Re: [V2 Patch net-next-2.6] netpoll: disable netpoll when enslave a device

[Neil Horman <nhorman@redhat.com>]

On Thu, May 19, 2011 at 07:31:27AM -0400, Andy Gospodarek wrote:
> On Thu, May 19, 2011 at 04:39:53PM +0800, Amerigo Wang wrote:
> [...]
> > diff --git a/include/linux/notifier.h b/include/linux/notifier.h
> > index 621dfa1..3d82867 100644
> > --- a/include/linux/notifier.h
> > +++ b/include/linux/notifier.h
> > @@ -211,6 +211,7 @@ static inline int notifier_to_errno(int ret)
> >  #define NETDEV_UNREGISTER_BATCH 0x0011
> >  #define NETDEV_BONDING_DESLAVE  0x0012
> >  #define NETDEV_NOTIFY_PEERS	0x0013
> > +#define NETDEV_ENSLAVE		0x0014
> >  
> >  #define SYS_DOWN	0x0001	/* Notify of system down */
> >  #define SYS_RESTART	SYS_DOWN
> 
> Neil just noted the same concern I had -- the asymmetry between
> NETDEV_ENSLAVE and NETDEV_BONDING_DESLAVE bothers me a bit.  I also
> don't really like the followup patch that uses 'ENSLAVE' in the bridging
> code when we typically use that language for bonding only.
> 
> What about changing NETDEV_BONDING_DESLAVE to NETDEV_RELEASE and create
> NETDEV_JOIN instead of NETDEV_ENSLAVE?  I would prefer that or something
> else that might use more generic language that could be applied to all
> for stacked interfaces.
JOIN and RELEASE (or perhaps LEAVE) sounds good to me.
Neil