Re: [PATCH net-next] devlink: Require devlink lock during device reload

[Jakub Kicinski]

On Sun, 7 Nov 2021 19:54:20 +0200 Leon Romanovsky wrote:
> > >  (3) should we let drivers take refs on the devlink instance?  
> > 
> > I think it's fine mainly because I don't expect it to be used by too
> > many drivers other than netdevsim which is somewhat special. Looking at
> > the call sites of devlink_get() in netdevsim, it is only called from
> > places (debugfs and trap workqueue) that shouldn't be present in real
> > drivers.  
> 
> Sorry, I'm obligated to ask. In which universe is it ok to create new
> set of API that no real driver should use?

I think it's common sense. We're just exporting something to make our
lives easier somewhere else in the three. Do you see a way in which
taking refs on devlink can help out-of-tree code?

BTW we can put the symbols in a namespace or under a kconfig, to aid
reviews of drivers using them if you want.