netdev - Re: [RFC PATCH net-next 0/3] ethtool: Add link down reason reporting

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20170622145347.4b5d7295@cakuba.netronome.com>
Date:   Thu, 22 Jun 2017 14:53:47 -0700
From:   Jakub Kicinski <kubakici@...pl>
To:     Gal Pressman <galp@...lanox.com>
Cc:     netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
        "John W. Linville" <linville@...driver.com>,
        Saeed Mahameed <saeedm@...lanox.com>,
        Vidya Sagar Ravipati <vidya@...ulusnetworks.com>,
        Jiri Pirko <jiri@...lanox.com>,
        David Decotigny <decot@...glers.com>, kernel-team@...com
Subject: Re: [RFC PATCH net-next 0/3] ethtool: Add link down reason
 reporting

On Thu, 22 Jun 2017 14:37:26 +0300, Gal Pressman wrote:
> > Any particular reason for implementing this ABI in ethtool rather than
> > via some netlink-based interface?  Devlink naturally comes to mind,
> > given that cabling problems are not really related to the L2 and netdev
> > shouldn't be required for diagnostics..  
> 
> ethtool is already used for reporting and handling of link related stuff through get/set_link_ksettings.
> How is reporting the link modes/port type/speed/etc different from this?
> This feature is only an extension of the already existent "link detected" field.
> 
> Implementing this ABI in devlink is a good idea, but it shouldn't be instead of ethtool but in addition to ethtool.
> Many users already use ethtool for this kind of info, we can tell them to use devlink to check why their link is down,
> but I think it's nicer to have it all in one place.

I understand where you're coming from, but it's a matter of user space
tooling.  If ethtool has to invoke devlink or know a little of netlink
to get this info, so be it.  Ethtool implements a lot of features, I'm
worried that if we select where things are added based on where the
similar features already live, we will be stuck adding ioctls for
ever :(

But I'm happy to let this go if others don't feel the same way ;)  

In general thanks for working on this feature, it's very useful!