Now that this pesky little problem with futexes is (hopefully) dealt with; https://lkml.kernel.org/r/20170322103547.756091...@infradead.org
We can get on with fixing the actual bug this all started out with. These patches, started by Xunlei Pang, rework the PI infrastructure a bit fixing various problems it has, most notable a NULL deref in SCHED_DEADLINE. Once this is sorted; we can look at improving the tracing thing (Daniel Bristot and Julien Desfossez were working on that) and have a better foundation to look at bandwidth inheritance (Juri and co.).