On Thu, 14 Oct 2010 06:20:39 -0700, Guenter Roeck wrote:
> On Thu, Oct 14, 2010 at 05:46:26AM -0400, Jean Delvare wrote:
> [ ... ]
> > + if ((reg_config1 & 0x2A) == 0x00
> > + && (reg_config2 & 0xF8) == 0x00) {
> > + if ((chip_id & 0xFE) == 0x10 /* W83L771AWG/ASG */
> > + && reg_convrate <= 0x08) {
> > + name = "w83l771";
> > + } else
> > + if (chip_id == 0x01 /* W83L771W/G */
> > + && reg_convrate <= 0x09) {
> > + name = "w83l771";
> > + }
>
> If the chips have a different maximum conversion rate, that should probably be reflected
> in .max_convrate as well. Unless you want to ignore it on purpose for simplicity ...
> which is fine with me.
Yes, I want to ignore it on purpose for simplicity. One reason being
that I have a doubt that the old W83L771W/G actually supports a higher
refresh rate than the new W83L771AWG/ASG does. I don't necessarily
trust the datasheet, as it is inconsistent: it says the default
register value is 08h, then later it claims that the default conversion
rate is 16 Hz, but 16 Hz corresponds to a value of 09h in the table.
Additionally, you'll notice that the W83L771W/G conversion rate table
is shifted by 1 compared to all the other compatible chips - which is
quite suspicious.
So I assume that the table is wrong in the datasheet. Only if someone
with a W83L771W/G at hand tests it and comes to the conclusion that the
table in the datasheet is correct, will I update the driver accordingly.
Thanks for the review, BTW :)
--
Jean Delvare
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
