Re: [PATCH] [PATCH v6 net-next] net: openvswitch: Add support to count upcall packets

From: wangchuanlei
Date: Sun Dec 04 2022 - 21:05:05 EST


Hi, Eelco,
Thank you for review again ! I will give a new version of patch based on your comments
today!

Best regards!
wangchuanlei

-------------------------------------------------------------------------


On 30 Nov 2022, at 10:15, wangchuanlei wrote:

> Add support to count upall packets, when kmod of openvswitch upcall to
> userspace , here count the number of packets for upcall succeed and
> failed, which is a better way to see how many packets upcalled to
> userspace(ovs-vswitchd) on every interfaces.
>
> Here modify format of code used by comments of v6.
>
> Changes since v4 & v5:
> - optimize the function used by comments
>
> Changes since v3:
> - use nested NLA_NESTED attribute in netlink message
>
> Changes since v2:
> - add count of upcall failed packets
>
> Changes since v1:
> - add count of upcall succeed packets
>
> Signed-off-by: wangchuanlei <wangchuanlei@xxxxxxxxxx>
> ---
> include/uapi/linux/openvswitch.h | 14 +++++++++
> net/openvswitch/datapath.c | 50 ++++++++++++++++++++++++++++++++
> net/openvswitch/vport.c | 44 ++++++++++++++++++++++++++++
> net/openvswitch/vport.h | 24 +++++++++++++++
> 4 files changed, 132 insertions(+)
>
> diff --git a/include/uapi/linux/openvswitch.h
> b/include/uapi/linux/openvswitch.h
> index 94066f87e9ee..8422ebf6885b 100644
> --- a/include/uapi/linux/openvswitch.h
> +++ b/include/uapi/linux/openvswitch.h
> @@ -277,11 +277,25 @@ enum ovs_vport_attr {
> OVS_VPORT_ATTR_PAD,
> OVS_VPORT_ATTR_IFINDEX,
> OVS_VPORT_ATTR_NETNSID,
> + OVS_VPORT_ATTR_UPCALL_STATS,
> __OVS_VPORT_ATTR_MAX
> };
>
> #define OVS_VPORT_ATTR_MAX (__OVS_VPORT_ATTR_MAX - 1)
>
> +/**
> + * enum ovs_vport_upcall_attr - attributes for %OVS_VPORT_UPCALL*
> +commands
> + * @OVS_VPORT_UPCALL_SUCCESS: 64-bit upcall success packets.
> + * @OVS_VPORT_UPCALL_FAIL: 64-bit upcall fail packets.
> + */
> +enum ovs_vport_upcall_attr {
> + OVS_VPORT_UPCALL_SUCCESS,
> + OVS_VPORT_UPCALL_FAIL,
> + __OVS_VPORT_UPCALL_MAX
> +};
> +
> +#define OVS_VPORT_UPCALL_MAX (__OVS_VPORT_UPCALL_MAX - 1)
> +
> enum {
> OVS_VXLAN_EXT_UNSPEC,
> OVS_VXLAN_EXT_GBP, /* Flag or __u32 */
> diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
> index c8a9075ddd0a..f9279aee2adb 100644
> --- a/net/openvswitch/datapath.c
> +++ b/net/openvswitch/datapath.c
> @@ -209,6 +209,26 @@ static struct vport *new_vport(const struct vport_parms *parms)
> return vport;
> }
>
> +static void ovs_vport_upcalls(struct sk_buff *skb,

This function name does not really represent what this function does.
It’s only taking care of statistics, so it should probably be called something like:

ovs_vport_update_upcall_stats() or ovs_vport_inc_upcall_stats()

> + const struct dp_upcall_info *upcall_info,
> + bool upcall_result)
> +{
> + struct vport *p = OVS_CB(skb)->input_vport;
> + struct vport_upcall_stats_percpu *vport_stats;

If you just call vport_stats, stats, the reverse Christmas tree order is achieved.

> +
> + if (upcall_info->cmd != OVS_PACKET_CMD_MISS &&
> + upcall_info->cmd != OVS_PACKET_CMD_ACTION)
> + return;
> +
> + vport_stats = this_cpu_ptr(p->upcall_stats);
> + u64_stats_update_begin(&vport_stats->syncp);
> + if (upcall_result)
> + u64_stats_inc(&vport_stats->n_success);
> + else
> + u64_stats_inc(&vport_stats->n_fail);
> + u64_stats_update_end(&vport_stats->syncp);
> +}
> +
> void ovs_dp_detach_port(struct vport *p) {
> ASSERT_OVSL();
> @@ -216,6 +236,9 @@ void ovs_dp_detach_port(struct vport *p)
> /* First drop references to device. */
> hlist_del_rcu(&p->dp_hash_node);
>
> + /* Free percpu memory */
> + free_percpu(p->upcall_stats);
> +
> /* Then destroy it. */
> ovs_vport_del(p);
> }
> @@ -305,6 +328,8 @@ int ovs_dp_upcall(struct datapath *dp, struct sk_buff *skb,
> err = queue_userspace_packet(dp, skb, key, upcall_info, cutlen);
> else
> err = queue_gso_packets(dp, skb, key, upcall_info, cutlen);
> +
> + ovs_vport_upcalls(skb, upcall_info, !err);
> if (err)
> goto err;
>
> @@ -1825,6 +1850,12 @@ static int ovs_dp_cmd_new(struct sk_buff *skb, struct genl_info *info)
> goto err_destroy_portids;
> }
>
> + vport->upcall_stats = netdev_alloc_pcpu_stats(struct vport_upcall_stats_percpu);
> + if (!vport->upcall_stats) {
> + err = -ENOMEM;
> + goto err_destroy_portids;
> + }
> +
> err = ovs_dp_cmd_fill_info(dp, reply, info->snd_portid,
> info->snd_seq, 0, OVS_DP_CMD_NEW);
> BUG_ON(err < 0);
> @@ -2068,6 +2099,8 @@ static int ovs_vport_cmd_fill_info(struct vport
> *vport, struct sk_buff *skb, {
> struct ovs_header *ovs_header;
> struct ovs_vport_stats vport_stats;
> + struct ovs_vport_upcall_stats stat;
> + struct nlattr *nla;
> int err;
>
> ovs_header = genlmsg_put(skb, portid, seq, &dp_vport_genl_family, @@
> -2097,6 +2130,15 @@ static int ovs_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb,
> OVS_VPORT_ATTR_PAD))
> goto nla_put_failure;
>
> + nla = nla_nest_start_noflag(skb, OVS_VPORT_ATTR_UPCALL_STATS);
> + if (!nla)
> + goto nla_put_failure;
> +
> + ovs_vport_get_upcall_stats(vport, &stat);
> + if (ovs_vport_put_upcall_stats(skb, &stat))
> + goto nla_put_failure;
> + nla_nest_end(skb, nla);
> +

See the comment below, as I think this all should be wrapped in ovs_vport_get_upcall_stats(vport, skb).

> if (ovs_vport_get_upcall_portids(vport, skb))
> goto nla_put_failure;
>
> @@ -2278,6 +2320,13 @@ static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info)
> goto exit_unlock_free;
> }
>
> + vport->upcall_stats = netdev_alloc_pcpu_stats(struct
> +vport_upcall_stats_percpu);
> +

nit: I think the extra new line is not needed.

> + if (!vport->upcall_stats) {
> + err = -ENOMEM;
> + goto exit_unlock_free;
> + }
> +
> err = ovs_vport_cmd_fill_info(vport, reply, genl_info_net(info),
> info->snd_portid, info->snd_seq, 0,
> OVS_VPORT_CMD_NEW, GFP_KERNEL); @@ -2507,6 +2556,7 @@
> static const struct nla_policy vport_policy[OVS_VPORT_ATTR_MAX + 1] = {
> [OVS_VPORT_ATTR_OPTIONS] = { .type = NLA_NESTED },
> [OVS_VPORT_ATTR_IFINDEX] = { .type = NLA_U32 },
> [OVS_VPORT_ATTR_NETNSID] = { .type = NLA_S32 },
> + [OVS_VPORT_ATTR_UPCALL_STATS] = { .type = NLA_NESTED },
> };
>
> static const struct genl_small_ops dp_vport_genl_ops[] = { diff --git
> a/net/openvswitch/vport.c b/net/openvswitch/vport.c index
> 82a74f998966..fd95536b35ef 100644
> --- a/net/openvswitch/vport.c
> +++ b/net/openvswitch/vport.c
> @@ -284,6 +284,50 @@ void ovs_vport_get_stats(struct vport *vport, struct ovs_vport_stats *stats)
> stats->tx_packets = dev_stats->tx_packets; }
>
> +/**
> + * ovs_vport_get_upcall_stats - retrieve upcall stats
> + *
> + * @vport: vport from which to retrieve the stats
> + * @ovs_vport_upcall_stats: location to store stats
> + *
> + * Retrieves upcall stats for the given device.
> + *
> + * Must be called with ovs_mutex or rcu_read_lock.
> + */
> +void ovs_vport_get_upcall_stats(struct vport *vport, struct
> +ovs_vport_upcall_stats *stats) {
> + int i;
> +
> + stats->tx_success = 0;
> + stats->tx_fail = 0;
> +
> + for_each_possible_cpu(i) {
> + const struct vport_upcall_stats_percpu *upcall_stats;
> + unsigned int start;
> +
> + upcall_stats = per_cpu_ptr(vport->upcall_stats, i);
> + do {
> + start = u64_stats_fetch_begin(&upcall_stats->syncp);
> + stats->tx_success += u64_stats_read(&upcall_stats->n_success);
> + stats->tx_fail += u64_stats_read(&upcall_stats->n_fail);
> + } while (u64_stats_fetch_retry(&upcall_stats->syncp, start));
> + }
> +}
> +
> +int ovs_vport_put_upcall_stats(struct sk_buff *skb,
> + struct ovs_vport_upcall_stats *stats) {
> + if (nla_put_u64_64bit(skb, OVS_VPORT_UPCALL_SUCCESS, stats->tx_success,
> + OVS_VPORT_ATTR_PAD))
> + return -EMSGSIZE;
> +
> + if (nla_put_u64_64bit(skb, OVS_VPORT_UPCALL_FAIL, stats->tx_fail,
> + OVS_VPORT_ATTR_PAD))
> + return -EMSGSIZE;
> +
> + return 0;
> +}

I think we should wrap ovs_vport_put_upcall_stats() into ovs_vport_get_upcall_stats(), so we have a single function. This would be similar to ovs_vport_get_options(). This way we will also get rid of the extra “struct ovs_vport_upcall_stats” definition, i.e.,

ovs_vport_get_upcall_stats(struct vport *vport, struct sk_buff *skb)

> +
> /**
> * ovs_vport_get_options - retrieve device options
> *
> diff --git a/net/openvswitch/vport.h b/net/openvswitch/vport.h index
> 7d276f60c000..5ba9f14df55a 100644
> --- a/net/openvswitch/vport.h
> +++ b/net/openvswitch/vport.h
> @@ -32,6 +32,16 @@ struct vport *ovs_vport_locate(const struct net
> *net, const char *name);
>
> void ovs_vport_get_stats(struct vport *, struct ovs_vport_stats *);
>
> +struct ovs_vport_upcall_stats {
> + __u64 tx_success; /* total packets upcalls succeed */
> + __u64 tx_fail; /* total packets upcalls failed */
> +};
> +
> +void ovs_vport_get_upcall_stats(struct vport *vport,
> + struct ovs_vport_upcall_stats *stats); int
> +ovs_vport_put_upcall_stats(struct sk_buff *skb,
> + struct ovs_vport_upcall_stats *stats);
> +
> int ovs_vport_set_options(struct vport *, struct nlattr *options);
> int ovs_vport_get_options(const struct vport *, struct sk_buff *);
>
> @@ -65,6 +75,7 @@ struct vport_portids {
> * @hash_node: Element in @dev_table hash table in vport.c.
> * @dp_hash_node: Element in @datapath->ports hash table in datapath.c.
> * @ops: Class structure.
> + * @upcall_stats: Upcall stats of every ports.
> * @detach_list: list used for detaching vport in net-exit call.
> * @rcu: RCU callback head for deferred destruction.
> */
> @@ -78,6 +89,7 @@ struct vport {
> struct hlist_node hash_node;
> struct hlist_node dp_hash_node;
> const struct vport_ops *ops;
> + struct vport_upcall_stats_percpu __percpu *upcall_stats;
>
> struct list_head detach_list;
> struct rcu_head rcu;
> @@ -137,6 +149,18 @@ struct vport_ops {
> struct list_head list;
> };
>
> +/**
> + * struct vport_upcall_stats_percpu - per-cpu packet upcall
> +statistics for
> + * a given vport.
> + * @n_success: Number of packets that upcall to userspace succeed.
> + * @n_fail: Number of packets that upcall to userspace failed.
> + */
> +struct vport_upcall_stats_percpu {
> + struct u64_stats_sync syncp;
> + u64_stats_t n_success;
> + u64_stats_t n_fail;
> +};
> +
> struct vport *ovs_vport_alloc(int priv_size, const struct vport_ops *,
> const struct vport_parms *); void ovs_vport_free(struct
> vport *);
> --
> 2.27.0