[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200914092658.6224045f@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date: Mon, 14 Sep 2020 09:26:58 -0700
From: Jakub Kicinski <kuba@...nel.org>
To: Andrew Lunn <andrew@...n.ch>
Cc: Vladimir Oltean <olteanv@...il.com>, davem@...emloft.net,
netdev@...r.kernel.org, mkubecek@...e.cz,
michael.chan@...adcom.com, tariqt@...dia.com, saeedm@...dia.com,
alexander.duyck@...il.com
Subject: Re: [PATCH net-next v2 0/8] ethtool: add pause frame stats
On Mon, 14 Sep 2020 04:08:14 +0200 Andrew Lunn wrote:
> > DSA used to override the "ethtool -S" callback of the host port, and
> > append its own CPU port counters to that.
>
> That was always a hack. It was bound to break sooner or later.
>
> Ido planned to add statistics to devlink. I hope we can make use of
> that to replace the CPU port statistics, and also add DSA port
> statistics, since these interfaces do exist in devlink.
I considered devlink but it really doesn't make much sense to me to
configure something via ethtool and have its stats in devlink. If
devlink was the way to go then the config interface should have been
added there, too. And it wasn't (we just merged ethtool-nl for pause
a couple of releases ago). Besides, doesn't it go against our "Linux
is in control policy" to facilitate ports that don't have netdevs?
Especially making a precedent like this for completely symmetrical
pause frame config and stats does not seem like the right trade off
to me.
Powered by blists - more mailing lists