[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAM_iQpV7sVimr+3BzA8ZYW-wixwxv29r4jGZFB1FfUwCGN8QxA@mail.gmail.com>
Date: Fri, 27 Feb 2015 22:17:27 -0800
From: Cong Wang <xiyou.wangcong@...il.com>
To: Nicolas Dichtel <nicolas.dichtel@...nd.com>
Cc: Linux Kernel Network Developers <netdev@...r.kernel.org>
Subject: Re: [Patch net] netns: avoid allocating idr when dumping info
(Removing Eric from discussion since he expressed he doesn't
care this code at all in a private email.)
On Fri, Feb 27, 2015 at 4:56 PM, Cong Wang <xiyou.wangcong@...il.com> wrote:
>
> Why? This relies on rtnetlink to dump the link after creation,
> which is not obvious at all. IOW, imagine if rtnetlink didn't do
> the dump after new link, what would happen?
>
> You are hiding it too deep, without a valid reason. It sounds
> like you want to defer the allocation as late as possible,
> but again it makes no sense: 1) we don't allocate too much,
> just one id in idr; 2) this loses the code readability and hides
> potential bugs.
>
Here we go:
This breaks veth, where veth peers are completely equal,
therefore they are links for each other. So we expect
link-netnsid shown on both side.
Your current impl. relies on rtnl_configure_link() to dump
the link to allocate the netns id after creating a link.
Unfortunately for veth, rtnl_configure_link() is called
before veth->peer is set, therefore the dump of the "peer"
does _not_ actually allocate the netns id, since its
->get_link_net() returns the same net.
I will send a patch together after I figure out another bug
I saw.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists