[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YZYoKuvrZmHEiQGx@Laptop-X1>
Date: Thu, 18 Nov 2021 18:17:14 +0800
From: Hangbin Liu <liuhangbin@...il.com>
To: Eric Dumazet <edumazet@...gle.com>
Cc: netdev@...r.kernel.org, Jay Vosburgh <jay.vosburgh@...onical.com>,
Veaceslav Falico <vfalico@...il.com>,
Andy Gospodarek <andy@...yhouse.net>,
Jarod Wilson <jarod@...hat.com>,
Jakub Kicinski <kuba@...nel.org>,
Jiri Pirko <jiri@...nulli.us>, davem@...emloft.net,
Denis Kirjanov <dkirjanov@...e.de>,
David Ahern <dsahern@...il.com>
Subject: Re: [DISCUSS] Bond arp monitor not works with veth due to flag
NETIF_F_LLTX.
On Wed, Nov 17, 2021 at 07:34:40PM -0800, Eric Dumazet wrote:
> Removing LLTX will have performance impact.
Yes, I also think so.
>
> Updating ->trans_start at most once per jiffy should be ok.
>
> Here is a patch against net-next
> @@ -4626,7 +4620,7 @@ static inline netdev_tx_t
> netdev_start_xmit(struct sk_buff *skb, struct net_devi
>
> rc = __netdev_start_xmit(ops, skb, dev, more);
> if (rc == NETDEV_TX_OK)
> - txq_trans_update(txq);
> + txq_trans_cond_update(txq);
>
> return rc;
> }
Awesome. This should resolve the veth's trans_start issue. Where is the patch?
I only find out the following one but could not find the fix in netdev_start_xmit()
https://patchwork.kernel.org/project/netdevbpf/patch/20211117032924.1740327-3-eric.dumazet@gmail.com/
Thanks
Hangbin
Powered by blists - more mailing lists