On Thu, 12 Aug 2010 10:40:12 -0700, Guenter Roeck wrote:
> On Wed, 2010-08-11 at 11:14 -0400, Jean Delvare wrote:
> > The sysfs file removal code is the same in the probe error path and in
> > the remove function, so move it to a separate function to avoid code
> > duplication.
> >
> > Signed-off-by: Jean Delvare <khali@xxxxxxxxxxxx>
>
> Acked-by: Guenter Roeck <guenter.roeck@xxxxxxxxxxxx>
>
> Another minor question below.
>
> > ---
> > drivers/hwmon/pc87427.c | 58 +++++++++++++++++++++--------------------------
> > 1 file changed, 26 insertions(+), 32 deletions(-)
> >
> > --- linux-2.6.35-rc4.orig/drivers/hwmon/pc87427.c 2010-07-11 10:13:07.000000000 +0200
> > +++ linux-2.6.35-rc4/drivers/hwmon/pc87427.c 2010-07-11 10:17:44.000000000 +0200
> > @@ -1045,6 +1045,29 @@ static void __devinit pc87427_init_devic
> > }
> > }
> >
> > +static void pc87427_remove_files(struct device *dev)
> > +{
> > + struct pc87427_data *data = dev_get_drvdata(dev);
> > + int i;
> > +
> > + device_remove_file(dev, &dev_attr_name);
>
> device_remove_file() is not called in the error path. Does that matter ?
No it doesn't matter. It's always OK to remove files that do not exist.
In fact, I could even remove all the tests on data->*_enabled below, it
would still work fine.
>
> > + for (i = 0; i < 8; i++) {
> > + if (!(data->fan_enabled & (1 << i)))
> > + continue;
> > + sysfs_remove_group(&dev->kobj, &pc87427_group_fan[i]);
> > + }
> > + for (i = 0; i < 4; i++) {
> > + if (!(data->pwm_enabled & (1 << i)))
> > + continue;
> > + sysfs_remove_group(&dev->kobj, &pc87427_group_pwm[i]);
> > + }
> > + for (i = 0; i < 6; i++) {
> > + if (!(data->temp_enabled & (1 << i)))
> > + continue;
> > + sysfs_remove_group(&dev->kobj, &pc87427_group_temp[i]);
> > + }
> > +}
--
Jean Delvare
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
