X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fvhost%2Frte_eth_vhost.c;h=848a3daff6fefcba7aebdae5e146b93d8f79dd02;hb=73db5badb042c2ba668cdc096168932c5d5e692d;hp=4ee10c1aa1bbb4edde23ae6a0c7a2fb0fd61cc69;hpb=ee584e9710b9abd60ee9faef664e106dcea10085;p=dpdk.git diff --git a/drivers/net/vhost/rte_eth_vhost.c b/drivers/net/vhost/rte_eth_vhost.c index 4ee10c1aa1..848a3daff6 100644 --- a/drivers/net/vhost/rte_eth_vhost.c +++ b/drivers/net/vhost/rte_eth_vhost.c @@ -41,7 +41,7 @@ #include #include #include -#include +#include #include #include #include @@ -50,12 +50,14 @@ #define ETH_VHOST_IFACE_ARG "iface" #define ETH_VHOST_QUEUES_ARG "queues" - -static const char *drivername = "VHOST PMD"; +#define ETH_VHOST_CLIENT_ARG "client" +#define ETH_VHOST_DEQUEUE_ZERO_COPY "dequeue-zero-copy" static const char *valid_arguments[] = { ETH_VHOST_IFACE_ARG, ETH_VHOST_QUEUES_ARG, + ETH_VHOST_CLIENT_ARG, + ETH_VHOST_DEQUEUE_ZERO_COPY, NULL }; @@ -70,26 +72,49 @@ static struct ether_addr base_eth_addr = { } }; +enum vhost_xstats_pkts { + VHOST_UNDERSIZE_PKT = 0, + VHOST_64_PKT, + VHOST_65_TO_127_PKT, + VHOST_128_TO_255_PKT, + VHOST_256_TO_511_PKT, + VHOST_512_TO_1023_PKT, + VHOST_1024_TO_1522_PKT, + VHOST_1523_TO_MAX_PKT, + VHOST_BROADCAST_PKT, + VHOST_MULTICAST_PKT, + VHOST_UNICAST_PKT, + VHOST_ERRORS_PKT, + VHOST_ERRORS_FRAGMENTED, + VHOST_ERRORS_JABBER, + VHOST_UNKNOWN_PROTOCOL, + VHOST_XSTATS_MAX, +}; + +struct vhost_stats { + uint64_t pkts; + uint64_t bytes; + uint64_t missed_pkts; + uint64_t xstats[VHOST_XSTATS_MAX]; +}; + struct vhost_queue { + int vid; rte_atomic32_t allow_queuing; rte_atomic32_t while_queuing; - struct virtio_net *device; struct pmd_internal *internal; struct rte_mempool *mb_pool; uint8_t port; uint16_t virtqueue_id; - uint64_t rx_pkts; - uint64_t tx_pkts; - uint64_t missed_pkts; - uint64_t rx_bytes; - uint64_t tx_bytes; + struct vhost_stats stats; }; struct pmd_internal { + rte_atomic32_t dev_attached; char *dev_name; char *iface_name; - - volatile uint16_t once; + uint16_t max_queues; + rte_atomic32_t started; }; struct internal_list { @@ -109,7 +134,7 @@ static pthread_t session_th; static struct rte_eth_link pmd_link = { .link_speed = 10000, .link_duplex = ETH_LINK_FULL_DUPLEX, - .link_status = 0 + .link_status = ETH_LINK_DOWN }; struct rte_vhost_vring_state { @@ -123,6 +148,244 @@ struct rte_vhost_vring_state { static struct rte_vhost_vring_state *vring_states[RTE_MAX_ETHPORTS]; +#define VHOST_XSTATS_NAME_SIZE 64 + +struct vhost_xstats_name_off { + char name[VHOST_XSTATS_NAME_SIZE]; + uint64_t offset; +}; + +/* [rx]_is prepended to the name string here */ +static const struct vhost_xstats_name_off vhost_rxport_stat_strings[] = { + {"good_packets", + offsetof(struct vhost_queue, stats.pkts)}, + {"total_bytes", + offsetof(struct vhost_queue, stats.bytes)}, + {"missed_pkts", + offsetof(struct vhost_queue, stats.missed_pkts)}, + {"broadcast_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_BROADCAST_PKT])}, + {"multicast_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_MULTICAST_PKT])}, + {"unicast_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_UNICAST_PKT])}, + {"undersize_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_UNDERSIZE_PKT])}, + {"size_64_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_64_PKT])}, + {"size_65_to_127_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_65_TO_127_PKT])}, + {"size_128_to_255_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_128_TO_255_PKT])}, + {"size_256_to_511_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_256_TO_511_PKT])}, + {"size_512_to_1023_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_512_TO_1023_PKT])}, + {"size_1024_to_1522_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_1024_TO_1522_PKT])}, + {"size_1523_to_max_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_1523_TO_MAX_PKT])}, + {"errors_with_bad_CRC", + offsetof(struct vhost_queue, stats.xstats[VHOST_ERRORS_PKT])}, + {"fragmented_errors", + offsetof(struct vhost_queue, stats.xstats[VHOST_ERRORS_FRAGMENTED])}, + {"jabber_errors", + offsetof(struct vhost_queue, stats.xstats[VHOST_ERRORS_JABBER])}, + {"unknown_protos_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_UNKNOWN_PROTOCOL])}, +}; + +/* [tx]_ is prepended to the name string here */ +static const struct vhost_xstats_name_off vhost_txport_stat_strings[] = { + {"good_packets", + offsetof(struct vhost_queue, stats.pkts)}, + {"total_bytes", + offsetof(struct vhost_queue, stats.bytes)}, + {"missed_pkts", + offsetof(struct vhost_queue, stats.missed_pkts)}, + {"broadcast_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_BROADCAST_PKT])}, + {"multicast_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_MULTICAST_PKT])}, + {"unicast_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_UNICAST_PKT])}, + {"undersize_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_UNDERSIZE_PKT])}, + {"size_64_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_64_PKT])}, + {"size_65_to_127_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_65_TO_127_PKT])}, + {"size_128_to_255_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_128_TO_255_PKT])}, + {"size_256_to_511_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_256_TO_511_PKT])}, + {"size_512_to_1023_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_512_TO_1023_PKT])}, + {"size_1024_to_1522_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_1024_TO_1522_PKT])}, + {"size_1523_to_max_packets", + offsetof(struct vhost_queue, stats.xstats[VHOST_1523_TO_MAX_PKT])}, + {"errors_with_bad_CRC", + offsetof(struct vhost_queue, stats.xstats[VHOST_ERRORS_PKT])}, +}; + +#define VHOST_NB_XSTATS_RXPORT (sizeof(vhost_rxport_stat_strings) / \ + sizeof(vhost_rxport_stat_strings[0])) + +#define VHOST_NB_XSTATS_TXPORT (sizeof(vhost_txport_stat_strings) / \ + sizeof(vhost_txport_stat_strings[0])) + +static void +vhost_dev_xstats_reset(struct rte_eth_dev *dev) +{ + struct vhost_queue *vq = NULL; + unsigned int i = 0; + + for (i = 0; i < dev->data->nb_rx_queues; i++) { + vq = dev->data->rx_queues[i]; + if (!vq) + continue; + memset(&vq->stats, 0, sizeof(vq->stats)); + } + for (i = 0; i < dev->data->nb_tx_queues; i++) { + vq = dev->data->tx_queues[i]; + if (!vq) + continue; + memset(&vq->stats, 0, sizeof(vq->stats)); + } +} + +static int +vhost_dev_xstats_get_names(struct rte_eth_dev *dev __rte_unused, + struct rte_eth_xstat_name *xstats_names, + unsigned int limit __rte_unused) +{ + unsigned int t = 0; + int count = 0; + int nstats = VHOST_NB_XSTATS_RXPORT + VHOST_NB_XSTATS_TXPORT; + + if (!xstats_names) + return nstats; + for (t = 0; t < VHOST_NB_XSTATS_RXPORT; t++) { + snprintf(xstats_names[count].name, + sizeof(xstats_names[count].name), + "rx_%s", vhost_rxport_stat_strings[t].name); + count++; + } + for (t = 0; t < VHOST_NB_XSTATS_TXPORT; t++) { + snprintf(xstats_names[count].name, + sizeof(xstats_names[count].name), + "tx_%s", vhost_txport_stat_strings[t].name); + count++; + } + return count; +} + +static int +vhost_dev_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *xstats, + unsigned int n) +{ + unsigned int i; + unsigned int t; + unsigned int count = 0; + struct vhost_queue *vq = NULL; + unsigned int nxstats = VHOST_NB_XSTATS_RXPORT + VHOST_NB_XSTATS_TXPORT; + + if (n < nxstats) + return nxstats; + + for (i = 0; i < dev->data->nb_rx_queues; i++) { + vq = dev->data->rx_queues[i]; + if (!vq) + continue; + vq->stats.xstats[VHOST_UNICAST_PKT] = vq->stats.pkts + - (vq->stats.xstats[VHOST_BROADCAST_PKT] + + vq->stats.xstats[VHOST_MULTICAST_PKT]); + } + for (i = 0; i < dev->data->nb_tx_queues; i++) { + vq = dev->data->tx_queues[i]; + if (!vq) + continue; + vq->stats.xstats[VHOST_UNICAST_PKT] = vq->stats.pkts + + vq->stats.missed_pkts + - (vq->stats.xstats[VHOST_BROADCAST_PKT] + + vq->stats.xstats[VHOST_MULTICAST_PKT]); + } + for (t = 0; t < VHOST_NB_XSTATS_RXPORT; t++) { + xstats[count].value = 0; + for (i = 0; i < dev->data->nb_rx_queues; i++) { + vq = dev->data->rx_queues[i]; + if (!vq) + continue; + xstats[count].value += + *(uint64_t *)(((char *)vq) + + vhost_rxport_stat_strings[t].offset); + } + xstats[count].id = count; + count++; + } + for (t = 0; t < VHOST_NB_XSTATS_TXPORT; t++) { + xstats[count].value = 0; + for (i = 0; i < dev->data->nb_tx_queues; i++) { + vq = dev->data->tx_queues[i]; + if (!vq) + continue; + xstats[count].value += + *(uint64_t *)(((char *)vq) + + vhost_txport_stat_strings[t].offset); + } + xstats[count].id = count; + count++; + } + return count; +} + +static inline void +vhost_count_multicast_broadcast(struct vhost_queue *vq, + struct rte_mbuf *mbuf) +{ + struct ether_addr *ea = NULL; + struct vhost_stats *pstats = &vq->stats; + + ea = rte_pktmbuf_mtod(mbuf, struct ether_addr *); + if (is_multicast_ether_addr(ea)) { + if (is_broadcast_ether_addr(ea)) + pstats->xstats[VHOST_BROADCAST_PKT]++; + else + pstats->xstats[VHOST_MULTICAST_PKT]++; + } +} + +static void +vhost_update_packet_xstats(struct vhost_queue *vq, + struct rte_mbuf **bufs, + uint16_t count) +{ + uint32_t pkt_len = 0; + uint64_t i = 0; + uint64_t index; + struct vhost_stats *pstats = &vq->stats; + + for (i = 0; i < count ; i++) { + pkt_len = bufs[i]->pkt_len; + if (pkt_len == 64) { + pstats->xstats[VHOST_64_PKT]++; + } else if (pkt_len > 64 && pkt_len < 1024) { + index = (sizeof(pkt_len) * 8) + - __builtin_clz(pkt_len) - 5; + pstats->xstats[index]++; + } else { + if (pkt_len < 64) + pstats->xstats[VHOST_UNDERSIZE_PKT]++; + else if (pkt_len <= 1522) + pstats->xstats[VHOST_1024_TO_1522_PKT]++; + else if (pkt_len > 1522) + pstats->xstats[VHOST_1523_TO_MAX_PKT]++; + } + vhost_count_multicast_broadcast(vq, bufs[i]); + } +} + static uint16_t eth_vhost_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs) { @@ -138,16 +401,18 @@ eth_vhost_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs) goto out; /* Dequeue packets from guest TX queue */ - nb_rx = rte_vhost_dequeue_burst(r->device, + nb_rx = rte_vhost_dequeue_burst(r->vid, r->virtqueue_id, r->mb_pool, bufs, nb_bufs); - r->rx_pkts += nb_rx; + r->stats.pkts += nb_rx; for (i = 0; likely(i < nb_rx); i++) { bufs[i]->port = r->port; - r->rx_bytes += bufs[i]->pkt_len; + r->stats.bytes += bufs[i]->pkt_len; } + vhost_update_packet_xstats(r, bufs, nb_rx); + out: rte_atomic32_set(&r->while_queuing, 0); @@ -169,14 +434,23 @@ eth_vhost_tx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs) goto out; /* Enqueue packets to guest RX queue */ - nb_tx = rte_vhost_enqueue_burst(r->device, + nb_tx = rte_vhost_enqueue_burst(r->vid, r->virtqueue_id, bufs, nb_bufs); - r->tx_pkts += nb_tx; - r->missed_pkts += nb_bufs - nb_tx; + r->stats.pkts += nb_tx; + r->stats.missed_pkts += nb_bufs - nb_tx; for (i = 0; likely(i < nb_tx); i++) - r->tx_bytes += bufs[i]->pkt_len; + r->stats.bytes += bufs[i]->pkt_len; + + vhost_update_packet_xstats(r, bufs, nb_tx); + + /* According to RFC2863 page42 section ifHCOutMulticastPkts and + * ifHCOutBroadcastPkts, the counters "multicast" and "broadcast" + * are increased when packets are not transmitted successfully. + */ + for (i = nb_tx; i < nb_bufs; i++) + vhost_count_multicast_broadcast(r, bufs[i]); for (i = 0; likely(i < nb_tx); i++) rte_pktmbuf_free(bufs[i]); @@ -220,166 +494,169 @@ find_internal_resource(char *ifname) return list; } +static void +update_queuing_status(struct rte_eth_dev *dev) +{ + struct pmd_internal *internal = dev->data->dev_private; + struct vhost_queue *vq; + unsigned int i; + int allow_queuing = 1; + + if (rte_atomic32_read(&internal->started) == 0 || + rte_atomic32_read(&internal->dev_attached) == 0) + allow_queuing = 0; + + /* Wait until rx/tx_pkt_burst stops accessing vhost device */ + for (i = 0; i < dev->data->nb_rx_queues; i++) { + vq = dev->data->rx_queues[i]; + if (vq == NULL) + continue; + rte_atomic32_set(&vq->allow_queuing, allow_queuing); + while (rte_atomic32_read(&vq->while_queuing)) + rte_pause(); + } + + for (i = 0; i < dev->data->nb_tx_queues; i++) { + vq = dev->data->tx_queues[i]; + if (vq == NULL) + continue; + rte_atomic32_set(&vq->allow_queuing, allow_queuing); + while (rte_atomic32_read(&vq->while_queuing)) + rte_pause(); + } +} + static int -new_device(struct virtio_net *dev) +new_device(int vid) { struct rte_eth_dev *eth_dev; struct internal_list *list; struct pmd_internal *internal; struct vhost_queue *vq; unsigned i; + char ifname[PATH_MAX]; +#ifdef RTE_LIBRTE_VHOST_NUMA + int newnode; +#endif - if (dev == NULL) { - RTE_LOG(INFO, PMD, "Invalid argument\n"); - return -1; - } - - list = find_internal_resource(dev->ifname); + rte_vhost_get_ifname(vid, ifname, sizeof(ifname)); + list = find_internal_resource(ifname); if (list == NULL) { - RTE_LOG(INFO, PMD, "Invalid device name\n"); + RTE_LOG(INFO, PMD, "Invalid device name: %s\n", ifname); return -1; } eth_dev = list->eth_dev; internal = eth_dev->data->dev_private; +#ifdef RTE_LIBRTE_VHOST_NUMA + newnode = rte_vhost_get_numa_node(vid); + if (newnode >= 0) + eth_dev->data->numa_node = newnode; +#endif + for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { vq = eth_dev->data->rx_queues[i]; if (vq == NULL) continue; - vq->device = dev; + vq->vid = vid; vq->internal = internal; vq->port = eth_dev->data->port_id; - rte_vhost_enable_guest_notification(dev, vq->virtqueue_id, 0); } for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { vq = eth_dev->data->tx_queues[i]; if (vq == NULL) continue; - vq->device = dev; + vq->vid = vid; vq->internal = internal; vq->port = eth_dev->data->port_id; - rte_vhost_enable_guest_notification(dev, vq->virtqueue_id, 0); } - dev->flags |= VIRTIO_DEV_RUNNING; - dev->priv = eth_dev; - eth_dev->data->dev_link.link_status = 1; + for (i = 0; i < rte_vhost_get_queue_num(vid) * VIRTIO_QNUM; i++) + rte_vhost_enable_guest_notification(vid, i, 0); - for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { - vq = eth_dev->data->rx_queues[i]; - if (vq == NULL) - continue; - rte_atomic32_set(&vq->allow_queuing, 1); - } - for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { - vq = eth_dev->data->tx_queues[i]; - if (vq == NULL) - continue; - rte_atomic32_set(&vq->allow_queuing, 1); - } + eth_dev->data->dev_link.link_status = ETH_LINK_UP; + + rte_atomic32_set(&internal->dev_attached, 1); + update_queuing_status(eth_dev); RTE_LOG(INFO, PMD, "New connection established\n"); - _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_INTR_LSC); + _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_INTR_LSC, NULL); return 0; } static void -destroy_device(volatile struct virtio_net *dev) +destroy_device(int vid) { struct rte_eth_dev *eth_dev; + struct pmd_internal *internal; struct vhost_queue *vq; + struct internal_list *list; + char ifname[PATH_MAX]; unsigned i; + struct rte_vhost_vring_state *state; - if (dev == NULL) { - RTE_LOG(INFO, PMD, "Invalid argument\n"); + rte_vhost_get_ifname(vid, ifname, sizeof(ifname)); + list = find_internal_resource(ifname); + if (list == NULL) { + RTE_LOG(ERR, PMD, "Invalid interface name: %s\n", ifname); return; } + eth_dev = list->eth_dev; + internal = eth_dev->data->dev_private; - eth_dev = (struct rte_eth_dev *)dev->priv; - if (eth_dev == NULL) { - RTE_LOG(INFO, PMD, "Failed to find a ethdev\n"); - return; - } + rte_atomic32_set(&internal->dev_attached, 0); + update_queuing_status(eth_dev); + + eth_dev->data->dev_link.link_status = ETH_LINK_DOWN; - /* Wait until rx/tx_pkt_burst stops accessing vhost device */ for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { vq = eth_dev->data->rx_queues[i]; if (vq == NULL) continue; - rte_atomic32_set(&vq->allow_queuing, 0); - while (rte_atomic32_read(&vq->while_queuing)) - rte_pause(); + vq->vid = -1; } for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { vq = eth_dev->data->tx_queues[i]; if (vq == NULL) continue; - rte_atomic32_set(&vq->allow_queuing, 0); - while (rte_atomic32_read(&vq->while_queuing)) - rte_pause(); + vq->vid = -1; } - eth_dev->data->dev_link.link_status = 0; - - dev->priv = NULL; - dev->flags &= ~VIRTIO_DEV_RUNNING; - - for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { - vq = eth_dev->data->rx_queues[i]; - if (vq == NULL) - continue; - vq->device = NULL; - } - for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { - vq = eth_dev->data->tx_queues[i]; - if (vq == NULL) - continue; - vq->device = NULL; + state = vring_states[eth_dev->data->port_id]; + rte_spinlock_lock(&state->lock); + for (i = 0; i <= state->max_vring; i++) { + state->cur[i] = false; + state->seen[i] = false; } + state->max_vring = 0; + rte_spinlock_unlock(&state->lock); RTE_LOG(INFO, PMD, "Connection closed\n"); - _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_INTR_LSC); + _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_INTR_LSC, NULL); } static int -vring_state_changed(struct virtio_net *dev, uint16_t vring, int enable) +vring_state_changed(int vid, uint16_t vring, int enable) { struct rte_vhost_vring_state *state; struct rte_eth_dev *eth_dev; struct internal_list *list; -#ifdef RTE_LIBRTE_VHOST_NUMA - int newnode, ret; -#endif - - if (dev == NULL) { - RTE_LOG(ERR, PMD, "Invalid argument\n"); - return -1; - } + char ifname[PATH_MAX]; - list = find_internal_resource(dev->ifname); + rte_vhost_get_ifname(vid, ifname, sizeof(ifname)); + list = find_internal_resource(ifname); if (list == NULL) { - RTE_LOG(ERR, PMD, "Invalid interface name: %s\n", dev->ifname); + RTE_LOG(ERR, PMD, "Invalid interface name: %s\n", ifname); return -1; } eth_dev = list->eth_dev; /* won't be NULL */ state = vring_states[eth_dev->data->port_id]; - -#ifdef RTE_LIBRTE_VHOST_NUMA - ret = get_mempolicy(&newnode, NULL, 0, dev, - MPOL_F_NODE | MPOL_F_ADDR); - if (ret < 0) { - RTE_LOG(ERR, PMD, "Unknown numa node\n"); - return -1; - } - - eth_dev->data->numa_node = newnode; -#endif rte_spinlock_lock(&state->lock); state->cur[vring] = enable; state->max_vring = RTE_MAX(vring, state->max_vring); @@ -388,7 +665,7 @@ vring_state_changed(struct virtio_net *dev, uint16_t vring, int enable) RTE_LOG(INFO, PMD, "vring%u is %s\n", vring, enable ? "enabled" : "disabled"); - _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_QUEUE_STATE); + _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_QUEUE_STATE, NULL); return 0; } @@ -430,6 +707,35 @@ rte_eth_vhost_get_queue_event(uint8_t port_id, return -1; } +int +rte_eth_vhost_get_vid_from_port_id(uint8_t port_id) +{ + struct internal_list *list; + struct rte_eth_dev *eth_dev; + struct vhost_queue *vq; + int vid = -1; + + if (!rte_eth_dev_is_valid_port(port_id)) + return -1; + + pthread_mutex_lock(&internal_list_lock); + + TAILQ_FOREACH(list, &internal_list, next) { + eth_dev = list->eth_dev; + if (eth_dev->data->port_id == port_id) { + vq = eth_dev->data->rx_queues[0]; + if (vq) { + vid = vq->vid; + } + break; + } + } + + pthread_mutex_unlock(&internal_list_lock); + + return vid; +} + static void * vhost_driver_session(void *param __rte_unused) { @@ -479,19 +785,11 @@ static int eth_dev_start(struct rte_eth_dev *dev) { struct pmd_internal *internal = dev->data->dev_private; - int ret = 0; - if (rte_atomic16_cmpset(&internal->once, 0, 1)) { - ret = rte_vhost_driver_register(internal->iface_name); - if (ret) - return ret; - } + rte_atomic32_set(&internal->started, 1); + update_queuing_status(dev); - /* We need only one message handling thread */ - if (rte_atomic16_add_return(&nb_started_ports, 1) == 1) - ret = vhost_driver_session_start(); - - return ret; + return 0; } static void @@ -499,11 +797,8 @@ eth_dev_stop(struct rte_eth_dev *dev) { struct pmd_internal *internal = dev->data->dev_private; - if (rte_atomic16_cmpset(&internal->once, 1, 0)) - rte_vhost_driver_unregister(internal->iface_name); - - if (rte_atomic16_sub_return(&nb_started_ports, 1) == 0) - vhost_driver_session_stop(); + rte_atomic32_set(&internal->started, 0); + update_queuing_status(dev); } static int @@ -554,11 +849,18 @@ static void eth_dev_info(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) { - dev_info->driver_name = drivername; + struct pmd_internal *internal; + + internal = dev->data->dev_private; + if (internal == NULL) { + RTE_LOG(ERR, PMD, "Invalid device specified\n"); + return; + } + dev_info->max_mac_addrs = 1; dev_info->max_rx_pktlen = (uint32_t)-1; - dev_info->max_rx_queues = dev->data->nb_rx_queues; - dev_info->max_tx_queues = dev->data->nb_tx_queues; + dev_info->max_rx_queues = internal->max_queues; + dev_info->max_tx_queues = internal->max_queues; dev_info->min_rx_bufsize = 0; } @@ -575,10 +877,10 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) if (dev->data->rx_queues[i] == NULL) continue; vq = dev->data->rx_queues[i]; - stats->q_ipackets[i] = vq->rx_pkts; + stats->q_ipackets[i] = vq->stats.pkts; rx_total += stats->q_ipackets[i]; - stats->q_ibytes[i] = vq->rx_bytes; + stats->q_ibytes[i] = vq->stats.bytes; rx_total_bytes += stats->q_ibytes[i]; } @@ -587,17 +889,17 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) if (dev->data->tx_queues[i] == NULL) continue; vq = dev->data->tx_queues[i]; - stats->q_opackets[i] = vq->tx_pkts; - tx_missed_total += vq->missed_pkts; + stats->q_opackets[i] = vq->stats.pkts; + tx_missed_total += vq->stats.missed_pkts; tx_total += stats->q_opackets[i]; - stats->q_obytes[i] = vq->tx_bytes; + stats->q_obytes[i] = vq->stats.bytes; tx_total_bytes += stats->q_obytes[i]; } stats->ipackets = rx_total; stats->opackets = tx_total; - stats->imissed = tx_missed_total; + stats->oerrors = tx_missed_total; stats->ibytes = rx_total_bytes; stats->obytes = tx_total_bytes; } @@ -612,16 +914,16 @@ eth_stats_reset(struct rte_eth_dev *dev) if (dev->data->rx_queues[i] == NULL) continue; vq = dev->data->rx_queues[i]; - vq->rx_pkts = 0; - vq->rx_bytes = 0; + vq->stats.pkts = 0; + vq->stats.bytes = 0; } for (i = 0; i < dev->data->nb_tx_queues; i++) { if (dev->data->tx_queues[i] == NULL) continue; vq = dev->data->tx_queues[i]; - vq->tx_pkts = 0; - vq->tx_bytes = 0; - vq->missed_pkts = 0; + vq->stats.pkts = 0; + vq->stats.bytes = 0; + vq->stats.missed_pkts = 0; } } @@ -675,11 +977,16 @@ static const struct eth_dev_ops ops = { .link_update = eth_link_update, .stats_get = eth_stats_get, .stats_reset = eth_stats_reset, + .xstats_reset = vhost_dev_xstats_reset, + .xstats_get = vhost_dev_xstats_get, + .xstats_get_names = vhost_dev_xstats_get_names, }; +static struct rte_vdev_driver pmd_vhost_drv; + static int eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, - const unsigned numa_node) + const unsigned numa_node, uint64_t flags) { struct rte_eth_dev_data *data = NULL; struct pmd_internal *internal = NULL; @@ -707,7 +1014,7 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, goto error; /* reserve an ethdev entry */ - eth_dev = rte_eth_dev_allocate(name, RTE_ETH_DEV_VIRTUAL); + eth_dev = rte_eth_dev_allocate(name); if (eth_dev == NULL) goto error; @@ -722,8 +1029,6 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, if (vring_state == NULL) goto error; - TAILQ_INIT(ð_dev->link_intr_cbs); - /* now put it all together * - store queue data in internal, * - store numa_node info in ethdev data @@ -750,6 +1055,7 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, memmove(data->name, eth_dev->data->name, sizeof(data->name)); data->nb_rx_queues = queues; data->nb_tx_queues = queues; + internal->max_queues = queues; data->dev_link = pmd_link; data->mac_addrs = eth_addr; @@ -762,13 +1068,22 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, data->dev_flags = RTE_ETH_DEV_DETACHABLE | RTE_ETH_DEV_INTR_LSC; data->kdrv = RTE_KDRV_NONE; - data->drv_name = internal->dev_name; + data->drv_name = pmd_vhost_drv.driver.name; data->numa_node = numa_node; /* finally assign rx and tx ops */ eth_dev->rx_pkt_burst = eth_vhost_rx; eth_dev->tx_pkt_burst = eth_vhost_tx; + if (rte_vhost_driver_register(iface_name, flags)) + goto error; + + /* We need only one message handling thread */ + if (rte_atomic16_add_return(&nb_started_ports, 1) == 1) { + if (vhost_driver_session_start()) + goto error; + } + return data->port_id; error: @@ -799,30 +1114,30 @@ open_iface(const char *key __rte_unused, const char *value, void *extra_args) } static inline int -open_queues(const char *key __rte_unused, const char *value, void *extra_args) +open_int(const char *key __rte_unused, const char *value, void *extra_args) { - uint16_t *q = extra_args; + uint16_t *n = extra_args; if (value == NULL || extra_args == NULL) return -EINVAL; - *q = (uint16_t)strtoul(value, NULL, 0); - if (*q == USHRT_MAX && errno == ERANGE) - return -1; - - if (*q > RTE_MAX_QUEUES_PER_PORT) + *n = (uint16_t)strtoul(value, NULL, 0); + if (*n == USHRT_MAX && errno == ERANGE) return -1; return 0; } static int -rte_pmd_vhost_devinit(const char *name, const char *params) +rte_pmd_vhost_probe(const char *name, const char *params) { struct rte_kvargs *kvlist = NULL; int ret = 0; char *iface_name; uint16_t queues; + uint64_t flags = 0; + int client_mode = 0; + int dequeue_zero_copy = 0; RTE_LOG(INFO, PMD, "Initializing pmd_vhost for %s\n", name); @@ -842,14 +1157,34 @@ rte_pmd_vhost_devinit(const char *name, const char *params) if (rte_kvargs_count(kvlist, ETH_VHOST_QUEUES_ARG) == 1) { ret = rte_kvargs_process(kvlist, ETH_VHOST_QUEUES_ARG, - &open_queues, &queues); - if (ret < 0) + &open_int, &queues); + if (ret < 0 || queues > RTE_MAX_QUEUES_PER_PORT) goto out_free; } else queues = 1; - eth_dev_vhost_create(name, iface_name, queues, rte_socket_id()); + if (rte_kvargs_count(kvlist, ETH_VHOST_CLIENT_ARG) == 1) { + ret = rte_kvargs_process(kvlist, ETH_VHOST_CLIENT_ARG, + &open_int, &client_mode); + if (ret < 0) + goto out_free; + + if (client_mode) + flags |= RTE_VHOST_USER_CLIENT; + } + + if (rte_kvargs_count(kvlist, ETH_VHOST_DEQUEUE_ZERO_COPY) == 1) { + ret = rte_kvargs_process(kvlist, ETH_VHOST_DEQUEUE_ZERO_COPY, + &open_int, &dequeue_zero_copy); + if (ret < 0) + goto out_free; + + if (dequeue_zero_copy) + flags |= RTE_VHOST_USER_DEQUEUE_ZERO_COPY; + } + + eth_dev_vhost_create(name, iface_name, queues, rte_socket_id(), flags); out_free: rte_kvargs_free(kvlist); @@ -857,7 +1192,7 @@ out_free: } static int -rte_pmd_vhost_devuninit(const char *name) +rte_pmd_vhost_remove(const char *name) { struct rte_eth_dev *eth_dev = NULL; struct pmd_internal *internal; @@ -886,6 +1221,11 @@ rte_pmd_vhost_devuninit(const char *name) eth_dev_stop(eth_dev); + rte_vhost_driver_unregister(internal->iface_name); + + if (rte_atomic16_sub_return(&nb_started_ports, 1) == 0) + vhost_driver_session_stop(); + rte_free(vring_states[eth_dev->data->port_id]); vring_states[eth_dev->data->port_id] = NULL; @@ -906,11 +1246,13 @@ rte_pmd_vhost_devuninit(const char *name) return 0; } -static struct rte_driver pmd_vhost_drv = { - .name = "eth_vhost", - .type = PMD_VDEV, - .init = rte_pmd_vhost_devinit, - .uninit = rte_pmd_vhost_devuninit, +static struct rte_vdev_driver pmd_vhost_drv = { + .probe = rte_pmd_vhost_probe, + .remove = rte_pmd_vhost_remove, }; -PMD_REGISTER_DRIVER(pmd_vhost_drv); +RTE_PMD_REGISTER_VDEV(net_vhost, pmd_vhost_drv); +RTE_PMD_REGISTER_ALIAS(net_vhost, eth_vhost); +RTE_PMD_REGISTER_PARAM_STRING(net_vhost, + "iface= " + "queues=");