MIME-Version: 1.0
In-Reply-To: <507DA245.9050709@am.sony.com>
References: <CALF0-+XGn5=QSE0bpa4RTag9CAJ63MKz1kvaYbpw34qUhViaZA@mail.gmail.com>
	<m27gqwtyu9.fsf@firstfloor.org>
	<alpine.DEB.2.00.1210111558290.6409@chino.kir.corp.google.com>
	<m2391ktxjj.fsf@firstfloor.org>
	<CALF0-+WLZWtwYY4taYW9D7j-abCJeY90JzcTQ2hGK64ftWsdxw@mail.gmail.com>
	<alpine.DEB.2.00.1210130252030.7462@chino.kir.corp.google.com>
	<CALF0-+Xp_P_NjZpifzDSWxz=aBzy_fwaTB3poGLEJA8yBPQb_Q@mail.gmail.com>
	<alpine.DEB.2.00.1210151745400.31712@chino.kir.corp.google.com>
	<CALF0-+WgfnNOOZwj+WLB397cgGX7YhNuoPXAK5E0DZ5v_BxxEA@mail.gmail.com>
	<1350392160.3954.986.camel@edumazet-glaptop>
	<507DA245.9050709@am.sony.com>
Date: Tue, 16 Oct 2012 15:36:44 -0300
Message-ID: <CALF0-+XLXAh3=OScQ=V0F80ZcnTGjHox68SApOwPUYVvmjdqPw@mail.gmail.com>
Subject: Re: [Q] Default SLAB allocator
From: Ezequiel Garcia <elezegarcia@gmail.com>
To: Tim Bird <tim.bird@am.sony.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>,
        David Rientjes <rientjes@google.com>, Andi Kleen <andi@firstfloor.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "linux-mm@kvack.org" <linux-mm@kvack.org>,
        "celinux-dev@lists.celinuxforum.org" 
	<celinux-dev@lists.celinuxforum.org>
Content-Type: text/plain; charset=ISO-8859-1
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1343
Lines: 33

On Tue, Oct 16, 2012 at 3:07 PM, Tim Bird <tim.bird@am.sony.com> wrote:
> On 10/16/2012 05:56 AM, Eric Dumazet wrote:
>> On Tue, 2012-10-16 at 09:35 -0300, Ezequiel Garcia wrote:
>>
>>> Now, returning to the fragmentation. The problem with SLAB is that
>>> its smaller cache available for kmalloced objects is 32 bytes;
>>> while SLUB allows 8, 16, 24 ...
>>>
>>> Perhaps adding smaller caches to SLAB might make sense?
>>> Is there any strong reason for NOT doing this?
>>
>> I would remove small kmalloc-XX caches, as sharing a cache line
>> is sometime dangerous for performance, because of false sharing.
>>
>> They make sense only for very small hosts.
>
> That's interesting...
>
> It would be good to measure the performance/size tradeoff here.
> I'm interested in very small systems, and it might be worth
> the tradeoff, depending on how bad the performance is.  Maybe
> a new config option would be useful (I can hear the groans now... :-)
>

It might be worth reminding that very small systems can use SLOB
allocator, which does not suffer from this kind of fragmentation.

    Ezequiel
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/