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 08:54:42 -0700
From:   Matthew Wilcox <willy@infradead.org>
To:     Christopher Lameter <cl@linux.com>
Cc:     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
Message-ID: <20180410155442.GA3614@bombadil.infradead.org>
References: <20180410125351.15837-1-willy@infradead.org>
 <alpine.DEB.2.20.1804100920110.27333@nuc-kabylake>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <alpine.DEB.2.20.1804100920110.27333@nuc-kabylake>
User-Agent: Mutt/1.9.2 (2017-12-15)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Apr 10, 2018 at 09:21:20AM -0500, Christopher Lameter wrote:
> On Tue, 10 Apr 2018, Matthew Wilcox wrote:
> 
> > __GFP_ZERO requests that the object be initialised to all-zeroes,
> > while the purpose of a constructor is to initialise an object to a
> > particular pattern.  We cannot do both.  Add a warning to catch any
> > users who mistakenly pass a __GFP_ZERO flag when allocating a slab with
> > a constructor.
> 
> Can we move this check out of the critical paths and check for
> a ctor and GFP_ZERO when calling the page allocator? F.e. in
> allocate_slab()?

Are you willing to have this kind of bug go uncaught for a while?
In this specific case, __GFP_ZERO was only being passed on a few of the
calls to kmem_cache_alloc.  So we'd happily trash the constructed object
any time we didn't allocate a page.

I appreciate it's a tradeoff, and we don't want to clutter the critical
path unnecessarily.