<div class="gmail_quote">On Tue, Feb 8, 2011 at 10:39 PM, Keith Whitwell <<a href="mailto:keithw@vmware.com" target="_blank">keithw@vmware.com</a>> wrote: <blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> <div>On Tue, 2011-02-08 at 22:29 +0100, Marek Olšák wrote: > > > Keith, > > redefine_user_buffer() would be a good compromise and I believe the > performance hit wouldn't be so noticable. It would also allow partial > uploads of constants in a user buffer, which is something we'd like to > have > too. > > Currently, st/mesa in st_draw_vbo is doing: > - N calls to resource_destroy > - some unnecessary computations > - N calls to user_buffer_create > - 1 call to set_vertex_buffers > - 1 call to set_vertex_elements_state > > If we can replace this by N calls to redefine_user_buffer, then I am > all for > it, provided neither _NEW_ARRAY nor _NEW_PROGRAM is dirty of course. > > Can the function look, let's say, like this? > > void redefine_user_buffer(struct pipe_context *, struct pipe_resource > *, > const struct pipe_box *); > </div>That looks good, but wouldn't you also want to be able to change the size of the userbuffer? I can see you might be able to implicitly grow the buffer this way (when box->x + box->width > resource->width0), but that's fairly obtuse and it doesn't permit shrinking. So perhaps as above with a "unsigned new_width0" parameter? </blockquote><div> </div></div>I had in mind that width0 would be ~0, but I guess that's not nice. What about this? void redefine_user_buffer( struct pipe_context*, struct pipe_resource*, unsigned offset, unsigned size); and new width0 would implicitly be offset+size; Marek