cell: Use code-gen for alpha blend

So far this is only tested when GL_BLEND is disabled.

src/gallium/drivers/cell/common.h

@@ -114,6 +114,16 @@ struct cell_command_depth_stencil_alpha_test {
 };
 
 
+/**
+ * Upload code to perform framebuffer blend operation
+ */
+struct cell_command_blend {
+   uint64_t base;               /**< Effective address of code start. */
+   unsigned size;               /**< Size in bytes of test code. */
+   unsigned read_fb;            /**< Flag: should framebuffer be read? */
+};
+
+
 /**
  * Tell SPUs about the framebuffer size, location
  */

src/gallium/drivers/cell/ppu/cell_pipe_state.c

@@ -54,14 +54,19 @@ cell_create_blend_state(struct pipe_context *pipe,
 
 
 static void
-cell_bind_blend_state(struct pipe_context *pipe, void *blend)
+cell_bind_blend_state(struct pipe_context *pipe, void *state)
 {
    struct cell_context *cell = cell_context(pipe);
+   struct cell_blend_state *blend = (struct cell_blend_state *) state;
+
 
    draw_flush(cell->draw);
 
-   cell->blend = (const struct cell_blend_state *)blend;
+   if ((blend != NULL) && (blend->code.store == NULL)) {
+      cell_generate_depth_stencil_test(blend);
+   }
 
+   cell->blend = blend;
    cell->dirty |= CELL_NEW_BLEND;
 }
 
@@ -70,7 +75,7 @@ static void
 cell_delete_blend_state(struct pipe_context *pipe, void *blend)
 {
    struct cell_blend_state *cb = (struct cell_blend_state *) blend;
-   
+
    spe_release_func(& cb->code);
    FREE(cb);
 }

src/gallium/drivers/cell/ppu/cell_state_emit.c

@@ -65,9 +65,20 @@ cell_emit_state(struct cell_context *cell)
    }
 
    if (cell->dirty & CELL_NEW_BLEND) {
-      emit_state_cmd(cell, CELL_CMD_STATE_BLEND,
-                     cell->blend,
-                     sizeof(struct pipe_blend_state));
+      struct cell_command_blend blend;
+
+      if (cell->blend != NULL) {
+         blend.base = (intptr_t) cell->blend->code.store;
+         blend.size = (char *) cell->blend->code.csr
+             - (char *) cell->blend->code.store;
+         blend.read_fb = TRUE;
+      } else {
+         blend.base = 0;
+         blend.size = 0;
+         blend.read_fb = FALSE;
+      }
+
+      emit_state_cmd(cell, CELL_CMD_STATE_BLEND, &blend, sizeof(blend));
    }
 
    if (cell->dirty & CELL_NEW_DEPTH_STENCIL) {

src/gallium/drivers/cell/spu/Makefile

@@ -17,7 +17,6 @@ PROG_SPU_EMBED_O = $(PROG)_spu-embed.o
 
 SOURCES = \
 	spu_main.c \
-	spu_blend.c \
 	spu_dcache.c \
 	spu_per_fragment_op.c \
 	spu_render.c \

src/gallium/drivers/cell/spu/spu_blend.c

@@ -1,62 +0,0 @@
-/**************************************************************************
- * 
- * 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.
- * 
- **************************************************************************/
-
-
-#include "spu_main.h"
-#include "spu_blend.h"
-#include "spu_colorpack.h"
-
-
-void
-blend_quad(uint itx, uint ity, vector float colors[4])
-{
-   /* simple SRC_ALPHA, ONE_MINUS_SRC_ALPHA blending */
-   vector float fbc00 = spu_unpack_color(spu.ctile.ui[ity][itx]);
-   vector float fbc01 = spu_unpack_color(spu.ctile.ui[ity][itx+1]);
-   vector float fbc10 = spu_unpack_color(spu.ctile.ui[ity+1][itx]);
-   vector float fbc11 = spu_unpack_color(spu.ctile.ui[ity+1][itx+1]);
-
-   vector float alpha00 = spu_splats(spu_extract(colors[0], 3));
-   vector float alpha01 = spu_splats(spu_extract(colors[1], 3));
-   vector float alpha10 = spu_splats(spu_extract(colors[2], 3));
-   vector float alpha11 = spu_splats(spu_extract(colors[3], 3));
-
-   vector float one_minus_alpha00 = spu_sub(spu_splats(1.0f), alpha00);
-   vector float one_minus_alpha01 = spu_sub(spu_splats(1.0f), alpha01);
-   vector float one_minus_alpha10 = spu_sub(spu_splats(1.0f), alpha10);
-   vector float one_minus_alpha11 = spu_sub(spu_splats(1.0f), alpha11);
-
-   colors[0] = spu_add(spu_mul(colors[0], alpha00),
-                       spu_mul(fbc00, one_minus_alpha00));
-   colors[1] = spu_add(spu_mul(colors[1], alpha01),
-                       spu_mul(fbc01, one_minus_alpha01));
-   colors[2] = spu_add(spu_mul(colors[2], alpha10),
-                       spu_mul(fbc10, one_minus_alpha10));
-   colors[3] = spu_add(spu_mul(colors[3], alpha11),
-                       spu_mul(fbc11, one_minus_alpha11));
-}
-

src/gallium/drivers/cell/spu/spu_blend.h

@@ -1,37 +0,0 @@
-/**************************************************************************
- * 
- * 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.
- * 
- **************************************************************************/
-
-
-#ifndef SPU_BLEND_H
-#define SPU_BLEND_H
-
-
-extern void
-blend_quad(uint itx, uint ity, vector float colors[4]);
-
-
-#endif /* SPU_BLEND_H */

src/gallium/drivers/cell/spu/spu_main.c

@@ -61,6 +61,25 @@ static unsigned char attribute_fetch_code_buffer[136 * PIPE_ATTRIB_MAX]
 static unsigned char depth_stencil_code_buffer[4 * 64]
     ALIGN16_ATTRIB;
 
+static unsigned char fb_blend_code_buffer[4 * 64]
+    ALIGN16_ATTRIB;
+
+static struct spu_blend_results
+default_blend(qword frag_r, qword frag_g, qword frag_b, qword frag_a,
+              qword pixel_r, qword pixel_g, qword pixel_b, qword pixel_a,
+              qword frag_mask)
+{
+   struct spu_blend_results result;
+
+   result.r = si_selb(pixel_r, frag_r, frag_mask);
+   result.g = si_selb(pixel_g, frag_g, frag_mask);
+   result.b = si_selb(pixel_b, frag_b, frag_mask);
+   result.a = si_selb(pixel_a, frag_a, frag_mask);
+
+   return result;
+}
+
+
 /**
  * Tell the PPU that this SPU has finished copying a buffer to
  * local store and that it may be reused by the PPU.
@@ -246,14 +265,31 @@ cmd_state_framebuffer(const struct cell_command_framebuffer *cmd)
 
 
 static void
-cmd_state_blend(const struct pipe_blend_state *state)
+cmd_state_blend(const struct cell_command_blend *state)
 {
    if (Debug)
       printf("SPU %u: BLEND: enabled %d\n",
              spu.init.id,
-             state->blend_enable);
+             (state->size != 0));
+
+   ASSERT_ALIGN16(state->base);
 
-   memcpy(&spu.blend, state, sizeof(*state));
+   if (state->size != 0) {
+      mfc_get(fb_blend_code_buffer,
+              (unsigned int) state->base,  /* src */
+              ROUNDUP16(state->size),
+              TAG_BATCH_BUFFER,
+              0, /* tid */
+              0  /* rid */);
+      wait_on_mask(1 << TAG_BATCH_BUFFER);
+      spu.blend = (blend_func) fb_blend_code_buffer;
+      spu.read_fb = state->read_fb;
+   } else {
+      /* If there is no code, use the default;
+       */
+      spu.blend = default_blend;
+      spu.read_fb = FALSE;
+   }
 }
 
 
@@ -441,9 +477,8 @@ cmd_batch(uint opcode)
          pos += 1;
          break;
       case CELL_CMD_STATE_BLEND:
-         cmd_state_blend((struct pipe_blend_state *)
-                                 &buffer[pos+1]);
-         pos += (1 + ROUNDUP8(sizeof(struct pipe_blend_state)) / 8);
+         cmd_state_blend((struct cell_command_blend *) &buffer[pos+1]);
+         pos += (1 + ROUNDUP8(sizeof(struct cell_command_blend)) / 8);
          break;
       case CELL_CMD_STATE_DEPTH_STENCIL:
          cmd_state_depth_stencil((struct cell_command_depth_stencil_alpha_test *)
@@ -587,6 +622,9 @@ one_time_init(void)
    memset(spu.ctile_status, TILE_STATUS_DEFINED, sizeof(spu.ctile_status));
    memset(spu.ztile_status, TILE_STATUS_DEFINED, sizeof(spu.ztile_status));
    invalidate_tex_cache();
+
+   spu.blend = default_blend;
+   spu.read_fb = FALSE;
 }
 
 

src/gallium/drivers/cell/spu/spu_main.h

@@ -67,6 +67,18 @@ typedef struct spu_frag_test_results (*frag_test_func)(qword frag_mask,
     qword frag_alpha, qword facing);
 
 
+struct spu_blend_results {
+   qword r;
+   qword g;
+   qword b;
+   qword a;
+};
+
+typedef struct spu_blend_results (*blend_func)(
+    qword frag_r, qword frag_g, qword frag_b, qword frag_a,
+    qword pixel_r, qword pixel_g, qword pixel_b, qword pixel_a,
+    qword frag_mask);
+
 struct spu_framebuffer {
    void *color_start;              /**< addr of color surface in main memory */
    void *depth_start;              /**< addr of depth surface in main memory */
@@ -93,7 +105,10 @@ struct spu_global
    boolean read_depth;
    boolean read_stencil;
    frag_test_func frag_test;
-   struct pipe_blend_state blend;
+   
+   boolean read_fb;
+   blend_func blend;
+
    struct pipe_sampler_state sampler[PIPE_MAX_SAMPLERS];
    struct cell_command_texture texture;
 

src/gallium/drivers/cell/spu/spu_tri.c

@@ -29,10 +29,10 @@
  * Triangle rendering within a tile.
  */
 
+#include <transpose_matrix4x4.h>
 #include "pipe/p_compiler.h"
 #include "pipe/p_format.h"
 #include "pipe/p_util.h"
-#include "spu_blend.h"
 #include "spu_colorpack.h"
 #include "spu_main.h"
 #include "spu_texture.h"
@@ -326,27 +326,45 @@ emit_quad( int x, int y, mask_t mask )
          eval_coeff(1, (float) x, (float) y, colors);
       }
 
-#if 1
-      if (spu.blend.blend_enable)
-         blend_quad(ix % TILE_SIZE, iy % TILE_SIZE, colors);
-#endif
 
-      if (spu_extract(mask, 0))
-         spu.ctile.ui[iy][ix] = spu_pack_color_shuffle(colors[0], shuffle);
-      if (spu_extract(mask, 1))
-         spu.ctile.ui[iy][ix+1] = spu_pack_color_shuffle(colors[1], shuffle);
-      if (spu_extract(mask, 2))
-         spu.ctile.ui[iy+1][ix] = spu_pack_color_shuffle(colors[2], shuffle);
-      if (spu_extract(mask, 3))
-         spu.ctile.ui[iy+1][ix+1] = spu_pack_color_shuffle(colors[3], shuffle);
+      /* Read the current framebuffer values.
+       *
+       * Ignore read_fb for now.  In the future we can use this to avoid
+       * reading the framebuffer if read_fb is false and the fragment mask is
+       * all 0xffffffff.  This is the common case, so it is probably worth
+       * the effort.  We'll have to profile to determine whether or not the
+       * extra conditional branches hurt overall performance.
+       */
+      vec_float4 aos_pix[4] = {
+         spu_unpack_A8R8G8B8(spu.ctile.ui[iy+0][ix+0]),
+         spu_unpack_A8R8G8B8(spu.ctile.ui[iy+0][ix+1]),
+         spu_unpack_A8R8G8B8(spu.ctile.ui[iy+1][ix+0]),
+         spu_unpack_A8R8G8B8(spu.ctile.ui[iy+1][ix+1]),
+      };
 
-#if 0
-      /* SIMD_Z with swizzled color buffer (someday) */
-      vector unsigned int uicolors = *((vector unsigned int *) &colors);
-      spu.ctile.ui4[iy/2][ix/2] = spu_sel(spu.ctile.ui4[iy/2][ix/2], uicolors, mask);
-#endif
-   }
+      qword soa_pix[4];
+      qword soa_frag[4];
 
+      /* Convert pixel and fragment data from AoS to SoA format.
+       */
+      _transpose_matrix4x4((vec_float4 *) soa_pix, aos_pix);
+      _transpose_matrix4x4((vec_float4 *) soa_frag, colors);
+
+      const struct spu_blend_results result =
+          (*spu.blend)(soa_frag[0], soa_frag[1], soa_frag[2], soa_frag[3],
+                       soa_pix[0], soa_pix[1], soa_pix[2], soa_pix[3],
+                       (qword) mask);
+
+
+      /* Convert final pixel data from SoA to AoS format.
+       */
+      _transpose_matrix4x4(aos_pix, (const vec_float4 *) &result);
+
+      spu.ctile.ui[iy+0][ix+0] = spu_pack_color_shuffle(aos_pix[0], shuffle);
+      spu.ctile.ui[iy+0][ix+1] = spu_pack_color_shuffle(aos_pix[1], shuffle);
+      spu.ctile.ui[iy+1][ix+0] = spu_pack_color_shuffle(aos_pix[2], shuffle);
+      spu.ctile.ui[iy+1][ix+1] = spu_pack_color_shuffle(aos_pix[3], shuffle);
+   }
 #endif
 }