[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190115073647.GC26670@unicorn.suse.cz>
Date: Tue, 15 Jan 2019 08:36:47 +0100
From: Michal Kubecek <mkubecek@...e.cz>
To: Jakub Kicinski <jakub.kicinski@...ronome.com>
Cc: Andrew Lunn <andrew@...n.ch>, davem@...emloft.net,
netdev@...r.kernel.org, oss-drivers@...ronome.com, jiri@...nulli.us
Subject: Re: [RFC net-next 0/6] devlink: add device (driver) information API
On Mon, Jan 14, 2019 at 07:27:34PM -0800, Jakub Kicinski wrote:
> On Tue, 15 Jan 2019 02:57:55 +0100, Andrew Lunn wrote:
> >
> > Using that argument, you should probably make the devlink core call
> > the ethtool .get_drvinfo op if the device does not implement the
> > devlink op.
>
> Could it possibly be done in user space? Have iproute2/devlink call old
> ethtool API on first netdev associated with one of the ports?
I suppose it could, as long as there is an interface to find the
corresponding netdev name (or ifindex). Some time ago I tried to find
a way to get a devlink handle for a given net device and I couldn't find
any straightforward way even in kernel. (I needed the opposite direction
as I was investigating the options for implementing ethtool dump/flash
commands using devlink regions.)
Michal Kubecek
Powered by blists - more mailing lists