ip: fix mc_loop checks for tunnels with multicast outer addresses
Subsystems (3)
PATCH 1/1 ip: fix mc_loop checks for tunnels with multicast outer addresses
When we have L3 tunnels with different inner/outer families (i.e. IPV4/IPV6) which use a multicast address as the outer tunnel destination address, multicast packets will be loopbacked back to the sending socket even if IP*_MULTICAST_LOOP is set to disabled.
The mc_loop flag is present in the family specific part of the socket (e.g. the IPv4 or IPv4 specific part). setsockopt sets the inner family mc_loop flag. When the packet is pushed through the L3 tunnel it will eventually be processed by the outer family which if different will check the flag in a different part of the socket then it was set.
---
This is a repost, original patch (and more details) in this thread:
http://marc.info/?l=linux-netdev&m=121010848426294&w=4
include/net/ip.h | 16 ++++++++++++++++ net/ipv4/ip_output.c | 2 +- net/ipv6/ip6_output.c | 3 +-- 3 files changed, 18 insertions(+), 3 deletions(-)
include/net/ip.h +16 -0
@@ -326,6 +326,22 @@ static __inline__ void inet_reset_saddr(struct sock *sk)#endif +static inline int sk_mc_loop (struct sock *sk) +{ +if (!sk) +return 1 ; +switch (sk->sk_family) { +case AF_INET : +return inet_sk (sk)->mc_loop; +#if defined(CONFIG_IPV6 ) || defined(CONFIG_IPV6_MODULE ) +case AF_INET6 : + return inet6_sk(sk)->mc_loop; + } +#endif +__WARN (); +return 1 ; +} +extern int ip_call_ra_chain (struct sk_buff *skb);/*
net/ipv4/ip_output.c +1 -1
@@ -254,7 +254,7 @@ int ip_mc_output(struct sk_buff *skb)*/ if (rt->rt_flags&RTCF_MULTICAST ) { - if ((!sk || inet_sk(sk)->mc_loop) + if (sk_mc_loop(sk)#ifdef CONFIG_IP_MROUTE /* Small optimization: do not loopback not local frames, which returned after forwarding; they will be dropped
net/ipv6/ip6_output.c +1 -2
@@ -121,10 +121,9 @@ static int ip6_output2(struct sk_buff *skb) skb->dev = dev; if (ipv6_addr_is_multicast(&ipv6_hdr(skb)->daddr)) { - struct ipv6_pinfo* np = skb->sk ? inet6_sk(skb->sk) : NULL; struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb)); - if (!(dev->flags & IFF_LOOPBACK) && (!np || np->mc_loop) && + if (!(dev->flags & IFF_LOOPBACK) && sk_mc_loop(skb->sk) && ((mroute6_socket(dev_net(dev)) && !(IP6CB(skb)->flags & IP6SKB_FORWARDED)) || ipv6_chk_mcast_addr(dev, &ipv6_hdr(skb)->daddr,
Signature · 1 line
1.5.6.5
David Miller
maintainerFrom: Octavian Purdila <[email protected]>
Date: Wed, 30 Dec 2009 22:01:37 +0200
include/net/ip.h static __inline__ void inet_reset_saddr(struct sock *sk) · line 344
Quoted 12 lines
include/net/ip.h +12
-0+switch (sk->sk_family) { +case AF_INET : +return inet_sk (sk)->mc_loop; +#if defined(CONFIG_IPV6 ) || defined(CONFIG_IPV6_MODULE ) +case AF_INET6 : + return inet6_sk(sk)->mc_loop; + } +#endif +__WARN (); +return 1 ; +} +
This won't compile with IPV6 disabled. Switch closing brace is erroneously inside of the ifdef test.
Quoted 19 lines · On Monday 04 January 2010 07:58:55 you wrote:
From: Octavian Purdila <[email protected]>
Date: Wed, 30 Dec 2009 22:01:37 +0200Review
Quoted 12 lines
(quoted diff) +12
-0+ switch (sk->sk_family) { + case AF_INET: + return inet_sk(sk)->mc_loop; +#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) + case AF_INET6: + return inet6_sk(sk)->mc_loop; + } +#endif + __WARN(); + return 1; +} +This won't compile with IPV6 disabled. Switch closing brace is erroneously inside of the ifdef test.
Oops... Here is a new version:
[PATCH] ip: fix mc_loop checks for tunnels with multicast outer addresses
When we have L3 tunnels with different inner/outer families (i.e. IPV4/IPV6) which use a multicast address as the outer tunnel destination address, multicast packets will be loopbacked back to the sending socket even if IP*_MULTICAST_LOOP is set to disabled.
The mc_loop flag is present in the family specific part of the socket (e.g. the IPv4 or IPv4 specific part). setsockopt sets the inner family mc_loop flag. When the packet is pushed through the L3 tunnel it will eventually be processed by the outer family which if different will check the flag in a different part of the socket then it was set.
--- include/net/ip.h | 16 ++++++++++++++++ net/ipv4/ip_output.c | 2 +- net/ipv6/ip6_output.c | 3 +-- 3 files changed, 18 insertions(+), 3 deletions(-)
include/net/ip.h +16 -0
@@ -326,6 +326,22 @@ static __inline__ void inet_reset_saddr(struct sock *sk)#endif +static inline int sk_mc_loop (struct sock *sk) +{ +if (!sk) +return 1 ; +switch (sk->sk_family) { +case AF_INET : +return inet_sk (sk)->mc_loop; +#if defined(CONFIG_IPV6 ) || defined(CONFIG_IPV6_MODULE ) +case AF_INET6 : +return inet6_sk (sk)->mc_loop; +#endif + } +__WARN (); +return 1 ; +} +extern int ip_call_ra_chain (struct sk_buff *skb);/*
net/ipv4/ip_output.c +1 -1
@@ -254,7 +254,7 @@ int ip_mc_output(struct sk_buff *skb)*/ if (rt->rt_flags&RTCF_MULTICAST ) { - if ((!sk || inet_sk(sk)->mc_loop) + if (sk_mc_loop(sk)#ifdef CONFIG_IP_MROUTE /* Small optimization: do not loopback not local frames, which returned after forwarding; they will be dropped
net/ipv6/ip6_output.c +1 -2
@@ -121,10 +121,9 @@ static int ip6_output2(struct sk_buff *skb) skb->dev = dev; if (ipv6_addr_is_multicast(&ipv6_hdr(skb)->daddr)) { - struct ipv6_pinfo* np = skb->sk ? inet6_sk(skb->sk) : NULL; struct inet6_dev *idev = ip6_dst_idev(skb_dst(skb)); - if (!(dev->flags & IFF_LOOPBACK) && (!np || np->mc_loop) && + if (!(dev->flags & IFF_LOOPBACK) && sk_mc_loop(skb->sk) && ((mroute6_socket(dev_net(dev)) && !(IP6CB(skb)->flags & IP6SKB_FORWARDED)) || ipv6_chk_mcast_addr(dev, &ipv6_hdr(skb)->daddr,
Signature · 1 line
1.5.6.5
From: Octavian Purdila <[email protected]>
Date: Mon, 4 Jan 2010 12:48:25 +0200
Quoted 15 lines · On Monday 04 January 2010 07:58:55 you wrote:
Quoted 4 lines
This won't compile with IPV6 disabled. Switch closing brace is erroneously inside of the ifdef test.
Oops... Here is a new version:
[PATCH] ip: fix
mc_loopchecks for tunnels with multicast outer addressesWhen we have L3 tunnels with different inner/outer families (i.e. IPV4/IPV6) which use a multicast address as the outer tunnel destination address, multicast packets will be loopbacked back to the sending socket even if IP*_MULTICAST_LOOP is set to disabled.
Quoted 7 lines
The
mc_loopflag is present in the family specific part of the socket (e.g. the IPv4 or IPv4 specific part). setsockopt sets the inner familymc_loopflag. When the packet is pushed through the L3 tunnel it will eventually be processed by the outer family which if different will check the flag in a different part of the socket then it was set.Signed-off-by Octavian Purdila <[email protected]>
Applied to net-2.6, thanks!