* [PATCH net-next 0/2] mpls: Packet stats @ 2017-01-13 18:14 Robert Shearman 2017-01-13 18:14 ` [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman ` (3 more replies) 0 siblings, 4 replies; 13+ messages in thread From: Robert Shearman @ 2017-01-13 18:14 UTC (permalink / raw) To: davem; +Cc: netdev, roopa, David Ahern, ebiederm, Robert Shearman This patchset records per-interface packet stats in the MPLS forwarding path and exports them using a nest of attributes root at a new IFLA_STATS_AF_SPEC attribute as part of RTM_GETSTATS messages: [IFLA_STATS_AF_SPEC] -> [AF_MPLS] -> [MPLS_STATS_LINK] -> struct mpls_link_stats The first patch adds the rtnl infrastructure for this, including a new callbacks to per-AF ops of fill_stats_af and get_stats_af_size. The second patch records MPLS stats and makes use of the infrastructure to export them. The rtnl infrastructure could also be used to export IPv6 stats in the future. Robert Shearman (2): net: AF-specific RTM_GETSTATS attributes mpls: Packet stats include/net/rtnetlink.h | 4 + include/uapi/linux/if_link.h | 1 + include/uapi/linux/mpls.h | 30 ++++++++ net/core/rtnetlink.c | 50 ++++++++++++ net/mpls/af_mpls.c | 179 +++++++++++++++++++++++++++++++++++++------ net/mpls/internal.h | 58 +++++++++++++- net/mpls/mpls_iptunnel.c | 11 ++- 7 files changed, 305 insertions(+), 28 deletions(-) -- 2.1.4 ^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes 2017-01-13 18:14 [PATCH net-next 0/2] mpls: Packet stats Robert Shearman @ 2017-01-13 18:14 ` Robert Shearman 2017-01-13 22:50 ` Roopa Prabhu 2017-01-13 18:14 ` [PATCH net-next 2/2] mpls: Packet stats Robert Shearman ` (2 subsequent siblings) 3 siblings, 1 reply; 13+ messages in thread From: Robert Shearman @ 2017-01-13 18:14 UTC (permalink / raw) To: davem; +Cc: netdev, roopa, David Ahern, ebiederm, Robert Shearman Add the functionality for including address-family-specific per-link stats in RTM_GETSTATS messages. This is done through adding a new IFLA_STATS_AF_SPEC attribute under which address family attributes are nested and then the AF-specific attributes can be further nested. This follows the model of IFLA_AF_SPEC on RTM_*LINK messages and it has the advantage of presenting an easily extended hierarchy. The rtnl_af_ops structure is extended to provide AFs with the opportunity to fill and provide the size of their stats attributes. One alternative would have been to provide AFs with the ability to add attributes directly into the RTM_GETSTATS message without a nested hierarchy. I discounted this approach as it increases the rate at which the 32 attribute number space is used up and it makes implementation a little more tricky for stats dump resuming (at the moment the order in which attributes are added to the message has to match the numeric order of the attributes). Another alternative would have been to register per-AF RTM_GETSTATS handlers. I discounted this approach as I perceived a common use-case to be getting all the stats for an interface and this approach would necessitate multiple requests/dumps to retrieve them all. Signed-off-by: Robert Shearman <rshearma@brocade.com> --- include/net/rtnetlink.h | 4 ++++ include/uapi/linux/if_link.h | 1 + net/core/rtnetlink.c | 50 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 55 insertions(+) diff --git a/include/net/rtnetlink.h b/include/net/rtnetlink.h index 4113916cc1bb..106de5f7bf06 100644 --- a/include/net/rtnetlink.h +++ b/include/net/rtnetlink.h @@ -139,6 +139,10 @@ struct rtnl_af_ops { const struct nlattr *attr); int (*set_link_af)(struct net_device *dev, const struct nlattr *attr); + + int (*fill_stats_af)(struct sk_buff *skb, + const struct net_device *dev); + size_t (*get_stats_af_size)(const struct net_device *dev); }; void __rtnl_af_unregister(struct rtnl_af_ops *ops); diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h index 6b13e591abc9..184b16ed2b84 100644 --- a/include/uapi/linux/if_link.h +++ b/include/uapi/linux/if_link.h @@ -847,6 +847,7 @@ enum { IFLA_STATS_LINK_XSTATS, IFLA_STATS_LINK_XSTATS_SLAVE, IFLA_STATS_LINK_OFFLOAD_XSTATS, + IFLA_STATS_AF_SPEC, __IFLA_STATS_MAX, }; diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 18b5aae99bec..4edc1bd7a735 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -3829,6 +3829,39 @@ static int rtnl_fill_statsinfo(struct sk_buff *skb, struct net_device *dev, *idxattr = 0; } + if (stats_attr_valid(filter_mask, IFLA_STATS_AF_SPEC, *idxattr)) { + struct rtnl_af_ops *af_ops; + + *idxattr = IFLA_STATS_AF_SPEC; + attr = nla_nest_start(skb, IFLA_STATS_AF_SPEC); + if (!attr) + goto nla_put_failure; + + list_for_each_entry(af_ops, &rtnl_af_ops, list) { + if (af_ops->fill_stats_af) { + struct nlattr *af; + int err; + + af = nla_nest_start(skb, af_ops->family); + if (!af) + goto nla_put_failure; + + err = af_ops->fill_stats_af(skb, dev); + + if (err == -ENODATA) + nla_nest_cancel(skb, af); + else if (err < 0) + goto nla_put_failure; + + nla_nest_end(skb, af); + } + } + + nla_nest_end(skb, attr); + + *idxattr = 0; + } + nlmsg_end(skb, nlh); return 0; @@ -3885,6 +3918,23 @@ static size_t if_nlmsg_stats_size(const struct net_device *dev, if (stats_attr_valid(filter_mask, IFLA_STATS_LINK_OFFLOAD_XSTATS, 0)) size += rtnl_get_offload_stats_size(dev); + if (stats_attr_valid(filter_mask, IFLA_STATS_AF_SPEC, 0)) { + struct rtnl_af_ops *af_ops; + + /* for IFLA_STATS_AF_SPEC */ + size += nla_total_size(0); + + list_for_each_entry(af_ops, &rtnl_af_ops, list) { + if (af_ops->get_stats_af_size) { + size += nla_total_size( + af_ops->get_stats_af_size(dev)); + + /* for AF_* */ + size += nla_total_size(0); + } + } + } + return size; } -- 2.1.4 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes 2017-01-13 18:14 ` [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman @ 2017-01-13 22:50 ` Roopa Prabhu 0 siblings, 0 replies; 13+ messages in thread From: Roopa Prabhu @ 2017-01-13 22:50 UTC (permalink / raw) To: Robert Shearman; +Cc: davem, netdev, David Ahern, ebiederm On 1/13/17, 10:14 AM, Robert Shearman wrote: > Add the functionality for including address-family-specific per-link > stats in RTM_GETSTATS messages. This is done through adding a new > IFLA_STATS_AF_SPEC attribute under which address family attributes are > nested and then the AF-specific attributes can be further nested. This > follows the model of IFLA_AF_SPEC on RTM_*LINK messages and it has the > advantage of presenting an easily extended hierarchy. The rtnl_af_ops > structure is extended to provide AFs with the opportunity to fill and > provide the size of their stats attributes. > > One alternative would have been to provide AFs with the ability to add > attributes directly into the RTM_GETSTATS message without a nested > hierarchy. I discounted this approach as it increases the rate at > which the 32 attribute number space is used up and it makes > implementation a little more tricky for stats dump resuming (at the > moment the order in which attributes are added to the message has to > match the numeric order of the attributes). > > Another alternative would have been to register per-AF RTM_GETSTATS > handlers. I discounted this approach as I perceived a common use-case > to be getting all the stats for an interface and this approach would > necessitate multiple requests/dumps to retrieve them all. > > Signed-off-by: Robert Shearman <rshearma@brocade.com> > Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com> ^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH net-next 2/2] mpls: Packet stats 2017-01-13 18:14 [PATCH net-next 0/2] mpls: Packet stats Robert Shearman 2017-01-13 18:14 ` [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman @ 2017-01-13 18:14 ` Robert Shearman 2017-01-13 22:51 ` Roopa Prabhu ` (2 more replies) 2017-01-13 22:50 ` [PATCH net-next 0/2] " Roopa Prabhu 2017-01-16 14:16 ` [PATCH net-next v2 " Robert Shearman 3 siblings, 3 replies; 13+ messages in thread From: Robert Shearman @ 2017-01-13 18:14 UTC (permalink / raw) To: davem; +Cc: netdev, roopa, David Ahern, ebiederm, Robert Shearman Having MPLS packet stats is useful for observing network operation and for diagnosing network problems. In the absence of anything better, RFC2863 and RFC3813 are used for guidance for which stats to expose and the semantics of them. In particular rx_noroutes maps to in unknown protos in RFC2863. The stats are exposed to userspace via AF_MPLS attributes embedded in the IFLA_STATS_AF_SPEC attribute of RTM_GETSTATS messages. All the introduced fields are 64-bit, even error ones, to ensure no overflow with long uptimes. Per-CPU counters are used to avoid cache-line contention on the commonly used fields. The other fields have also been made per-CPU for code to avoid performance problems in error conditions on the assumption that on some platforms the cost of atomic operations could be more expensive than sending the packet (which is what would be done in the success case). If that's not the case, we could instead not use per-CPU counters for these fields. Only unicast and non-fragment are exposed at the moment, but other counters can be exposed in the future either by adding to the end of struct mpls_link_stats or by additional netlink attributes in the AF_MPLS IFLA_STATS_AF_SPEC nested attribute. Signed-off-by: Robert Shearman <rshearma@brocade.com> --- include/uapi/linux/mpls.h | 30 ++++++++ net/mpls/af_mpls.c | 179 ++++++++++++++++++++++++++++++++++++++++------ net/mpls/internal.h | 58 ++++++++++++++- net/mpls/mpls_iptunnel.c | 11 ++- 4 files changed, 250 insertions(+), 28 deletions(-) diff --git a/include/uapi/linux/mpls.h b/include/uapi/linux/mpls.h index 24a6cb1aec86..77a19dfe3990 100644 --- a/include/uapi/linux/mpls.h +++ b/include/uapi/linux/mpls.h @@ -43,4 +43,34 @@ struct mpls_label { #define MPLS_LABEL_FIRST_UNRESERVED 16 /* RFC3032 */ +/* These are embedded into IFLA_STATS_AF_SPEC: + * [IFLA_STATS_AF_SPEC] + * -> [AF_MPLS] + * -> [MPLS_STATS_xxx] + * + * Attributes: + * [MPLS_STATS_LINK] = { + * struct mpls_link_stats + * } + */ +enum { + MPLS_STATS_UNSPEC, /* also used as 64bit pad attribute */ + MPLS_STATS_LINK, + __MPLS_STATS_MAX, +}; + +#define MPLS_STATS_MAX (__MPLS_STATS_MAX - 1) + +struct mpls_link_stats { + __u64 rx_packets; /* total packets received */ + __u64 tx_packets; /* total packets transmitted */ + __u64 rx_bytes; /* total bytes received */ + __u64 tx_bytes; /* total bytes transmitted */ + __u64 rx_errors; /* bad packets received */ + __u64 tx_errors; /* packet transmit problems */ + __u64 rx_dropped; /* packet dropped on receive */ + __u64 tx_dropped; /* packet dropped on transmit */ + __u64 rx_noroute; /* no route for packet dest */ +}; + #endif /* _UAPI_MPLS_H */ diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 15fe97644ffe..fb20941cdda2 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -8,6 +8,7 @@ #include <linux/ipv6.h> #include <linux/mpls.h> #include <linux/vmalloc.h> +#include <linux/percpu.h> #include <net/ip.h> #include <net/dst.h> #include <net/sock.h> @@ -17,8 +18,8 @@ #include <net/netns/generic.h> #if IS_ENABLED(CONFIG_IPV6) #include <net/ipv6.h> -#include <net/addrconf.h> #endif +#include <net/addrconf.h> #include <net/nexthop.h> #include "internal.h" @@ -48,11 +49,6 @@ static struct mpls_route *mpls_route_input_rcu(struct net *net, unsigned index) return rt; } -static inline struct mpls_dev *mpls_dev_get(const struct net_device *dev) -{ - return rcu_dereference_rtnl(dev->mpls_ptr); -} - bool mpls_output_possible(const struct net_device *dev) { return dev && (dev->flags & IFF_UP) && netif_carrier_ok(dev); @@ -98,6 +94,29 @@ bool mpls_pkt_too_big(const struct sk_buff *skb, unsigned int mtu) } EXPORT_SYMBOL_GPL(mpls_pkt_too_big); +void mpls_stats_inc_outucastpkts(struct net_device *dev, + const struct sk_buff *skb) +{ + struct mpls_dev *mdev; + struct inet6_dev *in6dev; + + if (skb->protocol == htons(ETH_P_MPLS_UC)) { + mdev = mpls_dev_get(dev); + if (mdev) + MPLS_INC_STATS_LEN(mdev, skb->len, + tx_packets, + tx_bytes); + } else if (skb->protocol == htons(ETH_P_IP)) { + IP_UPD_PO_STATS(dev_net(dev), IPSTATS_MIB_OUT, skb->len); + } else if (skb->protocol == htons(ETH_P_IPV6)) { + in6dev = __in6_dev_get(dev); + if (in6dev) + IP6_UPD_PO_STATS(dev_net(dev), in6dev, + IPSTATS_MIB_OUT, skb->len); + } +} +EXPORT_SYMBOL_GPL(mpls_stats_inc_outucastpkts); + static u32 mpls_multipath_hash(struct mpls_route *rt, struct sk_buff *skb, bool bos) { @@ -253,6 +272,7 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, struct mpls_nh *nh; struct mpls_entry_decoded dec; struct net_device *out_dev; + struct mpls_dev *out_mdev; struct mpls_dev *mdev; unsigned int hh_len; unsigned int new_header_size; @@ -262,17 +282,25 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, /* Careful this entire function runs inside of an rcu critical section */ mdev = mpls_dev_get(dev); - if (!mdev || !mdev->input_enabled) + if (!mdev) goto drop; - if (skb->pkt_type != PACKET_HOST) + MPLS_INC_STATS_LEN(mdev, skb->len, rx_packets, + rx_bytes); + + if (!mdev->input_enabled) { + MPLS_INC_STATS(mdev, rx_dropped); goto drop; + } + + if (skb->pkt_type != PACKET_HOST) + goto err; if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) - goto drop; + goto err; if (!pskb_may_pull(skb, sizeof(*hdr))) - goto drop; + goto err; /* Read and decode the label */ hdr = mpls_hdr(skb); @@ -285,33 +313,35 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, skb_orphan(skb); rt = mpls_route_input_rcu(net, dec.label); - if (!rt) + if (!rt) { + MPLS_INC_STATS(mdev, rx_noroute); goto drop; + } nh = mpls_select_multipath(rt, skb, dec.bos); if (!nh) - goto drop; - - /* Find the output device */ - out_dev = rcu_dereference(nh->nh_dev); - if (!mpls_output_possible(out_dev)) - goto drop; + goto err; if (skb_warn_if_lro(skb)) - goto drop; + goto err; skb_forward_csum(skb); /* Verify ttl is valid */ if (dec.ttl <= 1) - goto drop; + goto err; dec.ttl -= 1; + /* Find the output device */ + out_dev = rcu_dereference(nh->nh_dev); + if (!mpls_output_possible(out_dev)) + goto tx_err; + /* Verify the destination can hold the packet */ new_header_size = mpls_nh_header_size(nh); mtu = mpls_dev_mtu(out_dev); if (mpls_pkt_too_big(skb, mtu - new_header_size)) - goto drop; + goto tx_err; hh_len = LL_RESERVED_SPACE(out_dev); if (!out_dev->header_ops) @@ -319,7 +349,7 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, /* Ensure there is enough space for the headers in the skb */ if (skb_cow(skb, hh_len + new_header_size)) - goto drop; + goto tx_err; skb->dev = out_dev; skb->protocol = htons(ETH_P_MPLS_UC); @@ -327,7 +357,7 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, if (unlikely(!new_header_size && dec.bos)) { /* Penultimate hop popping */ if (!mpls_egress(rt, skb, dec)) - goto drop; + goto err; } else { bool bos; int i; @@ -343,6 +373,8 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, } } + mpls_stats_inc_outucastpkts(out_dev, skb); + /* If via wasn't specified then send out using device address */ if (nh->nh_via_table == MPLS_NEIGH_TABLE_UNSPEC) err = neigh_xmit(NEIGH_LINK_TABLE, out_dev, @@ -355,6 +387,13 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, __func__, err); return 0; +tx_err: + out_mdev = out_dev ? mpls_dev_get(out_dev) : NULL; + if (out_mdev) + MPLS_INC_STATS(out_mdev, tx_errors); + goto drop; +err: + MPLS_INC_STATS(mdev, rx_errors); drop: kfree_skb(skb); return NET_RX_DROP; @@ -853,6 +892,70 @@ static int mpls_route_del(struct mpls_route_config *cfg) return err; } +static void mpls_get_stats(struct mpls_dev *mdev, + struct mpls_link_stats *stats) +{ + struct mpls_pcpu_stats *p; + int i; + + memset(stats, 0, sizeof(*stats)); + + for_each_possible_cpu(i) { + struct mpls_link_stats local; + unsigned int start; + + p = per_cpu_ptr(mdev->stats, i); + do { + start = u64_stats_fetch_begin(&p->syncp); + local = p->stats; + } while (u64_stats_fetch_retry(&p->syncp, start)); + + stats->rx_packets += local.rx_packets; + stats->rx_bytes += local.rx_bytes; + stats->tx_packets += local.tx_packets; + stats->tx_bytes += local.tx_bytes; + stats->rx_errors += local.rx_errors; + stats->tx_errors += local.tx_errors; + stats->rx_dropped += local.rx_dropped; + stats->tx_dropped += local.tx_dropped; + stats->rx_noroute += local.rx_noroute; + } +} + +static int mpls_fill_stats_af(struct sk_buff *skb, + const struct net_device *dev) +{ + struct mpls_link_stats *stats; + struct mpls_dev *mdev; + struct nlattr *nla; + + mdev = mpls_dev_get(dev); + if (!mdev) + return -ENODATA; + + nla = nla_reserve_64bit(skb, MPLS_STATS_LINK, + sizeof(struct mpls_link_stats), + MPLS_STATS_UNSPEC); + if (!nla) + return -EMSGSIZE; + + stats = nla_data(nla); + mpls_get_stats(mdev, stats); + + return 0; +} + +static size_t mpls_get_stats_af_size(const struct net_device *dev) +{ + struct mpls_dev *mdev; + + mdev = mpls_dev_get(dev); + if (!mdev) + return 0; + + return nla_total_size_64bit(sizeof(struct mpls_link_stats)); +} + #define MPLS_PERDEV_SYSCTL_OFFSET(field) \ (&((struct mpls_dev *)0)->field) @@ -911,6 +1014,7 @@ static struct mpls_dev *mpls_add_dev(struct net_device *dev) { struct mpls_dev *mdev; int err = -ENOMEM; + int i; ASSERT_RTNL(); @@ -918,6 +1022,17 @@ static struct mpls_dev *mpls_add_dev(struct net_device *dev) if (!mdev) return ERR_PTR(err); + mdev->stats = alloc_percpu(struct mpls_pcpu_stats); + if (!mdev->stats) + goto free; + + for_each_possible_cpu(i) { + struct mpls_pcpu_stats *mpls_stats; + + mpls_stats = per_cpu_ptr(mdev->stats, i); + u64_stats_init(&mpls_stats->syncp); + } + err = mpls_dev_sysctl_register(dev, mdev); if (err) goto free; @@ -927,10 +1042,19 @@ static struct mpls_dev *mpls_add_dev(struct net_device *dev) return mdev; free: + free_percpu(mdev->stats); kfree(mdev); return ERR_PTR(err); } +static void mpls_dev_destroy_rcu(struct rcu_head *head) +{ + struct mpls_dev *mdev = container_of(head, struct mpls_dev, rcu); + + free_percpu(mdev->stats); + kfree(mdev); +} + static void mpls_ifdown(struct net_device *dev, int event) { struct mpls_route __rcu **platform_label; @@ -1045,7 +1169,7 @@ static int mpls_dev_notify(struct notifier_block *this, unsigned long event, if (mdev) { mpls_dev_sysctl_unregister(mdev); RCU_INIT_POINTER(dev->mpls_ptr, NULL); - kfree_rcu(mdev, rcu); + call_rcu(&mdev->rcu, mpls_dev_destroy_rcu); } break; case NETDEV_CHANGENAME: @@ -1706,6 +1830,12 @@ static struct pernet_operations mpls_net_ops = { .exit = mpls_net_exit, }; +static struct rtnl_af_ops mpls_af_ops __read_mostly = { + .family = AF_MPLS, + .fill_stats_af = mpls_fill_stats_af, + .get_stats_af_size = mpls_get_stats_af_size, +}; + static int __init mpls_init(void) { int err; @@ -1722,6 +1852,8 @@ static int __init mpls_init(void) dev_add_pack(&mpls_packet_type); + rtnl_af_register(&mpls_af_ops); + rtnl_register(PF_MPLS, RTM_NEWROUTE, mpls_rtm_newroute, NULL, NULL); rtnl_register(PF_MPLS, RTM_DELROUTE, mpls_rtm_delroute, NULL, NULL); rtnl_register(PF_MPLS, RTM_GETROUTE, NULL, mpls_dump_routes, NULL); @@ -1738,6 +1870,7 @@ module_init(mpls_init); static void __exit mpls_exit(void) { rtnl_unregister_all(PF_MPLS); + rtnl_af_unregister(&mpls_af_ops); dev_remove_pack(&mpls_packet_type); unregister_netdevice_notifier(&mpls_dev_notifier); unregister_pernet_subsys(&mpls_net_ops); diff --git a/net/mpls/internal.h b/net/mpls/internal.h index bdfef6c3271a..d97243034605 100644 --- a/net/mpls/internal.h +++ b/net/mpls/internal.h @@ -9,13 +9,58 @@ struct mpls_entry_decoded { u8 bos; }; +struct mpls_pcpu_stats { + struct mpls_link_stats stats; + struct u64_stats_sync syncp; +}; + struct mpls_dev { - int input_enabled; + int input_enabled; - struct ctl_table_header *sysctl; - struct rcu_head rcu; + struct mpls_pcpu_stats __percpu *stats; + + struct ctl_table_header *sysctl; + struct rcu_head rcu; }; +#if BITS_PER_LONG == 32 + +#define MPLS_INC_STATS_LEN(mdev, len, pkts_field, bytes_field) \ + do { \ + __typeof__(*(mdev)->stats) *ptr = \ + raw_cpu_ptr((mdev)->stats); \ + local_bh_disable(); \ + u64_stats_update_begin(&ptr->syncp); \ + ptr->stats.pkts_field++; \ + ptr->stats.bytes_field += (len); \ + u64_stats_update_end(&ptr->syncp); \ + local_bh_enable(); \ + } while (0) + +#define MPLS_INC_STATS(mdev, field) \ + do { \ + __typeof__(*(mdev)->stats) *ptr = \ + raw_cpu_ptr((mdev)->stats); \ + local_bh_disable(); \ + u64_stats_update_begin(&ptr->syncp); \ + ptr->stats.field++; \ + u64_stats_update_end(&ptr->syncp); \ + local_bh_enable(); \ + } while (0) + +#else + +#define MPLS_INC_STATS_LEN(mdev, len, pkts_field, bytes_field) \ + do { \ + this_cpu_inc((mdev)->stats->stats.pkts_field); \ + this_cpu_add((mdev)->stats->stats.bytes_field, (len)); \ + } while (0) + +#define MPLS_INC_STATS(mdev, field) \ + this_cpu_inc((mdev)->stats->stats.field) + +#endif + struct sk_buff; #define LABEL_NOT_SPECIFIED (1 << 20) @@ -114,6 +159,11 @@ static inline struct mpls_entry_decoded mpls_entry_decode(struct mpls_shim_hdr * return result; } +static inline struct mpls_dev *mpls_dev_get(const struct net_device *dev) +{ + return rcu_dereference_rtnl(dev->mpls_ptr); +} + int nla_put_labels(struct sk_buff *skb, int attrtype, u8 labels, const u32 label[]); int nla_get_labels(const struct nlattr *nla, u32 max_labels, u8 *labels, @@ -123,5 +173,7 @@ int nla_get_via(const struct nlattr *nla, u8 *via_alen, u8 *via_table, bool mpls_output_possible(const struct net_device *dev); unsigned int mpls_dev_mtu(const struct net_device *dev); bool mpls_pkt_too_big(const struct sk_buff *skb, unsigned int mtu); +void mpls_stats_inc_outucastpkts(struct net_device *dev, + const struct sk_buff *skb); #endif /* MPLS_INTERNAL_H */ diff --git a/net/mpls/mpls_iptunnel.c b/net/mpls/mpls_iptunnel.c index 2f7ccd934416..02531284bc49 100644 --- a/net/mpls/mpls_iptunnel.c +++ b/net/mpls/mpls_iptunnel.c @@ -48,11 +48,15 @@ static int mpls_xmit(struct sk_buff *skb) struct dst_entry *dst = skb_dst(skb); struct rtable *rt = NULL; struct rt6_info *rt6 = NULL; + struct mpls_dev *out_mdev; int err = 0; bool bos; int i; unsigned int ttl; + /* Find the output device */ + out_dev = dst->dev; + /* Obtain the ttl */ if (dst->ops->family == AF_INET) { ttl = ip_hdr(skb)->ttl; @@ -66,8 +70,6 @@ static int mpls_xmit(struct sk_buff *skb) skb_orphan(skb); - /* Find the output device */ - out_dev = dst->dev; if (!mpls_output_possible(out_dev) || !dst->lwtstate || skb_warn_if_lro(skb)) goto drop; @@ -109,6 +111,8 @@ static int mpls_xmit(struct sk_buff *skb) bos = false; } + mpls_stats_inc_outucastpkts(out_dev, skb); + if (rt) err = neigh_xmit(NEIGH_ARP_TABLE, out_dev, &rt->rt_gateway, skb); @@ -122,6 +126,9 @@ static int mpls_xmit(struct sk_buff *skb) return LWTUNNEL_XMIT_DONE; drop: + out_mdev = out_dev ? mpls_dev_get(out_dev) : NULL; + if (out_mdev) + MPLS_INC_STATS(out_mdev, tx_errors); kfree_skb(skb); return -EINVAL; } -- 2.1.4 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH net-next 2/2] mpls: Packet stats 2017-01-13 18:14 ` [PATCH net-next 2/2] mpls: Packet stats Robert Shearman @ 2017-01-13 22:51 ` Roopa Prabhu 2017-01-14 6:41 ` kbuild test robot 2017-01-14 6:58 ` kbuild test robot 2 siblings, 0 replies; 13+ messages in thread From: Roopa Prabhu @ 2017-01-13 22:51 UTC (permalink / raw) To: Robert Shearman; +Cc: davem, netdev, David Ahern, ebiederm On 1/13/17, 10:14 AM, Robert Shearman wrote: > Having MPLS packet stats is useful for observing network operation and > for diagnosing network problems. In the absence of anything better, > RFC2863 and RFC3813 are used for guidance for which stats to expose > and the semantics of them. In particular rx_noroutes maps to in > unknown protos in RFC2863. The stats are exposed to userspace via > AF_MPLS attributes embedded in the IFLA_STATS_AF_SPEC attribute of > RTM_GETSTATS messages. > > All the introduced fields are 64-bit, even error ones, to ensure no > overflow with long uptimes. Per-CPU counters are used to avoid > cache-line contention on the commonly used fields. The other fields > have also been made per-CPU for code to avoid performance problems in > error conditions on the assumption that on some platforms the cost of > atomic operations could be more expensive than sending the packet > (which is what would be done in the success case). If that's not the > case, we could instead not use per-CPU counters for these fields. > > Only unicast and non-fragment are exposed at the moment, but other > counters can be exposed in the future either by adding to the end of > struct mpls_link_stats or by additional netlink attributes in the > AF_MPLS IFLA_STATS_AF_SPEC nested attribute. > > Signed-off-by: Robert Shearman <rshearma@brocade.com> > --- > Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com> ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH net-next 2/2] mpls: Packet stats 2017-01-13 18:14 ` [PATCH net-next 2/2] mpls: Packet stats Robert Shearman 2017-01-13 22:51 ` Roopa Prabhu @ 2017-01-14 6:41 ` kbuild test robot 2017-01-16 10:30 ` Robert Shearman 2017-01-14 6:58 ` kbuild test robot 2 siblings, 1 reply; 13+ messages in thread From: kbuild test robot @ 2017-01-14 6:41 UTC (permalink / raw) To: Robert Shearman Cc: kbuild-all, davem, netdev, roopa, David Ahern, ebiederm, Robert Shearman [-- Attachment #1: Type: text/plain, Size: 10427 bytes --] Hi Robert, [auto build test ERROR on net-next/master] url: https://github.com/0day-ci/linux/commits/Robert-Shearman/net-AF-specific-RTM_GETSTATS-attributes/20170114-095819 config: x86_64-randconfig-u0-01141334 (attached as .config) compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901 reproduce: # save the attached .config to linux build tree make ARCH=x86_64 All error/warnings (new ones prefixed by >>): In file included from include/net/netns/mib.h:4:0, from include/net/net_namespace.h:14, from include/linux/netdevice.h:43, from include/uapi/linux/if_arp.h:26, from include/linux/if_arp.h:27, from net/mpls/af_mpls.c:7: net/mpls/af_mpls.c: In function 'mpls_stats_inc_outucastpkts': >> include/net/ipv6.h:163:35: error: 'struct netns_mib' has no member named 'ipv6_statistics'; did you mean 'ip_statistics'? mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/snmp.h:145:15: note: in definition of macro 'SNMP_UPD_PO_STATS' __typeof__((mib->mibs) + 0) ptr = mib->mibs; \ ^~~ >> include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^~~~~~~ >> net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^~~~~~~~~~~~~~~~ >> include/net/ipv6.h:163:35: error: 'struct netns_mib' has no member named 'ipv6_statistics'; did you mean 'ip_statistics'? mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/snmp.h:145:37: note: in definition of macro 'SNMP_UPD_PO_STATS' __typeof__((mib->mibs) + 0) ptr = mib->mibs; \ ^~~ >> include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^~~~~~~ >> net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^~~~~~~~~~~~~~~~ In file included from include/asm-generic/percpu.h:6:0, from arch/x86/include/asm/percpu.h:542, from arch/x86/include/asm/preempt.h:5, from include/linux/preempt.h:59, from include/linux/spinlock.h:50, from include/linux/mm_types.h:8, from include/linux/kmemcheck.h:4, from include/linux/skbuff.h:18, from net/mpls/af_mpls.c:2: >> include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/linux/percpu-defs.h:206:47: note: in definition of macro '__verify_pcpu_ptr' const void __percpu *__vpp_verify = (typeof((ptr) + 0))NULL; \ ^~~ include/linux/percpu-defs.h:496:33: note: in expansion of macro '__pcpu_size_call' #define this_cpu_add(pcp, val) __pcpu_size_call(this_cpu_add_, pcp, val) ^~~~~~~~~~~~~~~~ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^~~~~~~~~~~~ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^~~~~~~~~~~~ >> include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^~~~~~~~~~~~~~~~~ >> include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^~~~~~~ >> net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^~~~~~~~~~~~~~~~ >> include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/linux/percpu-defs.h:363:16: note: in definition of macro '__pcpu_size_call' switch(sizeof(variable)) { \ ^~~~~~~~ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^~~~~~~~~~~~ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^~~~~~~~~~~~ >> include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^~~~~~~~~~~~~~~~~ >> include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^~~~~~~ >> net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^~~~~~~~~~~~~~~~ In file included from arch/x86/include/asm/preempt.h:5:0, from include/linux/preempt.h:59, from include/linux/spinlock.h:50, from include/linux/mm_types.h:8, from include/linux/kmemcheck.h:4, from include/linux/skbuff.h:18, from net/mpls/af_mpls.c:2: >> include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ arch/x86/include/asm/percpu.h:128:17: note: in definition of macro 'percpu_add_op' typedef typeof(var) pao_T__; \ ^~~ include/linux/percpu-defs.h:364:11: note: in expansion of macro 'this_cpu_add_1' case 1: stem##1(variable, __VA_ARGS__);break; \ ^~~~ include/linux/percpu-defs.h:496:33: note: in expansion of macro '__pcpu_size_call' #define this_cpu_add(pcp, val) __pcpu_size_call(this_cpu_add_, pcp, val) ^~~~~~~~~~~~~~~~ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^~~~~~~~~~~~ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^~~~~~~~~~~~ >> include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^~~~~~~~~~~~~~~~~ >> include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^~~~~~~ >> net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^~~~~~~~~~~~~~~~ >> include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ arch/x86/include/asm/percpu.h:137:17: note: in definition of macro 'percpu_add_op' switch (sizeof(var)) { \ ^~~ include/linux/percpu-defs.h:364:11: note: in expansion of macro 'this_cpu_add_1' case 1: stem##1(variable, __VA_ARGS__);break; \ ^~~~ include/linux/percpu-defs.h:496:33: note: in expansion of macro '__pcpu_size_call' #define this_cpu_add(pcp, val) __pcpu_size_call(this_cpu_add_, pcp, val) ^~~~~~~~~~~~~~~~ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^~~~~~~~~~~~ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^~~~~~~~~~~~ >> include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^~~~~~~~~~~~~~~~~ vim +163 include/net/ipv6.h 8e7999c4 Pavel Emelyanov 2007-10-15 157 13415e46 Eric Dumazet 2016-04-27 158 #define _DEVUPD(net, statname, mod, idev, field, val) \ edf391ff Neil Horman 2009-04-27 159 ({ \ edf391ff Neil Horman 2009-04-27 160 struct inet6_dev *_idev = (idev); \ edf391ff Neil Horman 2009-04-27 161 if (likely(_idev != NULL)) \ 13415e46 Eric Dumazet 2016-04-27 162 mod##SNMP_UPD_PO_STATS((_idev)->stats.statname, field, (val)); \ 13415e46 Eric Dumazet 2016-04-27 @163 mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ edf391ff Neil Horman 2009-04-27 164 }) edf391ff Neil Horman 2009-04-27 165 14878f75 David L Stevens 2007-09-16 166 /* MIBs */ 14878f75 David L Stevens 2007-09-16 167 087fe240 Denis V. Lunev 2008-10-08 168 #define IP6_INC_STATS(net, idev,field) \ 13415e46 Eric Dumazet 2016-04-27 169 _DEVINC(net, ipv6, , idev, field) 1d015503 Eric Dumazet 2016-04-27 170 #define __IP6_INC_STATS(net, idev,field) \ 13415e46 Eric Dumazet 2016-04-27 171 _DEVINC(net, ipv6, __, idev, field) edf391ff Neil Horman 2009-04-27 172 #define IP6_ADD_STATS(net, idev,field,val) \ 13415e46 Eric Dumazet 2016-04-27 173 _DEVADD(net, ipv6, , idev, field, val) 1d015503 Eric Dumazet 2016-04-27 174 #define __IP6_ADD_STATS(net, idev,field,val) \ 13415e46 Eric Dumazet 2016-04-27 175 _DEVADD(net, ipv6, __, idev, field, val) edf391ff Neil Horman 2009-04-27 176 #define IP6_UPD_PO_STATS(net, idev,field,val) \ 13415e46 Eric Dumazet 2016-04-27 @177 _DEVUPD(net, ipv6, , idev, field, val) c2005eb0 Eric Dumazet 2016-04-27 178 #define __IP6_UPD_PO_STATS(net, idev,field,val) \ 13415e46 Eric Dumazet 2016-04-27 179 _DEVUPD(net, ipv6, __, idev, field, val) 087fe240 Denis V. Lunev 2008-10-08 180 #define ICMP6_INC_STATS(net, idev, field) \ :::::: The code at line 163 was first introduced by commit :::::: 13415e46c5915e2dac089de516369005fbc045f9 net: snmp: kill STATS_BH macros :::::: TO: Eric Dumazet <edumazet@google.com> :::::: CC: David S. Miller <davem@davemloft.net> --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation [-- Attachment #2: .config.gz --] [-- Type: application/gzip, Size: 28286 bytes --] ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH net-next 2/2] mpls: Packet stats 2017-01-14 6:41 ` kbuild test robot @ 2017-01-16 10:30 ` Robert Shearman 0 siblings, 0 replies; 13+ messages in thread From: Robert Shearman @ 2017-01-16 10:30 UTC (permalink / raw) To: kbuild test robot; +Cc: kbuild-all, davem, netdev, roopa, David Ahern, ebiederm On 14/01/17 06:41, kbuild test robot wrote: > Hi Robert, > > [auto build test ERROR on net-next/master] > > url: https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_0day-2Dci_linux_commits_Robert-2DShearman_net-2DAF-2Dspecific-2DRTM-5FGETSTATS-2Dattributes_20170114-2D095819&d=DwIBAg&c=IL_XqQWOjubgfqINi2jTzg&r=HbzsGgI7IidV3zRVGbZOYKAd0w1ch0IRBV2pqme4k9w&m=-FIZAjvyj_cLBHBTW14fNKp5IiJpkr1laQrxgxKnZ3g&s=_d4vequnI_QXJkUb0X3X8AUHdi6qD4R86g-mTBsEJ7g&e= > config: x86_64-randconfig-u0-01141334 (attached as .config) > compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901 > reproduce: > # save the attached .config to linux build tree > make ARCH=x86_64 > > All error/warnings (new ones prefixed by >>): > > In file included from include/net/netns/mib.h:4:0, > from include/net/net_namespace.h:14, > from include/linux/netdevice.h:43, > from include/uapi/linux/if_arp.h:26, > from include/linux/if_arp.h:27, > from net/mpls/af_mpls.c:7: > net/mpls/af_mpls.c: In function 'mpls_stats_inc_outucastpkts': >>> include/net/ipv6.h:163:35: error: 'struct netns_mib' has no member named 'ipv6_statistics'; did you mean 'ip_statistics'? > mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ Good catch kbuild test robot. Expect to see a v2 shortly with this CONFIG_IPV6=n build failure fixed. Thanks, Rob ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH net-next 2/2] mpls: Packet stats 2017-01-13 18:14 ` [PATCH net-next 2/2] mpls: Packet stats Robert Shearman 2017-01-13 22:51 ` Roopa Prabhu 2017-01-14 6:41 ` kbuild test robot @ 2017-01-14 6:58 ` kbuild test robot 2 siblings, 0 replies; 13+ messages in thread From: kbuild test robot @ 2017-01-14 6:58 UTC (permalink / raw) To: Robert Shearman Cc: kbuild-all, davem, netdev, roopa, David Ahern, ebiederm, Robert Shearman [-- Attachment #1: Type: text/plain, Size: 7847 bytes --] Hi Robert, [auto build test ERROR on net-next/master] url: https://github.com/0day-ci/linux/commits/Robert-Shearman/net-AF-specific-RTM_GETSTATS-attributes/20170114-095819 config: i386-randconfig-sb0-01141243 (attached as .config) compiler: gcc-5 (Debian 5.4.1-2) 5.4.1 20160904 reproduce: # save the attached .config to linux build tree make ARCH=i386 All errors (new ones prefixed by >>): In file included from include/net/netns/mib.h:4:0, from include/net/net_namespace.h:14, from include/linux/netdevice.h:43, from include/uapi/linux/if_arp.h:26, from include/linux/if_arp.h:27, from net/mpls/af_mpls.c:7: net/mpls/af_mpls.c: In function 'mpls_stats_inc_outucastpkts': >> include/net/ipv6.h:163:35: error: 'struct netns_mib' has no member named 'ipv6_statistics' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/snmp.h:145:15: note: in definition of macro 'SNMP_UPD_PO_STATS' __typeof__((mib->mibs) + 0) ptr = mib->mibs; \ ^ include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^ net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^ >> include/net/ipv6.h:163:35: error: 'struct netns_mib' has no member named 'ipv6_statistics' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/snmp.h:145:37: note: in definition of macro 'SNMP_UPD_PO_STATS' __typeof__((mib->mibs) + 0) ptr = mib->mibs; \ ^ include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^ net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^ In file included from include/asm-generic/percpu.h:6:0, from arch/x86/include/asm/percpu.h:542, from arch/x86/include/asm/preempt.h:5, from include/linux/preempt.h:59, from include/linux/spinlock.h:50, from include/linux/mm_types.h:8, from include/linux/kmemcheck.h:4, from include/linux/skbuff.h:18, from net/mpls/af_mpls.c:2: include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/linux/percpu-defs.h:206:47: note: in definition of macro '__verify_pcpu_ptr' const void __percpu *__vpp_verify = (typeof((ptr) + 0))NULL; \ ^ include/linux/percpu-defs.h:496:33: note: in expansion of macro '__pcpu_size_call' #define this_cpu_add(pcp, val) __pcpu_size_call(this_cpu_add_, pcp, val) ^ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^ net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^ include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/linux/percpu-defs.h:363:16: note: in definition of macro '__pcpu_size_call' switch(sizeof(variable)) { \ ^ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^ net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^ In file included from arch/x86/include/asm/preempt.h:5:0, from include/linux/preempt.h:59, from include/linux/spinlock.h:50, from include/linux/mm_types.h:8, from include/linux/kmemcheck.h:4, from include/linux/skbuff.h:18, from net/mpls/af_mpls.c:2: include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector this_cpu_inc(ptr[basefield##PKTS]); \ ^ arch/x86/include/asm/percpu.h:128:17: note: in definition of macro 'percpu_add_op' typedef typeof(var) pao_T__; \ ^ include/linux/percpu-defs.h:364:11: note: in expansion of macro 'this_cpu_add_1' case 1: stem##1(variable, __VA_ARGS__);break; \ ^ include/linux/percpu-defs.h:496:33: note: in expansion of macro '__pcpu_size_call' #define this_cpu_add(pcp, val) __pcpu_size_call(this_cpu_add_, pcp, val) ^ include/linux/percpu-defs.h:507:28: note: in expansion of macro 'this_cpu_add' #define this_cpu_inc(pcp) this_cpu_add(pcp, 1) ^ include/net/snmp.h:146:3: note: in expansion of macro 'this_cpu_inc' this_cpu_inc(ptr[basefield##PKTS]); \ ^ include/net/ipv6.h:163:7: note: in expansion of macro 'SNMP_UPD_PO_STATS' mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ ^ include/net/ipv6.h:177:3: note: in expansion of macro '_DEVUPD' _DEVUPD(net, ipv6, , idev, field, val) ^ net/mpls/af_mpls.c:114:4: note: in expansion of macro 'IP6_UPD_PO_STATS' IP6_UPD_PO_STATS(dev_net(dev), in6dev, ^ include/net/snmp.h:146:19: error: subscripted value is neither array nor pointer nor vector vim +163 include/net/ipv6.h 8e7999c4 Pavel Emelyanov 2007-10-15 157 13415e46 Eric Dumazet 2016-04-27 158 #define _DEVUPD(net, statname, mod, idev, field, val) \ edf391ff Neil Horman 2009-04-27 159 ({ \ edf391ff Neil Horman 2009-04-27 160 struct inet6_dev *_idev = (idev); \ edf391ff Neil Horman 2009-04-27 161 if (likely(_idev != NULL)) \ 13415e46 Eric Dumazet 2016-04-27 162 mod##SNMP_UPD_PO_STATS((_idev)->stats.statname, field, (val)); \ 13415e46 Eric Dumazet 2016-04-27 @163 mod##SNMP_UPD_PO_STATS((net)->mib.statname##_statistics, field, (val));\ edf391ff Neil Horman 2009-04-27 164 }) edf391ff Neil Horman 2009-04-27 165 14878f75 David L Stevens 2007-09-16 166 /* MIBs */ :::::: The code at line 163 was first introduced by commit :::::: 13415e46c5915e2dac089de516369005fbc045f9 net: snmp: kill STATS_BH macros :::::: TO: Eric Dumazet <edumazet@google.com> :::::: CC: David S. Miller <davem@davemloft.net> --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation [-- Attachment #2: .config.gz --] [-- Type: application/gzip, Size: 24141 bytes --] ^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH net-next 0/2] mpls: Packet stats 2017-01-13 18:14 [PATCH net-next 0/2] mpls: Packet stats Robert Shearman 2017-01-13 18:14 ` [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman 2017-01-13 18:14 ` [PATCH net-next 2/2] mpls: Packet stats Robert Shearman @ 2017-01-13 22:50 ` Roopa Prabhu 2017-01-16 14:16 ` [PATCH net-next v2 " Robert Shearman 3 siblings, 0 replies; 13+ messages in thread From: Roopa Prabhu @ 2017-01-13 22:50 UTC (permalink / raw) To: Robert Shearman; +Cc: davem, netdev, David Ahern, ebiederm On 1/13/17, 10:14 AM, Robert Shearman wrote: > This patchset records per-interface packet stats in the MPLS > forwarding path and exports them using a nest of attributes root at a > new IFLA_STATS_AF_SPEC attribute as part of RTM_GETSTATS messages: > > [IFLA_STATS_AF_SPEC] > -> [AF_MPLS] > -> [MPLS_STATS_LINK] > -> struct mpls_link_stats > > The first patch adds the rtnl infrastructure for this, including a new > callbacks to per-AF ops of fill_stats_af and get_stats_af_size. The > second patch records MPLS stats and makes use of the infrastructure to > export them. The rtnl infrastructure could also be used to export IPv6 > stats in the future. > > Robert Shearman (2): > net: AF-specific RTM_GETSTATS attributes > mpls: Packet stats > LGTM, thanks Robert ^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH net-next v2 0/2] mpls: Packet stats 2017-01-13 18:14 [PATCH net-next 0/2] mpls: Packet stats Robert Shearman ` (2 preceding siblings ...) 2017-01-13 22:50 ` [PATCH net-next 0/2] " Roopa Prabhu @ 2017-01-16 14:16 ` Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman ` (2 more replies) 3 siblings, 3 replies; 13+ messages in thread From: Robert Shearman @ 2017-01-16 14:16 UTC (permalink / raw) To: davem; +Cc: netdev, David Ahern, ebiederm, Roopa Prabhu, Robert Shearman This patchset records per-interface packet stats in the MPLS forwarding path and exports them using a nest of attributes root at a new IFLA_STATS_AF_SPEC attribute as part of RTM_GETSTATS messages: [IFLA_STATS_AF_SPEC] -> [AF_MPLS] -> [MPLS_STATS_LINK] -> struct mpls_link_stats The first patch adds the rtnl infrastructure for this, including a new callbacks to per-AF ops of fill_stats_af and get_stats_af_size. The second patch records MPLS stats and makes use of the infrastructure to export them. The rtnl infrastructure could also be used to export IPv6 stats in the future. Changes in v2: - make incrementing IPv6 stats in mpls_stats_inc_outucastpkts conditional on CONFIG_IPV6 to fix build with CONFIG_IPV6=n Robert Shearman (2): net: AF-specific RTM_GETSTATS attributes mpls: Packet stats include/net/rtnetlink.h | 4 + include/uapi/linux/if_link.h | 1 + include/uapi/linux/mpls.h | 30 +++++++ net/core/rtnetlink.c | 50 ++++++++++++ net/mpls/af_mpls.c | 181 +++++++++++++++++++++++++++++++++++++------ net/mpls/internal.h | 58 +++++++++++++- net/mpls/mpls_iptunnel.c | 11 ++- 7 files changed, 307 insertions(+), 28 deletions(-) -- 2.1.4 ^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH net-next v2 1/2] net: AF-specific RTM_GETSTATS attributes 2017-01-16 14:16 ` [PATCH net-next v2 " Robert Shearman @ 2017-01-16 14:16 ` Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 2/2] mpls: Packet stats Robert Shearman 2017-01-17 19:39 ` [PATCH net-next v2 0/2] " David Miller 2 siblings, 0 replies; 13+ messages in thread From: Robert Shearman @ 2017-01-16 14:16 UTC (permalink / raw) To: davem; +Cc: netdev, David Ahern, ebiederm, Roopa Prabhu, Robert Shearman Add the functionality for including address-family-specific per-link stats in RTM_GETSTATS messages. This is done through adding a new IFLA_STATS_AF_SPEC attribute under which address family attributes are nested and then the AF-specific attributes can be further nested. This follows the model of IFLA_AF_SPEC on RTM_*LINK messages and it has the advantage of presenting an easily extended hierarchy. The rtnl_af_ops structure is extended to provide AFs with the opportunity to fill and provide the size of their stats attributes. One alternative would have been to provide AFs with the ability to add attributes directly into the RTM_GETSTATS message without a nested hierarchy. I discounted this approach as it increases the rate at which the 32 attribute number space is used up and it makes implementation a little more tricky for stats dump resuming (at the moment the order in which attributes are added to the message has to match the numeric order of the attributes). Another alternative would have been to register per-AF RTM_GETSTATS handlers. I discounted this approach as I perceived a common use-case to be getting all the stats for an interface and this approach would necessitate multiple requests/dumps to retrieve them all. Signed-off-by: Robert Shearman <rshearma@brocade.com> Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com> --- include/net/rtnetlink.h | 4 ++++ include/uapi/linux/if_link.h | 1 + net/core/rtnetlink.c | 50 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 55 insertions(+) diff --git a/include/net/rtnetlink.h b/include/net/rtnetlink.h index 4113916cc1bb..106de5f7bf06 100644 --- a/include/net/rtnetlink.h +++ b/include/net/rtnetlink.h @@ -139,6 +139,10 @@ struct rtnl_af_ops { const struct nlattr *attr); int (*set_link_af)(struct net_device *dev, const struct nlattr *attr); + + int (*fill_stats_af)(struct sk_buff *skb, + const struct net_device *dev); + size_t (*get_stats_af_size)(const struct net_device *dev); }; void __rtnl_af_unregister(struct rtnl_af_ops *ops); diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h index 6b13e591abc9..184b16ed2b84 100644 --- a/include/uapi/linux/if_link.h +++ b/include/uapi/linux/if_link.h @@ -847,6 +847,7 @@ enum { IFLA_STATS_LINK_XSTATS, IFLA_STATS_LINK_XSTATS_SLAVE, IFLA_STATS_LINK_OFFLOAD_XSTATS, + IFLA_STATS_AF_SPEC, __IFLA_STATS_MAX, }; diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 18b5aae99bec..4edc1bd7a735 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -3829,6 +3829,39 @@ static int rtnl_fill_statsinfo(struct sk_buff *skb, struct net_device *dev, *idxattr = 0; } + if (stats_attr_valid(filter_mask, IFLA_STATS_AF_SPEC, *idxattr)) { + struct rtnl_af_ops *af_ops; + + *idxattr = IFLA_STATS_AF_SPEC; + attr = nla_nest_start(skb, IFLA_STATS_AF_SPEC); + if (!attr) + goto nla_put_failure; + + list_for_each_entry(af_ops, &rtnl_af_ops, list) { + if (af_ops->fill_stats_af) { + struct nlattr *af; + int err; + + af = nla_nest_start(skb, af_ops->family); + if (!af) + goto nla_put_failure; + + err = af_ops->fill_stats_af(skb, dev); + + if (err == -ENODATA) + nla_nest_cancel(skb, af); + else if (err < 0) + goto nla_put_failure; + + nla_nest_end(skb, af); + } + } + + nla_nest_end(skb, attr); + + *idxattr = 0; + } + nlmsg_end(skb, nlh); return 0; @@ -3885,6 +3918,23 @@ static size_t if_nlmsg_stats_size(const struct net_device *dev, if (stats_attr_valid(filter_mask, IFLA_STATS_LINK_OFFLOAD_XSTATS, 0)) size += rtnl_get_offload_stats_size(dev); + if (stats_attr_valid(filter_mask, IFLA_STATS_AF_SPEC, 0)) { + struct rtnl_af_ops *af_ops; + + /* for IFLA_STATS_AF_SPEC */ + size += nla_total_size(0); + + list_for_each_entry(af_ops, &rtnl_af_ops, list) { + if (af_ops->get_stats_af_size) { + size += nla_total_size( + af_ops->get_stats_af_size(dev)); + + /* for AF_* */ + size += nla_total_size(0); + } + } + } + return size; } -- 2.1.4 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next v2 2/2] mpls: Packet stats 2017-01-16 14:16 ` [PATCH net-next v2 " Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman @ 2017-01-16 14:16 ` Robert Shearman 2017-01-17 19:39 ` [PATCH net-next v2 0/2] " David Miller 2 siblings, 0 replies; 13+ messages in thread From: Robert Shearman @ 2017-01-16 14:16 UTC (permalink / raw) To: davem; +Cc: netdev, David Ahern, ebiederm, Roopa Prabhu, Robert Shearman Having MPLS packet stats is useful for observing network operation and for diagnosing network problems. In the absence of anything better, RFC2863 and RFC3813 are used for guidance for which stats to expose and the semantics of them. In particular rx_noroutes maps to in unknown protos in RFC2863. The stats are exposed to userspace via AF_MPLS attributes embedded in the IFLA_STATS_AF_SPEC attribute of RTM_GETSTATS messages. All the introduced fields are 64-bit, even error ones, to ensure no overflow with long uptimes. Per-CPU counters are used to avoid cache-line contention on the commonly used fields. The other fields have also been made per-CPU for code to avoid performance problems in error conditions on the assumption that on some platforms the cost of atomic operations could be more expensive than sending the packet (which is what would be done in the success case). If that's not the case, we could instead not use per-CPU counters for these fields. Only unicast and non-fragment are exposed at the moment, but other counters can be exposed in the future either by adding to the end of struct mpls_link_stats or by additional netlink attributes in the AF_MPLS IFLA_STATS_AF_SPEC nested attribute. Signed-off-by: Robert Shearman <rshearma@brocade.com> --- include/uapi/linux/mpls.h | 30 ++++++++ net/mpls/af_mpls.c | 181 ++++++++++++++++++++++++++++++++++++++++------ net/mpls/internal.h | 58 ++++++++++++++- net/mpls/mpls_iptunnel.c | 11 ++- 4 files changed, 252 insertions(+), 28 deletions(-) diff --git a/include/uapi/linux/mpls.h b/include/uapi/linux/mpls.h index 24a6cb1aec86..77a19dfe3990 100644 --- a/include/uapi/linux/mpls.h +++ b/include/uapi/linux/mpls.h @@ -43,4 +43,34 @@ struct mpls_label { #define MPLS_LABEL_FIRST_UNRESERVED 16 /* RFC3032 */ +/* These are embedded into IFLA_STATS_AF_SPEC: + * [IFLA_STATS_AF_SPEC] + * -> [AF_MPLS] + * -> [MPLS_STATS_xxx] + * + * Attributes: + * [MPLS_STATS_LINK] = { + * struct mpls_link_stats + * } + */ +enum { + MPLS_STATS_UNSPEC, /* also used as 64bit pad attribute */ + MPLS_STATS_LINK, + __MPLS_STATS_MAX, +}; + +#define MPLS_STATS_MAX (__MPLS_STATS_MAX - 1) + +struct mpls_link_stats { + __u64 rx_packets; /* total packets received */ + __u64 tx_packets; /* total packets transmitted */ + __u64 rx_bytes; /* total bytes received */ + __u64 tx_bytes; /* total bytes transmitted */ + __u64 rx_errors; /* bad packets received */ + __u64 tx_errors; /* packet transmit problems */ + __u64 rx_dropped; /* packet dropped on receive */ + __u64 tx_dropped; /* packet dropped on transmit */ + __u64 rx_noroute; /* no route for packet dest */ +}; + #endif /* _UAPI_MPLS_H */ diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 15fe97644ffe..4dc81963af8f 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -8,6 +8,7 @@ #include <linux/ipv6.h> #include <linux/mpls.h> #include <linux/vmalloc.h> +#include <linux/percpu.h> #include <net/ip.h> #include <net/dst.h> #include <net/sock.h> @@ -17,8 +18,8 @@ #include <net/netns/generic.h> #if IS_ENABLED(CONFIG_IPV6) #include <net/ipv6.h> -#include <net/addrconf.h> #endif +#include <net/addrconf.h> #include <net/nexthop.h> #include "internal.h" @@ -48,11 +49,6 @@ static struct mpls_route *mpls_route_input_rcu(struct net *net, unsigned index) return rt; } -static inline struct mpls_dev *mpls_dev_get(const struct net_device *dev) -{ - return rcu_dereference_rtnl(dev->mpls_ptr); -} - bool mpls_output_possible(const struct net_device *dev) { return dev && (dev->flags & IFF_UP) && netif_carrier_ok(dev); @@ -98,6 +94,31 @@ bool mpls_pkt_too_big(const struct sk_buff *skb, unsigned int mtu) } EXPORT_SYMBOL_GPL(mpls_pkt_too_big); +void mpls_stats_inc_outucastpkts(struct net_device *dev, + const struct sk_buff *skb) +{ + struct mpls_dev *mdev; + + if (skb->protocol == htons(ETH_P_MPLS_UC)) { + mdev = mpls_dev_get(dev); + if (mdev) + MPLS_INC_STATS_LEN(mdev, skb->len, + tx_packets, + tx_bytes); + } else if (skb->protocol == htons(ETH_P_IP)) { + IP_UPD_PO_STATS(dev_net(dev), IPSTATS_MIB_OUT, skb->len); +#if IS_ENABLED(CONFIG_IPV6) + } else if (skb->protocol == htons(ETH_P_IPV6)) { + struct inet6_dev *in6dev = __in6_dev_get(dev); + + if (in6dev) + IP6_UPD_PO_STATS(dev_net(dev), in6dev, + IPSTATS_MIB_OUT, skb->len); +#endif + } +} +EXPORT_SYMBOL_GPL(mpls_stats_inc_outucastpkts); + static u32 mpls_multipath_hash(struct mpls_route *rt, struct sk_buff *skb, bool bos) { @@ -253,6 +274,7 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, struct mpls_nh *nh; struct mpls_entry_decoded dec; struct net_device *out_dev; + struct mpls_dev *out_mdev; struct mpls_dev *mdev; unsigned int hh_len; unsigned int new_header_size; @@ -262,17 +284,25 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, /* Careful this entire function runs inside of an rcu critical section */ mdev = mpls_dev_get(dev); - if (!mdev || !mdev->input_enabled) + if (!mdev) goto drop; - if (skb->pkt_type != PACKET_HOST) + MPLS_INC_STATS_LEN(mdev, skb->len, rx_packets, + rx_bytes); + + if (!mdev->input_enabled) { + MPLS_INC_STATS(mdev, rx_dropped); goto drop; + } + + if (skb->pkt_type != PACKET_HOST) + goto err; if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL) - goto drop; + goto err; if (!pskb_may_pull(skb, sizeof(*hdr))) - goto drop; + goto err; /* Read and decode the label */ hdr = mpls_hdr(skb); @@ -285,33 +315,35 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, skb_orphan(skb); rt = mpls_route_input_rcu(net, dec.label); - if (!rt) + if (!rt) { + MPLS_INC_STATS(mdev, rx_noroute); goto drop; + } nh = mpls_select_multipath(rt, skb, dec.bos); if (!nh) - goto drop; - - /* Find the output device */ - out_dev = rcu_dereference(nh->nh_dev); - if (!mpls_output_possible(out_dev)) - goto drop; + goto err; if (skb_warn_if_lro(skb)) - goto drop; + goto err; skb_forward_csum(skb); /* Verify ttl is valid */ if (dec.ttl <= 1) - goto drop; + goto err; dec.ttl -= 1; + /* Find the output device */ + out_dev = rcu_dereference(nh->nh_dev); + if (!mpls_output_possible(out_dev)) + goto tx_err; + /* Verify the destination can hold the packet */ new_header_size = mpls_nh_header_size(nh); mtu = mpls_dev_mtu(out_dev); if (mpls_pkt_too_big(skb, mtu - new_header_size)) - goto drop; + goto tx_err; hh_len = LL_RESERVED_SPACE(out_dev); if (!out_dev->header_ops) @@ -319,7 +351,7 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, /* Ensure there is enough space for the headers in the skb */ if (skb_cow(skb, hh_len + new_header_size)) - goto drop; + goto tx_err; skb->dev = out_dev; skb->protocol = htons(ETH_P_MPLS_UC); @@ -327,7 +359,7 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, if (unlikely(!new_header_size && dec.bos)) { /* Penultimate hop popping */ if (!mpls_egress(rt, skb, dec)) - goto drop; + goto err; } else { bool bos; int i; @@ -343,6 +375,8 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, } } + mpls_stats_inc_outucastpkts(out_dev, skb); + /* If via wasn't specified then send out using device address */ if (nh->nh_via_table == MPLS_NEIGH_TABLE_UNSPEC) err = neigh_xmit(NEIGH_LINK_TABLE, out_dev, @@ -355,6 +389,13 @@ static int mpls_forward(struct sk_buff *skb, struct net_device *dev, __func__, err); return 0; +tx_err: + out_mdev = out_dev ? mpls_dev_get(out_dev) : NULL; + if (out_mdev) + MPLS_INC_STATS(out_mdev, tx_errors); + goto drop; +err: + MPLS_INC_STATS(mdev, rx_errors); drop: kfree_skb(skb); return NET_RX_DROP; @@ -853,6 +894,70 @@ static int mpls_route_del(struct mpls_route_config *cfg) return err; } +static void mpls_get_stats(struct mpls_dev *mdev, + struct mpls_link_stats *stats) +{ + struct mpls_pcpu_stats *p; + int i; + + memset(stats, 0, sizeof(*stats)); + + for_each_possible_cpu(i) { + struct mpls_link_stats local; + unsigned int start; + + p = per_cpu_ptr(mdev->stats, i); + do { + start = u64_stats_fetch_begin(&p->syncp); + local = p->stats; + } while (u64_stats_fetch_retry(&p->syncp, start)); + + stats->rx_packets += local.rx_packets; + stats->rx_bytes += local.rx_bytes; + stats->tx_packets += local.tx_packets; + stats->tx_bytes += local.tx_bytes; + stats->rx_errors += local.rx_errors; + stats->tx_errors += local.tx_errors; + stats->rx_dropped += local.rx_dropped; + stats->tx_dropped += local.tx_dropped; + stats->rx_noroute += local.rx_noroute; + } +} + +static int mpls_fill_stats_af(struct sk_buff *skb, + const struct net_device *dev) +{ + struct mpls_link_stats *stats; + struct mpls_dev *mdev; + struct nlattr *nla; + + mdev = mpls_dev_get(dev); + if (!mdev) + return -ENODATA; + + nla = nla_reserve_64bit(skb, MPLS_STATS_LINK, + sizeof(struct mpls_link_stats), + MPLS_STATS_UNSPEC); + if (!nla) + return -EMSGSIZE; + + stats = nla_data(nla); + mpls_get_stats(mdev, stats); + + return 0; +} + +static size_t mpls_get_stats_af_size(const struct net_device *dev) +{ + struct mpls_dev *mdev; + + mdev = mpls_dev_get(dev); + if (!mdev) + return 0; + + return nla_total_size_64bit(sizeof(struct mpls_link_stats)); +} + #define MPLS_PERDEV_SYSCTL_OFFSET(field) \ (&((struct mpls_dev *)0)->field) @@ -911,6 +1016,7 @@ static struct mpls_dev *mpls_add_dev(struct net_device *dev) { struct mpls_dev *mdev; int err = -ENOMEM; + int i; ASSERT_RTNL(); @@ -918,6 +1024,17 @@ static struct mpls_dev *mpls_add_dev(struct net_device *dev) if (!mdev) return ERR_PTR(err); + mdev->stats = alloc_percpu(struct mpls_pcpu_stats); + if (!mdev->stats) + goto free; + + for_each_possible_cpu(i) { + struct mpls_pcpu_stats *mpls_stats; + + mpls_stats = per_cpu_ptr(mdev->stats, i); + u64_stats_init(&mpls_stats->syncp); + } + err = mpls_dev_sysctl_register(dev, mdev); if (err) goto free; @@ -927,10 +1044,19 @@ static struct mpls_dev *mpls_add_dev(struct net_device *dev) return mdev; free: + free_percpu(mdev->stats); kfree(mdev); return ERR_PTR(err); } +static void mpls_dev_destroy_rcu(struct rcu_head *head) +{ + struct mpls_dev *mdev = container_of(head, struct mpls_dev, rcu); + + free_percpu(mdev->stats); + kfree(mdev); +} + static void mpls_ifdown(struct net_device *dev, int event) { struct mpls_route __rcu **platform_label; @@ -1045,7 +1171,7 @@ static int mpls_dev_notify(struct notifier_block *this, unsigned long event, if (mdev) { mpls_dev_sysctl_unregister(mdev); RCU_INIT_POINTER(dev->mpls_ptr, NULL); - kfree_rcu(mdev, rcu); + call_rcu(&mdev->rcu, mpls_dev_destroy_rcu); } break; case NETDEV_CHANGENAME: @@ -1706,6 +1832,12 @@ static struct pernet_operations mpls_net_ops = { .exit = mpls_net_exit, }; +static struct rtnl_af_ops mpls_af_ops __read_mostly = { + .family = AF_MPLS, + .fill_stats_af = mpls_fill_stats_af, + .get_stats_af_size = mpls_get_stats_af_size, +}; + static int __init mpls_init(void) { int err; @@ -1722,6 +1854,8 @@ static int __init mpls_init(void) dev_add_pack(&mpls_packet_type); + rtnl_af_register(&mpls_af_ops); + rtnl_register(PF_MPLS, RTM_NEWROUTE, mpls_rtm_newroute, NULL, NULL); rtnl_register(PF_MPLS, RTM_DELROUTE, mpls_rtm_delroute, NULL, NULL); rtnl_register(PF_MPLS, RTM_GETROUTE, NULL, mpls_dump_routes, NULL); @@ -1738,6 +1872,7 @@ module_init(mpls_init); static void __exit mpls_exit(void) { rtnl_unregister_all(PF_MPLS); + rtnl_af_unregister(&mpls_af_ops); dev_remove_pack(&mpls_packet_type); unregister_netdevice_notifier(&mpls_dev_notifier); unregister_pernet_subsys(&mpls_net_ops); diff --git a/net/mpls/internal.h b/net/mpls/internal.h index bdfef6c3271a..d97243034605 100644 --- a/net/mpls/internal.h +++ b/net/mpls/internal.h @@ -9,13 +9,58 @@ struct mpls_entry_decoded { u8 bos; }; +struct mpls_pcpu_stats { + struct mpls_link_stats stats; + struct u64_stats_sync syncp; +}; + struct mpls_dev { - int input_enabled; + int input_enabled; - struct ctl_table_header *sysctl; - struct rcu_head rcu; + struct mpls_pcpu_stats __percpu *stats; + + struct ctl_table_header *sysctl; + struct rcu_head rcu; }; +#if BITS_PER_LONG == 32 + +#define MPLS_INC_STATS_LEN(mdev, len, pkts_field, bytes_field) \ + do { \ + __typeof__(*(mdev)->stats) *ptr = \ + raw_cpu_ptr((mdev)->stats); \ + local_bh_disable(); \ + u64_stats_update_begin(&ptr->syncp); \ + ptr->stats.pkts_field++; \ + ptr->stats.bytes_field += (len); \ + u64_stats_update_end(&ptr->syncp); \ + local_bh_enable(); \ + } while (0) + +#define MPLS_INC_STATS(mdev, field) \ + do { \ + __typeof__(*(mdev)->stats) *ptr = \ + raw_cpu_ptr((mdev)->stats); \ + local_bh_disable(); \ + u64_stats_update_begin(&ptr->syncp); \ + ptr->stats.field++; \ + u64_stats_update_end(&ptr->syncp); \ + local_bh_enable(); \ + } while (0) + +#else + +#define MPLS_INC_STATS_LEN(mdev, len, pkts_field, bytes_field) \ + do { \ + this_cpu_inc((mdev)->stats->stats.pkts_field); \ + this_cpu_add((mdev)->stats->stats.bytes_field, (len)); \ + } while (0) + +#define MPLS_INC_STATS(mdev, field) \ + this_cpu_inc((mdev)->stats->stats.field) + +#endif + struct sk_buff; #define LABEL_NOT_SPECIFIED (1 << 20) @@ -114,6 +159,11 @@ static inline struct mpls_entry_decoded mpls_entry_decode(struct mpls_shim_hdr * return result; } +static inline struct mpls_dev *mpls_dev_get(const struct net_device *dev) +{ + return rcu_dereference_rtnl(dev->mpls_ptr); +} + int nla_put_labels(struct sk_buff *skb, int attrtype, u8 labels, const u32 label[]); int nla_get_labels(const struct nlattr *nla, u32 max_labels, u8 *labels, @@ -123,5 +173,7 @@ int nla_get_via(const struct nlattr *nla, u8 *via_alen, u8 *via_table, bool mpls_output_possible(const struct net_device *dev); unsigned int mpls_dev_mtu(const struct net_device *dev); bool mpls_pkt_too_big(const struct sk_buff *skb, unsigned int mtu); +void mpls_stats_inc_outucastpkts(struct net_device *dev, + const struct sk_buff *skb); #endif /* MPLS_INTERNAL_H */ diff --git a/net/mpls/mpls_iptunnel.c b/net/mpls/mpls_iptunnel.c index 2f7ccd934416..02531284bc49 100644 --- a/net/mpls/mpls_iptunnel.c +++ b/net/mpls/mpls_iptunnel.c @@ -48,11 +48,15 @@ static int mpls_xmit(struct sk_buff *skb) struct dst_entry *dst = skb_dst(skb); struct rtable *rt = NULL; struct rt6_info *rt6 = NULL; + struct mpls_dev *out_mdev; int err = 0; bool bos; int i; unsigned int ttl; + /* Find the output device */ + out_dev = dst->dev; + /* Obtain the ttl */ if (dst->ops->family == AF_INET) { ttl = ip_hdr(skb)->ttl; @@ -66,8 +70,6 @@ static int mpls_xmit(struct sk_buff *skb) skb_orphan(skb); - /* Find the output device */ - out_dev = dst->dev; if (!mpls_output_possible(out_dev) || !dst->lwtstate || skb_warn_if_lro(skb)) goto drop; @@ -109,6 +111,8 @@ static int mpls_xmit(struct sk_buff *skb) bos = false; } + mpls_stats_inc_outucastpkts(out_dev, skb); + if (rt) err = neigh_xmit(NEIGH_ARP_TABLE, out_dev, &rt->rt_gateway, skb); @@ -122,6 +126,9 @@ static int mpls_xmit(struct sk_buff *skb) return LWTUNNEL_XMIT_DONE; drop: + out_mdev = out_dev ? mpls_dev_get(out_dev) : NULL; + if (out_mdev) + MPLS_INC_STATS(out_mdev, tx_errors); kfree_skb(skb); return -EINVAL; } -- 2.1.4 ^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH net-next v2 0/2] mpls: Packet stats 2017-01-16 14:16 ` [PATCH net-next v2 " Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 2/2] mpls: Packet stats Robert Shearman @ 2017-01-17 19:39 ` David Miller 2 siblings, 0 replies; 13+ messages in thread From: David Miller @ 2017-01-17 19:39 UTC (permalink / raw) To: rshearma; +Cc: netdev, dsa, ebiederm, roopa From: Robert Shearman <rshearma@brocade.com> Date: Mon, 16 Jan 2017 14:16:35 +0000 > This patchset records per-interface packet stats in the MPLS > forwarding path and exports them using a nest of attributes root at a > new IFLA_STATS_AF_SPEC attribute as part of RTM_GETSTATS messages: > > [IFLA_STATS_AF_SPEC] > -> [AF_MPLS] > -> [MPLS_STATS_LINK] > -> struct mpls_link_stats > > The first patch adds the rtnl infrastructure for this, including a new > callbacks to per-AF ops of fill_stats_af and get_stats_af_size. The > second patch records MPLS stats and makes use of the infrastructure to > export them. The rtnl infrastructure could also be used to export IPv6 > stats in the future. > > Changes in v2: > - make incrementing IPv6 stats in mpls_stats_inc_outucastpkts > conditional on CONFIG_IPV6 to fix build with CONFIG_IPV6=n Series applied, thanks. ^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2017-01-17 19:39 UTC | newest] Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-01-13 18:14 [PATCH net-next 0/2] mpls: Packet stats Robert Shearman 2017-01-13 18:14 ` [PATCH net-next 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman 2017-01-13 22:50 ` Roopa Prabhu 2017-01-13 18:14 ` [PATCH net-next 2/2] mpls: Packet stats Robert Shearman 2017-01-13 22:51 ` Roopa Prabhu 2017-01-14 6:41 ` kbuild test robot 2017-01-16 10:30 ` Robert Shearman 2017-01-14 6:58 ` kbuild test robot 2017-01-13 22:50 ` [PATCH net-next 0/2] " Roopa Prabhu 2017-01-16 14:16 ` [PATCH net-next v2 " Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 1/2] net: AF-specific RTM_GETSTATS attributes Robert Shearman 2017-01-16 14:16 ` [PATCH net-next v2 2/2] mpls: Packet stats Robert Shearman 2017-01-17 19:39 ` [PATCH net-next v2 0/2] " David Miller
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.