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:   Tue, 10 Apr 2018 15:21:45 -0500 (CDT)
From:   Christopher Lameter <cl@linux.com>
To:     Matthew Wilcox <willy@infradead.org>
cc:     Eric Dumazet <eric.dumazet@gmail.com>, linux-mm@kvack.org,
        Matthew Wilcox <mawilcox@microsoft.com>,
        Pekka Enberg <penberg@kernel.org>,
        David Rientjes <rientjes@google.com>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, Jan Kara <jack@suse.cz>,
        Jeff Layton <jlayton@redhat.com>,
        Mel Gorman <mgorman@techsingularity.net>,
        stable@vger.kernel.org
Subject: Re: [PATCH 1/2] slab: __GFP_ZERO is incompatible with a
 constructor
In-Reply-To: <20180410175011.GE3614@bombadil.infradead.org>
Message-ID: <alpine.DEB.2.20.1804101512350.30225@nuc-kabylake>
References: <20180410125351.15837-1-willy@infradead.org> <fee8a8bc-3db5-a66a-33cb-0729143ba615@gmail.com> <20180410165054.GC3614@bombadil.infradead.org> <alpine.DEB.2.20.1804101228170.29384@nuc-kabylake> <20180410173841.GD3614@bombadil.infradead.org>
 <alpine.DEB.2.20.1804101244290.29559@nuc-kabylake> <20180410175011.GE3614@bombadil.infradead.org>
User-Agent: Alpine 2.20 (DEB 67 2015-01-07)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, 10 Apr 2018, Matthew Wilcox wrote:

> > Objects can be freed and reused and still be accessed from code that
> > thinks the object is the old and not the new object....
>
> Yes, I know, that's the point of RCU typesafety.  My point is that an
> object *which has never been used* can't be accessed.  So you don't *need*
> a constructor.

But the object needs to have the proper contents after it was released and
re-allocated. Some objects may rely on contents (like list heads)
surviving the realloc process because access must always be possible.

validate_slab() checks on proper metadata content in a slab
although it does not access the payload. So that may work you separate
the payload init from the metadata init.