Re: [PATCH] iio/adjd_s311: Fix potential memory leak in adjd_s311_update_scan_mode()

[Lars-Peter Clausen <lars@metafoo.de>]

On 08/08/2012 09:17 AM, Peter Meerwald wrote:
> 
>> Do not leak memory by updating pointer with potentially
>> NULL realloc return value.
> 
> I agree
> 
> use of krealloc() was suggested in driver review (see 
> http://www.spinics.net/lists/linux-iio/msg05930.html) to shorten the code; 
> unfortunately, I misunderstood the semantics of krealloc() in case 
> allocation fails

My fault I guess, sorry for that.

> 
> this is the original code:
> 
> 	kfree(data->buffer);
> 	data->buffer = kmalloc(indio_dev->scan_bytes, GFP_KERNEL);
> 	if (!data->buffer)
> 		return -ENOMEM;
> 
> I suggest to switch back to that original code, there is no need preserve 
> the data in the buffer as krealloc does

Agreed.

> 
> thanks, p.
> 
>> Found by Linux Driver Verification project (linuxtesting.org).
>>
>> Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
>> ---
>>  drivers/iio/light/adjd_s311.c |   14 ++++++++++----
>>  1 file changed, 10 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/iio/light/adjd_s311.c b/drivers/iio/light/adjd_s311.c
>> index 1cbb449..0adda5b 100644
>> --- a/drivers/iio/light/adjd_s311.c
>> +++ b/drivers/iio/light/adjd_s311.c
>> @@ -271,12 +271,18 @@ static int adjd_s311_update_scan_mode(struct iio_dev *indio_dev,
>>  	const unsigned long *scan_mask)
>>  {
>>  	struct adjd_s311_data *data = iio_priv(indio_dev);
>> -	data->buffer = krealloc(data->buffer, indio_dev->scan_bytes,
>> +	u16 *new_buffer;
>> +	int ret = 0;
>> +
>> +	new_buffer = krealloc(data->buffer, indio_dev->scan_bytes,
>>  				GFP_KERNEL);
>> -	if (!data->buffer)
>> -		return -ENOMEM;
>> +	if (new_buffer == NULL) {
>> +		kfree(data->buffer);
>> +		ret = -ENOMEM;
>> +	}
>> +	data->buffer = new_buffer;
>>  
>> -	return 0;
>> +	return ret;
>>  }
>>  
>>  static const struct iio_info adjd_s311_info = {
>>
>