+ gcmd.link_mode_masks_nwords = -gcmd.link_mode_masks_nwords;
+
+ alignas(struct ethtool_link_settings)
+ uint8_t data[offsetof(struct ethtool_link_settings, link_mode_masks) +
+ sizeof(uint32_t) * gcmd.link_mode_masks_nwords * 3];
+ struct ethtool_link_settings *ecmd = (void *)data;
+
+ *ecmd = gcmd;
+ ifr.ifr_data = (void *)ecmd;
+ ret = mlx5_ifreq(master ? master : dev, SIOCETHTOOL, &ifr);
+ if (ret) {
+ DRV_LOG(DEBUG,
+ "port %u ioctl(SIOCETHTOOL,"
+ "ETHTOOL_GLINKSETTINGS) failed: %s",
+ dev->data->port_id, strerror(rte_errno));
+ return ret;
+ }
+ dev_link.link_speed = (ecmd->speed == UINT32_MAX) ? ETH_SPEED_NUM_NONE :
+ ecmd->speed;
+ sc = ecmd->link_mode_masks[0] |
+ ((uint64_t)ecmd->link_mode_masks[1] << 32);
+ priv->link_speed_capa = 0;
+ if (sc & MLX5_BITSHIFT(ETHTOOL_LINK_MODE_Autoneg_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_AUTONEG;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_1000baseT_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_1000baseKX_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_1G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_10000baseKR_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_10000baseR_FEC_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_10G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_20000baseMLD2_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_20000baseKR2_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_20G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_40000baseKR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_40000baseCR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_40000baseSR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_40000baseLR4_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_40G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_56000baseKR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_56000baseCR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_56000baseSR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_56000baseLR4_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_56G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_25000baseCR_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_25000baseKR_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_25000baseSR_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_25G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_50000baseCR2_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_50000baseKR2_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_50G;
+ if (sc & (MLX5_BITSHIFT(ETHTOOL_LINK_MODE_100000baseKR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_100000baseSR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_100000baseCR4_Full_BIT) |
+ MLX5_BITSHIFT(ETHTOOL_LINK_MODE_100000baseLR4_ER4_Full_BIT)))
+ priv->link_speed_capa |= ETH_LINK_SPEED_100G;
+ dev_link.link_duplex = ((ecmd->duplex == DUPLEX_HALF) ?
+ ETH_LINK_HALF_DUPLEX : ETH_LINK_FULL_DUPLEX);
+ dev_link.link_autoneg = !(dev->data->dev_conf.link_speeds &
+ ETH_LINK_SPEED_FIXED);
+ if (((dev_link.link_speed && !dev_link.link_status) ||
+ (!dev_link.link_speed && dev_link.link_status))) {
+ rte_errno = EAGAIN;
+ return -rte_errno;
+ }
+ *link = dev_link;
+ return 0;
+}
+
+/**
+ * DPDK callback to retrieve physical link information.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param wait_to_complete
+ * Wait for request completion.
+ *
+ * @return
+ * 0 if link status was not updated, positive if it was, a negative errno
+ * value otherwise and rte_errno is set.
+ */
+int
+mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete)
+{
+ int ret;
+ struct rte_eth_link dev_link;
+ time_t start_time = time(NULL);
+ int retry = MLX5_GET_LINK_STATUS_RETRY_COUNT;
+
+ do {
+ ret = mlx5_link_update_unlocked_gs(dev, &dev_link);
+ if (ret == -ENOTSUP)
+ ret = mlx5_link_update_unlocked_gset(dev, &dev_link);
+ if (ret == 0)
+ break;
+ /* Handle wait to complete situation. */
+ if ((wait_to_complete || retry) && ret == -EAGAIN) {
+ if (abs((int)difftime(time(NULL), start_time)) <
+ MLX5_LINK_STATUS_TIMEOUT) {
+ usleep(0);
+ continue;
+ } else {
+ rte_errno = EBUSY;
+ return -rte_errno;
+ }
+ } else if (ret < 0) {
+ return ret;
+ }
+ } while (wait_to_complete || retry-- > 0);
+ ret = !!memcmp(&dev->data->dev_link, &dev_link,
+ sizeof(struct rte_eth_link));
+ dev->data->dev_link = dev_link;
+ return ret;
+}
+
+/**
+ * DPDK callback to change the MTU.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param in_mtu
+ * New MTU.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ uint16_t kern_mtu = 0;
+ int ret;
+
+ ret = mlx5_get_mtu(dev, &kern_mtu);
+ if (ret)
+ return ret;
+ /* Set kernel interface MTU first. */
+ ret = mlx5_set_mtu(dev, mtu);
+ if (ret)
+ return ret;
+ ret = mlx5_get_mtu(dev, &kern_mtu);
+ if (ret)
+ return ret;
+ if (kern_mtu == mtu) {
+ priv->mtu = mtu;
+ DRV_LOG(DEBUG, "port %u adapter MTU set to %u",
+ dev->data->port_id, mtu);
+ return 0;
+ }
+ rte_errno = EAGAIN;
+ return -rte_errno;
+}
+
+/**
+ * DPDK callback to get flow control status.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param[out] fc_conf
+ * Flow control output buffer.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
+{
+ struct ifreq ifr;
+ struct ethtool_pauseparam ethpause = {
+ .cmd = ETHTOOL_GPAUSEPARAM
+ };
+ int ret;
+
+ ifr.ifr_data = (void *)ðpause;
+ ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+ if (ret) {
+ DRV_LOG(WARNING,
+ "port %u ioctl(SIOCETHTOOL, ETHTOOL_GPAUSEPARAM) failed:"
+ " %s",
+ dev->data->port_id, strerror(rte_errno));
+ return ret;
+ }
+ fc_conf->autoneg = ethpause.autoneg;
+ if (ethpause.rx_pause && ethpause.tx_pause)
+ fc_conf->mode = RTE_FC_FULL;
+ else if (ethpause.rx_pause)
+ fc_conf->mode = RTE_FC_RX_PAUSE;
+ else if (ethpause.tx_pause)
+ fc_conf->mode = RTE_FC_TX_PAUSE;
+ else
+ fc_conf->mode = RTE_FC_NONE;
+ return 0;
+}
+
+/**
+ * DPDK callback to modify flow control parameters.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param[in] fc_conf
+ * Flow control parameters.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf)
+{
+ struct ifreq ifr;
+ struct ethtool_pauseparam ethpause = {
+ .cmd = ETHTOOL_SPAUSEPARAM
+ };
+ int ret;
+
+ ifr.ifr_data = (void *)ðpause;
+ ethpause.autoneg = fc_conf->autoneg;
+ if (((fc_conf->mode & RTE_FC_FULL) == RTE_FC_FULL) ||
+ (fc_conf->mode & RTE_FC_RX_PAUSE))
+ ethpause.rx_pause = 1;
+ else
+ ethpause.rx_pause = 0;
+
+ if (((fc_conf->mode & RTE_FC_FULL) == RTE_FC_FULL) ||
+ (fc_conf->mode & RTE_FC_TX_PAUSE))
+ ethpause.tx_pause = 1;
+ else
+ ethpause.tx_pause = 0;
+ ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+ if (ret) {
+ DRV_LOG(WARNING,
+ "port %u ioctl(SIOCETHTOOL, ETHTOOL_SPAUSEPARAM)"
+ " failed: %s",
+ dev->data->port_id, strerror(rte_errno));
+ return ret;
+ }
+ return 0;
+}
+
+/**
+ * Handle asynchronous removal event for entire multiport device.
+ *
+ * @param sh
+ * Infiniband device shared context.
+ */
+static void
+mlx5_dev_interrupt_device_fatal(struct mlx5_ibv_shared *sh)
+{
+ uint32_t i;
+
+ for (i = 0; i < sh->max_port; ++i) {
+ struct rte_eth_dev *dev;
+
+ if (sh->port[i].ih_port_id >= RTE_MAX_ETHPORTS) {
+ /*
+ * Or not existing port either no
+ * handler installed for this port.
+ */
+ continue;
+ }
+ dev = &rte_eth_devices[sh->port[i].ih_port_id];
+ MLX5_ASSERT(dev);
+ if (dev->data->dev_conf.intr_conf.rmv)
+ _rte_eth_dev_callback_process
+ (dev, RTE_ETH_EVENT_INTR_RMV, NULL);
+ }
+}
+
+/**
+ * Handle shared asynchronous events the NIC (removal event
+ * and link status change). Supports multiport IB device.
+ *
+ * @param cb_arg
+ * Callback argument.
+ */
+void
+mlx5_dev_interrupt_handler(void *cb_arg)
+{
+ struct mlx5_ibv_shared *sh = cb_arg;
+ struct ibv_async_event event;
+
+ /* Read all message from the IB device and acknowledge them. */
+ for (;;) {
+ struct rte_eth_dev *dev;
+ uint32_t tmp;
+
+ if (mlx5_glue->get_async_event(sh->ctx, &event))
+ break;
+ /* Retrieve and check IB port index. */
+ tmp = (uint32_t)event.element.port_num;
+ if (!tmp && event.event_type == IBV_EVENT_DEVICE_FATAL) {
+ /*
+ * The DEVICE_FATAL event is called once for
+ * entire device without port specifying.
+ * We should notify all existing ports.
+ */
+ mlx5_glue->ack_async_event(&event);
+ mlx5_dev_interrupt_device_fatal(sh);
+ continue;
+ }
+ MLX5_ASSERT(tmp && (tmp <= sh->max_port));
+ if (!tmp) {
+ /* Unsupported devive level event. */
+ mlx5_glue->ack_async_event(&event);
+ DRV_LOG(DEBUG,
+ "unsupported common event (type %d)",
+ event.event_type);
+ continue;
+ }
+ if (tmp > sh->max_port) {
+ /* Invalid IB port index. */
+ mlx5_glue->ack_async_event(&event);
+ DRV_LOG(DEBUG,
+ "cannot handle an event (type %d)"
+ "due to invalid IB port index (%u)",
+ event.event_type, tmp);
+ continue;
+ }
+ if (sh->port[tmp - 1].ih_port_id >= RTE_MAX_ETHPORTS) {
+ /* No handler installed. */
+ mlx5_glue->ack_async_event(&event);
+ DRV_LOG(DEBUG,
+ "cannot handle an event (type %d)"
+ "due to no handler installed for port %u",
+ event.event_type, tmp);
+ continue;
+ }
+ /* Retrieve ethernet device descriptor. */
+ tmp = sh->port[tmp - 1].ih_port_id;
+ dev = &rte_eth_devices[tmp];
+ MLX5_ASSERT(dev);
+ if ((event.event_type == IBV_EVENT_PORT_ACTIVE ||
+ event.event_type == IBV_EVENT_PORT_ERR) &&
+ dev->data->dev_conf.intr_conf.lsc) {
+ mlx5_glue->ack_async_event(&event);
+ if (mlx5_link_update(dev, 0) == -EAGAIN) {
+ usleep(0);
+ continue;
+ }
+ _rte_eth_dev_callback_process
+ (dev, RTE_ETH_EVENT_INTR_LSC, NULL);
+ continue;
+ }
+ DRV_LOG(DEBUG,
+ "port %u cannot handle an unknown event (type %d)",
+ dev->data->port_id, event.event_type);
+ mlx5_glue->ack_async_event(&event);
+ }
+}
+
+/*
+ * Unregister callback handler safely. The handler may be active
+ * while we are trying to unregister it, in this case code -EAGAIN
+ * is returned by rte_intr_callback_unregister(). This routine checks
+ * the return code and tries to unregister handler again.
+ *
+ * @param handle
+ * interrupt handle
+ * @param cb_fn
+ * pointer to callback routine
+ * @cb_arg
+ * opaque callback parameter
+ */
+void
+mlx5_intr_callback_unregister(const struct rte_intr_handle *handle,
+ rte_intr_callback_fn cb_fn, void *cb_arg)
+{
+ /*
+ * Try to reduce timeout management overhead by not calling
+ * the timer related routines on the first iteration. If the
+ * unregistering succeeds on first call there will be no
+ * timer calls at all.
+ */
+ uint64_t twait = 0;
+ uint64_t start = 0;
+
+ do {
+ int ret;
+
+ ret = rte_intr_callback_unregister(handle, cb_fn, cb_arg);
+ if (ret >= 0)
+ return;
+ if (ret != -EAGAIN) {
+ DRV_LOG(INFO, "failed to unregister interrupt"
+ " handler (error: %d)", ret);
+ MLX5_ASSERT(false);
+ return;
+ }
+ if (twait) {
+ struct timespec onems;
+
+ /* Wait one millisecond and try again. */
+ onems.tv_sec = 0;
+ onems.tv_nsec = NS_PER_S / MS_PER_S;
+ nanosleep(&onems, 0);
+ /* Check whether one second elapsed. */
+ if ((rte_get_timer_cycles() - start) <= twait)
+ continue;
+ } else {
+ /*
+ * We get the amount of timer ticks for one second.
+ * If this amount elapsed it means we spent one
+ * second in waiting. This branch is executed once
+ * on first iteration.
+ */
+ twait = rte_get_timer_hz();
+ MLX5_ASSERT(twait);
+ }
+ /*
+ * Timeout elapsed, show message (once a second) and retry.
+ * We have no other acceptable option here, if we ignore
+ * the unregistering return code the handler will not
+ * be unregistered, fd will be closed and we may get the
+ * crush. Hanging and messaging in the loop seems not to be
+ * the worst choice.
+ */
+ DRV_LOG(INFO, "Retrying to unregister interrupt handler");
+ start = rte_get_timer_cycles();
+ } while (true);
+}
+
+/**
+ * Handle DEVX interrupts from the NIC.
+ * This function is probably called from the DPDK host thread.
+ *
+ * @param cb_arg
+ * Callback argument.
+ */
+void
+mlx5_dev_interrupt_handler_devx(void *cb_arg)
+{
+#ifndef HAVE_IBV_DEVX_ASYNC
+ (void)cb_arg;
+ return;
+#else
+ struct mlx5_ibv_shared *sh = cb_arg;
+ union {
+ struct mlx5dv_devx_async_cmd_hdr cmd_resp;
+ uint8_t buf[MLX5_ST_SZ_BYTES(query_flow_counter_out) +
+ MLX5_ST_SZ_BYTES(traffic_counter) +
+ sizeof(struct mlx5dv_devx_async_cmd_hdr)];
+ } out;
+ uint8_t *buf = out.buf + sizeof(out.cmd_resp);
+
+ while (!mlx5_glue->devx_get_async_cmd_comp(sh->devx_comp,
+ &out.cmd_resp,
+ sizeof(out.buf)))
+ mlx5_flow_async_pool_query_handle
+ (sh, (uint64_t)out.cmd_resp.wr_id,
+ mlx5_devx_get_out_command_status(buf));
+#endif /* HAVE_IBV_DEVX_ASYNC */
+}
+
+/**
+ * Uninstall shared asynchronous device events handler.
+ * This function is implemented to support event sharing
+ * between multiple ports of single IB device.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+static void
+mlx5_dev_shared_handler_uninstall(struct rte_eth_dev *dev)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_ibv_shared *sh = priv->sh;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return;
+ pthread_mutex_lock(&sh->intr_mutex);
+ MLX5_ASSERT(priv->ibv_port);
+ MLX5_ASSERT(priv->ibv_port <= sh->max_port);
+ MLX5_ASSERT(dev->data->port_id < RTE_MAX_ETHPORTS);
+ if (sh->port[priv->ibv_port - 1].ih_port_id >= RTE_MAX_ETHPORTS)
+ goto exit;
+ MLX5_ASSERT(sh->port[priv->ibv_port - 1].ih_port_id ==
+ (uint32_t)dev->data->port_id);
+ MLX5_ASSERT(sh->intr_cnt);
+ sh->port[priv->ibv_port - 1].ih_port_id = RTE_MAX_ETHPORTS;
+ if (!sh->intr_cnt || --sh->intr_cnt)
+ goto exit;
+ mlx5_intr_callback_unregister(&sh->intr_handle,
+ mlx5_dev_interrupt_handler, sh);
+ sh->intr_handle.fd = 0;
+ sh->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN;
+exit:
+ pthread_mutex_unlock(&sh->intr_mutex);
+}
+
+/**
+ * Uninstall devx shared asynchronous device events handler.
+ * This function is implemeted to support event sharing
+ * between multiple ports of single IB device.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+static void
+mlx5_dev_shared_handler_devx_uninstall(struct rte_eth_dev *dev)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_ibv_shared *sh = priv->sh;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return;
+ pthread_mutex_lock(&sh->intr_mutex);
+ MLX5_ASSERT(priv->ibv_port);
+ MLX5_ASSERT(priv->ibv_port <= sh->max_port);
+ MLX5_ASSERT(dev->data->port_id < RTE_MAX_ETHPORTS);
+ if (sh->port[priv->ibv_port - 1].devx_ih_port_id >= RTE_MAX_ETHPORTS)
+ goto exit;
+ MLX5_ASSERT(sh->port[priv->ibv_port - 1].devx_ih_port_id ==
+ (uint32_t)dev->data->port_id);
+ sh->port[priv->ibv_port - 1].devx_ih_port_id = RTE_MAX_ETHPORTS;
+ if (!sh->devx_intr_cnt || --sh->devx_intr_cnt)
+ goto exit;
+ if (sh->intr_handle_devx.fd) {
+ rte_intr_callback_unregister(&sh->intr_handle_devx,
+ mlx5_dev_interrupt_handler_devx,
+ sh);
+ sh->intr_handle_devx.fd = 0;
+ sh->intr_handle_devx.type = RTE_INTR_HANDLE_UNKNOWN;
+ }
+ if (sh->devx_comp) {
+ mlx5_glue->devx_destroy_cmd_comp(sh->devx_comp);
+ sh->devx_comp = NULL;
+ }
+exit:
+ pthread_mutex_unlock(&sh->intr_mutex);
+}
+
+/**
+ * Install shared asynchronous device events handler.
+ * This function is implemented to support event sharing
+ * between multiple ports of single IB device.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+static void
+mlx5_dev_shared_handler_install(struct rte_eth_dev *dev)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_ibv_shared *sh = priv->sh;
+ int ret;
+ int flags;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return;
+ pthread_mutex_lock(&sh->intr_mutex);
+ MLX5_ASSERT(priv->ibv_port);
+ MLX5_ASSERT(priv->ibv_port <= sh->max_port);
+ MLX5_ASSERT(dev->data->port_id < RTE_MAX_ETHPORTS);
+ if (sh->port[priv->ibv_port - 1].ih_port_id < RTE_MAX_ETHPORTS) {
+ /* The handler is already installed for this port. */
+ MLX5_ASSERT(sh->intr_cnt);
+ goto exit;
+ }
+ if (sh->intr_cnt) {
+ sh->port[priv->ibv_port - 1].ih_port_id =
+ (uint32_t)dev->data->port_id;
+ sh->intr_cnt++;
+ goto exit;
+ }
+ /* No shared handler installed. */
+ MLX5_ASSERT(sh->ctx->async_fd > 0);
+ flags = fcntl(sh->ctx->async_fd, F_GETFL);
+ ret = fcntl(sh->ctx->async_fd, F_SETFL, flags | O_NONBLOCK);
+ if (ret) {
+ DRV_LOG(INFO, "failed to change file descriptor async event"
+ " queue");
+ /* Indicate there will be no interrupts. */
+ dev->data->dev_conf.intr_conf.lsc = 0;
+ dev->data->dev_conf.intr_conf.rmv = 0;
+ } else {
+ sh->intr_handle.fd = sh->ctx->async_fd;
+ sh->intr_handle.type = RTE_INTR_HANDLE_EXT;
+ rte_intr_callback_register(&sh->intr_handle,
+ mlx5_dev_interrupt_handler, sh);
+ sh->intr_cnt++;
+ sh->port[priv->ibv_port - 1].ih_port_id =
+ (uint32_t)dev->data->port_id;
+ }
+exit:
+ pthread_mutex_unlock(&sh->intr_mutex);
+}
+
+/**
+ * Install devx shared asyncronous device events handler.
+ * This function is implemeted to support event sharing
+ * between multiple ports of single IB device.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+static void
+mlx5_dev_shared_handler_devx_install(struct rte_eth_dev *dev)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_ibv_shared *sh = priv->sh;
+
+ if (rte_eal_process_type() != RTE_PROC_PRIMARY)
+ return;
+ pthread_mutex_lock(&sh->intr_mutex);
+ MLX5_ASSERT(priv->ibv_port);
+ MLX5_ASSERT(priv->ibv_port <= sh->max_port);
+ MLX5_ASSERT(dev->data->port_id < RTE_MAX_ETHPORTS);
+ if (sh->port[priv->ibv_port - 1].devx_ih_port_id < RTE_MAX_ETHPORTS) {
+ /* The handler is already installed for this port. */
+ MLX5_ASSERT(sh->devx_intr_cnt);
+ goto exit;
+ }
+ if (sh->devx_intr_cnt) {
+ sh->devx_intr_cnt++;
+ sh->port[priv->ibv_port - 1].devx_ih_port_id =
+ (uint32_t)dev->data->port_id;
+ goto exit;
+ }
+ if (priv->config.devx) {
+#ifndef HAVE_IBV_DEVX_ASYNC
+ goto exit;
+#else
+ sh->devx_comp = mlx5_glue->devx_create_cmd_comp(sh->ctx);
+ if (sh->devx_comp) {
+ int flags = fcntl(sh->devx_comp->fd, F_GETFL);
+ int ret = fcntl(sh->devx_comp->fd, F_SETFL,
+ flags | O_NONBLOCK);
+
+ if (ret) {
+ DRV_LOG(INFO, "failed to change file descriptor"
+ " devx async event queue");
+ } else {
+ sh->intr_handle_devx.fd = sh->devx_comp->fd;
+ sh->intr_handle_devx.type = RTE_INTR_HANDLE_EXT;
+ rte_intr_callback_register
+ (&sh->intr_handle_devx,
+ mlx5_dev_interrupt_handler_devx, sh);
+ sh->devx_intr_cnt++;
+ sh->port[priv->ibv_port - 1].devx_ih_port_id =
+ (uint32_t)dev->data->port_id;
+ }
+ }
+#endif /* HAVE_IBV_DEVX_ASYNC */
+ }
+exit:
+ pthread_mutex_unlock(&sh->intr_mutex);
+}
+
+/**
+ * Uninstall interrupt handler.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+void
+mlx5_dev_interrupt_handler_uninstall(struct rte_eth_dev *dev)
+{
+ mlx5_dev_shared_handler_uninstall(dev);
+}
+
+/**
+ * Install interrupt handler.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+void
+mlx5_dev_interrupt_handler_install(struct rte_eth_dev *dev)
+{
+ mlx5_dev_shared_handler_install(dev);
+}
+
+/**
+ * Devx uninstall interrupt handler.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+void
+mlx5_dev_interrupt_handler_devx_uninstall(struct rte_eth_dev *dev)
+{
+ mlx5_dev_shared_handler_devx_uninstall(dev);
+}
+
+/**
+ * Devx install interrupt handler.
+ *
+ * @param dev
+ * Pointer to Ethernet device.
+ */
+void
+mlx5_dev_interrupt_handler_devx_install(struct rte_eth_dev *dev)
+{
+ mlx5_dev_shared_handler_devx_install(dev);
+}
+
+/**
+ * DPDK callback to bring the link DOWN.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_set_link_down(struct rte_eth_dev *dev)
+{
+ return mlx5_set_flags(dev, ~IFF_UP, ~IFF_UP);
+}
+
+/**
+ * DPDK callback to bring the link UP.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_set_link_up(struct rte_eth_dev *dev)
+{
+ return mlx5_set_flags(dev, ~IFF_UP, IFF_UP);
+}
+
+/**
+ * Configure the RX function to use.
+ *
+ * @param dev
+ * Pointer to private data structure.
+ *
+ * @return
+ * Pointer to selected Rx burst function.
+ */
+eth_rx_burst_t
+mlx5_select_rx_function(struct rte_eth_dev *dev)
+{
+ eth_rx_burst_t rx_pkt_burst = mlx5_rx_burst;
+
+ MLX5_ASSERT(dev != NULL);
+ if (mlx5_check_vec_rx_support(dev) > 0) {
+ rx_pkt_burst = mlx5_rx_burst_vec;
+ DRV_LOG(DEBUG, "port %u selected Rx vectorized function",
+ dev->data->port_id);
+ } else if (mlx5_mprq_enabled(dev)) {
+ rx_pkt_burst = mlx5_rx_burst_mprq;
+ }
+ return rx_pkt_burst;
+}
+
+/**
+ * Check if mlx5 device was removed.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ *
+ * @return
+ * 1 when device is removed, otherwise 0.
+ */
+int
+mlx5_is_removed(struct rte_eth_dev *dev)
+{
+ struct ibv_device_attr device_attr;
+ struct mlx5_priv *priv = dev->data->dev_private;
+
+ if (mlx5_glue->query_device(priv->sh->ctx, &device_attr) == EIO)
+ return 1;
+ return 0;
+}
+
+/**
+ * Get the E-Switch parameters by port id.
+ *
+ * @param[in] port
+ * Device port id.
+ * @param[in] valid
+ * Device port id is valid, skip check. This flag is useful
+ * when trials are performed from probing and device is not
+ * flagged as valid yet (in attaching process).
+ * @param[out] es_domain_id
+ * E-Switch domain id.
+ * @param[out] es_port_id
+ * The port id of the port in the E-Switch.
+ *
+ * @return
+ * pointer to device private data structure containing data needed
+ * on success, NULL otherwise and rte_errno is set.
+ */
+struct mlx5_priv *
+mlx5_port_to_eswitch_info(uint16_t port, bool valid)
+{
+ struct rte_eth_dev *dev;
+ struct mlx5_priv *priv;
+
+ if (port >= RTE_MAX_ETHPORTS) {
+ rte_errno = EINVAL;
+ return NULL;
+ }
+ if (!valid && !rte_eth_dev_is_valid_port(port)) {
+ rte_errno = ENODEV;
+ return NULL;
+ }
+ dev = &rte_eth_devices[port];
+ priv = dev->data->dev_private;
+ if (!(priv->representor || priv->master)) {
+ rte_errno = EINVAL;
+ return NULL;
+ }
+ return priv;
+}
+
+/**
+ * Get the E-Switch parameters by device instance.
+ *
+ * @param[in] port
+ * Device port id.
+ * @param[out] es_domain_id
+ * E-Switch domain id.
+ * @param[out] es_port_id
+ * The port id of the port in the E-Switch.
+ *
+ * @return
+ * pointer to device private data structure containing data needed
+ * on success, NULL otherwise and rte_errno is set.
+ */
+struct mlx5_priv *
+mlx5_dev_to_eswitch_info(struct rte_eth_dev *dev)
+{
+ struct mlx5_priv *priv;
+
+ priv = dev->data->dev_private;
+ if (!(priv->representor || priv->master)) {
+ rte_errno = EINVAL;
+ return NULL;
+ }
+ return priv;
+}
+
+/**
+ * Get switch information associated with network interface.
+ *
+ * @param ifindex
+ * Network interface index.
+ * @param[out] info
+ * Switch information object, populated in case of success.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_sysfs_switch_info(unsigned int ifindex, struct mlx5_switch_info *info)
+{
+ char ifname[IF_NAMESIZE];
+ char port_name[IF_NAMESIZE];
+ FILE *file;
+ struct mlx5_switch_info data = {
+ .master = 0,
+ .representor = 0,
+ .name_type = MLX5_PHYS_PORT_NAME_TYPE_NOTSET,
+ .port_name = 0,
+ .switch_id = 0,
+ };
+ DIR *dir;
+ bool port_switch_id_set = false;
+ bool device_dir = false;
+ char c;
+ int ret;
+
+ if (!if_indextoname(ifindex, ifname)) {
+ rte_errno = errno;
+ return -rte_errno;
+ }
+
+ MKSTR(phys_port_name, "/sys/class/net/%s/phys_port_name",
+ ifname);
+ MKSTR(phys_switch_id, "/sys/class/net/%s/phys_switch_id",
+ ifname);
+ MKSTR(pci_device, "/sys/class/net/%s/device",
+ ifname);
+
+ file = fopen(phys_port_name, "rb");
+ if (file != NULL) {
+ ret = fscanf(file, "%s", port_name);
+ fclose(file);
+ if (ret == 1)
+ mlx5_translate_port_name(port_name, &data);
+ }
+ file = fopen(phys_switch_id, "rb");
+ if (file == NULL) {
+ rte_errno = errno;
+ return -rte_errno;
+ }
+ port_switch_id_set =
+ fscanf(file, "%" SCNx64 "%c", &data.switch_id, &c) == 2 &&
+ c == '\n';
+ fclose(file);
+ dir = opendir(pci_device);
+ if (dir != NULL) {
+ closedir(dir);
+ device_dir = true;
+ }
+ if (port_switch_id_set) {
+ /* We have some E-Switch configuration. */
+ mlx5_sysfs_check_switch_info(device_dir, &data);
+ }
+ *info = data;
+ MLX5_ASSERT(!(data.master && data.representor));
+ if (data.master && data.representor) {
+ DRV_LOG(ERR, "ifindex %u device is recognized as master"
+ " and as representor", ifindex);
+ rte_errno = ENODEV;
+ return -rte_errno;
+ }
+ return 0;
+}
+
+/**
+ * Analyze gathered port parameters via sysfs to recognize master
+ * and representor devices for E-Switch configuration.
+ *
+ * @param[in] device_dir
+ * flag of presence of "device" directory under port device key.
+ * @param[inout] switch_info
+ * Port information, including port name as a number and port name
+ * type if recognized
+ *
+ * @return
+ * master and representor flags are set in switch_info according to
+ * recognized parameters (if any).
+ */
+void
+mlx5_sysfs_check_switch_info(bool device_dir,
+ struct mlx5_switch_info *switch_info)
+{
+ switch (switch_info->name_type) {
+ case MLX5_PHYS_PORT_NAME_TYPE_UNKNOWN:
+ /*
+ * Name is not recognized, assume the master,
+ * check the device directory presence.
+ */
+ switch_info->master = device_dir;
+ break;
+ case MLX5_PHYS_PORT_NAME_TYPE_NOTSET:
+ /*
+ * Name is not set, this assumes the legacy naming
+ * schema for master, just check if there is
+ * a device directory.
+ */
+ switch_info->master = device_dir;
+ break;
+ case MLX5_PHYS_PORT_NAME_TYPE_UPLINK:
+ /* New uplink naming schema recognized. */
+ switch_info->master = 1;
+ break;
+ case MLX5_PHYS_PORT_NAME_TYPE_LEGACY:
+ /* Legacy representors naming schema. */
+ switch_info->representor = !device_dir;
+ break;
+ case MLX5_PHYS_PORT_NAME_TYPE_PFVF:
+ /* New representors naming schema. */
+ switch_info->representor = 1;
+ break;
+ }
+}
+
+/**
+ * DPDK callback to retrieve plug-in module EEPROM information (type and size).
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param[out] modinfo
+ * Storage for plug-in module EEPROM information.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int
+mlx5_get_module_info(struct rte_eth_dev *dev,
+ struct rte_eth_dev_module_info *modinfo)
+{
+ struct ethtool_modinfo info = {
+ .cmd = ETHTOOL_GMODULEINFO,
+ };
+ struct ifreq ifr = (struct ifreq) {
+ .ifr_data = (void *)&info,
+ };
+ int ret = 0;
+
+ if (!dev || !modinfo) {
+ DRV_LOG(WARNING, "missing argument, cannot get module info");
+ rte_errno = EINVAL;
+ return -rte_errno;
+ }
+ ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+ if (ret) {
+ DRV_LOG(WARNING, "port %u ioctl(SIOCETHTOOL) failed: %s",
+ dev->data->port_id, strerror(rte_errno));
+ return ret;
+ }
+ modinfo->type = info.type;
+ modinfo->eeprom_len = info.eeprom_len;
+ return ret;
+}
+
+/**
+ * DPDK callback to retrieve plug-in module EEPROM data.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param[out] info
+ * Storage for plug-in module EEPROM data.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int mlx5_get_module_eeprom(struct rte_eth_dev *dev,
+ struct rte_dev_eeprom_info *info)
+{
+ struct ethtool_eeprom *eeprom;
+ struct ifreq ifr;
+ int ret = 0;
+
+ if (!dev || !info) {
+ DRV_LOG(WARNING, "missing argument, cannot get module eeprom");
+ rte_errno = EINVAL;
+ return -rte_errno;
+ }
+ eeprom = rte_calloc(__func__, 1,
+ (sizeof(struct ethtool_eeprom) + info->length), 0);
+ if (!eeprom) {
+ DRV_LOG(WARNING, "port %u cannot allocate memory for "
+ "eeprom data", dev->data->port_id);
+ rte_errno = ENOMEM;
+ return -rte_errno;
+ }
+ eeprom->cmd = ETHTOOL_GMODULEEEPROM;
+ eeprom->offset = info->offset;
+ eeprom->len = info->length;
+ ifr = (struct ifreq) {
+ .ifr_data = (void *)eeprom,
+ };
+ ret = mlx5_ifreq(dev, SIOCETHTOOL, &ifr);
+ if (ret)
+ DRV_LOG(WARNING, "port %u ioctl(SIOCETHTOOL) failed: %s",
+ dev->data->port_id, strerror(rte_errno));
+ else
+ rte_memcpy(info->data, eeprom->data, info->length);
+ rte_free(eeprom);
+ return ret;
+}
+
+/**
+ * DPDK callback to retrieve hairpin capabilities.
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param[out] cap
+ * Storage for hairpin capability data.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+int mlx5_hairpin_cap_get(struct rte_eth_dev *dev,
+ struct rte_eth_hairpin_cap *cap)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+
+ if (priv->sh->devx == 0) {
+ rte_errno = ENOTSUP;
+ return -rte_errno;
+ }
+ cap->max_nb_queues = UINT16_MAX;
+ cap->max_rx_2_tx = 1;
+ cap->max_tx_2_rx = 1;
+ cap->max_nb_desc = 8192;