prog_execute.c

Mesa is an open-source implementation of the OpenGL specification - a system for rendering interacti

            GLfloat a[4], result[4];
            fetch_vector1(&inst->SrcReg[0], machine, a);
            result[0] = (GLfloat) _mesa_cos(a[0]);
            result[1] = (GLfloat) _mesa_sin(a[0]);
            result[2] = 0.0;    /* undefined! */
            result[3] = 0.0;    /* undefined! */
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_SEQ:         /* set on equal */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = (a[0] == b[0]) ? 1.0F : 0.0F;
            result[1] = (a[1] == b[1]) ? 1.0F : 0.0F;
            result[2] = (a[2] == b[2]) ? 1.0F : 0.0F;
            result[3] = (a[3] == b[3]) ? 1.0F : 0.0F;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SEQ (%g %g %g %g) = (%g %g %g %g) == (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3],
                      b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_SFL:         /* set false, operands ignored */
         {
            static const GLfloat result[4] = { 0.0F, 0.0F, 0.0F, 0.0F };
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_SGE:         /* set on greater or equal */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = (a[0] >= b[0]) ? 1.0F : 0.0F;
            result[1] = (a[1] >= b[1]) ? 1.0F : 0.0F;
            result[2] = (a[2] >= b[2]) ? 1.0F : 0.0F;
            result[3] = (a[3] >= b[3]) ? 1.0F : 0.0F;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SGE (%g %g %g %g) = (%g %g %g %g) >= (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3],
                      b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_SGT:         /* set on greater */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = (a[0] > b[0]) ? 1.0F : 0.0F;
            result[1] = (a[1] > b[1]) ? 1.0F : 0.0F;
            result[2] = (a[2] > b[2]) ? 1.0F : 0.0F;
            result[3] = (a[3] > b[3]) ? 1.0F : 0.0F;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SGT (%g %g %g %g) = (%g %g %g %g) > (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3],
                      b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_SIN:
         {
            GLfloat a[4], result[4];
            fetch_vector1(&inst->SrcReg[0], machine, a);
            result[0] = result[1] = result[2] = result[3]
               = (GLfloat) _mesa_sin(a[0]);
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_SLE:         /* set on less or equal */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = (a[0] <= b[0]) ? 1.0F : 0.0F;
            result[1] = (a[1] <= b[1]) ? 1.0F : 0.0F;
            result[2] = (a[2] <= b[2]) ? 1.0F : 0.0F;
            result[3] = (a[3] <= b[3]) ? 1.0F : 0.0F;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SLE (%g %g %g %g) = (%g %g %g %g) <= (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3],
                      b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_SLT:         /* set on less */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = (a[0] < b[0]) ? 1.0F : 0.0F;
            result[1] = (a[1] < b[1]) ? 1.0F : 0.0F;
            result[2] = (a[2] < b[2]) ? 1.0F : 0.0F;
            result[3] = (a[3] < b[3]) ? 1.0F : 0.0F;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SLT (%g %g %g %g) = (%g %g %g %g) < (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3],
                      b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_SNE:         /* set on not equal */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = (a[0] != b[0]) ? 1.0F : 0.0F;
            result[1] = (a[1] != b[1]) ? 1.0F : 0.0F;
            result[2] = (a[2] != b[2]) ? 1.0F : 0.0F;
            result[3] = (a[3] != b[3]) ? 1.0F : 0.0F;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SNE (%g %g %g %g) = (%g %g %g %g) != (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3],
                      b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_STR:         /* set true, operands ignored */
         {
            static const GLfloat result[4] = { 1.0F, 1.0F, 1.0F, 1.0F };
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_SUB:
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = a[0] - b[0];
            result[1] = a[1] - b[1];
            result[2] = a[2] - b[2];
            result[3] = a[3] - b[3];
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("SUB (%g %g %g %g) = (%g %g %g %g) - (%g %g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], a[3], b[0], b[1], b[2], b[3]);
            }
         }
         break;
      case OPCODE_SWZ:         /* extended swizzle */
         {
            const struct prog_src_register *source = &inst->SrcReg[0];
            const GLfloat *src = get_register_pointer(source, machine);
            GLfloat result[4];
            GLuint i;
            for (i = 0; i < 4; i++) {
               const GLuint swz = GET_SWZ(source->Swizzle, i);
               if (swz == SWIZZLE_ZERO)
                  result[i] = 0.0;
               else if (swz == SWIZZLE_ONE)
                  result[i] = 1.0;
               else {
                  ASSERT(swz >= 0);
                  ASSERT(swz <= 3);
                  result[i] = src[swz];
               }
               if (source->NegateBase & (1 << i))
                  result[i] = -result[i];
            }
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_TEX:         /* Both ARB and NV frag prog */
         /* Simple texel lookup */
         {
            GLfloat texcoord[4], color[4];
            fetch_vector4(&inst->SrcReg[0], machine, texcoord);

            fetch_texel(ctx, machine, inst, texcoord, 0.0, color);

            if (DEBUG_PROG) {
               printf("TEX (%g, %g, %g, %g) = texture[%d][%g, %g, %g, %g]\n",
                      color[0], color[1], color[2], color[3],
                      inst->TexSrcUnit,
                      texcoord[0], texcoord[1], texcoord[2], texcoord[3]);
            }
            store_vector4(inst, machine, color);
         }
         break;
      case OPCODE_TXB:         /* GL_ARB_fragment_program only */
         /* Texel lookup with LOD bias */
         {
            const struct gl_texture_unit *texUnit
               = &ctx->Texture.Unit[inst->TexSrcUnit];
            GLfloat texcoord[4], color[4], lodBias;

            fetch_vector4(&inst->SrcReg[0], machine, texcoord);

            /* texcoord[3] is the bias to add to lambda */
            lodBias = texUnit->LodBias + texcoord[3];
            if (texUnit->_Current) {
               lodBias += texUnit->_Current->LodBias;
            }

            fetch_texel(ctx, machine, inst, texcoord, lodBias, color);

            store_vector4(inst, machine, color);
         }
         break;
      case OPCODE_TXD:         /* GL_NV_fragment_program only */
         /* Texture lookup w/ partial derivatives for LOD */
         {
            GLfloat texcoord[4], dtdx[4], dtdy[4], color[4];
            fetch_vector4(&inst->SrcReg[0], machine, texcoord);
            fetch_vector4(&inst->SrcReg[1], machine, dtdx);
            fetch_vector4(&inst->SrcReg[2], machine, dtdy);
            machine->FetchTexelDeriv(ctx, texcoord, dtdx, dtdy,
                                     0.0, /* lodBias */
                                     inst->TexSrcUnit, color);
            store_vector4(inst, machine, color);
         }
         break;
      case OPCODE_TXP:         /* GL_ARB_fragment_program only */
         /* Texture lookup w/ projective divide */
         {
            GLfloat texcoord[4], color[4];

            fetch_vector4(&inst->SrcReg[0], machine, texcoord);
            /* Not so sure about this test - if texcoord[3] is
             * zero, we'd probably be fine except for an ASSERT in
             * IROUND_POS() which gets triggered by the inf values created.
             */
            if (texcoord[3] != 0.0) {
               texcoord[0] /= texcoord[3];
               texcoord[1] /= texcoord[3];
               texcoord[2] /= texcoord[3];
            }

            fetch_texel(ctx, machine, inst, texcoord, 0.0, color);

            store_vector4(inst, machine, color);
         }
         break;
      case OPCODE_TXP_NV:      /* GL_NV_fragment_program only */
         /* Texture lookup w/ projective divide, as above, but do not
          * do the divide by w if sampling from a cube map.
          */
         {
            GLfloat texcoord[4], color[4];

            fetch_vector4(&inst->SrcReg[0], machine, texcoord);
            if (inst->TexSrcTarget != TEXTURE_CUBE_INDEX &&
                texcoord[3] != 0.0) {
               texcoord[0] /= texcoord[3];
               texcoord[1] /= texcoord[3];
               texcoord[2] /= texcoord[3];
            }

            fetch_texel(ctx, machine, inst, texcoord, 0.0, color);

            store_vector4(inst, machine, color);
         }
         break;
      case OPCODE_UP2H:        /* unpack two 16-bit floats */
         {
            GLfloat a[4], result[4];
            const GLuint *rawBits = (const GLuint *) a;
            GLhalfNV hx, hy;
            fetch_vector1(&inst->SrcReg[0], machine, a);
            hx = rawBits[0] & 0xffff;
            hy = rawBits[0] >> 16;
            result[0] = result[2] = _mesa_half_to_float(hx);
            result[1] = result[3] = _mesa_half_to_float(hy);
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_UP2US:       /* unpack two GLushorts */
         {
            GLfloat a[4], result[4];
            const GLuint *rawBits = (const GLuint *) a;
            GLushort usx, usy;
            fetch_vector1(&inst->SrcReg[0], machine, a);
            usx = rawBits[0] & 0xffff;
            usy = rawBits[0] >> 16;
            result[0] = result[2] = usx * (1.0f / 65535.0f);
            result[1] = result[3] = usy * (1.0f / 65535.0f);
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_UP4B:        /* unpack four GLbytes */
         {
            GLfloat a[4], result[4];
            const GLuint *rawBits = (const GLuint *) a;
            fetch_vector1(&inst->SrcReg[0], machine, a);
            result[0] = (((rawBits[0] >> 0) & 0xff) - 128) / 127.0F;
            result[1] = (((rawBits[0] >> 8) & 0xff) - 128) / 127.0F;
            result[2] = (((rawBits[0] >> 16) & 0xff) - 128) / 127.0F;
            result[3] = (((rawBits[0] >> 24) & 0xff) - 128) / 127.0F;
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_UP4UB:       /* unpack four GLubytes */
         {
            GLfloat a[4], result[4];
            const GLuint *rawBits = (const GLuint *) a;
            fetch_vector1(&inst->SrcReg[0], machine, a);
            result[0] = ((rawBits[0] >> 0) & 0xff) / 255.0F;
            result[1] = ((rawBits[0] >> 8) & 0xff) / 255.0F;
            result[2] = ((rawBits[0] >> 16) & 0xff) / 255.0F;
            result[3] = ((rawBits[0] >> 24) & 0xff) / 255.0F;
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_XPD:         /* cross product */
         {
            GLfloat a[4], b[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            result[0] = a[1] * b[2] - a[2] * b[1];
            result[1] = a[2] * b[0] - a[0] * b[2];
            result[2] = a[0] * b[1] - a[1] * b[0];
            result[3] = 1.0;
            store_vector4(inst, machine, result);
            if (DEBUG_PROG) {
               printf("XPD (%g %g %g %g) = (%g %g %g) X (%g %g %g)\n",
                      result[0], result[1], result[2], result[3],
                      a[0], a[1], a[2], b[0], b[1], b[2]);
            }
         }
         break;
      case OPCODE_X2D:         /* 2-D matrix transform */
         {
            GLfloat a[4], b[4], c[4], result[4];
            fetch_vector4(&inst->SrcReg[0], machine, a);
            fetch_vector4(&inst->SrcReg[1], machine, b);
            fetch_vector4(&inst->SrcReg[2], machine, c);
            result[0] = a[0] + b[0] * c[0] + b[1] * c[1];
            result[1] = a[1] + b[0] * c[2] + b[1] * c[3];
            result[2] = a[2] + b[0] * c[0] + b[1] * c[1];
            result[3] = a[3] + b[0] * c[2] + b[1] * c[3];
            store_vector4(inst, machine, result);
         }
         break;
      case OPCODE_PRINT:
         {
            if (inst->SrcReg[0].File != -1) {
               GLfloat a[4];
               fetch_vector4(&inst->SrcReg[0], machine, a);
               _mesa_printf("%s%g, %g, %g, %g\n", (const char *) inst->Data,
                            a[0], a[1], a[2], a[3]);
            }
            else {
               _mesa_printf("%s\n", (const char *) inst->Data);
            }
         }
         break;
      case OPCODE_END:
         return GL_TRUE;
      default:
         _mesa_problem(ctx, "Bad opcode %d in _mesa_execute_program",
                       inst->Opcode);
         return GL_TRUE;        /* return value doesn't matter */
      }

      numExec++;
      if (numExec > maxExec) {
         _mesa_problem(ctx, "Infinite loop detected in fragment program");
         return GL_TRUE;
      }

   } /* for pc */

#if FEATURE_MESA_program_debug
   CurrentMachine = NULL;
#endif

   return GL_TRUE;
}