Re: [PATCH -v5 14/14] futex: futex_unlock_pi() determinism

[Peter Zijlstra]

On Mon, Mar 13, 2017 at 03:25:52PM +0100, Thomas Gleixner wrote:
> On Mon, 13 Mar 2017, Peter Zijlstra wrote:
> 
> > On Tue, Mar 07, 2017 at 03:31:50PM +0100, Thomas Gleixner wrote:
> > > On Sat, 4 Mar 2017, Peter Zijlstra wrote:
> > > 
> > > > The problem with returning -EAGAIN when the waiter state mismatches is
> > > > that it becomes very hard to proof a bounded execution time on the
> > > > operation. And seeing that this is a RT operation, this is somewhat
> > > > important.
> > > > 
> > > > While in practise it will be very unlikely to ever really take more
> > > > than one or two rounds, proving so becomes rather hard.
> > > 
> > > Oh no. Assume the following:
> > > 
> > > T1 and T2 are both pinned to CPU0. prio(T2) > prio(T1)
> > > 
> > > CPU0
> > > 
> > > T1 
> > >   lock_pi()
> > >   queue_me()  <- Waiter is visible
> > > 
> > > preemption
> > > 
> > > T2
> > >   unlock_pi()
> > >     loops with -EAGAIN forever
> > 
> > So this is true before the last patch; but if we look at the locking
> > changes brought by that (pasting its changelog here):
> 
> I was referring to the state before the last patch and your wording in the
> changelog of this being very unlikely.

Yeah, I understand that. Lemme see what I can do to clarify both
situations.