Nikolay Aleksandrov
2018-Oct-12 10:41 UTC
[Bridge] [PATCH net-next] net: bridge: add support for per-port vlan stats
This patch adds an option to have per-port vlan stats instead of the
default global stats. The option can be set only when there are no port
vlans in the bridge since we need to allocate the stats if it is set
when vlans are being added to ports (and respectively free them
when being deleted). Also bump RTNL_MAX_TYPE as the bridge is the
largest user of options. The current stats design allows us to add
these without any changes to the fast-path, it all comes down to
the per-vlan stats pointer which, if this option is enabled, will
be allocated for each port vlan instead of using the global bridge-wide
one.
CC: bridge at lists.linux-foundation.org
CC: Roopa Prabhu <roopa at cumulusnetworks.com>
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
include/uapi/linux/if_link.h | 1 +
net/bridge/br_netlink.c | 14 ++++++++++-
net/bridge/br_private.h | 2 ++
net/bridge/br_sysfs_br.c | 17 +++++++++++++
net/bridge/br_vlan.c | 49 ++++++++++++++++++++++++++++++++++--
net/core/rtnetlink.c | 2 +-
6 files changed, 81 insertions(+), 4 deletions(-)
diff --git a/include/uapi/linux/if_link.h b/include/uapi/linux/if_link.h
index 58faab897201..1debfa42cba1 100644
--- a/include/uapi/linux/if_link.h
+++ b/include/uapi/linux/if_link.h
@@ -287,6 +287,7 @@ enum {
IFLA_BR_MCAST_STATS_ENABLED,
IFLA_BR_MCAST_IGMP_VERSION,
IFLA_BR_MCAST_MLD_VERSION,
+ IFLA_BR_VLAN_STATS_PER_PORT,
__IFLA_BR_MAX,
};
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index e5a5bc5d5232..3345f1984542 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -1034,6 +1034,7 @@ static const struct nla_policy br_policy[IFLA_BR_MAX + 1]
= {
[IFLA_BR_MCAST_STATS_ENABLED] = { .type = NLA_U8 },
[IFLA_BR_MCAST_IGMP_VERSION] = { .type = NLA_U8 },
[IFLA_BR_MCAST_MLD_VERSION] = { .type = NLA_U8 },
+ [IFLA_BR_VLAN_STATS_PER_PORT] = { .type = NLA_U8 },
};
static int br_changelink(struct net_device *brdev, struct nlattr *tb[],
@@ -1114,6 +1115,14 @@ static int br_changelink(struct net_device *brdev, struct
nlattr *tb[],
if (err)
return err;
}
+
+ if (data[IFLA_BR_VLAN_STATS_PER_PORT]) {
+ __u8 per_port = nla_get_u8(data[IFLA_BR_VLAN_STATS_PER_PORT]);
+
+ err = br_vlan_set_stats_per_port(br, per_port);
+ if (err)
+ return err;
+ }
#endif
if (data[IFLA_BR_GROUP_FWD_MASK]) {
@@ -1327,6 +1336,7 @@ static size_t br_get_size(const struct net_device *brdev)
nla_total_size(sizeof(__be16)) + /* IFLA_BR_VLAN_PROTOCOL */
nla_total_size(sizeof(u16)) + /* IFLA_BR_VLAN_DEFAULT_PVID */
nla_total_size(sizeof(u8)) + /* IFLA_BR_VLAN_STATS_ENABLED */
+ nla_total_size(sizeof(u8)) + /* IFLA_BR_VLAN_STATS_PER_PORT */
#endif
nla_total_size(sizeof(u16)) + /* IFLA_BR_GROUP_FWD_MASK */
nla_total_size(sizeof(struct ifla_bridge_id)) + /* IFLA_BR_ROOT_ID */
@@ -1417,7 +1427,9 @@ static int br_fill_info(struct sk_buff *skb, const struct
net_device *brdev)
if (nla_put_be16(skb, IFLA_BR_VLAN_PROTOCOL, br->vlan_proto) ||
nla_put_u16(skb, IFLA_BR_VLAN_DEFAULT_PVID, br->default_pvid) ||
nla_put_u8(skb, IFLA_BR_VLAN_STATS_ENABLED,
- br_opt_get(br, BROPT_VLAN_STATS_ENABLED)))
+ br_opt_get(br, BROPT_VLAN_STATS_ENABLED)) ||
+ nla_put_u8(skb, IFLA_BR_VLAN_STATS_PER_PORT,
+ br_opt_get(br, IFLA_BR_VLAN_STATS_PER_PORT)))
return -EMSGSIZE;
#endif
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 57229b9d800f..10ee39fdca5c 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -320,6 +320,7 @@ enum net_bridge_opts {
BROPT_HAS_IPV6_ADDR,
BROPT_NEIGH_SUPPRESS_ENABLED,
BROPT_MTU_SET_BY_USER,
+ BROPT_VLAN_STATS_PER_PORT,
};
struct net_bridge {
@@ -859,6 +860,7 @@ int br_vlan_filter_toggle(struct net_bridge *br, unsigned
long val);
int __br_vlan_set_proto(struct net_bridge *br, __be16 proto);
int br_vlan_set_proto(struct net_bridge *br, unsigned long val);
int br_vlan_set_stats(struct net_bridge *br, unsigned long val);
+int br_vlan_set_stats_per_port(struct net_bridge *br, unsigned long val);
int br_vlan_init(struct net_bridge *br);
int br_vlan_set_default_pvid(struct net_bridge *br, unsigned long val);
int __br_vlan_set_default_pvid(struct net_bridge *br, u16 pvid);
diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c
index c93c5724609e..60182bef6341 100644
--- a/net/bridge/br_sysfs_br.c
+++ b/net/bridge/br_sysfs_br.c
@@ -803,6 +803,22 @@ static ssize_t vlan_stats_enabled_store(struct device *d,
return store_bridge_parm(d, buf, len, br_vlan_set_stats);
}
static DEVICE_ATTR_RW(vlan_stats_enabled);
+
+static ssize_t vlan_stats_per_port_show(struct device *d,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct net_bridge *br = to_bridge(d);
+ return sprintf(buf, "%u\n", br_opt_get(br,
BROPT_VLAN_STATS_PER_PORT));
+}
+
+static ssize_t vlan_stats_per_port_store(struct device *d,
+ struct device_attribute *attr,
+ const char *buf, size_t len)
+{
+ return store_bridge_parm(d, buf, len, br_vlan_set_stats_per_port);
+}
+static DEVICE_ATTR_RW(vlan_stats_per_port);
#endif
static struct attribute *bridge_attrs[] = {
@@ -856,6 +872,7 @@ static struct attribute *bridge_attrs[] = {
&dev_attr_vlan_protocol.attr,
&dev_attr_default_pvid.attr,
&dev_attr_vlan_stats_enabled.attr,
+ &dev_attr_vlan_stats_per_port.attr,
#endif
NULL
};
diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c
index 5942e03dd845..9b707234e4ae 100644
--- a/net/bridge/br_vlan.c
+++ b/net/bridge/br_vlan.c
@@ -190,6 +190,19 @@ static void br_vlan_put_master(struct net_bridge_vlan
*masterv)
}
}
+static void nbp_vlan_rcu_free(struct rcu_head *rcu)
+{
+ struct net_bridge_vlan *v;
+
+ v = container_of(rcu, struct net_bridge_vlan, rcu);
+ WARN_ON(br_vlan_is_master(v));
+ /* if we had per-port stats configured then free them here */
+ if (v->brvlan->stats != v->stats)
+ free_percpu(v->stats);
+ v->stats = NULL;
+ kfree(v);
+}
+
/* This is the shared VLAN add function which works for both ports and bridge
* devices. There are four possible calls to this function in terms of the
* vlan entry type:
@@ -245,7 +258,15 @@ static int __vlan_add(struct net_bridge_vlan *v, u16 flags)
if (!masterv)
goto out_filt;
v->brvlan = masterv;
- v->stats = masterv->stats;
+ if (br_opt_get(br, BROPT_VLAN_STATS_PER_PORT)) {
+ v->stats = netdev_alloc_pcpu_stats(struct br_vlan_stats);
+ if (!v->stats) {
+ err = -ENOMEM;
+ goto out_filt;
+ }
+ } else {
+ v->stats = masterv->stats;
+ }
} else {
err = br_switchdev_port_vlan_add(dev, v->vid, flags);
if (err && err != -EOPNOTSUPP)
@@ -329,7 +350,7 @@ static int __vlan_del(struct net_bridge_vlan *v)
rhashtable_remove_fast(&vg->vlan_hash, &v->vnode,
br_vlan_rht_params);
__vlan_del_list(v);
- kfree_rcu(v, rcu);
+ call_rcu(&v->rcu, nbp_vlan_rcu_free);
}
br_vlan_put_master(masterv);
@@ -830,6 +851,30 @@ int br_vlan_set_stats(struct net_bridge *br, unsigned long
val)
return 0;
}
+int br_vlan_set_stats_per_port(struct net_bridge *br, unsigned long val)
+{
+ struct net_bridge_port *p;
+
+ /* allow to change the option if there are no port vlans configured */
+ list_for_each_entry(p, &br->port_list, list) {
+ struct net_bridge_vlan_group *vg = nbp_vlan_group(p);
+
+ if (vg->num_vlans)
+ return -EBUSY;
+ }
+
+ switch (val) {
+ case 0:
+ case 1:
+ br_opt_toggle(br, BROPT_VLAN_STATS_PER_PORT, !!val);
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
static bool vlan_default_pvid(struct net_bridge_vlan_group *vg, u16 vid)
{
struct net_bridge_vlan *v;
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index 46328a10034a..0958c7be2c22 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -59,7 +59,7 @@
#include <net/rtnetlink.h>
#include <net/net_namespace.h>
-#define RTNL_MAX_TYPE 48
+#define RTNL_MAX_TYPE 49
#define RTNL_SLAVE_MAX_TYPE 36
struct rtnl_link {
--
2.17.2
Stephen Hemminger
2018-Oct-12 15:39 UTC
[Bridge] [PATCH net-next] net: bridge: add support for per-port vlan stats
On Fri, 12 Oct 2018 13:41:16 +0300 Nikolay Aleksandrov <nikolay at cumulusnetworks.com> wrote:> This patch adds an option to have per-port vlan stats instead of the > default global stats. The option can be set only when there are no port > vlans in the bridge since we need to allocate the stats if it is set > when vlans are being added to ports (and respectively free them > when being deleted). Also bump RTNL_MAX_TYPE as the bridge is the > largest user of options. The current stats design allows us to add > these without any changes to the fast-path, it all comes down to > the per-vlan stats pointer which, if this option is enabled, will > be allocated for each port vlan instead of using the global bridge-wide > one. > > CC: bridge at lists.linux-foundation.org > CC: Roopa Prabhu <roopa at cumulusnetworks.com> > Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>Yes, per-vlan stats could be quite useful. Most cases of statistics in the kernel are always on, and some API's get them (and skip others). Other than the additional memory overhead, why not make the statistics as always on. Also, is there any chance of creating too much data in a netlink message if there are 4K-1 VLAN's?
Nikolay Aleksandrov
2018-Oct-12 16:04 UTC
[Bridge] [PATCH net-next] net: bridge: add support for per-port vlan stats
On October 12, 2018 6:39:32 PM GMT+03:00, Stephen Hemminger <stephen at networkplumber.org> wrote:>On Fri, 12 Oct 2018 13:41:16 +0300 >Nikolay Aleksandrov <nikolay at cumulusnetworks.com> wrote: > >> This patch adds an option to have per-port vlan stats instead of the >> default global stats. The option can be set only when there are no >port >> vlans in the bridge since we need to allocate the stats if it is set >> when vlans are being added to ports (and respectively free them >> when being deleted). Also bump RTNL_MAX_TYPE as the bridge is the >> largest user of options. The current stats design allows us to add >> these without any changes to the fast-path, it all comes down to >> the per-vlan stats pointer which, if this option is enabled, will >> be allocated for each port vlan instead of using the global >bridge-wide >> one. >> >> CC: bridge at lists.linux-foundation.org >> CC: Roopa Prabhu <roopa at cumulusnetworks.com> >> Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com> > >Yes, per-vlan stats could be quite useful. >Most cases of statistics in the kernel are always on, and some API's >get them (and skip others). Other than the additional memory overhead, >why >not make the statistics as always on. >The additional overhead can be quite a lot and it also affects performance, besides that currently users expect the same accumulated stats.>Also, is there any chance of creating too much data in a netlink >message if there are 4K-1 VLAN's?No, that's okay because we expose the stats via the xstats API which continues the dump on filling the current buffer. We can export as much info as needed. Cheers, Nik
David Miller
2018-Oct-12 17:20 UTC
[Bridge] [PATCH net-next] net: bridge: add support for per-port vlan stats
From: Nikolay Aleksandrov <nikolay at cumulusnetworks.com> Date: Fri, 12 Oct 2018 13:41:16 +0300> This patch adds an option to have per-port vlan stats instead of the > default global stats. The option can be set only when there are no port > vlans in the bridge since we need to allocate the stats if it is set > when vlans are being added to ports (and respectively free them > when being deleted). Also bump RTNL_MAX_TYPE as the bridge is the > largest user of options. The current stats design allows us to add > these without any changes to the fast-path, it all comes down to > the per-vlan stats pointer which, if this option is enabled, will > be allocated for each port vlan instead of using the global bridge-wide > one. > > CC: bridge at lists.linux-foundation.org > CC: Roopa Prabhu <roopa at cumulusnetworks.com> > Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>Applied, thanks.