On Mon, 9 Jan 2012 09:02:56 -0800, Guenter Roeck wrote:
> On Mon, 2012-01-09 at 11:50 -0500, Jean Delvare wrote:
> > On Sun, 8 Jan 2012 16:21:22 -0800, Guenter Roeck wrote:
> > > Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx>
> > > ---
> > > CHANGES | 1 +
> > > prog/detect/sensors-detect | 6 ++++++
> > > 2 files changed, 7 insertions(+), 0 deletions(-)
> > >
> > > diff --git a/CHANGES b/CHANGES
> > > index 3ae863c..ce04e54 100644
> > > --- a/CHANGES
> > > +++ b/CHANGES
> > > @@ -10,6 +10,7 @@ SVN HEAD
> > > Fix loading of the cpuid module
> > > Make LM73 detection less problematic
> > > Add detection of National Semiconductor LM96163
> > > + Add detection of ITE IT8510E
> > >
> > > 3.3.1 (2011-07-21)
> > > isadump: Add support for word (16-bit) and long (32-bit) reads
> > > diff --git a/prog/detect/sensors-detect b/prog/detect/sensors-detect
> > > index 31a4619..1beb66b 100755
> > > --- a/prog/detect/sensors-detect
> > > +++ b/prog/detect/sensors-detect
> > > @@ -2076,6 +2076,12 @@ use constant FEAT_SMBUS => (1 << 7);
> > >
> > > @superio_ids_ite = (
> > > {
> > > + name => "ITE IT8510E Super IO Sensors",
> > > + driver => "to-be-written",
> > > + devid => 0x8510,
> > > + logdev => 0x04,
> >
> > Are you sure about the logical device? LD 4 is "System Wake-Up"
> > according to superiotool, this doesn't seem to be related to hardware
> > monitoring. I think I understand that for the IT8510E, monitoring is
> > done through an EC, this doesn't seem to be tied to a specific logical
> > device? I admit I'm a bit confused by these new chips still.
> >
> No, I am not sure. May be completely wrong, and is untested. I copied
> that value from one of the other ITE entries; they all use the same
> value for logdev. I don't even know what logdev is used for ;).
Logical devices refer to the different parts of Super-I/O chips, for
example parallel port, serial ports, hardware monitoring, RTC, watchdog
etc. Basically, every function which needs a separate I/O port range
corresponds to a logical device. Each logical device has a separate
configuration space, which can be selected using a global configuration
register.
> We should probably try to get some test coverage berfore committing the
> code.
Yes, please. Reading the datasheet should probably be a good idea as
well, even though it is no fun reading. These chips are so different
from what we support at the moment that we won't go anywhere without a
good amount of reading.
Note that the current superio detection code wants a logical device
number. If we start seeing devices which don't need/have one, we'll
have to revisit the code to make it optional.
--
Jean Delvare
_______________________________________________
lm-sensors mailing list
lm-sensors@xxxxxxxxxxxxxx
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
