Patch Detail
get:
Show a patch.
patch:
Update a patch.
put:
Update a patch.
GET /api/patches/89775/?format=api
{ "id": 89775, "url": "https://patches.dpdk.org/api/patches/89775/?format=api", "web_url": "https://patches.dpdk.org/project/dpdk/patch/20210324192830.5608-5-pbhagavatula@marvell.com/", "project": { "id": 1, "url": "https://patches.dpdk.org/api/projects/1/?format=api", "name": "DPDK", "link_name": "dpdk", "list_id": "dev.dpdk.org", "list_email": "dev@dpdk.org", "web_url": "http://core.dpdk.org", "scm_url": "git://dpdk.org/dpdk", "webscm_url": "http://git.dpdk.org/dpdk", "list_archive_url": "https://inbox.dpdk.org/dev", "list_archive_url_format": "https://inbox.dpdk.org/dev/{}", "commit_url_format": "" }, "msgid": "<20210324192830.5608-5-pbhagavatula@marvell.com>", "list_archive_url": "https://inbox.dpdk.org/dev/20210324192830.5608-5-pbhagavatula@marvell.com", "date": "2021-03-24T19:28:25", "name": "[v6,4/8] eventdev: add Rx adapter event vector support", "commit_ref": null, "pull_url": null, "state": "superseded", "archived": true, "hash": "12d2ab56af09a236d4c2633cd195d69d93dcac27", "submitter": { "id": 1183, "url": "https://patches.dpdk.org/api/people/1183/?format=api", "name": "Pavan Nikhilesh Bhagavatula", "email": "pbhagavatula@marvell.com" }, "delegate": { "id": 310, "url": "https://patches.dpdk.org/api/users/310/?format=api", "username": "jerin", "first_name": "Jerin", "last_name": "Jacob", "email": "jerinj@marvell.com" }, "mbox": "https://patches.dpdk.org/project/dpdk/patch/20210324192830.5608-5-pbhagavatula@marvell.com/mbox/", "series": [ { "id": 15860, "url": "https://patches.dpdk.org/api/series/15860/?format=api", "web_url": "https://patches.dpdk.org/project/dpdk/list/?series=15860", "date": "2021-03-24T19:28:21", "name": "Introduce event vectorization", "version": 6, "mbox": "https://patches.dpdk.org/series/15860/mbox/" } ], "comments": "https://patches.dpdk.org/api/patches/89775/comments/", "check": "success", "checks": "https://patches.dpdk.org/api/patches/89775/checks/", "tags": {}, "related": [], "headers": { "Return-Path": "<dev-bounces@dpdk.org>", "X-Original-To": "patchwork@inbox.dpdk.org", "Delivered-To": "patchwork@inbox.dpdk.org", "Received": [ "from mails.dpdk.org (mails.dpdk.org [217.70.189.124])\n\tby inbox.dpdk.org (Postfix) with ESMTP id 6ED99A0A02;\n\tWed, 24 Mar 2021 20:29:10 +0100 (CET)", "from [217.70.189.124] (localhost [127.0.0.1])\n\tby mails.dpdk.org (Postfix) with ESMTP id E80A5140FF4;\n\tWed, 24 Mar 2021 20:29:03 +0100 (CET)", "from mx0b-0016f401.pphosted.com (mx0b-0016f401.pphosted.com\n [67.231.156.173])\n by mails.dpdk.org (Postfix) with ESMTP id B387E140FF3\n for <dev@dpdk.org>; Wed, 24 Mar 2021 20:29:01 +0100 (CET)", "from pps.filterd (m0045851.ppops.net [127.0.0.1])\n by mx0b-0016f401.pphosted.com (8.16.0.43/8.16.0.43) with SMTP id\n 12OJK4dU018664; Wed, 24 Mar 2021 12:29:01 -0700", "from dc5-exch02.marvell.com ([199.233.59.182])\n by mx0b-0016f401.pphosted.com with ESMTP id 37dgjp65gb-1\n (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT);\n Wed, 24 Mar 2021 12:29:00 -0700", "from DC5-EXCH01.marvell.com (10.69.176.38) by DC5-EXCH02.marvell.com\n (10.69.176.39) with Microsoft SMTP Server (TLS) id 15.0.1497.2;\n Wed, 24 Mar 2021 12:28:58 -0700", "from maili.marvell.com (10.69.176.80) by DC5-EXCH01.marvell.com\n (10.69.176.38) with Microsoft SMTP Server id 15.0.1497.2 via Frontend\n Transport; Wed, 24 Mar 2021 12:28:58 -0700", "from BG-LT7430.marvell.com (BG-LT7430.marvell.com [10.28.177.176])\n by maili.marvell.com (Postfix) with ESMTP id CF8EE3F703F;\n Wed, 24 Mar 2021 12:28:54 -0700 (PDT)" ], "DKIM-Signature": "v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com;\n h=from : to : cc :\n subject : date : message-id : in-reply-to : references : mime-version :\n content-transfer-encoding : content-type; s=pfpt0220;\n bh=UWn9pMnuFBDxQuFkikOinmCUSsr2l31ZLxR8q46kR30=;\n b=Dku5AWTnnhnizOQflU8WTIbslDcmcQ06IOSs1syqZh96VXu3WuWxU7vIhUi9QZ7Fm5D1\n TZpvdGEyrSeJrAfJTRch7W+ePGqqvN6jfQALwjZFjKcH7NQ2TeQcMTqRd66n4mNHMUZq\n 68KjWjRKV80rGASuhEd2Lgn4ZmCP77LbLjhgYh1D8a2GrPTPHRQMS3KSMXWeTQR6ee1N\n bGTJ5k/u3vSpU8pOb6YwdXYn2SQ/qkFdLNERK8xsFbuQjYNrqjKS2CHqxuCae0oNg6Q6\n JTV12+g1IFQ8+zKaevSIhLfgkfJINdr1WPORAC7v/rDMNCbQYNT4PSYz0nKGCcR6TaQ3 nA==", "From": "<pbhagavatula@marvell.com>", "To": "<jerinj@marvell.com>, <jay.jayatheerthan@intel.com>,\n <erik.g.carrillo@intel.com>, <abhinandan.gujjar@intel.com>,\n <timothy.mcdaniel@intel.com>, <hemant.agrawal@nxp.com>,\n <harry.van.haaren@intel.com>, <mattias.ronnblom@ericsson.com>,\n <liang.j.ma@intel.com>", "CC": "<dev@dpdk.org>, Pavan Nikhilesh <pbhagavatula@marvell.com>", "Date": "Thu, 25 Mar 2021 00:58:25 +0530", "Message-ID": "<20210324192830.5608-5-pbhagavatula@marvell.com>", "X-Mailer": "git-send-email 2.17.1", "In-Reply-To": "<20210324192830.5608-1-pbhagavatula@marvell.com>", "References": "<20210324050525.4489-1-pbhagavatula@marvell.com>\n <20210324192830.5608-1-pbhagavatula@marvell.com>", "MIME-Version": "1.0", "Content-Transfer-Encoding": "8bit", "Content-Type": "text/plain", "X-Proofpoint-Virus-Version": "vendor=fsecure engine=2.50.10434:6.0.369, 18.0.761\n definitions=2021-03-24_13:2021-03-24,\n 2021-03-24 signatures=0", "Subject": "[dpdk-dev] [PATCH v6 4/8] eventdev: add Rx adapter event vector\n support", "X-BeenThere": "dev@dpdk.org", "X-Mailman-Version": "2.1.29", "Precedence": "list", "List-Id": "DPDK patches and discussions <dev.dpdk.org>", "List-Unsubscribe": "<https://mails.dpdk.org/options/dev>,\n <mailto:dev-request@dpdk.org?subject=unsubscribe>", "List-Archive": "<http://mails.dpdk.org/archives/dev/>", "List-Post": "<mailto:dev@dpdk.org>", "List-Help": "<mailto:dev-request@dpdk.org?subject=help>", "List-Subscribe": "<https://mails.dpdk.org/listinfo/dev>,\n <mailto:dev-request@dpdk.org?subject=subscribe>", "Errors-To": "dev-bounces@dpdk.org", "Sender": "\"dev\" <dev-bounces@dpdk.org>" }, "content": "From: Pavan Nikhilesh <pbhagavatula@marvell.com>\n\nAdd event vector support for event eth Rx adapter, the implementation\ncreates vector flows based on port and queue identifier of the received\nmbufs.\n\nSigned-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>\n---\n lib/librte_eventdev/eventdev_pmd.h | 7 +-\n .../rte_event_eth_rx_adapter.c | 257 ++++++++++++++++--\n lib/librte_eventdev/rte_eventdev.c | 6 +-\n 3 files changed, 250 insertions(+), 20 deletions(-)", "diff": "diff --git a/lib/librte_eventdev/eventdev_pmd.h b/lib/librte_eventdev/eventdev_pmd.h\nindex 9297f1433..0f724ac85 100644\n--- a/lib/librte_eventdev/eventdev_pmd.h\n+++ b/lib/librte_eventdev/eventdev_pmd.h\n@@ -69,9 +69,10 @@ extern \"C\" {\n \t} \\\n } while (0)\n \n-#define RTE_EVENT_ETH_RX_ADAPTER_SW_CAP \\\n-\t\t((RTE_EVENT_ETH_RX_ADAPTER_CAP_OVERRIDE_FLOW_ID) | \\\n-\t\t\t(RTE_EVENT_ETH_RX_ADAPTER_CAP_MULTI_EVENTQ))\n+#define RTE_EVENT_ETH_RX_ADAPTER_SW_CAP \\\n+\t((RTE_EVENT_ETH_RX_ADAPTER_CAP_OVERRIDE_FLOW_ID) | \\\n+\t (RTE_EVENT_ETH_RX_ADAPTER_CAP_MULTI_EVENTQ) | \\\n+\t (RTE_EVENT_ETH_RX_ADAPTER_CAP_EVENT_VECTOR))\n \n #define RTE_EVENT_CRYPTO_ADAPTER_SW_CAP \\\n \t\tRTE_EVENT_CRYPTO_ADAPTER_CAP_SESSION_PRIVATE_DATA\ndiff --git a/lib/librte_eventdev/rte_event_eth_rx_adapter.c b/lib/librte_eventdev/rte_event_eth_rx_adapter.c\nindex ac8ba5bf0..c71990078 100644\n--- a/lib/librte_eventdev/rte_event_eth_rx_adapter.c\n+++ b/lib/librte_eventdev/rte_event_eth_rx_adapter.c\n@@ -26,6 +26,10 @@\n #define BATCH_SIZE\t\t32\n #define BLOCK_CNT_THRESHOLD\t10\n #define ETH_EVENT_BUFFER_SIZE\t(4*BATCH_SIZE)\n+#define MAX_VECTOR_SIZE\t\t1024\n+#define MIN_VECTOR_SIZE\t\t4\n+#define MAX_VECTOR_NS\t\t1E9\n+#define MIN_VECTOR_NS\t\t1E5\n \n #define ETH_RX_ADAPTER_SERVICE_NAME_LEN\t32\n #define ETH_RX_ADAPTER_MEM_NAME_LEN\t32\n@@ -59,6 +63,20 @@ struct eth_rx_poll_entry {\n \tuint16_t eth_rx_qid;\n };\n \n+struct eth_rx_vector_data {\n+\tTAILQ_ENTRY(eth_rx_vector_data) next;\n+\tuint16_t port;\n+\tuint16_t queue;\n+\tuint16_t max_vector_count;\n+\tuint64_t event;\n+\tuint64_t ts;\n+\tuint64_t vector_timeout_ticks;\n+\tstruct rte_mempool *vector_pool;\n+\tstruct rte_event_vector *vector_ev;\n+} __rte_cache_aligned;\n+\n+TAILQ_HEAD(eth_rx_vector_data_list, eth_rx_vector_data);\n+\n /* Instance per adapter */\n struct rte_eth_event_enqueue_buffer {\n \t/* Count of events in this buffer */\n@@ -92,6 +110,14 @@ struct rte_event_eth_rx_adapter {\n \tuint32_t wrr_pos;\n \t/* Event burst buffer */\n \tstruct rte_eth_event_enqueue_buffer event_enqueue_buffer;\n+\t/* Vector enable flag */\n+\tuint8_t ena_vector;\n+\t/* Timestamp of previous vector expiry list traversal */\n+\tuint64_t prev_expiry_ts;\n+\t/* Minimum ticks to wait before traversing expiry list */\n+\tuint64_t vector_tmo_ticks;\n+\t/* vector list */\n+\tstruct eth_rx_vector_data_list vector_list;\n \t/* Per adapter stats */\n \tstruct rte_event_eth_rx_adapter_stats stats;\n \t/* Block count, counts up to BLOCK_CNT_THRESHOLD */\n@@ -198,9 +224,11 @@ struct eth_device_info {\n struct eth_rx_queue_info {\n \tint queue_enabled;\t/* True if added */\n \tint intr_enabled;\n+\tuint8_t ena_vector;\n \tuint16_t wt;\t\t/* Polling weight */\n \tuint32_t flow_id_mask;\t/* Set to ~0 if app provides flow id else 0 */\n \tuint64_t event;\n+\tstruct eth_rx_vector_data vector_data;\n };\n \n static struct rte_event_eth_rx_adapter **event_eth_rx_adapter;\n@@ -722,6 +750,9 @@ rxa_flush_event_buffer(struct rte_event_eth_rx_adapter *rx_adapter)\n \t &rx_adapter->event_enqueue_buffer;\n \tstruct rte_event_eth_rx_adapter_stats *stats = &rx_adapter->stats;\n \n+\tif (!buf->count)\n+\t\treturn 0;\n+\n \tuint16_t n = rte_event_enqueue_new_burst(rx_adapter->eventdev_id,\n \t\t\t\t\trx_adapter->event_port_id,\n \t\t\t\t\tbuf->events,\n@@ -742,6 +773,72 @@ rxa_flush_event_buffer(struct rte_event_eth_rx_adapter *rx_adapter)\n \treturn n;\n }\n \n+static inline uint16_t\n+rxa_create_event_vector(struct rte_event_eth_rx_adapter *rx_adapter,\n+\t\t\tstruct eth_rx_queue_info *queue_info,\n+\t\t\tstruct rte_eth_event_enqueue_buffer *buf,\n+\t\t\tstruct rte_mbuf **mbufs, uint16_t num)\n+{\n+\tstruct rte_event *ev = &buf->events[buf->count];\n+\tstruct eth_rx_vector_data *vec;\n+\tuint16_t filled, space, sz;\n+\n+\tfilled = 0;\n+\tvec = &queue_info->vector_data;\n+\twhile (num) {\n+\t\tif (vec->vector_ev == NULL) {\n+\t\t\tif (rte_mempool_get(vec->vector_pool,\n+\t\t\t\t\t (void **)&vec->vector_ev) < 0) {\n+\t\t\t\trte_pktmbuf_free_bulk(mbufs, num);\n+\t\t\t\treturn 0;\n+\t\t\t}\n+\t\t\tvec->vector_ev->nb_elem = 0;\n+\t\t\tvec->vector_ev->port = vec->port;\n+\t\t\tvec->vector_ev->queue = vec->queue;\n+\t\t\tvec->vector_ev->attr_valid = true;\n+\t\t\tTAILQ_INSERT_TAIL(&rx_adapter->vector_list, vec, next);\n+\t\t} else if (vec->vector_ev->nb_elem == vec->max_vector_count) {\n+\t\t\t/* Event ready. */\n+\t\t\tev->event = vec->event;\n+\t\t\tev->vec = vec->vector_ev;\n+\t\t\tev++;\n+\t\t\tfilled++;\n+\t\t\tvec->vector_ev = NULL;\n+\t\t\tTAILQ_REMOVE(&rx_adapter->vector_list, vec, next);\n+\t\t\tif (rte_mempool_get(vec->vector_pool,\n+\t\t\t\t\t (void **)&vec->vector_ev) < 0) {\n+\t\t\t\trte_pktmbuf_free_bulk(mbufs, num);\n+\t\t\t\treturn 0;\n+\t\t\t}\n+\t\t\tvec->vector_ev->nb_elem = 0;\n+\t\t\tvec->vector_ev->port = vec->port;\n+\t\t\tvec->vector_ev->queue = vec->queue;\n+\t\t\tvec->vector_ev->attr_valid = true;\n+\t\t\tTAILQ_INSERT_TAIL(&rx_adapter->vector_list, vec, next);\n+\t\t}\n+\n+\t\tspace = vec->max_vector_count - vec->vector_ev->nb_elem;\n+\t\tsz = num > space ? space : num;\n+\t\tmemcpy(vec->vector_ev->mbufs + vec->vector_ev->nb_elem, mbufs,\n+\t\t sizeof(void *) * sz);\n+\t\tvec->vector_ev->nb_elem += sz;\n+\t\tnum -= sz;\n+\t\tmbufs += sz;\n+\t\tvec->ts = rte_rdtsc();\n+\t}\n+\n+\tif (vec->vector_ev->nb_elem == vec->max_vector_count) {\n+\t\tev->event = vec->event;\n+\t\tev->vec = vec->vector_ev;\n+\t\tev++;\n+\t\tfilled++;\n+\t\tvec->vector_ev = NULL;\n+\t\tTAILQ_REMOVE(&rx_adapter->vector_list, vec, next);\n+\t}\n+\n+\treturn filled;\n+}\n+\n static inline void\n rxa_buffer_mbufs(struct rte_event_eth_rx_adapter *rx_adapter,\n \t\tuint16_t eth_dev_id,\n@@ -770,25 +867,30 @@ rxa_buffer_mbufs(struct rte_event_eth_rx_adapter *rx_adapter,\n \trss_mask = ~(((m->ol_flags & PKT_RX_RSS_HASH) != 0) - 1);\n \tdo_rss = !rss_mask && !eth_rx_queue_info->flow_id_mask;\n \n-\tfor (i = 0; i < num; i++) {\n-\t\tm = mbufs[i];\n-\n-\t\trss = do_rss ?\n-\t\t\trxa_do_softrss(m, rx_adapter->rss_key_be) :\n-\t\t\tm->hash.rss;\n-\t\tev->event = event;\n-\t\tev->flow_id = (rss & ~flow_id_mask) |\n-\t\t\t\t(ev->flow_id & flow_id_mask);\n-\t\tev->mbuf = m;\n-\t\tev++;\n+\tif (!eth_rx_queue_info->ena_vector) {\n+\t\tfor (i = 0; i < num; i++) {\n+\t\t\tm = mbufs[i];\n+\n+\t\t\trss = do_rss ? rxa_do_softrss(m, rx_adapter->rss_key_be)\n+\t\t\t\t : m->hash.rss;\n+\t\t\tev->event = event;\n+\t\t\tev->flow_id = (rss & ~flow_id_mask) |\n+\t\t\t\t (ev->flow_id & flow_id_mask);\n+\t\t\tev->mbuf = m;\n+\t\t\tev++;\n+\t\t}\n+\t} else {\n+\t\tnum = rxa_create_event_vector(rx_adapter, eth_rx_queue_info,\n+\t\t\t\t\t buf, mbufs, num);\n \t}\n \n-\tif (dev_info->cb_fn) {\n+\tif (num && dev_info->cb_fn) {\n \n \t\tdropped = 0;\n \t\tnb_cb = dev_info->cb_fn(eth_dev_id, rx_queue_id,\n-\t\t\t\t\tETH_EVENT_BUFFER_SIZE, buf->count, ev,\n-\t\t\t\t\tnum, dev_info->cb_arg, &dropped);\n+\t\t\t\t\tETH_EVENT_BUFFER_SIZE, buf->count,\n+\t\t\t\t\t&buf->events[buf->count], num,\n+\t\t\t\t\tdev_info->cb_arg, &dropped);\n \t\tif (unlikely(nb_cb > num))\n \t\t\tRTE_EDEV_LOG_ERR(\"Rx CB returned %d (> %d) events\",\n \t\t\t\tnb_cb, num);\n@@ -1124,6 +1226,30 @@ rxa_poll(struct rte_event_eth_rx_adapter *rx_adapter)\n \treturn nb_rx;\n }\n \n+static void\n+rxa_vector_expire(struct eth_rx_vector_data *vec, void *arg)\n+{\n+\tstruct rte_event_eth_rx_adapter *rx_adapter = arg;\n+\tstruct rte_eth_event_enqueue_buffer *buf =\n+\t\t&rx_adapter->event_enqueue_buffer;\n+\tstruct rte_event *ev;\n+\n+\tif (buf->count)\n+\t\trxa_flush_event_buffer(rx_adapter);\n+\n+\tif (vec->vector_ev->nb_elem == 0)\n+\t\treturn;\n+\tev = &buf->events[buf->count];\n+\n+\t/* Event ready. */\n+\tev->event = vec->event;\n+\tev->vec = vec->vector_ev;\n+\tbuf->count++;\n+\n+\tvec->vector_ev = NULL;\n+\tvec->ts = 0;\n+}\n+\n static int\n rxa_service_func(void *args)\n {\n@@ -1137,6 +1263,24 @@ rxa_service_func(void *args)\n \t\treturn 0;\n \t}\n \n+\tif (rx_adapter->ena_vector) {\n+\t\tif ((rte_rdtsc() - rx_adapter->prev_expiry_ts) >=\n+\t\t rx_adapter->vector_tmo_ticks) {\n+\t\t\tstruct eth_rx_vector_data *vec;\n+\n+\t\t\tTAILQ_FOREACH(vec, &rx_adapter->vector_list, next) {\n+\t\t\t\tuint64_t elapsed_time = rte_rdtsc() - vec->ts;\n+\n+\t\t\t\tif (elapsed_time >= vec->vector_timeout_ticks) {\n+\t\t\t\t\trxa_vector_expire(vec, rx_adapter);\n+\t\t\t\t\tTAILQ_REMOVE(&rx_adapter->vector_list,\n+\t\t\t\t\t\t vec, next);\n+\t\t\t\t}\n+\t\t\t}\n+\t\t\trx_adapter->prev_expiry_ts = rte_rdtsc();\n+\t\t}\n+\t}\n+\n \tstats = &rx_adapter->stats;\n \tstats->rx_packets += rxa_intr_ring_dequeue(rx_adapter);\n \tstats->rx_packets += rxa_poll(rx_adapter);\n@@ -1640,6 +1784,28 @@ rxa_update_queue(struct rte_event_eth_rx_adapter *rx_adapter,\n \t}\n }\n \n+static void\n+rxa_set_vector_data(struct eth_rx_queue_info *queue_info, uint16_t vector_count,\n+\t\t uint64_t vector_ns, struct rte_mempool *mp, int32_t qid,\n+\t\t uint16_t port_id)\n+{\n+#define NSEC2TICK(__ns, __freq) (((__ns) * (__freq)) / 1E9)\n+\tstruct eth_rx_vector_data *vector_data;\n+\tuint32_t flow_id;\n+\n+\tvector_data = &queue_info->vector_data;\n+\tvector_data->max_vector_count = vector_count;\n+\tvector_data->port = port_id;\n+\tvector_data->queue = qid;\n+\tvector_data->vector_pool = mp;\n+\tvector_data->vector_timeout_ticks =\n+\t\tNSEC2TICK(vector_ns, rte_get_timer_hz());\n+\tvector_data->ts = 0;\n+\tflow_id = queue_info->event & 0xFFFFF;\n+\tflow_id = flow_id == 0 ? (qid & 0xFF) | (port_id & 0xFFFF) : flow_id;\n+\tvector_data->event = (queue_info->event & ~0xFFFFF) | flow_id;\n+}\n+\n static void\n rxa_sw_del(struct rte_event_eth_rx_adapter *rx_adapter,\n \tstruct eth_device_info *dev_info,\n@@ -1741,6 +1907,44 @@ rxa_add_queue(struct rte_event_eth_rx_adapter *rx_adapter,\n \t}\n }\n \n+static void\n+rxa_sw_event_vector_configure(\n+\tstruct rte_event_eth_rx_adapter *rx_adapter, uint16_t eth_dev_id,\n+\tint rx_queue_id,\n+\tconst struct rte_event_eth_rx_adapter_event_vector_config *config)\n+{\n+\tstruct eth_device_info *dev_info = &rx_adapter->eth_devices[eth_dev_id];\n+\tstruct eth_rx_queue_info *queue_info;\n+\tstruct rte_event *qi_ev;\n+\n+\tif (rx_queue_id == -1) {\n+\t\tuint16_t nb_rx_queues;\n+\t\tuint16_t i;\n+\n+\t\tnb_rx_queues = dev_info->dev->data->nb_rx_queues;\n+\t\tfor (i = 0; i < nb_rx_queues; i++)\n+\t\t\trxa_sw_event_vector_configure(rx_adapter, eth_dev_id, i,\n+\t\t\t\t\t\t config);\n+\t\treturn;\n+\t}\n+\n+\tqueue_info = &dev_info->rx_queue[rx_queue_id];\n+\tqi_ev = (struct rte_event *)&queue_info->event;\n+\tqueue_info->ena_vector = 1;\n+\tqi_ev->event_type = RTE_EVENT_TYPE_ETH_RX_ADAPTER_VECTOR;\n+\trxa_set_vector_data(queue_info, config->vector_sz,\n+\t\t\t config->vector_timeout_ns, config->vector_mp,\n+\t\t\t rx_queue_id, dev_info->dev->data->port_id);\n+\trx_adapter->ena_vector = 1;\n+\trx_adapter->vector_tmo_ticks =\n+\t\trx_adapter->vector_tmo_ticks ?\n+\t\t\t RTE_MIN(config->vector_timeout_ns << 1,\n+\t\t\t\t rx_adapter->vector_tmo_ticks) :\n+\t\t\t config->vector_timeout_ns << 1;\n+\trx_adapter->prev_expiry_ts = 0;\n+\tTAILQ_INIT(&rx_adapter->vector_list);\n+}\n+\n static int rxa_sw_add(struct rte_event_eth_rx_adapter *rx_adapter,\n \t\tuint16_t eth_dev_id,\n \t\tint rx_queue_id,\n@@ -2081,6 +2285,15 @@ rte_event_eth_rx_adapter_queue_add(uint8_t id,\n \t\treturn -EINVAL;\n \t}\n \n+\tif ((cap & RTE_EVENT_ETH_RX_ADAPTER_CAP_EVENT_VECTOR) == 0 &&\n+\t (queue_conf->rx_queue_flags &\n+\t RTE_EVENT_ETH_RX_ADAPTER_QUEUE_EVENT_VECTOR)) {\n+\t\tRTE_EDEV_LOG_ERR(\"Event vectorization is not supported,\"\n+\t\t\t\t \" eth port: %\" PRIu16 \" adapter id: %\" PRIu8,\n+\t\t\t\t eth_dev_id, id);\n+\t\treturn -EINVAL;\n+\t}\n+\n \tif ((cap & RTE_EVENT_ETH_RX_ADAPTER_CAP_MULTI_EVENTQ) == 0 &&\n \t\t(rx_queue_id != -1)) {\n \t\tRTE_EDEV_LOG_ERR(\"Rx queues can only be connected to single \"\n@@ -2143,6 +2356,17 @@ rte_event_eth_rx_adapter_queue_add(uint8_t id,\n \treturn 0;\n }\n \n+static int\n+rxa_sw_vector_limits(struct rte_event_eth_rx_adapter_vector_limits *limits)\n+{\n+\tlimits->max_sz = MAX_VECTOR_SIZE;\n+\tlimits->min_sz = MIN_VECTOR_SIZE;\n+\tlimits->max_timeout_ns = MAX_VECTOR_NS;\n+\tlimits->min_timeout_ns = MIN_VECTOR_NS;\n+\n+\treturn 0;\n+}\n+\n int\n rte_event_eth_rx_adapter_queue_del(uint8_t id, uint16_t eth_dev_id,\n \t\t\t\tint32_t rx_queue_id)\n@@ -2333,7 +2557,8 @@ rte_event_eth_rx_adapter_queue_event_vector_config(\n \t\tret = dev->dev_ops->eth_rx_adapter_event_vector_config(\n \t\t\tdev, &rte_eth_devices[eth_dev_id], rx_queue_id, config);\n \t} else {\n-\t\tret = -ENOTSUP;\n+\t\trxa_sw_event_vector_configure(rx_adapter, eth_dev_id,\n+\t\t\t\t\t rx_queue_id, config);\n \t}\n \n \treturn ret;\n@@ -2371,7 +2596,7 @@ rte_event_eth_rx_adapter_vector_limits_get(\n \t\tret = dev->dev_ops->eth_rx_adapter_vector_limits_get(\n \t\t\tdev, &rte_eth_devices[eth_port_id], limits);\n \t} else {\n-\t\tret = -ENOTSUP;\n+\t\tret = rxa_sw_vector_limits(limits);\n \t}\n \n \treturn ret;\ndiff --git a/lib/librte_eventdev/rte_eventdev.c b/lib/librte_eventdev/rte_eventdev.c\nindex be0499c52..62824654b 100644\n--- a/lib/librte_eventdev/rte_eventdev.c\n+++ b/lib/librte_eventdev/rte_eventdev.c\n@@ -122,7 +122,11 @@ rte_event_eth_rx_adapter_caps_get(uint8_t dev_id, uint16_t eth_port_id,\n \n \tif (caps == NULL)\n \t\treturn -EINVAL;\n-\t*caps = 0;\n+\n+\tif (dev->dev_ops->eth_rx_adapter_caps_get == NULL)\n+\t\t*caps = RTE_EVENT_ETH_RX_ADAPTER_SW_CAP;\n+\telse\n+\t\t*caps = 0;\n \n \treturn dev->dev_ops->eth_rx_adapter_caps_get ?\n \t\t\t\t(*dev->dev_ops->eth_rx_adapter_caps_get)(dev,\n", "prefixes": [ "v6", "4/8" ] }