On 03/02/2018 11:08 AM, Russell King - ARM Linux wrote:
> On Fri, Mar 02, 2018 at 08:22:52AM -0600, Andrew F. Davis wrote:
>>> diff --git a/drivers/gpio/gpiolib-of.c b/drivers/gpio/gpiolib-of.c
>>> index 84e5a9d..f0fab26 100644
>>> --- a/drivers/gpio/gpiolib-of.c
>>> +++ b/drivers/gpio/gpiolib-of.c
>>> @@ -241,29 +241,17 @@ struct gpio_desc *of_find_gpio(struct device *dev,
>>> const char *con_id,
>>> desc = of_get_named_gpiod_flags(dev->of_node, prop_name, idx,
>>> - /*
>>> - * -EPROBE_DEFER in our case means that we found a
>>> - * valid GPIO property, but no controller has been
>>> - * registered so far.
>>> - *
>>> - * This means we don't need to look any further for
>>> - * alternate name conventions, and we should really
>>> - * preserve the return code for our user to be able to
>>> - * retry probing later.
>>> - */
>>> - if (IS_ERR(desc) && PTR_ERR(desc) == -EPROBE_DEFER)
>>> - return desc;
>>> - if (!IS_ERR(desc) || (PTR_ERR(desc) != -ENOENT))
>>> + if (!IS_ERR(desc) || PTR_ERR(desc) != -ENOENT)
>> I rather like the () so one doesn't always have to look up C operator
>> precedence to verify..
> Too many make it impossible to see which close paren ties up with
> which open paren. I've spent way too long in the past reformatting
> code, where people think that () are a good thing, to work out what
> the comparison is actually doing before then rewriting the damn
> thing with less () and in a much clearer way. I'm now convinced
> that unnecessary () are a very bad thing as they severely harm
> readability as test complexity increases.
Fair enough, I personally prefer always having a new line per condition
when doing chained conditionals:
if (something &&
this == that &&
>>> /* Special handling for SPI GPIOs if used */
>>> - if (IS_ERR(desc))
>>> + if (IS_ERR(desc) && PTR_ERR(desc) == -ENOENT)
> These can be simplified to:
> if (desc == ERR_PTR(-ENOENT))
> since error pointers are unique - ERR_PTR(-ENOENT) is what was
> returned as an error-pointer, and if IS_ERR(ERR_PTR(-ENOMENT)) etc
> were false, we'd have big problems as it'd mean we couldn't detect
> error pointers!
> As an added bonus, they don't involve any operator precedence
> questions either.
Looks like your suggestion clears up this one anyway.