1SDV is a combination of VV and VH polarizations and 1SDH one of HH and HV, hence, you have indeed different polarizations which throw the error. I would be surprised if that would be considered to be a bug because one would expect phase coherence to require the same polarization.
The Back-Geocoding operator is performing the check for same polarization and it is by design.
Back-Geocoding on its own does not require same polarization. However, since the processing step after Back-Geocoding is almost always interferogram (which requires same polarization) and Back-Geocoding is the most time consuming operator, it is better to alert the user to the same polarization requirement earlier than later.