Merge commit 'origin/master' into gallium-map-range

47 files changed, 437 insertions, 196 deletions

diff --git a/src/mesa/drivers/common/driverfuncs.c b/src/mesa/drivers/common/driverfuncs.c
index 986f751bdc1..9bee212c6a3 100644
--- a/src/mesa/drivers/common/driverfuncs.c
+++ b/src/mesa/drivers/common/driverfuncs.c
@@ -308,7 +308,7 @@ _mesa_init_driver_state(GLcontext *ctx)
    ctx->Driver.Enable(ctx, GL_LINE_SMOOTH, ctx->Line.SmoothFlag);
    ctx->Driver.Enable(ctx, GL_POLYGON_STIPPLE, ctx->Polygon.StippleFlag);
    ctx->Driver.Enable(ctx, GL_SCISSOR_TEST, ctx->Scissor.Enabled);
-   ctx->Driver.Enable(ctx, GL_STENCIL_TEST, ctx->Stencil.Enabled);
+   ctx->Driver.Enable(ctx, GL_STENCIL_TEST, ctx->Stencil._Enabled);
    ctx->Driver.Enable(ctx, GL_TEXTURE_1D, GL_FALSE);
    ctx->Driver.Enable(ctx, GL_TEXTURE_2D, GL_FALSE);
    ctx->Driver.Enable(ctx, GL_TEXTURE_RECTANGLE_NV, GL_FALSE);
diff --git a/src/mesa/drivers/dri/ffb/ffb_state.c b/src/mesa/drivers/dri/ffb/ffb_state.c
index ee0fe4e0dbe..5eb8f417ffb 100644
--- a/src/mesa/drivers/dri/ffb/ffb_state.c
+++ b/src/mesa/drivers/dri/ffb/ffb_state.c
@@ -275,7 +275,7 @@ ffbDDStencilFuncSeparate(GLcontext *ctx, GLenum face, GLenum func,
 	/* We will properly update sw/hw state when stenciling is
 	 * enabled.
 	 */
-	if (! ctx->Stencil.Enabled)
+	if (! ctx->Stencil._Enabled)
 		return;
 
 	stencilctl = fmesa->stencilctl;
@@ -333,7 +333,7 @@ ffbDDStencilOpSeparate(GLcontext *ctx, GLenum face, GLenum fail,
 	/* We will properly update sw/hw state when stenciling is
 	 * enabled.
 	 */
-	if (! ctx->Stencil.Enabled)
+	if (! ctx->Stencil._Enabled)
 		return;
 
 	stencilctl = fmesa->stencilctl;
diff --git a/src/mesa/drivers/dri/i965/brw_cc.c b/src/mesa/drivers/dri/i965/brw_cc.c
index 82370162f59..c724218cf56 100644
--- a/src/mesa/drivers/dri/i965/brw_cc.c
+++ b/src/mesa/drivers/dri/i965/brw_cc.c
@@ -88,7 +88,7 @@ cc_unit_populate_key(struct brw_context *brw, struct brw_cc_unit_key *key)
 
    memset(key, 0, sizeof(*key));
 
-   key->stencil = ctx->Stencil.Enabled;
+   key->stencil = ctx->Stencil._Enabled;
    key->stencil_two_side = ctx->Stencil._TestTwoSide;
 
    if (key->stencil) {
diff --git a/src/mesa/drivers/dri/i965/brw_fallback.c b/src/mesa/drivers/dri/i965/brw_fallback.c
index 5f4f2d515df..299357409ce 100644
--- a/src/mesa/drivers/dri/i965/brw_fallback.c
+++ b/src/mesa/drivers/dri/i965/brw_fallback.c
@@ -75,7 +75,7 @@ static GLboolean do_check_fallback(struct brw_context *brw)
    
    /* _NEW_STENCIL 
     */
-   if (ctx->Stencil.Enabled &&
+   if (ctx->Stencil._Enabled &&
        (ctx->DrawBuffer->Name == 0 && !brw->intel.hw_stencil)) {
       DBG("FALLBACK: stencil\n");
       return GL_TRUE;
diff --git a/src/mesa/drivers/dri/i965/brw_wm.c b/src/mesa/drivers/dri/i965/brw_wm.c
index 06a6f3f0f4e..879000644b2 100644
--- a/src/mesa/drivers/dri/i965/brw_wm.c
+++ b/src/mesa/drivers/dri/i965/brw_wm.c
@@ -210,7 +210,7 @@ static void brw_wm_populate_key( struct brw_context *brw,
       lookup |= IZ_DEPTH_WRITE_ENABLE_BIT;
 
    /* _NEW_STENCIL */
-   if (ctx->Stencil.Enabled) {
+   if (ctx->Stencil._Enabled) {
       lookup |= IZ_STENCIL_TEST_ENABLE_BIT;
 
       if (ctx->Stencil.WriteMask[0] ||
diff --git a/src/mesa/drivers/dri/intel/intel_buffers.c b/src/mesa/drivers/dri/intel/intel_buffers.c
index 7f2144abd4b..732bae5b5ae 100644
--- a/src/mesa/drivers/dri/intel/intel_buffers.c
+++ b/src/mesa/drivers/dri/intel/intel_buffers.c
@@ -263,7 +263,7 @@ intel_draw_buffer(GLcontext * ctx, struct gl_framebuffer *fb)
       }
    }
    else {
-      /* XXX FBO: instead of FALSE, pass ctx->Stencil.Enabled ??? */
+      /* XXX FBO: instead of FALSE, pass ctx->Stencil._Enabled ??? */
       FALLBACK(intel, INTEL_FALLBACK_STENCIL_BUFFER, GL_FALSE);
    }
 
@@ -274,7 +274,7 @@ intel_draw_buffer(GLcontext * ctx, struct gl_framebuffer *fb)
       ctx->Driver.Enable(ctx, GL_DEPTH_TEST,
                          (ctx->Depth.Test && fb->Visual.depthBits > 0));
       ctx->Driver.Enable(ctx, GL_STENCIL_TEST,
-                         (ctx->Stencil.Enabled && fb->Visual.stencilBits > 0));
+                         (ctx->Stencil._Enabled && fb->Visual.stencilBits > 0));
    }
    else {
       ctx->NewState |= (_NEW_DEPTH | _NEW_STENCIL);
diff --git a/src/mesa/drivers/dri/intel/intel_clear.c b/src/mesa/drivers/dri/intel/intel_clear.c
index c3ba50f3550..28281b38615 100644
--- a/src/mesa/drivers/dri/intel/intel_clear.c
+++ b/src/mesa/drivers/dri/intel/intel_clear.c
@@ -247,13 +247,10 @@ static const char *buffer_names[] = {
    [BUFFER_BACK_LEFT] = "back",
    [BUFFER_FRONT_RIGHT] = "front right",
    [BUFFER_BACK_RIGHT] = "back right",
-   [BUFFER_AUX0] = "aux0",
-   [BUFFER_AUX1] = "aux1",
-   [BUFFER_AUX2] = "aux2",
-   [BUFFER_AUX3] = "aux3",
    [BUFFER_DEPTH] = "depth",
    [BUFFER_STENCIL] = "stencil",
    [BUFFER_ACCUM] = "accum",
+   [BUFFER_AUX0] = "aux0",
    [BUFFER_COLOR0] = "color0",
    [BUFFER_COLOR1] = "color1",
    [BUFFER_COLOR2] = "color2",
diff --git a/src/mesa/drivers/dri/intel/intel_pixel.c b/src/mesa/drivers/dri/intel/intel_pixel.c
index 5e32288844c..9f4b4ff0ba0 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel.c
@@ -112,7 +112,7 @@ intel_check_blit_fragment_ops(GLcontext * ctx, GLboolean src_alpha_is_one)
       return GL_FALSE;
    }
 
-   if (ctx->Stencil.Enabled) {
+   if (ctx->Stencil._Enabled) {
       DBG("fallback due to image stencil\n");
       return GL_FALSE;
    }
diff --git a/src/mesa/drivers/dri/intel/intel_pixel_copy.c b/src/mesa/drivers/dri/intel/intel_pixel_copy.c
index 7c7aa6097c8..d50dd68092d 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel_copy.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel_copy.c
@@ -87,7 +87,7 @@ intel_check_copypixel_blit_fragment_ops(GLcontext * ctx)
             ctx->Color.AlphaEnabled ||
             ctx->Depth.Test ||
             ctx->Fog.Enabled ||
-            ctx->Stencil.Enabled ||
+            ctx->Stencil._Enabled ||
             !ctx->Color.ColorMask[0] ||
             !ctx->Color.ColorMask[1] ||
             !ctx->Color.ColorMask[2] ||
diff --git a/src/mesa/drivers/dri/intel/intel_pixel_draw.c b/src/mesa/drivers/dri/intel/intel_pixel_draw.c
index 7be7ea82b35..e8d5ac8569d 100644
--- a/src/mesa/drivers/dri/intel/intel_pixel_draw.c
+++ b/src/mesa/drivers/dri/intel/intel_pixel_draw.c
@@ -233,7 +233,7 @@ intel_stencil_drawpixels(GLcontext * ctx,
    }
 
    /* We don't support stencil testing/ops here */
-   if (ctx->Stencil.Enabled)
+   if (ctx->Stencil._Enabled)
       return GL_FALSE;
 
    /* We use FBOs for our wrapping of the depthbuffer into a color
diff --git a/src/mesa/drivers/dri/mga/mgapixel.c b/src/mesa/drivers/dri/mga/mgapixel.c
index 9f90047ba59..977dfa0b760 100644
--- a/src/mesa/drivers/dri/mga/mgapixel.c
+++ b/src/mesa/drivers/dri/mga/mgapixel.c
@@ -133,7 +133,7 @@ check_color_per_fragment_ops( const GLcontext *ctx )
 		    ctx->Depth.Test ||
 		    ctx->Fog.Enabled ||
 		    ctx->Scissor.Enabled ||
-		    ctx->Stencil.Enabled ||
+		    ctx->Stencil._Enabled ||
 		    !ctx->Color.ColorMask[0] ||
 		    !ctx->Color.ColorMask[1] ||
 		    !ctx->Color.ColorMask[2] ||
diff --git a/src/mesa/drivers/dri/r200/r200_pixel.c b/src/mesa/drivers/dri/r200/r200_pixel.c
index be68821dc1d..2797cbb3dc0 100644
--- a/src/mesa/drivers/dri/r200/r200_pixel.c
+++ b/src/mesa/drivers/dri/r200/r200_pixel.c
@@ -87,7 +87,7 @@ check_color_per_fragment_ops( const GLcontext *ctx )
 		    ctx->Depth.Test ||
 		    ctx->Fog.Enabled ||
 		    ctx->Scissor.Enabled ||
-		    ctx->Stencil.Enabled ||
+		    ctx->Stencil._Enabled ||
 		    !ctx->Color.ColorMask[0] ||
 		    !ctx->Color.ColorMask[1] ||
 		    !ctx->Color.ColorMask[2] ||
diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c
index ade45f581ca..37dae6c8869 100644
--- a/src/mesa/drivers/dri/r300/r300_state.c
+++ b/src/mesa/drivers/dri/r300/r300_state.c
@@ -2205,7 +2205,7 @@ static void r300ResetHwState(r300ContextPtr r300)
 	r300DepthFunc(ctx, ctx->Depth.Func);
 
 	/* stencil */
-	r300Enable(ctx, GL_STENCIL_TEST, ctx->Stencil.Enabled);
+	r300Enable(ctx, GL_STENCIL_TEST, ctx->Stencil._Enabled);
 	r300StencilMaskSeparate(ctx, 0, ctx->Stencil.WriteMask[0]);
 	r300StencilFuncSeparate(ctx, 0, ctx->Stencil.Function[0],
 				ctx->Stencil.Ref[0], ctx->Stencil.ValueMask[0]);
diff --git a/src/mesa/drivers/dri/savage/savagestate.c b/src/mesa/drivers/dri/savage/savagestate.c
index 73d85ed57b4..84e1b525854 100644
--- a/src/mesa/drivers/dri/savage/savagestate.c
+++ b/src/mesa/drivers/dri/savage/savagestate.c
@@ -514,7 +514,7 @@ static void savageDDDepthFunc_s4(GLcontext *ctx, GLenum func)
 	imesa->regs.s4.drawLocalCtrl.ni.flushPdZbufWrites = GL_TRUE;
 	imesa->regs.s4.zBufCtrl.ni.zBufEn = GL_TRUE;
     }
-    else if (imesa->glCtx->Stencil.Enabled && imesa->hw_stencil)
+    else if (imesa->glCtx->Stencil._Enabled && imesa->hw_stencil)
     {
         /* Need to keep Z on for Stencil. */
 	imesa->regs.s4.zBufCtrl.ni.zCmpFunc = CF_Always;
@@ -1092,7 +1092,7 @@ static void savageDDEnable_s4(GLcontext *ctx, GLenum cap, GLboolean state)
 		FALLBACK (ctx, SAVAGE_FALLBACK_STENCIL, state);
 	    else {
 		imesa->regs.s4.stencilCtrl.ni.stencilEn = state;
-		if (ctx->Stencil.Enabled &&
+		if (ctx->Stencil._Enabled &&
 		    imesa->regs.s4.zBufCtrl.ni.zBufEn != GL_TRUE)
 		{
 		    /* Stencil buffer requires Z enabled. */
diff --git a/src/mesa/drivers/dri/tdfx/tdfx_pixels.c b/src/mesa/drivers/dri/tdfx/tdfx_pixels.c
index 9ab9c05f2bd..18729d5ae08 100644
--- a/src/mesa/drivers/dri/tdfx/tdfx_pixels.c
+++ b/src/mesa/drivers/dri/tdfx/tdfx_pixels.c
@@ -610,7 +610,7 @@ tdfx_drawpixels_R8G8B8A8(GLcontext * ctx, GLint x, GLint y,
        ctx->Depth.Test ||
        ctx->Fog.Enabled ||
        ctx->Scissor.Enabled ||
-       ctx->Stencil.Enabled ||
+       ctx->Stencil._Enabled ||
        !ctx->Color.ColorMask[0] ||
        !ctx->Color.ColorMask[1] ||
        !ctx->Color.ColorMask[2] ||
diff --git a/src/mesa/drivers/dri/tdfx/tdfx_render.c b/src/mesa/drivers/dri/tdfx/tdfx_render.c
index cf840c57a7a..2cd8e12d955 100644
--- a/src/mesa/drivers/dri/tdfx/tdfx_render.c
+++ b/src/mesa/drivers/dri/tdfx/tdfx_render.c
@@ -740,7 +740,7 @@ void tdfxEmitHwStateLocked( tdfxContextPtr fxMesa )
    }
 
    if ( fxMesa->dirty & TDFX_UPLOAD_STENCIL ) {
-      if (fxMesa->glCtx->Stencil.Enabled) {
+      if (fxMesa->glCtx->Stencil._Enabled) {
          fxMesa->Glide.grEnable(GR_STENCIL_MODE_EXT);
          fxMesa->Glide.grStencilOp(fxMesa->Stencil.FailFunc,
                                    fxMesa->Stencil.ZFailFunc,
diff --git a/src/mesa/drivers/dri/tdfx/tdfx_state.c b/src/mesa/drivers/dri/tdfx/tdfx_state.c
index a2d7bcd97d9..591df8a905a 100644
--- a/src/mesa/drivers/dri/tdfx/tdfx_state.c
+++ b/src/mesa/drivers/dri/tdfx/tdfx_state.c
@@ -459,7 +459,7 @@ static void tdfxUpdateStencil( GLcontext *ctx )
    }
 
    if (fxMesa->haveHwStencil) {
-      if (ctx->Stencil.Enabled) {
+      if (ctx->Stencil._Enabled) {
          fxMesa->Stencil.Function = ctx->Stencil.Function[0] - GL_NEVER + GR_CMP_NEVER;
          fxMesa->Stencil.RefValue = ctx->Stencil.Ref[0] & 0xff;
          fxMesa->Stencil.ValueMask = ctx->Stencil.ValueMask[0] & 0xff;
diff --git a/src/mesa/drivers/dri/unichrome/via_state.c b/src/mesa/drivers/dri/unichrome/via_state.c
index 1cef01ab033..840e4e42da9 100644
--- a/src/mesa/drivers/dri/unichrome/via_state.c
+++ b/src/mesa/drivers/dri/unichrome/via_state.c
@@ -1342,7 +1342,7 @@ static void viaChooseStencilState(GLcontext *ctx)
 {
     struct via_context *vmesa = VIA_CONTEXT(ctx);
     
-    if (ctx->Stencil.Enabled) {
+    if (ctx->Stencil._Enabled) {
         GLuint temp;
 
         vmesa->regEnable |= HC_HenST_MASK;
diff --git a/src/mesa/drivers/ggi/default/stubs.c b/src/mesa/drivers/ggi/default/stubs.c
index 7b442b6d20b..62722972b2a 100644
--- a/src/mesa/drivers/ggi/default/stubs.c
+++ b/src/mesa/drivers/ggi/default/stubs.c
@@ -472,7 +472,7 @@ static void GGItriangle_flat_depth(GLcontext *ctx, const SWvertex *v0, const SWv
 
 static swrast_tri_func ggimesa_stubs_get_triangle_func(GLcontext *ctx)
 {
-	if (ctx->Stencil.Enabled) return NULL;
+	if (ctx->Stencil._Enabled) return NULL;
 	if (ctx->Polygon.SmoothFlag) return NULL;
 	if (ctx->Polygon.StippleFlag) return NULL;
 	if (ctx->Texture._ReallyEnabled) return NULL;  
diff --git a/src/mesa/drivers/glide/fxdd.c b/src/mesa/drivers/glide/fxdd.c
index 338cd37382a..1bcf2512a63 100644
--- a/src/mesa/drivers/glide/fxdd.c
+++ b/src/mesa/drivers/glide/fxdd.c
@@ -1948,7 +1948,7 @@ fx_check_IsInHardware(GLcontext * ctx)
       return FX_FALLBACK_RENDER_MODE;
    }
 
-   if (ctx->Stencil.Enabled && !fxMesa->haveHwStencil) {
+   if (ctx->Stencil._Enabled && !fxMesa->haveHwStencil) {
       return FX_FALLBACK_STENCIL;
    }
 
diff --git a/src/mesa/drivers/x11/xm_api.c b/src/mesa/drivers/x11/xm_api.c
index 62d30a69878..908fecc85a1 100644
--- a/src/mesa/drivers/x11/xm_api.c
+++ b/src/mesa/drivers/x11/xm_api.c
@@ -2410,11 +2410,8 @@ xbuffer_to_renderbuffer(int buffer)
    case GLX_AUX0_EXT:
       return BUFFER_AUX0;
    case GLX_AUX1_EXT:
-      return BUFFER_AUX1;
    case GLX_AUX2_EXT:
-      return BUFFER_AUX2;
    case GLX_AUX3_EXT:
-      return BUFFER_AUX3;
    case GLX_AUX4_EXT:
    case GLX_AUX5_EXT:
    case GLX_AUX6_EXT:
diff --git a/src/mesa/drivers/x11/xm_span.c b/src/mesa/drivers/x11/xm_span.c
index 57b57494483..309cefcb8e8 100644
--- a/src/mesa/drivers/x11/xm_span.c
+++ b/src/mesa/drivers/x11/xm_span.c
@@ -471,8 +471,26 @@ static void put_row_8R8G8B_pixmap( PUT_ROW_ARGS )
    if (mask) {
       for (i=0;i<n;i++,x++) {
          if (mask[i]) {
+#if 1
+            /*
+             * XXX Something funny is going on here.
+             * If we're drawing into a window that uses a depth 32 TrueColor
+             * visual, we see the right pixels on screen, but when we read
+             * them back with XGetImage() we get random colors.
+             * The alternative code below which uses XPutImage() instead
+             * seems to mostly fix the problem, but not always.
+             * We don't normally create windows with this visual, but glean
+             * does and we're seeing some failures there.
+             */
             XMesaSetForeground( dpy, gc, PACK_8R8G8B( rgba[i][RCOMP], rgba[i][GCOMP], rgba[i][BCOMP] ));
             XMesaDrawPoint( dpy, buffer, gc, (int) x, (int) y );
+#else
+            /* This code works more often, but not always */
+            XMesaImage *rowimg = XMESA_BUFFER(ctx->DrawBuffer)->rowimage;
+            GLuint *ptr4 = (GLuint *) rowimg->data;
+            *ptr4 = PACK_8R8G8B( rgba[i][RCOMP], rgba[i][GCOMP], rgba[i][BCOMP] );
+            XMesaPutImage( dpy, buffer, gc, rowimg, 0, 0, x, y, 1, 1 );
+#endif
          }
       }
    }
diff --git a/src/mesa/main/api_validate.c b/src/mesa/main/api_validate.c
index 5c8955d7c8d..42d1e579e08 100644
--- a/src/mesa/main/api_validate.c
+++ b/src/mesa/main/api_validate.c
@@ -87,11 +87,20 @@ check_valid_to_render(GLcontext *ctx, char *function)
       return GL_FALSE;
    }
 
-   /* Always need vertex positions, unless a vertex program is in use */
-   if (!ctx->VertexProgram._Current &&
-       !ctx->Array.ArrayObj->Vertex.Enabled &&
+#if FEATURE_es2_glsl
+   /* For ES2, we can draw if any vertex array is enabled (and we should
+    * always have a vertex program/shader).
+    */
+   if (ctx->Array.ArrayObj->_Enabled == 0x0 || !ctx->VertexProgram._Current)
+      return GL_FALSE;
+#else
+   /* For regular OpenGL, only draw if we have vertex positions (regardless
+    * of whether or not we have a vertex program/shader).
+    */
+   if (!ctx->Array.ArrayObj->Vertex.Enabled &&
        !ctx->Array.ArrayObj->VertexAttrib[0].Enabled)
       return GL_FALSE;
+#endif
 
    return GL_TRUE;
 }
diff --git a/src/mesa/main/buffers.c b/src/mesa/main/buffers.c
index 85db3868c49..1580487ffd6 100644
--- a/src/mesa/main/buffers.c
+++ b/src/mesa/main/buffers.c
@@ -119,12 +119,6 @@ draw_buffer_enum_to_bitmask(GLenum buffer)
          return BUFFER_BIT_FRONT_LEFT;
       case GL_AUX0:
          return BUFFER_BIT_AUX0;
-      case GL_AUX1:
-         return BUFFER_BIT_AUX1;
-      case GL_AUX2:
-         return BUFFER_BIT_AUX2;
-      case GL_AUX3:
-         return BUFFER_BIT_AUX3;
       case GL_COLOR_ATTACHMENT0_EXT:
          return BUFFER_BIT_COLOR0;
       case GL_COLOR_ATTACHMENT1_EXT:
@@ -176,12 +170,6 @@ read_buffer_enum_to_index(GLenum buffer)
          return BUFFER_FRONT_LEFT;
       case GL_AUX0:
          return BUFFER_AUX0;
-      case GL_AUX1:
-         return BUFFER_AUX1;
-      case GL_AUX2:
-         return BUFFER_AUX2;
-      case GL_AUX3:
-         return BUFFER_AUX3;
       case GL_COLOR_ATTACHMENT0_EXT:
          return BUFFER_COLOR0;
       case GL_COLOR_ATTACHMENT1_EXT:
diff --git a/src/mesa/main/config.h b/src/mesa/main/config.h
index bb3e980bfaa..282ad9514c2 100644
--- a/src/mesa/main/config.h
+++ b/src/mesa/main/config.h
@@ -68,7 +68,7 @@
 #define MAX_PIXEL_MAP_TABLE 256
 
 /** Maximum number of auxillary color buffers */
-#define MAX_AUX_BUFFERS 4
+#define MAX_AUX_BUFFERS 1
 
 /** Maximum order (degree) of curves */
 #ifdef AMIGA
diff --git a/src/mesa/main/dd.h b/src/mesa/main/dd.h
index 411b6a7b21f..a335f77479d 100644
--- a/src/mesa/main/dd.h
+++ b/src/mesa/main/dd.h
@@ -36,6 +36,22 @@
 struct gl_pixelstore_attrib;
 struct gl_display_list;
 
+#if FEATURE_ARB_vertex_buffer_object
+/* Modifies GL_MAP_UNSYNCHRONIZED_BIT to allow driver to fail (return
+ * NULL) if buffer is unavailable for immediate mapping.
+ *
+ * Does GL_MAP_INVALIDATE_RANGE_BIT do this?  It seems so, but it
+ * would require more book-keeping in the driver than seems necessary
+ * at this point.
+ *
+ * Does GL_MAP_INVALDIATE_BUFFER_BIT do this?  Not really -- we don't
+ * want to provoke the driver to throw away the old storage, we will
+ * respect the contents of already referenced data.
+ */
+#define MESA_MAP_NOWAIT_BIT       0x0040
+#endif
+
+
 /**
  * Device driver function table.
  * Core Mesa uses these function pointers to call into device drivers.
@@ -785,6 +801,16 @@ struct dd_function_table {
    void * (*MapBuffer)( GLcontext *ctx, GLenum target, GLenum access,
 			struct gl_buffer_object *obj );
 
+   /* May return NULL if MESA_MAP_NOWAIT_BIT is set in access:
+    */
+   void * (*MapBufferRange)( GLcontext *ctx, GLenum target,
+                             GLintptr offset, GLsizeiptr length, GLbitfield access,
+                             struct gl_buffer_object *obj);
+
+   void (*FlushMappedBufferRange) (GLcontext *ctx, GLenum target, 
+                                   GLintptr offset, GLsizeiptr length,
+                                   struct gl_buffer_object *obj);
+
    GLboolean (*UnmapBuffer)( GLcontext *ctx, GLenum target,
 			     struct gl_buffer_object *obj );
    /*@}*/
@@ -954,6 +980,12 @@ struct dd_function_table {
    GLuint NeedFlush;
    GLuint SaveNeedFlush;
 
+
+   /* Called prior to any of the GLvertexformat functions being
+    * called.  Paired with Driver.FlushVertices().
+    */
+   void (*BeginVertices)( GLcontext *ctx );
+
    /**
     * If inside glBegin()/glEnd(), it should ASSERT(0).  Otherwise, if
     * FLUSH_STORED_VERTICES bit in \p flags is set flushes any buffered
diff --git a/src/mesa/main/debug.c b/src/mesa/main/debug.c
index fcef093ac35..fdd10dd3074 100644
--- a/src/mesa/main/debug.c
+++ b/src/mesa/main/debug.c
@@ -23,6 +23,7 @@
  */
 
 #include "mtypes.h"
+#include "colormac.h"
 #include "context.h"
 #include "hash.h"
 #include "imports.h"
@@ -274,6 +275,27 @@ write_texture_image(struct gl_texture_object *texObj)
       case MESA_FORMAT_ARGB8888:
          write_ppm(s, img->Data, img->Width, img->Height, 4, 2, 1, 0);
          break;
+      case MESA_FORMAT_RGB888:
+         write_ppm(s, img->Data, img->Width, img->Height, 3, 2, 1, 0);
+         break;
+      case MESA_FORMAT_RGB565:
+         {
+            GLubyte *buf2 = (GLubyte *) _mesa_malloc(img->Width * img->Height * 3);
+            GLint i;
+            for (i = 0; i < img->Width * img->Height; i++) {
+               GLint r, g, b;
+               GLushort s = ((GLushort *) img->Data)[i];
+               r = UBYTE_TO_CHAN( ((s >> 8) & 0xf8) | ((s >> 13) & 0x7) );
+               g = UBYTE_TO_CHAN( ((s >> 3) & 0xfc) | ((s >>  9) & 0x3) );
+               b = UBYTE_TO_CHAN( ((s << 3) & 0xf8) | ((s >>  2) & 0x7) );
+               buf2[i*3+1] = r;
+               buf2[i*3+2] = g;
+               buf2[i*3+3] = b;
+            }
+            write_ppm(s, buf2, img->Width, img->Height, 3, 2, 1, 0);
+            _mesa_free(buf2);
+         }
+         break;
       default:
          printf("XXXX unsupported mesa tex format %d in %s\n",
                 img->TexFormat->MesaFormat, __FUNCTION__);
diff --git a/src/mesa/main/enable.c b/src/mesa/main/enable.c
index 7ff3b15c846..a824705bdcc 100644
--- a/src/mesa/main/enable.c
+++ b/src/mesa/main/enable.c
@@ -602,11 +602,6 @@ _mesa_set_enable(GLcontext *ctx, GLenum cap, GLboolean state)
          ctx->Texture.SharedPalette = state;
          break;
       case GL_STENCIL_TEST:
-         if (state && ctx->DrawBuffer->Visual.stencilBits == 0) {
-            _mesa_warning(ctx,
-                          "glEnable(GL_STENCIL_TEST) but no stencil buffer");
-            return;
-         }
          if (ctx->Stencil.Enabled == state)
             return;
          FLUSH_VERTICES(ctx, _NEW_STENCIL);
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index f17b9e1e717..f906de8357e 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -285,14 +285,11 @@ typedef enum
    BUFFER_BACK_LEFT,
    BUFFER_FRONT_RIGHT,
    BUFFER_BACK_RIGHT,
-   /* optional aux buffers */
-   BUFFER_AUX0,
-   BUFFER_AUX1,
-   BUFFER_AUX2,
-   BUFFER_AUX3,
    BUFFER_DEPTH,
    BUFFER_STENCIL,
    BUFFER_ACCUM,
+   /* optional aux buffer */
+   BUFFER_AUX0,
    /* generic renderbuffers */
    BUFFER_COLOR0,
    BUFFER_COLOR1,
@@ -336,9 +333,6 @@ typedef enum
                             BUFFER_BIT_FRONT_RIGHT | \
                             BUFFER_BIT_BACK_RIGHT | \
                             BUFFER_BIT_AUX0 | \
-                            BUFFER_BIT_AUX1 | \
-                            BUFFER_BIT_AUX2 | \
-                            BUFFER_BIT_AUX3 | \
                             BUFFER_BIT_COLOR0 | \
                             BUFFER_BIT_COLOR1 | \
                             BUFFER_BIT_COLOR2 | \
@@ -1018,6 +1012,7 @@ struct gl_stencil_attrib
    GLboolean Enabled;		/**< Enabled flag */
    GLboolean TestTwoSide;	/**< GL_EXT_stencil_two_side */
    GLubyte ActiveFace;		/**< GL_EXT_stencil_two_side (0 or 2) */
+   GLboolean _Enabled;          /**< Enabled and stencil buffer present */
    GLboolean _TestTwoSide;
    GLubyte _BackFace;           /**< Current back stencil state (1 or 2) */
    GLenum Function[3];		/**< Stencil function */
@@ -1446,14 +1441,20 @@ struct gl_texture_attrib
    GLboolean SharedPalette;
    struct gl_color_table Palette;
 
-   /** Per-unit flags */
-   /*@{*/
-   GLbitfield _EnabledUnits;  /**< one bit set for each really-enabled unit */
-   GLbitfield _EnabledCoordUnits;   /**< one bit per enabled coordinate unit */
-   GLbitfield _GenFlags;            /**< for texgen */
-   GLbitfield _TexGenEnabled;  /**< Mask of ENABLE_TEXGEN flags */
-   GLbitfield _TexMatEnabled;  /**< Mask of ENABLE_TEXMAT flags */
-   /*@}*/
+   /** Texture units/samplers used by vertex or fragment texturing */
+   GLbitfield _EnabledUnits;
+
+   /** Texture coord units/sets used for fragment texturing */
+   GLbitfield _EnabledCoordUnits;
+
+   /** Texture coord units that have texgen enabled */
+   GLbitfield _TexGenEnabled;
+
+   /** Texture coord units that have non-identity matrices */
+   GLbitfield _TexMatEnabled;
+
+   /** Bitwise-OR of all Texture.Unit[i]._GenFlags */
+   GLbitfield _GenFlags;
 };
 
 
diff --git a/src/mesa/main/state.c b/src/mesa/main/state.c
index 0a39279bff3..9ea932ea962 100644
--- a/src/mesa/main/state.c
+++ b/src/mesa/main/state.c
@@ -490,7 +490,7 @@ _mesa_update_state_locked( GLcontext *ctx )
    if (new_state & _NEW_LIGHT)
       _mesa_update_lighting( ctx );
 
-   if (new_state & _NEW_STENCIL)
+   if (new_state & (_NEW_STENCIL | _NEW_BUFFERS))
       _mesa_update_stencil( ctx );
 
 #if FEATURE_pixel_transfer
diff --git a/src/mesa/main/stencil.c b/src/mesa/main/stencil.c
index e4a255d0a7e..15c98e20156 100644
--- a/src/mesa/main/stencil.c
+++ b/src/mesa/main/stencil.c
@@ -536,7 +536,11 @@ _mesa_update_stencil(GLcontext *ctx)
 {
    const GLint face = ctx->Stencil._BackFace;
 
-   ctx->Stencil._TestTwoSide =
+   ctx->Stencil._Enabled = (ctx->Stencil.Enabled &&
+                            ctx->DrawBuffer->Visual.stencilBits > 0);
+
+    ctx->Stencil._TestTwoSide =
+       ctx->Stencil._Enabled &&
        (ctx->Stencil.Function[0] != ctx->Stencil.Function[face] ||
 	ctx->Stencil.FailFunc[0] != ctx->Stencil.FailFunc[face] ||
 	ctx->Stencil.ZPassFunc[0] != ctx->Stencil.ZPassFunc[face] ||
diff --git a/src/mesa/main/texstate.c b/src/mesa/main/texstate.c
index e25c9e732c3..2b07da805c0 100644
--- a/src/mesa/main/texstate.c
+++ b/src/mesa/main/texstate.c
@@ -385,28 +385,6 @@ update_texture_compare_function(GLcontext *ctx,
 
 
 /**
- * Helper function for determining which texture object (1D, 2D, cube, etc)
- * should actually be used.
- */
-static void
-texture_override(GLcontext *ctx,
-                 struct gl_texture_unit *texUnit, GLbitfield enableBits,
-                 struct gl_texture_object *texObj, GLuint textureBit)
-{
-   if (!texUnit->_ReallyEnabled && (enableBits & textureBit)) {
-      if (!texObj->_Complete) {
-         _mesa_test_texobj_completeness(ctx, texObj);
-      }
-      if (texObj->_Complete) {
-         texUnit->_ReallyEnabled = textureBit;
-         texUnit->_Current = texObj;
-         update_texture_compare_function(ctx, texObj);
-      }
-   }
-}
-
-
-/**
  * Examine texture unit's combine/env state to update derived state.
  */
 static void
@@ -514,6 +492,7 @@ update_texture_state( GLcontext *ctx )
    GLuint unit;
    struct gl_fragment_program *fprog = NULL;
    struct gl_vertex_program *vprog = NULL;
+   GLbitfield enabledFragUnits = 0x0;
 
    if (ctx->Shader.CurrentProgram &&
        ctx->Shader.CurrentProgram->LinkStatus) {
@@ -545,44 +524,60 @@ update_texture_state( GLcontext *ctx )
     */
    for (unit = 0; unit < ctx->Const.MaxTextureImageUnits; unit++) {
       struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
-      GLbitfield enableBits;
+      GLbitfield enabledVertTargets = 0x0;
+      GLbitfield enabledFragTargets = 0x0;
+      GLbitfield enabledTargets = 0x0;
       GLuint texIndex;
 
-      texUnit->_Current = NULL;
-      texUnit->_ReallyEnabled = 0x0;
-
       /* Get the bitmask of texture target enables.
        * enableBits will be a mask of the TEXTURE_*_BIT flags indicating
        * which texture targets are enabled (fixed function) or referenced
        * by a fragment shader/program.  When multiple flags are set, we'll
-       * settle on the one with highest priority (see texture_override below).
+       * settle on the one with highest priority (see below).
        */
-      enableBits = 0x0;
       if (vprog) {
-         enableBits |= vprog->Base.TexturesUsed[unit];
+         enabledVertTargets |= vprog->Base.TexturesUsed[unit];
       }
+
       if (fprog) {
-         enableBits |= fprog->Base.TexturesUsed[unit];
+         enabledFragTargets |= fprog->Base.TexturesUsed[unit];
       }
       else {
          /* fixed-function fragment program */
-         enableBits |= texUnit->Enabled;
+         enabledFragTargets |= texUnit->Enabled;
       }
 
-      if (enableBits == 0x0)
+      enabledTargets = enabledVertTargets | enabledFragTargets;
+
+      texUnit->_ReallyEnabled = 0x0;
+
+      if (enabledTargets == 0x0) {
+         /* neither vertex nor fragment processing uses this unit */
          continue;
+      }
 
-      /* Look for the highest-priority texture target that's enabled and
-       * complete.  That's the one we'll use for texturing.  If we're using
-       * a fragment program we're guaranteed that bitcount(enabledBits) <= 1.
+      /* Look for the highest priority texture target that's enabled (or used
+       * by the vert/frag shaders) and "complete".  That's the one we'll use
+       * for texturing.  If we're using vert/frag program we're guaranteed
+       * that bitcount(enabledBits) <= 1.
        * Note that the TEXTURE_x_INDEX values are in high to low priority.
        */
       for (texIndex = 0; texIndex < NUM_TEXTURE_TARGETS; texIndex++) {
-         texture_override(ctx, texUnit, enableBits,
-                          texUnit->CurrentTex[texIndex], 1 << texIndex);
+         if (enabledTargets & (1 << texIndex)) {
+            struct gl_texture_object *texObj = texUnit->CurrentTex[texIndex];
+            if (!texObj->_Complete) {
+               _mesa_test_texobj_completeness(ctx, texObj);
+            }
+            if (texObj->_Complete) {
+               texUnit->_ReallyEnabled = 1 << texIndex;
+               _mesa_reference_texobj(&texUnit->_Current, texObj);
+               break;
+            }
+         }
       }
 
       if (!texUnit->_ReallyEnabled) {
+         _mesa_reference_texobj(&texUnit->_Current, NULL);
          continue;
       }
 
@@ -590,7 +585,11 @@ update_texture_state( GLcontext *ctx )
 
       ctx->Texture._EnabledUnits |= (1 << unit);
 
+      if (enabledFragTargets)
+         enabledFragUnits |= (1 << unit);
+
       update_tex_combine(ctx, texUnit);
+      update_texture_compare_function(ctx, texUnit->_Current);
    }
 
 
@@ -601,7 +600,7 @@ update_texture_state( GLcontext *ctx )
          = (fprog->Base.InputsRead >> FRAG_ATTRIB_TEX0) & coordMask;
    }
    else {
-      ctx->Texture._EnabledCoordUnits = ctx->Texture._EnabledUnits;
+      ctx->Texture._EnabledCoordUnits = enabledFragUnits;
    }
 
    /* Setup texgen for those texture coordinate sets that are in use */
diff --git a/src/mesa/main/vtxfmt.c b/src/mesa/main/vtxfmt.c
index 0204979003e..1f807dc3dc3 100644
--- a/src/mesa/main/vtxfmt.c
+++ b/src/mesa/main/vtxfmt.c
@@ -54,9 +54,12 @@
    ASSERT( tnl->Current );						\
    ASSERT( tnl->SwapCount < NUM_VERTEX_FORMAT_ENTRIES );		\
    ASSERT( tmp_offset >= 0 );						\
-									\
-   /* Save the swapped function's dispatch entry so it can be */	\
-   /* restored later. */						\
+                                                                        \
+   if (tnl->SwapCount == 0)                                             \
+      ctx->Driver.BeginVertices( ctx );                                 \
+                                                                        \
+   /* Save the swapped function's dispatch entry so it can be */        \
+   /* restored later. */                                                \
    tnl->Swapped[tnl->SwapCount].location = & (((_glapi_proc *)ctx->Exec)[tmp_offset]); \
    tnl->Swapped[tnl->SwapCount].function = (_glapi_proc)TAG(FUNC);	\
    tnl->SwapCount++;							\
diff --git a/src/mesa/sources.mak b/src/mesa/sources.mak
index 357661456ad..92c2183a942 100644
--- a/src/mesa/sources.mak
+++ b/src/mesa/sources.mak
@@ -91,7 +91,6 @@ MATH_SOURCES = \
 	math/m_vector.c
 
 MATH_XFORM_SOURCES = \
-	$(MATH_SOURCES) \
 	math/m_xform.c
 
 SWRAST_SOURCES = \
@@ -165,11 +164,6 @@ VBO_SOURCES = \
 	vbo/vbo_save_draw.c \
 	vbo/vbo_save_loopback.c 
 
-VF_SOURCES = \
-	vf/vf.c \
-	vf/vf_generic.c \
-	vf/vf_sse.c
-
 STATETRACKER_SOURCES = \
 	state_tracker/st_atom.c \
 	state_tracker/st_atom_blend.c \
@@ -311,6 +305,7 @@ COMMON_DRIVER_SOURCES =			\
 # Sources for building non-Gallium drivers
 MESA_SOURCES = \
 	$(MAIN_SOURCES)		\
+	$(MATH_SOURCES)		\
 	$(MATH_XFORM_SOURCES)	\
 	$(VBO_SOURCES)		\
 	$(TNL_SOURCES)		\
diff --git a/src/mesa/state_tracker/st_atom_depth.c b/src/mesa/state_tracker/st_atom_depth.c
index 4ea62dda184..0aa128f9479 100644
--- a/src/mesa/state_tracker/st_atom_depth.c
+++ b/src/mesa/state_tracker/st_atom_depth.c
@@ -98,9 +98,11 @@ update_depth_stencil_alpha(struct st_context *st)
 
    memset(dsa, 0, sizeof(*dsa));
 
-   dsa->depth.enabled = ctx->Depth.Test;
-   dsa->depth.writemask = ctx->Depth.Mask;
-   dsa->depth.func = st_compare_func_to_pipe(ctx->Depth.Func);
+   if (ctx->Depth.Test && ctx->DrawBuffer->Visual.depthBits > 0) {
+      dsa->depth.enabled = 1;
+      dsa->depth.writemask = ctx->Depth.Mask;
+      dsa->depth.func = st_compare_func_to_pipe(ctx->Depth.Func);
+   }
 
    if (ctx->Query.CurrentOcclusionObject &&
        ctx->Query.CurrentOcclusionObject->Active)
diff --git a/src/mesa/state_tracker/st_cb_bufferobjects.c b/src/mesa/state_tracker/st_cb_bufferobjects.c
index 562ac6c65c5..28e387c399d 100644
--- a/src/mesa/state_tracker/st_cb_bufferobjects.c
+++ b/src/mesa/state_tracker/st_cb_bufferobjects.c
@@ -206,8 +206,40 @@ st_bufferobj_map(GLcontext *ctx, GLenum target, GLenum access,
 }
 
 
+
 /**
- * Called via glMapBufferARB().
+ * Called via glMapBufferRange().
+ */
+static void *
+st_bufferobj_map_range(GLcontext *ctx, GLenum target, 
+                       GLintptr offset, GLsizeiptr length, GLbitfield access,
+                       struct gl_buffer_object *obj)
+{
+   struct pipe_context *pipe = st_context(ctx)->pipe;
+   struct st_buffer_object *st_obj = st_buffer_object(obj);
+   GLuint flags = 0;
+
+   if (access & GL_MAP_WRITE_BIT)
+      flags |= PIPE_BUFFER_USAGE_CPU_WRITE;
+
+   if (access & GL_MAP_READ_BIT)
+      flags |= PIPE_BUFFER_USAGE_CPU_READ;
+
+   /* ... other flags ...
+    */
+
+   if (access & MESA_MAP_NOWAIT_BIT)
+      flags |= PIPE_BUFFER_USAGE_DONTBLOCK;
+
+   obj->Pointer = pipe_buffer_map(pipe->screen, st_obj->buffer, flags);
+   return obj->Pointer;
+}
+
+
+
+
+/**
+ * Called via glUnmapBufferARB().
  */
 static GLboolean
 st_bufferobj_unmap(GLcontext *ctx, GLenum target, struct gl_buffer_object *obj)
@@ -230,5 +262,6 @@ st_init_bufferobject_functions(struct dd_function_table *functions)
    functions->BufferSubData = st_bufferobj_subdata;
    functions->GetBufferSubData = st_bufferobj_get_subdata;
    functions->MapBuffer = st_bufferobj_map;
+   functions->MapBufferRange = st_bufferobj_map_range;
    functions->UnmapBuffer = st_bufferobj_unmap;
 }
diff --git a/src/mesa/swrast/s_buffers.c b/src/mesa/swrast/s_buffers.c
index af475ad8cb5..34ae1b4fccf 100644
--- a/src/mesa/swrast/s_buffers.c
+++ b/src/mesa/swrast/s_buffers.c
@@ -317,10 +317,7 @@ _swrast_Clear(GLcontext *ctx, GLbitfield buffers)
 	 BUFFER_BIT_DEPTH |
 	 BUFFER_BIT_STENCIL |
 	 BUFFER_BIT_ACCUM |
-         BUFFER_BIT_AUX0 |
-         BUFFER_BIT_AUX1 |
-         BUFFER_BIT_AUX2 |
-         BUFFER_BIT_AUX3;
+         BUFFER_BIT_AUX0;
       assert((buffers & (~legalBits)) == 0);
    }
 #endif
diff --git a/src/mesa/swrast/s_context.c b/src/mesa/swrast/s_context.c
index 719e6b82964..4dbccbb2d59 100644
--- a/src/mesa/swrast/s_context.c
+++ b/src/mesa/swrast/s_context.c
@@ -61,7 +61,7 @@ _swrast_update_rasterflags( GLcontext *ctx )
    if (ctx->Depth.Test)                   rasterMask |= DEPTH_BIT;
    if (swrast->_FogEnabled)               rasterMask |= FOG_BIT;
    if (ctx->Scissor.Enabled)              rasterMask |= CLIP_BIT;
-   if (ctx->Stencil.Enabled)              rasterMask |= STENCIL_BIT;
+   if (ctx->Stencil._Enabled)             rasterMask |= STENCIL_BIT;
    if (ctx->Visual.rgbMode) {
       const GLuint colorMask = *((GLuint *) &ctx->Color.ColorMask);
       if (colorMask != 0xffffffff)        rasterMask |= MASKING_BIT;
diff --git a/src/mesa/swrast/s_span.c b/src/mesa/swrast/s_span.c
index ab7b82b19d9..5e1f412d6bf 100644
--- a/src/mesa/swrast/s_span.c
+++ b/src/mesa/swrast/s_span.c
@@ -846,11 +846,11 @@ _swrast_write_index_span( GLcontext *ctx, SWspan *span)
    }
 
    /* Stencil and Z testing */
-   if (ctx->Depth.Test || ctx->Stencil.Enabled) {
+   if (ctx->Stencil._Enabled || ctx->Depth.Test) {
       if (!(span->arrayMask & SPAN_Z))
          _swrast_span_interpolate_z(ctx, span);
 
-      if (ctx->Stencil.Enabled) {
+      if (ctx->Stencil._Enabled) {
          if (!_swrast_stencil_and_ztest_span(ctx, span)) {
             span->arrayMask = origArrayMask;
             return;
@@ -1211,7 +1211,7 @@ shade_texture_span(GLcontext *ctx, SWspan *span)
          _swrast_exec_fragment_shader(ctx, span);
       }
    }
-   else if (ctx->Texture._EnabledUnits) {
+   else if (ctx->Texture._EnabledCoordUnits) {
       /* conventional texturing */
 
 #if CHAN_BITS == 32
@@ -1250,7 +1250,7 @@ _swrast_write_rgba_span( GLcontext *ctx, SWspan *span)
    void * const origRgba = span->array->rgba;
    const GLboolean shader = (ctx->FragmentProgram._Current
                              || ctx->ATIFragmentShader._Enabled);
-   const GLboolean shaderOrTexture = shader || ctx->Texture._EnabledUnits;
+   const GLboolean shaderOrTexture = shader || ctx->Texture._EnabledCoordUnits;
    struct gl_framebuffer *fb = ctx->DrawBuffer;
 
    /*
@@ -1317,11 +1317,11 @@ _swrast_write_rgba_span( GLcontext *ctx, SWspan *span)
    }
 
    /* Stencil and Z testing */
-   if (ctx->Stencil.Enabled || ctx->Depth.Test) {
+   if (ctx->Stencil._Enabled || ctx->Depth.Test) {
       if (!(span->arrayMask & SPAN_Z))
          _swrast_span_interpolate_z(ctx, span);
 
-      if (ctx->Stencil.Enabled && fb->Visual.stencilBits > 0) {
+      if (ctx->Stencil._Enabled) {
          /* Combined Z/stencil tests */
          if (!_swrast_stencil_and_ztest_span(ctx, span)) {
             /* all fragments failed test */
diff --git a/src/mesa/swrast/s_triangle.c b/src/mesa/swrast/s_triangle.c
index ab08b7325a5..9260e35066f 100644
--- a/src/mesa/swrast/s_triangle.c
+++ b/src/mesa/swrast/s_triangle.c
@@ -266,7 +266,7 @@ affine_span(GLcontext *ctx, SWspan *span,
             struct affine_info *info)
 {
    GLchan sample[4];  /* the filtered texture sample */
-   const GLuint texEnableSave = ctx->Texture._EnabledUnits;
+   const GLuint texEnableSave = ctx->Texture._EnabledCoordUnits;
 
    /* Instead of defining a function for each mode, a test is done
     * between the outer and inner loops. This is to reduce code size
@@ -397,7 +397,7 @@ affine_span(GLcontext *ctx, SWspan *span,
    GLchan *dest = span->array->rgba[0];
 
    /* Disable tex units so they're not re-applied in swrast_write_rgba_span */
-   ctx->Texture._EnabledUnits = 0x0;
+   ctx->Texture._EnabledCoordUnits = 0x0;
 
    span->intTex[0] -= FIXED_HALF;
    span->intTex[1] -= FIXED_HALF;
@@ -504,7 +504,7 @@ affine_span(GLcontext *ctx, SWspan *span,
    _swrast_write_rgba_span(ctx, span);
 
    /* re-enable texture units */
-   ctx->Texture._EnabledUnits = texEnableSave;
+   ctx->Texture._EnabledCoordUnits = texEnableSave;
 
 #undef SPAN_NEAREST
 #undef SPAN_LINEAR
@@ -664,8 +664,8 @@ fast_persp_span(GLcontext *ctx, SWspan *span,
    GLfloat tex_coord[3], tex_step[3];
    GLchan *dest = span->array->rgba[0];
 
-   const GLuint savedTexEnable = ctx->Texture._EnabledUnits;
-   ctx->Texture._EnabledUnits = 0;
+   const GLuint texEnableSave = ctx->Texture._EnabledCoordUnits;
+   ctx->Texture._EnabledCoordUnits = 0;
 
    tex_coord[0] = span->attrStart[FRAG_ATTRIB_TEX0][0]  * (info->smask + 1);
    tex_step[0] = span->attrStepX[FRAG_ATTRIB_TEX0][0] * (info->smask + 1);
@@ -778,7 +778,7 @@ fast_persp_span(GLcontext *ctx, SWspan *span,
 #undef SPAN_LINEAR
 
    /* restore state */
-   ctx->Texture._EnabledUnits = savedTexEnable;
+   ctx->Texture._EnabledCoordUnits = texEnableSave;
 }
 
 
@@ -1022,7 +1022,7 @@ _swrast_choose_triangle( GLcontext *ctx )
           ctx->Depth.Test &&
           ctx->Depth.Mask == GL_FALSE &&
           ctx->Depth.Func == GL_LESS &&
-          !ctx->Stencil.Enabled) {
+          !ctx->Stencil._Enabled) {
          if ((rgbmode &&
               ctx->Color.ColorMask[0] == 0 &&
               ctx->Color.ColorMask[1] == 0 &&
diff --git a/src/mesa/vbo/vbo_exec.c b/src/mesa/vbo/vbo_exec.c
index 635f239acc8..e168a89ea5e 100644
--- a/src/mesa/vbo/vbo_exec.c
+++ b/src/mesa/vbo/vbo_exec.c
@@ -57,6 +57,7 @@ void vbo_exec_init( GLcontext *ctx )
 
    ctx->Driver.NeedFlush = 0;
    ctx->Driver.CurrentExecPrimitive = PRIM_OUTSIDE_BEGIN_END;
+   ctx->Driver.BeginVertices = vbo_exec_BeginVertices;
    ctx->Driver.FlushVertices = vbo_exec_FlushVertices;
 
    vbo_exec_invalidate_state( ctx, ~0 );
diff --git a/src/mesa/vbo/vbo_exec.h b/src/mesa/vbo/vbo_exec.h
index ddbcbe11814..100bb8a5dec 100644
--- a/src/mesa/vbo/vbo_exec.h
+++ b/src/mesa/vbo/vbo_exec.h
@@ -43,7 +43,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 /* Wierd implementation stuff:
  */
-#define VBO_VERT_BUFFER_SIZE (1024*16)	/* dwords == 64k */
+#define VBO_VERT_BUFFER_SIZE (1024*64)	/* bytes */
 #define VBO_MAX_ATTR_CODEGEN 16 
 #define ERROR_ATTRIB 16
 
@@ -78,14 +78,15 @@ struct vbo_exec_context
 
    struct {
       struct gl_buffer_object *bufferobj;
-      GLubyte *buffer_map;
 
-      GLuint vertex_size;
+      GLuint vertex_size;       /* in dwords */
 
       struct _mesa_prim prim[VBO_MAX_PRIM];
       GLuint prim_count;
 
-      GLfloat *vbptr;		     /* cursor, points into buffer */
+      GLfloat *buffer_map;
+      GLfloat *buffer_ptr;              /* cursor, points into buffer */
+      GLuint   buffer_used;             /* in bytes */
       GLfloat vertex[VBO_ATTRIB_MAX*4]; /* current vertex */
 
       GLuint vert_count;
@@ -140,6 +141,9 @@ struct vbo_exec_context
 void vbo_exec_init( GLcontext *ctx );
 void vbo_exec_destroy( GLcontext *ctx );
 void vbo_exec_invalidate_state( GLcontext *ctx, GLuint new_state );
+void vbo_exec_FlushVertices_internal( GLcontext *ctx, GLboolean unmap );
+
+void vbo_exec_BeginVertices( GLcontext *ctx );
 void vbo_exec_FlushVertices( GLcontext *ctx, GLuint flags );
 
 
@@ -151,7 +155,8 @@ void vbo_exec_array_destroy( struct vbo_exec_context *exec );
 
 void vbo_exec_vtx_init( struct vbo_exec_context *exec );
 void vbo_exec_vtx_destroy( struct vbo_exec_context *exec );
-void vbo_exec_vtx_flush( struct vbo_exec_context *exec );
+void vbo_exec_vtx_flush( struct vbo_exec_context *exec, GLboolean unmap );
+void vbo_exec_vtx_map( struct vbo_exec_context *exec );
 void vbo_exec_vtx_wrap( struct vbo_exec_context *exec );
 
 void vbo_exec_eval_update( struct vbo_exec_context *exec );
diff --git a/src/mesa/vbo/vbo_exec_api.c b/src/mesa/vbo/vbo_exec_api.c
index 2743bf6b55c..c0ffdb55e47 100644
--- a/src/mesa/vbo/vbo_exec_api.c
+++ b/src/mesa/vbo/vbo_exec_api.c
@@ -62,7 +62,7 @@ static void vbo_exec_wrap_buffers( struct vbo_exec_context *exec )
    if (exec->vtx.prim_count == 0) {
       exec->vtx.copied.nr = 0;
       exec->vtx.vert_count = 0;
-      exec->vtx.vbptr = (GLfloat *)exec->vtx.buffer_map;
+      exec->vtx.buffer_ptr = exec->vtx.buffer_map;
    }
    else {
       GLuint last_begin = exec->vtx.prim[exec->vtx.prim_count-1].begin;
@@ -80,7 +80,7 @@ static void vbo_exec_wrap_buffers( struct vbo_exec_context *exec )
       /* Execute the buffer and save copied vertices.
        */
       if (exec->vtx.vert_count)
-	 vbo_exec_vtx_flush( exec );
+	 vbo_exec_vtx_flush( exec, GL_FALSE );
       else {
 	 exec->vtx.prim_count = 0;
 	 exec->vtx.copied.nr = 0;
@@ -121,9 +121,9 @@ void vbo_exec_vtx_wrap( struct vbo_exec_context *exec )
    assert(exec->vtx.max_vert - exec->vtx.vert_count > exec->vtx.copied.nr);
 
    for (i = 0 ; i < exec->vtx.copied.nr ; i++) {
-      _mesa_memcpy( exec->vtx.vbptr, data, 
+      _mesa_memcpy( exec->vtx.buffer_ptr, data, 
 		    exec->vtx.vertex_size * sizeof(GLfloat));
-      exec->vtx.vbptr += exec->vtx.vertex_size;
+      exec->vtx.buffer_ptr += exec->vtx.vertex_size;
       data += exec->vtx.vertex_size;
       exec->vtx.vert_count++;
    }
@@ -251,9 +251,10 @@ static void vbo_exec_wrap_upgrade_vertex( struct vbo_exec_context *exec,
    exec->vtx.attrsz[attr] = newsz;
 
    exec->vtx.vertex_size += newsz - oldsz;
-   exec->vtx.max_vert = VBO_VERT_BUFFER_SIZE / exec->vtx.vertex_size;
+   exec->vtx.max_vert = ((VBO_VERT_BUFFER_SIZE - exec->vtx.buffer_used) / 
+                         (exec->vtx.vertex_size * sizeof(GLfloat)));
    exec->vtx.vert_count = 0;
-   exec->vtx.vbptr = (GLfloat *)exec->vtx.buffer_map;
+   exec->vtx.buffer_ptr = exec->vtx.buffer_map;
    
 
    /* Recalculate all the attrptr[] values
@@ -279,10 +280,10 @@ static void vbo_exec_wrap_upgrade_vertex( struct vbo_exec_context *exec,
    if (exec->vtx.copied.nr)
    {
       GLfloat *data = exec->vtx.copied.buffer;
-      GLfloat *dest = exec->vtx.vbptr;
+      GLfloat *dest = exec->vtx.buffer_ptr;
       GLuint j;
 
-      assert(exec->vtx.vbptr == (GLfloat *)exec->vtx.buffer_map);
+      assert(exec->vtx.buffer_ptr == exec->vtx.buffer_map);
       
       for (i = 0 ; i < exec->vtx.copied.nr ; i++) {
 	 for (j = 0 ; j < VBO_ATTRIB_MAX ; j++) {
@@ -308,7 +309,7 @@ static void vbo_exec_wrap_upgrade_vertex( struct vbo_exec_context *exec,
 	 }
       }
 
-      exec->vtx.vbptr = dest;
+      exec->vtx.buffer_ptr = dest;
       exec->vtx.vert_count += exec->vtx.copied.nr;
       exec->vtx.copied.nr = 0;
    }
@@ -373,9 +374,9 @@ do {								\
       GLuint i;							\
 								\
       for (i = 0; i < exec->vtx.vertex_size; i++)		\
-	 exec->vtx.vbptr[i] = exec->vtx.vertex[i];		\
+	 exec->vtx.buffer_ptr[i] = exec->vtx.vertex[i];		\
 								\
-      exec->vtx.vbptr += exec->vtx.vertex_size;			\
+      exec->vtx.buffer_ptr += exec->vtx.vertex_size;			\
       exec->ctx->Driver.NeedFlush |= FLUSH_STORED_VERTICES;	\
 								\
       if (++exec->vtx.vert_count >= exec->vtx.max_vert)		\
@@ -532,7 +533,7 @@ static void GLAPIENTRY vbo_exec_Begin( GLenum mode )
        * begin/end pairs.
        */
       if (exec->vtx.vertex_size && !exec->vtx.attrsz[0]) 
-	 vbo_exec_FlushVertices( ctx, ~0 );
+	 vbo_exec_FlushVertices_internal( ctx, GL_FALSE );
 
       i = exec->vtx.prim_count++;
       exec->vtx.prim[i].mode = mode;
@@ -566,7 +567,7 @@ static void GLAPIENTRY vbo_exec_End( void )
       ctx->Driver.CurrentExecPrimitive = PRIM_OUTSIDE_BEGIN_END;
 
       if (exec->vtx.prim_count == VBO_MAX_PRIM)
-	 vbo_exec_vtx_flush( exec );	
+	 vbo_exec_vtx_flush( exec, GL_FALSE );
    }
    else 
       _mesa_error( ctx, GL_INVALID_OPERATION, "glEnd" );
@@ -672,23 +673,19 @@ void vbo_use_buffer_objects(GLcontext *ctx)
     */
    GLuint bufName = 0xaabbccdd;
    GLenum target = GL_ARRAY_BUFFER_ARB;
-   GLenum access = GL_READ_WRITE_ARB;
    GLenum usage = GL_STREAM_DRAW_ARB;
-   GLsizei size = VBO_VERT_BUFFER_SIZE * sizeof(GLfloat);
+   GLsizei size = VBO_VERT_BUFFER_SIZE;
 
    /* Make sure this func is only used once */
    assert(exec->vtx.bufferobj == ctx->Array.NullBufferObj);
    if (exec->vtx.buffer_map) {
       _mesa_align_free(exec->vtx.buffer_map);
+      exec->vtx.buffer_map = NULL;
    }
 
    /* Allocate a real buffer object now */
    exec->vtx.bufferobj = ctx->Driver.NewBufferObject(ctx, bufName, target);
    ctx->Driver.BufferData(ctx, target, size, NULL, usage, exec->vtx.bufferobj);
-
-   /* and map it */
-   exec->vtx.buffer_map
-      = ctx->Driver.MapBuffer(ctx, target, access, exec->vtx.bufferobj);
 }
 
 
@@ -708,7 +705,8 @@ void vbo_exec_vtx_init( struct vbo_exec_context *exec )
                                  ctx->Array.NullBufferObj);
 
    ASSERT(!exec->vtx.buffer_map);
-   exec->vtx.buffer_map = ALIGN_MALLOC(VBO_VERT_BUFFER_SIZE * sizeof(GLfloat), 64);
+   exec->vtx.buffer_map = (GLfloat *)ALIGN_MALLOC(VBO_VERT_BUFFER_SIZE, 64);
+
    vbo_exec_vtxfmt_init( exec );
 
    /* Hook our functions into the dispatch table.
@@ -747,22 +745,45 @@ void vbo_exec_vtx_destroy( struct vbo_exec_context *exec )
    }
 }
 
-
-void vbo_exec_FlushVertices( GLcontext *ctx, GLuint flags )
+void vbo_exec_BeginVertices( GLcontext *ctx )
 {
    struct vbo_exec_context *exec = &vbo_context(ctx)->exec;
+   if (0) _mesa_printf("%s\n", __FUNCTION__);
+   vbo_exec_vtx_map( exec );
+}
 
-   if (exec->ctx->Driver.CurrentExecPrimitive != PRIM_OUTSIDE_BEGIN_END)
-      return;
+void vbo_exec_FlushVertices_internal( GLcontext *ctx, GLboolean unmap )
+{
+   struct vbo_exec_context *exec = &vbo_context(ctx)->exec;
 
-   if (exec->vtx.vert_count) {
-      vbo_exec_vtx_flush( exec );
+   if (exec->vtx.vert_count || unmap) {
+      vbo_exec_vtx_flush( exec, unmap );
    }
 
    if (exec->vtx.vertex_size) {
       vbo_exec_copy_to_current( exec );
       reset_attrfv( exec );
    }
+}
+
+
+
+void vbo_exec_FlushVertices( GLcontext *ctx, GLuint flags )
+{
+   struct vbo_exec_context *exec = &vbo_context(ctx)->exec;
+
+   if (0) _mesa_printf("%s\n", __FUNCTION__);
+
+   if (exec->ctx->Driver.CurrentExecPrimitive != PRIM_OUTSIDE_BEGIN_END) {
+      if (0) _mesa_printf("%s - inside begin/end\n", __FUNCTION__);
+      return;
+   }
+
+   vbo_exec_FlushVertices_internal( ctx, GL_TRUE );
+
+   /* Need to do this to ensure BeginVertices gets called again:
+    */
+   _mesa_restore_exec_vtxfmt( ctx );
 
    exec->ctx->Driver.NeedFlush = 0;
 }
diff --git a/src/mesa/vbo/vbo_exec_array.c b/src/mesa/vbo/vbo_exec_array.c
index 8871e10cf60..cf7c7056c5a 100644
--- a/src/mesa/vbo/vbo_exec_array.c
+++ b/src/mesa/vbo/vbo_exec_array.c
@@ -31,6 +31,7 @@
 #include "main/api_validate.h"
 #include "main/api_noop.h"
 #include "main/varray.h"
+#include "main/bufferobj.h"
 #include "glapi/dispatch.h"
 
 #include "vbo_context.h"
@@ -291,6 +292,47 @@ vbo_exec_DrawArrays(GLenum mode, GLint start, GLsizei count)
    prim[0].indexed = 0;
 
    vbo->draw_prims( ctx, exec->array.inputs, prim, 1, NULL, start, start + count - 1 );
+
+#if 0
+   {
+      int i;
+
+      _mesa_printf("vbo_exec_DrawArrays(mode 0x%x, start %d, count %d):\n",
+                   mode, start, count);
+
+      for (i = 0; i < 32; i++) {
+         GLuint bufName = exec->array.inputs[i]->BufferObj->Name;
+         GLint stride = exec->array.inputs[i]->Stride;
+         _mesa_printf("attr %2d: size %d stride %d  enabled %d  "
+                      "ptr %p  Bufobj %u\n",
+                      i,
+                      exec->array.inputs[i]->Size,
+                      stride,
+                      /*exec->array.inputs[i]->Enabled,*/
+                      exec->array.legacy_array[i]->Enabled,
+                      exec->array.inputs[i]->Ptr,
+                      bufName);
+         
+         if (bufName) {
+            struct gl_buffer_object *buf = _mesa_lookup_bufferobj(ctx, bufName);
+            GLubyte *p = ctx->Driver.MapBuffer(ctx, GL_ARRAY_BUFFER_ARB,
+                                            GL_READ_ONLY_ARB, buf);
+            int offset = (int) exec->array.inputs[i]->Ptr;
+            float *f = (float *) (p + offset);
+            int *k = (int *) f;
+            int i;
+            int n = (count * stride) / 4;
+            if (n > 32)
+               n = 32;
+            _mesa_printf("  Data at offset %d:\n", offset);
+            for (i = 0; i < n; i++) {
+               _mesa_printf("    float[%d] = 0x%08x %f\n", i, k[i], f[i]);
+            }
+            ctx->Driver.UnmapBuffer(ctx, GL_ARRAY_BUFFER_ARB, buf);
+         }
+      }
+   }
+#endif
 }
 
 
diff --git a/src/mesa/vbo/vbo_exec_draw.c b/src/mesa/vbo/vbo_exec_draw.c
index 45133fa4f32..38b6c56f479 100644
--- a/src/mesa/vbo/vbo_exec_draw.c
+++ b/src/mesa/vbo/vbo_exec_draw.c
@@ -70,7 +70,7 @@ static GLuint vbo_copy_vertices( struct vbo_exec_context *exec )
    GLuint ovf, i;
    GLuint sz = exec->vtx.vertex_size;
    GLfloat *dst = exec->vtx.copied.buffer;
-   GLfloat *src = ((GLfloat *)exec->vtx.buffer_map + 
+   GLfloat *src = (exec->vtx.buffer_map + 
 		   exec->vtx.prim[exec->vtx.prim_count-1].start * 
 		   exec->vtx.vertex_size);
 
@@ -147,7 +147,7 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
    struct vbo_exec_context *exec = &vbo->exec;
    struct gl_client_array *arrays = exec->vtx.arrays;
    GLuint count = exec->vtx.vert_count;
-   GLubyte *data = exec->vtx.buffer_map;
+   GLubyte *data = (GLubyte *)exec->vtx.buffer_map;
    const GLuint *map;
    GLuint attr;
    GLbitfield varying_inputs = 0x0;
@@ -234,10 +234,78 @@ static void vbo_exec_bind_arrays( GLcontext *ctx )
 }
 
 
+static void vbo_exec_vtx_unmap( struct vbo_exec_context *exec )
+{
+   GLenum target = GL_ARRAY_BUFFER_ARB;
+
+   if (exec->vtx.bufferobj->Name) {
+      GLcontext *ctx = exec->ctx;
+
+      exec->vtx.buffer_used += (exec->vtx.buffer_ptr -
+                                exec->vtx.buffer_map) * sizeof(float);
+            
+      ctx->Driver.UnmapBuffer(ctx, target, exec->vtx.bufferobj);
+      exec->vtx.buffer_map = NULL;
+      exec->vtx.buffer_ptr = NULL;
+      exec->vtx.max_vert = 0;
+   }
+}
+
+void vbo_exec_vtx_map( struct vbo_exec_context *exec )
+{
+   GLcontext *ctx = exec->ctx;
+   GLenum target = GL_ARRAY_BUFFER_ARB;
+   GLenum access = GL_READ_WRITE_ARB;
+   GLenum usage = GL_STREAM_DRAW_ARB;
+   
+   if (exec->vtx.bufferobj->Name == 0)
+      return;
+
+   if (exec->vtx.buffer_map != NULL) {
+      assert(0);
+      exec->vtx.buffer_map = NULL;
+   }
+
+   if (VBO_VERT_BUFFER_SIZE > exec->vtx.buffer_used + 1024 &&
+       ctx->Driver.MapBufferRange)
+   {
+      exec->vtx.buffer_map = 
+         (GLfloat *)ctx->Driver.MapBufferRange(ctx, 
+                                               target, 
+                                               exec->vtx.buffer_used,
+                                               (VBO_VERT_BUFFER_SIZE - 
+                                                exec->vtx.buffer_used),
+                                               (GL_MAP_WRITE_BIT |
+                                                GL_MAP_INVALIDATE_RANGE_BIT | 
+                                                GL_MAP_UNSYNCHRONIZED_BIT | 
+                                                MESA_MAP_NOWAIT_BIT),
+                                               exec->vtx.bufferobj);
+   }
+
+   if (exec->vtx.buffer_map) {
+      exec->vtx.buffer_map += exec->vtx.buffer_used / sizeof(float);
+   }
+   else {
+      exec->vtx.buffer_used = 0;
+
+      ctx->Driver.BufferData(ctx, target, 
+                             VBO_VERT_BUFFER_SIZE, 
+                             NULL, usage, exec->vtx.bufferobj);
+
+      exec->vtx.buffer_map = 
+         (GLfloat *)ctx->Driver.MapBuffer(ctx, target, access, exec->vtx.bufferobj);
+   }
+
+   if (0) _mesa_printf("map %d..\n", exec->vtx.buffer_used);
+}
+
+
+
 /**
  * Execute the buffer and save copied verts.
  */
-void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
+void vbo_exec_vtx_flush( struct vbo_exec_context *exec,
+                         GLboolean unmap )
 {
    if (0)
       vbo_exec_debug_verts( exec );
@@ -250,25 +318,22 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
 
       if (exec->vtx.copied.nr != exec->vtx.vert_count) {
 	 GLcontext *ctx = exec->ctx;
-
-	 GLenum target = GL_ARRAY_BUFFER_ARB;
-	 GLenum access = GL_READ_WRITE_ARB;
-	 GLenum usage = GL_STREAM_DRAW_ARB;
-	 GLsizei size = VBO_VERT_BUFFER_SIZE * sizeof(GLfloat);
 	 
-	 /* Before the unmap (why?)
+	 /* Before the update_state() as this may raise _NEW_ARRAY
+          * from _mesa_set_varying_vp_inputs().
 	  */
 	 vbo_exec_bind_arrays( ctx );
 
          if (ctx->NewState)
             _mesa_update_state( ctx );
 
-         /* if using a real VBO, unmap it before drawing */
          if (exec->vtx.bufferobj->Name) {
-            ctx->Driver.UnmapBuffer(ctx, target, exec->vtx.bufferobj);
-            exec->vtx.buffer_map = NULL;
+            vbo_exec_vtx_unmap( exec );
          }
 
+         if (0) _mesa_printf("%s %d %d\n", __FUNCTION__, exec->vtx.prim_count,
+                      exec->vtx.vert_count);
+
 	 vbo_context(ctx)->draw_prims( ctx, 
 				       exec->vtx.inputs, 
 				       exec->vtx.prim, 
@@ -277,16 +342,31 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
 				       0,
 				       exec->vtx.vert_count - 1);
 
-	 /* If using a real VBO, get new storage */
-         if (exec->vtx.bufferobj->Name) {
-            ctx->Driver.BufferData(ctx, target, size, NULL, usage, exec->vtx.bufferobj);
-            exec->vtx.buffer_map = 
-               ctx->Driver.MapBuffer(ctx, target, access, exec->vtx.bufferobj);
-         }
+	 /* If using a real VBO, get new storage -- unless asked not to.
+          */
+         if (exec->vtx.bufferobj->Name && !unmap) {
+            vbo_exec_vtx_map( exec );
+          }
       }
    }
 
+   /* May have to unmap explicitly if we didn't draw:
+    */
+   if (unmap && 
+       exec->vtx.bufferobj->Name &&
+       exec->vtx.buffer_map) {
+      vbo_exec_vtx_unmap( exec );
+   }
+
+
+   if (unmap) 
+      exec->vtx.max_vert = 0;
+   else
+      exec->vtx.max_vert = ((VBO_VERT_BUFFER_SIZE - exec->vtx.buffer_used) / 
+                            (exec->vtx.vertex_size * sizeof(GLfloat)));
+
+
+   exec->vtx.buffer_ptr = exec->vtx.buffer_map;
    exec->vtx.prim_count = 0;
    exec->vtx.vert_count = 0;
-   exec->vtx.vbptr = (GLfloat *)exec->vtx.buffer_map;
 }
diff --git a/src/mesa/vbo/vbo_save.h b/src/mesa/vbo/vbo_save.h
index 7dda54af48b..9558f838837 100644
--- a/src/mesa/vbo/vbo_save.h
+++ b/src/mesa/vbo/vbo_save.h
@@ -130,7 +130,7 @@ struct vbo_save_context {
    struct vbo_save_vertex_store *vertex_store;
    struct vbo_save_primitive_store *prim_store;
 
-   GLfloat *vbptr;		   /* cursor, points into buffer */
+   GLfloat *buffer_ptr;		   /* cursor, points into buffer */
    GLfloat vertex[VBO_ATTRIB_MAX*4];	   /* current values */
    GLfloat *attrptr[VBO_ATTRIB_MAX];
    GLuint vert_count;
diff --git a/src/mesa/vbo/vbo_save_api.c b/src/mesa/vbo/vbo_save_api.c
index ddfd2765774..52b6f1884ee 100644
--- a/src/mesa/vbo/vbo_save_api.c
+++ b/src/mesa/vbo/vbo_save_api.c
@@ -241,7 +241,7 @@ static void _save_reset_counters( GLcontext *ctx )
    save->buffer = (save->vertex_store->buffer + 
 		   save->vertex_store->used);
 
-   assert(save->buffer == save->vbptr);
+   assert(save->buffer == save->buffer_ptr);
 
    if (save->vertex_size)
       save->max_vert = ((VBO_SAVE_BUFFER_SIZE - save->vertex_store->used) / 
@@ -343,7 +343,7 @@ static void _save_compile_vertex_list( GLcontext *ctx )
       /* Allocate and map new store:
        */
       save->vertex_store = alloc_vertex_store( ctx );
-      save->vbptr = map_vertex_store( ctx, save->vertex_store );
+      save->buffer_ptr = map_vertex_store( ctx, save->vertex_store );
    } 
 
    if (save->prim_store->used > VBO_SAVE_PRIM_SIZE - 6) {
@@ -414,9 +414,9 @@ static void _save_wrap_filled_vertex( GLcontext *ctx )
    assert(save->max_vert - save->vert_count > save->copied.nr);
 
    for (i = 0 ; i < save->copied.nr ; i++) {
-      _mesa_memcpy( save->vbptr, data, save->vertex_size * sizeof(GLfloat));
+      _mesa_memcpy( save->buffer_ptr, data, save->vertex_size * sizeof(GLfloat));
       data += save->vertex_size;
-      save->vbptr += save->vertex_size;
+      save->buffer_ptr += save->vertex_size;
       save->vert_count++;
    }
 }
@@ -550,7 +550,7 @@ static void _save_upgrade_vertex( GLcontext *ctx,
 	 }
       }
 
-      save->vbptr = dest;
+      save->buffer_ptr = dest;
       save->vert_count += save->copied.nr;
    }
 }
@@ -622,9 +622,9 @@ do {								\
       GLuint i;							\
 								\
       for (i = 0; i < save->vertex_size; i++)			\
-	 save->vbptr[i] = save->vertex[i];			\
+	 save->buffer_ptr[i] = save->vertex[i];			\
 								\
-      save->vbptr += save->vertex_size;				\
+      save->buffer_ptr += save->vertex_size;				\
 								\
       if (++save->vert_count >= save->max_vert)			\
 	 _save_wrap_filled_vertex( ctx );			\
@@ -1035,7 +1035,7 @@ void vbo_save_NewList( GLcontext *ctx, GLuint list, GLenum mode )
    if (!save->vertex_store) 
       save->vertex_store = alloc_vertex_store( ctx );
       
-   save->vbptr = map_vertex_store( ctx, save->vertex_store );
+   save->buffer_ptr = map_vertex_store( ctx, save->vertex_store );
    
    _save_reset_vertex( ctx );
    _save_reset_counters( ctx );