From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1753913Ab2ISHtY (ORCPT <rfc822;w@1wt.eu>);
	Wed, 19 Sep 2012 03:49:24 -0400
Received: from mx2.parallels.com ([64.131.90.16]:47468 "EHLO mx2.parallels.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1751790Ab2ISHtU (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Wed, 19 Sep 2012 03:49:20 -0400
Message-ID: <5059777E.8060906@parallels.com>
Date: Wed, 19 Sep 2012 11:42:54 +0400
From: Glauber Costa <glommer@parallels.com>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:15.0) Gecko/20120828 Thunderbird/15.0
MIME-Version: 1.0
To: Christoph Lameter <cl@linux.com>
CC: <linux-kernel@vger.kernel.org>, <cgroups@vger.kernel.org>,
        <kamezawa.hiroyu@jp.fujitsu.com>, <devel@openvz.org>,
        Tejun Heo <tj@kernel.org>, <linux-mm@kvack.org>,
        Suleiman Souhlal <suleiman@google.com>,
        Frederic Weisbecker <fweisbec@gmail.com>, Mel Gorman <mgorman@suse.de>,
        David Rientjes <rientjes@google.com>,
        Pekka Enberg <penberg@cs.helsinki.fi>
Subject: Re: [PATCH v3 09/16] sl[au]b: always get the cache from its page
 in kfree
References: <1347977530-29755-1-git-send-email-glommer@parallels.com> <1347977530-29755-10-git-send-email-glommer@parallels.com> <00000139d9fe8595-8905906d-18ed-4d41-afdb-f4c632c2d50a-000000@email.amazonses.com>
In-Reply-To: <00000139d9fe8595-8905906d-18ed-4d41-afdb-f4c632c2d50a-000000@email.amazonses.com>
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 09/18/2012 07:28 PM, Christoph Lameter wrote:
> On Tue, 18 Sep 2012, Glauber Costa wrote:
> 
>> index f2d760c..18de3f6 100644
>> --- a/mm/slab.c
>> +++ b/mm/slab.c
>> @@ -3938,9 +3938,12 @@ EXPORT_SYMBOL(__kmalloc);
>>   * Free an object which was previously allocated from this
>>   * cache.
>>   */
>> -void kmem_cache_free(struct kmem_cache *cachep, void *objp)
>> +void kmem_cache_free(struct kmem_cache *s, void *objp)
>>  {
>>  	unsigned long flags;
>> +	struct kmem_cache *cachep = virt_to_cache(objp);
>> +
>> +	VM_BUG_ON(!slab_equal_or_parent(cachep, s));
>>
> 
> This is an extremely hot path of the kernel and you are adding significant
> processing. Check how the benchmarks are influenced by this change.
> virt_to_cache can be a bit expensive.
> 

Would it be enough for you to have a separate code path for
!CONFIG_MEMCG_KMEM?

I don't really see another way to do it, aside from deriving the cache
from the object in our case. I am open to suggestions if you do.

>> diff --git a/mm/slub.c b/mm/slub.c
>> index 4778548..a045dfc 100644
>> --- a/mm/slub.c
>> +++ b/mm/slub.c
>> @@ -2604,7 +2604,9 @@ void kmem_cache_free(struct kmem_cache *s, void *x)
>>
>>  	page = virt_to_head_page(x);
>>
>> -	slab_free(s, page, x, _RET_IP_);
>> +	VM_BUG_ON(!slab_equal_or_parent(page->slab, s));
>> +
>> +	slab_free(page->slab, page, x, _RET_IP_);
>>
> 
> Less of a problem here but you are eroding one advantage that slab has had
> in the past over slub in terms of freeing objects.
> 
likewise.