From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga04.intel.com (mga04.intel.com [192.55.52.120]) by dpdk.org (Postfix) with ESMTP id BA0027CF8 for ; Mon, 4 Sep 2017 17:55:15 +0200 (CEST) Received: from orsmga004.jf.intel.com ([10.7.209.38]) by fmsmga104.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 04 Sep 2017 08:55:14 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.41,475,1498546800"; d="scan'208";a="125444261" Received: from irsmsx108.ger.corp.intel.com ([163.33.3.3]) by orsmga004.jf.intel.com with ESMTP; 04 Sep 2017 08:55:13 -0700 Received: from irsmsx156.ger.corp.intel.com (10.108.20.68) by IRSMSX108.ger.corp.intel.com (163.33.3.3) with Microsoft SMTP Server (TLS) id 14.3.319.2; Mon, 4 Sep 2017 16:55:12 +0100 Received: from irsmsx105.ger.corp.intel.com ([169.254.7.75]) by IRSMSX156.ger.corp.intel.com ([169.254.3.130]) with mapi id 14.03.0319.002; Mon, 4 Sep 2017 16:55:12 +0100 From: "Ananyev, Konstantin" To: Shahaf Shuler , Thomas Monjalon CC: "dev@dpdk.org" Thread-Topic: [dpdk-dev] [PATCH 4/4] ethdev: add helpers to move to the new offloads API Thread-Index: AQHTJU1fVpbLAok7P06kzXdRtIICjKKkr2EAgAACQgCAACmKQA== Date: Mon, 4 Sep 2017 15:55:11 +0000 Message-ID: <2601191342CEEE43887BDE71AB9772584F246317@irsmsx105.ger.corp.intel.com> References: <810c1d26724f82f0d9fc9d6684dc4b1c62fd5f62.1504508375.git.shahafs@mellanox.com> <2601191342CEEE43887BDE71AB9772584F24602F@irsmsx105.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: dlp-product: dlpe-windows dlp-version: 11.0.0.116 dlp-reaction: no-action x-originating-ip: [163.33.239.180] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH 4/4] ethdev: add helpers to move to the new offloads API 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, 04 Sep 2017 15:55:16 -0000 > -----Original Message----- > From: Shahaf Shuler [mailto:shahafs@mellanox.com] > Sent: Monday, September 4, 2017 3:03 PM > To: Ananyev, Konstantin ; Thomas Monjalon <= thomas@monjalon.net> > Cc: dev@dpdk.org > Subject: RE: [dpdk-dev] [PATCH 4/4] ethdev: add helpers to move to the ne= w offloads API >=20 > Hi Konstantin, >=20 > Monday, September 4, 2017 4:25 PM, Ananyev, Konstantin: > > > > Hi Shahaf, > > > > > } > > > > > > +/** > > > + * A conversion function from rxmode offloads API to rte_eth_rxq_con= f > > > + * offloads API. > > > + */ > > > +static void > > > +rte_eth_convert_rxmode_offloads(struct rte_eth_rxmode *rxmode, > > > + struct rte_eth_rxq_conf *rxq_conf) { > > > + if (rxmode->header_split =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_HEADER_SPLIT; > > > + if (rxmode->hw_ip_checksum =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_CHECKSUM; > > > + if (rxmode->hw_vlan_filter =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_VLAN_FILTER; > > > > Thinking on it a bit more: > > VLAN_FILTER is definitely one per device, as it would affect VFs also. > > At least that's what we have for Intel devices (ixgbe, i40e) right now. >=20 > This is vendor specific. For Mellanox this is one per device (regardless = if it is a vf/pf). >=20 > > For Intel devices VLAN_STRIP is also per device and will also be appli= ed to all > > corresponding VFs. >=20 > Again - vendor specific. For Mellanox is per queue. Yep, I understand it varies quite a lot from vendor from vendor. That's why I started to think we need to have to allow user to specify rx_o= ffloads for both port and queue basis. >=20 > > In fact, right now it is possible to query/change these 3 vlan offload = flags on > > the fly (after dev_start) on port basis by > > rte_eth_dev_(get|set)_vlan_offload API. > > So, I think at least these 3 flags need to be remained on a port basis. >=20 > Am not sure I agree. > Why, for example, block from application the option to set some queues wi= th vlan strip and some without if device allows? > Also how will we decide which offloads should stay per port and which are= allowed to move per queue? this much depends on the > underlying PMD. >=20 > Looks like i missed that part on ethdev, and if Rx offload will be per qu= eue I will need to change it also. >=20 >=20 > > In fact, why can't we have both per port and per queue RX offload: > > - dev_configure() will accept RX_OFFLOAD_* flags and apply them on a po= rt > > basis. > > - rx_queue_setup() will also accept RX_OFFLOAD_* flags and apply them o= n > > a queue basis. > > - if particular RX_OFFLOAD flag for that device couldn't be setup on a = queue > > basis - > > rx_queue_setup() will return an error. >=20 > Why not taking the per port configuration as a sub-case of per queue conf= iguration? > For per-port offloads as long as the same configuration applies the queue= setup succeeds. Do you mean that as long as queue config for offload X matches port config = for the same offload - it is ok, even if for that particular device offload X is per port not per queue? Let say: ... rxconf.offloads =3D DEV_RX_OFFLOAD_VLAN_STRIP; ... rte_eth_dev_configure(port, ...); ... /* queue offloads matches port offloads - always */ rte_eth_rx_queue_setup(port, ..., &rcxonf); ... rxconf.offloads &=3D ~ DEV_RX_OFFLOAD_VLAN_STRIP; /* ok for devices where vlan_stip per queue is supported, * fails for devices with vlan_strip offload is per device. */=20 rte_eth_rx_queue_setup(port, ..., &rcxonf); ? Konstantin =20 >=20 >=20 > > - rte_eth_rxq_info can be extended to provide information which > > RX_OFFLOADs > > can be configured on a per queue basis. >=20 > I am OK with the info suggestion. >=20 > > BTW - in that case we probably wouldn't need ignore flag inside rx_conf > > anymore. > > > > > > > + if (rxmode->hw_vlan_strip =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_VLAN_STRIP; > > > + if (rxmode->hw_vlan_extend =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_VLAN_EXTEND; > > > + if (rxmode->jumbo_frame =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_JUMBO_FRAME; > > > > There are some extra checks for that flag inside rte_eth_dev_configure(= ). > > If we going so support it per queue - then it probably need to be updat= ed. > > > > > + if (rxmode->hw_strip_crc =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_CRC_STRIP; > > > + if (rxmode->enable_scatter =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_SCATTER; > > > + if (rxmode->enable_lro =3D=3D 1) > > > + rxq_conf->offloads |=3D DEV_RX_OFFLOAD_TCP_LRO; } > > > + > > > > Konstantin