linux-kernel - Re: [PATCH net-next v2] net: dsa: make "label" property optional for dsa2

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:   Sun, 08 Jan 2017 21:56:19 -0500
From:   Vivien Didelot <vivien.didelot@...oirfairelinux.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        kernel@...oirfairelinux.com,
        "David S. Miller" <davem@...emloft.net>,
        Florian Fainelli <f.fainelli@...il.com>,
        Uwe Kleine-König <uwe@...ine-koenig.org>,
        Andrey Smirnov <andrew.smirnov@...il.com>,
        Jiri Pirko <jiri@...nulli.us>
Subject: Re: [PATCH net-next v2] net: dsa: make "label" property optional for dsa2

Hi Andrew,

Andrew Lunn <andrew@...n.ch> writes:

>> Until the printing of netdev_phys_item_id structures is fixed in
>> net/core/net-sysfs.c, an external helper can be used like this:
>
> As Florian pointed out, this cannot be changed. It is now part of the
> ABI. We have to live with it printing little endian numbers as big
> endian.

I totally understand the fact that ABI must not be changed. However we
should be aware that the current phys_switch_id of DSA is broken.

In addition to the minor issue of being hardly useable, it does not meet
the requirement described in the switchdev documentation of being unique
on a system. A switch ID in DSA is currently unique only to a DSA tree.
A system with two disjoint switch trees will have two switches with a
phys_switch_id of "00000000".

> Rather than recommending something, it might be better to point to the
> Free Desktop "Predictable Network Interface Names" which is what most
> people will end up with, if they rename:
>
> https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/
>
> It would also be good to test on a recent systemd system and see what
> happens. What names does it pick?

Note that the udev rules I gave in this commit message were only there
as examples of renaming DSA slave interfaces from userspace. This is
orthogonal with the purpose of this patch.

Thanks,

        Vivien