The driver exports a sysfs attrubte to change the coalescing timeout. A new
library funciton can be used to set/get this value; and the test program allows
the configuration before an acquisition.

Keep in mind that, for FIFOs the configuration is global (users are warned on
dmesg) while for DMA it is per-channel. I though it was better to have a common
interface for both. Users are warnd in the documentation as well. In the long
run, perhaps also the FIFO can have its own register.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

It is not an IRQ coalescing size, but the size that triggers the dump
from the FIFO to the DDR. so, it is internal stuff that does not need
to get exposed to the user
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>

This way the name is properly terminated. Perhaps there is a bug in ZIO
that expect a \0 and it does not check any upper limit.
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>

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

They should be implemented in ZIO
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

the HW_BUSY flag must be cleared before doing data_done in order to hold
only during DMA.
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 happened that ZIO was taking some extra characters after the name.
to play safe I put a terminator there
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>

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>

This also fix a bug. The bug was that I was coping data
at the wrong offset withing the ZIO block. It starts always
at offset 0, but I was shifing by "start" bytes
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

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

For the time being is used only to change the BURST size using POST-SAMPLEs

This solves also a bug with the ZIO trigger "user" which does arm for us, but
this is not exactly what we want because we need to arm only when we know the
number of samples to acquire
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>

Not reallyused at this point
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

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

This allows bigger read from the driver, which reduces the number of syscall
and make the system faster
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>

We need to change the nsample according to the current number of timestamps
available
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

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