Re: [PATCH 5/8] gpio: exar: Fix reading of directions and values

On Thu, May 18, 2017 at 8:20 AM, Jan Kiszka [off-list ref] wrote:

On 2017-05-13 15:36, Andy Shevchenko wrote:

quoted

On Sat, May 13, 2017 at 10:29 AM, Jan Kiszka [off-list ref] wrote:

quoted

First, the logic for translating a register bit to the return code of
exar_get_direction and exar_get_value were wrong. And second, there was
a flip regarding the register bank in exar_get_direction.

Again, I wish it was tested in the first place.

After addressing below:
FWIW:
Reviewed-by: Andy Shevchenko <redacted>

quoted

@@ -68,7 +68,7 @@ static int exar_get(struct gpio_chip *chip, unsigned int reg)
        value = readb(exar_gpio->regs + reg);
        mutex_unlock(&exar_gpio->lock);

-       return !!value;
+       return value;

This one is correct.

quoted

quoted

@@ -80,7 +80,7 @@ static int exar_get_direction(struct gpio_chip *chip, unsigned int offset)
        addr = bank ? EXAR_OFFSET_MPIOSEL_HI : EXAR_OFFSET_MPIOSEL_LO;
        val = exar_get(chip, addr) >> (offset % 8);

-       return !!val;
+       return val & 1;

It should be rather

        val = exar_get(chip, addr) & BIT(offset % 8);

That won't give us 0 or 1 as return value, thus would be incorrect.

Full picture:

 val = exar_get(chip, addr) & BIT(offset % 8);

 return !!val;

How it could be non-(1 or 0)?

-- 
With Best Regards,
Andy Shevchenko