i965: use pack/unpackDouble lowering

[mesa.git] / src / gallium / drivers / svga / svga_state_constants.c

diff --git a/src/gallium/drivers/svga/svga_state_constants.c b/src/gallium/drivers/svga/svga_state_constants.c
index b6d6de0dca329fd486244d4cb3c1dcd42872c9e5..8784f47413f12c254e80ce883907831cb8558e41 100644 (file)
--- a/src/gallium/drivers/svga/svga_state_constants.c
+++ b/src/gallium/drivers/svga/svga_state_constants.c
@@ -301,6 +301,8 @@ emit_const(struct svga_context *svga, unsigned shader, unsigned i,
          return ret;
 
       memcpy(svga->state.hw_draw.cb[shader][i], value, 4 * sizeof(float));
+
+      svga->hud.num_const_updates++;
    }
 
    return ret;
@@ -420,6 +422,9 @@ emit_const_range(struct svga_context *svga,
                 (j - i) * 4 * sizeof(float));
 
          i = j + 1;
+
+         svga->hud.num_const_updates++;
+
       } else {
          ++i;
       }
@@ -459,7 +464,7 @@ emit_consts_vgpu9(struct svga_context *svga, unsigned shader)
       data = (const float (*)[4])
          pipe_buffer_map(&svga->pipe, svga->curr.constbufs[shader][0].buffer,
                          PIPE_TRANSFER_READ, &transfer);
-      if (data == NULL) {
+      if (!data) {
          return PIPE_ERROR_OUT_OF_MEMORY;
       }
 
@@ -513,7 +518,7 @@ emit_consts_vgpu9(struct svga_context *svga, unsigned shader)
 
       assert(variant);
       offset = variant->shader->info.file_max[TGSI_FILE_CONSTANT] + 1;
-      assert(count <= Elements(extras));
+      assert(count <= ARRAY_SIZE(extras));
 
       if (count > 0) {
          if (ss->hw_version >= SVGA3D_HWVERSION_WS8_B1) {
@@ -549,6 +554,7 @@ emit_constbuf_vgpu10(struct svga_context *svga, unsigned shader)
    void *src_map = NULL, *dst_map;
    unsigned offset;
    const struct svga_shader_variant *variant;
+   unsigned alloc_buf_size;
 
    assert(shader == PIPE_SHADER_VERTEX ||
           shader == PIPE_SHADER_GEOMETRY ||
@@ -613,7 +619,17 @@ emit_constbuf_vgpu10(struct svga_context *svga, unsigned shader)
     */
    new_buf_size = align(new_buf_size, 16);
 
-   u_upload_alloc(svga->const0_upload, 0, new_buf_size, &offset,
+   /* Constant buffer size in the upload buffer must be in multiples of 256.
+    * In order to maximize the chance of merging the upload buffer chunks
+    * when svga_buffer_add_range() is called,
+    * the allocate buffer size needs to be in multiples of 256 as well.
+    * Otherwise, since there is gap between each dirty range of the upload buffer,
+    * each dirty range will end up in its own UPDATE_GB_IMAGE command.
+    */
+   alloc_buf_size = align(new_buf_size, CONST0_UPLOAD_ALIGNMENT);
+
+   u_upload_alloc(svga->const0_upload, 0, alloc_buf_size,
+                  CONST0_UPLOAD_ALIGNMENT, &offset,
                   &dst_buffer, &dst_map);
    if (!dst_map) {
       if (src_map)
@@ -663,6 +679,8 @@ emit_constbuf_vgpu10(struct svga_context *svga, unsigned shader)
 
    pipe_resource_reference(&dst_buffer, NULL);
 
+   svga->hud.num_const_buf_updates++;
+
    return ret;
 }
 
@@ -704,6 +722,24 @@ emit_consts_vgpu10(struct svga_context *svga, unsigned shader)
          assert(size == 0);
       }
 
+      if (size % 16 != 0) {
+         /* GL's buffer range sizes can be any number of bytes but the
+          * SVGA3D device requires a multiple of 16 bytes.
+          */
+         const unsigned total_size = buffer->b.b.width0;
+
+         if (offset + align(size, 16) <= total_size) {
+            /* round up size to multiple of 16 */
+            size = align(size, 16);
+         }
+         else {
+            /* round down to mulitple of 16 (this may cause rendering problems
+             * but should avoid a device error).
+             */
+            size &= ~15;
+         }
+      }
+
       assert(size % 16 == 0);
       ret = SVGA3D_vgpu10_SetSingleConstantBuffer(svga->swc,
                                                   index,
@@ -713,6 +749,8 @@ emit_consts_vgpu10(struct svga_context *svga, unsigned shader)
                                                   size);
       if (ret != PIPE_OK)
          return ret;
+
+      svga->hud.num_const_buf_updates++;
    }
 
    svga->state.hw_draw.enabled_constbufs[shader] = enabled_constbufs;
@@ -729,7 +767,7 @@ emit_fs_consts(struct svga_context *svga, unsigned dirty)
 
    /* SVGA_NEW_FS_VARIANT
     */
-   if (variant == NULL)
+   if (!variant)
       return PIPE_OK;
 
    /* SVGA_NEW_FS_CONST_BUFFER
@@ -750,7 +788,7 @@ struct svga_tracked_state svga_hw_fs_constants =
    "hw fs params",
    (SVGA_NEW_FS_CONST_BUFFER |
     SVGA_NEW_FS_VARIANT |
-    SVGA_NEW_TEXTURE_BINDING),
+    SVGA_NEW_TEXTURE_CONSTS),
    emit_fs_consts
 };
 
@@ -764,7 +802,7 @@ emit_vs_consts(struct svga_context *svga, unsigned dirty)
 
    /* SVGA_NEW_VS_VARIANT
     */
-   if (variant == NULL)
+   if (!variant)
       return PIPE_OK;
 
    /* SVGA_NEW_VS_CONST_BUFFER
@@ -798,7 +836,7 @@ emit_gs_consts(struct svga_context *svga, unsigned dirty)
 
    /* SVGA_NEW_GS_VARIANT
     */
-   if (variant == NULL)
+   if (!variant)
       return PIPE_OK;
 
    /* SVGA_NEW_GS_CONST_BUFFER