[PATCH] IIO: Add basic MXS LRADC driver

[Marek Vasut]

Dear Lars-Peter Clausen,

> > On 07/19/2012 04:23 PM, Marek Vasut wrote:
> > > Dear Lars-Peter Clausen,
> > > 
> > > Sorry for my late reply, got busy with other crazy stuff.
> > > 
> > > [...]
> > > 
> > >>>> Is there any reason not to use the mxs_lradc struct directly has the
> > >>>> iio data?
> > >>> 
> > >>> I explained it below in the patch. I hope we'll eventually support
> > >>> the delay triggers, which will need 4 separate IIO devices. And this
> > >>> is where this structure will be augmented by other components.
> > >> 
> > >> Ok I saw the comment, but it wasn't obvious to me that delay channels
> > >> will require multiple IIO devices. Might make sense to state this
> > >> explicitly.
> > > 
> > > Fixed.
> > > 
> > >>>>> [...]
> > >>>>> 
> > >>>>> +
> > >>>>> +/*
> > >>>>> + * Channel management
> > >>>>> + *
> > >>>>> + * This involves crazy mapping between 8 virtual channels the
> > >>>>> CTRL4 register + * can harbor and 16 channels total this hardware
> > >>>>> supports. + */
> > >>>> 
> > >>>> I suppose this means only a maximum 8 channels can be active at a
> > >>>> time. I've recently posted a patch which makes it easy to implement
> > >>>> such a restriction.
> > >>>> http://www.spinics.net/lists/linux-iio/msg05936.html and
> > >>>> http://www.spinics.net/lists/linux-iio/msg05935.html for an example
> > >>>> validate callback implementation. In your case you'd check for
> > >>>> bitmap_weight(...) <= 8. Those patches are not yet in IIO though.
> > >>> 
> > >>> This is good. When do you expect this to hit linux-next possibly? Or
> > >>> at least linux-iio?
> > >> 
> > >> soon, hopefully.
> > > 
> > > So I checked this, not sure how it'll help me though.
> > 
> > Right now with your driver you can enable any combination of channels. If
> > more than 8 channels are enabled and you start sampling it will fail,
> > since not all channels can be mapped. With those patch you can implement
> > a validate callback and check for bitmap_weight(scan_mask) <= 8. This
> > will ensure that it is not possible to select more than 8 channels at
> > once, which again means that starting sampling can't fail of this.
> 
> Ok, I fixed this one. One last thing that I don't really understand is
> this. I run generic_buffer.c to source samples from the LRADC. Is there
> any way to continuously sample them? Because right now, I get one sample
> and that's it, no more samples happen later on (I can 0 data in subsequent
> read() call).

One more thing I'm curious about. There's another ADC block on the CPU, called 
HSADC (high-speed ADC). It can sample even up to 2Msamples/s. If I were to, say 
-- sample at 100kHz and be able to DMA the results into memory -- is there any 
way to push such results into userland somehow? Or how to operate such fast 
beast?

Best regards,
Marek Vasut