Re: [Minios-devel] [UNIKRAFT PATCH v3 05/12] lib/uksched: Add support for waiting threads

[Florian Schmidt <Florian.Schmidt@xxxxxxxxx>]

Hi Costin, all,

actually, I forgot to raise an important point here:

On 3/10/19 9:09 PM, Costin Lupu wrote:
> +int uk_thread_wait(struct uk_thread *thread)
> +{
> +       UK_ASSERT(thread);
> +
> +       /* TODO critical region */
> +
> +       if (thread->detached)
> +               return -1;
> +
> +       uk_waitq_wait_event(&thread->waiting_threads, is_exited(thread));
> +
> +       thread->detached = true;
> +
> +       uk_sched_thread_destroy(thread->sched, thread);
> +
> +       return 0;
> +}
If more than one thread waits on the same other thread, then all of 
those waiting threads will wait in uk_waitq_event(). The first thread 
that wakes up after the thread has finished will then proceed to destroy 
the thread management structure. Every other waiting thread will try to 
do the same after waking up, ending up with duplicate free's and a crash 
of that thread.
That's not very nice.

I'm willing to let that slide for now, because the decision was that 
those thread management functions should align to behavior of the 
pthread API, and for pthread_join, the documentation explicitly says: 
"If multiple threads simultaneously try to join with the same thread, 
the results are undefined." So, technically, this is in line with the 
design of that function.
Nevertheless, I think we ought to revisit that soon, and either don't 
allow more than one thread to wait (confusingly, even those the 
specification of pthread_join speaks of undefined behavior, it specifies 
a return values of EINVAL for "another thread is already waiting to join 
with this thread"); or to make it so that only the last thrad waking up 
from the wait queue cleans up.
Cheers,
Florian

--
Dr. Florian Schmidt
フローリアン・シュミット
Research Scientist,
Systems and Machine Learning Group
NEC Laboratories Europe
Kurfürsten-Anlage 36, D-69115 Heidelberg
Tel.     +49 (0)6221 4342-265
Fax:     +49 (0)6221 4342-155
e-mail:  florian.schmidt@xxxxxxxxx
============================================================
Registered at Amtsgericht Mannheim, Germany, HRB728558

_______________________________________________
Minios-devel mailing list
Minios-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/minios-devel