From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-lb0-x22b.google.com (mail-lb0-x22b.google.com [IPv6:2a00:1450:4010:c04::22b]) by dpdk.org (Postfix) with ESMTP id 1F1812F4 for ; Wed, 22 Jan 2014 04:48:22 +0100 (CET) Received: by mail-lb0-f171.google.com with SMTP id c11so6561080lbj.2 for ; Tue, 21 Jan 2014 19:49:39 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:date:message-id:subject:from:to:content-type; bh=tHWjDpkBfPGvfZd6Hutx4Dj35riAw5B+dqrV/np/KGw=; b=j18twnVF31EF7X/c0mmYDRYTboFg642QxunnHdfih7oBxNffIHuRNn4cjK4/mstN5p LspDpZ2Avmeuhh41TjdRKb6bNdsSPwWDvq4+yceeIWcHnoe3R8nVKJHLQPRVPHaSbBX5 YM/aDnqYA+af9jHNWutdq/cnr/bFyXuyVpKlMlRny63H/eNKFgagW0HW/s8MKfhkU/s5 0w9diBHgZ2K/a979SRHGp2fGmigdhVnmjFB/oacM1oIoR0Vz3VpfK66pA1XolLkxgW6d m+VEYC4DA+WQN+TAHWbhA0xZ5b6oXEEK7UOMpekHJa1j6Jq13Twy0FM3pSGLnHYW7SF7 p/fQ== MIME-Version: 1.0 X-Received: by 10.112.188.135 with SMTP id ga7mr18037510lbc.5.1390362579076; Tue, 21 Jan 2014 19:49:39 -0800 (PST) Received: by 10.112.143.231 with HTTP; Tue, 21 Jan 2014 19:49:39 -0800 (PST) Date: Wed, 22 Jan 2014 11:49:39 +0800 Message-ID: From: Chao Pei To: dev Content-Type: text/plain; charset=ISO-8859-1 X-Content-Filtered-By: Mailman/MimeDel 2.1.15 Subject: [dpdk-dev] KNI with multiple rxq/freeqs 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: Wed, 22 Jan 2014 03:48:23 -0000 Hi, Currently we are using run-to-complete model with multiple workers in our project. there is also a KNI running with a fixed ip address. However, a packet with a destination IP of the KNI may go to any worker that owns a rxq of the physical NIC(without Flow Director). So all the workers have to enqueue the packets to the same rxq of the KNI. Which may cause a performance issue. I think it's necessary to have multiple rxq/freeqs for KNI, so that each worker can have its own rxq/freeq. As for txq/allocq, it does not have the issue I mentioned above.