From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+willy=40w.ods.org-S1423591AbWKFJWz@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1423591AbWKFJWz (ORCPT <rfc822;willy@w.ods.org>);
	Mon, 6 Nov 2006 04:22:55 -0500
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1423561AbWKFJWz
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Mon, 6 Nov 2006 04:22:55 -0500
Received: from pfx2.jmh.fr ([194.153.89.55]:4053 "EHLO pfx2.jmh.fr")
	by vger.kernel.org with ESMTP id S1422805AbWKFJWy (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Mon, 6 Nov 2006 04:22:54 -0500
From: Eric Dumazet <dada1@cosmosbay.com>
To: "Zhao Xiaoming" <xiaoming.nj@gmail.com>
Subject: Re: ZONE_NORMAL memory exhausted by 4000 TCP sockets
Date: Mon, 6 Nov 2006 10:22:57 +0100
User-Agent: KMail/1.9.5
Cc: linux-kernel@vger.kernel.org, "Linux Netdev List" <netdev@vger.kernel.org>
References: <f55850a70611052207j384e1d3flaf40bb9dd74df7c5@mail.gmail.com> <454EE580.5040506@cosmosbay.com> <f55850a70611060059p28d6baeco9f6f0c0f81b54ba6@mail.gmail.com>
In-Reply-To: <f55850a70611060059p28d6baeco9f6f0c0f81b54ba6@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200611061022.57840.dada1@cosmosbay.com>
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

On Monday 06 November 2006 09:59, Zhao Xiaoming wrote:

> Thank you again for your help. To have more detailed statistic data, I
> did another round of test and gathered some data.  I give the overall
> description here and detailed /proc/net/sockstat, /proc/meminfo,
> /proc/slabinfo and /proc/buddyinfo follows.
> =====================================================
>                            slab mem cost        tcp mem pages       lowmem
> free with traffic:             254668KB                 34693
>       38772KB
> without traffic:       104080KB                           1
>        702652KB
> =====================================================

Thank you for detailed infos.

It appears you have an extensive use of threads (about 10000), since :

> task_struct        10095  10095   1360    3    1 : tunables   24   12
>   8 : slabdata   3365   3365      0

Each thread has a kernel stack, 8KB (ie 2 pages, order-1 allocation), plus a 
user vma

> vm_area_struct     21346  21504     92   42    1 : tunables  120   60
>   8 : slabdata    512    512      0

Most likely you dont need that much threads. A program with fewer threads will 
perform better and use less ram.