[<prev] [next>] [day] [month] [year] [list]
Message-ID: <b978cadc-935e-8086-0389-60068677843e@gmail.com>
Date: Mon, 28 Sep 2020 20:06:32 -0700
From: David Ahern <dsahern@...il.com>
To: Qingtao Cao <qingtao.cao.au@...il.com>
Cc: netdev@...r.kernel.org, davem@...emloft.net, kuba@...nel.org,
kuznet@....inr.ac.ru, yoshfuji@...ux-ipv6.org, corbet@....net,
Qingtao Cao <qingtao.cao@...i.com>,
David Leonard <david.leonard@...i.com>
Subject: Re: [PATCH 1/1] Network: support default route metric per interface
On 9/28/20 6:54 PM, Qingtao Cao wrote:
> Hi David,
>
> I understand I can use the "metric" parameter along with the "ip addr
> add/change" and "ip route add/change" commands to make use of the
> IFA_RT_PRIORITY attribute to explicitly specify the metric for routes
> created directly or associated with added addresses, and the network
> manager should do this.
>
> But what if userspace can bypass the network manager? or forgetting to
> apply the "metric" parameter for example directly on the CLI?
use a sticky note? :-)
>
> Then the kernel will fall back on the same default metric (such as 0 for
> ipv4 and IP6_RT_PRIO_ADDRCONF(256) or USER(1024) for ipv6) for ALL
> interfaces. So I think the def_rt_metric patchset can be regarded as
> having the kernel providing a mechanism for the uerspace to specify
> DIFFERENT default routes per interface, which does not necessarily
> indulge the userspace network manager skipping over specifying the metic
> explicitly
>
> What do you think?
>
I think it is a userspace, configuration management problem.
Powered by blists - more mailing lists