In a scenario where PCM gets upsampled to DSD128 and DSD64 content gets left alone. Can I just use the SDM gain setting to deal with intersample clipping instead of enabling the global headroom management?
Sure.
The only issue is–then DSD content that is not being upsampled will be louder than PCM content.
If you don’t use volume leveling, that’s no big deal–tracks vary quite a bit in volume naturally. If you do use volume leveling, then it’s probably better to manage clipping using the headroom mgmt section.