From: Ajit Khaparde Date: Wed, 15 Jun 2016 21:23:14 +0000 (-0700) Subject: net/bnxt: add initial Rx code X-Git-Tag: spdx-start~6488 X-Git-Url: http://git.droids-corp.org/?a=commitdiff_plain;h=2eb53b134aae1301c1f90e345c00a94c2c13cabb;p=dpdk.git net/bnxt: add initial Rx code This patch adds initial implementation of rx_pkt_burst() function for Rx. bnxt_recv_pkts() is the top level function for doing Rx. This patch also adds code to allocate rings in the ASIC. For each Rx queue allocated in the PMD driver, a corresponding ring in hardware will be created. Every time a frame is received a Rx ring is selected based on the hardware configuration like RSS, MAC or VLAN, COS and such. The hardware uses a completion ring to indicate the availability of a packet. This patch also brings in functions like bnxt_init_one_rx_ring() bnxt_init_rx_ring_struct() which initializes various structures before a Rx can begin. bnxt_init_rxbds() initializes the Rx Buffer Descriptors while bnxt_alloc_rx_data() allocates a buffer in the host to receive the incoming packet. Signed-off-by: Ajit Khaparde Signed-off-by: Stephen Hurd Reviewed-by: David Christensen --- diff --git a/drivers/net/bnxt/Makefile b/drivers/net/bnxt/Makefile index 842b2fdc24..d9c5a4cb5a 100644 --- a/drivers/net/bnxt/Makefile +++ b/drivers/net/bnxt/Makefile @@ -54,6 +54,7 @@ SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_filter.c SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_hwrm.c SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_ring.c SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_rxq.c +SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_rxr.c SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_stats.c SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_txq.c SRCS-$(CONFIG_RTE_LIBRTE_BNXT_PMD) += bnxt_txr.c diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 4ace5432a0..68883637fe 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -42,6 +42,7 @@ #include "bnxt.h" #include "bnxt_hwrm.h" #include "bnxt_rxq.h" +#include "bnxt_rxr.h" #include "bnxt_stats.h" #include "bnxt_txq.h" #include "bnxt_txr.h" @@ -269,7 +270,7 @@ bnxt_dev_init(struct rte_eth_dev *eth_dev) goto error; } eth_dev->dev_ops = &bnxt_dev_ops; - /* eth_dev->rx_pkt_burst = &bnxt_recv_pkts; */ + eth_dev->rx_pkt_burst = &bnxt_recv_pkts; eth_dev->tx_pkt_burst = &bnxt_xmit_pkts; rc = bnxt_alloc_hwrm_resources(bp); diff --git a/drivers/net/bnxt/bnxt_ring.c b/drivers/net/bnxt/bnxt_ring.c index 2645ddac4d..f547a9e407 100644 --- a/drivers/net/bnxt/bnxt_ring.c +++ b/drivers/net/bnxt/bnxt_ring.c @@ -36,6 +36,7 @@ #include "bnxt.h" #include "bnxt_cpr.h" #include "bnxt_ring.h" +#include "bnxt_rxr.h" #include "bnxt_txr.h" #include "hsi_struct_def_dpdk.h" @@ -74,8 +75,7 @@ int bnxt_alloc_rings(struct bnxt *bp, uint16_t qidx, { struct bnxt_ring *cp_ring = cp_ring_info->cp_ring_struct; struct bnxt_ring *tx_ring; - /* TODO: RX ring */ - /* struct bnxt_ring *rx_ring; */ + struct bnxt_ring *rx_ring; struct rte_pci_device *pdev = bp->pdev; const struct rte_memzone *mz = NULL; char mz_name[RTE_MEMZONE_NAMESIZE]; @@ -92,12 +92,9 @@ int bnxt_alloc_rings(struct bnxt *bp, uint16_t qidx, tx_ring_struct->vmem_size) : 0; int rx_vmem_start = tx_vmem_start + tx_vmem_len; - /* TODO: RX ring */ - int rx_vmem_len = 0; - /* - * rx_ring_info ? RTE_CACHE_LINE_ROUNDUP(rx_ring_info-> - * rx_ring_struct->vmem_size) : 0; - */ + int rx_vmem_len = rx_ring_info ? + RTE_CACHE_LINE_ROUNDUP(rx_ring_info-> + rx_ring_struct->vmem_size) : 0; int cp_ring_start = rx_vmem_start + rx_vmem_len; int cp_ring_len = RTE_CACHE_LINE_ROUNDUP(cp_ring->ring_size * @@ -109,13 +106,9 @@ int bnxt_alloc_rings(struct bnxt *bp, uint16_t qidx, sizeof(struct tx_bd_long)) : 0; int rx_ring_start = tx_ring_start + tx_ring_len; - /* TODO: RX ring */ - int rx_ring_len = 0; - /* - * rx_ring_info ? - * RTE_CACHE_LINE_ROUNDUP(rx_ring_info->rx_ring_struct->ring_size * - * sizeof(struct rx_prod_pkt_bd)) : 0; - */ + int rx_ring_len = rx_ring_info ? + RTE_CACHE_LINE_ROUNDUP(rx_ring_info->rx_ring_struct->ring_size * + sizeof(struct rx_prod_pkt_bd)) : 0; int total_alloc_len = rx_ring_start + rx_ring_len; @@ -153,26 +146,24 @@ int bnxt_alloc_rings(struct bnxt *bp, uint16_t qidx, } } -/* - * if (rx_ring_info) { - * rx_ring = &rx_ring_info->rx_ring_struct; - * - * rx_ring->bd = ((char *)mz->addr + rx_ring_start); - * rx_ring_info->rx_desc_ring = - * (struct rx_prod_pkt_bd *)rx_ring->bd; - * rx_ring->bd_dma = mz->phys_addr + rx_ring_start; - * rx_ring_info->rx_desc_mapping = rx_ring->bd_dma; - * - * if (!rx_ring->bd) - * return -ENOMEM; - * if (rx_ring->vmem_size) { - * rx_ring->vmem = - * (void **)((char *)mz->addr + rx_vmem_start); - * rx_ring_info->rx_buf_ring = - * (struct bnxt_sw_rx_bd *)rx_ring->vmem; - * } - * } - */ + if (rx_ring_info) { + rx_ring = rx_ring_info->rx_ring_struct; + + rx_ring->bd = ((char *)mz->addr + rx_ring_start); + rx_ring_info->rx_desc_ring = + (struct rx_prod_pkt_bd *)rx_ring->bd; + rx_ring->bd_dma = mz->phys_addr + rx_ring_start; + rx_ring_info->rx_desc_mapping = rx_ring->bd_dma; + + if (!rx_ring->bd) + return -ENOMEM; + if (rx_ring->vmem_size) { + rx_ring->vmem = + (void **)((char *)mz->addr + rx_vmem_start); + rx_ring_info->rx_buf_ring = + (struct bnxt_sw_rx_bd *)rx_ring->vmem; + } + } cp_ring->bd = ((char *)mz->addr + cp_ring_start); cp_ring->bd_dma = mz->phys_addr + cp_ring_start; diff --git a/drivers/net/bnxt/bnxt_rxq.c b/drivers/net/bnxt/bnxt_rxq.c index 19c0e3f647..eeda631ff8 100644 --- a/drivers/net/bnxt/bnxt_rxq.c +++ b/drivers/net/bnxt/bnxt_rxq.c @@ -41,6 +41,7 @@ #include "bnxt_hwrm.h" #include "bnxt_ring.h" #include "bnxt_rxq.h" +#include "bnxt_rxr.h" #include "bnxt_vnic.h" #include "hsi_struct_def_dpdk.h" @@ -216,7 +217,20 @@ err_out: static void bnxt_rx_queue_release_mbufs(struct bnxt_rx_queue *rxq __rte_unused) { - /* TODO: Requires interaction with TX ring */ + struct bnxt_sw_rx_bd *sw_ring; + uint16_t i; + + if (rxq) { + sw_ring = rxq->rx_ring->rx_buf_ring; + if (sw_ring) { + for (i = 0; i < rxq->nb_rx_desc; i++) { + if (sw_ring[i].mbuf) { + rte_pktmbuf_free_seg(sw_ring[i].mbuf); + sw_ring[i].mbuf = NULL; + } + } + } + } } void bnxt_free_rx_mbufs(struct bnxt *bp) @@ -237,7 +251,13 @@ void bnxt_rx_queue_release_op(void *rx_queue) if (rxq) { bnxt_rx_queue_release_mbufs(rxq); - /* TODO: Free ring and stats here */ + /* Free RX ring hardware descriptors */ + bnxt_free_ring(rxq->rx_ring->rx_ring_struct); + + /* Free RX completion ring hardware descriptors */ + bnxt_free_ring(rxq->cp_ring->cp_ring_struct); + + bnxt_free_rxq_stats(rxq); rte_free(rxq); } @@ -274,7 +294,7 @@ int bnxt_rx_queue_setup_op(struct rte_eth_dev *eth_dev, rxq->nb_rx_desc = nb_desc; rxq->rx_free_thresh = rx_conf->rx_free_thresh; - /* TODO: Initialize ring structure */ + bnxt_init_rx_ring_struct(rxq); rxq->queue_id = queue_idx; rxq->port_id = eth_dev->data->port_id; @@ -282,7 +302,13 @@ int bnxt_rx_queue_setup_op(struct rte_eth_dev *eth_dev, 0 : ETHER_CRC_LEN); eth_dev->data->rx_queues[queue_idx] = rxq; - /* TODO: Allocate RX ring hardware descriptors */ + /* Allocate RX ring hardware descriptors */ + if (bnxt_alloc_rings(bp, queue_idx, NULL, rxq->rx_ring, rxq->cp_ring, + "rxr")) { + RTE_LOG(ERR, PMD, "ring_dma_zone_reserve for rx_ring failed!"); + bnxt_rx_queue_release_op(rxq); + return -ENOMEM; + } return 0; } diff --git a/drivers/net/bnxt/bnxt_rxr.c b/drivers/net/bnxt/bnxt_rxr.c new file mode 100644 index 0000000000..b02395ba8c --- /dev/null +++ b/drivers/net/bnxt/bnxt_rxr.c @@ -0,0 +1,341 @@ +/*- + * BSD LICENSE + * + * Copyright(c) Broadcom Limited. + * 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 Broadcom 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. + */ + +#include +#include + +#include +#include +#include + +#include "bnxt.h" +#include "bnxt_cpr.h" +#include "bnxt_ring.h" +#include "bnxt_rxr.h" +#include "bnxt_rxq.h" +#include "hsi_struct_def_dpdk.h" + +/* + * RX Ring handling + */ + +static inline struct rte_mbuf *__bnxt_alloc_rx_data(struct rte_mempool *mb) +{ + struct rte_mbuf *data; + + data = rte_mbuf_raw_alloc(mb); + + return data; +} + +static inline int bnxt_alloc_rx_data(struct bnxt_rx_queue *rxq, + struct bnxt_rx_ring_info *rxr, + uint16_t prod) +{ + struct rx_prod_pkt_bd *rxbd = &rxr->rx_desc_ring[prod]; + struct bnxt_sw_rx_bd *rx_buf = &rxr->rx_buf_ring[prod]; + struct rte_mbuf *data; + + data = __bnxt_alloc_rx_data(rxq->mb_pool); + if (!data) + return -ENOMEM; + + rx_buf->mbuf = data; + + rxbd->addr = rte_cpu_to_le_64(RTE_MBUF_DATA_DMA_ADDR(rx_buf->mbuf)); + + return 0; +} + +static void bnxt_reuse_rx_mbuf(struct bnxt_rx_ring_info *rxr, uint16_t cons, + struct rte_mbuf *mbuf) +{ + uint16_t prod = rxr->rx_prod; + struct bnxt_sw_rx_bd *prod_rx_buf; + struct rx_prod_pkt_bd *prod_bd, *cons_bd; + + prod_rx_buf = &rxr->rx_buf_ring[prod]; + + prod_rx_buf->mbuf = mbuf; + + prod_bd = &rxr->rx_desc_ring[prod]; + cons_bd = &rxr->rx_desc_ring[cons]; + + prod_bd->addr = cons_bd->addr; +} + +static uint16_t bnxt_rx_pkt(struct rte_mbuf **rx_pkt, + struct bnxt_rx_queue *rxq, uint32_t *raw_cons) +{ + struct bnxt_cp_ring_info *cpr = rxq->cp_ring; + struct bnxt_rx_ring_info *rxr = rxq->rx_ring; + struct rx_pkt_cmpl *rxcmp; + struct rx_pkt_cmpl_hi *rxcmp1; + uint32_t tmp_raw_cons = *raw_cons; + uint16_t cons, prod, cp_cons = + RING_CMP(cpr->cp_ring_struct, tmp_raw_cons); + struct bnxt_sw_rx_bd *rx_buf; + struct rte_mbuf *mbuf; + int rc = 0; + + rxcmp = (struct rx_pkt_cmpl *) + &cpr->cp_desc_ring[cp_cons]; + + tmp_raw_cons = NEXT_RAW_CMP(tmp_raw_cons); + cp_cons = RING_CMP(cpr->cp_ring_struct, tmp_raw_cons); + rxcmp1 = (struct rx_pkt_cmpl_hi *)&cpr->cp_desc_ring[cp_cons]; + + if (!CMP_VALID(rxcmp1, tmp_raw_cons, cpr->cp_ring_struct)) + return -EBUSY; + + prod = rxr->rx_prod; + + /* EW - GRO deferred to phase 3 */ + cons = rxcmp->opaque; + rx_buf = &rxr->rx_buf_ring[cons]; + mbuf = rx_buf->mbuf; + rte_prefetch0(mbuf); + + mbuf->nb_segs = 1; + mbuf->next = NULL; + mbuf->pkt_len = rxcmp->len; + mbuf->data_len = mbuf->pkt_len; + mbuf->port = rxq->port_id; + mbuf->ol_flags = 0; + if (rxcmp->flags_type & RX_PKT_CMPL_FLAGS_RSS_VALID) { + mbuf->hash.rss = rxcmp->rss_hash; + mbuf->ol_flags |= PKT_RX_RSS_HASH; + } else { + mbuf->hash.fdir.id = rxcmp1->cfa_code; + mbuf->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID; + } + if (rxcmp1->flags2 & RX_PKT_CMPL_FLAGS2_META_FORMAT_VLAN) { + mbuf->vlan_tci = rxcmp1->metadata & + (RX_PKT_CMPL_METADATA_VID_MASK | + RX_PKT_CMPL_METADATA_DE | + RX_PKT_CMPL_METADATA_PRI_MASK); + mbuf->ol_flags |= PKT_RX_VLAN_PKT; + } + + rx_buf->mbuf = NULL; + if (rxcmp1->errors_v2 & RX_CMP_L2_ERRORS) { + /* Re-install the mbuf back to the rx ring */ + bnxt_reuse_rx_mbuf(rxr, cons, mbuf); + + rc = -EIO; + goto next_rx; + } + /* + * TODO: Redesign this.... + * If the allocation fails, the packet does not get received. + * Simply returning this will result in slowly falling behind + * on the producer ring buffers. + * Instead, "filling up" the producer just before ringing the + * doorbell could be a better solution since it will let the + * producer ring starve until memory is available again pushing + * the drops into hardware and getting them out of the driver + * allowing recovery to a full producer ring. + * + * This could also help with cache usage by preventing per-packet + * calls in favour of a tight loop with the same function being called + * in it. + */ + if (bnxt_alloc_rx_data(rxq, rxr, prod)) { + RTE_LOG(ERR, PMD, "mbuf alloc failed with prod=0x%x\n", prod); + rc = -ENOMEM; + goto next_rx; + } + + /* + * All MBUFs are allocated with the same size under DPDK, + * no optimization for rx_copy_thresh + */ + + /* AGG buf operation is deferred */ + + /* EW - VLAN reception. Must compare against the ol_flags */ + + *rx_pkt = mbuf; +next_rx: + rxr->rx_prod = RING_NEXT(rxr->rx_ring_struct, prod); + + *raw_cons = tmp_raw_cons; + + return rc; +} + +uint16_t bnxt_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts) +{ + struct bnxt_rx_queue *rxq = rx_queue; + struct bnxt_cp_ring_info *cpr = rxq->cp_ring; + struct bnxt_rx_ring_info *rxr = rxq->rx_ring; + uint32_t raw_cons = cpr->cp_raw_cons; + uint32_t cons; + int nb_rx_pkts = 0; + bool rx_event = false; + struct rx_pkt_cmpl *rxcmp; + + /* Handle RX burst request */ + while (1) { + int rc; + + cons = RING_CMP(cpr->cp_ring_struct, raw_cons); + rte_prefetch0(&cpr->cp_desc_ring[cons]); + rxcmp = (struct rx_pkt_cmpl *)&cpr->cp_desc_ring[cons]; + + if (!CMP_VALID(rxcmp, raw_cons, cpr->cp_ring_struct)) + break; + + /* TODO: Avoid magic numbers... */ + if ((CMP_TYPE(rxcmp) & 0x30) == 0x10) { + rc = bnxt_rx_pkt(&rx_pkts[nb_rx_pkts], rxq, &raw_cons); + if (likely(!rc)) + nb_rx_pkts++; + else if (rc == -EBUSY) /* partial completion */ + break; + rx_event = true; + } + raw_cons = NEXT_RAW_CMP(raw_cons); + if (nb_rx_pkts == nb_pkts) + break; + } + if (raw_cons == cpr->cp_raw_cons) { + /* + * For PMD, there is no need to keep on pushing to REARM + * the doorbell if there are no new completions + */ + return nb_rx_pkts; + } + cpr->cp_raw_cons = raw_cons; + + B_CP_DIS_DB(cpr, cpr->cp_raw_cons); + if (rx_event) + B_RX_DB(rxr->rx_doorbell, rxr->rx_prod); + return nb_rx_pkts; +} + +void bnxt_free_rx_rings(struct bnxt *bp) +{ + int i; + + for (i = 0; i < (int)bp->rx_nr_rings; i++) { + struct bnxt_rx_queue *rxq = bp->rx_queues[i]; + + if (!rxq) + continue; + + /* TODO: free() rxq->rx_ring and rxq->rx_ring->rx_ring_struct */ + bnxt_free_ring(rxq->rx_ring->rx_ring_struct); + /* TODO: free() rxq->cp_ring and rxq->cp_ring->cp_ring_struct */ + bnxt_free_ring(rxq->cp_ring->cp_ring_struct); + + rte_free(rxq); + bp->rx_queues[i] = NULL; + } +} + +void bnxt_init_rx_ring_struct(struct bnxt_rx_queue *rxq) +{ + struct bnxt *bp = rxq->bp; + struct bnxt_cp_ring_info *cpr; + struct bnxt_rx_ring_info *rxr; + struct bnxt_ring *ring; + + rxq->rx_buf_use_size = bp->eth_dev->data->mtu + + ETHER_HDR_LEN + ETHER_CRC_LEN + + (2 * VLAN_TAG_SIZE); + rxq->rx_buf_size = rxq->rx_buf_use_size + sizeof(struct rte_mbuf); + + rxr = rxq->rx_ring; + ring = rxr->rx_ring_struct; + ring->ring_size = rte_align32pow2(rxq->nb_rx_desc); + ring->ring_mask = ring->ring_size - 1; + ring->bd = (void *)rxr->rx_desc_ring; + ring->bd_dma = rxr->rx_desc_mapping; + ring->vmem_size = ring->ring_size * sizeof(struct bnxt_sw_rx_bd); + ring->vmem = (void **)&rxr->rx_buf_ring; + + cpr = rxq->cp_ring; + ring = cpr->cp_ring_struct; + ring->ring_size = rxr->rx_ring_struct->ring_size * 2; + ring->ring_mask = ring->ring_size - 1; + ring->bd = (void *)cpr->cp_desc_ring; + ring->bd_dma = cpr->cp_desc_mapping; + ring->vmem_size = 0; + ring->vmem = NULL; +} + +static void bnxt_init_rxbds(struct bnxt_ring *ring, uint32_t type, + uint16_t len) +{ + uint32_t j; + struct rx_prod_pkt_bd *rx_bd_ring = (struct rx_prod_pkt_bd *)ring->bd; + + if (!rx_bd_ring) + return; + for (j = 0; j < ring->ring_size; j++) { + rx_bd_ring[j].flags_type = type; + rx_bd_ring[j].len = len; + rx_bd_ring[j].opaque = j; + } +} + +int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq) +{ + struct bnxt_rx_ring_info *rxr; + struct bnxt_ring *ring; + uint32_t prod, type; + unsigned int i; + + type = RX_PROD_PKT_BD_TYPE_RX_PROD_PKT | RX_PROD_PKT_BD_FLAGS_EOP_PAD; + + /* TODO: These need to be allocated */ + rxr = rxq->rx_ring; + ring = rxr->rx_ring_struct; + bnxt_init_rxbds(ring, type, rxq->rx_buf_use_size); + + prod = rxr->rx_prod; + for (i = 0; i < ring->ring_size; i++) { + if (bnxt_alloc_rx_data(rxq, rxr, prod) != 0) { + RTE_LOG(WARNING, PMD, + "init'ed rx ring %d with %d/%d mbufs only\n", + rxq->queue_id, i, ring->ring_size); + break; + } + rxr->rx_prod = prod; + prod = RING_NEXT(rxr->rx_ring_struct, prod); + } + + return 0; +} diff --git a/drivers/net/bnxt/bnxt_rxr.h b/drivers/net/bnxt/bnxt_rxr.h new file mode 100644 index 0000000000..95c61a085b --- /dev/null +++ b/drivers/net/bnxt/bnxt_rxr.h @@ -0,0 +1,62 @@ +/*- + * BSD LICENSE + * + * Copyright(c) Broadcom Limited. + * 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 Broadcom 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. + */ + +#ifndef _BNXT_RXR_H_ +#define _BNXT_RXR_H_ + +#define B_RX_DB(db, prod) \ + (*(uint32_t *)db = (DB_KEY_RX | prod)) + +struct bnxt_sw_rx_bd { + struct rte_mbuf *mbuf; /* data associated with RX descriptor */ +}; + +struct bnxt_rx_ring_info { + uint16_t rx_prod; + void *rx_doorbell; + + struct rx_prod_pkt_bd *rx_desc_ring; + struct bnxt_sw_rx_bd *rx_buf_ring; /* sw ring */ + + phys_addr_t rx_desc_mapping; + + struct bnxt_ring *rx_ring_struct; +}; + +uint16_t bnxt_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts, + uint16_t nb_pkts); +void bnxt_free_rx_rings(struct bnxt *bp); +void bnxt_init_rx_ring_struct(struct bnxt_rx_queue *rxq); +int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq); + +#endif diff --git a/drivers/net/bnxt/hsi_struct_def_dpdk.h b/drivers/net/bnxt/hsi_struct_def_dpdk.h index 1543f20240..8b30787470 100644 --- a/drivers/net/bnxt/hsi_struct_def_dpdk.h +++ b/drivers/net/bnxt/hsi_struct_def_dpdk.h @@ -446,6 +446,79 @@ struct tx_bd_long_hi { uint32_t cfa_meta; } __attribute__((packed)); +/* RX Producer Packet BD (16 bytes) */ +struct rx_prod_pkt_bd { + /* This value identifies the type of buffer descriptor. */ + #define RX_PROD_PKT_BD_TYPE_MASK UINT32_C(0x3f) + #define RX_PROD_PKT_BD_TYPE_SFT 0 + /* + * Indicates that this BD is 16B long and is an RX Producer (ie. + * empty) buffer descriptor. + */ + #define RX_PROD_PKT_BD_TYPE_RX_PROD_PKT (UINT32_C(0x4) << 0) + /* + * If set to 1, the packet will be placed at the address plus 2B. The 2 + * Bytes of padding will be written as zero. + */ + /* + * This is intended to be used when the host buffer is cache-line + * aligned to produce packets that are easy to parse in host memory + * while still allowing writes to be cache line aligned. + */ + #define RX_PROD_PKT_BD_FLAGS_SOP_PAD UINT32_C(0x40) + /* + * If set to 1, the packet write will be padded out to the nearest + * cache-line with zero value padding. + */ + /* + * If receive buffers start/end on cache-line boundaries, this feature + * will ensure that all data writes on the PCI bus start/end on cache + * line boundaries. + */ + #define RX_PROD_PKT_BD_FLAGS_EOP_PAD UINT32_C(0x80) + /* + * This value is the number of additional buffers in the ring that + * describe the buffer space to be consumed for the this packet. If the + * value is zero, then the packet must fit within the space described by + * this BD. If this value is 1 or more, it indicates how many additional + * "buffer" BDs are in the ring immediately following this BD to be used + * for the same network packet. Even if the packet to be placed does not + * need all the additional buffers, they will be consumed anyway. + */ + #define RX_PROD_PKT_BD_FLAGS_BUFFERS_MASK UINT32_C(0x300) + #define RX_PROD_PKT_BD_FLAGS_BUFFERS_SFT 8 + #define RX_PROD_PKT_BD_FLAGS_MASK UINT32_C(0xffc0) + #define RX_PROD_PKT_BD_FLAGS_SFT 6 + uint16_t flags_type; + + /* + * This is the length in Bytes of the host physical buffer where data + * for the packet may be placed in host memory. + */ + /* + * While this is a Byte resolution value, it is often advantageous to + * ensure that the buffers provided end on a host cache line. + */ + uint16_t len; + + /* + * The opaque data field is pass through to the completion and can be + * used for any data that the driver wants to associate with this + * receive buffer set. + */ + uint32_t opaque; + + /* + * This is the host physical address where data for the packet may by + * placed in host memory. + */ + /* + * While this is a Byte resolution value, it is often advantageous to + * ensure that the buffers provide start on a host cache line. + */ + uint64_t addr; +} __attribute__((packed)); + /* Completion Ring Structures */ /* Note: This structure is used by the HWRM to communicate HWRM Error. */ /* Base Completion Record (16 bytes) */ @@ -611,6 +684,407 @@ struct tx_cmpl { uint32_t unused_2; } __attribute__((packed)) tx_cmpl_t, *ptx_cmpl_t; +/* RX Packet Completion Record (32 bytes split to 2 16-byte struct) */ +struct rx_pkt_cmpl { + /* + * This field indicates the exact type of the completion. By convention, + * the LSB identifies the length of the record in 16B units. Even values + * indicate 16B records. Odd values indicate 32B records. + */ + #define RX_PKT_CMPL_TYPE_MASK UINT32_C(0x3f) + #define RX_PKT_CMPL_TYPE_SFT 0 + /* + * RX L2 completion: Completion of and L2 RX packet. + * Length = 32B + */ + #define RX_PKT_CMPL_TYPE_RX_L2 (UINT32_C(0x11) << 0) + /* + * When this bit is '1', it indicates a packet that has an error of some + * type. Type of error is indicated in error_flags. + */ + #define RX_PKT_CMPL_FLAGS_ERROR UINT32_C(0x40) + /* This field indicates how the packet was placed in the buffer. */ + #define RX_PKT_CMPL_FLAGS_PLACEMENT_MASK UINT32_C(0x380) + #define RX_PKT_CMPL_FLAGS_PLACEMENT_SFT 7 + /* Normal: Packet was placed using normal algorithm. */ + #define RX_PKT_CMPL_FLAGS_PLACEMENT_NORMAL (UINT32_C(0x0) << 7) + /* Jumbo: Packet was placed using jumbo algorithm. */ + #define RX_PKT_CMPL_FLAGS_PLACEMENT_JUMBO (UINT32_C(0x1) << 7) + /* + * Header/Data Separation: Packet was placed using Header/Data + * separation algorithm. The separation location is indicated by + * the itype field. + */ + #define RX_PKT_CMPL_FLAGS_PLACEMENT_HDS (UINT32_C(0x2) << 7) + #define RX_PKT_CMPL_FLAGS_PLACEMENT_LAST \ + RX_PKT_CMPL_FLAGS_PLACEMENT_HDS + /* This bit is '1' if the RSS field in this completion is valid. */ + #define RX_PKT_CMPL_FLAGS_RSS_VALID UINT32_C(0x400) + /* + * This value indicates what the inner packet determined for the packet + * was. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_MASK UINT32_C(0xf000) + #define RX_PKT_CMPL_FLAGS_ITYPE_SFT 12 + /* Not Known: Indicates that the packet type was not known. */ + #define RX_PKT_CMPL_FLAGS_ITYPE_NOT_KNOWN (UINT32_C(0x0) << 12) + /* + * IP Packet: Indicates that the packet was an IP packet, but + * further classification was not possible. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_IP (UINT32_C(0x1) << 12) + /* + * TCP Packet: Indicates that the packet was IP and TCP. This + * indicates that the payload_offset field is valid. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_TCP (UINT32_C(0x2) << 12) + /* + * UDP Packet: Indicates that the packet was IP and UDP. This + * indicates that the payload_offset field is valid. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_UDP (UINT32_C(0x3) << 12) + /* + * FCoE Packet: Indicates that the packet was recognized as a + * FCoE. This also indicates that the payload_offset field is + * valid. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_FCOE (UINT32_C(0x4) << 12) + /* + * RoCE Packet: Indicates that the packet was recognized as a + * RoCE. This also indicates that the payload_offset field is + * valid. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_ROCE (UINT32_C(0x5) << 12) + /* + * ICMP Packet: Indicates that the packet was recognized as + * ICMP. This indicates that the payload_offset field is valid. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_ICMP (UINT32_C(0x7) << 12) + /* + * PtP packet wo/timestamp: Indicates that the packet was + * recognized as a PtP packet. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_PTP_WO_TIMESTAMP \ + (UINT32_C(0x8) << 12) + /* + * PtP packet w/timestamp: Indicates that the packet was + * recognized as a PtP packet and that a timestamp was taken for + * the packet. + */ + #define RX_PKT_CMPL_FLAGS_ITYPE_PTP_W_TIMESTAMP (UINT32_C(0x9) << 12) + #define RX_PKT_CMPL_FLAGS_ITYPE_LAST \ + RX_PKT_CMPL_FLAGS_ITYPE_PTP_W_TIMESTAMP + #define RX_PKT_CMPL_FLAGS_MASK UINT32_C(0xffc0) + #define RX_PKT_CMPL_FLAGS_SFT 6 + uint16_t flags_type; + + /* + * This is the length of the data for the packet stored in the buffer(s) + * identified by the opaque value. This includes the packet BD and any + * associated buffer BDs. This does not include the the length of any + * data places in aggregation BDs. + */ + uint16_t len; + + /* + * This is a copy of the opaque field from the RX BD this completion + * corresponds to. + */ + uint32_t opaque; + + /* + * This value is written by the NIC such that it will be different for + * each pass through the completion queue. The even passes will write 1. + * The odd passes will write 0. + */ + #define RX_PKT_CMPL_V1 UINT32_C(0x1) + /* + * This value is the number of aggregation buffers that follow this + * entry in the completion ring that are a part of this packet. If the + * value is zero, then the packet is completely contained in the buffer + * space provided for the packet in the RX ring. + */ + #define RX_PKT_CMPL_AGG_BUFS_MASK UINT32_C(0x3e) + #define RX_PKT_CMPL_AGG_BUFS_SFT 1 + uint8_t agg_bufs_v1; + + /* + * This is the RSS hash type for the packet. The value is packed + * {tuple_extrac_op[1:0],rss_profile_id[4:0],tuple_extrac_op[2]}. + */ + uint8_t rss_hash_type; + + /* + * This value indicates the offset from the beginning of the packet + * where the inner payload starts. This value is valid for TCP, UDP, + * FCoE, and RoCE packets. + */ + uint8_t payload_offset; + + uint8_t unused_1; + + /* + * This value is the RSS hash value calculated for the packet based on + * the mode bits and key value in the VNIC. + */ + uint32_t rss_hash; +} __attribute__((packed)); + +/* last 16 bytes of RX Packet Completion Record */ +struct rx_pkt_cmpl_hi { + /* + * This indicates that the ip checksum was calculated for the inner + * packet and that the ip_cs_error field indicates if there was an + * error. + */ + #define RX_PKT_CMPL_FLAGS2_IP_CS_CALC UINT32_C(0x1) + /* + * This indicates that the TCP, UDP or ICMP checksum was calculated for + * the inner packet and that the l4_cs_error field indicates if there + * was an error. + */ + #define RX_PKT_CMPL_FLAGS2_L4_CS_CALC UINT32_C(0x2) + /* + * This indicates that the ip checksum was calculated for the tunnel + * header and that the t_ip_cs_error field indicates if there was an + * error. + */ + #define RX_PKT_CMPL_FLAGS2_T_IP_CS_CALC UINT32_C(0x4) + /* + * This indicates that the UDP checksum was calculated for the tunnel + * packet and that the t_l4_cs_error field indicates if there was an + * error. + */ + #define RX_PKT_CMPL_FLAGS2_T_L4_CS_CALC UINT32_C(0x8) + /* This value indicates what format the metadata field is. */ + #define RX_PKT_CMPL_FLAGS2_META_FORMAT_MASK UINT32_C(0xf0) + #define RX_PKT_CMPL_FLAGS2_META_FORMAT_SFT 4 + /* No metadata informtaion. Value is zero. */ + #define RX_PKT_CMPL_FLAGS2_META_FORMAT_NONE (UINT32_C(0x0) << 4) + /* + * The metadata field contains the VLAN tag and TPID value. - + * metadata[11:0] contains the vlan VID value. - metadata[12] + * contains the vlan DE value. - metadata[15:13] contains the + * vlan PRI value. - metadata[31:16] contains the vlan TPID + * value. + */ + #define RX_PKT_CMPL_FLAGS2_META_FORMAT_VLAN (UINT32_C(0x1) << 4) + #define RX_PKT_CMPL_FLAGS2_META_FORMAT_LAST \ + RX_PKT_CMPL_FLAGS2_META_FORMAT_VLAN + /* + * This field indicates the IP type for the inner-most IP header. A + * value of '0' indicates IPv4. A value of '1' indicates IPv6. This + * value is only valid if itype indicates a packet with an IP header. + */ + #define RX_PKT_CMPL_FLAGS2_IP_TYPE UINT32_C(0x100) + uint32_t flags2; + + /* + * This is data from the CFA block as indicated by the meta_format + * field. + */ + /* When meta_format=1, this value is the VLAN VID. */ + #define RX_PKT_CMPL_METADATA_VID_MASK UINT32_C(0xfff) + #define RX_PKT_CMPL_METADATA_VID_SFT 0 + /* When meta_format=1, this value is the VLAN DE. */ + #define RX_PKT_CMPL_METADATA_DE UINT32_C(0x1000) + /* When meta_format=1, this value is the VLAN PRI. */ + #define RX_PKT_CMPL_METADATA_PRI_MASK UINT32_C(0xe000) + #define RX_PKT_CMPL_METADATA_PRI_SFT 13 + /* When meta_format=1, this value is the VLAN TPID. */ + #define RX_PKT_CMPL_METADATA_TPID_MASK UINT32_C(0xffff0000) + #define RX_PKT_CMPL_METADATA_TPID_SFT 16 + uint32_t metadata; + + /* + * This value is written by the NIC such that it will be different for + * each pass through the completion queue. The even passes will write 1. + * The odd passes will write 0. + */ + #define RX_PKT_CMPL_V2 UINT32_C(0x1) + /* + * This error indicates that there was some sort of problem with the BDs + * for the packet that was found after part of the packet was already + * placed. The packet should be treated as invalid. + */ + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_MASK UINT32_C(0xe) + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_SFT 1 + /* No buffer error */ + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_NO_BUFFER \ + (UINT32_C(0x0) << 1) + /* + * Did Not Fit: Packet did not fit into packet buffer provided. + * For regular placement, this means the packet did not fit in + * the buffer provided. For HDS and jumbo placement, this means + * that the packet could not be placed into 7 physical buffers + * or less. + */ + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_DID_NOT_FIT \ + (UINT32_C(0x1) << 1) + /* + * Not On Chip: All BDs needed for the packet were not on-chip + * when the packet arrived. + */ + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_NOT_ON_CHIP \ + (UINT32_C(0x2) << 1) + /* Bad Format: BDs were not formatted correctly. */ + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_BAD_FORMAT \ + (UINT32_C(0x3) << 1) + #define RX_PKT_CMPL_ERRORS_BUFFER_ERROR_LAST \ + RX_PKT_CMPL_ERRORS_BUFFER_ERROR_BAD_FORMAT + /* This indicates that there was an error in the IP header checksum. */ + #define RX_PKT_CMPL_ERRORS_IP_CS_ERROR UINT32_C(0x10) + /* + * This indicates that there was an error in the TCP, UDP or ICMP + * checksum. + */ + #define RX_PKT_CMPL_ERRORS_L4_CS_ERROR UINT32_C(0x20) + /* + * This indicates that there was an error in the tunnel IP header + * checksum. + */ + #define RX_PKT_CMPL_ERRORS_T_IP_CS_ERROR UINT32_C(0x40) + /* This indicates that there was an error in the tunnel UDP checksum. */ + #define RX_PKT_CMPL_ERRORS_T_L4_CS_ERROR UINT32_C(0x80) + /* + * This indicates that there was a CRC error on either an FCoE or RoCE + * packet. The itype indicates the packet type. + */ + #define RX_PKT_CMPL_ERRORS_CRC_ERROR UINT32_C(0x100) + /* + * This indicates that there was an error in the tunnel portion of the + * packet when this field is non-zero. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_MASK UINT32_C(0xe00) + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_SFT 9 + /* + * No additional error occurred on the tunnel portion of the + * packet of the packet does not have a tunnel. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_NO_ERROR (UINT32_C(0x0) << 9) + /* + * Indicates that IP header version does not match expectation + * from L2 Ethertype for IPv4 and IPv6 in the tunnel header. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_T_L3_BAD_VERSION \ + (UINT32_C(0x1) << 9) + /* + * Indicates that header length is out of range in the tunnel + * header. Valid for IPv4. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_T_L3_BAD_HDR_LEN \ + (UINT32_C(0x2) << 9) + /* + * Indicates that the physical packet is shorter than that + * claimed by the PPPoE header length for a tunnel PPPoE packet. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_TUNNEL_TOTAL_ERROR \ + (UINT32_C(0x3) << 9) + /* + * Indicates that physical packet is shorter than that claimed + * by the tunnel l3 header length. Valid for IPv4, or IPv6 + * tunnel packet packets. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_T_IP_TOTAL_ERROR \ + (UINT32_C(0x4) << 9) + /* + * Indicates that the physical packet is shorter than that + * claimed by the tunnel UDP header length for a tunnel UDP + * packet that is not fragmented. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_T_UDP_TOTAL_ERROR \ + (UINT32_C(0x5) << 9) + /* + * indicates that the IPv4 TTL or IPv6 hop limit check have + * failed (e.g. TTL = 0) in the tunnel header. Valid for IPv4, + * and IPv6. + */ + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_T_L3_BAD_TTL \ + (UINT32_C(0x6) << 9) + #define RX_PKT_CMPL_ERRORS_T_PKT_ERROR_LAST \ + RX_PKT_CMPL_ERRORS_T_PKT_ERROR_T_L3_BAD_TTL + /* + * This indicates that there was an error in the inner portion of the + * packet when this field is non-zero. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_MASK UINT32_C(0xf000) + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_SFT 12 + /* + * No additional error occurred on the tunnel portion of the + * packet of the packet does not have a tunnel. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_NO_ERROR (UINT32_C(0x0) << 12) + /* + * Indicates that IP header version does not match expectation + * from L2 Ethertype for IPv4 and IPv6 or that option other than + * VFT was parsed on FCoE packet. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_L3_BAD_VERSION \ + (UINT32_C(0x1) << 12) + /* + * indicates that header length is out of range. Valid for IPv4 + * and RoCE + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_L3_BAD_HDR_LEN \ + (UINT32_C(0x2) << 12) + /* + * indicates that the IPv4 TTL or IPv6 hop limit check have + * failed (e.g. TTL = 0). Valid for IPv4, and IPv6 + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_L3_BAD_TTL (UINT32_C(0x3) << 12) + /* + * Indicates that physical packet is shorter than that claimed + * by the l3 header length. Valid for IPv4, IPv6 packet or RoCE + * packets. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_IP_TOTAL_ERROR \ + (UINT32_C(0x4) << 12) + /* + * Indicates that the physical packet is shorter than that + * claimed by the UDP header length for a UDP packet that is not + * fragmented. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_UDP_TOTAL_ERROR \ + (UINT32_C(0x5) << 12) + /* + * Indicates that TCP header length > IP payload. Valid for TCP + * packets only. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_L4_BAD_HDR_LEN \ + (UINT32_C(0x6) << 12) + /* Indicates that TCP header length < 5. Valid for TCP. */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_L4_BAD_HDR_LEN_TOO_SMALL \ + (UINT32_C(0x7) << 12) + /* + * Indicates that TCP option headers result in a TCP header size + * that does not match data offset in TCP header. Valid for TCP. + */ + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_L4_BAD_OPT_LEN \ + (UINT32_C(0x8) << 12) + #define RX_PKT_CMPL_ERRORS_PKT_ERROR_LAST \ + RX_PKT_CMPL_ERRORS_PKT_ERROR_L4_BAD_OPT_LEN + #define RX_PKT_CMPL_ERRORS_MASK UINT32_C(0xfffe) + #define RX_PKT_CMPL_ERRORS_SFT 1 + uint16_t errors_v2; + + /* + * This field identifies the CFA action rule that was used for this + * packet. + */ + uint16_t cfa_code; + + /* + * This value holds the reordering sequence number for the packet. If + * the reordering sequence is not valid, then this value is zero. The + * reordering domain for the packet is in the bottom 8 to 10b of the + * rss_hash value. The bottom 20b of this value contain the ordering + * domain value for the packet. + */ + #define RX_PKT_CMPL_REORDER_MASK UINT32_C(0xffffff) + #define RX_PKT_CMPL_REORDER_SFT 0 + uint32_t reorder; +} __attribute__((packed)); + /* HWRM Forwarded Request (16 bytes) */ struct hwrm_fwd_req_cmpl { /* Length of forwarded request in bytes. */