Merge remote branch 'main/master' into radeon-rewrite

Conflicts:
	src/mesa/drivers/dri/r300/r300_cmdbuf.c
	src/mesa/drivers/dri/r300/r300_state.c
	src/mesa/drivers/dri/r300/r300_swtcl.c
	src/mesa/drivers/dri/r300/radeon_ioctl.c
	src/mesa/drivers/dri/radeon/radeon_screen.c

101 files changed, 4802 insertions, 996 deletions

diff --git a/src/gallium/auxiliary/Makefile b/src/gallium/auxiliary/Makefile
index eaa0f2fe4e..5446eb68a9 100644
--- a/src/gallium/auxiliary/Makefile
+++ b/src/gallium/auxiliary/Makefile
@@ -1,20 +1,12 @@
+# src/gallium/auxiliary/Makefile
 TOP = ../../..
 include $(TOP)/configs/current
 
-
 SUBDIRS = $(GALLIUM_AUXILIARY_DIRS)
 
-
-default: subdirs
-
-
-subdirs:
+default install clean:
 	@for dir in $(SUBDIRS) ; do \
 		if [ -d $$dir ] ; then \
-			(cd $$dir && $(MAKE)) || exit 1 ; \
+			(cd $$dir && $(MAKE) $@) || exit 1; \
 		fi \
 	done
-
-
-clean:
-	rm -f `find . -name \*.[oa]`
diff --git a/src/gallium/auxiliary/cso_cache/Makefile b/src/gallium/auxiliary/cso_cache/Makefile
index 6bd6602088..8726afcd94 100644
--- a/src/gallium/auxiliary/cso_cache/Makefile
+++ b/src/gallium/auxiliary/cso_cache/Makefile
@@ -9,6 +9,3 @@ C_SOURCES = \
 	cso_hash.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/cso_cache/cso_cache.c b/src/gallium/auxiliary/cso_cache/cso_cache.c
index 6b1754ea00..0bc77a5728 100644
--- a/src/gallium/auxiliary/cso_cache/cso_cache.c
+++ b/src/gallium/auxiliary/cso_cache/cso_cache.c
@@ -28,7 +28,7 @@
 /* Authors:  Zack Rusin <zack@tungstengraphics.com>
  */
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "util/u_memory.h"
 
diff --git a/src/gallium/auxiliary/cso_cache/cso_context.c b/src/gallium/auxiliary/cso_cache/cso_context.c
index a9157aad71..f388bf5d95 100644
--- a/src/gallium/auxiliary/cso_cache/cso_context.c
+++ b/src/gallium/auxiliary/cso_cache/cso_context.c
@@ -37,7 +37,6 @@
 
 #include "pipe/p_state.h"
 #include "util/u_memory.h"
-#include "pipe/p_inlines.h"
 #include "tgsi/tgsi_parse.h"
 
 #include "cso_cache/cso_context.h"
diff --git a/src/gallium/auxiliary/cso_cache/cso_hash.c b/src/gallium/auxiliary/cso_cache/cso_hash.c
index 4e7664f9bf..288cef7b6f 100644
--- a/src/gallium/auxiliary/cso_cache/cso_hash.c
+++ b/src/gallium/auxiliary/cso_cache/cso_hash.c
@@ -30,7 +30,7 @@
   *   Zack Rusin <zack@tungstengraphics.com>
   */
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_memory.h"
 
 #include "cso_hash.h"
diff --git a/src/gallium/auxiliary/draw/Makefile b/src/gallium/auxiliary/draw/Makefile
index bdbf5a08ed..5041dcc072 100644
--- a/src/gallium/auxiliary/draw/Makefile
+++ b/src/gallium/auxiliary/draw/Makefile
@@ -43,8 +43,4 @@ C_SOURCES = \
 	draw_vs_ppc.c  \
 	draw_vs_sse.c 
 
-
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/draw/draw_context.c b/src/gallium/auxiliary/draw/draw_context.c
index 7bd4a2e221..a4f1fcddc1 100644
--- a/src/gallium/auxiliary/draw/draw_context.c
+++ b/src/gallium/auxiliary/draw/draw_context.c
@@ -127,7 +127,7 @@ void draw_set_rasterizer_state( struct draw_context *draw,
 
    draw->rasterizer = raster;
    draw->bypass_clipping =
-      ((draw->rasterizer && draw->rasterizer->bypass_clipping) ||
+      ((draw->rasterizer && draw->rasterizer->bypass_vs_clip_and_viewport) ||
        draw->driver.bypass_clipping);
 }
 
@@ -138,8 +138,9 @@ void draw_set_driver_clipping( struct draw_context *draw,
    draw_do_flush( draw, DRAW_FLUSH_STATE_CHANGE );
 
    draw->driver.bypass_clipping = bypass_clipping;
-   draw->bypass_clipping = (draw->rasterizer->bypass_clipping || 
-                            draw->driver.bypass_clipping);
+   draw->bypass_clipping =
+      ((draw->rasterizer && draw->rasterizer->bypass_vs_clip_and_viewport) ||
+       draw->driver.bypass_clipping);
 }
 
 
diff --git a/src/gallium/auxiliary/draw/draw_pipe_aaline.c b/src/gallium/auxiliary/draw/draw_pipe_aaline.c
index 20841bb5d6..ca69f0b95e 100644
--- a/src/gallium/auxiliary/draw/draw_pipe_aaline.c
+++ b/src/gallium/auxiliary/draw/draw_pipe_aaline.c
@@ -32,7 +32,6 @@
  */
 
 
-#include "pipe/p_inlines.h"
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_shader_tokens.h"
@@ -410,7 +409,7 @@ aaline_create_texture(struct aaline_stage *aaline)
     * texels which are zero.  Special case the 1x1 and 2x2 levels.
     */
    for (level = 0; level <= MAX_TEXTURE_LEVEL; level++) {
-      struct pipe_surface *surface;
+      struct pipe_transfer *transfer;
       const uint size = aaline->texture->width[level];
       ubyte *data;
       uint i, j;
@@ -419,9 +418,9 @@ aaline_create_texture(struct aaline_stage *aaline)
 
       /* This texture is new, no need to flush. 
        */
-      surface = screen->get_tex_surface(screen, aaline->texture, 0, level, 0,
-                                        PIPE_BUFFER_USAGE_CPU_WRITE);
-      data = screen->surface_map(screen, surface, PIPE_BUFFER_USAGE_CPU_WRITE);
+      transfer = screen->get_tex_transfer(screen, aaline->texture, 0, level, 0,
+                                         PIPE_TRANSFER_WRITE, 0, 0, size, size);
+      data = screen->transfer_map(screen, transfer);
       if (data == NULL)
          return FALSE;
 
@@ -440,13 +439,13 @@ aaline_create_texture(struct aaline_stage *aaline)
             else {
                d = 255;
             }
-            data[i * surface->stride + j] = d;
+            data[i * transfer->stride + j] = d;
          }
       }
 
       /* unmap */
-      screen->surface_unmap(screen, surface);
-      screen->tex_surface_release(screen, &surface);
+      screen->transfer_unmap(screen, transfer);
+      screen->tex_transfer_destroy(transfer);
    }
    return TRUE;
 }
@@ -728,7 +727,7 @@ aaline_destroy(struct draw_stage *stage)
       aaline->pipe->delete_sampler_state(aaline->pipe, aaline->sampler_cso);
 
    if (aaline->texture)
-      pipe_texture_release(&aaline->texture);
+      pipe_texture_reference(&aaline->texture, NULL);
 
    draw_free_temp_verts( stage );
 
diff --git a/src/gallium/auxiliary/draw/draw_pipe_aapoint.c b/src/gallium/auxiliary/draw/draw_pipe_aapoint.c
index 2c1cacbdb4..5008086cf2 100644
--- a/src/gallium/auxiliary/draw/draw_pipe_aapoint.c
+++ b/src/gallium/auxiliary/draw/draw_pipe_aapoint.c
@@ -38,7 +38,6 @@
  */
 
 
-#include "pipe/p_inlines.h"
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_shader_tokens.h"
diff --git a/src/gallium/auxiliary/draw/draw_pipe_pstipple.c b/src/gallium/auxiliary/draw/draw_pipe_pstipple.c
index a0f9716dac..04e59152c5 100644
--- a/src/gallium/auxiliary/draw/draw_pipe_pstipple.c
+++ b/src/gallium/auxiliary/draw/draw_pipe_pstipple.c
@@ -34,7 +34,6 @@
  */
 
 
-#include "pipe/p_inlines.h"
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_shader_tokens.h"
@@ -372,7 +371,7 @@ pstip_update_texture(struct pstip_stage *pstip)
    static const uint bit31 = 1 << 31;
    struct pipe_context *pipe = pstip->pipe;
    struct pipe_screen *screen = pipe->screen;
-   struct pipe_surface *surface;
+   struct pipe_transfer *transfer;
    const uint *stipple = pstip->state.stipple->stipple;
    uint i, j;
    ubyte *data;
@@ -381,10 +380,9 @@ pstip_update_texture(struct pstip_stage *pstip)
     */
    pipe->flush( pipe, PIPE_FLUSH_TEXTURE_CACHE, NULL );
 
-   surface = screen->get_tex_surface(screen, pstip->texture, 0, 0, 0,
-                                     PIPE_BUFFER_USAGE_CPU_WRITE);
-   data = screen->surface_map(screen, surface,
-                              PIPE_BUFFER_USAGE_CPU_WRITE);
+   transfer = screen->get_tex_transfer(screen, pstip->texture, 0, 0, 0,
+                                       PIPE_TRANSFER_WRITE, 0, 0, 32, 32);
+   data = screen->transfer_map(screen, transfer);
 
    /*
     * Load alpha texture.
@@ -396,18 +394,18 @@ pstip_update_texture(struct pstip_stage *pstip)
       for (j = 0; j < 32; j++) {
          if (stipple[i] & (bit31 >> j)) {
             /* fragment "on" */
-            data[i * surface->stride + j] = 0;
+            data[i * transfer->stride + j] = 0;
          }
          else {
             /* fragment "off" */
-            data[i * surface->stride + j] = 255;
+            data[i * transfer->stride + j] = 255;
          }
       }
    }
 
    /* unmap */
-   screen->surface_unmap(screen, surface);
-   screen->tex_surface_release(screen, &surface);
+   screen->transfer_unmap(screen, transfer);
+   screen->tex_transfer_destroy(transfer);
 }
 
 
@@ -573,7 +571,7 @@ pstip_destroy(struct draw_stage *stage)
 
    pstip->pipe->delete_sampler_state(pstip->pipe, pstip->sampler_cso);
 
-   pipe_texture_release(&pstip->texture);
+   pipe_texture_reference(&pstip->texture, NULL);
 
    draw_free_temp_verts( stage );
    FREE( stage );
diff --git a/src/gallium/auxiliary/draw/draw_pipe_vbuf.c b/src/gallium/auxiliary/draw/draw_pipe_vbuf.c
index 9153bc2f86..12325d30d6 100644
--- a/src/gallium/auxiliary/draw/draw_pipe_vbuf.c
+++ b/src/gallium/auxiliary/draw/draw_pipe_vbuf.c
@@ -29,12 +29,12 @@
  * \file
  * Vertex buffer drawing stage.
  * 
- * \author José Fonseca <jrfonsec@tungstengraphics.com>
+ * \author Jose Fonseca <jrfonsec@tungstengraphics.com>
  * \author Keith Whitwell <keith@tungstengraphics.com>
  */
 
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -93,7 +93,6 @@ vbuf_stage( struct draw_stage *stage )
 }
 
 
-static void vbuf_flush_indices( struct vbuf_stage *vbuf );
 static void vbuf_flush_vertices( struct vbuf_stage *vbuf );
 static void vbuf_alloc_vertices( struct vbuf_stage *vbuf );
 
@@ -109,13 +108,12 @@ overflow( void *map, void *ptr, unsigned bytes, unsigned bufsz )
 static INLINE void 
 check_space( struct vbuf_stage *vbuf, unsigned nr )
 {
-   if (vbuf->nr_vertices + nr > vbuf->max_vertices ) {
-      vbuf_flush_vertices(vbuf);
-      vbuf_alloc_vertices(vbuf);
+   if (vbuf->nr_vertices + nr > vbuf->max_vertices ||
+       vbuf->nr_indices + nr > vbuf->max_indices)
+   {
+      vbuf_flush_vertices( vbuf );
+      vbuf_alloc_vertices( vbuf );
    }
-
-   if (vbuf->nr_indices + nr > vbuf->max_indices )
-      vbuf_flush_indices(vbuf);
 }
 
 
@@ -202,7 +200,7 @@ vbuf_point( struct draw_stage *stage,
  * will be flushed if needed and a new one allocated.
  */
 static void
-vbuf_set_prim( struct vbuf_stage *vbuf, uint prim )
+vbuf_start_prim( struct vbuf_stage *vbuf, uint prim )
 {
    struct translate_key hw_key;
    unsigned dst_offset;
@@ -217,11 +215,7 @@ vbuf_set_prim( struct vbuf_stage *vbuf, uint prim )
     * state change.
     */
    vbuf->vinfo = vbuf->render->get_vertex_info(vbuf->render);
-
-   if (vbuf->vertex_size != vbuf->vinfo->size * sizeof(float)) {
-      vbuf_flush_vertices(vbuf);
-      vbuf->vertex_size = vbuf->vinfo->size * sizeof(float);
-   }
+   vbuf->vertex_size = vbuf->vinfo->size * sizeof(float);
 
    /* Translate from pipeline vertices to hw vertices.
     */
@@ -294,8 +288,8 @@ vbuf_set_prim( struct vbuf_stage *vbuf, uint prim )
 
    /* Allocate new buffer?
     */
-   if (!vbuf->vertices)
-      vbuf_alloc_vertices(vbuf);
+   assert(vbuf->vertices == NULL);
+   vbuf_alloc_vertices(vbuf);
 }
 
 
@@ -305,9 +299,9 @@ vbuf_first_tri( struct draw_stage *stage,
 {
    struct vbuf_stage *vbuf = vbuf_stage( stage );
 
-   vbuf_flush_indices( vbuf );   
+   vbuf_flush_vertices( vbuf );
+   vbuf_start_prim(vbuf, PIPE_PRIM_TRIANGLES);
    stage->tri = vbuf_tri;
-   vbuf_set_prim(vbuf, PIPE_PRIM_TRIANGLES);
    stage->tri( stage, prim );
 }
 
@@ -318,9 +312,9 @@ vbuf_first_line( struct draw_stage *stage,
 {
    struct vbuf_stage *vbuf = vbuf_stage( stage );
 
-   vbuf_flush_indices( vbuf );
+   vbuf_flush_vertices( vbuf );
+   vbuf_start_prim(vbuf, PIPE_PRIM_LINES);
    stage->line = vbuf_line;
-   vbuf_set_prim(vbuf, PIPE_PRIM_LINES);
    stage->line( stage, prim );
 }
 
@@ -331,53 +325,42 @@ vbuf_first_point( struct draw_stage *stage,
 {
    struct vbuf_stage *vbuf = vbuf_stage( stage );
 
-   vbuf_flush_indices( vbuf );
+   vbuf_flush_vertices(vbuf);
+   vbuf_start_prim(vbuf, PIPE_PRIM_POINTS);
    stage->point = vbuf_point;
-   vbuf_set_prim(vbuf, PIPE_PRIM_POINTS);
    stage->point( stage, prim );
 }
 
 
-static void 
-vbuf_flush_indices( struct vbuf_stage *vbuf ) 
-{
-   if(!vbuf->nr_indices)
-      return;
-   
-   assert((uint) (vbuf->vertex_ptr - vbuf->vertices) == 
-          vbuf->nr_vertices * vbuf->vertex_size / sizeof(unsigned));
-
-   vbuf->render->draw(vbuf->render, vbuf->indices, vbuf->nr_indices);
-   
-   vbuf->nr_indices = 0;
-}
-
 
 /**
  * Flush existing vertex buffer and allocate a new one.
- * 
- * XXX: We separate flush-on-index-full and flush-on-vb-full, but may 
- * raise issues uploading vertices if the hardware wants to flush when
- * we flush.
  */
 static void 
 vbuf_flush_vertices( struct vbuf_stage *vbuf )
 {
-   if(vbuf->vertices) {      
-      vbuf_flush_indices(vbuf);
-      
+   if(vbuf->vertices) {
+
+      vbuf->render->unmap_vertices( vbuf->render, 0, vbuf->nr_vertices - 1 );
+
+      if (vbuf->nr_indices) 
+      {
+         vbuf->render->draw(vbuf->render, 
+                            vbuf->indices, 
+                            vbuf->nr_indices );
+   
+         vbuf->nr_indices = 0;
+      }
+     
       /* Reset temporary vertices ids */
       if(vbuf->nr_vertices)
 	 draw_reset_vertex_ids( vbuf->stage.draw );
       
       /* Free the vertex buffer */
-      vbuf->render->release_vertices(vbuf->render,
-                                     vbuf->vertices,
-                                     vbuf->vertex_size,
-                                     vbuf->nr_vertices);
+      vbuf->render->release_vertices( vbuf->render );
+
       vbuf->max_vertices = vbuf->nr_vertices = 0;
       vbuf->vertex_ptr = vbuf->vertices = NULL;
-      
    }
 }
    
@@ -402,9 +385,12 @@ vbuf_alloc_vertices( struct vbuf_stage *vbuf )
     * and it will flush itself if necessary to do so.  If this does
     * fail, we are basically without usable hardware.
     */
-   vbuf->vertices = (uint *) vbuf->render->allocate_vertices(vbuf->render,
-							     (ushort) vbuf->vertex_size,
-							     (ushort) vbuf->max_vertices);
+   vbuf->render->allocate_vertices(vbuf->render,
+                                   (ushort) vbuf->vertex_size,
+                                   (ushort) vbuf->max_vertices);
+
+   vbuf->vertices = (uint *) vbuf->render->map_vertices( vbuf->render );
+   
    vbuf->vertex_ptr = vbuf->vertices;
 }
 
@@ -415,14 +401,11 @@ vbuf_flush( struct draw_stage *stage, unsigned flags )
 {
    struct vbuf_stage *vbuf = vbuf_stage( stage );
 
-   vbuf_flush_indices( vbuf );
+   vbuf_flush_vertices( vbuf );
 
    stage->point = vbuf_first_point;
    stage->line = vbuf_first_line;
    stage->tri = vbuf_first_tri;
-
-   if (flags & DRAW_FLUSH_BACKEND)
-      vbuf_flush_vertices( vbuf );
 }
 
 
diff --git a/src/gallium/auxiliary/draw/draw_pt.c b/src/gallium/auxiliary/draw/draw_pt.c
index 4e5ffa0930..9ea0cbe599 100644
--- a/src/gallium/auxiliary/draw/draw_pt.c
+++ b/src/gallium/auxiliary/draw/draw_pt.c
@@ -87,7 +87,7 @@ draw_pt_arrays(struct draw_context *draw,
          opt |= PT_CLIPTEST;
       }
       
-      if (!draw->rasterizer->bypass_vs) {
+      if (!draw->rasterizer->bypass_vs_clip_and_viewport) {
          opt |= PT_SHADE;
       }
    }
diff --git a/src/gallium/auxiliary/draw/draw_pt.h b/src/gallium/auxiliary/draw/draw_pt.h
index c02f229110..6f3e1e0289 100644
--- a/src/gallium/auxiliary/draw/draw_pt.h
+++ b/src/gallium/auxiliary/draw/draw_pt.h
@@ -173,9 +173,7 @@ void draw_pt_emit( struct pt_emit *emit,
 
 void draw_pt_emit_linear( struct pt_emit *emit,
                           const float (*vertex_data)[4],
-                          unsigned vertex_count,
                           unsigned stride,
-                          unsigned start,
                           unsigned count );
 
 void draw_pt_emit_destroy( struct pt_emit *emit );
@@ -217,7 +215,7 @@ boolean draw_pt_post_vs_run( struct pt_post_vs *pvs,
 
 void draw_pt_post_vs_prepare( struct pt_post_vs *pvs,
 			      boolean bypass_clipping,
-			      boolean identity_viewport,
+			      boolean bypass_viewport,
 			      boolean opengl );
 
 struct pt_post_vs *draw_pt_post_vs_create( struct draw_context *draw );
diff --git a/src/gallium/auxiliary/draw/draw_pt_emit.c b/src/gallium/auxiliary/draw/draw_pt_emit.c
index 232dfdaed2..064e16c295 100644
--- a/src/gallium/auxiliary/draw/draw_pt_emit.c
+++ b/src/gallium/auxiliary/draw/draw_pt_emit.c
@@ -165,6 +165,9 @@ void draw_pt_emit( struct pt_emit *emit,
     */
    draw_do_flush( draw, DRAW_FLUSH_BACKEND );
 
+   if (vertex_count == 0)
+      return;
+
    if (vertex_count >= UNDEFINED_VERTEX_ID) {
       assert(0);
       return;
@@ -178,9 +181,11 @@ void draw_pt_emit( struct pt_emit *emit,
       return;
    }
 
-   hw_verts = render->allocate_vertices(render,
-					(ushort)translate->key.output_stride,
-					(ushort)vertex_count);
+   render->allocate_vertices(render,
+                             (ushort)translate->key.output_stride,
+                             (ushort)vertex_count);
+
+   hw_verts = render->map_vertices( render );
    if (!hw_verts) {
       assert(0);
       return;
@@ -201,22 +206,21 @@ void draw_pt_emit( struct pt_emit *emit,
 		   vertex_count,
 		   hw_verts );
 
+   render->unmap_vertices( render, 
+                           0, 
+                           vertex_count - 1 );
+
    render->draw(render,
 		elts,
 		count);
 
-   render->release_vertices(render,
-			    hw_verts,
-			    translate->key.output_stride,
-			    vertex_count);
+   render->release_vertices(render);
 }
 
 
 void draw_pt_emit_linear(struct pt_emit *emit,
                          const float (*vertex_data)[4],
-                         unsigned vertex_count,
                          unsigned stride,
-                         unsigned start,
                          unsigned count)
 {
    struct draw_context *draw = emit->draw;
@@ -231,26 +235,23 @@ void draw_pt_emit_linear(struct pt_emit *emit,
     */
    draw_do_flush( draw, DRAW_FLUSH_BACKEND );
 
-   if (count >= UNDEFINED_VERTEX_ID) {
-      assert(0);
-      return;
-   }
+   if (count >= UNDEFINED_VERTEX_ID)
+      goto fail;
 
    /* XXX: and work out some way to coordinate the render primitive
     * between vbuf.c and here...
     */
-   if (!draw->render->set_primitive(draw->render, emit->prim)) {
-      assert(0);
-      return;
-   }
+   if (!draw->render->set_primitive(draw->render, emit->prim)) 
+      goto fail;
 
-   hw_verts = render->allocate_vertices(render,
-					(ushort)translate->key.output_stride,
-					(ushort)count);
-   if (!hw_verts) {
-      assert(0);
-      return;
-   }
+   if (!render->allocate_vertices(render,
+                                  (ushort)translate->key.output_stride,
+                                  (ushort)count))
+      goto fail;
+
+   hw_verts = render->map_vertices( render );
+   if (!hw_verts)
+      goto fail;
 
    translate->set_buffer(translate, 0,
 			 vertex_data, stride);
@@ -261,12 +262,12 @@ void draw_pt_emit_linear(struct pt_emit *emit,
 
    translate->run(translate,
                   0,
-                  vertex_count,
+                  count,
                   hw_verts);
 
    if (0) {
       unsigned i;
-      for (i = 0; i < vertex_count; i++) {
+      for (i = 0; i < count; i++) {
          debug_printf("\n\n%s vertex %d:\n", __FUNCTION__, i);
          draw_dump_emitted_vertex( emit->vinfo, 
                                    (const uint8_t *)hw_verts + 
@@ -274,13 +275,17 @@ void draw_pt_emit_linear(struct pt_emit *emit,
       }
    }
 
+   render->unmap_vertices( render, 0, count - 1 );
+
+   render->draw_arrays(render, 0, count);
+
+   render->release_vertices(render);
 
-   render->draw_arrays(render, start, count);
+   return;
 
-   render->release_vertices(render,
-			    hw_verts,
-			    translate->key.output_stride,
-			    vertex_count);
+fail:
+   assert(0);
+   return;
 }
 
 struct pt_emit *draw_pt_emit_create( struct draw_context *draw )
diff --git a/src/gallium/auxiliary/draw/draw_pt_fetch_emit.c b/src/gallium/auxiliary/draw/draw_pt_fetch_emit.c
index dcb7744b17..6b7d02a19b 100644
--- a/src/gallium/auxiliary/draw/draw_pt_fetch_emit.c
+++ b/src/gallium/auxiliary/draw/draw_pt_fetch_emit.c
@@ -234,9 +234,11 @@ static void fetch_emit_run( struct draw_pt_middle_end *middle,
       return;
    }
 
-   hw_verts = draw->render->allocate_vertices( draw->render,
-                                               (ushort)feme->translate->key.output_stride,
-                                               (ushort)fetch_count );
+   draw->render->allocate_vertices( draw->render,
+                                    (ushort)feme->translate->key.output_stride,
+                                    (ushort)fetch_count );
+
+   hw_verts = draw->render->map_vertices( draw->render );
    if (!hw_verts) {
       assert(0);
       return;
@@ -259,6 +261,10 @@ static void fetch_emit_run( struct draw_pt_middle_end *middle,
       }
    }
 
+   draw->render->unmap_vertices( draw->render, 
+                                 0, 
+                                 (ushort)(fetch_count - 1) );
+
    /* XXX: Draw arrays path to avoid re-emitting index list again and
     * again.
     */
@@ -268,10 +274,7 @@ static void fetch_emit_run( struct draw_pt_middle_end *middle,
 
    /* Done -- that was easy, wasn't it: 
     */
-   draw->render->release_vertices( draw->render, 
-                                   hw_verts, 
-                                   feme->translate->key.output_stride, 
-                                   fetch_count );
+   draw->render->release_vertices( draw->render );
 
 }
 
@@ -288,18 +291,17 @@ static void fetch_emit_run_linear( struct draw_pt_middle_end *middle,
     */
    draw_do_flush( draw, DRAW_FLUSH_BACKEND );
 
-   if (count >= UNDEFINED_VERTEX_ID) {
-      assert(0);
-      return;
-   }
+   if (count >= UNDEFINED_VERTEX_ID) 
+      goto fail;
 
-   hw_verts = draw->render->allocate_vertices( draw->render,
-                                               (ushort)feme->translate->key.output_stride,
-                                               (ushort)count );
-   if (!hw_verts) {
-      assert(0);
-      return;
-   }
+   if (!draw->render->allocate_vertices( draw->render,
+                                         (ushort)feme->translate->key.output_stride,
+                                         (ushort)count )) 
+      goto fail;
+
+   hw_verts = draw->render->map_vertices( draw->render );
+   if (!hw_verts) 
+      goto fail;
 
    /* Single routine to fetch vertices and emit HW verts.
     */
@@ -317,20 +319,21 @@ static void fetch_emit_run_linear( struct draw_pt_middle_end *middle,
       }
    }
 
+   draw->render->unmap_vertices( draw->render, 0, count - 1 );
+
    /* XXX: Draw arrays path to avoid re-emitting index list again and
     * again.
     */
-   draw->render->draw_arrays( draw->render,
-                              0, /*start*/
-                              count );
+   draw->render->draw_arrays( draw->render, 0, count );
 
    /* Done -- that was easy, wasn't it:
     */
-   draw->render->release_vertices( draw->render,
-                                   hw_verts,
-                                   feme->translate->key.output_stride,
-                                   count );
+   draw->render->release_vertices( draw->render );
+   return;
 
+fail:
+   assert(0);
+   return;
 }
 
 
@@ -351,9 +354,12 @@ static boolean fetch_emit_run_linear_elts( struct draw_pt_middle_end *middle,
    if (count >= UNDEFINED_VERTEX_ID)
       return FALSE;
 
-   hw_verts = draw->render->allocate_vertices( draw->render,
-                                               (ushort)feme->translate->key.output_stride,
-                                               (ushort)count );
+   if (!draw->render->allocate_vertices( draw->render,
+                                         (ushort)feme->translate->key.output_stride,
+                                         (ushort)count ))
+      return FALSE;
+
+   hw_verts = draw->render->map_vertices( draw->render );
    if (!hw_verts) 
       return FALSE;
 
@@ -364,6 +370,8 @@ static boolean fetch_emit_run_linear_elts( struct draw_pt_middle_end *middle,
                          count,
                          hw_verts );
 
+   draw->render->unmap_vertices( draw->render, 0, (ushort)(count - 1) );
+
    /* XXX: Draw arrays path to avoid re-emitting index list again and
     * again.
     */
@@ -373,10 +381,7 @@ static boolean fetch_emit_run_linear_elts( struct draw_pt_middle_end *middle,
 
    /* Done -- that was easy, wasn't it:
     */
-   draw->render->release_vertices( draw->render,
-                                   hw_verts,
-                                   feme->translate->key.output_stride,
-                                   count );
+   draw->render->release_vertices( draw->render );
 
    return TRUE;
 }
diff --git a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_emit.c b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_emit.c
index 84ffe3296a..44147aed9b 100644
--- a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_emit.c
+++ b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_emit.c
@@ -101,7 +101,8 @@ static void fse_prepare( struct draw_pt_middle_end *middle,
    fse->key.nr_elements = MAX2(fse->key.nr_outputs,     /* outputs - translate to hw format */
                                fse->key.nr_inputs);     /* inputs - fetch from api format */
 
-   fse->key.viewport = !draw->identity_viewport;
+   fse->key.viewport = (!draw->rasterizer->bypass_vs_clip_and_viewport &&
+                        !draw->identity_viewport);
    fse->key.clip = !draw->bypass_clipping;
    fse->key.const_vbuffers = 0;
 
@@ -234,19 +235,17 @@ static void fse_run_linear( struct draw_pt_middle_end *middle,
     */
    draw_do_flush( draw, DRAW_FLUSH_BACKEND );
 
-   if (count >= UNDEFINED_VERTEX_ID) {
-      assert(0);
-      return;
-   }
+   if (count >= UNDEFINED_VERTEX_ID) 
+      goto fail;
 
-   hw_verts = draw->render->allocate_vertices( draw->render,
-                                               (ushort)fse->key.output_stride,
-                                               (ushort)count );
+   if (!draw->render->allocate_vertices( draw->render,
+                                         (ushort)fse->key.output_stride,
+                                         (ushort)count ))
+      goto fail;
 
-   if (!hw_verts) {
-      assert(0);
-      return;
-   }
+   hw_verts = draw->render->map_vertices( draw->render );
+   if (!hw_verts)
+      goto fail;
 
    /* Single routine to fetch vertices, run shader and emit HW verts.
     * Clipping is done elsewhere -- either by the API or on hardware,
@@ -256,13 +255,7 @@ static void fse_run_linear( struct draw_pt_middle_end *middle,
                             start, count,
                             hw_verts );
 
-   /* Draw arrays path to avoid re-emitting index list again and
-    * again.
-    */
-   draw->render->draw_arrays( draw->render,
-                              0,
-                              count );
-   
+
    if (0) {
       unsigned i;
       for (i = 0; i < count; i++) {
@@ -274,12 +267,24 @@ static void fse_run_linear( struct draw_pt_middle_end *middle,
                                    (const uint8_t *)hw_verts + fse->key.output_stride * i );
       }
    }
+   
+   draw->render->unmap_vertices( draw->render, 0, (ushort)(count - 1) );
 
+   /* Draw arrays path to avoid re-emitting index list again and
+    * again.
+    */
+   draw->render->draw_arrays( draw->render,
+                              0,
+                              count );
+   
+
+   draw->render->release_vertices( draw->render );
+
+   return;
 
-   draw->render->release_vertices( draw->render, 
-				   hw_verts, 
-				   fse->key.output_stride, 
-				   count );
+fail:
+   assert(0);
+   return;
 }
 
 
@@ -298,18 +303,17 @@ fse_run(struct draw_pt_middle_end *middle,
     */
    draw_do_flush( draw, DRAW_FLUSH_BACKEND );
 
-   if (fetch_count >= UNDEFINED_VERTEX_ID) {
-      assert(0);
-      return;
-   }
+   if (fetch_count >= UNDEFINED_VERTEX_ID) 
+      goto fail;
 
-   hw_verts = draw->render->allocate_vertices( draw->render,
-                                               (ushort)fse->key.output_stride,
-                                               (ushort)fetch_count );
-   if (!hw_verts) {
-      assert(0);
-      return;
-   }
+   if (!draw->render->allocate_vertices( draw->render,
+                                         (ushort)fse->key.output_stride,
+                                         (ushort)fetch_count ))
+      goto fail;
+
+   hw_verts = draw->render->map_vertices( draw->render ); 
+   if (!hw_verts) 
+      goto fail;
          
 					
    /* Single routine to fetch vertices, run shader and emit HW verts.
@@ -319,9 +323,6 @@ fse_run(struct draw_pt_middle_end *middle,
                           fetch_count,
                           hw_verts );
 
-   draw->render->draw( draw->render, 
-                       draw_elts, 
-                       draw_count );
 
    if (0) {
       unsigned i;
@@ -333,12 +334,19 @@ fse_run(struct draw_pt_middle_end *middle,
       }
    }
 
+   draw->render->unmap_vertices( draw->render, 0, (ushort)(fetch_count - 1) );
+   
+   draw->render->draw( draw->render, 
+                       draw_elts, 
+                       draw_count );
+
 
-   draw->render->release_vertices( draw->render, 
-                                   hw_verts, 
-                                   fse->key.output_stride, 
-                                   fetch_count );
+   draw->render->release_vertices( draw->render );
+   return;
 
+fail:
+   assert(0);
+   return;
 }
 
 
@@ -360,13 +368,14 @@ static boolean fse_run_linear_elts( struct draw_pt_middle_end *middle,
    if (count >= UNDEFINED_VERTEX_ID)
       return FALSE;
 
-   hw_verts = draw->render->allocate_vertices( draw->render,
-                                               (ushort)fse->key.output_stride,
-                                               (ushort)count );
+   if (!draw->render->allocate_vertices( draw->render,
+                                         (ushort)fse->key.output_stride,
+                                         (ushort)count ))
+      return FALSE;
 
-   if (!hw_verts) {
+   hw_verts = draw->render->map_vertices( draw->render );
+   if (!hw_verts) 
       return FALSE;
-   }
 
    /* Single routine to fetch vertices, run shader and emit HW verts.
     * Clipping is done elsewhere -- either by the API or on hardware,
@@ -382,11 +391,9 @@ static boolean fse_run_linear_elts( struct draw_pt_middle_end *middle,
                        draw_count );
    
 
+   draw->render->unmap_vertices( draw->render, 0, (ushort)(count - 1) );
 
-   draw->render->release_vertices( draw->render, 
-				   hw_verts, 
-				   fse->key.output_stride, 
-				   count );
+   draw->render->release_vertices( draw->render );
 
    return TRUE;
 }
diff --git a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c
index ec3b41c320..11ac90fc56 100644
--- a/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c
+++ b/src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c
@@ -84,7 +84,8 @@ static void fetch_pipeline_prepare( struct draw_pt_middle_end *middle,
     */
    draw_pt_post_vs_prepare( fpme->post_vs,
 			    (boolean)draw->bypass_clipping,
-			    (boolean)draw->identity_viewport,
+			    (boolean)(draw->identity_viewport ||
+                                      draw->rasterizer->bypass_vs_clip_and_viewport),
 			    (boolean)draw->rasterizer->gl_rasterization_rules );
 			    
 
@@ -140,9 +141,9 @@ static void fetch_pipeline_run( struct draw_pt_middle_end *middle,
 		      (char *)pipeline_verts );
 
    /* Run the shader, note that this overwrites the data[] parts of
-    * the pipeline verts.  If there is no shader, ie a bypass shader,
-    * then the inputs == outputs, and are already in the correct
-    * place.
+    * the pipeline verts.  If there is no shader, eg if
+    * bypass_vs_clip_and_viewport, then the inputs == outputs, and are
+    * already in the correct place.
     */
    if (opt & PT_SHADE)
    {
@@ -216,9 +217,9 @@ static void fetch_pipeline_linear_run( struct draw_pt_middle_end *middle,
                              (char *)pipeline_verts );
 
    /* Run the shader, note that this overwrites the data[] parts of
-    * the pipeline verts.  If there is no shader, ie a bypass shader,
-    * then the inputs == outputs, and are already in the correct
-    * place.
+    * the pipeline verts.  If there is no shader, ie if
+    * bypass_vs_clip_and_viewport, then the inputs == outputs, and are
+    * already in the correct place.
     */
    if (opt & PT_SHADE)
    {
@@ -251,9 +252,7 @@ static void fetch_pipeline_linear_run( struct draw_pt_middle_end *middle,
    else {
       draw_pt_emit_linear( fpme->emit,
                            (const float (*)[4])pipeline_verts->data,
-                           count,
                            fpme->vertex_size,
-                           0, /*start*/
                            count );
    }
 
@@ -288,9 +287,9 @@ static boolean fetch_pipeline_linear_run_elts( struct draw_pt_middle_end *middle
                              (char *)pipeline_verts );
 
    /* Run the shader, note that this overwrites the data[] parts of
-    * the pipeline verts.  If there is no shader, ie a bypass shader,
-    * then the inputs == outputs, and are already in the correct
-    * place.
+    * the pipeline verts.  If there is no shader, ie if
+    * bypass_vs_clip_and_viewport, then the inputs == outputs, and are
+    * already in the correct place.
     */
    if (opt & PT_SHADE)
    {
diff --git a/src/gallium/auxiliary/draw/draw_pt_post_vs.c b/src/gallium/auxiliary/draw/draw_pt_post_vs.c
index 96dc706b99..00d7197b13 100644
--- a/src/gallium/auxiliary/draw/draw_pt_post_vs.c
+++ b/src/gallium/auxiliary/draw/draw_pt_post_vs.c
@@ -200,11 +200,11 @@ boolean draw_pt_post_vs_run( struct pt_post_vs *pvs,
 
 void draw_pt_post_vs_prepare( struct pt_post_vs *pvs,
 			      boolean bypass_clipping,
-			      boolean identity_viewport,
+			      boolean bypass_viewport,
 			      boolean opengl )
 {
    if (bypass_clipping) {
-      if (identity_viewport)
+      if (bypass_viewport)
 	 pvs->run = post_vs_none;
       else
 	 pvs->run = post_vs_viewport;
diff --git a/src/gallium/auxiliary/draw/draw_vbuf.h b/src/gallium/auxiliary/draw/draw_vbuf.h
index a1c4c14445..cccd3bf435 100644
--- a/src/gallium/auxiliary/draw/draw_vbuf.h
+++ b/src/gallium/auxiliary/draw/draw_vbuf.h
@@ -80,9 +80,14 @@ struct vbuf_render {
     * Hardware renderers will use ttm memory, others will just malloc
     * something.
     */
-   void *(*allocate_vertices)( struct vbuf_render *,
-			       ushort vertex_size,
-			       ushort nr_vertices );
+   boolean (*allocate_vertices)( struct vbuf_render *,
+                                 ushort vertex_size,
+                                 ushort nr_vertices );
+   
+   void *(*map_vertices)( struct vbuf_render * );
+   void (*unmap_vertices)( struct vbuf_render *, 
+                           ushort min_index,
+                           ushort max_index );
 
    /**
     * Notify the renderer of the current primitive when it changes.
@@ -109,10 +114,7 @@ struct vbuf_render {
    /**
     * Called when vbuf is done with this set of vertices:
     */
-   void (*release_vertices)( struct vbuf_render *,
-			     void *vertices, 
-			     unsigned vertex_size,
-			     unsigned vertices_used );
+   void (*release_vertices)( struct vbuf_render * );
 
    void (*destroy)( struct vbuf_render * );
 };
diff --git a/src/gallium/auxiliary/draw/draw_vs_aos.c b/src/gallium/auxiliary/draw/draw_vs_aos.c
index 0c693a4a65..9e37a26c1e 100644
--- a/src/gallium/auxiliary/draw/draw_vs_aos.c
+++ b/src/gallium/auxiliary/draw/draw_vs_aos.c
@@ -32,7 +32,7 @@
 #include "util/u_memory.h"
 #include "util/u_math.h"
 #include "pipe/p_shader_tokens.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "tgsi/tgsi_parse.h"
 #include "tgsi/tgsi_util.h"
 #include "tgsi/tgsi_exec.h"
@@ -143,7 +143,7 @@ static struct x86_reg get_reg_ptr(struct aos_compilation *cp,
       return x86_make_disp(aos_get_x86(cp, 1, X86_CONSTANTS), idx * 4 * sizeof(float));
 
    default:
-      ERROR(cp, "unknown reg file");
+      AOS_ERROR(cp, "unknown reg file");
       return x86_make_reg(0,0);
    }
 }
@@ -177,7 +177,7 @@ static void spill( struct aos_compilation *cp, unsigned idx )
        (cp->xmm[idx].file != TGSI_FILE_INPUT && /* inputs are fetched into xmm & set dirty */
         cp->xmm[idx].file != TGSI_FILE_OUTPUT &&
         cp->xmm[idx].file != TGSI_FILE_TEMPORARY)) {
-      ERROR(cp, "invalid spill");
+      AOS_ERROR(cp, "invalid spill");
       return;
    }
    else {
@@ -284,6 +284,15 @@ void aos_release_xmm_reg( struct aos_compilation *cp,
 }
 
 
+static void aos_soft_release_xmm( struct aos_compilation *cp,
+                                  struct x86_reg reg )
+{
+   if (reg.file == file_XMM) {
+      assert(cp->xmm[reg.idx].last_used == cp->insn_counter);
+      cp->xmm[reg.idx].last_used = cp->insn_counter - 1;
+   }
+}
+
 
      
 /* Mark an xmm reg as holding the current copy of a shader reg.
@@ -534,7 +543,7 @@ static struct x86_reg fetch_src( struct aos_compilation *cp,
       switch (swizzle) {
       case TGSI_EXTSWIZZLE_ZERO:
       case TGSI_EXTSWIZZLE_ONE:
-         ERROR(cp, "not supporting full swizzles yet in tgsi_aos_sse2");
+         AOS_ERROR(cp, "not supporting full swizzles yet in tgsi_aos_sse2");
          break;
 
       default:
@@ -555,7 +564,7 @@ static struct x86_reg fetch_src( struct aos_compilation *cp,
          break;
 
       default:
-         ERROR(cp, "unsupported sign-mode");
+         AOS_ERROR(cp, "unsupported sign-mode");
          break;
       }
    }
@@ -584,15 +593,17 @@ static struct x86_reg fetch_src( struct aos_compilation *cp,
          sse_mulps(cp->func, dst, tmp);
 
          aos_release_xmm_reg(cp, tmp.idx);
+         aos_soft_release_xmm(cp, imm_swz);
       }
       else if (negs) {
          struct x86_reg imm_negs = aos_get_internal_xmm(cp, IMM_NEGS);
          sse_mulps(cp->func, dst, imm_negs);
+         aos_soft_release_xmm(cp, imm_negs);
       }
 
 
       if (abs && abs != 0xf) {
-         ERROR(cp, "unsupported partial abs");
+         AOS_ERROR(cp, "unsupported partial abs");
       }
       else if (abs) {
          struct x86_reg neg = aos_get_internal(cp, IMM_NEGS);
@@ -603,8 +614,10 @@ static struct x86_reg fetch_src( struct aos_compilation *cp,
          sse_maxps(cp->func, dst, tmp);
 
          aos_release_xmm_reg(cp, tmp.idx);
+         aos_soft_release_xmm(cp, neg);
       }
 
+      aos_soft_release_xmm(cp, arg0);
       return dst;
    }
       
@@ -657,7 +670,7 @@ static void x87_fld_src( struct aos_compilation *cp,
       break;
 
    default:
-      ERROR(cp, "unsupported sign-mode");
+      AOS_ERROR(cp, "unsupported sign-mode");
       break;
    }
 }
@@ -1559,7 +1572,6 @@ static boolean emit_RCP( struct aos_compilation *cp, const struct tgsi_full_inst
  */
 static boolean emit_RSQ( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-
    if (0) {
       struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
       struct x86_reg r = aos_get_xmm_reg(cp);
@@ -1568,21 +1580,30 @@ static boolean emit_RSQ( struct aos_compilation *cp, const struct tgsi_full_inst
       return TRUE;
    }
    else {
-      struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-      struct x86_reg r = aos_get_xmm_reg(cp);
+      struct x86_reg arg0           = fetch_src(cp, &op->FullSrcRegisters[0]);
+      struct x86_reg r              = aos_get_xmm_reg(cp);
 
       struct x86_reg neg_half       = get_reg_ptr( cp, AOS_FILE_INTERNAL, IMM_RSQ );
       struct x86_reg one_point_five = x86_make_disp( neg_half, 4 );
       struct x86_reg src            = get_xmm_writable( cp, arg0 );
-      
-      sse_rsqrtss( cp->func, r, src  );             /* rsqrtss(a) */
-      sse_mulss(   cp->func, src, neg_half  );      /* -.5 * a */
-      sse_mulss(   cp->func, src,  r );             /* -.5 * a * r */
-      sse_mulss(   cp->func, src,  r );             /* -.5 * a * r * r */
-      sse_addss(   cp->func, src, one_point_five ); /* 1.5 - .5 * a * r * r */
-      sse_mulss(   cp->func, r,  src );             /* r * (1.5 - .5 * a * r * r) */
+      struct x86_reg neg            = aos_get_internal(cp, IMM_NEGS);
+      struct x86_reg tmp            = aos_get_xmm_reg(cp);
+
+      sse_movaps(cp->func, tmp, src);
+      sse_mulps(cp->func, tmp, neg);
+      sse_maxps(cp->func, tmp, src);
+   
+      sse_rsqrtss( cp->func, r, tmp  );             /* rsqrtss(a) */
+      sse_mulss(   cp->func, tmp, neg_half  );      /* -.5 * a */
+      sse_mulss(   cp->func, tmp,  r );             /* -.5 * a * r */
+      sse_mulss(   cp->func, tmp,  r );             /* -.5 * a * r * r */
+      sse_addss(   cp->func, tmp, one_point_five ); /* 1.5 - .5 * a * r * r */
+      sse_mulss(   cp->func, r,  tmp );             /* r * (1.5 - .5 * a * r * r) */
 
       store_scalar_dest(cp, &op->FullDstRegisters[0], r);
+
+      aos_release_xmm_reg(cp, tmp.idx);
+
       return TRUE;
    }
 }
diff --git a/src/gallium/auxiliary/draw/draw_vs_aos.h b/src/gallium/auxiliary/draw/draw_vs_aos.h
index 264387517b..2cf72ddf7b 100644
--- a/src/gallium/auxiliary/draw/draw_vs_aos.h
+++ b/src/gallium/auxiliary/draw/draw_vs_aos.h
@@ -204,7 +204,7 @@ struct x86_reg aos_get_internal_xmm( struct aos_compilation *cp,
                                      unsigned imm );
 
 
-#define ERROR(cp, msg)                                                  \
+#define AOS_ERROR(cp, msg)                                                  \
 do {                                                                    \
    if (0) debug_printf("%s: x86 translation failed: %s\n", __FUNCTION__, msg); \
    cp->error = 1;                                                       \
diff --git a/src/gallium/auxiliary/draw/draw_vs_aos_io.c b/src/gallium/auxiliary/draw/draw_vs_aos_io.c
index 39f75b50b7..a6eb37d128 100644
--- a/src/gallium/auxiliary/draw/draw_vs_aos_io.c
+++ b/src/gallium/auxiliary/draw/draw_vs_aos_io.c
@@ -199,7 +199,7 @@ static boolean load_input( struct aos_compilation *cp,
       emit_load_R8G8B8A8_UNORM(cp, dataXMM, src);
       break;
    default:
-      ERROR(cp, "unhandled input format");
+      AOS_ERROR(cp, "unhandled input format");
       return FALSE;
    }
 
@@ -410,7 +410,7 @@ static boolean emit_output( struct aos_compilation *cp,
       }
       break;
    default:
-      ERROR(cp, "unhandled output format");
+      AOS_ERROR(cp, "unhandled output format");
       return FALSE;
    }
 
diff --git a/src/gallium/auxiliary/gallivm/storagesoa.cpp b/src/gallium/auxiliary/gallivm/storagesoa.cpp
index e1e5cabcf5..4984ce985c 100644
--- a/src/gallium/auxiliary/gallivm/storagesoa.cpp
+++ b/src/gallium/auxiliary/gallivm/storagesoa.cpp
@@ -30,7 +30,7 @@
 #include "gallivm_p.h"
 
 #include "pipe/p_shader_tokens.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include <llvm/BasicBlock.h>
 #include <llvm/Module.h>
diff --git a/src/gallium/auxiliary/indices/Makefile b/src/gallium/auxiliary/indices/Makefile
index 8fa61d265e..f2ebc3f410 100644
--- a/src/gallium/auxiliary/indices/Makefile
+++ b/src/gallium/auxiliary/indices/Makefile
@@ -4,13 +4,13 @@ include $(TOP)/configs/current
 LIBNAME = indices
 
 C_SOURCES = \
-	u_indices_gen.c 
+	u_indices_gen.c \
+	u_unfilled_gen.c 
 
 include ../../Makefile.template
 
 u_indices_gen.c: u_indices_gen.py
 	python $< > $@
 
-
-symlinks:
-
+u_unfilled_gen.c: u_unfilled_gen.py
+	python $< > $@
diff --git a/src/gallium/auxiliary/indices/SConscript b/src/gallium/auxiliary/indices/SConscript
index 65a43a9f64..e5f7ee9484 100644
--- a/src/gallium/auxiliary/indices/SConscript
+++ b/src/gallium/auxiliary/indices/SConscript
@@ -7,11 +7,20 @@ env.CodeGenerate(
 	command = 'python $SCRIPT > $TARGET'
 )
 
+env.CodeGenerate(
+	target = 'u_unfilled_gen.c', 
+	script = 'u_unfilled_gen.py', 
+	source = [],
+	command = 'python $SCRIPT > $TARGET'
+)
+
 indices = env.ConvenienceLibrary(
 	target = 'indices',
 	source = [
 #               'u_indices.c',
+#               'u_unfilled_indices.c',
                'u_indices_gen.c',
+               'u_unfilled_gen.c',
 	])
 
 auxiliaries.insert(0, indices)
diff --git a/src/gallium/auxiliary/indices/u_indices.h b/src/gallium/auxiliary/indices/u_indices.h
index abf5a3037d..be522c6725 100644
--- a/src/gallium/auxiliary/indices/u_indices.h
+++ b/src/gallium/auxiliary/indices/u_indices.h
@@ -80,4 +80,27 @@ int u_index_generator( unsigned hw_mask,
                        u_generate_func *out_generate );
 
 
+void u_unfilled_init( void );
+
+int u_unfilled_translator( unsigned prim,
+                           unsigned in_index_size,
+                           unsigned nr,
+                           unsigned unfilled_mode,
+                           unsigned *out_prim,
+                           unsigned *out_index_size,
+                           unsigned *out_nr,
+                           u_translate_func *out_translate );
+
+int u_unfilled_generator( unsigned prim,
+                          unsigned start,
+                          unsigned nr,
+                          unsigned unfilled_mode,
+                          unsigned *out_prim,
+                          unsigned *out_index_size,
+                          unsigned *out_nr,
+                          u_generate_func *out_generate );
+
+
+
+
 #endif
diff --git a/src/gallium/auxiliary/indices/u_indices_gen.c b/src/gallium/auxiliary/indices/u_indices_gen.c
index 4c05b3eedb..3c981e5d7f 100644
--- a/src/gallium/auxiliary/indices/u_indices_gen.c
+++ b/src/gallium/auxiliary/indices/u_indices_gen.c
@@ -34,7 +34,7 @@
 #include "indices/u_indices.h"
 #include "indices/u_indices_priv.h"
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_defines.h"
 #include "util/u_memory.h"
 
diff --git a/src/gallium/auxiliary/indices/u_indices_gen.py b/src/gallium/auxiliary/indices/u_indices_gen.py
index 0dc58d0cd0..af63d09930 100644
--- a/src/gallium/auxiliary/indices/u_indices_gen.py
+++ b/src/gallium/auxiliary/indices/u_indices_gen.py
@@ -72,7 +72,7 @@ def prolog():
 #include "indices/u_indices.h"
 #include "indices/u_indices_priv.h"
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_defines.h"
 #include "util/u_memory.h"
 
diff --git a/src/gallium/auxiliary/indices/u_unfilled_gen.c b/src/gallium/auxiliary/indices/u_unfilled_gen.c
new file mode 100644
index 0000000000..fd082ebbb3
--- /dev/null
+++ b/src/gallium/auxiliary/indices/u_unfilled_gen.c
@@ -0,0 +1,1152 @@
+/* File automatically generated by u_unfilled_gen.py */
+
+/*
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * VMWARE AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+
+/**
+ * @file
+ * Functions to translate and generate index lists
+ */
+
+#include "indices/u_indices.h"
+#include "indices/u_indices_priv.h"
+#include "pipe/p_compiler.h"
+#include "util/u_debug.h"
+#include "pipe/p_defines.h"
+#include "util/u_memory.h"
+
+
+static unsigned out_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return OUT_UINT;
+   case 2: return OUT_USHORT;
+   default: assert(0); return OUT_USHORT;
+   }
+}
+
+static unsigned in_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return IN_UINT;
+   case 2: return IN_USHORT;
+   case 1: return IN_UBYTE;
+   default: assert(0); return IN_UBYTE;
+   }
+}
+
+
+static u_generate_func generate_line[OUT_COUNT][PRIM_COUNT];
+static u_translate_func translate_line[IN_COUNT][OUT_COUNT][PRIM_COUNT];
+
+
+static void generate_tris_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(i);
+   }
+}
+static void generate_tristrip_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1/*+(i&1)*/);
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (ushort)(i+1/*+(i&1)*/);
+      (out+j+2)[1] = (ushort)(i+2/*-(i&1)*/);
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (ushort)(i+2/*-(i&1)*/);
+      (out+j+4)[1] = (ushort)(i);
+   }
+}
+static void generate_trifan_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(0);
+   }
+}
+static void generate_quads_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (ushort)(i+0);
+      (out+j)[1] = (ushort)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(i+3);
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (ushort)(i+3);
+      (out+j+6)[1] = (ushort)(i+0);
+   }
+}
+static void generate_quadstrip_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (ushort)(i+2);
+      (out+j)[1] = (ushort)(i+0);
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (ushort)(i+0);
+      (out+j+2)[1] = (ushort)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (ushort)(i+1);
+      (out+j+4)[1] = (ushort)(i+3);
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (ushort)(i+3);
+      (out+j+6)[1] = (ushort)(i+2);
+   }
+}
+static void generate_polygon_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(0);
+   }
+}
+static void generate_tris_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(i);
+   }
+}
+static void generate_tristrip_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1/*+(i&1)*/);
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (uint)(i+1/*+(i&1)*/);
+      (out+j+2)[1] = (uint)(i+2/*-(i&1)*/);
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (uint)(i+2/*-(i&1)*/);
+      (out+j+4)[1] = (uint)(i);
+   }
+}
+static void generate_trifan_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(0);
+   }
+}
+static void generate_quads_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (uint)(i+0);
+      (out+j)[1] = (uint)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(i+3);
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (uint)(i+3);
+      (out+j+6)[1] = (uint)(i+0);
+   }
+}
+static void generate_quadstrip_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (uint)(i+2);
+      (out+j)[1] = (uint)(i+0);
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (uint)(i+0);
+      (out+j+2)[1] = (uint)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (uint)(i+1);
+      (out+j+4)[1] = (uint)(i+3);
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (uint)(i+3);
+      (out+j+6)[1] = (uint)(i+2);
+   }
+}
+static void generate_polygon_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(0);
+   }
+}
+static void translate_tris_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_trifan_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_quads_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (ushort)in[i+0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+0];
+   }
+}
+static void translate_quadstrip_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+0];
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (ushort)in[i+0];
+      (out+j+2)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (ushort)in[i+1];
+      (out+j+4)[1] = (ushort)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+2];
+   }
+}
+static void translate_polygon_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_tris_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_tristrip_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_trifan_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_quads_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (uint)in[i+0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+0];
+   }
+}
+static void translate_quadstrip_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+0];
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (uint)in[i+0];
+      (out+j+2)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (uint)in[i+1];
+      (out+j+4)[1] = (uint)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+2];
+   }
+}
+static void translate_polygon_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_tris_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_trifan_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_quads_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (ushort)in[i+0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+0];
+   }
+}
+static void translate_quadstrip_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+0];
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (ushort)in[i+0];
+      (out+j+2)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (ushort)in[i+1];
+      (out+j+4)[1] = (ushort)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+2];
+   }
+}
+static void translate_polygon_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_tris_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_tristrip_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_trifan_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_quads_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (uint)in[i+0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+0];
+   }
+}
+static void translate_quadstrip_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+0];
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (uint)in[i+0];
+      (out+j+2)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (uint)in[i+1];
+      (out+j+4)[1] = (uint)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+2];
+   }
+}
+static void translate_polygon_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_tris_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_trifan_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_quads_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (ushort)in[i+0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+0];
+   }
+}
+static void translate_quadstrip_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+0];
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (ushort)in[i+0];
+      (out+j+2)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (ushort)in[i+1];
+      (out+j+4)[1] = (ushort)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+2];
+   }
+}
+static void translate_polygon_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_tris_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1);
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_tristrip_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)i, (int)i+1/*+(i&1)*/);
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+1/*+(i&1)*/, (int)i+2/*-(i&1)*/);
+      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
+      debug_printf("  line %d %d\n", (int)i+2/*-(i&1)*/, (int)i);
+      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_trifan_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_quads_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j)[0] = (uint)in[i+0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+3);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+0);
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+0];
+   }
+}
+static void translate_quadstrip_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      debug_printf("  line %d %d\n", (int)i+2, (int)i+0);
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+0];
+      debug_printf("  line %d %d\n", (int)i+0, (int)i+1);
+      (out+j+2)[0] = (uint)in[i+0];
+      (out+j+2)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+3);
+      (out+j+4)[0] = (uint)in[i+1];
+      (out+j+4)[1] = (uint)in[i+3];
+      debug_printf("  line %d %d\n", (int)i+3, (int)i+2);
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+2];
+   }
+}
+static void translate_polygon_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      debug_printf("  line %d %d\n", (int)0, (int)i+1);
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      debug_printf("  line %d %d\n", (int)i+1, (int)i+2);
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      debug_printf("  line %d %d\n", (int)i+2, (int)0);
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+void u_unfilled_init( void )
+{
+  static int firsttime = 1;
+  if (!firsttime) return;
+  firsttime = 0;
+generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLES] = generate_tris_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_QUADS] = generate_quads_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_POLYGON] = generate_polygon_ushort;
+generate_line[OUT_UINT][PIPE_PRIM_TRIANGLES] = generate_tris_uint;
+generate_line[OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint;
+generate_line[OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint;
+generate_line[OUT_UINT][PIPE_PRIM_QUADS] = generate_quads_uint;
+generate_line[OUT_UINT][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint;
+generate_line[OUT_UINT][PIPE_PRIM_POLYGON] = generate_polygon_uint;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint;
+}
+#include "indices/u_unfilled_indices.c"
diff --git a/src/gallium/auxiliary/indices/u_unfilled_gen.py b/src/gallium/auxiliary/indices/u_unfilled_gen.py
new file mode 100644
index 0000000000..d0344fe313
--- /dev/null
+++ b/src/gallium/auxiliary/indices/u_unfilled_gen.py
@@ -0,0 +1,247 @@
+#!/usr/bin/env python
+copyright = '''
+/*
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * VMWARE AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+'''
+
+GENERATE, UBYTE, USHORT, UINT = 'generate', 'ubyte', 'ushort', 'uint'
+FIRST, LAST = 'first', 'last'
+
+INTYPES = (GENERATE, UBYTE, USHORT, UINT)
+OUTTYPES = (USHORT, UINT)
+PRIMS=('tris', 
+       'trifan', 
+       'tristrip', 
+       'quads', 
+       'quadstrip', 
+       'polygon')
+
+LONGPRIMS=('PIPE_PRIM_TRIANGLES', 
+           'PIPE_PRIM_TRIANGLE_FAN', 
+           'PIPE_PRIM_TRIANGLE_STRIP', 
+           'PIPE_PRIM_QUADS', 
+           'PIPE_PRIM_QUAD_STRIP', 
+           'PIPE_PRIM_POLYGON')
+
+longprim = dict(zip(PRIMS, LONGPRIMS))
+intype_idx = dict(ubyte='IN_UBYTE', ushort='IN_USHORT', uint='IN_UINT')
+outtype_idx = dict(ushort='OUT_USHORT', uint='OUT_UINT')
+
+
+def prolog():
+    print '''/* File automatically generated by u_unfilled_gen.py */'''
+    print copyright
+    print r'''
+
+/**
+ * @file
+ * Functions to translate and generate index lists
+ */
+
+#include "indices/u_indices.h"
+#include "indices/u_indices_priv.h"
+#include "pipe/p_compiler.h"
+#include "util/u_debug.h"
+#include "pipe/p_defines.h"
+#include "util/u_memory.h"
+
+
+static unsigned out_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return OUT_UINT;
+   case 2: return OUT_USHORT;
+   default: assert(0); return OUT_USHORT;
+   }
+}
+
+static unsigned in_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return IN_UINT;
+   case 2: return IN_USHORT;
+   case 1: return IN_UBYTE;
+   default: assert(0); return IN_UBYTE;
+   }
+}
+
+
+static u_generate_func generate_line[OUT_COUNT][PRIM_COUNT];
+static u_translate_func translate_line[IN_COUNT][OUT_COUNT][PRIM_COUNT];
+
+'''
+
+def vert( intype, outtype, v0 ):
+    if intype == GENERATE:
+        return '(' + outtype + ')(' + v0 + ')'
+    else:
+        return '(' + outtype + ')in[' + v0 + ']'
+
+def line( intype, outtype, ptr, v0, v1 ):
+    print '      debug_printf("  line %d %d\\n", (int)' + v0 + ', (int)' + v1 + ');'
+    print '      (' + ptr + ')[0] = ' + vert( intype, outtype, v0 ) + ';'
+    print '      (' + ptr + ')[1] = ' + vert( intype, outtype, v1 ) + ';'
+
+# XXX: have the opportunity here to avoid over-drawing shared lines in
+# tristrips, fans, etc, by integrating this into the calling functions
+# and only emitting each line at most once.
+# 
+def do_tri( intype, outtype, ptr, v0, v1, v2 ):
+    line( intype, outtype, ptr, v0, v1 )
+    line( intype, outtype, ptr + '+2', v1, v2 )
+    line( intype, outtype, ptr + '+4', v2, v0 )
+
+def do_quad( intype, outtype, ptr, v0, v1, v2, v3 ):
+    line( intype, outtype, ptr, v0, v1 )
+    line( intype, outtype, ptr + '+2', v1, v2 )
+    line( intype, outtype, ptr + '+4', v2, v3 )
+    line( intype, outtype, ptr + '+6', v3, v0 )
+
+def name(intype, outtype, prim):
+    if intype == GENERATE:
+        return 'generate_' + prim + '_' + outtype
+    else:
+        return 'translate_' + prim + '_' + intype + '2' + outtype
+
+def preamble(intype, outtype, prim):
+    print 'static void ' + name( intype, outtype, prim ) + '('
+    if intype != GENERATE:
+        print '    const void * _in,'
+    print '    unsigned nr,'
+    print '    void *_out )'
+    print '{'
+    if intype != GENERATE:
+        print '  const ' + intype + '*in = (const ' + intype + '*)_in;'
+    print '  ' + outtype + ' *out = (' + outtype + '*)_out;'
+    print '  unsigned i, j;'
+    print '  (void)j;'
+
+def postamble():
+    print '}'
+
+
+def tris(intype, outtype):
+    preamble(intype, outtype, prim='tris')
+    print '  for (j = i = 0; j < nr; j+=6, i+=3) { '
+    do_tri( intype, outtype, 'out+j',  'i', 'i+1', 'i+2' );
+    print '   }'
+    postamble()
+
+
+def tristrip(intype, outtype):
+    preamble(intype, outtype, prim='tristrip')
+    print '  for (j = i = 0; j < nr; j+=6, i++) { '
+    do_tri( intype, outtype, 'out+j',  'i', 'i+1/*+(i&1)*/', 'i+2/*-(i&1)*/' );
+    print '   }'
+    postamble()
+
+
+def trifan(intype, outtype):
+    preamble(intype, outtype, prim='trifan')
+    print '  for (j = i = 0; j < nr; j+=6, i++) { '
+    do_tri( intype, outtype, 'out+j',  '0', 'i+1', 'i+2' );
+    print '   }'
+    postamble()
+
+
+
+def polygon(intype, outtype):
+    preamble(intype, outtype, prim='polygon')
+    print '  for (j = i = 0; j < nr; j+=6, i++) { '
+    do_tri( intype, outtype, 'out+j',  '0', 'i+1', 'i+2' );
+    print '   }'
+    postamble()
+
+
+def quads(intype, outtype):
+    preamble(intype, outtype, prim='quads')
+    print '  for (j = i = 0; j < nr; j+=8, i+=4) { '
+    do_quad( intype, outtype, 'out+j', 'i+0', 'i+1', 'i+2', 'i+3' );
+    print '   }'
+    postamble()
+
+
+def quadstrip(intype, outtype):
+    preamble(intype, outtype, prim='quadstrip')
+    print '  for (j = i = 0; j < nr; j+=8, i+=2) { '
+    do_quad( intype, outtype, 'out+j', 'i+2', 'i+0', 'i+1', 'i+3' );
+    print '   }'
+    postamble()
+
+
+def emit_funcs():
+    for intype in INTYPES:
+        for outtype in OUTTYPES:
+            tris(intype, outtype)
+            tristrip(intype, outtype)
+            trifan(intype, outtype)
+            quads(intype, outtype)
+            quadstrip(intype, outtype)
+            polygon(intype, outtype)
+
+def init(intype, outtype, prim):
+    if intype == GENERATE:
+        print ('generate_line[' + 
+               outtype_idx[outtype] + 
+               '][' + longprim[prim] + 
+               '] = ' + name( intype, outtype, prim ) + ';')
+    else:
+        print ('translate_line[' + 
+               intype_idx[intype] + 
+               '][' + outtype_idx[outtype] + 
+               '][' + longprim[prim] + 
+               '] = ' + name( intype, outtype, prim ) + ';')
+
+
+def emit_all_inits():
+    for intype in INTYPES:
+        for outtype in OUTTYPES:
+            for prim in PRIMS:
+                init(intype, outtype, prim)
+
+def emit_init():
+    print 'void u_unfilled_init( void )'
+    print '{'
+    print '  static int firsttime = 1;'
+    print '  if (!firsttime) return;'
+    print '  firsttime = 0;'
+    emit_all_inits()
+    print '}'
+
+
+    
+
+def epilog():
+    print '#include "indices/u_unfilled_indices.c"'
+
+
+def main():
+    prolog()
+    emit_funcs()
+    emit_init()
+    epilog()
+
+
+if __name__ == '__main__':
+    main()
diff --git a/src/gallium/auxiliary/indices/u_unfilled_indices.c b/src/gallium/auxiliary/indices/u_unfilled_indices.c
new file mode 100644
index 0000000000..26c5d4d4c7
--- /dev/null
+++ b/src/gallium/auxiliary/indices/u_unfilled_indices.c
@@ -0,0 +1,186 @@
+/*
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * VMWARE AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "u_indices.h"
+#include "u_indices_priv.h"
+
+
+static void translate_ubyte_ushort( const void *in,
+                                    unsigned nr,
+                                    void *out )
+{
+   const ubyte *in_ub = (const ubyte *)in;
+   ushort *out_us = (ushort *)out;
+   unsigned i;
+   for (i = 0; i < nr; i++)
+      out_us[i] = (ushort) in_ub[i];
+}
+
+static void translate_memcpy_ushort( const void *in,
+                                     unsigned nr,
+                                     void *out )
+{
+   memcpy(out, in, nr*sizeof(short));
+}
+                              
+static void translate_memcpy_uint( const void *in,
+                                   unsigned nr,
+                                   void *out )
+{
+   memcpy(out, in, nr*sizeof(int));
+}
+
+
+static void generate_linear_ushort( unsigned nr,
+                                    void *out )
+{
+   ushort *out_us = (ushort *)out;
+   unsigned i;
+   for (i = 0; i < nr; i++)
+      out_us[i] = (ushort) i;
+}
+                              
+static void generate_linear_uint( unsigned nr,
+                                  void *out )
+{
+   unsigned *out_ui = (unsigned *)out;
+   unsigned i;
+   for (i = 0; i < nr; i++)
+      out_ui[i] = i;
+}
+
+
+static unsigned nr_lines( unsigned prim,
+                          unsigned nr )
+{
+   switch (prim) {
+   case PIPE_PRIM_TRIANGLES:
+      return (nr / 3) * 6; 
+   case PIPE_PRIM_TRIANGLE_STRIP:
+      return (nr - 2) * 6;
+   case PIPE_PRIM_TRIANGLE_FAN:
+      return (nr - 2)  * 6;
+   case PIPE_PRIM_QUADS:
+      return (nr / 4) * 8;
+   case PIPE_PRIM_QUAD_STRIP:
+      return (nr - 2) / 2 * 8;
+   case PIPE_PRIM_POLYGON:
+      return (nr - 2) * 6;
+   default:
+      assert(0);
+      return 0;
+   }
+}
+                              
+
+
+int u_unfilled_translator( unsigned prim,
+                        unsigned in_index_size,
+                        unsigned nr,
+                        unsigned unfilled_mode,
+                        unsigned *out_prim,
+                        unsigned *out_index_size,
+                        unsigned *out_nr,
+                        u_translate_func *out_translate )
+{
+   unsigned in_idx;
+   unsigned out_idx;
+
+   u_unfilled_init();
+
+   in_idx = in_size_idx(in_index_size);
+   *out_index_size = (in_index_size == 4) ? 4 : 2;
+   out_idx = out_size_idx(*out_index_size);
+
+   if (unfilled_mode == PIPE_POLYGON_MODE_POINT) 
+   {
+      *out_prim = PIPE_PRIM_POINTS;
+      *out_nr = nr;
+
+      switch (in_index_size)
+      {
+      case 1:
+         *out_translate = translate_ubyte_ushort;
+         return U_TRANSLATE_NORMAL;
+      case 2:
+         *out_translate = translate_memcpy_uint;
+         return U_TRANSLATE_MEMCPY;
+      case 4:
+         *out_translate = translate_memcpy_ushort;
+         return U_TRANSLATE_MEMCPY;
+      default:
+         *out_translate = translate_memcpy_uint;
+         *out_nr = 0;
+         assert(0);
+         return U_TRANSLATE_ERROR;
+      }
+   }
+   else {
+      assert(unfilled_mode == PIPE_POLYGON_MODE_LINE);
+      *out_prim = PIPE_PRIM_LINES;
+      *out_translate = translate_line[in_idx][out_idx][prim];
+      *out_nr = nr_lines( prim, nr );
+      return U_TRANSLATE_NORMAL;
+   }
+}
+
+
+
+int u_unfilled_generator( unsigned prim,
+                          unsigned start,
+                          unsigned nr,
+                          unsigned unfilled_mode,
+                          unsigned *out_prim,
+                          unsigned *out_index_size,
+                          unsigned *out_nr,
+                          u_generate_func *out_generate )
+{
+   unsigned out_idx;
+
+   u_unfilled_init();
+
+   *out_index_size = ((start + nr) > 0xfffe) ? 4 : 2;
+   out_idx = out_size_idx(*out_index_size);
+
+   if (unfilled_mode == PIPE_POLYGON_MODE_POINT) {
+
+      if (*out_index_size == 4)
+         *out_generate = generate_linear_uint;
+      else
+         *out_generate = generate_linear_ushort;
+
+      *out_prim = PIPE_PRIM_POINTS;
+      *out_nr = nr;
+      return U_GENERATE_LINEAR;
+   }
+   else {
+      assert(unfilled_mode == PIPE_POLYGON_MODE_LINE);
+      *out_prim = PIPE_PRIM_LINES;
+      *out_generate = generate_line[out_idx][prim];
+      *out_nr = nr_lines( prim, nr );
+
+      return U_GENERATE_REUSABLE;
+   }
+}
+
diff --git a/src/gallium/auxiliary/pipebuffer/Makefile b/src/gallium/auxiliary/pipebuffer/Makefile
index 3b501c51df..1c00ba8d98 100644
--- a/src/gallium/auxiliary/pipebuffer/Makefile
+++ b/src/gallium/auxiliary/pipebuffer/Makefile
@@ -17,6 +17,3 @@ C_SOURCES = \
 	pb_validate.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer.h b/src/gallium/auxiliary/pipebuffer/pb_buffer.h
index d8f1f02d68..2a1315922a 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_buffer.h
+++ b/src/gallium/auxiliary/pipebuffer/pb_buffer.h
@@ -45,10 +45,9 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_error.h"
 #include "pipe/p_state.h"
-#include "pipe/p_inlines.h"
 
 
 #ifdef __cplusplus
@@ -159,7 +158,7 @@ pb_map(struct pb_buffer *buf,
    assert(buf);
    if(!buf)
       return NULL;
-   assert(buf->base.refcount > 0);
+   assert(p_atomic_read(&buf->base.reference.count) > 0);
    return buf->vtbl->map(buf, flags);
 }
 
@@ -170,7 +169,7 @@ pb_unmap(struct pb_buffer *buf)
    assert(buf);
    if(!buf)
       return;
-   assert(buf->base.refcount > 0);
+   assert(p_atomic_read(&buf->base.reference.count) > 0);
    buf->vtbl->unmap(buf);
 }
 
@@ -186,7 +185,7 @@ pb_get_base_buffer( struct pb_buffer *buf,
       offset = 0;
       return;
    }
-   assert(buf->base.refcount > 0);
+   assert(p_atomic_read(&buf->base.reference.count) > 0);
    assert(buf->vtbl->get_base_buffer);
    buf->vtbl->get_base_buffer(buf, base_buf, offset);
    assert(*base_buf);
@@ -222,29 +221,18 @@ pb_destroy(struct pb_buffer *buf)
    assert(buf);
    if(!buf)
       return;
-   assert(buf->base.refcount == 0);
+   assert(p_atomic_read(&buf->base.reference.count) == 0);
    buf->vtbl->destroy(buf);
 }
 
-
-/* XXX: thread safety issues!
- */
 static INLINE void
 pb_reference(struct pb_buffer **dst,
              struct pb_buffer *src)
 {
-   if (src) {
-      assert(src->base.refcount);
-      src->base.refcount++;
-   }
-
-   if (*dst) {
-      assert((*dst)->base.refcount);
-      if(--(*dst)->base.refcount == 0)
-         pb_destroy( *dst );
-   }
+   struct pb_buffer *old = *dst;
 
-   *dst = src;
+   if (pipe_reference((struct pipe_reference**)dst, &src->base.reference))
+      pb_destroy( old );
 }
 
 
diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c b/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c
index f9e6226436..1bdf7a0b2d 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c
@@ -36,14 +36,14 @@
 
 #include "pipe/p_config.h"
 
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
 #include <unistd.h>
 #include <sched.h>
 #endif
 
 #include "pipe/p_compiler.h"
 #include "pipe/p_error.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "util/u_memory.h"
 #include "util/u_double_list.h"
@@ -66,8 +66,12 @@ struct fenced_buffer_list
    struct pb_fence_ops *ops;
    
    size_t numDelayed;
-   
    struct list_head delayed;
+   
+#ifdef DEBUG
+   size_t numUnfenced;
+   struct list_head unfenced;
+#endif
 };
 
 
@@ -111,12 +115,15 @@ _fenced_buffer_add(struct fenced_buffer *fenced_buf)
 {
    struct fenced_buffer_list *fenced_list = fenced_buf->list;
 
-   assert(fenced_buf->base.base.refcount);
+   assert(p_atomic_read(&fenced_buf->base.base.reference.count));
    assert(fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_READ_WRITE);
    assert(fenced_buf->fence);
 
-   assert(!fenced_buf->head.prev);
-   assert(!fenced_buf->head.next);
+#ifdef DEBUG
+   LIST_DEL(&fenced_buf->head);
+   assert(fenced_list->numUnfenced);
+   --fenced_list->numUnfenced;
+#endif
    LIST_ADDTAIL(&fenced_buf->head, &fenced_list->delayed);
    ++fenced_list->numDelayed;
 }
@@ -128,8 +135,19 @@ _fenced_buffer_add(struct fenced_buffer *fenced_buf)
 static INLINE void
 _fenced_buffer_destroy(struct fenced_buffer *fenced_buf)
 {
-   assert(!fenced_buf->base.base.refcount);
+   struct fenced_buffer_list *fenced_list = fenced_buf->list;
+   
+   assert(p_atomic_read(&fenced_buf->base.base.reference.count) == 0);
    assert(!fenced_buf->fence);
+#ifdef DEBUG
+   assert(fenced_buf->head.prev);
+   assert(fenced_buf->head.next);
+   LIST_DEL(&fenced_buf->head);
+   assert(fenced_list->numUnfenced);
+   --fenced_list->numUnfenced;
+#else
+   (void)fenced_list;
+#endif
    pb_reference(&fenced_buf->buffer, NULL);
    FREE(fenced_buf);
 }
@@ -149,16 +167,21 @@ _fenced_buffer_remove(struct fenced_buffer_list *fenced_list,
    
    assert(fenced_buf->head.prev);
    assert(fenced_buf->head.next);
-   LIST_DEL(&fenced_buf->head);
-#ifdef DEBUG
-   fenced_buf->head.prev = NULL;
-   fenced_buf->head.next = NULL;
-#endif
    
+   LIST_DEL(&fenced_buf->head);
    assert(fenced_list->numDelayed);
    --fenced_list->numDelayed;
    
-   if(!fenced_buf->base.base.refcount)
+#ifdef DEBUG
+   LIST_ADDTAIL(&fenced_buf->head, &fenced_list->unfenced);
+   ++fenced_list->numUnfenced;
+#endif
+   
+   /**
+    * FIXME!!!
+    */
+
+   if(!p_atomic_read(&fenced_buf->base.base.reference.count))
       _fenced_buffer_destroy(fenced_buf);
 }
 
@@ -234,7 +257,7 @@ fenced_buffer_destroy(struct pb_buffer *buf)
    struct fenced_buffer_list *fenced_list = fenced_buf->list;
 
    pipe_mutex_lock(fenced_list->mutex);
-   assert(fenced_buf->base.base.refcount == 0);
+   assert(p_atomic_read(&fenced_buf->base.base.reference.count) == 0);
    if (fenced_buf->fence) {
       struct pb_fence_ops *ops = fenced_list->ops;
       if(ops->fence_signalled(ops, fenced_buf->fence, 0) == 0) {
@@ -265,16 +288,26 @@ fenced_buffer_map(struct pb_buffer *buf,
                   unsigned flags)
 {
    struct fenced_buffer *fenced_buf = fenced_buffer(buf);
+   struct fenced_buffer_list *fenced_list = fenced_buf->list;
+   struct pb_fence_ops *ops = fenced_list->ops;
    void *map;
 
-   assert(flags & PIPE_BUFFER_USAGE_CPU_READ_WRITE);
-   assert(!(flags & ~PIPE_BUFFER_USAGE_CPU_READ_WRITE));
-   flags &= PIPE_BUFFER_USAGE_CPU_READ_WRITE;
+   assert(!(flags & PIPE_BUFFER_USAGE_GPU_READ_WRITE));
    
-   /* Check for GPU read/write access */
-   if(fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_WRITE) {
-      /* Wait for the GPU to finish writing */
-      _fenced_buffer_finish(fenced_buf);
+   /* Serialize writes */
+   if((fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_WRITE) ||
+      ((fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_READ) && (flags & PIPE_BUFFER_USAGE_CPU_WRITE))) {
+      if(flags & PIPE_BUFFER_USAGE_DONTBLOCK) {
+         /* Don't wait for the GPU to finish writing */
+         if(ops->fence_signalled(ops, fenced_buf->fence, 0) == 0)
+            _fenced_buffer_remove(fenced_list, fenced_buf);
+         else
+            return NULL;
+      }
+      else {
+         /* Wait for the GPU to finish writing */
+         _fenced_buffer_finish(fenced_buf);
+      }
    }
 
 #if 0
@@ -288,7 +321,7 @@ fenced_buffer_map(struct pb_buffer *buf,
    map = pb_map(fenced_buf->buffer, flags);
    if(map) {
       ++fenced_buf->mapcount;
-      fenced_buf->flags |= flags;
+      fenced_buf->flags |= flags & PIPE_BUFFER_USAGE_CPU_READ_WRITE;
    }
 
    return map;
@@ -432,7 +465,7 @@ fenced_buffer_create(struct fenced_buffer_list *fenced_list,
       return NULL;
    }
    
-   buf->base.base.refcount = 1;
+   pipe_reference_init(&buf->base.base.reference, 1);
    buf->base.base.alignment = buffer->base.alignment;
    buf->base.base.usage = buffer->base.usage;
    buf->base.base.size = buffer->base.size;
@@ -441,6 +474,13 @@ fenced_buffer_create(struct fenced_buffer_list *fenced_list,
    buf->buffer = buffer;
    buf->list = fenced_list;
    
+#ifdef DEBUG
+   pipe_mutex_lock(fenced_list->mutex);
+   LIST_ADDTAIL(&buf->head, &fenced_list->unfenced);
+   ++fenced_list->numUnfenced;
+   pipe_mutex_unlock(fenced_list->mutex);
+#endif
+
    return &buf->base;
 }
 
@@ -457,9 +497,13 @@ fenced_buffer_list_create(struct pb_fence_ops *ops)
    fenced_list->ops = ops;
 
    LIST_INITHEAD(&fenced_list->delayed);
-
    fenced_list->numDelayed = 0;
    
+#ifdef DEBUG
+   LIST_INITHEAD(&fenced_list->unfenced);
+   fenced_list->numUnfenced = 0;
+#endif
+
    pipe_mutex_init(fenced_list->mutex);
 
    return fenced_list;
@@ -476,6 +520,51 @@ fenced_buffer_list_check_free(struct fenced_buffer_list *fenced_list,
 }
 
 
+#ifdef DEBUG
+void
+fenced_buffer_list_dump(struct fenced_buffer_list *fenced_list)
+{
+   struct pb_fence_ops *ops = fenced_list->ops;
+   struct list_head *curr, *next;
+   struct fenced_buffer *fenced_buf;
+
+   pipe_mutex_lock(fenced_list->mutex);
+
+   debug_printf("%10s %7s %10s %s\n",
+                "buffer", "reference.count", "fence", "signalled");
+   
+   curr = fenced_list->unfenced.next;
+   next = curr->next;
+   while(curr != &fenced_list->unfenced) {
+      fenced_buf = LIST_ENTRY(struct fenced_buffer, curr, head);
+      assert(!fenced_buf->fence);
+      debug_printf("%10p %7u\n",
+                   fenced_buf,
+                   fenced_buf->base.base.reference.count);
+      curr = next; 
+      next = curr->next;
+   }
+   
+   curr = fenced_list->delayed.next;
+   next = curr->next;
+   while(curr != &fenced_list->delayed) {
+      int signaled;
+      fenced_buf = LIST_ENTRY(struct fenced_buffer, curr, head);
+      signaled = ops->fence_signalled(ops, fenced_buf->fence, 0);
+      debug_printf("%10p %7u %10p %s\n",
+                   fenced_buf,
+                   fenced_buf->base.base.reference.count,
+                   fenced_buf->fence,
+                   signaled == 0 ? "y" : "n");
+      curr = next; 
+      next = curr->next;
+   }
+   
+   pipe_mutex_unlock(fenced_list->mutex);
+}
+#endif
+
+
 void
 fenced_buffer_list_destroy(struct fenced_buffer_list *fenced_list)
 {
@@ -484,13 +573,17 @@ fenced_buffer_list_destroy(struct fenced_buffer_list *fenced_list)
    /* Wait on outstanding fences */
    while (fenced_list->numDelayed) {
       pipe_mutex_unlock(fenced_list->mutex);
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
       sched_yield();
 #endif
       _fenced_buffer_list_check_free(fenced_list, 1);
       pipe_mutex_lock(fenced_list->mutex);
    }
 
+#ifdef DEBUG
+   //assert(!fenced_list->numUnfenced);
+#endif
+      
    pipe_mutex_unlock(fenced_list->mutex);
    
    fenced_list->ops->destroy(fenced_list->ops);
diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.h b/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.h
index d1c9d4c17d..034ca1e024 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.h
+++ b/src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.h
@@ -51,7 +51,7 @@
 #define PB_BUFFER_FENCED_H_
 
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 
 #ifdef __cplusplus
@@ -114,6 +114,13 @@ void
 fenced_buffer_list_check_free(struct fenced_buffer_list *fenced_list, 
                               int wait);
 
+
+#ifdef DEBUG
+void
+fenced_buffer_list_dump(struct fenced_buffer_list *fenced_list);
+#endif
+
+
 void
 fenced_buffer_list_destroy(struct fenced_buffer_list *fenced_list);
 
diff --git a/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c b/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c
index 53f497cfb0..689fd74771 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_buffer_malloc.c
@@ -34,7 +34,7 @@
  */
 
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_memory.h"
 #include "pb_buffer.h"
 #include "pb_bufmgr.h"
@@ -132,7 +132,7 @@ pb_malloc_buffer_create(size_t size,
    if(!buf)
       return NULL;
 
-   buf->base.base.refcount = 1;
+   pipe_reference_init(&buf->base.base.reference, 1);
    buf->base.base.alignment = desc->alignment;
    buf->base.base.usage = desc->usage;
    buf->base.base.size = size;
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr.h b/src/gallium/auxiliary/pipebuffer/pb_bufmgr.h
index fec8db91c7..74077f8277 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr.h
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr.h
@@ -200,10 +200,11 @@ pb_ondemand_manager_create(struct pb_manager *provider);
 /** 
  * Debug buffer manager to detect buffer under- and overflows.
  *
- * Band size should be a multiple of the largest alignment
+ * Under/overflow sizes should be a multiple of the largest alignment
  */
 struct pb_manager *
-pb_debug_manager_create(struct pb_manager *provider, size_t band_size); 
+pb_debug_manager_create(struct pb_manager *provider,
+                        size_t underflow_size, size_t overflow_size); 
 
 
 #ifdef __cplusplus
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c
index c956924cc7..db67d46c56 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_alt.c
@@ -34,7 +34,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_memory.h"
 
 #include "pb_buffer.h"
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c
index a168853713..010a2ecc1f 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c
@@ -35,7 +35,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "util/u_memory.h"
 #include "util/u_double_list.h"
@@ -112,7 +112,7 @@ _pb_cache_buffer_destroy(struct pb_cache_buffer *buf)
    LIST_DEL(&buf->head);
    assert(mgr->numDelayed);
    --mgr->numDelayed;
-   assert(!buf->base.base.refcount);
+   assert(p_atomic_read(&buf->base.base.reference.count) == 0);
    pb_reference(&buf->buffer, NULL);
    FREE(buf);
 }
@@ -153,7 +153,7 @@ pb_cache_buffer_destroy(struct pb_buffer *_buf)
    struct pb_cache_manager *mgr = buf->mgr;
 
    pipe_mutex_lock(mgr->mutex);
-   assert(buf->base.base.refcount == 0);
+   assert(p_atomic_read(&buf->base.base.reference.count) == 0);
    
    _pb_cache_buffer_list_check_free(mgr);
    
@@ -293,7 +293,8 @@ pb_cache_manager_create_buffer(struct pb_manager *_mgr,
    if(buf) {
       LIST_DEL(&buf->head);
       pipe_mutex_unlock(mgr->mutex);
-      ++buf->base.base.refcount;
+      /* Increase refcount */
+      pb_reference((struct pb_buffer**)&buf, &buf->base);
       return &buf->base;
    }
    
@@ -309,12 +310,12 @@ pb_cache_manager_create_buffer(struct pb_manager *_mgr,
       return NULL;
    }
    
-   assert(buf->buffer->base.refcount >= 1);
+   assert(p_atomic_read(&buf->buffer->base.reference.count) >= 1);
    assert(pb_check_alignment(desc->alignment, buf->buffer->base.alignment));
    assert(pb_check_usage(desc->usage, buf->buffer->base.usage));
    assert(buf->buffer->base.size >= size);
    
-   buf->base.base.refcount = 1;
+   pipe_reference_init(&buf->base.base.reference, 1);
    buf->base.base.alignment = buf->buffer->base.alignment;
    buf->base.base.usage = buf->buffer->base.usage;
    buf->base.base.size = buf->buffer->base.size;
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c
index 26d9c24aec..478682dbee 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_debug.c
@@ -34,7 +34,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
@@ -78,7 +78,8 @@ struct pb_debug_manager
 
    struct pb_manager *provider;
 
-   size_t band_size;
+   size_t underflow_size;
+   size_t overflow_size;
 };
 
 
@@ -205,9 +206,9 @@ pb_debug_buffer_check(struct pb_debug_buffer *buf)
 static void
 pb_debug_buffer_destroy(struct pb_buffer *_buf)
 {
-   struct pb_debug_buffer *buf = pb_debug_buffer(_buf);  
+   struct pb_debug_buffer *buf = pb_debug_buffer(_buf);
    
-   assert(!buf->base.base.refcount);
+   assert(p_atomic_read(&buf->base.base.reference.count) == 0);
    
    pb_debug_buffer_check(buf);
 
@@ -301,7 +302,7 @@ pb_debug_manager_create_buffer(struct pb_manager *_mgr,
    if(!buf)
       return NULL;
    
-   real_size = size + 2*mgr->band_size;
+   real_size = mgr->underflow_size + size + mgr->overflow_size;
    real_desc = *desc;
    real_desc.usage |= PIPE_BUFFER_USAGE_CPU_WRITE;
    real_desc.usage |= PIPE_BUFFER_USAGE_CPU_READ;
@@ -314,12 +315,12 @@ pb_debug_manager_create_buffer(struct pb_manager *_mgr,
       return NULL;
    }
    
-   assert(buf->buffer->base.refcount >= 1);
+   assert(p_atomic_read(&buf->buffer->base.reference.count) >= 1);
    assert(pb_check_alignment(real_desc.alignment, buf->buffer->base.alignment));
    assert(pb_check_usage(real_desc.usage, buf->buffer->base.usage));
    assert(buf->buffer->base.size >= real_size);
    
-   buf->base.base.refcount = 1;
+   pipe_reference_init(&buf->base.base.reference, 1);
    buf->base.base.alignment = desc->alignment;
    buf->base.base.usage = desc->usage;
    buf->base.base.size = size;
@@ -327,7 +328,7 @@ pb_debug_manager_create_buffer(struct pb_manager *_mgr,
    buf->base.vtbl = &pb_debug_buffer_vtbl;
    buf->mgr = mgr;
 
-   buf->underflow_size = mgr->band_size;
+   buf->underflow_size = mgr->underflow_size;
    buf->overflow_size = buf->buffer->base.size - buf->underflow_size - size;
    
    pb_debug_buffer_fill(buf);
@@ -356,7 +357,8 @@ pb_debug_manager_destroy(struct pb_manager *_mgr)
 
 
 struct pb_manager *
-pb_debug_manager_create(struct pb_manager *provider, size_t band_size) 
+pb_debug_manager_create(struct pb_manager *provider, 
+                        size_t underflow_size, size_t overflow_size) 
 {
    struct pb_debug_manager *mgr;
 
@@ -371,7 +373,8 @@ pb_debug_manager_create(struct pb_manager *provider, size_t band_size)
    mgr->base.create_buffer = pb_debug_manager_create_buffer;
    mgr->base.flush = pb_debug_manager_flush;
    mgr->provider = provider;
-   mgr->band_size = band_size;
+   mgr->underflow_size = underflow_size;
+   mgr->overflow_size = overflow_size;
       
    return &mgr->base;
 }
@@ -381,7 +384,8 @@ pb_debug_manager_create(struct pb_manager *provider, size_t band_size)
 
 
 struct pb_manager *
-pb_debug_manager_create(struct pb_manager *provider, size_t band_size) 
+pb_debug_manager_create(struct pb_manager *provider, 
+                        size_t underflow_size, size_t overflow_size) 
 {
    return provider;
 }
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_fenced.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_fenced.c
index 47e9fee533..144db5669b 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_fenced.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_fenced.c
@@ -34,7 +34,7 @@
  */
 
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_memory.h"
 
 #include "pb_buffer.h"
@@ -79,6 +79,10 @@ fenced_bufmgr_create_buffer(struct pb_manager *mgr,
       
       buf = fenced_mgr->provider->create_buffer(fenced_mgr->provider, size, desc);
       if(!buf) {
+#if 0
+         fenced_buffer_list_dump(fenced_mgr->fenced_list);
+#endif
+         
          /* give up */
          return NULL;
       }
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c
index 2f5a5d8ea0..fb18dcc5dc 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_mm.c
@@ -34,7 +34,7 @@
 
 
 #include "pipe/p_defines.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "util/u_memory.h"
 #include "util/u_double_list.h"
@@ -97,11 +97,11 @@ mm_buffer_destroy(struct pb_buffer *buf)
    struct mm_buffer *mm_buf = mm_buffer(buf);
    struct mm_pb_manager *mm = mm_buf->mgr;
    
-   assert(buf->base.refcount == 0);
+   assert(p_atomic_read(&mm_buf->base.base.reference.count) == 0);
    
    pipe_mutex_lock(mm->mutex);
    u_mmFreeMem(mm_buf->block);
-   FREE(buf);
+   FREE(mm_buf);
    pipe_mutex_unlock(mm->mutex);
 }
 
@@ -189,7 +189,7 @@ mm_bufmgr_create_buffer(struct pb_manager *mgr,
       return NULL;
    }
 
-   mm_buf->base.base.refcount = 1;
+   pipe_reference_init(&mm_buf->base.base.reference, 1);
    mm_buf->base.base.alignment = desc->alignment;
    mm_buf->base.base.usage = desc->usage;
    mm_buf->base.base.size = size;
@@ -204,13 +204,9 @@ mm_bufmgr_create_buffer(struct pb_manager *mgr,
 #if 0
       mmDumpMemInfo(mm->heap);
 #endif
-      
-      mm_buf->block = u_mmAllocMem(mm->heap, size, mm->align2, 0);
-      if(!mm_buf->block) {
-         FREE(mm_buf);
-         pipe_mutex_unlock(mm->mutex);
-         return NULL;
-      }
+      FREE(mm_buf);
+      pipe_mutex_unlock(mm->mutex);
+      return NULL;
    }
    
    /* Some sanity checks */
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c
index ba02a84e62..4f7e6b1c4d 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_ondemand.c
@@ -34,7 +34,7 @@
  */
 
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_memory.h"
 #include "pb_buffer.h"
 #include "pb_bufmgr.h"
@@ -242,7 +242,7 @@ pb_ondemand_manager_create_buffer(struct pb_manager *_mgr,
    if(!buf)
       return NULL;
 
-   buf->base.base.refcount = 1;
+   pipe_reference_init(&buf->base.base.reference, 1);
    buf->base.base.alignment = desc->alignment;
    buf->base.base.usage = desc->usage;
    buf->base.base.size = size;
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c
index a6ff37653e..75b95e132e 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_pool.c
@@ -36,7 +36,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "pipe/p_defines.h"
 #include "util/u_memory.h"
@@ -108,7 +108,7 @@ pool_buffer_destroy(struct pb_buffer *buf)
    struct pool_buffer *pool_buf = pool_buffer(buf);
    struct pool_pb_manager *pool = pool_buf->mgr;
    
-   assert(pool_buf->base.base.refcount == 0);
+   assert(p_atomic_read(&pool_buf->base.base.reference.count) == 0);
 
    pipe_mutex_lock(pool->mutex);
    LIST_ADD(&pool_buf->head, &pool->free);
@@ -216,8 +216,8 @@ pool_bufmgr_create_buffer(struct pb_manager *mgr,
    pipe_mutex_unlock(pool->mutex);
    
    pool_buf = LIST_ENTRY(struct pool_buffer, item, head);
-   assert(pool_buf->base.base.refcount == 0);
-   pool_buf->base.base.refcount = 1;
+   assert(p_atomic_read(&pool_buf->base.base.reference.count) == 0);
+   pipe_reference_init(&pool_buf->base.base.reference, 1);
    pool_buf->base.base.alignment = desc->alignment;
    pool_buf->base.base.usage = desc->usage;
    
@@ -295,7 +295,7 @@ pool_bufmgr_create(struct pb_manager *provider,
 
    pool_buf = pool->bufs;
    for (i = 0; i < numBufs; ++i) {
-      pool_buf->base.base.refcount = 0;
+      pipe_reference_init(&pool_buf->base.base.reference, 0);
       pool_buf->base.base.alignment = 0;
       pool_buf->base.base.usage = 0;
       pool_buf->base.base.size = bufSize;
diff --git a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c
index 9b9fedccb4..a431fd5211 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_bufmgr_slab.c
@@ -38,7 +38,7 @@
 
 #include "pipe/p_compiler.h"
 #include "pipe/p_error.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "pipe/p_defines.h"
 #include "util/u_memory.h"
@@ -202,7 +202,7 @@ pb_slab_buffer_destroy(struct pb_buffer *_buf)
 
    pipe_mutex_lock(mgr->mutex);
    
-   assert(buf->base.base.refcount == 0);
+   assert(p_atomic_read(&buf->base.base.reference.count) == 0);
    
    buf->mapCount = 0;
 
@@ -340,7 +340,7 @@ pb_slab_create(struct pb_slab_manager *mgr)
 
    buf = slab->buffers;
    for (i=0; i < numBuffers; ++i) {
-      buf->base.base.refcount = 0;
+      pipe_reference_init(&buf->base.base.reference, 0);
       buf->base.base.size = mgr->bufSize;
       buf->base.base.alignment = 0;
       buf->base.base.usage = 0;
@@ -419,7 +419,7 @@ pb_slab_manager_create_buffer(struct pb_manager *_mgr,
    pipe_mutex_unlock(mgr->mutex);
    buf = LIST_ENTRY(struct pb_slab_buffer, list, head);
    
-   ++buf->base.base.refcount;
+   pipe_reference_init(&buf->base.base.reference, 1);
    buf->base.base.alignment = desc->alignment;
    buf->base.base.usage = desc->usage;
    
diff --git a/src/gallium/auxiliary/pipebuffer/pb_validate.c b/src/gallium/auxiliary/pipebuffer/pb_validate.c
index 94532bb4ce..150fd50618 100644
--- a/src/gallium/auxiliary/pipebuffer/pb_validate.c
+++ b/src/gallium/auxiliary/pipebuffer/pb_validate.c
@@ -36,7 +36,7 @@
 #include "pipe/p_compiler.h"
 #include "pipe/p_error.h"
 #include "util/u_memory.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "pb_buffer.h"
 #include "pb_buffer_fenced.h"
diff --git a/src/gallium/auxiliary/rtasm/Makefile b/src/gallium/auxiliary/rtasm/Makefile
index 252dc5274a..ab8ea464c6 100644
--- a/src/gallium/auxiliary/rtasm/Makefile
+++ b/src/gallium/auxiliary/rtasm/Makefile
@@ -11,6 +11,3 @@ C_SOURCES = \
 	rtasm_ppc_spe.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/rtasm/rtasm_cpu.c b/src/gallium/auxiliary/rtasm/rtasm_cpu.c
index 5499018b21..03bdd47238 100644
--- a/src/gallium/auxiliary/rtasm/rtasm_cpu.c
+++ b/src/gallium/auxiliary/rtasm/rtasm_cpu.c
@@ -26,7 +26,7 @@
  **************************************************************************/
 
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "rtasm_cpu.h"
 
 
diff --git a/src/gallium/auxiliary/rtasm/rtasm_execmem.c b/src/gallium/auxiliary/rtasm/rtasm_execmem.c
index be7433baf8..1f0923b683 100644
--- a/src/gallium/auxiliary/rtasm/rtasm_execmem.c
+++ b/src/gallium/auxiliary/rtasm/rtasm_execmem.c
@@ -31,14 +31,18 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_thread.h"
 #include "util/u_memory.h"
 
 #include "rtasm_execmem.h"
 
+#if defined(PIPE_OS_BSD)
+#define MAP_ANONYMOUS MAP_ANON
+#endif
 
-#if defined(PIPE_OS_LINUX)
+
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
 
 
 /*
@@ -114,7 +118,7 @@ rtasm_exec_free(void *addr)
 }
 
 
-#else /* PIPE_OS_LINUX */
+#else /* PIPE_OS_LINUX || PIPE_OS_BSD */
 
 /*
  * Just use regular memory.
@@ -134,4 +138,4 @@ rtasm_exec_free(void *addr)
 }
 
 
-#endif /* PIPE_OS_LINUX */
+#endif /* PIPE_OS_LINUX || PIPE_OS_BSD */
diff --git a/src/gallium/auxiliary/rtasm/rtasm_ppc.c b/src/gallium/auxiliary/rtasm/rtasm_ppc.c
index 1bb9026205..e3586482db 100644
--- a/src/gallium/auxiliary/rtasm/rtasm_ppc.c
+++ b/src/gallium/auxiliary/rtasm/rtasm_ppc.c
@@ -38,7 +38,7 @@
 
 #include <stdio.h>
 #include "util/u_memory.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "rtasm_execmem.h"
 #include "rtasm_ppc.h"
 
diff --git a/src/gallium/auxiliary/rtasm/rtasm_x86sse.c b/src/gallium/auxiliary/rtasm/rtasm_x86sse.c
index 99ee74cf14..57fcf6de2a 100644
--- a/src/gallium/auxiliary/rtasm/rtasm_x86sse.c
+++ b/src/gallium/auxiliary/rtasm/rtasm_x86sse.c
@@ -26,7 +26,7 @@
 #if defined(PIPE_ARCH_X86)
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_pointer.h"
 
 #include "rtasm_execmem.h"
diff --git a/src/gallium/auxiliary/sct/Makefile b/src/gallium/auxiliary/sct/Makefile
index 516d1756cf..a7d111b689 100644
--- a/src/gallium/auxiliary/sct/Makefile
+++ b/src/gallium/auxiliary/sct/Makefile
@@ -7,6 +7,3 @@ C_SOURCES = \
 	sct.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/sct/sct.c b/src/gallium/auxiliary/sct/sct.c
index 49bb7ea92e..722d2b7e66 100644
--- a/src/gallium/auxiliary/sct/sct.c
+++ b/src/gallium/auxiliary/sct/sct.c
@@ -28,7 +28,6 @@
 
 #include "util/u_memory.h"
 #include "pipe/p_state.h"
-#include "pipe/p_inlines.h"
 #include "sct.h"
 
 
@@ -372,7 +371,7 @@ sct_flush_textures(struct surface_context_tracker *sct,
 
    for (tl = ci->textures_used; tl; tl = next) {
       next = tl->next;
-      pipe_texture_release(&tl->texture);
+      pipe_texture_reference(&tl->texture, NULL);
       FREE(tl);
    }
    ci->textures_used = NULL;
diff --git a/src/gallium/auxiliary/tgsi/Makefile b/src/gallium/auxiliary/tgsi/Makefile
index d7df9490cf..b4900e8dba 100644
--- a/src/gallium/auxiliary/tgsi/Makefile
+++ b/src/gallium/auxiliary/tgsi/Makefile
@@ -19,6 +19,3 @@ C_SOURCES = \
 	tgsi_util.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/tgsi/tgsi-instruction-set.txt b/src/gallium/auxiliary/tgsi/tgsi-instruction-set.txt
new file mode 100644
index 0000000000..b83abd4093
--- /dev/null
+++ b/src/gallium/auxiliary/tgsi/tgsi-instruction-set.txt
@@ -0,0 +1,1036 @@
+TGSI Instruction Specification
+==============================
+==============================
+
+
+1  Instruction Set Operations
+=============================
+
+
+1.1  GL_NV_vertex_program
+-------------------------
+
+
+1.1.1  ARL - Address Register Load
+
+  dst.x = floor(src.x)
+  dst.y = floor(src.y)
+  dst.z = floor(src.z)
+  dst.w = floor(src.w)
+
+
+1.1.2  MOV - Move
+
+  dst.x = src.x
+  dst.y = src.y
+  dst.z = src.z
+  dst.w = src.w
+
+
+1.1.3  LIT - Light Coefficients
+
+  dst.x = 1.0
+  dst.y = max(src.x, 0.0)
+  dst.z = (src.x > 0.0) ? pow(max(src.y, 0.0), clamp(src.w, -128.0, 128.0)) : 0.0
+  dst.w = 1.0
+
+
+1.1.4  RCP - Reciprocal
+
+  dst.x = 1.0 / src.x
+  dst.y = 1.0 / src.x
+  dst.z = 1.0 / src.x
+  dst.w = 1.0 / src.x
+
+
+1.1.5  RSQ - Reciprocal Square Root
+
+  dst.x = 1.0 / sqrt(abs(src.x))
+  dst.y = 1.0 / sqrt(abs(src.x))
+  dst.z = 1.0 / sqrt(abs(src.x))
+  dst.w = 1.0 / sqrt(abs(src.x))
+
+
+1.1.6  EXP - Approximate Exponential Base 2
+
+  dst.x = pow(2.0, floor(src.x))
+  dst.y = src.x - floor(src.x)
+  dst.z = pow(2.0, src.x)
+  dst.w = 1.0
+
+
+1.1.7  LOG - Approximate Logarithm Base 2
+
+  dst.x = floor(lg2(abs(src.x)))
+  dst.y = abs(src.x) / pow(2.0, floor(lg2(abs(src.x))))
+  dst.z = lg2(abs(src.x))
+  dst.w = 1.0
+
+
+1.1.8  MUL - Multiply
+
+  dst.x = src0.x * src1.x
+  dst.y = src0.y * src1.y
+  dst.z = src0.z * src1.z
+  dst.w = src0.w * src1.w
+
+
+1.1.9  ADD - Add
+
+  dst.x = src0.x + src1.x
+  dst.y = src0.y + src1.y
+  dst.z = src0.z + src1.z
+  dst.w = src0.w + src1.w
+
+
+1.1.10  DP3 - 3-component Dot Product
+
+  dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
+  dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
+  dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
+  dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z
+
+
+1.1.11  DP4 - 4-component Dot Product
+
+  dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
+  dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
+  dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
+  dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src0.w * src1.w
+
+
+1.1.12  DST - Distance Vector
+
+  dst.x = 1.0
+  dst.y = src0.y * src1.y
+  dst.z = src0.z
+  dst.w = src1.w
+
+
+1.1.13  MIN - Minimum
+
+  dst.x = min(src0.x, src1.x)
+  dst.y = min(src0.y, src1.y)
+  dst.z = min(src0.z, src1.z)
+  dst.w = min(src0.w, src1.w)
+
+
+1.1.14  MAX - Maximum
+
+  dst.x = max(src0.x, src1.x)
+  dst.y = max(src0.y, src1.y)
+  dst.z = max(src0.z, src1.z)
+  dst.w = max(src0.w, src1.w)
+
+
+1.1.15  SLT - Set On Less Than
+
+  dst.x = (src0.x < src1.x) ? 1.0 : 0.0
+  dst.y = (src0.y < src1.y) ? 1.0 : 0.0
+  dst.z = (src0.z < src1.z) ? 1.0 : 0.0
+  dst.w = (src0.w < src1.w) ? 1.0 : 0.0
+
+
+1.1.16  SGE - Set On Greater Equal Than
+
+  dst.x = (src0.x >= src1.x) ? 1.0 : 0.0
+  dst.y = (src0.y >= src1.y) ? 1.0 : 0.0
+  dst.z = (src0.z >= src1.z) ? 1.0 : 0.0
+  dst.w = (src0.w >= src1.w) ? 1.0 : 0.0
+
+
+1.1.17  MAD - Multiply And Add
+
+  dst.x = src0.x * src1.x + src2.x
+  dst.y = src0.y * src1.y + src2.y
+  dst.z = src0.z * src1.z + src2.z
+  dst.w = src0.w * src1.w + src2.w
+
+
+1.2  GL_ATI_fragment_shader
+---------------------------
+
+
+1.2.1  SUB - Subtract
+
+  dst.x = src0.x - src1.x
+  dst.y = src0.y - src1.y
+  dst.z = src0.z - src1.z
+  dst.w = src0.w - src1.w
+
+
+1.2.2  DOT3 - 3-component Dot Product
+
+  Alias for DP3.
+
+
+1.2.3  DOT4 - 4-component Dot Product
+
+  Alias for DP4.
+
+
+1.2.4  LERP - Linear Interpolate
+
+  dst.x = src0.x * (src1.x - src2.x) + src2.x
+  dst.y = src0.y * (src1.y - src2.y) + src2.y
+  dst.z = src0.z * (src1.z - src2.z) + src2.z
+  dst.w = src0.w * (src1.w - src2.w) + src2.w
+
+
+1.2.5  CND - Condition
+
+  dst.x = (src2.x > 0.5) ? src0.x : src1.x
+  dst.y = (src2.y > 0.5) ? src0.y : src1.y
+  dst.z = (src2.z > 0.5) ? src0.z : src1.z
+  dst.w = (src2.w > 0.5) ? src0.w : src1.w
+
+
+1.2.6  CND0 - Condition Zero
+
+  dst.x = (src2.x >= 0.0) ? src0.x : src1.x
+  dst.y = (src2.y >= 0.0) ? src0.y : src1.y
+  dst.z = (src2.z >= 0.0) ? src0.z : src1.z
+  dst.w = (src2.w >= 0.0) ? src0.w : src1.w
+
+
+1.2.7  DOT2ADD - 2-component Dot Product And Add
+
+  dst.x = src0.x * src1.x + src0.y * src1.y + src2.x
+  dst.y = src0.x * src1.x + src0.y * src1.y + src2.x
+  dst.z = src0.x * src1.x + src0.y * src1.y + src2.x
+  dst.w = src0.x * src1.x + src0.y * src1.y + src2.x
+
+
+1.3  GL_EXT_vertex_shader
+-------------------------
+
+
+1.3.1  INDEX - Array Lookup
+
+  Considered for removal from language.
+
+
+1.3.2  NEGATE - Negate
+
+  Considered for removal from language.
+
+
+1.3.3  MADD - Multiply And Add
+
+  Alias for MAD.
+
+
+1.3.4  FRAC - Fraction
+
+  dst.x = src.x - floor(src.x)
+  dst.y = src.y - floor(src.y)
+  dst.z = src.z - floor(src.z)
+  dst.w = src.w - floor(src.w)
+
+
+1.3.5  SETGE - Set On Greater Equal
+
+  Alias for SGE.
+
+
+1.3.6  SETLT - Set On Less Than
+
+  Alias for SLT.
+
+
+1.3.7  CLAMP - Clamp
+
+  dst.x = clamp(src0.x, src1.x, src2.x)
+  dst.y = clamp(src0.y, src1.y, src2.y)
+  dst.z = clamp(src0.z, src1.z, src2.z)
+  dst.w = clamp(src0.w, src1.w, src2.w)
+
+
+1.3.8  FLOOR - Floor
+
+  dst.x = floor(src.x)
+  dst.y = floor(src.y)
+  dst.z = floor(src.z)
+  dst.w = floor(src.w)
+
+
+1.3.9  ROUND - Round
+
+  dst.x = round(src.x)
+  dst.y = round(src.y)
+  dst.z = round(src.z)
+  dst.w = round(src.w)
+
+
+1.3.10  EXPBASE2 - Exponent Base 2
+
+  dst.x = pow(2.0, src.x)
+  dst.y = pow(2.0, src.x)
+  dst.z = pow(2.0, src.x)
+  dst.w = pow(2.0, src.x)
+
+
+1.3.11  LOGBASE2 - Logarithm Base 2
+
+  dst.x = lg2(src.x)
+  dst.y = lg2(src.x)
+  dst.z = lg2(src.x)
+  dst.w = lg2(src.x)
+
+
+1.3.12  POWER - Power
+
+  dst.x = pow(src0.x, src1.x)
+  dst.y = pow(src0.x, src1.x)
+  dst.z = pow(src0.x, src1.x)
+  dst.w = pow(src0.x, src1.x)
+
+
+1.3.13  RECIP - Reciprocal
+
+  Alias for RCP.
+
+
+1.3.14  RECIPSQRT - Reciprocal Square Root
+
+  Alias for RSQ.
+
+
+1.3.15  CROSSPRODUCT - Cross Product
+
+  dst.x = src0.y * src1.z - src1.y * src0.z
+  dst.y = src0.z * src1.x - src1.z * src0.x
+  dst.z = src0.x * src1.y - src1.x * src0.y
+  dst.w = 1.0
+
+
+1.3.16  MULTIPLYMATRIX - Multiply Matrix
+
+  Considered for removal from language.
+
+
+1.4  GL_NV_vertex_program1_1
+----------------------------
+
+
+1.4.1  ABS - Absolute
+
+  dst.x = abs(src.x)
+  dst.y = abs(src.y)
+  dst.z = abs(src.z)
+  dst.w = abs(src.w)
+
+
+1.4.2  RCC - Reciprocal Clamped
+
+  dst.x = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
+  dst.y = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
+  dst.z = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
+  dst.w = (1.0 / src.x) > 0.0 ? clamp(1.0 / src.x, 5.42101e-020, 1.884467e+019) : clamp(1.0 / src.x, -1.884467e+019, -5.42101e-020)
+
+
+1.4.3  DPH - Homogeneous Dot Product
+
+  dst.x = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
+  dst.y = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
+  dst.z = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
+  dst.w = src0.x * src1.x + src0.y * src1.y + src0.z * src1.z + src1.w
+
+
+1.5  GL_NV_fragment_program
+---------------------------
+
+
+1.5.1  COS - Cosine
+
+  dst.x = cos(src.x)
+  dst.y = cos(src.x)
+  dst.z = cos(src.x)
+  dst.w = cos(src.w)
+
+
+1.5.2  DDX - Derivative Relative To X
+
+  dst.x = partialx(src.x)
+  dst.y = partialx(src.y)
+  dst.z = partialx(src.z)
+  dst.w = partialx(src.w)
+
+
+1.5.3  DDY - Derivative Relative To Y
+
+  dst.x = partialy(src.x)
+  dst.y = partialy(src.y)
+  dst.z = partialy(src.z)
+  dst.w = partialy(src.w)
+
+
+1.5.4  EX2 - Exponential Base 2
+
+  Alias for EXPBASE2.
+
+
+1.5.5  FLR - Floor
+
+  Alias for FLOOR.
+
+
+1.5.6  FRC - Fraction
+
+  Alias for FRAC.
+
+
+1.5.7  KILP - Predicated Discard
+
+  TBD
+
+
+1.5.8  LG2 - Logarithm Base 2
+
+  Alias for LOGBASE2.
+
+
+1.5.9  LRP - Linear Interpolate
+
+  Alias for LERP.
+
+
+1.5.10  PK2H - Pack Two 16-bit Floats
+
+  TBD
+
+
+1.5.11  PK2US - Pack Two Unsigned 16-bit Scalars
+
+  TBD
+
+
+1.5.12  PK4B - Pack Four Signed 8-bit Scalars
+
+  TBD
+
+
+1.5.13  PK4UB - Pack Four Unsigned 8-bit Scalars
+
+  TBD
+
+
+1.5.14  POW - Power
+
+  Alias for POWER.
+
+
+1.5.15  RFL - Reflection Vector
+
+  dst.x = 2.0 * (src0.x * src1.x + src0.y * src1.y + src0.z * src1.z) / (src0.x * src0.x + src0.y * src0.y + src0.z * src0.z) * src0.x - src1.x
+  dst.y = 2.0 * (src0.x * src1.x + src0.y * src1.y + src0.z * src1.z) / (src0.x * src0.x + src0.y * src0.y + src0.z * src0.z) * src0.y - src1.y
+  dst.z = 2.0 * (src0.x * src1.x + src0.y * src1.y + src0.z * src1.z) / (src0.x * src0.x + src0.y * src0.y + src0.z * src0.z) * src0.z - src1.z
+  dst.w = 1.0
+
+
+1.5.16  SEQ - Set On Equal
+
+  dst.x = (src0.x == src1.x) ? 1.0 : 0.0
+  dst.y = (src0.y == src1.y) ? 1.0 : 0.0
+  dst.z = (src0.z == src1.z) ? 1.0 : 0.0
+  dst.w = (src0.w == src1.w) ? 1.0 : 0.0
+
+
+1.5.17  SFL - Set On False
+
+  dst.x = 0.0
+  dst.y = 0.0
+  dst.z = 0.0
+  dst.w = 0.0
+
+
+1.5.18  SGT - Set On Greater Than
+
+  dst.x = (src0.x > src1.x) ? 1.0 : 0.0
+  dst.y = (src0.y > src1.y) ? 1.0 : 0.0
+  dst.z = (src0.z > src1.z) ? 1.0 : 0.0
+  dst.w = (src0.w > src1.w) ? 1.0 : 0.0
+
+
+1.5.19  SIN - Sine
+
+  dst.x = sin(src.x)
+  dst.y = sin(src.x)
+  dst.z = sin(src.x)
+  dst.w = sin(src.w)
+
+
+1.5.20  SLE - Set On Less Equal Than
+
+  dst.x = (src0.x <= src1.x) ? 1.0 : 0.0
+  dst.y = (src0.y <= src1.y) ? 1.0 : 0.0
+  dst.z = (src0.z <= src1.z) ? 1.0 : 0.0
+  dst.w = (src0.w <= src1.w) ? 1.0 : 0.0
+
+
+1.5.21  SNE - Set On Not Equal
+
+  dst.x = (src0.x != src1.x) ? 1.0 : 0.0
+  dst.y = (src0.y != src1.y) ? 1.0 : 0.0
+  dst.z = (src0.z != src1.z) ? 1.0 : 0.0
+  dst.w = (src0.w != src1.w) ? 1.0 : 0.0
+
+
+1.5.22  STR - Set On True
+
+  dst.x = 1.0
+  dst.y = 1.0
+  dst.z = 1.0
+  dst.w = 1.0
+
+
+1.5.23  TEX - Texture Lookup
+
+  TBD
+
+
+1.5.24  TXD - Texture Lookup with Derivatives
+
+  TBD
+
+
+1.5.25  TXP - Projective Texture Lookup
+
+  TBD
+
+
+1.5.26  UP2H - Unpack Two 16-Bit Floats
+
+  TBD
+
+
+1.5.27  UP2US - Unpack Two Unsigned 16-Bit Scalars
+
+  TBD
+
+
+1.5.28  UP4B - Unpack Four Signed 8-Bit Values
+
+  TBD
+
+
+1.5.29  UP4UB - Unpack Four Unsigned 8-Bit Scalars
+
+  TBD
+
+
+1.5.30  X2D - 2D Coordinate Transformation
+
+  dst.x = src0.x + src1.x * src2.x + src1.y * src2.y
+  dst.y = src0.y + src1.x * src2.z + src1.y * src2.w
+  dst.z = src0.x + src1.x * src2.x + src1.y * src2.y
+  dst.w = src0.y + src1.x * src2.z + src1.y * src2.w
+
+
+1.6  GL_NV_vertex_program2
+--------------------------
+
+
+1.6.1  ARA - Address Register Add
+
+  TBD
+
+
+1.6.2  ARR - Address Register Load With Round
+
+  dst.x = round(src.x)
+  dst.y = round(src.y)
+  dst.z = round(src.z)
+  dst.w = round(src.w)
+
+
+1.6.3  BRA - Branch
+
+  TBD
+
+
+1.6.4  CAL - Subroutine Call
+
+  push(pc)
+  pc = target
+
+
+1.6.5  RET - Subroutine Call Return
+
+  pc = pop()
+
+
+1.6.6  SSG - Set Sign
+
+  dst.x = (src.x > 0.0) ? 1.0 : (src.x < 0.0) ? -1.0 : 0.0
+  dst.y = (src.y > 0.0) ? 1.0 : (src.y < 0.0) ? -1.0 : 0.0
+  dst.z = (src.z > 0.0) ? 1.0 : (src.z < 0.0) ? -1.0 : 0.0
+  dst.w = (src.w > 0.0) ? 1.0 : (src.w < 0.0) ? -1.0 : 0.0
+
+
+1.7  GL_ARB_vertex_program
+--------------------------
+
+
+1.7.1  SWZ - Extended Swizzle
+
+  dst.x = src.x
+  dst.y = src.y
+  dst.z = src.z
+  dst.w = src.w
+
+
+1.7.2  XPD - Cross Product
+
+  Alias for CROSSPRODUCT.
+
+
+1.8  GL_ARB_fragment_program
+----------------------------
+
+
+1.8.1  CMP - Compare
+
+  dst.x = (src0.x < 0.0) ? src1.x : src2.x
+  dst.y = (src0.y < 0.0) ? src1.y : src2.y
+  dst.z = (src0.z < 0.0) ? src1.z : src2.z
+  dst.w = (src0.w < 0.0) ? src1.w : src2.w
+
+
+1.8.2  KIL - Conditional Discard
+
+  TBD
+
+
+1.8.3  SCS - Sine Cosine
+
+  dst.x = cos(src.x)
+  dst.y = sin(src.x)
+  dst.z = 0.0
+  dst.y = 1.0
+
+
+1.8.4  TXB - Texture Lookup With Bias
+
+  TBD
+
+
+1.9  GL_NV_fragment_program2
+----------------------------
+
+
+1.9.1  NRM - 3-component Vector Normalise
+
+  dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z)
+  dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z)
+  dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z)
+  dst.w = 1.0
+
+
+1.9.2  DIV - Divide
+
+  dst.x = src0.x / src1.x
+  dst.y = src0.y / src1.y
+  dst.z = src0.z / src1.z
+  dst.w = src0.w / src1.w
+
+
+1.9.3  DP2 - 2-component Dot Product
+
+  dst.x = src0.x * src1.x + src0.y * src1.y
+  dst.y = src0.x * src1.x + src0.y * src1.y
+  dst.z = src0.x * src1.x + src0.y * src1.y
+  dst.w = src0.x * src1.x + src0.y * src1.y
+
+
+1.9.4  DP2A - 2-component Dot Product And Add
+
+  Alias for DOT2ADD.
+
+
+1.9.5  TXL - Texture Lookup With LOD
+
+  TBD
+
+
+1.9.6  BRK - Break
+
+  TBD
+
+
+1.9.7  IF - If
+
+  TBD
+
+
+1.9.8  LOOP - Loop
+
+  TBD
+
+
+1.9.9  REP - Repeat
+
+  TBD
+
+
+1.9.10  ELSE - Else
+
+  TBD
+
+
+1.9.11  ENDIF - End If
+
+  TBD
+
+
+1.9.12  ENDLOOP - End Loop
+
+  TBD
+
+
+1.9.13  ENDREP - End Repeat
+
+  TBD
+
+
+1.10  GL_NV_vertex_program3
+---------------------------
+
+
+1.10.1  PUSHA - Push Address Register On Stack
+
+  push(src.x)
+  push(src.y)
+  push(src.z)
+  push(src.w)
+
+
+1.10.2  POPA - Pop Address Register From Stack
+
+  dst.w = pop()
+  dst.z = pop()
+  dst.y = pop()
+  dst.x = pop()
+
+
+1.11  GL_NV_gpu_program4
+------------------------
+
+
+1.11.1  CEIL - Ceiling
+
+  dst.x = ceil(src.x)
+  dst.y = ceil(src.y)
+  dst.z = ceil(src.z)
+  dst.w = ceil(src.w)
+
+
+1.11.2  I2F - Integer To Float
+
+  dst.x = (float) src.x
+  dst.y = (float) src.y
+  dst.z = (float) src.z
+  dst.w = (float) src.w
+
+
+1.11.3  NOT - Bitwise Not
+
+  dst.x = ~src.x
+  dst.y = ~src.y
+  dst.z = ~src.z
+  dst.w = ~src.w
+
+
+1.11.4  TRUNC - Truncate
+
+  dst.x = trunc(src.x)
+  dst.y = trunc(src.y)
+  dst.z = trunc(src.z)
+  dst.w = trunc(src.w)
+
+
+1.11.5  SHL - Shift Left
+
+  dst.x = src0.x << src1.x
+  dst.y = src0.y << src1.x
+  dst.z = src0.z << src1.x
+  dst.w = src0.w << src1.x
+
+
+1.11.6  SHR - Shift Right
+
+  dst.x = src0.x >> src1.x
+  dst.y = src0.y >> src1.x
+  dst.z = src0.z >> src1.x
+  dst.w = src0.w >> src1.x
+
+
+1.11.7  AND - Bitwise And
+
+  dst.x = src0.x & src1.x
+  dst.y = src0.y & src1.y
+  dst.z = src0.z & src1.z
+  dst.w = src0.w & src1.w
+
+
+1.11.8  OR - Bitwise Or
+
+  dst.x = src0.x | src1.x
+  dst.y = src0.y | src1.y
+  dst.z = src0.z | src1.z
+  dst.w = src0.w | src1.w
+
+
+1.11.9  MOD - Modulus
+
+  dst.x = src0.x % src1.x
+  dst.y = src0.y % src1.y
+  dst.z = src0.z % src1.z
+  dst.w = src0.w % src1.w
+
+
+1.11.10  XOR - Bitwise Xor
+
+  dst.x = src0.x ^ src1.x
+  dst.y = src0.y ^ src1.y
+  dst.z = src0.z ^ src1.z
+  dst.w = src0.w ^ src1.w
+
+
+1.11.11  SAD - Sum Of Absolute Differences
+
+  dst.x = abs(src0.x - src1.x) + src2.x
+  dst.y = abs(src0.y - src1.y) + src2.y
+  dst.z = abs(src0.z - src1.z) + src2.z
+  dst.w = abs(src0.w - src1.w) + src2.w
+
+
+1.11.12  TXF - Texel Fetch
+
+  TBD
+
+
+1.11.13  TXQ - Texture Size Query
+
+  TBD
+
+
+1.11.14  CONT - Continue
+
+  TBD
+
+
+1.12  GL_NV_geometry_program4
+-----------------------------
+
+
+1.12.1  EMIT - Emit
+
+  TBD
+
+
+1.12.2  ENDPRIM - End Primitive
+
+  TBD
+
+
+1.13  GLSL
+----------
+
+
+1.13.1  BGNLOOP2 - Begin Loop
+
+  TBD
+
+
+1.13.2  BGNSUB - Begin Subroutine
+
+  TBD
+
+
+1.13.3  ENDLOOP2 - End Loop
+
+  TBD
+
+
+1.13.4  ENDSUB - End Subroutine
+
+  TBD
+
+
+1.13.5  INT - Truncate
+
+  Alias for TRUNC.
+
+
+1.13.6  NOISE1 - 1D Noise
+
+  TBD
+
+
+1.13.7  NOISE2 - 2D Noise
+
+  TBD
+
+
+1.13.8  NOISE3 - 3D Noise
+
+  TBD
+
+
+1.13.9  NOISE4 - 4D Noise
+
+  TBD
+
+
+1.13.10  NOP - No Operation
+
+  Do nothing.
+
+
+1.14  ps_1_1
+------------
+
+
+1.14.1  TEXKILL - Conditional Discard
+
+  Alias for KIL.
+
+
+1.15  ps_1_4
+------------
+
+
+1.15.1  TEXLD - Texture Lookup
+
+  Alias for TEX.
+
+
+1.16  ps_2_0
+------------
+
+
+1.16.1  M4X4 - Multiply Matrix
+
+  Alias for MULTIPLYMATRIX.
+
+
+1.16.2  M4X3 - Multiply Matrix
+
+  Considered for removal from language.
+
+
+1.16.3  M3X4 - Multiply Matrix
+
+  Considered for removal from language.
+
+
+1.16.4  M3X3 - Multiply Matrix
+
+  Considered for removal from language.
+
+
+1.16.5  M3X2 - Multiply Matrix
+
+  Considered for removal from language.
+
+
+1.16.6  CRS - Cross Product
+
+  Alias for XPD.
+
+
+1.16.7  NRM4 - 4-component Vector Normalise
+
+  dst.x = src.x / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
+  dst.y = src.y / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
+  dst.z = src.z / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
+  dst.w = src.w / (src.x * src.x + src.y * src.y + src.z * src.z + src.w * src.w)
+
+
+1.16.8  SINCOS - Sine Cosine
+
+  Alias for SCS.
+
+
+1.16.9  TEXLDB - Texture Lookup With Bias
+
+  Alias for TXB.
+
+
+1.16.10  DP2ADD - 2-component Dot Product And Add
+
+  Alias for DP2A.
+
+
+1.17  ps_2_x
+------------
+
+
+1.17.1  CALL - Subroutine Call
+
+  Alias for CAL.
+
+
+1.17.2  CALLNZ - Subroutine Call If Not Zero
+
+  TBD
+
+
+1.17.3  IFC - If
+
+  TBD
+
+
+1.17.4  BREAK - Break
+
+  Alias for BRK.
+
+
+1.17.5  BREAKC - Break Conditional
+
+  TBD
+
+
+1.17.6  DSX - Derivative Relative To X
+
+  Alias for DDX.
+
+
+1.17.7  DSY - Derivative Relative To Y
+
+  Alias for DDY.
+
+
+1.17.8  TEXLDD - Texture Lookup with Derivatives
+
+  Alias for TXD.
+
+
+1.18  vs_1_1
+------------
+
+
+1.18.1  EXPP - Approximate Exponential Base 2
+
+  Alias for EXP.
+
+
+1.18.2  LOGP - Logarithm Base 2
+
+  Alias for LG2.
+
+
+1.19  vs_2_0
+------------
+
+
+1.19.1  SGN - Set Sign
+
+  Alias for SSG.
+
+
+1.19.2  MOVA - Move Address Register
+
+  Alias for ARR.
+
diff --git a/src/gallium/auxiliary/tgsi/tgsi_build.c b/src/gallium/auxiliary/tgsi/tgsi_build.c
index 17886540cf..a1891a140a 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_build.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_build.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_shader_tokens.h"
 #include "tgsi_build.h"
 #include "tgsi_parse.h"
diff --git a/src/gallium/auxiliary/tgsi/tgsi_dump.c b/src/gallium/auxiliary/tgsi/tgsi_dump.c
index ab2b1f2c58..a784b7cc3c 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_dump.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_dump.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_string.h"
 #include "tgsi_dump.h"
 #include "tgsi_info.h"
@@ -245,8 +245,12 @@ iter_declaration(
       }
    }
 
-   TXT( ", " );
-   ENM( decl->Declaration.Interpolate, interpolate_names );
+   if (iter->processor.Processor == TGSI_PROCESSOR_FRAGMENT &&
+       decl->Declaration.File == TGSI_FILE_INPUT)
+   {
+      TXT( ", " );
+      ENM( decl->Declaration.Interpolate, interpolate_names );
+   }
 
    if (decl->Declaration.Centroid) {
       TXT( ", CENTROID" );
diff --git a/src/gallium/auxiliary/tgsi/tgsi_dump_c.c b/src/gallium/auxiliary/tgsi/tgsi_dump_c.c
index 2ecf1e2f14..3dc61c48ca 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_dump_c.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_dump_c.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "util/u_string.h"
 #include "tgsi_dump_c.h"
 #include "tgsi_build.h"
diff --git a/src/gallium/auxiliary/tgsi/tgsi_exec.c b/src/gallium/auxiliary/tgsi/tgsi_exec.c
index ab641efb60..ba807e498f 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_exec.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_exec.c
@@ -486,6 +486,32 @@ micro_f2ut(
 #endif
 
 static void
+micro_float_clamp(union tgsi_exec_channel *dst,
+                  const union tgsi_exec_channel *src)
+{
+   uint i;
+
+   for (i = 0; i < 4; i++) {
+      if (src->f[i] > 0.0f) {
+         if (src->f[i] > 1.884467e+019f)
+            dst->f[i] = 1.884467e+019f;
+         else if (src->f[i] < 5.42101e-020f)
+            dst->f[i] = 5.42101e-020f;
+         else
+            dst->f[i] = src->f[i];
+      }
+      else {
+         if (src->f[i] < -1.884467e+019f)
+            dst->f[i] = -1.884467e+019f;
+         else if (src->f[i] > -5.42101e-020f)
+            dst->f[i] = -5.42101e-020f;
+         else
+            dst->f[i] = src->f[i];
+      }
+   }
+}
+
+static void
 micro_flr(
    union tgsi_exec_channel *dst,
    const union tgsi_exec_channel *src )
@@ -508,20 +534,6 @@ micro_frc(
 }
 
 static void
-micro_ge(
-   union tgsi_exec_channel *dst,
-   const union tgsi_exec_channel *src0,
-   const union tgsi_exec_channel *src1,
-   const union tgsi_exec_channel *src2,
-   const union tgsi_exec_channel *src3 )
-{
-   dst->f[0] = src0->f[0] >= src1->f[0] ? src2->f[0] : src3->f[0];
-   dst->f[1] = src0->f[1] >= src1->f[1] ? src2->f[1] : src3->f[1];
-   dst->f[2] = src0->f[2] >= src1->f[2] ? src2->f[2] : src3->f[2];
-   dst->f[3] = src0->f[3] >= src1->f[3] ? src2->f[3] : src3->f[3];
-}
-
-static void
 micro_i2f(
    union tgsi_exec_channel *dst,
    const union tgsi_exec_channel *src )
@@ -1614,6 +1626,7 @@ exec_tex(struct tgsi_exec_machine *mach,
 
    switch (inst->InstructionExtTexture.Texture) {
    case TGSI_TEXTURE_1D:
+   case TGSI_TEXTURE_SHADOW1D:
 
       FETCH(&r[0], 0, CHAN_X);
 
@@ -1636,6 +1649,8 @@ exec_tex(struct tgsi_exec_machine *mach,
 
    case TGSI_TEXTURE_2D:
    case TGSI_TEXTURE_RECT:
+   case TGSI_TEXTURE_SHADOW2D:
+   case TGSI_TEXTURE_SHADOWRECT:
 
       FETCH(&r[0], 0, CHAN_X);
       FETCH(&r[1], 0, CHAN_Y);
@@ -1826,12 +1841,14 @@ exec_instruction(
    int *pc )
 {
    uint chan_index;
-   union tgsi_exec_channel r[8];
+   union tgsi_exec_channel r[10];
 
    (*pc)++;
 
    switch (inst->Instruction.Opcode) {
    case TGSI_OPCODE_ARL:
+   /* TGSI_OPCODE_FLOOR */
+   /* TGSI_OPCODE_FLR */
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          micro_flr( &r[0], &r[0] );
@@ -1849,31 +1866,31 @@ exec_instruction(
 
    case TGSI_OPCODE_LIT:
       if (IS_CHANNEL_ENABLED( *inst, CHAN_X )) {
-	 STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
+         STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
       }
 
       if (IS_CHANNEL_ENABLED( *inst, CHAN_Y ) || IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
-	 FETCH( &r[0], 0, CHAN_X );
-	 if (IS_CHANNEL_ENABLED( *inst, CHAN_Y )) {
-	    micro_max( &r[0], &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-	    STORE( &r[0], 0, CHAN_Y );
-	 }
+         FETCH( &r[0], 0, CHAN_X );
+         if (IS_CHANNEL_ENABLED( *inst, CHAN_Y )) {
+            micro_max( &r[0], &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
+            STORE( &r[0], 0, CHAN_Y );
+         }
 
-	 if (IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
-	    FETCH( &r[1], 0, CHAN_Y );
-	    micro_max( &r[1], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
+         if (IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
+            FETCH( &r[1], 0, CHAN_Y );
+            micro_max( &r[1], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
 
-	    FETCH( &r[2], 0, CHAN_W );
-	    micro_min( &r[2], &r[2], &mach->Temps[TEMP_128_I].xyzw[TEMP_128_C] );
-	    micro_max( &r[2], &r[2], &mach->Temps[TEMP_M128_I].xyzw[TEMP_M128_C] );
-	    micro_pow( &r[1], &r[1], &r[2] );
-	    micro_lt( &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-	    STORE( &r[0], 0, CHAN_Z );
-	 }
+            FETCH( &r[2], 0, CHAN_W );
+            micro_min( &r[2], &r[2], &mach->Temps[TEMP_128_I].xyzw[TEMP_128_C] );
+            micro_max( &r[2], &r[2], &mach->Temps[TEMP_M128_I].xyzw[TEMP_M128_C] );
+            micro_pow( &r[1], &r[1], &r[2] );
+            micro_lt( &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
+            STORE( &r[0], 0, CHAN_Z );
+         }
       }
 
       if (IS_CHANNEL_ENABLED( *inst, CHAN_W )) {
-	 STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_W );
+         STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_W );
       }
       break;
 
@@ -1882,17 +1899,18 @@ exec_instruction(
       FETCH( &r[0], 0, CHAN_X );
       micro_div( &r[0], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &r[0] );
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
    case TGSI_OPCODE_RSQ:
    /* TGSI_OPCODE_RECIPSQRT */
       FETCH( &r[0], 0, CHAN_X );
+      micro_abs( &r[0], &r[0] );
       micro_sqrt( &r[0], &r[0] );
       micro_div( &r[0], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &r[0] );
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
@@ -2005,30 +2023,30 @@ exec_instruction(
        micro_add( &r[0], &r[0], &r[1] );
 
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
    case TGSI_OPCODE_DST:
       if (IS_CHANNEL_ENABLED( *inst, CHAN_X )) {
-	 STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
+         STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
       }
 
       if (IS_CHANNEL_ENABLED( *inst, CHAN_Y )) {
-	 FETCH( &r[0], 0, CHAN_Y );
-	 FETCH( &r[1], 1, CHAN_Y);
-	 micro_mul( &r[0], &r[0], &r[1] );
-	 STORE( &r[0], 0, CHAN_Y );
+         FETCH( &r[0], 0, CHAN_Y );
+         FETCH( &r[1], 1, CHAN_Y);
+         micro_mul( &r[0], &r[0], &r[1] );
+         STORE( &r[0], 0, CHAN_Y );
       }
 
       if (IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
-	 FETCH( &r[0], 0, CHAN_Z );
-	 STORE( &r[0], 0, CHAN_Z );
+         FETCH( &r[0], 0, CHAN_Z );
+         STORE( &r[0], 0, CHAN_Z );
       }
 
       if (IS_CHANNEL_ENABLED( *inst, CHAN_W )) {
-	 FETCH( &r[0], 1, CHAN_W );
-	 STORE( &r[0], 0, CHAN_W );
+         FETCH( &r[0], 1, CHAN_W );
+         STORE( &r[0], 0, CHAN_W );
       }
       break;
 
@@ -2071,7 +2089,7 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_ge( &r[0], &r[0], &r[1], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
+         micro_le( &r[0], &r[1], &r[0], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
          STORE( &r[0], 0, chan_index );
       }
       break;
@@ -2115,15 +2133,27 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_CND:
-      assert (0);
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         FETCH(&r[0], 0, chan_index);
+         FETCH(&r[1], 1, chan_index);
+         FETCH(&r[2], 2, chan_index);
+         micro_lt(&r[0], &mach->Temps[TEMP_HALF_I].xyzw[TEMP_HALF_C], &r[2], &r[0], &r[1]);
+         STORE(&r[0], 0, chan_index);
+      }
       break;
 
    case TGSI_OPCODE_CND0:
-      assert (0);
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         FETCH(&r[0], 0, chan_index);
+         FETCH(&r[1], 1, chan_index);
+         FETCH(&r[2], 2, chan_index);
+         micro_le(&r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[2], &r[0], &r[1]);
+         STORE(&r[0], 0, chan_index);
+      }
       break;
 
    case TGSI_OPCODE_DOT2ADD:
-      /* TGSI_OPCODE_DP2A */
+   /* TGSI_OPCODE_DP2A */
       FETCH( &r[0], 0, CHAN_X );
       FETCH( &r[1], 1, CHAN_X );
       micro_mul( &r[0], &r[0], &r[1] );
@@ -2142,10 +2172,12 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_INDEX:
+      /* XXX: considered for removal */
       assert (0);
       break;
 
    case TGSI_OPCODE_NEGATE:
+      /* XXX: considered for removal */
       assert (0);
       break;
 
@@ -2159,15 +2191,13 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_CLAMP:
-      assert (0);
-      break;
-
-   case TGSI_OPCODE_FLOOR:
-   /* TGSI_OPCODE_FLR */
-      FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-         FETCH( &r[0], 0, chan_index );
-         micro_flr( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         FETCH(&r[0], 0, chan_index);
+         FETCH(&r[1], 1, chan_index);
+         micro_max(&r[0], &r[0], &r[1]);
+         FETCH(&r[1], 2, chan_index);
+         micro_min(&r[0], &r[0], &r[1]);
+         STORE(&r[0], 0, chan_index);
       }
       break;
 
@@ -2181,7 +2211,7 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_EXPBASE2:
-    /* TGSI_OPCODE_EX2 */
+   /* TGSI_OPCODE_EX2 */
       FETCH(&r[0], 0, CHAN_X);
 
 #if FAST_MATH
@@ -2191,7 +2221,7 @@ exec_instruction(
 #endif
 
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
@@ -2205,19 +2235,19 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_POWER:
-      /* TGSI_OPCODE_POW */
+   /* TGSI_OPCODE_POW */
       FETCH(&r[0], 0, CHAN_X);
       FETCH(&r[1], 1, CHAN_X);
 
       micro_pow( &r[0], &r[0], &r[1] );
 
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
    case TGSI_OPCODE_CROSSPRODUCT:
-      /* TGSI_OPCODE_XPD */
+   /* TGSI_OPCODE_XPD */
       FETCH(&r[0], 0, CHAN_Y);
       FETCH(&r[1], 1, CHAN_Z);
 
@@ -2260,6 +2290,7 @@ exec_instruction(
       break;
 
     case TGSI_OPCODE_MULTIPLYMATRIX:
+       /* XXX: considered for removal */
        assert (0);
        break;
 
@@ -2274,7 +2305,12 @@ exec_instruction(
        break;
 
    case TGSI_OPCODE_RCC:
-      assert (0);
+      FETCH(&r[0], 0, CHAN_X);
+      micro_div(&r[0], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &r[0]);
+      micro_float_clamp(&r[0], &r[0]);
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&r[0], 0, chan_index);
+      }
       break;
 
    case TGSI_OPCODE_DPH:
@@ -2300,7 +2336,7 @@ exec_instruction(
       micro_add( &r[0], &r[0], &r[1] );
 
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
@@ -2310,7 +2346,7 @@ exec_instruction(
       micro_cos( &r[0], &r[0] );
 
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-	 STORE( &r[0], 0, chan_index );
+         STORE( &r[0], 0, chan_index );
       }
       break;
 
@@ -2355,7 +2391,52 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_RFL:
-      assert (0);
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_X) ||
+          IS_CHANNEL_ENABLED(*inst, CHAN_Y) ||
+          IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+         /* r0 = dp3(src0, src0) */
+         FETCH(&r[2], 0, CHAN_X);
+         micro_mul(&r[0], &r[2], &r[2]);
+         FETCH(&r[4], 0, CHAN_Y);
+         micro_mul(&r[8], &r[4], &r[4]);
+         micro_add(&r[0], &r[0], &r[8]);
+         FETCH(&r[6], 0, CHAN_Z);
+         micro_mul(&r[8], &r[6], &r[6]);
+         micro_add(&r[0], &r[0], &r[8]);
+
+         /* r1 = dp3(src0, src1) */
+         FETCH(&r[3], 1, CHAN_X);
+         micro_mul(&r[1], &r[2], &r[3]);
+         FETCH(&r[5], 1, CHAN_Y);
+         micro_mul(&r[8], &r[4], &r[5]);
+         micro_add(&r[1], &r[1], &r[8]);
+         FETCH(&r[7], 1, CHAN_Z);
+         micro_mul(&r[8], &r[6], &r[7]);
+         micro_add(&r[1], &r[1], &r[8]);
+
+         /* r1 = 2 * r1 / r0 */
+         micro_add(&r[1], &r[1], &r[1]);
+         micro_div(&r[1], &r[1], &r[0]);
+
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+            micro_mul(&r[2], &r[2], &r[1]);
+            micro_sub(&r[2], &r[2], &r[3]);
+            STORE(&r[2], 0, CHAN_X);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+            micro_mul(&r[4], &r[4], &r[1]);
+            micro_sub(&r[4], &r[4], &r[5]);
+            STORE(&r[4], 0, CHAN_Y);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+            micro_mul(&r[6], &r[6], &r[1]);
+            micro_sub(&r[6], &r[6], &r[7]);
+            STORE(&r[6], 0, CHAN_Z);
+         }
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
+         STORE(&mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_W);
+      }
       break;
 
    case TGSI_OPCODE_SEQ:
@@ -2370,7 +2451,9 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_SFL:
-      assert (0);
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], 0, chan_index);
+      }
       break;
 
    case TGSI_OPCODE_SGT:
@@ -2409,7 +2492,9 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_STR:
-      assert (0);
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, chan_index);
+      }
       break;
 
    case TGSI_OPCODE_TEX:
@@ -2466,7 +2551,40 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_X2D:
-      assert (0);
+      FETCH(&r[0], 1, CHAN_X);
+      FETCH(&r[1], 1, CHAN_Y);
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_X) ||
+          IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+         FETCH(&r[2], 2, CHAN_X);
+         micro_mul(&r[2], &r[2], &r[0]);
+         FETCH(&r[3], 2, CHAN_Y);
+         micro_mul(&r[3], &r[3], &r[1]);
+         micro_add(&r[2], &r[2], &r[3]);
+         FETCH(&r[3], 0, CHAN_X);
+         micro_add(&r[2], &r[2], &r[3]);
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+            STORE(&r[2], 0, CHAN_X);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+            STORE(&r[2], 0, CHAN_Z);
+         }
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Y) ||
+          IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
+         FETCH(&r[2], 2, CHAN_Z);
+         micro_mul(&r[2], &r[2], &r[0]);
+         FETCH(&r[3], 2, CHAN_W);
+         micro_mul(&r[3], &r[3], &r[1]);
+         micro_add(&r[2], &r[2], &r[3]);
+         FETCH(&r[3], 0, CHAN_Y);
+         micro_add(&r[2], &r[2], &r[3]);
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+            STORE(&r[2], 0, CHAN_Y);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
+            STORE(&r[2], 0, CHAN_W);
+         }
+      }
       break;
 
    case TGSI_OPCODE_ARA:
@@ -2551,14 +2669,14 @@ exec_instruction(
    case TGSI_OPCODE_SCS:
       if( IS_CHANNEL_ENABLED( *inst, CHAN_X ) || IS_CHANNEL_ENABLED( *inst, CHAN_Y ) ) {
          FETCH( &r[0], 0, CHAN_X );
-      }
-      if( IS_CHANNEL_ENABLED( *inst, CHAN_X ) ) {
-         micro_cos( &r[1], &r[0] );
-         STORE( &r[1], 0, CHAN_X );
-      }
-      if( IS_CHANNEL_ENABLED( *inst, CHAN_Y ) ) {
-         micro_sin( &r[1], &r[0] );
-         STORE( &r[1], 0, CHAN_Y );
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+            micro_cos(&r[1], &r[0]);
+            STORE(&r[1], 0, CHAN_X);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+            micro_sin(&r[1], &r[0]);
+            STORE(&r[1], 0, CHAN_Y);
+         }
       }
       if( IS_CHANNEL_ENABLED( *inst, CHAN_Z ) ) {
          STORE( &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], 0, CHAN_Z );
@@ -2570,32 +2688,36 @@ exec_instruction(
 
    case TGSI_OPCODE_NRM:
       /* 3-component vector normalize */
-      {
-         union tgsi_exec_channel tmp, dot;
-
-         /* tmp = dp3(src0, src0): */
-         FETCH( &r[0], 0, CHAN_X );
-         micro_mul( &tmp, &r[0], &r[0] );
-
-         FETCH( &r[1], 0, CHAN_Y );
-         micro_mul( &dot, &r[1], &r[1] );
-         micro_add( &tmp, &tmp, &dot );
-
-         FETCH( &r[2], 0, CHAN_Z );
-         micro_mul( &dot, &r[2], &r[2] );
-         micro_add( &tmp, &tmp, &dot );
-
-         /* tmp = 1 / sqrt(tmp) */
-         micro_sqrt( &tmp, &tmp );
-         micro_div( &tmp, &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &tmp );
-
-         /* note: w channel is undefined */
-         FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-            /* chan = chan * tmp */
-            micro_mul( &r[chan_index], &tmp, &r[chan_index] );
-            STORE( &r[chan_index], 0, chan_index );
+      if(IS_CHANNEL_ENABLED(*inst, CHAN_X) ||
+         IS_CHANNEL_ENABLED(*inst, CHAN_Y) ||
+         IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+         /* r3 = sqrt(dp3(src0, src0)) */
+         FETCH(&r[0], 0, CHAN_X);
+         micro_mul(&r[3], &r[0], &r[0]);
+         FETCH(&r[1], 0, CHAN_Y);
+         micro_mul(&r[4], &r[1], &r[1]);
+         micro_add(&r[3], &r[3], &r[4]);
+         FETCH(&r[2], 0, CHAN_Z);
+         micro_mul(&r[4], &r[2], &r[2]);
+         micro_add(&r[3], &r[3], &r[4]);
+         micro_sqrt(&r[3], &r[3]);
+
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+            micro_div(&r[0], &r[0], &r[3]);
+            STORE(&r[0], 0, CHAN_X);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+            micro_div(&r[1], &r[1], &r[3]);
+            STORE(&r[1], 0, CHAN_Y);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+            micro_div(&r[2], &r[2], &r[3]);
+            STORE(&r[2], 0, CHAN_Z);
          }
       }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
+         STORE(&mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_W);
+      }
       break;
 
    case TGSI_OPCODE_NRM4:
@@ -2956,5 +3078,3 @@ tgsi_exec_machine_run( struct tgsi_exec_machine *mach )
 
    return ~mach->Temps[TEMP_KILMASK_I].xyzw[TEMP_KILMASK_C].u[0];
 }
-
-
diff --git a/src/gallium/auxiliary/tgsi/tgsi_info.c b/src/gallium/auxiliary/tgsi/tgsi_info.c
index 68c7a6b7f5..2b8a6f0fb1 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_info.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_info.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "tgsi_info.h"
 
 static const struct tgsi_opcode_info opcode_info[TGSI_OPCODE_LAST] =
diff --git a/src/gallium/auxiliary/tgsi/tgsi_iterate.c b/src/gallium/auxiliary/tgsi/tgsi_iterate.c
index 5371a88b96..d88c2558d8 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_iterate.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_iterate.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "tgsi_iterate.h"
 
 boolean
diff --git a/src/gallium/auxiliary/tgsi/tgsi_parse.c b/src/gallium/auxiliary/tgsi/tgsi_parse.c
index d374b16f9a..0081f74ffc 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_parse.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_parse.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_shader_tokens.h"
 #include "tgsi_parse.h"
 #include "tgsi_build.h"
@@ -154,10 +154,17 @@ tgsi_parse_token(
 
       switch (imm->Immediate.DataType) {
       case TGSI_IMM_FLOAT32:
-         imm->u.Pointer = MALLOC(
-            sizeof( struct tgsi_immediate_float32 ) * (imm->Immediate.NrTokens - 1) );
-         for( i = 0; i < imm->Immediate.NrTokens - 1; i++ ) {
-            next_token( ctx, (struct tgsi_immediate_float32 *) &imm->u.ImmediateFloat32[i] );
+         {
+            uint imm_count = imm->Immediate.NrTokens - 1;
+            struct tgsi_immediate_float32 *data;
+
+            data = (struct tgsi_immediate_float32 *) MALLOC(sizeof(struct tgsi_immediate_float32) * imm_count);
+            if (data) {
+               for (i = 0; i < imm_count; i++) {
+                  next_token(ctx, &data[i]);
+               }
+               imm->u.ImmediateFloat32 = data;
+            }
          }
          break;
 
diff --git a/src/gallium/auxiliary/tgsi/tgsi_ppc.c b/src/gallium/auxiliary/tgsi/tgsi_ppc.c
index f365030e52..0c64ae5713 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_ppc.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_ppc.c
@@ -33,7 +33,7 @@
 
 #if defined(PIPE_ARCH_PPC)
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_shader_tokens.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
diff --git a/src/gallium/auxiliary/tgsi/tgsi_sanity.c b/src/gallium/auxiliary/tgsi/tgsi_sanity.c
index bc7b941b78..6f1f5c2b4b 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_sanity.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_sanity.c
@@ -25,23 +25,24 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "tgsi_sanity.h"
 #include "tgsi_info.h"
 #include "tgsi_iterate.h"
 
-#define MAX_REGISTERS 256
-
 typedef uint reg_flag;
 
 #define BITS_IN_REG_FLAG (sizeof( reg_flag ) * 8)
 
+#define MAX_REGISTERS 256
+#define MAX_REG_FLAGS ((MAX_REGISTERS + BITS_IN_REG_FLAG - 1) / BITS_IN_REG_FLAG)
+
 struct sanity_check_ctx
 {
    struct tgsi_iterate_context iter;
 
-   reg_flag regs_decl[TGSI_FILE_COUNT][MAX_REGISTERS / BITS_IN_REG_FLAG];
-   reg_flag regs_used[TGSI_FILE_COUNT][MAX_REGISTERS / BITS_IN_REG_FLAG];
+   reg_flag regs_decl[TGSI_FILE_COUNT][MAX_REG_FLAGS];
+   reg_flag regs_used[TGSI_FILE_COUNT][MAX_REG_FLAGS];
    boolean regs_ind_used[TGSI_FILE_COUNT];
    uint num_imms;
    uint num_instructions;
@@ -89,7 +90,7 @@ check_file_name(
    uint file )
 {
    if (file <= TGSI_FILE_NULL || file >= TGSI_FILE_COUNT) {
-      report_error( ctx, "Invalid register file name" );
+      report_error( ctx, "(%u): Invalid register file name", file );
       return FALSE;
    }
    return TRUE;
@@ -113,7 +114,7 @@ is_any_register_declared(
 {
    uint i;
 
-   for (i = 0; i < MAX_REGISTERS / BITS_IN_REG_FLAG; i++)
+   for (i = 0; i < MAX_REG_FLAGS; i++)
       if (ctx->regs_decl[file][i])
          return TRUE;
    return FALSE;
@@ -162,9 +163,8 @@ check_register_usage(
       ctx->regs_ind_used[file] = TRUE;
    }
    else {
-      if (index < 0 || index > MAX_REGISTERS) {
-         report_error( ctx, "%s[%i]: Invalid index %s",
-                       file_names[file], index, name );
+      if (index < 0 || index >= MAX_REGISTERS) {
+         report_error( ctx, "%s[%d]: Invalid %s index", file_names[file], index, name );
          return FALSE;
       }
 
@@ -193,15 +193,15 @@ iter_instruction(
 
    info = tgsi_get_opcode_info( inst->Instruction.Opcode );
    if (info == NULL) {
-      report_error( ctx, "Invalid instruction opcode" );
+      report_error( ctx, "(%u): Invalid instruction opcode", inst->Instruction.Opcode );
       return TRUE;
    }
 
    if (info->num_dst != inst->Instruction.NumDstRegs) {
-      report_error( ctx, "Invalid number of destination operands" );
+      report_error( ctx, "Invalid number of destination operands, should be %u", info->num_dst );
    }
    if (info->num_src != inst->Instruction.NumSrcRegs) {
-      report_error( ctx, "Invalid number of source operands" );
+      report_error( ctx, "Invalid number of source operands, should be %u", info->num_src );
    }
 
    /* Check destination and source registers' validity.
@@ -266,7 +266,7 @@ iter_declaration(
       return TRUE;
    for (i = decl->DeclarationRange.First; i <= decl->DeclarationRange.Last; i++) {
       if (is_register_declared( ctx, file, i ))
-         report_error( ctx, "The same register declared twice" );
+         report_error( ctx, "%s[%u]: The same register declared more than once", file_names[file], i );
       ctx->regs_decl[file][i / BITS_IN_REG_FLAG] |= (1 << (i % BITS_IN_REG_FLAG));
    }
 
@@ -295,7 +295,7 @@ iter_immediate(
    /* Check data type validity.
     */
    if (imm->Immediate.DataType != TGSI_IMM_FLOAT32) {
-      report_error( ctx, "Invalid immediate data type" );
+      report_error( ctx, "(%u): Invalid immediate data type", imm->Immediate.DataType );
       return TRUE;
    }
 
@@ -322,7 +322,7 @@ epilog(
 
       for (i = 0; i < MAX_REGISTERS; i++) {
          if (is_register_declared( ctx, file, i ) && !is_register_used( ctx, file, i ) && !ctx->regs_ind_used[file]) {
-            report_warning( ctx, "Register never used" );
+            report_warning( ctx, "%s[%u]: Register never used", file_names[file], i );
          }
       }
    }
diff --git a/src/gallium/auxiliary/tgsi/tgsi_scan.c b/src/gallium/auxiliary/tgsi/tgsi_scan.c
index d02205a63e..c535788819 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_scan.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_scan.c
@@ -151,7 +151,14 @@ tgsi_scan_shader(const struct tgsi_token *tokens,
          break;
 
       case TGSI_TOKEN_TYPE_IMMEDIATE:
-         info->immediate_count++;
+         {
+            uint reg = info->immediate_count++;
+            uint file = TGSI_FILE_IMMEDIATE;
+
+            info->file_mask[file] |= (1 << reg);
+            info->file_count[file]++;
+            info->file_max[file] = MAX2(info->file_max[file], (int)reg);
+         }
          break;
 
       default:
diff --git a/src/gallium/auxiliary/tgsi/tgsi_sse2.c b/src/gallium/auxiliary/tgsi/tgsi_sse2.c
index 481ba89c5e..d70bcd03c5 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_sse2.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_sse2.c
@@ -29,7 +29,7 @@
 
 #if defined(PIPE_ARCH_X86)
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_shader_tokens.h"
 #include "util/u_math.h"
 #if defined(PIPE_ARCH_SSE)
@@ -1575,6 +1575,7 @@ emit_instruction(
    case TGSI_OPCODE_RSQ:
    /* TGSI_OPCODE_RECIPSQRT */
       FETCH( func, *inst, 0, 0, CHAN_X );
+      emit_abs( func, 0 );
       emit_rsqrt( func, 1, 0 );
       FOR_EACH_DST0_ENABLED_CHANNEL( *inst, chan_index ) {
          STORE( func, *inst, 1, 0, chan_index );
diff --git a/src/gallium/auxiliary/tgsi/tgsi_text.c b/src/gallium/auxiliary/tgsi/tgsi_text.c
index 1e822fbbea..58fe07c11d 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_text.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_text.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "tgsi_text.h"
 #include "tgsi_build.h"
 #include "tgsi_info.h"
diff --git a/src/gallium/auxiliary/tgsi/tgsi_transform.c b/src/gallium/auxiliary/tgsi/tgsi_transform.c
index ea87da31e5..062c1be938 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_transform.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_transform.c
@@ -31,7 +31,7 @@
  * Authors:  Brian Paul
  */
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "tgsi_transform.h"
 
diff --git a/src/gallium/auxiliary/tgsi/tgsi_util.c b/src/gallium/auxiliary/tgsi/tgsi_util.c
index 50101a9bb0..71f8a6ca40 100644
--- a/src/gallium/auxiliary/tgsi/tgsi_util.c
+++ b/src/gallium/auxiliary/tgsi/tgsi_util.c
@@ -25,7 +25,7 @@
  * 
  **************************************************************************/
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_shader_tokens.h"
 #include "tgsi_parse.h"
 #include "tgsi_build.h"
diff --git a/src/gallium/auxiliary/translate/Makefile b/src/gallium/auxiliary/translate/Makefile
index ad2a5b705e..3c82f8ae03 100644
--- a/src/gallium/auxiliary/translate/Makefile
+++ b/src/gallium/auxiliary/translate/Makefile
@@ -10,6 +10,3 @@ C_SOURCES = \
         translate_cache.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/util/Makefile b/src/gallium/auxiliary/util/Makefile
index 44c2377721..160df8dfa7 100644
--- a/src/gallium/auxiliary/util/Makefile
+++ b/src/gallium/auxiliary/util/Makefile
@@ -4,7 +4,7 @@ include $(TOP)/configs/current
 LIBNAME = util
 
 C_SOURCES = \
-	p_debug.c \
+	u_debug.c \
 	u_blit.c \
 	u_cache.c \
 	u_draw_quad.c \
@@ -27,6 +27,3 @@ C_SOURCES = \
 	u_simple_screen.c
 
 include ../../Makefile.template
-
-symlinks:
-
diff --git a/src/gallium/auxiliary/util/SConscript b/src/gallium/auxiliary/util/SConscript
index 35f683fb8e..9d5dd006f0 100644
--- a/src/gallium/auxiliary/util/SConscript
+++ b/src/gallium/auxiliary/util/SConscript
@@ -3,11 +3,13 @@ Import('*')
 util = env.ConvenienceLibrary(
 	target = 'util',
 	source = [
-		'p_debug.c',
-		'p_debug_mem.c',
-		'p_debug_prof.c',
+		'u_bitmask.c',
 		'u_blit.c',
 		'u_cache.c',
+		'u_debug.c',
+		'u_debug_memory.c',
+		'u_debug_profile.c',
+		'u_debug_stack.c',
 		'u_draw_quad.c',
 		'u_gen_mipmap.c',
 		'u_handle_table.c',
diff --git a/src/gallium/auxiliary/util/u_bitmask.c b/src/gallium/auxiliary/util/u_bitmask.c
new file mode 100644
index 0000000000..77587c07ec
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_bitmask.c
@@ -0,0 +1,320 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+/**
+ * @file
+ * Generic bitmask implementation.
+ *  
+ * @author Jose Fonseca <jfonseca@vmware.com>
+ */
+
+
+#include "pipe/p_compiler.h"
+#include "util/u_debug.h"
+
+#include "util/u_memory.h"
+#include "util/u_bitmask.h"
+
+
+typedef uint32_t util_bitmask_word;  
+
+
+#define UTIL_BITMASK_INITIAL_WORDS 16
+#define UTIL_BITMASK_BITS_PER_BYTE 8
+#define UTIL_BITMASK_BITS_PER_WORD (sizeof(util_bitmask_word) * UTIL_BITMASK_BITS_PER_BYTE)
+
+
+struct util_bitmask
+{
+   util_bitmask_word *words;
+   
+   /** Number of bits we can currently hold */
+   unsigned size;
+   
+   /** Number of consecutive bits set at the start of the bitmask */
+   unsigned filled;
+};
+
+
+struct util_bitmask *
+util_bitmask_create(void)
+{
+   struct util_bitmask *bm;
+   
+   bm = MALLOC_STRUCT(util_bitmask);
+   if(!bm)
+      return NULL;
+   
+   bm->words = (util_bitmask_word *)CALLOC(UTIL_BITMASK_INITIAL_WORDS, sizeof(util_bitmask_word));
+   if(!bm->words) {
+      FREE(bm);
+      return NULL;
+   }
+   
+   bm->size = UTIL_BITMASK_INITIAL_WORDS * UTIL_BITMASK_BITS_PER_WORD;
+   bm->filled = 0;
+   
+   return bm;
+}
+
+
+/**
+ * Resize the bitmask if necessary 
+ */
+static INLINE boolean
+util_bitmask_resize(struct util_bitmask *bm,
+                    unsigned minimum_index)
+{
+   unsigned minimum_size = minimum_index + 1;
+   unsigned new_size;
+   util_bitmask_word *new_words;
+
+   /* Check integer overflow */
+   if(!minimum_size)
+      return FALSE;
+      
+   if(bm->size > minimum_size)
+      return TRUE;
+
+   assert(bm->size % UTIL_BITMASK_BITS_PER_WORD == 0);
+   new_size = bm->size;
+   while(!(new_size > minimum_size)) {
+      new_size *= 2;
+      /* Check integer overflow */
+      if(new_size < bm->size)
+         return FALSE;
+   }
+   assert(new_size);
+   assert(new_size % UTIL_BITMASK_BITS_PER_WORD == 0);
+   
+   new_words = (util_bitmask_word *)REALLOC((void *)bm->words,
+                                            bm->size / UTIL_BITMASK_BITS_PER_BYTE,
+                                            new_size / UTIL_BITMASK_BITS_PER_BYTE);
+   if(!new_words)
+      return FALSE;
+   
+   memset(new_words + bm->size/UTIL_BITMASK_BITS_PER_WORD, 
+          0, 
+          (new_size - bm->size)/UTIL_BITMASK_BITS_PER_BYTE);
+   
+   bm->size = new_size;
+   bm->words = new_words;
+   
+   return TRUE;
+}
+
+
+/**
+ * Lazily update the filled.
+ */
+static INLINE void
+util_bitmask_filled_set(struct util_bitmask *bm,
+                        unsigned index)
+{
+   assert(bm->filled <= bm->size);
+   assert(index <= bm->size);
+   
+   if(index == bm->filled) {
+      ++bm->filled;
+      assert(bm->filled <= bm->size);
+   }
+}
+
+static INLINE void
+util_bitmask_filled_unset(struct util_bitmask *bm,
+                          unsigned index)
+{
+   assert(bm->filled <= bm->size);
+   assert(index <= bm->size);
+   
+   if(index < bm->filled)
+      bm->filled = index;
+}
+
+
+unsigned
+util_bitmask_add(struct util_bitmask *bm)
+{
+   unsigned word;
+   unsigned bit;
+   util_bitmask_word mask;
+   
+   assert(bm);
+
+   /* linear search for an empty index */
+   word = bm->filled / UTIL_BITMASK_BITS_PER_WORD;
+   bit  = bm->filled % UTIL_BITMASK_BITS_PER_WORD;
+   mask = 1 << bit;
+   while(word < bm->size / UTIL_BITMASK_BITS_PER_WORD) {
+      while(bit < UTIL_BITMASK_BITS_PER_WORD) {
+         if(!(bm->words[word] & mask))
+            goto found;
+         ++bm->filled;
+         ++bit;
+         mask <<= 1;
+      }
+      ++word;
+      bit = 0;
+      mask = 1;
+   }
+found:
+   
+   /* grow the bitmask if necessary */
+   if(!util_bitmask_resize(bm, bm->filled))
+      return UTIL_BITMASK_INVALID_INDEX;
+
+   assert(!(bm->words[word] & mask));
+   bm->words[word] |= mask;
+
+   return bm->filled++;
+}
+
+
+unsigned
+util_bitmask_set(struct util_bitmask *bm, 
+                 unsigned index)
+{
+   unsigned word = index / UTIL_BITMASK_BITS_PER_WORD;
+   unsigned bit  = index % UTIL_BITMASK_BITS_PER_WORD;
+   util_bitmask_word mask = 1 << bit;
+   
+   assert(bm);
+   
+   /* grow the bitmask if necessary */
+   if(!util_bitmask_resize(bm, index))
+      return UTIL_BITMASK_INVALID_INDEX;
+
+   bm->words[word] |= mask;
+
+   util_bitmask_filled_set(bm, index);
+   
+   return index;
+}
+
+
+void
+util_bitmask_clear(struct util_bitmask *bm, 
+                   unsigned index)
+{
+   unsigned word = index / UTIL_BITMASK_BITS_PER_WORD;
+   unsigned bit  = index % UTIL_BITMASK_BITS_PER_WORD;
+   util_bitmask_word mask = 1 << bit;
+   
+   assert(bm);
+   
+   if(index >= bm->size)
+      return;
+
+   bm->words[word] &= ~mask;
+   
+   util_bitmask_filled_unset(bm, index);
+}
+
+
+boolean
+util_bitmask_get(struct util_bitmask *bm, 
+                 unsigned index)
+{
+   unsigned word = index / UTIL_BITMASK_BITS_PER_WORD;
+   unsigned bit  = index % UTIL_BITMASK_BITS_PER_WORD;
+   util_bitmask_word mask = 1 << bit;
+   
+   assert(bm);
+   
+   if(index < bm->filled) {
+      assert(bm->words[word] & mask);
+      return TRUE;
+   }
+
+   if(index > bm->size)
+      return FALSE;
+
+   if(bm->words[word] & mask) {
+      util_bitmask_filled_set(bm, index);
+      return TRUE;
+   }
+   else
+      return FALSE;
+}
+
+
+unsigned
+util_bitmask_get_next_index(struct util_bitmask *bm, 
+                            unsigned index)
+{
+   unsigned word = index / UTIL_BITMASK_BITS_PER_WORD;
+   unsigned bit  = index % UTIL_BITMASK_BITS_PER_WORD;
+   util_bitmask_word mask = 1 << bit;
+
+   if(index < bm->filled) {
+      assert(bm->words[word] & mask);
+      return index;
+   }
+
+   if(index >= bm->size) {
+      return UTIL_BITMASK_INVALID_INDEX;
+   }
+
+   /* Do a linear search */
+   while(word < bm->size / UTIL_BITMASK_BITS_PER_WORD) {
+      while(bit < UTIL_BITMASK_BITS_PER_WORD) {
+         if(bm->words[word] & mask) {
+            if(index == bm->filled) {
+               ++bm->filled;
+               assert(bm->filled <= bm->size);
+            }
+            return index;
+         }
+         ++index;
+         ++bit;
+         mask <<= 1;
+      }
+      ++word;
+      bit = 0;
+      mask = 1;
+   }
+   
+   return UTIL_BITMASK_INVALID_INDEX;
+}
+
+
+unsigned
+util_bitmask_get_first_index(struct util_bitmask *bm)
+{
+   return util_bitmask_get_next_index(bm, 0);
+}
+
+
+void
+util_bitmask_destroy(struct util_bitmask *bm)
+{
+   assert(bm);
+
+   FREE(bm->words);
+   FREE(bm);
+}
+
diff --git a/src/gallium/auxiliary/util/u_bitmask.h b/src/gallium/auxiliary/util/u_bitmask.h
new file mode 100644
index 0000000000..87f1110296
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_bitmask.h
@@ -0,0 +1,114 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+/**
+ * @file
+ * Generic bitmask.
+ *  
+ * @author Jose Fonseca <jfonseca@vmware.com>
+ */
+
+#ifndef U_HANDLE_BITMASK_H_
+#define U_HANDLE_BITMASK_H_
+
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+
+#define UTIL_BITMASK_INVALID_INDEX (~0U)
+   
+   
+/**
+ * Abstract data type to represent arbitrary set of bits.
+ */
+struct util_bitmask;
+
+
+struct util_bitmask *
+util_bitmask_create(void);
+
+
+/**
+ * Search a cleared bit and set it.
+ * 
+ * It searches for the first cleared bit.
+ * 
+ * Returns the bit index on success, or UTIL_BITMASK_INVALID_INDEX on out of 
+ * memory growing the bitmask.
+ */
+unsigned
+util_bitmask_add(struct util_bitmask *bm);
+
+/**
+ * Set a bit.
+ * 
+ * Returns the input index on success, or UTIL_BITMASK_INVALID_INDEX on out of 
+ * memory growing the bitmask.
+ */
+unsigned
+util_bitmask_set(struct util_bitmask *bm, 
+                 unsigned index);
+
+void
+util_bitmask_clear(struct util_bitmask *bm, 
+                   unsigned index);
+
+boolean
+util_bitmask_get(struct util_bitmask *bm, 
+                 unsigned index);
+
+
+void
+util_bitmask_destroy(struct util_bitmask *bm);
+
+
+/**
+ * Search for the first set bit.
+ * 
+ * Returns UTIL_BITMASK_INVALID_INDEX if a set bit cannot be found. 
+ */
+unsigned
+util_bitmask_get_first_index(struct util_bitmask *bm);
+
+
+/**
+ * Search for the first set bit, starting from the giving index.
+ * 
+ * Returns UTIL_BITMASK_INVALID_INDEX if a set bit cannot be found. 
+ */
+unsigned
+util_bitmask_get_next_index(struct util_bitmask *bm,
+                            unsigned index);
+
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* U_HANDLE_BITMASK_H_ */
diff --git a/src/gallium/auxiliary/util/u_blit.c b/src/gallium/auxiliary/util/u_blit.c
index 841e9c01e7..deb68c43a6 100644
--- a/src/gallium/auxiliary/util/u_blit.c
+++ b/src/gallium/auxiliary/util/u_blit.c
@@ -34,10 +34,11 @@
 
 
 #include "pipe/p_context.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/p_shader_tokens.h"
+#include "pipe/p_state.h"
 
 #include "util/u_blit.h"
 #include "util/u_draw_quad.h"
@@ -59,8 +60,6 @@ struct blit_state
    struct pipe_sampler_state sampler;
    struct pipe_viewport_state viewport;
 
-   struct pipe_shader_state vert_shader;
-   struct pipe_shader_state frag_shader;
    void *vs;
    void *fs;
 
@@ -103,8 +102,7 @@ util_create_blit(struct pipe_context *pipe, struct cso_context *cso)
    memset(&ctx->rasterizer, 0, sizeof(ctx->rasterizer));
    ctx->rasterizer.front_winding = PIPE_WINDING_CW;
    ctx->rasterizer.cull_mode = PIPE_WINDING_NONE;
-   ctx->rasterizer.bypass_clipping = 1;
-   /*ctx->rasterizer.bypass_vs = 1;*/
+   ctx->rasterizer.bypass_vs_clip_and_viewport = 1;
    ctx->rasterizer.gl_rasterization_rules = 1;
 
    /* samplers */
@@ -117,28 +115,20 @@ util_create_blit(struct pipe_context *pipe, struct cso_context *cso)
    ctx->sampler.mag_img_filter = 0; /* set later */
    ctx->sampler.normalized_coords = 1;
 
-   /* viewport (identity, we setup vertices in wincoords) */
-   ctx->viewport.scale[0] = 1.0;
-   ctx->viewport.scale[1] = 1.0;
-   ctx->viewport.scale[2] = 1.0;
-   ctx->viewport.scale[3] = 1.0;
-   ctx->viewport.translate[0] = 0.0;
-   ctx->viewport.translate[1] = 0.0;
-   ctx->viewport.translate[2] = 0.0;
-   ctx->viewport.translate[3] = 0.0;
-
-   /* vertex shader */
+
+   /* vertex shader - still required to provide the linkage between
+    * fragment shader input semantics and vertex_element/buffers.
+    */
    {
       const uint semantic_names[] = { TGSI_SEMANTIC_POSITION,
                                       TGSI_SEMANTIC_GENERIC };
       const uint semantic_indexes[] = { 0, 0 };
       ctx->vs = util_make_vertex_passthrough_shader(pipe, 2, semantic_names,
-                                                    semantic_indexes,
-                                                    &ctx->vert_shader);
+                                                    semantic_indexes);
    }
 
    /* fragment shader */
-   ctx->fs = util_make_fragment_tex_shader(pipe, &ctx->frag_shader);
+   ctx->fs = util_make_fragment_tex_shader(pipe);
    ctx->vbuf = NULL;
 
    /* init vertex data that doesn't change */
@@ -163,10 +153,7 @@ util_destroy_blit(struct blit_state *ctx)
    pipe->delete_vs_state(pipe, ctx->vs);
    pipe->delete_fs_state(pipe, ctx->fs);
 
-   FREE((void*) ctx->vert_shader.tokens);
-   FREE((void*) ctx->frag_shader.tokens);
-
-   pipe_buffer_reference(pipe->screen, &ctx->vbuf, NULL);
+   pipe_buffer_reference(&ctx->vbuf, NULL);
 
    FREE(ctx);
 }
@@ -199,7 +186,6 @@ static unsigned
 setup_vertex_data(struct blit_state *ctx,
                   float x0, float y0, float x1, float y1, float z)
 {
-   void *buf;
    unsigned offset;
 
    ctx->vertices[0][0][0] = x0;
@@ -228,12 +214,8 @@ setup_vertex_data(struct blit_state *ctx,
 
    offset = get_next_slot( ctx );
 
-   buf = pipe_buffer_map(ctx->pipe->screen, ctx->vbuf,
-                         PIPE_BUFFER_USAGE_CPU_WRITE);
-
-   memcpy((char *)buf + offset, ctx->vertices, sizeof(ctx->vertices));
-
-   pipe_buffer_unmap(ctx->pipe->screen, ctx->vbuf);
+   pipe_buffer_write(ctx->pipe->screen, ctx->vbuf,
+                     offset, sizeof(ctx->vertices), ctx->vertices);
 
    return offset;
 }
@@ -249,7 +231,6 @@ setup_vertex_data_tex(struct blit_state *ctx,
                       float s0, float t0, float s1, float t1,
                       float z)
 {
-   void *buf;
    unsigned offset;
 
    ctx->vertices[0][0][0] = x0;
@@ -278,12 +259,8 @@ setup_vertex_data_tex(struct blit_state *ctx,
 
    offset = get_next_slot( ctx );
 
-   buf = pipe_buffer_map(ctx->pipe->screen, ctx->vbuf,
-                         PIPE_BUFFER_USAGE_CPU_WRITE);
-
-   memcpy((char *)buf + offset, ctx->vertices, sizeof(ctx->vertices));
-
-   pipe_buffer_unmap(ctx->pipe->screen, ctx->vbuf);
+   pipe_buffer_write(ctx->pipe->screen, ctx->vbuf,
+                     offset, sizeof(ctx->vertices), ctx->vertices);
 
    return offset;
 }
@@ -337,7 +314,7 @@ util_blit_pixels(struct blit_state *ctx,
 
    if(dst->format == src->format && (dstX1 - dstX0) == srcW && (dstY1 - dstY0) == srcH) {
       /* FIXME: this will most surely fail for overlapping rectangles */
-      pipe->surface_copy(pipe, FALSE,
+      pipe->surface_copy(pipe,
 			 dst, dstX0, dstY0,   /* dest */
 			 src, srcX0, srcY0, /* src */
 			 srcW, srcH);     /* size */
@@ -371,14 +348,14 @@ util_blit_pixels(struct blit_state *ctx,
                                      PIPE_BUFFER_USAGE_GPU_WRITE);
 
    /* load temp texture */
-   pipe->surface_copy(pipe, FALSE,
+   pipe->surface_copy(pipe,
                       texSurf, 0, 0,   /* dest */
                       src, srcLeft, srcTop, /* src */
                       srcW, srcH);     /* size */
 
    /* free the surface, update the texture if necessary.
     */
-   screen->tex_surface_release(screen, &texSurf);
+   pipe_surface_reference(&texSurf, NULL);
 
    /* save state (restored below) */
    cso_save_blend(ctx->cso);
@@ -389,13 +366,11 @@ util_blit_pixels(struct blit_state *ctx,
    cso_save_framebuffer(ctx->cso);
    cso_save_fragment_shader(ctx->cso);
    cso_save_vertex_shader(ctx->cso);
-   cso_save_viewport(ctx->cso);
 
    /* set misc state we care about */
    cso_set_blend(ctx->cso, &ctx->blend);
    cso_set_depth_stencil_alpha(ctx->cso, &ctx->depthstencil);
    cso_set_rasterizer(ctx->cso, &ctx->rasterizer);
-   cso_set_viewport(ctx->cso, &ctx->viewport);
 
    /* sampler */
    ctx->sampler.min_img_filter = filter;
@@ -437,9 +412,8 @@ util_blit_pixels(struct blit_state *ctx,
    cso_restore_framebuffer(ctx->cso);
    cso_restore_fragment_shader(ctx->cso);
    cso_restore_vertex_shader(ctx->cso);
-   cso_restore_viewport(ctx->cso);
 
-   screen->texture_release(screen, &tex);
+   pipe_texture_reference(&tex, NULL);
 }
 
 
@@ -448,7 +422,7 @@ util_blit_pixels(struct blit_state *ctx,
  */
 void util_blit_flush( struct blit_state *ctx )
 {
-   pipe_buffer_reference(ctx->pipe->screen, &ctx->vbuf, NULL);
+   pipe_buffer_reference(&ctx->vbuf, NULL);
    ctx->vbuf_slot = 0;
 } 
 
@@ -471,8 +445,6 @@ util_blit_pixels_tex(struct blit_state *ctx,
                  int dstX1, int dstY1,
                  float z, uint filter)
 {
-   struct pipe_context *pipe = ctx->pipe;
-   struct pipe_screen *screen = pipe->screen;
    struct pipe_framebuffer_state fb;
    float s0, t0, s1, t1;
    unsigned offset;
@@ -488,8 +460,10 @@ util_blit_pixels_tex(struct blit_state *ctx,
    t0 = srcY0 / (float)tex->height[0];
    t1 = srcY1 / (float)tex->height[0];
 
-   assert(screen->is_format_supported(screen, dst->format, PIPE_TEXTURE_2D,
-                                      PIPE_TEXTURE_USAGE_RENDER_TARGET, 0));
+   assert(ctx->pipe->screen->is_format_supported(ctx->pipe->screen, dst->format,
+                                                 PIPE_TEXTURE_2D,
+                                                 PIPE_TEXTURE_USAGE_RENDER_TARGET,
+                                                 0));
 
    /* save state (restored below) */
    cso_save_blend(ctx->cso);
@@ -500,13 +474,11 @@ util_blit_pixels_tex(struct blit_state *ctx,
    cso_save_framebuffer(ctx->cso);
    cso_save_fragment_shader(ctx->cso);
    cso_save_vertex_shader(ctx->cso);
-   cso_save_viewport(ctx->cso);
 
    /* set misc state we care about */
    cso_set_blend(ctx->cso, &ctx->blend);
    cso_set_depth_stencil_alpha(ctx->cso, &ctx->depthstencil);
    cso_set_rasterizer(ctx->cso, &ctx->rasterizer);
-   cso_set_viewport(ctx->cso, &ctx->viewport);
 
    /* sampler */
    ctx->sampler.min_img_filter = filter;
@@ -551,5 +523,4 @@ util_blit_pixels_tex(struct blit_state *ctx,
    cso_restore_framebuffer(ctx->cso);
    cso_restore_fragment_shader(ctx->cso);
    cso_restore_vertex_shader(ctx->cso);
-   cso_restore_viewport(ctx->cso);
 }
diff --git a/src/gallium/auxiliary/util/u_cache.c b/src/gallium/auxiliary/util/u_cache.c
index 0a1a64259f..41cd38171f 100644
--- a/src/gallium/auxiliary/util/u_cache.c
+++ b/src/gallium/auxiliary/util/u_cache.c
@@ -36,7 +36,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "util/u_math.h"
 #include "util/u_memory.h"
diff --git a/src/gallium/auxiliary/util/p_debug.c b/src/gallium/auxiliary/util/u_debug.c
index f373f941dd..f96e27e09f 100644
--- a/src/gallium/auxiliary/util/p_debug.c
+++ b/src/gallium/auxiliary/util/u_debug.c
@@ -50,6 +50,7 @@
 #define WIN32_LEAN_AND_MEAN      // Exclude rarely-used stuff from Windows headers
 #endif
 #include <windows.h>
+#include <stdio.h>
 
 #else
 
@@ -59,7 +60,7 @@
 #endif
 
 #include "pipe/p_compiler.h" 
-#include "pipe/p_debug.h" 
+#include "util/u_debug.h" 
 #include "pipe/p_format.h" 
 #include "pipe/p_state.h" 
 #include "pipe/p_inlines.h" 
@@ -106,6 +107,12 @@ void _debug_vprintf(const char *format, va_list ap)
       OutputDebugStringA(buf);
       buf[0] = '\0';
    }
+   
+   if(GetConsoleWindow() && !IsDebuggerPresent()) {
+      vfprintf(stderr, format, ap);
+      fflush(stderr);
+   }
+   
 #elif defined(PIPE_SUBSYSTEM_WINDOWS_CE)
    wchar_t *wide_format;
    long wide_str_len;   
@@ -643,34 +650,37 @@ void debug_dump_image(const char *prefix,
 void debug_dump_surface(const char *prefix,
                         struct pipe_surface *surface)     
 {
-   unsigned surface_usage;
+   struct pipe_texture *texture;
+   struct pipe_screen *screen;
+   struct pipe_transfer *transfer;
    void *data;
 
    if (!surface)
-      goto error1;
+      return;
+
+   texture = surface->texture;
+   screen = texture->screen;
 
-   /* XXX: force mappable surface */
-   surface_usage = surface->usage;
-   surface->usage |= PIPE_BUFFER_USAGE_CPU_READ;
+   transfer = screen->get_tex_transfer(screen, texture, surface->face,
+                                       surface->level, surface->zslice,
+                                       PIPE_TRANSFER_READ, 0, 0, surface->width,
+                                       surface->height);
    
-   data = pipe_surface_map(surface, 
-                           PIPE_BUFFER_USAGE_CPU_READ);
+   data = screen->transfer_map(screen, transfer);
    if(!data)
-      goto error2;
+      goto error;
    
    debug_dump_image(prefix, 
-                    surface->format,
-                    surface->block.size, 
-                    surface->nblocksx,
-                    surface->nblocksy,
-                    surface->stride,
+                    transfer->format,
+                    transfer->block.size, 
+                    transfer->nblocksx,
+                    transfer->nblocksy,
+                    transfer->stride,
                     data);
    
-   pipe_surface_unmap(surface);
-error2:
-   surface->usage = surface_usage;
-error1:
-   ;
+   screen->transfer_unmap(screen, transfer);
+error:
+   screen->tex_transfer_destroy(transfer);
 }
 
 
@@ -710,8 +720,10 @@ debug_dump_surface_bmp(const char *filename,
                        struct pipe_surface *surface)
 {
 #ifndef PIPE_SUBSYSTEM_WINDOWS_MINIPORT
+   struct pipe_texture *texture;
+   struct pipe_screen *screen;
    struct util_stream *stream;
-   unsigned surface_usage;
+   struct pipe_transfer *transfer;
    struct bmp_file_header bmfh;
    struct bmp_info_header bmih;
    float *rgba;
@@ -748,14 +760,18 @@ debug_dump_surface_bmp(const char *filename,
    
    util_stream_write(stream, &bmfh, 14);
    util_stream_write(stream, &bmih, 40);
+
+   texture = surface->texture;
+   screen = texture->screen;
    
-   /* XXX: force mappable surface */
-   surface_usage = surface->usage;
-   surface->usage |= PIPE_BUFFER_USAGE_CPU_READ;
+   transfer = screen->get_tex_transfer(screen, texture, surface->face,
+                                       surface->level, surface->zslice,
+                                       PIPE_TRANSFER_READ, 0, 0, surface->width,
+                                       surface->height);
 
    y = surface->height;
    while(y--) {
-      pipe_get_tile_rgba(surface,
+      pipe_get_tile_rgba(transfer,
                          0, y, surface->width, 1,
                          rgba);
       for(x = 0; x < surface->width; ++x)
@@ -768,9 +784,9 @@ debug_dump_surface_bmp(const char *filename,
          util_stream_write(stream, &pixel, 4);
       }  
    }
-   
-   surface->usage = surface_usage;
 
+   screen->tex_transfer_destroy(transfer);
+   
    util_stream_close(stream);
 error2:
    FREE(rgba);
diff --git a/src/gallium/auxiliary/util/u_debug.h b/src/gallium/auxiliary/util/u_debug.h
new file mode 100644
index 0000000000..7c829707b2
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_debug.h
@@ -0,0 +1,361 @@
+/**************************************************************************
+ * 
+ * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+/**
+ * @file
+ * Cross-platform debugging helpers.
+ * 
+ * For now it just has assert and printf replacements, but it might be extended 
+ * with stack trace reports and more advanced logging in the near future. 
+ * 
+ * @author Jose Fonseca <jrfonseca@tungstengraphics.com>
+ */
+
+#ifndef U_DEBUG_H_
+#define U_DEBUG_H_
+
+
+#include <stdarg.h>
+
+#include "pipe/p_compiler.h"
+
+
+#ifdef	__cplusplus
+extern "C" {
+#endif
+
+
+#if defined(DBG) || defined(DEBUG)
+#ifndef DEBUG
+#define DEBUG 1
+#endif
+#else
+#ifndef NDEBUG
+#define NDEBUG 1
+#endif
+#endif
+
+   
+/* MSVC bebore VC7 does not have the __FUNCTION__ macro */
+#if defined(_MSC_VER) && _MSC_VER < 1300
+#define __FUNCTION__ "???"
+#endif
+
+
+void _debug_vprintf(const char *format, va_list ap);
+   
+
+static INLINE void
+_debug_printf(const char *format, ...)
+{
+   va_list ap;
+   va_start(ap, format);
+   _debug_vprintf(format, ap);
+   va_end(ap);
+}
+
+
+/**
+ * Print debug messages.
+ *
+ * The actual channel used to output debug message is platform specific. To 
+ * avoid misformating or truncation, follow these rules of thumb:   
+ * - output whole lines
+ * - avoid outputing large strings (512 bytes is the current maximum length 
+ * that is guaranteed to be printed in all platforms)
+ */
+static INLINE void
+debug_printf(const char *format, ...)
+{
+#ifdef DEBUG
+   va_list ap;
+   va_start(ap, format);
+   _debug_vprintf(format, ap);
+   va_end(ap);
+#else
+   (void) format; /* silence warning */
+#endif
+}
+
+
+#ifdef DEBUG
+#define debug_vprintf(_format, _ap) _debug_vprintf(_format, _ap)
+#else
+#define debug_vprintf(_format, _ap) ((void)0)
+#endif
+
+
+#ifdef DEBUG
+/**
+ * Dump a blob in hex to the same place that debug_printf sends its
+ * messages.
+ */
+void debug_print_blob( const char *name, const void *blob, unsigned size );
+
+/* Print a message along with a prettified format string
+ */
+void debug_print_format(const char *msg, unsigned fmt );
+#else
+#define debug_print_blob(_name, _blob, _size) ((void)0)
+#define debug_print_format(_msg, _fmt) ((void)0)
+#endif
+
+
+void _debug_break(void);
+
+
+/**
+ * Hard-coded breakpoint.
+ */
+#ifdef DEBUG
+#if defined(PIPE_ARCH_X86) && defined(PIPE_CC_GCC)
+#define debug_break() __asm("int3")
+#elif defined(PIPE_ARCH_X86) && defined(PIPE_CC_MSVC)
+#define debug_break()  do { _asm {int 3} } while(0)
+#else
+#define debug_break() _debug_break()
+#endif
+#else /* !DEBUG */
+#define debug_break() ((void)0)
+#endif /* !DEBUG */
+
+
+long
+debug_get_num_option(const char *name, long dfault);
+
+void _debug_assert_fail(const char *expr, 
+                        const char *file, 
+                        unsigned line, 
+                        const char *function);
+
+
+/** 
+ * Assert macro
+ * 
+ * Do not expect that the assert call terminates -- errors must be handled 
+ * regardless of assert behavior.
+ */
+#ifdef DEBUG
+#define debug_assert(expr) ((expr) ? (void)0 : _debug_assert_fail(#expr, __FILE__, __LINE__, __FUNCTION__))
+#else
+#define debug_assert(expr) ((void)(expr))
+#endif
+
+
+/** Override standard assert macro */
+#ifdef assert
+#undef assert
+#endif
+#define assert(expr) debug_assert(expr)
+
+
+/**
+ * Output the current function name.
+ */
+#ifdef DEBUG
+#define debug_checkpoint() \
+   _debug_printf("%s\n", __FUNCTION__)
+#else
+#define debug_checkpoint() \
+   ((void)0) 
+#endif
+
+
+/**
+ * Output the full source code position.
+ */
+#ifdef DEBUG
+#define debug_checkpoint_full() \
+   _debug_printf("%s:%u:%s", __FILE__, __LINE__, __FUNCTION__) 
+#else
+#define debug_checkpoint_full() \
+   ((void)0) 
+#endif
+
+
+/**
+ * Output a warning message. Muted on release version.
+ */
+#ifdef DEBUG
+#define debug_warning(__msg) \
+   _debug_printf("%s:%u:%s: warning: %s\n", __FILE__, __LINE__, __FUNCTION__, __msg)
+#else
+#define debug_warning(__msg) \
+   ((void)0) 
+#endif
+
+
+/**
+ * Output an error message. Not muted on release version.
+ */
+#ifdef DEBUG
+#define debug_error(__msg) \
+   _debug_printf("%s:%u:%s: error: %s\n", __FILE__, __LINE__, __FUNCTION__, __msg) 
+#else
+#define debug_error(__msg) \
+   _debug_printf("error: %s\n", __msg)
+#endif
+
+
+/**
+ * Used by debug_dump_enum and debug_dump_flags to describe symbols.
+ */
+struct debug_named_value
+{
+   const char *name;
+   unsigned long value;
+};
+
+
+/**
+ * Some C pre-processor magic to simplify creating named values.
+ * 
+ * Example:
+ * @code
+ * static const debug_named_value my_names[] = {
+ *    DEBUG_NAMED_VALUE(MY_ENUM_VALUE_X),
+ *    DEBUG_NAMED_VALUE(MY_ENUM_VALUE_Y),
+ *    DEBUG_NAMED_VALUE(MY_ENUM_VALUE_Z),
+ *    DEBUG_NAMED_VALUE_END
+ * };
+ * 
+ *    ...
+ *    debug_printf("%s = %s\n", 
+ *                 name,
+ *                 debug_dump_enum(my_names, my_value));
+ *    ...
+ * @endcode
+ */
+#define DEBUG_NAMED_VALUE(__symbol) {#__symbol, (unsigned long)__symbol} 
+#define DEBUG_NAMED_VALUE_END {NULL, 0} 
+
+
+/**
+ * Convert a enum value to a string.
+ */
+const char *
+debug_dump_enum(const struct debug_named_value *names, 
+                unsigned long value);
+
+const char *
+debug_dump_enum_noprefix(const struct debug_named_value *names, 
+                         const char *prefix,
+                         unsigned long value);
+
+
+/**
+ * Convert binary flags value to a string.
+ */
+const char *
+debug_dump_flags(const struct debug_named_value *names, 
+                 unsigned long value);
+
+
+/**
+ * Get option.
+ * 
+ * It is an alias for getenv on Linux. 
+ * 
+ * On Windows it reads C:\gallium.cfg, which is a text file with CR+LF line 
+ * endings with one option per line as
+ *  
+ *   NAME=value
+ * 
+ * This file must be terminated with an extra empty line.
+ */
+const char *
+debug_get_option(const char *name, const char *dfault);
+
+boolean
+debug_get_bool_option(const char *name, boolean dfault);
+
+long
+debug_get_num_option(const char *name, long dfault);
+
+unsigned long
+debug_get_flags_option(const char *name, 
+                       const struct debug_named_value *flags,
+                       unsigned long dfault);
+
+
+void *
+debug_malloc(const char *file, unsigned line, const char *function,
+             size_t size);
+
+void
+debug_free(const char *file, unsigned line, const char *function,
+           void *ptr);
+
+void *
+debug_calloc(const char *file, unsigned line, const char *function,
+             size_t count, size_t size );
+
+void *
+debug_realloc(const char *file, unsigned line, const char *function,
+              void *old_ptr, size_t old_size, size_t new_size );
+
+unsigned long
+debug_memory_begin(void);
+
+void 
+debug_memory_end(unsigned long beginning);
+
+
+#if defined(PROFILE) && defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY)
+
+void
+debug_profile_start(void);
+
+void 
+debug_profile_stop(void);
+
+#endif
+
+
+#ifdef DEBUG
+struct pipe_surface;
+void debug_dump_image(const char *prefix,
+                      unsigned format, unsigned cpp,
+                      unsigned width, unsigned height,
+                      unsigned stride,
+                      const void *data);
+void debug_dump_surface(const char *prefix,
+                        struct pipe_surface *surface);   
+void debug_dump_surface_bmp(const char *filename,
+                            struct pipe_surface *surface);
+#else
+#define debug_dump_image(prefix, format, cpp, width, height, stride, data) ((void)0)
+#define debug_dump_surface(prefix, surface) ((void)0)
+#define debug_dump_surface_bmp(filename, surface) ((void)0)
+#endif
+
+
+#ifdef	__cplusplus
+}
+#endif
+
+#endif /* U_DEBUG_H_ */
diff --git a/src/gallium/auxiliary/util/p_debug_mem.c b/src/gallium/auxiliary/util/u_debug_memory.c
index 250fd60f63..7623cb9398 100644
--- a/src/gallium/auxiliary/util/p_debug_mem.c
+++ b/src/gallium/auxiliary/util/u_debug_memory.c
@@ -44,11 +44,13 @@
 #include <stdlib.h>
 #endif
 
-#include "pipe/p_debug.h" 
+#include "util/u_debug.h" 
+#include "util/u_debug_stack.h" 
 #include "util/u_double_list.h" 
 
 
 #define DEBUG_MEMORY_MAGIC 0x6e34090aU 
+#define DEBUG_MEMORY_STACK 0 /* XXX: disabled until we have symbol lookup */
 
 
 #if defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY) && !defined(WINCE)
@@ -71,7 +73,11 @@ struct debug_memory_header
    const char *file;
    unsigned line;
    const char *function;
+#if DEBUG_MEMORY_STACK
+   struct debug_stack_frame backtrace[DEBUG_MEMORY_STACK];
+#endif
    size_t size;
+   
    unsigned magic;
 };
 
@@ -136,6 +142,10 @@ debug_malloc(const char *file, unsigned line, const char *function,
    hdr->size = size;
    hdr->magic = DEBUG_MEMORY_MAGIC;
 
+#if DEBUG_MEMORY_STACK
+   debug_backtrace_capture(hdr->backtrace, 0, DEBUG_MEMORY_STACK);
+#endif
+
    ftr = footer_from_header(hdr);
    ftr->magic = DEBUG_MEMORY_MAGIC;
    
@@ -290,6 +300,9 @@ debug_memory_end(unsigned long start_no)
 	 debug_printf("%s:%u:%s: %u bytes at %p not freed\n",
 		      hdr->file, hdr->line, hdr->function,
 		      hdr->size, ptr);
+#if DEBUG_MEMORY_STACK
+	 debug_backtrace_dump(hdr->backtrace, DEBUG_MEMORY_STACK);
+#endif
 	 total_size += hdr->size;
       }
 
diff --git a/src/gallium/auxiliary/util/p_debug_prof.c b/src/gallium/auxiliary/util/u_debug_profile.c
index 5f9772ef91..6d8b244c3a 100644
--- a/src/gallium/auxiliary/util/p_debug_prof.c
+++ b/src/gallium/auxiliary/util/u_debug_profile.c
@@ -42,7 +42,7 @@
 #include <windows.h>
 #include <winddi.h>
 
-#include "pipe/p_debug.h" 
+#include "util/u_debug.h" 
 #include "util/u_string.h" 
 
 
diff --git a/src/gallium/auxiliary/util/u_debug_stack.c b/src/gallium/auxiliary/util/u_debug_stack.c
new file mode 100644
index 0000000000..76068a6509
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_debug_stack.c
@@ -0,0 +1,97 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+/**
+ * @file
+ * Stack backtracing.
+ * 
+ * @author Jose Fonseca <jfonseca@vmware.com>
+ */
+
+#include "u_debug.h"
+#include "u_debug_stack.h"
+
+
+void
+debug_backtrace_capture(struct debug_stack_frame *backtrace,
+                        unsigned start_frame, 
+                        unsigned nr_frames)
+{
+   const void **frame_pointer = NULL;
+   unsigned i = 0;
+
+   if(!nr_frames)
+      return;
+
+#if defined(PIPE_CC_GCC)
+   frame_pointer = ((const void **)__builtin_frame_address(1));
+#elif defined(PIPE_CC_MSVC)
+   __asm {
+      mov frame_pointer, ebp
+   }
+   frame_pointer = (const void **)frame_pointer[0];
+#else
+   frame_pointer = NULL;
+#endif
+  
+   
+#ifdef PIPE_ARCH_X86
+   while(nr_frames) {
+      if(!frame_pointer)
+         break;
+      
+      if(start_frame)
+         --start_frame;
+      else {
+         backtrace[i++].function = frame_pointer[1];
+         --nr_frames;
+      }
+      
+      frame_pointer = (const void **)frame_pointer[0];
+   }
+#endif
+   
+   while(nr_frames) {
+      backtrace[i++].function = NULL;
+      --nr_frames;
+   }
+}
+   
+
+void
+debug_backtrace_dump(const struct debug_stack_frame *backtrace, 
+                     unsigned nr_frames)
+{
+   unsigned i;
+   
+   for(i = 0; i < nr_frames; ++i) {
+      if(!backtrace[i].function)
+         break;
+      debug_printf("\t%p\n", backtrace[i].function);
+   }
+}
+
diff --git a/src/gallium/auxiliary/util/u_debug_stack.h b/src/gallium/auxiliary/util/u_debug_stack.h
new file mode 100644
index 0000000000..f50f04e0f7
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_debug_stack.h
@@ -0,0 +1,65 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef U_DEBUG_STACK_H_
+#define U_DEBUG_STACK_H_
+
+
+/**
+ * @file
+ * Stack backtracing.
+ * 
+ * @author Jose Fonseca <jfonseca@vmware.com>
+ */
+
+
+#ifdef	__cplusplus
+extern "C" {
+#endif
+
+
+struct debug_stack_frame 
+{
+   const void *function;
+};
+   
+
+void
+debug_backtrace_capture(struct debug_stack_frame *backtrace,
+                        unsigned start_frame, 
+                        unsigned nr_frames);
+
+void
+debug_backtrace_dump(const struct debug_stack_frame *backtrace, 
+                     unsigned nr_frames);
+
+
+#ifdef	__cplusplus
+}
+#endif
+
+#endif /* U_DEBUG_STACK_H_ */
diff --git a/src/gallium/auxiliary/util/u_draw_quad.c b/src/gallium/auxiliary/util/u_draw_quad.c
index f282f3d289..4110485fb1 100644
--- a/src/gallium/auxiliary/util/u_draw_quad.c
+++ b/src/gallium/auxiliary/util/u_draw_quad.c
@@ -51,9 +51,11 @@ util_draw_vertex_buffer(struct pipe_context *pipe,
    assert(num_attribs <= PIPE_MAX_ATTRIBS);
 
    /* tell pipe about the vertex buffer */
+   memset(&vbuffer, 0, sizeof(vbuffer));
    vbuffer.buffer = vbuf;
    vbuffer.stride = num_attribs * 4 * sizeof(float);  /* vertex size */
    vbuffer.buffer_offset = offset;
+   vbuffer.max_index = num_verts - 1;
    pipe->set_vertex_buffers(pipe, 1, &vbuffer);
 
    /* tell pipe about the vertex attributes */
@@ -127,6 +129,6 @@ util_draw_texquad(struct pipe_context *pipe,
          util_draw_vertex_buffer(pipe, vbuf, 0, PIPE_PRIM_TRIANGLE_FAN, 4, 2);
       }
 
-      pipe_buffer_reference(pipe->screen, &vbuf, NULL);
+      pipe_buffer_reference(&vbuf, NULL);
    }
 }
diff --git a/src/gallium/auxiliary/util/u_gen_mipmap.c b/src/gallium/auxiliary/util/u_gen_mipmap.c
index 2b4cdab6cf..b32ad1cbe9 100644
--- a/src/gallium/auxiliary/util/u_gen_mipmap.c
+++ b/src/gallium/auxiliary/util/u_gen_mipmap.c
@@ -35,10 +35,11 @@
 
 
 #include "pipe/p_context.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/p_shader_tokens.h"
+#include "pipe/p_state.h"
 
 #include "util/u_memory.h"
 #include "util/u_draw_quad.h"
@@ -61,10 +62,7 @@ struct gen_mipmap_state
    struct pipe_depth_stencil_alpha_state depthstencil;
    struct pipe_rasterizer_state rasterizer;
    struct pipe_sampler_state sampler;
-   struct pipe_viewport_state viewport;
 
-   struct pipe_shader_state vert_shader;
-   struct pipe_shader_state frag_shader;
    void *vs;
    void *fs;
 
@@ -78,15 +76,15 @@ struct gen_mipmap_state
 
 enum dtype
 {
-   UBYTE,
-   UBYTE_3_3_2,
-   USHORT,
-   USHORT_4_4_4_4,
-   USHORT_5_6_5,
-   USHORT_1_5_5_5_REV,
-   UINT,
-   FLOAT,
-   HALF_FLOAT
+   DTYPE_UBYTE,
+   DTYPE_UBYTE_3_3_2,
+   DTYPE_USHORT,
+   DTYPE_USHORT_4_4_4_4,
+   DTYPE_USHORT_5_6_5,
+   DTYPE_USHORT_1_5_5_5_REV,
+   DTYPE_UINT,
+   DTYPE_FLOAT,
+   DTYPE_HALF_FLOAT
 };
 
 
@@ -194,7 +192,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
    assert(srcWidth == dstWidth || srcWidth == 2 * dstWidth);
    */
 
-   if (datatype == UBYTE && comps == 4) {
+   if (datatype == DTYPE_UBYTE && comps == 4) {
       uint i, j, k;
       const ubyte(*rowA)[4] = (const ubyte(*)[4]) srcRowA;
       const ubyte(*rowB)[4] = (const ubyte(*)[4]) srcRowB;
@@ -207,7 +205,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i][3] = (rowA[j][3] + rowA[k][3] + rowB[j][3] + rowB[k][3]) / 4;
       }
    }
-   else if (datatype == UBYTE && comps == 3) {
+   else if (datatype == DTYPE_UBYTE && comps == 3) {
       uint i, j, k;
       const ubyte(*rowA)[3] = (const ubyte(*)[3]) srcRowA;
       const ubyte(*rowB)[3] = (const ubyte(*)[3]) srcRowB;
@@ -219,7 +217,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
       }
    }
-   else if (datatype == UBYTE && comps == 2) {
+   else if (datatype == DTYPE_UBYTE && comps == 2) {
       uint i, j, k;
       const ubyte(*rowA)[2] = (const ubyte(*)[2]) srcRowA;
       const ubyte(*rowB)[2] = (const ubyte(*)[2]) srcRowB;
@@ -230,7 +228,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) >> 2;
       }
    }
-   else if (datatype == UBYTE && comps == 1) {
+   else if (datatype == DTYPE_UBYTE && comps == 1) {
       uint i, j, k;
       const ubyte *rowA = (const ubyte *) srcRowA;
       const ubyte *rowB = (const ubyte *) srcRowB;
@@ -241,7 +239,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
       }
    }
 
-   else if (datatype == USHORT && comps == 4) {
+   else if (datatype == DTYPE_USHORT && comps == 4) {
       uint i, j, k;
       const ushort(*rowA)[4] = (const ushort(*)[4]) srcRowA;
       const ushort(*rowB)[4] = (const ushort(*)[4]) srcRowB;
@@ -254,7 +252,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i][3] = (rowA[j][3] + rowA[k][3] + rowB[j][3] + rowB[k][3]) / 4;
       }
    }
-   else if (datatype == USHORT && comps == 3) {
+   else if (datatype == DTYPE_USHORT && comps == 3) {
       uint i, j, k;
       const ushort(*rowA)[3] = (const ushort(*)[3]) srcRowA;
       const ushort(*rowB)[3] = (const ushort(*)[3]) srcRowB;
@@ -266,7 +264,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i][2] = (rowA[j][2] + rowA[k][2] + rowB[j][2] + rowB[k][2]) / 4;
       }
    }
-   else if (datatype == USHORT && comps == 2) {
+   else if (datatype == DTYPE_USHORT && comps == 2) {
       uint i, j, k;
       const ushort(*rowA)[2] = (const ushort(*)[2]) srcRowA;
       const ushort(*rowB)[2] = (const ushort(*)[2]) srcRowB;
@@ -277,7 +275,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i][1] = (rowA[j][1] + rowA[k][1] + rowB[j][1] + rowB[k][1]) / 4;
       }
    }
-   else if (datatype == USHORT && comps == 1) {
+   else if (datatype == DTYPE_USHORT && comps == 1) {
       uint i, j, k;
       const ushort *rowA = (const ushort *) srcRowA;
       const ushort *rowB = (const ushort *) srcRowB;
@@ -288,7 +286,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
       }
    }
 
-   else if (datatype == FLOAT && comps == 4) {
+   else if (datatype == DTYPE_FLOAT && comps == 4) {
       uint i, j, k;
       const float(*rowA)[4] = (const float(*)[4]) srcRowA;
       const float(*rowB)[4] = (const float(*)[4]) srcRowB;
@@ -305,7 +303,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
                       rowB[j][3] + rowB[k][3]) * 0.25F;
       }
    }
-   else if (datatype == FLOAT && comps == 3) {
+   else if (datatype == DTYPE_FLOAT && comps == 3) {
       uint i, j, k;
       const float(*rowA)[3] = (const float(*)[3]) srcRowA;
       const float(*rowB)[3] = (const float(*)[3]) srcRowB;
@@ -320,7 +318,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
                       rowB[j][2] + rowB[k][2]) * 0.25F;
       }
    }
-   else if (datatype == FLOAT && comps == 2) {
+   else if (datatype == DTYPE_FLOAT && comps == 2) {
       uint i, j, k;
       const float(*rowA)[2] = (const float(*)[2]) srcRowA;
       const float(*rowB)[2] = (const float(*)[2]) srcRowB;
@@ -333,7 +331,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
                       rowB[j][1] + rowB[k][1]) * 0.25F;
       }
    }
-   else if (datatype == FLOAT && comps == 1) {
+   else if (datatype == DTYPE_FLOAT && comps == 1) {
       uint i, j, k;
       const float *rowA = (const float *) srcRowA;
       const float *rowB = (const float *) srcRowB;
@@ -345,7 +343,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
    }
 
 #if 0
-   else if (datatype == HALF_FLOAT && comps == 4) {
+   else if (datatype == HALF_DTYPE_FLOAT && comps == 4) {
       uint i, j, k, comp;
       const half_float(*rowA)[4] = (const half_float(*)[4]) srcRowA;
       const half_float(*rowB)[4] = (const half_float(*)[4]) srcRowB;
@@ -362,7 +360,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          }
       }
    }
-   else if (datatype == HALF_FLOAT && comps == 3) {
+   else if (datatype == DTYPE_HALF_FLOAT && comps == 3) {
       uint i, j, k, comp;
       const half_float(*rowA)[3] = (const half_float(*)[3]) srcRowA;
       const half_float(*rowB)[3] = (const half_float(*)[3]) srcRowB;
@@ -379,7 +377,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          }
       }
    }
-   else if (datatype == HALF_FLOAT && comps == 2) {
+   else if (datatype == DTYPE_HALF_FLOAT && comps == 2) {
       uint i, j, k, comp;
       const half_float(*rowA)[2] = (const half_float(*)[2]) srcRowA;
       const half_float(*rowB)[2] = (const half_float(*)[2]) srcRowB;
@@ -396,7 +394,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          }
       }
    }
-   else if (datatype == HALF_FLOAT && comps == 1) {
+   else if (datatype == DTYPE_HALF_FLOAT && comps == 1) {
       uint i, j, k;
       const half_float *rowA = (const half_float *) srcRowA;
       const half_float *rowB = (const half_float *) srcRowB;
@@ -413,7 +411,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
    }
 #endif
 
-   else if (datatype == UINT && comps == 1) {
+   else if (datatype == DTYPE_UINT && comps == 1) {
       uint i, j, k;
       const uint *rowA = (const uint *) srcRowA;
       const uint *rowB = (const uint *) srcRowB;
@@ -424,7 +422,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
       }
    }
 
-   else if (datatype == USHORT_5_6_5 && comps == 3) {
+   else if (datatype == DTYPE_USHORT_5_6_5 && comps == 3) {
       uint i, j, k;
       const ushort *rowA = (const ushort *) srcRowA;
       const ushort *rowB = (const ushort *) srcRowB;
@@ -449,7 +447,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (blue << 11) | (green << 5) | red;
       }
    }
-   else if (datatype == USHORT_4_4_4_4 && comps == 4) {
+   else if (datatype == DTYPE_USHORT_4_4_4_4 && comps == 4) {
       uint i, j, k;
       const ushort *rowA = (const ushort *) srcRowA;
       const ushort *rowB = (const ushort *) srcRowB;
@@ -479,7 +477,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (alpha << 12) | (blue << 8) | (green << 4) | red;
       }
    }
-   else if (datatype == USHORT_1_5_5_5_REV && comps == 4) {
+   else if (datatype == DTYPE_USHORT_1_5_5_5_REV && comps == 4) {
       uint i, j, k;
       const ushort *rowA = (const ushort *) srcRowA;
       const ushort *rowB = (const ushort *) srcRowB;
@@ -509,7 +507,7 @@ do_row(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (alpha << 15) | (blue << 10) | (green << 5) | red;
       }
    }
-   else if (datatype == UBYTE_3_3_2 && comps == 3) {
+   else if (datatype == DTYPE_UBYTE_3_3_2 && comps == 3) {
       uint i, j, k;
       const ubyte *rowA = (const ubyte *) srcRowA;
       const ubyte *rowB = (const ubyte *) srcRowB;
@@ -570,7 +568,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
    assert(comps >= 1);
    assert(comps <= 4);
 
-   if ((datatype == UBYTE) && (comps == 4)) {
+   if ((datatype == DTYPE_UBYTE) && (comps == 4)) {
       DECLARE_ROW_POINTERS(ubyte, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -581,7 +579,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(3);
       }
    }
-   else if ((datatype == UBYTE) && (comps == 3)) {
+   else if ((datatype == DTYPE_UBYTE) && (comps == 3)) {
       DECLARE_ROW_POINTERS(ubyte, 3);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -591,7 +589,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(2);
       }
    }
-   else if ((datatype == UBYTE) && (comps == 2)) {
+   else if ((datatype == DTYPE_UBYTE) && (comps == 2)) {
       DECLARE_ROW_POINTERS(ubyte, 2);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -600,7 +598,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(1);
       }
    }
-   else if ((datatype == UBYTE) && (comps == 1)) {
+   else if ((datatype == DTYPE_UBYTE) && (comps == 1)) {
       DECLARE_ROW_POINTERS(ubyte, 1);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -608,7 +606,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(0);
       }
    }
-   else if ((datatype == USHORT) && (comps == 4)) {
+   else if ((datatype == DTYPE_USHORT) && (comps == 4)) {
       DECLARE_ROW_POINTERS(ushort, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -619,7 +617,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(3);
       }
    }
-   else if ((datatype == USHORT) && (comps == 3)) {
+   else if ((datatype == DTYPE_USHORT) && (comps == 3)) {
       DECLARE_ROW_POINTERS(ushort, 3);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -629,7 +627,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(2);
       }
    }
-   else if ((datatype == USHORT) && (comps == 2)) {
+   else if ((datatype == DTYPE_USHORT) && (comps == 2)) {
       DECLARE_ROW_POINTERS(ushort, 2);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -638,7 +636,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(1);
       }
    }
-   else if ((datatype == USHORT) && (comps == 1)) {
+   else if ((datatype == DTYPE_USHORT) && (comps == 1)) {
       DECLARE_ROW_POINTERS(ushort, 1);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -646,7 +644,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_3D(0);
       }
    }
-   else if ((datatype == FLOAT) && (comps == 4)) {
+   else if ((datatype == DTYPE_FLOAT) && (comps == 4)) {
       DECLARE_ROW_POINTERS(float, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -657,7 +655,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_F_3D(3);
       }
    }
-   else if ((datatype == FLOAT) && (comps == 3)) {
+   else if ((datatype == DTYPE_FLOAT) && (comps == 3)) {
       DECLARE_ROW_POINTERS(float, 3);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -667,7 +665,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_F_3D(2);
       }
    }
-   else if ((datatype == FLOAT) && (comps == 2)) {
+   else if ((datatype == DTYPE_FLOAT) && (comps == 2)) {
       DECLARE_ROW_POINTERS(float, 2);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -676,7 +674,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_F_3D(1);
       }
    }
-   else if ((datatype == FLOAT) && (comps == 1)) {
+   else if ((datatype == DTYPE_FLOAT) && (comps == 1)) {
       DECLARE_ROW_POINTERS(float, 1);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -684,7 +682,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_F_3D(0);
       }
    }
-   else if ((datatype == HALF_FLOAT) && (comps == 4)) {
+   else if ((datatype == DTYPE_HALF_FLOAT) && (comps == 4)) {
       DECLARE_ROW_POINTERS(half_float, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -695,7 +693,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_HF_3D(3);
       }
    }
-   else if ((datatype == HALF_FLOAT) && (comps == 3)) {
+   else if ((datatype == DTYPE_HALF_FLOAT) && (comps == 3)) {
       DECLARE_ROW_POINTERS(half_float, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -705,7 +703,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_HF_3D(2);
       }
    }
-   else if ((datatype == HALF_FLOAT) && (comps == 2)) {
+   else if ((datatype == DTYPE_HALF_FLOAT) && (comps == 2)) {
       DECLARE_ROW_POINTERS(half_float, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -714,7 +712,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_HF_3D(1);
       }
    }
-   else if ((datatype == HALF_FLOAT) && (comps == 1)) {
+   else if ((datatype == DTYPE_HALF_FLOAT) && (comps == 1)) {
       DECLARE_ROW_POINTERS(half_float, 4);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -722,7 +720,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          FILTER_HF_3D(0);
       }
    }
-   else if ((datatype == UINT) && (comps == 1)) {
+   else if ((datatype == DTYPE_UINT) && (comps == 1)) {
       const uint *rowA = (const uint *) srcRowA;
       const uint *rowB = (const uint *) srcRowB;
       const uint *rowC = (const uint *) srcRowC;
@@ -738,7 +736,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (float)((double) tmp * 0.125);
       }
    }
-   else if ((datatype == USHORT_5_6_5) && (comps == 3)) {
+   else if ((datatype == DTYPE_USHORT_5_6_5) && (comps == 3)) {
       DECLARE_ROW_POINTERS0(ushort);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -776,7 +774,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (b << 11) | (g << 5) | r;
       }
    }
-   else if ((datatype == USHORT_4_4_4_4) && (comps == 4)) {
+   else if ((datatype == DTYPE_USHORT_4_4_4_4) && (comps == 4)) {
       DECLARE_ROW_POINTERS0(ushort);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -825,7 +823,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (a << 12) | (b << 8) | (g << 4) | r;
       }
    }
-   else if ((datatype == USHORT_1_5_5_5_REV) && (comps == 4)) {
+   else if ((datatype == DTYPE_USHORT_1_5_5_5_REV) && (comps == 4)) {
       DECLARE_ROW_POINTERS0(ushort);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -874,7 +872,7 @@ do_row_3D(enum dtype datatype, uint comps, int srcWidth,
          dst[i] = (a << 15) | (b << 10) | (g << 5) | r;
       }
    }
-   else if ((datatype == UBYTE_3_3_2) && (comps == 3)) {
+   else if ((datatype == DTYPE_UBYTE_3_3_2) && (comps == 3)) {
       DECLARE_ROW_POINTERS0(ushort);
 
       for (i = j = 0, k = k0; i < (uint) dstWidth;
@@ -928,34 +926,34 @@ format_to_type_comps(enum pipe_format pformat,
    case PIPE_FORMAT_X8R8G8B8_UNORM:
    case PIPE_FORMAT_B8G8R8A8_UNORM:
    case PIPE_FORMAT_B8G8R8X8_UNORM:
-      *datatype = UBYTE;
+      *datatype = DTYPE_UBYTE;
       *comps = 4;
       return;
    case PIPE_FORMAT_A1R5G5B5_UNORM:
-      *datatype = USHORT_1_5_5_5_REV;
+      *datatype = DTYPE_USHORT_1_5_5_5_REV;
       *comps = 4;
       return;
    case PIPE_FORMAT_A4R4G4B4_UNORM:
-      *datatype = USHORT_4_4_4_4;
+      *datatype = DTYPE_USHORT_4_4_4_4;
       *comps = 4;
       return;
    case PIPE_FORMAT_R5G6B5_UNORM:
-      *datatype = USHORT_5_6_5;
+      *datatype = DTYPE_USHORT_5_6_5;
       *comps = 3;
       return;
    case PIPE_FORMAT_L8_UNORM:
    case PIPE_FORMAT_A8_UNORM:
    case PIPE_FORMAT_I8_UNORM:
-      *datatype = UBYTE;
+      *datatype = DTYPE_UBYTE;
       *comps = 1;
       return;
    case PIPE_FORMAT_A8L8_UNORM:
-      *datatype = UBYTE;
+      *datatype = DTYPE_UBYTE;
       *comps = 2;
       return;
    default:
       assert(0);
-      *datatype = UBYTE;
+      *datatype = DTYPE_UBYTE;
       *comps = 0;
       break;
    }
@@ -1116,31 +1114,30 @@ make_1d_mipmap(struct gen_mipmap_state *ctx,
 
    for (dstLevel = baseLevel + 1; dstLevel <= lastLevel; dstLevel++) {
       const uint srcLevel = dstLevel - 1;
-      struct pipe_surface *srcSurf, *dstSurf;
+      struct pipe_transfer *srcTrans, *dstTrans;
       void *srcMap, *dstMap;
       
-      srcSurf = screen->get_tex_surface(screen, pt, face, srcLevel, zslice,
-                                        PIPE_BUFFER_USAGE_CPU_READ);
-
-      dstSurf = screen->get_tex_surface(screen, pt, face, dstLevel, zslice,
-                                        PIPE_BUFFER_USAGE_CPU_WRITE);
-
-      srcMap = ((ubyte *) pipe_surface_map(srcSurf,
-                                           PIPE_BUFFER_USAGE_CPU_READ)
-                + srcSurf->offset);
-      dstMap = ((ubyte *) pipe_surface_map(dstSurf,
-                                           PIPE_BUFFER_USAGE_CPU_WRITE)
-                + dstSurf->offset);
+      srcTrans = screen->get_tex_transfer(screen, pt, face, srcLevel, zslice,
+                                          PIPE_TRANSFER_READ, 0, 0,
+                                          pt->width[srcLevel],
+                                          pt->height[srcLevel]);
+      dstTrans = screen->get_tex_transfer(screen, pt, face, dstLevel, zslice,
+                                          PIPE_TRANSFER_WRITE, 0, 0,
+                                          pt->width[dstLevel],
+                                          pt->height[dstLevel]);
+
+      srcMap = (ubyte *) screen->transfer_map(screen, srcTrans);
+      dstMap = (ubyte *) screen->transfer_map(screen, dstTrans);
 
       reduce_1d(pt->format,
-                srcSurf->width, srcMap,
-                dstSurf->width, dstMap);
+                srcTrans->width, srcMap,
+                dstTrans->width, dstMap);
 
-      pipe_surface_unmap(srcSurf);
-      pipe_surface_unmap(dstSurf);
+      screen->transfer_unmap(screen, srcTrans);
+      screen->transfer_unmap(screen, dstTrans);
 
-      pipe_surface_reference(&srcSurf, NULL);
-      pipe_surface_reference(&dstSurf, NULL);
+      screen->tex_transfer_destroy(srcTrans);
+      screen->tex_transfer_destroy(dstTrans);
    }
 }
 
@@ -1160,32 +1157,32 @@ make_2d_mipmap(struct gen_mipmap_state *ctx,
 
    for (dstLevel = baseLevel + 1; dstLevel <= lastLevel; dstLevel++) {
       const uint srcLevel = dstLevel - 1;
-      struct pipe_surface *srcSurf, *dstSurf;
+      struct pipe_transfer *srcTrans, *dstTrans;
       ubyte *srcMap, *dstMap;
       
-      srcSurf = screen->get_tex_surface(screen, pt, face, srcLevel, zslice,
-                                        PIPE_BUFFER_USAGE_CPU_READ);
-      dstSurf = screen->get_tex_surface(screen, pt, face, dstLevel, zslice,
-                                        PIPE_BUFFER_USAGE_CPU_WRITE);
-
-      srcMap = ((ubyte *) pipe_surface_map(srcSurf,
-                                           PIPE_BUFFER_USAGE_CPU_READ)
-                + srcSurf->offset);
-      dstMap = ((ubyte *) pipe_surface_map(dstSurf,
-                                           PIPE_BUFFER_USAGE_CPU_WRITE)
-                + dstSurf->offset);
+      srcTrans = screen->get_tex_transfer(screen, pt, face, srcLevel, zslice,
+                                          PIPE_TRANSFER_READ, 0, 0,
+                                          pt->width[srcLevel],
+                                          pt->height[srcLevel]);
+      dstTrans = screen->get_tex_transfer(screen, pt, face, dstLevel, zslice,
+                                          PIPE_TRANSFER_WRITE, 0, 0,
+                                          pt->width[dstLevel],
+                                          pt->height[dstLevel]);
+
+      srcMap = (ubyte *) screen->transfer_map(screen, srcTrans);
+      dstMap = (ubyte *) screen->transfer_map(screen, dstTrans);
 
       reduce_2d(pt->format,
-                srcSurf->width, srcSurf->height,
-                srcSurf->stride, srcMap,
-                dstSurf->width, dstSurf->height,
-                dstSurf->stride, dstMap);
+                srcTrans->width, srcTrans->height,
+                srcTrans->stride, srcMap,
+                dstTrans->width, dstTrans->height,
+                dstTrans->stride, dstMap);
 
-      pipe_surface_unmap(srcSurf);
-      pipe_surface_unmap(dstSurf);
+      screen->transfer_unmap(screen, srcTrans);
+      screen->transfer_unmap(screen, dstTrans);
 
-      pipe_surface_reference(&srcSurf, NULL);
-      pipe_surface_reference(&dstSurf, NULL);
+      screen->tex_transfer_destroy(srcTrans);
+      screen->tex_transfer_destroy(dstTrans);
    }
 }
 
@@ -1195,6 +1192,7 @@ make_3d_mipmap(struct gen_mipmap_state *ctx,
                struct pipe_texture *pt,
                uint face, uint baseLevel, uint lastLevel)
 {
+#if 0
    struct pipe_context *pipe = ctx->pipe;
    struct pipe_screen *screen = pipe->screen;
    uint dstLevel, zslice = 0;
@@ -1204,37 +1202,36 @@ make_3d_mipmap(struct gen_mipmap_state *ctx,
 
    for (dstLevel = baseLevel + 1; dstLevel <= lastLevel; dstLevel++) {
       const uint srcLevel = dstLevel - 1;
-      struct pipe_surface *srcSurf, *dstSurf;
+      struct pipe_transfer *srcTrans, *dstTrans;
       ubyte *srcMap, *dstMap;
       
-      srcSurf = screen->get_tex_surface(screen, pt, face, srcLevel, zslice,
-                                        PIPE_BUFFER_USAGE_CPU_READ);
-      dstSurf = screen->get_tex_surface(screen, pt, face, dstLevel, zslice,
-                                        PIPE_BUFFER_USAGE_CPU_WRITE);
-
-      srcMap = ((ubyte *) pipe_surface_map(srcSurf,
-                                           PIPE_BUFFER_USAGE_CPU_READ)
-                + srcSurf->offset);
-      dstMap = ((ubyte *) pipe_surface_map(dstSurf,
-                                           PIPE_BUFFER_USAGE_CPU_WRITE)
-                + dstSurf->offset);
+      srcTrans = screen->get_tex_transfer(screen, pt, face, srcLevel, zslice,
+                                          PIPE_TRANSFER_READ, 0, 0,
+                                          pt->width[srcLevel],
+                                          pt->height[srcLevel]);
+      dstTrans = screen->get_tex_transfer(screen, pt, face, dstLevel, zslice,
+                                          PIPE_TRANSFER_WRITE, 0, 0,
+                                          pt->width[dstLevel],
+                                          pt->height[dstLevel]);
+
+      srcMap = (ubyte *) screen->transfer_map(screen, srcTrans);
+      dstMap = (ubyte *) screen->transfer_map(screen, dstTrans);
 
-#if 0
       reduce_3d(pt->format,
-                srcSurf->width, srcSurf->height,
-                srcSurf->stride, srcMap,
-                dstSurf->width, dstSurf->height,
-                dstSurf->stride, dstMap);
-#else
-      (void) reduce_3d;
-#endif
+                srcTrans->width, srcTrans->height,
+                srcTrans->stride, srcMap,
+                dstTrans->width, dstTrans->height,
+                dstTrans->stride, dstMap);
 
-      pipe_surface_unmap(srcSurf);
-      pipe_surface_unmap(dstSurf);
+      screen->transfer_unmap(screen, srcTrans);
+      screen->transfer_unmap(screen, dstTrans);
 
-      pipe_surface_reference(&srcSurf, NULL);
-      pipe_surface_reference(&dstSurf, NULL);
+      screen->tex_transfer_destroy(srcTrans);
+      screen->tex_transfer_destroy(dstTrans);
    }
+#else
+   (void) reduce_3d;
+#endif
 }
 
 
@@ -1294,8 +1291,7 @@ util_create_gen_mipmap(struct pipe_context *pipe,
    memset(&ctx->rasterizer, 0, sizeof(ctx->rasterizer));
    ctx->rasterizer.front_winding = PIPE_WINDING_CW;
    ctx->rasterizer.cull_mode = PIPE_WINDING_NONE;
-   ctx->rasterizer.bypass_clipping = 1;
-   /*ctx->rasterizer.bypass_vs = 1;*/
+   ctx->rasterizer.bypass_vs_clip_and_viewport = 1;
    ctx->rasterizer.gl_rasterization_rules = 1;
 
    /* sampler state */
@@ -1306,28 +1302,19 @@ util_create_gen_mipmap(struct pipe_context *pipe,
    ctx->sampler.min_mip_filter = PIPE_TEX_MIPFILTER_NEAREST;
    ctx->sampler.normalized_coords = 1;
 
-   /* viewport state (identity, verts are in wincoords) */
-   ctx->viewport.scale[0] = 1.0;
-   ctx->viewport.scale[1] = 1.0;
-   ctx->viewport.scale[2] = 1.0;
-   ctx->viewport.scale[3] = 1.0;
-   ctx->viewport.translate[0] = 0.0;
-   ctx->viewport.translate[1] = 0.0;
-   ctx->viewport.translate[2] = 0.0;
-   ctx->viewport.translate[3] = 0.0;
-
-   /* vertex shader */
+   /* vertex shader - still needed to specify mapping from fragment
+    * shader input semantics to vertex elements 
+    */
    {
       const uint semantic_names[] = { TGSI_SEMANTIC_POSITION,
                                       TGSI_SEMANTIC_GENERIC };
       const uint semantic_indexes[] = { 0, 0 };
       ctx->vs = util_make_vertex_passthrough_shader(pipe, 2, semantic_names,
-                                                    semantic_indexes,
-                                                    &ctx->vert_shader);
+                                                    semantic_indexes);
    }
 
    /* fragment shader */
-   ctx->fs = util_make_fragment_tex_shader(pipe, &ctx->frag_shader);
+   ctx->fs = util_make_fragment_tex_shader(pipe);
 
    /* vertex data that doesn't change */
    for (i = 0; i < 4; i++) {
@@ -1369,7 +1356,6 @@ get_next_slot(struct gen_mipmap_state *ctx)
 static unsigned
 set_vertex_data(struct gen_mipmap_state *ctx, float width, float height)
 {
-   void *buf;
    unsigned offset;
 
    ctx->vertices[0][0][0] = 0.0f; /*x*/
@@ -1394,12 +1380,8 @@ set_vertex_data(struct gen_mipmap_state *ctx, float width, float height)
 
    offset = get_next_slot( ctx );
 
-   buf = pipe_buffer_map(ctx->pipe->screen, ctx->vbuf,
-                         PIPE_BUFFER_USAGE_CPU_WRITE);
-
-   memcpy((char *)buf + offset, ctx->vertices, sizeof(ctx->vertices));
-
-   pipe_buffer_unmap(ctx->pipe->screen, ctx->vbuf);
+   pipe_buffer_write(ctx->pipe->screen, ctx->vbuf,
+                     offset, sizeof(ctx->vertices), ctx->vertices);
 
    return offset;
 }
@@ -1417,10 +1399,7 @@ util_destroy_gen_mipmap(struct gen_mipmap_state *ctx)
    pipe->delete_vs_state(pipe, ctx->vs);
    pipe->delete_fs_state(pipe, ctx->fs);
 
-   FREE((void*) ctx->vert_shader.tokens);
-   FREE((void*) ctx->frag_shader.tokens);
-
-   pipe_buffer_reference(pipe->screen, &ctx->vbuf, NULL);
+   pipe_buffer_reference(&ctx->vbuf, NULL);
 
    FREE(ctx);
 }
@@ -1432,7 +1411,7 @@ util_destroy_gen_mipmap(struct gen_mipmap_state *ctx)
  */
 void util_gen_mipmap_flush( struct gen_mipmap_state *ctx )
 {
-   pipe_buffer_reference(ctx->pipe->screen, &ctx->vbuf, NULL);
+   pipe_buffer_reference(&ctx->vbuf, NULL);
    ctx->vbuf_slot = 0;
 } 
 
@@ -1476,13 +1455,11 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
    cso_save_framebuffer(ctx->cso);
    cso_save_fragment_shader(ctx->cso);
    cso_save_vertex_shader(ctx->cso);
-   cso_save_viewport(ctx->cso);
 
    /* bind our state */
    cso_set_blend(ctx->cso, &ctx->blend);
    cso_set_depth_stencil_alpha(ctx->cso, &ctx->depthstencil);
    cso_set_rasterizer(ctx->cso, &ctx->rasterizer);
-   cso_set_viewport(ctx->cso, &ctx->viewport);
 
    cso_set_fragment_shader_handle(ctx->cso, ctx->fs);
    cso_set_vertex_shader_handle(ctx->cso, ctx->vs);
@@ -1528,7 +1505,7 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
 
       cso_set_sampler_textures(ctx->cso, 1, &pt);
 
-      /* quad coords in window coords (bypassing clipping, viewport mapping) */
+      /* quad coords in window coords (bypassing vs, clip and viewport) */
       offset = set_vertex_data(ctx,
                                (float) pt->width[dstLevel],
                                (float) pt->height[dstLevel]);
@@ -1555,5 +1532,4 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
    cso_restore_framebuffer(ctx->cso);
    cso_restore_fragment_shader(ctx->cso);
    cso_restore_vertex_shader(ctx->cso);
-   cso_restore_viewport(ctx->cso);
 }
diff --git a/src/gallium/auxiliary/util/u_handle_table.c b/src/gallium/auxiliary/util/u_handle_table.c
index 2d15932ce3..6da7353e25 100644
--- a/src/gallium/auxiliary/util/u_handle_table.c
+++ b/src/gallium/auxiliary/util/u_handle_table.c
@@ -34,7 +34,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "util/u_memory.h"
 #include "util/u_handle_table.h"
diff --git a/src/gallium/auxiliary/util/u_hash_table.c b/src/gallium/auxiliary/util/u_hash_table.c
index 0bc8de9632..2f83e318e4 100644
--- a/src/gallium/auxiliary/util/u_hash_table.c
+++ b/src/gallium/auxiliary/util/u_hash_table.c
@@ -39,7 +39,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "cso_cache/cso_hash.h"
 
diff --git a/src/gallium/auxiliary/util/u_keymap.c b/src/gallium/auxiliary/util/u_keymap.c
index 01b17ddb1b..3f70809efd 100644
--- a/src/gallium/auxiliary/util/u_keymap.c
+++ b/src/gallium/auxiliary/util/u_keymap.c
@@ -35,7 +35,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_error.h"
 
 #include "cso_cache/cso_hash.h"
diff --git a/src/gallium/auxiliary/util/u_linear.c b/src/gallium/auxiliary/util/u_linear.c
index e999cefe74..6be365e53b 100644
--- a/src/gallium/auxiliary/util/u_linear.c
+++ b/src/gallium/auxiliary/util/u_linear.c
@@ -1,5 +1,5 @@
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "u_linear.h"
 
 void
diff --git a/src/gallium/auxiliary/util/u_math.h b/src/gallium/auxiliary/util/u_math.h
index ab6f39ac31..1ecde7a912 100644
--- a/src/gallium/auxiliary/util/u_math.h
+++ b/src/gallium/auxiliary/util/u_math.h
@@ -40,7 +40,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 
 #ifdef __cplusplus
diff --git a/src/gallium/auxiliary/util/u_memory.h b/src/gallium/auxiliary/util/u_memory.h
index 1a6b596421..0b18d043ad 100644
--- a/src/gallium/auxiliary/util/u_memory.h
+++ b/src/gallium/auxiliary/util/u_memory.h
@@ -36,7 +36,7 @@
 
 
 #include "util/u_pointer.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 
 #ifdef __cplusplus
@@ -52,11 +52,11 @@ extern "C" {
 #endif
 
 
-#if defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY) && defined(DEBUG) 
+#if defined(PIPE_OS_WINDOWS) && defined(DEBUG) 
 
 /* memory debugging */
 
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #define MALLOC( _size ) \
    debug_malloc( __FILE__, __LINE__, __FUNCTION__, _size )
@@ -191,9 +191,11 @@ align_free(void *ptr)
 #if defined(HAVE_POSIX_MEMALIGN)
    FREE(ptr);
 #else
-   void **cubbyHole = (void **) ((char *) ptr - sizeof(void *));
-   void *realAddr = *cubbyHole;
-   FREE(realAddr);
+   if (ptr) {
+      void **cubbyHole = (void **) ((char *) ptr - sizeof(void *));
+      void *realAddr = *cubbyHole;
+      FREE(realAddr);
+   }
 #endif /* defined(HAVE_POSIX_MEMALIGN) */
 }
 
diff --git a/src/gallium/auxiliary/util/u_mm.c b/src/gallium/auxiliary/util/u_mm.c
index 45ce257b5e..151a480d34 100644
--- a/src/gallium/auxiliary/util/u_mm.c
+++ b/src/gallium/auxiliary/util/u_mm.c
@@ -24,7 +24,7 @@
 
 
 #include "pipe/p_compiler.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 
 #include "util/u_memory.h"
 #include "util/u_mm.h"
diff --git a/src/gallium/auxiliary/util/u_rect.c b/src/gallium/auxiliary/util/u_rect.c
index fe81a685be..74259d453b 100644
--- a/src/gallium/auxiliary/util/u_rect.c
+++ b/src/gallium/auxiliary/util/u_rect.c
@@ -169,46 +169,35 @@ util_surface_copy(struct pipe_context *pipe,
                   unsigned w, unsigned h)
 {
    struct pipe_screen *screen = pipe->screen;
-   struct pipe_surface *new_src = NULL, *new_dst = NULL;
+   struct pipe_transfer *src_trans, *dst_trans;
    void *dst_map;
    const void *src_map;
 
-   assert(dst->block.size == src->block.size);
-   assert(dst->block.width == src->block.width);
-   assert(dst->block.height == src->block.height);
-
-   if ((src->usage & PIPE_BUFFER_USAGE_CPU_READ) == 0) {
-      /* Need to create new src surface which is CPU readable */
-      assert(src->texture);
-      if (!src->texture)
-         return;
-      new_src = screen->get_tex_surface(screen,
+   assert(src->texture && dst->texture);
+   if (!src->texture || !dst->texture)
+      return;
+   src_trans = screen->get_tex_transfer(screen,
                                         src->texture,
                                         src->face,
                                         src->level,
                                         src->zslice,
-                                        PIPE_BUFFER_USAGE_CPU_READ);
-      src = new_src;
-   }
+                                        PIPE_TRANSFER_READ,
+                                        src_x, src_y, w, h);
 
-   if ((dst->usage & PIPE_BUFFER_USAGE_CPU_WRITE) == 0) {
-      /* Need to create new dst surface which is CPU writable */
-      assert(dst->texture);
-      if (!dst->texture)
-         return;
-      new_dst = screen->get_tex_surface(screen,
+   dst_trans = screen->get_tex_transfer(screen,
                                         dst->texture,
                                         dst->face,
                                         dst->level,
                                         dst->zslice,
-                                        PIPE_BUFFER_USAGE_CPU_WRITE);
-      dst = new_dst;
-   }
+                                        PIPE_TRANSFER_WRITE,
+                                        dst_x, dst_y, w, h);
 
-   src_map = pipe->screen->surface_map(screen,
-                                       src, PIPE_BUFFER_USAGE_CPU_READ);
-   dst_map = pipe->screen->surface_map(screen,
-                                       dst, PIPE_BUFFER_USAGE_CPU_WRITE);
+   assert(dst_trans->block.size == src_trans->block.size);
+   assert(dst_trans->block.width == src_trans->block.width);
+   assert(dst_trans->block.height == src_trans->block.height);
+
+   src_map = pipe->screen->transfer_map(screen, src_trans);
+   dst_map = pipe->screen->transfer_map(screen, dst_trans);
 
    assert(src_map);
    assert(dst_map);
@@ -216,36 +205,25 @@ util_surface_copy(struct pipe_context *pipe,
    if (src_map && dst_map) {
       /* If do_flip, invert src_y position and pass negative src stride */
       pipe_copy_rect(dst_map,
-                     &dst->block,
-                     dst->stride,
-                     dst_x, dst_y,
+                     &dst_trans->block,
+                     dst_trans->stride,
+                     0, 0,
                      w, h,
                      src_map,
-                     do_flip ? -(int) src->stride : src->stride,
-                     src_x,
-                     do_flip ? src_y + h - 1 : src_y);
+                     do_flip ? -(int) src_trans->stride : src_trans->stride,
+                     0,
+                     do_flip ? h - 1 : 0);
    }
 
-   pipe->screen->surface_unmap(pipe->screen, src);
-   pipe->screen->surface_unmap(pipe->screen, dst);
+   pipe->screen->transfer_unmap(pipe->screen, src_trans);
+   pipe->screen->transfer_unmap(pipe->screen, dst_trans);
 
-   if (new_src)
-      screen->tex_surface_release(screen, &new_src);
-   if (new_dst)
-      screen->tex_surface_release(screen, &new_dst);
+   screen->tex_transfer_destroy(src_trans);
+   screen->tex_transfer_destroy(dst_trans);
 }
 
 
 
-static void *
-get_pointer(struct pipe_surface *dst, void *dst_map, unsigned x, unsigned y)
-{
-   return (char *)dst_map
-      + y / dst->block.height * dst->stride
-      + x / dst->block.width * dst->block.size;
-}
-
-
 #define UBYTE_TO_USHORT(B) ((B) | ((B) << 8))
 
 
@@ -260,42 +238,38 @@ util_surface_fill(struct pipe_context *pipe,
                   unsigned width, unsigned height, unsigned value)
 {
    struct pipe_screen *screen = pipe->screen;
-   struct pipe_surface *new_dst = NULL;
+   struct pipe_transfer *dst_trans;
    void *dst_map;
 
-   if ((dst->usage & PIPE_BUFFER_USAGE_CPU_WRITE) == 0) {
-      /* Need to create new dst surface which is CPU writable */
-      assert(dst->texture);
-      if (!dst->texture)
-         return;
-      new_dst = screen->get_tex_surface(screen,
+   assert(dst->texture);
+   if (!dst->texture)
+      return;
+   dst_trans = screen->get_tex_transfer(screen,
                                         dst->texture,
                                         dst->face,
                                         dst->level,
                                         dst->zslice,
-                                        PIPE_BUFFER_USAGE_CPU_WRITE);
-      dst = new_dst;
-   }
+                                        PIPE_TRANSFER_WRITE,
+                                        dstx, dsty, width, height);
 
-   dst_map = pipe->screen->surface_map(screen,
-                                       dst, PIPE_BUFFER_USAGE_CPU_WRITE);
+   dst_map = pipe->screen->transfer_map(screen, dst_trans);
 
    assert(dst_map);
 
    if (dst_map) {
-      assert(dst->stride > 0);
+      assert(dst_trans->stride > 0);
 
-      switch (dst->block.size) {
+      switch (dst_trans->block.size) {
       case 1:
       case 2:
       case 4:
-         pipe_fill_rect(dst_map, &dst->block, dst->stride,
-                        dstx, dsty, width, height, value);
+         pipe_fill_rect(dst_map, &dst_trans->block, dst_trans->stride,
+                        0, 0, width, height, value);
          break;
       case 8:
          {
             /* expand the 4-byte clear value to an 8-byte value */
-            ushort *row = (ushort *) get_pointer(dst, dst_map, dstx, dsty);
+            ushort *row = (ushort *) dst_map;
             ushort val0 = UBYTE_TO_USHORT((value >>  0) & 0xff);
             ushort val1 = UBYTE_TO_USHORT((value >>  8) & 0xff);
             ushort val2 = UBYTE_TO_USHORT((value >> 16) & 0xff);
@@ -312,7 +286,7 @@ util_surface_fill(struct pipe_context *pipe,
                   row[j*4+2] = val2;
                   row[j*4+3] = val3;
                }
-               row += dst->stride/2;
+               row += dst_trans->stride/2;
             }
          }
          break;
@@ -322,8 +296,6 @@ util_surface_fill(struct pipe_context *pipe,
       }
    }
 
-   pipe->screen->surface_unmap(pipe->screen, dst);
-
-   if (new_dst)
-      screen->tex_surface_release(screen, &new_dst);
+   pipe->screen->transfer_unmap(pipe->screen, dst_trans);
+   screen->tex_transfer_destroy(dst_trans);
 }
diff --git a/src/gallium/auxiliary/util/u_simple_screen.c b/src/gallium/auxiliary/util/u_simple_screen.c
index 089bbbc48a..8114b53cd0 100644
--- a/src/gallium/auxiliary/util/u_simple_screen.c
+++ b/src/gallium/auxiliary/util/u_simple_screen.c
@@ -28,6 +28,7 @@
 #include "u_simple_screen.h"
 
 #include "pipe/p_screen.h"
+#include "pipe/p_state.h"
 #include "pipe/internal/p_winsys_screen.h"
 
 
@@ -37,8 +38,12 @@ pass_buffer_create(struct pipe_screen *screen,
                    unsigned usage,
                    unsigned size)
 {
-   return screen->winsys->buffer_create(screen->winsys,
-                                        alignment, usage, size);
+   struct pipe_buffer *buffer =
+      screen->winsys->buffer_create(screen->winsys, alignment, usage, size);
+
+   buffer->screen = screen;
+
+   return buffer;
 }
 
 static struct pipe_buffer *
@@ -46,8 +51,13 @@ pass_user_buffer_create(struct pipe_screen *screen,
                         void *ptr,
                         unsigned bytes)
 {
-   return screen->winsys->user_buffer_create(screen->winsys,
+   struct pipe_buffer *buffer =
+      screen->winsys->user_buffer_create(screen->winsys,
                                              ptr, bytes);
+
+   buffer->screen = screen;
+
+   return buffer;
 }
 
 static struct pipe_buffer *
@@ -57,9 +67,14 @@ pass_surface_buffer_create(struct pipe_screen *screen,
                            unsigned usage,
                            unsigned *stride)
 {
-   return screen->winsys->surface_buffer_create(screen->winsys,
+   struct pipe_buffer *buffer =
+      screen->winsys->surface_buffer_create(screen->winsys,
                                                 width, height,
                                                 format, usage, stride);
+
+   buffer->screen = screen;
+
+   return buffer;
 }
 
 static void *
@@ -79,10 +94,9 @@ pass_buffer_unmap(struct pipe_screen *screen,
 }
 
 static void
-pass_buffer_destroy(struct pipe_screen *screen,
-                    struct pipe_buffer *buf)
+pass_buffer_destroy(struct pipe_buffer *buf)
 {
-   screen->winsys->buffer_destroy(screen->winsys, buf);
+   buf->screen->winsys->buffer_destroy(buf);
 }
 
 
diff --git a/src/gallium/auxiliary/util/u_simple_shaders.c b/src/gallium/auxiliary/util/u_simple_shaders.c
index 706155e99a..e519c354d2 100644
--- a/src/gallium/auxiliary/util/u_simple_shaders.c
+++ b/src/gallium/auxiliary/util/u_simple_shaders.c
@@ -34,9 +34,8 @@
 
 
 #include "pipe/p_context.h"
-#include "pipe/p_debug.h"
+#include "util/u_debug.h"
 #include "pipe/p_defines.h"
-#include "pipe/p_inlines.h"
 #include "pipe/p_screen.h"
 #include "pipe/p_shader_tokens.h"
 
@@ -56,12 +55,11 @@ void *
 util_make_vertex_passthrough_shader(struct pipe_context *pipe,
                                     uint num_attribs,
                                     const uint *semantic_names,
-                                    const uint *semantic_indexes,
-                                    struct pipe_shader_state *shader)
+                                    const uint *semantic_indexes)
                                     
 {
-   uint maxTokens = 100;
-   struct tgsi_token *tokens;
+   struct pipe_shader_state shader;
+   struct tgsi_token tokens[100];
    struct tgsi_header *header;
    struct tgsi_processor *processor;
    struct tgsi_full_declaration decl;
@@ -69,8 +67,6 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
    const uint procType = TGSI_PROCESSOR_VERTEX;
    uint ti, i;
 
-   tokens = (struct tgsi_token *) MALLOC(maxTokens * sizeof(tokens[0]));
-
    /* shader header
     */
    *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
@@ -97,7 +93,7 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
       ti += tgsi_build_full_declaration(&decl,
                                         &tokens[ti],
                                         header,
-                                        maxTokens - ti);
+                                        Elements(tokens) - ti);
    }
 
    /* declare outputs */
@@ -112,7 +108,7 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
       ti += tgsi_build_full_declaration(&decl,
                                         &tokens[ti],
                                         header,
-                                        maxTokens - ti);
+                                        Elements(tokens) - ti);
    }
 
    /* emit MOV instructions */
@@ -129,7 +125,7 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
       ti += tgsi_build_full_instruction(&inst,
                                         &tokens[ti],
                                         header,
-                                        maxTokens - ti );
+                                        Elements(tokens) - ti );
    }
 
    /* END instruction */
@@ -140,16 +136,15 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_instruction(&inst,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti );
+                                     Elements(tokens) - ti );
 
 #if 0 /*debug*/
    tgsi_dump(tokens, 0);
 #endif
 
-   shader->tokens = tokens;
-   /*shader->num_tokens = ti;*/
+   shader.tokens = tokens;
 
-   return pipe->create_vs_state(pipe, shader);
+   return pipe->create_vs_state(pipe, &shader);
 }
 
 
@@ -161,11 +156,10 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
  *  END;
  */
 void *
-util_make_fragment_tex_shader(struct pipe_context *pipe,
-                              struct pipe_shader_state *shader)
+util_make_fragment_tex_shader(struct pipe_context *pipe)
 {
-   uint maxTokens = 100;
-   struct tgsi_token *tokens;
+   struct pipe_shader_state shader;
+   struct tgsi_token tokens[100];
    struct tgsi_header *header;
    struct tgsi_processor *processor;
    struct tgsi_full_declaration decl;
@@ -173,8 +167,6 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
    const uint procType = TGSI_PROCESSOR_FRAGMENT;
    uint ti;
 
-   tokens = (struct tgsi_token *) MALLOC(maxTokens * sizeof(tokens[0]));
-
    /* shader header
     */
    *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
@@ -200,7 +192,7 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_declaration(&decl,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti);
+                                     Elements(tokens) - ti);
 
    /* declare color[0] output */
    decl = tgsi_default_full_declaration();
@@ -213,7 +205,7 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_declaration(&decl,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti);
+                                     Elements(tokens) - ti);
 
    /* declare sampler */
    decl = tgsi_default_full_declaration();
@@ -223,7 +215,7 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_declaration(&decl,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti);
+                                     Elements(tokens) - ti);
 
    /* TEX instruction */
    inst = tgsi_default_full_instruction();
@@ -240,7 +232,7 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_instruction(&inst,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti );
+                                     Elements(tokens) - ti );
 
    /* END instruction */
    inst = tgsi_default_full_instruction();
@@ -250,16 +242,15 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_instruction(&inst,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti );
+                                     Elements(tokens) - ti );
 
 #if 0 /*debug*/
    tgsi_dump(tokens, 0);
 #endif
 
-   shader->tokens = tokens;
-   /*shader->num_tokens = ti;*/
+   shader.tokens = tokens;
 
-   return pipe->create_fs_state(pipe, shader);
+   return pipe->create_fs_state(pipe, &shader);
 }
 
 
@@ -270,11 +261,10 @@ util_make_fragment_tex_shader(struct pipe_context *pipe,
  * Make simple fragment color pass-through shader.
  */
 void *
-util_make_fragment_passthrough_shader(struct pipe_context *pipe,
-                                      struct pipe_shader_state *shader)
+util_make_fragment_passthrough_shader(struct pipe_context *pipe)
 {
-   uint maxTokens = 40;
-   struct tgsi_token *tokens;
+   struct pipe_shader_state shader;
+   struct tgsi_token tokens[40];
    struct tgsi_header *header;
    struct tgsi_processor *processor;
    struct tgsi_full_declaration decl;
@@ -282,8 +272,6 @@ util_make_fragment_passthrough_shader(struct pipe_context *pipe,
    const uint procType = TGSI_PROCESSOR_FRAGMENT;
    uint ti;
 
-   tokens = (struct tgsi_token *) MALLOC(maxTokens * sizeof(tokens[0]));
-
    /* shader header
     */
    *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
@@ -307,7 +295,7 @@ util_make_fragment_passthrough_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_declaration(&decl,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti);
+                                     Elements(tokens) - ti);
 
    /* declare output */
    decl = tgsi_default_full_declaration();
@@ -320,7 +308,7 @@ util_make_fragment_passthrough_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_declaration(&decl,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti);
+                                     Elements(tokens) - ti);
 
 
    /* MOVE out[0], in[0]; */
@@ -335,7 +323,7 @@ util_make_fragment_passthrough_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_instruction(&inst,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti );
+                                     Elements(tokens) - ti );
 
    /* END instruction */
    inst = tgsi_default_full_instruction();
@@ -345,17 +333,17 @@ util_make_fragment_passthrough_shader(struct pipe_context *pipe,
    ti += tgsi_build_full_instruction(&inst,
                                      &tokens[ti],
                                      header,
-                                     maxTokens - ti );
+                                     Elements(tokens) - ti );
 
-   assert(ti < maxTokens);
+   assert(ti < Elements(tokens));
 
 #if 0 /*debug*/
    tgsi_dump(tokens, 0);
 #endif
 
-   shader->tokens = tokens;
-   /*shader->num_tokens = ti;*/
+   shader.tokens = tokens;
 
-   return pipe->create_fs_state(pipe, shader);
+   return pipe->create_fs_state(pipe, &shader);
 }
 
+
diff --git a/src/gallium/auxiliary/util/u_simple_shaders.h b/src/gallium/auxiliary/util/u_simple_shaders.h
index 8ca4977d71..6f8d96af9b 100644
--- a/src/gallium/auxiliary/util/u_simple_shaders.h
+++ b/src/gallium/auxiliary/util/u_simple_shaders.h
@@ -46,18 +46,15 @@ extern void *
 util_make_vertex_passthrough_shader(struct pipe_context *pipe,
                                     uint num_attribs,
                                     const uint *semantic_names,
-                                    const uint *semantic_indexes,
-                                    struct pipe_shader_state *shader);
+                                    const uint *semantic_indexes);
 
 
 extern void *
-util_make_fragment_tex_shader(struct pipe_context *pipe,
-                              struct pipe_shader_state *shader);
+util_make_fragment_tex_shader(struct pipe_context *pipe);
 
 
 extern void *
-util_make_fragment_passthrough_shader(struct pipe_context *pipe,
-                                      struct pipe_shader_state *shader);
+util_make_fragment_passthrough_shader(struct pipe_context *pipe);
 
 
 #ifdef __cplusplus
diff --git a/src/gallium/auxiliary/util/u_stream_stdc.c b/src/gallium/auxiliary/util/u_stream_stdc.c
index ca80bef0f3..0ead45a749 100644
--- a/src/gallium/auxiliary/util/u_stream_stdc.c
+++ b/src/gallium/auxiliary/util/u_stream_stdc.c
@@ -32,7 +32,7 @@
 
 #include "pipe/p_config.h"
 
-#if defined(PIPE_OS_LINUX) || defined(PIPE_SUBSYSTEM_WINDOWS_USER)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_SUBSYSTEM_WINDOWS_USER)
 
 #include <stdio.h>
 
diff --git a/src/gallium/auxiliary/util/u_tile.c b/src/gallium/auxiliary/util/u_tile.c
index 32f6b072a0..d31ca9c029 100644
--- a/src/gallium/auxiliary/util/u_tile.c
+++ b/src/gallium/auxiliary/util/u_tile.c
@@ -28,7 +28,6 @@
 /**
  * RGBA/float tile get/put functions.
  * Usable both by drivers and state trackers.
- * Surfaces should already be in a mapped state.
  */
 
 
@@ -42,58 +41,58 @@
 
 
 /**
- * Move raw block of pixels from surface to user memory.
- * This should be usable by any hw driver that has mappable surfaces.
+ * Move raw block of pixels from transfer object to user memory.
  */
 void
-pipe_get_tile_raw(struct pipe_surface *ps,
+pipe_get_tile_raw(struct pipe_transfer *pt,
                   uint x, uint y, uint w, uint h,
                   void *dst, int dst_stride)
 {
+   struct pipe_screen *screen = pt->texture->screen;
    const void *src;
 
    if (dst_stride == 0)
-      dst_stride = pf_get_nblocksx(&ps->block, w) * ps->block.size;
+      dst_stride = pf_get_nblocksx(&pt->block, w) * pt->block.size;
 
-   if (pipe_clip_tile(x, y, &w, &h, ps))
+   if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   src = pipe_surface_map(ps, PIPE_BUFFER_USAGE_CPU_READ);
+   src = screen->transfer_map(screen, pt);
    assert(src);
    if(!src)
       return;
 
-   pipe_copy_rect(dst, &ps->block, dst_stride, 0, 0, w, h, src, ps->stride, x, y);
+   pipe_copy_rect(dst, &pt->block, dst_stride, 0, 0, w, h, src, pt->stride, x, y);
 
-   pipe_surface_unmap(ps);
+   screen->transfer_unmap(screen, pt);
 }
 
 
 /**
- * Move raw block of pixels from user memory to surface.
- * This should be usable by any hw driver that has mappable surfaces.
+ * Move raw block of pixels from user memory to transfer object.
  */
 void
-pipe_put_tile_raw(struct pipe_surface *ps,
+pipe_put_tile_raw(struct pipe_transfer *pt,
                   uint x, uint y, uint w, uint h,
                   const void *src, int src_stride)
 {
+   struct pipe_screen *screen = pt->texture->screen;
    void *dst;
 
    if (src_stride == 0)
-      src_stride = pf_get_nblocksx(&ps->block, w) * ps->block.size;
+      src_stride = pf_get_nblocksx(&pt->block, w) * pt->block.size;
 
-   if (pipe_clip_tile(x, y, &w, &h, ps))
+   if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   dst = pipe_surface_map(ps, PIPE_BUFFER_USAGE_CPU_WRITE);
+   dst = screen->transfer_map(screen, pt);
    assert(dst);
    if(!dst)
       return;
 
-   pipe_copy_rect(dst, &ps->block, ps->stride, x, y, w, h, src, src_stride, 0, 0);
+   pipe_copy_rect(dst, &pt->block, pt->stride, x, y, w, h, src, src_stride, 0, 0);
 
-   pipe_surface_unmap(ps);
+   screen->transfer_unmap(screen, pt);
 }
 
 
@@ -883,6 +882,27 @@ ycbcr_get_tile_rgba(const ushort *src,
 }
 
 
+static void
+fake_get_tile_rgba(const ushort *src,
+                   unsigned w, unsigned h,
+                   float *p,
+                   unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (i ^ j) & 1 ? 1.0f : 0.0f;
+      }
+      p += dst_stride;
+   }
+}
+
+
 void
 pipe_tile_raw_to_rgba(enum pipe_format format,
                       void *src,
@@ -949,55 +969,56 @@ pipe_tile_raw_to_rgba(enum pipe_format format,
       ycbcr_get_tile_rgba((ushort *) src, w, h, dst, dst_stride, TRUE);
       break;
    default:
-      assert(0);
+      debug_printf("%s: unsupported format %s\n", __FUNCTION__, pf_name(format));
+      fake_get_tile_rgba(src, w, h, dst, dst_stride);
    }
 }
 
 
 void
-pipe_get_tile_rgba(struct pipe_surface *ps,
+pipe_get_tile_rgba(struct pipe_transfer *pt,
                    uint x, uint y, uint w, uint h,
                    float *p)
 {
    unsigned dst_stride = w * 4;
    void *packed;
 
-   if (pipe_clip_tile(x, y, &w, &h, ps))
+   if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   packed = MALLOC(pf_get_nblocks(&ps->block, w, h) * ps->block.size);
+   packed = MALLOC(pf_get_nblocks(&pt->block, w, h) * pt->block.size);
 
    if (!packed)
       return;
 
-   if(ps->format == PIPE_FORMAT_YCBCR || ps->format == PIPE_FORMAT_YCBCR_REV)
+   if(pt->format == PIPE_FORMAT_YCBCR || pt->format == PIPE_FORMAT_YCBCR_REV)
       assert((x & 1) == 0);
 
-   pipe_get_tile_raw(ps, x, y, w, h, packed, 0);
+   pipe_get_tile_raw(pt, x, y, w, h, packed, 0);
 
-   pipe_tile_raw_to_rgba(ps->format, packed, w, h, p, dst_stride);
+   pipe_tile_raw_to_rgba(pt->format, packed, w, h, p, dst_stride);
 
    FREE(packed);
 }
 
 
 void
-pipe_put_tile_rgba(struct pipe_surface *ps,
+pipe_put_tile_rgba(struct pipe_transfer *pt,
                    uint x, uint y, uint w, uint h,
                    const float *p)
 {
    unsigned src_stride = w * 4;
    void *packed;
 
-   if (pipe_clip_tile(x, y, &w, &h, ps))
+   if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   packed = MALLOC(pf_get_nblocks(&ps->block, w, h) * ps->block.size);
+   packed = MALLOC(pf_get_nblocks(&pt->block, w, h) * pt->block.size);
 
    if (!packed)
       return;
 
-   switch (ps->format) {
+   switch (pt->format) {
    case PIPE_FORMAT_A8R8G8B8_UNORM:
       a8r8g8b8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);
       break;
@@ -1051,10 +1072,10 @@ pipe_put_tile_rgba(struct pipe_surface *ps,
       /*z24s8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
       break;
    default:
-      assert(0);
+      debug_printf("%s: unsupported format %s\n", __FUNCTION__, pf_name(pt->format));
    }
 
-   pipe_put_tile_raw(ps, x, y, w, h, packed, 0);
+   pipe_put_tile_raw(pt, x, y, w, h, packed, 0);
 
    FREE(packed);
 }
@@ -1064,62 +1085,63 @@ pipe_put_tile_rgba(struct pipe_surface *ps,
  * Get a block of Z values, converted to 32-bit range.
  */
 void
-pipe_get_tile_z(struct pipe_surface *ps,
+pipe_get_tile_z(struct pipe_transfer *pt,
                 uint x, uint y, uint w, uint h,
                 uint *z)
 {
+   struct pipe_screen *screen = pt->texture->screen;
    const uint dstStride = w;
    ubyte *map;
    uint *pDest = z;
    uint i, j;
 
-   if (pipe_clip_tile(x, y, &w, &h, ps))
+   if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   map = (ubyte *)pipe_surface_map(ps, PIPE_BUFFER_USAGE_CPU_READ);
+   map = (ubyte *)screen->transfer_map(screen, pt);
    if (!map) {
       assert(0);
       return;
    }
 
-   switch (ps->format) {
+   switch (pt->format) {
    case PIPE_FORMAT_Z32_UNORM:
       {
-         const uint *pSrc
-            = (const uint *)(map  + y * ps->stride + x*4);
+         const uint *ptrc
+            = (const uint *)(map  + y * pt->stride + x*4);
          for (i = 0; i < h; i++) {
-            memcpy(pDest, pSrc, 4 * w);
+            memcpy(pDest, ptrc, 4 * w);
             pDest += dstStride;
-            pSrc += ps->stride/4;
+            ptrc += pt->stride/4;
          }
       }
       break;
    case PIPE_FORMAT_S8Z24_UNORM:
    case PIPE_FORMAT_X8Z24_UNORM:
       {
-         const uint *pSrc
-            = (const uint *)(map + y * ps->stride + x*4);
+         const uint *ptrc
+            = (const uint *)(map + y * pt->stride + x*4);
          for (i = 0; i < h; i++) {
             for (j = 0; j < w; j++) {
                /* convert 24-bit Z to 32-bit Z */
-               pDest[j] = (pSrc[j] << 8) | (pSrc[j] & 0xff);
+               pDest[j] = (ptrc[j] << 8) | (ptrc[j] & 0xff);
             }
             pDest += dstStride;
-            pSrc += ps->stride/4;
+            ptrc += pt->stride/4;
          }
       }
       break;
    case PIPE_FORMAT_Z16_UNORM:
       {
-         const ushort *pSrc
-            = (const ushort *)(map + y * ps->stride + x*2);
+         const ushort *ptrc
+            = (const ushort *)(map + y * pt->stride + x*2);
          for (i = 0; i < h; i++) {
             for (j = 0; j < w; j++) {
                /* convert 16-bit Z to 32-bit Z */
-               pDest[j] = (pSrc[j] << 16) | pSrc[j];
+               pDest[j] = (ptrc[j] << 16) | ptrc[j];
             }
             pDest += dstStride;
-            pSrc += ps->stride/2;
+            ptrc += pt->stride/2;
          }
       }
       break;
@@ -1127,64 +1149,65 @@ pipe_get_tile_z(struct pipe_surface *ps,
       assert(0);
    }
 
-   pipe_surface_unmap(ps);
+   screen->transfer_unmap(screen, pt);
 }
 
 
 void
-pipe_put_tile_z(struct pipe_surface *ps,
+pipe_put_tile_z(struct pipe_transfer *pt,
                 uint x, uint y, uint w, uint h,
                 const uint *zSrc)
 {
+   struct pipe_screen *screen = pt->texture->screen;
    const uint srcStride = w;
-   const uint *pSrc = zSrc;
+   const uint *ptrc = zSrc;
    ubyte *map;
    uint i, j;
 
-   if (pipe_clip_tile(x, y, &w, &h, ps))
+   if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   map = (ubyte *)pipe_surface_map(ps, PIPE_BUFFER_USAGE_CPU_WRITE);
+   map = (ubyte *)screen->transfer_map(screen, pt);
    if (!map) {
       assert(0);
       return;
    }
 
-   switch (ps->format) {
+   switch (pt->format) {
    case PIPE_FORMAT_Z32_UNORM:
       {
-         uint *pDest = (uint *) (map + y * ps->stride + x*4);
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
          for (i = 0; i < h; i++) {
-            memcpy(pDest, pSrc, 4 * w);
-            pDest += ps->stride/4;
-            pSrc += srcStride;
+            memcpy(pDest, ptrc, 4 * w);
+            pDest += pt->stride/4;
+            ptrc += srcStride;
          }
       }
       break;
    case PIPE_FORMAT_S8Z24_UNORM:
    case PIPE_FORMAT_X8Z24_UNORM:
       {
-         uint *pDest = (uint *) (map + y * ps->stride + x*4);
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
          for (i = 0; i < h; i++) {
             for (j = 0; j < w; j++) {
                /* convert 32-bit Z to 24-bit Z (0 stencil) */
-               pDest[j] = pSrc[j] >> 8;
+               pDest[j] = ptrc[j] >> 8;
             }
-            pDest += ps->stride/4;
-            pSrc += srcStride;
+            pDest += pt->stride/4;
+            ptrc += srcStride;
          }
       }
       break;
    case PIPE_FORMAT_Z16_UNORM:
       {
-         ushort *pDest = (ushort *) (map + y * ps->stride + x*2);
+         ushort *pDest = (ushort *) (map + y * pt->stride + x*2);
          for (i = 0; i < h; i++) {
             for (j = 0; j < w; j++) {
                /* convert 32-bit Z to 16-bit Z */
-               pDest[j] = pSrc[j] >> 16;
+               pDest[j] = ptrc[j] >> 16;
             }
-            pDest += ps->stride/2;
-            pSrc += srcStride;
+            pDest += pt->stride/2;
+            ptrc += srcStride;
          }
       }
       break;
@@ -1192,7 +1215,7 @@ pipe_put_tile_z(struct pipe_surface *ps,
       assert(0);
    }
 
-   pipe_surface_unmap(ps);
+   screen->transfer_unmap(screen, pt);
 }
 
 
diff --git a/src/gallium/auxiliary/util/u_tile.h b/src/gallium/auxiliary/util/u_tile.h
index a8ac805308..1453af38b8 100644
--- a/src/gallium/auxiliary/util/u_tile.h
+++ b/src/gallium/auxiliary/util/u_tile.h
@@ -30,24 +30,24 @@
 
 #include "pipe/p_compiler.h"
 
-struct pipe_surface;
+struct pipe_transfer;
 
 
 /**
- * Clip tile against surface dims.
+ * Clip tile against transfer dims.
  * \return TRUE if tile is totally clipped, FALSE otherwise
  */
 static INLINE boolean
-pipe_clip_tile(uint x, uint y, uint *w, uint *h, const struct pipe_surface *ps)
+pipe_clip_tile(uint x, uint y, uint *w, uint *h, const struct pipe_transfer *pt)
 {
-   if (x >= ps->width)
+   if (x >= pt->width)
       return TRUE;
-   if (y >= ps->height)
+   if (y >= pt->height)
       return TRUE;
-   if (x + *w > ps->width)
-      *w = ps->width - x;
-   if (y + *h > ps->height)
-      *h = ps->height - y;
+   if (x + *w > pt->width)
+      *w = pt->width - x;
+   if (y + *h > pt->height)
+      *h = pt->height - y;
    return FALSE;
 }
 
@@ -56,34 +56,34 @@ extern "C" {
 #endif
 
 void
-pipe_get_tile_raw(struct pipe_surface *ps,
+pipe_get_tile_raw(struct pipe_transfer *pt,
                   uint x, uint y, uint w, uint h,
                   void *p, int dst_stride);
 
 void
-pipe_put_tile_raw(struct pipe_surface *ps,
+pipe_put_tile_raw(struct pipe_transfer *pt,
                   uint x, uint y, uint w, uint h,
                   const void *p, int src_stride);
 
 
 void
-pipe_get_tile_rgba(struct pipe_surface *ps,
+pipe_get_tile_rgba(struct pipe_transfer *pt,
                    uint x, uint y, uint w, uint h,
                    float *p);
 
 void
-pipe_put_tile_rgba(struct pipe_surface *ps,
+pipe_put_tile_rgba(struct pipe_transfer *pt,
                    uint x, uint y, uint w, uint h,
                    const float *p);
 
 
 void
-pipe_get_tile_z(struct pipe_surface *ps,
+pipe_get_tile_z(struct pipe_transfer *pt,
                 uint x, uint y, uint w, uint h,
                 uint *z);
 
 void
-pipe_put_tile_z(struct pipe_surface *ps,
+pipe_put_tile_z(struct pipe_transfer *pt,
                 uint x, uint y, uint w, uint h,
                 const uint *z);
 
diff --git a/src/gallium/auxiliary/util/u_time.c b/src/gallium/auxiliary/util/u_time.c
index dde2c74fa8..357d9360c9 100644
--- a/src/gallium/auxiliary/util/u_time.c
+++ b/src/gallium/auxiliary/util/u_time.c
@@ -35,7 +35,7 @@
 
 #include "pipe/p_config.h"
 
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
 #include <sys/time.h>
 #elif defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY)
 #include <windows.h>
@@ -77,7 +77,7 @@ util_time_get_frequency(void)
 void 
 util_time_get(struct util_time *t)
 {
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
    gettimeofday(&t->tv, NULL);
 #elif defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY)
    LONGLONG temp;
@@ -102,7 +102,7 @@ util_time_add(const struct util_time *t1,
               int64_t usecs,
               struct util_time *t2)
 {
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
    t2->tv.tv_sec = t1->tv.tv_sec + usecs / 1000000;
    t2->tv.tv_usec = t1->tv.tv_usec + usecs % 1000000;
 #elif defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY) || defined(PIPE_SUBSYSTEM_WINDOWS_USER) || defined(PIPE_SUBSYSTEM_WINDOWS_CE)
@@ -124,7 +124,7 @@ int64_t
 util_time_diff(const struct util_time *t1, 
                const struct util_time *t2)
 {
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
    return (t2->tv.tv_usec - t1->tv.tv_usec) + 
           (t2->tv.tv_sec - t1->tv.tv_sec)*1000000;
 #elif defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY) || defined(PIPE_SUBSYSTEM_WINDOWS_USER) || defined(PIPE_SUBSYSTEM_WINDOWS_CE)
@@ -144,7 +144,7 @@ util_time_micros( void )
    
    util_time_get(&t1);
    
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
    return t1.tv.tv_usec + t1.tv.tv_sec*1000000LL;
 #elif defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY) || defined(PIPE_SUBSYSTEM_WINDOWS_USER) || defined(PIPE_SUBSYSTEM_WINDOWS_CE)
    util_time_get_frequency();
@@ -166,7 +166,7 @@ static INLINE int
 util_time_compare(const struct util_time *t1, 
                   const struct util_time *t2)
 {
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
    if (t1->tv.tv_sec < t2->tv.tv_sec)
       return -1;
    else if(t1->tv.tv_sec > t2->tv.tv_sec)
diff --git a/src/gallium/auxiliary/util/u_time.h b/src/gallium/auxiliary/util/u_time.h
index 35d97d16c7..4346ce1fa4 100644
--- a/src/gallium/auxiliary/util/u_time.h
+++ b/src/gallium/auxiliary/util/u_time.h
@@ -38,7 +38,7 @@
 
 #include "pipe/p_config.h"
 
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
 #include <time.h> /* timeval */
 #include <unistd.h> /* usleep */
 #endif
@@ -58,7 +58,7 @@ extern "C" {
  */
 struct util_time 
 {
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
    struct timeval tv;
 #else
    int64_t counter;
@@ -89,7 +89,7 @@ util_time_timeout(const struct util_time *start,
                   const struct util_time *end,
                   const struct util_time *curr);
 
-#if defined(PIPE_OS_LINUX)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD)
 #define util_time_sleep usleep
 #else
 void
diff --git a/src/gallium/auxiliary/util/u_timed_winsys.c b/src/gallium/auxiliary/util/u_timed_winsys.c
index f237e12d73..77b2a3a1c8 100644
--- a/src/gallium/auxiliary/util/u_timed_winsys.c
+++ b/src/gallium/auxiliary/util/u_timed_winsys.c
@@ -29,6 +29,7 @@
  * Authors: Keith Whitwell <keithw-at-tungstengraphics-dot-com>
  */
 
+#include "pipe/p_state.h"
 #include "pipe/internal/p_winsys_screen.h"
 #include "u_timed_winsys.h"
 #include "util/u_memory.h"
@@ -178,13 +179,13 @@ timed_buffer_unmap(struct pipe_winsys *winsys,
 
 
 static void
-timed_buffer_destroy(struct pipe_winsys *winsys,
-                         struct pipe_buffer *buf)
+timed_buffer_destroy(struct pipe_buffer *buf)
 {
+   struct pipe_winsys *winsys = buf->screen->winsys;
    struct pipe_winsys *backend = timed_winsys(winsys)->backend;
    uint64_t start = time_start();
 
-   backend->buffer_destroy( backend, buf );
+   backend->buffer_destroy( buf );
 
    time_finish(winsys, start, 4, __FUNCTION__);
 }