Re: [lm-sensors] Ambient Light sensor for Intersil-ISL29020 device

[Jonathan Cameron <jic23@cam.ac.uk>]

Zhang Rui wrote:
> On Wed, 2009-09-02 at 23:26 +0800, Jonathan Cameron wrote:
>> Jean Delvare wrote:
>>> On Thu, 13 Aug 2009 16:22:29 +0530, Trisal, Kalhan wrote:
>>>> This driver used Hwmon interface ,
>>> No, it doesn't. It registers as a hwmon device but it doesn't implement
>>> _any_ feature listed in Documentation/hwmon/sysfs-interface. Which can
>>> be easily explained: your device is not a hardware monitoring device.
>>>
> I tried to bind the ACPI ALS device driver with the hwmon sysfs I/F at
> the beginning, then stopped for the same reason.
> that's why the generic ALS sysfs class is introduced.
> 
>>>> I don't find the ils29003 registering with hwmon. I still believe
>>>> this should be part of hwmon group.
>>> No, it shouldn't. hwmon is for hardware monitoring. Other sensor types
>>> must find a different home.
>> For info current options I know of:
>>
>> 1) drivers/misc (where the isl290003 is, previously intention was to possibly
>> move this to iio when / if that gets merged)
>>
>> 2) drivers/staging/iio/light (tsl2561) I'm happy with more light sensors
>> in there, though the interface might need some hammering out.  Currently none
>> of them are using any iio specific features so can certainly go elsewhere
>> if people would prefer. In my personal view a lot of the processing currently
>> in the various drivers ought not to be in the kernel, but that's a different
>> matter)  IIO is currently in Greg KH's tree.
>>
>> 3) drivers/als (acpi ambient light sensor - latest version posted to lkml a couple
>> of days ago - now in acpi-testing I think) 
> 
> it's not ACPI ambient light sensor.
> drivers/als/als.c is a new sysfs class (als_class) which provides a
> generic sysfs I/F for all the ALS devices.
> ACPI ALS device driver is just the first user of this class.
> 
>>  This one is a bit different, but perhaps
>> a conversation needs to be opened with them to see if the requirements overlap
>> sufficiently to use a shared framework.)
> 
> right.
> As I'm not familiar with other ALS devices, I just introduced two sysfs
> I/F which IMO are needed for all the ALS devices.
Unfortunately things aren't that simple, I'll give more details in initial lkml
posting.

> We can add any attributes which are generic and useful. :)
> 
> thanks,
> rui
Lets move this discussion to a new thread on LKML to pick up anyone else who is
interested in the discussion. I'll try and draw up a summary of previous discussions
and will cc everyone in this thread (yell if you'd rather I didn't!)

Jonathan

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors