[PATCH] pinctrl: use kasprintf() in pinmux_request_gpio()

[Colin Cross]

On Thu, Sep 13, 2012 at 10:19 PM, Joe Perches <joe at perches.com> wrote:
> On Thu, 2012-09-13 at 21:49 +0200, Thomas Petazzoni wrote:
>> Instead of using a temporary buffer, snprintf() and kstrdup(), just
>> use kasprintf() that does the same thing in just oneline.
>>
>> Signed-off-by: Thomas Petazzoni <thomas.petazzoni at free-electrons.com>
>> ---
>>  drivers/pinctrl/pinmux.c |    5 +----
>>  1 file changed, 1 insertion(+), 4 deletions(-)
>>
>> diff --git a/drivers/pinctrl/pinmux.c b/drivers/pinctrl/pinmux.c
>> index 3d5ac73..9301a7a 100644
>> --- a/drivers/pinctrl/pinmux.c
>> +++ b/drivers/pinctrl/pinmux.c
>> @@ -232,14 +232,11 @@ int pinmux_request_gpio(struct pinctrl_dev *pctldev,
>>                       struct pinctrl_gpio_range *range,
>>                       unsigned pin, unsigned gpio)
>>  {
>> -     char gpiostr[16];
>>       const char *owner;
>>       int ret;
>>
>>       /* Conjure some name stating what chip and pin this is taken by */
>> -     snprintf(gpiostr, 15, "%s:%d", range->name, gpio);
>> -
>> -     owner = kstrdup(gpiostr, GFP_KERNEL);
>> +     owner = kasprintf(GFP_KERNEL, "%s:%d", range->name, gpio);
>>       if (!owner)
>>               return -EINVAL;
>>
>
> No not really.  It's a bit different because the first
> snprintf is length limited but the kasprintf is not.

The one it's replacing is worse, it never sets gpiostr[15] to 0 and
kstrdup will read past the end of the buffer if the string is
truncated.