[hdl] manually drive cheby-generated WB read data to 'X' for non-defined addresses. This helps with meeting timing because of smaller and simpler multiplexer logic.

Fixes https://www.ohwr.org/project/fmc-adc-100m14b4cha/issues/3.

Re-aligment was performed using the WR-enabled FMC-ADC and the PPS signal coming out from the WR
switch as signal input on channel #1, as well as external trigger input. The PPS signal was
delivered to the FMC-ADC by means of a 40ns coaxial cable.

External, internal and time triggering well all tested and calibrated using the above setup.

The new scheme (2x BUFIO2) is causing some routing troubles on the SVEC when WR is also present, so
we provide a generic that allows to use the previous clocking scheme as well.

This is done in order to:

a) Stop using an extra PLL for the ADC serdes and follow the clocking scheme proposed in UG382, v1.10, page 32, Figure 1-15.

b) Decouple the LTC2174-specific code, as a first step to a more generalized acquisition core.

c) Not distribute coregen-generated sources.

Also generate the doc.