[RFC] Moving magnetometer HMC6352 into IIO subsystem

[RFC] Moving magnetometer HMC6352 into IIO subsystem

[Jonathan Kunkee]

Hello all,

Should the HMC6352 digital compass driver (drivers/misc/hmc6352.c)
be moved under the IIO subsystem (drivers/staging/iio/magnetometer)?

Magnetometers act as ADCs and can have high sampling rates and
complex internals, so I think that according to
drivers/staging/iio/Documentation/overview.txt they should almost always
land in IIO land.

I would be willing to work on the change as I want to submit a similar
driver for the HMC6343 magnetometer/accelerometer as well.

Thanks!
Jon Kunkee

Re: [RFC] Moving magnetometer HMC6352 into IIO subsystem

[Alan Cox]

> Should the HMC6352 digital compass driver (drivers/misc/hmc6352.c)
> be moved under the IIO subsystem (drivers/staging/iio/magnetometer)?

No - that would be a regression .. but read on

> Magnetometers act as ADCs and can have high sampling rates and
> complex internals, so I think that according to
> drivers/staging/iio/Documentation/overview.txt they should almost always
> land in IIO land.

IIO is a staging project that may never go anywhere - some of us are
using magnetometers in product so this alone isn't an answer that works
right now.

> I would be willing to work on the change as I want to submit a similar
> driver for the HMC6343 magnetometer/accelerometer as well.

It would be great if it got added into the IIO subystem *as well*, and in
time when and if IIO goes out of staging it would then eventually replace
the existing drivers.

Alan

Re: [RFC] Moving magnetometer HMC6352 into IIO subsystem

[Jonathan Kunkee]

>> Should the HMC6352 digital compass driver be moved [...]?
>
> No - that would be a regression .. but read on

>> [...] they should almost always land in IIO land.
>
> IIO is a staging project that may never go anywhere - some of us are
> using magnetometers in product so this alone isn't an answer that works
> right now.

In my reading I failed to catch the how transient 'staging' is--I'll
keep it mind, as I'm hoping my driver will not be so transient.

>> I would be willing to work on the change as I want to submit a similar
>> driver for the HMC6343 magnetometer/accelerometer as well.
>
> It would be great if it got added into the IIO subystem *as well*, and in
> time when and if IIO goes out of staging it would then eventually replace
> the existing drivers.

This would then necessitate maintaining both chunks of code, so I think
for now I will develop my module in drivers/misc. After it has stabilized,
I can add both modules to IIO as well with this in mind.

Thanks!
--Jon Kunkee

Re: [RFC] Moving magnetometer HMC6352 into IIO subsystem

[Jonathan Cameron]

On 08/16/11 00:22, Jonathan Kunkee wrote:
>>> Should the HMC6352 digital compass driver be moved [...]?
>>
>> No - that would be a regression .. but read on
> 
>>> [...] they should almost always land in IIO land.
>>
>> IIO is a staging project that may never go anywhere - some of us are
>> using magnetometers in product so this alone isn't an answer that works
>> right now.
> 
> In my reading I failed to catch the how transient 'staging' is--I'll
> keep it mind, as I'm hoping my driver will not be so transient.
> 
>>> I would be willing to work on the change as I want to submit a similar
>>> driver for the HMC6343 magnetometer/accelerometer as well.
>>
>> It would be great if it got added into the IIO subystem *as well*, and in
>> time when and if IIO goes out of staging it would then eventually replace
>> the existing drivers.
> 
> This would then necessitate maintaining both chunks of code, so I think
> for now I will develop my module in drivers/misc. After it has stabilized,
> I can add both modules to IIO as well with this in mind.
> 
Certainly a valid option right now.   Moving things over tends to be
straight forward anyway.

IIO is hoping to make first steps out of staging shortly, but it will be a
good while before anywhere near everything moves out into the main tree
(lots of code and some of it is probably going to be controversial!)

Jonathan