From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from EUR02-AM5-obe.outbound.protection.outlook.com (mail-eopbgr00083.outbound.protection.outlook.com [40.107.0.83]) by dpdk.org (Postfix) with ESMTP id CFEBC5F25 for ; Wed, 3 Oct 2018 13:47:21 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=Mellanox.com; s=selector1; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-SenderADCheck; bh=tJaZikCJRUdJ99HWrkJ9STel7cFe+GdTdBw/9zajTvk=; b=mi4qH5kvQpWIBm1fzeIJ1dcnReKTGn45d30xL5oBdfIvN5PTsSyKQIRcMgDA3TBg7DGgea08ju7wjpMXn8km26UlDAZsv37espu2ANUDvpBJwQ2w3y6ehjfWVCxNVNQZHc0LMJY/c+m1nvHMzHtO+MD3cnUPpZI8w1zZNFeP80Y= Received: from VI1PR05MB4224.eurprd05.prod.outlook.com (52.133.12.13) by VI1PR05MB1423.eurprd05.prod.outlook.com (10.162.124.154) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.20.1185.25; Wed, 3 Oct 2018 11:47:19 +0000 Received: from VI1PR05MB4224.eurprd05.prod.outlook.com ([fe80::f911:fe8:927d:400b]) by VI1PR05MB4224.eurprd05.prod.outlook.com ([fe80::f911:fe8:927d:400b%3]) with mapi id 15.20.1207.018; Wed, 3 Oct 2018 11:47:19 +0000 From: Dekel Peled To: Yongseok Koh CC: "dev@dpdk.org" , Shahaf Shuler , Ori Kam Thread-Topic: [PATCH 3/4] net/mlx5: add DV encap L2 and L3 operations Thread-Index: AQHUWuZ81hz3HEl6d0WhwfOEWORtC6UNJfewgABAGNA= Date: Wed, 3 Oct 2018 11:47:18 +0000 Message-ID: References: <1538059845-35896-1-git-send-email-dekelp@mellanox.com> <1538059845-35896-4-git-send-email-dekelp@mellanox.com> <20181003065817.GD10028@mtidpdk.mti.labs.mlnx> In-Reply-To: Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: authentication-results: spf=none (sender IP is ) smtp.mailfrom=dekelp@mellanox.com; x-originating-ip: [193.47.165.251] x-ms-publictraffictype: Email x-microsoft-exchange-diagnostics: 1; VI1PR05MB1423; 6:J/mD735uDsgDu9wcNJTxLP9OG631R/MNOBLeuZsaidgnuUCyir4ZleFA3t5ujAnDTyQ+Nvz5Efx6CR+lqq0qGTkT2P1w0TUth0FPkGsJwvYxeK+9iCRyWD4t3L40njh+2ycFPep3Yyr4DmaZE6tFXcV/ncCI/RysgteJvWaBhHrsAjgybgIEY15740YUe1TuOY3vHAEU13VlQD8u8fGq6L6/+wtpp1nQylGhOm0n48o7qZhcunKkv0S3fP5O3XwukeHddWp+tgpSZ09c3gGyeotfEUJWVsfUvvGCmeiG/xNh7E+Y43H/NNFpJTl+Ia/RseXokK6v+mTVnCKLtKpiUazDrOHS2hCE75o/Rw0FFWXSqbbr9pqGs/ZN3pvTBVi+TRMnhYZeR2Ih/noXBnER2vx+lzchmbX/bxtgvDNfHEhMcKBnzop6C+QlWx1c8Lc1/gu5haqF013CM4UDNOuUtA==; 5:I4waisZDkIzvu1+X3EWhaiHQQJCcKGoCJ4kfgiMCOrE2/3SPPbhqBfMsGur/epe6ysSexzin2eHSCE595bGteO2mOpjb8xemU2kSc9NTim+qstE+5PrS6AG7vI6z8jixfoEgTsEbwmJ7frU8BAcPBa2c7Uxot2vvMcU9YGtzLIc=; 7:wtVjdBEDrN7MMlk5p9zY9Y7mRw8EB7bo5Z2Bz3SVJ8marFDZxDbI6cDo/s5JZSryN3rS4QcfHwFKxp5oIDKIdSi7m/4KiXQDR6WhE9cIaIP2eT2qYcEpuD1THPv2I08NK5oNCCnhXdmu1CI4NZXCBxaOqVtvzgRcvlwrswOhnEUlsksKPhPuDvtjMqrIa+HvmwCePxCd1PSPAswxYUWyH5L5jDGFIqU644zRGJzOycwrNVgjQt8TbEcSo9OxuCsN x-ms-exchange-antispam-srfa-diagnostics: SOS; x-ms-office365-filtering-correlation-id: d623b22c-5e6e-401d-58ae-08d62925f932 x-ms-office365-filtering-ht: Tenant x-microsoft-antispam: BCL:0; PCL:0; RULEID:(7020095)(4652040)(8989299)(4534185)(4627221)(201703031133081)(201702281549075)(8990200)(5600074)(711020)(4618075)(2017052603328)(7153060)(7193020); SRVR:VI1PR05MB1423; x-ms-traffictypediagnostic: VI1PR05MB1423: x-microsoft-antispam-prvs: x-exchange-antispam-report-test: UriScan:; x-ms-exchange-senderadcheck: 1 x-exchange-antispam-report-cfa-test: BCL:0; PCL:0; RULEID:(8211001083)(6040522)(2401047)(8121501046)(5005006)(93006095)(93001095)(3231355)(944501410)(52105095)(10201501046)(3002001)(6055026)(149066)(150057)(6041310)(20161123558120)(20161123562045)(20161123564045)(20161123560045)(201703131423095)(201702281528075)(20161123555045)(201703061421075)(201703061406153)(201708071742011)(7699051); SRVR:VI1PR05MB1423; BCL:0; PCL:0; RULEID:; SRVR:VI1PR05MB1423; x-forefront-prvs: 0814A2C7A3 x-forefront-antispam-report: SFV:NSPM; SFS:(10009020)(39860400002)(346002)(376002)(136003)(396003)(366004)(189003)(199004)(13464003)(6636002)(3846002)(6246003)(6116002)(14444005)(81156014)(6862004)(81166006)(2906002)(66066001)(478600001)(97736004)(256004)(71200400001)(71190400001)(476003)(8936002)(486006)(229853002)(33656002)(11346002)(2940100002)(14454004)(446003)(4744004)(93156006)(4326008)(55016002)(54906003)(25786009)(106356001)(186003)(53546011)(6506007)(2900100001)(93886005)(7696005)(99286004)(68736007)(26005)(9686003)(105586002)(316002)(305945005)(7736002)(53936002)(6436002)(107886003)(5660300001)(8676002)(102836004)(76176011)(5250100002)(86362001)(74316002); DIR:OUT; SFP:1101; SCL:1; SRVR:VI1PR05MB1423; H:VI1PR05MB4224.eurprd05.prod.outlook.com; FPR:; SPF:None; LANG:en; PTR:InfoNoRecords; MX:1; A:1; received-spf: None (protection.outlook.com: mellanox.com does not designate permitted sender hosts) x-microsoft-antispam-message-info: xn04V2DzPtDMrxcPBMCjFNQTIVFQYekTQQNs1CN/SbXZ4xN/V5Y+USWhkmz/s++Z6PDdS61oAbeJmNBw1O/Lj1KWsrSaFg23F3B9/N8VlfdSuFTX4qpTgLFLqnuHH4ASDh3AezVENMsIqhbnlxzSjAeTu/yGUirgZD0+RxywkXbq4kq/0EAuG1jH3V4JcCrpzSFKXvvEZ0m0rOEEshgVM9AXxMiMq6+O7H9Hi1wSjeV/wpfPSMLwxIwmrI94cVX1zoSCEQRf2SRIHCtsn2vhwcwulUZhZ4aolbAlL/Sh9IgPii6oVBZj6gc5aoC1utPTOQDIKE1VIRRY7XkcJEd6yGuifXB29SiGY3pgvpnhuI4= spamdiagnosticoutput: 1:99 spamdiagnosticmetadata: NSPM Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-OriginatorOrg: Mellanox.com X-MS-Exchange-CrossTenant-Network-Message-Id: d623b22c-5e6e-401d-58ae-08d62925f932 X-MS-Exchange-CrossTenant-originalarrivaltime: 03 Oct 2018 11:47:18.9452 (UTC) X-MS-Exchange-CrossTenant-fromentityheader: Hosted X-MS-Exchange-CrossTenant-id: a652971c-7d2e-4d9b-a6a4-d149256f461b X-MS-Exchange-Transport-CrossTenantHeadersStamped: VI1PR05MB1423 Subject: Re: [dpdk-dev] [PATCH 3/4] net/mlx5: add DV encap L2 and L3 operations 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: Wed, 03 Oct 2018 11:47:22 -0000 One more issue, PSB. > -----Original Message----- > From: Dekel Peled > Sent: Wednesday, October 3, 2018 11:35 AM > To: Yongseok Koh > Cc: dev@dpdk.org; Shahaf Shuler ; Ori Kam > > Subject: RE: [PATCH 3/4] net/mlx5: add DV encap L2 and L3 operations >=20 > Thanks, PSB. >=20 > > -----Original Message----- > > From: Yongseok Koh > > Sent: Wednesday, October 3, 2018 9:58 AM > > To: Dekel Peled > > Cc: dev@dpdk.org; Shahaf Shuler ; Ori Kam > > > > Subject: Re: [PATCH 3/4] net/mlx5: add DV encap L2 and L3 operations > > > > On Thu, Sep 27, 2018 at 05:50:44PM +0300, Dekel Peled wrote: > > > This patch adds support for Direct Verbs encap operations, L2 and L3. > > > > > > Signed-off-by: Dekel Peled > > > --- > > > drivers/net/mlx5/mlx5_flow_dv.c | 249 > > > +++++++++++++++++++++++++++++++++++++++- > > > 1 file changed, 244 insertions(+), 5 deletions(-) > > > > > > diff --git a/drivers/net/mlx5/mlx5_flow_dv.c > > > b/drivers/net/mlx5/mlx5_flow_dv.c index 1f3fcb8..50925ac 100644 > > > --- a/drivers/net/mlx5/mlx5_flow_dv.c > > > +++ b/drivers/net/mlx5/mlx5_flow_dv.c > > > @@ -85,6 +85,178 @@ > > > } > > > > > > /** > > > + * Validate the tunnel encap action. > > > + * > > > + * @param[in] action_flags > > > + * holds the actions detected until now. > > > + * @param[in] action > > > + * Pointer to the encap action. > > > + * @param[in] attr > > > + * Pointer to flow attributes > > > + * @param[out] error > > > + * Pointer to error structure. > > > + * > > > + * @return > > > + * 0 on success, a negative errno value otherwise and rte_ernno is= set. > > > + */ > > > +static int > > > +mlx5_flow_validate_action_tunnel_encap(uint64_t action_flags, > > > > A static func in this file should start from flow_dv_. Or, you can > > move these to mlx5_flow.c as general validation helpers. I prefer the l= atter. >=20 > Renamed functions to flow_dv_validate_action_tunnel_... > Left it as static in mlx5_flow_dv.c, since it is relevant for DV only. >=20 > > > > > + const struct rte_flow_action *action, > > > + const struct rte_flow_attr *attr, > > > + struct rte_flow_error *error) { > > > + const struct rte_flow_action_tunnel_encap *tunnel_encap =3D > > > +action->conf; > > > + > > > + if (!tunnel_encap || !(tunnel_encap->buf)) > > > + return rte_flow_error_set(error, EINVAL, > > > + RTE_FLOW_ERROR_TYPE_ACTION, > > > + action, > > > > Two lines above can be merged? Please check others and merge lines as > > much as possible. >=20 > Done. >=20 > > > > > + "configuration cannot be null"); > > > + if (action_flags & MLX5_ACTION_DROP) > > > + return rte_flow_error_set(error, > > > + ENOTSUP, > > > + RTE_FLOW_ERROR_TYPE_ACTION, > > > + NULL, > > > + "can't drop and encap in same > > flow"); > > > + if (action_flags & (MLX5_ACTION_TUNNEL_ENCAP | > > > + MLX5_ACTION_TUNNEL_ENCAP_L3)) > > > + return rte_flow_error_set(error, > > > + ENOTSUP, > > > + RTE_FLOW_ERROR_TYPE_ACTION, > > > + NULL, > > > + "can't have 2 encap actions in same" > > > + " flow"); > > > + if (attr->ingress) > > > + return rte_flow_error_set(error, ENOTSUP, > > > + > > RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, > > > + NULL, > > > + "encap action not supported for " > > > + "ingress"); > > > + return 0; > > > +} > > > + > > > +/** > > > + * Validate the tunnel encap L3 action > > > + * > > > + * @param[in] action_flags > > > + * holds the actions detected until now. > > > + * @param[in] action > > > + * Pointer to the encap action. > > > + * @param[in] attr > > > + * Pointer to flow attributes > > > + * @param[out] error > > > + * Pointer to error structure. > > > + * > > > + * @return > > > + * 0 on success, a negative errno value otherwise and rte_ernno is= set. > > > + */ > > > +static int > > > +mlx5_flow_validate_action_tunnel_encap_l3(uint64_t action_flags, > > > + const struct rte_flow_action *action, > > > + const struct rte_flow_attr *attr, > > > + struct rte_flow_error *error) { > > > + const struct rte_flow_action_tunnel_encap_l3 *tunnel_encap_l3 =3D > > > + action->conf; > > > > Indentation. > > > > const struct rte_flow_action_tunnel_encap_l3 *tunnel_encap_l3 =3D > > action->conf; >=20 > Done. >=20 > > > > > + > > > + if (!tunnel_encap_l3 || !(tunnel_encap_l3->buf)) > > > + return rte_flow_error_set(error, EINVAL, > > > + RTE_FLOW_ERROR_TYPE_ACTION, > > > + action, > > > + "configuration cannot be null"); > > > + if (action_flags & MLX5_ACTION_DROP) > > > + return rte_flow_error_set(error, > > > + ENOTSUP, > > > + RTE_FLOW_ERROR_TYPE_ACTION, > > > + NULL, > > > + "can't drop and encap in same > > flow"); > > > + if (action_flags & (MLX5_ACTION_TUNNEL_ENCAP | > > > + MLX5_ACTION_TUNNEL_ENCAP_L3)) > > > + return rte_flow_error_set(error, > > > + ENOTSUP, > > > + RTE_FLOW_ERROR_TYPE_ACTION, > > > + NULL, > > > + "can't have 2 encap actions in same" > > > + " flow"); > > > + if (attr->ingress) > > > + return rte_flow_error_set(error, ENOTSUP, > > > + > > RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, > > > + NULL, > > > + "encap action not supported for " > > > + "ingress"); > > > + return 0; > > > +} > > > + > > > +/** > > > + * Convert encap action to DV specification. > > > + * > > > + * @param[in] action > > > + * Pointer to action structure. > > > + * @param[in] ctx > > > + * Verbs context. > > > + * @param[out] error > > > + * Pointer to the error structure. > > > + * > > > + * @return > > > + * Pointer to action on success, NULL otherwise and rte_errno is s= et. > > > + */ > > > +static struct ibv_flow_action * > > > +mlx5_flow_dv_create_encap(const struct rte_flow_action *action, > > > > Change to flow_dv_create_encap() >=20 > Done. >=20 > > > > > + struct ibv_context *ctx, > > > + struct rte_flow_error *error) > > > +{ > > > + struct ibv_flow_action *encap_verb =3D NULL; > > > + const struct rte_flow_action_tunnel_encap *encap_data; > > > + > > > + encap_data =3D (const struct rte_flow_action_tunnel_encap *)action- > > >conf; > > > + encap_verb =3D mlx5_glue- > > >dv_create_flow_action_packet_reformat(ctx, > > > + encap_data->size, > > > + encap_data->size ? encap_data->buf : > > > + NULL, > > > + > > MLX5DV_FLOW_ACTION_PACKET_REFORMAT_TYPE_L2_TO_L2_TU > > NNEL, > > > + MLX5DV_FLOW_TABLE_TYPE_NIC_TX); > > > > Indentation. >=20 > I'm using very long MLX5DV_... names defined in rdma-core. > If I use the required indentation I get illegal line length. >=20 > > > > encap_verb =3D mlx5_glue->dv_create_flow_action_packet_reformat > > (ctx, encap_data->size, > > encap_data->size ? encap_data->buf : NULL, > > > > MLX5DV_FLOW_ACTION_PACKET_REFORMAT_TYPE_L2_TO_L2_TUNNEL, > > MLX5DV_FLOW_TABLE_TYPE_NIC_TX); > > > > Please make the same change to others. > > > > > + if (!encap_verb) > > > + rte_flow_error_set(error, EINVAL, > > RTE_FLOW_ERROR_TYPE_ACTION, > > > + NULL, "cannot create encap action"); > > > + return encap_verb; > > > +} > > > + > > > +/** > > > + * Convert encap L3 action to DV specification. > > > + * > > > + * @param[in] action > > > + * Pointer to action structure. > > > + * @param[in] ctx > > > + * Verbs context. > > > + * @param[out] error > > > + * Pointer to the error structure. > > > + * > > > + * @return > > > + * Pointer to action on success, NULL otherwise and rte_errno is s= et. > > > + */ > > > +static struct ibv_flow_action * > > > +mlx5_flow_dv_create_encap_l3(const struct rte_flow_action *action, > > > > Change to flow_dv_create_encap_l3() >=20 > Done. >=20 > > > > > + struct ibv_context *ctx, > > > + struct rte_flow_error *error) { > > > + struct ibv_flow_action *encap_l3_verb =3D NULL; > > > + const struct rte_flow_action_tunnel_encap_l3 *encap_l3_data; > > > + > > > + encap_l3_data =3D > > > + (const struct rte_flow_action_tunnel_encap_l3 *)action- > > >conf; > > > + encap_l3_verb =3D mlx5_glue- > > >dv_create_flow_action_packet_reformat(ctx, > > > + encap_l3_data->size, > > > + encap_l3_data->size ? encap_l3_data->buf : > > > + NULL, > > > + > > MLX5DV_FLOW_ACTION_PACKET_REFORMAT_TYPE_L2_TO_L3_TU > > NNEL, > > > + MLX5DV_FLOW_TABLE_TYPE_NIC_TX); > > > + if (!encap_l3_verb) > > > + rte_flow_error_set(error, EINVAL, > > RTE_FLOW_ERROR_TYPE_ACTION, > > > + NULL, "cannot create encap L3 action"); > > > + return encap_l3_verb; > > > +} > > > + > > > +/** > > > * Verify the @p attributes will be correctly understood by the NIC > > > and > > store > > > * them in the @p flow if everything is correct. > > > * > > > @@ -337,6 +509,26 @@ > > > action_flags |=3D MLX5_ACTION_COUNT; > > > ++actions_n; > > > break; > > > + case RTE_FLOW_ACTION_TYPE_TUNNEL_ENCAP: > > > + ret =3D mlx5_flow_validate_action_tunnel_encap > > > + (action_flags, > > > + actions, > > > + attr, > > > + error); > > > > Can be one line. >=20 > Done. >=20 > > > > > + if (ret < 0) > > > + return ret; > > > + action_flags |=3D MLX5_ACTION_TUNNEL_ENCAP; > > > + break; > > > + case RTE_FLOW_ACTION_TYPE_TUNNEL_ENCAP_L3: > > > + ret =3D mlx5_flow_validate_action_tunnel_encap_l3 > > > + (action_flags, > > > + actions, > > > + attr, > > > + error); > > > + if (ret < 0) > > > + return ret; > > > + action_flags |=3D MLX5_ACTION_TUNNEL_ENCAP_L3; > > > + break; > > > default: > > > return rte_flow_error_set(error, ENOTSUP, > > > > > RTE_FLOW_ERROR_TYPE_ACTION, > > > @@ -1047,10 +1239,19 @@ > > > * Flow action to translate. > > > * @param[in, out] dev_flow > > > * Pointer to the mlx5_flow. > > > + * @param[in] ctx > > > + * Verbs context. > > > + * @param[out] error > > > + * Pointer to the error structure. > > > + * > > > + * @return > > > + * 0 on success, a negative errno value otherwise and rte_ernno is= set. > > > */ > > > -static void > > > +static int > > > flow_dv_create_action(const struct rte_flow_action *action, > > > - struct mlx5_flow *dev_flow) > > > + struct mlx5_flow *dev_flow, > > > + struct ibv_context *ctx, > > > > If it is just priv->ctx, it would be better to get dev as an arg and > > make mlx5_flow_dv_create_encap*(dev, ...) gets priv->ctx from dev. >=20 > I considered it during implementation, but preferred to give the function= s > only what they need. >=20 > > > > > + struct rte_flow_error *error) > > > { > > > const struct rte_flow_action_queue *queue; > > > const struct rte_flow_action_rss *rss; @@ -1095,10 +1296,35 @@ > > > /* Added to array only in apply since we need the QP */ > > > flow->actions |=3D MLX5_ACTION_RSS; > > > break; > > > + case RTE_FLOW_ACTION_TYPE_TUNNEL_ENCAP: > > > + dev_flow->dv.actions[actions_n].type =3D > > > + MLX5DV_FLOW_ACTION_IBV_FLOW_ACTION; > > > + dev_flow->dv.actions[actions_n].action =3D > > > + mlx5_flow_dv_create_encap(action, ctx, error); > > > + if (!(dev_flow->dv.actions[actions_n].action)) > > > + return -rte_errno; > > > + dev_flow->dv.encap_verb =3D > > > + dev_flow->dv.actions[actions_n].action; > > > + flow->actions |=3D MLX5_ACTION_TUNNEL_ENCAP; > > > + actions_n++; > > > + break; > > > + case RTE_FLOW_ACTION_TYPE_TUNNEL_ENCAP_L3: > > > + dev_flow->dv.actions[actions_n].type =3D > > > + MLX5DV_FLOW_ACTION_IBV_FLOW_ACTION; > > > + dev_flow->dv.actions[actions_n].action =3D > > > + mlx5_flow_dv_create_encap_l3(action, ctx, error); > > > + if (!(dev_flow->dv.actions[actions_n].action)) > > > + return -rte_errno; > > > + dev_flow->dv.encap_verb =3D > > > + dev_flow->dv.actions[actions_n].action; > > > + flow->actions |=3D MLX5_ACTION_TUNNEL_ENCAP_L3; > > > + actions_n++; > > > + break; > > > default: > > > break; > > > } > > > dev_flow->dv.actions_n =3D actions_n; > > > + return 0; > > > } > > > > > > static uint32_t matcher_zero[MLX5_ST_SZ_DW(fte_match_param)] =3D { 0 > > > }; @@ -1247,6 +1473,7 @@ > > > }; > > > void *match_value =3D dev_flow->dv.value.buf; > > > uint8_t inner =3D 0; > > > + int retval =3D 0; > > > > > > if (priority =3D=3D MLX5_FLOW_PRIO_RSVD) > > > priority =3D priv->config.flow_prio - 1; @@ -1262,9 +1489,14 > > @@ > > > matcher.egress =3D attr->egress; > > > if (flow_dv_matcher_register(dev, &matcher, dev_flow, error)) > > > return -rte_errno; > > > - for (; actions->type !=3D RTE_FLOW_ACTION_TYPE_END; actions++) > > > - flow_dv_create_action(actions, dev_flow); > > > - return 0; > > > + for (; actions->type !=3D RTE_FLOW_ACTION_TYPE_END; actions++) { > > > + retval =3D > > > + flow_dv_create_action(actions, dev_flow, priv->ctx, error); > > > + if (retval) > > > + break; > > > > The following looks better matched? > > > > if (flow_dv_create_action(dev, actions, dev_flow, error)) > > return -rte_errno; > > > > And leave 'return 0;' as is below. >=20 > OK. >=20 > > > > BTW, flow_dv_matcher_register() allocates mlx5_flow_dv_matcher or > > increase the refcnt. Then, if it fails here, doesn't it need to clean t= hat up? Missed this in the first time. Moved the create_action loop earlier, right after the create_item loop, and before the matcher_register call. > > > > > + } > > > + > > > > No blank line is allowed. Please fix other occurrences too. >=20 > Done. >=20 > > > > > + return retval; > > > } > > > > > > /** > > > @@ -1447,6 +1679,13 @@ > > > LIST_REMOVE(dev_flow, next); > > > if (dev_flow->dv.matcher) > > > flow_dv_matcher_release(dev, dev_flow- > dv.matcher); > > > + > > > + if (dev_flow->dv.encap_verb) { > > > + claim_zero(mlx5_glue->destroy_flow_action > > > + (dev_flow->dv.encap_verb)); > > > + dev_flow->dv.encap_verb =3D NULL; > > > + } > > > + > > > rte_free(dev_flow); > > > } > > > } > > > -- > > > 1.8.3.1 > > >