ZIO does not support start and stop, so it wrong to provide a common
ZIO feature about this.

Moreover, its usage in the FMC-ADC100M14B4CH device is buggy because
it forces a flush while the acquisition is running; instead, we want
to flush *before* starting the acquisition.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Before this patch, flushing was a special option available during
opening or starting. Users may want to flush independently at any time.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Actually, GCC is wrong, but we like when it is happy
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

The API change is quite small, but the specific code that handles it for
the FMC-ADC-100M is big.

FMC-ADC-100M
- it computes zero-offset for all channels and for all possibile ranges
- split manual and automatic mode for better readability and maintenance
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Get the total number of samples directly from the applied configuration
instead of asking the user to provide this value.

It is true that this can be redundant, but it is more reliable
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

When the ZERO flag is set, the function will remove any zero-offset applied.
This is used to reset the auto-clear status
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

With this flag we can detect if any of the configuration parameter failed.
But we will not know which. It does not change much from the current
situation but, it if useful when applying many configuration in a raw.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Also during manual configuration offset-zero must be zero.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

It adds support for the offset auto-clear
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

It adds a new board operation for auto-clear any offset. If a board
does not support this feature in its hardware or driver, then it can
use the software implementation. The software implementation does an
acquisition on disconnected channel and compute the signal average;
we assume that this average is a constant noise, then we apply this
value to the DAC to clear it on the ADC input.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

I am going to introduce the offset auto-clear feature.
The purpre is to compensate any external (constant) offset
on a given channel. It may happen because of temperature or other
external factors.

Within this commit I present the API and the documentation of it.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

- Better sphinx+breathe usage
- Improve the actual content
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

If there are known issues with an acquisition, the library tries to
compensate them by manipulating, somehow, an acquired buffer by
using the function adc_buffer_fixup(). To keep track of these manipulation
I introduced a private pointer that can be used by the library to specify
how the buffer has been manipulated.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>