diff mbox series

gpiolib: fix the size of the ngpios variable

Message ID 20220305154256.278394-1-brgl@bgdev.pl
State New
Headers show
Series gpiolib: fix the size of the ngpios variable | expand

Commit Message

Bartosz Golaszewski March 5, 2022, 3:42 p.m. UTC
The ngpio fields both in struct gpio_device as well as gpio_chip are
16 bit unsigned integers. Let's not risk an overflow and use the correct
size when reading the device property.

Fixes: 9dbd1ab20509 ("gpiolib: check the 'ngpios' property in core gpiolib code")
Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
---
 drivers/gpio/gpiolib.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Andy Shevchenko March 5, 2022, 9:26 p.m. UTC | #1
On Sat, Mar 5, 2022 at 10:14 PM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
>
> The ngpio fields both in struct gpio_device as well as gpio_chip are
> 16 bit unsigned integers. Let's not risk an overflow and use the correct
> size when reading the device property.

...

> -               ret = device_property_read_u32(&gdev->dev, "ngpios", &ngpios);
> +               ret = device_property_read_u16(&gdev->dev, "ngpios", &ngpios);

AFAIU DT format you may not do this. The property is ABI-like and it's
32-bit (BE!). Reading as u16 will bring you 2 MSB (bytes) out of 4
from the 32-bit value.
Bartosz Golaszewski March 6, 2022, 7:22 p.m. UTC | #2
On Sat, Mar 5, 2022 at 10:27 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
>
> On Sat, Mar 5, 2022 at 10:14 PM Bartosz Golaszewski <brgl@bgdev.pl> wrote:
> >
> > The ngpio fields both in struct gpio_device as well as gpio_chip are
> > 16 bit unsigned integers. Let's not risk an overflow and use the correct
> > size when reading the device property.
>
> ...
>
> > -               ret = device_property_read_u32(&gdev->dev, "ngpios", &ngpios);
> > +               ret = device_property_read_u16(&gdev->dev, "ngpios", &ngpios);
>
> AFAIU DT format you may not do this. The property is ABI-like and it's
> 32-bit (BE!). Reading as u16 will bring you 2 MSB (bytes) out of 4
> from the 32-bit value.
>

Ah right. It's not very obvious from the docs but the DT syntax must
be different for 16 bits. In that case our best shot is to read a
32-bit unsigned integer and check if its value doesn't exceed
UINT16_MAX.

Bart
diff mbox series

Patch

diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index a3d14277f17c..147b86db16aa 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -607,7 +607,7 @@  int gpiochip_add_data_with_key(struct gpio_chip *gc, void *data,
 	int base = gc->base;
 	unsigned int i;
 	int ret = 0;
-	u32 ngpios;
+	u16 ngpios;
 
 	if (gc->fwnode)
 		fwnode = gc->fwnode;
@@ -665,7 +665,7 @@  int gpiochip_add_data_with_key(struct gpio_chip *gc, void *data,
 	 * of GPIO lines.
 	 */
 	if (gc->ngpio == 0) {
-		ret = device_property_read_u32(&gdev->dev, "ngpios", &ngpios);
+		ret = device_property_read_u16(&gdev->dev, "ngpios", &ngpios);
 		if (ret == -ENODATA)
 			/*
 			 * -ENODATA means that there is no property found and