Hello,
I'm requesting this commit to be back-ported to v4.14:
---
commit 5b18f1289808fee5d04a7e6ecf200189f41a4db6
Author: Stephen Suryaputra <[email protected]>
Date: Wed Jun 26 02:21:16 2019 -0400
ipv4: reset rt_iif for recirculated mcast/bcast out pkts
Multicast or broadcast egress packets have rt_iif set to the oif. These
packets might be recirculated back as input and lookup to the raw
sockets may fail because they are bound to the incoming interface
(skb_iif). If rt_iif is not zero, during the lookup, inet_iif() function
returns rt_iif instead of skb_iif. Hence, the lookup fails.
v2: Make it non vrf specific (David Ahern). Reword the changelog to
reflect it.
Signed-off-by: Stephen Suryaputra <[email protected]>
Reviewed-by: David Ahern <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
---
We found the issue in that release and the above commit is on
linux-stable. On the discussion behind this commit, please see:
https://www.spinics.net/lists/netdev/msg581045.html
I think after the following diff is needed on top of the above commit
for v4.14:
---
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 4d85a4fdfdb0..ad2718c1624e 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1623,11 +1623,8 @@ struct rtable *rt_dst_clone(struct net_device *dev, struct rtable *rt)
new_rt->rt_iif = rt->rt_iif;
new_rt->rt_pmtu = rt->rt_pmtu;
new_rt->rt_mtu_locked = rt->rt_mtu_locked;
- new_rt->rt_gw_family = rt->rt_gw_family;
- if (rt->rt_gw_family == AF_INET)
- new_rt->rt_gw4 = rt->rt_gw4;
- else if (rt->rt_gw_family == AF_INET6)
- new_rt->rt_gw6 = rt->rt_gw6;
+ new_rt->rt_gateway = rt->rt_gateway;
+ new_rt->rt_table_id = rt->rt_table_id;
INIT_LIST_HEAD(&new_rt->rt_uncached);
new_rt->dst.flags |= DST_HOST;
---
Thank you,
Stephen.
On Mon, Jul 29, 2019 at 11:42:34AM -0400, Stephen Suryaputra wrote:
> Hello,
>
> I'm requesting this commit to be back-ported to v4.14:
> ---
> commit 5b18f1289808fee5d04a7e6ecf200189f41a4db6
> Author: Stephen Suryaputra <[email protected]>
> Date: Wed Jun 26 02:21:16 2019 -0400
>
> ipv4: reset rt_iif for recirculated mcast/bcast out pkts
>
> Multicast or broadcast egress packets have rt_iif set to the oif. These
> packets might be recirculated back as input and lookup to the raw
> sockets may fail because they are bound to the incoming interface
> (skb_iif). If rt_iif is not zero, during the lookup, inet_iif() function
> returns rt_iif instead of skb_iif. Hence, the lookup fails.
>
> v2: Make it non vrf specific (David Ahern). Reword the changelog to
> reflect it.
> Signed-off-by: Stephen Suryaputra <[email protected]>
> Reviewed-by: David Ahern <[email protected]>
> Signed-off-by: David S. Miller <[email protected]>
> ---
>
> We found the issue in that release and the above commit is on
> linux-stable. On the discussion behind this commit, please see:
> https://www.spinics.net/lists/netdev/msg581045.html
>
> I think after the following diff is needed on top of the above commit
> for v4.14:
>
> ---
> diff --git a/net/ipv4/route.c b/net/ipv4/route.c
> index 4d85a4fdfdb0..ad2718c1624e 100644
> --- a/net/ipv4/route.c
> +++ b/net/ipv4/route.c
> @@ -1623,11 +1623,8 @@ struct rtable *rt_dst_clone(struct net_device *dev, struct rtable *rt)
> new_rt->rt_iif = rt->rt_iif;
> new_rt->rt_pmtu = rt->rt_pmtu;
> new_rt->rt_mtu_locked = rt->rt_mtu_locked;
> - new_rt->rt_gw_family = rt->rt_gw_family;
> - if (rt->rt_gw_family == AF_INET)
> - new_rt->rt_gw4 = rt->rt_gw4;
> - else if (rt->rt_gw_family == AF_INET6)
> - new_rt->rt_gw6 = rt->rt_gw6;
> + new_rt->rt_gateway = rt->rt_gateway;
> + new_rt->rt_table_id = rt->rt_table_id;
> INIT_LIST_HEAD(&new_rt->rt_uncached);
>
> new_rt->dst.flags |= DST_HOST;
> ---
This only worked for 4.14, this also needs to go to 4.19. Can you
provide a working backport for both trees so that I can apply it? I'll
go drop the one I added, as it breaks the build :(
thanks,
greg k-h