Commit 7ba91ecb authored by Jesper Dangaard Brouer's avatar Jesper Dangaard Brouer Committed by David S. Miller

net: for rate-limited ICMP replies save one atomic operation

It is possible to avoid the atomic operation in icmp{v6,}_xmit_lock,
by checking the sysctl_icmp_msgs_per_sec ratelimit before these calls,
as pointed out by Eric Dumazet, but the BH disabled state must be correct.

The icmp_global_allow() call states it must be called with BH
disabled.  This protection was given by the calls icmp_xmit_lock and
icmpv6_xmit_lock.  Thus, split out local_bh_disable/enable from these
functions and maintain it explicitly at callers.
Suggested-by: default avatarEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: default avatarJesper Dangaard Brouer <brouer@redhat.com>
Acked-by: default avatarEric Dumazet <edumazet@google.com>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent c0303efe
...@@ -209,19 +209,17 @@ static struct sock *icmp_sk(struct net *net) ...@@ -209,19 +209,17 @@ static struct sock *icmp_sk(struct net *net)
return *this_cpu_ptr(net->ipv4.icmp_sk); return *this_cpu_ptr(net->ipv4.icmp_sk);
} }
/* Called with BH disabled */
static inline struct sock *icmp_xmit_lock(struct net *net) static inline struct sock *icmp_xmit_lock(struct net *net)
{ {
struct sock *sk; struct sock *sk;
local_bh_disable();
sk = icmp_sk(net); sk = icmp_sk(net);
if (unlikely(!spin_trylock(&sk->sk_lock.slock))) { if (unlikely(!spin_trylock(&sk->sk_lock.slock))) {
/* This can happen if the output path signals a /* This can happen if the output path signals a
* dst_link_failure() for an outgoing ICMP packet. * dst_link_failure() for an outgoing ICMP packet.
*/ */
local_bh_enable();
return NULL; return NULL;
} }
return sk; return sk;
...@@ -229,7 +227,7 @@ static inline struct sock *icmp_xmit_lock(struct net *net) ...@@ -229,7 +227,7 @@ static inline struct sock *icmp_xmit_lock(struct net *net)
static inline void icmp_xmit_unlock(struct sock *sk) static inline void icmp_xmit_unlock(struct sock *sk)
{ {
spin_unlock_bh(&sk->sk_lock.slock); spin_unlock(&sk->sk_lock.slock);
} }
int sysctl_icmp_msgs_per_sec __read_mostly = 1000; int sysctl_icmp_msgs_per_sec __read_mostly = 1000;
...@@ -417,14 +415,17 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb) ...@@ -417,14 +415,17 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
if (ip_options_echo(&icmp_param->replyopts.opt.opt, skb)) if (ip_options_echo(&icmp_param->replyopts.opt.opt, skb))
return; return;
sk = icmp_xmit_lock(net); /* Needed by both icmp_global_allow and icmp_xmit_lock */
if (!sk) local_bh_disable();
return;
inet = inet_sk(sk);
/* global icmp_msgs_per_sec */ /* global icmp_msgs_per_sec */
if (!icmpv4_global_allow(net, type, code)) if (!icmpv4_global_allow(net, type, code))
goto out_unlock; goto out_bh_enable;
sk = icmp_xmit_lock(net);
if (!sk)
goto out_bh_enable;
inet = inet_sk(sk);
icmp_param->data.icmph.checksum = 0; icmp_param->data.icmph.checksum = 0;
...@@ -459,6 +460,8 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb) ...@@ -459,6 +460,8 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
ip_rt_put(rt); ip_rt_put(rt);
out_unlock: out_unlock:
icmp_xmit_unlock(sk); icmp_xmit_unlock(sk);
out_bh_enable:
local_bh_enable();
} }
#ifdef CONFIG_IP_ROUTE_MULTIPATH #ifdef CONFIG_IP_ROUTE_MULTIPATH
...@@ -668,13 +671,16 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info) ...@@ -668,13 +671,16 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
} }
} }
sk = icmp_xmit_lock(net); /* Needed by both icmp_global_allow and icmp_xmit_lock */
if (!sk) local_bh_disable();
goto out;
/* Check global sysctl_icmp_msgs_per_sec ratelimit */ /* Check global sysctl_icmp_msgs_per_sec ratelimit */
if (!icmpv4_global_allow(net, type, code)) if (!icmpv4_global_allow(net, type, code))
goto out_unlock; goto out_bh_enable;
sk = icmp_xmit_lock(net);
if (!sk)
goto out_bh_enable;
/* /*
* Construct source address and options. * Construct source address and options.
...@@ -750,6 +756,8 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info) ...@@ -750,6 +756,8 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
ip_rt_put(rt); ip_rt_put(rt);
out_unlock: out_unlock:
icmp_xmit_unlock(sk); icmp_xmit_unlock(sk);
out_bh_enable:
local_bh_enable();
out:; out:;
} }
EXPORT_SYMBOL(icmp_send); EXPORT_SYMBOL(icmp_send);
......
...@@ -110,19 +110,17 @@ static const struct inet6_protocol icmpv6_protocol = { ...@@ -110,19 +110,17 @@ static const struct inet6_protocol icmpv6_protocol = {
.flags = INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL, .flags = INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
}; };
/* Called with BH disabled */
static __inline__ struct sock *icmpv6_xmit_lock(struct net *net) static __inline__ struct sock *icmpv6_xmit_lock(struct net *net)
{ {
struct sock *sk; struct sock *sk;
local_bh_disable();
sk = icmpv6_sk(net); sk = icmpv6_sk(net);
if (unlikely(!spin_trylock(&sk->sk_lock.slock))) { if (unlikely(!spin_trylock(&sk->sk_lock.slock))) {
/* This can happen if the output path (f.e. SIT or /* This can happen if the output path (f.e. SIT or
* ip6ip6 tunnel) signals dst_link_failure() for an * ip6ip6 tunnel) signals dst_link_failure() for an
* outgoing ICMP6 packet. * outgoing ICMP6 packet.
*/ */
local_bh_enable();
return NULL; return NULL;
} }
return sk; return sk;
...@@ -130,7 +128,7 @@ static __inline__ struct sock *icmpv6_xmit_lock(struct net *net) ...@@ -130,7 +128,7 @@ static __inline__ struct sock *icmpv6_xmit_lock(struct net *net)
static __inline__ void icmpv6_xmit_unlock(struct sock *sk) static __inline__ void icmpv6_xmit_unlock(struct sock *sk)
{ {
spin_unlock_bh(&sk->sk_lock.slock); spin_unlock(&sk->sk_lock.slock);
} }
/* /*
...@@ -489,6 +487,13 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info, ...@@ -489,6 +487,13 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
return; return;
} }
/* Needed by both icmp_global_allow and icmpv6_xmit_lock */
local_bh_disable();
/* Check global sysctl_icmp_msgs_per_sec ratelimit */
if (!icmpv6_global_allow(type))
goto out_bh_enable;
mip6_addr_swap(skb); mip6_addr_swap(skb);
memset(&fl6, 0, sizeof(fl6)); memset(&fl6, 0, sizeof(fl6));
...@@ -507,10 +512,7 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info, ...@@ -507,10 +512,7 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
sk = icmpv6_xmit_lock(net); sk = icmpv6_xmit_lock(net);
if (!sk) if (!sk)
return; goto out_bh_enable;
if (!icmpv6_global_allow(type))
goto out;
sk->sk_mark = mark; sk->sk_mark = mark;
np = inet6_sk(sk); np = inet6_sk(sk);
...@@ -571,6 +573,8 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info, ...@@ -571,6 +573,8 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
dst_release(dst); dst_release(dst);
out: out:
icmpv6_xmit_unlock(sk); icmpv6_xmit_unlock(sk);
out_bh_enable:
local_bh_enable();
} }
/* Slightly more convenient version of icmp6_send. /* Slightly more convenient version of icmp6_send.
...@@ -684,9 +688,10 @@ static void icmpv6_echo_reply(struct sk_buff *skb) ...@@ -684,9 +688,10 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
fl6.flowi6_uid = sock_net_uid(net, NULL); fl6.flowi6_uid = sock_net_uid(net, NULL);
security_skb_classify_flow(skb, flowi6_to_flowi(&fl6)); security_skb_classify_flow(skb, flowi6_to_flowi(&fl6));
local_bh_disable();
sk = icmpv6_xmit_lock(net); sk = icmpv6_xmit_lock(net);
if (!sk) if (!sk)
return; goto out_bh_enable;
sk->sk_mark = mark; sk->sk_mark = mark;
np = inet6_sk(sk); np = inet6_sk(sk);
...@@ -728,6 +733,8 @@ static void icmpv6_echo_reply(struct sk_buff *skb) ...@@ -728,6 +733,8 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
dst_release(dst); dst_release(dst);
out: out:
icmpv6_xmit_unlock(sk); icmpv6_xmit_unlock(sk);
out_bh_enable:
local_bh_enable();
} }
void icmpv6_notify(struct sk_buff *skb, u8 type, u8 code, __be32 info) void icmpv6_notify(struct sk_buff *skb, u8 type, u8 code, __be32 info)
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment