Re: [net-next: PATCH v2 5/8] device property: introduce fwnode_dev_node_match

[Andy Shevchenko <andriy.shevchenko@linux.intel.com>]

On Sat, Jul 16, 2022 at 01:15:55AM +0200, Marcin Wojtas wrote:
> pt., 15 lip 2022 o 21:42 Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> napisał(a):
> >
> > On Fri, Jul 15, 2022 at 10:36:29PM +0300, Andy Shevchenko wrote:
> > > On Fri, Jul 15, 2022 at 10:50:09AM +0200, Marcin Wojtas wrote:
> > > > This patch adds a new generic routine fwnode_dev_node_match
> > > > that can be used e.g. as a callback for class_find_device().
> > > > It searches for the struct device corresponding to a
> > > > struct fwnode_handle by iterating over device and
> > > > its parents.
> > >
> > > Implementation
> > > 1) misses the word 'parent';
> 
> I'm not sure. We don't necessarily look for parent device(s). We start
> with a struct device and if it matches the fwnode, success is returned
> immediately. Only otherwise we iterate over parent devices to find a
> match.

Yes, you iterate over parents. 0 iterations doesn't change semantics of
all cases, right?

> > > 2) located outside of the group of fwnode APIs operating on parents.
> 
> I can shift it right below fwnode_get_nth_parent if you prefer.

Yes, please do.

> > > I would suggest to rename to fwnode_get_next_parent_node() and place
> > > near to fwnode_get_next_parent_dev() (either before or after, where
> > > it makes more sense).
> >
> > And matching function will be after that:
> >
> >         return fwnode_get_next_parent_node(...) != NULL;
> >
> > Think about it. Maybe current solution is good enough, just needs better
> > naming (fwnode_match_parent_node()? Dunno).
> >
> > P.S. Actually _get maybe misleading as we won't bump reference counting,
> >      rather _find?
> 
> How about the following name:
> fwnode_find_dev_match()
> ?

fwnode_find_parent_dev_match() LGTM, thanks!

You iterate over parents.

-- 
With Best Regards,
Andy Shevchenko