Re: [PATCH] of: make for_each_property_of_node() available to to !OF

[Bartosz Golaszewski]

On Wed, Mar 6, 2024 at 8:34 PM Rob Herring <robh+dt@xxxxxxxxxx> wrote:
>
> On Tue, Mar 5, 2024 at 12:46 PM Bartosz Golaszewski <brgl@xxxxxxxx> wrote:
> >
> > On Tue, 5 Mar 2024 18:56:04 +0100, Rob Herring <robh+dt@xxxxxxxxxx> said:
> > >
> > > Long term, I want to make struct device_node opaque. So if we really
> > > want to fix this, I think we'd want to convert this to use an iterator
> > > function. Though I guess any user would be mucking with struct
> > > property too, so the whole loop would need to be reworked. So in
> > > conclusion, don't use for_each_property_of_node(). :) Shrug.
> > >
> >
> > I basically just need to get the list of all properties of a node. Even just
> > names. I'm working on a testing driver that needs to request all GPIOs assigned
> > to it over DT so it must find all `foo-gpios` properties.
> >
> > How about:
> >
> > int of_node_for_each_property(struct device_node *dn, int
> > (*func)(struct property *, void *), void *data)
> >
> > as the iterator?
>
> Why would we make the caller provide the iterator? We just need a
> function call like the other iterators rather than directly
> dereferencing the pointers: of_next_property_iter(node, last_prop)
>
> > You didn't say if you want to make struct property opaque as
> > well but even then it can be used with provided interfaces.
>
> Yes, I'd like to make struct property opaque as well. That's probably
> the first step.
>
> Rob

Or maybe we should implement it at the fwnode_operations level?
Although not sure how to handle it as fwnode doesn't have a separate
structure for properties.

Bart