More GLSL code:

- use macros to access and modify render inputs bit-field;
- un-alias generic vertex attributes for ARB vertex calls;
- use MAX_VERTEX_PROGRAM_ATTRIBS (NV code) or MAX_VERTEX_ATTRIBS
  (ARB code) in place of VERT_ATTRIB_MAX;
- define VERT_ATTRIB_GENERIC0..15 for un-aliased vertex
  attributes for ARB_vertex_shader;
- fix generic attribute index range check in arbprogparse.c;
- interface GLSL varyings between vertex and fragment shader;
- use 64-bit optimised bitset (bitset.h) for render inputs;

39 files changed, 882 insertions, 583 deletions

diff --git a/src/mesa/drivers/dri/i915/i830_context.h b/src/mesa/drivers/dri/i915/i830_context.h
index 5bdcc21c5a..d5811e6c34 100644
--- a/src/mesa/drivers/dri/i915/i830_context.h
+++ b/src/mesa/drivers/dri/i915/i830_context.h
@@ -118,7 +118,7 @@ struct i830_context
 {
    struct intel_context intel;
    
-   GLuint last_index;
+   DECLARE_RENDERINPUTS(last_index_bitset);
 
    struct i830_hw_state meta, initial, state, *current;
 };
diff --git a/src/mesa/drivers/dri/i915/i830_vtbl.c b/src/mesa/drivers/dri/i915/i830_vtbl.c
index e8913e8063..9e71b11109 100644
--- a/src/mesa/drivers/dri/i915/i830_vtbl.c
+++ b/src/mesa/drivers/dri/i915/i830_vtbl.c
@@ -65,11 +65,13 @@ static void i830_render_start( intelContextPtr intel )
    i830ContextPtr i830 = I830_CONTEXT(intel);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLuint v0 = _3DSTATE_VFT0_CMD;
    GLuint v2 = _3DSTATE_VFT1_CMD;
    GLuint mcsb1 = 0;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
    /* Important:
     */
    VB->AttribPtr[VERT_ATTRIB_POS] = VB->NdcPtr;
@@ -78,7 +80,7 @@ static void i830_render_start( intelContextPtr intel )
    /* EMIT_ATTR's must be in order as they tell t_vertex.c how to
     * build up a hardware vertex.
     */
-   if (index & _TNL_BITS_TEX_ANY) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, VFT0_XYZW );
       intel->coloroffset = 4;
    }
@@ -87,36 +89,37 @@ static void i830_render_start( intelContextPtr intel )
       intel->coloroffset = 3;
    }
 
-   if (index & _TNL_BIT_POINTSIZE) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POINTSIZE )) {
       EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F, VFT0_POINT_WIDTH );
    }
 
    EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4UB_4F_BGRA, VFT0_DIFFUSE );
       
    intel->specoffset = 0;
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
-      if (index & _TNL_BIT_COLOR1) {
-	 intel->specoffset = intel->coloroffset + 1;
-	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR, VFT0_SPEC );
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
+         intel->specoffset = intel->coloroffset + 1;
+         EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR, VFT0_SPEC );
       }
-      else 
-	 EMIT_PAD( 3 );
-      
-      if (index & _TNL_BIT_FOG)
-	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, VFT0_SPEC );
       else
-	 EMIT_PAD( 1 );
+         EMIT_PAD( 3 );
+
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
+         EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, VFT0_SPEC );
+      else
+         EMIT_PAD( 1 );
    }
 
-   if (index & _TNL_BITS_TEX_ANY) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       int i, count = 0;
 
       for (i = 0; i < I830_TEX_UNITS; i++) {
- 	 if (index & _TNL_BIT_TEX(i)) {
-	    GLuint sz = VB->TexCoordPtr[i]->size;
-	    GLuint emit;
-	    GLuint mcs = (i830->state.Tex[i][I830_TEXREG_MCS] & 
-			  ~TEXCOORDTYPE_MASK);
+         if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
+            GLuint sz = VB->TexCoordPtr[i]->size;
+            GLuint emit;
+            GLuint mcs = (i830->state.Tex[i][I830_TEXREG_MCS] & 
+                          ~TEXCOORDTYPE_MASK);
 
 	    switch (sz) {
 	    case 1: 
@@ -162,7 +165,7 @@ static void i830_render_start( intelContextPtr intel )
    if (v0 != i830->state.Ctx[I830_CTXREG_VF] ||
        v2 != i830->state.Ctx[I830_CTXREG_VF2] ||
        mcsb1 != i830->state.Ctx[I830_CTXREG_MCSB1] ||
-       index != i830->last_index) {
+       !RENDERINPUTS_EQUAL( index_bitset, i830->last_index_bitset )) {
     
       I830_STATECHANGE( i830, I830_UPLOAD_CTX );
 
@@ -180,7 +183,7 @@ static void i830_render_start( intelContextPtr intel )
       i830->state.Ctx[I830_CTXREG_VF] = v0;
       i830->state.Ctx[I830_CTXREG_VF2] = v2;
       i830->state.Ctx[I830_CTXREG_MCSB1] = mcsb1;
-      i830->last_index = index;
+      RENDERINPUTS_COPY( i830->last_index_bitset, index_bitset );
 
       assert(i830_check_vertex_size( intel, intel->vertex_size ));
    }
diff --git a/src/mesa/drivers/dri/i915/i915_texprog.c b/src/mesa/drivers/dri/i915/i915_texprog.c
index 74ece96f8b..4fbce34ba2 100644
--- a/src/mesa/drivers/dri/i915/i915_texprog.c
+++ b/src/mesa/drivers/dri/i915/i915_texprog.c
@@ -576,11 +576,13 @@ void i915ValidateTextureProgram( i915ContextPtr i915 )
    GLcontext *ctx = &intel->ctx;
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    int i, offset;
    GLuint s4 = i915->state.Ctx[I915_CTXREG_LIS4] & ~S4_VFMT_MASK;
    GLuint s2 = S2_TEXCOORD_NONE;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
    /* Important:
     */
    VB->AttribPtr[VERT_ATTRIB_POS] = VB->NdcPtr;
@@ -591,9 +593,9 @@ void i915ValidateTextureProgram( i915ContextPtr i915 )
 
    if (i915->vertex_fog == I915_FOG_PIXEL) {
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, S4_VFMT_XYZW, 16 );
-      index &= ~_TNL_BIT_FOG;
+      RENDERINPUTS_CLEAR( index_bitset, _TNL_ATTRIB_FOG );
    }
-   else if (index & _TNL_BITS_TEX_ANY) {
+   else if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, S4_VFMT_XYZW, 16 );
    }
    else {
@@ -601,29 +603,30 @@ void i915ValidateTextureProgram( i915ContextPtr i915 )
    }
 
    /* How undefined is undefined? */
-   if (index & _TNL_BIT_POINTSIZE) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POINTSIZE )) {
       EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F, S4_VFMT_POINT_WIDTH, 4 );
    }
       
    intel->coloroffset = offset / 4;
    EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4UB_4F_BGRA, S4_VFMT_COLOR, 4 );
             
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
-      if (index & _TNL_BIT_COLOR1) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 intel->specoffset = offset / 4;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR, S4_VFMT_SPEC_FOG, 3 );
       } else 
 	 EMIT_PAD( 3 );
       
-      if (index & _TNL_BIT_FOG)
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, S4_VFMT_SPEC_FOG, 1 );
       else
 	 EMIT_PAD( 1 );
    }
 
-   if (index & _TNL_BITS_TEX_ANY) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       for (i = 0; i < 8; i++) {
-	 if (index & _TNL_BIT_TEX(i)) {
+	 if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
 	    int sz = VB->TexCoordPtr[i]->size;
 	    
 	    s2 &= ~S2_TEXCOORD_FMT(i, S2_TEXCOORD_FMT0_MASK);
diff --git a/src/mesa/drivers/dri/r128/r128_context.c b/src/mesa/drivers/dri/r128/r128_context.c
index 6194d128f4..2f30bd253b 100644
--- a/src/mesa/drivers/dri/r128/r128_context.c
+++ b/src/mesa/drivers/dri/r128/r128_context.c
@@ -190,7 +190,7 @@ GLboolean r128CreateContext( const __GLcontextModes *glVisual,
    rmesa->RenderIndex = -1;		/* Impossible value */
    rmesa->vert_buf = NULL;
    rmesa->num_verts = 0;
-   rmesa->tnl_state = ~0;
+   RENDERINPUTS_ONES( rmesa->tnl_state_bitset );
 
    /* Set the maximum texture size small enough that we can guarentee that
     * all texture units can bind a maximal texture and have them both in
diff --git a/src/mesa/drivers/dri/r128/r128_context.h b/src/mesa/drivers/dri/r128/r128_context.h
index b3552ac763..c51dd7fa58 100644
--- a/src/mesa/drivers/dri/r128/r128_context.h
+++ b/src/mesa/drivers/dri/r128/r128_context.h
@@ -130,7 +130,7 @@ struct r128_context {
    char *verts;			/* points to tnl->clipspace.vertex_buf */
    GLuint num_verts;
    int coloroffset, specoffset;
-   int tnl_state;	/* tnl->render_inputs for this _tnl_install_attrs */
+   DECLARE_RENDERINPUTS(tnl_state_bitset);	/* tnl->render_inputs for this _tnl_install_attrs */
 
    GLuint NewGLState;
    GLuint Fallback;
diff --git a/src/mesa/drivers/dri/r128/r128_lock.c b/src/mesa/drivers/dri/r128/r128_lock.c
index ef67bc6a43..393dd1ed74 100644
--- a/src/mesa/drivers/dri/r128/r128_lock.c
+++ b/src/mesa/drivers/dri/r128/r128_lock.c
@@ -89,7 +89,7 @@ void r128GetLock( r128ContextPtr rmesa, GLuint flags )
       driUpdateFramebufferSize(rmesa->glCtx, dPriv);
       rmesa->lastStamp = dPriv->lastStamp;
       rmesa->new_state |= R128_NEW_CLIP;
-      rmesa->tnl_state = ~0;
+      RENDERINPUTS_ONES( rmesa->tnl_state_bitset );
    }
 
    rmesa->dirty |= R128_UPLOAD_CONTEXT | R128_UPLOAD_CLIPRECTS;
diff --git a/src/mesa/drivers/dri/r128/r128_tris.c b/src/mesa/drivers/dri/r128/r128_tris.c
index 64f54682b6..4631522516 100644
--- a/src/mesa/drivers/dri/r128/r128_tris.c
+++ b/src/mesa/drivers/dri/r128/r128_tris.c
@@ -564,11 +564,13 @@ static void r128RenderStart( GLcontext *ctx )
    r128ContextPtr rmesa = R128_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLuint vc_frmt = 0;
    GLboolean fallback_projtex = GL_FALSE;
    GLuint offset = 0;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
    /* Important: */
    VB->AttribPtr[VERT_ATTRIB_POS] = VB->NdcPtr;
    rmesa->vertex_attr_count = 0;
@@ -577,7 +579,7 @@ static void r128RenderStart( GLcontext *ctx )
    /* EMIT_ATTR's must be in order as they tell t_vertex.c how to
     * build up a hardware vertex.
     */
-   if ( index & _TNL_BITS_TEX_ANY )
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX ))
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, R128_CCE_VC_FRMT_RHW, 16 );
    else
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_3F_VIEWPORT, 0, 12 );
@@ -591,28 +593,29 @@ static void r128RenderStart( GLcontext *ctx )
       R128_CCE_VC_FRMT_DIFFUSE_ARGB, 4 );
 #endif
 
-   if ( index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG) ) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 #if MESA_LITTLE_ENDIAN
-      if ( index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 rmesa->specoffset = offset;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR,
 	    R128_CCE_VC_FRMT_SPEC_FRGB, 3 );
       } else 
 	 EMIT_PAD( 3 );
 
-      if (index & _TNL_BIT_FOG)
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, R128_CCE_VC_FRMT_SPEC_FRGB,
 		    1 );
       else
 	 EMIT_PAD( 1 );
 #else
-      if (index & _TNL_BIT_FOG)
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, R128_CCE_VC_FRMT_SPEC_FRGB,
 		    1 );
       else
 	 EMIT_PAD( 1 );
 
-      if ( index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 rmesa->specoffset = offset;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_RGB,
 	    R128_CCE_VC_FRMT_SPEC_FRGB, 3 );
@@ -621,12 +624,12 @@ static void r128RenderStart( GLcontext *ctx )
 #endif
    }
 
-   if ( index & _TNL_BIT_TEX(rmesa->tmu_source[0]) ) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(rmesa->tmu_source[0]) )) {
       if ( VB->TexCoordPtr[rmesa->tmu_source[0]]->size > 2 )
 	 fallback_projtex = GL_TRUE;
       EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_2F, R128_CCE_VC_FRMT_S_T, 8 );
    }
-   if ( index & _TNL_BIT_TEX(rmesa->tmu_source[1]) ) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(rmesa->tmu_source[1]) )) {
       if ( VB->TexCoordPtr[rmesa->tmu_source[1]]->size > 2 )
 	 fallback_projtex = GL_TRUE;
       EMIT_ATTR( _TNL_ATTRIB_TEX1, EMIT_2F, R128_CCE_VC_FRMT_S2_T2, 8 );
@@ -638,7 +641,7 @@ static void r128RenderStart( GLcontext *ctx )
    /* Only need to change the vertex emit code if there has been a
     * statechange to a TNL index.
     */
-   if ( index != rmesa->tnl_state ) {
+   if (!RENDERINPUTS_EQUAL( index_bitset, rmesa->tnl_state_bitset )) {
       FLUSH_BATCH( rmesa );
       rmesa->dirty |= R128_UPLOAD_CONTEXT;
 
@@ -763,7 +766,7 @@ void r128InitTriFuncs( GLcontext *ctx )
    _tnl_init_vertices( ctx, ctx->Const.MaxArrayLockSize + 12, 
 		       (6 + 2 * ctx->Const.MaxTextureUnits) * sizeof(GLfloat) );
    rmesa->verts = (char *)tnl->clipspace.vertex_buf;
-   rmesa->tnl_state = -1;
+   RENDERINPUTS_ONES( rmesa->tnl_state_bitset );
 
    rmesa->NewGLState |= _R128_NEW_RENDER_STATE;
 }
diff --git a/src/mesa/drivers/dri/r200/r200_context.h b/src/mesa/drivers/dri/r200/r200_context.h
index faf1b96a2b..f6709d3d7f 100644
--- a/src/mesa/drivers/dri/r200/r200_context.h
+++ b/src/mesa/drivers/dri/r200/r200_context.h
@@ -897,7 +897,7 @@ struct r200_context {
    GLuint TclFallback;
    GLuint Fallback;
    GLuint NewGLState;
-   GLuint tnl_index;	/* index of bits for last tnl_install_attrs */
+   DECLARE_RENDERINPUTS(tnl_index_bitset);	/* index of bits for last tnl_install_attrs */
 
    /* Vertex buffers
     */
diff --git a/src/mesa/drivers/dri/r200/r200_swtcl.c b/src/mesa/drivers/dri/r200/r200_swtcl.c
index 7ab3f7cd03..aa78f38286 100644
--- a/src/mesa/drivers/dri/r200/r200_swtcl.c
+++ b/src/mesa/drivers/dri/r200/r200_swtcl.c
@@ -85,11 +85,12 @@ static void r200SetVertexFormat( GLcontext *ctx )
    r200ContextPtr rmesa = R200_CONTEXT( ctx );
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    int fmt_0 = 0;
    int fmt_1 = 0;
    int offset = 0;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
 
    /* Important:
     */
@@ -106,7 +107,8 @@ static void r200SetVertexFormat( GLcontext *ctx )
    /* EMIT_ATTR's must be in order as they tell t_vertex.c how to
     * build up a hardware vertex.
     */
-   if ( !rmesa->swtcl.needproj || (index & _TNL_BITS_TEX_ANY)) { /* need w coord for projected textures */
+   if ( !rmesa->swtcl.needproj ||
+       RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) { /* need w coord for projected textures */
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F, R200_VTX_XY | R200_VTX_Z0 | R200_VTX_W0 );
       offset = 4;
    }
@@ -124,10 +126,11 @@ static void r200SetVertexFormat( GLcontext *ctx )
    offset += 1;
 
    rmesa->swtcl.specoffset = 0;
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 
 #if MESA_LITTLE_ENDIAN 
-      if (index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 rmesa->swtcl.specoffset = offset;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_RGB, (R200_VTX_PK_RGBA << R200_VTX_COLOR_1_SHIFT) );
       }
@@ -135,21 +138,21 @@ static void r200SetVertexFormat( GLcontext *ctx )
 	 EMIT_PAD( 3 );
       }
 
-      if (index & _TNL_BIT_FOG) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, (R200_VTX_PK_RGBA << R200_VTX_COLOR_1_SHIFT) );
       }
       else {
 	 EMIT_PAD( 1 );
       }
 #else
-      if (index & _TNL_BIT_FOG) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, (R200_VTX_PK_RGBA << R200_VTX_COLOR_1_SHIFT) );
       }
       else {
 	 EMIT_PAD( 1 );
       }
 
-      if (index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 rmesa->swtcl.specoffset = offset;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR, (R200_VTX_PK_RGBA << R200_VTX_COLOR_1_SHIFT) );
       }
@@ -159,11 +162,11 @@ static void r200SetVertexFormat( GLcontext *ctx )
 #endif
    }
 
-   if (index & _TNL_BITS_TEX_ANY) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       int i;
 
       for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
-	 if (index & _TNL_BIT_TEX(i)) {
+	 if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
 	    GLuint sz = VB->TexCoordPtr[i]->size;
 
 	    fmt_1 |= sz << (3 * i);
@@ -179,7 +182,7 @@ static void r200SetVertexFormat( GLcontext *ctx )
       rmesa->hw.ctx.cmd[CTX_PP_FOG_COLOR] |= R200_FOG_USE_SPEC_ALPHA;
    }
 
-   if ( rmesa->tnl_index != index ||
+   if (!RENDERINPUTS_EQUAL( rmesa->tnl_index_bitset, index_bitset ) ||
 	(rmesa->hw.vtx.cmd[VTX_VTXFMT_0] != fmt_0) ||
 	(rmesa->hw.vtx.cmd[VTX_VTXFMT_1] != fmt_1) ) {
       R200_NEWPRIM(rmesa);
@@ -193,7 +196,7 @@ static void r200SetVertexFormat( GLcontext *ctx )
 			      rmesa->swtcl.vertex_attr_count,
 			      NULL, 0 );
       rmesa->swtcl.vertex_size /= 4;
-      rmesa->tnl_index = index;
+      RENDERINPUTS_COPY( rmesa->tnl_index_bitset, index_bitset );
    }
 }
 
@@ -235,12 +238,12 @@ void r200ChooseVertexState( GLcontext *ctx )
    /* HW perspective divide is a win, but tiny vertex formats are a
     * bigger one.
     */
-   if ( ((tnl->render_inputs & _TNL_BITS_TEX_ANY) == 0)
+   if (!RENDERINPUTS_TEST_RANGE( tnl->render_inputs_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )
 	|| (ctx->_TriangleCaps & (DD_TRI_LIGHT_TWOSIDE|DD_TRI_UNFILLED))) {
       rmesa->swtcl.needproj = GL_TRUE;
       vte |= R200_VTX_XY_FMT | R200_VTX_Z_FMT;
       vte &= ~R200_VTX_W0_FMT;
-      if (tnl->render_inputs & _TNL_BITS_TEX_ANY) {
+      if (RENDERINPUTS_TEST_RANGE( tnl->render_inputs_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
 	 vap &= ~R200_VAP_FORCE_W_TO_ONE;
       }
       else {
@@ -719,7 +722,7 @@ void r200Fallback( GLcontext *ctx, GLuint bit, GLboolean mode )
 	     */
 	    _tnl_invalidate_vertex_state( ctx, ~0 );
 	    _tnl_invalidate_vertices( ctx, ~0 );
-	    rmesa->tnl_index = 0;
+	    RENDERINPUTS_ZERO( rmesa->tnl_index_bitset );
 	    r200ChooseVertexState( ctx );
 	    r200ChooseRenderState( ctx );
 	 }
diff --git a/src/mesa/drivers/dri/r300/r300_context.h b/src/mesa/drivers/dri/r300/r300_context.h
index 449459b5b1..7504bd894a 100644
--- a/src/mesa/drivers/dri/r300/r300_context.h
+++ b/src/mesa/drivers/dri/r300/r300_context.h
@@ -776,7 +776,7 @@ struct r300_state {
 	GLuint *Elts;
 	struct r300_dma_region elt_dma;
 	
-	GLuint render_inputs; /* actual render inputs that R300 was configured for. 
+	DECLARE_RENDERINPUTS(render_inputs_bitset); /* actual render inputs that R300 was configured for. 
 				 They are the same as tnl->render_inputs for fixed pipeline */	
 	
 	struct {
diff --git a/src/mesa/drivers/dri/r300/r300_maos.c b/src/mesa/drivers/dri/r300/r300_maos.c
index 1aa005c720..290ffb44ea 100644
--- a/src/mesa/drivers/dri/r300/r300_maos.c
+++ b/src/mesa/drivers/dri/r300/r300_maos.c
@@ -268,8 +268,10 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 	GLuint vic_1 = 0;	/* R300_VAP_INPUT_CNTL_1 */
 	GLuint aa_vap_reg = 0; /* VAP register assignment */
 	GLuint i;
-	GLuint inputs = 0;
-	
+	DECLARE_RENDERINPUTS(inputs_bitset);
+
+	RENDERINPUTS_ZERO( inputs_bitset );
+
 #define CONFIGURE_AOS(r, f, v, sz, cn) { \
 		if (RADEON_DEBUG & DEBUG_STATE) \
 			fprintf(stderr, "Enabling "#v "\n"); \
@@ -300,23 +302,23 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 		GLuint InputsRead = CURRENT_VERTEX_SHADER(ctx)->Base.InputsRead;
 		struct r300_vertex_program *prog=(struct r300_vertex_program *)CURRENT_VERTEX_SHADER(ctx);
 		if (InputsRead & (1<<VERT_ATTRIB_POS)) {
-			inputs |= _TNL_BIT_POS;
+			RENDERINPUTS_SET( inputs_bitset, _TNL_ATTRIB_POS );
 			rmesa->state.aos[nr++].aos_reg = prog->inputs[VERT_ATTRIB_POS];
 		}
 		if (InputsRead & (1<<VERT_ATTRIB_NORMAL)) {
-			inputs |= _TNL_BIT_NORMAL;
+			RENDERINPUTS_SET( inputs_bitset, _TNL_ATTRIB_NORMAL );
 			rmesa->state.aos[nr++].aos_reg = prog->inputs[VERT_ATTRIB_NORMAL];
 		}
 		if (InputsRead & (1<<VERT_ATTRIB_COLOR0)) {
-			inputs |= _TNL_BIT_COLOR0;
+			RENDERINPUTS_SET( inputs_bitset, _TNL_ATTRIB_COLOR0 );
 			rmesa->state.aos[nr++].aos_reg = prog->inputs[VERT_ATTRIB_COLOR0];
 		}
 		if (InputsRead & (1<<VERT_ATTRIB_COLOR1)) {
-			inputs |= _TNL_BIT_COLOR1;
+			RENDERINPUTS_SET( inputs_bitset, _TNL_ATTRIB_COLOR1 );
 			rmesa->state.aos[nr++].aos_reg = prog->inputs[VERT_ATTRIB_COLOR1];
 		}
 		if (InputsRead & (1<<VERT_ATTRIB_FOG)) {
-			inputs |= _TNL_BIT_FOG;
+			RENDERINPUTS_SET( inputs_bitset, _TNL_ATTRIB_FOG );
 			rmesa->state.aos[nr++].aos_reg = prog->inputs[VERT_ATTRIB_FOG];
 		}
 		if(ctx->Const.MaxTextureUnits > 8) { /* Not sure if this can even happen... */
@@ -325,17 +327,17 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 		}
 		for (i=0;i<ctx->Const.MaxTextureUnits;i++) {
 			if (InputsRead & (1<<(VERT_ATTRIB_TEX0+i))) {
-				inputs |= _TNL_BIT_TEX0<<i;
+				RENDERINPUTS_SET( inputs_bitset, _TNL_ATTRIB_TEX(i) );
 				rmesa->state.aos[nr++].aos_reg = prog->inputs[VERT_ATTRIB_TEX0+i];
 			}
 		}
 		nr = 0;
 	} else {
-		inputs = TNL_CONTEXT(ctx)->render_inputs;
+		RENDERINPUTS_COPY( inputs_bitset, TNL_CONTEXT(ctx)->render_inputs_bitset );
 	}
-	rmesa->state.render_inputs = inputs;
+	RENDERINPUTS_COPY( rmesa->state.render_inputs_bitset, inputs_bitset );
 
-	if (inputs & _TNL_BIT_POS) {
+	if (RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_POS )) {
 		CONFIGURE_AOS(i_coords,	AOS_FORMAT_FLOAT,
 						VB->AttribPtr[VERT_ATTRIB_POS],
 						immd ? 4 : VB->AttribPtr[VERT_ATTRIB_POS].size,
@@ -344,7 +346,7 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 		vic_1 |= R300_INPUT_CNTL_POS;
 	}
 
-	if (inputs & _TNL_BIT_NORMAL) {
+	if (RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_NORMAL )) {
 		CONFIGURE_AOS(i_normal,	AOS_FORMAT_FLOAT,
 						VB->AttribPtr[VERT_ATTRIB_NORMAL],
 						immd ? 4 : VB->AttribPtr[VERT_ATTRIB_NORMAL].size,
@@ -353,7 +355,7 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 		vic_1 |= R300_INPUT_CNTL_NORMAL;
 	}
 
-	if (inputs & _TNL_BIT_COLOR0) {
+	if (RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_COLOR0 )) {
 		int emitsize=4;
 
 		if (!immd) {
@@ -376,7 +378,7 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 		vic_1 |= R300_INPUT_CNTL_COLOR;
 	}
 
-	if (inputs & _TNL_BIT_COLOR1) {
+	if (RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_COLOR1 )) {
 		int emitsize=4;
 
 		if (!immd) {
@@ -398,7 +400,7 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 	}
 
 #if 0
-	if (inputs & _TNL_BIT_FOG) {
+	if (RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_FOG )) {
 		CONFIGURE_AOS(	AOS_FORMAT_FLOAT,
 						VB->FogCoordPtr,
 						immd ? 4 : VB->FogCoordPtr->size,
@@ -408,7 +410,7 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 
 	r300->state.texture.tc_count = 0;
 	for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
-		if (inputs & (_TNL_BIT_TEX0 << i)) {
+		if (RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_TEX(i) )) {
 			CONFIGURE_AOS(i_tex[i], AOS_FORMAT_FLOAT,
 							VB->AttribPtr[VERT_ATTRIB_TEX0+i],
 							immd ? 4 : VB->AttribPtr[VERT_ATTRIB_TEX0+i].size,
@@ -531,17 +533,17 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 #if 0
 	r300->hw.vic.cmd[R300_VIC_CNTL_1]=0;
 
-	if(r300->state.render_inputs & _TNL_BIT_POS)
+	if(RENDERINPUTS_TEST( r300->state.render_inputs_bitset, _TNL_ATTRIB_POS ))
 		r300->hw.vic.cmd[R300_VIC_CNTL_1]|=R300_INPUT_CNTL_POS;
 
-	if(r300->state.render_inputs & _TNL_BIT_NORMAL)
+	if(RENDERINPUTS_TEST( r300->state.render_inputs_bitset, _TNL_ATTRIB_NORMAL ))
 		r300->hw.vic.cmd[R300_VIC_CNTL_1]|=R300_INPUT_CNTL_NORMAL;
 
-	if(r300->state.render_inputs & _TNL_BIT_COLOR0)
+	if(RENDERINPUTS_TEST( r300->state.render_inputs_bitset, _TNL_ATTRIB_COLOR0 ))
 		r300->hw.vic.cmd[R300_VIC_CNTL_1]|=R300_INPUT_CNTL_COLOR;
 
 	for(i=0;i < ctx->Const.MaxTextureUnits;i++)
-		if(r300->state.render_inputs & (_TNL_BIT_TEX0<<i))
+		if(RENDERINPUTS_TEST( r300->state.render_inputs_bitset, _TNL_ATTRIB_TEX(i) ))
 			r300->hw.vic.cmd[R300_VIC_CNTL_1]|=(R300_INPUT_CNTL_TC0<<i);
 #endif
 
@@ -573,15 +575,15 @@ void r300EmitArrays(GLcontext * ctx, GLboolean immd)
 			if(OutputsWritten & (1<<(VERT_RESULT_TEX0+i)))
 				r300->hw.vof.cmd[R300_VOF_CNTL_1] |= (4<<(3*i));
 	} else {
-		if(inputs & _TNL_BIT_POS)
+		if(RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_POS ))
 			r300->hw.vof.cmd[R300_VOF_CNTL_0] |= R300_VAP_OUTPUT_VTX_FMT_0__POS_PRESENT;
-		if(inputs & _TNL_BIT_COLOR0)
+		if(RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_COLOR0 ))
 			r300->hw.vof.cmd[R300_VOF_CNTL_0] |= R300_VAP_OUTPUT_VTX_FMT_0__COLOR_PRESENT;
-		if(inputs & _TNL_BIT_COLOR1)
+		if(RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_COLOR1 ))
 			r300->hw.vof.cmd[R300_VOF_CNTL_0] |= R300_VAP_OUTPUT_VTX_FMT_0__COLOR_1_PRESENT;
 
 		for(i=0;i < ctx->Const.MaxTextureUnits;i++)
-			if(inputs & (_TNL_BIT_TEX0<<i))
+			if(RENDERINPUTS_TEST( inputs_bitset, _TNL_ATTRIB_TEX(i) ))
 				r300->hw.vof.cmd[R300_VOF_CNTL_1]|=(4<<(3*i));
 	}
 
diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c
index 56052d64d2..bb0cb78382 100644
--- a/src/mesa/drivers/dri/r300/r300_state.c
+++ b/src/mesa/drivers/dri/r300/r300_state.c
@@ -1270,6 +1270,19 @@ void r300_setup_textures(GLcontext *ctx)
 		fprintf(stderr, "TX_ENABLE: %08x  last_hw_tmu=%d\n", r300->hw.txe.cmd[R300_TXE_ENABLE], last_hw_tmu);
 }
 
+union r300_outputs_written {
+	GLuint vp_outputs;                       /* hw_tcl_on */
+	DECLARE_RENDERINPUTS(index_bitset);      /* !hw_tcl_on */
+};
+
+static GLboolean r300_outputs_written_test (r300_outputs_written *ow, GLuint vp_result,
+                                            GLuint tnl_attrib)
+{
+	if (hw_tcl_on)
+		return ow->vp_outputs & (1 << vp_result);
+	return RENDERINPUTS_TEST( ow->index_bitset, tnl_attrib );
+}
+
 void r300_setup_rs_unit(GLcontext *ctx)
 {
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
@@ -1284,16 +1297,16 @@ void r300_setup_rs_unit(GLcontext *ctx)
 		0x00,
 		0x00
 	};
-	GLuint OutputsWritten;
+	union r300_outputs_written OutputsWritten;
 	GLuint InputsRead;
 	int fp_reg, high_rr;
 	int in_texcoords, col_interp_nr;
 	int i;
 
 	if(hw_tcl_on)
-		OutputsWritten = CURRENT_VERTEX_SHADER(ctx)->Base.OutputsWritten;
+		OutputsWritten.vp_outputs = CURRENT_VERTEX_SHADER(ctx)->Base.OutputsWritten;
 	else
-		OutputsWritten = r300->state.render_inputs;
+		RENDERINPUTS_COPY( OutputsWritten.index_bitset, r300->state.render_inputs_bitset );
 
 	if (ctx->FragmentProgram._Current)
 		InputsRead = ctx->FragmentProgram._Current->Base.InputsRead;
@@ -1324,7 +1337,7 @@ void r300_setup_rs_unit(GLcontext *ctx)
 					| (fp_reg << R300_RS_ROUTE_DEST_SHIFT);
 			high_rr = fp_reg;
 
-			if (!(OutputsWritten & (hw_tcl_on ? (1 << (VERT_RESULT_TEX0+i)) : (_TNL_BIT_TEX0<<i)))) {
+			if (!r300_outputs_written_test( &OutputsWritten, VERT_RESULT_TEX0+i, _TNL_ATTRIB_TEX(i) )) {
 				/* Passing invalid data here can lock the GPU. */
 				WARN_ONCE("fragprog wants coords for tex%d, vp doesn't provide them!\n", i);
 				//_mesa_print_program(&CURRENT_VERTEX_SHADER(ctx)->Base);
@@ -1334,12 +1347,12 @@ void r300_setup_rs_unit(GLcontext *ctx)
 			fp_reg++;
 		} 
 		/* Need to count all coords enabled at vof */
-		if (OutputsWritten & (hw_tcl_on ? (1 << (VERT_RESULT_TEX0+i)) : (_TNL_BIT_TEX0<<i)))
+		if (r300_outputs_written_test( &OutputsWritten, VERT_RESULT_TEX0+i, _TNL_ATTRIB_TEX(i) ))
 			in_texcoords++;
 	}
 
 	if (InputsRead & FRAG_BIT_COL0) {
-		if (!(OutputsWritten & (hw_tcl_on ? (1<<VERT_RESULT_COL0) : _TNL_BIT_COLOR0))) {
+		if (!r300_outputs_written_test( &OutputsWritten, VERT_RESULT_COL0, _TNL_ATTRIB_COLOR0 )) {
 			WARN_ONCE("fragprog wants col0, vp doesn't provide it\n");
 			goto out; /* FIXME */
 			//_mesa_print_program(&CURRENT_VERTEX_SHADER(ctx)->Base);
@@ -1355,7 +1368,7 @@ void r300_setup_rs_unit(GLcontext *ctx)
 	out:
 	
 	if (InputsRead & FRAG_BIT_COL1) {
-		if (!(OutputsWritten & (hw_tcl_on ? (1<<VERT_RESULT_COL1) : _TNL_BIT_COLOR1))) {
+		if (!r300_outputs_written_test( &OutputsWritten, VERT_RESULT_COL1, _TNL_ATTRIB_COLOR1 )) {
 			WARN_ONCE("fragprog wants col1, vp doesn't provide it\n");
 			//exit(-1);
 		}
@@ -1515,7 +1528,7 @@ static void r300GenerateSimpleVertexShader(r300ContextPtr r300)
 		)
 	o_reg += 2;
 	
-	if (r300->state.render_inputs & _TNL_BIT_COLOR1) {
+	if (RENDERINPUTS_TEST( r300->state.render_inputs_bitset, _TNL_ATTRIB_COLOR1 )) {
 		WRITE_OP(
 			EASY_VSF_OP(MUL, o_reg++, ALL, RESULT),
 			VSF_REG(r300->state.vap_reg.i_color[1]),
@@ -1526,7 +1539,7 @@ static void r300GenerateSimpleVertexShader(r300ContextPtr r300)
 	
 	/* Pass through texture coordinates, if any */
 	for(i=0;i < r300->radeon.glCtx->Const.MaxTextureUnits;i++)
-		if(r300->state.render_inputs & (_TNL_BIT_TEX0<<i)){
+		if (RENDERINPUTS_TEST( r300->state.render_inputs_bitset, _TNL_ATTRIB_TEX(i) )){
 			// fprintf(stderr, "i_tex[%d]=%d\n", i, r300->state.vap_reg.i_tex[i]);
 			WRITE_OP(
 				EASY_VSF_OP(MUL, o_reg++ /* 2+i */, ALL, RESULT),
diff --git a/src/mesa/drivers/dri/radeon/radeon_context.h b/src/mesa/drivers/dri/radeon/radeon_context.h
index 9abd866730..9902e60c59 100644
--- a/src/mesa/drivers/dri/radeon/radeon_context.h
+++ b/src/mesa/drivers/dri/radeon/radeon_context.h
@@ -735,7 +735,7 @@ struct radeon_context {
    GLuint TclFallback;
    GLuint Fallback;
    GLuint NewGLState;
-   GLuint tnl_index;	/* index of bits for last tnl_install_attrs */
+   DECLARE_RENDERINPUTS(tnl_index_bitset);	/* index of bits for last tnl_install_attrs */
 
    /* Vertex buffers
     */
diff --git a/src/mesa/drivers/dri/radeon/radeon_swtcl.c b/src/mesa/drivers/dri/radeon/radeon_swtcl.c
index 9924931e1a..4d5bbbd1f1 100644
--- a/src/mesa/drivers/dri/radeon/radeon_swtcl.c
+++ b/src/mesa/drivers/dri/radeon/radeon_swtcl.c
@@ -92,10 +92,11 @@ static void radeonSetVertexFormat( GLcontext *ctx )
    radeonContextPtr rmesa = RADEON_CONTEXT( ctx );
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    int fmt_0 = 0;
    int offset = 0;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
 
    /* Important:
     */
@@ -113,7 +114,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
     * build up a hardware vertex.
     */
    if ( !rmesa->swtcl.needproj ||
-        (index & _TNL_BITS_TEX_ANY)) {	/* for projtex */
+        RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {	/* for projtex */
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F, 
 		 RADEON_CP_VC_FRMT_XY |	RADEON_CP_VC_FRMT_Z | RADEON_CP_VC_FRMT_W0 );
       offset = 4;
@@ -135,10 +136,11 @@ static void radeonSetVertexFormat( GLcontext *ctx )
    offset += 1;
 
    rmesa->swtcl.specoffset = 0;
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 
 #if MESA_LITTLE_ENDIAN 
-      if (index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 rmesa->swtcl.specoffset = offset;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_RGB,
 	 	    RADEON_CP_VC_FRMT_PKSPEC );
@@ -147,7 +149,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
 	 EMIT_PAD( 3 );
       }
 
-      if (index & _TNL_BIT_FOG) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F,
 	 	    RADEON_CP_VC_FRMT_PKSPEC );
       }
@@ -155,7 +157,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
 	 EMIT_PAD( 1 );
       }
 #else
-      if (index & _TNL_BIT_FOG) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F,
 	 	    RADEON_CP_VC_FRMT_PKSPEC );
       }
@@ -163,7 +165,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
 	 EMIT_PAD( 1 );
       }
 
-      if (index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 rmesa->swtcl.specoffset = offset;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR,
 	 	    RADEON_CP_VC_FRMT_PKSPEC );
@@ -174,11 +176,11 @@ static void radeonSetVertexFormat( GLcontext *ctx )
 #endif
    }
 
-   if (index & _TNL_BITS_TEX_ANY) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       int i;
 
       for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
-	 if (index & _TNL_BIT_TEX(i)) {
+	 if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
 	    GLuint sz = VB->TexCoordPtr[i]->size;
 
 	    switch (sz) {
@@ -204,7 +206,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
       }
    }
 
-   if ( rmesa->tnl_index != index ||
+   if (!RENDERINPUTS_EQUAL( rmesa->tnl_index_bitset, index_bitset ) ||
 	fmt_0 != rmesa->swtcl.vertex_format) {
       RADEON_NEWPRIM(rmesa);
       rmesa->swtcl.vertex_format = fmt_0;
@@ -214,7 +216,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
 			      rmesa->swtcl.vertex_attr_count,
 			      NULL, 0 );
       rmesa->swtcl.vertex_size /= 4;
-      rmesa->tnl_index = index;
+      RENDERINPUTS_COPY( rmesa->tnl_index_bitset, index_bitset );
       if (RADEON_DEBUG & DEBUG_VERTS)
 	 fprintf( stderr, "%s: vertex_size= %d floats\n",
 		  __FUNCTION__, rmesa->swtcl.vertex_size);
@@ -257,8 +259,9 @@ void radeonChooseVertexState( GLcontext *ctx )
     * bigger one.
     */
 
-   if ( ((tnl->render_inputs & (_TNL_BITS_TEX_ANY|_TNL_BIT_COLOR1) ) == 0)
-	|| (ctx->_TriangleCaps & (DD_TRI_LIGHT_TWOSIDE|DD_TRI_UNFILLED))) {
+   if ((!RENDERINPUTS_TEST_RANGE( tnl->render_inputs_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX ) &&
+       !RENDERINPUTS_TEST( tnl->render_inputs_bitset, _TNL_ATTRIB_COLOR1 ))
+       || (ctx->_TriangleCaps & (DD_TRI_LIGHT_TWOSIDE|DD_TRI_UNFILLED))) {
       rmesa->swtcl.needproj = GL_TRUE;
       se_coord_fmt = (RADEON_VTX_XY_PRE_MULT_1_OVER_W0 |
 		      RADEON_VTX_Z_PRE_MULT_1_OVER_W0 |
@@ -938,7 +941,7 @@ void radeonFallback( GLcontext *ctx, GLuint bit, GLboolean mode )
 	     */
 	    _tnl_invalidate_vertex_state( ctx, ~0 );
 	    _tnl_invalidate_vertices( ctx, ~0 );
-	    rmesa->tnl_index = 0;
+	    RENDERINPUTS_ZERO( rmesa->tnl_index_bitset );
 	    radeonChooseVertexState( ctx );
 	    radeonChooseRenderState( ctx );
 	 }
diff --git a/src/mesa/drivers/dri/savage/savagetris.c b/src/mesa/drivers/dri/savage/savagetris.c
index 1847040137..3dd821a4d3 100644
--- a/src/mesa/drivers/dri/savage/savagetris.c
+++ b/src/mesa/drivers/dri/savage/savagetris.c
@@ -867,15 +867,17 @@ static GLboolean savageCheckPTexHack( GLcontext *ctx )
 {
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
 
-   if (index & _TNL_BIT_TEX(0) && VB->TexCoordPtr[0]->size == 4) {
-      if ((index & _TNL_BITS_TEX_ANY) == _TNL_BIT_TEX(0))
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 ) && VB->TexCoordPtr[0]->size == 4) {
+      if (!RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_ATTRIB_TEX1, _TNL_LAST_TEX ))
 	 return GL_TRUE; /* apply ptex hack */
       else
 	 FALLBACK(ctx, SAVAGE_FALLBACK_PROJ_TEXTURE, GL_TRUE);
    }
-   if ((index & _TNL_BIT_TEX(1)) && VB->TexCoordPtr[1]->size == 4)
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 ) && VB->TexCoordPtr[1]->size == 4)
       FALLBACK(ctx, SAVAGE_FALLBACK_PROJ_TEXTURE, GL_TRUE);
 
    return GL_FALSE; /* don't apply ptex hack */
@@ -929,10 +931,11 @@ static __inline__ GLuint savageChooseVertexFormat_s3d( GLcontext *ctx )
    savageContextPtr imesa = SAVAGE_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLuint setupIndex = SAVAGE_EMIT_XYZ;
    GLubyte skip;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
    imesa->vertex_attr_count = 0;
 
    skip = SAVAGE_SKIP_ALL_S3D;
@@ -941,7 +944,7 @@ static __inline__ GLuint savageChooseVertexFormat_s3d( GLcontext *ctx )
    /* EMIT_ATTR's must be in order as they tell t_vertex.c how to
     * build up a hardware vertex.
     */
-   if ((index & _TNL_BITS_TEX_ANY) || !(ctx->_TriangleCaps & DD_FLATSHADE))
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX ) || !(ctx->_TriangleCaps & DD_FLATSHADE))
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, SAVAGE_EMIT_W, SAVAGE_SKIP_W );
    else {
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_3F_VIEWPORT, 0, 0 );
@@ -952,17 +955,17 @@ static __inline__ GLuint savageChooseVertexFormat_s3d( GLcontext *ctx )
    /* t_context.c always includes a diffuse color */
    EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4UB_4F_BGRA, SAVAGE_EMIT_C0, SAVAGE_SKIP_C0 );
 
-   if ((index & _TNL_BIT_COLOR1))
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ))
       EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR, SAVAGE_EMIT_C1, SAVAGE_SKIP_C1 );
    else
       EMIT_PAD( 3 );
-   if ((index & _TNL_BIT_FOG))
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
       EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, SAVAGE_EMIT_FOG, SAVAGE_SKIP_C1 );
    else
       EMIT_PAD( 1 );
    skip &= ~SAVAGE_SKIP_C1;
 
-   if (index & _TNL_BIT_TEX(0)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
       if (imesa->ptexHack)
 	 EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_3F_XYW, SAVAGE_EMIT_STQ0, SAVAGE_SKIP_ST0);
       else if (VB->TexCoordPtr[0]->size == 4)
@@ -991,28 +994,27 @@ static __inline__ GLuint savageChooseVertexFormat_s4( GLcontext *ctx )
    savageContextPtr imesa = SAVAGE_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLuint setupIndex = SAVAGE_EMIT_XYZ;
    GLubyte skip;
    GLuint size, mask;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
    skip = SAVAGE_SKIP_ALL_S4;
    skip &= ~SAVAGE_SKIP_Z; /* all mesa vertices have a z coordinate */
 
-   if ((index & _TNL_BITS_TEX_ANY) || !(ctx->_TriangleCaps & DD_FLATSHADE))
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX ) || !(ctx->_TriangleCaps & DD_FLATSHADE))
       NEED_ATTR( SAVAGE_EMIT_W, SAVAGE_SKIP_W );
 
    /* t_context.c always includes a diffuse color */
    NEED_ATTR( SAVAGE_EMIT_C0, SAVAGE_SKIP_C0 );
-      
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
-      if ((index & _TNL_BIT_COLOR1))
-	 NEED_ATTR( SAVAGE_EMIT_C1, SAVAGE_SKIP_C1 );
-      if ((index & _TNL_BIT_FOG))
-	 NEED_ATTR( SAVAGE_EMIT_FOG, SAVAGE_SKIP_C1 );
-   }
 
-   if (index & _TNL_BIT_TEX(0)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ))
+      NEED_ATTR( SAVAGE_EMIT_C1, SAVAGE_SKIP_C1 );
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
+      NEED_ATTR( SAVAGE_EMIT_FOG, SAVAGE_SKIP_C1 );
+
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
       if (imesa->ptexHack)
 	 NEED_ATTR( SAVAGE_EMIT_STQ0, SAVAGE_SKIP_ST0);
       else if (VB->TexCoordPtr[0]->size == 4)
@@ -1024,7 +1026,7 @@ static __inline__ GLuint savageChooseVertexFormat_s4( GLcontext *ctx )
       else
 	 NEED_ATTR( SAVAGE_EMIT_S0, SAVAGE_SKIP_S0 );
    }
-   if (index & _TNL_BIT_TEX(1)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 )) {
       if (VB->TexCoordPtr[1]->size == 4)
 	 /* projective textures are not supported by the hardware */
 	 assert (0); /* should be caught by savageCheckPTexHack */
diff --git a/src/mesa/drivers/dri/sis/sis_context.h b/src/mesa/drivers/dri/sis/sis_context.h
index ead4a26678..c349bf96ed 100644
--- a/src/mesa/drivers/dri/sis/sis_context.h
+++ b/src/mesa/drivers/dri/sis/sis_context.h
@@ -332,7 +332,7 @@ struct sis_context
   GLint drawableID;
 
   GLint GlobalFlag;
-  GLuint last_tcl_state;
+  DECLARE_RENDERINPUTS(last_tcl_state_bitset);
 
   /* Stereo */
   GLboolean useStereo;
diff --git a/src/mesa/drivers/dri/sis/sis_tris.c b/src/mesa/drivers/dri/sis/sis_tris.c
index 24f6cb9aa8..a0e39dcd3c 100644
--- a/src/mesa/drivers/dri/sis/sis_tris.c
+++ b/src/mesa/drivers/dri/sis/sis_tris.c
@@ -842,10 +842,12 @@ static void sisRenderStart( GLcontext *ctx )
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    sisContextPtr smesa = SIS_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLuint AGPParseSet = smesa->AGPParseSet;
    GLboolean tex_fallback = GL_FALSE;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
    if (ctx->DrawBuffer->_ColorDrawBufferMask[0] == BUFFER_BIT_FRONT_LEFT && 
       smesa->driDrawable->numClipRects != 0)
    {
@@ -869,7 +871,7 @@ static void sisRenderStart( GLcontext *ctx )
 
    AGPParseSet &= ~(MASK_VertexDWSize | MASK_VertexDataFormat);
    AGPParseSet |= SiS_PS_HAS_XYZ | SiS_PS_HAS_DIFFUSE;
-   if (index & _TNL_BITS_TEX_ANY) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
       EMIT_ATTR(_TNL_ATTRIB_POS, EMIT_4F_VIEWPORT);
       AGPParseSet |= SiS_PS_HAS_W;
       smesa->coloroffset = 4;
@@ -881,17 +883,18 @@ static void sisRenderStart( GLcontext *ctx )
    EMIT_ATTR(_TNL_ATTRIB_COLOR0, EMIT_4UB_4F_BGRA);
 
    smesa->specoffset = 0;
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
       AGPParseSet |= SiS_PS_HAS_SPECULAR;
 
-      if (index & _TNL_BIT_COLOR1) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 EMIT_ATTR(_TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR);
 	 smesa->specoffset = smesa->coloroffset + 1;
       } else {
 	 EMIT_PAD(3);
       }
 
-      if (index & _TNL_BIT_FOG) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
 	 EMIT_ATTR(_TNL_ATTRIB_FOG, EMIT_1UB_1F);
       } else {
 	 EMIT_PAD(1);
@@ -899,14 +902,14 @@ static void sisRenderStart( GLcontext *ctx )
    }
 
    /* projective textures are not supported by the hardware */
-   if (index & _TNL_BIT_TEX(0)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
       if (VB->TexCoordPtr[0]->size > 2)
 	 tex_fallback = GL_TRUE;
       EMIT_ATTR(_TNL_ATTRIB_TEX0, EMIT_2F);
       AGPParseSet |= SiS_PS_HAS_UV0;
    }
    /* Will only hit tex1 on SiS300 */
-   if (index & _TNL_BIT_TEX(1)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 )) {
       if (VB->TexCoordPtr[1]->size > 2)
 	 tex_fallback = GL_TRUE;
       EMIT_ATTR(_TNL_ATTRIB_TEX1, EMIT_2F);
@@ -914,7 +917,7 @@ static void sisRenderStart( GLcontext *ctx )
    }
    FALLBACK(smesa, SIS_FALLBACK_TEXTURE, tex_fallback);
 
-   if (smesa->last_tcl_state != index) {
+   if (!RENDERINPUTS_EQUAL( smesa->last_tcl_state_bitset, index_bitset )) {
       smesa->AGPParseSet = AGPParseSet;
 
       smesa->vertex_size =  _tnl_install_attrs( ctx, smesa->vertex_attrs, 
diff --git a/src/mesa/drivers/dri/unichrome/via_tris.c b/src/mesa/drivers/dri/unichrome/via_tris.c
index 9cb88ae239..4cc7942b1b 100644
--- a/src/mesa/drivers/dri/unichrome/via_tris.c
+++ b/src/mesa/drivers/dri/unichrome/via_tris.c
@@ -744,16 +744,18 @@ static void viaChooseVertexState( GLcontext *ctx )
 {
    struct via_context *vmesa = VIA_CONTEXT(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLuint regCmdB = HC_HVPMSK_X | HC_HVPMSK_Y | HC_HVPMSK_Z;
    GLuint setupIndex = 0;
 
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
    vmesa->vertex_attr_count = 0;
  
    /* EMIT_ATTR's must be in order as they tell t_vertex.c how to
     * build up a hardware vertex.
     */
-   if (index & (_TNL_BITS_TEX_ANY|_TNL_BIT_FOG)) {
+   if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, VIA_EMIT_W, HC_HVPMSK_W );
       vmesa->coloroffset = 4;
    }
@@ -767,8 +769,9 @@ static void viaChooseVertexState( GLcontext *ctx )
 	      HC_HVPMSK_Cd );
       
    vmesa->specoffset = 0;
-   if (index & (_TNL_BIT_COLOR1|_TNL_BIT_FOG)) {
-      if ((index & _TNL_BIT_COLOR1)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ) ||
+       RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG )) {
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 )) {
 	 vmesa->specoffset = vmesa->coloroffset + 1;
 	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_3UB_3F_BGR, VIA_EMIT_SPEC, 
 		    HC_HVPMSK_Cs );
@@ -776,13 +779,13 @@ static void viaChooseVertexState( GLcontext *ctx )
       else
 	 EMIT_PAD( 3 );
 
-      if ((index & _TNL_BIT_FOG))
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
 	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1UB_1F, VIA_EMIT_FOG, HC_HVPMSK_Cs );
       else
 	 EMIT_PAD( 1 );
    }
 
-   if (index & _TNL_BIT_TEX(0)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
       if (vmesa->ptexHack)
 	 EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_3F_XYW, VIA_EMIT_PTEX0, 
 		    (HC_HVPMSK_S | HC_HVPMSK_T) );
@@ -791,7 +794,7 @@ static void viaChooseVertexState( GLcontext *ctx )
 		    (HC_HVPMSK_S | HC_HVPMSK_T) );
    }
 
-   if (index & _TNL_BIT_TEX(1)) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 )) {
       EMIT_ATTR( _TNL_ATTRIB_TEX1, EMIT_2F, VIA_EMIT_TEX1, 
 		 (HC_HVPMSK_S | HC_HVPMSK_T) );
    }
@@ -824,17 +827,19 @@ static GLboolean viaCheckPTexHack( GLcontext *ctx )
 {
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLuint index = tnl->render_inputs;
+   DECLARE_RENDERINPUTS(index_bitset);
    GLboolean fallback = GL_FALSE;
    GLboolean ptexHack = GL_FALSE;
 
-   if (index & _TNL_BIT_TEX(0) && VB->TexCoordPtr[0]->size == 4) {
-      if ((index & _TNL_BITS_TEX_ANY) == _TNL_BIT_TEX(0))
+   RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 ) && VB->TexCoordPtr[0]->size == 4) {
+      if (!RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_ATTRIB_TEX1, _TNL_LAST_TEX ))
 	 ptexHack = GL_TRUE; 
       else
 	 fallback = GL_TRUE;
    }
-   if ((index & _TNL_BIT_TEX(1)) && VB->TexCoordPtr[1]->size == 4)
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 ) && VB->TexCoordPtr[1]->size == 4)
       fallback = GL_TRUE;
 
    FALLBACK(VIA_CONTEXT(ctx), VIA_FALLBACK_PROJ_TEXTURE, fallback);
diff --git a/src/mesa/main/api_noop.c b/src/mesa/main/api_noop.c
index af7e1f9ac7..acb9341514 100644
--- a/src/mesa/main/api_noop.c
+++ b/src/mesa/main/api_noop.c
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -391,7 +391,7 @@ void GLAPIENTRY _mesa_noop_TexCoord4fv( const GLfloat *v )
 void GLAPIENTRY _mesa_noop_VertexAttrib1fNV( GLuint index, GLfloat x )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], x, 0, 0, 1);
    }
    else
@@ -401,7 +401,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib1fNV( GLuint index, GLfloat x )
 void GLAPIENTRY _mesa_noop_VertexAttrib1fvNV( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], v[0], 0, 0, 1);
    }
    else
@@ -411,7 +411,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib1fvNV( GLuint index, const GLfloat *v )
 void GLAPIENTRY _mesa_noop_VertexAttrib2fNV( GLuint index, GLfloat x, GLfloat y )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], x, y, 0, 1);
    }
    else
@@ -421,7 +421,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib2fNV( GLuint index, GLfloat x, GLfloat y
 void GLAPIENTRY _mesa_noop_VertexAttrib2fvNV( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], v[0], v[1], 0, 1);
    }
    else
@@ -432,7 +432,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib3fNV( GLuint index, GLfloat x,
                                   GLfloat y, GLfloat z )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], x, y, z, 1);
    }
    else
@@ -442,7 +442,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib3fNV( GLuint index, GLfloat x,
 void GLAPIENTRY _mesa_noop_VertexAttrib3fvNV( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], v[0], v[1], v[2], 1);
    }
    else
@@ -453,7 +453,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib4fNV( GLuint index, GLfloat x,
                                   GLfloat y, GLfloat z, GLfloat w )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], x, y, z, w);
    }
    else
@@ -463,7 +463,7 @@ void GLAPIENTRY _mesa_noop_VertexAttrib4fNV( GLuint index, GLfloat x,
 void GLAPIENTRY _mesa_noop_VertexAttrib4fvNV( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS) {
       ASSIGN_4V(ctx->Current.Attrib[index], v[0], v[1], v[2], v[3]);
    }
    else
@@ -471,15 +471,12 @@ void GLAPIENTRY _mesa_noop_VertexAttrib4fvNV( GLuint index, const GLfloat *v )
 }
 
 
-/*
- * XXX Un-alias attribs here
- */
 
 void GLAPIENTRY _mesa_noop_VertexAttrib1fARB( GLuint index, GLfloat x )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], x, 0, 0, 1);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], x, 0, 0, 1);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib1fARB" );
@@ -488,8 +485,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib1fARB( GLuint index, GLfloat x )
 void GLAPIENTRY _mesa_noop_VertexAttrib1fvARB( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], v[0], 0, 0, 1);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], v[0], 0, 0, 1);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib1fvARB" );
@@ -498,8 +495,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib1fvARB( GLuint index, const GLfloat *v )
 void GLAPIENTRY _mesa_noop_VertexAttrib2fARB( GLuint index, GLfloat x, GLfloat y )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], x, y, 0, 1);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], x, y, 0, 1);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib2fARB" );
@@ -508,8 +505,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib2fARB( GLuint index, GLfloat x, GLfloat y
 void GLAPIENTRY _mesa_noop_VertexAttrib2fvARB( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], v[0], v[1], 0, 1);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], v[0], v[1], 0, 1);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib2fvARB" );
@@ -519,8 +516,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib3fARB( GLuint index, GLfloat x,
                                   GLfloat y, GLfloat z )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], x, y, z, 1);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], x, y, z, 1);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib3fARB" );
@@ -529,8 +526,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib3fARB( GLuint index, GLfloat x,
 void GLAPIENTRY _mesa_noop_VertexAttrib3fvARB( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], v[0], v[1], v[2], 1);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], v[0], v[1], v[2], 1);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib3fvARB" );
@@ -540,8 +537,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib4fARB( GLuint index, GLfloat x,
                                   GLfloat y, GLfloat z, GLfloat w )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], x, y, z, w);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], x, y, z, w);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib4fARB" );
@@ -550,8 +547,8 @@ void GLAPIENTRY _mesa_noop_VertexAttrib4fARB( GLuint index, GLfloat x,
 void GLAPIENTRY _mesa_noop_VertexAttrib4fvARB( GLuint index, const GLfloat *v )
 {
    GET_CURRENT_CONTEXT(ctx);
-   if (index < VERT_ATTRIB_MAX) {
-      ASSIGN_4V(ctx->Current.Attrib[index], v[0], v[1], v[2], v[3]);
+   if (index < MAX_VERTEX_ATTRIBS) {
+      ASSIGN_4V(ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index], v[0], v[1], v[2], v[3]);
    }
    else
       _mesa_error( ctx, GL_INVALID_ENUM, "glVertexAttrib4fvARB" );
diff --git a/src/mesa/main/bitset.h b/src/mesa/main/bitset.h
index 1fd548bfda..8bd4526cb6 100644
--- a/src/mesa/main/bitset.h
+++ b/src/mesa/main/bitset.h
@@ -1,72 +1,122 @@
-/*

- * Mesa 3-D graphics library

- * Version:  6.5

- *

- * Copyright (C) 2006  Brian Paul   All Rights Reserved.

- *

- * Permission is hereby granted, free of charge, to any person obtaining a

- * copy of this software and associated documentation files (the "Software"),

- * to deal in the Software without restriction, including without limitation

- * the rights to use, copy, modify, merge, publish, distribute, sublicense,

- * and/or sell copies of the Software, and to permit persons to whom the

- * Software is furnished to do so, subject to the following conditions:

- *

- * The above copyright notice and this permission notice shall be included

- * in all copies or substantial portions of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- */

-

-/**

- * \file bitset.h

- * \brief Bitset of arbitrary size definitions.

- * \author Michal Krol

- */

-

-#define BITSET_WORD GLuint

-#define BITSET_WORDBITS (sizeof (BITSET_WORD) * 8)

-

-/* bitset declarations

- */

-#define BITSET_DECLARE(name, size) \

-   BITSET_WORD name[((size) + BITSET_WORDBITS - 1) / BITSET_WORDBITS]

-

-/* bitset operations

- */

-#define BITSET_COPY(x, y) _mesa_memcpy( (x), (y), sizeof (x) )

-#define BITSET_EQUAL(x, y) (_mesa_memcmp( (x), (y), sizeof (x) ) == 0)

-#define BITSET_ZERO(x) _mesa_memset( (x), 0, sizeof (x) )

-#define BITSET_ONES(x) _mesa_memset( (x), 0xff, sizeof (x) )

-

-#define BITSET_BITWORD(b) ((b) / BITSET_WORDBITS)

-#define BITSET_BIT(b) (1 << ((b) % BITSET_WORDBITS))

-

-/* single bit operations

- */

-#define BITSET_TEST(x, b) ((x)[BITSET_BITWORD(b)] & BITSET_BIT(b))

-#define BITSET_SET(x, b) ((x)[BITSET_BITWORD(b)] |= BITSET_BIT(b))

-#define BITSET_CLEAR(x, b) ((x)[BITSET_BITWORD(b)] &= ~BITSET_BIT(b))

-

-#define BITSET_MASK(b) ((b) == BITSET_WORDBITS ? ~0 : BITSET_BIT(b) - 1)

-#define BITSET_RANGE(b, e) (BITSET_MASK((e) + 1) & ~BITSET_MASK(b))

-

-/* bit range operations

- */

-#define BITSET_TEST_RANGE(x, b, e) \

-   (BITSET_BITWORD(b) == BITSET_BITWORD(e) ? \

-   ((x)[BITSET_BITWORD(b)] & BITSET_RANGE(b, e)) : \

-   (assert (!"BITSET_TEST_RANGE: bit range crosses word boundary"), 0))

-#define BITSET_SET_RANGE(x, b, e) \

-   (BITSET_BITWORD(b) == BITSET_BITWORD(e) ? \

-   ((x)[BITSET_BITWORD(b)] |= BITSET_RANGE(b, e)) : \

-   (assert (!"BITSET_SET_RANGE: bit range crosses word boundary"), 0))

-#define BITSET_CLEAR_RANGE(x, b, e) \

-   (BITSET_BITWORD(b) == BITSET_BITWORD(e) ? \

-   ((x)[BITSET_BITWORD(b)] &= ~BITSET_RANGE(b, e)) : \

-   (assert (!"BITSET_CLEAR_RANGE: bit range crosses word boundary"), 0))

-

+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * \file bitset.h
+ * \brief Bitset of arbitrary size definitions.
+ * \author Michal Krol
+ */
+ 
+/****************************************************************************
+ * generic bitset implementation
+ */
+
+#define BITSET_WORD GLuint
+#define BITSET_WORDBITS (sizeof (BITSET_WORD) * 8)
+
+/* bitset declarations
+ */
+#define BITSET_DECLARE(name, size) \
+   BITSET_WORD name[((size) + BITSET_WORDBITS - 1) / BITSET_WORDBITS]
+
+/* bitset operations
+ */
+#define BITSET_COPY(x, y) _mesa_memcpy( (x), (y), sizeof (x) )
+#define BITSET_EQUAL(x, y) (_mesa_memcmp( (x), (y), sizeof (x) ) == 0)
+#define BITSET_ZERO(x) _mesa_memset( (x), 0, sizeof (x) )
+#define BITSET_ONES(x) _mesa_memset( (x), 0xff, sizeof (x) )
+
+#define BITSET_BITWORD(b) ((b) / BITSET_WORDBITS)
+#define BITSET_BIT(b) (1 << ((b) % BITSET_WORDBITS))
+
+/* single bit operations
+ */
+#define BITSET_TEST(x, b) ((x)[BITSET_BITWORD(b)] & BITSET_BIT(b))
+#define BITSET_SET(x, b) ((x)[BITSET_BITWORD(b)] |= BITSET_BIT(b))
+#define BITSET_CLEAR(x, b) ((x)[BITSET_BITWORD(b)] &= ~BITSET_BIT(b))
+
+#define BITSET_MASK(b) ((b) == BITSET_WORDBITS ? ~0 : BITSET_BIT(b) - 1)
+#define BITSET_RANGE(b, e) (BITSET_MASK((e) + 1) & ~BITSET_MASK(b))
+
+/* bit range operations
+ */
+#define BITSET_TEST_RANGE(x, b, e) \
+   (BITSET_BITWORD(b) == BITSET_BITWORD(e) ? \
+   ((x)[BITSET_BITWORD(b)] & BITSET_RANGE(b, e)) : \
+   (assert (!"BITSET_TEST_RANGE: bit range crosses word boundary"), 0))
+#define BITSET_SET_RANGE(x, b, e) \
+   (BITSET_BITWORD(b) == BITSET_BITWORD(e) ? \
+   ((x)[BITSET_BITWORD(b)] |= BITSET_RANGE(b, e)) : \
+   (assert (!"BITSET_SET_RANGE: bit range crosses word boundary"), 0))
+#define BITSET_CLEAR_RANGE(x, b, e) \
+   (BITSET_BITWORD(b) == BITSET_BITWORD(e) ? \
+   ((x)[BITSET_BITWORD(b)] &= ~BITSET_RANGE(b, e)) : \
+   (assert (!"BITSET_CLEAR_RANGE: bit range crosses word boundary"), 0))
+
+/****************************************************************************
+ * 64-bit bitset implementation
+ */
+ 
+#define BITSET64_WORD GLuint
+#define BITSET64_WORDBITS (sizeof (BITSET64_WORD) * 8)
+
+/* bitset declarations
+ */
+#define BITSET64_DECLARE(name, size) \
+   GLuint name[2]
+
+/* bitset operations
+ */
+#define BITSET64_COPY(x, y) do { (x)[0] = (y)[0]; (x)[1] = (y)[1]; } while (0)
+#define BITSET64_EQUAL(x, y) ( (x)[0] == (y)[0] && (x)[1] == (y)[1] )
+#define BITSET64_ZERO(x) do { (x)[0] = 0; (x)[1] = 0; } while (0)
+#define BITSET64_ONES(x) do { (x)[0] = 0xFF; (x)[1] = 0xFF; } while (0)
+
+#define BITSET64_BITWORD(b) ((b) / BITSET64_WORDBITS)
+#define BITSET64_BIT(b) (1 << ((b) % BITSET64_WORDBITS))
+
+/* single bit operations
+ */
+#define BITSET64_TEST(x, b) ((x)[BITSET64_BITWORD(b)] & BITSET64_BIT(b))
+#define BITSET64_SET(x, b) ((x)[BITSET64_BITWORD(b)] |= BITSET64_BIT(b))
+#define BITSET64_CLEAR(x, b) ((x)[BITSET64_BITWORD(b)] &= ~BITSET64_BIT(b))
+
+#define BITSET64_MASK(b) ((b) == BITSET64_WORDBITS ? ~0 : BITSET64_BIT(b) - 1)
+#define BITSET64_RANGE(b, e) (BITSET64_MASK((e) + 1) & ~BITSET64_MASK(b))
+
+/* bit range operations
+ */
+#define BITSET64_TEST_RANGE(x, b, e) \
+   (BITSET64_BITWORD(b) == BITSET64_BITWORD(e) ? \
+   ((x)[BITSET64_BITWORD(b)] & BITSET64_RANGE(b, e)) : \
+   (assert (!"BITSET64_TEST_RANGE: bit range crosses word boundary"), 0))
+#define BITSET64_SET_RANGE(x, b, e) \
+   (BITSET64_BITWORD(b) == BITSET64_BITWORD(e) ? \
+   ((x)[BITSET64_BITWORD(b)] |= BITSET64_RANGE(b, e)) : \
+   (assert (!"BITSET64_SET_RANGE: bit range crosses word boundary"), 0))
+#define BITSET64_CLEAR_RANGE(x, b, e) \
+   (BITSET64_BITWORD(b) == BITSET64_BITWORD(e) ? \
+   ((x)[BITSET64_BITWORD(b)] &= ~BITSET64_RANGE(b, e)) : \
+   (assert (!"BITSET64_CLEAR_RANGE: bit range crosses word boundary"), 0))
+
diff --git a/src/mesa/main/dlist.c b/src/mesa/main/dlist.c
index dba2008ef7..4fe1bfddcc 100644
--- a/src/mesa/main/dlist.c
+++ b/src/mesa/main/dlist.c
@@ -4777,7 +4777,7 @@ save_Attr1fNV(GLenum attr, GLfloat x)
       n[2].f = x;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_PROGRAM_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 1;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, 0, 0, 1);
 
@@ -4799,7 +4799,7 @@ save_Attr2fNV(GLenum attr, GLfloat x, GLfloat y)
       n[3].f = y;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_PROGRAM_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 2;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, 0, 1);
 
@@ -4822,7 +4822,7 @@ save_Attr3fNV(GLenum attr, GLfloat x, GLfloat y, GLfloat z)
       n[4].f = z;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_PROGRAM_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 3;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, 1);
 
@@ -4846,7 +4846,7 @@ save_Attr4fNV(GLenum attr, GLfloat x, GLfloat y, GLfloat z, GLfloat w)
       n[5].f = w;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_PROGRAM_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 4;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, w);
 
@@ -4868,7 +4868,7 @@ save_Attr1fARB(GLenum attr, GLfloat x)
       n[2].f = x;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 1;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, 0, 0, 1);
 
@@ -4890,7 +4890,7 @@ save_Attr2fARB(GLenum attr, GLfloat x, GLfloat y)
       n[3].f = y;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 2;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, 0, 1);
 
@@ -4913,7 +4913,7 @@ save_Attr3fARB(GLenum attr, GLfloat x, GLfloat y, GLfloat z)
       n[4].f = z;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 3;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, 1);
 
@@ -4937,7 +4937,7 @@ save_Attr4fARB(GLenum attr, GLfloat x, GLfloat y, GLfloat z, GLfloat w)
       n[5].f = w;
    }
 
-   ASSERT(attr < VERT_ATTRIB_MAX);
+   ASSERT(attr < MAX_VERTEX_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 4;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, w);
 
@@ -5432,7 +5432,7 @@ index_error(void)
 static void GLAPIENTRY
 save_VertexAttrib1fNV(GLuint index, GLfloat x)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr1fNV(index, x);
    else
       index_error();
@@ -5441,7 +5441,7 @@ save_VertexAttrib1fNV(GLuint index, GLfloat x)
 static void GLAPIENTRY
 save_VertexAttrib1fvNV(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr1fNV(index, v[0]);
    else
       index_error();
@@ -5450,7 +5450,7 @@ save_VertexAttrib1fvNV(GLuint index, const GLfloat * v)
 static void GLAPIENTRY
 save_VertexAttrib2fNV(GLuint index, GLfloat x, GLfloat y)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr2fNV(index, x, y);
    else
       index_error();
@@ -5459,7 +5459,7 @@ save_VertexAttrib2fNV(GLuint index, GLfloat x, GLfloat y)
 static void GLAPIENTRY
 save_VertexAttrib2fvNV(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr2fNV(index, v[0], v[1]);
    else
       index_error();
@@ -5468,7 +5468,7 @@ save_VertexAttrib2fvNV(GLuint index, const GLfloat * v)
 static void GLAPIENTRY
 save_VertexAttrib3fNV(GLuint index, GLfloat x, GLfloat y, GLfloat z)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr3fNV(index, x, y, z);
    else
       index_error();
@@ -5477,7 +5477,7 @@ save_VertexAttrib3fNV(GLuint index, GLfloat x, GLfloat y, GLfloat z)
 static void GLAPIENTRY
 save_VertexAttrib3fvNV(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr3fNV(index, v[0], v[1], v[2]);
    else
       index_error();
@@ -5487,7 +5487,7 @@ static void GLAPIENTRY
 save_VertexAttrib4fNV(GLuint index, GLfloat x, GLfloat y,
                       GLfloat z, GLfloat w)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr4fNV(index, x, y, z, w);
    else
       index_error();
@@ -5496,7 +5496,7 @@ save_VertexAttrib4fNV(GLuint index, GLfloat x, GLfloat y,
 static void GLAPIENTRY
 save_VertexAttrib4fvNV(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       save_Attr4fNV(index, v[0], v[1], v[2], v[3]);
    else
       index_error();
@@ -5508,7 +5508,7 @@ save_VertexAttrib4fvNV(GLuint index, const GLfloat * v)
 static void GLAPIENTRY
 save_VertexAttrib1fARB(GLuint index, GLfloat x)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr1fARB(index, x);
    else
       index_error();
@@ -5517,7 +5517,7 @@ save_VertexAttrib1fARB(GLuint index, GLfloat x)
 static void GLAPIENTRY
 save_VertexAttrib1fvARB(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr1fARB(index, v[0]);
    else
       index_error();
@@ -5526,7 +5526,7 @@ save_VertexAttrib1fvARB(GLuint index, const GLfloat * v)
 static void GLAPIENTRY
 save_VertexAttrib2fARB(GLuint index, GLfloat x, GLfloat y)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr2fARB(index, x, y);
    else
       index_error();
@@ -5535,7 +5535,7 @@ save_VertexAttrib2fARB(GLuint index, GLfloat x, GLfloat y)
 static void GLAPIENTRY
 save_VertexAttrib2fvARB(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr2fARB(index, v[0], v[1]);
    else
       index_error();
@@ -5544,7 +5544,7 @@ save_VertexAttrib2fvARB(GLuint index, const GLfloat * v)
 static void GLAPIENTRY
 save_VertexAttrib3fARB(GLuint index, GLfloat x, GLfloat y, GLfloat z)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr3fARB(index, x, y, z);
    else
       index_error();
@@ -5553,7 +5553,7 @@ save_VertexAttrib3fARB(GLuint index, GLfloat x, GLfloat y, GLfloat z)
 static void GLAPIENTRY
 save_VertexAttrib3fvARB(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr3fARB(index, v[0], v[1], v[2]);
    else
       index_error();
@@ -5563,7 +5563,7 @@ static void GLAPIENTRY
 save_VertexAttrib4fARB(GLuint index, GLfloat x, GLfloat y, GLfloat z,
                        GLfloat w)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr4fARB(index, x, y, z, w);
    else
       index_error();
@@ -5572,7 +5572,7 @@ save_VertexAttrib4fARB(GLuint index, GLfloat x, GLfloat y, GLfloat z,
 static void GLAPIENTRY
 save_VertexAttrib4fvARB(GLuint index, const GLfloat * v)
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       save_Attr4fARB(index, v[0], v[1], v[2], v[3]);
    else
       index_error();
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index 7876cff2ed..c443411b62 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -41,6 +41,7 @@
 #include "glapitable.h"
 #include "glthread.h"
 #include "math/m_matrix.h"	/* GLmatrix */
+#include "bitset.h"
 
 
 /**
@@ -155,7 +156,19 @@ enum
    VERT_ATTRIB_GENERIC1 = 17,
    VERT_ATTRIB_GENERIC2 = 18,
    VERT_ATTRIB_GENERIC3 = 19,
-   VERT_ATTRIB_MAX = 16  /* XXX not counting generic attribs yet */
+   VERT_ATTRIB_GENERIC4 = 20,
+   VERT_ATTRIB_GENERIC5 = 21,
+   VERT_ATTRIB_GENERIC6 = 22,
+   VERT_ATTRIB_GENERIC7 = 23,
+   VERT_ATTRIB_GENERIC8 = 24,
+   VERT_ATTRIB_GENERIC9 = 25,
+   VERT_ATTRIB_GENERIC10 = 26,
+   VERT_ATTRIB_GENERIC11 = 27,
+   VERT_ATTRIB_GENERIC12 = 28,
+   VERT_ATTRIB_GENERIC13 = 29,
+   VERT_ATTRIB_GENERIC14 = 30,
+   VERT_ATTRIB_GENERIC15 = 31,
+   VERT_ATTRIB_MAX = 32
 };
 
 /**
@@ -183,12 +196,41 @@ enum
 #define VERT_BIT_GENERIC1 (1 << VERT_ATTRIB_GENERIC1)
 #define VERT_BIT_GENERIC2 (1 << VERT_ATTRIB_GENERIC2)
 #define VERT_BIT_GENERIC3 (1 << VERT_ATTRIB_GENERIC3)
+#define VERT_BIT_GENERIC4 (1 << VERT_ATTRIB_GENERIC4)
+#define VERT_BIT_GENERIC5 (1 << VERT_ATTRIB_GENERIC5)
+#define VERT_BIT_GENERIC6 (1 << VERT_ATTRIB_GENERIC6)
+#define VERT_BIT_GENERIC7 (1 << VERT_ATTRIB_GENERIC7)
+#define VERT_BIT_GENERIC8 (1 << VERT_ATTRIB_GENERIC8)
+#define VERT_BIT_GENERIC9 (1 << VERT_ATTRIB_GENERIC9)
+#define VERT_BIT_GENERIC10 (1 << VERT_ATTRIB_GENERIC10)
+#define VERT_BIT_GENERIC11 (1 << VERT_ATTRIB_GENERIC11)
+#define VERT_BIT_GENERIC12 (1 << VERT_ATTRIB_GENERIC12)
+#define VERT_BIT_GENERIC13 (1 << VERT_ATTRIB_GENERIC13)
+#define VERT_BIT_GENERIC14 (1 << VERT_ATTRIB_GENERIC14)
+#define VERT_BIT_GENERIC15 (1 << VERT_ATTRIB_GENERIC15)
 
 #define VERT_BIT_TEX(u)  (1 << (VERT_ATTRIB_TEX0 + (u)))
 #define VERT_BIT_GENERIC(g)  (1 << (VERT_ATTRIB_GENERIC0 + (g)))
 /*@}*/
 
 /**
+ * GLSL allows shader writers to allocate vertex result attributes (varyings) in
+ * single float component granularity. This is in contrast to vertex / fragment
+ * programs, where result attributes (actually texcoords) were allocated
+ * in 4-component vectors of floats granularity.
+ * For performance reasons, it would be optimal to stick with this scheme on a scalar
+ * processor. Varyings will likely be allocated as 3-component vectors, so statistically
+ * we win 2 floats.
+ * The constant VARYINGS_PER_VECTOR tells us how much of float components we pack into
+ * one result vector. For scalar processor it would be 1, for vector processor - 4.
+ * 
+ * NOTE: Currently we pack varyings into vertex attributes.
+ */
+#define VARYINGS_PER_VECTOR 2
+#define VARYING_EMIT_STYLE  EMIT_2F
+#define MAX_VARYING_VECTORS ((MAX_VARYING_FLOATS + VARYINGS_PER_VECTOR - 1) / VARYINGS_PER_VECTOR)
+
+/**
  * Indexes for vertex program result attributes
  */
 /*@{*/
diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index 7e452bdfd7..761a45f2bd 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -820,7 +820,11 @@ update_arrays( GLcontext *ctx )
    /* find min of _MaxElement values for all enabled arrays */
 
    /* 0 */
-   if (ctx->VertexProgram._Enabled
+   if (ctx->ShaderObjects._VertexShaderPresent
+       && ctx->Array.VertexAttrib[VERT_ATTRIB_GENERIC0].Enabled) {
+      min = ctx->Array.VertexAttrib[VERT_ATTRIB_GENERIC0]._MaxElement;
+   }
+   else if (ctx->VertexProgram._Enabled
        && ctx->Array.VertexAttrib[VERT_ATTRIB_POS].Enabled) {
       min = ctx->Array.VertexAttrib[VERT_ATTRIB_POS]._MaxElement;
    }
@@ -888,7 +892,7 @@ update_arrays( GLcontext *ctx )
    }
 
    /* 8..15 */
-   for (i = VERT_ATTRIB_TEX0; i < VERT_ATTRIB_MAX; i++) {
+   for (i = VERT_ATTRIB_TEX0; i <= VERT_ATTRIB_TEX7; i++) {
       if (ctx->VertexProgram._Enabled
           && ctx->Array.VertexAttrib[i].Enabled) {
          min = MIN2(min, ctx->Array.VertexAttrib[i]._MaxElement);
@@ -899,6 +903,15 @@ update_arrays( GLcontext *ctx )
       }
    }
 
+   /* 16..31 */
+   if (ctx->ShaderObjects._VertexShaderPresent) {
+      for (i = VERT_ATTRIB_GENERIC0; i < VERT_ATTRIB_MAX; i++) {
+         if (ctx->Array.VertexAttrib[i].Enabled) {
+            min = MIN2(min, ctx->Array.VertexAttrib[i]._MaxElement);
+         }
+      }
+   }
+
    if (ctx->Array.Index.Enabled) {
       min = MIN2(min, ctx->Array.Index._MaxElement);
    }
diff --git a/src/mesa/main/varray.c b/src/mesa/main/varray.c
index cf64fb23df..9f8dc766c9 100644
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.1
+ * Version:  6.5
  *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -441,7 +441,7 @@ _mesa_VertexAttribPointerNV(GLuint index, GLint size, GLenum type,
    GET_CURRENT_CONTEXT(ctx);
    ASSERT_OUTSIDE_BEGIN_END(ctx);
 
-   if (index >= VERT_ATTRIB_MAX) {
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glVertexAttribPointerNV(index)");
       return;
    }
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index a7068e6dbc..f8809c576f 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -856,7 +856,7 @@ parse_generic_attrib_num(GLcontext *ctx, GLubyte ** inst,
 {
    GLint i = parse_integer(inst, Program);
 
-   if ((i < 0) || (i > MAX_VERTEX_PROGRAM_ATTRIBS))
+   if ((i < 0) || (i >= MAX_VERTEX_PROGRAM_ATTRIBS))
    {
       _mesa_set_program_error (ctx, Program->Position,
                                "Invalid generic vertex attribute index");
diff --git a/src/mesa/shader/arbprogram.c b/src/mesa/shader/arbprogram.c
index a7b26e72b5..9cc5488b2b 100644
--- a/src/mesa/shader/arbprogram.c
+++ b/src/mesa/shader/arbprogram.c
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -123,11 +123,8 @@ _mesa_GetVertexAttribfvARB(GLuint index, GLenum pname, GLfloat *params)
          params[0] = ctx->Array.VertexAttrib[index].Normalized;
          break;
       case GL_CURRENT_VERTEX_ATTRIB_ARB:
-	 FLUSH_CURRENT(ctx, 0);
-         /* XXX should read:
-            COPY_4V(params, ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index]);
-          */
-         COPY_4V(params, ctx->Current.Attrib[index]);
+         FLUSH_CURRENT(ctx, 0);
+         COPY_4V(params, ctx->Current.Attrib[VERT_ATTRIB_GENERIC0 + index]);
          break;
       case GL_VERTEX_ATTRIB_ARRAY_BUFFER_BINDING_ARB:
          if (!ctx->Extensions.ARB_vertex_buffer_object) {
diff --git a/src/mesa/shader/nvvertexec.c b/src/mesa/shader/nvvertexec.c
index 0e0e459b28..fc5837812e 100644
--- a/src/mesa/shader/nvvertexec.c
+++ b/src/mesa/shader/nvvertexec.c
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -51,7 +51,7 @@ _mesa_init_vp_per_vertex_registers(GLcontext *ctx)
 {
    /* Input registers get initialized from the current vertex attribs */
    MEMCPY(ctx->VertexProgram.Inputs, ctx->Current.Attrib,
-          VERT_ATTRIB_MAX * 4 * sizeof(GLfloat));
+          MAX_VERTEX_PROGRAM_ATTRIBS * 4 * sizeof(GLfloat));
 
    if (ctx->VertexProgram.Current->IsNVProgram) {
       GLuint i;
diff --git a/src/mesa/swrast/s_arbshader.c b/src/mesa/swrast/s_arbshader.c
index ba146dd1bb..9cd96c5a36 100644
--- a/src/mesa/swrast/s_arbshader.c
+++ b/src/mesa/swrast/s_arbshader.c
@@ -1,99 +1,109 @@
-/*

- * Mesa 3-D graphics library

- * Version:  6.5

- *

- * Copyright (C) 2006  Brian Paul   All Rights Reserved.

- *

- * Permission is hereby granted, free of charge, to any person obtaining a

- * copy of this software and associated documentation files (the "Software"),

- * to deal in the Software without restriction, including without limitation

- * the rights to use, copy, modify, merge, publish, distribute, sublicense,

- * and/or sell copies of the Software, and to permit persons to whom the

- * Software is furnished to do so, subject to the following conditions:

- *

- * The above copyright notice and this permission notice shall be included

- * in all copies or substantial portions of the Software.

- *

- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

- *

- * Authors:

- *    Michal Krol

- */

-

-#include "glheader.h"

-#include "context.h"

-#include "colormac.h"

-#include "s_arbshader.h"

-#include "s_context.h"

-#include "shaderobjects.h"

-#include "shaderobjects_3dlabs.h"

-#include "slang_utility.h"

-#include "slang_link.h"

-

-void _swrast_exec_arbshader (GLcontext *ctx, struct sw_span *span)

-{

-	struct gl2_program_intf **pro;

-	GLuint i;

-

-	if (!ctx->ShaderObjects._FragmentShaderPresent)

-		return;

-	pro = ctx->ShaderObjects.CurrentProgram;

-	if (!ctx->ShaderObjects._VertexShaderPresent)

-		(**pro).UpdateFixedUniforms (pro);

-

-	for (i = span->start; i < span->end; i++)

-	{

-		/* only run shader on active fragments */

-		if (span->array->mask[i]) {

-		GLfloat vec[4];

-		GLuint j;

-		GLboolean discard;

-

-		vec[0] = (GLfloat) span->x + i;

-		vec[1] = (GLfloat) span->y;

-		vec[2] = (GLfloat) span->array->z[i] / ctx->DrawBuffer->_DepthMaxF;

-		vec[3] = span->w + span->dwdx * i;

-		(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_FRAGCOORD, vec, 0,

-			4 * sizeof (GLfloat), GL_TRUE);

-		vec[0] = CHAN_TO_FLOAT(span->array->rgba[i][RCOMP]);

-		vec[1] = CHAN_TO_FLOAT(span->array->rgba[i][GCOMP]);

-		vec[2] = CHAN_TO_FLOAT(span->array->rgba[i][BCOMP]);

-		vec[3] = CHAN_TO_FLOAT(span->array->rgba[i][ACOMP]);

-		(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_COLOR, vec, 0, 4 * sizeof (GLfloat),

-			GL_TRUE);

-		for (j = 0; j < ctx->Const.MaxTextureCoordUnits; j++)

-		{

-			vec[0] = span->array->texcoords[j][i][0];

-			vec[1] = span->array->texcoords[j][i][1];

-			vec[2] = span->array->texcoords[j][i][2];

-			vec[3] = span->array->texcoords[j][i][3];

-			(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_TEXCOORD, vec, j,

-				4 * sizeof (GLfloat), GL_TRUE);

-		}

-

-		_slang_exec_fragment_shader (pro);

-

-		_slang_fetch_discard (pro, &discard);

-		if (discard)

-		{

-			span->array->mask[i] = GL_FALSE;

-			span->writeAll = GL_FALSE;

-		}

-		else

-		{

-			(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_FRAGCOLOR, vec, 0,

-				4 * sizeof (GLfloat), GL_FALSE);

-			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][RCOMP], vec[0]);

-			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][GCOMP], vec[1]);

-			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][BCOMP], vec[2]);

-			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][ACOMP], vec[3]);

-		}

-		}

-	}

-}

-

+/*
+ * Mesa 3-D graphics library
+ * Version:  6.5
+ *
+ * Copyright (C) 2006  Brian Paul   All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included
+ * in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Michal Krol
+ */
+
+#include "glheader.h"
+#include "context.h"
+#include "colormac.h"
+#include "s_arbshader.h"
+#include "s_context.h"
+#include "shaderobjects.h"
+#include "shaderobjects_3dlabs.h"
+#include "slang_utility.h"
+#include "slang_link.h"
+
+void _swrast_exec_arbshader (GLcontext *ctx, struct sw_span *span)
+{
+	struct gl2_program_intf **pro;
+	GLuint i;
+
+	if (!ctx->ShaderObjects._FragmentShaderPresent)
+		return;
+	pro = ctx->ShaderObjects.CurrentProgram;
+	if (!ctx->ShaderObjects._VertexShaderPresent)
+		(**pro).UpdateFixedUniforms (pro);
+
+	for (i = span->start; i < span->end; i++)
+	{
+		/* only run shader on active fragments */
+		if (span->array->mask[i]) {
+		GLfloat vec[4];
+		GLuint j;
+		GLboolean discard;
+
+		vec[0] = (GLfloat) span->x + i;
+		vec[1] = (GLfloat) span->y;
+		vec[2] = (GLfloat) span->array->z[i] / ctx->DrawBuffer->_DepthMaxF;
+		vec[3] = span->w + span->dwdx * i;
+		(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_FRAGCOORD, vec, 0,
+			4 * sizeof (GLfloat), GL_TRUE);
+		vec[0] = CHAN_TO_FLOAT(span->array->rgba[i][RCOMP]);
+		vec[1] = CHAN_TO_FLOAT(span->array->rgba[i][GCOMP]);
+		vec[2] = CHAN_TO_FLOAT(span->array->rgba[i][BCOMP]);
+		vec[3] = CHAN_TO_FLOAT(span->array->rgba[i][ACOMP]);
+		(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_COLOR, vec, 0, 4 * sizeof (GLfloat),
+			GL_TRUE);
+		for (j = 0; j < ctx->Const.MaxTextureCoordUnits; j++)
+		{
+			vec[0] = span->array->texcoords[j][i][0];
+			vec[1] = span->array->texcoords[j][i][1];
+			vec[2] = span->array->texcoords[j][i][2];
+			vec[3] = span->array->texcoords[j][i][3];
+			(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_TEXCOORD, vec, j,
+				4 * sizeof (GLfloat), GL_TRUE);
+		}
+		for (j = 0; j < MAX_VARYING_VECTORS; j++)
+		{
+			GLuint k;
+
+			for (k = 0; k < VARYINGS_PER_VECTOR; k++)
+			{
+				(**pro).UpdateVarying (pro, j * VARYINGS_PER_VECTOR + k,
+					&span->array->varying[i][j][k], GL_FALSE);
+			}
+		}
+
+		_slang_exec_fragment_shader (pro);
+
+		_slang_fetch_discard (pro, &discard);
+		if (discard)
+		{
+			span->array->mask[i] = GL_FALSE;
+			span->writeAll = GL_FALSE;
+		}
+		else
+		{
+			(**pro).UpdateFixedVarying (pro, SLANG_FRAGMENT_FIXED_FRAGCOLOR, vec, 0,
+				4 * sizeof (GLfloat), GL_FALSE);
+			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][RCOMP], vec[0]);
+			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][GCOMP], vec[1]);
+			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][BCOMP], vec[2]);
+			UNCLAMPED_FLOAT_TO_CHAN(span->array->rgba[i][ACOMP], vec[3]);
+		}
+		}
+	}
+}
+
diff --git a/src/mesa/swrast/s_context.h b/src/mesa/swrast/s_context.h
index 5d9a35043b..3c5a4c3222 100644
--- a/src/mesa/swrast/s_context.h
+++ b/src/mesa/swrast/s_context.h
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -66,6 +66,7 @@
 #define SPAN_FLAT         0x400  /**< flat shading? */
 #define SPAN_XY           0x800
 #define SPAN_MASK        0x1000
+#define SPAN_VARYING     0x2000
 /*@}*/
 
 
@@ -90,6 +91,7 @@ struct span_arrays {
    GLfloat texcoords[MAX_TEXTURE_COORD_UNITS][MAX_WIDTH][4];
    GLfloat lambda[MAX_TEXTURE_COORD_UNITS][MAX_WIDTH];
    GLfloat coverage[MAX_WIDTH];
+   GLfloat varying[MAX_WIDTH][MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
 
    /** This mask indicates which fragments are alive or culled */
    GLubyte mask[MAX_WIDTH];
@@ -167,6 +169,9 @@ struct sw_span {
    GLfloat texStepX[MAX_TEXTURE_COORD_UNITS][4];
    GLfloat texStepY[MAX_TEXTURE_COORD_UNITS][4];
    GLfixed intTex[2], intTexStep[2];  /* s, t only */
+   GLfloat var[MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
+   GLfloat varStepX[MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
+   GLfloat varStepY[MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
 
    /* partial derivatives wrt X and Y. */
    GLfloat dzdx, dzdy;
diff --git a/src/mesa/swrast/s_span.c b/src/mesa/swrast/s_span.c
index 02901a7ed9..11457723b0 100644
--- a/src/mesa/swrast/s_span.c
+++ b/src/mesa/swrast/s_span.c
@@ -37,7 +37,7 @@
 #include "imports.h"
 
 #include "s_atifragshader.h"
-#include "s_alpha.h"

+#include "s_alpha.h"
 #include "s_arbshader.h"
 #include "s_blend.h"
 #include "s_context.h"
@@ -434,7 +434,8 @@ interpolate_texcoords(GLcontext *ctx, struct sw_span *span)
                GLfloat r = span->tex[u][2];
                GLfloat q = span->tex[u][3];
                GLuint i;
-               if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled) {
+               if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled ||
+                   ctx->ShaderObjects._FragmentShaderPresent) {
                   /* do perspective correction but don't divide s, t, r by q */
                   const GLfloat dwdx = span->dwdx;
                   GLfloat w = span->w;
@@ -485,7 +486,8 @@ interpolate_texcoords(GLcontext *ctx, struct sw_span *span)
                GLfloat r = span->tex[u][2];
                GLfloat q = span->tex[u][3];
                GLuint i;
-               if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled) {
+               if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled ||
+                   ctx->ShaderObjects._FragmentShaderPresent) {
                   /* do perspective correction but don't divide s, t, r by q */
                   const GLfloat dwdx = span->dwdx;
                   GLfloat w = span->w;
@@ -568,7 +570,8 @@ interpolate_texcoords(GLcontext *ctx, struct sw_span *span)
          GLfloat r = span->tex[0][2];
          GLfloat q = span->tex[0][3];
          GLuint i;
-         if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled) {
+         if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled ||
+             ctx->ShaderObjects._FragmentShaderPresent) {
             /* do perspective correction but don't divide s, t, r by q */
             const GLfloat dwdx = span->dwdx;
             GLfloat w = span->w;
@@ -619,7 +622,8 @@ interpolate_texcoords(GLcontext *ctx, struct sw_span *span)
          GLfloat r = span->tex[0][2];
          GLfloat q = span->tex[0][3];
          GLuint i;
-         if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled) {
+         if (ctx->FragmentProgram._Active || ctx->ATIFragmentShader._Enabled ||
+             ctx->ShaderObjects._FragmentShaderPresent) {
             /* do perspective correction but don't divide s, t, r by q */
             const GLfloat dwdx = span->dwdx;
             GLfloat w = span->w;
@@ -668,6 +672,38 @@ interpolate_texcoords(GLcontext *ctx, struct sw_span *span)
 
 
 /**
+ * Fill in the span.varying array from the interpolation values.
+ */
+static void
+interpolate_varying(GLcontext *ctx, struct sw_span *span)
+{
+   GLuint i, j;
+
+   ASSERT(span->interpMask & SPAN_VARYING);
+   ASSERT(!(span->arrayMask & SPAN_VARYING));
+
+   span->arrayMask |= SPAN_VARYING;
+
+   for (i = 0; i < MAX_VARYING_VECTORS; i++) {
+      for (j = 0; j < VARYINGS_PER_VECTOR; j++) {
+         const GLfloat dvdx = span->varStepX[i][j];
+         GLfloat v = span->var[i][j];
+         const GLfloat dwdx = span->dwdx;
+         GLfloat w = span->w;
+         GLuint k;
+
+         for (k = 0; k < span->end; k++) {
+            GLfloat invW = 1.0f / w;
+            span->array->varying[k][i][j] = v * invW;
+            v += dvdx;
+            w += dwdx;
+         }
+      }
+   }
+}
+
+
+/**
  * Apply the current polygon stipple pattern to a span of pixels.
  */
 static void
@@ -1139,6 +1175,10 @@ _swrast_write_rgba_span( GLcontext *ctx, struct sw_span *span)
       interpolate_texcoords(ctx, span);
    }
 
+   if (ctx->ShaderObjects._FragmentShaderPresent) {
+      interpolate_varying(ctx, span);
+   }
+
    /* This is the normal place to compute the resulting fragment color/Z.
     * As an optimization, we try to defer this until after Z/stencil
     * testing in order to try to avoid computing colors that we won't
@@ -1155,11 +1195,11 @@ _swrast_write_rgba_span( GLcontext *ctx, struct sw_span *span)
       if (span->interpMask & SPAN_FOG)
          interpolate_fog(ctx, span);
 
-      /* Compute fragment colors with fragment program or texture lookups */

-      if (ctx->ShaderObjects._FragmentShaderPresent) {

-         if (span->interpMask & SPAN_Z)

-            _swrast_span_interpolate_z (ctx, span);

-         _swrast_exec_arbshader (ctx, span);

+      /* Compute fragment colors with fragment program or texture lookups */
+      if (ctx->ShaderObjects._FragmentShaderPresent) {
+         if (span->interpMask & SPAN_Z)
+            _swrast_span_interpolate_z (ctx, span);
+         _swrast_exec_arbshader (ctx, span);
       }
       else if (ctx->FragmentProgram._Active) {
          /* frag prog may need Z values */
@@ -1240,11 +1280,11 @@ _swrast_write_rgba_span( GLcontext *ctx, struct sw_span *span)
       if (span->interpMask & SPAN_FOG)
          interpolate_fog(ctx, span);
 
-      if (ctx->ShaderObjects._FragmentShaderPresent) {

-         if (span->interpMask & SPAN_Z)

-            _swrast_span_interpolate_z (ctx, span);

-         _swrast_exec_arbshader (ctx, span);

-      }

+      if (ctx->ShaderObjects._FragmentShaderPresent) {
+         if (span->interpMask & SPAN_Z)
+            _swrast_span_interpolate_z (ctx, span);
+         _swrast_exec_arbshader (ctx, span);
+      }
       else if (ctx->FragmentProgram._Active)
          _swrast_exec_fragment_program( ctx, span );
       else if (ctx->ATIFragmentShader._Enabled)
diff --git a/src/mesa/swrast/s_tritemp.h b/src/mesa/swrast/s_tritemp.h
index 7f581561ec..29a7a94da1 100644
--- a/src/mesa/swrast/s_tritemp.h
+++ b/src/mesa/swrast/s_tritemp.h
@@ -28,7 +28,7 @@
  * This file is #include'd to generate custom triangle rasterizers.
  *
  * The following macros may be defined to indicate what auxillary information
- * must be interplated across the triangle:
+ * must be interpolated across the triangle:
  *    INTERP_Z        - if defined, interpolate vertex Z values
  *    INTERP_W        - if defined, interpolate vertex W values
  *    INTERP_FOG      - if defined, interpolate fog values
@@ -41,6 +41,7 @@
  *    INTERP_TEX      - if defined, interpolate set 0 float STRQ texcoords
  *                         NOTE:  OpenGL STRQ = Mesa STUV (R was taken for red)
  *    INTERP_MULTITEX - if defined, interpolate N units of STRQ texcoords
+ *    INTERP_VARYING  - if defined, interpolate M floats of GLSL varyings
  *
  * When one can directly address pixels in the color buffer the following
  * macros can be defined and used to compute pixel addresses during
@@ -142,6 +143,20 @@
 
 
 
+#ifdef INTERP_VARYING
+#define VARYING_LOOP(CODE)\
+   {\
+      GLuint iv, ic;\
+      for (iv = 0; iv < MAX_VARYING_VECTORS; iv++) {\
+         for (ic = 0; ic < VARYINGS_PER_VECTOR; ic++) {\
+            CODE\
+         }\
+      }\
+   }
+#endif
+
+
+
 /*
  * Some code we unfortunately need to prevent negative interpolated colors.
  */
@@ -648,6 +663,19 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
          )
       }
 #endif
+#ifdef INTERP_VARYING
+      span.interpMask |= SPAN_VARYING;
+      {
+         /* win[3] is 1/W */
+         const GLfloat wMax = vMax->win[3], wMin = vMin->win[3], wMid = vMid->win[3];
+         VARYING_LOOP(
+            GLfloat eMaj_dvar = vMax->attribute[iv][ic] * wMax - vMin->attribute[iv][ic] * wMin;
+            GLfloat eBot_dvar = vMid->attribute[iv][ic] * wMid - vMin->attribute[iv][ic] * wMin;
+            span.varStepX[iv][ic] = oneOverArea * (eMaj_dvar * eBot.dy - eMaj.dy * eBot_dvar);
+            span.varStepY[iv][ic] = oneOverArea * (eMaj.dx * eBot_dvar - eMaj_dvar * eBot.dx);
+         )
+      }
+#endif
 
       /*
        * We always sample at pixel centers.  However, we avoid
@@ -751,6 +779,11 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
          GLfloat duOuter[MAX_TEXTURE_COORD_UNITS], duInner[MAX_TEXTURE_COORD_UNITS];
          GLfloat dvOuter[MAX_TEXTURE_COORD_UNITS], dvInner[MAX_TEXTURE_COORD_UNITS];
 #endif
+#ifdef INTERP_VARYING
+         GLfloat varLeft[MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
+         GLfloat dvarOuter[MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
+         GLfloat dvarInner[MAX_VARYING_VECTORS][VARYINGS_PER_VECTOR];
+#endif
 
          for (subTriangle=0; subTriangle<=1; subTriangle++) {
             EdgeT *eLeft, *eRight;
@@ -1024,6 +1057,15 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
                   dvOuter[u] = span.texStepY[u][3] + dxOuter * span.texStepX[u][3];
                )
 #endif
+#ifdef INTERP_VARYING
+               VARYING_LOOP(
+                  const GLfloat invW = vLower->win[3];
+                  const GLfloat var0 = vLower->attribute[iv][ic] * invW;
+                  varLeft[iv][ic] = var0 + (span.varStepX[iv][ic] * adjx +
+                     span.varStepY[iv][ic] * adjy) * (1.0f / FIXED_SCALE);
+                  dvarOuter[iv][ic] = span.varStepY[iv][ic] + dxOuter * span.varStepX[iv][ic];
+               )
+#endif
             } /*if setupLeft*/
 
 
@@ -1086,6 +1128,11 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
                dvInner[u] = dvOuter[u] + span.texStepX[u][3];
             )
 #endif
+#ifdef INTERP_VARYING
+            VARYING_LOOP(
+               dvarInner[iv][ic] = dvarOuter[iv][ic] + span.varStepX[iv][ic];
+            )
+#endif
 
             while (lines > 0) {
                /* initialize the span interpolants to the leftmost value */
@@ -1135,6 +1182,11 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
                   span.tex[u][3] = vLeft[u];
                )
 #endif
+#ifdef INTERP_VARYING
+               VARYING_LOOP(
+                  span.var[iv][ic] = varLeft[iv][ic];
+               )
+#endif
 
                /* This is where we actually generate fragments */
                /* XXX the test for span.y > 0 _shouldn't_ be needed but
@@ -1223,6 +1275,11 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
                      vLeft[u] += dvOuter[u];
                   )
 #endif
+#ifdef INTERP_VARYING
+                  VARYING_LOOP(
+                     varLeft[iv][ic] += dvarOuter[iv][ic];
+                  )
+#endif
                }
                else {
 #ifdef PIXEL_ADDRESS
@@ -1268,6 +1325,11 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
                      vLeft[u] += dvInner[u];
                   )
 #endif
+#ifdef INTERP_VARYING
+                  VARYING_LOOP(
+                     varLeft[iv][ic] += dvarInner[iv][ic];
+                  )
+#endif
                }
             } /*while lines>0*/
 
@@ -1299,7 +1361,9 @@ static void NAME(GLcontext *ctx, const SWvertex *v0,
 #undef INTERP_INT_TEX
 #undef INTERP_TEX
 #undef INTERP_MULTITEX
+#undef INTERP_VARYING
 #undef TEX_UNIT_LOOP
+#undef VARYING_LOOP
 
 #undef S_SCALE
 #undef T_SCALE
diff --git a/src/mesa/swrast/swrast.h b/src/mesa/swrast/swrast.h
index 4422195a52..2a21234253 100644
--- a/src/mesa/swrast/swrast.h
+++ b/src/mesa/swrast/swrast.h
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -73,6 +73,7 @@ typedef struct {
    GLfloat fog;
    GLfloat index;
    GLfloat pointSize;
+   GLfloat attribute[MAX_VERTEX_ATTRIBS][4];
 } SWvertex;
 
 
diff --git a/src/mesa/swrast_setup/ss_context.c b/src/mesa/swrast_setup/ss_context.c
index c63646ca97..7b25d14a0c 100644
--- a/src/mesa/swrast_setup/ss_context.c
+++ b/src/mesa/swrast_setup/ss_context.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.1
+ * Version:  6.5
  *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -93,7 +93,6 @@ do {						\
    e++;						\
 } while (0)
 
-
 /*
  * We patch this function into tnl->Driver.Render.Start.
  * It's called when we start rendering a vertex buffer.
@@ -119,43 +118,52 @@ _swsetup_RenderStart( GLcontext *ctx )
    VB->AttribPtr[VERT_ATTRIB_POS] = VB->NdcPtr;
 
 
-   if (tnl->render_inputs != swsetup->last_index) {
-      GLuint index = tnl->render_inputs;
+   if (!RENDERINPUTS_EQUAL(tnl->render_inputs_bitset, swsetup->last_index_bitset)) {
+      DECLARE_RENDERINPUTS(index_bitset);
       struct tnl_attr_map map[_TNL_ATTRIB_MAX];
       int i, e = 0;
 
+      RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
+
       EMIT_ATTR( _TNL_ATTRIB_POS, EMIT_4F_VIEWPORT, win );
-   
-      if (index & _TNL_BIT_COLOR0)
-	 EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4CHAN_4F_RGBA, color );
-
-      if (index & _TNL_BIT_COLOR1)
-	 EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_4CHAN_4F_RGBA, specular);
-
-      if (index & _TNL_BIT_FOG) 
-	 EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1F, fog);
-	 
-      if (index & _TNL_BITS_TEX_ANY) {
-	 for (i = 0; i < MAX_TEXTURE_COORD_UNITS; i++) {
-	    if (index & _TNL_BIT_TEX(i)) {
-	       EMIT_ATTR( _TNL_ATTRIB_TEX0+i, EMIT_4F, texcoord[i] );
-	    }
-	 }
+
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR0 ))
+         EMIT_ATTR( _TNL_ATTRIB_COLOR0, EMIT_4CHAN_4F_RGBA, color );
+
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_COLOR1 ))
+         EMIT_ATTR( _TNL_ATTRIB_COLOR1, EMIT_4CHAN_4F_RGBA, specular);
+
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_FOG ))
+         EMIT_ATTR( _TNL_ATTRIB_FOG, EMIT_1F, fog);
+
+      if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_TEX, _TNL_LAST_TEX )) {
+         for (i = 0; i < MAX_TEXTURE_COORD_UNITS; i++) {
+            if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
+               EMIT_ATTR( _TNL_ATTRIB_TEX(i), EMIT_4F, texcoord[i] );
+            }
+         }
       }
-      
-      if (index & _TNL_BIT_INDEX) 
-	 EMIT_ATTR( _TNL_ATTRIB_INDEX, EMIT_1F, index );
- 
-      if (index & _TNL_BIT_POINTSIZE)
-	 EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F, pointSize );
-   
+
+      if (RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_FIRST_ATTRIBUTE, _TNL_LAST_ATTRIBUTE )) {
+          for (i = 0; i < MAX_VERTEX_ATTRIBS; i++) {
+              if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_ATTRIBUTE(i) )) {
+                  EMIT_ATTR( _TNL_ATTRIB_ATTRIBUTE(i), VARYING_EMIT_STYLE, attribute[i] );
+              }
+          }
+      }
+
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_INDEX ))
+         EMIT_ATTR( _TNL_ATTRIB_INDEX, EMIT_1F, index );
+
+      if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_POINTSIZE ))
+         EMIT_ATTR( _TNL_ATTRIB_POINTSIZE, EMIT_1F, pointSize );
+
       _tnl_install_attrs( ctx, map, e,
-			  ctx->Viewport._WindowMap.m,
-			  sizeof(SWvertex) ); 
-      
-      swsetup->last_index = index;
-   }
+                          ctx->Viewport._WindowMap.m,
+                          sizeof(SWvertex) );
 
+      RENDERINPUTS_COPY( swsetup->last_index_bitset, index_bitset );
+   }
 }
 
 /*
@@ -205,7 +213,7 @@ _swsetup_Wakeup( GLcontext *ctx )
    _swsetup_InvalidateState( ctx, ~0 );
 
    swsetup->verts = (SWvertex *)tnl->clipspace.vertex_buf;
-   swsetup->last_index = 0;
+   RENDERINPUTS_ZERO( swsetup->last_index_bitset );
 }
 
 
diff --git a/src/mesa/swrast_setup/ss_context.h b/src/mesa/swrast_setup/ss_context.h
index 2c6e4faf3d..e5d890447a 100644
--- a/src/mesa/swrast_setup/ss_context.h
+++ b/src/mesa/swrast_setup/ss_context.h
@@ -1,9 +1,9 @@
 
 /*
  * Mesa 3-D graphics library
- * Version:  4.1
+ * Version:  6.5
  *
- * Copyright (C) 1999-2002  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -32,11 +32,12 @@
 #include "mtypes.h"
 #include "swrast/swrast.h"
 #include "swrast_setup.h"
+#include "tnl/t_context.h"
 
 typedef struct {
    GLuint NewState;
    GLenum render_prim;
-   GLuint last_index;
+   DECLARE_RENDERINPUTS(last_index_bitset);
    SWvertex *verts;
 } SScontext;
 
diff --git a/src/mesa/tnl/t_context.c b/src/mesa/tnl/t_context.c
index d9777bfe6d..ad972fa7b7 100644
--- a/src/mesa/tnl/t_context.c
+++ b/src/mesa/tnl/t_context.c
@@ -2,7 +2,7 @@
  * Mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -170,32 +170,43 @@ _tnl_InvalidateState( GLcontext *ctx, GLuint new_state )
    /* Calculate tnl->render_inputs:
     */
    if (ctx->Visual.rgbMode) {
-      tnl->render_inputs = (_TNL_BIT_POS|
-			    _TNL_BIT_COLOR0|
-			    (ctx->Texture._EnabledCoordUnits << _TNL_ATTRIB_TEX0));
+      GLuint i;
+
+      RENDERINPUTS_ZERO( tnl->render_inputs_bitset );
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_POS );
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_COLOR0 );
+      for (i = 0; i < ctx->Const.MaxTextureCoordUnits; i++) {
+         if (ctx->Texture._EnabledCoordUnits & (1 << i)) {
+            RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_TEX(i) );
+         }
+      }
 
       if (NEED_SECONDARY_COLOR(ctx))
-	 tnl->render_inputs |= _TNL_BIT_COLOR1;
+         RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_COLOR1 );
    }
    else {
-      tnl->render_inputs |= (_TNL_BIT_POS|_TNL_BIT_INDEX);
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_POS );
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_INDEX );
    }
-    
+
    if (ctx->Fog.Enabled ||
        (ctx->FragmentProgram._Active &&
-        ctx->FragmentProgram._Current->FogOption != GL_NONE))
-      tnl->render_inputs |= _TNL_BIT_FOG;
+       ctx->FragmentProgram._Current->FogOption != GL_NONE))
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_FOG );
 
    if (ctx->Polygon.FrontMode != GL_FILL || 
        ctx->Polygon.BackMode != GL_FILL)
-      tnl->render_inputs |= _TNL_BIT_EDGEFLAG;
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_EDGEFLAG );
 
    if (ctx->RenderMode == GL_FEEDBACK)
-      tnl->render_inputs |= _TNL_BIT_TEX0;
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_TEX0 );
 
    if (ctx->Point._Attenuated ||
        (ctx->VertexProgram._Enabled && ctx->VertexProgram.PointSizeEnabled))
-      tnl->render_inputs |= _TNL_BIT_POINTSIZE;
+      RENDERINPUTS_SET( tnl->render_inputs_bitset, _TNL_ATTRIB_POINTSIZE );
+
+   if (ctx->ShaderObjects._VertexShaderPresent || ctx->ShaderObjects._FragmentShaderPresent)
+      RENDERINPUTS_SET_RANGE( tnl->render_inputs_bitset, _TNL_FIRST_ATTRIBUTE, _TNL_LAST_ATTRIBUTE );
 }
 
 
diff --git a/src/mesa/tnl/t_context.h b/src/mesa/tnl/t_context.h
index 753f59257a..7a0ca6f411 100644
--- a/src/mesa/tnl/t_context.h
+++ b/src/mesa/tnl/t_context.h
@@ -2,7 +2,7 @@
  * mesa 3-D graphics library
  * Version:  6.5
  *
- * Copyright (C) 1999-2005  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -92,6 +92,12 @@
  * number of bits allocated for these numbers in places like vertex
  * program instruction formats and register layouts.
  */
+/* The bit space exhaustion is a fact now, done by _TNL_ATTRIB_ATTRIBUTE* for
+ * GLSL vertex shader which cannot be aliased with conventional vertex attribs.
+ * Compacting _TNL_ATTRIB_MAT_* attribs would not work, they would not give
+ * as many free bits (11 plus already 1 free bit) as _TNL_ATTRIB_ATTRIBUTE*
+ * attribs want (16).
+ */
 enum {
 	_TNL_ATTRIB_POS = 0,
 	_TNL_ATTRIB_WEIGHT = 1,
@@ -109,98 +115,56 @@ enum {
 	_TNL_ATTRIB_TEX5 = 13,
 	_TNL_ATTRIB_TEX6 = 14,
 	_TNL_ATTRIB_TEX7 = 15,
-	_TNL_ATTRIB_MAT_FRONT_AMBIENT = 16,
-	_TNL_ATTRIB_MAT_BACK_AMBIENT = 17,
-	_TNL_ATTRIB_MAT_FRONT_DIFFUSE = 18,
-	_TNL_ATTRIB_MAT_BACK_DIFFUSE = 19,
-	_TNL_ATTRIB_MAT_FRONT_SPECULAR = 20,
-	_TNL_ATTRIB_MAT_BACK_SPECULAR = 21,
-	_TNL_ATTRIB_MAT_FRONT_EMISSION = 22,
-	_TNL_ATTRIB_MAT_BACK_EMISSION = 23,
-	_TNL_ATTRIB_MAT_FRONT_SHININESS = 24,
-	_TNL_ATTRIB_MAT_BACK_SHININESS = 25,
-	_TNL_ATTRIB_MAT_FRONT_INDEXES = 26,
-	_TNL_ATTRIB_MAT_BACK_INDEXES = 27, 
-	_TNL_ATTRIB_INDEX = 28,        
-	_TNL_ATTRIB_EDGEFLAG = 29,     
-	_TNL_ATTRIB_POINTSIZE = 30,
-	_TNL_ATTRIB_MAX = 31
+	_TNL_ATTRIB_ATTRIBUTE0 = 16,
+	_TNL_ATTRIB_ATTRIBUTE1 = 17,
+	_TNL_ATTRIB_ATTRIBUTE2 = 18,
+	_TNL_ATTRIB_ATTRIBUTE3 = 19,
+	_TNL_ATTRIB_ATTRIBUTE4 = 20,
+	_TNL_ATTRIB_ATTRIBUTE5 = 21,
+	_TNL_ATTRIB_ATTRIBUTE6 = 22,
+	_TNL_ATTRIB_ATTRIBUTE7 = 23,
+	_TNL_ATTRIB_ATTRIBUTE8 = 24,
+	_TNL_ATTRIB_ATTRIBUTE9 = 25,
+	_TNL_ATTRIB_ATTRIBUTE10 = 26,
+	_TNL_ATTRIB_ATTRIBUTE11 = 27,
+	_TNL_ATTRIB_ATTRIBUTE12 = 28,
+	_TNL_ATTRIB_ATTRIBUTE13 = 29,
+	_TNL_ATTRIB_ATTRIBUTE14 = 30,
+	_TNL_ATTRIB_ATTRIBUTE15 = 31,
+	_TNL_ATTRIB_MAT_FRONT_AMBIENT = 32,
+	_TNL_ATTRIB_MAT_BACK_AMBIENT = 33,
+	_TNL_ATTRIB_MAT_FRONT_DIFFUSE = 34,
+	_TNL_ATTRIB_MAT_BACK_DIFFUSE = 35,
+	_TNL_ATTRIB_MAT_FRONT_SPECULAR = 36,
+	_TNL_ATTRIB_MAT_BACK_SPECULAR = 37,
+	_TNL_ATTRIB_MAT_FRONT_EMISSION = 38,
+	_TNL_ATTRIB_MAT_BACK_EMISSION = 39,
+	_TNL_ATTRIB_MAT_FRONT_SHININESS = 40,
+	_TNL_ATTRIB_MAT_BACK_SHININESS = 41,
+	_TNL_ATTRIB_MAT_FRONT_INDEXES = 42,
+	_TNL_ATTRIB_MAT_BACK_INDEXES = 43,
+	_TNL_ATTRIB_INDEX = 44,
+	_TNL_ATTRIB_EDGEFLAG = 45,
+	_TNL_ATTRIB_POINTSIZE = 46,
+	_TNL_ATTRIB_MAX = 47
 } ;
 
-/* Will probably have to revise this scheme fairly shortly, eg. by
- * compacting all the MAT flags down to one bit, or by using two
- * dwords to store the flags.
+#define _TNL_ATTRIB_TEX(u)       (_TNL_ATTRIB_TEX0 + (u))
+#define _TNL_ATTRIB_ATTRIBUTE(n) (_TNL_ATTRIB_ATTRIBUTE0 + (n))
+
+/* Define bit ranges instead of bit masks.
  */
-#define _TNL_BIT_POS                 (1<<0)
-#define _TNL_BIT_WEIGHT              (1<<1)
-#define _TNL_BIT_NORMAL              (1<<2)
-#define _TNL_BIT_COLOR0              (1<<3)
-#define _TNL_BIT_COLOR1              (1<<4)
-#define _TNL_BIT_FOG                 (1<<5)
-#define _TNL_BIT_SIX                 (1<<6)
-#define _TNL_BIT_SEVEN               (1<<7)
-#define _TNL_BIT_TEX0                (1<<8)
-#define _TNL_BIT_TEX1                (1<<9)
-#define _TNL_BIT_TEX2                (1<<10)
-#define _TNL_BIT_TEX3                (1<<11)
-#define _TNL_BIT_TEX4                (1<<12)
-#define _TNL_BIT_TEX5                (1<<13)
-#define _TNL_BIT_TEX6                (1<<14)
-#define _TNL_BIT_TEX7                (1<<15)
-#define _TNL_BIT_MAT_FRONT_AMBIENT   (1<<16)
-#define _TNL_BIT_MAT_BACK_AMBIENT    (1<<17)
-#define _TNL_BIT_MAT_FRONT_DIFFUSE   (1<<18)
-#define _TNL_BIT_MAT_BACK_DIFFUSE    (1<<19)
-#define _TNL_BIT_MAT_FRONT_SPECULAR  (1<<20)
-#define _TNL_BIT_MAT_BACK_SPECULAR   (1<<21)
-#define _TNL_BIT_MAT_FRONT_EMISSION  (1<<22)
-#define _TNL_BIT_MAT_BACK_EMISSION   (1<<23)
-#define _TNL_BIT_MAT_FRONT_SHININESS (1<<24)
-#define _TNL_BIT_MAT_BACK_SHININESS  (1<<25)
-#define _TNL_BIT_MAT_FRONT_INDEXES   (1<<26)
-#define _TNL_BIT_MAT_BACK_INDEXES    (1<<27)
-#define _TNL_BIT_INDEX               (1<<28)
-#define _TNL_BIT_EDGEFLAG            (1<<29)
-#define _TNL_BIT_POINTSIZE           (1<<30)
-
-#define _TNL_BIT_TEX(u)  (1 << (_TNL_ATTRIB_TEX0 + (u)))
-
-
-
-#define _TNL_BITS_MAT_ANY  (_TNL_BIT_MAT_FRONT_AMBIENT   | 	\
-			    _TNL_BIT_MAT_BACK_AMBIENT    | 	\
-			    _TNL_BIT_MAT_FRONT_DIFFUSE   | 	\
-			    _TNL_BIT_MAT_BACK_DIFFUSE    | 	\
-			    _TNL_BIT_MAT_FRONT_SPECULAR  | 	\
-			    _TNL_BIT_MAT_BACK_SPECULAR   | 	\
-			    _TNL_BIT_MAT_FRONT_EMISSION  | 	\
-			    _TNL_BIT_MAT_BACK_EMISSION   | 	\
-			    _TNL_BIT_MAT_FRONT_SHININESS | 	\
-			    _TNL_BIT_MAT_BACK_SHININESS  | 	\
-			    _TNL_BIT_MAT_FRONT_INDEXES   | 	\
-			    _TNL_BIT_MAT_BACK_INDEXES)
-
-
-#define _TNL_BITS_TEX_ANY  (_TNL_BIT_TEX0 |	\
-                            _TNL_BIT_TEX1 |	\
-                            _TNL_BIT_TEX2 |	\
-                            _TNL_BIT_TEX3 |	\
-                            _TNL_BIT_TEX4 |	\
-                            _TNL_BIT_TEX5 |	\
-                            _TNL_BIT_TEX6 |	\
-                            _TNL_BIT_TEX7)
-
-
-#define _TNL_BITS_PROG_ANY   (_TNL_BIT_POS    |		\
-			      _TNL_BIT_WEIGHT |		\
-			      _TNL_BIT_NORMAL |		\
-			      _TNL_BIT_COLOR0 |		\
-			      _TNL_BIT_COLOR1 |		\
-			      _TNL_BIT_FOG    |		\
-			      _TNL_BIT_SIX    |		\
-			      _TNL_BIT_SEVEN  |		\
-			      _TNL_BITS_TEX_ANY)
+#define _TNL_FIRST_PROG      _TNL_ATTRIB_WEIGHT
+#define _TNL_LAST_PROG       _TNL_ATTRIB_TEX7
+
+#define _TNL_FIRST_TEX       _TNL_ATTRIB_TEX0
+#define _TNL_LAST_TEX        _TNL_ATTRIB_TEX7
+
+#define _TNL_FIRST_ATTRIBUTE _TNL_ATTRIB_ATTRIBUTE0
+#define _TNL_LAST_ATTRIBUTE  _TNL_ATTRIB_ATTRIBUTE15
 
+#define _TNL_FIRST_MAT       _TNL_ATTRIB_MAT_FRONT_AMBIENT
+#define _TNL_LAST_MAT        _TNL_ATTRIB_MAT_BACK_INDEXES
 
 
 #define PRIM_BEGIN     0x10
@@ -444,6 +408,7 @@ struct vertex_buffer
    GLvector4f  *SecondaryColorPtr[2];           /* _TNL_BIT_COLOR1 */
    GLvector4f  *PointSizePtr;	                /* _TNL_BIT_POS */
    GLvector4f  *FogCoordPtr;	                /* _TNL_BIT_FOG */
+   GLvector4f  *VaryingPtr[MAX_VARYING_VECTORS];
 
    struct tnl_prim  *Primitive;	              
    GLuint      PrimitiveCount;	      
@@ -730,7 +695,20 @@ struct tnl_device_driver
        */
    } Render;
 };
-   
+
+
+#define DECLARE_RENDERINPUTS(name) BITSET64_DECLARE(name, _TNL_ATTRIB_MAX)
+#define RENDERINPUTS_COPY BITSET64_COPY
+#define RENDERINPUTS_EQUAL BITSET64_EQUAL
+#define RENDERINPUTS_ZERO BITSET64_ZERO
+#define RENDERINPUTS_ONES BITSET64_ONES
+#define RENDERINPUTS_TEST BITSET64_TEST
+#define RENDERINPUTS_SET BITSET64_SET
+#define RENDERINPUTS_CLEAR BITSET64_CLEAR
+#define RENDERINPUTS_TEST_RANGE BITSET64_TEST_RANGE
+#define RENDERINPUTS_SET_RANGE BITSET64_SET_RANGE
+#define RENDERINPUTS_CLEAR_RANGE BITSET64_CLEAR_RANGE
+
 
 /**
  * Context state for T&L context.
@@ -783,7 +761,7 @@ typedef struct
     */
    GLboolean DiscardPrimitive;
 
-   GLuint render_inputs;
+   DECLARE_RENDERINPUTS(render_inputs_bitset);
 
    GLvertexformat exec_vtxfmt;
    GLvertexformat save_vtxfmt;
diff --git a/src/mesa/tnl/t_save_api.c b/src/mesa/tnl/t_save_api.c
index 9788468d50..5765652b7e 100644
--- a/src/mesa/tnl/t_save_api.c
+++ b/src/mesa/tnl/t_save_api.c
@@ -936,7 +936,7 @@ static void GLAPIENTRY _save_MultiTexCoord4fv( GLenum target, const GLfloat *v )
 
 static void GLAPIENTRY _save_VertexAttrib1fNV( GLuint index, GLfloat x )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR1F( index, x );
    else
       enum_error(); 
@@ -944,7 +944,7 @@ static void GLAPIENTRY _save_VertexAttrib1fNV( GLuint index, GLfloat x )
 
 static void GLAPIENTRY _save_VertexAttrib1fvNV( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR1FV( index, v );
    else
       enum_error();
@@ -952,7 +952,7 @@ static void GLAPIENTRY _save_VertexAttrib1fvNV( GLuint index, const GLfloat *v )
 
 static void GLAPIENTRY _save_VertexAttrib2fNV( GLuint index, GLfloat x, GLfloat y )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR2F( index, x, y );
    else
       enum_error();
@@ -960,7 +960,7 @@ static void GLAPIENTRY _save_VertexAttrib2fNV( GLuint index, GLfloat x, GLfloat
 
 static void GLAPIENTRY _save_VertexAttrib2fvNV( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR2FV( index, v );
    else
       enum_error();
@@ -969,7 +969,7 @@ static void GLAPIENTRY _save_VertexAttrib2fvNV( GLuint index, const GLfloat *v )
 static void GLAPIENTRY _save_VertexAttrib3fNV( GLuint index, GLfloat x, GLfloat y, 
 				  GLfloat z )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR3F( index, x, y, z );
    else
       enum_error();
@@ -977,7 +977,7 @@ static void GLAPIENTRY _save_VertexAttrib3fNV( GLuint index, GLfloat x, GLfloat
 
 static void GLAPIENTRY _save_VertexAttrib3fvNV( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR3FV( index, v );
    else
       enum_error();
@@ -986,7 +986,7 @@ static void GLAPIENTRY _save_VertexAttrib3fvNV( GLuint index, const GLfloat *v )
 static void GLAPIENTRY _save_VertexAttrib4fNV( GLuint index, GLfloat x, GLfloat y,
 				  GLfloat z, GLfloat w )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR4F( index, x, y, z, w );
    else
       enum_error();
@@ -994,7 +994,7 @@ static void GLAPIENTRY _save_VertexAttrib4fNV( GLuint index, GLfloat x, GLfloat
 
 static void GLAPIENTRY _save_VertexAttrib4fvNV( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_PROGRAM_ATTRIBS)
       DISPATCH_ATTR4FV( index, v );
    else
       enum_error();
@@ -1004,7 +1004,7 @@ static void GLAPIENTRY _save_VertexAttrib4fvNV( GLuint index, const GLfloat *v )
 static void GLAPIENTRY
 _save_VertexAttrib1fARB( GLuint index, GLfloat x )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR1F( index, x );
    else
       enum_error(); 
@@ -1013,7 +1013,7 @@ _save_VertexAttrib1fARB( GLuint index, GLfloat x )
 static void GLAPIENTRY
 _save_VertexAttrib1fvARB( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR1FV( index, v );
    else
       enum_error();
@@ -1022,7 +1022,7 @@ _save_VertexAttrib1fvARB( GLuint index, const GLfloat *v )
 static void GLAPIENTRY
 _save_VertexAttrib2fARB( GLuint index, GLfloat x, GLfloat y )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR2F( index, x, y );
    else
       enum_error();
@@ -1031,7 +1031,7 @@ _save_VertexAttrib2fARB( GLuint index, GLfloat x, GLfloat y )
 static void GLAPIENTRY
 _save_VertexAttrib2fvARB( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR2FV( index, v );
    else
       enum_error();
@@ -1040,7 +1040,7 @@ _save_VertexAttrib2fvARB( GLuint index, const GLfloat *v )
 static void GLAPIENTRY
 _save_VertexAttrib3fARB( GLuint index, GLfloat x, GLfloat y, GLfloat z )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR3F( index, x, y, z );
    else
       enum_error();
@@ -1049,7 +1049,7 @@ _save_VertexAttrib3fARB( GLuint index, GLfloat x, GLfloat y, GLfloat z )
 static void GLAPIENTRY
 _save_VertexAttrib3fvARB( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR3FV( index, v );
    else
       enum_error();
@@ -1058,7 +1058,7 @@ _save_VertexAttrib3fvARB( GLuint index, const GLfloat *v )
 static void GLAPIENTRY
 _save_VertexAttrib4fARB( GLuint index, GLfloat x, GLfloat y, GLfloat z, GLfloat w )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR4F( index, x, y, z, w );
    else
       enum_error();
@@ -1067,7 +1067,7 @@ _save_VertexAttrib4fARB( GLuint index, GLfloat x, GLfloat y, GLfloat z, GLfloat
 static void GLAPIENTRY
 _save_VertexAttrib4fvARB( GLuint index, const GLfloat *v )
 {
-   if (index < VERT_ATTRIB_MAX)
+   if (index < MAX_VERTEX_ATTRIBS)
       DISPATCH_ATTR4FV( index, v );
    else
       enum_error();
diff --git a/src/mesa/tnl/t_vb_arbshader.c b/src/mesa/tnl/t_vb_arbshader.c
index 88d7f6ac8c..b1d024e0cc 100644
--- a/src/mesa/tnl/t_vb_arbshader.c
+++ b/src/mesa/tnl/t_vb_arbshader.c
@@ -37,6 +37,7 @@
 typedef struct

 {

 	GLvector4f outputs[VERT_RESULT_MAX];

+	GLvector4f varyings[MAX_VARYING_VECTORS];

 	GLvector4f ndc_coords;

 	GLubyte *clipmask;

 	GLubyte ormask;

@@ -63,6 +64,11 @@ static GLboolean construct_arb_vertex_shader (GLcontext *ctx, struct tnl_pipelin
 		_mesa_vector4f_alloc (&store->outputs[i], 0, size, 32);

 		store->outputs[i].size = 4;

 	}

+	for (i = 0; i < MAX_VARYING_VECTORS; i++)

+	{

+		_mesa_vector4f_alloc (&store->varyings[i], 0, size, 32);

+		store->varyings[i].size = 4;

+	}

 	_mesa_vector4f_alloc (&store->ndc_coords, 0, size, 32);

 	store->clipmask = (GLubyte *) ALIGN_MALLOC (size, 32);

 

@@ -79,6 +85,8 @@ static void destruct_arb_vertex_shader (struct tnl_pipeline_stage *stage)
 

 		for (i = 0; i < VERT_RESULT_MAX; i++)

 			_mesa_vector4f_free (&store->outputs[i]);

+		for (i = 0; i < MAX_VARYING_VECTORS; i++)

+			_mesa_vector4f_free (&store->varyings[i]);

 		_mesa_vector4f_free (&store->ndc_coords);

 		ALIGN_FREE (store->clipmask);

 

@@ -206,6 +214,17 @@ static GLboolean run_arb_vertex_shader (GLcontext *ctx, struct tnl_pipeline_stag
 		fetch_output_vec4 (pro, SLANG_VERTEX_FIXED_BACKCOLOR, VERT_RESULT_BFC0, i, 0, store);

 		fetch_output_vec4 (pro, SLANG_VERTEX_FIXED_BACKSECONDARYCOLOR, VERT_RESULT_BFC1, i, 0, store);

 		/* XXX: fetch output SLANG_VERTEX_FIXED_CLIPVERTEX */

+

+		for (j = 0; j < MAX_VARYING_VECTORS; j++)

+		{

+			GLuint k;

+

+			for (k = 0; k < VARYINGS_PER_VECTOR; k++)

+			{

+				(**pro).UpdateVarying (pro, j * VARYINGS_PER_VECTOR + k,

+					&store->varyings[j].data[i][k], GL_TRUE);

+			}

+		}

 	}

 

 	vb->ClipPtr = &store->outputs[VERT_RESULT_HPOS];

@@ -218,6 +237,8 @@ static GLboolean run_arb_vertex_shader (GLcontext *ctx, struct tnl_pipeline_stag
 	vb->SecondaryColorPtr[1] = &store->outputs[VERT_RESULT_BFC1];

 	vb->FogCoordPtr = &store->outputs[VERT_RESULT_FOGC];

 	vb->PointSizePtr = &store->outputs[VERT_RESULT_PSIZ];

+	for (i = 0; i < MAX_VARYING_VECTORS; i++)

+		vb->VaryingPtr[i] = &store->varyings[i];

 

 	vb->AttribPtr[VERT_ATTRIB_COLOR0] = vb->ColorPtr[0];

 	vb->AttribPtr[VERT_ATTRIB_COLOR1] = vb->SecondaryColorPtr[0];

@@ -225,6 +246,8 @@ static GLboolean run_arb_vertex_shader (GLcontext *ctx, struct tnl_pipeline_stag
 	for (i = 0; i < ctx->Const.MaxTextureUnits; i++)

 		vb->AttribPtr[VERT_ATTRIB_TEX0 + i] = vb->TexCoordPtr[i];

 	vb->AttribPtr[_TNL_ATTRIB_POINTSIZE] = &store->outputs[VERT_RESULT_PSIZ];

+	for (i = 0; i < MAX_VARYING_VECTORS; i++)

+		vb->AttribPtr[_TNL_ATTRIB_ATTRIBUTE0 + i] = vb->VaryingPtr[i];

 

 	store->ormask = 0;

 	store->andmask = CLIP_FRUSTUM_BITS;

diff --git a/src/mesa/tnl/t_vtx_generic.c b/src/mesa/tnl/t_vtx_generic.c
index 0422fcd45d..d59ce90c6d 100644
--- a/src/mesa/tnl/t_vtx_generic.c
+++ b/src/mesa/tnl/t_vtx_generic.c
@@ -364,42 +364,42 @@ static void GLAPIENTRY _tnl_MultiTexCoord4fv( GLenum target,
 
 static void GLAPIENTRY _tnl_VertexAttrib1fNV( GLuint index, GLfloat x )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR1F( index, x );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib1fvNV( GLuint index, 
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR1FV( index, v );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib2fNV( GLuint index, GLfloat x, 
 					      GLfloat y )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR2F( index, x, y );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib2fvNV( GLuint index,
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR2FV( index, v );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib3fNV( GLuint index, GLfloat x,
 					      GLfloat y, GLfloat z )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR3F( index, x, y, z );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib3fvNV( GLuint index,
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR3FV( index, v );
 }
 
@@ -407,60 +407,73 @@ static void GLAPIENTRY _tnl_VertexAttrib4fNV( GLuint index, GLfloat x,
 					      GLfloat y, GLfloat z,
 					      GLfloat w )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR4F( index, x, y, z, w );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib4fvNV( GLuint index, 
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_PROGRAM_ATTRIBS) index = ERROR_ATTRIB;
    DISPATCH_ATTR4FV( index, v );
 }
 
-
-/*
- * XXX adjust index
- */
-
 static void GLAPIENTRY _tnl_VertexAttrib1fARB( GLuint index, GLfloat x )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR1F( index, x );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib1fvARB( GLuint index, 
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR1FV( index, v );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib2fARB( GLuint index, GLfloat x, 
 					      GLfloat y )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR2F( index, x, y );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib2fvARB( GLuint index,
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR2FV( index, v );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib3fARB( GLuint index, GLfloat x,
 					      GLfloat y, GLfloat z )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR3F( index, x, y, z );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib3fvARB( GLuint index,
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR3FV( index, v );
 }
 
@@ -468,14 +481,20 @@ static void GLAPIENTRY _tnl_VertexAttrib4fARB( GLuint index, GLfloat x,
 					      GLfloat y, GLfloat z,
 					      GLfloat w )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR4F( index, x, y, z, w );
 }
 
 static void GLAPIENTRY _tnl_VertexAttrib4fvARB( GLuint index, 
 					       const GLfloat *v )
 {
-   if (index >= VERT_ATTRIB_MAX) index = ERROR_ATTRIB;
+   if (index >= MAX_VERTEX_ATTRIBS)
+      index = ERROR_ATTRIB;
+   else
+      index += VERT_ATTRIB_GENERIC0;
    DISPATCH_ATTR4FV( index, v );
 }