From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wi0-f181.google.com (mail-wi0-f181.google.com [209.85.212.181]) by dpdk.org (Postfix) with ESMTP id 320571F3 for ; Mon, 2 Sep 2013 15:24:48 +0200 (CEST) Received: by mail-wi0-f181.google.com with SMTP id c10so2006113wiw.8 for ; Mon, 02 Sep 2013 06:25:21 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:from:organization:to:subject:date:user-agent:cc :references:in-reply-to:mime-version:content-type :content-transfer-encoding:message-id; bh=11e0tKUSkGkF1KiMB4qgTQXn7EvP2yRJNFccWsd5ck0=; b=ZXqMbnhoyIlxRvbHa8wTmLVJoXDcr4+Z2dznuyPiWFrV8x5KYBlTqIcHRMLWOPqgnx xM1/sTL7i5EiqpPpjNvKuojxSSljzm48XFrdIlR9WCpl9clFdQoWqI3lm3H3gCBqMtld Q8zJxQ3OzkBuBW03BRfGlxtyhuXqcRR05ojq08bC/v2xHPafp3o3AVbkHMvSse8GR2C+ XOptV+K564aiL42OM9okI3edXu8tCgXarHiQ+tgGBkH0o87LCtEzb9UBwCH+XZAi+NQ9 Ow/AAQRLmgG6K5TZqHQDN1ZykEYMQSmtlLUpPUokXvPTSOciSnOkjQ2KANil+L0LyC28 caiQ== X-Gm-Message-State: ALoCoQmOiH1Sz44G9MScCpdS2qwSyk5DaDLKfTLO4ox0Yxsfu6PVRGoHxyVz5zNZ8jXtkxXNVF0B X-Received: by 10.194.202.230 with SMTP id kl6mr23606953wjc.9.1378128321774; Mon, 02 Sep 2013 06:25:21 -0700 (PDT) Received: from angus.localnet (6wind.net2.nerim.net. [213.41.180.237]) by mx.google.com with ESMTPSA id i12sm17778973wiw.3.1969.12.31.16.00.00 (version=TLSv1.2 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Mon, 02 Sep 2013 06:25:21 -0700 (PDT) From: Thomas Monjalon Organization: 6WIND To: Nulik Nol Date: Mon, 2 Sep 2013 15:25:19 +0200 User-Agent: KMail/1.13.7 (Linux/3.2.0-4-amd64; KDE/4.8.4; x86_64; ; ) References: In-Reply-To: MIME-Version: 1.0 Content-Type: Text/Plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Message-Id: <201309021525.19720.thomas.monjalon@6wind.com> Cc: dev@dpdk.org Subject: Re: [dpdk-dev] About interrupts and packet loss X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 02 Sep 2013 13:24:48 -0000 01/09/2013 09:01, Nulik Nol : > I have read in a pdf document about DPDK that the driver that comes > within the DPDK distribution was modified to disable interrupt > handling. This explains why DPDK is so fast. Yes, DPDK drivers poll the NICs. > However , this means the app in user space must constantly watch for > new packets, so the CPU load average will be almost 100% all the time, > correct? Well, then what happens if the app executes a busy process > and at that time all RX rings get filled, will the incoming network > traffic be lost? It's not specific to DPDK. If CPU don't read packets fast enough in the NIC, the buffer will be full, and yes, some packets will be lost. By using DPDK with the right number of CPUs, you can achieve high rates without loss. -- Thomas