vbo: optimize vertex copying when 'wrapping'

Instead of calling memcpy() 'n' times, we can do it all at once since
the source and dest regions are all contiguous.

Reviewed-by: Matt Turner <mattst88@gmail.com>

src/mesa/vbo/vbo_exec_api.c

 static void
 vbo_exec_vtx_wrap(struct vbo_exec_context *exec)
 {
-   fi_type *data = exec->vtx.copied.buffer;
-   GLuint i;
+   unsigned numComponents;
 
    /* Run pipeline on current vertices, copy wrapped vertices
     * to exec->vtx.copied.
     */
    assert(exec->vtx.max_vert - exec->vtx.vert_count > exec->vtx.copied.nr);
 
-   for (i = 0 ; i < exec->vtx.copied.nr ; i++) {
-      memcpy( exec->vtx.buffer_ptr, data, 
-	      exec->vtx.vertex_size * sizeof(GLfloat));
-      exec->vtx.buffer_ptr += exec->vtx.vertex_size;
-      data += exec->vtx.vertex_size;
-      exec->vtx.vert_count++;
-   }
+   numComponents = exec->vtx.copied.nr * exec->vtx.vertex_size;
+   memcpy(exec->vtx.buffer_ptr,
+          exec->vtx.copied.buffer,
+          numComponents * sizeof(fi_type));
+   exec->vtx.buffer_ptr += numComponents;
+   exec->vtx.vert_count += exec->vtx.copied.nr;
 
    exec->vtx.copied.nr = 0;
 }

src/mesa/vbo/vbo_save_api.c

 _save_wrap_filled_vertex(struct gl_context *ctx)
 {
    struct vbo_save_context *save = &vbo_context(ctx)->save;
-   fi_type *data = save->copied.buffer;
-   GLuint i;
+   unsigned numComponents;
 
    /* Emit a glEnd to close off the last vertex list.
     */
     */
    assert(save->max_vert - save->vert_count > save->copied.nr);
 
-   for (i = 0; i < save->copied.nr; i++) {
-      memcpy(save->buffer_ptr, data, save->vertex_size * sizeof(GLfloat));
-      data += save->vertex_size;
-      save->buffer_ptr += save->vertex_size;
-      save->vert_count++;
-   }
+   numComponents = save->copied.nr * save->vertex_size;
+   memcpy(save->buffer_ptr,
+          save->copied.buffer,
+          numComponents * sizeof(fi_type));
+   save->buffer_ptr += numComponents;
+   save->vert_count += save->copied.nr;
 }