#include <sys/socket.h>
#include <linux/if.h>
+#include <rte_bus_pci.h>
#include <rte_config.h>
#include <rte_cycles.h>
#include <rte_errno.h>
#include <rte_meter.h>
#include <rte_pci.h>
#include <rte_vhost.h>
+#include <rte_version.h>
#include "dirs.h"
#include "dp-packet.h"
#define NETDEV_DPDK_MBUF_ALIGN 1024
#define NETDEV_DPDK_MAX_PKT_LEN 9728
-/* Max and min number of packets in the mempool. OVS tries to allocate a
- * mempool with MAX_NB_MBUF: if this fails (because the system doesn't have
- * enough hugepages) we keep halving the number until the allocation succeeds
- * or we reach MIN_NB_MBUF */
-
-#define MAX_NB_MBUF (4096 * 64)
+/* Min number of packets in the mempool. OVS tries to allocate a mempool with
+ * roughly estimated number of mbufs: if this fails (because the system doesn't
+ * have enough hugepages) we keep halving the number until the allocation
+ * succeeds or we reach MIN_NB_MBUF */
#define MIN_NB_MBUF (4096 * 4)
#define MP_CACHE_SZ RTE_MEMPOOL_CACHE_MAX_SIZE
-/* MAX_NB_MBUF can be divided by 2 many times, until MIN_NB_MBUF */
-BUILD_ASSERT_DECL(MAX_NB_MBUF % ROUND_DOWN_POW2(MAX_NB_MBUF/MIN_NB_MBUF) == 0);
-
-/* The smallest possible NB_MBUF that we're going to try should be a multiple
- * of MP_CACHE_SZ. This is advised by DPDK documentation. */
-BUILD_ASSERT_DECL((MAX_NB_MBUF / ROUND_DOWN_POW2(MAX_NB_MBUF/MIN_NB_MBUF))
- % MP_CACHE_SZ == 0);
-
/*
* DPDK XSTATS Counter names definition
*/
#define DPDK_ETH_PORT_ID_INVALID RTE_MAX_ETHPORTS
-/* DPDK library uses uint8_t for port_id. */
-typedef uint8_t dpdk_port_t;
+/* DPDK library uses uint16_t for port_id. */
+typedef uint16_t dpdk_port_t;
#define VHOST_ENQ_RETRY_NUM 8
#define IF_NAME_SZ (PATH_MAX > IFNAMSIZ ? PATH_MAX : IFNAMSIZ)
* For all QoS implementations it should always be non-null.
*/
int (*qos_run)(struct qos_conf *qos_conf, struct rte_mbuf **pkts,
- int pkt_cnt);
+ int pkt_cnt, bool may_steal);
};
/* dpdk_qos_ops for each type of user space QoS implementation */
static struct ovs_mutex dpdk_mp_mutex OVS_ACQ_AFTER(dpdk_mutex)
= OVS_MUTEX_INITIALIZER;
-static struct ovs_list dpdk_mp_list OVS_GUARDED_BY(dpdk_mp_mutex)
- = OVS_LIST_INITIALIZER(&dpdk_mp_list);
-
-struct dpdk_mp {
- struct rte_mempool *mp;
- int mtu;
- int socket_id;
- int refcount;
- struct ovs_list list_node OVS_GUARDED_BY(dpdk_mp_mutex);
-};
-
/* There should be one 'struct dpdk_tx_queue' created for
* each cpu core. */
struct dpdk_tx_queue {
NETDEV_RX_CHECKSUM_OFFLOAD = 1 << 0,
};
-struct netdev_dpdk {
- struct netdev up;
- dpdk_port_t port_id;
- int max_packet_len;
- enum dpdk_dev_type type;
-
- struct dpdk_tx_queue *tx_q;
-
- struct ovs_mutex mutex OVS_ACQ_AFTER(dpdk_mutex);
-
- struct dpdk_mp *dpdk_mp;
- int mtu;
- int socket_id;
- int buf_size;
- struct netdev_stats stats;
- /* Protects stats */
- rte_spinlock_t stats_lock;
-
- struct eth_addr hwaddr;
- enum netdev_flags flags;
-
- struct rte_eth_link link;
- int link_reset_cnt;
-
- /* virtio identifier for vhost devices */
- ovsrcu_index vid;
-
- /* True if vHost device is 'up' and has been reconfigured at least once */
- bool vhost_reconfigured;
-
- /* Identifier used to distinguish vhost devices from each other. */
- char vhost_id[PATH_MAX];
-
- /* Device arguments for dpdk ports */
- char *devargs;
-
- /* If true, device was attached by rte_eth_dev_attach(). */
- bool attached;
-
- /* In dpdk_list. */
- struct ovs_list list_node OVS_GUARDED_BY(dpdk_mutex);
-
- /* QoS configuration and lock for the device */
- OVSRCU_TYPE(struct qos_conf *) qos_conf;
-
- /* The following properties cannot be changed when a device is running,
- * so we remember the request and update them next time
- * netdev_dpdk*_reconfigure() is called */
- int requested_mtu;
- int requested_n_txq;
- int requested_n_rxq;
- int requested_rxq_size;
- int requested_txq_size;
-
- /* Number of rx/tx descriptors for physical devices */
- int rxq_size;
- int txq_size;
-
- /* Socket ID detected when vHost device is brought up */
- int requested_socket_id;
-
- /* Denotes whether vHost port is client/server mode */
- uint64_t vhost_driver_flags;
-
- /* Ingress Policer */
- OVSRCU_TYPE(struct ingress_policer *) ingress_policer;
- uint32_t policer_rate;
- uint32_t policer_burst;
-
- /* DPDK-ETH Flow control */
- struct rte_eth_fc_conf fc_conf;
+/*
+ * In order to avoid confusion in variables names, following naming convention
+ * should be used, if possible:
+ *
+ * 'struct netdev' : 'netdev'
+ * 'struct netdev_dpdk' : 'dev'
+ * 'struct netdev_rxq' : 'rxq'
+ * 'struct netdev_rxq_dpdk' : 'rx'
+ *
+ * Example:
+ * struct netdev *netdev = netdev_from_name(name);
+ * struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
+ *
+ * Also, 'netdev' should be used instead of 'dev->up', where 'netdev' was
+ * already defined.
+ */
- /* DPDK-ETH hardware offload features,
- * from the enum set 'dpdk_hw_ol_features' */
- uint32_t hw_ol_features;
+struct netdev_dpdk {
+ PADDED_MEMBERS_CACHELINE_MARKER(CACHE_LINE_SIZE, cacheline0,
+ dpdk_port_t port_id;
+
+ /* If true, device was attached by rte_eth_dev_attach(). */
+ bool attached;
+ struct eth_addr hwaddr;
+ int mtu;
+ int socket_id;
+ int buf_size;
+ int max_packet_len;
+ enum dpdk_dev_type type;
+ enum netdev_flags flags;
+ char *devargs; /* Device arguments for dpdk ports */
+ struct dpdk_tx_queue *tx_q;
+ struct rte_eth_link link;
+ int link_reset_cnt;
+ /* 4 pad bytes here. */
+ );
+
+ PADDED_MEMBERS_CACHELINE_MARKER(CACHE_LINE_SIZE, cacheline1,
+ struct ovs_mutex mutex OVS_ACQ_AFTER(dpdk_mutex);
+ struct rte_mempool *mp;
+
+ /* virtio identifier for vhost devices */
+ ovsrcu_index vid;
+
+ /* True if vHost device is 'up' and has been reconfigured at least once */
+ bool vhost_reconfigured;
+ /* 3 pad bytes here. */
+ );
+
+ PADDED_MEMBERS(CACHE_LINE_SIZE,
+ /* Identifier used to distinguish vhost devices from each other. */
+ char vhost_id[PATH_MAX];
+ );
+
+ PADDED_MEMBERS(CACHE_LINE_SIZE,
+ struct netdev up;
+ /* In dpdk_list. */
+ struct ovs_list list_node OVS_GUARDED_BY(dpdk_mutex);
+
+ /* QoS configuration and lock for the device */
+ OVSRCU_TYPE(struct qos_conf *) qos_conf;
+
+ /* Ingress Policer */
+ OVSRCU_TYPE(struct ingress_policer *) ingress_policer;
+ uint32_t policer_rate;
+ uint32_t policer_burst;
+ );
+
+ PADDED_MEMBERS(CACHE_LINE_SIZE,
+ struct netdev_stats stats;
+ /* Protects stats */
+ rte_spinlock_t stats_lock;
+ /* 44 pad bytes here. */
+ );
+
+ PADDED_MEMBERS(CACHE_LINE_SIZE,
+ /* The following properties cannot be changed when a device is running,
+ * so we remember the request and update them next time
+ * netdev_dpdk*_reconfigure() is called */
+ int requested_mtu;
+ int requested_n_txq;
+ int requested_n_rxq;
+ int requested_rxq_size;
+ int requested_txq_size;
+
+ /* Number of rx/tx descriptors for physical devices */
+ int rxq_size;
+ int txq_size;
+
+ /* Socket ID detected when vHost device is brought up */
+ int requested_socket_id;
+
+ /* Denotes whether vHost port is client/server mode */
+ uint64_t vhost_driver_flags;
+
+ /* DPDK-ETH Flow control */
+ struct rte_eth_fc_conf fc_conf;
+
+ /* DPDK-ETH hardware offload features,
+ * from the enum set 'dpdk_hw_ol_features' */
+ uint32_t hw_ol_features;
+ );
};
struct netdev_rxq_dpdk {
dp_packet_init_dpdk((struct dp_packet *) pkt, pkt->buf_len);
}
-static struct dpdk_mp *
-dpdk_mp_create(int socket_id, int mtu)
-{
- struct dpdk_mp *dmp;
- unsigned mp_size;
- char *mp_name;
-
- dmp = dpdk_rte_mzalloc(sizeof *dmp);
- if (!dmp) {
- return NULL;
- }
- dmp->socket_id = socket_id;
- dmp->mtu = mtu;
- dmp->refcount = 1;
- /* XXX: this is a really rough method of provisioning memory.
- * It's impossible to determine what the exact memory requirements are
- * when the number of ports and rxqs that utilize a particular mempool can
- * change dynamically at runtime. For now, use this rough heurisitic.
+/* Returns a valid pointer when either of the following is true:
+ * - a new mempool was just created;
+ * - a matching mempool already exists. */
+static struct rte_mempool *
+dpdk_mp_create(struct netdev_dpdk *dev, int mtu)
+{
+ char mp_name[RTE_MEMPOOL_NAMESIZE];
+ const char *netdev_name = netdev_get_name(&dev->up);
+ int socket_id = dev->requested_socket_id;
+ uint32_t n_mbufs;
+ uint32_t hash = hash_string(netdev_name, 0);
+ struct rte_mempool *mp = NULL;
+
+ /*
+ * XXX: rough estimation of number of mbufs required for this port:
+ * <packets required to fill the device rxqs>
+ * + <packets that could be stuck on other ports txqs>
+ * + <packets in the pmd threads>
+ * + <additional memory for corner cases>
*/
- if (mtu >= ETHER_MTU) {
- mp_size = MAX_NB_MBUF;
- } else {
- mp_size = MIN_NB_MBUF;
- }
+ n_mbufs = dev->requested_n_rxq * dev->requested_rxq_size
+ + dev->requested_n_txq * dev->requested_txq_size
+ + MIN(RTE_MAX_LCORE, dev->requested_n_rxq) * NETDEV_MAX_BURST
+ + MIN_NB_MBUF;
+ ovs_mutex_lock(&dpdk_mp_mutex);
do {
- mp_name = xasprintf("ovs_mp_%d_%d_%u", dmp->mtu, dmp->socket_id,
- mp_size);
-
- dmp->mp = rte_pktmbuf_pool_create(mp_name, mp_size,
- MP_CACHE_SZ,
- sizeof (struct dp_packet)
- - sizeof (struct rte_mbuf),
- MBUF_SIZE(mtu)
- - sizeof(struct dp_packet),
- socket_id);
- if (dmp->mp) {
- VLOG_DBG("Allocated \"%s\" mempool with %u mbufs",
- mp_name, mp_size);
- }
- free(mp_name);
- if (dmp->mp) {
- /* rte_pktmbuf_pool_create has done some initialization of the
- * rte_mbuf part of each dp_packet, while ovs_rte_pktmbuf_init
- * initializes some OVS specific fields of dp_packet.
- */
- rte_mempool_obj_iter(dmp->mp, ovs_rte_pktmbuf_init, NULL);
- return dmp;
+ /* Full DPDK memory pool name must be unique and cannot be
+ * longer than RTE_MEMPOOL_NAMESIZE. */
+ int ret = snprintf(mp_name, RTE_MEMPOOL_NAMESIZE,
+ "ovs%08x%02d%05d%07u",
+ hash, socket_id, mtu, n_mbufs);
+ if (ret < 0 || ret >= RTE_MEMPOOL_NAMESIZE) {
+ VLOG_DBG("snprintf returned %d. "
+ "Failed to generate a mempool name for \"%s\". "
+ "Hash:0x%x, socket_id: %d, mtu:%d, mbufs:%u.",
+ ret, netdev_name, hash, socket_id, mtu, n_mbufs);
+ break;
}
- } while (rte_errno == ENOMEM && (mp_size /= 2) >= MIN_NB_MBUF);
- rte_free(dmp);
- return NULL;
-}
+ VLOG_DBG("Port %s: Requesting a mempool of %u mbufs "
+ "on socket %d for %d Rx and %d Tx queues.",
+ netdev_name, n_mbufs, socket_id,
+ dev->requested_n_rxq, dev->requested_n_txq);
-static struct dpdk_mp *
-dpdk_mp_get(int socket_id, int mtu)
-{
- struct dpdk_mp *dmp;
+ mp = rte_pktmbuf_pool_create(mp_name, n_mbufs, MP_CACHE_SZ,
+ sizeof (struct dp_packet) - sizeof (struct rte_mbuf),
+ MBUF_SIZE(mtu) - sizeof(struct dp_packet), socket_id);
- ovs_mutex_lock(&dpdk_mp_mutex);
- LIST_FOR_EACH (dmp, list_node, &dpdk_mp_list) {
- if (dmp->socket_id == socket_id && dmp->mtu == mtu) {
- dmp->refcount++;
- goto out;
+ if (mp) {
+ VLOG_DBG("Allocated \"%s\" mempool with %u mbufs",
+ mp_name, n_mbufs);
+ /* rte_pktmbuf_pool_create has done some initialization of the
+ * rte_mbuf part of each dp_packet. Some OvS specific fields
+ * of the packet still need to be initialized by
+ * ovs_rte_pktmbuf_init. */
+ rte_mempool_obj_iter(mp, ovs_rte_pktmbuf_init, NULL);
+ } else if (rte_errno == EEXIST) {
+ /* A mempool with the same name already exists. We just
+ * retrieve its pointer to be returned to the caller. */
+ mp = rte_mempool_lookup(mp_name);
+ /* As the mempool create returned EEXIST we can expect the
+ * lookup has returned a valid pointer. If for some reason
+ * that's not the case we keep track of it. */
+ VLOG_DBG("A mempool with name \"%s\" already exists at %p.",
+ mp_name, mp);
+ } else {
+ VLOG_ERR("Failed mempool \"%s\" create request of %u mbufs",
+ mp_name, n_mbufs);
}
- }
+ } while (!mp && rte_errno == ENOMEM && (n_mbufs /= 2) >= MIN_NB_MBUF);
- dmp = dpdk_mp_create(socket_id, mtu);
- if (dmp) {
- ovs_list_push_back(&dpdk_mp_list, &dmp->list_node);
- }
-
-out:
ovs_mutex_unlock(&dpdk_mp_mutex);
-
- return dmp;
+ return mp;
}
+/* Release an existing mempool. */
static void
-dpdk_mp_put(struct dpdk_mp *dmp)
+dpdk_mp_free(struct rte_mempool *mp)
{
- if (!dmp) {
+ if (!mp) {
return;
}
ovs_mutex_lock(&dpdk_mp_mutex);
- ovs_assert(dmp->refcount);
-
- if (!--dmp->refcount) {
- ovs_list_remove(&dmp->list_node);
- rte_mempool_free(dmp->mp);
- rte_free(dmp);
- }
+ VLOG_DBG("Releasing \"%s\" mempool", mp->name);
+ rte_mempool_free(mp);
ovs_mutex_unlock(&dpdk_mp_mutex);
}
-/* Tries to allocate new mempool on requested_socket_id with
- * mbuf size corresponding to requested_mtu.
- * On success new configuration will be applied.
+/* Tries to allocate a new mempool - or re-use an existing one where
+ * appropriate - on requested_socket_id with a size determined by
+ * requested_mtu and requested Rx/Tx queues.
+ * On success - or when re-using an existing mempool - the new configuration
+ * will be applied.
* On error, device will be left unchanged. */
static int
netdev_dpdk_mempool_configure(struct netdev_dpdk *dev)
OVS_REQUIRES(dev->mutex)
{
uint32_t buf_size = dpdk_buf_size(dev->requested_mtu);
- struct dpdk_mp *mp;
+ struct rte_mempool *mp;
+ int ret = 0;
- mp = dpdk_mp_get(dev->requested_socket_id, FRAME_LEN_TO_MTU(buf_size));
+ mp = dpdk_mp_create(dev, FRAME_LEN_TO_MTU(buf_size));
if (!mp) {
VLOG_ERR("Failed to create memory pool for netdev "
"%s, with MTU %d on socket %d: %s\n",
dev->up.name, dev->requested_mtu, dev->requested_socket_id,
rte_strerror(rte_errno));
- return rte_errno;
+ ret = rte_errno;
} else {
- dpdk_mp_put(dev->dpdk_mp);
- dev->dpdk_mp = mp;
+ /* If a new MTU was requested and its rounded value equals the one
+ * that is currently used, then the existing mempool is returned. */
+ if (dev->mp != mp) {
+ /* A new mempool was created, release the previous one. */
+ dpdk_mp_free(dev->mp);
+ } else {
+ ret = EEXIST;
+ }
+ dev->mp = mp;
dev->mtu = dev->requested_mtu;
dev->socket_id = dev->requested_socket_id;
dev->max_packet_len = MTU_TO_FRAME_LEN(dev->mtu);
}
- return 0;
+ return ret;
}
static void
for (i = 0; i < n_rxq; i++) {
diag = rte_eth_rx_queue_setup(dev->port_id, i, dev->rxq_size,
- dev->socket_id, NULL,
- dev->dpdk_mp->mp);
+ dev->socket_id, NULL, dev->mp);
if (diag) {
VLOG_INFO("Interface %s rxq(%d) setup error: %s",
dev->up.name, i, rte_strerror(-diag));
memcpy(dev->hwaddr.ea, eth_addr.addr_bytes, ETH_ADDR_LEN);
rte_eth_link_get_nowait(dev->port_id, &dev->link);
- mbp_priv = rte_mempool_get_priv(dev->dpdk_mp->mp);
+ mbp_priv = rte_mempool_get_priv(dev->mp);
dev->buf_size = mbp_priv->mbuf_data_room_size - RTE_PKTMBUF_HEADROOM;
/* Get the Flow control configuration for DPDK-ETH */
OVS_EXCLUDED(dev->mutex)
{
rte_free(dev->tx_q);
- dpdk_mp_put(dev->dpdk_mp);
+ dpdk_mp_free(dev->mp);
ovs_list_remove(&dev->list_node);
free(ovsrcu_get_protected(struct ingress_policer *,
char *name = xmemdup0(devargs, strcspn(devargs, ","));
dpdk_port_t new_port_id = DPDK_ETH_PORT_ID_INVALID;
- if (!rte_eth_dev_count()
- || rte_eth_dev_get_port_by_name(name, &new_port_id)
+ if (rte_eth_dev_get_port_by_name(name, &new_port_id)
|| !rte_eth_dev_is_valid_port(new_port_id)) {
/* Device not found in DPDK, attempt to attach it */
if (!rte_eth_dev_attach(devargs, &new_port_id)) {
static int
netdev_dpdk_policer_run(struct rte_meter_srtcm *meter,
- struct rte_mbuf **pkts, int pkt_cnt)
+ struct rte_mbuf **pkts, int pkt_cnt,
+ bool may_steal)
{
int i = 0;
int cnt = 0;
}
cnt++;
} else {
- rte_pktmbuf_free(pkt);
+ if (may_steal) {
+ rte_pktmbuf_free(pkt);
+ }
}
}
static int
ingress_policer_run(struct ingress_policer *policer, struct rte_mbuf **pkts,
- int pkt_cnt)
+ int pkt_cnt, bool may_steal)
{
int cnt = 0;
rte_spinlock_lock(&policer->policer_lock);
- cnt = netdev_dpdk_policer_run(&policer->in_policer, pkts, pkt_cnt);
+ cnt = netdev_dpdk_policer_run(&policer->in_policer, pkts,
+ pkt_cnt, may_steal);
rte_spinlock_unlock(&policer->policer_lock);
return cnt;
struct dp_packet_batch *batch)
{
struct netdev_dpdk *dev = netdev_dpdk_cast(rxq->netdev);
- int qid = rxq->queue_id;
struct ingress_policer *policer = netdev_dpdk_get_ingress_policer(dev);
uint16_t nb_rx = 0;
uint16_t dropped = 0;
+ int qid = rxq->queue_id;
+ int vid = netdev_dpdk_get_vid(dev);
- if (OVS_UNLIKELY(!is_vhost_running(dev)
+ if (OVS_UNLIKELY(vid < 0 || !dev->vhost_reconfigured
|| !(dev->flags & NETDEV_UP))) {
return EAGAIN;
}
- nb_rx = rte_vhost_dequeue_burst(netdev_dpdk_get_vid(dev),
- qid * VIRTIO_QNUM + VIRTIO_TXQ,
- dev->dpdk_mp->mp,
+ nb_rx = rte_vhost_dequeue_burst(vid, qid * VIRTIO_QNUM + VIRTIO_TXQ,
+ dev->mp,
(struct rte_mbuf **) batch->packets,
NETDEV_MAX_BURST);
if (!nb_rx) {
dropped = nb_rx;
nb_rx = ingress_policer_run(policer,
(struct rte_mbuf **) batch->packets,
- nb_rx);
+ nb_rx, true);
dropped -= nb_rx;
}
nb_rx, dropped);
rte_spinlock_unlock(&dev->stats_lock);
- dp_packet_batch_init_cutlen(batch);
- batch->count = (int) nb_rx;
+ batch->count = nb_rx;
+ dp_packet_batch_init_packet_fields(batch);
+
return 0;
}
dropped = nb_rx;
nb_rx = ingress_policer_run(policer,
(struct rte_mbuf **) batch->packets,
- nb_rx);
+ nb_rx, true);
dropped -= nb_rx;
}
rte_spinlock_unlock(&dev->stats_lock);
}
- dp_packet_batch_init_cutlen(batch);
batch->count = nb_rx;
+ dp_packet_batch_init_packet_fields(batch);
return 0;
}
static inline int
netdev_dpdk_qos_run(struct netdev_dpdk *dev, struct rte_mbuf **pkts,
- int cnt)
+ int cnt, bool may_steal)
{
struct qos_conf *qos_conf = ovsrcu_get(struct qos_conf *, &dev->qos_conf);
if (qos_conf) {
rte_spinlock_lock(&qos_conf->lock);
- cnt = qos_conf->ops->qos_run(qos_conf, pkts, cnt);
+ cnt = qos_conf->ops->qos_run(qos_conf, pkts, cnt, may_steal);
rte_spinlock_unlock(&qos_conf->lock);
}
unsigned int total_pkts = cnt;
unsigned int dropped = 0;
int i, retries = 0;
+ int vid = netdev_dpdk_get_vid(dev);
qid = dev->tx_q[qid % netdev->n_txq].map;
- if (OVS_UNLIKELY(!is_vhost_running(dev) || qid < 0
+ if (OVS_UNLIKELY(vid < 0 || !dev->vhost_reconfigured || qid < 0
|| !(dev->flags & NETDEV_UP))) {
rte_spinlock_lock(&dev->stats_lock);
dev->stats.tx_dropped+= cnt;
cnt = netdev_dpdk_filter_packet_len(dev, cur_pkts, cnt);
/* Check has QoS has been configured for the netdev */
- cnt = netdev_dpdk_qos_run(dev, cur_pkts, cnt);
+ cnt = netdev_dpdk_qos_run(dev, cur_pkts, cnt, true);
dropped = total_pkts - cnt;
do {
int vhost_qid = qid * VIRTIO_QNUM + VIRTIO_RXQ;
unsigned int tx_pkts;
- tx_pkts = rte_vhost_enqueue_burst(netdev_dpdk_get_vid(dev),
- vhost_qid, cur_pkts, cnt);
+ tx_pkts = rte_vhost_enqueue_burst(vid, vhost_qid, cur_pkts, cnt);
if (OVS_LIKELY(tx_pkts)) {
/* Packets have been sent.*/
cnt -= tx_pkts;
dpdk_do_tx_copy(struct netdev *netdev, int qid, struct dp_packet_batch *batch)
OVS_NO_THREAD_SAFETY_ANALYSIS
{
+ const size_t batch_cnt = dp_packet_batch_size(batch);
#if !defined(__CHECKER__) && !defined(_WIN32)
- const size_t PKT_ARRAY_SIZE = batch->count;
+ const size_t PKT_ARRAY_SIZE = batch_cnt;
#else
/* Sparse or MSVC doesn't like variable length array. */
enum { PKT_ARRAY_SIZE = NETDEV_MAX_BURST };
#endif
struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
struct rte_mbuf *pkts[PKT_ARRAY_SIZE];
- int dropped = 0;
- int newcnt = 0;
- int i;
+ uint32_t cnt = batch_cnt;
+ uint32_t dropped = 0;
+
+ if (dev->type != DPDK_DEV_VHOST) {
+ /* Check if QoS has been configured for this netdev. */
+ cnt = netdev_dpdk_qos_run(dev, (struct rte_mbuf **) batch->packets,
+ batch_cnt, false);
+ dropped += batch_cnt - cnt;
+ }
- dp_packet_batch_apply_cutlen(batch);
+ uint32_t txcnt = 0;
- for (i = 0; i < batch->count; i++) {
- int size = dp_packet_size(batch->packets[i]);
+ for (uint32_t i = 0; i < cnt; i++) {
+ struct dp_packet *packet = batch->packets[i];
+ uint32_t size = dp_packet_size(packet);
if (OVS_UNLIKELY(size > dev->max_packet_len)) {
- VLOG_WARN_RL(&rl, "Too big size %d max_packet_len %d",
- (int) size, dev->max_packet_len);
+ VLOG_WARN_RL(&rl, "Too big size %u max_packet_len %d",
+ size, dev->max_packet_len);
dropped++;
continue;
}
- pkts[newcnt] = rte_pktmbuf_alloc(dev->dpdk_mp->mp);
-
- if (!pkts[newcnt]) {
- dropped += batch->count - i;
+ pkts[txcnt] = rte_pktmbuf_alloc(dev->mp);
+ if (OVS_UNLIKELY(!pkts[txcnt])) {
+ dropped += cnt - i;
break;
}
/* We have to do a copy for now */
- memcpy(rte_pktmbuf_mtod(pkts[newcnt], void *),
- dp_packet_data(batch->packets[i]), size);
-
- rte_pktmbuf_data_len(pkts[newcnt]) = size;
- rte_pktmbuf_pkt_len(pkts[newcnt]) = size;
+ memcpy(rte_pktmbuf_mtod(pkts[txcnt], void *),
+ dp_packet_data(packet), size);
+ dp_packet_set_size((struct dp_packet *)pkts[txcnt], size);
- newcnt++;
+ txcnt++;
}
- if (dev->type == DPDK_DEV_VHOST) {
- __netdev_dpdk_vhost_send(netdev, qid, (struct dp_packet **) pkts,
- newcnt);
- } else {
- unsigned int qos_pkts = newcnt;
-
- /* Check if QoS has been configured for this netdev. */
- newcnt = netdev_dpdk_qos_run(dev, pkts, newcnt);
-
- dropped += qos_pkts - newcnt;
- dropped += netdev_dpdk_eth_tx_burst(dev, qid, pkts, newcnt);
+ if (OVS_LIKELY(txcnt)) {
+ if (dev->type == DPDK_DEV_VHOST) {
+ __netdev_dpdk_vhost_send(netdev, qid, (struct dp_packet **) pkts,
+ txcnt);
+ } else {
+ dropped += netdev_dpdk_eth_tx_burst(dev, qid, pkts, txcnt);
+ }
}
if (OVS_UNLIKELY(dropped)) {
static int
netdev_dpdk_vhost_send(struct netdev *netdev, int qid,
struct dp_packet_batch *batch,
- bool may_steal, bool concurrent_txq OVS_UNUSED)
+ bool concurrent_txq OVS_UNUSED)
{
- if (OVS_UNLIKELY(!may_steal || batch->packets[0]->source != DPBUF_DPDK)) {
+ if (OVS_UNLIKELY(batch->packets[0]->source != DPBUF_DPDK)) {
dpdk_do_tx_copy(netdev, qid, batch);
- dp_packet_delete_batch(batch, may_steal);
+ dp_packet_delete_batch(batch, true);
} else {
- dp_packet_batch_apply_cutlen(batch);
__netdev_dpdk_vhost_send(netdev, qid, batch->packets, batch->count);
}
return 0;
static inline void
netdev_dpdk_send__(struct netdev_dpdk *dev, int qid,
- struct dp_packet_batch *batch, bool may_steal,
+ struct dp_packet_batch *batch,
bool concurrent_txq)
{
if (OVS_UNLIKELY(!(dev->flags & NETDEV_UP))) {
- dp_packet_delete_batch(batch, may_steal);
+ dp_packet_delete_batch(batch, true);
return;
}
rte_spinlock_lock(&dev->tx_q[qid].tx_lock);
}
- if (OVS_UNLIKELY(!may_steal ||
- batch->packets[0]->source != DPBUF_DPDK)) {
+ if (OVS_UNLIKELY(batch->packets[0]->source != DPBUF_DPDK)) {
struct netdev *netdev = &dev->up;
dpdk_do_tx_copy(netdev, qid, batch);
- dp_packet_delete_batch(batch, may_steal);
+ dp_packet_delete_batch(batch, true);
} else {
- int dropped;
- int cnt = batch->count;
+ int tx_cnt, dropped;
+ int batch_cnt = dp_packet_batch_size(batch);
struct rte_mbuf **pkts = (struct rte_mbuf **) batch->packets;
- dp_packet_batch_apply_cutlen(batch);
+ tx_cnt = netdev_dpdk_filter_packet_len(dev, pkts, batch_cnt);
+ tx_cnt = netdev_dpdk_qos_run(dev, pkts, tx_cnt, true);
+ dropped = batch_cnt - tx_cnt;
- cnt = netdev_dpdk_filter_packet_len(dev, pkts, cnt);
- cnt = netdev_dpdk_qos_run(dev, pkts, cnt);
- dropped = batch->count - cnt;
-
- dropped += netdev_dpdk_eth_tx_burst(dev, qid, pkts, cnt);
+ dropped += netdev_dpdk_eth_tx_burst(dev, qid, pkts, tx_cnt);
if (OVS_UNLIKELY(dropped)) {
rte_spinlock_lock(&dev->stats_lock);
static int
netdev_dpdk_eth_send(struct netdev *netdev, int qid,
- struct dp_packet_batch *batch, bool may_steal,
- bool concurrent_txq)
+ struct dp_packet_batch *batch, bool concurrent_txq)
{
struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
- netdev_dpdk_send__(dev, qid, batch, may_steal, concurrent_txq);
+ netdev_dpdk_send__(dev, qid, batch, concurrent_txq);
return 0;
}
smap_add_format(args, "max_vfs", "%u", dev_info.max_vfs);
smap_add_format(args, "max_vmdq_pools", "%u", dev_info.max_vmdq_pools);
+ /* Querying the DPDK library for iftype may be done in future, pending
+ * support; cf. RFC 3635 Section 3.2.4. */
+ enum { IF_TYPE_ETHERNETCSMACD = 6 };
+
+ smap_add_format(args, "if_type", "%"PRIu32, IF_TYPE_ETHERNETCSMACD);
+ smap_add_format(args, "if_descr", "%s %s", rte_version(),
+ dev_info.driver_name);
+
if (dev_info.pci_dev) {
smap_add_format(args, "pci-vendor_id", "0x%u",
dev_info.pci_dev->id.vendor_id);
if (argc > 2) {
struct netdev *netdev = netdev_from_name(argv[1]);
+
if (netdev && is_dpdk_class(netdev->netdev_class)) {
- struct netdev_dpdk *dpdk_dev = netdev_dpdk_cast(netdev);
+ struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
- ovs_mutex_lock(&dpdk_dev->mutex);
- netdev_dpdk_set_admin_state__(dpdk_dev, up);
- ovs_mutex_unlock(&dpdk_dev->mutex);
+ ovs_mutex_lock(&dev->mutex);
+ netdev_dpdk_set_admin_state__(dev, up);
+ ovs_mutex_unlock(&dev->mutex);
netdev_close(netdev);
} else {
return;
}
} else {
- struct netdev_dpdk *netdev;
+ struct netdev_dpdk *dev;
ovs_mutex_lock(&dpdk_mutex);
- LIST_FOR_EACH (netdev, list_node, &dpdk_list) {
- ovs_mutex_lock(&netdev->mutex);
- netdev_dpdk_set_admin_state__(netdev, up);
- ovs_mutex_unlock(&netdev->mutex);
+ LIST_FOR_EACH (dev, list_node, &dpdk_list) {
+ ovs_mutex_lock(&dev->mutex);
+ netdev_dpdk_set_admin_state__(dev, up);
+ ovs_mutex_unlock(&dev->mutex);
}
ovs_mutex_unlock(&dpdk_mutex);
}
{
int ret;
char *response;
- uint8_t port_id;
+ dpdk_port_t port_id;
char devname[RTE_ETH_NAME_MAX_LEN];
struct netdev_dpdk *dev;
ovs_mutex_lock(&dpdk_mutex);
- if (!rte_eth_dev_count() || rte_eth_dev_get_port_by_name(argv[1],
- &port_id)) {
+ if (rte_eth_dev_get_port_by_name(argv[1], &port_id)) {
response = xasprintf("Device '%s' not found in DPDK", argv[1]);
goto error;
}
free(response);
}
+static void
+netdev_dpdk_get_mempool_info(struct unixctl_conn *conn,
+ int argc, const char *argv[],
+ void *aux OVS_UNUSED)
+{
+ size_t size;
+ FILE *stream;
+ char *response = NULL;
+ struct netdev *netdev = NULL;
+
+ if (argc == 2) {
+ netdev = netdev_from_name(argv[1]);
+ if (!netdev || !is_dpdk_class(netdev->netdev_class)) {
+ unixctl_command_reply_error(conn, "Not a DPDK Interface");
+ goto out;
+ }
+ }
+
+ stream = open_memstream(&response, &size);
+ if (!stream) {
+ response = xasprintf("Unable to open memstream: %s.",
+ ovs_strerror(errno));
+ unixctl_command_reply_error(conn, response);
+ goto out;
+ }
+
+ if (netdev) {
+ struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
+
+ ovs_mutex_lock(&dev->mutex);
+ ovs_mutex_lock(&dpdk_mp_mutex);
+
+ rte_mempool_dump(stream, dev->mp);
+
+ ovs_mutex_unlock(&dpdk_mp_mutex);
+ ovs_mutex_unlock(&dev->mutex);
+ } else {
+ ovs_mutex_lock(&dpdk_mp_mutex);
+ rte_mempool_list_dump(stream);
+ ovs_mutex_unlock(&dpdk_mp_mutex);
+ }
+
+ fclose(stream);
+
+ unixctl_command_reply(conn, response);
+out:
+ free(response);
+ netdev_close(netdev);
+}
+
/*
* Set virtqueue flags so that we do not receive interrupts.
*/
"pci address of device", 1, 1,
netdev_dpdk_detach, NULL);
+ unixctl_command_register("netdev-dpdk/get-mempool-info",
+ "[netdev]", 0, 1,
+ netdev_dpdk_get_mempool_info, NULL);
+
ovsthread_once_done(&once);
}
static int
netdev_dpdk_ring_send(struct netdev *netdev, int qid,
- struct dp_packet_batch *batch, bool may_steal,
- bool concurrent_txq)
+ struct dp_packet_batch *batch, bool concurrent_txq)
{
struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
- unsigned i;
+ struct dp_packet *packet;
/* When using 'dpdkr' and sending to a DPDK ring, we want to ensure that
* the rss hash field is clear. This is because the same mbuf may be
* modified by the consumer of the ring and return into the datapath
* without recalculating the RSS hash. */
- for (i = 0; i < batch->count; i++) {
- dp_packet_mbuf_rss_flag_reset(batch->packets[i]);
+ DP_PACKET_BATCH_FOR_EACH (packet, batch) {
+ dp_packet_mbuf_rss_flag_reset(packet);
}
- netdev_dpdk_send__(dev, qid, batch, may_steal, concurrent_txq);
+ netdev_dpdk_send__(dev, qid, batch, concurrent_txq);
return 0;
}
}
static int
-egress_policer_run(struct qos_conf *conf, struct rte_mbuf **pkts, int pkt_cnt)
+egress_policer_run(struct qos_conf *conf, struct rte_mbuf **pkts, int pkt_cnt,
+ bool may_steal)
{
int cnt = 0;
struct egress_policer *policer =
CONTAINER_OF(conf, struct egress_policer, qos_conf);
- cnt = netdev_dpdk_policer_run(&policer->egress_meter, pkts, pkt_cnt);
+ cnt = netdev_dpdk_policer_run(&policer->egress_meter, pkts,
+ pkt_cnt, may_steal);
return cnt;
}
rte_eth_dev_stop(dev->port_id);
- if (dev->mtu != dev->requested_mtu
- || dev->socket_id != dev->requested_socket_id) {
- err = netdev_dpdk_mempool_configure(dev);
- if (err) {
- goto out;
- }
+ err = netdev_dpdk_mempool_configure(dev);
+ if (err && err != EEXIST) {
+ goto out;
}
netdev->n_txq = dev->requested_n_txq;
netdev_dpdk_remap_txqs(dev);
- if (dev->requested_socket_id != dev->socket_id
- || dev->requested_mtu != dev->mtu) {
- err = netdev_dpdk_mempool_configure(dev);
- if (err) {
- return err;
- } else {
- netdev_change_seq_changed(&dev->up);
- }
+ err = netdev_dpdk_mempool_configure(dev);
+ if (!err) {
+ /* A new mempool was created. */
+ netdev_change_seq_changed(&dev->up);
+ } else if (err != EEXIST){
+ return err;
}
-
if (netdev_dpdk_get_vid(dev) >= 0) {
if (dev->vhost_reconfigured == false) {
dev->vhost_reconfigured = true;
{
struct netdev_dpdk *dev = netdev_dpdk_cast(netdev);
int err;
+ uint64_t vhost_flags = 0;
ovs_mutex_lock(&dev->mutex);
*/
if (!(dev->vhost_driver_flags & RTE_VHOST_USER_CLIENT)
&& strlen(dev->vhost_id)) {
- /* Register client-mode device */
- err = rte_vhost_driver_register(dev->vhost_id,
- RTE_VHOST_USER_CLIENT);
+ /* Register client-mode device. */
+ vhost_flags |= RTE_VHOST_USER_CLIENT;
+
+ /* Enable IOMMU support, if explicitly requested. */
+ if (dpdk_vhost_iommu_enabled()) {
+ vhost_flags |= RTE_VHOST_USER_IOMMU_SUPPORT;
+ }
+ err = rte_vhost_driver_register(dev->vhost_id, vhost_flags);
if (err) {
VLOG_ERR("vhost-user device setup failure for device %s\n",
dev->vhost_id);
goto unlock;
} else {
/* Configuration successful */
- dev->vhost_driver_flags |= RTE_VHOST_USER_CLIENT;
+ dev->vhost_driver_flags |= vhost_flags;
VLOG_INFO("vHost User device '%s' created in 'client' mode, "
"using client socket '%s'",
dev->up.name, dev->vhost_id);