X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fi40e%2Fi40e_rxtx.c;h=788ffb51c272ad00a7b76ff486f4765e22a0263d;hb=66689dc09f0cf92af03abcba58b7474fa8ac847e;hp=ad06b71e6db52fc40d727d52427cd52d172e6ce0;hpb=6c293ffd63b3f61a95d8c492d007160874e7a057;p=dpdk.git diff --git a/drivers/net/i40e/i40e_rxtx.c b/drivers/net/i40e/i40e_rxtx.c index ad06b71e6d..788ffb51c2 100644 --- a/drivers/net/i40e/i40e_rxtx.c +++ b/drivers/net/i40e/i40e_rxtx.c @@ -1,34 +1,5 @@ -/*- - * BSD LICENSE - * - * Copyright(c) 2010-2016 Intel Corporation. All rights reserved. - * All rights reserved. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * - * * Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * * Redistributions in binary form must reproduce the above copyright - * notice, this list of conditions and the following disclaimer in - * the documentation and/or other materials provided with the - * distribution. - * * Neither the name of Intel Corporation nor the names of its - * contributors may be used to endorse or promote products derived - * from this software without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS - * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT - * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR - * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT - * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, - * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT - * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, - * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY - * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT - * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE - * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright(c) 2010-2016 Intel Corporation */ #include @@ -46,12 +17,13 @@ #include #include #include -#include +#include #include #include #include #include #include +#include #include "i40e_logs.h" #include "base/i40e_prototype.h" @@ -69,46 +41,83 @@ /* Base address of the HW descriptor ring should be 128B aligned. */ #define I40E_RING_BASE_ALIGN 128 -#define I40E_SIMPLE_FLAGS ((uint32_t)ETH_TXQ_FLAGS_NOMULTSEGS | \ - ETH_TXQ_FLAGS_NOOFFLOADS) - #define I40E_TXD_CMD (I40E_TX_DESC_CMD_EOP | I40E_TX_DESC_CMD_RS) #ifdef RTE_LIBRTE_IEEE1588 -#define I40E_TX_IEEE1588_TMST PKT_TX_IEEE1588_TMST +#define I40E_TX_IEEE1588_TMST RTE_MBUF_F_TX_IEEE1588_TMST #else #define I40E_TX_IEEE1588_TMST 0 #endif -#define I40E_TX_CKSUM_OFFLOAD_MASK ( \ - PKT_TX_IP_CKSUM | \ - PKT_TX_L4_MASK | \ - PKT_TX_TCP_SEG | \ - PKT_TX_OUTER_IP_CKSUM) - -#define I40E_TX_OFFLOAD_MASK ( \ - PKT_TX_IP_CKSUM | \ - PKT_TX_L4_MASK | \ - PKT_TX_OUTER_IP_CKSUM | \ - PKT_TX_TCP_SEG | \ - PKT_TX_QINQ_PKT | \ - PKT_TX_VLAN_PKT | \ - PKT_TX_TUNNEL_MASK | \ +#define I40E_TX_CKSUM_OFFLOAD_MASK (RTE_MBUF_F_TX_IP_CKSUM | \ + RTE_MBUF_F_TX_L4_MASK | \ + RTE_MBUF_F_TX_TCP_SEG | \ + RTE_MBUF_F_TX_OUTER_IP_CKSUM) + +#define I40E_TX_OFFLOAD_MASK (RTE_MBUF_F_TX_OUTER_IPV4 | \ + RTE_MBUF_F_TX_OUTER_IPV6 | \ + RTE_MBUF_F_TX_IPV4 | \ + RTE_MBUF_F_TX_IPV6 | \ + RTE_MBUF_F_TX_IP_CKSUM | \ + RTE_MBUF_F_TX_L4_MASK | \ + RTE_MBUF_F_TX_OUTER_IP_CKSUM | \ + RTE_MBUF_F_TX_TCP_SEG | \ + RTE_MBUF_F_TX_QINQ | \ + RTE_MBUF_F_TX_VLAN | \ + RTE_MBUF_F_TX_TUNNEL_MASK | \ + RTE_MBUF_F_TX_OUTER_UDP_CKSUM | \ I40E_TX_IEEE1588_TMST) #define I40E_TX_OFFLOAD_NOTSUP_MASK \ - (PKT_TX_OFFLOAD_MASK ^ I40E_TX_OFFLOAD_MASK) + (RTE_MBUF_F_TX_OFFLOAD_MASK ^ I40E_TX_OFFLOAD_MASK) + +#define I40E_TX_OFFLOAD_SIMPLE_SUP_MASK (RTE_MBUF_F_TX_IPV4 | \ + RTE_MBUF_F_TX_IPV6 | \ + RTE_MBUF_F_TX_OUTER_IPV4 | \ + RTE_MBUF_F_TX_OUTER_IPV6) + +#define I40E_TX_OFFLOAD_SIMPLE_NOTSUP_MASK \ + (RTE_MBUF_F_TX_OFFLOAD_MASK ^ I40E_TX_OFFLOAD_SIMPLE_SUP_MASK) -static uint16_t i40e_xmit_pkts_simple(void *tx_queue, - struct rte_mbuf **tx_pkts, - uint16_t nb_pkts); +static int +i40e_monitor_callback(const uint64_t value, + const uint64_t arg[RTE_POWER_MONITOR_OPAQUE_SZ] __rte_unused) +{ + const uint64_t m = rte_cpu_to_le_64(1 << I40E_RX_DESC_STATUS_DD_SHIFT); + /* + * we expect the DD bit to be set to 1 if this descriptor was already + * written to. + */ + return (value & m) == m ? -1 : 0; +} + +int +i40e_get_monitor_addr(void *rx_queue, struct rte_power_monitor_cond *pmc) +{ + struct i40e_rx_queue *rxq = rx_queue; + volatile union i40e_rx_desc *rxdp; + uint16_t desc; + + desc = rxq->rx_tail; + rxdp = &rxq->rx_ring[desc]; + /* watch for changes in status bit */ + pmc->addr = &rxdp->wb.qword1.status_error_len; + + /* comparison callback */ + pmc->fn = i40e_monitor_callback; + + /* registers are 64-bit */ + pmc->size = sizeof(uint64_t); + + return 0; +} static inline void i40e_rxd_to_vlan_tci(struct rte_mbuf *mb, volatile union i40e_rx_desc *rxdp) { if (rte_le_to_cpu_64(rxdp->wb.qword1.status_error_len) & (1 << I40E_RX_DESC_STATUS_L2TAG1P_SHIFT)) { - mb->ol_flags |= PKT_RX_VLAN | PKT_RX_VLAN_STRIPPED; + mb->ol_flags |= RTE_MBUF_F_RX_VLAN | RTE_MBUF_F_RX_VLAN_STRIPPED; mb->vlan_tci = rte_le_to_cpu_16(rxdp->wb.qword0.lo_dword.l2tag1); PMD_RX_LOG(DEBUG, "Descriptor l2tag1: %u", @@ -119,7 +128,8 @@ i40e_rxd_to_vlan_tci(struct rte_mbuf *mb, volatile union i40e_rx_desc *rxdp) #ifndef RTE_LIBRTE_I40E_16BYTE_RX_DESC if (rte_le_to_cpu_16(rxdp->wb.qword2.ext_status) & (1 << I40E_RX_DESC_EXT_STATUS_L2TAG2P_SHIFT)) { - mb->ol_flags |= PKT_RX_QINQ_STRIPPED; + mb->ol_flags |= RTE_MBUF_F_RX_QINQ_STRIPPED | RTE_MBUF_F_RX_QINQ | + RTE_MBUF_F_RX_VLAN_STRIPPED | RTE_MBUF_F_RX_VLAN; mb->vlan_tci_outer = mb->vlan_tci; mb->vlan_tci = rte_le_to_cpu_16(rxdp->wb.qword2.l2tag2_2); PMD_RX_LOG(DEBUG, "Descriptor l2tag2_1: %u, l2tag2_2: %u", @@ -142,11 +152,11 @@ i40e_rxd_status_to_pkt_flags(uint64_t qword) /* Check if RSS_HASH */ flags = (((qword >> I40E_RX_DESC_STATUS_FLTSTAT_SHIFT) & I40E_RX_DESC_FLTSTAT_RSS_HASH) == - I40E_RX_DESC_FLTSTAT_RSS_HASH) ? PKT_RX_RSS_HASH : 0; + I40E_RX_DESC_FLTSTAT_RSS_HASH) ? RTE_MBUF_F_RX_RSS_HASH : 0; /* Check if FDIR Match */ flags |= (qword & (1 << I40E_RX_DESC_STATUS_FLM_SHIFT) ? - PKT_RX_FDIR : 0); + RTE_MBUF_F_RX_FDIR : 0); return flags; } @@ -159,22 +169,22 @@ i40e_rxd_error_to_pkt_flags(uint64_t qword) #define I40E_RX_ERR_BITS 0x3f if (likely((error_bits & I40E_RX_ERR_BITS) == 0)) { - flags |= (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_GOOD); + flags |= (RTE_MBUF_F_RX_IP_CKSUM_GOOD | RTE_MBUF_F_RX_L4_CKSUM_GOOD); return flags; } if (unlikely(error_bits & (1 << I40E_RX_DESC_ERROR_IPE_SHIFT))) - flags |= PKT_RX_IP_CKSUM_BAD; + flags |= RTE_MBUF_F_RX_IP_CKSUM_BAD; else - flags |= PKT_RX_IP_CKSUM_GOOD; + flags |= RTE_MBUF_F_RX_IP_CKSUM_GOOD; if (unlikely(error_bits & (1 << I40E_RX_DESC_ERROR_L4E_SHIFT))) - flags |= PKT_RX_L4_CKSUM_BAD; + flags |= RTE_MBUF_F_RX_L4_CKSUM_BAD; else - flags |= PKT_RX_L4_CKSUM_GOOD; + flags |= RTE_MBUF_F_RX_L4_CKSUM_GOOD; if (unlikely(error_bits & (1 << I40E_RX_DESC_ERROR_EIPE_SHIFT))) - flags |= PKT_RX_EIP_CKSUM_BAD; + flags |= RTE_MBUF_F_RX_OUTER_IP_CKSUM_BAD; return flags; } @@ -193,9 +203,9 @@ i40e_get_iee15888_flags(struct rte_mbuf *mb, uint64_t qword) if ((mb->packet_type & RTE_PTYPE_L2_MASK) == RTE_PTYPE_L2_ETHER_TIMESYNC) - pkt_flags = PKT_RX_IEEE1588_PTP; + pkt_flags = RTE_MBUF_F_RX_IEEE1588_PTP; if (tsyn & 0x04) { - pkt_flags |= PKT_RX_IEEE1588_TMST; + pkt_flags |= RTE_MBUF_F_RX_IEEE1588_TMST; mb->timesync = tsyn & 0x03; } @@ -203,12 +213,6 @@ i40e_get_iee15888_flags(struct rte_mbuf *mb, uint64_t qword) } #endif -#define I40E_RX_DESC_EXT_STATUS_FLEXBH_MASK 0x03 -#define I40E_RX_DESC_EXT_STATUS_FLEXBH_FD_ID 0x01 -#define I40E_RX_DESC_EXT_STATUS_FLEXBH_FLEX 0x02 -#define I40E_RX_DESC_EXT_STATUS_FLEXBL_MASK 0x03 -#define I40E_RX_DESC_EXT_STATUS_FLEXBL_FLEX 0x01 - static inline uint64_t i40e_rxd_build_fdir(volatile union i40e_rx_desc *rxdp, struct rte_mbuf *mb) { @@ -227,21 +231,21 @@ i40e_rxd_build_fdir(volatile union i40e_rx_desc *rxdp, struct rte_mbuf *mb) if (flexbh == I40E_RX_DESC_EXT_STATUS_FLEXBH_FD_ID) { mb->hash.fdir.hi = rte_le_to_cpu_32(rxdp->wb.qword3.hi_dword.fd_id); - flags |= PKT_RX_FDIR_ID; + flags |= RTE_MBUF_F_RX_FDIR_ID; } else if (flexbh == I40E_RX_DESC_EXT_STATUS_FLEXBH_FLEX) { mb->hash.fdir.hi = rte_le_to_cpu_32(rxdp->wb.qword3.hi_dword.flex_bytes_hi); - flags |= PKT_RX_FDIR_FLX; + flags |= RTE_MBUF_F_RX_FDIR_FLX; } if (flexbl == I40E_RX_DESC_EXT_STATUS_FLEXBL_FLEX) { mb->hash.fdir.lo = rte_le_to_cpu_32(rxdp->wb.qword3.lo_dword.flex_bytes_lo); - flags |= PKT_RX_FDIR_FLX; + flags |= RTE_MBUF_F_RX_FDIR_FLX; } #else mb->hash.fdir.hi = rte_le_to_cpu_32(rxdp->wb.qword0.hi_dword.fd_id); - flags |= PKT_RX_FDIR_ID; + flags |= RTE_MBUF_F_RX_FDIR_ID; #endif return flags; } @@ -252,11 +256,11 @@ i40e_parse_tunneling_params(uint64_t ol_flags, uint32_t *cd_tunneling) { /* EIPT: External (outer) IP header type */ - if (ol_flags & PKT_TX_OUTER_IP_CKSUM) + if (ol_flags & RTE_MBUF_F_TX_OUTER_IP_CKSUM) *cd_tunneling |= I40E_TX_CTX_EXT_IP_IPV4; - else if (ol_flags & PKT_TX_OUTER_IPV4) + else if (ol_flags & RTE_MBUF_F_TX_OUTER_IPV4) *cd_tunneling |= I40E_TX_CTX_EXT_IP_IPV4_NO_CSUM; - else if (ol_flags & PKT_TX_OUTER_IPV6) + else if (ol_flags & RTE_MBUF_F_TX_OUTER_IPV6) *cd_tunneling |= I40E_TX_CTX_EXT_IP_IPV6; /* EIPLEN: External (outer) IP header length, in DWords */ @@ -264,15 +268,15 @@ i40e_parse_tunneling_params(uint64_t ol_flags, I40E_TXD_CTX_QW0_EXT_IPLEN_SHIFT; /* L4TUNT: L4 Tunneling Type */ - switch (ol_flags & PKT_TX_TUNNEL_MASK) { - case PKT_TX_TUNNEL_IPIP: + switch (ol_flags & RTE_MBUF_F_TX_TUNNEL_MASK) { + case RTE_MBUF_F_TX_TUNNEL_IPIP: /* for non UDP / GRE tunneling, set to 00b */ break; - case PKT_TX_TUNNEL_VXLAN: - case PKT_TX_TUNNEL_GENEVE: + case RTE_MBUF_F_TX_TUNNEL_VXLAN: + case RTE_MBUF_F_TX_TUNNEL_GENEVE: *cd_tunneling |= I40E_TXD_CTX_UDP_TUNNELING; break; - case PKT_TX_TUNNEL_GRE: + case RTE_MBUF_F_TX_TUNNEL_GRE: *cd_tunneling |= I40E_TXD_CTX_GRE_TUNNELING; break; default: @@ -300,7 +304,7 @@ i40e_txd_enable_checksum(uint64_t ol_flags, union i40e_tx_offload tx_offload) { /* Set MACLEN */ - if (ol_flags & PKT_TX_TUNNEL_MASK) + if (ol_flags & RTE_MBUF_F_TX_TUNNEL_MASK) *td_offset |= (tx_offload.outer_l2_len >> 1) << I40E_TX_DESC_LENGTH_MACLEN_SHIFT; else @@ -308,21 +312,21 @@ i40e_txd_enable_checksum(uint64_t ol_flags, << I40E_TX_DESC_LENGTH_MACLEN_SHIFT; /* Enable L3 checksum offloads */ - if (ol_flags & PKT_TX_IP_CKSUM) { + if (ol_flags & RTE_MBUF_F_TX_IP_CKSUM) { *td_cmd |= I40E_TX_DESC_CMD_IIPT_IPV4_CSUM; *td_offset |= (tx_offload.l3_len >> 2) << I40E_TX_DESC_LENGTH_IPLEN_SHIFT; - } else if (ol_flags & PKT_TX_IPV4) { + } else if (ol_flags & RTE_MBUF_F_TX_IPV4) { *td_cmd |= I40E_TX_DESC_CMD_IIPT_IPV4; *td_offset |= (tx_offload.l3_len >> 2) << I40E_TX_DESC_LENGTH_IPLEN_SHIFT; - } else if (ol_flags & PKT_TX_IPV6) { + } else if (ol_flags & RTE_MBUF_F_TX_IPV6) { *td_cmd |= I40E_TX_DESC_CMD_IIPT_IPV6; *td_offset |= (tx_offload.l3_len >> 2) << I40E_TX_DESC_LENGTH_IPLEN_SHIFT; } - if (ol_flags & PKT_TX_TCP_SEG) { + if (ol_flags & RTE_MBUF_F_TX_TCP_SEG) { *td_cmd |= I40E_TX_DESC_CMD_L4T_EOFT_TCP; *td_offset |= (tx_offload.l4_len >> 2) << I40E_TX_DESC_LENGTH_L4_FC_LEN_SHIFT; @@ -330,20 +334,20 @@ i40e_txd_enable_checksum(uint64_t ol_flags, } /* Enable L4 checksum offloads */ - switch (ol_flags & PKT_TX_L4_MASK) { - case PKT_TX_TCP_CKSUM: + switch (ol_flags & RTE_MBUF_F_TX_L4_MASK) { + case RTE_MBUF_F_TX_TCP_CKSUM: *td_cmd |= I40E_TX_DESC_CMD_L4T_EOFT_TCP; - *td_offset |= (sizeof(struct tcp_hdr) >> 2) << + *td_offset |= (sizeof(struct rte_tcp_hdr) >> 2) << I40E_TX_DESC_LENGTH_L4_FC_LEN_SHIFT; break; - case PKT_TX_SCTP_CKSUM: + case RTE_MBUF_F_TX_SCTP_CKSUM: *td_cmd |= I40E_TX_DESC_CMD_L4T_EOFT_SCTP; - *td_offset |= (sizeof(struct sctp_hdr) >> 2) << + *td_offset |= (sizeof(struct rte_sctp_hdr) >> 2) << I40E_TX_DESC_LENGTH_L4_FC_LEN_SHIFT; break; - case PKT_TX_UDP_CKSUM: + case RTE_MBUF_F_TX_UDP_CKSUM: *td_cmd |= I40E_TX_DESC_CMD_L4T_EOFT_UDP; - *td_offset |= (sizeof(struct udp_hdr) >> 2) << + *td_offset |= (sizeof(struct rte_udp_hdr) >> 2) << I40E_TX_DESC_LENGTH_L4_FC_LEN_SHIFT; break; default: @@ -383,9 +387,9 @@ i40e_xmit_cleanup(struct i40e_tx_queue *txq) if ((txd[desc_to_clean_to].cmd_type_offset_bsz & rte_cpu_to_le_64(I40E_TXD_QW1_DTYPE_MASK)) != rte_cpu_to_le_64(I40E_TX_DESC_DTYPE_DESC_DONE)) { - PMD_TX_FREE_LOG(DEBUG, "TX descriptor %4u is not done " - "(port=%d queue=%d)", desc_to_clean_to, - txq->port_id, txq->queue_id); + 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; } @@ -454,7 +458,7 @@ i40e_rx_scan_hw_ring(struct i40e_rx_queue *rxq) uint16_t pkt_len; uint64_t qword1; uint32_t rx_status; - int32_t s[I40E_LOOK_AHEAD], nb_dd; + int32_t s[I40E_LOOK_AHEAD], var, nb_dd; int32_t i, j, nb_rx = 0; uint64_t pkt_flags; uint32_t *ptype_tbl = rxq->vsi->adapter->ptype_tbl; @@ -484,11 +488,22 @@ i40e_rx_scan_hw_ring(struct i40e_rx_queue *rxq) I40E_RXD_QW1_STATUS_SHIFT; } - rte_smp_rmb(); + /* This barrier is to order loads of different words in the descriptor */ + rte_atomic_thread_fence(__ATOMIC_ACQUIRE); /* Compute how many status bits were set */ - for (j = 0, nb_dd = 0; j < I40E_LOOK_AHEAD; j++) - nb_dd += s[j] & (1 << I40E_RX_DESC_STATUS_DD_SHIFT); + for (j = 0, nb_dd = 0; j < I40E_LOOK_AHEAD; j++) { + var = s[j] & (1 << I40E_RX_DESC_STATUS_DD_SHIFT); +#ifdef RTE_ARCH_ARM + /* For Arm platforms, only compute continuous status bits */ + if (var) + nb_dd += 1; + else + break; +#else + nb_dd += var; +#endif + } nb_rx += nb_dd; @@ -509,10 +524,10 @@ i40e_rx_scan_hw_ring(struct i40e_rx_queue *rxq) ptype_tbl[(uint8_t)((qword1 & I40E_RXD_QW1_PTYPE_MASK) >> I40E_RXD_QW1_PTYPE_SHIFT)]; - if (pkt_flags & PKT_RX_RSS_HASH) + if (pkt_flags & RTE_MBUF_F_RX_RSS_HASH) mb->hash.rss = rte_le_to_cpu_32(\ rxdp[j].wb.qword0.hi_dword.rss); - if (pkt_flags & PKT_RX_FDIR) + if (pkt_flags & RTE_MBUF_F_RX_FDIR) pkt_flags |= i40e_rxd_build_fdir(&rxdp[j], mb); #ifdef RTE_LIBRTE_IEEE1588 @@ -594,9 +609,8 @@ i40e_rx_alloc_bufs(struct i40e_rx_queue *rxq) rxdp[i].read.pkt_addr = dma_addr; } - /* Update rx tail regsiter */ - rte_wmb(); - I40E_PCI_REG_WRITE_RELAXED(rxq->qrx_tail, rxq->rx_free_trigger); + /* Update rx tail register */ + I40E_PCI_REG_WRITE(rxq->qrx_tail, rxq->rx_free_trigger); rxq->rx_free_trigger = (uint16_t)(rxq->rx_free_trigger + rxq->rx_free_thresh); @@ -730,6 +744,12 @@ i40e_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts) break; } + /** + * Use acquire fence to ensure that qword1 which includes DD + * bit is loaded before loading of other descriptor words. + */ + rte_atomic_thread_fence(__ATOMIC_ACQUIRE); + rxd = *rxdp; nb_hold++; rxe = &sw_ring[rx_id]; @@ -773,10 +793,10 @@ i40e_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts) rxm->packet_type = ptype_tbl[(uint8_t)((qword1 & I40E_RXD_QW1_PTYPE_MASK) >> I40E_RXD_QW1_PTYPE_SHIFT)]; - if (pkt_flags & PKT_RX_RSS_HASH) + if (pkt_flags & RTE_MBUF_F_RX_RSS_HASH) rxm->hash.rss = rte_le_to_cpu_32(rxd.wb.qword0.hi_dword.rss); - if (pkt_flags & PKT_RX_FDIR) + if (pkt_flags & RTE_MBUF_F_RX_FDIR) pkt_flags |= i40e_rxd_build_fdir(&rxd, rxm); #ifdef RTE_LIBRTE_IEEE1588 @@ -798,7 +818,7 @@ i40e_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts) if (nb_hold > rxq->rx_free_thresh) { rx_id = (uint16_t) ((rx_id == 0) ? (rxq->nb_rx_desc - 1) : (rx_id - 1)); - I40E_PCI_REG_WRITE(rxq->qrx_tail, rx_id); + I40E_PCI_REG_WC_WRITE(rxq->qrx_tail, rx_id); nb_hold = 0; } rxq->nb_rx_hold = nb_hold; @@ -846,6 +866,12 @@ i40e_recv_scattered_pkts(void *rx_queue, break; } + /** + * Use acquire fence to ensure that qword1 which includes DD + * bit is loaded before loading of other descriptor words. + */ + rte_atomic_thread_fence(__ATOMIC_ACQUIRE); + rxd = *rxdp; nb_hold++; rxe = &sw_ring[rx_id]; @@ -920,17 +946,17 @@ i40e_recv_scattered_pkts(void *rx_queue, */ rxm->next = NULL; if (unlikely(rxq->crc_len > 0)) { - first_seg->pkt_len -= ETHER_CRC_LEN; - if (rx_packet_len <= ETHER_CRC_LEN) { + first_seg->pkt_len -= RTE_ETHER_CRC_LEN; + if (rx_packet_len <= RTE_ETHER_CRC_LEN) { rte_pktmbuf_free_seg(rxm); first_seg->nb_segs--; last_seg->data_len = (uint16_t)(last_seg->data_len - - (ETHER_CRC_LEN - rx_packet_len)); + (RTE_ETHER_CRC_LEN - rx_packet_len)); last_seg->next = NULL; } else rxm->data_len = (uint16_t)(rx_packet_len - - ETHER_CRC_LEN); + RTE_ETHER_CRC_LEN); } first_seg->port = rxq->port_id; @@ -941,10 +967,10 @@ i40e_recv_scattered_pkts(void *rx_queue, first_seg->packet_type = ptype_tbl[(uint8_t)((qword1 & I40E_RXD_QW1_PTYPE_MASK) >> I40E_RXD_QW1_PTYPE_SHIFT)]; - if (pkt_flags & PKT_RX_RSS_HASH) + if (pkt_flags & RTE_MBUF_F_RX_RSS_HASH) first_seg->hash.rss = rte_le_to_cpu_32(rxd.wb.qword0.hi_dword.rss); - if (pkt_flags & PKT_RX_FDIR) + if (pkt_flags & RTE_MBUF_F_RX_FDIR) pkt_flags |= i40e_rxd_build_fdir(&rxd, first_seg); #ifdef RTE_LIBRTE_IEEE1588 @@ -969,14 +995,14 @@ i40e_recv_scattered_pkts(void *rx_queue, * threshold of the queue, advance the Receive Descriptor Tail (RDT) * register. Update the RDT with the value of the last processed RX * descriptor minus 1, to guarantee that the RDT register is never - * equal to the RDH register, which creates a "full" ring situtation + * equal to the RDH register, which creates a "full" ring situation * from the hardware point of view. */ nb_hold = (uint16_t)(nb_hold + rxq->nb_rx_hold); if (nb_hold > rxq->rx_free_thresh) { rx_id = (uint16_t)(rx_id == 0 ? (rxq->nb_rx_desc - 1) : (rx_id - 1)); - I40E_PCI_REG_WRITE(rxq->qrx_tail, rx_id); + I40E_PCI_REG_WC_WRITE(rxq->qrx_tail, rx_id); nb_hold = 0; } rxq->nb_rx_hold = nb_hold; @@ -988,13 +1014,13 @@ i40e_recv_scattered_pkts(void *rx_queue, static inline uint16_t i40e_calc_context_desc(uint64_t flags) { - static uint64_t mask = PKT_TX_OUTER_IP_CKSUM | - PKT_TX_TCP_SEG | - PKT_TX_QINQ_PKT | - PKT_TX_TUNNEL_MASK; + static uint64_t mask = RTE_MBUF_F_TX_OUTER_IP_CKSUM | + RTE_MBUF_F_TX_TCP_SEG | + RTE_MBUF_F_TX_QINQ | + RTE_MBUF_F_TX_TUNNEL_MASK; #ifdef RTE_LIBRTE_IEEE1588 - mask |= PKT_TX_IEEE1588_TMST; + mask |= RTE_MBUF_F_TX_IEEE1588_TMST; #endif return (flags & mask) ? 1 : 0; @@ -1012,15 +1038,9 @@ i40e_set_tso_ctx(struct rte_mbuf *mbuf, union i40e_tx_offload tx_offload) return ctx_desc; } - /** - * in case of non tunneling packet, the outer_l2_len and - * outer_l3_len must be 0. - */ - hdr_len = tx_offload.outer_l2_len + - tx_offload.outer_l3_len + - tx_offload.l2_len + - tx_offload.l3_len + - tx_offload.l4_len; + hdr_len = tx_offload.l2_len + tx_offload.l3_len + tx_offload.l4_len; + hdr_len += (mbuf->ol_flags & RTE_MBUF_F_TX_TUNNEL_MASK) ? + tx_offload.outer_l2_len + tx_offload.outer_l3_len : 0; cd_cmd = I40E_TX_CTX_DESC_TSO; cd_tso_len = mbuf->pkt_len - hdr_len; @@ -1033,6 +1053,24 @@ i40e_set_tso_ctx(struct rte_mbuf *mbuf, union i40e_tx_offload tx_offload) return ctx_desc; } +/* HW requires that Tx buffer size ranges from 1B up to (16K-1)B. */ +#define I40E_MAX_DATA_PER_TXD \ + (I40E_TXD_QW1_TX_BUF_SZ_MASK >> I40E_TXD_QW1_TX_BUF_SZ_SHIFT) +/* Calculate the number of TX descriptors needed for each pkt */ +static inline uint16_t +i40e_calc_pkt_desc(struct rte_mbuf *tx_pkt) +{ + struct rte_mbuf *txd = tx_pkt; + uint16_t count = 0; + + while (txd != NULL) { + count += DIV_ROUND_UP(txd->data_len, I40E_MAX_DATA_PER_TXD); + txd = txd->next; + } + + return count; +} + uint16_t i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) { @@ -1065,7 +1103,7 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) /* Check if the descriptor ring needs to be cleaned. */ if (txq->nb_tx_free < txq->tx_free_thresh) - i40e_xmit_cleanup(txq); + (void)i40e_xmit_cleanup(txq); for (nb_tx = 0; nb_tx < nb_pkts; nb_tx++) { td_cmd = 0; @@ -1090,8 +1128,15 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) * 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. + * Recalculate the needed tx descs when TSO enabled in case + * the mbuf data size exceeds max data size that hw allows + * per tx desc. */ - nb_used = (uint16_t)(tx_pkt->nb_segs + nb_ctx); + if (ol_flags & RTE_MBUF_F_TX_TCP_SEG) + nb_used = (uint16_t)(i40e_calc_pkt_desc(tx_pkt) + + nb_ctx); + else + nb_used = (uint16_t)(tx_pkt->nb_segs + nb_ctx); tx_last = (uint16_t)(tx_id + nb_used - 1); /* Circular ring */ @@ -1116,7 +1161,7 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) } /* Descriptor based VLAN insertion */ - if (ol_flags & (PKT_TX_VLAN_PKT | PKT_TX_QINQ_PKT)) { + if (ol_flags & (RTE_MBUF_F_TX_VLAN | RTE_MBUF_F_TX_QINQ)) { td_cmd |= I40E_TX_DESC_CMD_IL2TAG1; td_tag = tx_pkt->vlan_tci; } @@ -1126,7 +1171,7 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) /* Fill in tunneling parameters if necessary */ cd_tunneling_params = 0; - if (ol_flags & PKT_TX_TUNNEL_MASK) + if (ol_flags & RTE_MBUF_F_TX_TUNNEL_MASK) i40e_parse_tunneling_params(ol_flags, tx_offload, &cd_tunneling_params); /* Enable checksum offloading */ @@ -1151,12 +1196,12 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) } /* TSO enabled means no timestamp */ - if (ol_flags & PKT_TX_TCP_SEG) + if (ol_flags & RTE_MBUF_F_TX_TCP_SEG) cd_type_cmd_tso_mss |= i40e_set_tso_ctx(tx_pkt, tx_offload); else { #ifdef RTE_LIBRTE_IEEE1588 - if (ol_flags & PKT_TX_IEEE1588_TMST) + if (ol_flags & RTE_MBUF_F_TX_IEEE1588_TMST) cd_type_cmd_tso_mss |= ((uint64_t)I40E_TX_CTX_DESC_TSYN << I40E_TXD_CTX_QW1_CMD_SHIFT); @@ -1165,7 +1210,7 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) ctx_txd->tunneling_params = rte_cpu_to_le_32(cd_tunneling_params); - if (ol_flags & PKT_TX_QINQ_PKT) { + if (ol_flags & RTE_MBUF_F_TX_QINQ) { cd_l2tag2 = tx_pkt->vlan_tci_outer; cd_type_cmd_tso_mss |= ((uint64_t)I40E_TX_CTX_DESC_IL2TAG2 << @@ -1204,6 +1249,24 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) slen = m_seg->data_len; buf_dma_addr = rte_mbuf_data_iova(m_seg); + while ((ol_flags & RTE_MBUF_F_TX_TCP_SEG) && + unlikely(slen > I40E_MAX_DATA_PER_TXD)) { + txd->buffer_addr = + rte_cpu_to_le_64(buf_dma_addr); + txd->cmd_type_offset_bsz = + i40e_build_ctob(td_cmd, + td_offset, I40E_MAX_DATA_PER_TXD, + td_tag); + + buf_dma_addr += I40E_MAX_DATA_PER_TXD; + slen -= I40E_MAX_DATA_PER_TXD; + + txe->last_id = tx_last; + tx_id = txe->next_id; + txe = txn; + txd = &txr[tx_id]; + txn = &sw_ring[txe->next_id]; + } PMD_TX_LOG(DEBUG, "mbuf: %p, TDD[%u]:\n" "buf_dma_addr: %#"PRIx64";\n" "td_cmd: %#x;\n" @@ -1228,10 +1291,10 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) txq->nb_tx_free = (uint16_t)(txq->nb_tx_free - nb_used); if (txq->nb_tx_used >= txq->tx_rs_thresh) { - PMD_TX_FREE_LOG(DEBUG, - "Setting RS bit on TXD id=" - "%4u (port=%d queue=%d)", - tx_last, txq->port_id, txq->queue_id); + 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 |= I40E_TX_DESC_CMD_RS; @@ -1245,13 +1308,12 @@ i40e_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts, uint16_t nb_pkts) } end_of_tx: - rte_wmb(); - PMD_TX_LOG(DEBUG, "port_id=%u queue_id=%u tx_tail=%u nb_tx=%u", (unsigned) txq->port_id, (unsigned) txq->queue_id, (unsigned) tx_id, (unsigned) nb_tx); - I40E_PCI_REG_WRITE_RELAXED(txq->qtx_tail, tx_id); + rte_io_wmb(); + I40E_PCI_REG_WC_WRITE_RELAXED(txq->qtx_tail, tx_id); txq->tx_tail = tx_id; return nb_tx; @@ -1261,22 +1323,40 @@ static __rte_always_inline int i40e_tx_free_bufs(struct i40e_tx_queue *txq) { struct i40e_tx_entry *txep; - uint16_t i; + uint16_t tx_rs_thresh = txq->tx_rs_thresh; + uint16_t i = 0, j = 0; + struct rte_mbuf *free[RTE_I40E_TX_MAX_FREE_BUF_SZ]; + const uint16_t k = RTE_ALIGN_FLOOR(tx_rs_thresh, RTE_I40E_TX_MAX_FREE_BUF_SZ); + const uint16_t m = tx_rs_thresh % RTE_I40E_TX_MAX_FREE_BUF_SZ; if ((txq->tx_ring[txq->tx_next_dd].cmd_type_offset_bsz & rte_cpu_to_le_64(I40E_TXD_QW1_DTYPE_MASK)) != rte_cpu_to_le_64(I40E_TX_DESC_DTYPE_DESC_DONE)) return 0; - txep = &(txq->sw_ring[txq->tx_next_dd - (txq->tx_rs_thresh - 1)]); + txep = &txq->sw_ring[txq->tx_next_dd - (tx_rs_thresh - 1)]; - for (i = 0; i < txq->tx_rs_thresh; i++) + for (i = 0; i < tx_rs_thresh; i++) rte_prefetch0((txep + i)->mbuf); - if (txq->txq_flags & (uint32_t)ETH_TXQ_FLAGS_NOREFCOUNT) { - for (i = 0; i < txq->tx_rs_thresh; ++i, ++txep) { - rte_mempool_put(txep->mbuf->pool, txep->mbuf); - txep->mbuf = NULL; + if (txq->offloads & RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE) { + if (k) { + for (j = 0; j != k; j += RTE_I40E_TX_MAX_FREE_BUF_SZ) { + for (i = 0; i < RTE_I40E_TX_MAX_FREE_BUF_SZ; ++i, ++txep) { + free[i] = txep->mbuf; + txep->mbuf = NULL; + } + rte_mempool_put_bulk(free[0]->pool, (void **)free, + RTE_I40E_TX_MAX_FREE_BUF_SZ); + } + } + + if (m) { + for (i = 0; i < m; ++i, ++txep) { + free[i] = txep->mbuf; + txep->mbuf = NULL; + } + rte_mempool_put_bulk(free[0]->pool, (void **)free, m); } } else { for (i = 0; i < txq->tx_rs_thresh; ++i, ++txep) { @@ -1387,7 +1467,7 @@ tx_xmit_pkts(struct i40e_tx_queue *txq, i40e_tx_fill_hw_ring(txq, tx_pkts + n, (uint16_t)(nb_pkts - n)); txq->tx_tail = (uint16_t)(txq->tx_tail + (nb_pkts - n)); - /* Determin if RS bit needs to be set */ + /* Determine if RS bit needs to be set */ if (txq->tx_tail > txq->tx_next_rs) { txr[txq->tx_next_rs].cmd_type_offset_bsz |= rte_cpu_to_le_64(((uint64_t)I40E_TX_DESC_CMD_RS) << @@ -1402,8 +1482,7 @@ tx_xmit_pkts(struct i40e_tx_queue *txq, txq->tx_tail = 0; /* Update the tx tail register */ - rte_wmb(); - I40E_PCI_REG_WRITE_RELAXED(txq->qtx_tail, txq->tx_tail); + I40E_PCI_REG_WC_WRITE(txq->qtx_tail, txq->tx_tail); return nb_pkts; } @@ -1444,6 +1523,7 @@ i40e_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts, while (nb_pkts) { uint16_t ret, num; + /* cross rs_thresh boundary is not allowed */ num = (uint16_t)RTE_MIN(nb_pkts, txq->tx_rs_thresh); ret = i40e_xmit_fixed_burst_vec(tx_queue, &tx_pkts[nb_tx], num); @@ -1456,6 +1536,43 @@ i40e_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts, return nb_tx; } +/********************************************************************* + * + * TX simple prep functions + * + **********************************************************************/ +uint16_t +i40e_simple_prep_pkts(__rte_unused void *tx_queue, struct rte_mbuf **tx_pkts, + uint16_t nb_pkts) +{ + int i; + uint64_t ol_flags; + struct rte_mbuf *m; + + for (i = 0; i < nb_pkts; i++) { + m = tx_pkts[i]; + ol_flags = m->ol_flags; + + if (m->nb_segs != 1) { + rte_errno = EINVAL; + return i; + } + + if (ol_flags & I40E_TX_OFFLOAD_SIMPLE_NOTSUP_MASK) { + rte_errno = ENOTSUP; + return i; + } + + /* check the size of packet */ + if (m->pkt_len < I40E_TX_MIN_PKT_LEN || + m->pkt_len > I40E_FRAME_SIZE_MAX) { + rte_errno = EINVAL; + return i; + } + } + return i; +} + /********************************************************************* * * TX prep functions @@ -1474,36 +1591,44 @@ i40e_prep_pkts(__rte_unused void *tx_queue, struct rte_mbuf **tx_pkts, ol_flags = m->ol_flags; /* Check for m->nb_segs to not exceed the limits. */ - if (!(ol_flags & PKT_TX_TCP_SEG)) { - if (m->nb_segs > I40E_TX_MAX_SEG || - m->nb_segs > I40E_TX_MAX_MTU_SEG) { - rte_errno = -EINVAL; + if (!(ol_flags & RTE_MBUF_F_TX_TCP_SEG)) { + if (m->nb_segs > I40E_TX_MAX_MTU_SEG || + m->pkt_len > I40E_FRAME_SIZE_MAX) { + rte_errno = EINVAL; return i; } - } else if ((m->tso_segsz < I40E_MIN_TSO_MSS) || - (m->tso_segsz > I40E_MAX_TSO_MSS)) { + } else if (m->nb_segs > I40E_TX_MAX_SEG || + m->tso_segsz < I40E_MIN_TSO_MSS || + m->tso_segsz > I40E_MAX_TSO_MSS || + m->pkt_len > I40E_TSO_FRAME_SIZE_MAX) { /* MSS outside the range (256B - 9674B) are considered * malicious */ - rte_errno = -EINVAL; + rte_errno = EINVAL; return i; } if (ol_flags & I40E_TX_OFFLOAD_NOTSUP_MASK) { - rte_errno = -ENOTSUP; + rte_errno = ENOTSUP; + return i; + } + + /* check the size of packet */ + if (m->pkt_len < I40E_TX_MIN_PKT_LEN) { + rte_errno = EINVAL; return i; } -#ifdef RTE_LIBRTE_ETHDEV_DEBUG +#ifdef RTE_ETHDEV_DEBUG_TX ret = rte_validate_tx_offload(m); if (ret != 0) { - rte_errno = ret; + rte_errno = -ret; return i; } #endif ret = rte_net_intel_cksum_prepare(m); if (ret != 0) { - rte_errno = ret; + rte_errno = -ret; return i; } } @@ -1560,38 +1685,43 @@ int i40e_dev_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id) { struct i40e_rx_queue *rxq; - int err = -1; + int err; struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); PMD_INIT_FUNC_TRACE(); - if (rx_queue_id < dev->data->nb_rx_queues) { - rxq = dev->data->rx_queues[rx_queue_id]; - - err = i40e_alloc_rx_queue_mbufs(rxq); - if (err) { - PMD_DRV_LOG(ERR, "Failed to allocate RX queue mbuf"); - return err; - } + rxq = dev->data->rx_queues[rx_queue_id]; + if (!rxq || !rxq->q_set) { + PMD_DRV_LOG(ERR, "RX queue %u not available or setup", + rx_queue_id); + return -EINVAL; + } - rte_wmb(); + if (rxq->rx_deferred_start) + PMD_DRV_LOG(WARNING, "RX queue %u is deferred start", + rx_queue_id); - /* Init the RX tail regieter. */ - I40E_PCI_REG_WRITE(rxq->qrx_tail, rxq->nb_rx_desc - 1); + err = i40e_alloc_rx_queue_mbufs(rxq); + if (err) { + PMD_DRV_LOG(ERR, "Failed to allocate RX queue mbuf"); + return err; + } - err = i40e_switch_rx_queue(hw, rxq->reg_idx, TRUE); + /* Init the RX tail register. */ + I40E_PCI_REG_WRITE(rxq->qrx_tail, rxq->nb_rx_desc - 1); - if (err) { - PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on", - rx_queue_id); + err = i40e_switch_rx_queue(hw, rxq->reg_idx, TRUE); + if (err) { + PMD_DRV_LOG(ERR, "Failed to switch RX queue %u on", + rx_queue_id); - i40e_rx_queue_release_mbufs(rxq); - i40e_reset_rx_queue(rxq); - } else - dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED; + i40e_rx_queue_release_mbufs(rxq); + i40e_reset_rx_queue(rxq); + return err; } + dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED; - return err; + return 0; } int @@ -1601,24 +1731,26 @@ i40e_dev_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id) int err; struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); - if (rx_queue_id < dev->data->nb_rx_queues) { - rxq = dev->data->rx_queues[rx_queue_id]; - - /* - * rx_queue_id is queue id application refers to, while - * rxq->reg_idx is the real queue index. - */ - err = i40e_switch_rx_queue(hw, rxq->reg_idx, FALSE); + rxq = dev->data->rx_queues[rx_queue_id]; + if (!rxq || !rxq->q_set) { + PMD_DRV_LOG(ERR, "RX queue %u not available or setup", + rx_queue_id); + return -EINVAL; + } - if (err) { - PMD_DRV_LOG(ERR, "Failed to switch RX queue %u off", - rx_queue_id); - return err; - } - i40e_rx_queue_release_mbufs(rxq); - i40e_reset_rx_queue(rxq); - dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED; + /* + * rx_queue_id is queue id application refers to, while + * rxq->reg_idx is the real queue index. + */ + err = i40e_switch_rx_queue(hw, rxq->reg_idx, FALSE); + if (err) { + PMD_DRV_LOG(ERR, "Failed to switch RX queue %u off", + rx_queue_id); + return err; } + i40e_rx_queue_release_mbufs(rxq); + i40e_reset_rx_queue(rxq); + dev->data->rx_queue_state[rx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED; return 0; } @@ -1626,28 +1758,36 @@ i40e_dev_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id) int i40e_dev_tx_queue_start(struct rte_eth_dev *dev, uint16_t tx_queue_id) { - int err = -1; + int err; struct i40e_tx_queue *txq; struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); PMD_INIT_FUNC_TRACE(); - if (tx_queue_id < dev->data->nb_tx_queues) { - txq = dev->data->tx_queues[tx_queue_id]; - - /* - * tx_queue_id is queue id application refers to, while - * rxq->reg_idx is the real queue index. - */ - err = i40e_switch_tx_queue(hw, txq->reg_idx, TRUE); - if (err) - PMD_DRV_LOG(ERR, "Failed to switch TX queue %u on", - tx_queue_id); - else - dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED; + txq = dev->data->tx_queues[tx_queue_id]; + if (!txq || !txq->q_set) { + PMD_DRV_LOG(ERR, "TX queue %u is not available or setup", + tx_queue_id); + return -EINVAL; } - return err; + if (txq->tx_deferred_start) + PMD_DRV_LOG(WARNING, "TX queue %u is deferred start", + tx_queue_id); + + /* + * tx_queue_id is queue id application refers to, while + * rxq->reg_idx is the real queue index. + */ + err = i40e_switch_tx_queue(hw, txq->reg_idx, TRUE); + if (err) { + PMD_DRV_LOG(ERR, "Failed to switch TX queue %u on", + tx_queue_id); + return err; + } + dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STARTED; + + return 0; } int @@ -1657,26 +1797,28 @@ i40e_dev_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id) int err; struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); - if (tx_queue_id < dev->data->nb_tx_queues) { - txq = dev->data->tx_queues[tx_queue_id]; - - /* - * tx_queue_id is queue id application refers to, while - * txq->reg_idx is the real queue index. - */ - err = i40e_switch_tx_queue(hw, txq->reg_idx, FALSE); - - if (err) { - PMD_DRV_LOG(ERR, "Failed to switch TX queue %u of", - tx_queue_id); - return err; - } + txq = dev->data->tx_queues[tx_queue_id]; + if (!txq || !txq->q_set) { + PMD_DRV_LOG(ERR, "TX queue %u is not available or setup", + tx_queue_id); + return -EINVAL; + } - i40e_tx_queue_release_mbufs(txq); - i40e_reset_tx_queue(txq); - dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED; + /* + * tx_queue_id is queue id application refers to, while + * txq->reg_idx is the real queue index. + */ + err = i40e_switch_tx_queue(hw, txq->reg_idx, FALSE); + if (err) { + PMD_DRV_LOG(ERR, "Failed to switch TX queue %u of", + tx_queue_id); + return err; } + i40e_tx_queue_release_mbufs(txq); + i40e_reset_tx_queue(txq); + dev->data->tx_queue_state[tx_queue_id] = RTE_ETH_QUEUE_STATE_STOPPED; + return 0; } @@ -1718,11 +1860,91 @@ i40e_dev_supported_ptypes_get(struct rte_eth_dev *dev) #endif dev->rx_pkt_burst == i40e_recv_scattered_pkts || dev->rx_pkt_burst == i40e_recv_scattered_pkts_vec || - dev->rx_pkt_burst == i40e_recv_pkts_vec) + dev->rx_pkt_burst == i40e_recv_pkts_vec || +#ifdef CC_AVX512_SUPPORT + dev->rx_pkt_burst == i40e_recv_scattered_pkts_vec_avx512 || + dev->rx_pkt_burst == i40e_recv_pkts_vec_avx512 || +#endif + dev->rx_pkt_burst == i40e_recv_scattered_pkts_vec_avx2 || + dev->rx_pkt_burst == i40e_recv_pkts_vec_avx2) return ptypes; return NULL; } +static int +i40e_dev_first_queue(uint16_t idx, void **queues, int num) +{ + uint16_t i; + + for (i = 0; i < num; i++) { + if (i != idx && queues[i]) + return 0; + } + + return 1; +} + +static int +i40e_dev_rx_queue_setup_runtime(struct rte_eth_dev *dev, + struct i40e_rx_queue *rxq) +{ + struct i40e_adapter *ad = + I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private); + int use_def_burst_func = + check_rx_burst_bulk_alloc_preconditions(rxq); + uint16_t buf_size = + (uint16_t)(rte_pktmbuf_data_room_size(rxq->mp) - + RTE_PKTMBUF_HEADROOM); + int use_scattered_rx = + (rxq->max_pkt_len > buf_size); + + if (i40e_rx_queue_init(rxq) != I40E_SUCCESS) { + PMD_DRV_LOG(ERR, + "Failed to do RX queue initialization"); + return -EINVAL; + } + + if (i40e_dev_first_queue(rxq->queue_id, + dev->data->rx_queues, + dev->data->nb_rx_queues)) { + /** + * If it is the first queue to setup, + * set all flags to default and call + * i40e_set_rx_function. + */ + ad->rx_bulk_alloc_allowed = true; + ad->rx_vec_allowed = true; + dev->data->scattered_rx = use_scattered_rx; + if (use_def_burst_func) + ad->rx_bulk_alloc_allowed = false; + i40e_set_rx_function(dev); + return 0; + } else if (ad->rx_vec_allowed && !rte_is_power_of_2(rxq->nb_rx_desc)) { + PMD_DRV_LOG(ERR, "Vector mode is allowed, but descriptor" + " number %d of queue %d isn't power of 2", + rxq->nb_rx_desc, rxq->queue_id); + return -EINVAL; + } + + /* check bulk alloc conflict */ + if (ad->rx_bulk_alloc_allowed && use_def_burst_func) { + PMD_DRV_LOG(ERR, "Can't use default burst."); + return -EINVAL; + } + /* check scattered conflict */ + if (!dev->data->scattered_rx && use_scattered_rx) { + PMD_DRV_LOG(ERR, "Scattered rx is required."); + return -EINVAL; + } + /* check vector conflict */ + if (ad->rx_vec_allowed && i40e_rxq_vec_setup(rxq)) { + PMD_DRV_LOG(ERR, "Failed vector rx setup."); + return -EINVAL; + } + + return 0; +} + int i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, uint16_t queue_idx, @@ -1731,35 +1953,28 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, const struct rte_eth_rxconf *rx_conf, struct rte_mempool *mp) { - struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); struct i40e_adapter *ad = I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private); struct i40e_vsi *vsi; struct i40e_pf *pf = NULL; - struct i40e_vf *vf = NULL; struct i40e_rx_queue *rxq; const struct rte_memzone *rz; uint32_t ring_size; uint16_t len, i; uint16_t reg_idx, base, bsf, tc_mapping; int q_offset, use_def_burst_func = 1; + uint64_t offloads; - if (hw->mac.type == I40E_MAC_VF || hw->mac.type == I40E_MAC_X722_VF) { - vf = I40EVF_DEV_PRIVATE_TO_VF(dev->data->dev_private); - vsi = &vf->vsi; - if (!vsi) - return -EINVAL; - reg_idx = queue_idx; - } else { - pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); - vsi = i40e_pf_get_vsi_by_qindex(pf, queue_idx); - if (!vsi) - return -EINVAL; - q_offset = i40e_get_queue_offset_by_qindex(pf, queue_idx); - if (q_offset < 0) - return -EINVAL; - reg_idx = vsi->base_queue + q_offset; - } + offloads = rx_conf->offloads | dev->data->dev_conf.rxmode.offloads; + + pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); + vsi = i40e_pf_get_vsi_by_qindex(pf, queue_idx); + if (!vsi) + return -EINVAL; + q_offset = i40e_get_queue_offset_by_qindex(pf, queue_idx); + if (q_offset < 0) + return -EINVAL; + reg_idx = vsi->base_queue + q_offset; if (nb_desc % I40E_ALIGN_RING_DESC != 0 || (nb_desc > I40E_MAX_RING_DESC) || @@ -1771,7 +1986,7 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, /* Free memory if needed */ if (dev->data->rx_queues[queue_idx]) { - i40e_dev_rx_queue_release(dev->data->rx_queues[queue_idx]); + i40e_rx_queue_release(dev->data->rx_queues[queue_idx]); dev->data->rx_queues[queue_idx] = NULL; } @@ -1791,13 +2006,16 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, rxq->queue_id = queue_idx; rxq->reg_idx = reg_idx; rxq->port_id = dev->data->port_id; - rxq->crc_len = (uint8_t) ((dev->data->dev_conf.rxmode.hw_strip_crc) ? - 0 : ETHER_CRC_LEN); + if (dev->data->dev_conf.rxmode.offloads & RTE_ETH_RX_OFFLOAD_KEEP_CRC) + rxq->crc_len = RTE_ETHER_CRC_LEN; + else + rxq->crc_len = 0; rxq->drop_en = rx_conf->rx_drop_en; rxq->vsi = vsi; rxq->rx_deferred_start = rx_conf->rx_deferred_start; + rxq->offloads = offloads; - /* Allocate the maximun number of RX ring hardware descriptor. */ + /* Allocate the maximum number of RX ring hardware descriptor. */ len = I40E_MAX_RING_DESC; /** @@ -1812,11 +2030,12 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, rz = rte_eth_dma_zone_reserve(dev, "rx_ring", queue_idx, ring_size, I40E_RING_BASE_ALIGN, socket_id); if (!rz) { - i40e_dev_rx_queue_release(rxq); + i40e_rx_queue_release(rxq); PMD_DRV_LOG(ERR, "Failed to reserve DMA memory for RX"); return -ENOMEM; } + rxq->mz = rz; /* Zero all the descriptors in the ring. */ memset(rz->addr, 0, ring_size); @@ -1832,32 +2051,13 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, RTE_CACHE_LINE_SIZE, socket_id); if (!rxq->sw_ring) { - i40e_dev_rx_queue_release(rxq); + i40e_rx_queue_release(rxq); PMD_DRV_LOG(ERR, "Failed to allocate memory for SW ring"); return -ENOMEM; } i40e_reset_rx_queue(rxq); rxq->q_set = TRUE; - dev->data->rx_queues[queue_idx] = rxq; - - use_def_burst_func = check_rx_burst_bulk_alloc_preconditions(rxq); - - if (!use_def_burst_func) { -#ifdef RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC - PMD_INIT_LOG(DEBUG, "Rx Burst Bulk Alloc Preconditions are " - "satisfied. Rx Burst Bulk Alloc function will be " - "used on port=%d, queue=%d.", - rxq->port_id, rxq->queue_id); -#endif /* RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC */ - } else { - PMD_INIT_LOG(DEBUG, "Rx Burst Bulk Alloc Preconditions are " - "not satisfied, Scattered Rx is requested, " - "or RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC is " - "not enabled on port=%d, queue=%d.", - rxq->port_id, rxq->queue_id); - ad->rx_bulk_alloc_allowed = false; - } for (i = 0; i < I40E_MAX_TRAFFIC_CLASS; i++) { if (!(vsi->enabled_tc & (1 << i))) @@ -1872,11 +2072,51 @@ i40e_dev_rx_queue_setup(struct rte_eth_dev *dev, rxq->dcb_tc = i; } + if (dev->data->dev_started) { + if (i40e_dev_rx_queue_setup_runtime(dev, rxq)) { + i40e_rx_queue_release(rxq); + return -EINVAL; + } + } else { + use_def_burst_func = + check_rx_burst_bulk_alloc_preconditions(rxq); + if (!use_def_burst_func) { +#ifdef RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC + PMD_INIT_LOG(DEBUG, + "Rx Burst Bulk Alloc Preconditions are " + "satisfied. Rx Burst Bulk Alloc function will be " + "used on port=%d, queue=%d.", + rxq->port_id, rxq->queue_id); +#endif /* RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC */ + } else { + PMD_INIT_LOG(DEBUG, + "Rx Burst Bulk Alloc Preconditions are " + "not satisfied, Scattered Rx is requested, " + "or RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC is " + "not enabled on port=%d, queue=%d.", + rxq->port_id, rxq->queue_id); + ad->rx_bulk_alloc_allowed = false; + } + } + + dev->data->rx_queues[queue_idx] = rxq; return 0; } void -i40e_dev_rx_queue_release(void *rxq) +i40e_dev_rx_queue_release(struct rte_eth_dev *dev, uint16_t qid) +{ + i40e_rx_queue_release(dev->data->rx_queues[qid]); +} + +void +i40e_dev_tx_queue_release(struct rte_eth_dev *dev, uint16_t qid) +{ + i40e_tx_queue_release(dev->data->tx_queues[qid]); +} + +void +i40e_rx_queue_release(void *rxq) { struct i40e_rx_queue *q = (struct i40e_rx_queue *)rxq; @@ -1887,18 +2127,19 @@ i40e_dev_rx_queue_release(void *rxq) i40e_rx_queue_release_mbufs(q); rte_free(q->sw_ring); + rte_memzone_free(q->mz); rte_free(q); } uint32_t -i40e_dev_rx_queue_count(struct rte_eth_dev *dev, uint16_t rx_queue_id) +i40e_dev_rx_queue_count(void *rx_queue) { #define I40E_RXQ_SCAN_INTERVAL 4 volatile union i40e_rx_desc *rxdp; struct i40e_rx_queue *rxq; uint16_t desc = 0; - rxq = dev->data->rx_queues[rx_queue_id]; + rxq = rx_queue; rxdp = &(rxq->rx_ring[rxq->rx_tail]); while ((desc < rxq->nb_rx_desc) && ((rte_le_to_cpu_64(rxdp->wb.qword1.status_error_len) & @@ -1919,32 +2160,6 @@ i40e_dev_rx_queue_count(struct rte_eth_dev *dev, uint16_t rx_queue_id) return desc; } -int -i40e_dev_rx_descriptor_done(void *rx_queue, uint16_t offset) -{ - volatile union i40e_rx_desc *rxdp; - struct i40e_rx_queue *rxq = rx_queue; - uint16_t desc; - int ret; - - if (unlikely(offset >= rxq->nb_rx_desc)) { - PMD_DRV_LOG(ERR, "Invalid RX descriptor id %u", offset); - return 0; - } - - desc = rxq->rx_tail + offset; - if (desc >= rxq->nb_rx_desc) - desc -= rxq->nb_rx_desc; - - rxdp = &(rxq->rx_ring[desc]); - - ret = !!(((rte_le_to_cpu_64(rxdp->wb.qword1.status_error_len) & - I40E_RXD_QW1_STATUS_MASK) >> I40E_RXD_QW1_STATUS_SHIFT) & - (1 << I40E_RX_DESC_STATUS_DD_SHIFT)); - - return ret; -} - int i40e_dev_rx_descriptor_status(void *rx_queue, uint16_t offset) { @@ -2003,6 +2218,52 @@ i40e_dev_tx_descriptor_status(void *tx_queue, uint16_t offset) return RTE_ETH_TX_DESC_FULL; } +static int +i40e_dev_tx_queue_setup_runtime(struct rte_eth_dev *dev, + struct i40e_tx_queue *txq) +{ + struct i40e_adapter *ad = + I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private); + + if (i40e_tx_queue_init(txq) != I40E_SUCCESS) { + PMD_DRV_LOG(ERR, + "Failed to do TX queue initialization"); + return -EINVAL; + } + + if (i40e_dev_first_queue(txq->queue_id, + dev->data->tx_queues, + dev->data->nb_tx_queues)) { + /** + * If it is the first queue to setup, + * set all flags and call + * i40e_set_tx_function. + */ + i40e_set_tx_function_flag(dev, txq); + i40e_set_tx_function(dev); + return 0; + } + + /* check vector conflict */ + if (ad->tx_vec_allowed) { + if (txq->tx_rs_thresh > RTE_I40E_TX_MAX_FREE_BUF_SZ || + i40e_txq_vec_setup(txq)) { + PMD_DRV_LOG(ERR, "Failed vector tx setup."); + return -EINVAL; + } + } + /* check simple tx conflict */ + if (ad->tx_simple_allowed) { + if ((txq->offloads & ~RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE) != 0 || + txq->tx_rs_thresh < RTE_PMD_I40E_TX_MAX_BURST) { + PMD_DRV_LOG(ERR, "No-simple tx is required."); + return -EINVAL; + } + } + + return 0; +} + int i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, uint16_t queue_idx, @@ -2010,33 +2271,26 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, unsigned int socket_id, const struct rte_eth_txconf *tx_conf) { - struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data->dev_private); struct i40e_vsi *vsi; struct i40e_pf *pf = NULL; - struct i40e_vf *vf = NULL; struct i40e_tx_queue *txq; const struct rte_memzone *tz; uint32_t ring_size; uint16_t tx_rs_thresh, tx_free_thresh; uint16_t reg_idx, i, base, bsf, tc_mapping; int q_offset; + uint64_t offloads; - if (hw->mac.type == I40E_MAC_VF || hw->mac.type == I40E_MAC_X722_VF) { - vf = I40EVF_DEV_PRIVATE_TO_VF(dev->data->dev_private); - vsi = &vf->vsi; - if (!vsi) - return -EINVAL; - reg_idx = queue_idx; - } else { - pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); - vsi = i40e_pf_get_vsi_by_qindex(pf, queue_idx); - if (!vsi) - return -EINVAL; - q_offset = i40e_get_queue_offset_by_qindex(pf, queue_idx); - if (q_offset < 0) - return -EINVAL; - reg_idx = vsi->base_queue + q_offset; - } + offloads = tx_conf->offloads | dev->data->dev_conf.txmode.offloads; + + pf = I40E_DEV_PRIVATE_TO_PF(dev->data->dev_private); + vsi = i40e_pf_get_vsi_by_qindex(pf, queue_idx); + if (!vsi) + return -EINVAL; + q_offset = i40e_get_queue_offset_by_qindex(pf, queue_idx); + if (q_offset < 0) + return -EINVAL; + reg_idx = vsi->base_queue + q_offset; if (nb_desc % I40E_ALIGN_RING_DESC != 0 || (nb_desc > I40E_MAX_RING_DESC) || @@ -2061,15 +2315,30 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, * - tx_rs_thresh must be a divisor of the ring size. * - tx_free_thresh must be greater than 0. * - tx_free_thresh must be less than the size of the ring minus 3. + * - tx_free_thresh + tx_rs_thresh must not exceed nb_desc. * * One descriptor in the TX ring is used as a sentinel to avoid a H/W * race condition, hence the maximum threshold constraints. When set * to zero use default values. */ - tx_rs_thresh = (uint16_t)((tx_conf->tx_rs_thresh) ? - tx_conf->tx_rs_thresh : DEFAULT_TX_RS_THRESH); tx_free_thresh = (uint16_t)((tx_conf->tx_free_thresh) ? tx_conf->tx_free_thresh : DEFAULT_TX_FREE_THRESH); + /* force tx_rs_thresh to adapt an aggressive tx_free_thresh */ + tx_rs_thresh = (DEFAULT_TX_RS_THRESH + tx_free_thresh > nb_desc) ? + nb_desc - tx_free_thresh : DEFAULT_TX_RS_THRESH; + if (tx_conf->tx_rs_thresh > 0) + tx_rs_thresh = tx_conf->tx_rs_thresh; + if (tx_rs_thresh + tx_free_thresh > nb_desc) { + PMD_INIT_LOG(ERR, "tx_rs_thresh + tx_free_thresh must not " + "exceed nb_desc. (tx_rs_thresh=%u " + "tx_free_thresh=%u nb_desc=%u port=%d queue=%d)", + (unsigned int)tx_rs_thresh, + (unsigned int)tx_free_thresh, + (unsigned int)nb_desc, + (int)dev->data->port_id, + (int)queue_idx); + return I40E_ERR_PARAM; + } if (tx_rs_thresh >= (nb_desc - 2)) { PMD_INIT_LOG(ERR, "tx_rs_thresh must be less than the " "number of TX descriptors minus 2. " @@ -2119,7 +2388,7 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, /* Free memory if needed. */ if (dev->data->tx_queues[queue_idx]) { - i40e_dev_tx_queue_release(dev->data->tx_queues[queue_idx]); + i40e_tx_queue_release(dev->data->tx_queues[queue_idx]); dev->data->tx_queues[queue_idx] = NULL; } @@ -2140,11 +2409,12 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, tz = rte_eth_dma_zone_reserve(dev, "tx_ring", queue_idx, ring_size, I40E_RING_BASE_ALIGN, socket_id); if (!tz) { - i40e_dev_tx_queue_release(txq); + i40e_tx_queue_release(txq); PMD_DRV_LOG(ERR, "Failed to reserve DMA memory for TX"); return -ENOMEM; } + txq->mz = tz; txq->nb_tx_desc = nb_desc; txq->tx_rs_thresh = tx_rs_thresh; txq->tx_free_thresh = tx_free_thresh; @@ -2154,7 +2424,7 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, txq->queue_id = queue_idx; txq->reg_idx = reg_idx; txq->port_id = dev->data->port_id; - txq->txq_flags = tx_conf->txq_flags; + txq->offloads = offloads; txq->vsi = vsi; txq->tx_deferred_start = tx_conf->tx_deferred_start; @@ -2168,17 +2438,13 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, RTE_CACHE_LINE_SIZE, socket_id); if (!txq->sw_ring) { - i40e_dev_tx_queue_release(txq); + i40e_tx_queue_release(txq); PMD_DRV_LOG(ERR, "Failed to allocate memory for SW TX ring"); return -ENOMEM; } i40e_reset_tx_queue(txq); txq->q_set = TRUE; - dev->data->tx_queues[queue_idx] = txq; - - /* Use a simple TX queue without offloads or multi segs if possible */ - i40e_set_tx_function_flag(dev, txq); for (i = 0; i < I40E_MAX_TRAFFIC_CLASS; i++) { if (!(vsi->enabled_tc & (1 << i))) @@ -2193,11 +2459,25 @@ i40e_dev_tx_queue_setup(struct rte_eth_dev *dev, txq->dcb_tc = i; } + if (dev->data->dev_started) { + if (i40e_dev_tx_queue_setup_runtime(dev, txq)) { + i40e_tx_queue_release(txq); + return -EINVAL; + } + } else { + /** + * Use a simple TX queue without offloads or + * multi segs if possible + */ + i40e_set_tx_function_flag(dev, txq); + } + dev->data->tx_queues[queue_idx] = txq; + return 0; } void -i40e_dev_tx_queue_release(void *txq) +i40e_tx_queue_release(void *txq) { struct i40e_tx_queue *q = (struct i40e_tx_queue *)txq; @@ -2208,6 +2488,7 @@ i40e_dev_tx_queue_release(void *txq) i40e_tx_queue_release_mbufs(q); rte_free(q->sw_ring); + rte_memzone_free(q->mz); rte_free(q); } @@ -2220,8 +2501,8 @@ i40e_memzone_reserve(const char *name, uint32_t len, int socket_id) if (mz) return mz; - mz = rte_memzone_reserve_aligned(name, len, - socket_id, 0, I40E_RING_BASE_ALIGN); + mz = rte_memzone_reserve_aligned(name, len, socket_id, + RTE_MEMZONE_IOVA_CONTIG, I40E_RING_BASE_ALIGN); return mz; } @@ -2292,6 +2573,9 @@ i40e_reset_rx_queue(struct i40e_rx_queue *rxq) #endif /* RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC */ rxq->rx_tail = 0; rxq->nb_rx_hold = 0; + + rte_pktmbuf_free(rxq->pkt_first_seg); + rxq->pkt_first_seg = NULL; rxq->pkt_last_seg = NULL; @@ -2305,18 +2589,38 @@ i40e_tx_queue_release_mbufs(struct i40e_tx_queue *txq) struct rte_eth_dev *dev; uint16_t i; - dev = &rte_eth_devices[txq->port_id]; - if (!txq || !txq->sw_ring) { - PMD_DRV_LOG(DEBUG, "Pointer to rxq or sw_ring is NULL"); + PMD_DRV_LOG(DEBUG, "Pointer to txq or sw_ring is NULL"); return; } + dev = &rte_eth_devices[txq->port_id]; + /** * vPMD tx will not set sw_ring's mbuf to NULL after free, * so need to free remains more carefully. */ - if (dev->tx_pkt_burst == i40e_xmit_pkts_vec) { +#ifdef CC_AVX512_SUPPORT + if (dev->tx_pkt_burst == i40e_xmit_pkts_vec_avx512) { + struct i40e_vec_tx_entry *swr = (void *)txq->sw_ring; + + i = txq->tx_next_dd - txq->tx_rs_thresh + 1; + if (txq->tx_tail < i) { + for (; i < txq->nb_tx_desc; i++) { + rte_pktmbuf_free_seg(swr[i].mbuf); + swr[i].mbuf = NULL; + } + i = 0; + } + for (; i < txq->tx_tail; i++) { + rte_pktmbuf_free_seg(swr[i].mbuf); + swr[i].mbuf = NULL; + } + return; + } +#endif + if (dev->tx_pkt_burst == i40e_xmit_pkts_vec_avx2 || + dev->tx_pkt_burst == i40e_xmit_pkts_vec) { i = txq->tx_next_dd - txq->tx_rs_thresh + 1; if (txq->tx_tail < i) { for (; i < txq->nb_tx_desc; i++) { @@ -2339,6 +2643,113 @@ i40e_tx_queue_release_mbufs(struct i40e_tx_queue *txq) } } +static int +i40e_tx_done_cleanup_full(struct i40e_tx_queue *txq, + uint32_t free_cnt) +{ + struct i40e_tx_entry *swr_ring = txq->sw_ring; + uint16_t i, tx_last, tx_id; + uint16_t nb_tx_free_last; + uint16_t nb_tx_to_clean; + uint32_t pkt_cnt; + + /* Start free mbuf from the next of tx_tail */ + tx_last = txq->tx_tail; + tx_id = swr_ring[tx_last].next_id; + + if (txq->nb_tx_free == 0 && i40e_xmit_cleanup(txq)) + return 0; + + nb_tx_to_clean = txq->nb_tx_free; + nb_tx_free_last = txq->nb_tx_free; + if (!free_cnt) + free_cnt = txq->nb_tx_desc; + + /* Loop through swr_ring to count the amount of + * freeable mubfs and packets. + */ + for (pkt_cnt = 0; pkt_cnt < free_cnt; ) { + for (i = 0; i < nb_tx_to_clean && + pkt_cnt < free_cnt && + tx_id != tx_last; i++) { + if (swr_ring[tx_id].mbuf != NULL) { + rte_pktmbuf_free_seg(swr_ring[tx_id].mbuf); + swr_ring[tx_id].mbuf = NULL; + + /* + * last segment in the packet, + * increment packet count + */ + pkt_cnt += (swr_ring[tx_id].last_id == tx_id); + } + + tx_id = swr_ring[tx_id].next_id; + } + + if (txq->tx_rs_thresh > txq->nb_tx_desc - + txq->nb_tx_free || tx_id == tx_last) + break; + + if (pkt_cnt < free_cnt) { + if (i40e_xmit_cleanup(txq)) + break; + + nb_tx_to_clean = txq->nb_tx_free - nb_tx_free_last; + nb_tx_free_last = txq->nb_tx_free; + } + } + + return (int)pkt_cnt; +} + +static int +i40e_tx_done_cleanup_simple(struct i40e_tx_queue *txq, + uint32_t free_cnt) +{ + int i, n, cnt; + + if (free_cnt == 0 || free_cnt > txq->nb_tx_desc) + free_cnt = txq->nb_tx_desc; + + cnt = free_cnt - free_cnt % txq->tx_rs_thresh; + + for (i = 0; i < cnt; i += n) { + if (txq->nb_tx_desc - txq->nb_tx_free < txq->tx_rs_thresh) + break; + + n = i40e_tx_free_bufs(txq); + + if (n == 0) + break; + } + + return i; +} + +static int +i40e_tx_done_cleanup_vec(struct i40e_tx_queue *txq __rte_unused, + uint32_t free_cnt __rte_unused) +{ + return -ENOTSUP; +} +int +i40e_tx_done_cleanup(void *txq, uint32_t free_cnt) +{ + struct i40e_tx_queue *q = (struct i40e_tx_queue *)txq; + struct rte_eth_dev *dev = &rte_eth_devices[q->port_id]; + struct i40e_adapter *ad = + I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private); + + if (ad->tx_simple_allowed) { + if (ad->tx_vec_allowed) + return i40e_tx_done_cleanup_vec(q, free_cnt); + else + return i40e_tx_done_cleanup_simple(q, free_cnt); + } else { + return i40e_tx_done_cleanup_full(q, free_cnt); + } +} + void i40e_reset_tx_queue(struct i40e_tx_queue *txq) { @@ -2474,7 +2885,7 @@ i40e_rx_queue_config(struct i40e_rx_queue *rxq) struct i40e_pf *pf = I40E_VSI_TO_PF(rxq->vsi); struct i40e_hw *hw = I40E_VSI_TO_HW(rxq->vsi); struct rte_eth_dev_data *data = pf->dev_data; - uint16_t buf_size, len; + uint16_t buf_size; buf_size = (uint16_t)(rte_pktmbuf_data_room_size(rxq->mp) - RTE_PKTMBUF_HEADROOM); @@ -2497,28 +2908,16 @@ i40e_rx_queue_config(struct i40e_rx_queue *rxq) break; } - len = hw->func_caps.rx_buf_chain_len * rxq->rx_buf_len; - rxq->max_pkt_len = RTE_MIN(len, data->dev_conf.rxmode.max_rx_pkt_len); - if (data->dev_conf.rxmode.jumbo_frame == 1) { - if (rxq->max_pkt_len <= ETHER_MAX_LEN || - rxq->max_pkt_len > I40E_FRAME_SIZE_MAX) { - PMD_DRV_LOG(ERR, "maximum packet length must " - "be larger than %u and smaller than %u," - "as jumbo frame is enabled", - (uint32_t)ETHER_MAX_LEN, - (uint32_t)I40E_FRAME_SIZE_MAX); - return I40E_ERR_CONFIG; - } - } else { - if (rxq->max_pkt_len < ETHER_MIN_LEN || - rxq->max_pkt_len > ETHER_MAX_LEN) { - PMD_DRV_LOG(ERR, "maximum packet length must be " - "larger than %u and smaller than %u, " - "as jumbo frame is disabled", - (uint32_t)ETHER_MIN_LEN, - (uint32_t)ETHER_MAX_LEN); - return I40E_ERR_CONFIG; - } + rxq->max_pkt_len = + RTE_MIN(hw->func_caps.rx_buf_chain_len * rxq->rx_buf_len, + data->mtu + I40E_ETH_OVERHEAD); + if (rxq->max_pkt_len < RTE_ETHER_MIN_LEN || + rxq->max_pkt_len > I40E_FRAME_SIZE_MAX) { + PMD_DRV_LOG(ERR, "maximum packet length must be " + "larger than %u and smaller than %u", + (uint32_t)RTE_ETHER_MIN_LEN, + (uint32_t)I40E_FRAME_SIZE_MAX); + return I40E_ERR_CONFIG; } return 0; @@ -2589,11 +2988,10 @@ i40e_rx_queue_init(struct i40e_rx_queue *rxq) RTE_PKTMBUF_HEADROOM); /* Check if scattered RX needs to be used. */ - if ((rxq->max_pkt_len + 2 * I40E_VLAN_TAG_SIZE) > buf_size) { + if (rxq->max_pkt_len > buf_size) dev_data->scattered_rx = 1; - } - /* Init the RX tail regieter. */ + /* Init the RX tail register. */ I40E_PCI_REG_WRITE(rxq->qrx_tail, rxq->nb_rx_desc - 1); return 0; @@ -2631,37 +3029,32 @@ i40e_dev_free_queues(struct rte_eth_dev *dev) for (i = 0; i < dev->data->nb_rx_queues; i++) { if (!dev->data->rx_queues[i]) continue; - i40e_dev_rx_queue_release(dev->data->rx_queues[i]); + i40e_rx_queue_release(dev->data->rx_queues[i]); dev->data->rx_queues[i] = NULL; } - dev->data->nb_rx_queues = 0; for (i = 0; i < dev->data->nb_tx_queues; i++) { if (!dev->data->tx_queues[i]) continue; - i40e_dev_tx_queue_release(dev->data->tx_queues[i]); + i40e_tx_queue_release(dev->data->tx_queues[i]); dev->data->tx_queues[i] = NULL; } - dev->data->nb_tx_queues = 0; } -#define I40E_FDIR_NUM_TX_DESC I40E_MIN_RING_DESC -#define I40E_FDIR_NUM_RX_DESC I40E_MIN_RING_DESC - enum i40e_status_code i40e_fdir_setup_tx_resources(struct i40e_pf *pf) { struct i40e_tx_queue *txq; const struct rte_memzone *tz = NULL; - uint32_t ring_size; struct rte_eth_dev *dev; + uint32_t ring_size; if (!pf) { PMD_DRV_LOG(ERR, "PF is not available"); return I40E_ERR_BAD_PTR; } - dev = pf->adapter->eth_dev; + dev = &rte_eth_devices[pf->dev_data->port_id]; /* Allocate the TX queue data structure. */ txq = rte_zmalloc_socket("i40e fdir tx queue", @@ -2682,11 +3075,12 @@ i40e_fdir_setup_tx_resources(struct i40e_pf *pf) I40E_FDIR_QUEUE_ID, ring_size, I40E_RING_BASE_ALIGN, SOCKET_ID_ANY); if (!tz) { - i40e_dev_tx_queue_release(txq); + i40e_tx_queue_release(txq); PMD_DRV_LOG(ERR, "Failed to reserve DMA memory for TX."); return I40E_ERR_NO_MEMORY; } + txq->mz = tz; txq->nb_tx_desc = I40E_FDIR_NUM_TX_DESC; txq->queue_id = I40E_FDIR_QUEUE_ID; txq->reg_idx = pf->fdir.fdir_vsi->base_queue; @@ -2694,12 +3088,14 @@ i40e_fdir_setup_tx_resources(struct i40e_pf *pf) txq->tx_ring_phys_addr = tz->iova; txq->tx_ring = (struct i40e_tx_desc *)tz->addr; + /* * don't need to allocate software ring and reset for the fdir * program queue just set the queue has been configured. */ txq->q_set = TRUE; pf->fdir.txq = txq; + pf->fdir.txq_available_buf_count = I40E_FDIR_PRG_PKT_CNT; return I40E_SUCCESS; } @@ -2717,7 +3113,7 @@ i40e_fdir_setup_rx_resources(struct i40e_pf *pf) return I40E_ERR_BAD_PTR; } - dev = pf->adapter->eth_dev; + dev = &rte_eth_devices[pf->dev_data->port_id]; /* Allocate the RX queue data structure. */ rxq = rte_zmalloc_socket("i40e fdir rx queue", @@ -2738,17 +3134,19 @@ i40e_fdir_setup_rx_resources(struct i40e_pf *pf) I40E_FDIR_QUEUE_ID, ring_size, I40E_RING_BASE_ALIGN, SOCKET_ID_ANY); if (!rz) { - i40e_dev_rx_queue_release(rxq); + i40e_rx_queue_release(rxq); PMD_DRV_LOG(ERR, "Failed to reserve DMA memory for RX."); return I40E_ERR_NO_MEMORY; } + rxq->mz = rz; rxq->nb_rx_desc = I40E_FDIR_NUM_RX_DESC; rxq->queue_id = I40E_FDIR_QUEUE_ID; rxq->reg_idx = pf->fdir.fdir_vsi->base_queue; rxq->vsi = pf->fdir.fdir_vsi; rxq->rx_ring_phys_addr = rz->iova; + memset(rz->addr, 0, I40E_FDIR_NUM_RX_DESC * sizeof(union i40e_rx_desc)); rxq->rx_ring = (union i40e_rx_desc *)rz->addr; /* @@ -2776,6 +3174,7 @@ i40e_rxq_info_get(struct rte_eth_dev *dev, uint16_t queue_id, qinfo->conf.rx_free_thresh = rxq->rx_free_thresh; qinfo->conf.rx_drop_en = rxq->drop_en; qinfo->conf.rx_deferred_start = rxq->rx_deferred_start; + qinfo->conf.offloads = rxq->offloads; } void @@ -2794,11 +3193,44 @@ i40e_txq_info_get(struct rte_eth_dev *dev, uint16_t queue_id, qinfo->conf.tx_free_thresh = txq->tx_free_thresh; qinfo->conf.tx_rs_thresh = txq->tx_rs_thresh; - qinfo->conf.txq_flags = txq->txq_flags; qinfo->conf.tx_deferred_start = txq->tx_deferred_start; + qinfo->conf.offloads = txq->offloads; } -void __attribute__((cold)) +#ifdef RTE_ARCH_X86 +static inline bool +get_avx_supported(bool request_avx512) +{ + if (request_avx512) { + if (rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_512 && + rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX512F) == 1 && + rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX512BW) == 1) +#ifdef CC_AVX512_SUPPORT + return true; +#else + PMD_DRV_LOG(NOTICE, + "AVX512 is not supported in build env"); + return false; +#endif + } else { + if (rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_256 && + rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX2) == 1 && + rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX512F) == 1) +#ifdef CC_AVX2_SUPPORT + return true; +#else + PMD_DRV_LOG(NOTICE, + "AVX2 is not supported in build env"); + return false; +#endif + } + + return false; +} +#endif /* RTE_ARCH_X86 */ + + +void __rte_cold i40e_set_rx_function(struct rte_eth_dev *dev) { struct i40e_adapter *ad = @@ -2808,6 +3240,10 @@ i40e_set_rx_function(struct rte_eth_dev *dev) * conditions to be met and Rx Bulk Allocation should be allowed. */ if (rte_eal_process_type() == RTE_PROC_PRIMARY) { +#ifdef RTE_ARCH_X86 + ad->rx_use_avx512 = false; + ad->rx_use_avx2 = false; +#endif if (i40e_rx_vec_dev_conf_condition_check(dev) || !ad->rx_bulk_alloc_allowed) { PMD_INIT_LOG(DEBUG, "Port[%d] doesn't meet" @@ -2826,39 +3262,68 @@ i40e_set_rx_function(struct rte_eth_dev *dev) break; } } +#ifdef RTE_ARCH_X86 + ad->rx_use_avx512 = get_avx_supported(1); + + if (!ad->rx_use_avx512) + ad->rx_use_avx2 = get_avx_supported(0); +#endif } } - if (dev->data->scattered_rx) { - /* Set the non-LRO scattered callback: there are Vector and - * single allocation versions. - */ - if (ad->rx_vec_allowed) { - PMD_INIT_LOG(DEBUG, "Using Vector Scattered Rx " - "callback (port=%d).", + if (ad->rx_vec_allowed && + rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_128) { +#ifdef RTE_ARCH_X86 + if (dev->data->scattered_rx) { + if (ad->rx_use_avx512) { +#ifdef CC_AVX512_SUPPORT + PMD_DRV_LOG(NOTICE, + "Using AVX512 Vector Scattered Rx (port %d).", + dev->data->port_id); + dev->rx_pkt_burst = + i40e_recv_scattered_pkts_vec_avx512; +#endif + } else { + PMD_INIT_LOG(DEBUG, + "Using %sVector Scattered Rx (port %d).", + ad->rx_use_avx2 ? "avx2 " : "", + dev->data->port_id); + dev->rx_pkt_burst = ad->rx_use_avx2 ? + i40e_recv_scattered_pkts_vec_avx2 : + i40e_recv_scattered_pkts_vec; + } + } else { + if (ad->rx_use_avx512) { +#ifdef CC_AVX512_SUPPORT + PMD_DRV_LOG(NOTICE, + "Using AVX512 Vector Rx (port %d).", + dev->data->port_id); + dev->rx_pkt_burst = + i40e_recv_pkts_vec_avx512; +#endif + } else { + PMD_INIT_LOG(DEBUG, + "Using %sVector Rx (port %d).", + ad->rx_use_avx2 ? "avx2 " : "", + dev->data->port_id); + dev->rx_pkt_burst = ad->rx_use_avx2 ? + i40e_recv_pkts_vec_avx2 : + i40e_recv_pkts_vec; + } + } +#else /* RTE_ARCH_X86 */ + if (dev->data->scattered_rx) { + PMD_INIT_LOG(DEBUG, + "Using Vector Scattered Rx (port %d).", dev->data->port_id); - dev->rx_pkt_burst = i40e_recv_scattered_pkts_vec; } else { - PMD_INIT_LOG(DEBUG, "Using a Scattered with bulk " - "allocation callback (port=%d).", + PMD_INIT_LOG(DEBUG, "Using Vector Rx (port %d).", dev->data->port_id); - dev->rx_pkt_burst = i40e_recv_scattered_pkts; + dev->rx_pkt_burst = i40e_recv_pkts_vec; } - /* If parameters allow we are going to choose between the following - * callbacks: - * - Vector - * - Bulk Allocation - * - Single buffer allocation (the simplest one) - */ - } else if (ad->rx_vec_allowed) { - PMD_INIT_LOG(DEBUG, "Vector rx enabled, please make sure RX " - "burst size no less than %d (port=%d).", - RTE_I40E_DESCS_PER_LOOP, - dev->data->port_id); - - dev->rx_pkt_burst = i40e_recv_pkts_vec; - } else if (ad->rx_bulk_alloc_allowed) { +#endif /* RTE_ARCH_X86 */ + } else if (!dev->data->scattered_rx && ad->rx_bulk_alloc_allowed) { PMD_INIT_LOG(DEBUG, "Rx Burst Bulk Alloc Preconditions are " "satisfied. Rx Burst Bulk Alloc function " "will be used on port=%d.", @@ -2866,19 +3331,25 @@ i40e_set_rx_function(struct rte_eth_dev *dev) dev->rx_pkt_burst = i40e_recv_pkts_bulk_alloc; } else { - PMD_INIT_LOG(DEBUG, "Rx Burst Bulk Alloc Preconditions are not " - "satisfied, or Scattered Rx is requested " - "(port=%d).", + /* Simple Rx Path. */ + PMD_INIT_LOG(DEBUG, "Simple Rx path will be used on port=%d.", dev->data->port_id); - - dev->rx_pkt_burst = i40e_recv_pkts; + dev->rx_pkt_burst = dev->data->scattered_rx ? + i40e_recv_scattered_pkts : + i40e_recv_pkts; } /* Propagate information about RX function choice through all queues. */ if (rte_eal_process_type() == RTE_PROC_PRIMARY) { rx_using_sse = (dev->rx_pkt_burst == i40e_recv_scattered_pkts_vec || - dev->rx_pkt_burst == i40e_recv_pkts_vec); + dev->rx_pkt_burst == i40e_recv_pkts_vec || +#ifdef CC_AVX512_SUPPORT + dev->rx_pkt_burst == i40e_recv_scattered_pkts_vec_avx512 || + dev->rx_pkt_burst == i40e_recv_pkts_vec_avx512 || +#endif + dev->rx_pkt_burst == i40e_recv_scattered_pkts_vec_avx2 || + dev->rx_pkt_burst == i40e_recv_pkts_vec_avx2); for (i = 0; i < dev->data->nb_rx_queues; i++) { struct i40e_rx_queue *rxq = dev->data->rx_queues[i]; @@ -2889,28 +3360,78 @@ i40e_set_rx_function(struct rte_eth_dev *dev) } } -void __attribute__((cold)) +static const struct { + eth_rx_burst_t pkt_burst; + const char *info; +} i40e_rx_burst_infos[] = { + { i40e_recv_scattered_pkts, "Scalar Scattered" }, + { i40e_recv_pkts_bulk_alloc, "Scalar Bulk Alloc" }, + { i40e_recv_pkts, "Scalar" }, +#ifdef RTE_ARCH_X86 +#ifdef CC_AVX512_SUPPORT + { i40e_recv_scattered_pkts_vec_avx512, "Vector AVX512 Scattered" }, + { i40e_recv_pkts_vec_avx512, "Vector AVX512" }, +#endif + { i40e_recv_scattered_pkts_vec_avx2, "Vector AVX2 Scattered" }, + { i40e_recv_pkts_vec_avx2, "Vector AVX2" }, + { i40e_recv_scattered_pkts_vec, "Vector SSE Scattered" }, + { i40e_recv_pkts_vec, "Vector SSE" }, +#elif defined(RTE_ARCH_ARM64) + { i40e_recv_scattered_pkts_vec, "Vector Neon Scattered" }, + { i40e_recv_pkts_vec, "Vector Neon" }, +#elif defined(RTE_ARCH_PPC_64) + { i40e_recv_scattered_pkts_vec, "Vector AltiVec Scattered" }, + { i40e_recv_pkts_vec, "Vector AltiVec" }, +#endif +}; + +int +i40e_rx_burst_mode_get(struct rte_eth_dev *dev, __rte_unused uint16_t queue_id, + struct rte_eth_burst_mode *mode) +{ + eth_rx_burst_t pkt_burst = dev->rx_pkt_burst; + int ret = -EINVAL; + unsigned int i; + + for (i = 0; i < RTE_DIM(i40e_rx_burst_infos); ++i) { + if (pkt_burst == i40e_rx_burst_infos[i].pkt_burst) { + snprintf(mode->info, sizeof(mode->info), "%s", + i40e_rx_burst_infos[i].info); + ret = 0; + break; + } + } + + return ret; +} + +void __rte_cold i40e_set_tx_function_flag(struct rte_eth_dev *dev, struct i40e_tx_queue *txq) { struct i40e_adapter *ad = I40E_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private); - /* Use a simple Tx queue (no offloads, no multi segs) if possible */ - if (((txq->txq_flags & I40E_SIMPLE_FLAGS) == I40E_SIMPLE_FLAGS) - && (txq->tx_rs_thresh >= RTE_PMD_I40E_TX_MAX_BURST)) { - if (txq->tx_rs_thresh <= RTE_I40E_TX_MAX_FREE_BUF_SZ) { - PMD_INIT_LOG(DEBUG, "Vector tx" - " can be enabled on this txq."); - - } else { - ad->tx_vec_allowed = false; - } - } else { - ad->tx_simple_allowed = false; - } + /* Use a simple Tx queue if possible (only fast free is allowed) */ + ad->tx_simple_allowed = + (txq->offloads == + (txq->offloads & RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE) && + txq->tx_rs_thresh >= RTE_PMD_I40E_TX_MAX_BURST); + ad->tx_vec_allowed = (ad->tx_simple_allowed && + txq->tx_rs_thresh <= RTE_I40E_TX_MAX_FREE_BUF_SZ); + + if (ad->tx_vec_allowed) + PMD_INIT_LOG(DEBUG, "Vector Tx can be enabled on Tx queue %u.", + txq->queue_id); + else if (ad->tx_simple_allowed) + PMD_INIT_LOG(DEBUG, "Simple Tx can be enabled on Tx queue %u.", + txq->queue_id); + else + PMD_INIT_LOG(DEBUG, + "Neither simple nor vector Tx enabled on Tx queue %u\n", + txq->queue_id); } -void __attribute__((cold)) +void __rte_cold i40e_set_tx_function(struct rte_eth_dev *dev) { struct i40e_adapter *ad = @@ -2918,6 +3439,10 @@ i40e_set_tx_function(struct rte_eth_dev *dev) int i; if (rte_eal_process_type() == RTE_PROC_PRIMARY) { +#ifdef RTE_ARCH_X86 + ad->tx_use_avx2 = false; + ad->tx_use_avx512 = false; +#endif if (ad->tx_vec_allowed) { for (i = 0; i < dev->data->nb_tx_queues; i++) { struct i40e_tx_queue *txq = @@ -2928,18 +3453,43 @@ i40e_set_tx_function(struct rte_eth_dev *dev) break; } } +#ifdef RTE_ARCH_X86 + ad->tx_use_avx512 = get_avx_supported(1); + + if (!ad->tx_use_avx512) + ad->tx_use_avx2 = get_avx_supported(0); +#endif } } if (ad->tx_simple_allowed) { - if (ad->tx_vec_allowed) { - PMD_INIT_LOG(DEBUG, "Vector tx finally be used."); + if (ad->tx_vec_allowed && + rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_128) { +#ifdef RTE_ARCH_X86 + if (ad->tx_use_avx512) { +#ifdef CC_AVX512_SUPPORT + PMD_DRV_LOG(NOTICE, "Using AVX512 Vector Tx (port %d).", + dev->data->port_id); + dev->tx_pkt_burst = i40e_xmit_pkts_vec_avx512; +#endif + } else { + PMD_INIT_LOG(DEBUG, "Using %sVector Tx (port %d).", + ad->tx_use_avx2 ? "avx2 " : "", + dev->data->port_id); + dev->tx_pkt_burst = ad->tx_use_avx2 ? + i40e_xmit_pkts_vec_avx2 : + i40e_xmit_pkts_vec; + } +#else /* RTE_ARCH_X86 */ + PMD_INIT_LOG(DEBUG, "Using Vector Tx (port %d).", + dev->data->port_id); dev->tx_pkt_burst = i40e_xmit_pkts_vec; +#endif /* RTE_ARCH_X86 */ } else { PMD_INIT_LOG(DEBUG, "Simple tx finally be used."); dev->tx_pkt_burst = i40e_xmit_pkts_simple; } - dev->tx_pkt_prepare = NULL; + dev->tx_pkt_prepare = i40e_simple_prep_pkts; } else { PMD_INIT_LOG(DEBUG, "Xmit tx finally be used."); dev->tx_pkt_burst = i40e_xmit_pkts; @@ -2947,7 +3497,46 @@ i40e_set_tx_function(struct rte_eth_dev *dev) } } -void __attribute__((cold)) +static const struct { + eth_tx_burst_t pkt_burst; + const char *info; +} i40e_tx_burst_infos[] = { + { i40e_xmit_pkts_simple, "Scalar Simple" }, + { i40e_xmit_pkts, "Scalar" }, +#ifdef RTE_ARCH_X86 +#ifdef CC_AVX512_SUPPORT + { i40e_xmit_pkts_vec_avx512, "Vector AVX512" }, +#endif + { i40e_xmit_pkts_vec_avx2, "Vector AVX2" }, + { i40e_xmit_pkts_vec, "Vector SSE" }, +#elif defined(RTE_ARCH_ARM64) + { i40e_xmit_pkts_vec, "Vector Neon" }, +#elif defined(RTE_ARCH_PPC_64) + { i40e_xmit_pkts_vec, "Vector AltiVec" }, +#endif +}; + +int +i40e_tx_burst_mode_get(struct rte_eth_dev *dev, __rte_unused uint16_t queue_id, + struct rte_eth_burst_mode *mode) +{ + eth_tx_burst_t pkt_burst = dev->tx_pkt_burst; + int ret = -EINVAL; + unsigned int i; + + for (i = 0; i < RTE_DIM(i40e_tx_burst_infos); ++i) { + if (pkt_burst == i40e_tx_burst_infos[i].pkt_burst) { + snprintf(mode->info, sizeof(mode->info), "%s", + i40e_tx_burst_infos[i].info); + ret = 0; + break; + } + } + + return ret; +} + +void __rte_cold i40e_set_default_ptype_table(struct rte_eth_dev *dev) { struct i40e_adapter *ad = @@ -2958,7 +3547,7 @@ i40e_set_default_ptype_table(struct rte_eth_dev *dev) ad->ptype_tbl[i] = i40e_get_default_pkt_type(i); } -void __attribute__((cold)) +void __rte_cold i40e_set_default_pctype_table(struct rte_eth_dev *dev) { struct i40e_adapter *ad = @@ -2994,7 +3583,8 @@ i40e_set_default_pctype_table(struct rte_eth_dev *dev) ad->pctypes_tbl[RTE_ETH_FLOW_L2_PAYLOAD] = (1ULL << I40E_FILTER_PCTYPE_L2_PAYLOAD); - if (hw->mac.type == I40E_MAC_X722) { + if (hw->mac.type == I40E_MAC_X722 || + hw->mac.type == I40E_MAC_X722_VF) { ad->pctypes_tbl[RTE_ETH_FLOW_NONFRAG_IPV4_UDP] |= (1ULL << I40E_FILTER_PCTYPE_NONF_UNICAST_IPV4_UDP); ad->pctypes_tbl[RTE_ETH_FLOW_NONFRAG_IPV4_UDP] |= @@ -3016,53 +3606,28 @@ i40e_set_default_pctype_table(struct rte_eth_dev *dev) } } -/* Stubs needed for linkage when CONFIG_RTE_I40E_INC_VECTOR is set to 'n' */ -int __attribute__((weak)) -i40e_rx_vec_dev_conf_condition_check(struct rte_eth_dev __rte_unused *dev) -{ - return -1; -} - -uint16_t __attribute__((weak)) -i40e_recv_pkts_vec( - void __rte_unused *rx_queue, - struct rte_mbuf __rte_unused **rx_pkts, - uint16_t __rte_unused nb_pkts) +#ifndef CC_AVX2_SUPPORT +uint16_t +i40e_recv_pkts_vec_avx2(void __rte_unused *rx_queue, + struct rte_mbuf __rte_unused **rx_pkts, + uint16_t __rte_unused nb_pkts) { return 0; } -uint16_t __attribute__((weak)) -i40e_recv_scattered_pkts_vec( - void __rte_unused *rx_queue, - struct rte_mbuf __rte_unused **rx_pkts, - uint16_t __rte_unused nb_pkts) +uint16_t +i40e_recv_scattered_pkts_vec_avx2(void __rte_unused *rx_queue, + struct rte_mbuf __rte_unused **rx_pkts, + uint16_t __rte_unused nb_pkts) { return 0; } -int __attribute__((weak)) -i40e_rxq_vec_setup(struct i40e_rx_queue __rte_unused *rxq) -{ - return -1; -} - -int __attribute__((weak)) -i40e_txq_vec_setup(struct i40e_tx_queue __rte_unused *txq) -{ - return -1; -} - -void __attribute__((weak)) -i40e_rx_queue_release_mbufs_vec(struct i40e_rx_queue __rte_unused*rxq) -{ - return; -} - -uint16_t __attribute__((weak)) -i40e_xmit_fixed_burst_vec(void __rte_unused * tx_queue, +uint16_t +i40e_xmit_pkts_vec_avx2(void __rte_unused * tx_queue, struct rte_mbuf __rte_unused **tx_pkts, uint16_t __rte_unused nb_pkts) { return 0; } +#endif /* ifndef CC_AVX2_SUPPORT */