Decimation and interpolation parameter removal

[tfcollins]

I think we may want to remove these parameters from the block masks from FMComms and Pluto since they can cause aliasing based on how they are implemented.

gr-iio/lib/device_sink_impl.cc

Line 162 in 2997f46

ptrdiff_t buf_step = iio_buffer_step(buf) * (interpolation + 1);

gr-iio/lib/device_source_impl.cc

Line 237 in 2997f46

ptrdiff_t buf_step = iio_buffer_step(buf) * (decimation + 1);

Is there a strong use case for them?

[mhennerich]

It was introduced with this commit:
8b9eacc

I don't remember the reason - but simply skipping samples in SW is not going to make a strong use case. This can be better done outside the block.

At some point we were talking about adding support for the Pluto Int/Dec by 8 HW filter.
But this is something different.

[lclausen-adi]

It was done to safe CPU resources so some of the designs were capable of running on the ARM processor on the ZYNQ.

[tfcollins]

I'm fine with leaving it for the iio_device blocks, but having it in the FMComms/Pluto blocks is just asking for trouble. I think this was more of a concern before the baseband_auto_rate function existed, but now we can get down to much lower rates. Especially if we add the hooks to control Pluto's FPGA filters

[tfcollins]

Are the FMComms/Pluto blocks required for scopy? Doing a quick search of the repo I couldn't find anything.

[lclausen-adi]

no