On Thu, 2011-07-07 at 05:38 -0400, Jean Delvare wrote:
> On Tue, 5 Jul 2011 07:36:32 -0700, Guenter Roeck wrote:
> > Negative temperatures were returned in degrees C instead of milli-Degrees C.
> > Fix by multiplying with 1000.
> >
> > Signed-off-by: Guenter Roeck <guenter.roeck@xxxxxxxxxxxx>
> > ---
> > drivers/hwmon/lm95241.c | 2 +-
> > 1 files changed, 1 insertions(+), 1 deletions(-)
> >
> > diff --git a/drivers/hwmon/lm95241.c b/drivers/hwmon/lm95241.c
> > index f2cfecf..d1bbdf8 100644
> > --- a/drivers/hwmon/lm95241.c
> > +++ b/drivers/hwmon/lm95241.c
> > @@ -101,7 +101,7 @@ struct lm95241_data {
> > static int TempFromReg(u8 val_h, u8 val_l)
> > {
> > if (val_h & 0x80)
> > - return val_h - 0x100;
> > + return (val_h - 0x100) * 1000;
> > return val_h * 1000 + val_l * 1000 / 256;
> > }
> >
>
> The fix is needed but is it sufficient? How comes that positive
> temperatures would be returned with a sub-degree resolution and negative
> ones with 1°C resolution only? I don't see any such limitation in the
> datasheet.
>
> Also, negative values are really only 2's complement format so a simple
> cast should do (untested):
>
> s16 val_hl = (val_h << 8) | val_l;
> return val_hl * 1000 / 256;
>
Actually, it is getting more complicated than that. Like the LM95245,
the LM95241 supports both unsigned and signed temperature formats, and
the driver configures unsigned mode when loaded - meaning there are
actually no negative temperature values to report by default. I'll have
to change the code to take the chip configuration into account when
doing a conversion, similar to the LM95245 code.
Guenter
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
