On Thu, Sep 01, 2016 at 07:26:58PM +0200, Oleg Nesterov wrote: > On 09/01, Peter Zijlstra wrote: > > > > So mixing INTERRUPTIBLE and UNINTERRUPTIBLE and then not using > > TASK_NORMAL for wakeups is a mis-feature/abuse of waitqueues IMO. > > Heh, agreed. When I was doing this fix I suddenly realize that I do > not understand why do we have, say, wake_up_interruptible(). > > I mean, I can't imagine the "real" use-case when you actually want > to wake up only the INTERRUPTIBLE tasks and leave the UNINTERRUPTIBLE > sleeping. Exclusive or not. > > It seems that wake_up_interruptible() is mostly used simply because > the caller knows that UNINTERRUPTIBLE waiters are not possible, this > is often the case.
I suspect the same. > Actually, I think that TASK_NORMAL should be used even if wq mixes > UNINTERRUPTIBLE and KILLABLE waiters. The fact that TASK_KILLABLE > includes TASK_UNINTERRUPTIBLE is just "implementation detail" even > if I do not think this will be ever changed. I think its a fairly fundamental thing, its part of the semantics of TASK_KILLABLE. Namely its UNINTERRUPTIBLE, except you can interrupt it with fatal. A TASK_NORMAL wake should very much wake a TASK_KILLABLE sleep, and for that to happen they need to share a bit, namely UNINTERRUPTIBLE. But I agree with you, all waitqueue wakeups _should_ simply be TASK_NORMAL. Like said, I don't see it ever makes sense to play games with it. Now, let me try and get back to making sense of your abort abortion ;-)
