Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 21 Jan 2019 15:44:46 +0100
From:   Andrea Parri <andrea.parri@amarulasolutions.com>
To:     Dmitry Vyukov <dvyukov@google.com>
Cc:     Elena Reshetova <elena.reshetova@intel.com>,
        Kees Cook <keescook@chromium.org>,
        Peter Zijlstra <peterz@infradead.org>,
        Alan Stern <stern@rowland.harvard.edu>,
        "Paul E. McKenney" <paulmck@linux.ibm.com>,
        Will Deacon <will.deacon@arm.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Andrey Ryabinin <aryabinin@virtuozzo.com>,
        Anders Roxell <anders.roxell@linaro.org>,
        Mark Rutland <mark.rutland@arm.com>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] kcov: convert kcov.refcount to refcount_t
Message-ID: <20190121144446.GA8926@andrea>
References: <1547634429-772-1-git-send-email-elena.reshetova@intel.com>
 <CACT4Y+bWgcS=c=KrthWyyjjBpc72DEK-=czLYK8=SkmOsZ_-jg@mail.gmail.com>
 <CACT4Y+a6-YyrwZ9qX+yhhMuoPoZ5=B-NKxkn-EyFqijNBf4N+g@mail.gmail.com>
 <20190121114505.GA7307@andrea>
 <CACT4Y+aF8+RqEk8RJn3G4nN8bCAxuci5W9YXa=cFq4c-QC3kgg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CACT4Y+aF8+RqEk8RJn3G4nN8bCAxuci5W9YXa=cFq4c-QC3kgg@mail.gmail.com>
User-Agent: Mutt/1.9.4 (2018-02-28)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jan 21, 2019 at 01:29:11PM +0100, Dmitry Vyukov wrote:
>  On Mon, Jan 21, 2019 at 12:45 PM Andrea Parri
> <andrea.parri@amarulasolutions.com> wrote:
> >
> > On Mon, Jan 21, 2019 at 10:52:37AM +0100, Dmitry Vyukov wrote:
> >
> > [...]
> >
> > > > Am I missing something or refcount_dec_and_test does not in fact
> > > > provide ACQUIRE ordering?
> > > >
> > > > +case 5) - decrement-based RMW ops that return a value
> > > > +-----------------------------------------------------
> > > > +
> > > > +Function changes:
> > > > +                atomic_dec_and_test() --> refcount_dec_and_test()
> > > > +                atomic_sub_and_test() --> refcount_sub_and_test()
> > > > +                no atomic counterpart --> refcount_dec_if_one()
> > > > +                atomic_add_unless(&var, -1, 1) --> refcount_dec_not_one(&var)
> > > > +
> > > > +Memory ordering guarantees changes:
> > > > +                fully ordered --> RELEASE ordering + control dependency
> > > >
> > > > I think that's against the expected refcount guarantees. When I
> > > > privatize an  atomic_dec_and_test I would expect that not only stores,
> > > > but also loads act on a quiescent object. But loads can hoist outside
> > > > of the control dependency.
> > > >
> > > > Consider the following example, is it the case that the BUG_ON can still fire?
> >
> > Can't it fire in an SC world? (wrong example, or a Monday morning? ;D)
> 
> I don't see how. Maybe there is a stupid off-by-one, but overall
> that's the example I wanted to show. refcount is 2, each thread sets
> own done flag, drops refcount, last thread checks done flag of the
> other thread.

You're right: looking at the example again, I think that the BUG_ON()
in your example can indeed trigger with a CTRL+RELEASE semantics (but
_not with the fully-ordered semantics).

I apologize for the confusion (it must have been _my Monday...  ;-/).

  Andrea


> 
> 
> 
> > > > struct X {
> > > >   refcount_t rc; // == 2
> > > >   int done1, done2; // == 0
> > > > };
> > > >
> > > > // thread 1:
> > > > x->done1 = 1;
> > > > if (refcount_dec_and_test(&x->rc))
> > > >   BUG_ON(!x->done2);
> > > >
> > > > // thread 2:
> > > > x->done2 = 1;
> > > > if (refcount_dec_and_test(&x->rc))
> > > >   BUG_ON(!x->done1);