diff mbox

[2/2] openvswitch: Allocate memory for ovs internal device stats.

Message ID 1444920855-16409-1-git-send-email-james.morse@arm.com
State Changes Requested, archived
Delegated to: David Miller
Headers show

Commit Message

James Morse Oct. 15, 2015, 2:54 p.m. UTC 
"openvswitch: Remove vport stats" removed the per-vport statistics, in
order to use the netdev's statistics fields.
"openvswitch: Fix ovs_vport_get_stats()" fixed the export of these stats
to user-space, by using the provided netdev_ops to collate them - but ovs
internal devices still use an unallocated dev->tstats field to count
packets, which are not exported by this api.

Allocate the dev->tstats field for ovs internal devices, and wire up
ndo_get_stats64 with the original implementation of
ovs_vport_get_stats().

On its own, "openvswitch: Fix ovs_vport_get_stats()" fixes the OOPs,
unmasking a full-on panic on arm64:

Comments

Pravin B Shelar Oct. 15, 2015, 5:19 p.m. UTC | #1 
On Thu, Oct 15, 2015 at 7:54 AM, James Morse <james.morse@arm.com> wrote:
> "openvswitch: Remove vport stats" removed the per-vport statistics, in
> order to use the netdev's statistics fields.
> "openvswitch: Fix ovs_vport_get_stats()" fixed the export of these stats
> to user-space, by using the provided netdev_ops to collate them - but ovs
> internal devices still use an unallocated dev->tstats field to count
> packets, which are not exported by this api.
>
> Allocate the dev->tstats field for ovs internal devices, and wire up
> ndo_get_stats64 with the original implementation of
> ovs_vport_get_stats().
>
> On its own, "openvswitch: Fix ovs_vport_get_stats()" fixes the OOPs,
> unmasking a full-on panic on arm64:
> =============%<==============
> [<ffffffbffc00ce4c>] internal_dev_recv+0xa8/0x170 [openvswitch]
> [<ffffffbffc0008b4>] do_output.isra.31+0x60/0x19c [openvswitch]
> [<ffffffbffc000bf8>] do_execute_actions+0x208/0x11c0 [openvswitch]
> [<ffffffbffc001c78>] ovs_execute_actions+0xc8/0x238 [openvswitch]
> [<ffffffbffc003dfc>] ovs_packet_cmd_execute+0x21c/0x288 [openvswitch]
> [<ffffffc0005e8c5c>] genl_family_rcv_msg+0x1b0/0x310
> [<ffffffc0005e8e60>] genl_rcv_msg+0xa4/0xe4
> [<ffffffc0005e7ddc>] netlink_rcv_skb+0xb0/0xdc
> [<ffffffc0005e8a94>] genl_rcv+0x38/0x50
> [<ffffffc0005e76c0>] netlink_unicast+0x164/0x210
> [<ffffffc0005e7b70>] netlink_sendmsg+0x304/0x368
> [<ffffffc0005a21c0>] sock_sendmsg+0x30/0x4c
> [SNIP]
> Kernel panic - not syncing: Fatal exception in interrupt
> =============%<==============
>
> Signed-off-by: James Morse <james.morse@arm.com>
> Fixes: 8c876639c985 ("openvswitch: Remove vport stats.")
> ---
> Hi netdev,
>
> "openvswitch: Fix ovs_vport_get_stats()"  - already applied according to
> [0] (where?), is an incomplete fix for the issue in "openvswitch: Remove
> vport stats.". Use of an unallocated dev->tstats remains for ovs internal
> devices, which causes a panic on arm64. Could this patch and "openvswitch:
> Fix ovs_vport_get_stats()" be considered as fixes for v4.3?
>
You need to target this patch for net branch. So that the fix will get
into v4.3.

> Thanks!
>
> James Morse
>
>
> [0] https://patchwork.ozlabs.org/patch/525827/
>
>  net/openvswitch/vport-internal_dev.c | 43 +++++++++++++++++++++++++++++++++++-
>  1 file changed, 42 insertions(+), 1 deletion(-)
>
> diff --git a/net/openvswitch/vport-internal_dev.c b/net/openvswitch/vport-internal_dev.c
> index 388b8a6bf112..758e53cb7a81 100644
> --- a/net/openvswitch/vport-internal_dev.c
> +++ b/net/openvswitch/vport-internal_dev.c
> @@ -106,12 +106,48 @@ static void internal_dev_destructor(struct net_device *dev)
>         free_netdev(dev);
>  }
>
> +static struct rtnl_link_stats64 *
> +internal_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
> +{
> +       int i;
> +
> +       memset(stats, 0, sizeof(*stats));
> +       stats->rx_errors  = dev->stats.rx_errors;
> +       stats->tx_errors  = dev->stats.tx_errors;
> +       stats->tx_dropped = dev->stats.tx_dropped;
> +       stats->rx_dropped = dev->stats.rx_dropped;
> +
> +       stats->rx_dropped += atomic_long_read(&dev->rx_dropped);
> +       stats->tx_dropped += atomic_long_read(&dev->tx_dropped);
> +
dev_get_stats() already account for dev->rx and tx dropped packets. So
there is no need to add it here.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

=============%<==============
[<ffffffbffc00ce4c>] internal_dev_recv+0xa8/0x170 [openvswitch]
[<ffffffbffc0008b4>] do_output.isra.31+0x60/0x19c [openvswitch]
[<ffffffbffc000bf8>] do_execute_actions+0x208/0x11c0 [openvswitch]
[<ffffffbffc001c78>] ovs_execute_actions+0xc8/0x238 [openvswitch]
[<ffffffbffc003dfc>] ovs_packet_cmd_execute+0x21c/0x288 [openvswitch]
[<ffffffc0005e8c5c>] genl_family_rcv_msg+0x1b0/0x310
[<ffffffc0005e8e60>] genl_rcv_msg+0xa4/0xe4
[<ffffffc0005e7ddc>] netlink_rcv_skb+0xb0/0xdc
[<ffffffc0005e8a94>] genl_rcv+0x38/0x50
[<ffffffc0005e76c0>] netlink_unicast+0x164/0x210
[<ffffffc0005e7b70>] netlink_sendmsg+0x304/0x368
[<ffffffc0005a21c0>] sock_sendmsg+0x30/0x4c
[SNIP]
Kernel panic - not syncing: Fatal exception in interrupt
=============%<==============

Signed-off-by: James Morse <james.morse@arm.com>
Fixes: 8c876639c985 ("openvswitch: Remove vport stats.")
---
Hi netdev,

"openvswitch: Fix ovs_vport_get_stats()"  - already applied according to
[0] (where?), is an incomplete fix for the issue in "openvswitch: Remove
vport stats.". Use of an unallocated dev->tstats remains for ovs internal
devices, which causes a panic on arm64. Could this patch and "openvswitch:
Fix ovs_vport_get_stats()" be considered as fixes for v4.3?

Thanks!

James Morse


[0] https://patchwork.ozlabs.org/patch/525827/

 net/openvswitch/vport-internal_dev.c | 43 +++++++++++++++++++++++++++++++++++-
 1 file changed, 42 insertions(+), 1 deletion(-)

diff --git a/net/openvswitch/vport-internal_dev.c b/net/openvswitch/vport-internal_dev.c
index 388b8a6bf112..758e53cb7a81 100644
--- a/net/openvswitch/vport-internal_dev.c
+++ b/net/openvswitch/vport-internal_dev.c
@@ -106,12 +106,48 @@  static void internal_dev_destructor(struct net_device *dev)
 	free_netdev(dev);
 }
 
+static struct rtnl_link_stats64 *
+internal_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+	int i;
+
+	memset(stats, 0, sizeof(*stats));
+	stats->rx_errors  = dev->stats.rx_errors;
+	stats->tx_errors  = dev->stats.tx_errors;
+	stats->tx_dropped = dev->stats.tx_dropped;
+	stats->rx_dropped = dev->stats.rx_dropped;
+
+	stats->rx_dropped += atomic_long_read(&dev->rx_dropped);
+	stats->tx_dropped += atomic_long_read(&dev->tx_dropped);
+
+	for_each_possible_cpu(i) {
+		const struct pcpu_sw_netstats *percpu_stats;
+		struct pcpu_sw_netstats local_stats;
+		unsigned int start;
+
+		percpu_stats = per_cpu_ptr(dev->tstats, i);
+
+		do {
+			start = u64_stats_fetch_begin_irq(&percpu_stats->syncp);
+			local_stats = *percpu_stats;
+		} while (u64_stats_fetch_retry_irq(&percpu_stats->syncp, start));
+
+		stats->rx_bytes         += local_stats.rx_bytes;
+		stats->rx_packets       += local_stats.rx_packets;
+		stats->tx_bytes         += local_stats.tx_bytes;
+		stats->tx_packets       += local_stats.tx_packets;
+	}
+
+	return stats;
+}
+
 static const struct net_device_ops internal_dev_netdev_ops = {
 	.ndo_open = internal_dev_open,
 	.ndo_stop = internal_dev_stop,
 	.ndo_start_xmit = internal_dev_xmit,
 	.ndo_set_mac_address = eth_mac_addr,
 	.ndo_change_mtu = internal_dev_change_mtu,
+	.ndo_get_stats64 = internal_get_stats,
 };
 
 static struct rtnl_link_ops internal_dev_link_ops __read_mostly = {
@@ -161,6 +197,11 @@  static struct vport *internal_dev_create(const struct vport_parms *parms)
 		err = -ENOMEM;
 		goto error_free_vport;
 	}
+	vport->dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
+	if (!vport->dev->tstats) {
+		err = -ENOMEM;
+		goto error_free_vport;
+	}
 
 	dev_net_set(vport->dev, ovs_dp_get_net(vport->dp));
 	internal_dev = internal_dev_priv(vport->dev);
@@ -198,7 +239,7 @@  static void internal_dev_destroy(struct vport *vport)
 
 	/* unregister_netdevice() waits for an RCU grace period. */
 	unregister_netdevice(vport->dev);
-
+	free_percpu(vport->dev->tstats);
 	rtnl_unlock();
 }