> On May 22, 2019 at 1:59 PM Simon Horman <[email protected]> wrote: > > > On Mon, May 20, 2019 at 02:09:54PM +0200, Wolfram Sang wrote: > > > > > > > > > > How about the code below instead? > > > > > > > > > > > > > > > > if (netif_running(ndev)) > > > > > > > > ravb_close(ndev); > > > > > > > > > > > > > > > > ndev->mtu = new_mtu; > > > > > > > > netdev_update_features(ndev); > > > > > > > > > > > > > > Is there a need to call netdev_update_features() even if the if > > > > > > > is not > > > > > > > running? > > > > > > > > > > > > In my testing, it didn't seem so. > > > > > > > > > > That may be because your testing doesn't cover cases where it would > > > > > make > > > > > any difference. > > > > > > > > Cases other than changing the MTU while the device is up? > > > > > > I was thinking of cases where listeners are registered for the > > > notifier that netdev_update_features() triggers. > > > > Where are we here? Is this a blocker? > > I don't think this is a blocker but I would lean towards leaving > netdev_update_features() in unless we are certain its not needed. >
I have read through the code and have indeed not found any indication that netdev_update_features() or something equivalent is called implicitly when the device is closed and reopened. I'll send a v2. CU Uli
