Merge branch 'master' of ssh+git://idr@git.freedesktop.org/git/mesa/mesa

63 files changed, 577 insertions, 650 deletions

diff --git a/src/mesa/drivers/allegro/amesa.c b/src/mesa/drivers/allegro/amesa.c
index 594668affd..518211ccf4 100644
--- a/src/mesa/drivers/allegro/amesa.c
+++ b/src/mesa/drivers/allegro/amesa.c
@@ -338,7 +338,7 @@ void AMesaDestroyBuffer(AMesaBuffer buffer)
 {

    if (buffer->Screen)     destroy_bitmap(buffer->Screen);

    if (buffer->Background) destroy_bitmap(buffer->Background);

-   _mesa_destroy_framebuffer(buffer->GLBuffer);

+   _mesa_unreference_framebuffer(&buffer->GLBuffer);

    free(buffer);

 }

 

diff --git a/src/mesa/drivers/dri/common/dri_util.c b/src/mesa/drivers/dri/common/dri_util.c
index cc3dcf9d8d..07ac4c7cd5 100644
--- a/src/mesa/drivers/dri/common/dri_util.c
+++ b/src/mesa/drivers/dri/common/dri_util.c
@@ -340,7 +340,7 @@ static GLboolean DoBindContext(__DRInativeDisplay *dpy,
 	DRM_SPINUNLOCK(&psp->pSAREA->drawable_lock, psp->drawLockID);
     }
 
-    if ((pdp != prp) && (!pdp->pStamp || *pdp->pStamp != pdp->lastStamp)) {
+    if ((pdp != prp) && (!prp->pStamp || *prp->pStamp != prp->lastStamp)) {
 	DRM_SPINLOCK(&psp->pSAREA->drawable_lock, psp->drawLockID);
 	__driUtilUpdateDrawableInfo(prp);
 	DRM_SPINUNLOCK(&psp->pSAREA->drawable_lock, psp->drawLockID);
@@ -420,7 +420,7 @@ __driUtilUpdateDrawableInfo(__DRIdrawablePrivate *pdp)
     psp = pdp->driScreenPriv;
     if (!psp) {
 	/* ERROR!!! */
-       _mesa_problem("Warning! Possible infinite loop due to bug "
+       _mesa_problem(NULL, "Warning! Possible infinite loop due to bug "
 		     "in file %s, line %d\n",
 		     __FILE__, __LINE__);
 	return;
diff --git a/src/mesa/drivers/dri/fb/fb_dri.c b/src/mesa/drivers/dri/fb/fb_dri.c
index 08b52b4d8f..a6d7590eff 100644
--- a/src/mesa/drivers/dri/fb/fb_dri.c
+++ b/src/mesa/drivers/dri/fb/fb_dri.c
@@ -480,11 +480,7 @@ fbCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 fbDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   struct gl_framebuffer *mesa_framebuffer = (struct gl_framebuffer *)driDrawPriv->driverPrivate;
-   
-   _mesa_free(mesa_framebuffer->Attachment[BUFFER_BACK_LEFT].Renderbuffer->Data);
-   _mesa_destroy_framebuffer(mesa_framebuffer);
-   driDrawPriv->driverPrivate = NULL;
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/ffb/ffb_xmesa.c b/src/mesa/drivers/dri/ffb/ffb_xmesa.c
index 215aaf8ffb..4c5323d230 100644
--- a/src/mesa/drivers/dri/ffb/ffb_xmesa.c
+++ b/src/mesa/drivers/dri/ffb/ffb_xmesa.c
@@ -392,7 +392,7 @@ ffbCreateBuffer(__DRIscreenPrivate *driScrnPriv,
 static void
 ffbDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/gamma/gamma_xmesa.c b/src/mesa/drivers/dri/gamma/gamma_xmesa.c
index e8922b1503..f41682cea7 100644
--- a/src/mesa/drivers/dri/gamma/gamma_xmesa.c
+++ b/src/mesa/drivers/dri/gamma/gamma_xmesa.c
@@ -97,7 +97,7 @@ gammaCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 gammaDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 static void
diff --git a/src/mesa/drivers/dri/i810/i810screen.c b/src/mesa/drivers/dri/i810/i810screen.c
index d0e58f667c..f64c10a9ae 100644
--- a/src/mesa/drivers/dri/i810/i810screen.c
+++ b/src/mesa/drivers/dri/i810/i810screen.c
@@ -398,7 +398,7 @@ i810CreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 i810DestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+    _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/i915/intel_screen.c b/src/mesa/drivers/dri/i915/intel_screen.c
index 107cf137ff..67e176a1c6 100644
--- a/src/mesa/drivers/dri/i915/intel_screen.c
+++ b/src/mesa/drivers/dri/i915/intel_screen.c
@@ -449,7 +449,7 @@ static GLboolean intelCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 
 static void intelDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/i915tex/i915_tex_layout.c b/src/mesa/drivers/dri/i915tex/i915_tex_layout.c
index 333fefef85..c844f5351d 100644
--- a/src/mesa/drivers/dri/i915tex/i915_tex_layout.c
+++ b/src/mesa/drivers/dri/i915tex/i915_tex_layout.c
@@ -62,15 +62,23 @@ i915_miptree_layout(struct intel_mipmap_tree * mt)
    case GL_TEXTURE_CUBE_MAP:{
          const GLuint dim = mt->width0;
          GLuint face;
+         GLuint lvlWidth = mt->width0, lvlHeight = mt->height0;
+
+         assert(lvlWidth == lvlHeight); /* cubemap images are square */
 
          /* double pitch for cube layouts */
          mt->pitch = ((dim * mt->cpp * 2 + 3) & ~3) / mt->cpp;
          mt->total_height = dim * 4;
 
-         for (level = mt->first_level; level <= mt->last_level; level++)
+         for (level = mt->first_level; level <= mt->last_level; level++) {
             intel_miptree_set_level_info(mt, level, 6,
                                          0, 0,
-                                         mt->pitch, mt->total_height, 1);
+                                         /*OLD: mt->pitch, mt->total_height,*/
+                                         lvlWidth, lvlHeight,
+                                         1);
+            lvlWidth /= 2;
+            lvlHeight /= 2;
+         }
 
          for (face = 0; face < 6; face++) {
             GLuint x = initial_offsets[face][0] * dim;
diff --git a/src/mesa/drivers/dri/i915tex/intel_mipmap_tree.c b/src/mesa/drivers/dri/i915tex/intel_mipmap_tree.c
index 2b1077aee0..93372f4dcc 100644
--- a/src/mesa/drivers/dri/i915tex/intel_mipmap_tree.c
+++ b/src/mesa/drivers/dri/i915tex/intel_mipmap_tree.c
@@ -309,8 +309,13 @@ intel_miptree_image_data(struct intel_context *intel,
       height = dst->level[level].height;
       if(dst->compressed)
 	 height /= 4;
-      intel_region_data(intel->intelScreen, dst->region, dst_offset + dst_depth_offset[i], 0, 0, src, src_row_pitch, 0, 0,   /* source x,y */
-                        dst->level[level].width, height);
+      intel_region_data(intel->intelScreen, dst->region,
+                        dst_offset + dst_depth_offset[i], /* dst_offset */
+                        0, 0,                             /* dstx, dsty */
+                        src,
+                        src_row_pitch,
+                        0, 0,                             /* source x, y */
+                        dst->level[level].width, height); /* width, height */
 
       src += src_image_pitch;
    }
diff --git a/src/mesa/drivers/dri/i915tex/intel_regions.c b/src/mesa/drivers/dri/i915tex/intel_regions.c
index 1205b180ca..a114bdf896 100644
--- a/src/mesa/drivers/dri/i915tex/intel_regions.c
+++ b/src/mesa/drivers/dri/i915tex/intel_regions.c
@@ -217,7 +217,8 @@ _mesa_copy_rect(GLubyte * dst,
                 GLuint dst_y,
                 GLuint width,
                 GLuint height,
-                GLubyte * src, GLuint src_pitch, GLuint src_x, GLuint src_y)
+                const GLubyte * src,
+                GLuint src_pitch, GLuint src_x, GLuint src_y)
 {
    GLuint i;
 
@@ -253,7 +254,7 @@ intel_region_data(intelScreenPrivate *intelScreen,
                   struct intel_region *dst,
                   GLuint dst_offset,
                   GLuint dstx, GLuint dsty,
-                  void *src, GLuint src_pitch,
+                  const void *src, GLuint src_pitch,
                   GLuint srcx, GLuint srcy, GLuint width, GLuint height)
 {
    struct intel_context *intel = intelScreenContext(intelScreen);
diff --git a/src/mesa/drivers/dri/i915tex/intel_regions.h b/src/mesa/drivers/dri/i915tex/intel_regions.h
index 1415bace69..d938c107a4 100644
--- a/src/mesa/drivers/dri/i915tex/intel_regions.h
+++ b/src/mesa/drivers/dri/i915tex/intel_regions.h
@@ -103,7 +103,7 @@ void intel_region_data(intelScreenPrivate *intelScreen,
                        struct intel_region *dest,
                        GLuint dest_offset,
                        GLuint destx, GLuint desty,
-                       void *src, GLuint src_stride,
+                       const void *src, GLuint src_stride,
                        GLuint srcx, GLuint srcy, GLuint width, GLuint height);
 
 /* Copy rectangular sub-regions
diff --git a/src/mesa/drivers/dri/i915tex/intel_screen.c b/src/mesa/drivers/dri/i915tex/intel_screen.c
index efa1b014a6..ac83254593 100644
--- a/src/mesa/drivers/dri/i915tex/intel_screen.c
+++ b/src/mesa/drivers/dri/i915tex/intel_screen.c
@@ -609,7 +609,7 @@ intelCreateBuffer(__DRIscreenPrivate * driScrnPriv,
 static void
 intelDestroyBuffer(__DRIdrawablePrivate * driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/i965/brw_draw.c b/src/mesa/drivers/dri/i965/brw_draw.c
index 092c6bafc2..4eb2eedd42 100644
--- a/src/mesa/drivers/dri/i965/brw_draw.c
+++ b/src/mesa/drivers/dri/i965/brw_draw.c
@@ -458,6 +458,7 @@ void brw_draw_prims( GLcontext *ctx,
     * swrast to do the drawing.
     */
    if (!retval) {
+       _swsetup_Wakeup(ctx);
       _tnl_draw_prims(ctx, arrays, prim, nr_prims, ib, min_index, max_index);
    }
 
diff --git a/src/mesa/drivers/dri/i965/intel_screen.c b/src/mesa/drivers/dri/i965/intel_screen.c
index 08f0bb340f..5dac50df32 100644
--- a/src/mesa/drivers/dri/i965/intel_screen.c
+++ b/src/mesa/drivers/dri/i965/intel_screen.c
@@ -457,7 +457,7 @@ static GLboolean intelCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 
 static void intelDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/mach64/mach64_screen.c b/src/mesa/drivers/dri/mach64/mach64_screen.c
index 1014b8acd5..4e9e216e7d 100644
--- a/src/mesa/drivers/dri/mach64/mach64_screen.c
+++ b/src/mesa/drivers/dri/mach64/mach64_screen.c
@@ -435,7 +435,7 @@ mach64CreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 mach64DestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/mga/mga_xmesa.c b/src/mesa/drivers/dri/mga/mga_xmesa.c
index f024f73eb6..67a6f8bdf0 100644
--- a/src/mesa/drivers/dri/mga/mga_xmesa.c
+++ b/src/mesa/drivers/dri/mga/mga_xmesa.c
@@ -831,7 +831,7 @@ mgaCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 mgaDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 static void
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_context.c b/src/mesa/drivers/dri/nouveau/nouveau_context.c
index f811dc1b72..aec7b19771 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_context.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_context.c
@@ -141,9 +141,15 @@ GLboolean nouveauCreateContext( const __GLcontextModes *glVisual,
 	if (!nouveauDRMGetParam(nmesa, NOUVEAU_GETPARAM_FB_PHYSICAL,
 		 		&nmesa->vram_phys))
 	   return GL_FALSE;
+	if (!nouveauDRMGetParam(nmesa, NOUVEAU_GETPARAM_FB_SIZE,
+		 		&nmesa->vram_size))
+	   return GL_FALSE;
 	if (!nouveauDRMGetParam(nmesa, NOUVEAU_GETPARAM_AGP_PHYSICAL,
 		 		&nmesa->agp_phys))
 	   return GL_FALSE;
+	if (!nouveauDRMGetParam(nmesa, NOUVEAU_GETPARAM_AGP_SIZE,
+		 		&nmesa->agp_size))
+	   return GL_FALSE;
 	if (!nouveauFifoInit(nmesa))
 	   return GL_FALSE;
 	nouveauObjectInit(nmesa);
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_context.h b/src/mesa/drivers/dri/nouveau/nouveau_context.h
index f61fcbb48c..f79a8675f4 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_context.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_context.h
@@ -100,7 +100,9 @@ typedef struct nouveau_context {
 
 	/* Physical addresses of AGP/VRAM apertures */
 	uint64_t vram_phys;
+	uint64_t vram_size;
 	uint64_t agp_phys;
+	uint64_t agp_size;
 
 	/* Channel synchronisation */
 	nouveau_notifier *syncNotifier;
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_object.c b/src/mesa/drivers/dri/nouveau/nouveau_object.c
index 302009c8b1..468b18e6d9 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_object.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_object.c
@@ -4,17 +4,14 @@
 #include "nouveau_reg.h"
 
 
-GLboolean nouveauCreateContextObject(nouveauContextPtr nmesa, int handle, int class, uint32_t flags, uint32_t dma_in, uint32_t dma_out, uint32_t dma_notifier)
+GLboolean nouveauCreateContextObject(nouveauContextPtr nmesa,
+				     uint32_t handle, int class)
 {
 	drm_nouveau_object_init_t cto;
 	int ret;
 
 	cto.handle = handle;
 	cto.class  = class;
-	cto.flags  = flags;
-	cto.dma0= dma_in;
-	cto.dma1= dma_out;
-	cto.dma_notifier = dma_notifier;
 	ret = drmCommandWrite(nmesa->driFd, DRM_NOUVEAU_OBJECT_INIT, &cto, sizeof(cto));
 
 	return ret == 0;
@@ -22,6 +19,7 @@ GLboolean nouveauCreateContextObject(nouveauContextPtr nmesa, int handle, int cl
 
 GLboolean nouveauCreateDmaObject(nouveauContextPtr nmesa,
       				 uint32_t handle,
+				 int      class,
 				 uint32_t offset,
 				 uint32_t size,
 				 int	  target,
@@ -30,6 +28,7 @@ GLboolean nouveauCreateDmaObject(nouveauContextPtr nmesa,
 	drm_nouveau_dma_object_init_t dma;
 	int ret;
 
+	dma.class  = class;
 	dma.handle = handle;
 	dma.target = target;
 	dma.access = access;
@@ -40,6 +39,27 @@ GLboolean nouveauCreateDmaObject(nouveauContextPtr nmesa,
 	return ret == 0;
 }
 
+GLboolean nouveauCreateDmaObjectFromMem(nouveauContextPtr nmesa,
+					uint32_t handle, int class,
+					nouveau_mem *mem,
+					int access)
+{
+	uint32_t offset = mem->offset;
+	int target = mem->type & (NOUVEAU_MEM_FB | NOUVEAU_MEM_AGP);
+
+	if (!target)
+		return GL_FALSE;
+
+	if (target & NOUVEAU_MEM_FB)
+		offset -= nmesa->vram_phys;
+	else if (target & NOUVEAU_MEM_AGP)
+		offset -= nmesa->agp_phys;
+
+	return nouveauCreateDmaObject(nmesa, handle, class,
+				      offset, mem->size,
+				      target, access);
+}
+
 void nouveauObjectOnSubchannel(nouveauContextPtr nmesa, int subchannel, int handle)
 {
 	BEGIN_RING_SIZE(subchannel, 0, 1);
@@ -53,31 +73,25 @@ void nouveauObjectInit(nouveauContextPtr nmesa)
 #endif
 
 /* We need to know vram size.. and AGP size (and even if the card is AGP..) */
-	nouveauCreateDmaObject( nmesa, NvDmaFB,
-				0, (256*1024*1024),
-				0 /*NV_DMA_TARGET_FB*/, 0 /*NV_DMA_ACCESS_RW*/);
-	nouveauCreateDmaObject( nmesa, NvDmaAGP,
-	      			nmesa->agp_phys, (128*1024*1024),
-				3 /* AGP */, 0 /* RW */);
-
-	nouveauCreateContextObject(nmesa, Nv3D, nmesa->screen->card->class_3d,
-	      			   0, 0, 0, 0);
+	nouveauCreateDmaObject( nmesa, NvDmaFB, NV_DMA_IN_MEMORY,
+				0, nmesa->vram_size,
+				NOUVEAU_MEM_FB,
+				NOUVEAU_MEM_ACCESS_RW);
+	nouveauCreateDmaObject( nmesa, NvDmaAGP, NV_DMA_IN_MEMORY,
+	      			0, nmesa->agp_size,
+				NOUVEAU_MEM_AGP,
+				NOUVEAU_MEM_ACCESS_RW);
+
+	nouveauCreateContextObject(nmesa, Nv3D, nmesa->screen->card->class_3d);
 	if (nmesa->screen->card->type>=NV_10) {
-		nouveauCreateContextObject(nmesa, NvCtxSurf2D, NV10_CONTEXT_SURFACES_2D,
-	      			   0, 0, 0, 0);
-		nouveauCreateContextObject(nmesa, NvImageBlit, NV10_IMAGE_BLIT,
-	      			   NV_DMA_CONTEXT_FLAGS_PATCH_SRCCOPY, 0, 0, 0);
+		nouveauCreateContextObject(nmesa, NvCtxSurf2D, NV10_CONTEXT_SURFACES_2D);
+		nouveauCreateContextObject(nmesa, NvImageBlit, NV10_IMAGE_BLIT);
 	} else {
-		nouveauCreateContextObject(nmesa, NvCtxSurf2D, NV04_CONTEXT_SURFACES_2D,
-	      			   0, 0, 0, 0);
-		nouveauCreateContextObject(nmesa, NvCtxSurf3D, NV04_CONTEXT_SURFACES_3D,
-	      			   0, 0, 0, 0);
-		nouveauCreateContextObject(nmesa, NvImageBlit, NV_IMAGE_BLIT,
-	      			   NV_DMA_CONTEXT_FLAGS_PATCH_SRCCOPY, 0, 0, 0);
+		nouveauCreateContextObject(nmesa, NvCtxSurf2D, NV04_CONTEXT_SURFACES_2D);
+		nouveauCreateContextObject(nmesa, NvCtxSurf3D, NV04_CONTEXT_SURFACES_3D);
+		nouveauCreateContextObject(nmesa, NvImageBlit, NV_IMAGE_BLIT);
 	}
-	nouveauCreateContextObject(nmesa, NvMemFormat,
-	      			   NV_MEMORY_TO_MEMORY_FORMAT,
-	      			   0, 0, 0, 0);
+	nouveauCreateContextObject(nmesa, NvMemFormat, NV_MEMORY_TO_MEMORY_FORMAT);
 
 #ifdef ALLOW_MULTI_SUBCHANNEL
 	nouveauObjectOnSubchannel(nmesa, NvSubCtxSurf2D, NvCtxSurf2D);
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_object.h b/src/mesa/drivers/dri/nouveau/nouveau_object.h
index daad281029..0be9b4309c 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_object.h
+++ b/src/mesa/drivers/dri/nouveau/nouveau_object.h
@@ -30,15 +30,18 @@ enum DMASubchannel {
 extern void nouveauObjectOnSubchannel(nouveauContextPtr nmesa, int subchannel, int handle);
 
 extern GLboolean nouveauCreateContextObject(nouveauContextPtr nmesa,
-      					    int handle, int class,
-					    uint32_t flags,
-					    uint32_t dma_in,
-					    uint32_t dma_out,
-					    uint32_t dma_notifier);
+      					    uint32_t handle, int class);
 extern GLboolean nouveauCreateDmaObject(nouveauContextPtr nmesa,
       					uint32_t handle,
+					int      class,
 					uint32_t offset,
 					uint32_t size,
 					int      target,
 					int      access);
+extern GLboolean nouveauCreateDmaObjectFromMem(nouveauContextPtr nmesa,
+					       uint32_t     handle,
+					       int          class,
+					       nouveau_mem *mem,
+					       int          access);
+
 #endif
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_screen.c b/src/mesa/drivers/dri/nouveau/nouveau_screen.c
index 881b20149f..e00080fce1 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_screen.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_screen.c
@@ -202,7 +202,7 @@ nouveauCreateBuffer(__DRIscreenPrivate *driScrnPriv,
 static void
 nouveauDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-	_mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+	_mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 static int
@@ -327,8 +327,10 @@ void * __driCreateNewScreen_20050727( __DRInativeDisplay *dpy, int scrn, __DRIsc
 	__DRIscreenPrivate *psp;
 	static const __DRIversion ddx_expected = { 1, 2, 0 };
 	static const __DRIversion dri_expected = { 4, 0, 0 };
-	static const __DRIversion drm_expected = { 0, 0, 3 };
-
+	static const __DRIversion drm_expected = { 0, 0, NOUVEAU_DRM_HEADER_PATCHLEVEL };
+#if NOUVEAU_DRM_HEADER_PATCHLEVEL != 4
+#error nouveau_drm.h version doesn't match expected version
+#endif
 	dri_interface = interface;
 
 	if (!driCheckDriDdxDrmVersions2("nouveau",
diff --git a/src/mesa/drivers/dri/nouveau/nouveau_sync.c b/src/mesa/drivers/dri/nouveau/nouveau_sync.c
index c47ff3a985..428b19b46e 100644
--- a/src/mesa/drivers/dri/nouveau/nouveau_sync.c
+++ b/src/mesa/drivers/dri/nouveau/nouveau_sync.c
@@ -58,10 +58,9 @@ nouveau_notifier_new(GLcontext *ctx, GLuint handle, GLuint count)
 		return NULL;
 	}
 
-	if (!nouveauCreateDmaObject(nmesa, handle, notifier->mem->offset,
-						   notifier->mem->size,
-						   0 /* NV_DMA_TARGET_FB */,
-						   0 /* NV_DMA_ACCESS_RW */)) {
+	if (!nouveauCreateDmaObjectFromMem(nmesa, handle, NV_DMA_IN_MEMORY,
+					   notifier->mem,
+					   NOUVEAU_MEM_ACCESS_RW)) {
 		nouveau_mem_free(ctx, notifier->mem);
 		FREE(notifier);
 		return NULL;
diff --git a/src/mesa/drivers/dri/r128/r128_screen.c b/src/mesa/drivers/dri/r128/r128_screen.c
index 4f1b20aadd..880dee85c2 100644
--- a/src/mesa/drivers/dri/r128/r128_screen.c
+++ b/src/mesa/drivers/dri/r128/r128_screen.c
@@ -357,7 +357,7 @@ r128CreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 r128DestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/r200/r200_maos_arrays.c b/src/mesa/drivers/dri/r200/r200_maos_arrays.c
index 6a6c30a2b0..3162b508c2 100644
--- a/src/mesa/drivers/dri/r200/r200_maos_arrays.c
+++ b/src/mesa/drivers/dri/r200/r200_maos_arrays.c
@@ -420,14 +420,14 @@ void r200EmitArrays( GLcontext *ctx, GLubyte *vimap_rev )
 	    if (!rmesa->tcl.vertex_data[i].buf) {
 	       if (ctx->VertexProgram._Enabled)
 		  emit_vector( ctx,
-			 &(rmesa->tcl.vertex_data[attrib]),
+			 &(rmesa->tcl.vertex_data[i]),
 			 (char *)VB->AttribPtr[attrib]->data,
 			 1,
 			 VB->AttribPtr[attrib]->stride,
 			 count);
 	       else
 		  emit_vecfog( ctx,
-			 &(rmesa->tcl.vertex_data[attrib]),
+			 &(rmesa->tcl.vertex_data[i]),
 			 (char *)VB->AttribPtr[attrib]->data,
 			 VB->AttribPtr[attrib]->stride,
 			 count);
diff --git a/src/mesa/drivers/dri/r300/r300_cmdbuf.c b/src/mesa/drivers/dri/r300/r300_cmdbuf.c
index ace9b66724..2c7b5aa011 100644
--- a/src/mesa/drivers/dri/r300/r300_cmdbuf.c
+++ b/src/mesa/drivers/dri/r300/r300_cmdbuf.c
@@ -292,13 +292,13 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	ALLOC_STATE( vpt, always, R300_VPT_CMDSIZE, "vpt", 0 );
 		r300->hw.vpt.cmd[R300_VPT_CMD_0] = cmdpacket0(R300_SE_VPORT_XSCALE, 6);
 	ALLOC_STATE( unk2080, always, 2, "unk2080", 0 );
-		r300->hw.unk2080.cmd[0] = cmdpacket0(0x2080, 1);
+		r300->hw.unk2080.cmd[0] = cmdpacket0(R300_VAP_CNTL, 1);
 	ALLOC_STATE( vte, always, 3, "vte", 0 );
 		r300->hw.vte.cmd[0] = cmdpacket0(R300_SE_VTE_CNTL, 2);
 	ALLOC_STATE( unk2134, always, 3, "unk2134", 0 );
 		r300->hw.unk2134.cmd[0] = cmdpacket0(0x2134, 2);
 	ALLOC_STATE( unk2140, always, 2, "unk2140", 0 );
-		r300->hw.unk2140.cmd[0] = cmdpacket0(0x2140, 1);
+		r300->hw.unk2140.cmd[0] = cmdpacket0(R300_VAP_CNTL_STATUS, 1);
 	ALLOC_STATE( vir[0], variable, R300_VIR_CMDSIZE, "vir/0", 0 );
 		r300->hw.vir[0].cmd[R300_VIR_CMD_0] = cmdpacket0(R300_VAP_INPUT_ROUTE_0_0, 1);
 	ALLOC_STATE( vir[1], variable, R300_VIR_CMDSIZE, "vir/1", 1 );
@@ -308,11 +308,11 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	ALLOC_STATE( unk21DC, always, 2, "unk21DC", 0 );
 		r300->hw.unk21DC.cmd[0] = cmdpacket0(0x21DC, 1);
 	ALLOC_STATE( unk221C, always, 2, "unk221C", 0 );
-		r300->hw.unk221C.cmd[0] = cmdpacket0(0x221C, 1);
+		r300->hw.unk221C.cmd[0] = cmdpacket0(R300_VAP_UNKNOWN_221C, 1);
 	ALLOC_STATE( unk2220, always, 5, "unk2220", 0 );
 		r300->hw.unk2220.cmd[0] = cmdpacket0(0x2220, 4);
 	ALLOC_STATE( unk2288, always, 2, "unk2288", 0 );
-		r300->hw.unk2288.cmd[0] = cmdpacket0(0x2288, 1);
+		r300->hw.unk2288.cmd[0] = cmdpacket0(R300_VAP_UNKNOWN_2288, 1);
 	ALLOC_STATE( vof, always, R300_VOF_CMDSIZE, "vof", 0 );
 		r300->hw.vof.cmd[R300_VOF_CMD_0] = cmdpacket0(R300_VAP_OUTPUT_VTX_FMT_0, 2);
 	ALLOC_STATE( pvs, always, R300_PVS_CMDSIZE, "pvs", 0 );
@@ -336,9 +336,9 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	ALLOC_STATE( unk4260, always, 4, "unk4260", 0 );
 		r300->hw.unk4260.cmd[0] = cmdpacket0(0x4260, 3);
 	ALLOC_STATE( unk4274, always, 5, "unk4274", 0 );
-		r300->hw.unk4274.cmd[0] = cmdpacket0(0x4274, 4);
+		r300->hw.unk4274.cmd[0] = cmdpacket0(R300_RE_SHADE, 4);
 	ALLOC_STATE( unk4288, always, 4, "unk4288", 0 );
-		r300->hw.unk4288.cmd[0] = cmdpacket0(0x4288, 3);
+		r300->hw.unk4288.cmd[0] = cmdpacket0(R300_RE_POLYGON_MODE, 3);
 	ALLOC_STATE( fogp, always, 3, "fogp", 0 );
 		r300->hw.fogp.cmd[0] = cmdpacket0(R300_RE_FOG_SCALE, 2);
 	ALLOC_STATE( unk42A0, always, 2, "unk42A0", 0 );
@@ -346,7 +346,7 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	ALLOC_STATE( zbs, always, R300_ZBS_CMDSIZE, "zbs", 0 );
 		r300->hw.zbs.cmd[R300_ZBS_CMD_0] = cmdpacket0(R300_RE_ZBIAS_T_FACTOR, 4);
 	ALLOC_STATE( unk42B4, always, 2, "unk42B4", 0 );
-		r300->hw.unk42B4.cmd[0] = cmdpacket0(0x42B4, 1);
+		r300->hw.unk42B4.cmd[0] = cmdpacket0(R300_RE_OCCLUSION_CNTL, 1);
 	ALLOC_STATE( cul, always, R300_CUL_CMDSIZE, "cul", 0 );
 		r300->hw.cul.cmd[R300_CUL_CMD_0] = cmdpacket0(R300_RE_CULL_CNTL, 1);
 	ALLOC_STATE( unk42C0, always, 3, "unk42C0", 0 );
@@ -393,7 +393,7 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	ALLOC_STATE( cmk, always, R300_CMK_CMDSIZE, "cmk", 0 );
 		r300->hw.cmk.cmd[R300_CMK_CMD_0] = cmdpacket0(R300_RB3D_COLORMASK, 1);
 	ALLOC_STATE( unk4E10, always, 4, "unk4E10", 0 );
-		r300->hw.unk4E10.cmd[0] = cmdpacket0(0x4E10, 3);
+		r300->hw.unk4E10.cmd[0] = cmdpacket0(R300_RB3D_BLEND_COLOR, 3);
 	ALLOC_STATE( cb, always, R300_CB_CMDSIZE, "cb", 0 );
 		r300->hw.cb.cmd[R300_CB_CMD_0] = cmdpacket0(R300_RB3D_COLOROFFSET0, 1);
 		r300->hw.cb.cmd[R300_CB_CMD_1] = cmdpacket0(R300_RB3D_COLORPITCH0, 1);
@@ -406,7 +406,7 @@ void r300InitCmdBuf(r300ContextPtr r300)
 	ALLOC_STATE( zs, always, R300_ZS_CMDSIZE, "zstencil", 0 );
 		r300->hw.zs.cmd[R300_ZS_CMD_0] = cmdpacket0(R300_RB3D_ZSTENCIL_CNTL_0, 3);
 	ALLOC_STATE( unk4F10, always, 5, "unk4F10", 0 );
-		r300->hw.unk4F10.cmd[0] = cmdpacket0(0x4F10, 4);
+		r300->hw.unk4F10.cmd[0] = cmdpacket0(R300_RB3D_ZSTENCIL_FORMAT, 4);
 	ALLOC_STATE( zb, always, R300_ZB_CMDSIZE, "zb", 0 );
 		r300->hw.zb.cmd[R300_ZB_CMD_0] = cmdpacket0(R300_RB3D_DEPTHOFFSET, 2);
 	ALLOC_STATE( unk4F28, always, 2, "unk4F28", 0 );
diff --git a/src/mesa/drivers/dri/r300/r300_context.h b/src/mesa/drivers/dri/r300/r300_context.h
index 48b50bca65..a339b2f9c1 100644
--- a/src/mesa/drivers/dri/r300/r300_context.h
+++ b/src/mesa/drivers/dri/r300/r300_context.h
@@ -60,7 +60,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #endif
 
 //#define OPTIMIZE_ELTS
-#define CB_DPATH
 
 struct r300_context;
 typedef struct r300_context r300ContextRec;
diff --git a/src/mesa/drivers/dri/r300/r300_fragprog.c b/src/mesa/drivers/dri/r300/r300_fragprog.c
index 8e45bd5403..a1c634a54d 100644
--- a/src/mesa/drivers/dri/r300/r300_fragprog.c
+++ b/src/mesa/drivers/dri/r300/r300_fragprog.c
@@ -1214,8 +1214,8 @@ static void make_sin_const(struct r300_fragment_program *rp)
 	    cnstv[3] = 0.2225;      // weight
 	    rp->const_sin[0] = emit_const4fv(rp, cnstv);
 
-	    cnstv[0] = 0.5;
-	    cnstv[1] = -1.5;
+	    cnstv[0] = 0.75;
+	    cnstv[1] = 0.0;
 	    cnstv[2] = 0.159154943; // 1/(2*PI)
 	    cnstv[3] = 6.283185307; // 2*PI
 	    rp->const_sin[1] = emit_const4fv(rp, cnstv);
@@ -1227,7 +1227,7 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 	struct gl_fragment_program *mp = &rp->mesa_program;
 	const struct prog_instruction *inst = mp->Base.Instructions;
 	struct prog_instruction *fpi;
-	GLuint src[3], dest, temp;
+	GLuint src[3], dest, temp[2];
 	GLuint cnst;
 	int flags, mask = 0;
 	GLfloat cnstv[4] = {0.0, 0.0, 0.0, 0.0};
@@ -1277,70 +1277,63 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			/*
 			 * cos using a parabola (see SIN):
 			 * cos(x):
-			 *   x += PI/2
-			 *   x = (x/(2*PI))+0.5
+			 *   x = (x/(2*PI))+0.75
 			 *   x = frac(x)
 			 *   x = (x*2*PI)-PI
 			 *   result = sin(x)
 			 */
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			make_sin_const(rp);
 			src[0] = t_scalar_src(rp, fpi->SrcReg[0]);
 
 			/* add 0.5*PI and do range reduction */
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
-				   swizzle(rp->const_sin[0], Z, Z, Z, Z), //PI
-				   pfs_half,
-				   swizzle(keep(src[0]), X, X, X, X),
-				   0);
-
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X,
+				   swizzle(src[0], X, X, X, X),
 				   swizzle(rp->const_sin[1], Z, Z, Z, Z),
-				   pfs_half,
+				   swizzle(rp->const_sin[1], X, X, X, X),
 				   0);
 
-			emit_arith(rp, PFS_OP_FRC, temp, WRITEMASK_X,
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_FRC, temp[0], WRITEMASK_X,
+				   swizzle(temp[0], X, X, X, X),
 				   undef,
 				   undef,
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_Z,
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_Z,
+				   swizzle(temp[0], X, X, X, X),
 				   swizzle(rp->const_sin[1], W, W, W, W), //2*PI
 				   negate(swizzle(rp->const_sin[0], Z, Z, Z, Z)), //-PI
 				   0);
 
 			/* SIN */
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X | WRITEMASK_Y,
-				   swizzle(temp, Z, Z, Z, Z),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X | WRITEMASK_Y,
+				   swizzle(temp[0], Z, Z, Z, Z),
 				   rp->const_sin[0],
 				   pfs_zero,
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
-				   swizzle(temp, Y, Y, Y, Y),
-				   absolute(swizzle(temp, Z, Z, Z, Z)),
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X,
+				   swizzle(temp[0], Y, Y, Y, Y),
+				   absolute(swizzle(temp[0], Z, Z, Z, Z)),
+				   swizzle(temp[0], X, X, X, X),
 				   0);
 			
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_Y,
-				   swizzle(temp, X, X, X, X),
-				   absolute(swizzle(temp, X, X, X, X)),
-				   negate(swizzle(temp, X, X, X, X)),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_Y,
+				   swizzle(temp[0], X, X, X, X),
+				   absolute(swizzle(temp[0], X, X, X, X)),
+				   negate(swizzle(temp[0], X, X, X, X)),
 				   0);
 
 
 	    		emit_arith(rp, PFS_OP_MAD, dest, mask,
-				   swizzle(temp, Y, Y, Y, Y),
+				   swizzle(temp[0], Y, Y, Y, Y),
 				   swizzle(rp->const_sin[0], W, W, W, W),
-				   swizzle(temp, X, X, X, X),
+				   swizzle(temp[0], X, X, X, X),
 				   flags);
 
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_DP3:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
@@ -1363,15 +1356,15 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			 * DP4 dest, temp, src1
 			 */
 #if 0
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			src[0].s_swz = SWIZZLE_ONE;
-			emit_arith(rp, PFS_OP_MAD, temp, mask,
+			emit_arith(rp, PFS_OP_MAD, temp[0], mask,
 				   src[0], pfs_one, pfs_zero,
 				   0);
 			emit_arith(rp, PFS_OP_DP4, dest, mask,
-				   temp, src[1], undef,
+				   temp[0], src[1], undef,
 				   flags);	
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 #else
 			emit_arith(rp, PFS_OP_DP4, dest, mask,
 				   swizzle(src[0], X, Y, Z, ONE), src[1],
@@ -1408,17 +1401,17 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			break;
 		case OPCODE_FLR:		
 			src[0] = t_src(rp, fpi->SrcReg[0]);
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			/* FRC temp, src0
 			 * MAD dest, src0, 1.0, -temp
 			 */
-			emit_arith(rp, PFS_OP_FRC, temp, mask,
+			emit_arith(rp, PFS_OP_FRC, temp[0], mask,
 				   keep(src[0]), undef, undef,
 				   0);
 			emit_arith(rp, PFS_OP_MAD, dest, mask,
-				   src[0], pfs_one, negate(temp),
+				   src[0], pfs_one, negate(temp[0]),
 				   flags);
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_FRC:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
@@ -1450,44 +1443,44 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			 */
 			cnstv[0] = cnstv[1] = cnstv[2] = cnstv[3] = 0.50001;
 			src[0] = t_src(rp, fpi->SrcReg[0]);
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			cnst = emit_const4fv(rp, cnstv);
-			emit_arith(rp, PFS_OP_CMP, temp,
+			emit_arith(rp, PFS_OP_CMP, temp[0],
 				   WRITEMASK_X | WRITEMASK_Y,
 				   src[0], pfs_zero, src[0], flags);
-			emit_arith(rp, PFS_OP_MIN, temp, WRITEMASK_Z,
+			emit_arith(rp, PFS_OP_MIN, temp[0], WRITEMASK_Z,
 				   swizzle(keep(src[0]), W, W, W, W),
 				   cnst, undef, flags);
-			emit_arith(rp, PFS_OP_LG2, temp, WRITEMASK_W,
-				   swizzle(temp, Y, Y, Y, Y),
+			emit_arith(rp, PFS_OP_LG2, temp[0], WRITEMASK_W,
+				   swizzle(temp[0], Y, Y, Y, Y),
 				   undef, undef, flags);
-			emit_arith(rp, PFS_OP_MAX, temp, WRITEMASK_Z,
-				   temp, negate(cnst), undef, flags);
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_W,
-				   temp, swizzle(temp, Z, Z, Z, Z),
+			emit_arith(rp, PFS_OP_MAX, temp[0], WRITEMASK_Z,
+				   temp[0], negate(cnst), undef, flags);
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_W,
+				   temp[0], swizzle(temp[0], Z, Z, Z, Z),
 				   pfs_zero, flags);
-			emit_arith(rp, PFS_OP_EX2, temp, WRITEMASK_W,
-				   temp, undef, undef, flags);
+			emit_arith(rp, PFS_OP_EX2, temp[0], WRITEMASK_W,
+				   temp[0], undef, undef, flags);
 			emit_arith(rp, PFS_OP_MAD, dest, WRITEMASK_Y,
-				   swizzle(keep(temp), X, X, X, X),
+				   swizzle(keep(temp[0]), X, X, X, X),
 				   pfs_one, pfs_zero, flags);
 #if 0
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
-				   temp, pfs_one, pfs_half, flags);
-			emit_arith(rp, PFS_OP_CMPH, temp, WRITEMASK_Z,
-				   swizzle(keep(temp), W, W, W, W),
-				   pfs_zero, swizzle(keep(temp), X, X, X, X),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X,
+				   temp[0], pfs_one, pfs_half, flags);
+			emit_arith(rp, PFS_OP_CMPH, temp[0], WRITEMASK_Z,
+				   swizzle(keep(temp[0]), W, W, W, W),
+				   pfs_zero, swizzle(keep(temp[0]), X, X, X, X),
 				   flags);
 #else
-			emit_arith(rp, PFS_OP_CMP, temp, WRITEMASK_Z,
+			emit_arith(rp, PFS_OP_CMP, temp[0], WRITEMASK_Z,
 				   pfs_zero,
-				   swizzle(keep(temp), W, W, W, W),
-				   negate(swizzle(keep(temp), X, X, X, X)),
+				   swizzle(keep(temp[0]), W, W, W, W),
+				   negate(swizzle(keep(temp[0]), X, X, X, X)),
 				   flags);
 #endif
 			emit_arith(rp, PFS_OP_CMP, dest, WRITEMASK_Z,
-				   pfs_zero, temp,
-				   negate(swizzle(keep(temp), Y, Y, Y, Y)),
+				   pfs_zero, temp[0],
+				   negate(swizzle(keep(temp[0]), Y, Y, Y, Y)),
 				   flags);
 			emit_arith(rp, PFS_OP_MAD, dest,
 				   WRITEMASK_X | WRITEMASK_W,
@@ -1495,7 +1488,7 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 				   pfs_one,
 				   pfs_zero,
 				   flags);
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_LRP:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
@@ -1506,14 +1499,14 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			 *     MAD temp, -tmp0, tmp2, tmp2
 			 *     MAD result, tmp0, tmp1, temp
 			 */
-			temp = get_temp_reg(rp);
-			emit_arith(rp, PFS_OP_MAD, temp, mask,
+			temp[0] = get_temp_reg(rp);
+			emit_arith(rp, PFS_OP_MAD, temp[0], mask,
 				   negate(keep(src[0])), keep(src[2]), src[2],
 				   0);
 			emit_arith(rp, PFS_OP_MAD, dest, mask,
-				   src[0], src[1], temp,
+				   src[0], src[1], temp[0],
 				   flags);
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;			
 		case OPCODE_MAD:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
@@ -1554,17 +1547,17 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 		case OPCODE_POW:
 			src[0] = t_scalar_src(rp, fpi->SrcReg[0]);
 			src[1] = t_scalar_src(rp, fpi->SrcReg[1]);
-			temp = get_temp_reg(rp);	
-			emit_arith(rp, PFS_OP_LG2, temp, WRITEMASK_W,
+			temp[0] = get_temp_reg(rp);	
+			emit_arith(rp, PFS_OP_LG2, temp[0], WRITEMASK_W,
 				   src[0], undef, undef,
 				   0);
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_W,
-				   temp, src[1], pfs_zero,
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_W,
+				   temp[0], src[1], pfs_zero,
 				   0);
 			emit_arith(rp, PFS_OP_EX2, dest, fpi->DstReg.WriteMask,
-				   temp, undef, undef,
+				   temp[0], undef, undef,
 				   0);
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_RCP:
 			src[0] = t_scalar_src(rp, fpi->SrcReg[0]);
@@ -1580,107 +1573,91 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			break;
 		case OPCODE_SCS:
 			/*
-			 * cos using a parabola (see SIN):
-			 * cos(x):
-			 *   x += PI/2
-			 *   x = (x/(2*PI))+0.5
-			 *   x = frac(x)
-			 *   x = (x*2*PI)-PI
-			 *   result = sin(x)
+			 * scs using a parabola :
+			 * scs(x):
+			 *   result.x = sin(-abs(x)+0.5*PI)  (cos)
+			 *   result.y = sin(x)               (sin)
+			 *
 			 */
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
+			temp[1] = get_temp_reg(rp);
 			make_sin_const(rp);
 			src[0] = t_scalar_src(rp, fpi->SrcReg[0]);
 
-			/* add 0.5*PI and do range reduction */
+			/* x = -abs(x)+0.5*PI */
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_Z,
+				   swizzle(rp->const_sin[0], Z, Z, Z, Z), //PI
+				   pfs_half,
+				   negate(abs(swizzle(keep(src[0]), X, X, X, X))),
+				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X|WRITEMASK_Y,
-				   swizzle(rp->const_sin[0], Z, Z, Z, Z),
-				   rp->const_sin[1],
+			/* C*x (sin) */
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_W,
+				   swizzle(rp->const_sin[0], Y, Y, Y, Y),
 				   swizzle(keep(src[0]), X, X, X, X),
+				   pfs_zero,
 				   0);
 
-			emit_arith(rp, PFS_OP_CMP, temp, WRITEMASK_W,
-				   swizzle(rp->const_sin[0], Z, Z, Z, Z),
-				   negate(pfs_half),
-				   swizzle(keep(src[0]), X, X, X, X),
+			/* B*x, C*x (cos) */
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X | WRITEMASK_Y,
+			           swizzle(temp[0], Z, Z, Z, Z),
+				   rp->const_sin[0],
+			           pfs_zero,
 				   0);
 
-			emit_arith(rp, PFS_OP_CMP, temp, WRITEMASK_Z,
-				   swizzle(temp, X, X, X, X),
-				   swizzle(temp, Y, Y, Y, Y),
-				   swizzle(temp, W, W, W, W),
+			/* B*x (sin) */
+			emit_arith(rp, PFS_OP_MAD, temp[1], WRITEMASK_W,
+				   swizzle(rp->const_sin[0], X, X, X, X),
+				   keep(src[0]),
+				   pfs_zero,
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X | WRITEMASK_Y,
-			           swizzle(temp, Z, Z, Z, Z),
-				   rp->const_sin[0],
-			           pfs_zero,
+			/* y = B*x + C*x*abs(x) (sin)*/
+		    	emit_arith(rp, PFS_OP_MAD, temp[1], WRITEMASK_Z,
+				   absolute(src[0]),
+				   swizzle(temp[0], W, W, W, W),
+				   swizzle(temp[1], W, W, W, W),
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_W,
-				   swizzle(temp, Y, Y, Y, Y),
-				   absolute(swizzle(temp, Z, Z, Z, Z)),
-				   swizzle(temp, X, X, X, X),
+			/* y = B*x + C*x*abs(x) (cos)*/
+			emit_arith(rp, PFS_OP_MAD, temp[1], WRITEMASK_W,
+				   swizzle(temp[0], Y, Y, Y, Y),
+				   absolute(swizzle(temp[0], Z, Z, Z, Z)),
+				   swizzle(temp[0], X, X, X, X),
 				   0);
 
-			if(mask & WRITEMASK_Y)
-			{
-			    emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X | WRITEMASK_Y,
-				       swizzle(keep(src[0]), X, X, X, X),
-				       rp->const_sin[0],
-				       pfs_zero,
-				       0);
-
-			    emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
-				       swizzle(temp, Y, Y, Y, Y),
-				       absolute(swizzle(keep(src[0]), X, X, X, X)),
-				       swizzle(temp, X, X, X, X),
-				       0);
-			}
+			/* y*abs(y) - y (cos), y*abs(y) - y (sin) */
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X | WRITEMASK_Y,
+			           swizzle(temp[1], W, Z, Y, X),
+				   absolute(swizzle(temp[1], W, Z, Y, X)),
+				   negate(swizzle(temp[1], W, Z, Y, X)),
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_Z,
-				   swizzle(temp, W, W, W, W),
-				   absolute(swizzle(temp, W, W, W, W)),
-				   negate(swizzle(temp, W, W, W, W)),
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, dest, WRITEMASK_X,
-				   swizzle(temp, Z, Z, Z, Z),
+			/* dest.xy = mad(temp.xy, P, temp2.wz) */
+			emit_arith(rp, PFS_OP_MAD, dest, mask & (WRITEMASK_X | WRITEMASK_Y),
+				   temp[0],
 				   swizzle(rp->const_sin[0], W, W, W, W),
-				   swizzle(temp, W, W, W, W),
+				   swizzle(temp[1], W, Z, Y, X),
 				   flags);
 
-			if(mask & WRITEMASK_Y)
-			{
-			    emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_W,
-				       swizzle(temp, X, X, X, X),
-				       absolute(swizzle(temp, X, X, X, X)),
-				       negate(swizzle(temp, X, X, X, X)),
-				       0);
-
-	    		    emit_arith(rp, PFS_OP_MAD, dest, WRITEMASK_Y,
-				       swizzle(temp, W, W, W, W),
-				       swizzle(rp->const_sin[0], W, W, W, W),
-				       swizzle(temp, X, X, X, X),
-				       flags);
-			}
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
+			free_temp(rp, temp[1]);
 			break;
 		case OPCODE_SGE:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
 			src[1] = t_src(rp, fpi->SrcReg[1]);
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			/* temp = src0 - src1
 			 * dest.c = (temp.c < 0.0) ? 0 : 1
 			 */
-			emit_arith(rp, PFS_OP_MAD, temp, mask,
+			emit_arith(rp, PFS_OP_MAD, temp[0], mask,
 				   src[0], pfs_one, negate(src[1]),
 				   0);
 			emit_arith(rp, PFS_OP_CMP, dest, mask,
-				   pfs_one, pfs_zero, temp,
+				   pfs_one, pfs_zero, temp[0],
 				   0);
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_SIN:
 			/*
@@ -1690,73 +1667,74 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			 * itself squared.
 			 */
 
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			make_sin_const(rp);
 			src[0] = t_scalar_src(rp, fpi->SrcReg[0]);
 
+
 			/* do range reduction */
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X,
 				   swizzle(keep(src[0]), X, X, X, X),
 				   swizzle(rp->const_sin[1], Z, Z, Z, Z),
 				   pfs_half,
 				   0);
 
-			emit_arith(rp, PFS_OP_FRC, temp, WRITEMASK_X,
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_FRC, temp[0], WRITEMASK_X,
+				   swizzle(temp[0], X, X, X, X),
 				   undef,
 				   undef,
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_Z,
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_Z,
+				   swizzle(temp[0], X, X, X, X),
 				   swizzle(rp->const_sin[1], W, W, W, W), //2*PI
 				   negate(swizzle(rp->const_sin[0], Z, Z, Z, Z)), //PI
 				   0);
 
 			/* SIN */
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X | WRITEMASK_Y,
-				   swizzle(temp, Z, Z, Z, Z),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X | WRITEMASK_Y,
+				   swizzle(temp[0], Z, Z, Z, Z),
 				   rp->const_sin[0],
 				   pfs_zero,
 				   0);
 
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_X,
-				   swizzle(temp, Y, Y, Y, Y),
-				   absolute(swizzle(temp, Z, Z, Z, Z)),
-				   swizzle(temp, X, X, X, X),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_X,
+				   swizzle(temp[0], Y, Y, Y, Y),
+				   absolute(swizzle(temp[0], Z, Z, Z, Z)),
+				   swizzle(temp[0], X, X, X, X),
 				   0);
 			
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_Y,
-				   swizzle(temp, X, X, X, X),
-				   absolute(swizzle(temp, X, X, X, X)),
-				   negate(swizzle(temp, X, X, X, X)),
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_Y,
+				   swizzle(temp[0], X, X, X, X),
+				   absolute(swizzle(temp[0], X, X, X, X)),
+				   negate(swizzle(temp[0], X, X, X, X)),
 				   0);
 
 
 	    		emit_arith(rp, PFS_OP_MAD, dest, mask,
-				   swizzle(temp, Y, Y, Y, Y),
+				   swizzle(temp[0], Y, Y, Y, Y),
 				   swizzle(rp->const_sin[0], W, W, W, W),
-				   swizzle(temp, X, X, X, X),
+				   swizzle(temp[0], X, X, X, X),
 				   flags);
 
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_SLT:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
 			src[1] = t_src(rp, fpi->SrcReg[1]);
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			/* temp = src0 - src1
 			 * dest.c = (temp.c < 0.0) ? 1 : 0
 			 */
-			emit_arith(rp, PFS_OP_MAD, temp, mask,
+			emit_arith(rp, PFS_OP_MAD, temp[0], mask,
 				   src[0], pfs_one, negate(src[1]),
 				   0);
 			emit_arith(rp, PFS_OP_CMP, dest, mask,
-				   pfs_zero, pfs_one, temp,
+				   pfs_zero, pfs_one, temp[0],
 				   0);
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		case OPCODE_SUB:
 			src[0] = t_src(rp, fpi->SrcReg[0]);
@@ -1777,9 +1755,9 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 		case OPCODE_XPD: {
 			src[0] = t_src(rp, fpi->SrcReg[0]);
 			src[1] = t_src(rp, fpi->SrcReg[1]);
-			temp = get_temp_reg(rp);
+			temp[0] = get_temp_reg(rp);
 			/* temp = src0.zxy * src1.yzx */
-			emit_arith(rp, PFS_OP_MAD, temp, WRITEMASK_XYZ,
+			emit_arith(rp, PFS_OP_MAD, temp[0], WRITEMASK_XYZ,
 				   swizzle(keep(src[0]), Z, X, Y, W),
 				   swizzle(keep(src[1]), Y, Z, X, W),
 				   pfs_zero,
@@ -1790,10 +1768,10 @@ static GLboolean parse_program(struct r300_fragment_program *rp)
 			emit_arith(rp, PFS_OP_MAD, dest, mask & WRITEMASK_XYZ,
 				   swizzle(src[0], Y, Z, X, W),
 				   swizzle(src[1], Z, X, Y, W),
-				   negate(temp),
+				   negate(temp[0]),
 				   flags);
 			/* cleanup */
-			free_temp(rp, temp);
+			free_temp(rp, temp[0]);
 			break;
 		}
 		default:
diff --git a/src/mesa/drivers/dri/r300/r300_ioctl.c b/src/mesa/drivers/dri/r300/r300_ioctl.c
index 0774a2af6b..6fa34ee482 100644
--- a/src/mesa/drivers/dri/r300/r300_ioctl.c
+++ b/src/mesa/drivers/dri/r300/r300_ioctl.c
@@ -70,10 +70,6 @@ static void r300ClearBuffer(r300ContextPtr r300, int flags, int buffer)
 	drm_radeon_cmd_header_t *cmd = NULL;
 	r300ContextPtr rmesa=r300;
 
-#ifndef CB_DPATH
-	int i;
-#endif
-	
 	if (RADEON_DEBUG & DEBUG_IOCTL)
 		fprintf(stderr, "%s: %s buffer (%i,%i %ix%i)\n",
 			__FUNCTION__, buffer ? "back" : "front",
@@ -89,168 +85,8 @@ static void r300ClearBuffer(r300ContextPtr r300, int flags, int buffer)
 
 	cboffset += r300->radeon.radeonScreen->fbLocation;
 
-#ifndef CB_DPATH
-	R300_STATECHANGE(r300, vir[0]);
-	((drm_r300_cmd_header_t*)r300->hw.vir[0].cmd)->packet0.count = 1;
-	r300->hw.vir[0].cmd[1] = 0x21030003;
-
-	R300_STATECHANGE(r300, vir[1]);
-	((drm_r300_cmd_header_t*)r300->hw.vir[1].cmd)->packet0.count = 1;
-	r300->hw.vir[1].cmd[1] = 0xF688F688;
-
-	R300_STATECHANGE(r300, vic);
-	r300->hw.vic.cmd[R300_VIC_CNTL_0] = 0x00000001;
-	r300->hw.vic.cmd[R300_VIC_CNTL_1] = 0x00000405;
-	
-	R300_STATECHANGE(r300, vof);
-	r300->hw.vof.cmd[R300_VOF_CNTL_0] = R300_VAP_OUTPUT_VTX_FMT_0__POS_PRESENT
-				| R300_VAP_OUTPUT_VTX_FMT_0__COLOR_PRESENT;
-	r300->hw.vof.cmd[R300_VOF_CNTL_1] = 0; /* no textures */
-	
-	R300_STATECHANGE(r300, txe);
-	r300->hw.txe.cmd[R300_TXE_ENABLE] = 0;
-	
-	R300_STATECHANGE(r300, vpt);
-	r300->hw.vpt.cmd[R300_VPT_XSCALE] = r300PackFloat32(1.0);
-	r300->hw.vpt.cmd[R300_VPT_XOFFSET] = r300PackFloat32(dPriv->x);
-	r300->hw.vpt.cmd[R300_VPT_YSCALE] = r300PackFloat32(1.0);
-	r300->hw.vpt.cmd[R300_VPT_YOFFSET] = r300PackFloat32(dPriv->y);
-	r300->hw.vpt.cmd[R300_VPT_ZSCALE] = r300PackFloat32(1.0);
-	r300->hw.vpt.cmd[R300_VPT_ZOFFSET] = r300PackFloat32(0.0);
-
-	R300_STATECHANGE(r300, at);
-	r300->hw.at.cmd[R300_AT_ALPHA_TEST] = 0;
-	
-	R300_STATECHANGE(r300, bld);
-	r300->hw.bld.cmd[R300_BLD_CBLEND] = 0;
-	r300->hw.bld.cmd[R300_BLD_ABLEND] = 0;
-	
-	if (r300->radeon.radeonScreen->cpp == 4)
-		cbpitch |= R300_COLOR_FORMAT_ARGB8888;
-	else
-		cbpitch |= R300_COLOR_FORMAT_RGB565;
-	
-	if (r300->radeon.sarea->tiling_enabled)
-		cbpitch |= R300_COLOR_TILE_ENABLE;
-	
-	R300_STATECHANGE(r300, cb);
-	r300->hw.cb.cmd[R300_CB_OFFSET] = cboffset;
-	r300->hw.cb.cmd[R300_CB_PITCH] = cbpitch;
-
-	R300_STATECHANGE(r300, unk221C);
-	r300->hw.unk221C.cmd[1] = R300_221C_CLEAR;
-
-	R300_STATECHANGE(r300, ps);
-	r300->hw.ps.cmd[R300_PS_POINTSIZE] =
-		((dPriv->w * 6) << R300_POINTSIZE_X_SHIFT) |
-		((dPriv->h * 6) << R300_POINTSIZE_Y_SHIFT);
-
-	R300_STATECHANGE(r300, ri);
-	for(i = 1; i <= 8; ++i)
-		r300->hw.ri.cmd[i] = R300_RS_INTERP_USED;
-
-	R300_STATECHANGE(r300, rc);
-	/* The second constant is needed to get glxgears display anything .. */
-	r300->hw.rc.cmd[1] = (1 << R300_RS_CNTL_CI_CNT_SHIFT) | R300_RS_CNTL_0_UNKNOWN_18;
-	r300->hw.rc.cmd[2] = 0;
-	
-	R300_STATECHANGE(r300, rr);
-	((drm_r300_cmd_header_t*)r300->hw.rr.cmd)->packet0.count = 1;
-	r300->hw.rr.cmd[1] = 0x00004000;
-
-	R300_STATECHANGE(r300, cmk);
-	if (flags & CLEARBUFFER_COLOR) {
-		r300->hw.cmk.cmd[R300_CMK_COLORMASK] =
-			(ctx->Color.ColorMask[BCOMP] ? R300_COLORMASK0_B : 0) |
-			(ctx->Color.ColorMask[GCOMP] ? R300_COLORMASK0_G : 0) |
-			(ctx->Color.ColorMask[RCOMP] ? R300_COLORMASK0_R : 0) |
-			(ctx->Color.ColorMask[ACOMP] ? R300_COLORMASK0_A : 0);
-	} else {
-		r300->hw.cmk.cmd[R300_CMK_COLORMASK] = 0;
-	}
-
-	R300_STATECHANGE(r300, fp);
-	r300->hw.fp.cmd[R300_FP_CNTL0] = 0; /* 1 pass, no textures */
-	r300->hw.fp.cmd[R300_FP_CNTL1] = 0; /* no temporaries */
-	r300->hw.fp.cmd[R300_FP_CNTL2] = 0; /* no offset, one ALU instr */
-	r300->hw.fp.cmd[R300_FP_NODE0] = 0;
-	r300->hw.fp.cmd[R300_FP_NODE1] = 0;
-	r300->hw.fp.cmd[R300_FP_NODE2] = 0;
-	r300->hw.fp.cmd[R300_FP_NODE3] = R300_PFS_NODE_OUTPUT_COLOR;
-
-	R300_STATECHANGE(r300, fpi[0]);
-	R300_STATECHANGE(r300, fpi[1]);
-	R300_STATECHANGE(r300, fpi[2]);
-	R300_STATECHANGE(r300, fpi[3]);
-	((drm_r300_cmd_header_t*)r300->hw.fpi[0].cmd)->packet0.count = 1;
-	((drm_r300_cmd_header_t*)r300->hw.fpi[1].cmd)->packet0.count = 1;
-	((drm_r300_cmd_header_t*)r300->hw.fpi[2].cmd)->packet0.count = 1;
-	((drm_r300_cmd_header_t*)r300->hw.fpi[3].cmd)->packet0.count = 1;
-
-	/* MOV o0, t0 */
-	r300->hw.fpi[0].cmd[1] = FP_INSTRC(MAD, FP_ARGC(SRC0C_XYZ), FP_ARGC(ONE), FP_ARGC(ZERO));
-	r300->hw.fpi[1].cmd[1] = FP_SELC(0,NO,XYZ,FP_TMP(0),0,0);
-	r300->hw.fpi[2].cmd[1] = FP_INSTRA(MAD, FP_ARGA(SRC0A), FP_ARGA(ONE), FP_ARGA(ZERO));
-	r300->hw.fpi[3].cmd[1] = FP_SELA(0,NO,W,FP_TMP(0),0,0);
-
-	R300_STATECHANGE(r300, pvs);
-	r300->hw.pvs.cmd[R300_PVS_CNTL_1] =
-		(0 << R300_PVS_CNTL_1_PROGRAM_START_SHIFT) |
-		(0 << R300_PVS_CNTL_1_POS_END_SHIFT) |
-		(1 << R300_PVS_CNTL_1_PROGRAM_END_SHIFT);
-	r300->hw.pvs.cmd[R300_PVS_CNTL_2] = 0; /* no parameters */
-	r300->hw.pvs.cmd[R300_PVS_CNTL_3] =
-		(1 << R300_PVS_CNTL_3_PROGRAM_UNKNOWN_SHIFT);
-
-	R300_STATECHANGE(r300, vpi);
-	((drm_r300_cmd_header_t*)r300->hw.vpi.cmd)->packet0.count = 8;
-
-	/* MOV o0, i0; */
-	r300->hw.vpi.cmd[1] = VP_OUT(ADD,OUT,0,XYZW);
-	r300->hw.vpi.cmd[2] = VP_IN(IN,0);
-	r300->hw.vpi.cmd[3] = VP_ZERO();
-	r300->hw.vpi.cmd[4] = 0;
-
-	/* MOV o1, i1; */
-	r300->hw.vpi.cmd[5] = VP_OUT(ADD,OUT,1,XYZW);
-	r300->hw.vpi.cmd[6] = VP_IN(IN,1);
-	r300->hw.vpi.cmd[7] = VP_ZERO();
-	r300->hw.vpi.cmd[8] = 0;
-
-	R300_STATECHANGE(r300, zs);
-	r300->hw.zs.cmd[R300_ZS_CNTL_0] = 0;
-	r300->hw.zs.cmd[R300_ZS_CNTL_1] = 0;
-	if (flags & CLEARBUFFER_DEPTH) {
-		r300->hw.zs.cmd[R300_ZS_CNTL_0] |= R300_RB3D_Z_WRITE_ONLY;
-		r300->hw.zs.cmd[R300_ZS_CNTL_1] |= (R300_ZS_ALWAYS<<R300_RB3D_ZS1_DEPTH_FUNC_SHIFT);
-	} else {
-		r300->hw.zs.cmd[R300_ZS_CNTL_0] |= R300_RB3D_Z_DISABLED_1; // disable
-	}
-	
-	R300_STATECHANGE(r300, zs);
-	if (flags & CLEARBUFFER_STENCIL) {
-		r300->hw.zs.cmd[R300_ZS_CNTL_0] |= R300_RB3D_STENCIL_ENABLE;
-		r300->hw.zs.cmd[R300_ZS_CNTL_1] |= 
-		    (R300_ZS_ALWAYS<<R300_RB3D_ZS1_FRONT_FUNC_SHIFT) | 
-		    (R300_ZS_REPLACE<<R300_RB3D_ZS1_FRONT_FAIL_OP_SHIFT) |
-		    (R300_ZS_REPLACE<<R300_RB3D_ZS1_FRONT_ZPASS_OP_SHIFT) |
-		    (R300_ZS_REPLACE<<R300_RB3D_ZS1_FRONT_ZFAIL_OP_SHIFT) |
-		    (R300_ZS_ALWAYS<<R300_RB3D_ZS1_BACK_FUNC_SHIFT) |
-		    (R300_ZS_REPLACE<<R300_RB3D_ZS1_BACK_FAIL_OP_SHIFT) |
-		    (R300_ZS_REPLACE<<R300_RB3D_ZS1_BACK_ZPASS_OP_SHIFT) |
-		    (R300_ZS_REPLACE<<R300_RB3D_ZS1_BACK_ZFAIL_OP_SHIFT) ;
-		r300->hw.zs.cmd[R300_ZS_CNTL_2] = r300->state.stencil.clear;
-	}
-			
-	/* Make sure we have enough space */
-	r300EnsureCmdBufSpace(r300, r300->hw.max_state_size + 9+8, __FUNCTION__);
-
-	r300EmitState(r300);
-#else
-#if 1
 	cp_wait(r300, R300_WAIT_3D | R300_WAIT_3D_CLEAN);
 	end_3d(rmesa);
-#endif
 	
 	R300_STATECHANGE(r300, cb);
 	reg_start(R300_RB3D_COLOROFFSET0, 0);
@@ -313,8 +149,6 @@ static void r300ClearBuffer(r300ContextPtr r300, int flags, int buffer)
 	e32(r300->state.stencil.clear);
 	}
 	
-#endif
-
 	cmd2 = (drm_r300_cmd_header_t*)r300AllocCmdBuf(r300, 9, __FUNCTION__);
 	cmd2[0].packet3.cmd_type = R300_CMD_PACKET3;
 	cmd2[0].packet3.packet = R300_CMD_PACKET3_CLEAR;
@@ -327,7 +161,6 @@ static void r300ClearBuffer(r300ContextPtr r300, int flags, int buffer)
 	cmd2[7].u = r300PackFloat32(ctx->Color.ClearColor[2]);
 	cmd2[8].u = r300PackFloat32(ctx->Color.ClearColor[3]);
 
-#if 1
 	reg_start(R300_RB3D_DSTCACHE_CTLSTAT,0);
 	e32(0x0000000a);
 	  
@@ -335,10 +168,8 @@ static void r300ClearBuffer(r300ContextPtr r300, int flags, int buffer)
 	reg_start(0x4f18,0);
 	e32(0x00000003);
 	cp_wait(rmesa, R300_WAIT_3D | R300_WAIT_3D_CLEAN);
-#endif
 }
 
-#ifdef CB_DPATH
 static void r300EmitClearState(GLcontext * ctx)
 {
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
@@ -397,7 +228,7 @@ static void r300EmitClearState(GLcontext * ctx)
 	e32(0);
 	
 	R300_STATECHANGE(r300, unk221C);
-	reg_start(0x221C, 0);
+	reg_start(R300_VAP_UNKNOWN_221C, 0);
 	e32(R300_221C_CLEAR);
 	
 	R300_STATECHANGE(r300, ps);
@@ -472,7 +303,6 @@ static void r300EmitClearState(GLcontext * ctx)
 	/*reg_start(0x4500,0);
 	e32(2560-1);*/
 }
-#endif
 
 /**
  * Buffer clear
@@ -524,12 +354,10 @@ static void r300Clear(GLcontext * ctx, GLbitfield mask)
 
 	swapped = r300->radeon.doPageFlip && (r300->radeon.sarea->pfCurrentPage == 1);
 
-#ifdef CB_DPATH
 	/* Make sure it fits there. */
 	r300EnsureCmdBufSpace(r300, 421*3, __FUNCTION__);
 	if(flags || bits)
 		r300EmitClearState(ctx);
-#endif
 
 	if (flags & BUFFER_BIT_FRONT_LEFT) {
 		r300ClearBuffer(r300, bits | CLEARBUFFER_COLOR, swapped);
@@ -544,15 +372,6 @@ static void r300Clear(GLcontext * ctx, GLbitfield mask)
 	if (bits)
 		r300ClearBuffer(r300, bits, 0);
 
-#ifndef CB_DPATH
-	/* Recalculate the hardware state. This could be done more efficiently,
-	 * but do keep it like this for now.
-	 */
-	r300ResetHwState(r300);
-	
-	/* r300ClearBuffer has trampled all over the hardware state.. */
-	r300->hw.all_dirty=GL_TRUE;
-#endif
 }
 
 
diff --git a/src/mesa/drivers/dri/r300/r300_reg.h b/src/mesa/drivers/dri/r300/r300_reg.h
index 9f636ec1d2..69bc994cf6 100644
--- a/src/mesa/drivers/dri/r300/r300_reg.h
+++ b/src/mesa/drivers/dri/r300/r300_reg.h
@@ -63,6 +63,12 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define R300_SE_VPORT_ZOFFSET               0x1DAC
 
 
+/*
+ * Vertex Array Processing (VAP) Control
+ * Stolen from r200 code from Christoph Brill (It's a guess!)
+ */
+#define R300_VAP_CNTL	0x2080
+
 /* This register is written directly and also starts data section
  * in many 3d CP_PACKET3's
  */
@@ -135,7 +141,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 /* gap */
 
-#define R300_VAP_CNTL                     0x2140
+#define R300_VAP_CNTL_STATUS              0x2140
 #	define R300_VC_NO_SWAP                  (0 << 0)
 #	define R300_VC_16BIT_SWAP               (1 << 0)
 #	define R300_VC_32BIT_SWAP               (2 << 0)
@@ -538,6 +544,9 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 /* Some sort of scale or clamp value for texcoordless textures. */
 #define R300_RE_UNK4238                       0x4238
 
+/* Something shade related */
+#define R300_RE_SHADE                         0x4274
+
 #define R300_RE_SHADE_MODEL                   0x4278
 #	define R300_RE_SHADE_MODEL_SMOOTH     0x3aaaa
 #	define R300_RE_SHADE_MODEL_FLAT       0x39595
@@ -1273,6 +1282,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #       define R300_BLEND_MASK                       (63)
 #       define R300_SRC_BLEND_SHIFT                  (16)
 #       define R300_DST_BLEND_SHIFT                  (24)
+#define R300_RB3D_BLEND_COLOR               0x4E10
 #define R300_RB3D_COLORMASK                 0x4E0C
 #       define R300_COLORMASK0_B                 (1<<0)
 #       define R300_COLORMASK0_G                 (1<<1)
@@ -1378,6 +1388,10 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 	/* 16 bit format or some aditional bit ? */
 #	define R300_DEPTH_FORMAT_UNK32          (32 << 0)
 
+#define R300_RB3D_EARLY_Z                           0x4F14
+#	define R300_EARLY_Z_DISABLE              (0 << 0)
+#	define R300_EARLY_Z_ENABLE               (1 << 0)
+
 /* gap */
 
 #define R300_RB3D_DEPTHOFFSET               0x4F20
diff --git a/src/mesa/drivers/dri/r300/r300_state.c b/src/mesa/drivers/dri/r300/r300_state.c
index 906dfceb48..0e33e51ef3 100644
--- a/src/mesa/drivers/dri/r300/r300_state.c
+++ b/src/mesa/drivers/dri/r300/r300_state.c
@@ -328,24 +328,24 @@ static void r300UpdateCulling(GLcontext* ctx)
 
 static void update_early_z(GLcontext *ctx)
 {
-	/* updates register 0x4f14 
-	   if depth test is not enabled it should be 0x00000000
-	   if depth is enabled and alpha not it should be 0x00000001
-	   if depth and alpha is enabled it should be 0x00000000
+	/* updates register R300_RB3D_EARLY_Z (0x4F14)
+	   if depth test is not enabled it should be R300_EARLY_Z_DISABLE
+	   if depth is enabled and alpha not it should be R300_EARLY_Z_ENABLE
+	   if depth and alpha is enabled it should be R300_EARLY_Z_DISABLE
 	*/
 	r300ContextPtr r300 = R300_CONTEXT(ctx);
 
 	R300_STATECHANGE(r300, unk4F10);
 	if (ctx->Color.AlphaEnabled && ctx->Color.AlphaFunc != GL_ALWAYS)
 		/* disable early Z */
-		r300->hw.unk4F10.cmd[2] = 0x00000000;
+		r300->hw.unk4F10.cmd[2] = R300_EARLY_Z_DISABLE;
 	else {
 		if (ctx->Depth.Test && ctx->Depth.Func != GL_NEVER)
 			/* enable early Z */
-			r300->hw.unk4F10.cmd[2] = 0x00000001;
+			r300->hw.unk4F10.cmd[2] = R300_EARLY_Z_ENABLE;
 		else
 			/* disable early Z */
-			r300->hw.unk4F10.cmd[2] = 0x00000000;
+			r300->hw.unk4F10.cmd[2] = R300_EARLY_Z_DISABLE;
 	}
 }
 
@@ -1768,9 +1768,7 @@ void r300UpdateShaderStates(r300ContextPtr rmesa)
 	GLcontext *ctx;
 	ctx = rmesa->radeon.glCtx;
 	
-#ifdef CB_DPATH
 	r300UpdateTextureState(ctx);
-#endif
 
 	r300SetupPixelShader(rmesa);
 	r300_setup_textures(ctx);
@@ -1895,10 +1893,6 @@ static void r300InvalidateState(GLcontext * ctx, GLuint new_state)
 
 	r300UpdateStateParameters(ctx, new_state);
 
-#ifndef CB_DPATH
-	/* Go inefficiency! */
-	r300ResetHwState(r300);
-#endif
 #ifdef HW_VBOS
 	if(new_state & _NEW_ARRAY)
 		r300->state.VB.lock_uptodate = GL_FALSE;
diff --git a/src/mesa/drivers/dri/r300/radeon_context.c b/src/mesa/drivers/dri/r300/radeon_context.c
index 3a6bde8fc3..d66f1dc49e 100644
--- a/src/mesa/drivers/dri/r300/radeon_context.c
+++ b/src/mesa/drivers/dri/r300/radeon_context.c
@@ -51,6 +51,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "radeon_macros.h"
 #include "radeon_reg.h"
 
+#include "radeon_state.h"
 #include "r300_state.h"
 
 #include "utils.h"
@@ -279,6 +280,8 @@ GLboolean radeonMakeCurrent(__DRIcontextPrivate * driContextPriv,
 
 			r300UpdateWindow(radeon->glCtx);
 			r300UpdateViewportOffset(radeon->glCtx);
+
+			radeonSetCliprects(radeon);
 		}
 
 		_mesa_make_current(radeon->glCtx,
diff --git a/src/mesa/drivers/dri/r300/radeon_lock.c b/src/mesa/drivers/dri/r300/radeon_lock.c
index a00da6c8bc..1a2dfca7b0 100644
--- a/src/mesa/drivers/dri/r300/radeon_lock.c
+++ b/src/mesa/drivers/dri/r300/radeon_lock.c
@@ -90,7 +90,6 @@ static void r300RegainedLock(radeonContextPtr radeon)
 #else
 		radeonUpdateScissor(radeon->glCtx);
 #endif
-		radeon->lastStamp = drawable->lastStamp;
 	}
 
 	if (sarea->ctx_owner != radeon->dri.hwContext) {
diff --git a/src/mesa/drivers/dri/r300/radeon_mm.c b/src/mesa/drivers/dri/r300/radeon_mm.c
index f86a1b4e72..32ed1f4393 100644
--- a/src/mesa/drivers/dri/r300/radeon_mm.c
+++ b/src/mesa/drivers/dri/r300/radeon_mm.c
@@ -283,7 +283,7 @@ static void emit_lin_cp(r300ContextPtr rmesa, unsigned long dst, unsigned long s
 		size -= cp_size;
 	}
 	
-	reg_start(0x4e4c,0);
+	reg_start(R300_RB3D_DSTCACHE_CTLSTAT,0);
 	e32(0x0000000a);
 	
 	reg_start(0x342c,0);
diff --git a/src/mesa/drivers/dri/r300/radeon_state.c b/src/mesa/drivers/dri/r300/radeon_state.c
index ddadf83a00..902e9583a6 100644
--- a/src/mesa/drivers/dri/r300/radeon_state.c
+++ b/src/mesa/drivers/dri/r300/radeon_state.c
@@ -185,6 +185,8 @@ void radeonSetCliprects(radeonContextPtr radeon)
 
 	if (radeon->state.scissor.enabled)
 		radeonRecalcScissorRects(radeon);
+
+	radeon->lastStamp = drawable->lastStamp;
 }
 
 
diff --git a/src/mesa/drivers/dri/radeon/radeon_screen.c b/src/mesa/drivers/dri/radeon/radeon_screen.c
index fc5aa11462..abb14faaa1 100644
--- a/src/mesa/drivers/dri/radeon/radeon_screen.c
+++ b/src/mesa/drivers/dri/radeon/radeon_screen.c
@@ -901,7 +901,7 @@ radeonCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 radeonDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 #if RADEON_COMMON && defined(RADEON_COMMON_FOR_R300)
diff --git a/src/mesa/drivers/dri/s3v/s3v_xmesa.c b/src/mesa/drivers/dri/s3v/s3v_xmesa.c
index c451f7452b..c66fd6dac3 100644
--- a/src/mesa/drivers/dri/s3v/s3v_xmesa.c
+++ b/src/mesa/drivers/dri/s3v/s3v_xmesa.c
@@ -131,7 +131,7 @@ s3vCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 s3vDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 static void
diff --git a/src/mesa/drivers/dri/savage/savage_xmesa.c b/src/mesa/drivers/dri/savage/savage_xmesa.c
index ad79b9235c..f859217069 100644
--- a/src/mesa/drivers/dri/savage/savage_xmesa.c
+++ b/src/mesa/drivers/dri/savage/savage_xmesa.c
@@ -710,7 +710,7 @@ savageCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 savageDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 #if 0
diff --git a/src/mesa/drivers/dri/sis/sis_screen.c b/src/mesa/drivers/dri/sis/sis_screen.c
index 8f52cfe098..89d734ba78 100644
--- a/src/mesa/drivers/dri/sis/sis_screen.c
+++ b/src/mesa/drivers/dri/sis/sis_screen.c
@@ -233,7 +233,7 @@ sisCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 sisDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 static void sisCopyBuffer( __DRIdrawablePrivate *dPriv )
diff --git a/src/mesa/drivers/dri/tdfx/tdfx_screen.c b/src/mesa/drivers/dri/tdfx/tdfx_screen.c
index 646f5126eb..1f9ff4e30c 100644
--- a/src/mesa/drivers/dri/tdfx/tdfx_screen.c
+++ b/src/mesa/drivers/dri/tdfx/tdfx_screen.c
@@ -233,7 +233,7 @@ tdfxCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 tdfxDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/dri/trident/trident_context.c b/src/mesa/drivers/dri/trident/trident_context.c
index dbbd1ac0c6..8dc7f0dc78 100644
--- a/src/mesa/drivers/dri/trident/trident_context.c
+++ b/src/mesa/drivers/dri/trident/trident_context.c
@@ -279,7 +279,7 @@ tridentCreateBuffer( __DRIscreenPrivate *driScrnPriv,
 static void
 tridentDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-   _mesa_destroy_framebuffer((GLframebuffer *) (driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 static void
diff --git a/src/mesa/drivers/dri/unichrome/via_screen.c b/src/mesa/drivers/dri/unichrome/via_screen.c
index 28e1f9451e..90f76be44d 100644
--- a/src/mesa/drivers/dri/unichrome/via_screen.c
+++ b/src/mesa/drivers/dri/unichrome/via_screen.c
@@ -320,7 +320,7 @@ viaCreateBuffer(__DRIscreenPrivate *driScrnPriv,
 static void
 viaDestroyBuffer(__DRIdrawablePrivate *driDrawPriv)
 {
-    _mesa_destroy_framebuffer((GLframebuffer *)(driDrawPriv->driverPrivate));
+   _mesa_unreference_framebuffer((GLframebuffer **)(&(driDrawPriv->driverPrivate)));
 }
 
 
diff --git a/src/mesa/drivers/glide/fxapi.c b/src/mesa/drivers/glide/fxapi.c
index e535e739b3..00b9d29fac 100644
--- a/src/mesa/drivers/glide/fxapi.c
+++ b/src/mesa/drivers/glide/fxapi.c
@@ -728,7 +728,7 @@ errorhandler:
        FREE(fxMesa->fogTable);
     }
     if (fxMesa->glBuffer) {
-       _mesa_destroy_framebuffer(fxMesa->glBuffer);
+       _mesa_unreference_framebuffer(&fxMesa->glBuffer);
     }
     if (fxMesa->glVis) {
        _mesa_destroy_visual(fxMesa->glVis);
@@ -828,7 +828,7 @@ fxMesaDestroyContext(fxMesaContext fxMesa)
    fxDDDestroyFxMesaContext(fxMesa); /* must be before _mesa_destroy_context */
    _mesa_destroy_visual(fxMesa->glVis);
    _mesa_destroy_context(fxMesa->glCtx);
-   _mesa_destroy_framebuffer(fxMesa->glBuffer);
+   _mesa_unreference_framebuffer(&fxMesa->glBuffer);
    fxTMClose(fxMesa); /* must be after _mesa_destroy_context */
 
    FREE(fxMesa);
diff --git a/src/mesa/drivers/osmesa/osmesa.c b/src/mesa/drivers/osmesa/osmesa.c
index c4fc8821e5..96b9b5c174 100644
--- a/src/mesa/drivers/osmesa/osmesa.c
+++ b/src/mesa/drivers/osmesa/osmesa.c
@@ -1303,7 +1303,8 @@ OSMesaDestroyContext( OSMesaContext osmesa )
       _swrast_DestroyContext( &osmesa->mesa );
 
       _mesa_destroy_visual( osmesa->gl_visual );
-      _mesa_destroy_framebuffer( osmesa->gl_buffer );
+      _mesa_unreference_framebuffer( &osmesa->gl_buffer );
+
       _mesa_free_context_data( &osmesa->mesa );
       _mesa_free( osmesa );
    }
diff --git a/src/mesa/drivers/svga/svgamesa.c b/src/mesa/drivers/svga/svgamesa.c
index 0dd9a144fe..d138587569 100644
--- a/src/mesa/drivers/svga/svgamesa.c
+++ b/src/mesa/drivers/svga/svgamesa.c
@@ -433,7 +433,6 @@ void SVGAMesaDestroyContext( SVGAMesaContext ctx )
    if (ctx) {
       _mesa_destroy_visual( ctx->gl_vis );
       _mesa_destroy_context( ctx->gl_ctx );
-      _mesa_destroy_framebuffer( ctx->gl_buffer );
       free( ctx );
       if (ctx==SVGAMesa) {
          SVGAMesa = NULL;
diff --git a/src/mesa/drivers/x11/xm_api.c b/src/mesa/drivers/x11/xm_api.c
index b0ef422de2..a42de72827 100644
--- a/src/mesa/drivers/x11/xm_api.c
+++ b/src/mesa/drivers/x11/xm_api.c
@@ -345,7 +345,7 @@ xmesa_get_window_size(XMesaDisplay *dpy, XMesaBuffer b,
 /*****                Linked list of XMesaBuffers                 *****/
 /**********************************************************************/
 
-static XMesaBuffer XMesaBufferList = NULL;
+XMesaBuffer XMesaBufferList = NULL;
 
 
 /**
@@ -378,6 +378,7 @@ create_xmesa_buffer(XMesaDrawable d, BufferType type,
    b->cmap = cmap;
 
    _mesa_initialize_framebuffer(&b->mesa_buffer, &vis->mesa_visual);
+   b->mesa_buffer.Delete = xmesa_delete_framebuffer;
 
    /*
     * Front renderbuffer
@@ -451,8 +452,8 @@ create_xmesa_buffer(XMesaDrawable d, BufferType type,
  * Find an XMesaBuffer by matching X display and colormap but NOT matching
  * the notThis buffer.
  */
-static XMesaBuffer
-find_xmesa_buffer(XMesaDisplay *dpy, XMesaColormap cmap, XMesaBuffer notThis)
+XMesaBuffer
+xmesa_find_buffer(XMesaDisplay *dpy, XMesaColormap cmap, XMesaBuffer notThis)
 {
    XMesaBuffer b;
    for (b=XMesaBufferList; b; b=b->Next) {
@@ -465,38 +466,27 @@ find_xmesa_buffer(XMesaDisplay *dpy, XMesaColormap cmap, XMesaBuffer notThis)
 
 
 /**
- * Free an XMesaBuffer, remove from linked list, perhaps free X colormap
- * entries.
+ * Remove buffer from linked list, delete if no longer referenced.
  */
 static void
-free_xmesa_buffer(int client, XMesaBuffer buffer)
+xmesa_free_buffer(XMesaBuffer buffer)
 {
    XMesaBuffer prev = NULL, b;
-   (void) client;
-   for (b=XMesaBufferList; b; b=b->Next) {
-      if (b==buffer) {
-         /* unlink bufer from list */
+
+   for (b = XMesaBufferList; b; b = b->Next) {
+      if (b == buffer) {
+         struct gl_framebuffer *fb = &buffer->mesa_buffer;
+
+         /* unlink buffer from list */
          if (prev)
             prev->Next = buffer->Next;
          else
             XMesaBufferList = buffer->Next;
-         /* Check to free X colors */
-         if (buffer->num_alloced>0) {
-            /* If no other buffer uses this X colormap then free the colors. */
-            if (!find_xmesa_buffer(buffer->display, buffer->cmap, buffer)) {
-#ifdef XFree86Server
-               (void)FreeColors(buffer->cmap, client,
-				buffer->num_alloced, buffer->alloced_colors,
-				0);
-#else
-               XFreeColors(buffer->display, buffer->cmap,
-                           buffer->alloced_colors, buffer->num_alloced, 0);
-#endif
-            }
-         }
 
-         _mesa_free_framebuffer_data(&buffer->mesa_buffer);
-         _mesa_free(buffer);
+         /* mark as delete pending */
+         fb->DeletePending = GL_TRUE;
+         /* Unreference.  If count = zero we'll really delete the buffer */
+         _mesa_unreference_framebuffer(&fb);
 
          return;
       }
@@ -504,7 +494,7 @@ free_xmesa_buffer(int client, XMesaBuffer buffer)
       prev = b;
    }
    /* buffer not found in XMesaBufferList */
-   _mesa_problem(NULL,"free_xmesa_buffer() - buffer not found\n");
+   _mesa_problem(NULL,"xmesa_free_buffer() - buffer not found\n");
 }
 
 
@@ -686,7 +676,7 @@ setup_grayscale(int client, XMesaVisual v,
          return GL_FALSE;
       }
 
-      prevBuffer = find_xmesa_buffer(v->display, cmap, buffer);
+      prevBuffer = xmesa_find_buffer(v->display, cmap, buffer);
       if (prevBuffer &&
           (buffer->xm_visual->mesa_visual.rgbMode ==
            prevBuffer->xm_visual->mesa_visual.rgbMode)) {
@@ -775,7 +765,7 @@ setup_dithered_color(int client, XMesaVisual v,
          return GL_FALSE;
       }
 
-      prevBuffer = find_xmesa_buffer(v->display, cmap, buffer);
+      prevBuffer = xmesa_find_buffer(v->display, cmap, buffer);
       if (prevBuffer &&
           (buffer->xm_visual->mesa_visual.rgbMode ==
            prevBuffer->xm_visual->mesa_visual.rgbMode)) {
@@ -1666,7 +1656,7 @@ XMesaCreateWindowBuffer2(XMesaVisual v, XMesaWindow w, XMesaContext c)
 
    if (!initialize_visual_and_buffer( client, v, b, v->mesa_visual.rgbMode,
                                       (XMesaDrawable) w, cmap )) {
-      free_xmesa_buffer(client, b);
+      xmesa_free_buffer(b);
       return NULL;
    }
 
@@ -1787,7 +1777,7 @@ XMesaCreatePixmapBuffer(XMesaVisual v, XMesaPixmap p, XMesaColormap cmap)
 
    if (!initialize_visual_and_buffer(client, v, b, v->mesa_visual.rgbMode,
 				     (XMesaDrawable) p, cmap)) {
-      free_xmesa_buffer(client, b);
+      xmesa_free_buffer(b);
       return NULL;
    }
 
@@ -1821,7 +1811,7 @@ XMesaCreatePBuffer(XMesaVisual v, XMesaColormap cmap,
 
    if (!initialize_visual_and_buffer(client, v, b, v->mesa_visual.rgbMode,
 				     drawable, cmap)) {
-      free_xmesa_buffer(client, b);
+      xmesa_free_buffer(b);
       return NULL;
    }
 
@@ -1834,48 +1824,10 @@ XMesaCreatePBuffer(XMesaVisual v, XMesaColormap cmap,
 /*
  * Deallocate an XMesaBuffer structure and all related info.
  */
-void XMesaDestroyBuffer( XMesaBuffer b )
+void
+XMesaDestroyBuffer(XMesaBuffer b)
 {
-   int client = 0;
-
-#ifdef XFree86Server
-   if (b->frontxrb->drawable)
-       client = CLIENT_ID(b->frontxrb->drawable->id);
-#endif
-
-   if (b->gc)  XMesaFreeGC( b->xm_visual->display, b->gc );
-   if (b->cleargc)  XMesaFreeGC( b->xm_visual->display, b->cleargc );
-   if (b->swapgc)  XMesaFreeGC( b->xm_visual->display, b->swapgc );
-
-   if (b->xm_visual->mesa_visual.doubleBufferMode)
-   {
-      if (b->backxrb->ximage) {
-#if defined(USE_XSHM) && !defined(XFree86Server)
-         if (b->shm) {
-            XShmDetach( b->xm_visual->display, &b->shminfo );
-            XDestroyImage( b->backxrb->ximage );
-            shmdt( b->shminfo.shmaddr );
-         }
-         else
-#endif
-            XMesaDestroyImage( b->backxrb->ximage );
-      }
-      if (b->backxrb->pixmap) {
-         XMesaFreePixmap( b->xm_visual->display, b->backxrb->pixmap );
-         if (b->xm_visual->hpcr_clear_flag) {
-            XMesaFreePixmap( b->xm_visual->display,
-                             b->xm_visual->hpcr_clear_pixmap );
-            XMesaDestroyImage( b->xm_visual->hpcr_clear_ximage );
-         }
-      }
-   }
-   if (b->rowimage) {
-      _mesa_free( b->rowimage->data );
-      b->rowimage->data = NULL;
-      XMesaDestroyImage( b->rowimage );
-   }
-
-   free_xmesa_buffer(client, b);
+   xmesa_free_buffer(b);
 }
 
 
@@ -2436,7 +2388,7 @@ void xmesa_destroy_buffers_on_display(XMesaDisplay *dpy)
    for (b = XMesaBufferList; b; b = next) {
       next = b->Next;
       if (b->display == dpy) {
-         free_xmesa_buffer(0, b);
+         xmesa_free_buffer(b);
       }
    }
 }
diff --git a/src/mesa/drivers/x11/xm_buffer.c b/src/mesa/drivers/x11/xm_buffer.c
index 490c479676..a358ec25ae 100644
--- a/src/mesa/drivers/x11/xm_buffer.c
+++ b/src/mesa/drivers/x11/xm_buffer.c
@@ -33,6 +33,7 @@
 #include "GL/xmesa.h"
 #include "xmesaP.h"
 #include "imports.h"
+#include "framebuffer.h"
 #include "renderbuffer.h"
 
 
@@ -352,5 +353,71 @@ xmesa_new_renderbuffer(GLcontext *ctx, GLuint name, const GLvisual *visual,
 }
 
 
+/**
+ * Called via gl_framebuffer::Delete() method when this buffer
+ * is _really_ being deleted.
+ */
+void
+xmesa_delete_framebuffer(struct gl_framebuffer *fb)
+{
+   XMesaBuffer b = XMESA_BUFFER(fb);
 
+#ifdef XFree86Server
+   int client = 0;
+   if (b->frontxrb->drawable)
+       client = CLIENT_ID(b->frontxrb->drawable->id);
+#endif
 
+   if (b->num_alloced > 0) {
+      /* If no other buffer uses this X colormap then free the colors. */
+      if (!xmesa_find_buffer(b->display, b->cmap, b)) {
+#ifdef XFree86Server
+         (void)FreeColors(b->cmap, client,
+                          b->num_alloced, b->alloced_colors, 0);
+#else
+         XFreeColors(b->display, b->cmap,
+                     b->alloced_colors, b->num_alloced, 0);
+#endif
+      }
+   }
+
+   if (b->gc)
+      XMesaFreeGC(b->xm_visual->display, b->gc);
+   if (b->cleargc)
+      XMesaFreeGC(b->xm_visual->display, b->cleargc);
+   if (b->swapgc)
+      XMesaFreeGC(b->xm_visual->display, b->swapgc);
+
+   if (b->xm_visual->mesa_visual.doubleBufferMode) {
+      /* free back ximage/pixmap/shmregion */
+      if (b->backxrb->ximage) {
+#if defined(USE_XSHM) && !defined(XFree86Server)
+         if (b->shm) {
+            XShmDetach( b->xm_visual->display, &b->shminfo );
+            XDestroyImage( b->backxrb->ximage );
+            shmdt( b->shminfo.shmaddr );
+         }
+         else
+#endif
+            XMesaDestroyImage( b->backxrb->ximage );
+         b->backxrb->ximage = NULL;
+      }
+      if (b->backxrb->pixmap) {
+         XMesaFreePixmap( b->xm_visual->display, b->backxrb->pixmap );
+         if (b->xm_visual->hpcr_clear_flag) {
+            XMesaFreePixmap( b->xm_visual->display,
+                             b->xm_visual->hpcr_clear_pixmap );
+            XMesaDestroyImage( b->xm_visual->hpcr_clear_ximage );
+         }
+      }
+   }
+
+   if (b->rowimage) {
+      _mesa_free( b->rowimage->data );
+      b->rowimage->data = NULL;
+      XMesaDestroyImage( b->rowimage );
+   }
+
+   _mesa_free_framebuffer_data(fb);
+   _mesa_free(fb);
+}
diff --git a/src/mesa/drivers/x11/xmesaP.h b/src/mesa/drivers/x11/xmesaP.h
index e332fb5480..5516031ffe 100644
--- a/src/mesa/drivers/x11/xmesaP.h
+++ b/src/mesa/drivers/x11/xmesaP.h
@@ -42,6 +42,7 @@
 
 extern _glthread_Mutex _xmesa_lock;
 
+extern XMesaBuffer XMesaBufferList;
 
 /* for PF_8R8G8B24 pixel format */
 typedef struct {
@@ -489,6 +490,12 @@ extern struct xmesa_renderbuffer *
 xmesa_new_renderbuffer(GLcontext *ctx, GLuint name, const GLvisual *visual,
                        GLboolean backBuffer);
 
+extern void
+xmesa_delete_framebuffer(struct gl_framebuffer *fb);
+
+extern XMesaBuffer
+xmesa_find_buffer(XMesaDisplay *dpy, XMesaColormap cmap, XMesaBuffer notThis);
+
 extern unsigned long
 xmesa_color_to_pixel( GLcontext *ctx,
                       GLubyte r, GLubyte g, GLubyte b, GLubyte a,
diff --git a/src/mesa/glapi/gl_table.py b/src/mesa/glapi/gl_table.py
index 52b4be1d2b..69f7bd7c7b 100644
--- a/src/mesa/glapi/gl_table.py
+++ b/src/mesa/glapi/gl_table.py
@@ -49,7 +49,11 @@ class PrintGlTable(gl_XML.gl_print_base):
 
 	def printRealHeader(self):
 		print '#ifndef GLAPIENTRYP'
-		print '#define GLAPIENTRYP'
+		print '# ifndef GLAPIENTRY'
+		print '#  define GLAPIENTRY'
+		print '# endif'
+		print ''
+		print '# define GLAPIENTRYP GLAPIENTRY *'
 		print '#endif'
 		print ''
 		print 'typedef void (*_glapi_proc)(void); /* generic function pointer */'
diff --git a/src/mesa/glapi/glapitable.h b/src/mesa/glapi/glapitable.h
index 55548adcd3..4af0c2d43b 100644
--- a/src/mesa/glapi/glapitable.h
+++ b/src/mesa/glapi/glapitable.h
@@ -30,7 +30,11 @@
 #  define _GLAPI_TABLE_H_
 
 #ifndef GLAPIENTRYP
-#define GLAPIENTRYP
+# ifndef GLAPIENTRY
+#  define GLAPIENTRY
+# endif
+
+# define GLAPIENTRYP GLAPIENTRY *
 #endif
 
 typedef void (*_glapi_proc)(void); /* generic function pointer */
diff --git a/src/mesa/main/context.c b/src/mesa/main/context.c
index 9b3759b6c8..135c814c0a 100644
--- a/src/mesa/main/context.c
+++ b/src/mesa/main/context.c
@@ -95,6 +95,7 @@
 #include "fbobject.h"
 #include "feedback.h"
 #include "fog.h"
+#include "framebuffer.h"
 #include "get.h"
 #include "glthread.h"
 #include "glapioffsets.h"
@@ -1407,6 +1408,13 @@ _mesa_free_context_data( GLcontext *ctx )
    if (ctx == _mesa_get_current_context()) {
       _mesa_make_current(NULL, NULL, NULL);
    }
+   else {
+      /* unreference WinSysDraw/Read buffers */
+      _mesa_unreference_framebuffer(&ctx->WinSysDrawBuffer);
+      _mesa_unreference_framebuffer(&ctx->WinSysReadBuffer);
+      _mesa_unreference_framebuffer(&ctx->DrawBuffer);
+      _mesa_unreference_framebuffer(&ctx->ReadBuffer);
+   }
 
    _mesa_free_lighting_data( ctx );
    _mesa_free_eval_data( ctx );
@@ -1666,6 +1674,8 @@ void
 _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer,
                     GLframebuffer *readBuffer )
 {
+   GET_CURRENT_CONTEXT(oldCtx);
+
    if (MESA_VERBOSE & VERBOSE_API)
       _mesa_debug(newCtx, "_mesa_make_current()\n");
 
@@ -1690,6 +1700,11 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer,
    _glapi_set_context((void *) newCtx);
    ASSERT(_mesa_get_current_context() == newCtx);
 
+   if (oldCtx) {
+      _mesa_unreference_framebuffer(&oldCtx->WinSysDrawBuffer);
+      _mesa_unreference_framebuffer(&oldCtx->WinSysReadBuffer);
+   }
+         
    if (!newCtx) {
       _glapi_set_dispatch(NULL);  /* none current */
    }
@@ -1701,18 +1716,18 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer,
 
          ASSERT(drawBuffer->Name == 0);
          ASSERT(readBuffer->Name == 0);
-         newCtx->WinSysDrawBuffer = drawBuffer;
-         newCtx->WinSysReadBuffer = readBuffer;
+         _mesa_reference_framebuffer(&newCtx->WinSysDrawBuffer, drawBuffer);
+         _mesa_reference_framebuffer(&newCtx->WinSysReadBuffer, readBuffer);
 
          /*
           * Only set the context's Draw/ReadBuffer fields if they're NULL
           * or not bound to a user-created FBO.
           */
          if (!newCtx->DrawBuffer || newCtx->DrawBuffer->Name == 0) {
-            newCtx->DrawBuffer = drawBuffer;
+            _mesa_reference_framebuffer(&newCtx->DrawBuffer, drawBuffer);
          }
          if (!newCtx->ReadBuffer || newCtx->ReadBuffer->Name == 0) {
-            newCtx->ReadBuffer = readBuffer;
+            _mesa_reference_framebuffer(&newCtx->ReadBuffer, readBuffer);
          }
 
 	 newCtx->NewState |= _NEW_BUFFERS;
diff --git a/src/mesa/main/fbobject.c b/src/mesa/main/fbobject.c
index cf8de1e0cb..6608eefc6c 100644
--- a/src/mesa/main/fbobject.c
+++ b/src/mesa/main/fbobject.c
@@ -602,7 +602,7 @@ _mesa_BindRenderbufferEXT(GLenum target, GLuint renderbuffer)
 
    oldRb = ctx->CurrentRenderbuffer;
    if (oldRb) {
-      _mesa_dereference_renderbuffer(&oldRb);
+      _mesa_unreference_renderbuffer(&oldRb);
    }
 
    ASSERT(newRb != &DummyRenderbuffer);
@@ -639,7 +639,7 @@ _mesa_DeleteRenderbuffersEXT(GLsizei n, const GLuint *renderbuffers)
                /* But the object will not be freed until it's no longer
                 * bound in any context.
                 */
-               _mesa_dereference_renderbuffer(&rb);
+               _mesa_unreference_renderbuffer(&rb);
 	    }
 	 }
       }
@@ -938,7 +938,7 @@ check_end_texture_render(GLcontext *ctx, struct gl_framebuffer *fb)
 void GLAPIENTRY
 _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
 {
-   struct gl_framebuffer *newFb, *oldFb;
+   struct gl_framebuffer *newFb;
    GLboolean bindReadBuf, bindDrawBuf;
    GET_CURRENT_CONTEXT(ctx);
 
@@ -998,12 +998,6 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
 	 }
          _mesa_HashInsert(ctx->Shared->FrameBuffers, framebuffer, newFb);
       }
-      _glthread_LOCK_MUTEX(newFb->Mutex);
-      if (bindReadBuf)
-         newFb->RefCount++;
-      if (bindDrawBuf)
-         newFb->RefCount++;
-      _glthread_UNLOCK_MUTEX(newFb->Mutex);
    }
    else {
       /* Binding the window system framebuffer (which was originally set
@@ -1020,22 +1014,16 @@ _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer)
     */
 
    if (bindReadBuf) {
-      oldFb = ctx->ReadBuffer;
-      if (oldFb && oldFb->Name != 0) {
-         _mesa_dereference_framebuffer(&oldFb);
-      }
-      ctx->ReadBuffer = newFb;
+      _mesa_unreference_framebuffer(&ctx->ReadBuffer);
+      _mesa_reference_framebuffer(&ctx->ReadBuffer, newFb);
    }
 
    if (bindDrawBuf) {
-      oldFb = ctx->DrawBuffer;
-      if (oldFb && oldFb->Name != 0) {
-         /* check if old FB had any texture attachments */
-         check_end_texture_render(ctx, oldFb);
-         /* check if time to delete this framebuffer */
-         _mesa_dereference_framebuffer(&oldFb);
-      }
-      ctx->DrawBuffer = newFb;
+      /* check if old FB had any texture attachments */
+      check_end_texture_render(ctx, ctx->DrawBuffer);
+      /* check if time to delete this framebuffer */
+      _mesa_unreference_framebuffer(&ctx->DrawBuffer);
+      _mesa_reference_framebuffer(&ctx->DrawBuffer, newFb);
       if (newFb->Name != 0) {
          /* check if newly bound framebuffer has any texture attachments */
          check_begin_texture_render(ctx, newFb);
@@ -1083,7 +1071,7 @@ _mesa_DeleteFramebuffersEXT(GLsizei n, const GLuint *framebuffers)
                /* But the object will not be freed until it's no longer
                 * bound in any context.
                 */
-               _mesa_dereference_framebuffer(&fb);
+               _mesa_unreference_framebuffer(&fb);
 	    }
 	 }
       }
diff --git a/src/mesa/main/framebuffer.c b/src/mesa/main/framebuffer.c
index 465197401b..c97d2f0077 100644
--- a/src/mesa/main/framebuffer.c
+++ b/src/mesa/main/framebuffer.c
@@ -78,7 +78,7 @@ set_depth_renderbuffer(struct gl_framebuffer *fb,
                        struct gl_renderbuffer *rb)
 {
    if (fb->_DepthBuffer) {
-      _mesa_dereference_renderbuffer(&fb->_DepthBuffer);
+      _mesa_unreference_renderbuffer(&fb->_DepthBuffer);
    }
    fb->_DepthBuffer = rb;
    if (rb) {
@@ -96,7 +96,7 @@ set_stencil_renderbuffer(struct gl_framebuffer *fb,
                          struct gl_renderbuffer *rb)
 {
    if (fb->_StencilBuffer) {
-      _mesa_dereference_renderbuffer(&fb->_StencilBuffer);
+      _mesa_unreference_renderbuffer(&fb->_StencilBuffer);
    }
    fb->_StencilBuffer = rb;
    if (rb) {
@@ -166,6 +166,8 @@ _mesa_initialize_framebuffer(struct gl_framebuffer *fb, const GLvisual *visual)
 
    _glthread_INIT_MUTEX(fb->Mutex);
 
+   fb->RefCount = 1;
+
    /* save the visual */
    fb->Visual = *visual;
 
@@ -198,7 +200,6 @@ void
 _mesa_destroy_framebuffer(struct gl_framebuffer *fb)
 {
    if (fb) {
-      _glthread_DESTROY_MUTEX(fb->Mutex);
       _mesa_free_framebuffer_data(fb);
       _mesa_free(fb);
    }
@@ -215,20 +216,27 @@ _mesa_free_framebuffer_data(struct gl_framebuffer *fb)
    GLuint i;
 
    assert(fb);
+   assert(fb->RefCount == 0);
+
+   _glthread_DESTROY_MUTEX(fb->Mutex);
 
    for (i = 0; i < BUFFER_COUNT; i++) {
       struct gl_renderbuffer_attachment *att = &fb->Attachment[i];
       if (att->Renderbuffer) {
-         struct gl_renderbuffer *rb = att->Renderbuffer;
-         /* remove framebuffer's reference to renderbuffer */
-         _mesa_dereference_renderbuffer(&rb);
-         if (rb && rb->Name == 0) {
-            /* delete window system renderbuffer */
-            _mesa_dereference_renderbuffer(&rb);
+         _mesa_unreference_renderbuffer(&att->Renderbuffer);
+      }
+      if (att->Texture) {
+         /* render to texture */
+         att->Texture->RefCount--;
+         if (att->Texture->RefCount == 0) {
+            GET_CURRENT_CONTEXT(ctx);
+            if (ctx) {
+               ctx->Driver.DeleteTexture(ctx, att->Texture);
+            }
          }
       }
       att->Type = GL_NONE;
-      att->Renderbuffer = NULL;
+      att->Texture = NULL;
    }
 
    /* unbind depth/stencil to decr ref counts */
@@ -238,25 +246,44 @@ _mesa_free_framebuffer_data(struct gl_framebuffer *fb)
 
 
 /**
- * Decrement the reference count on a framebuffer and delete it when
+ * Set *ptr to point to fb, with refcounting and locking.
+ */
+void
+_mesa_reference_framebuffer(struct gl_framebuffer **ptr,
+                            struct gl_framebuffer *fb)
+{
+   assert(ptr);
+   assert(!*ptr);
+   assert(fb);
+   _glthread_LOCK_MUTEX(fb->Mutex);
+   fb->RefCount++;
+   _glthread_UNLOCK_MUTEX(fb->Mutex);
+   *ptr = fb;
+}
+
+
+/**
+ * Undo/remove a reference to a framebuffer object.
+ * Decrement the framebuffer object's reference count and delete it when
  * the refcount hits zero.
- * Note: we pass the address of a pointer and set it to NULL if we delete it.
+ * Note: we pass the address of a pointer and set it to NULL.
  */
 void
-_mesa_dereference_framebuffer(struct gl_framebuffer **fb)
+_mesa_unreference_framebuffer(struct gl_framebuffer **fb)
 {
-   GLboolean deleteFlag = GL_FALSE;
+   assert(fb);
+   if (*fb) {
+      GLboolean deleteFlag = GL_FALSE;
 
-   _glthread_LOCK_MUTEX((*fb)->Mutex);
-   {
+      _glthread_LOCK_MUTEX((*fb)->Mutex);
       ASSERT((*fb)->RefCount > 0);
       (*fb)->RefCount--;
       deleteFlag = ((*fb)->RefCount == 0);
-   }
-   _glthread_UNLOCK_MUTEX((*fb)->Mutex);
+      _glthread_UNLOCK_MUTEX((*fb)->Mutex);
+      
+      if (deleteFlag)
+         (*fb)->Delete(*fb);
 
-   if (deleteFlag) {
-      (*fb)->Delete(*fb);
       *fb = NULL;
    }
 }
@@ -605,21 +632,25 @@ update_color_draw_buffers(GLcontext *ctx, struct gl_framebuffer *fb)
       GLbitfield bufferMask = fb->_ColorDrawBufferMask[output];
       GLuint count = 0;
       GLuint i;
-      /* We need the inner loop here because glDrawBuffer(GL_FRONT_AND_BACK)
-       * can specify writing to two or four color buffers (for example).
-       */
-      for (i = 0; bufferMask && i < BUFFER_COUNT; i++) {
-         const GLuint bufferBit = 1 << i;
-         if (bufferBit & bufferMask) {
-            struct gl_renderbuffer *rb = fb->Attachment[i].Renderbuffer;
-            if (rb) {
-               fb->_ColorDrawBuffers[output][count] = rb;
-               count++;
-            }
-            else {
-               /*_mesa_warning(ctx, "DrawBuffer names a missing buffer!\n");*/
+      if (!fb->DeletePending) {
+         /* We need the inner loop here because glDrawBuffer(GL_FRONT_AND_BACK)
+          * can specify writing to two or four color buffers (for example).
+          */
+         for (i = 0; bufferMask && i < BUFFER_COUNT; i++) {
+            const GLuint bufferBit = 1 << i;
+            if (bufferBit & bufferMask) {
+               struct gl_renderbuffer *rb = fb->Attachment[i].Renderbuffer;
+               if (rb) {
+                  fb->_ColorDrawBuffers[output][count] = rb;
+                  count++;
+               }
+               else {
+                  /*
+                  _mesa_warning(ctx, "DrawBuffer names a missing buffer!\n");
+                  */
+               }
+               bufferMask &= ~bufferBit;
             }
-            bufferMask &= ~bufferBit;
          }
       }
       fb->_NumColorDrawBuffers[output] = count;
@@ -635,7 +666,7 @@ static void
 update_color_read_buffer(GLcontext *ctx, struct gl_framebuffer *fb)
 {
    (void) ctx;
-   if (fb->_ColorReadBufferIndex == -1) {
+   if (fb->_ColorReadBufferIndex == -1 || fb->DeletePending) {
       fb->_ColorReadBuffer = NULL; /* legal! */
    }
    else {
diff --git a/src/mesa/main/framebuffer.h b/src/mesa/main/framebuffer.h
index 7f3254fe8e..4d76f3a90f 100644
--- a/src/mesa/main/framebuffer.h
+++ b/src/mesa/main/framebuffer.h
@@ -43,7 +43,11 @@ extern void
 _mesa_free_framebuffer_data(struct gl_framebuffer *buffer);
 
 extern void
-_mesa_dereference_framebuffer(struct gl_framebuffer **fb);
+_mesa_reference_framebuffer(struct gl_framebuffer **ptr,
+                            struct gl_framebuffer *fb);
+
+extern void
+_mesa_unreference_framebuffer(struct gl_framebuffer **fb);
 
 extern void
 _mesa_resize_framebuffer(GLcontext *ctx, struct gl_framebuffer *fb,
diff --git a/src/mesa/main/mtypes.h b/src/mesa/main/mtypes.h
index e8f0f45d39..422d176c25 100644
--- a/src/mesa/main/mtypes.h
+++ b/src/mesa/main/mtypes.h
@@ -2243,6 +2243,7 @@ struct gl_framebuffer
    _glthread_Mutex Mutex;		   /**< for thread safety */
    GLuint Name;      /* if zero, this is a window system framebuffer */
    GLint RefCount;
+   GLboolean DeletePending;
 
    GLvisual Visual;	/**< The framebuffer's visual.
                              Immutable if this is a window system buffer.
diff --git a/src/mesa/main/rastpos.c b/src/mesa/main/rastpos.c
index ddf2ac51d3..d4f0fa2a71 100644
--- a/src/mesa/main/rastpos.c
+++ b/src/mesa/main/rastpos.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.3
+ * Version:  6.5.3
  *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  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"),
@@ -112,9 +112,7 @@ userclip_point( GLcontext *ctx, const GLfloat v[] )
 
 
 /**
- * This has been split off to allow the normal shade routines to
- * get a little closer to the vertex buffer, and to use the
- * GLvector objects directly.
+ * Compute lighting for the raster position.  Both RGB and CI modes computed.
  * \param ctx the context
  * \param vertex vertex location
  * \param normal normal vector
@@ -130,10 +128,10 @@ shade_rastpos(GLcontext *ctx,
               GLfloat Rspec[4],
               GLfloat *Rindex)
 {
-   GLfloat (*base)[3] = ctx->Light._BaseColor;
-   struct gl_light *light;
-   GLfloat diffuseColor[4], specularColor[4];
-   GLfloat diffuse = 0, specular = 0;
+   /*const*/ GLfloat (*base)[3] = ctx->Light._BaseColor;
+   const struct gl_light *light;
+   GLfloat diffuseColor[4], specularColor[4];  /* for RGB mode only */
+   GLfloat diffuseCI = 0.0, specularCI = 0.0;  /* for CI mode only */
 
    if (!ctx->_ShineTable[0] || !ctx->_ShineTable[1])
       _mesa_validate_all_lighting_tables( ctx );
@@ -144,28 +142,31 @@ shade_rastpos(GLcontext *ctx,
    ASSIGN_4V(specularColor, 0.0, 0.0, 0.0, 0.0);
 
    foreach (light, &ctx->Light.EnabledList) {
-      GLfloat n_dot_h;
       GLfloat attenuation = 1.0;
-      GLfloat VP[3];
+      GLfloat VP[3]; /* vector from vertex to light pos */
       GLfloat n_dot_VP;
-      GLfloat *h;
       GLfloat diffuseContrib[3], specularContrib[3];
-      GLboolean normalized;
 
       if (!(light->_Flags & LIGHT_POSITIONAL)) {
+         /* light at infinity */
 	 COPY_3V(VP, light->_VP_inf_norm);
 	 attenuation = light->_VP_inf_spot_attenuation;
       }
       else {
+         /* local/positional light */
 	 GLfloat d;
 
+         /* VP = vector from vertex pos to light[i].pos */
 	 SUB_3V(VP, light->_Position, vertex);
+         /* d = length(VP) */
 	 d = (GLfloat) LEN_3FV( VP );
-
-	 if ( d > 1e-6) {
+	 if (d > 1.0e-6) {
+            /* normalize VP */
 	    GLfloat invd = 1.0F / d;
 	    SELF_SCALE_SCALAR_3V(VP, invd);
 	 }
+
+         /* atti */
 	 attenuation = 1.0F / (light->ConstantAttenuation + d *
 			       (light->LinearAttenuation + d *
 				light->QuadraticAttenuation));
@@ -196,43 +197,39 @@ shade_rastpos(GLcontext *ctx,
 	 continue;
       }
 
+      /* Ambient + diffuse */
       COPY_3V(diffuseContrib, light->_MatAmbient[0]);
       ACC_SCALE_SCALAR_3V(diffuseContrib, n_dot_VP, light->_MatDiffuse[0]);
-      diffuse += n_dot_VP * light->_dli * attenuation;
-      ASSIGN_3V(specularContrib, 0.0, 0.0, 0.0);
+      diffuseCI += n_dot_VP * light->_dli * attenuation;
 
+      /* Specular */
       {
+         const GLfloat *h;
+         GLfloat n_dot_h;
+
+         ASSIGN_3V(specularContrib, 0.0, 0.0, 0.0);
+
 	 if (ctx->Light.Model.LocalViewer) {
 	    GLfloat v[3];
 	    COPY_3V(v, vertex);
 	    NORMALIZE_3FV(v);
 	    SUB_3V(VP, VP, v);
+            NORMALIZE_3FV(VP);
 	    h = VP;
-	    normalized = 0;
 	 }
 	 else if (light->_Flags & LIGHT_POSITIONAL) {
+	    ACC_3V(VP, ctx->_EyeZDir);
+            NORMALIZE_3FV(VP);
 	    h = VP;
-	    ACC_3V(h, ctx->_EyeZDir);
-	    normalized = 0;
 	 }
          else {
 	    h = light->_h_inf_norm;
-	    normalized = 1;
 	 }
 
 	 n_dot_h = DOT3(normal, h);
 
 	 if (n_dot_h > 0.0F) {
-	    GLfloat (*mat)[4] = ctx->Light.Material.Attrib;
 	    GLfloat spec_coef;
-	    GLfloat shininess = mat[MAT_ATTRIB_FRONT_SHININESS][0];
-
-	    if (!normalized) {
-	       n_dot_h *= n_dot_h;
-	       n_dot_h /= LEN_SQUARED_3FV( h );
-	       shininess *= .5;
-	    }
-
 	    GET_SHINE_TAB_ENTRY( ctx->_ShineTable[0], n_dot_h, spec_coef );
 
 	    if (spec_coef > 1.0e-10) {
@@ -244,7 +241,8 @@ shade_rastpos(GLcontext *ctx,
                   ACC_SCALE_SCALAR_3V( diffuseContrib, spec_coef,
                                        light->_MatSpecular[0]);
                }
-	       specular += spec_coef * light->_sli * attenuation;
+               /*assert(light->_sli > 0.0);*/
+               specularCI += spec_coef * light->_sli * attenuation;
 	    }
 	 }
       }
@@ -268,8 +266,8 @@ shade_rastpos(GLcontext *ctx,
       GLfloat d_a = ind[MAT_INDEX_DIFFUSE] - ind[MAT_INDEX_AMBIENT];
       GLfloat s_a = ind[MAT_INDEX_SPECULAR] - ind[MAT_INDEX_AMBIENT];
       GLfloat i = (ind[MAT_INDEX_AMBIENT]
-		   + diffuse * (1.0F-specular) * d_a
-		   + specular * s_a);
+		   + diffuseCI * (1.0F-specularCI) * d_a
+		   + specularCI * s_a);
       if (i > ind[MAT_INDEX_SPECULAR]) {
 	 i = ind[MAT_INDEX_SPECULAR];
       }
diff --git a/src/mesa/main/rbadaptors.c b/src/mesa/main/rbadaptors.c
index 313c8d43d6..60f4948bec 100644
--- a/src/mesa/main/rbadaptors.c
+++ b/src/mesa/main/rbadaptors.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.1
+ * Version:  6.5.3
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  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"),
@@ -45,7 +45,7 @@ Delete_wrapper(struct gl_renderbuffer *rb)
    /* Decrement reference count on the buffer we're wrapping and delete
     * it if refcount hits zero.
     */
-   _mesa_dereference_renderbuffer(&rb->Wrapped);
+   _mesa_unreference_renderbuffer(&rb->Wrapped);
 
    /* delete myself */
    _mesa_delete_renderbuffer(rb);
diff --git a/src/mesa/main/renderbuffer.c b/src/mesa/main/renderbuffer.c
index 6b18d60baf..1cc95a7d3b 100644
--- a/src/mesa/main/renderbuffer.c
+++ b/src/mesa/main/renderbuffer.c
@@ -2089,32 +2089,33 @@ _mesa_remove_renderbuffer(struct gl_framebuffer *fb, GLuint bufferName)
    if (!rb)
       return;
 
-   _mesa_dereference_renderbuffer(&rb);
+   _mesa_unreference_renderbuffer(&rb);
 
    fb->Attachment[bufferName].Renderbuffer = NULL;
 }
 
 
 /**
- * Decrement the reference count on a renderbuffer and delete it when
+ * Decrement a renderbuffer object's reference count and delete it when
  * the refcount hits zero.
- * Note: we pass the address of a pointer and set it to NULL if we delete it.
+ * Note: we pass the address of a pointer.
  */
 void
-_mesa_dereference_renderbuffer(struct gl_renderbuffer **rb)
+_mesa_unreference_renderbuffer(struct gl_renderbuffer **rb)
 {
-   GLboolean deleteFlag = GL_FALSE;
+   assert(rb);
+   if (*rb) {
+      GLboolean deleteFlag = GL_FALSE;
 
-   _glthread_LOCK_MUTEX((*rb)->Mutex);
-   {
+      _glthread_LOCK_MUTEX((*rb)->Mutex);
       ASSERT((*rb)->RefCount > 0);
       (*rb)->RefCount--;
       deleteFlag = ((*rb)->RefCount == 0);
-   }
-   _glthread_UNLOCK_MUTEX((*rb)->Mutex);
+      _glthread_UNLOCK_MUTEX((*rb)->Mutex);
+
+      if (deleteFlag)
+         (*rb)->Delete(*rb);
 
-   if (deleteFlag) {
-      (*rb)->Delete(*rb);
       *rb = NULL;
    }
 }
diff --git a/src/mesa/main/renderbuffer.h b/src/mesa/main/renderbuffer.h
index 74ca43c57a..e1a0a55979 100644
--- a/src/mesa/main/renderbuffer.h
+++ b/src/mesa/main/renderbuffer.h
@@ -99,7 +99,7 @@ extern void
 _mesa_remove_renderbuffer(struct gl_framebuffer *fb, GLuint bufferName);
 
 extern void
-_mesa_dereference_renderbuffer(struct gl_renderbuffer **rb);
+_mesa_unreference_renderbuffer(struct gl_renderbuffer **rb);
 
 extern struct gl_renderbuffer *
 _mesa_new_depthstencil_renderbuffer(GLcontext *ctx, GLuint name);
diff --git a/src/mesa/shader/arbprogparse.c b/src/mesa/shader/arbprogparse.c
index b9ff08de5f..c6ea5c4f97 100644
--- a/src/mesa/shader/arbprogparse.c
+++ b/src/mesa/shader/arbprogparse.c
@@ -4038,6 +4038,7 @@ _mesa_parse_arb_fragment_program(GLcontext* ctx, GLenum target,
    for (i = 0; i < MAX_TEXTURE_IMAGE_UNITS; i++)
       program->TexturesUsed[i] = ap.TexturesUsed[i];
    program->FogOption          = ap.FogOption;
+   program->UsesKill          = ap.UsesKill;
 
    if (program->Base.Instructions)
       _mesa_free(program->Base.Instructions);
diff --git a/src/mesa/swrast/s_readpix.c b/src/mesa/swrast/s_readpix.c
index 27f4736c39..15dc8106b4 100644
--- a/src/mesa/swrast/s_readpix.c
+++ b/src/mesa/swrast/s_readpix.c
@@ -1,8 +1,8 @@
 /*
  * Mesa 3-D graphics library
- * Version:  6.5.2
+ * Version:  6.5.3
  *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2007  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"),
@@ -54,7 +54,8 @@ read_index_pixels( GLcontext *ctx,
    struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
    GLint i;
 
-   ASSERT(rb);
+   if (!rb)
+      return;
 
    /* width should never be > MAX_WIDTH since we did clipping earlier */
    ASSERT(width <= MAX_WIDTH);
@@ -91,6 +92,9 @@ read_depth_pixels( GLcontext *ctx,
    const GLboolean biasOrScale
       = ctx->Pixel.DepthScale != 1.0 || ctx->Pixel.DepthBias != 0.0;
 
+   if (!rb)
+      return;
+
    /* clipping should have been done already */
    ASSERT(x >= 0);
    ASSERT(y >= 0);
@@ -99,8 +103,6 @@ read_depth_pixels( GLcontext *ctx,
    /* width should never be > MAX_WIDTH since we did clipping earlier */
    ASSERT(width <= MAX_WIDTH);
 
-   ASSERT(rb);
-
    if (type == GL_UNSIGNED_SHORT && fb->Visual.depthBits == 16
        && !biasOrScale && !packing->SwapBytes) {
       /* Special case: directly read 16-bit unsigned depth values. */
@@ -171,7 +173,8 @@ read_stencil_pixels( GLcontext *ctx,
    struct gl_renderbuffer *rb = fb->_StencilBuffer;
    GLint j;
 
-   ASSERT(rb);
+   if (!rb)
+      return;
 
    /* width should never be > MAX_WIDTH since we did clipping earlier */
    ASSERT(width <= MAX_WIDTH);
@@ -195,6 +198,7 @@ read_stencil_pixels( GLcontext *ctx,
 /**
  * Optimized glReadPixels for particular pixel formats when pixel
  * scaling, biasing, mapping, etc. are disabled.
+ * \return GL_TRUE if success, GL_FALSE if unable to do the readpixels
  */
 static GLboolean
 fast_read_rgba_pixels( GLcontext *ctx,
@@ -207,6 +211,9 @@ fast_read_rgba_pixels( GLcontext *ctx,
 {
    struct gl_renderbuffer *rb = ctx->ReadBuffer->_ColorReadBuffer;
 
+   if (!rb)
+      return GL_FALSE;
+
    ASSERT(rb->_BaseFormat == GL_RGBA || rb->_BaseFormat == GL_RGB);
 
    /* clipping should have already been done */
@@ -316,7 +323,8 @@ read_rgba_pixels( GLcontext *ctx,
    struct gl_framebuffer *fb = ctx->ReadBuffer;
    struct gl_renderbuffer *rb = fb->_ColorReadBuffer;
 
-   ASSERT(rb);
+   if (!rb)
+      return;
 
    if (type == GL_FLOAT && ((ctx->Color.ClampReadColor == GL_TRUE) ||
                             (ctx->Color.ClampReadColor == GL_FIXED_ONLY_ARB &&
@@ -457,8 +465,8 @@ read_depth_stencil_pixels(GLcontext *ctx,
    depthRb = ctx->ReadBuffer->_DepthBuffer;
    stencilRb = ctx->ReadBuffer->_StencilBuffer;
 
-   ASSERT(depthRb);
-   ASSERT(stencilRb);
+   if (!depthRb || !stencilRb)
+      return;
 
    depthRb = ctx->ReadBuffer->Attachment[BUFFER_DEPTH].Renderbuffer;
    stencilRb = ctx->ReadBuffer->Attachment[BUFFER_STENCIL].Renderbuffer;
diff --git a/src/mesa/vbo/vbo_context.c b/src/mesa/vbo/vbo_context.c
index 65998e7ff7..f64f59d11e 100644
--- a/src/mesa/vbo/vbo_context.c
+++ b/src/mesa/vbo/vbo_context.c
@@ -241,6 +241,7 @@ void _vbo_DestroyContext( GLcontext *ctx )
       ctx->aelt_context = NULL;
    }
 
+   vbo_exec_destroy(ctx);
    FREE(vbo_context(ctx));
    ctx->swtnl_im = NULL;
 }
diff --git a/src/mesa/vbo/vbo_exec_draw.c b/src/mesa/vbo/vbo_exec_draw.c
index 2b52d7a1ff..d8f167b357 100644
--- a/src/mesa/vbo/vbo_exec_draw.c
+++ b/src/mesa/vbo/vbo_exec_draw.c
@@ -218,7 +218,7 @@ void vbo_exec_vtx_flush( struct vbo_exec_context *exec )
 				       exec->vtx.prim_count,
 				       NULL,
 				       0,
-				       exec->vtx.vert_count );
+				       exec->vtx.vert_count - 1);
       }
    }
 
diff --git a/src/mesa/vbo/vbo_save_draw.c b/src/mesa/vbo/vbo_save_draw.c
index a50d1cd904..bc59a0d821 100644
--- a/src/mesa/vbo/vbo_save_draw.c
+++ b/src/mesa/vbo/vbo_save_draw.c
@@ -218,7 +218,7 @@ void vbo_save_playback_vertex_list( GLcontext *ctx, void *data )
 				    node->prim_count,
 				    NULL,
 				    0,	/* Node is a VBO, so this is ok */
-				    node->count );
+				    node->count - 1);
    }
 
    /* Copy to current?