Hi Eric, On 20 October 2016 at 00:09, Eric Engestrom <e...@engestrom.ch> wrote: > I'm not sure this is the right fix. The other thing I'm starting to lean > toward is to just remove these last two instruction (ie. everything after > `return VK_SUCCESS`). > Also, this is untested. It compiles, but that's it.
Indeed, that would be the right fix. Since Wayland is fully asynchronous, create_prime_buffer will either succeed, or generate a protocol error much further down the line. That being said, the use of wl_display_roundtrip is really broken in the presence of multiple threads: much like Jonas's fix for wayland-egl, all the Wayland objects created by the WSI _must_ use a separate event queue. Cheers, Daniel _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
