In one of our internal tests we have found this deadlock:

0) Various threads are waiting for the GC to hand out freed memory.

1) The GC is blocked on the Service_lock:

  Mutex::lock_without_safepoint_check()
  OopMapCache::trigger_cleanup()
  ZGenerationYoung::pause_relocate_start()
  ZGenerationYoung::collect(ZYoungType, ConcurrentGCTimer*)
  ZDriverMinor::gc(ZDriverRequest const&)
  ZThread::run_service()

2) The Service_lock is held by the ServiceThread, which is stalling because it is waiting for the GC to give it memory

  PlatformMonitor::wait(unsigned long)
  ZRelocateQueue::add_and_wait(ZForwarding*)
  ZRelocate::relocate_object(ZForwarding*, zaddress_unsafe)
  ZUncoloredRootProcessWeakOopClosure::do_root(zaddress_unsafe*)
  ZNMethod::nmethod_oops_do_inner(nmethod*, OopClosure*)
  ZBarrierSetNMethod::nmethod_entry_barrier(nmethod*)
  JvmtiDeferredEventQueue::oops_do(OopClosure*, NMethodClosure*)
  // vvv Lock taken in this frame vvv
  ServiceThread::oops_do_no_frames(OopClosure*, NMethodClosure*)
  ZStackWatermark::process_head(void*)
  ZStackWatermark::start_processing_impl(void*)
  StackWatermark::start_processing()
  StackWatermark::on_safepoint()
  SafepointMechanism::process(JavaThread*, bool, bool)
  StringTable::clean_dead_entries(JavaThread*)
  StringTable::do_concurrent_work(JavaThread*)
  ServiceThread::service_thread_entry