radv: prefetch VBO descriptors at the right place

Just after the vertex shader.

This seems to give a minor boost for, at least, Serious Sam
Fusion 2017 and Dawn of War 3. I don't see any real impacts
with The Talos Principle.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>

src/amd/vulkan/radv_cmd_buffer.c

@@ -668,6 +668,17 @@ radv_emit_prefetch_TC_L2_async(struct radv_cmd_buffer *cmd_buffer, uint64_t va,
 		si_cp_dma_prefetch(cmd_buffer, va, size);
 }
 
+static void
+radv_emit_VBO_descriptors_prefetch(struct radv_cmd_buffer *cmd_buffer)
+{
+	if (cmd_buffer->state.vb_prefetch_dirty) {
+		radv_emit_prefetch_TC_L2_async(cmd_buffer,
+					       cmd_buffer->state.vb_va,
+					       cmd_buffer->state.vb_size);
+		cmd_buffer->state.vb_prefetch_dirty = false;
+	}
+}
+
 static void
 radv_emit_shader_prefetch(struct radv_cmd_buffer *cmd_buffer,
 			  struct radv_shader_variant *shader)
@@ -691,6 +702,7 @@ radv_emit_prefetch(struct radv_cmd_buffer *cmd_buffer,
 {
 	radv_emit_shader_prefetch(cmd_buffer,
 				  pipeline->shaders[MESA_SHADER_VERTEX]);
+	radv_emit_VBO_descriptors_prefetch(cmd_buffer);
 	radv_emit_shader_prefetch(cmd_buffer,
 				  pipeline->shaders[MESA_SHADER_TESS_CTRL]);
 	radv_emit_shader_prefetch(cmd_buffer,
@@ -1814,6 +1826,10 @@ radv_cmd_buffer_update_vertex_descriptors(struct radv_cmd_buffer *cmd_buffer, bo
 
 		radv_emit_userdata_address(cmd_buffer, cmd_buffer->state.pipeline, MESA_SHADER_VERTEX,
 					   AC_UD_VS_VERTEX_BUFFERS, va);
+
+		cmd_buffer->state.vb_va = va;
+		cmd_buffer->state.vb_size = count * 16;
+		cmd_buffer->state.vb_prefetch_dirty = true;
 	}
 	cmd_buffer->state.vb_dirty = false;
 

src/amd/vulkan/radv_private.h

@@ -809,7 +809,12 @@ struct radv_attachment_state {
 };
 
 struct radv_cmd_state {
+	/* Vertex descriptors */
 	bool                                          vb_dirty;
+	bool                                          vb_prefetch_dirty;
+	uint64_t                                      vb_va;
+	unsigned                                      vb_size;
+
 	bool                                          push_descriptors_dirty;
 	bool predicating;
 	radv_cmd_dirty_mask_t                         dirty;