| 1 |
/* |
| 2 |
* |
| 3 |
* Copyright (C) 2004 David Airlie All Rights Reserved. |
| 4 |
* |
| 5 |
* Permission is hereby granted, free of charge, to any person obtaining a |
| 6 |
* copy of this software and associated documentation files (the "Software"), |
| 7 |
* to deal in the Software without restriction, including without limitation |
| 8 |
* the rights to use, copy, modify, merge, publish, distribute, sublicense, |
| 9 |
* and/or sell copies of the Software, and to permit persons to whom the |
| 10 |
* Software is furnished to do so, subject to the following conditions: |
| 11 |
* |
| 12 |
* The above copyright notice and this permission notice shall be included |
| 13 |
* in all copies or substantial portions of the Software. |
| 14 |
* |
| 15 |
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS |
| 16 |
* OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 17 |
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL |
| 18 |
* DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN |
| 19 |
* AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN |
| 20 |
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. |
| 21 |
*/ |
| 22 |
|
| 23 |
#include "glheader.h" |
| 24 |
#include "colormac.h" |
| 25 |
#include "context.h" |
| 26 |
#include "atifragshader.h" |
| 27 |
#include "macros.h" |
| 28 |
#include "program.h" |
| 29 |
|
| 30 |
#include "s_atifragshader.h" |
| 31 |
#include "s_nvfragprog.h" |
| 32 |
#include "s_span.h" |
| 33 |
#include "s_texture.h" |
| 34 |
|
| 35 |
/** |
| 36 |
* Fetch a texel. |
| 37 |
*/ |
| 38 |
static void |
| 39 |
fetch_texel(GLcontext * ctx, const GLfloat texcoord[4], GLfloat lambda, |
| 40 |
GLuint unit, GLfloat color[4]) |
| 41 |
{ |
| 42 |
GLchan rgba[4]; |
| 43 |
SWcontext *swrast = SWRAST_CONTEXT(ctx); |
| 44 |
|
| 45 |
/* XXX use a float-valued TextureSample routine here!!! */ |
| 46 |
swrast->TextureSample[unit] (ctx, unit, ctx->Texture.Unit[unit]._Current, |
| 47 |
1, (const GLfloat(*)[4]) texcoord, |
| 48 |
&lambda, &rgba); |
| 49 |
color[0] = CHAN_TO_FLOAT(rgba[0]); |
| 50 |
color[1] = CHAN_TO_FLOAT(rgba[1]); |
| 51 |
color[2] = CHAN_TO_FLOAT(rgba[2]); |
| 52 |
color[3] = CHAN_TO_FLOAT(rgba[3]); |
| 53 |
} |
| 54 |
|
| 55 |
static void |
| 56 |
apply_swizzle(struct atifs_machine *machine, GLuint reg, GLuint swizzle) |
| 57 |
{ |
| 58 |
GLfloat s, t, r, q; |
| 59 |
|
| 60 |
s = machine->Registers[reg][0]; |
| 61 |
t = machine->Registers[reg][1]; |
| 62 |
r = machine->Registers[reg][2]; |
| 63 |
q = machine->Registers[reg][3]; |
| 64 |
|
| 65 |
switch (swizzle) { |
| 66 |
case GL_SWIZZLE_STR_ATI: |
| 67 |
machine->Registers[reg][0] = s; |
| 68 |
machine->Registers[reg][1] = t; |
| 69 |
machine->Registers[reg][2] = r; |
| 70 |
break; |
| 71 |
case GL_SWIZZLE_STQ_ATI: |
| 72 |
machine->Registers[reg][0] = s; |
| 73 |
machine->Registers[reg][1] = t; |
| 74 |
machine->Registers[reg][2] = q; |
| 75 |
break; |
| 76 |
case GL_SWIZZLE_STR_DR_ATI: |
| 77 |
machine->Registers[reg][0] = s / r; |
| 78 |
machine->Registers[reg][1] = t / r; |
| 79 |
machine->Registers[reg][2] = 1 / r; |
| 80 |
break; |
| 81 |
case GL_SWIZZLE_STQ_DQ_ATI: |
| 82 |
machine->Registers[reg][0] = s / q; |
| 83 |
machine->Registers[reg][1] = t / q; |
| 84 |
machine->Registers[reg][2] = 1 / q; |
| 85 |
break; |
| 86 |
} |
| 87 |
machine->Registers[reg][3] = 0.0; |
| 88 |
} |
| 89 |
|
| 90 |
static void |
| 91 |
apply_src_rep(GLint optype, GLuint rep, GLfloat * val) |
| 92 |
{ |
| 93 |
GLint i; |
| 94 |
GLint start, end; |
| 95 |
if (!rep) |
| 96 |
return; |
| 97 |
|
| 98 |
start = optype ? 3 : 0; |
| 99 |
end = optype ? 4 : 3; |
| 100 |
|
| 101 |
for (i = start; i < end; i++) { |
| 102 |
switch (rep) { |
| 103 |
case GL_RED: |
| 104 |
val[i] = val[0]; |
| 105 |
break; |
| 106 |
case GL_GREEN: |
| 107 |
val[i] = val[1]; |
| 108 |
break; |
| 109 |
case GL_BLUE: |
| 110 |
val[i] = val[2]; |
| 111 |
break; |
| 112 |
case GL_ALPHA: |
| 113 |
val[i] = val[3]; |
| 114 |
break; |
| 115 |
} |
| 116 |
} |
| 117 |
} |
| 118 |
|
| 119 |
static void |
| 120 |
apply_src_mod(GLint optype, GLuint mod, GLfloat * val) |
| 121 |
{ |
| 122 |
GLint i; |
| 123 |
GLint start, end; |
| 124 |
|
| 125 |
if (!mod) |
| 126 |
return; |
| 127 |
|
| 128 |
start = optype ? 3 : 0; |
| 129 |
end = optype ? 4 : 3; |
| 130 |
|
| 131 |
for (i = start; i < end; i++) { |
| 132 |
if (mod & GL_COMP_BIT_ATI) |
| 133 |
val[i] = 1 - val[i]; |
| 134 |
|
| 135 |
if (mod & GL_BIAS_BIT_ATI) |
| 136 |
val[i] = val[i] - 0.5; |
| 137 |
|
| 138 |
if (mod & GL_2X_BIT_ATI) |
| 139 |
val[i] = 2 * val[i]; |
| 140 |
|
| 141 |
if (mod & GL_NEGATE_BIT_ATI) |
| 142 |
val[i] = -val[i]; |
| 143 |
} |
| 144 |
} |
| 145 |
|
| 146 |
static void |
| 147 |
apply_dst_mod(GLuint optype, GLuint mod, GLfloat * val) |
| 148 |
{ |
| 149 |
GLint i; |
| 150 |
GLint has_sat = mod & GL_SATURATE_BIT_ATI; |
| 151 |
GLint start, end; |
| 152 |
|
| 153 |
mod &= ~GL_SATURATE_BIT_ATI; |
| 154 |
|
| 155 |
start = optype ? 3 : 0; |
| 156 |
end = optype ? 4 : 3; |
| 157 |
|
| 158 |
for (i = start; i < end; i++) { |
| 159 |
switch (mod) { |
| 160 |
case GL_2X_BIT_ATI: |
| 161 |
val[i] = 2 * val[i]; |
| 162 |
break; |
| 163 |
case GL_4X_BIT_ATI: |
| 164 |
val[i] = 4 * val[i]; |
| 165 |
break; |
| 166 |
case GL_8X_BIT_ATI: |
| 167 |
val[i] = 8 * val[i]; |
| 168 |
break; |
| 169 |
case GL_HALF_BIT_ATI: |
| 170 |
val[i] = val[i] * 0.5; |
| 171 |
break; |
| 172 |
case GL_QUARTER_BIT_ATI: |
| 173 |
val[i] = val[i] * 0.25; |
| 174 |
break; |
| 175 |
case GL_EIGHTH_BIT_ATI: |
| 176 |
val[i] = val[i] * 0.125; |
| 177 |
break; |
| 178 |
} |
| 179 |
|
| 180 |
if (has_sat) { |
| 181 |
if (val[i] < 0.0) |
| 182 |
val[i] = 0; |
| 183 |
else if (val[i] > 1.0) |
| 184 |
val[i] = 1.0; |
| 185 |
} |
| 186 |
else { |
| 187 |
if (val[i] < -8.0) |
| 188 |
val[i] = -8.0; |
| 189 |
else if (val[i] > 8.0) |
| 190 |
val[i] = 8.0; |
| 191 |
} |
| 192 |
} |
| 193 |
} |
| 194 |
|
| 195 |
|
| 196 |
static void |
| 197 |
write_dst_addr(GLuint optype, GLuint mod, GLuint mask, GLfloat * src, |
| 198 |
GLfloat * dst) |
| 199 |
{ |
| 200 |
GLint i; |
| 201 |
apply_dst_mod(optype, mod, src); |
| 202 |
|
| 203 |
if (optype == ATI_FRAGMENT_SHADER_COLOR_OP) { |
| 204 |
if (mask) { |
| 205 |
if (mask & GL_RED_BIT_ATI) |
| 206 |
dst[0] = src[0]; |
| 207 |
|
| 208 |
if (mask & GL_GREEN_BIT_ATI) |
| 209 |
dst[1] = src[1]; |
| 210 |
|
| 211 |
if (mask & GL_BLUE_BIT_ATI) |
| 212 |
dst[2] = src[2]; |
| 213 |
} |
| 214 |
else { |
| 215 |
for (i = 0; i < 3; i++) |
| 216 |
dst[i] = src[i]; |
| 217 |
} |
| 218 |
} |
| 219 |
else |
| 220 |
dst[3] = src[3]; |
| 221 |
} |
| 222 |
|
| 223 |
static void |
| 224 |
finish_pass(struct atifs_machine *machine) |
| 225 |
{ |
| 226 |
GLint i; |
| 227 |
|
| 228 |
for (i = 0; i < 6; i++) { |
| 229 |
COPY_4V(machine->PrevPassRegisters[i], machine->Registers[i]); |
| 230 |
} |
| 231 |
} |
| 232 |
|
| 233 |
/** |
| 234 |
* Execute the given fragment shader |
| 235 |
* NOTE: we do everything in single-precision floating point; we don't |
| 236 |
* currently observe the single/half/fixed-precision qualifiers. |
| 237 |
* \param ctx - rendering context |
| 238 |
* \param program - the fragment program to execute |
| 239 |
* \param machine - machine state (register file) |
| 240 |
* \param maxInst - max number of instructions to execute |
| 241 |
* \return GL_TRUE if program completed or GL_FALSE if program executed KIL. |
| 242 |
*/ |
| 243 |
|
| 244 |
struct ati_fs_opcode_st ati_fs_opcodes[] = { |
| 245 |
{GL_ADD_ATI, 2}, |
| 246 |
{GL_SUB_ATI, 2}, |
| 247 |
{GL_MUL_ATI, 2}, |
| 248 |
{GL_MAD_ATI, 3}, |
| 249 |
{GL_LERP_ATI, 3}, |
| 250 |
{GL_MOV_ATI, 1}, |
| 251 |
{GL_CND_ATI, 3}, |
| 252 |
{GL_CND0_ATI, 3}, |
| 253 |
{GL_DOT2_ADD_ATI, 3}, |
| 254 |
{GL_DOT3_ATI, 2}, |
| 255 |
{GL_DOT4_ATI, 2} |
| 256 |
}; |
| 257 |
|
| 258 |
|
| 259 |
|
| 260 |
static void |
| 261 |
handle_pass_op(struct atifs_machine *machine, struct atifs_instruction *inst, |
| 262 |
const struct sw_span *span, GLuint column) |
| 263 |
{ |
| 264 |
GLuint idx = inst->DstReg[0].Index - GL_REG_0_ATI; |
| 265 |
GLuint swizzle = inst->DstReg[0].Swizzle; |
| 266 |
GLuint pass_tex = inst->SrcReg[0][0].Index; |
| 267 |
|
| 268 |
/* if we get here after passing pass one then we are starting pass two - backup the registers */ |
| 269 |
if (machine->pass == 1) { |
| 270 |
finish_pass(machine); |
| 271 |
machine->pass = 2; |
| 272 |
} |
| 273 |
if (pass_tex >= GL_TEXTURE0_ARB && pass_tex <= GL_TEXTURE7_ARB) { |
| 274 |
pass_tex -= GL_TEXTURE0_ARB; |
| 275 |
COPY_4V(machine->Registers[idx], |
| 276 |
span->array->texcoords[pass_tex][column]); |
| 277 |
} |
| 278 |
else if (pass_tex >= GL_REG_0_ATI && pass_tex <= GL_REG_5_ATI |
| 279 |
&& machine->pass == 2) { |
| 280 |
pass_tex -= GL_REG_0_ATI; |
| 281 |
COPY_4V(machine->Registers[idx], machine->PrevPassRegisters[pass_tex]); |
| 282 |
} |
| 283 |
apply_swizzle(machine, idx, swizzle); |
| 284 |
|
| 285 |
} |
| 286 |
|
| 287 |
static void |
| 288 |
handle_sample_op(GLcontext * ctx, struct atifs_machine *machine, |
| 289 |
struct atifs_instruction *inst, const struct sw_span *span, |
| 290 |
GLuint column) |
| 291 |
{ |
| 292 |
GLuint idx = inst->DstReg[0].Index - GL_REG_0_ATI; |
| 293 |
GLuint swizzle = inst->DstReg[0].Swizzle; |
| 294 |
GLuint sample_tex = inst->SrcReg[0][0].Index; |
| 295 |
|
| 296 |
/* if we get here after passing pass one then we are starting pass two - backup the registers */ |
| 297 |
if (machine->pass == 1) { |
| 298 |
finish_pass(machine); |
| 299 |
machine->pass = 2; |
| 300 |
} |
| 301 |
|
| 302 |
if (sample_tex >= GL_TEXTURE0_ARB && sample_tex <= GL_TEXTURE7_ARB) { |
| 303 |
sample_tex -= GL_TEXTURE0_ARB; |
| 304 |
fetch_texel(ctx, span->array->texcoords[sample_tex][column], 0.0F, |
| 305 |
sample_tex, machine->Registers[idx]); |
| 306 |
} |
| 307 |
else if (sample_tex >= GL_REG_0_ATI && sample_tex <= GL_REG_5_ATI) { |
| 308 |
/* this is wrong... */ |
| 309 |
sample_tex -= GL_REG_0_ATI; |
| 310 |
fetch_texel(ctx, machine->Registers[sample_tex], 0, sample_tex, |
| 311 |
machine->Registers[idx]); |
| 312 |
} |
| 313 |
|
| 314 |
apply_swizzle(machine, idx, swizzle); |
| 315 |
} |
| 316 |
|
| 317 |
#define SETUP_SRC_REG(optype, i, x) do { \ |
| 318 |
if (optype) \ |
| 319 |
src[optype][i][3] = x[3]; \ |
| 320 |
else \ |
| 321 |
COPY_3V(src[optype][i], x); \ |
| 322 |
} while (0) |
| 323 |
|
| 324 |
static GLboolean |
| 325 |
execute_shader(GLcontext * ctx, |
| 326 |
const struct ati_fragment_shader *shader, GLuint maxInst, |
| 327 |
struct atifs_machine *machine, const struct sw_span *span, |
| 328 |
GLuint column) |
| 329 |
{ |
| 330 |
GLuint pc; |
| 331 |
struct atifs_instruction *inst; |
| 332 |
GLint optype; |
| 333 |
GLint i; |
| 334 |
GLint dstreg; |
| 335 |
GLfloat src[2][3][4]; |
| 336 |
GLfloat zeros[4] = { 0.0, 0.0, 0.0, 0.0 }; |
| 337 |
GLfloat ones[4] = { 1.0, 1.0, 1.0, 1.0 }; |
| 338 |
GLfloat dst[2][4], *dstp; |
| 339 |
|
| 340 |
for (pc = 0; pc < shader->Base.NumInstructions; pc++) { |
| 341 |
inst = &shader->Instructions[pc]; |
| 342 |
|
| 343 |
if (inst->Opcode[0] == ATI_FRAGMENT_SHADER_PASS_OP) |
| 344 |
handle_pass_op(machine, inst, span, column); |
| 345 |
else if (inst->Opcode[0] == ATI_FRAGMENT_SHADER_SAMPLE_OP) |
| 346 |
handle_sample_op(ctx, machine, inst, span, column); |
| 347 |
else { |
| 348 |
if (machine->pass == 0) |
| 349 |
machine->pass = 1; |
| 350 |
|
| 351 |
/* setup the source registers for color and alpha ops */ |
| 352 |
for (optype = 0; optype < 2; optype++) { |
| 353 |
for (i = 0; i < inst->ArgCount[optype]; i++) { |
| 354 |
GLint index = inst->SrcReg[optype][i].Index; |
| 355 |
|
| 356 |
if (index >= GL_REG_0_ATI && index <= GL_REG_5_ATI) |
| 357 |
SETUP_SRC_REG(optype, i, |
| 358 |
machine->Registers[index - GL_REG_0_ATI]); |
| 359 |
else if (index >= GL_CON_0_ATI && index <= GL_CON_7_ATI) |
| 360 |
SETUP_SRC_REG(optype, i, |
| 361 |
shader->Constants[index - GL_CON_0_ATI]); |
| 362 |
else if (index == GL_ONE) |
| 363 |
SETUP_SRC_REG(optype, i, ones); |
| 364 |
else if (index == GL_ZERO) |
| 365 |
SETUP_SRC_REG(optype, i, zeros); |
| 366 |
else if (index == GL_PRIMARY_COLOR_EXT) |
| 367 |
SETUP_SRC_REG(optype, i, |
| 368 |
machine->Inputs[ATI_FS_INPUT_PRIMARY]); |
| 369 |
else if (index == GL_SECONDARY_INTERPOLATOR_ATI) |
| 370 |
SETUP_SRC_REG(optype, i, |
| 371 |
machine->Inputs[ATI_FS_INPUT_SECONDARY]); |
| 372 |
|
| 373 |
apply_src_rep(optype, inst->SrcReg[optype][i].argRep, |
| 374 |
src[optype][i]); |
| 375 |
apply_src_mod(optype, inst->SrcReg[optype][i].argMod, |
| 376 |
src[optype][i]); |
| 377 |
} |
| 378 |
} |
| 379 |
|
| 380 |
/* Execute the operations - color then alpha */ |
| 381 |
for (optype = 0; optype < 2; optype++) { |
| 382 |
if (inst->Opcode[optype]) { |
| 383 |
switch (inst->Opcode[optype]) { |
| 384 |
case GL_ADD_ATI: |
| 385 |
if (!optype) |
| 386 |
for (i = 0; i < 3; i++) { |
| 387 |
dst[optype][i] = |
| 388 |
src[optype][0][i] + src[optype][1][i]; |
| 389 |
} |
| 390 |
else |
| 391 |
dst[optype][3] = src[optype][0][3] + src[optype][1][3]; |
| 392 |
break; |
| 393 |
case GL_SUB_ATI: |
| 394 |
if (!optype) |
| 395 |
for (i = 0; i < 3; i++) { |
| 396 |
dst[optype][i] = |
| 397 |
src[optype][0][i] - src[optype][1][i]; |
| 398 |
} |
| 399 |
else |
| 400 |
dst[optype][3] = src[optype][0][3] - src[optype][1][3]; |
| 401 |
break; |
| 402 |
case GL_MUL_ATI: |
| 403 |
if (!optype) |
| 404 |
for (i = 0; i < 3; i++) { |
| 405 |
dst[optype][i] = |
| 406 |
src[optype][0][i] * src[optype][1][i]; |
| 407 |
} |
| 408 |
else |
| 409 |
dst[optype][3] = src[optype][0][3] * src[optype][1][3]; |
| 410 |
break; |
| 411 |
case GL_MAD_ATI: |
| 412 |
if (!optype) |
| 413 |
for (i = 0; i < 3; i++) { |
| 414 |
dst[optype][i] = |
| 415 |
src[optype][0][i] * src[optype][1][i] + |
| 416 |
src[optype][2][i]; |
| 417 |
} |
| 418 |
else |
| 419 |
dst[optype][3] = |
| 420 |
src[optype][0][3] * src[optype][1][3] + |
| 421 |
src[optype][2][3]; |
| 422 |
break; |
| 423 |
case GL_LERP_ATI: |
| 424 |
if (!optype) |
| 425 |
for (i = 0; i < 3; i++) { |
| 426 |
dst[optype][i] = |
| 427 |
src[optype][0][i] * src[optype][1][i] + (1 - |
| 428 |
src |
| 429 |
[optype] |
| 430 |
[0][i]) * |
| 431 |
src[optype][2][i]; |
| 432 |
} |
| 433 |
else |
| 434 |
dst[optype][3] = |
| 435 |
src[optype][0][3] * src[optype][1][3] + (1 - |
| 436 |
src[optype] |
| 437 |
[0][3]) * |
| 438 |
src[optype][2][3]; |
| 439 |
break; |
| 440 |
|
| 441 |
case GL_MOV_ATI: |
| 442 |
if (!optype) |
| 443 |
for (i = 0; i < 3; i++) { |
| 444 |
dst[optype][i] = src[optype][0][i]; |
| 445 |
} |
| 446 |
else |
| 447 |
dst[optype][3] = src[optype][0][3]; |
| 448 |
break; |
| 449 |
case GL_CND_ATI: |
| 450 |
if (!optype) { |
| 451 |
for (i = 0; i < 3; i++) { |
| 452 |
dst[optype][i] = |
| 453 |
(src[optype][2][i] > |
| 454 |
0.5) ? src[optype][0][i] : src[optype][1][i]; |
| 455 |
} |
| 456 |
} |
| 457 |
else { |
| 458 |
dst[optype][3] = |
| 459 |
(src[optype][2][3] > |
| 460 |
0.5) ? src[optype][0][3] : src[optype][1][3]; |
| 461 |
} |
| 462 |
break; |
| 463 |
|
| 464 |
case GL_CND0_ATI: |
| 465 |
if (!optype) |
| 466 |
for (i = 0; i < 3; i++) { |
| 467 |
dst[optype][i] = |
| 468 |
(src[optype][2][i] >= |
| 469 |
0) ? src[optype][0][i] : src[optype][1][i]; |
| 470 |
} |
| 471 |
else { |
| 472 |
dst[optype][3] = |
| 473 |
(src[optype][2][3] >= |
| 474 |
0) ? src[optype][0][3] : src[optype][1][3]; |
| 475 |
} |
| 476 |
break; |
| 477 |
case GL_DOT2_ADD_ATI: |
| 478 |
{ |
| 479 |
GLfloat result; |
| 480 |
|
| 481 |
/* DOT 2 always uses the source from the color op */ |
| 482 |
result = src[0][0][0] * src[0][1][0] + |
| 483 |
src[0][0][1] * src[0][1][1] + src[0][2][2]; |
| 484 |
if (!optype) { |
| 485 |
for (i = 0; i < 3; i++) { |
| 486 |
dst[optype][i] = result; |
| 487 |
} |
| 488 |
} |
| 489 |
else |
| 490 |
dst[optype][3] = result; |
| 491 |
|
| 492 |
} |
| 493 |
break; |
| 494 |
case GL_DOT3_ATI: |
| 495 |
{ |
| 496 |
GLfloat result; |
| 497 |
|
| 498 |
/* DOT 3 always uses the source from the color op */ |
| 499 |
result = src[0][0][0] * src[0][1][0] + |
| 500 |
src[0][0][1] * src[0][1][1] + |
| 501 |
src[0][0][2] * src[0][1][2]; |
| 502 |
|
| 503 |
if (!optype) { |
| 504 |
for (i = 0; i < 3; i++) { |
| 505 |
dst[optype][i] = result; |
| 506 |
} |
| 507 |
} |
| 508 |
else |
| 509 |
dst[optype][3] = result; |
| 510 |
} |
| 511 |
break; |
| 512 |
case GL_DOT4_ATI: |
| 513 |
{ |
| 514 |
GLfloat result; |
| 515 |
|
| 516 |
/* DOT 4 always uses the source from the color op */ |
| 517 |
result = src[optype][0][0] * src[0][1][0] + |
| 518 |
src[0][0][1] * src[0][1][1] + |
| 519 |
src[0][0][2] * src[0][1][2] + |
| 520 |
src[0][0][3] * src[0][1][3]; |
| 521 |
if (!optype) { |
| 522 |
for (i = 0; i < 3; i++) { |
| 523 |
dst[optype][i] = result; |
| 524 |
} |
| 525 |
} |
| 526 |
else |
| 527 |
dst[optype][3] = result; |
| 528 |
} |
| 529 |
break; |
| 530 |
|
| 531 |
} |
| 532 |
} |
| 533 |
} |
| 534 |
|
| 535 |
/* write out the destination registers */ |
| 536 |
for (optype = 0; optype < 2; optype++) { |
| 537 |
if (inst->Opcode[optype]) { |
| 538 |
dstreg = inst->DstReg[optype].Index; |
| 539 |
dstp = machine->Registers[dstreg - GL_REG_0_ATI]; |
| 540 |
|
| 541 |
write_dst_addr(optype, inst->DstReg[optype].dstMod, |
| 542 |
inst->DstReg[optype].dstMask, dst[optype], |
| 543 |
dstp); |
| 544 |
} |
| 545 |
} |
| 546 |
} |
| 547 |
} |
| 548 |
return GL_TRUE; |
| 549 |
} |
| 550 |
|
| 551 |
static void |
| 552 |
init_machine(GLcontext * ctx, struct atifs_machine *machine, |
| 553 |
const struct ati_fragment_shader *shader, |
| 554 |
const struct sw_span *span, GLuint col) |
| 555 |
{ |
| 556 |
GLint i, j; |
| 557 |
|
| 558 |
for (i = 0; i < 6; i++) { |
| 559 |
for (j = 0; j < 4; j++) |
| 560 |
ctx->ATIFragmentShader.Machine.Registers[i][j] = 0.0; |
| 561 |
|
| 562 |
} |
| 563 |
|
| 564 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_PRIMARY][0] = |
| 565 |
CHAN_TO_FLOAT(span->array->rgba[col][0]); |
| 566 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_PRIMARY][1] = |
| 567 |
CHAN_TO_FLOAT(span->array->rgba[col][1]); |
| 568 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_PRIMARY][2] = |
| 569 |
CHAN_TO_FLOAT(span->array->rgba[col][2]); |
| 570 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_PRIMARY][3] = |
| 571 |
CHAN_TO_FLOAT(span->array->rgba[col][3]); |
| 572 |
|
| 573 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_SECONDARY][0] = |
| 574 |
CHAN_TO_FLOAT(span->array->spec[col][0]); |
| 575 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_SECONDARY][1] = |
| 576 |
CHAN_TO_FLOAT(span->array->spec[col][1]); |
| 577 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_SECONDARY][2] = |
| 578 |
CHAN_TO_FLOAT(span->array->spec[col][2]); |
| 579 |
ctx->ATIFragmentShader.Machine.Inputs[ATI_FS_INPUT_SECONDARY][3] = |
| 580 |
CHAN_TO_FLOAT(span->array->spec[col][3]); |
| 581 |
|
| 582 |
ctx->ATIFragmentShader.Machine.pass = 0; |
| 583 |
} |
| 584 |
|
| 585 |
|
| 586 |
|
| 587 |
/** |
| 588 |
* Execute the current fragment program, operating on the given span. |
| 589 |
*/ |
| 590 |
void |
| 591 |
_swrast_exec_fragment_shader(GLcontext * ctx, struct sw_span *span) |
| 592 |
{ |
| 593 |
const struct ati_fragment_shader *shader = ctx->ATIFragmentShader.Current; |
| 594 |
GLuint i; |
| 595 |
|
| 596 |
ctx->_CurrentProgram = GL_FRAGMENT_SHADER_ATI; |
| 597 |
|
| 598 |
for (i = 0; i < span->end; i++) { |
| 599 |
if (span->array->mask[i]) { |
| 600 |
init_machine(ctx, &ctx->ATIFragmentShader.Machine, |
| 601 |
ctx->ATIFragmentShader.Current, span, i); |
| 602 |
|
| 603 |
if (execute_shader(ctx, shader, ~0, |
| 604 |
&ctx->ATIFragmentShader.Machine, span, i)) { |
| 605 |
span->array->mask[i] = GL_FALSE; |
| 606 |
} |
| 607 |
|
| 608 |
{ |
| 609 |
const GLfloat *colOut = |
| 610 |
ctx->ATIFragmentShader.Machine.Registers[0]; |
| 611 |
|
| 612 |
/*fprintf(stderr,"outputs %f %f %f %f\n", colOut[0], colOut[1], colOut[2], colOut[3]); */ |
| 613 |
UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][RCOMP], colOut[0]); |
| 614 |
UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][GCOMP], colOut[1]); |
| 615 |
UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][BCOMP], colOut[2]); |
| 616 |
UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][ACOMP], colOut[3]); |
| 617 |
} |
| 618 |
} |
| 619 |
|
| 620 |
} |
| 621 |
|
| 622 |
|
| 623 |
ctx->_CurrentProgram = 0; |
| 624 |
|
| 625 |
} |