Subject: Re: Should tap(4) and etherip(4) really clutter dmesg?
To: Nino Dehne <>
From: Quentin Garnier <>
List: current-users
Date: 02/27/2007 00:42:36
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Feb 26, 2007 at 11:23:15PM +0100, Nino Dehne wrote:
> I'm using 3 tap(4) instances on a system. When a tap(4) interface is
> created, the following kernel messages appear:
>    tap0: faking Ethernet device
>    tap0: Ethernet address f2:0b:a4:91:b1:0e
> Now I'm not sure I really need that information cluttering dmesg per
> default, since it provides no real information. I know that an
> Ethernet device is being faked by virtue of issuing the ifconfig create
> command. If that were to fail, ifconfig would be telling me. So I don't
> need to check the kernel messages. Also, I can query the resulting random
> MAC address via sysctl.
> One thing that I thought about was having these messages as a kind of
> audit logging in case the log-reading admin wasn't the one creating the
> interfaces. But then all interface creation should have a message logged.
> Thus I really think this should be debug output. Similarly, etherip(4)
> devices output the following:
>    etherip0: Ethernet address f2:0b:a5:18:83:0b
> It's a bit more terse, but nevertheless redundant IMHO. At the very
> least, they both should be consistent.
> Any opinions?

Well, the original idea of tap was to pretend to be an Ethernet device
and kind of serve as a basic skeleton for a driver.  It's grown a bit
outside that scope, though.

What you say makes perfect sense;  I'll remove the first line from tap
output and aprint_verbose-ify the other.

Quentin Garnier - -
"You could have made it, spitting out benchmarks
Owe it to yourself not to fail"
Amplifico, Spitting Out Benchmarks, Hometakes Vol. 2, 2005.

Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.4.6 (NetBSD)