From owner-soc-status@FreeBSD.ORG Fri Jul 30 16:40:39 2010 Return-Path: Delivered-To: soc-status@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 198EF106567D for ; Fri, 30 Jul 2010 16:40:39 +0000 (UTC) (envelope-from alexandre@net.t-labs.tu-berlin.de) Received: from mail.net.t-labs.tu-berlin.de (mail.net.t-labs.tu-berlin.de [130.149.220.252]) by mx1.freebsd.org (Postfix) with ESMTP id D04B18FC1F for ; Fri, 30 Jul 2010 16:40:38 +0000 (UTC) Received: from raven.net.t-labs.tu-berlin.de (raven.net.t-labs.tu-berlin.de [130.149.220.18]) by mail.net.t-labs.tu-berlin.de (Postfix) with ESMTP id AF5B6700D46E for ; Fri, 30 Jul 2010 18:40:37 +0200 (CEST) From: Alexandre Fiveg Organization: t-net To: soc-status@freebsd.org Date: Fri, 30 Jul 2010 18:40:36 +0200 User-Agent: KMail/1.9.10 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <201007301840.36736.alexandre@net.t-labs.tu-berlin.de> Subject: ringmap #8,#9 X-BeenThere: soc-status@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Summer of Code Status Reports and Discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 30 Jul 2010 16:40:39 -0000 Ringmap is ported to the 10Gb (integrated with ixgbe driver) with the following restrictions: - Only one queue will be used. All the traffic will be pushed into the only one predefined queue. Currently, there is no interface to the user-space for managing multi-queues in the ixgbe driver. That is why it is problematic to use multiple queues in the ringmap. I would like to implement this interface, but due to the lack of the time until the end of GSoC-2010 it seems to be impossible (there is a lot of work that I should manage until the end of GSOC: debugging, testing, evaluation, documentation, etc...). - The packet-filtering doesn't work yet. All the packets will be captured by ringmap. I will set filtering in user-space using libpcap-BPF. But I think, ringmap should be able to filter the packets also in the kernel using kernel-BPF. - The coherence-problem is still relevant. The kernel and user-thread see sometimes different data in the same mapped memory buffer.