[PATCH] net: igmp: Use ingress interface rather than vrf device

[PATCH] net: igmp: Use ingress interface rather than vrf device

[David Ahern]

Anuradha reported that statically added groups for interfaces enslaved
to a VRF device were not persisting. The problem is that igmp queries
and reports need to use the data in the in_dev for the real ingress
device rather than the VRF device. Update igmp_rcv accordingly.

Fixes: e58e41596811 ("net: Enable support for VRF with ipv4 multicast")
Reported-by: Anuradha Karuppiah <anuradhak@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
---
 net/ipv4/igmp.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index 9f86b5133605..ab183af0b5b6 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -1007,10 +1007,18 @@ int igmp_rcv(struct sk_buff *skb)
 {
 	/* This basically follows the spec line by line -- see RFC1112 */
 	struct igmphdr *ih;
-	struct in_device *in_dev = __in_dev_get_rcu(skb->dev);
+	struct net_device *dev = skb->dev;
+	struct in_device *in_dev;
 	int len = skb->len;
 	bool dropped = true;
 
+	if (netif_is_l3_master(dev)) {
+		dev = dev_get_by_index_rcu(dev_net(dev), IPCB(skb)->iif);
+		if (!dev)
+			goto drop;
+	}
+
+	in_dev = __in_dev_get_rcu(dev);
 	if (!in_dev)
 		goto drop;
 
-- 
2.1.4

Re: [PATCH] net: igmp: Use ingress interface rather than vrf device

[Nikolay Aleksandrov]

On 16/08/17 04:38, David Ahern wrote:
> Anuradha reported that statically added groups for interfaces enslaved
> to a VRF device were not persisting. The problem is that igmp queries
> and reports need to use the data in the in_dev for the real ingress
> device rather than the VRF device. Update igmp_rcv accordingly.
> 
> Fixes: e58e41596811 ("net: Enable support for VRF with ipv4 multicast")
> Reported-by: Anuradha Karuppiah <anuradhak@cumulusnetworks.com>
> Signed-off-by: David Ahern <dsahern@gmail.com>
> ---
>  net/ipv4/igmp.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
> index 9f86b5133605..ab183af0b5b6 100644
> --- a/net/ipv4/igmp.c
> +++ b/net/ipv4/igmp.c
> @@ -1007,10 +1007,18 @@ int igmp_rcv(struct sk_buff *skb)
>  {
>  	/* This basically follows the spec line by line -- see RFC1112 */
>  	struct igmphdr *ih;
> -	struct in_device *in_dev = __in_dev_get_rcu(skb->dev);
> +	struct net_device *dev = skb->dev;
> +	struct in_device *in_dev;
>  	int len = skb->len;
>  	bool dropped = true;
>  
> +	if (netif_is_l3_master(dev)) {
> +		dev = dev_get_by_index_rcu(dev_net(dev), IPCB(skb)->iif);
> +		if (!dev)
> +			goto drop;
> +	}
> +
> +	in_dev = __in_dev_get_rcu(dev);
>  	if (!in_dev)
>  		goto drop;
>  
> 

Reviewed-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>

Re: [PATCH] net: igmp: Use ingress interface rather than vrf device

[David Miller]

From: David Ahern <dsahern@gmail.com>
Date: Tue, 15 Aug 2017 18:38:42 -0700

> Anuradha reported that statically added groups for interfaces enslaved
> to a VRF device were not persisting. The problem is that igmp queries
> and reports need to use the data in the in_dev for the real ingress
> device rather than the VRF device. Update igmp_rcv accordingly.
> 
> Fixes: e58e41596811 ("net: Enable support for VRF with ipv4 multicast")
> Reported-by: Anuradha Karuppiah <anuradhak@cumulusnetworks.com>
> Signed-off-by: David Ahern <dsahern@gmail.com>

Applied and queued up for -stable, thanks.