X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;ds=sidebyside;f=drivers%2Fnet%2Fvhost%2Frte_eth_vhost.c;h=11b607650b06e50ecf4d08c51d4cdf4eb28ec232;hb=30a701a53737;hp=2a38b194f1da5e77d7bb851d4b25af6cd36286be;hpb=7cd1dbf4fff0e612673594e7055fe25724a148bb;p=dpdk.git diff --git a/drivers/net/vhost/rte_eth_vhost.c b/drivers/net/vhost/rte_eth_vhost.c index 2a38b194f1..11b607650b 100644 --- a/drivers/net/vhost/rte_eth_vhost.c +++ b/drivers/net/vhost/rte_eth_vhost.c @@ -33,25 +33,26 @@ #include #include #include -#ifdef RTE_LIBRTE_VHOST_NUMA -#include -#endif #include -#include +#include +#include #include #include -#include +#include #include -#include +#include #include #include "rte_eth_vhost.h" +enum {VIRTIO_RXQ, VIRTIO_TXQ, VIRTIO_QNUM}; + #define ETH_VHOST_IFACE_ARG "iface" #define ETH_VHOST_QUEUES_ARG "queues" #define ETH_VHOST_CLIENT_ARG "client" #define ETH_VHOST_DEQUEUE_ZERO_COPY "dequeue-zero-copy" +#define ETH_VHOST_IOMMU_SUPPORT "iommu-support" #define VHOST_MAX_PKT_BURST 32 static const char *valid_arguments[] = { @@ -59,6 +60,7 @@ static const char *valid_arguments[] = { ETH_VHOST_QUEUES_ARG, ETH_VHOST_CLIENT_ARG, ETH_VHOST_DEQUEUE_ZERO_COPY, + ETH_VHOST_IOMMU_SUPPORT, NULL }; @@ -105,7 +107,7 @@ struct vhost_queue { rte_atomic32_t while_queuing; struct pmd_internal *internal; struct rte_mempool *mb_pool; - uint8_t port; + uint16_t port; uint16_t virtqueue_id; struct vhost_stats stats; }; @@ -115,6 +117,7 @@ struct pmd_internal { char *dev_name; char *iface_name; uint16_t max_queues; + uint16_t vid; rte_atomic32_t started; }; @@ -129,9 +132,6 @@ static struct internal_list_head internal_list = static pthread_mutex_t internal_list_lock = PTHREAD_MUTEX_INITIALIZER; -static rte_atomic16_t nb_started_ports; -static pthread_t session_th; - static struct rte_eth_link pmd_link = { .link_speed = 10000, .link_duplex = ETH_LINK_FULL_DUPLEX, @@ -392,6 +392,7 @@ eth_vhost_rx(void *q, struct rte_mbuf **bufs, uint16_t nb_bufs) { struct vhost_queue *r = q; uint16_t i, nb_rx = 0; + uint16_t nb_receive = nb_bufs; if (unlikely(rte_atomic32_read(&r->allow_queuing) == 0)) return 0; @@ -402,8 +403,20 @@ 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->vid, - r->virtqueue_id, r->mb_pool, bufs, nb_bufs); + while (nb_receive) { + uint16_t nb_pkts; + uint16_t num = (uint16_t)RTE_MIN(nb_receive, + VHOST_MAX_PKT_BURST); + + nb_pkts = rte_vhost_dequeue_burst(r->vid, r->virtqueue_id, + r->mb_pool, &bufs[nb_rx], + num); + + nb_rx += nb_pkts; + nb_receive -= nb_pkts; + if (nb_pkts < num) + break; + } r->stats.pkts += nb_rx; @@ -515,8 +528,10 @@ update_queuing_status(struct rte_eth_dev *dev) unsigned int i; int allow_queuing = 1; - if (rte_atomic32_read(&internal->started) == 0 || - rte_atomic32_read(&internal->dev_attached) == 0) + if (rte_atomic32_read(&internal->dev_attached) == 0) + return; + + if (rte_atomic32_read(&internal->started) == 0) allow_queuing = 0; /* Wait until rx/tx_pkt_burst stops accessing vhost device */ @@ -539,13 +554,36 @@ update_queuing_status(struct rte_eth_dev *dev) } } +static void +queue_setup(struct rte_eth_dev *eth_dev, struct pmd_internal *internal) +{ + struct vhost_queue *vq; + int i; + + for (i = 0; i < eth_dev->data->nb_rx_queues; i++) { + vq = eth_dev->data->rx_queues[i]; + if (!vq) + continue; + vq->vid = internal->vid; + vq->internal = internal; + vq->port = eth_dev->data->port_id; + } + for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { + vq = eth_dev->data->tx_queues[i]; + if (!vq) + continue; + vq->vid = internal->vid; + vq->internal = internal; + vq->port = eth_dev->data->port_id; + } +} + static int 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 @@ -568,32 +606,25 @@ new_device(int vid) 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->vid = vid; - vq->internal = internal; - vq->port = eth_dev->data->port_id; - } - for (i = 0; i < eth_dev->data->nb_tx_queues; i++) { - vq = eth_dev->data->tx_queues[i]; - if (vq == NULL) - continue; - vq->vid = vid; - vq->internal = internal; - vq->port = eth_dev->data->port_id; + internal->vid = vid; + if (eth_dev->data->rx_queues && eth_dev->data->tx_queues) { + queue_setup(eth_dev, internal); + rte_atomic32_set(&internal->dev_attached, 1); + } else { + RTE_LOG(INFO, PMD, "RX/TX queues have not setup yet\n"); + rte_atomic32_set(&internal->dev_attached, 0); } - for (i = 0; i < rte_vhost_get_queue_num(vid) * VIRTIO_QNUM; i++) + for (i = 0; i < rte_vhost_get_vring_num(vid); i++) rte_vhost_enable_guest_notification(vid, i, 0); + rte_vhost_get_mtu(vid, ð_dev->data->mtu); + 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_LOG(INFO, PMD, "Vhost device %d created\n", vid); _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_INTR_LSC, NULL); @@ -620,8 +651,9 @@ destroy_device(int vid) eth_dev = list->eth_dev; internal = eth_dev->data->dev_private; - rte_atomic32_set(&internal->dev_attached, 0); + rte_atomic32_set(&internal->started, 0); update_queuing_status(eth_dev); + rte_atomic32_set(&internal->dev_attached, 0); eth_dev->data->dev_link.link_status = ETH_LINK_DOWN; @@ -647,7 +679,7 @@ destroy_device(int vid) state->max_vring = 0; rte_spinlock_unlock(&state->lock); - RTE_LOG(INFO, PMD, "Connection closed\n"); + RTE_LOG(INFO, PMD, "Vhost device %d destroyed\n", vid); _rte_eth_dev_callback_process(eth_dev, RTE_ETH_EVENT_INTR_LSC, NULL); } @@ -683,8 +715,14 @@ vring_state_changed(int vid, uint16_t vring, int enable) return 0; } +static struct vhost_device_ops vhost_ops = { + .new_device = new_device, + .destroy_device = destroy_device, + .vring_state_changed = vring_state_changed, +}; + int -rte_eth_vhost_get_queue_event(uint8_t port_id, +rte_eth_vhost_get_queue_event(uint16_t port_id, struct rte_eth_vhost_queue_event *event) { struct rte_vhost_vring_state *state; @@ -721,7 +759,7 @@ rte_eth_vhost_get_queue_event(uint8_t port_id, } int -rte_eth_vhost_get_vid_from_port_id(uint8_t port_id) +rte_eth_vhost_get_vid_from_port_id(uint16_t port_id) { struct internal_list *list; struct rte_eth_dev *eth_dev; @@ -749,58 +787,18 @@ rte_eth_vhost_get_vid_from_port_id(uint8_t port_id) return vid; } -static void * -vhost_driver_session(void *param __rte_unused) -{ - static struct virtio_net_device_ops vhost_ops; - - /* set vhost arguments */ - vhost_ops.new_device = new_device; - vhost_ops.destroy_device = destroy_device; - vhost_ops.vring_state_changed = vring_state_changed; - if (rte_vhost_driver_callback_register(&vhost_ops) < 0) - RTE_LOG(ERR, PMD, "Can't register callbacks\n"); - - /* start event handling */ - rte_vhost_driver_session_start(); - - return NULL; -} - static int -vhost_driver_session_start(void) +eth_dev_start(struct rte_eth_dev *eth_dev) { - int ret; - - ret = pthread_create(&session_th, - NULL, vhost_driver_session, NULL); - if (ret) - RTE_LOG(ERR, PMD, "Can't create a thread\n"); + struct pmd_internal *internal = eth_dev->data->dev_private; - return ret; -} - -static void -vhost_driver_session_stop(void) -{ - int ret; - - ret = pthread_cancel(session_th); - if (ret) - RTE_LOG(ERR, PMD, "Can't cancel the thread\n"); - - ret = pthread_join(session_th, NULL); - if (ret) - RTE_LOG(ERR, PMD, "Can't join the thread\n"); -} - -static int -eth_dev_start(struct rte_eth_dev *dev) -{ - struct pmd_internal *internal = dev->data->dev_private; + if (unlikely(rte_atomic32_read(&internal->dev_attached) == 0)) { + queue_setup(eth_dev, internal); + rte_atomic32_set(&internal->dev_attached, 1); + } rte_atomic32_set(&internal->started, 1); - update_queuing_status(dev); + update_queuing_status(eth_dev); return 0; } @@ -819,11 +817,14 @@ eth_dev_close(struct rte_eth_dev *dev) { struct pmd_internal *internal; struct internal_list *list; + unsigned int i; internal = dev->data->dev_private; if (!internal) return; + eth_dev_stop(dev); + rte_vhost_driver_unregister(internal->iface_name); list = find_internal_resource(internal->iface_name); @@ -835,9 +836,17 @@ eth_dev_close(struct rte_eth_dev *dev) pthread_mutex_unlock(&internal_list_lock); rte_free(list); + for (i = 0; i < dev->data->nb_rx_queues; i++) + rte_free(dev->data->rx_queues[i]); + for (i = 0; i < dev->data->nb_tx_queues; i++) + rte_free(dev->data->tx_queues[i]); + + rte_free(dev->data->mac_addrs); free(internal->dev_name); free(internal->iface_name); rte_free(internal); + + dev->data->dev_private = NULL; } static int @@ -903,7 +912,7 @@ eth_dev_info(struct rte_eth_dev *dev, dev_info->min_rx_bufsize = 0; } -static void +static int eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) { unsigned i; @@ -941,6 +950,8 @@ eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) stats->oerrors = tx_missed_total; stats->ibytes = rx_total_bytes; stats->obytes = tx_total_bytes; + + return 0; } static void @@ -989,29 +1000,16 @@ eth_link_update(struct rte_eth_dev *dev __rte_unused, return 0; } -/** - * Disable features in feature_mask. Returns 0 on success. - */ -int -rte_eth_vhost_feature_disable(uint64_t feature_mask) +static uint32_t +eth_rx_queue_count(struct rte_eth_dev *dev, uint16_t rx_queue_id) { - return rte_vhost_feature_disable(feature_mask); -} + struct vhost_queue *vq; -/** - * Enable features in feature_mask. Returns 0 on success. - */ -int -rte_eth_vhost_feature_enable(uint64_t feature_mask) -{ - return rte_vhost_feature_enable(feature_mask); -} + vq = dev->data->rx_queues[rx_queue_id]; + if (vq == NULL) + return 0; -/* Returns currently supported vhost features */ -uint64_t -rte_eth_vhost_feature_get(void) -{ - return rte_vhost_feature_get(); + return rte_vhost_rx_queue_count(vq->vid, vq->virtqueue_id); } static const struct eth_dev_ops ops = { @@ -1025,6 +1023,7 @@ static const struct eth_dev_ops ops = { .rx_queue_release = eth_queue_release, .tx_queue_release = eth_queue_release, .tx_done_cleanup = eth_tx_done_cleanup, + .rx_queue_count = eth_rx_queue_count, .link_update = eth_link_update, .stats_get = eth_stats_get, .stats_reset = eth_stats_reset, @@ -1036,9 +1035,10 @@ static const struct eth_dev_ops ops = { 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, uint64_t flags) +eth_dev_vhost_create(struct rte_vdev_device *dev, char *iface_name, + int16_t queues, const unsigned int numa_node, uint64_t flags) { + const char *name = rte_vdev_device_name(dev); struct rte_eth_dev_data *data = NULL; struct pmd_internal *internal = NULL; struct rte_eth_dev *eth_dev = NULL; @@ -1049,23 +1049,19 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, RTE_LOG(INFO, PMD, "Creating VHOST-USER backend on numa socket %u\n", numa_node); - /* now do all data allocation - for eth_dev structure, dummy pci driver - * and internal (private) data + /* now do all data allocation - for eth_dev structure and internal + * (private) data */ data = rte_zmalloc_socket(name, sizeof(*data), 0, numa_node); if (data == NULL) goto error; - internal = rte_zmalloc_socket(name, sizeof(*internal), 0, numa_node); - if (internal == NULL) - goto error; - list = rte_zmalloc_socket(name, sizeof(*list), 0, numa_node); if (list == NULL) goto error; /* reserve an ethdev entry */ - eth_dev = rte_eth_dev_allocate(name); + eth_dev = rte_eth_vdev_allocate(dev, sizeof(*internal)); if (eth_dev == NULL) goto error; @@ -1082,10 +1078,10 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, /* now put it all together * - store queue data in internal, - * - store numa_node info in ethdev data * - point eth_dev_data to internals * - and point eth_dev structure to new eth_dev_data structure */ + internal = eth_dev->data->dev_private; internal->dev_name = strdup(name); if (internal->dev_name == NULL) goto error; @@ -1101,26 +1097,20 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, rte_spinlock_init(&vring_state->lock); vring_states[eth_dev->data->port_id] = vring_state; - data->dev_private = internal; - data->port_id = eth_dev->data->port_id; - memmove(data->name, eth_dev->data->name, sizeof(data->name)); + /* We'll replace the 'data' originally allocated by eth_dev. So the + * vhost PMD resources won't be shared between multi processes. + */ + rte_memcpy(data, eth_dev->data, sizeof(*data)); + eth_dev->data = data; + data->nb_rx_queues = queues; data->nb_tx_queues = queues; internal->max_queues = queues; data->dev_link = pmd_link; data->mac_addrs = eth_addr; + data->dev_flags = RTE_ETH_DEV_INTR_LSC; - /* We'll replace the 'data' originally allocated by eth_dev. So the - * vhost PMD resources won't be shared between multi processes. - */ - eth_dev->data = data; eth_dev->dev_ops = &ops; - eth_dev->driver = NULL; - data->dev_flags = - RTE_ETH_DEV_DETACHABLE | RTE_ETH_DEV_INTR_LSC; - data->kdrv = RTE_KDRV_NONE; - 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; @@ -1129,17 +1119,24 @@ eth_dev_vhost_create(const char *name, char *iface_name, int16_t queues, 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; + if (rte_vhost_driver_callback_register(iface_name, &vhost_ops) < 0) { + RTE_LOG(ERR, PMD, "Can't register callbacks\n"); + goto error; + } + + if (rte_vhost_driver_start(iface_name) < 0) { + RTE_LOG(ERR, PMD, "Failed to start driver for %s\n", + iface_name); + goto error; } return data->port_id; error: - if (internal) + if (internal) { + free(internal->iface_name); free(internal->dev_name); + } rte_free(vring_state); rte_free(eth_addr); if (eth_dev) @@ -1180,7 +1177,7 @@ open_int(const char *key __rte_unused, const char *value, void *extra_args) } static int -rte_pmd_vhost_probe(const char *name, const char *params) +rte_pmd_vhost_probe(struct rte_vdev_device *dev) { struct rte_kvargs *kvlist = NULL; int ret = 0; @@ -1189,10 +1186,12 @@ rte_pmd_vhost_probe(const char *name, const char *params) uint64_t flags = 0; int client_mode = 0; int dequeue_zero_copy = 0; + int iommu_support = 0; - RTE_LOG(INFO, PMD, "Initializing pmd_vhost for %s\n", name); + RTE_LOG(INFO, PMD, "Initializing pmd_vhost for %s\n", + rte_vdev_device_name(dev)); - kvlist = rte_kvargs_parse(params, valid_arguments); + kvlist = rte_kvargs_parse(rte_vdev_device_args(dev), valid_arguments); if (kvlist == NULL) return -1; @@ -1235,7 +1234,21 @@ rte_pmd_vhost_probe(const char *name, const char *params) flags |= RTE_VHOST_USER_DEQUEUE_ZERO_COPY; } - eth_dev_vhost_create(name, iface_name, queues, rte_socket_id(), flags); + if (rte_kvargs_count(kvlist, ETH_VHOST_IOMMU_SUPPORT) == 1) { + ret = rte_kvargs_process(kvlist, ETH_VHOST_IOMMU_SUPPORT, + &open_int, &iommu_support); + if (ret < 0) + goto out_free; + + if (iommu_support) + flags |= RTE_VHOST_USER_IOMMU_SUPPORT; + } + + if (dev->device.numa_node == SOCKET_ID_ANY) + dev->device.numa_node = rte_socket_id(); + + eth_dev_vhost_create(dev, iface_name, queues, dev->device.numa_node, + flags); out_free: rte_kvargs_free(kvlist); @@ -1243,11 +1256,12 @@ out_free: } static int -rte_pmd_vhost_remove(const char *name) +rte_pmd_vhost_remove(struct rte_vdev_device *dev) { + const char *name; struct rte_eth_dev *eth_dev = NULL; - unsigned int i; + name = rte_vdev_device_name(dev); RTE_LOG(INFO, PMD, "Un-Initializing pmd_vhost for %s\n", name); /* find an ethdev entry */ @@ -1255,22 +1269,11 @@ rte_pmd_vhost_remove(const char *name) if (eth_dev == NULL) return -ENODEV; - eth_dev_stop(eth_dev); - eth_dev_close(eth_dev); - 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; - for (i = 0; i < eth_dev->data->nb_rx_queues; i++) - rte_free(eth_dev->data->rx_queues[i]); - for (i = 0; i < eth_dev->data->nb_tx_queues; i++) - rte_free(eth_dev->data->tx_queues[i]); - - rte_free(eth_dev->data->mac_addrs); rte_free(eth_dev->data); rte_eth_dev_release_port(eth_dev);