On Mon, Feb 15, 2021 at 05:15:36PM +0200, Andy Shevchenko wrote:
> On Mon, Feb 15, 2021 at 5:13 PM Andy Shevchenko
> <andy.shevche...@gmail.com> wrote:
> >
> > On Mon, Feb 15, 2021 at 2:33 PM Calvin Johnson
> > <calvin.john...@oss.nxp.com> wrote:
> > > On Mon, Feb 08, 2021 at 04:28:31PM +0000, Russell King - ARM Linux admin
> > > wrote:
> >
> > ...
> >
> > > I think of_phy_is_fixed_link() needs to be fixed. I'll add below fix.
> > >
> > > --- a/drivers/net/mdio/of_mdio.c
> > > +++ b/drivers/net/mdio/of_mdio.c
> > > @@ -439,6 +439,9 @@ bool of_phy_is_fixed_link(struct device_node *np)
> > > int len, err;
> > > const char *managed;
> > >
> > > + if (!np)
> > > + return false;
> >
> > AFAICS this doesn't add anything: all of the of_* APIs should handle
> > OF nodes being NULL below.
> >
> > > /* New binding */
> > > dn = of_get_child_by_name(np, "fixed-link");
> > > if (dn) {
>
> Yes, of_get_next_child() and of_get_property() are NULL aware.
>
> So, the check is redundant.
Yes, all the function calls in of_phy_is_fixed_link() handles NULL properly.
I don't see any way this function can oops.
Regards
Calvin
