From: Martin Spinler <[email protected]> Enable configuration of the MTU on the Rx MAC assigned to the port. The max MTU value is also reported in dev_info.
Signed-off-by: Martin Spinler <[email protected]> --- doc/guides/rel_notes/release_26_03.rst | 4 +-- drivers/net/nfb/nfb.h | 1 + drivers/net/nfb/nfb_ethdev.c | 45 +++++++++++++++++++++++++- 3 files changed, 47 insertions(+), 3 deletions(-) diff --git a/doc/guides/rel_notes/release_26_03.rst b/doc/guides/rel_notes/release_26_03.rst index ac76ac5b67..7ce852e349 100644 --- a/doc/guides/rel_notes/release_26_03.rst +++ b/doc/guides/rel_notes/release_26_03.rst @@ -64,10 +64,10 @@ New Features * The timestamp value has been updated to make it usable. * The DPDK port now represents just one Ethernet port instead of all Ethernet ports on the NIC. * All ports are used by default, but a subset can be selected using the ``port`` argument. - * Report firmware version and correct Ethernet link speed. + * Report firmware version, correct Ethernet link speed and maximum capable MTU value. * Common CESNET-NDK-based adapters have been added, including the FB2CGHH (Silicom Denmark) and XpressSX AGI-FH400G (Reflex CES). - * Added support for configuration of the RS-FEC mode, Link Up / Link Down state. + * Added support for configuration of the RS-FEC mode, Link Up / Link Down state, and the Rx MTU. Removed Items diff --git a/drivers/net/nfb/nfb.h b/drivers/net/nfb/nfb.h index 3b905bbb3b..7d192946bc 100644 --- a/drivers/net/nfb/nfb.h +++ b/drivers/net/nfb/nfb.h @@ -84,6 +84,7 @@ struct pmd_priv { int *queue_map_rx; /** Mapping from DPDK TX queue index to firmware queue ID */ int *queue_map_tx; + uint16_t frame_len_max_cap; /**< Maximum length obtained from Rx MAC */ bool ready; /**< This structure is initialized for usage in secondary process */ }; diff --git a/drivers/net/nfb/nfb_ethdev.c b/drivers/net/nfb/nfb_ethdev.c index a556ceca67..1093c6edf4 100644 --- a/drivers/net/nfb/nfb_ethdev.c +++ b/drivers/net/nfb/nfb_ethdev.c @@ -51,6 +51,7 @@ static struct nfb_pmd_internals_head nfb_eth_dev_list = TAILQ_HEAD_INITIALIZER(nfb_eth_dev_list); static int nfb_eth_dev_uninit(struct rte_eth_dev *dev); +static int nfb_eth_mtu_set(struct rte_eth_dev *dev, uint16_t mtu); static int nfb_mdio_read(void *priv, int prtad, int devad, uint16_t addr) @@ -271,6 +272,10 @@ nfb_eth_dev_configure(struct rte_eth_dev *dev __rte_unused) int ret; struct rte_eth_conf *dev_conf = &dev->data->dev_conf; + ret = nfb_eth_mtu_set(dev, dev_conf->rxmode.mtu); + if (ret) + goto err_mtu_set; + if (dev_conf->rxmode.offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP) { ret = rte_mbuf_dyn_rx_timestamp_register (&nfb_timestamp_dynfield_offset, @@ -285,6 +290,7 @@ nfb_eth_dev_configure(struct rte_eth_dev *dev __rte_unused) return 0; err_ts_register: +err_mtu_set: nfb_eth_dev_uninit(dev); return ret; } @@ -330,7 +336,8 @@ nfb_eth_dev_info(struct rte_eth_dev *dev, dev_info->max_mac_addrs = nfb_eth_get_max_mac_address_count(dev); - dev_info->max_rx_pktlen = (uint32_t)-1; + dev_info->max_rx_pktlen = priv->frame_len_max_cap; + dev_info->max_mtu = dev_info->max_rx_pktlen - (RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN); dev_info->max_rx_queues = priv->max_rx_queues; dev_info->max_tx_queues = priv->max_tx_queues; dev_info->speed_capa = RTE_ETH_LINK_SPEED_FIXED; @@ -575,6 +582,39 @@ nfb_eth_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index) nc_rxmac_set_mac(internals->rxmac[i], index, 0, 0); } +static uint16_t +nfb_mac_read_frame_len_max_cap(struct pmd_internals *intl) +{ + unsigned int i; + uint16_t ret = UINT16_MAX; + struct nc_rxmac_status status; + + for (i = 0; i < intl->max_rxmac; ++i) { + status.frame_length_max_capable = 0; + nc_rxmac_read_status(intl->rxmac[i], &status); + if (status.frame_length_max_capable && ret > status.frame_length_max_capable) + ret = status.frame_length_max_capable; + } + return ret; +} + +static int +nfb_eth_mtu_set(struct rte_eth_dev *dev, uint16_t mtu) +{ + unsigned int i; + struct pmd_internals *intl = dev->process_private; + struct pmd_priv *priv = dev->data->dev_private; + + mtu += RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN; + if (mtu > priv->frame_len_max_cap || mtu < RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN) + return -EINVAL; + + for (i = 0; i < intl->max_rxmac; ++i) + nc_rxmac_set_frame_length(intl->rxmac[i], mtu, RXMAC_FRAME_LENGTH_MAX); + + return 0; +} + static int nfb_eth_fw_version_get(struct rte_eth_dev *dev, char *fw_version, size_t fw_size) @@ -690,6 +730,7 @@ static const struct eth_dev_ops ops = { .mac_addr_set = nfb_eth_mac_addr_set, .mac_addr_add = nfb_eth_mac_addr_add, .mac_addr_remove = nfb_eth_mac_addr_remove, + .mtu_set = nfb_eth_mtu_set, .fw_version_get = nfb_eth_fw_version_get, .fec_get = nfb_eth_fec_get, .fec_set = nfb_eth_fec_set, @@ -793,6 +834,8 @@ nfb_eth_dev_init(struct rte_eth_dev *dev, void *init_data) } } + priv->frame_len_max_cap = nfb_mac_read_frame_len_max_cap(internals); + /* Allocate space for MAC addresses */ mac_count = nfb_eth_get_max_mac_address_count(dev); data->mac_addrs = rte_zmalloc(data->name, -- 2.53.0
