[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231207205939.GI50400@kernel.org>
Date: Thu, 7 Dec 2023 20:59:39 +0000
From: Simon Horman <horms@...nel.org>
To: Pedro Tammela <pctammela@...atatu.com>
Cc: netdev@...r.kernel.org, davem@...emloft.net, edumazet@...gle.com,
kuba@...nel.org, pabeni@...hat.com, jhs@...atatu.com,
xiyou.wangcong@...il.com, jiri@...nulli.us,
marcelo.leitner@...il.com, vladbu@...dia.com,
Jiri Pirko <jiri@...dia.com>
Subject: Re: [PATCH net-next v3 5/5] net/sched: cls_api: conditional
notification of events
On Wed, Dec 06, 2023 at 01:44:16PM -0300, Pedro Tammela wrote:
> As of today tc-filter/chain events are unconditionally built and sent to
> RTNLGRP_TC. As with the introduction of rtnl_notify_needed we can check
> before-hand if they are really needed. This will help to alleviate
> system pressure when filters are concurrently added without the rtnl
> lock as in tc-flower.
>
> Reviewed-by: Jiri Pirko <jiri@...dia.com>
> Signed-off-by: Pedro Tammela <pctammela@...atatu.com>
Thanks Pedro,
my comment below notwithstanding, this looks good to me.
Reviewed-by: Simon Horman <horms@...nel.org>
> ---
> net/sched/cls_api.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
> index 1976bd163986..123185907ebd 100644
> --- a/net/sched/cls_api.c
> +++ b/net/sched/cls_api.c
> @@ -2053,6 +2053,9 @@ static int tfilter_notify(struct net *net, struct sk_buff *oskb,
> u32 portid = oskb ? NETLINK_CB(oskb).portid : 0;
> int err = 0;
>
> + if (!unicast && !rtnl_notify_needed(net, n->nlmsg_flags, RTNLGRP_TC))
I don't feel strongly about this, but
as the above condition appears 3 times in this patch,
perhaps it could be a helper?
> + return 0;
> +
> skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
> if (!skb)
> return -ENOBUFS;
> @@ -2082,6 +2085,9 @@ static int tfilter_del_notify(struct net *net, struct sk_buff *oskb,
> u32 portid = oskb ? NETLINK_CB(oskb).portid : 0;
> int err;
>
> + if (!unicast && !rtnl_notify_needed(net, n->nlmsg_flags, RTNLGRP_TC))
> + return tp->ops->delete(tp, fh, last, rtnl_held, extack);
> +
> skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
> if (!skb)
> return -ENOBUFS;
> @@ -2906,6 +2912,9 @@ static int tc_chain_notify(struct tcf_chain *chain, struct sk_buff *oskb,
> struct sk_buff *skb;
> int err = 0;
>
> + if (!unicast && !rtnl_notify_needed(net, flags, RTNLGRP_TC))
> + return 0;
> +
> skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
> if (!skb)
> return -ENOBUFS;
> @@ -2935,6 +2944,9 @@ static int tc_chain_notify_delete(const struct tcf_proto_ops *tmplt_ops,
> struct net *net = block->net;
> struct sk_buff *skb;
>
> + if (!unicast && !rtnl_notify_needed(net, flags, RTNLGRP_TC))
> + return 0;
> +
> skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
> if (!skb)
> return -ENOBUFS;
> --
> 2.40.1
>
Powered by blists - more mailing lists