X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fliquidio%2Flio_ethdev.c;h=de234e903bc6c7a4b46c9a49ee1734bc7f814521;hb=6d13ea8e8e49ab957deae2bba5ecf4a4bfe747d1;hp=5d7d5a7d4c67eae46e8be9eab0ebb466b522f09a;hpb=b8b4c621804ca86774ad6d7e64ad30560ec90ed8;p=dpdk.git diff --git a/drivers/net/liquidio/lio_ethdev.c b/drivers/net/liquidio/lio_ethdev.c index 5d7d5a7d4c..de234e903b 100644 --- a/drivers/net/liquidio/lio_ethdev.c +++ b/drivers/net/liquidio/lio_ethdev.c @@ -1,46 +1,1916 @@ -/* - * BSD LICENSE - * - * Copyright(c) 2017 Cavium, Inc.. 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 Cavium, Inc. 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(S) 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) 2017 Cavium, Inc */ -#include +#include +#include +#include #include #include #include +#include #include "lio_logs.h" #include "lio_23xx_vf.h" #include "lio_ethdev.h" #include "lio_rxtx.h" +int lio_logtype_init; +int lio_logtype_driver; + +/* Default RSS key in use */ +static uint8_t lio_rss_key[40] = { + 0x6D, 0x5A, 0x56, 0xDA, 0x25, 0x5B, 0x0E, 0xC2, + 0x41, 0x67, 0x25, 0x3D, 0x43, 0xA3, 0x8F, 0xB0, + 0xD0, 0xCA, 0x2B, 0xCB, 0xAE, 0x7B, 0x30, 0xB4, + 0x77, 0xCB, 0x2D, 0xA3, 0x80, 0x30, 0xF2, 0x0C, + 0x6A, 0x42, 0xB7, 0x3B, 0xBE, 0xAC, 0x01, 0xFA, +}; + +static const struct rte_eth_desc_lim lio_rx_desc_lim = { + .nb_max = CN23XX_MAX_OQ_DESCRIPTORS, + .nb_min = CN23XX_MIN_OQ_DESCRIPTORS, + .nb_align = 1, +}; + +static const struct rte_eth_desc_lim lio_tx_desc_lim = { + .nb_max = CN23XX_MAX_IQ_DESCRIPTORS, + .nb_min = CN23XX_MIN_IQ_DESCRIPTORS, + .nb_align = 1, +}; + +/* Wait for control command to reach nic. */ +static uint16_t +lio_wait_for_ctrl_cmd(struct lio_device *lio_dev, + struct lio_dev_ctrl_cmd *ctrl_cmd) +{ + uint16_t timeout = LIO_MAX_CMD_TIMEOUT; + + while ((ctrl_cmd->cond == 0) && --timeout) { + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + rte_delay_ms(1); + } + + return !timeout; +} + +/** + * \brief Send Rx control command + * @param eth_dev Pointer to the structure rte_eth_dev + * @param start_stop whether to start or stop + */ +static int +lio_send_rx_ctrl_cmd(struct rte_eth_dev *eth_dev, int start_stop) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_RX_CTL; + ctrl_pkt.ncmd.s.param1 = start_stop; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send RX Control message\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "RX Control command timed out\n"); + return -1; + } + + return 0; +} + +/* store statistics names and its offset in stats structure */ +struct rte_lio_xstats_name_off { + char name[RTE_ETH_XSTATS_NAME_SIZE]; + unsigned int offset; +}; + +static const struct rte_lio_xstats_name_off rte_lio_stats_strings[] = { + {"rx_pkts", offsetof(struct octeon_rx_stats, total_rcvd)}, + {"rx_bytes", offsetof(struct octeon_rx_stats, bytes_rcvd)}, + {"rx_broadcast_pkts", offsetof(struct octeon_rx_stats, total_bcst)}, + {"rx_multicast_pkts", offsetof(struct octeon_rx_stats, total_mcst)}, + {"rx_flow_ctrl_pkts", offsetof(struct octeon_rx_stats, ctl_rcvd)}, + {"rx_fifo_err", offsetof(struct octeon_rx_stats, fifo_err)}, + {"rx_dmac_drop", offsetof(struct octeon_rx_stats, dmac_drop)}, + {"rx_fcs_err", offsetof(struct octeon_rx_stats, fcs_err)}, + {"rx_jabber_err", offsetof(struct octeon_rx_stats, jabber_err)}, + {"rx_l2_err", offsetof(struct octeon_rx_stats, l2_err)}, + {"rx_vxlan_pkts", offsetof(struct octeon_rx_stats, fw_rx_vxlan)}, + {"rx_vxlan_err", offsetof(struct octeon_rx_stats, fw_rx_vxlan_err)}, + {"rx_lro_pkts", offsetof(struct octeon_rx_stats, fw_lro_pkts)}, + {"tx_pkts", (offsetof(struct octeon_tx_stats, total_pkts_sent)) + + sizeof(struct octeon_rx_stats)}, + {"tx_bytes", (offsetof(struct octeon_tx_stats, total_bytes_sent)) + + sizeof(struct octeon_rx_stats)}, + {"tx_broadcast_pkts", + (offsetof(struct octeon_tx_stats, bcast_pkts_sent)) + + sizeof(struct octeon_rx_stats)}, + {"tx_multicast_pkts", + (offsetof(struct octeon_tx_stats, mcast_pkts_sent)) + + sizeof(struct octeon_rx_stats)}, + {"tx_flow_ctrl_pkts", (offsetof(struct octeon_tx_stats, ctl_sent)) + + sizeof(struct octeon_rx_stats)}, + {"tx_fifo_err", (offsetof(struct octeon_tx_stats, fifo_err)) + + sizeof(struct octeon_rx_stats)}, + {"tx_total_collisions", (offsetof(struct octeon_tx_stats, + total_collisions)) + + sizeof(struct octeon_rx_stats)}, + {"tx_tso", (offsetof(struct octeon_tx_stats, fw_tso)) + + sizeof(struct octeon_rx_stats)}, + {"tx_vxlan_pkts", (offsetof(struct octeon_tx_stats, fw_tx_vxlan)) + + sizeof(struct octeon_rx_stats)}, +}; + +#define LIO_NB_XSTATS RTE_DIM(rte_lio_stats_strings) + +/* Get hw stats of the port */ +static int +lio_dev_xstats_get(struct rte_eth_dev *eth_dev, struct rte_eth_xstat *xstats, + unsigned int n) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + uint16_t timeout = LIO_MAX_CMD_TIMEOUT; + struct octeon_link_stats *hw_stats; + struct lio_link_stats_resp *resp; + struct lio_soft_command *sc; + uint32_t resp_size; + unsigned int i; + int retval; + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down\n", + lio_dev->port_id); + return -EINVAL; + } + + if (n < LIO_NB_XSTATS) + return LIO_NB_XSTATS; + + resp_size = sizeof(struct lio_link_stats_resp); + sc = lio_alloc_soft_command(lio_dev, 0, resp_size, 0); + if (sc == NULL) + return -ENOMEM; + + resp = (struct lio_link_stats_resp *)sc->virtrptr; + lio_prepare_soft_command(lio_dev, sc, LIO_OPCODE, + LIO_OPCODE_PORT_STATS, 0, 0, 0); + + /* Setting wait time in seconds */ + sc->wait_time = LIO_MAX_CMD_TIMEOUT / 1000; + + retval = lio_send_soft_command(lio_dev, sc); + if (retval == LIO_IQ_SEND_FAILED) { + lio_dev_err(lio_dev, "failed to get port stats from firmware. status: %x\n", + retval); + goto get_stats_fail; + } + + while ((*sc->status_word == LIO_COMPLETION_WORD_INIT) && --timeout) { + lio_flush_iq(lio_dev, lio_dev->instr_queue[sc->iq_no]); + lio_process_ordered_list(lio_dev); + rte_delay_ms(1); + } + + retval = resp->status; + if (retval) { + lio_dev_err(lio_dev, "failed to get port stats from firmware\n"); + goto get_stats_fail; + } + + lio_swap_8B_data((uint64_t *)(&resp->link_stats), + sizeof(struct octeon_link_stats) >> 3); + + hw_stats = &resp->link_stats; + + for (i = 0; i < LIO_NB_XSTATS; i++) { + xstats[i].id = i; + xstats[i].value = + *(uint64_t *)(((char *)hw_stats) + + rte_lio_stats_strings[i].offset); + } + + lio_free_soft_command(sc); + + return LIO_NB_XSTATS; + +get_stats_fail: + lio_free_soft_command(sc); + + return -1; +} + +static int +lio_dev_xstats_get_names(struct rte_eth_dev *eth_dev, + struct rte_eth_xstat_name *xstats_names, + unsigned limit __rte_unused) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + unsigned int i; + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down\n", + lio_dev->port_id); + return -EINVAL; + } + + if (xstats_names == NULL) + return LIO_NB_XSTATS; + + /* Note: limit checked in rte_eth_xstats_names() */ + + for (i = 0; i < LIO_NB_XSTATS; i++) { + snprintf(xstats_names[i].name, sizeof(xstats_names[i].name), + "%s", rte_lio_stats_strings[i].name); + } + + return LIO_NB_XSTATS; +} + +/* Reset hw stats for the port */ +static void +lio_dev_xstats_reset(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down\n", + lio_dev->port_id); + return; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_CLEAR_STATS; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send clear stats command\n"); + return; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "Clear stats command timed out\n"); + return; + } + + /* clear stored per queue stats */ + RTE_FUNC_PTR_OR_RET(*eth_dev->dev_ops->stats_reset); + (*eth_dev->dev_ops->stats_reset)(eth_dev); +} + +/* Retrieve the device statistics (# packets in/out, # bytes in/out, etc */ +static int +lio_dev_stats_get(struct rte_eth_dev *eth_dev, + struct rte_eth_stats *stats) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_droq_stats *oq_stats; + struct lio_iq_stats *iq_stats; + struct lio_instr_queue *txq; + struct lio_droq *droq; + int i, iq_no, oq_no; + uint64_t bytes = 0; + uint64_t pkts = 0; + uint64_t drop = 0; + + for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { + iq_no = lio_dev->linfo.txpciq[i].s.q_no; + txq = lio_dev->instr_queue[iq_no]; + if (txq != NULL) { + iq_stats = &txq->stats; + pkts += iq_stats->tx_done; + drop += iq_stats->tx_dropped; + bytes += iq_stats->tx_tot_bytes; + } + } + + stats->opackets = pkts; + stats->obytes = bytes; + stats->oerrors = drop; + + pkts = 0; + drop = 0; + bytes = 0; + + for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { + oq_no = lio_dev->linfo.rxpciq[i].s.q_no; + droq = lio_dev->droq[oq_no]; + if (droq != NULL) { + oq_stats = &droq->stats; + pkts += oq_stats->rx_pkts_received; + drop += (oq_stats->rx_dropped + + oq_stats->dropped_toomany + + oq_stats->dropped_nomem); + bytes += oq_stats->rx_bytes_received; + } + } + stats->ibytes = bytes; + stats->ipackets = pkts; + stats->ierrors = drop; + + return 0; +} + +static void +lio_dev_stats_reset(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_droq_stats *oq_stats; + struct lio_iq_stats *iq_stats; + struct lio_instr_queue *txq; + struct lio_droq *droq; + int i, iq_no, oq_no; + + for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { + iq_no = lio_dev->linfo.txpciq[i].s.q_no; + txq = lio_dev->instr_queue[iq_no]; + if (txq != NULL) { + iq_stats = &txq->stats; + memset(iq_stats, 0, sizeof(struct lio_iq_stats)); + } + } + + for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { + oq_no = lio_dev->linfo.rxpciq[i].s.q_no; + droq = lio_dev->droq[oq_no]; + if (droq != NULL) { + oq_stats = &droq->stats; + memset(oq_stats, 0, sizeof(struct lio_droq_stats)); + } + } +} + +static void +lio_dev_info_get(struct rte_eth_dev *eth_dev, + struct rte_eth_dev_info *devinfo) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev); + + switch (pci_dev->id.subsystem_device_id) { + /* CN23xx 10G cards */ + case PCI_SUBSYS_DEV_ID_CN2350_210: + case PCI_SUBSYS_DEV_ID_CN2360_210: + case PCI_SUBSYS_DEV_ID_CN2350_210SVPN3: + case PCI_SUBSYS_DEV_ID_CN2360_210SVPN3: + case PCI_SUBSYS_DEV_ID_CN2350_210SVPT: + case PCI_SUBSYS_DEV_ID_CN2360_210SVPT: + devinfo->speed_capa = ETH_LINK_SPEED_10G; + break; + /* CN23xx 25G cards */ + case PCI_SUBSYS_DEV_ID_CN2350_225: + case PCI_SUBSYS_DEV_ID_CN2360_225: + devinfo->speed_capa = ETH_LINK_SPEED_25G; + break; + default: + devinfo->speed_capa = ETH_LINK_SPEED_10G; + lio_dev_err(lio_dev, + "Unknown CN23XX subsystem device id. Setting 10G as default link speed.\n"); + } + + devinfo->max_rx_queues = lio_dev->max_rx_queues; + devinfo->max_tx_queues = lio_dev->max_tx_queues; + + devinfo->min_rx_bufsize = LIO_MIN_RX_BUF_SIZE; + devinfo->max_rx_pktlen = LIO_MAX_RX_PKTLEN; + + devinfo->max_mac_addrs = 1; + + devinfo->rx_offload_capa = (DEV_RX_OFFLOAD_IPV4_CKSUM | + DEV_RX_OFFLOAD_UDP_CKSUM | + DEV_RX_OFFLOAD_TCP_CKSUM | + DEV_RX_OFFLOAD_VLAN_STRIP); + devinfo->tx_offload_capa = (DEV_TX_OFFLOAD_IPV4_CKSUM | + DEV_TX_OFFLOAD_UDP_CKSUM | + DEV_TX_OFFLOAD_TCP_CKSUM | + DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM); + + devinfo->rx_desc_lim = lio_rx_desc_lim; + devinfo->tx_desc_lim = lio_tx_desc_lim; + + devinfo->reta_size = LIO_RSS_MAX_TABLE_SZ; + devinfo->hash_key_size = LIO_RSS_MAX_KEY_SZ; + devinfo->flow_type_rss_offloads = (ETH_RSS_IPV4 | + ETH_RSS_NONFRAG_IPV4_TCP | + ETH_RSS_IPV6 | + ETH_RSS_NONFRAG_IPV6_TCP | + ETH_RSS_IPV6_EX | + ETH_RSS_IPV6_TCP_EX); +} + +static int +lio_dev_mtu_set(struct rte_eth_dev *eth_dev, uint16_t mtu) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + uint16_t pf_mtu = lio_dev->linfo.link.s.mtu; + uint32_t frame_len = mtu + ETHER_HDR_LEN + ETHER_CRC_LEN; + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + PMD_INIT_FUNC_TRACE(); + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't set MTU\n", + lio_dev->port_id); + return -EINVAL; + } + + /* check if VF MTU is within allowed range. + * New value should not exceed PF MTU. + */ + if ((mtu < ETHER_MIN_MTU) || (mtu > pf_mtu)) { + lio_dev_err(lio_dev, "VF MTU should be >= %d and <= %d\n", + ETHER_MIN_MTU, pf_mtu); + return -EINVAL; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_CHANGE_MTU; + ctrl_pkt.ncmd.s.param1 = mtu; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send command to change MTU\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "Command to change MTU timed out\n"); + return -1; + } + + if (frame_len > ETHER_MAX_LEN) + eth_dev->data->dev_conf.rxmode.offloads |= + DEV_RX_OFFLOAD_JUMBO_FRAME; + else + eth_dev->data->dev_conf.rxmode.offloads &= + ~DEV_RX_OFFLOAD_JUMBO_FRAME; + + eth_dev->data->dev_conf.rxmode.max_rx_pkt_len = frame_len; + eth_dev->data->mtu = mtu; + + return 0; +} + +static int +lio_dev_rss_reta_update(struct rte_eth_dev *eth_dev, + struct rte_eth_rss_reta_entry64 *reta_conf, + uint16_t reta_size) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_rss_ctx *rss_state = &lio_dev->rss_state; + struct lio_rss_set *rss_param; + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + int i, j, index; + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't update reta\n", + lio_dev->port_id); + return -EINVAL; + } + + if (reta_size != LIO_RSS_MAX_TABLE_SZ) { + lio_dev_err(lio_dev, + "The size of hash lookup table configured (%d) doesn't match the number hardware can supported (%d)\n", + reta_size, LIO_RSS_MAX_TABLE_SZ); + return -EINVAL; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + rss_param = (struct lio_rss_set *)&ctrl_pkt.udd[0]; + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_SET_RSS; + ctrl_pkt.ncmd.s.more = sizeof(struct lio_rss_set) >> 3; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + rss_param->param.flags = 0xF; + rss_param->param.flags &= ~LIO_RSS_PARAM_ITABLE_UNCHANGED; + rss_param->param.itablesize = LIO_RSS_MAX_TABLE_SZ; + + for (i = 0; i < (reta_size / RTE_RETA_GROUP_SIZE); i++) { + for (j = 0; j < RTE_RETA_GROUP_SIZE; j++) { + if ((reta_conf[i].mask) & ((uint64_t)1 << j)) { + index = (i * RTE_RETA_GROUP_SIZE) + j; + rss_state->itable[index] = reta_conf[i].reta[j]; + } + } + } + + rss_state->itable_size = LIO_RSS_MAX_TABLE_SZ; + memcpy(rss_param->itable, rss_state->itable, rss_state->itable_size); + + lio_swap_8B_data((uint64_t *)rss_param, LIO_RSS_PARAM_SIZE >> 3); + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to set rss hash\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "Set rss hash timed out\n"); + return -1; + } + + return 0; +} + +static int +lio_dev_rss_reta_query(struct rte_eth_dev *eth_dev, + struct rte_eth_rss_reta_entry64 *reta_conf, + uint16_t reta_size) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_rss_ctx *rss_state = &lio_dev->rss_state; + int i, num; + + if (reta_size != LIO_RSS_MAX_TABLE_SZ) { + lio_dev_err(lio_dev, + "The size of hash lookup table configured (%d) doesn't match the number hardware can supported (%d)\n", + reta_size, LIO_RSS_MAX_TABLE_SZ); + return -EINVAL; + } + + num = reta_size / RTE_RETA_GROUP_SIZE; + + for (i = 0; i < num; i++) { + memcpy(reta_conf->reta, + &rss_state->itable[i * RTE_RETA_GROUP_SIZE], + RTE_RETA_GROUP_SIZE); + reta_conf++; + } + + return 0; +} + +static int +lio_dev_rss_hash_conf_get(struct rte_eth_dev *eth_dev, + struct rte_eth_rss_conf *rss_conf) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_rss_ctx *rss_state = &lio_dev->rss_state; + uint8_t *hash_key = NULL; + uint64_t rss_hf = 0; + + if (rss_state->hash_disable) { + lio_dev_info(lio_dev, "RSS disabled in nic\n"); + rss_conf->rss_hf = 0; + return 0; + } + + /* Get key value */ + hash_key = rss_conf->rss_key; + if (hash_key != NULL) + memcpy(hash_key, rss_state->hash_key, rss_state->hash_key_size); + + if (rss_state->ip) + rss_hf |= ETH_RSS_IPV4; + if (rss_state->tcp_hash) + rss_hf |= ETH_RSS_NONFRAG_IPV4_TCP; + if (rss_state->ipv6) + rss_hf |= ETH_RSS_IPV6; + if (rss_state->ipv6_tcp_hash) + rss_hf |= ETH_RSS_NONFRAG_IPV6_TCP; + if (rss_state->ipv6_ex) + rss_hf |= ETH_RSS_IPV6_EX; + if (rss_state->ipv6_tcp_ex_hash) + rss_hf |= ETH_RSS_IPV6_TCP_EX; + + rss_conf->rss_hf = rss_hf; + + return 0; +} + +static int +lio_dev_rss_hash_update(struct rte_eth_dev *eth_dev, + struct rte_eth_rss_conf *rss_conf) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_rss_ctx *rss_state = &lio_dev->rss_state; + struct lio_rss_set *rss_param; + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't update hash\n", + lio_dev->port_id); + return -EINVAL; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + rss_param = (struct lio_rss_set *)&ctrl_pkt.udd[0]; + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_SET_RSS; + ctrl_pkt.ncmd.s.more = sizeof(struct lio_rss_set) >> 3; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + rss_param->param.flags = 0xF; + + if (rss_conf->rss_key) { + rss_param->param.flags &= ~LIO_RSS_PARAM_HASH_KEY_UNCHANGED; + rss_state->hash_key_size = LIO_RSS_MAX_KEY_SZ; + rss_param->param.hashkeysize = LIO_RSS_MAX_KEY_SZ; + memcpy(rss_state->hash_key, rss_conf->rss_key, + rss_state->hash_key_size); + memcpy(rss_param->key, rss_state->hash_key, + rss_state->hash_key_size); + } + + if ((rss_conf->rss_hf & LIO_RSS_OFFLOAD_ALL) == 0) { + /* Can't disable rss through hash flags, + * if it is enabled by default during init + */ + if (!rss_state->hash_disable) + return -EINVAL; + + /* This is for --disable-rss during testpmd launch */ + rss_param->param.flags |= LIO_RSS_PARAM_DISABLE_RSS; + } else { + uint32_t hashinfo = 0; + + /* Can't enable rss if disabled by default during init */ + if (rss_state->hash_disable) + return -EINVAL; + + if (rss_conf->rss_hf & ETH_RSS_IPV4) { + hashinfo |= LIO_RSS_HASH_IPV4; + rss_state->ip = 1; + } else { + rss_state->ip = 0; + } + + if (rss_conf->rss_hf & ETH_RSS_NONFRAG_IPV4_TCP) { + hashinfo |= LIO_RSS_HASH_TCP_IPV4; + rss_state->tcp_hash = 1; + } else { + rss_state->tcp_hash = 0; + } + + if (rss_conf->rss_hf & ETH_RSS_IPV6) { + hashinfo |= LIO_RSS_HASH_IPV6; + rss_state->ipv6 = 1; + } else { + rss_state->ipv6 = 0; + } + + if (rss_conf->rss_hf & ETH_RSS_NONFRAG_IPV6_TCP) { + hashinfo |= LIO_RSS_HASH_TCP_IPV6; + rss_state->ipv6_tcp_hash = 1; + } else { + rss_state->ipv6_tcp_hash = 0; + } + + if (rss_conf->rss_hf & ETH_RSS_IPV6_EX) { + hashinfo |= LIO_RSS_HASH_IPV6_EX; + rss_state->ipv6_ex = 1; + } else { + rss_state->ipv6_ex = 0; + } + + if (rss_conf->rss_hf & ETH_RSS_IPV6_TCP_EX) { + hashinfo |= LIO_RSS_HASH_TCP_IPV6_EX; + rss_state->ipv6_tcp_ex_hash = 1; + } else { + rss_state->ipv6_tcp_ex_hash = 0; + } + + rss_param->param.flags &= ~LIO_RSS_PARAM_HASH_INFO_UNCHANGED; + rss_param->param.hashinfo = hashinfo; + } + + lio_swap_8B_data((uint64_t *)rss_param, LIO_RSS_PARAM_SIZE >> 3); + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to set rss hash\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "Set rss hash timed out\n"); + return -1; + } + + return 0; +} + +/** + * Add vxlan dest udp port for an interface. + * + * @param eth_dev + * Pointer to the structure rte_eth_dev + * @param udp_tnl + * udp tunnel conf + * + * @return + * On success return 0 + * On failure return -1 + */ +static int +lio_dev_udp_tunnel_add(struct rte_eth_dev *eth_dev, + struct rte_eth_udp_tunnel *udp_tnl) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + if (udp_tnl == NULL) + return -EINVAL; + + if (udp_tnl->prot_type != RTE_TUNNEL_TYPE_VXLAN) { + lio_dev_err(lio_dev, "Unsupported tunnel type\n"); + return -1; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_VXLAN_PORT_CONFIG; + ctrl_pkt.ncmd.s.param1 = udp_tnl->udp_port; + ctrl_pkt.ncmd.s.more = LIO_CMD_VXLAN_PORT_ADD; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send VXLAN_PORT_ADD command\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "VXLAN_PORT_ADD command timed out\n"); + return -1; + } + + return 0; +} + +/** + * Remove vxlan dest udp port for an interface. + * + * @param eth_dev + * Pointer to the structure rte_eth_dev + * @param udp_tnl + * udp tunnel conf + * + * @return + * On success return 0 + * On failure return -1 + */ +static int +lio_dev_udp_tunnel_del(struct rte_eth_dev *eth_dev, + struct rte_eth_udp_tunnel *udp_tnl) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + if (udp_tnl == NULL) + return -EINVAL; + + if (udp_tnl->prot_type != RTE_TUNNEL_TYPE_VXLAN) { + lio_dev_err(lio_dev, "Unsupported tunnel type\n"); + return -1; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_VXLAN_PORT_CONFIG; + ctrl_pkt.ncmd.s.param1 = udp_tnl->udp_port; + ctrl_pkt.ncmd.s.more = LIO_CMD_VXLAN_PORT_DEL; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send VXLAN_PORT_DEL command\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "VXLAN_PORT_DEL command timed out\n"); + return -1; + } + + return 0; +} + +static int +lio_dev_vlan_filter_set(struct rte_eth_dev *eth_dev, uint16_t vlan_id, int on) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + if (lio_dev->linfo.vlan_is_admin_assigned) + return -EPERM; + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = on ? + LIO_CMD_ADD_VLAN_FILTER : LIO_CMD_DEL_VLAN_FILTER; + ctrl_pkt.ncmd.s.param1 = vlan_id; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to %s VLAN port\n", + on ? "add" : "remove"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "Command to %s VLAN port timed out\n", + on ? "add" : "remove"); + return -1; + } + + return 0; +} + +static uint64_t +lio_hweight64(uint64_t w) +{ + uint64_t res = w - ((w >> 1) & 0x5555555555555555ul); + + res = + (res & 0x3333333333333333ul) + ((res >> 2) & 0x3333333333333333ul); + res = (res + (res >> 4)) & 0x0F0F0F0F0F0F0F0Ful; + res = res + (res >> 8); + res = res + (res >> 16); + + return (res + (res >> 32)) & 0x00000000000000FFul; +} + +static int +lio_dev_link_update(struct rte_eth_dev *eth_dev, + int wait_to_complete __rte_unused) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct rte_eth_link link; + + /* Initialize */ + memset(&link, 0, sizeof(link)); + link.link_status = ETH_LINK_DOWN; + link.link_speed = ETH_SPEED_NUM_NONE; + link.link_duplex = ETH_LINK_HALF_DUPLEX; + link.link_autoneg = ETH_LINK_AUTONEG; + + /* Return what we found */ + if (lio_dev->linfo.link.s.link_up == 0) { + /* Interface is down */ + return rte_eth_linkstatus_set(eth_dev, &link); + } + + link.link_status = ETH_LINK_UP; /* Interface is up */ + link.link_duplex = ETH_LINK_FULL_DUPLEX; + switch (lio_dev->linfo.link.s.speed) { + case LIO_LINK_SPEED_10000: + link.link_speed = ETH_SPEED_NUM_10G; + break; + case LIO_LINK_SPEED_25000: + link.link_speed = ETH_SPEED_NUM_25G; + break; + default: + link.link_speed = ETH_SPEED_NUM_NONE; + link.link_duplex = ETH_LINK_HALF_DUPLEX; + } + + return rte_eth_linkstatus_set(eth_dev, &link); +} + +/** + * \brief Net device enable, disable allmulticast + * @param eth_dev Pointer to the structure rte_eth_dev + */ +static void +lio_change_dev_flag(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + /* Create a ctrl pkt command to be sent to core app. */ + ctrl_pkt.ncmd.s.cmd = LIO_CMD_CHANGE_DEVFLAGS; + ctrl_pkt.ncmd.s.param1 = lio_dev->ifflags; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send change flag message\n"); + return; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) + lio_dev_err(lio_dev, "Change dev flag command timed out\n"); +} + +static void +lio_dev_promiscuous_enable(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (strcmp(lio_dev->firmware_version, LIO_VF_TRUST_MIN_VERSION) < 0) { + lio_dev_err(lio_dev, "Require firmware version >= %s\n", + LIO_VF_TRUST_MIN_VERSION); + return; + } + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't enable promiscuous\n", + lio_dev->port_id); + return; + } + + lio_dev->ifflags |= LIO_IFFLAG_PROMISC; + lio_change_dev_flag(eth_dev); +} + +static void +lio_dev_promiscuous_disable(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (strcmp(lio_dev->firmware_version, LIO_VF_TRUST_MIN_VERSION) < 0) { + lio_dev_err(lio_dev, "Require firmware version >= %s\n", + LIO_VF_TRUST_MIN_VERSION); + return; + } + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't disable promiscuous\n", + lio_dev->port_id); + return; + } + + lio_dev->ifflags &= ~LIO_IFFLAG_PROMISC; + lio_change_dev_flag(eth_dev); +} + +static void +lio_dev_allmulticast_enable(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't enable multicast\n", + lio_dev->port_id); + return; + } + + lio_dev->ifflags |= LIO_IFFLAG_ALLMULTI; + lio_change_dev_flag(eth_dev); +} + +static void +lio_dev_allmulticast_disable(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (!lio_dev->intf_open) { + lio_dev_err(lio_dev, "Port %d down, can't disable multicast\n", + lio_dev->port_id); + return; + } + + lio_dev->ifflags &= ~LIO_IFFLAG_ALLMULTI; + lio_change_dev_flag(eth_dev); +} + +static void +lio_dev_rss_configure(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_rss_ctx *rss_state = &lio_dev->rss_state; + struct rte_eth_rss_reta_entry64 reta_conf[8]; + struct rte_eth_rss_conf rss_conf; + uint16_t i; + + /* Configure the RSS key and the RSS protocols used to compute + * the RSS hash of input packets. + */ + rss_conf = eth_dev->data->dev_conf.rx_adv_conf.rss_conf; + if ((rss_conf.rss_hf & LIO_RSS_OFFLOAD_ALL) == 0) { + rss_state->hash_disable = 1; + lio_dev_rss_hash_update(eth_dev, &rss_conf); + return; + } + + if (rss_conf.rss_key == NULL) + rss_conf.rss_key = lio_rss_key; /* Default hash key */ + + lio_dev_rss_hash_update(eth_dev, &rss_conf); + + memset(reta_conf, 0, sizeof(reta_conf)); + for (i = 0; i < LIO_RSS_MAX_TABLE_SZ; i++) { + uint8_t q_idx, conf_idx, reta_idx; + + q_idx = (uint8_t)((eth_dev->data->nb_rx_queues > 1) ? + i % eth_dev->data->nb_rx_queues : 0); + conf_idx = i / RTE_RETA_GROUP_SIZE; + reta_idx = i % RTE_RETA_GROUP_SIZE; + reta_conf[conf_idx].reta[reta_idx] = q_idx; + reta_conf[conf_idx].mask |= ((uint64_t)1 << reta_idx); + } + + lio_dev_rss_reta_update(eth_dev, reta_conf, LIO_RSS_MAX_TABLE_SZ); +} + +static void +lio_dev_mq_rx_configure(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_rss_ctx *rss_state = &lio_dev->rss_state; + struct rte_eth_rss_conf rss_conf; + + switch (eth_dev->data->dev_conf.rxmode.mq_mode) { + case ETH_MQ_RX_RSS: + lio_dev_rss_configure(eth_dev); + break; + case ETH_MQ_RX_NONE: + /* if mq_mode is none, disable rss mode. */ + default: + memset(&rss_conf, 0, sizeof(rss_conf)); + rss_state->hash_disable = 1; + lio_dev_rss_hash_update(eth_dev, &rss_conf); + } +} + +/** + * Setup our receive queue/ringbuffer. This is the + * queue the Octeon uses to send us packets and + * responses. We are given a memory pool for our + * packet buffers that are used to populate the receive + * queue. + * + * @param eth_dev + * Pointer to the structure rte_eth_dev + * @param q_no + * Queue number + * @param num_rx_descs + * Number of entries in the queue + * @param socket_id + * Where to allocate memory + * @param rx_conf + * Pointer to the struction rte_eth_rxconf + * @param mp + * Pointer to the packet pool + * + * @return + * - On success, return 0 + * - On failure, return -1 + */ +static int +lio_dev_rx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t q_no, + uint16_t num_rx_descs, unsigned int socket_id, + const struct rte_eth_rxconf *rx_conf __rte_unused, + struct rte_mempool *mp) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct rte_pktmbuf_pool_private *mbp_priv; + uint32_t fw_mapped_oq; + uint16_t buf_size; + + if (q_no >= lio_dev->nb_rx_queues) { + lio_dev_err(lio_dev, "Invalid rx queue number %u\n", q_no); + return -EINVAL; + } + + lio_dev_dbg(lio_dev, "setting up rx queue %u\n", q_no); + + fw_mapped_oq = lio_dev->linfo.rxpciq[q_no].s.q_no; + + /* Free previous allocation if any */ + if (eth_dev->data->rx_queues[q_no] != NULL) { + lio_dev_rx_queue_release(eth_dev->data->rx_queues[q_no]); + eth_dev->data->rx_queues[q_no] = NULL; + } + + mbp_priv = rte_mempool_get_priv(mp); + buf_size = mbp_priv->mbuf_data_room_size - RTE_PKTMBUF_HEADROOM; + + if (lio_setup_droq(lio_dev, fw_mapped_oq, num_rx_descs, buf_size, mp, + socket_id)) { + lio_dev_err(lio_dev, "droq allocation failed\n"); + return -1; + } + + eth_dev->data->rx_queues[q_no] = lio_dev->droq[fw_mapped_oq]; + + return 0; +} + +/** + * Release the receive queue/ringbuffer. Called by + * the upper layers. + * + * @param rxq + * Opaque pointer to the receive queue to release + * + * @return + * - nothing + */ +void +lio_dev_rx_queue_release(void *rxq) +{ + struct lio_droq *droq = rxq; + int oq_no; + + if (droq) { + oq_no = droq->q_no; + lio_delete_droq_queue(droq->lio_dev, oq_no); + } +} + +/** + * Allocate and initialize SW ring. Initialize associated HW registers. + * + * @param eth_dev + * Pointer to structure rte_eth_dev + * + * @param q_no + * Queue number + * + * @param num_tx_descs + * Number of ringbuffer descriptors + * + * @param socket_id + * NUMA socket id, used for memory allocations + * + * @param tx_conf + * Pointer to the structure rte_eth_txconf + * + * @return + * - On success, return 0 + * - On failure, return -errno value + */ +static int +lio_dev_tx_queue_setup(struct rte_eth_dev *eth_dev, uint16_t q_no, + uint16_t num_tx_descs, unsigned int socket_id, + const struct rte_eth_txconf *tx_conf __rte_unused) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + int fw_mapped_iq = lio_dev->linfo.txpciq[q_no].s.q_no; + int retval; + + if (q_no >= lio_dev->nb_tx_queues) { + lio_dev_err(lio_dev, "Invalid tx queue number %u\n", q_no); + return -EINVAL; + } + + lio_dev_dbg(lio_dev, "setting up tx queue %u\n", q_no); + + /* Free previous allocation if any */ + if (eth_dev->data->tx_queues[q_no] != NULL) { + lio_dev_tx_queue_release(eth_dev->data->tx_queues[q_no]); + eth_dev->data->tx_queues[q_no] = NULL; + } + + retval = lio_setup_iq(lio_dev, q_no, lio_dev->linfo.txpciq[q_no], + num_tx_descs, lio_dev, socket_id); + + if (retval) { + lio_dev_err(lio_dev, "Runtime IQ(TxQ) creation failed.\n"); + return retval; + } + + retval = lio_setup_sglists(lio_dev, q_no, fw_mapped_iq, + lio_dev->instr_queue[fw_mapped_iq]->nb_desc, + socket_id); + + if (retval) { + lio_delete_instruction_queue(lio_dev, fw_mapped_iq); + return retval; + } + + eth_dev->data->tx_queues[q_no] = lio_dev->instr_queue[fw_mapped_iq]; + + return 0; +} + +/** + * Release the transmit queue/ringbuffer. Called by + * the upper layers. + * + * @param txq + * Opaque pointer to the transmit queue to release + * + * @return + * - nothing + */ +void +lio_dev_tx_queue_release(void *txq) +{ + struct lio_instr_queue *tq = txq; + uint32_t fw_mapped_iq_no; + + + if (tq) { + /* Free sg_list */ + lio_delete_sglist(tq); + + fw_mapped_iq_no = tq->txpciq.s.q_no; + lio_delete_instruction_queue(tq->lio_dev, fw_mapped_iq_no); + } +} + +/** + * Api to check link state. + */ +static void +lio_dev_get_link_status(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + uint16_t timeout = LIO_MAX_CMD_TIMEOUT; + struct lio_link_status_resp *resp; + union octeon_link_status *ls; + struct lio_soft_command *sc; + uint32_t resp_size; + + if (!lio_dev->intf_open) + return; + + resp_size = sizeof(struct lio_link_status_resp); + sc = lio_alloc_soft_command(lio_dev, 0, resp_size, 0); + if (sc == NULL) + return; + + resp = (struct lio_link_status_resp *)sc->virtrptr; + lio_prepare_soft_command(lio_dev, sc, LIO_OPCODE, + LIO_OPCODE_INFO, 0, 0, 0); + + /* Setting wait time in seconds */ + sc->wait_time = LIO_MAX_CMD_TIMEOUT / 1000; + + if (lio_send_soft_command(lio_dev, sc) == LIO_IQ_SEND_FAILED) + goto get_status_fail; + + while ((*sc->status_word == LIO_COMPLETION_WORD_INIT) && --timeout) { + lio_flush_iq(lio_dev, lio_dev->instr_queue[sc->iq_no]); + rte_delay_ms(1); + } + + if (resp->status) + goto get_status_fail; + + ls = &resp->link_info.link; + + lio_swap_8B_data((uint64_t *)ls, sizeof(union octeon_link_status) >> 3); + + if (lio_dev->linfo.link.link_status64 != ls->link_status64) { + if (ls->s.mtu < eth_dev->data->mtu) { + lio_dev_info(lio_dev, "Lowered VF MTU to %d as PF MTU dropped\n", + ls->s.mtu); + eth_dev->data->mtu = ls->s.mtu; + } + lio_dev->linfo.link.link_status64 = ls->link_status64; + lio_dev_link_update(eth_dev, 0); + } + + lio_free_soft_command(sc); + + return; + +get_status_fail: + lio_free_soft_command(sc); +} + +/* This function will be invoked every LSC_TIMEOUT ns (100ms) + * and will update link state if it changes. + */ +static void +lio_sync_link_state_check(void *eth_dev) +{ + struct lio_device *lio_dev = + (((struct rte_eth_dev *)eth_dev)->data->dev_private); + + if (lio_dev->port_configured) + lio_dev_get_link_status(eth_dev); + + /* Schedule periodic link status check. + * Stop check if interface is close and start again while opening. + */ + if (lio_dev->intf_open) + rte_eal_alarm_set(LIO_LSC_TIMEOUT, lio_sync_link_state_check, + eth_dev); +} + +static int +lio_dev_start(struct rte_eth_dev *eth_dev) +{ + uint16_t mtu; + uint32_t frame_len = eth_dev->data->dev_conf.rxmode.max_rx_pkt_len; + struct lio_device *lio_dev = LIO_DEV(eth_dev); + uint16_t timeout = LIO_MAX_CMD_TIMEOUT; + int ret = 0; + + lio_dev_info(lio_dev, "Starting port %d\n", eth_dev->data->port_id); + + if (lio_dev->fn_list.enable_io_queues(lio_dev)) + return -1; + + if (lio_send_rx_ctrl_cmd(eth_dev, 1)) + return -1; + + /* Ready for link status updates */ + lio_dev->intf_open = 1; + rte_mb(); + + /* Configure RSS if device configured with multiple RX queues. */ + lio_dev_mq_rx_configure(eth_dev); + + /* Before update the link info, + * must set linfo.link.link_status64 to 0. + */ + lio_dev->linfo.link.link_status64 = 0; + + /* start polling for lsc */ + ret = rte_eal_alarm_set(LIO_LSC_TIMEOUT, + lio_sync_link_state_check, + eth_dev); + if (ret) { + lio_dev_err(lio_dev, + "link state check handler creation failed\n"); + goto dev_lsc_handle_error; + } + + while ((lio_dev->linfo.link.link_status64 == 0) && (--timeout)) + rte_delay_ms(1); + + if (lio_dev->linfo.link.link_status64 == 0) { + ret = -1; + goto dev_mtu_set_error; + } + + mtu = (uint16_t)(frame_len - ETHER_HDR_LEN - ETHER_CRC_LEN); + if (mtu < ETHER_MIN_MTU) + mtu = ETHER_MIN_MTU; + + if (eth_dev->data->mtu != mtu) { + ret = lio_dev_mtu_set(eth_dev, mtu); + if (ret) + goto dev_mtu_set_error; + } + + return 0; + +dev_mtu_set_error: + rte_eal_alarm_cancel(lio_sync_link_state_check, eth_dev); + +dev_lsc_handle_error: + lio_dev->intf_open = 0; + lio_send_rx_ctrl_cmd(eth_dev, 0); + + return ret; +} + +/* Stop device and disable input/output functions */ +static void +lio_dev_stop(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + lio_dev_info(lio_dev, "Stopping port %d\n", eth_dev->data->port_id); + lio_dev->intf_open = 0; + rte_mb(); + + /* Cancel callback if still running. */ + rte_eal_alarm_cancel(lio_sync_link_state_check, eth_dev); + + lio_send_rx_ctrl_cmd(eth_dev, 0); + + lio_wait_for_instr_fetch(lio_dev); + + /* Clear recorded link status */ + lio_dev->linfo.link.link_status64 = 0; +} + +static int +lio_dev_set_link_up(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (!lio_dev->intf_open) { + lio_dev_info(lio_dev, "Port is stopped, Start the port first\n"); + return 0; + } + + if (lio_dev->linfo.link.s.link_up) { + lio_dev_info(lio_dev, "Link is already UP\n"); + return 0; + } + + if (lio_send_rx_ctrl_cmd(eth_dev, 1)) { + lio_dev_err(lio_dev, "Unable to set Link UP\n"); + return -1; + } + + lio_dev->linfo.link.s.link_up = 1; + eth_dev->data->dev_link.link_status = ETH_LINK_UP; + + return 0; +} + +static int +lio_dev_set_link_down(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (!lio_dev->intf_open) { + lio_dev_info(lio_dev, "Port is stopped, Start the port first\n"); + return 0; + } + + if (!lio_dev->linfo.link.s.link_up) { + lio_dev_info(lio_dev, "Link is already DOWN\n"); + return 0; + } + + lio_dev->linfo.link.s.link_up = 0; + eth_dev->data->dev_link.link_status = ETH_LINK_DOWN; + + if (lio_send_rx_ctrl_cmd(eth_dev, 0)) { + lio_dev->linfo.link.s.link_up = 1; + eth_dev->data->dev_link.link_status = ETH_LINK_UP; + lio_dev_err(lio_dev, "Unable to set Link Down\n"); + return -1; + } + + return 0; +} + +/** + * Reset and stop the device. This occurs on the first + * call to this routine. Subsequent calls will simply + * return. NB: This will require the NIC to be rebooted. + * + * @param eth_dev + * Pointer to the structure rte_eth_dev + * + * @return + * - nothing + */ +static void +lio_dev_close(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + lio_dev_info(lio_dev, "closing port %d\n", eth_dev->data->port_id); + + if (lio_dev->intf_open) + lio_dev_stop(eth_dev); + + /* Reset ioq regs */ + lio_dev->fn_list.setup_device_regs(lio_dev); + + if (lio_dev->pci_dev->kdrv == RTE_KDRV_IGB_UIO) { + cn23xx_vf_ask_pf_to_do_flr(lio_dev); + rte_delay_ms(LIO_PCI_FLR_WAIT); + } + + /* lio_free_mbox */ + lio_dev->fn_list.free_mbox(lio_dev); + + /* Free glist resources */ + rte_free(lio_dev->glist_head); + rte_free(lio_dev->glist_lock); + lio_dev->glist_head = NULL; + lio_dev->glist_lock = NULL; + + lio_dev->port_configured = 0; + + /* Delete all queues */ + lio_dev_clear_queues(eth_dev); +} + +/** + * Enable tunnel rx checksum verification from firmware. + */ +static void +lio_enable_hw_tunnel_rx_checksum(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_TNL_RX_CSUM_CTL; + ctrl_pkt.ncmd.s.param1 = LIO_CMD_RXCSUM_ENABLE; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send TNL_RX_CSUM command\n"); + return; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) + lio_dev_err(lio_dev, "TNL_RX_CSUM command timed out\n"); +} + +/** + * Enable checksum calculation for inner packet in a tunnel. + */ +static void +lio_enable_hw_tunnel_tx_checksum(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_TNL_TX_CSUM_CTL; + ctrl_pkt.ncmd.s.param1 = LIO_CMD_TXCSUM_ENABLE; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send TNL_TX_CSUM command\n"); + return; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) + lio_dev_err(lio_dev, "TNL_TX_CSUM command timed out\n"); +} + +static int +lio_send_queue_count_update(struct rte_eth_dev *eth_dev, int num_txq, + int num_rxq) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + struct lio_dev_ctrl_cmd ctrl_cmd; + struct lio_ctrl_pkt ctrl_pkt; + + if (strcmp(lio_dev->firmware_version, LIO_Q_RECONF_MIN_VERSION) < 0) { + lio_dev_err(lio_dev, "Require firmware version >= %s\n", + LIO_Q_RECONF_MIN_VERSION); + return -ENOTSUP; + } + + /* flush added to prevent cmd failure + * incase the queue is full + */ + lio_flush_iq(lio_dev, lio_dev->instr_queue[0]); + + memset(&ctrl_pkt, 0, sizeof(struct lio_ctrl_pkt)); + memset(&ctrl_cmd, 0, sizeof(struct lio_dev_ctrl_cmd)); + + ctrl_cmd.eth_dev = eth_dev; + ctrl_cmd.cond = 0; + + ctrl_pkt.ncmd.s.cmd = LIO_CMD_QUEUE_COUNT_CTL; + ctrl_pkt.ncmd.s.param1 = num_txq; + ctrl_pkt.ncmd.s.param2 = num_rxq; + ctrl_pkt.ctrl_cmd = &ctrl_cmd; + + if (lio_send_ctrl_pkt(lio_dev, &ctrl_pkt)) { + lio_dev_err(lio_dev, "Failed to send queue count control command\n"); + return -1; + } + + if (lio_wait_for_ctrl_cmd(lio_dev, &ctrl_cmd)) { + lio_dev_err(lio_dev, "Queue count control command timed out\n"); + return -1; + } + + return 0; +} + +static int +lio_reconf_queues(struct rte_eth_dev *eth_dev, int num_txq, int num_rxq) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + + if (lio_dev->nb_rx_queues != num_rxq || + lio_dev->nb_tx_queues != num_txq) { + if (lio_send_queue_count_update(eth_dev, num_txq, num_rxq)) + return -1; + lio_dev->nb_rx_queues = num_rxq; + lio_dev->nb_tx_queues = num_txq; + } + + if (lio_dev->intf_open) + lio_dev_stop(eth_dev); + + /* Reset ioq registers */ + if (lio_dev->fn_list.setup_device_regs(lio_dev)) { + lio_dev_err(lio_dev, "Failed to configure device registers\n"); + return -1; + } + + return 0; +} + +static int +lio_dev_configure(struct rte_eth_dev *eth_dev) +{ + struct lio_device *lio_dev = LIO_DEV(eth_dev); + uint16_t timeout = LIO_MAX_CMD_TIMEOUT; + int retval, num_iqueues, num_oqueues; + uint8_t mac[ETHER_ADDR_LEN], i; + struct lio_if_cfg_resp *resp; + struct lio_soft_command *sc; + union lio_if_cfg if_cfg; + uint32_t resp_size; + + PMD_INIT_FUNC_TRACE(); + + /* Inform firmware about change in number of queues to use. + * Disable IO queues and reset registers for re-configuration. + */ + if (lio_dev->port_configured) + return lio_reconf_queues(eth_dev, + eth_dev->data->nb_tx_queues, + eth_dev->data->nb_rx_queues); + + lio_dev->nb_rx_queues = eth_dev->data->nb_rx_queues; + lio_dev->nb_tx_queues = eth_dev->data->nb_tx_queues; + + /* Set max number of queues which can be re-configured. */ + lio_dev->max_rx_queues = eth_dev->data->nb_rx_queues; + lio_dev->max_tx_queues = eth_dev->data->nb_tx_queues; + + resp_size = sizeof(struct lio_if_cfg_resp); + sc = lio_alloc_soft_command(lio_dev, 0, resp_size, 0); + if (sc == NULL) + return -ENOMEM; + + resp = (struct lio_if_cfg_resp *)sc->virtrptr; + + /* Firmware doesn't have capability to reconfigure the queues, + * Claim all queues, and use as many required + */ + if_cfg.if_cfg64 = 0; + if_cfg.s.num_iqueues = lio_dev->nb_tx_queues; + if_cfg.s.num_oqueues = lio_dev->nb_rx_queues; + if_cfg.s.base_queue = 0; + + if_cfg.s.gmx_port_id = lio_dev->pf_num; + + lio_prepare_soft_command(lio_dev, sc, LIO_OPCODE, + LIO_OPCODE_IF_CFG, 0, + if_cfg.if_cfg64, 0); + + /* Setting wait time in seconds */ + sc->wait_time = LIO_MAX_CMD_TIMEOUT / 1000; + + retval = lio_send_soft_command(lio_dev, sc); + if (retval == LIO_IQ_SEND_FAILED) { + lio_dev_err(lio_dev, "iq/oq config failed status: %x\n", + retval); + /* Soft instr is freed by driver in case of failure. */ + goto nic_config_fail; + } + + /* Sleep on a wait queue till the cond flag indicates that the + * response arrived or timed-out. + */ + while ((*sc->status_word == LIO_COMPLETION_WORD_INIT) && --timeout) { + lio_flush_iq(lio_dev, lio_dev->instr_queue[sc->iq_no]); + lio_process_ordered_list(lio_dev); + rte_delay_ms(1); + } + + retval = resp->status; + if (retval) { + lio_dev_err(lio_dev, "iq/oq config failed\n"); + goto nic_config_fail; + } + + strlcpy(lio_dev->firmware_version, + resp->cfg_info.lio_firmware_version, LIO_FW_VERSION_LENGTH); + + lio_swap_8B_data((uint64_t *)(&resp->cfg_info), + sizeof(struct octeon_if_cfg_info) >> 3); + + num_iqueues = lio_hweight64(resp->cfg_info.iqmask); + num_oqueues = lio_hweight64(resp->cfg_info.oqmask); + + if (!(num_iqueues) || !(num_oqueues)) { + lio_dev_err(lio_dev, + "Got bad iqueues (%016lx) or oqueues (%016lx) from firmware.\n", + (unsigned long)resp->cfg_info.iqmask, + (unsigned long)resp->cfg_info.oqmask); + goto nic_config_fail; + } + + lio_dev_dbg(lio_dev, + "interface %d, iqmask %016lx, oqmask %016lx, numiqueues %d, numoqueues %d\n", + eth_dev->data->port_id, + (unsigned long)resp->cfg_info.iqmask, + (unsigned long)resp->cfg_info.oqmask, + num_iqueues, num_oqueues); + + lio_dev->linfo.num_rxpciq = num_oqueues; + lio_dev->linfo.num_txpciq = num_iqueues; + + for (i = 0; i < num_oqueues; i++) { + lio_dev->linfo.rxpciq[i].rxpciq64 = + resp->cfg_info.linfo.rxpciq[i].rxpciq64; + lio_dev_dbg(lio_dev, "index %d OQ %d\n", + i, lio_dev->linfo.rxpciq[i].s.q_no); + } + + for (i = 0; i < num_iqueues; i++) { + lio_dev->linfo.txpciq[i].txpciq64 = + resp->cfg_info.linfo.txpciq[i].txpciq64; + lio_dev_dbg(lio_dev, "index %d IQ %d\n", + i, lio_dev->linfo.txpciq[i].s.q_no); + } + + lio_dev->linfo.hw_addr = resp->cfg_info.linfo.hw_addr; + lio_dev->linfo.gmxport = resp->cfg_info.linfo.gmxport; + lio_dev->linfo.link.link_status64 = + resp->cfg_info.linfo.link.link_status64; + + /* 64-bit swap required on LE machines */ + lio_swap_8B_data(&lio_dev->linfo.hw_addr, 1); + for (i = 0; i < ETHER_ADDR_LEN; i++) + mac[i] = *((uint8_t *)(((uint8_t *)&lio_dev->linfo.hw_addr) + + 2 + i)); + + /* Copy the permanent MAC address */ + ether_addr_copy((struct rte_ether_addr *)mac, + ð_dev->data->mac_addrs[0]); + + /* enable firmware checksum support for tunnel packets */ + lio_enable_hw_tunnel_rx_checksum(eth_dev); + lio_enable_hw_tunnel_tx_checksum(eth_dev); + + lio_dev->glist_lock = + rte_zmalloc(NULL, sizeof(*lio_dev->glist_lock) * num_iqueues, 0); + if (lio_dev->glist_lock == NULL) + return -ENOMEM; + + lio_dev->glist_head = + rte_zmalloc(NULL, sizeof(*lio_dev->glist_head) * num_iqueues, + 0); + if (lio_dev->glist_head == NULL) { + rte_free(lio_dev->glist_lock); + lio_dev->glist_lock = NULL; + return -ENOMEM; + } + + lio_dev_link_update(eth_dev, 0); + + lio_dev->port_configured = 1; + + lio_free_soft_command(sc); + + /* Reset ioq regs */ + lio_dev->fn_list.setup_device_regs(lio_dev); + + /* Free iq_0 used during init */ + lio_free_instr_queue0(lio_dev); + + return 0; + +nic_config_fail: + lio_dev_err(lio_dev, "Failed retval %d\n", retval); + lio_free_soft_command(sc); + lio_free_instr_queue0(lio_dev); + + return -ENODEV; +} + +/* Define our ethernet definitions */ +static const struct eth_dev_ops liovf_eth_dev_ops = { + .dev_configure = lio_dev_configure, + .dev_start = lio_dev_start, + .dev_stop = lio_dev_stop, + .dev_set_link_up = lio_dev_set_link_up, + .dev_set_link_down = lio_dev_set_link_down, + .dev_close = lio_dev_close, + .promiscuous_enable = lio_dev_promiscuous_enable, + .promiscuous_disable = lio_dev_promiscuous_disable, + .allmulticast_enable = lio_dev_allmulticast_enable, + .allmulticast_disable = lio_dev_allmulticast_disable, + .link_update = lio_dev_link_update, + .stats_get = lio_dev_stats_get, + .xstats_get = lio_dev_xstats_get, + .xstats_get_names = lio_dev_xstats_get_names, + .stats_reset = lio_dev_stats_reset, + .xstats_reset = lio_dev_xstats_reset, + .dev_infos_get = lio_dev_info_get, + .vlan_filter_set = lio_dev_vlan_filter_set, + .rx_queue_setup = lio_dev_rx_queue_setup, + .rx_queue_release = lio_dev_rx_queue_release, + .tx_queue_setup = lio_dev_tx_queue_setup, + .tx_queue_release = lio_dev_tx_queue_release, + .reta_update = lio_dev_rss_reta_update, + .reta_query = lio_dev_rss_reta_query, + .rss_hash_conf_get = lio_dev_rss_hash_conf_get, + .rss_hash_update = lio_dev_rss_hash_update, + .udp_tunnel_port_add = lio_dev_udp_tunnel_add, + .udp_tunnel_port_del = lio_dev_udp_tunnel_del, + .mtu_set = lio_dev_mtu_set, +}; + static void lio_check_pf_hs_response(void *lio_dev) { @@ -101,6 +1971,17 @@ lio_first_time_init(struct lio_device *lio_dev, return -1; } + /* Initialize soft command buffer pool */ + if (lio_setup_sc_buffer_pool(lio_dev)) { + lio_dev_err(lio_dev, "sc buffer pool allocation failed\n"); + return -1; + } + + /* Initialize lists to manage the requests of different types that + * arrive from applications for this lio device. + */ + lio_setup_response_list(lio_dev); + if (lio_dev->fn_list.setup_mbox(lio_dev)) { lio_dev_err(lio_dev, "Mailbox setup failed\n"); goto error; @@ -113,14 +1994,11 @@ lio_first_time_init(struct lio_device *lio_dev, if (cn23xx_pfvf_handshake(lio_dev)) goto error; - /* Initial reset */ - cn23xx_vf_ask_pf_to_do_flr(lio_dev); - /* Wait for FLR for 100ms per SRIOV specification */ - rte_delay_ms(100); - - if (cn23xx_vf_set_io_queues_off(lio_dev)) { - lio_dev_err(lio_dev, "Setting io queues off failed\n"); - goto error; + /* Request and wait for device reset. */ + if (pdev->kdrv == RTE_KDRV_IGB_UIO) { + cn23xx_vf_ask_pf_to_do_flr(lio_dev); + /* FLR wait time doubled as a precaution. */ + rte_delay_ms(LIO_PCI_FLR_WAIT * 2); } if (lio_dev->fn_list.setup_device_regs(lio_dev)) { @@ -138,9 +2016,14 @@ lio_first_time_init(struct lio_device *lio_dev, lio_dev->max_tx_queues = dpdk_queues; lio_dev->max_rx_queues = dpdk_queues; + /* Enable input and output queues for this device */ + if (lio_dev->fn_list.enable_io_queues(lio_dev)) + goto error; + return 0; error: + lio_free_sc_buffer_pool(lio_dev); if (lio_dev->mbox[0]) lio_dev->fn_list.free_mbox(lio_dev); if (lio_dev->instr_queue[0]) @@ -152,13 +2035,19 @@ error: static int lio_eth_dev_uninit(struct rte_eth_dev *eth_dev) { + struct lio_device *lio_dev = LIO_DEV(eth_dev); + PMD_INIT_FUNC_TRACE(); if (rte_eal_process_type() != RTE_PROC_PRIMARY) - return -EPERM; + return 0; + + /* lio_free_sc_buffer_pool */ + lio_free_sc_buffer_pool(lio_dev); - rte_free(eth_dev->data->mac_addrs); - eth_dev->data->mac_addrs = NULL; + eth_dev->dev_ops = NULL; + eth_dev->rx_pkt_burst = NULL; + eth_dev->tx_pkt_burst = NULL; return 0; } @@ -166,17 +2055,19 @@ lio_eth_dev_uninit(struct rte_eth_dev *eth_dev) static int lio_eth_dev_init(struct rte_eth_dev *eth_dev) { - struct rte_pci_device *pdev = RTE_DEV_TO_PCI(eth_dev->device); + struct rte_pci_device *pdev = RTE_ETH_DEV_TO_PCI(eth_dev); struct lio_device *lio_dev = LIO_DEV(eth_dev); PMD_INIT_FUNC_TRACE(); + eth_dev->rx_pkt_burst = &lio_dev_recv_pkts; + eth_dev->tx_pkt_burst = &lio_dev_xmit_pkts; + /* Primary does the initialization. */ if (rte_eal_process_type() != RTE_PROC_PRIMARY) return 0; rte_eth_copy_pci_info(eth_dev, pdev); - eth_dev->data->dev_flags |= RTE_ETH_DEV_DETACHABLE; if (pdev->mem_resource[0].addr) { lio_dev->hw_addr = pdev->mem_resource[0].addr; @@ -198,34 +2089,65 @@ lio_eth_dev_init(struct rte_eth_dev *eth_dev) return -EINVAL; } + eth_dev->dev_ops = &liovf_eth_dev_ops; eth_dev->data->mac_addrs = rte_zmalloc("lio", ETHER_ADDR_LEN, 0); if (eth_dev->data->mac_addrs == NULL) { lio_dev_err(lio_dev, "MAC addresses memory allocation failed\n"); + eth_dev->dev_ops = NULL; + eth_dev->rx_pkt_burst = NULL; + eth_dev->tx_pkt_burst = NULL; return -ENOMEM; } + rte_atomic64_set(&lio_dev->status, LIO_DEV_RUNNING); + rte_wmb(); + + lio_dev->port_configured = 0; + /* Always allow unicast packets */ + lio_dev->ifflags |= LIO_IFFLAG_UNICAST; + return 0; } +static int +lio_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 lio_device), + lio_eth_dev_init); +} + +static int +lio_eth_dev_pci_remove(struct rte_pci_device *pci_dev) +{ + return rte_eth_dev_pci_generic_remove(pci_dev, + lio_eth_dev_uninit); +} + /* Set of PCI devices this driver supports */ static const struct rte_pci_id pci_id_liovf_map[] = { { RTE_PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, LIO_CN23XX_VF_VID) }, { .vendor_id = 0, /* sentinel */ } }; -static struct eth_driver rte_liovf_pmd = { - .pci_drv = { - .id_table = pci_id_liovf_map, - .drv_flags = RTE_PCI_DRV_NEED_MAPPING, - .probe = rte_eth_dev_pci_probe, - .remove = rte_eth_dev_pci_remove, - }, - .eth_dev_init = lio_eth_dev_init, - .eth_dev_uninit = lio_eth_dev_uninit, - .dev_private_size = sizeof(struct lio_device), +static struct rte_pci_driver rte_liovf_pmd = { + .id_table = pci_id_liovf_map, + .drv_flags = RTE_PCI_DRV_NEED_MAPPING, + .probe = lio_eth_dev_pci_probe, + .remove = lio_eth_dev_pci_remove, }; -RTE_PMD_REGISTER_PCI(net_liovf, rte_liovf_pmd.pci_drv); +RTE_PMD_REGISTER_PCI(net_liovf, rte_liovf_pmd); RTE_PMD_REGISTER_PCI_TABLE(net_liovf, pci_id_liovf_map); -RTE_PMD_REGISTER_KMOD_DEP(net_liovf, "* igb_uio | vfio"); +RTE_PMD_REGISTER_KMOD_DEP(net_liovf, "* igb_uio | vfio-pci"); + +RTE_INIT(lio_init_log) +{ + lio_logtype_init = rte_log_register("pmd.net.liquidio.init"); + if (lio_logtype_init >= 0) + rte_log_set_level(lio_logtype_init, RTE_LOG_NOTICE); + lio_logtype_driver = rte_log_register("pmd.net.liquidio.driver"); + if (lio_logtype_driver >= 0) + rte_log_set_level(lio_logtype_driver, RTE_LOG_NOTICE); +}