https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69347
--- Comment #4 from Jeffrey A. Law <law at redhat dot com> --- We've got function with something like 15k blocks. One particular block as ~3k predecessors. Clearly something in the FSM bits isn't scaling. I've already fixed one bug in this space a few months ago and I'm not surprised we're finding more now that we're using the FSM threader more aggressively. Still investigating.
