netdev - Re: [patch net-next rfc 3/7] net: rtnetlink: add commands to add and delete alternative ifnames

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 27 Aug 2019 08:14:49 -0700
From:   Roopa Prabhu <roopa@...ulusnetworks.com>
To:     Jiri Pirko <jiri@...nulli.us>
Cc:     David Miller <davem@...emloft.net>,
        Jakub Kicinski <jakub.kicinski@...ronome.com>,
        David Ahern <dsahern@...il.com>,
        netdev <netdev@...r.kernel.org>,
        Stephen Hemminger <sthemmin@...rosoft.com>, dcbw@...hat.com,
        Michal Kubecek <mkubecek@...e.cz>,
        Andrew Lunn <andrew@...n.ch>, parav@...lanox.com,
        Saeed Mahameed <saeedm@...lanox.com>,
        mlxsw <mlxsw@...lanox.com>
Subject: Re: [patch net-next rfc 3/7] net: rtnetlink: add commands to add and
 delete alternative ifnames

On Tue, Aug 27, 2019 at 2:35 AM Jiri Pirko <jiri@...nulli.us> wrote:
>
> Tue, Aug 27, 2019 at 10:22:42AM CEST, davem@...emloft.net wrote:
> >From: Jiri Pirko <jiri@...nulli.us>
> >Date: Tue, 27 Aug 2019 09:08:08 +0200
> >
> >> Okay, so if I understand correctly, on top of separate commands for
> >> add/del of alternative names, you suggest also get/dump to be separate
> >> command and don't fill this up in existing newling/getlink command.
> >
> >I'm not sure what to do yet.
> >
> >David has a point, because the only way these ifnames are useful is
> >as ways to specify and choose net devices.  So based upon that I'm
> >slightly learning towards not using separate commands.
>
> Well yeah, one can use it to handle existing commands instead of
> IFLA_NAME.
>
> But why does it rule out separate commands? I think it is cleaner than
> to put everything in poor setlink messages :/ The fact that we would
> need to add "OP" to the setlink message just feels of. Other similar
> needs may show up in the future and we may endup in ridiculous messages
> like:
>
> SETLINK
>   IFLA_NAME eth0
>   IFLA_ATLNAME_LIST (nest)
>       IFLA_ALTNAME_OP add
>       IFLA_ALTNAME somereallylongname
>       IFLA_ALTNAME_OP del
>       IFLA_ALTNAME somereallyreallylongname
>       IFLA_ALTNAME_OP add
>       IFLA_ALTNAME someotherreallylongname
>   IFLA_SOMETHING_ELSE_LIST (nest)
>       IFLA_SOMETHING_ELSE_OP add
>       ...
>       IFLA_SOMETHING_ELSE_OP del
>       ...
>       IFLA_SOMETHING_ELSE_OP add
>       ...
>
> I don't know what to think about it. Rollbacks are going to be pure hell :/

I don't see a huge problem with the above. We need a way to solve this
anyways for other list types in the future correct ?.
The approach taken by this series will not scale if we have to add a
new msg type and header for every such list attribute in the future.

A good parallel here is bridge vlan which uses RTM_SETLINK and
RTM_DELLINK for vlan add and deletes. But it does have an advantage of
a separate
msg space under AF_BRIDGE which makes it cleaner. Maybe something
closer to that  can be made to work (possibly with a msg flag) ?.

Would be good to have a consistent way to update list attributes for
future needs too.