> On June 6, 2019 at 4:08 AM David Miller <davem@xxxxxxxxxxxxx> wrote:
>
>
> From: Ulrich Hecht <uli+renesas@xxxxxxxx>
> Date: Wed, 5 Jun 2019 17:14:20 +0200
>
> > @@ -1811,11 +1811,14 @@ static int ravb_do_ioctl(struct net_device *ndev, struct ifreq *req, int cmd)
> > static int ravb_change_mtu(struct net_device *ndev, int new_mtu)
> > {
> > if (netif_running(ndev))
> > - return -EBUSY;
> > + ravb_close(ndev);
> >
> > ndev->mtu = new_mtu;
> > netdev_update_features(ndev);
> >
> > + if (netif_running(ndev))
> > + return ravb_open(ndev);
> > +
>
> And if ravb_open() fails? The user sees a failure, but to the user the failure
> means the MTU change can't be done, yet the device has the new MTU set still.
>
> This really is terrible behavior.
>
> If you must do a prepare/commit kind of sequence for this to work properly if
> you are going to go down the road of taking the device down to change the MTU
> when the device is UP.
So would rolling back the MTU change in case of a re-open failure be acceptable?
If so, is there additional action that needs to be taken if a device unexpectedly goes down as a side-effect of an MTU change?
CU
Uli
