From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 60C7AA0A02; Thu, 20 Oct 2022 08:33:33 +0200 (CEST) Received: from [217.70.189.124] (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 1E16A42CF0; Thu, 20 Oct 2022 08:32:10 +0200 (CEST) Received: from mga06.intel.com (mga06b.intel.com [134.134.136.31]) by mails.dpdk.org (Postfix) with ESMTP id 9FB4042CC0 for ; Thu, 20 Oct 2022 08:32:05 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1666247525; x=1697783525; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=QZzYqwQ20o0ovYo3itvRmiiICGlXkgcx591a0TZpLM4=; b=Cb8g9oqPCJcZU2FHNhYHTdRlfRHdSU0SAqpD73wxCK4R/ROOPHJD8pZy 3zZfq0FLfUT048hoeRDBk/xIVNGz12WUS2Pp3Gm7SR5w5AvolfgmZ7eTV f6j0DF9EjFzU5zOYw5FZOD+AnEeoJkiMYCT8PKq6xdlBWDoXlTOsL0Jlf HdAgjx8RGdNln5DFxhg4K2IPnJ38h9zs7qbcZOn5Jn9lBvyf0jEoXuf/t NGxlwNMRKzmvkb6AU9Ie3T83qc8kxoqQ8oLaK60ovWYfP1CKEYkMHMOC1 zPn024hCDPU0MudTGMU+41ypgbvDO4g17ufIh5j+FuI8e9ercyszY5H0v A==; X-IronPort-AV: E=McAfee;i="6500,9779,10505"; a="368667054" X-IronPort-AV: E=Sophos;i="5.95,198,1661842800"; d="scan'208";a="368667054" Received: from orsmga006.jf.intel.com ([10.7.209.51]) by orsmga104.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 19 Oct 2022 23:32:05 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=McAfee;i="6500,9779,10505"; a="607481563" X-IronPort-AV: E=Sophos;i="5.95,198,1661842800"; d="scan'208";a="607481563" Received: from dpdk-jf-ntb-one.sh.intel.com ([10.67.111.104]) by orsmga006.jf.intel.com with ESMTP; 19 Oct 2022 23:32:02 -0700 From: Junfeng Guo To: andrew.rybchenko@oktetlabs.ru, qi.z.zhang@intel.com, jingjing.wu@intel.com, beilei.xing@intel.com Cc: dev@dpdk.org, Junfeng Guo , Xiaoyun Li Subject: [PATCH v8 08/14] net/idpf: add support for basic Rx/Tx datapath Date: Thu, 20 Oct 2022 14:29:45 +0800 Message-Id: <20221020062951.645121-9-junfeng.guo@intel.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20221020062951.645121-1-junfeng.guo@intel.com> References: <20221020024135.338280-15-junfeng.guo@intel.com> <20221020062951.645121-1-junfeng.guo@intel.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Add basic Rx & Tx support in split queue mode and single queue mode. Split queue mode is selected by default. Signed-off-by: Beilei Xing Signed-off-by: Xiaoyun Li Signed-off-by: Junfeng Guo --- drivers/net/idpf/idpf_ethdev.c | 8 +- drivers/net/idpf/idpf_rxtx.c | 667 +++++++++++++++++++++++++++++++++ drivers/net/idpf/idpf_rxtx.h | 16 +- 3 files changed, 687 insertions(+), 4 deletions(-) diff --git a/drivers/net/idpf/idpf_ethdev.c b/drivers/net/idpf/idpf_ethdev.c index 20f0f39640..c6d6c87274 100644 --- a/drivers/net/idpf/idpf_ethdev.c +++ b/drivers/net/idpf/idpf_ethdev.c @@ -313,6 +313,9 @@ idpf_dev_start(struct rte_eth_dev *dev) goto err_mtu; } + idpf_set_rx_function(dev); + idpf_set_tx_function(dev); + if (idpf_vc_ena_dis_vport(vport, true)) { PMD_DRV_LOG(ERR, "Failed to enable vport"); goto err_vport; @@ -764,8 +767,11 @@ idpf_dev_init(struct rte_eth_dev *dev, void *init_params) /* for secondary processes, we don't initialise any further as primary * has already done this work. */ - if (rte_eal_process_type() != RTE_PROC_PRIMARY) + if (rte_eal_process_type() != RTE_PROC_PRIMARY) { + idpf_set_rx_function(dev); + idpf_set_tx_function(dev); return ret; + } dev->data->dev_flags |= RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS; diff --git a/drivers/net/idpf/idpf_rxtx.c b/drivers/net/idpf/idpf_rxtx.c index e92ea09c3b..6280b7887b 100644 --- a/drivers/net/idpf/idpf_rxtx.c +++ b/drivers/net/idpf/idpf_rxtx.c @@ -1268,3 +1268,670 @@ idpf_stop_queues(struct rte_eth_dev *dev) PMD_DRV_LOG(WARNING, "Fail to stop Tx queue %d", i); } } + +static void +idpf_split_rx_bufq_refill(struct idpf_rx_queue *rx_bufq) +{ + volatile struct virtchnl2_splitq_rx_buf_desc *rx_buf_ring; + volatile struct virtchnl2_splitq_rx_buf_desc *rx_buf_desc; + uint16_t nb_refill = rx_bufq->rx_free_thresh; + uint16_t nb_desc = rx_bufq->nb_rx_desc; + uint16_t next_avail = rx_bufq->rx_tail; + struct rte_mbuf *nmb[rx_bufq->rx_free_thresh]; + struct rte_eth_dev *dev; + uint64_t dma_addr; + uint16_t delta; + int i; + + if (rx_bufq->nb_rx_hold < rx_bufq->rx_free_thresh) + return; + + rx_buf_ring = + (volatile struct virtchnl2_splitq_rx_buf_desc *)rx_bufq->rx_ring; + delta = nb_desc - next_avail; + if (unlikely(delta < nb_refill)) { + if (likely(!rte_pktmbuf_alloc_bulk(rx_bufq->mp, nmb, delta))) { + for (i = 0; i < delta; i++) { + rx_buf_desc = &rx_buf_ring[next_avail + i]; + rx_bufq->sw_ring[next_avail + i] = nmb[i]; + dma_addr = rte_cpu_to_le_64(rte_mbuf_data_iova_default(nmb[i])); + rx_buf_desc->hdr_addr = 0; + rx_buf_desc->pkt_addr = dma_addr; + } + nb_refill -= delta; + next_avail = 0; + rx_bufq->nb_rx_hold -= delta; + } else { + dev = &rte_eth_devices[rx_bufq->port_id]; + dev->data->rx_mbuf_alloc_failed += nb_desc - next_avail; + PMD_RX_LOG(DEBUG, "RX mbuf alloc failed port_id=%u queue_id=%u", + rx_bufq->port_id, rx_bufq->queue_id); + return; + } + } + + if (nb_desc - next_avail >= nb_refill) { + if (likely(!rte_pktmbuf_alloc_bulk(rx_bufq->mp, nmb, nb_refill))) { + for (i = 0; i < nb_refill; i++) { + rx_buf_desc = &rx_buf_ring[next_avail + i]; + rx_bufq->sw_ring[next_avail + i] = nmb[i]; + dma_addr = rte_cpu_to_le_64(rte_mbuf_data_iova_default(nmb[i])); + rx_buf_desc->hdr_addr = 0; + rx_buf_desc->pkt_addr = dma_addr; + } + next_avail += nb_refill; + rx_bufq->nb_rx_hold -= nb_refill; + } else { + dev = &rte_eth_devices[rx_bufq->port_id]; + dev->data->rx_mbuf_alloc_failed += nb_desc - next_avail; + PMD_RX_LOG(DEBUG, "RX mbuf alloc failed port_id=%u queue_id=%u", + rx_bufq->port_id, rx_bufq->queue_id); + } + } + + IDPF_PCI_REG_WRITE(rx_bufq->qrx_tail, next_avail); + + rx_bufq->rx_tail = next_avail; +} + +uint16_t +idpf_splitq_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts) +{ + volatile struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc_ring; + volatile struct virtchnl2_rx_flex_desc_adv_nic_3 *rx_desc; + uint16_t pktlen_gen_bufq_id; + struct idpf_rx_queue *rxq; + const uint32_t *ptype_tbl; + struct rte_mbuf *rxm; + uint16_t rx_id_bufq1; + uint16_t rx_id_bufq2; + uint16_t pkt_len; + uint16_t bufq_id; + uint16_t gen_id; + uint16_t rx_id; + uint16_t nb_rx; + + nb_rx = 0; + rxq = (struct idpf_rx_queue *)rx_queue; + + if (unlikely(!rxq) || unlikely(!rxq->q_started)) + return nb_rx; + + rx_id = rxq->rx_tail; + rx_id_bufq1 = rxq->bufq1->rx_next_avail; + rx_id_bufq2 = rxq->bufq2->rx_next_avail; + rx_desc_ring = + (volatile struct virtchnl2_rx_flex_desc_adv_nic_3 *)rxq->rx_ring; + ptype_tbl = rxq->adapter->ptype_tbl; + + if (rxq->offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP) + rxq->hw_register_set = 1; + + while (nb_rx < nb_pkts) { + rx_desc = &rx_desc_ring[rx_id]; + + pktlen_gen_bufq_id = + rte_le_to_cpu_16(rx_desc->pktlen_gen_bufq_id); + gen_id = (pktlen_gen_bufq_id & + VIRTCHNL2_RX_FLEX_DESC_ADV_GEN_M) >> + VIRTCHNL2_RX_FLEX_DESC_ADV_GEN_S; + if (gen_id != rxq->expected_gen_id) + break; + + pkt_len = (pktlen_gen_bufq_id & + VIRTCHNL2_RX_FLEX_DESC_ADV_LEN_PBUF_M) >> + VIRTCHNL2_RX_FLEX_DESC_ADV_LEN_PBUF_S; + if (!pkt_len) + PMD_RX_LOG(ERR, "Packet length is 0"); + + rx_id++; + if (unlikely(rx_id == rxq->nb_rx_desc)) { + rx_id = 0; + rxq->expected_gen_id ^= 1; + } + + bufq_id = (pktlen_gen_bufq_id & + VIRTCHNL2_RX_FLEX_DESC_ADV_BUFQ_ID_M) >> + VIRTCHNL2_RX_FLEX_DESC_ADV_BUFQ_ID_S; + if (!bufq_id) { + rxm = rxq->bufq1->sw_ring[rx_id_bufq1]; + rx_id_bufq1++; + if (unlikely(rx_id_bufq1 == rxq->bufq1->nb_rx_desc)) + rx_id_bufq1 = 0; + rxq->bufq1->nb_rx_hold++; + } else { + rxm = rxq->bufq2->sw_ring[rx_id_bufq2]; + rx_id_bufq2++; + if (unlikely(rx_id_bufq2 == rxq->bufq2->nb_rx_desc)) + rx_id_bufq2 = 0; + rxq->bufq2->nb_rx_hold++; + } + + rxm->pkt_len = pkt_len; + rxm->data_len = pkt_len; + rxm->data_off = RTE_PKTMBUF_HEADROOM; + rxm->next = NULL; + rxm->nb_segs = 1; + rxm->port = rxq->port_id; + rxm->ol_flags = 0; + rxm->packet_type = + ptype_tbl[(rte_le_to_cpu_16(rx_desc->ptype_err_fflags0) & + VIRTCHNL2_RX_FLEX_DESC_ADV_PTYPE_M) >> + VIRTCHNL2_RX_FLEX_DESC_ADV_PTYPE_S]; + + rx_pkts[nb_rx++] = rxm; + } + + if (nb_rx) { + rxq->rx_tail = rx_id; + if (rx_id_bufq1 != rxq->bufq1->rx_next_avail) + rxq->bufq1->rx_next_avail = rx_id_bufq1; + if (rx_id_bufq2 != rxq->bufq2->rx_next_avail) + rxq->bufq2->rx_next_avail = rx_id_bufq2; + + idpf_split_rx_bufq_refill(rxq->bufq1); + idpf_split_rx_bufq_refill(rxq->bufq2); + } + + return nb_rx; +} + +static inline void +idpf_split_tx_free(struct idpf_tx_queue *cq) +{ + volatile struct idpf_splitq_tx_compl_desc *compl_ring = cq->compl_ring; + volatile struct idpf_splitq_tx_compl_desc *txd; + uint16_t next = cq->tx_tail; + struct idpf_tx_entry *txe; + struct idpf_tx_queue *txq; + uint16_t gen, qid, q_head; + uint8_t ctype; + + txd = &compl_ring[next]; + gen = (rte_le_to_cpu_16(txd->qid_comptype_gen) & + IDPF_TXD_COMPLQ_GEN_M) >> IDPF_TXD_COMPLQ_GEN_S; + if (gen != cq->expected_gen_id) + return; + + ctype = (rte_le_to_cpu_16(txd->qid_comptype_gen) & + IDPF_TXD_COMPLQ_COMPL_TYPE_M) >> IDPF_TXD_COMPLQ_COMPL_TYPE_S; + qid = (rte_le_to_cpu_16(txd->qid_comptype_gen) & + IDPF_TXD_COMPLQ_QID_M) >> IDPF_TXD_COMPLQ_QID_S; + q_head = rte_le_to_cpu_16(txd->q_head_compl_tag.compl_tag); + txq = cq->txqs[qid - cq->tx_start_qid]; + + switch (ctype) { + case IDPF_TXD_COMPLT_RE: + if (q_head == 0) + txq->last_desc_cleaned = txq->nb_tx_desc - 1; + else + txq->last_desc_cleaned = q_head - 1; + if (unlikely(!(txq->last_desc_cleaned % 32))) { + PMD_DRV_LOG(ERR, "unexpected desc (head = %u) completion.", + q_head); + return; + } + + break; + case IDPF_TXD_COMPLT_RS: + txq->nb_free++; + txq->nb_used--; + txe = &txq->sw_ring[q_head]; + if (txe->mbuf) { + rte_pktmbuf_free_seg(txe->mbuf); + txe->mbuf = NULL; + } + break; + default: + PMD_DRV_LOG(ERR, "unknown completion type."); + return; + } + + if (++next == cq->nb_tx_desc) { + next = 0; + cq->expected_gen_id ^= 1; + } + + cq->tx_tail = next; +} + +uint16_t +idpf_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts) +{ + struct idpf_tx_queue *txq = (struct idpf_tx_queue *)tx_queue; + volatile struct idpf_flex_tx_sched_desc *txr; + volatile struct idpf_flex_tx_sched_desc *txd; + struct idpf_tx_entry *sw_ring; + struct idpf_tx_entry *txe, *txn; + uint16_t nb_used, tx_id, sw_id; + struct rte_mbuf *tx_pkt; + uint16_t nb_to_clean; + uint16_t nb_tx = 0; + + if (unlikely(!txq) || unlikely(!txq->q_started)) + return nb_tx; + + txr = txq->desc_ring; + sw_ring = txq->sw_ring; + tx_id = txq->tx_tail; + sw_id = txq->sw_tail; + txe = &sw_ring[sw_id]; + + for (nb_tx = 0; nb_tx < nb_pkts; nb_tx++) { + tx_pkt = tx_pkts[nb_tx]; + + if (txq->nb_free <= txq->free_thresh) { + /* TODO: Need to refine + * 1. free and clean: Better to decide a clean destination instead of + * loop times. And don't free mbuf when RS got immediately, free when + * transmit or according to the clean destination. + * Now, just ignore the RE write back, free mbuf when get RS + * 2. out-of-order rewrite back haven't be supported, SW head and HW head + * need to be separated. + **/ + nb_to_clean = 2 * txq->rs_thresh; + while (nb_to_clean--) + idpf_split_tx_free(txq->complq); + } + + if (txq->nb_free < tx_pkt->nb_segs) + break; + nb_used = tx_pkt->nb_segs; + + + do { + txd = &txr[tx_id]; + txn = &sw_ring[txe->next_id]; + txe->mbuf = tx_pkt; + + /* Setup TX descriptor */ + txd->buf_addr = + rte_cpu_to_le_64(rte_mbuf_data_iova(tx_pkt)); + txd->qw1.cmd_dtype = + rte_cpu_to_le_16(IDPF_TX_DESC_DTYPE_FLEX_FLOW_SCHE); + txd->qw1.rxr_bufsize = tx_pkt->data_len; + txd->qw1.compl_tag = sw_id; + tx_id++; + if (tx_id == txq->nb_tx_desc) + tx_id = 0; + sw_id = txe->next_id; + txe = txn; + tx_pkt = tx_pkt->next; + } while (tx_pkt); + + /* fill the last descriptor with End of Packet (EOP) bit */ + txd->qw1.cmd_dtype |= IDPF_TXD_FLEX_FLOW_CMD_EOP; + + if (unlikely(!(tx_id % 32))) + txd->qw1.cmd_dtype |= IDPF_TXD_FLEX_FLOW_CMD_RE; + txq->nb_free = (uint16_t)(txq->nb_free - nb_used); + txq->nb_used = (uint16_t)(txq->nb_used + nb_used); + } + + /* update the tail pointer if any packets were processed */ + if (likely(nb_tx)) { + IDPF_PCI_REG_WRITE(txq->qtx_tail, tx_id); + txq->tx_tail = tx_id; + txq->sw_tail = sw_id; + } + + return nb_tx; +} + +static inline void +idpf_update_rx_tail(struct idpf_rx_queue *rxq, uint16_t nb_hold, + uint16_t rx_id) +{ + nb_hold = (uint16_t)(nb_hold + rxq->nb_rx_hold); + + if (nb_hold > rxq->rx_free_thresh) { + PMD_RX_LOG(DEBUG, + "port_id=%u queue_id=%u rx_tail=%u nb_hold=%u", + rxq->port_id, rxq->queue_id, rx_id, nb_hold); + rx_id = (uint16_t)((rx_id == 0) ? + (rxq->nb_rx_desc - 1) : (rx_id - 1)); + IDPF_PCI_REG_WRITE(rxq->qrx_tail, rx_id); + nb_hold = 0; + } + rxq->nb_rx_hold = nb_hold; +} + +uint16_t +idpf_singleq_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts) +{ + volatile union virtchnl2_rx_desc *rx_ring; + volatile union virtchnl2_rx_desc *rxdp; + union virtchnl2_rx_desc rxd; + struct idpf_rx_queue *rxq; + const uint32_t *ptype_tbl; + uint16_t rx_id, nb_hold; + struct rte_eth_dev *dev; + uint16_t rx_packet_len; + struct rte_mbuf *rxm; + struct rte_mbuf *nmb; + uint16_t rx_status0; + uint64_t dma_addr; + uint16_t nb_rx; + + nb_rx = 0; + nb_hold = 0; + rxq = rx_queue; + + if (unlikely(!rxq) || unlikely(!rxq->q_started)) + return nb_rx; + + rx_id = rxq->rx_tail; + rx_ring = rxq->rx_ring; + ptype_tbl = rxq->adapter->ptype_tbl; + + if (rxq->offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP) + rxq->hw_register_set = 1; + + while (nb_rx < nb_pkts) { + rxdp = &rx_ring[rx_id]; + rx_status0 = rte_le_to_cpu_16(rxdp->flex_nic_wb.status_error0); + + /* Check the DD bit first */ + if (!(rx_status0 & (1 << VIRTCHNL2_RX_FLEX_DESC_STATUS0_DD_S))) + break; + + nmb = rte_mbuf_raw_alloc(rxq->mp); + if (unlikely(!nmb)) { + dev = &rte_eth_devices[rxq->port_id]; + dev->data->rx_mbuf_alloc_failed++; + PMD_RX_LOG(DEBUG, "RX mbuf alloc failed port_id=%u " + "queue_id=%u", rxq->port_id, rxq->queue_id); + break; + } + rxd = *rxdp; /* copy descriptor in ring to temp variable*/ + + nb_hold++; + rxm = rxq->sw_ring[rx_id]; + rxq->sw_ring[rx_id] = nmb; + rx_id++; + if (unlikely(rx_id == rxq->nb_rx_desc)) + rx_id = 0; + + /* Prefetch next mbuf */ + rte_prefetch0(rxq->sw_ring[rx_id]); + + /* When next RX descriptor is on a cache line boundary, + * prefetch the next 4 RX descriptors and next 8 pointers + * to mbufs. + */ + if ((rx_id & 0x3) == 0) { + rte_prefetch0(&rx_ring[rx_id]); + rte_prefetch0(rxq->sw_ring[rx_id]); + } + dma_addr = + rte_cpu_to_le_64(rte_mbuf_data_iova_default(nmb)); + rxdp->read.hdr_addr = 0; + rxdp->read.pkt_addr = dma_addr; + + rx_packet_len = (rte_cpu_to_le_16(rxd.flex_nic_wb.pkt_len) & + VIRTCHNL2_RX_FLEX_DESC_PKT_LEN_M); + + rxm->data_off = RTE_PKTMBUF_HEADROOM; + rte_prefetch0(RTE_PTR_ADD(rxm->buf_addr, RTE_PKTMBUF_HEADROOM)); + rxm->nb_segs = 1; + rxm->next = NULL; + rxm->pkt_len = rx_packet_len; + rxm->data_len = rx_packet_len; + rxm->port = rxq->port_id; + rxm->ol_flags = 0; + rxm->packet_type = + ptype_tbl[(uint8_t)(rte_cpu_to_le_16(rxd.flex_nic_wb.ptype_flex_flags0) & + VIRTCHNL2_RX_FLEX_DESC_PTYPE_M)]; + + + rx_pkts[nb_rx++] = rxm; + } + rxq->rx_tail = rx_id; + + idpf_update_rx_tail(rxq, nb_hold, rx_id); + + return nb_rx; +} + +static inline int +idpf_xmit_cleanup(struct idpf_tx_queue *txq) +{ + uint16_t last_desc_cleaned = txq->last_desc_cleaned; + struct idpf_tx_entry *sw_ring = txq->sw_ring; + uint16_t nb_tx_desc = txq->nb_tx_desc; + uint16_t desc_to_clean_to; + uint16_t nb_tx_to_clean; + uint16_t i; + + volatile struct idpf_flex_tx_desc *txd = txq->tx_ring; + + desc_to_clean_to = (uint16_t)(last_desc_cleaned + txq->rs_thresh); + if (desc_to_clean_to >= nb_tx_desc) + desc_to_clean_to = (uint16_t)(desc_to_clean_to - nb_tx_desc); + + desc_to_clean_to = sw_ring[desc_to_clean_to].last_id; + /* In the writeback Tx desccriptor, the only significant fields are the 4-bit DTYPE */ + if ((txd[desc_to_clean_to].qw1.cmd_dtype & + rte_cpu_to_le_16(IDPF_TXD_QW1_DTYPE_M)) != + rte_cpu_to_le_16(IDPF_TX_DESC_DTYPE_DESC_DONE)) { + PMD_TX_LOG(DEBUG, "TX descriptor %4u is not done " + "(port=%d queue=%d)", desc_to_clean_to, + txq->port_id, txq->queue_id); + return -1; + } + + if (last_desc_cleaned > desc_to_clean_to) + nb_tx_to_clean = (uint16_t)((nb_tx_desc - last_desc_cleaned) + + desc_to_clean_to); + else + nb_tx_to_clean = (uint16_t)(desc_to_clean_to - + last_desc_cleaned); + + txd[desc_to_clean_to].qw1.cmd_dtype = 0; + txd[desc_to_clean_to].qw1.buf_size = 0; + for (i = 0; i < RTE_DIM(txd[desc_to_clean_to].qw1.flex.raw); i++) + txd[desc_to_clean_to].qw1.flex.raw[i] = 0; + + txq->last_desc_cleaned = desc_to_clean_to; + txq->nb_free = (uint16_t)(txq->nb_free + nb_tx_to_clean); + + return 0; +} + +/* TX function */ +uint16_t +idpf_singleq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts) +{ + volatile struct idpf_flex_tx_desc *txd; + volatile struct idpf_flex_tx_desc *txr; + struct idpf_tx_entry *txe, *txn; + struct idpf_tx_entry *sw_ring; + struct idpf_tx_queue *txq; + struct rte_mbuf *tx_pkt; + struct rte_mbuf *m_seg; + uint64_t buf_dma_addr; + uint16_t tx_last; + uint16_t nb_used; + uint16_t td_cmd; + uint16_t tx_id; + uint16_t nb_tx; + uint16_t slen; + + nb_tx = 0; + txq = tx_queue; + + if (unlikely(!txq) || unlikely(!txq->q_started)) + return nb_tx; + + sw_ring = txq->sw_ring; + txr = txq->tx_ring; + tx_id = txq->tx_tail; + txe = &sw_ring[tx_id]; + + /* Check if the descriptor ring needs to be cleaned. */ + if (txq->nb_free < txq->free_thresh) + (void)idpf_xmit_cleanup(txq); + + for (nb_tx = 0; nb_tx < nb_pkts; nb_tx++) { + td_cmd = 0; + + tx_pkt = *tx_pkts++; + RTE_MBUF_PREFETCH_TO_FREE(txe->mbuf); + + /* The number of descriptors that must be allocated for + * a packet equals to the number of the segments of that + * packet plus 1 context descriptor if needed. + */ + nb_used = (uint16_t)(tx_pkt->nb_segs); + tx_last = (uint16_t)(tx_id + nb_used - 1); + + /* Circular ring */ + if (tx_last >= txq->nb_tx_desc) + tx_last = (uint16_t)(tx_last - txq->nb_tx_desc); + + PMD_TX_LOG(DEBUG, "port_id=%u queue_id=%u" + " tx_first=%u tx_last=%u", + txq->port_id, txq->queue_id, tx_id, tx_last); + + if (nb_used > txq->nb_free) { + if (idpf_xmit_cleanup(txq)) { + if (nb_tx == 0) + return 0; + goto end_of_tx; + } + if (unlikely(nb_used > txq->rs_thresh)) { + while (nb_used > txq->nb_free) { + if (idpf_xmit_cleanup(txq)) { + if (nb_tx == 0) + return 0; + goto end_of_tx; + } + } + } + } + + + m_seg = tx_pkt; + do { + txd = &txr[tx_id]; + txn = &sw_ring[txe->next_id]; + + if (txe->mbuf) + rte_pktmbuf_free_seg(txe->mbuf); + txe->mbuf = m_seg; + + /* Setup TX Descriptor */ + slen = m_seg->data_len; + buf_dma_addr = rte_mbuf_data_iova(m_seg); + txd->buf_addr = rte_cpu_to_le_64(buf_dma_addr); + txd->qw1.buf_size = slen; + txd->qw1.cmd_dtype = rte_cpu_to_le_16(IDPF_TX_DESC_DTYPE_FLEX_DATA << + IDPF_FLEX_TXD_QW1_DTYPE_S); + + txe->last_id = tx_last; + tx_id = txe->next_id; + txe = txn; + m_seg = m_seg->next; + } while (m_seg); + + /* The last packet data descriptor needs End Of Packet (EOP) */ + td_cmd |= IDPF_TX_FLEX_DESC_CMD_EOP; + txq->nb_used = (uint16_t)(txq->nb_used + nb_used); + txq->nb_free = (uint16_t)(txq->nb_free - nb_used); + + if (txq->nb_used >= txq->rs_thresh) { + PMD_TX_LOG(DEBUG, "Setting RS bit on TXD id=" + "%4u (port=%d queue=%d)", + tx_last, txq->port_id, txq->queue_id); + + td_cmd |= IDPF_TX_FLEX_DESC_CMD_RS; + + /* Update txq RS bit counters */ + txq->nb_used = 0; + } + + txd->qw1.cmd_dtype |= rte_cpu_to_le_16(td_cmd << IDPF_FLEX_TXD_QW1_CMD_S); + } + +end_of_tx: + rte_wmb(); + + PMD_TX_LOG(DEBUG, "port_id=%u queue_id=%u tx_tail=%u nb_tx=%u", + txq->port_id, txq->queue_id, tx_id, nb_tx); + + IDPF_PCI_REG_WRITE(txq->qtx_tail, tx_id); + txq->tx_tail = tx_id; + + return nb_tx; +} + +/* TX prep functions */ +uint16_t +idpf_prep_pkts(__rte_unused void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts) +{ + int i, ret; + uint64_t ol_flags; + struct rte_mbuf *m; + + for (i = 0; i < nb_pkts; i++) { + m = tx_pkts[i]; + ol_flags = m->ol_flags; + + /* Check condition for nb_segs > IDPF_TX_MAX_MTU_SEG. */ + if (!(ol_flags & RTE_MBUF_F_TX_TCP_SEG)) { + if (m->nb_segs > IDPF_TX_MAX_MTU_SEG) { + rte_errno = EINVAL; + return i; + } + } else if ((m->tso_segsz < IDPF_MIN_TSO_MSS) || + (m->tso_segsz > IDPF_MAX_TSO_MSS) || + (m->pkt_len > IDPF_MAX_TSO_FRAME_SIZE)) { + /* MSS outside the range are considered malicious */ + rte_errno = EINVAL; + return i; + } + + if (m->pkt_len < IDPF_MIN_FRAME_SIZE) { + rte_errno = EINVAL; + return i; + } + + ret = rte_net_intel_cksum_prepare(m); + if (ret != 0) { + rte_errno = -ret; + return i; + } + } + + return i; +} + +void +idpf_set_rx_function(struct rte_eth_dev *dev) +{ + struct idpf_vport *vport = dev->data->dev_private; + + if (vport->rxq_model == VIRTCHNL2_QUEUE_MODEL_SPLIT) + dev->rx_pkt_burst = idpf_splitq_recv_pkts; + else + dev->rx_pkt_burst = idpf_singleq_recv_pkts; +} + +void +idpf_set_tx_function(struct rte_eth_dev *dev) +{ + struct idpf_vport *vport = dev->data->dev_private; + + if (vport->txq_model == VIRTCHNL2_QUEUE_MODEL_SPLIT) { + dev->tx_pkt_burst = idpf_splitq_xmit_pkts; + dev->tx_pkt_prepare = idpf_prep_pkts; + } else { + dev->tx_pkt_burst = idpf_singleq_xmit_pkts; + dev->tx_pkt_prepare = idpf_prep_pkts; + } +} diff --git a/drivers/net/idpf/idpf_rxtx.h b/drivers/net/idpf/idpf_rxtx.h index cd1dda4688..bd3ebe2f50 100644 --- a/drivers/net/idpf/idpf_rxtx.h +++ b/drivers/net/idpf/idpf_rxtx.h @@ -44,9 +44,6 @@ #define IDPF_MAX_TSO_FRAME_SIZE 262143 #define IDPF_TX_MAX_MTU_SEG 10 -#define IDPF_TX_OFFLOAD_NOTSUP_MASK \ - (RTE_MBUF_F_TX_OFFLOAD_MASK ^ IDPF_TX_OFFLOAD_MASK) - #define IDPF_GET_PTYPE_SIZE(p) \ (sizeof(struct virtchnl2_ptype) + \ (((p)->proto_id_count ? ((p)->proto_id_count - 1) : 0) * sizeof((p)->proto_id[0]))) @@ -191,8 +188,21 @@ int idpf_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id); int idpf_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id); void idpf_dev_tx_queue_release(struct rte_eth_dev *dev, uint16_t qid); +uint16_t idpf_singleq_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts); +uint16_t idpf_splitq_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts); +uint16_t idpf_singleq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts); +uint16_t idpf_splitq_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts); +uint16_t idpf_prep_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts); void idpf_stop_queues(struct rte_eth_dev *dev); +void idpf_set_rx_function(struct rte_eth_dev *dev); +void idpf_set_tx_function(struct rte_eth_dev *dev); + const uint32_t *idpf_dev_supported_ptypes_get(struct rte_eth_dev *dev); #endif /* _IDPF_RXTX_H_ */ -- 2.34.1