From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from dpdk.org (dpdk.org [92.243.14.124]) by inbox.dpdk.org (Postfix) with ESMTP id 0E28AA0487 for ; Wed, 3 Jul 2019 10:53:56 +0200 (CEST) Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id 1480E1BE9B; Wed, 3 Jul 2019 10:49:12 +0200 (CEST) Received: from mx0b-0016f401.pphosted.com (mx0a-0016f401.pphosted.com [67.231.148.174]) by dpdk.org (Postfix) with ESMTP id E593F1BDEF for ; Wed, 3 Jul 2019 10:49:07 +0200 (CEST) Received: from pps.filterd (m0045849.ppops.net [127.0.0.1]) by mx0a-0016f401.pphosted.com (8.16.0.27/8.16.0.27) with SMTP id x638j1Qr030527 for ; Wed, 3 Jul 2019 01:49:07 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com; h=from : to : cc : subject : date : message-id : in-reply-to : references : mime-version : content-transfer-encoding : content-type; s=pfpt0818; bh=IIJK/6cCxUizjJfcQWB9ffeZaen1XTnjZFq2LB+TIu8=; b=r0pDY1egZ4NmgeWTVHYDAga6LtLbe4yDzklDC/qflGYZWA+f6bf6eqStoBSNkL4+ausq flPQ5an6HmkU07yzPiriuhD/jAPkxpZkyqUhOswO8QoRHf1UhtWweeRYB5QGh6TEa3bO +tU03niB3BfIOgyKwDwePXIAKFEKoekiJ9RQV0QG/mgN2ziQJARST1IoVts3m6AXGKrG MIgxSGcBJPYjenvg30uwwQE3EZzgFP5MJ3x1SfyyL84DI0IXXti0f6WGKM5OcnQErngA toStvPxs9xIS1caivkORdxceOo4i3u+XB2bYCYxnBpHbtk+uFOcZJs14UKTgCQgyghtw OA== Received: from sc-exch03.marvell.com ([199.233.58.183]) by mx0a-0016f401.pphosted.com with ESMTP id 2tgrv182p4-4 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT) for ; Wed, 03 Jul 2019 01:49:07 -0700 Received: from SC-EXCH03.marvell.com (10.93.176.83) by SC-EXCH03.marvell.com (10.93.176.83) with Microsoft SMTP Server (TLS) id 15.0.1367.3; Wed, 3 Jul 2019 01:45:17 -0700 Received: from maili.marvell.com (10.93.176.43) by SC-EXCH03.marvell.com (10.93.176.83) with Microsoft SMTP Server id 15.0.1367.3 via Frontend Transport; Wed, 3 Jul 2019 01:45:17 -0700 Received: from jerin-lab.marvell.com (jerin-lab.marvell.com [10.28.34.14]) by maili.marvell.com (Postfix) with ESMTP id DE7483F7041; Wed, 3 Jul 2019 01:45:15 -0700 (PDT) From: To: , Jerin Jacob , Nithin Dabilpuram , Kiran Kumar K CC: Vivek Sharma Date: Wed, 3 Jul 2019 14:12:24 +0530 Message-ID: <20190703084244.33553-39-jerinj@marvell.com> X-Mailer: git-send-email 2.21.0 In-Reply-To: <20190703084244.33553-1-jerinj@marvell.com> References: <20190630180609.36705-1-jerinj@marvell.com> <20190703084244.33553-1-jerinj@marvell.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:, , definitions=2019-07-03_03:, , signatures=0 Subject: [dpdk-dev] [PATCH v3 38/58] net/octeontx2: add flow parse actions support 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" From: Kiran Kumar K Adding support to parse flow actions like drop, count, mark, rss, queue. On egress side, only drop and count actions were supported. Signed-off-by: Kiran Kumar K Signed-off-by: Vivek Sharma --- drivers/net/octeontx2/otx2_flow_parse.c | 276 ++++++++++++++++++++++++ drivers/net/octeontx2/otx2_rx.h | 1 + 2 files changed, 277 insertions(+) diff --git a/drivers/net/octeontx2/otx2_flow_parse.c b/drivers/net/octeontx2/otx2_flow_parse.c index b46fdd258..7f997ab74 100644 --- a/drivers/net/octeontx2/otx2_flow_parse.c +++ b/drivers/net/octeontx2/otx2_flow_parse.c @@ -679,3 +679,279 @@ otx2_flow_parse_la(struct otx2_parse_state *pst) /* Update pst if not validate only? clash check? */ return otx2_flow_update_parse_state(pst, &info, lid, lt, 0); } + +static int +parse_rss_action(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr, + const struct rte_flow_action *act, + struct rte_flow_error *error) +{ + struct otx2_eth_dev *hw = dev->data->dev_private; + struct otx2_rss_info *rss_info = &hw->rss_info; + const struct rte_flow_action_rss *rss; + uint32_t i; + + rss = (const struct rte_flow_action_rss *)act->conf; + + /* Not supported */ + if (attr->egress) { + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, + attr, "No support of RSS in egress"); + } + + if (dev->data->dev_conf.rxmode.mq_mode != ETH_MQ_RX_RSS) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, + act, "multi-queue mode is disabled"); + + /* Parse RSS related parameters from configuration */ + if (!rss || !rss->queue_num) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, "no valid queues"); + + if (rss->func != RTE_ETH_HASH_FUNCTION_DEFAULT) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, act, + "non-default RSS hash functions" + " are not supported"); + + if (rss->key_len && rss->key_len > RTE_DIM(rss_info->key)) + return rte_flow_error_set(error, ENOTSUP, + RTE_FLOW_ERROR_TYPE_ACTION, act, + "RSS hash key too large"); + + if (rss->queue_num > rss_info->rss_size) + return rte_flow_error_set + (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION, act, + "too many queues for RSS context"); + + for (i = 0; i < rss->queue_num; i++) { + if (rss->queue[i] >= dev->data->nb_rx_queues) + return rte_flow_error_set(error, EINVAL, + RTE_FLOW_ERROR_TYPE_ACTION, + act, + "queue id > max number" + " of queues"); + } + + return 0; +} + +int +otx2_flow_parse_actions(struct rte_eth_dev *dev, + const struct rte_flow_attr *attr, + const struct rte_flow_action actions[], + struct rte_flow_error *error, + struct rte_flow *flow) +{ + struct otx2_eth_dev *hw = dev->data->dev_private; + struct otx2_npc_flow_info *npc = &hw->npc_flow; + const struct rte_flow_action_count *act_count; + const struct rte_flow_action_mark *act_mark; + const struct rte_flow_action_queue *act_q; + const char *errmsg = NULL; + int sel_act, req_act = 0; + uint16_t pf_func; + int errcode = 0; + int mark = 0; + int rq = 0; + + /* Initialize actions */ + flow->ctr_id = NPC_COUNTER_NONE; + + for (; actions->type != RTE_FLOW_ACTION_TYPE_END; actions++) { + otx2_npc_dbg("Action type = %d", actions->type); + + switch (actions->type) { + case RTE_FLOW_ACTION_TYPE_VOID: + break; + case RTE_FLOW_ACTION_TYPE_MARK: + act_mark = + (const struct rte_flow_action_mark *)actions->conf; + + /* We have only 16 bits. Use highest val for flag */ + if (act_mark->id > (OTX2_FLOW_FLAG_VAL - 2)) { + errmsg = "mark value must be < 0xfffe"; + errcode = ENOTSUP; + goto err_exit; + } + mark = act_mark->id + 1; + req_act |= OTX2_FLOW_ACT_MARK; + rte_atomic32_inc(&npc->mark_actions); + break; + + case RTE_FLOW_ACTION_TYPE_FLAG: + mark = OTX2_FLOW_FLAG_VAL; + req_act |= OTX2_FLOW_ACT_FLAG; + rte_atomic32_inc(&npc->mark_actions); + break; + + case RTE_FLOW_ACTION_TYPE_COUNT: + act_count = + (const struct rte_flow_action_count *) + actions->conf; + + if (act_count->shared == 1) { + errmsg = "Shared Counters not supported"; + errcode = ENOTSUP; + goto err_exit; + } + /* Indicates, need a counter */ + flow->ctr_id = 1; + req_act |= OTX2_FLOW_ACT_COUNT; + break; + + case RTE_FLOW_ACTION_TYPE_DROP: + req_act |= OTX2_FLOW_ACT_DROP; + break; + + case RTE_FLOW_ACTION_TYPE_QUEUE: + /* Applicable only to ingress flow */ + act_q = (const struct rte_flow_action_queue *) + actions->conf; + rq = act_q->index; + if (rq >= dev->data->nb_rx_queues) { + errmsg = "invalid queue index"; + errcode = EINVAL; + goto err_exit; + } + req_act |= OTX2_FLOW_ACT_QUEUE; + break; + + case RTE_FLOW_ACTION_TYPE_RSS: + errcode = parse_rss_action(dev, attr, actions, error); + if (errcode) + return -rte_errno; + + req_act |= OTX2_FLOW_ACT_RSS; + break; + + case RTE_FLOW_ACTION_TYPE_SECURITY: + /* Assumes user has already configured security + * session for this flow. Associated conf is + * opaque. When RTE security is implemented for otx2, + * we need to verify that for specified security + * session: + * action_type == + * RTE_SECURITY_ACTION_TYPE_INLINE_PROTOCOL && + * session_protocol == + * RTE_SECURITY_PROTOCOL_IPSEC + * + * RSS is not supported with inline ipsec. Get the + * rq from associated conf, or make + * RTE_FLOW_ACTION_TYPE_QUEUE compulsory with this + * action. + * Currently, rq = 0 is assumed. + */ + req_act |= OTX2_FLOW_ACT_SEC; + rq = 0; + break; + default: + errmsg = "Unsupported action specified"; + errcode = ENOTSUP; + goto err_exit; + } + } + + /* Check if actions specified are compatible */ + if (attr->egress) { + /* Only DROP/COUNT is supported */ + if (!(req_act & OTX2_FLOW_ACT_DROP)) { + errmsg = "DROP is required action for egress"; + errcode = EINVAL; + goto err_exit; + } else if (req_act & ~(OTX2_FLOW_ACT_DROP | + OTX2_FLOW_ACT_COUNT)) { + errmsg = "Unsupported action specified"; + errcode = ENOTSUP; + goto err_exit; + } + flow->npc_action = NIX_TX_ACTIONOP_DROP; + goto set_pf_func; + } + + /* We have already verified the attr, this is ingress. + * - Exactly one terminating action is supported + * - Exactly one of MARK or FLAG is supported + * - If terminating action is DROP, only count is valid. + */ + sel_act = req_act & OTX2_FLOW_ACT_TERM; + if ((sel_act & (sel_act - 1)) != 0) { + errmsg = "Only one terminating action supported"; + errcode = EINVAL; + goto err_exit; + } + + if (req_act & OTX2_FLOW_ACT_DROP) { + sel_act = req_act & ~OTX2_FLOW_ACT_COUNT; + if ((sel_act & (sel_act - 1)) != 0) { + errmsg = "Only COUNT action is supported " + "with DROP ingress action"; + errcode = ENOTSUP; + goto err_exit; + } + } + + if ((req_act & (OTX2_FLOW_ACT_FLAG | OTX2_FLOW_ACT_MARK)) + == (OTX2_FLOW_ACT_FLAG | OTX2_FLOW_ACT_MARK)) { + errmsg = "Only one of FLAG or MARK action is supported"; + errcode = ENOTSUP; + goto err_exit; + } + + /* Set NIX_RX_ACTIONOP */ + if (req_act & OTX2_FLOW_ACT_DROP) { + flow->npc_action = NIX_RX_ACTIONOP_DROP; + } else if (req_act & OTX2_FLOW_ACT_QUEUE) { + flow->npc_action = NIX_RX_ACTIONOP_UCAST; + flow->npc_action |= (uint64_t)rq << 20; + } else if (req_act & OTX2_FLOW_ACT_RSS) { + /* When user added a rule for rss, first we will add the + *rule in MCAM and then update the action, once if we have + *FLOW_KEY_ALG index. So, till we update the action with + *flow_key_alg index, set the action to drop. + */ + if (dev->data->dev_conf.rxmode.mq_mode == ETH_MQ_RX_RSS) + flow->npc_action = NIX_RX_ACTIONOP_DROP; + else + flow->npc_action = NIX_RX_ACTIONOP_UCAST; + } else if (req_act & OTX2_FLOW_ACT_SEC) { + flow->npc_action = NIX_RX_ACTIONOP_UCAST_IPSEC; + flow->npc_action |= (uint64_t)rq << 20; + } else if (req_act & (OTX2_FLOW_ACT_FLAG | OTX2_FLOW_ACT_MARK)) { + flow->npc_action = NIX_RX_ACTIONOP_UCAST; + } else if (req_act & OTX2_FLOW_ACT_COUNT) { + /* Keep OTX2_FLOW_ACT_COUNT always at the end + * This is default action, when user specify only + * COUNT ACTION + */ + flow->npc_action = NIX_RX_ACTIONOP_UCAST; + } else { + /* Should never reach here */ + errmsg = "Invalid action specified"; + errcode = EINVAL; + goto err_exit; + } + + if (mark) + flow->npc_action |= (uint64_t)mark << 40; + + if (rte_atomic32_read(&npc->mark_actions) == 1) + hw->rx_offload_flags |= + NIX_RX_OFFLOAD_MARK_UPDATE_F; + +set_pf_func: + /* Ideally AF must ensure that correct pf_func is set */ + pf_func = otx2_pfvf_func(hw->pf, hw->vf); + flow->npc_action |= (uint64_t)pf_func << 4; + + return 0; + +err_exit: + rte_flow_error_set(error, errcode, + RTE_FLOW_ERROR_TYPE_ACTION_NUM, NULL, + errmsg); + return -rte_errno; +} diff --git a/drivers/net/octeontx2/otx2_rx.h b/drivers/net/octeontx2/otx2_rx.h index 0c3627c12..db79451b9 100644 --- a/drivers/net/octeontx2/otx2_rx.h +++ b/drivers/net/octeontx2/otx2_rx.h @@ -13,6 +13,7 @@ sizeof(uint16_t)) #define NIX_RX_OFFLOAD_PTYPE_F BIT(1) +#define NIX_RX_OFFLOAD_MARK_UPDATE_F BIT(4) #define NIX_RX_OFFLOAD_TSTAMP_F BIT(5) #define NIX_TIMESYNC_RX_OFFSET 8 -- 2.21.0