The sensor pattern doesn't directly relate to chroma sub-sampling on a codec. The two are different things, and to co-mingle them is misleading. What you need to do is measure the output. If you have a 4:2:2 output codec, you shouldn't be able to measure chroma resolution to be more than half, but it still could measure less, just as luma could very well measure less than full resolution. It's still a 4:2:2 codec though. Similarly, on a 4:4:4 system (like the F35) you could measure less on the resolution than you'd expect even on a 4:4:4 recording.
I don't understand why s-log costs more. It's math is little different from REDLog and ViperLog, or Panalog. I can see why a 4:4:4 output costs more, but in this case it sounds to me like a firmware turn-on rather than a new output block on the camera.