Hi Markus,
On 09/01/17 21:44, SF Markus Elfring wrote:
> From: Markus Elfring
> Date: Fri, 1 Sep 2017 21:08:38 +0200
>
> * Adjust a jump target so that a null pointer will not be passed to a call
> of the function "kfree".
>
> * Move this function call into an if branch.
>
> Signed-off-by: Markus Elfring
> ---
> drivers/media/dvb-frontends/sp2.c | 13 +++--
> 1 file changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/media/dvb-frontends/sp2.c
> b/drivers/media/dvb-frontends/sp2.c
> index dd556012ceb6..b2a7a54174ae 100644
> --- a/drivers/media/dvb-frontends/sp2.c
> +++ b/drivers/media/dvb-frontends/sp2.c
> @@ -384,7 +384,7 @@ static int sp2_probe(struct i2c_client *client,
> s = kzalloc(sizeof(*s), GFP_KERNEL);
> if (!s) {
> ret = -ENOMEM;
> - goto err;
> + goto report_failure;
> }
>
> s->client = client;
> @@ -395,15 +395,16 @@ static int sp2_probe(struct i2c_client *client,
> i2c_set_clientdata(client, s);
>
> ret = sp2_init(s);
> - if (ret)
> - goto err;
> + if (ret) {
> + kfree(s);
> + goto report_failure;
> + }
>
> dev_info(&s->client->dev, "CIMaX SP2 successfully attached\n");
> return 0;
> -err:
> - dev_dbg(&client->dev, "init failed=%d\n", ret);
> - kfree(s);
>
> +report_failure:
> + dev_dbg(&client->dev, "init failed=%d\n", ret);
> return ret;
> }
>
>
It's perfectly fine to call kfree() with a NULL pointer, and I don't think
that this patch makes the code more readable, so I'm dropping this patch.
Regards,
Hans
