netvsc: report per-channel stats in ethtool statistics

Report packets and bytes transferred through a vmbus channel via ethtool.
This supersedes need for per-cpu statistics.

Example:
$ ethtool -S eth0
NIC statistics:
...
     tx_queue_0_packets: 3523179
     tx_queue_0_bytes: 505370920
     rx_queue_0_packets: 41430490
     rx_queue_0_bytes: 62714661254
     tx_queue_1_packets: 0
     tx_queue_1_bytes: 0
     rx_queue_1_packets: 0
     rx_queue_1_bytes: 0
...

Reviewed-by: Long Li <longli@microsoft.com>
Reviewed-by: K. Y. Srinivasan <kys@microsoft.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Simon Xiao <sixiao@microsoft.com>
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/drivers/net/hyperv/hyperv_net.h b/drivers/net/hyperv/hyperv_net.h
index 340f64233e2a803d9f5438e327e7a7d5763d2f1f..d3e73ac158aee6d3958b18618b85f0180f22abf7 100644 (file)
--- a/drivers/net/hyperv/hyperv_net.h
+++ b/drivers/net/hyperv/hyperv_net.h
@@ -699,8 +699,6 @@ struct net_device_context {
        u32 msg_enable; /* debug level */
 
        u32 tx_checksum_mask;
-       struct netvsc_stats __percpu *tx_stats;
-       struct netvsc_stats __percpu *rx_stats;
 
        /* Ethtool settings */
        u8 duplex;
@@ -725,6 +723,9 @@ struct netvsc_channel {
        struct multi_send_data msd;
        struct multi_recv_comp mrc;
        atomic_t queue_sends;
+
+       struct netvsc_stats tx_stats;
+       struct netvsc_stats rx_stats;
 };
 
 /* Per netvsc device */

diff --git a/drivers/net/hyperv/netvsc.c b/drivers/net/hyperv/netvsc.c
index 397aa1d88a6f49afa57e3edc2160f983ec6a4d0a..c70c4ac77b7470c87436874ee637ce0d2398d8f4 100644 (file)
--- a/drivers/net/hyperv/netvsc.c
+++ b/drivers/net/hyperv/netvsc.c
@@ -621,7 +621,7 @@ static void netvsc_send_tx_complete(struct netvsc_device *net_device,
                q_idx = packet->q_idx;
                channel = incoming_channel;
 
-               tx_stats = this_cpu_ptr(net_device_ctx->tx_stats);
+               tx_stats = &net_device->chan_table[q_idx].tx_stats;
 
                u64_stats_update_begin(&tx_stats->syncp);
                tx_stats->packets += packet->total_packets;

diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c
index 1dd13da79f02668feeb4a849076532d3277d36a4..fe0df72532a3c1d8a73afba1b44012532f18dd58 100644 (file)
--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c
@@ -641,9 +641,12 @@ int netvsc_recv_callback(struct net_device *net,
                         const struct ndis_pkt_8021q_info *vlan)
 {
        struct net_device_context *net_device_ctx = netdev_priv(net);
+       struct netvsc_device *net_device = net_device_ctx->nvdev;
        struct net_device *vf_netdev;
        struct sk_buff *skb;
        struct netvsc_stats *rx_stats;
+       u16 q_idx = channel->offermsg.offer.sub_channel_index;
+
 
        if (net->reg_state != NETREG_REGISTERED)
                return NVSP_STAT_FAIL;
@@ -669,15 +672,14 @@ int netvsc_recv_callback(struct net_device *net,
        }
 
        if (net != vf_netdev)
-               skb_record_rx_queue(skb,
-                                   channel->offermsg.offer.sub_channel_index);
+               skb_record_rx_queue(skb, q_idx);
 
        /*
         * Even if injecting the packet, record the statistics
         * on the synthetic device because modifying the VF device
         * statistics will not work correctly.
         */
-       rx_stats = this_cpu_ptr(net_device_ctx->rx_stats);
+       rx_stats = &net_device->chan_table[q_idx].rx_stats;
        u64_stats_update_begin(&rx_stats->syncp);
        rx_stats->packets++;
        rx_stats->bytes += len;
@@ -882,34 +884,39 @@ static void netvsc_get_stats64(struct net_device *net,
                               struct rtnl_link_stats64 *t)
 {
        struct net_device_context *ndev_ctx = netdev_priv(net);
-       int cpu;
-
-       for_each_possible_cpu(cpu) {
-               struct netvsc_stats *tx_stats = per_cpu_ptr(ndev_ctx->tx_stats,
-                                                           cpu);
-               struct netvsc_stats *rx_stats = per_cpu_ptr(ndev_ctx->rx_stats,
-                                                           cpu);
-               u64 tx_packets, tx_bytes, rx_packets, rx_bytes, rx_multicast;
+       struct netvsc_device *nvdev = ndev_ctx->nvdev;
+       int i;
+
+       if (!nvdev)
+               return;
+
+       for (i = 0; i < nvdev->num_chn; i++) {
+               const struct netvsc_channel *nvchan = &nvdev->chan_table[i];
+               const struct netvsc_stats *stats;
+               u64 packets, bytes, multicast;
                unsigned int start;
 
+               stats = &nvchan->tx_stats;
                do {
-                       start = u64_stats_fetch_begin_irq(&tx_stats->syncp);
-                       tx_packets = tx_stats->packets;
-                       tx_bytes = tx_stats->bytes;
-               } while (u64_stats_fetch_retry_irq(&tx_stats->syncp, start));
+                       start = u64_stats_fetch_begin_irq(&stats->syncp);
+                       packets = stats->packets;
+                       bytes = stats->bytes;
+               } while (u64_stats_fetch_retry_irq(&stats->syncp, start));
+
+               t->tx_bytes     += bytes;
+               t->tx_packets   += packets;
 
+               stats = &nvchan->rx_stats;
                do {
-                       start = u64_stats_fetch_begin_irq(&rx_stats->syncp);
-                       rx_packets = rx_stats->packets;
-                       rx_bytes = rx_stats->bytes;
-                       rx_multicast = rx_stats->multicast + rx_stats->broadcast;
-               } while (u64_stats_fetch_retry_irq(&rx_stats->syncp, start));
-
-               t->tx_bytes     += tx_bytes;
-               t->tx_packets   += tx_packets;
-               t->rx_bytes     += rx_bytes;
-               t->rx_packets   += rx_packets;
-               t->multicast    += rx_multicast;
+                       start = u64_stats_fetch_begin_irq(&stats->syncp);
+                       packets = stats->packets;
+                       bytes = stats->bytes;
+                       multicast = stats->multicast + stats->broadcast;
+               } while (u64_stats_fetch_retry_irq(&stats->syncp, start));
+
+               t->rx_bytes     += bytes;
+               t->rx_packets   += packets;
+               t->multicast    += multicast;
        }
 
        t->tx_dropped   = net->stats.tx_dropped;
@@ -954,11 +961,19 @@ static const struct {
        { "tx_busy",      offsetof(struct netvsc_ethtool_stats, tx_busy) },
 };
 
+#define NETVSC_GLOBAL_STATS_LEN        ARRAY_SIZE(netvsc_stats)
+
+/* 4 statistics per queue (rx/tx packets/bytes) */
+#define NETVSC_QUEUE_STATS_LEN(dev) ((dev)->num_chn * 4)
+
 static int netvsc_get_sset_count(struct net_device *dev, int string_set)
 {
+       struct net_device_context *ndc = netdev_priv(dev);
+       struct netvsc_device *nvdev = ndc->nvdev;
+
        switch (string_set) {
        case ETH_SS_STATS:
-               return ARRAY_SIZE(netvsc_stats);
+               return NETVSC_GLOBAL_STATS_LEN + NETVSC_QUEUE_STATS_LEN(nvdev);
        default:
                return -EINVAL;
        }
@@ -968,22 +983,63 @@ static void netvsc_get_ethtool_stats(struct net_device *dev,
                                     struct ethtool_stats *stats, u64 *data)
 {
        struct net_device_context *ndc = netdev_priv(dev);
+       struct netvsc_device *nvdev = ndc->nvdev;
        const void *nds = &ndc->eth_stats;
-       int i;
+       const struct netvsc_stats *qstats;
+       unsigned int start;
+       u64 packets, bytes;
+       int i, j;
 
-       for (i = 0; i < ARRAY_SIZE(netvsc_stats); i++)
+       for (i = 0; i < NETVSC_GLOBAL_STATS_LEN; i++)
                data[i] = *(unsigned long *)(nds + netvsc_stats[i].offset);
+
+       for (j = 0; j < nvdev->num_chn; j++) {
+               qstats = &nvdev->chan_table[j].tx_stats;
+
+               do {
+                       start = u64_stats_fetch_begin_irq(&qstats->syncp);
+                       packets = qstats->packets;
+                       bytes = qstats->bytes;
+               } while (u64_stats_fetch_retry_irq(&qstats->syncp, start));
+               data[i++] = packets;
+               data[i++] = bytes;
+
+               qstats = &nvdev->chan_table[j].rx_stats;
+               do {
+                       start = u64_stats_fetch_begin_irq(&qstats->syncp);
+                       packets = qstats->packets;
+                       bytes = qstats->bytes;
+               } while (u64_stats_fetch_retry_irq(&qstats->syncp, start));
+               data[i++] = packets;
+               data[i++] = bytes;
+       }
 }
 
 static void netvsc_get_strings(struct net_device *dev, u32 stringset, u8 *data)
 {
+       struct net_device_context *ndc = netdev_priv(dev);
+       struct netvsc_device *nvdev = ndc->nvdev;
+       u8 *p = data;
        int i;
 
        switch (stringset) {
        case ETH_SS_STATS:
                for (i = 0; i < ARRAY_SIZE(netvsc_stats); i++)
-                       memcpy(data + i * ETH_GSTRING_LEN,
+                       memcpy(p + i * ETH_GSTRING_LEN,
                               netvsc_stats[i].name, ETH_GSTRING_LEN);
+
+               p += i * ETH_GSTRING_LEN;
+               for (i = 0; i < nvdev->num_chn; i++) {
+                       sprintf(p, "tx_queue_%u_packets", i);
+                       p += ETH_GSTRING_LEN;
+                       sprintf(p, "tx_queue_%u_bytes", i);
+                       p += ETH_GSTRING_LEN;
+                       sprintf(p, "rx_queue_%u_packets", i);
+                       p += ETH_GSTRING_LEN;
+                       sprintf(p, "rx_queue_%u_bytes", i);
+                       p += ETH_GSTRING_LEN;
+               }
+
                break;
        }
 }
@@ -1237,15 +1293,6 @@ out_unlock:
        rtnl_unlock();
 }
 
-static void netvsc_free_netdev(struct net_device *netdev)
-{
-       struct net_device_context *net_device_ctx = netdev_priv(netdev);
-
-       free_percpu(net_device_ctx->tx_stats);
-       free_percpu(net_device_ctx->rx_stats);
-       free_netdev(netdev);
-}
-
 static struct net_device *get_netvsc_bymac(const u8 *mac)
 {
        struct net_device *dev;
@@ -1423,18 +1470,6 @@ static int netvsc_probe(struct hv_device *dev,
                netdev_dbg(net, "netvsc msg_enable: %d\n",
                           net_device_ctx->msg_enable);
 
-       net_device_ctx->tx_stats = netdev_alloc_pcpu_stats(struct netvsc_stats);
-       if (!net_device_ctx->tx_stats) {
-               free_netdev(net);
-               return -ENOMEM;
-       }
-       net_device_ctx->rx_stats = netdev_alloc_pcpu_stats(struct netvsc_stats);
-       if (!net_device_ctx->rx_stats) {
-               free_percpu(net_device_ctx->tx_stats);
-               free_netdev(net);
-               return -ENOMEM;
-       }
-
        hv_set_drvdata(dev, net);
 
        net_device_ctx->start_remove = false;
@@ -1460,7 +1495,7 @@ static int netvsc_probe(struct hv_device *dev,
        ret = rndis_filter_device_add(dev, &device_info);
        if (ret != 0) {
                netdev_err(net, "unable to add netvsc device (ret %d)\n", ret);
-               netvsc_free_netdev(net);
+               free_netdev(net);
                hv_set_drvdata(dev, NULL);
                return ret;
        }
@@ -1487,7 +1522,7 @@ static int netvsc_probe(struct hv_device *dev,
        if (ret != 0) {
                pr_err("Unable to register netdev.\n");
                rndis_filter_device_remove(dev, nvdev);
-               netvsc_free_netdev(net);
+               free_netdev(net);
        }
 
        return ret;
@@ -1530,7 +1565,7 @@ static int netvsc_remove(struct hv_device *dev)
 
        hv_set_drvdata(dev, NULL);
 
-       netvsc_free_netdev(net);
+       free_netdev(net);
        return 0;
 }