Re: [PATCH v2 net] net: bridge: Do not offload IGMP/MLD messages
From: Ido Schimmel <idosch@nvidia.com>
Date: 2025-07-16 14:50:42
Also in:
bridge, lkml
On Wed, Jul 16, 2025 at 10:29:55AM -0400, Joseph Huang wrote:
Talking about these packets being low rate, should I add unlikely() like so:
I don't think it's needed and probably not measurable in most (all?) deployments. The entire thing is hidden behind static_branch_unlikely(). It just seemed weird to perform the checks about the Tx offload and only later check if it's even a packet for which you want this offload.
quoted hunk ↗ jump to hunk
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 structnet_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