From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH net-next] team: Notify state change on ports Date: Thu, 12 Mar 2015 15:29:07 -0400 (EDT) Message-ID: <20150312.152907.1881565281406694368.davem@redhat.com> References: <1426176236-28255-1-git-send-email-jonas.johansson@gmail.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: jiri@resnulli.us, netdev@vger.kernel.org, ogerlitz@mellanox.com, monis@mellanox.com, jonas.johansson@westermo.se To: jonasj76@gmail.com Return-path: Received: from mx1.redhat.com ([209.132.183.28]:57296 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752614AbbCLT3S (ORCPT ); Thu, 12 Mar 2015 15:29:18 -0400 In-Reply-To: <1426176236-28255-1-git-send-email-jonas.johansson@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Jonas Johansson Date: Thu, 12 Mar 2015 17:03:56 +0100 > +static void team_notify_port_work(struct work_struct *work) > +{ > + struct team *team; > + struct team_port_notify *notify; > + > + team = container_of(work, struct team, notify_ports.dw.work); > + > + if (!rtnl_trylock()) { > + schedule_delayed_work(&team->notify_ports.dw, 0); > + return; > + } > + rcu_read_lock(); This endless retry logic has the same problems as the issue being worked on right now in the bonding notifiers. We need something more sane, therefore I suggest we wait for the bonding case to be worked out and then we can apply that solution here as well.