https://bugs.freedesktop.org/show_bug.cgi?id=91750
--- Comment #5 from Pekka Paalanen <[email protected]> --- (In reply to Jonas Ådahl from comment #4) > The difference to your example below is the wl_display_prepare_read(display) > before creating the proxy, and wl_display_cancel_read(display) after setting > the queue. The effect is that the other thread will not dispatch it, since > it will only the dispatch events on the queue associated with it. > > Now, wl_display_roundtrip_queue() is a function that currently exposes the > bug, but I'm not sure what unforseen side effects would be to add a > work-around like the above, and whether it is a good idea to do it. The > obvious side effect would be stalling other dispatcher threads a bit more > than before. > > Pekka, any opinions whether we should temporarly "fix" > wl_display_roundtrip_queue() until we introduce some new API? It would only fix the roundtrip and none of the other problems. IMHO it would be a wasted effort, time would be better spent solving the underlying problem. Using wl_display_prepare_read() is problematic, because it requires the wl_event_queue to be empty. You need a dispatch loop to drain the wl_event_queue until prepare_read() succeeds. Maybe that is not *that* bad since wl_display_roundtrip_queue() *is* a dispatching function. However, wl_display_roundtrip_queue() explicitly documents itself to be incompatible with prepare_read() because roundtrip_queue() uses wl_display_dispatch_queue(). I really do not want to review whether we could break that rule in this specific case or not. Because of the above, using wl_display_roundtrip{,_queue}() in a program using the same wl_display from multiple threads is a mistake to begin with. All this makes me wonder if we could rewrite the roundtrip functions on top the prepare_read() API, and just deprecate wl_display_dispatch{,_queue}() completely. Multi-threaded programs using wl_display_dispatch{,_queue}() are broken to start with, and single-threaded programs using wl_display_dispatch{,_queue}() cannot use the mutually exclusive APIs at the same time by definition. Yes, Mesa EGL Wayland platform implementation is broken. It uses wl_display_dispatch_queue(). -- You are receiving this mail because: You are the assignee for the bug.
_______________________________________________ wayland-bugs mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/wayland-bugs
