Re: [PATCH v2] gpio: swnode: restore the swnode-name-against-chip-label matching

[Andy Shevchenko]

On Wed, Feb 18, 2026 at 11:22:07AM -0800, Dmitry Torokhov wrote:
> On Wed, Feb 18, 2026 at 07:08:22PM +0100, Bartosz Golaszewski wrote:
> > On Wed, Feb 18, 2026 at 6:15 PM Dmitry Torokhov
> > <dmitry.torokhov@xxxxxxxxx> wrote:
> > > On Wed, Feb 18, 2026 at 09:42:28AM +0100, Bartosz Golaszewski wrote:
> > > > On Wed, Feb 18, 2026 at 1:31 AM Dmitry Torokhov
> > > > <dmitry.torokhov@xxxxxxxxx> wrote:
> > > > > On Wed, Feb 11, 2026 at 09:53:13AM +0100, Bartosz Golaszewski wrote:
> > > > > > Using the remote firmware node for software node lookup is the right
> > > > > > thing to do. The GPIO controller we want to resolve should have the
> > > > > > software node we scooped out of the reference attached to it. However,
> > > > > > there are existing users who abuse the software node API by creating
> > > > > > dummy swnodes whose name is set to the expected label string of the GPIO
> > > > > > controller whose pins they want to control and use them in their local
> > > > > > swnode references as GPIO properties.
> > > > > >
> > > > > > This used to work when we compared the software node's name to the
> > > > > > chip's label. When we switched to using a real fwnode lookup, these
> > > > > > users broke down because the firmware nodes in question were never
> > > > > > attached to the controllers they were looking for.
> > > > > >
> > > > > > Restore the label matching as a fallback to fix the broken users but add
> > > > > > a big FIXME urging for a better solution.
> > > > > >
> > > > > > Cc: stable@xxxxxxxxxxxxxxx # v6.18, v6.19
> > > > > > Fixes: 216c12047571 ("gpio: swnode: allow referencing GPIO chips by firmware nodes")
> > > > > > Link: https://lore.kernel.org/all/aYkdKfP5fg6iywgr@jekhomev/
> > > > > > Acked-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> > > > > > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxxxxxxxx>
> > > > > > ---
> > > > > > Changes in v2:
> > > > > > - check if gdev_node and gdev_node->name are not NULL before trying to
> > > > > >   match the label (Hans & Dan)
> > > > > > - use the right link
> > > > > > - collect tags
> > > > > >
> > > > > >  drivers/gpio/gpiolib-swnode.c | 19 +++++++++++++++++++
> > > > > >  1 file changed, 19 insertions(+)
> > > > > >
> > > > > > diff --git a/drivers/gpio/gpiolib-swnode.c b/drivers/gpio/gpiolib-swnode.c
> > > > > > index 21478b45c127d..0d7f3f09a0b4b 100644
> > > > > > --- a/drivers/gpio/gpiolib-swnode.c
> > > > > > +++ b/drivers/gpio/gpiolib-swnode.c
> > > > > > @@ -42,6 +42,25 @@ static struct gpio_device *swnode_get_gpio_device(struct fwnode_handle *fwnode)
> > > > > >
> > > > > >  fwnode_lookup:
> > > > > >       gdev = gpio_device_find_by_fwnode(fwnode);
> > > > >
> > > > > By the way, should we extend gpio_device_find_by_fwnode() to use both
> > > > > primary and secondary nodes?
> > > > >
> > > >
> > > > That's already done on a higher lever for all fwnodes in gpiod_fwnode_lookup().
> > >
> > > How exactly? I am not talking about checking secondary node for the
> > > fwnode that is used in the reference, I am talking about secondary
> > > fwnode that might be assigned to the gpio chip and you need to check
> > > both primary and secondary if they match with the fwnode that you call
> > > gpio_device_find_by_fwnode() with.
> > >
> > 
> > Right, I didn't quite get it. I was surprised to find out
> > device_match_fwnode() - which we use in gpiolib - does not do it
> > already. I'm wondering if this is something we should change in device
> > core or only locally.
> 
> Yes, I think so (change in driver core). There is only a handful of
> calls to device_match_fwnode() that they can be audited to make sure the
> conversion is safe.

I believe there was already similar discussion and IIRC conclusion was to
distinguish current device_match_fwnode() with something new that will check
also secondary one. This way we can in a usual way convert drivers one-by-one
if even required.

-- 
With Best Regards,
Andy Shevchenko