Hi Jean,
On Sat, Mar 24, 2012 at 04:37:44AM -0400, Jean Delvare wrote:
> On Fri, 23 Mar 2012 21:18:32 -0700, Guenter Roeck wrote:
> > On Fri, Mar 23, 2012 at 05:17:40PM -0400, Jean Delvare wrote:
> > > On Wed, 7 Mar 2012 20:25:39 -0800, Guenter Roeck wrote:
> > > > else
> > > > return val * 12;
> > > > - } else
> > > > - return val * 16;
> > > > + } else {
> > > > + if (data->in_scaled & (1 << nr))
> > > > + return val * 32;
> > > > + else
> > > > + return val * 16;
> > > > + }
> > > > }
> > >
> > > I think both functions can then be rewritten more efficiently. I'll do
> > > some testing and send a patch later.
> >
> > I know, I was a bit lazy ;).
> >
> > How about the following ?
> >
> > static int adc_lsb(const struct it87_data *data)
> > {
> > int lsb = has_12mv_adc(data) ? 12 : 16;
> > if (data->in_scaled & (1 << nr))
> > lsb <<= 1;
> > return lsb;
> > }
> >
> > and:
> > val = DIV_ROUND_CLOSEST(val, adc_lsb(data));
> >
> > and:
> > return val * adc_lsb(data);
>
> Except that you have to pass nr around as a parameter, but otherwise
> yes, that's what I had in mind.
>
Yes, I noticed that when I tried to compile it. Dumb compiler, should have
known what I meant and auto-fixed it ;).
Thanks,
Guenter
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
