Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 0/9] net: bridge: convert bool options to bits
Hi,
A lot of boolean bridge options have been added around the net_bridge
structure resulting in holes and more importantly different cache lines
that need to be fetched in the fast path. This set moves all of those
to bits in a bitfield which resides in a hot cache line thus reducing
the size of net_bridge, the number of holes and the number of cache
lines needed for the fast path.
The set is also sent in preparation for new boolean options to avoid
spreading them in the structure and making new holes.
One nice side-effect is that we avoid potential race conditions by using
the bitops since some of the options were bits being directly set in
parallel risking hard to debug issues (has_ipv6_addr).
Before:
size: 1184, holes: 8, sum holes: 30
After:
size: 1160, holes: 2, sum holes: 3
Patch 01 is a trivial style fix
Patch 02 adds the new options bitfield and converts the vlan boolean
options to bits
Patches 03-08 convert the rest of the boolean options to bits
Patch 09 re-arranges a few fields in net_bridge to avoid fetching one
extra cache line in fast path and to further reduce size
Thanks,
Nik
Nikolay Aleksandrov (9):
net: bridge: make struct opening bracket consistent
net: bridge: add bitfield for options and convert vlan opts
net: bridge: convert nf call options to bits
net: bridge: convert group_addr_set option to a bit
net: bridge: convert and rename mcast disabled
net: bridge: convert mcast options to bits
net: bridge: convert neigh_suppress_enabled option to a bit
net: bridge: convert mtu_set_by_user to a bit
net: bridge: pack net_bridge better
net/bridge/br.c | 16 +++++++++
net/bridge/br_arp_nd_proxy.c | 13 +++++---
net/bridge/br_device.c | 6 ++--
net/bridge/br_if.c | 4 +--
net/bridge/br_input.c | 2 +-
net/bridge/br_mdb.c | 6 ++--
net/bridge/br_multicast.c | 54 +++++++++++++++---------------
net/bridge/br_netfilter_hooks.c | 7 ++--
net/bridge/br_netlink.c | 31 +++++++++--------
net/bridge/br_private.h | 74 +++++++++++++++++++++++------------------
net/bridge/br_sysfs_br.c | 32 +++++++++---------
net/bridge/br_vlan.c | 30 +++++++++--------
12 files changed, 155 insertions(+), 120 deletions(-)
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 1/9] net: bridge: make struct opening bracket consistent
Currently we have a mix of opening brackets on new lines and on the same
line, let's move them all on the same line.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_private.h | 12 ++++--------
1 file changed, 4 insertions(+), 8 deletions(-)
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index d21035a17f4c..522d707cc533 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -54,14 +54,12 @@ typedef struct bridge_id bridge_id;
typedef struct mac_addr mac_addr;
typedef __u16 port_id;
-struct bridge_id
-{
+struct bridge_id {
unsigned char prio[2];
unsigned char addr[ETH_ALEN];
};
-struct mac_addr
-{
+struct mac_addr {
unsigned char addr[ETH_ALEN];
};
@@ -207,8 +205,7 @@ struct net_bridge_port_group {
unsigned char eth_addr[ETH_ALEN];
};
-struct net_bridge_mdb_entry
-{
+struct net_bridge_mdb_entry {
struct hlist_node hlist[2];
struct net_bridge *br;
struct net_bridge_port_group __rcu *ports;
@@ -218,8 +215,7 @@ struct net_bridge_mdb_entry
bool host_joined;
};
-struct net_bridge_mdb_htable
-{
+struct net_bridge_mdb_htable {
struct hlist_head *mhash;
struct rcu_head rcu;
struct net_bridge_mdb_htable *old;
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 2/9] net: bridge: add bitfield for options and convert vlan opts
Bridge options have usually been added as separate fields all over the
net_bridge struct taking up space and ending up in different cache lines.
Let's move them to a single bitfield to save up space and speedup lookups.
This patch adds a simple API for option modifying and retrieving using
bitops and converts the first user of the API - the bridge vlan options
(vlan_enabled and vlan_stats_enabled).
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br.c | 16 ++++++++++++++++
net/bridge/br_netlink.c | 3 ++-
net/bridge/br_private.h | 16 ++++++++++++++--
net/bridge/br_sysfs_br.c | 4 ++--
net/bridge/br_vlan.c | 28 +++++++++++++++-------------
5 files changed, 49 insertions(+), 18 deletions(-)
diff --git a/net/bridge/br.c b/net/bridge/br.c
index b0a0b82e2d91..e411e40333e2 100644
--- a/net/bridge/br.c
+++ b/net/bridge/br.c
@@ -175,6 +175,22 @@ static struct notifier_block br_switchdev_notifier = {
.notifier_call = br_switchdev_event,
};
+void br_opt_toggle(struct net_bridge *br, enum net_bridge_opts opt, bool on)
+{
+ bool cur = !!br_opt_get(br, opt);
+
+ br_debug(br, "toggle option: %d state: %d -> %d\n",
+ opt, cur, on);
+
+ if (cur == on)
+ return;
+
+ if (on)
+ set_bit(opt, &br->options);
+ else
+ clear_bit(opt, &br->options);
+}
+
static void __net_exit br_net_exit(struct net *net)
{
struct net_device *dev;
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index ec2b58a09f76..6a53caff2d31 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -1416,7 +1416,8 @@ static int br_fill_info(struct sk_buff *skb, const struct
net_device *brdev)
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
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->vlan_stats_enabled))
+ nla_put_u8(skb, IFLA_BR_VLAN_STATS_ENABLED,
+ br_opt_get(br, BROPT_VLAN_STATS_ENABLED)))
return -EMSGSIZE;
#endif
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 522d707cc533..0abb632283ff 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -306,16 +306,20 @@ static inline struct net_bridge_port
*br_port_get_rtnl_rcu(const struct net_devi
rcu_dereference_rtnl(dev->rx_handler_data) : NULL;
}
+enum net_bridge_opts {
+ BROPT_VLAN_ENABLED,
+ BROPT_VLAN_STATS_ENABLED,
+};
+
struct net_bridge {
spinlock_t lock;
spinlock_t hash_lock;
struct list_head port_list;
struct net_device *dev;
struct pcpu_sw_netstats __percpu *stats;
+ unsigned long options;
/* These fields are accessed on each packet */
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
- u8 vlan_enabled;
- u8 vlan_stats_enabled;
__be16 vlan_proto;
u16 default_pvid;
struct net_bridge_vlan_group __rcu *vlgrp;
@@ -489,6 +493,14 @@ static inline bool br_vlan_should_use(const struct
net_bridge_vlan *v)
return true;
}
+static inline int br_opt_get(const struct net_bridge *br,
+ enum net_bridge_opts opt)
+{
+ return test_bit(opt, &br->options);
+}
+
+void br_opt_toggle(struct net_bridge *br, enum net_bridge_opts opt, bool on);
+
/* br_device.c */
void br_dev_setup(struct net_device *dev);
void br_dev_delete(struct net_device *dev, struct list_head *list);
diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c
index 0318a69888d4..9f1f3866c833 100644
--- a/net/bridge/br_sysfs_br.c
+++ b/net/bridge/br_sysfs_br.c
@@ -743,7 +743,7 @@ static ssize_t vlan_filtering_show(struct device *d,
char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%d\n", br->vlan_enabled);
+ return sprintf(buf, "%d\n", br_opt_get(br, BROPT_VLAN_ENABLED));
}
static ssize_t vlan_filtering_store(struct device *d,
@@ -791,7 +791,7 @@ static ssize_t vlan_stats_enabled_show(struct device *d,
char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%u\n", br->vlan_stats_enabled);
+ return sprintf(buf, "%u\n", br_opt_get(br,
BROPT_VLAN_STATS_ENABLED));
}
static ssize_t vlan_stats_enabled_store(struct device *d,
diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c
index bb6ba794864f..61d698bfe2b7 100644
--- a/net/bridge/br_vlan.c
+++ b/net/bridge/br_vlan.c
@@ -386,7 +386,7 @@ struct sk_buff *br_handle_vlan(struct net_bridge *br,
return NULL;
}
}
- if (br->vlan_stats_enabled) {
+ if (br_opt_get(br, BROPT_VLAN_STATS_ENABLED)) {
stats = this_cpu_ptr(v->stats);
u64_stats_update_begin(&stats->syncp);
stats->tx_bytes += skb->len;
@@ -475,14 +475,14 @@ static bool __allowed_ingress(const struct net_bridge *br,
skb->vlan_tci |= pvid;
/* if stats are disabled we can avoid the lookup */
- if (!br->vlan_stats_enabled)
+ if (!br_opt_get(br, BROPT_VLAN_STATS_ENABLED))
return true;
}
v = br_vlan_find(vg, *vid);
if (!v || !br_vlan_should_use(v))
goto drop;
- if (br->vlan_stats_enabled) {
+ if (br_opt_get(br, BROPT_VLAN_STATS_ENABLED)) {
stats = this_cpu_ptr(v->stats);
u64_stats_update_begin(&stats->syncp);
stats->rx_bytes += skb->len;
@@ -504,7 +504,7 @@ bool br_allowed_ingress(const struct net_bridge *br,
/* If VLAN filtering is disabled on the bridge, all packets are
* permitted.
*/
- if (!br->vlan_enabled) {
+ if (!br_opt_get(br, BROPT_VLAN_ENABLED)) {
BR_INPUT_SKB_CB(skb)->vlan_filtered = false;
return true;
}
@@ -538,7 +538,7 @@ bool br_should_learn(struct net_bridge_port *p, struct
sk_buff *skb, u16 *vid)
struct net_bridge *br = p->br;
/* If filtering was disabled at input, let it pass. */
- if (!br->vlan_enabled)
+ if (!br_opt_get(br, BROPT_VLAN_ENABLED))
return true;
vg = nbp_vlan_group_rcu(p);
@@ -699,7 +699,8 @@ static void recalculate_group_addr(struct net_bridge *br)
return;
spin_lock_bh(&br->lock);
- if (!br->vlan_enabled || br->vlan_proto == htons(ETH_P_8021Q)) {
+ if (!br_opt_get(br, BROPT_VLAN_ENABLED) ||
+ br->vlan_proto == htons(ETH_P_8021Q)) {
/* Bridge Group Address */
br->group_addr[5] = 0x00;
} else { /* vlan_enabled && ETH_P_8021AD */
@@ -712,7 +713,8 @@ static void recalculate_group_addr(struct net_bridge *br)
/* Must be protected by RTNL. */
void br_recalculate_fwd_mask(struct net_bridge *br)
{
- if (!br->vlan_enabled || br->vlan_proto == htons(ETH_P_8021Q))
+ if (!br_opt_get(br, BROPT_VLAN_ENABLED) ||
+ br->vlan_proto == htons(ETH_P_8021Q))
br->group_fwd_mask_required = BR_GROUPFWD_DEFAULT;
else /* vlan_enabled && ETH_P_8021AD */
br->group_fwd_mask_required = BR_GROUPFWD_8021AD &
@@ -729,14 +731,14 @@ int __br_vlan_filter_toggle(struct net_bridge *br,
unsigned long val)
};
int err;
- if (br->vlan_enabled == val)
+ if (br_opt_get(br, BROPT_VLAN_ENABLED) == !!val)
return 0;
err = switchdev_port_attr_set(br->dev, &attr);
if (err && err != -EOPNOTSUPP)
return err;
- br->vlan_enabled = val;
+ br_opt_toggle(br, BROPT_VLAN_ENABLED, !!val);
br_manage_promisc(br);
recalculate_group_addr(br);
br_recalculate_fwd_mask(br);
@@ -753,7 +755,7 @@ bool br_vlan_enabled(const struct net_device *dev)
{
struct net_bridge *br = netdev_priv(dev);
- return !!br->vlan_enabled;
+ return br_opt_get(br, BROPT_VLAN_ENABLED);
}
EXPORT_SYMBOL_GPL(br_vlan_enabled);
@@ -819,7 +821,7 @@ int br_vlan_set_stats(struct net_bridge *br, unsigned long
val)
switch (val) {
case 0:
case 1:
- br->vlan_stats_enabled = val;
+ br_opt_toggle(br, BROPT_VLAN_STATS_ENABLED, !!val);
break;
default:
return -EINVAL;
@@ -964,7 +966,7 @@ int br_vlan_set_default_pvid(struct net_bridge *br, unsigned
long val)
goto out;
/* Only allow default pvid change when filtering is disabled */
- if (br->vlan_enabled) {
+ if (br_opt_get(br, BROPT_VLAN_ENABLED)) {
pr_info_once("Please disable vlan filtering to change
default_pvid\n");
err = -EPERM;
goto out;
@@ -1018,7 +1020,7 @@ int nbp_vlan_init(struct net_bridge_port *p)
.orig_dev = p->br->dev,
.id = SWITCHDEV_ATTR_ID_BRIDGE_VLAN_FILTERING,
.flags = SWITCHDEV_F_SKIP_EOPNOTSUPP,
- .u.vlan_filtering = p->br->vlan_enabled,
+ .u.vlan_filtering = br_opt_get(p->br, BROPT_VLAN_ENABLED),
};
struct net_bridge_vlan_group *vg;
int ret = -ENOMEM;
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 3/9] net: bridge: convert nf call options to bits
No functional change, convert of nf_call_[ip|ip6|arp]tables to bits.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_netfilter_hooks.c | 7 ++++---
net/bridge/br_netlink.c | 12 ++++++------
net/bridge/br_private.h | 6 +++---
net/bridge/br_sysfs_br.c | 12 ++++++------
4 files changed, 19 insertions(+), 18 deletions(-)
diff --git a/net/bridge/br_netfilter_hooks.c b/net/bridge/br_netfilter_hooks.c
index 6e0dc6bcd32a..e0a3b038d052 100644
--- a/net/bridge/br_netfilter_hooks.c
+++ b/net/bridge/br_netfilter_hooks.c
@@ -487,14 +487,15 @@ static unsigned int br_nf_pre_routing(void *priv,
br = p->br;
if (IS_IPV6(skb) || IS_VLAN_IPV6(skb) || IS_PPPOE_IPV6(skb)) {
- if (!brnf_call_ip6tables && !br->nf_call_ip6tables)
+ if (!brnf_call_ip6tables &&
+ !br_opt_get(br, BROPT_NF_CALL_IP6TABLES))
return NF_ACCEPT;
nf_bridge_pull_encap_header_rcsum(skb);
return br_nf_pre_routing_ipv6(priv, skb, state);
}
- if (!brnf_call_iptables && !br->nf_call_iptables)
+ if (!brnf_call_iptables && !br_opt_get(br, BROPT_NF_CALL_IPTABLES))
return NF_ACCEPT;
if (!IS_IP(skb) && !IS_VLAN_IP(skb) && !IS_PPPOE_IP(skb))
@@ -636,7 +637,7 @@ static unsigned int br_nf_forward_arp(void *priv,
return NF_ACCEPT;
br = p->br;
- if (!brnf_call_arptables && !br->nf_call_arptables)
+ if (!brnf_call_arptables && !br_opt_get(br, BROPT_NF_CALL_ARPTABLES))
return NF_ACCEPT;
if (!IS_ARP(skb)) {
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index 6a53caff2d31..521893df0c2a 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -1271,19 +1271,19 @@ static int br_changelink(struct net_device *brdev,
struct nlattr *tb[],
if (data[IFLA_BR_NF_CALL_IPTABLES]) {
u8 val = nla_get_u8(data[IFLA_BR_NF_CALL_IPTABLES]);
- br->nf_call_iptables = val ? true : false;
+ br_opt_toggle(br, BROPT_NF_CALL_IPTABLES, !!val);
}
if (data[IFLA_BR_NF_CALL_IP6TABLES]) {
u8 val = nla_get_u8(data[IFLA_BR_NF_CALL_IP6TABLES]);
- br->nf_call_ip6tables = val ? true : false;
+ br_opt_toggle(br, BROPT_NF_CALL_IP6TABLES, !!val);
}
if (data[IFLA_BR_NF_CALL_ARPTABLES]) {
u8 val = nla_get_u8(data[IFLA_BR_NF_CALL_ARPTABLES]);
- br->nf_call_arptables = val ? true : false;
+ br_opt_toggle(br, BROPT_NF_CALL_ARPTABLES, !!val);
}
#endif
@@ -1470,11 +1470,11 @@ static int br_fill_info(struct sk_buff *skb, const
struct net_device *brdev)
#endif
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
if (nla_put_u8(skb, IFLA_BR_NF_CALL_IPTABLES,
- br->nf_call_iptables ? 1 : 0) ||
+ br_opt_get(br, BROPT_NF_CALL_IPTABLES) ? 1 : 0) ||
nla_put_u8(skb, IFLA_BR_NF_CALL_IP6TABLES,
- br->nf_call_ip6tables ? 1 : 0) ||
+ br_opt_get(br, BROPT_NF_CALL_IP6TABLES) ? 1 : 0) ||
nla_put_u8(skb, IFLA_BR_NF_CALL_ARPTABLES,
- br->nf_call_arptables ? 1 : 0))
+ br_opt_get(br, BROPT_NF_CALL_ARPTABLES) ? 1 : 0))
return -EMSGSIZE;
#endif
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 0abb632283ff..d8eadab09d7d 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -309,6 +309,9 @@ static inline struct net_bridge_port
*br_port_get_rtnl_rcu(const struct net_devi
enum net_bridge_opts {
BROPT_VLAN_ENABLED,
BROPT_VLAN_STATS_ENABLED,
+ BROPT_NF_CALL_IPTABLES,
+ BROPT_NF_CALL_IP6TABLES,
+ BROPT_NF_CALL_ARPTABLES,
};
struct net_bridge {
@@ -331,9 +334,6 @@ struct net_bridge {
struct rtable fake_rtable;
struct rt6_info fake_rt6_info;
};
- bool nf_call_iptables;
- bool nf_call_ip6tables;
- bool nf_call_arptables;
#endif
u16 group_fwd_mask;
u16 group_fwd_mask_required;
diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c
index 9f1f3866c833..0f8545c94948 100644
--- a/net/bridge/br_sysfs_br.c
+++ b/net/bridge/br_sysfs_br.c
@@ -678,12 +678,12 @@ static ssize_t nf_call_iptables_show(
struct device *d, struct device_attribute *attr, char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%u\n", br->nf_call_iptables);
+ return sprintf(buf, "%u\n", br_opt_get(br, BROPT_NF_CALL_IPTABLES));
}
static int set_nf_call_iptables(struct net_bridge *br, unsigned long val)
{
- br->nf_call_iptables = val ? true : false;
+ br_opt_toggle(br, BROPT_NF_CALL_IPTABLES, !!val);
return 0;
}
@@ -699,12 +699,12 @@ static ssize_t nf_call_ip6tables_show(
struct device *d, struct device_attribute *attr, char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%u\n", br->nf_call_ip6tables);
+ return sprintf(buf, "%u\n", br_opt_get(br,
BROPT_NF_CALL_IP6TABLES));
}
static int set_nf_call_ip6tables(struct net_bridge *br, unsigned long val)
{
- br->nf_call_ip6tables = val ? true : false;
+ br_opt_toggle(br, BROPT_NF_CALL_IP6TABLES, !!val);
return 0;
}
@@ -720,12 +720,12 @@ static ssize_t nf_call_arptables_show(
struct device *d, struct device_attribute *attr, char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%u\n", br->nf_call_arptables);
+ return sprintf(buf, "%u\n", br_opt_get(br,
BROPT_NF_CALL_ARPTABLES));
}
static int set_nf_call_arptables(struct net_bridge *br, unsigned long val)
{
- br->nf_call_arptables = val ? true : false;
+ br_opt_toggle(br, BROPT_NF_CALL_ARPTABLES, !!val);
return 0;
}
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 4/9] net: bridge: convert group_addr_set option to a bit
Convert group_addr_set internal bridge opt to a bit.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_netlink.c | 2 +-
net/bridge/br_private.h | 2 +-
net/bridge/br_sysfs_br.c | 2 +-
net/bridge/br_vlan.c | 2 +-
4 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index 521893df0c2a..b00ccf35c35d 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -1139,7 +1139,7 @@ static int br_changelink(struct net_device *brdev, struct
nlattr *tb[],
spin_lock_bh(&br->lock);
memcpy(br->group_addr, new_addr, sizeof(br->group_addr));
spin_unlock_bh(&br->lock);
- br->group_addr_set = true;
+ br_opt_toggle(br, BROPT_GROUP_ADDR_SET, true);
br_recalculate_fwd_mask(br);
}
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index d8eadab09d7d..dda117ce386e 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -312,6 +312,7 @@ enum net_bridge_opts {
BROPT_NF_CALL_IPTABLES,
BROPT_NF_CALL_IP6TABLES,
BROPT_NF_CALL_ARPTABLES,
+ BROPT_GROUP_ADDR_SET,
};
struct net_bridge {
@@ -355,7 +356,6 @@ struct net_bridge {
unsigned long bridge_ageing_time;
u8 group_addr[ETH_ALEN];
- bool group_addr_set;
enum {
BR_NO_STP, /* no spanning tree */
diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c
index 0f8545c94948..4c2d56354d89 100644
--- a/net/bridge/br_sysfs_br.c
+++ b/net/bridge/br_sysfs_br.c
@@ -303,7 +303,7 @@ static ssize_t group_addr_store(struct device *d,
ether_addr_copy(br->group_addr, new_addr);
spin_unlock_bh(&br->lock);
- br->group_addr_set = true;
+ br_opt_toggle(br, BROPT_GROUP_ADDR_SET, true);
br_recalculate_fwd_mask(br);
netdev_state_change(br->dev);
diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c
index 61d698bfe2b7..5942e03dd845 100644
--- a/net/bridge/br_vlan.c
+++ b/net/bridge/br_vlan.c
@@ -695,7 +695,7 @@ struct net_bridge_vlan *br_vlan_find(struct
net_bridge_vlan_group *vg, u16 vid)
/* Must be protected by RTNL. */
static void recalculate_group_addr(struct net_bridge *br)
{
- if (br->group_addr_set)
+ if (br_opt_get(br, BROPT_GROUP_ADDR_SET))
return;
spin_lock_bh(&br->lock);
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 5/9] net: bridge: convert and rename mcast disabled
Convert mcast disabled to an option bit and while doing so convert the
logic to check if multicast is enabled instead. That is make the logic
follow the option value - if it's set then mcast is enabled and vice versa.
This avoids a few confusing places where we inverted the value that's being
set to follow the mcast_disabled logic.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_mdb.c | 6 +++---
net/bridge/br_multicast.c | 32 +++++++++++++++++---------------
net/bridge/br_netlink.c | 3 ++-
net/bridge/br_private.h | 2 +-
net/bridge/br_sysfs_br.c | 2 +-
5 files changed, 24 insertions(+), 21 deletions(-)
diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c
index 6d9f48bd374a..a4a848bf827b 100644
--- a/net/bridge/br_mdb.c
+++ b/net/bridge/br_mdb.c
@@ -84,7 +84,7 @@ static int br_mdb_fill_info(struct sk_buff *skb, struct
netlink_callback *cb,
int i, err = 0;
int idx = 0, s_idx = cb->args[1];
- if (br->multicast_disabled)
+ if (!br_opt_get(br, BROPT_MULTICAST_ENABLED))
return 0;
mdb = rcu_dereference(br->mdb);
@@ -598,7 +598,7 @@ static int __br_mdb_add(struct net *net, struct net_bridge
*br,
struct net_bridge_port *p;
int ret;
- if (!netif_running(br->dev) || br->multicast_disabled)
+ if (!netif_running(br->dev) || !br_opt_get(br, BROPT_MULTICAST_ENABLED))
return -EINVAL;
dev = __dev_get_by_index(net, entry->ifindex);
@@ -673,7 +673,7 @@ static int __br_mdb_del(struct net_bridge *br, struct
br_mdb_entry *entry)
struct br_ip ip;
int err = -EINVAL;
- if (!netif_running(br->dev) || br->multicast_disabled)
+ if (!netif_running(br->dev) || !br_opt_get(br, BROPT_MULTICAST_ENABLED))
return -EINVAL;
__mdb_entry_to_br_ip(entry, &ip);
diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index 20ed7adcf1cc..ed6c33adc747 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -158,7 +158,7 @@ struct net_bridge_mdb_entry *br_mdb_get(struct net_bridge
*br,
struct net_bridge_mdb_htable *mdb = rcu_dereference(br->mdb);
struct br_ip ip;
- if (br->multicast_disabled)
+ if (!br_opt_get(br, BROPT_MULTICAST_ENABLED))
return NULL;
if (BR_INPUT_SKB_CB(skb)->igmp)
@@ -628,7 +628,7 @@ static struct net_bridge_mdb_entry *br_multicast_get_group(
port ? port->dev->name : br->dev->name);
err = -E2BIG;
disable:
- br->multicast_disabled = 1;
+ br_opt_toggle(br, BROPT_MULTICAST_ENABLED, false);
goto err;
}
}
@@ -894,7 +894,7 @@ static void br_multicast_querier_expired(struct net_bridge
*br,
struct bridge_mcast_own_query *query)
{
spin_lock(&br->multicast_lock);
- if (!netif_running(br->dev) || br->multicast_disabled)
+ if (!netif_running(br->dev) || !br_opt_get(br, BROPT_MULTICAST_ENABLED))
goto out;
br_multicast_start_querier(br, query);
@@ -965,7 +965,8 @@ static void br_multicast_send_query(struct net_bridge *br,
struct br_ip br_group;
unsigned long time;
- if (!netif_running(br->dev) || br->multicast_disabled ||
+ if (!netif_running(br->dev) ||
+ !br_opt_get(br, BROPT_MULTICAST_ENABLED) ||
!br->multicast_querier)
return;
@@ -1036,7 +1037,7 @@ static void br_mc_disabled_update(struct net_device *dev,
bool value)
.orig_dev = dev,
.id = SWITCHDEV_ATTR_ID_BRIDGE_MC_DISABLED,
.flags = SWITCHDEV_F_DEFER,
- .u.mc_disabled = value,
+ .u.mc_disabled = !value,
};
switchdev_port_attr_set(dev, &attr);
@@ -1054,7 +1055,8 @@ int br_multicast_add_port(struct net_bridge_port *port)
timer_setup(&port->ip6_own_query.timer,
br_ip6_multicast_port_query_expired, 0);
#endif
- br_mc_disabled_update(port->dev, port->br->multicast_disabled);
+ br_mc_disabled_update(port->dev,
+ br_opt_get(port->br, BROPT_MULTICAST_ENABLED));
port->mcast_stats = netdev_alloc_pcpu_stats(struct bridge_mcast_stats);
if (!port->mcast_stats)
@@ -1091,7 +1093,7 @@ static void __br_multicast_enable_port(struct
net_bridge_port *port)
{
struct net_bridge *br = port->br;
- if (br->multicast_disabled || !netif_running(br->dev))
+ if (!br_opt_get(br, BROPT_MULTICAST_ENABLED) || !netif_running(br->dev))
return;
br_multicast_enable(&port->ip4_own_query);
@@ -1904,7 +1906,7 @@ int br_multicast_rcv(struct net_bridge *br, struct
net_bridge_port *port,
BR_INPUT_SKB_CB(skb)->igmp = 0;
BR_INPUT_SKB_CB(skb)->mrouters_only = 0;
- if (br->multicast_disabled)
+ if (!br_opt_get(br, BROPT_MULTICAST_ENABLED))
return 0;
switch (skb->protocol) {
@@ -1998,7 +2000,7 @@ static void __br_multicast_open(struct net_bridge *br,
{
query->startup_sent = 0;
- if (br->multicast_disabled)
+ if (!br_opt_get(br, BROPT_MULTICAST_ENABLED))
return;
mod_timer(&query->timer, jiffies);
@@ -2173,12 +2175,12 @@ int br_multicast_toggle(struct net_bridge *br, unsigned
long val)
int err = 0;
spin_lock_bh(&br->multicast_lock);
- if (br->multicast_disabled == !val)
+ if (!!br_opt_get(br, BROPT_MULTICAST_ENABLED) == !!val)
goto unlock;
- br_mc_disabled_update(br->dev, !val);
- br->multicast_disabled = !val;
- if (br->multicast_disabled)
+ br_mc_disabled_update(br->dev, val);
+ br_opt_toggle(br, BROPT_MULTICAST_ENABLED, !!val);
+ if (!br_opt_get(br, BROPT_MULTICAST_ENABLED))
goto unlock;
if (!netif_running(br->dev))
@@ -2189,7 +2191,7 @@ int br_multicast_toggle(struct net_bridge *br, unsigned
long val)
if (mdb->old) {
err = -EEXIST;
rollback:
- br->multicast_disabled = !!val;
+ br_opt_toggle(br, BROPT_MULTICAST_ENABLED, false);
goto unlock;
}
@@ -2213,7 +2215,7 @@ bool br_multicast_enabled(const struct net_device *dev)
{
struct net_bridge *br = netdev_priv(dev);
- return !br->multicast_disabled;
+ return !!br_opt_get(br, BROPT_MULTICAST_ENABLED);
}
EXPORT_SYMBOL_GPL(br_multicast_enabled);
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index b00ccf35c35d..1d276b7b27e0 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -1422,7 +1422,8 @@ static int br_fill_info(struct sk_buff *skb, const struct
net_device *brdev)
#endif
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
if (nla_put_u8(skb, IFLA_BR_MCAST_ROUTER, br->multicast_router) ||
- nla_put_u8(skb, IFLA_BR_MCAST_SNOOPING, !br->multicast_disabled) ||
+ nla_put_u8(skb, IFLA_BR_MCAST_SNOOPING,
+ br_opt_get(br, BROPT_MULTICAST_ENABLED)) ||
nla_put_u8(skb, IFLA_BR_MCAST_QUERY_USE_IFADDR,
br->multicast_query_use_ifaddr) ||
nla_put_u8(skb, IFLA_BR_MCAST_QUERIER, br->multicast_querier) ||
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index dda117ce386e..fb1a266cdac8 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -313,6 +313,7 @@ enum net_bridge_opts {
BROPT_NF_CALL_IP6TABLES,
BROPT_NF_CALL_ARPTABLES,
BROPT_GROUP_ADDR_SET,
+ BROPT_MULTICAST_ENABLED,
};
struct net_bridge {
@@ -366,7 +367,6 @@ struct net_bridge {
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
unsigned char multicast_router;
- u8 multicast_disabled:1;
u8 multicast_querier:1;
u8 multicast_query_use_ifaddr:1;
u8 has_ipv6_addr:1;
diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c
index 4c2d56354d89..bf6f409a83e1 100644
--- a/net/bridge/br_sysfs_br.c
+++ b/net/bridge/br_sysfs_br.c
@@ -349,7 +349,7 @@ static ssize_t multicast_snooping_show(struct device *d,
char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%d\n", !br->multicast_disabled);
+ return sprintf(buf, "%d\n", br_opt_get(br,
BROPT_MULTICAST_ENABLED));
}
static ssize_t multicast_snooping_store(struct device *d,
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:17 UTC
[Bridge] [PATCH net-next 6/9] net: bridge: convert mcast options to bits
This patch converts the rest of the mcast options to bits. It also packs
the mcast options a little better by moving multicast_mld_version to an
existing hole, reducing the net_bridge size by 8 bytes.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_multicast.c | 22 ++++++++++------------
net/bridge/br_netlink.c | 11 ++++++-----
net/bridge/br_private.h | 20 ++++++++++----------
net/bridge/br_sysfs_br.c | 12 +++++++-----
4 files changed, 33 insertions(+), 32 deletions(-)
diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c
index ed6c33adc747..928024d8360d 100644
--- a/net/bridge/br_multicast.c
+++ b/net/bridge/br_multicast.c
@@ -411,7 +411,7 @@ static struct sk_buff *br_ip4_multicast_alloc_query(struct
net_bridge *br,
iph->frag_off = htons(IP_DF);
iph->ttl = 1;
iph->protocol = IPPROTO_IGMP;
- iph->saddr = br->multicast_query_use_ifaddr ?
+ iph->saddr = br_opt_get(br, BROPT_MULTICAST_QUERY_USE_IFADDR) ?
inet_select_addr(br->dev, 0, RT_SCOPE_LINK) : 0;
iph->daddr = htonl(INADDR_ALLHOSTS_GROUP);
((u8 *)&iph[1])[0] = IPOPT_RA;
@@ -503,11 +503,11 @@ static struct sk_buff *br_ip6_multicast_alloc_query(struct
net_bridge *br,
if (ipv6_dev_get_saddr(dev_net(br->dev), br->dev, &ip6h->daddr,
0,
&ip6h->saddr)) {
kfree_skb(skb);
- br->has_ipv6_addr = 0;
+ br_opt_toggle(br, BROPT_HAS_IPV6_ADDR, false);
return NULL;
}
- br->has_ipv6_addr = 1;
+ br_opt_toggle(br, BROPT_HAS_IPV6_ADDR, true);
ipv6_eth_mc_map(&ip6h->daddr, eth->h_dest);
hopopt = (u8 *)(ip6h + 1);
@@ -967,7 +967,7 @@ static void br_multicast_send_query(struct net_bridge *br,
if (!netif_running(br->dev) ||
!br_opt_get(br, BROPT_MULTICAST_ENABLED) ||
- !br->multicast_querier)
+ !br_opt_get(br, BROPT_MULTICAST_QUERIER))
return;
memset(&br_group.u, 0, sizeof(br_group.u));
@@ -1636,7 +1636,7 @@ br_multicast_leave_group(struct net_bridge *br,
if (timer_pending(&other_query->timer))
goto out;
- if (br->multicast_querier) {
+ if (br_opt_get(br, BROPT_MULTICAST_QUERIER)) {
__br_multicast_send_query(br, port, &mp->addr);
time = jiffies + br->multicast_last_member_count *
@@ -1748,7 +1748,7 @@ static void br_multicast_err_count(const struct net_bridge
*br,
struct bridge_mcast_stats __percpu *stats;
struct bridge_mcast_stats *pstats;
- if (!br->multicast_stats_enabled)
+ if (!br_opt_get(br, BROPT_MULTICAST_STATS_ENABLED))
return;
if (p)
@@ -1958,8 +1958,6 @@ void br_multicast_init(struct net_bridge *br)
br->hash_max = 512;
br->multicast_router = MDB_RTR_TYPE_TEMP_QUERY;
- br->multicast_querier = 0;
- br->multicast_query_use_ifaddr = 0;
br->multicast_last_member_count = 2;
br->multicast_startup_query_count = 2;
@@ -1978,7 +1976,7 @@ void br_multicast_init(struct net_bridge *br)
br->ip6_other_query.delay_time = 0;
br->ip6_querier.port = NULL;
#endif
- br->has_ipv6_addr = 1;
+ br_opt_toggle(br, BROPT_HAS_IPV6_ADDR, true);
spin_lock_init(&br->multicast_lock);
timer_setup(&br->multicast_router_timer,
@@ -2238,10 +2236,10 @@ int br_multicast_set_querier(struct net_bridge *br,
unsigned long val)
val = !!val;
spin_lock_bh(&br->multicast_lock);
- if (br->multicast_querier == val)
+ if (br_opt_get(br, BROPT_MULTICAST_QUERIER) == val)
goto unlock;
- br->multicast_querier = val;
+ br_opt_toggle(br, BROPT_MULTICAST_QUERIER, !!val);
if (!val)
goto unlock;
@@ -2562,7 +2560,7 @@ void br_multicast_count(struct net_bridge *br, const
struct net_bridge_port *p,
struct bridge_mcast_stats __percpu *stats;
/* if multicast_disabled is true then igmp type can't be set */
- if (!type || !br->multicast_stats_enabled)
+ if (!type || !br_opt_get(br, BROPT_MULTICAST_STATS_ENABLED))
return;
if (p)
diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index 1d276b7b27e0..e5a5bc5d5232 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -1167,7 +1167,7 @@ static int br_changelink(struct net_device *brdev, struct
nlattr *tb[],
u8 val;
val = nla_get_u8(data[IFLA_BR_MCAST_QUERY_USE_IFADDR]);
- br->multicast_query_use_ifaddr = !!val;
+ br_opt_toggle(br, BROPT_MULTICAST_QUERY_USE_IFADDR, !!val);
}
if (data[IFLA_BR_MCAST_QUERIER]) {
@@ -1244,7 +1244,7 @@ static int br_changelink(struct net_device *brdev, struct
nlattr *tb[],
__u8 mcast_stats;
mcast_stats = nla_get_u8(data[IFLA_BR_MCAST_STATS_ENABLED]);
- br->multicast_stats_enabled = !!mcast_stats;
+ br_opt_toggle(br, BROPT_MULTICAST_STATS_ENABLED, !!mcast_stats);
}
if (data[IFLA_BR_MCAST_IGMP_VERSION]) {
@@ -1425,10 +1425,11 @@ static int br_fill_info(struct sk_buff *skb, const
struct net_device *brdev)
nla_put_u8(skb, IFLA_BR_MCAST_SNOOPING,
br_opt_get(br, BROPT_MULTICAST_ENABLED)) ||
nla_put_u8(skb, IFLA_BR_MCAST_QUERY_USE_IFADDR,
- br->multicast_query_use_ifaddr) ||
- nla_put_u8(skb, IFLA_BR_MCAST_QUERIER, br->multicast_querier) ||
+ br_opt_get(br, BROPT_MULTICAST_QUERY_USE_IFADDR)) ||
+ nla_put_u8(skb, IFLA_BR_MCAST_QUERIER,
+ br_opt_get(br, BROPT_MULTICAST_QUERIER)) ||
nla_put_u8(skb, IFLA_BR_MCAST_STATS_ENABLED,
- br->multicast_stats_enabled) ||
+ br_opt_get(br, BROPT_MULTICAST_STATS_ENABLED)) ||
nla_put_u32(skb, IFLA_BR_MCAST_HASH_ELASTICITY,
br->hash_elasticity) ||
nla_put_u32(skb, IFLA_BR_MCAST_HASH_MAX, br->hash_max) ||
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index fb1a266cdac8..e6bdb107fe04 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -314,6 +314,10 @@ enum net_bridge_opts {
BROPT_NF_CALL_ARPTABLES,
BROPT_GROUP_ADDR_SET,
BROPT_MULTICAST_ENABLED,
+ BROPT_MULTICAST_QUERIER,
+ BROPT_MULTICAST_QUERY_USE_IFADDR,
+ BROPT_MULTICAST_STATS_ENABLED,
+ BROPT_HAS_IPV6_ADDR,
};
struct net_bridge {
@@ -365,12 +369,6 @@ struct net_bridge {
} stp_enabled;
#ifdef CONFIG_BRIDGE_IGMP_SNOOPING
- unsigned char multicast_router;
-
- u8 multicast_querier:1;
- u8 multicast_query_use_ifaddr:1;
- u8 has_ipv6_addr:1;
- u8 multicast_stats_enabled:1;
u32 hash_elasticity;
u32 hash_max;
@@ -379,7 +377,10 @@ struct net_bridge {
u32 multicast_startup_query_count;
u8 multicast_igmp_version;
-
+ u8 multicast_router;
+#if IS_ENABLED(CONFIG_IPV6)
+ u8 multicast_mld_version;
+#endif
unsigned long multicast_last_member_interval;
unsigned long multicast_membership_interval;
unsigned long multicast_querier_interval;
@@ -400,7 +401,6 @@ struct net_bridge {
struct bridge_mcast_other_query ip6_other_query;
struct bridge_mcast_own_query ip6_own_query;
struct bridge_mcast_querier ip6_querier;
- u8 multicast_mld_version;
#endif /* IS_ENABLED(CONFIG_IPV6) */
#endif
@@ -707,8 +707,8 @@ __br_multicast_querier_exists(struct net_bridge *br,
{
bool own_querier_enabled;
- if (br->multicast_querier) {
- if (is_ipv6 && !br->has_ipv6_addr)
+ if (br_opt_get(br, BROPT_MULTICAST_QUERIER)) {
+ if (is_ipv6 && !br_opt_get(br, BROPT_HAS_IPV6_ADDR))
own_querier_enabled = false;
else
own_querier_enabled = true;
diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c
index bf6f409a83e1..c93c5724609e 100644
--- a/net/bridge/br_sysfs_br.c
+++ b/net/bridge/br_sysfs_br.c
@@ -365,12 +365,13 @@ static ssize_t multicast_query_use_ifaddr_show(struct
device *d,
char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%d\n", br->multicast_query_use_ifaddr);
+ return sprintf(buf, "%d\n",
+ br_opt_get(br, BROPT_MULTICAST_QUERY_USE_IFADDR));
}
static int set_query_use_ifaddr(struct net_bridge *br, unsigned long val)
{
- br->multicast_query_use_ifaddr = !!val;
+ br_opt_toggle(br, BROPT_MULTICAST_QUERY_USE_IFADDR, !!val);
return 0;
}
@@ -388,7 +389,7 @@ static ssize_t multicast_querier_show(struct device *d,
char *buf)
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%d\n", br->multicast_querier);
+ return sprintf(buf, "%d\n", br_opt_get(br,
BROPT_MULTICAST_QUERIER));
}
static ssize_t multicast_querier_store(struct device *d,
@@ -636,12 +637,13 @@ static ssize_t multicast_stats_enabled_show(struct device
*d,
{
struct net_bridge *br = to_bridge(d);
- return sprintf(buf, "%u\n", br->multicast_stats_enabled);
+ return sprintf(buf, "%d\n",
+ br_opt_get(br, BROPT_MULTICAST_STATS_ENABLED));
}
static int set_stats_enabled(struct net_bridge *br, unsigned long val)
{
- br->multicast_stats_enabled = !!val;
+ br_opt_toggle(br, BROPT_MULTICAST_STATS_ENABLED, !!val);
return 0;
}
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:18 UTC
[Bridge] [PATCH net-next 7/9] net: bridge: convert neigh_suppress_enabled option to a bit
Convert the neigh_suppress_enabled option to a bit.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_arp_nd_proxy.c | 13 ++++++++-----
net/bridge/br_device.c | 4 ++--
net/bridge/br_input.c | 2 +-
net/bridge/br_private.h | 2 +-
4 files changed, 12 insertions(+), 9 deletions(-)
diff --git a/net/bridge/br_arp_nd_proxy.c b/net/bridge/br_arp_nd_proxy.c
index d42e3904b498..6b78e6351719 100644
--- a/net/bridge/br_arp_nd_proxy.c
+++ b/net/bridge/br_arp_nd_proxy.c
@@ -39,7 +39,7 @@ void br_recalculate_neigh_suppress_enabled(struct net_bridge
*br)
}
}
- br->neigh_suppress_enabled = neigh_suppress;
+ br_opt_toggle(br, BROPT_NEIGH_SUPPRESS_ENABLED, neigh_suppress);
}
#if IS_ENABLED(CONFIG_INET)
@@ -155,7 +155,7 @@ void br_do_proxy_suppress_arp(struct sk_buff *skb, struct
net_bridge *br,
ipv4_is_multicast(tip))
return;
- if (br->neigh_suppress_enabled) {
+ if (br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED)) {
if (p && (p->flags & BR_NEIGH_SUPPRESS))
return;
if (ipv4_is_zeronet(sip) || sip == tip) {
@@ -175,7 +175,8 @@ void br_do_proxy_suppress_arp(struct sk_buff *skb, struct
net_bridge *br,
return;
}
- if (br->neigh_suppress_enabled && br_is_local_ip(vlandev, tip)) {
+ if (br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED) &&
+ br_is_local_ip(vlandev, tip)) {
/* its our local ip, so don't proxy reply
* and don't forward to neigh suppress ports
*/
@@ -213,7 +214,8 @@ void br_do_proxy_suppress_arp(struct sk_buff *skb, struct
net_bridge *br,
/* If we have replied or as long as we know the
* mac, indicate to arp replied
*/
- if (replied || br->neigh_suppress_enabled)
+ if (replied ||
+ br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED))
BR_INPUT_SKB_CB(skb)->proxyarp_replied = true;
}
@@ -460,7 +462,8 @@ void br_do_suppress_nd(struct sk_buff *skb, struct
net_bridge *br,
* mac, indicate to NEIGH_SUPPRESS ports that we
* have replied
*/
- if (replied || br->neigh_suppress_enabled)
+ if (replied ||
+ br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED))
BR_INPUT_SKB_CB(skb)->proxyarp_replied = true;
}
neigh_release(n);
diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c
index e682a668ce57..a0afdca73d8f 100644
--- a/net/bridge/br_device.c
+++ b/net/bridge/br_device.c
@@ -67,11 +67,11 @@ netdev_tx_t br_dev_xmit(struct sk_buff *skb, struct
net_device *dev)
if (IS_ENABLED(CONFIG_INET) &&
(eth->h_proto == htons(ETH_P_ARP) ||
eth->h_proto == htons(ETH_P_RARP)) &&
- br->neigh_suppress_enabled) {
+ br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED)) {
br_do_proxy_suppress_arp(skb, br, vid, NULL);
} else if (IS_ENABLED(CONFIG_IPV6) &&
skb->protocol == htons(ETH_P_IPV6) &&
- br->neigh_suppress_enabled &&
+ br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED) &&
pskb_may_pull(skb, sizeof(struct ipv6hdr) +
sizeof(struct nd_msg)) &&
ipv6_hdr(skb)->nexthdr == IPPROTO_ICMPV6) {
diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c
index 72074276c088..3ddca11f44c2 100644
--- a/net/bridge/br_input.c
+++ b/net/bridge/br_input.c
@@ -122,7 +122,7 @@ int br_handle_frame_finish(struct net *net, struct sock *sk,
struct sk_buff *skb
br_do_proxy_suppress_arp(skb, br, vid, p);
} else if (IS_ENABLED(CONFIG_IPV6) &&
skb->protocol == htons(ETH_P_IPV6) &&
- br->neigh_suppress_enabled &&
+ br_opt_get(br, BROPT_NEIGH_SUPPRESS_ENABLED) &&
pskb_may_pull(skb, sizeof(struct ipv6hdr) +
sizeof(struct nd_msg)) &&
ipv6_hdr(skb)->nexthdr == IPPROTO_ICMPV6) {
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index e6bdb107fe04..6bdcc701bc71 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -318,6 +318,7 @@ enum net_bridge_opts {
BROPT_MULTICAST_QUERY_USE_IFADDR,
BROPT_MULTICAST_STATS_ENABLED,
BROPT_HAS_IPV6_ADDR,
+ BROPT_NEIGH_SUPPRESS_ENABLED,
};
struct net_bridge {
@@ -414,7 +415,6 @@ struct net_bridge {
#ifdef CONFIG_NET_SWITCHDEV
int offload_fwd_mark;
#endif
- bool neigh_suppress_enabled;
bool mtu_set_by_user;
struct hlist_head fdb_list;
};
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:18 UTC
[Bridge] [PATCH net-next 8/9] net: bridge: convert mtu_set_by_user to a bit
Convert the last remaining bool option to a bit thus reducing the overall
net_bridge size further by 8 bytes.
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_device.c | 2 +-
net/bridge/br_if.c | 4 ++--
net/bridge/br_private.h | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/bridge/br_device.c b/net/bridge/br_device.c
index a0afdca73d8f..e053a4e43758 100644
--- a/net/bridge/br_device.c
+++ b/net/bridge/br_device.c
@@ -228,7 +228,7 @@ static int br_change_mtu(struct net_device *dev, int
new_mtu)
dev->mtu = new_mtu;
/* this flag will be cleared if the MTU was automatically adjusted */
- br->mtu_set_by_user = true;
+ br_opt_toggle(br, BROPT_MTU_SET_BY_USER, true);
#if IS_ENABLED(CONFIG_BRIDGE_NETFILTER)
/* remember the MTU in the rtable for PMTU */
dst_metric_set(&br->fake_rtable.dst, RTAX_MTU, new_mtu);
diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
index 3bb66508f07d..9b46d2dc4c22 100644
--- a/net/bridge/br_if.c
+++ b/net/bridge/br_if.c
@@ -508,14 +508,14 @@ void br_mtu_auto_adjust(struct net_bridge *br)
ASSERT_RTNL();
/* if the bridge MTU was manually configured don't mess with it */
- if (br->mtu_set_by_user)
+ if (br_opt_get(br, BROPT_MTU_SET_BY_USER))
return;
/* change to the minimum MTU and clear the flag which was set by
* the bridge ndo_change_mtu callback
*/
dev_set_mtu(br->dev, br_mtu_min(br));
- br->mtu_set_by_user = false;
+ br_opt_toggle(br, BROPT_MTU_SET_BY_USER, false);
}
static void br_set_gso_limits(struct net_bridge *br)
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 6bdcc701bc71..92d08f8c7dd4 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -319,6 +319,7 @@ enum net_bridge_opts {
BROPT_MULTICAST_STATS_ENABLED,
BROPT_HAS_IPV6_ADDR,
BROPT_NEIGH_SUPPRESS_ENABLED,
+ BROPT_MTU_SET_BY_USER,
};
struct net_bridge {
@@ -415,7 +416,6 @@ struct net_bridge {
#ifdef CONFIG_NET_SWITCHDEV
int offload_fwd_mark;
#endif
- bool mtu_set_by_user;
struct hlist_head fdb_list;
};
--
2.11.0
Nikolay Aleksandrov
2018-Sep-26 12:18 UTC
[Bridge] [PATCH net-next 9/9] net: bridge: pack net_bridge better
Further reduce the size of net_bridge with 8 bytes and reduce the number of
holes in it:
Before: holes: 5, sum holes: 15
After: holes: 2, sum holes: 3
Also as a bonus we get offload_fwd_mark in the first cache line and remove
the need to pull the last cache line of net_bridge just for that member in
fast path (should_deliver -> nbp_switchdev_allowed_egress).
Signed-off-by: Nikolay Aleksandrov <nikolay at cumulusnetworks.com>
---
net/bridge/br_private.h | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h
index 92d08f8c7dd4..e50506225666 100644
--- a/net/bridge/br_private.h
+++ b/net/bridge/br_private.h
@@ -329,6 +329,9 @@ struct net_bridge {
struct net_device *dev;
struct pcpu_sw_netstats __percpu *stats;
unsigned long options;
+#ifdef CONFIG_NET_SWITCHDEV
+ int offload_fwd_mark;
+#endif
/* These fields are accessed on each packet */
#ifdef CONFIG_BRIDGE_VLAN_FILTERING
__be16 vlan_proto;
@@ -349,7 +352,6 @@ struct net_bridge {
/* STP */
bridge_id designated_root;
bridge_id bridge_id;
- u32 root_path_cost;
unsigned char topology_change;
unsigned char topology_change_detected;
u16 root_port;
@@ -361,6 +363,7 @@ struct net_bridge {
unsigned long bridge_hello_time;
unsigned long bridge_forward_delay;
unsigned long bridge_ageing_time;
+ u32 root_path_cost;
u8 group_addr[ETH_ALEN];
@@ -383,6 +386,7 @@ struct net_bridge {
#if IS_ENABLED(CONFIG_IPV6)
u8 multicast_mld_version;
#endif
+ spinlock_t multicast_lock;
unsigned long multicast_last_member_interval;
unsigned long multicast_membership_interval;
unsigned long multicast_querier_interval;
@@ -390,7 +394,6 @@ struct net_bridge {
unsigned long multicast_query_response_interval;
unsigned long multicast_startup_query_interval;
- spinlock_t multicast_lock;
struct net_bridge_mdb_htable __rcu *mdb;
struct hlist_head router_list;
@@ -411,12 +414,9 @@ struct net_bridge {
struct timer_list topology_change_timer;
struct delayed_work gc_work;
struct kobject *ifobj;
- u32 auto_cnt;
-#ifdef CONFIG_NET_SWITCHDEV
- int offload_fwd_mark;
-#endif
struct hlist_head fdb_list;
+ u32 auto_cnt;
};
struct br_input_skb_cb {
--
2.11.0
Stephen Hemminger
2018-Sep-26 12:52 UTC
[Bridge] [PATCH net-next 0/9] net: bridge: convert bool options to bits
On Wed, 26 Sep 2018 15:17:53 +0300 Nikolay Aleksandrov <nikolay at cumulusnetworks.com> wrote:> Hi, > A lot of boolean bridge options have been added around the net_bridge > structure resulting in holes and more importantly different cache lines > that need to be fetched in the fast path. This set moves all of those > to bits in a bitfield which resides in a hot cache line thus reducing > the size of net_bridge, the number of holes and the number of cache > lines needed for the fast path. > The set is also sent in preparation for new boolean options to avoid > spreading them in the structure and making new holes. > One nice side-effect is that we avoid potential race conditions by using > the bitops since some of the options were bits being directly set in > parallel risking hard to debug issues (has_ipv6_addr). > > Before: > size: 1184, holes: 8, sum holes: 30 > After: > size: 1160, holes: 2, sum holes: 3 > > Patch 01 is a trivial style fix > Patch 02 adds the new options bitfield and converts the vlan boolean > options to bits > Patches 03-08 convert the rest of the boolean options to bits > Patch 09 re-arranges a few fields in net_bridge to avoid fetching one > extra cache line in fast path and to further reduce size > > Thanks, > Nik > > > Nikolay Aleksandrov (9): > net: bridge: make struct opening bracket consistent > net: bridge: add bitfield for options and convert vlan opts > net: bridge: convert nf call options to bits > net: bridge: convert group_addr_set option to a bit > net: bridge: convert and rename mcast disabled > net: bridge: convert mcast options to bits > net: bridge: convert neigh_suppress_enabled option to a bit > net: bridge: convert mtu_set_by_user to a bit > net: bridge: pack net_bridge better > > net/bridge/br.c | 16 +++++++++ > net/bridge/br_arp_nd_proxy.c | 13 +++++--- > net/bridge/br_device.c | 6 ++-- > net/bridge/br_if.c | 4 +-- > net/bridge/br_input.c | 2 +- > net/bridge/br_mdb.c | 6 ++-- > net/bridge/br_multicast.c | 54 +++++++++++++++--------------- > net/bridge/br_netfilter_hooks.c | 7 ++-- > net/bridge/br_netlink.c | 31 +++++++++-------- > net/bridge/br_private.h | 74 +++++++++++++++++++++++------------------ > net/bridge/br_sysfs_br.c | 32 +++++++++--------- > net/bridge/br_vlan.c | 30 +++++++++-------- > 12 files changed, 155 insertions(+), 120 deletions(-) >Looks good, these flags grew over time and cleanup was overdue. Reviewed-by: Stephen Hemminger <stephen at networkplumber.org>