Merge remote branch 'origin/lp-binning'

Conflicts:
	src/gallium/auxiliary/util/u_dl.c
	src/gallium/auxiliary/util/u_time.h
	src/gallium/drivers/llvmpipe/lp_state_derived.c
	src/gallium/drivers/llvmpipe/lp_state_surface.c
	src/gallium/drivers/llvmpipe/lp_tex_cache.c
	src/gallium/drivers/llvmpipe/lp_tile_cache.c

1 files changed, 1023 insertions, 0 deletions

diff --git a/src/gallium/drivers/llvmpipe/lp_rast.c b/src/gallium/drivers/llvmpipe/lp_rast.c
new file mode 100644
index 0000000000..54af850467
--- /dev/null
+++ b/src/gallium/drivers/llvmpipe/lp_rast.c
@@ -0,0 +1,1023 @@
+/**************************************************************************
+ *
+ * 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.
+ *
+ **************************************************************************/
+
+#include <limits.h>
+#include "util/u_memory.h"
+#include "util/u_math.h"
+#include "util/u_cpu_detect.h"
+#include "util/u_surface.h"
+
+#include "lp_scene_queue.h"
+#include "lp_debug.h"
+#include "lp_fence.h"
+#include "lp_rast.h"
+#include "lp_rast_priv.h"
+#include "lp_tile_soa.h"
+#include "lp_bld_debug.h"
+#include "lp_scene.h"
+
+
+/**
+ * Begin the rasterization phase.
+ * Map the framebuffer surfaces.  Initialize the 'rast' state.
+ */
+static boolean
+lp_rast_begin( struct lp_rasterizer *rast,
+               const struct pipe_framebuffer_state *fb,
+               boolean write_color,
+               boolean write_zstencil )
+{
+   struct pipe_screen *screen = rast->screen;
+   struct pipe_surface *cbuf, *zsbuf;
+   int i;
+
+   LP_DBG(DEBUG_RAST, "%s\n", __FUNCTION__);
+
+   util_copy_framebuffer_state(&rast->state.fb, fb);
+
+   rast->state.write_zstencil = write_zstencil;
+   rast->state.write_color = write_color;
+
+   rast->check_for_clipped_tiles = (fb->width % TILE_SIZE != 0 ||
+                                    fb->height % TILE_SIZE != 0);
+
+   
+   for (i = 0; i < rast->state.fb.nr_cbufs; i++) {
+      cbuf = rast->state.fb.cbufs[i];
+      if (cbuf) {
+	 rast->cbuf_transfer[i] = screen->get_tex_transfer(rast->screen,
+							   cbuf->texture,
+							   cbuf->face,
+							   cbuf->level,
+							   cbuf->zslice,
+							   PIPE_TRANSFER_READ_WRITE,
+							   0, 0,
+							   cbuf->width, 
+							   cbuf->height);
+	 if (!rast->cbuf_transfer[i])
+	    goto fail;
+
+	 rast->cbuf_map[i] = screen->transfer_map(rast->screen, 
+						  rast->cbuf_transfer[i]);
+	 if (!rast->cbuf_map[i])
+	    goto fail;
+      }
+   }
+
+   zsbuf = rast->state.fb.zsbuf;
+   if (zsbuf) {
+      rast->zsbuf_transfer = screen->get_tex_transfer(rast->screen,
+                                                      zsbuf->texture,
+                                                      zsbuf->face,
+                                                      zsbuf->level,
+                                                      zsbuf->zslice,
+                                                      PIPE_TRANSFER_READ_WRITE,
+                                                      0, 0,
+                                                      zsbuf->width,
+						      zsbuf->height);
+      if (!rast->zsbuf_transfer)
+         goto fail;
+
+      rast->zsbuf_map = screen->transfer_map(rast->screen, 
+                                            rast->zsbuf_transfer);
+      if (!rast->zsbuf_map)
+	 goto fail;
+   }
+
+   return TRUE;
+
+fail:
+   /* Unmap and release transfers?
+    */
+   return FALSE;
+}
+
+
+/**
+ * Finish the rasterization phase.
+ * Unmap framebuffer surfaces.
+ */
+static void
+lp_rast_end( struct lp_rasterizer *rast )
+{
+   struct pipe_screen *screen = rast->screen;
+   unsigned i;
+
+   for (i = 0; i < rast->state.fb.nr_cbufs; i++) {
+      if (rast->cbuf_map[i]) 
+	 screen->transfer_unmap(screen, rast->cbuf_transfer[i]);
+
+      if (rast->cbuf_transfer[i])
+	 screen->tex_transfer_destroy(rast->cbuf_transfer[i]);
+
+      rast->cbuf_transfer[i] = NULL;
+      rast->cbuf_map[i] = NULL;
+   }
+
+   if (rast->zsbuf_map) 
+      screen->transfer_unmap(screen, rast->zsbuf_transfer);
+
+   if (rast->zsbuf_transfer)
+      screen->tex_transfer_destroy(rast->zsbuf_transfer);
+
+   rast->zsbuf_transfer = NULL;
+   rast->zsbuf_map = NULL;
+}
+
+
+/**
+ * Begining rasterization of a tile.
+ * \param x  window X position of the tile, in pixels
+ * \param y  window Y position of the tile, in pixels
+ */
+static void
+lp_rast_start_tile( struct lp_rasterizer *rast,
+                    unsigned thread_index,
+                    unsigned x, unsigned y )
+{
+   LP_DBG(DEBUG_RAST, "%s %d,%d\n", __FUNCTION__, x, y);
+
+   rast->tasks[thread_index].x = x;
+   rast->tasks[thread_index].y = y;
+}
+
+
+/**
+ * Clear the rasterizer's current color tile.
+ * This is a bin command called during bin processing.
+ */
+void lp_rast_clear_color( struct lp_rasterizer *rast,
+                          unsigned thread_index,
+                          const union lp_rast_cmd_arg arg )
+{
+   const uint8_t *clear_color = arg.clear_color;
+   uint8_t **color_tile = rast->tasks[thread_index].tile.color;
+   unsigned i;
+
+   LP_DBG(DEBUG_RAST, "%s 0x%x,0x%x,0x%x,0x%x\n", __FUNCTION__, 
+              clear_color[0],
+              clear_color[1],
+              clear_color[2],
+              clear_color[3]);
+
+   if (clear_color[0] == clear_color[1] &&
+       clear_color[1] == clear_color[2] &&
+       clear_color[2] == clear_color[3]) {
+      /* clear to grayscale value {x, x, x, x} */
+      for (i = 0; i < rast->state.fb.nr_cbufs; i++) {
+	 memset(color_tile[i], clear_color[0], TILE_SIZE * TILE_SIZE * 4);
+      }
+   }
+   else {
+      /* Non-gray color.
+       * Note: if the swizzled tile layout changes (see TILE_PIXEL) this code
+       * will need to change.  It'll be pretty obvious when clearing no longer
+       * works.
+       */
+      const unsigned chunk = TILE_SIZE / 4;
+      for (i = 0; i < rast->state.fb.nr_cbufs; i++) {
+         uint8_t *c = color_tile[i];
+         unsigned j;
+         for (j = 0; j < 4 * TILE_SIZE; j++) {
+            memset(c, clear_color[0], chunk);
+            c += chunk;
+            memset(c, clear_color[1], chunk);
+            c += chunk;
+            memset(c, clear_color[2], chunk);
+            c += chunk;
+            memset(c, clear_color[3], chunk);
+            c += chunk;
+         }
+         assert(c - color_tile[i] == TILE_SIZE * TILE_SIZE * 4);
+      }
+   }
+}
+
+
+/**
+ * Clear the rasterizer's current z/stencil tile.
+ * This is a bin command called during bin processing.
+ */
+void lp_rast_clear_zstencil( struct lp_rasterizer *rast,
+                             unsigned thread_index,
+                             const union lp_rast_cmd_arg arg)
+{
+   unsigned i;
+   uint32_t *depth_tile = rast->tasks[thread_index].tile.depth;
+   
+   LP_DBG(DEBUG_RAST, "%s 0x%x\n", __FUNCTION__, arg.clear_zstencil);
+
+   for (i = 0; i < TILE_SIZE * TILE_SIZE; i++)
+      depth_tile[i] = arg.clear_zstencil;
+}
+
+
+/**
+ * Load tile color from the framebuffer surface.
+ * This is a bin command called during bin processing.
+ */
+void lp_rast_load_color( struct lp_rasterizer *rast,
+                         unsigned thread_index,
+                         const union lp_rast_cmd_arg arg)
+{
+   struct lp_rasterizer_task *task = &rast->tasks[thread_index];
+   const unsigned x = task->x;
+   const unsigned y = task->y;
+   unsigned i;
+
+   LP_DBG(DEBUG_RAST, "%s at %u, %u\n", __FUNCTION__, x, y);
+
+   for (i = 0; i < rast->state.fb.nr_cbufs; i++) {
+      struct pipe_transfer *transfer = rast->cbuf_transfer[i];
+      int w = TILE_SIZE;
+      int h = TILE_SIZE;
+
+      if (x >= transfer->width)
+	 continue;
+
+      if (y >= transfer->height)
+	 continue;
+
+      assert(w >= 0);
+      assert(h >= 0);
+      assert(w <= TILE_SIZE);
+      assert(h <= TILE_SIZE);
+
+      lp_tile_read_4ub(transfer->texture->format,
+		       task->tile.color[i],
+		       rast->cbuf_map[i], 
+		       transfer->stride,
+		       x, y,
+		       w, h);
+   }
+}
+
+
+static void
+lp_tile_read_z32(uint32_t *tile,
+                 const uint8_t *map,
+                 unsigned map_stride,
+                 unsigned x0, unsigned y0, unsigned w, unsigned h)
+{
+   unsigned x, y;
+   const uint8_t *map_row = map + y0*map_stride;
+   for (y = 0; y < h; ++y) {
+      const uint32_t *map_pixel = (uint32_t *)(map_row + x0*4);
+      for (x = 0; x < w; ++x) {
+         *tile++ = *map_pixel++;
+      }
+      map_row += map_stride;
+   }
+}
+
+/**
+ * Load tile z/stencil from the framebuffer surface.
+ * This is a bin command called during bin processing.
+ */
+void lp_rast_load_zstencil( struct lp_rasterizer *rast,
+                            unsigned thread_index,
+                            const union lp_rast_cmd_arg arg )
+{
+   struct lp_rasterizer_task *task = &rast->tasks[thread_index];
+   const unsigned x = task->x;
+   const unsigned y = task->y;
+   unsigned w = TILE_SIZE;
+   unsigned h = TILE_SIZE;
+
+   if (x + w > rast->state.fb.width)
+      w -= x + w - rast->state.fb.width;
+
+   if (y + h > rast->state.fb.height)
+      h -= y + h - rast->state.fb.height;
+
+   LP_DBG(DEBUG_RAST, "%s %d,%d %dx%d\n", __FUNCTION__, x, y, w, h);
+
+   assert(rast->zsbuf_transfer->texture->format == PIPE_FORMAT_Z32_UNORM);
+   lp_tile_read_z32(task->tile.depth,
+                    rast->zsbuf_map, 
+                    rast->zsbuf_transfer->stride,
+                    x, y, w, h);
+}
+
+
+void lp_rast_set_state( struct lp_rasterizer *rast,
+                        unsigned thread_index,
+                        const union lp_rast_cmd_arg arg )
+{
+   const struct lp_rast_state *state = arg.set_state;
+
+   LP_DBG(DEBUG_RAST, "%s %p\n", __FUNCTION__, (void *) state);
+
+   /* just set the current state pointer for this rasterizer */
+   rast->tasks[thread_index].current_state = state;
+}
+
+
+
+/**
+ * Run the shader on all blocks in a tile.  This is used when a tile is
+ * completely contained inside a triangle.
+ * This is a bin command called during bin processing.
+ */
+void lp_rast_shade_tile( struct lp_rasterizer *rast,
+                         unsigned thread_index,
+                         const union lp_rast_cmd_arg arg )
+{
+   struct lp_rasterizer_task *task = &rast->tasks[thread_index];
+   const struct lp_rast_state *state = task->current_state;
+   struct lp_rast_tile *tile = &task->tile;
+   const struct lp_rast_shader_inputs *inputs = arg.shade_tile;
+   const unsigned tile_x = task->x;
+   const unsigned tile_y = task->y;
+   unsigned x, y;
+
+   LP_DBG(DEBUG_RAST, "%s\n", __FUNCTION__);
+
+   /* render the whole 64x64 tile in 4x4 chunks */
+   for (y = 0; y < TILE_SIZE; y += 4){
+      for (x = 0; x < TILE_SIZE; x += 4) {
+         uint8_t *color[PIPE_MAX_COLOR_BUFS];
+         uint32_t *depth;
+         unsigned block_offset, i;
+
+         /* offset of the 16x16 pixel block within the tile */
+         block_offset = ((y / 4) * (16 * 16) + (x / 4) * 16);
+
+         /* color buffer */
+         for (i = 0; i < rast->state.fb.nr_cbufs; i++)
+            color[i] = tile->color[i] + 4 * block_offset;
+
+         /* depth buffer */
+         depth = tile->depth + block_offset;
+
+         /* run shader */
+         state->jit_function[0]( &state->jit_context,
+                                 tile_x + x, tile_y + y,
+                                 inputs->a0,
+                                 inputs->dadx,
+                                 inputs->dady,
+                                 color,
+                                 depth,
+                                 INT_MIN, INT_MIN, INT_MIN,
+                                 NULL, NULL, NULL );
+      }
+   }
+}
+
+
+/**
+ * Compute shading for a 4x4 block of pixels.
+ * This is a bin command called during bin processing.
+ */
+void lp_rast_shade_quads( struct lp_rasterizer *rast,
+                          unsigned thread_index,
+                          const struct lp_rast_shader_inputs *inputs,
+                          unsigned x, unsigned y,
+                          int32_t c1, int32_t c2, int32_t c3)
+{
+   struct lp_rasterizer_task *task = &rast->tasks[thread_index];
+   const struct lp_rast_state *state = task->current_state;
+   struct lp_rast_tile *tile = &task->tile;
+   uint8_t *color[PIPE_MAX_COLOR_BUFS];
+   void *depth;
+   unsigned i;
+   unsigned ix, iy;
+   int block_offset;
+
+#ifdef DEBUG
+   assert(state);
+
+   /* Sanity checks */
+   assert(x % TILE_VECTOR_WIDTH == 0);
+   assert(y % TILE_VECTOR_HEIGHT == 0);
+
+   assert((x % 4) == 0);
+   assert((y % 4) == 0);
+#endif
+
+   ix = x % TILE_SIZE;
+   iy = y % TILE_SIZE;
+
+   /* offset of the 16x16 pixel block within the tile */
+   block_offset = ((iy / 4) * (16 * 16) + (ix / 4) * 16);
+
+   /* color buffer */
+   for (i = 0; i < rast->state.fb.nr_cbufs; i++)
+      color[i] = tile->color[i] + 4 * block_offset;
+
+   /* depth buffer */
+   depth = tile->depth + block_offset;
+
+
+
+#ifdef DEBUG
+   assert(lp_check_alignment(tile->depth, 16));
+   assert(lp_check_alignment(tile->color[0], 16));
+   assert(lp_check_alignment(state->jit_context.blend_color, 16));
+
+   assert(lp_check_alignment(inputs->step[0], 16));
+   assert(lp_check_alignment(inputs->step[1], 16));
+   assert(lp_check_alignment(inputs->step[2], 16));
+#endif
+
+   /* run shader */
+   state->jit_function[1]( &state->jit_context,
+                        x, y,
+                        inputs->a0,
+                        inputs->dadx,
+                        inputs->dady,
+                        color,
+                        depth,
+                        c1, c2, c3,
+                        inputs->step[0], inputs->step[1], inputs->step[2]);
+}
+
+
+/**
+ * Set top row and left column of the tile's pixels to white.  For debugging.
+ */
+static void
+outline_tile(uint8_t *tile)
+{
+   const uint8_t val = 0xff;
+   unsigned i;
+
+   for (i = 0; i < TILE_SIZE; i++) {
+      TILE_PIXEL(tile, i, 0, 0) = val;
+      TILE_PIXEL(tile, i, 0, 1) = val;
+      TILE_PIXEL(tile, i, 0, 2) = val;
+      TILE_PIXEL(tile, i, 0, 3) = val;
+
+      TILE_PIXEL(tile, 0, i, 0) = val;
+      TILE_PIXEL(tile, 0, i, 1) = val;
+      TILE_PIXEL(tile, 0, i, 2) = val;
+      TILE_PIXEL(tile, 0, i, 3) = val;
+   }
+}
+
+
+/**
+ * Draw grid of gray lines at 16-pixel intervals across the tile to
+ * show the sub-tile boundaries.  For debugging.
+ */
+static void
+outline_subtiles(uint8_t *tile)
+{
+   const uint8_t val = 0x80;
+   const unsigned step = 16;
+   unsigned i, j;
+
+   for (i = 0; i < TILE_SIZE; i += step) {
+      for (j = 0; j < TILE_SIZE; j++) {
+         TILE_PIXEL(tile, i, j, 0) = val;
+         TILE_PIXEL(tile, i, j, 1) = val;
+         TILE_PIXEL(tile, i, j, 2) = val;
+         TILE_PIXEL(tile, i, j, 3) = val;
+
+         TILE_PIXEL(tile, j, i, 0) = val;
+         TILE_PIXEL(tile, j, i, 1) = val;
+         TILE_PIXEL(tile, j, i, 2) = val;
+         TILE_PIXEL(tile, j, i, 3) = val;
+      }
+   }
+
+   outline_tile(tile);
+}
+
+
+
+/**
+ * Write the rasterizer's color tile to the framebuffer.
+ */
+static void lp_rast_store_color( struct lp_rasterizer *rast,
+                                 unsigned thread_index)
+{
+   struct lp_rasterizer_task *task = &rast->tasks[thread_index];
+   const unsigned x = task->x;
+   const unsigned y = task->y;
+   unsigned i;
+
+   for (i = 0; i < rast->state.fb.nr_cbufs; i++) {
+      struct pipe_transfer *transfer = rast->cbuf_transfer[i];
+      int w = TILE_SIZE;
+      int h = TILE_SIZE;
+
+      if (x >= transfer->width)
+	 continue;
+
+      if (y >= transfer->height)
+	 continue;
+
+      LP_DBG(DEBUG_RAST, "%s [%u] %d,%d %dx%d\n", __FUNCTION__,
+	     thread_index, x, y, w, h);
+
+      if (LP_DEBUG & DEBUG_SHOW_SUBTILES)
+         outline_subtiles(task->tile.color[i]);
+      else if (LP_DEBUG & DEBUG_SHOW_TILES)
+         outline_tile(task->tile.color[i]);
+
+      lp_tile_write_4ub(transfer->texture->format,
+			task->tile.color[i],
+			rast->cbuf_map[i], 
+			transfer->stride,
+			x, y,
+			w, h);
+   }
+}
+
+
+static void
+lp_tile_write_z32(const uint32_t *src, uint8_t *dst, unsigned dst_stride,
+                  unsigned x0, unsigned y0, unsigned w, unsigned h)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < h; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*4);
+      for (x = 0; x < w; ++x) {
+         *dst_pixel++ = *src++;
+      }
+      dst_row += dst_stride;
+   }
+}
+
+/**
+ * Write the rasterizer's z/stencil tile to the framebuffer.
+ */
+static void lp_rast_store_zstencil( struct lp_rasterizer *rast,
+                                    unsigned thread_index )
+{
+   struct lp_rasterizer_task *task = &rast->tasks[thread_index];
+   const unsigned x = task->x;
+   const unsigned y = task->y;
+   unsigned w = TILE_SIZE;
+   unsigned h = TILE_SIZE;
+
+   if (x + w > rast->state.fb.width)
+      w -= x + w - rast->state.fb.width;
+
+   if (y + h > rast->state.fb.height)
+      h -= y + h - rast->state.fb.height;
+
+   LP_DBG(DEBUG_RAST, "%s %d,%d %dx%d\n", __FUNCTION__, x, y, w, h);
+
+   assert(rast->zsbuf_transfer->texture->format == PIPE_FORMAT_Z32_UNORM);
+   lp_tile_write_z32(task->tile.depth,
+                     rast->zsbuf_map, 
+                     rast->zsbuf_transfer->stride,
+                     x, y, w, h);
+}
+
+
+/**
+ * Write the rasterizer's tiles to the framebuffer.
+ */
+static void
+lp_rast_end_tile( struct lp_rasterizer *rast,
+                  unsigned thread_index )
+{
+   LP_DBG(DEBUG_RAST, "%s\n", __FUNCTION__);
+
+   if (rast->state.write_color)
+      lp_rast_store_color(rast, thread_index);
+
+   if (rast->state.write_zstencil)
+      lp_rast_store_zstencil(rast, thread_index);
+}
+
+
+/**
+ * Signal on a fence.  This is called during bin execution/rasterization.
+ * Called per thread.
+ */
+void lp_rast_fence( struct lp_rasterizer *rast,
+                    unsigned thread_index,
+                    const union lp_rast_cmd_arg arg )
+{
+   struct lp_fence *fence = arg.fence;
+
+   pipe_mutex_lock( fence->mutex );
+
+   fence->count++;
+   assert(fence->count <= fence->rank);
+
+   LP_DBG(DEBUG_RAST, "%s count=%u rank=%u\n", __FUNCTION__,
+          fence->count, fence->rank);
+
+   pipe_condvar_signal( fence->signalled );
+
+   pipe_mutex_unlock( fence->mutex );
+}
+
+
+/**
+ * When all the threads are done rasterizing a scene, one thread will
+ * call this function to reset the scene and put it onto the empty queue.
+ */
+static void
+release_scene( struct lp_rasterizer *rast,
+	       struct lp_scene *scene )
+{
+   util_unreference_framebuffer_state( &scene->fb );
+
+   lp_scene_reset( scene );
+   lp_scene_enqueue( rast->empty_scenes, scene );
+   rast->curr_scene = NULL;
+}
+
+
+/**
+ * Rasterize commands for a single bin.
+ * \param x, y  position of the bin's tile in the framebuffer
+ * Must be called between lp_rast_begin() and lp_rast_end().
+ * Called per thread.
+ */
+static void
+rasterize_bin( struct lp_rasterizer *rast,
+               unsigned thread_index,
+               const struct cmd_bin *bin,
+               int x, int y)
+{
+   const struct cmd_block_list *commands = &bin->commands;
+   struct cmd_block *block;
+   unsigned k;
+
+   lp_rast_start_tile( rast, thread_index, x, y );
+
+   /* simply execute each of the commands in the block list */
+   for (block = commands->head; block; block = block->next) {
+      for (k = 0; k < block->count; k++) {
+         block->cmd[k]( rast, thread_index, block->arg[k] );
+      }
+   }
+
+   lp_rast_end_tile( rast, thread_index );
+}
+
+
+#define RAST(x) { lp_rast_##x, #x }
+
+static struct {
+   lp_rast_cmd cmd;
+   const char *name;
+} cmd_names[] = 
+{
+   RAST(load_color),
+   RAST(load_zstencil),
+   RAST(clear_color),
+   RAST(clear_zstencil),
+   RAST(triangle),
+   RAST(shade_tile),
+   RAST(set_state),
+   RAST(fence),
+};
+
+static void
+debug_bin( const struct cmd_bin *bin )
+{
+   const struct cmd_block *head = bin->commands.head;
+   int i, j;
+
+   for (i = 0; i < head->count; i++) {
+      debug_printf("%d: ", i);
+      for (j = 0; j < Elements(cmd_names); j++) {
+         if (head->cmd[i] == cmd_names[j].cmd) {
+            debug_printf("%s\n", cmd_names[j].name);
+            break;
+         }
+      }
+      if (j == Elements(cmd_names))
+         debug_printf("...other\n");
+   }
+
+}
+
+/* An empty bin is one that just loads the contents of the tile and
+ * stores them again unchanged.  This typically happens when bins have
+ * been flushed for some reason in the middle of a frame, or when
+ * incremental updates are being made to a render target.
+ * 
+ * Try to avoid doing pointless work in this case.
+ */
+static boolean
+is_empty_bin( const struct cmd_bin *bin )
+{
+   const struct cmd_block *head = bin->commands.head;
+   int i;
+   
+   if (0)
+      debug_bin(bin);
+   
+   /* We emit at most two load-tile commands at the start of the first
+    * command block.  In addition we seem to emit a couple of
+    * set-state commands even in empty bins.
+    *
+    * As a heuristic, if a bin has more than 4 commands, consider it
+    * non-empty.
+    */
+   if (head->next != NULL ||
+       head->count > 4) {
+      return FALSE;
+   }
+
+   for (i = 0; i < head->count; i++)
+      if (head->cmd[i] != lp_rast_load_color &&
+          head->cmd[i] != lp_rast_load_zstencil &&
+          head->cmd[i] != lp_rast_set_state) {
+         return FALSE;
+      }
+
+   return TRUE;
+}
+
+
+
+/**
+ * Rasterize/execute all bins within a scene.
+ * Called per thread.
+ */
+static void
+rasterize_scene( struct lp_rasterizer *rast,
+                unsigned thread_index,
+                struct lp_scene *scene,
+                bool write_depth )
+{
+   /* loop over scene bins, rasterize each */
+#if 0
+   {
+      unsigned i, j;
+      for (i = 0; i < scene->tiles_x; i++) {
+         for (j = 0; j < scene->tiles_y; j++) {
+            struct cmd_bin *bin = lp_get_bin(scene, i, j);
+            rasterize_bin( rast, thread_index,
+                           bin, i * TILE_SIZE, j * TILE_SIZE );
+         }
+      }
+   }
+#else
+   {
+      struct cmd_bin *bin;
+      int x, y;
+
+      assert(scene);
+      while ((bin = lp_scene_bin_iter_next(scene, &x, &y))) {
+         if (!is_empty_bin( bin ))
+            rasterize_bin( rast, thread_index, bin, x * TILE_SIZE, y * TILE_SIZE);
+      }
+   }
+#endif
+}
+
+
+/**
+ * Called by setup module when it has something for us to render.
+ */
+void
+lp_rasterize_scene( struct lp_rasterizer *rast,
+                   struct lp_scene *scene,
+                   const struct pipe_framebuffer_state *fb,
+                   bool write_depth )
+{
+   boolean debug = false;
+
+   LP_DBG(DEBUG_SETUP, "%s\n", __FUNCTION__);
+
+   if (debug) {
+      unsigned x, y;
+      debug_printf("rasterize scene:\n");
+      debug_printf("  data size: %u\n", lp_scene_data_size(scene));
+      for (y = 0; y < scene->tiles_y; y++) {
+         for (x = 0; x < scene->tiles_x; x++) {
+            debug_printf("  bin %u, %u size: %u\n", x, y,
+                         lp_scene_bin_size(scene, x, y));
+         }
+      }
+   }
+
+   /* save framebuffer state in the bin */
+   util_copy_framebuffer_state(&scene->fb, fb);
+   scene->write_depth = write_depth;
+
+   if (rast->num_threads == 0) {
+      /* no threading */
+
+      lp_rast_begin( rast, fb,
+                     fb->nr_cbufs != 0, /* always write color if cbufs present */
+                     fb->zsbuf != NULL && write_depth );
+
+      lp_scene_bin_iter_begin( scene );
+      rasterize_scene( rast, 0, scene, write_depth );
+
+      release_scene( rast, scene );
+
+      lp_rast_end( rast );
+   }
+   else {
+      /* threaded rendering! */
+      unsigned i;
+
+      lp_scene_enqueue( rast->full_scenes, scene );
+
+      /* signal the threads that there's work to do */
+      for (i = 0; i < rast->num_threads; i++) {
+         pipe_semaphore_signal(&rast->tasks[i].work_ready);
+      }
+
+      /* wait for work to complete */
+      for (i = 0; i < rast->num_threads; i++) {
+         pipe_semaphore_wait(&rast->tasks[i].work_done);
+      }
+   }
+
+   LP_DBG(DEBUG_SETUP, "%s done \n", __FUNCTION__);
+}
+
+
+/**
+ * This is the thread's main entrypoint.
+ * It's a simple loop:
+ *   1. wait for work
+ *   2. do work
+ *   3. signal that we're done
+ */
+static PIPE_THREAD_ROUTINE( thread_func, init_data )
+{
+   struct lp_rasterizer_task *task = (struct lp_rasterizer_task *) init_data;
+   struct lp_rasterizer *rast = task->rast;
+   boolean debug = false;
+
+   while (1) {
+      /* wait for work */
+      if (debug)
+         debug_printf("thread %d waiting for work\n", task->thread_index);
+      pipe_semaphore_wait(&task->work_ready);
+
+      if (task->thread_index == 0) {
+         /* thread[0]:
+          *  - get next scene to rasterize
+          *  - map the framebuffer surfaces
+          */
+         const struct pipe_framebuffer_state *fb;
+         boolean write_depth;
+
+         rast->curr_scene = lp_scene_dequeue( rast->full_scenes, TRUE );
+
+         lp_scene_bin_iter_begin( rast->curr_scene );
+
+         fb = &rast->curr_scene->fb;
+         write_depth = rast->curr_scene->write_depth;
+
+         lp_rast_begin( rast, fb,
+                        fb->nr_cbufs != 0,
+                        fb->zsbuf != NULL && write_depth );
+      }
+
+      /* Wait for all threads to get here so that threads[1+] don't
+       * get a null rast->curr_scene pointer.
+       */
+      pipe_barrier_wait( &rast->barrier );
+
+      /* do work */
+      if (debug)
+         debug_printf("thread %d doing work\n", task->thread_index);
+      rasterize_scene(rast, 
+		     task->thread_index,
+                     rast->curr_scene, 
+		     rast->curr_scene->write_depth);
+      
+      /* wait for all threads to finish with this scene */
+      pipe_barrier_wait( &rast->barrier );
+
+      if (task->thread_index == 0) {
+         /* thread[0]:
+          * - release the scene object
+          * - unmap the framebuffer surfaces
+          */
+         release_scene( rast, rast->curr_scene );
+         lp_rast_end( rast );
+      }
+
+      /* signal done with work */
+      if (debug)
+         debug_printf("thread %d done working\n", task->thread_index);
+      pipe_semaphore_signal(&task->work_done);
+   }
+
+   return NULL;
+}
+
+
+/**
+ * Initialize semaphores and spawn the threads.
+ */
+static void
+create_rast_threads(struct lp_rasterizer *rast)
+{
+   unsigned i;
+
+   rast->num_threads = util_cpu_caps.nr_cpus;
+   rast->num_threads = debug_get_num_option("LP_NUM_THREADS", rast->num_threads);
+   rast->num_threads = MIN2(rast->num_threads, MAX_THREADS);
+
+   /* NOTE: if num_threads is zero, we won't use any threads */
+   for (i = 0; i < rast->num_threads; i++) {
+      pipe_semaphore_init(&rast->tasks[i].work_ready, 0);
+      pipe_semaphore_init(&rast->tasks[i].work_done, 0);
+      rast->threads[i] = pipe_thread_create(thread_func,
+                                            (void *) &rast->tasks[i]);
+   }
+}
+
+
+
+/**
+ * Create new lp_rasterizer.
+ * \param empty  the queue to put empty scenes on after we've finished
+ *               processing them.
+ */
+struct lp_rasterizer *
+lp_rast_create( struct pipe_screen *screen, struct lp_scene_queue *empty )
+{
+   struct lp_rasterizer *rast;
+   unsigned i, cbuf;
+
+   rast = CALLOC_STRUCT(lp_rasterizer);
+   if(!rast)
+      return NULL;
+
+   rast->screen = screen;
+
+   rast->empty_scenes = empty;
+   rast->full_scenes = lp_scene_queue_create();
+
+   for (i = 0; i < Elements(rast->tasks); i++) {
+      struct lp_rasterizer_task *task = &rast->tasks[i];
+
+      for (cbuf = 0; cbuf < PIPE_MAX_COLOR_BUFS; cbuf++ )
+	 task->tile.color[cbuf] = align_malloc(TILE_SIZE * TILE_SIZE * 4, 16);
+
+      task->tile.depth = align_malloc(TILE_SIZE * TILE_SIZE * 4, 16);
+      task->rast = rast;
+      task->thread_index = i;
+   }
+
+   create_rast_threads(rast);
+
+   /* for synchronizing rasterization threads */
+   pipe_barrier_init( &rast->barrier, rast->num_threads );
+
+   return rast;
+}
+
+
+/* Shutdown:
+ */
+void lp_rast_destroy( struct lp_rasterizer *rast )
+{
+   unsigned i, cbuf;
+
+   util_unreference_framebuffer_state(&rast->state.fb);
+
+   for (i = 0; i < Elements(rast->tasks); i++) {
+      align_free(rast->tasks[i].tile.depth);
+      for (cbuf = 0; cbuf < PIPE_MAX_COLOR_BUFS; cbuf++ )
+	 align_free(rast->tasks[i].tile.color[cbuf]);
+   }
+
+   /* for synchronizing rasterization threads */
+   pipe_barrier_destroy( &rast->barrier );
+
+   FREE(rast);
+}
+
+
+/** Return number of rasterization threads */
+unsigned
+lp_rast_get_num_threads( struct lp_rasterizer *rast )
+{
+   return rast->num_threads;
+}