* Netronome vNIC DPDK Poll-Mode Driver: Main entry point
*/
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-#include <unistd.h>
-#include <string.h>
-#include <sys/mman.h>
-#include <sys/socket.h>
-#include <sys/io.h>
-#include <assert.h>
-#include <time.h>
#include <math.h>
-#include <inttypes.h>
#include <rte_byteorder.h>
#include <rte_common.h>
#include <rte_version.h>
#include <rte_string_fns.h>
#include <rte_alarm.h>
+#include <rte_spinlock.h>
#include "nfp_net_pmd.h"
#include "nfp_net_logs.h"
/* Prototypes */
static void nfp_net_close(struct rte_eth_dev *dev);
static int nfp_net_configure(struct rte_eth_dev *dev);
+static void nfp_net_dev_interrupt_handler(struct rte_intr_handle *handle,
+ void *param);
+static void nfp_net_dev_interrupt_delayed_handler(void *param);
static int nfp_net_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu);
static void nfp_net_infos_get(struct rte_eth_dev *dev,
struct rte_eth_dev_info *dev_info);
txq->wr_p = 0;
txq->rd_p = 0;
txq->tail = 0;
+ txq->qcp_rd_p = 0;
}
static int
PMD_DRV_LOG(DEBUG, "nfp_net_reconfig: ctrl=%08x update=%08x\n",
ctrl, update);
+ rte_spinlock_lock(&hw->reconfig_lock);
+
nn_cfg_writel(hw, NFP_NET_CFG_CTRL, ctrl);
nn_cfg_writel(hw, NFP_NET_CFG_UPDATE, update);
err = __nfp_net_reconfig(hw, update);
+ rte_spinlock_unlock(&hw->reconfig_lock);
+
if (!err)
return 0;
nfp_net_stop(dev);
+ rte_intr_disable(&dev->pci_dev->intr_handle);
nn_cfg_writeb(hw, NFP_NET_CFG_LSC, 0xff);
/*
memset(&link, 0, sizeof(struct rte_eth_link));
if (nn_link_status & NFP_NET_CFG_STS_LINK)
- link.link_status = 1;
+ link.link_status = ETH_LINK_UP;
link.link_duplex = ETH_LINK_FULL_DUPLEX;
/* Other cards can limit the tx and rx rate per VF */
- link.link_speed = ETH_LINK_SPEED_40G;
+ link.link_speed = ETH_SPEED_NUM_40G;
if (old.link_status != link.link_status) {
nfp_net_dev_atomic_write_link_status(dev, &link);
nfp_dev_stats.obytes -= hw->eth_stats_base.obytes;
- nfp_dev_stats.imcasts =
- nn_cfg_readq(hw, NFP_NET_CFG_STATS_RX_MC_FRAMES);
-
- nfp_dev_stats.imcasts -= hw->eth_stats_base.imcasts;
-
/* reading general device stats */
nfp_dev_stats.ierrors =
nn_cfg_readq(hw, NFP_NET_CFG_STATS_RX_ERRORS);
nfp_dev_stats.oerrors -= hw->eth_stats_base.oerrors;
- /* Multicast frames received */
- nfp_dev_stats.imcasts =
- nn_cfg_readq(hw, NFP_NET_CFG_STATS_RX_MC_FRAMES);
-
- nfp_dev_stats.imcasts -= hw->eth_stats_base.imcasts;
-
/* RX ring mbuf allocation failures */
nfp_dev_stats.rx_nombuf = dev->data->rx_mbuf_alloc_failed;
hw->eth_stats_base.obytes =
nn_cfg_readq(hw, NFP_NET_CFG_STATS_TX_OCTETS);
- hw->eth_stats_base.imcasts =
- nn_cfg_readq(hw, NFP_NET_CFG_STATS_RX_MC_FRAMES);
-
/* reading general device stats */
hw->eth_stats_base.ierrors =
nn_cfg_readq(hw, NFP_NET_CFG_STATS_RX_ERRORS);
hw->eth_stats_base.oerrors =
nn_cfg_readq(hw, NFP_NET_CFG_STATS_TX_ERRORS);
- /* Multicast frames received */
- hw->eth_stats_base.imcasts =
- nn_cfg_readq(hw, NFP_NET_CFG_STATS_RX_MC_FRAMES);
-
/* RX ring mbuf allocation failures */
dev->data->rx_mbuf_alloc_failed = 0;
};
dev_info->reta_size = NFP_NET_CFG_RSS_ITBL_SZ;
-#if RTE_VER_MAJOR == 2 && RTE_VER_MINOR >= 1
dev_info->hash_key_size = NFP_NET_CFG_RSS_KEY_SZ;
-#endif
+
+ dev_info->speed_capa = ETH_LINK_SPEED_40G | ETH_LINK_SPEED_100G;
+}
+
+static const uint32_t *
+nfp_net_supported_ptypes_get(struct rte_eth_dev *dev)
+{
+ static const uint32_t ptypes[] = {
+ /* refers to nfp_net_set_hash() */
+ RTE_PTYPE_INNER_L3_IPV4,
+ RTE_PTYPE_INNER_L3_IPV6,
+ RTE_PTYPE_INNER_L3_IPV6_EXT,
+ RTE_PTYPE_INNER_L4_MASK,
+ RTE_PTYPE_UNKNOWN
+ };
+
+ if (dev->rx_pkt_burst == nfp_net_recv_pkts)
+ return ptypes;
+ return NULL;
}
static uint32_t
return count;
}
+static void
+nfp_net_dev_link_status_print(struct rte_eth_dev *dev)
+{
+ struct rte_eth_link link;
+
+ memset(&link, 0, sizeof(link));
+ nfp_net_dev_atomic_read_link_status(dev, &link);
+ if (link.link_status)
+ RTE_LOG(INFO, PMD, "Port %d: Link Up - speed %u Mbps - %s\n",
+ (int)(dev->data->port_id), (unsigned)link.link_speed,
+ link.link_duplex == ETH_LINK_FULL_DUPLEX
+ ? "full-duplex" : "half-duplex");
+ else
+ RTE_LOG(INFO, PMD, " Port %d: Link Down\n",
+ (int)(dev->data->port_id));
+
+ RTE_LOG(INFO, PMD, "PCI Address: %04d:%02d:%02d:%d\n",
+ dev->pci_dev->addr.domain, dev->pci_dev->addr.bus,
+ dev->pci_dev->addr.devid, dev->pci_dev->addr.function);
+}
+
+/* Interrupt configuration and handling */
+
+/*
+ * nfp_net_irq_unmask - Unmask an interrupt
+ *
+ * If MSI-X auto-masking is enabled clear the mask bit, otherwise
+ * clear the ICR for the entry.
+ */
+static void
+nfp_net_irq_unmask(struct rte_eth_dev *dev)
+{
+ struct nfp_net_hw *hw;
+
+ hw = NFP_NET_DEV_PRIVATE_TO_HW(dev->data->dev_private);
+
+ if (hw->ctrl & NFP_NET_CFG_CTRL_MSIXAUTO) {
+ /* If MSI-X auto-masking is used, clear the entry */
+ rte_wmb();
+ rte_intr_enable(&dev->pci_dev->intr_handle);
+ } else {
+ /* Make sure all updates are written before un-masking */
+ rte_wmb();
+ nn_cfg_writeb(hw, NFP_NET_CFG_ICR(NFP_NET_IRQ_LSC_IDX),
+ NFP_NET_CFG_ICR_UNMASKED);
+ }
+}
+
+static void
+nfp_net_dev_interrupt_handler(__rte_unused struct rte_intr_handle *handle,
+ void *param)
+{
+ int64_t timeout;
+ struct rte_eth_link link;
+ struct rte_eth_dev *dev = (struct rte_eth_dev *)param;
+
+ PMD_DRV_LOG(DEBUG, "We got a LSC interrupt!!!\n");
+
+ /* get the link status */
+ memset(&link, 0, sizeof(link));
+ nfp_net_dev_atomic_read_link_status(dev, &link);
+
+ nfp_net_link_update(dev, 0);
+
+ /* likely to up */
+ if (!link.link_status) {
+ /* handle it 1 sec later, wait it being stable */
+ timeout = NFP_NET_LINK_UP_CHECK_TIMEOUT;
+ /* likely to down */
+ } else {
+ /* handle it 4 sec later, wait it being stable */
+ timeout = NFP_NET_LINK_DOWN_CHECK_TIMEOUT;
+ }
+
+ if (rte_eal_alarm_set(timeout * 1000,
+ nfp_net_dev_interrupt_delayed_handler,
+ (void *)dev) < 0) {
+ RTE_LOG(ERR, PMD, "Error setting alarm");
+ /* Unmasking */
+ nfp_net_irq_unmask(dev);
+ }
+}
+
+/*
+ * Interrupt handler which shall be registered for alarm callback for delayed
+ * handling specific interrupt to wait for the stable nic state. As the NIC
+ * interrupt state is not stable for nfp after link is just down, it needs
+ * to wait 4 seconds to get the stable status.
+ *
+ * @param handle Pointer to interrupt handle.
+ * @param param The address of parameter (struct rte_eth_dev *)
+ *
+ * @return void
+ */
+static void
+nfp_net_dev_interrupt_delayed_handler(void *param)
+{
+ struct rte_eth_dev *dev = (struct rte_eth_dev *)param;
+
+ nfp_net_link_update(dev, 0);
+ _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC);
+
+ nfp_net_dev_link_status_print(dev);
+
+ /* Unmasking */
+ nfp_net_irq_unmask(dev);
+}
+
static int
nfp_net_dev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu)
{
(nb_desc > NFP_NET_MAX_RX_DESC) ||
(nb_desc < NFP_NET_MIN_RX_DESC)) {
RTE_LOG(ERR, PMD, "Wrong nb_desc value\n");
- return (-EINVAL);
+ return -EINVAL;
}
/*
rxq = rte_zmalloc_socket("ethdev RX queue", sizeof(struct nfp_net_rxq),
RTE_CACHE_LINE_SIZE, socket_id);
if (rxq == NULL)
- return (-ENOMEM);
+ return -ENOMEM;
/* Hw queues mapping based on firmware confifguration */
rxq->qidx = queue_idx;
if (tz == NULL) {
RTE_LOG(ERR, PMD, "Error allocatig rx dma\n");
nfp_net_rx_queue_release(rxq);
- return (-ENOMEM);
+ return -ENOMEM;
}
/* Saving physical and virtual addresses for the RX ring */
RTE_CACHE_LINE_SIZE, socket_id);
if (rxq->rxbufs == NULL) {
nfp_net_rx_queue_release(rxq);
- return (-ENOMEM);
+ return -ENOMEM;
}
PMD_RX_LOG(DEBUG, "rxbufs=%p hw_ring=%p dma_addr=0x%" PRIx64 "\n",
if (mbuf == NULL) {
RTE_LOG(ERR, PMD, "RX mbuf alloc failed queue_id=%u\n",
(unsigned)rxq->qidx);
- return (-ENOMEM);
+ return -ENOMEM;
}
dma_addr = rte_cpu_to_le_64(RTE_MBUF_DMA_ADDR_DEFAULT(mbuf));
RTE_CACHE_LINE_SIZE, socket_id);
if (txq == NULL) {
RTE_LOG(ERR, PMD, "Error allocating tx dma\n");
- return (-ENOMEM);
+ return -ENOMEM;
}
/*
if (tz == NULL) {
RTE_LOG(ERR, PMD, "Error allocating tx dma\n");
nfp_net_tx_queue_release(txq);
- return (-ENOMEM);
+ return -ENOMEM;
}
txq->tx_count = nb_desc;
RTE_CACHE_LINE_SIZE, socket_id);
if (txq->txbufs == NULL) {
nfp_net_tx_queue_release(txq);
- return (-ENOMEM);
+ return -ENOMEM;
}
PMD_TX_LOG(DEBUG, "txbufs=%p hw_ring=%p dma_addr=0x%" PRIx64 "\n",
txq->txbufs, txq->txds, (unsigned long int)txq->dma);
nfp_net_tx_cksum(struct nfp_net_txq *txq, struct nfp_net_tx_desc *txd,
struct rte_mbuf *mb)
{
- uint16_t ol_flags;
+ uint64_t ol_flags;
struct nfp_net_hw *hw = txq->hw;
if (!(hw->cap & NFP_NET_CFG_CTRL_TXCSUM))
break;
}
- txd->flags |= PCIE_DESC_TX_CSUM;
+ if (ol_flags & (PKT_TX_IP_CKSUM | PKT_TX_L4_MASK))
+ txd->flags |= PCIE_DESC_TX_CSUM;
}
/* nfp_net_rx_cksum - set mbuf checksum flags based on RX descriptor flags */
rte_pktmbuf_free_seg(*lmbuf);
dma_size = pkt->data_len;
- dma_addr = RTE_MBUF_DATA_DMA_ADDR(pkt);
+ dma_addr = rte_mbuf_data_dma_addr(pkt);
PMD_TX_LOG(DEBUG, "Working with mbuf at dma address:"
"%" PRIx64 "\n", dma_addr);
}
/* Initialise and register driver with DPDK Application */
-static struct eth_dev_ops nfp_net_eth_dev_ops = {
+static const struct eth_dev_ops nfp_net_eth_dev_ops = {
.dev_configure = nfp_net_configure,
.dev_start = nfp_net_start,
.dev_stop = nfp_net_stop,
.stats_get = nfp_net_stats_get,
.stats_reset = nfp_net_stats_reset,
.dev_infos_get = nfp_net_infos_get,
+ .dev_supported_ptypes_get = nfp_net_supported_ptypes_get,
.mtu_set = nfp_net_dev_mtu_set,
.vlan_offload_set = nfp_net_vlan_offload_set,
.reta_update = nfp_net_reta_update,
return 0;
pci_dev = eth_dev->pci_dev;
+ rte_eth_copy_pci_info(eth_dev, pci_dev);
+
hw->device_id = pci_dev->id.device_id;
hw->vendor_id = pci_dev->id.vendor_id;
hw->subsystem_device_id = pci_dev->id.subsystem_device_id;
PMD_INIT_LOG(INFO, "max_rx_queues: %u, max_tx_queues: %u\n",
hw->max_rx_queues, hw->max_tx_queues);
+ /* Initializing spinlock for reconfigs */
+ rte_spinlock_init(&hw->reconfig_lock);
+
/* Allocating memory for mac addr */
eth_dev->data->mac_addrs = rte_zmalloc("mac_addr", ETHER_ADDR_LEN, 0);
if (eth_dev->data->mac_addrs == NULL) {
hw->mac_addr[0], hw->mac_addr[1], hw->mac_addr[2],
hw->mac_addr[3], hw->mac_addr[4], hw->mac_addr[5]);
+ /* Registering LSC interrupt handler */
+ rte_intr_callback_register(&pci_dev->intr_handle,
+ nfp_net_dev_interrupt_handler,
+ (void *)eth_dev);
+
+ /* enable uio intr after callback register */
+ rte_intr_enable(&pci_dev->intr_handle);
+
+ /* Telling the firmware about the LSC interrupt entry */
+ nn_cfg_writeb(hw, NFP_NET_CFG_LSC, NFP_NET_IRQ_LSC_IDX);
+
/* Recording current stats counters values */
nfp_net_stats_reset(eth_dev);