Re: [PATCHv2 3/3] IIO: hid-sensor-magn-3d: Add in support for True/Magnetic North HID usages

[Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>]

On 06/04/2014 08:23 AM, Reyad Attiyat wrote:
> Hey Srinivas,
>
> On Tue, Jun 3, 2014 at 12:43 PM, Srinivas Pandruvada
> <srinivas.pandruvada@linux.intel.com> wrote:
>
>>
>> May be we should have a field in const struct iio_chan_spec, so that we
>> can dynamically enable disable channels. In this way we can statically
>> define channels, but can be enabled/disabled dynamically.
>>
> Would this require changing iio subsystem to create sysfs entries only
> when enabled? Would we need to add functions to disable and enable
> later on?

This is just a thought. You don't have to change it. I am sure Jonathan 
will have some opinion this.

>>
>> I think we need to present angle in radians. Are you basing change present
>> in linux-next? This will automatically do in this function.
>>
> I'll look into this. What function should I use to make the iio chanel
> to report radians?
I think it will work if the existing sequence is maintained
  st->scale_precision = hid_sensor_format_scale(
                                 HID_USAGE_SENSOR_COMPASS_3D,
                                 &st->magn[CHANNEL_SCAN_INDEX_X],
                                 &st->scale_pre_decml, 
&st->scale_post_decml);

So as long as you call this function, the scale value to user space will
be returned correctly.

>>
>>
>> I don't see kfree. Try devm_kcalloc?
>>
> I changed kmemdup to kcalloc so there is still a kfree in the exsiting
> code. I can change this to devm_kcalloc but only if we don't go with
> static channels that are enabled as found.
>
Since you are changing this part, devm_ calls are preferred, I think.

Thanks,
Srinivas


> Thanks,
> Reyad Attiyat
>