RE: [PATCH v1 2/9] pktcdvd: replace sscanf() by kstrtoul()

From: David Laight
Date: Thu Jan 19 2023 - 17:57:34 EST


From: Andy Shevchenko
> Sent: 19 January 2023 22:08
>
> The checkpatch.pl warns: "Prefer kstrto<type> to single variable sscanf".
> Fix the code accordingly.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>
> ---
> drivers/block/pktcdvd.c | 34 ++++++++++++++++++----------------
> 1 file changed, 18 insertions(+), 16 deletions(-)
>
> diff --git a/drivers/block/pktcdvd.c b/drivers/block/pktcdvd.c
> index 0ec8dc8ee5ed..ad4336ae9927 100644
> --- a/drivers/block/pktcdvd.c
> +++ b/drivers/block/pktcdvd.c
> @@ -236,15 +236,16 @@ static ssize_t congestion_off_store(struct device *dev,
> const char *buf, size_t len)
> {
> struct pktcdvd_device *pd = dev_get_drvdata(dev);
> - int val;
> + int val, ret;
>
> - if (sscanf(buf, "%d", &val) == 1) {
> - spin_lock(&pd->lock);
> - pd->write_congestion_off = val;
> - init_write_congestion_marks(&pd->write_congestion_off,
> - &pd->write_congestion_on);
> - spin_unlock(&pd->lock);
> - }
> + ret = kstrtoint(buf, 10, &val);
> + if (ret)
> + return ret;
> +
> + spin_lock(&pd->lock);
> + pd->write_congestion_off = val;
> + init_write_congestion_marks(&pd->write_congestion_off, &pd->write_congestion_on);
> + spin_unlock(&pd->lock);
> return len;
> }

These don't look directly equivalent.
The sscanf() version silently ignores trailing characters.
I think kstrtoint() will generate an error.
Have you actually checked that the caller allows for
an error return.

David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)