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>

Recent compilers detected that we were not checking the return value
of some functions that we should have checked.

``asprintf`` may fail durint allocation
``read`` may fail in that context because of a ZIO bug

Both are unlikely to happen, but it is better to be informed when they do
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
Signed-off-by: Dimitris Lampridis <dimitris.lampridis@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>
Signed-off-by: Milosz Malczak <milosz.malczak@cern.ch>

Fixed hasattr - the string has to be passed as an argument

Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
Signed-off-by: Milosz Malczak <milosz.malczak@cern.ch>

Import form current module

This class serves different types of ADC, so it can't open only
the FMC-ADC-100M
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

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

Signed-off-by: Milosz Malczak <milosz.malczak@cern.ch>
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

This is not really necessary for C/C++ programs, but it is
foundamental for Python programs who do not have access to
the header files.
Signed-off-by: Milosz Malczak <milosz.malczak@cern.ch>
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Extern inline functions moved from adc-lib.h to lib.c to make them visible in .so file

This is not really necessary for C/C++ programs, but it is
foundamental for Python programs who do not have access to
the header files.
Signed-off-by: Milosz Malczak <milosz.malczak@cern.ch>
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

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

The way this patch has been written is a bit inconsistent with other
parts, but this is the way to go: values should be clear for the user
and not to be interpreted according to the board type. So, for timer
trigger we talk about seconds and nanoseconds, instead of seconds and
coarse.

Another thing to be noted here is that the FMCADC100M has 64bit seconds
while the library supports only 32bit. I leave it like this before to
get something working as soon as possible.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

Select the channel to acquire instead of defaulting to 'interleave'
using a completely wrong flag
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>

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>