#include <net/if.h>
#include <sys/ioctl.h>
#include <sys/socket.h>
+#include <sys/utsname.h>
#include <netinet/in.h>
#include <linux/ethtool.h>
#include <linux/sockios.h>
+#include <linux/version.h>
#include <fcntl.h>
/* DPDK headers don't like -pedantic. */
#include "mlx5_rxtx.h"
#include "mlx5_utils.h"
+/* Add defines in case the running kernel is not the same as user headers. */
+#ifndef ETHTOOL_GLINKSETTINGS
+struct ethtool_link_settings {
+ uint32_t cmd;
+ uint32_t speed;
+ uint8_t duplex;
+ uint8_t port;
+ uint8_t phy_address;
+ uint8_t autoneg;
+ uint8_t mdio_support;
+ uint8_t eth_to_mdix;
+ uint8_t eth_tp_mdix_ctrl;
+ int8_t link_mode_masks_nwords;
+ uint32_t reserved[8];
+ uint32_t link_mode_masks[];
+};
+
+#define ETHTOOL_GLINKSETTINGS 0x0000004c
+#define ETHTOOL_LINK_MODE_1000baseT_Full_BIT 5
+#define ETHTOOL_LINK_MODE_Autoneg_BIT 6
+#define ETHTOOL_LINK_MODE_1000baseKX_Full_BIT 17
+#define ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT 18
+#define ETHTOOL_LINK_MODE_10000baseKR_Full_BIT 19
+#define ETHTOOL_LINK_MODE_10000baseR_FEC_BIT 20
+#define ETHTOOL_LINK_MODE_20000baseMLD2_Full_BIT 21
+#define ETHTOOL_LINK_MODE_20000baseKR2_Full_BIT 22
+#define ETHTOOL_LINK_MODE_40000baseKR4_Full_BIT 23
+#define ETHTOOL_LINK_MODE_40000baseCR4_Full_BIT 24
+#define ETHTOOL_LINK_MODE_40000baseSR4_Full_BIT 25
+#define ETHTOOL_LINK_MODE_40000baseLR4_Full_BIT 26
+#define ETHTOOL_LINK_MODE_56000baseKR4_Full_BIT 27
+#define ETHTOOL_LINK_MODE_56000baseCR4_Full_BIT 28
+#define ETHTOOL_LINK_MODE_56000baseSR4_Full_BIT 29
+#define ETHTOOL_LINK_MODE_56000baseLR4_Full_BIT 30
+#endif
+#ifndef HAVE_ETHTOOL_LINK_MODE_25G
+#define ETHTOOL_LINK_MODE_25000baseCR_Full_BIT 31
+#define ETHTOOL_LINK_MODE_25000baseKR_Full_BIT 32
+#define ETHTOOL_LINK_MODE_25000baseSR_Full_BIT 33
+#endif
+#ifndef HAVE_ETHTOOL_LINK_MODE_50G
+#define ETHTOOL_LINK_MODE_50000baseCR2_Full_BIT 34
+#define ETHTOOL_LINK_MODE_50000baseKR2_Full_BIT 35
+#endif
+#ifndef HAVE_ETHTOOL_LINK_MODE_100G
+#define ETHTOOL_LINK_MODE_100000baseKR4_Full_BIT 36
+#define ETHTOOL_LINK_MODE_100000baseSR4_Full_BIT 37
+#define ETHTOOL_LINK_MODE_100000baseCR4_Full_BIT 38
+#define ETHTOOL_LINK_MODE_100000baseLR4_ER4_Full_BIT 39
+#endif
+#define ETHTOOL_LINK_MODE_MASK_MAX_KERNEL_NU32 (SCHAR_MAX)
+
/**
* Return private structure associated with an Ethernet device.
*
return 0;
}
+/**
+ * Check if the counter is located on ib counters file.
+ *
+ * @param[in] cntr
+ * Counter name.
+ *
+ * @return
+ * 1 if counter is located on ib counters file , 0 otherwise.
+ */
+int
+priv_is_ib_cntr(const char *cntr)
+{
+ if (!strcmp(cntr, "out_of_buffer"))
+ return 1;
+ return 0;
+}
+
/**
* Read from sysfs entry.
*
if (priv_get_ifname(priv, &ifname))
return -1;
- MKSTR(path, "%s/device/net/%s/%s", priv->ctx->device->ibdev_path,
- ifname, entry);
-
- file = fopen(path, "rb");
+ if (priv_is_ib_cntr(entry)) {
+ MKSTR(path, "%s/ports/1/hw_counters/%s",
+ priv->ctx->device->ibdev_path, entry);
+ file = fopen(path, "rb");
+ } else {
+ MKSTR(path, "%s/device/net/%s/%s",
+ priv->ctx->device->ibdev_path, ifname, entry);
+ file = fopen(path, "rb");
+ }
if (file == NULL)
return -1;
ret = fread(buf, 1, size, file);
return 0;
}
+/**
+ * Read device counter from sysfs.
+ *
+ * @param priv
+ * Pointer to private structure.
+ * @param name
+ * Counter name.
+ * @param[out] cntr
+ * Counter output buffer.
+ *
+ * @return
+ * 0 on success, -1 on failure and errno is set.
+ */
+int
+priv_get_cntr_sysfs(struct priv *priv, const char *name, uint64_t *cntr)
+{
+ unsigned long ulong_ctr;
+
+ if (priv_get_sysfs_ulong(priv, name, &ulong_ctr) == -1)
+ return -1;
+ *cntr = ulong_ctr;
+ return 0;
+}
+
/**
* Set device MTU.
*
unsigned int max;
char ifname[IF_NAMESIZE];
+ info->pci_dev = RTE_ETH_DEV_TO_PCI(dev);
+
priv_lock(priv);
/* FIXME: we should ask the device for these values. */
info->min_rx_bufsize = 32;
(DEV_TX_OFFLOAD_IPV4_CKSUM |
DEV_TX_OFFLOAD_UDP_CKSUM |
DEV_TX_OFFLOAD_TCP_CKSUM);
+ if (priv->tso)
+ info->tx_offload_capa |= DEV_TX_OFFLOAD_TCP_TSO;
+ if (priv->tunnel_en)
+ info->tx_offload_capa |= (DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM |
+ DEV_TX_OFFLOAD_VXLAN_TNL_TSO |
+ DEV_TX_OFFLOAD_GRE_TNL_TSO);
if (priv_get_ifname(priv, &ifname) == 0)
info->if_index = if_nametoindex(ifname);
- /* FIXME: RETA update/query API expects the callee to know the size of
- * the indirection table, for this PMD the size varies depending on
- * the number of RX queues, it becomes impossible to find the correct
- * size if it is not fixed.
- * The API should be updated to solve this problem. */
- info->reta_size = priv->ind_table_max_size;
+ info->reta_size = priv->reta_idx_n ?
+ priv->reta_idx_n : priv->ind_table_max_size;
info->hash_key_size = ((*priv->rss_conf) ?
(*priv->rss_conf)[0]->rss_key_len :
0);
{
static const uint32_t ptypes[] = {
/* refers to rxq_cq_to_pkt_type() */
- RTE_PTYPE_L3_IPV4,
- RTE_PTYPE_L3_IPV6,
- RTE_PTYPE_INNER_L3_IPV4,
- RTE_PTYPE_INNER_L3_IPV6,
+ RTE_PTYPE_L3_IPV4_EXT_UNKNOWN,
+ RTE_PTYPE_L3_IPV6_EXT_UNKNOWN,
+ RTE_PTYPE_INNER_L3_IPV4_EXT_UNKNOWN,
+ RTE_PTYPE_INNER_L3_IPV6_EXT_UNKNOWN,
RTE_PTYPE_UNKNOWN
};
- if (dev->rx_pkt_burst == mlx5_rx_burst)
+ if (dev->rx_pkt_burst == mlx5_rx_burst ||
+ dev->rx_pkt_burst == mlx5_rx_burst_vec)
return ptypes;
return NULL;
}
/**
- * DPDK callback to retrieve physical link information (unlocked version).
+ * DPDK callback to retrieve physical link information.
*
* @param dev
* Pointer to Ethernet device structure.
* @param wait_to_complete
* Wait for request completion (ignored).
*/
-int
-mlx5_link_update_unlocked(struct rte_eth_dev *dev, int wait_to_complete)
+static int
+mlx5_link_update_unlocked_gset(struct rte_eth_dev *dev, int wait_to_complete)
{
struct priv *priv = mlx5_get_priv(dev);
struct ethtool_cmd edata = {
struct rte_eth_link dev_link;
int link_speed = 0;
+ /* priv_lock() is not taken to allow concurrent calls. */
+
(void)wait_to_complete;
if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
return -1;
}
+/**
+ * Retrieve physical link information (unlocked version using new ioctl).
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param wait_to_complete
+ * Wait for request completion (ignored).
+ */
+static int
+mlx5_link_update_unlocked_gs(struct rte_eth_dev *dev, int wait_to_complete)
+{
+ struct priv *priv = mlx5_get_priv(dev);
+ __extension__ struct {
+ struct ethtool_link_settings edata;
+ uint32_t link_mode_data[3 *
+ ETHTOOL_LINK_MODE_MASK_MAX_KERNEL_NU32];
+ } ecmd;
+
+ struct ifreq ifr;
+ struct rte_eth_link dev_link;
+ uint64_t sc;
+
+ (void)wait_to_complete;
+ if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) {
+ WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno));
+ return -1;
+ }
+ memset(&dev_link, 0, sizeof(dev_link));
+ dev_link.link_status = ((ifr.ifr_flags & IFF_UP) &&
+ (ifr.ifr_flags & IFF_RUNNING));
+ memset(&ecmd, 0, sizeof(ecmd));
+ ecmd.edata.cmd = ETHTOOL_GLINKSETTINGS;
+ ifr.ifr_data = (void *)&ecmd;
+ if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+ DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
+ strerror(errno));
+ return -1;
+ }
+ ecmd.edata.link_mode_masks_nwords = -ecmd.edata.link_mode_masks_nwords;
+ if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) {
+ DEBUG("ioctl(SIOCETHTOOL, ETHTOOL_GLINKSETTINGS) failed: %s",
+ strerror(errno));
+ return -1;
+ }
+ dev_link.link_speed = ecmd.edata.speed;
+ sc = ecmd.edata.link_mode_masks[0] |
+ ((uint64_t)ecmd.edata.link_mode_masks[1] << 32);
+ priv->link_speed_capa = 0;
+ if (sc & ETHTOOL_LINK_MODE_Autoneg_BIT)
+ priv->link_speed_capa |= ETH_LINK_SPEED_AUTONEG;
+ if (sc & (ETHTOOL_LINK_MODE_1000baseT_Full_BIT |
+ ETHTOOL_LINK_MODE_1000baseKX_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_1G;
+ if (sc & (ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT |
+ ETHTOOL_LINK_MODE_10000baseKR_Full_BIT |
+ ETHTOOL_LINK_MODE_10000baseR_FEC_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_10G;
+ if (sc & (ETHTOOL_LINK_MODE_20000baseMLD2_Full_BIT |
+ ETHTOOL_LINK_MODE_20000baseKR2_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_20G;
+ if (sc & (ETHTOOL_LINK_MODE_40000baseKR4_Full_BIT |
+ ETHTOOL_LINK_MODE_40000baseCR4_Full_BIT |
+ ETHTOOL_LINK_MODE_40000baseSR4_Full_BIT |
+ ETHTOOL_LINK_MODE_40000baseLR4_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_40G;
+ if (sc & (ETHTOOL_LINK_MODE_56000baseKR4_Full_BIT |
+ ETHTOOL_LINK_MODE_56000baseCR4_Full_BIT |
+ ETHTOOL_LINK_MODE_56000baseSR4_Full_BIT |
+ ETHTOOL_LINK_MODE_56000baseLR4_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_56G;
+ if (sc & (ETHTOOL_LINK_MODE_25000baseCR_Full_BIT |
+ ETHTOOL_LINK_MODE_25000baseKR_Full_BIT |
+ ETHTOOL_LINK_MODE_25000baseSR_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_25G;
+ if (sc & (ETHTOOL_LINK_MODE_50000baseCR2_Full_BIT |
+ ETHTOOL_LINK_MODE_50000baseKR2_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_50G;
+ if (sc & (ETHTOOL_LINK_MODE_100000baseKR4_Full_BIT |
+ ETHTOOL_LINK_MODE_100000baseSR4_Full_BIT |
+ ETHTOOL_LINK_MODE_100000baseCR4_Full_BIT |
+ ETHTOOL_LINK_MODE_100000baseLR4_ER4_Full_BIT))
+ priv->link_speed_capa |= ETH_LINK_SPEED_100G;
+ dev_link.link_duplex = ((ecmd.edata.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 (memcmp(&dev_link, &dev->data->dev_link, sizeof(dev_link))) {
+ /* Link status changed. */
+ dev->data->dev_link = dev_link;
+ return 0;
+ }
+ /* Link status is still the same. */
+ return -1;
+}
+
/**
* DPDK callback to retrieve physical link information.
*
int
mlx5_link_update(struct rte_eth_dev *dev, int wait_to_complete)
{
- struct priv *priv = mlx5_get_priv(dev);
- int ret;
-
- priv_lock(priv);
- ret = mlx5_link_update_unlocked(dev, wait_to_complete);
- priv_unlock(priv);
- return ret;
+ struct utsname utsname;
+ int ver[3];
+
+ if (uname(&utsname) == -1 ||
+ sscanf(utsname.release, "%d.%d.%d",
+ &ver[0], &ver[1], &ver[2]) != 3 ||
+ KERNEL_VERSION(ver[0], ver[1], ver[2]) < KERNEL_VERSION(4, 9, 0))
+ return mlx5_link_update_unlocked_gset(dev, wait_to_complete);
+ return mlx5_link_update_unlocked_gs(dev, wait_to_complete);
}
/**
struct priv *priv = dev->data->dev_private;
int ret = 0;
unsigned int i;
- uint16_t (*rx_func)(void *, struct rte_mbuf **, uint16_t) =
- mlx5_rx_burst;
unsigned int max_frame_len;
int rehash;
int restart = priv->started;
/* Temporarily replace RX handler with a fake one, assuming it has not
* been copied elsewhere. */
dev->rx_pkt_burst = removed_rx_burst;
- /* Make sure everyone has left mlx5_rx_burst() and uses
+ /* Make sure everyone has left dev->rx_pkt_burst() and uses
* removed_rx_burst() instead. */
rte_wmb();
usleep(1000);
struct rxq *rxq = (*priv->rxqs)[i];
struct rxq_ctrl *rxq_ctrl =
container_of(rxq, struct rxq_ctrl, rxq);
- int sp;
unsigned int mb_len;
unsigned int tmp;
continue;
mb_len = rte_pktmbuf_data_room_size(rxq->mp);
assert(mb_len >= RTE_PKTMBUF_HEADROOM);
- /* Toggle scattered support (sp) if necessary. */
- sp = (max_frame_len > (mb_len - RTE_PKTMBUF_HEADROOM));
/* Provide new values to rxq_setup(). */
- dev->data->dev_conf.rxmode.jumbo_frame = sp;
+ dev->data->dev_conf.rxmode.jumbo_frame =
+ (max_frame_len > ETHER_MAX_LEN);
dev->data->dev_conf.rxmode.max_rx_pkt_len = max_frame_len;
if (rehash)
ret = rxq_rehash(dev, rxq_ctrl);
/* Double fault, disable RX. */
break;
}
- /*
- * Use a safe RX burst function in case of error, otherwise mimic
- * mlx5_dev_start().
- */
+ /* Mimic mlx5_dev_start(). */
if (ret) {
ERROR("unable to reconfigure RX queues, RX disabled");
- rx_func = removed_rx_burst;
} else if (restart &&
- !rehash &&
- !priv_create_hash_rxqs(priv) &&
- !priv_rehash_flows(priv)) {
+ !rehash &&
+ !priv_create_hash_rxqs(priv) &&
+ !priv_rehash_flows(priv)) {
if (dev->data->dev_conf.fdir_conf.mode == RTE_FDIR_MODE_NONE)
priv_fdir_enable(priv);
priv_dev_interrupt_handler_install(priv, dev);
priv->mtu = mtu;
/* Burst functions can now be called again. */
rte_wmb();
- dev->rx_pkt_burst = rx_func;
+ /*
+ * Use a safe RX burst function in case of error, otherwise select RX
+ * burst function again.
+ */
+ if (!ret)
+ priv_select_rx_function(priv);
out:
priv_unlock(priv);
assert(ret >= 0);
/* Extract information. */
if (sscanf(line,
"PCI_SLOT_NAME="
- "%" SCNx16 ":%" SCNx8 ":%" SCNx8 ".%" SCNx8 "\n",
+ "%" SCNx32 ":%" SCNx8 ":%" SCNx8 ".%" SCNx8 "\n",
&pci_addr->domain,
&pci_addr->bus,
&pci_addr->devid,
priv_dev_link_status_handler(struct priv *priv, struct rte_eth_dev *dev)
{
struct ibv_async_event event;
- int port_change = 0;
+ struct rte_eth_link *link = &dev->data->dev_link;
int ret = 0;
/* Read all message and acknowledge them. */
if (ibv_get_async_event(priv->ctx, &event))
break;
- if (event.event_type == IBV_EVENT_PORT_ACTIVE ||
- event.event_type == IBV_EVENT_PORT_ERR)
- port_change = 1;
- else
+ if (event.event_type != IBV_EVENT_PORT_ACTIVE &&
+ event.event_type != IBV_EVENT_PORT_ERR)
DEBUG("event type %d on port %d not handled",
event.event_type, event.element.port_num);
ibv_ack_async_event(&event);
}
-
- if (port_change ^ priv->pending_alarm) {
- struct rte_eth_link *link = &dev->data->dev_link;
-
- priv->pending_alarm = 0;
- mlx5_link_update_unlocked(dev, 0);
- if (((link->link_speed == 0) && link->link_status) ||
- ((link->link_speed != 0) && !link->link_status)) {
+ mlx5_link_update(dev, 0);
+ if (((link->link_speed == 0) && link->link_status) ||
+ ((link->link_speed != 0) && !link->link_status)) {
+ if (!priv->pending_alarm) {
/* Inconsistent status, check again later. */
priv->pending_alarm = 1;
rte_eal_alarm_set(MLX5_ALARM_TIMEOUT_US,
mlx5_dev_link_status_handler,
dev);
- } else
- ret = 1;
+ }
+ } else {
+ ret = 1;
}
return ret;
}
priv_lock(priv);
assert(priv->pending_alarm == 1);
+ priv->pending_alarm = 0;
ret = priv_dev_link_status_handler(priv, dev);
priv_unlock(priv);
if (ret)
- _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL);
+ _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
+ NULL);
}
/**
* Callback argument.
*/
void
-mlx5_dev_interrupt_handler(struct rte_intr_handle *intr_handle, void *cb_arg)
+mlx5_dev_interrupt_handler(void *cb_arg)
{
struct rte_eth_dev *dev = cb_arg;
struct priv *priv = dev->data->dev_private;
int ret;
- (void)intr_handle;
priv_lock(priv);
ret = priv_dev_link_status_handler(priv, dev);
priv_unlock(priv);
if (ret)
- _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL);
+ _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL,
+ NULL);
}
/**
priv_select_tx_function(struct priv *priv)
{
priv->dev->tx_pkt_burst = mlx5_tx_burst;
- /* Display warning for unsupported configurations. */
- if (priv->sriov && priv->mps)
- WARN("multi-packet send WQE cannot be used on a SR-IOV setup");
/* Select appropriate TX function. */
- if ((priv->sriov == 0) && priv->mps && priv->txq_inline) {
+ if (priv->mps == MLX5_MPW_ENHANCED) {
+ if (priv_check_vec_tx_support(priv) > 0) {
+ if (priv_check_raw_vec_tx_support(priv) > 0)
+ priv->dev->tx_pkt_burst = mlx5_tx_burst_raw_vec;
+ else
+ priv->dev->tx_pkt_burst = mlx5_tx_burst_vec;
+ DEBUG("selected Enhanced MPW TX vectorized function");
+ } else {
+ priv->dev->tx_pkt_burst = mlx5_tx_burst_empw;
+ DEBUG("selected Enhanced MPW TX function");
+ }
+ } else if (priv->mps && priv->txq_inline) {
priv->dev->tx_pkt_burst = mlx5_tx_burst_mpw_inline;
DEBUG("selected MPW inline TX function");
- } else if ((priv->sriov == 0) && priv->mps) {
+ } else if (priv->mps) {
priv->dev->tx_pkt_burst = mlx5_tx_burst_mpw;
DEBUG("selected MPW TX function");
}
void
priv_select_rx_function(struct priv *priv)
{
- priv->dev->rx_pkt_burst = mlx5_rx_burst;
+ if (priv_check_vec_rx_support(priv) > 0) {
+ priv_prep_vec_rx_function(priv);
+ priv->dev->rx_pkt_burst = mlx5_rx_burst_vec;
+ DEBUG("selected RX vectorized function");
+ } else {
+ priv->dev->rx_pkt_burst = mlx5_rx_burst;
+ }
}
git.droids-corp.org / dpdk.git / blobdiff