[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMpxmJXbwrE-X7zsnNy0DzmhWADR0GRXZy_RFK4RuOnCv=p7OA@mail.gmail.com>
Date: Fri, 25 Sep 2020 11:43:55 +0200
From: Bartosz Golaszewski <bgolaszewski@...libre.com>
To: "Liu, Yongxin" <Yongxin.Liu@...driver.com>
Cc: "David S . Miller" <davem@...emloft.net>,
netdev <netdev@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] Revert "net: ethernet: ixgbe: check the return value of ixgbe_mii_bus_init()"
On Fri, Sep 25, 2020 at 10:51 AM Liu, Yongxin <Yongxin.Liu@...driver.com> wrote:
>
[snip]
> > > true);
> > >
> > > - err = ixgbe_mii_bus_init(hw);
> > > - if (err)
> > > - goto err_netdev;
> > > + ixgbe_mii_bus_init(hw);
> > >
> > > return 0;
> > >
> > > -err_netdev:
> > > - unregister_netdev(netdev);
> > > err_register:
> > > ixgbe_release_hw_control(adapter);
> > > ixgbe_clear_interrupt_scheme(adapter);
> > > --
> > > 2.14.4
> > >
> >
> > Then we should check if err == -ENODEV, not outright ignore all potential
> > errors, right?
> >
>
> Hm, it is weird to take -ENODEV as a no error.
> How about just return 0 instead of -ENODEV in the following function?
>
No, it's perfectly fine. -ENODEV means there's no device and this can
happen. The caller can then act accordingly - for example: ignore that
fact. We do it in several places[1].
Bartosz
[snip]
[1] https://elixir.bootlin.com/linux/latest/source/drivers/misc/eeprom/at24.c#L714
Powered by blists - more mailing lists