* [PATCH v2] gpiolib: fix hogs with multiple lines
@ 2026-03-26 14:18 Bartosz Golaszewski
2026-03-27 8:58 ` Linus Walleij
2026-03-27 9:08 ` Geert Uytterhoeven
0 siblings, 2 replies; 3+ messages in thread
From: Bartosz Golaszewski @ 2026-03-26 14:18 UTC (permalink / raw)
To: Linus Walleij, Bartosz Golaszewski, Mika Westerberg,
Andy Shevchenko, Geert Uytterhoeven
Cc: linux-gpio, linux-kernel, Geert Uytterhoeven, Bartosz Golaszewski
After moving GPIO hog handling into GPIOLIB core, we accidentally stopped
supporting devicetree hog definitions with multiple lines like so:
hog {
gpio-hog;
gpios = <3 0>, <4 GPIO_ACTIVE_LOW>;
output-high;
line-name = "foo";
};
Restore this functionality to fix reported regressions.
Fixes: d1d564ec4992 ("gpio: move hogs into GPIO core")
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Closes: https://lore.kernel.org/all/CAMuHMdX6RuZXAozrF5m625ZepJTVVr4pcyKczSk12MedWvoejw@mail.gmail.com/
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
---
Changes in v2:
- Use a default value of 2 for the number of GPIO cells
- Use kzalloc_objs() instead of kcalloc()
- Propagate the error number from fwnode_property_read_u32()
- Remove special cases forced by the lack of default cells value
- Link to v1: https://patch.msgid.link/20260325-gpio-hogs-multiple-v1-1-7e3a7347f0d6@oss.qualcomm.com
---
drivers/gpio/gpiolib.c | 92 +++++++++++++++++++++++++++++---------------------
1 file changed, 54 insertions(+), 38 deletions(-)
diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 7ee45ccdf1265e94b68fe8fa0d10dbc6adccd93a..cc36f1d1f3edb2b5c8609ab5ba72b23e6db480bd 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -938,12 +938,18 @@ int gpiochip_add_hog(struct gpio_chip *gc, struct fwnode_handle *fwnode)
struct fwnode_handle *gc_node = dev_fwnode(&gc->gpiodev->dev);
struct fwnode_reference_args gpiospec;
enum gpiod_flags dflags;
+ const char *name = NULL;
struct gpio_desc *desc;
unsigned long lflags;
- const char *name;
+ size_t num_hogs;
int ret, argc;
- u32 gpios[3]; /* We support up to three-cell bindings. */
- u32 cells;
+ /*
+ * For devicetree-based systems, this needs to be defined in bindings
+ * and there's no real default value. For other firmware descriptions
+ * it makes the most sense to use 2 cells for the GPIO offset and
+ * request flags.
+ */
+ u32 cells = 2;
lflags = GPIO_LOOKUP_FLAGS_DEFAULT;
dflags = GPIOD_ASIS;
@@ -952,42 +958,22 @@ int gpiochip_add_hog(struct gpio_chip *gc, struct fwnode_handle *fwnode)
argc = fwnode_property_count_u32(fwnode, "gpios");
if (argc < 0)
return argc;
- if (argc > 3)
- return -EINVAL;
- ret = fwnode_property_read_u32_array(fwnode, "gpios", gpios, argc);
- if (ret < 0)
+ ret = fwnode_property_read_u32(gc_node, "#gpio-cells", &cells);
+ if (ret && is_of_node(fwnode))
return ret;
+ if (!ret && (argc % cells))
+ return -EINVAL;
- if (is_of_node(fwnode)) {
- /*
- * OF-nodes need some additional special handling for
- * translating of devicetree flags.
- */
- ret = fwnode_property_read_u32(gc_node, "#gpio-cells", &cells);
- if (ret)
- return ret;
- if (!ret && argc != cells)
- return -EINVAL;
-
- memset(&gpiospec, 0, sizeof(gpiospec));
- gpiospec.fwnode = fwnode;
- gpiospec.nargs = argc;
+ num_hogs = argc / cells;
- for (int i = 0; i < argc; i++)
- gpiospec.args[i] = gpios[i];
+ u32 *gpios __free(kfree) = kzalloc_objs(*gpios, argc);
+ if (!gpios)
+ return -ENOMEM;
- ret = of_gpiochip_get_lflags(gc, &gpiospec, &lflags);
- if (ret)
- return ret;
- } else {
- /*
- * GPIO_ACTIVE_LOW is currently the only lookup flag
- * supported for non-OF firmware nodes.
- */
- if (gpios[1])
- lflags |= GPIO_ACTIVE_LOW;
- }
+ ret = fwnode_property_read_u32_array(fwnode, "gpios", gpios, argc);
+ if (ret < 0)
+ return ret;
if (fwnode_property_present(fwnode, "input"))
dflags |= GPIOD_IN;
@@ -1000,11 +986,41 @@ int gpiochip_add_hog(struct gpio_chip *gc, struct fwnode_handle *fwnode)
fwnode_property_read_string(fwnode, "line-name", &name);
- desc = gpiochip_get_desc(gc, gpios[0]);
- if (IS_ERR(desc))
- return PTR_ERR(desc);
+ for (unsigned int i = 0; i < num_hogs; i++) {
+ if (is_of_node(fwnode)) {
+ /*
+ * OF-nodes need some additional special handling for
+ * translating of devicetree flags.
+ */
+ memset(&gpiospec, 0, sizeof(gpiospec));
+ gpiospec.fwnode = fwnode;
+ gpiospec.nargs = cells;
+
+ for (int j = 0; j < cells; j++)
+ gpiospec.args[j] = gpios[i * cells + j];
+
+ ret = of_gpiochip_get_lflags(gc, &gpiospec, &lflags);
+ if (ret)
+ return ret;
+ } else {
+ /*
+ * GPIO_ACTIVE_LOW is currently the only lookup flag
+ * supported for non-OF firmware nodes.
+ */
+ if (gpios[i * cells + 1])
+ lflags |= GPIO_ACTIVE_LOW;
+ }
+
+ desc = gpiochip_get_desc(gc, gpios[i * cells]);
+ if (IS_ERR(desc))
+ return PTR_ERR(desc);
- return gpiod_hog(desc, name, lflags, dflags);
+ ret = gpiod_hog(desc, name, lflags, dflags);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
}
static int gpiochip_hog_lines(struct gpio_chip *gc)
---
base-commit: 880e76d276421113e3017d1b950bf932dc4e3ce3
change-id: 20260325-gpio-hogs-multiple-1718d38c1f5b
Best regards,
--
Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH v2] gpiolib: fix hogs with multiple lines
2026-03-26 14:18 [PATCH v2] gpiolib: fix hogs with multiple lines Bartosz Golaszewski
@ 2026-03-27 8:58 ` Linus Walleij
2026-03-27 9:08 ` Geert Uytterhoeven
1 sibling, 0 replies; 3+ messages in thread
From: Linus Walleij @ 2026-03-27 8:58 UTC (permalink / raw)
To: Bartosz Golaszewski
Cc: Bartosz Golaszewski, Mika Westerberg, Andy Shevchenko,
Geert Uytterhoeven, linux-gpio, linux-kernel, Geert Uytterhoeven
On Thu, Mar 26, 2026 at 3:18 PM Bartosz Golaszewski
<bartosz.golaszewski@oss.qualcomm.com> wrote:
> After moving GPIO hog handling into GPIOLIB core, we accidentally stopped
> supporting devicetree hog definitions with multiple lines like so:
>
> hog {
> gpio-hog;
> gpios = <3 0>, <4 GPIO_ACTIVE_LOW>;
> output-high;
> line-name = "foo";
> };
>
> Restore this functionality to fix reported regressions.
>
> Fixes: d1d564ec4992 ("gpio: move hogs into GPIO core")
> Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Closes: https://lore.kernel.org/all/CAMuHMdX6RuZXAozrF5m625ZepJTVVr4pcyKczSk12MedWvoejw@mail.gmail.com/
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
Oups, nice catch!
Reviewed-by: Linus Walleij <linusw@kernel.org>
Yours,
Linus Walleij
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH v2] gpiolib: fix hogs with multiple lines
2026-03-26 14:18 [PATCH v2] gpiolib: fix hogs with multiple lines Bartosz Golaszewski
2026-03-27 8:58 ` Linus Walleij
@ 2026-03-27 9:08 ` Geert Uytterhoeven
1 sibling, 0 replies; 3+ messages in thread
From: Geert Uytterhoeven @ 2026-03-27 9:08 UTC (permalink / raw)
To: Bartosz Golaszewski
Cc: Linus Walleij, Bartosz Golaszewski, Mika Westerberg,
Andy Shevchenko, Geert Uytterhoeven, linux-gpio, linux-kernel
Hi Bartosz,
On Thu, 26 Mar 2026 at 15:18, Bartosz Golaszewski
<bartosz.golaszewski@oss.qualcomm.com> wrote:
> After moving GPIO hog handling into GPIOLIB core, we accidentally stopped
> supporting devicetree hog definitions with multiple lines like so:
>
> hog {
> gpio-hog;
> gpios = <3 0>, <4 GPIO_ACTIVE_LOW>;
> output-high;
> line-name = "foo";
> };
>
> Restore this functionality to fix reported regressions.
>
> Fixes: d1d564ec4992 ("gpio: move hogs into GPIO core")
> Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
> Closes: https://lore.kernel.org/all/CAMuHMdX6RuZXAozrF5m625ZepJTVVr4pcyKczSk12MedWvoejw@mail.gmail.com/
> Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@oss.qualcomm.com>
> ---
> Changes in v2:
> - Use a default value of 2 for the number of GPIO cells
> - Use kzalloc_objs() instead of kcalloc()
> - Propagate the error number from fwnode_property_read_u32()
> - Remove special cases forced by the lack of default cells value
> - Link to v1: https://patch.msgid.link/20260325-gpio-hogs-multiple-v1-1-7e3a7347f0d6@oss.qualcomm.com
Thanks for the update!
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -938,12 +938,18 @@ int gpiochip_add_hog(struct gpio_chip *gc, struct fwnode_handle *fwnode)
> struct fwnode_handle *gc_node = dev_fwnode(&gc->gpiodev->dev);
> struct fwnode_reference_args gpiospec;
> enum gpiod_flags dflags;
> + const char *name = NULL;
> struct gpio_desc *desc;
> unsigned long lflags;
> - const char *name;
> + size_t num_hogs;
unsigned int should be sufficient to store the quotient of two positive
32-bit numbers.
> int ret, argc;
> - u32 gpios[3]; /* We support up to three-cell bindings. */
> - u32 cells;
> + /*
> + * For devicetree-based systems, this needs to be defined in bindings
> + * and there's no real default value. For other firmware descriptions
> + * it makes the most sense to use 2 cells for the GPIO offset and
> + * request flags.
> + */
> + u32 cells = 2;
>
> lflags = GPIO_LOOKUP_FLAGS_DEFAULT;
> dflags = GPIOD_ASIS;
> @@ -952,42 +958,22 @@ int gpiochip_add_hog(struct gpio_chip *gc, struct fwnode_handle *fwnode)
> argc = fwnode_property_count_u32(fwnode, "gpios");
> if (argc < 0)
> return argc;
> - if (argc > 3)
> - return -EINVAL;
>
> - ret = fwnode_property_read_u32_array(fwnode, "gpios", gpios, argc);
> - if (ret < 0)
> + ret = fwnode_property_read_u32(gc_node, "#gpio-cells", &cells);
> + if (ret && is_of_node(fwnode))
> return ret;
> + if (!ret && (argc % cells))
No need to check for !ret, the %-test should be valid for non-DT, too.
I guess we can't have #gpio-cells = <0> (single-GPIO provider not
supporting flags), and it would be very difficult to hog it.
> + return -EINVAL;
>
> - if (is_of_node(fwnode)) {
> - /*
> - * OF-nodes need some additional special handling for
> - * translating of devicetree flags.
> - */
> - ret = fwnode_property_read_u32(gc_node, "#gpio-cells", &cells);
> - if (ret)
> - return ret;
> - if (!ret && argc != cells)
> - return -EINVAL;
> -
> - memset(&gpiospec, 0, sizeof(gpiospec));
> - gpiospec.fwnode = fwnode;
> - gpiospec.nargs = argc;
> + num_hogs = argc / cells;
>
> - for (int i = 0; i < argc; i++)
> - gpiospec.args[i] = gpios[i];
> + u32 *gpios __free(kfree) = kzalloc_objs(*gpios, argc);
> + if (!gpios)
> + return -ENOMEM;
>
> - ret = of_gpiochip_get_lflags(gc, &gpiospec, &lflags);
> - if (ret)
> - return ret;
> - } else {
> - /*
> - * GPIO_ACTIVE_LOW is currently the only lookup flag
> - * supported for non-OF firmware nodes.
> - */
> - if (gpios[1])
> - lflags |= GPIO_ACTIVE_LOW;
> - }
> + ret = fwnode_property_read_u32_array(fwnode, "gpios", gpios, argc);
> + if (ret < 0)
> + return ret;
>
> if (fwnode_property_present(fwnode, "input"))
> dflags |= GPIOD_IN;
> @@ -1000,11 +986,41 @@ int gpiochip_add_hog(struct gpio_chip *gc, struct fwnode_handle *fwnode)
>
> fwnode_property_read_string(fwnode, "line-name", &name);
>
> - desc = gpiochip_get_desc(gc, gpios[0]);
> - if (IS_ERR(desc))
> - return PTR_ERR(desc);
> + for (unsigned int i = 0; i < num_hogs; i++) {
> + if (is_of_node(fwnode)) {
> + /*
> + * OF-nodes need some additional special handling for
> + * translating of devicetree flags.
> + */
> + memset(&gpiospec, 0, sizeof(gpiospec));
> + gpiospec.fwnode = fwnode;
> + gpiospec.nargs = cells;
> +
> + for (int j = 0; j < cells; j++)
unsigned int
> + gpiospec.args[j] = gpios[i * cells + j];
> +
> + ret = of_gpiochip_get_lflags(gc, &gpiospec, &lflags);
> + if (ret)
> + return ret;
> + } else {
> + /*
> + * GPIO_ACTIVE_LOW is currently the only lookup flag
> + * supported for non-OF firmware nodes.
> + */
> + if (gpios[i * cells + 1])
> + lflags |= GPIO_ACTIVE_LOW;
> + }
> +
> + desc = gpiochip_get_desc(gc, gpios[i * cells]);
> + if (IS_ERR(desc))
> + return PTR_ERR(desc);
>
> - return gpiod_hog(desc, name, lflags, dflags);
> + ret = gpiod_hog(desc, name, lflags, dflags);
> + if (ret)
> + return ret;
> + }
> +
> + return 0;
> }
>
> static int gpiochip_hog_lines(struct gpio_chip *gc)
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-03-27 9:08 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-03-26 14:18 [PATCH v2] gpiolib: fix hogs with multiple lines Bartosz Golaszewski
2026-03-27 8:58 ` Linus Walleij
2026-03-27 9:08 ` Geert Uytterhoeven
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®