From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by dpdk.org (Postfix) with ESMTP id 464EAAAC8 for ; Tue, 13 Mar 2018 16:52:16 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga002.fm.intel.com ([10.253.24.26]) by fmsmga103.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 13 Mar 2018 08:52:15 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.47,465,1515484800"; d="scan'208";a="27723818" Received: from irsmsx103.ger.corp.intel.com ([163.33.3.157]) by fmsmga002.fm.intel.com with ESMTP; 13 Mar 2018 08:52:13 -0700 Received: from irsmsx101.ger.corp.intel.com ([169.254.1.5]) by IRSMSX103.ger.corp.intel.com ([169.254.3.197]) with mapi id 14.03.0319.002; Tue, 13 Mar 2018 15:52:13 +0000 From: "Trahe, Fiona" To: "Verma, Shally" , Ahmed Mansour , "dev@dpdk.org" CC: "De Lara Guarch, Pablo" , "Athreya, Narayana Prasad" , "Gupta, Ashish" , "Sahu, Sunila" , "Challa, Mahipal" , "Jain, Deepak K" , Hemant Agrawal , "Roy Pledge" , Youri Querry , "Daly, Lee" , "Jozwiak, TomaszX" , "Trahe, Fiona" Thread-Topic: [dpdk-dev] [PATCH] compressdev: implement API - mbuf alternative Thread-Index: AdO6FqZGVQESVGJPRLukdFdKpdlnigAioyWQABAgLEA= Date: Tue, 13 Mar 2018 15:52:12 +0000 Message-ID: <348A99DA5F5B7549AA880327E580B4358934A600@IRSMSX101.ger.corp.intel.com> References: <348A99DA5F5B7549AA880327E580B435893478BA@IRSMSX101.ger.corp.intel.com> In-Reply-To: Accept-Language: en-IE, en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-titus-metadata-40: eyJDYXRlZ29yeUxhYmVscyI6IiIsIk1ldGFkYXRhIjp7Im5zIjoiaHR0cDpcL1wvd3d3LnRpdHVzLmNvbVwvbnNcL0ludGVsMyIsImlkIjoiNmUzZmI3YjEtZmY2Yi00MTlmLWJlZDAtMDczNDdhNDRjYzYyIiwicHJvcHMiOlt7Im4iOiJDVFBDbGFzc2lmaWNhdGlvbiIsInZhbHMiOlt7InZhbHVlIjoiQ1RQX05UIn1dfV19LCJTdWJqZWN0TGFiZWxzIjpbXSwiVE1DVmVyc2lvbiI6IjE2LjUuOS4zIiwiVHJ1c3RlZExhYmVsSGFzaCI6Im5BK1dURWVzOWlkQWo3K00xSjJFeUpMNFZvN0Y3VGQ3NXl0SnB0MFQ3RGc9In0= x-ctpclassification: CTP_NT dlp-product: dlpe-windows dlp-version: 11.0.0.116 dlp-reaction: no-action x-originating-ip: [163.33.239.182] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH] compressdev: implement API - mbuf alternative 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: Tue, 13 Mar 2018 15:52:17 -0000 Hi Shally, > -----Original Message----- > From: Verma, Shally [mailto:Shally.Verma@cavium.com] > Sent: Tuesday, March 13, 2018 8:15 AM > To: Trahe, Fiona ; Ahmed Mansour ; dev@dpdk.org > Cc: De Lara Guarch, Pablo ; Athreya, Nara= yana Prasad > ; Gupta, Ashish ; Sahu, Sunila > ; Challa, Mahipal ; Ja= in, Deepak K > ; Hemant Agrawal ; Roy P= ledge > ; Youri Querry ; fiona.trahe@= gmail.com; Daly, Lee > ; Jozwiak, TomaszX > Subject: RE: [dpdk-dev] [PATCH] compressdev: implement API - mbuf alterna= tive >=20 > HI Fiona >=20 > So I understand we're moving away from mbufs because of its size limitati= on (64k) and cacheline overhead > and their more suitability to n/w applications. Given that, I understand = benefit of having another structure > to input data but then what is proposal for ipcomp like application where= mbuf usage may be a better > option? Should we keep support for both (mbuf and this structure) so that= apps can use appropriate data > structure depending on their requirement. [Fiona] An application can use pass buffers from an mbuf or mbuf chain to c= ompressdev by filling in the=20 compressdev struct fields with the mbuf meta-data, using rte_pktmbuf_data_l= en(),=20 rte_pktmbuf_mtod(), rte_pktmbuf_mtophys(), etc For simplicity I'd prefer to offer only 1 rather than 2 data formats on the= API. We see storage applications rather than IPComp as the main use-case for com= pressdev, so would prefer to optimise for that. Do you think otherwise? >=20 > Further comments, on github. >=20 > Thanks > Shally >=20 > >-----Original Message----- > >From: Trahe, Fiona [mailto:fiona.trahe@intel.com] > >Sent: 12 March 2018 21:31 > >To: Ahmed Mansour ; Verma, Shally ; > dev@dpdk.org > >Cc: De Lara Guarch, Pablo ; Athreya, Nar= ayana Prasad > ; > >Gupta, Ashish ; Sahu, Sunila ; Challa, Mahipal > >; Jain, Deepak K ; H= emant Agrawal > ; Roy > >Pledge ; Youri Querry ; fion= a.trahe@gmail.com; Daly, > Lee ; > >Jozwiak, TomaszX > >Subject: RE: [dpdk-dev] [PATCH] compressdev: implement API - mbuf altern= ative > > > >Hi Shally, Ahmed, and anyone else interested in compressdev, > > > >I mentioned last week that we've been exploring using something other th= an mbufs to pass src/dst > buffers to compressdev PMDs. > > > >Reasons: > > - mbuf data is limited to 64k-1 in each segment of a chained mbuf. Data= for compression > > can be greater and it would add cycles to have to break up into smal= ler segments. > > - data may originate in mbufs, but is more likely, particularly for sto= rage use-cases, to > > originate in other data structures. > > - There's a 2 cache-line overhead for every segment in a chain, most of= this data > > is network-related, not needed by compressdev > >So moving to a custom structure would minimise memory overhead, remove r= estriction on 64k-1 size and > give more flexibility if > >compressdev ever needs any comp-specific meta-data. > > > >We've come up with a compressdev-specific structure using the struct iov= ec from sys/uio.h, which is > commonly used by storage > >applications. This would replace the src and dest mbufs in the op. > >I'll not include the code here - Pablo will push that to github shortly = and we'd appreciate review > comments there. > >https://github.com/pablodelara/dpdk-draft-compressdev > >Just posting on the mailing list to give a heads-up and ensure this reac= hes a wider audience than may see > it on github. > > > >Note : We also considered having no data structures in the op, instead t= he application > >would supply a callback which the PMD would use to retrieve meta-data (v= irt address, iova, length) > >for each next segment as needed. While this is quite flexible and allow = the application > >to keep its data in its native structures, it's likely to cost more cycl= es. > >So we're not proposing this at the moment, but hope to benchmark it late= r while the API is still > experimental. > > > >General feedback on direction is welcome here on the mailing list. > >For feedback on the details of implementation we would appreciate commen= ts on github. > > > >Regards, > >Fiona.