FFT Zero Padding

The fairly new guy in the RIL department (which I just was moved to) has a master's in signal processing.  Since I didn't have an even number of samples for the Fourier transform, I was using the DFT, but he claimed that zero padding to power-of-2 FFT lengths is legit.  Apparently it is.  However, there are bin shift concerns [ni.com] when doing so because the bin number and size are determined by the number of samples used.