2020-01-25 22:15:56

by Paul Thomas

Subject: [PATCH] GPIO, Fix bug where the wrong GPIO register is written to

Care is taken with "index", however with the current version
the actual xgpio_writereg is using index for data but
xgpio_regoffset(chip, i) for the offset. And since i is already
incremented it is incorrect. This patch fixes it so that index
is used for the offset too.

Signed-off-by: Paul Thomas <[email protected]>
---
There are many different ways to correct this, I'd just like it to get
fixed. I've tested this with a 5.2 kernel, but this patch is against
5.5rc7.

drivers/gpio/gpio-xilinx.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpio/gpio-xilinx.c b/drivers/gpio/gpio-xilinx.c
index a9748b5198e6..67f9f82e0db0 100644
--- a/drivers/gpio/gpio-xilinx.c
+++ b/drivers/gpio/gpio-xilinx.c
@@ -147,9 +147,10 @@ static void xgpio_set_multiple(struct gpio_chip *gc, unsigned long *mask,
for (i = 0; i < gc->ngpio; i++) {
if (*mask == 0)
break;
+ /* Once finished with an index write it out to the register */
if (index != xgpio_index(chip, i)) {
xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET +
- xgpio_regoffset(chip, i),
+ index * XGPIO_CHANNEL_OFFSET,
chip->gpio_state[index]);
spin_unlock_irqrestore(&chip->gpio_lock[index], flags);
index = xgpio_index(chip, i);
@@ -165,7 +166,7 @@ static void xgpio_set_multiple(struct gpio_chip *gc, unsigned long *mask,
}

xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET +
- xgpio_regoffset(chip, i), chip->gpio_state[index]);
+ index * XGPIO_CHANNEL_OFFSET, chip->gpio_state[index]);

spin_unlock_irqrestore(&chip->gpio_lock[index], flags);
}
--
2.17.1

2020-01-26 21:33:17

by Andrew Lunn

Subject: Re: [PATCH] GPIO, Fix bug where the wrong GPIO register is written to

On Sat, Jan 25, 2020 at 05:14:10PM -0500, Paul Thomas wrote:
> Care is taken with "index", however with the current version
> the actual xgpio_writereg is using index for data but
> xgpio_regoffset(chip, i) for the offset. And since i is already
> incremented it is incorrect. This patch fixes it so that index
> is used for the offset too.
>
> Signed-off-by: Paul Thomas <[email protected]>

Hi Paul

Please put Xilinx into the subject line. I had to actually look at the
patch to decide it was not relevant to me.

Andrew

2020-01-27 07:45:48

by Michal Simek

Subject: Re: [PATCH] GPIO, Fix bug where the wrong GPIO register is written to

On 25. 01. 20 23:14, Paul Thomas wrote:
> Care is taken with "index", however with the current version
> the actual xgpio_writereg is using index for data but
> xgpio_regoffset(chip, i) for the offset. And since i is already
> incremented it is incorrect. This patch fixes it so that index
> is used for the offset too.
>
> Signed-off-by: Paul Thomas <[email protected]>
> ---
> There are many different ways to correct this, I'd just like it to get
> fixed. I've tested this with a 5.2 kernel, but this patch is against
> 5.5rc7.
>
> drivers/gpio/gpio-xilinx.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpio/gpio-xilinx.c b/drivers/gpio/gpio-xilinx.c
> index a9748b5198e6..67f9f82e0db0 100644
> --- a/drivers/gpio/gpio-xilinx.c
> +++ b/drivers/gpio/gpio-xilinx.c
> @@ -147,9 +147,10 @@ static void xgpio_set_multiple(struct gpio_chip *gc, unsigned long *mask,
> for (i = 0; i < gc->ngpio; i++) {
> if (*mask == 0)
> break;
> + /* Once finished with an index write it out to the register */
> if (index != xgpio_index(chip, i)) {
> xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET +
> - xgpio_regoffset(chip, i),
> + index * XGPIO_CHANNEL_OFFSET,
> chip->gpio_state[index]);
> spin_unlock_irqrestore(&chip->gpio_lock[index], flags);
> index = xgpio_index(chip, i);
> @@ -165,7 +166,7 @@ static void xgpio_set_multiple(struct gpio_chip *gc, unsigned long *mask,
> }
>
> xgpio_writereg(chip->regs + XGPIO_DATA_OFFSET +
> - xgpio_regoffset(chip, i), chip->gpio_state[index]);
> + index * XGPIO_CHANNEL_OFFSET, chip->gpio_state[index]);
>
> spin_unlock_irqrestore(&chip->gpio_lock[index], flags);
> }
>

Shubhrajyoti: Please take a look.

Thanks,
Michal

2020-02-10 11:53:02

by Linus Walleij

Subject: Re: [PATCH] GPIO, Fix bug where the wrong GPIO register is written to

On Sat, Jan 25, 2020 at 11:14 PM Paul Thomas <[email protected]> wrote:

> Care is taken with "index", however with the current version
> the actual xgpio_writereg is using index for data but
> xgpio_regoffset(chip, i) for the offset. And since i is already
> incremented it is incorrect. This patch fixes it so that index
> is used for the offset too.
>
> Signed-off-by: Paul Thomas <[email protected]>
> ---
> There are many different ways to correct this, I'd just like it to get
> fixed. I've tested this with a 5.2 kernel, but this patch is against
> 5.5rc7.

Fixed up the subject and applied, added a tag for stable.

Thanks!
Linus Walleij

2020-02-18 07:29:15

by Paul Thomas

Subject: Re: [PATCH] GPIO, Fix bug where the wrong GPIO register is written to

> Fixed up the subject and applied, added a tag for stable.
>
> Thanks!
> Linus Walleij

Thanks everyone, next time I'll do the subject better.

-Paul