
| | 2108 | ldrh y, [ span_edge_data, #edge_data_y_offset ]; \ |
| | 2109 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ]; \ |
| | 2110 | \ |
| | 2111 | cmp span_num_blocks, #0; \ |
| | 2112 | beq 1f; \ |
| | 2113 | \ |
| | 2114 | ldrh left_x, [ span_edge_data, #edge_data_left_x_offset ]; \ |
| | 2115 | add num_blocks, span_num_blocks, num_blocks; \ |
| | 2116 | \ |
| | 2117 | cmp num_blocks, #MAX_BLOCKS; \ |
| | 2118 | bgt 2f; \ |
| | 2119 | \ |
| | 2120 | 3: \ |
| | 2121 | ldr b, [ span_b_offset ]; \ |
| | 2122 | add fb_ptr, fb_ptr, y, lsl #11; \ |
| | 2123 | \ |
| | 2124 | vdup.u32 v_left_x, left_x; \ |
| | 2125 | and y, y, #0x3; \ |
| | 2126 | \ |
| | 2127 | ldr dither_row, [ dither_offset_ptr, y, lsl #2 ]; \ |
| | 2128 | add fb_ptr, fb_ptr, left_x, lsl #1; \ |
| | 2129 | \ |
| | 2130 | mla b, b_dx, left_x, b; \ |
| | 2131 | and dither_shift, left_x, #0x03; \ |
| | 2132 | \ |
| | 2133 | vld1.u32 { uvrg }, [ span_uvrg_offset, :128 ]; \ |
| | 2134 | vshr.u32 rg_dx, rg_dx4, #2; \ |
| | 2135 | \ |
| | 2136 | mov dither_shift, dither_shift, lsl #3; \ |
| | 2137 | vmla.u32 rg, rg_dx, v_left_x; \ |
| | 2138 | \ |
| | 2139 | mov c_64, #64; \ |
| | 2140 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 2141 | \ |
| | 2142 | mov dither_row, dither_row, ror dither_shift; \ |
| | 2143 | mov b_dx4, b_dx, lsl #2; \ |
| | 2144 | \ |
| | 2145 | vdup.u32 dither_offsets, dither_row; \ |
| | 2146 | add block_span_ptr, psx_gpu, #psx_gpu_r_block_span_offset; \ |
| | 2147 | \ |
| | 2148 | vdup.u32 b_block, b; \ |
| | 2149 | vadd.u8 dither_offsets, dither_offsets, d128_4; \ |
| | 2150 | \ |
| | 2151 | mov b_dx8, b_dx, lsl #3; \ |
| | 2152 | vdup.u32 r_block, rg[0]; \ |
| | 2153 | vdup.u32 g_block, rg[1]; \ |
| | 2154 | \ |
| | 2155 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 2156 | \ |
| | 2157 | vadd.u32 r_block, r_block, block_span; \ |
| | 2158 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 2159 | \ |
| | 2160 | vadd.u32 g_block, g_block, block_span; \ |
| | 2161 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]; \ |
| | 2162 | \ |
| | 2163 | vadd.u32 b_block, b_block, block_span; \ |
| | 2164 | add block_ptr_b, block_ptr_a, #16; \ |
| | 2165 | \ |
| | 2166 | vshrn.u32 r_whole_low, r_block, #16; \ |
| | 2167 | vshrn.u32 g_whole_low, g_block, #16; \ |
| | 2168 | vshrn.u32 b_whole_low, b_block, #16; \ |
| | 2169 | vdup.u32 dx4, rg_dx4[0]; \ |
| | 2170 | \ |
| | 2171 | vaddhn.u32 r_whole_high, r_block, dx4; \ |
| | 2172 | vdup.u32 dx4, rg_dx4[1]; \ |
| | 2173 | \ |
| | 2174 | vaddhn.u32 g_whole_high, g_block, dx4; \ |
| | 2175 | vdup.u32 dx4, b_dx4; \ |
| | 2176 | \ |
| | 2177 | vaddhn.u32 b_whole_high, b_block, dx4; \ |
| | 2178 | vdup.u32 dx8, rg_dx8[0]; \ |
| | 2179 | \ |
| | 2180 | vadd.u32 r_block, r_block, dx8; \ |
| | 2181 | vdup.u32 dx8, rg_dx8[1]; \ |
| | 2182 | \ |
| | 2183 | vadd.u32 g_block, g_block, dx8; \ |
| | 2184 | vdup.u32 dx8, b_dx8; \ |
| | 2185 | \ |
| | 2186 | vadd.u32 b_block, b_block, dx8; \ |
| | 2187 | \ |
| | 2188 | vmovn.u16 r_whole_8, r_whole; \ |
| | 2189 | vmovn.u16 g_whole_8, g_whole; \ |
| | 2190 | vmovn.u16 b_whole_8, b_whole; \ |
| | 2191 | \ |
| | 2192 | beq 5f; \ |
| | 2193 | veor.u32 draw_mask, draw_mask, draw_mask; \ |
| | 2194 | \ |
| | 2195 | 4: \ |
| | 2196 | setup_blocks_shaded_untextured_dither_a_##dithering(); \ |
| | 2197 | vshrn.u32 r_whole_low, r_block, #16; \ |
| | 2198 | \ |
| | 2199 | setup_blocks_shaded_untextured_dither_b_##dithering(); \ |
| | 2200 | vshrn.u32 g_whole_low, g_block, #16; \ |
| | 2201 | \ |
| | 2202 | vshrn.u32 b_whole_low, b_block, #16; \ |
| | 2203 | str fb_ptr, [ block_ptr_a, #44 ]; \ |
| | 2204 | \ |
| | 2205 | vdup.u32 dx4, rg_dx4[0]; \ |
| | 2206 | vshr.u8 r_whole_8, r_whole_8, #3; \ |
| | 2207 | vbic.u8 gb_whole_8, gb_whole_8, d128_0x7; \ |
| | 2208 | \ |
| | 2209 | vaddhn.u32 r_whole_high, r_block, dx4; \ |
| | 2210 | vdup.u32 dx4, rg_dx4[1]; \ |
| | 2211 | \ |
| | 2212 | vaddhn.u32 g_whole_high, g_block, dx4; \ |
| | 2213 | vdup.u32 dx4, b_dx4; \ |
| | 2214 | \ |
| | 2215 | vaddhn.u32 b_whole_high, b_block, dx4; \ |
| | 2216 | vdup.u32 dx8, rg_dx8[0]; \ |
| | 2217 | \ |
| | 2218 | vmull.u8 pixels, r_whole_8, d64_1; \ |
| | 2219 | vmlal.u8 pixels, g_whole_8, d64_4; \ |
| | 2220 | vmlal.u8 pixels, b_whole_8, d64_128; \ |
| | 2221 | \ |
| | 2222 | vadd.u32 r_block, r_block, dx8; \ |
| | 2223 | vdup.u32 dx8, rg_dx8[1]; \ |
| | 2224 | \ |
| | 2225 | vadd.u32 g_block, g_block, dx8; \ |
| | 2226 | vdup.u32 dx8, b_dx8; \ |
| | 2227 | \ |
| | 2228 | vadd.u32 b_block, b_block, dx8; \ |
| | 2229 | add fb_ptr, fb_ptr, #16; \ |
| | 2230 | \ |
| | 2231 | vmovn.u16 r_whole_8, r_whole; \ |
| | 2232 | vmovn.u16 g_whole_8, g_whole; \ |
| | 2233 | vmovn.u16 b_whole_8, b_whole; \ |
| | 2234 | \ |
| | 2235 | vst1.u32 { draw_mask }, [ block_ptr_a, :128 ], c_64; \ |
| | 2236 | vst1.u32 { pixels }, [ block_ptr_b, :128 ], c_64; \ |
| | 2237 | \ |
| | 2238 | pld [ fb_ptr ]; \ |
| | 2239 | \ |
| | 2240 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 2241 | bne 4b; \ |
| | 2242 | \ |
| | 2243 | 5: \ |
| | 2244 | str fb_ptr, [ block_ptr_a, #44 ]; \ |
| | 2245 | setup_blocks_shaded_untextured_dither_a_##dithering(); \ |
| | 2246 | \ |
| | 2247 | ldrh right_mask, [ span_edge_data, #edge_data_right_mask_offset ]; \ |
| | 2248 | setup_blocks_shaded_untextured_dither_b_##dithering(); \ |
| | 2249 | \ |
| | 2250 | vshr.u8 r_whole_8, r_whole_8, #3; \ |
| | 2251 | vdup.u8 draw_mask, right_mask; \ |
| | 2252 | \ |
| | 2253 | vbic.u8 gb_whole_8, gb_whole_8, d128_0x7; \ |
| | 2254 | vld1.u32 { test_mask }, [ psx_gpu, :128 ]; \ |
| | 2255 | \ |
| | 2256 | vtst.u16 draw_mask, draw_mask, test_mask; \ |
| | 2257 | \ |
| | 2258 | vmull.u8 pixels, r_whole_8, d64_1; \ |
| | 2259 | vmlal.u8 pixels, g_whole_8, d64_4; \ |
| | 2260 | vmlal.u8 pixels, b_whole_8, d64_128; \ |
| | 2261 | \ |
| | 2262 | vst1.u32 { draw_mask }, [ block_ptr_a, :128 ], c_64; \ |
| | 2263 | vst1.u32 { pixels }, [ block_ptr_b, :128 ], c_64; \ |
| | 2264 | \ |
| | 2265 | 1: \ |
| | 2266 | add span_uvrg_offset, span_uvrg_offset, #16; \ |
| | 2267 | add span_b_offset, span_b_offset, #4; \ |
| | 2268 | \ |
| | 2269 | add span_edge_data, span_edge_data, #8; \ |
| | 2270 | subs num_spans, num_spans, #1; \ |
| | 2271 | \ |
| | 2272 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 2273 | bne 0b; \ |
| | 2274 | \ |
| | 2275 | ldmia sp!, { r4 - r11, pc }; \ |
| | 2276 | \ |
| | 2277 | 2: \ |
| | 2278 | /* TODO: Load from psx_gpu instead of saving/restoring these */\ |
| | 2279 | vpush { rg_dx4 }; \ |
| | 2280 | \ |
| | 2281 | stmdb sp!, { r0 - r3, r12, r14 }; \ |
| | 2282 | bl flush_render_block_buffer; \ |
| | 2283 | ldmia sp!, { r0 - r3, r12, r14 }; \ |
| | 2284 | \ |
| | 2285 | vpop { rg_dx4 }; \ |
| | 2286 | \ |
| | 2287 | vmov.u8 d64_1, #1; \ |
| | 2288 | vmov.u8 d128_4, #4; \ |
| | 2289 | vmov.u8 d64_128, #128; \ |
| | 2290 | vmov.u8 d128_0x7, #0x7; \ |
| | 2291 | \ |
| | 2292 | vadd.u32 rg_dx8, rg_dx4, rg_dx4; \ |
| | 2293 | \ |
| | 2294 | mov num_blocks, span_num_blocks; \ |
| | 2295 | add block_ptr_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 2296 | bal 3b \ |
| | 2297 | |
| | 2298 | |
| | 2299 | setup_blocks_shaded_untextured_indirect_builder(undithered) |
| | 2300 | setup_blocks_shaded_untextured_indirect_builder(dithered) |
| | 2301 | |
| | 2302 | |
| | 2303 | #undef draw_mask |
| | 2304 | |
| | 2305 | #define mask_msb_ptr r14 |
| | 2306 | |
| | 2307 | #define draw_mask q0 |
| | 2308 | #define pixels_low d16 |
| | 2309 | #define pixels_high d17 |
| | 2310 | |
| | 2311 | |
| | 2312 | |
| | 2313 | #define setup_blocks_shaded_untextured_direct_builder(dithering) \ |
| | 2314 | .align 3; \ |
| | 2315 | \ |
| | 2316 | function(setup_blocks_shaded_untextured_##dithering##_unswizzled_direct) \ |
| | 2317 | ldrh num_spans, [ psx_gpu, #psx_gpu_num_spans_offset ]; \ |
| | 2318 | add rg_dx_ptr, psx_gpu, #(psx_gpu_uvrg_dx_offset + 8); \ |
| | 2319 | \ |
| | 2320 | vld1.u32 { rg_dx }, [ rg_dx_ptr, :64 ]; \ |
| | 2321 | \ |
| | 2322 | cmp num_spans, #0; \ |
| | 2323 | bxeq lr; \ |
| | 2324 | \ |
| | 2325 | stmdb sp!, { r4 - r11, r14 }; \ |
| | 2326 | vshl.u32 rg_dx4, rg_dx, #2; \ |
| | 2327 | \ |
| | 2328 | ldr b_dx, [ psx_gpu, #psx_gpu_b_dx_offset ]; \ |
| | 2329 | vshl.u32 rg_dx8, rg_dx, #3; \ |
| | 2330 | \ |
| | 2331 | add span_uvrg_offset, psx_gpu, #psx_gpu_span_uvrg_offset_offset; \ |
| | 2332 | add span_edge_data, psx_gpu, #psx_gpu_span_edge_data_offset; \ |
| | 2333 | \ |
| | 2334 | add span_b_offset, psx_gpu, #psx_gpu_span_b_offset_offset; \ |
| | 2335 | vmov.u8 d64_1, #1; \ |
| | 2336 | \ |
| | 2337 | vmov.u8 d128_4, #4; \ |
| | 2338 | vmov.u8 d64_128, #128; \ |
| | 2339 | \ |
| | 2340 | vmov.u8 d128_0x7, #0x7; \ |
| | 2341 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 2342 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 2343 | \ |
| | 2344 | 0: \ |
| | 2345 | ldrh span_num_blocks, [ span_edge_data, #edge_data_num_blocks_offset ]; \ |
| | 2346 | add dither_offset_ptr, psx_gpu, #psx_gpu_dither_table_offset; \ |
| | 2347 | \ |
| | 2348 | ldrh y, [ span_edge_data, #edge_data_y_offset ]; \ |
| | 2349 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ]; \ |
| | 2350 | \ |
| | 2351 | cmp span_num_blocks, #0; \ |
| | 2352 | beq 1f; \ |
| | 2353 | \ |
| | 2354 | ldrh left_x, [ span_edge_data, #edge_data_left_x_offset ]; \ |
| | 2355 | add fb_ptr, fb_ptr, y, lsl #11; \ |
| | 2356 | \ |
| | 2357 | ldr b, [ span_b_offset ]; \ |
| | 2358 | vdup.u32 v_left_x, left_x; \ |
| | 2359 | and y, y, #0x3; \ |
| | 2360 | \ |
| | 2361 | ldr dither_row, [ dither_offset_ptr, y, lsl #2 ]; \ |
| | 2362 | add fb_ptr, fb_ptr, left_x, lsl #1; \ |
| | 2363 | \ |
| | 2364 | mla b, b_dx, left_x, b; \ |
| | 2365 | and dither_shift, left_x, #0x03; \ |
| | 2366 | \ |
| | 2367 | vld1.u32 { uvrg }, [ span_uvrg_offset, :128 ]; \ |
| | 2368 | vshr.u32 rg_dx, rg_dx4, #2; \ |
| | 2369 | \ |
| | 2370 | mov dither_shift, dither_shift, lsl #3; \ |
| | 2371 | vmla.u32 rg, rg_dx, v_left_x; \ |
| | 2372 | \ |
| | 2373 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 2374 | \ |
| | 2375 | mov dither_row, dither_row, ror dither_shift; \ |
| | 2376 | mov b_dx4, b_dx, lsl #2; \ |
| | 2377 | \ |
| | 2378 | vdup.u32 dither_offsets, dither_row; \ |
| | 2379 | add block_span_ptr, psx_gpu, #psx_gpu_r_block_span_offset; \ |
| | 2380 | \ |
| | 2381 | vdup.u32 b_block, b; \ |
| | 2382 | vadd.u8 dither_offsets, dither_offsets, d128_4; \ |
| | 2383 | \ |
| | 2384 | mov b_dx8, b_dx, lsl #3; \ |
| | 2385 | vdup.u32 r_block, rg[0]; \ |
| | 2386 | vdup.u32 g_block, rg[1]; \ |
| | 2387 | \ |
| | 2388 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 2389 | \ |
| | 2390 | vadd.u32 r_block, r_block, block_span; \ |
| | 2391 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]!; \ |
| | 2392 | \ |
| | 2393 | vadd.u32 g_block, g_block, block_span; \ |
| | 2394 | vld1.u32 { block_span }, [ block_span_ptr, :128 ]; \ |
| | 2395 | \ |
| | 2396 | vadd.u32 b_block, b_block, block_span; \ |
| | 2397 | add block_ptr_b, block_ptr_a, #16; \ |
| | 2398 | \ |
| | 2399 | vshrn.u32 r_whole_low, r_block, #16; \ |
| | 2400 | vshrn.u32 g_whole_low, g_block, #16; \ |
| | 2401 | vshrn.u32 b_whole_low, b_block, #16; \ |
| | 2402 | vdup.u32 dx4, rg_dx4[0]; \ |
| | 2403 | \ |
| | 2404 | vaddhn.u32 r_whole_high, r_block, dx4; \ |
| | 2405 | vdup.u32 dx4, rg_dx4[1]; \ |
| | 2406 | \ |
| | 2407 | vaddhn.u32 g_whole_high, g_block, dx4; \ |
| | 2408 | vdup.u32 dx4, b_dx4; \ |
| | 2409 | \ |
| | 2410 | vaddhn.u32 b_whole_high, b_block, dx4; \ |
| | 2411 | vdup.u32 dx8, rg_dx8[0]; \ |
| | 2412 | \ |
| | 2413 | vadd.u32 r_block, r_block, dx8; \ |
| | 2414 | vdup.u32 dx8, rg_dx8[1]; \ |
| | 2415 | \ |
| | 2416 | vadd.u32 g_block, g_block, dx8; \ |
| | 2417 | vdup.u32 dx8, b_dx8; \ |
| | 2418 | \ |
| | 2419 | vadd.u32 b_block, b_block, dx8; \ |
| | 2420 | \ |
| | 2421 | vmovn.u16 r_whole_8, r_whole; \ |
| | 2422 | vmovn.u16 g_whole_8, g_whole; \ |
| | 2423 | vmovn.u16 b_whole_8, b_whole; \ |
| | 2424 | \ |
| | 2425 | beq 3f; \ |
| | 2426 | \ |
| | 2427 | 2: \ |
| | 2428 | setup_blocks_shaded_untextured_dither_a_##dithering(); \ |
| | 2429 | vshrn.u32 r_whole_low, r_block, #16; \ |
| | 2430 | \ |
| | 2431 | setup_blocks_shaded_untextured_dither_b_##dithering(); \ |
| | 2432 | vshrn.u32 g_whole_low, g_block, #16; \ |
| | 2433 | \ |
| | 2434 | vshrn.u32 b_whole_low, b_block, #16; \ |
| | 2435 | \ |
| | 2436 | vdup.u32 dx4, rg_dx4[0]; \ |
| | 2437 | vshr.u8 r_whole_8, r_whole_8, #3; \ |
| | 2438 | vbic.u8 gb_whole_8, gb_whole_8, d128_0x7; \ |
| | 2439 | \ |
| | 2440 | vaddhn.u32 r_whole_high, r_block, dx4; \ |
| | 2441 | vdup.u32 dx4, rg_dx4[1]; \ |
| | 2442 | \ |
| | 2443 | vmov pixels, msb_mask; \ |
| | 2444 | vaddhn.u32 g_whole_high, g_block, dx4; \ |
| | 2445 | vdup.u32 dx4, b_dx4; \ |
| | 2446 | \ |
| | 2447 | vaddhn.u32 b_whole_high, b_block, dx4; \ |
| | 2448 | vdup.u32 dx8, rg_dx8[0]; \ |
| | 2449 | \ |
| | 2450 | vmlal.u8 pixels, r_whole_8, d64_1; \ |
| | 2451 | vmlal.u8 pixels, g_whole_8, d64_4; \ |
| | 2452 | vmlal.u8 pixels, b_whole_8, d64_128; \ |
| | 2453 | \ |
| | 2454 | vadd.u32 r_block, r_block, dx8; \ |
| | 2455 | vdup.u32 dx8, rg_dx8[1]; \ |
| | 2456 | \ |
| | 2457 | vadd.u32 g_block, g_block, dx8; \ |
| | 2458 | vdup.u32 dx8, b_dx8; \ |
| | 2459 | \ |
| | 2460 | vadd.u32 b_block, b_block, dx8; \ |
| | 2461 | \ |
| | 2462 | vmovn.u16 r_whole_8, r_whole; \ |
| | 2463 | vmovn.u16 g_whole_8, g_whole; \ |
| | 2464 | vmovn.u16 b_whole_8, b_whole; \ |
| | 2465 | \ |
| | 2466 | vst1.u32 { pixels }, [ fb_ptr ]!; \ |
| | 2467 | subs span_num_blocks, span_num_blocks, #1; \ |
| | 2468 | bne 2b; \ |
| | 2469 | \ |
| | 2470 | 3: \ |
| | 2471 | setup_blocks_shaded_untextured_dither_a_##dithering(); \ |
| | 2472 | \ |
| | 2473 | ldrh right_mask, [ span_edge_data, #edge_data_right_mask_offset ]; \ |
| | 2474 | setup_blocks_shaded_untextured_dither_b_##dithering(); \ |
| | 2475 | \ |
| | 2476 | vshr.u8 r_whole_8, r_whole_8, #3; \ |
| | 2477 | rbit right_mask, right_mask; \ |
| | 2478 | vmov pixels, msb_mask; \ |
| | 2479 | vbic.u8 gb_whole_8, gb_whole_8, d128_0x7; \ |
| | 2480 | clz right_mask, right_mask; \ |
| | 2481 | \ |
| | 2482 | vmlal.u8 pixels, r_whole_8, d64_1; \ |
| | 2483 | vmlal.u8 pixels, g_whole_8, d64_4; \ |
| | 2484 | vmlal.u8 pixels, b_whole_8, d64_128; \ |
| | 2485 | \ |
| | 2486 | ldr pc, [ pc, right_mask, lsl #2 ]; \ |
| | 2487 | nop; \ |
| | 2488 | nop; \ |
| | 2489 | .word 4f; \ |
| | 2490 | .word 5f; \ |
| | 2491 | .word 6f; \ |
| | 2492 | .word 7f; \ |
| | 2493 | .word 8f; \ |
| | 2494 | .word 9f; \ |
| | 2495 | .word 10f; \ |
| | 2496 | .word 11f; \ |
| | 2497 | \ |
| | 2498 | 4: \ |
| | 2499 | vst1.u16 { pixels_low[0] }, [ fb_ptr ]; \ |
| | 2500 | bal 1f; \ |
| | 2501 | \ |
| | 2502 | 5: \ |
| | 2503 | vst1.u32 { pixels_low[0] }, [ fb_ptr ]; \ |
| | 2504 | bal 1f; \ |
| | 2505 | \ |
| | 2506 | 6: \ |
| | 2507 | vst1.u32 { pixels_low[0] }, [ fb_ptr ]!; \ |
| | 2508 | vst1.u16 { pixels_low[2] }, [ fb_ptr ]; \ |
| | 2509 | bal 1f; \ |
| | 2510 | \ |
| | 2511 | 7: \ |
| | 2512 | vst1.u32 { pixels_low }, [ fb_ptr ]; \ |
| | 2513 | bal 1f; \ |
| | 2514 | \ |
| | 2515 | 8: \ |
| | 2516 | vst1.u32 { pixels_low }, [ fb_ptr ]!; \ |
| | 2517 | vst1.u16 { pixels_high[0] }, [ fb_ptr ]; \ |
| | 2518 | bal 1f; \ |
| | 2519 | \ |
| | 2520 | 9: \ |
| | 2521 | vst1.u32 { pixels_low }, [ fb_ptr ]!; \ |
| | 2522 | vst1.u32 { pixels_high[0] }, [ fb_ptr ]!; \ |
| | 2523 | bal 1f; \ |
| | 2524 | \ |
| | 2525 | 10: \ |
| | 2526 | vst1.u32 { pixels_low }, [ fb_ptr ]!; \ |
| | 2527 | vst1.u32 { pixels_high[0] }, [ fb_ptr ]!; \ |
| | 2528 | vst1.u16 { pixels_high[2] }, [ fb_ptr ]; \ |
| | 2529 | bal 1f; \ |
| | 2530 | \ |
| | 2531 | 11: \ |
| | 2532 | vst1.u32 { pixels }, [ fb_ptr ]; \ |
| | 2533 | bal 1f; \ |
| | 2534 | \ |
| | 2535 | 1: \ |
| | 2536 | add span_uvrg_offset, span_uvrg_offset, #16; \ |
| | 2537 | add span_b_offset, span_b_offset, #4; \ |
| | 2538 | \ |
| | 2539 | add span_edge_data, span_edge_data, #8; \ |
| | 2540 | subs num_spans, num_spans, #1; \ |
| | 2541 | \ |
| | 2542 | bne 0b; \ |
| | 2543 | \ |
| | 2544 | ldmia sp!, { r4 - r11, pc } \ |
| | 2545 | |
| | 2546 | setup_blocks_shaded_untextured_direct_builder(undithered) |
| | 2547 | setup_blocks_shaded_untextured_direct_builder(dithered) |
| | 2548 | |
| | 2549 | |
| | 2550 | #undef psx_gpu |
| | 2551 | #undef num_blocks |
| | 2552 | #undef triangle |
| | 2553 | #undef c_64 |
| | 2554 | |
| | 2555 | #define psx_gpu r0 |
| | 2556 | #define block_ptr r1 |
| | 2557 | #define num_blocks r2 |
| | 2558 | #define uv_01 r3 |
| | 2559 | #define uv_23 r4 |
| | 2560 | #define uv_45 r5 |
| | 2561 | #define uv_67 r6 |
| | 2562 | #define uv_0 r7 |
| | 2563 | #define uv_1 r3 |
| | 2564 | #define uv_2 r8 |
| | 2565 | #define uv_3 r4 |
| | 2566 | #define uv_4 r9 |
| | 2567 | #define uv_5 r5 |
| | 2568 | #define uv_6 r10 |
| | 2569 | #define uv_7 r6 |
| | 2570 | #define texture_ptr r11 |
| | 2571 | |
| | 2572 | #define pixel_0 r7 |
| | 2573 | #define pixel_1 r3 |
| | 2574 | #define pixel_2 r8 |
| | 2575 | #define pixel_3 r4 |
| | 2576 | #define pixel_4 r9 |
| | 2577 | #define pixel_5 r5 |
| | 2578 | #define pixel_6 r10 |
| | 2579 | #define pixel_7 r6 |
| | 2580 | |
| | 2581 | #define pixels_a r7 |
| | 2582 | #define pixels_b r9 |
| | 2583 | #define pixels_c r8 |
| | 2584 | #define pixels_d r10 |
| | 2585 | |
| | 2586 | #define c_64 r0 |
| | 2587 | |
| | 2588 | #define clut_ptr r12 |
| | 2589 | #define current_texture_mask r5 |
| | 2590 | #define dirty_textures_mask r6 |
| | 2591 | |
| | 2592 | #define texels d0 |
| | 2593 | |
| | 2594 | #define clut_low_a d2 |
| | 2595 | #define clut_low_b d3 |
| | 2596 | #define clut_high_a d4 |
| | 2597 | #define clut_high_b d5 |
| | 2598 | |
| | 2599 | #define clut_a q1 |
| | 2600 | #define clut_b q2 |
| | 2601 | |
| | 2602 | #define texels_low d6 |
| | 2603 | #define texels_high d7 |
| | 2604 | |
| | 2605 | .align 3 |
| | 2606 | |
| | 2607 | function(texture_blocks_untextured) |
| | 2608 | bx lr |
| | 2609 | |
| | 2610 | |
| | 2611 | .align 3 |
| | 2612 | |
| | 2613 | function(texture_blocks_4bpp) |
| | 2614 | stmdb sp!, { r3 - r11, r14 } |
| | 2615 | add block_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 2616 | |
| | 2617 | ldr texture_ptr, [ psx_gpu, #psx_gpu_texture_page_ptr_offset ] |
| | 2618 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 2619 | |
| | 2620 | ldr clut_ptr, [ psx_gpu, #psx_gpu_clut_ptr_offset ] |
| | 2621 | vld1.u32 { clut_a, clut_b }, [ clut_ptr, :128 ] |
| | 2622 | |
| | 2623 | ldr current_texture_mask, [ psx_gpu, #psx_gpu_current_texture_mask_offset ] |
| | 2624 | vuzp.u8 clut_a, clut_b |
| | 2625 | |
| | 2626 | ldr dirty_textures_mask, [ psx_gpu, #psx_gpu_dirty_textures_4bpp_mask_offset ] |
| | 2627 | tst dirty_textures_mask, current_texture_mask |
| | 2628 | |
| | 2629 | bne 1f |
| | 2630 | mov c_64, #64 |
| | 2631 | |
| | 2632 | 0: |
| | 2633 | ldm block_ptr, { uv_01, uv_23, uv_45, uv_67 } |
| | 2634 | |
| | 2635 | uxtah uv_0, texture_ptr, uv_01 |
| | 2636 | uxtah uv_1, texture_ptr, uv_01, ror #16 |
| | 2637 | |
| | 2638 | uxtah uv_2, texture_ptr, uv_23 |
| | 2639 | uxtah uv_3, texture_ptr, uv_23, ror #16 |
| | 2640 | |
| | 2641 | uxtah uv_4, texture_ptr, uv_45 |
| | 2642 | ldrb pixel_0, [ uv_0 ] |
| | 2643 | |
| | 2644 | uxtah uv_5, texture_ptr, uv_45, ror #16 |
| | 2645 | ldrb pixel_1, [ uv_1 ] |
| | 2646 | |
| | 2647 | uxtah uv_6, texture_ptr, uv_67 |
| | 2648 | ldrb pixel_2, [ uv_2 ] |
| | 2649 | |
| | 2650 | uxtah uv_7, texture_ptr, uv_67, ror #16 |
| | 2651 | ldrb pixel_3, [ uv_3 ] |
| | 2652 | |
| | 2653 | ldrb pixel_4, [ uv_4 ] |
| | 2654 | subs num_blocks, num_blocks, #1 |
| | 2655 | |
| | 2656 | ldrb pixel_5, [ uv_5 ] |
| | 2657 | orr pixels_a, pixel_0, pixel_1, lsl #8 |
| | 2658 | |
| | 2659 | ldrb pixel_6, [ uv_6 ] |
| | 2660 | orr pixels_b, pixel_4, pixel_5, lsl #8 |
| | 2661 | |
| | 2662 | ldrb pixel_7, [ uv_7 ] |
| | 2663 | orr pixels_a, pixels_a, pixel_2, lsl #16 |
| | 2664 | |
| | 2665 | orr pixels_b, pixels_b, pixel_6, lsl #16 |
| | 2666 | orr pixels_a, pixels_a, pixel_3, lsl #24 |
| | 2667 | |
| | 2668 | orr pixels_b, pixels_b, pixel_7, lsl #24 |
| | 2669 | vmov.u32 texels, pixels_a, pixels_b |
| | 2670 | |
| | 2671 | vtbl.8 texels_low, { clut_low_a, clut_low_b }, texels |
| | 2672 | vtbl.8 texels_high, { clut_high_a, clut_high_b }, texels |
| | 2673 | |
| | 2674 | vst2.u8 { texels_low, texels_high }, [ block_ptr, :128 ], c_64 |
| | 2675 | bne 0b |
| | 2676 | |
| | 2677 | ldmia sp!, { r3 - r11, pc } |
| | 2678 | |
| | 2679 | 1: |
| | 2680 | stmdb sp!, { r1 - r2 } |
| | 2681 | bl update_texture_4bpp_cache |
| | 2682 | |
| | 2683 | mov c_64, #64 |
| | 2684 | ldmia sp!, { r1 - r2 } |
| | 2685 | bal 0b |
| | 2686 | |
| | 2687 | |
| | 2688 | .align 3 |
| | 2689 | |
| | 2690 | function(texture_blocks_8bpp) |
| | 2691 | stmdb sp!, { r3 - r11, r14 } |
| | 2692 | add block_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 2693 | |
| | 2694 | ldr texture_ptr, [ psx_gpu, #psx_gpu_texture_page_ptr_offset ] |
| | 2695 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 2696 | |
| | 2697 | ldr clut_ptr, [ psx_gpu, #psx_gpu_clut_ptr_offset ] |
| | 2698 | ldr current_texture_mask, [ psx_gpu, #psx_gpu_current_texture_mask_offset ] |
| | 2699 | |
| | 2700 | ldr dirty_textures_mask, [ psx_gpu, #psx_gpu_dirty_textures_8bpp_mask_offset ] |
| | 2701 | tst dirty_textures_mask, current_texture_mask |
| | 2702 | |
| | 2703 | bne 1f |
| | 2704 | nop |
| | 2705 | |
| | 2706 | 0: |
| | 2707 | ldm block_ptr, { uv_01, uv_23, uv_45, uv_67 } |
| | 2708 | |
| | 2709 | uxtah uv_0, texture_ptr, uv_01 |
| | 2710 | uxtah uv_1, texture_ptr, uv_01, ror #16 |
| | 2711 | |
| | 2712 | uxtah uv_2, texture_ptr, uv_23 |
| | 2713 | uxtah uv_3, texture_ptr, uv_23, ror #16 |
| | 2714 | |
| | 2715 | uxtah uv_4, texture_ptr, uv_45 |
| | 2716 | ldrb pixel_0, [ uv_0 ] |
| | 2717 | |
| | 2718 | uxtah uv_5, texture_ptr, uv_45, ror #16 |
| | 2719 | ldrb pixel_1, [ uv_1 ] |
| | 2720 | |
| | 2721 | uxtah uv_6, texture_ptr, uv_67 |
| | 2722 | ldrb pixel_2, [ uv_2 ] |
| | 2723 | |
| | 2724 | uxtah uv_7, texture_ptr, uv_67, ror #16 |
| | 2725 | ldrb pixel_3, [ uv_3 ] |
| | 2726 | |
| | 2727 | ldrb pixel_4, [ uv_4 ] |
| | 2728 | add pixel_0, pixel_0, pixel_0 |
| | 2729 | |
| | 2730 | ldrb pixel_5, [ uv_5 ] |
| | 2731 | add pixel_1, pixel_1, pixel_1 |
| | 2732 | |
| | 2733 | ldrb pixel_6, [ uv_6 ] |
| | 2734 | add pixel_2, pixel_2, pixel_2 |
| | 2735 | |
| | 2736 | ldrb pixel_7, [ uv_7 ] |
| | 2737 | add pixel_3, pixel_3, pixel_3 |
| | 2738 | |
| | 2739 | ldrh pixel_0, [ clut_ptr, pixel_0 ] |
| | 2740 | add pixel_4, pixel_4, pixel_4 |
| | 2741 | |
| | 2742 | ldrh pixel_1, [ clut_ptr, pixel_1 ] |
| | 2743 | add pixel_5, pixel_5, pixel_5 |
| | 2744 | |
| | 2745 | ldrh pixel_2, [ clut_ptr, pixel_2 ] |
| | 2746 | add pixel_6, pixel_6, pixel_6 |
| | 2747 | |
| | 2748 | ldrh pixel_3, [ clut_ptr, pixel_3 ] |
| | 2749 | add pixel_7, pixel_7, pixel_7 |
| | 2750 | |
| | 2751 | ldrh pixel_4, [ clut_ptr, pixel_4 ] |
| | 2752 | orr pixels_a, pixel_0, pixel_1, lsl #16 |
| | 2753 | |
| | 2754 | ldrh pixel_5, [ clut_ptr, pixel_5 ] |
| | 2755 | orr pixels_c, pixel_2, pixel_3, lsl #16 |
| | 2756 | |
| | 2757 | ldrh pixel_6, [ clut_ptr, pixel_6 ] |
| | 2758 | subs num_blocks, num_blocks, #1 |
| | 2759 | |
| | 2760 | ldrh pixel_7, [ clut_ptr, pixel_7 ] |
| | 2761 | orr pixels_b, pixel_4, pixel_5, lsl #16 |
| | 2762 | |
| | 2763 | orr pixels_d, pixel_6, pixel_7, lsl #16 |
| | 2764 | stm block_ptr, { pixels_a, pixels_c, pixels_b, pixels_d } |
| | 2765 | |
| | 2766 | add block_ptr, block_ptr, #64 |
| | 2767 | bne 0b |
| | 2768 | |
| | 2769 | ldmia sp!, { r3 - r11, pc } |
| | 2770 | |
| | 2771 | 1: |
| | 2772 | stmdb sp!, { r1 - r2, r12 } |
| | 2773 | |
| | 2774 | bl update_texture_8bpp_cache |
| | 2775 | |
| | 2776 | ldmia sp!, { r1 - r2, r12 } |
| | 2777 | bal 0b |
| | 2778 | |
| | 2779 | |
| | 2780 | #undef uv_0 |
| | 2781 | #undef uv_1 |
| | 2782 | #undef uv_2 |
| | 2783 | #undef uv_3 |
| | 2784 | #undef uv_4 |
| | 2785 | #undef uv_5 |
| | 2786 | #undef uv_6 |
| | 2787 | #undef uv_7 |
| | 2788 | |
| | 2789 | #undef pixel_0 |
| | 2790 | #undef pixel_1 |
| | 2791 | #undef pixel_2 |
| | 2792 | #undef pixel_3 |
| | 2793 | #undef pixel_4 |
| | 2794 | #undef pixel_5 |
| | 2795 | #undef pixel_6 |
| | 2796 | #undef pixel_7 |
| | 2797 | |
| | 2798 | #undef texture_ptr |
| | 2799 | |
| | 2800 | #undef pixels_a |
| | 2801 | #undef pixels_b |
| | 2802 | #undef pixels_c |
| | 2803 | #undef pixels_d |
| | 2804 | |
| | 2805 | #define psx_gpu r0 |
| | 2806 | #define block_ptr r1 |
| | 2807 | #define num_blocks r2 |
| | 2808 | |
| | 2809 | #define uv_0 r3 |
| | 2810 | #define uv_1 r4 |
| | 2811 | #define u_0 r3 |
| | 2812 | #define u_1 r4 |
| | 2813 | #define v_0 r5 |
| | 2814 | #define v_1 r6 |
| | 2815 | |
| | 2816 | #define uv_2 r5 |
| | 2817 | #define uv_3 r6 |
| | 2818 | #define u_2 r5 |
| | 2819 | #define u_3 r6 |
| | 2820 | #define v_2 r7 |
| | 2821 | #define v_3 r8 |
| | 2822 | |
| | 2823 | #define uv_4 r7 |
| | 2824 | #define uv_5 r8 |
| | 2825 | #define u_4 r7 |
| | 2826 | #define u_5 r8 |
| | 2827 | #define v_4 r9 |
| | 2828 | #define v_5 r10 |
| | 2829 | |
| | 2830 | #define uv_6 r9 |
| | 2831 | #define uv_7 r10 |
| | 2832 | #define u_6 r9 |
| | 2833 | #define u_7 r10 |
| | 2834 | #define v_6 r11 |
| | 2835 | #define v_7 r0 |
| | 2836 | |
| | 2837 | #define pixel_0 r3 |
| | 2838 | #define pixel_1 r4 |
| | 2839 | #define pixel_2 r5 |
| | 2840 | #define pixel_3 r6 |
| | 2841 | #define pixel_4 r7 |
| | 2842 | #define pixel_5 r8 |
| | 2843 | #define pixel_6 r9 |
| | 2844 | #define pixel_7 r10 |
| | 2845 | |
| | 2846 | #define pixels_a r3 |
| | 2847 | #define pixels_b r5 |
| | 2848 | #define pixels_c r7 |
| | 2849 | #define pixels_d r9 |
| | 2850 | |
| | 2851 | #define texture_ptr r12 |
| | 2852 | |
| | 2853 | |
| | 2854 | .align 3 |
| | 2855 | |
| | 2856 | function(texture_blocks_16bpp) |
| | 2857 | stmdb sp!, { r3 - r11, r14 } |
| | 2858 | add block_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 2859 | |
| | 2860 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 2861 | ldr texture_ptr, [ psx_gpu, #psx_gpu_texture_page_ptr_offset ] |
| | 2862 | |
| | 2863 | 0: |
| | 2864 | ldrh uv_0, [ block_ptr ] |
| | 2865 | subs num_blocks, num_blocks, #1 |
| | 2866 | |
| | 2867 | ldrh uv_1, [ block_ptr, #2 ] |
| | 2868 | |
| | 2869 | and v_0, uv_0, #0xFF00 |
| | 2870 | and v_1, uv_1, #0xFF00 |
| | 2871 | |
| | 2872 | and u_0, uv_0, #0xFF |
| | 2873 | and u_1, uv_1, #0xFF |
| | 2874 | |
| | 2875 | add uv_0, u_0, v_0, lsl #2 |
| | 2876 | ldrh uv_2, [ block_ptr, #4 ] |
| | 2877 | |
| | 2878 | add uv_1, u_1, v_1, lsl #2 |
| | 2879 | ldrh uv_3, [ block_ptr, #6 ] |
| | 2880 | |
| | 2881 | add uv_0, uv_0, uv_0 |
| | 2882 | add uv_1, uv_1, uv_1 |
| | 2883 | |
| | 2884 | and v_2, uv_2, #0xFF00 |
| | 2885 | and v_3, uv_3, #0xFF00 |
| | 2886 | |
| | 2887 | and u_2, uv_2, #0xFF |
| | 2888 | and u_3, uv_3, #0xFF |
| | 2889 | |
| | 2890 | add uv_2, u_2, v_2, lsl #2 |
| | 2891 | ldrh uv_4, [ block_ptr, #8 ] |
| | 2892 | |
| | 2893 | add uv_3, u_3, v_3, lsl #2 |
| | 2894 | ldrh uv_5, [ block_ptr, #10 ] |
| | 2895 | |
| | 2896 | add uv_2, uv_2, uv_2 |
| | 2897 | add uv_3, uv_3, uv_3 |
| | 2898 | |
| | 2899 | and v_4, uv_4, #0xFF00 |
| | 2900 | and v_5, uv_5, #0xFF00 |
| | 2901 | |
| | 2902 | and u_4, uv_4, #0xFF |
| | 2903 | and u_5, uv_5, #0xFF |
| | 2904 | |
| | 2905 | add uv_4, u_4, v_4, lsl #2 |
| | 2906 | ldrh uv_6, [ block_ptr, #12 ] |
| | 2907 | |
| | 2908 | add uv_5, u_5, v_5, lsl #2 |
| | 2909 | ldrh uv_7, [ block_ptr, #14 ] |
| | 2910 | |
| | 2911 | add uv_4, uv_4, uv_4 |
| | 2912 | ldrh pixel_0, [ texture_ptr, uv_0 ] |
| | 2913 | |
| | 2914 | add uv_5, uv_5, uv_5 |
| | 2915 | ldrh pixel_1, [ texture_ptr, uv_1 ] |
| | 2916 | |
| | 2917 | and v_6, uv_6, #0xFF00 |
| | 2918 | ldrh pixel_2, [ texture_ptr, uv_2 ] |
| | 2919 | |
| | 2920 | and v_7, uv_7, #0xFF00 |
| | 2921 | ldrh pixel_3, [ texture_ptr, uv_3 ] |
| | 2922 | |
| | 2923 | and u_6, uv_6, #0xFF |
| | 2924 | ldrh pixel_4, [ texture_ptr, uv_4 ] |
| | 2925 | |
| | 2926 | and u_7, uv_7, #0xFF |
| | 2927 | ldrh pixel_5, [ texture_ptr, uv_5 ] |
| | 2928 | |
| | 2929 | add uv_6, u_6, v_6, lsl #2 |
| | 2930 | add uv_7, u_7, v_7, lsl #2 |
| | 2931 | |
| | 2932 | add uv_6, uv_6, uv_6 |
| | 2933 | add uv_7, uv_7, uv_7 |
| | 2934 | |
| | 2935 | orr pixels_a, pixel_0, pixel_1, lsl #16 |
| | 2936 | orr pixels_b, pixel_2, pixel_3, lsl #16 |
| | 2937 | |
| | 2938 | ldrh pixel_6, [ texture_ptr, uv_6 ] |
| | 2939 | orr pixels_c, pixel_4, pixel_5, lsl #16 |
| | 2940 | |
| | 2941 | ldrh pixel_7, [ texture_ptr, uv_7 ] |
| | 2942 | orr pixels_d, pixel_6, pixel_7, lsl #16 |
| | 2943 | |
| | 2944 | stm block_ptr, { pixels_a, pixels_b, pixels_c, pixels_d } |
| | 2945 | add block_ptr, block_ptr, #64 |
| | 2946 | |
| | 2947 | bne 0b |
| | 2948 | |
| | 2949 | ldmia sp!, { r3 - r11, pc } |
| | 2950 | |
| | 2951 | |
| | 2952 | #undef num_blocks |
| | 2953 | |
| | 2954 | #undef test_mask |
| | 2955 | #undef texels |
| | 2956 | #undef pixels_b |
| | 2957 | #undef pixels |
| | 2958 | #undef d64_1 |
| | 2959 | #undef d64_4 |
| | 2960 | #undef d64_128 |
| | 2961 | #undef draw_mask |
| | 2962 | #undef msb_mask |
| | 2963 | #undef msb_mask_low |
| | 2964 | #undef msb_mask_high |
| | 2965 | #undef fb_pixels |
| | 2966 | |
| | 2967 | #undef c_32 |
| | 2968 | #undef fb_ptr |
| | 2969 | #undef mask_msb_ptr |
| | 2970 | |
| | 2971 | #define psx_gpu r0 |
| | 2972 | #define num_blocks r1 |
| | 2973 | #define color_ptr r2 |
| | 2974 | #define colors_scalar r2 |
| | 2975 | #define colors_scalar_compare r3 |
| | 2976 | #define mask_msb_ptr r2 |
| | 2977 | |
| | 2978 | #define block_ptr_load_a r0 |
| | 2979 | #define block_ptr_store r3 |
| | 2980 | #define block_ptr_load_b r12 |
| | 2981 | #define c_32 r2 |
| | 2982 | |
| | 2983 | #define c_48 r4 |
| | 2984 | #define fb_ptr r14 |
| | 2985 | #define draw_mask_bits_scalar r5 |
| | 2986 | |
| | 2987 | #define d128_0x07 q0 |
| | 2988 | #define d128_0x1F q1 |
| | 2989 | #define d128_0x8000 q2 |
| | 2990 | #define test_mask q3 |
| | 2991 | #define texels q4 |
| | 2992 | #define colors_rg q5 |
| | 2993 | #define colors_b_dm_bits q6 |
| | 2994 | #define texels_rg q7 |
| | 2995 | #define pixels_r q8 |
| | 2996 | #define pixels_g q9 |
| | 2997 | #define pixels_b q10 |
| | 2998 | #define pixels q11 |
| | 2999 | #define zero_mask q4 |
| | 3000 | #define draw_mask q12 |
| | 3001 | #define msb_mask q13 |
| | 3002 | |
| | 3003 | #define fb_pixels q8 |
| | 3004 | |
| | 3005 | #define pixels_gb_low q9 |
| | 3006 | |
| | 3007 | #define colors_r d10 |
| | 3008 | #define colors_g d11 |
| | 3009 | #define colors_b d12 |
| | 3010 | #define draw_mask_bits d13 |
| | 3011 | #define texels_r d14 |
| | 3012 | #define texels_g d15 |
| | 3013 | #define pixels_r_low d16 |
| | 3014 | #define pixels_g_low d18 |
| | 3015 | #define pixels_b_low d19 |
| | 3016 | #define msb_mask_low d26 |
| | 3017 | #define msb_mask_high d27 |
| | 3018 | |
| | 3019 | #define d64_1 d28 |
| | 3020 | #define d64_4 d29 |
| | 3021 | #define d64_128 d30 |
| | 3022 | #define texels_b d31 |
| | 3023 | |
| | 3024 | #define shade_blocks_textured_modulated_prologue_indirect() \ |
| | 3025 | mov c_48, #48; \ |
| | 3026 | add block_ptr_store, psx_gpu, #psx_gpu_blocks_offset \ |
| | 3027 | |
| | 3028 | #define shade_blocks_textured_modulated_prologue_direct() \ |
| | 3029 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 3030 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ] \ |
| | 3031 | |
| | 3032 | |
| | 3033 | #define shade_blocks_textured_modulated_prologue_shaded(dithering, target) \ |
| | 3034 | |
| | 3035 | #define shade_blocks_textured_false_modulation_check_undithered(target) \ |
| | 3036 | ldr colors_scalar, [ psx_gpu, #psx_gpu_triangle_color_offset ]; \ |
| | 3037 | movw colors_scalar_compare, #0x8080; \ |
| | 3038 | \ |
| | 3039 | movt colors_scalar_compare, #0x80; \ |
| | 3040 | cmp colors_scalar, colors_scalar_compare; \ |
| | 3041 | beq shade_blocks_textured_unmodulated_##target \ |
| | 3042 | |
| | 3043 | #define shade_blocks_textured_false_modulation_check_dithered(target) \ |
| | 3044 | |
| | 3045 | #define shade_blocks_textured_modulated_prologue_unshaded(dithering, target) \ |
| | 3046 | shade_blocks_textured_false_modulation_check_##dithering(target); \ |
| | 3047 | add color_ptr, psx_gpu, #psx_gpu_triangle_color_offset; \ |
| | 3048 | vld1.u32 { colors_r[] }, [ color_ptr, :32 ]; \ |
| | 3049 | vdup.u8 colors_g, colors_r[1]; \ |
| | 3050 | vdup.u8 colors_b, colors_r[2]; \ |
| | 3051 | vdup.u8 colors_r, colors_r[0] \ |
| | 3052 | |
| | 3053 | |
| | 3054 | #define shade_blocks_textured_modulated_load_dithered(target) \ |
| | 3055 | vld1.u32 { target }, [ block_ptr_load_b, :128 ] \ |
| | 3056 | |
| | 3057 | #define shade_blocks_textured_modulated_load_last_dithered(target) \ |
| | 3058 | vld1.u32 { target }, [ block_ptr_load_b, :128 ], c_32 \ |
| | 3059 | |
| | 3060 | #define shade_blocks_textured_modulated_load_undithered(target) \ |
| | 3061 | |
| | 3062 | #define shade_blocks_textured_modulated_load_last_undithered(target) \ |
| | 3063 | add block_ptr_load_b, block_ptr_load_b, #32 \ |
| | 3064 | |
| | 3065 | #define shade_blocks_textured_modulate_dithered(channel) \ |
| | 3066 | vmlal.u8 pixels_##channel, texels_##channel, colors_##channel \ |
| | 3067 | |
| | 3068 | #define shade_blocks_textured_modulate_undithered(channel) \ |
| | 3069 | vmull.u8 pixels_##channel, texels_##channel, colors_##channel \ |
| | 3070 | |
| | 3071 | |
| | 3072 | #define shade_blocks_textured_modulated_store_draw_mask_indirect(offset) \ |
| | 3073 | vst1.u32 { draw_mask }, [ block_ptr_store, :128 ]! \ |
| | 3074 | |
| | 3075 | #define shade_blocks_textured_modulated_store_draw_mask_direct(offset) \ |
| | 3076 | ldr fb_ptr, [ block_ptr_load_b, #(offset - 64) ]; \ |
| | 3077 | vld1.u32 { fb_pixels }, [ fb_ptr ]; \ |
| | 3078 | vbit.u16 pixels, fb_pixels, draw_mask \ |
| | 3079 | |
| | 3080 | #define shade_blocks_textured_modulated_store_pixels_indirect() \ |
| | 3081 | vst1.u32 { pixels }, [ block_ptr_store, :128 ], c_48 \ |
| | 3082 | |
| | 3083 | #define shade_blocks_textured_modulated_store_pixels_direct() \ |
| | 3084 | vst1.u32 { pixels }, [ fb_ptr ] \ |
| | 3085 | |
| | 3086 | |
| | 3087 | #define shade_blocks_textured_modulated_load_rg_shaded() \ |
| | 3088 | vld1.u32 { colors_r, colors_g }, [ block_ptr_load_b, :128 ], c_32 \ |
| | 3089 | |
| | 3090 | #define shade_blocks_textured_modulated_load_rg_unshaded() \ |
| | 3091 | add block_ptr_load_b, block_ptr_load_b, #32 \ |
| | 3092 | |
| | 3093 | #define shade_blocks_textured_modulated_load_bdm_shaded() \ |
| | 3094 | vld1.u32 { colors_b, draw_mask_bits }, [ block_ptr_load_a, :128 ], c_32 \ |
| | 3095 | |
| | 3096 | #define shade_blocks_textured_modulated_load_bdm_unshaded() \ |
| | 3097 | ldr draw_mask_bits_scalar, [ block_ptr_load_a, #8 ]; \ |
| | 3098 | add block_ptr_load_a, block_ptr_load_a, #32 \ |
| | 3099 | |
| | 3100 | #define shade_blocks_textured_modulated_expand_draw_mask_shaded() \ |
| | 3101 | vdup.u16 draw_mask, draw_mask_bits[0] \ |
| | 3102 | |
| | 3103 | #define shade_blocks_textured_modulated_expand_draw_mask_unshaded() \ |
| | 3104 | vdup.u16 draw_mask, draw_mask_bits_scalar \ |
| | 3105 | |
| | 3106 | |
| | 3107 | #define shade_blocks_textured_modulated_apply_msb_mask_indirect() \ |
| | 3108 | |
| | 3109 | #define shade_blocks_textured_modulated_apply_msb_mask_direct() \ |
| | 3110 | vorr.u16 pixels, pixels, msb_mask \ |
| | 3111 | |
| | 3112 | |
| | 3113 | #define shade_blocks_textured_modulated_builder(shading, dithering, target) \ |
| | 3114 | .align 3; \ |
| | 3115 | \ |
| | 3116 | function(shade_blocks_##shading##_textured_modulated_##dithering##_##target) \ |
| | 3117 | shade_blocks_textured_modulated_prologue_##shading(dithering, target); \ |
| | 3118 | stmdb sp!, { r4 - r5, lr }; \ |
| | 3119 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 3120 | \ |
| | 3121 | vld1.u32 { test_mask }, [ psx_gpu, :128 ]; \ |
| | 3122 | \ |
| | 3123 | shade_blocks_textured_modulated_prologue_##target(); \ |
| | 3124 | \ |
| | 3125 | add block_ptr_load_a, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 3126 | mov c_32, #32; \ |
| | 3127 | \ |
| | 3128 | add block_ptr_load_b, block_ptr_load_a, #16; \ |
| | 3129 | vmov.u8 d64_1, #1; \ |
| | 3130 | vmov.u8 d64_4, #4; \ |
| | 3131 | vmov.u8 d64_128, #128; \ |
| | 3132 | \ |
| | 3133 | vld1.u32 { texels }, [ block_ptr_load_a, :128 ], c_32; \ |
| | 3134 | vmov.u8 d128_0x07, #0x07; \ |
| | 3135 | \ |
| | 3136 | shade_blocks_textured_modulated_load_rg_##shading(); \ |
| | 3137 | vmov.u8 d128_0x1F, #0x1F; \ |
| | 3138 | \ |
| | 3139 | shade_blocks_textured_modulated_load_bdm_##shading(); \ |
| | 3140 | vmov.u16 d128_0x8000, #0x8000; \ |
| | 3141 | \ |
| | 3142 | vmovn.u16 texels_r, texels; \ |
| | 3143 | vshrn.u16 texels_g, texels, #5; \ |
| | 3144 | \ |
| | 3145 | vshrn.u16 texels_b, texels, #7; \ |
| | 3146 | shade_blocks_textured_modulated_expand_draw_mask_##shading(); \ |
| | 3147 | \ |
| | 3148 | shade_blocks_textured_modulated_load_##dithering(pixels_r); \ |
| | 3149 | vtst.u16 draw_mask, draw_mask, test_mask; \ |
| | 3150 | \ |
| | 3151 | shade_blocks_textured_modulated_load_##dithering(pixels_g); \ |
| | 3152 | vand.u8 texels_rg, texels_rg, d128_0x1F; \ |
| | 3153 | \ |
| | 3154 | shade_blocks_textured_modulated_load_last_##dithering(pixels_b); \ |
| | 3155 | vshr.u8 texels_b, texels_b, #3; \ |
| | 3156 | \ |
| | 3157 | shade_blocks_textured_modulate_##dithering(r); \ |
| | 3158 | shade_blocks_textured_modulate_##dithering(g); \ |
| | 3159 | shade_blocks_textured_modulate_##dithering(b); \ |
| | 3160 | \ |
| | 3161 | vand.u16 pixels, texels, d128_0x8000; \ |
| | 3162 | vceq.u16 zero_mask, texels, #0; \ |
| | 3163 | \ |
| | 3164 | vqshrun.s16 pixels_r_low, pixels_r, #4; \ |
| | 3165 | vqshrun.s16 pixels_g_low, pixels_g, #4; \ |
| | 3166 | vqshrun.s16 pixels_b_low, pixels_b, #4; \ |
| | 3167 | \ |
| | 3168 | shade_blocks_textured_modulated_apply_msb_mask_##target(); \ |
| | 3169 | vorr.u16 draw_mask, draw_mask, zero_mask; \ |
| | 3170 | vshr.u8 pixels_r_low, pixels_r_low, #3; \ |
| | 3171 | vbic.u8 pixels_gb_low, pixels_gb_low, d128_0x07; \ |
| | 3172 | \ |
| | 3173 | subs num_blocks, num_blocks, #1; \ |
| | 3174 | beq 1f; \ |
| | 3175 | \ |
| | 3176 | .align 3; \ |
| | 3177 | \ |
| | 3178 | 0: \ |
| | 3179 | vld1.u32 { texels }, [ block_ptr_load_a, :128 ], c_32; \ |
| | 3180 | shade_blocks_textured_modulated_load_rg_##shading(); \ |
| | 3181 | vshrn.u16 texels_g, texels, #5; \ |
| | 3182 | \ |
| | 3183 | shade_blocks_textured_modulated_load_bdm_##shading(); \ |
| | 3184 | vshrn.u16 texels_b, texels, #7; \ |
| | 3185 | \ |
| | 3186 | vmovn.u16 texels_r, texels; \ |
| | 3187 | vmlal.u8 pixels, pixels_r_low, d64_1; \ |
| | 3188 | \ |
| | 3189 | vmlal.u8 pixels, pixels_g_low, d64_4; \ |
| | 3190 | vmlal.u8 pixels, pixels_b_low, d64_128; \ |
| | 3191 | shade_blocks_textured_modulated_store_draw_mask_##target(-4); \ |
| | 3192 | \ |
| | 3193 | shade_blocks_textured_modulated_load_##dithering(pixels_r); \ |
| | 3194 | shade_blocks_textured_modulated_expand_draw_mask_##shading(); \ |
| | 3195 | \ |
| | 3196 | shade_blocks_textured_modulated_load_##dithering(pixels_g); \ |
| | 3197 | vand.u8 texels_rg, texels_rg, d128_0x1F; \ |
| | 3198 | \ |
| | 3199 | shade_blocks_textured_modulated_load_last_##dithering(pixels_b); \ |
| | 3200 | vtst.u16 draw_mask, draw_mask, test_mask; \ |
| | 3201 | \ |
| | 3202 | shade_blocks_textured_modulated_store_pixels_##target(); \ |
| | 3203 | vshr.u8 texels_b, texels_b, #3; \ |
| | 3204 | \ |
| | 3205 | shade_blocks_textured_modulate_##dithering(r); \ |
| | 3206 | shade_blocks_textured_modulate_##dithering(g); \ |
| | 3207 | shade_blocks_textured_modulate_##dithering(b); \ |
| | 3208 | \ |
| | 3209 | vand.u16 pixels, texels, d128_0x8000; \ |
| | 3210 | vceq.u16 zero_mask, texels, #0; \ |
| | 3211 | \ |
| | 3212 | subs num_blocks, num_blocks, #1; \ |
| | 3213 | \ |
| | 3214 | vqshrun.s16 pixels_r_low, pixels_r, #4; \ |
| | 3215 | vqshrun.s16 pixels_g_low, pixels_g, #4; \ |
| | 3216 | vqshrun.s16 pixels_b_low, pixels_b, #4; \ |
| | 3217 | \ |
| | 3218 | shade_blocks_textured_modulated_apply_msb_mask_##target(); \ |
| | 3219 | vorr.u16 draw_mask, draw_mask, zero_mask; \ |
| | 3220 | vshr.u8 pixels_r_low, pixels_r_low, #3; \ |
| | 3221 | vbic.u8 pixels_gb_low, pixels_gb_low, d128_0x07; \ |
| | 3222 | \ |
| | 3223 | bne 0b; \ |
| | 3224 | \ |
| | 3225 | 1: \ |
| | 3226 | vmlal.u8 pixels, pixels_r_low, d64_1; \ |
| | 3227 | vmlal.u8 pixels, pixels_g_low, d64_4; \ |
| | 3228 | vmlal.u8 pixels, pixels_b_low, d64_128; \ |
| | 3229 | \ |
| | 3230 | shade_blocks_textured_modulated_store_draw_mask_##target(28); \ |
| | 3231 | shade_blocks_textured_modulated_store_pixels_##target(); \ |
| | 3232 | \ |
| | 3233 | ldmia sp!, { r4 - r5, pc } \ |
| | 3234 | |
| | 3235 | |
| | 3236 | shade_blocks_textured_modulated_builder(shaded, dithered, direct); |
| | 3237 | shade_blocks_textured_modulated_builder(shaded, undithered, direct); |
| | 3238 | shade_blocks_textured_modulated_builder(unshaded, dithered, direct); |
| | 3239 | shade_blocks_textured_modulated_builder(unshaded, undithered, direct); |
| | 3240 | |
| | 3241 | shade_blocks_textured_modulated_builder(shaded, dithered, indirect); |
| | 3242 | shade_blocks_textured_modulated_builder(shaded, undithered, indirect); |
| | 3243 | shade_blocks_textured_modulated_builder(unshaded, dithered, indirect); |
| | 3244 | shade_blocks_textured_modulated_builder(unshaded, undithered, indirect); |
| | 3245 | |
| | 3246 | |
| | 3247 | #undef c_64 |
| | 3248 | #undef fb_ptr |
| | 3249 | #undef color_ptr |
| | 3250 | |
| | 3251 | #undef color_r |
| | 3252 | #undef color_g |
| | 3253 | #undef color_b |
| | 3254 | |
| | 3255 | #undef test_mask |
| | 3256 | #undef pixels |
| | 3257 | #undef draw_mask |
| | 3258 | #undef zero_mask |
| | 3259 | #undef fb_pixels |
| | 3260 | #undef msb_mask |
| | 3261 | #undef msb_mask_low |
| | 3262 | #undef msb_mask_high |
| | 3263 | |
| | 3264 | #define psx_gpu r0 |
| | 3265 | #define num_blocks r1 |
| | 3266 | #define mask_msb_ptr r2 |
| | 3267 | #define color_ptr r3 |
| | 3268 | |
| | 3269 | #define block_ptr_load r0 |
| | 3270 | #define draw_mask_store_ptr r3 |
| | 3271 | #define draw_mask_bits_ptr r12 |
| | 3272 | #define draw_mask_ptr r12 |
| | 3273 | #define pixel_store_ptr r14 |
| | 3274 | |
| | 3275 | #define fb_ptr_cmp r4 |
| | 3276 | |
| | 3277 | #define fb_ptr r3 |
| | 3278 | #define fb_ptr_next r14 |
| | 3279 | |
| | 3280 | #define c_64 r2 |
| | 3281 | |
| | 3282 | #define test_mask q0 |
| | 3283 | #define pixels q1 |
| | 3284 | #define draw_mask q2 |
| | 3285 | #define zero_mask q3 |
| | 3286 | #define draw_mask_combined q4 |
| | 3287 | #define fb_pixels q5 |
| | 3288 | #define fb_pixels_next q6 |
| | 3289 | #define msb_mask q7 |
| | 3290 | |
| | 3291 | #define draw_mask_low d4 |
| | 3292 | #define draw_mask_high d5 |
| | 3293 | #define msb_mask_low d14 |
| | 3294 | #define msb_mask_high d15 |
| | 3295 | |
| | 3296 | .align 3 |
| | 3297 | function(shade_blocks_textured_unmodulated_indirect) |
| | 3298 | str r14, [ sp, #-4 ] |
| | 3299 | add draw_mask_bits_ptr, psx_gpu, #(psx_gpu_blocks_offset + 40) |
| | 3300 | |
| | 3301 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 3302 | add pixel_store_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16) |
| | 3303 | |
| | 3304 | vld1.u32 { test_mask }, [ psx_gpu, :128 ] |
| | 3305 | add draw_mask_store_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 3306 | |
| | 3307 | mov c_64, #64 |
| | 3308 | add block_ptr_load, psx_gpu, #psx_gpu_blocks_offset |
| | 3309 | |
| | 3310 | vld1.u32 { pixels }, [ block_ptr_load, :128 ], c_64 |
| | 3311 | vld1.u16 { draw_mask_low[], draw_mask_high[] }, \ |
| | 3312 | [ draw_mask_bits_ptr, :16 ], c_64 |
| | 3313 | vceq.u16 zero_mask, pixels, #0 |
| | 3314 | |
| | 3315 | vtst.u16 draw_mask, draw_mask, test_mask |
| | 3316 | vst1.u32 { pixels }, [ pixel_store_ptr, :128 ], c_64 |
| | 3317 | |
| | 3318 | subs num_blocks, num_blocks, #1 |
| | 3319 | beq 1f |
| | 3320 | |
| | 3321 | 0: |
| | 3322 | vld1.u32 { pixels }, [ block_ptr_load, :128 ], c_64 |
| | 3323 | vorr.u16 draw_mask_combined, draw_mask, zero_mask |
| | 3324 | |
| | 3325 | vld1.u16 { draw_mask_low[], draw_mask_high[] }, \ |
| | 3326 | [ draw_mask_bits_ptr, :16 ], c_64 |
| | 3327 | vceq.u16 zero_mask, pixels, #0 |
| | 3328 | |
| | 3329 | vtst.u16 draw_mask, draw_mask, test_mask |
| | 3330 | vst1.u32 { pixels }, [ pixel_store_ptr, :128 ], c_64 |
| | 3331 | |
| | 3332 | vst1.u32 { draw_mask_combined }, [ draw_mask_store_ptr, :128 ], c_64 |
| | 3333 | subs num_blocks, num_blocks, #1 |
| | 3334 | |
| | 3335 | bne 0b |
| | 3336 | |
| | 3337 | 1: |
| | 3338 | vorr.u16 draw_mask_combined, draw_mask, zero_mask |
| | 3339 | vst1.u32 { draw_mask_combined }, [ draw_mask_store_ptr, :128 ], c_64 |
| | 3340 | |
| | 3341 | ldr pc, [ sp, #-4 ] |
| | 3342 | |
| | 3343 | |
| | 3344 | .align 3 |
| | 3345 | |
| | 3346 | function(shade_blocks_textured_unmodulated_direct) |
| | 3347 | stmdb sp!, { r4, r14 } |
| | 3348 | add draw_mask_bits_ptr, psx_gpu, #(psx_gpu_blocks_offset + 40) |
| | 3349 | |
| | 3350 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 3351 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset |
| | 3352 | |
| | 3353 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ] |
| | 3354 | mov c_64, #64 |
| | 3355 | |
| | 3356 | vld1.u32 { test_mask }, [ psx_gpu, :128 ] |
| | 3357 | add block_ptr_load, psx_gpu, #psx_gpu_blocks_offset |
| | 3358 | |
| | 3359 | vld1.u16 { draw_mask_low[], draw_mask_high[] }, \ |
| | 3360 | [ draw_mask_bits_ptr, :16 ], c_64 |
| | 3361 | ldr fb_ptr_next, [ block_ptr_load, #44 ] |
| | 3362 | |
| | 3363 | vld1.u32 { pixels }, [ block_ptr_load, :128 ], c_64 |
| | 3364 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3365 | vceq.u16 zero_mask, pixels, #0 |
| | 3366 | vtst.u16 draw_mask, draw_mask, test_mask |
| | 3367 | |
| | 3368 | subs num_blocks, num_blocks, #1 |
| | 3369 | beq 1f |
| | 3370 | |
| | 3371 | 0: |
| | 3372 | mov fb_ptr, fb_ptr_next |
| | 3373 | ldr fb_ptr_next, [ block_ptr_load, #44 ] |
| | 3374 | |
| | 3375 | vorr.u16 pixels, pixels, msb_mask |
| | 3376 | |
| | 3377 | vorr.u16 draw_mask_combined, draw_mask, zero_mask |
| | 3378 | vmov fb_pixels, fb_pixels_next |
| | 3379 | |
| | 3380 | vld1.u16 { draw_mask_low[], draw_mask_high[] }, \ |
| | 3381 | [ draw_mask_bits_ptr, :16 ], c_64 |
| | 3382 | vbif.u16 fb_pixels, pixels, draw_mask_combined |
| | 3383 | |
| | 3384 | vld1.u32 { pixels }, [ block_ptr_load, :128 ], c_64 |
| | 3385 | |
| | 3386 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr |
| | 3387 | add fb_ptr_cmp, fb_ptr_cmp, #14 |
| | 3388 | cmp fb_ptr_cmp, #28 |
| | 3389 | bls 4f |
| | 3390 | |
| | 3391 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3392 | vceq.u16 zero_mask, pixels, #0 |
| | 3393 | |
| | 3394 | vst1.u16 { fb_pixels }, [ fb_ptr ] |
| | 3395 | vtst.u16 draw_mask, draw_mask, test_mask |
| | 3396 | |
| | 3397 | 3: |
| | 3398 | subs num_blocks, num_blocks, #1 |
| | 3399 | bne 0b |
| | 3400 | |
| | 3401 | 1: |
| | 3402 | vorr.u16 draw_mask_combined, draw_mask, zero_mask |
| | 3403 | vbif.u16 fb_pixels_next, pixels, draw_mask_combined |
| | 3404 | |
| | 3405 | vst1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3406 | |
| | 3407 | ldmia sp!, { r4, pc } |
| | 3408 | |
| | 3409 | 4: |
| | 3410 | vst1.u16 { fb_pixels }, [ fb_ptr ] |
| | 3411 | vceq.u16 zero_mask, pixels, #0 |
| | 3412 | |
| | 3413 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3414 | vtst.u16 draw_mask, draw_mask, test_mask |
| | 3415 | |
| | 3416 | bal 3b |
| | 3417 | |
| | 3418 | |
| | 3419 | function(shade_blocks_unshaded_untextured_indirect) |
| | 3420 | bx lr |
| | 3421 | |
| | 3422 | .align 3 |
| | 3423 | |
| | 3424 | function(shade_blocks_unshaded_untextured_direct) |
| | 3425 | stmdb sp!, { r4, r14 } |
| | 3426 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 3427 | |
| | 3428 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 3429 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset |
| | 3430 | |
| | 3431 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ] |
| | 3432 | add color_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16) |
| | 3433 | |
| | 3434 | add block_ptr_load, psx_gpu, #(psx_gpu_blocks_offset + 44) |
| | 3435 | vld1.u16 { pixels }, [ color_ptr, :128 ] |
| | 3436 | |
| | 3437 | mov c_64, #64 |
| | 3438 | vld1.u16 { draw_mask }, [ draw_mask_ptr, :128 ], c_64 |
| | 3439 | |
| | 3440 | vorr.u16 pixels, pixels, msb_mask |
| | 3441 | subs num_blocks, num_blocks, #1 |
| | 3442 | |
| | 3443 | ldr fb_ptr_next, [ block_ptr_load ], #64 |
| | 3444 | |
| | 3445 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3446 | beq 1f |
| | 3447 | |
| | 3448 | 0: |
| | 3449 | vmov fb_pixels, fb_pixels_next |
| | 3450 | mov fb_ptr, fb_ptr_next |
| | 3451 | ldr fb_ptr_next, [ block_ptr_load ], #64 |
| | 3452 | |
| | 3453 | vbif.u16 fb_pixels, pixels, draw_mask |
| | 3454 | vld1.u16 { draw_mask }, [ draw_mask_ptr, :128 ], c_64 |
| | 3455 | |
| | 3456 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr |
| | 3457 | add fb_ptr_cmp, fb_ptr_cmp, #14 |
| | 3458 | cmp fb_ptr_cmp, #28 |
| | 3459 | bls 4f |
| | 3460 | |
| | 3461 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3462 | vst1.u16 { fb_pixels }, [ fb_ptr ] |
| | 3463 | |
| | 3464 | 3: |
| | 3465 | subs num_blocks, num_blocks, #1 |
| | 3466 | bne 0b |
| | 3467 | |
| | 3468 | 1: |
| | 3469 | vbif.u16 fb_pixels_next, pixels, draw_mask |
| | 3470 | vst1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3471 | |
| | 3472 | ldmia sp!, { r4, pc } |
| | 3473 | |
| | 3474 | 4: |
| | 3475 | vst1.u16 { fb_pixels }, [ fb_ptr ] |
| | 3476 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ] |
| | 3477 | bal 3b |
| | 3478 | |
| | 3479 | |
| | 3480 | #undef draw_mask_ptr |
| | 3481 | #undef c_64 |
| | 3482 | #undef fb_ptr |
| | 3483 | #undef fb_ptr_next |
| | 3484 | #undef fb_ptr_cmp |
| | 3485 | |
| | 3486 | #define psx_gpu r0 |
| | 3487 | #define num_blocks r1 |
| | 3488 | #define msb_mask_ptr r2 |
| | 3489 | #define pixel_ptr r3 |
| | 3490 | #define draw_mask_ptr r0 |
| | 3491 | #define c_64 r2 |
| | 3492 | #define fb_ptr r12 |
| | 3493 | #define fb_ptr_next r14 |
| | 3494 | #define fb_ptr_cmp r4 |
| | 3495 | |
| | 3496 | #undef msb_mask |
| | 3497 | #undef draw_mask |
| | 3498 | #undef pixels |
| | 3499 | #undef fb_pixels |
| | 3500 | #undef d128_0x8000 |
| | 3501 | #undef msb_mask_low |
| | 3502 | #undef msb_mask_high |
| | 3503 | #undef draw_mask_next |
| | 3504 | #undef pixels_g |
| | 3505 | #undef blend_pixels |
| | 3506 | #undef fb_pixels_next |
| | 3507 | |
| | 3508 | #define msb_mask q0 |
| | 3509 | #define draw_mask q1 |
| | 3510 | #define pixels q2 |
| | 3511 | #define fb_pixels q3 |
| | 3512 | #define blend_pixels q4 |
| | 3513 | #define pixels_no_msb q5 |
| | 3514 | #define blend_mask q6 |
| | 3515 | #define fb_pixels_no_msb q7 |
| | 3516 | #define d128_0x8000 q8 |
| | 3517 | #define d128_0x0421 q9 |
| | 3518 | #define fb_pixels_next q10 |
| | 3519 | #define blend_pixels_next q11 |
| | 3520 | #define pixels_next q12 |
| | 3521 | #define draw_mask_next q13 |
| | 3522 | #define write_mask q14 |
| | 3523 | |
| | 3524 | #define pixels_rb q5 |
| | 3525 | #define pixels_mg q7 |
| | 3526 | #define pixels_g q7 |
| | 3527 | #define d128_0x7C1F q8 |
| | 3528 | #define d128_0x03E0 q9 |
| | 3529 | #define fb_pixels_rb q10 |
| | 3530 | #define fb_pixels_g q11 |
| | 3531 | #define fb_pixels_masked q11 |
| | 3532 | #define d128_0x83E0 q15 |
| | 3533 | #define pixels_fourth q7 |
| | 3534 | #define d128_0x1C07 q12 |
| | 3535 | #define d128_0x00E0 q13 |
| | 3536 | #define d128_0x80E0 q13 |
| | 3537 | |
| | 3538 | #define msb_mask_low d0 |
| | 3539 | #define msb_mask_high d1 |
| | 3540 | |
| | 3541 | #define blend_blocks_average_set_blend_mask_textured(source) \ |
| | 3542 | vclt.s16 blend_mask, source, #0 \ |
| | 3543 | |
| | 3544 | #define blend_blocks_average_set_stp_bit_textured() \ |
| | 3545 | vorr.u16 blend_pixels, #0x8000 \ |
| | 3546 | |
| | 3547 | #define blend_blocks_average_combine_textured(source) \ |
| | 3548 | vbif.u16 blend_pixels, source, blend_mask \ |
| | 3549 | |
| | 3550 | #define blend_blocks_average_set_blend_mask_untextured(source) \ |
| | 3551 | |
| | 3552 | #define blend_blocks_average_set_stp_bit_untextured() \ |
| | 3553 | |
| | 3554 | #define blend_blocks_average_combine_untextured(source) \ |
| | 3555 | |
| | 3556 | #define blend_blocks_average_mask_set_on() \ |
| | 3557 | vclt.s16 write_mask, fb_pixels_next, #0 \ |
| | 3558 | |
| | 3559 | #define blend_blocks_average_mask_copy_on() \ |
| | 3560 | vorr.u16 draw_mask, draw_mask_next, write_mask \ |
| | 3561 | |
| | 3562 | #define blend_blocks_average_mask_copy_b_on() \ |
| | 3563 | vorr.u16 draw_mask_next, draw_mask_next, write_mask \ |
| | 3564 | |
| | 3565 | #define blend_blocks_average_mask_set_off() \ |
| | 3566 | |
| | 3567 | #define blend_blocks_average_mask_copy_off() \ |
| | 3568 | vmov draw_mask, draw_mask_next \ |
| | 3569 | |
| | 3570 | #define blend_blocks_average_mask_copy_b_off() \ |
| | 3571 | |
| | 3572 | #define blend_blocks_average_builder(texturing, mask_evaluate) \ |
| | 3573 | .align 3; \ |
| | 3574 | \ |
| | 3575 | function(blend_blocks_##texturing##_average_##mask_evaluate) \ |
| | 3576 | stmdb sp!, { r4, r14 }; \ |
| | 3577 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 3578 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 3579 | \ |
| | 3580 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16); \ |
| | 3581 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 3582 | \ |
| | 3583 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 3584 | mov c_64, #64; \ |
| | 3585 | \ |
| | 3586 | vmov.u16 d128_0x8000, #0x8000; \ |
| | 3587 | vld1.u32 { draw_mask_next }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3588 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3589 | \ |
| | 3590 | vmov.u16 d128_0x0421, #0x0400; \ |
| | 3591 | vld1.u32 { pixels_next }, [ pixel_ptr, :128 ], c_64; \ |
| | 3592 | \ |
| | 3593 | vorr.u16 d128_0x0421, #0x0021; \ |
| | 3594 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ]; \ |
| | 3595 | \ |
| | 3596 | veor.u16 blend_pixels_next, pixels_next, fb_pixels_next; \ |
| | 3597 | vbic.u16 pixels_no_msb, pixels_next, d128_0x8000; \ |
| | 3598 | vand.u16 blend_pixels_next, blend_pixels_next, d128_0x0421; \ |
| | 3599 | vsub.u16 blend_pixels_next, pixels_no_msb, blend_pixels_next; \ |
| | 3600 | blend_blocks_average_mask_set_##mask_evaluate(); \ |
| | 3601 | vbic.u16 fb_pixels_no_msb, fb_pixels_next, d128_0x8000; \ |
| | 3602 | \ |
| | 3603 | subs num_blocks, num_blocks, #1; \ |
| | 3604 | beq 1f; \ |
| | 3605 | \ |
| | 3606 | 0: \ |
| | 3607 | mov fb_ptr, fb_ptr_next; \ |
| | 3608 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3609 | \ |
| | 3610 | vmov pixels, pixels_next; \ |
| | 3611 | vld1.u32 { pixels_next }, [ pixel_ptr, :128 ], c_64; \ |
| | 3612 | \ |
| | 3613 | vhadd.u16 blend_pixels, fb_pixels_no_msb, blend_pixels_next; \ |
| | 3614 | \ |
| | 3615 | blend_blocks_average_mask_copy_##mask_evaluate(); \ |
| | 3616 | vld1.u32 { draw_mask_next }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3617 | \ |
| | 3618 | blend_blocks_average_set_blend_mask_##texturing(pixels); \ |
| | 3619 | blend_blocks_average_set_stp_bit_##texturing(); \ |
| | 3620 | vmov fb_pixels, fb_pixels_next; \ |
| | 3621 | blend_blocks_average_combine_##texturing(pixels); \ |
| | 3622 | \ |
| | 3623 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr; \ |
| | 3624 | add fb_ptr_cmp, fb_ptr_cmp, #14; \ |
| | 3625 | cmp fb_ptr_cmp, #28; \ |
| | 3626 | bls 2f; \ |
| | 3627 | \ |
| | 3628 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ]; \ |
| | 3629 | veor.u16 blend_pixels_next, pixels_next, fb_pixels_next; \ |
| | 3630 | \ |
| | 3631 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 3632 | vbic.u16 pixels_no_msb, pixels_next, d128_0x8000; \ |
| | 3633 | \ |
| | 3634 | vand.u16 blend_pixels_next, blend_pixels_next, d128_0x0421; \ |
| | 3635 | vbif.u16 fb_pixels, blend_pixels, draw_mask; \ |
| | 3636 | \ |
| | 3637 | vbic.u16 fb_pixels_no_msb, fb_pixels_next, d128_0x8000; \ |
| | 3638 | vsub.u16 blend_pixels_next, pixels_no_msb, blend_pixels_next; \ |
| | 3639 | blend_blocks_average_mask_set_##mask_evaluate(); \ |
| | 3640 | vst1.u16 { fb_pixels }, [ fb_ptr ]; \ |
| | 3641 | \ |
| | 3642 | 3: \ |
| | 3643 | subs num_blocks, num_blocks, #1; \ |
| | 3644 | bne 0b; \ |
| | 3645 | \ |
| | 3646 | 1: \ |
| | 3647 | blend_blocks_average_mask_copy_b_##mask_evaluate(); \ |
| | 3648 | vhadd.u16 blend_pixels, fb_pixels_no_msb, blend_pixels_next; \ |
| | 3649 | \ |
| | 3650 | blend_blocks_average_set_blend_mask_##texturing(pixels_next); \ |
| | 3651 | blend_blocks_average_set_stp_bit_##texturing(); \ |
| | 3652 | blend_blocks_average_combine_##texturing(pixels_next); \ |
| | 3653 | \ |
| | 3654 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 3655 | vbif.u16 fb_pixels_next, blend_pixels, draw_mask_next; \ |
| | 3656 | vst1.u16 { fb_pixels_next }, [ fb_ptr_next ]; \ |
| | 3657 | \ |
| | 3658 | ldmia sp!, { r4, pc }; \ |
| | 3659 | \ |
| | 3660 | 2: \ |
| | 3661 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 3662 | vbif.u16 fb_pixels, blend_pixels, draw_mask; \ |
| | 3663 | vst1.u16 { fb_pixels }, [ fb_ptr ]; \ |
| | 3664 | \ |
| | 3665 | vld1.u16 { fb_pixels_next }, [ fb_ptr_next ]; \ |
| | 3666 | veor.u16 blend_pixels_next, pixels_next, fb_pixels_next; \ |
| | 3667 | vbic.u16 pixels_no_msb, pixels_next, d128_0x8000; \ |
| | 3668 | vand.u16 blend_pixels_next, blend_pixels_next, d128_0x0421; \ |
| | 3669 | vsub.u16 blend_pixels_next, pixels_no_msb, blend_pixels_next; \ |
| | 3670 | vbic.u16 fb_pixels_no_msb, fb_pixels_next, d128_0x8000; \ |
| | 3671 | \ |
| | 3672 | bal 3b \ |
| | 3673 | |
| | 3674 | blend_blocks_average_builder(textured, off) |
| | 3675 | blend_blocks_average_builder(untextured, off) |
| | 3676 | blend_blocks_average_builder(textured, on) |
| | 3677 | blend_blocks_average_builder(untextured, on) |
| | 3678 | |
| | 3679 | |
| | 3680 | #define blend_blocks_add_mask_set_on() \ |
| | 3681 | vclt.s16 write_mask, fb_pixels, #0 \ |
| | 3682 | |
| | 3683 | #define blend_blocks_add_mask_copy_on() \ |
| | 3684 | vorr.u16 draw_mask, draw_mask, write_mask \ |
| | 3685 | |
| | 3686 | #define blend_blocks_add_mask_set_off() \ |
| | 3687 | |
| | 3688 | #define blend_blocks_add_mask_copy_off() \ |
| | 3689 | |
| | 3690 | |
| | 3691 | #define blend_blocks_add_textured_builder(mask_evaluate) \ |
| | 3692 | .align 3; \ |
| | 3693 | \ |
| | 3694 | function(blend_blocks_textured_add_##mask_evaluate) \ |
| | 3695 | stmdb sp!, { r4, r14 }; \ |
| | 3696 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 3697 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 3698 | \ |
| | 3699 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16); \ |
| | 3700 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 3701 | \ |
| | 3702 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 3703 | mov c_64, #64; \ |
| | 3704 | \ |
| | 3705 | vmov.u16 d128_0x7C1F, #0x7C00; \ |
| | 3706 | vmov.u16 d128_0x03E0, #0x0300; \ |
| | 3707 | vmov.u16 d128_0x83E0, #0x8000; \ |
| | 3708 | vorr.u16 d128_0x03E0, #0x00E0; \ |
| | 3709 | vorr.u16 d128_0x7C1F, #0x001F; \ |
| | 3710 | vorr.u16 d128_0x83E0, d128_0x83E0, d128_0x03E0; \ |
| | 3711 | \ |
| | 3712 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3713 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3714 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 3715 | vclt.s16 blend_mask, pixels, #0; \ |
| | 3716 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3717 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 3718 | vand.u16 pixels_rb, pixels, d128_0x7C1F; \ |
| | 3719 | \ |
| | 3720 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 3721 | vorr.u16 pixels, pixels, msb_mask; \ |
| | 3722 | vand.u16 fb_pixels_masked, fb_pixels, blend_mask; \ |
| | 3723 | vand.u16 pixels_mg, pixels, d128_0x83E0; \ |
| | 3724 | vand.u16 fb_pixels_rb, fb_pixels_masked, d128_0x7C1F; \ |
| | 3725 | vand.u16 fb_pixels_g, fb_pixels_masked, d128_0x03E0; \ |
| | 3726 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 3727 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_mg; \ |
| | 3728 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 3729 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x83E0; \ |
| | 3730 | \ |
| | 3731 | subs num_blocks, num_blocks, #1; \ |
| | 3732 | beq 1f; \ |
| | 3733 | \ |
| | 3734 | 0: \ |
| | 3735 | mov fb_ptr, fb_ptr_next; \ |
| | 3736 | \ |
| | 3737 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3738 | \ |
| | 3739 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 3740 | vclt.s16 blend_mask, pixels, #0; \ |
| | 3741 | \ |
| | 3742 | vorr.u16 pixels, pixels, msb_mask; \ |
| | 3743 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 3744 | vand.u16 pixels_mg, pixels, d128_0x83E0; \ |
| | 3745 | \ |
| | 3746 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 3747 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3748 | \ |
| | 3749 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr; \ |
| | 3750 | add fb_ptr_cmp, fb_ptr_cmp, #14; \ |
| | 3751 | cmp fb_ptr_cmp, #28; \ |
| | 3752 | bls 2f; \ |
| | 3753 | \ |
| | 3754 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3755 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 3756 | vand.u16 fb_pixels_masked, fb_pixels, blend_mask; \ |
| | 3757 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 3758 | vand.u16 pixels_rb, pixels, d128_0x7C1F; \ |
| | 3759 | vand.u16 fb_pixels_rb, fb_pixels_masked, d128_0x7C1F; \ |
| | 3760 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 3761 | \ |
| | 3762 | 3: \ |
| | 3763 | vand.u16 fb_pixels_g, fb_pixels_masked, d128_0x03E0; \ |
| | 3764 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 3765 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_mg; \ |
| | 3766 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 3767 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x83E0; \ |
| | 3768 | \ |
| | 3769 | subs num_blocks, num_blocks, #1; \ |
| | 3770 | bne 0b; \ |
| | 3771 | \ |
| | 3772 | 1: \ |
| | 3773 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 3774 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 3775 | vst1.u16 { blend_pixels }, [ fb_ptr_next ]; \ |
| | 3776 | \ |
| | 3777 | ldmia sp!, { r4, pc }; \ |
| | 3778 | \ |
| | 3779 | 2: \ |
| | 3780 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 3781 | vand.u16 pixels_rb, pixels, d128_0x7C1F; \ |
| | 3782 | \ |
| | 3783 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3784 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 3785 | vand.u16 fb_pixels_masked, fb_pixels, blend_mask; \ |
| | 3786 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 3787 | vand.u16 fb_pixels_rb, fb_pixels_masked, d128_0x7C1F; \ |
| | 3788 | bal 3b \ |
| | 3789 | |
| | 3790 | |
| | 3791 | #define blend_blocks_add_untextured_builder(mask_evaluate) \ |
| | 3792 | .align 3; \ |
| | 3793 | \ |
| | 3794 | function(blend_blocks_untextured_add_##mask_evaluate) \ |
| | 3795 | stmdb sp!, { r4, r14 }; \ |
| | 3796 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 3797 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 3798 | \ |
| | 3799 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16); \ |
| | 3800 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 3801 | \ |
| | 3802 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 3803 | mov c_64, #64; \ |
| | 3804 | \ |
| | 3805 | vmov.u16 d128_0x7C1F, #0x7C00; \ |
| | 3806 | vmov.u16 d128_0x03E0, #0x0300; \ |
| | 3807 | vorr.u16 d128_0x7C1F, #0x001F; \ |
| | 3808 | vorr.u16 d128_0x03E0, #0x00E0; \ |
| | 3809 | \ |
| | 3810 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3811 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3812 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 3813 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3814 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 3815 | vand.u16 pixels_rb, pixels, d128_0x7C1F; \ |
| | 3816 | \ |
| | 3817 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 3818 | vand.u16 pixels_g, pixels, d128_0x03E0; \ |
| | 3819 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 3820 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 3821 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 3822 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 3823 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 3824 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x03E0; \ |
| | 3825 | \ |
| | 3826 | subs num_blocks, num_blocks, #1; \ |
| | 3827 | beq 1f; \ |
| | 3828 | \ |
| | 3829 | 0: \ |
| | 3830 | mov fb_ptr, fb_ptr_next; \ |
| | 3831 | \ |
| | 3832 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3833 | \ |
| | 3834 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 3835 | \ |
| | 3836 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 3837 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 3838 | vand.u16 pixels_g, pixels, d128_0x03E0; \ |
| | 3839 | \ |
| | 3840 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 3841 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3842 | \ |
| | 3843 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr; \ |
| | 3844 | add fb_ptr_cmp, fb_ptr_cmp, #14; \ |
| | 3845 | cmp fb_ptr_cmp, #28; \ |
| | 3846 | bls 2f; \ |
| | 3847 | \ |
| | 3848 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3849 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 3850 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 3851 | vand.u16 pixels_rb, pixels, d128_0x7C1F; \ |
| | 3852 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 3853 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 3854 | \ |
| | 3855 | 3: \ |
| | 3856 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 3857 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 3858 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 3859 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 3860 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x03E0; \ |
| | 3861 | \ |
| | 3862 | subs num_blocks, num_blocks, #1; \ |
| | 3863 | bne 0b; \ |
| | 3864 | \ |
| | 3865 | 1: \ |
| | 3866 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 3867 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 3868 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 3869 | vst1.u16 { blend_pixels }, [ fb_ptr_next ]; \ |
| | 3870 | \ |
| | 3871 | ldmia sp!, { r4, pc }; \ |
| | 3872 | \ |
| | 3873 | 2: \ |
| | 3874 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 3875 | vand.u16 pixels_rb, pixels, d128_0x7C1F; \ |
| | 3876 | \ |
| | 3877 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3878 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 3879 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 3880 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 3881 | bal 3b \ |
| | 3882 | |
| | 3883 | |
| | 3884 | blend_blocks_add_textured_builder(off) |
| | 3885 | blend_blocks_add_textured_builder(on) |
| | 3886 | blend_blocks_add_untextured_builder(off) |
| | 3887 | blend_blocks_add_untextured_builder(on) |
| | 3888 | |
| | 3889 | #define blend_blocks_subtract_set_blend_mask_textured() \ |
| | 3890 | vclt.s16 blend_mask, pixels_next, #0 \ |
| | 3891 | |
| | 3892 | #define blend_blocks_subtract_combine_textured() \ |
| | 3893 | vbif.u16 blend_pixels, pixels, blend_mask \ |
| | 3894 | |
| | 3895 | #define blend_blocks_subtract_set_stb_textured() \ |
| | 3896 | vorr.u16 blend_pixels, #0x8000 \ |
| | 3897 | |
| | 3898 | #define blend_blocks_subtract_msb_mask_textured() \ |
| | 3899 | vorr.u16 pixels, pixels_next, msb_mask \ |
| | 3900 | |
| | 3901 | #define blend_blocks_subtract_set_blend_mask_untextured() \ |
| | 3902 | |
| | 3903 | #define blend_blocks_subtract_combine_untextured() \ |
| | 3904 | |
| | 3905 | #define blend_blocks_subtract_set_stb_untextured() \ |
| | 3906 | vorr.u16 blend_pixels, blend_pixels, msb_mask \ |
| | 3907 | |
| | 3908 | #define blend_blocks_subtract_msb_mask_untextured() \ |
| | 3909 | |
| | 3910 | |
| | 3911 | #define blend_blocks_subtract_mask_set_on() \ |
| | 3912 | vclt.s16 write_mask, fb_pixels, #0 \ |
| | 3913 | |
| | 3914 | #define blend_blocks_subtract_mask_copy_on() \ |
| | 3915 | vorr.u16 draw_mask, draw_mask_next, write_mask \ |
| | 3916 | |
| | 3917 | #define blend_blocks_subtract_mask_set_off() \ |
| | 3918 | |
| | 3919 | #define blend_blocks_subtract_mask_copy_off() \ |
| | 3920 | vmov draw_mask, draw_mask_next \ |
| | 3921 | |
| | 3922 | |
| | 3923 | #define blend_blocks_subtract_builder(texturing, mask_evaluate) \ |
| | 3924 | .align 3; \ |
| | 3925 | \ |
| | 3926 | function(blend_blocks_##texturing##_subtract_##mask_evaluate) \ |
| | 3927 | stmdb sp!, { r4, r14 }; \ |
| | 3928 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 3929 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 3930 | \ |
| | 3931 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16); \ |
| | 3932 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 3933 | \ |
| | 3934 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 3935 | mov c_64, #64; \ |
| | 3936 | \ |
| | 3937 | vmov.u16 d128_0x7C1F, #0x7C00; \ |
| | 3938 | vmov.u16 d128_0x03E0, #0x0300; \ |
| | 3939 | vorr.u16 d128_0x7C1F, #0x001F; \ |
| | 3940 | vorr.u16 d128_0x03E0, #0x00E0; \ |
| | 3941 | \ |
| | 3942 | vld1.u32 { draw_mask_next }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3943 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3944 | vld1.u32 { pixels_next }, [ pixel_ptr, :128 ], c_64; \ |
| | 3945 | blend_blocks_subtract_set_blend_mask_##texturing(); \ |
| | 3946 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3947 | blend_blocks_subtract_mask_set_##mask_evaluate(); \ |
| | 3948 | vand.u16 pixels_rb, pixels_next, d128_0x7C1F; \ |
| | 3949 | \ |
| | 3950 | vand.u16 pixels_g, pixels_next, d128_0x03E0; \ |
| | 3951 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 3952 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 3953 | vqsub.u8 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 3954 | vqsub.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 3955 | \ |
| | 3956 | subs num_blocks, num_blocks, #1; \ |
| | 3957 | beq 1f; \ |
| | 3958 | \ |
| | 3959 | 0: \ |
| | 3960 | blend_blocks_subtract_mask_copy_##mask_evaluate(); \ |
| | 3961 | mov fb_ptr, fb_ptr_next; \ |
| | 3962 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 3963 | \ |
| | 3964 | vld1.u32 { draw_mask_next }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 3965 | blend_blocks_subtract_msb_mask_##texturing(); \ |
| | 3966 | \ |
| | 3967 | vld1.u32 { pixels_next }, [ pixel_ptr, :128 ], c_64; \ |
| | 3968 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 3969 | vand.u16 pixels_rb, pixels_next, d128_0x7C1F; \ |
| | 3970 | blend_blocks_subtract_set_stb_##texturing(); \ |
| | 3971 | vand.u16 pixels_g, pixels_next, d128_0x03E0; \ |
| | 3972 | blend_blocks_subtract_combine_##texturing(); \ |
| | 3973 | blend_blocks_subtract_set_blend_mask_##texturing(); \ |
| | 3974 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 3975 | \ |
| | 3976 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr; \ |
| | 3977 | add fb_ptr_cmp, fb_ptr_cmp, #14; \ |
| | 3978 | cmp fb_ptr_cmp, #28; \ |
| | 3979 | bls 2f; \ |
| | 3980 | \ |
| | 3981 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 3982 | blend_blocks_subtract_mask_set_##mask_evaluate(); \ |
| | 3983 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 3984 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 3985 | vqsub.u8 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 3986 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 3987 | vqsub.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 3988 | \ |
| | 3989 | 3: \ |
| | 3990 | subs num_blocks, num_blocks, #1; \ |
| | 3991 | bne 0b; \ |
| | 3992 | \ |
| | 3993 | 1: \ |
| | 3994 | blend_blocks_subtract_mask_copy_##mask_evaluate(); \ |
| | 3995 | \ |
| | 3996 | blend_blocks_subtract_msb_mask_##texturing(); \ |
| | 3997 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 3998 | blend_blocks_subtract_set_stb_##texturing(); \ |
| | 3999 | blend_blocks_subtract_combine_##texturing(); \ |
| | 4000 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 4001 | vst1.u16 { blend_pixels }, [ fb_ptr_next ]; \ |
| | 4002 | \ |
| | 4003 | ldmia sp!, { r4, pc }; \ |
| | 4004 | \ |
| | 4005 | 2: \ |
| | 4006 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 4007 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4008 | blend_blocks_subtract_mask_set_##mask_evaluate(); \ |
| | 4009 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4010 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 4011 | vqsub.u8 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 4012 | vqsub.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 4013 | bal 3b \ |
| | 4014 | |
| | 4015 | |
| | 4016 | blend_blocks_subtract_builder(textured, off) |
| | 4017 | blend_blocks_subtract_builder(textured, on) |
| | 4018 | blend_blocks_subtract_builder(untextured, off) |
| | 4019 | blend_blocks_subtract_builder(untextured, on) |
| | 4020 | |
| | 4021 | |
| | 4022 | #define blend_blocks_add_fourth_textured_builder(mask_evaluate) \ |
| | 4023 | .align 3; \ |
| | 4024 | \ |
| | 4025 | function(blend_blocks_textured_add_fourth_##mask_evaluate) \ |
| | 4026 | stmdb sp!, { r4, r14 }; \ |
| | 4027 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 4028 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 4029 | \ |
| | 4030 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16); \ |
| | 4031 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 4032 | \ |
| | 4033 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 4034 | mov c_64, #64; \ |
| | 4035 | \ |
| | 4036 | vmov.u16 d128_0x7C1F, #0x7C00; \ |
| | 4037 | vmov.u16 d128_0x03E0, #0x0300; \ |
| | 4038 | vmov.u16 d128_0x1C07, #0x1C00; \ |
| | 4039 | vmov.u16 d128_0x00E0, #0x00E0; \ |
| | 4040 | vorr.u16 d128_0x7C1F, #0x001F; \ |
| | 4041 | vorr.u16 d128_0x03E0, #0x00E0; \ |
| | 4042 | vorr.u16 d128_0x1C07, #0x0007; \ |
| | 4043 | \ |
| | 4044 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 4045 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 4046 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 4047 | vclt.s16 blend_mask, pixels, #0; \ |
| | 4048 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4049 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 4050 | vshr.s16 pixels_fourth, pixels, #2; \ |
| | 4051 | vand.u16 pixels_rb, pixels_fourth, d128_0x1C07; \ |
| | 4052 | \ |
| | 4053 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 4054 | vand.u16 pixels_g, pixels_fourth, d128_0x00E0; \ |
| | 4055 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4056 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 4057 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 4058 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 4059 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 4060 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x03E0; \ |
| | 4061 | \ |
| | 4062 | subs num_blocks, num_blocks, #1; \ |
| | 4063 | beq 1f; \ |
| | 4064 | \ |
| | 4065 | 0: \ |
| | 4066 | mov fb_ptr, fb_ptr_next; \ |
| | 4067 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 4068 | \ |
| | 4069 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 4070 | vbif.u16 blend_pixels, pixels, blend_mask; \ |
| | 4071 | \ |
| | 4072 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 4073 | vclt.s16 blend_mask, pixels, #0; \ |
| | 4074 | vshr.s16 pixels_fourth, pixels, #2; \ |
| | 4075 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 4076 | vand.u16 pixels_rb, pixels_fourth, d128_0x1C07; \ |
| | 4077 | \ |
| | 4078 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 4079 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 4080 | \ |
| | 4081 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr; \ |
| | 4082 | add fb_ptr_cmp, fb_ptr_cmp, #14; \ |
| | 4083 | cmp fb_ptr_cmp, #28; \ |
| | 4084 | bls 2f; \ |
| | 4085 | \ |
| | 4086 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4087 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 4088 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 4089 | vand.u16 pixels_g, pixels_fourth, d128_0x00E0; \ |
| | 4090 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4091 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 4092 | \ |
| | 4093 | 3: \ |
| | 4094 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 4095 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 4096 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 4097 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 4098 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x03E0; \ |
| | 4099 | \ |
| | 4100 | subs num_blocks, num_blocks, #1; \ |
| | 4101 | bne 0b; \ |
| | 4102 | \ |
| | 4103 | 1: \ |
| | 4104 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 4105 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 4106 | vbif.u16 blend_pixels, pixels, blend_mask; \ |
| | 4107 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 4108 | vst1.u16 { blend_pixels }, [ fb_ptr_next ]; \ |
| | 4109 | \ |
| | 4110 | ldmia sp!, { r4, pc }; \ |
| | 4111 | \ |
| | 4112 | 2: \ |
| | 4113 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 4114 | vand.u16 pixels_g, pixels_fourth, d128_0x00E0; \ |
| | 4115 | \ |
| | 4116 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4117 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 4118 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 4119 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4120 | bal 3b \ |
| | 4121 | |
| | 4122 | |
| | 4123 | |
| | 4124 | #define blend_blocks_add_fourth_untextured_builder(mask_evaluate) \ |
| | 4125 | .align 3; \ |
| | 4126 | \ |
| | 4127 | function(blend_blocks_untextured_add_fourth_##mask_evaluate) \ |
| | 4128 | stmdb sp!, { r4, r14 }; \ |
| | 4129 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset; \ |
| | 4130 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 4131 | \ |
| | 4132 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16); \ |
| | 4133 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ]; \ |
| | 4134 | \ |
| | 4135 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 4136 | mov c_64, #64; \ |
| | 4137 | \ |
| | 4138 | vmov.u16 d128_0x7C1F, #0x7C00; \ |
| | 4139 | vmov.u16 d128_0x03E0, #0x0300; \ |
| | 4140 | vmov.u16 d128_0x1C07, #0x1C00; \ |
| | 4141 | vmov.u16 d128_0x00E0, #0x00E0; \ |
| | 4142 | vorr.u16 d128_0x7C1F, #0x001F; \ |
| | 4143 | vorr.u16 d128_0x03E0, #0x00E0; \ |
| | 4144 | vorr.u16 d128_0x1C07, #0x0007; \ |
| | 4145 | \ |
| | 4146 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 4147 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 4148 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 4149 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4150 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 4151 | vshr.s16 pixels_fourth, pixels, #2; \ |
| | 4152 | vand.u16 pixels_rb, pixels_fourth, d128_0x1C07; \ |
| | 4153 | \ |
| | 4154 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 4155 | vand.u16 pixels_g, pixels_fourth, d128_0x00E0; \ |
| | 4156 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4157 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 4158 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 4159 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 4160 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 4161 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x03E0; \ |
| | 4162 | \ |
| | 4163 | subs num_blocks, num_blocks, #1; \ |
| | 4164 | beq 1f; \ |
| | 4165 | \ |
| | 4166 | 0: \ |
| | 4167 | mov fb_ptr, fb_ptr_next; \ |
| | 4168 | ldr fb_ptr_next, [ pixel_ptr, #28 ]; \ |
| | 4169 | \ |
| | 4170 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64; \ |
| | 4171 | \ |
| | 4172 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 4173 | vshr.s16 pixels_fourth, pixels, #2; \ |
| | 4174 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 4175 | vand.u16 pixels_rb, pixels_fourth, d128_0x1C07; \ |
| | 4176 | \ |
| | 4177 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 4178 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64; \ |
| | 4179 | \ |
| | 4180 | sub fb_ptr_cmp, fb_ptr_next, fb_ptr; \ |
| | 4181 | add fb_ptr_cmp, fb_ptr_cmp, #14; \ |
| | 4182 | cmp fb_ptr_cmp, #28; \ |
| | 4183 | bls 2f; \ |
| | 4184 | \ |
| | 4185 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4186 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 4187 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 4188 | vand.u16 pixels_g, pixels_fourth, d128_0x00E0; \ |
| | 4189 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4190 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 4191 | \ |
| | 4192 | 3: \ |
| | 4193 | vand.u16 fb_pixels_g, fb_pixels, d128_0x03E0; \ |
| | 4194 | vadd.u16 fb_pixels_rb, fb_pixels_rb, pixels_rb; \ |
| | 4195 | vadd.u16 fb_pixels_g, fb_pixels_g, pixels_g; \ |
| | 4196 | vmin.u8 fb_pixels_rb, fb_pixels_rb, d128_0x7C1F; \ |
| | 4197 | vmin.u16 fb_pixels_g, fb_pixels_g, d128_0x03E0; \ |
| | 4198 | \ |
| | 4199 | subs num_blocks, num_blocks, #1; \ |
| | 4200 | bne 0b; \ |
| | 4201 | \ |
| | 4202 | 1: \ |
| | 4203 | vorr.u16 blend_pixels, fb_pixels_rb, fb_pixels_g; \ |
| | 4204 | vorr.u16 blend_pixels, blend_pixels, msb_mask; \ |
| | 4205 | vbit.u16 blend_pixels, fb_pixels, draw_mask; \ |
| | 4206 | vst1.u16 { blend_pixels }, [ fb_ptr_next ]; \ |
| | 4207 | \ |
| | 4208 | ldmia sp!, { r4, pc }; \ |
| | 4209 | \ |
| | 4210 | 2: \ |
| | 4211 | vst1.u16 { blend_pixels }, [ fb_ptr ]; \ |
| | 4212 | vand.u16 pixels_g, pixels_fourth, d128_0x00E0; \ |
| | 4213 | \ |
| | 4214 | vld1.u16 { fb_pixels }, [ fb_ptr_next ]; \ |
| | 4215 | blend_blocks_add_mask_set_##mask_evaluate(); \ |
| | 4216 | blend_blocks_add_mask_copy_##mask_evaluate(); \ |
| | 4217 | vand.u16 fb_pixels_rb, fb_pixels, d128_0x7C1F; \ |
| | 4218 | bal 3b \ |
| | 4219 | |
| | 4220 | |
| | 4221 | blend_blocks_add_fourth_textured_builder(off) |
| | 4222 | blend_blocks_add_fourth_textured_builder(on) |
| | 4223 | blend_blocks_add_fourth_untextured_builder(off) |
| | 4224 | blend_blocks_add_fourth_untextured_builder(on) |
| | 4225 | |
| | 4226 | // TODO: Optimize this more. Need a scene that actually uses it for |
| | 4227 | // confirmation.. |
| | 4228 | |
| | 4229 | .align 3 |
| | 4230 | |
| | 4231 | function(blend_blocks_textured_unblended_on) |
| | 4232 | stmdb sp!, { r4, r14 } |
| | 4233 | add mask_msb_ptr, psx_gpu, #psx_gpu_mask_msb_offset |
| | 4234 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 4235 | |
| | 4236 | add pixel_ptr, psx_gpu, #(psx_gpu_blocks_offset + 16) |
| | 4237 | vld1.u16 { msb_mask_low[], msb_mask_high[] }, [ mask_msb_ptr, :16 ] |
| | 4238 | |
| | 4239 | add draw_mask_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 4240 | mov c_64, #64 |
| | 4241 | |
| | 4242 | ldr fb_ptr, [ pixel_ptr, #28 ] |
| | 4243 | vld1.u16 { fb_pixels }, [ fb_ptr ] |
| | 4244 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64 |
| | 4245 | vclt.s16 write_mask, fb_pixels, #0 |
| | 4246 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64 |
| | 4247 | |
| | 4248 | subs num_blocks, num_blocks, #1 |
| | 4249 | beq 1f |
| | 4250 | |
| | 4251 | 0: |
| | 4252 | vorr.u16 draw_mask, draw_mask, write_mask |
| | 4253 | vbif.u16 fb_pixels, pixels, draw_mask |
| | 4254 | vst1.u16 { fb_pixels }, [ fb_ptr ] |
| | 4255 | |
| | 4256 | ldr fb_ptr, [ pixel_ptr, #28 ] |
| | 4257 | vld1.u16 { fb_pixels }, [ fb_ptr ] |
| | 4258 | vld1.u32 { draw_mask }, [ draw_mask_ptr, :128 ], c_64 |
| | 4259 | vclt.s16 write_mask, fb_pixels, #0 |
| | 4260 | vld1.u32 { pixels }, [ pixel_ptr, :128 ], c_64 |
| | 4261 | |
| | 4262 | subs num_blocks, num_blocks, #1 |
| | 4263 | bne 0b |
| | 4264 | |
| | 4265 | 1: |
| | 4266 | vorr.u16 draw_mask, draw_mask, write_mask |
| | 4267 | vbif.u16 fb_pixels, pixels, draw_mask |
| | 4268 | vst1.u16 { fb_pixels }, [ fb_ptr ] |
| | 4269 | |
| | 4270 | ldmia sp!, { r4, pc } |
| | 4271 | |
| | 4272 | |
| | 4273 | function(blend_blocks_textured_unblended_off) |
| | 4274 | bx lr |
| | 4275 | |
| | 4276 | |
| | 4277 | function(warmup) |
| | 4278 | mov r3, #64 |
| | 4279 | cmp r0, #0 |
| | 4280 | bxeq lr |
| | 4281 | |
| | 4282 | 0: |
| | 4283 | vld1.u32 { u_whole_8, v_whole_8 }, [ r1, :128 ], r3 |
| | 4284 | |
| | 4285 | subs r0, r0, #1 |
| | 4286 | bne 0b |
| | 4287 | |
| | 4288 | bx lr |
| | 4289 | |
| | 4290 | #undef vram_ptr |
| | 4291 | #undef color |
| | 4292 | #undef width |
| | 4293 | #undef height |
| | 4294 | #undef pitch |
| | 4295 | |
| | 4296 | #define vram_ptr r0 |
| | 4297 | #define color r1 |
| | 4298 | #define width r2 |
| | 4299 | #define height r3 |
| | 4300 | |
| | 4301 | #define pitch r1 |
| | 4302 | |
| | 4303 | #define num_width r12 |
| | 4304 | |
| | 4305 | #undef colors_a |
| | 4306 | #undef colors_b |
| | 4307 | |
| | 4308 | #define colors_a q0 |
| | 4309 | #define colors_b q1 |
| | 4310 | |
| | 4311 | .align 3 |
| | 4312 | |
| | 4313 | function(render_block_fill_body) |
| | 4314 | vdup.u16 colors_a, color |
| | 4315 | mov pitch, #2048 |
| | 4316 | |
| | 4317 | vmov colors_b, colors_a |
| | 4318 | sub pitch, pitch, width, lsl #1 |
| | 4319 | |
| | 4320 | mov num_width, width |
| | 4321 | |
| | 4322 | 0: |
| | 4323 | vst1.u32 { colors_a, colors_b }, [ vram_ptr, :256 ]! |
| | 4324 | |
| | 4325 | subs num_width, num_width, #16 |
| | 4326 | bne 0b |
| | 4327 | |
| | 4328 | add vram_ptr, vram_ptr, pitch |
| | 4329 | mov num_width, width |
| | 4330 | |
| | 4331 | subs height, height, #1 |
| | 4332 | bne 0b |
| | 4333 | |
| | 4334 | bx lr |
| | 4335 | |
| | 4336 | |
| | 4337 | #undef x |
| | 4338 | #undef y |
| | 4339 | #undef width |
| | 4340 | #undef height |
| | 4341 | #undef fb_ptr |
| | 4342 | #undef texture_mask |
| | 4343 | #undef num_blocks |
| | 4344 | #undef temp |
| | 4345 | #undef dirty_textures_mask |
| | 4346 | #undef clut_ptr |
| | 4347 | #undef current_texture_mask |
| | 4348 | |
| | 4349 | #define psx_gpu r0 |
| | 4350 | #define x r1 |
| | 4351 | #define y r2 |
| | 4352 | #define u r3 |
| | 4353 | #define v r4 |
| | 4354 | #define width r5 |
| | 4355 | #define height r6 |
| | 4356 | #define offset_u r8 |
| | 4357 | #define offset_v r9 |
| | 4358 | #define offset_u_right r10 |
| | 4359 | #define width_rounded r11 |
| | 4360 | #define height_rounded r12 |
| | 4361 | |
| | 4362 | #define texture_offset_base r1 |
| | 4363 | #define tile_width r2 |
| | 4364 | #define tile_height r3 |
| | 4365 | #define num_blocks r4 |
| | 4366 | #define block r5 |
| | 4367 | #define sub_tile_height r6 |
| | 4368 | #define fb_ptr r7 |
| | 4369 | #define texture_mask r8 |
| | 4370 | #define column_data r9 |
| | 4371 | #define texture_offset r10 |
| | 4372 | #define tiles_remaining r11 |
| | 4373 | #define fb_ptr_advance_column r12 |
| | 4374 | #define texture_block_ptr r14 |
| | 4375 | |
| | 4376 | #define texture_page_ptr r3 |
| | 4377 | #define left_block_mask r4 |
| | 4378 | #define right_block_mask r5 |
| | 4379 | #define texture_mask_rev r10 |
| | 4380 | #define control_mask r11 |
| | 4381 | |
| | 4382 | #define dirty_textures_mask r4 |
| | 4383 | #define clut_ptr r5 |
| | 4384 | #define current_texture_mask r6 |
| | 4385 | |
| | 4386 | |
| | 4387 | #undef texels |
| | 4388 | #undef clut_low_a |
| | 4389 | #undef clut_low_b |
| | 4390 | #undef clut_high_a |
| | 4391 | #undef clut_high_b |
| | 4392 | #undef clut_a |
| | 4393 | #undef clut_b |
| | 4394 | #undef texels_low |
| | 4395 | #undef texels_high |
| | 4396 | |
| | 4397 | #define texels d0 |
| | 4398 | #define draw_masks_fb_ptrs q1 |
| | 4399 | |
| | 4400 | #define draw_mask_fb_ptr_left d2 |
| | 4401 | #define draw_mask_fb_ptr_right d3 |
| | 4402 | |
| | 4403 | #define clut_low_a d4 |
| | 4404 | #define clut_low_b d5 |
| | 4405 | #define clut_high_a d6 |
| | 4406 | #define clut_high_b d7 |
| | 4407 | |
| | 4408 | #define block_masks d8 |
| | 4409 | #define block_masks_shifted d9 |
| | 4410 | |
| | 4411 | #define clut_a q2 |
| | 4412 | #define clut_b q3 |
| | 4413 | |
| | 4414 | #define texels_low d10 |
| | 4415 | #define texels_high d11 |
| | 4416 | |
| | 4417 | |
| | 4418 | setup_sprite_flush_blocks_single: |
| | 4419 | vpush { q1 - q4 } |
| | 4420 | |
| | 4421 | stmdb sp!, { r0 - r3, r12, r14 } |
| | 4422 | bl flush_render_block_buffer |
| | 4423 | ldmia sp!, { r0 - r3, r12, r14 } |
| | 4424 | |
| | 4425 | vpop { q1 - q4 } |
| | 4426 | |
| | 4427 | add block, psx_gpu, #psx_gpu_blocks_offset |
| | 4428 | |
| | 4429 | mov num_blocks, sub_tile_height |
| | 4430 | bx lr |
| | 4431 | |
| | 4432 | |
| | 4433 | setup_sprite_flush_blocks_double: |
| | 4434 | vpush { q1 - q4 } |
| | 4435 | |
| | 4436 | stmdb sp!, { r0 - r3, r12, r14 } |
| | 4437 | bl flush_render_block_buffer |
| | 4438 | ldmia sp!, { r0 - r3, r12, r14 } |
| | 4439 | |
| | 4440 | vpop { q1 - q4 } |
| | 4441 | |
| | 4442 | add block, psx_gpu, #psx_gpu_blocks_offset |
| | 4443 | |
| | 4444 | mov num_blocks, sub_tile_height, lsl #1 |
| | 4445 | bx lr |
| | 4446 | |
| | 4447 | |
| | 4448 | setup_sprite_update_texture_4bpp_cache: |
| | 4449 | stmdb sp!, { r0 - r3, r14 } |
| | 4450 | bl update_texture_4bpp_cache |
| | 4451 | ldmia sp!, { r0 - r3, pc } |
| | 4452 | |
| | 4453 | |
| | 4454 | setup_sprite_update_texture_8bpp_cache: |
| | 4455 | stmdb sp!, { r0 - r3, r14 } |
| | 4456 | bl update_texture_8bpp_cache |
| | 4457 | ldmia sp!, { r0 - r3, pc } |
| | 4458 | |
| | 4459 | |
| | 4460 | #define setup_sprite_tiled_initialize_4bpp() \ |
| | 4461 | ldr dirty_textures_mask, \ |
| | 4462 | [ psx_gpu, #psx_gpu_dirty_textures_4bpp_mask_offset ]; \ |
| | 4463 | ldr clut_ptr, [ psx_gpu, #psx_gpu_clut_ptr_offset ]; \ |
| | 4464 | \ |
| | 4465 | ldr current_texture_mask, [ psx_gpu, #psx_gpu_current_texture_mask_offset ]; \ |
| | 4466 | vld1.u32 { clut_a, clut_b }, [ clut_ptr, :128 ]; \ |
| | 4467 | \ |
| | 4468 | tst current_texture_mask, dirty_textures_mask; \ |
| | 4469 | vuzp.u8 clut_a, clut_b; \ |
| | 4470 | \ |
| | 4471 | blne setup_sprite_update_texture_4bpp_cache \ |
| | 4472 | |
| | 4473 | #define setup_sprite_tiled_initialize_8bpp() \ |
| | 4474 | ldr dirty_textures_mask, \ |
| | 4475 | [ psx_gpu, #psx_gpu_dirty_textures_8bpp_mask_offset ]; \ |
| | 4476 | ldr current_texture_mask, [ psx_gpu, #psx_gpu_current_texture_mask_offset ]; \ |
| | 4477 | \ |
| | 4478 | tst current_texture_mask, dirty_textures_mask; \ |
| | 4479 | blne setup_sprite_update_texture_8bpp_cache \ |
| | 4480 | |
| | 4481 | |
| | 4482 | #define setup_sprite_tile_setup_block_no(side, offset, texture_mode) \ |
| | 4483 | |
| | 4484 | #define setup_sprite_block_count_single() \ |
| | 4485 | sub_tile_height \ |
| | 4486 | |
| | 4487 | #define setup_sprite_block_count_double() \ |
| | 4488 | sub_tile_height, lsl #1 \ |
| | 4489 | |
| | 4490 | #define setup_sprite_tile_add_blocks(type) \ |
| | 4491 | add num_blocks, num_blocks, setup_sprite_block_count_##type(); \ |
| | 4492 | cmp num_blocks, #MAX_BLOCKS; \ |
| | 4493 | \ |
| | 4494 | blgt setup_sprite_flush_blocks_##type \ |
| | 4495 | |
| | 4496 | |
| | 4497 | #define setup_sprite_tile_full_4bpp(edge) \ |
| | 4498 | setup_sprite_tile_add_blocks(double); \ |
| | 4499 | \ |
| | 4500 | 4: \ |
| | 4501 | and texture_block_ptr, texture_offset, texture_mask; \ |
| | 4502 | vmov.u32 draw_mask_fb_ptr_left[1], fb_ptr; \ |
| | 4503 | \ |
| | 4504 | pld [ fb_ptr ]; \ |
| | 4505 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4506 | vld1.u32 { texels }, [ texture_block_ptr, :64 ]; \ |
| | 4507 | \ |
| | 4508 | vtbl.8 texels_low, { clut_low_a, clut_low_b }, texels; \ |
| | 4509 | vtbl.8 texels_high, { clut_high_a, clut_high_b }, texels; \ |
| | 4510 | \ |
| | 4511 | vst2.u8 { texels_low, texels_high }, [ block, :128 ]; \ |
| | 4512 | add texture_block_ptr, texture_offset, #8; \ |
| | 4513 | \ |
| | 4514 | and texture_block_ptr, texture_block_ptr, texture_mask; \ |
| | 4515 | add block, block, #40; \ |
| | 4516 | \ |
| | 4517 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4518 | add fb_ptr, fb_ptr, #16; \ |
| | 4519 | \ |
| | 4520 | vst1.u32 { draw_mask_fb_ptr_left }, [ block, :64 ]; \ |
| | 4521 | add block, block, #24; \ |
| | 4522 | \ |
| | 4523 | vld1.u32 { texels }, [ texture_block_ptr, :64 ]; \ |
| | 4524 | vtbl.8 texels_low, { clut_low_a, clut_low_b }, texels; \ |
| | 4525 | \ |
| | 4526 | pld [ fb_ptr ]; \ |
| | 4527 | vmov.u32 draw_mask_fb_ptr_right[1], fb_ptr; \ |
| | 4528 | vtbl.8 texels_high, { clut_high_a, clut_high_b }, texels; \ |
| | 4529 | \ |
| | 4530 | vst2.u8 { texels_low, texels_high }, [ block, :128 ]; \ |
| | 4531 | add block, block, #40; \ |
| | 4532 | \ |
| | 4533 | add texture_offset, texture_offset, #0x10; \ |
| | 4534 | add fb_ptr, fb_ptr, #(2048 - 16); \ |
| | 4535 | \ |
| | 4536 | vst1.u32 { draw_mask_fb_ptr_right }, [ block, :64 ]; \ |
| | 4537 | add block, block, #24; \ |
| | 4538 | \ |
| | 4539 | subs sub_tile_height, sub_tile_height, #1; \ |
| | 4540 | bne 4b; \ |
| | 4541 | \ |
| | 4542 | add texture_offset, texture_offset, #0xF00; \ |
| | 4543 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] \ |
| | 4544 | |
| | 4545 | |
| | 4546 | #define setup_sprite_tile_half_4bpp(edge) \ |
| | 4547 | setup_sprite_tile_add_blocks(single); \ |
| | 4548 | \ |
| | 4549 | 4: \ |
| | 4550 | and texture_block_ptr, texture_offset, texture_mask; \ |
| | 4551 | vmov.u32 draw_mask_fb_ptr_##edge[1], fb_ptr; \ |
| | 4552 | \ |
| | 4553 | pld [ fb_ptr ]; \ |
| | 4554 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4555 | vld1.u32 { texels }, [ texture_block_ptr, :64 ]; \ |
| | 4556 | \ |
| | 4557 | vtbl.8 texels_low, { clut_low_a, clut_low_b }, texels; \ |
| | 4558 | vtbl.8 texels_high, { clut_high_a, clut_high_b }, texels; \ |
| | 4559 | \ |
| | 4560 | vst2.u8 { texels_low, texels_high }, [ block, :128 ]; \ |
| | 4561 | add block, block, #40; \ |
| | 4562 | \ |
| | 4563 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4564 | vst1.u32 { draw_mask_fb_ptr_##edge }, [ block, :64 ]; \ |
| | 4565 | \ |
| | 4566 | add block, block, #24; \ |
| | 4567 | add texture_offset, texture_offset, #0x10; \ |
| | 4568 | \ |
| | 4569 | add fb_ptr, fb_ptr, #2048; \ |
| | 4570 | subs sub_tile_height, sub_tile_height, #1; \ |
| | 4571 | \ |
| | 4572 | bne 4b; \ |
| | 4573 | \ |
| | 4574 | add texture_offset, texture_offset, #0xF00; \ |
| | 4575 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] \ |
| | 4576 | |
| | 4577 | |
| | 4578 | #define setup_sprite_tile_full_8bpp(edge) \ |
| | 4579 | setup_sprite_tile_add_blocks(double); \ |
| | 4580 | add block, block, #16; \ |
| | 4581 | \ |
| | 4582 | 4: \ |
| | 4583 | and texture_block_ptr, texture_offset, texture_mask; \ |
| | 4584 | vmov.u32 draw_mask_fb_ptr_left[1], fb_ptr; \ |
| | 4585 | \ |
| | 4586 | pld [ fb_ptr ]; \ |
| | 4587 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4588 | vld1.u32 { texels }, [ texture_block_ptr, :64 ]; \ |
| | 4589 | \ |
| | 4590 | add texture_block_ptr, texture_offset, #8; \ |
| | 4591 | vst1.u32 { texels }, [ block, :64 ]; \ |
| | 4592 | \ |
| | 4593 | and texture_block_ptr, texture_block_ptr, texture_mask; \ |
| | 4594 | add block, block, #24; \ |
| | 4595 | \ |
| | 4596 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4597 | \ |
| | 4598 | add fb_ptr, fb_ptr, #16; \ |
| | 4599 | vst1.u32 { draw_mask_fb_ptr_left }, [ block, :64 ]; \ |
| | 4600 | \ |
| | 4601 | add block, block, #40; \ |
| | 4602 | vld1.u32 { texels }, [ texture_block_ptr, :64 ]; \ |
| | 4603 | pld [ fb_ptr ]; \ |
| | 4604 | \ |
| | 4605 | vmov.u32 draw_mask_fb_ptr_right[1], fb_ptr; \ |
| | 4606 | vst1.u32 { texels }, [ block, :64 ]; \ |
| | 4607 | add block, block, #24; \ |
| | 4608 | \ |
| | 4609 | add texture_offset, texture_offset, #0x10; \ |
| | 4610 | add fb_ptr, fb_ptr, #(2048 - 16); \ |
| | 4611 | \ |
| | 4612 | vst1.u32 { draw_mask_fb_ptr_right }, [ block, :64 ]; \ |
| | 4613 | add block, block, #40; \ |
| | 4614 | \ |
| | 4615 | subs sub_tile_height, sub_tile_height, #1; \ |
| | 4616 | bne 4b; \ |
| | 4617 | \ |
| | 4618 | sub block, block, #16; \ |
| | 4619 | add texture_offset, texture_offset, #0xF00; \ |
| | 4620 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] \ |
| | 4621 | |
| | 4622 | |
| | 4623 | #define setup_sprite_tile_half_8bpp(edge) \ |
| | 4624 | setup_sprite_tile_add_blocks(single); \ |
| | 4625 | add block, block, #16; \ |
| | 4626 | \ |
| | 4627 | 4: \ |
| | 4628 | and texture_block_ptr, texture_offset, texture_mask; \ |
| | 4629 | vmov.u32 draw_mask_fb_ptr_##edge[1], fb_ptr; \ |
| | 4630 | pld [ fb_ptr ]; \ |
| | 4631 | \ |
| | 4632 | add texture_block_ptr, texture_page_ptr, texture_block_ptr; \ |
| | 4633 | vld1.u32 { texels }, [ texture_block_ptr, :64 ]; \ |
| | 4634 | \ |
| | 4635 | vst1.u32 { texels }, [ block, :64 ]; \ |
| | 4636 | add block, block, #24; \ |
| | 4637 | \ |
| | 4638 | vst1.u32 { draw_mask_fb_ptr_##edge }, [ block, :64 ]; \ |
| | 4639 | add block, block, #40; \ |
| | 4640 | \ |
| | 4641 | add texture_offset, texture_offset, #0x10; \ |
| | 4642 | add fb_ptr, fb_ptr, #2048; \ |
| | 4643 | \ |
| | 4644 | subs sub_tile_height, sub_tile_height, #1; \ |
| | 4645 | bne 4b; \ |
| | 4646 | \ |
| | 4647 | sub block, block, #16; \ |
| | 4648 | add texture_offset, texture_offset, #0xF00; \ |
| | 4649 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] \ |
| | 4650 | |
| | 4651 | |
| | 4652 | #define setup_sprite_tile_column_edge_pre_adjust_half_right() \ |
| | 4653 | add texture_offset, texture_offset_base, #8; \ |
| | 4654 | add fb_ptr, fb_ptr, #16 \ |
| | 4655 | |
| | 4656 | #define setup_sprite_tile_column_edge_pre_adjust_half_left() \ |
| | 4657 | mov texture_offset, texture_offset_base \ |
| | 4658 | |
| | 4659 | #define setup_sprite_tile_column_edge_pre_adjust_half(edge) \ |
| | 4660 | setup_sprite_tile_column_edge_pre_adjust_half_##edge() \ |
| | 4661 | |
| | 4662 | #define setup_sprite_tile_column_edge_pre_adjust_full(edge) \ |
| | 4663 | mov texture_offset, texture_offset_base \ |
| | 4664 | |
| | 4665 | #define setup_sprite_tile_column_edge_post_adjust_half_right() \ |
| | 4666 | sub fb_ptr, fb_ptr, #16 \ |
| | 4667 | |
| | 4668 | #define setup_sprite_tile_column_edge_post_adjust_half_left() \ |
| | 4669 | |
| | 4670 | #define setup_sprite_tile_column_edge_post_adjust_half(edge) \ |
| | 4671 | setup_sprite_tile_column_edge_post_adjust_half_##edge() \ |
| | 4672 | |
| | 4673 | #define setup_sprite_tile_column_edge_post_adjust_full(edge) \ |
| | 4674 | |
| | 4675 | |
| | 4676 | #define setup_sprite_tile_column_height_single(edge_mode, edge, texture_mode) \ |
| | 4677 | mov sub_tile_height, column_data; \ |
| | 4678 | setup_sprite_tile_column_edge_pre_adjust_##edge_mode(edge); \ |
| | 4679 | setup_sprite_tile_##edge_mode##_##texture_mode(edge); \ |
| | 4680 | setup_sprite_tile_column_edge_post_adjust_##edge_mode(edge) \ |
| | 4681 | |
| | 4682 | #define setup_sprite_tile_column_height_multi(edge_mode, edge, texture_mode) \ |
| | 4683 | and sub_tile_height, column_data, #0xFF; \ |
| | 4684 | mov tiles_remaining, column_data, lsr #16; \ |
| | 4685 | setup_sprite_tile_column_edge_pre_adjust_##edge_mode(edge); \ |
| | 4686 | setup_sprite_tile_##edge_mode##_##texture_mode(edge); \ |
| | 4687 | \ |
| | 4688 | subs tiles_remaining, tiles_remaining, #1; \ |
| | 4689 | beq 2f; \ |
| | 4690 | \ |
| | 4691 | 3: \ |
| | 4692 | mov sub_tile_height, #16; \ |
| | 4693 | setup_sprite_tile_##edge_mode##_##texture_mode(edge); \ |
| | 4694 | subs tiles_remaining, tiles_remaining, #1; \ |
| | 4695 | bne 3b; \ |
| | 4696 | \ |
| | 4697 | 2: \ |
| | 4698 | uxtb sub_tile_height, column_data, ror #8; \ |
| | 4699 | setup_sprite_tile_##edge_mode##_##texture_mode(edge); \ |
| | 4700 | setup_sprite_tile_column_edge_post_adjust_##edge_mode(edge) \ |
| | 4701 | |
| | 4702 | |
| | 4703 | #define setup_sprite_column_data_single() \ |
| | 4704 | mov column_data, height; \ |
| | 4705 | ldr texture_page_ptr, [ psx_gpu, #psx_gpu_texture_page_ptr_offset ] \ |
| | 4706 | |
| | 4707 | #define setup_sprite_column_data_multi() \ |
| | 4708 | and height_rounded, height_rounded, #0xF; \ |
| | 4709 | rsb column_data, offset_v, #16; \ |
| | 4710 | \ |
| | 4711 | add height_rounded, height_rounded, #1; \ |
| | 4712 | sub tile_height, tile_height, #1; \ |
| | 4713 | \ |
| | 4714 | orr column_data, column_data, tile_height, lsl #16; \ |
| | 4715 | ldr texture_page_ptr, [ psx_gpu, #psx_gpu_texture_page_ptr_offset ]; \ |
| | 4716 | \ |
| | 4717 | orr column_data, column_data, height_rounded, lsl #8 \ |
| | 4718 | |
| | 4719 | #define setup_sprite_tile_column_width_single(texture_mode, multi_height, \ |
| | 4720 | edge_mode, edge) \ |
| | 4721 | setup_sprite_##texture_mode##_single_##multi_height##_##edge_mode##_##edge: \ |
| | 4722 | setup_sprite_column_data_##multi_height(); \ |
| | 4723 | vext.32 block_masks_shifted, block_masks, block_masks, #1; \ |
| | 4724 | vorr.u32 block_masks, block_masks, block_masks_shifted; \ |
| | 4725 | vdup.u8 draw_mask_fb_ptr_left, block_masks[0]; \ |
| | 4726 | vdup.u8 draw_mask_fb_ptr_right, block_masks[1]; \ |
| | 4727 | \ |
| | 4728 | setup_sprite_tile_column_height_##multi_height(edge_mode, edge, \ |
| | 4729 | texture_mode); \ |
| | 4730 | ldmia sp!, { r4 - r11, pc } \ |
| | 4731 | |
| | 4732 | #define setup_sprite_tiled_advance_column() \ |
| | 4733 | add texture_offset_base, texture_offset_base, #0x100; \ |
| | 4734 | tst texture_offset_base, #0xF00; \ |
| | 4735 | subeq texture_offset_base, texture_offset_base, #(0x100 + 0xF00) \ |
| | 4736 | |
| | 4737 | #define setup_sprite_tile_column_width_multi(tm, multi_height, left_mode, \ |
| | 4738 | right_mode) \ |
| | 4739 | setup_sprite_##tm##_multi_##multi_height##_##left_mode##_##right_mode: \ |
| | 4740 | setup_sprite_column_data_##multi_height(); \ |
| | 4741 | mov fb_ptr_advance_column, #32; \ |
| | 4742 | \ |
| | 4743 | sub fb_ptr_advance_column, height, lsl #11; \ |
| | 4744 | vdup.u8 draw_mask_fb_ptr_left, block_masks[0]; \ |
| | 4745 | \ |
| | 4746 | vdup.u8 draw_mask_fb_ptr_right, block_masks[1]; \ |
| | 4747 | setup_sprite_tile_column_height_##multi_height(left_mode, right, tm); \ |
| | 4748 | \ |
| | 4749 | subs tile_width, tile_width, #2; \ |
| | 4750 | add fb_ptr, fb_ptr, fb_ptr_advance_column; \ |
| | 4751 | \ |
| | 4752 | vmov.u8 draw_masks_fb_ptrs, #0; \ |
| | 4753 | beq 1f; \ |
| | 4754 | \ |
| | 4755 | 0: \ |
| | 4756 | setup_sprite_tiled_advance_column(); \ |
| | 4757 | setup_sprite_tile_column_height_##multi_height(full, none, tm); \ |
| | 4758 | add fb_ptr, fb_ptr, fb_ptr_advance_column; \ |
| | 4759 | subs tile_width, tile_width, #1; \ |
| | 4760 | bne 0b; \ |
| | 4761 | \ |
| | 4762 | 1: \ |
| | 4763 | vdup.u8 draw_mask_fb_ptr_left, block_masks[4]; \ |
| | 4764 | vdup.u8 draw_mask_fb_ptr_right, block_masks[5]; \ |
| | 4765 | \ |
| | 4766 | setup_sprite_tiled_advance_column(); \ |
| | 4767 | setup_sprite_tile_column_height_##multi_height(right_mode, left, tm); \ |
| | 4768 | ldmia sp!, { r4 - r11, pc } \ |
| | 4769 | |
| | 4770 | |
| | 4771 | // r0: psx_gpu |
| | 4772 | // r1: x |
| | 4773 | // r2: y |
| | 4774 | // r3: u |
| | 4775 | // [ sp ]: v |
| | 4776 | // [ sp + 4 ]: width |
| | 4777 | // [ sp + 8 ]: height |
| | 4778 | // [ sp + 12 ]: color (unused) |
| | 4779 | |
| | 4780 | #define setup_sprite_tiled_builder(texture_mode) \ |
| | 4781 | \ |
| | 4782 | setup_sprite_tile_column_width_multi(texture_mode, multi, full, full); \ |
| | 4783 | setup_sprite_tile_column_width_single(texture_mode, multi, full, none); \ |
| | 4784 | setup_sprite_tile_column_width_multi(texture_mode, single, full, full); \ |
| | 4785 | setup_sprite_tile_column_width_single(texture_mode, single, full, none); \ |
| | 4786 | setup_sprite_tile_column_width_multi(texture_mode, multi, half, full); \ |
| | 4787 | setup_sprite_tile_column_width_single(texture_mode, multi, half, right); \ |
| | 4788 | setup_sprite_tile_column_width_multi(texture_mode, single, half, full); \ |
| | 4789 | setup_sprite_tile_column_width_single(texture_mode, single, half, right); \ |
| | 4790 | setup_sprite_tile_column_width_multi(texture_mode, multi, full, half); \ |
| | 4791 | setup_sprite_tile_column_width_single(texture_mode, multi, half, left); \ |
| | 4792 | setup_sprite_tile_column_width_multi(texture_mode, single, full, half); \ |
| | 4793 | setup_sprite_tile_column_width_single(texture_mode, single, half, left); \ |
| | 4794 | setup_sprite_tile_column_width_multi(texture_mode, multi, half, half); \ |
| | 4795 | setup_sprite_tile_column_width_multi(texture_mode, single, half, half); \ |
| | 4796 | \ |
| | 4797 | .align 4; \ |
| | 4798 | \ |
| | 4799 | function(setup_sprite_##texture_mode) \ |
| | 4800 | stmdb sp!, { r4 - r11, r14 }; \ |
| | 4801 | setup_sprite_tiled_initialize_##texture_mode(); \ |
| | 4802 | \ |
| | 4803 | ldr v, [ sp, #36 ]; \ |
| | 4804 | and offset_u, u, #0xF; \ |
| | 4805 | \ |
| | 4806 | ldr width, [ sp, #40 ]; \ |
| | 4807 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ]; \ |
| | 4808 | \ |
| | 4809 | ldr height, [ sp, #44 ]; \ |
| | 4810 | add fb_ptr, fb_ptr, y, lsl #11; \ |
| | 4811 | \ |
| | 4812 | add fb_ptr, fb_ptr, x, lsl #1; \ |
| | 4813 | and offset_v, v, #0xF; \ |
| | 4814 | \ |
| | 4815 | sub fb_ptr, fb_ptr, offset_u, lsl #1; \ |
| | 4816 | add width_rounded, offset_u, width; \ |
| | 4817 | \ |
| | 4818 | add height_rounded, offset_v, height; \ |
| | 4819 | add width_rounded, width_rounded, #15; \ |
| | 4820 | \ |
| | 4821 | add height_rounded, height_rounded, #15; \ |
| | 4822 | mov tile_width, width_rounded, lsr #4; \ |
| | 4823 | \ |
| | 4824 | /* texture_offset_base = VH-VL-00-00 */\ |
| | 4825 | mov texture_offset_base, v, lsl #8; \ |
| | 4826 | and offset_u_right, width_rounded, #0xF; \ |
| | 4827 | \ |
| | 4828 | /* texture_offset_base = VH-UH-UL-00 */\ |
| | 4829 | bfi texture_offset_base, u, #4, #8; \ |
| | 4830 | movw right_block_mask, #0xFFFE; \ |
| | 4831 | \ |
| | 4832 | /* texture_offset_base = VH-UH-VL-00 */\ |
| | 4833 | bfi texture_offset_base, v, #4, #4; \ |
| | 4834 | movw left_block_mask, #0xFFFF; \ |
| | 4835 | \ |
| | 4836 | mov tile_height, height_rounded, lsr #4; \ |
| | 4837 | mvn left_block_mask, left_block_mask, lsl offset_u; \ |
| | 4838 | \ |
| | 4839 | /* texture_mask = HH-HL-WH-WL */\ |
| | 4840 | ldrh texture_mask, [ psx_gpu, #psx_gpu_texture_mask_width_offset ]; \ |
| | 4841 | mov right_block_mask, right_block_mask, lsl offset_u_right; \ |
| | 4842 | \ |
| | 4843 | /* texture_mask_rev = WH-WL-HH-HL */\ |
| | 4844 | rev16 texture_mask_rev, texture_mask; \ |
| | 4845 | vmov block_masks, left_block_mask, right_block_mask; \ |
| | 4846 | \ |
| | 4847 | /* texture_mask = HH-HL-HL-WL */\ |
| | 4848 | bfi texture_mask, texture_mask_rev, #4, #4; \ |
| | 4849 | /* texture_mask_rev = 00-00-00-WH */\ |
| | 4850 | mov texture_mask_rev, texture_mask_rev, lsr #12; \ |
| | 4851 | \ |
| | 4852 | /* texture_mask = HH-WH-HL-WL */\ |
| | 4853 | bfi texture_mask, texture_mask_rev, #8, #4; \ |
| | 4854 | and left_block_mask, left_block_mask, #0xFF; \ |
| | 4855 | \ |
| | 4856 | mov control_mask, #0; \ |
| | 4857 | cmp left_block_mask, #0xFF; \ |
| | 4858 | \ |
| | 4859 | uxtb right_block_mask, right_block_mask, ror #8; \ |
| | 4860 | orreq control_mask, control_mask, #0x4; \ |
| | 4861 | \ |
| | 4862 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ]; \ |
| | 4863 | cmp right_block_mask, #0xFF; \ |
| | 4864 | \ |
| | 4865 | orreq control_mask, control_mask, #0x8; \ |
| | 4866 | cmp tile_width, #1; \ |
| | 4867 | \ |
| | 4868 | add block, psx_gpu, #psx_gpu_blocks_offset; \ |
| | 4869 | orreq control_mask, control_mask, #0x1; \ |
| | 4870 | \ |
| | 4871 | cmp tile_height, #1; \ |
| | 4872 | add block, block, num_blocks, lsl #6; \ |
| | 4873 | \ |
| | 4874 | orreq control_mask, control_mask, #0x2; \ |
| | 4875 | ldr pc, [ pc, control_mask, lsl #2 ]; \ |
| | 4876 | nop; \ |
| | 4877 | \ |
| | 4878 | .word setup_sprite_##texture_mode##_multi_multi_full_full; \ |
| | 4879 | .word setup_sprite_##texture_mode##_single_multi_full_none; \ |
| | 4880 | .word setup_sprite_##texture_mode##_multi_single_full_full; \ |
| | 4881 | .word setup_sprite_##texture_mode##_single_single_full_none; \ |
| | 4882 | .word setup_sprite_##texture_mode##_multi_multi_half_full; \ |
| | 4883 | .word setup_sprite_##texture_mode##_single_multi_half_right; \ |
| | 4884 | .word setup_sprite_##texture_mode##_multi_single_half_full; \ |
| | 4885 | .word setup_sprite_##texture_mode##_single_single_half_right; \ |
| | 4886 | .word setup_sprite_##texture_mode##_multi_multi_full_half; \ |
| | 4887 | .word setup_sprite_##texture_mode##_single_multi_half_left; \ |
| | 4888 | .word setup_sprite_##texture_mode##_multi_single_full_half; \ |
| | 4889 | .word setup_sprite_##texture_mode##_single_single_half_left; \ |
| | 4890 | .word setup_sprite_##texture_mode##_multi_multi_half_half; \ |
| | 4891 | .word 0x00000000; \ |
| | 4892 | .word setup_sprite_##texture_mode##_multi_single_half_half \ |
| | 4893 | |
| | 4894 | |
| | 4895 | setup_sprite_tiled_builder(4bpp); |
| | 4896 | setup_sprite_tiled_builder(8bpp); |
| | 4897 | |
| | 4898 | |
| | 4899 | #undef block_ptr |
| | 4900 | #undef num_blocks |
| | 4901 | #undef clut_ptr |
| | 4902 | |
| | 4903 | #define psx_gpu r0 |
| | 4904 | #define block_ptr r0 |
| | 4905 | #define num_blocks r1 |
| | 4906 | #define clut_ptr r2 |
| | 4907 | #define texel_shift_mask r3 |
| | 4908 | #define block_pixels_a r4 |
| | 4909 | #define block_pixels_b r5 |
| | 4910 | #define texel_0 r6 |
| | 4911 | #define texel_2 r7 |
| | 4912 | #define texel_4 r8 |
| | 4913 | #define texel_6 r9 |
| | 4914 | #define texel_1 r10 |
| | 4915 | #define texel_3 r11 |
| | 4916 | #define texel_5 r12 |
| | 4917 | #define texel_7 r14 |
| | 4918 | #define texels_01 r6 |
| | 4919 | #define texels_23 r7 |
| | 4920 | #define texels_45 r8 |
| | 4921 | #define texels_67 r9 |
| | 4922 | |
| | 4923 | function(texture_sprite_blocks_8bpp) |
| | 4924 | stmdb sp!, { r4 - r11, r14 } |
| | 4925 | movw texel_shift_mask, #(0xFF << 1) |
| | 4926 | |
| | 4927 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 4928 | ldr clut_ptr, [ psx_gpu, #psx_gpu_clut_ptr_offset ] |
| | 4929 | |
| | 4930 | add block_ptr, psx_gpu, #psx_gpu_blocks_offset |
| | 4931 | ldr block_pixels_a, [ block_ptr, #16 ] |
| | 4932 | |
| | 4933 | 0: |
| | 4934 | and texel_0, texel_shift_mask, block_pixels_a, lsl #1 |
| | 4935 | ldr block_pixels_b, [ block_ptr, #20 ] |
| | 4936 | |
| | 4937 | and texel_1, texel_shift_mask, block_pixels_a, lsr #7 |
| | 4938 | ldrh texel_0, [ clut_ptr, texel_0 ] |
| | 4939 | |
| | 4940 | and texel_2, texel_shift_mask, block_pixels_a, lsr #15 |
| | 4941 | ldrh texel_1, [ clut_ptr, texel_1 ] |
| | 4942 | |
| | 4943 | and texel_3, texel_shift_mask, block_pixels_a, lsr #23 |
| | 4944 | ldr block_pixels_a, [ block_ptr, #(64 + 16) ] |
| | 4945 | |
| | 4946 | ldrh texel_2, [ clut_ptr, texel_2 ] |
| | 4947 | and texel_4, texel_shift_mask, block_pixels_b, lsl #1 |
| | 4948 | |
| | 4949 | ldrh texel_3, [ clut_ptr, texel_3 ] |
| | 4950 | and texel_5, texel_shift_mask, block_pixels_b, lsr #7 |
| | 4951 | |
| | 4952 | ldrh texel_4, [ clut_ptr, texel_4 ] |
| | 4953 | and texel_6, texel_shift_mask, block_pixels_b, lsr #15 |
| | 4954 | |
| | 4955 | ldrh texel_5, [ clut_ptr, texel_5 ] |
| | 4956 | and texel_7, texel_shift_mask, block_pixels_b, lsr #23 |
| | 4957 | |
| | 4958 | ldrh texel_6, [ clut_ptr, texel_6 ] |
| | 4959 | orr texels_01, texel_0, texel_1, lsl #16 |
| | 4960 | |
| | 4961 | ldrh texel_7, [ clut_ptr, texel_7 ] |
| | 4962 | orr texels_23, texel_2, texel_3, lsl #16 |
| | 4963 | |
| | 4964 | orr texels_45, texel_4, texel_5, lsl #16 |
| | 4965 | str texels_01, [ block_ptr, #0 ] |
| | 4966 | |
| | 4967 | orr texels_67, texel_6, texel_7, lsl #16 |
| | 4968 | str texels_23, [ block_ptr, #4 ] |
| | 4969 | |
| | 4970 | subs num_blocks, num_blocks, #1 |
| | 4971 | str texels_45, [ block_ptr, #8 ] |
| | 4972 | |
| | 4973 | str texels_67, [ block_ptr, #12 ] |
| | 4974 | add block_ptr, block_ptr, #64 |
| | 4975 | |
| | 4976 | bne 0b |
| | 4977 | |
| | 4978 | ldmia sp!, { r4 - r11, pc } |
| | 4979 | |
| | 4980 | |
| | 4981 | #undef width_rounded |
| | 4982 | #undef texture_mask |
| | 4983 | #undef num_blocks |
| | 4984 | #undef texture_offset |
| | 4985 | |
| | 4986 | #define psx_gpu r0 |
| | 4987 | #define x r1 |
| | 4988 | #define y r2 |
| | 4989 | #define u r3 |
| | 4990 | #define v r4 |
| | 4991 | #define width r5 |
| | 4992 | #define height r6 |
| | 4993 | #define left_offset r8 |
| | 4994 | #define width_rounded r9 |
| | 4995 | #define right_width r10 |
| | 4996 | #define block_width r11 |
| | 4997 | |
| | 4998 | #define texture_offset_base r1 |
| | 4999 | #define texture_mask r2 |
| | 5000 | #define texture_page_ptr r3 |
| | 5001 | #define num_blocks r4 |
| | 5002 | #define block r5 |
| | 5003 | #define fb_ptr r7 |
| | 5004 | #define texture_offset r8 |
| | 5005 | #define blocks_remaining r9 |
| | 5006 | #define fb_ptr_pitch r12 |
| | 5007 | #define texture_block_ptr r14 |
| | 5008 | |
| | 5009 | #define texture_mask_width r2 |
| | 5010 | #define texture_mask_height r3 |
| | 5011 | #define left_mask_bits r4 |
| | 5012 | #define right_mask_bits r5 |
| | 5013 | |
| | 5014 | |
| | 5015 | #undef block_masks |
| | 5016 | #undef block_masks_shifted |
| | 5017 | #undef texels |
| | 5018 | |
| | 5019 | #define block_masks d0 |
| | 5020 | #define block_masks_shifted d1 |
| | 5021 | #define draw_mask_fb_ptr d2 |
| | 5022 | #define texels q2 |
| | 5023 | |
| | 5024 | |
| | 5025 | setup_sprites_16bpp_flush_single: |
| | 5026 | vpush { d0 - d2 } |
| | 5027 | |
| | 5028 | stmdb sp!, { r0 - r3, r12, r14 } |
| | 5029 | bl flush_render_block_buffer |
| | 5030 | ldmia sp!, { r0 - r3, r12, r14 } |
| | 5031 | |
| | 5032 | vpop { d0 - d2 } |
| | 5033 | |
| | 5034 | add block, psx_gpu, #psx_gpu_blocks_offset |
| | 5035 | mov num_blocks, #1 |
| | 5036 | |
| | 5037 | bx lr |
| | 5038 | |
| | 5039 | setup_sprites_16bpp_flush_row: |
| | 5040 | vpush { d0 - d2 } |
| | 5041 | |
| | 5042 | stmdb sp!, { r0 - r3, r12, r14 } |
| | 5043 | bl flush_render_block_buffer |
| | 5044 | ldmia sp!, { r0 - r3, r12, r14 } |
| | 5045 | |
| | 5046 | vpop { d0 - d2 } |
| | 5047 | |
| | 5048 | add block, psx_gpu, #psx_gpu_blocks_offset |
| | 5049 | mov num_blocks, block_width |
| | 5050 | |
| | 5051 | bx lr |
| | 5052 | |
| | 5053 | function(setup_sprite_16bpp) |
| | 5054 | stmdb sp!, { r4 - r11, r14 } |
| | 5055 | ldr fb_ptr, [ psx_gpu, #psx_gpu_vram_ptr_offset ] |
| | 5056 | |
| | 5057 | ldr v, [ sp, #36 ] |
| | 5058 | add fb_ptr, fb_ptr, y, lsl #11 |
| | 5059 | |
| | 5060 | ldr width, [ sp, #40 ] |
| | 5061 | add fb_ptr, fb_ptr, x, lsl #1 |
| | 5062 | |
| | 5063 | ldr height, [ sp, #44 ] |
| | 5064 | and left_offset, u, #0x7 |
| | 5065 | |
| | 5066 | add texture_offset_base, u, u |
| | 5067 | add width_rounded, width, #7 |
| | 5068 | |
| | 5069 | add texture_offset_base, v, lsl #11 |
| | 5070 | mov left_mask_bits, #0xFF |
| | 5071 | |
| | 5072 | ldrb texture_mask_width, [ psx_gpu, #psx_gpu_texture_mask_width_offset ] |
| | 5073 | add width_rounded, width_rounded, left_offset |
| | 5074 | |
| | 5075 | ldrb texture_mask_height, [ psx_gpu, #psx_gpu_texture_mask_height_offset ] |
| | 5076 | sub fb_ptr, fb_ptr, left_offset, lsl #1 |
| | 5077 | |
| | 5078 | add texture_mask, texture_mask_width, texture_mask_width |
| | 5079 | mov right_mask_bits, #0xFE |
| | 5080 | |
| | 5081 | and right_width, width_rounded, #0x7 |
| | 5082 | mvn left_mask_bits, left_mask_bits, lsl left_offset |
| | 5083 | |
| | 5084 | add texture_mask, texture_mask_height, lsl #11 |
| | 5085 | mov block_width, width_rounded, lsr #3 |
| | 5086 | |
| | 5087 | mov right_mask_bits, right_mask_bits, lsl right_width |
| | 5088 | movw fb_ptr_pitch, #(2048 + 16) |
| | 5089 | |
| | 5090 | sub fb_ptr_pitch, fb_ptr_pitch, block_width, lsl #4 |
| | 5091 | vmov block_masks, left_mask_bits, right_mask_bits |
| | 5092 | |
| | 5093 | ldrh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 5094 | add block, psx_gpu, #psx_gpu_blocks_offset |
| | 5095 | |
| | 5096 | bic texture_offset_base, texture_offset_base, #0xF |
| | 5097 | cmp block_width, #1 |
| | 5098 | |
| | 5099 | ldr texture_page_ptr, [ psx_gpu, #psx_gpu_texture_page_ptr_offset ] |
| | 5100 | add block, block, num_blocks, lsl #6 |
| | 5101 | |
| | 5102 | bne 0f |
| | 5103 | |
| | 5104 | vext.32 block_masks_shifted, block_masks, block_masks, #1 |
| | 5105 | vorr.u32 block_masks, block_masks, block_masks_shifted |
| | 5106 | vdup.u8 draw_mask_fb_ptr, block_masks[0] |
| | 5107 | |
| | 5108 | 1: |
| | 5109 | add num_blocks, num_blocks, #1 |
| | 5110 | cmp num_blocks, #MAX_BLOCKS |
| | 5111 | blgt setup_sprites_16bpp_flush_single |
| | 5112 | |
| | 5113 | and texture_block_ptr, texture_offset_base, texture_mask |
| | 5114 | subs height, height, #1 |
| | 5115 | |
| | 5116 | add texture_block_ptr, texture_page_ptr, texture_block_ptr |
| | 5117 | vld1.u32 { texels }, [ texture_block_ptr, :128 ] |
| | 5118 | |
| | 5119 | vst1.u32 { texels }, [ block, :128 ] |
| | 5120 | add block, block, #40 |
| | 5121 | |
| | 5122 | vmov.u32 draw_mask_fb_ptr[1], fb_ptr |
| | 5123 | pld [ fb_ptr ] |
| | 5124 | |
| | 5125 | vst1.u32 { draw_mask_fb_ptr }, [ block, :64 ] |
| | 5126 | |
| | 5127 | add block, block, #24 |
| | 5128 | add texture_offset_base, texture_offset_base, #2048 |
| | 5129 | add fb_ptr, fb_ptr, #2048 |
| | 5130 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 5131 | bne 1b |
| | 5132 | |
| | 5133 | ldmia sp!, { r4 - r11, pc } |
| | 5134 | |
| | 5135 | 0: |
| | 5136 | add num_blocks, num_blocks, block_width |
| | 5137 | mov texture_offset, texture_offset_base |
| | 5138 | |
| | 5139 | cmp num_blocks, #MAX_BLOCKS |
| | 5140 | blgt setup_sprites_16bpp_flush_row |
| | 5141 | |
| | 5142 | add texture_offset_base, texture_offset_base, #2048 |
| | 5143 | and texture_block_ptr, texture_offset, texture_mask |
| | 5144 | |
| | 5145 | add texture_block_ptr, texture_page_ptr, texture_block_ptr |
| | 5146 | vld1.u32 { texels }, [ texture_block_ptr, :128 ] |
| | 5147 | |
| | 5148 | vst1.u32 { texels }, [ block, :128 ] |
| | 5149 | add block, block, #40 |
| | 5150 | |
| | 5151 | vdup.u8 draw_mask_fb_ptr, block_masks[0] |
| | 5152 | vmov.u32 draw_mask_fb_ptr[1], fb_ptr |
| | 5153 | pld [ fb_ptr ] |
| | 5154 | |
| | 5155 | vst1.u32 { draw_mask_fb_ptr }, [ block, :64 ] |
| | 5156 | subs blocks_remaining, block_width, #2 |
| | 5157 | |
| | 5158 | add texture_offset, texture_offset, #16 |
| | 5159 | add fb_ptr, fb_ptr, #16 |
| | 5160 | |
| | 5161 | vmov.u8 draw_mask_fb_ptr, #0 |
| | 5162 | |
| | 5163 | add block, block, #24 |
| | 5164 | beq 2f |
| | 5165 | |
| | 5166 | 1: |
| | 5167 | and texture_block_ptr, texture_offset, texture_mask |
| | 5168 | subs blocks_remaining, blocks_remaining, #1 |
| | 5169 | |
| | 5170 | add texture_block_ptr, texture_page_ptr, texture_block_ptr |
| | 5171 | vld1.u32 { texels }, [ texture_block_ptr, :128 ] |
| | 5172 | |
| | 5173 | vst1.u32 { texels }, [ block, :128 ] |
| | 5174 | add block, block, #40 |
| | 5175 | |
| | 5176 | vmov.u32 draw_mask_fb_ptr[1], fb_ptr |
| | 5177 | pld [ fb_ptr ] |
| | 5178 | |
| | 5179 | vst1.u32 { draw_mask_fb_ptr }, [ block, :64 ] |
| | 5180 | |
| | 5181 | add texture_offset, texture_offset, #16 |
| | 5182 | add fb_ptr, fb_ptr, #16 |
| | 5183 | |
| | 5184 | add block, block, #24 |
| | 5185 | bne 1b |
| | 5186 | |
| | 5187 | 2: |
| | 5188 | and texture_block_ptr, texture_offset, texture_mask |
| | 5189 | add texture_block_ptr, texture_page_ptr, texture_block_ptr |
| | 5190 | |
| | 5191 | vld1.u32 { texels }, [ texture_block_ptr, :128 ] |
| | 5192 | vdup.u8 draw_mask_fb_ptr, block_masks[4] |
| | 5193 | |
| | 5194 | vst1.u32 { texels }, [ block, :128 ] |
| | 5195 | add block, block, #40 |
| | 5196 | |
| | 5197 | vmov.u32 draw_mask_fb_ptr[1], fb_ptr |
| | 5198 | vst1.u32 { draw_mask_fb_ptr }, [ block, :64 ] |
| | 5199 | |
| | 5200 | add block, block, #24 |
| | 5201 | subs height, height, #1 |
| | 5202 | |
| | 5203 | add fb_ptr, fb_ptr, fb_ptr_pitch |
| | 5204 | strh num_blocks, [ psx_gpu, #psx_gpu_num_blocks_offset ] |
| | 5205 | |
| | 5206 | bne 0b |
| | 5207 | |
| | 5208 | ldmia sp!, { r4 - r11, pc } |
| | 5209 | |
| | 5210 | |
| | 5211 | #undef texture_page_ptr |
| | 5212 | #undef vram_ptr |
| | 5213 | #undef dirty_textures_mask |
| | 5214 | #undef current_texture_mask |
| | 5215 | |
| | 5216 | #define psx_gpu r0 |
| | 5217 | #define current_texture_page r1 |
| | 5218 | #define texture_page_ptr r2 |
| | 5219 | #define vram_ptr_a r3 |
| | 5220 | #define current_texture_page_x r12 |
| | 5221 | #define current_texture_page_y r4 |
| | 5222 | #define dirty_textures_mask r5 |
| | 5223 | #define tile_y r6 |
| | 5224 | #define tile_x r7 |
| | 5225 | #define sub_y r8 |
| | 5226 | #define current_texture_mask r9 |
| | 5227 | #define c_4096 r10 |
| | 5228 | #define vram_ptr_b r11 |
| | 5229 | |
| | 5230 | #define texel_block_a d0 |
| | 5231 | #define texel_block_b d1 |
| | 5232 | #define texel_block_expanded_a q1 |
| | 5233 | #define texel_block_expanded_b q2 |
| | 5234 | #define texel_block_expanded_ab q2 |
| | 5235 | #define texel_block_expanded_c q3 |
| | 5236 | #define texel_block_expanded_d q4 |
| | 5237 | #define texel_block_expanded_cd q3 |
| | 5238 | |
| | 5239 | function(update_texture_4bpp_cache) |
| | 5240 | stmdb sp!, { r4 - r11, r14 } |
| | 5241 | vpush { q0 - q3 } |
| | 5242 | |
| | 5243 | ldrb current_texture_page, [ psx_gpu, #psx_gpu_current_texture_page_offset ] |
| | 5244 | |
| | 5245 | ldr texture_page_ptr, [ psx_gpu, #psx_gpu_texture_page_base_offset ] |
| | 5246 | ldr vram_ptr_a, [ psx_gpu, #psx_gpu_vram_ptr_offset ] |
| | 5247 | |
| | 5248 | and current_texture_page_x, current_texture_page, #0xF |
| | 5249 | ldr current_texture_mask, [ psx_gpu, #psx_gpu_current_texture_mask_offset ] |
| | 5250 | |
| | 5251 | mov current_texture_page_y, current_texture_page, lsr #4 |
| | 5252 | ldr dirty_textures_mask, [ psx_gpu, #psx_gpu_dirty_textures_4bpp_mask_offset ] |
| | 5253 | |
| | 5254 | add vram_ptr_a, vram_ptr_a, current_texture_page_y, lsl #19 |
| | 5255 | mov tile_y, #16 |
| | 5256 | |
| | 5257 | add vram_ptr_a, vram_ptr_a, current_texture_page_x, lsl #7 |
| | 5258 | bic dirty_textures_mask, current_texture_mask |
| | 5259 | |
| | 5260 | mov tile_x, #16 |
| | 5261 | str dirty_textures_mask, [ psx_gpu, #psx_gpu_dirty_textures_4bpp_mask_offset ] |
| | 5262 | |
| | 5263 | mov sub_y, #8 |
| | 5264 | movw c_4096, #4096 |
| | 5265 | |
| | 5266 | add vram_ptr_b, vram_ptr_a, #2048 |
| | 5267 | |
| | 5268 | 0: |
| | 5269 | vld1.u32 { texel_block_a }, [ vram_ptr_a, :64 ], c_4096 |
| | 5270 | vld1.u32 { texel_block_b }, [ vram_ptr_b, :64 ], c_4096 |
| | 5271 | |
| | 5272 | vmovl.u8 texel_block_expanded_a, texel_block_a |
| | 5273 | vshll.u8 texel_block_expanded_b, texel_block_a, #4 |
| | 5274 | vmovl.u8 texel_block_expanded_c, texel_block_b |
| | 5275 | vshll.u8 texel_block_expanded_d, texel_block_b, #4 |
| | 5276 | |
| | 5277 | vbic.u16 texel_block_expanded_a, #0x00F0 |
| | 5278 | vbic.u16 texel_block_expanded_b, #0x00F0 |
| | 5279 | vbic.u16 texel_block_expanded_c, #0x00F0 |
| | 5280 | vbic.u16 texel_block_expanded_d, #0x00F0 |
| | 5281 | |
| | 5282 | vorr.u16 texel_block_expanded_ab, texel_block_expanded_a, \ |
| | 5283 | texel_block_expanded_b |
| | 5284 | vorr.u16 texel_block_expanded_cd, texel_block_expanded_c, \ |
| | 5285 | texel_block_expanded_d |
| | 5286 | |
| | 5287 | vst1.u32 { texel_block_expanded_ab, texel_block_expanded_cd }, \ |
| | 5288 | [ texture_page_ptr, :256 ]! |
| | 5289 | |
| | 5290 | subs sub_y, sub_y, #1 |
| | 5291 | bne 0b |
| | 5292 | |
| | 5293 | mov sub_y, #8 |
| | 5294 | add vram_ptr_a, vram_ptr_a, #8 |
| | 5295 | add vram_ptr_b, vram_ptr_b, #8 |
| | 5296 | |
| | 5297 | sub vram_ptr_a, vram_ptr_a, #(16 * 2048) |
| | 5298 | sub vram_ptr_b, vram_ptr_b, #(16 * 2048) |
| | 5299 | |
| | 5300 | subs tile_x, tile_x, #1 |
| | 5301 | bne 0b |
| | 5302 | |
| | 5303 | mov tile_x, #16 |
| | 5304 | add vram_ptr_a, vram_ptr_a, #(16 * 2048) |
| | 5305 | add vram_ptr_b, vram_ptr_b, #(16 * 2048) |
| | 5306 | |
| | 5307 | sub vram_ptr_a, vram_ptr_a, #(8 * 16) |
| | 5308 | sub vram_ptr_b, vram_ptr_b, #(8 * 16) |
| | 5309 | |
| | 5310 | subs tile_y, tile_y, #1 |
| | 5311 | bne 0b |
| | 5312 | |
| | 5313 | vpop { q0 - q3 } |
| | 5314 | ldmia sp!, { r4 - r11, pc } |
| | 5315 | |
| | 5316 | |
| | 5317 | #undef current_texture_page |
| | 5318 | |
| | 5319 | #define psx_gpu r0 |
| | 5320 | #define texture_page r1 |
| | 5321 | #define texture_page_ptr r2 |
| | 5322 | #define vram_ptr_a r3 |
| | 5323 | #define texture_page_x r12 |
| | 5324 | #define texture_page_y r4 |
| | 5325 | #define current_texture_page r5 |
| | 5326 | #define tile_y r6 |
| | 5327 | #define tile_x r7 |
| | 5328 | #define sub_y r8 |
| | 5329 | #define c_4096 r10 |
| | 5330 | #define vram_ptr_b r11 |
| | 5331 | |
| | 5332 | |
| | 5333 | #undef texels_a |
| | 5334 | #undef texels_b |
| | 5335 | |
| | 5336 | #define texels_a q0 |
| | 5337 | #define texels_b q1 |
| | 5338 | #define texels_c q2 |
| | 5339 | #define texels_d q3 |
| | 5340 | |
| | 5341 | |
| | 5342 | function(update_texture_8bpp_cache_slice) |
| | 5343 | stmdb sp!, { r4 - r11, r14 } |
| | 5344 | vpush { q0 - q3 } |
| | 5345 | |
| | 5346 | ldrb current_texture_page, [ psx_gpu, #psx_gpu_current_texture_page_offset ] |
| | 5347 | ldr vram_ptr_a, [ psx_gpu, #psx_gpu_vram_ptr_offset ] |
| | 5348 | |
| | 5349 | ldr texture_page_ptr, [ psx_gpu, #psx_gpu_texture_page_base_offset ] |
| | 5350 | mov tile_y, #16 |
| | 5351 | |
| | 5352 | and texture_page_x, texture_page, #0xF |
| | 5353 | mov texture_page_y, texture_page, lsr #4 |
| | 5354 | |
| | 5355 | add vram_ptr_a, vram_ptr_a, texture_page_x, lsl #7 |
| | 5356 | mov tile_x, #8 |
| | 5357 | |
| | 5358 | add vram_ptr_a, vram_ptr_a, texture_page_y, lsl #19 |
| | 5359 | eor current_texture_page, current_texture_page, texture_page |
| | 5360 | |
| | 5361 | ands current_texture_page, current_texture_page, #0x1 |
| | 5362 | mov sub_y, #4 |
| | 5363 | |
| | 5364 | addne texture_page_ptr, texture_page_ptr, #(8 * 16 * 16) |
| | 5365 | movw c_4096, #4096 |
| | 5366 | |
| | 5367 | add vram_ptr_b, vram_ptr_a, #2048 |
| | 5368 | |
| | 5369 | 0: |
| | 5370 | vld1.u32 { texels_a }, [ vram_ptr_a, :128 ], c_4096 |
| | 5371 | vld1.u32 { texels_b }, [ vram_ptr_b, :128 ], c_4096 |
| | 5372 | vld1.u32 { texels_c }, [ vram_ptr_a, :128 ], c_4096 |
| | 5373 | vld1.u32 { texels_d }, [ vram_ptr_b, :128 ], c_4096 |
| | 5374 | |
| | 5375 | vst1.u32 { texels_a, texels_b }, [ texture_page_ptr, :256 ]! |
| | 5376 | vst1.u32 { texels_c, texels_d }, [ texture_page_ptr, :256 ]! |
| | 5377 | |
| | 5378 | subs sub_y, sub_y, #1 |
| | 5379 | bne 0b |
| | 5380 | |
| | 5381 | mov sub_y, #4 |
| | 5382 | |
| | 5383 | add vram_ptr_a, vram_ptr_a, #16 |
| | 5384 | add vram_ptr_b, vram_ptr_b, #16 |
| | 5385 | |
| | 5386 | sub vram_ptr_a, vram_ptr_a, #(16 * 2048) |
| | 5387 | sub vram_ptr_b, vram_ptr_b, #(16 * 2048) |
| | 5388 | |
| | 5389 | subs tile_x, tile_x, #1 |
| | 5390 | bne 0b |
| | 5391 | |
| | 5392 | mov tile_x, #8 |
| | 5393 | |
| | 5394 | add vram_ptr_a, vram_ptr_a, #(16 * 2048) |
| | 5395 | add vram_ptr_b, vram_ptr_b, #(16 * 2048) |
| | 5396 | |
| | 5397 | sub vram_ptr_a, vram_ptr_a, #(8 * 16) |
| | 5398 | sub vram_ptr_b, vram_ptr_b, #(8 * 16) |
| | 5399 | |
| | 5400 | subs tile_y, tile_y, #1 |
| | 5401 | add texture_page_ptr, texture_page_ptr, #(8 * 16 * 16) |
| | 5402 | |
| | 5403 | bne 0b |
| | 5404 | |
| | 5405 | vpop { q0 - q3 } |
| | 5406 | ldmia sp!, { r4 - r11, pc } |
| | 5407 | |
notaz.gp2x.de / pcsx_rearmed.git / blame_incremental