Re: [PATCH/RFC net-next] rocker: forward packets to CPU when a port in promiscuous mode

[Simon Horman <simon.horman@netronome.com>]

On Wed, Jul 15, 2015 at 07:50:32AM -0700, Scott Feldman wrote:
> On Wed, Jul 15, 2015 at 12:54 AM, Simon Horman
> <simon.horman@netronome.com> wrote:
> > On Wed, Jul 15, 2015 at 12:18:20AM -0700, Scott Feldman wrote:
> >> On Tue, Jul 14, 2015 at 11:34 PM, Simon Horman
> >> <simon.horman@netronome.com> wrote:
> >> > On Tue, Jul 14, 2015 at 10:32:54PM -0700, Scott Feldman wrote:
> >>
> >> >> > @@ -5263,11 +5301,16 @@ static int rocker_port_master_changed(struct net_device *dev)
> >> >> >          * 3. Other, e.g. being added to or removed from a bond or openvswitch,
> >> >> >          *    in which case nothing is done
> >> >> >          */
> >> >>
> >> >> Maybe comment above needs adjusting?
> >> >
> >> > Indeed, sorry for missing that.
> >> > How about this?
> >> >
> >> >
> >> >         /* There are currently five cases handled here:
> >> >          * 1. Joining a bridge
> >> >          * 2. Joining a Open vSwitch datapath
> >> >          * 3. Leaving a previously joined bridge
> >> >          * 4. Leaving a previously joined Open vSwitch datapath
> >> >          * 5. Other, e.g. being added to or removed from a bond,
> >> >          *    in which case nothing is done
> >> >          */
> >>
> >> Seems like one of those comments that needs adjusting each time the
> >> code changes, which isn't good.  Maybe just kill the comment or write
> >> in a generic way saying what code is doing.  Code seems obvious enough
> >> to me to not require a comment.
> >
> > My purpose in adding the comment in the first place was
> > to document the "other" case, as it wasn't handled and thus
> > didn't seem obvious.
> >
> > Perhaps something like this would work.
> >
> >         /* N.B: Do nothing if the type of master is not supported */
> 
> Ack

Thanks. I've made a formal submission of the path with the
changes you suggested earlier in this thread.