Not resuming Virtual Threads?

[Alan Bateman]

On 03/07/2020 03:06, David Holmes wrote:
>
> A Thread sleeping forever does not allow the Thread or its Runnable or 
> any other reachable object to garbage-collected. Can you please 
> clarify what the lifecycle of a virtual thread is and how that relates 
> to its reachability?
>
> For reference a regular Thread is always logically reachable from its 
> ThreadGroup, until it terminates. And ThreadGroups with active threads 
> are reachable from their parent ThreadGroups. Hence all live threads 
> are always reachable.
A virtual Thread is just an object in the heap. It will always be 
reachable when mounted of course. If it parks waiting on a lock then 
it's the lock's waiter list that keeps the Thread reachable. It it parks 
in a blocking I/O operation then there is equivalent that keeps the 
Thread reachable so it can be unparked when the I/O operation can 
continue. If it sleeps then it will be scheduled to be unparked after 
the sleep time so that keeps the Thread reachable when sleeping. If 
there is code executing in a virtual thread that uses LockSupport.park() 
when it will be GC'ed if there is anything the Thread reachable.

Regular thread are GC roots in that their stacks are scanned for 
references. You are correct that the ThreadGroup also keeps Thread 
objects reachable too but we are on a mission to deprecate and 
eventually remove most of the TG API. This includes getting rid of the 
internal "threads" array. It will take a few releases and there will be 
interim stages where legacy APIs (such as "enumerate" methods) will work 
by reconstituting group membership from the thread list in the VM. JVMTI 
will do the same.

-Alan