From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga05.intel.com (mga05.intel.com [192.55.52.43]) by dpdk.org (Postfix) with ESMTP id 0DD1E2A5E for ; Mon, 9 Oct 2017 04:39:48 +0200 (CEST) Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga105.fm.intel.com with ESMTP; 08 Oct 2017 19:39:47 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.42,498,1500966000"; d="scan'208";a="908005992" Received: from fyigit-mobl1.ger.corp.intel.com (HELO [10.254.114.24]) ([10.254.114.24]) by FMSMGA003.fm.intel.com with ESMTP; 08 Oct 2017 19:39:47 -0700 To: Yongseok Koh , adrien.mazarguil@6wind.com, nelio.laranjeiro@6wind.com Cc: dev@dpdk.org, Raslan Darawsheh , Shahaf Shuler , Xueming Li References: <20171005230032.7548-1-yskoh@mellanox.com> From: Ferruh Yigit Message-ID: <51bda660-2653-1416-539d-1333e4f0da04@intel.com> Date: Mon, 9 Oct 2017 03:39:47 +0100 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.3.0 MIME-Version: 1.0 In-Reply-To: <20171005230032.7548-1-yskoh@mellanox.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 8bit Subject: Re: [dpdk-dev] [PATCH v1 0/7] net/mlx5: add vectorized Rx/Tx burst for ARM X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 09 Oct 2017 02:39:49 -0000 On 10/6/2017 12:00 AM, Yongseok Koh wrote: > Add dataplane functions using ARM NEON instructions. To modularize vectorized > functions for different architectures, the existing files having x86 SSE support > is reorganized. > > This patchset has dependency with the following patches: > - net/mlx5: fix overflow of Rx SW ring > - Nelio's mlx5 flow cleanup patchset, the last one is: > [dpdk-dev,v2,30/30] net/mlx5: add new operations for isolated mode - Nelio's set was waiting Xueming's which is merged now. - Nelio's set will have v3 because of your comments. - And Raslan's patch is dependent on yours. So, if I don't miss anything - first there will be Nelio's v3 - Later this patchset on top of it - Later Raslan's on top of yours With multiple developers working on same drivers, mlx drivers become hard to manage/trace. next-net-mlx sound like good idea :) Any comment on this? > > Yongseok Koh (7): > net/mlx5: cleanup memory barriers > net/mlx5: rename a file of SSE Rx/Tx > net/mlx5: use static assert for compile-time sanity checks > net/mlx5: separate shareable vector functions > net/mlx5: match Rx completion entry size to cacheline > net/mlx5: fix configuration of Rx CQE compression > net/mlx5: add vectorized Rx/Tx burst for ARM > > drivers/net/mlx5/Makefile | 10 +- > drivers/net/mlx5/mlx5.c | 19 +- > drivers/net/mlx5/mlx5_rxq.c | 20 +- > drivers/net/mlx5/mlx5_rxtx.c | 4 +- > drivers/net/mlx5/mlx5_rxtx.h | 2 +- > drivers/net/mlx5/mlx5_rxtx_vec.c | 388 ++++++++ > drivers/net/mlx5/mlx5_rxtx_vec.h | 126 +++ > drivers/net/mlx5/mlx5_rxtx_vec_neon.h | 1028 ++++++++++++++++++++ > .../{mlx5_rxtx_vec_sse.c => mlx5_rxtx_vec_sse.h} | 414 +------- > 9 files changed, 1597 insertions(+), 414 deletions(-) > create mode 100644 drivers/net/mlx5/mlx5_rxtx_vec.c > create mode 100644 drivers/net/mlx5/mlx5_rxtx_vec.h > create mode 100644 drivers/net/mlx5/mlx5_rxtx_vec_neon.h > rename drivers/net/mlx5/{mlx5_rxtx_vec_sse.c => mlx5_rxtx_vec_sse.h} (76%) <...>