+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ fetch_vector4( ctx, &inst->SrcReg[1], machine, program, b );
+ result[0] = MAX2(a[0], b[0]);
+ result[1] = MAX2(a[1], b[1]);
+ result[2] = MAX2(a[2], b[2]);
+ result[3] = MAX2(a[3], b[3]);
+ store_vector4( inst, machine, result );
+ if (DEBUG_FRAG) {
+ printf("MAX (%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_MIN:
+ {
+ GLfloat a[4], b[4], result[4];
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ fetch_vector4( ctx, &inst->SrcReg[1], machine, program, b );
+ result[0] = MIN2(a[0], b[0]);
+ result[1] = MIN2(a[1], b[1]);
+ result[2] = MIN2(a[2], b[2]);
+ result[3] = MIN2(a[3], b[3]);
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_MOV:
+ {
+ GLfloat result[4];
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, result );
+ store_vector4( inst, machine, result );
+ if (DEBUG_FRAG) {
+ printf("MOV (%g %g %g %g)\n",
+ result[0], result[1], result[2], result[3]);
+ }
+ }
+ break;
+ case OPCODE_MUL:
+ {
+ GLfloat a[4], b[4], result[4];
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ fetch_vector4( ctx, &inst->SrcReg[1], machine, program, 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_FRAG) {
+ printf("MUL (%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_PK2H: /* pack two 16-bit floats in one 32-bit float */
+ {
+ GLfloat a[4], result[4];
+ GLhalfNV hx, hy;
+ GLuint *rawResult = (GLuint *) result;
+ GLuint twoHalves;
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ hx = _mesa_float_to_half(a[0]);
+ hy = _mesa_float_to_half(a[1]);
+ twoHalves = hx | (hy << 16);
+ rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
+ = twoHalves;
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_PK2US: /* pack two GLushorts into one 32-bit float */
+ {
+ GLfloat a[4], result[4];
+ GLuint usx, usy, *rawResult = (GLuint *) result;
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ a[0] = CLAMP(a[0], 0.0F, 1.0F);
+ a[1] = CLAMP(a[1], 0.0F, 1.0F);
+ usx = IROUND(a[0] * 65535.0F);
+ usy = IROUND(a[1] * 65535.0F);
+ rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
+ = usx | (usy << 16);
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_PK4B: /* pack four GLbytes into one 32-bit float */
+ {
+ GLfloat a[4], result[4];
+ GLuint ubx, uby, ubz, ubw, *rawResult = (GLuint *) result;
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ a[0] = CLAMP(a[0], -128.0F / 127.0F, 1.0F);
+ a[1] = CLAMP(a[1], -128.0F / 127.0F, 1.0F);
+ a[2] = CLAMP(a[2], -128.0F / 127.0F, 1.0F);
+ a[3] = CLAMP(a[3], -128.0F / 127.0F, 1.0F);
+ ubx = IROUND(127.0F * a[0] + 128.0F);
+ uby = IROUND(127.0F * a[1] + 128.0F);
+ ubz = IROUND(127.0F * a[2] + 128.0F);
+ ubw = IROUND(127.0F * a[3] + 128.0F);
+ rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
+ = ubx | (uby << 8) | (ubz << 16) | (ubw << 24);
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_PK4UB: /* pack four GLubytes into one 32-bit float */
+ {
+ GLfloat a[4], result[4];
+ GLuint ubx, uby, ubz, ubw, *rawResult = (GLuint *) result;
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ a[0] = CLAMP(a[0], 0.0F, 1.0F);
+ a[1] = CLAMP(a[1], 0.0F, 1.0F);
+ a[2] = CLAMP(a[2], 0.0F, 1.0F);
+ a[3] = CLAMP(a[3], 0.0F, 1.0F);
+ ubx = IROUND(255.0F * a[0]);
+ uby = IROUND(255.0F * a[1]);
+ ubz = IROUND(255.0F * a[2]);
+ ubw = IROUND(255.0F * a[3]);
+ rawResult[0] = rawResult[1] = rawResult[2] = rawResult[3]
+ = ubx | (uby << 8) | (ubz << 16) | (ubw << 24);
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_POW:
+ {
+ GLfloat a[4], b[4], result[4];
+ fetch_vector1( ctx, &inst->SrcReg[0], machine, program, a );
+ fetch_vector1( ctx, &inst->SrcReg[1], machine, program, b );
+ result[0] = result[1] = result[2] = result[3]
+ = (GLfloat)_mesa_pow(a[0], b[0]);
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_RCP:
+ {
+ GLfloat a[4], result[4];
+ fetch_vector1( ctx, &inst->SrcReg[0], machine, program, a );
+ if (DEBUG_FRAG) {
+ if (a[0] == 0)
+ printf("RCP(0)\n");
+ else if (IS_INF_OR_NAN(a[0]))
+ printf("RCP(inf)\n");
+ }
+ result[0] = result[1] = result[2] = result[3] = 1.0F / a[0];
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_RET: /* return from subroutine */
+ {
+ /* NOTE: The return is conditional! */
+ const GLuint swizzle = inst->DstReg.CondSwizzle;
+ const GLuint condMask = inst->DstReg.CondMask;
+ if (test_cc(machine->CondCodes[GET_SWZ(swizzle, 0)], condMask) ||
+ test_cc(machine->CondCodes[GET_SWZ(swizzle, 1)], condMask) ||
+ test_cc(machine->CondCodes[GET_SWZ(swizzle, 2)], condMask) ||
+ test_cc(machine->CondCodes[GET_SWZ(swizzle, 3)], condMask)) {
+ if (machine->StackDepth == 0) {
+ return GL_TRUE; /* Per GL_NV_vertex_program2 spec */
+ }
+ pc = machine->CallStack[--machine->StackDepth];
+ }
+ }
+ break;
+ case OPCODE_RFL: /* reflection vector */
+ {
+ GLfloat axis[4], dir[4], result[4], tmpX, tmpW;
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, axis );
+ fetch_vector4( ctx, &inst->SrcReg[1], machine, program, dir );
+ tmpW = DOT3(axis, axis);
+ tmpX = (2.0F * DOT3(axis, dir)) / tmpW;
+ result[0] = tmpX * axis[0] - dir[0];
+ result[1] = tmpX * axis[1] - dir[1];
+ result[2] = tmpX * axis[2] - dir[2];
+ /* result[3] is never written! XXX enforce in parser! */
+ store_vector4( inst, machine, result );
+ }
+ break;
+ case OPCODE_RSQ: /* 1 / sqrt() */
+ {
+ GLfloat a[4], result[4];
+ fetch_vector1( ctx, &inst->SrcReg[0], machine, program, a );
+ a[0] = FABSF(a[0]);
+ result[0] = result[1] = result[2] = result[3] = INV_SQRTF(a[0]);
+ store_vector4( inst, machine, result );
+ if (DEBUG_FRAG) {
+ printf("RSQ %g = 1/sqrt(|%g|)\n", result[0], a[0]);
+ }
+ }
+ break;
+ case OPCODE_SCS: /* sine and cos */
+ {
+ GLfloat a[4], result[4];
+ fetch_vector1( ctx, &inst->SrcReg[0], machine, program, 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( ctx, &inst->SrcReg[0], machine, program, a );
+ fetch_vector4( ctx, &inst->SrcReg[1], machine, program, b );