Hi Christian,
On Mon, 13 Aug 2007 10:39:00 +0200, Christian Hohnstaedt wrote:
> On Sun, Aug 12, 2007 at 01:13:52PM +0200, Jean Delvare wrote:
> > On Tue, 24 Jul 2007 15:43:28 +0200, Christian Hohnstaedt wrote:
> > > Then LM75_TEMP_FROM_REG() should also return a long instead of an int.
> >
> > This doesn't really matter. A long is needed for LM75_TEMP_TO_REG()
> > because that's what simple_strtol() gives, but in the other direction,
> > we use sprintf, which can handle both an int or a long.
>
> I will keep the "static inline int LM75_TEMP_FROM_REG(u16 reg)" ,
> because I don't want to change all the sprintf("%d") to sprintf("%ld").
Agreed.
> What about w83627hf.c
> return sprintf(buf,"%ld\n", \
> (long)LM75_TEMP_FROM_REG(data->reg##_add[nr-2]));
> ?
The cast to long could be removed together with the "l" length modifier
in the format string, independently of the rest of your patch.
> > > Other drivers using LM75_TEMP_TO_REG:
> > >
> > > driver temp.type func
> > > ------------------------------------
> > > asb100.c long strtoul
> > > ds1621.c direct strtoul
> > > w83627e?hf.c u32 strtoul
> > > w83781d.c s32 strtol
> > >
> > > Looks like there is some kind of general inconsistency :-)
> >
> > Indeed. Can you please send a patch fixing all the affected drivers?
>
> I could, but I only can test the LM75 driver.
I can test the ds1621, w83627hf, w83627ehf and w83781d drivers (given
enough time) when you're done.
--
Jean Delvare
