We've found in GL that an actual end-of-pipe sync is required before invalidating the aux tables and that a simple CS stall is insufficient. If we're about to modify the actual AUX table entries from the GPU, we should definitely make sure it's stopped dead before we do so.
I have no proof yet that this actually fixes hangs. (I just started my first CI run with it). However, we've got some mystery hangs in CI and this seems like a thing that could cause significant grief.