[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20170131.154922.106030849115718252.davem@davemloft.net>
Date: Tue, 31 Jan 2017 15:49:22 -0500 (EST)
From: David Miller <davem@...emloft.net>
To: dsa@...ulusnetworks.com
Cc: xiyou.wangcong@...il.com, theuns.verwoerd@...iedtelesis.co.nz,
netdev@...r.kernel.org
Subject: Re: [PATCH] rtnetlink: Handle IFLA_MASTER parameter when
processing rtnl_newlink
From: David Ahern <dsa@...ulusnetworks.com>
Date: Tue, 31 Jan 2017 13:25:32 -0700
> On 1/30/17 11:49 PM, Cong Wang wrote:
>> On Mon, Jan 30, 2017 at 3:23 PM, Theuns Verwoerd
>> <theuns.verwoerd@...iedtelesis.co.nz> wrote:
>>> @@ -2653,6 +2653,11 @@ static int rtnl_newlink(struct sk_buff *skb, struct nlmsghdr *nlh)
>>> if (err < 0)
>>> goto out_unregister;
>>> }
>>> + if (tb[IFLA_MASTER]) {
>>> + err = do_set_master(dev, nla_get_u32(tb[IFLA_MASTER]));
>>> + if (err)
>>> + goto out_unregister;
>>> + }
>>> out:
>>> if (link_net)
>>> put_net(link_net);
>>
>> Not sure if it is too late to call do_set_master() after
>> dev_change_net_namespace().
>>
>
> The master device index is relative to a namespace. If both are given then the namespace change must be done first.
Right and this is exactly how do_setlink() handles this.
Powered by blists - more mailing lists