[lm-sensors] [PATCH RESEND] lm-sensors: add dme1737 support

Juerg Haefliger juergh at gmail.com
Tue Mar 27 17:35:13 CEST 2007


Hi Jean,


> > +     name => "SMSC DME1737 Super IO",
> > +     # The DME1737 shows up twice in this list because it can return both
>
> I think you mean "either", not "both". I doubt a given chip would
> return two different values?

Yes, sure.


> > +#   0x73: Read-only test register (4 test bits)
> > +#   0x8A: Read-only test register (7 test bits)
> >  sub dme1737_detect
> >  {
> >    my ($file, $addr) = @_;
> > -  return unless i2c_smbus_read_byte_data($file, 0x3E) == 0x55
> > +  return unless i2c_smbus_read_byte_data($file, 0x3E) == 0x5c
> >             and (i2c_smbus_read_byte_data($file, 0x3F) & 0xF8) == 0x88
> > -           and (i2c_smbus_read_byte_data($file, 0x40) & 0xC4) == 0x04
> > -           and (i2c_smbus_read_byte_data($file, 0x42) & 0x02) == 0x00
> > -           and (i2c_smbus_read_byte_data($file, 0x43) & 0xC0) == 0x00;
> > +           and (i2c_smbus_read_byte_data($file, 0x73) & 0x0F) == 0x09
> > +           and (i2c_smbus_read_byte_data($file, 0x8A) & 0x7F) == 0x4D;
> >    return ($addr == 0x2e ? 6 : 5);
> >  }
>
> I'm fine with using register 0x8A, even though I don't think we ever
> used a test register for identification purposes. But I am skeptical
> about 0x73, it's essentially not documented in the datasheet I have,
> but the only think they say is that it's read/write, that doesn't sound
> right for identification purposes. Assuming that your datasheet is more
> complete than mine, you'll know better, but please confirm it's really
> that register you want to use. If not, I still think that register 0x40
> is good for identification purposes.

Yes, 0x73 is the one I wanted. My datasheet is at rev 0.4
If you don't feel comfortable with the new registers I don't have a
problem going back to the original register set. I just thought
something a little bit more unique and meaningful makes sense.

Thanks
...juerg



> Other than that, your patch looks good and I am willing to apply it.
>
> Thanks,
> --
> Jean Delvare
>




More information about the lm-sensors mailing list