Re: [RFC] hv_netvsc: automatically name slave VF network device

[Jiri Pirko <jiri@resnulli.us>]

Wed, Dec 20, 2017 at 12:53:53AM CET, jakub.kicinski@netronome.com wrote:
>On Tue, 19 Dec 2017 15:44:05 -0800, Stephen Hemminger wrote:
>> On Tue, 19 Dec 2017 15:20:57 -0800
>> Jakub Kicinski <jakub.kicinski@netronome.com> wrote:
>> 
>> > On Tue, 19 Dec 2017 14:50:17 -0800, Stephen Hemminger wrote:  
>> > > On Tue, 19 Dec 2017 14:44:37 -0800
>> > > "Samudrala, Sridhar" <sridhar.samudrala@intel.com> wrote:
>> > >     
>> > > >  -static void __netvsc_vf_setup(struct net_device *ndev,      
>> > > > > -			      struct net_device *vf_netdev)
>> > > > > -{
>> > > > > -	int ret;
>> > > > > +	/* set the name of VF device based on upper device name */
>> > > > > +	snprintf(vf_name, IFNAMSIZ, "%s_vf", ndev->name);
>> > > > > +	ret = dev_change_name(vf_netdev, vf_name);
>> > > > > +	if (ret != 0)
>> > > > > +		netdev_warn(vf_netdev,
>> > > > > +			    "can not rename device: (%d)\n", ret);        
>> > > > 
>> > > > It is possible that upper device name can change after this call.  I 
>> > > > noticed this
>> > > > when i tried this approach with virtio_net.
>> > > > 
>> > > > Also, what should happen if the upper device is unloaded? Should we rename
>> > > > the VF name?      
>> > > 
>> > > Yes upper device can change name. So sure, netdevice could trap that
>> > > in callback (it already has notifier) and rename VF. Will add that in V2.
>> > > 
>> > > If upper device is unloaded then it is already decoupled from the VF.
>> > > There is no good value to change it back to. The orignal name probably
>> > > has been reused by then.    
>> > 
>> > Both of those issues would be solved by just exposing phys_port_name
>> > from the VF driver, and letting systemd do its thing independent of 
>> > magic bonds.
>> > 
>> > Reluctance to do driver work aside :/  
>> 
>> The port name for Mellanox driver is already set in the driver as a numeric value.
>> It indicates which port is used.
>> This won't work.
>
>You must be looking at representor ndos.  I don't think MLX NIC drivers
>are implementing phys_port_name for normal netdevs at all today.

You are right Jakub. I plan to generate phys_port_name according to the
netdev flavours (vf,pf,qp,vfrep,pfrep,qprep,etc). That would resolve
this.