Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 2 Jun 2022 19:46:22 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     Kaarel =?UTF-8?B?UMOkcnRlbA==?= <kaarelp2rtel@gmail.com>
Cc:     Wolfgang Grandegger <wg@grandegger.com>,
        Marc Kleine-Budde <mkl@pengutronix.de>,
        "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Paolo Abeni <pabeni@redhat.com>, linux-can@vger.kernel.org,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] net: add operstate for vcan and dummy
Message-ID: <20220602194622.0c54a256@kernel.org>
In-Reply-To: <20220602081929.21929-1-kaarelp2rtel@gmail.com>
References: <20220602081929.21929-1-kaarelp2rtel@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu,  2 Jun 2022 11:19:29 +0300 Kaarel P=C3=A4rtel wrote:
> The idea here is simple. The vcan and the dummy network devices
> currently do not set the operational state of the interface.
> The result is that the interface state will be UNKNOWN.
>=20
> The kernel considers the unknown state to be the same as up:
> https://elixir.bootlin.com/linux/latest/source/include/linux/netdevice.h#=
L4125
>=20
> However for users this creates confusion:
> https://serverfault.com/questions/629676/dummy-network-interface-in-linux
>=20
> The change in this patch is very simple. When the interface is set up, the
> operational state is set to IF_OPER_UP.
>=20
> Signed-off-by: Kaarel P=C3=A4rtel <kaarelp2rtel@gmail.com>

You can change the carrier state from user space on a dummy device,
that will inform the kernel of the operstate:

# ip link add type dummy
# ip link show dev dummy0
8: dummy0: <BROADCAST,NOARP> mtu 1500 qdisc noop state DOWN mode DEFAULT gr=
oup default qlen 1000
    link/ether 8e:35:15:22:e3:d7 brd ff:ff:ff:ff:ff:ff
# ip link set dev dummy0 up
# ip link show dev dummy0
8: dummy0: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNO=
WN mode DEFAULT group default qlen 1000
    link/ether 8e:35:15:22:e3:d7 brd ff:ff:ff:ff:ff:ff
# ip link set dev dummy0 carrier off
# ip link set dev dummy0 carrier on
# ip link show dev dummy0
8: dummy0: <BROADCAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP mo=
de DEFAULT group default qlen 1000
    link/ether 8e:35:15:22:e3:d7 brd ff:ff:ff:ff:ff:ff


Flipping all soft devices which don't have a lower or don't expect user
space management to UP is fine but doing it one by one feels icky.
Yet another random thing a driver author has to know to flip.

If people are confused about seeing UNKNOWN in ip link output maybe we
should move displaying that under the -d flag (detailed output)?
Saves space, and nobody will get confused.