reindent

1 files changed, 924 insertions, 928 deletions

diff --git a/src/mesa/shader/prog_execute.c b/src/mesa/shader/prog_execute.c
index d75a55b95b..82c578f983 100644
--- a/src/mesa/shader/prog_execute.c
+++ b/src/mesa/shader/prog_execute.c
@@ -73,7 +73,7 @@ _mesa_get_program_register(GLcontext *ctx, enum register_file file,
          }
          else {
             COPY_4V(val,
-                    CurrentMachine->Attribs[index][CurrentMachine->CurElement]);
+                   CurrentMachine->Attribs[index][CurrentMachine->CurElement]);
          }
          break;
       case PROGRAM_OUTPUT:
@@ -97,9 +97,9 @@ _mesa_get_program_register(GLcontext *ctx, enum register_file file,
  * source register.
  */
 static INLINE const GLfloat *
-get_register_pointer( GLcontext *ctx,
-                      const struct prog_src_register *source,
-                      const struct gl_program_machine *machine)
+get_register_pointer(GLcontext * ctx,
+                     const struct prog_src_register *source,
+                     const struct gl_program_machine *machine)
 {
    /* XXX relative addressing... */
    switch (source->File) {
@@ -158,10 +158,9 @@ get_register_pointer( GLcontext *ctx,
  * Apply swizzling and negating as needed.
  */
 static void
-fetch_vector4( GLcontext *ctx,
-               const struct prog_src_register *source,
-               const struct gl_program_machine *machine,
-               GLfloat result[4] )
+fetch_vector4(GLcontext * ctx,
+              const struct prog_src_register *source,
+              const struct gl_program_machine *machine, GLfloat result[4])
 {
    const GLfloat *src = get_register_pointer(ctx, source, machine);
    ASSERT(src);
@@ -208,10 +207,10 @@ fetch_vector4( GLcontext *ctx,
  * need to execute another instance of the program (ugh)!
  */
 static GLboolean
-fetch_vector4_deriv( GLcontext *ctx,
-                     const struct prog_src_register *source,
-                     const SWspan *span,
-                     char xOrY, GLint column, GLfloat result[4] )
+fetch_vector4_deriv(GLcontext * ctx,
+                    const struct prog_src_register *source,
+                    const SWspan * span,
+                    char xOrY, GLint column, GLfloat result[4])
 {
    GLfloat src[4];
 
@@ -223,14 +222,14 @@ fetch_vector4_deriv( GLcontext *ctx,
          src[0] = 1.0;
          src[1] = 0.0;
          src[2] = span->attrStepX[FRAG_ATTRIB_WPOS][2]
-                / ctx->DrawBuffer->_DepthMaxF;
+            / ctx->DrawBuffer->_DepthMaxF;
          src[3] = span->attrStepX[FRAG_ATTRIB_WPOS][3];
       }
       else {
          src[0] = 0.0;
          src[1] = 1.0;
          src[2] = span->attrStepY[FRAG_ATTRIB_WPOS][2]
-                / ctx->DrawBuffer->_DepthMaxF;
+            / ctx->DrawBuffer->_DepthMaxF;
          src[3] = span->attrStepY[FRAG_ATTRIB_WPOS][3];
       }
       break;
@@ -269,7 +268,9 @@ fetch_vector4_deriv( GLcontext *ctx,
       if (xOrY == 'X') {
          /* this is a little tricky - I think I've got it right */
          const GLfloat invQ = 1.0f / (span->attrStart[source->Index][3]
-                               + span->attrStepX[source->Index][3] * column);
+                                      +
+                                      span->attrStepX[source->Index][3] *
+                                      column);
          src[0] = span->attrStepX[source->Index][0] * invQ;
          src[1] = span->attrStepX[source->Index][1] * invQ;
          src[2] = span->attrStepX[source->Index][2] * invQ;
@@ -319,10 +320,9 @@ fetch_vector4_deriv( GLcontext *ctx,
  * As above, but only return result[0] element.
  */
 static void
-fetch_vector1( GLcontext *ctx,
-               const struct prog_src_register *source,
-               const struct gl_program_machine *machine,
-               GLfloat result[4] )
+fetch_vector1(GLcontext * ctx,
+              const struct prog_src_register *source,
+              const struct gl_program_machine *machine, GLfloat result[4])
 {
    const GLfloat *src = get_register_pointer(ctx, source, machine);
    ASSERT(src);
@@ -345,10 +345,10 @@ fetch_vector1( GLcontext *ctx,
  * Test value against zero and return GT, LT, EQ or UN if NaN.
  */
 static INLINE GLuint
-generate_cc( float value )
+generate_cc(float value)
 {
    if (value != value)
-      return COND_UN;  /* NaN */
+      return COND_UN;           /* NaN */
    if (value > 0.0F)
       return COND_GT;
    if (value < 0.0F)
@@ -406,9 +406,8 @@ eval_condition(const struct gl_program_machine *machine,
  * set-condition-code flags.
  */
 static void
-store_vector4( const struct prog_instruction *inst,
-               struct gl_program_machine *machine,
-               const GLfloat value[4] )
+store_vector4(const struct prog_instruction *inst,
+              struct gl_program_machine *machine, const GLfloat value[4])
 {
    const struct prog_dst_register *dest = &(inst->DstReg);
    const GLboolean clamp = inst->SaturateMode == SATURATE_ZERO_ONE;
@@ -418,26 +417,25 @@ store_vector4( const struct prog_instruction *inst,
    GLuint writeMask = dest->WriteMask;
 
    switch (dest->File) {
-      case PROGRAM_OUTPUT:
-         dstReg = machine->Outputs[dest->Index];
-         break;
-      case PROGRAM_TEMPORARY:
-         dstReg = machine->Temporaries[dest->Index];
-         break;
-      case PROGRAM_WRITE_ONLY:
-         dstReg = dummyReg;
-         return;
-      default:
-         _mesa_problem(NULL, "bad register file in store_vector4(fp)");
-         return;
+   case PROGRAM_OUTPUT:
+      dstReg = machine->Outputs[dest->Index];
+      break;
+   case PROGRAM_TEMPORARY:
+      dstReg = machine->Temporaries[dest->Index];
+      break;
+   case PROGRAM_WRITE_ONLY:
+      dstReg = dummyReg;
+      return;
+   default:
+      _mesa_problem(NULL, "bad register file in store_vector4(fp)");
+      return;
    }
 
 #if 0
    if (value[0] > 1.0e10 ||
        IS_INF_OR_NAN(value[0]) ||
        IS_INF_OR_NAN(value[1]) ||
-       IS_INF_OR_NAN(value[2]) ||
-       IS_INF_OR_NAN(value[3])  )
+       IS_INF_OR_NAN(value[2]) || IS_INF_OR_NAN(value[3]))
       printf("store %g %g %g %g\n", value[0], value[1], value[2], value[3]);
 #endif
 
@@ -502,11 +500,11 @@ store_vector4( const struct prog_instruction *inst,
  * Used to implement DDX and DDY instructions in non-trivial cases.
  */
 static void
-init_machine_deriv( GLcontext *ctx,
-                    const struct gl_program_machine *machine,
-                    const struct gl_fragment_program *program,
-                    const SWspan *span, char xOrY,
-                    struct gl_program_machine *dMachine )
+init_machine_deriv(GLcontext * ctx,
+                   const struct gl_program_machine *machine,
+                   const struct gl_fragment_program *program,
+                   const SWspan * span, char xOrY,
+                   struct gl_program_machine *dMachine)
 {
    GLuint attr;
 
@@ -518,8 +516,8 @@ init_machine_deriv( GLcontext *ctx,
    if (program->Base.Target == GL_FRAGMENT_PROGRAM_NV) {
       /* XXX also need to do this when using valgrind */
       /* Clear temporary registers (undefined for ARB_f_p) */
-      _mesa_bzero( (void*) machine->Temporaries,
-                   MAX_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
+      _mesa_bzero((void *) machine->Temporaries,
+                  MAX_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
    }
 
    /* Add derivatives */
@@ -605,7 +603,7 @@ init_machine_deriv( GLcontext *ctx,
  * \return GL_TRUE if program completed or GL_FALSE if program executed KIL.
  */
 GLboolean
-_mesa_execute_program(GLcontext *ctx,
+_mesa_execute_program(GLcontext * ctx,
                       const struct gl_program *program, GLuint maxInst,
                       struct gl_program_machine *machine, GLuint element)
 {
@@ -639,941 +637,939 @@ _mesa_execute_program(GLcontext *ctx,
       }
 
       switch (inst->Opcode) {
-         case OPCODE_ABS:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = FABSF(a[0]);
-               result[1] = FABSF(a[1]);
-               result[2] = FABSF(a[2]);
-               result[3] = FABSF(a[3]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_ADD:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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("ADD (%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_BGNLOOP:
-            /* no-op */
-            break;
-         case OPCODE_ENDLOOP:
-            /* subtract 1 here since pc is incremented by for(pc) loop */
-            pc = inst->BranchTarget - 1; /* go to matching BNGLOOP */
-            break;
-         case OPCODE_BGNSUB: /* begin subroutine */
-            break;
-         case OPCODE_ENDSUB: /* end subroutine */
-            break;
-         case OPCODE_BRA: /* branch (conditional) */
-            /* fall-through */
-         case OPCODE_BRK: /* break out of loop (conditional) */
-            /* fall-through */
-         case OPCODE_CONT: /* continue loop (conditional) */
-            if (eval_condition(machine, inst)) {
-               /* take branch */
-               /* Subtract 1 here since we'll do pc++ at end of for-loop */
-               pc = inst->BranchTarget - 1;
-            }
-            break;
-         case OPCODE_CAL: /* Call subroutine (conditional) */
-            if (eval_condition(machine, inst)) {
-               /* call the subroutine */
-               if (machine->StackDepth >= MAX_PROGRAM_CALL_DEPTH) {
-                  return GL_TRUE; /* Per GL_NV_vertex_program2 spec */
-               }
-               machine->CallStack[machine->StackDepth++] = pc + 1;
-               pc = inst->BranchTarget; /* XXX - 1 ??? */
-            }
-            break;
-         case OPCODE_CMP:
-            {
-               GLfloat a[4], b[4], c[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               fetch_vector4( ctx, &inst->SrcReg[2], machine, c );
-               result[0] = a[0] < 0.0F ? b[0] : c[0];
-               result[1] = a[1] < 0.0F ? b[1] : c[1];
-               result[2] = a[2] < 0.0F ? b[2] : c[2];
-               result[3] = a[3] < 0.0F ? b[3] : c[3];
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_COS:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector1( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = result[1] = result[2] = result[3]
-                  = (GLfloat) _mesa_cos(a[0]);
-               store_vector4( inst, machine, result );
+      case OPCODE_ABS:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = FABSF(a[0]);
+            result[1] = FABSF(a[1]);
+            result[2] = FABSF(a[2]);
+            result[3] = FABSF(a[3]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_ADD:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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("ADD (%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_DDX: /* Partial derivative with respect to X */
-            {
+         }
+         break;
+      case OPCODE_BGNLOOP:
+         /* no-op */
+         break;
+      case OPCODE_ENDLOOP:
+         /* subtract 1 here since pc is incremented by for(pc) loop */
+         pc = inst->BranchTarget - 1;   /* go to matching BNGLOOP */
+         break;
+      case OPCODE_BGNSUB:      /* begin subroutine */
+         break;
+      case OPCODE_ENDSUB:      /* end subroutine */
+         break;
+      case OPCODE_BRA:         /* branch (conditional) */
+         /* fall-through */
+      case OPCODE_BRK:         /* break out of loop (conditional) */
+         /* fall-through */
+      case OPCODE_CONT:        /* continue loop (conditional) */
+         if (eval_condition(machine, inst)) {
+            /* take branch */
+            /* Subtract 1 here since we'll do pc++ at end of for-loop */
+            pc = inst->BranchTarget - 1;
+         }
+         break;
+      case OPCODE_CAL:         /* Call subroutine (conditional) */
+         if (eval_condition(machine, inst)) {
+            /* call the subroutine */
+            if (machine->StackDepth >= MAX_PROGRAM_CALL_DEPTH) {
+               return GL_TRUE;  /* Per GL_NV_vertex_program2 spec */
+            }
+            machine->CallStack[machine->StackDepth++] = pc + 1;
+            pc = inst->BranchTarget;    /* XXX - 1 ??? */
+         }
+         break;
+      case OPCODE_CMP:
+         {
+            GLfloat a[4], b[4], c[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            fetch_vector4(ctx, &inst->SrcReg[2], machine, c);
+            result[0] = a[0] < 0.0F ? b[0] : c[0];
+            result[1] = a[1] < 0.0F ? b[1] : c[1];
+            result[2] = a[2] < 0.0F ? b[2] : c[2];
+            result[3] = a[3] < 0.0F ? b[3] : c[3];
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_COS:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector1(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = result[1] = result[2] = result[3]
+               = (GLfloat) _mesa_cos(a[0]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_DDX:         /* Partial derivative with respect to X */
+         {
 #if 0
-               GLfloat a[4], aNext[4], result[4];
-               struct gl_program_machine dMachine;
-               if (!fetch_vector4_deriv(ctx, &inst->SrcReg[0], span, 'X',
-                                        column, result)) {
-                  /* This is tricky.  Make a copy of the current machine state,
-                   * increment the input registers by the dx or dy partial
-                   * derivatives, then re-execute the program up to the
-                   * preceeding instruction, then fetch the source register.
-                   * Finally, find the difference in the register values for
-                   * the original and derivative runs.
-                   */
-                  fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a);
-                  init_machine_deriv(ctx, machine, program, span,
-                                     'X', &dMachine);
-                  execute_program(ctx, program, pc, &dMachine, span, column);
-                  fetch_vector4( ctx, &inst->SrcReg[0], &dMachine, program, aNext );
-                  result[0] = aNext[0] - a[0];
-                  result[1] = aNext[1] - a[1];
-                  result[2] = aNext[2] - a[2];
-                  result[3] = aNext[3] - a[3];
-               }
-               store_vector4( inst, machine, result );
+            GLfloat a[4], aNext[4], result[4];
+            struct gl_program_machine dMachine;
+            if (!fetch_vector4_deriv(ctx, &inst->SrcReg[0], span, 'X',
+                                     column, result)) {
+               /* This is tricky.  Make a copy of the current machine state,
+                * increment the input registers by the dx or dy partial
+                * derivatives, then re-execute the program up to the
+                * preceeding instruction, then fetch the source register.
+                * Finally, find the difference in the register values for
+                * the original and derivative runs.
+                */
+               fetch_vector4(ctx, &inst->SrcReg[0], machine, program, a);
+               init_machine_deriv(ctx, machine, program, span,
+                                  'X', &dMachine);
+               execute_program(ctx, program, pc, &dMachine, span, column);
+               fetch_vector4(ctx, &inst->SrcReg[0], &dMachine, program,
+                             aNext);
+               result[0] = aNext[0] - a[0];
+               result[1] = aNext[1] - a[1];
+               result[2] = aNext[2] - a[2];
+               result[3] = aNext[3] - a[3];
+            }
+            store_vector4(inst, machine, result);
 #else
-               static const GLfloat result[4] = { 0, 0, 0, 0 };
-               store_vector4( inst, machine, result );
+            static const GLfloat result[4] = { 0, 0, 0, 0 };
+            store_vector4(inst, machine, result);
 #endif
-            }
-            break;
-         case OPCODE_DDY: /* Partial derivative with respect to Y */
-            {
+         }
+         break;
+      case OPCODE_DDY:         /* Partial derivative with respect to Y */
+         {
 #if 0
-               GLfloat a[4], aNext[4], result[4];
-               struct gl_program_machine dMachine;
-               if (!fetch_vector4_deriv(ctx, &inst->SrcReg[0], span, 'Y',
-                                        column, result)) {
-                  init_machine_deriv(ctx, machine, program, span,
-                                     'Y', &dMachine);
-                  fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a);
-                  execute_program(ctx, program, pc, &dMachine, span, column);
-                  fetch_vector4( ctx, &inst->SrcReg[0], &dMachine, program, aNext );
-                  result[0] = aNext[0] - a[0];
-                  result[1] = aNext[1] - a[1];
-                  result[2] = aNext[2] - a[2];
-                  result[3] = aNext[3] - a[3];
-               }
-               store_vector4( inst, machine, result );
+            GLfloat a[4], aNext[4], result[4];
+            struct gl_program_machine dMachine;
+            if (!fetch_vector4_deriv(ctx, &inst->SrcReg[0], span, 'Y',
+                                     column, result)) {
+               init_machine_deriv(ctx, machine, program, span,
+                                  'Y', &dMachine);
+               fetch_vector4(ctx, &inst->SrcReg[0], machine, program, a);
+               execute_program(ctx, program, pc, &dMachine, span, column);
+               fetch_vector4(ctx, &inst->SrcReg[0], &dMachine, program,
+                             aNext);
+               result[0] = aNext[0] - a[0];
+               result[1] = aNext[1] - a[1];
+               result[2] = aNext[2] - a[2];
+               result[3] = aNext[3] - a[3];
+            }
+            store_vector4(inst, machine, result);
 #else
-               static const GLfloat result[4] = { 0, 0, 0, 0 };
-               store_vector4( inst, machine, result );
+            static const GLfloat result[4] = { 0, 0, 0, 0 };
+            store_vector4(inst, machine, result);
 #endif
+         }
+         break;
+      case OPCODE_DP3:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            result[0] = result[1] = result[2] = result[3] = DOT3(a, b);
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("DP3 %g = (%g %g %g) . (%g %g %g)\n",
+                      result[0], a[0], a[1], a[2], b[0], b[1], b[2]);
             }
-            break;
-         case OPCODE_DP3:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               result[0] = result[1] = result[2] = result[3] = DOT3(a, b);
-               store_vector4( inst, machine, result );
-               if (DEBUG_PROG) {
-                  printf("DP3 %g = (%g %g %g) . (%g %g %g)\n",
-                         result[0], a[0], a[1], a[2], b[0], b[1], b[2]);
-               }
-            }
-            break;
-         case OPCODE_DP4:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               result[0] = result[1] = result[2] = result[3] = DOT4(a,b);
-               store_vector4( inst, machine, result );
-               if (DEBUG_PROG) {
-                  printf("DP4 %g = (%g, %g %g %g) . (%g, %g %g %g)\n",
-                         result[0], a[0], a[1], a[2], a[3],
-                         b[0], b[1], b[2], b[3]);
-               }
-            }
-            break;
-         case OPCODE_DPH:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               result[0] = result[1] = result[2] = result[3] = 
-                  a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + b[3];
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_DST: /* Distance vector */
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               result[0] = 1.0F;
-               result[1] = a[1] * b[1];
-               result[2] = a[2];
-               result[3] = b[3];
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_EX2: /* Exponential base 2 */
-            {
-               GLfloat a[4], result[4];
-               fetch_vector1( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = result[1] = result[2] = result[3] =
-                  (GLfloat) _mesa_pow(2.0, a[0]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_FLR:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = FLOORF(a[0]);
-               result[1] = FLOORF(a[1]);
-               result[2] = FLOORF(a[2]);
-               result[3] = FLOORF(a[3]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_FRC:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = a[0] - FLOORF(a[0]);
-               result[1] = a[1] - FLOORF(a[1]);
-               result[2] = a[2] - FLOORF(a[2]);
-               result[3] = a[3] - FLOORF(a[3]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_IF:
-            if (eval_condition(machine, inst)) {
-               /* do if-clause (just continue execution) */
-            }
-            else {
-               /* go to the instruction after ELSE or ENDIF */
-               assert(inst->BranchTarget >= 0);
-               pc = inst->BranchTarget - 1;
+         }
+         break;
+      case OPCODE_DP4:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            result[0] = result[1] = result[2] = result[3] = DOT4(a, b);
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("DP4 %g = (%g, %g %g %g) . (%g, %g %g %g)\n",
+                      result[0], a[0], a[1], a[2], a[3],
+                      b[0], b[1], b[2], b[3]);
             }
-            break;
-         case OPCODE_ELSE:
-            /* goto ENDIF */
+         }
+         break;
+      case OPCODE_DPH:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            result[0] = result[1] = result[2] = result[3] =
+               a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + b[3];
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_DST:         /* Distance vector */
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            result[0] = 1.0F;
+            result[1] = a[1] * b[1];
+            result[2] = a[2];
+            result[3] = b[3];
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_EX2:         /* Exponential base 2 */
+         {
+            GLfloat a[4], result[4];
+            fetch_vector1(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = result[1] = result[2] = result[3] =
+               (GLfloat) _mesa_pow(2.0, a[0]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_FLR:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = FLOORF(a[0]);
+            result[1] = FLOORF(a[1]);
+            result[2] = FLOORF(a[2]);
+            result[3] = FLOORF(a[3]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_FRC:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = a[0] - FLOORF(a[0]);
+            result[1] = a[1] - FLOORF(a[1]);
+            result[2] = a[2] - FLOORF(a[2]);
+            result[3] = a[3] - FLOORF(a[3]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_IF:
+         if (eval_condition(machine, inst)) {
+            /* do if-clause (just continue execution) */
+         }
+         else {
+            /* go to the instruction after ELSE or ENDIF */
             assert(inst->BranchTarget >= 0);
             pc = inst->BranchTarget - 1;
-            break;
-         case OPCODE_ENDIF:
-            /* nothing */
-            break;
-         case OPCODE_INT: /* float to int */
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = (GLfloat) (GLint) a[0];
-               result[1] = (GLfloat) (GLint) a[1];
-               result[2] = (GLfloat) (GLint) a[2];
-               result[3] = (GLfloat) (GLint) a[3];
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_KIL_NV: /* NV_f_p only (conditional) */
-            if (eval_condition(machine, inst)) {
+         }
+         break;
+      case OPCODE_ELSE:
+         /* goto ENDIF */
+         assert(inst->BranchTarget >= 0);
+         pc = inst->BranchTarget - 1;
+         break;
+      case OPCODE_ENDIF:
+         /* nothing */
+         break;
+      case OPCODE_INT:         /* float to int */
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = (GLfloat) (GLint) a[0];
+            result[1] = (GLfloat) (GLint) a[1];
+            result[2] = (GLfloat) (GLint) a[2];
+            result[3] = (GLfloat) (GLint) a[3];
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_KIL_NV:      /* NV_f_p only (conditional) */
+         if (eval_condition(machine, inst)) {
+            return GL_FALSE;
+         }
+         break;
+      case OPCODE_KIL:         /* ARB_f_p only */
+         {
+            GLfloat a[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            if (a[0] < 0.0F || a[1] < 0.0F || a[2] < 0.0F || a[3] < 0.0F) {
                return GL_FALSE;
             }
-            break;
-         case OPCODE_KIL: /* ARB_f_p only */
-            {
-               GLfloat a[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               if (a[0] < 0.0F || a[1] < 0.0F || a[2] < 0.0F || a[3] < 0.0F) {
-                  return GL_FALSE;
-               }
-            }
-            break;
-         case OPCODE_LG2:  /* log base 2 */
-            {
-               GLfloat a[4], result[4];
-               fetch_vector1( ctx, &inst->SrcReg[0], machine, a );
-               result[0] = result[1] = result[2] = result[3] = LOG2(a[0]);
-               store_vector4( inst, machine, result );
+         }
+         break;
+      case OPCODE_LG2:         /* log base 2 */
+         {
+            GLfloat a[4], result[4];
+            fetch_vector1(ctx, &inst->SrcReg[0], machine, a);
+            result[0] = result[1] = result[2] = result[3] = LOG2(a[0]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_LIT:
+         {
+            const GLfloat epsilon = 1.0F / 256.0F;      /* from NV VP spec */
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            a[0] = MAX2(a[0], 0.0F);
+            a[1] = MAX2(a[1], 0.0F);
+            /* XXX ARB version clamps a[3], NV version doesn't */
+            a[3] = CLAMP(a[3], -(128.0F - epsilon), (128.0F - epsilon));
+            result[0] = 1.0F;
+            result[1] = a[0];
+            /* XXX we could probably just use pow() here */
+            if (a[0] > 0.0F) {
+               if (a[1] == 0.0 && a[3] == 0.0)
+                  result[2] = 1.0;
+               else
+                  result[2] = EXPF(a[3] * LOGF(a[1]));
             }
-            break;
-         case OPCODE_LIT:
-            {
-               const GLfloat epsilon = 1.0F / 256.0F; /* from NV VP spec */
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               a[0] = MAX2(a[0], 0.0F);
-               a[1] = MAX2(a[1], 0.0F);
-               /* XXX ARB version clamps a[3], NV version doesn't */
-               a[3] = CLAMP(a[3], -(128.0F - epsilon), (128.0F - epsilon));
-               result[0] = 1.0F;
-               result[1] = a[0];
-               /* XXX we could probably just use pow() here */
-               if (a[0] > 0.0F) {
-                  if (a[1] == 0.0 && a[3] == 0.0)
-                     result[2] = 1.0;
-                  else
-                     result[2] = EXPF(a[3] * LOGF(a[1]));
-               }
-               else {
-                  result[2] = 0.0;
-               }
-               result[3] = 1.0F;
-               store_vector4( inst, machine, result );
-               if (DEBUG_PROG) {
-                  printf("LIT (%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]);
-               }
+            else {
+               result[2] = 0.0;
             }
-            break;
-         case OPCODE_LRP:
-            {
-               GLfloat a[4], b[4], c[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               fetch_vector4( ctx, &inst->SrcReg[2], machine, c );
-               result[0] = a[0] * b[0] + (1.0F - a[0]) * c[0];
-               result[1] = a[1] * b[1] + (1.0F - a[1]) * c[1];
-               result[2] = a[2] * b[2] + (1.0F - a[2]) * c[2];
-               result[3] = a[3] * b[3] + (1.0F - a[3]) * c[3];
-               store_vector4( inst, machine, result );
-               if (DEBUG_PROG) {
-                  printf("LRP (%g %g %g %g) = (%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],
-                         c[0], c[1], c[2], c[3]);
-               }
+            result[3] = 1.0F;
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("LIT (%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]);
             }
-            break;
-         case OPCODE_MAD:
-            {
-               GLfloat a[4], b[4], c[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               fetch_vector4( ctx, &inst->SrcReg[2], machine, c );
-               result[0] = a[0] * b[0] + c[0];
-               result[1] = a[1] * b[1] + c[1];
-               result[2] = a[2] * b[2] + c[2];
-               result[3] = a[3] * b[3] + c[3];
-               store_vector4( inst, machine, result );
-               if (DEBUG_PROG) {
-                  printf("MAD (%g %g %g %g) = (%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],
-                         c[0], c[1], c[2], c[3]);
-               }
+         }
+         break;
+      case OPCODE_LRP:
+         {
+            GLfloat a[4], b[4], c[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            fetch_vector4(ctx, &inst->SrcReg[2], machine, c);
+            result[0] = a[0] * b[0] + (1.0F - a[0]) * c[0];
+            result[1] = a[1] * b[1] + (1.0F - a[1]) * c[1];
+            result[2] = a[2] * b[2] + (1.0F - a[2]) * c[2];
+            result[3] = a[3] * b[3] + (1.0F - a[3]) * c[3];
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("LRP (%g %g %g %g) = (%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], c[0], c[1], c[2], c[3]);
             }
-            break;
-         case OPCODE_MAX:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, 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_PROG) {
-                  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_MAD:
+         {
+            GLfloat a[4], b[4], c[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            fetch_vector4(ctx, &inst->SrcReg[2], machine, c);
+            result[0] = a[0] * b[0] + c[0];
+            result[1] = a[1] * b[1] + c[1];
+            result[2] = a[2] * b[2] + c[2];
+            result[3] = a[3] * b[3] + c[3];
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               printf("MAD (%g %g %g %g) = (%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], c[0], c[1], c[2], c[3]);
             }
-            break;
-         case OPCODE_MIN:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, 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_MAX:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, 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_PROG) {
+               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_MOV:
-            {
-               GLfloat result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, result );
-               store_vector4( inst, machine, result );
-               if (DEBUG_PROG) {
-                  printf("MOV (%g %g %g %g)\n",
-                         result[0], result[1], result[2], result[3]);
-               }
+         }
+         break;
+      case OPCODE_MIN:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, 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, result);
+            store_vector4(inst, machine, result);
+            if (DEBUG_PROG) {
+               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, a );
-               fetch_vector4( ctx, &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("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_MUL:
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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("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_NOISE1:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector1( ctx, &inst->SrcReg[0], machine, a );
-               result[0] =
+         }
+         break;
+      case OPCODE_NOISE1:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector1(ctx, &inst->SrcReg[0], machine, a);
+            result[0] =
                result[1] =
-               result[2] =
-               result[3] = _slang_library_noise1(a[0]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_NOISE2:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] =
+               result[2] = result[3] = _slang_library_noise1(a[0]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_NOISE2:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] =
                result[1] =
-               result[2] =
-               result[3] = _slang_library_noise2(a[0], a[1]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_NOISE3:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] =
+               result[2] = result[3] = _slang_library_noise2(a[0], a[1]);
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_NOISE3:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] =
                result[1] =
                result[2] =
                result[3] = _slang_library_noise3(a[0], a[1], a[2]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_NOISE4:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               result[0] =
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_NOISE4:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            result[0] =
                result[1] =
                result[2] =
                result[3] = _slang_library_noise4(a[0], a[1], a[2], a[3]);
-               store_vector4( inst, machine, result );
-            }
-            break;
-         case OPCODE_NOP:
-            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, 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, 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, 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, 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, a );
-               fetch_vector1( ctx, &inst->SrcReg[1], machine, 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, a );
-               if (DEBUG_PROG) {
-                  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 (conditional) */
-            if (eval_condition(machine, inst)) {
-               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, axis );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, 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, 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_PROG) {
-                  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, 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, a );
-               fetch_vector4( ctx, &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 );
-            }
-            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( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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 );
-            }
-            break;
-         case OPCODE_SGT: /* set on greater */
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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\n",
-                         result[0], result[1], result[2], result[3]);
-               }
-            }
-            break;
-         case OPCODE_SIN:
-            {
-               GLfloat a[4], result[4];
-               fetch_vector1( ctx, &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( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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 );
-            }
-            break;
-         case OPCODE_SLT: /* set on less */
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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 );
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_NOP:
+         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, 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, 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, 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, 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, a);
+            fetch_vector1(ctx, &inst->SrcReg[1], machine, 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, a);
+            if (DEBUG_PROG) {
+               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 (conditional) */
+         if (eval_condition(machine, inst)) {
+            if (machine->StackDepth == 0) {
+               return GL_TRUE;  /* Per GL_NV_vertex_program2 spec */
             }
-            break;
-         case OPCODE_SNE: /* set on not equal */
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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 );
+            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, axis);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, 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, 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_PROG) {
+               printf("RSQ %g = 1/sqrt(|%g|)\n", result[0], a[0]);
             }
-            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_SCS:         /* sine and cos */
+         {
+            GLfloat a[4], result[4];
+            fetch_vector1(ctx, &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(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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);
+         }
+         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(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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);
+         }
+         break;
+      case OPCODE_SGT:         /* set on greater */
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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\n",
+                      result[0], result[1], result[2], result[3]);
             }
-            break;
-         case OPCODE_SUB:
-            {
-               GLfloat a[4], b[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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_SIN:
+         {
+            GLfloat a[4], result[4];
+            fetch_vector1(ctx, &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(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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);
+         }
+         break;
+      case OPCODE_SLT:         /* set on less */
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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);
+         }
+         break;
+      case OPCODE_SNE:         /* set on not equal */
+         {
+            GLfloat a[4], b[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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);
+         }
+         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(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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(ctx, 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];
+         }
+         break;
+      case OPCODE_SWZ:         /* extended swizzle */
+         {
+            const struct prog_src_register *source = &inst->SrcReg[0];
+            const GLfloat *src = get_register_pointer(ctx, 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];
                }
-               store_vector4( inst, machine, result );
+               if (source->NegateBase & (1 << i))
+                  result[i] = -result[i];
             }
-            break;
-         case OPCODE_TEX: /* Both ARB and NV frag prog */
-            /* Texel lookup */
-            {
-               /* Note: only use the precomputed lambda value when we're
-                * sampling texture unit [K] with texcoord[K].
-                * Otherwise, the lambda value may have no relation to the
-                * instruction's texcoord or texture image.  Using the wrong
-                * lambda is usually bad news.
-                * The rest of the time, just use zero (until we get a more
-                * sophisticated way of computing lambda).
-                */
-               GLfloat coord[4], color[4], lambda;
+            store_vector4(inst, machine, result);
+         }
+         break;
+      case OPCODE_TEX:         /* Both ARB and NV frag prog */
+         /* Texel lookup */
+         {
+            /* Note: only use the precomputed lambda value when we're
+             * sampling texture unit [K] with texcoord[K].
+             * Otherwise, the lambda value may have no relation to the
+             * instruction's texcoord or texture image.  Using the wrong
+             * lambda is usually bad news.
+             * The rest of the time, just use zero (until we get a more
+             * sophisticated way of computing lambda).
+             */
+            GLfloat coord[4], color[4], lambda;
 #if 0
-               if (inst->SrcReg[0].File == PROGRAM_INPUT &&
-                   inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0+inst->TexSrcUnit)
-                  lambda = span->array->lambda[inst->TexSrcUnit][column];
-               else
+            if (inst->SrcReg[0].File == PROGRAM_INPUT &&
+                inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
+               lambda = span->array->lambda[inst->TexSrcUnit][column];
+            else
 #endif
-                  lambda = 0.0;
-               fetch_vector4(ctx, &inst->SrcReg[0], machine, coord);
-               machine->FetchTexelLod(ctx, coord, lambda, inst->TexSrcUnit, color);
-               if (DEBUG_PROG) {
-                  printf("TEX (%g, %g, %g, %g) = texture[%d][%g, %g, %g, %g], "
-                         "lod %f\n",
-                         color[0], color[1], color[2], color[3],
-                         inst->TexSrcUnit,
-                         coord[0], coord[1], coord[2], coord[3], lambda);
-               }
-               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 coord[4], color[4], lambda, bias;
+               lambda = 0.0;
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, coord);
+            machine->FetchTexelLod(ctx, coord, lambda, inst->TexSrcUnit,
+                                   color);
+            if (DEBUG_PROG) {
+               printf("TEX (%g, %g, %g, %g) = texture[%d][%g, %g, %g, %g], "
+                      "lod %f\n",
+                      color[0], color[1], color[2], color[3],
+                      inst->TexSrcUnit,
+                      coord[0], coord[1], coord[2], coord[3], lambda);
+            }
+            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 coord[4], color[4], lambda, bias;
 #if 0
-               if (inst->SrcReg[0].File == PROGRAM_INPUT &&
-                   inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0+inst->TexSrcUnit)
-                  lambda = span->array->lambda[inst->TexSrcUnit][column];
-               else
+            if (inst->SrcReg[0].File == PROGRAM_INPUT &&
+                inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
+               lambda = span->array->lambda[inst->TexSrcUnit][column];
+            else
 #endif
-                  lambda = 0.0;
-               fetch_vector4(ctx, &inst->SrcReg[0], machine, coord);
-               /* coord[3] is the bias to add to lambda */
-               bias = texUnit->LodBias + coord[3];
-               if (texUnit->_Current)
-                  bias += texUnit->_Current->LodBias;
-               machine->FetchTexelLod(ctx, coord, lambda + bias,
-                                      inst->TexSrcUnit, 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( ctx, &inst->SrcReg[0], machine, texcoord );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, dtdx );
-               fetch_vector4( ctx, &inst->SrcReg[2], machine, dtdy );
-               machine->FetchTexelDeriv(ctx, texcoord, dtdx, dtdy,
-                                        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], lambda;
+               lambda = 0.0;
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, coord);
+            /* coord[3] is the bias to add to lambda */
+            bias = texUnit->LodBias + coord[3];
+            if (texUnit->_Current)
+               bias += texUnit->_Current->LodBias;
+            machine->FetchTexelLod(ctx, coord, lambda + bias,
+                                   inst->TexSrcUnit, 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(ctx, &inst->SrcReg[0], machine, texcoord);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, dtdx);
+            fetch_vector4(ctx, &inst->SrcReg[2], machine, dtdy);
+            machine->FetchTexelDeriv(ctx, texcoord, dtdx, dtdy,
+                                     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], lambda;
 #if 0
-               if (inst->SrcReg[0].File == PROGRAM_INPUT &&
-                   inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0+inst->TexSrcUnit)
-                  lambda = span->array->lambda[inst->TexSrcUnit][column];
-               else
+            if (inst->SrcReg[0].File == PROGRAM_INPUT &&
+                inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
+               lambda = span->array->lambda[inst->TexSrcUnit][column];
+            else
 #endif
-                  lambda = 0.0;
-               fetch_vector4(ctx, &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];
-	       }
-               machine->FetchTexelLod(ctx, texcoord, lambda,
-                                      inst->TexSrcUnit, color);
-               store_vector4( inst, machine, color );
-            }
-            break;
-         case OPCODE_TXP_NV: /* GL_NV_fragment_program only */
-            /* Texture lookup w/ projective divide */
-            {
-               GLfloat texcoord[4], color[4], lambda;
+               lambda = 0.0;
+            fetch_vector4(ctx, &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];
+            }
+            machine->FetchTexelLod(ctx, texcoord, lambda,
+                                   inst->TexSrcUnit, color);
+            store_vector4(inst, machine, color);
+         }
+         break;
+      case OPCODE_TXP_NV:      /* GL_NV_fragment_program only */
+         /* Texture lookup w/ projective divide */
+         {
+            GLfloat texcoord[4], color[4], lambda;
 #if 0
-               if (inst->SrcReg[0].File == PROGRAM_INPUT &&
-                   inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0+inst->TexSrcUnit)
-                  lambda = span->array->lambda[inst->TexSrcUnit][column];
-               else
+            if (inst->SrcReg[0].File == PROGRAM_INPUT &&
+                inst->SrcReg[0].Index == FRAG_ATTRIB_TEX0 + inst->TexSrcUnit)
+               lambda = span->array->lambda[inst->TexSrcUnit][column];
+            else
 #endif
-                  lambda = 0.0;
-               fetch_vector4(ctx, &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];
-               }
-               machine->FetchTexelLod(ctx, texcoord, lambda,
-                                      inst->TexSrcUnit, 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( ctx, &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( ctx, &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( ctx, &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( ctx, &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( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &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 );
-            }
-            break;
-         case OPCODE_X2D: /* 2-D matrix transform */
-            {
-               GLfloat a[4], b[4], c[4], result[4];
-               fetch_vector4( ctx, &inst->SrcReg[0], machine, a );
-               fetch_vector4( ctx, &inst->SrcReg[1], machine, b );
-               fetch_vector4( ctx, &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 );
+               lambda = 0.0;
+            fetch_vector4(ctx, &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];
+            }
+            machine->FetchTexelLod(ctx, texcoord, lambda,
+                                   inst->TexSrcUnit, 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(ctx, &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(ctx, &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(ctx, &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(ctx, &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(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &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);
+         }
+         break;
+      case OPCODE_X2D:         /* 2-D matrix transform */
+         {
+            GLfloat a[4], b[4], c[4], result[4];
+            fetch_vector4(ctx, &inst->SrcReg[0], machine, a);
+            fetch_vector4(ctx, &inst->SrcReg[1], machine, b);
+            fetch_vector4(ctx, &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(ctx, &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]);
             }
-            break;
-         case OPCODE_PRINT:
-            {
-               if (inst->SrcReg[0].File != -1) {
-                  GLfloat a[4];
-                  fetch_vector4( ctx, &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);
-               }
+            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_exec_fragment_program",
-                          inst->Opcode);
-            return GL_TRUE; /* return value doesn't matter */
+         }
+         break;
+      case OPCODE_END:
+         return GL_TRUE;
+      default:
+         _mesa_problem(ctx, "Bad opcode %d in _mesa_exec_fragment_program",
+                       inst->Opcode);
+         return GL_TRUE;        /* return value doesn't matter */
 
       }
+
       total++;
       if (total > MAX_EXEC) {
          _mesa_problem(ctx, "Infinite loop detected in fragment program");
          return GL_TRUE;
          abort();
       }
-   }
+
+   } /* for pc */
 
 #if FEATURE_MESA_program_debug
    CurrentMachine = NULL;