[PATCH 1/5] vhost: add per-virtqueue statistics support

[Maxime Coquelin <maxime.coquelin@redhat.com>]

This patch introduces new APIs for the application
to query and reset per-virtqueue statistics. The
patch also introduces generic counters.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
---
 doc/guides/prog_guide/vhost_lib.rst |  24 ++++++
 lib/vhost/rte_vhost.h               |  99 ++++++++++++++++++++++++
 lib/vhost/socket.c                  |   4 +-
 lib/vhost/version.map               |   4 +-
 lib/vhost/vhost.c                   | 112 +++++++++++++++++++++++++++-
 lib/vhost/vhost.h                   |  18 ++++-
 lib/vhost/virtio_net.c              |  55 ++++++++++++++
 7 files changed, 311 insertions(+), 5 deletions(-)

diff --git a/doc/guides/prog_guide/vhost_lib.rst b/doc/guides/prog_guide/vhost_lib.rst
index f287b76ebf..0f91215a90 100644
--- a/doc/guides/prog_guide/vhost_lib.rst
+++ b/doc/guides/prog_guide/vhost_lib.rst
@@ -130,6 +130,15 @@ The following is an overview of some key Vhost API functions:
 
     It is disabled by default.
 
+  - ``RTE_VHOST_USER_NET_STATS_ENABLE``
+
+  Per-virtqueue statistics collection will be enabled when this flag is set.
+  When enabled, the application may use rte_vhost_stats_get_names() and
+  rte_vhost_stats_get() to collect statistics, and rte_vhost_stats_reset() to
+  reset them.
+
+  It is disabled by default
+
 * ``rte_vhost_driver_set_features(path, features)``
 
   This function sets the feature bits the vhost-user driver supports. The
@@ -282,6 +291,21 @@ The following is an overview of some key Vhost API functions:
   Clear inflight packets which are submitted to DMA engine in vhost async data
   path. Completed packets are returned to applications through ``pkts``.
 
+* ``rte_vhost_vring_stats_get_names(int vid, uint16_t queue_id, struct rte_vhost_stat_name *names, unsigned int size)``
+
+  This function returns the names of the queue statistics. It requires
+  statistics collection to be enabled at registration time.
+
+* ``rte_vhost_vring_stats_get(int vid, uint16_t queue_id, struct rte_vhost_stat *stats, unsigned int n)``
+
+  This function returns the queue statistics. It requires statistics
+  collection to be enabled at registration time.
+
+* ``rte_vhost_vring_stats_reset(int vid, uint16_t queue_id)``
+
+  This function resets the queue statistics. It requires statistics
+  collection to be enabled at registration time.
+
 Vhost-user Implementations
 --------------------------
 
diff --git a/lib/vhost/rte_vhost.h b/lib/vhost/rte_vhost.h
index c733f857c6..266048f3e5 100644
--- a/lib/vhost/rte_vhost.h
+++ b/lib/vhost/rte_vhost.h
@@ -39,6 +39,7 @@ extern "C" {
 #define RTE_VHOST_USER_LINEARBUF_SUPPORT	(1ULL << 6)
 #define RTE_VHOST_USER_ASYNC_COPY	(1ULL << 7)
 #define RTE_VHOST_USER_NET_COMPLIANT_OL_FLAGS	(1ULL << 8)
+#define RTE_VHOST_USER_NET_STATS_ENABLE	(1ULL << 9)
 
 /* Features. */
 #ifndef VIRTIO_NET_F_GUEST_ANNOUNCE
@@ -321,6 +322,32 @@ struct rte_vhost_power_monitor_cond {
 	uint8_t match;
 };
 
+/** Maximum name length for the statistics counters */
+#define RTE_VHOST_STATS_NAME_SIZE 64
+
+/**
+ * Vhost virtqueue statistics structure
+ *
+ * This structure is used by rte_vhost_vring_stats_get() to provide
+ * virtqueue statistics to the calling application.
+ * It maps a name ID, corresponding to an index in the array returned
+ * by rte_vhost_vring_stats_get_names(), to a statistic value.
+ */
+struct rte_vhost_stat {
+	uint64_t id;    /**< The index in xstats name array. */
+	uint64_t value; /**< The statistic counter value. */
+};
+
+/**
+ * Vhost virtqueue statistic name element
+ *
+ * This structure is used by rte_vhost_vring_stats_get_names() to
+ * provide virtqueue statistics names to the calling application.
+ */
+struct rte_vhost_stat_name {
+	char name[RTE_VHOST_STATS_NAME_SIZE]; /**< The statistic name. */
+};
+
 /**
  * Convert guest physical address to host virtual address
  *
@@ -1063,6 +1090,78 @@ __rte_experimental
 int
 rte_vhost_slave_config_change(int vid, bool need_reply);
 
+/**
+ * Retrieve names of statistics of a Vhost virtqueue.
+ *
+ * There is an assumption that 'stat_names' and 'stats' arrays are matched
+ * by array index: stats_names[i].name => stats[i].value
+ *
+ * @param vid
+ *   vhost device ID
+ * @param queue_id
+ *   vhost queue index
+ * @param name
+ *   array of at least size elements to be filled.
+ *   If set to NULL, the function returns the required number of elements.
+ * @param size
+ *   The number of elements in stats_names array.
+ * @return
+ *  - Success if greater than 0 and lower or equal to *size*. The return value
+ *    indicates the number of elements filled in the *names* array.
+ *  - Failure if greater than *size*. The return value indicates the number of
+ *    elements the *names* array that should be given to succeed.
+ *  - Failure if lower than 0. The device ID or queue ID is invalid or
+ +    statistics collection is not enabled.
+ */
+__rte_experimental
+int
+rte_vhost_vring_stats_get_names(int vid, uint16_t queue_id,
+		struct rte_vhost_stat_name *name, unsigned int size);
+
+/**
+ * Retrieve statistics of a Vhost virtqueue.
+ *
+ * There is an assumption that 'stat_names' and 'stats' arrays are matched
+ * by array index: stats_names[i].name => stats[i].value
+ *
+ * @param vid
+ *   vhost device ID
+ * @param queue_id
+ *   vhost queue index
+ * @param stats
+ *   A pointer to a table of structure of type rte_vhost_stat to be filled with
+ *   virtqueue statistics ids and values.
+ * @param n
+ *   The number of elements in stats array.
+ * @return
+ *  - Success if greater than 0 and lower or equal to *n*. The return value
+ *    indicates the number of elements filled in the *stats* array.
+ *  - Failure if greater than *n*. The return value indicates the number of
+ *    elements the *stats* array that should be given to succeed.
+ *  - Failure if lower than 0. The device ID or queue ID is invalid, or
+ *    statistics collection is not enabled.
+ */
+__rte_experimental
+int
+rte_vhost_vring_stats_get(int vid, uint16_t queue_id,
+		struct rte_vhost_stat *stats, unsigned int n);
+
+/**
+ * Reset statistics of a Vhost virtqueue.
+ *
+ * @param vid
+ *   vhost device ID
+ * @param queue_id
+ *   vhost queue index
+ * @return
+ *  - Success if 0. Statistics have been reset.
+ *  - Failure if lower than 0. The device ID or queue ID is invalid, or
+ *    statistics collection is not enabled.
+ */
+__rte_experimental
+int
+rte_vhost_vring_stats_reset(int vid, uint16_t queue_id);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/lib/vhost/socket.c b/lib/vhost/socket.c
index b304339de9..baf6c338d3 100644
--- a/lib/vhost/socket.c
+++ b/lib/vhost/socket.c
@@ -42,6 +42,7 @@ struct vhost_user_socket {
 	bool linearbuf;
 	bool async_copy;
 	bool net_compliant_ol_flags;
+	bool stats_enabled;
 
 	/*
 	 * The "supported_features" indicates the feature bits the
@@ -227,7 +228,7 @@ vhost_user_add_connection(int fd, struct vhost_user_socket *vsocket)
 	vhost_set_ifname(vid, vsocket->path, size);
 
 	vhost_setup_virtio_net(vid, vsocket->use_builtin_virtio_net,
-		vsocket->net_compliant_ol_flags);
+		vsocket->net_compliant_ol_flags, vsocket->stats_enabled);
 
 	vhost_attach_vdpa_device(vid, vsocket->vdpa_dev);
 
@@ -863,6 +864,7 @@ rte_vhost_driver_register(const char *path, uint64_t flags)
 	vsocket->linearbuf = flags & RTE_VHOST_USER_LINEARBUF_SUPPORT;
 	vsocket->async_copy = flags & RTE_VHOST_USER_ASYNC_COPY;
 	vsocket->net_compliant_ol_flags = flags & RTE_VHOST_USER_NET_COMPLIANT_OL_FLAGS;
+	vsocket->stats_enabled = flags & RTE_VHOST_USER_NET_STATS_ENABLE;
 
 	if (vsocket->async_copy &&
 		(flags & (RTE_VHOST_USER_IOMMU_SUPPORT |
diff --git a/lib/vhost/version.map b/lib/vhost/version.map
index 5841315386..34df41556d 100644
--- a/lib/vhost/version.map
+++ b/lib/vhost/version.map
@@ -90,7 +90,9 @@ EXPERIMENTAL {
 
 	# added in 22.07
 	rte_vhost_async_get_inflight_thread_unsafe;
-
+	rte_vhost_vring_stats_get_names;
+	rte_vhost_vring_stats_get;
+	rte_vhost_vring_stats_reset;
 };
 
 INTERNAL {
diff --git a/lib/vhost/vhost.c b/lib/vhost/vhost.c
index df0bb9d043..b22f10de72 100644
--- a/lib/vhost/vhost.c
+++ b/lib/vhost/vhost.c
@@ -24,6 +24,28 @@
 struct virtio_net *vhost_devices[RTE_MAX_VHOST_DEVICE];
 pthread_mutex_t vhost_dev_lock = PTHREAD_MUTEX_INITIALIZER;
 
+struct vhost_vq_stats_name_off {
+	char name[RTE_VHOST_STATS_NAME_SIZE];
+	unsigned int offset;
+};
+
+static const struct vhost_vq_stats_name_off vhost_vq_stat_strings[] = {
+	{"good_packets",           offsetof(struct vhost_virtqueue, stats.packets)},
+	{"good_bytes",             offsetof(struct vhost_virtqueue, stats.bytes)},
+	{"multicast_packets",      offsetof(struct vhost_virtqueue, stats.multicast)},
+	{"broadcast_packets",      offsetof(struct vhost_virtqueue, stats.broadcast)},
+	{"undersize_packets",      offsetof(struct vhost_virtqueue, stats.size_bins[0])},
+	{"size_64_packets",        offsetof(struct vhost_virtqueue, stats.size_bins[1])},
+	{"size_65_127_packets",    offsetof(struct vhost_virtqueue, stats.size_bins[2])},
+	{"size_128_255_packets",   offsetof(struct vhost_virtqueue, stats.size_bins[3])},
+	{"size_256_511_packets",   offsetof(struct vhost_virtqueue, stats.size_bins[4])},
+	{"size_512_1023_packets",  offsetof(struct vhost_virtqueue, stats.size_bins[5])},
+	{"size_1024_1518_packets", offsetof(struct vhost_virtqueue, stats.size_bins[6])},
+	{"size_1519_max_packets",  offsetof(struct vhost_virtqueue, stats.size_bins[7])},
+};
+
+#define VHOST_NB_VQ_STATS RTE_DIM(vhost_vq_stat_strings)
+
 /* Called with iotlb_lock read-locked */
 uint64_t
 __vhost_iova_to_vva(struct virtio_net *dev, struct vhost_virtqueue *vq,
@@ -755,7 +777,7 @@ vhost_set_ifname(int vid, const char *if_name, unsigned int if_len)
 }
 
 void
-vhost_setup_virtio_net(int vid, bool enable, bool compliant_ol_flags)
+vhost_setup_virtio_net(int vid, bool enable, bool compliant_ol_flags, bool stats_enabled)
 {
 	struct virtio_net *dev = get_device(vid);
 
@@ -770,6 +792,10 @@ vhost_setup_virtio_net(int vid, bool enable, bool compliant_ol_flags)
 		dev->flags |= VIRTIO_DEV_LEGACY_OL_FLAGS;
 	else
 		dev->flags &= ~VIRTIO_DEV_LEGACY_OL_FLAGS;
+	if (stats_enabled)
+		dev->flags |= VIRTIO_DEV_STATS_ENABLED;
+	else
+		dev->flags &= ~VIRTIO_DEV_STATS_ENABLED;
 }
 
 void
@@ -1971,5 +1997,89 @@ rte_vhost_get_monitor_addr(int vid, uint16_t queue_id,
 	return 0;
 }
 
+
+int
+rte_vhost_vring_stats_get_names(int vid, uint16_t queue_id,
+		struct rte_vhost_stat_name *name, unsigned int size)
+{
+	struct virtio_net *dev = get_device(vid);
+	unsigned int i;
+
+	if (dev == NULL)
+		return -1;
+
+	if (queue_id >= dev->nr_vring)
+		return -1;
+
+	if (!(dev->flags & VIRTIO_DEV_STATS_ENABLED))
+		return -1;
+
+	if (name == NULL || size < VHOST_NB_VQ_STATS)
+		return VHOST_NB_VQ_STATS;
+
+	for (i = 0; i < VHOST_NB_VQ_STATS; i++)
+		snprintf(name[i].name, sizeof(name[i].name), "%s_q%u_%s",
+				(queue_id & 1) ? "rx" : "tx",
+				queue_id / 2, vhost_vq_stat_strings[i].name);
+
+	return VHOST_NB_VQ_STATS;
+}
+
+int
+rte_vhost_vring_stats_get(int vid, uint16_t queue_id,
+		struct rte_vhost_stat *stats, unsigned int n)
+{
+	struct virtio_net *dev = get_device(vid);
+	struct vhost_virtqueue *vq;
+	unsigned int i;
+
+	if (dev == NULL)
+		return -1;
+
+	if (queue_id >= dev->nr_vring)
+		return -1;
+
+	if (!(dev->flags & VIRTIO_DEV_STATS_ENABLED))
+		return -1;
+
+	if (stats == NULL || n < VHOST_NB_VQ_STATS)
+		return VHOST_NB_VQ_STATS;
+
+	vq = dev->virtqueue[queue_id];
+
+	rte_spinlock_lock(&vq->access_lock);
+	for (i = 0; i < VHOST_NB_VQ_STATS; i++) {
+		stats[i].value =
+			*(uint64_t *)(((char *)vq) + vhost_vq_stat_strings[i].offset);
+		stats[i].id = i;
+	}
+	rte_spinlock_unlock(&vq->access_lock);
+
+	return VHOST_NB_VQ_STATS;
+}
+
+int rte_vhost_vring_stats_reset(int vid, uint16_t queue_id)
+{
+	struct virtio_net *dev = get_device(vid);
+	struct vhost_virtqueue *vq;
+
+	if (dev == NULL)
+		return -1;
+
+	if (queue_id >= dev->nr_vring)
+		return -1;
+
+	if (!(dev->flags & VIRTIO_DEV_STATS_ENABLED))
+		return -1;
+
+	vq = dev->virtqueue[queue_id];
+
+	rte_spinlock_lock(&vq->access_lock);
+	memset(&vq->stats, 0, sizeof(vq->stats));
+	rte_spinlock_unlock(&vq->access_lock);
+
+	return 0;
+}
+
 RTE_LOG_REGISTER_SUFFIX(vhost_config_log_level, config, INFO);
 RTE_LOG_REGISTER_SUFFIX(vhost_data_log_level, data, WARNING);
diff --git a/lib/vhost/vhost.h b/lib/vhost/vhost.h
index a9edc271aa..01b97011aa 100644
--- a/lib/vhost/vhost.h
+++ b/lib/vhost/vhost.h
@@ -37,6 +37,8 @@
 #define VIRTIO_DEV_FEATURES_FAILED ((uint32_t)1 << 4)
 /* Used to indicate that the virtio_net tx code should fill TX ol_flags */
 #define VIRTIO_DEV_LEGACY_OL_FLAGS ((uint32_t)1 << 5)
+/*  Used to indicate the application has requested statistics collection */
+#define VIRTIO_DEV_STATS_ENABLED ((uint32_t)1 << 6)
 
 /* Backend value set by guest. */
 #define VIRTIO_DEV_STOPPED -1
@@ -121,6 +123,18 @@ struct vring_used_elem_packed {
 	uint32_t count;
 };
 
+/**
+ * Virtqueue statistics
+ */
+struct virtqueue_stats {
+	uint64_t packets;
+	uint64_t bytes;
+	uint64_t multicast;
+	uint64_t broadcast;
+	/* Size bins in array as RFC 2819, undersized [0], 64 [1], etc */
+	uint64_t size_bins[8];
+};
+
 /**
  * iovec
  */
@@ -305,6 +319,7 @@ struct vhost_virtqueue {
 #define VIRTIO_UNINITIALIZED_NOTIF	(-1)
 
 	struct vhost_vring_addr ring_addrs;
+	struct virtqueue_stats	stats;
 } __rte_cache_aligned;
 
 /* Virtio device status as per Virtio specification */
@@ -780,7 +795,7 @@ int alloc_vring_queue(struct virtio_net *dev, uint32_t vring_idx);
 void vhost_attach_vdpa_device(int vid, struct rte_vdpa_device *dev);
 
 void vhost_set_ifname(int, const char *if_name, unsigned int if_len);
-void vhost_setup_virtio_net(int vid, bool enable, bool legacy_ol_flags);
+void vhost_setup_virtio_net(int vid, bool enable, bool legacy_ol_flags, bool stats_enabled);
 void vhost_enable_extbuf(int vid);
 void vhost_enable_linearbuf(int vid);
 int vhost_enable_guest_notification(struct virtio_net *dev,
@@ -957,5 +972,4 @@ mbuf_is_consumed(struct rte_mbuf *m)
 
 	return true;
 }
-
 #endif /* _VHOST_NET_CDEV_H_ */
diff --git a/lib/vhost/virtio_net.c b/lib/vhost/virtio_net.c
index 5f432b0d77..b1ea9fa4a5 100644
--- a/lib/vhost/virtio_net.c
+++ b/lib/vhost/virtio_net.c
@@ -47,6 +47,54 @@ is_valid_virt_queue_idx(uint32_t idx, int is_tx, uint32_t nr_vring)
 	return (is_tx ^ (idx & 1)) == 0 && idx < nr_vring;
 }
 
+/*
+ * This function must be called with virtqueue's access_lock taken.
+ */
+static inline void
+vhost_queue_stats_update(struct virtio_net *dev, struct vhost_virtqueue *vq,
+		struct rte_mbuf **pkts, uint16_t count)
+{
+	struct virtqueue_stats *stats = &vq->stats;
+	int i;
+
+	if (!(dev->flags & VIRTIO_DEV_STATS_ENABLED))
+		return;
+
+	for (i = 0; i < count; i++) {
+		struct rte_ether_addr *ea;
+		struct rte_mbuf *pkt = pkts[i];
+		uint32_t pkt_len = rte_pktmbuf_pkt_len(pkt);
+
+		stats->packets++;
+		stats->bytes += pkt_len;
+
+		if (pkt_len == 64) {
+			stats->size_bins[1]++;
+		} else if (pkt_len > 64 && pkt_len < 1024) {
+			uint32_t bin;
+
+			/* count zeros, and offset into correct bin */
+			bin = (sizeof(pkt_len) * 8) - __builtin_clz(pkt_len) - 5;
+			stats->size_bins[bin]++;
+		} else {
+			if (pkt_len < 64)
+				stats->size_bins[0]++;
+			else if (pkt_len < 1519)
+				stats->size_bins[6]++;
+			else
+				stats->size_bins[7]++;
+		}
+
+		ea = rte_pktmbuf_mtod(pkt, struct rte_ether_addr *);
+		if (rte_is_multicast_ether_addr(ea)) {
+			if (rte_is_broadcast_ether_addr(ea))
+				stats->broadcast++;
+			else
+				stats->multicast++;
+		}
+	}
+}
+
 static __rte_always_inline int64_t
 vhost_async_dma_transfer_one(struct virtio_net *dev, struct vhost_virtqueue *vq,
 		int16_t dma_id, uint16_t vchan_id, uint16_t flag_idx,
@@ -1509,6 +1557,8 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id,
 	else
 		nb_tx = virtio_dev_rx_split(dev, vq, pkts, count);
 
+	vhost_queue_stats_update(dev, vq, pkts, nb_tx);
+
 out:
 	if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM))
 		vhost_user_iotlb_rd_unlock(vq);
@@ -2064,6 +2114,8 @@ rte_vhost_poll_enqueue_completed(int vid, uint16_t queue_id,
 
 	n_pkts_cpl = vhost_poll_enqueue_completed(dev, queue_id, pkts, count, dma_id, vchan_id);
 
+	vhost_queue_stats_update(dev, vq, pkts, n_pkts_cpl);
+
 out:
 	rte_spinlock_unlock(&vq->access_lock);
 
@@ -3113,6 +3165,7 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
 		 * learning table will get updated first.
 		 */
 		pkts[0] = rarp_mbuf;
+		vhost_queue_stats_update(dev, vq, pkts, 1);
 		pkts++;
 		count -= 1;
 	}
@@ -3129,6 +3182,8 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
 			count = virtio_dev_tx_split_compliant(dev, vq, mbuf_pool, pkts, count);
 	}
 
+	vhost_queue_stats_update(dev, vq, pkts, count);
+
 out:
 	if (dev->features & (1ULL << VIRTIO_F_IOMMU_PLATFORM))
 		vhost_user_iotlb_rd_unlock(vq);
-- 
2.35.1