X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fsfc%2Fsfc_ethdev.c;h=47d7a8609b30edebf4911f4e76e1d89d093e2bed;hb=caccf8b318cafcdafe39faa3c5ce3eef67007621;hp=3d7483dd7a5949e01273a309d98ddec5d593ea02;hpb=82faef5076085c81573b1c6c62307d13acaf330c;p=dpdk.git diff --git a/drivers/net/sfc/sfc_ethdev.c b/drivers/net/sfc/sfc_ethdev.c index 3d7483dd7a..47d7a8609b 100644 --- a/drivers/net/sfc/sfc_ethdev.c +++ b/drivers/net/sfc/sfc_ethdev.c @@ -1,35 +1,18 @@ -/*- - * Copyright (c) 2016 Solarflare Communications Inc. +/* SPDX-License-Identifier: BSD-3-Clause + * + * Copyright (c) 2016-2018 Solarflare Communications Inc. * All rights reserved. * * This software was jointly developed between OKTET Labs (under contract * for Solarflare) and Solarflare Communications, Inc. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions are met: - * - * 1. Redistributions of source code must retain the above copyright notice, - * this list of conditions and the following disclaimer. - * 2. 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. - * - * 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 +#include #include "efx.h" @@ -40,17 +23,72 @@ #include "sfc_ev.h" #include "sfc_rx.h" #include "sfc_tx.h" +#include "sfc_flow.h" +#include "sfc_dp.h" +#include "sfc_dp_rx.h" + +uint32_t sfc_logtype_driver; + +static struct sfc_dp_list sfc_dp_head = + TAILQ_HEAD_INITIALIZER(sfc_dp_head); + +static int +sfc_fw_version_get(struct rte_eth_dev *dev, char *fw_version, size_t fw_size) +{ + struct sfc_adapter *sa = dev->data->dev_private; + efx_nic_fw_info_t enfi; + int ret; + int rc; + + /* + * Return value of the callback is likely supposed to be + * equal to or greater than 0, nevertheless, if an error + * occurs, it will be desirable to pass it to the caller + */ + if ((fw_version == NULL) || (fw_size == 0)) + return -EINVAL; + + rc = efx_nic_get_fw_version(sa->nic, &enfi); + if (rc != 0) + return -rc; + + ret = snprintf(fw_version, fw_size, + "%" PRIu16 ".%" PRIu16 ".%" PRIu16 ".%" PRIu16, + enfi.enfi_mc_fw_version[0], enfi.enfi_mc_fw_version[1], + enfi.enfi_mc_fw_version[2], enfi.enfi_mc_fw_version[3]); + if (ret < 0) + return ret; + + if (enfi.enfi_dpcpu_fw_ids_valid) { + size_t dpcpu_fw_ids_offset = MIN(fw_size - 1, (size_t)ret); + int ret_extra; + + ret_extra = snprintf(fw_version + dpcpu_fw_ids_offset, + fw_size - dpcpu_fw_ids_offset, + " rx%" PRIx16 " tx%" PRIx16, + enfi.enfi_rx_dpcpu_fw_id, + enfi.enfi_tx_dpcpu_fw_id); + if (ret_extra < 0) + return ret_extra; + + ret += ret_extra; + } + if (fw_size < (size_t)(++ret)) + return ret; + else + return 0; +} static void sfc_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) { struct sfc_adapter *sa = dev->data->dev_private; const efx_nic_cfg_t *encp = efx_nic_cfg_get(sa->nic); + uint64_t txq_offloads_def = 0; sfc_log_init(sa, "entry"); - dev_info->pci_dev = RTE_DEV_TO_PCI(dev->device); dev_info->max_rx_pktlen = EFX_MAC_PDU_MAX; /* Autonegotiation may be disabled */ @@ -59,8 +97,14 @@ sfc_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) dev_info->speed_capa |= ETH_LINK_SPEED_1G; if (sa->port.phy_adv_cap_mask & EFX_PHY_CAP_10000FDX) dev_info->speed_capa |= ETH_LINK_SPEED_10G; + if (sa->port.phy_adv_cap_mask & EFX_PHY_CAP_25000FDX) + dev_info->speed_capa |= ETH_LINK_SPEED_25G; if (sa->port.phy_adv_cap_mask & EFX_PHY_CAP_40000FDX) dev_info->speed_capa |= ETH_LINK_SPEED_40G; + if (sa->port.phy_adv_cap_mask & EFX_PHY_CAP_50000FDX) + dev_info->speed_capa |= ETH_LINK_SPEED_50G; + if (sa->port.phy_adv_cap_mask & EFX_PHY_CAP_100000FDX) + dev_info->speed_capa |= ETH_LINK_SPEED_100G; dev_info->max_rx_queues = sa->rxq_max; dev_info->max_tx_queues = sa->txq_max; @@ -68,30 +112,54 @@ sfc_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) /* By default packets are dropped if no descriptors are available */ dev_info->default_rxconf.rx_drop_en = 1; - dev_info->rx_offload_capa = - DEV_RX_OFFLOAD_IPV4_CKSUM | - DEV_RX_OFFLOAD_UDP_CKSUM | - DEV_RX_OFFLOAD_TCP_CKSUM; + dev_info->rx_queue_offload_capa = sfc_rx_get_queue_offload_caps(sa); + + /* + * rx_offload_capa includes both device and queue offloads since + * the latter may be requested on a per device basis which makes + * sense when some offloads are needed to be set on all queues. + */ + dev_info->rx_offload_capa = sfc_rx_get_dev_offload_caps(sa) | + dev_info->rx_queue_offload_capa; + + dev_info->tx_queue_offload_capa = sfc_tx_get_queue_offload_caps(sa); + + /* + * tx_offload_capa includes both device and queue offloads since + * the latter may be requested on a per device basis which makes + * sense when some offloads are needed to be set on all queues. + */ + dev_info->tx_offload_capa = sfc_tx_get_dev_offload_caps(sa) | + dev_info->tx_queue_offload_capa; - dev_info->tx_offload_capa = - DEV_TX_OFFLOAD_IPV4_CKSUM | - DEV_TX_OFFLOAD_UDP_CKSUM | - DEV_TX_OFFLOAD_TCP_CKSUM; + if (dev_info->tx_offload_capa & DEV_TX_OFFLOAD_MBUF_FAST_FREE) + txq_offloads_def |= DEV_TX_OFFLOAD_MBUF_FAST_FREE; + + dev_info->default_txconf.offloads |= txq_offloads_def; dev_info->default_txconf.txq_flags = ETH_TXQ_FLAGS_NOXSUMSCTP; - if (!encp->enc_hw_tx_insert_vlan_enabled) + if ((~sa->dp_tx->features & SFC_DP_TX_FEAT_VLAN_INSERT) || + !encp->enc_hw_tx_insert_vlan_enabled) dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOVLANOFFL; - else - dev_info->tx_offload_capa |= DEV_TX_OFFLOAD_VLAN_INSERT; + + if (~sa->dp_tx->features & SFC_DP_TX_FEAT_MULTI_SEG) + dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOMULTSEGS; + + if (~sa->dp_tx->features & SFC_DP_TX_FEAT_MULTI_POOL) + dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOMULTMEMP; + + if (~sa->dp_tx->features & SFC_DP_TX_FEAT_REFCNT) + dev_info->default_txconf.txq_flags |= ETH_TXQ_FLAGS_NOREFCOUNT; #if EFSYS_OPT_RX_SCALE if (sa->rss_support != EFX_RX_SCALE_UNAVAILABLE) { dev_info->reta_size = EFX_RSS_TBL_SIZE; - dev_info->hash_key_size = SFC_RSS_KEY_SIZE; + dev_info->hash_key_size = EFX_RSS_KEY_SIZE; dev_info->flow_type_rss_offloads = SFC_RSS_OFFLOADS; } #endif + /* Initialize to hardware limits */ dev_info->rx_desc_lim.nb_max = EFX_RXQ_MAXNDESCS; dev_info->rx_desc_lim.nb_min = EFX_RXQ_MINNDESCS; /* The RXQ hardware requires that the descriptor count is a power @@ -99,6 +167,7 @@ sfc_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) */ dev_info->rx_desc_lim.nb_align = EFX_RXQ_MINNDESCS; + /* Initialize to hardware limits */ dev_info->tx_desc_lim.nb_max = sa->txq_max_entries; dev_info->tx_desc_lim.nb_min = EFX_TXQ_MINNDESCS; /* @@ -106,24 +175,21 @@ sfc_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) * of 2, but tx_desc_lim cannot properly describe that constraint */ dev_info->tx_desc_lim.nb_align = EFX_TXQ_MINNDESCS; + + if (sa->dp_rx->get_dev_info != NULL) + sa->dp_rx->get_dev_info(dev_info); + if (sa->dp_tx->get_dev_info != NULL) + sa->dp_tx->get_dev_info(dev_info); } static const uint32_t * sfc_dev_supported_ptypes_get(struct rte_eth_dev *dev) { - static const uint32_t ptypes[] = { - RTE_PTYPE_L2_ETHER, - RTE_PTYPE_L3_IPV4_EXT_UNKNOWN, - RTE_PTYPE_L3_IPV6_EXT_UNKNOWN, - RTE_PTYPE_L4_TCP, - RTE_PTYPE_L4_UDP, - RTE_PTYPE_UNKNOWN - }; - - if (dev->rx_pkt_burst == sfc_recv_pkts) - return ptypes; - - return NULL; + struct sfc_adapter *sa = dev->data->dev_private; + const efx_nic_cfg_t *encp = efx_nic_cfg_get(sa->nic); + uint32_t tunnel_encaps = encp->enc_tunnel_encapsulations_supported; + + return sa->dp_rx->supported_ptypes_get(tunnel_encaps); } static int @@ -139,8 +205,6 @@ sfc_dev_configure(struct rte_eth_dev *dev) sfc_adapter_lock(sa); switch (sa->state) { case SFC_ADAPTER_CONFIGURED: - sfc_close(sa); - SFC_ASSERT(sa->state == SFC_ADAPTER_INITIALIZED); /* FALLTHROUGH */ case SFC_ADAPTER_INITIALIZED: rc = sfc_configure(sa); @@ -179,40 +243,31 @@ static int sfc_dev_link_update(struct rte_eth_dev *dev, int wait_to_complete) { struct sfc_adapter *sa = dev->data->dev_private; - struct rte_eth_link *dev_link = &dev->data->dev_link; - struct rte_eth_link old_link; struct rte_eth_link current_link; + int ret; sfc_log_init(sa, "entry"); - if (sa->state != SFC_ADAPTER_STARTED) - return 0; - -retry: - EFX_STATIC_ASSERT(sizeof(*dev_link) == sizeof(rte_atomic64_t)); - *(int64_t *)&old_link = rte_atomic64_read((rte_atomic64_t *)dev_link); - - if (wait_to_complete) { + if (sa->state != SFC_ADAPTER_STARTED) { + sfc_port_link_mode_to_info(EFX_LINK_UNKNOWN, ¤t_link); + } else if (wait_to_complete) { efx_link_mode_t link_mode; - efx_port_poll(sa->nic, &link_mode); + if (efx_port_poll(sa->nic, &link_mode) != 0) + link_mode = EFX_LINK_UNKNOWN; sfc_port_link_mode_to_info(link_mode, ¤t_link); - if (!rte_atomic64_cmpset((volatile uint64_t *)dev_link, - *(uint64_t *)&old_link, - *(uint64_t *)¤t_link)) - goto retry; } else { sfc_ev_mgmt_qpoll(sa); - *(int64_t *)¤t_link = - rte_atomic64_read((rte_atomic64_t *)dev_link); + rte_eth_linkstatus_get(dev, ¤t_link); } - if (old_link.link_status != current_link.link_status) - sfc_info(sa, "Link status is %s", - current_link.link_status ? "UP" : "DOWN"); + ret = rte_eth_linkstatus_set(dev, ¤t_link); + if (ret == 0) + sfc_notice(sa, "Link status is %s", + current_link.link_status ? "UP" : "DOWN"); - return old_link.link_status == current_link.link_status ? 0 : -1; + return ret; } static void @@ -305,8 +360,13 @@ sfc_dev_filter_set(struct rte_eth_dev *dev, enum sfc_dev_filter_mode mode, if (*toggle != enabled) { *toggle = enabled; - if ((sa->state == SFC_ADAPTER_STARTED) && - (sfc_set_rx_mode(sa) != 0)) { + if (port->isolated) { + sfc_warn(sa, "isolated mode is active on the port"); + sfc_warn(sa, "the change is to be applied on the next " + "start provided that isolated mode is " + "disabled prior the next start"); + } else if ((sa->state == SFC_ADAPTER_STARTED) && + (sfc_set_rx_mode(sa) != 0)) { *toggle = !(enabled); sfc_warn(sa, "Failed to %s %s mode", ((enabled) ? "enable" : "disable"), desc); @@ -359,7 +419,7 @@ sfc_rx_queue_setup(struct rte_eth_dev *dev, uint16_t rx_queue_id, if (rc != 0) goto fail_rx_qinit; - dev->data->rx_queues[rx_queue_id] = sa->rxq_info[rx_queue_id].rxq; + dev->data->rx_queues[rx_queue_id] = sa->rxq_info[rx_queue_id].rxq->dp; sfc_adapter_unlock(sa); @@ -374,13 +434,15 @@ fail_rx_qinit: static void sfc_rx_queue_release(void *queue) { - struct sfc_rxq *rxq = queue; + struct sfc_dp_rxq *dp_rxq = queue; + struct sfc_rxq *rxq; struct sfc_adapter *sa; unsigned int sw_index; - if (rxq == NULL) + if (dp_rxq == NULL) return; + rxq = sfc_rxq_by_dp_rxq(dp_rxq); sa = rxq->evq->sa; sfc_adapter_lock(sa); @@ -412,7 +474,7 @@ sfc_tx_queue_setup(struct rte_eth_dev *dev, uint16_t tx_queue_id, if (rc != 0) goto fail_tx_qinit; - dev->data->tx_queues[tx_queue_id] = sa->txq_info[tx_queue_id].txq; + dev->data->tx_queues[tx_queue_id] = sa->txq_info[tx_queue_id].txq->dp; sfc_adapter_unlock(sa); return 0; @@ -426,13 +488,15 @@ fail_tx_qinit: static void sfc_tx_queue_release(void *queue) { - struct sfc_txq *txq = queue; + struct sfc_dp_txq *dp_txq = queue; + struct sfc_txq *txq; unsigned int sw_index; struct sfc_adapter *sa; - if (txq == NULL) + if (dp_txq == NULL) return; + txq = sfc_txq_by_dp_txq(dp_txq); sw_index = sfc_txq_sw_index(txq); SFC_ASSERT(txq->evq != NULL); @@ -450,16 +514,18 @@ sfc_tx_queue_release(void *queue) sfc_adapter_unlock(sa); } -static void +static int sfc_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) { struct sfc_adapter *sa = dev->data->dev_private; struct sfc_port *port = &sa->port; uint64_t *mac_stats; + int ret; rte_spinlock_lock(&port->mac_stats_lock); - if (sfc_port_update_mac_stats(sa) != 0) + ret = sfc_port_update_mac_stats(sa); + if (ret != 0) goto unlock; mac_stats = port->mac_stats_buf; @@ -516,6 +582,29 @@ sfc_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) unlock: rte_spinlock_unlock(&port->mac_stats_lock); + SFC_ASSERT(ret >= 0); + return -ret; +} + +static void +sfc_stats_reset(struct rte_eth_dev *dev) +{ + struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; + int rc; + + if (sa->state != SFC_ADAPTER_STARTED) { + /* + * The operation cannot be done if port is not started; it + * will be scheduled to be done during the next port start + */ + port->mac_stats_reset_pending = B_TRUE; + return; + } + + rc = sfc_port_reset_mac_stats(sa); + if (rc != 0) + sfc_err(sa, "failed to reset statistics (rc = %d)", rc); } static int @@ -579,6 +668,85 @@ sfc_xstats_get_names(struct rte_eth_dev *dev, return nstats; } +static int +sfc_xstats_get_by_id(struct rte_eth_dev *dev, const uint64_t *ids, + uint64_t *values, unsigned int n) +{ + struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; + uint64_t *mac_stats; + unsigned int nb_supported = 0; + unsigned int nb_written = 0; + unsigned int i; + int ret; + int rc; + + if (unlikely(values == NULL) || + unlikely((ids == NULL) && (n < port->mac_stats_nb_supported))) + return port->mac_stats_nb_supported; + + rte_spinlock_lock(&port->mac_stats_lock); + + rc = sfc_port_update_mac_stats(sa); + if (rc != 0) { + SFC_ASSERT(rc > 0); + ret = -rc; + goto unlock; + } + + mac_stats = port->mac_stats_buf; + + for (i = 0; (i < EFX_MAC_NSTATS) && (nb_written < n); ++i) { + if (!EFX_MAC_STAT_SUPPORTED(port->mac_stats_mask, i)) + continue; + + if ((ids == NULL) || (ids[nb_written] == nb_supported)) + values[nb_written++] = mac_stats[i]; + + ++nb_supported; + } + + ret = nb_written; + +unlock: + rte_spinlock_unlock(&port->mac_stats_lock); + + return ret; +} + +static int +sfc_xstats_get_names_by_id(struct rte_eth_dev *dev, + struct rte_eth_xstat_name *xstats_names, + const uint64_t *ids, unsigned int size) +{ + struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; + unsigned int nb_supported = 0; + unsigned int nb_written = 0; + unsigned int i; + + if (unlikely(xstats_names == NULL) || + unlikely((ids == NULL) && (size < port->mac_stats_nb_supported))) + return port->mac_stats_nb_supported; + + for (i = 0; (i < EFX_MAC_NSTATS) && (nb_written < size); ++i) { + if (!EFX_MAC_STAT_SUPPORTED(port->mac_stats_mask, i)) + continue; + + if ((ids == NULL) || (ids[nb_written] == nb_supported)) { + char *name = xstats_names[nb_written++].name; + + strncpy(name, efx_mac_stat_name(sa->nic, i), + sizeof(xstats_names[0].name)); + name[sizeof(xstats_names[0].name) - 1] = '\0'; + } + + ++nb_supported; + } + + return nb_written; +} + static int sfc_flow_ctrl_get(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf) { @@ -719,7 +887,13 @@ sfc_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu) * The driver does not use it, but other PMDs update jumbo_frame * flag and max_rx_pkt_len when MTU is set. */ - dev->data->dev_conf.rxmode.jumbo_frame = (mtu > ETHER_MAX_LEN); + if (mtu > ETHER_MAX_LEN) { + struct rte_eth_rxmode *rxmode = &dev->data->dev_conf.rxmode; + + rxmode->offloads |= DEV_RX_OFFLOAD_JUMBO_FRAME; + rxmode->jumbo_frame = 1; + } + dev->data->dev_conf.rxmode.max_rx_pkt_len = sa->port.pdu; sfc_adapter_unlock(sa); @@ -740,18 +914,39 @@ fail_inval: SFC_ASSERT(rc > 0); return -rc; } -static void +static int sfc_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) { struct sfc_adapter *sa = dev->data->dev_private; const efx_nic_cfg_t *encp = efx_nic_cfg_get(sa->nic); - int rc; + struct sfc_port *port = &sa->port; + struct ether_addr *old_addr = &dev->data->mac_addrs[0]; + int rc = 0; sfc_adapter_lock(sa); + /* + * Copy the address to the device private data so that + * it could be recalled in the case of adapter restart. + */ + ether_addr_copy(mac_addr, &port->default_mac_addr); + + /* + * Neither of the two following checks can return + * an error. The new MAC address is preserved in + * the device private data and can be activated + * on the next port start if the user prevents + * isolated mode from being enabled. + */ + if (port->isolated) { + sfc_warn(sa, "isolated mode is active on the port"); + sfc_warn(sa, "will not set MAC address"); + goto unlock; + } + if (sa->state != SFC_ADAPTER_STARTED) { - sfc_info(sa, "the port is not started"); - sfc_info(sa, "the new MAC address will be set on port start"); + sfc_notice(sa, "the port is not started"); + sfc_notice(sa, "the new MAC address will be set on port start"); goto unlock; } @@ -769,8 +964,12 @@ sfc_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) * we also need to update unicast filters */ rc = sfc_set_rx_mode(sa); - if (rc != 0) + if (rc != 0) { sfc_err(sa, "cannot set filter (rc = %u)", rc); + /* Rollback the old address */ + (void)efx_mac_addr_set(sa->nic, old_addr->addr_bytes); + (void)sfc_set_rx_mode(sa); + } } else { sfc_warn(sa, "cannot set MAC address with filters installed"); sfc_warn(sa, "adapter will be restarted to pick the new MAC"); @@ -778,9 +977,9 @@ sfc_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) /* * Since setting MAC address with filters installed is not - * allowed on the adapter, one needs to simply restart adapter - * so that the new MAC address will be taken from an outer - * storage and set flawlessly by means of sfc_start() call + * allowed on the adapter, the new MAC address will be set + * by means of adapter restart. sfc_start() shall retrieve + * the new address from the device private data and set it. */ sfc_stop(sa); rc = sfc_start(sa); @@ -789,7 +988,13 @@ sfc_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) } unlock: + if (rc != 0) + ether_addr_copy(old_addr, &port->default_mac_addr); + sfc_adapter_unlock(sa); + + SFC_ASSERT(rc >= 0); + return -rc; } @@ -798,33 +1003,39 @@ sfc_set_mc_addr_list(struct rte_eth_dev *dev, struct ether_addr *mc_addr_set, uint32_t nb_mc_addr) { struct sfc_adapter *sa = dev->data->dev_private; - uint8_t *mc_addrs_p; - uint8_t *mc_addrs; + struct sfc_port *port = &sa->port; + uint8_t *mc_addrs = port->mcast_addrs; int rc; unsigned int i; - if (nb_mc_addr > EFX_MAC_MULTICAST_LIST_MAX) { - sfc_err(sa, "too many multicast addresses: %u > %u", - nb_mc_addr, EFX_MAC_MULTICAST_LIST_MAX); - return -EINVAL; + if (port->isolated) { + sfc_err(sa, "isolated mode is active on the port"); + sfc_err(sa, "will not set multicast address list"); + return -ENOTSUP; } - mc_addrs_p = rte_calloc("mc-addrs", nb_mc_addr, EFX_MAC_ADDR_LEN, 0); - if (mc_addrs_p == NULL) - return -ENOMEM; + if (mc_addrs == NULL) + return -ENOBUFS; - mc_addrs = mc_addrs_p; + if (nb_mc_addr > port->max_mcast_addrs) { + sfc_err(sa, "too many multicast addresses: %u > %u", + nb_mc_addr, port->max_mcast_addrs); + return -EINVAL; + } for (i = 0; i < nb_mc_addr; ++i) { - (void)rte_memcpy(mc_addrs, mc_addr_set[i].addr_bytes, + rte_memcpy(mc_addrs, mc_addr_set[i].addr_bytes, EFX_MAC_ADDR_LEN); mc_addrs += EFX_MAC_ADDR_LEN; } - rc = efx_mac_multicast_list_set(sa->nic, mc_addrs_p, nb_mc_addr); + port->nb_mcast_addrs = nb_mc_addr; - rte_free(mc_addrs_p); + if (sa->state != SFC_ADAPTER_STARTED) + return 0; + rc = efx_mac_multicast_list_set(sa->nic, port->mcast_addrs, + port->nb_mcast_addrs); if (rc != 0) sfc_err(sa, "cannot set multicast address list (rc = %u)", rc); @@ -832,6 +1043,10 @@ sfc_set_mc_addr_list(struct rte_eth_dev *dev, struct ether_addr *mc_addr_set, return -rc; } +/* + * The function is used by the secondary process as well. It must not + * use any process-local pointers from the adapter data. + */ static void sfc_rx_queue_info_get(struct rte_eth_dev *dev, uint16_t rx_queue_id, struct rte_eth_rxq_info *qinfo) @@ -852,12 +1067,22 @@ sfc_rx_queue_info_get(struct rte_eth_dev *dev, uint16_t rx_queue_id, qinfo->conf.rx_free_thresh = rxq->refill_threshold; qinfo->conf.rx_drop_en = 1; qinfo->conf.rx_deferred_start = rxq_info->deferred_start; - qinfo->scattered_rx = (rxq_info->type == EFX_RXQ_TYPE_SCATTER); + qinfo->conf.offloads = DEV_RX_OFFLOAD_IPV4_CKSUM | + DEV_RX_OFFLOAD_UDP_CKSUM | + DEV_RX_OFFLOAD_TCP_CKSUM; + if (rxq_info->type_flags & EFX_RXQ_FLAG_SCATTER) { + qinfo->conf.offloads |= DEV_RX_OFFLOAD_SCATTER; + qinfo->scattered_rx = 1; + } qinfo->nb_desc = rxq_info->entries; sfc_adapter_unlock(sa); } +/* + * The function is used by the secondary process as well. It must not + * use any process-local pointers from the adapter data. + */ static void sfc_tx_queue_info_get(struct rte_eth_dev *dev, uint16_t tx_queue_id, struct rte_eth_txq_info *qinfo) @@ -875,6 +1100,7 @@ sfc_tx_queue_info_get(struct rte_eth_dev *dev, uint16_t tx_queue_id, memset(qinfo, 0, sizeof(*qinfo)); qinfo->conf.txq_flags = txq_info->txq->flags; + qinfo->conf.offloads = txq_info->txq->offloads; qinfo->conf.tx_free_thresh = txq_info->txq->free_thresh; qinfo->conf.tx_deferred_start = txq_info->deferred_start; qinfo->nb_desc = txq_info->entries; @@ -895,9 +1121,27 @@ sfc_rx_queue_count(struct rte_eth_dev *dev, uint16_t rx_queue_id) static int sfc_rx_descriptor_done(void *queue, uint16_t offset) { - struct sfc_rxq *rxq = queue; + struct sfc_dp_rxq *dp_rxq = queue; - return sfc_rx_qdesc_done(rxq, offset); + return sfc_rx_qdesc_done(dp_rxq, offset); +} + +static int +sfc_rx_descriptor_status(void *queue, uint16_t offset) +{ + struct sfc_dp_rxq *dp_rxq = queue; + struct sfc_rxq *rxq = sfc_rxq_by_dp_rxq(dp_rxq); + + return rxq->evq->sa->dp_rx->qdesc_status(dp_rxq, offset); +} + +static int +sfc_tx_descriptor_status(void *queue, uint16_t offset) +{ + struct sfc_dp_txq *dp_txq = queue; + struct sfc_txq *txq = sfc_txq_by_dp_txq(dp_txq); + + return txq->evq->sa->dp_tx->qdesc_status(dp_txq, offset); } static int @@ -996,17 +1240,137 @@ sfc_tx_queue_stop(struct rte_eth_dev *dev, uint16_t tx_queue_id) return 0; } +static efx_tunnel_protocol_t +sfc_tunnel_rte_type_to_efx_udp_proto(enum rte_eth_tunnel_type rte_type) +{ + switch (rte_type) { + case RTE_TUNNEL_TYPE_VXLAN: + return EFX_TUNNEL_PROTOCOL_VXLAN; + case RTE_TUNNEL_TYPE_GENEVE: + return EFX_TUNNEL_PROTOCOL_GENEVE; + default: + return EFX_TUNNEL_NPROTOS; + } +} + +enum sfc_udp_tunnel_op_e { + SFC_UDP_TUNNEL_ADD_PORT, + SFC_UDP_TUNNEL_DEL_PORT, +}; + +static int +sfc_dev_udp_tunnel_op(struct rte_eth_dev *dev, + struct rte_eth_udp_tunnel *tunnel_udp, + enum sfc_udp_tunnel_op_e op) +{ + struct sfc_adapter *sa = dev->data->dev_private; + efx_tunnel_protocol_t tunnel_proto; + int rc; + + sfc_log_init(sa, "%s udp_port=%u prot_type=%u", + (op == SFC_UDP_TUNNEL_ADD_PORT) ? "add" : + (op == SFC_UDP_TUNNEL_DEL_PORT) ? "delete" : "unknown", + tunnel_udp->udp_port, tunnel_udp->prot_type); + + tunnel_proto = + sfc_tunnel_rte_type_to_efx_udp_proto(tunnel_udp->prot_type); + if (tunnel_proto >= EFX_TUNNEL_NPROTOS) { + rc = ENOTSUP; + goto fail_bad_proto; + } + + sfc_adapter_lock(sa); + + switch (op) { + case SFC_UDP_TUNNEL_ADD_PORT: + rc = efx_tunnel_config_udp_add(sa->nic, + tunnel_udp->udp_port, + tunnel_proto); + break; + case SFC_UDP_TUNNEL_DEL_PORT: + rc = efx_tunnel_config_udp_remove(sa->nic, + tunnel_udp->udp_port, + tunnel_proto); + break; + default: + rc = EINVAL; + goto fail_bad_op; + } + + if (rc != 0) + goto fail_op; + + if (sa->state == SFC_ADAPTER_STARTED) { + rc = efx_tunnel_reconfigure(sa->nic); + if (rc == EAGAIN) { + /* + * Configuration is accepted by FW and MC reboot + * is initiated to apply the changes. MC reboot + * will be handled in a usual way (MC reboot + * event on management event queue and adapter + * restart). + */ + rc = 0; + } else if (rc != 0) { + goto fail_reconfigure; + } + } + + sfc_adapter_unlock(sa); + return 0; + +fail_reconfigure: + /* Remove/restore entry since the change makes the trouble */ + switch (op) { + case SFC_UDP_TUNNEL_ADD_PORT: + (void)efx_tunnel_config_udp_remove(sa->nic, + tunnel_udp->udp_port, + tunnel_proto); + break; + case SFC_UDP_TUNNEL_DEL_PORT: + (void)efx_tunnel_config_udp_add(sa->nic, + tunnel_udp->udp_port, + tunnel_proto); + break; + } + +fail_op: +fail_bad_op: + sfc_adapter_unlock(sa); + +fail_bad_proto: + SFC_ASSERT(rc > 0); + return -rc; +} + +static int +sfc_dev_udp_tunnel_port_add(struct rte_eth_dev *dev, + struct rte_eth_udp_tunnel *tunnel_udp) +{ + return sfc_dev_udp_tunnel_op(dev, tunnel_udp, SFC_UDP_TUNNEL_ADD_PORT); +} + +static int +sfc_dev_udp_tunnel_port_del(struct rte_eth_dev *dev, + struct rte_eth_udp_tunnel *tunnel_udp) +{ + return sfc_dev_udp_tunnel_op(dev, tunnel_udp, SFC_UDP_TUNNEL_DEL_PORT); +} + #if EFSYS_OPT_RX_SCALE static int sfc_dev_rss_hash_conf_get(struct rte_eth_dev *dev, struct rte_eth_rss_conf *rss_conf) { struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; - if ((sa->rss_channels == 1) || - (sa->rss_support != EFX_RX_SCALE_EXCLUSIVE)) + if ((sa->rss_support != EFX_RX_SCALE_EXCLUSIVE) || port->isolated) return -ENOTSUP; + if (sa->rss_channels == 0) + return -EINVAL; + sfc_adapter_lock(sa); /* @@ -1016,9 +1380,9 @@ sfc_dev_rss_hash_conf_get(struct rte_eth_dev *dev, * locally in 'sfc_adapter' and kept up-to-date */ rss_conf->rss_hf = sfc_efx_to_rte_hash_type(sa->rss_hash_types); - rss_conf->rss_key_len = SFC_RSS_KEY_SIZE; + rss_conf->rss_key_len = EFX_RSS_KEY_SIZE; if (rss_conf->rss_key != NULL) - rte_memcpy(rss_conf->rss_key, sa->rss_key, SFC_RSS_KEY_SIZE); + rte_memcpy(rss_conf->rss_key, sa->rss_key, EFX_RSS_KEY_SIZE); sfc_adapter_unlock(sa); @@ -1030,15 +1394,23 @@ sfc_dev_rss_hash_update(struct rte_eth_dev *dev, struct rte_eth_rss_conf *rss_conf) { struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; unsigned int efx_hash_types; int rc = 0; - if ((sa->rss_channels == 1) || - (sa->rss_support != EFX_RX_SCALE_EXCLUSIVE)) { + if (port->isolated) + return -ENOTSUP; + + if (sa->rss_support != EFX_RX_SCALE_EXCLUSIVE) { sfc_err(sa, "RSS is not available"); return -ENOTSUP; } + if (sa->rss_channels == 0) { + sfc_err(sa, "RSS is not configured"); + return -EINVAL; + } + if ((rss_conf->rss_key != NULL) && (rss_conf->rss_key_len != sizeof(sa->rss_key))) { sfc_err(sa, "RSS key size is wrong (should be %lu)", @@ -1055,14 +1427,17 @@ sfc_dev_rss_hash_update(struct rte_eth_dev *dev, efx_hash_types = sfc_rte_to_efx_hash_type(rss_conf->rss_hf); - rc = efx_rx_scale_mode_set(sa->nic, EFX_RX_HASHALG_TOEPLITZ, + rc = efx_rx_scale_mode_set(sa->nic, EFX_RSS_CONTEXT_DEFAULT, + EFX_RX_HASHALG_TOEPLITZ, efx_hash_types, B_TRUE); if (rc != 0) goto fail_scale_mode_set; if (rss_conf->rss_key != NULL) { if (sa->state == SFC_ADAPTER_STARTED) { - rc = efx_rx_scale_key_set(sa->nic, rss_conf->rss_key, + rc = efx_rx_scale_key_set(sa->nic, + EFX_RSS_CONTEXT_DEFAULT, + rss_conf->rss_key, sizeof(sa->rss_key)); if (rc != 0) goto fail_scale_key_set; @@ -1078,7 +1453,8 @@ sfc_dev_rss_hash_update(struct rte_eth_dev *dev, return 0; fail_scale_key_set: - if (efx_rx_scale_mode_set(sa->nic, EFX_RX_HASHALG_TOEPLITZ, + if (efx_rx_scale_mode_set(sa->nic, EFX_RSS_CONTEXT_DEFAULT, + EFX_RX_HASHALG_TOEPLITZ, sa->rss_hash_types, B_TRUE) != 0) sfc_err(sa, "failed to restore RSS mode"); @@ -1086,8 +1462,170 @@ fail_scale_mode_set: sfc_adapter_unlock(sa); return -rc; } + +static int +sfc_dev_rss_reta_query(struct rte_eth_dev *dev, + struct rte_eth_rss_reta_entry64 *reta_conf, + uint16_t reta_size) +{ + struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; + int entry; + + if ((sa->rss_support != EFX_RX_SCALE_EXCLUSIVE) || port->isolated) + return -ENOTSUP; + + if (sa->rss_channels == 0) + return -EINVAL; + + if (reta_size != EFX_RSS_TBL_SIZE) + return -EINVAL; + + sfc_adapter_lock(sa); + + for (entry = 0; entry < reta_size; entry++) { + int grp = entry / RTE_RETA_GROUP_SIZE; + int grp_idx = entry % RTE_RETA_GROUP_SIZE; + + if ((reta_conf[grp].mask >> grp_idx) & 1) + reta_conf[grp].reta[grp_idx] = sa->rss_tbl[entry]; + } + + sfc_adapter_unlock(sa); + + return 0; +} + +static int +sfc_dev_rss_reta_update(struct rte_eth_dev *dev, + struct rte_eth_rss_reta_entry64 *reta_conf, + uint16_t reta_size) +{ + struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_port *port = &sa->port; + unsigned int *rss_tbl_new; + uint16_t entry; + int rc = 0; + + + if (port->isolated) + return -ENOTSUP; + + if (sa->rss_support != EFX_RX_SCALE_EXCLUSIVE) { + sfc_err(sa, "RSS is not available"); + return -ENOTSUP; + } + + if (sa->rss_channels == 0) { + sfc_err(sa, "RSS is not configured"); + return -EINVAL; + } + + if (reta_size != EFX_RSS_TBL_SIZE) { + sfc_err(sa, "RETA size is wrong (should be %u)", + EFX_RSS_TBL_SIZE); + return -EINVAL; + } + + rss_tbl_new = rte_zmalloc("rss_tbl_new", sizeof(sa->rss_tbl), 0); + if (rss_tbl_new == NULL) + return -ENOMEM; + + sfc_adapter_lock(sa); + + rte_memcpy(rss_tbl_new, sa->rss_tbl, sizeof(sa->rss_tbl)); + + for (entry = 0; entry < reta_size; entry++) { + int grp_idx = entry % RTE_RETA_GROUP_SIZE; + struct rte_eth_rss_reta_entry64 *grp; + + grp = &reta_conf[entry / RTE_RETA_GROUP_SIZE]; + + if (grp->mask & (1ull << grp_idx)) { + if (grp->reta[grp_idx] >= sa->rss_channels) { + rc = EINVAL; + goto bad_reta_entry; + } + rss_tbl_new[entry] = grp->reta[grp_idx]; + } + } + + if (sa->state == SFC_ADAPTER_STARTED) { + rc = efx_rx_scale_tbl_set(sa->nic, EFX_RSS_CONTEXT_DEFAULT, + rss_tbl_new, EFX_RSS_TBL_SIZE); + if (rc != 0) + goto fail_scale_tbl_set; + } + + rte_memcpy(sa->rss_tbl, rss_tbl_new, sizeof(sa->rss_tbl)); + +fail_scale_tbl_set: +bad_reta_entry: + sfc_adapter_unlock(sa); + + rte_free(rss_tbl_new); + + SFC_ASSERT(rc >= 0); + return -rc; +} #endif +static int +sfc_dev_filter_ctrl(struct rte_eth_dev *dev, enum rte_filter_type filter_type, + enum rte_filter_op filter_op, + void *arg) +{ + struct sfc_adapter *sa = dev->data->dev_private; + int rc = ENOTSUP; + + sfc_log_init(sa, "entry"); + + switch (filter_type) { + case RTE_ETH_FILTER_NONE: + sfc_err(sa, "Global filters configuration not supported"); + break; + case RTE_ETH_FILTER_MACVLAN: + sfc_err(sa, "MACVLAN filters not supported"); + break; + case RTE_ETH_FILTER_ETHERTYPE: + sfc_err(sa, "EtherType filters not supported"); + break; + case RTE_ETH_FILTER_FLEXIBLE: + sfc_err(sa, "Flexible filters not supported"); + break; + case RTE_ETH_FILTER_SYN: + sfc_err(sa, "SYN filters not supported"); + break; + case RTE_ETH_FILTER_NTUPLE: + sfc_err(sa, "NTUPLE filters not supported"); + break; + case RTE_ETH_FILTER_TUNNEL: + sfc_err(sa, "Tunnel filters not supported"); + break; + case RTE_ETH_FILTER_FDIR: + sfc_err(sa, "Flow Director filters not supported"); + break; + case RTE_ETH_FILTER_HASH: + sfc_err(sa, "Hash filters not supported"); + break; + case RTE_ETH_FILTER_GENERIC: + if (filter_op != RTE_ETH_FILTER_GET) { + rc = EINVAL; + } else { + *(const void **)arg = &sfc_flow_ops; + rc = 0; + } + break; + default: + sfc_err(sa, "Unknown filter type %u", filter_type); + break; + } + + sfc_log_init(sa, "exit: %d", -rc); + SFC_ASSERT(rc >= 0); + return -rc; +} + static const struct eth_dev_ops sfc_eth_dev_ops = { .dev_configure = sfc_dev_configure, .dev_start = sfc_dev_start, @@ -1101,7 +1639,9 @@ static const struct eth_dev_ops sfc_eth_dev_ops = { .allmulticast_disable = sfc_dev_allmulti_disable, .link_update = sfc_dev_link_update, .stats_get = sfc_stats_get, + .stats_reset = sfc_stats_reset, .xstats_get = sfc_xstats_get, + .xstats_reset = sfc_stats_reset, .xstats_get_names = sfc_xstats_get_names, .dev_infos_get = sfc_dev_infos_get, .dev_supported_ptypes_get = sfc_dev_supported_ptypes_get, @@ -1114,44 +1654,300 @@ static const struct eth_dev_ops sfc_eth_dev_ops = { .rx_queue_release = sfc_rx_queue_release, .rx_queue_count = sfc_rx_queue_count, .rx_descriptor_done = sfc_rx_descriptor_done, + .rx_descriptor_status = sfc_rx_descriptor_status, + .tx_descriptor_status = sfc_tx_descriptor_status, .tx_queue_setup = sfc_tx_queue_setup, .tx_queue_release = sfc_tx_queue_release, .flow_ctrl_get = sfc_flow_ctrl_get, .flow_ctrl_set = sfc_flow_ctrl_set, .mac_addr_set = sfc_mac_addr_set, + .udp_tunnel_port_add = sfc_dev_udp_tunnel_port_add, + .udp_tunnel_port_del = sfc_dev_udp_tunnel_port_del, #if EFSYS_OPT_RX_SCALE + .reta_update = sfc_dev_rss_reta_update, + .reta_query = sfc_dev_rss_reta_query, .rss_hash_update = sfc_dev_rss_hash_update, .rss_hash_conf_get = sfc_dev_rss_hash_conf_get, #endif + .filter_ctrl = sfc_dev_filter_ctrl, .set_mc_addr_list = sfc_set_mc_addr_list, .rxq_info_get = sfc_rx_queue_info_get, .txq_info_get = sfc_tx_queue_info_get, + .fw_version_get = sfc_fw_version_get, + .xstats_get_by_id = sfc_xstats_get_by_id, + .xstats_get_names_by_id = sfc_xstats_get_names_by_id, }; +/** + * Duplicate a string in potentially shared memory required for + * multi-process support. + * + * strdup() allocates from process-local heap/memory. + */ +static char * +sfc_strdup(const char *str) +{ + size_t size; + char *copy; + + if (str == NULL) + return NULL; + + size = strlen(str) + 1; + copy = rte_malloc(__func__, size, 0); + if (copy != NULL) + rte_memcpy(copy, str, size); + + return copy; +} + +static int +sfc_eth_dev_set_ops(struct rte_eth_dev *dev) +{ + struct sfc_adapter *sa = dev->data->dev_private; + unsigned int avail_caps = 0; + const char *rx_name = NULL; + const char *tx_name = NULL; + int rc; + + switch (sa->family) { + case EFX_FAMILY_HUNTINGTON: + case EFX_FAMILY_MEDFORD: + case EFX_FAMILY_MEDFORD2: + avail_caps |= SFC_DP_HW_FW_CAP_EF10; + break; + default: + break; + } + + rc = sfc_kvargs_process(sa, SFC_KVARG_RX_DATAPATH, + sfc_kvarg_string_handler, &rx_name); + if (rc != 0) + goto fail_kvarg_rx_datapath; + + if (rx_name != NULL) { + sa->dp_rx = sfc_dp_find_rx_by_name(&sfc_dp_head, rx_name); + if (sa->dp_rx == NULL) { + sfc_err(sa, "Rx datapath %s not found", rx_name); + rc = ENOENT; + goto fail_dp_rx; + } + if (!sfc_dp_match_hw_fw_caps(&sa->dp_rx->dp, avail_caps)) { + sfc_err(sa, + "Insufficient Hw/FW capabilities to use Rx datapath %s", + rx_name); + rc = EINVAL; + goto fail_dp_rx_caps; + } + } else { + sa->dp_rx = sfc_dp_find_rx_by_caps(&sfc_dp_head, avail_caps); + if (sa->dp_rx == NULL) { + sfc_err(sa, "Rx datapath by caps %#x not found", + avail_caps); + rc = ENOENT; + goto fail_dp_rx; + } + } + + sa->dp_rx_name = sfc_strdup(sa->dp_rx->dp.name); + if (sa->dp_rx_name == NULL) { + rc = ENOMEM; + goto fail_dp_rx_name; + } + + sfc_notice(sa, "use %s Rx datapath", sa->dp_rx_name); + + dev->rx_pkt_burst = sa->dp_rx->pkt_burst; + + rc = sfc_kvargs_process(sa, SFC_KVARG_TX_DATAPATH, + sfc_kvarg_string_handler, &tx_name); + if (rc != 0) + goto fail_kvarg_tx_datapath; + + if (tx_name != NULL) { + sa->dp_tx = sfc_dp_find_tx_by_name(&sfc_dp_head, tx_name); + if (sa->dp_tx == NULL) { + sfc_err(sa, "Tx datapath %s not found", tx_name); + rc = ENOENT; + goto fail_dp_tx; + } + if (!sfc_dp_match_hw_fw_caps(&sa->dp_tx->dp, avail_caps)) { + sfc_err(sa, + "Insufficient Hw/FW capabilities to use Tx datapath %s", + tx_name); + rc = EINVAL; + goto fail_dp_tx_caps; + } + } else { + sa->dp_tx = sfc_dp_find_tx_by_caps(&sfc_dp_head, avail_caps); + if (sa->dp_tx == NULL) { + sfc_err(sa, "Tx datapath by caps %#x not found", + avail_caps); + rc = ENOENT; + goto fail_dp_tx; + } + } + + sa->dp_tx_name = sfc_strdup(sa->dp_tx->dp.name); + if (sa->dp_tx_name == NULL) { + rc = ENOMEM; + goto fail_dp_tx_name; + } + + sfc_notice(sa, "use %s Tx datapath", sa->dp_tx_name); + + dev->tx_pkt_burst = sa->dp_tx->pkt_burst; + + dev->dev_ops = &sfc_eth_dev_ops; + + return 0; + +fail_dp_tx_name: +fail_dp_tx_caps: + sa->dp_tx = NULL; + +fail_dp_tx: +fail_kvarg_tx_datapath: + rte_free(sa->dp_rx_name); + sa->dp_rx_name = NULL; + +fail_dp_rx_name: +fail_dp_rx_caps: + sa->dp_rx = NULL; + +fail_dp_rx: +fail_kvarg_rx_datapath: + return rc; +} + +static void +sfc_eth_dev_clear_ops(struct rte_eth_dev *dev) +{ + struct sfc_adapter *sa = dev->data->dev_private; + + dev->dev_ops = NULL; + dev->rx_pkt_burst = NULL; + dev->tx_pkt_burst = NULL; + + rte_free(sa->dp_tx_name); + sa->dp_tx_name = NULL; + sa->dp_tx = NULL; + + rte_free(sa->dp_rx_name); + sa->dp_rx_name = NULL; + sa->dp_rx = NULL; +} + +static const struct eth_dev_ops sfc_eth_dev_secondary_ops = { + .rxq_info_get = sfc_rx_queue_info_get, + .txq_info_get = sfc_tx_queue_info_get, +}; + +static int +sfc_eth_dev_secondary_set_ops(struct rte_eth_dev *dev) +{ + /* + * Device private data has really many process-local pointers. + * Below code should be extremely careful to use data located + * in shared memory only. + */ + struct sfc_adapter *sa = dev->data->dev_private; + const struct sfc_dp_rx *dp_rx; + const struct sfc_dp_tx *dp_tx; + int rc; + + dp_rx = sfc_dp_find_rx_by_name(&sfc_dp_head, sa->dp_rx_name); + if (dp_rx == NULL) { + sfc_err(sa, "cannot find %s Rx datapath", sa->dp_tx_name); + rc = ENOENT; + goto fail_dp_rx; + } + if (~dp_rx->features & SFC_DP_RX_FEAT_MULTI_PROCESS) { + sfc_err(sa, "%s Rx datapath does not support multi-process", + sa->dp_tx_name); + rc = EINVAL; + goto fail_dp_rx_multi_process; + } + + dp_tx = sfc_dp_find_tx_by_name(&sfc_dp_head, sa->dp_tx_name); + if (dp_tx == NULL) { + sfc_err(sa, "cannot find %s Tx datapath", sa->dp_tx_name); + rc = ENOENT; + goto fail_dp_tx; + } + if (~dp_tx->features & SFC_DP_TX_FEAT_MULTI_PROCESS) { + sfc_err(sa, "%s Tx datapath does not support multi-process", + sa->dp_tx_name); + rc = EINVAL; + goto fail_dp_tx_multi_process; + } + + dev->rx_pkt_burst = dp_rx->pkt_burst; + dev->tx_pkt_burst = dp_tx->pkt_burst; + dev->dev_ops = &sfc_eth_dev_secondary_ops; + + return 0; + +fail_dp_tx_multi_process: +fail_dp_tx: +fail_dp_rx_multi_process: +fail_dp_rx: + return rc; +} + +static void +sfc_eth_dev_secondary_clear_ops(struct rte_eth_dev *dev) +{ + dev->dev_ops = NULL; + dev->tx_pkt_burst = NULL; + dev->rx_pkt_burst = NULL; +} + +static void +sfc_register_dp(void) +{ + /* Register once */ + if (TAILQ_EMPTY(&sfc_dp_head)) { + /* Prefer EF10 datapath */ + sfc_dp_register(&sfc_dp_head, &sfc_ef10_rx.dp); + sfc_dp_register(&sfc_dp_head, &sfc_efx_rx.dp); + + sfc_dp_register(&sfc_dp_head, &sfc_ef10_tx.dp); + sfc_dp_register(&sfc_dp_head, &sfc_efx_tx.dp); + sfc_dp_register(&sfc_dp_head, &sfc_ef10_simple_tx.dp); + } +} + static int sfc_eth_dev_init(struct rte_eth_dev *dev) { struct sfc_adapter *sa = dev->data->dev_private; - struct rte_pci_device *pci_dev = SFC_DEV_TO_PCI(dev); + struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); int rc; const efx_nic_cfg_t *encp; const struct ether_addr *from; + sfc_register_dp(); + + if (rte_eal_process_type() != RTE_PROC_PRIMARY) + return -sfc_eth_dev_secondary_set_ops(dev); + /* Required for logging */ + sa->pci_addr = pci_dev->addr; + sa->port_id = dev->data->port_id; + sa->eth_dev = dev; /* Copy PCI device info to the dev->data */ rte_eth_copy_pci_info(dev, pci_dev); + sa->logtype_main = sfc_register_logtype(sa, SFC_LOGTYPE_MAIN_STR, + RTE_LOG_NOTICE); + rc = sfc_kvargs_parse(sa); if (rc != 0) goto fail_kvargs_parse; - rc = sfc_kvargs_process(sa, SFC_KVARG_DEBUG_INIT, - sfc_kvarg_bool_handler, &sa->debug_init); - if (rc != 0) - goto fail_kvarg_debug_init; - sfc_log_init(sa, "entry"); dev->data->mac_addrs = rte_zmalloc("sfc", ETHER_ADDR_LEN, 0); @@ -1163,6 +1959,16 @@ sfc_eth_dev_init(struct rte_eth_dev *dev) sfc_adapter_lock_init(sa); sfc_adapter_lock(sa); + sfc_log_init(sa, "probing"); + rc = sfc_probe(sa); + if (rc != 0) + goto fail_probe; + + sfc_log_init(sa, "set device ops"); + rc = sfc_eth_dev_set_ops(dev); + if (rc != 0) + goto fail_set_ops; + sfc_log_init(sa, "attaching"); rc = sfc_attach(sa); if (rc != 0) @@ -1177,23 +1983,24 @@ sfc_eth_dev_init(struct rte_eth_dev *dev) from = (const struct ether_addr *)(encp->enc_mac_addr); ether_addr_copy(from, &dev->data->mac_addrs[0]); - dev->dev_ops = &sfc_eth_dev_ops; - dev->rx_pkt_burst = &sfc_recv_pkts; - dev->tx_pkt_burst = &sfc_xmit_pkts; - sfc_adapter_unlock(sa); sfc_log_init(sa, "done"); return 0; fail_attach: + sfc_eth_dev_clear_ops(dev); + +fail_set_ops: + sfc_unprobe(sa); + +fail_probe: sfc_adapter_unlock(sa); sfc_adapter_lock_fini(sa); rte_free(dev->data->mac_addrs); dev->data->mac_addrs = NULL; fail_mac_addrs: -fail_kvarg_debug_init: sfc_kvargs_cleanup(sa); fail_kvargs_parse: @@ -1205,21 +2012,26 @@ fail_kvargs_parse: static int sfc_eth_dev_uninit(struct rte_eth_dev *dev) { - struct sfc_adapter *sa = dev->data->dev_private; + struct sfc_adapter *sa; + + if (rte_eal_process_type() != RTE_PROC_PRIMARY) { + sfc_eth_dev_secondary_clear_ops(dev); + return 0; + } + sa = dev->data->dev_private; sfc_log_init(sa, "entry"); sfc_adapter_lock(sa); + sfc_eth_dev_clear_ops(dev); + sfc_detach(sa); + sfc_unprobe(sa); rte_free(dev->data->mac_addrs); dev->data->mac_addrs = NULL; - dev->dev_ops = NULL; - dev->rx_pkt_burst = NULL; - dev->tx_pkt_burst = NULL; - sfc_kvargs_cleanup(sa); sfc_adapter_unlock(sa); @@ -1234,28 +2046,54 @@ sfc_eth_dev_uninit(struct rte_eth_dev *dev) static const struct rte_pci_id pci_id_sfc_efx_map[] = { { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_FARMINGDALE) }, + { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_FARMINGDALE_VF) }, { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_GREENPORT) }, + { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_GREENPORT_VF) }, { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_MEDFORD) }, + { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_MEDFORD_VF) }, + { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_MEDFORD2) }, + { RTE_PCI_DEVICE(EFX_PCI_VENID_SFC, EFX_PCI_DEVID_MEDFORD2_VF) }, { .vendor_id = 0 /* sentinel */ } }; -static struct eth_driver sfc_efx_pmd = { - .pci_drv = { - .id_table = pci_id_sfc_efx_map, - .drv_flags = - RTE_PCI_DRV_INTR_LSC | - RTE_PCI_DRV_NEED_MAPPING, - .probe = rte_eth_dev_pci_probe, - .remove = rte_eth_dev_pci_remove, - }, - .eth_dev_init = sfc_eth_dev_init, - .eth_dev_uninit = sfc_eth_dev_uninit, - .dev_private_size = sizeof(struct sfc_adapter), +static int sfc_eth_dev_pci_probe(struct rte_pci_driver *pci_drv __rte_unused, + struct rte_pci_device *pci_dev) +{ + return rte_eth_dev_pci_generic_probe(pci_dev, + sizeof(struct sfc_adapter), sfc_eth_dev_init); +} + +static int sfc_eth_dev_pci_remove(struct rte_pci_device *pci_dev) +{ + return rte_eth_dev_pci_generic_remove(pci_dev, sfc_eth_dev_uninit); +} + +static struct rte_pci_driver sfc_efx_pmd = { + .id_table = pci_id_sfc_efx_map, + .drv_flags = + RTE_PCI_DRV_INTR_LSC | + RTE_PCI_DRV_NEED_MAPPING, + .probe = sfc_eth_dev_pci_probe, + .remove = sfc_eth_dev_pci_remove, }; -RTE_PMD_REGISTER_PCI(net_sfc_efx, sfc_efx_pmd.pci_drv); +RTE_PMD_REGISTER_PCI(net_sfc_efx, sfc_efx_pmd); RTE_PMD_REGISTER_PCI_TABLE(net_sfc_efx, pci_id_sfc_efx_map); +RTE_PMD_REGISTER_KMOD_DEP(net_sfc_efx, "* igb_uio | uio_pci_generic | vfio-pci"); RTE_PMD_REGISTER_PARAM_STRING(net_sfc_efx, + SFC_KVARG_RX_DATAPATH "=" SFC_KVARG_VALUES_RX_DATAPATH " " + SFC_KVARG_TX_DATAPATH "=" SFC_KVARG_VALUES_TX_DATAPATH " " SFC_KVARG_PERF_PROFILE "=" SFC_KVARG_VALUES_PERF_PROFILE " " - SFC_KVARG_MCDI_LOGGING "=" SFC_KVARG_VALUES_BOOL " " - SFC_KVARG_DEBUG_INIT "=" SFC_KVARG_VALUES_BOOL); + SFC_KVARG_FW_VARIANT "=" SFC_KVARG_VALUES_FW_VARIANT " " + SFC_KVARG_STATS_UPDATE_PERIOD_MS "="); + +RTE_INIT(sfc_driver_register_logtype); +static void +sfc_driver_register_logtype(void) +{ + int ret; + + ret = rte_log_register_type_and_pick_level(SFC_LOGTYPE_PREFIX "driver", + RTE_LOG_NOTICE); + sfc_logtype_driver = (ret < 0) ? RTE_LOGTYPE_PMD : ret; +}