Re: [PATCH 1/2] iio: fix possible buffer overflow

[Jonathan Cameron <jic23@kernel.org>]

On 02/05/14 23:40, Alexandre Belloni wrote:
> Found using smatch:
> drivers/iio/industrialio-core.c:719 iio_device_add_info_mask_type() error:
> buffer overflow 'iio_chan_info_postfix' 17 <= 63
>
> It was probably never hit because the info_mask_* members are filled by using
> the BIT() macro with values from the iio_chan_info_enum enum that also serve as
> the index of the iio_chan_info_postfix array.
>
> Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
See
ef4b4856593fc3d9d169bededdaf7acf62f83a52
iio:core: Fix bug in length of event info_mask and catch unhandled bits set in masks.

Which fixes the same issue in a slightly different way.

Pretty recent patch though and this was there for ages before that.
Better to have two fixes than none.

Thanks,

J
> ---
>   drivers/iio/industrialio-core.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iio/industrialio-core.c b/drivers/iio/industrialio-core.c
> index ede16aec20fb..5e7a67e53879 100644
> --- a/drivers/iio/industrialio-core.c
> +++ b/drivers/iio/industrialio-core.c
> @@ -715,7 +715,8 @@ static int iio_device_add_info_mask_type(struct iio_dev *indio_dev,
>   {
>   	int i, ret, attrcount = 0;
>
> -	for_each_set_bit(i, infomask, sizeof(infomask)*8) {
> +	for_each_set_bit(i, infomask, min(sizeof(infomask)*8,
> +					  ARRAY_SIZE(iio_chan_info_postfix))) {
>   		ret = __iio_add_chan_devattr(iio_chan_info_postfix[i],
>   					     chan,
>   					     &iio_read_channel_info,
>