Re: [PATCH v2] mac80211: Supporting of IFLA_INFO_KIND rtnl attribute

[vadim4j@gmail.com]

On Mon, Dec 08, 2014 at 11:17:03AM +0100, Johannes Berg wrote:
> On Mon, 2014-12-08 at 12:07 +0200, Vadim Kochan wrote:
> > You meant here ?
> > 
> > net/wireless/core.c : static int cfg80211_netdev_notifier_call(struct
> > notifier_block *nb,
> >                                           unsigned long state, void *ptr)
> > {
> > 
> >     ...
> >     switch (state) {
> >              case NETDEV_POST_INIT:
> >                      SET_NETDEV_DEVTYPE(dev, &wiphy_type);
> >                      break;
> >              case NETDEV_REGISTER:
> >                      ...
> >                      dev->rtnl_link_ops = &wireless_link_ops;
> >                      ...
> >     }
> 
> Well it seems it could go more with the SET_NETDEV_TYPE()?
> 
> johannes
> 

What I worry about is this peace of code in net/core/dev.c : 

    int register_netdevice(struct net_device *dev)
    {
            ...

            if (!dev->rtnl_link_ops ||
                dev->rtnl_link_state == RTNL_LINK_INITIALIZED)
                    rtmsg_ifinfo(RTM_NEWLINK, dev, ~0U, GFP_KERNEL);
            ...
    }

which will skip sending RTM_NEWLINK message if it has rtnl_link_ops ...
Actually meanwhile I dont see any problems as my wifi works OK with
these changes, but I want to be sure in 100% that it does not break
anything ...

So I am under investigating this ... Maybe you can suggest ?

Thanks,