On Fri, Jun 22, 2018 at 02:19:48PM +0300, Dan Carpenter wrote: > If info->pdata is NULL then we would oops on the next line. And we can > flip the "ret" test around and give up if a failure has already occured. > > Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> > > diff --git a/drivers/media/i2c/sr030pc30.c b/drivers/media/i2c/sr030pc30.c > index 2a4882cddc51..4ebd00198d34 100644 > --- a/drivers/media/i2c/sr030pc30.c > +++ b/drivers/media/i2c/sr030pc30.c > @@ -569,8 +569,8 @@ static int sr030pc30_base_config(struct v4l2_subdev *sd) > if (!ret) > ret = sr030pc30_pwr_ctrl(sd, false, false); > > - if (!ret && !info->pdata) > - return ret; > + if (ret || !info->pdata) > + return -EIO; There seem to be a couple of other places checking pdata is not NULL, including the driver's probe() function; doing the same here seems redundant. Just checking ret and failing if it's non-zero should suffice: if (ret) return ret; Let me know if you'd like to respin; I can do that as well. > > expmin = EXPOS_MIN_MS * info->pdata->clk_rate / (8 * 1000); > expmax = EXPOS_MAX_MS * info->pdata->clk_rate / (8 * 1000); -- Sakari Ailus e-mail: sakari.ailus@xxxxxx
