Re: [PATCH v6 8/8] reset: gpio: use software nodes to setup the GPIO lookup

[Philipp Zabel]

On Mi, 2025-11-19 at 00:28 -0800, Bartosz Golaszewski wrote:
> On Wed, 19 Nov 2025 09:19:30 +0100, Philipp Zabel <p.zabel@xxxxxxxxxxxxxx> said:
> > On Di, 2025-11-18 at 18:08 +0100, Bartosz Golaszewski wrote:
> > > On Tue, Nov 18, 2025 at 5:44 PM Philipp Zabel <p.zabel@xxxxxxxxxxxxxx> wrote:
> > > > 
> > > > On Do, 2025-11-06 at 15:32 +0100, Bartosz Golaszewski wrote:
> > > > > From: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> > > > > 
> > > > > GPIO machine lookup is a nice mechanism for associating GPIOs with
> > > > > consumers if we don't know what kind of device the GPIO provider is or
> > > > > when it will become available. However in the case of the reset-gpio, we
> > > > > are already holding a reference to the device and so can reference its
> > > > > firmware node. Let's setup a software node that references the relevant
> > > > > GPIO and attach it to the auxiliary device we're creating.
> > > > > 
> > > > > Reviewed-by: Philipp Zabel <p.zabel@xxxxxxxxxxxxxx>
> > > > > Acked-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
> > > > > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
> > > > > ---
> > > > 
> > > > I'll apply this with the following patch squashed in:
> > 
> > Strike that, I'll have to wait for the SPI issue to be resolved.
> > 
> > > > diff --git a/drivers/reset/core.c b/drivers/reset/core.c
> > > > index 3edf04ae8a95..8a7b112a9a77 100644
> > > > --- a/drivers/reset/core.c
> > > > +++ b/drivers/reset/core.c
> > > > @@ -945,7 +945,7 @@ static int __reset_add_reset_gpio_device(const struct of_phandle_args *args)
> > > >         of_node_get(rgpio_dev->of_args.np);
> > > > 
> > > >         rgpio_dev->swnode = fwnode_create_software_node(properties, NULL);
> > > > -       ret = PTR_ERR(rgpio_dev->swnode);
> > > > +       ret = PTR_ERR_OR_ZERO(rgpio_dev->swnode);
> > > >         if (ret)
> > > >                 goto err_put_of_node;
> > > 
> > > Huh? Why?
> > 
> > PTR_ERR(ptr) is just (long)ptr, so a valid swnode pointer makes ret
> > non-zero and takes us into the error path. PTR_ERR_OR_ZERO() includes
> > the IS_ERR() check and returns 0 for non-error pointers.
> > 
> > And there is a (false-positive) sparse warning:
> > 
> >  drivers/reset/core.c:978 __reset_add_reset_gpio_device() warn: passing zero to 'PTR_ERR'
> > 
> > I think it would be better to return to the explicit IS_ERR() check
> > from v5.
> > 
> 
> Yes, it was like this in my initial submission and seems like it's more
> readable and doesn't cause this confusion. I will go back to it. Though
> I'm not sure if the SPI issue will require a v5

v7 :)

If there are no other changes required, I can fix it up:

  https://git.pengutronix.de/cgit/pza/linux/commit/?id=8e72a48fc1df

>  - I'm looking into fixing it in the affected driver.

Great, can we make sure the fix is applied first, to avoid breaking git
bisect?

regards
Philipp