Re: [PATCH v2 net] net: bridge: Do not offload IGMP/MLD messages

[Joseph Huang]

On 7/16/2025 2:26 AM, Ido Schimmel wrote:
> On Mon, Jul 14, 2025 at 11:01:00AM -0400, Joseph Huang wrote:
> > Do not offload IGMP/MLD messages as it could lead to IGMP/MLD Reports
> > being unintentionally flooded to Hosts. Instead, let the bridge decide
> > where to send these IGMP/MLD messages.
> >
> > Consider the case where the local host is sending out reports in response
> > to a remote querier like the following:
> >
> >         mcast-listener-process (IP_ADD_MEMBERSHIP)
> >            \
> >            br0
> >           /   \
> >        swp1   swp2
> >          |     |
> >    QUERIER     SOME-OTHER-HOST
> >
> > In the above setup, br0 will want to br_forward() reports for
> > mcast-listener-process's group(s) via swp1 to QUERIER; but since the
> > source hwdom is 0, the report is eligible for tx offloading, and is
> > flooded by hardware to both swp1 and swp2, reaching SOME-OTHER-HOST as
> > well. (Example and illustration provided by Tobias.)
> >
> > Fixes: 472111920f1c ("net: bridge: switchdev: allow the TX data plane forwarding to be offloaded")
> > Signed-off-by: Joseph Huang <Joseph.Huang@xxxxxxxxxx>
>
> I don't have personal experience with this offload, but it makes sense
> to not offload the replication of control packets to the underlying
> device and instead let the CPU handle it. These shouldn't be sent at an
> high rate anyway.
>
>
> I think you can just early return if the packet is IGMP/MLD. Something
> like:
>
> diff --git a/net/bridge/br_switchdev.c b/net/bridge/br_switchdev.c
> index 95d7355a0407..9a910cf0256e 100644
> --- a/net/bridge/br_switchdev.c
> +++ b/net/bridge/br_switchdev.c
> @@ -17,6 +17,9 @@ static bool nbp_switchdev_can_offload_tx_fwd(const struct net_bridge_port *p,
>   	if (!static_branch_unlikely(&br_switchdev_tx_fwd_offload))
>   		return false;
>   
> +	if (br_multicast_igmp_type(skb))
> +		return false;
> +
>   	return (p->flags & BR_TX_FWD_OFFLOAD) &&
>   	       (p->hwdom != BR_INPUT_SKB_CB(skb)->src_hwdom);
>   }

Talking about these packets being low rate, should I add unlikely() like so:

diff --git a/net/bridge/br_switchdev.c b/net/bridge/br_switchdev.c
index 95d7355a0407..9a910cf0256e 100644
--- a/net/bridge/br_switchdev.c
+++ b/net/bridge/br_switchdev.c
@@ -17,6 +17,9 @@ static bool nbp_switchdev_can_offload_tx_fwd(const 
struct net_bridge_port *p,
  	if (!static_branch_unlikely(&br_switchdev_tx_fwd_offload))
  		return false;

+	if (unlikely(br_multicast_igmp_type(skb)))
+		return false;
+
  	return (p->flags & BR_TX_FWD_OFFLOAD) &&
  	       (p->hwdom != BR_INPUT_SKB_CB(skb)->src_hwdom);
  }

Thanks,
Joseph