On Thu, 10 Sep 2020 17:09:50 +0200 Andrew Lunn wrote:
> On Thu, Sep 10, 2020 at 07:38:16AM -0700, Jakub Kicinski wrote:
> > On Thu, 10 Sep 2020 01:58:20 +0200 Andrew Lunn wrote:  
> > > diff --git a/drivers/net/ethernet/intel/ice/ice_devlink.c 
> > > b/drivers/net/ethernet/intel/ice/ice_devlink.c
> > > index 111d6bfe4222..eb189d2070ae 100644
> > > --- a/drivers/net/ethernet/intel/ice/ice_devlink.c
> > > +++ b/drivers/net/ethernet/intel/ice/ice_devlink.c
> > > @@ -413,6 +413,7 @@ void ice_devlink_destroy_port(struct ice_pf *pf)
> > >   * error code on failure.
> > >   */
> > >  static int ice_devlink_nvm_snapshot(struct devlink *devlink,
> > > +                             const struct devlink_region_ops *ops,
> > >                               struct netlink_ext_ack *extack, u8 **data)
> > >  {
> > >   struct ice_pf *pf = devlink_priv(devlink);
> > > @@ -468,6 +469,7 @@ static int ice_devlink_nvm_snapshot(struct devlink 
> > > *devlink,
> > >   */
> > >  static int
> > >  ice_devlink_devcaps_snapshot(struct devlink *devlink,
> > > +                      const struct devlink_region_ops *ops,
> > >                        struct netlink_ext_ack *extack, u8 **data)
> > >  {
> > >   struct ice_pf *pf = devlink_priv(devlink);  
> > 
> > 
> > drivers/net/ethernet/intel/ice/ice_devlink.c:418: warning: Function 
> > parameter or member 'ops' not described in 'ice_devlink_nvm_snapshot'
> > drivers/net/ethernet/intel/ice/ice_devlink.c:474: warning: Function 
> > parameter or member 'ops' not described in 'ice_devlink_devcaps_snapshot'  
> 
> Thanks Jakub
> 
> I did try a W=1 build, but there are so many warnings it is hard to
> pick out the new ones. I assume you have some sort of automation for
> this? Could you share it?

I knew you'd ask :)

The scripts are here: https://github.com/kuba-moo/nipa
but they run off a private patchwork instance, I was planning to make
it possible for it to consume a local mdir:

https://github.com/kuba-moo/nipa/blob/master/ingest_mdir.py

But that has probably bit rotted completely by now.

I submit my own patches to netdev and check if I broke anything later.

> How far are we from just enabling W=1 by default under drivers/net ?

I think pretty far :(

> What is the best way to do this in the Makefiles? I think
> drivers/net/phy is now W=1 clean, or very close. So it would be nice
> to enable that checking by default.

No idea 🤔

Reply via email to