src/mesa/swrast/s_fragprog.c

   1 /*
   2  * Mesa 3-D graphics library
   3  * Version:  7.0.3
   4  *
   5  * Copyright (C) 1999-2007  Brian Paul   All Rights Reserved.
   6  *
   7  * Permission is hereby granted, free of charge, to any person obtaining a
   8  * copy of this software and associated documentation files (the "Software"),
   9  * to deal in the Software without restriction, including without limitation
  10  * the rights to use, copy, modify, merge, publish, distribute, sublicense,
  11  * and/or sell copies of the Software, and to permit persons to whom the
  12  * Software is furnished to do so, subject to the following conditions:
  13  *
  14  * The above copyright notice and this permission notice shall be included
  15  * in all copies or substantial portions of the Software.
  16  *
  17  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  18  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  19  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
  20  * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
  21  * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
  22  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  23  */
  24
  25 #include "main/glheader.h"
  26 #include "main/colormac.h"
  27 #include "main/context.h"
  28 #include "main/texstate.h"
  29 #include "shader/prog_instruction.h"
  30
  31 #include "s_fragprog.h"
  32 #include "s_span.h"
  33
  34
  35 /**
  36  * Apply texture object's swizzle (X/Y/Z/W/0/1) to incoming 'texel'
  37  * and return results in 'colorOut'.
  38  */
  39 static INLINE void
  40 swizzle_texel(const GLfloat texel[4], GLfloat colorOut[4], GLuint swizzle)
  41 {
  42    if (swizzle == SWIZZLE_NOOP) {
  43       COPY_4V(colorOut, texel);
  44    }
  45    else {
  46       GLfloat vector[6];
  47       vector[SWIZZLE_X] = texel[0];
  48       vector[SWIZZLE_Y] = texel[1];
  49       vector[SWIZZLE_Z] = texel[2];
  50       vector[SWIZZLE_W] = texel[3];
  51       vector[SWIZZLE_ZERO] = 0.0F;
  52       vector[SWIZZLE_ONE] = 1.0F;
  53       colorOut[0] = vector[GET_SWZ(swizzle, 0)];
  54       colorOut[1] = vector[GET_SWZ(swizzle, 1)];
  55       colorOut[2] = vector[GET_SWZ(swizzle, 2)];
  56       colorOut[3] = vector[GET_SWZ(swizzle, 3)];
  57    }
  58 }
  59
  60
  61 /**
  62  * Fetch a texel with given lod.
  63  * Called via machine->FetchTexelLod()
  64  */
  65 static void
  66 fetch_texel_lod( GLcontext *ctx, const GLfloat texcoord[4], GLfloat lambda,
  67                  GLuint unit, GLfloat color[4] )
  68 {
  69    const struct gl_texture_object *texObj = ctx->Texture.Unit[unit]._Current;
  70
  71    if (texObj) {
  72       SWcontext *swrast = SWRAST_CONTEXT(ctx);
  73       GLfloat rgba[4];
  74
  75       lambda = CLAMP(lambda, texObj->MinLod, texObj->MaxLod);
  76
  77       swrast->TextureSample[unit](ctx, texObj, 1,
  78                                   (const GLfloat (*)[4]) texcoord,
  79                                   &lambda, &rgba);
  80       swizzle_texel(rgba, color, texObj->_Swizzle);
  81    }
  82    else {
  83       ASSIGN_4V(color, 0.0F, 0.0F, 0.0F, 1.0F);
  84    }
  85 }
  86
  87
  88 /**
  89  * Fetch a texel with the given partial derivatives to compute a level
  90  * of detail in the mipmap.
  91  * Called via machine->FetchTexelDeriv()
  92  * \param lodBias  the lod bias which may be specified by a TXB instruction,
  93  *                 otherwise zero.
  94  */
  95 static void
  96 fetch_texel_deriv( GLcontext *ctx, const GLfloat texcoord[4],
  97                    const GLfloat texdx[4], const GLfloat texdy[4],
  98                    GLfloat lodBias, GLuint unit, GLfloat color[4] )
  99 {
 100    SWcontext *swrast = SWRAST_CONTEXT(ctx);
 101    const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
 102    const struct gl_texture_object *texObj = texUnit->_Current;
 103
 104    if (texObj) {
 105       const struct gl_texture_image *texImg =
 106          texObj->Image[0][texObj->BaseLevel];
 107       const GLfloat texW = (GLfloat) texImg->WidthScale;
 108       const GLfloat texH = (GLfloat) texImg->HeightScale;
 109       GLfloat lambda;
 110       GLfloat rgba[4];
 111
 112       lambda = _swrast_compute_lambda(texdx[0], texdy[0], /* ds/dx, ds/dy */
 113                                       texdx[1], texdy[1], /* dt/dx, dt/dy */
 114                                       texdx[3], texdy[3], /* dq/dx, dq/dy */
 115                                       texW, texH,
 116                                       texcoord[0], texcoord[1], texcoord[3],
 117                                       1.0F / texcoord[3]);
 118
 119       lambda += lodBias + texUnit->LodBias + texObj->LodBias;
 120
 121       lambda = CLAMP(lambda, texObj->MinLod, texObj->MaxLod);
 122
 123       swrast->TextureSample[unit](ctx, texObj, 1,
 124                                   (const GLfloat (*)[4]) texcoord,
 125                                   &lambda, &rgba);
 126       swizzle_texel(rgba, color, texObj->_Swizzle);
 127    }
 128    else {
 129       ASSIGN_4V(color, 0.0F, 0.0F, 0.0F, 1.0F);
 130    }
 131 }
 132
 133
 134 /**
 135  * Initialize the virtual fragment program machine state prior to running
 136  * fragment program on a fragment.  This involves initializing the input
 137  * registers, condition codes, etc.
 138  * \param machine  the virtual machine state to init
 139  * \param program  the fragment program we're about to run
 140  * \param span  the span of pixels we'll operate on
 141  * \param col  which element (column) of the span we'll operate on
 142  */
 143 static void
 144 init_machine(GLcontext *ctx, struct gl_program_machine *machine,
 145              const struct gl_fragment_program *program,
 146              const SWspan *span, GLuint col)
 147 {
 148    if (program->Base.Target == GL_FRAGMENT_PROGRAM_NV) {
 149       /* Clear temporary registers (undefined for ARB_f_p) */
 150       _mesa_bzero(machine->Temporaries,
 151                   MAX_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
 152    }
 153
 154    /* Setup pointer to input attributes */
 155    machine->Attribs = span->array->attribs;
 156
 157    machine->DerivX = (GLfloat (*)[4]) span->attrStepX;
 158    machine->DerivY = (GLfloat (*)[4]) span->attrStepY;
 159    machine->NumDeriv = FRAG_ATTRIB_MAX;
 160
 161    machine->Samplers = program->Base.SamplerUnits;
 162
 163    /* if running a GLSL program (not ARB_fragment_program) */
 164    if (ctx->Shader.CurrentProgram) {
 165       /* Store front/back facing value */
 166       machine->Attribs[FRAG_ATTRIB_FACE][col][0] = 1.0 - span->facing;
 167    }
 168
 169    machine->CurElement = col;
 170
 171    /* init condition codes */
 172    machine->CondCodes[0] = COND_EQ;
 173    machine->CondCodes[1] = COND_EQ;
 174    machine->CondCodes[2] = COND_EQ;
 175    machine->CondCodes[3] = COND_EQ;
 176
 177    /* init call stack */
 178    machine->StackDepth = 0;
 179
 180    machine->FetchTexelLod = fetch_texel_lod;
 181    machine->FetchTexelDeriv = fetch_texel_deriv;
 182 }
 183
 184
 185 /**
 186  * Run fragment program on the pixels in span from 'start' to 'end' - 1.
 187  */
 188 static void
 189 run_program(GLcontext *ctx, SWspan *span, GLuint start, GLuint end)
 190 {
 191    SWcontext *swrast = SWRAST_CONTEXT(ctx);
 192    const struct gl_fragment_program *program = ctx->FragmentProgram._Current;
 193    const GLbitfield64 outputsWritten = program->Base.OutputsWritten;
 194    struct gl_program_machine *machine = &swrast->FragProgMachine;
 195    GLuint i;
 196
 197    for (i = start; i < end; i++) {
 198       if (span->array->mask[i]) {
 199          init_machine(ctx, machine, program, span, i);
 200
 201          if (_mesa_execute_program(ctx, &program->Base, machine)) {
 202
 203             /* Store result color */
 204             if (outputsWritten & BITFIELD64_BIT(FRAG_RESULT_COLOR)) {
 205                COPY_4V(span->array->attribs[FRAG_ATTRIB_COL0][i],
 206                        machine->Outputs[FRAG_RESULT_COLOR]);
 207             }
 208             else {
 209                /* Multiple drawbuffers / render targets
 210                 * Note that colors beyond 0 and 1 will overwrite other
 211                 * attributes, such as FOGC, TEX0, TEX1, etc.  That's OK.
 212                 */
 213                GLuint buf;
 214                for (buf = 0; buf < ctx->DrawBuffer->_NumColorDrawBuffers; buf++) {
 215                   if (outputsWritten & BITFIELD64_BIT(FRAG_RESULT_DATA0 + buf)) {
 216                      COPY_4V(span->array->attribs[FRAG_ATTRIB_COL0 + buf][i],
 217                              machine->Outputs[FRAG_RESULT_DATA0 + buf]);
 218                   }
 219                }
 220             }
 221
 222             /* Store result depth/z */
 223             if (outputsWritten & BITFIELD64_BIT(FRAG_RESULT_DEPTH)) {
 224                const GLfloat depth = machine->Outputs[FRAG_RESULT_DEPTH][2];
 225                if (depth <= 0.0)
 226                   span->array->z[i] = 0;
 227                else if (depth >= 1.0)
 228                   span->array->z[i] = ctx->DrawBuffer->_DepthMax;
 229                else
 230                   span->array->z[i] = IROUND(depth * ctx->DrawBuffer->_DepthMaxF);
 231             }
 232          }
 233          else {
 234             /* killed fragment */
 235             span->array->mask[i] = GL_FALSE;
 236             span->writeAll = GL_FALSE;
 237          }
 238       }
 239    }
 240 }
 241
 242
 243 /**
 244  * Execute the current fragment program for all the fragments
 245  * in the given span.
 246  */
 247 void
 248 _swrast_exec_fragment_program( GLcontext *ctx, SWspan *span )
 249 {
 250    const struct gl_fragment_program *program = ctx->FragmentProgram._Current;
 251
 252    /* incoming colors should be floats */
 253    if (program->Base.InputsRead & FRAG_BIT_COL0) {
 254       ASSERT(span->array->ChanType == GL_FLOAT);
 255    }
 256
 257    run_program(ctx, span, 0, span->end);
 258
 259    if (program->Base.OutputsWritten & BITFIELD64_BIT(FRAG_RESULT_COLOR)) {
 260       span->interpMask &= ~SPAN_RGBA;
 261       span->arrayMask |= SPAN_RGBA;
 262    }
 263
 264    if (program->Base.OutputsWritten & BITFIELD64_BIT(FRAG_RESULT_DEPTH)) {
 265       span->interpMask &= ~SPAN_Z;
 266       span->arrayMask |= SPAN_Z;
 267    }
 268 }
 269