[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Z9wuVAyFYcziY4-D@krikkit>
Date: Thu, 20 Mar 2025 16:03:48 +0100
From: Sabrina Dubroca <sd@...asysnail.net>
To: Paolo Abeni <pabeni@...hat.com>
Cc: netdev@...r.kernel.org, Andrew Lunn <andrew+netdev@...n.ch>,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>, Simon Horman <horms@...nel.org>
Subject: Re: [PATCH net-next v3] net: introduce per netns packet chains
2025-03-19, 17:24:03 +0100, Paolo Abeni wrote:
> Currently network taps unbound to any interface are linked in the
> global ptype_all list, affecting the performance in all the network
> namespaces.
>
> Add per netns ptypes chains, so that in the mentioned case only
> the netns owning the packet socket(s) is affected.
>
> While at that drop the global ptype_all list: no in kernel user
> registers a tap on "any" type without specifying either the target
> device or the target namespace (and IMHO doing that would not make
> any sense).
>
> Note that this adds a conditional in the fast path (to check for
> per netns ptype_specific list) and increases the dataset size by
> a cacheline (owing the per netns lists).
>
> Signed-off-by: Paolo Abeni <pabeni@...hat.com>
Reviewed-by: Sabrina Dubroca <sd@...asysnail.net>
Just two tiny nits:
> /**
> - * dev_nit_active - return true if any network interface taps are in use
> + * dev_nit_active_rcu - return true if any network interface taps are in use
> + *
> + * The caller must hold the RCU lock
> *
> * @dev: network device to check for the presence of taps
> */
> -bool dev_nit_active(struct net_device *dev)
> +bool dev_nit_active_rcu(struct net_device *dev)
I guess that would have been a good time to make dev const? (same in
the new wrapper)
> {
> - return !list_empty(&net_hotdata.ptype_all) ||
> + /* Callers may hold either RCU or RCU BH lock */
> + WARN_ON_ONCE(!rcu_read_lock_held() && !rcu_read_lock_bh_held());
> +
> + return !list_empty(&dev_net(dev)->ptype_all) ||
> !list_empty(&dev->ptype_all);
> }
[...]
> @@ -5830,6 +5851,14 @@ static int __netif_receive_skb_core(struct sk_buff **pskb, bool pfmemalloc,
> deliver_ptype_list_skb(skb, &pt_prev, orig_dev, type,
> &ptype_base[ntohs(type) &
> PTYPE_HASH_MASK]);
> +
> + /* orig_dev and skb->dev could belong to different netns;
> + * Even is such case we need to traverse only the list
I think there's a typo here: "Even in"
--
Sabrina
Powered by blists - more mailing lists