On Fri, Jan 13, 2012 at 03:47:10AM -0500, Jean Delvare wrote: > On Thu, 12 Jan 2012 14:23:35 -0800, Guenter Roeck wrote: > > On LM96163, the external temperature sensor type is configurable to > > either a thermal diode or a 3904 transistor. The chip reports a wrong > > temperature if misconfigured. Add writable attribute to support it. > > > > Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx> > > --- > > v2: > > - snprintf -> sprintf > > - set temp sensor type to 1(CPU diode) instead of 3 (thermal diode) if trutherm > > is enabled > > - return error code from kstrtoul() > > - When configuring the diode type, only set trutherm enable bit > > - When detecting configured diode type, only check trutherm enable bit > > - Document that configuring the thermal sensor type is supported for LM96163 > > I don't see such documentation addition. Forgot to add the file to the > patch? > That I did. > > drivers/hwmon/lm63.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++ > > 1 files changed, 50 insertions(+), 0 deletions(-) > > > > diff --git a/drivers/hwmon/lm63.c b/drivers/hwmon/lm63.c > > (...) > > + reg = i2c_smbus_read_byte_data(client, LM96163_REG_TRUTHERM) & ~0x02; > > + i2c_smbus_write_byte_data(client, LM96163_REG_TRUTHERM, > > + reg | data->trutherm ? 0x02 : 0x00); > > This lacks parentheses to do the right thing. | has higher precedence > than ?:. It works by chance because reg has value 0 in practice. > Sigh. You'll get another one in a minute. Guenter _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
