From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by dpdk.space (Postfix) with ESMTP id 7EFA6A045E for ; Tue, 28 May 2019 11:40:14 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 7557B1B05; Tue, 28 May 2019 11:40:13 +0200 (CEST) Received: from mga17.intel.com (mga17.intel.com [192.55.52.151]) by dpdk.org (Postfix) with ESMTP id 17496137D for ; Tue, 28 May 2019 11:40:11 +0200 (CEST) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga006.jf.intel.com ([10.7.209.51]) by fmsmga107.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 28 May 2019 02:40:10 -0700 X-ExtLoop1: 1 Received: from aburakov-mobl1.ger.corp.intel.com (HELO [10.237.220.117]) ([10.237.220.117]) by orsmga006.jf.intel.com with ESMTP; 28 May 2019 02:40:09 -0700 To: Yutaro Hayakawa , dev@dpdk.org References: <678FFC0B-20D6-4913-AF2B-CB08A10F5A7E@gmail.com> From: "Burakov, Anatoly" Message-ID: Date: Tue, 28 May 2019 10:40:09 +0100 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.7.0 MIME-Version: 1.0 In-Reply-To: <678FFC0B-20D6-4913-AF2B-CB08A10F5A7E@gmail.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 7bit Subject: Re: [dpdk-dev] AF_XDP PMD cannot use with software NICs 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: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On 25-May-19 1:45 AM, Yutaro Hayakawa wrote: > Hello, > > In current AF_XDP PMD implementation, we cannot use it with software NICs like VETH. Because it allocates umem with RTE_MEMZONE_IOVA_CONTIG. > > Since AF_XDP itself is usable with any network devices, I think it is better to have option to take umem with RTE_MEMZONE_SIZE_HINT_ONLY. It is also useful for testing perpose or using DPDK from containers. I'm not sure i understand the problem. IOVA_CONTIG has no relation to SIZE_HINT_ONLY - the former forces the allocation to be IOVA-contiguous. The latter allows to allocate from other page sizes when the requested one is not available, e.g. if you request memory with RTE_MEMZONE_1GB and RTE_MEMZONE_SIZE_HINT_ONLY, it will try to allocate from 1GB pages first, and if that fails, will allocate from any other page size available - but says nothing about whether that memory has to be IOVA-contiguous. What is the problem with allocating IOVA-contiguous memory for AF_XDP and VETH? > > Any thought or future plan for this problem? > > Thanks, > Yutaro > -- Thanks, Anatoly