[UPSTREAM] Attempt at improving ADC access

[Werner Almesberger]

Balaji Rao wrote:
> Ah, now I remember! From the proposed sync callback, we wouldn't know
> which completion to complete as we wouldn't be passed the adc_request
> object. We could very well look at the queue head, but that would not be
> a clean way to do it. What do you say ?

How about passing the request ? Or a void pointer that's passed to
the callback function, which then does with it whatever it wants,
i.e., the usual

	..., void (*fn)(void *arg), void *arg, ...

idiom.

- Werner