#include <assert.h>
#include <stdint.h>
#include <stdlib.h>
+#include <errno.h>
#include <net/if.h>
/* Verbs header. */
/* ISO C doesn't support unnamed structs/unions, disabling -pedantic. */
#ifdef PEDANTIC
-#pragma GCC diagnostic ignored "-pedantic"
+#pragma GCC diagnostic ignored "-Wpedantic"
#endif
#include <infiniband/verbs.h>
#ifdef PEDANTIC
-#pragma GCC diagnostic error "-pedantic"
+#pragma GCC diagnostic error "-Wpedantic"
#endif
/* DPDK headers don't like -pedantic. */
#ifdef PEDANTIC
-#pragma GCC diagnostic ignored "-pedantic"
+#pragma GCC diagnostic ignored "-Wpedantic"
#endif
#include <rte_malloc.h>
#include <rte_ethdev.h>
#include <rte_pci.h>
#include <rte_common.h>
+#include <rte_kvargs.h>
#ifdef PEDANTIC
-#pragma GCC diagnostic error "-pedantic"
+#pragma GCC diagnostic error "-Wpedantic"
#endif
#include "mlx5.h"
#include "mlx5_autoconf.h"
#include "mlx5_defs.h"
+/* Device parameter to enable RX completion queue compression. */
+#define MLX5_RXQ_CQE_COMP_EN "rxq_cqe_comp_en"
+
+/* Device parameter to configure inline send. */
+#define MLX5_TXQ_INLINE "txq_inline"
+
+/*
+ * Device parameter to configure the number of TX queues threshold for
+ * enabling inline send.
+ */
+#define MLX5_TXQS_MIN_INLINE "txqs_min_inline"
+
+/* Device parameter to enable multi-packet send WQEs. */
+#define MLX5_TXQ_MPW_EN "txq_mpw_en"
+
+/**
+ * Retrieve integer value from environment variable.
+ *
+ * @param[in] name
+ * Environment variable name.
+ *
+ * @return
+ * Integer value, 0 if the variable is not set.
+ */
+int
+mlx5_getenv_int(const char *name)
+{
+ const char *val = getenv(name);
+
+ if (val == NULL)
+ return 0;
+ return atoi(val);
+}
+
/**
* DPDK callback to close the device.
*
static void
mlx5_dev_close(struct rte_eth_dev *dev)
{
- struct priv *priv = dev->data->dev_private;
- void *tmp;
+ struct priv *priv = mlx5_get_priv(dev);
unsigned int i;
priv_lock(priv);
((priv->ctx != NULL) ? priv->ctx->device->name : ""));
/* In case mlx5_dev_stop() has not been called. */
priv_dev_interrupt_handler_uninstall(priv, dev);
- priv_special_flow_disable(priv, HASH_RXQ_FLOW_TYPE_ALLMULTI);
- priv_special_flow_disable(priv, HASH_RXQ_FLOW_TYPE_PROMISC);
- priv_special_flow_disable(priv, HASH_RXQ_FLOW_TYPE_BROADCAST);
- priv_special_flow_disable(priv, HASH_RXQ_FLOW_TYPE_IPV6MULTI);
+ priv_special_flow_disable_all(priv);
priv_mac_addrs_disable(priv);
priv_destroy_hash_rxqs(priv);
/* XXX race condition if mlx5_rx_burst() is still running. */
usleep(1000);
for (i = 0; (i != priv->rxqs_n); ++i) {
- tmp = (*priv->rxqs)[i];
- if (tmp == NULL)
+ struct rxq *rxq = (*priv->rxqs)[i];
+ struct rxq_ctrl *rxq_ctrl;
+
+ if (rxq == NULL)
continue;
+ rxq_ctrl = container_of(rxq, struct rxq_ctrl, rxq);
(*priv->rxqs)[i] = NULL;
- rxq_cleanup(tmp);
- rte_free(tmp);
+ rxq_cleanup(rxq_ctrl);
+ rte_free(rxq_ctrl);
}
priv->rxqs_n = 0;
priv->rxqs = NULL;
/* XXX race condition if mlx5_tx_burst() is still running. */
usleep(1000);
for (i = 0; (i != priv->txqs_n); ++i) {
- tmp = (*priv->txqs)[i];
- if (tmp == NULL)
+ struct txq *txq = (*priv->txqs)[i];
+ struct txq_ctrl *txq_ctrl;
+
+ if (txq == NULL)
continue;
+ txq_ctrl = container_of(txq, struct txq_ctrl, txq);
(*priv->txqs)[i] = NULL;
- txq_cleanup(tmp);
- rte_free(tmp);
+ txq_cleanup(txq_ctrl);
+ rte_free(txq_ctrl);
}
priv->txqs_n = 0;
priv->txqs = NULL;
.dev_configure = mlx5_dev_configure,
.dev_start = mlx5_dev_start,
.dev_stop = mlx5_dev_stop,
+ .dev_set_link_down = mlx5_set_link_down,
+ .dev_set_link_up = mlx5_set_link_up,
.dev_close = mlx5_dev_close,
.promiscuous_enable = mlx5_promiscuous_enable,
.promiscuous_disable = mlx5_promiscuous_disable,
.link_update = mlx5_link_update,
.stats_get = mlx5_stats_get,
.stats_reset = mlx5_stats_reset,
+ .xstats_get = mlx5_xstats_get,
+ .xstats_reset = mlx5_xstats_reset,
+ .xstats_get_names = mlx5_xstats_get_names,
.dev_infos_get = mlx5_dev_infos_get,
+ .dev_supported_ptypes_get = mlx5_dev_supported_ptypes_get,
.vlan_filter_set = mlx5_vlan_filter_set,
.rx_queue_setup = mlx5_rx_queue_setup,
.tx_queue_setup = mlx5_tx_queue_setup,
.mac_addr_add = mlx5_mac_addr_add,
.mac_addr_set = mlx5_mac_addr_set,
.mtu_set = mlx5_dev_set_mtu,
+ .vlan_strip_queue_set = mlx5_vlan_strip_queue_set,
+ .vlan_offload_set = mlx5_vlan_offload_set,
.reta_update = mlx5_dev_rss_reta_update,
.reta_query = mlx5_dev_rss_reta_query,
.rss_hash_update = mlx5_rss_hash_update,
.rss_hash_conf_get = mlx5_rss_hash_conf_get,
-#ifdef MLX5_FDIR_SUPPORT
.filter_ctrl = mlx5_dev_filter_ctrl,
-#endif /* MLX5_FDIR_SUPPORT */
+ .rx_descriptor_status = mlx5_rx_descriptor_status,
+ .tx_descriptor_status = mlx5_tx_descriptor_status,
};
static struct {
return ret;
}
+/**
+ * Verify and store value for device argument.
+ *
+ * @param[in] key
+ * Key argument to verify.
+ * @param[in] val
+ * Value associated with key.
+ * @param opaque
+ * User data.
+ *
+ * @return
+ * 0 on success, negative errno value on failure.
+ */
+static int
+mlx5_args_check(const char *key, const char *val, void *opaque)
+{
+ struct priv *priv = opaque;
+ unsigned long tmp;
+
+ errno = 0;
+ tmp = strtoul(val, NULL, 0);
+ if (errno) {
+ WARN("%s: \"%s\" is not a valid integer", key, val);
+ return errno;
+ }
+ if (strcmp(MLX5_RXQ_CQE_COMP_EN, key) == 0) {
+ priv->cqe_comp = !!tmp;
+ } else if (strcmp(MLX5_TXQ_INLINE, key) == 0) {
+ priv->txq_inline = tmp;
+ } else if (strcmp(MLX5_TXQS_MIN_INLINE, key) == 0) {
+ priv->txqs_inline = tmp;
+ } else if (strcmp(MLX5_TXQ_MPW_EN, key) == 0) {
+ priv->mps &= !!tmp; /* Enable MPW only if HW supports */
+ } else {
+ WARN("%s: unknown parameter", key);
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * Parse device parameters.
+ *
+ * @param priv
+ * Pointer to private structure.
+ * @param devargs
+ * Device arguments structure.
+ *
+ * @return
+ * 0 on success, errno value on failure.
+ */
+static int
+mlx5_args(struct priv *priv, struct rte_devargs *devargs)
+{
+ const char **params = (const char *[]){
+ MLX5_RXQ_CQE_COMP_EN,
+ MLX5_TXQ_INLINE,
+ MLX5_TXQS_MIN_INLINE,
+ MLX5_TXQ_MPW_EN,
+ NULL,
+ };
+ struct rte_kvargs *kvlist;
+ int ret = 0;
+ int i;
+
+ if (devargs == NULL)
+ return 0;
+ /* Following UGLY cast is done to pass checkpatch. */
+ kvlist = rte_kvargs_parse(devargs->args, params);
+ if (kvlist == NULL)
+ return 0;
+ /* Process parameters. */
+ for (i = 0; (params[i] != NULL); ++i) {
+ if (rte_kvargs_count(kvlist, params[i])) {
+ ret = rte_kvargs_process(kvlist, params[i],
+ mlx5_args_check, priv);
+ if (ret != 0) {
+ rte_kvargs_free(kvlist);
+ return ret;
+ }
+ }
+ }
+ rte_kvargs_free(kvlist);
+ return 0;
+}
+
static struct eth_driver mlx5_driver;
/**
* 0 on success, negative errno value on failure.
*/
static int
-mlx5_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
+mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
{
struct ibv_device **list;
struct ibv_device *ibv_dev;
int err = 0;
struct ibv_context *attr_ctx = NULL;
struct ibv_device_attr device_attr;
- unsigned int vf;
+ unsigned int sriov;
+ unsigned int mps;
int idx;
int i;
(pci_dev->addr.devid != pci_addr.devid) ||
(pci_dev->addr.function != pci_addr.function))
continue;
- vf = ((pci_dev->id.device_id ==
+ sriov = ((pci_dev->id.device_id ==
PCI_DEVICE_ID_MELLANOX_CONNECTX4VF) ||
(pci_dev->id.device_id ==
- PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF));
- INFO("PCI information matches, using device \"%s\" (VF: %s)",
- list[i]->name, (vf ? "true" : "false"));
+ PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF) ||
+ (pci_dev->id.device_id ==
+ PCI_DEVICE_ID_MELLANOX_CONNECTX5VF) ||
+ (pci_dev->id.device_id ==
+ PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF));
+ /*
+ * Multi-packet send is supported by ConnectX-4 Lx PF as well
+ * as all ConnectX-5 devices.
+ */
+ switch (pci_dev->id.device_id) {
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX4LX:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5VF:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5EX:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF:
+ mps = 1;
+ break;
+ default:
+ mps = 0;
+ }
+ INFO("PCI information matches, using device \"%s\""
+ " (SR-IOV: %s, MPS: %s)",
+ list[i]->name,
+ sriov ? "true" : "false",
+ mps ? "true" : "false");
attr_ctx = ibv_open_device(list[i]);
err = errno;
break;
struct ibv_pd *pd = NULL;
struct priv *priv = NULL;
struct rte_eth_dev *eth_dev;
-#ifdef HAVE_EXP_QUERY_DEVICE
struct ibv_exp_device_attr exp_device_attr;
-#endif /* HAVE_EXP_QUERY_DEVICE */
struct ether_addr mac;
+ uint16_t num_vfs = 0;
-#ifdef HAVE_EXP_QUERY_DEVICE
exp_device_attr.comp_mask =
IBV_EXP_DEVICE_ATTR_EXP_CAP_FLAGS |
- IBV_EXP_DEVICE_ATTR_RX_HASH;
-#endif /* HAVE_EXP_QUERY_DEVICE */
+ IBV_EXP_DEVICE_ATTR_RX_HASH |
+ IBV_EXP_DEVICE_ATTR_VLAN_OFFLOADS |
+ IBV_EXP_DEVICE_ATTR_RX_PAD_END_ALIGN |
+ 0;
DEBUG("using port %u (%08" PRIx32 ")", port, test);
ERROR("port query failed: %s", strerror(err));
goto port_error;
}
+
+ if (port_attr.link_layer != IBV_LINK_LAYER_ETHERNET) {
+ ERROR("port %d is not configured in Ethernet mode",
+ port);
+ goto port_error;
+ }
+
if (port_attr.state != IBV_PORT_ACTIVE)
DEBUG("port %d is not active: \"%s\" (%d)",
port, ibv_port_state_str(port_attr.state),
priv->port = port;
priv->pd = pd;
priv->mtu = ETHER_MTU;
-#ifdef HAVE_EXP_QUERY_DEVICE
+ priv->mps = mps; /* Enable MPW by default if supported. */
+ priv->cqe_comp = 1; /* Enable compression by default. */
+ err = mlx5_args(priv, pci_dev->device.devargs);
+ if (err) {
+ ERROR("failed to process device arguments: %s",
+ strerror(err));
+ goto port_error;
+ }
if (ibv_exp_query_device(ctx, &exp_device_attr)) {
ERROR("ibv_exp_query_device() failed");
goto port_error;
priv->ind_table_max_size = exp_device_attr.rx_hash_caps.max_rwq_indirection_table_size;
/* Remove this check once DPDK supports larger/variable
* indirection tables. */
- if (priv->ind_table_max_size > (unsigned int)RSS_INDIRECTION_TABLE_SIZE)
- priv->ind_table_max_size = RSS_INDIRECTION_TABLE_SIZE;
+ if (priv->ind_table_max_size >
+ (unsigned int)ETH_RSS_RETA_SIZE_512)
+ priv->ind_table_max_size = ETH_RSS_RETA_SIZE_512;
DEBUG("maximum RX indirection table size is %u",
priv->ind_table_max_size);
-
-#else /* HAVE_EXP_QUERY_DEVICE */
- priv->ind_table_max_size = RSS_INDIRECTION_TABLE_SIZE;
-#endif /* HAVE_EXP_QUERY_DEVICE */
-
- priv->vf = vf;
+ priv->hw_vlan_strip = !!(exp_device_attr.wq_vlan_offloads_cap &
+ IBV_EXP_RECEIVE_WQ_CVLAN_STRIP);
+ DEBUG("VLAN stripping is %ssupported",
+ (priv->hw_vlan_strip ? "" : "not "));
+
+ priv->hw_fcs_strip = !!(exp_device_attr.exp_device_cap_flags &
+ IBV_EXP_DEVICE_SCATTER_FCS);
+ DEBUG("FCS stripping configuration is %ssupported",
+ (priv->hw_fcs_strip ? "" : "not "));
+
+ priv->hw_padding = !!exp_device_attr.rx_pad_end_addr_align;
+ DEBUG("hardware RX end alignment padding is %ssupported",
+ (priv->hw_padding ? "" : "not "));
+
+ priv_get_num_vfs(priv, &num_vfs);
+ priv->sriov = (num_vfs || sriov);
+ if (priv->mps && !mps) {
+ ERROR("multi-packet send not supported on this device"
+ " (" MLX5_TXQ_MPW_EN ")");
+ err = ENOTSUP;
+ goto port_error;
+ }
/* Allocate and register default RSS hash keys. */
priv->rss_conf = rte_calloc(__func__, hash_rxq_init_n,
sizeof((*priv->rss_conf)[0]), 0);
snprintf(name, sizeof(name), "%s port %u",
ibv_get_device_name(ibv_dev), port);
- eth_dev = rte_eth_dev_allocate(name, RTE_ETH_DEV_PCI);
+ eth_dev = rte_eth_dev_allocate(name);
}
if (eth_dev == NULL) {
ERROR("can not allocate rte ethdev");
goto port_error;
}
- eth_dev->data->dev_private = priv;
- eth_dev->pci_dev = pci_dev;
+ /* Secondary processes have to use local storage for their
+ * private data as well as a copy of eth_dev->data, but this
+ * pointer must not be modified before burst functions are
+ * actually called. */
+ if (mlx5_is_secondary()) {
+ struct mlx5_secondary_data *sd =
+ &mlx5_secondary_data[eth_dev->data->port_id];
+ sd->primary_priv = eth_dev->data->dev_private;
+ if (sd->primary_priv == NULL) {
+ ERROR("no private data for port %u",
+ eth_dev->data->port_id);
+ err = EINVAL;
+ goto port_error;
+ }
+ sd->shared_dev_data = eth_dev->data;
+ rte_spinlock_init(&sd->lock);
+ memcpy(sd->data.name, sd->shared_dev_data->name,
+ sizeof(sd->data.name));
+ sd->data.dev_private = priv;
+ sd->data.rx_mbuf_alloc_failed = 0;
+ sd->data.mtu = ETHER_MTU;
+ sd->data.port_id = sd->shared_dev_data->port_id;
+ sd->data.mac_addrs = priv->mac;
+ eth_dev->tx_pkt_burst = mlx5_tx_burst_secondary_setup;
+ eth_dev->rx_pkt_burst = mlx5_rx_burst_secondary_setup;
+ } else {
+ eth_dev->data->dev_private = priv;
+ eth_dev->data->mac_addrs = priv->mac;
+ }
+ eth_dev->device = &pci_dev->device;
rte_eth_copy_pci_info(eth_dev, pci_dev);
-
eth_dev->driver = &mlx5_driver;
- eth_dev->data->rx_mbuf_alloc_failed = 0;
- eth_dev->data->mtu = ETHER_MTU;
-
priv->dev = eth_dev;
eth_dev->dev_ops = &mlx5_dev_ops;
- eth_dev->data->mac_addrs = priv->mac;
- TAILQ_INIT(ð_dev->link_intr_cbs);
/* Bring Ethernet device up. */
DEBUG("forcing Ethernet interface up");
priv_set_flags(priv, ~IFF_UP, IFF_UP);
+ mlx5_link_update(priv->dev, 1);
continue;
port_error:
- rte_free(priv->rss_conf);
- rte_free(priv);
+ if (priv) {
+ rte_free(priv->rss_conf);
+ rte_free(priv);
+ }
if (pd)
claim_zero(ibv_dealloc_pd(pd));
if (ctx)
static const struct rte_pci_id mlx5_pci_id_map[] = {
{
- .vendor_id = PCI_VENDOR_ID_MELLANOX,
- .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX4,
- .subsystem_vendor_id = PCI_ANY_ID,
- .subsystem_device_id = PCI_ANY_ID
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX4)
+ },
+ {
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX4VF)
+ },
+ {
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX4LX)
},
{
- .vendor_id = PCI_VENDOR_ID_MELLANOX,
- .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX4VF,
- .subsystem_vendor_id = PCI_ANY_ID,
- .subsystem_device_id = PCI_ANY_ID
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF)
},
{
- .vendor_id = PCI_VENDOR_ID_MELLANOX,
- .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX4LX,
- .subsystem_vendor_id = PCI_ANY_ID,
- .subsystem_device_id = PCI_ANY_ID
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX5)
},
{
- .vendor_id = PCI_VENDOR_ID_MELLANOX,
- .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF,
- .subsystem_vendor_id = PCI_ANY_ID,
- .subsystem_device_id = PCI_ANY_ID
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX5VF)
+ },
+ {
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX5EX)
+ },
+ {
+ RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX,
+ PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF)
},
{
.vendor_id = 0
static struct eth_driver mlx5_driver = {
.pci_drv = {
- .name = MLX5_DRIVER_NAME,
+ .driver = {
+ .name = MLX5_DRIVER_NAME
+ },
.id_table = mlx5_pci_id_map,
- .devinit = mlx5_pci_devinit,
+ .probe = mlx5_pci_probe,
.drv_flags = RTE_PCI_DRV_INTR_LSC,
},
.dev_private_size = sizeof(struct priv)
/**
* Driver initialization routine.
*/
-static int
-rte_mlx5_pmd_init(const char *name, const char *args)
+RTE_INIT(rte_mlx5_pmd_init);
+static void
+rte_mlx5_pmd_init(void)
{
- (void)name;
- (void)args;
/*
* RDMAV_HUGEPAGES_SAFE tells ibv_fork_init() we intend to use
* huge pages. Calling ibv_fork_init() during init allows
setenv("RDMAV_HUGEPAGES_SAFE", "1", 1);
ibv_fork_init();
rte_eal_pci_register(&mlx5_driver.pci_drv);
- return 0;
}
-static struct rte_driver rte_mlx5_driver = {
- .type = PMD_PDEV,
- .name = MLX5_DRIVER_NAME,
- .init = rte_mlx5_pmd_init,
-};
-
-PMD_REGISTER_DRIVER(rte_mlx5_driver)
+RTE_PMD_EXPORT_NAME(net_mlx5, __COUNTER__);
+RTE_PMD_REGISTER_PCI_TABLE(net_mlx5, mlx5_pci_id_map);
+RTE_PMD_REGISTER_KMOD_DEP(net_mlx5, "* ib_uverbs & mlx5_core & mlx5_ib");