Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

printfing dma_addr_t as %lli requires a cast, since it may be a 32-bit
item, like it is on my computer. I got meaningless numbers on printk.
Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Do not start acquisition if the trigger is not armed
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

In ZIO commit 65b98d18b73dfc41c14b3c5ecac61968f1b94aea the return
value of alloc_block() change. Fix allocation error detection:
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

There is a little window of time where it happens that the
acquisition_done interrupt come after reading the irq_status but before
the irq acknowledgment. The result is that also acquisition_done is
acked

To solve the problem, the irq handler give the irq acknowledgment only
after it handles all active interrupts. Then, it reads again for active
interrupt and it handles them
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

In single-shot mode, the ADC use the DDR memory as a circular buffer
and it continuously writes into the memory until the trigger fire
occurs. When the trigger occur we know where our samples are stored
thanks to the trigger_position register.
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>
Signed-off-by: Michel Arruat <michel.arruat@cern.ch>

Reset pre-samples, post-samples and nshots register on trigger instance
destroy. This allow a software trigger to work with the ADC with a known
configuration. The software trigger will be able to change only the
post-samples register
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

The cset operations init() and exit() are removed from the ZIO framework
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

The library was checking the flags field, but it must check the mask field
to verify if there is a configuration to apply. To avoid situations where
the user is convinced to apply a configuration but he forgots to set
the bit mask, now the library return an error if the user tries to configure
with an empty mask.

Thank to "Michel Arruat <michel.arruat@cern.ch>" who found the problem
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

The channel index is zero-based, so the first channel is "channel 0". The
help string was ambiguous.
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

The program allow the use of flags: --external and --negative-edge, but it
does not apply the correspondent configuration on the trigger.
Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>

Signed-off-by: Federico Vaga <federico.vaga@gmail.com>