On Wed, 21 Oct 2009 18:01:37 -0700, Darrick J. Wong wrote: > acpi-power-meter: Don't leak ACPI error codes to userspace > > If the ACPI methods return an error code, we must return -EINVAL to userspace > to flag the error. Right now we pass the (positive) number right through, > which causes echo to keep writing bogus values. > > Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx> Acked-by: Jean Delvare <khali@xxxxxxxxxxxx> > --- > > drivers/acpi/power_meter.c | 6 +++++- > 1 files changed, 5 insertions(+), 1 deletions(-) > > > diff --git a/drivers/acpi/power_meter.c b/drivers/acpi/power_meter.c > index e6bfd77..2ef7030 100644 > --- a/drivers/acpi/power_meter.c > +++ b/drivers/acpi/power_meter.c > @@ -294,7 +294,11 @@ static int set_acpi_trip(struct acpi_power_meter_resource *resource) > return -EINVAL; > } > > - return data; > + /* _PTP returns 0 on success, nonzero otherwise */ > + if (data) > + return -EINVAL; > + > + return 0; > } > > static ssize_t set_trip(struct device *dev, struct device_attribute *devattr, -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
