From mboxrd@z Thu Jan  1 00:00:00 1970
From: Al Viro <viro@ZenIV.linux.org.uk>
Subject: Re: Potential data race in dput and __d_lookup
Date: Thu, 20 Nov 2014 18:05:03 +0000
Message-ID: <20141120180503.GN7996@ZenIV.linux.org.uk>
References: <CAAeHK+zQZ=sJYQA=BKLS2i7hu8NyXn6GXngvDqG9WD96yd9acg@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: Ingo Molnar <mingo@kernel.org>, Will Deacon <will.deacon@arm.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Davidlohr Bueso <davidlohr@hp.com>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	Dmitry Vyukov <dvyukov@google.com>,
	Kostya Serebryany <kcc@google.com>, ktsan@googlegroups.com
To: Andrey Konovalov <andreyknvl@google.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
Content-Disposition: inline
In-Reply-To: <CAAeHK+zQZ=sJYQA=BKLS2i7hu8NyXn6GXngvDqG9WD96yd9acg@mail.gmail.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

On Thu, Nov 20, 2014 at 09:20:14PM +0400, Andrey Konovalov wrote:

> It seems that one thread increments 'dentry->d_lockref.count', while
> other does 'lockref_put_or_lock(&dentry->d_lockref)'  without any
> synchronization.
> 
> Could you confirm if this is a real race?

I can not.  Increment in __d_lookup() is under ->d_lockref.lock (aka
->d_lock) and lockref_put_or_lock() is atomic wrt that.