[Bug 50292] Process-local synchronization objects use private interfaces into the Unix library

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=50292

Rémi Bernon  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #1 from Rémi Bernon  ---
Hi, I had a quick look at the patch series and here's a few nitpicks
that I've found:
> In 0004-ntdll-Implement-NtAlertThreadByThreadId-and-NtWaitFo.patch:
> -
>    if (teb->ClientId.UniqueThread == tid)
> -
>    {
> -
>        pthread_rwlock_unlock( &teb_list_lock );
> -
>        NtSetEvent( thread_data->tid_alert_event, NULL );
> -
>        return STATUS_SUCCESS;
> -
>    }

I think there's a race condition here, were the thread could potentially
be interrupted after the TEB lock is released, but before the event is
set.

The other thread that thread_data refers to may then terminate, the
NtSetEvent call may set an non-existing event, or worse if the TEB is
reused, and the new thread waiting itself, wake a wrong thread.

It's probably unlikely to happen but from a correctness point of view I
think it's wrong.