[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <269b7f80-e11f-4b93-9b89-cc75293b0039@lunn.ch>
Date: Sat, 10 May 2025 15:47:47 +0200
From: Andrew Lunn <andrew@...n.ch>
To: Jay Vosburgh <jv@...sburgh.net>
Cc: tonghao@...aicloud.com, netdev@...r.kernel.org,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
Simon Horman <horms@...nel.org>, Jonathan Corbet <corbet@....net>,
Andrew Lunn <andrew+netdev@...n.ch>
Subject: Re: [PATCH net-next 1/4] net: bonding: add broadcast_neighbor option
for 802.3ad
> >+static netdev_tx_t bond_3ad_xor_xmit(struct sk_buff *skb,
> >+ struct net_device *dev)
> >+{
> >+ struct bonding *bond = netdev_priv(dev);
> >+ struct bond_up_slave *slaves;
> >+ struct slave *slave;
> >+
> >+ if (bond_should_broadcast_neighbor(bond, skb))
> >+ return bond_xmit_broadcast(skb, dev);
>
> I feel like there has to be a way to implement this that won't
> add two or three branches to the transmit fast path for every packet
> when this option is not enabled (which will be the vast majority of
> cases). I'm not sure what that would look like, needs some thought.
Maybe include/linux/jump_label.h. It is what netfilter uses to hook
into the IP stack at various points.
Andrew
Powered by blists - more mailing lists